diff options
Diffstat (limited to 'target')
27 files changed, 64425 insertions, 19 deletions
diff --git a/target/alix1c/device.mk b/target/alix1c/device.mk index fb0f58477..fc3479ab4 100644 --- a/target/alix1c/device.mk +++ b/target/alix1c/device.mk @@ -1,6 +1,6 @@ ARCH:= x86 CPU_ARCH:= i586 -KERNEL_VERSION:= 2.6.30.5 +KERNEL_VERSION:= 2.6.31.4 KERNEL_RELEASE:= 1 KERNEL_MD5SUM:= be9c3a697a54ac099c910d068ff0dc03 TARGET_OPTIMIZATION:= -Os -pipe diff --git a/target/foxboard/device.mk b/target/foxboard/device.mk index e9be810b6..b2eb19f57 100644 --- a/target/foxboard/device.mk +++ b/target/foxboard/device.mk @@ -1,6 +1,6 @@ ARCH:= cris CPU_ARCH:= cris -KERNEL_VERSION:= 2.6.30.5 +KERNEL_VERSION:= 2.6.31.4 KERNEL_RELEASE:= 1 KERNEL_MD5SUM:= be9c3a697a54ac099c910d068ff0dc03 TARGET_OPTIMIZATION:= -Os -pipe -fno-peephole2 diff --git a/target/g1/device.mk b/target/g1/device.mk index 1d5c2cd72..933c74e18 100644 --- a/target/g1/device.mk +++ b/target/g1/device.mk @@ -1,6 +1,6 @@ ARCH:= arm CPU_ARCH:= armel -KERNEL_VERSION:= 2.6.30.5 +KERNEL_VERSION:= 2.6.31.4 KERNEL_RELEASE:= 1 KERNEL_MD5SUM:= be9c3a697a54ac099c910d068ff0dc03 TARGET_OPTIMIZATION:= -Os -pipe diff --git a/target/lemote/device.mk b/target/lemote/device.mk index 52cc097fc..a9037d8ed 100644 --- a/target/lemote/device.mk +++ b/target/lemote/device.mk @@ -1,6 +1,6 @@ ARCH:= mips CPU_ARCH:= mips64el -KERNEL_VERSION:= 2.6.30.5 +KERNEL_VERSION:= 2.6.31.4 KERNEL_RELEASE:= 1 KERNEL_MD5SUM:= be9c3a697a54ac099c910d068ff0dc03 TARGET_OPTIMIZATION:= -Os -pipe diff --git a/target/linux/patches/2.6.31.4/aufs2.patch b/target/linux/patches/2.6.31.4/aufs2.patch new file mode 100644 index 000000000..7a6d3bac6 --- /dev/null +++ b/target/linux/patches/2.6.31.4/aufs2.patch @@ -0,0 +1,24453 @@ +diff -Nur linux-2.6.31.4.orig/Documentation/ABI/testing/debugfs-aufs linux-2.6.31.4/Documentation/ABI/testing/debugfs-aufs +--- linux-2.6.31.4.orig/Documentation/ABI/testing/debugfs-aufs 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/Documentation/ABI/testing/debugfs-aufs 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,40 @@ ++What: /debug/aufs/si_<id>/ ++Date: March 2009 ++Contact: J. R. Okajima <hooanon05@yahoo.co.jp> ++Description: ++ Under /debug/aufs, a directory named si_<id> is created ++ per aufs mount, where <id> is a unique id generated ++ internally. ++ ++What: /debug/aufs/si_<id>/xib ++Date: March 2009 ++Contact: J. R. Okajima <hooanon05@yahoo.co.jp> ++Description: ++ It shows the consumed blocks by xib (External Inode Number ++ Bitmap), its block size and file size. ++ When the aufs mount option 'noxino' is specified, it ++ will be empty. About XINO files, see ++ Documentation/filesystems/aufs/aufs.5 in detail. ++ ++What: /debug/aufs/si_<id>/xino0, xino1 ... xinoN ++Date: March 2009 ++Contact: J. R. Okajima <hooanon05@yahoo.co.jp> ++Description: ++ It shows the consumed blocks by xino (External Inode Number ++ Translation Table), its link count, block size and file ++ size. ++ When the aufs mount option 'noxino' is specified, it ++ will be empty. About XINO files, see ++ Documentation/filesystems/aufs/aufs.5 in detail. ++ ++What: /debug/aufs/si_<id>/xigen ++Date: March 2009 ++Contact: J. R. Okajima <hooanon05@yahoo.co.jp> ++Description: ++ It shows the consumed blocks by xigen (External Inode ++ Generation Table), its block size and file size. ++ If CONFIG_AUFS_EXPORT is disabled, this entry will not ++ be created. ++ When the aufs mount option 'noxino' is specified, it ++ will be empty. About XINO files, see ++ Documentation/filesystems/aufs/aufs.5 in detail. +diff -Nur linux-2.6.31.4.orig/Documentation/ABI/testing/sysfs-aufs linux-2.6.31.4/Documentation/ABI/testing/sysfs-aufs +--- linux-2.6.31.4.orig/Documentation/ABI/testing/sysfs-aufs 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/Documentation/ABI/testing/sysfs-aufs 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,25 @@ ++What: /sys/fs/aufs/si_<id>/ ++Date: March 2009 ++Contact: J. R. Okajima <hooanon05@yahoo.co.jp> ++Description: ++ Under /sys/fs/aufs, a directory named si_<id> is created ++ per aufs mount, where <id> is a unique id generated ++ internally. ++ ++What: /sys/fs/aufs/si_<id>/br0, br1 ... brN ++Date: March 2009 ++Contact: J. R. Okajima <hooanon05@yahoo.co.jp> ++Description: ++ It shows the abolute path of a member directory (which ++ is called branch) in aufs, and its permission. ++ ++What: /sys/fs/aufs/si_<id>/xi_path ++Date: March 2009 ++Contact: J. R. Okajima <hooanon05@yahoo.co.jp> ++Description: ++ It shows the abolute path of XINO (External Inode Number ++ Bitmap, Translation Table and Generation Table) file ++ even if it is the default path. ++ When the aufs mount option 'noxino' is specified, it ++ will be empty. About XINO files, see ++ Documentation/filesystems/aufs/aufs.5 in detail. +diff -Nur linux-2.6.31.4.orig/fs/aufs/aufs.h linux-2.6.31.4/fs/aufs/aufs.h +--- linux-2.6.31.4.orig/fs/aufs/aufs.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/aufs.h 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,51 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * all header files ++ */ ++ ++#ifndef __AUFS_H__ ++#define __AUFS_H__ ++ ++#ifdef __KERNEL__ ++ ++#include "debug.h" ++ ++#include "branch.h" ++#include "cpup.h" ++#include "dcsub.h" ++#include "dbgaufs.h" ++#include "dentry.h" ++#include "dir.h" ++#include "file.h" ++#include "fstype.h" ++#include "inode.h" ++#include "loop.h" ++#include "module.h" ++#include "opts.h" ++#include "rwsem.h" ++#include "spl.h" ++#include "super.h" ++#include "sysaufs.h" ++#include "vfsub.h" ++#include "whout.h" ++#include "wkq.h" ++ ++#endif /* __KERNEL__ */ ++#endif /* __AUFS_H__ */ +diff -Nur linux-2.6.31.4.orig/fs/aufs/branch.c linux-2.6.31.4/fs/aufs/branch.c +--- linux-2.6.31.4.orig/fs/aufs/branch.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/branch.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,974 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * branch management ++ */ ++ ++#include <linux/file.h> ++#include "aufs.h" ++ ++/* ++ * free a single branch ++ */ ++static void au_br_do_free(struct au_branch *br) ++{ ++ int i; ++ struct au_wbr *wbr; ++ ++ if (br->br_xino.xi_file) ++ fput(br->br_xino.xi_file); ++ mutex_destroy(&br->br_xino.xi_nondir_mtx); ++ ++ AuDebugOn(atomic_read(&br->br_count)); ++ ++ wbr = br->br_wbr; ++ if (wbr) { ++ for (i = 0; i < AuBrWh_Last; i++) ++ dput(wbr->wbr_wh[i]); ++ AuDebugOn(atomic_read(&wbr->wbr_wh_running)); ++ AuRwDestroy(&wbr->wbr_wh_rwsem); ++ } ++ ++ /* some filesystems acquire extra lock */ ++ lockdep_off(); ++ mntput(br->br_mnt); ++ lockdep_on(); ++ ++ kfree(wbr); ++ kfree(br); ++} ++ ++/* ++ * frees all branches ++ */ ++void au_br_free(struct au_sbinfo *sbinfo) ++{ ++ aufs_bindex_t bmax; ++ struct au_branch **br; ++ ++ AuRwMustWriteLock(&sbinfo->si_rwsem); ++ ++ bmax = sbinfo->si_bend + 1; ++ br = sbinfo->si_branch; ++ while (bmax--) ++ au_br_do_free(*br++); ++} ++ ++/* ++ * find the index of a branch which is specified by @br_id. ++ */ ++int au_br_index(struct super_block *sb, aufs_bindex_t br_id) ++{ ++ aufs_bindex_t bindex, bend; ++ ++ bend = au_sbend(sb); ++ for (bindex = 0; bindex <= bend; bindex++) ++ if (au_sbr_id(sb, bindex) == br_id) ++ return bindex; ++ return -1; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* ++ * add a branch ++ */ ++ ++static int test_overlap(struct super_block *sb, struct dentry *h_d1, ++ struct dentry *h_d2) ++{ ++ if (unlikely(h_d1 == h_d2)) ++ return 1; ++ return !!au_test_subdir(h_d1, h_d2) ++ || !!au_test_subdir(h_d2, h_d1) ++ || au_test_loopback_overlap(sb, h_d1, h_d2) ++ || au_test_loopback_overlap(sb, h_d2, h_d1); ++} ++ ++/* ++ * returns a newly allocated branch. @new_nbranch is a number of branches ++ * after adding a branch. ++ */ ++static struct au_branch *au_br_alloc(struct super_block *sb, int new_nbranch, ++ int perm) ++{ ++ struct au_branch *add_branch; ++ struct dentry *root; ++ ++ root = sb->s_root; ++ add_branch = kmalloc(sizeof(*add_branch), GFP_NOFS); ++ if (unlikely(!add_branch)) ++ goto out; ++ ++ add_branch->br_wbr = NULL; ++ if (au_br_writable(perm)) { ++ /* may be freed separately at changing the branch permission */ ++ add_branch->br_wbr = kmalloc(sizeof(*add_branch->br_wbr), ++ GFP_NOFS); ++ if (unlikely(!add_branch->br_wbr)) ++ goto out_br; ++ } ++ ++ if (unlikely(au_sbr_realloc(au_sbi(sb), new_nbranch) ++ || au_di_realloc(au_di(root), new_nbranch) ++ || au_ii_realloc(au_ii(root->d_inode), new_nbranch))) ++ goto out_wbr; ++ return add_branch; /* success */ ++ ++ out_wbr: ++ kfree(add_branch->br_wbr); ++ out_br: ++ kfree(add_branch); ++ out: ++ return ERR_PTR(-ENOMEM); ++} ++ ++/* ++ * test if the branch permission is legal or not. ++ */ ++static int test_br(struct inode *inode, int brperm, char *path) ++{ ++ int err; ++ ++ err = 0; ++ if (unlikely(au_br_writable(brperm) && IS_RDONLY(inode))) { ++ AuErr("write permission for readonly mount or inode, %s\n", ++ path); ++ err = -EINVAL; ++ } ++ ++ return err; ++} ++ ++/* ++ * returns: ++ * 0: success, the caller will add it ++ * plus: success, it is already unified, the caller should ignore it ++ * minus: error ++ */ ++static int test_add(struct super_block *sb, struct au_opt_add *add, int remount) ++{ ++ int err; ++ aufs_bindex_t bend, bindex; ++ struct dentry *root; ++ struct inode *inode, *h_inode; ++ ++ root = sb->s_root; ++ bend = au_sbend(sb); ++ if (unlikely(bend >= 0 ++ && au_find_dbindex(root, add->path.dentry) >= 0)) { ++ err = 1; ++ if (!remount) { ++ err = -EINVAL; ++ AuErr("%s duplicated\n", add->pathname); ++ } ++ goto out; ++ } ++ ++ err = -ENOSPC; /* -E2BIG; */ ++ if (unlikely(AUFS_BRANCH_MAX <= add->bindex ++ || AUFS_BRANCH_MAX - 1 <= bend)) { ++ AuErr("number of branches exceeded %s\n", add->pathname); ++ goto out; ++ } ++ ++ err = -EDOM; ++ if (unlikely(add->bindex < 0 || bend + 1 < add->bindex)) { ++ AuErr("bad index %d\n", add->bindex); ++ goto out; ++ } ++ ++ inode = add->path.dentry->d_inode; ++ err = -ENOENT; ++ if (unlikely(!inode->i_nlink)) { ++ AuErr("no existence %s\n", add->pathname); ++ goto out; ++ } ++ ++ err = -EINVAL; ++ if (unlikely(inode->i_sb == sb)) { ++ AuErr("%s must be outside\n", add->pathname); ++ goto out; ++ } ++ ++ if (unlikely(au_test_fs_unsuppoted(inode->i_sb))) { ++ AuErr("unsupported filesystem, %s (%s)\n", ++ add->pathname, au_sbtype(inode->i_sb)); ++ goto out; ++ } ++ ++ err = test_br(add->path.dentry->d_inode, add->perm, add->pathname); ++ if (unlikely(err)) ++ goto out; ++ ++ if (bend < 0) ++ return 0; /* success */ ++ ++ err = -EINVAL; ++ for (bindex = 0; bindex <= bend; bindex++) ++ if (unlikely(test_overlap(sb, add->path.dentry, ++ au_h_dptr(root, bindex)))) { ++ AuErr("%s is overlapped\n", add->pathname); ++ goto out; ++ } ++ ++ err = 0; ++ if (au_opt_test(au_mntflags(sb), WARN_PERM)) { ++ h_inode = au_h_dptr(root, 0)->d_inode; ++ if ((h_inode->i_mode & S_IALLUGO) != (inode->i_mode & S_IALLUGO) ++ || h_inode->i_uid != inode->i_uid ++ || h_inode->i_gid != inode->i_gid) ++ AuWarn("uid/gid/perm %s %u/%u/0%o, %u/%u/0%o\n", ++ add->pathname, ++ inode->i_uid, inode->i_gid, ++ (inode->i_mode & S_IALLUGO), ++ h_inode->i_uid, h_inode->i_gid, ++ (h_inode->i_mode & S_IALLUGO)); ++ } ++ ++ out: ++ return err; ++} ++ ++/* ++ * initialize or clean the whiteouts for an adding branch ++ */ ++static int au_br_init_wh(struct super_block *sb, struct au_branch *br, ++ int new_perm, struct dentry *h_root) ++{ ++ int err, old_perm; ++ aufs_bindex_t bindex; ++ struct mutex *h_mtx; ++ struct au_wbr *wbr; ++ struct au_hinode *hdir; ++ ++ wbr = br->br_wbr; ++ old_perm = br->br_perm; ++ br->br_perm = new_perm; ++ hdir = NULL; ++ h_mtx = NULL; ++ bindex = au_br_index(sb, br->br_id); ++ if (0 <= bindex) { ++ hdir = au_hi(sb->s_root->d_inode, bindex); ++ au_hin_imtx_lock_nested(hdir, AuLsc_I_PARENT); ++ } else { ++ h_mtx = &h_root->d_inode->i_mutex; ++ mutex_lock_nested(h_mtx, AuLsc_I_PARENT); ++ } ++ if (!wbr) ++ err = au_wh_init(h_root, br, sb); ++ else { ++ wbr_wh_write_lock(wbr); ++ err = au_wh_init(h_root, br, sb); ++ wbr_wh_write_unlock(wbr); ++ } ++ if (hdir) ++ au_hin_imtx_unlock(hdir); ++ else ++ mutex_unlock(h_mtx); ++ br->br_perm = old_perm; ++ ++ if (!err && wbr && !au_br_writable(new_perm)) { ++ kfree(wbr); ++ br->br_wbr = NULL; ++ } ++ ++ return err; ++} ++ ++static int au_wbr_init(struct au_branch *br, struct super_block *sb, ++ int perm, struct path *path) ++{ ++ int err; ++ struct au_wbr *wbr; ++ ++ wbr = br->br_wbr; ++ au_rw_init(&wbr->wbr_wh_rwsem); ++ memset(wbr->wbr_wh, 0, sizeof(wbr->wbr_wh)); ++ atomic_set(&wbr->wbr_wh_running, 0); ++ wbr->wbr_bytes = 0; ++ ++ err = au_br_init_wh(sb, br, perm, path->dentry); ++ ++ return err; ++} ++ ++/* intialize a new branch */ ++static int au_br_init(struct au_branch *br, struct super_block *sb, ++ struct au_opt_add *add) ++{ ++ int err; ++ ++ err = 0; ++ memset(&br->br_xino, 0, sizeof(br->br_xino)); ++ mutex_init(&br->br_xino.xi_nondir_mtx); ++ br->br_perm = add->perm; ++ br->br_mnt = add->path.mnt; /* set first, mntget() later */ ++ atomic_set(&br->br_count, 0); ++ br->br_xino_upper = AUFS_XINO_TRUNC_INIT; ++ atomic_set(&br->br_xino_running, 0); ++ br->br_id = au_new_br_id(sb); ++ ++ if (au_br_writable(add->perm)) { ++ err = au_wbr_init(br, sb, add->perm, &add->path); ++ if (unlikely(err)) ++ goto out; ++ } ++ ++ if (au_opt_test(au_mntflags(sb), XINO)) { ++ err = au_xino_br(sb, br, add->path.dentry->d_inode->i_ino, ++ au_sbr(sb, 0)->br_xino.xi_file, /*do_test*/1); ++ if (unlikely(err)) { ++ AuDebugOn(br->br_xino.xi_file); ++ goto out; ++ } ++ } ++ ++ sysaufs_br_init(br); ++ mntget(add->path.mnt); ++ ++ out: ++ return err; ++} ++ ++static void au_br_do_add_brp(struct au_sbinfo *sbinfo, aufs_bindex_t bindex, ++ struct au_branch *br, aufs_bindex_t bend, ++ aufs_bindex_t amount) ++{ ++ struct au_branch **brp; ++ ++ AuRwMustWriteLock(&sbinfo->si_rwsem); ++ ++ brp = sbinfo->si_branch + bindex; ++ memmove(brp + 1, brp, sizeof(*brp) * amount); ++ *brp = br; ++ sbinfo->si_bend++; ++ if (unlikely(bend < 0)) ++ sbinfo->si_bend = 0; ++} ++ ++static void au_br_do_add_hdp(struct au_dinfo *dinfo, aufs_bindex_t bindex, ++ aufs_bindex_t bend, aufs_bindex_t amount) ++{ ++ struct au_hdentry *hdp; ++ ++ AuRwMustWriteLock(&dinfo->di_rwsem); ++ ++ hdp = dinfo->di_hdentry + bindex; ++ memmove(hdp + 1, hdp, sizeof(*hdp) * amount); ++ au_h_dentry_init(hdp); ++ dinfo->di_bend++; ++ if (unlikely(bend < 0)) ++ dinfo->di_bstart = 0; ++} ++ ++static void au_br_do_add_hip(struct au_iinfo *iinfo, aufs_bindex_t bindex, ++ aufs_bindex_t bend, aufs_bindex_t amount) ++{ ++ struct au_hinode *hip; ++ ++ AuRwMustWriteLock(&iinfo->ii_rwsem); ++ ++ hip = iinfo->ii_hinode + bindex; ++ memmove(hip + 1, hip, sizeof(*hip) * amount); ++ hip->hi_inode = NULL; ++ au_hin_init(hip, NULL); ++ iinfo->ii_bend++; ++ if (unlikely(bend < 0)) ++ iinfo->ii_bstart = 0; ++} ++ ++static void au_br_do_add(struct super_block *sb, struct dentry *h_dentry, ++ struct au_branch *br, aufs_bindex_t bindex) ++{ ++ struct dentry *root; ++ struct inode *root_inode; ++ aufs_bindex_t bend, amount; ++ ++ root = sb->s_root; ++ root_inode = root->d_inode; ++ au_plink_block_maintain(sb); ++ bend = au_sbend(sb); ++ amount = bend + 1 - bindex; ++ au_br_do_add_brp(au_sbi(sb), bindex, br, bend, amount); ++ au_br_do_add_hdp(au_di(root), bindex, bend, amount); ++ au_br_do_add_hip(au_ii(root_inode), bindex, bend, amount); ++ au_set_h_dptr(root, bindex, dget(h_dentry)); ++ au_set_h_iptr(root_inode, bindex, au_igrab(h_dentry->d_inode), ++ /*flags*/0); ++} ++ ++int au_br_add(struct super_block *sb, struct au_opt_add *add, int remount) ++{ ++ int err; ++ unsigned long long maxb; ++ aufs_bindex_t bend, add_bindex; ++ struct dentry *root, *h_dentry; ++ struct inode *root_inode; ++ struct au_branch *add_branch; ++ ++ root = sb->s_root; ++ root_inode = root->d_inode; ++ IMustLock(root_inode); ++ err = test_add(sb, add, remount); ++ if (unlikely(err < 0)) ++ goto out; ++ if (err) { ++ err = 0; ++ goto out; /* success */ ++ } ++ ++ bend = au_sbend(sb); ++ add_branch = au_br_alloc(sb, bend + 2, add->perm); ++ err = PTR_ERR(add_branch); ++ if (IS_ERR(add_branch)) ++ goto out; ++ ++ err = au_br_init(add_branch, sb, add); ++ if (unlikely(err)) { ++ au_br_do_free(add_branch); ++ goto out; ++ } ++ ++ add_bindex = add->bindex; ++ h_dentry = add->path.dentry; ++ if (!remount) ++ au_br_do_add(sb, h_dentry, add_branch, add_bindex); ++ else { ++ sysaufs_brs_del(sb, add_bindex); ++ au_br_do_add(sb, h_dentry, add_branch, add_bindex); ++ sysaufs_brs_add(sb, add_bindex); ++ } ++ ++ if (!add_bindex) ++ au_cpup_attr_all(root_inode, /*force*/1); ++ else ++ au_add_nlink(root_inode, h_dentry->d_inode); ++ maxb = h_dentry->d_sb->s_maxbytes; ++ if (sb->s_maxbytes < maxb) ++ sb->s_maxbytes = maxb; ++ ++ /* ++ * this test/set prevents aufs from handling unnecesary inotify events ++ * of xino files, in a case of re-adding a writable branch which was ++ * once detached from aufs. ++ */ ++ if (au_xino_brid(sb) < 0 ++ && au_br_writable(add_branch->br_perm) ++ && !au_test_fs_bad_xino(h_dentry->d_sb) ++ && add_branch->br_xino.xi_file ++ && add_branch->br_xino.xi_file->f_dentry->d_parent == h_dentry) ++ au_xino_brid_set(sb, add_branch->br_id); ++ ++ out: ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* ++ * delete a branch ++ */ ++ ++/* to show the line number, do not make it inlined function */ ++#define AuVerbose(do_info, fmt, args...) do { \ ++ if (do_info) \ ++ AuInfo(fmt, ##args); \ ++} while (0) ++ ++/* ++ * test if the branch is deletable or not. ++ */ ++static int test_dentry_busy(struct dentry *root, aufs_bindex_t bindex, ++ unsigned int sigen) ++{ ++ int err, i, j, ndentry; ++ aufs_bindex_t bstart, bend; ++ unsigned char verbose; ++ struct au_dcsub_pages dpages; ++ struct au_dpage *dpage; ++ struct dentry *d; ++ struct inode *inode; ++ ++ err = au_dpages_init(&dpages, GFP_NOFS); ++ if (unlikely(err)) ++ goto out; ++ err = au_dcsub_pages(&dpages, root, NULL, NULL); ++ if (unlikely(err)) ++ goto out_dpages; ++ ++ verbose = !!au_opt_test(au_mntflags(root->d_sb), VERBOSE); ++ for (i = 0; !err && i < dpages.ndpage; i++) { ++ dpage = dpages.dpages + i; ++ ndentry = dpage->ndentry; ++ for (j = 0; !err && j < ndentry; j++) { ++ d = dpage->dentries[j]; ++ AuDebugOn(!atomic_read(&d->d_count)); ++ inode = d->d_inode; ++ if (au_digen(d) == sigen && au_iigen(inode) == sigen) ++ di_read_lock_child(d, AuLock_IR); ++ else { ++ di_write_lock_child(d); ++ err = au_reval_dpath(d, sigen); ++ if (!err) ++ di_downgrade_lock(d, AuLock_IR); ++ else { ++ di_write_unlock(d); ++ break; ++ } ++ } ++ ++ bstart = au_dbstart(d); ++ bend = au_dbend(d); ++ if (bstart <= bindex ++ && bindex <= bend ++ && au_h_dptr(d, bindex) ++ && (!S_ISDIR(inode->i_mode) || bstart == bend)) { ++ err = -EBUSY; ++ AuVerbose(verbose, "busy %.*s\n", AuDLNPair(d)); ++ } ++ di_read_unlock(d, AuLock_IR); ++ } ++ } ++ ++ out_dpages: ++ au_dpages_free(&dpages); ++ out: ++ return err; ++} ++ ++static int test_inode_busy(struct super_block *sb, aufs_bindex_t bindex, ++ unsigned int sigen) ++{ ++ int err; ++ struct inode *i; ++ aufs_bindex_t bstart, bend; ++ unsigned char verbose; ++ ++ err = 0; ++ verbose = !!au_opt_test(au_mntflags(sb), VERBOSE); ++ list_for_each_entry(i, &sb->s_inodes, i_sb_list) { ++ AuDebugOn(!atomic_read(&i->i_count)); ++ if (!list_empty(&i->i_dentry)) ++ continue; ++ ++ if (au_iigen(i) == sigen) ++ ii_read_lock_child(i); ++ else { ++ ii_write_lock_child(i); ++ err = au_refresh_hinode_self(i, /*do_attr*/1); ++ if (!err) ++ ii_downgrade_lock(i); ++ else { ++ ii_write_unlock(i); ++ break; ++ } ++ } ++ ++ bstart = au_ibstart(i); ++ bend = au_ibend(i); ++ if (bstart <= bindex ++ && bindex <= bend ++ && au_h_iptr(i, bindex) ++ && (!S_ISDIR(i->i_mode) || bstart == bend)) { ++ err = -EBUSY; ++ AuVerbose(verbose, "busy i%lu\n", i->i_ino); ++ ii_read_unlock(i); ++ break; ++ } ++ ii_read_unlock(i); ++ } ++ ++ return err; ++} ++ ++static int test_children_busy(struct dentry *root, aufs_bindex_t bindex) ++{ ++ int err; ++ unsigned int sigen; ++ ++ sigen = au_sigen(root->d_sb); ++ DiMustNoWaiters(root); ++ IiMustNoWaiters(root->d_inode); ++ di_write_unlock(root); ++ err = test_dentry_busy(root, bindex, sigen); ++ if (!err) ++ err = test_inode_busy(root->d_sb, bindex, sigen); ++ di_write_lock_child(root); /* aufs_write_lock() calls ..._child() */ ++ ++ return err; ++} ++ ++static void au_br_do_del_brp(struct au_sbinfo *sbinfo, ++ const aufs_bindex_t bindex, ++ const aufs_bindex_t bend) ++{ ++ struct au_branch **brp, **p; ++ ++ AuRwMustWriteLock(&sbinfo->si_rwsem); ++ ++ brp = sbinfo->si_branch + bindex; ++ if (bindex < bend) ++ memmove(brp, brp + 1, sizeof(*brp) * (bend - bindex)); ++ sbinfo->si_branch[0 + bend] = NULL; ++ sbinfo->si_bend--; ++ ++ p = krealloc(sbinfo->si_branch, sizeof(*p) * bend, GFP_NOFS); ++ if (p) ++ sbinfo->si_branch = p; ++} ++ ++static void au_br_do_del_hdp(struct au_dinfo *dinfo, const aufs_bindex_t bindex, ++ const aufs_bindex_t bend) ++{ ++ struct au_hdentry *hdp, *p; ++ ++ AuRwMustWriteLock(&dinfo->di_rwsem); ++ ++ hdp = dinfo->di_hdentry + bindex; ++ if (bindex < bend) ++ memmove(hdp, hdp + 1, sizeof(*hdp) * (bend - bindex)); ++ dinfo->di_hdentry[0 + bend].hd_dentry = NULL; ++ dinfo->di_bend--; ++ ++ p = krealloc(dinfo->di_hdentry, sizeof(*p) * bend, GFP_NOFS); ++ if (p) ++ dinfo->di_hdentry = p; ++} ++ ++static void au_br_do_del_hip(struct au_iinfo *iinfo, const aufs_bindex_t bindex, ++ const aufs_bindex_t bend) ++{ ++ struct au_hinode *hip, *p; ++ ++ AuRwMustWriteLock(&iinfo->ii_rwsem); ++ ++ hip = iinfo->ii_hinode + bindex; ++ if (bindex < bend) ++ memmove(hip, hip + 1, sizeof(*hip) * (bend - bindex)); ++ iinfo->ii_hinode[0 + bend].hi_inode = NULL; ++ au_hin_init(iinfo->ii_hinode + bend, NULL); ++ iinfo->ii_bend--; ++ ++ p = krealloc(iinfo->ii_hinode, sizeof(*p) * bend, GFP_NOFS); ++ if (p) ++ iinfo->ii_hinode = p; ++} ++ ++static void au_br_do_del(struct super_block *sb, aufs_bindex_t bindex, ++ struct au_branch *br) ++{ ++ aufs_bindex_t bend; ++ struct au_sbinfo *sbinfo; ++ struct dentry *root; ++ struct inode *inode; ++ ++ SiMustWriteLock(sb); ++ ++ root = sb->s_root; ++ inode = root->d_inode; ++ au_plink_block_maintain(sb); ++ sbinfo = au_sbi(sb); ++ bend = sbinfo->si_bend; ++ ++ dput(au_h_dptr(root, bindex)); ++ au_hiput(au_hi(inode, bindex)); ++ au_br_do_free(br); ++ ++ au_br_do_del_brp(sbinfo, bindex, bend); ++ au_br_do_del_hdp(au_di(root), bindex, bend); ++ au_br_do_del_hip(au_ii(inode), bindex, bend); ++} ++ ++int au_br_del(struct super_block *sb, struct au_opt_del *del, int remount) ++{ ++ int err, rerr, i; ++ unsigned int mnt_flags; ++ aufs_bindex_t bindex, bend, br_id; ++ unsigned char do_wh, verbose; ++ struct au_branch *br; ++ struct au_wbr *wbr; ++ ++ err = 0; ++ bindex = au_find_dbindex(sb->s_root, del->h_path.dentry); ++ if (bindex < 0) { ++ if (remount) ++ goto out; /* success */ ++ err = -ENOENT; ++ AuErr("%s no such branch\n", del->pathname); ++ goto out; ++ } ++ AuDbg("bindex b%d\n", bindex); ++ ++ err = -EBUSY; ++ mnt_flags = au_mntflags(sb); ++ verbose = !!au_opt_test(mnt_flags, VERBOSE); ++ bend = au_sbend(sb); ++ if (unlikely(!bend)) { ++ AuVerbose(verbose, "no more branches left\n"); ++ goto out; ++ } ++ br = au_sbr(sb, bindex); ++ i = atomic_read(&br->br_count); ++ if (unlikely(i)) { ++ AuVerbose(verbose, "%d file(s) opened\n", i); ++ goto out; ++ } ++ ++ wbr = br->br_wbr; ++ do_wh = wbr && (wbr->wbr_whbase || wbr->wbr_plink || wbr->wbr_orph); ++ if (do_wh) { ++ /* instead of WbrWhMustWriteLock(wbr) */ ++ SiMustWriteLock(sb); ++ for (i = 0; i < AuBrWh_Last; i++) { ++ dput(wbr->wbr_wh[i]); ++ wbr->wbr_wh[i] = NULL; ++ } ++ } ++ ++ err = test_children_busy(sb->s_root, bindex); ++ if (unlikely(err)) { ++ if (do_wh) ++ goto out_wh; ++ goto out; ++ } ++ ++ err = 0; ++ br_id = br->br_id; ++ if (!remount) ++ au_br_do_del(sb, bindex, br); ++ else { ++ sysaufs_brs_del(sb, bindex); ++ au_br_do_del(sb, bindex, br); ++ sysaufs_brs_add(sb, bindex); ++ } ++ ++ if (!bindex) ++ au_cpup_attr_all(sb->s_root->d_inode, /*force*/1); ++ else ++ au_sub_nlink(sb->s_root->d_inode, del->h_path.dentry->d_inode); ++ if (au_opt_test(mnt_flags, PLINK)) ++ au_plink_half_refresh(sb, br_id); ++ ++ if (sb->s_maxbytes == del->h_path.dentry->d_sb->s_maxbytes) { ++ bend--; ++ sb->s_maxbytes = 0; ++ for (bindex = 0; bindex <= bend; bindex++) { ++ unsigned long long maxb; ++ ++ maxb = au_sbr_sb(sb, bindex)->s_maxbytes; ++ if (sb->s_maxbytes < maxb) ++ sb->s_maxbytes = maxb; ++ } ++ } ++ ++ if (au_xino_brid(sb) == br->br_id) ++ au_xino_brid_set(sb, -1); ++ goto out; /* success */ ++ ++ out_wh: ++ /* revert */ ++ rerr = au_br_init_wh(sb, br, br->br_perm, del->h_path.dentry); ++ if (rerr) ++ AuWarn("failed re-creating base whiteout, %s. (%d)\n", ++ del->pathname, rerr); ++ out: ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* ++ * change a branch permission ++ */ ++ ++static int do_need_sigen_inc(int a, int b) ++{ ++ return au_br_whable(a) && !au_br_whable(b); ++} ++ ++static int need_sigen_inc(int old, int new) ++{ ++ return do_need_sigen_inc(old, new) ++ || do_need_sigen_inc(new, old); ++} ++ ++static int au_br_mod_files_ro(struct super_block *sb, aufs_bindex_t bindex) ++{ ++ int err; ++ unsigned long n, ul, bytes, files; ++ aufs_bindex_t bstart; ++ struct file *file, *hf, **a; ++ const int step_bytes = 1024, /* memory allocation unit */ ++ step_files = step_bytes / sizeof(*a); ++ ++ err = -ENOMEM; ++ n = 0; ++ bytes = step_bytes; ++ files = step_files; ++ a = kmalloc(bytes, GFP_NOFS); ++ if (unlikely(!a)) ++ goto out; ++ ++ /* no need file_list_lock() since sbinfo is locked? defered? */ ++ list_for_each_entry(file, &sb->s_files, f_u.fu_list) { ++ if (special_file(file->f_dentry->d_inode->i_mode)) ++ continue; ++ ++ AuDbg("%.*s\n", AuDLNPair(file->f_dentry)); ++ fi_read_lock(file); ++ if (unlikely(au_test_mmapped(file))) { ++ err = -EBUSY; ++ FiMustNoWaiters(file); ++ fi_read_unlock(file); ++ goto out_free; ++ } ++ ++ bstart = au_fbstart(file); ++ if (!S_ISREG(file->f_dentry->d_inode->i_mode) ++ || !(file->f_mode & FMODE_WRITE) ++ || bstart != bindex) { ++ FiMustNoWaiters(file); ++ fi_read_unlock(file); ++ continue; ++ } ++ ++ hf = au_h_fptr(file, bstart); ++ FiMustNoWaiters(file); ++ fi_read_unlock(file); ++ ++ if (n < files) ++ a[n++] = hf; ++ else { ++ void *p; ++ ++ err = -ENOMEM; ++ bytes += step_bytes; ++ files += step_files; ++ p = krealloc(a, bytes, GFP_NOFS); ++ if (p) { ++ a = p; ++ a[n++] = hf; ++ } else ++ goto out_free; ++ } ++ } ++ ++ err = 0; ++ for (ul = 0; ul < n; ul++) { ++ /* todo: already flushed? */ ++ /* cf. fs/super.c:mark_files_ro() */ ++ hf = a[ul]; ++ hf->f_mode &= ~FMODE_WRITE; ++ if (!file_check_writeable(hf)) { ++ file_release_write(hf); ++ mnt_drop_write(hf->f_vfsmnt); ++ } ++ } ++ ++ out_free: ++ kfree(a); ++ out: ++ return err; ++} ++ ++int au_br_mod(struct super_block *sb, struct au_opt_mod *mod, int remount, ++ int *do_update) ++{ ++ int err, rerr; ++ aufs_bindex_t bindex; ++ struct dentry *root; ++ struct au_branch *br; ++ ++ root = sb->s_root; ++ au_plink_block_maintain(sb); ++ bindex = au_find_dbindex(root, mod->h_root); ++ if (bindex < 0) { ++ if (remount) ++ return 0; /* success */ ++ err = -ENOENT; ++ AuErr("%s no such branch\n", mod->path); ++ goto out; ++ } ++ AuDbg("bindex b%d\n", bindex); ++ ++ err = test_br(mod->h_root->d_inode, mod->perm, mod->path); ++ if (unlikely(err)) ++ goto out; ++ ++ br = au_sbr(sb, bindex); ++ if (br->br_perm == mod->perm) ++ return 0; /* success */ ++ ++ if (au_br_writable(br->br_perm)) { ++ /* remove whiteout base */ ++ err = au_br_init_wh(sb, br, mod->perm, mod->h_root); ++ if (unlikely(err)) ++ goto out; ++ ++ if (!au_br_writable(mod->perm)) { ++ /* rw --> ro, file might be mmapped */ ++ DiMustNoWaiters(root); ++ IiMustNoWaiters(root->d_inode); ++ di_write_unlock(root); ++ err = au_br_mod_files_ro(sb, bindex); ++ /* aufs_write_lock() calls ..._child() */ ++ di_write_lock_child(root); ++ ++ if (unlikely(err)) { ++ rerr = -ENOMEM; ++ br->br_wbr = kmalloc(sizeof(*br->br_wbr), ++ GFP_NOFS); ++ if (br->br_wbr) ++ rerr = au_br_init_wh ++ (sb, br, br->br_perm, ++ mod->h_root); ++ if (unlikely(rerr)) { ++ AuIOErr("nested error %d (%d)\n", ++ rerr, err); ++ br->br_perm = mod->perm; ++ } ++ } ++ } ++ } else if (au_br_writable(mod->perm)) { ++ /* ro --> rw */ ++ err = -ENOMEM; ++ br->br_wbr = kmalloc(sizeof(*br->br_wbr), GFP_NOFS); ++ if (br->br_wbr) { ++ struct path path = { ++ .mnt = br->br_mnt, ++ .dentry = mod->h_root ++ }; ++ ++ err = au_wbr_init(br, sb, mod->perm, &path); ++ if (unlikely(err)) { ++ kfree(br->br_wbr); ++ br->br_wbr = NULL; ++ } ++ } ++ } ++ ++ if (!err) { ++ *do_update |= need_sigen_inc(br->br_perm, mod->perm); ++ br->br_perm = mod->perm; ++ } ++ ++ out: ++ return err; ++} +diff -Nur linux-2.6.31.4.orig/fs/aufs/branch.h linux-2.6.31.4/fs/aufs/branch.h +--- linux-2.6.31.4.orig/fs/aufs/branch.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/branch.h 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,219 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * branch filesystems and xino for them ++ */ ++ ++#ifndef __AUFS_BRANCH_H__ ++#define __AUFS_BRANCH_H__ ++ ++#ifdef __KERNEL__ ++ ++#include <linux/fs.h> ++#include <linux/mount.h> ++#include <linux/aufs_type.h> ++#include "rwsem.h" ++#include "super.h" ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* a xino file */ ++struct au_xino_file { ++ struct file *xi_file; ++ struct mutex xi_nondir_mtx; ++ ++ /* todo: make xino files an array to support huge inode number */ ++ ++#ifdef CONFIG_DEBUG_FS ++ struct dentry *xi_dbgaufs; ++#endif ++}; ++ ++/* members for writable branch only */ ++enum {AuBrWh_BASE, AuBrWh_PLINK, AuBrWh_ORPH, AuBrWh_Last}; ++struct au_wbr { ++ struct au_rwsem wbr_wh_rwsem; ++ struct dentry *wbr_wh[AuBrWh_Last]; ++ atomic_t wbr_wh_running; ++#define wbr_whbase wbr_wh[AuBrWh_BASE] /* whiteout base */ ++#define wbr_plink wbr_wh[AuBrWh_PLINK] /* pseudo-link dir */ ++#define wbr_orph wbr_wh[AuBrWh_ORPH] /* dir for orphans */ ++ ++ /* mfs mode */ ++ unsigned long long wbr_bytes; ++}; ++ ++/* protected by superblock rwsem */ ++struct au_branch { ++ struct au_xino_file br_xino; ++ ++ aufs_bindex_t br_id; ++ ++ int br_perm; ++ struct vfsmount *br_mnt; ++ atomic_t br_count; ++ ++ struct au_wbr *br_wbr; ++ ++ /* xino truncation */ ++ blkcnt_t br_xino_upper; /* watermark in blocks */ ++ atomic_t br_xino_running; ++ ++#ifdef CONFIG_SYSFS ++ /* an entry under sysfs per mount-point */ ++ char br_name[8]; ++ struct attribute br_attr; ++#endif ++}; ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* branch permission and attribute */ ++enum { ++ AuBrPerm_RW, /* writable, linkable wh */ ++ AuBrPerm_RO, /* readonly, no wh */ ++ AuBrPerm_RR, /* natively readonly, no wh */ ++ ++ AuBrPerm_RWNoLinkWH, /* un-linkable whiteouts */ ++ ++ AuBrPerm_ROWH, /* whiteout-able */ ++ AuBrPerm_RRWH, /* whiteout-able */ ++ ++ AuBrPerm_Last ++}; ++ ++static inline int au_br_writable(int brperm) ++{ ++ return brperm == AuBrPerm_RW || brperm == AuBrPerm_RWNoLinkWH; ++} ++ ++static inline int au_br_whable(int brperm) ++{ ++ return brperm == AuBrPerm_RW ++ || brperm == AuBrPerm_ROWH ++ || brperm == AuBrPerm_RRWH; ++} ++ ++static inline int au_br_rdonly(struct au_branch *br) ++{ ++ return ((br->br_mnt->mnt_sb->s_flags & MS_RDONLY) ++ || !au_br_writable(br->br_perm)) ++ ? -EROFS : 0; ++} ++ ++static inline int au_br_hinotifyable(int brperm __maybe_unused) ++{ ++#ifdef CONFIG_AUFS_HINOTIFY ++ return brperm != AuBrPerm_RR && brperm != AuBrPerm_RRWH; ++#else ++ return 0; ++#endif ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* branch.c */ ++struct au_sbinfo; ++void au_br_free(struct au_sbinfo *sinfo); ++int au_br_index(struct super_block *sb, aufs_bindex_t br_id); ++struct au_opt_add; ++int au_br_add(struct super_block *sb, struct au_opt_add *add, int remount); ++struct au_opt_del; ++int au_br_del(struct super_block *sb, struct au_opt_del *del, int remount); ++struct au_opt_mod; ++int au_br_mod(struct super_block *sb, struct au_opt_mod *mod, int remount, ++ int *do_update); ++ ++/* xino.c */ ++static const loff_t au_loff_max = LLONG_MAX; ++ ++int au_xib_trunc(struct super_block *sb); ++ssize_t xino_fread(au_readf_t func, struct file *file, void *buf, size_t size, ++ loff_t *pos); ++ssize_t xino_fwrite(au_writef_t func, struct file *file, void *buf, size_t size, ++ loff_t *pos); ++struct file *au_xino_create2(struct file *base_file, struct file *copy_src); ++struct file *au_xino_create(struct super_block *sb, char *fname, int silent); ++ino_t au_xino_new_ino(struct super_block *sb); ++int au_xino_write0(struct super_block *sb, aufs_bindex_t bindex, ino_t h_ino, ++ ino_t ino); ++int au_xino_write(struct super_block *sb, aufs_bindex_t bindex, ino_t h_ino, ++ ino_t ino); ++int au_xino_read(struct super_block *sb, aufs_bindex_t bindex, ino_t h_ino, ++ ino_t *ino); ++int au_xino_br(struct super_block *sb, struct au_branch *br, ino_t hino, ++ struct file *base_file, int do_test); ++int au_xino_trunc(struct super_block *sb, aufs_bindex_t bindex); ++ ++struct au_opt_xino; ++int au_xino_set(struct super_block *sb, struct au_opt_xino *xino, int remount); ++void au_xino_clr(struct super_block *sb); ++struct file *au_xino_def(struct super_block *sb); ++int au_xino_path(struct seq_file *seq, struct file *file); ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* Superblock to branch */ ++static inline ++aufs_bindex_t au_sbr_id(struct super_block *sb, aufs_bindex_t bindex) ++{ ++ return au_sbr(sb, bindex)->br_id; ++} ++ ++static inline ++struct vfsmount *au_sbr_mnt(struct super_block *sb, aufs_bindex_t bindex) ++{ ++ return au_sbr(sb, bindex)->br_mnt; ++} ++ ++static inline ++struct super_block *au_sbr_sb(struct super_block *sb, aufs_bindex_t bindex) ++{ ++ return au_sbr_mnt(sb, bindex)->mnt_sb; ++} ++ ++static inline void au_sbr_put(struct super_block *sb, aufs_bindex_t bindex) ++{ ++ atomic_dec_return(&au_sbr(sb, bindex)->br_count); ++} ++ ++static inline int au_sbr_perm(struct super_block *sb, aufs_bindex_t bindex) ++{ ++ return au_sbr(sb, bindex)->br_perm; ++} ++ ++static inline int au_sbr_whable(struct super_block *sb, aufs_bindex_t bindex) ++{ ++ return au_br_whable(au_sbr_perm(sb, bindex)); ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* ++ * wbr_wh_read_lock, wbr_wh_write_lock ++ * wbr_wh_read_unlock, wbr_wh_write_unlock, wbr_wh_downgrade_lock ++ */ ++AuSimpleRwsemFuncs(wbr_wh, struct au_wbr *wbr, &wbr->wbr_wh_rwsem); ++ ++#define WbrWhMustNoWaiters(wbr) AuRwMustNoWaiters(&wbr->wbr_wh_rwsem) ++#define WbrWhMustAnyLock(wbr) AuRwMustAnyLock(&wbr->wbr_wh_rwsem) ++#define WbrWhMustWriteLock(wbr) AuRwMustWriteLock(&wbr->wbr_wh_rwsem) ++ ++#endif /* __KERNEL__ */ ++#endif /* __AUFS_BRANCH_H__ */ +diff -Nur linux-2.6.31.4.orig/fs/aufs/cpup.c linux-2.6.31.4/fs/aufs/cpup.c +--- linux-2.6.31.4.orig/fs/aufs/cpup.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/cpup.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,1048 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * copy-up functions, see wbr_policy.c for copy-down ++ */ ++ ++#include <linux/file.h> ++#include <linux/fs_stack.h> ++#include <linux/mm.h> ++#include <linux/uaccess.h> ++#include "aufs.h" ++ ++void au_cpup_attr_flags(struct inode *dst, struct inode *src) ++{ ++ const unsigned int mask = S_DEAD | S_SWAPFILE | S_PRIVATE ++ | S_NOATIME | S_NOCMTIME; ++ ++ dst->i_flags |= src->i_flags & ~mask; ++ if (au_test_fs_notime(dst->i_sb)) ++ dst->i_flags |= S_NOATIME | S_NOCMTIME; ++} ++ ++void au_cpup_attr_timesizes(struct inode *inode) ++{ ++ struct inode *h_inode; ++ ++ h_inode = au_h_iptr(inode, au_ibstart(inode)); ++ fsstack_copy_attr_times(inode, h_inode); ++ vfsub_copy_inode_size(inode, h_inode); ++} ++ ++void au_cpup_attr_nlink(struct inode *inode, int force) ++{ ++ struct inode *h_inode; ++ struct super_block *sb; ++ aufs_bindex_t bindex, bend; ++ ++ sb = inode->i_sb; ++ bindex = au_ibstart(inode); ++ h_inode = au_h_iptr(inode, bindex); ++ if (!force ++ && !S_ISDIR(h_inode->i_mode) ++ && au_opt_test(au_mntflags(sb), PLINK) ++ && au_plink_test(inode)) ++ return; ++ ++ inode->i_nlink = h_inode->i_nlink; ++ ++ /* ++ * fewer nlink makes find(1) noisy, but larger nlink doesn't. ++ * it may includes whplink directory. ++ */ ++ if (S_ISDIR(h_inode->i_mode)) { ++ bend = au_ibend(inode); ++ for (bindex++; bindex <= bend; bindex++) { ++ h_inode = au_h_iptr(inode, bindex); ++ if (h_inode) ++ au_add_nlink(inode, h_inode); ++ } ++ } ++} ++ ++void au_cpup_attr_changeable(struct inode *inode) ++{ ++ struct inode *h_inode; ++ ++ h_inode = au_h_iptr(inode, au_ibstart(inode)); ++ inode->i_mode = h_inode->i_mode; ++ inode->i_uid = h_inode->i_uid; ++ inode->i_gid = h_inode->i_gid; ++ au_cpup_attr_timesizes(inode); ++ au_cpup_attr_flags(inode, h_inode); ++} ++ ++void au_cpup_igen(struct inode *inode, struct inode *h_inode) ++{ ++ struct au_iinfo *iinfo = au_ii(inode); ++ ++ IiMustWriteLock(inode); ++ ++ iinfo->ii_higen = h_inode->i_generation; ++ iinfo->ii_hsb1 = h_inode->i_sb; ++} ++ ++void au_cpup_attr_all(struct inode *inode, int force) ++{ ++ struct inode *h_inode; ++ ++ h_inode = au_h_iptr(inode, au_ibstart(inode)); ++ au_cpup_attr_changeable(inode); ++ if (inode->i_nlink > 0) ++ au_cpup_attr_nlink(inode, force); ++ inode->i_rdev = h_inode->i_rdev; ++ inode->i_blkbits = h_inode->i_blkbits; ++ au_cpup_igen(inode, h_inode); ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* Note: dt_dentry and dt_h_dentry are not dget/dput-ed */ ++ ++/* keep the timestamps of the parent dir when cpup */ ++void au_dtime_store(struct au_dtime *dt, struct dentry *dentry, ++ struct path *h_path) ++{ ++ struct inode *h_inode; ++ ++ dt->dt_dentry = dentry; ++ dt->dt_h_path = *h_path; ++ h_inode = h_path->dentry->d_inode; ++ dt->dt_atime = h_inode->i_atime; ++ dt->dt_mtime = h_inode->i_mtime; ++ /* smp_mb(); */ ++} ++ ++void au_dtime_revert(struct au_dtime *dt) ++{ ++ struct iattr attr; ++ int err; ++ ++ attr.ia_atime = dt->dt_atime; ++ attr.ia_mtime = dt->dt_mtime; ++ attr.ia_valid = ATTR_FORCE | ATTR_MTIME | ATTR_MTIME_SET ++ | ATTR_ATIME | ATTR_ATIME_SET; ++ ++ err = vfsub_notify_change(&dt->dt_h_path, &attr); ++ if (unlikely(err)) ++ AuWarn("restoring timestamps failed(%d). ignored\n", err); ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static noinline_for_stack ++int cpup_iattr(struct dentry *dst, aufs_bindex_t bindex, struct dentry *h_src) ++{ ++ int err, sbits; ++ struct iattr ia; ++ struct path h_path; ++ struct inode *h_isrc, *h_idst; ++ ++ h_path.dentry = au_h_dptr(dst, bindex); ++ h_idst = h_path.dentry->d_inode; ++ h_path.mnt = au_sbr_mnt(dst->d_sb, bindex); ++ h_isrc = h_src->d_inode; ++ ia.ia_valid = ATTR_FORCE | ATTR_UID | ATTR_GID ++ | ATTR_ATIME | ATTR_MTIME ++ | ATTR_ATIME_SET | ATTR_MTIME_SET; ++ ia.ia_uid = h_isrc->i_uid; ++ ia.ia_gid = h_isrc->i_gid; ++ ia.ia_atime = h_isrc->i_atime; ++ ia.ia_mtime = h_isrc->i_mtime; ++ if (h_idst->i_mode != h_isrc->i_mode ++ && !S_ISLNK(h_idst->i_mode)) { ++ ia.ia_valid |= ATTR_MODE; ++ ia.ia_mode = h_isrc->i_mode; ++ } ++ sbits = !!(h_isrc->i_mode & (S_ISUID | S_ISGID)); ++ au_cpup_attr_flags(h_idst, h_isrc); ++ err = vfsub_notify_change(&h_path, &ia); ++ ++ /* is this nfs only? */ ++ if (!err && sbits && au_test_nfs(h_path.dentry->d_sb)) { ++ ia.ia_valid = ATTR_FORCE | ATTR_MODE; ++ ia.ia_mode = h_isrc->i_mode; ++ err = vfsub_notify_change(&h_path, &ia); ++ } ++ ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static int au_do_copy_file(struct file *dst, struct file *src, loff_t len, ++ char *buf, unsigned long blksize) ++{ ++ int err; ++ size_t sz, rbytes, wbytes; ++ unsigned char all_zero; ++ char *p, *zp; ++ struct mutex *h_mtx; ++ /* reduce stack usage */ ++ struct iattr *ia; ++ ++ zp = page_address(ZERO_PAGE(0)); ++ if (unlikely(!zp)) ++ return -ENOMEM; /* possible? */ ++ ++ err = 0; ++ all_zero = 0; ++ while (len) { ++ AuDbg("len %lld\n", len); ++ sz = blksize; ++ if (len < blksize) ++ sz = len; ++ ++ rbytes = 0; ++ /* todo: signal_pending? */ ++ while (!rbytes || err == -EAGAIN || err == -EINTR) { ++ rbytes = vfsub_read_k(src, buf, sz, &src->f_pos); ++ err = rbytes; ++ } ++ if (unlikely(err < 0)) ++ break; ++ ++ all_zero = 0; ++ if (len >= rbytes && rbytes == blksize) ++ all_zero = !memcmp(buf, zp, rbytes); ++ if (!all_zero) { ++ wbytes = rbytes; ++ p = buf; ++ while (wbytes) { ++ size_t b; ++ ++ b = vfsub_write_k(dst, p, wbytes, &dst->f_pos); ++ err = b; ++ /* todo: signal_pending? */ ++ if (unlikely(err == -EAGAIN || err == -EINTR)) ++ continue; ++ if (unlikely(err < 0)) ++ break; ++ wbytes -= b; ++ p += b; ++ } ++ } else { ++ loff_t res; ++ ++ AuLabel(hole); ++ res = vfsub_llseek(dst, rbytes, SEEK_CUR); ++ err = res; ++ if (unlikely(res < 0)) ++ break; ++ } ++ len -= rbytes; ++ err = 0; ++ } ++ ++ /* the last block may be a hole */ ++ if (!err && all_zero) { ++ AuLabel(last hole); ++ ++ err = 1; ++ if (au_test_nfs(dst->f_dentry->d_sb)) { ++ /* nfs requires this step to make last hole */ ++ /* is this only nfs? */ ++ do { ++ /* todo: signal_pending? */ ++ err = vfsub_write_k(dst, "\0", 1, &dst->f_pos); ++ } while (err == -EAGAIN || err == -EINTR); ++ if (err == 1) ++ dst->f_pos--; ++ } ++ ++ if (err == 1) { ++ ia = (void *)buf; ++ ia->ia_size = dst->f_pos; ++ ia->ia_valid = ATTR_SIZE | ATTR_FILE; ++ ia->ia_file = dst; ++ h_mtx = &dst->f_dentry->d_inode->i_mutex; ++ mutex_lock_nested(h_mtx, AuLsc_I_CHILD2); ++ err = vfsub_notify_change(&dst->f_path, ia); ++ mutex_unlock(h_mtx); ++ } ++ } ++ ++ return err; ++} ++ ++int au_copy_file(struct file *dst, struct file *src, loff_t len) ++{ ++ int err; ++ unsigned long blksize; ++ unsigned char do_kfree; ++ char *buf; ++ ++ err = -ENOMEM; ++ blksize = dst->f_dentry->d_sb->s_blocksize; ++ if (!blksize || PAGE_SIZE < blksize) ++ blksize = PAGE_SIZE; ++ AuDbg("blksize %lu\n", blksize); ++ do_kfree = (blksize != PAGE_SIZE && blksize >= sizeof(struct iattr *)); ++ if (do_kfree) ++ buf = kmalloc(blksize, GFP_NOFS); ++ else ++ buf = (void *)__get_free_page(GFP_NOFS); ++ if (unlikely(!buf)) ++ goto out; ++ ++ if (len > (1 << 22)) ++ AuDbg("copying a large file %lld\n", (long long)len); ++ ++ src->f_pos = 0; ++ dst->f_pos = 0; ++ err = au_do_copy_file(dst, src, len, buf, blksize); ++ if (do_kfree) ++ kfree(buf); ++ else ++ free_page((unsigned long)buf); ++ ++ out: ++ return err; ++} ++ ++/* ++ * to support a sparse file which is opened with O_APPEND, ++ * we need to close the file. ++ */ ++static int au_cp_regular(struct dentry *dentry, aufs_bindex_t bdst, ++ aufs_bindex_t bsrc, loff_t len) ++{ ++ int err, i; ++ enum { SRC, DST }; ++ struct { ++ aufs_bindex_t bindex; ++ unsigned int flags; ++ struct dentry *dentry; ++ struct file *file; ++ void *label, *label_file; ++ } *f, file[] = { ++ { ++ .bindex = bsrc, ++ .flags = O_RDONLY | O_NOATIME | O_LARGEFILE, ++ .file = NULL, ++ .label = &&out, ++ .label_file = &&out_src ++ }, ++ { ++ .bindex = bdst, ++ .flags = O_WRONLY | O_NOATIME | O_LARGEFILE, ++ .file = NULL, ++ .label = &&out_src, ++ .label_file = &&out_dst ++ } ++ }; ++ struct super_block *sb; ++ ++ /* bsrc branch can be ro/rw. */ ++ sb = dentry->d_sb; ++ f = file; ++ for (i = 0; i < 2; i++, f++) { ++ f->dentry = au_h_dptr(dentry, f->bindex); ++ f->file = au_h_open(dentry, f->bindex, f->flags, /*file*/NULL); ++ err = PTR_ERR(f->file); ++ if (IS_ERR(f->file)) ++ goto *f->label; ++ err = -EINVAL; ++ if (unlikely(!f->file->f_op)) ++ goto *f->label_file; ++ } ++ ++ /* try stopping to update while we copyup */ ++ IMustLock(file[SRC].dentry->d_inode); ++ err = au_copy_file(file[DST].file, file[SRC].file, len); ++ ++ out_dst: ++ fput(file[DST].file); ++ au_sbr_put(sb, file[DST].bindex); ++ out_src: ++ fput(file[SRC].file); ++ au_sbr_put(sb, file[SRC].bindex); ++ out: ++ return err; ++} ++ ++static int au_do_cpup_regular(struct dentry *dentry, aufs_bindex_t bdst, ++ aufs_bindex_t bsrc, loff_t len, ++ struct inode *h_dir, struct path *h_path) ++{ ++ int err, rerr; ++ loff_t l; ++ ++ err = 0; ++ l = i_size_read(au_h_iptr(dentry->d_inode, bsrc)); ++ if (len == -1 || l < len) ++ len = l; ++ if (len) ++ err = au_cp_regular(dentry, bdst, bsrc, len); ++ if (!err) ++ goto out; /* success */ ++ ++ rerr = vfsub_unlink(h_dir, h_path, /*force*/0); ++ if (rerr) { ++ AuIOErr("failed unlinking cpup-ed %.*s(%d, %d)\n", ++ AuDLNPair(h_path->dentry), err, rerr); ++ err = -EIO; ++ } ++ ++ out: ++ return err; ++} ++ ++static int au_do_cpup_symlink(struct path *h_path, struct dentry *h_src, ++ struct inode *h_dir) ++{ ++ int err, symlen; ++ mm_segment_t old_fs; ++ char *sym; ++ ++ err = -ENOSYS; ++ if (unlikely(!h_src->d_inode->i_op->readlink)) ++ goto out; ++ ++ err = -ENOMEM; ++ sym = __getname(); ++ if (unlikely(!sym)) ++ goto out; ++ ++ old_fs = get_fs(); ++ set_fs(KERNEL_DS); ++ symlen = h_src->d_inode->i_op->readlink(h_src, (char __user *)sym, ++ PATH_MAX); ++ err = symlen; ++ set_fs(old_fs); ++ ++ if (symlen > 0) { ++ sym[symlen] = 0; ++ err = vfsub_symlink(h_dir, h_path, sym); ++ } ++ __putname(sym); ++ ++ out: ++ return err; ++} ++ ++/* return with the lower dst inode is locked */ ++static noinline_for_stack ++int cpup_entry(struct dentry *dentry, aufs_bindex_t bdst, ++ aufs_bindex_t bsrc, loff_t len, unsigned int flags, ++ struct dentry *dst_parent) ++{ ++ int err; ++ umode_t mode; ++ unsigned int mnt_flags; ++ unsigned char isdir; ++ const unsigned char do_dt = !!au_ftest_cpup(flags, DTIME); ++ struct au_dtime dt; ++ struct path h_path; ++ struct dentry *h_src, *h_dst, *h_parent; ++ struct inode *h_inode, *h_dir; ++ struct super_block *sb; ++ ++ /* bsrc branch can be ro/rw. */ ++ h_src = au_h_dptr(dentry, bsrc); ++ h_inode = h_src->d_inode; ++ AuDebugOn(h_inode != au_h_iptr(dentry->d_inode, bsrc)); ++ ++ /* try stopping to be referenced while we are creating */ ++ h_dst = au_h_dptr(dentry, bdst); ++ h_parent = h_dst->d_parent; /* dir inode is locked */ ++ h_dir = h_parent->d_inode; ++ IMustLock(h_dir); ++ AuDebugOn(h_parent != h_dst->d_parent); ++ ++ sb = dentry->d_sb; ++ h_path.mnt = au_sbr_mnt(sb, bdst); ++ if (do_dt) { ++ h_path.dentry = h_parent; ++ au_dtime_store(&dt, dst_parent, &h_path); ++ } ++ h_path.dentry = h_dst; ++ ++ isdir = 0; ++ mode = h_inode->i_mode; ++ switch (mode & S_IFMT) { ++ case S_IFREG: ++ /* try stopping to update while we are referencing */ ++ IMustLock(h_inode); ++ err = vfsub_create(h_dir, &h_path, mode | S_IWUSR); ++ if (!err) ++ err = au_do_cpup_regular ++ (dentry, bdst, bsrc, len, ++ au_h_iptr(dst_parent->d_inode, bdst), &h_path); ++ break; ++ case S_IFDIR: ++ isdir = 1; ++ err = vfsub_mkdir(h_dir, &h_path, mode); ++ if (!err) { ++ /* ++ * strange behaviour from the users view, ++ * particularry setattr case ++ */ ++ if (au_ibstart(dst_parent->d_inode) == bdst) ++ au_cpup_attr_nlink(dst_parent->d_inode, ++ /*force*/1); ++ au_cpup_attr_nlink(dentry->d_inode, /*force*/1); ++ } ++ break; ++ case S_IFLNK: ++ err = au_do_cpup_symlink(&h_path, h_src, h_dir); ++ break; ++ case S_IFCHR: ++ case S_IFBLK: ++ AuDebugOn(!capable(CAP_MKNOD)); ++ /*FALLTHROUGH*/ ++ case S_IFIFO: ++ case S_IFSOCK: ++ err = vfsub_mknod(h_dir, &h_path, mode, h_inode->i_rdev); ++ break; ++ default: ++ AuIOErr("Unknown inode type 0%o\n", mode); ++ err = -EIO; ++ } ++ ++ mnt_flags = au_mntflags(sb); ++ if (!au_opt_test(mnt_flags, UDBA_NONE) ++ && !isdir ++ && au_opt_test(mnt_flags, XINO) ++ && h_inode->i_nlink == 1 ++ /* todo: unnecessary? */ ++ /* && dentry->d_inode->i_nlink == 1 */ ++ && bdst < bsrc ++ && !au_ftest_cpup(flags, KEEPLINO)) ++ au_xino_write(sb, bsrc, h_inode->i_ino, /*ino*/0); ++ /* ignore this error */ ++ ++ if (do_dt) ++ au_dtime_revert(&dt); ++ return err; ++} ++ ++/* ++ * copyup the @dentry from @bsrc to @bdst. ++ * the caller must set the both of lower dentries. ++ * @len is for truncating when it is -1 copyup the entire file. ++ * in link/rename cases, @dst_parent may be different from the real one. ++ */ ++static int au_cpup_single(struct dentry *dentry, aufs_bindex_t bdst, ++ aufs_bindex_t bsrc, loff_t len, unsigned int flags, ++ struct dentry *dst_parent) ++{ ++ int err, rerr; ++ aufs_bindex_t old_ibstart; ++ unsigned char isdir, plink; ++ struct au_dtime dt; ++ struct path h_path; ++ struct dentry *h_src, *h_dst, *h_parent; ++ struct inode *dst_inode, *h_dir, *inode; ++ struct super_block *sb; ++ ++ AuDebugOn(bsrc <= bdst); ++ ++ sb = dentry->d_sb; ++ h_path.mnt = au_sbr_mnt(sb, bdst); ++ h_dst = au_h_dptr(dentry, bdst); ++ h_parent = h_dst->d_parent; /* dir inode is locked */ ++ h_dir = h_parent->d_inode; ++ IMustLock(h_dir); ++ ++ h_src = au_h_dptr(dentry, bsrc); ++ inode = dentry->d_inode; ++ ++ if (!dst_parent) ++ dst_parent = dget_parent(dentry); ++ else ++ dget(dst_parent); ++ ++ plink = !!au_opt_test(au_mntflags(sb), PLINK); ++ dst_inode = au_h_iptr(inode, bdst); ++ if (dst_inode) { ++ if (unlikely(!plink)) { ++ err = -EIO; ++ AuIOErr("i%lu exists on a upper branch " ++ "but plink is disabled\n", inode->i_ino); ++ goto out; ++ } ++ ++ if (dst_inode->i_nlink) { ++ const int do_dt = au_ftest_cpup(flags, DTIME); ++ ++ h_src = au_plink_lkup(inode, bdst); ++ err = PTR_ERR(h_src); ++ if (IS_ERR(h_src)) ++ goto out; ++ if (unlikely(!h_src->d_inode)) { ++ err = -EIO; ++ AuIOErr("i%lu exists on a upper branch " ++ "but plink is broken\n", inode->i_ino); ++ dput(h_src); ++ goto out; ++ } ++ ++ if (do_dt) { ++ h_path.dentry = h_parent; ++ au_dtime_store(&dt, dst_parent, &h_path); ++ } ++ h_path.dentry = h_dst; ++ err = vfsub_link(h_src, h_dir, &h_path); ++ if (do_dt) ++ au_dtime_revert(&dt); ++ dput(h_src); ++ goto out; ++ } else ++ /* todo: cpup_wh_file? */ ++ /* udba work */ ++ au_update_brange(inode, 1); ++ } ++ ++ old_ibstart = au_ibstart(inode); ++ err = cpup_entry(dentry, bdst, bsrc, len, flags, dst_parent); ++ if (unlikely(err)) ++ goto out; ++ dst_inode = h_dst->d_inode; ++ mutex_lock_nested(&dst_inode->i_mutex, AuLsc_I_CHILD2); ++ ++ err = cpup_iattr(dentry, bdst, h_src); ++ isdir = S_ISDIR(dst_inode->i_mode); ++ if (!err) { ++ if (bdst < old_ibstart) ++ au_set_ibstart(inode, bdst); ++ au_set_h_iptr(inode, bdst, au_igrab(dst_inode), ++ au_hi_flags(inode, isdir)); ++ mutex_unlock(&dst_inode->i_mutex); ++ if (!isdir ++ && h_src->d_inode->i_nlink > 1 ++ && plink) ++ au_plink_append(inode, bdst, h_dst); ++ goto out; /* success */ ++ } ++ ++ /* revert */ ++ h_path.dentry = h_parent; ++ mutex_unlock(&dst_inode->i_mutex); ++ au_dtime_store(&dt, dst_parent, &h_path); ++ h_path.dentry = h_dst; ++ if (!isdir) ++ rerr = vfsub_unlink(h_dir, &h_path, /*force*/0); ++ else ++ rerr = vfsub_rmdir(h_dir, &h_path); ++ au_dtime_revert(&dt); ++ if (rerr) { ++ AuIOErr("failed removing broken entry(%d, %d)\n", err, rerr); ++ err = -EIO; ++ } ++ ++ out: ++ dput(dst_parent); ++ return err; ++} ++ ++struct au_cpup_single_args { ++ int *errp; ++ struct dentry *dentry; ++ aufs_bindex_t bdst, bsrc; ++ loff_t len; ++ unsigned int flags; ++ struct dentry *dst_parent; ++}; ++ ++static void au_call_cpup_single(void *args) ++{ ++ struct au_cpup_single_args *a = args; ++ *a->errp = au_cpup_single(a->dentry, a->bdst, a->bsrc, a->len, ++ a->flags, a->dst_parent); ++} ++ ++int au_sio_cpup_single(struct dentry *dentry, aufs_bindex_t bdst, ++ aufs_bindex_t bsrc, loff_t len, unsigned int flags, ++ struct dentry *dst_parent) ++{ ++ int err, wkq_err; ++ umode_t mode; ++ struct dentry *h_dentry; ++ ++ h_dentry = au_h_dptr(dentry, bsrc); ++ mode = h_dentry->d_inode->i_mode & S_IFMT; ++ if ((mode != S_IFCHR && mode != S_IFBLK) ++ || capable(CAP_MKNOD)) ++ err = au_cpup_single(dentry, bdst, bsrc, len, flags, ++ dst_parent); ++ else { ++ struct au_cpup_single_args args = { ++ .errp = &err, ++ .dentry = dentry, ++ .bdst = bdst, ++ .bsrc = bsrc, ++ .len = len, ++ .flags = flags, ++ .dst_parent = dst_parent ++ }; ++ wkq_err = au_wkq_wait(au_call_cpup_single, &args); ++ if (unlikely(wkq_err)) ++ err = wkq_err; ++ } ++ ++ return err; ++} ++ ++/* ++ * copyup the @dentry from the first active lower branch to @bdst, ++ * using au_cpup_single(). ++ */ ++static int au_cpup_simple(struct dentry *dentry, aufs_bindex_t bdst, loff_t len, ++ unsigned int flags) ++{ ++ int err; ++ aufs_bindex_t bsrc, bend; ++ ++ bend = au_dbend(dentry); ++ for (bsrc = bdst + 1; bsrc <= bend; bsrc++) ++ if (au_h_dptr(dentry, bsrc)) ++ break; ++ ++ err = au_lkup_neg(dentry, bdst); ++ if (!err) { ++ err = au_cpup_single(dentry, bdst, bsrc, len, flags, NULL); ++ if (!err) ++ return 0; /* success */ ++ ++ /* revert */ ++ au_set_h_dptr(dentry, bdst, NULL); ++ au_set_dbstart(dentry, bsrc); ++ } ++ ++ return err; ++} ++ ++struct au_cpup_simple_args { ++ int *errp; ++ struct dentry *dentry; ++ aufs_bindex_t bdst; ++ loff_t len; ++ unsigned int flags; ++}; ++ ++static void au_call_cpup_simple(void *args) ++{ ++ struct au_cpup_simple_args *a = args; ++ *a->errp = au_cpup_simple(a->dentry, a->bdst, a->len, a->flags); ++} ++ ++int au_sio_cpup_simple(struct dentry *dentry, aufs_bindex_t bdst, loff_t len, ++ unsigned int flags) ++{ ++ int err, wkq_err; ++ unsigned char do_sio; ++ struct dentry *parent; ++ struct inode *h_dir; ++ ++ parent = dget_parent(dentry); ++ h_dir = au_h_iptr(parent->d_inode, bdst); ++ do_sio = !!au_test_h_perm_sio(h_dir, MAY_EXEC | MAY_WRITE); ++ if (!do_sio) { ++ /* ++ * testing CAP_MKNOD is for generic fs, ++ * but CAP_FSETID is for xfs only, currently. ++ */ ++ umode_t mode = dentry->d_inode->i_mode; ++ do_sio = (((mode & (S_IFCHR | S_IFBLK)) ++ && !capable(CAP_MKNOD)) ++ || ((mode & (S_ISUID | S_ISGID)) ++ && !capable(CAP_FSETID))); ++ } ++ if (!do_sio) ++ err = au_cpup_simple(dentry, bdst, len, flags); ++ else { ++ struct au_cpup_simple_args args = { ++ .errp = &err, ++ .dentry = dentry, ++ .bdst = bdst, ++ .len = len, ++ .flags = flags ++ }; ++ wkq_err = au_wkq_wait(au_call_cpup_simple, &args); ++ if (unlikely(wkq_err)) ++ err = wkq_err; ++ } ++ ++ dput(parent); ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* ++ * copyup the deleted file for writing. ++ */ ++static int au_do_cpup_wh(struct dentry *dentry, aufs_bindex_t bdst, ++ struct dentry *wh_dentry, struct file *file, ++ loff_t len) ++{ ++ int err; ++ aufs_bindex_t bstart; ++ struct au_dinfo *dinfo; ++ struct dentry *h_d_dst, *h_d_start; ++ ++ dinfo = au_di(dentry); ++ AuRwMustWriteLock(&dinfo->di_rwsem); ++ ++ bstart = dinfo->di_bstart; ++ h_d_dst = dinfo->di_hdentry[0 + bdst].hd_dentry; ++ dinfo->di_bstart = bdst; ++ dinfo->di_hdentry[0 + bdst].hd_dentry = wh_dentry; ++ h_d_start = dinfo->di_hdentry[0 + bstart].hd_dentry; ++ if (file) ++ dinfo->di_hdentry[0 + bstart].hd_dentry ++ = au_h_fptr(file, au_fbstart(file))->f_dentry; ++ err = au_cpup_single(dentry, bdst, bstart, len, !AuCpup_DTIME, ++ /*h_parent*/NULL); ++ if (!err && file) { ++ err = au_reopen_nondir(file); ++ dinfo->di_hdentry[0 + bstart].hd_dentry = h_d_start; ++ } ++ dinfo->di_hdentry[0 + bdst].hd_dentry = h_d_dst; ++ dinfo->di_bstart = bstart; ++ ++ return err; ++} ++ ++static int au_cpup_wh(struct dentry *dentry, aufs_bindex_t bdst, loff_t len, ++ struct file *file) ++{ ++ int err; ++ struct au_dtime dt; ++ struct dentry *parent, *h_parent, *wh_dentry; ++ struct au_branch *br; ++ struct path h_path; ++ ++ br = au_sbr(dentry->d_sb, bdst); ++ parent = dget_parent(dentry); ++ h_parent = au_h_dptr(parent, bdst); ++ wh_dentry = au_whtmp_lkup(h_parent, br, &dentry->d_name); ++ err = PTR_ERR(wh_dentry); ++ if (IS_ERR(wh_dentry)) ++ goto out; ++ ++ h_path.dentry = h_parent; ++ h_path.mnt = br->br_mnt; ++ au_dtime_store(&dt, parent, &h_path); ++ err = au_do_cpup_wh(dentry, bdst, wh_dentry, file, len); ++ if (unlikely(err)) ++ goto out_wh; ++ ++ dget(wh_dentry); ++ h_path.dentry = wh_dentry; ++ err = vfsub_unlink(h_parent->d_inode, &h_path, /*force*/0); ++ if (unlikely(err)) { ++ AuIOErr("failed remove copied-up tmp file %.*s(%d)\n", ++ AuDLNPair(wh_dentry), err); ++ err = -EIO; ++ } ++ au_dtime_revert(&dt); ++ au_set_hi_wh(dentry->d_inode, bdst, wh_dentry); ++ ++ out_wh: ++ dput(wh_dentry); ++ out: ++ dput(parent); ++ return err; ++} ++ ++struct au_cpup_wh_args { ++ int *errp; ++ struct dentry *dentry; ++ aufs_bindex_t bdst; ++ loff_t len; ++ struct file *file; ++}; ++ ++static void au_call_cpup_wh(void *args) ++{ ++ struct au_cpup_wh_args *a = args; ++ *a->errp = au_cpup_wh(a->dentry, a->bdst, a->len, a->file); ++} ++ ++int au_sio_cpup_wh(struct dentry *dentry, aufs_bindex_t bdst, loff_t len, ++ struct file *file) ++{ ++ int err, wkq_err; ++ struct dentry *parent, *h_orph, *h_parent, *h_dentry; ++ struct inode *dir, *h_dir, *h_tmpdir, *h_inode; ++ struct au_wbr *wbr; ++ ++ parent = dget_parent(dentry); ++ dir = parent->d_inode; ++ h_orph = NULL; ++ h_parent = NULL; ++ h_dir = au_igrab(au_h_iptr(dir, bdst)); ++ h_tmpdir = h_dir; ++ if (!h_dir->i_nlink) { ++ wbr = au_sbr(dentry->d_sb, bdst)->br_wbr; ++ h_orph = wbr->wbr_orph; ++ ++ h_parent = dget(au_h_dptr(parent, bdst)); ++ au_set_h_dptr(parent, bdst, NULL); ++ au_set_h_dptr(parent, bdst, dget(h_orph)); ++ h_tmpdir = h_orph->d_inode; ++ au_set_h_iptr(dir, bdst, NULL, 0); ++ au_set_h_iptr(dir, bdst, au_igrab(h_tmpdir), /*flags*/0); ++ ++ /* this temporary unlock is safe */ ++ if (file) ++ h_dentry = au_h_fptr(file, au_fbstart(file))->f_dentry; ++ else ++ h_dentry = au_h_dptr(dentry, au_dbstart(dentry)); ++ h_inode = h_dentry->d_inode; ++ IMustLock(h_inode); ++ mutex_unlock(&h_inode->i_mutex); ++ mutex_lock_nested(&h_tmpdir->i_mutex, AuLsc_I_PARENT3); ++ mutex_lock_nested(&h_inode->i_mutex, AuLsc_I_CHILD); ++ } ++ ++ if (!au_test_h_perm_sio(h_tmpdir, MAY_EXEC | MAY_WRITE)) ++ err = au_cpup_wh(dentry, bdst, len, file); ++ else { ++ struct au_cpup_wh_args args = { ++ .errp = &err, ++ .dentry = dentry, ++ .bdst = bdst, ++ .len = len, ++ .file = file ++ }; ++ wkq_err = au_wkq_wait(au_call_cpup_wh, &args); ++ if (unlikely(wkq_err)) ++ err = wkq_err; ++ } ++ ++ if (h_orph) { ++ mutex_unlock(&h_tmpdir->i_mutex); ++ au_set_h_iptr(dir, bdst, NULL, 0); ++ au_set_h_iptr(dir, bdst, au_igrab(h_dir), /*flags*/0); ++ au_set_h_dptr(parent, bdst, NULL); ++ au_set_h_dptr(parent, bdst, h_parent); ++ } ++ iput(h_dir); ++ dput(parent); ++ ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* ++ * generic routine for both of copy-up and copy-down. ++ */ ++/* cf. revalidate function in file.c */ ++int au_cp_dirs(struct dentry *dentry, aufs_bindex_t bdst, ++ int (*cp)(struct dentry *dentry, aufs_bindex_t bdst, ++ struct dentry *h_parent, void *arg), ++ void *arg) ++{ ++ int err; ++ struct au_pin pin; ++ struct dentry *d, *parent, *h_parent, *real_parent; ++ ++ err = 0; ++ parent = dget_parent(dentry); ++ if (IS_ROOT(parent)) ++ goto out; ++ ++ au_pin_init(&pin, dentry, bdst, AuLsc_DI_PARENT2, AuLsc_I_PARENT2, ++ au_opt_udba(dentry->d_sb), AuPin_MNT_WRITE); ++ ++ /* do not use au_dpage */ ++ real_parent = parent; ++ while (1) { ++ dput(parent); ++ parent = dget_parent(dentry); ++ h_parent = au_h_dptr(parent, bdst); ++ if (h_parent) ++ goto out; /* success */ ++ ++ /* find top dir which is necessary to cpup */ ++ do { ++ d = parent; ++ dput(parent); ++ parent = dget_parent(d); ++ di_read_lock_parent3(parent, !AuLock_IR); ++ h_parent = au_h_dptr(parent, bdst); ++ di_read_unlock(parent, !AuLock_IR); ++ } while (!h_parent); ++ ++ if (d != real_parent) ++ di_write_lock_child3(d); ++ ++ /* somebody else might create while we were sleeping */ ++ if (!au_h_dptr(d, bdst) || !au_h_dptr(d, bdst)->d_inode) { ++ if (au_h_dptr(d, bdst)) ++ au_update_dbstart(d); ++ ++ au_pin_set_dentry(&pin, d); ++ err = au_do_pin(&pin); ++ if (!err) { ++ err = cp(d, bdst, h_parent, arg); ++ au_unpin(&pin); ++ } ++ } ++ ++ if (d != real_parent) ++ di_write_unlock(d); ++ if (unlikely(err)) ++ break; ++ } ++ ++ out: ++ dput(parent); ++ return err; ++} ++ ++static int au_cpup_dir(struct dentry *dentry, aufs_bindex_t bdst, ++ struct dentry *h_parent __maybe_unused , ++ void *arg __maybe_unused) ++{ ++ return au_sio_cpup_simple(dentry, bdst, -1, AuCpup_DTIME); ++} ++ ++int au_cpup_dirs(struct dentry *dentry, aufs_bindex_t bdst) ++{ ++ return au_cp_dirs(dentry, bdst, au_cpup_dir, NULL); ++} ++ ++int au_test_and_cpup_dirs(struct dentry *dentry, aufs_bindex_t bdst) ++{ ++ int err; ++ struct dentry *parent; ++ struct inode *dir; ++ ++ parent = dget_parent(dentry); ++ dir = parent->d_inode; ++ err = 0; ++ if (au_h_iptr(dir, bdst)) ++ goto out; ++ ++ di_read_unlock(parent, AuLock_IR); ++ di_write_lock_parent(parent); ++ /* someone else might change our inode while we were sleeping */ ++ if (!au_h_iptr(dir, bdst)) ++ err = au_cpup_dirs(dentry, bdst); ++ di_downgrade_lock(parent, AuLock_IR); ++ ++ out: ++ dput(parent); ++ return err; ++} +diff -Nur linux-2.6.31.4.orig/fs/aufs/cpup.h linux-2.6.31.4/fs/aufs/cpup.h +--- linux-2.6.31.4.orig/fs/aufs/cpup.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/cpup.h 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,81 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * copy-up/down functions ++ */ ++ ++#ifndef __AUFS_CPUP_H__ ++#define __AUFS_CPUP_H__ ++ ++#ifdef __KERNEL__ ++ ++#include <linux/path.h> ++#include <linux/time.h> ++#include <linux/aufs_type.h> ++ ++struct inode; ++struct file; ++ ++void au_cpup_attr_flags(struct inode *dst, struct inode *src); ++void au_cpup_attr_timesizes(struct inode *inode); ++void au_cpup_attr_nlink(struct inode *inode, int force); ++void au_cpup_attr_changeable(struct inode *inode); ++void au_cpup_igen(struct inode *inode, struct inode *h_inode); ++void au_cpup_attr_all(struct inode *inode, int force); ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* cpup flags */ ++#define AuCpup_DTIME 1 /* do dtime_store/revert */ ++#define AuCpup_KEEPLINO (1 << 1) /* do not clear the lower xino, ++ for link(2) */ ++#define au_ftest_cpup(flags, name) ((flags) & AuCpup_##name) ++#define au_fset_cpup(flags, name) { (flags) |= AuCpup_##name; } ++#define au_fclr_cpup(flags, name) { (flags) &= ~AuCpup_##name; } ++ ++int au_copy_file(struct file *dst, struct file *src, loff_t len); ++int au_sio_cpup_single(struct dentry *dentry, aufs_bindex_t bdst, ++ aufs_bindex_t bsrc, loff_t len, unsigned int flags, ++ struct dentry *dst_parent); ++int au_sio_cpup_simple(struct dentry *dentry, aufs_bindex_t bdst, loff_t len, ++ unsigned int flags); ++int au_sio_cpup_wh(struct dentry *dentry, aufs_bindex_t bdst, loff_t len, ++ struct file *file); ++ ++int au_cp_dirs(struct dentry *dentry, aufs_bindex_t bdst, ++ int (*cp)(struct dentry *dentry, aufs_bindex_t bdst, ++ struct dentry *h_parent, void *arg), ++ void *arg); ++int au_cpup_dirs(struct dentry *dentry, aufs_bindex_t bdst); ++int au_test_and_cpup_dirs(struct dentry *dentry, aufs_bindex_t bdst); ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* keep timestamps when copyup */ ++struct au_dtime { ++ struct dentry *dt_dentry; ++ struct path dt_h_path; ++ struct timespec dt_atime, dt_mtime; ++}; ++void au_dtime_store(struct au_dtime *dt, struct dentry *dentry, ++ struct path *h_path); ++void au_dtime_revert(struct au_dtime *dt); ++ ++#endif /* __KERNEL__ */ ++#endif /* __AUFS_CPUP_H__ */ +diff -Nur linux-2.6.31.4.orig/fs/aufs/dbgaufs.c linux-2.6.31.4/fs/aufs/dbgaufs.c +--- linux-2.6.31.4.orig/fs/aufs/dbgaufs.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/dbgaufs.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,331 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * debugfs interface ++ */ ++ ++#include <linux/debugfs.h> ++#include "aufs.h" ++ ++#ifndef CONFIG_SYSFS ++#error DEBUG_FS depends upon SYSFS ++#endif ++ ++static struct dentry *dbgaufs; ++static const mode_t dbgaufs_mode = S_IRUSR | S_IRGRP | S_IROTH; ++ ++/* 20 is max digits length of ulong 64 */ ++struct dbgaufs_arg { ++ int n; ++ char a[20 * 4]; ++}; ++ ++/* ++ * common function for all XINO files ++ */ ++static int dbgaufs_xi_release(struct inode *inode __maybe_unused, ++ struct file *file) ++{ ++ kfree(file->private_data); ++ return 0; ++} ++ ++static int dbgaufs_xi_open(struct file *xf, struct file *file, int do_fcnt) ++{ ++ int err; ++ struct kstat st; ++ struct dbgaufs_arg *p; ++ ++ err = -ENOMEM; ++ p = kmalloc(sizeof(*p), GFP_NOFS); ++ if (unlikely(!p)) ++ goto out; ++ ++ err = 0; ++ p->n = 0; ++ file->private_data = p; ++ if (!xf) ++ goto out; ++ ++ err = vfs_getattr(xf->f_vfsmnt, xf->f_dentry, &st); ++ if (!err) { ++ if (do_fcnt) ++ p->n = snprintf ++ (p->a, sizeof(p->a), "%ld, %llux%lu %lld\n", ++ (long)file_count(xf), st.blocks, st.blksize, ++ (long long)st.size); ++ else ++ p->n = snprintf(p->a, sizeof(p->a), "%llux%lu %lld\n", ++ st.blocks, st.blksize, ++ (long long)st.size); ++ AuDebugOn(p->n >= sizeof(p->a)); ++ } else { ++ p->n = snprintf(p->a, sizeof(p->a), "err %d\n", err); ++ err = 0; ++ } ++ ++ out: ++ return err; ++ ++} ++ ++static ssize_t dbgaufs_xi_read(struct file *file, char __user *buf, ++ size_t count, loff_t *ppos) ++{ ++ struct dbgaufs_arg *p; ++ ++ p = file->private_data; ++ return simple_read_from_buffer(buf, count, ppos, p->a, p->n); ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static int dbgaufs_xib_open(struct inode *inode, struct file *file) ++{ ++ int err; ++ struct au_sbinfo *sbinfo; ++ struct super_block *sb; ++ ++ sbinfo = inode->i_private; ++ sb = sbinfo->si_sb; ++ si_noflush_read_lock(sb); ++ err = dbgaufs_xi_open(sbinfo->si_xib, file, /*do_fcnt*/0); ++ si_read_unlock(sb); ++ return err; ++} ++ ++static const struct file_operations dbgaufs_xib_fop = { ++ .open = dbgaufs_xib_open, ++ .release = dbgaufs_xi_release, ++ .read = dbgaufs_xi_read ++}; ++ ++/* ---------------------------------------------------------------------- */ ++ ++#define DbgaufsXi_PREFIX "xi" ++ ++static int dbgaufs_xino_open(struct inode *inode, struct file *file) ++{ ++ int err; ++ long l; ++ struct au_sbinfo *sbinfo; ++ struct super_block *sb; ++ struct file *xf; ++ struct qstr *name; ++ ++ err = -ENOENT; ++ xf = NULL; ++ name = &file->f_dentry->d_name; ++ if (unlikely(name->len < sizeof(DbgaufsXi_PREFIX) ++ || memcmp(name->name, DbgaufsXi_PREFIX, ++ sizeof(DbgaufsXi_PREFIX) - 1))) ++ goto out; ++ err = strict_strtol(name->name + sizeof(DbgaufsXi_PREFIX) - 1, 10, &l); ++ if (unlikely(err)) ++ goto out; ++ ++ sbinfo = inode->i_private; ++ sb = sbinfo->si_sb; ++ si_noflush_read_lock(sb); ++ if (l <= au_sbend(sb)) { ++ xf = au_sbr(sb, (aufs_bindex_t)l)->br_xino.xi_file; ++ err = dbgaufs_xi_open(xf, file, /*do_fcnt*/1); ++ } else ++ err = -ENOENT; ++ si_read_unlock(sb); ++ ++ out: ++ return err; ++} ++ ++static const struct file_operations dbgaufs_xino_fop = { ++ .open = dbgaufs_xino_open, ++ .release = dbgaufs_xi_release, ++ .read = dbgaufs_xi_read ++}; ++ ++void dbgaufs_brs_del(struct super_block *sb, aufs_bindex_t bindex) ++{ ++ aufs_bindex_t bend; ++ struct au_branch *br; ++ struct au_xino_file *xi; ++ ++ if (!au_sbi(sb)->si_dbgaufs) ++ return; ++ ++ bend = au_sbend(sb); ++ for (; bindex <= bend; bindex++) { ++ br = au_sbr(sb, bindex); ++ xi = &br->br_xino; ++ if (xi->xi_dbgaufs) { ++ debugfs_remove(xi->xi_dbgaufs); ++ xi->xi_dbgaufs = NULL; ++ } ++ } ++} ++ ++void dbgaufs_brs_add(struct super_block *sb, aufs_bindex_t bindex) ++{ ++ struct au_sbinfo *sbinfo; ++ struct dentry *parent; ++ struct au_branch *br; ++ struct au_xino_file *xi; ++ aufs_bindex_t bend; ++ char name[sizeof(DbgaufsXi_PREFIX) + 5]; /* "xi" bindex NULL */ ++ ++ sbinfo = au_sbi(sb); ++ parent = sbinfo->si_dbgaufs; ++ if (!parent) ++ return; ++ ++ bend = au_sbend(sb); ++ for (; bindex <= bend; bindex++) { ++ snprintf(name, sizeof(name), DbgaufsXi_PREFIX "%d", bindex); ++ br = au_sbr(sb, bindex); ++ xi = &br->br_xino; ++ AuDebugOn(xi->xi_dbgaufs); ++ xi->xi_dbgaufs = debugfs_create_file(name, dbgaufs_mode, parent, ++ sbinfo, &dbgaufs_xino_fop); ++ /* ignore an error */ ++ if (unlikely(!xi->xi_dbgaufs)) ++ AuWarn1("failed %s under debugfs\n", name); ++ } ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++#ifdef CONFIG_AUFS_EXPORT ++static int dbgaufs_xigen_open(struct inode *inode, struct file *file) ++{ ++ int err; ++ struct au_sbinfo *sbinfo; ++ struct super_block *sb; ++ ++ sbinfo = inode->i_private; ++ sb = sbinfo->si_sb; ++ si_noflush_read_lock(sb); ++ err = dbgaufs_xi_open(sbinfo->si_xigen, file, /*do_fcnt*/0); ++ si_read_unlock(sb); ++ return err; ++} ++ ++static const struct file_operations dbgaufs_xigen_fop = { ++ .open = dbgaufs_xigen_open, ++ .release = dbgaufs_xi_release, ++ .read = dbgaufs_xi_read ++}; ++ ++static int dbgaufs_xigen_init(struct au_sbinfo *sbinfo) ++{ ++ int err; ++ ++ /* ++ * This function is a dynamic '__init' fucntion actually, ++ * so the tiny check for si_rwsem is unnecessary. ++ */ ++ /* AuRwMustWriteLock(&sbinfo->si_rwsem); */ ++ ++ err = -EIO; ++ sbinfo->si_dbgaufs_xigen = debugfs_create_file ++ ("xigen", dbgaufs_mode, sbinfo->si_dbgaufs, sbinfo, ++ &dbgaufs_xigen_fop); ++ if (sbinfo->si_dbgaufs_xigen) ++ err = 0; ++ ++ return err; ++} ++#else ++static int dbgaufs_xigen_init(struct au_sbinfo *sbinfo) ++{ ++ return 0; ++} ++#endif /* CONFIG_AUFS_EXPORT */ ++ ++/* ---------------------------------------------------------------------- */ ++ ++void dbgaufs_si_fin(struct au_sbinfo *sbinfo) ++{ ++ /* ++ * This function is a dynamic '__init' fucntion actually, ++ * so the tiny check for si_rwsem is unnecessary. ++ */ ++ /* AuRwMustWriteLock(&sbinfo->si_rwsem); */ ++ ++ debugfs_remove_recursive(sbinfo->si_dbgaufs); ++ sbinfo->si_dbgaufs = NULL; ++ kobject_put(&sbinfo->si_kobj); ++} ++ ++int dbgaufs_si_init(struct au_sbinfo *sbinfo) ++{ ++ int err; ++ char name[SysaufsSiNameLen]; ++ ++ /* ++ * This function is a dynamic '__init' fucntion actually, ++ * so the tiny check for si_rwsem is unnecessary. ++ */ ++ /* AuRwMustWriteLock(&sbinfo->si_rwsem); */ ++ ++ err = -ENOENT; ++ if (!dbgaufs) { ++ AuErr1("/debug/aufs is uninitialized\n"); ++ goto out; ++ } ++ ++ err = -EIO; ++ sysaufs_name(sbinfo, name); ++ sbinfo->si_dbgaufs = debugfs_create_dir(name, dbgaufs); ++ if (unlikely(!sbinfo->si_dbgaufs)) ++ goto out; ++ kobject_get(&sbinfo->si_kobj); ++ ++ sbinfo->si_dbgaufs_xib = debugfs_create_file ++ ("xib", dbgaufs_mode, sbinfo->si_dbgaufs, sbinfo, ++ &dbgaufs_xib_fop); ++ if (unlikely(!sbinfo->si_dbgaufs_xib)) ++ goto out_dir; ++ ++ err = dbgaufs_xigen_init(sbinfo); ++ if (!err) ++ goto out; /* success */ ++ ++ out_dir: ++ dbgaufs_si_fin(sbinfo); ++ out: ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++void dbgaufs_fin(void) ++{ ++ debugfs_remove(dbgaufs); ++} ++ ++int __init dbgaufs_init(void) ++{ ++ int err; ++ ++ err = -EIO; ++ dbgaufs = debugfs_create_dir(AUFS_NAME, NULL); ++ if (dbgaufs) ++ err = 0; ++ return err; ++} +diff -Nur linux-2.6.31.4.orig/fs/aufs/dbgaufs.h linux-2.6.31.4/fs/aufs/dbgaufs.h +--- linux-2.6.31.4.orig/fs/aufs/dbgaufs.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/dbgaufs.h 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,79 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * debugfs interface ++ */ ++ ++#ifndef __DBGAUFS_H__ ++#define __DBGAUFS_H__ ++ ++#ifdef __KERNEL__ ++ ++#include <linux/init.h> ++#include <linux/aufs_type.h> ++ ++struct super_block; ++struct au_sbinfo; ++ ++#ifdef CONFIG_DEBUG_FS ++/* dbgaufs.c */ ++void dbgaufs_brs_del(struct super_block *sb, aufs_bindex_t bindex); ++void dbgaufs_brs_add(struct super_block *sb, aufs_bindex_t bindex); ++void dbgaufs_si_fin(struct au_sbinfo *sbinfo); ++int dbgaufs_si_init(struct au_sbinfo *sbinfo); ++void dbgaufs_fin(void); ++int __init dbgaufs_init(void); ++ ++#else ++ ++static inline ++void dbgaufs_brs_del(struct super_block *sb, aufs_bindex_t bindex) ++{ ++ /* empty */ ++} ++ ++static inline ++void dbgaufs_brs_add(struct super_block *sb, aufs_bindex_t bindex) ++{ ++ /* empty */ ++} ++ ++static inline ++void dbgaufs_si_fin(struct au_sbinfo *sbinfo) ++{ ++ /* empty */ ++} ++ ++static inline ++int dbgaufs_si_init(struct au_sbinfo *sbinfo) ++{ ++ return 0; ++} ++ ++#define dbgaufs_fin() do {} while (0) ++ ++static inline ++int __init dbgaufs_init(void) ++{ ++ return 0; ++} ++#endif /* CONFIG_DEBUG_FS */ ++ ++#endif /* __KERNEL__ */ ++#endif /* __DBGAUFS_H__ */ +diff -Nur linux-2.6.31.4.orig/fs/aufs/dcsub.c linux-2.6.31.4/fs/aufs/dcsub.c +--- linux-2.6.31.4.orig/fs/aufs/dcsub.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/dcsub.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,223 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * sub-routines for dentry cache ++ */ ++ ++#include "aufs.h" ++ ++static void au_dpage_free(struct au_dpage *dpage) ++{ ++ int i; ++ struct dentry **p; ++ ++ p = dpage->dentries; ++ for (i = 0; i < dpage->ndentry; i++) ++ dput(*p++); ++ free_page((unsigned long)dpage->dentries); ++} ++ ++int au_dpages_init(struct au_dcsub_pages *dpages, gfp_t gfp) ++{ ++ int err; ++ void *p; ++ ++ err = -ENOMEM; ++ dpages->dpages = kmalloc(sizeof(*dpages->dpages), gfp); ++ if (unlikely(!dpages->dpages)) ++ goto out; ++ ++ p = (void *)__get_free_page(gfp); ++ if (unlikely(!p)) ++ goto out_dpages; ++ ++ dpages->dpages[0].ndentry = 0; ++ dpages->dpages[0].dentries = p; ++ dpages->ndpage = 1; ++ return 0; /* success */ ++ ++ out_dpages: ++ kfree(dpages->dpages); ++ out: ++ return err; ++} ++ ++void au_dpages_free(struct au_dcsub_pages *dpages) ++{ ++ int i; ++ struct au_dpage *p; ++ ++ p = dpages->dpages; ++ for (i = 0; i < dpages->ndpage; i++) ++ au_dpage_free(p++); ++ kfree(dpages->dpages); ++} ++ ++static int au_dpages_append(struct au_dcsub_pages *dpages, ++ struct dentry *dentry, gfp_t gfp) ++{ ++ int err, sz; ++ struct au_dpage *dpage; ++ void *p; ++ ++ dpage = dpages->dpages + dpages->ndpage - 1; ++ sz = PAGE_SIZE / sizeof(dentry); ++ if (unlikely(dpage->ndentry >= sz)) { ++ AuLabel(new dpage); ++ err = -ENOMEM; ++ sz = dpages->ndpage * sizeof(*dpages->dpages); ++ p = au_kzrealloc(dpages->dpages, sz, ++ sz + sizeof(*dpages->dpages), gfp); ++ if (unlikely(!p)) ++ goto out; ++ ++ dpages->dpages = p; ++ dpage = dpages->dpages + dpages->ndpage; ++ p = (void *)__get_free_page(gfp); ++ if (unlikely(!p)) ++ goto out; ++ ++ dpage->ndentry = 0; ++ dpage->dentries = p; ++ dpages->ndpage++; ++ } ++ ++ dpage->dentries[dpage->ndentry++] = dget(dentry); ++ return 0; /* success */ ++ ++ out: ++ return err; ++} ++ ++int au_dcsub_pages(struct au_dcsub_pages *dpages, struct dentry *root, ++ au_dpages_test test, void *arg) ++{ ++ int err; ++ struct dentry *this_parent = root; ++ struct list_head *next; ++ struct super_block *sb = root->d_sb; ++ ++ err = 0; ++ spin_lock(&dcache_lock); ++ repeat: ++ next = this_parent->d_subdirs.next; ++ resume: ++ if (this_parent->d_sb == sb ++ && !IS_ROOT(this_parent) ++ && atomic_read(&this_parent->d_count) ++ && this_parent->d_inode ++ && (!test || test(this_parent, arg))) { ++ err = au_dpages_append(dpages, this_parent, GFP_ATOMIC); ++ if (unlikely(err)) ++ goto out; ++ } ++ ++ while (next != &this_parent->d_subdirs) { ++ struct list_head *tmp = next; ++ struct dentry *dentry = list_entry(tmp, struct dentry, ++ d_u.d_child); ++ next = tmp->next; ++ if (/*d_unhashed(dentry) || */!dentry->d_inode) ++ continue; ++ if (!list_empty(&dentry->d_subdirs)) { ++ this_parent = dentry; ++ goto repeat; ++ } ++ if (dentry->d_sb == sb ++ && atomic_read(&dentry->d_count) ++ && (!test || test(dentry, arg))) { ++ err = au_dpages_append(dpages, dentry, GFP_ATOMIC); ++ if (unlikely(err)) ++ goto out; ++ } ++ } ++ ++ if (this_parent != root) { ++ next = this_parent->d_u.d_child.next; ++ this_parent = this_parent->d_parent; /* dcache_lock is locked */ ++ goto resume; ++ } ++ out: ++ spin_unlock(&dcache_lock); ++ return err; ++} ++ ++int au_dcsub_pages_rev(struct au_dcsub_pages *dpages, struct dentry *dentry, ++ int do_include, au_dpages_test test, void *arg) ++{ ++ int err; ++ ++ err = 0; ++ spin_lock(&dcache_lock); ++ if (do_include && (!test || test(dentry, arg))) { ++ err = au_dpages_append(dpages, dentry, GFP_ATOMIC); ++ if (unlikely(err)) ++ goto out; ++ } ++ while (!IS_ROOT(dentry)) { ++ dentry = dentry->d_parent; /* dcache_lock is locked */ ++ if (!test || test(dentry, arg)) { ++ err = au_dpages_append(dpages, dentry, GFP_ATOMIC); ++ if (unlikely(err)) ++ break; ++ } ++ } ++ ++ out: ++ spin_unlock(&dcache_lock); ++ ++ return err; ++} ++ ++struct dentry *au_test_subdir(struct dentry *d1, struct dentry *d2) ++{ ++ struct dentry *trap, **dentries; ++ int err, i, j; ++ struct au_dcsub_pages dpages; ++ struct au_dpage *dpage; ++ ++ trap = ERR_PTR(-ENOMEM); ++ err = au_dpages_init(&dpages, GFP_NOFS); ++ if (unlikely(err)) ++ goto out; ++ err = au_dcsub_pages_rev(&dpages, d1, /*do_include*/1, NULL, NULL); ++ if (unlikely(err)) ++ goto out_dpages; ++ ++ trap = d1; ++ for (i = 0; !err && i < dpages.ndpage; i++) { ++ dpage = dpages.dpages + i; ++ dentries = dpage->dentries; ++ for (j = 0; !err && j < dpage->ndentry; j++) { ++ struct dentry *d; ++ ++ d = dentries[j]; ++ err = (d == d2); ++ if (!err) ++ trap = d; ++ } ++ } ++ if (!err) ++ trap = NULL; ++ ++ out_dpages: ++ au_dpages_free(&dpages); ++ out: ++ return trap; ++} +diff -Nur linux-2.6.31.4.orig/fs/aufs/dcsub.h linux-2.6.31.4/fs/aufs/dcsub.h +--- linux-2.6.31.4.orig/fs/aufs/dcsub.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/dcsub.h 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,54 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * sub-routines for dentry cache ++ */ ++ ++#ifndef __AUFS_DCSUB_H__ ++#define __AUFS_DCSUB_H__ ++ ++#ifdef __KERNEL__ ++ ++#include <linux/types.h> ++ ++struct dentry; ++ ++struct au_dpage { ++ int ndentry; ++ struct dentry **dentries; ++}; ++ ++struct au_dcsub_pages { ++ int ndpage; ++ struct au_dpage *dpages; ++}; ++ ++/* ---------------------------------------------------------------------- */ ++ ++int au_dpages_init(struct au_dcsub_pages *dpages, gfp_t gfp); ++void au_dpages_free(struct au_dcsub_pages *dpages); ++typedef int (*au_dpages_test)(struct dentry *dentry, void *arg); ++int au_dcsub_pages(struct au_dcsub_pages *dpages, struct dentry *root, ++ au_dpages_test test, void *arg); ++int au_dcsub_pages_rev(struct au_dcsub_pages *dpages, struct dentry *dentry, ++ int do_include, au_dpages_test test, void *arg); ++struct dentry *au_test_subdir(struct dentry *d1, struct dentry *d2); ++ ++#endif /* __KERNEL__ */ ++#endif /* __AUFS_DCSUB_H__ */ +diff -Nur linux-2.6.31.4.orig/fs/aufs/debug.c linux-2.6.31.4/fs/aufs/debug.c +--- linux-2.6.31.4.orig/fs/aufs/debug.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/debug.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,427 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * debug print functions ++ */ ++ ++#include <linux/module.h> ++#include <linux/vt_kern.h> ++#include "aufs.h" ++ ++int aufs_debug; ++MODULE_PARM_DESC(debug, "debug print"); ++module_param_named(debug, aufs_debug, int, S_IRUGO | S_IWUSR | S_IWGRP); ++ ++char *au_plevel = KERN_DEBUG; ++#define dpri(fmt, arg...) do { \ ++ if (au_debug_test()) \ ++ printk("%s" fmt, au_plevel, ##arg); \ ++} while (0) ++ ++/* ---------------------------------------------------------------------- */ ++ ++void au_dpri_whlist(struct au_nhash *whlist) ++{ ++ unsigned long ul, n; ++ struct hlist_head *head; ++ struct au_vdir_wh *tpos; ++ struct hlist_node *pos; ++ ++ n = whlist->nh_num; ++ head = whlist->nh_head; ++ for (ul = 0; ul < n; ul++) { ++ hlist_for_each_entry(tpos, pos, head, wh_hash) ++ dpri("b%d, %.*s, %d\n", ++ tpos->wh_bindex, ++ tpos->wh_str.len, tpos->wh_str.name, ++ tpos->wh_str.len); ++ head++; ++ } ++} ++ ++void au_dpri_vdir(struct au_vdir *vdir) ++{ ++ unsigned long ul; ++ union au_vdir_deblk_p p; ++ unsigned char *o; ++ ++ if (!vdir || IS_ERR(vdir)) { ++ dpri("err %ld\n", PTR_ERR(vdir)); ++ return; ++ } ++ ++ dpri("deblk %u, nblk %lu, deblk %p, last{%lu, %p}, ver %lu\n", ++ vdir->vd_deblk_sz, vdir->vd_nblk, vdir->vd_deblk, ++ vdir->vd_last.ul, vdir->vd_last.p.deblk, vdir->vd_version); ++ for (ul = 0; ul < vdir->vd_nblk; ul++) { ++ p.deblk = vdir->vd_deblk[ul]; ++ o = p.deblk; ++ dpri("[%lu]: %p\n", ul, o); ++ } ++} ++ ++static int do_pri_inode(aufs_bindex_t bindex, struct inode *inode, ++ struct dentry *wh) ++{ ++ char *n = NULL; ++ int l = 0; ++ ++ if (!inode || IS_ERR(inode)) { ++ dpri("i%d: err %ld\n", bindex, PTR_ERR(inode)); ++ return -1; ++ } ++ ++ /* the type of i_blocks depends upon CONFIG_LSF */ ++ BUILD_BUG_ON(sizeof(inode->i_blocks) != sizeof(unsigned long) ++ && sizeof(inode->i_blocks) != sizeof(u64)); ++ if (wh) { ++ n = (void *)wh->d_name.name; ++ l = wh->d_name.len; ++ } ++ ++ dpri("i%d: i%lu, %s, cnt %d, nl %u, 0%o, sz %llu, blk %llu," ++ " ct %lld, np %lu, st 0x%lx, f 0x%x, g %x%s%.*s\n", ++ bindex, ++ inode->i_ino, inode->i_sb ? au_sbtype(inode->i_sb) : "??", ++ atomic_read(&inode->i_count), inode->i_nlink, inode->i_mode, ++ i_size_read(inode), (unsigned long long)inode->i_blocks, ++ (long long)timespec_to_ns(&inode->i_ctime) & 0x0ffff, ++ inode->i_mapping ? inode->i_mapping->nrpages : 0, ++ inode->i_state, inode->i_flags, inode->i_generation, ++ l ? ", wh " : "", l, n); ++ return 0; ++} ++ ++void au_dpri_inode(struct inode *inode) ++{ ++ struct au_iinfo *iinfo; ++ aufs_bindex_t bindex; ++ int err; ++ ++ err = do_pri_inode(-1, inode, NULL); ++ if (err || !au_test_aufs(inode->i_sb)) ++ return; ++ ++ iinfo = au_ii(inode); ++ if (!iinfo) ++ return; ++ dpri("i-1: bstart %d, bend %d, gen %d\n", ++ iinfo->ii_bstart, iinfo->ii_bend, au_iigen(inode)); ++ if (iinfo->ii_bstart < 0) ++ return; ++ for (bindex = iinfo->ii_bstart; bindex <= iinfo->ii_bend; bindex++) ++ do_pri_inode(bindex, iinfo->ii_hinode[0 + bindex].hi_inode, ++ iinfo->ii_hinode[0 + bindex].hi_whdentry); ++} ++ ++static int do_pri_dentry(aufs_bindex_t bindex, struct dentry *dentry) ++{ ++ struct dentry *wh = NULL; ++ ++ if (!dentry || IS_ERR(dentry)) { ++ dpri("d%d: err %ld\n", bindex, PTR_ERR(dentry)); ++ return -1; ++ } ++ /* do not call dget_parent() here */ ++ dpri("d%d: %.*s?/%.*s, %s, cnt %d, flags 0x%x\n", ++ bindex, ++ AuDLNPair(dentry->d_parent), AuDLNPair(dentry), ++ dentry->d_sb ? au_sbtype(dentry->d_sb) : "??", ++ atomic_read(&dentry->d_count), dentry->d_flags); ++ if (bindex >= 0 && dentry->d_inode && au_test_aufs(dentry->d_sb)) { ++ struct au_iinfo *iinfo = au_ii(dentry->d_inode); ++ if (iinfo) ++ wh = iinfo->ii_hinode[0 + bindex].hi_whdentry; ++ } ++ do_pri_inode(bindex, dentry->d_inode, wh); ++ return 0; ++} ++ ++void au_dpri_dentry(struct dentry *dentry) ++{ ++ struct au_dinfo *dinfo; ++ aufs_bindex_t bindex; ++ int err; ++ ++ err = do_pri_dentry(-1, dentry); ++ if (err || !au_test_aufs(dentry->d_sb)) ++ return; ++ ++ dinfo = au_di(dentry); ++ if (!dinfo) ++ return; ++ dpri("d-1: bstart %d, bend %d, bwh %d, bdiropq %d, gen %d\n", ++ dinfo->di_bstart, dinfo->di_bend, ++ dinfo->di_bwh, dinfo->di_bdiropq, au_digen(dentry)); ++ if (dinfo->di_bstart < 0) ++ return; ++ for (bindex = dinfo->di_bstart; bindex <= dinfo->di_bend; bindex++) ++ do_pri_dentry(bindex, dinfo->di_hdentry[0 + bindex].hd_dentry); ++} ++ ++static int do_pri_file(aufs_bindex_t bindex, struct file *file) ++{ ++ char a[32]; ++ ++ if (!file || IS_ERR(file)) { ++ dpri("f%d: err %ld\n", bindex, PTR_ERR(file)); ++ return -1; ++ } ++ a[0] = 0; ++ if (bindex < 0 ++ && file->f_dentry ++ && au_test_aufs(file->f_dentry->d_sb) ++ && au_fi(file)) ++ snprintf(a, sizeof(a), ", mmapped %d", au_test_mmapped(file)); ++ dpri("f%d: mode 0x%x, flags 0%o, cnt %ld, pos %llu%s\n", ++ bindex, file->f_mode, file->f_flags, (long)file_count(file), ++ file->f_pos, a); ++ if (file->f_dentry) ++ do_pri_dentry(bindex, file->f_dentry); ++ return 0; ++} ++ ++void au_dpri_file(struct file *file) ++{ ++ struct au_finfo *finfo; ++ aufs_bindex_t bindex; ++ int err; ++ ++ err = do_pri_file(-1, file); ++ if (err || !file->f_dentry || !au_test_aufs(file->f_dentry->d_sb)) ++ return; ++ ++ finfo = au_fi(file); ++ if (!finfo) ++ return; ++ if (finfo->fi_bstart < 0) ++ return; ++ for (bindex = finfo->fi_bstart; bindex <= finfo->fi_bend; bindex++) { ++ struct au_hfile *hf; ++ ++ hf = finfo->fi_hfile + bindex; ++ do_pri_file(bindex, hf ? hf->hf_file : NULL); ++ } ++} ++ ++static int do_pri_br(aufs_bindex_t bindex, struct au_branch *br) ++{ ++ struct vfsmount *mnt; ++ struct super_block *sb; ++ ++ if (!br || IS_ERR(br)) ++ goto out; ++ mnt = br->br_mnt; ++ if (!mnt || IS_ERR(mnt)) ++ goto out; ++ sb = mnt->mnt_sb; ++ if (!sb || IS_ERR(sb)) ++ goto out; ++ ++ dpri("s%d: {perm 0x%x, cnt %d, wbr %p}, " ++ "%s, dev 0x%02x%02x, flags 0x%lx, cnt(BIAS) %d, active %d, " ++ "xino %d\n", ++ bindex, br->br_perm, atomic_read(&br->br_count), br->br_wbr, ++ au_sbtype(sb), MAJOR(sb->s_dev), MINOR(sb->s_dev), ++ sb->s_flags, sb->s_count - S_BIAS, ++ atomic_read(&sb->s_active), !!br->br_xino.xi_file); ++ return 0; ++ ++ out: ++ dpri("s%d: err %ld\n", bindex, PTR_ERR(br)); ++ return -1; ++} ++ ++void au_dpri_sb(struct super_block *sb) ++{ ++ struct au_sbinfo *sbinfo; ++ aufs_bindex_t bindex; ++ int err; ++ /* to reuduce stack size */ ++ struct { ++ struct vfsmount mnt; ++ struct au_branch fake; ++ } *a; ++ ++ /* this function can be called from magic sysrq */ ++ a = kzalloc(sizeof(*a), GFP_ATOMIC); ++ if (unlikely(!a)) { ++ dpri("no memory\n"); ++ return; ++ } ++ ++ a->mnt.mnt_sb = sb; ++ a->fake.br_perm = 0; ++ a->fake.br_mnt = &a->mnt; ++ a->fake.br_xino.xi_file = NULL; ++ atomic_set(&a->fake.br_count, 0); ++ smp_mb(); /* atomic_set */ ++ err = do_pri_br(-1, &a->fake); ++ kfree(a); ++ dpri("dev 0x%x\n", sb->s_dev); ++ if (err || !au_test_aufs(sb)) ++ return; ++ ++ sbinfo = au_sbi(sb); ++ if (!sbinfo) ++ return; ++ dpri("nw %d, gen %u, kobj %d\n", ++ atomic_read(&sbinfo->si_nowait.nw_len), sbinfo->si_generation, ++ atomic_read(&sbinfo->si_kobj.kref.refcount)); ++ for (bindex = 0; bindex <= sbinfo->si_bend; bindex++) ++ do_pri_br(bindex, sbinfo->si_branch[0 + bindex]); ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++void au_dbg_sleep_jiffy(int jiffy) ++{ ++ while (jiffy) ++ jiffy = schedule_timeout_uninterruptible(jiffy); ++} ++ ++void au_dbg_iattr(struct iattr *ia) ++{ ++#define AuBit(name) if (ia->ia_valid & ATTR_ ## name) \ ++ dpri(#name "\n") ++ AuBit(MODE); ++ AuBit(UID); ++ AuBit(GID); ++ AuBit(SIZE); ++ AuBit(ATIME); ++ AuBit(MTIME); ++ AuBit(CTIME); ++ AuBit(ATIME_SET); ++ AuBit(MTIME_SET); ++ AuBit(FORCE); ++ AuBit(ATTR_FLAG); ++ AuBit(KILL_SUID); ++ AuBit(KILL_SGID); ++ AuBit(FILE); ++ AuBit(KILL_PRIV); ++ AuBit(OPEN); ++ AuBit(TIMES_SET); ++#undef AuBit ++ dpri("ia_file %p\n", ia->ia_file); ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++void au_dbg_verify_dir_parent(struct dentry *dentry, unsigned int sigen) ++{ ++ struct dentry *parent; ++ ++ parent = dget_parent(dentry); ++ AuDebugOn(!S_ISDIR(dentry->d_inode->i_mode) ++ || IS_ROOT(dentry) ++ || au_digen(parent) != sigen); ++ dput(parent); ++} ++ ++void au_dbg_verify_nondir_parent(struct dentry *dentry, unsigned int sigen) ++{ ++ struct dentry *parent; ++ ++ parent = dget_parent(dentry); ++ AuDebugOn(S_ISDIR(dentry->d_inode->i_mode) ++ || au_digen(parent) != sigen); ++ dput(parent); ++} ++ ++void au_dbg_verify_gen(struct dentry *parent, unsigned int sigen) ++{ ++ int err, i, j; ++ struct au_dcsub_pages dpages; ++ struct au_dpage *dpage; ++ struct dentry **dentries; ++ ++ err = au_dpages_init(&dpages, GFP_NOFS); ++ AuDebugOn(err); ++ err = au_dcsub_pages_rev(&dpages, parent, /*do_include*/1, NULL, NULL); ++ AuDebugOn(err); ++ for (i = dpages.ndpage - 1; !err && i >= 0; i--) { ++ dpage = dpages.dpages + i; ++ dentries = dpage->dentries; ++ for (j = dpage->ndentry - 1; !err && j >= 0; j--) ++ AuDebugOn(au_digen(dentries[j]) != sigen); ++ } ++ au_dpages_free(&dpages); ++} ++ ++void au_dbg_verify_hf(struct au_finfo *finfo) ++{ ++ struct au_hfile *hf; ++ aufs_bindex_t bend, bindex; ++ ++ if (finfo->fi_bstart >= 0) { ++ bend = finfo->fi_bend; ++ for (bindex = finfo->fi_bstart; bindex <= bend; bindex++) { ++ hf = finfo->fi_hfile + bindex; ++ AuDebugOn(hf->hf_file || hf->hf_br); ++ } ++ } ++} ++ ++void au_dbg_verify_kthread(void) ++{ ++ if (au_test_wkq(current)) { ++ au_dbg_blocked(); ++ BUG(); ++ } ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++void au_debug_sbinfo_init(struct au_sbinfo *sbinfo __maybe_unused) ++{ ++#ifdef AuForceNoPlink ++ au_opt_clr(sbinfo->si_mntflags, PLINK); ++#endif ++#ifdef AuForceNoXino ++ au_opt_clr(sbinfo->si_mntflags, XINO); ++#endif ++#ifdef AuForceNoRefrof ++ au_opt_clr(sbinfo->si_mntflags, REFROF); ++#endif ++#ifdef AuForceHinotify ++ au_opt_set_udba(sbinfo->si_mntflags, UDBA_HINOTIFY); ++#endif ++} ++ ++int __init au_debug_init(void) ++{ ++ aufs_bindex_t bindex; ++ struct au_vdir_destr destr; ++ ++ bindex = -1; ++ AuDebugOn(bindex >= 0); ++ ++ destr.len = -1; ++ AuDebugOn(destr.len < NAME_MAX); ++ ++#ifdef CONFIG_4KSTACKS ++ AuWarn("CONFIG_4KSTACKS is defined.\n"); ++#endif ++ ++#ifdef AuForceNoBrs ++ sysaufs_brs = 0; ++#endif ++ ++ return 0; ++} +diff -Nur linux-2.6.31.4.orig/fs/aufs/debug.h linux-2.6.31.4/fs/aufs/debug.h +--- linux-2.6.31.4.orig/fs/aufs/debug.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/debug.h 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,260 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * debug print functions ++ */ ++ ++#ifndef __AUFS_DEBUG_H__ ++#define __AUFS_DEBUG_H__ ++ ++#ifdef __KERNEL__ ++ ++#include <linux/bug.h> ++/* #include <linux/err.h> */ ++/* #include <linux/init.h> */ ++/* #include <linux/kernel.h> */ ++#include <linux/delay.h> ++/* #include <linux/kd.h> */ ++/* #include <linux/vt_kern.h> */ ++#include <linux/sysrq.h> ++#include <linux/aufs_type.h> ++ ++#ifdef CONFIG_AUFS_DEBUG ++#define AuDebugOn(a) BUG_ON(a) ++ ++/* module parameter */ ++extern int aufs_debug; ++static inline void au_debug(int n) ++{ ++ aufs_debug = n; ++ smp_mb(); ++} ++ ++static inline int au_debug_test(void) ++{ ++ return aufs_debug; ++} ++#else ++#define AuDebugOn(a) do {} while (0) ++#define au_debug() do {} while (0) ++static inline int au_debug_test(void) ++{ ++ return 0; ++} ++#endif /* CONFIG_AUFS_DEBUG */ ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* debug print */ ++ ++#define AuDpri(lvl, fmt, arg...) \ ++ printk(lvl AUFS_NAME " %s:%d:%s[%d]: " fmt, \ ++ __func__, __LINE__, current->comm, current->pid, ##arg) ++#define AuDbg(fmt, arg...) do { \ ++ if (au_debug_test()) \ ++ AuDpri(KERN_DEBUG, "DEBUG: " fmt, ##arg); \ ++} while (0) ++#define AuLabel(l) AuDbg(#l "\n") ++#define AuInfo(fmt, arg...) AuDpri(KERN_INFO, fmt, ##arg) ++#define AuWarn(fmt, arg...) AuDpri(KERN_WARNING, fmt, ##arg) ++#define AuErr(fmt, arg...) AuDpri(KERN_ERR, fmt, ##arg) ++#define AuIOErr(fmt, arg...) AuErr("I/O Error, " fmt, ##arg) ++#define AuWarn1(fmt, arg...) do { \ ++ static unsigned char _c; \ ++ if (!_c++) \ ++ AuWarn(fmt, ##arg); \ ++} while (0) ++ ++#define AuErr1(fmt, arg...) do { \ ++ static unsigned char _c; \ ++ if (!_c++) \ ++ AuErr(fmt, ##arg); \ ++} while (0) ++ ++#define AuIOErr1(fmt, arg...) do { \ ++ static unsigned char _c; \ ++ if (!_c++) \ ++ AuIOErr(fmt, ##arg); \ ++} while (0) ++ ++#define AuUnsupportMsg "This operation is not supported." \ ++ " Please report this application to aufs-users ML." ++#define AuUnsupport(fmt, args...) do { \ ++ AuErr(AuUnsupportMsg "\n" fmt, ##args); \ ++ dump_stack(); \ ++} while (0) ++ ++#define AuTraceErr(e) do { \ ++ if (unlikely((e) < 0)) \ ++ AuDbg("err %d\n", (int)(e)); \ ++} while (0) ++ ++#define AuTraceErrPtr(p) do { \ ++ if (IS_ERR(p)) \ ++ AuDbg("err %ld\n", PTR_ERR(p)); \ ++} while (0) ++ ++/* dirty macros for debug print, use with "%.*s" and caution */ ++#define AuLNPair(qstr) (qstr)->len, (qstr)->name ++#define AuDLNPair(d) AuLNPair(&(d)->d_name) ++ ++/* ---------------------------------------------------------------------- */ ++ ++struct au_sbinfo; ++struct au_finfo; ++struct dentry; ++#ifdef CONFIG_AUFS_DEBUG ++extern char *au_plevel; ++struct au_nhash; ++void au_dpri_whlist(struct au_nhash *whlist); ++struct au_vdir; ++void au_dpri_vdir(struct au_vdir *vdir); ++struct inode; ++void au_dpri_inode(struct inode *inode); ++void au_dpri_dentry(struct dentry *dentry); ++struct file; ++void au_dpri_file(struct file *filp); ++struct super_block; ++void au_dpri_sb(struct super_block *sb); ++ ++void au_dbg_sleep_jiffy(int jiffy); ++struct iattr; ++void au_dbg_iattr(struct iattr *ia); ++ ++void au_dbg_verify_dir_parent(struct dentry *dentry, unsigned int sigen); ++void au_dbg_verify_nondir_parent(struct dentry *dentry, unsigned int sigen); ++void au_dbg_verify_gen(struct dentry *parent, unsigned int sigen); ++void au_dbg_verify_hf(struct au_finfo *finfo); ++void au_dbg_verify_kthread(void); ++ ++int __init au_debug_init(void); ++void au_debug_sbinfo_init(struct au_sbinfo *sbinfo); ++#define AuDbgWhlist(w) do { \ ++ AuDbg(#w "\n"); \ ++ au_dpri_whlist(w); \ ++} while (0) ++ ++#define AuDbgVdir(v) do { \ ++ AuDbg(#v "\n"); \ ++ au_dpri_vdir(v); \ ++} while (0) ++ ++#define AuDbgInode(i) do { \ ++ AuDbg(#i "\n"); \ ++ au_dpri_inode(i); \ ++} while (0) ++ ++#define AuDbgDentry(d) do { \ ++ AuDbg(#d "\n"); \ ++ au_dpri_dentry(d); \ ++} while (0) ++ ++#define AuDbgFile(f) do { \ ++ AuDbg(#f "\n"); \ ++ au_dpri_file(f); \ ++} while (0) ++ ++#define AuDbgSb(sb) do { \ ++ AuDbg(#sb "\n"); \ ++ au_dpri_sb(sb); \ ++} while (0) ++ ++#define AuDbgSleep(sec) do { \ ++ AuDbg("sleep %d sec\n", sec); \ ++ ssleep(sec); \ ++} while (0) ++ ++#define AuDbgSleepJiffy(jiffy) do { \ ++ AuDbg("sleep %d jiffies\n", jiffy); \ ++ au_dbg_sleep_jiffy(jiffy); \ ++} while (0) ++ ++#define AuDbgIAttr(ia) do { \ ++ AuDbg("ia_valid 0x%x\n", (ia)->ia_valid); \ ++ au_dbg_iattr(ia); \ ++} while (0) ++#else ++static inline void au_dbg_verify_dir_parent(struct dentry *dentry, ++ unsigned int sigen) ++{ ++ /* empty */ ++} ++static inline void au_dbg_verify_nondir_parent(struct dentry *dentry, ++ unsigned int sigen) ++{ ++ /* empty */ ++} ++static inline void au_dbg_verify_gen(struct dentry *parent, unsigned int sigen) ++{ ++ /* empty */ ++} ++static inline void au_dbg_verify_hf(struct au_finfo *finfo) ++{ ++ /* empty */ ++} ++static inline void au_dbg_verify_kthread(void) ++{ ++ /* empty */ ++} ++ ++static inline int au_debug_init(void) ++{ ++ return 0; ++} ++static inline void au_debug_sbinfo_init(struct au_sbinfo *sbinfo) ++{ ++ /* empty */ ++} ++#define AuDbgWhlist(w) do {} while (0) ++#define AuDbgVdir(v) do {} while (0) ++#define AuDbgInode(i) do {} while (0) ++#define AuDbgDentry(d) do {} while (0) ++#define AuDbgFile(f) do {} while (0) ++#define AuDbgSb(sb) do {} while (0) ++#define AuDbgSleep(sec) do {} while (0) ++#define AuDbgSleepJiffy(jiffy) do {} while (0) ++#define AuDbgIAttr(ia) do {} while (0) ++#endif /* CONFIG_AUFS_DEBUG */ ++ ++/* ---------------------------------------------------------------------- */ ++ ++#ifdef CONFIG_AUFS_MAGIC_SYSRQ ++int __init au_sysrq_init(void); ++void au_sysrq_fin(void); ++ ++#ifdef CONFIG_HW_CONSOLE ++#define au_dbg_blocked() do { \ ++ WARN_ON(1); \ ++ handle_sysrq('w', vc_cons[fg_console].d->vc_tty); \ ++} while (0) ++#else ++#define au_dbg_blocked() do {} while (0) ++#endif ++ ++#else ++static inline int au_sysrq_init(void) ++{ ++ return 0; ++} ++#define au_sysrq_fin() do {} while (0) ++#define au_dbg_blocked() do {} while (0) ++#endif /* CONFIG_AUFS_MAGIC_SYSRQ */ ++ ++#endif /* __KERNEL__ */ ++#endif /* __AUFS_DEBUG_H__ */ +diff -Nur linux-2.6.31.4.orig/fs/aufs/dentry.c linux-2.6.31.4/fs/aufs/dentry.c +--- linux-2.6.31.4.orig/fs/aufs/dentry.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/dentry.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,880 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * lookup and dentry operations ++ */ ++ ++#include <linux/namei.h> ++#include "aufs.h" ++ ++static void au_h_nd(struct nameidata *h_nd, struct nameidata *nd) ++{ ++ if (nd) { ++ *h_nd = *nd; ++ ++ /* ++ * gave up supporting LOOKUP_CREATE/OPEN for lower fs, ++ * due to whiteout and branch permission. ++ */ ++ h_nd->flags &= ~(/*LOOKUP_PARENT |*/ LOOKUP_OPEN | LOOKUP_CREATE ++ | LOOKUP_FOLLOW); ++ /* unnecessary? */ ++ h_nd->intent.open.file = NULL; ++ } else ++ memset(h_nd, 0, sizeof(*h_nd)); ++} ++ ++struct au_lkup_one_args { ++ struct dentry **errp; ++ struct qstr *name; ++ struct dentry *h_parent; ++ struct au_branch *br; ++ struct nameidata *nd; ++}; ++ ++struct dentry *au_lkup_one(struct qstr *name, struct dentry *h_parent, ++ struct au_branch *br, struct nameidata *nd) ++{ ++ struct dentry *h_dentry; ++ int err; ++ struct nameidata h_nd; ++ ++ if (au_test_fs_null_nd(h_parent->d_sb)) ++ return vfsub_lookup_one_len(name->name, h_parent, name->len); ++ ++ au_h_nd(&h_nd, nd); ++ h_nd.path.dentry = h_parent; ++ h_nd.path.mnt = br->br_mnt; ++ ++ err = __lookup_one_len(name->name, &h_nd.last, NULL, name->len); ++ h_dentry = ERR_PTR(err); ++ if (!err) { ++ path_get(&h_nd.path); ++ h_dentry = vfsub_lookup_hash(&h_nd); ++ path_put(&h_nd.path); ++ } ++ ++ return h_dentry; ++} ++ ++static void au_call_lkup_one(void *args) ++{ ++ struct au_lkup_one_args *a = args; ++ *a->errp = au_lkup_one(a->name, a->h_parent, a->br, a->nd); ++} ++ ++#define AuLkup_ALLOW_NEG 1 ++#define au_ftest_lkup(flags, name) ((flags) & AuLkup_##name) ++#define au_fset_lkup(flags, name) { (flags) |= AuLkup_##name; } ++#define au_fclr_lkup(flags, name) { (flags) &= ~AuLkup_##name; } ++ ++struct au_do_lookup_args { ++ unsigned int flags; ++ mode_t type; ++ struct nameidata *nd; ++}; ++ ++/* ++ * returns positive/negative dentry, NULL or an error. ++ * NULL means whiteout-ed or not-found. ++ */ ++static struct dentry* ++au_do_lookup(struct dentry *h_parent, struct dentry *dentry, ++ aufs_bindex_t bindex, struct qstr *wh_name, ++ struct au_do_lookup_args *args) ++{ ++ struct dentry *h_dentry; ++ struct inode *h_inode, *inode; ++ struct qstr *name; ++ struct au_branch *br; ++ int wh_found, opq; ++ unsigned char wh_able; ++ const unsigned char allow_neg = !!au_ftest_lkup(args->flags, ALLOW_NEG); ++ ++ name = &dentry->d_name; ++ wh_found = 0; ++ br = au_sbr(dentry->d_sb, bindex); ++ wh_able = !!au_br_whable(br->br_perm); ++ if (wh_able) ++ wh_found = au_wh_test(h_parent, wh_name, br, /*try_sio*/0); ++ h_dentry = ERR_PTR(wh_found); ++ if (!wh_found) ++ goto real_lookup; ++ if (unlikely(wh_found < 0)) ++ goto out; ++ ++ /* We found a whiteout */ ++ /* au_set_dbend(dentry, bindex); */ ++ au_set_dbwh(dentry, bindex); ++ if (!allow_neg) ++ return NULL; /* success */ ++ ++ real_lookup: ++ h_dentry = au_lkup_one(name, h_parent, br, args->nd); ++ if (IS_ERR(h_dentry)) ++ goto out; ++ ++ h_inode = h_dentry->d_inode; ++ if (!h_inode) { ++ if (!allow_neg) ++ goto out_neg; ++ } else if (wh_found ++ || (args->type && args->type != (h_inode->i_mode & S_IFMT))) ++ goto out_neg; ++ ++ if (au_dbend(dentry) <= bindex) ++ au_set_dbend(dentry, bindex); ++ if (au_dbstart(dentry) < 0 || bindex < au_dbstart(dentry)) ++ au_set_dbstart(dentry, bindex); ++ au_set_h_dptr(dentry, bindex, h_dentry); ++ ++ inode = dentry->d_inode; ++ if (!h_inode || !S_ISDIR(h_inode->i_mode) || !wh_able ++ || (inode && !S_ISDIR(inode->i_mode))) ++ goto out; /* success */ ++ ++ mutex_lock_nested(&h_inode->i_mutex, AuLsc_I_CHILD); ++ opq = au_diropq_test(h_dentry, br); ++ mutex_unlock(&h_inode->i_mutex); ++ if (opq > 0) ++ au_set_dbdiropq(dentry, bindex); ++ else if (unlikely(opq < 0)) { ++ au_set_h_dptr(dentry, bindex, NULL); ++ h_dentry = ERR_PTR(opq); ++ } ++ goto out; ++ ++ out_neg: ++ dput(h_dentry); ++ h_dentry = NULL; ++ out: ++ return h_dentry; ++} ++ ++static int au_test_shwh(struct super_block *sb, const struct qstr *name) ++{ ++ if (unlikely(!au_opt_test(au_mntflags(sb), SHWH) ++ && !strncmp(name->name, AUFS_WH_PFX, AUFS_WH_PFX_LEN))) ++ return -EPERM; ++ return 0; ++} ++ ++/* ++ * returns the number of lower positive dentries, ++ * otherwise an error. ++ * can be called at unlinking with @type is zero. ++ */ ++int au_lkup_dentry(struct dentry *dentry, aufs_bindex_t bstart, mode_t type, ++ struct nameidata *nd) ++{ ++ int npositive, err; ++ aufs_bindex_t bindex, btail, bdiropq; ++ unsigned char isdir; ++ struct qstr whname; ++ struct au_do_lookup_args args = { ++ .flags = 0, ++ .type = type, ++ .nd = nd ++ }; ++ const struct qstr *name = &dentry->d_name; ++ struct dentry *parent; ++ struct inode *inode; ++ ++ parent = dget_parent(dentry); ++ err = au_test_shwh(dentry->d_sb, name); ++ if (unlikely(err)) ++ goto out; ++ ++ err = au_wh_name_alloc(&whname, name); ++ if (unlikely(err)) ++ goto out; ++ ++ inode = dentry->d_inode; ++ isdir = !!(inode && S_ISDIR(inode->i_mode)); ++ if (!type) ++ au_fset_lkup(args.flags, ALLOW_NEG); ++ ++ npositive = 0; ++ btail = au_dbtaildir(parent); ++ for (bindex = bstart; bindex <= btail; bindex++) { ++ struct dentry *h_parent, *h_dentry; ++ struct inode *h_inode, *h_dir; ++ ++ h_dentry = au_h_dptr(dentry, bindex); ++ if (h_dentry) { ++ if (h_dentry->d_inode) ++ npositive++; ++ if (type != S_IFDIR) ++ break; ++ continue; ++ } ++ h_parent = au_h_dptr(parent, bindex); ++ if (!h_parent) ++ continue; ++ h_dir = h_parent->d_inode; ++ if (!h_dir || !S_ISDIR(h_dir->i_mode)) ++ continue; ++ ++ mutex_lock_nested(&h_dir->i_mutex, AuLsc_I_PARENT); ++ h_dentry = au_do_lookup(h_parent, dentry, bindex, &whname, ++ &args); ++ mutex_unlock(&h_dir->i_mutex); ++ err = PTR_ERR(h_dentry); ++ if (IS_ERR(h_dentry)) ++ goto out_wh; ++ au_fclr_lkup(args.flags, ALLOW_NEG); ++ ++ if (au_dbwh(dentry) >= 0) ++ break; ++ if (!h_dentry) ++ continue; ++ h_inode = h_dentry->d_inode; ++ if (!h_inode) ++ continue; ++ npositive++; ++ if (!args.type) ++ args.type = h_inode->i_mode & S_IFMT; ++ if (args.type != S_IFDIR) ++ break; ++ else if (isdir) { ++ /* the type of lower may be different */ ++ bdiropq = au_dbdiropq(dentry); ++ if (bdiropq >= 0 && bdiropq <= bindex) ++ break; ++ } ++ } ++ ++ if (npositive) { ++ AuLabel(positive); ++ au_update_dbstart(dentry); ++ } ++ err = npositive; ++ if (unlikely(!au_opt_test(au_mntflags(dentry->d_sb), UDBA_NONE) ++ && au_dbstart(dentry) < 0)) ++ /* both of real entry and whiteout found */ ++ err = -EIO; ++ ++ out_wh: ++ kfree(whname.name); ++ out: ++ dput(parent); ++ return err; ++} ++ ++struct dentry *au_sio_lkup_one(struct qstr *name, struct dentry *parent, ++ struct au_branch *br) ++{ ++ struct dentry *dentry; ++ int wkq_err; ++ ++ if (!au_test_h_perm_sio(parent->d_inode, MAY_EXEC)) ++ dentry = au_lkup_one(name, parent, br, /*nd*/NULL); ++ else { ++ struct au_lkup_one_args args = { ++ .errp = &dentry, ++ .name = name, ++ .h_parent = parent, ++ .br = br, ++ .nd = NULL ++ }; ++ ++ wkq_err = au_wkq_wait(au_call_lkup_one, &args); ++ if (unlikely(wkq_err)) ++ dentry = ERR_PTR(wkq_err); ++ } ++ ++ return dentry; ++} ++ ++/* ++ * lookup @dentry on @bindex which should be negative. ++ */ ++int au_lkup_neg(struct dentry *dentry, aufs_bindex_t bindex) ++{ ++ int err; ++ struct dentry *parent, *h_parent, *h_dentry; ++ struct qstr *name; ++ ++ name = &dentry->d_name; ++ parent = dget_parent(dentry); ++ h_parent = au_h_dptr(parent, bindex); ++ h_dentry = au_sio_lkup_one(name, h_parent, ++ au_sbr(dentry->d_sb, bindex)); ++ err = PTR_ERR(h_dentry); ++ if (IS_ERR(h_dentry)) ++ goto out; ++ if (unlikely(h_dentry->d_inode)) { ++ err = -EIO; ++ AuIOErr("b%d %.*s should be negative.\n", ++ bindex, AuDLNPair(h_dentry)); ++ dput(h_dentry); ++ goto out; ++ } ++ ++ if (bindex < au_dbstart(dentry)) ++ au_set_dbstart(dentry, bindex); ++ if (au_dbend(dentry) < bindex) ++ au_set_dbend(dentry, bindex); ++ au_set_h_dptr(dentry, bindex, h_dentry); ++ err = 0; ++ ++ out: ++ dput(parent); ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* subset of struct inode */ ++struct au_iattr { ++ unsigned long i_ino; ++ /* unsigned int i_nlink; */ ++ uid_t i_uid; ++ gid_t i_gid; ++ u64 i_version; ++/* ++ loff_t i_size; ++ blkcnt_t i_blocks; ++*/ ++ umode_t i_mode; ++}; ++ ++static void au_iattr_save(struct au_iattr *ia, struct inode *h_inode) ++{ ++ ia->i_ino = h_inode->i_ino; ++ /* ia->i_nlink = h_inode->i_nlink; */ ++ ia->i_uid = h_inode->i_uid; ++ ia->i_gid = h_inode->i_gid; ++ ia->i_version = h_inode->i_version; ++/* ++ ia->i_size = h_inode->i_size; ++ ia->i_blocks = h_inode->i_blocks; ++*/ ++ ia->i_mode = (h_inode->i_mode & S_IFMT); ++} ++ ++static int au_iattr_test(struct au_iattr *ia, struct inode *h_inode) ++{ ++ return ia->i_ino != h_inode->i_ino ++ /* || ia->i_nlink != h_inode->i_nlink */ ++ || ia->i_uid != h_inode->i_uid ++ || ia->i_gid != h_inode->i_gid ++ || ia->i_version != h_inode->i_version ++/* ++ || ia->i_size != h_inode->i_size ++ || ia->i_blocks != h_inode->i_blocks ++*/ ++ || ia->i_mode != (h_inode->i_mode & S_IFMT); ++} ++ ++static int au_h_verify_dentry(struct dentry *h_dentry, struct dentry *h_parent, ++ struct au_branch *br) ++{ ++ int err; ++ struct au_iattr ia; ++ struct inode *h_inode; ++ struct dentry *h_d; ++ struct super_block *h_sb; ++ ++ err = 0; ++ memset(&ia, -1, sizeof(ia)); ++ h_sb = h_dentry->d_sb; ++ h_inode = h_dentry->d_inode; ++ if (h_inode) ++ au_iattr_save(&ia, h_inode); ++ else if (au_test_nfs(h_sb) || au_test_fuse(h_sb)) ++ /* nfs d_revalidate may return 0 for negative dentry */ ++ /* fuse d_revalidate always return 0 for negative dentry */ ++ goto out; ++ ++ /* main purpose is namei.c:cached_lookup() and d_revalidate */ ++ h_d = au_lkup_one(&h_dentry->d_name, h_parent, br, /*nd*/NULL); ++ err = PTR_ERR(h_d); ++ if (IS_ERR(h_d)) ++ goto out; ++ ++ err = 0; ++ if (unlikely(h_d != h_dentry ++ || h_d->d_inode != h_inode ++ || (h_inode && au_iattr_test(&ia, h_inode)))) ++ err = au_busy_or_stale(); ++ dput(h_d); ++ ++ out: ++ AuTraceErr(err); ++ return err; ++} ++ ++int au_h_verify(struct dentry *h_dentry, unsigned int udba, struct inode *h_dir, ++ struct dentry *h_parent, struct au_branch *br) ++{ ++ int err; ++ ++ err = 0; ++ if (udba == AuOpt_UDBA_REVAL) { ++ IMustLock(h_dir); ++ err = (h_dentry->d_parent->d_inode != h_dir); ++ } else if (udba == AuOpt_UDBA_HINOTIFY) ++ err = au_h_verify_dentry(h_dentry, h_parent, br); ++ ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static void au_do_refresh_hdentry(struct au_hdentry *p, struct au_dinfo *dinfo, ++ struct dentry *parent) ++{ ++ struct dentry *h_d, *h_dp; ++ struct au_hdentry tmp, *q; ++ struct super_block *sb; ++ aufs_bindex_t new_bindex, bindex, bend, bwh, bdiropq; ++ ++ AuRwMustWriteLock(&dinfo->di_rwsem); ++ ++ bend = dinfo->di_bend; ++ bwh = dinfo->di_bwh; ++ bdiropq = dinfo->di_bdiropq; ++ for (bindex = dinfo->di_bstart; bindex <= bend; bindex++, p++) { ++ h_d = p->hd_dentry; ++ if (!h_d) ++ continue; ++ ++ h_dp = dget_parent(h_d); ++ if (h_dp == au_h_dptr(parent, bindex)) { ++ dput(h_dp); ++ continue; ++ } ++ ++ new_bindex = au_find_dbindex(parent, h_dp); ++ dput(h_dp); ++ if (dinfo->di_bwh == bindex) ++ bwh = new_bindex; ++ if (dinfo->di_bdiropq == bindex) ++ bdiropq = new_bindex; ++ if (new_bindex < 0) { ++ au_hdput(p); ++ p->hd_dentry = NULL; ++ continue; ++ } ++ ++ /* swap two lower dentries, and loop again */ ++ q = dinfo->di_hdentry + new_bindex; ++ tmp = *q; ++ *q = *p; ++ *p = tmp; ++ if (tmp.hd_dentry) { ++ bindex--; ++ p--; ++ } ++ } ++ ++ sb = parent->d_sb; ++ dinfo->di_bwh = -1; ++ if (bwh >= 0 && bwh <= au_sbend(sb) && au_sbr_whable(sb, bwh)) ++ dinfo->di_bwh = bwh; ++ ++ dinfo->di_bdiropq = -1; ++ if (bdiropq >= 0 ++ && bdiropq <= au_sbend(sb) ++ && au_sbr_whable(sb, bdiropq)) ++ dinfo->di_bdiropq = bdiropq; ++ ++ bend = au_dbend(parent); ++ p = dinfo->di_hdentry; ++ for (bindex = 0; bindex <= bend; bindex++, p++) ++ if (p->hd_dentry) { ++ dinfo->di_bstart = bindex; ++ break; ++ } ++ ++ p = dinfo->di_hdentry + bend; ++ for (bindex = bend; bindex >= 0; bindex--, p--) ++ if (p->hd_dentry) { ++ dinfo->di_bend = bindex; ++ break; ++ } ++} ++ ++/* ++ * returns the number of found lower positive dentries, ++ * otherwise an error. ++ */ ++int au_refresh_hdentry(struct dentry *dentry, mode_t type) ++{ ++ int npositive, err; ++ unsigned int sigen; ++ aufs_bindex_t bstart; ++ struct au_dinfo *dinfo; ++ struct super_block *sb; ++ struct dentry *parent; ++ ++ DiMustWriteLock(dentry); ++ ++ sb = dentry->d_sb; ++ AuDebugOn(IS_ROOT(dentry)); ++ sigen = au_sigen(sb); ++ parent = dget_parent(dentry); ++ AuDebugOn(au_digen(parent) != sigen ++ || au_iigen(parent->d_inode) != sigen); ++ ++ dinfo = au_di(dentry); ++ err = au_di_realloc(dinfo, au_sbend(sb) + 1); ++ npositive = err; ++ if (unlikely(err)) ++ goto out; ++ au_do_refresh_hdentry(dinfo->di_hdentry + dinfo->di_bstart, dinfo, ++ parent); ++ ++ npositive = 0; ++ bstart = au_dbstart(parent); ++ if (type != S_IFDIR && dinfo->di_bstart == bstart) ++ goto out_dgen; /* success */ ++ ++ npositive = au_lkup_dentry(dentry, bstart, type, /*nd*/NULL); ++ if (npositive < 0) ++ goto out; ++ if (dinfo->di_bwh >= 0 && dinfo->di_bwh <= dinfo->di_bstart) ++ d_drop(dentry); ++ ++ out_dgen: ++ au_update_digen(dentry); ++ out: ++ dput(parent); ++ AuTraceErr(npositive); ++ return npositive; ++} ++ ++static noinline_for_stack ++int au_do_h_d_reval(struct dentry *h_dentry, struct nameidata *nd, ++ struct dentry *dentry, aufs_bindex_t bindex) ++{ ++ int err, valid; ++ int (*reval)(struct dentry *, struct nameidata *); ++ ++ err = 0; ++ reval = NULL; ++ if (h_dentry->d_op) ++ reval = h_dentry->d_op->d_revalidate; ++ if (!reval) ++ goto out; ++ ++ AuDbg("b%d\n", bindex); ++ if (au_test_fs_null_nd(h_dentry->d_sb)) ++ /* it may return tri-state */ ++ valid = reval(h_dentry, NULL); ++ else { ++ struct nameidata h_nd; ++ int locked; ++ struct dentry *parent; ++ ++ au_h_nd(&h_nd, nd); ++ parent = nd->path.dentry; ++ locked = (nd && nd->path.dentry != dentry); ++ if (locked) ++ di_read_lock_parent(parent, AuLock_IR); ++ BUG_ON(bindex > au_dbend(parent)); ++ h_nd.path.dentry = au_h_dptr(parent, bindex); ++ BUG_ON(!h_nd.path.dentry); ++ h_nd.path.mnt = au_sbr(parent->d_sb, bindex)->br_mnt; ++ path_get(&h_nd.path); ++ valid = reval(h_dentry, &h_nd); ++ path_put(&h_nd.path); ++ if (locked) ++ di_read_unlock(parent, AuLock_IR); ++ } ++ ++ if (unlikely(valid < 0)) ++ err = valid; ++ else if (!valid) ++ err = -EINVAL; ++ ++ out: ++ AuTraceErr(err); ++ return err; ++} ++ ++/* todo: remove this */ ++static int h_d_revalidate(struct dentry *dentry, struct inode *inode, ++ struct nameidata *nd, int do_udba) ++{ ++ int err; ++ umode_t mode, h_mode; ++ aufs_bindex_t bindex, btail, bstart, ibs, ibe; ++ unsigned char plus, unhashed, is_root, h_plus; ++ struct inode *first, *h_inode, *h_cached_inode; ++ struct dentry *h_dentry; ++ struct qstr *name, *h_name; ++ ++ err = 0; ++ plus = 0; ++ mode = 0; ++ first = NULL; ++ ibs = -1; ++ ibe = -1; ++ unhashed = !!d_unhashed(dentry); ++ is_root = !!IS_ROOT(dentry); ++ name = &dentry->d_name; ++ ++ /* ++ * Theoretically, REVAL test should be unnecessary in case of INOTIFY. ++ * But inotify doesn't fire some necessary events, ++ * IN_ATTRIB for atime/nlink/pageio ++ * IN_DELETE for NFS dentry ++ * Let's do REVAL test too. ++ */ ++ if (do_udba && inode) { ++ mode = (inode->i_mode & S_IFMT); ++ plus = (inode->i_nlink > 0); ++ first = au_h_iptr(inode, au_ibstart(inode)); ++ ibs = au_ibstart(inode); ++ ibe = au_ibend(inode); ++ } ++ ++ bstart = au_dbstart(dentry); ++ btail = bstart; ++ if (inode && S_ISDIR(inode->i_mode)) ++ btail = au_dbtaildir(dentry); ++ for (bindex = bstart; bindex <= btail; bindex++) { ++ h_dentry = au_h_dptr(dentry, bindex); ++ if (!h_dentry) ++ continue; ++ ++ AuDbg("b%d, %.*s\n", bindex, AuDLNPair(h_dentry)); ++ h_name = &h_dentry->d_name; ++ if (unlikely(do_udba ++ && !is_root ++ && (unhashed != !!d_unhashed(h_dentry) ++ || name->len != h_name->len ++ || memcmp(name->name, h_name->name, name->len)) ++ )) { ++ AuDbg("unhash 0x%x 0x%x, %.*s %.*s\n", ++ unhashed, d_unhashed(h_dentry), ++ AuDLNPair(dentry), AuDLNPair(h_dentry)); ++ goto err; ++ } ++ ++ err = au_do_h_d_reval(h_dentry, nd, dentry, bindex); ++ if (unlikely(err)) ++ /* do not goto err, to keep the errno */ ++ break; ++ ++ /* todo: plink too? */ ++ if (!do_udba) ++ continue; ++ ++ /* UDBA tests */ ++ h_inode = h_dentry->d_inode; ++ if (unlikely(!!inode != !!h_inode)) ++ goto err; ++ ++ h_plus = plus; ++ h_mode = mode; ++ h_cached_inode = h_inode; ++ if (h_inode) { ++ h_mode = (h_inode->i_mode & S_IFMT); ++ h_plus = (h_inode->i_nlink > 0); ++ } ++ if (inode && ibs <= bindex && bindex <= ibe) ++ h_cached_inode = au_h_iptr(inode, bindex); ++ ++ if (unlikely(plus != h_plus ++ || mode != h_mode ++ || h_cached_inode != h_inode)) ++ goto err; ++ continue; ++ ++ err: ++ err = -EINVAL; ++ break; ++ } ++ ++ return err; ++} ++ ++static int simple_reval_dpath(struct dentry *dentry, unsigned int sigen) ++{ ++ int err; ++ struct dentry *parent; ++ struct inode *inode; ++ ++ inode = dentry->d_inode; ++ if (au_digen(dentry) == sigen && au_iigen(inode) == sigen) ++ return 0; ++ ++ parent = dget_parent(dentry); ++ di_read_lock_parent(parent, AuLock_IR); ++ AuDebugOn(au_digen(parent) != sigen ++ || au_iigen(parent->d_inode) != sigen); ++ au_dbg_verify_gen(parent, sigen); ++ ++ /* returns a number of positive dentries */ ++ err = au_refresh_hdentry(dentry, inode->i_mode & S_IFMT); ++ if (err >= 0) ++ err = au_refresh_hinode(inode, dentry); ++ ++ di_read_unlock(parent, AuLock_IR); ++ dput(parent); ++ return err; ++} ++ ++int au_reval_dpath(struct dentry *dentry, unsigned int sigen) ++{ ++ int err; ++ struct dentry *d, *parent; ++ struct inode *inode; ++ ++ if (!au_ftest_si(au_sbi(dentry->d_sb), FAILED_REFRESH_DIRS)) ++ return simple_reval_dpath(dentry, sigen); ++ ++ /* slow loop, keep it simple and stupid */ ++ /* cf: au_cpup_dirs() */ ++ err = 0; ++ parent = NULL; ++ while (au_digen(dentry) != sigen ++ || au_iigen(dentry->d_inode) != sigen) { ++ d = dentry; ++ while (1) { ++ dput(parent); ++ parent = dget_parent(d); ++ if (au_digen(parent) == sigen ++ && au_iigen(parent->d_inode) == sigen) ++ break; ++ d = parent; ++ } ++ ++ inode = d->d_inode; ++ if (d != dentry) ++ di_write_lock_child(d); ++ ++ /* someone might update our dentry while we were sleeping */ ++ if (au_digen(d) != sigen || au_iigen(d->d_inode) != sigen) { ++ di_read_lock_parent(parent, AuLock_IR); ++ /* returns a number of positive dentries */ ++ err = au_refresh_hdentry(d, inode->i_mode & S_IFMT); ++ if (err >= 0) ++ err = au_refresh_hinode(inode, d); ++ di_read_unlock(parent, AuLock_IR); ++ } ++ ++ if (d != dentry) ++ di_write_unlock(d); ++ dput(parent); ++ if (unlikely(err)) ++ break; ++ } ++ ++ return err; ++} ++ ++/* ++ * if valid returns 1, otherwise 0. ++ */ ++static int aufs_d_revalidate(struct dentry *dentry, struct nameidata *nd) ++{ ++ int valid, err; ++ unsigned int sigen; ++ unsigned char do_udba; ++ struct super_block *sb; ++ struct inode *inode; ++ ++ err = -EINVAL; ++ sb = dentry->d_sb; ++ inode = dentry->d_inode; ++ aufs_read_lock(dentry, AuLock_FLUSH | AuLock_DW); ++ sigen = au_sigen(sb); ++ if (au_digen(dentry) != sigen) { ++ AuDebugOn(IS_ROOT(dentry)); ++ if (inode) ++ err = au_reval_dpath(dentry, sigen); ++ if (unlikely(err)) ++ goto out_dgrade; ++ AuDebugOn(au_digen(dentry) != sigen); ++ } ++ if (inode && au_iigen(inode) != sigen) { ++ AuDebugOn(IS_ROOT(dentry)); ++ err = au_refresh_hinode(inode, dentry); ++ if (unlikely(err)) ++ goto out_dgrade; ++ AuDebugOn(au_iigen(inode) != sigen); ++ } ++ di_downgrade_lock(dentry, AuLock_IR); ++ ++ AuDebugOn(au_digen(dentry) != sigen); ++ AuDebugOn(inode && au_iigen(inode) != sigen); ++ err = -EINVAL; ++ do_udba = !au_opt_test(au_mntflags(sb), UDBA_NONE); ++ if (do_udba && inode) { ++ aufs_bindex_t bstart = au_ibstart(inode); ++ ++ if (bstart >= 0 ++ && au_test_higen(inode, au_h_iptr(inode, bstart))) ++ goto out; ++ } ++ ++ err = h_d_revalidate(dentry, inode, nd, do_udba); ++ if (unlikely(!err && do_udba && au_dbstart(dentry) < 0)) ++ /* both of real entry and whiteout found */ ++ err = -EIO; ++ goto out; ++ ++ out_dgrade: ++ di_downgrade_lock(dentry, AuLock_IR); ++ out: ++ au_store_oflag(nd, inode); ++ aufs_read_unlock(dentry, AuLock_IR); ++ AuTraceErr(err); ++ valid = !err; ++ if (!valid) ++ AuDbg("%.*s invalid\n", AuDLNPair(dentry)); ++ return valid; ++} ++ ++static void aufs_d_release(struct dentry *dentry) ++{ ++ struct au_dinfo *dinfo; ++ aufs_bindex_t bend, bindex; ++ ++ dinfo = dentry->d_fsdata; ++ if (!dinfo) ++ return; ++ ++ /* dentry may not be revalidated */ ++ bindex = dinfo->di_bstart; ++ if (bindex >= 0) { ++ struct au_hdentry *p; ++ ++ bend = dinfo->di_bend; ++ p = dinfo->di_hdentry + bindex; ++ while (bindex++ <= bend) { ++ if (p->hd_dentry) ++ au_hdput(p); ++ p++; ++ } ++ } ++ kfree(dinfo->di_hdentry); ++ AuRwDestroy(&dinfo->di_rwsem); ++ au_cache_free_dinfo(dinfo); ++ au_hin_di_reinit(dentry); ++} ++ ++struct dentry_operations aufs_dop = { ++ .d_revalidate = aufs_d_revalidate, ++ .d_release = aufs_d_release ++}; +diff -Nur linux-2.6.31.4.orig/fs/aufs/dentry.h linux-2.6.31.4/fs/aufs/dentry.h +--- linux-2.6.31.4.orig/fs/aufs/dentry.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/dentry.h 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,231 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * lookup and dentry operations ++ */ ++ ++#ifndef __AUFS_DENTRY_H__ ++#define __AUFS_DENTRY_H__ ++ ++#ifdef __KERNEL__ ++ ++#include <linux/dcache.h> ++#include <linux/aufs_type.h> ++#include "rwsem.h" ++ ++/* make a single member structure for future use */ ++/* todo: remove this structure */ ++struct au_hdentry { ++ struct dentry *hd_dentry; ++}; ++ ++struct au_dinfo { ++ atomic_t di_generation; ++ ++ struct au_rwsem di_rwsem; ++ aufs_bindex_t di_bstart, di_bend, di_bwh, di_bdiropq; ++ struct au_hdentry *di_hdentry; ++}; ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* dentry.c */ ++extern struct dentry_operations aufs_dop; ++struct au_branch; ++struct dentry *au_lkup_one(struct qstr *name, struct dentry *h_parent, ++ struct au_branch *br, struct nameidata *nd); ++struct dentry *au_sio_lkup_one(struct qstr *name, struct dentry *parent, ++ struct au_branch *br); ++int au_h_verify(struct dentry *h_dentry, unsigned int udba, struct inode *h_dir, ++ struct dentry *h_parent, struct au_branch *br); ++ ++int au_lkup_dentry(struct dentry *dentry, aufs_bindex_t bstart, mode_t type, ++ struct nameidata *nd); ++int au_lkup_neg(struct dentry *dentry, aufs_bindex_t bindex); ++int au_refresh_hdentry(struct dentry *dentry, mode_t type); ++int au_reval_dpath(struct dentry *dentry, unsigned int sigen); ++ ++/* dinfo.c */ ++int au_alloc_dinfo(struct dentry *dentry); ++int au_di_realloc(struct au_dinfo *dinfo, int nbr); ++ ++void di_read_lock(struct dentry *d, int flags, unsigned int lsc); ++void di_read_unlock(struct dentry *d, int flags); ++void di_downgrade_lock(struct dentry *d, int flags); ++void di_write_lock(struct dentry *d, unsigned int lsc); ++void di_write_unlock(struct dentry *d); ++void di_write_lock2_child(struct dentry *d1, struct dentry *d2, int isdir); ++void di_write_lock2_parent(struct dentry *d1, struct dentry *d2, int isdir); ++void di_write_unlock2(struct dentry *d1, struct dentry *d2); ++ ++struct dentry *au_h_dptr(struct dentry *dentry, aufs_bindex_t bindex); ++aufs_bindex_t au_dbtail(struct dentry *dentry); ++aufs_bindex_t au_dbtaildir(struct dentry *dentry); ++ ++void au_set_h_dptr(struct dentry *dentry, aufs_bindex_t bindex, ++ struct dentry *h_dentry); ++void au_update_digen(struct dentry *dentry); ++void au_update_dbrange(struct dentry *dentry, int do_put_zero); ++void au_update_dbstart(struct dentry *dentry); ++void au_update_dbend(struct dentry *dentry); ++int au_find_dbindex(struct dentry *dentry, struct dentry *h_dentry); ++ ++/* ---------------------------------------------------------------------- */ ++ ++static inline struct au_dinfo *au_di(struct dentry *dentry) ++{ ++ return dentry->d_fsdata; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* lock subclass for dinfo */ ++enum { ++ AuLsc_DI_CHILD, /* child first */ ++ AuLsc_DI_CHILD2, /* rename(2), link(2), and cpup at hinotify */ ++ AuLsc_DI_CHILD3, /* copyup dirs */ ++ AuLsc_DI_PARENT, ++ AuLsc_DI_PARENT2, ++ AuLsc_DI_PARENT3 ++}; ++ ++/* ++ * di_read_lock_child, di_write_lock_child, ++ * di_read_lock_child2, di_write_lock_child2, ++ * di_read_lock_child3, di_write_lock_child3, ++ * di_read_lock_parent, di_write_lock_parent, ++ * di_read_lock_parent2, di_write_lock_parent2, ++ * di_read_lock_parent3, di_write_lock_parent3, ++ */ ++#define AuReadLockFunc(name, lsc) \ ++static inline void di_read_lock_##name(struct dentry *d, int flags) \ ++{ di_read_lock(d, flags, AuLsc_DI_##lsc); } ++ ++#define AuWriteLockFunc(name, lsc) \ ++static inline void di_write_lock_##name(struct dentry *d) \ ++{ di_write_lock(d, AuLsc_DI_##lsc); } ++ ++#define AuRWLockFuncs(name, lsc) \ ++ AuReadLockFunc(name, lsc) \ ++ AuWriteLockFunc(name, lsc) ++ ++AuRWLockFuncs(child, CHILD); ++AuRWLockFuncs(child2, CHILD2); ++AuRWLockFuncs(child3, CHILD3); ++AuRWLockFuncs(parent, PARENT); ++AuRWLockFuncs(parent2, PARENT2); ++AuRWLockFuncs(parent3, PARENT3); ++ ++#undef AuReadLockFunc ++#undef AuWriteLockFunc ++#undef AuRWLockFuncs ++ ++#define DiMustNoWaiters(d) AuRwMustNoWaiters(&au_di(d)->di_rwsem) ++#define DiMustAnyLock(d) AuRwMustAnyLock(&au_di(d)->di_rwsem) ++#define DiMustWriteLock(d) AuRwMustWriteLock(&au_di(d)->di_rwsem) ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* todo: memory barrier? */ ++static inline unsigned int au_digen(struct dentry *d) ++{ ++ return atomic_read(&au_di(d)->di_generation); ++} ++ ++static inline void au_h_dentry_init(struct au_hdentry *hdentry) ++{ ++ hdentry->hd_dentry = NULL; ++} ++ ++static inline void au_hdput(struct au_hdentry *hd) ++{ ++ dput(hd->hd_dentry); ++} ++ ++static inline aufs_bindex_t au_dbstart(struct dentry *dentry) ++{ ++ DiMustAnyLock(dentry); ++ return au_di(dentry)->di_bstart; ++} ++ ++static inline aufs_bindex_t au_dbend(struct dentry *dentry) ++{ ++ DiMustAnyLock(dentry); ++ return au_di(dentry)->di_bend; ++} ++ ++static inline aufs_bindex_t au_dbwh(struct dentry *dentry) ++{ ++ DiMustAnyLock(dentry); ++ return au_di(dentry)->di_bwh; ++} ++ ++static inline aufs_bindex_t au_dbdiropq(struct dentry *dentry) ++{ ++ DiMustAnyLock(dentry); ++ return au_di(dentry)->di_bdiropq; ++} ++ ++/* todo: hard/soft set? */ ++static inline void au_set_dbstart(struct dentry *dentry, aufs_bindex_t bindex) ++{ ++ DiMustWriteLock(dentry); ++ au_di(dentry)->di_bstart = bindex; ++} ++ ++static inline void au_set_dbend(struct dentry *dentry, aufs_bindex_t bindex) ++{ ++ DiMustWriteLock(dentry); ++ au_di(dentry)->di_bend = bindex; ++} ++ ++static inline void au_set_dbwh(struct dentry *dentry, aufs_bindex_t bindex) ++{ ++ DiMustWriteLock(dentry); ++ /* dbwh can be outside of bstart - bend range */ ++ au_di(dentry)->di_bwh = bindex; ++} ++ ++static inline void au_set_dbdiropq(struct dentry *dentry, aufs_bindex_t bindex) ++{ ++ DiMustWriteLock(dentry); ++ au_di(dentry)->di_bdiropq = bindex; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++#ifdef CONFIG_AUFS_HINOTIFY ++static inline void au_digen_dec(struct dentry *d) ++{ ++ atomic_dec_return(&au_di(d)->di_generation); ++} ++ ++static inline void au_hin_di_reinit(struct dentry *dentry) ++{ ++ dentry->d_fsdata = NULL; ++} ++#else ++static inline void au_hin_di_reinit(struct dentry *dentry __maybe_unused) ++{ ++ /* empty */ ++} ++#endif /* CONFIG_AUFS_HINOTIFY */ ++ ++#endif /* __KERNEL__ */ ++#endif /* __AUFS_DENTRY_H__ */ +diff -Nur linux-2.6.31.4.orig/fs/aufs/dinfo.c linux-2.6.31.4/fs/aufs/dinfo.c +--- linux-2.6.31.4.orig/fs/aufs/dinfo.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/dinfo.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,367 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * dentry private data ++ */ ++ ++#include "aufs.h" ++ ++int au_alloc_dinfo(struct dentry *dentry) ++{ ++ struct au_dinfo *dinfo; ++ struct super_block *sb; ++ int nbr; ++ ++ dinfo = au_cache_alloc_dinfo(); ++ if (unlikely(!dinfo)) ++ goto out; ++ ++ sb = dentry->d_sb; ++ nbr = au_sbend(sb) + 1; ++ if (nbr <= 0) ++ nbr = 1; ++ dinfo->di_hdentry = kcalloc(nbr, sizeof(*dinfo->di_hdentry), GFP_NOFS); ++ if (unlikely(!dinfo->di_hdentry)) ++ goto out_dinfo; ++ ++ atomic_set(&dinfo->di_generation, au_sigen(sb)); ++ /* smp_mb(); */ /* atomic_set */ ++ au_rw_init_wlock_nested(&dinfo->di_rwsem, AuLsc_DI_CHILD); ++ dinfo->di_bstart = -1; ++ dinfo->di_bend = -1; ++ dinfo->di_bwh = -1; ++ dinfo->di_bdiropq = -1; ++ ++ dentry->d_fsdata = dinfo; ++ dentry->d_op = &aufs_dop; ++ return 0; /* success */ ++ ++ out_dinfo: ++ au_cache_free_dinfo(dinfo); ++ out: ++ return -ENOMEM; ++} ++ ++int au_di_realloc(struct au_dinfo *dinfo, int nbr) ++{ ++ int err, sz; ++ struct au_hdentry *hdp; ++ ++ AuRwMustWriteLock(&dinfo->di_rwsem); ++ ++ err = -ENOMEM; ++ sz = sizeof(*hdp) * (dinfo->di_bend + 1); ++ if (!sz) ++ sz = sizeof(*hdp); ++ hdp = au_kzrealloc(dinfo->di_hdentry, sz, sizeof(*hdp) * nbr, GFP_NOFS); ++ if (hdp) { ++ dinfo->di_hdentry = hdp; ++ err = 0; ++ } ++ ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static void do_ii_write_lock(struct inode *inode, unsigned int lsc) ++{ ++ switch (lsc) { ++ case AuLsc_DI_CHILD: ++ ii_write_lock_child(inode); ++ break; ++ case AuLsc_DI_CHILD2: ++ ii_write_lock_child2(inode); ++ break; ++ case AuLsc_DI_CHILD3: ++ ii_write_lock_child3(inode); ++ break; ++ case AuLsc_DI_PARENT: ++ ii_write_lock_parent(inode); ++ break; ++ case AuLsc_DI_PARENT2: ++ ii_write_lock_parent2(inode); ++ break; ++ case AuLsc_DI_PARENT3: ++ ii_write_lock_parent3(inode); ++ break; ++ default: ++ BUG(); ++ } ++} ++ ++static void do_ii_read_lock(struct inode *inode, unsigned int lsc) ++{ ++ switch (lsc) { ++ case AuLsc_DI_CHILD: ++ ii_read_lock_child(inode); ++ break; ++ case AuLsc_DI_CHILD2: ++ ii_read_lock_child2(inode); ++ break; ++ case AuLsc_DI_CHILD3: ++ ii_read_lock_child3(inode); ++ break; ++ case AuLsc_DI_PARENT: ++ ii_read_lock_parent(inode); ++ break; ++ case AuLsc_DI_PARENT2: ++ ii_read_lock_parent2(inode); ++ break; ++ case AuLsc_DI_PARENT3: ++ ii_read_lock_parent3(inode); ++ break; ++ default: ++ BUG(); ++ } ++} ++ ++void di_read_lock(struct dentry *d, int flags, unsigned int lsc) ++{ ++ au_rw_read_lock_nested(&au_di(d)->di_rwsem, lsc); ++ if (d->d_inode) { ++ if (au_ftest_lock(flags, IW)) ++ do_ii_write_lock(d->d_inode, lsc); ++ else if (au_ftest_lock(flags, IR)) ++ do_ii_read_lock(d->d_inode, lsc); ++ } ++} ++ ++void di_read_unlock(struct dentry *d, int flags) ++{ ++ if (d->d_inode) { ++ if (au_ftest_lock(flags, IW)) ++ ii_write_unlock(d->d_inode); ++ else if (au_ftest_lock(flags, IR)) ++ ii_read_unlock(d->d_inode); ++ } ++ au_rw_read_unlock(&au_di(d)->di_rwsem); ++} ++ ++void di_downgrade_lock(struct dentry *d, int flags) ++{ ++ if (d->d_inode && au_ftest_lock(flags, IR)) ++ ii_downgrade_lock(d->d_inode); ++ au_rw_dgrade_lock(&au_di(d)->di_rwsem); ++} ++ ++void di_write_lock(struct dentry *d, unsigned int lsc) ++{ ++ au_rw_write_lock_nested(&au_di(d)->di_rwsem, lsc); ++ if (d->d_inode) ++ do_ii_write_lock(d->d_inode, lsc); ++} ++ ++void di_write_unlock(struct dentry *d) ++{ ++ if (d->d_inode) ++ ii_write_unlock(d->d_inode); ++ au_rw_write_unlock(&au_di(d)->di_rwsem); ++} ++ ++void di_write_lock2_child(struct dentry *d1, struct dentry *d2, int isdir) ++{ ++ AuDebugOn(d1 == d2 ++ || d1->d_inode == d2->d_inode ++ || d1->d_sb != d2->d_sb); ++ ++ if (isdir && au_test_subdir(d1, d2)) { ++ di_write_lock_child(d1); ++ di_write_lock_child2(d2); ++ } else { ++ /* there should be no races */ ++ di_write_lock_child(d2); ++ di_write_lock_child2(d1); ++ } ++} ++ ++void di_write_lock2_parent(struct dentry *d1, struct dentry *d2, int isdir) ++{ ++ AuDebugOn(d1 == d2 ++ || d1->d_inode == d2->d_inode ++ || d1->d_sb != d2->d_sb); ++ ++ if (isdir && au_test_subdir(d1, d2)) { ++ di_write_lock_parent(d1); ++ di_write_lock_parent2(d2); ++ } else { ++ /* there should be no races */ ++ di_write_lock_parent(d2); ++ di_write_lock_parent2(d1); ++ } ++} ++ ++void di_write_unlock2(struct dentry *d1, struct dentry *d2) ++{ ++ di_write_unlock(d1); ++ if (d1->d_inode == d2->d_inode) ++ au_rw_write_unlock(&au_di(d2)->di_rwsem); ++ else ++ di_write_unlock(d2); ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++struct dentry *au_h_dptr(struct dentry *dentry, aufs_bindex_t bindex) ++{ ++ struct dentry *d; ++ ++ DiMustAnyLock(dentry); ++ ++ if (au_dbstart(dentry) < 0 || bindex < au_dbstart(dentry)) ++ return NULL; ++ AuDebugOn(bindex < 0); ++ d = au_di(dentry)->di_hdentry[0 + bindex].hd_dentry; ++ AuDebugOn(d && (atomic_read(&d->d_count) <= 0)); ++ return d; ++} ++ ++aufs_bindex_t au_dbtail(struct dentry *dentry) ++{ ++ aufs_bindex_t bend, bwh; ++ ++ bend = au_dbend(dentry); ++ if (0 <= bend) { ++ bwh = au_dbwh(dentry); ++ if (!bwh) ++ return bwh; ++ if (0 < bwh && bwh < bend) ++ return bwh - 1; ++ } ++ return bend; ++} ++ ++aufs_bindex_t au_dbtaildir(struct dentry *dentry) ++{ ++ aufs_bindex_t bend, bopq; ++ ++ bend = au_dbtail(dentry); ++ if (0 <= bend) { ++ bopq = au_dbdiropq(dentry); ++ if (0 <= bopq && bopq < bend) ++ bend = bopq; ++ } ++ return bend; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++void au_set_h_dptr(struct dentry *dentry, aufs_bindex_t bindex, ++ struct dentry *h_dentry) ++{ ++ struct au_hdentry *hd = au_di(dentry)->di_hdentry + bindex; ++ ++ DiMustWriteLock(dentry); ++ ++ if (hd->hd_dentry) ++ au_hdput(hd); ++ hd->hd_dentry = h_dentry; ++} ++ ++void au_update_digen(struct dentry *dentry) ++{ ++ atomic_set(&au_di(dentry)->di_generation, au_sigen(dentry->d_sb)); ++ /* smp_mb(); */ /* atomic_set */ ++} ++ ++void au_update_dbrange(struct dentry *dentry, int do_put_zero) ++{ ++ struct au_dinfo *dinfo; ++ struct dentry *h_d; ++ ++ DiMustWriteLock(dentry); ++ ++ dinfo = au_di(dentry); ++ if (!dinfo || dinfo->di_bstart < 0) ++ return; ++ ++ if (do_put_zero) { ++ aufs_bindex_t bindex, bend; ++ ++ bend = dinfo->di_bend; ++ for (bindex = dinfo->di_bstart; bindex <= bend; bindex++) { ++ h_d = dinfo->di_hdentry[0 + bindex].hd_dentry; ++ if (h_d && !h_d->d_inode) ++ au_set_h_dptr(dentry, bindex, NULL); ++ } ++ } ++ ++ dinfo->di_bstart = -1; ++ while (++dinfo->di_bstart <= dinfo->di_bend) ++ if (dinfo->di_hdentry[0 + dinfo->di_bstart].hd_dentry) ++ break; ++ if (dinfo->di_bstart > dinfo->di_bend) { ++ dinfo->di_bstart = -1; ++ dinfo->di_bend = -1; ++ return; ++ } ++ ++ dinfo->di_bend++; ++ while (0 <= --dinfo->di_bend) ++ if (dinfo->di_hdentry[0 + dinfo->di_bend].hd_dentry) ++ break; ++ AuDebugOn(dinfo->di_bstart > dinfo->di_bend || dinfo->di_bend < 0); ++} ++ ++void au_update_dbstart(struct dentry *dentry) ++{ ++ aufs_bindex_t bindex, bend; ++ struct dentry *h_dentry; ++ ++ bend = au_dbend(dentry); ++ for (bindex = au_dbstart(dentry); bindex <= bend; bindex++) { ++ h_dentry = au_h_dptr(dentry, bindex); ++ if (!h_dentry) ++ continue; ++ if (h_dentry->d_inode) { ++ au_set_dbstart(dentry, bindex); ++ return; ++ } ++ au_set_h_dptr(dentry, bindex, NULL); ++ } ++} ++ ++void au_update_dbend(struct dentry *dentry) ++{ ++ aufs_bindex_t bindex, bstart; ++ struct dentry *h_dentry; ++ ++ bstart = au_dbstart(dentry); ++ for (bindex = au_dbend(dentry); bindex <= bstart; bindex--) { ++ h_dentry = au_h_dptr(dentry, bindex); ++ if (!h_dentry) ++ continue; ++ if (h_dentry->d_inode) { ++ au_set_dbend(dentry, bindex); ++ return; ++ } ++ au_set_h_dptr(dentry, bindex, NULL); ++ } ++} ++ ++int au_find_dbindex(struct dentry *dentry, struct dentry *h_dentry) ++{ ++ aufs_bindex_t bindex, bend; ++ ++ bend = au_dbend(dentry); ++ for (bindex = au_dbstart(dentry); bindex <= bend; bindex++) ++ if (au_h_dptr(dentry, bindex) == h_dentry) ++ return bindex; ++ return -1; ++} +diff -Nur linux-2.6.31.4.orig/fs/aufs/dir.c linux-2.6.31.4/fs/aufs/dir.c +--- linux-2.6.31.4.orig/fs/aufs/dir.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/dir.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,538 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * directory operations ++ */ ++ ++#include <linux/file.h> ++#include <linux/fs_stack.h> ++#include "aufs.h" ++ ++void au_add_nlink(struct inode *dir, struct inode *h_dir) ++{ ++ AuDebugOn(!S_ISDIR(dir->i_mode) || !S_ISDIR(h_dir->i_mode)); ++ ++ dir->i_nlink += h_dir->i_nlink - 2; ++ if (h_dir->i_nlink < 2) ++ dir->i_nlink += 2; ++} ++ ++void au_sub_nlink(struct inode *dir, struct inode *h_dir) ++{ ++ AuDebugOn(!S_ISDIR(dir->i_mode) || !S_ISDIR(h_dir->i_mode)); ++ ++ dir->i_nlink -= h_dir->i_nlink - 2; ++ if (h_dir->i_nlink < 2) ++ dir->i_nlink -= 2; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static int reopen_dir(struct file *file) ++{ ++ int err; ++ unsigned int flags; ++ aufs_bindex_t bindex, btail, bstart; ++ struct dentry *dentry, *h_dentry; ++ struct file *h_file; ++ ++ /* open all lower dirs */ ++ dentry = file->f_dentry; ++ bstart = au_dbstart(dentry); ++ for (bindex = au_fbstart(file); bindex < bstart; bindex++) ++ au_set_h_fptr(file, bindex, NULL); ++ au_set_fbstart(file, bstart); ++ ++ btail = au_dbtaildir(dentry); ++ for (bindex = au_fbend(file); btail < bindex; bindex--) ++ au_set_h_fptr(file, bindex, NULL); ++ au_set_fbend(file, btail); ++ ++ flags = file->f_flags; ++ for (bindex = bstart; bindex <= btail; bindex++) { ++ h_dentry = au_h_dptr(dentry, bindex); ++ if (!h_dentry) ++ continue; ++ h_file = au_h_fptr(file, bindex); ++ if (h_file) ++ continue; ++ ++ h_file = au_h_open(dentry, bindex, flags, file); ++ err = PTR_ERR(h_file); ++ if (IS_ERR(h_file)) ++ goto out; /* close all? */ ++ au_set_h_fptr(file, bindex, h_file); ++ } ++ au_update_figen(file); ++ /* todo: necessary? */ ++ /* file->f_ra = h_file->f_ra; */ ++ err = 0; ++ ++ out: ++ return err; ++} ++ ++static int do_open_dir(struct file *file, int flags) ++{ ++ int err; ++ aufs_bindex_t bindex, btail; ++ struct dentry *dentry, *h_dentry; ++ struct file *h_file; ++ ++ FiMustWriteLock(file); ++ ++ err = 0; ++ dentry = file->f_dentry; ++ au_set_fvdir_cache(file, NULL); ++ au_fi(file)->fi_maintain_plink = 0; ++ file->f_version = dentry->d_inode->i_version; ++ bindex = au_dbstart(dentry); ++ au_set_fbstart(file, bindex); ++ btail = au_dbtaildir(dentry); ++ au_set_fbend(file, btail); ++ for (; !err && bindex <= btail; bindex++) { ++ h_dentry = au_h_dptr(dentry, bindex); ++ if (!h_dentry) ++ continue; ++ ++ h_file = au_h_open(dentry, bindex, flags, file); ++ if (IS_ERR(h_file)) { ++ err = PTR_ERR(h_file); ++ break; ++ } ++ au_set_h_fptr(file, bindex, h_file); ++ } ++ au_update_figen(file); ++ /* todo: necessary? */ ++ /* file->f_ra = h_file->f_ra; */ ++ if (!err) ++ return 0; /* success */ ++ ++ /* close all */ ++ for (bindex = au_fbstart(file); bindex <= btail; bindex++) ++ au_set_h_fptr(file, bindex, NULL); ++ au_set_fbstart(file, -1); ++ au_set_fbend(file, -1); ++ return err; ++} ++ ++static int aufs_open_dir(struct inode *inode __maybe_unused, ++ struct file *file) ++{ ++ return au_do_open(file, do_open_dir); ++} ++ ++static int aufs_release_dir(struct inode *inode __maybe_unused, ++ struct file *file) ++{ ++ struct au_vdir *vdir_cache; ++ struct super_block *sb; ++ struct au_sbinfo *sbinfo; ++ ++ sb = file->f_dentry->d_sb; ++ si_noflush_read_lock(sb); ++ fi_write_lock(file); ++ vdir_cache = au_fvdir_cache(file); ++ if (vdir_cache) ++ au_vdir_free(vdir_cache); ++ if (au_fi(file)->fi_maintain_plink) { ++ sbinfo = au_sbi(sb); ++ /* clear the flag without write-lock */ ++ sbinfo->au_si_status &= ~AuSi_MAINTAIN_PLINK; ++ smp_mb(); ++ wake_up_all(&sbinfo->si_plink_wq); ++ } ++ fi_write_unlock(file); ++ au_finfo_fin(file); ++ si_read_unlock(sb); ++ return 0; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static int au_do_fsync_dir_no_file(struct dentry *dentry, int datasync) ++{ ++ int err; ++ aufs_bindex_t bend, bindex; ++ struct inode *inode; ++ struct super_block *sb; ++ ++ err = 0; ++ sb = dentry->d_sb; ++ inode = dentry->d_inode; ++ IMustLock(inode); ++ bend = au_dbend(dentry); ++ for (bindex = au_dbstart(dentry); !err && bindex <= bend; bindex++) { ++ struct path h_path; ++ struct inode *h_inode; ++ ++ if (au_test_ro(sb, bindex, inode)) ++ continue; ++ h_path.dentry = au_h_dptr(dentry, bindex); ++ if (!h_path.dentry) ++ continue; ++ h_inode = h_path.dentry->d_inode; ++ if (!h_inode) ++ continue; ++ ++ /* no mnt_want_write() */ ++ /* cf. fs/nsfd/vfs.c and fs/nfsd/nfs4recover.c */ ++ /* todo: inotiry fired? */ ++ h_path.mnt = au_sbr_mnt(sb, bindex); ++ mutex_lock(&h_inode->i_mutex); ++ err = filemap_fdatawrite(h_inode->i_mapping); ++ AuDebugOn(!h_inode->i_fop); ++ if (!err && h_inode->i_fop->fsync) ++ err = h_inode->i_fop->fsync(NULL, h_path.dentry, ++ datasync); ++ if (!err) ++ err = filemap_fdatawrite(h_inode->i_mapping); ++ if (!err) ++ vfsub_update_h_iattr(&h_path, /*did*/NULL); /*ignore*/ ++ mutex_unlock(&h_inode->i_mutex); ++ } ++ ++ return err; ++} ++ ++static int au_do_fsync_dir(struct file *file, int datasync) ++{ ++ int err; ++ aufs_bindex_t bend, bindex; ++ struct file *h_file; ++ struct super_block *sb; ++ struct inode *inode; ++ struct mutex *h_mtx; ++ ++ err = au_reval_and_lock_fdi(file, reopen_dir, /*wlock*/1); ++ if (unlikely(err)) ++ goto out; ++ ++ sb = file->f_dentry->d_sb; ++ inode = file->f_dentry->d_inode; ++ bend = au_fbend(file); ++ for (bindex = au_fbstart(file); !err && bindex <= bend; bindex++) { ++ h_file = au_h_fptr(file, bindex); ++ if (!h_file || au_test_ro(sb, bindex, inode)) ++ continue; ++ ++ err = vfs_fsync(h_file, h_file->f_dentry, datasync); ++ if (!err) { ++ h_mtx = &h_file->f_dentry->d_inode->i_mutex; ++ mutex_lock(h_mtx); ++ vfsub_update_h_iattr(&h_file->f_path, /*did*/NULL); ++ /*ignore*/ ++ mutex_unlock(h_mtx); ++ } ++ } ++ ++ out: ++ return err; ++} ++ ++/* ++ * @file may be NULL ++ */ ++static int aufs_fsync_dir(struct file *file, struct dentry *dentry, ++ int datasync) ++{ ++ int err; ++ struct super_block *sb; ++ ++ IMustLock(dentry->d_inode); ++ ++ err = 0; ++ sb = dentry->d_sb; ++ si_noflush_read_lock(sb); ++ if (file) ++ err = au_do_fsync_dir(file, datasync); ++ else { ++ di_write_lock_child(dentry); ++ err = au_do_fsync_dir_no_file(dentry, datasync); ++ } ++ au_cpup_attr_timesizes(dentry->d_inode); ++ di_write_unlock(dentry); ++ if (file) ++ fi_write_unlock(file); ++ ++ si_read_unlock(sb); ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static int aufs_readdir(struct file *file, void *dirent, filldir_t filldir) ++{ ++ int err; ++ struct dentry *dentry; ++ struct inode *inode; ++ struct super_block *sb; ++ ++ dentry = file->f_dentry; ++ inode = dentry->d_inode; ++ IMustLock(inode); ++ ++ sb = dentry->d_sb; ++ si_read_lock(sb, AuLock_FLUSH); ++ err = au_reval_and_lock_fdi(file, reopen_dir, /*wlock*/1); ++ if (unlikely(err)) ++ goto out; ++ err = au_vdir_init(file); ++ di_downgrade_lock(dentry, AuLock_IR); ++ if (unlikely(err)) ++ goto out_unlock; ++ ++ if (!au_test_nfsd(current)) { ++ err = au_vdir_fill_de(file, dirent, filldir); ++ fsstack_copy_attr_atime(inode, ++ au_h_iptr(inode, au_ibstart(inode))); ++ } else { ++ /* ++ * nfsd filldir may call lookup_one_len(), vfs_getattr(), ++ * encode_fh() and others. ++ */ ++ struct inode *h_inode = au_h_iptr(inode, au_ibstart(inode)); ++ ++ di_read_unlock(dentry, AuLock_IR); ++ si_read_unlock(sb); ++ lockdep_off(); ++ err = au_vdir_fill_de(file, dirent, filldir); ++ lockdep_on(); ++ fsstack_copy_attr_atime(inode, h_inode); ++ fi_write_unlock(file); ++ ++ AuTraceErr(err); ++ return err; ++ } ++ ++ out_unlock: ++ di_read_unlock(dentry, AuLock_IR); ++ fi_write_unlock(file); ++ out: ++ si_read_unlock(sb); ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++#define AuTestEmpty_WHONLY 1 ++#define AuTestEmpty_CALLED (1 << 1) ++#define AuTestEmpty_SHWH (1 << 2) ++#define au_ftest_testempty(flags, name) ((flags) & AuTestEmpty_##name) ++#define au_fset_testempty(flags, name) { (flags) |= AuTestEmpty_##name; } ++#define au_fclr_testempty(flags, name) { (flags) &= ~AuTestEmpty_##name; } ++ ++#ifndef CONFIG_AUFS_SHWH ++#undef AuTestEmpty_SHWH ++#define AuTestEmpty_SHWH 0 ++#endif ++ ++struct test_empty_arg { ++ struct au_nhash whlist; ++ unsigned int flags; ++ int err; ++ aufs_bindex_t bindex; ++}; ++ ++static int test_empty_cb(void *__arg, const char *__name, int namelen, ++ loff_t offset __maybe_unused, u64 ino, ++ unsigned int d_type) ++{ ++ struct test_empty_arg *arg = __arg; ++ char *name = (void *)__name; ++ ++ arg->err = 0; ++ au_fset_testempty(arg->flags, CALLED); ++ /* smp_mb(); */ ++ if (name[0] == '.' ++ && (namelen == 1 || (name[1] == '.' && namelen == 2))) ++ goto out; /* success */ ++ ++ if (namelen <= AUFS_WH_PFX_LEN ++ || memcmp(name, AUFS_WH_PFX, AUFS_WH_PFX_LEN)) { ++ if (au_ftest_testempty(arg->flags, WHONLY) ++ && !au_nhash_test_known_wh(&arg->whlist, name, namelen)) ++ arg->err = -ENOTEMPTY; ++ goto out; ++ } ++ ++ name += AUFS_WH_PFX_LEN; ++ namelen -= AUFS_WH_PFX_LEN; ++ if (!au_nhash_test_known_wh(&arg->whlist, name, namelen)) ++ arg->err = au_nhash_append_wh ++ (&arg->whlist, name, namelen, ino, d_type, arg->bindex, ++ au_ftest_testempty(arg->flags, SHWH)); ++ ++ out: ++ /* smp_mb(); */ ++ AuTraceErr(arg->err); ++ return arg->err; ++} ++ ++static int do_test_empty(struct dentry *dentry, struct test_empty_arg *arg) ++{ ++ int err; ++ struct file *h_file; ++ ++ h_file = au_h_open(dentry, arg->bindex, ++ O_RDONLY | O_NONBLOCK | O_DIRECTORY | O_LARGEFILE, ++ /*file*/NULL); ++ err = PTR_ERR(h_file); ++ if (IS_ERR(h_file)) ++ goto out; ++ ++ err = 0; ++ if (!au_opt_test(au_mntflags(dentry->d_sb), UDBA_NONE) ++ && !h_file->f_dentry->d_inode->i_nlink) ++ goto out_put; ++ ++ do { ++ arg->err = 0; ++ au_fclr_testempty(arg->flags, CALLED); ++ /* smp_mb(); */ ++ err = vfsub_readdir(h_file, test_empty_cb, arg); ++ if (err >= 0) ++ err = arg->err; ++ } while (!err && au_ftest_testempty(arg->flags, CALLED)); ++ ++ out_put: ++ fput(h_file); ++ au_sbr_put(dentry->d_sb, arg->bindex); ++ out: ++ return err; ++} ++ ++struct do_test_empty_args { ++ int *errp; ++ struct dentry *dentry; ++ struct test_empty_arg *arg; ++}; ++ ++static void call_do_test_empty(void *args) ++{ ++ struct do_test_empty_args *a = args; ++ *a->errp = do_test_empty(a->dentry, a->arg); ++} ++ ++static int sio_test_empty(struct dentry *dentry, struct test_empty_arg *arg) ++{ ++ int err, wkq_err; ++ struct dentry *h_dentry; ++ struct inode *h_inode; ++ ++ h_dentry = au_h_dptr(dentry, arg->bindex); ++ h_inode = h_dentry->d_inode; ++ mutex_lock_nested(&h_inode->i_mutex, AuLsc_I_CHILD); ++ err = au_test_h_perm_sio(h_inode, MAY_EXEC | MAY_READ); ++ mutex_unlock(&h_inode->i_mutex); ++ if (!err) ++ err = do_test_empty(dentry, arg); ++ else { ++ struct do_test_empty_args args = { ++ .errp = &err, ++ .dentry = dentry, ++ .arg = arg ++ }; ++ unsigned int flags = arg->flags; ++ ++ wkq_err = au_wkq_wait(call_do_test_empty, &args); ++ if (unlikely(wkq_err)) ++ err = wkq_err; ++ arg->flags = flags; ++ } ++ ++ return err; ++} ++ ++int au_test_empty_lower(struct dentry *dentry) ++{ ++ int err; ++ aufs_bindex_t bindex, bstart, btail; ++ struct test_empty_arg arg; ++ ++ SiMustAnyLock(dentry->d_sb); ++ ++ err = au_nhash_alloc(&arg.whlist, au_sbi(dentry->d_sb)->si_rdhash, ++ GFP_NOFS); ++ if (unlikely(err)) ++ goto out; ++ ++ bstart = au_dbstart(dentry); ++ arg.flags = 0; ++ if (au_opt_test(au_mntflags(dentry->d_sb), SHWH)) ++ au_fset_testempty(arg.flags, SHWH); ++ arg.bindex = bstart; ++ err = do_test_empty(dentry, &arg); ++ if (unlikely(err)) ++ goto out_whlist; ++ ++ au_fset_testempty(arg.flags, WHONLY); ++ btail = au_dbtaildir(dentry); ++ for (bindex = bstart + 1; !err && bindex <= btail; bindex++) { ++ struct dentry *h_dentry; ++ ++ h_dentry = au_h_dptr(dentry, bindex); ++ if (h_dentry && h_dentry->d_inode) { ++ arg.bindex = bindex; ++ err = do_test_empty(dentry, &arg); ++ } ++ } ++ ++ out_whlist: ++ au_nhash_wh_free(&arg.whlist); ++ out: ++ return err; ++} ++ ++int au_test_empty(struct dentry *dentry, struct au_nhash *whlist) ++{ ++ int err; ++ struct test_empty_arg arg; ++ aufs_bindex_t bindex, btail; ++ ++ err = 0; ++ arg.whlist = *whlist; ++ arg.flags = AuTestEmpty_WHONLY; ++ if (au_opt_test(au_mntflags(dentry->d_sb), SHWH)) ++ au_fset_testempty(arg.flags, SHWH); ++ btail = au_dbtaildir(dentry); ++ for (bindex = au_dbstart(dentry); !err && bindex <= btail; bindex++) { ++ struct dentry *h_dentry; ++ ++ h_dentry = au_h_dptr(dentry, bindex); ++ if (h_dentry && h_dentry->d_inode) { ++ arg.bindex = bindex; ++ err = sio_test_empty(dentry, &arg); ++ } ++ } ++ ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++const struct file_operations aufs_dir_fop = { ++ .read = generic_read_dir, ++ .readdir = aufs_readdir, ++ .unlocked_ioctl = aufs_ioctl_dir, ++ .open = aufs_open_dir, ++ .release = aufs_release_dir, ++ .flush = aufs_flush, ++ .fsync = aufs_fsync_dir ++}; +diff -Nur linux-2.6.31.4.orig/fs/aufs/dir.h linux-2.6.31.4/fs/aufs/dir.h +--- linux-2.6.31.4.orig/fs/aufs/dir.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/dir.h 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,114 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * directory operations ++ */ ++ ++#ifndef __AUFS_DIR_H__ ++#define __AUFS_DIR_H__ ++ ++#ifdef __KERNEL__ ++ ++#include <linux/fs.h> ++#include <linux/aufs_type.h> ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* need to be faster and smaller */ ++ ++struct au_nhash { ++ unsigned int nh_num; ++ struct hlist_head *nh_head; ++}; ++ ++struct au_vdir_destr { ++ unsigned char len; ++ unsigned char name[0]; ++} __packed; ++ ++struct au_vdir_dehstr { ++ struct hlist_node hash; ++ struct au_vdir_destr *str; ++}; ++ ++struct au_vdir_de { ++ ino_t de_ino; ++ unsigned char de_type; ++ /* caution: packed */ ++ struct au_vdir_destr de_str; ++} __packed; ++ ++struct au_vdir_wh { ++ struct hlist_node wh_hash; ++#ifdef CONFIG_AUFS_SHWH ++ ino_t wh_ino; ++ aufs_bindex_t wh_bindex; ++ unsigned char wh_type; ++#else ++ aufs_bindex_t wh_bindex; ++#endif ++ /* caution: packed */ ++ struct au_vdir_destr wh_str; ++} __packed; ++ ++union au_vdir_deblk_p { ++ unsigned char *deblk; ++ struct au_vdir_de *de; ++}; ++ ++struct au_vdir { ++ unsigned char **vd_deblk; ++ unsigned long vd_nblk; ++ struct { ++ unsigned long ul; ++ union au_vdir_deblk_p p; ++ } vd_last; ++ ++ unsigned long vd_version; ++ unsigned int vd_deblk_sz; ++ unsigned long vd_jiffy; ++}; ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* dir.c */ ++extern const struct file_operations aufs_dir_fop; ++void au_add_nlink(struct inode *dir, struct inode *h_dir); ++void au_sub_nlink(struct inode *dir, struct inode *h_dir); ++int au_test_empty_lower(struct dentry *dentry); ++int au_test_empty(struct dentry *dentry, struct au_nhash *whlist); ++ ++/* vdir.c */ ++int au_nhash_alloc(struct au_nhash *nhash, unsigned int num_hash, gfp_t gfp); ++void au_nhash_wh_free(struct au_nhash *whlist); ++int au_nhash_test_longer_wh(struct au_nhash *whlist, aufs_bindex_t btgt, ++ int limit); ++int au_nhash_test_known_wh(struct au_nhash *whlist, char *name, int nlen); ++int au_nhash_append_wh(struct au_nhash *whlist, char *name, int nlen, ino_t ino, ++ unsigned int d_type, aufs_bindex_t bindex, ++ unsigned char shwh); ++void au_vdir_free(struct au_vdir *vdir); ++int au_vdir_init(struct file *file); ++int au_vdir_fill_de(struct file *file, void *dirent, filldir_t filldir); ++ ++/* ioctl.c */ ++long aufs_ioctl_dir(struct file *file, unsigned int cmd, unsigned long arg); ++ ++#endif /* __KERNEL__ */ ++#endif /* __AUFS_DIR_H__ */ +diff -Nur linux-2.6.31.4.orig/fs/aufs/export.c linux-2.6.31.4/fs/aufs/export.c +--- linux-2.6.31.4.orig/fs/aufs/export.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/export.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,745 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * export via nfs ++ */ ++ ++#include <linux/exportfs.h> ++#include <linux/file.h> ++#include <linux/mnt_namespace.h> ++#include <linux/namei.h> ++#include <linux/random.h> ++#include "aufs.h" ++ ++union conv { ++#ifdef CONFIG_AUFS_INO_T_64 ++ __u32 a[2]; ++#else ++ __u32 a[1]; ++#endif ++ ino_t ino; ++}; ++ ++static ino_t decode_ino(__u32 *a) ++{ ++ union conv u; ++ ++ BUILD_BUG_ON(sizeof(u.ino) != sizeof(u.a)); ++ u.a[0] = a[0]; ++#ifdef CONFIG_AUFS_INO_T_64 ++ u.a[1] = a[1]; ++#endif ++ return u.ino; ++} ++ ++static void encode_ino(__u32 *a, ino_t ino) ++{ ++ union conv u; ++ ++ u.ino = ino; ++ a[0] = u.a[0]; ++#ifdef CONFIG_AUFS_INO_T_64 ++ a[1] = u.a[1]; ++#endif ++} ++ ++/* NFS file handle */ ++enum { ++ Fh_br_id, ++ Fh_sigen, ++#ifdef CONFIG_AUFS_INO_T_64 ++ /* support 64bit inode number */ ++ Fh_ino1, ++ Fh_ino2, ++ Fh_dir_ino1, ++ Fh_dir_ino2, ++#else ++ Fh_ino1, ++ Fh_dir_ino1, ++#endif ++ Fh_igen, ++ Fh_h_type, ++ Fh_tail, ++ ++ Fh_ino = Fh_ino1, ++ Fh_dir_ino = Fh_dir_ino1 ++}; ++ ++static int au_test_anon(struct dentry *dentry) ++{ ++ return !!(dentry->d_flags & DCACHE_DISCONNECTED); ++} ++ ++/* ---------------------------------------------------------------------- */ ++/* inode generation external table */ ++ ++int au_xigen_inc(struct inode *inode) ++{ ++ int err; ++ loff_t pos; ++ ssize_t sz; ++ __u32 igen; ++ struct super_block *sb; ++ struct au_sbinfo *sbinfo; ++ ++ err = 0; ++ sb = inode->i_sb; ++ sbinfo = au_sbi(sb); ++ /* ++ * temporary workaround for escaping from SiMustAnyLock() in ++ * au_mntflags(), since this function is called from au_iinfo_fin(). ++ */ ++ if (unlikely(!au_opt_test(sbinfo->si_mntflags, XINO))) ++ goto out; ++ ++ pos = inode->i_ino; ++ pos *= sizeof(igen); ++ igen = inode->i_generation + 1; ++ sz = xino_fwrite(sbinfo->si_xwrite, sbinfo->si_xigen, &igen, ++ sizeof(igen), &pos); ++ if (sz == sizeof(igen)) ++ goto out; /* success */ ++ ++ err = sz; ++ if (unlikely(sz >= 0)) { ++ err = -EIO; ++ AuIOErr("xigen error (%zd)\n", sz); ++ } ++ ++ out: ++ return err; ++} ++ ++int au_xigen_new(struct inode *inode) ++{ ++ int err; ++ loff_t pos; ++ ssize_t sz; ++ struct super_block *sb; ++ struct au_sbinfo *sbinfo; ++ struct file *file; ++ ++ err = 0; ++ /* todo: dirty, at mount time */ ++ if (inode->i_ino == AUFS_ROOT_INO) ++ goto out; ++ sb = inode->i_sb; ++ SiMustAnyLock(sb); ++ if (unlikely(!au_opt_test(au_mntflags(sb), XINO))) ++ goto out; ++ ++ err = -EFBIG; ++ pos = inode->i_ino; ++ if (unlikely(au_loff_max / sizeof(inode->i_generation) - 1 < pos)) { ++ AuIOErr1("too large i%lld\n", pos); ++ goto out; ++ } ++ pos *= sizeof(inode->i_generation); ++ ++ err = 0; ++ sbinfo = au_sbi(sb); ++ file = sbinfo->si_xigen; ++ BUG_ON(!file); ++ ++ if (i_size_read(file->f_dentry->d_inode) ++ < pos + sizeof(inode->i_generation)) { ++ inode->i_generation = atomic_inc_return(&sbinfo->si_xigen_next); ++ sz = xino_fwrite(sbinfo->si_xwrite, file, &inode->i_generation, ++ sizeof(inode->i_generation), &pos); ++ } else ++ sz = xino_fread(sbinfo->si_xread, file, &inode->i_generation, ++ sizeof(inode->i_generation), &pos); ++ if (sz == sizeof(inode->i_generation)) ++ goto out; /* success */ ++ ++ err = sz; ++ if (unlikely(sz >= 0)) { ++ err = -EIO; ++ AuIOErr("xigen error (%zd)\n", sz); ++ } ++ ++ out: ++ return err; ++} ++ ++int au_xigen_set(struct super_block *sb, struct file *base) ++{ ++ int err; ++ struct au_sbinfo *sbinfo; ++ struct file *file; ++ ++ SiMustWriteLock(sb); ++ ++ sbinfo = au_sbi(sb); ++ file = au_xino_create2(base, sbinfo->si_xigen); ++ err = PTR_ERR(file); ++ if (IS_ERR(file)) ++ goto out; ++ err = 0; ++ if (sbinfo->si_xigen) ++ fput(sbinfo->si_xigen); ++ sbinfo->si_xigen = file; ++ ++ out: ++ return err; ++} ++ ++void au_xigen_clr(struct super_block *sb) ++{ ++ struct au_sbinfo *sbinfo; ++ ++ SiMustWriteLock(sb); ++ ++ sbinfo = au_sbi(sb); ++ if (sbinfo->si_xigen) { ++ fput(sbinfo->si_xigen); ++ sbinfo->si_xigen = NULL; ++ } ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static struct dentry *decode_by_ino(struct super_block *sb, ino_t ino, ++ ino_t dir_ino) ++{ ++ struct dentry *dentry, *d; ++ struct inode *inode; ++ unsigned int sigen; ++ ++ dentry = NULL; ++ inode = ilookup(sb, ino); ++ if (!inode) ++ goto out; ++ ++ dentry = ERR_PTR(-ESTALE); ++ sigen = au_sigen(sb); ++ if (unlikely(is_bad_inode(inode) ++ || IS_DEADDIR(inode) ++ || sigen != au_iigen(inode))) ++ goto out_iput; ++ ++ dentry = NULL; ++ if (!dir_ino || S_ISDIR(inode->i_mode)) ++ dentry = d_find_alias(inode); ++ else { ++ spin_lock(&dcache_lock); ++ list_for_each_entry(d, &inode->i_dentry, d_alias) ++ if (!au_test_anon(d) ++ && d->d_parent->d_inode->i_ino == dir_ino) { ++ dentry = dget_locked(d); ++ break; ++ } ++ spin_unlock(&dcache_lock); ++ } ++ if (unlikely(dentry && sigen != au_digen(dentry))) { ++ dput(dentry); ++ dentry = ERR_PTR(-ESTALE); ++ } ++ ++ out_iput: ++ iput(inode); ++ out: ++ return dentry; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* todo: dirty? */ ++/* if exportfs_decode_fh() passed vfsmount*, we could be happy */ ++static struct vfsmount *au_mnt_get(struct super_block *sb) ++{ ++ struct mnt_namespace *ns; ++ struct vfsmount *pos, *mnt; ++ ++ spin_lock(&vfsmount_lock); ++ /* no get/put ?? */ ++ AuDebugOn(!current->nsproxy); ++ ns = current->nsproxy->mnt_ns; ++ AuDebugOn(!ns); ++ mnt = NULL; ++ /* the order (reverse) will not be a problem */ ++ list_for_each_entry(pos, &ns->list, mnt_list) ++ if (pos->mnt_sb == sb) { ++ mnt = mntget(pos); ++ break; ++ } ++ spin_unlock(&vfsmount_lock); ++ AuDebugOn(!mnt); ++ ++ return mnt; ++} ++ ++struct au_nfsd_si_lock { ++ const unsigned int sigen; ++ const aufs_bindex_t br_id; ++ unsigned char force_lock; ++}; ++ ++static aufs_bindex_t si_nfsd_read_lock(struct super_block *sb, ++ struct au_nfsd_si_lock *nsi_lock) ++{ ++ aufs_bindex_t bindex; ++ ++ si_read_lock(sb, AuLock_FLUSH); ++ ++ /* branch id may be wrapped around */ ++ bindex = au_br_index(sb, nsi_lock->br_id); ++ if (bindex >= 0 && nsi_lock->sigen + AUFS_BRANCH_MAX > au_sigen(sb)) ++ goto out; /* success */ ++ ++ if (!nsi_lock->force_lock) ++ si_read_unlock(sb); ++ bindex = -1; ++ ++ out: ++ return bindex; ++} ++ ++struct find_name_by_ino { ++ int called, found; ++ ino_t ino; ++ char *name; ++ int namelen; ++}; ++ ++static int ++find_name_by_ino(void *arg, const char *name, int namelen, loff_t offset, ++ u64 ino, unsigned int d_type) ++{ ++ struct find_name_by_ino *a = arg; ++ ++ a->called++; ++ if (a->ino != ino) ++ return 0; ++ ++ memcpy(a->name, name, namelen); ++ a->namelen = namelen; ++ a->found = 1; ++ return 1; ++} ++ ++static struct dentry *au_lkup_by_ino(struct path *path, ino_t ino, ++ struct au_nfsd_si_lock *nsi_lock) ++{ ++ struct dentry *dentry, *parent; ++ struct file *file; ++ struct inode *dir; ++ struct find_name_by_ino arg; ++ int err; ++ ++ parent = path->dentry; ++ if (nsi_lock) ++ si_read_unlock(parent->d_sb); ++ path_get(path); ++ file = dentry_open(parent, path->mnt, au_dir_roflags, current_cred()); ++ dentry = (void *)file; ++ if (IS_ERR(file)) ++ goto out; ++ ++ dentry = ERR_PTR(-ENOMEM); ++ arg.name = __getname(); ++ if (unlikely(!arg.name)) ++ goto out_file; ++ arg.ino = ino; ++ arg.found = 0; ++ do { ++ arg.called = 0; ++ /* smp_mb(); */ ++ err = vfsub_readdir(file, find_name_by_ino, &arg); ++ } while (!err && !arg.found && arg.called); ++ dentry = ERR_PTR(err); ++ if (unlikely(err)) ++ goto out_name; ++ dentry = ERR_PTR(-ENOENT); ++ if (!arg.found) ++ goto out_name; ++ ++ /* do not call au_lkup_one() */ ++ dir = parent->d_inode; ++ mutex_lock(&dir->i_mutex); ++ dentry = vfsub_lookup_one_len(arg.name, parent, arg.namelen); ++ mutex_unlock(&dir->i_mutex); ++ AuTraceErrPtr(dentry); ++ if (IS_ERR(dentry)) ++ goto out_name; ++ AuDebugOn(au_test_anon(dentry)); ++ if (unlikely(!dentry->d_inode)) { ++ dput(dentry); ++ dentry = ERR_PTR(-ENOENT); ++ } ++ ++ out_name: ++ __putname(arg.name); ++ out_file: ++ fput(file); ++ out: ++ if (unlikely(nsi_lock ++ && si_nfsd_read_lock(parent->d_sb, nsi_lock) < 0)) ++ if (!IS_ERR(dentry)) { ++ dput(dentry); ++ dentry = ERR_PTR(-ESTALE); ++ } ++ AuTraceErrPtr(dentry); ++ return dentry; ++} ++ ++static struct dentry *decode_by_dir_ino(struct super_block *sb, ino_t ino, ++ ino_t dir_ino, ++ struct au_nfsd_si_lock *nsi_lock) ++{ ++ struct dentry *dentry; ++ struct path path; ++ ++ if (dir_ino != AUFS_ROOT_INO) { ++ path.dentry = decode_by_ino(sb, dir_ino, 0); ++ dentry = path.dentry; ++ if (!path.dentry || IS_ERR(path.dentry)) ++ goto out; ++ AuDebugOn(au_test_anon(path.dentry)); ++ } else ++ path.dentry = dget(sb->s_root); ++ ++ path.mnt = au_mnt_get(sb); ++ dentry = au_lkup_by_ino(&path, ino, nsi_lock); ++ path_put(&path); ++ ++ out: ++ AuTraceErrPtr(dentry); ++ return dentry; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static int h_acceptable(void *expv, struct dentry *dentry) ++{ ++ return 1; ++} ++ ++static char *au_build_path(struct dentry *h_parent, struct path *h_rootpath, ++ char *buf, int len, struct super_block *sb) ++{ ++ char *p; ++ int n; ++ struct path path; ++ ++ p = d_path(h_rootpath, buf, len); ++ if (IS_ERR(p)) ++ goto out; ++ n = strlen(p); ++ ++ path.mnt = h_rootpath->mnt; ++ path.dentry = h_parent; ++ p = d_path(&path, buf, len); ++ if (IS_ERR(p)) ++ goto out; ++ if (n != 1) ++ p += n; ++ ++ path.mnt = au_mnt_get(sb); ++ path.dentry = sb->s_root; ++ p = d_path(&path, buf, len - strlen(p)); ++ mntput(path.mnt); ++ if (IS_ERR(p)) ++ goto out; ++ if (n != 1) ++ p[strlen(p)] = '/'; ++ ++ out: ++ AuTraceErrPtr(p); ++ return p; ++} ++ ++static ++struct dentry *decode_by_path(struct super_block *sb, aufs_bindex_t bindex, ++ ino_t ino, __u32 *fh, int fh_len, ++ struct au_nfsd_si_lock *nsi_lock) ++{ ++ struct dentry *dentry, *h_parent, *root; ++ struct super_block *h_sb; ++ char *pathname, *p; ++ struct vfsmount *h_mnt; ++ struct au_branch *br; ++ int err; ++ struct path path; ++ ++ br = au_sbr(sb, bindex); ++ /* au_br_get(br); */ ++ h_mnt = br->br_mnt; ++ h_sb = h_mnt->mnt_sb; ++ /* todo: call lower fh_to_dentry()? fh_to_parent()? */ ++ h_parent = exportfs_decode_fh(h_mnt, (void *)(fh + Fh_tail), ++ fh_len - Fh_tail, fh[Fh_h_type], ++ h_acceptable, /*context*/NULL); ++ dentry = h_parent; ++ if (unlikely(!h_parent || IS_ERR(h_parent))) { ++ AuWarn1("%s decode_fh failed, %ld\n", ++ au_sbtype(h_sb), PTR_ERR(h_parent)); ++ goto out; ++ } ++ dentry = NULL; ++ if (unlikely(au_test_anon(h_parent))) { ++ AuWarn1("%s decode_fh returned a disconnected dentry\n", ++ au_sbtype(h_sb)); ++ goto out_h_parent; ++ } ++ ++ dentry = ERR_PTR(-ENOMEM); ++ pathname = (void *)__get_free_page(GFP_NOFS); ++ if (unlikely(!pathname)) ++ goto out_h_parent; ++ ++ root = sb->s_root; ++ path.mnt = h_mnt; ++ di_read_lock_parent(root, !AuLock_IR); ++ path.dentry = au_h_dptr(root, bindex); ++ di_read_unlock(root, !AuLock_IR); ++ p = au_build_path(h_parent, &path, pathname, PAGE_SIZE, sb); ++ dentry = (void *)p; ++ if (IS_ERR(p)) ++ goto out_pathname; ++ ++ si_read_unlock(sb); ++ err = vfsub_kern_path(p, LOOKUP_FOLLOW | LOOKUP_DIRECTORY, &path); ++ dentry = ERR_PTR(err); ++ if (unlikely(err)) ++ goto out_relock; ++ ++ dentry = ERR_PTR(-ENOENT); ++ AuDebugOn(au_test_anon(path.dentry)); ++ if (unlikely(!path.dentry->d_inode)) ++ goto out_path; ++ ++ if (ino != path.dentry->d_inode->i_ino) ++ dentry = au_lkup_by_ino(&path, ino, /*nsi_lock*/NULL); ++ else ++ dentry = dget(path.dentry); ++ ++ out_path: ++ path_put(&path); ++ out_relock: ++ if (unlikely(si_nfsd_read_lock(sb, nsi_lock) < 0)) ++ if (!IS_ERR(dentry)) { ++ dput(dentry); ++ dentry = ERR_PTR(-ESTALE); ++ } ++ out_pathname: ++ free_page((unsigned long)pathname); ++ out_h_parent: ++ dput(h_parent); ++ out: ++ /* au_br_put(br); */ ++ AuTraceErrPtr(dentry); ++ return dentry; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static struct dentry * ++aufs_fh_to_dentry(struct super_block *sb, struct fid *fid, int fh_len, ++ int fh_type) ++{ ++ struct dentry *dentry; ++ __u32 *fh = fid->raw; ++ ino_t ino, dir_ino; ++ aufs_bindex_t bindex; ++ struct au_nfsd_si_lock nsi_lock = { ++ .sigen = fh[Fh_sigen], ++ .br_id = fh[Fh_br_id], ++ .force_lock = 0 ++ }; ++ ++ AuDebugOn(fh_len < Fh_tail); ++ ++ dentry = ERR_PTR(-ESTALE); ++ /* branch id may be wrapped around */ ++ bindex = si_nfsd_read_lock(sb, &nsi_lock); ++ if (unlikely(bindex < 0)) ++ goto out; ++ nsi_lock.force_lock = 1; ++ ++ /* is this inode still cached? */ ++ ino = decode_ino(fh + Fh_ino); ++ AuDebugOn(ino == AUFS_ROOT_INO); ++ dir_ino = decode_ino(fh + Fh_dir_ino); ++ dentry = decode_by_ino(sb, ino, dir_ino); ++ if (IS_ERR(dentry)) ++ goto out_unlock; ++ if (dentry) ++ goto accept; ++ ++ /* is the parent dir cached? */ ++ dentry = decode_by_dir_ino(sb, ino, dir_ino, &nsi_lock); ++ if (IS_ERR(dentry)) ++ goto out_unlock; ++ if (dentry) ++ goto accept; ++ ++ /* lookup path */ ++ dentry = decode_by_path(sb, bindex, ino, fh, fh_len, &nsi_lock); ++ if (IS_ERR(dentry)) ++ goto out_unlock; ++ if (unlikely(!dentry)) ++ /* todo?: make it ESTALE */ ++ goto out_unlock; ++ ++ accept: ++ if (dentry->d_inode->i_generation == fh[Fh_igen]) ++ goto out_unlock; /* success */ ++ ++ dput(dentry); ++ dentry = ERR_PTR(-ESTALE); ++ out_unlock: ++ si_read_unlock(sb); ++ out: ++ AuTraceErrPtr(dentry); ++ return dentry; ++} ++ ++#if 0 /* reserved for future use */ ++/* support subtreecheck option */ ++static struct dentry *aufs_fh_to_parent(struct super_block *sb, struct fid *fid, ++ int fh_len, int fh_type) ++{ ++ struct dentry *parent; ++ __u32 *fh = fid->raw; ++ ino_t dir_ino; ++ ++ dir_ino = decode_ino(fh + Fh_dir_ino); ++ parent = decode_by_ino(sb, dir_ino, 0); ++ if (IS_ERR(parent)) ++ goto out; ++ if (!parent) ++ parent = decode_by_path(sb, au_br_index(sb, fh[Fh_br_id]), ++ dir_ino, fh, fh_len); ++ ++ out: ++ AuTraceErrPtr(parent); ++ return parent; ++} ++#endif ++ ++/* ---------------------------------------------------------------------- */ ++ ++static int aufs_encode_fh(struct dentry *dentry, __u32 *fh, int *max_len, ++ int connectable) ++{ ++ int err; ++ aufs_bindex_t bindex, bend; ++ struct super_block *sb, *h_sb; ++ struct inode *inode; ++ struct dentry *parent, *h_parent; ++ struct au_branch *br; ++ ++ AuDebugOn(au_test_anon(dentry)); ++ ++ parent = NULL; ++ err = -ENOSPC; ++ if (unlikely(*max_len <= Fh_tail)) { ++ AuWarn1("NFSv2 client (max_len %d)?\n", *max_len); ++ goto out; ++ } ++ ++ err = FILEID_ROOT; ++ if (IS_ROOT(dentry)) { ++ AuDebugOn(dentry->d_inode->i_ino != AUFS_ROOT_INO); ++ goto out; ++ } ++ ++ err = -EIO; ++ h_parent = NULL; ++ sb = dentry->d_sb; ++ aufs_read_lock(dentry, AuLock_FLUSH | AuLock_IR); ++ parent = dget_parent(dentry); ++ di_read_lock_parent(parent, !AuLock_IR); ++ inode = dentry->d_inode; ++ AuDebugOn(!inode); ++#ifdef CONFIG_AUFS_DEBUG ++ if (unlikely(!au_opt_test(au_mntflags(sb), XINO))) ++ AuWarn1("NFS-exporting requires xino\n"); ++#endif ++ ++ bend = au_dbtaildir(parent); ++ for (bindex = au_dbstart(parent); bindex <= bend; bindex++) { ++ h_parent = au_h_dptr(parent, bindex); ++ if (h_parent) { ++ dget(h_parent); ++ break; ++ } ++ } ++ if (unlikely(!h_parent)) ++ goto out_unlock; ++ ++ err = -EPERM; ++ br = au_sbr(sb, bindex); ++ h_sb = br->br_mnt->mnt_sb; ++ if (unlikely(!h_sb->s_export_op)) { ++ AuErr1("%s branch is not exportable\n", au_sbtype(h_sb)); ++ goto out_dput; ++ } ++ ++ fh[Fh_br_id] = br->br_id; ++ fh[Fh_sigen] = au_sigen(sb); ++ encode_ino(fh + Fh_ino, inode->i_ino); ++ encode_ino(fh + Fh_dir_ino, parent->d_inode->i_ino); ++ fh[Fh_igen] = inode->i_generation; ++ ++ *max_len -= Fh_tail; ++ fh[Fh_h_type] = exportfs_encode_fh(h_parent, (void *)(fh + Fh_tail), ++ max_len, ++ /*connectable or subtreecheck*/0); ++ err = fh[Fh_h_type]; ++ *max_len += Fh_tail; ++ /* todo: macros? */ ++ if (err != 255) ++ err = 99; ++ else ++ AuWarn1("%s encode_fh failed\n", au_sbtype(h_sb)); ++ ++ out_dput: ++ dput(h_parent); ++ out_unlock: ++ di_read_unlock(parent, !AuLock_IR); ++ dput(parent); ++ aufs_read_unlock(dentry, AuLock_IR); ++ out: ++ if (unlikely(err < 0)) ++ err = 255; ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static struct export_operations aufs_export_op = { ++ .fh_to_dentry = aufs_fh_to_dentry, ++ /* .fh_to_parent = aufs_fh_to_parent, */ ++ .encode_fh = aufs_encode_fh ++}; ++ ++void au_export_init(struct super_block *sb) ++{ ++ struct au_sbinfo *sbinfo; ++ __u32 u; ++ ++ sb->s_export_op = &aufs_export_op; ++ sbinfo = au_sbi(sb); ++ sbinfo->si_xigen = NULL; ++ get_random_bytes(&u, sizeof(u)); ++ BUILD_BUG_ON(sizeof(u) != sizeof(int)); ++ atomic_set(&sbinfo->si_xigen_next, u); ++} +diff -Nur linux-2.6.31.4.orig/fs/aufs/file.c linux-2.6.31.4/fs/aufs/file.c +--- linux-2.6.31.4.orig/fs/aufs/file.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/file.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,578 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * handling file/dir, and address_space operation ++ */ ++ ++#include <linux/file.h> ++#include <linux/fsnotify.h> ++#include <linux/namei.h> ++#include <linux/pagemap.h> ++#include "aufs.h" ++ ++/* ++ * a dirty trick for handling deny_write_access(). ++ * because FMODE_EXEC flag is not passed to f_op->open(), ++ * set it to file->private_data temporary. ++ */ ++void au_store_oflag(struct nameidata *nd, struct inode *inode) ++{ ++ if (nd ++ /* && !(nd->flags & LOOKUP_CONTINUE) */ ++ && (nd->flags & LOOKUP_OPEN) ++ && (nd->intent.open.flags & vfsub_fmode_to_uint(FMODE_EXEC)) ++ && inode ++ && S_ISREG(inode->i_mode)) { ++ /* suppress a warning in lp64 */ ++ unsigned long flags = nd->intent.open.flags; ++ nd->intent.open.file->private_data = (void *)flags; ++ /* smp_mb(); */ ++ } ++} ++ ++/* drop flags for writing */ ++unsigned int au_file_roflags(unsigned int flags) ++{ ++ flags &= ~(O_WRONLY | O_RDWR | O_APPEND | O_CREAT | O_TRUNC); ++ flags |= O_RDONLY | O_NOATIME; ++ return flags; ++} ++ ++/* common functions to regular file and dir */ ++struct file *au_h_open(struct dentry *dentry, aufs_bindex_t bindex, int flags, ++ struct file *file) ++{ ++ struct file *h_file; ++ struct dentry *h_dentry; ++ struct inode *h_inode; ++ struct super_block *sb; ++ struct au_branch *br; ++ int err; ++ ++ /* a race condition can happen between open and unlink/rmdir */ ++ h_file = ERR_PTR(-ENOENT); ++ h_dentry = au_h_dptr(dentry, bindex); ++ if (au_test_nfsd(current) && !h_dentry) ++ goto out; ++ h_inode = h_dentry->d_inode; ++ if (au_test_nfsd(current) && !h_inode) ++ goto out; ++ if (unlikely((!d_unhashed(dentry) && d_unhashed(h_dentry)) ++ || !h_inode)) ++ goto out; ++ ++ sb = dentry->d_sb; ++ br = au_sbr(sb, bindex); ++ h_file = ERR_PTR(-EACCES); ++ if (file && (file->f_mode & FMODE_EXEC) ++ && (br->br_mnt->mnt_flags & MNT_NOEXEC)) ++ goto out; ++ ++ /* drop flags for writing */ ++ if (au_test_ro(sb, bindex, dentry->d_inode)) ++ flags = au_file_roflags(flags); ++ flags &= ~O_CREAT; ++ atomic_inc(&br->br_count); ++ h_file = dentry_open(dget(h_dentry), mntget(br->br_mnt), flags, ++ current_cred()); ++ if (IS_ERR(h_file)) ++ goto out_br; ++ ++ if (file && (file->f_mode & FMODE_EXEC)) { ++ h_file->f_mode |= FMODE_EXEC; ++ err = deny_write_access(h_file); ++ if (unlikely(err)) { ++ fput(h_file); ++ h_file = ERR_PTR(err); ++ goto out_br; ++ } ++ } ++ fsnotify_open(h_dentry); ++ goto out; /* success */ ++ ++ out_br: ++ atomic_dec(&br->br_count); ++ out: ++ return h_file; ++} ++ ++int au_do_open(struct file *file, int (*open)(struct file *file, int flags)) ++{ ++ int err; ++ struct dentry *dentry; ++ struct super_block *sb; ++ ++ dentry = file->f_dentry; ++ sb = dentry->d_sb; ++ si_read_lock(sb, AuLock_FLUSH); ++ err = au_finfo_init(file); ++ if (unlikely(err)) ++ goto out; ++ ++ di_read_lock_child(dentry, AuLock_IR); ++ err = open(file, file->f_flags); ++ di_read_unlock(dentry, AuLock_IR); ++ ++ fi_write_unlock(file); ++ if (unlikely(err)) ++ au_finfo_fin(file); ++ out: ++ si_read_unlock(sb); ++ return err; ++} ++ ++int au_reopen_nondir(struct file *file) ++{ ++ int err; ++ aufs_bindex_t bstart, bindex, bend; ++ struct dentry *dentry; ++ struct file *h_file, *h_file_tmp; ++ ++ dentry = file->f_dentry; ++ bstart = au_dbstart(dentry); ++ h_file_tmp = NULL; ++ if (au_fbstart(file) == bstart) { ++ h_file = au_h_fptr(file, bstart); ++ if (file->f_mode == h_file->f_mode) ++ return 0; /* success */ ++ h_file_tmp = h_file; ++ get_file(h_file_tmp); ++ au_set_h_fptr(file, bstart, NULL); ++ } ++ AuDebugOn(au_fbstart(file) < bstart ++ || au_fi(file)->fi_hfile[0 + bstart].hf_file); ++ ++ h_file = au_h_open(dentry, bstart, file->f_flags & ~O_TRUNC, file); ++ err = PTR_ERR(h_file); ++ if (IS_ERR(h_file)) ++ goto out; /* todo: close all? */ ++ ++ err = 0; ++ au_set_fbstart(file, bstart); ++ au_set_h_fptr(file, bstart, h_file); ++ au_update_figen(file); ++ /* todo: necessary? */ ++ /* file->f_ra = h_file->f_ra; */ ++ ++ /* close lower files */ ++ bend = au_fbend(file); ++ for (bindex = bstart + 1; bindex <= bend; bindex++) ++ au_set_h_fptr(file, bindex, NULL); ++ au_set_fbend(file, bstart); ++ ++ out: ++ if (h_file_tmp) ++ fput(h_file_tmp); ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static int au_reopen_wh(struct file *file, aufs_bindex_t btgt, ++ struct dentry *hi_wh) ++{ ++ int err; ++ aufs_bindex_t bstart; ++ struct au_dinfo *dinfo; ++ struct dentry *h_dentry; ++ ++ dinfo = au_di(file->f_dentry); ++ AuRwMustWriteLock(&dinfo->di_rwsem); ++ ++ bstart = dinfo->di_bstart; ++ dinfo->di_bstart = btgt; ++ h_dentry = dinfo->di_hdentry[0 + btgt].hd_dentry; ++ dinfo->di_hdentry[0 + btgt].hd_dentry = hi_wh; ++ err = au_reopen_nondir(file); ++ dinfo->di_hdentry[0 + btgt].hd_dentry = h_dentry; ++ dinfo->di_bstart = bstart; ++ ++ return err; ++} ++ ++static int au_ready_to_write_wh(struct file *file, loff_t len, ++ aufs_bindex_t bcpup) ++{ ++ int err; ++ struct inode *inode; ++ struct dentry *dentry, *hi_wh; ++ struct super_block *sb; ++ ++ dentry = file->f_dentry; ++ inode = dentry->d_inode; ++ hi_wh = au_hi_wh(inode, bcpup); ++ if (!hi_wh) ++ err = au_sio_cpup_wh(dentry, bcpup, len, file); ++ else ++ /* already copied-up after unlink */ ++ err = au_reopen_wh(file, bcpup, hi_wh); ++ ++ sb = dentry->d_sb; ++ if (!err && inode->i_nlink > 1 && au_opt_test(au_mntflags(sb), PLINK)) ++ au_plink_append(inode, bcpup, au_h_dptr(dentry, bcpup)); ++ ++ return err; ++} ++ ++/* ++ * prepare the @file for writing. ++ */ ++int au_ready_to_write(struct file *file, loff_t len, struct au_pin *pin) ++{ ++ int err; ++ aufs_bindex_t bstart, bcpup; ++ struct dentry *dentry, *parent, *h_dentry; ++ struct inode *h_inode, *inode; ++ struct super_block *sb; ++ ++ dentry = file->f_dentry; ++ sb = dentry->d_sb; ++ bstart = au_fbstart(file); ++ inode = dentry->d_inode; ++ err = au_test_ro(sb, bstart, inode); ++ if (!err && (au_h_fptr(file, bstart)->f_mode & FMODE_WRITE)) { ++ err = au_pin(pin, dentry, bstart, AuOpt_UDBA_NONE, /*flags*/0); ++ goto out; ++ } ++ ++ /* need to cpup */ ++ parent = dget_parent(dentry); ++ di_write_lock_parent(parent); ++ err = AuWbrCopyup(au_sbi(sb), dentry); ++ bcpup = err; ++ if (unlikely(err < 0)) ++ goto out_dgrade; ++ err = 0; ++ ++ if (!au_h_dptr(parent, bcpup)) { ++ err = au_cpup_dirs(dentry, bcpup); ++ if (unlikely(err)) ++ goto out_dgrade; ++ } ++ ++ err = au_pin(pin, dentry, bcpup, AuOpt_UDBA_NONE, ++ AuPin_DI_LOCKED | AuPin_MNT_WRITE); ++ if (unlikely(err)) ++ goto out_dgrade; ++ ++ h_dentry = au_h_fptr(file, bstart)->f_dentry; ++ h_inode = h_dentry->d_inode; ++ mutex_lock_nested(&h_inode->i_mutex, AuLsc_I_CHILD); ++ if (d_unhashed(dentry) /* || d_unhashed(h_dentry) */ ++ /* || !h_inode->i_nlink */) { ++ err = au_ready_to_write_wh(file, len, bcpup); ++ di_downgrade_lock(parent, AuLock_IR); ++ } else { ++ di_downgrade_lock(parent, AuLock_IR); ++ if (!au_h_dptr(dentry, bcpup)) ++ err = au_sio_cpup_simple(dentry, bcpup, len, ++ AuCpup_DTIME); ++ if (!err) ++ err = au_reopen_nondir(file); ++ } ++ mutex_unlock(&h_inode->i_mutex); ++ ++ if (!err) { ++ au_pin_set_parent_lflag(pin, /*lflag*/0); ++ goto out_dput; /* success */ ++ } ++ au_unpin(pin); ++ goto out_unlock; ++ ++ out_dgrade: ++ di_downgrade_lock(parent, AuLock_IR); ++ out_unlock: ++ di_read_unlock(parent, AuLock_IR); ++ out_dput: ++ dput(parent); ++ out: ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static int au_file_refresh_by_inode(struct file *file, int *need_reopen) ++{ ++ int err; ++ aufs_bindex_t bstart; ++ struct au_pin pin; ++ struct au_finfo *finfo; ++ struct dentry *dentry, *parent, *hi_wh; ++ struct inode *inode; ++ struct super_block *sb; ++ ++ FiMustWriteLock(file); ++ ++ err = 0; ++ finfo = au_fi(file); ++ dentry = file->f_dentry; ++ sb = dentry->d_sb; ++ inode = dentry->d_inode; ++ bstart = au_ibstart(inode); ++ if (bstart == finfo->fi_bstart) ++ goto out; ++ ++ parent = dget_parent(dentry); ++ if (au_test_ro(sb, bstart, inode)) { ++ di_read_lock_parent(parent, !AuLock_IR); ++ err = AuWbrCopyup(au_sbi(sb), dentry); ++ bstart = err; ++ di_read_unlock(parent, !AuLock_IR); ++ if (unlikely(err < 0)) ++ goto out_parent; ++ err = 0; ++ } ++ ++ di_read_lock_parent(parent, AuLock_IR); ++ hi_wh = au_hi_wh(inode, bstart); ++ if (au_opt_test(au_mntflags(sb), PLINK) ++ && au_plink_test(inode) ++ && !d_unhashed(dentry)) { ++ err = au_test_and_cpup_dirs(dentry, bstart); ++ if (unlikely(err)) ++ goto out_unlock; ++ ++ /* always superio. */ ++ err = au_pin(&pin, dentry, bstart, AuOpt_UDBA_NONE, ++ AuPin_DI_LOCKED | AuPin_MNT_WRITE); ++ if (!err) ++ err = au_sio_cpup_simple(dentry, bstart, -1, ++ AuCpup_DTIME); ++ au_unpin(&pin); ++ } else if (hi_wh) { ++ /* already copied-up after unlink */ ++ err = au_reopen_wh(file, bstart, hi_wh); ++ *need_reopen = 0; ++ } ++ ++ out_unlock: ++ di_read_unlock(parent, AuLock_IR); ++ out_parent: ++ dput(parent); ++ out: ++ return err; ++} ++ ++static void au_do_refresh_file(struct file *file) ++{ ++ aufs_bindex_t bindex, bend, new_bindex, brid; ++ struct au_hfile *p, tmp, *q; ++ struct au_finfo *finfo; ++ struct super_block *sb; ++ ++ FiMustWriteLock(file); ++ ++ sb = file->f_dentry->d_sb; ++ finfo = au_fi(file); ++ p = finfo->fi_hfile + finfo->fi_bstart; ++ brid = p->hf_br->br_id; ++ bend = finfo->fi_bend; ++ for (bindex = finfo->fi_bstart; bindex <= bend; bindex++, p++) { ++ if (!p->hf_file) ++ continue; ++ ++ new_bindex = au_br_index(sb, p->hf_br->br_id); ++ if (new_bindex == bindex) ++ continue; ++ if (new_bindex < 0) { ++ au_set_h_fptr(file, bindex, NULL); ++ continue; ++ } ++ ++ /* swap two lower inode, and loop again */ ++ q = finfo->fi_hfile + new_bindex; ++ tmp = *q; ++ *q = *p; ++ *p = tmp; ++ if (tmp.hf_file) { ++ bindex--; ++ p--; ++ } ++ } ++ ++ p = finfo->fi_hfile; ++ if (!au_test_mmapped(file) && !d_unhashed(file->f_dentry)) { ++ bend = au_sbend(sb); ++ for (finfo->fi_bstart = 0; finfo->fi_bstart <= bend; ++ finfo->fi_bstart++, p++) ++ if (p->hf_file) { ++ if (p->hf_file->f_dentry ++ && p->hf_file->f_dentry->d_inode) ++ break; ++ else ++ au_hfput(p, file); ++ } ++ } else { ++ bend = au_br_index(sb, brid); ++ for (finfo->fi_bstart = 0; finfo->fi_bstart < bend; ++ finfo->fi_bstart++, p++) ++ if (p->hf_file) ++ au_hfput(p, file); ++ bend = au_sbend(sb); ++ } ++ ++ p = finfo->fi_hfile + bend; ++ for (finfo->fi_bend = bend; finfo->fi_bend >= finfo->fi_bstart; ++ finfo->fi_bend--, p--) ++ if (p->hf_file) { ++ if (p->hf_file->f_dentry ++ && p->hf_file->f_dentry->d_inode) ++ break; ++ else ++ au_hfput(p, file); ++ } ++ AuDebugOn(finfo->fi_bend < finfo->fi_bstart); ++} ++ ++/* ++ * after branch manipulating, refresh the file. ++ */ ++static int refresh_file(struct file *file, int (*reopen)(struct file *file)) ++{ ++ int err, need_reopen; ++ struct dentry *dentry; ++ aufs_bindex_t bend, bindex; ++ ++ dentry = file->f_dentry; ++ err = au_fi_realloc(au_fi(file), au_sbend(dentry->d_sb) + 1); ++ if (unlikely(err)) ++ goto out; ++ au_do_refresh_file(file); ++ ++ err = 0; ++ need_reopen = 1; ++ if (!au_test_mmapped(file)) ++ err = au_file_refresh_by_inode(file, &need_reopen); ++ if (!err && need_reopen && !d_unhashed(dentry)) ++ err = reopen(file); ++ if (!err) { ++ au_update_figen(file); ++ return 0; /* success */ ++ } ++ ++ /* error, close all lower files */ ++ bend = au_fbend(file); ++ for (bindex = au_fbstart(file); bindex <= bend; bindex++) ++ au_set_h_fptr(file, bindex, NULL); ++ ++ out: ++ return err; ++} ++ ++/* common function to regular file and dir */ ++int au_reval_and_lock_fdi(struct file *file, int (*reopen)(struct file *file), ++ int wlock) ++{ ++ int err; ++ unsigned int sigen, figen; ++ aufs_bindex_t bstart; ++ unsigned char pseudo_link; ++ struct dentry *dentry; ++ ++ err = 0; ++ dentry = file->f_dentry; ++ sigen = au_sigen(dentry->d_sb); ++ fi_write_lock(file); ++ figen = au_figen(file); ++ di_write_lock_child(dentry); ++ bstart = au_dbstart(dentry); ++ pseudo_link = (bstart != au_ibstart(dentry->d_inode)); ++ if (sigen == figen && !pseudo_link && au_fbstart(file) == bstart) { ++ if (!wlock) { ++ di_downgrade_lock(dentry, AuLock_IR); ++ fi_downgrade_lock(file); ++ } ++ goto out; /* success */ ++ } ++ ++ AuDbg("sigen %d, figen %d\n", sigen, figen); ++ if (sigen != au_digen(dentry) ++ || sigen != au_iigen(dentry->d_inode)) { ++ err = au_reval_dpath(dentry, sigen); ++ if (unlikely(err < 0)) ++ goto out; ++ AuDebugOn(au_digen(dentry) != sigen ++ || au_iigen(dentry->d_inode) != sigen); ++ } ++ ++ err = refresh_file(file, reopen); ++ if (!err) { ++ if (!wlock) { ++ di_downgrade_lock(dentry, AuLock_IR); ++ fi_downgrade_lock(file); ++ } ++ } else { ++ di_write_unlock(dentry); ++ fi_write_unlock(file); ++ } ++ ++ out: ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* cf. aufs_nopage() */ ++/* for madvise(2) */ ++static int aufs_readpage(struct file *file __maybe_unused, struct page *page) ++{ ++ unlock_page(page); ++ return 0; ++} ++ ++/* they will never be called. */ ++#ifdef CONFIG_AUFS_DEBUG ++static int aufs_write_begin(struct file *file, struct address_space *mapping, ++ loff_t pos, unsigned len, unsigned flags, ++ struct page **pagep, void **fsdata) ++{ AuUnsupport(); return 0; } ++static int aufs_write_end(struct file *file, struct address_space *mapping, ++ loff_t pos, unsigned len, unsigned copied, ++ struct page *page, void *fsdata) ++{ AuUnsupport(); return 0; } ++static int aufs_writepage(struct page *page, struct writeback_control *wbc) ++{ AuUnsupport(); return 0; } ++static void aufs_sync_page(struct page *page) ++{ AuUnsupport(); } ++ ++static int aufs_set_page_dirty(struct page *page) ++{ AuUnsupport(); return 0; } ++static void aufs_invalidatepage(struct page *page, unsigned long offset) ++{ AuUnsupport(); } ++static int aufs_releasepage(struct page *page, gfp_t gfp) ++{ AuUnsupport(); return 0; } ++static ssize_t aufs_direct_IO(int rw, struct kiocb *iocb, ++ const struct iovec *iov, loff_t offset, ++ unsigned long nr_segs) ++{ AuUnsupport(); return 0; } ++#endif /* CONFIG_AUFS_DEBUG */ ++ ++struct address_space_operations aufs_aop = { ++ .readpage = aufs_readpage, ++#ifdef CONFIG_AUFS_DEBUG ++ .writepage = aufs_writepage, ++ .sync_page = aufs_sync_page, ++ .set_page_dirty = aufs_set_page_dirty, ++ .write_begin = aufs_write_begin, ++ .write_end = aufs_write_end, ++ .invalidatepage = aufs_invalidatepage, ++ .releasepage = aufs_releasepage, ++ .direct_IO = aufs_direct_IO, ++#endif /* CONFIG_AUFS_DEBUG */ ++}; +diff -Nur linux-2.6.31.4.orig/fs/aufs/file.h linux-2.6.31.4/fs/aufs/file.h +--- linux-2.6.31.4.orig/fs/aufs/file.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/file.h 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,175 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * file operations ++ */ ++ ++#ifndef __AUFS_FILE_H__ ++#define __AUFS_FILE_H__ ++ ++#ifdef __KERNEL__ ++ ++#include <linux/fs.h> ++#include <linux/poll.h> ++#include <linux/aufs_type.h> ++#include "rwsem.h" ++ ++struct au_branch; ++struct au_hfile { ++ struct file *hf_file; ++ struct au_branch *hf_br; ++}; ++ ++struct au_vdir; ++struct au_finfo { ++ atomic_t fi_generation; ++ ++ struct au_rwsem fi_rwsem; ++ struct au_hfile *fi_hfile; ++ aufs_bindex_t fi_bstart, fi_bend; ++ ++ union { ++ /* non-dir only */ ++ struct { ++ struct vm_operations_struct *fi_h_vm_ops; ++ struct vm_operations_struct *fi_vm_ops; ++ }; ++ ++ /* dir only */ ++ struct { ++ struct au_vdir *fi_vdir_cache; ++ int fi_maintain_plink; ++ }; ++ }; ++}; ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* file.c */ ++extern struct address_space_operations aufs_aop; ++void au_store_oflag(struct nameidata *nd, struct inode *inode); ++unsigned int au_file_roflags(unsigned int flags); ++struct file *au_h_open(struct dentry *dentry, aufs_bindex_t bindex, int flags, ++ struct file *file); ++int au_do_open(struct file *file, int (*open)(struct file *file, int flags)); ++int au_reopen_nondir(struct file *file); ++struct au_pin; ++int au_ready_to_write(struct file *file, loff_t len, struct au_pin *pin); ++int au_reval_and_lock_fdi(struct file *file, int (*reopen)(struct file *file), ++ int wlock); ++ ++/* poll.c */ ++#ifdef CONFIG_AUFS_POLL ++unsigned int aufs_poll(struct file *file, poll_table *wait); ++#endif ++ ++/* f_op.c */ ++extern const struct file_operations aufs_file_fop; ++int aufs_flush(struct file *file, fl_owner_t id); ++ ++/* finfo.c */ ++void au_hfput(struct au_hfile *hf, struct file *file); ++void au_set_h_fptr(struct file *file, aufs_bindex_t bindex, ++ struct file *h_file); ++ ++void au_update_figen(struct file *file); ++ ++void au_finfo_fin(struct file *file); ++int au_finfo_init(struct file *file); ++int au_fi_realloc(struct au_finfo *finfo, int nbr); ++ ++/* ---------------------------------------------------------------------- */ ++ ++static inline struct au_finfo *au_fi(struct file *file) ++{ ++ return file->private_data; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* ++ * fi_read_lock, fi_write_lock, ++ * fi_read_unlock, fi_write_unlock, fi_downgrade_lock ++ */ ++AuSimpleRwsemFuncs(fi, struct file *f, &au_fi(f)->fi_rwsem); ++ ++#define FiMustNoWaiters(f) AuRwMustNoWaiters(&au_fi(f)->fi_rwsem) ++#define FiMustAnyLock(f) AuRwMustAnyLock(&au_fi(f)->fi_rwsem) ++#define FiMustWriteLock(f) AuRwMustWriteLock(&au_fi(f)->fi_rwsem) ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* todo: hard/soft set? */ ++static inline aufs_bindex_t au_fbstart(struct file *file) ++{ ++ FiMustAnyLock(file); ++ return au_fi(file)->fi_bstart; ++} ++ ++static inline aufs_bindex_t au_fbend(struct file *file) ++{ ++ FiMustAnyLock(file); ++ return au_fi(file)->fi_bend; ++} ++ ++static inline struct au_vdir *au_fvdir_cache(struct file *file) ++{ ++ FiMustAnyLock(file); ++ return au_fi(file)->fi_vdir_cache; ++} ++ ++static inline void au_set_fbstart(struct file *file, aufs_bindex_t bindex) ++{ ++ FiMustWriteLock(file); ++ au_fi(file)->fi_bstart = bindex; ++} ++ ++static inline void au_set_fbend(struct file *file, aufs_bindex_t bindex) ++{ ++ FiMustWriteLock(file); ++ au_fi(file)->fi_bend = bindex; ++} ++ ++static inline void au_set_fvdir_cache(struct file *file, ++ struct au_vdir *vdir_cache) ++{ ++ FiMustWriteLock(file); ++ au_fi(file)->fi_vdir_cache = vdir_cache; ++} ++ ++static inline struct file *au_h_fptr(struct file *file, aufs_bindex_t bindex) ++{ ++ FiMustAnyLock(file); ++ return au_fi(file)->fi_hfile[0 + bindex].hf_file; ++} ++ ++/* todo: memory barrier? */ ++static inline unsigned int au_figen(struct file *f) ++{ ++ return atomic_read(&au_fi(f)->fi_generation); ++} ++ ++static inline int au_test_mmapped(struct file *f) ++{ ++ FiMustAnyLock(f); ++ return !!(au_fi(f)->fi_h_vm_ops); ++} ++ ++#endif /* __KERNEL__ */ ++#endif /* __AUFS_FILE_H__ */ +diff -Nur linux-2.6.31.4.orig/fs/aufs/finfo.c linux-2.6.31.4/fs/aufs/finfo.c +--- linux-2.6.31.4.orig/fs/aufs/finfo.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/finfo.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,133 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * file private data ++ */ ++ ++#include <linux/file.h> ++#include "aufs.h" ++ ++void au_hfput(struct au_hfile *hf, struct file *file) ++{ ++ if (file->f_mode & FMODE_EXEC) ++ allow_write_access(hf->hf_file); ++ fput(hf->hf_file); ++ hf->hf_file = NULL; ++ atomic_dec_return(&hf->hf_br->br_count); ++ hf->hf_br = NULL; ++} ++ ++void au_set_h_fptr(struct file *file, aufs_bindex_t bindex, struct file *val) ++{ ++ struct au_finfo *finfo = au_fi(file); ++ struct au_hfile *hf; ++ ++ hf = finfo->fi_hfile + bindex; ++ if (hf->hf_file) ++ au_hfput(hf, file); ++ if (val) { ++ hf->hf_file = val; ++ hf->hf_br = au_sbr(file->f_dentry->d_sb, bindex); ++ } ++} ++ ++void au_update_figen(struct file *file) ++{ ++ atomic_set(&au_fi(file)->fi_generation, au_digen(file->f_dentry)); ++ /* smp_mb(); */ /* atomic_set */ ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++void au_finfo_fin(struct file *file) ++{ ++ struct au_finfo *finfo; ++ aufs_bindex_t bindex, bend; ++ ++ fi_write_lock(file); ++ bend = au_fbend(file); ++ bindex = au_fbstart(file); ++ if (bindex >= 0) ++ /* ++ * calls fput() instead of filp_close(), ++ * since no dnotify or lock for the lower file. ++ */ ++ for (; bindex <= bend; bindex++) ++ au_set_h_fptr(file, bindex, NULL); ++ ++ finfo = au_fi(file); ++ au_dbg_verify_hf(finfo); ++ kfree(finfo->fi_hfile); ++ fi_write_unlock(file); ++ AuRwDestroy(&finfo->fi_rwsem); ++ au_cache_free_finfo(finfo); ++} ++ ++int au_finfo_init(struct file *file) ++{ ++ struct au_finfo *finfo; ++ struct dentry *dentry; ++ unsigned long ul; ++ ++ dentry = file->f_dentry; ++ finfo = au_cache_alloc_finfo(); ++ if (unlikely(!finfo)) ++ goto out; ++ ++ finfo->fi_hfile = kcalloc(au_sbend(dentry->d_sb) + 1, ++ sizeof(*finfo->fi_hfile), GFP_NOFS); ++ if (unlikely(!finfo->fi_hfile)) ++ goto out_finfo; ++ ++ au_rw_init_wlock(&finfo->fi_rwsem); ++ finfo->fi_bstart = -1; ++ finfo->fi_bend = -1; ++ atomic_set(&finfo->fi_generation, au_digen(dentry)); ++ /* smp_mb(); */ /* atomic_set */ ++ ++ /* cf. au_store_oflag() */ ++ /* suppress a warning in lp64 */ ++ ul = (unsigned long)file->private_data; ++ file->f_mode |= (vfsub_uint_to_fmode(ul) & FMODE_EXEC); ++ file->private_data = finfo; ++ return 0; /* success */ ++ ++ out_finfo: ++ au_cache_free_finfo(finfo); ++ out: ++ return -ENOMEM; ++} ++ ++int au_fi_realloc(struct au_finfo *finfo, int nbr) ++{ ++ int err, sz; ++ struct au_hfile *hfp; ++ ++ err = -ENOMEM; ++ sz = sizeof(*hfp) * (finfo->fi_bend + 1); ++ if (!sz) ++ sz = sizeof(*hfp); ++ hfp = au_kzrealloc(finfo->fi_hfile, sz, sizeof(*hfp) * nbr, GFP_NOFS); ++ if (hfp) { ++ finfo->fi_hfile = hfp; ++ err = 0; ++ } ++ ++ return err; ++} +diff -Nur linux-2.6.31.4.orig/fs/aufs/f_op.c linux-2.6.31.4/fs/aufs/f_op.c +--- linux-2.6.31.4.orig/fs/aufs/f_op.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/f_op.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,802 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * file and vm operations ++ */ ++ ++#include <linux/file.h> ++#include <linux/fs_stack.h> ++#include <linux/mm.h> ++#include <linux/security.h> ++#include "aufs.h" ++ ++/* common function to regular file and dir */ ++int aufs_flush(struct file *file, fl_owner_t id) ++{ ++ int err; ++ aufs_bindex_t bindex, bend; ++ struct dentry *dentry; ++ struct file *h_file; ++ ++ dentry = file->f_dentry; ++ si_noflush_read_lock(dentry->d_sb); ++ fi_read_lock(file); ++ di_read_lock_child(dentry, AuLock_IW); ++ ++ err = 0; ++ bend = au_fbend(file); ++ for (bindex = au_fbstart(file); !err && bindex <= bend; bindex++) { ++ h_file = au_h_fptr(file, bindex); ++ if (!h_file || !h_file->f_op || !h_file->f_op->flush) ++ continue; ++ ++ err = h_file->f_op->flush(h_file, id); ++ if (!err) ++ vfsub_update_h_iattr(&h_file->f_path, /*did*/NULL); ++ /*ignore*/ ++ } ++ au_cpup_attr_timesizes(dentry->d_inode); ++ ++ di_read_unlock(dentry, AuLock_IW); ++ fi_read_unlock(file); ++ si_read_unlock(dentry->d_sb); ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static int do_open_nondir(struct file *file, int flags) ++{ ++ int err; ++ aufs_bindex_t bindex; ++ struct file *h_file; ++ struct dentry *dentry; ++ struct au_finfo *finfo; ++ ++ FiMustWriteLock(file); ++ ++ err = 0; ++ dentry = file->f_dentry; ++ finfo = au_fi(file); ++ finfo->fi_h_vm_ops = NULL; ++ finfo->fi_vm_ops = NULL; ++ bindex = au_dbstart(dentry); ++ /* O_TRUNC is processed already */ ++ BUG_ON(au_test_ro(dentry->d_sb, bindex, dentry->d_inode) ++ && (flags & O_TRUNC)); ++ ++ h_file = au_h_open(dentry, bindex, flags, file); ++ if (IS_ERR(h_file)) ++ err = PTR_ERR(h_file); ++ else { ++ au_set_fbstart(file, bindex); ++ au_set_fbend(file, bindex); ++ au_set_h_fptr(file, bindex, h_file); ++ au_update_figen(file); ++ /* todo: necessary? */ ++ /* file->f_ra = h_file->f_ra; */ ++ } ++ return err; ++} ++ ++static int aufs_open_nondir(struct inode *inode __maybe_unused, ++ struct file *file) ++{ ++ return au_do_open(file, do_open_nondir); ++} ++ ++static int aufs_release_nondir(struct inode *inode __maybe_unused, ++ struct file *file) ++{ ++ struct super_block *sb = file->f_dentry->d_sb; ++ ++ si_noflush_read_lock(sb); ++ kfree(au_fi(file)->fi_vm_ops); ++ au_finfo_fin(file); ++ si_read_unlock(sb); ++ return 0; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static ssize_t aufs_read(struct file *file, char __user *buf, size_t count, ++ loff_t *ppos) ++{ ++ ssize_t err; ++ struct dentry *dentry; ++ struct file *h_file; ++ struct super_block *sb; ++ ++ dentry = file->f_dentry; ++ sb = dentry->d_sb; ++ si_read_lock(sb, AuLock_FLUSH); ++ err = au_reval_and_lock_fdi(file, au_reopen_nondir, /*wlock*/0); ++ if (unlikely(err)) ++ goto out; ++ ++ h_file = au_h_fptr(file, au_fbstart(file)); ++ err = vfsub_read_u(h_file, buf, count, ppos); ++ /* todo: necessary? */ ++ /* file->f_ra = h_file->f_ra; */ ++ fsstack_copy_attr_atime(dentry->d_inode, h_file->f_dentry->d_inode); ++ ++ di_read_unlock(dentry, AuLock_IR); ++ fi_read_unlock(file); ++ out: ++ si_read_unlock(sb); ++ return err; ++} ++ ++static ssize_t aufs_write(struct file *file, const char __user *ubuf, ++ size_t count, loff_t *ppos) ++{ ++ ssize_t err; ++ aufs_bindex_t bstart; ++ struct au_pin pin; ++ struct dentry *dentry; ++ struct inode *inode; ++ struct super_block *sb; ++ struct file *h_file; ++ char __user *buf = (char __user *)ubuf; ++ ++ dentry = file->f_dentry; ++ sb = dentry->d_sb; ++ inode = dentry->d_inode; ++ mutex_lock(&inode->i_mutex); ++ si_read_lock(sb, AuLock_FLUSH); ++ ++ err = au_reval_and_lock_fdi(file, au_reopen_nondir, /*wlock*/1); ++ if (unlikely(err)) ++ goto out; ++ ++ err = au_ready_to_write(file, -1, &pin); ++ di_downgrade_lock(dentry, AuLock_IR); ++ if (unlikely(err)) ++ goto out_unlock; ++ ++ bstart = au_fbstart(file); ++ h_file = au_h_fptr(file, bstart); ++ au_unpin(&pin); ++ err = vfsub_write_u(h_file, buf, count, ppos); ++ au_cpup_attr_timesizes(inode); ++ inode->i_mode = h_file->f_dentry->d_inode->i_mode; ++ ++ out_unlock: ++ di_read_unlock(dentry, AuLock_IR); ++ fi_write_unlock(file); ++ out: ++ si_read_unlock(sb); ++ mutex_unlock(&inode->i_mutex); ++ return err; ++} ++ ++static ssize_t aufs_aio_read(struct kiocb *kio, const struct iovec *iov, ++ unsigned long nv, loff_t pos) ++{ ++ ssize_t err; ++ struct file *file, *h_file; ++ struct dentry *dentry; ++ struct super_block *sb; ++ ++ file = kio->ki_filp; ++ dentry = file->f_dentry; ++ sb = dentry->d_sb; ++ si_read_lock(sb, AuLock_FLUSH); ++ err = au_reval_and_lock_fdi(file, au_reopen_nondir, /*wlock*/0); ++ if (unlikely(err)) ++ goto out; ++ ++ err = -ENOSYS; ++ h_file = au_h_fptr(file, au_fbstart(file)); ++ if (h_file->f_op && h_file->f_op->aio_read) { ++ err = security_file_permission(h_file, MAY_READ); ++ if (unlikely(err)) ++ goto out_unlock; ++ if (!is_sync_kiocb(kio)) { ++ get_file(h_file); ++ fput(file); ++ } ++ kio->ki_filp = h_file; ++ err = h_file->f_op->aio_read(kio, iov, nv, pos); ++ /* todo: necessary? */ ++ /* file->f_ra = h_file->f_ra; */ ++ fsstack_copy_attr_atime(dentry->d_inode, ++ h_file->f_dentry->d_inode); ++ } else ++ /* currently there is no such fs */ ++ WARN_ON_ONCE(h_file->f_op && h_file->f_op->read); ++ ++ out_unlock: ++ di_read_unlock(dentry, AuLock_IR); ++ fi_read_unlock(file); ++ out: ++ si_read_unlock(sb); ++ return err; ++} ++ ++static ssize_t aufs_aio_write(struct kiocb *kio, const struct iovec *iov, ++ unsigned long nv, loff_t pos) ++{ ++ ssize_t err; ++ aufs_bindex_t bstart; ++ struct au_pin pin; ++ struct dentry *dentry; ++ struct inode *inode; ++ struct super_block *sb; ++ struct file *file, *h_file; ++ ++ file = kio->ki_filp; ++ dentry = file->f_dentry; ++ sb = dentry->d_sb; ++ inode = dentry->d_inode; ++ mutex_lock(&inode->i_mutex); ++ si_read_lock(sb, AuLock_FLUSH); ++ ++ err = au_reval_and_lock_fdi(file, au_reopen_nondir, /*wlock*/1); ++ if (unlikely(err)) ++ goto out; ++ ++ err = au_ready_to_write(file, -1, &pin); ++ di_downgrade_lock(dentry, AuLock_IR); ++ if (unlikely(err)) ++ goto out_unlock; ++ ++ err = -ENOSYS; ++ bstart = au_fbstart(file); ++ h_file = au_h_fptr(file, bstart); ++ au_unpin(&pin); ++ if (h_file->f_op && h_file->f_op->aio_write) { ++ err = security_file_permission(h_file, MAY_WRITE); ++ if (unlikely(err)) ++ goto out_unlock; ++ if (!is_sync_kiocb(kio)) { ++ get_file(h_file); ++ fput(file); ++ } ++ kio->ki_filp = h_file; ++ err = h_file->f_op->aio_write(kio, iov, nv, pos); ++ au_cpup_attr_timesizes(inode); ++ inode->i_mode = h_file->f_dentry->d_inode->i_mode; ++ } else ++ /* currently there is no such fs */ ++ WARN_ON_ONCE(h_file->f_op && h_file->f_op->write); ++ ++ out_unlock: ++ di_read_unlock(dentry, AuLock_IR); ++ fi_write_unlock(file); ++ out: ++ si_read_unlock(sb); ++ mutex_unlock(&inode->i_mutex); ++ return err; ++} ++ ++static ssize_t aufs_splice_read(struct file *file, loff_t *ppos, ++ struct pipe_inode_info *pipe, size_t len, ++ unsigned int flags) ++{ ++ ssize_t err; ++ struct file *h_file; ++ struct dentry *dentry; ++ struct super_block *sb; ++ ++ dentry = file->f_dentry; ++ sb = dentry->d_sb; ++ si_read_lock(sb, AuLock_FLUSH); ++ err = au_reval_and_lock_fdi(file, au_reopen_nondir, /*wlock*/0); ++ if (unlikely(err)) ++ goto out; ++ ++ err = -EINVAL; ++ h_file = au_h_fptr(file, au_fbstart(file)); ++ if (au_test_loopback_kthread()) { ++ file->f_mapping = h_file->f_mapping; ++ smp_mb(); /* unnecessary? */ ++ } ++ err = vfsub_splice_to(h_file, ppos, pipe, len, flags); ++ /* todo: necessasry? */ ++ /* file->f_ra = h_file->f_ra; */ ++ fsstack_copy_attr_atime(dentry->d_inode, h_file->f_dentry->d_inode); ++ ++ di_read_unlock(dentry, AuLock_IR); ++ fi_read_unlock(file); ++ ++ out: ++ si_read_unlock(sb); ++ return err; ++} ++ ++static ssize_t ++aufs_splice_write(struct pipe_inode_info *pipe, struct file *file, loff_t *ppos, ++ size_t len, unsigned int flags) ++{ ++ ssize_t err; ++ struct au_pin pin; ++ struct dentry *dentry; ++ struct inode *inode; ++ struct super_block *sb; ++ struct file *h_file; ++ ++ dentry = file->f_dentry; ++ inode = dentry->d_inode; ++ mutex_lock(&inode->i_mutex); ++ sb = dentry->d_sb; ++ si_read_lock(sb, AuLock_FLUSH); ++ ++ err = au_reval_and_lock_fdi(file, au_reopen_nondir, /*wlock*/1); ++ if (unlikely(err)) ++ goto out; ++ ++ err = au_ready_to_write(file, -1, &pin); ++ di_downgrade_lock(dentry, AuLock_IR); ++ if (unlikely(err)) ++ goto out_unlock; ++ ++ h_file = au_h_fptr(file, au_fbstart(file)); ++ au_unpin(&pin); ++ err = vfsub_splice_from(pipe, h_file, ppos, len, flags); ++ au_cpup_attr_timesizes(inode); ++ inode->i_mode = h_file->f_dentry->d_inode->i_mode; ++ ++ out_unlock: ++ di_read_unlock(dentry, AuLock_IR); ++ fi_write_unlock(file); ++ out: ++ si_read_unlock(sb); ++ mutex_unlock(&inode->i_mutex); ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static struct file *au_safe_file(struct vm_area_struct *vma) ++{ ++ struct file *file; ++ ++ file = vma->vm_file; ++ if (file->private_data && au_test_aufs(file->f_dentry->d_sb)) ++ return file; ++ return NULL; ++} ++ ++static void au_reset_file(struct vm_area_struct *vma, struct file *file) ++{ ++ vma->vm_file = file; ++ /* smp_mb(); */ /* flush vm_file */ ++} ++ ++static int aufs_fault(struct vm_area_struct *vma, struct vm_fault *vmf) ++{ ++ int err; ++ static DECLARE_WAIT_QUEUE_HEAD(wq); ++ struct file *file, *h_file; ++ struct au_finfo *finfo; ++ ++ /* todo: non-robr mode, user vm_file as it is? */ ++ wait_event(wq, (file = au_safe_file(vma))); ++ ++ /* do not revalidate, no si lock */ ++ finfo = au_fi(file); ++ h_file = finfo->fi_hfile[0 + finfo->fi_bstart].hf_file; ++ AuDebugOn(!h_file || !finfo->fi_h_vm_ops); ++ ++ fi_write_lock(file); ++ vma->vm_file = h_file; ++ err = finfo->fi_h_vm_ops->fault(vma, vmf); ++ /* todo: necessary? */ ++ /* file->f_ra = h_file->f_ra; */ ++ au_reset_file(vma, file); ++ fi_write_unlock(file); ++#if 0 /* def CONFIG_SMP */ ++ /* wake_up_nr(&wq, online_cpu - 1); */ ++ wake_up_all(&wq); ++#else ++ wake_up(&wq); ++#endif ++ ++ return err; ++} ++ ++static int aufs_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) ++{ ++ int err; ++ static DECLARE_WAIT_QUEUE_HEAD(wq); ++ struct file *file, *h_file; ++ struct au_finfo *finfo; ++ ++ wait_event(wq, (file = au_safe_file(vma))); ++ ++ finfo = au_fi(file); ++ h_file = finfo->fi_hfile[0 + finfo->fi_bstart].hf_file; ++ AuDebugOn(!h_file || !finfo->fi_h_vm_ops); ++ ++ fi_write_lock(file); ++ vma->vm_file = h_file; ++ err = finfo->fi_h_vm_ops->page_mkwrite(vma, vmf); ++ au_reset_file(vma, file); ++ fi_write_unlock(file); ++ wake_up(&wq); ++ ++ return err; ++} ++ ++static void aufs_vm_close(struct vm_area_struct *vma) ++{ ++ static DECLARE_WAIT_QUEUE_HEAD(wq); ++ struct file *file, *h_file; ++ struct au_finfo *finfo; ++ ++ wait_event(wq, (file = au_safe_file(vma))); ++ ++ finfo = au_fi(file); ++ h_file = finfo->fi_hfile[0 + finfo->fi_bstart].hf_file; ++ AuDebugOn(!h_file || !finfo->fi_h_vm_ops); ++ ++ fi_write_lock(file); ++ vma->vm_file = h_file; ++ finfo->fi_h_vm_ops->close(vma); ++ au_reset_file(vma, file); ++ fi_write_unlock(file); ++ wake_up(&wq); ++} ++ ++static struct vm_operations_struct aufs_vm_ops = { ++ /* .close and .page_mkwrite are not set by default */ ++ .fault = aufs_fault, ++}; ++ ++/* ---------------------------------------------------------------------- */ ++ ++static struct vm_operations_struct *au_vm_ops(struct file *h_file, ++ struct vm_area_struct *vma) ++{ ++ struct vm_operations_struct *vm_ops; ++ int err; ++ ++ vm_ops = ERR_PTR(-ENODEV); ++ if (!h_file->f_op || !h_file->f_op->mmap) ++ goto out; ++ ++ err = h_file->f_op->mmap(h_file, vma); ++ vm_ops = ERR_PTR(err); ++ if (unlikely(err)) ++ goto out; ++ ++ vm_ops = vma->vm_ops; ++ err = do_munmap(current->mm, vma->vm_start, ++ vma->vm_end - vma->vm_start); ++ if (unlikely(err)) { ++ AuIOErr("failed internal unmapping %.*s, %d\n", ++ AuDLNPair(h_file->f_dentry), err); ++ vm_ops = ERR_PTR(-EIO); ++ } ++ ++ out: ++ return vm_ops; ++} ++ ++static int au_custom_vm_ops(struct au_finfo *finfo, struct vm_area_struct *vma) ++{ ++ int err; ++ struct vm_operations_struct *h_ops; ++ ++ AuRwMustAnyLock(&finfo->fi_rwsem); ++ ++ err = 0; ++ h_ops = finfo->fi_h_vm_ops; ++ AuDebugOn(!h_ops); ++ if ((!h_ops->page_mkwrite && !h_ops->close) ++ || finfo->fi_vm_ops) ++ goto out; ++ ++ err = -ENOMEM; ++ finfo->fi_vm_ops = kmemdup(&aufs_vm_ops, sizeof(aufs_vm_ops), GFP_NOFS); ++ if (unlikely(!finfo->fi_vm_ops)) ++ goto out; ++ ++ err = 0; ++ if (h_ops->page_mkwrite) ++ finfo->fi_vm_ops->page_mkwrite = aufs_page_mkwrite; ++ if (h_ops->close) ++ finfo->fi_vm_ops->close = aufs_vm_close; ++ ++ vma->vm_ops = finfo->fi_vm_ops; ++ ++ out: ++ return err; ++} ++ ++static int aufs_mmap(struct file *file, struct vm_area_struct *vma) ++{ ++ int err; ++ unsigned char wlock, mmapped; ++ struct dentry *dentry; ++ struct super_block *sb; ++ struct file *h_file; ++ struct vm_operations_struct *vm_ops; ++ ++ dentry = file->f_dentry; ++ wlock = !!(file->f_mode & FMODE_WRITE) && (vma->vm_flags & VM_SHARED); ++ sb = dentry->d_sb; ++ si_read_lock(sb, AuLock_FLUSH); ++ err = au_reval_and_lock_fdi(file, au_reopen_nondir, /*wlock*/1); ++ if (unlikely(err)) ++ goto out; ++ ++ mmapped = !!au_test_mmapped(file); ++ if (wlock) { ++ struct au_pin pin; ++ ++ err = au_ready_to_write(file, -1, &pin); ++ di_downgrade_lock(dentry, AuLock_IR); ++ if (unlikely(err)) ++ goto out_unlock; ++ au_unpin(&pin); ++ } else ++ di_downgrade_lock(dentry, AuLock_IR); ++ ++ h_file = au_h_fptr(file, au_fbstart(file)); ++ if (!mmapped && au_test_fs_bad_mapping(h_file->f_dentry->d_sb)) { ++ /* ++ * by this assignment, f_mapping will differs from aufs inode ++ * i_mapping. ++ * if someone else mixes the use of f_dentry->d_inode and ++ * f_mapping->host, then a problem may arise. ++ */ ++ file->f_mapping = h_file->f_mapping; ++ } ++ ++ vm_ops = NULL; ++ if (!mmapped) { ++ vm_ops = au_vm_ops(h_file, vma); ++ err = PTR_ERR(vm_ops); ++ if (IS_ERR(vm_ops)) ++ goto out_unlock; ++ } ++ ++ /* ++ * unnecessary to handle MAP_DENYWRITE and deny_write_access()? ++ * currently MAP_DENYWRITE from userspace is ignored, but elf loader ++ * sets it. when FMODE_EXEC is set (by open_exec() or sys_uselib()), ++ * both of the aufs file and the lower file is deny_write_access()-ed. ++ * finally I hope we can skip handlling MAP_DENYWRITE here. ++ */ ++ err = generic_file_mmap(file, vma); ++ if (unlikely(err)) ++ goto out_unlock; ++ ++ vma->vm_ops = &aufs_vm_ops; ++ /* test again */ ++ if (!au_test_mmapped(file)) ++ au_fi(file)->fi_h_vm_ops = vm_ops; ++ ++ err = au_custom_vm_ops(au_fi(file), vma); ++ if (unlikely(err)) ++ goto out_unlock; ++ ++ vfsub_file_accessed(h_file); ++ fsstack_copy_attr_atime(dentry->d_inode, h_file->f_dentry->d_inode); ++ ++ out_unlock: ++ di_read_unlock(dentry, AuLock_IR); ++ fi_write_unlock(file); ++ out: ++ si_read_unlock(sb); ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static int aufs_fsync_nondir(struct file *file, struct dentry *dentry, ++ int datasync) ++{ ++ int err; ++ struct au_pin pin; ++ struct inode *inode; ++ struct file *h_file; ++ struct super_block *sb; ++ ++ inode = dentry->d_inode; ++ IMustLock(file->f_mapping->host); ++ if (inode != file->f_mapping->host) { ++ mutex_unlock(&file->f_mapping->host->i_mutex); ++ mutex_lock(&inode->i_mutex); ++ } ++ IMustLock(inode); ++ ++ sb = dentry->d_sb; ++ si_read_lock(sb, AuLock_FLUSH); ++ ++ err = 0; /* -EBADF; */ /* posix? */ ++ if (unlikely(!(file->f_mode & FMODE_WRITE))) ++ goto out; ++ err = au_reval_and_lock_fdi(file, au_reopen_nondir, /*wlock*/1); ++ if (unlikely(err)) ++ goto out; ++ ++ err = au_ready_to_write(file, -1, &pin); ++ di_downgrade_lock(dentry, AuLock_IR); ++ if (unlikely(err)) ++ goto out_unlock; ++ au_unpin(&pin); ++ ++ err = -EINVAL; ++ h_file = au_h_fptr(file, au_fbstart(file)); ++ if (h_file->f_op && h_file->f_op->fsync) { ++ struct dentry *h_d; ++ struct mutex *h_mtx; ++ ++ /* ++ * no filemap_fdatawrite() since aufs file has no its own ++ * mapping, but dir. ++ */ ++ h_d = h_file->f_dentry; ++ h_mtx = &h_d->d_inode->i_mutex; ++ mutex_lock_nested(h_mtx, AuLsc_I_CHILD); ++ err = h_file->f_op->fsync(h_file, h_d, datasync); ++ if (!err) ++ vfsub_update_h_iattr(&h_file->f_path, /*did*/NULL); ++ /*ignore*/ ++ au_cpup_attr_timesizes(inode); ++ mutex_unlock(h_mtx); ++ } ++ ++ out_unlock: ++ di_read_unlock(dentry, AuLock_IR); ++ fi_write_unlock(file); ++ out: ++ si_read_unlock(sb); ++ if (inode != file->f_mapping->host) { ++ mutex_unlock(&inode->i_mutex); ++ mutex_lock(&file->f_mapping->host->i_mutex); ++ } ++ return err; ++} ++ ++/* no one supports this operation, currently */ ++#if 0 ++static int aufs_aio_fsync_nondir(struct kiocb *kio, int datasync) ++{ ++ int err; ++ struct au_pin pin; ++ struct dentry *dentry; ++ struct inode *inode; ++ struct file *file, *h_file; ++ struct super_block *sb; ++ ++ file = kio->ki_filp; ++ dentry = file->f_dentry; ++ inode = dentry->d_inode; ++ mutex_lock(&inode->i_mutex); ++ ++ sb = dentry->d_sb; ++ si_read_lock(sb, AuLock_FLUSH); ++ ++ err = 0; /* -EBADF; */ /* posix? */ ++ if (unlikely(!(file->f_mode & FMODE_WRITE))) ++ goto out; ++ err = au_reval_and_lock_fdi(file, au_reopen_nondir, /*wlock*/1); ++ if (unlikely(err)) ++ goto out; ++ ++ err = au_ready_to_write(file, -1, &pin); ++ di_downgrade_lock(dentry, AuLock_IR); ++ if (unlikely(err)) ++ goto out_unlock; ++ au_unpin(&pin); ++ ++ err = -ENOSYS; ++ h_file = au_h_fptr(file, au_fbstart(file)); ++ if (h_file->f_op && h_file->f_op->aio_fsync) { ++ struct dentry *h_d; ++ struct mutex *h_mtx; ++ ++ h_d = h_file->f_dentry; ++ h_mtx = &h_d->d_inode->i_mutex; ++ if (!is_sync_kiocb(kio)) { ++ get_file(h_file); ++ fput(file); ++ } ++ kio->ki_filp = h_file; ++ err = h_file->f_op->aio_fsync(kio, datasync); ++ mutex_lock_nested(h_mtx, AuLsc_I_CHILD); ++ if (!err) ++ vfsub_update_h_iattr(&h_file->f_path, /*did*/NULL); ++ /*ignore*/ ++ au_cpup_attr_timesizes(inode); ++ mutex_unlock(h_mtx); ++ } ++ ++ out_unlock: ++ di_read_unlock(dentry, AuLock_IR); ++ fi_write_unlock(file); ++ out: ++ si_read_unlock(sb); ++ mutex_unlock(&inode->i_mutex); ++ return err; ++} ++#endif ++ ++static int aufs_fasync(int fd, struct file *file, int flag) ++{ ++ int err; ++ struct file *h_file; ++ struct dentry *dentry; ++ struct super_block *sb; ++ ++ dentry = file->f_dentry; ++ sb = dentry->d_sb; ++ si_read_lock(sb, AuLock_FLUSH); ++ err = au_reval_and_lock_fdi(file, au_reopen_nondir, /*wlock*/0); ++ if (unlikely(err)) ++ goto out; ++ ++ h_file = au_h_fptr(file, au_fbstart(file)); ++ if (h_file->f_op && h_file->f_op->fasync) ++ err = h_file->f_op->fasync(fd, h_file, flag); ++ ++ di_read_unlock(dentry, AuLock_IR); ++ fi_read_unlock(file); ++ ++ out: ++ si_read_unlock(sb); ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* no one supports this operation, currently */ ++#if 0 ++static ssize_t aufs_sendpage(struct file *file, struct page *page, int offset, ++ size_t len, loff_t *pos , int more) ++{ ++} ++#endif ++ ++/* ---------------------------------------------------------------------- */ ++ ++const struct file_operations aufs_file_fop = { ++ /* ++ * while generic_file_llseek/_unlocked() don't use BKL, ++ * don't use it since it operates file->f_mapping->host. ++ * in aufs, it may be a real file and may confuse users by UDBA. ++ */ ++ /* .llseek = generic_file_llseek, */ ++ ++ .read = aufs_read, ++ .write = aufs_write, ++ .aio_read = aufs_aio_read, ++ .aio_write = aufs_aio_write, ++#ifdef CONFIG_AUFS_POLL ++ .poll = aufs_poll, ++#endif ++ .mmap = aufs_mmap, ++ .open = aufs_open_nondir, ++ .flush = aufs_flush, ++ .release = aufs_release_nondir, ++ .fsync = aufs_fsync_nondir, ++ /* .aio_fsync = aufs_aio_fsync_nondir, */ ++ .fasync = aufs_fasync, ++ /* .sendpage = aufs_sendpage, */ ++ .splice_write = aufs_splice_write, ++ .splice_read = aufs_splice_read, ++#if 0 ++ .aio_splice_write = aufs_aio_splice_write, ++ .aio_splice_read = aufs_aio_splice_read ++#endif ++}; +diff -Nur linux-2.6.31.4.orig/fs/aufs/fstype.h linux-2.6.31.4/fs/aufs/fstype.h +--- linux-2.6.31.4.orig/fs/aufs/fstype.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/fstype.h 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,474 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * judging filesystem type ++ */ ++ ++#ifndef __AUFS_FSTYPE_H__ ++#define __AUFS_FSTYPE_H__ ++ ++#ifdef __KERNEL__ ++ ++#include <linux/cramfs_fs.h> ++#include <linux/fs.h> ++#include <linux/magic.h> ++#include <linux/romfs_fs.h> ++#include <linux/aufs_type.h> ++ ++static inline int au_test_aufs(struct super_block *sb) ++{ ++ return sb->s_magic == AUFS_SUPER_MAGIC; ++} ++ ++static inline const char *au_sbtype(struct super_block *sb) ++{ ++ return sb->s_type->name; ++} ++ ++static inline int au_test_iso9660(struct super_block *sb __maybe_unused) ++{ ++#if defined(CONFIG_ROMFS_FS) || defined(CONFIG_ROMFS_FS_MODULE) ++ return sb->s_magic == ROMFS_MAGIC; ++#else ++ return 0; ++#endif ++} ++ ++static inline int au_test_romfs(struct super_block *sb __maybe_unused) ++{ ++#if defined(CONFIG_ISO9660_FS) || defined(CONFIG_ISO9660_FS_MODULE) ++ return sb->s_magic == ISOFS_SUPER_MAGIC; ++#else ++ return 0; ++#endif ++} ++ ++static inline int au_test_cramfs(struct super_block *sb __maybe_unused) ++{ ++#if defined(CONFIG_CRAMFS) || defined(CONFIG_CRAMFS_MODULE) ++ return sb->s_magic == CRAMFS_MAGIC; ++#endif ++ return 0; ++} ++ ++static inline int au_test_nfs(struct super_block *sb __maybe_unused) ++{ ++#if defined(CONFIG_NFS_FS) || defined(CONFIG_NFS_FS_MODULE) ++ return sb->s_magic == NFS_SUPER_MAGIC; ++#else ++ return 0; ++#endif ++} ++ ++static inline int au_test_fuse(struct super_block *sb __maybe_unused) ++{ ++#if defined(CONFIG_FUSE_FS) || defined(CONFIG_FUSE_FS_MODULE) ++ return sb->s_magic == FUSE_SUPER_MAGIC; ++#else ++ return 0; ++#endif ++} ++ ++static inline int au_test_xfs(struct super_block *sb __maybe_unused) ++{ ++#if defined(CONFIG_XFS_FS) || defined(CONFIG_XFS_FS_MODULE) ++ return sb->s_magic == XFS_SB_MAGIC; ++#else ++ return 0; ++#endif ++} ++ ++static inline int au_test_tmpfs(struct super_block *sb __maybe_unused) ++{ ++#ifdef CONFIG_TMPFS ++ return sb->s_magic == TMPFS_MAGIC; ++#else ++ return 0; ++#endif ++} ++ ++static inline int au_test_ecryptfs(struct super_block *sb __maybe_unused) ++{ ++#if defined(CONFIG_ECRYPT_FS) || defined(CONFIG_ECRYPT_FS_MODULE) ++ return !strcmp(au_sbtype(sb), "ecryptfs"); ++#else ++ return 0; ++#endif ++} ++ ++static inline int au_test_smbfs(struct super_block *sb __maybe_unused) ++{ ++#if defined(CONFIG_SMB_FS) || defined(CONFIG_SMB_FS_MODULE) ++ return sb->s_magic == SMB_SUPER_MAGIC; ++#else ++ return 0; ++#endif ++} ++ ++static inline int au_test_ocfs2(struct super_block *sb __maybe_unused) ++{ ++#if defined(CONFIG_OCFS2_FS) || defined(CONFIG_OCFS2_FS_MODULE) ++ return sb->s_magic == OCFS2_SUPER_MAGIC; ++#else ++ return 0; ++#endif ++} ++ ++static inline int au_test_ocfs2_dlmfs(struct super_block *sb __maybe_unused) ++{ ++#if defined(CONFIG_OCFS2_FS_O2CB) || defined(CONFIG_OCFS2_FS_O2CB_MODULE) ++ return sb->s_magic == DLMFS_MAGIC; ++#else ++ return 0; ++#endif ++} ++ ++static inline int au_test_coda(struct super_block *sb __maybe_unused) ++{ ++#if defined(CONFIG_CODA_FS) || defined(CONFIG_CODA_FS_MODULE) ++ return sb->s_magic == CODA_SUPER_MAGIC; ++#else ++ return 0; ++#endif ++} ++ ++static inline int au_test_v9fs(struct super_block *sb __maybe_unused) ++{ ++#if defined(CONFIG_9P_FS) || defined(CONFIG_9P_FS_MODULE) ++ return sb->s_magic == V9FS_MAGIC; ++#else ++ return 0; ++#endif ++} ++ ++static inline int au_test_ext4(struct super_block *sb __maybe_unused) ++{ ++#if defined(CONFIG_EXT4DEV_FS) || defined(CONFIG_EXT4DEV_FS_MODULE) ++ return sb->s_magic == EXT4_SUPER_MAGIC; ++#else ++ return 0; ++#endif ++} ++ ++static inline int au_test_sysv(struct super_block *sb __maybe_unused) ++{ ++#if defined(CONFIG_SYSV_FS) || defined(CONFIG_SYSV_FS_MODULE) ++ return !strcmp(au_sbtype(sb), "sysv"); ++#else ++ return 0; ++#endif ++} ++ ++static inline int au_test_ramfs(struct super_block *sb) ++{ ++ return sb->s_magic == RAMFS_MAGIC; ++} ++ ++static inline int au_test_ubifs(struct super_block *sb __maybe_unused) ++{ ++#if defined(CONFIG_UBIFS_FS) || defined(CONFIG_UBIFS_FS_MODULE) ++ return sb->s_magic == UBIFS_SUPER_MAGIC; ++#else ++ return 0; ++#endif ++} ++ ++static inline int au_test_procfs(struct super_block *sb __maybe_unused) ++{ ++#ifdef CONFIG_PROC_FS ++ return sb->s_magic == PROC_SUPER_MAGIC; ++#else ++ return 0; ++#endif ++} ++ ++static inline int au_test_sysfs(struct super_block *sb __maybe_unused) ++{ ++#ifdef CONFIG_SYSFS ++ return sb->s_magic == SYSFS_MAGIC; ++#else ++ return 0; ++#endif ++} ++ ++static inline int au_test_configfs(struct super_block *sb __maybe_unused) ++{ ++#if defined(CONFIG_CONFIGFS_FS) || defined(CONFIG_CONFIGFS_FS_MODULE) ++ return sb->s_magic == CONFIGFS_MAGIC; ++#else ++ return 0; ++#endif ++} ++ ++static inline int au_test_minix(struct super_block *sb __maybe_unused) ++{ ++#if defined(CONFIG_MINIX_FS) || defined(CONFIG_MINIX_FS_MODULE) ++ return sb->s_magic == MINIX3_SUPER_MAGIC ++ || sb->s_magic == MINIX2_SUPER_MAGIC ++ || sb->s_magic == MINIX2_SUPER_MAGIC2 ++ || sb->s_magic == MINIX_SUPER_MAGIC ++ || sb->s_magic == MINIX_SUPER_MAGIC2; ++#else ++ return 0; ++#endif ++} ++ ++static inline int au_test_cifs(struct super_block *sb __maybe_unused) ++{ ++#if defined(CONFIG_CIFS_FS) || defined(CONFIGCIFS_FS_MODULE) ++ return sb->s_magic == CIFS_MAGIC_NUMBER; ++#else ++ return 0; ++#endif ++} ++ ++static inline int au_test_fat(struct super_block *sb __maybe_unused) ++{ ++#if defined(CONFIG_FAT_FS) || defined(CONFIG_FAT_FS_MODULE) ++ return sb->s_magic == MSDOS_SUPER_MAGIC; ++#else ++ return 0; ++#endif ++} ++ ++static inline int au_test_msdos(struct super_block *sb) ++{ ++ return au_test_fat(sb); ++} ++ ++static inline int au_test_vfat(struct super_block *sb) ++{ ++ return au_test_fat(sb); ++} ++ ++static inline int au_test_securityfs(struct super_block *sb __maybe_unused) ++{ ++#ifdef CONFIG_SECURITYFS ++ return sb->s_magic == SECURITYFS_MAGIC; ++#else ++ return 0; ++#endif ++} ++ ++static inline int au_test_squashfs(struct super_block *sb __maybe_unused) ++{ ++#if defined(CONFIG_SQUASHFS) || defined(CONFIG_SQUASHFS_MODULE) ++ return sb->s_magic == SQUASHFS_MAGIC; ++#else ++ return 0; ++#endif ++} ++ ++static inline int au_test_btrfs(struct super_block *sb __maybe_unused) ++{ ++#if defined(CONFIG_BTRFS_FS) || defined(CONFIG_BTRFS_FS_MODULE) ++ return sb->s_magic == BTRFS_SUPER_MAGIC; ++#else ++ return 0; ++#endif ++} ++ ++static inline int au_test_xenfs(struct super_block *sb __maybe_unused) ++{ ++#if defined(CONFIG_XENFS) || defined(CONFIG_XENFS_MODULE) ++ return sb->s_magic == XENFS_SUPER_MAGIC; ++#else ++ return 0; ++#endif ++} ++ ++static inline int au_test_debugfs(struct super_block *sb __maybe_unused) ++{ ++#ifdef CONFIG_DEBUG_FS ++ return sb->s_magic == DEBUGFS_MAGIC; ++#else ++ return 0; ++#endif ++} ++ ++/* ---------------------------------------------------------------------- */ ++/* ++ * they can't be an aufs branch. ++ */ ++static inline int au_test_fs_unsuppoted(struct super_block *sb) ++{ ++ return ++#ifndef CONFIG_AUFS_BR_RAMFS ++ au_test_ramfs(sb) || ++#endif ++ au_test_procfs(sb) ++ || au_test_sysfs(sb) ++ || au_test_configfs(sb) ++ || au_test_debugfs(sb) ++ || au_test_securityfs(sb) ++ || au_test_xenfs(sb) ++ /* || !strcmp(au_sbtype(sb), "unionfs") */ ++ || au_test_aufs(sb); /* will be supported in next version */ ++} ++ ++/* ++ * If the filesystem supports NFS-export, then it has to support NULL as ++ * a nameidata parameter for ->create(), ->lookup() and ->d_revalidate(). ++ * We can apply this principle when we handle a lower filesystem. ++ */ ++static inline int au_test_fs_null_nd(struct super_block *sb) ++{ ++ return !!sb->s_export_op; ++} ++ ++static inline int au_test_fs_remote(struct super_block *sb) ++{ ++ return !au_test_tmpfs(sb) ++#ifdef CONFIG_AUFS_BR_RAMFS ++ && !au_test_ramfs(sb) ++#endif ++ && !(sb->s_type->fs_flags & FS_REQUIRES_DEV); ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* ++ * Note: these functions (below) are created after reading ->getattr() in all ++ * filesystems under linux/fs. it means we have to do so in every update... ++ */ ++ ++/* ++ * some filesystems require getattr to refresh the inode attributes before ++ * referencing. ++ * in most cases, we can rely on the inode attribute in NFS (or every remote fs) ++ * and leave the work for d_revalidate() ++ */ ++static inline int au_test_fs_refresh_iattr(struct super_block *sb) ++{ ++ return au_test_nfs(sb) ++ || au_test_fuse(sb) ++ /* || au_test_smbfs(sb) */ /* untested */ ++ /* || au_test_ocfs2(sb) */ /* untested */ ++ /* || au_test_btrfs(sb) */ /* untested */ ++ /* || au_test_coda(sb) */ /* untested */ ++ /* || au_test_v9fs(sb) */ /* untested */ ++ ; ++} ++ ++/* ++ * filesystems which don't maintain i_size or i_blocks. ++ */ ++static inline int au_test_fs_bad_iattr_size(struct super_block *sb) ++{ ++ return au_test_xfs(sb) ++ /* || au_test_ext4(sb) */ /* untested */ ++ /* || au_test_ocfs2(sb) */ /* untested */ ++ /* || au_test_ocfs2_dlmfs(sb) */ /* untested */ ++ /* || au_test_sysv(sb) */ /* untested */ ++ /* || au_test_ubifs(sb) */ /* untested */ ++ /* || au_test_minix(sb) */ /* untested */ ++ ; ++} ++ ++/* ++ * filesystems which don't store the correct value in some of their inode ++ * attributes. ++ */ ++static inline int au_test_fs_bad_iattr(struct super_block *sb) ++{ ++ return au_test_fs_bad_iattr_size(sb) ++ /* || au_test_cifs(sb) */ /* untested */ ++ || au_test_fat(sb) ++ || au_test_msdos(sb) ++ || au_test_vfat(sb); ++} ++ ++/* they don't check i_nlink in link(2) */ ++static inline int au_test_fs_no_limit_nlink(struct super_block *sb) ++{ ++ return au_test_tmpfs(sb) ++#ifdef CONFIG_AUFS_BR_RAMFS ++ || au_test_ramfs(sb) ++#endif ++ || au_test_ubifs(sb); ++} ++ ++/* ++ * filesystems which sets S_NOATIME and S_NOCMTIME. ++ */ ++static inline int au_test_fs_notime(struct super_block *sb) ++{ ++ return au_test_nfs(sb) ++ || au_test_fuse(sb) ++ || au_test_ubifs(sb) ++ /* || au_test_cifs(sb) */ /* untested */ ++ ; ++} ++ ++/* ++ * filesystems which requires replacing i_mapping. ++ */ ++static inline int au_test_fs_bad_mapping(struct super_block *sb) ++{ ++ return au_test_fuse(sb) ++ || au_test_ubifs(sb); ++} ++ ++/* temporary support for i#1 in cramfs */ ++static inline int au_test_fs_unique_ino(struct inode *inode) ++{ ++ if (au_test_cramfs(inode->i_sb)) ++ return inode->i_ino != 1; ++ return 1; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* ++ * the filesystem where the xino files placed must support i/o after unlink and ++ * maintain i_size and i_blocks. ++ */ ++static inline int au_test_fs_bad_xino(struct super_block *sb) ++{ ++ return au_test_fs_remote(sb) ++ || au_test_fs_bad_iattr_size(sb) ++#ifdef CONFIG_AUFS_BR_RAMFS ++ || !(au_test_ramfs(sb) || au_test_fs_null_nd(sb)) ++#else ++ || !au_test_fs_null_nd(sb) /* to keep xino code simple */ ++#endif ++ /* don't want unnecessary work for xino */ ++ || au_test_aufs(sb) ++ || au_test_ecryptfs(sb); ++} ++ ++static inline int au_test_fs_trunc_xino(struct super_block *sb) ++{ ++ return au_test_tmpfs(sb) ++ || au_test_ramfs(sb); ++} ++ ++/* ++ * test if the @sb is real-readonly. ++ */ ++static inline int au_test_fs_rr(struct super_block *sb) ++{ ++ return au_test_squashfs(sb) ++ || au_test_iso9660(sb) ++ || au_test_cramfs(sb) ++ || au_test_romfs(sb); ++} ++ ++#endif /* __KERNEL__ */ ++#endif /* __AUFS_FSTYPE_H__ */ +diff -Nur linux-2.6.31.4.orig/fs/aufs/hinotify.c linux-2.6.31.4/fs/aufs/hinotify.c +--- linux-2.6.31.4.orig/fs/aufs/hinotify.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/hinotify.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,755 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * inotify for the lower directories ++ */ ++ ++#include "aufs.h" ++ ++static const __u32 AuHinMask = (IN_MOVE | IN_DELETE | IN_CREATE); ++static struct inotify_handle *au_hin_handle; ++ ++AuCacheFuncs(hinotify, HINOTIFY); ++ ++int au_hin_alloc(struct au_hinode *hinode, struct inode *inode, ++ struct inode *h_inode) ++{ ++ int err; ++ struct au_hinotify *hin; ++ s32 wd; ++ ++ err = -ENOMEM; ++ hin = au_cache_alloc_hinotify(); ++ if (hin) { ++ AuDebugOn(hinode->hi_notify); ++ hinode->hi_notify = hin; ++ hin->hin_aufs_inode = inode; ++ ++ inotify_init_watch(&hin->hin_watch); ++ wd = inotify_add_watch(au_hin_handle, &hin->hin_watch, h_inode, ++ AuHinMask); ++ if (wd >= 0) ++ return 0; /* success */ ++ ++ err = wd; ++ put_inotify_watch(&hin->hin_watch); ++ au_cache_free_hinotify(hin); ++ hinode->hi_notify = NULL; ++ } ++ ++ return err; ++} ++ ++void au_hin_free(struct au_hinode *hinode) ++{ ++ int err; ++ struct au_hinotify *hin; ++ ++ hin = hinode->hi_notify; ++ if (hin) { ++ err = 0; ++ if (atomic_read(&hin->hin_watch.count)) ++ err = inotify_rm_watch(au_hin_handle, &hin->hin_watch); ++ if (unlikely(err)) ++ /* it means the watch is already removed */ ++ AuWarn("failed inotify_rm_watch() %d\n", err); ++ au_cache_free_hinotify(hin); ++ hinode->hi_notify = NULL; ++ } ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++void au_hin_ctl(struct au_hinode *hinode, int do_set) ++{ ++ struct inode *h_inode; ++ struct inotify_watch *watch; ++ ++ if (!hinode->hi_notify) ++ return; ++ ++ h_inode = hinode->hi_inode; ++ IMustLock(h_inode); ++ ++ /* todo: try inotify_find_update_watch()? */ ++ watch = &hinode->hi_notify->hin_watch; ++ mutex_lock(&h_inode->inotify_mutex); ++ /* mutex_lock(&watch->ih->mutex); */ ++ if (do_set) { ++ AuDebugOn(watch->mask & AuHinMask); ++ watch->mask |= AuHinMask; ++ } else { ++ AuDebugOn(!(watch->mask & AuHinMask)); ++ watch->mask &= ~AuHinMask; ++ } ++ /* mutex_unlock(&watch->ih->mutex); */ ++ mutex_unlock(&h_inode->inotify_mutex); ++} ++ ++void au_reset_hinotify(struct inode *inode, unsigned int flags) ++{ ++ aufs_bindex_t bindex, bend; ++ struct inode *hi; ++ struct dentry *iwhdentry; ++ ++ bend = au_ibend(inode); ++ for (bindex = au_ibstart(inode); bindex <= bend; bindex++) { ++ hi = au_h_iptr(inode, bindex); ++ if (!hi) ++ continue; ++ ++ /* mutex_lock_nested(&hi->i_mutex, AuLsc_I_CHILD); */ ++ iwhdentry = au_hi_wh(inode, bindex); ++ if (iwhdentry) ++ dget(iwhdentry); ++ au_igrab(hi); ++ au_set_h_iptr(inode, bindex, NULL, 0); ++ au_set_h_iptr(inode, bindex, au_igrab(hi), ++ flags & ~AuHi_XINO); ++ iput(hi); ++ dput(iwhdentry); ++ /* mutex_unlock(&hi->i_mutex); */ ++ } ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static int hin_xino(struct inode *inode, struct inode *h_inode) ++{ ++ int err; ++ aufs_bindex_t bindex, bend, bfound, bstart; ++ struct inode *h_i; ++ ++ err = 0; ++ if (unlikely(inode->i_ino == AUFS_ROOT_INO)) { ++ AuWarn("branch root dir was changed\n"); ++ goto out; ++ } ++ ++ bfound = -1; ++ bend = au_ibend(inode); ++ bstart = au_ibstart(inode); ++#if 0 /* reserved for future use */ ++ if (bindex == bend) { ++ /* keep this ino in rename case */ ++ goto out; ++ } ++#endif ++ for (bindex = bstart; bindex <= bend; bindex++) { ++ if (au_h_iptr(inode, bindex) == h_inode) { ++ bfound = bindex; ++ break; ++ } ++ } ++ if (bfound < 0) ++ goto out; ++ ++ for (bindex = bstart; bindex <= bend; bindex++) { ++ h_i = au_h_iptr(inode, bindex); ++ if (!h_i) ++ continue; ++ ++ err = au_xino_write(inode->i_sb, bindex, h_i->i_ino, /*ino*/0); ++ /* ignore this error */ ++ /* bad action? */ ++ } ++ ++ /* children inode number will be broken */ ++ ++ out: ++ AuTraceErr(err); ++ return err; ++} ++ ++static int hin_gen_tree(struct dentry *dentry) ++{ ++ int err, i, j, ndentry; ++ struct au_dcsub_pages dpages; ++ struct au_dpage *dpage; ++ struct dentry **dentries; ++ ++ err = au_dpages_init(&dpages, GFP_NOFS); ++ if (unlikely(err)) ++ goto out; ++ err = au_dcsub_pages(&dpages, dentry, NULL, NULL); ++ if (unlikely(err)) ++ goto out_dpages; ++ ++ for (i = 0; i < dpages.ndpage; i++) { ++ dpage = dpages.dpages + i; ++ dentries = dpage->dentries; ++ ndentry = dpage->ndentry; ++ for (j = 0; j < ndentry; j++) { ++ struct dentry *d; ++ ++ d = dentries[j]; ++ if (IS_ROOT(d)) ++ continue; ++ ++ d_drop(d); ++ au_digen_dec(d); ++ if (d->d_inode) ++ /* todo: reset children xino? ++ cached children only? */ ++ au_iigen_dec(d->d_inode); ++ } ++ } ++ ++ out_dpages: ++ au_dpages_free(&dpages); ++ ++ /* discard children */ ++ dentry_unhash(dentry); ++ dput(dentry); ++ out: ++ return err; ++} ++ ++/* ++ * return 0 if processed. ++ */ ++static int hin_gen_by_inode(char *name, unsigned int nlen, struct inode *inode, ++ const unsigned int isdir) ++{ ++ int err; ++ struct dentry *d; ++ struct qstr *dname; ++ ++ err = 1; ++ if (unlikely(inode->i_ino == AUFS_ROOT_INO)) { ++ AuWarn("branch root dir was changed\n"); ++ err = 0; ++ goto out; ++ } ++ ++ if (!isdir) { ++ AuDebugOn(!name); ++ au_iigen_dec(inode); ++ spin_lock(&dcache_lock); ++ list_for_each_entry(d, &inode->i_dentry, d_alias) { ++ dname = &d->d_name; ++ if (dname->len != nlen ++ && memcmp(dname->name, name, nlen)) ++ continue; ++ err = 0; ++ spin_lock(&d->d_lock); ++ __d_drop(d); ++ au_digen_dec(d); ++ spin_unlock(&d->d_lock); ++ break; ++ } ++ spin_unlock(&dcache_lock); ++ } else { ++ au_fset_si(au_sbi(inode->i_sb), FAILED_REFRESH_DIRS); ++ d = d_find_alias(inode); ++ if (!d) { ++ au_iigen_dec(inode); ++ goto out; ++ } ++ ++ dname = &d->d_name; ++ if (dname->len == nlen && !memcmp(dname->name, name, nlen)) ++ err = hin_gen_tree(d); ++ dput(d); ++ } ++ ++ out: ++ AuTraceErr(err); ++ return err; ++} ++ ++static int hin_gen_by_name(struct dentry *dentry, const unsigned int isdir) ++{ ++ int err; ++ struct inode *inode; ++ ++ inode = dentry->d_inode; ++ if (IS_ROOT(dentry) ++ /* || (inode && inode->i_ino == AUFS_ROOT_INO) */ ++ ) { ++ AuWarn("branch root dir was changed\n"); ++ return 0; ++ } ++ ++ err = 0; ++ if (!isdir) { ++ d_drop(dentry); ++ au_digen_dec(dentry); ++ if (inode) ++ au_iigen_dec(inode); ++ } else { ++ au_fset_si(au_sbi(dentry->d_sb), FAILED_REFRESH_DIRS); ++ if (inode) ++ err = hin_gen_tree(dentry); ++ } ++ ++ AuTraceErr(err); ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* hinotify job flags */ ++#define AuHinJob_XINO0 1 ++#define AuHinJob_GEN (1 << 1) ++#define AuHinJob_DIRENT (1 << 2) ++#define AuHinJob_ISDIR (1 << 3) ++#define AuHinJob_TRYXINO0 (1 << 4) ++#define AuHinJob_MNTPNT (1 << 5) ++#define au_ftest_hinjob(flags, name) ((flags) & AuHinJob_##name) ++#define au_fset_hinjob(flags, name) { (flags) |= AuHinJob_##name; } ++#define au_fclr_hinjob(flags, name) { (flags) &= ~AuHinJob_##name; } ++ ++struct hin_job_args { ++ unsigned int flags; ++ struct inode *inode, *h_inode, *dir, *h_dir; ++ struct dentry *dentry; ++ char *h_name; ++ int h_nlen; ++}; ++ ++static int hin_job(struct hin_job_args *a) ++{ ++ const unsigned int isdir = au_ftest_hinjob(a->flags, ISDIR); ++ ++ /* reset xino */ ++ if (au_ftest_hinjob(a->flags, XINO0) && a->inode) ++ hin_xino(a->inode, a->h_inode); /* ignore this error */ ++ ++ if (au_ftest_hinjob(a->flags, TRYXINO0) ++ && a->inode ++ && a->h_inode) { ++ mutex_lock_nested(&a->h_inode->i_mutex, AuLsc_I_CHILD); ++ if (!a->h_inode->i_nlink) ++ hin_xino(a->inode, a->h_inode); /* ignore this error */ ++ mutex_unlock(&a->h_inode->i_mutex); ++ } ++ ++ /* make the generation obsolete */ ++ if (au_ftest_hinjob(a->flags, GEN)) { ++ int err = -1; ++ if (a->inode) ++ err = hin_gen_by_inode(a->h_name, a->h_nlen, a->inode, ++ isdir); ++ if (err && a->dentry) ++ hin_gen_by_name(a->dentry, isdir); ++ /* ignore this error */ ++ } ++ ++ /* make dir entries obsolete */ ++ if (au_ftest_hinjob(a->flags, DIRENT) && a->inode) { ++ struct au_vdir *vdir; ++ ++ vdir = au_ivdir(a->inode); ++ if (vdir) ++ vdir->vd_jiffy = 0; ++ /* IMustLock(a->inode); */ ++ /* a->inode->i_version++; */ ++ } ++ ++ /* can do nothing but warn */ ++ if (au_ftest_hinjob(a->flags, MNTPNT) ++ && a->dentry ++ && d_mountpoint(a->dentry)) ++ AuWarn("mount-point %.*s is removed or renamed\n", ++ AuDLNPair(a->dentry)); ++ ++ return 0; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static char *in_name(u32 mask) ++{ ++#ifdef CONFIG_AUFS_DEBUG ++#define test_ret(flag) if (mask & flag) \ ++ return #flag; ++ test_ret(IN_ACCESS); ++ test_ret(IN_MODIFY); ++ test_ret(IN_ATTRIB); ++ test_ret(IN_CLOSE_WRITE); ++ test_ret(IN_CLOSE_NOWRITE); ++ test_ret(IN_OPEN); ++ test_ret(IN_MOVED_FROM); ++ test_ret(IN_MOVED_TO); ++ test_ret(IN_CREATE); ++ test_ret(IN_DELETE); ++ test_ret(IN_DELETE_SELF); ++ test_ret(IN_MOVE_SELF); ++ test_ret(IN_UNMOUNT); ++ test_ret(IN_Q_OVERFLOW); ++ test_ret(IN_IGNORED); ++ return ""; ++#undef test_ret ++#else ++ return "??"; ++#endif ++} ++ ++static struct dentry *lookup_wlock_by_name(char *name, unsigned int nlen, ++ struct inode *dir) ++{ ++ struct dentry *dentry, *d, *parent; ++ struct qstr *dname; ++ ++ parent = d_find_alias(dir); ++ if (!parent) ++ return NULL; ++ ++ dentry = NULL; ++ spin_lock(&dcache_lock); ++ list_for_each_entry(d, &parent->d_subdirs, d_u.d_child) { ++ /* AuDbg("%.*s\n", AuDLNPair(d)); */ ++ dname = &d->d_name; ++ if (dname->len != nlen || memcmp(dname->name, name, nlen)) ++ continue; ++ if (!atomic_read(&d->d_count) || !d->d_fsdata) { ++ spin_lock(&d->d_lock); ++ __d_drop(d); ++ spin_unlock(&d->d_lock); ++ continue; ++ } ++ ++ dentry = dget(d); ++ break; ++ } ++ spin_unlock(&dcache_lock); ++ dput(parent); ++ ++ if (dentry) ++ di_write_lock_child(dentry); ++ ++ return dentry; ++} ++ ++static struct inode *lookup_wlock_by_ino(struct super_block *sb, ++ aufs_bindex_t bindex, ino_t h_ino) ++{ ++ struct inode *inode; ++ ino_t ino; ++ int err; ++ ++ inode = NULL; ++ err = au_xino_read(sb, bindex, h_ino, &ino); ++ if (!err && ino) ++ inode = ilookup(sb, ino); ++ if (!inode) ++ goto out; ++ ++ if (unlikely(inode->i_ino == AUFS_ROOT_INO)) { ++ AuWarn("wrong root branch\n"); ++ iput(inode); ++ inode = NULL; ++ goto out; ++ } ++ ++ ii_write_lock_child(inode); ++ ++ out: ++ return inode; ++} ++ ++enum { CHILD, PARENT }; ++struct postproc_args { ++ struct inode *h_dir, *dir, *h_child_inode; ++ u32 mask; ++ unsigned int flags[2]; ++ unsigned int h_child_nlen; ++ char h_child_name[]; ++}; ++ ++static void postproc(void *_args) ++{ ++ struct postproc_args *a = _args; ++ struct super_block *sb; ++ aufs_bindex_t bindex, bend, bfound; ++ unsigned char xino, try_iput; ++ int err; ++ struct inode *inode; ++ ino_t h_ino; ++ struct hin_job_args args; ++ struct dentry *dentry; ++ struct au_sbinfo *sbinfo; ++ ++ AuDebugOn(!_args); ++ AuDebugOn(!a->h_dir); ++ AuDebugOn(!a->dir); ++ AuDebugOn(!a->mask); ++ AuDbg("mask 0x%x %s, i%lu, hi%lu, hci%lu\n", ++ a->mask, in_name(a->mask), a->dir->i_ino, a->h_dir->i_ino, ++ a->h_child_inode ? a->h_child_inode->i_ino : 0); ++ ++ inode = NULL; ++ dentry = NULL; ++ /* ++ * do not lock a->dir->i_mutex here ++ * because of d_revalidate() may cause a deadlock. ++ */ ++ sb = a->dir->i_sb; ++ AuDebugOn(!sb); ++ sbinfo = au_sbi(sb); ++ AuDebugOn(!sbinfo); ++ /* big aufs lock */ ++ si_noflush_write_lock(sb); ++ ++ ii_read_lock_parent(a->dir); ++ bfound = -1; ++ bend = au_ibend(a->dir); ++ for (bindex = au_ibstart(a->dir); bindex <= bend; bindex++) ++ if (au_h_iptr(a->dir, bindex) == a->h_dir) { ++ bfound = bindex; ++ break; ++ } ++ ii_read_unlock(a->dir); ++ if (unlikely(bfound < 0)) ++ goto out; ++ ++ xino = !!au_opt_test(au_mntflags(sb), XINO); ++ h_ino = 0; ++ if (a->h_child_inode) ++ h_ino = a->h_child_inode->i_ino; ++ ++ if (a->h_child_nlen ++ && (au_ftest_hinjob(a->flags[CHILD], GEN) ++ || au_ftest_hinjob(a->flags[CHILD], MNTPNT))) ++ dentry = lookup_wlock_by_name(a->h_child_name, a->h_child_nlen, ++ a->dir); ++ try_iput = 0; ++ if (dentry) ++ inode = dentry->d_inode; ++ if (xino && !inode && h_ino ++ && (au_ftest_hinjob(a->flags[CHILD], XINO0) ++ || au_ftest_hinjob(a->flags[CHILD], TRYXINO0) ++ || au_ftest_hinjob(a->flags[CHILD], GEN))) { ++ inode = lookup_wlock_by_ino(sb, bfound, h_ino); ++ try_iput = 1; ++ } ++ ++ args.flags = a->flags[CHILD]; ++ args.dentry = dentry; ++ args.inode = inode; ++ args.h_inode = a->h_child_inode; ++ args.dir = a->dir; ++ args.h_dir = a->h_dir; ++ args.h_name = a->h_child_name; ++ args.h_nlen = a->h_child_nlen; ++ err = hin_job(&args); ++ if (dentry) { ++ if (dentry->d_fsdata) ++ di_write_unlock(dentry); ++ dput(dentry); ++ } ++ if (inode && try_iput) { ++ ii_write_unlock(inode); ++ iput(inode); ++ } ++ ++ ii_write_lock_parent(a->dir); ++ args.flags = a->flags[PARENT]; ++ args.dentry = NULL; ++ args.inode = a->dir; ++ args.h_inode = a->h_dir; ++ args.dir = NULL; ++ args.h_dir = NULL; ++ args.h_name = NULL; ++ args.h_nlen = 0; ++ err = hin_job(&args); ++ ii_write_unlock(a->dir); ++ ++ out: ++ au_nwt_done(&sbinfo->si_nowait); ++ si_write_unlock(sb); ++ ++ iput(a->h_child_inode); ++ iput(a->h_dir); ++ iput(a->dir); ++ kfree(a); ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static void aufs_inotify(struct inotify_watch *watch, u32 wd __maybe_unused, ++ u32 mask, u32 cookie __maybe_unused, ++ const char *h_child_name, struct inode *h_child_inode) ++{ ++ struct au_hinotify *hinotify; ++ struct postproc_args *args; ++ int len, wkq_err; ++ unsigned char isdir, isroot, wh; ++ char *p; ++ struct inode *dir; ++ unsigned int flags[2]; ++ ++ /* if IN_UNMOUNT happens, there must be another bug */ ++ AuDebugOn(mask & IN_UNMOUNT); ++ if (mask & (IN_IGNORED | IN_UNMOUNT)) { ++ put_inotify_watch(watch); ++ return; ++ } ++#ifdef AuDbgHinotify ++ au_debug(1); ++ if (1 || !h_child_name || strcmp(h_child_name, AUFS_XINO_FNAME)) { ++ AuDbg("i%lu, wd %d, mask 0x%x %s, cookie 0x%x, hcname %s," ++ " hi%lu\n", ++ watch->inode->i_ino, wd, mask, in_name(mask), cookie, ++ h_child_name ? h_child_name : "", ++ h_child_inode ? h_child_inode->i_ino : 0); ++ WARN_ON(1); ++ } ++ au_debug(0); ++#endif ++ ++ hinotify = container_of(watch, struct au_hinotify, hin_watch); ++ AuDebugOn(!hinotify || !hinotify->hin_aufs_inode); ++ dir = igrab(hinotify->hin_aufs_inode); ++ if (!dir) ++ return; ++ ++ isroot = (dir->i_ino == AUFS_ROOT_INO); ++ len = 0; ++ wh = 0; ++ if (h_child_name) { ++ len = strlen(h_child_name); ++ if (!memcmp(h_child_name, AUFS_WH_PFX, AUFS_WH_PFX_LEN)) { ++ h_child_name += AUFS_WH_PFX_LEN; ++ len -= AUFS_WH_PFX_LEN; ++ wh = 1; ++ } ++ } ++ ++ isdir = 0; ++ if (h_child_inode) ++ isdir = !!S_ISDIR(h_child_inode->i_mode); ++ flags[PARENT] = AuHinJob_ISDIR; ++ flags[CHILD] = 0; ++ if (isdir) ++ flags[CHILD] = AuHinJob_ISDIR; ++ switch (mask & IN_ALL_EVENTS) { ++ case IN_MOVED_FROM: ++ case IN_MOVED_TO: ++ AuDebugOn(!h_child_name || !h_child_inode); ++ au_fset_hinjob(flags[CHILD], GEN); ++ au_fset_hinjob(flags[CHILD], XINO0); ++ au_fset_hinjob(flags[CHILD], MNTPNT); ++ au_fset_hinjob(flags[PARENT], DIRENT); ++ break; ++ ++ case IN_CREATE: ++ AuDebugOn(!h_child_name || !h_child_inode); ++ au_fset_hinjob(flags[PARENT], DIRENT); ++ au_fset_hinjob(flags[CHILD], GEN); ++ break; ++ ++ case IN_DELETE: ++ /* ++ * aufs never be able to get this child inode. ++ * revalidation should be in d_revalidate() ++ * by checking i_nlink, i_generation or d_unhashed(). ++ */ ++ AuDebugOn(!h_child_name); ++ au_fset_hinjob(flags[PARENT], DIRENT); ++ au_fset_hinjob(flags[CHILD], GEN); ++ au_fset_hinjob(flags[CHILD], TRYXINO0); ++ au_fset_hinjob(flags[CHILD], MNTPNT); ++ break; ++ ++ default: ++ AuDebugOn(1); ++ } ++ ++ if (wh) ++ h_child_inode = NULL; ++ ++ /* iput() and kfree() will be called in postproc() */ ++ /* ++ * inotify_mutex is already acquired and kmalloc/prune_icache may lock ++ * iprune_mutex. strange. ++ */ ++ lockdep_off(); ++ args = kmalloc(sizeof(*args) + len + 1, GFP_NOFS); ++ lockdep_on(); ++ if (unlikely(!args)) { ++ AuErr1("no memory\n"); ++ iput(dir); ++ return; ++ } ++ args->flags[PARENT] = flags[PARENT]; ++ args->flags[CHILD] = flags[CHILD]; ++ args->mask = mask; ++ args->dir = dir; ++ args->h_dir = igrab(watch->inode); ++ if (h_child_inode) ++ h_child_inode = igrab(h_child_inode); /* can be NULL */ ++ args->h_child_inode = h_child_inode; ++ args->h_child_nlen = len; ++ if (len) { ++ p = (void *)args; ++ p += sizeof(*args); ++ memcpy(p, h_child_name, len + 1); ++ } ++ ++ lockdep_off(); ++ wkq_err = au_wkq_nowait(postproc, args, dir->i_sb); ++ lockdep_on(); ++ if (unlikely(wkq_err)) ++ AuErr("wkq %d\n", wkq_err); ++} ++ ++static void aufs_inotify_destroy(struct inotify_watch *watch __maybe_unused) ++{ ++ return; ++} ++ ++static struct inotify_operations aufs_inotify_ops = { ++ .handle_event = aufs_inotify, ++ .destroy_watch = aufs_inotify_destroy ++}; ++ ++/* ---------------------------------------------------------------------- */ ++ ++static void au_hin_destroy_cache(void) ++{ ++ kmem_cache_destroy(au_cachep[AuCache_HINOTIFY]); ++ au_cachep[AuCache_HINOTIFY] = NULL; ++} ++ ++int __init au_hinotify_init(void) ++{ ++ int err; ++ ++ err = -ENOMEM; ++ au_cachep[AuCache_HINOTIFY] = AuCache(au_hinotify); ++ if (au_cachep[AuCache_HINOTIFY]) { ++ err = 0; ++ au_hin_handle = inotify_init(&aufs_inotify_ops); ++ if (IS_ERR(au_hin_handle)) { ++ err = PTR_ERR(au_hin_handle); ++ au_hin_destroy_cache(); ++ } ++ } ++ AuTraceErr(err); ++ return err; ++} ++ ++void au_hinotify_fin(void) ++{ ++ inotify_destroy(au_hin_handle); ++ if (au_cachep[AuCache_HINOTIFY]) ++ au_hin_destroy_cache(); ++} +diff -Nur linux-2.6.31.4.orig/fs/aufs/iinfo.c linux-2.6.31.4/fs/aufs/iinfo.c +--- linux-2.6.31.4.orig/fs/aufs/iinfo.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/iinfo.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,283 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * inode private data ++ */ ++ ++#include "aufs.h" ++ ++struct inode *au_h_iptr(struct inode *inode, aufs_bindex_t bindex) ++{ ++ struct inode *h_inode; ++ ++ IiMustAnyLock(inode); ++ ++ h_inode = au_ii(inode)->ii_hinode[0 + bindex].hi_inode; ++ AuDebugOn(h_inode && atomic_read(&h_inode->i_count) <= 0); ++ return h_inode; ++} ++ ++/* todo: hard/soft set? */ ++void au_set_ibstart(struct inode *inode, aufs_bindex_t bindex) ++{ ++ struct au_iinfo *iinfo = au_ii(inode); ++ struct inode *h_inode; ++ ++ IiMustWriteLock(inode); ++ ++ iinfo->ii_bstart = bindex; ++ h_inode = iinfo->ii_hinode[bindex + 0].hi_inode; ++ if (h_inode) ++ au_cpup_igen(inode, h_inode); ++} ++ ++void au_hiput(struct au_hinode *hinode) ++{ ++ au_hin_free(hinode); ++ dput(hinode->hi_whdentry); ++ iput(hinode->hi_inode); ++} ++ ++unsigned int au_hi_flags(struct inode *inode, int isdir) ++{ ++ unsigned int flags; ++ const unsigned int mnt_flags = au_mntflags(inode->i_sb); ++ ++ flags = 0; ++ if (au_opt_test(mnt_flags, XINO)) ++ au_fset_hi(flags, XINO); ++ if (isdir && au_opt_test(mnt_flags, UDBA_HINOTIFY)) ++ au_fset_hi(flags, HINOTIFY); ++ return flags; ++} ++ ++void au_set_h_iptr(struct inode *inode, aufs_bindex_t bindex, ++ struct inode *h_inode, unsigned int flags) ++{ ++ struct au_hinode *hinode; ++ struct inode *hi; ++ struct au_iinfo *iinfo = au_ii(inode); ++ ++ IiMustWriteLock(inode); ++ ++ hinode = iinfo->ii_hinode + bindex; ++ hi = hinode->hi_inode; ++ AuDebugOn(h_inode && atomic_read(&h_inode->i_count) <= 0); ++ AuDebugOn(h_inode && hi); ++ ++ if (hi) ++ au_hiput(hinode); ++ hinode->hi_inode = h_inode; ++ if (h_inode) { ++ int err; ++ struct super_block *sb = inode->i_sb; ++ struct au_branch *br; ++ ++ if (bindex == iinfo->ii_bstart) ++ au_cpup_igen(inode, h_inode); ++ br = au_sbr(sb, bindex); ++ hinode->hi_id = br->br_id; ++ if (au_ftest_hi(flags, XINO)) { ++ err = au_xino_write(sb, bindex, h_inode->i_ino, ++ inode->i_ino); ++ if (unlikely(err)) ++ AuIOErr1("failed au_xino_write() %d\n", err); ++ } ++ ++ if (au_ftest_hi(flags, HINOTIFY) ++ && au_br_hinotifyable(br->br_perm)) { ++ err = au_hin_alloc(hinode, inode, h_inode); ++ if (unlikely(err)) ++ AuIOErr1("au_hin_alloc() %d\n", err); ++ } ++ } ++} ++ ++void au_set_hi_wh(struct inode *inode, aufs_bindex_t bindex, ++ struct dentry *h_wh) ++{ ++ struct au_hinode *hinode; ++ ++ IiMustWriteLock(inode); ++ ++ hinode = au_ii(inode)->ii_hinode + bindex; ++ AuDebugOn(hinode->hi_whdentry); ++ hinode->hi_whdentry = h_wh; ++} ++ ++void au_update_iigen(struct inode *inode) ++{ ++ atomic_set(&au_ii(inode)->ii_generation, au_sigen(inode->i_sb)); ++ /* smp_mb(); */ /* atomic_set */ ++} ++ ++/* it may be called at remount time, too */ ++void au_update_brange(struct inode *inode, int do_put_zero) ++{ ++ struct au_iinfo *iinfo; ++ ++ iinfo = au_ii(inode); ++ if (!iinfo || iinfo->ii_bstart < 0) ++ return; ++ ++ IiMustWriteLock(inode); ++ ++ if (do_put_zero) { ++ aufs_bindex_t bindex; ++ ++ for (bindex = iinfo->ii_bstart; bindex <= iinfo->ii_bend; ++ bindex++) { ++ struct inode *h_i; ++ ++ h_i = iinfo->ii_hinode[0 + bindex].hi_inode; ++ if (h_i && !h_i->i_nlink) ++ au_set_h_iptr(inode, bindex, NULL, 0); ++ } ++ } ++ ++ iinfo->ii_bstart = -1; ++ while (++iinfo->ii_bstart <= iinfo->ii_bend) ++ if (iinfo->ii_hinode[0 + iinfo->ii_bstart].hi_inode) ++ break; ++ if (iinfo->ii_bstart > iinfo->ii_bend) { ++ iinfo->ii_bstart = -1; ++ iinfo->ii_bend = -1; ++ return; ++ } ++ ++ iinfo->ii_bend++; ++ while (0 <= --iinfo->ii_bend) ++ if (iinfo->ii_hinode[0 + iinfo->ii_bend].hi_inode) ++ break; ++ AuDebugOn(iinfo->ii_bstart > iinfo->ii_bend || iinfo->ii_bend < 0); ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++int au_iinfo_init(struct inode *inode) ++{ ++ struct au_iinfo *iinfo; ++ struct super_block *sb; ++ int nbr, i; ++ ++ sb = inode->i_sb; ++ iinfo = &(container_of(inode, struct au_icntnr, vfs_inode)->iinfo); ++ nbr = au_sbend(sb) + 1; ++ if (unlikely(nbr <= 0)) ++ nbr = 1; ++ iinfo->ii_hinode = kcalloc(nbr, sizeof(*iinfo->ii_hinode), GFP_NOFS); ++ if (iinfo->ii_hinode) { ++ for (i = 0; i < nbr; i++) ++ iinfo->ii_hinode[i].hi_id = -1; ++ ++ atomic_set(&iinfo->ii_generation, au_sigen(sb)); ++ /* smp_mb(); */ /* atomic_set */ ++ au_rw_init(&iinfo->ii_rwsem); ++ iinfo->ii_bstart = -1; ++ iinfo->ii_bend = -1; ++ iinfo->ii_vdir = NULL; ++ return 0; ++ } ++ return -ENOMEM; ++} ++ ++int au_ii_realloc(struct au_iinfo *iinfo, int nbr) ++{ ++ int err, sz; ++ struct au_hinode *hip; ++ ++ AuRwMustWriteLock(&iinfo->ii_rwsem); ++ ++ err = -ENOMEM; ++ sz = sizeof(*hip) * (iinfo->ii_bend + 1); ++ if (!sz) ++ sz = sizeof(*hip); ++ hip = au_kzrealloc(iinfo->ii_hinode, sz, sizeof(*hip) * nbr, GFP_NOFS); ++ if (hip) { ++ iinfo->ii_hinode = hip; ++ err = 0; ++ } ++ ++ return err; ++} ++ ++static int au_iinfo_write0(struct super_block *sb, struct au_hinode *hinode, ++ ino_t ino) ++{ ++ int err; ++ aufs_bindex_t bindex; ++ unsigned char locked; ++ ++ err = 0; ++ locked = !!si_noflush_read_trylock(sb); ++ bindex = au_br_index(sb, hinode->hi_id); ++ if (bindex >= 0) ++ err = au_xino_write0(sb, bindex, hinode->hi_inode->i_ino, ino); ++ /* error action? */ ++ if (locked) ++ si_read_unlock(sb); ++ return err; ++} ++ ++void au_iinfo_fin(struct inode *inode) ++{ ++ ino_t ino; ++ aufs_bindex_t bend; ++ unsigned char unlinked = !inode->i_nlink; ++ struct au_iinfo *iinfo; ++ struct au_hinode *hi; ++ struct super_block *sb; ++ ++ if (unlinked) { ++ int err = au_xigen_inc(inode); ++ if (unlikely(err)) ++ AuWarn1("failed resetting i_generation, %d\n", err); ++ } ++ ++ iinfo = au_ii(inode); ++ /* bad_inode case */ ++ if (!iinfo) ++ return; ++ ++ if (iinfo->ii_vdir) ++ au_vdir_free(iinfo->ii_vdir); ++ ++ if (iinfo->ii_bstart >= 0) { ++ sb = inode->i_sb; ++ ino = 0; ++ if (unlinked) ++ ino = inode->i_ino; ++ hi = iinfo->ii_hinode + iinfo->ii_bstart; ++ bend = iinfo->ii_bend; ++ while (iinfo->ii_bstart++ <= bend) { ++ if (hi->hi_inode) { ++ if (unlinked || !hi->hi_inode->i_nlink) { ++ au_iinfo_write0(sb, hi, ino); ++ /* ignore this error */ ++ ino = 0; ++ } ++ au_hiput(hi); ++ } ++ hi++; ++ } ++ } ++ ++ kfree(iinfo->ii_hinode); ++ AuRwDestroy(&iinfo->ii_rwsem); ++} +diff -Nur linux-2.6.31.4.orig/fs/aufs/inode.c linux-2.6.31.4/fs/aufs/inode.c +--- linux-2.6.31.4.orig/fs/aufs/inode.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/inode.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,380 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * inode functions ++ */ ++ ++#include "aufs.h" ++ ++struct inode *au_igrab(struct inode *inode) ++{ ++ if (inode) { ++ AuDebugOn(!atomic_read(&inode->i_count)); ++ atomic_inc_return(&inode->i_count); ++ } ++ return inode; ++} ++ ++static void au_refresh_hinode_attr(struct inode *inode, int do_version) ++{ ++ au_cpup_attr_all(inode, /*force*/0); ++ au_update_iigen(inode); ++ if (do_version) ++ inode->i_version++; ++} ++ ++int au_refresh_hinode_self(struct inode *inode, int do_attr) ++{ ++ int err; ++ aufs_bindex_t bindex, new_bindex; ++ unsigned char update; ++ struct inode *first; ++ struct au_hinode *p, *q, tmp; ++ struct super_block *sb; ++ struct au_iinfo *iinfo; ++ ++ IiMustWriteLock(inode); ++ ++ update = 0; ++ sb = inode->i_sb; ++ iinfo = au_ii(inode); ++ err = au_ii_realloc(iinfo, au_sbend(sb) + 1); ++ if (unlikely(err)) ++ goto out; ++ ++ p = iinfo->ii_hinode + iinfo->ii_bstart; ++ first = p->hi_inode; ++ err = 0; ++ for (bindex = iinfo->ii_bstart; bindex <= iinfo->ii_bend; ++ bindex++, p++) { ++ if (!p->hi_inode) ++ continue; ++ ++ new_bindex = au_br_index(sb, p->hi_id); ++ if (new_bindex == bindex) ++ continue; ++ ++ if (new_bindex < 0) { ++ update++; ++ au_hiput(p); ++ p->hi_inode = NULL; ++ continue; ++ } ++ ++ if (new_bindex < iinfo->ii_bstart) ++ iinfo->ii_bstart = new_bindex; ++ if (iinfo->ii_bend < new_bindex) ++ iinfo->ii_bend = new_bindex; ++ /* swap two lower inode, and loop again */ ++ q = iinfo->ii_hinode + new_bindex; ++ tmp = *q; ++ *q = *p; ++ *p = tmp; ++ if (tmp.hi_inode) { ++ bindex--; ++ p--; ++ } ++ } ++ au_update_brange(inode, /*do_put_zero*/0); ++ if (do_attr) ++ au_refresh_hinode_attr(inode, update && S_ISDIR(inode->i_mode)); ++ ++ out: ++ return err; ++} ++ ++int au_refresh_hinode(struct inode *inode, struct dentry *dentry) ++{ ++ int err, update; ++ unsigned int flags; ++ aufs_bindex_t bindex, bend; ++ unsigned char isdir; ++ struct inode *first; ++ struct au_hinode *p; ++ struct au_iinfo *iinfo; ++ ++ err = au_refresh_hinode_self(inode, /*do_attr*/0); ++ if (unlikely(err)) ++ goto out; ++ ++ update = 0; ++ iinfo = au_ii(inode); ++ p = iinfo->ii_hinode + iinfo->ii_bstart; ++ first = p->hi_inode; ++ isdir = S_ISDIR(inode->i_mode); ++ flags = au_hi_flags(inode, isdir); ++ bend = au_dbend(dentry); ++ for (bindex = au_dbstart(dentry); bindex <= bend; bindex++) { ++ struct inode *h_i; ++ struct dentry *h_d; ++ ++ h_d = au_h_dptr(dentry, bindex); ++ if (!h_d || !h_d->d_inode) ++ continue; ++ ++ if (iinfo->ii_bstart <= bindex && bindex <= iinfo->ii_bend) { ++ h_i = au_h_iptr(inode, bindex); ++ if (h_i) { ++ if (h_i == h_d->d_inode) ++ continue; ++ err = -EIO; ++ break; ++ } ++ } ++ if (bindex < iinfo->ii_bstart) ++ iinfo->ii_bstart = bindex; ++ if (iinfo->ii_bend < bindex) ++ iinfo->ii_bend = bindex; ++ au_set_h_iptr(inode, bindex, au_igrab(h_d->d_inode), flags); ++ update = 1; ++ } ++ au_update_brange(inode, /*do_put_zero*/0); ++ ++ if (unlikely(err)) ++ goto out; ++ ++ au_refresh_hinode_attr(inode, update && isdir); ++ ++ out: ++ return err; ++} ++ ++static int set_inode(struct inode *inode, struct dentry *dentry) ++{ ++ int err; ++ unsigned int flags; ++ umode_t mode; ++ aufs_bindex_t bindex, bstart, btail; ++ unsigned char isdir; ++ struct dentry *h_dentry; ++ struct inode *h_inode; ++ struct au_iinfo *iinfo; ++ ++ IiMustWriteLock(inode); ++ ++ err = 0; ++ isdir = 0; ++ bstart = au_dbstart(dentry); ++ h_inode = au_h_dptr(dentry, bstart)->d_inode; ++ mode = h_inode->i_mode; ++ switch (mode & S_IFMT) { ++ case S_IFREG: ++ btail = au_dbtail(dentry); ++ inode->i_op = &aufs_iop; ++ inode->i_fop = &aufs_file_fop; ++ inode->i_mapping->a_ops = &aufs_aop; ++ break; ++ case S_IFDIR: ++ isdir = 1; ++ btail = au_dbtaildir(dentry); ++ inode->i_op = &aufs_dir_iop; ++ inode->i_fop = &aufs_dir_fop; ++ break; ++ case S_IFLNK: ++ btail = au_dbtail(dentry); ++ inode->i_op = &aufs_symlink_iop; ++ break; ++ case S_IFBLK: ++ case S_IFCHR: ++ case S_IFIFO: ++ case S_IFSOCK: ++ btail = au_dbtail(dentry); ++ inode->i_op = &aufs_iop; ++ init_special_inode(inode, mode, h_inode->i_rdev); ++ break; ++ default: ++ AuIOErr("Unknown file type 0%o\n", mode); ++ err = -EIO; ++ goto out; ++ } ++ ++ /* do not set inotify for whiteouted dirs (SHWH mode) */ ++ flags = au_hi_flags(inode, isdir); ++ if (au_opt_test(au_mntflags(dentry->d_sb), SHWH) ++ && au_ftest_hi(flags, HINOTIFY) ++ && dentry->d_name.len > AUFS_WH_PFX_LEN ++ && !memcmp(dentry->d_name.name, AUFS_WH_PFX, AUFS_WH_PFX_LEN)) ++ au_fclr_hi(flags, HINOTIFY); ++ iinfo = au_ii(inode); ++ iinfo->ii_bstart = bstart; ++ iinfo->ii_bend = btail; ++ for (bindex = bstart; bindex <= btail; bindex++) { ++ h_dentry = au_h_dptr(dentry, bindex); ++ if (h_dentry) ++ au_set_h_iptr(inode, bindex, ++ au_igrab(h_dentry->d_inode), flags); ++ } ++ au_cpup_attr_all(inode, /*force*/1); ++ ++ out: ++ return err; ++} ++ ++/* successful returns with iinfo write_locked */ ++static int reval_inode(struct inode *inode, struct dentry *dentry, int *matched) ++{ ++ int err; ++ aufs_bindex_t bindex, bend; ++ struct inode *h_inode, *h_dinode; ++ ++ *matched = 0; ++ ++ /* ++ * before this function, if aufs got any iinfo lock, it must be only ++ * one, the parent dir. ++ * it can happen by UDBA and the obsoleted inode number. ++ */ ++ err = -EIO; ++ if (unlikely(inode->i_ino == parent_ino(dentry))) ++ goto out; ++ ++ err = 0; ++ ii_write_lock_new_child(inode); ++ h_dinode = au_h_dptr(dentry, au_dbstart(dentry))->d_inode; ++ bend = au_ibend(inode); ++ for (bindex = au_ibstart(inode); bindex <= bend; bindex++) { ++ h_inode = au_h_iptr(inode, bindex); ++ if (h_inode && h_inode == h_dinode) { ++ *matched = 1; ++ err = 0; ++ if (au_iigen(inode) != au_digen(dentry)) ++ err = au_refresh_hinode(inode, dentry); ++ break; ++ } ++ } ++ ++ if (unlikely(err)) ++ ii_write_unlock(inode); ++ out: ++ return err; ++} ++ ++/* successful returns with iinfo write_locked */ ++/* todo: return with unlocked? */ ++struct inode *au_new_inode(struct dentry *dentry, int must_new) ++{ ++ struct inode *inode; ++ struct dentry *h_dentry; ++ struct super_block *sb; ++ ino_t h_ino, ino; ++ int err, match; ++ aufs_bindex_t bstart; ++ ++ sb = dentry->d_sb; ++ bstart = au_dbstart(dentry); ++ h_dentry = au_h_dptr(dentry, bstart); ++ h_ino = h_dentry->d_inode->i_ino; ++ err = au_xino_read(sb, bstart, h_ino, &ino); ++ inode = ERR_PTR(err); ++ if (unlikely(err)) ++ goto out; ++ new_ino: ++ if (!ino) { ++ ino = au_xino_new_ino(sb); ++ if (unlikely(!ino)) { ++ inode = ERR_PTR(-EIO); ++ goto out; ++ } ++ } ++ ++ AuDbg("i%lu\n", (unsigned long)ino); ++ inode = au_iget_locked(sb, ino); ++ err = PTR_ERR(inode); ++ if (IS_ERR(inode)) ++ goto out; ++ ++ AuDbg("%lx, new %d\n", inode->i_state, !!(inode->i_state & I_NEW)); ++ if (inode->i_state & I_NEW) { ++ ii_write_lock_new_child(inode); ++ err = set_inode(inode, dentry); ++ unlock_new_inode(inode); ++ if (!err) ++ goto out; /* success */ ++ ++ iget_failed(inode); ++ ii_write_unlock(inode); ++ goto out_iput; ++ } else if (!must_new) { ++ err = reval_inode(inode, dentry, &match); ++ if (!err) ++ goto out; /* success */ ++ else if (match) ++ goto out_iput; ++ } ++ ++ if (unlikely(au_test_fs_unique_ino(h_dentry->d_inode))) ++ AuWarn1("Warning: Un-notified UDBA or repeatedly renamed dir," ++ " b%d, %s, %.*s, hi%lu, i%lu.\n", ++ bstart, au_sbtype(h_dentry->d_sb), AuDLNPair(dentry), ++ (unsigned long)h_ino, (unsigned long)ino); ++ ino = 0; ++ err = au_xino_write(sb, bstart, h_ino, /*ino*/0); ++ if (!err) { ++ iput(inode); ++ goto new_ino; ++ } ++ ++ out_iput: ++ iput(inode); ++ inode = ERR_PTR(err); ++ out: ++ return inode; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++int au_test_ro(struct super_block *sb, aufs_bindex_t bindex, ++ struct inode *inode) ++{ ++ int err; ++ ++ err = au_br_rdonly(au_sbr(sb, bindex)); ++ ++ /* pseudo-link after flushed may happen out of bounds */ ++ if (!err ++ && inode ++ && au_ibstart(inode) <= bindex ++ && bindex <= au_ibend(inode)) { ++ /* ++ * permission check is unnecessary since vfsub routine ++ * will be called later ++ */ ++ struct inode *hi = au_h_iptr(inode, bindex); ++ if (hi) ++ err = IS_IMMUTABLE(hi) ? -EROFS : 0; ++ } ++ ++ return err; ++} ++ ++int au_test_h_perm(struct inode *h_inode, int mask) ++{ ++ if (!current_fsuid()) ++ return 0; ++ return inode_permission(h_inode, mask); ++} ++ ++int au_test_h_perm_sio(struct inode *h_inode, int mask) ++{ ++ if (au_test_nfs(h_inode->i_sb) ++ && (mask & MAY_WRITE) ++ && S_ISDIR(h_inode->i_mode)) ++ mask |= MAY_READ; /* force permission check */ ++ return au_test_h_perm(h_inode, mask); ++} +diff -Nur linux-2.6.31.4.orig/fs/aufs/inode.h linux-2.6.31.4/fs/aufs/inode.h +--- linux-2.6.31.4.orig/fs/aufs/inode.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/inode.h 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,484 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * inode operations ++ */ ++ ++#ifndef __AUFS_INODE_H__ ++#define __AUFS_INODE_H__ ++ ++#ifdef __KERNEL__ ++ ++#include <linux/fs.h> ++#include <linux/inotify.h> ++#include <linux/aufs_type.h> ++#include "rwsem.h" ++ ++struct vfsmount; ++ ++struct au_hinotify { ++#ifdef CONFIG_AUFS_HINOTIFY ++ struct inotify_watch hin_watch; ++ struct inode *hin_aufs_inode; /* no get/put */ ++#endif ++}; ++ ++struct au_hinode { ++ struct inode *hi_inode; ++ aufs_bindex_t hi_id; ++#ifdef CONFIG_AUFS_HINOTIFY ++ struct au_hinotify *hi_notify; ++#endif ++ ++ /* reference to the copied-up whiteout with get/put */ ++ struct dentry *hi_whdentry; ++}; ++ ++struct au_vdir; ++struct au_iinfo { ++ atomic_t ii_generation; ++ struct super_block *ii_hsb1; /* no get/put */ ++ ++ struct au_rwsem ii_rwsem; ++ aufs_bindex_t ii_bstart, ii_bend; ++ __u32 ii_higen; ++ struct au_hinode *ii_hinode; ++ struct au_vdir *ii_vdir; ++}; ++ ++struct au_icntnr { ++ struct au_iinfo iinfo; ++ struct inode vfs_inode; ++}; ++ ++/* au_pin flags */ ++#define AuPin_DI_LOCKED 1 ++#define AuPin_MNT_WRITE (1 << 1) ++#define au_ftest_pin(flags, name) ((flags) & AuPin_##name) ++#define au_fset_pin(flags, name) { (flags) |= AuPin_##name; } ++#define au_fclr_pin(flags, name) { (flags) &= ~AuPin_##name; } ++ ++struct au_pin { ++ /* input */ ++ struct dentry *dentry; ++ unsigned int udba; ++ unsigned char lsc_di, lsc_hi, flags; ++ aufs_bindex_t bindex; ++ ++ /* output */ ++ struct dentry *parent; ++ struct au_hinode *hdir; ++ struct vfsmount *h_mnt; ++}; ++ ++/* ---------------------------------------------------------------------- */ ++ ++static inline struct au_iinfo *au_ii(struct inode *inode) ++{ ++ struct au_iinfo *iinfo; ++ ++ iinfo = &(container_of(inode, struct au_icntnr, vfs_inode)->iinfo); ++ if (iinfo->ii_hinode) ++ return iinfo; ++ return NULL; /* debugging bad_inode case */ ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* inode.c */ ++struct inode *au_igrab(struct inode *inode); ++int au_refresh_hinode_self(struct inode *inode, int do_attr); ++int au_refresh_hinode(struct inode *inode, struct dentry *dentry); ++struct inode *au_new_inode(struct dentry *dentry, int must_new); ++int au_test_ro(struct super_block *sb, aufs_bindex_t bindex, ++ struct inode *inode); ++int au_test_h_perm(struct inode *h_inode, int mask); ++int au_test_h_perm_sio(struct inode *h_inode, int mask); ++ ++/* i_op.c */ ++extern struct inode_operations aufs_iop, aufs_symlink_iop, aufs_dir_iop; ++ ++/* au_wr_dir flags */ ++#define AuWrDir_ADD_ENTRY 1 ++#define AuWrDir_ISDIR (1 << 1) ++#define au_ftest_wrdir(flags, name) ((flags) & AuWrDir_##name) ++#define au_fset_wrdir(flags, name) { (flags) |= AuWrDir_##name; } ++#define au_fclr_wrdir(flags, name) { (flags) &= ~AuWrDir_##name; } ++ ++struct au_wr_dir_args { ++ aufs_bindex_t force_btgt; ++ unsigned char flags; ++}; ++int au_wr_dir(struct dentry *dentry, struct dentry *src_dentry, ++ struct au_wr_dir_args *args); ++ ++struct dentry *au_pinned_h_parent(struct au_pin *pin); ++void au_pin_init(struct au_pin *pin, struct dentry *dentry, ++ aufs_bindex_t bindex, int lsc_di, int lsc_hi, ++ unsigned int udba, unsigned char flags); ++int au_pin(struct au_pin *pin, struct dentry *dentry, aufs_bindex_t bindex, ++ unsigned int udba, unsigned char flags) __must_check; ++int au_do_pin(struct au_pin *pin) __must_check; ++void au_unpin(struct au_pin *pin); ++ ++/* i_op_add.c */ ++int au_may_add(struct dentry *dentry, aufs_bindex_t bindex, ++ struct dentry *h_parent, int isdir); ++int aufs_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t dev); ++int aufs_symlink(struct inode *dir, struct dentry *dentry, const char *symname); ++int aufs_create(struct inode *dir, struct dentry *dentry, int mode, ++ struct nameidata *nd); ++int aufs_link(struct dentry *src_dentry, struct inode *dir, ++ struct dentry *dentry); ++int aufs_mkdir(struct inode *dir, struct dentry *dentry, int mode); ++ ++/* i_op_del.c */ ++int au_wr_dir_need_wh(struct dentry *dentry, int isdir, aufs_bindex_t *bcpup); ++int au_may_del(struct dentry *dentry, aufs_bindex_t bindex, ++ struct dentry *h_parent, int isdir); ++int aufs_unlink(struct inode *dir, struct dentry *dentry); ++int aufs_rmdir(struct inode *dir, struct dentry *dentry); ++ ++/* i_op_ren.c */ ++int au_wbr(struct dentry *dentry, aufs_bindex_t btgt); ++int aufs_rename(struct inode *src_dir, struct dentry *src_dentry, ++ struct inode *dir, struct dentry *dentry); ++ ++/* iinfo.c */ ++struct inode *au_h_iptr(struct inode *inode, aufs_bindex_t bindex); ++void au_hiput(struct au_hinode *hinode); ++void au_set_ibstart(struct inode *inode, aufs_bindex_t bindex); ++void au_set_hi_wh(struct inode *inode, aufs_bindex_t bindex, ++ struct dentry *h_wh); ++unsigned int au_hi_flags(struct inode *inode, int isdir); ++ ++/* hinode flags */ ++#define AuHi_XINO 1 ++#define AuHi_HINOTIFY (1 << 1) ++#define au_ftest_hi(flags, name) ((flags) & AuHi_##name) ++#define au_fset_hi(flags, name) { (flags) |= AuHi_##name; } ++#define au_fclr_hi(flags, name) { (flags) &= ~AuHi_##name; } ++ ++#ifndef CONFIG_AUFS_HINOTIFY ++#undef AuHi_HINOTIFY ++#define AuHi_HINOTIFY 0 ++#endif ++ ++void au_set_h_iptr(struct inode *inode, aufs_bindex_t bindex, ++ struct inode *h_inode, unsigned int flags); ++ ++void au_update_iigen(struct inode *inode); ++void au_update_brange(struct inode *inode, int do_put_zero); ++ ++int au_iinfo_init(struct inode *inode); ++void au_iinfo_fin(struct inode *inode); ++int au_ii_realloc(struct au_iinfo *iinfo, int nbr); ++ ++/* plink.c */ ++void au_plink_block_maintain(struct super_block *sb); ++#ifdef CONFIG_AUFS_DEBUG ++void au_plink_list(struct super_block *sb); ++#else ++static inline void au_plink_list(struct super_block *sb) ++{ ++ /* nothing */ ++} ++#endif ++int au_plink_test(struct inode *inode); ++struct dentry *au_plink_lkup(struct inode *inode, aufs_bindex_t bindex); ++void au_plink_append(struct inode *inode, aufs_bindex_t bindex, ++ struct dentry *h_dentry); ++void au_plink_put(struct super_block *sb); ++void au_plink_half_refresh(struct super_block *sb, aufs_bindex_t br_id); ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* lock subclass for iinfo */ ++enum { ++ AuLsc_II_CHILD, /* child first */ ++ AuLsc_II_CHILD2, /* rename(2), link(2), and cpup at hinotify */ ++ AuLsc_II_CHILD3, /* copyup dirs */ ++ AuLsc_II_PARENT, /* see AuLsc_I_PARENT in vfsub.h */ ++ AuLsc_II_PARENT2, ++ AuLsc_II_PARENT3, /* copyup dirs */ ++ AuLsc_II_NEW_CHILD ++}; ++ ++/* ++ * ii_read_lock_child, ii_write_lock_child, ++ * ii_read_lock_child2, ii_write_lock_child2, ++ * ii_read_lock_child3, ii_write_lock_child3, ++ * ii_read_lock_parent, ii_write_lock_parent, ++ * ii_read_lock_parent2, ii_write_lock_parent2, ++ * ii_read_lock_parent3, ii_write_lock_parent3, ++ * ii_read_lock_new_child, ii_write_lock_new_child, ++ */ ++#define AuReadLockFunc(name, lsc) \ ++static inline void ii_read_lock_##name(struct inode *i) \ ++{ \ ++ au_rw_read_lock_nested(&au_ii(i)->ii_rwsem, AuLsc_II_##lsc); \ ++} ++ ++#define AuWriteLockFunc(name, lsc) \ ++static inline void ii_write_lock_##name(struct inode *i) \ ++{ \ ++ au_rw_write_lock_nested(&au_ii(i)->ii_rwsem, AuLsc_II_##lsc); \ ++} ++ ++#define AuRWLockFuncs(name, lsc) \ ++ AuReadLockFunc(name, lsc) \ ++ AuWriteLockFunc(name, lsc) ++ ++AuRWLockFuncs(child, CHILD); ++AuRWLockFuncs(child2, CHILD2); ++AuRWLockFuncs(child3, CHILD3); ++AuRWLockFuncs(parent, PARENT); ++AuRWLockFuncs(parent2, PARENT2); ++AuRWLockFuncs(parent3, PARENT3); ++AuRWLockFuncs(new_child, NEW_CHILD); ++ ++#undef AuReadLockFunc ++#undef AuWriteLockFunc ++#undef AuRWLockFuncs ++ ++/* ++ * ii_read_unlock, ii_write_unlock, ii_downgrade_lock ++ */ ++AuSimpleUnlockRwsemFuncs(ii, struct inode *i, &au_ii(i)->ii_rwsem); ++ ++#define IiMustNoWaiters(i) AuRwMustNoWaiters(&au_ii(i)->ii_rwsem) ++#define IiMustAnyLock(i) AuRwMustAnyLock(&au_ii(i)->ii_rwsem) ++#define IiMustWriteLock(i) AuRwMustWriteLock(&au_ii(i)->ii_rwsem) ++ ++/* ---------------------------------------------------------------------- */ ++ ++static inline unsigned int au_iigen(struct inode *inode) ++{ ++ return atomic_read(&au_ii(inode)->ii_generation); ++} ++ ++/* tiny test for inode number */ ++/* tmpfs generation is too rough */ ++static inline int au_test_higen(struct inode *inode, struct inode *h_inode) ++{ ++ struct au_iinfo *iinfo; ++ ++ iinfo = au_ii(inode); ++ AuRwMustAnyLock(&iinfo->ii_rwsem); ++ return !(iinfo->ii_hsb1 == h_inode->i_sb ++ && iinfo->ii_higen == h_inode->i_generation); ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static inline aufs_bindex_t au_ii_br_id(struct inode *inode, ++ aufs_bindex_t bindex) ++{ ++ IiMustAnyLock(inode); ++ return au_ii(inode)->ii_hinode[0 + bindex].hi_id; ++} ++ ++static inline aufs_bindex_t au_ibstart(struct inode *inode) ++{ ++ IiMustAnyLock(inode); ++ return au_ii(inode)->ii_bstart; ++} ++ ++static inline aufs_bindex_t au_ibend(struct inode *inode) ++{ ++ IiMustAnyLock(inode); ++ return au_ii(inode)->ii_bend; ++} ++ ++static inline struct au_vdir *au_ivdir(struct inode *inode) ++{ ++ IiMustAnyLock(inode); ++ return au_ii(inode)->ii_vdir; ++} ++ ++static inline struct dentry *au_hi_wh(struct inode *inode, aufs_bindex_t bindex) ++{ ++ IiMustAnyLock(inode); ++ return au_ii(inode)->ii_hinode[0 + bindex].hi_whdentry; ++} ++ ++static inline void au_set_ibend(struct inode *inode, aufs_bindex_t bindex) ++{ ++ IiMustWriteLock(inode); ++ au_ii(inode)->ii_bend = bindex; ++} ++ ++static inline void au_set_ivdir(struct inode *inode, struct au_vdir *vdir) ++{ ++ IiMustWriteLock(inode); ++ au_ii(inode)->ii_vdir = vdir; ++} ++ ++static inline struct au_hinode *au_hi(struct inode *inode, aufs_bindex_t bindex) ++{ ++ IiMustAnyLock(inode); ++ return au_ii(inode)->ii_hinode + bindex; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static inline struct dentry *au_pinned_parent(struct au_pin *pin) ++{ ++ if (pin) ++ return pin->parent; ++ return NULL; ++} ++ ++static inline struct inode *au_pinned_h_dir(struct au_pin *pin) ++{ ++ if (pin && pin->hdir) ++ return pin->hdir->hi_inode; ++ return NULL; ++} ++ ++static inline struct au_hinode *au_pinned_hdir(struct au_pin *pin) ++{ ++ if (pin) ++ return pin->hdir; ++ return NULL; ++} ++ ++static inline void au_pin_set_dentry(struct au_pin *pin, struct dentry *dentry) ++{ ++ if (pin) ++ pin->dentry = dentry; ++} ++ ++static inline void au_pin_set_parent_lflag(struct au_pin *pin, ++ unsigned char lflag) ++{ ++ if (pin) { ++ /* dirty macros require brackets */ ++ if (lflag) { ++ au_fset_pin(pin->flags, DI_LOCKED); ++ } else { ++ au_fclr_pin(pin->flags, DI_LOCKED); ++ } ++ } ++} ++ ++static inline void au_pin_set_parent(struct au_pin *pin, struct dentry *parent) ++{ ++ if (pin) { ++ dput(pin->parent); ++ pin->parent = dget(parent); ++ } ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++#ifdef CONFIG_AUFS_HINOTIFY ++/* hinotify.c */ ++int au_hin_alloc(struct au_hinode *hinode, struct inode *inode, ++ struct inode *h_inode); ++void au_hin_free(struct au_hinode *hinode); ++void au_hin_ctl(struct au_hinode *hinode, int do_set); ++void au_reset_hinotify(struct inode *inode, unsigned int flags); ++ ++int __init au_hinotify_init(void); ++void au_hinotify_fin(void); ++ ++static inline ++void au_hin_init(struct au_hinode *hinode, struct au_hinotify *val) ++{ ++ hinode->hi_notify = val; ++} ++ ++static inline void au_iigen_dec(struct inode *inode) ++{ ++ atomic_dec_return(&au_ii(inode)->ii_generation); ++} ++ ++#else ++static inline ++int au_hin_alloc(struct au_hinode *hinode __maybe_unused, ++ struct inode *inode __maybe_unused, ++ struct inode *h_inode __maybe_unused) ++{ ++ return -EOPNOTSUPP; ++} ++ ++static inline void au_hin_free(struct au_hinode *hinode __maybe_unused) ++{ ++ /* nothing */ ++} ++ ++static inline void au_hin_ctl(struct au_hinode *hinode __maybe_unused, ++ int do_set __maybe_unused) ++{ ++ /* nothing */ ++} ++ ++static inline void au_reset_hinotify(struct inode *inode __maybe_unused, ++ unsigned int flags __maybe_unused) ++{ ++ /* nothing */ ++} ++ ++static inline int au_hinotify_init(void) ++{ ++ return 0; ++} ++ ++#define au_hinotify_fin() do {} while (0) ++ ++static inline ++void au_hin_init(struct au_hinode *hinode __maybe_unused, ++ struct au_hinotify *val __maybe_unused) ++{ ++ /* empty */ ++} ++#endif /* CONFIG_AUFS_HINOTIFY */ ++ ++static inline void au_hin_suspend(struct au_hinode *hdir) ++{ ++ au_hin_ctl(hdir, /*do_set*/0); ++} ++ ++static inline void au_hin_resume(struct au_hinode *hdir) ++{ ++ au_hin_ctl(hdir, /*do_set*/1); ++} ++ ++static inline void au_hin_imtx_lock(struct au_hinode *hdir) ++{ ++ mutex_lock(&hdir->hi_inode->i_mutex); ++ au_hin_suspend(hdir); ++} ++ ++static inline void au_hin_imtx_lock_nested(struct au_hinode *hdir, ++ unsigned int sc __maybe_unused) ++{ ++ mutex_lock_nested(&hdir->hi_inode->i_mutex, sc); ++ au_hin_suspend(hdir); ++} ++ ++static inline void au_hin_imtx_unlock(struct au_hinode *hdir) ++{ ++ au_hin_resume(hdir); ++ mutex_unlock(&hdir->hi_inode->i_mutex); ++} ++ ++#endif /* __KERNEL__ */ ++#endif /* __AUFS_INODE_H__ */ +diff -Nur linux-2.6.31.4.orig/fs/aufs/ioctl.c linux-2.6.31.4/fs/aufs/ioctl.c +--- linux-2.6.31.4.orig/fs/aufs/ioctl.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/ioctl.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,67 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * ioctl ++ * currently plink-management only. ++ */ ++ ++#include <linux/uaccess.h> ++#include "aufs.h" ++ ++long aufs_ioctl_dir(struct file *file, unsigned int cmd, ++ unsigned long arg __maybe_unused) ++{ ++ long err; ++ struct super_block *sb; ++ struct au_sbinfo *sbinfo; ++ ++ err = -EACCES; ++ if (!capable(CAP_SYS_ADMIN)) ++ goto out; ++ ++ err = 0; ++ sb = file->f_dentry->d_sb; ++ sbinfo = au_sbi(sb); ++ switch (cmd) { ++ case AUFS_CTL_PLINK_MAINT: ++ /* ++ * pseudo-link maintenance mode, ++ * cleared by aufs_release_dir() ++ */ ++ si_write_lock(sb); ++ if (!au_ftest_si(sbinfo, MAINTAIN_PLINK)) { ++ au_fset_si(sbinfo, MAINTAIN_PLINK); ++ au_fi(file)->fi_maintain_plink = 1; ++ } else ++ err = -EBUSY; ++ si_write_unlock(sb); ++ break; ++ case AUFS_CTL_PLINK_CLEAN: ++ aufs_write_lock(sb->s_root); ++ if (au_opt_test(sbinfo->si_mntflags, PLINK)) ++ au_plink_put(sb); ++ aufs_write_unlock(sb->s_root); ++ break; ++ default: ++ err = -EINVAL; ++ } ++ ++ out: ++ return err; ++} +diff -Nur linux-2.6.31.4.orig/fs/aufs/i_op_add.c linux-2.6.31.4/fs/aufs/i_op_add.c +--- linux-2.6.31.4.orig/fs/aufs/i_op_add.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/i_op_add.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,649 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * inode operations (add entry) ++ */ ++ ++#include "aufs.h" ++ ++/* ++ * final procedure of adding a new entry, except link(2). ++ * remove whiteout, instantiate, copyup the parent dir's times and size ++ * and update version. ++ * if it failed, re-create the removed whiteout. ++ */ ++static int epilog(struct inode *dir, aufs_bindex_t bindex, ++ struct dentry *wh_dentry, struct dentry *dentry) ++{ ++ int err, rerr; ++ aufs_bindex_t bwh; ++ struct path h_path; ++ struct inode *inode, *h_dir; ++ struct dentry *wh; ++ ++ bwh = -1; ++ if (wh_dentry) { ++ h_dir = wh_dentry->d_parent->d_inode; /* dir inode is locked */ ++ IMustLock(h_dir); ++ AuDebugOn(au_h_iptr(dir, bindex) != h_dir); ++ bwh = au_dbwh(dentry); ++ h_path.dentry = wh_dentry; ++ h_path.mnt = au_sbr_mnt(dir->i_sb, bindex); ++ err = au_wh_unlink_dentry(au_h_iptr(dir, bindex), &h_path, ++ dentry); ++ if (unlikely(err)) ++ goto out; ++ } ++ ++ inode = au_new_inode(dentry, /*must_new*/1); ++ if (!IS_ERR(inode)) { ++ d_instantiate(dentry, inode); ++ dir = dentry->d_parent->d_inode; /* dir inode is locked */ ++ IMustLock(dir); ++ if (au_ibstart(dir) == au_dbstart(dentry)) ++ au_cpup_attr_timesizes(dir); ++ dir->i_version++; ++ return 0; /* success */ ++ } ++ ++ err = PTR_ERR(inode); ++ if (!wh_dentry) ++ goto out; ++ ++ /* revert */ ++ /* dir inode is locked */ ++ wh = au_wh_create(dentry, bwh, wh_dentry->d_parent); ++ rerr = PTR_ERR(wh); ++ if (IS_ERR(wh)) { ++ AuIOErr("%.*s reverting whiteout failed(%d, %d)\n", ++ AuDLNPair(dentry), err, rerr); ++ err = -EIO; ++ } else ++ dput(wh); ++ ++ out: ++ return err; ++} ++ ++/* ++ * simple tests for the adding inode operations. ++ * following the checks in vfs, plus the parent-child relationship. ++ */ ++int au_may_add(struct dentry *dentry, aufs_bindex_t bindex, ++ struct dentry *h_parent, int isdir) ++{ ++ int err; ++ umode_t h_mode; ++ struct dentry *h_dentry; ++ struct inode *h_inode; ++ ++ h_dentry = au_h_dptr(dentry, bindex); ++ h_inode = h_dentry->d_inode; ++ if (!dentry->d_inode) { ++ err = -EEXIST; ++ if (unlikely(h_inode)) ++ goto out; ++ } else { ++ /* rename(2) case */ ++ err = -EIO; ++ if (unlikely(!h_inode || !h_inode->i_nlink)) ++ goto out; ++ ++ h_mode = h_inode->i_mode; ++ if (!isdir) { ++ err = -EISDIR; ++ if (unlikely(S_ISDIR(h_mode))) ++ goto out; ++ } else if (unlikely(!S_ISDIR(h_mode))) { ++ err = -ENOTDIR; ++ goto out; ++ } ++ } ++ ++ err = -EIO; ++ /* expected parent dir is locked */ ++ if (unlikely(h_parent != h_dentry->d_parent)) ++ goto out; ++ err = 0; ++ ++ out: ++ return err; ++} ++ ++/* ++ * initial procedure of adding a new entry. ++ * prepare writable branch and the parent dir, lock it, ++ * and lookup whiteout for the new entry. ++ */ ++static struct dentry* ++lock_hdir_lkup_wh(struct dentry *dentry, struct au_dtime *dt, ++ struct dentry *src_dentry, struct au_pin *pin, ++ struct au_wr_dir_args *wr_dir_args) ++{ ++ struct dentry *wh_dentry, *h_parent; ++ struct super_block *sb; ++ struct au_branch *br; ++ int err; ++ unsigned int udba; ++ aufs_bindex_t bcpup; ++ ++ err = au_wr_dir(dentry, src_dentry, wr_dir_args); ++ bcpup = err; ++ wh_dentry = ERR_PTR(err); ++ if (unlikely(err < 0)) ++ goto out; ++ ++ sb = dentry->d_sb; ++ udba = au_opt_udba(sb); ++ err = au_pin(pin, dentry, bcpup, udba, ++ AuPin_DI_LOCKED | AuPin_MNT_WRITE); ++ wh_dentry = ERR_PTR(err); ++ if (unlikely(err)) ++ goto out; ++ ++ h_parent = au_pinned_h_parent(pin); ++ if (udba != AuOpt_UDBA_NONE ++ && au_dbstart(dentry) == bcpup) { ++ err = au_may_add(dentry, bcpup, h_parent, ++ au_ftest_wrdir(wr_dir_args->flags, ISDIR)); ++ wh_dentry = ERR_PTR(err); ++ if (unlikely(err)) ++ goto out_unpin; ++ } ++ ++ br = au_sbr(sb, bcpup); ++ if (dt) { ++ struct path tmp = { ++ .dentry = h_parent, ++ .mnt = br->br_mnt ++ }; ++ au_dtime_store(dt, au_pinned_parent(pin), &tmp); ++ } ++ ++ wh_dentry = NULL; ++ if (bcpup != au_dbwh(dentry)) ++ goto out; /* success */ ++ ++ wh_dentry = au_wh_lkup(h_parent, &dentry->d_name, br); ++ ++ out_unpin: ++ if (IS_ERR(wh_dentry)) ++ au_unpin(pin); ++ out: ++ return wh_dentry; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++enum { Mknod, Symlink, Creat }; ++struct simple_arg { ++ int type; ++ union { ++ struct { ++ int mode; ++ struct nameidata *nd; ++ } c; ++ struct { ++ const char *symname; ++ } s; ++ struct { ++ int mode; ++ dev_t dev; ++ } m; ++ } u; ++}; ++ ++static int add_simple(struct inode *dir, struct dentry *dentry, ++ struct simple_arg *arg) ++{ ++ int err; ++ aufs_bindex_t bstart; ++ unsigned char created; ++ struct au_dtime dt; ++ struct au_pin pin; ++ struct path h_path; ++ struct dentry *wh_dentry, *parent; ++ struct inode *h_dir; ++ struct au_wr_dir_args wr_dir_args = { ++ .force_btgt = -1, ++ .flags = AuWrDir_ADD_ENTRY ++ }; ++ ++ IMustLock(dir); ++ ++ parent = dentry->d_parent; /* dir inode is locked */ ++ aufs_read_lock(dentry, AuLock_DW); ++ di_write_lock_parent(parent); ++ wh_dentry = lock_hdir_lkup_wh(dentry, &dt, /*src_dentry*/NULL, &pin, ++ &wr_dir_args); ++ err = PTR_ERR(wh_dentry); ++ if (IS_ERR(wh_dentry)) ++ goto out; ++ ++ bstart = au_dbstart(dentry); ++ h_path.dentry = au_h_dptr(dentry, bstart); ++ h_path.mnt = au_sbr_mnt(dentry->d_sb, bstart); ++ h_dir = au_pinned_h_dir(&pin); ++ switch (arg->type) { ++ case Creat: ++ err = vfsub_create(h_dir, &h_path, arg->u.c.mode); ++ break; ++ case Symlink: ++ err = vfsub_symlink(h_dir, &h_path, arg->u.s.symname); ++ break; ++ case Mknod: ++ err = vfsub_mknod(h_dir, &h_path, arg->u.m.mode, arg->u.m.dev); ++ break; ++ default: ++ BUG(); ++ } ++ created = !err; ++ if (!err) ++ err = epilog(dir, bstart, wh_dentry, dentry); ++ ++ /* revert */ ++ if (unlikely(created && err && h_path.dentry->d_inode)) { ++ int rerr; ++ rerr = vfsub_unlink(h_dir, &h_path, /*force*/0); ++ if (rerr) { ++ AuIOErr("%.*s revert failure(%d, %d)\n", ++ AuDLNPair(dentry), err, rerr); ++ err = -EIO; ++ } ++ au_dtime_revert(&dt); ++ d_drop(dentry); ++ } ++ ++ au_unpin(&pin); ++ dput(wh_dentry); ++ ++ out: ++ if (unlikely(err)) { ++ au_update_dbstart(dentry); ++ d_drop(dentry); ++ } ++ di_write_unlock(parent); ++ aufs_read_unlock(dentry, AuLock_DW); ++ return err; ++} ++ ++int aufs_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t dev) ++{ ++ struct simple_arg arg = { ++ .type = Mknod, ++ .u.m = { ++ .mode = mode, ++ .dev = dev ++ } ++ }; ++ return add_simple(dir, dentry, &arg); ++} ++ ++int aufs_symlink(struct inode *dir, struct dentry *dentry, const char *symname) ++{ ++ struct simple_arg arg = { ++ .type = Symlink, ++ .u.s.symname = symname ++ }; ++ return add_simple(dir, dentry, &arg); ++} ++ ++int aufs_create(struct inode *dir, struct dentry *dentry, int mode, ++ struct nameidata *nd) ++{ ++ struct simple_arg arg = { ++ .type = Creat, ++ .u.c = { ++ .mode = mode, ++ .nd = nd ++ } ++ }; ++ return add_simple(dir, dentry, &arg); ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++struct au_link_args { ++ aufs_bindex_t bdst, bsrc; ++ struct au_pin pin; ++ struct path h_path; ++ struct dentry *src_parent, *parent; ++}; ++ ++static int au_cpup_before_link(struct dentry *src_dentry, ++ struct au_link_args *a) ++{ ++ int err; ++ struct dentry *h_src_dentry; ++ struct mutex *h_mtx; ++ ++ di_read_lock_parent(a->src_parent, AuLock_IR); ++ err = au_test_and_cpup_dirs(src_dentry, a->bdst); ++ if (unlikely(err)) ++ goto out; ++ ++ h_src_dentry = au_h_dptr(src_dentry, a->bsrc); ++ h_mtx = &h_src_dentry->d_inode->i_mutex; ++ err = au_pin(&a->pin, src_dentry, a->bdst, ++ au_opt_udba(src_dentry->d_sb), ++ AuPin_DI_LOCKED | AuPin_MNT_WRITE); ++ if (unlikely(err)) ++ goto out; ++ mutex_lock_nested(h_mtx, AuLsc_I_CHILD); ++ err = au_sio_cpup_simple(src_dentry, a->bdst, -1, ++ AuCpup_DTIME /* | AuCpup_KEEPLINO */); ++ mutex_unlock(h_mtx); ++ au_unpin(&a->pin); ++ ++ out: ++ di_read_unlock(a->src_parent, AuLock_IR); ++ return err; ++} ++ ++static int au_cpup_or_link(struct dentry *src_dentry, struct au_link_args *a) ++{ ++ int err; ++ unsigned char plink; ++ struct inode *h_inode, *inode; ++ struct dentry *h_src_dentry; ++ struct super_block *sb; ++ ++ plink = 0; ++ h_inode = NULL; ++ sb = src_dentry->d_sb; ++ inode = src_dentry->d_inode; ++ if (au_ibstart(inode) <= a->bdst) ++ h_inode = au_h_iptr(inode, a->bdst); ++ if (!h_inode || !h_inode->i_nlink) { ++ /* copyup src_dentry as the name of dentry. */ ++ au_set_dbstart(src_dentry, a->bdst); ++ au_set_h_dptr(src_dentry, a->bdst, dget(a->h_path.dentry)); ++ h_inode = au_h_dptr(src_dentry, a->bsrc)->d_inode; ++ mutex_lock_nested(&h_inode->i_mutex, AuLsc_I_CHILD); ++ err = au_sio_cpup_single(src_dentry, a->bdst, a->bsrc, -1, ++ AuCpup_KEEPLINO, a->parent); ++ mutex_unlock(&h_inode->i_mutex); ++ au_set_h_dptr(src_dentry, a->bdst, NULL); ++ au_set_dbstart(src_dentry, a->bsrc); ++ } else { ++ /* the inode of src_dentry already exists on a.bdst branch */ ++ h_src_dentry = d_find_alias(h_inode); ++ if (!h_src_dentry && au_plink_test(inode)) { ++ plink = 1; ++ h_src_dentry = au_plink_lkup(inode, a->bdst); ++ err = PTR_ERR(h_src_dentry); ++ if (IS_ERR(h_src_dentry)) ++ goto out; ++ ++ if (unlikely(!h_src_dentry->d_inode)) { ++ dput(h_src_dentry); ++ h_src_dentry = NULL; ++ } ++ ++ } ++ if (h_src_dentry) { ++ err = vfsub_link(h_src_dentry, au_pinned_h_dir(&a->pin), ++ &a->h_path); ++ dput(h_src_dentry); ++ } else { ++ AuIOErr("no dentry found for hi%lu on b%d\n", ++ h_inode->i_ino, a->bdst); ++ err = -EIO; ++ } ++ } ++ ++ if (!err && !plink) ++ au_plink_append(inode, a->bdst, a->h_path.dentry); ++ ++out: ++ return err; ++} ++ ++int aufs_link(struct dentry *src_dentry, struct inode *dir, ++ struct dentry *dentry) ++{ ++ int err, rerr; ++ struct au_dtime dt; ++ struct au_link_args *a; ++ struct dentry *wh_dentry, *h_src_dentry; ++ struct inode *inode; ++ struct super_block *sb; ++ struct au_wr_dir_args wr_dir_args = { ++ /* .force_btgt = -1, */ ++ .flags = AuWrDir_ADD_ENTRY ++ }; ++ ++ IMustLock(dir); ++ inode = src_dentry->d_inode; ++ IMustLock(inode); ++ ++ err = -ENOENT; ++ if (unlikely(!inode->i_nlink)) ++ goto out; ++ ++ err = -ENOMEM; ++ a = kzalloc(sizeof(*a), GFP_NOFS); ++ if (unlikely(!a)) ++ goto out; ++ ++ a->parent = dentry->d_parent; /* dir inode is locked */ ++ aufs_read_and_write_lock2(dentry, src_dentry, /*AuLock_FLUSH*/0); ++ a->src_parent = dget_parent(src_dentry); ++ wr_dir_args.force_btgt = au_dbstart(src_dentry); ++ ++ di_write_lock_parent(a->parent); ++ wr_dir_args.force_btgt = au_wbr(dentry, wr_dir_args.force_btgt); ++ wh_dentry = lock_hdir_lkup_wh(dentry, &dt, src_dentry, &a->pin, ++ &wr_dir_args); ++ err = PTR_ERR(wh_dentry); ++ if (IS_ERR(wh_dentry)) ++ goto out_unlock; ++ ++ err = 0; ++ sb = dentry->d_sb; ++ a->bdst = au_dbstart(dentry); ++ a->h_path.dentry = au_h_dptr(dentry, a->bdst); ++ a->h_path.mnt = au_sbr_mnt(sb, a->bdst); ++ a->bsrc = au_dbstart(src_dentry); ++ if (au_opt_test(au_mntflags(sb), PLINK)) { ++ if (a->bdst < a->bsrc ++ /* && h_src_dentry->d_sb != a->h_path.dentry->d_sb */) ++ err = au_cpup_or_link(src_dentry, a); ++ else { ++ h_src_dentry = au_h_dptr(src_dentry, a->bdst); ++ err = vfsub_link(h_src_dentry, au_pinned_h_dir(&a->pin), ++ &a->h_path); ++ } ++ } else { ++ /* ++ * copyup src_dentry to the branch we process, ++ * and then link(2) to it. ++ */ ++ if (a->bdst < a->bsrc ++ /* && h_src_dentry->d_sb != a->h_path.dentry->d_sb */) { ++ au_unpin(&a->pin); ++ di_write_unlock(a->parent); ++ err = au_cpup_before_link(src_dentry, a); ++ di_write_lock_parent(a->parent); ++ if (!err) ++ err = au_pin(&a->pin, dentry, a->bdst, ++ au_opt_udba(sb), ++ AuPin_DI_LOCKED | AuPin_MNT_WRITE); ++ if (unlikely(err)) ++ goto out_wh; ++ } ++ if (!err) { ++ h_src_dentry = au_h_dptr(src_dentry, a->bdst); ++ err = -ENOENT; ++ if (h_src_dentry && h_src_dentry->d_inode) ++ err = vfsub_link(h_src_dentry, ++ au_pinned_h_dir(&a->pin), ++ &a->h_path); ++ } ++ } ++ if (unlikely(err)) ++ goto out_unpin; ++ ++ if (wh_dentry) { ++ a->h_path.dentry = wh_dentry; ++ err = au_wh_unlink_dentry(au_pinned_h_dir(&a->pin), &a->h_path, ++ dentry); ++ if (unlikely(err)) ++ goto out_revert; ++ } ++ ++ dir->i_version++; ++ if (au_ibstart(dir) == au_dbstart(dentry)) ++ au_cpup_attr_timesizes(dir); ++ inc_nlink(inode); ++ inode->i_ctime = dir->i_ctime; ++ if (!d_unhashed(a->h_path.dentry)) ++ d_instantiate(dentry, au_igrab(inode)); ++ else ++ /* some filesystem calls d_drop() */ ++ d_drop(dentry); ++ goto out_unpin; /* success */ ++ ++ out_revert: ++ rerr = vfsub_unlink(au_pinned_h_dir(&a->pin), &a->h_path, /*force*/0); ++ if (!rerr) ++ goto out_dt; ++ AuIOErr("%.*s reverting failed(%d, %d)\n", ++ AuDLNPair(dentry), err, rerr); ++ err = -EIO; ++ out_dt: ++ d_drop(dentry); ++ au_dtime_revert(&dt); ++ out_unpin: ++ au_unpin(&a->pin); ++ out_wh: ++ dput(wh_dentry); ++ out_unlock: ++ if (unlikely(err)) { ++ au_update_dbstart(dentry); ++ d_drop(dentry); ++ } ++ di_write_unlock(a->parent); ++ dput(a->src_parent); ++ aufs_read_and_write_unlock2(dentry, src_dentry); ++ kfree(a); ++ out: ++ return err; ++} ++ ++int aufs_mkdir(struct inode *dir, struct dentry *dentry, int mode) ++{ ++ int err, rerr; ++ aufs_bindex_t bindex; ++ unsigned char diropq; ++ struct path h_path; ++ struct dentry *wh_dentry, *parent, *opq_dentry; ++ struct mutex *h_mtx; ++ struct super_block *sb; ++ struct { ++ struct au_pin pin; ++ struct au_dtime dt; ++ } *a; /* reduce the stack usage */ ++ struct au_wr_dir_args wr_dir_args = { ++ .force_btgt = -1, ++ .flags = AuWrDir_ADD_ENTRY | AuWrDir_ISDIR ++ }; ++ ++ IMustLock(dir); ++ ++ err = -ENOMEM; ++ a = kmalloc(sizeof(*a), GFP_NOFS); ++ if (unlikely(!a)) ++ goto out; ++ ++ aufs_read_lock(dentry, AuLock_DW); ++ parent = dentry->d_parent; /* dir inode is locked */ ++ di_write_lock_parent(parent); ++ wh_dentry = lock_hdir_lkup_wh(dentry, &a->dt, /*src_dentry*/NULL, ++ &a->pin, &wr_dir_args); ++ err = PTR_ERR(wh_dentry); ++ if (IS_ERR(wh_dentry)) ++ goto out_free; ++ ++ sb = dentry->d_sb; ++ bindex = au_dbstart(dentry); ++ h_path.dentry = au_h_dptr(dentry, bindex); ++ h_path.mnt = au_sbr_mnt(sb, bindex); ++ err = vfsub_mkdir(au_pinned_h_dir(&a->pin), &h_path, mode); ++ if (unlikely(err)) ++ goto out_unlock; ++ ++ /* make the dir opaque */ ++ diropq = 0; ++ h_mtx = &h_path.dentry->d_inode->i_mutex; ++ if (wh_dentry ++ || au_opt_test(au_mntflags(sb), ALWAYS_DIROPQ)) { ++ mutex_lock_nested(h_mtx, AuLsc_I_CHILD); ++ opq_dentry = au_diropq_create(dentry, bindex); ++ mutex_unlock(h_mtx); ++ err = PTR_ERR(opq_dentry); ++ if (IS_ERR(opq_dentry)) ++ goto out_dir; ++ dput(opq_dentry); ++ diropq = 1; ++ } ++ ++ err = epilog(dir, bindex, wh_dentry, dentry); ++ if (!err) { ++ inc_nlink(dir); ++ goto out_unlock; /* success */ ++ } ++ ++ /* revert */ ++ if (diropq) { ++ AuLabel(revert opq); ++ mutex_lock_nested(h_mtx, AuLsc_I_CHILD); ++ rerr = au_diropq_remove(dentry, bindex); ++ mutex_unlock(h_mtx); ++ if (rerr) { ++ AuIOErr("%.*s reverting diropq failed(%d, %d)\n", ++ AuDLNPair(dentry), err, rerr); ++ err = -EIO; ++ } ++ } ++ ++ out_dir: ++ AuLabel(revert dir); ++ rerr = vfsub_rmdir(au_pinned_h_dir(&a->pin), &h_path); ++ if (rerr) { ++ AuIOErr("%.*s reverting dir failed(%d, %d)\n", ++ AuDLNPair(dentry), err, rerr); ++ err = -EIO; ++ } ++ d_drop(dentry); ++ au_dtime_revert(&a->dt); ++ out_unlock: ++ au_unpin(&a->pin); ++ dput(wh_dentry); ++ out_free: ++ if (unlikely(err)) { ++ au_update_dbstart(dentry); ++ d_drop(dentry); ++ } ++ di_write_unlock(parent); ++ aufs_read_unlock(dentry, AuLock_DW); ++ kfree(a); ++ out: ++ return err; ++} +diff -Nur linux-2.6.31.4.orig/fs/aufs/i_op.c linux-2.6.31.4/fs/aufs/i_op.c +--- linux-2.6.31.4.orig/fs/aufs/i_op.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/i_op.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,872 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * inode operations (except add/del/rename) ++ */ ++ ++#include <linux/device_cgroup.h> ++#include <linux/fs_stack.h> ++#include <linux/mm.h> ++#include <linux/namei.h> ++#include <linux/security.h> ++#include <linux/uaccess.h> ++#include "aufs.h" ++ ++static int h_permission(struct inode *h_inode, int mask, ++ struct vfsmount *h_mnt, int brperm) ++{ ++ int err; ++ const unsigned char write_mask = !!(mask & (MAY_WRITE | MAY_APPEND)); ++ ++ err = -EACCES; ++ if ((write_mask && IS_IMMUTABLE(h_inode)) ++ || ((mask & MAY_EXEC) ++ && S_ISREG(h_inode->i_mode) ++ && ((h_mnt->mnt_flags & MNT_NOEXEC) ++ || !(h_inode->i_mode & S_IXUGO)))) ++ goto out; ++ ++ /* ++ * - skip the lower fs test in the case of write to ro branch. ++ * - nfs dir permission write check is optimized, but a policy for ++ * link/rename requires a real check. ++ */ ++ if ((write_mask && !au_br_writable(brperm)) ++ || (au_test_nfs(h_inode->i_sb) && S_ISDIR(h_inode->i_mode) ++ && write_mask && !(mask & MAY_READ)) ++ || !h_inode->i_op->permission) { ++ /* AuLabel(generic_permission); */ ++ err = generic_permission(h_inode, mask, NULL); ++ } else { ++ /* AuLabel(h_inode->permission); */ ++ err = h_inode->i_op->permission(h_inode, mask); ++ AuTraceErr(err); ++ } ++ ++ if (!err) ++ err = devcgroup_inode_permission(h_inode, mask); ++ if (!err) ++ err = security_inode_permission ++ (h_inode, mask & (MAY_READ | MAY_WRITE | MAY_EXEC ++ | MAY_APPEND)); ++ ++ out: ++ return err; ++} ++ ++static int aufs_permission(struct inode *inode, int mask) ++{ ++ int err; ++ aufs_bindex_t bindex, bend; ++ const unsigned char isdir = !!S_ISDIR(inode->i_mode); ++ const unsigned char write_mask = !!(mask & (MAY_WRITE | MAY_APPEND)); ++ struct inode *h_inode; ++ struct super_block *sb; ++ struct au_branch *br; ++ ++ sb = inode->i_sb; ++ si_read_lock(sb, AuLock_FLUSH); ++ ii_read_lock_child(inode); ++ ++ if (!isdir || write_mask) { ++ h_inode = au_h_iptr(inode, au_ibstart(inode)); ++ AuDebugOn(!h_inode ++ || ((h_inode->i_mode & S_IFMT) ++ != (inode->i_mode & S_IFMT))); ++ err = 0; ++ bindex = au_ibstart(inode); ++ br = au_sbr(sb, bindex); ++ err = h_permission(h_inode, mask, br->br_mnt, br->br_perm); ++ ++ if (write_mask && !err) { ++ /* test whether the upper writable branch exists */ ++ err = -EROFS; ++ for (; bindex >= 0; bindex--) ++ if (!au_br_rdonly(au_sbr(sb, bindex))) { ++ err = 0; ++ break; ++ } ++ } ++ goto out; ++ } ++ ++ /* non-write to dir */ ++ err = 0; ++ bend = au_ibend(inode); ++ for (bindex = au_ibstart(inode); !err && bindex <= bend; bindex++) { ++ h_inode = au_h_iptr(inode, bindex); ++ if (h_inode) { ++ AuDebugOn(!S_ISDIR(h_inode->i_mode)); ++ br = au_sbr(sb, bindex); ++ err = h_permission(h_inode, mask, br->br_mnt, ++ br->br_perm); ++ } ++ } ++ ++ out: ++ ii_read_unlock(inode); ++ si_read_unlock(sb); ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static struct dentry *aufs_lookup(struct inode *dir, struct dentry *dentry, ++ struct nameidata *nd) ++{ ++ struct dentry *ret, *parent; ++ struct inode *inode, *h_inode; ++ struct mutex *mtx; ++ struct super_block *sb; ++ int err, npositive; ++ aufs_bindex_t bstart; ++ ++ /* temporary workaround for a bug in NFSD readdir */ ++ if (!au_test_nfsd(current)) ++ IMustLock(dir); ++ else ++ WARN_ONCE(!mutex_is_locked(&dir->i_mutex), ++ "a known problem of NFSD readdir since 2.6.28\n"); ++ ++ sb = dir->i_sb; ++ si_read_lock(sb, AuLock_FLUSH); ++ err = au_alloc_dinfo(dentry); ++ ret = ERR_PTR(err); ++ if (unlikely(err)) ++ goto out; ++ ++ parent = dentry->d_parent; /* dir inode is locked */ ++ di_read_lock_parent(parent, AuLock_IR); ++ npositive = au_lkup_dentry(dentry, au_dbstart(parent), /*type*/0, nd); ++ di_read_unlock(parent, AuLock_IR); ++ err = npositive; ++ ret = ERR_PTR(err); ++ if (unlikely(err < 0)) ++ goto out_unlock; ++ ++ inode = NULL; ++ if (npositive) { ++ bstart = au_dbstart(dentry); ++ h_inode = au_h_dptr(dentry, bstart)->d_inode; ++ if (!S_ISDIR(h_inode->i_mode)) { ++ /* ++ * stop 'race'-ing between hardlinks under different ++ * parents. ++ */ ++ mtx = &au_sbr(sb, bstart)->br_xino.xi_nondir_mtx; ++ mutex_lock(mtx); ++ inode = au_new_inode(dentry, /*must_new*/0); ++ mutex_unlock(mtx); ++ } else ++ inode = au_new_inode(dentry, /*must_new*/0); ++ ret = (void *)inode; ++ } ++ if (IS_ERR(inode)) ++ goto out_unlock; ++ ++ ret = d_splice_alias(inode, dentry); ++ if (unlikely(IS_ERR(ret) && inode)) ++ ii_write_unlock(inode); ++ au_store_oflag(nd, inode); ++ ++ out_unlock: ++ di_write_unlock(dentry); ++ out: ++ si_read_unlock(sb); ++ return ret; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static int au_wr_dir_cpup(struct dentry *dentry, struct dentry *parent, ++ const unsigned char add_entry, aufs_bindex_t bcpup, ++ aufs_bindex_t bstart) ++{ ++ int err; ++ struct dentry *h_parent; ++ struct inode *h_dir; ++ ++ if (add_entry) { ++ au_update_dbstart(dentry); ++ IMustLock(parent->d_inode); ++ } else ++ di_write_lock_parent(parent); ++ ++ err = 0; ++ if (!au_h_dptr(parent, bcpup)) { ++ if (bstart < bcpup) ++ err = au_cpdown_dirs(dentry, bcpup); ++ else ++ err = au_cpup_dirs(dentry, bcpup); ++ } ++ if (!err && add_entry) { ++ h_parent = au_h_dptr(parent, bcpup); ++ h_dir = h_parent->d_inode; ++ mutex_lock_nested(&h_dir->i_mutex, AuLsc_I_PARENT); ++ err = au_lkup_neg(dentry, bcpup); ++ /* todo: no unlock here */ ++ mutex_unlock(&h_dir->i_mutex); ++ if (bstart < bcpup && au_dbstart(dentry) < 0) { ++ au_set_dbstart(dentry, 0); ++ au_update_dbrange(dentry, /*do_put_zero*/0); ++ } ++ } ++ ++ if (!add_entry) ++ di_write_unlock(parent); ++ if (!err) ++ err = bcpup; /* success */ ++ ++ return err; ++} ++ ++/* ++ * decide the branch and the parent dir where we will create a new entry. ++ * returns new bindex or an error. ++ * copyup the parent dir if needed. ++ */ ++int au_wr_dir(struct dentry *dentry, struct dentry *src_dentry, ++ struct au_wr_dir_args *args) ++{ ++ int err; ++ aufs_bindex_t bcpup, bstart, src_bstart; ++ const unsigned char add_entry = !!au_ftest_wrdir(args->flags, ++ ADD_ENTRY); ++ struct super_block *sb; ++ struct dentry *parent; ++ struct au_sbinfo *sbinfo; ++ ++ sb = dentry->d_sb; ++ sbinfo = au_sbi(sb); ++ parent = dget_parent(dentry); ++ bstart = au_dbstart(dentry); ++ bcpup = bstart; ++ if (args->force_btgt < 0) { ++ if (src_dentry) { ++ src_bstart = au_dbstart(src_dentry); ++ if (src_bstart < bstart) ++ bcpup = src_bstart; ++ } else if (add_entry) { ++ err = AuWbrCreate(sbinfo, dentry, ++ au_ftest_wrdir(args->flags, ISDIR)); ++ bcpup = err; ++ } ++ ++ if (bcpup < 0 || au_test_ro(sb, bcpup, dentry->d_inode)) { ++ if (add_entry) ++ err = AuWbrCopyup(sbinfo, dentry); ++ else { ++ if (!IS_ROOT(dentry)) { ++ di_read_lock_parent(parent, !AuLock_IR); ++ err = AuWbrCopyup(sbinfo, dentry); ++ di_read_unlock(parent, !AuLock_IR); ++ } else ++ err = AuWbrCopyup(sbinfo, dentry); ++ } ++ bcpup = err; ++ if (unlikely(err < 0)) ++ goto out; ++ } ++ } else { ++ bcpup = args->force_btgt; ++ AuDebugOn(au_test_ro(sb, bcpup, dentry->d_inode)); ++ } ++ AuDbg("bstart %d, bcpup %d\n", bstart, bcpup); ++ if (bstart < bcpup) ++ au_update_dbrange(dentry, /*do_put_zero*/1); ++ ++ err = bcpup; ++ if (bcpup == bstart) ++ goto out; /* success */ ++ ++ /* copyup the new parent into the branch we process */ ++ err = au_wr_dir_cpup(dentry, parent, add_entry, bcpup, bstart); ++ ++ out: ++ dput(parent); ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++struct dentry *au_pinned_h_parent(struct au_pin *pin) ++{ ++ if (pin && pin->parent) ++ return au_h_dptr(pin->parent, pin->bindex); ++ return NULL; ++} ++ ++void au_unpin(struct au_pin *p) ++{ ++ if (au_ftest_pin(p->flags, MNT_WRITE)) ++ mnt_drop_write(p->h_mnt); ++ if (!p->hdir) ++ return; ++ ++ au_hin_imtx_unlock(p->hdir); ++ if (!au_ftest_pin(p->flags, DI_LOCKED)) ++ di_read_unlock(p->parent, AuLock_IR); ++ iput(p->hdir->hi_inode); ++ dput(p->parent); ++ p->parent = NULL; ++ p->hdir = NULL; ++ p->h_mnt = NULL; ++} ++ ++int au_do_pin(struct au_pin *p) ++{ ++ int err; ++ struct super_block *sb; ++ struct dentry *h_dentry, *h_parent; ++ struct au_branch *br; ++ struct inode *h_dir; ++ ++ err = 0; ++ sb = p->dentry->d_sb; ++ br = au_sbr(sb, p->bindex); ++ if (IS_ROOT(p->dentry)) { ++ if (au_ftest_pin(p->flags, MNT_WRITE)) { ++ p->h_mnt = br->br_mnt; ++ err = mnt_want_write(p->h_mnt); ++ if (unlikely(err)) { ++ au_fclr_pin(p->flags, MNT_WRITE); ++ goto out_err; ++ } ++ } ++ goto out; ++ } ++ ++ h_dentry = NULL; ++ if (p->bindex <= au_dbend(p->dentry)) ++ h_dentry = au_h_dptr(p->dentry, p->bindex); ++ ++ p->parent = dget_parent(p->dentry); ++ if (!au_ftest_pin(p->flags, DI_LOCKED)) ++ di_read_lock(p->parent, AuLock_IR, p->lsc_di); ++ ++ h_dir = NULL; ++ h_parent = au_h_dptr(p->parent, p->bindex); ++ p->hdir = au_hi(p->parent->d_inode, p->bindex); ++ if (p->hdir) ++ h_dir = p->hdir->hi_inode; ++ ++ /* udba case */ ++ if (unlikely(!p->hdir || !h_dir)) { ++ if (!au_ftest_pin(p->flags, DI_LOCKED)) ++ di_read_unlock(p->parent, AuLock_IR); ++ dput(p->parent); ++ p->parent = NULL; ++ goto out_err; ++ } ++ ++ au_igrab(h_dir); ++ au_hin_imtx_lock_nested(p->hdir, p->lsc_hi); ++ ++ if (unlikely(p->hdir->hi_inode != h_parent->d_inode)) { ++ err = -EBUSY; ++ goto out_unpin; ++ } ++ if (h_dentry) { ++ err = au_h_verify(h_dentry, p->udba, h_dir, h_parent, br); ++ if (unlikely(err)) { ++ au_fclr_pin(p->flags, MNT_WRITE); ++ goto out_unpin; ++ } ++ } ++ ++ if (au_ftest_pin(p->flags, MNT_WRITE)) { ++ p->h_mnt = br->br_mnt; ++ err = mnt_want_write(p->h_mnt); ++ if (unlikely(err)) { ++ au_fclr_pin(p->flags, MNT_WRITE); ++ goto out_unpin; ++ } ++ } ++ goto out; /* success */ ++ ++ out_unpin: ++ au_unpin(p); ++ out_err: ++ AuErr("err %d\n", err); ++ err = au_busy_or_stale(); ++ out: ++ return err; ++} ++ ++void au_pin_init(struct au_pin *p, struct dentry *dentry, ++ aufs_bindex_t bindex, int lsc_di, int lsc_hi, ++ unsigned int udba, unsigned char flags) ++{ ++ p->dentry = dentry; ++ p->udba = udba; ++ p->lsc_di = lsc_di; ++ p->lsc_hi = lsc_hi; ++ p->flags = flags; ++ p->bindex = bindex; ++ ++ p->parent = NULL; ++ p->hdir = NULL; ++ p->h_mnt = NULL; ++} ++ ++int au_pin(struct au_pin *pin, struct dentry *dentry, aufs_bindex_t bindex, ++ unsigned int udba, unsigned char flags) ++{ ++ au_pin_init(pin, dentry, bindex, AuLsc_DI_PARENT, AuLsc_I_PARENT2, ++ udba, flags); ++ return au_do_pin(pin); ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++#define AuIcpup_DID_CPUP 1 ++#define au_ftest_icpup(flags, name) ((flags) & AuIcpup_##name) ++#define au_fset_icpup(flags, name) { (flags) |= AuIcpup_##name; } ++#define au_fclr_icpup(flags, name) { (flags) &= ~AuIcpup_##name; } ++ ++struct au_icpup_args { ++ unsigned char flags; ++ unsigned char pin_flags; ++ aufs_bindex_t btgt; ++ struct au_pin pin; ++ struct path h_path; ++ struct inode *h_inode; ++}; ++ ++static int au_lock_and_icpup(struct dentry *dentry, struct iattr *ia, ++ struct au_icpup_args *a) ++{ ++ int err; ++ unsigned int udba; ++ loff_t sz; ++ aufs_bindex_t bstart; ++ struct dentry *hi_wh, *parent; ++ struct inode *inode; ++ struct au_wr_dir_args wr_dir_args = { ++ .force_btgt = -1, ++ .flags = 0 ++ }; ++ ++ di_write_lock_child(dentry); ++ bstart = au_dbstart(dentry); ++ inode = dentry->d_inode; ++ if (S_ISDIR(inode->i_mode)) ++ au_fset_wrdir(wr_dir_args.flags, ISDIR); ++ /* plink or hi_wh() case */ ++ if (bstart != au_ibstart(inode)) ++ wr_dir_args.force_btgt = au_ibstart(inode); ++ err = au_wr_dir(dentry, /*src_dentry*/NULL, &wr_dir_args); ++ if (unlikely(err < 0)) ++ goto out_dentry; ++ a->btgt = err; ++ if (err != bstart) ++ au_fset_icpup(a->flags, DID_CPUP); ++ ++ err = 0; ++ a->pin_flags = AuPin_MNT_WRITE; ++ parent = NULL; ++ if (!IS_ROOT(dentry)) { ++ au_fset_pin(a->pin_flags, DI_LOCKED); ++ parent = dget_parent(dentry); ++ di_write_lock_parent(parent); ++ } ++ ++ udba = au_opt_udba(dentry->d_sb); ++ if (d_unhashed(dentry) || (ia->ia_valid & ATTR_FILE)) ++ udba = AuOpt_UDBA_NONE; ++ err = au_pin(&a->pin, dentry, a->btgt, udba, a->pin_flags); ++ if (unlikely(err)) { ++ if (parent) { ++ di_write_unlock(parent); ++ dput(parent); ++ } ++ goto out_dentry; ++ } ++ a->h_path.dentry = au_h_dptr(dentry, bstart); ++ a->h_inode = a->h_path.dentry->d_inode; ++ mutex_lock_nested(&a->h_inode->i_mutex, AuLsc_I_CHILD); ++ sz = -1; ++ if ((ia->ia_valid & ATTR_SIZE) && ia->ia_size < i_size_read(a->h_inode)) ++ sz = ia->ia_size; ++ ++ hi_wh = NULL; ++ if (au_ftest_icpup(a->flags, DID_CPUP) && d_unhashed(dentry)) { ++ hi_wh = au_hi_wh(inode, a->btgt); ++ if (!hi_wh) { ++ err = au_sio_cpup_wh(dentry, a->btgt, sz, /*file*/NULL); ++ if (unlikely(err)) ++ goto out_unlock; ++ hi_wh = au_hi_wh(inode, a->btgt); ++ /* todo: revalidate hi_wh? */ ++ } ++ } ++ ++ if (parent) { ++ au_pin_set_parent_lflag(&a->pin, /*lflag*/0); ++ di_downgrade_lock(parent, AuLock_IR); ++ dput(parent); ++ } ++ if (!au_ftest_icpup(a->flags, DID_CPUP)) ++ goto out; /* success */ ++ ++ if (!d_unhashed(dentry)) { ++ err = au_sio_cpup_simple(dentry, a->btgt, sz, AuCpup_DTIME); ++ if (!err) ++ a->h_path.dentry = au_h_dptr(dentry, a->btgt); ++ } else if (!hi_wh) ++ a->h_path.dentry = au_h_dptr(dentry, a->btgt); ++ else ++ a->h_path.dentry = hi_wh; /* do not dget here */ ++ ++ out_unlock: ++ mutex_unlock(&a->h_inode->i_mutex); ++ a->h_inode = a->h_path.dentry->d_inode; ++ if (!err) { ++ mutex_lock_nested(&a->h_inode->i_mutex, AuLsc_I_CHILD); ++ goto out; /* success */ ++ } ++ ++ au_unpin(&a->pin); ++ ++ out_dentry: ++ di_write_unlock(dentry); ++ out: ++ return err; ++} ++ ++static int aufs_setattr(struct dentry *dentry, struct iattr *ia) ++{ ++ int err; ++ struct inode *inode; ++ struct super_block *sb; ++ struct file *file; ++ struct au_icpup_args *a; ++ ++ err = -ENOMEM; ++ a = kzalloc(sizeof(*a), GFP_NOFS); ++ if (unlikely(!a)) ++ goto out; ++ ++ inode = dentry->d_inode; ++ IMustLock(inode); ++ sb = dentry->d_sb; ++ si_read_lock(sb, AuLock_FLUSH); ++ ++ file = NULL; ++ if (ia->ia_valid & ATTR_FILE) { ++ /* currently ftruncate(2) only */ ++ file = ia->ia_file; ++ fi_write_lock(file); ++ ia->ia_file = au_h_fptr(file, au_fbstart(file)); ++ } ++ ++ if (ia->ia_valid & (ATTR_KILL_SUID | ATTR_KILL_SGID)) ++ ia->ia_valid &= ~ATTR_MODE; ++ ++ err = au_lock_and_icpup(dentry, ia, a); ++ if (unlikely(err < 0)) ++ goto out_si; ++ if (au_ftest_icpup(a->flags, DID_CPUP)) { ++ ia->ia_file = NULL; ++ ia->ia_valid &= ~ATTR_FILE; ++ } ++ ++ a->h_path.mnt = au_sbr_mnt(sb, a->btgt); ++ if (ia->ia_valid & ATTR_SIZE) { ++ struct file *f; ++ ++ if (ia->ia_size < i_size_read(inode)) { ++ /* unmap only */ ++ err = vmtruncate(inode, ia->ia_size); ++ if (unlikely(err)) ++ goto out_unlock; ++ } ++ ++ f = NULL; ++ if (ia->ia_valid & ATTR_FILE) ++ f = ia->ia_file; ++ mutex_unlock(&a->h_inode->i_mutex); ++ err = vfsub_trunc(&a->h_path, ia->ia_size, ia->ia_valid, f); ++ mutex_lock_nested(&a->h_inode->i_mutex, AuLsc_I_CHILD); ++ } else ++ err = vfsub_notify_change(&a->h_path, ia); ++ if (!err) ++ au_cpup_attr_changeable(inode); ++ ++ out_unlock: ++ mutex_unlock(&a->h_inode->i_mutex); ++ au_unpin(&a->pin); ++ di_write_unlock(dentry); ++ out_si: ++ if (file) { ++ fi_write_unlock(file); ++ ia->ia_file = file; ++ ia->ia_valid |= ATTR_FILE; ++ } ++ si_read_unlock(sb); ++ kfree(a); ++ out: ++ return err; ++} ++ ++static int au_getattr_lock_reval(struct dentry *dentry, unsigned int sigen) ++{ ++ int err; ++ struct inode *inode; ++ struct dentry *parent; ++ ++ err = 0; ++ inode = dentry->d_inode; ++ di_write_lock_child(dentry); ++ if (au_digen(dentry) != sigen || au_iigen(inode) != sigen) { ++ parent = dget_parent(dentry); ++ di_read_lock_parent(parent, AuLock_IR); ++ /* returns a number of positive dentries */ ++ err = au_refresh_hdentry(dentry, inode->i_mode & S_IFMT); ++ if (err > 0) ++ err = au_refresh_hinode(inode, dentry); ++ di_read_unlock(parent, AuLock_IR); ++ dput(parent); ++ if (unlikely(!err)) ++ err = -EIO; ++ } ++ di_downgrade_lock(dentry, AuLock_IR); ++ if (unlikely(err)) ++ di_read_unlock(dentry, AuLock_IR); ++ ++ return err; ++} ++ ++static void au_refresh_iattr(struct inode *inode, struct kstat *st, ++ unsigned int nlink) ++{ ++ inode->i_mode = st->mode; ++ inode->i_uid = st->uid; ++ inode->i_gid = st->gid; ++ inode->i_atime = st->atime; ++ inode->i_mtime = st->mtime; ++ inode->i_ctime = st->ctime; ++ ++ au_cpup_attr_nlink(inode, /*force*/0); ++ if (S_ISDIR(inode->i_mode)) { ++ inode->i_nlink -= nlink; ++ inode->i_nlink += st->nlink; ++ } ++ ++ spin_lock(&inode->i_lock); ++ inode->i_blocks = st->blocks; ++ i_size_write(inode, st->size); ++ spin_unlock(&inode->i_lock); ++} ++ ++static int aufs_getattr(struct vfsmount *mnt __maybe_unused, ++ struct dentry *dentry, struct kstat *st) ++{ ++ int err; ++ unsigned int mnt_flags; ++ aufs_bindex_t bindex; ++ unsigned char udba_none, positive; ++ struct super_block *sb, *h_sb; ++ struct inode *inode; ++ struct vfsmount *h_mnt; ++ struct dentry *h_dentry; ++ ++ err = 0; ++ sb = dentry->d_sb; ++ inode = dentry->d_inode; ++ si_read_lock(sb, AuLock_FLUSH); ++ mnt_flags = au_mntflags(sb); ++ udba_none = !!au_opt_test(mnt_flags, UDBA_NONE); ++ ++ /* support fstat(2) */ ++ if (!d_unhashed(dentry) && !udba_none) { ++ unsigned int sigen = au_sigen(sb); ++ if (au_digen(dentry) == sigen && au_iigen(inode) == sigen) ++ di_read_lock_child(dentry, AuLock_IR); ++ else { ++ AuDebugOn(!IS_ROOT(dentry)); ++ err = au_getattr_lock_reval(dentry, sigen); ++ if (unlikely(err)) ++ goto out; ++ } ++ } else ++ di_read_lock_child(dentry, AuLock_IR); ++ ++ bindex = au_ibstart(inode); ++ h_mnt = au_sbr_mnt(sb, bindex); ++ h_sb = h_mnt->mnt_sb; ++ if (!au_test_fs_bad_iattr(h_sb) && udba_none) ++ goto out_fill; /* success */ ++ ++ h_dentry = NULL; ++ if (au_dbstart(dentry) == bindex) ++ h_dentry = dget(au_h_dptr(dentry, bindex)); ++ else if (au_opt_test(mnt_flags, PLINK) && au_plink_test(inode)) { ++ h_dentry = au_plink_lkup(inode, bindex); ++ if (IS_ERR(h_dentry)) ++ goto out_fill; /* pretending success */ ++ } ++ /* illegally overlapped or something */ ++ if (unlikely(!h_dentry)) ++ goto out_fill; /* pretending success */ ++ ++ positive = !!h_dentry->d_inode; ++ if (positive) ++ err = vfs_getattr(h_mnt, h_dentry, st); ++ dput(h_dentry); ++ if (!err) { ++ if (positive) ++ au_refresh_iattr(inode, st, h_dentry->d_inode->i_nlink); ++ goto out_fill; /* success */ ++ } ++ goto out_unlock; ++ ++ out_fill: ++ generic_fillattr(inode, st); ++ out_unlock: ++ di_read_unlock(dentry, AuLock_IR); ++ out: ++ si_read_unlock(sb); ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static int h_readlink(struct dentry *dentry, int bindex, char __user *buf, ++ int bufsiz) ++{ ++ int err; ++ struct super_block *sb; ++ struct dentry *h_dentry; ++ ++ err = -EINVAL; ++ h_dentry = au_h_dptr(dentry, bindex); ++ if (unlikely(/* !h_dentry ++ || !h_dentry->d_inode ++ || !h_dentry->d_inode->i_op ++ || */ !h_dentry->d_inode->i_op->readlink)) ++ goto out; ++ ++ err = security_inode_readlink(h_dentry); ++ if (unlikely(err)) ++ goto out; ++ ++ sb = dentry->d_sb; ++ if (!au_test_ro(sb, bindex, dentry->d_inode)) { ++ vfsub_touch_atime(au_sbr_mnt(sb, bindex), h_dentry); ++ fsstack_copy_attr_atime(dentry->d_inode, h_dentry->d_inode); ++ } ++ err = h_dentry->d_inode->i_op->readlink(h_dentry, buf, bufsiz); ++ ++ out: ++ return err; ++} ++ ++static int aufs_readlink(struct dentry *dentry, char __user *buf, int bufsiz) ++{ ++ int err; ++ ++ aufs_read_lock(dentry, AuLock_IR); ++ err = h_readlink(dentry, au_dbstart(dentry), buf, bufsiz); ++ aufs_read_unlock(dentry, AuLock_IR); ++ ++ return err; ++} ++ ++static void *aufs_follow_link(struct dentry *dentry, struct nameidata *nd) ++{ ++ int err; ++ char *buf; ++ mm_segment_t old_fs; ++ ++ err = -ENOMEM; ++ buf = __getname(); ++ if (unlikely(!buf)) ++ goto out; ++ ++ aufs_read_lock(dentry, AuLock_IR); ++ old_fs = get_fs(); ++ set_fs(KERNEL_DS); ++ err = h_readlink(dentry, au_dbstart(dentry), (char __user *)buf, ++ PATH_MAX); ++ set_fs(old_fs); ++ aufs_read_unlock(dentry, AuLock_IR); ++ ++ if (err >= 0) { ++ buf[err] = 0; ++ /* will be freed by put_link */ ++ nd_set_link(nd, buf); ++ return NULL; /* success */ ++ } ++ __putname(buf); ++ ++ out: ++ path_put(&nd->path); ++ AuTraceErr(err); ++ return ERR_PTR(err); ++} ++ ++static void aufs_put_link(struct dentry *dentry __maybe_unused, ++ struct nameidata *nd, void *cookie __maybe_unused) ++{ ++ __putname(nd_get_link(nd)); ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static void aufs_truncate_range(struct inode *inode __maybe_unused, ++ loff_t start __maybe_unused, ++ loff_t end __maybe_unused) ++{ ++ AuUnsupport(); ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++struct inode_operations aufs_symlink_iop = { ++ .permission = aufs_permission, ++ .setattr = aufs_setattr, ++ .getattr = aufs_getattr, ++ .readlink = aufs_readlink, ++ .follow_link = aufs_follow_link, ++ .put_link = aufs_put_link ++}; ++ ++struct inode_operations aufs_dir_iop = { ++ .create = aufs_create, ++ .lookup = aufs_lookup, ++ .link = aufs_link, ++ .unlink = aufs_unlink, ++ .symlink = aufs_symlink, ++ .mkdir = aufs_mkdir, ++ .rmdir = aufs_rmdir, ++ .mknod = aufs_mknod, ++ .rename = aufs_rename, ++ ++ .permission = aufs_permission, ++ .setattr = aufs_setattr, ++ .getattr = aufs_getattr ++}; ++ ++struct inode_operations aufs_iop = { ++ .permission = aufs_permission, ++ .setattr = aufs_setattr, ++ .getattr = aufs_getattr, ++ .truncate_range = aufs_truncate_range ++}; +diff -Nur linux-2.6.31.4.orig/fs/aufs/i_op_del.c linux-2.6.31.4/fs/aufs/i_op_del.c +--- linux-2.6.31.4.orig/fs/aufs/i_op_del.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/i_op_del.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,468 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * inode operations (del entry) ++ */ ++ ++#include "aufs.h" ++ ++/* ++ * decide if a new whiteout for @dentry is necessary or not. ++ * when it is necessary, prepare the parent dir for the upper branch whose ++ * branch index is @bcpup for creation. the actual creation of the whiteout will ++ * be done by caller. ++ * return value: ++ * 0: wh is unnecessary ++ * plus: wh is necessary ++ * minus: error ++ */ ++int au_wr_dir_need_wh(struct dentry *dentry, int isdir, aufs_bindex_t *bcpup) ++{ ++ int need_wh, err; ++ aufs_bindex_t bstart; ++ struct super_block *sb; ++ ++ sb = dentry->d_sb; ++ bstart = au_dbstart(dentry); ++ if (*bcpup < 0) { ++ *bcpup = bstart; ++ if (au_test_ro(sb, bstart, dentry->d_inode)) { ++ err = AuWbrCopyup(au_sbi(sb), dentry); ++ *bcpup = err; ++ if (unlikely(err < 0)) ++ goto out; ++ } ++ } else ++ AuDebugOn(bstart < *bcpup ++ || au_test_ro(sb, *bcpup, dentry->d_inode)); ++ AuDbg("bcpup %d, bstart %d\n", *bcpup, bstart); ++ ++ if (*bcpup != bstart) { ++ err = au_cpup_dirs(dentry, *bcpup); ++ if (unlikely(err)) ++ goto out; ++ need_wh = 1; ++ } else { ++ aufs_bindex_t old_bend, new_bend, bdiropq = -1; ++ ++ old_bend = au_dbend(dentry); ++ if (isdir) { ++ bdiropq = au_dbdiropq(dentry); ++ au_set_dbdiropq(dentry, -1); ++ } ++ need_wh = au_lkup_dentry(dentry, bstart + 1, /*type*/0, ++ /*nd*/NULL); ++ err = need_wh; ++ if (isdir) ++ au_set_dbdiropq(dentry, bdiropq); ++ if (unlikely(err < 0)) ++ goto out; ++ new_bend = au_dbend(dentry); ++ if (!need_wh && old_bend != new_bend) { ++ au_set_h_dptr(dentry, new_bend, NULL); ++ au_set_dbend(dentry, old_bend); ++ } ++ } ++ AuDbg("need_wh %d\n", need_wh); ++ err = need_wh; ++ ++ out: ++ return err; ++} ++ ++/* ++ * simple tests for the del-entry operations. ++ * following the checks in vfs, plus the parent-child relationship. ++ */ ++int au_may_del(struct dentry *dentry, aufs_bindex_t bindex, ++ struct dentry *h_parent, int isdir) ++{ ++ int err; ++ umode_t h_mode; ++ struct dentry *h_dentry, *h_latest; ++ struct inode *h_inode; ++ ++ h_dentry = au_h_dptr(dentry, bindex); ++ h_inode = h_dentry->d_inode; ++ if (dentry->d_inode) { ++ err = -ENOENT; ++ if (unlikely(!h_inode || !h_inode->i_nlink)) ++ goto out; ++ ++ h_mode = h_inode->i_mode; ++ if (!isdir) { ++ err = -EISDIR; ++ if (unlikely(S_ISDIR(h_mode))) ++ goto out; ++ } else if (unlikely(!S_ISDIR(h_mode))) { ++ err = -ENOTDIR; ++ goto out; ++ } ++ } else { ++ /* rename(2) case */ ++ err = -EIO; ++ if (unlikely(h_inode)) ++ goto out; ++ } ++ ++ err = -ENOENT; ++ /* expected parent dir is locked */ ++ if (unlikely(h_parent != h_dentry->d_parent)) ++ goto out; ++ err = 0; ++ ++ /* ++ * rmdir a dir may break the consistency on some filesystem. ++ * let's try heavy test. ++ */ ++ err = -EACCES; ++ if (unlikely(au_test_h_perm(h_parent->d_inode, MAY_EXEC | MAY_WRITE))) ++ goto out; ++ ++ h_latest = au_sio_lkup_one(&dentry->d_name, h_parent, ++ au_sbr(dentry->d_sb, bindex)); ++ err = -EIO; ++ if (IS_ERR(h_latest)) ++ goto out; ++ if (h_latest == h_dentry) ++ err = 0; ++ dput(h_latest); ++ ++ out: ++ return err; ++} ++ ++/* ++ * decide the branch where we operate for @dentry. the branch index will be set ++ * @rbcpup. after diciding it, 'pin' it and store the timestamps of the parent ++ * dir for reverting. ++ * when a new whiteout is necessary, create it. ++ */ ++static struct dentry* ++lock_hdir_create_wh(struct dentry *dentry, int isdir, aufs_bindex_t *rbcpup, ++ struct au_dtime *dt, struct au_pin *pin) ++{ ++ struct dentry *wh_dentry; ++ struct super_block *sb; ++ struct path h_path; ++ int err, need_wh; ++ unsigned int udba; ++ aufs_bindex_t bcpup; ++ ++ need_wh = au_wr_dir_need_wh(dentry, isdir, rbcpup); ++ wh_dentry = ERR_PTR(need_wh); ++ if (unlikely(need_wh < 0)) ++ goto out; ++ ++ sb = dentry->d_sb; ++ udba = au_opt_udba(sb); ++ bcpup = *rbcpup; ++ err = au_pin(pin, dentry, bcpup, udba, ++ AuPin_DI_LOCKED | AuPin_MNT_WRITE); ++ wh_dentry = ERR_PTR(err); ++ if (unlikely(err)) ++ goto out; ++ ++ h_path.dentry = au_pinned_h_parent(pin); ++ if (udba != AuOpt_UDBA_NONE ++ && au_dbstart(dentry) == bcpup) { ++ err = au_may_del(dentry, bcpup, h_path.dentry, isdir); ++ wh_dentry = ERR_PTR(err); ++ if (unlikely(err)) ++ goto out_unpin; ++ } ++ ++ h_path.mnt = au_sbr_mnt(sb, bcpup); ++ au_dtime_store(dt, au_pinned_parent(pin), &h_path); ++ wh_dentry = NULL; ++ if (!need_wh) ++ goto out; /* success, no need to create whiteout */ ++ ++ wh_dentry = au_wh_create(dentry, bcpup, h_path.dentry); ++ if (!IS_ERR(wh_dentry)) ++ goto out; /* success */ ++ /* returns with the parent is locked and wh_dentry is dget-ed */ ++ ++ out_unpin: ++ au_unpin(pin); ++ out: ++ return wh_dentry; ++} ++ ++/* ++ * when removing a dir, rename it to a unique temporary whiteout-ed name first ++ * in order to be revertible and save time for removing many child whiteouts ++ * under the dir. ++ * returns 1 when there are too many child whiteout and caller should remove ++ * them asynchronously. returns 0 when the number of children is enough small to ++ * remove now or the branch fs is a remote fs. ++ * otherwise return an error. ++ */ ++static int renwh_and_rmdir(struct dentry *dentry, aufs_bindex_t bindex, ++ struct au_nhash *whlist, struct inode *dir) ++{ ++ int rmdir_later, err, dirwh; ++ struct dentry *h_dentry; ++ struct super_block *sb; ++ ++ sb = dentry->d_sb; ++ SiMustAnyLock(sb); ++ h_dentry = au_h_dptr(dentry, bindex); ++ err = au_whtmp_ren(h_dentry, au_sbr(sb, bindex)); ++ if (unlikely(err)) ++ goto out; ++ ++ /* stop monitoring */ ++ au_hin_free(au_hi(dentry->d_inode, bindex)); ++ ++ if (!au_test_fs_remote(h_dentry->d_sb)) { ++ dirwh = au_sbi(sb)->si_dirwh; ++ rmdir_later = (dirwh <= 1); ++ if (!rmdir_later) ++ rmdir_later = au_nhash_test_longer_wh(whlist, bindex, ++ dirwh); ++ if (rmdir_later) ++ return rmdir_later; ++ } ++ ++ err = au_whtmp_rmdir(dir, bindex, h_dentry, whlist); ++ if (unlikely(err)) { ++ AuIOErr("rmdir %.*s, b%d failed, %d. ignored\n", ++ AuDLNPair(h_dentry), bindex, err); ++ err = 0; ++ } ++ ++ out: ++ return err; ++} ++ ++/* ++ * final procedure for deleting a entry. ++ * maintain dentry and iattr. ++ */ ++static void epilog(struct inode *dir, struct dentry *dentry, ++ aufs_bindex_t bindex) ++{ ++ struct inode *inode; ++ ++ inode = dentry->d_inode; ++ d_drop(dentry); ++ inode->i_ctime = dir->i_ctime; ++ ++ if (atomic_read(&dentry->d_count) == 1) { ++ au_set_h_dptr(dentry, au_dbstart(dentry), NULL); ++ au_update_dbstart(dentry); ++ } ++ if (au_ibstart(dir) == bindex) ++ au_cpup_attr_timesizes(dir); ++ dir->i_version++; ++} ++ ++/* ++ * when an error happened, remove the created whiteout and revert everything. ++ */ ++static int do_revert(int err, struct inode *dir, aufs_bindex_t bwh, ++ struct dentry *wh_dentry, struct dentry *dentry, ++ struct au_dtime *dt) ++{ ++ int rerr; ++ struct path h_path = { ++ .dentry = wh_dentry, ++ .mnt = au_sbr_mnt(dir->i_sb, bwh) ++ }; ++ ++ rerr = au_wh_unlink_dentry(au_h_iptr(dir, bwh), &h_path, dentry); ++ if (!rerr) { ++ au_set_dbwh(dentry, bwh); ++ au_dtime_revert(dt); ++ return 0; ++ } ++ ++ AuIOErr("%.*s reverting whiteout failed(%d, %d)\n", ++ AuDLNPair(dentry), err, rerr); ++ return -EIO; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++int aufs_unlink(struct inode *dir, struct dentry *dentry) ++{ ++ int err; ++ aufs_bindex_t bwh, bindex, bstart; ++ struct au_dtime dt; ++ struct au_pin pin; ++ struct path h_path; ++ struct inode *inode, *h_dir; ++ struct dentry *parent, *wh_dentry; ++ ++ IMustLock(dir); ++ inode = dentry->d_inode; ++ if (unlikely(!inode)) ++ return -ENOENT; /* possible? */ ++ IMustLock(inode); ++ ++ aufs_read_lock(dentry, AuLock_DW); ++ parent = dentry->d_parent; /* dir inode is locked */ ++ di_write_lock_parent(parent); ++ ++ bstart = au_dbstart(dentry); ++ bwh = au_dbwh(dentry); ++ bindex = -1; ++ wh_dentry = lock_hdir_create_wh(dentry, /*isdir*/0, &bindex, &dt, &pin); ++ err = PTR_ERR(wh_dentry); ++ if (IS_ERR(wh_dentry)) ++ goto out; ++ ++ h_path.mnt = au_sbr_mnt(dentry->d_sb, bstart); ++ h_path.dentry = au_h_dptr(dentry, bstart); ++ dget(h_path.dentry); ++ if (bindex == bstart) { ++ h_dir = au_pinned_h_dir(&pin); ++ err = vfsub_unlink(h_dir, &h_path, /*force*/0); ++ } else { ++ /* dir inode is locked */ ++ h_dir = wh_dentry->d_parent->d_inode; ++ IMustLock(h_dir); ++ err = 0; ++ } ++ ++ if (!err) { ++ drop_nlink(inode); ++ epilog(dir, dentry, bindex); ++ ++ /* update target timestamps */ ++ if (bindex == bstart) { ++ vfsub_update_h_iattr(&h_path, /*did*/NULL); /*ignore*/ ++ inode->i_ctime = h_path.dentry->d_inode->i_ctime; ++ } else ++ /* todo: this timestamp may be reverted later */ ++ inode->i_ctime = h_dir->i_ctime; ++ goto out_unlock; /* success */ ++ } ++ ++ /* revert */ ++ if (wh_dentry) { ++ int rerr; ++ ++ rerr = do_revert(err, dir, bwh, wh_dentry, dentry, &dt); ++ if (rerr) ++ err = rerr; ++ } ++ ++ out_unlock: ++ au_unpin(&pin); ++ dput(wh_dentry); ++ dput(h_path.dentry); ++ out: ++ di_write_unlock(parent); ++ aufs_read_unlock(dentry, AuLock_DW); ++ return err; ++} ++ ++int aufs_rmdir(struct inode *dir, struct dentry *dentry) ++{ ++ int err, rmdir_later; ++ aufs_bindex_t bwh, bindex, bstart; ++ struct au_dtime dt; ++ struct au_pin pin; ++ struct inode *inode; ++ struct dentry *parent, *wh_dentry, *h_dentry; ++ struct au_whtmp_rmdir *args; ++ ++ IMustLock(dir); ++ inode = dentry->d_inode; ++ err = -ENOENT; /* possible? */ ++ if (unlikely(!inode)) ++ goto out; ++ IMustLock(inode); ++ ++ aufs_read_lock(dentry, AuLock_DW | AuLock_FLUSH); ++ err = -ENOMEM; ++ args = au_whtmp_rmdir_alloc(dir->i_sb, GFP_NOFS); ++ if (unlikely(!args)) ++ goto out_unlock; ++ ++ parent = dentry->d_parent; /* dir inode is locked */ ++ di_write_lock_parent(parent); ++ err = au_test_empty(dentry, &args->whlist); ++ if (unlikely(err)) ++ goto out_args; ++ ++ bstart = au_dbstart(dentry); ++ bwh = au_dbwh(dentry); ++ bindex = -1; ++ wh_dentry = lock_hdir_create_wh(dentry, /*isdir*/1, &bindex, &dt, &pin); ++ err = PTR_ERR(wh_dentry); ++ if (IS_ERR(wh_dentry)) ++ goto out_args; ++ ++ h_dentry = au_h_dptr(dentry, bstart); ++ dget(h_dentry); ++ rmdir_later = 0; ++ if (bindex == bstart) { ++ err = renwh_and_rmdir(dentry, bstart, &args->whlist, dir); ++ if (err > 0) { ++ rmdir_later = err; ++ err = 0; ++ } ++ } else { ++ /* stop monitoring */ ++ au_hin_free(au_hi(inode, bstart)); ++ ++ /* dir inode is locked */ ++ IMustLock(wh_dentry->d_parent->d_inode); ++ err = 0; ++ } ++ ++ if (!err) { ++ clear_nlink(inode); ++ au_set_dbdiropq(dentry, -1); ++ epilog(dir, dentry, bindex); ++ ++ if (rmdir_later) { ++ au_whtmp_kick_rmdir(dir, bstart, h_dentry, args); ++ args = NULL; ++ } ++ ++ goto out_unpin; /* success */ ++ } ++ ++ /* revert */ ++ AuLabel(revert); ++ if (wh_dentry) { ++ int rerr; ++ ++ rerr = do_revert(err, dir, bwh, wh_dentry, dentry, &dt); ++ if (rerr) ++ err = rerr; ++ } ++ ++ out_unpin: ++ au_unpin(&pin); ++ dput(wh_dentry); ++ dput(h_dentry); ++ out_args: ++ di_write_unlock(parent); ++ if (args) ++ au_whtmp_rmdir_free(args); ++ out_unlock: ++ aufs_read_unlock(dentry, AuLock_DW); ++ out: ++ return err; ++} +diff -Nur linux-2.6.31.4.orig/fs/aufs/i_op_ren.c linux-2.6.31.4/fs/aufs/i_op_ren.c +--- linux-2.6.31.4.orig/fs/aufs/i_op_ren.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/i_op_ren.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,948 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * inode operation (rename entry) ++ * todo: this is crazy monster ++ */ ++ ++#include "aufs.h" ++ ++enum { AuSRC, AuDST, AuSrcDst }; ++enum { AuPARENT, AuCHILD, AuParentChild }; ++ ++#define AuRen_ISDIR 1 ++#define AuRen_ISSAMEDIR (1 << 1) ++#define AuRen_WHSRC (1 << 2) ++#define AuRen_WHDST (1 << 3) ++#define AuRen_MNT_WRITE (1 << 4) ++#define AuRen_DT_DSTDIR (1 << 5) ++#define AuRen_DIROPQ (1 << 6) ++#define AuRen_CPUP (1 << 7) ++#define au_ftest_ren(flags, name) ((flags) & AuRen_##name) ++#define au_fset_ren(flags, name) { (flags) |= AuRen_##name; } ++#define au_fclr_ren(flags, name) { (flags) &= ~AuRen_##name; } ++ ++struct au_ren_args { ++ struct { ++ struct dentry *dentry, *h_dentry, *parent, *h_parent, ++ *wh_dentry; ++ struct inode *dir, *inode; ++ struct au_hinode *hdir; ++ struct au_dtime dt[AuParentChild]; ++ aufs_bindex_t bstart; ++ } sd[AuSrcDst]; ++ ++#define src_dentry sd[AuSRC].dentry ++#define src_dir sd[AuSRC].dir ++#define src_inode sd[AuSRC].inode ++#define src_h_dentry sd[AuSRC].h_dentry ++#define src_parent sd[AuSRC].parent ++#define src_h_parent sd[AuSRC].h_parent ++#define src_wh_dentry sd[AuSRC].wh_dentry ++#define src_hdir sd[AuSRC].hdir ++#define src_h_dir sd[AuSRC].hdir->hi_inode ++#define src_dt sd[AuSRC].dt ++#define src_bstart sd[AuSRC].bstart ++ ++#define dst_dentry sd[AuDST].dentry ++#define dst_dir sd[AuDST].dir ++#define dst_inode sd[AuDST].inode ++#define dst_h_dentry sd[AuDST].h_dentry ++#define dst_parent sd[AuDST].parent ++#define dst_h_parent sd[AuDST].h_parent ++#define dst_wh_dentry sd[AuDST].wh_dentry ++#define dst_hdir sd[AuDST].hdir ++#define dst_h_dir sd[AuDST].hdir->hi_inode ++#define dst_dt sd[AuDST].dt ++#define dst_bstart sd[AuDST].bstart ++ ++ struct dentry *h_trap; ++ struct au_branch *br; ++ struct au_hinode *src_hinode; ++ struct path h_path; ++ struct au_nhash whlist; ++ aufs_bindex_t btgt; ++ ++ unsigned int flags; ++ ++ struct au_whtmp_rmdir *thargs; ++ struct dentry *h_dst; ++}; ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* ++ * functions for reverting. ++ * when an error happened in a single rename systemcall, we should revert ++ * everything as if nothing happend. ++ * we don't need to revert the copied-up/down the parent dir since they are ++ * harmless. ++ */ ++ ++#define RevertFailure(fmt, args...) do { \ ++ AuIOErr("revert failure: " fmt " (%d, %d)\n", \ ++ ##args, err, rerr); \ ++ err = -EIO; \ ++} while (0) ++ ++static void au_ren_rev_diropq(int err, struct au_ren_args *a) ++{ ++ int rerr; ++ ++ au_hin_imtx_lock_nested(a->src_hinode, AuLsc_I_CHILD); ++ rerr = au_diropq_remove(a->src_dentry, a->btgt); ++ au_hin_imtx_unlock(a->src_hinode); ++ if (rerr) ++ RevertFailure("remove diropq %.*s", AuDLNPair(a->src_dentry)); ++} ++ ++ ++static void au_ren_rev_rename(int err, struct au_ren_args *a) ++{ ++ int rerr; ++ ++ a->h_path.dentry = au_lkup_one(&a->src_dentry->d_name, a->src_h_parent, ++ a->br, /*nd*/NULL); ++ rerr = PTR_ERR(a->h_path.dentry); ++ if (IS_ERR(a->h_path.dentry)) { ++ RevertFailure("au_lkup_one %.*s", AuDLNPair(a->src_dentry)); ++ return; ++ } ++ ++ rerr = vfsub_rename(a->dst_h_dir, ++ au_h_dptr(a->src_dentry, a->btgt), ++ a->src_h_dir, &a->h_path); ++ d_drop(a->h_path.dentry); ++ dput(a->h_path.dentry); ++ /* au_set_h_dptr(a->src_dentry, a->btgt, NULL); */ ++ if (rerr) ++ RevertFailure("rename %.*s", AuDLNPair(a->src_dentry)); ++} ++ ++static void au_ren_rev_cpup(int err, struct au_ren_args *a) ++{ ++ int rerr; ++ ++ a->h_path.dentry = a->dst_h_dentry; ++ rerr = vfsub_unlink(a->dst_h_dir, &a->h_path, /*force*/0); ++ au_set_h_dptr(a->src_dentry, a->btgt, NULL); ++ au_set_dbstart(a->src_dentry, a->src_bstart); ++ if (rerr) ++ RevertFailure("unlink %.*s", AuDLNPair(a->dst_h_dentry)); ++} ++ ++ ++static void au_ren_rev_whtmp(int err, struct au_ren_args *a) ++{ ++ int rerr; ++ ++ a->h_path.dentry = au_lkup_one(&a->dst_dentry->d_name, a->dst_h_parent, ++ a->br, /*nd*/NULL); ++ rerr = PTR_ERR(a->h_path.dentry); ++ if (IS_ERR(a->h_path.dentry)) { ++ RevertFailure("lookup %.*s", AuDLNPair(a->dst_dentry)); ++ return; ++ } ++ if (a->h_path.dentry->d_inode) { ++ d_drop(a->h_path.dentry); ++ dput(a->h_path.dentry); ++ return; ++ } ++ ++ rerr = vfsub_rename(a->dst_h_dir, a->h_dst, a->dst_h_dir, &a->h_path); ++ d_drop(a->h_path.dentry); ++ dput(a->h_path.dentry); ++ if (!rerr) { ++ au_set_h_dptr(a->dst_dentry, a->btgt, NULL); ++ au_set_h_dptr(a->dst_dentry, a->btgt, dget(a->h_dst)); ++ } else ++ RevertFailure("rename %.*s", AuDLNPair(a->h_dst)); ++} ++ ++static void au_ren_rev_whsrc(int err, struct au_ren_args *a) ++{ ++ int rerr; ++ ++ a->h_path.dentry = a->src_wh_dentry; ++ rerr = au_wh_unlink_dentry(a->src_h_dir, &a->h_path, a->src_dentry); ++ if (rerr) ++ RevertFailure("unlink %.*s", AuDLNPair(a->src_wh_dentry)); ++} ++ ++static void au_ren_rev_drop(struct au_ren_args *a) ++{ ++ struct dentry *d, *h_d; ++ int i; ++ aufs_bindex_t bend, bindex; ++ ++ for (i = 0; i < AuSrcDst; i++) { ++ d = a->sd[i].dentry; ++ d_drop(d); ++ bend = au_dbend(d); ++ for (bindex = au_dbstart(d); bindex <= bend; bindex++) { ++ h_d = au_h_dptr(d, bindex); ++ if (h_d) ++ d_drop(h_d); ++ } ++ } ++ ++ au_update_dbstart(a->dst_dentry); ++ if (a->thargs) ++ d_drop(a->h_dst); ++} ++#undef RevertFailure ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* ++ * when we have to copyup the renaming entry, do it with the rename-target name ++ * in order to minimize the cost (the later actual rename is unnecessary). ++ * otherwise rename it on the target branch. ++ */ ++static int au_ren_or_cpup(struct au_ren_args *a) ++{ ++ int err; ++ struct dentry *d; ++ ++ d = a->src_dentry; ++ if (au_dbstart(d) == a->btgt) { ++ a->h_path.dentry = a->dst_h_dentry; ++ if (au_ftest_ren(a->flags, DIROPQ) ++ && au_dbdiropq(d) == a->btgt) ++ au_fclr_ren(a->flags, DIROPQ); ++ AuDebugOn(au_dbstart(d) != a->btgt); ++ err = vfsub_rename(a->src_h_dir, au_h_dptr(d, a->btgt), ++ a->dst_h_dir, &a->h_path); ++ } else { ++ struct mutex *h_mtx = &a->src_h_dentry->d_inode->i_mutex; ++ ++ au_fset_ren(a->flags, CPUP); ++ mutex_lock_nested(h_mtx, AuLsc_I_CHILD); ++ au_set_dbstart(d, a->btgt); ++ au_set_h_dptr(d, a->btgt, dget(a->dst_h_dentry)); ++ err = au_sio_cpup_single(d, a->btgt, a->src_bstart, -1, ++ !AuCpup_DTIME, a->dst_parent); ++ if (unlikely(err)) { ++ au_set_h_dptr(d, a->btgt, NULL); ++ au_set_dbstart(d, a->src_bstart); ++ } ++ mutex_unlock(h_mtx); ++ } ++ ++ return err; ++} ++ ++/* cf. aufs_rmdir() */ ++static int au_ren_del_whtmp(struct au_ren_args *a) ++{ ++ int err; ++ struct inode *dir; ++ ++ dir = a->dst_dir; ++ SiMustAnyLock(dir->i_sb); ++ if (!au_nhash_test_longer_wh(&a->whlist, a->btgt, ++ au_sbi(dir->i_sb)->si_dirwh) ++ || au_test_fs_remote(a->h_dst->d_sb)) { ++ err = au_whtmp_rmdir(dir, a->btgt, a->h_dst, &a->whlist); ++ if (unlikely(err)) ++ AuWarn("failed removing whtmp dir %.*s (%d), " ++ "ignored.\n", AuDLNPair(a->h_dst), err); ++ } else { ++ au_nhash_wh_free(&a->thargs->whlist); ++ a->thargs->whlist = a->whlist; ++ a->whlist.nh_num = 0; ++ au_whtmp_kick_rmdir(dir, a->btgt, a->h_dst, a->thargs); ++ dput(a->h_dst); ++ a->thargs = NULL; ++ } ++ ++ return 0; ++} ++ ++/* make it 'opaque' dir. */ ++static int au_ren_diropq(struct au_ren_args *a) ++{ ++ int err; ++ struct dentry *diropq; ++ ++ err = 0; ++ a->src_hinode = au_hi(a->src_inode, a->btgt); ++ au_hin_imtx_lock_nested(a->src_hinode, AuLsc_I_CHILD); ++ diropq = au_diropq_create(a->src_dentry, a->btgt); ++ au_hin_imtx_unlock(a->src_hinode); ++ if (IS_ERR(diropq)) ++ err = PTR_ERR(diropq); ++ dput(diropq); ++ ++ return err; ++} ++ ++static int do_rename(struct au_ren_args *a) ++{ ++ int err; ++ struct dentry *d, *h_d; ++ ++ /* prepare workqueue args for asynchronous rmdir */ ++ h_d = a->dst_h_dentry; ++ if (au_ftest_ren(a->flags, ISDIR) && h_d->d_inode) { ++ err = -ENOMEM; ++ a->thargs = au_whtmp_rmdir_alloc(a->src_dentry->d_sb, GFP_NOFS); ++ if (unlikely(!a->thargs)) ++ goto out; ++ a->h_dst = dget(h_d); ++ } ++ ++ /* create whiteout for src_dentry */ ++ if (au_ftest_ren(a->flags, WHSRC)) { ++ a->src_wh_dentry ++ = au_wh_create(a->src_dentry, a->btgt, a->src_h_parent); ++ err = PTR_ERR(a->src_wh_dentry); ++ if (IS_ERR(a->src_wh_dentry)) ++ goto out_thargs; ++ } ++ ++ /* lookup whiteout for dentry */ ++ if (au_ftest_ren(a->flags, WHDST)) { ++ h_d = au_wh_lkup(a->dst_h_parent, &a->dst_dentry->d_name, ++ a->br); ++ err = PTR_ERR(h_d); ++ if (IS_ERR(h_d)) ++ goto out_whsrc; ++ if (!h_d->d_inode) ++ dput(h_d); ++ else ++ a->dst_wh_dentry = h_d; ++ } ++ ++ /* rename dentry to tmpwh */ ++ if (a->thargs) { ++ err = au_whtmp_ren(a->dst_h_dentry, a->br); ++ if (unlikely(err)) ++ goto out_whdst; ++ ++ d = a->dst_dentry; ++ au_set_h_dptr(d, a->btgt, NULL); ++ err = au_lkup_neg(d, a->btgt); ++ if (unlikely(err)) ++ goto out_whtmp; ++ a->dst_h_dentry = au_h_dptr(d, a->btgt); ++ } ++ ++ /* cpup src */ ++ if (a->dst_h_dentry->d_inode && a->src_bstart != a->btgt) { ++ struct mutex *h_mtx = &a->src_h_dentry->d_inode->i_mutex; ++ ++ mutex_lock_nested(h_mtx, AuLsc_I_CHILD); ++ err = au_sio_cpup_simple(a->src_dentry, a->btgt, -1, ++ !AuCpup_DTIME); ++ mutex_unlock(h_mtx); ++ if (unlikely(err)) ++ goto out_whtmp; ++ } ++ ++ /* rename by vfs_rename or cpup */ ++ d = a->dst_dentry; ++ if (au_ftest_ren(a->flags, ISDIR) ++ && (a->dst_wh_dentry ++ || au_dbdiropq(d) == a->btgt ++ /* hide the lower to keep xino */ ++ || a->btgt < au_dbend(d) ++ || au_opt_test(au_mntflags(d->d_sb), ALWAYS_DIROPQ))) ++ au_fset_ren(a->flags, DIROPQ); ++ err = au_ren_or_cpup(a); ++ if (unlikely(err)) ++ /* leave the copied-up one */ ++ goto out_whtmp; ++ ++ /* make dir opaque */ ++ if (au_ftest_ren(a->flags, DIROPQ)) { ++ err = au_ren_diropq(a); ++ if (unlikely(err)) ++ goto out_rename; ++ } ++ ++ /* update target timestamps */ ++ AuDebugOn(au_dbstart(a->src_dentry) != a->btgt); ++ a->h_path.dentry = au_h_dptr(a->src_dentry, a->btgt); ++ vfsub_update_h_iattr(&a->h_path, /*did*/NULL); /*ignore*/ ++ a->src_inode->i_ctime = a->h_path.dentry->d_inode->i_ctime; ++ ++ /* remove whiteout for dentry */ ++ if (a->dst_wh_dentry) { ++ a->h_path.dentry = a->dst_wh_dentry; ++ err = au_wh_unlink_dentry(a->dst_h_dir, &a->h_path, ++ a->dst_dentry); ++ if (unlikely(err)) ++ goto out_diropq; ++ } ++ ++ /* remove whtmp */ ++ if (a->thargs) ++ au_ren_del_whtmp(a); /* ignore this error */ ++ ++ err = 0; ++ goto out_success; ++ ++ out_diropq: ++ if (au_ftest_ren(a->flags, DIROPQ)) ++ au_ren_rev_diropq(err, a); ++ out_rename: ++ if (!au_ftest_ren(a->flags, CPUP)) ++ au_ren_rev_rename(err, a); ++ else ++ au_ren_rev_cpup(err, a); ++ out_whtmp: ++ if (a->thargs) ++ au_ren_rev_whtmp(err, a); ++ out_whdst: ++ dput(a->dst_wh_dentry); ++ a->dst_wh_dentry = NULL; ++ out_whsrc: ++ if (a->src_wh_dentry) ++ au_ren_rev_whsrc(err, a); ++ au_ren_rev_drop(a); ++ out_success: ++ dput(a->src_wh_dentry); ++ dput(a->dst_wh_dentry); ++ out_thargs: ++ if (a->thargs) { ++ dput(a->h_dst); ++ au_whtmp_rmdir_free(a->thargs); ++ a->thargs = NULL; ++ } ++ out: ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* ++ * test if @dentry dir can be rename destination or not. ++ * success means, it is a logically empty dir. ++ */ ++static int may_rename_dstdir(struct dentry *dentry, struct au_nhash *whlist) ++{ ++ return au_test_empty(dentry, whlist); ++} ++ ++/* ++ * test if @dentry dir can be rename source or not. ++ * if it can, return 0 and @children is filled. ++ * success means, ++ * - it is a logically empty dir. ++ * - or, it exists on writable branch and has no children including whiteouts ++ * on the lower branch. ++ */ ++static int may_rename_srcdir(struct dentry *dentry, aufs_bindex_t btgt) ++{ ++ int err; ++ aufs_bindex_t bstart; ++ ++ bstart = au_dbstart(dentry); ++ if (bstart != btgt) { ++ struct au_nhash whlist; ++ ++ SiMustAnyLock(dentry->d_sb); ++ err = au_nhash_alloc(&whlist, au_sbi(dentry->d_sb)->si_rdhash, ++ GFP_NOFS); ++ if (unlikely(err)) ++ goto out; ++ err = au_test_empty(dentry, &whlist); ++ au_nhash_wh_free(&whlist); ++ goto out; ++ } ++ ++ if (bstart == au_dbtaildir(dentry)) ++ return 0; /* success */ ++ ++ err = au_test_empty_lower(dentry); ++ ++ out: ++ if (err == -ENOTEMPTY) { ++ AuWarn1("renaming dir who has child(ren) on multiple branches," ++ " is not supported\n"); ++ err = -EXDEV; ++ } ++ return err; ++} ++ ++/* side effect: sets whlist and h_dentry */ ++static int au_ren_may_dir(struct au_ren_args *a) ++{ ++ int err; ++ struct dentry *d; ++ ++ d = a->dst_dentry; ++ SiMustAnyLock(d->d_sb); ++ err = au_nhash_alloc(&a->whlist, au_sbi(d->d_sb)->si_rdhash, GFP_NOFS); ++ if (unlikely(err)) ++ goto out; ++ ++ err = 0; ++ if (au_ftest_ren(a->flags, ISDIR) && a->dst_inode) { ++ au_set_dbstart(d, a->dst_bstart); ++ err = may_rename_dstdir(d, &a->whlist); ++ au_set_dbstart(d, a->btgt); ++ } ++ a->dst_h_dentry = au_h_dptr(d, au_dbstart(d)); ++ if (unlikely(err)) ++ goto out; ++ ++ d = a->src_dentry; ++ a->src_h_dentry = au_h_dptr(d, au_dbstart(d)); ++ if (au_ftest_ren(a->flags, ISDIR)) { ++ err = may_rename_srcdir(d, a->btgt); ++ if (unlikely(err)) { ++ au_nhash_wh_free(&a->whlist); ++ a->whlist.nh_num = 0; ++ } ++ } ++ out: ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* ++ * simple tests for rename. ++ * following the checks in vfs, plus the parent-child relationship. ++ */ ++static int au_may_ren(struct au_ren_args *a) ++{ ++ int err, isdir; ++ struct inode *h_inode; ++ ++ if (a->src_bstart == a->btgt) { ++ err = au_may_del(a->src_dentry, a->btgt, a->src_h_parent, ++ au_ftest_ren(a->flags, ISDIR)); ++ if (unlikely(err)) ++ goto out; ++ err = -EINVAL; ++ if (unlikely(a->src_h_dentry == a->h_trap)) ++ goto out; ++ } ++ ++ err = 0; ++ if (a->dst_bstart != a->btgt) ++ goto out; ++ ++ err = -EIO; ++ h_inode = a->dst_h_dentry->d_inode; ++ isdir = !!au_ftest_ren(a->flags, ISDIR); ++ if (!a->dst_dentry->d_inode) { ++ if (unlikely(h_inode)) ++ goto out; ++ err = au_may_add(a->dst_dentry, a->btgt, a->dst_h_parent, ++ isdir); ++ } else { ++ if (unlikely(!h_inode || !h_inode->i_nlink)) ++ goto out; ++ err = au_may_del(a->dst_dentry, a->btgt, a->dst_h_parent, ++ isdir); ++ if (unlikely(err)) ++ goto out; ++ err = -ENOTEMPTY; ++ if (unlikely(a->dst_h_dentry == a->h_trap)) ++ goto out; ++ err = 0; ++ } ++ ++ out: ++ if (unlikely(err == -ENOENT || err == -EEXIST)) ++ err = -EIO; ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* ++ * locking order ++ * (VFS) ++ * - src_dir and dir by lock_rename() ++ * - inode if exitsts ++ * (aufs) ++ * - lock all ++ * + src_dentry and dentry by aufs_read_and_write_lock2() which calls, ++ * + si_read_lock ++ * + di_write_lock2_child() ++ * + di_write_lock_child() ++ * + ii_write_lock_child() ++ * + di_write_lock_child2() ++ * + ii_write_lock_child2() ++ * + src_parent and parent ++ * + di_write_lock_parent() ++ * + ii_write_lock_parent() ++ * + di_write_lock_parent2() ++ * + ii_write_lock_parent2() ++ * + lower src_dir and dir by vfsub_lock_rename() ++ * + verify the every relationships between child and parent. if any ++ * of them failed, unlock all and return -EBUSY. ++ */ ++static void au_ren_unlock(struct au_ren_args *a) ++{ ++ struct super_block *sb; ++ ++ sb = a->dst_dentry->d_sb; ++ if (au_ftest_ren(a->flags, MNT_WRITE)) ++ mnt_drop_write(a->br->br_mnt); ++ vfsub_unlock_rename(a->src_h_parent, a->src_hdir, ++ a->dst_h_parent, a->dst_hdir); ++} ++ ++static int au_ren_lock(struct au_ren_args *a) ++{ ++ int err; ++ unsigned int udba; ++ ++ err = 0; ++ a->src_h_parent = au_h_dptr(a->src_parent, a->btgt); ++ a->src_hdir = au_hi(a->src_dir, a->btgt); ++ a->dst_h_parent = au_h_dptr(a->dst_parent, a->btgt); ++ a->dst_hdir = au_hi(a->dst_dir, a->btgt); ++ a->h_trap = vfsub_lock_rename(a->src_h_parent, a->src_hdir, ++ a->dst_h_parent, a->dst_hdir); ++ udba = au_opt_udba(a->src_dentry->d_sb); ++ if (unlikely(a->src_hdir->hi_inode != a->src_h_parent->d_inode ++ || a->dst_hdir->hi_inode != a->dst_h_parent->d_inode)) ++ err = au_busy_or_stale(); ++ if (!err && au_dbstart(a->src_dentry) == a->btgt) ++ err = au_h_verify(a->src_h_dentry, udba, ++ a->src_h_parent->d_inode, a->src_h_parent, ++ a->br); ++ if (!err && au_dbstart(a->dst_dentry) == a->btgt) ++ err = au_h_verify(a->dst_h_dentry, udba, ++ a->dst_h_parent->d_inode, a->dst_h_parent, ++ a->br); ++ if (!err) { ++ err = mnt_want_write(a->br->br_mnt); ++ if (unlikely(err)) ++ goto out_unlock; ++ au_fset_ren(a->flags, MNT_WRITE); ++ goto out; /* success */ ++ } ++ ++ err = au_busy_or_stale(); ++ ++ out_unlock: ++ au_ren_unlock(a); ++ out: ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static void au_ren_refresh_dir(struct au_ren_args *a) ++{ ++ struct inode *dir; ++ ++ dir = a->dst_dir; ++ dir->i_version++; ++ if (au_ftest_ren(a->flags, ISDIR)) { ++ /* is this updating defined in POSIX? */ ++ au_cpup_attr_timesizes(a->src_inode); ++ au_cpup_attr_nlink(dir, /*force*/1); ++ if (a->dst_inode) { ++ clear_nlink(a->dst_inode); ++ au_cpup_attr_timesizes(a->dst_inode); ++ } ++ } ++ if (au_ibstart(dir) == a->btgt) ++ au_cpup_attr_timesizes(dir); ++ ++ if (au_ftest_ren(a->flags, ISSAMEDIR)) ++ return; ++ ++ dir = a->src_dir; ++ dir->i_version++; ++ if (au_ftest_ren(a->flags, ISDIR)) ++ au_cpup_attr_nlink(dir, /*force*/1); ++ if (au_ibstart(dir) == a->btgt) ++ au_cpup_attr_timesizes(dir); ++} ++ ++static void au_ren_refresh(struct au_ren_args *a) ++{ ++ aufs_bindex_t bend, bindex; ++ struct dentry *d, *h_d; ++ struct inode *i, *h_i; ++ struct super_block *sb; ++ ++ d = a->src_dentry; ++ au_set_dbwh(d, -1); ++ bend = au_dbend(d); ++ for (bindex = a->btgt + 1; bindex <= bend; bindex++) { ++ h_d = au_h_dptr(d, bindex); ++ if (h_d) ++ au_set_h_dptr(d, bindex, NULL); ++ } ++ au_set_dbend(d, a->btgt); ++ ++ sb = d->d_sb; ++ i = a->src_inode; ++ if (au_opt_test(au_mntflags(sb), PLINK) && au_plink_test(i)) ++ return; /* success */ ++ ++ bend = au_ibend(i); ++ for (bindex = a->btgt + 1; bindex <= bend; bindex++) { ++ h_i = au_h_iptr(i, bindex); ++ if (h_i) { ++ au_xino_write(sb, bindex, h_i->i_ino, /*ino*/0); ++ /* ignore this error */ ++ au_set_h_iptr(i, bindex, NULL, 0); ++ } ++ } ++ au_set_ibend(i, a->btgt); ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* mainly for link(2) and rename(2) */ ++int au_wbr(struct dentry *dentry, aufs_bindex_t btgt) ++{ ++ aufs_bindex_t bdiropq, bwh; ++ struct dentry *parent; ++ struct au_branch *br; ++ ++ parent = dentry->d_parent; ++ IMustLock(parent->d_inode); /* dir is locked */ ++ ++ bdiropq = au_dbdiropq(parent); ++ bwh = au_dbwh(dentry); ++ br = au_sbr(dentry->d_sb, btgt); ++ if (au_br_rdonly(br) ++ || (0 <= bdiropq && bdiropq < btgt) ++ || (0 <= bwh && bwh < btgt)) ++ btgt = -1; ++ ++ AuDbg("btgt %d\n", btgt); ++ return btgt; ++} ++ ++/* sets src_bstart, dst_bstart and btgt */ ++static int au_ren_wbr(struct au_ren_args *a) ++{ ++ int err; ++ struct au_wr_dir_args wr_dir_args = { ++ /* .force_btgt = -1, */ ++ .flags = AuWrDir_ADD_ENTRY ++ }; ++ ++ a->src_bstart = au_dbstart(a->src_dentry); ++ a->dst_bstart = au_dbstart(a->dst_dentry); ++ if (au_ftest_ren(a->flags, ISDIR)) ++ au_fset_wrdir(wr_dir_args.flags, ISDIR); ++ wr_dir_args.force_btgt = a->src_bstart; ++ if (a->dst_inode && a->dst_bstart < a->src_bstart) ++ wr_dir_args.force_btgt = a->dst_bstart; ++ wr_dir_args.force_btgt = au_wbr(a->dst_dentry, wr_dir_args.force_btgt); ++ err = au_wr_dir(a->dst_dentry, a->src_dentry, &wr_dir_args); ++ a->btgt = err; ++ ++ return err; ++} ++ ++static void au_ren_dt(struct au_ren_args *a) ++{ ++ a->h_path.dentry = a->src_h_parent; ++ au_dtime_store(a->src_dt + AuPARENT, a->src_parent, &a->h_path); ++ if (!au_ftest_ren(a->flags, ISSAMEDIR)) { ++ a->h_path.dentry = a->dst_h_parent; ++ au_dtime_store(a->dst_dt + AuPARENT, a->dst_parent, &a->h_path); ++ } ++ ++ au_fclr_ren(a->flags, DT_DSTDIR); ++ if (!au_ftest_ren(a->flags, ISDIR)) ++ return; ++ ++ a->h_path.dentry = a->src_h_dentry; ++ au_dtime_store(a->src_dt + AuCHILD, a->src_dentry, &a->h_path); ++ if (a->dst_h_dentry->d_inode) { ++ au_fset_ren(a->flags, DT_DSTDIR); ++ a->h_path.dentry = a->dst_h_dentry; ++ au_dtime_store(a->dst_dt + AuCHILD, a->dst_dentry, &a->h_path); ++ } ++} ++ ++static void au_ren_rev_dt(int err, struct au_ren_args *a) ++{ ++ struct dentry *h_d; ++ struct mutex *h_mtx; ++ ++ au_dtime_revert(a->src_dt + AuPARENT); ++ if (!au_ftest_ren(a->flags, ISSAMEDIR)) ++ au_dtime_revert(a->dst_dt + AuPARENT); ++ ++ if (au_ftest_ren(a->flags, ISDIR) && err != -EIO) { ++ h_d = a->src_dt[AuCHILD].dt_h_path.dentry; ++ h_mtx = &h_d->d_inode->i_mutex; ++ mutex_lock_nested(h_mtx, AuLsc_I_CHILD); ++ au_dtime_revert(a->src_dt + AuCHILD); ++ mutex_unlock(h_mtx); ++ ++ if (au_ftest_ren(a->flags, DT_DSTDIR)) { ++ h_d = a->dst_dt[AuCHILD].dt_h_path.dentry; ++ h_mtx = &h_d->d_inode->i_mutex; ++ mutex_lock_nested(h_mtx, AuLsc_I_CHILD); ++ au_dtime_revert(a->dst_dt + AuCHILD); ++ mutex_unlock(h_mtx); ++ } ++ } ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++int aufs_rename(struct inode *_src_dir, struct dentry *_src_dentry, ++ struct inode *_dst_dir, struct dentry *_dst_dentry) ++{ ++ int err; ++ /* reduce stack space */ ++ struct au_ren_args *a; ++ ++ IMustLock(_src_dir); ++ IMustLock(_dst_dir); ++ ++ err = -ENOMEM; ++ BUILD_BUG_ON(sizeof(*a) > PAGE_SIZE); ++ a = kzalloc(sizeof(*a), GFP_NOFS); ++ if (unlikely(!a)) ++ goto out; ++ ++ a->src_dir = _src_dir; ++ a->src_dentry = _src_dentry; ++ a->src_inode = a->src_dentry->d_inode; ++ a->src_parent = a->src_dentry->d_parent; /* dir inode is locked */ ++ a->dst_dir = _dst_dir; ++ a->dst_dentry = _dst_dentry; ++ a->dst_inode = a->dst_dentry->d_inode; ++ a->dst_parent = a->dst_dentry->d_parent; /* dir inode is locked */ ++ if (a->dst_inode) { ++ IMustLock(a->dst_inode); ++ au_igrab(a->dst_inode); ++ } ++ ++ err = -ENOTDIR; ++ if (S_ISDIR(a->src_inode->i_mode)) { ++ au_fset_ren(a->flags, ISDIR); ++ if (unlikely(a->dst_inode && !S_ISDIR(a->dst_inode->i_mode))) ++ goto out_free; ++ aufs_read_and_write_lock2(a->dst_dentry, a->src_dentry, ++ AuLock_DIR | AuLock_FLUSH); ++ } else ++ aufs_read_and_write_lock2(a->dst_dentry, a->src_dentry, ++ AuLock_FLUSH); ++ ++ au_fset_ren(a->flags, ISSAMEDIR); /* temporary */ ++ di_write_lock_parent(a->dst_parent); ++ ++ /* which branch we process */ ++ err = au_ren_wbr(a); ++ if (unlikely(err < 0)) ++ goto out_unlock; ++ a->br = au_sbr(a->dst_dentry->d_sb, a->btgt); ++ a->h_path.mnt = a->br->br_mnt; ++ ++ /* are they available to be renamed */ ++ err = au_ren_may_dir(a); ++ if (unlikely(err)) ++ goto out_unlock; ++ ++ /* prepare the writable parent dir on the same branch */ ++ if (a->dst_bstart == a->btgt) { ++ au_fset_ren(a->flags, WHDST); ++ } else { ++ err = au_cpup_dirs(a->dst_dentry, a->btgt); ++ if (unlikely(err)) ++ goto out_children; ++ } ++ ++ if (a->src_dir != a->dst_dir) { ++ /* ++ * this temporary unlock is safe, ++ * because both dir->i_mutex are locked. ++ */ ++ di_write_unlock(a->dst_parent); ++ di_write_lock_parent(a->src_parent); ++ err = au_wr_dir_need_wh(a->src_dentry, ++ au_ftest_ren(a->flags, ISDIR), ++ &a->btgt); ++ di_write_unlock(a->src_parent); ++ di_write_lock2_parent(a->src_parent, a->dst_parent, /*isdir*/1); ++ au_fclr_ren(a->flags, ISSAMEDIR); ++ } else ++ err = au_wr_dir_need_wh(a->src_dentry, ++ au_ftest_ren(a->flags, ISDIR), ++ &a->btgt); ++ if (unlikely(err < 0)) ++ goto out_children; ++ if (err) ++ au_fset_ren(a->flags, WHSRC); ++ ++ /* lock them all */ ++ err = au_ren_lock(a); ++ if (unlikely(err)) ++ goto out_children; ++ ++ if (!au_opt_test(au_mntflags(a->dst_dir->i_sb), UDBA_NONE)) { ++ err = au_may_ren(a); ++ if (unlikely(err)) ++ goto out_hdir; ++ } ++ ++ /* store timestamps to be revertible */ ++ au_ren_dt(a); ++ ++ /* here we go */ ++ err = do_rename(a); ++ if (unlikely(err)) ++ goto out_dt; ++ ++ /* update dir attributes */ ++ au_ren_refresh_dir(a); ++ ++ /* dput/iput all lower dentries */ ++ au_ren_refresh(a); ++ ++ goto out_hdir; /* success */ ++ ++ out_dt: ++ au_ren_rev_dt(err, a); ++ out_hdir: ++ au_ren_unlock(a); ++ out_children: ++ au_nhash_wh_free(&a->whlist); ++ out_unlock: ++ if (unlikely(err && au_ftest_ren(a->flags, ISDIR))) { ++ au_update_dbstart(a->dst_dentry); ++ d_drop(a->dst_dentry); ++ } ++ if (!err) ++ d_move(a->src_dentry, a->dst_dentry); ++ if (au_ftest_ren(a->flags, ISSAMEDIR)) ++ di_write_unlock(a->dst_parent); ++ else ++ di_write_unlock2(a->src_parent, a->dst_parent); ++ aufs_read_and_write_unlock2(a->dst_dentry, a->src_dentry); ++ out_free: ++ iput(a->dst_inode); ++ if (a->thargs) ++ au_whtmp_rmdir_free(a->thargs); ++ kfree(a); ++ out: ++ return err; ++} +diff -Nur linux-2.6.31.4.orig/fs/aufs/Kconfig linux-2.6.31.4/fs/aufs/Kconfig +--- linux-2.6.31.4.orig/fs/aufs/Kconfig 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/Kconfig 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,132 @@ ++config AUFS_FS ++ tristate "Aufs (Advanced multi layered unification filesystem) support" ++ depends on EXPERIMENTAL ++ help ++ Aufs is a stackable unification filesystem such as Unionfs, ++ which unifies several directories and provides a merged single ++ directory. ++ In the early days, aufs was entirely re-designed and ++ re-implemented Unionfs Version 1.x series. Introducing many ++ original ideas, approaches and improvements, it becomes totally ++ different from Unionfs while keeping the basic features. ++ ++if AUFS_FS ++choice ++ prompt "Maximum number of branches" ++ default AUFS_BRANCH_MAX_127 ++ help ++ Specifies the maximum number of branches (or member directories) ++ in a single aufs. The larger value consumes more system ++ resources and has a minor impact to performance. ++config AUFS_BRANCH_MAX_127 ++ bool "127" ++ help ++ Specifies the maximum number of branches (or member directories) ++ in a single aufs. The larger value consumes more system ++ resources and has a minor impact to performance. ++config AUFS_BRANCH_MAX_511 ++ bool "511" ++ help ++ Specifies the maximum number of branches (or member directories) ++ in a single aufs. The larger value consumes more system ++ resources and has a minor impact to performance. ++config AUFS_BRANCH_MAX_1023 ++ bool "1023" ++ help ++ Specifies the maximum number of branches (or member directories) ++ in a single aufs. The larger value consumes more system ++ resources and has a minor impact to performance. ++config AUFS_BRANCH_MAX_32767 ++ bool "32767" ++ help ++ Specifies the maximum number of branches (or member directories) ++ in a single aufs. The larger value consumes more system ++ resources and has a minor impact to performance. ++endchoice ++ ++config AUFS_HINOTIFY ++ bool "Use inotify to detect actions on a branch" ++ depends on INOTIFY ++ help ++ If you want to modify files on branches directly, eg. bypassing aufs, ++ and want aufs to detect the changes of them fully, then enable this ++ option and use 'udba=inotify' mount option. ++ It will have a negative impact to the performance. ++ See detail in aufs.5. ++ ++config AUFS_EXPORT ++ bool "NFS-exportable aufs" ++ depends on (AUFS_FS = y && EXPORTFS = y) || (AUFS_FS = m && EXPORTFS) ++ help ++ If you want to export your mounted aufs via NFS, then enable this ++ option. There are several requirements for this configuration. ++ See detail in aufs.5. ++ ++config AUFS_SHWH ++ bool "Show whiteouts" ++ help ++ If you want to make the whiteouts in aufs visible, then enable ++ this option and specify 'shwh' mount option. Although it may ++ sounds like philosophy or something, but in technically it ++ simply shows the name of whiteout with keeping its behaviour. ++ ++config AUFS_BR_RAMFS ++ bool "Ramfs (initramfs/rootfs) as an aufs branch" ++ help ++ If you want to use ramfs as an aufs branch fs, then enable this ++ option. Generally tmpfs is recommended. ++ Aufs prohibited them to be a branch fs by default, because ++ initramfs becomes unusable after switch_root or something ++ generally. If you sets initramfs as an aufs branch and boot your ++ system by switch_root, you will meet a problem easily since the ++ files in initramfs may be inaccessible. ++ Unless you are going to use ramfs as an aufs branch fs without ++ switch_root or something, leave it N. ++ ++config AUFS_BR_FUSE ++ bool "Fuse fs as an aufs branch" ++ depends on FUSE_FS ++ select AUFS_POLL ++ help ++ If you want to use fuse-based userspace filesystem as an aufs ++ branch fs, then enable this option. ++ It implements the internal poll(2) operation which is ++ implemented by fuse only (curretnly). ++ ++config AUFS_DEBUG ++ bool "Debug aufs" ++ help ++ Enable this to compile aufs internal debug code. ++ It will have a negative impact to the performance. ++ ++config AUFS_MAGIC_SYSRQ ++ bool ++ depends on AUFS_DEBUG && MAGIC_SYSRQ ++ default y ++ help ++ Automatic configuration for internal use. ++ When aufs supports Magic SysRq, enabled automatically. ++ ++config AUFS_BDEV_LOOP ++ bool ++ depends on BLK_DEV_LOOP ++ default y ++ help ++ Automatic configuration for internal use. ++ Convert =[ym] into =y. ++ ++config AUFS_INO_T_64 ++ bool ++ depends on AUFS_EXPORT ++ depends on 64BIT && !(ALPHA || S390) ++ default y ++ help ++ Automatic configuration for internal use. ++ /* typedef unsigned long/int __kernel_ino_t */ ++ /* alpha and s390x are int */ ++ ++config AUFS_POLL ++ bool ++ help ++ Automatic configuration for internal use. ++endif +diff -Nur linux-2.6.31.4.orig/fs/aufs/loop.c linux-2.6.31.4/fs/aufs/loop.c +--- linux-2.6.31.4.orig/fs/aufs/loop.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/loop.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,55 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * support for loopback block device as a branch ++ */ ++ ++#include <linux/loop.h> ++#include "aufs.h" ++ ++/* ++ * test if two lower dentries have overlapping branches. ++ */ ++int au_test_loopback_overlap(struct super_block *sb, struct dentry *h_d1, ++ struct dentry *h_d2) ++{ ++ struct inode *h_inode; ++ struct loop_device *l; ++ ++ h_inode = h_d1->d_inode; ++ if (MAJOR(h_inode->i_sb->s_dev) != LOOP_MAJOR) ++ return 0; ++ ++ l = h_inode->i_sb->s_bdev->bd_disk->private_data; ++ h_d1 = l->lo_backing_file->f_dentry; ++ /* h_d1 can be local NFS. in this case aufs cannot detect the loop */ ++ if (unlikely(h_d1->d_sb == sb)) ++ return 1; ++ return !!au_test_subdir(h_d1, h_d2); ++} ++ ++/* true if a kernel thread named 'loop[0-9].*' accesses a file */ ++int au_test_loopback_kthread(void) ++{ ++ const char c = current->comm[4]; ++ ++ return current->mm == NULL ++ && '0' <= c && c <= '9' ++ && strncmp(current->comm, "loop", 4) == 0; ++} +diff -Nur linux-2.6.31.4.orig/fs/aufs/loop.h linux-2.6.31.4/fs/aufs/loop.h +--- linux-2.6.31.4.orig/fs/aufs/loop.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/loop.h 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,51 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * support for loopback mount as a branch ++ */ ++ ++#ifndef __AUFS_LOOP_H__ ++#define __AUFS_LOOP_H__ ++ ++#ifdef __KERNEL__ ++ ++struct dentry; ++struct super_block; ++ ++#ifdef CONFIG_AUFS_BDEV_LOOP ++/* loop.c */ ++int au_test_loopback_overlap(struct super_block *sb, struct dentry *h_d1, ++ struct dentry *h_d2); ++int au_test_loopback_kthread(void); ++#else ++static inline ++int au_test_loopback_overlap(struct super_block *sb, struct dentry *h_d1, ++ struct dentry *h_d2) ++{ ++ return 0; ++} ++ ++static inline int au_test_loopback_kthread(void) ++{ ++ return 0; ++} ++#endif /* BLK_DEV_LOOP */ ++ ++#endif /* __KERNEL__ */ ++#endif /* __AUFS_LOOP_H__ */ +diff -Nur linux-2.6.31.4.orig/fs/aufs/magic.mk linux-2.6.31.4/fs/aufs/magic.mk +--- linux-2.6.31.4.orig/fs/aufs/magic.mk 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/magic.mk 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,52 @@ ++ ++# defined in ${srctree}/fs/fuse/inode.c ++# tristate ++ifdef CONFIG_FUSE_FS ++ccflags-y += -DFUSE_SUPER_MAGIC=0x65735546 ++endif ++ ++# defined in ${srctree}/fs/ocfs2/ocfs2_fs.h ++# tristate ++ifdef CONFIG_OCFS2_FS ++ccflags-y += -DOCFS2_SUPER_MAGIC=0x7461636f ++endif ++ ++# defined in ${srctree}/fs/ocfs2/dlm/userdlm.h ++# tristate ++ifdef CONFIG_OCFS2_FS_O2CB ++ccflags-y += -DDLMFS_MAGIC=0x76a9f425 ++endif ++ ++# defined in ${srctree}/fs/ramfs/inode.c ++# always true ++ccflags-y += -DRAMFS_MAGIC=0x858458f6 ++ ++# defined in ${srctree}/fs/cifs/cifsfs.c ++# tristate ++ifdef CONFIG_CIFS_FS ++ccflags-y += -DCIFS_MAGIC_NUMBER=0xFF534D42 ++endif ++ ++# defined in ${srctree}/fs/xfs/xfs_sb.h ++# tristate ++ifdef CONFIG_XFS_FS ++ccflags-y += -DXFS_SB_MAGIC=0x58465342 ++endif ++ ++# defined in ${srctree}/fs/configfs/mount.c ++# tristate ++ifdef CONFIG_CONFIGFS_FS ++ccflags-y += -DCONFIGFS_MAGIC=0x62656570 ++endif ++ ++# defined in ${srctree}/fs/9p/v9fs.h ++# tristate ++ifdef CONFIG_9P_FS ++ccflags-y += -DV9FS_MAGIC=0x01021997 ++endif ++ ++# defined in ${srctree}/fs/ubifs/ubifs.h ++# tristate ++ifdef CONFIG_UBIFS_FS ++ccflags-y += -DUBIFS_SUPER_MAGIC=0x24051905 ++endif +diff -Nur linux-2.6.31.4.orig/fs/aufs/Makefile linux-2.6.31.4/fs/aufs/Makefile +--- linux-2.6.31.4.orig/fs/aufs/Makefile 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/Makefile 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,23 @@ ++ ++include ${src}/magic.mk ++-include ${src}/priv_def.mk ++ ++obj-$(CONFIG_AUFS_FS) += aufs.o ++aufs-y := module.o sbinfo.o super.o branch.o xino.o sysaufs.o opts.o \ ++ wkq.o vfsub.o dcsub.o \ ++ cpup.o whout.o plink.o wbr_policy.o \ ++ dinfo.o dentry.o \ ++ finfo.o file.o f_op.o \ ++ dir.o vdir.o \ ++ iinfo.o inode.o i_op.o i_op_add.o i_op_del.o i_op_ren.o \ ++ ioctl.o ++ ++# all are boolean ++aufs-$(CONFIG_SYSFS) += sysfs.o ++aufs-$(CONFIG_DEBUG_FS) += dbgaufs.o ++aufs-$(CONFIG_AUFS_BDEV_LOOP) += loop.o ++aufs-$(CONFIG_AUFS_HINOTIFY) += hinotify.o ++aufs-$(CONFIG_AUFS_EXPORT) += export.o ++aufs-$(CONFIG_AUFS_POLL) += poll.o ++aufs-$(CONFIG_AUFS_DEBUG) += debug.o ++aufs-$(CONFIG_AUFS_MAGIC_SYSRQ) += sysrq.o +diff -Nur linux-2.6.31.4.orig/fs/aufs/module.c linux-2.6.31.4/fs/aufs/module.c +--- linux-2.6.31.4.orig/fs/aufs/module.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/module.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,173 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * module global variables and operations ++ */ ++ ++#include <linux/module.h> ++#include <linux/seq_file.h> ++#include "aufs.h" ++ ++void *au_kzrealloc(void *p, unsigned int nused, unsigned int new_sz, gfp_t gfp) ++{ ++ if (new_sz <= nused) ++ return p; ++ ++ p = krealloc(p, new_sz, gfp); ++ if (p) ++ memset(p + nused, 0, new_sz - nused); ++ return p; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* ++ * aufs caches ++ */ ++struct kmem_cache *au_cachep[AuCache_Last]; ++static int __init au_cache_init(void) ++{ ++ au_cachep[AuCache_DINFO] = AuCache(au_dinfo); ++ if (au_cachep[AuCache_DINFO]) ++ au_cachep[AuCache_ICNTNR] = AuCache(au_icntnr); ++ if (au_cachep[AuCache_ICNTNR]) ++ au_cachep[AuCache_FINFO] = AuCache(au_finfo); ++ if (au_cachep[AuCache_FINFO]) ++ au_cachep[AuCache_VDIR] = AuCache(au_vdir); ++ if (au_cachep[AuCache_VDIR]) ++ au_cachep[AuCache_DEHSTR] = AuCache(au_vdir_dehstr); ++ if (au_cachep[AuCache_DEHSTR]) ++ return 0; ++ ++ return -ENOMEM; ++} ++ ++static void au_cache_fin(void) ++{ ++ int i; ++ for (i = 0; i < AuCache_Last; i++) ++ if (au_cachep[i]) { ++ kmem_cache_destroy(au_cachep[i]); ++ au_cachep[i] = NULL; ++ } ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++int au_dir_roflags; ++ ++/* ++ * functions for module interface. ++ */ ++MODULE_LICENSE("GPL"); ++/* MODULE_LICENSE("GPL v2"); */ ++MODULE_AUTHOR("Junjiro R. Okajima <aufs-users@lists.sourceforge.net>"); ++MODULE_DESCRIPTION(AUFS_NAME ++ " -- Advanced multi layered unification filesystem"); ++MODULE_VERSION(AUFS_VERSION); ++ ++/* it should be 'byte', but param_set_byte() prints it by "%c" */ ++short aufs_nwkq = AUFS_NWKQ_DEF; ++MODULE_PARM_DESC(nwkq, "the number of workqueue thread, " AUFS_WKQ_NAME); ++module_param_named(nwkq, aufs_nwkq, short, S_IRUGO); ++ ++/* this module parameter has no meaning when SYSFS is disabled */ ++int sysaufs_brs = 1; ++MODULE_PARM_DESC(brs, "use <sysfs>/fs/aufs/si_*/brN"); ++module_param_named(brs, sysaufs_brs, int, S_IRUGO); ++ ++/* ---------------------------------------------------------------------- */ ++ ++static char au_esc_chars[0x20 + 3]; /* 0x01-0x20, backslash, del, and NULL */ ++ ++int au_seq_path(struct seq_file *seq, struct path *path) ++{ ++ return seq_path(seq, path, au_esc_chars); ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static int __init aufs_init(void) ++{ ++ int err, i; ++ char *p; ++ ++ p = au_esc_chars; ++ for (i = 1; i <= ' '; i++) ++ *p++ = i; ++ *p++ = '\\'; ++ *p++ = '\x7f'; ++ *p = 0; ++ ++ au_dir_roflags = au_file_roflags(O_DIRECTORY | O_LARGEFILE); ++ ++ sysaufs_brs_init(); ++ au_debug_init(); ++ ++ err = -EINVAL; ++ if (unlikely(aufs_nwkq <= 0)) ++ goto out; ++ ++ err = sysaufs_init(); ++ if (unlikely(err)) ++ goto out; ++ err = au_wkq_init(); ++ if (unlikely(err)) ++ goto out_sysaufs; ++ err = au_hinotify_init(); ++ if (unlikely(err)) ++ goto out_wkq; ++ err = au_sysrq_init(); ++ if (unlikely(err)) ++ goto out_hin; ++ err = au_cache_init(); ++ if (unlikely(err)) ++ goto out_sysrq; ++ err = register_filesystem(&aufs_fs_type); ++ if (unlikely(err)) ++ goto out_cache; ++ pr_info(AUFS_NAME " " AUFS_VERSION "\n"); ++ goto out; /* success */ ++ ++ out_cache: ++ au_cache_fin(); ++ out_sysrq: ++ au_sysrq_fin(); ++ out_hin: ++ au_hinotify_fin(); ++ out_wkq: ++ au_wkq_fin(); ++ out_sysaufs: ++ sysaufs_fin(); ++ out: ++ return err; ++} ++ ++static void __exit aufs_exit(void) ++{ ++ unregister_filesystem(&aufs_fs_type); ++ au_cache_fin(); ++ au_sysrq_fin(); ++ au_hinotify_fin(); ++ au_wkq_fin(); ++ sysaufs_fin(); ++} ++ ++module_init(aufs_init); ++module_exit(aufs_exit); +diff -Nur linux-2.6.31.4.orig/fs/aufs/module.h linux-2.6.31.4/fs/aufs/module.h +--- linux-2.6.31.4.orig/fs/aufs/module.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/module.h 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,78 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * module initialization and module-global ++ */ ++ ++#ifndef __AUFS_MODULE_H__ ++#define __AUFS_MODULE_H__ ++ ++#ifdef __KERNEL__ ++ ++#include <linux/slab.h> ++ ++struct path; ++struct seq_file; ++ ++/* module parameters */ ++extern short aufs_nwkq; ++extern int sysaufs_brs; ++ ++/* ---------------------------------------------------------------------- */ ++ ++extern int au_dir_roflags; ++ ++void *au_kzrealloc(void *p, unsigned int nused, unsigned int new_sz, gfp_t gfp); ++int au_seq_path(struct seq_file *seq, struct path *path); ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* kmem cache */ ++enum { ++ AuCache_DINFO, ++ AuCache_ICNTNR, ++ AuCache_FINFO, ++ AuCache_VDIR, ++ AuCache_DEHSTR, ++#ifdef CONFIG_AUFS_HINOTIFY ++ AuCache_HINOTIFY, ++#endif ++ AuCache_Last ++}; ++ ++#define AuCache(type) KMEM_CACHE(type, SLAB_RECLAIM_ACCOUNT) ++ ++extern struct kmem_cache *au_cachep[]; ++ ++#define AuCacheFuncs(name, index) \ ++static inline void *au_cache_alloc_##name(void) \ ++{ return kmem_cache_alloc(au_cachep[AuCache_##index], GFP_NOFS); } \ ++static inline void au_cache_free_##name(void *p) \ ++{ kmem_cache_free(au_cachep[AuCache_##index], p); } ++ ++AuCacheFuncs(dinfo, DINFO); ++AuCacheFuncs(icntnr, ICNTNR); ++AuCacheFuncs(finfo, FINFO); ++AuCacheFuncs(vdir, VDIR); ++AuCacheFuncs(dehstr, DEHSTR); ++ ++/* ---------------------------------------------------------------------- */ ++ ++#endif /* __KERNEL__ */ ++#endif /* __AUFS_MODULE_H__ */ +diff -Nur linux-2.6.31.4.orig/fs/aufs/opts.c linux-2.6.31.4/fs/aufs/opts.c +--- linux-2.6.31.4.orig/fs/aufs/opts.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/opts.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,1543 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * mount options/flags ++ */ ++ ++#include <linux/file.h> ++#include <linux/namei.h> ++#include <linux/types.h> /* a distribution requires */ ++#include <linux/parser.h> ++#include "aufs.h" ++ ++/* ---------------------------------------------------------------------- */ ++ ++enum { ++ Opt_br, ++ Opt_add, Opt_del, Opt_mod, Opt_reorder, Opt_append, Opt_prepend, ++ Opt_idel, Opt_imod, Opt_ireorder, ++ Opt_dirwh, Opt_rdcache, Opt_rdblk, Opt_rdhash, Opt_rendir, ++ Opt_rdblk_def, Opt_rdhash_def, ++ Opt_xino, Opt_zxino, Opt_noxino, ++ Opt_trunc_xino, Opt_trunc_xino_v, Opt_notrunc_xino, ++ Opt_trunc_xino_path, Opt_itrunc_xino, ++ Opt_trunc_xib, Opt_notrunc_xib, ++ Opt_shwh, Opt_noshwh, ++ Opt_plink, Opt_noplink, Opt_list_plink, ++ Opt_udba, ++ /* Opt_lock, Opt_unlock, */ ++ Opt_cmd, Opt_cmd_args, ++ Opt_diropq_a, Opt_diropq_w, ++ Opt_warn_perm, Opt_nowarn_perm, ++ Opt_wbr_copyup, Opt_wbr_create, ++ Opt_refrof, Opt_norefrof, ++ Opt_verbose, Opt_noverbose, ++ Opt_sum, Opt_nosum, Opt_wsum, ++ Opt_tail, Opt_ignore, Opt_ignore_silent, Opt_err ++}; ++ ++static match_table_t options = { ++ {Opt_br, "br=%s"}, ++ {Opt_br, "br:%s"}, ++ ++ {Opt_add, "add=%d:%s"}, ++ {Opt_add, "add:%d:%s"}, ++ {Opt_add, "ins=%d:%s"}, ++ {Opt_add, "ins:%d:%s"}, ++ {Opt_append, "append=%s"}, ++ {Opt_append, "append:%s"}, ++ {Opt_prepend, "prepend=%s"}, ++ {Opt_prepend, "prepend:%s"}, ++ ++ {Opt_del, "del=%s"}, ++ {Opt_del, "del:%s"}, ++ /* {Opt_idel, "idel:%d"}, */ ++ {Opt_mod, "mod=%s"}, ++ {Opt_mod, "mod:%s"}, ++ /* {Opt_imod, "imod:%d:%s"}, */ ++ ++ {Opt_dirwh, "dirwh=%d"}, ++ ++ {Opt_xino, "xino=%s"}, ++ {Opt_noxino, "noxino"}, ++ {Opt_trunc_xino, "trunc_xino"}, ++ {Opt_trunc_xino_v, "trunc_xino_v=%d:%d"}, ++ {Opt_notrunc_xino, "notrunc_xino"}, ++ {Opt_trunc_xino_path, "trunc_xino=%s"}, ++ {Opt_itrunc_xino, "itrunc_xino=%d"}, ++ /* {Opt_zxino, "zxino=%s"}, */ ++ {Opt_trunc_xib, "trunc_xib"}, ++ {Opt_notrunc_xib, "notrunc_xib"}, ++ ++ {Opt_plink, "plink"}, ++ {Opt_noplink, "noplink"}, ++#ifdef CONFIG_AUFS_DEBUG ++ {Opt_list_plink, "list_plink"}, ++#endif ++ ++ {Opt_udba, "udba=%s"}, ++ ++ {Opt_diropq_a, "diropq=always"}, ++ {Opt_diropq_a, "diropq=a"}, ++ {Opt_diropq_w, "diropq=whiteouted"}, ++ {Opt_diropq_w, "diropq=w"}, ++ ++ {Opt_warn_perm, "warn_perm"}, ++ {Opt_nowarn_perm, "nowarn_perm"}, ++ ++ /* keep them temporary */ ++ {Opt_ignore_silent, "coo=%s"}, ++ {Opt_ignore_silent, "nodlgt"}, ++ {Opt_ignore_silent, "nodirperm1"}, ++ {Opt_ignore_silent, "clean_plink"}, ++ ++#ifdef CONFIG_AUFS_SHWH ++ {Opt_shwh, "shwh"}, ++#endif ++ {Opt_noshwh, "noshwh"}, ++ ++ {Opt_rendir, "rendir=%d"}, ++ ++ {Opt_refrof, "refrof"}, ++ {Opt_norefrof, "norefrof"}, ++ ++ {Opt_verbose, "verbose"}, ++ {Opt_verbose, "v"}, ++ {Opt_noverbose, "noverbose"}, ++ {Opt_noverbose, "quiet"}, ++ {Opt_noverbose, "q"}, ++ {Opt_noverbose, "silent"}, ++ ++ {Opt_sum, "sum"}, ++ {Opt_nosum, "nosum"}, ++ {Opt_wsum, "wsum"}, ++ ++ {Opt_rdcache, "rdcache=%d"}, ++ {Opt_rdblk, "rdblk=%d"}, ++ {Opt_rdblk_def, "rdblk=def"}, ++ {Opt_rdhash, "rdhash=%d"}, ++ {Opt_rdhash_def, "rdhash=def"}, ++ ++ {Opt_wbr_create, "create=%s"}, ++ {Opt_wbr_create, "create_policy=%s"}, ++ {Opt_wbr_copyup, "cpup=%s"}, ++ {Opt_wbr_copyup, "copyup=%s"}, ++ {Opt_wbr_copyup, "copyup_policy=%s"}, ++ ++ /* internal use for the scripts */ ++ {Opt_ignore_silent, "si=%s"}, ++ ++ {Opt_br, "dirs=%s"}, ++ {Opt_ignore, "debug=%d"}, ++ {Opt_ignore, "delete=whiteout"}, ++ {Opt_ignore, "delete=all"}, ++ {Opt_ignore, "imap=%s"}, ++ ++ {Opt_err, NULL} ++}; ++ ++/* ---------------------------------------------------------------------- */ ++ ++static const char *au_parser_pattern(int val, struct match_token *token) ++{ ++ while (token->pattern) { ++ if (token->token == val) ++ return token->pattern; ++ token++; ++ } ++ BUG(); ++ return "??"; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static match_table_t brperms = { ++ {AuBrPerm_RO, AUFS_BRPERM_RO}, ++ {AuBrPerm_RR, AUFS_BRPERM_RR}, ++ {AuBrPerm_RW, AUFS_BRPERM_RW}, ++ ++ {AuBrPerm_ROWH, AUFS_BRPERM_ROWH}, ++ {AuBrPerm_RRWH, AUFS_BRPERM_RRWH}, ++ {AuBrPerm_RWNoLinkWH, AUFS_BRPERM_RWNLWH}, ++ ++ {AuBrPerm_ROWH, "nfsro"}, ++ {AuBrPerm_RO, NULL} ++}; ++ ++static int br_perm_val(char *perm) ++{ ++ int val; ++ substring_t args[MAX_OPT_ARGS]; ++ ++ val = match_token(perm, brperms, args); ++ return val; ++} ++ ++const char *au_optstr_br_perm(int brperm) ++{ ++ return au_parser_pattern(brperm, (void *)brperms); ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static match_table_t udbalevel = { ++ {AuOpt_UDBA_REVAL, "reval"}, ++ {AuOpt_UDBA_NONE, "none"}, ++#ifdef CONFIG_AUFS_HINOTIFY ++ {AuOpt_UDBA_HINOTIFY, "inotify"}, ++#endif ++ {-1, NULL} ++}; ++ ++static int udba_val(char *str) ++{ ++ substring_t args[MAX_OPT_ARGS]; ++ ++ return match_token(str, udbalevel, args); ++} ++ ++const char *au_optstr_udba(int udba) ++{ ++ return au_parser_pattern(udba, (void *)udbalevel); ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static match_table_t au_wbr_create_policy = { ++ {AuWbrCreate_TDP, "tdp"}, ++ {AuWbrCreate_TDP, "top-down-parent"}, ++ {AuWbrCreate_RR, "rr"}, ++ {AuWbrCreate_RR, "round-robin"}, ++ {AuWbrCreate_MFS, "mfs"}, ++ {AuWbrCreate_MFS, "most-free-space"}, ++ {AuWbrCreate_MFSV, "mfs:%d"}, ++ {AuWbrCreate_MFSV, "most-free-space:%d"}, ++ ++ {AuWbrCreate_MFSRR, "mfsrr:%d"}, ++ {AuWbrCreate_MFSRRV, "mfsrr:%d:%d"}, ++ {AuWbrCreate_PMFS, "pmfs"}, ++ {AuWbrCreate_PMFSV, "pmfs:%d"}, ++ ++ {-1, NULL} ++}; ++ ++/* ++ * cf. linux/lib/parser.c and cmdline.c ++ * gave up calling memparse() since it uses simple_strtoull() instead of ++ * strict_...(). ++ */ ++static int au_match_ull(substring_t *s, unsigned long long *result) ++{ ++ int err; ++ unsigned int len; ++ char a[32]; ++ ++ err = -ERANGE; ++ len = s->to - s->from; ++ if (len + 1 <= sizeof(a)) { ++ memcpy(a, s->from, len); ++ a[len] = '\0'; ++ err = strict_strtoull(a, 0, result); ++ } ++ return err; ++} ++ ++static int au_wbr_mfs_wmark(substring_t *arg, char *str, ++ struct au_opt_wbr_create *create) ++{ ++ int err; ++ unsigned long long ull; ++ ++ err = 0; ++ if (!au_match_ull(arg, &ull)) ++ create->mfsrr_watermark = ull; ++ else { ++ AuErr("bad integer in %s\n", str); ++ err = -EINVAL; ++ } ++ ++ return err; ++} ++ ++static int au_wbr_mfs_sec(substring_t *arg, char *str, ++ struct au_opt_wbr_create *create) ++{ ++ int n, err; ++ ++ err = 0; ++ if (!match_int(arg, &n) && 0 <= n) ++ create->mfs_second = n; ++ else { ++ AuErr("bad integer in %s\n", str); ++ err = -EINVAL; ++ } ++ ++ return err; ++} ++ ++static int au_wbr_create_val(char *str, struct au_opt_wbr_create *create) ++{ ++ int err, e; ++ substring_t args[MAX_OPT_ARGS]; ++ ++ err = match_token(str, au_wbr_create_policy, args); ++ create->wbr_create = err; ++ switch (err) { ++ case AuWbrCreate_MFSRRV: ++ e = au_wbr_mfs_wmark(&args[0], str, create); ++ if (!e) ++ e = au_wbr_mfs_sec(&args[1], str, create); ++ if (unlikely(e)) ++ err = e; ++ break; ++ case AuWbrCreate_MFSRR: ++ e = au_wbr_mfs_wmark(&args[0], str, create); ++ if (unlikely(e)) { ++ err = e; ++ break; ++ } ++ /*FALLTHROUGH*/ ++ case AuWbrCreate_MFS: ++ case AuWbrCreate_PMFS: ++ create->mfs_second = AUFS_MFS_SECOND_DEF; ++ break; ++ case AuWbrCreate_MFSV: ++ case AuWbrCreate_PMFSV: ++ e = au_wbr_mfs_sec(&args[0], str, create); ++ if (unlikely(e)) ++ err = e; ++ break; ++ } ++ ++ return err; ++} ++ ++const char *au_optstr_wbr_create(int wbr_create) ++{ ++ return au_parser_pattern(wbr_create, (void *)au_wbr_create_policy); ++} ++ ++static match_table_t au_wbr_copyup_policy = { ++ {AuWbrCopyup_TDP, "tdp"}, ++ {AuWbrCopyup_TDP, "top-down-parent"}, ++ {AuWbrCopyup_BUP, "bup"}, ++ {AuWbrCopyup_BUP, "bottom-up-parent"}, ++ {AuWbrCopyup_BU, "bu"}, ++ {AuWbrCopyup_BU, "bottom-up"}, ++ {-1, NULL} ++}; ++ ++static int au_wbr_copyup_val(char *str) ++{ ++ substring_t args[MAX_OPT_ARGS]; ++ ++ return match_token(str, au_wbr_copyup_policy, args); ++} ++ ++const char *au_optstr_wbr_copyup(int wbr_copyup) ++{ ++ return au_parser_pattern(wbr_copyup, (void *)au_wbr_copyup_policy); ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static const int lkup_dirflags = LOOKUP_FOLLOW | LOOKUP_DIRECTORY; ++ ++static void dump_opts(struct au_opts *opts) ++{ ++#ifdef CONFIG_AUFS_DEBUG ++ /* reduce stack space */ ++ union { ++ struct au_opt_add *add; ++ struct au_opt_del *del; ++ struct au_opt_mod *mod; ++ struct au_opt_xino *xino; ++ struct au_opt_xino_itrunc *xino_itrunc; ++ struct au_opt_wbr_create *create; ++ } u; ++ struct au_opt *opt; ++ ++ opt = opts->opt; ++ while (opt->type != Opt_tail) { ++ switch (opt->type) { ++ case Opt_add: ++ u.add = &opt->add; ++ AuDbg("add {b%d, %s, 0x%x, %p}\n", ++ u.add->bindex, u.add->pathname, u.add->perm, ++ u.add->path.dentry); ++ break; ++ case Opt_del: ++ case Opt_idel: ++ u.del = &opt->del; ++ AuDbg("del {%s, %p}\n", ++ u.del->pathname, u.del->h_path.dentry); ++ break; ++ case Opt_mod: ++ case Opt_imod: ++ u.mod = &opt->mod; ++ AuDbg("mod {%s, 0x%x, %p}\n", ++ u.mod->path, u.mod->perm, u.mod->h_root); ++ break; ++ case Opt_append: ++ u.add = &opt->add; ++ AuDbg("append {b%d, %s, 0x%x, %p}\n", ++ u.add->bindex, u.add->pathname, u.add->perm, ++ u.add->path.dentry); ++ break; ++ case Opt_prepend: ++ u.add = &opt->add; ++ AuDbg("prepend {b%d, %s, 0x%x, %p}\n", ++ u.add->bindex, u.add->pathname, u.add->perm, ++ u.add->path.dentry); ++ break; ++ case Opt_dirwh: ++ AuDbg("dirwh %d\n", opt->dirwh); ++ break; ++ case Opt_rdcache: ++ AuDbg("rdcache %d\n", opt->rdcache); ++ break; ++ case Opt_rdblk: ++ AuDbg("rdblk %u\n", opt->rdblk); ++ break; ++ case Opt_rdblk_def: ++ AuDbg("rdblk_def\n"); ++ break; ++ case Opt_rdhash: ++ AuDbg("rdhash %u\n", opt->rdhash); ++ break; ++ case Opt_rdhash_def: ++ AuDbg("rdhash_def\n"); ++ break; ++ case Opt_xino: ++ u.xino = &opt->xino; ++ AuDbg("xino {%s %.*s}\n", ++ u.xino->path, ++ AuDLNPair(u.xino->file->f_dentry)); ++ break; ++ case Opt_trunc_xino: ++ AuLabel(trunc_xino); ++ break; ++ case Opt_notrunc_xino: ++ AuLabel(notrunc_xino); ++ break; ++ case Opt_trunc_xino_path: ++ case Opt_itrunc_xino: ++ u.xino_itrunc = &opt->xino_itrunc; ++ AuDbg("trunc_xino %d\n", u.xino_itrunc->bindex); ++ break; ++ ++ case Opt_noxino: ++ AuLabel(noxino); ++ break; ++ case Opt_trunc_xib: ++ AuLabel(trunc_xib); ++ break; ++ case Opt_notrunc_xib: ++ AuLabel(notrunc_xib); ++ break; ++ case Opt_shwh: ++ AuLabel(shwh); ++ break; ++ case Opt_noshwh: ++ AuLabel(noshwh); ++ break; ++ case Opt_plink: ++ AuLabel(plink); ++ break; ++ case Opt_noplink: ++ AuLabel(noplink); ++ break; ++ case Opt_list_plink: ++ AuLabel(list_plink); ++ break; ++ case Opt_udba: ++ AuDbg("udba %d, %s\n", ++ opt->udba, au_optstr_udba(opt->udba)); ++ break; ++ case Opt_diropq_a: ++ AuLabel(diropq_a); ++ break; ++ case Opt_diropq_w: ++ AuLabel(diropq_w); ++ break; ++ case Opt_warn_perm: ++ AuLabel(warn_perm); ++ break; ++ case Opt_nowarn_perm: ++ AuLabel(nowarn_perm); ++ break; ++ case Opt_refrof: ++ AuLabel(refrof); ++ break; ++ case Opt_norefrof: ++ AuLabel(norefrof); ++ break; ++ case Opt_verbose: ++ AuLabel(verbose); ++ break; ++ case Opt_noverbose: ++ AuLabel(noverbose); ++ break; ++ case Opt_sum: ++ AuLabel(sum); ++ break; ++ case Opt_nosum: ++ AuLabel(nosum); ++ break; ++ case Opt_wsum: ++ AuLabel(wsum); ++ break; ++ case Opt_wbr_create: ++ u.create = &opt->wbr_create; ++ AuDbg("create %d, %s\n", u.create->wbr_create, ++ au_optstr_wbr_create(u.create->wbr_create)); ++ switch (u.create->wbr_create) { ++ case AuWbrCreate_MFSV: ++ case AuWbrCreate_PMFSV: ++ AuDbg("%d sec\n", u.create->mfs_second); ++ break; ++ case AuWbrCreate_MFSRR: ++ AuDbg("%llu watermark\n", ++ u.create->mfsrr_watermark); ++ break; ++ case AuWbrCreate_MFSRRV: ++ AuDbg("%llu watermark, %d sec\n", ++ u.create->mfsrr_watermark, ++ u.create->mfs_second); ++ break; ++ } ++ break; ++ case Opt_wbr_copyup: ++ AuDbg("copyup %d, %s\n", opt->wbr_copyup, ++ au_optstr_wbr_copyup(opt->wbr_copyup)); ++ break; ++ default: ++ BUG(); ++ } ++ opt++; ++ } ++#endif ++} ++ ++void au_opts_free(struct au_opts *opts) ++{ ++ struct au_opt *opt; ++ ++ opt = opts->opt; ++ while (opt->type != Opt_tail) { ++ switch (opt->type) { ++ case Opt_add: ++ case Opt_append: ++ case Opt_prepend: ++ path_put(&opt->add.path); ++ break; ++ case Opt_del: ++ case Opt_idel: ++ path_put(&opt->del.h_path); ++ break; ++ case Opt_mod: ++ case Opt_imod: ++ dput(opt->mod.h_root); ++ break; ++ case Opt_xino: ++ fput(opt->xino.file); ++ break; ++ } ++ opt++; ++ } ++} ++ ++static int opt_add(struct au_opt *opt, char *opt_str, unsigned long sb_flags, ++ aufs_bindex_t bindex) ++{ ++ int err; ++ struct au_opt_add *add = &opt->add; ++ char *p; ++ ++ add->bindex = bindex; ++ add->perm = AuBrPerm_Last; ++ add->pathname = opt_str; ++ p = strchr(opt_str, '='); ++ if (p) { ++ *p++ = 0; ++ if (*p) ++ add->perm = br_perm_val(p); ++ } ++ ++ err = vfsub_kern_path(add->pathname, lkup_dirflags, &add->path); ++ if (!err) { ++ if (!p) { ++ add->perm = AuBrPerm_RO; ++ if (au_test_fs_rr(add->path.dentry->d_sb)) ++ add->perm = AuBrPerm_RR; ++ else if (!bindex && !(sb_flags & MS_RDONLY)) ++ add->perm = AuBrPerm_RW; ++ } ++ opt->type = Opt_add; ++ goto out; ++ } ++ AuErr("lookup failed %s (%d)\n", add->pathname, err); ++ err = -EINVAL; ++ ++ out: ++ return err; ++} ++ ++static int au_opts_parse_del(struct au_opt_del *del, substring_t args[]) ++{ ++ int err; ++ ++ del->pathname = args[0].from; ++ AuDbg("del path %s\n", del->pathname); ++ ++ err = vfsub_kern_path(del->pathname, lkup_dirflags, &del->h_path); ++ if (unlikely(err)) ++ AuErr("lookup failed %s (%d)\n", del->pathname, err); ++ ++ return err; ++} ++ ++#if 0 /* reserved for future use */ ++static int au_opts_parse_idel(struct super_block *sb, aufs_bindex_t bindex, ++ struct au_opt_del *del, substring_t args[]) ++{ ++ int err; ++ struct dentry *root; ++ ++ err = -EINVAL; ++ root = sb->s_root; ++ aufs_read_lock(root, AuLock_FLUSH); ++ if (bindex < 0 || au_sbend(sb) < bindex) { ++ AuErr("out of bounds, %d\n", bindex); ++ goto out; ++ } ++ ++ err = 0; ++ del->h_path.dentry = dget(au_h_dptr(root, bindex)); ++ del->h_path.mnt = mntget(au_sbr_mnt(sb, bindex)); ++ ++ out: ++ aufs_read_unlock(root, !AuLock_IR); ++ return err; ++} ++#endif ++ ++static int au_opts_parse_mod(struct au_opt_mod *mod, substring_t args[]) ++{ ++ int err; ++ struct path path; ++ char *p; ++ ++ err = -EINVAL; ++ mod->path = args[0].from; ++ p = strchr(mod->path, '='); ++ if (unlikely(!p)) { ++ AuErr("no permssion %s\n", args[0].from); ++ goto out; ++ } ++ ++ *p++ = 0; ++ err = vfsub_kern_path(mod->path, lkup_dirflags, &path); ++ if (unlikely(err)) { ++ AuErr("lookup failed %s (%d)\n", mod->path, err); ++ goto out; ++ } ++ ++ mod->perm = br_perm_val(p); ++ AuDbg("mod path %s, perm 0x%x, %s\n", mod->path, mod->perm, p); ++ mod->h_root = dget(path.dentry); ++ path_put(&path); ++ ++ out: ++ return err; ++} ++ ++#if 0 /* reserved for future use */ ++static int au_opts_parse_imod(struct super_block *sb, aufs_bindex_t bindex, ++ struct au_opt_mod *mod, substring_t args[]) ++{ ++ int err; ++ struct dentry *root; ++ ++ err = -EINVAL; ++ root = sb->s_root; ++ aufs_read_lock(root, AuLock_FLUSH); ++ if (bindex < 0 || au_sbend(sb) < bindex) { ++ AuErr("out of bounds, %d\n", bindex); ++ goto out; ++ } ++ ++ err = 0; ++ mod->perm = br_perm_val(args[1].from); ++ AuDbg("mod path %s, perm 0x%x, %s\n", ++ mod->path, mod->perm, args[1].from); ++ mod->h_root = dget(au_h_dptr(root, bindex)); ++ ++ out: ++ aufs_read_unlock(root, !AuLock_IR); ++ return err; ++} ++#endif ++ ++static int au_opts_parse_xino(struct super_block *sb, struct au_opt_xino *xino, ++ substring_t args[]) ++{ ++ int err; ++ struct file *file; ++ ++ file = au_xino_create(sb, args[0].from, /*silent*/0); ++ err = PTR_ERR(file); ++ if (IS_ERR(file)) ++ goto out; ++ ++ err = -EINVAL; ++ if (unlikely(file->f_dentry->d_sb == sb)) { ++ fput(file); ++ AuErr("%s must be outside\n", args[0].from); ++ goto out; ++ } ++ ++ err = 0; ++ xino->file = file; ++ xino->path = args[0].from; ++ ++ out: ++ return err; ++} ++ ++static ++int au_opts_parse_xino_itrunc_path(struct super_block *sb, ++ struct au_opt_xino_itrunc *xino_itrunc, ++ substring_t args[]) ++{ ++ int err; ++ aufs_bindex_t bend, bindex; ++ struct path path; ++ struct dentry *root; ++ ++ err = vfsub_kern_path(args[0].from, lkup_dirflags, &path); ++ if (unlikely(err)) { ++ AuErr("lookup failed %s (%d)\n", args[0].from, err); ++ goto out; ++ } ++ ++ xino_itrunc->bindex = -1; ++ root = sb->s_root; ++ aufs_read_lock(root, AuLock_FLUSH); ++ bend = au_sbend(sb); ++ for (bindex = 0; bindex <= bend; bindex++) { ++ if (au_h_dptr(root, bindex) == path.dentry) { ++ xino_itrunc->bindex = bindex; ++ break; ++ } ++ } ++ aufs_read_unlock(root, !AuLock_IR); ++ path_put(&path); ++ ++ if (unlikely(xino_itrunc->bindex < 0)) { ++ AuErr("no such branch %s\n", args[0].from); ++ err = -EINVAL; ++ } ++ ++ out: ++ return err; ++} ++ ++/* called without aufs lock */ ++int au_opts_parse(struct super_block *sb, char *str, struct au_opts *opts) ++{ ++ int err, n, token; ++ aufs_bindex_t bindex; ++ unsigned char skipped; ++ struct dentry *root; ++ struct au_opt *opt, *opt_tail; ++ char *opt_str; ++ /* reduce the stack space */ ++ union { ++ struct au_opt_xino_itrunc *xino_itrunc; ++ struct au_opt_wbr_create *create; ++ } u; ++ struct { ++ substring_t args[MAX_OPT_ARGS]; ++ } *a; ++ ++ err = -ENOMEM; ++ a = kmalloc(sizeof(*a), GFP_NOFS); ++ if (unlikely(!a)) ++ goto out; ++ ++ root = sb->s_root; ++ err = 0; ++ bindex = 0; ++ opt = opts->opt; ++ opt_tail = opt + opts->max_opt - 1; ++ opt->type = Opt_tail; ++ while (!err && (opt_str = strsep(&str, ",")) && *opt_str) { ++ err = -EINVAL; ++ skipped = 0; ++ token = match_token(opt_str, options, a->args); ++ switch (token) { ++ case Opt_br: ++ err = 0; ++ while (!err && (opt_str = strsep(&a->args[0].from, ":")) ++ && *opt_str) { ++ err = opt_add(opt, opt_str, opts->sb_flags, ++ bindex++); ++ if (unlikely(!err && ++opt > opt_tail)) { ++ err = -E2BIG; ++ break; ++ } ++ opt->type = Opt_tail; ++ skipped = 1; ++ } ++ break; ++ case Opt_add: ++ if (unlikely(match_int(&a->args[0], &n))) { ++ AuErr("bad integer in %s\n", opt_str); ++ break; ++ } ++ bindex = n; ++ err = opt_add(opt, a->args[1].from, opts->sb_flags, ++ bindex); ++ if (!err) ++ opt->type = token; ++ break; ++ case Opt_append: ++ err = opt_add(opt, a->args[0].from, opts->sb_flags, ++ /*dummy bindex*/1); ++ if (!err) ++ opt->type = token; ++ break; ++ case Opt_prepend: ++ err = opt_add(opt, a->args[0].from, opts->sb_flags, ++ /*bindex*/0); ++ if (!err) ++ opt->type = token; ++ break; ++ case Opt_del: ++ err = au_opts_parse_del(&opt->del, a->args); ++ if (!err) ++ opt->type = token; ++ break; ++#if 0 /* reserved for future use */ ++ case Opt_idel: ++ del->pathname = "(indexed)"; ++ if (unlikely(match_int(&args[0], &n))) { ++ AuErr("bad integer in %s\n", opt_str); ++ break; ++ } ++ err = au_opts_parse_idel(sb, n, &opt->del, a->args); ++ if (!err) ++ opt->type = token; ++ break; ++#endif ++ case Opt_mod: ++ err = au_opts_parse_mod(&opt->mod, a->args); ++ if (!err) ++ opt->type = token; ++ break; ++#ifdef IMOD /* reserved for future use */ ++ case Opt_imod: ++ u.mod->path = "(indexed)"; ++ if (unlikely(match_int(&a->args[0], &n))) { ++ AuErr("bad integer in %s\n", opt_str); ++ break; ++ } ++ err = au_opts_parse_imod(sb, n, &opt->mod, a->args); ++ if (!err) ++ opt->type = token; ++ break; ++#endif ++ case Opt_xino: ++ err = au_opts_parse_xino(sb, &opt->xino, a->args); ++ if (!err) ++ opt->type = token; ++ break; ++ ++ case Opt_trunc_xino_path: ++ err = au_opts_parse_xino_itrunc_path ++ (sb, &opt->xino_itrunc, a->args); ++ if (!err) ++ opt->type = token; ++ break; ++ ++ case Opt_itrunc_xino: ++ u.xino_itrunc = &opt->xino_itrunc; ++ if (unlikely(match_int(&a->args[0], &n))) { ++ AuErr("bad integer in %s\n", opt_str); ++ break; ++ } ++ u.xino_itrunc->bindex = n; ++ aufs_read_lock(root, AuLock_FLUSH); ++ if (n < 0 || au_sbend(sb) < n) { ++ AuErr("out of bounds, %d\n", n); ++ aufs_read_unlock(root, !AuLock_IR); ++ break; ++ } ++ aufs_read_unlock(root, !AuLock_IR); ++ err = 0; ++ opt->type = token; ++ break; ++ ++ case Opt_dirwh: ++ if (unlikely(match_int(&a->args[0], &opt->dirwh))) ++ break; ++ err = 0; ++ opt->type = token; ++ break; ++ ++ case Opt_rdcache: ++ if (unlikely(match_int(&a->args[0], &opt->rdcache))) ++ break; ++ err = 0; ++ opt->type = token; ++ break; ++ case Opt_rdblk: ++ if (unlikely(match_int(&a->args[0], &n) ++ || n <= 0 ++ || n > KMALLOC_MAX_SIZE)) { ++ AuErr("bad integer in %s\n", opt_str); ++ break; ++ } ++ if (unlikely(n < NAME_MAX)) { ++ AuErr("rdblk must be larger than %d\n", ++ NAME_MAX); ++ break; ++ } ++ opt->rdblk = n; ++ err = 0; ++ opt->type = token; ++ break; ++ case Opt_rdhash: ++ if (unlikely(match_int(&a->args[0], &n) ++ || n <= 0 ++ || n * sizeof(struct hlist_head) ++ > KMALLOC_MAX_SIZE)) { ++ AuErr("bad integer in %s\n", opt_str); ++ break; ++ } ++ opt->rdhash = n; ++ err = 0; ++ opt->type = token; ++ break; ++ ++ case Opt_trunc_xino: ++ case Opt_notrunc_xino: ++ case Opt_noxino: ++ case Opt_trunc_xib: ++ case Opt_notrunc_xib: ++ case Opt_shwh: ++ case Opt_noshwh: ++ case Opt_plink: ++ case Opt_noplink: ++ case Opt_list_plink: ++ case Opt_diropq_a: ++ case Opt_diropq_w: ++ case Opt_warn_perm: ++ case Opt_nowarn_perm: ++ case Opt_refrof: ++ case Opt_norefrof: ++ case Opt_verbose: ++ case Opt_noverbose: ++ case Opt_sum: ++ case Opt_nosum: ++ case Opt_wsum: ++ case Opt_rdblk_def: ++ case Opt_rdhash_def: ++ err = 0; ++ opt->type = token; ++ break; ++ ++ case Opt_udba: ++ opt->udba = udba_val(a->args[0].from); ++ if (opt->udba >= 0) { ++ err = 0; ++ opt->type = token; ++ } else ++ AuErr("wrong value, %s\n", opt_str); ++ break; ++ ++ case Opt_wbr_create: ++ u.create = &opt->wbr_create; ++ u.create->wbr_create ++ = au_wbr_create_val(a->args[0].from, u.create); ++ if (u.create->wbr_create >= 0) { ++ err = 0; ++ opt->type = token; ++ } else ++ AuErr("wrong value, %s\n", opt_str); ++ break; ++ case Opt_wbr_copyup: ++ opt->wbr_copyup = au_wbr_copyup_val(a->args[0].from); ++ if (opt->wbr_copyup >= 0) { ++ err = 0; ++ opt->type = token; ++ } else ++ AuErr("wrong value, %s\n", opt_str); ++ break; ++ ++ case Opt_ignore: ++ AuWarn("ignored %s\n", opt_str); ++ /*FALLTHROUGH*/ ++ case Opt_ignore_silent: ++ skipped = 1; ++ err = 0; ++ break; ++ case Opt_err: ++ AuErr("unknown option %s\n", opt_str); ++ break; ++ } ++ ++ if (!err && !skipped) { ++ if (unlikely(++opt > opt_tail)) { ++ err = -E2BIG; ++ opt--; ++ opt->type = Opt_tail; ++ break; ++ } ++ opt->type = Opt_tail; ++ } ++ } ++ ++ kfree(a); ++ dump_opts(opts); ++ if (unlikely(err)) ++ au_opts_free(opts); ++ ++ out: ++ return err; ++} ++ ++static int au_opt_wbr_create(struct super_block *sb, ++ struct au_opt_wbr_create *create) ++{ ++ int err; ++ struct au_sbinfo *sbinfo; ++ ++ SiMustWriteLock(sb); ++ ++ err = 1; /* handled */ ++ sbinfo = au_sbi(sb); ++ if (sbinfo->si_wbr_create_ops->fin) { ++ err = sbinfo->si_wbr_create_ops->fin(sb); ++ if (!err) ++ err = 1; ++ } ++ ++ sbinfo->si_wbr_create = create->wbr_create; ++ sbinfo->si_wbr_create_ops = au_wbr_create_ops + create->wbr_create; ++ switch (create->wbr_create) { ++ case AuWbrCreate_MFSRRV: ++ case AuWbrCreate_MFSRR: ++ sbinfo->si_wbr_mfs.mfsrr_watermark = create->mfsrr_watermark; ++ /*FALLTHROUGH*/ ++ case AuWbrCreate_MFS: ++ case AuWbrCreate_MFSV: ++ case AuWbrCreate_PMFS: ++ case AuWbrCreate_PMFSV: ++ sbinfo->si_wbr_mfs.mfs_expire = create->mfs_second * HZ; ++ break; ++ } ++ ++ if (sbinfo->si_wbr_create_ops->init) ++ sbinfo->si_wbr_create_ops->init(sb); /* ignore */ ++ ++ return err; ++} ++ ++/* ++ * returns, ++ * plus: processed without an error ++ * zero: unprocessed ++ */ ++static int au_opt_simple(struct super_block *sb, struct au_opt *opt, ++ struct au_opts *opts) ++{ ++ int err; ++ struct au_sbinfo *sbinfo; ++ ++ SiMustWriteLock(sb); ++ ++ err = 1; /* handled */ ++ sbinfo = au_sbi(sb); ++ switch (opt->type) { ++ case Opt_udba: ++ sbinfo->si_mntflags &= ~AuOptMask_UDBA; ++ sbinfo->si_mntflags |= opt->udba; ++ opts->given_udba |= opt->udba; ++ break; ++ ++ case Opt_plink: ++ au_opt_set(sbinfo->si_mntflags, PLINK); ++ break; ++ case Opt_noplink: ++ if (au_opt_test(sbinfo->si_mntflags, PLINK)) ++ au_plink_put(sb); ++ au_opt_clr(sbinfo->si_mntflags, PLINK); ++ break; ++ case Opt_list_plink: ++ if (au_opt_test(sbinfo->si_mntflags, PLINK)) ++ au_plink_list(sb); ++ break; ++ ++ case Opt_diropq_a: ++ au_opt_set(sbinfo->si_mntflags, ALWAYS_DIROPQ); ++ break; ++ case Opt_diropq_w: ++ au_opt_clr(sbinfo->si_mntflags, ALWAYS_DIROPQ); ++ break; ++ ++ case Opt_warn_perm: ++ au_opt_set(sbinfo->si_mntflags, WARN_PERM); ++ break; ++ case Opt_nowarn_perm: ++ au_opt_clr(sbinfo->si_mntflags, WARN_PERM); ++ break; ++ ++ case Opt_refrof: ++ au_opt_set(sbinfo->si_mntflags, REFROF); ++ break; ++ case Opt_norefrof: ++ au_opt_clr(sbinfo->si_mntflags, REFROF); ++ break; ++ ++ case Opt_verbose: ++ au_opt_set(sbinfo->si_mntflags, VERBOSE); ++ break; ++ case Opt_noverbose: ++ au_opt_clr(sbinfo->si_mntflags, VERBOSE); ++ break; ++ ++ case Opt_sum: ++ au_opt_set(sbinfo->si_mntflags, SUM); ++ break; ++ case Opt_wsum: ++ au_opt_clr(sbinfo->si_mntflags, SUM); ++ au_opt_set(sbinfo->si_mntflags, SUM_W); ++ case Opt_nosum: ++ au_opt_clr(sbinfo->si_mntflags, SUM); ++ au_opt_clr(sbinfo->si_mntflags, SUM_W); ++ break; ++ ++ case Opt_wbr_create: ++ err = au_opt_wbr_create(sb, &opt->wbr_create); ++ break; ++ case Opt_wbr_copyup: ++ sbinfo->si_wbr_copyup = opt->wbr_copyup; ++ sbinfo->si_wbr_copyup_ops = au_wbr_copyup_ops + opt->wbr_copyup; ++ break; ++ ++ case Opt_dirwh: ++ sbinfo->si_dirwh = opt->dirwh; ++ break; ++ ++ case Opt_rdcache: ++ sbinfo->si_rdcache = opt->rdcache * HZ; ++ break; ++ case Opt_rdblk: ++ sbinfo->si_rdblk = opt->rdblk; ++ break; ++ case Opt_rdblk_def: ++ sbinfo->si_rdblk = AUFS_RDBLK_DEF; ++ break; ++ case Opt_rdhash: ++ sbinfo->si_rdhash = opt->rdhash; ++ break; ++ case Opt_rdhash_def: ++ sbinfo->si_rdhash = AUFS_RDHASH_DEF; ++ break; ++ ++ case Opt_shwh: ++ au_opt_set(sbinfo->si_mntflags, SHWH); ++ break; ++ case Opt_noshwh: ++ au_opt_clr(sbinfo->si_mntflags, SHWH); ++ break; ++ ++ case Opt_trunc_xino: ++ au_opt_set(sbinfo->si_mntflags, TRUNC_XINO); ++ break; ++ case Opt_notrunc_xino: ++ au_opt_clr(sbinfo->si_mntflags, TRUNC_XINO); ++ break; ++ ++ case Opt_trunc_xino_path: ++ case Opt_itrunc_xino: ++ err = au_xino_trunc(sb, opt->xino_itrunc.bindex); ++ if (!err) ++ err = 1; ++ break; ++ ++ case Opt_trunc_xib: ++ au_fset_opts(opts->flags, TRUNC_XIB); ++ break; ++ case Opt_notrunc_xib: ++ au_fclr_opts(opts->flags, TRUNC_XIB); ++ break; ++ ++ default: ++ err = 0; ++ break; ++ } ++ ++ return err; ++} ++ ++/* ++ * returns tri-state. ++ * plus: processed without an error ++ * zero: unprocessed ++ * minus: error ++ */ ++static int au_opt_br(struct super_block *sb, struct au_opt *opt, ++ struct au_opts *opts) ++{ ++ int err, do_refresh; ++ ++ err = 0; ++ switch (opt->type) { ++ case Opt_append: ++ opt->add.bindex = au_sbend(sb) + 1; ++ if (opt->add.bindex < 0) ++ opt->add.bindex = 0; ++ goto add; ++ case Opt_prepend: ++ opt->add.bindex = 0; ++ add: ++ case Opt_add: ++ err = au_br_add(sb, &opt->add, ++ au_ftest_opts(opts->flags, REMOUNT)); ++ if (!err) { ++ err = 1; ++ au_fset_opts(opts->flags, REFRESH_DIR); ++ if (au_br_whable(opt->add.perm)) ++ au_fset_opts(opts->flags, REFRESH_NONDIR); ++ } ++ break; ++ ++ case Opt_del: ++ case Opt_idel: ++ err = au_br_del(sb, &opt->del, ++ au_ftest_opts(opts->flags, REMOUNT)); ++ if (!err) { ++ err = 1; ++ au_fset_opts(opts->flags, TRUNC_XIB); ++ au_fset_opts(opts->flags, REFRESH_DIR); ++ au_fset_opts(opts->flags, REFRESH_NONDIR); ++ } ++ break; ++ ++ case Opt_mod: ++ case Opt_imod: ++ err = au_br_mod(sb, &opt->mod, ++ au_ftest_opts(opts->flags, REMOUNT), ++ &do_refresh); ++ if (!err) { ++ err = 1; ++ if (do_refresh) { ++ au_fset_opts(opts->flags, REFRESH_DIR); ++ au_fset_opts(opts->flags, REFRESH_NONDIR); ++ } ++ } ++ break; ++ } ++ ++ return err; ++} ++ ++static int au_opt_xino(struct super_block *sb, struct au_opt *opt, ++ struct au_opt_xino **opt_xino, ++ struct au_opts *opts) ++{ ++ int err; ++ aufs_bindex_t bend, bindex; ++ struct dentry *root, *parent, *h_root; ++ ++ err = 0; ++ switch (opt->type) { ++ case Opt_xino: ++ err = au_xino_set(sb, &opt->xino, ++ !!au_ftest_opts(opts->flags, REMOUNT)); ++ if (unlikely(err)) ++ break; ++ ++ *opt_xino = &opt->xino; ++ au_xino_brid_set(sb, -1); ++ ++ /* safe d_parent access */ ++ parent = opt->xino.file->f_dentry->d_parent; ++ root = sb->s_root; ++ bend = au_sbend(sb); ++ for (bindex = 0; bindex <= bend; bindex++) { ++ h_root = au_h_dptr(root, bindex); ++ if (h_root == parent) { ++ au_xino_brid_set(sb, au_sbr_id(sb, bindex)); ++ break; ++ } ++ } ++ break; ++ ++ case Opt_noxino: ++ au_xino_clr(sb); ++ au_xino_brid_set(sb, -1); ++ *opt_xino = (void *)-1; ++ break; ++ } ++ ++ return err; ++} ++ ++int au_opts_verify(struct super_block *sb, unsigned long sb_flags, ++ unsigned int pending) ++{ ++ int err; ++ aufs_bindex_t bindex, bend; ++ unsigned char do_plink, skip, do_free; ++ struct au_branch *br; ++ struct au_wbr *wbr; ++ struct dentry *root; ++ struct inode *dir, *h_dir; ++ struct au_sbinfo *sbinfo; ++ struct au_hinode *hdir; ++ ++ SiMustAnyLock(sb); ++ ++ sbinfo = au_sbi(sb); ++ AuDebugOn(!(sbinfo->si_mntflags & AuOptMask_UDBA)); ++ ++ if (!(sb_flags & MS_RDONLY)) { ++ if (unlikely(!au_br_writable(au_sbr_perm(sb, 0)))) ++ AuWarn("first branch should be rw\n"); ++ if (unlikely(au_opt_test(sbinfo->si_mntflags, SHWH))) ++ AuWarn("shwh should be used with ro\n"); ++ } ++ ++ if (au_opt_test((sbinfo->si_mntflags | pending), UDBA_HINOTIFY) ++ && !au_opt_test(sbinfo->si_mntflags, XINO)) ++ AuWarn("udba=inotify requires xino\n"); ++ ++ err = 0; ++ root = sb->s_root; ++ dir = sb->s_root->d_inode; ++ do_plink = !!au_opt_test(sbinfo->si_mntflags, PLINK); ++ bend = au_sbend(sb); ++ for (bindex = 0; !err && bindex <= bend; bindex++) { ++ skip = 0; ++ h_dir = au_h_iptr(dir, bindex); ++ br = au_sbr(sb, bindex); ++ do_free = 0; ++ ++ wbr = br->br_wbr; ++ if (wbr) ++ wbr_wh_read_lock(wbr); ++ ++ switch (br->br_perm) { ++ case AuBrPerm_RO: ++ case AuBrPerm_ROWH: ++ case AuBrPerm_RR: ++ case AuBrPerm_RRWH: ++ do_free = !!wbr; ++ skip = (!wbr ++ || (!wbr->wbr_whbase ++ && !wbr->wbr_plink ++ && !wbr->wbr_orph)); ++ break; ++ ++ case AuBrPerm_RWNoLinkWH: ++ /* skip = (!br->br_whbase && !br->br_orph); */ ++ skip = (!wbr || !wbr->wbr_whbase); ++ if (skip && wbr) { ++ if (do_plink) ++ skip = !!wbr->wbr_plink; ++ else ++ skip = !wbr->wbr_plink; ++ } ++ break; ++ ++ case AuBrPerm_RW: ++ /* skip = (br->br_whbase && br->br_ohph); */ ++ skip = (wbr && wbr->wbr_whbase); ++ if (skip) { ++ if (do_plink) ++ skip = !!wbr->wbr_plink; ++ else ++ skip = !wbr->wbr_plink; ++ } ++ break; ++ ++ default: ++ BUG(); ++ } ++ if (wbr) ++ wbr_wh_read_unlock(wbr); ++ ++ if (skip) ++ continue; ++ ++ hdir = au_hi(dir, bindex); ++ au_hin_imtx_lock_nested(hdir, AuLsc_I_PARENT); ++ if (wbr) ++ wbr_wh_write_lock(wbr); ++ err = au_wh_init(au_h_dptr(root, bindex), br, sb); ++ if (wbr) ++ wbr_wh_write_unlock(wbr); ++ au_hin_imtx_unlock(hdir); ++ ++ if (!err && do_free) { ++ kfree(wbr); ++ br->br_wbr = NULL; ++ } ++ } ++ ++ return err; ++} ++ ++int au_opts_mount(struct super_block *sb, struct au_opts *opts) ++{ ++ int err; ++ unsigned int tmp; ++ aufs_bindex_t bend; ++ struct au_opt *opt; ++ struct au_opt_xino *opt_xino, xino; ++ struct au_sbinfo *sbinfo; ++ ++ SiMustWriteLock(sb); ++ ++ err = 0; ++ opt_xino = NULL; ++ opt = opts->opt; ++ while (err >= 0 && opt->type != Opt_tail) ++ err = au_opt_simple(sb, opt++, opts); ++ if (err > 0) ++ err = 0; ++ else if (unlikely(err < 0)) ++ goto out; ++ ++ /* disable xino and udba temporary */ ++ sbinfo = au_sbi(sb); ++ tmp = sbinfo->si_mntflags; ++ au_opt_clr(sbinfo->si_mntflags, XINO); ++ au_opt_set_udba(sbinfo->si_mntflags, UDBA_REVAL); ++ ++ opt = opts->opt; ++ while (err >= 0 && opt->type != Opt_tail) ++ err = au_opt_br(sb, opt++, opts); ++ if (err > 0) ++ err = 0; ++ else if (unlikely(err < 0)) ++ goto out; ++ ++ bend = au_sbend(sb); ++ if (unlikely(bend < 0)) { ++ err = -EINVAL; ++ AuErr("no branches\n"); ++ goto out; ++ } ++ ++ if (au_opt_test(tmp, XINO)) ++ au_opt_set(sbinfo->si_mntflags, XINO); ++ opt = opts->opt; ++ while (!err && opt->type != Opt_tail) ++ err = au_opt_xino(sb, opt++, &opt_xino, opts); ++ if (unlikely(err)) ++ goto out; ++ ++ err = au_opts_verify(sb, sb->s_flags, tmp); ++ if (unlikely(err)) ++ goto out; ++ ++ /* restore xino */ ++ if (au_opt_test(tmp, XINO) && !opt_xino) { ++ xino.file = au_xino_def(sb); ++ err = PTR_ERR(xino.file); ++ if (IS_ERR(xino.file)) ++ goto out; ++ ++ err = au_xino_set(sb, &xino, /*remount*/0); ++ fput(xino.file); ++ if (unlikely(err)) ++ goto out; ++ } ++ ++ /* restore udba */ ++ sbinfo->si_mntflags &= ~AuOptMask_UDBA; ++ sbinfo->si_mntflags |= (tmp & AuOptMask_UDBA); ++ if (au_opt_test(tmp, UDBA_HINOTIFY)) { ++ struct inode *dir = sb->s_root->d_inode; ++ au_reset_hinotify(dir, ++ au_hi_flags(dir, /*isdir*/1) & ~AuHi_XINO); ++ } ++ ++ out: ++ return err; ++} ++ ++int au_opts_remount(struct super_block *sb, struct au_opts *opts) ++{ ++ int err, rerr; ++ struct inode *dir; ++ struct au_opt_xino *opt_xino; ++ struct au_opt *opt; ++ struct au_sbinfo *sbinfo; ++ ++ SiMustWriteLock(sb); ++ ++ dir = sb->s_root->d_inode; ++ sbinfo = au_sbi(sb); ++ err = 0; ++ opt_xino = NULL; ++ opt = opts->opt; ++ while (err >= 0 && opt->type != Opt_tail) { ++ err = au_opt_simple(sb, opt, opts); ++ if (!err) ++ err = au_opt_br(sb, opt, opts); ++ if (!err) ++ err = au_opt_xino(sb, opt, &opt_xino, opts); ++ opt++; ++ } ++ if (err > 0) ++ err = 0; ++ AuTraceErr(err); ++ /* go on even err */ ++ ++ rerr = au_opts_verify(sb, opts->sb_flags, /*pending*/0); ++ if (unlikely(rerr && !err)) ++ err = rerr; ++ ++ if (au_ftest_opts(opts->flags, TRUNC_XIB)) { ++ rerr = au_xib_trunc(sb); ++ if (unlikely(rerr && !err)) ++ err = rerr; ++ } ++ ++ /* will be handled by the caller */ ++ if (!au_ftest_opts(opts->flags, REFRESH_DIR) ++ && (opts->given_udba || au_opt_test(sbinfo->si_mntflags, XINO))) ++ au_fset_opts(opts->flags, REFRESH_DIR); ++ ++ AuDbg("status 0x%x\n", opts->flags); ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++unsigned int au_opt_udba(struct super_block *sb) ++{ ++ return au_mntflags(sb) & AuOptMask_UDBA; ++} +diff -Nur linux-2.6.31.4.orig/fs/aufs/opts.h linux-2.6.31.4/fs/aufs/opts.h +--- linux-2.6.31.4.orig/fs/aufs/opts.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/opts.h 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,196 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * mount options/flags ++ */ ++ ++#ifndef __AUFS_OPTS_H__ ++#define __AUFS_OPTS_H__ ++ ++#ifdef __KERNEL__ ++ ++#include <linux/path.h> ++#include <linux/aufs_type.h> ++ ++struct file; ++struct super_block; ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* mount flags */ ++#define AuOpt_XINO 1 /* external inode number bitmap ++ and translation table */ ++#define AuOpt_TRUNC_XINO (1 << 1) /* truncate xino files */ ++#define AuOpt_UDBA_NONE (1 << 2) /* users direct branch access */ ++#define AuOpt_UDBA_REVAL (1 << 3) ++#define AuOpt_UDBA_HINOTIFY (1 << 4) ++#define AuOpt_SHWH (1 << 5) /* show whiteout */ ++#define AuOpt_PLINK (1 << 6) /* pseudo-link */ ++#define AuOpt_DIRPERM1 (1 << 7) /* unimplemented */ ++#define AuOpt_REFROF (1 << 8) /* unimplemented */ ++#define AuOpt_ALWAYS_DIROPQ (1 << 9) /* policy to creating diropq */ ++#define AuOpt_SUM (1 << 10) /* summation for statfs(2) */ ++#define AuOpt_SUM_W (1 << 11) /* unimplemented */ ++#define AuOpt_WARN_PERM (1 << 12) /* warn when add-branch */ ++#define AuOpt_VERBOSE (1 << 13) /* busy inode when del-branch */ ++ ++#ifndef CONFIG_AUFS_HINOTIFY ++#undef AuOpt_UDBA_HINOTIFY ++#define AuOpt_UDBA_HINOTIFY 0 ++#endif ++#ifndef CONFIG_AUFS_SHWH ++#undef AuOpt_SHWH ++#define AuOpt_SHWH 0 ++#endif ++ ++#define AuOpt_Def (AuOpt_XINO \ ++ | AuOpt_UDBA_REVAL \ ++ | AuOpt_PLINK \ ++ /* | AuOpt_DIRPERM1 */ \ ++ | AuOpt_WARN_PERM) ++#define AuOptMask_UDBA (AuOpt_UDBA_NONE \ ++ | AuOpt_UDBA_REVAL \ ++ | AuOpt_UDBA_HINOTIFY) ++ ++#define au_opt_test(flags, name) (flags & AuOpt_##name) ++#define au_opt_set(flags, name) do { \ ++ BUILD_BUG_ON(AuOpt_##name & AuOptMask_UDBA); \ ++ ((flags) |= AuOpt_##name); \ ++} while (0) ++#define au_opt_set_udba(flags, name) do { \ ++ (flags) &= ~AuOptMask_UDBA; \ ++ ((flags) |= AuOpt_##name); \ ++} while (0) ++#define au_opt_clr(flags, name) { ((flags) &= ~AuOpt_##name); } ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* policies to select one among multiple writable branches */ ++enum { ++ AuWbrCreate_TDP, /* top down parent */ ++ AuWbrCreate_RR, /* round robin */ ++ AuWbrCreate_MFS, /* most free space */ ++ AuWbrCreate_MFSV, /* mfs with seconds */ ++ AuWbrCreate_MFSRR, /* mfs then rr */ ++ AuWbrCreate_MFSRRV, /* mfs then rr with seconds */ ++ AuWbrCreate_PMFS, /* parent and mfs */ ++ AuWbrCreate_PMFSV, /* parent and mfs with seconds */ ++ ++ AuWbrCreate_Def = AuWbrCreate_TDP ++}; ++ ++enum { ++ AuWbrCopyup_TDP, /* top down parent */ ++ AuWbrCopyup_BUP, /* bottom up parent */ ++ AuWbrCopyup_BU, /* bottom up */ ++ ++ AuWbrCopyup_Def = AuWbrCopyup_TDP ++}; ++ ++/* ---------------------------------------------------------------------- */ ++ ++struct au_opt_add { ++ aufs_bindex_t bindex; ++ char *pathname; ++ int perm; ++ struct path path; ++}; ++ ++struct au_opt_del { ++ char *pathname; ++ struct path h_path; ++}; ++ ++struct au_opt_mod { ++ char *path; ++ int perm; ++ struct dentry *h_root; ++}; ++ ++struct au_opt_xino { ++ char *path; ++ struct file *file; ++}; ++ ++struct au_opt_xino_itrunc { ++ aufs_bindex_t bindex; ++}; ++ ++struct au_opt_wbr_create { ++ int wbr_create; ++ int mfs_second; ++ unsigned long long mfsrr_watermark; ++}; ++ ++struct au_opt { ++ int type; ++ union { ++ struct au_opt_xino xino; ++ struct au_opt_xino_itrunc xino_itrunc; ++ struct au_opt_add add; ++ struct au_opt_del del; ++ struct au_opt_mod mod; ++ int dirwh; ++ int rdcache; ++ unsigned int rdblk; ++ unsigned int rdhash; ++ int udba; ++ struct au_opt_wbr_create wbr_create; ++ int wbr_copyup; ++ }; ++}; ++ ++/* opts flags */ ++#define AuOpts_REMOUNT 1 ++#define AuOpts_REFRESH_DIR (1 << 1) ++#define AuOpts_REFRESH_NONDIR (1 << 2) ++#define AuOpts_TRUNC_XIB (1 << 3) ++#define au_ftest_opts(flags, name) ((flags) & AuOpts_##name) ++#define au_fset_opts(flags, name) { (flags) |= AuOpts_##name; } ++#define au_fclr_opts(flags, name) { (flags) &= ~AuOpts_##name; } ++ ++struct au_opts { ++ struct au_opt *opt; ++ int max_opt; ++ ++ unsigned int given_udba; ++ unsigned int flags; ++ unsigned long sb_flags; ++}; ++ ++/* ---------------------------------------------------------------------- */ ++ ++const char *au_optstr_br_perm(int brperm); ++const char *au_optstr_udba(int udba); ++const char *au_optstr_wbr_copyup(int wbr_copyup); ++const char *au_optstr_wbr_create(int wbr_create); ++ ++void au_opts_free(struct au_opts *opts); ++int au_opts_parse(struct super_block *sb, char *str, struct au_opts *opts); ++int au_opts_verify(struct super_block *sb, unsigned long sb_flags, ++ unsigned int pending); ++int au_opts_mount(struct super_block *sb, struct au_opts *opts); ++int au_opts_remount(struct super_block *sb, struct au_opts *opts); ++ ++unsigned int au_opt_udba(struct super_block *sb); ++ ++/* ---------------------------------------------------------------------- */ ++ ++#endif /* __KERNEL__ */ ++#endif /* __AUFS_OPTS_H__ */ +diff -Nur linux-2.6.31.4.orig/fs/aufs/plink.c linux-2.6.31.4/fs/aufs/plink.c +--- linux-2.6.31.4.orig/fs/aufs/plink.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/plink.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,354 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * pseudo-link ++ */ ++ ++#include "aufs.h" ++ ++/* ++ * during a user process maintains the pseudo-links, ++ * prohibit adding a new plink and branch manipulation. ++ */ ++void au_plink_block_maintain(struct super_block *sb) ++{ ++ struct au_sbinfo *sbi = au_sbi(sb); ++ ++ SiMustAnyLock(sb); ++ ++ /* gave up wake_up_bit() */ ++ wait_event(sbi->si_plink_wq, !au_ftest_si(sbi, MAINTAIN_PLINK)); ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++struct pseudo_link { ++ struct list_head list; ++ struct inode *inode; ++}; ++ ++#ifdef CONFIG_AUFS_DEBUG ++void au_plink_list(struct super_block *sb) ++{ ++ struct au_sbinfo *sbinfo; ++ struct list_head *plink_list; ++ struct pseudo_link *plink; ++ ++ SiMustAnyLock(sb); ++ ++ sbinfo = au_sbi(sb); ++ AuDebugOn(!au_opt_test(au_mntflags(sb), PLINK)); ++ ++ plink_list = &sbinfo->si_plink.head; ++ spin_lock(&sbinfo->si_plink.spin); ++ list_for_each_entry(plink, plink_list, list) ++ AuDbg("%lu\n", plink->inode->i_ino); ++ spin_unlock(&sbinfo->si_plink.spin); ++} ++#endif ++ ++/* is the inode pseudo-linked? */ ++int au_plink_test(struct inode *inode) ++{ ++ int found; ++ struct au_sbinfo *sbinfo; ++ struct list_head *plink_list; ++ struct pseudo_link *plink; ++ ++ sbinfo = au_sbi(inode->i_sb); ++ AuRwMustAnyLock(&sbinfo->si_rwsem); ++ AuDebugOn(!au_opt_test(au_mntflags(inode->i_sb), PLINK)); ++ ++ found = 0; ++ plink_list = &sbinfo->si_plink.head; ++ spin_lock(&sbinfo->si_plink.spin); ++ list_for_each_entry(plink, plink_list, list) ++ if (plink->inode == inode) { ++ found = 1; ++ break; ++ } ++ spin_unlock(&sbinfo->si_plink.spin); ++ return found; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* ++ * generate a name for plink. ++ * the file will be stored under AUFS_WH_PLINKDIR. ++ */ ++/* 20 is max digits length of ulong 64 */ ++#define PLINK_NAME_LEN ((20 + 1) * 2) ++ ++static int plink_name(char *name, int len, struct inode *inode, ++ aufs_bindex_t bindex) ++{ ++ int rlen; ++ struct inode *h_inode; ++ ++ h_inode = au_h_iptr(inode, bindex); ++ rlen = snprintf(name, len, "%lu.%lu", inode->i_ino, h_inode->i_ino); ++ return rlen; ++} ++ ++/* lookup the plink-ed @inode under the branch at @bindex */ ++struct dentry *au_plink_lkup(struct inode *inode, aufs_bindex_t bindex) ++{ ++ struct dentry *h_dentry, *h_parent; ++ struct au_branch *br; ++ struct inode *h_dir; ++ char a[PLINK_NAME_LEN]; ++ struct qstr tgtname = { ++ .name = a ++ }; ++ ++ br = au_sbr(inode->i_sb, bindex); ++ h_parent = br->br_wbr->wbr_plink; ++ h_dir = h_parent->d_inode; ++ tgtname.len = plink_name(a, sizeof(a), inode, bindex); ++ ++ /* always superio. */ ++ mutex_lock_nested(&h_dir->i_mutex, AuLsc_I_CHILD2); ++ h_dentry = au_sio_lkup_one(&tgtname, h_parent, br); ++ mutex_unlock(&h_dir->i_mutex); ++ return h_dentry; ++} ++ ++/* create a pseudo-link */ ++static int do_whplink(struct qstr *tgt, struct dentry *h_parent, ++ struct dentry *h_dentry, struct au_branch *br) ++{ ++ int err; ++ struct path h_path = { ++ .mnt = br->br_mnt ++ }; ++ struct inode *h_dir; ++ ++ h_dir = h_parent->d_inode; ++ again: ++ h_path.dentry = au_lkup_one(tgt, h_parent, br, /*nd*/NULL); ++ err = PTR_ERR(h_path.dentry); ++ if (IS_ERR(h_path.dentry)) ++ goto out; ++ ++ err = 0; ++ /* wh.plink dir is not monitored */ ++ if (h_path.dentry->d_inode ++ && h_path.dentry->d_inode != h_dentry->d_inode) { ++ err = vfsub_unlink(h_dir, &h_path, /*force*/0); ++ dput(h_path.dentry); ++ h_path.dentry = NULL; ++ if (!err) ++ goto again; ++ } ++ if (!err && !h_path.dentry->d_inode) ++ err = vfsub_link(h_dentry, h_dir, &h_path); ++ dput(h_path.dentry); ++ ++ out: ++ return err; ++} ++ ++struct do_whplink_args { ++ int *errp; ++ struct qstr *tgt; ++ struct dentry *h_parent; ++ struct dentry *h_dentry; ++ struct au_branch *br; ++}; ++ ++static void call_do_whplink(void *args) ++{ ++ struct do_whplink_args *a = args; ++ *a->errp = do_whplink(a->tgt, a->h_parent, a->h_dentry, a->br); ++} ++ ++static int whplink(struct dentry *h_dentry, struct inode *inode, ++ aufs_bindex_t bindex, struct au_branch *br) ++{ ++ int err, wkq_err; ++ struct au_wbr *wbr; ++ struct dentry *h_parent; ++ struct inode *h_dir; ++ char a[PLINK_NAME_LEN]; ++ struct qstr tgtname = { ++ .name = a ++ }; ++ ++ wbr = au_sbr(inode->i_sb, bindex)->br_wbr; ++ h_parent = wbr->wbr_plink; ++ h_dir = h_parent->d_inode; ++ tgtname.len = plink_name(a, sizeof(a), inode, bindex); ++ ++ /* always superio. */ ++ mutex_lock_nested(&h_dir->i_mutex, AuLsc_I_CHILD2); ++ if (!au_test_wkq(current)) { ++ struct do_whplink_args args = { ++ .errp = &err, ++ .tgt = &tgtname, ++ .h_parent = h_parent, ++ .h_dentry = h_dentry, ++ .br = br ++ }; ++ wkq_err = au_wkq_wait(call_do_whplink, &args); ++ if (unlikely(wkq_err)) ++ err = wkq_err; ++ } else ++ err = do_whplink(&tgtname, h_parent, h_dentry, br); ++ mutex_unlock(&h_dir->i_mutex); ++ ++ return err; ++} ++ ++/* free a single plink */ ++static void do_put_plink(struct pseudo_link *plink, int do_del) ++{ ++ iput(plink->inode); ++ if (do_del) ++ list_del(&plink->list); ++ kfree(plink); ++} ++ ++/* ++ * create a new pseudo-link for @h_dentry on @bindex. ++ * the linked inode is held in aufs @inode. ++ */ ++void au_plink_append(struct inode *inode, aufs_bindex_t bindex, ++ struct dentry *h_dentry) ++{ ++ struct super_block *sb; ++ struct au_sbinfo *sbinfo; ++ struct list_head *plink_list; ++ struct pseudo_link *plink; ++ int found, err, cnt; ++ ++ sb = inode->i_sb; ++ sbinfo = au_sbi(sb); ++ AuDebugOn(!au_opt_test(au_mntflags(sb), PLINK)); ++ ++ err = 0; ++ cnt = 0; ++ found = 0; ++ plink_list = &sbinfo->si_plink.head; ++ spin_lock(&sbinfo->si_plink.spin); ++ list_for_each_entry(plink, plink_list, list) { ++ cnt++; ++ if (plink->inode == inode) { ++ found = 1; ++ break; ++ } ++ } ++ if (found) { ++ spin_unlock(&sbinfo->si_plink.spin); ++ return; ++ } ++ ++ plink = NULL; ++ if (!found) { ++ plink = kmalloc(sizeof(*plink), GFP_ATOMIC); ++ if (plink) { ++ plink->inode = au_igrab(inode); ++ list_add(&plink->list, plink_list); ++ cnt++; ++ } else ++ err = -ENOMEM; ++ } ++ spin_unlock(&sbinfo->si_plink.spin); ++ ++ if (!err) { ++ au_plink_block_maintain(sb); ++ err = whplink(h_dentry, inode, bindex, au_sbr(sb, bindex)); ++ } ++ ++ if (unlikely(cnt > AUFS_PLINK_WARN)) ++ AuWarn1("unexpectedly many pseudo links, %d\n", cnt); ++ if (unlikely(err)) { ++ AuWarn("err %d, damaged pseudo link.\n", err); ++ if (!found && plink) ++ do_put_plink(plink, /*do_del*/1); ++ } ++} ++ ++/* free all plinks */ ++void au_plink_put(struct super_block *sb) ++{ ++ struct au_sbinfo *sbinfo; ++ struct list_head *plink_list; ++ struct pseudo_link *plink, *tmp; ++ ++ SiMustWriteLock(sb); ++ ++ sbinfo = au_sbi(sb); ++ AuDebugOn(!au_opt_test(au_mntflags(sb), PLINK)); ++ ++ plink_list = &sbinfo->si_plink.head; ++ /* no spin_lock since sbinfo is write-locked */ ++ list_for_each_entry_safe(plink, tmp, plink_list, list) ++ do_put_plink(plink, 0); ++ INIT_LIST_HEAD(plink_list); ++} ++ ++/* free the plinks on a branch specified by @br_id */ ++void au_plink_half_refresh(struct super_block *sb, aufs_bindex_t br_id) ++{ ++ struct au_sbinfo *sbinfo; ++ struct list_head *plink_list; ++ struct pseudo_link *plink, *tmp; ++ struct inode *inode; ++ aufs_bindex_t bstart, bend, bindex; ++ unsigned char do_put; ++ ++ SiMustWriteLock(sb); ++ ++ sbinfo = au_sbi(sb); ++ AuDebugOn(!au_opt_test(au_mntflags(sb), PLINK)); ++ ++ plink_list = &sbinfo->si_plink.head; ++ /* no spin_lock since sbinfo is write-locked */ ++ list_for_each_entry_safe(plink, tmp, plink_list, list) { ++ do_put = 0; ++ inode = au_igrab(plink->inode); ++ ii_write_lock_child(inode); ++ bstart = au_ibstart(inode); ++ bend = au_ibend(inode); ++ if (bstart >= 0) { ++ for (bindex = bstart; bindex <= bend; bindex++) { ++ if (!au_h_iptr(inode, bindex) ++ || au_ii_br_id(inode, bindex) != br_id) ++ continue; ++ au_set_h_iptr(inode, bindex, NULL, 0); ++ do_put = 1; ++ break; ++ } ++ } else ++ do_put_plink(plink, 1); ++ ++ if (do_put) { ++ for (bindex = bstart; bindex <= bend; bindex++) ++ if (au_h_iptr(inode, bindex)) { ++ do_put = 0; ++ break; ++ } ++ if (do_put) ++ do_put_plink(plink, 1); ++ } ++ ii_write_unlock(inode); ++ iput(inode); ++ } ++} +diff -Nur linux-2.6.31.4.orig/fs/aufs/poll.c linux-2.6.31.4/fs/aufs/poll.c +--- linux-2.6.31.4.orig/fs/aufs/poll.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/poll.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,56 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * poll operation ++ * There is only one filesystem which implements ->poll operation, currently. ++ */ ++ ++#include "aufs.h" ++ ++unsigned int aufs_poll(struct file *file, poll_table *wait) ++{ ++ unsigned int mask; ++ int err; ++ struct file *h_file; ++ struct dentry *dentry; ++ struct super_block *sb; ++ ++ /* We should pretend an error happened. */ ++ mask = POLLERR /* | POLLIN | POLLOUT */; ++ dentry = file->f_dentry; ++ sb = dentry->d_sb; ++ si_read_lock(sb, AuLock_FLUSH); ++ err = au_reval_and_lock_fdi(file, au_reopen_nondir, /*wlock*/0); ++ if (unlikely(err)) ++ goto out; ++ ++ /* it is not an error if h_file has no operation */ ++ mask = DEFAULT_POLLMASK; ++ h_file = au_h_fptr(file, au_fbstart(file)); ++ if (h_file->f_op && h_file->f_op->poll) ++ mask = h_file->f_op->poll(h_file, wait); ++ ++ di_read_unlock(dentry, AuLock_IR); ++ fi_read_unlock(file); ++ ++ out: ++ si_read_unlock(sb); ++ AuTraceErr((int)mask); ++ return mask; ++} +diff -Nur linux-2.6.31.4.orig/fs/aufs/rwsem.h linux-2.6.31.4/fs/aufs/rwsem.h +--- linux-2.6.31.4.orig/fs/aufs/rwsem.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/rwsem.h 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,186 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * simple read-write semaphore wrappers ++ */ ++ ++#ifndef __AUFS_RWSEM_H__ ++#define __AUFS_RWSEM_H__ ++ ++#ifdef __KERNEL__ ++ ++#include <linux/rwsem.h> ++ ++struct au_rwsem { ++ struct rw_semaphore rwsem; ++#ifdef CONFIG_AUFS_DEBUG ++ /* just for debugging, not almighty counter */ ++ atomic_t rcnt, wcnt; ++#endif ++}; ++ ++#ifdef CONFIG_AUFS_DEBUG ++#define AuDbgCntInit(rw) do { \ ++ atomic_set(&(rw)->rcnt, 0); \ ++ atomic_set(&(rw)->wcnt, 0); \ ++ smp_mb(); /* atomic set */ \ ++} while (0) ++ ++#define AuDbgRcntInc(rw) atomic_inc_return(&(rw)->rcnt) ++#define AuDbgRcntDec(rw) WARN_ON(atomic_dec_return(&(rw)->rcnt) < 0) ++#define AuDbgWcntInc(rw) WARN_ON(atomic_inc_return(&(rw)->wcnt) > 1) ++#define AuDbgWcntDec(rw) WARN_ON(atomic_dec_return(&(rw)->wcnt) < 0) ++#else ++#define AuDbgCntInit(rw) do {} while (0) ++#define AuDbgRcntInc(rw) do {} while (0) ++#define AuDbgRcntDec(rw) do {} while (0) ++#define AuDbgWcntInc(rw) do {} while (0) ++#define AuDbgWcntDec(rw) do {} while (0) ++#endif /* CONFIG_AUFS_DEBUG */ ++ ++/* to debug easier, do not make them inlined functions */ ++#define AuRwMustNoWaiters(rw) AuDebugOn(!list_empty(&(rw)->rwsem.wait_list)) ++/* rwsem_is_locked() is unusable */ ++#define AuRwMustReadLock(rw) AuDebugOn(atomic_read(&(rw)->rcnt) <= 0) ++#define AuRwMustWriteLock(rw) AuDebugOn(atomic_read(&(rw)->wcnt) <= 0) ++#define AuRwMustAnyLock(rw) AuDebugOn(atomic_read(&(rw)->rcnt) <= 0 \ ++ && atomic_read(&(rw)->wcnt) <= 0) ++#define AuRwDestroy(rw) AuDebugOn(atomic_read(&(rw)->rcnt) \ ++ || atomic_read(&(rw)->wcnt)) ++ ++static inline void au_rw_init(struct au_rwsem *rw) ++{ ++ AuDbgCntInit(rw); ++ init_rwsem(&rw->rwsem); ++} ++ ++static inline void au_rw_init_wlock(struct au_rwsem *rw) ++{ ++ au_rw_init(rw); ++ down_write(&rw->rwsem); ++ AuDbgWcntInc(rw); ++} ++ ++static inline void au_rw_init_wlock_nested(struct au_rwsem *rw, ++ unsigned int lsc) ++{ ++ au_rw_init(rw); ++ down_write_nested(&rw->rwsem, lsc); ++ AuDbgWcntInc(rw); ++} ++ ++static inline void au_rw_read_lock(struct au_rwsem *rw) ++{ ++ down_read(&rw->rwsem); ++ AuDbgRcntInc(rw); ++} ++ ++static inline void au_rw_read_lock_nested(struct au_rwsem *rw, unsigned int lsc) ++{ ++ down_read_nested(&rw->rwsem, lsc); ++ AuDbgRcntInc(rw); ++} ++ ++static inline void au_rw_read_unlock(struct au_rwsem *rw) ++{ ++ AuRwMustReadLock(rw); ++ AuDbgRcntDec(rw); ++ up_read(&rw->rwsem); ++} ++ ++static inline void au_rw_dgrade_lock(struct au_rwsem *rw) ++{ ++ AuRwMustWriteLock(rw); ++ AuDbgRcntInc(rw); ++ AuDbgWcntDec(rw); ++ downgrade_write(&rw->rwsem); ++} ++ ++static inline void au_rw_write_lock(struct au_rwsem *rw) ++{ ++ down_write(&rw->rwsem); ++ AuDbgWcntInc(rw); ++} ++ ++static inline void au_rw_write_lock_nested(struct au_rwsem *rw, ++ unsigned int lsc) ++{ ++ down_write_nested(&rw->rwsem, lsc); ++ AuDbgWcntInc(rw); ++} ++ ++static inline void au_rw_write_unlock(struct au_rwsem *rw) ++{ ++ AuRwMustWriteLock(rw); ++ AuDbgWcntDec(rw); ++ up_write(&rw->rwsem); ++} ++ ++/* why is not _nested version defined */ ++static inline int au_rw_read_trylock(struct au_rwsem *rw) ++{ ++ int ret = down_read_trylock(&rw->rwsem); ++ if (ret) ++ AuDbgRcntInc(rw); ++ return ret; ++} ++ ++static inline int au_rw_write_trylock(struct au_rwsem *rw) ++{ ++ int ret = down_write_trylock(&rw->rwsem); ++ if (ret) ++ AuDbgWcntInc(rw); ++ return ret; ++} ++ ++#undef AuDbgCntInit ++#undef AuDbgRcntInc ++#undef AuDbgRcntDec ++#undef AuDbgWcntInc ++#undef AuDbgWcntDec ++ ++#define AuSimpleLockRwsemFuncs(prefix, param, rwsem) \ ++static inline void prefix##_read_lock(param) \ ++{ au_rw_read_lock(rwsem); } \ ++static inline void prefix##_write_lock(param) \ ++{ au_rw_write_lock(rwsem); } \ ++static inline int prefix##_read_trylock(param) \ ++{ return au_rw_read_trylock(rwsem); } \ ++static inline int prefix##_write_trylock(param) \ ++{ return au_rw_write_trylock(rwsem); } ++/* why is not _nested version defined */ ++/* static inline void prefix##_read_trylock_nested(param, lsc) ++{ au_rw_read_trylock_nested(rwsem, lsc)); } ++static inline void prefix##_write_trylock_nestd(param, lsc) ++{ au_rw_write_trylock_nested(rwsem, lsc); } */ ++ ++#define AuSimpleUnlockRwsemFuncs(prefix, param, rwsem) \ ++static inline void prefix##_read_unlock(param) \ ++{ au_rw_read_unlock(rwsem); } \ ++static inline void prefix##_write_unlock(param) \ ++{ au_rw_write_unlock(rwsem); } \ ++static inline void prefix##_downgrade_lock(param) \ ++{ au_rw_dgrade_lock(rwsem); } ++ ++#define AuSimpleRwsemFuncs(prefix, param, rwsem) \ ++ AuSimpleLockRwsemFuncs(prefix, param, rwsem) \ ++ AuSimpleUnlockRwsemFuncs(prefix, param, rwsem) ++ ++#endif /* __KERNEL__ */ ++#endif /* __AUFS_RWSEM_H__ */ +diff -Nur linux-2.6.31.4.orig/fs/aufs/sbinfo.c linux-2.6.31.4/fs/aufs/sbinfo.c +--- linux-2.6.31.4.orig/fs/aufs/sbinfo.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/sbinfo.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,208 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * superblock private data ++ */ ++ ++#include "aufs.h" ++ ++/* ++ * they are necessary regardless sysfs is disabled. ++ */ ++void au_si_free(struct kobject *kobj) ++{ ++ struct au_sbinfo *sbinfo; ++ struct super_block *sb; ++ ++ sbinfo = container_of(kobj, struct au_sbinfo, si_kobj); ++ AuDebugOn(!list_empty(&sbinfo->si_plink.head)); ++ ++ sb = sbinfo->si_sb; ++ si_write_lock(sb); ++ au_xino_clr(sb); ++ au_br_free(sbinfo); ++ kfree(sbinfo->si_branch); ++ mutex_destroy(&sbinfo->si_xib_mtx); ++ si_write_unlock(sb); ++ AuRwDestroy(&sbinfo->si_rwsem); ++ ++ kfree(sbinfo); ++} ++ ++int au_si_alloc(struct super_block *sb) ++{ ++ int err; ++ struct au_sbinfo *sbinfo; ++ ++ err = -ENOMEM; ++ sbinfo = kmalloc(sizeof(*sbinfo), GFP_NOFS); ++ if (unlikely(!sbinfo)) ++ goto out; ++ ++ /* will be reallocated separately */ ++ sbinfo->si_branch = kzalloc(sizeof(*sbinfo->si_branch), GFP_NOFS); ++ if (unlikely(!sbinfo->si_branch)) ++ goto out_sbinfo; ++ ++ memset(&sbinfo->si_kobj, 0, sizeof(sbinfo->si_kobj)); ++ err = sysaufs_si_init(sbinfo); ++ if (unlikely(err)) ++ goto out_br; ++ ++ au_nwt_init(&sbinfo->si_nowait); ++ au_rw_init_wlock(&sbinfo->si_rwsem); ++ sbinfo->si_generation = 0; ++ sbinfo->au_si_status = 0; ++ sbinfo->si_bend = -1; ++ sbinfo->si_last_br_id = 0; ++ ++ sbinfo->si_wbr_copyup = AuWbrCopyup_Def; ++ sbinfo->si_wbr_create = AuWbrCreate_Def; ++ sbinfo->si_wbr_copyup_ops = au_wbr_copyup_ops + AuWbrCopyup_Def; ++ sbinfo->si_wbr_create_ops = au_wbr_create_ops + AuWbrCreate_Def; ++ ++ sbinfo->si_mntflags = AuOpt_Def; ++ ++ sbinfo->si_xread = NULL; ++ sbinfo->si_xwrite = NULL; ++ sbinfo->si_xib = NULL; ++ mutex_init(&sbinfo->si_xib_mtx); ++ sbinfo->si_xib_buf = NULL; ++ sbinfo->si_xino_brid = -1; ++ /* leave si_xib_last_pindex and si_xib_next_bit */ ++ ++ sbinfo->si_rdcache = AUFS_RDCACHE_DEF * HZ; ++ sbinfo->si_rdblk = AUFS_RDBLK_DEF; ++ sbinfo->si_rdhash = AUFS_RDHASH_DEF; ++ sbinfo->si_dirwh = AUFS_DIRWH_DEF; ++ ++ au_spl_init(&sbinfo->si_plink); ++ init_waitqueue_head(&sbinfo->si_plink_wq); ++ ++ /* leave other members for sysaufs and si_mnt. */ ++ sbinfo->si_sb = sb; ++ sb->s_fs_info = sbinfo; ++ au_debug_sbinfo_init(sbinfo); ++ return 0; /* success */ ++ ++ out_br: ++ kfree(sbinfo->si_branch); ++ out_sbinfo: ++ kfree(sbinfo); ++ out: ++ return err; ++} ++ ++int au_sbr_realloc(struct au_sbinfo *sbinfo, int nbr) ++{ ++ int err, sz; ++ struct au_branch **brp; ++ ++ AuRwMustWriteLock(&sbinfo->si_rwsem); ++ ++ err = -ENOMEM; ++ sz = sizeof(*brp) * (sbinfo->si_bend + 1); ++ if (unlikely(!sz)) ++ sz = sizeof(*brp); ++ brp = au_kzrealloc(sbinfo->si_branch, sz, sizeof(*brp) * nbr, GFP_NOFS); ++ if (brp) { ++ sbinfo->si_branch = brp; ++ err = 0; ++ } ++ ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++unsigned int au_sigen_inc(struct super_block *sb) ++{ ++ unsigned int gen; ++ ++ SiMustWriteLock(sb); ++ ++ gen = ++au_sbi(sb)->si_generation; ++ au_update_digen(sb->s_root); ++ au_update_iigen(sb->s_root->d_inode); ++ sb->s_root->d_inode->i_version++; ++ return gen; ++} ++ ++aufs_bindex_t au_new_br_id(struct super_block *sb) ++{ ++ aufs_bindex_t br_id; ++ int i; ++ struct au_sbinfo *sbinfo; ++ ++ SiMustWriteLock(sb); ++ ++ sbinfo = au_sbi(sb); ++ for (i = 0; i <= AUFS_BRANCH_MAX; i++) { ++ br_id = ++sbinfo->si_last_br_id; ++ if (br_id && au_br_index(sb, br_id) < 0) ++ return br_id; ++ } ++ ++ return -1; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* dentry and super_block lock. call at entry point */ ++void aufs_read_lock(struct dentry *dentry, int flags) ++{ ++ si_read_lock(dentry->d_sb, flags); ++ if (au_ftest_lock(flags, DW)) ++ di_write_lock_child(dentry); ++ else ++ di_read_lock_child(dentry, flags); ++} ++ ++void aufs_read_unlock(struct dentry *dentry, int flags) ++{ ++ if (au_ftest_lock(flags, DW)) ++ di_write_unlock(dentry); ++ else ++ di_read_unlock(dentry, flags); ++ si_read_unlock(dentry->d_sb); ++} ++ ++void aufs_write_lock(struct dentry *dentry) ++{ ++ si_write_lock(dentry->d_sb); ++ di_write_lock_child(dentry); ++} ++ ++void aufs_write_unlock(struct dentry *dentry) ++{ ++ di_write_unlock(dentry); ++ si_write_unlock(dentry->d_sb); ++} ++ ++void aufs_read_and_write_lock2(struct dentry *d1, struct dentry *d2, int flags) ++{ ++ si_read_lock(d1->d_sb, flags); ++ di_write_lock2_child(d1, d2, au_ftest_lock(flags, DIR)); ++} ++ ++void aufs_read_and_write_unlock2(struct dentry *d1, struct dentry *d2) ++{ ++ di_write_unlock2(d1, d2); ++ si_read_unlock(d1->d_sb); ++} +diff -Nur linux-2.6.31.4.orig/fs/aufs/spl.h linux-2.6.31.4/fs/aufs/spl.h +--- linux-2.6.31.4.orig/fs/aufs/spl.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/spl.h 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,57 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * simple list protected by a spinlock ++ */ ++ ++#ifndef __AUFS_SPL_H__ ++#define __AUFS_SPL_H__ ++ ++#ifdef __KERNEL__ ++ ++#include <linux/spinlock.h> ++#include <linux/list.h> ++ ++struct au_splhead { ++ spinlock_t spin; ++ struct list_head head; ++}; ++ ++static inline void au_spl_init(struct au_splhead *spl) ++{ ++ spin_lock_init(&spl->spin); ++ INIT_LIST_HEAD(&spl->head); ++} ++ ++static inline void au_spl_add(struct list_head *list, struct au_splhead *spl) ++{ ++ spin_lock(&spl->spin); ++ list_add(list, &spl->head); ++ spin_unlock(&spl->spin); ++} ++ ++static inline void au_spl_del(struct list_head *list, struct au_splhead *spl) ++{ ++ spin_lock(&spl->spin); ++ list_del(list); ++ spin_unlock(&spl->spin); ++} ++ ++#endif /* __KERNEL__ */ ++#endif /* __AUFS_SPL_H__ */ +diff -Nur linux-2.6.31.4.orig/fs/aufs/super.c linux-2.6.31.4/fs/aufs/super.c +--- linux-2.6.31.4.orig/fs/aufs/super.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/super.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,874 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * mount and super_block operations ++ */ ++ ++#include <linux/buffer_head.h> ++#include <linux/module.h> ++#include <linux/seq_file.h> ++#include <linux/statfs.h> ++#include "aufs.h" ++ ++/* ++ * super_operations ++ */ ++static struct inode *aufs_alloc_inode(struct super_block *sb __maybe_unused) ++{ ++ struct au_icntnr *c; ++ ++ c = au_cache_alloc_icntnr(); ++ if (c) { ++ inode_init_once(&c->vfs_inode); ++ c->vfs_inode.i_version = 1; /* sigen(sb); */ ++ c->iinfo.ii_hinode = NULL; ++ return &c->vfs_inode; ++ } ++ return NULL; ++} ++ ++static void aufs_destroy_inode(struct inode *inode) ++{ ++ au_iinfo_fin(inode); ++ au_cache_free_icntnr(container_of(inode, struct au_icntnr, vfs_inode)); ++} ++ ++struct inode *au_iget_locked(struct super_block *sb, ino_t ino) ++{ ++ struct inode *inode; ++ int err; ++ ++ inode = iget_locked(sb, ino); ++ if (unlikely(!inode)) { ++ inode = ERR_PTR(-ENOMEM); ++ goto out; ++ } ++ if (!(inode->i_state & I_NEW)) ++ goto out; ++ ++ err = au_xigen_new(inode); ++ if (!err) ++ err = au_iinfo_init(inode); ++ if (!err) ++ inode->i_version++; ++ else { ++ iget_failed(inode); ++ inode = ERR_PTR(err); ++ } ++ ++ out: ++ /* never return NULL */ ++ AuDebugOn(!inode); ++ AuTraceErrPtr(inode); ++ return inode; ++} ++ ++/* lock free root dinfo */ ++static int au_show_brs(struct seq_file *seq, struct super_block *sb) ++{ ++ int err; ++ aufs_bindex_t bindex, bend; ++ struct path path; ++ struct au_hdentry *hd; ++ struct au_branch *br; ++ ++ err = 0; ++ bend = au_sbend(sb); ++ hd = au_di(sb->s_root)->di_hdentry; ++ for (bindex = 0; !err && bindex <= bend; bindex++) { ++ br = au_sbr(sb, bindex); ++ path.mnt = br->br_mnt; ++ path.dentry = hd[bindex].hd_dentry; ++ err = au_seq_path(seq, &path); ++ if (err > 0) ++ err = seq_printf(seq, "=%s", ++ au_optstr_br_perm(br->br_perm)); ++ if (!err && bindex != bend) ++ err = seq_putc(seq, ':'); ++ } ++ ++ return err; ++} ++ ++static void au_show_wbr_create(struct seq_file *m, int v, ++ struct au_sbinfo *sbinfo) ++{ ++ const char *pat; ++ ++ AuRwMustAnyLock(&sbinfo->si_rwsem); ++ ++ seq_printf(m, ",create="); ++ pat = au_optstr_wbr_create(v); ++ switch (v) { ++ case AuWbrCreate_TDP: ++ case AuWbrCreate_RR: ++ case AuWbrCreate_MFS: ++ case AuWbrCreate_PMFS: ++ seq_printf(m, pat); ++ break; ++ case AuWbrCreate_MFSV: ++ seq_printf(m, /*pat*/"mfs:%lu", ++ sbinfo->si_wbr_mfs.mfs_expire / HZ); ++ break; ++ case AuWbrCreate_PMFSV: ++ seq_printf(m, /*pat*/"pmfs:%lu", ++ sbinfo->si_wbr_mfs.mfs_expire / HZ); ++ break; ++ case AuWbrCreate_MFSRR: ++ seq_printf(m, /*pat*/"mfsrr:%llu", ++ sbinfo->si_wbr_mfs.mfsrr_watermark); ++ break; ++ case AuWbrCreate_MFSRRV: ++ seq_printf(m, /*pat*/"mfsrr:%llu:%lu", ++ sbinfo->si_wbr_mfs.mfsrr_watermark, ++ sbinfo->si_wbr_mfs.mfs_expire / HZ); ++ break; ++ } ++} ++ ++static int au_show_xino(struct seq_file *seq, struct vfsmount *mnt) ++{ ++#ifdef CONFIG_SYSFS ++ return 0; ++#else ++ int err; ++ const int len = sizeof(AUFS_XINO_FNAME) - 1; ++ aufs_bindex_t bindex, brid; ++ struct super_block *sb; ++ struct qstr *name; ++ struct file *f; ++ struct dentry *d, *h_root; ++ ++ AuRwMustAnyLock(&sbinfo->si_rwsem); ++ ++ err = 0; ++ sb = mnt->mnt_sb; ++ f = au_sbi(sb)->si_xib; ++ if (!f) ++ goto out; ++ ++ /* stop printing the default xino path on the first writable branch */ ++ h_root = NULL; ++ brid = au_xino_brid(sb); ++ if (brid >= 0) { ++ bindex = au_br_index(sb, brid); ++ h_root = au_di(sb->s_root)->di_hdentry[0 + bindex].hd_dentry; ++ } ++ d = f->f_dentry; ++ name = &d->d_name; ++ /* safe ->d_parent because the file is unlinked */ ++ if (d->d_parent == h_root ++ && name->len == len ++ && !memcmp(name->name, AUFS_XINO_FNAME, len)) ++ goto out; ++ ++ seq_puts(seq, ",xino="); ++ err = au_xino_path(seq, f); ++ ++ out: ++ return err; ++#endif ++} ++ ++/* seq_file will re-call me in case of too long string */ ++static int aufs_show_options(struct seq_file *m, struct vfsmount *mnt) ++{ ++ int err, n; ++ unsigned int mnt_flags, v; ++ struct super_block *sb; ++ struct au_sbinfo *sbinfo; ++ ++#define AuBool(name, str) do { \ ++ v = au_opt_test(mnt_flags, name); \ ++ if (v != au_opt_test(AuOpt_Def, name)) \ ++ seq_printf(m, ",%s" #str, v ? "" : "no"); \ ++} while (0) ++ ++#define AuStr(name, str) do { \ ++ v = mnt_flags & AuOptMask_##name; \ ++ if (v != (AuOpt_Def & AuOptMask_##name)) \ ++ seq_printf(m, "," #str "=%s", au_optstr_##str(v)); \ ++} while (0) ++ ++#define AuUInt(name, str, val) do { \ ++ if (val != AUFS_##name##_DEF) \ ++ seq_printf(m, "," #str "=%u", val); \ ++} while (0) ++ ++ /* lock free root dinfo */ ++ sb = mnt->mnt_sb; ++ si_noflush_read_lock(sb); ++ sbinfo = au_sbi(sb); ++ seq_printf(m, ",si=%lx", sysaufs_si_id(sbinfo)); ++ ++ mnt_flags = au_mntflags(sb); ++ if (au_opt_test(mnt_flags, XINO)) { ++ err = au_show_xino(m, mnt); ++ if (unlikely(err)) ++ goto out; ++ } else ++ seq_puts(m, ",noxino"); ++ ++ AuBool(TRUNC_XINO, trunc_xino); ++ AuStr(UDBA, udba); ++ AuBool(SHWH, shwh); ++ AuBool(PLINK, plink); ++ /* AuBool(DIRPERM1, dirperm1); */ ++ /* AuBool(REFROF, refrof); */ ++ ++ v = sbinfo->si_wbr_create; ++ if (v != AuWbrCreate_Def) ++ au_show_wbr_create(m, v, sbinfo); ++ ++ v = sbinfo->si_wbr_copyup; ++ if (v != AuWbrCopyup_Def) ++ seq_printf(m, ",cpup=%s", au_optstr_wbr_copyup(v)); ++ ++ v = au_opt_test(mnt_flags, ALWAYS_DIROPQ); ++ if (v != au_opt_test(AuOpt_Def, ALWAYS_DIROPQ)) ++ seq_printf(m, ",diropq=%c", v ? 'a' : 'w'); ++ ++ AuUInt(DIRWH, dirwh, sbinfo->si_dirwh); ++ ++ n = sbinfo->si_rdcache / HZ; ++ AuUInt(RDCACHE, rdcache, n); ++ ++ AuUInt(RDBLK, rdblk, sbinfo->si_rdblk); ++ AuUInt(RDHASH, rdhash, sbinfo->si_rdhash); ++ ++ AuBool(SUM, sum); ++ /* AuBool(SUM_W, wsum); */ ++ AuBool(WARN_PERM, warn_perm); ++ AuBool(VERBOSE, verbose); ++ ++ out: ++ /* be sure to print "br:" last */ ++ if (!sysaufs_brs) { ++ seq_puts(m, ",br:"); ++ au_show_brs(m, sb); ++ } ++ si_read_unlock(sb); ++ return 0; ++ ++#undef Deleted ++#undef AuBool ++#undef AuStr ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* sum mode which returns the summation for statfs(2) */ ++ ++static u64 au_add_till_max(u64 a, u64 b) ++{ ++ u64 old; ++ ++ old = a; ++ a += b; ++ if (old < a) ++ return a; ++ return ULLONG_MAX; ++} ++ ++static int au_statfs_sum(struct super_block *sb, struct kstatfs *buf) ++{ ++ int err; ++ u64 blocks, bfree, bavail, files, ffree; ++ aufs_bindex_t bend, bindex, i; ++ unsigned char shared; ++ struct vfsmount *h_mnt; ++ struct super_block *h_sb; ++ ++ blocks = 0; ++ bfree = 0; ++ bavail = 0; ++ files = 0; ++ ffree = 0; ++ ++ err = 0; ++ bend = au_sbend(sb); ++ for (bindex = bend; bindex >= 0; bindex--) { ++ h_mnt = au_sbr_mnt(sb, bindex); ++ h_sb = h_mnt->mnt_sb; ++ shared = 0; ++ for (i = bindex + 1; !shared && i <= bend; i++) ++ shared = (au_sbr_sb(sb, i) == h_sb); ++ if (shared) ++ continue; ++ ++ /* sb->s_root for NFS is unreliable */ ++ err = vfs_statfs(h_mnt->mnt_root, buf); ++ if (unlikely(err)) ++ goto out; ++ ++ blocks = au_add_till_max(blocks, buf->f_blocks); ++ bfree = au_add_till_max(bfree, buf->f_bfree); ++ bavail = au_add_till_max(bavail, buf->f_bavail); ++ files = au_add_till_max(files, buf->f_files); ++ ffree = au_add_till_max(ffree, buf->f_ffree); ++ } ++ ++ buf->f_blocks = blocks; ++ buf->f_bfree = bfree; ++ buf->f_bavail = bavail; ++ buf->f_files = files; ++ buf->f_ffree = ffree; ++ ++ out: ++ return err; ++} ++ ++static int aufs_statfs(struct dentry *dentry, struct kstatfs *buf) ++{ ++ int err; ++ struct super_block *sb; ++ ++ /* lock free root dinfo */ ++ sb = dentry->d_sb; ++ si_noflush_read_lock(sb); ++ if (!au_opt_test(au_mntflags(sb), SUM)) ++ /* sb->s_root for NFS is unreliable */ ++ err = vfs_statfs(au_sbr_mnt(sb, 0)->mnt_root, buf); ++ else ++ err = au_statfs_sum(sb, buf); ++ si_read_unlock(sb); ++ ++ if (!err) { ++ buf->f_type = AUFS_SUPER_MAGIC; ++ buf->f_namelen -= AUFS_WH_PFX_LEN; ++ memset(&buf->f_fsid, 0, sizeof(buf->f_fsid)); ++ } ++ /* buf->f_bsize = buf->f_blocks = buf->f_bfree = buf->f_bavail = -1; */ ++ ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* try flushing the lower fs at aufs remount/unmount time */ ++ ++static void au_fsync_br(struct super_block *sb) ++{ ++ aufs_bindex_t bend, bindex; ++ int brperm; ++ struct au_branch *br; ++ struct super_block *h_sb; ++ ++ bend = au_sbend(sb); ++ for (bindex = 0; bindex < bend; bindex++) { ++ br = au_sbr(sb, bindex); ++ brperm = br->br_perm; ++ if (brperm == AuBrPerm_RR || brperm == AuBrPerm_RRWH) ++ continue; ++ h_sb = br->br_mnt->mnt_sb; ++ if (bdev_read_only(h_sb->s_bdev)) ++ continue; ++ ++ lockdep_off(); ++ down_write(&h_sb->s_umount); ++ shrink_dcache_sb(h_sb); ++ fsync_super(h_sb); ++ up_write(&h_sb->s_umount); ++ lockdep_on(); ++ } ++} ++ ++/* ++ * this IS NOT for super_operations. ++ * I guess it will be reverted someday. ++ */ ++static void aufs_umount_begin(struct super_block *sb) ++{ ++ struct au_sbinfo *sbinfo; ++ ++ sbinfo = au_sbi(sb); ++ if (!sbinfo) ++ return; ++ ++ si_write_lock(sb); ++ au_fsync_br(sb); ++ if (au_opt_test(au_mntflags(sb), PLINK)) ++ au_plink_put(sb); ++ if (sbinfo->si_wbr_create_ops->fin) ++ sbinfo->si_wbr_create_ops->fin(sb); ++ si_write_unlock(sb); ++} ++ ++/* final actions when unmounting a file system */ ++static void aufs_put_super(struct super_block *sb) ++{ ++ struct au_sbinfo *sbinfo; ++ ++ sbinfo = au_sbi(sb); ++ if (!sbinfo) ++ return; ++ ++ aufs_umount_begin(sb); ++ dbgaufs_si_fin(sbinfo); ++ kobject_put(&sbinfo->si_kobj); ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* ++ * refresh dentry and inode at remount time. ++ */ ++static int do_refresh(struct dentry *dentry, mode_t type, ++ unsigned int dir_flags) ++{ ++ int err; ++ struct dentry *parent; ++ ++ di_write_lock_child(dentry); ++ parent = dget_parent(dentry); ++ di_read_lock_parent(parent, AuLock_IR); ++ ++ /* returns the number of positive dentries */ ++ err = au_refresh_hdentry(dentry, type); ++ if (err >= 0) { ++ struct inode *inode = dentry->d_inode; ++ err = au_refresh_hinode(inode, dentry); ++ if (!err && type == S_IFDIR) ++ au_reset_hinotify(inode, dir_flags); ++ } ++ if (unlikely(err)) ++ AuErr("unrecoverable error %d, %.*s\n", err, AuDLNPair(dentry)); ++ ++ di_read_unlock(parent, AuLock_IR); ++ dput(parent); ++ di_write_unlock(dentry); ++ ++ return err; ++} ++ ++static int test_dir(struct dentry *dentry, void *arg __maybe_unused) ++{ ++ return S_ISDIR(dentry->d_inode->i_mode); ++} ++ ++/* gave up consolidating with refresh_nondir() */ ++static int refresh_dir(struct dentry *root, unsigned int sigen) ++{ ++ int err, i, j, ndentry, e; ++ struct au_dcsub_pages dpages; ++ struct au_dpage *dpage; ++ struct dentry **dentries; ++ struct inode *inode; ++ const unsigned int flags = au_hi_flags(root->d_inode, /*isdir*/1); ++ ++ err = 0; ++ list_for_each_entry(inode, &root->d_sb->s_inodes, i_sb_list) ++ if (S_ISDIR(inode->i_mode) && au_iigen(inode) != sigen) { ++ ii_write_lock_child(inode); ++ e = au_refresh_hinode_self(inode, /*do_attr*/1); ++ ii_write_unlock(inode); ++ if (unlikely(e)) { ++ AuDbg("e %d, i%lu\n", e, inode->i_ino); ++ if (!err) ++ err = e; ++ /* go on even if err */ ++ } ++ } ++ ++ e = au_dpages_init(&dpages, GFP_NOFS); ++ if (unlikely(e)) { ++ if (!err) ++ err = e; ++ goto out; ++ } ++ e = au_dcsub_pages(&dpages, root, test_dir, NULL); ++ if (unlikely(e)) { ++ if (!err) ++ err = e; ++ goto out_dpages; ++ } ++ ++ for (i = 0; !e && i < dpages.ndpage; i++) { ++ dpage = dpages.dpages + i; ++ dentries = dpage->dentries; ++ ndentry = dpage->ndentry; ++ for (j = 0; !e && j < ndentry; j++) { ++ struct dentry *d; ++ ++ d = dentries[j]; ++ au_dbg_verify_dir_parent(d, sigen); ++ if (au_digen(d) != sigen) { ++ e = do_refresh(d, S_IFDIR, flags); ++ if (unlikely(e && !err)) ++ err = e; ++ /* break on err */ ++ } ++ } ++ } ++ ++ out_dpages: ++ au_dpages_free(&dpages); ++ out: ++ return err; ++} ++ ++static int test_nondir(struct dentry *dentry, void *arg __maybe_unused) ++{ ++ return !S_ISDIR(dentry->d_inode->i_mode); ++} ++ ++static int refresh_nondir(struct dentry *root, unsigned int sigen, ++ int do_dentry) ++{ ++ int err, i, j, ndentry, e; ++ struct au_dcsub_pages dpages; ++ struct au_dpage *dpage; ++ struct dentry **dentries; ++ struct inode *inode; ++ ++ err = 0; ++ list_for_each_entry(inode, &root->d_sb->s_inodes, i_sb_list) ++ if (!S_ISDIR(inode->i_mode) && au_iigen(inode) != sigen) { ++ ii_write_lock_child(inode); ++ e = au_refresh_hinode_self(inode, /*do_attr*/1); ++ ii_write_unlock(inode); ++ if (unlikely(e)) { ++ AuDbg("e %d, i%lu\n", e, inode->i_ino); ++ if (!err) ++ err = e; ++ /* go on even if err */ ++ } ++ } ++ ++ if (!do_dentry) ++ goto out; ++ ++ e = au_dpages_init(&dpages, GFP_NOFS); ++ if (unlikely(e)) { ++ if (!err) ++ err = e; ++ goto out; ++ } ++ e = au_dcsub_pages(&dpages, root, test_nondir, NULL); ++ if (unlikely(e)) { ++ if (!err) ++ err = e; ++ goto out_dpages; ++ } ++ ++ for (i = 0; i < dpages.ndpage; i++) { ++ dpage = dpages.dpages + i; ++ dentries = dpage->dentries; ++ ndentry = dpage->ndentry; ++ for (j = 0; j < ndentry; j++) { ++ struct dentry *d; ++ ++ d = dentries[j]; ++ au_dbg_verify_nondir_parent(d, sigen); ++ inode = d->d_inode; ++ if (inode && au_digen(d) != sigen) { ++ e = do_refresh(d, inode->i_mode & S_IFMT, ++ /*dir_flags*/0); ++ if (unlikely(e && !err)) ++ err = e; ++ /* go on even err */ ++ } ++ } ++ } ++ ++ out_dpages: ++ au_dpages_free(&dpages); ++ out: ++ return err; ++} ++ ++static void au_remount_refresh(struct super_block *sb, unsigned int flags) ++{ ++ int err; ++ unsigned int sigen; ++ struct au_sbinfo *sbinfo; ++ struct dentry *root; ++ struct inode *inode; ++ ++ au_sigen_inc(sb); ++ sigen = au_sigen(sb); ++ sbinfo = au_sbi(sb); ++ au_fclr_si(sbinfo, FAILED_REFRESH_DIRS); ++ ++ root = sb->s_root; ++ DiMustNoWaiters(root); ++ inode = root->d_inode; ++ IiMustNoWaiters(inode); ++ au_reset_hinotify(inode, au_hi_flags(inode, /*isdir*/1)); ++ di_write_unlock(root); ++ ++ err = refresh_dir(root, sigen); ++ if (unlikely(err)) { ++ au_fset_si(sbinfo, FAILED_REFRESH_DIRS); ++ AuWarn("Refreshing directories failed, ignored (%d)\n", err); ++ } ++ ++ if (au_ftest_opts(flags, REFRESH_NONDIR)) { ++ err = refresh_nondir(root, sigen, !err); ++ if (unlikely(err)) ++ AuWarn("Refreshing non-directories failed, ignored" ++ "(%d)\n", err); ++ } ++ ++ /* aufs_write_lock() calls ..._child() */ ++ di_write_lock_child(root); ++ au_cpup_attr_all(root->d_inode, /*force*/1); ++} ++ ++/* stop extra interpretation of errno in mount(8), and strange error messages */ ++static int cvt_err(int err) ++{ ++ AuTraceErr(err); ++ ++ switch (err) { ++ case -ENOENT: ++ case -ENOTDIR: ++ case -EEXIST: ++ case -EIO: ++ err = -EINVAL; ++ } ++ return err; ++} ++ ++static int aufs_remount_fs(struct super_block *sb, int *flags, char *data) ++{ ++ int err; ++ struct au_opts opts; ++ struct dentry *root; ++ struct inode *inode; ++ struct au_sbinfo *sbinfo; ++ ++ err = 0; ++ root = sb->s_root; ++ if (!data || !*data) { ++ aufs_write_lock(root); ++ err = au_opts_verify(sb, *flags, /*pending*/0); ++ if (!err) ++ au_fsync_br(sb); ++ aufs_write_unlock(root); ++ goto out; ++ } ++ ++ err = -ENOMEM; ++ memset(&opts, 0, sizeof(opts)); ++ opts.opt = (void *)__get_free_page(GFP_NOFS); ++ if (unlikely(!opts.opt)) ++ goto out; ++ opts.max_opt = PAGE_SIZE / sizeof(*opts.opt); ++ opts.flags = AuOpts_REMOUNT; ++ opts.sb_flags = *flags; ++ ++ /* parse it before aufs lock */ ++ err = au_opts_parse(sb, data, &opts); ++ if (unlikely(err)) ++ goto out_opts; ++ ++ sbinfo = au_sbi(sb); ++ inode = root->d_inode; ++ mutex_lock(&inode->i_mutex); ++ aufs_write_lock(root); ++ au_fsync_br(sb); ++ ++ /* au_opts_remount() may return an error */ ++ err = au_opts_remount(sb, &opts); ++ au_opts_free(&opts); ++ ++ if (au_ftest_opts(opts.flags, REFRESH_DIR) ++ || au_ftest_opts(opts.flags, REFRESH_NONDIR)) ++ au_remount_refresh(sb, opts.flags); ++ ++ aufs_write_unlock(root); ++ mutex_unlock(&inode->i_mutex); ++ ++ out_opts: ++ free_page((unsigned long)opts.opt); ++ out: ++ err = cvt_err(err); ++ AuTraceErr(err); ++ return err; ++} ++ ++static struct super_operations aufs_sop = { ++ .alloc_inode = aufs_alloc_inode, ++ .destroy_inode = aufs_destroy_inode, ++ .drop_inode = generic_delete_inode, ++ .show_options = aufs_show_options, ++ .statfs = aufs_statfs, ++ .put_super = aufs_put_super, ++ .remount_fs = aufs_remount_fs ++}; ++ ++/* ---------------------------------------------------------------------- */ ++ ++static int alloc_root(struct super_block *sb) ++{ ++ int err; ++ struct inode *inode; ++ struct dentry *root; ++ ++ err = -ENOMEM; ++ inode = au_iget_locked(sb, AUFS_ROOT_INO); ++ err = PTR_ERR(inode); ++ if (IS_ERR(inode)) ++ goto out; ++ ++ inode->i_op = &aufs_dir_iop; ++ inode->i_fop = &aufs_dir_fop; ++ inode->i_mode = S_IFDIR; ++ inode->i_nlink = 2; ++ unlock_new_inode(inode); ++ ++ root = d_alloc_root(inode); ++ if (unlikely(!root)) ++ goto out_iput; ++ err = PTR_ERR(root); ++ if (IS_ERR(root)) ++ goto out_iput; ++ ++ err = au_alloc_dinfo(root); ++ if (!err) { ++ sb->s_root = root; ++ return 0; /* success */ ++ } ++ dput(root); ++ goto out; /* do not iput */ ++ ++ out_iput: ++ iget_failed(inode); ++ iput(inode); ++ out: ++ return err; ++ ++} ++ ++static int aufs_fill_super(struct super_block *sb, void *raw_data, ++ int silent __maybe_unused) ++{ ++ int err; ++ struct au_opts opts; ++ struct dentry *root; ++ struct inode *inode; ++ char *arg = raw_data; ++ ++ if (unlikely(!arg || !*arg)) { ++ err = -EINVAL; ++ AuErr("no arg\n"); ++ goto out; ++ } ++ ++ err = -ENOMEM; ++ memset(&opts, 0, sizeof(opts)); ++ opts.opt = (void *)__get_free_page(GFP_NOFS); ++ if (unlikely(!opts.opt)) ++ goto out; ++ opts.max_opt = PAGE_SIZE / sizeof(*opts.opt); ++ opts.sb_flags = sb->s_flags; ++ ++ err = au_si_alloc(sb); ++ if (unlikely(err)) ++ goto out_opts; ++ ++ /* all timestamps always follow the ones on the branch */ ++ sb->s_flags |= MS_NOATIME | MS_NODIRATIME; ++ sb->s_op = &aufs_sop; ++ sb->s_magic = AUFS_SUPER_MAGIC; ++ sb->s_maxbytes = 0; ++ au_export_init(sb); ++ ++ err = alloc_root(sb); ++ if (unlikely(err)) { ++ si_write_unlock(sb); ++ goto out_info; ++ } ++ root = sb->s_root; ++ inode = root->d_inode; ++ ++ /* ++ * actually we can parse options regardless aufs lock here. ++ * but at remount time, parsing must be done before aufs lock. ++ * so we follow the same rule. ++ */ ++ ii_write_lock_parent(inode); ++ aufs_write_unlock(root); ++ err = au_opts_parse(sb, arg, &opts); ++ if (unlikely(err)) ++ goto out_root; ++ ++ /* lock vfs_inode first, then aufs. */ ++ mutex_lock(&inode->i_mutex); ++ inode->i_op = &aufs_dir_iop; ++ inode->i_fop = &aufs_dir_fop; ++ aufs_write_lock(root); ++ err = au_opts_mount(sb, &opts); ++ au_opts_free(&opts); ++ if (unlikely(err)) ++ goto out_unlock; ++ aufs_write_unlock(root); ++ mutex_unlock(&inode->i_mutex); ++ goto out_opts; /* success */ ++ ++ out_unlock: ++ aufs_write_unlock(root); ++ mutex_unlock(&inode->i_mutex); ++ out_root: ++ dput(root); ++ sb->s_root = NULL; ++ out_info: ++ kobject_put(&au_sbi(sb)->si_kobj); ++ sb->s_fs_info = NULL; ++ out_opts: ++ free_page((unsigned long)opts.opt); ++ out: ++ AuTraceErr(err); ++ err = cvt_err(err); ++ AuTraceErr(err); ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static int aufs_get_sb(struct file_system_type *fs_type, int flags, ++ const char *dev_name __maybe_unused, void *raw_data, ++ struct vfsmount *mnt) ++{ ++ int err; ++ struct super_block *sb; ++ ++ /* all timestamps always follow the ones on the branch */ ++ /* mnt->mnt_flags |= MNT_NOATIME | MNT_NODIRATIME; */ ++ err = get_sb_nodev(fs_type, flags, raw_data, aufs_fill_super, mnt); ++ if (!err) { ++ sb = mnt->mnt_sb; ++ si_write_lock(sb); ++ sysaufs_brs_add(sb, 0); ++ si_write_unlock(sb); ++ } ++ return err; ++} ++ ++struct file_system_type aufs_fs_type = { ++ .name = AUFS_FSTYPE, ++ .fs_flags = ++ FS_RENAME_DOES_D_MOVE /* a race between rename and others */ ++ | FS_REVAL_DOT, /* for NFS branch and udba */ ++ .get_sb = aufs_get_sb, ++ .kill_sb = generic_shutdown_super, ++ /* no need to __module_get() and module_put(). */ ++ .owner = THIS_MODULE, ++}; +diff -Nur linux-2.6.31.4.orig/fs/aufs/super.h linux-2.6.31.4/fs/aufs/super.h +--- linux-2.6.31.4.orig/fs/aufs/super.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/super.h 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,384 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * super_block operations ++ */ ++ ++#ifndef __AUFS_SUPER_H__ ++#define __AUFS_SUPER_H__ ++ ++#ifdef __KERNEL__ ++ ++#include <linux/fs.h> ++#include <linux/aufs_type.h> ++#include "rwsem.h" ++#include "spl.h" ++#include "wkq.h" ++ ++typedef ssize_t (*au_readf_t)(struct file *, char __user *, size_t, loff_t *); ++typedef ssize_t (*au_writef_t)(struct file *, const char __user *, size_t, ++ loff_t *); ++ ++/* policies to select one among multiple writable branches */ ++struct au_wbr_copyup_operations { ++ int (*copyup)(struct dentry *dentry); ++}; ++ ++struct au_wbr_create_operations { ++ int (*create)(struct dentry *dentry, int isdir); ++ int (*init)(struct super_block *sb); ++ int (*fin)(struct super_block *sb); ++}; ++ ++struct au_wbr_mfs { ++ struct mutex mfs_lock; /* protect this structure */ ++ unsigned long mfs_jiffy; ++ unsigned long mfs_expire; ++ aufs_bindex_t mfs_bindex; ++ ++ unsigned long long mfsrr_bytes; ++ unsigned long long mfsrr_watermark; ++}; ++ ++struct au_branch; ++struct au_sbinfo { ++ /* nowait tasks in the system-wide workqueue */ ++ struct au_nowait_tasks si_nowait; ++ ++ struct au_rwsem si_rwsem; ++ ++ /* branch management */ ++ unsigned int si_generation; ++ ++ /* see above flags */ ++ unsigned char au_si_status; ++ ++ aufs_bindex_t si_bend; ++ aufs_bindex_t si_last_br_id; ++ struct au_branch **si_branch; ++ ++ /* policy to select a writable branch */ ++ unsigned char si_wbr_copyup; ++ unsigned char si_wbr_create; ++ struct au_wbr_copyup_operations *si_wbr_copyup_ops; ++ struct au_wbr_create_operations *si_wbr_create_ops; ++ ++ /* round robin */ ++ atomic_t si_wbr_rr_next; ++ ++ /* most free space */ ++ struct au_wbr_mfs si_wbr_mfs; ++ ++ /* mount flags */ ++ /* include/asm-ia64/siginfo.h defines a macro named si_flags */ ++ unsigned int si_mntflags; ++ ++ /* external inode number (bitmap and translation table) */ ++ au_readf_t si_xread; ++ au_writef_t si_xwrite; ++ struct file *si_xib; ++ struct mutex si_xib_mtx; /* protect xib members */ ++ unsigned long *si_xib_buf; ++ unsigned long si_xib_last_pindex; ++ int si_xib_next_bit; ++ aufs_bindex_t si_xino_brid; ++ /* reserved for future use */ ++ /* unsigned long long si_xib_limit; */ /* Max xib file size */ ++ ++#ifdef CONFIG_AUFS_EXPORT ++ /* i_generation */ ++ struct file *si_xigen; ++ atomic_t si_xigen_next; ++#endif ++ ++ /* vdir parameters */ ++ unsigned long si_rdcache; /* max cache time in HZ */ ++ unsigned int si_rdblk; /* deblk size */ ++ unsigned int si_rdhash; /* hash size */ ++ ++ /* ++ * If the number of whiteouts are larger than si_dirwh, leave all of ++ * them after au_whtmp_ren to reduce the cost of rmdir(2). ++ * future fsck.aufs or kernel thread will remove them later. ++ * Otherwise, remove all whiteouts and the dir in rmdir(2). ++ */ ++ unsigned int si_dirwh; ++ ++ /* ++ * rename(2) a directory with all children. ++ */ ++ /* reserved for future use */ ++ /* int si_rendir; */ ++ ++ /* pseudo_link list */ ++ struct au_splhead si_plink; ++ wait_queue_head_t si_plink_wq; ++ ++ /* ++ * sysfs and lifetime management. ++ * this is not a small structure and it may be a waste of memory in case ++ * of sysfs is disabled, particulary when many aufs-es are mounted. ++ * but using sysfs is majority. ++ */ ++ struct kobject si_kobj; ++#ifdef CONFIG_DEBUG_FS ++ struct dentry *si_dbgaufs, *si_dbgaufs_xib; ++#ifdef CONFIG_AUFS_EXPORT ++ struct dentry *si_dbgaufs_xigen; ++#endif ++#endif ++ ++ /* dirty, necessary for unmounting, sysfs and sysrq */ ++ struct super_block *si_sb; ++}; ++ ++/* sbinfo status flags */ ++/* ++ * set true when refresh_dirs() failed at remount time. ++ * then try refreshing dirs at access time again. ++ * if it is false, refreshing dirs at access time is unnecesary ++ */ ++#define AuSi_FAILED_REFRESH_DIRS 1 ++#define AuSi_MAINTAIN_PLINK (1 << 1) /* ioctl */ ++static inline unsigned char au_do_ftest_si(struct au_sbinfo *sbi, ++ unsigned int flag) ++{ ++ AuRwMustAnyLock(&sbi->si_rwsem); ++ return sbi->au_si_status & flag; ++} ++#define au_ftest_si(sbinfo, name) au_do_ftest_si(sbinfo, AuSi_##name) ++#define au_fset_si(sbinfo, name) do { \ ++ AuRwMustWriteLock(&(sbinfo)->si_rwsem); \ ++ (sbinfo)->au_si_status |= AuSi_##name; \ ++} while (0) ++#define au_fclr_si(sbinfo, name) do { \ ++ AuRwMustWriteLock(&(sbinfo)->si_rwsem); \ ++ (sbinfo)->au_si_status &= ~AuSi_##name; \ ++} while (0) ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* policy to select one among writable branches */ ++#define AuWbrCopyup(sbinfo, args...) \ ++ ((sbinfo)->si_wbr_copyup_ops->copyup(args)) ++#define AuWbrCreate(sbinfo, args...) \ ++ ((sbinfo)->si_wbr_create_ops->create(args)) ++ ++/* flags for si_read_lock()/aufs_read_lock()/di_read_lock() */ ++#define AuLock_DW 1 /* write-lock dentry */ ++#define AuLock_IR (1 << 1) /* read-lock inode */ ++#define AuLock_IW (1 << 2) /* write-lock inode */ ++#define AuLock_FLUSH (1 << 3) /* wait for 'nowait' tasks */ ++#define AuLock_DIR (1 << 4) /* target is a dir */ ++#define au_ftest_lock(flags, name) ((flags) & AuLock_##name) ++#define au_fset_lock(flags, name) { (flags) |= AuLock_##name; } ++#define au_fclr_lock(flags, name) { (flags) &= ~AuLock_##name; } ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* super.c */ ++extern struct file_system_type aufs_fs_type; ++struct inode *au_iget_locked(struct super_block *sb, ino_t ino); ++ ++/* sbinfo.c */ ++void au_si_free(struct kobject *kobj); ++int au_si_alloc(struct super_block *sb); ++int au_sbr_realloc(struct au_sbinfo *sbinfo, int nbr); ++ ++unsigned int au_sigen_inc(struct super_block *sb); ++aufs_bindex_t au_new_br_id(struct super_block *sb); ++ ++void aufs_read_lock(struct dentry *dentry, int flags); ++void aufs_read_unlock(struct dentry *dentry, int flags); ++void aufs_write_lock(struct dentry *dentry); ++void aufs_write_unlock(struct dentry *dentry); ++void aufs_read_and_write_lock2(struct dentry *d1, struct dentry *d2, int isdir); ++void aufs_read_and_write_unlock2(struct dentry *d1, struct dentry *d2); ++ ++/* wbr_policy.c */ ++extern struct au_wbr_copyup_operations au_wbr_copyup_ops[]; ++extern struct au_wbr_create_operations au_wbr_create_ops[]; ++int au_cpdown_dirs(struct dentry *dentry, aufs_bindex_t bdst); ++ ++/* ---------------------------------------------------------------------- */ ++ ++static inline struct au_sbinfo *au_sbi(struct super_block *sb) ++{ ++ return sb->s_fs_info; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++#ifdef CONFIG_AUFS_EXPORT ++void au_export_init(struct super_block *sb); ++ ++static inline int au_test_nfsd(struct task_struct *tsk) ++{ ++ return !tsk->mm && !strcmp(tsk->comm, "nfsd"); ++} ++ ++int au_xigen_inc(struct inode *inode); ++int au_xigen_new(struct inode *inode); ++int au_xigen_set(struct super_block *sb, struct file *base); ++void au_xigen_clr(struct super_block *sb); ++ ++static inline int au_busy_or_stale(void) ++{ ++ if (!au_test_nfsd(current)) ++ return -EBUSY; ++ return -ESTALE; ++} ++#else ++static inline void au_export_init(struct super_block *sb) ++{ ++ /* nothing */ ++} ++ ++static inline int au_test_nfsd(struct task_struct *tsk) ++{ ++ return 0; ++} ++ ++static inline int au_xigen_inc(struct inode *inode) ++{ ++ return 0; ++} ++ ++static inline int au_xigen_new(struct inode *inode) ++{ ++ return 0; ++} ++ ++static inline int au_xigen_set(struct super_block *sb, struct file *base) ++{ ++ return 0; ++} ++ ++static inline void au_xigen_clr(struct super_block *sb) ++{ ++ /* empty */ ++} ++ ++static inline int au_busy_or_stale(void) ++{ ++ return -EBUSY; ++} ++#endif /* CONFIG_AUFS_EXPORT */ ++ ++/* ---------------------------------------------------------------------- */ ++ ++static inline void dbgaufs_si_null(struct au_sbinfo *sbinfo) ++{ ++ /* ++ * This function is a dynamic '__init' fucntion actually, ++ * so the tiny check for si_rwsem is unnecessary. ++ */ ++ /* AuRwMustWriteLock(&sbinfo->si_rwsem); */ ++#ifdef CONFIG_DEBUG_FS ++ sbinfo->si_dbgaufs = NULL; ++ sbinfo->si_dbgaufs_xib = NULL; ++#ifdef CONFIG_AUFS_EXPORT ++ sbinfo->si_dbgaufs_xigen = NULL; ++#endif ++#endif ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* lock superblock. mainly for entry point functions */ ++/* ++ * si_noflush_read_lock, si_noflush_write_lock, ++ * si_read_unlock, si_write_unlock, si_downgrade_lock ++ */ ++AuSimpleLockRwsemFuncs(si_noflush, struct super_block *sb, ++ &au_sbi(sb)->si_rwsem); ++AuSimpleUnlockRwsemFuncs(si, struct super_block *sb, &au_sbi(sb)->si_rwsem); ++ ++#define SiMustNoWaiters(sb) AuRwMustNoWaiters(&au_sbi(sb)->si_rwsem) ++#define SiMustAnyLock(sb) AuRwMustAnyLock(&au_sbi(sb)->si_rwsem) ++#define SiMustWriteLock(sb) AuRwMustWriteLock(&au_sbi(sb)->si_rwsem) ++ ++static inline void si_read_lock(struct super_block *sb, int flags) ++{ ++ if (au_ftest_lock(flags, FLUSH)) ++ au_nwt_flush(&au_sbi(sb)->si_nowait); ++ si_noflush_read_lock(sb); ++} ++ ++static inline void si_write_lock(struct super_block *sb) ++{ ++ au_nwt_flush(&au_sbi(sb)->si_nowait); ++ si_noflush_write_lock(sb); ++} ++ ++static inline int si_read_trylock(struct super_block *sb, int flags) ++{ ++ if (au_ftest_lock(flags, FLUSH)) ++ au_nwt_flush(&au_sbi(sb)->si_nowait); ++ return si_noflush_read_trylock(sb); ++} ++ ++static inline int si_write_trylock(struct super_block *sb, int flags) ++{ ++ if (au_ftest_lock(flags, FLUSH)) ++ au_nwt_flush(&au_sbi(sb)->si_nowait); ++ return si_noflush_write_trylock(sb); ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static inline aufs_bindex_t au_sbend(struct super_block *sb) ++{ ++ SiMustAnyLock(sb); ++ return au_sbi(sb)->si_bend; ++} ++ ++static inline unsigned int au_mntflags(struct super_block *sb) ++{ ++ SiMustAnyLock(sb); ++ return au_sbi(sb)->si_mntflags; ++} ++ ++static inline unsigned int au_sigen(struct super_block *sb) ++{ ++ SiMustAnyLock(sb); ++ return au_sbi(sb)->si_generation; ++} ++ ++static inline struct au_branch *au_sbr(struct super_block *sb, ++ aufs_bindex_t bindex) ++{ ++ SiMustAnyLock(sb); ++ return au_sbi(sb)->si_branch[0 + bindex]; ++} ++ ++static inline void au_xino_brid_set(struct super_block *sb, aufs_bindex_t brid) ++{ ++ SiMustWriteLock(sb); ++ au_sbi(sb)->si_xino_brid = brid; ++} ++ ++static inline aufs_bindex_t au_xino_brid(struct super_block *sb) ++{ ++ SiMustAnyLock(sb); ++ return au_sbi(sb)->si_xino_brid; ++} ++ ++#endif /* __KERNEL__ */ ++#endif /* __AUFS_SUPER_H__ */ +diff -Nur linux-2.6.31.4.orig/fs/aufs/sysaufs.c linux-2.6.31.4/fs/aufs/sysaufs.c +--- linux-2.6.31.4.orig/fs/aufs/sysaufs.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/sysaufs.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,104 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * sysfs interface and lifetime management ++ * they are necessary regardless sysfs is disabled. ++ */ ++ ++#include <linux/fs.h> ++#include <linux/random.h> ++#include <linux/sysfs.h> ++#include "aufs.h" ++ ++unsigned long sysaufs_si_mask; ++struct kset *sysaufs_ket; ++ ++#define AuSiAttr(_name) { \ ++ .attr = { .name = __stringify(_name), .mode = 0444 }, \ ++ .show = sysaufs_si_##_name, \ ++} ++ ++static struct sysaufs_si_attr sysaufs_si_attr_xi_path = AuSiAttr(xi_path); ++struct attribute *sysaufs_si_attrs[] = { ++ &sysaufs_si_attr_xi_path.attr, ++ NULL, ++}; ++ ++static struct sysfs_ops au_sbi_ops = { ++ .show = sysaufs_si_show ++}; ++ ++static struct kobj_type au_sbi_ktype = { ++ .release = au_si_free, ++ .sysfs_ops = &au_sbi_ops, ++ .default_attrs = sysaufs_si_attrs ++}; ++ ++/* ---------------------------------------------------------------------- */ ++ ++int sysaufs_si_init(struct au_sbinfo *sbinfo) ++{ ++ int err; ++ ++ sbinfo->si_kobj.kset = sysaufs_ket; ++ /* cf. sysaufs_name() */ ++ err = kobject_init_and_add ++ (&sbinfo->si_kobj, &au_sbi_ktype, /*&sysaufs_ket->kobj*/NULL, ++ SysaufsSiNamePrefix "%lx", sysaufs_si_id(sbinfo)); ++ ++ dbgaufs_si_null(sbinfo); ++ if (!err) { ++ err = dbgaufs_si_init(sbinfo); ++ if (unlikely(err)) ++ kobject_put(&sbinfo->si_kobj); ++ } ++ return err; ++} ++ ++void sysaufs_fin(void) ++{ ++ dbgaufs_fin(); ++ sysfs_remove_group(&sysaufs_ket->kobj, sysaufs_attr_group); ++ kset_unregister(sysaufs_ket); ++} ++ ++int __init sysaufs_init(void) ++{ ++ int err; ++ ++ do { ++ get_random_bytes(&sysaufs_si_mask, sizeof(sysaufs_si_mask)); ++ } while (!sysaufs_si_mask); ++ ++ sysaufs_ket = kset_create_and_add(AUFS_NAME, NULL, fs_kobj); ++ err = PTR_ERR(sysaufs_ket); ++ if (IS_ERR(sysaufs_ket)) ++ goto out; ++ err = sysfs_create_group(&sysaufs_ket->kobj, sysaufs_attr_group); ++ if (unlikely(err)) { ++ kset_unregister(sysaufs_ket); ++ goto out; ++ } ++ ++ err = dbgaufs_init(); ++ if (unlikely(err)) ++ sysaufs_fin(); ++ out: ++ return err; ++} +diff -Nur linux-2.6.31.4.orig/fs/aufs/sysaufs.h linux-2.6.31.4/fs/aufs/sysaufs.h +--- linux-2.6.31.4.orig/fs/aufs/sysaufs.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/sysaufs.h 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,120 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * sysfs interface and mount lifetime management ++ */ ++ ++#ifndef __SYSAUFS_H__ ++#define __SYSAUFS_H__ ++ ++#ifdef __KERNEL__ ++ ++#include <linux/sysfs.h> ++#include <linux/aufs_type.h> ++#include "module.h" ++ ++struct super_block; ++struct au_sbinfo; ++ ++struct sysaufs_si_attr { ++ struct attribute attr; ++ int (*show)(struct seq_file *seq, struct super_block *sb); ++}; ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* sysaufs.c */ ++extern unsigned long sysaufs_si_mask; ++extern struct kset *sysaufs_ket; ++extern struct attribute *sysaufs_si_attrs[]; ++int sysaufs_si_init(struct au_sbinfo *sbinfo); ++int __init sysaufs_init(void); ++void sysaufs_fin(void); ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* some people doesn't like to show a pointer in kernel */ ++static inline unsigned long sysaufs_si_id(struct au_sbinfo *sbinfo) ++{ ++ return sysaufs_si_mask ^ (unsigned long)sbinfo; ++} ++ ++#define SysaufsSiNamePrefix "si_" ++#define SysaufsSiNameLen (sizeof(SysaufsSiNamePrefix) + 16) ++static inline void sysaufs_name(struct au_sbinfo *sbinfo, char *name) ++{ ++ snprintf(name, SysaufsSiNameLen, SysaufsSiNamePrefix "%lx", ++ sysaufs_si_id(sbinfo)); ++} ++ ++struct au_branch; ++#ifdef CONFIG_SYSFS ++/* sysfs.c */ ++extern struct attribute_group *sysaufs_attr_group; ++ ++int sysaufs_si_xi_path(struct seq_file *seq, struct super_block *sb); ++ssize_t sysaufs_si_show(struct kobject *kobj, struct attribute *attr, ++ char *buf); ++ ++void sysaufs_br_init(struct au_branch *br); ++void sysaufs_brs_add(struct super_block *sb, aufs_bindex_t bindex); ++void sysaufs_brs_del(struct super_block *sb, aufs_bindex_t bindex); ++ ++#define sysaufs_brs_init() do {} while (0) ++ ++#else ++#define sysaufs_attr_group NULL ++ ++static inline ++int sysaufs_si_xi_path(struct seq_file *seq, struct super_block *sb) ++{ ++ return 0; ++} ++ ++static inline ++ssize_t sysaufs_si_show(struct kobject *kobj, struct attribute *attr, ++ char *buf) ++{ ++ return 0; ++} ++ ++static inline void sysaufs_br_init(struct au_branch *br) ++{ ++ /* empty */ ++} ++ ++static inline void sysaufs_brs_add(struct super_block *sb, aufs_bindex_t bindex) ++{ ++ /* nothing */ ++} ++ ++static inline void sysaufs_brs_del(struct super_block *sb, aufs_bindex_t bindex) ++{ ++ /* nothing */ ++} ++ ++static inline void sysaufs_brs_init(void) ++{ ++ sysaufs_brs = 0; ++} ++ ++#endif /* CONFIG_SYSFS */ ++ ++#endif /* __KERNEL__ */ ++#endif /* __SYSAUFS_H__ */ +diff -Nur linux-2.6.31.4.orig/fs/aufs/sysfs.c linux-2.6.31.4/fs/aufs/sysfs.c +--- linux-2.6.31.4.orig/fs/aufs/sysfs.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/sysfs.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,210 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * sysfs interface ++ */ ++ ++#include <linux/fs.h> ++#include <linux/module.h> ++#include <linux/seq_file.h> ++#include <linux/sysfs.h> ++#include "aufs.h" ++ ++static struct attribute *au_attr[] = { ++ NULL, /* need to NULL terminate the list of attributes */ ++}; ++ ++static struct attribute_group sysaufs_attr_group_body = { ++ .attrs = au_attr ++}; ++ ++struct attribute_group *sysaufs_attr_group = &sysaufs_attr_group_body; ++ ++/* ---------------------------------------------------------------------- */ ++ ++int sysaufs_si_xi_path(struct seq_file *seq, struct super_block *sb) ++{ ++ int err; ++ ++ SiMustAnyLock(sb); ++ ++ err = 0; ++ if (au_opt_test(au_mntflags(sb), XINO)) { ++ err = au_xino_path(seq, au_sbi(sb)->si_xib); ++ seq_putc(seq, '\n'); ++ } ++ return err; ++} ++ ++/* ++ * the lifetime of branch is independent from the entry under sysfs. ++ * sysfs handles the lifetime of the entry, and never call ->show() after it is ++ * unlinked. ++ */ ++static int sysaufs_si_br(struct seq_file *seq, struct super_block *sb, ++ aufs_bindex_t bindex) ++{ ++ struct path path; ++ struct dentry *root; ++ struct au_branch *br; ++ ++ AuDbg("b%d\n", bindex); ++ ++ root = sb->s_root; ++ di_read_lock_parent(root, !AuLock_IR); ++ br = au_sbr(sb, bindex); ++ path.mnt = br->br_mnt; ++ path.dentry = au_h_dptr(root, bindex); ++ au_seq_path(seq, &path); ++ di_read_unlock(root, !AuLock_IR); ++ seq_printf(seq, "=%s\n", au_optstr_br_perm(br->br_perm)); ++ return 0; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static struct seq_file *au_seq(char *p, ssize_t len) ++{ ++ struct seq_file *seq; ++ ++ seq = kzalloc(sizeof(*seq), GFP_NOFS); ++ if (seq) { ++ /* mutex_init(&seq.lock); */ ++ seq->buf = p; ++ seq->size = len; ++ return seq; /* success */ ++ } ++ ++ seq = ERR_PTR(-ENOMEM); ++ return seq; ++} ++ ++#define SysaufsBr_PREFIX "br" ++ ++/* todo: file size may exceed PAGE_SIZE */ ++ssize_t sysaufs_si_show(struct kobject *kobj, struct attribute *attr, ++ char *buf) ++{ ++ ssize_t err; ++ long l; ++ aufs_bindex_t bend; ++ struct au_sbinfo *sbinfo; ++ struct super_block *sb; ++ struct seq_file *seq; ++ char *name; ++ struct attribute **cattr; ++ ++ sbinfo = container_of(kobj, struct au_sbinfo, si_kobj); ++ sb = sbinfo->si_sb; ++ si_noflush_read_lock(sb); ++ ++ seq = au_seq(buf, PAGE_SIZE); ++ err = PTR_ERR(seq); ++ if (IS_ERR(seq)) ++ goto out; ++ ++ name = (void *)attr->name; ++ cattr = sysaufs_si_attrs; ++ while (*cattr) { ++ if (!strcmp(name, (*cattr)->name)) { ++ err = container_of(*cattr, struct sysaufs_si_attr, attr) ++ ->show(seq, sb); ++ goto out_seq; ++ } ++ cattr++; ++ } ++ ++ bend = au_sbend(sb); ++ if (!strncmp(name, SysaufsBr_PREFIX, sizeof(SysaufsBr_PREFIX) - 1)) { ++ name += sizeof(SysaufsBr_PREFIX) - 1; ++ err = strict_strtol(name, 10, &l); ++ if (!err) { ++ if (l <= bend) ++ err = sysaufs_si_br(seq, sb, (aufs_bindex_t)l); ++ else ++ err = -ENOENT; ++ } ++ goto out_seq; ++ } ++ BUG(); ++ ++ out_seq: ++ if (!err) { ++ err = seq->count; ++ /* sysfs limit */ ++ if (unlikely(err == PAGE_SIZE)) ++ err = -EFBIG; ++ } ++ kfree(seq); ++ out: ++ si_read_unlock(sb); ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++void sysaufs_br_init(struct au_branch *br) ++{ ++ br->br_attr.name = br->br_name; ++ br->br_attr.mode = S_IRUGO; ++ br->br_attr.owner = THIS_MODULE; ++} ++ ++void sysaufs_brs_del(struct super_block *sb, aufs_bindex_t bindex) ++{ ++ struct au_branch *br; ++ struct kobject *kobj; ++ aufs_bindex_t bend; ++ ++ dbgaufs_brs_del(sb, bindex); ++ ++ if (!sysaufs_brs) ++ return; ++ ++ kobj = &au_sbi(sb)->si_kobj; ++ bend = au_sbend(sb); ++ for (; bindex <= bend; bindex++) { ++ br = au_sbr(sb, bindex); ++ sysfs_remove_file(kobj, &br->br_attr); ++ } ++} ++ ++void sysaufs_brs_add(struct super_block *sb, aufs_bindex_t bindex) ++{ ++ int err; ++ aufs_bindex_t bend; ++ struct kobject *kobj; ++ struct au_branch *br; ++ ++ dbgaufs_brs_add(sb, bindex); ++ ++ if (!sysaufs_brs) ++ return; ++ ++ kobj = &au_sbi(sb)->si_kobj; ++ bend = au_sbend(sb); ++ for (; bindex <= bend; bindex++) { ++ br = au_sbr(sb, bindex); ++ snprintf(br->br_name, sizeof(br->br_name), SysaufsBr_PREFIX ++ "%d", bindex); ++ err = sysfs_create_file(kobj, &br->br_attr); ++ if (unlikely(err)) ++ AuWarn("failed %s under sysfs(%d)\n", br->br_name, err); ++ } ++} +diff -Nur linux-2.6.31.4.orig/fs/aufs/sysrq.c linux-2.6.31.4/fs/aufs/sysrq.c +--- linux-2.6.31.4.orig/fs/aufs/sysrq.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/sysrq.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,115 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * magic sysrq hanlder ++ */ ++ ++#include <linux/fs.h> ++#include <linux/module.h> ++#include <linux/moduleparam.h> ++/* #include <linux/sysrq.h> */ ++#include "aufs.h" ++ ++/* ---------------------------------------------------------------------- */ ++ ++static void sysrq_sb(struct super_block *sb) ++{ ++ char *plevel; ++ struct au_sbinfo *sbinfo; ++ struct file *file; ++ ++ plevel = au_plevel; ++ au_plevel = KERN_WARNING; ++ au_debug(1); ++ ++ sbinfo = au_sbi(sb); ++ pr_warning("si=%lx\n", sysaufs_si_id(sbinfo)); ++ pr_warning(AUFS_NAME ": superblock\n"); ++ au_dpri_sb(sb); ++ pr_warning(AUFS_NAME ": root dentry\n"); ++ au_dpri_dentry(sb->s_root); ++ pr_warning(AUFS_NAME ": root inode\n"); ++ au_dpri_inode(sb->s_root->d_inode); ++#if 0 ++ struct inode *i; ++ pr_warning(AUFS_NAME ": isolated inode\n"); ++ list_for_each_entry(i, &sb->s_inodes, i_sb_list) ++ if (list_empty(&i->i_dentry)) ++ au_dpri_inode(i); ++#endif ++ pr_warning(AUFS_NAME ": files\n"); ++ list_for_each_entry(file, &sb->s_files, f_u.fu_list) ++ if (!special_file(file->f_dentry->d_inode->i_mode)) ++ au_dpri_file(file); ++ ++ au_plevel = plevel; ++ au_debug(0); ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* module parameter */ ++static char *aufs_sysrq_key = "a"; ++module_param_named(sysrq, aufs_sysrq_key, charp, S_IRUGO); ++MODULE_PARM_DESC(sysrq, "MagicSysRq key for " AUFS_NAME); ++ ++static void au_sysrq(int key __maybe_unused, ++ struct tty_struct *tty __maybe_unused) ++{ ++ struct kobject *kobj; ++ struct au_sbinfo *sbinfo; ++ ++ /* spin_lock(&sysaufs_ket->list_lock); */ ++ list_for_each_entry(kobj, &sysaufs_ket->list, entry) { ++ sbinfo = container_of(kobj, struct au_sbinfo, si_kobj); ++ sysrq_sb(sbinfo->si_sb); ++ } ++ /* spin_unlock(&sysaufs_ket->list_lock); */ ++} ++ ++static struct sysrq_key_op au_sysrq_op = { ++ .handler = au_sysrq, ++ .help_msg = "Aufs", ++ .action_msg = "Aufs", ++ .enable_mask = SYSRQ_ENABLE_DUMP ++}; ++ ++/* ---------------------------------------------------------------------- */ ++ ++int __init au_sysrq_init(void) ++{ ++ int err; ++ char key; ++ ++ err = -1; ++ key = *aufs_sysrq_key; ++ if ('a' <= key && key <= 'z') ++ err = register_sysrq_key(key, &au_sysrq_op); ++ if (unlikely(err)) ++ AuErr("err %d, sysrq=%c\n", err, key); ++ return err; ++} ++ ++void au_sysrq_fin(void) ++{ ++ int err; ++ err = unregister_sysrq_key(*aufs_sysrq_key, &au_sysrq_op); ++ if (unlikely(err)) ++ AuErr("err %d (ignored)\n", err); ++} +diff -Nur linux-2.6.31.4.orig/fs/aufs/vdir.c linux-2.6.31.4/fs/aufs/vdir.c +--- linux-2.6.31.4.orig/fs/aufs/vdir.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/vdir.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,882 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * virtual or vertical directory ++ */ ++ ++#include <linux/hash.h> ++#include "aufs.h" ++ ++static unsigned int calc_size(int nlen) ++{ ++ BUILD_BUG_ON(sizeof(ino_t) != sizeof(long)); ++ return ALIGN(sizeof(struct au_vdir_de) + nlen, sizeof(ino_t)); ++} ++ ++static int set_deblk_end(union au_vdir_deblk_p *p, ++ union au_vdir_deblk_p *deblk_end) ++{ ++ if (calc_size(0) <= deblk_end->deblk - p->deblk) { ++ p->de->de_str.len = 0; ++ /* smp_mb(); */ ++ return 0; ++ } ++ return -1; /* error */ ++} ++ ++/* returns true or false */ ++static int is_deblk_end(union au_vdir_deblk_p *p, ++ union au_vdir_deblk_p *deblk_end) ++{ ++ if (calc_size(0) <= deblk_end->deblk - p->deblk) ++ return !p->de->de_str.len; ++ return 1; ++} ++ ++static unsigned char *last_deblk(struct au_vdir *vdir) ++{ ++ return vdir->vd_deblk[vdir->vd_nblk - 1]; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* ++ * the allocated memory has to be freed by ++ * au_nhash_wh_free() or au_nhash_de_free(). ++ */ ++int au_nhash_alloc(struct au_nhash *nhash, unsigned int num_hash, gfp_t gfp) ++{ ++ struct hlist_head *head; ++ unsigned int u; ++ ++ head = kmalloc(sizeof(*nhash->nh_head) * num_hash, gfp); ++ if (head) { ++ nhash->nh_num = num_hash; ++ nhash->nh_head = head; ++ for (u = 0; u < num_hash; u++) ++ INIT_HLIST_HEAD(head++); ++ return 0; /* success */ ++ } ++ ++ return -ENOMEM; ++} ++ ++static void nhash_count(struct hlist_head *head) ++{ ++#if 0 ++ unsigned long n; ++ struct hlist_node *pos; ++ ++ n = 0; ++ hlist_for_each(pos, head) ++ n++; ++ AuInfo("%lu\n", n); ++#endif ++} ++ ++static void au_nhash_wh_do_free(struct hlist_head *head) ++{ ++ struct au_vdir_wh *tpos; ++ struct hlist_node *pos, *node; ++ ++ hlist_for_each_entry_safe(tpos, pos, node, head, wh_hash) { ++ /* hlist_del(pos); */ ++ kfree(tpos); ++ } ++} ++ ++static void au_nhash_de_do_free(struct hlist_head *head) ++{ ++ struct au_vdir_dehstr *tpos; ++ struct hlist_node *pos, *node; ++ ++ hlist_for_each_entry_safe(tpos, pos, node, head, hash) { ++ /* hlist_del(pos); */ ++ au_cache_free_dehstr(tpos); ++ } ++} ++ ++static void au_nhash_do_free(struct au_nhash *nhash, ++ void (*free)(struct hlist_head *head)) ++{ ++ unsigned int u, n; ++ struct hlist_head *head; ++ ++ n = nhash->nh_num; ++ head = nhash->nh_head; ++ for (u = 0; u < n; u++) { ++ nhash_count(head); ++ free(head++); ++ } ++ kfree(nhash->nh_head); ++} ++ ++void au_nhash_wh_free(struct au_nhash *whlist) ++{ ++ au_nhash_do_free(whlist, au_nhash_wh_do_free); ++} ++ ++static void au_nhash_de_free(struct au_nhash *delist) ++{ ++ au_nhash_do_free(delist, au_nhash_de_do_free); ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++int au_nhash_test_longer_wh(struct au_nhash *whlist, aufs_bindex_t btgt, ++ int limit) ++{ ++ int num; ++ unsigned int u, n; ++ struct hlist_head *head; ++ struct au_vdir_wh *tpos; ++ struct hlist_node *pos; ++ ++ num = 0; ++ n = whlist->nh_num; ++ head = whlist->nh_head; ++ for (u = 0; u < n; u++) { ++ hlist_for_each_entry(tpos, pos, head, wh_hash) ++ if (tpos->wh_bindex == btgt && ++num > limit) ++ return 1; ++ head++; ++ } ++ return 0; ++} ++ ++static struct hlist_head *au_name_hash(struct au_nhash *nhash, ++ unsigned char *name, ++ unsigned int len) ++{ ++ unsigned int v; ++ /* const unsigned int magic_bit = 12; */ ++ ++ v = 0; ++ while (len--) ++ v += *name++; ++ /* v = hash_long(v, magic_bit); */ ++ v %= nhash->nh_num; ++ return nhash->nh_head + v; ++} ++ ++static int au_nhash_test_name(struct au_vdir_destr *str, const char *name, ++ int nlen) ++{ ++ return str->len == nlen && !memcmp(str->name, name, nlen); ++} ++ ++/* returns found or not */ ++int au_nhash_test_known_wh(struct au_nhash *whlist, char *name, int nlen) ++{ ++ struct hlist_head *head; ++ struct au_vdir_wh *tpos; ++ struct hlist_node *pos; ++ struct au_vdir_destr *str; ++ ++ head = au_name_hash(whlist, name, nlen); ++ hlist_for_each_entry(tpos, pos, head, wh_hash) { ++ str = &tpos->wh_str; ++ AuDbg("%.*s\n", str->len, str->name); ++ if (au_nhash_test_name(str, name, nlen)) ++ return 1; ++ } ++ return 0; ++} ++ ++/* returns found(true) or not */ ++static int test_known(struct au_nhash *delist, char *name, int nlen) ++{ ++ struct hlist_head *head; ++ struct au_vdir_dehstr *tpos; ++ struct hlist_node *pos; ++ struct au_vdir_destr *str; ++ ++ head = au_name_hash(delist, name, nlen); ++ hlist_for_each_entry(tpos, pos, head, hash) { ++ str = tpos->str; ++ AuDbg("%.*s\n", str->len, str->name); ++ if (au_nhash_test_name(str, name, nlen)) ++ return 1; ++ } ++ return 0; ++} ++ ++static void au_shwh_init_wh(struct au_vdir_wh *wh, ino_t ino, ++ unsigned char d_type) ++{ ++#ifdef CONFIG_AUFS_SHWH ++ wh->wh_ino = ino; ++ wh->wh_type = d_type; ++#endif ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++int au_nhash_append_wh(struct au_nhash *whlist, char *name, int nlen, ino_t ino, ++ unsigned int d_type, aufs_bindex_t bindex, ++ unsigned char shwh) ++{ ++ int err; ++ struct au_vdir_destr *str; ++ struct au_vdir_wh *wh; ++ ++ AuDbg("%.*s\n", nlen, name); ++ err = -ENOMEM; ++ wh = kmalloc(sizeof(*wh) + nlen, GFP_NOFS); ++ if (unlikely(!wh)) ++ goto out; ++ ++ err = 0; ++ wh->wh_bindex = bindex; ++ if (shwh) ++ au_shwh_init_wh(wh, ino, d_type); ++ str = &wh->wh_str; ++ str->len = nlen; ++ memcpy(str->name, name, nlen); ++ hlist_add_head(&wh->wh_hash, au_name_hash(whlist, name, nlen)); ++ /* smp_mb(); */ ++ ++ out: ++ return err; ++} ++ ++static int append_deblk(struct au_vdir *vdir) ++{ ++ int err; ++ unsigned long ul; ++ const unsigned int deblk_sz = vdir->vd_deblk_sz; ++ union au_vdir_deblk_p p, deblk_end; ++ unsigned char **o; ++ ++ err = -ENOMEM; ++ o = krealloc(vdir->vd_deblk, sizeof(*o) * (vdir->vd_nblk + 1), ++ GFP_NOFS); ++ if (unlikely(!o)) ++ goto out; ++ ++ vdir->vd_deblk = o; ++ p.deblk = kmalloc(deblk_sz, GFP_NOFS); ++ if (p.deblk) { ++ ul = vdir->vd_nblk++; ++ vdir->vd_deblk[ul] = p.deblk; ++ vdir->vd_last.ul = ul; ++ vdir->vd_last.p.deblk = p.deblk; ++ deblk_end.deblk = p.deblk + deblk_sz; ++ err = set_deblk_end(&p, &deblk_end); ++ } ++ ++ out: ++ return err; ++} ++ ++static int append_de(struct au_vdir *vdir, char *name, int nlen, ino_t ino, ++ unsigned int d_type, struct au_nhash *delist) ++{ ++ int err; ++ unsigned int sz; ++ const unsigned int deblk_sz = vdir->vd_deblk_sz; ++ union au_vdir_deblk_p p, *room, deblk_end; ++ struct au_vdir_dehstr *dehstr; ++ ++ p.deblk = last_deblk(vdir); ++ deblk_end.deblk = p.deblk + deblk_sz; ++ room = &vdir->vd_last.p; ++ AuDebugOn(room->deblk < p.deblk || deblk_end.deblk <= room->deblk ++ || !is_deblk_end(room, &deblk_end)); ++ ++ sz = calc_size(nlen); ++ if (unlikely(sz > deblk_end.deblk - room->deblk)) { ++ err = append_deblk(vdir); ++ if (unlikely(err)) ++ goto out; ++ ++ p.deblk = last_deblk(vdir); ++ deblk_end.deblk = p.deblk + deblk_sz; ++ /* smp_mb(); */ ++ AuDebugOn(room->deblk != p.deblk); ++ } ++ ++ err = -ENOMEM; ++ dehstr = au_cache_alloc_dehstr(); ++ if (unlikely(!dehstr)) ++ goto out; ++ ++ dehstr->str = &room->de->de_str; ++ hlist_add_head(&dehstr->hash, au_name_hash(delist, name, nlen)); ++ room->de->de_ino = ino; ++ room->de->de_type = d_type; ++ room->de->de_str.len = nlen; ++ memcpy(room->de->de_str.name, name, nlen); ++ ++ err = 0; ++ room->deblk += sz; ++ if (unlikely(set_deblk_end(room, &deblk_end))) ++ err = append_deblk(vdir); ++ /* smp_mb(); */ ++ ++ out: ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++void au_vdir_free(struct au_vdir *vdir) ++{ ++ unsigned char **deblk; ++ ++ deblk = vdir->vd_deblk; ++ while (vdir->vd_nblk--) ++ kfree(*deblk++); ++ kfree(vdir->vd_deblk); ++ au_cache_free_vdir(vdir); ++} ++ ++static struct au_vdir *alloc_vdir(struct super_block *sb) ++{ ++ struct au_vdir *vdir; ++ int err; ++ ++ SiMustAnyLock(sb); ++ ++ err = -ENOMEM; ++ vdir = au_cache_alloc_vdir(); ++ if (unlikely(!vdir)) ++ goto out; ++ ++ vdir->vd_deblk = kzalloc(sizeof(*vdir->vd_deblk), GFP_NOFS); ++ if (unlikely(!vdir->vd_deblk)) ++ goto out_free; ++ ++ vdir->vd_deblk_sz = au_sbi(sb)->si_rdblk; ++ vdir->vd_nblk = 0; ++ vdir->vd_version = 0; ++ vdir->vd_jiffy = 0; ++ err = append_deblk(vdir); ++ if (!err) ++ return vdir; /* success */ ++ ++ kfree(vdir->vd_deblk); ++ ++ out_free: ++ au_cache_free_vdir(vdir); ++ out: ++ vdir = ERR_PTR(err); ++ return vdir; ++} ++ ++static int reinit_vdir(struct au_vdir *vdir) ++{ ++ int err; ++ union au_vdir_deblk_p p, deblk_end; ++ ++ while (vdir->vd_nblk > 1) { ++ kfree(vdir->vd_deblk[vdir->vd_nblk - 1]); ++ /* vdir->vd_deblk[vdir->vd_nblk - 1] = NULL; */ ++ vdir->vd_nblk--; ++ } ++ p.deblk = vdir->vd_deblk[0]; ++ deblk_end.deblk = p.deblk + vdir->vd_deblk_sz; ++ err = set_deblk_end(&p, &deblk_end); ++ /* keep vd_dblk_sz */ ++ vdir->vd_last.ul = 0; ++ vdir->vd_last.p.deblk = vdir->vd_deblk[0]; ++ vdir->vd_version = 0; ++ vdir->vd_jiffy = 0; ++ /* smp_mb(); */ ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static int au_ino(struct super_block *sb, aufs_bindex_t bindex, ino_t h_ino, ++ unsigned int d_type, ino_t *ino) ++{ ++ int err; ++ struct mutex *mtx; ++ const int isdir = (d_type == DT_DIR); ++ ++ /* prevent hardlinks from race condition */ ++ mtx = NULL; ++ if (!isdir) { ++ mtx = &au_sbr(sb, bindex)->br_xino.xi_nondir_mtx; ++ mutex_lock(mtx); ++ } ++ err = au_xino_read(sb, bindex, h_ino, ino); ++ if (unlikely(err)) ++ goto out; ++ ++ if (!*ino) { ++ err = -EIO; ++ *ino = au_xino_new_ino(sb); ++ if (unlikely(!*ino)) ++ goto out; ++ err = au_xino_write(sb, bindex, h_ino, *ino); ++ if (unlikely(err)) ++ goto out; ++ } ++ ++ out: ++ if (!isdir) ++ mutex_unlock(mtx); ++ return err; ++} ++ ++static int au_wh_ino(struct super_block *sb, aufs_bindex_t bindex, ino_t h_ino, ++ unsigned int d_type, ino_t *ino) ++{ ++#ifdef CONFIG_AUFS_SHWH ++ return au_ino(sb, bindex, h_ino, d_type, ino); ++#else ++ return 0; ++#endif ++} ++ ++#define AuFillVdir_CALLED 1 ++#define AuFillVdir_WHABLE (1 << 1) ++#define AuFillVdir_SHWH (1 << 2) ++#define au_ftest_fillvdir(flags, name) ((flags) & AuFillVdir_##name) ++#define au_fset_fillvdir(flags, name) { (flags) |= AuFillVdir_##name; } ++#define au_fclr_fillvdir(flags, name) { (flags) &= ~AuFillVdir_##name; } ++ ++#ifndef CONFIG_AUFS_SHWH ++#undef AuFillVdir_SHWH ++#define AuFillVdir_SHWH 0 ++#endif ++ ++struct fillvdir_arg { ++ struct file *file; ++ struct au_vdir *vdir; ++ struct au_nhash delist; ++ struct au_nhash whlist; ++ aufs_bindex_t bindex; ++ unsigned int flags; ++ int err; ++}; ++ ++static int fillvdir(void *__arg, const char *__name, int nlen, ++ loff_t offset __maybe_unused, u64 h_ino, ++ unsigned int d_type) ++{ ++ struct fillvdir_arg *arg = __arg; ++ char *name = (void *)__name; ++ struct super_block *sb; ++ ino_t ino; ++ const unsigned char shwh = !!au_ftest_fillvdir(arg->flags, SHWH); ++ ++ arg->err = 0; ++ sb = arg->file->f_dentry->d_sb; ++ au_fset_fillvdir(arg->flags, CALLED); ++ /* smp_mb(); */ ++ if (nlen <= AUFS_WH_PFX_LEN ++ || memcmp(name, AUFS_WH_PFX, AUFS_WH_PFX_LEN)) { ++ if (test_known(&arg->delist, name, nlen) ++ || au_nhash_test_known_wh(&arg->whlist, name, nlen)) ++ goto out; /* already exists or whiteouted */ ++ ++ sb = arg->file->f_dentry->d_sb; ++ arg->err = au_ino(sb, arg->bindex, h_ino, d_type, &ino); ++ if (!arg->err) ++ arg->err = append_de(arg->vdir, name, nlen, ino, ++ d_type, &arg->delist); ++ } else if (au_ftest_fillvdir(arg->flags, WHABLE)) { ++ name += AUFS_WH_PFX_LEN; ++ nlen -= AUFS_WH_PFX_LEN; ++ if (au_nhash_test_known_wh(&arg->whlist, name, nlen)) ++ goto out; /* already whiteouted */ ++ ++ if (shwh) ++ arg->err = au_wh_ino(sb, arg->bindex, h_ino, d_type, ++ &ino); ++ if (!arg->err) ++ arg->err = au_nhash_append_wh ++ (&arg->whlist, name, nlen, ino, d_type, ++ arg->bindex, shwh); ++ } ++ ++ out: ++ if (!arg->err) ++ arg->vdir->vd_jiffy = jiffies; ++ /* smp_mb(); */ ++ AuTraceErr(arg->err); ++ return arg->err; ++} ++ ++static int au_handle_shwh(struct super_block *sb, struct au_vdir *vdir, ++ struct au_nhash *whlist, struct au_nhash *delist) ++{ ++#ifdef CONFIG_AUFS_SHWH ++ int err; ++ unsigned int nh, u; ++ struct hlist_head *head; ++ struct au_vdir_wh *tpos; ++ struct hlist_node *pos, *n; ++ char *p, *o; ++ struct au_vdir_destr *destr; ++ ++ AuDebugOn(!au_opt_test(au_mntflags(sb), SHWH)); ++ ++ err = -ENOMEM; ++ o = p = __getname(); ++ if (unlikely(!p)) ++ goto out; ++ ++ err = 0; ++ nh = whlist->nh_num; ++ memcpy(p, AUFS_WH_PFX, AUFS_WH_PFX_LEN); ++ p += AUFS_WH_PFX_LEN; ++ for (u = 0; u < nh; u++) { ++ head = whlist->nh_head + u; ++ hlist_for_each_entry_safe(tpos, pos, n, head, wh_hash) { ++ destr = &tpos->wh_str; ++ memcpy(p, destr->name, destr->len); ++ err = append_de(vdir, o, destr->len + AUFS_WH_PFX_LEN, ++ tpos->wh_ino, tpos->wh_type, delist); ++ if (unlikely(err)) ++ break; ++ } ++ } ++ ++ __putname(o); ++ ++ out: ++ AuTraceErr(err); ++ return err; ++#else ++ return 0; ++#endif ++} ++ ++static int au_do_read_vdir(struct fillvdir_arg *arg) ++{ ++ int err; ++ unsigned int rdhash; ++ loff_t offset; ++ aufs_bindex_t bend, bindex, bstart; ++ unsigned char shwh; ++ struct file *hf, *file; ++ struct super_block *sb; ++ ++ file = arg->file; ++ sb = file->f_dentry->d_sb; ++ SiMustAnyLock(sb); ++ ++ rdhash = au_sbi(sb)->si_rdhash; ++ err = au_nhash_alloc(&arg->delist, rdhash, GFP_NOFS); ++ if (unlikely(err)) ++ goto out; ++ err = au_nhash_alloc(&arg->whlist, rdhash, GFP_NOFS); ++ if (unlikely(err)) ++ goto out_delist; ++ ++ err = 0; ++ arg->flags = 0; ++ shwh = 0; ++ if (au_opt_test(au_mntflags(sb), SHWH)) { ++ shwh = 1; ++ au_fset_fillvdir(arg->flags, SHWH); ++ } ++ bstart = au_fbstart(file); ++ bend = au_fbend(file); ++ for (bindex = bstart; !err && bindex <= bend; bindex++) { ++ hf = au_h_fptr(file, bindex); ++ if (!hf) ++ continue; ++ ++ offset = vfsub_llseek(hf, 0, SEEK_SET); ++ err = offset; ++ if (unlikely(offset)) ++ break; ++ ++ arg->bindex = bindex; ++ au_fclr_fillvdir(arg->flags, WHABLE); ++ if (shwh ++ || (bindex != bend ++ && au_br_whable(au_sbr_perm(sb, bindex)))) ++ au_fset_fillvdir(arg->flags, WHABLE); ++ do { ++ arg->err = 0; ++ au_fclr_fillvdir(arg->flags, CALLED); ++ /* smp_mb(); */ ++ err = vfsub_readdir(hf, fillvdir, arg); ++ if (err >= 0) ++ err = arg->err; ++ } while (!err && au_ftest_fillvdir(arg->flags, CALLED)); ++ } ++ ++ if (!err && shwh) ++ err = au_handle_shwh(sb, arg->vdir, &arg->whlist, &arg->delist); ++ ++ au_nhash_wh_free(&arg->whlist); ++ ++ out_delist: ++ au_nhash_de_free(&arg->delist); ++ out: ++ return err; ++} ++ ++static int read_vdir(struct file *file, int may_read) ++{ ++ int err; ++ unsigned long expire; ++ unsigned char do_read; ++ struct fillvdir_arg arg; ++ struct inode *inode; ++ struct au_vdir *vdir, *allocated; ++ ++ err = 0; ++ inode = file->f_dentry->d_inode; ++ IMustLock(inode); ++ SiMustAnyLock(inode->i_sb); ++ ++ allocated = NULL; ++ do_read = 0; ++ expire = au_sbi(inode->i_sb)->si_rdcache; ++ vdir = au_ivdir(inode); ++ if (!vdir) { ++ do_read = 1; ++ vdir = alloc_vdir(inode->i_sb); ++ err = PTR_ERR(vdir); ++ if (IS_ERR(vdir)) ++ goto out; ++ err = 0; ++ allocated = vdir; ++ } else if (may_read ++ && (inode->i_version != vdir->vd_version ++ || time_after(jiffies, vdir->vd_jiffy + expire))) { ++ do_read = 1; ++ err = reinit_vdir(vdir); ++ if (unlikely(err)) ++ goto out; ++ } ++ ++ if (!do_read) ++ return 0; /* success */ ++ ++ arg.file = file; ++ arg.vdir = vdir; ++ err = au_do_read_vdir(&arg); ++ if (!err) { ++ /* file->f_pos = 0; */ ++ vdir->vd_version = inode->i_version; ++ vdir->vd_last.ul = 0; ++ vdir->vd_last.p.deblk = vdir->vd_deblk[0]; ++ if (allocated) ++ au_set_ivdir(inode, allocated); ++ } else if (allocated) ++ au_vdir_free(allocated); ++ ++ out: ++ return err; ++} ++ ++static int copy_vdir(struct au_vdir *tgt, struct au_vdir *src) ++{ ++ int err, rerr; ++ unsigned long ul, n; ++ const unsigned int deblk_sz = src->vd_deblk_sz; ++ ++ AuDebugOn(tgt->vd_nblk != 1); ++ ++ err = -ENOMEM; ++ if (tgt->vd_nblk < src->vd_nblk) { ++ unsigned char **p; ++ ++ p = krealloc(tgt->vd_deblk, sizeof(*p) * src->vd_nblk, ++ GFP_NOFS); ++ if (unlikely(!p)) ++ goto out; ++ tgt->vd_deblk = p; ++ } ++ ++ tgt->vd_nblk = src->vd_nblk; ++ tgt->vd_deblk_sz = deblk_sz; ++ memcpy(tgt->vd_deblk[0], src->vd_deblk[0], deblk_sz); ++ /* tgt->vd_last.i = 0; */ ++ /* tgt->vd_last.p.deblk = tgt->vd_deblk[0]; */ ++ tgt->vd_version = src->vd_version; ++ tgt->vd_jiffy = src->vd_jiffy; ++ ++ n = src->vd_nblk; ++ for (ul = 1; ul < n; ul++) { ++ tgt->vd_deblk[ul] = kmemdup(src->vd_deblk[ul], deblk_sz, ++ GFP_NOFS); ++ if (unlikely(!tgt->vd_deblk[ul])) ++ goto out; ++ } ++ /* smp_mb(); */ ++ return 0; /* success */ ++ ++ out: ++ rerr = reinit_vdir(tgt); ++ BUG_ON(rerr); ++ return err; ++} ++ ++int au_vdir_init(struct file *file) ++{ ++ int err; ++ struct inode *inode; ++ struct au_vdir *vdir_cache, *allocated; ++ ++ err = read_vdir(file, !file->f_pos); ++ if (unlikely(err)) ++ goto out; ++ ++ allocated = NULL; ++ vdir_cache = au_fvdir_cache(file); ++ if (!vdir_cache) { ++ vdir_cache = alloc_vdir(file->f_dentry->d_sb); ++ err = PTR_ERR(vdir_cache); ++ if (IS_ERR(vdir_cache)) ++ goto out; ++ allocated = vdir_cache; ++ } else if (!file->f_pos && vdir_cache->vd_version != file->f_version) { ++ err = reinit_vdir(vdir_cache); ++ if (unlikely(err)) ++ goto out; ++ } else ++ return 0; /* success */ ++ ++ inode = file->f_dentry->d_inode; ++ err = copy_vdir(vdir_cache, au_ivdir(inode)); ++ if (!err) { ++ file->f_version = inode->i_version; ++ if (allocated) ++ au_set_fvdir_cache(file, allocated); ++ } else if (allocated) ++ au_vdir_free(allocated); ++ ++ out: ++ return err; ++} ++ ++static loff_t calc_offset(struct au_vdir *vdir) ++{ ++ loff_t offset; ++ union au_vdir_deblk_p p; ++ ++ p.deblk = vdir->vd_deblk[vdir->vd_last.ul]; ++ offset = vdir->vd_last.p.deblk - p.deblk; ++ offset += vdir->vd_deblk_sz * vdir->vd_last.ul; ++ return offset; ++} ++ ++/* returns true or false */ ++static int seek_vdir(struct file *file) ++{ ++ int valid; ++ unsigned int deblk_sz; ++ unsigned long ul, n; ++ loff_t offset; ++ union au_vdir_deblk_p p, deblk_end; ++ struct au_vdir *vdir_cache; ++ ++ valid = 1; ++ vdir_cache = au_fvdir_cache(file); ++ offset = calc_offset(vdir_cache); ++ AuDbg("offset %lld\n", offset); ++ if (file->f_pos == offset) ++ goto out; ++ ++ vdir_cache->vd_last.ul = 0; ++ vdir_cache->vd_last.p.deblk = vdir_cache->vd_deblk[0]; ++ if (!file->f_pos) ++ goto out; ++ ++ valid = 0; ++ deblk_sz = vdir_cache->vd_deblk_sz; ++ ul = div64_u64(file->f_pos, deblk_sz); ++ AuDbg("ul %lu\n", ul); ++ if (ul >= vdir_cache->vd_nblk) ++ goto out; ++ ++ n = vdir_cache->vd_nblk; ++ for (; ul < n; ul++) { ++ p.deblk = vdir_cache->vd_deblk[ul]; ++ deblk_end.deblk = p.deblk + deblk_sz; ++ offset = ul; ++ offset *= deblk_sz; ++ while (!is_deblk_end(&p, &deblk_end) && offset < file->f_pos) { ++ unsigned int l; ++ ++ l = calc_size(p.de->de_str.len); ++ offset += l; ++ p.deblk += l; ++ } ++ if (!is_deblk_end(&p, &deblk_end)) { ++ valid = 1; ++ vdir_cache->vd_last.ul = ul; ++ vdir_cache->vd_last.p = p; ++ break; ++ } ++ } ++ ++ out: ++ /* smp_mb(); */ ++ AuTraceErr(!valid); ++ return valid; ++} ++ ++int au_vdir_fill_de(struct file *file, void *dirent, filldir_t filldir) ++{ ++ int err; ++ unsigned int l, deblk_sz; ++ union au_vdir_deblk_p deblk_end; ++ struct au_vdir *vdir_cache; ++ struct au_vdir_de *de; ++ ++ vdir_cache = au_fvdir_cache(file); ++ if (!seek_vdir(file)) ++ return 0; ++ ++ deblk_sz = vdir_cache->vd_deblk_sz; ++ while (1) { ++ deblk_end.deblk = vdir_cache->vd_deblk[vdir_cache->vd_last.ul]; ++ deblk_end.deblk += deblk_sz; ++ while (!is_deblk_end(&vdir_cache->vd_last.p, &deblk_end)) { ++ de = vdir_cache->vd_last.p.de; ++ AuDbg("%.*s, off%lld, i%lu, dt%d\n", ++ de->de_str.len, de->de_str.name, file->f_pos, ++ (unsigned long)de->de_ino, de->de_type); ++ err = filldir(dirent, de->de_str.name, de->de_str.len, ++ file->f_pos, de->de_ino, de->de_type); ++ if (unlikely(err)) { ++ AuTraceErr(err); ++ /* todo: ignore the error caused by udba? */ ++ /* return err; */ ++ return 0; ++ } ++ ++ l = calc_size(de->de_str.len); ++ vdir_cache->vd_last.p.deblk += l; ++ file->f_pos += l; ++ } ++ if (vdir_cache->vd_last.ul < vdir_cache->vd_nblk - 1) { ++ vdir_cache->vd_last.ul++; ++ vdir_cache->vd_last.p.deblk ++ = vdir_cache->vd_deblk[vdir_cache->vd_last.ul]; ++ file->f_pos = deblk_sz * vdir_cache->vd_last.ul; ++ continue; ++ } ++ break; ++ } ++ ++ /* smp_mb(); */ ++ return 0; ++} +diff -Nur linux-2.6.31.4.orig/fs/aufs/vfsub.c linux-2.6.31.4/fs/aufs/vfsub.c +--- linux-2.6.31.4.orig/fs/aufs/vfsub.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/vfsub.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,740 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * sub-routines for VFS ++ */ ++ ++#include <linux/namei.h> ++#include <linux/security.h> ++#include <linux/splice.h> ++#include <linux/uaccess.h> ++#include "aufs.h" ++ ++int vfsub_update_h_iattr(struct path *h_path, int *did) ++{ ++ int err; ++ struct kstat st; ++ struct super_block *h_sb; ++ ++ /* for remote fs, leave work for its getattr or d_revalidate */ ++ /* for bad i_attr fs, handle them in aufs_getattr() */ ++ /* still some fs may acquire i_mutex. we need to skip them */ ++ err = 0; ++ if (!did) ++ did = &err; ++ h_sb = h_path->dentry->d_sb; ++ *did = (!au_test_fs_remote(h_sb) && au_test_fs_refresh_iattr(h_sb)); ++ if (*did) ++ err = vfs_getattr(h_path->mnt, h_path->dentry, &st); ++ ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++#ifdef CONFIG_IMA ++#error IMA is not supported since it does not work well. Wait for their fixing. ++#endif ++ ++struct file *vfsub_filp_open(const char *path, int oflags, int mode) ++{ ++ struct file *file; ++ ++ lockdep_off(); ++ file = filp_open(path, oflags, mode); ++ lockdep_on(); ++ if (IS_ERR(file)) ++ goto out; ++ vfsub_update_h_iattr(&file->f_path, /*did*/NULL); /*ignore*/ ++ ++ out: ++ return file; ++} ++ ++int vfsub_kern_path(const char *name, unsigned int flags, struct path *path) ++{ ++ int err; ++ ++ /* lockdep_off(); */ ++ err = kern_path(name, flags, path); ++ /* lockdep_on(); */ ++ if (!err && path->dentry->d_inode) ++ vfsub_update_h_iattr(path, /*did*/NULL); /*ignore*/ ++ return err; ++} ++ ++struct dentry *vfsub_lookup_one_len(const char *name, struct dentry *parent, ++ int len) ++{ ++ struct path path = { ++ .mnt = NULL ++ }; ++ ++ IMustLock(parent->d_inode); ++ ++ path.dentry = lookup_one_len(name, parent, len); ++ if (IS_ERR(path.dentry)) ++ goto out; ++ if (path.dentry->d_inode) ++ vfsub_update_h_iattr(&path, /*did*/NULL); /*ignore*/ ++ ++ out: ++ return path.dentry; ++} ++ ++struct dentry *vfsub_lookup_hash(struct nameidata *nd) ++{ ++ struct path path = { ++ .mnt = nd->path.mnt ++ }; ++ ++ IMustLock(nd->path.dentry->d_inode); ++ ++ path.dentry = lookup_hash(nd); ++ if (!IS_ERR(path.dentry) && path.dentry->d_inode) ++ vfsub_update_h_iattr(&path, /*did*/NULL); /*ignore*/ ++ ++ return path.dentry; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++struct dentry *vfsub_lock_rename(struct dentry *d1, struct au_hinode *hdir1, ++ struct dentry *d2, struct au_hinode *hdir2) ++{ ++ struct dentry *d; ++ ++ lockdep_off(); ++ d = lock_rename(d1, d2); ++ lockdep_on(); ++ au_hin_suspend(hdir1); ++ if (hdir1 != hdir2) ++ au_hin_suspend(hdir2); ++ ++ return d; ++} ++ ++void vfsub_unlock_rename(struct dentry *d1, struct au_hinode *hdir1, ++ struct dentry *d2, struct au_hinode *hdir2) ++{ ++ au_hin_resume(hdir1); ++ if (hdir1 != hdir2) ++ au_hin_resume(hdir2); ++ lockdep_off(); ++ unlock_rename(d1, d2); ++ lockdep_on(); ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++int vfsub_create(struct inode *dir, struct path *path, int mode) ++{ ++ int err; ++ struct dentry *d; ++ ++ IMustLock(dir); ++ ++ d = path->dentry; ++ path->dentry = d->d_parent; ++ err = security_path_mknod(path, path->dentry, mode, 0); ++ path->dentry = d; ++ if (unlikely(err)) ++ goto out; ++ ++ if (au_test_fs_null_nd(dir->i_sb)) ++ err = vfs_create(dir, path->dentry, mode, NULL); ++ else { ++ struct nameidata h_nd; ++ ++ memset(&h_nd, 0, sizeof(h_nd)); ++ h_nd.flags = LOOKUP_CREATE; ++ h_nd.intent.open.flags = O_CREAT ++ | vfsub_fmode_to_uint(FMODE_READ); ++ h_nd.intent.open.create_mode = mode; ++ h_nd.path.dentry = path->dentry->d_parent; ++ h_nd.path.mnt = path->mnt; ++ path_get(&h_nd.path); ++ err = vfs_create(dir, path->dentry, mode, &h_nd); ++ path_put(&h_nd.path); ++ } ++ ++ if (!err) { ++ struct path tmp = *path; ++ int did; ++ ++ vfsub_update_h_iattr(&tmp, &did); ++ if (did) { ++ tmp.dentry = path->dentry->d_parent; ++ vfsub_update_h_iattr(&tmp, /*did*/NULL); ++ } ++ /*ignore*/ ++ } ++ ++ out: ++ return err; ++} ++ ++int vfsub_symlink(struct inode *dir, struct path *path, const char *symname) ++{ ++ int err; ++ struct dentry *d; ++ ++ IMustLock(dir); ++ ++ d = path->dentry; ++ path->dentry = d->d_parent; ++ err = security_path_symlink(path, path->dentry, symname); ++ path->dentry = d; ++ if (unlikely(err)) ++ goto out; ++ ++ err = vfs_symlink(dir, path->dentry, symname); ++ if (!err) { ++ struct path tmp = *path; ++ int did; ++ ++ vfsub_update_h_iattr(&tmp, &did); ++ if (did) { ++ tmp.dentry = path->dentry->d_parent; ++ vfsub_update_h_iattr(&tmp, /*did*/NULL); ++ } ++ /*ignore*/ ++ } ++ ++ out: ++ return err; ++} ++ ++int vfsub_mknod(struct inode *dir, struct path *path, int mode, dev_t dev) ++{ ++ int err; ++ struct dentry *d; ++ ++ IMustLock(dir); ++ ++ d = path->dentry; ++ path->dentry = d->d_parent; ++ err = security_path_mknod(path, path->dentry, mode, dev); ++ path->dentry = d; ++ if (unlikely(err)) ++ goto out; ++ ++ err = vfs_mknod(dir, path->dentry, mode, dev); ++ if (!err) { ++ struct path tmp = *path; ++ int did; ++ ++ vfsub_update_h_iattr(&tmp, &did); ++ if (did) { ++ tmp.dentry = path->dentry->d_parent; ++ vfsub_update_h_iattr(&tmp, /*did*/NULL); ++ } ++ /*ignore*/ ++ } ++ ++ out: ++ return err; ++} ++ ++static int au_test_nlink(struct inode *inode) ++{ ++ const unsigned int link_max = UINT_MAX >> 1; /* rough margin */ ++ ++ if (!au_test_fs_no_limit_nlink(inode->i_sb) ++ || inode->i_nlink < link_max) ++ return 0; ++ return -EMLINK; ++} ++ ++int vfsub_link(struct dentry *src_dentry, struct inode *dir, struct path *path) ++{ ++ int err; ++ struct dentry *d; ++ ++ IMustLock(dir); ++ ++ err = au_test_nlink(src_dentry->d_inode); ++ if (unlikely(err)) ++ return err; ++ ++ d = path->dentry; ++ path->dentry = d->d_parent; ++ err = security_path_link(src_dentry, path, path->dentry); ++ path->dentry = d; ++ if (unlikely(err)) ++ goto out; ++ ++ lockdep_off(); ++ err = vfs_link(src_dentry, dir, path->dentry); ++ lockdep_on(); ++ if (!err) { ++ struct path tmp = *path; ++ int did; ++ ++ /* fuse has different memory inode for the same inumber */ ++ vfsub_update_h_iattr(&tmp, &did); ++ if (did) { ++ tmp.dentry = path->dentry->d_parent; ++ vfsub_update_h_iattr(&tmp, /*did*/NULL); ++ tmp.dentry = src_dentry; ++ vfsub_update_h_iattr(&tmp, /*did*/NULL); ++ } ++ /*ignore*/ ++ } ++ ++ out: ++ return err; ++} ++ ++int vfsub_rename(struct inode *src_dir, struct dentry *src_dentry, ++ struct inode *dir, struct path *path) ++{ ++ int err; ++ struct path tmp = { ++ .mnt = path->mnt ++ }; ++ struct dentry *d; ++ ++ IMustLock(dir); ++ IMustLock(src_dir); ++ ++ d = path->dentry; ++ path->dentry = d->d_parent; ++ tmp.dentry = src_dentry->d_parent; ++ err = security_path_rename(&tmp, src_dentry, path, path->dentry); ++ path->dentry = d; ++ if (unlikely(err)) ++ goto out; ++ ++ lockdep_off(); ++ err = vfs_rename(src_dir, src_dentry, dir, path->dentry); ++ lockdep_on(); ++ if (!err) { ++ int did; ++ ++ tmp.dentry = d->d_parent; ++ vfsub_update_h_iattr(&tmp, &did); ++ if (did) { ++ tmp.dentry = src_dentry; ++ vfsub_update_h_iattr(&tmp, /*did*/NULL); ++ tmp.dentry = src_dentry->d_parent; ++ vfsub_update_h_iattr(&tmp, /*did*/NULL); ++ } ++ /*ignore*/ ++ } ++ ++ out: ++ return err; ++} ++ ++int vfsub_mkdir(struct inode *dir, struct path *path, int mode) ++{ ++ int err; ++ struct dentry *d; ++ ++ IMustLock(dir); ++ ++ d = path->dentry; ++ path->dentry = d->d_parent; ++ err = security_path_mkdir(path, path->dentry, mode); ++ path->dentry = d; ++ if (unlikely(err)) ++ goto out; ++ ++ err = vfs_mkdir(dir, path->dentry, mode); ++ if (!err) { ++ struct path tmp = *path; ++ int did; ++ ++ vfsub_update_h_iattr(&tmp, &did); ++ if (did) { ++ tmp.dentry = path->dentry->d_parent; ++ vfsub_update_h_iattr(&tmp, /*did*/NULL); ++ } ++ /*ignore*/ ++ } ++ ++ out: ++ return err; ++} ++ ++int vfsub_rmdir(struct inode *dir, struct path *path) ++{ ++ int err; ++ struct dentry *d; ++ ++ IMustLock(dir); ++ ++ d = path->dentry; ++ path->dentry = d->d_parent; ++ err = security_path_rmdir(path, path->dentry); ++ path->dentry = d; ++ if (unlikely(err)) ++ goto out; ++ ++ lockdep_off(); ++ err = vfs_rmdir(dir, path->dentry); ++ lockdep_on(); ++ if (!err) { ++ struct path tmp = { ++ .dentry = path->dentry->d_parent, ++ .mnt = path->mnt ++ }; ++ ++ vfsub_update_h_iattr(&tmp, /*did*/NULL); /*ignore*/ ++ } ++ ++ out: ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++ssize_t vfsub_read_u(struct file *file, char __user *ubuf, size_t count, ++ loff_t *ppos) ++{ ++ ssize_t err; ++ ++ err = vfs_read(file, ubuf, count, ppos); ++ if (err >= 0) ++ vfsub_update_h_iattr(&file->f_path, /*did*/NULL); /*ignore*/ ++ return err; ++} ++ ++/* todo: kernel_read()? */ ++ssize_t vfsub_read_k(struct file *file, void *kbuf, size_t count, ++ loff_t *ppos) ++{ ++ ssize_t err; ++ mm_segment_t oldfs; ++ ++ oldfs = get_fs(); ++ set_fs(KERNEL_DS); ++ err = vfsub_read_u(file, (char __user *)kbuf, count, ppos); ++ set_fs(oldfs); ++ return err; ++} ++ ++ssize_t vfsub_write_u(struct file *file, const char __user *ubuf, size_t count, ++ loff_t *ppos) ++{ ++ ssize_t err; ++ ++ lockdep_off(); ++ err = vfs_write(file, ubuf, count, ppos); ++ lockdep_on(); ++ if (err >= 0) ++ vfsub_update_h_iattr(&file->f_path, /*did*/NULL); /*ignore*/ ++ return err; ++} ++ ++ssize_t vfsub_write_k(struct file *file, void *kbuf, size_t count, loff_t *ppos) ++{ ++ ssize_t err; ++ mm_segment_t oldfs; ++ ++ oldfs = get_fs(); ++ set_fs(KERNEL_DS); ++ err = vfsub_write_u(file, (const char __user *)kbuf, count, ppos); ++ set_fs(oldfs); ++ return err; ++} ++ ++int vfsub_readdir(struct file *file, filldir_t filldir, void *arg) ++{ ++ int err; ++ ++ lockdep_off(); ++ err = vfs_readdir(file, filldir, arg); ++ lockdep_on(); ++ if (err >= 0) ++ vfsub_update_h_iattr(&file->f_path, /*did*/NULL); /*ignore*/ ++ return err; ++} ++ ++long vfsub_splice_to(struct file *in, loff_t *ppos, ++ struct pipe_inode_info *pipe, size_t len, ++ unsigned int flags) ++{ ++ long err; ++ ++ lockdep_off(); ++ err = do_splice_to(in, ppos, pipe, len, flags); ++ lockdep_on(); ++ if (err >= 0) ++ vfsub_update_h_iattr(&in->f_path, /*did*/NULL); /*ignore*/ ++ return err; ++} ++ ++long vfsub_splice_from(struct pipe_inode_info *pipe, struct file *out, ++ loff_t *ppos, size_t len, unsigned int flags) ++{ ++ long err; ++ ++ lockdep_off(); ++ err = do_splice_from(pipe, out, ppos, len, flags); ++ lockdep_on(); ++ if (err >= 0) ++ vfsub_update_h_iattr(&out->f_path, /*did*/NULL); /*ignore*/ ++ return err; ++} ++ ++/* cf. open.c:do_sys_truncate() and do_sys_ftruncate() */ ++int vfsub_trunc(struct path *h_path, loff_t length, unsigned int attr, ++ struct file *h_file) ++{ ++ int err; ++ struct inode *h_inode; ++ ++ h_inode = h_path->dentry->d_inode; ++ if (!h_file) { ++ err = mnt_want_write(h_path->mnt); ++ if (err) ++ goto out; ++ err = inode_permission(h_inode, MAY_WRITE); ++ if (err) ++ goto out_mnt; ++ err = get_write_access(h_inode); ++ if (err) ++ goto out_mnt; ++ err = break_lease(h_inode, vfsub_fmode_to_uint(FMODE_WRITE)); ++ if (err) ++ goto out_inode; ++ } ++ ++ err = locks_verify_truncate(h_inode, h_file, length); ++ if (!err) ++ err = security_path_truncate(h_path, length, attr); ++ if (!err) { ++ lockdep_off(); ++ err = do_truncate(h_path->dentry, length, attr, h_file); ++ lockdep_on(); ++ } ++ ++ out_inode: ++ if (!h_file) ++ put_write_access(h_inode); ++ out_mnt: ++ if (!h_file) ++ mnt_drop_write(h_path->mnt); ++ out: ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++struct au_vfsub_mkdir_args { ++ int *errp; ++ struct inode *dir; ++ struct path *path; ++ int mode; ++}; ++ ++static void au_call_vfsub_mkdir(void *args) ++{ ++ struct au_vfsub_mkdir_args *a = args; ++ *a->errp = vfsub_mkdir(a->dir, a->path, a->mode); ++} ++ ++int vfsub_sio_mkdir(struct inode *dir, struct path *path, int mode) ++{ ++ int err, do_sio, wkq_err; ++ ++ do_sio = au_test_h_perm_sio(dir, MAY_EXEC | MAY_WRITE); ++ if (!do_sio) ++ err = vfsub_mkdir(dir, path, mode); ++ else { ++ struct au_vfsub_mkdir_args args = { ++ .errp = &err, ++ .dir = dir, ++ .path = path, ++ .mode = mode ++ }; ++ wkq_err = au_wkq_wait(au_call_vfsub_mkdir, &args); ++ if (unlikely(wkq_err)) ++ err = wkq_err; ++ } ++ ++ return err; ++} ++ ++struct au_vfsub_rmdir_args { ++ int *errp; ++ struct inode *dir; ++ struct path *path; ++}; ++ ++static void au_call_vfsub_rmdir(void *args) ++{ ++ struct au_vfsub_rmdir_args *a = args; ++ *a->errp = vfsub_rmdir(a->dir, a->path); ++} ++ ++int vfsub_sio_rmdir(struct inode *dir, struct path *path) ++{ ++ int err, do_sio, wkq_err; ++ ++ do_sio = au_test_h_perm_sio(dir, MAY_EXEC | MAY_WRITE); ++ if (!do_sio) ++ err = vfsub_rmdir(dir, path); ++ else { ++ struct au_vfsub_rmdir_args args = { ++ .errp = &err, ++ .dir = dir, ++ .path = path ++ }; ++ wkq_err = au_wkq_wait(au_call_vfsub_rmdir, &args); ++ if (unlikely(wkq_err)) ++ err = wkq_err; ++ } ++ ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++struct notify_change_args { ++ int *errp; ++ struct path *path; ++ struct iattr *ia; ++}; ++ ++static void call_notify_change(void *args) ++{ ++ struct notify_change_args *a = args; ++ struct inode *h_inode; ++ ++ h_inode = a->path->dentry->d_inode; ++ IMustLock(h_inode); ++ ++ *a->errp = -EPERM; ++ if (!IS_IMMUTABLE(h_inode) && !IS_APPEND(h_inode)) { ++ lockdep_off(); ++ *a->errp = notify_change(a->path->dentry, a->ia); ++ lockdep_on(); ++ if (!*a->errp) ++ vfsub_update_h_iattr(a->path, /*did*/NULL); /*ignore*/ ++ } ++ AuTraceErr(*a->errp); ++} ++ ++int vfsub_notify_change(struct path *path, struct iattr *ia) ++{ ++ int err; ++ struct notify_change_args args = { ++ .errp = &err, ++ .path = path, ++ .ia = ia ++ }; ++ ++ call_notify_change(&args); ++ ++ return err; ++} ++ ++int vfsub_sio_notify_change(struct path *path, struct iattr *ia) ++{ ++ int err, wkq_err; ++ struct notify_change_args args = { ++ .errp = &err, ++ .path = path, ++ .ia = ia ++ }; ++ ++ wkq_err = au_wkq_wait(call_notify_change, &args); ++ if (unlikely(wkq_err)) ++ err = wkq_err; ++ ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++struct unlink_args { ++ int *errp; ++ struct inode *dir; ++ struct path *path; ++}; ++ ++static void call_unlink(void *args) ++{ ++ struct unlink_args *a = args; ++ struct dentry *d = a->path->dentry; ++ struct inode *h_inode; ++ const int stop_sillyrename = (au_test_nfs(d->d_sb) ++ && atomic_read(&d->d_count) == 1); ++ ++ IMustLock(a->dir); ++ ++ a->path->dentry = d->d_parent; ++ *a->errp = security_path_unlink(a->path, d); ++ a->path->dentry = d; ++ if (unlikely(*a->errp)) ++ return; ++ ++ if (!stop_sillyrename) ++ dget(d); ++ h_inode = d->d_inode; ++ if (h_inode) ++ atomic_inc(&h_inode->i_count); ++ ++ lockdep_off(); ++ *a->errp = vfs_unlink(a->dir, d); ++ lockdep_on(); ++ if (!*a->errp) { ++ struct path tmp = { ++ .dentry = d->d_parent, ++ .mnt = a->path->mnt ++ }; ++ vfsub_update_h_iattr(&tmp, /*did*/NULL); /*ignore*/ ++ } ++ ++ if (!stop_sillyrename) ++ dput(d); ++ if (h_inode) ++ iput(h_inode); ++ ++ AuTraceErr(*a->errp); ++} ++ ++/* ++ * @dir: must be locked. ++ * @dentry: target dentry. ++ */ ++int vfsub_unlink(struct inode *dir, struct path *path, int force) ++{ ++ int err; ++ struct unlink_args args = { ++ .errp = &err, ++ .dir = dir, ++ .path = path ++ }; ++ ++ if (!force) ++ call_unlink(&args); ++ else { ++ int wkq_err; ++ ++ wkq_err = au_wkq_wait(call_unlink, &args); ++ if (unlikely(wkq_err)) ++ err = wkq_err; ++ } ++ ++ return err; ++} +diff -Nur linux-2.6.31.4.orig/fs/aufs/vfsub.h linux-2.6.31.4/fs/aufs/vfsub.h +--- linux-2.6.31.4.orig/fs/aufs/vfsub.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/vfsub.h 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,172 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * sub-routines for VFS ++ */ ++ ++#ifndef __AUFS_VFSUB_H__ ++#define __AUFS_VFSUB_H__ ++ ++#ifdef __KERNEL__ ++ ++#include <linux/fs.h> ++#include <linux/fs_stack.h> ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* lock subclass for lower inode */ ++/* default MAX_LOCKDEP_SUBCLASSES(8) is not enough */ ++/* reduce? gave up. */ ++enum { ++ AuLsc_I_Begin = I_MUTEX_QUOTA, /* 4 */ ++ AuLsc_I_PARENT, /* lower inode, parent first */ ++ AuLsc_I_PARENT2, /* copyup dirs */ ++ AuLsc_I_PARENT3, /* copyup wh */ ++ AuLsc_I_CHILD, ++ AuLsc_I_CHILD2, ++ AuLsc_I_End ++}; ++ ++/* to debug easier, do not make them inlined functions */ ++#define MtxMustLock(mtx) AuDebugOn(!mutex_is_locked(mtx)) ++#define IMustLock(i) MtxMustLock(&(i)->i_mutex) ++ ++/* ---------------------------------------------------------------------- */ ++ ++static inline void vfsub_copy_inode_size(struct inode *inode, ++ struct inode *h_inode) ++{ ++ spin_lock(&inode->i_lock); ++ fsstack_copy_inode_size(inode, h_inode); ++ spin_unlock(&inode->i_lock); ++} ++ ++int vfsub_update_h_iattr(struct path *h_path, int *did); ++struct file *vfsub_filp_open(const char *path, int oflags, int mode); ++int vfsub_kern_path(const char *name, unsigned int flags, struct path *path); ++struct dentry *vfsub_lookup_one_len(const char *name, struct dentry *parent, ++ int len); ++struct dentry *vfsub_lookup_hash(struct nameidata *nd); ++ ++/* ---------------------------------------------------------------------- */ ++ ++struct au_hinode; ++struct dentry *vfsub_lock_rename(struct dentry *d1, struct au_hinode *hdir1, ++ struct dentry *d2, struct au_hinode *hdir2); ++void vfsub_unlock_rename(struct dentry *d1, struct au_hinode *hdir1, ++ struct dentry *d2, struct au_hinode *hdir2); ++ ++int vfsub_create(struct inode *dir, struct path *path, int mode); ++int vfsub_symlink(struct inode *dir, struct path *path, ++ const char *symname); ++int vfsub_mknod(struct inode *dir, struct path *path, int mode, dev_t dev); ++int vfsub_link(struct dentry *src_dentry, struct inode *dir, ++ struct path *path); ++int vfsub_rename(struct inode *src_hdir, struct dentry *src_dentry, ++ struct inode *hdir, struct path *path); ++int vfsub_mkdir(struct inode *dir, struct path *path, int mode); ++int vfsub_rmdir(struct inode *dir, struct path *path); ++ ++/* ---------------------------------------------------------------------- */ ++ ++ssize_t vfsub_read_u(struct file *file, char __user *ubuf, size_t count, ++ loff_t *ppos); ++ssize_t vfsub_read_k(struct file *file, void *kbuf, size_t count, ++ loff_t *ppos); ++ssize_t vfsub_write_u(struct file *file, const char __user *ubuf, size_t count, ++ loff_t *ppos); ++ssize_t vfsub_write_k(struct file *file, void *kbuf, size_t count, ++ loff_t *ppos); ++int vfsub_readdir(struct file *file, filldir_t filldir, void *arg); ++ ++static inline void vfsub_file_accessed(struct file *h_file) ++{ ++ file_accessed(h_file); ++ vfsub_update_h_iattr(&h_file->f_path, /*did*/NULL); /*ignore*/ ++} ++ ++static inline void vfsub_touch_atime(struct vfsmount *h_mnt, ++ struct dentry *h_dentry) ++{ ++ struct path h_path = { ++ .dentry = h_dentry, ++ .mnt = h_mnt ++ }; ++ touch_atime(h_mnt, h_dentry); ++ vfsub_update_h_iattr(&h_path, /*did*/NULL); /*ignore*/ ++} ++ ++long vfsub_splice_to(struct file *in, loff_t *ppos, ++ struct pipe_inode_info *pipe, size_t len, ++ unsigned int flags); ++long vfsub_splice_from(struct pipe_inode_info *pipe, struct file *out, ++ loff_t *ppos, size_t len, unsigned int flags); ++int vfsub_trunc(struct path *h_path, loff_t length, unsigned int attr, ++ struct file *h_file); ++ ++/* ---------------------------------------------------------------------- */ ++ ++static inline loff_t vfsub_llseek(struct file *file, loff_t offset, int origin) ++{ ++ loff_t err; ++ ++ lockdep_off(); ++ err = vfs_llseek(file, offset, origin); ++ lockdep_on(); ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* dirty workaround for strict type of fmode_t */ ++union vfsub_fmu { ++ fmode_t fm; ++ unsigned int ui; ++}; ++ ++static inline unsigned int vfsub_fmode_to_uint(fmode_t fm) ++{ ++ union vfsub_fmu u = { ++ .fm = fm ++ }; ++ ++ BUILD_BUG_ON(sizeof(u.fm) != sizeof(u.ui)); ++ ++ return u.ui; ++} ++ ++static inline fmode_t vfsub_uint_to_fmode(unsigned int ui) ++{ ++ union vfsub_fmu u = { ++ .ui = ui ++ }; ++ ++ return u.fm; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++int vfsub_sio_mkdir(struct inode *dir, struct path *path, int mode); ++int vfsub_sio_rmdir(struct inode *dir, struct path *path); ++int vfsub_sio_notify_change(struct path *path, struct iattr *ia); ++int vfsub_notify_change(struct path *path, struct iattr *ia); ++int vfsub_unlink(struct inode *dir, struct path *path, int force); ++ ++#endif /* __KERNEL__ */ ++#endif /* __AUFS_VFSUB_H__ */ +diff -Nur linux-2.6.31.4.orig/fs/aufs/wbr_policy.c linux-2.6.31.4/fs/aufs/wbr_policy.c +--- linux-2.6.31.4.orig/fs/aufs/wbr_policy.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/wbr_policy.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,641 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * policies for selecting one among multiple writable branches ++ */ ++ ++#include <linux/statfs.h> ++#include "aufs.h" ++ ++/* subset of cpup_attr() */ ++static noinline_for_stack ++int au_cpdown_attr(struct path *h_path, struct dentry *h_src) ++{ ++ int err, sbits; ++ struct iattr ia; ++ struct inode *h_isrc; ++ ++ h_isrc = h_src->d_inode; ++ ia.ia_valid = ATTR_FORCE | ATTR_MODE | ATTR_UID | ATTR_GID; ++ ia.ia_mode = h_isrc->i_mode; ++ ia.ia_uid = h_isrc->i_uid; ++ ia.ia_gid = h_isrc->i_gid; ++ sbits = !!(ia.ia_mode & (S_ISUID | S_ISGID)); ++ au_cpup_attr_flags(h_path->dentry->d_inode, h_isrc); ++ err = vfsub_sio_notify_change(h_path, &ia); ++ ++ /* is this nfs only? */ ++ if (!err && sbits && au_test_nfs(h_path->dentry->d_sb)) { ++ ia.ia_valid = ATTR_FORCE | ATTR_MODE; ++ ia.ia_mode = h_isrc->i_mode; ++ err = vfsub_sio_notify_change(h_path, &ia); ++ } ++ ++ return err; ++} ++ ++#define AuCpdown_PARENT_OPQ 1 ++#define AuCpdown_WHED (1 << 1) ++#define AuCpdown_MADE_DIR (1 << 2) ++#define AuCpdown_DIROPQ (1 << 3) ++#define au_ftest_cpdown(flags, name) ((flags) & AuCpdown_##name) ++#define au_fset_cpdown(flags, name) { (flags) |= AuCpdown_##name; } ++#define au_fclr_cpdown(flags, name) { (flags) &= ~AuCpdown_##name; } ++ ++struct au_cpdown_dir_args { ++ struct dentry *parent; ++ unsigned int flags; ++}; ++ ++static int au_cpdown_dir_opq(struct dentry *dentry, aufs_bindex_t bdst, ++ struct au_cpdown_dir_args *a) ++{ ++ int err; ++ struct dentry *opq_dentry; ++ ++ opq_dentry = au_diropq_create(dentry, bdst); ++ err = PTR_ERR(opq_dentry); ++ if (IS_ERR(opq_dentry)) ++ goto out; ++ dput(opq_dentry); ++ au_fset_cpdown(a->flags, DIROPQ); ++ ++ out: ++ return err; ++} ++ ++static int au_cpdown_dir_wh(struct dentry *dentry, struct dentry *h_parent, ++ struct inode *dir, aufs_bindex_t bdst) ++{ ++ int err; ++ struct path h_path; ++ struct au_branch *br; ++ ++ br = au_sbr(dentry->d_sb, bdst); ++ h_path.dentry = au_wh_lkup(h_parent, &dentry->d_name, br); ++ err = PTR_ERR(h_path.dentry); ++ if (IS_ERR(h_path.dentry)) ++ goto out; ++ ++ err = 0; ++ if (h_path.dentry->d_inode) { ++ h_path.mnt = br->br_mnt; ++ err = au_wh_unlink_dentry(au_h_iptr(dir, bdst), &h_path, ++ dentry); ++ } ++ dput(h_path.dentry); ++ ++ out: ++ return err; ++} ++ ++static int au_cpdown_dir(struct dentry *dentry, aufs_bindex_t bdst, ++ struct dentry *h_parent, void *arg) ++{ ++ int err, rerr; ++ aufs_bindex_t bend, bopq, bstart; ++ unsigned char parent_opq; ++ struct path h_path; ++ struct dentry *parent; ++ struct inode *h_dir, *h_inode, *inode, *dir; ++ struct au_cpdown_dir_args *args = arg; ++ ++ bstart = au_dbstart(dentry); ++ /* dentry is di-locked */ ++ parent = dget_parent(dentry); ++ dir = parent->d_inode; ++ h_dir = h_parent->d_inode; ++ AuDebugOn(h_dir != au_h_iptr(dir, bdst)); ++ IMustLock(h_dir); ++ ++ err = au_lkup_neg(dentry, bdst); ++ if (unlikely(err < 0)) ++ goto out; ++ h_path.dentry = au_h_dptr(dentry, bdst); ++ h_path.mnt = au_sbr_mnt(dentry->d_sb, bdst); ++ err = vfsub_sio_mkdir(au_h_iptr(dir, bdst), &h_path, ++ S_IRWXU | S_IRUGO | S_IXUGO); ++ if (unlikely(err)) ++ goto out_put; ++ au_fset_cpdown(args->flags, MADE_DIR); ++ ++ bend = au_dbend(dentry); ++ bopq = au_dbdiropq(dentry); ++ au_fclr_cpdown(args->flags, WHED); ++ au_fclr_cpdown(args->flags, DIROPQ); ++ if (au_dbwh(dentry) == bdst) ++ au_fset_cpdown(args->flags, WHED); ++ if (!au_ftest_cpdown(args->flags, PARENT_OPQ) && bopq <= bdst) ++ au_fset_cpdown(args->flags, PARENT_OPQ); ++ parent_opq = (au_ftest_cpdown(args->flags, PARENT_OPQ) ++ && args->parent == dentry); ++ h_inode = h_path.dentry->d_inode; ++ mutex_lock_nested(&h_inode->i_mutex, AuLsc_I_CHILD); ++ if (au_ftest_cpdown(args->flags, WHED)) { ++ err = au_cpdown_dir_opq(dentry, bdst, args); ++ if (unlikely(err)) { ++ mutex_unlock(&h_inode->i_mutex); ++ goto out_dir; ++ } ++ } ++ ++ err = au_cpdown_attr(&h_path, au_h_dptr(dentry, bstart)); ++ mutex_unlock(&h_inode->i_mutex); ++ if (unlikely(err)) ++ goto out_opq; ++ ++ if (au_ftest_cpdown(args->flags, WHED)) { ++ err = au_cpdown_dir_wh(dentry, h_parent, dir, bdst); ++ if (unlikely(err)) ++ goto out_opq; ++ } ++ ++ inode = dentry->d_inode; ++ if (au_ibend(inode) < bdst) ++ au_set_ibend(inode, bdst); ++ au_set_h_iptr(inode, bdst, au_igrab(h_inode), ++ au_hi_flags(inode, /*isdir*/1)); ++ goto out; /* success */ ++ ++ /* revert */ ++ out_opq: ++ if (au_ftest_cpdown(args->flags, DIROPQ)) { ++ mutex_lock_nested(&h_inode->i_mutex, AuLsc_I_CHILD); ++ rerr = au_diropq_remove(dentry, bdst); ++ mutex_unlock(&h_inode->i_mutex); ++ if (unlikely(rerr)) { ++ AuIOErr("failed removing diropq for %.*s b%d (%d)\n", ++ AuDLNPair(dentry), bdst, rerr); ++ err = -EIO; ++ goto out; ++ } ++ } ++ out_dir: ++ if (au_ftest_cpdown(args->flags, MADE_DIR)) { ++ rerr = vfsub_sio_rmdir(au_h_iptr(dir, bdst), &h_path); ++ if (unlikely(rerr)) { ++ AuIOErr("failed removing %.*s b%d (%d)\n", ++ AuDLNPair(dentry), bdst, rerr); ++ err = -EIO; ++ } ++ } ++ out_put: ++ au_set_h_dptr(dentry, bdst, NULL); ++ if (au_dbend(dentry) == bdst) ++ au_update_dbend(dentry); ++ out: ++ dput(parent); ++ return err; ++} ++ ++int au_cpdown_dirs(struct dentry *dentry, aufs_bindex_t bdst) ++{ ++ int err; ++ struct au_cpdown_dir_args args = { ++ .parent = dget_parent(dentry), ++ .flags = 0 ++ }; ++ ++ err = au_cp_dirs(dentry, bdst, au_cpdown_dir, &args); ++ dput(args.parent); ++ ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* policies for create */ ++ ++static int au_wbr_bu(struct super_block *sb, aufs_bindex_t bindex) ++{ ++ for (; bindex >= 0; bindex--) ++ if (!au_br_rdonly(au_sbr(sb, bindex))) ++ return bindex; ++ return -EROFS; ++} ++ ++/* top down parent */ ++static int au_wbr_create_tdp(struct dentry *dentry, int isdir __maybe_unused) ++{ ++ int err; ++ aufs_bindex_t bstart, bindex; ++ struct super_block *sb; ++ struct dentry *parent, *h_parent; ++ ++ sb = dentry->d_sb; ++ bstart = au_dbstart(dentry); ++ err = bstart; ++ if (!au_br_rdonly(au_sbr(sb, bstart))) ++ goto out; ++ ++ err = -EROFS; ++ parent = dget_parent(dentry); ++ for (bindex = au_dbstart(parent); bindex < bstart; bindex++) { ++ h_parent = au_h_dptr(parent, bindex); ++ if (!h_parent || !h_parent->d_inode) ++ continue; ++ ++ if (!au_br_rdonly(au_sbr(sb, bindex))) { ++ err = bindex; ++ break; ++ } ++ } ++ dput(parent); ++ ++ /* bottom up here */ ++ if (unlikely(err < 0)) ++ err = au_wbr_bu(sb, bstart - 1); ++ ++ out: ++ AuDbg("b%d\n", err); ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* an exception for the policy other than tdp */ ++static int au_wbr_create_exp(struct dentry *dentry) ++{ ++ int err; ++ aufs_bindex_t bwh, bdiropq; ++ struct dentry *parent; ++ ++ err = -1; ++ bwh = au_dbwh(dentry); ++ parent = dget_parent(dentry); ++ bdiropq = au_dbdiropq(parent); ++ if (bwh >= 0) { ++ if (bdiropq >= 0) ++ err = min(bdiropq, bwh); ++ else ++ err = bwh; ++ AuDbg("%d\n", err); ++ } else if (bdiropq >= 0) { ++ err = bdiropq; ++ AuDbg("%d\n", err); ++ } ++ dput(parent); ++ ++ if (err >= 0 && au_br_rdonly(au_sbr(dentry->d_sb, err))) ++ err = -1; ++ ++ AuDbg("%d\n", err); ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* round robin */ ++static int au_wbr_create_init_rr(struct super_block *sb) ++{ ++ int err; ++ ++ err = au_wbr_bu(sb, au_sbend(sb)); ++ atomic_set(&au_sbi(sb)->si_wbr_rr_next, -err); /* less important */ ++ /* smp_mb(); */ ++ ++ AuDbg("b%d\n", err); ++ return err; ++} ++ ++static int au_wbr_create_rr(struct dentry *dentry, int isdir) ++{ ++ int err, nbr; ++ unsigned int u; ++ aufs_bindex_t bindex, bend; ++ struct super_block *sb; ++ atomic_t *next; ++ ++ err = au_wbr_create_exp(dentry); ++ if (err >= 0) ++ goto out; ++ ++ sb = dentry->d_sb; ++ next = &au_sbi(sb)->si_wbr_rr_next; ++ bend = au_sbend(sb); ++ nbr = bend + 1; ++ for (bindex = 0; bindex <= bend; bindex++) { ++ if (!isdir) { ++ err = atomic_dec_return(next) + 1; ++ /* modulo for 0 is meaningless */ ++ if (unlikely(!err)) ++ err = atomic_dec_return(next) + 1; ++ } else ++ err = atomic_read(next); ++ AuDbg("%d\n", err); ++ u = err; ++ err = u % nbr; ++ AuDbg("%d\n", err); ++ if (!au_br_rdonly(au_sbr(sb, err))) ++ break; ++ err = -EROFS; ++ } ++ ++ out: ++ AuDbg("%d\n", err); ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* most free space */ ++static void au_mfs(struct dentry *dentry) ++{ ++ struct super_block *sb; ++ struct au_branch *br; ++ struct au_wbr_mfs *mfs; ++ aufs_bindex_t bindex, bend; ++ int err; ++ unsigned long long b, bavail; ++ /* reduce the stack usage */ ++ struct kstatfs *st; ++ ++ st = kmalloc(sizeof(*st), GFP_NOFS); ++ if (unlikely(!st)) { ++ AuWarn1("failed updating mfs(%d), ignored\n", -ENOMEM); ++ return; ++ } ++ ++ bavail = 0; ++ sb = dentry->d_sb; ++ mfs = &au_sbi(sb)->si_wbr_mfs; ++ MtxMustLock(&mfs->mfs_lock); ++ mfs->mfs_bindex = -EROFS; ++ mfs->mfsrr_bytes = 0; ++ bend = au_sbend(sb); ++ for (bindex = 0; bindex <= bend; bindex++) { ++ br = au_sbr(sb, bindex); ++ if (au_br_rdonly(br)) ++ continue; ++ ++ /* sb->s_root for NFS is unreliable */ ++ err = vfs_statfs(br->br_mnt->mnt_root, st); ++ if (unlikely(err)) { ++ AuWarn1("failed statfs, b%d, %d\n", bindex, err); ++ continue; ++ } ++ ++ /* when the available size is equal, select the lower one */ ++ BUILD_BUG_ON(sizeof(b) < sizeof(st->f_bavail) ++ || sizeof(b) < sizeof(st->f_bsize)); ++ b = st->f_bavail * st->f_bsize; ++ br->br_wbr->wbr_bytes = b; ++ if (b >= bavail) { ++ bavail = b; ++ mfs->mfs_bindex = bindex; ++ mfs->mfs_jiffy = jiffies; ++ } ++ } ++ ++ mfs->mfsrr_bytes = bavail; ++ AuDbg("b%d\n", mfs->mfs_bindex); ++ kfree(st); ++} ++ ++static int au_wbr_create_mfs(struct dentry *dentry, int isdir __maybe_unused) ++{ ++ int err; ++ struct super_block *sb; ++ struct au_wbr_mfs *mfs; ++ ++ err = au_wbr_create_exp(dentry); ++ if (err >= 0) ++ goto out; ++ ++ sb = dentry->d_sb; ++ mfs = &au_sbi(sb)->si_wbr_mfs; ++ mutex_lock(&mfs->mfs_lock); ++ if (time_after(jiffies, mfs->mfs_jiffy + mfs->mfs_expire) ++ || mfs->mfs_bindex < 0 ++ || au_br_rdonly(au_sbr(sb, mfs->mfs_bindex))) ++ au_mfs(dentry); ++ mutex_unlock(&mfs->mfs_lock); ++ err = mfs->mfs_bindex; ++ ++ out: ++ AuDbg("b%d\n", err); ++ return err; ++} ++ ++static int au_wbr_create_init_mfs(struct super_block *sb) ++{ ++ struct au_wbr_mfs *mfs; ++ ++ mfs = &au_sbi(sb)->si_wbr_mfs; ++ mutex_init(&mfs->mfs_lock); ++ mfs->mfs_jiffy = 0; ++ mfs->mfs_bindex = -EROFS; ++ ++ return 0; ++} ++ ++static int au_wbr_create_fin_mfs(struct super_block *sb __maybe_unused) ++{ ++ mutex_destroy(&au_sbi(sb)->si_wbr_mfs.mfs_lock); ++ return 0; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* most free space and then round robin */ ++static int au_wbr_create_mfsrr(struct dentry *dentry, int isdir) ++{ ++ int err; ++ struct au_wbr_mfs *mfs; ++ ++ err = au_wbr_create_mfs(dentry, isdir); ++ if (err >= 0) { ++ mfs = &au_sbi(dentry->d_sb)->si_wbr_mfs; ++ mutex_lock(&mfs->mfs_lock); ++ if (mfs->mfsrr_bytes < mfs->mfsrr_watermark) ++ err = au_wbr_create_rr(dentry, isdir); ++ mutex_unlock(&mfs->mfs_lock); ++ } ++ ++ AuDbg("b%d\n", err); ++ return err; ++} ++ ++static int au_wbr_create_init_mfsrr(struct super_block *sb) ++{ ++ int err; ++ ++ au_wbr_create_init_mfs(sb); /* ignore */ ++ err = au_wbr_create_init_rr(sb); ++ ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* top down parent and most free space */ ++static int au_wbr_create_pmfs(struct dentry *dentry, int isdir) ++{ ++ int err, e2; ++ unsigned long long b; ++ aufs_bindex_t bindex, bstart, bend; ++ struct super_block *sb; ++ struct dentry *parent, *h_parent; ++ struct au_branch *br; ++ ++ err = au_wbr_create_tdp(dentry, isdir); ++ if (unlikely(err < 0)) ++ goto out; ++ parent = dget_parent(dentry); ++ bstart = au_dbstart(parent); ++ bend = au_dbtaildir(parent); ++ if (bstart == bend) ++ goto out_parent; /* success */ ++ ++ e2 = au_wbr_create_mfs(dentry, isdir); ++ if (e2 < 0) ++ goto out_parent; /* success */ ++ ++ /* when the available size is equal, select upper one */ ++ sb = dentry->d_sb; ++ br = au_sbr(sb, err); ++ b = br->br_wbr->wbr_bytes; ++ AuDbg("b%d, %llu\n", err, b); ++ ++ for (bindex = bstart; bindex <= bend; bindex++) { ++ h_parent = au_h_dptr(parent, bindex); ++ if (!h_parent || !h_parent->d_inode) ++ continue; ++ ++ br = au_sbr(sb, bindex); ++ if (!au_br_rdonly(br) && br->br_wbr->wbr_bytes > b) { ++ b = br->br_wbr->wbr_bytes; ++ err = bindex; ++ AuDbg("b%d, %llu\n", err, b); ++ } ++ } ++ ++ out_parent: ++ dput(parent); ++ out: ++ AuDbg("b%d\n", err); ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* policies for copyup */ ++ ++/* top down parent */ ++static int au_wbr_copyup_tdp(struct dentry *dentry) ++{ ++ return au_wbr_create_tdp(dentry, /*isdir, anything is ok*/0); ++} ++ ++/* bottom up parent */ ++static int au_wbr_copyup_bup(struct dentry *dentry) ++{ ++ int err; ++ aufs_bindex_t bindex, bstart; ++ struct dentry *parent, *h_parent; ++ struct super_block *sb; ++ ++ err = -EROFS; ++ sb = dentry->d_sb; ++ parent = dget_parent(dentry); ++ bstart = au_dbstart(parent); ++ for (bindex = au_dbstart(dentry); bindex >= bstart; bindex--) { ++ h_parent = au_h_dptr(parent, bindex); ++ if (!h_parent || !h_parent->d_inode) ++ continue; ++ ++ if (!au_br_rdonly(au_sbr(sb, bindex))) { ++ err = bindex; ++ break; ++ } ++ } ++ dput(parent); ++ ++ /* bottom up here */ ++ if (unlikely(err < 0)) ++ err = au_wbr_bu(sb, bstart - 1); ++ ++ AuDbg("b%d\n", err); ++ return err; ++} ++ ++/* bottom up */ ++static int au_wbr_copyup_bu(struct dentry *dentry) ++{ ++ int err; ++ ++ err = au_wbr_bu(dentry->d_sb, au_dbstart(dentry)); ++ ++ AuDbg("b%d\n", err); ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++struct au_wbr_copyup_operations au_wbr_copyup_ops[] = { ++ [AuWbrCopyup_TDP] = { ++ .copyup = au_wbr_copyup_tdp ++ }, ++ [AuWbrCopyup_BUP] = { ++ .copyup = au_wbr_copyup_bup ++ }, ++ [AuWbrCopyup_BU] = { ++ .copyup = au_wbr_copyup_bu ++ } ++}; ++ ++struct au_wbr_create_operations au_wbr_create_ops[] = { ++ [AuWbrCreate_TDP] = { ++ .create = au_wbr_create_tdp ++ }, ++ [AuWbrCreate_RR] = { ++ .create = au_wbr_create_rr, ++ .init = au_wbr_create_init_rr ++ }, ++ [AuWbrCreate_MFS] = { ++ .create = au_wbr_create_mfs, ++ .init = au_wbr_create_init_mfs, ++ .fin = au_wbr_create_fin_mfs ++ }, ++ [AuWbrCreate_MFSV] = { ++ .create = au_wbr_create_mfs, ++ .init = au_wbr_create_init_mfs, ++ .fin = au_wbr_create_fin_mfs ++ }, ++ [AuWbrCreate_MFSRR] = { ++ .create = au_wbr_create_mfsrr, ++ .init = au_wbr_create_init_mfsrr, ++ .fin = au_wbr_create_fin_mfs ++ }, ++ [AuWbrCreate_MFSRRV] = { ++ .create = au_wbr_create_mfsrr, ++ .init = au_wbr_create_init_mfsrr, ++ .fin = au_wbr_create_fin_mfs ++ }, ++ [AuWbrCreate_PMFS] = { ++ .create = au_wbr_create_pmfs, ++ .init = au_wbr_create_init_mfs, ++ .fin = au_wbr_create_fin_mfs ++ }, ++ [AuWbrCreate_PMFSV] = { ++ .create = au_wbr_create_pmfs, ++ .init = au_wbr_create_init_mfs, ++ .fin = au_wbr_create_fin_mfs ++ } ++}; +diff -Nur linux-2.6.31.4.orig/fs/aufs/whout.c linux-2.6.31.4/fs/aufs/whout.c +--- linux-2.6.31.4.orig/fs/aufs/whout.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/whout.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,1048 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * whiteout for logical deletion and opaque directory ++ */ ++ ++#include <linux/fs.h> ++#include "aufs.h" ++ ++#define WH_MASK S_IRUGO ++ ++/* ++ * If a directory contains this file, then it is opaque. We start with the ++ * .wh. flag so that it is blocked by lookup. ++ */ ++static struct qstr diropq_name = { ++ .name = AUFS_WH_DIROPQ, ++ .len = sizeof(AUFS_WH_DIROPQ) - 1 ++}; ++ ++/* ++ * generate whiteout name, which is NOT terminated by NULL. ++ * @name: original d_name.name ++ * @len: original d_name.len ++ * @wh: whiteout qstr ++ * returns zero when succeeds, otherwise error. ++ * succeeded value as wh->name should be freed by kfree(). ++ */ ++int au_wh_name_alloc(struct qstr *wh, const struct qstr *name) ++{ ++ char *p; ++ ++ if (unlikely(name->len > PATH_MAX - AUFS_WH_PFX_LEN)) ++ return -ENAMETOOLONG; ++ ++ wh->len = name->len + AUFS_WH_PFX_LEN; ++ p = kmalloc(wh->len, GFP_NOFS); ++ wh->name = p; ++ if (p) { ++ memcpy(p, AUFS_WH_PFX, AUFS_WH_PFX_LEN); ++ memcpy(p + AUFS_WH_PFX_LEN, name->name, name->len); ++ /* smp_mb(); */ ++ return 0; ++ } ++ return -ENOMEM; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* ++ * test if the @wh_name exists under @h_parent. ++ * @try_sio specifies the necessary of super-io. ++ */ ++int au_wh_test(struct dentry *h_parent, struct qstr *wh_name, ++ struct au_branch *br, int try_sio) ++{ ++ int err; ++ struct dentry *wh_dentry; ++ struct inode *h_dir; ++ ++ h_dir = h_parent->d_inode; ++ if (!try_sio) ++ wh_dentry = au_lkup_one(wh_name, h_parent, br, /*nd*/NULL); ++ else ++ wh_dentry = au_sio_lkup_one(wh_name, h_parent, br); ++ err = PTR_ERR(wh_dentry); ++ if (IS_ERR(wh_dentry)) ++ goto out; ++ ++ err = 0; ++ if (!wh_dentry->d_inode) ++ goto out_wh; /* success */ ++ ++ err = 1; ++ if (S_ISREG(wh_dentry->d_inode->i_mode)) ++ goto out_wh; /* success */ ++ ++ err = -EIO; ++ AuIOErr("%.*s Invalid whiteout entry type 0%o.\n", ++ AuDLNPair(wh_dentry), wh_dentry->d_inode->i_mode); ++ ++ out_wh: ++ dput(wh_dentry); ++ out: ++ return err; ++} ++ ++/* ++ * test if the @h_dentry sets opaque or not. ++ */ ++int au_diropq_test(struct dentry *h_dentry, struct au_branch *br) ++{ ++ int err; ++ struct inode *h_dir; ++ ++ h_dir = h_dentry->d_inode; ++ err = au_wh_test(h_dentry, &diropq_name, br, ++ au_test_h_perm_sio(h_dir, MAY_EXEC)); ++ return err; ++} ++ ++/* ++ * returns a negative dentry whose name is unique and temporary. ++ */ ++struct dentry *au_whtmp_lkup(struct dentry *h_parent, struct au_branch *br, ++ struct qstr *prefix) ++{ ++#define HEX_LEN 4 ++ struct dentry *dentry; ++ int i; ++ char defname[AUFS_WH_PFX_LEN * 2 + DNAME_INLINE_LEN_MIN + 1 ++ + HEX_LEN + 1], *name, *p; ++ static unsigned short cnt; ++ struct qstr qs; ++ ++ name = defname; ++ qs.len = sizeof(defname) - DNAME_INLINE_LEN_MIN + prefix->len - 1; ++ if (unlikely(prefix->len > DNAME_INLINE_LEN_MIN)) { ++ dentry = ERR_PTR(-ENAMETOOLONG); ++ if (unlikely(qs.len >= PATH_MAX)) ++ goto out; ++ dentry = ERR_PTR(-ENOMEM); ++ name = kmalloc(qs.len + 1, GFP_NOFS); ++ if (unlikely(!name)) ++ goto out; ++ } ++ ++ /* doubly whiteout-ed */ ++ memcpy(name, AUFS_WH_PFX AUFS_WH_PFX, AUFS_WH_PFX_LEN * 2); ++ p = name + AUFS_WH_PFX_LEN * 2; ++ memcpy(p, prefix->name, prefix->len); ++ p += prefix->len; ++ *p++ = '.'; ++ AuDebugOn(name + qs.len + 1 - p <= HEX_LEN); ++ ++ qs.name = name; ++ for (i = 0; i < 3; i++) { ++ sprintf(p, "%.*d", HEX_LEN, cnt++); ++ dentry = au_sio_lkup_one(&qs, h_parent, br); ++ if (IS_ERR(dentry) || !dentry->d_inode) ++ goto out_name; ++ dput(dentry); ++ } ++ /* AuWarn("could not get random name\n"); */ ++ dentry = ERR_PTR(-EEXIST); ++ AuDbg("%.*s\n", AuLNPair(&qs)); ++ BUG(); ++ ++ out_name: ++ if (name != defname) ++ kfree(name); ++ out: ++ return dentry; ++#undef HEX_LEN ++} ++ ++/* ++ * rename the @h_dentry on @br to the whiteouted temporary name. ++ */ ++int au_whtmp_ren(struct dentry *h_dentry, struct au_branch *br) ++{ ++ int err; ++ struct path h_path = { ++ .mnt = br->br_mnt ++ }; ++ struct inode *h_dir; ++ struct dentry *h_parent; ++ ++ h_parent = h_dentry->d_parent; /* dir inode is locked */ ++ h_dir = h_parent->d_inode; ++ IMustLock(h_dir); ++ ++ h_path.dentry = au_whtmp_lkup(h_parent, br, &h_dentry->d_name); ++ err = PTR_ERR(h_path.dentry); ++ if (IS_ERR(h_path.dentry)) ++ goto out; ++ ++ /* under the same dir, no need to lock_rename() */ ++ err = vfsub_rename(h_dir, h_dentry, h_dir, &h_path); ++ AuTraceErr(err); ++ dput(h_path.dentry); ++ ++ out: ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++/* ++ * functions for removing a whiteout ++ */ ++ ++static int do_unlink_wh(struct inode *h_dir, struct path *h_path) ++{ ++ int force; ++ ++ /* ++ * forces superio when the dir has a sticky bit. ++ * this may be a violation of unix fs semantics. ++ */ ++ force = (h_dir->i_mode & S_ISVTX) ++ && h_path->dentry->d_inode->i_uid != current_fsuid(); ++ return vfsub_unlink(h_dir, h_path, force); ++} ++ ++int au_wh_unlink_dentry(struct inode *h_dir, struct path *h_path, ++ struct dentry *dentry) ++{ ++ int err; ++ ++ err = do_unlink_wh(h_dir, h_path); ++ if (!err && dentry) ++ au_set_dbwh(dentry, -1); ++ ++ return err; ++} ++ ++static int unlink_wh_name(struct dentry *h_parent, struct qstr *wh, ++ struct au_branch *br) ++{ ++ int err; ++ struct path h_path = { ++ .mnt = br->br_mnt ++ }; ++ ++ err = 0; ++ h_path.dentry = au_lkup_one(wh, h_parent, br, /*nd*/NULL); ++ if (IS_ERR(h_path.dentry)) ++ err = PTR_ERR(h_path.dentry); ++ else { ++ if (h_path.dentry->d_inode ++ && S_ISREG(h_path.dentry->d_inode->i_mode)) ++ err = do_unlink_wh(h_parent->d_inode, &h_path); ++ dput(h_path.dentry); ++ } ++ ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++/* ++ * initialize/clean whiteout for a branch ++ */ ++ ++static void au_wh_clean(struct inode *h_dir, struct path *whpath, ++ const int isdir) ++{ ++ int err; ++ ++ if (!whpath->dentry->d_inode) ++ return; ++ ++ err = mnt_want_write(whpath->mnt); ++ if (!err) { ++ if (isdir) ++ err = vfsub_rmdir(h_dir, whpath); ++ else ++ err = vfsub_unlink(h_dir, whpath, /*force*/0); ++ mnt_drop_write(whpath->mnt); ++ } ++ if (unlikely(err)) ++ AuWarn("failed removing %.*s (%d), ignored.\n", ++ AuDLNPair(whpath->dentry), err); ++} ++ ++static int test_linkable(struct dentry *h_root) ++{ ++ struct inode *h_dir = h_root->d_inode; ++ ++ if (h_dir->i_op->link) ++ return 0; ++ ++ AuErr("%.*s (%s) doesn't support link(2), use noplink and rw+nolwh\n", ++ AuDLNPair(h_root), au_sbtype(h_root->d_sb)); ++ return -ENOSYS; ++} ++ ++/* todo: should this mkdir be done in /sbin/mount.aufs helper? */ ++static int au_whdir(struct inode *h_dir, struct path *path) ++{ ++ int err; ++ ++ err = -EEXIST; ++ if (!path->dentry->d_inode) { ++ int mode = S_IRWXU; ++ ++ if (au_test_nfs(path->dentry->d_sb)) ++ mode |= S_IXUGO; ++ err = mnt_want_write(path->mnt); ++ if (!err) { ++ err = vfsub_mkdir(h_dir, path, mode); ++ mnt_drop_write(path->mnt); ++ } ++ } else if (S_ISDIR(path->dentry->d_inode->i_mode)) ++ err = 0; ++ else ++ AuErr("unknown %.*s exists\n", AuDLNPair(path->dentry)); ++ ++ return err; ++} ++ ++struct au_wh_base { ++ const struct qstr *name; ++ struct dentry *dentry; ++}; ++ ++static void au_wh_init_ro(struct inode *h_dir, struct au_wh_base base[], ++ struct path *h_path) ++{ ++ h_path->dentry = base[AuBrWh_BASE].dentry; ++ au_wh_clean(h_dir, h_path, /*isdir*/0); ++ h_path->dentry = base[AuBrWh_PLINK].dentry; ++ au_wh_clean(h_dir, h_path, /*isdir*/1); ++ h_path->dentry = base[AuBrWh_ORPH].dentry; ++ au_wh_clean(h_dir, h_path, /*isdir*/1); ++} ++ ++/* ++ * returns tri-state, ++ * minus: error, caller should print the mesage ++ * zero: succuess ++ * plus: error, caller should NOT print the mesage ++ */ ++static int au_wh_init_rw_nolink(struct dentry *h_root, struct au_wbr *wbr, ++ int do_plink, struct au_wh_base base[], ++ struct path *h_path) ++{ ++ int err; ++ struct inode *h_dir; ++ ++ h_dir = h_root->d_inode; ++ h_path->dentry = base[AuBrWh_BASE].dentry; ++ au_wh_clean(h_dir, h_path, /*isdir*/0); ++ h_path->dentry = base[AuBrWh_PLINK].dentry; ++ if (do_plink) { ++ err = test_linkable(h_root); ++ if (unlikely(err)) { ++ err = 1; ++ goto out; ++ } ++ ++ err = au_whdir(h_dir, h_path); ++ if (unlikely(err)) ++ goto out; ++ wbr->wbr_plink = dget(base[AuBrWh_PLINK].dentry); ++ } else ++ au_wh_clean(h_dir, h_path, /*isdir*/1); ++ h_path->dentry = base[AuBrWh_ORPH].dentry; ++ err = au_whdir(h_dir, h_path); ++ if (unlikely(err)) ++ goto out; ++ wbr->wbr_orph = dget(base[AuBrWh_ORPH].dentry); ++ ++ out: ++ return err; ++} ++ ++/* ++ * for the moment, aufs supports the branch filesystem which does not support ++ * link(2). testing on FAT which does not support i_op->setattr() fully either, ++ * copyup failed. finally, such filesystem will not be used as the writable ++ * branch. ++ * ++ * returns tri-state, see above. ++ */ ++static int au_wh_init_rw(struct dentry *h_root, struct au_wbr *wbr, ++ int do_plink, struct au_wh_base base[], ++ struct path *h_path) ++{ ++ int err; ++ struct inode *h_dir; ++ ++ WbrWhMustWriteLock(wbr); ++ ++ err = test_linkable(h_root); ++ if (unlikely(err)) { ++ err = 1; ++ goto out; ++ } ++ ++ /* ++ * todo: should this create be done in /sbin/mount.aufs helper? ++ */ ++ err = -EEXIST; ++ h_dir = h_root->d_inode; ++ if (!base[AuBrWh_BASE].dentry->d_inode) { ++ err = mnt_want_write(h_path->mnt); ++ if (!err) { ++ h_path->dentry = base[AuBrWh_BASE].dentry; ++ err = vfsub_create(h_dir, h_path, WH_MASK); ++ mnt_drop_write(h_path->mnt); ++ } ++ } else if (S_ISREG(base[AuBrWh_BASE].dentry->d_inode->i_mode)) ++ err = 0; ++ else ++ AuErr("unknown %.*s/%.*s exists\n", ++ AuDLNPair(h_root), AuDLNPair(base[AuBrWh_BASE].dentry)); ++ if (unlikely(err)) ++ goto out; ++ ++ h_path->dentry = base[AuBrWh_PLINK].dentry; ++ if (do_plink) { ++ err = au_whdir(h_dir, h_path); ++ if (unlikely(err)) ++ goto out; ++ wbr->wbr_plink = dget(base[AuBrWh_PLINK].dentry); ++ } else ++ au_wh_clean(h_dir, h_path, /*isdir*/1); ++ wbr->wbr_whbase = dget(base[AuBrWh_BASE].dentry); ++ ++ h_path->dentry = base[AuBrWh_ORPH].dentry; ++ err = au_whdir(h_dir, h_path); ++ if (unlikely(err)) ++ goto out; ++ wbr->wbr_orph = dget(base[AuBrWh_ORPH].dentry); ++ ++ out: ++ return err; ++} ++ ++/* ++ * initialize the whiteout base file/dir for @br. ++ */ ++int au_wh_init(struct dentry *h_root, struct au_branch *br, ++ struct super_block *sb) ++{ ++ int err, i; ++ const unsigned char do_plink ++ = !!au_opt_test(au_mntflags(sb), PLINK); ++ struct path path = { ++ .mnt = br->br_mnt ++ }; ++ struct inode *h_dir; ++ struct au_wbr *wbr = br->br_wbr; ++ static const struct qstr base_name[] = { ++ [AuBrWh_BASE] = { ++ .name = AUFS_BASE_NAME, ++ .len = sizeof(AUFS_BASE_NAME) - 1 ++ }, ++ [AuBrWh_PLINK] = { ++ .name = AUFS_PLINKDIR_NAME, ++ .len = sizeof(AUFS_PLINKDIR_NAME) - 1 ++ }, ++ [AuBrWh_ORPH] = { ++ .name = AUFS_ORPHDIR_NAME, ++ .len = sizeof(AUFS_ORPHDIR_NAME) - 1 ++ } ++ }; ++ struct au_wh_base base[] = { ++ [AuBrWh_BASE] = { ++ .name = base_name + AuBrWh_BASE, ++ .dentry = NULL ++ }, ++ [AuBrWh_PLINK] = { ++ .name = base_name + AuBrWh_PLINK, ++ .dentry = NULL ++ }, ++ [AuBrWh_ORPH] = { ++ .name = base_name + AuBrWh_ORPH, ++ .dentry = NULL ++ } ++ }; ++ ++ if (wbr) ++ WbrWhMustWriteLock(wbr); ++ ++ h_dir = h_root->d_inode; ++ for (i = 0; i < AuBrWh_Last; i++) { ++ /* doubly whiteouted */ ++ struct dentry *d; ++ ++ d = au_wh_lkup(h_root, (void *)base[i].name, br); ++ err = PTR_ERR(d); ++ if (IS_ERR(d)) ++ goto out; ++ ++ base[i].dentry = d; ++ AuDebugOn(wbr ++ && wbr->wbr_wh[i] ++ && wbr->wbr_wh[i] != base[i].dentry); ++ } ++ ++ if (wbr) ++ for (i = 0; i < AuBrWh_Last; i++) { ++ dput(wbr->wbr_wh[i]); ++ wbr->wbr_wh[i] = NULL; ++ } ++ ++ err = 0; ++ ++ switch (br->br_perm) { ++ case AuBrPerm_RO: ++ case AuBrPerm_ROWH: ++ case AuBrPerm_RR: ++ case AuBrPerm_RRWH: ++ au_wh_init_ro(h_dir, base, &path); ++ break; ++ ++ case AuBrPerm_RWNoLinkWH: ++ err = au_wh_init_rw_nolink(h_root, wbr, do_plink, base, &path); ++ if (err > 0) ++ goto out; ++ else if (err) ++ goto out_err; ++ break; ++ ++ case AuBrPerm_RW: ++ err = au_wh_init_rw(h_root, wbr, do_plink, base, &path); ++ if (err > 0) ++ goto out; ++ else if (err) ++ goto out_err; ++ break; ++ ++ default: ++ BUG(); ++ } ++ goto out; /* success */ ++ ++ out_err: ++ AuErr("an error(%d) on the writable branch %.*s(%s)\n", ++ err, AuDLNPair(h_root), au_sbtype(h_root->d_sb)); ++ out: ++ for (i = 0; i < AuBrWh_Last; i++) ++ dput(base[i].dentry); ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++/* ++ * whiteouts are all hard-linked usually. ++ * when its link count reaches a ceiling, we create a new whiteout base ++ * asynchronously. ++ */ ++ ++struct reinit_br_wh { ++ struct super_block *sb; ++ struct au_branch *br; ++}; ++ ++static void reinit_br_wh(void *arg) ++{ ++ int err; ++ aufs_bindex_t bindex; ++ struct path h_path; ++ struct reinit_br_wh *a = arg; ++ struct au_wbr *wbr; ++ struct inode *dir; ++ struct dentry *h_root; ++ struct au_hinode *hdir; ++ ++ err = 0; ++ wbr = a->br->br_wbr; ++ /* big aufs lock */ ++ si_noflush_write_lock(a->sb); ++ if (!au_br_writable(a->br->br_perm)) ++ goto out; ++ bindex = au_br_index(a->sb, a->br->br_id); ++ if (unlikely(bindex < 0)) ++ goto out; ++ ++ di_read_lock_parent(a->sb->s_root, AuLock_IR); ++ dir = a->sb->s_root->d_inode; ++ hdir = au_hi(dir, bindex); ++ h_root = au_h_dptr(a->sb->s_root, bindex); ++ ++ au_hin_imtx_lock_nested(hdir, AuLsc_I_PARENT); ++ wbr_wh_write_lock(wbr); ++ err = au_h_verify(wbr->wbr_whbase, au_opt_udba(a->sb), hdir->hi_inode, ++ h_root, a->br); ++ if (!err) { ++ err = mnt_want_write(a->br->br_mnt); ++ if (!err) { ++ h_path.dentry = wbr->wbr_whbase; ++ h_path.mnt = a->br->br_mnt; ++ err = vfsub_unlink(hdir->hi_inode, &h_path, /*force*/0); ++ mnt_drop_write(a->br->br_mnt); ++ } ++ } else { ++ AuWarn("%.*s is moved, ignored\n", AuDLNPair(wbr->wbr_whbase)); ++ err = 0; ++ } ++ dput(wbr->wbr_whbase); ++ wbr->wbr_whbase = NULL; ++ if (!err) ++ err = au_wh_init(h_root, a->br, a->sb); ++ wbr_wh_write_unlock(wbr); ++ au_hin_imtx_unlock(hdir); ++ di_read_unlock(a->sb->s_root, AuLock_IR); ++ ++ out: ++ if (wbr) ++ atomic_dec(&wbr->wbr_wh_running); ++ atomic_dec(&a->br->br_count); ++ au_nwt_done(&au_sbi(a->sb)->si_nowait); ++ si_write_unlock(a->sb); ++ kfree(arg); ++ if (unlikely(err)) ++ AuIOErr("err %d\n", err); ++} ++ ++static void kick_reinit_br_wh(struct super_block *sb, struct au_branch *br) ++{ ++ int do_dec, wkq_err; ++ struct reinit_br_wh *arg; ++ ++ do_dec = 1; ++ if (atomic_inc_return(&br->br_wbr->wbr_wh_running) != 1) ++ goto out; ++ ++ /* ignore ENOMEM */ ++ arg = kmalloc(sizeof(*arg), GFP_NOFS); ++ if (arg) { ++ /* ++ * dec(wh_running), kfree(arg) and dec(br_count) ++ * in reinit function ++ */ ++ arg->sb = sb; ++ arg->br = br; ++ atomic_inc(&br->br_count); ++ wkq_err = au_wkq_nowait(reinit_br_wh, arg, sb); ++ if (unlikely(wkq_err)) { ++ atomic_dec(&br->br_wbr->wbr_wh_running); ++ atomic_dec(&br->br_count); ++ kfree(arg); ++ } ++ do_dec = 0; ++ } ++ ++ out: ++ if (do_dec) ++ atomic_dec(&br->br_wbr->wbr_wh_running); ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* ++ * create the whiteout @wh. ++ */ ++static int link_or_create_wh(struct super_block *sb, aufs_bindex_t bindex, ++ struct dentry *wh) ++{ ++ int err; ++ struct path h_path = { ++ .dentry = wh ++ }; ++ struct au_branch *br; ++ struct au_wbr *wbr; ++ struct dentry *h_parent; ++ struct inode *h_dir; ++ ++ h_parent = wh->d_parent; /* dir inode is locked */ ++ h_dir = h_parent->d_inode; ++ IMustLock(h_dir); ++ ++ br = au_sbr(sb, bindex); ++ h_path.mnt = br->br_mnt; ++ wbr = br->br_wbr; ++ wbr_wh_read_lock(wbr); ++ if (wbr->wbr_whbase) { ++ err = vfsub_link(wbr->wbr_whbase, h_dir, &h_path); ++ if (!err || err != -EMLINK) ++ goto out; ++ ++ /* link count full. re-initialize br_whbase. */ ++ kick_reinit_br_wh(sb, br); ++ } ++ ++ /* return this error in this context */ ++ err = vfsub_create(h_dir, &h_path, WH_MASK); ++ ++ out: ++ wbr_wh_read_unlock(wbr); ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* ++ * create or remove the diropq. ++ */ ++static struct dentry *do_diropq(struct dentry *dentry, aufs_bindex_t bindex, ++ unsigned int flags) ++{ ++ struct dentry *opq_dentry, *h_dentry; ++ struct super_block *sb; ++ struct au_branch *br; ++ int err; ++ ++ sb = dentry->d_sb; ++ br = au_sbr(sb, bindex); ++ h_dentry = au_h_dptr(dentry, bindex); ++ opq_dentry = au_lkup_one(&diropq_name, h_dentry, br, /*nd*/NULL); ++ if (IS_ERR(opq_dentry)) ++ goto out; ++ ++ if (au_ftest_diropq(flags, CREATE)) { ++ err = link_or_create_wh(sb, bindex, opq_dentry); ++ if (!err) { ++ au_set_dbdiropq(dentry, bindex); ++ goto out; /* success */ ++ } ++ } else { ++ struct path tmp = { ++ .dentry = opq_dentry, ++ .mnt = br->br_mnt ++ }; ++ err = do_unlink_wh(au_h_iptr(dentry->d_inode, bindex), &tmp); ++ if (!err) ++ au_set_dbdiropq(dentry, -1); ++ } ++ dput(opq_dentry); ++ opq_dentry = ERR_PTR(err); ++ ++ out: ++ return opq_dentry; ++} ++ ++struct do_diropq_args { ++ struct dentry **errp; ++ struct dentry *dentry; ++ aufs_bindex_t bindex; ++ unsigned int flags; ++}; ++ ++static void call_do_diropq(void *args) ++{ ++ struct do_diropq_args *a = args; ++ *a->errp = do_diropq(a->dentry, a->bindex, a->flags); ++} ++ ++struct dentry *au_diropq_sio(struct dentry *dentry, aufs_bindex_t bindex, ++ unsigned int flags) ++{ ++ struct dentry *diropq, *h_dentry; ++ ++ h_dentry = au_h_dptr(dentry, bindex); ++ if (!au_test_h_perm_sio(h_dentry->d_inode, MAY_EXEC | MAY_WRITE)) ++ diropq = do_diropq(dentry, bindex, flags); ++ else { ++ int wkq_err; ++ struct do_diropq_args args = { ++ .errp = &diropq, ++ .dentry = dentry, ++ .bindex = bindex, ++ .flags = flags ++ }; ++ ++ wkq_err = au_wkq_wait(call_do_diropq, &args); ++ if (unlikely(wkq_err)) ++ diropq = ERR_PTR(wkq_err); ++ } ++ ++ return diropq; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* ++ * lookup whiteout dentry. ++ * @h_parent: lower parent dentry which must exist and be locked ++ * @base_name: name of dentry which will be whiteouted ++ * returns dentry for whiteout. ++ */ ++struct dentry *au_wh_lkup(struct dentry *h_parent, struct qstr *base_name, ++ struct au_branch *br) ++{ ++ int err; ++ struct qstr wh_name; ++ struct dentry *wh_dentry; ++ ++ err = au_wh_name_alloc(&wh_name, base_name); ++ wh_dentry = ERR_PTR(err); ++ if (!err) { ++ wh_dentry = au_lkup_one(&wh_name, h_parent, br, /*nd*/NULL); ++ kfree(wh_name.name); ++ } ++ return wh_dentry; ++} ++ ++/* ++ * link/create a whiteout for @dentry on @bindex. ++ */ ++struct dentry *au_wh_create(struct dentry *dentry, aufs_bindex_t bindex, ++ struct dentry *h_parent) ++{ ++ struct dentry *wh_dentry; ++ struct super_block *sb; ++ int err; ++ ++ sb = dentry->d_sb; ++ wh_dentry = au_wh_lkup(h_parent, &dentry->d_name, au_sbr(sb, bindex)); ++ if (!IS_ERR(wh_dentry) && !wh_dentry->d_inode) { ++ err = link_or_create_wh(sb, bindex, wh_dentry); ++ if (!err) ++ au_set_dbwh(dentry, bindex); ++ else { ++ dput(wh_dentry); ++ wh_dentry = ERR_PTR(err); ++ } ++ } ++ ++ return wh_dentry; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* Delete all whiteouts in this directory on branch bindex. */ ++static int del_wh_children(struct dentry *h_dentry, struct au_nhash *whlist, ++ aufs_bindex_t bindex, struct au_branch *br) ++{ ++ int err; ++ unsigned long ul, n; ++ struct qstr wh_name; ++ char *p; ++ struct hlist_head *head; ++ struct au_vdir_wh *tpos; ++ struct hlist_node *pos; ++ struct au_vdir_destr *str; ++ ++ err = -ENOMEM; ++ p = __getname(); ++ wh_name.name = p; ++ if (unlikely(!wh_name.name)) ++ goto out; ++ ++ err = 0; ++ memcpy(p, AUFS_WH_PFX, AUFS_WH_PFX_LEN); ++ p += AUFS_WH_PFX_LEN; ++ n = whlist->nh_num; ++ head = whlist->nh_head; ++ for (ul = 0; !err && ul < n; ul++, head++) { ++ hlist_for_each_entry(tpos, pos, head, wh_hash) { ++ if (tpos->wh_bindex != bindex) ++ continue; ++ ++ str = &tpos->wh_str; ++ if (str->len + AUFS_WH_PFX_LEN <= PATH_MAX) { ++ memcpy(p, str->name, str->len); ++ wh_name.len = AUFS_WH_PFX_LEN + str->len; ++ err = unlink_wh_name(h_dentry, &wh_name, br); ++ if (!err) ++ continue; ++ break; ++ } ++ AuIOErr("whiteout name too long %.*s\n", ++ str->len, str->name); ++ err = -EIO; ++ break; ++ } ++ } ++ __putname(wh_name.name); ++ ++ out: ++ return err; ++} ++ ++struct del_wh_children_args { ++ int *errp; ++ struct dentry *h_dentry; ++ struct au_nhash whlist; ++ aufs_bindex_t bindex; ++ struct au_branch *br; ++}; ++ ++static void call_del_wh_children(void *args) ++{ ++ struct del_wh_children_args *a = args; ++ *a->errp = del_wh_children(a->h_dentry, &a->whlist, a->bindex, a->br); ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++struct au_whtmp_rmdir *au_whtmp_rmdir_alloc(struct super_block *sb, gfp_t gfp) ++{ ++ struct au_whtmp_rmdir *whtmp; ++ int err; ++ ++ SiMustAnyLock(sb); ++ ++ whtmp = kmalloc(sizeof(*whtmp), gfp); ++ if (unlikely(!whtmp)) { ++ whtmp = ERR_PTR(-ENOMEM); ++ goto out; ++ } ++ ++ whtmp->dir = NULL; ++ whtmp->wh_dentry = NULL; ++ err = au_nhash_alloc(&whtmp->whlist, au_sbi(sb)->si_rdhash, gfp); ++ if (!err) ++ return whtmp; /* success */ ++ ++ kfree(whtmp); ++ whtmp = ERR_PTR(err); ++ ++ out: ++ return whtmp; ++} ++ ++void au_whtmp_rmdir_free(struct au_whtmp_rmdir *whtmp) ++{ ++ dput(whtmp->wh_dentry); ++ iput(whtmp->dir); ++ au_nhash_wh_free(&whtmp->whlist); ++ kfree(whtmp); ++} ++ ++/* ++ * rmdir the whiteouted temporary named dir @h_dentry. ++ * @whlist: whiteouted children. ++ */ ++int au_whtmp_rmdir(struct inode *dir, aufs_bindex_t bindex, ++ struct dentry *wh_dentry, struct au_nhash *whlist) ++{ ++ int err; ++ struct path h_tmp; ++ struct inode *wh_inode, *h_dir; ++ struct au_branch *br; ++ ++ h_dir = wh_dentry->d_parent->d_inode; /* dir inode is locked */ ++ IMustLock(h_dir); ++ ++ br = au_sbr(dir->i_sb, bindex); ++ wh_inode = wh_dentry->d_inode; ++ mutex_lock_nested(&wh_inode->i_mutex, AuLsc_I_CHILD); ++ ++ /* ++ * someone else might change some whiteouts while we were sleeping. ++ * it means this whlist may have an obsoleted entry. ++ */ ++ if (!au_test_h_perm_sio(wh_inode, MAY_EXEC | MAY_WRITE)) ++ err = del_wh_children(wh_dentry, whlist, bindex, br); ++ else { ++ int wkq_err; ++ struct del_wh_children_args args = { ++ .errp = &err, ++ .h_dentry = wh_dentry, ++ .whlist = *whlist, ++ .bindex = bindex, ++ .br = br ++ }; ++ ++ wkq_err = au_wkq_wait(call_del_wh_children, &args); ++ if (unlikely(wkq_err)) ++ err = wkq_err; ++ } ++ mutex_unlock(&wh_inode->i_mutex); ++ ++ if (!err) { ++ h_tmp.dentry = wh_dentry; ++ h_tmp.mnt = br->br_mnt; ++ err = vfsub_rmdir(h_dir, &h_tmp); ++ /* d_drop(h_dentry); */ ++ } ++ ++ if (!err) { ++ if (au_ibstart(dir) == bindex) { ++ au_cpup_attr_timesizes(dir); ++ drop_nlink(dir); ++ } ++ return 0; /* success */ ++ } ++ ++ AuWarn("failed removing %.*s(%d), ignored\n", ++ AuDLNPair(wh_dentry), err); ++ return err; ++} ++ ++static void call_rmdir_whtmp(void *args) ++{ ++ int err; ++ struct au_whtmp_rmdir *a = args; ++ struct super_block *sb; ++ struct dentry *h_parent; ++ struct inode *h_dir; ++ struct au_branch *br; ++ struct au_hinode *hdir; ++ ++ /* rmdir by nfsd may cause deadlock with this i_mutex */ ++ /* mutex_lock(&a->dir->i_mutex); */ ++ sb = a->dir->i_sb; ++ si_noflush_read_lock(sb); ++ err = au_test_ro(sb, a->bindex, NULL); ++ if (unlikely(err)) ++ goto out; ++ ++ err = -EIO; ++ br = au_sbr(sb, a->bindex); ++ ii_write_lock_parent(a->dir); ++ h_parent = dget_parent(a->wh_dentry); ++ h_dir = h_parent->d_inode; ++ hdir = au_hi(a->dir, a->bindex); ++ au_hin_imtx_lock_nested(hdir, AuLsc_I_PARENT); ++ err = au_h_verify(a->wh_dentry, au_opt_udba(sb), h_dir, h_parent, br); ++ if (!err) { ++ err = mnt_want_write(br->br_mnt); ++ if (!err) { ++ err = au_whtmp_rmdir(a->dir, a->bindex, a->wh_dentry, ++ &a->whlist); ++ mnt_drop_write(br->br_mnt); ++ } ++ } ++ au_hin_imtx_unlock(hdir); ++ dput(h_parent); ++ ii_write_unlock(a->dir); ++ ++ out: ++ /* mutex_unlock(&a->dir->i_mutex); */ ++ au_nwt_done(&au_sbi(sb)->si_nowait); ++ si_read_unlock(sb); ++ au_whtmp_rmdir_free(a); ++ if (unlikely(err)) ++ AuIOErr("err %d\n", err); ++} ++ ++void au_whtmp_kick_rmdir(struct inode *dir, aufs_bindex_t bindex, ++ struct dentry *wh_dentry, struct au_whtmp_rmdir *args) ++{ ++ int wkq_err; ++ ++ IMustLock(dir); ++ ++ /* all post-process will be done in do_rmdir_whtmp(). */ ++ args->dir = au_igrab(dir); ++ args->bindex = bindex; ++ args->wh_dentry = dget(wh_dentry); ++ wkq_err = au_wkq_nowait(call_rmdir_whtmp, args, dir->i_sb); ++ if (unlikely(wkq_err)) { ++ AuWarn("rmdir error %.*s (%d), ignored\n", ++ AuDLNPair(wh_dentry), wkq_err); ++ au_whtmp_rmdir_free(args); ++ } ++} +diff -Nur linux-2.6.31.4.orig/fs/aufs/whout.h linux-2.6.31.4/fs/aufs/whout.h +--- linux-2.6.31.4.orig/fs/aufs/whout.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/whout.h 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,87 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * whiteout for logical deletion and opaque directory ++ */ ++ ++#ifndef __AUFS_WHOUT_H__ ++#define __AUFS_WHOUT_H__ ++ ++#ifdef __KERNEL__ ++ ++#include <linux/aufs_type.h> ++#include "dir.h" ++ ++/* whout.c */ ++int au_wh_name_alloc(struct qstr *wh, const struct qstr *name); ++struct au_branch; ++int au_wh_test(struct dentry *h_parent, struct qstr *wh_name, ++ struct au_branch *br, int try_sio); ++int au_diropq_test(struct dentry *h_dentry, struct au_branch *br); ++struct dentry *au_whtmp_lkup(struct dentry *h_parent, struct au_branch *br, ++ struct qstr *prefix); ++int au_whtmp_ren(struct dentry *h_dentry, struct au_branch *br); ++int au_wh_unlink_dentry(struct inode *h_dir, struct path *h_path, ++ struct dentry *dentry); ++int au_wh_init(struct dentry *h_parent, struct au_branch *br, ++ struct super_block *sb); ++ ++/* diropq flags */ ++#define AuDiropq_CREATE 1 ++#define au_ftest_diropq(flags, name) ((flags) & AuDiropq_##name) ++#define au_fset_diropq(flags, name) { (flags) |= AuDiropq_##name; } ++#define au_fclr_diropq(flags, name) { (flags) &= ~AuDiropq_##name; } ++ ++struct dentry *au_diropq_sio(struct dentry *dentry, aufs_bindex_t bindex, ++ unsigned int flags); ++struct dentry *au_wh_lkup(struct dentry *h_parent, struct qstr *base_name, ++ struct au_branch *br); ++struct dentry *au_wh_create(struct dentry *dentry, aufs_bindex_t bindex, ++ struct dentry *h_parent); ++ ++/* real rmdir for the whiteout-ed dir */ ++struct au_whtmp_rmdir { ++ struct inode *dir; ++ aufs_bindex_t bindex; ++ struct dentry *wh_dentry; ++ struct au_nhash whlist; ++}; ++ ++struct au_whtmp_rmdir *au_whtmp_rmdir_alloc(struct super_block *sb, gfp_t gfp); ++void au_whtmp_rmdir_free(struct au_whtmp_rmdir *whtmp); ++int au_whtmp_rmdir(struct inode *dir, aufs_bindex_t bindex, ++ struct dentry *wh_dentry, struct au_nhash *whlist); ++void au_whtmp_kick_rmdir(struct inode *dir, aufs_bindex_t bindex, ++ struct dentry *wh_dentry, struct au_whtmp_rmdir *args); ++ ++/* ---------------------------------------------------------------------- */ ++ ++static inline struct dentry *au_diropq_create(struct dentry *dentry, ++ aufs_bindex_t bindex) ++{ ++ return au_diropq_sio(dentry, bindex, AuDiropq_CREATE); ++} ++ ++static inline int au_diropq_remove(struct dentry *dentry, aufs_bindex_t bindex) ++{ ++ return PTR_ERR(au_diropq_sio(dentry, bindex, !AuDiropq_CREATE)); ++} ++ ++#endif /* __KERNEL__ */ ++#endif /* __AUFS_WHOUT_H__ */ +diff -Nur linux-2.6.31.4.orig/fs/aufs/wkq.c linux-2.6.31.4/fs/aufs/wkq.c +--- linux-2.6.31.4.orig/fs/aufs/wkq.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/wkq.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,259 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * workqueue for asynchronous/super-io operations ++ * todo: try new dredential scheme ++ */ ++ ++#include <linux/module.h> ++#include "aufs.h" ++ ++/* internal workqueue named AUFS_WKQ_NAME */ ++static struct au_wkq { ++ struct workqueue_struct *q; ++ ++ /* balancing */ ++ atomic_t busy; ++} *au_wkq; ++ ++struct au_wkinfo { ++ struct work_struct wk; ++ struct super_block *sb; ++ ++ unsigned int flags; /* see wkq.h */ ++ ++ au_wkq_func_t func; ++ void *args; ++ ++ atomic_t *busyp; ++ struct completion *comp; ++}; ++ ++/* ---------------------------------------------------------------------- */ ++ ++static int enqueue(struct au_wkq *wkq, struct au_wkinfo *wkinfo) ++{ ++ wkinfo->busyp = &wkq->busy; ++ if (au_ftest_wkq(wkinfo->flags, WAIT)) ++ return !queue_work(wkq->q, &wkinfo->wk); ++ else ++ return !schedule_work(&wkinfo->wk); ++} ++ ++static void do_wkq(struct au_wkinfo *wkinfo) ++{ ++ unsigned int idle, n; ++ int i, idle_idx; ++ ++ while (1) { ++ if (au_ftest_wkq(wkinfo->flags, WAIT)) { ++ idle_idx = 0; ++ idle = UINT_MAX; ++ for (i = 0; i < aufs_nwkq; i++) { ++ n = atomic_inc_return(&au_wkq[i].busy); ++ if (n == 1 && !enqueue(au_wkq + i, wkinfo)) ++ return; /* success */ ++ ++ if (n < idle) { ++ idle_idx = i; ++ idle = n; ++ } ++ atomic_dec(&au_wkq[i].busy); ++ } ++ } else ++ idle_idx = aufs_nwkq; ++ ++ atomic_inc(&au_wkq[idle_idx].busy); ++ if (!enqueue(au_wkq + idle_idx, wkinfo)) ++ return; /* success */ ++ ++ /* impossible? */ ++ AuWarn1("failed to queue_work()\n"); ++ yield(); ++ } ++} ++ ++static void wkq_func(struct work_struct *wk) ++{ ++ struct au_wkinfo *wkinfo = container_of(wk, struct au_wkinfo, wk); ++ ++ wkinfo->func(wkinfo->args); ++ atomic_dec_return(wkinfo->busyp); ++ if (au_ftest_wkq(wkinfo->flags, WAIT)) ++ complete(wkinfo->comp); ++ else { ++ kobject_put(&au_sbi(wkinfo->sb)->si_kobj); ++ module_put(THIS_MODULE); ++ kfree(wkinfo); ++ } ++} ++ ++/* ++ * Since struct completion is large, try allocating it dynamically. ++ */ ++#if defined(CONFIG_4KSTACKS) || defined(AuTest4KSTACKS) ++#define AuWkqCompDeclare(name) struct completion *comp = NULL ++ ++static int au_wkq_comp_alloc(struct au_wkinfo *wkinfo, struct completion **comp) ++{ ++ *comp = kmalloc(sizeof(**comp), GFP_NOFS); ++ if (*comp) { ++ init_completion(*comp); ++ wkinfo->comp = *comp; ++ return 0; ++ } ++ return -ENOMEM; ++} ++ ++static void au_wkq_comp_free(struct completion *comp) ++{ ++ kfree(comp); ++} ++ ++#else ++ ++/* no braces */ ++#define AuWkqCompDeclare(name) \ ++ DECLARE_COMPLETION_ONSTACK(_ ## name); \ ++ struct completion *comp = &_ ## name ++ ++static int au_wkq_comp_alloc(struct au_wkinfo *wkinfo, struct completion **comp) ++{ ++ wkinfo->comp = *comp; ++ return 0; ++} ++ ++static void au_wkq_comp_free(struct completion *comp __maybe_unused) ++{ ++ /* empty */ ++} ++#endif /* 4KSTACKS */ ++ ++static void au_wkq_run(struct au_wkinfo *wkinfo) ++{ ++ au_dbg_verify_kthread(); ++ INIT_WORK(&wkinfo->wk, wkq_func); ++ do_wkq(wkinfo); ++} ++ ++int au_wkq_wait(au_wkq_func_t func, void *args) ++{ ++ int err; ++ AuWkqCompDeclare(comp); ++ struct au_wkinfo wkinfo = { ++ .flags = AuWkq_WAIT, ++ .func = func, ++ .args = args ++ }; ++ ++ err = au_wkq_comp_alloc(&wkinfo, &comp); ++ if (!err) { ++ au_wkq_run(&wkinfo); ++ /* no timeout, no interrupt */ ++ wait_for_completion(wkinfo.comp); ++ au_wkq_comp_free(comp); ++ } ++ ++ return err; ++ ++} ++ ++int au_wkq_nowait(au_wkq_func_t func, void *args, struct super_block *sb) ++{ ++ int err; ++ struct au_wkinfo *wkinfo; ++ ++ atomic_inc(&au_sbi(sb)->si_nowait.nw_len); ++ ++ /* ++ * wkq_func() must free this wkinfo. ++ * it highly depends upon the implementation of workqueue. ++ */ ++ err = 0; ++ wkinfo = kmalloc(sizeof(*wkinfo), GFP_NOFS); ++ if (wkinfo) { ++ wkinfo->sb = sb; ++ wkinfo->flags = !AuWkq_WAIT; ++ wkinfo->func = func; ++ wkinfo->args = args; ++ wkinfo->comp = NULL; ++ kobject_get(&au_sbi(sb)->si_kobj); ++ __module_get(THIS_MODULE); ++ ++ au_wkq_run(wkinfo); ++ } else { ++ err = -ENOMEM; ++ atomic_dec(&au_sbi(sb)->si_nowait.nw_len); ++ } ++ ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++void au_nwt_init(struct au_nowait_tasks *nwt) ++{ ++ atomic_set(&nwt->nw_len, 0); ++ /* smp_mb();*/ /* atomic_set */ ++ init_waitqueue_head(&nwt->nw_wq); ++} ++ ++void au_wkq_fin(void) ++{ ++ int i; ++ ++ for (i = 0; i < aufs_nwkq; i++) ++ if (au_wkq[i].q && !IS_ERR(au_wkq[i].q)) ++ destroy_workqueue(au_wkq[i].q); ++ kfree(au_wkq); ++} ++ ++int __init au_wkq_init(void) ++{ ++ int err, i; ++ struct au_wkq *nowaitq; ++ ++ /* '+1' is for accounting of nowait queue */ ++ err = -ENOMEM; ++ au_wkq = kcalloc(aufs_nwkq + 1, sizeof(*au_wkq), GFP_NOFS); ++ if (unlikely(!au_wkq)) ++ goto out; ++ ++ err = 0; ++ for (i = 0; i < aufs_nwkq; i++) { ++ au_wkq[i].q = create_singlethread_workqueue(AUFS_WKQ_NAME); ++ if (au_wkq[i].q && !IS_ERR(au_wkq[i].q)) { ++ atomic_set(&au_wkq[i].busy, 0); ++ continue; ++ } ++ ++ err = PTR_ERR(au_wkq[i].q); ++ au_wkq_fin(); ++ goto out; ++ } ++ ++ /* nowait accounting */ ++ nowaitq = au_wkq + aufs_nwkq; ++ atomic_set(&nowaitq->busy, 0); ++ nowaitq->q = NULL; ++ /* smp_mb(); */ /* atomic_set */ ++ ++ out: ++ return err; ++} +diff -Nur linux-2.6.31.4.orig/fs/aufs/wkq.h linux-2.6.31.4/fs/aufs/wkq.h +--- linux-2.6.31.4.orig/fs/aufs/wkq.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/wkq.h 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,82 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * workqueue for asynchronous/super-io operations ++ * todo: try new credentials management scheme ++ */ ++ ++#ifndef __AUFS_WKQ_H__ ++#define __AUFS_WKQ_H__ ++ ++#ifdef __KERNEL__ ++ ++#include <linux/sched.h> ++#include <linux/wait.h> ++#include <linux/aufs_type.h> ++ ++struct super_block; ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* ++ * in the next operation, wait for the 'nowait' tasks in system-wide workqueue ++ */ ++struct au_nowait_tasks { ++ atomic_t nw_len; ++ wait_queue_head_t nw_wq; ++}; ++ ++/* ---------------------------------------------------------------------- */ ++ ++typedef void (*au_wkq_func_t)(void *args); ++ ++/* wkq flags */ ++#define AuWkq_WAIT 1 ++#define au_ftest_wkq(flags, name) ((flags) & AuWkq_##name) ++#define au_fset_wkq(flags, name) { (flags) |= AuWkq_##name; } ++#define au_fclr_wkq(flags, name) { (flags) &= ~AuWkq_##name; } ++ ++/* wkq.c */ ++int au_wkq_wait(au_wkq_func_t func, void *args); ++int au_wkq_nowait(au_wkq_func_t func, void *args, struct super_block *sb); ++void au_nwt_init(struct au_nowait_tasks *nwt); ++int __init au_wkq_init(void); ++void au_wkq_fin(void); ++ ++/* ---------------------------------------------------------------------- */ ++ ++static inline int au_test_wkq(struct task_struct *tsk) ++{ ++ return !tsk->mm && !strcmp(tsk->comm, AUFS_WKQ_NAME); ++} ++ ++static inline void au_nwt_done(struct au_nowait_tasks *nwt) ++{ ++ if (!atomic_dec_return(&nwt->nw_len)) ++ wake_up_all(&nwt->nw_wq); ++} ++ ++static inline int au_nwt_flush(struct au_nowait_tasks *nwt) ++{ ++ wait_event(nwt->nw_wq, !atomic_read(&nwt->nw_len)); ++ return 0; ++} ++ ++#endif /* __KERNEL__ */ ++#endif /* __AUFS_WKQ_H__ */ +diff -Nur linux-2.6.31.4.orig/fs/aufs/xino.c linux-2.6.31.4/fs/aufs/xino.c +--- linux-2.6.31.4.orig/fs/aufs/xino.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/fs/aufs/xino.c 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,1200 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++/* ++ * external inode number translation table and bitmap ++ */ ++ ++#include <linux/file.h> ++#include <linux/seq_file.h> ++#include <linux/uaccess.h> ++#include "aufs.h" ++ ++ssize_t xino_fread(au_readf_t func, struct file *file, void *buf, size_t size, ++ loff_t *pos) ++{ ++ ssize_t err; ++ mm_segment_t oldfs; ++ ++ oldfs = get_fs(); ++ set_fs(KERNEL_DS); ++ do { ++ /* todo: signal_pending? */ ++ err = func(file, (char __user *)buf, size, pos); ++ } while (err == -EAGAIN || err == -EINTR); ++ set_fs(oldfs); ++ ++#if 0 /* reserved for future use */ ++ if (err > 0) ++ fsnotify_access(file->f_dentry); ++#endif ++ ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static ssize_t do_xino_fwrite(au_writef_t func, struct file *file, void *buf, ++ size_t size, loff_t *pos) ++{ ++ ssize_t err; ++ mm_segment_t oldfs; ++ ++ oldfs = get_fs(); ++ set_fs(KERNEL_DS); ++ lockdep_off(); ++ do { ++ /* todo: signal_pending? */ ++ err = func(file, (const char __user *)buf, size, pos); ++ } while (err == -EAGAIN || err == -EINTR); ++ lockdep_on(); ++ set_fs(oldfs); ++ ++#if 0 /* reserved for future use */ ++ if (err > 0) ++ fsnotify_modify(file->f_dentry); ++#endif ++ ++ return err; ++} ++ ++struct do_xino_fwrite_args { ++ ssize_t *errp; ++ au_writef_t func; ++ struct file *file; ++ void *buf; ++ size_t size; ++ loff_t *pos; ++}; ++ ++static void call_do_xino_fwrite(void *args) ++{ ++ struct do_xino_fwrite_args *a = args; ++ *a->errp = do_xino_fwrite(a->func, a->file, a->buf, a->size, a->pos); ++} ++ ++ssize_t xino_fwrite(au_writef_t func, struct file *file, void *buf, size_t size, ++ loff_t *pos) ++{ ++ ssize_t err; ++ ++ /* todo: signal block and no wkq? */ ++ /* todo: new credential scheme */ ++ /* ++ * it breaks RLIMIT_FSIZE and normal user's limit, ++ * users should care about quota and real 'filesystem full.' ++ */ ++ if (!au_test_wkq(current)) { ++ int wkq_err; ++ struct do_xino_fwrite_args args = { ++ .errp = &err, ++ .func = func, ++ .file = file, ++ .buf = buf, ++ .size = size, ++ .pos = pos ++ }; ++ ++ wkq_err = au_wkq_wait(call_do_xino_fwrite, &args); ++ if (unlikely(wkq_err)) ++ err = wkq_err; ++ } else ++ err = do_xino_fwrite(func, file, buf, size, pos); ++ ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* ++ * create a new xinofile at the same place/path as @base_file. ++ */ ++struct file *au_xino_create2(struct file *base_file, struct file *copy_src) ++{ ++ struct file *file; ++ struct dentry *base, *dentry, *parent; ++ struct inode *dir; ++ struct qstr *name; ++ int err; ++ ++ base = base_file->f_dentry; ++ parent = base->d_parent; /* dir inode is locked */ ++ dir = parent->d_inode; ++ IMustLock(dir); ++ ++ file = ERR_PTR(-EINVAL); ++ name = &base->d_name; ++ dentry = vfsub_lookup_one_len(name->name, parent, name->len); ++ if (IS_ERR(dentry)) { ++ file = (void *)dentry; ++ AuErr("%.*s lookup err %ld\n", AuLNPair(name), PTR_ERR(dentry)); ++ goto out; ++ } ++ ++ /* no need to mnt_want_write() since we call dentry_open() later */ ++ err = vfs_create(dir, dentry, S_IRUGO | S_IWUGO, NULL); ++ if (unlikely(err)) { ++ file = ERR_PTR(err); ++ AuErr("%.*s create err %d\n", AuLNPair(name), err); ++ goto out_dput; ++ } ++ ++ file = dentry_open(dget(dentry), mntget(base_file->f_vfsmnt), ++ O_RDWR | O_CREAT | O_EXCL | O_LARGEFILE, ++ current_cred()); ++ if (IS_ERR(file)) { ++ AuErr("%.*s open err %ld\n", AuLNPair(name), PTR_ERR(file)); ++ goto out_dput; ++ } ++ ++ err = vfsub_unlink(dir, &file->f_path, /*force*/0); ++ if (unlikely(err)) { ++ AuErr("%.*s unlink err %d\n", AuLNPair(name), err); ++ goto out_fput; ++ } ++ ++ if (copy_src) { ++ /* no one can touch copy_src xino */ ++ err = au_copy_file(file, copy_src, ++ i_size_read(copy_src->f_dentry->d_inode)); ++ if (unlikely(err)) { ++ AuErr("%.*s copy err %d\n", AuLNPair(name), err); ++ goto out_fput; ++ } ++ } ++ goto out_dput; /* success */ ++ ++ out_fput: ++ fput(file); ++ file = ERR_PTR(err); ++ out_dput: ++ dput(dentry); ++ out: ++ return file; ++} ++ ++struct au_xino_lock_dir { ++ struct au_hinode *hdir; ++ struct dentry *parent; ++ struct mutex *mtx; ++}; ++ ++static void au_xino_lock_dir(struct super_block *sb, struct file *xino, ++ struct au_xino_lock_dir *ldir) ++{ ++ aufs_bindex_t brid, bindex; ++ ++ ldir->hdir = NULL; ++ bindex = -1; ++ brid = au_xino_brid(sb); ++ if (brid >= 0) ++ bindex = au_br_index(sb, brid); ++ if (bindex >= 0) { ++ ldir->hdir = au_hi(sb->s_root->d_inode, bindex); ++ au_hin_imtx_lock_nested(ldir->hdir, AuLsc_I_PARENT); ++ } else { ++ ldir->parent = dget_parent(xino->f_dentry); ++ ldir->mtx = &ldir->parent->d_inode->i_mutex; ++ mutex_lock_nested(ldir->mtx, AuLsc_I_PARENT); ++ } ++} ++ ++static void au_xino_unlock_dir(struct au_xino_lock_dir *ldir) ++{ ++ if (ldir->hdir) ++ au_hin_imtx_unlock(ldir->hdir); ++ else { ++ mutex_unlock(ldir->mtx); ++ dput(ldir->parent); ++ } ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* trucate xino files asynchronously */ ++ ++int au_xino_trunc(struct super_block *sb, aufs_bindex_t bindex) ++{ ++ int err; ++ aufs_bindex_t bi, bend; ++ struct au_branch *br; ++ struct file *new_xino, *file; ++ struct super_block *h_sb; ++ struct au_xino_lock_dir ldir; ++ ++ err = -EINVAL; ++ bend = au_sbend(sb); ++ if (unlikely(bindex < 0 || bend < bindex)) ++ goto out; ++ br = au_sbr(sb, bindex); ++ file = br->br_xino.xi_file; ++ if (!file) ++ goto out; ++ ++ au_xino_lock_dir(sb, file, &ldir); ++ /* mnt_want_write() is unnecessary here */ ++ new_xino = au_xino_create2(file, file); ++ au_xino_unlock_dir(&ldir); ++ err = PTR_ERR(new_xino); ++ if (IS_ERR(new_xino)) ++ goto out; ++ err = 0; ++ fput(file); ++ br->br_xino.xi_file = new_xino; ++ ++ h_sb = br->br_mnt->mnt_sb; ++ for (bi = 0; bi <= bend; bi++) { ++ if (unlikely(bi == bindex)) ++ continue; ++ br = au_sbr(sb, bi); ++ if (br->br_mnt->mnt_sb != h_sb) ++ continue; ++ ++ fput(br->br_xino.xi_file); ++ br->br_xino.xi_file = new_xino; ++ get_file(new_xino); ++ } ++ ++ out: ++ return err; ++} ++ ++struct xino_do_trunc_args { ++ struct super_block *sb; ++ struct au_branch *br; ++}; ++ ++static void xino_do_trunc(void *_args) ++{ ++ struct xino_do_trunc_args *args = _args; ++ struct super_block *sb; ++ struct au_branch *br; ++ struct inode *dir; ++ int err; ++ aufs_bindex_t bindex; ++ ++ err = 0; ++ sb = args->sb; ++ dir = sb->s_root->d_inode; ++ br = args->br; ++ ++ si_noflush_write_lock(sb); ++ ii_read_lock_parent(dir); ++ bindex = au_br_index(sb, br->br_id); ++ err = au_xino_trunc(sb, bindex); ++ if (!err ++ && br->br_xino.xi_file->f_dentry->d_inode->i_blocks ++ >= br->br_xino_upper) ++ br->br_xino_upper += AUFS_XINO_TRUNC_STEP; ++ ++ ii_read_unlock(dir); ++ if (unlikely(err)) ++ AuWarn("err b%d, (%d)\n", bindex, err); ++ atomic_dec(&br->br_xino_running); ++ atomic_dec(&br->br_count); ++ au_nwt_done(&au_sbi(sb)->si_nowait); ++ si_write_unlock(sb); ++ kfree(args); ++} ++ ++static void xino_try_trunc(struct super_block *sb, struct au_branch *br) ++{ ++ struct xino_do_trunc_args *args; ++ int wkq_err; ++ ++ if (br->br_xino.xi_file->f_dentry->d_inode->i_blocks ++ < br->br_xino_upper) ++ return; ++ ++ if (atomic_inc_return(&br->br_xino_running) > 1) ++ goto out; ++ ++ /* lock and kfree() will be called in trunc_xino() */ ++ args = kmalloc(sizeof(*args), GFP_NOFS); ++ if (unlikely(!args)) { ++ AuErr1("no memory\n"); ++ goto out_args; ++ } ++ ++ atomic_inc_return(&br->br_count); ++ args->sb = sb; ++ args->br = br; ++ wkq_err = au_wkq_nowait(xino_do_trunc, args, sb); ++ if (!wkq_err) ++ return; /* success */ ++ ++ AuErr("wkq %d\n", wkq_err); ++ atomic_dec_return(&br->br_count); ++ ++ out_args: ++ kfree(args); ++ out: ++ atomic_dec_return(&br->br_xino_running); ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++static int au_xino_do_write(au_writef_t write, struct file *file, ++ ino_t h_ino, ino_t ino) ++{ ++ loff_t pos; ++ ssize_t sz; ++ ++ pos = h_ino; ++ if (unlikely(au_loff_max / sizeof(ino) - 1 < pos)) { ++ AuIOErr1("too large hi%lu\n", (unsigned long)h_ino); ++ return -EFBIG; ++ } ++ pos *= sizeof(ino); ++ sz = xino_fwrite(write, file, &ino, sizeof(ino), &pos); ++ if (sz == sizeof(ino)) ++ return 0; /* success */ ++ ++ AuIOErr("write failed (%zd)\n", sz); ++ return -EIO; ++} ++ ++/* ++ * write @ino to the xinofile for the specified branch{@sb, @bindex} ++ * at the position of @h_ino. ++ * even if @ino is zero, it is written to the xinofile and means no entry. ++ * if the size of the xino file on a specific filesystem exceeds the watermark, ++ * try truncating it. ++ */ ++int au_xino_write(struct super_block *sb, aufs_bindex_t bindex, ino_t h_ino, ++ ino_t ino) ++{ ++ int err; ++ unsigned int mnt_flags; ++ struct au_branch *br; ++ ++ BUILD_BUG_ON(sizeof(long long) != sizeof(au_loff_max) ++ || ((loff_t)-1) > 0); ++ SiMustAnyLock(sb); ++ ++ mnt_flags = au_mntflags(sb); ++ if (!au_opt_test(mnt_flags, XINO)) ++ return 0; ++ ++ br = au_sbr(sb, bindex); ++ err = au_xino_do_write(au_sbi(sb)->si_xwrite, br->br_xino.xi_file, ++ h_ino, ino); ++ if (!err) { ++ if (au_opt_test(mnt_flags, TRUNC_XINO) ++ && au_test_fs_trunc_xino(br->br_mnt->mnt_sb)) ++ xino_try_trunc(sb, br); ++ return 0; /* success */ ++ } ++ ++ AuIOErr("write failed (%d)\n", err); ++ return -EIO; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* aufs inode number bitmap */ ++ ++static const int page_bits = (int)PAGE_SIZE * BITS_PER_BYTE; ++static ino_t xib_calc_ino(unsigned long pindex, int bit) ++{ ++ ino_t ino; ++ ++ AuDebugOn(bit < 0 || page_bits <= bit); ++ ino = AUFS_FIRST_INO + pindex * page_bits + bit; ++ return ino; ++} ++ ++static void xib_calc_bit(ino_t ino, unsigned long *pindex, int *bit) ++{ ++ AuDebugOn(ino < AUFS_FIRST_INO); ++ ino -= AUFS_FIRST_INO; ++ *pindex = ino / page_bits; ++ *bit = ino % page_bits; ++} ++ ++static int xib_pindex(struct super_block *sb, unsigned long pindex) ++{ ++ int err; ++ loff_t pos; ++ ssize_t sz; ++ struct au_sbinfo *sbinfo; ++ struct file *xib; ++ unsigned long *p; ++ ++ sbinfo = au_sbi(sb); ++ MtxMustLock(&sbinfo->si_xib_mtx); ++ AuDebugOn(pindex > ULONG_MAX / PAGE_SIZE ++ || !au_opt_test(sbinfo->si_mntflags, XINO)); ++ ++ if (pindex == sbinfo->si_xib_last_pindex) ++ return 0; ++ ++ xib = sbinfo->si_xib; ++ p = sbinfo->si_xib_buf; ++ pos = sbinfo->si_xib_last_pindex; ++ pos *= PAGE_SIZE; ++ sz = xino_fwrite(sbinfo->si_xwrite, xib, p, PAGE_SIZE, &pos); ++ if (unlikely(sz != PAGE_SIZE)) ++ goto out; ++ ++ pos = pindex; ++ pos *= PAGE_SIZE; ++ if (i_size_read(xib->f_dentry->d_inode) >= pos + PAGE_SIZE) ++ sz = xino_fread(sbinfo->si_xread, xib, p, PAGE_SIZE, &pos); ++ else { ++ memset(p, 0, PAGE_SIZE); ++ sz = xino_fwrite(sbinfo->si_xwrite, xib, p, PAGE_SIZE, &pos); ++ } ++ if (sz == PAGE_SIZE) { ++ sbinfo->si_xib_last_pindex = pindex; ++ return 0; /* success */ ++ } ++ ++ out: ++ AuIOErr1("write failed (%zd)\n", sz); ++ err = sz; ++ if (sz >= 0) ++ err = -EIO; ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++int au_xino_write0(struct super_block *sb, aufs_bindex_t bindex, ino_t h_ino, ++ ino_t ino) ++{ ++ int err, bit; ++ unsigned long pindex; ++ struct au_sbinfo *sbinfo; ++ ++ if (!au_opt_test(au_mntflags(sb), XINO)) ++ return 0; ++ ++ err = 0; ++ if (ino) { ++ sbinfo = au_sbi(sb); ++ xib_calc_bit(ino, &pindex, &bit); ++ AuDebugOn(page_bits <= bit); ++ mutex_lock(&sbinfo->si_xib_mtx); ++ err = xib_pindex(sb, pindex); ++ if (!err) { ++ clear_bit(bit, sbinfo->si_xib_buf); ++ sbinfo->si_xib_next_bit = bit; ++ } ++ mutex_unlock(&sbinfo->si_xib_mtx); ++ } ++ ++ if (!err) ++ err = au_xino_write(sb, bindex, h_ino, 0); ++ return err; ++} ++ ++/* get an unused inode number from bitmap */ ++ino_t au_xino_new_ino(struct super_block *sb) ++{ ++ ino_t ino; ++ unsigned long *p, pindex, ul, pend; ++ struct au_sbinfo *sbinfo; ++ struct file *file; ++ int free_bit, err; ++ ++ if (!au_opt_test(au_mntflags(sb), XINO)) ++ return iunique(sb, AUFS_FIRST_INO); ++ ++ sbinfo = au_sbi(sb); ++ mutex_lock(&sbinfo->si_xib_mtx); ++ p = sbinfo->si_xib_buf; ++ free_bit = sbinfo->si_xib_next_bit; ++ if (free_bit < page_bits && !test_bit(free_bit, p)) ++ goto out; /* success */ ++ free_bit = find_first_zero_bit(p, page_bits); ++ if (free_bit < page_bits) ++ goto out; /* success */ ++ ++ pindex = sbinfo->si_xib_last_pindex; ++ for (ul = pindex - 1; ul < ULONG_MAX; ul--) { ++ err = xib_pindex(sb, ul); ++ if (unlikely(err)) ++ goto out_err; ++ free_bit = find_first_zero_bit(p, page_bits); ++ if (free_bit < page_bits) ++ goto out; /* success */ ++ } ++ ++ file = sbinfo->si_xib; ++ pend = i_size_read(file->f_dentry->d_inode) / PAGE_SIZE; ++ for (ul = pindex + 1; ul <= pend; ul++) { ++ err = xib_pindex(sb, ul); ++ if (unlikely(err)) ++ goto out_err; ++ free_bit = find_first_zero_bit(p, page_bits); ++ if (free_bit < page_bits) ++ goto out; /* success */ ++ } ++ BUG(); ++ ++ out: ++ set_bit(free_bit, p); ++ sbinfo->si_xib_next_bit++; ++ pindex = sbinfo->si_xib_last_pindex; ++ mutex_unlock(&sbinfo->si_xib_mtx); ++ ino = xib_calc_ino(pindex, free_bit); ++ AuDbg("i%lu\n", (unsigned long)ino); ++ return ino; ++ out_err: ++ mutex_unlock(&sbinfo->si_xib_mtx); ++ AuDbg("i0\n"); ++ return 0; ++} ++ ++/* ++ * read @ino from xinofile for the specified branch{@sb, @bindex} ++ * at the position of @h_ino. ++ * if @ino does not exist and @do_new is true, get new one. ++ */ ++int au_xino_read(struct super_block *sb, aufs_bindex_t bindex, ino_t h_ino, ++ ino_t *ino) ++{ ++ int err; ++ ssize_t sz; ++ loff_t pos; ++ struct file *file; ++ struct au_sbinfo *sbinfo; ++ ++ *ino = 0; ++ if (!au_opt_test(au_mntflags(sb), XINO)) ++ return 0; /* no xino */ ++ ++ err = 0; ++ sbinfo = au_sbi(sb); ++ pos = h_ino; ++ if (unlikely(au_loff_max / sizeof(*ino) - 1 < pos)) { ++ AuIOErr1("too large hi%lu\n", (unsigned long)h_ino); ++ return -EFBIG; ++ } ++ pos *= sizeof(*ino); ++ ++ file = au_sbr(sb, bindex)->br_xino.xi_file; ++ if (i_size_read(file->f_dentry->d_inode) < pos + sizeof(*ino)) ++ return 0; /* no ino */ ++ ++ sz = xino_fread(sbinfo->si_xread, file, ino, sizeof(*ino), &pos); ++ if (sz == sizeof(*ino)) ++ return 0; /* success */ ++ ++ err = sz; ++ if (unlikely(sz >= 0)) { ++ err = -EIO; ++ AuIOErr("xino read error (%zd)\n", sz); ++ } ++ ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* create and set a new xino file */ ++ ++struct file *au_xino_create(struct super_block *sb, char *fname, int silent) ++{ ++ struct file *file; ++ struct dentry *h_parent, *d; ++ struct inode *h_dir; ++ int err; ++ ++ /* ++ * at mount-time, and the xino file is the default path, ++ * hinotify is disabled so we have no inotify events to ignore. ++ * when a user specified the xino, we cannot get au_hdir to be ignored. ++ */ ++ file = vfsub_filp_open(fname, O_RDWR | O_CREAT | O_EXCL | O_LARGEFILE, ++ S_IRUGO | S_IWUGO); ++ if (IS_ERR(file)) { ++ if (!silent) ++ AuErr("open %s(%ld)\n", fname, PTR_ERR(file)); ++ return file; ++ } ++ ++ /* keep file count */ ++ h_parent = dget_parent(file->f_dentry); ++ h_dir = h_parent->d_inode; ++ mutex_lock_nested(&h_dir->i_mutex, AuLsc_I_PARENT); ++ /* mnt_want_write() is unnecessary here */ ++ err = vfsub_unlink(h_dir, &file->f_path, /*force*/0); ++ mutex_unlock(&h_dir->i_mutex); ++ dput(h_parent); ++ if (unlikely(err)) { ++ if (!silent) ++ AuErr("unlink %s(%d)\n", fname, err); ++ goto out; ++ } ++ ++ err = -EINVAL; ++ d = file->f_dentry; ++ if (unlikely(sb == d->d_sb)) { ++ if (!silent) ++ AuErr("%s must be outside\n", fname); ++ goto out; ++ } ++ if (unlikely(au_test_fs_bad_xino(d->d_sb))) { ++ if (!silent) ++ AuErr("xino doesn't support %s(%s)\n", ++ fname, au_sbtype(d->d_sb)); ++ goto out; ++ } ++ return file; /* success */ ++ ++ out: ++ fput(file); ++ file = ERR_PTR(err); ++ return file; ++} ++ ++/* ++ * find another branch who is on the same filesystem of the specified ++ * branch{@btgt}. search until @bend. ++ */ ++static int is_sb_shared(struct super_block *sb, aufs_bindex_t btgt, ++ aufs_bindex_t bend) ++{ ++ aufs_bindex_t bindex; ++ struct super_block *tgt_sb = au_sbr_sb(sb, btgt); ++ ++ for (bindex = 0; bindex < btgt; bindex++) ++ if (unlikely(tgt_sb == au_sbr_sb(sb, bindex))) ++ return bindex; ++ for (bindex++; bindex <= bend; bindex++) ++ if (unlikely(tgt_sb == au_sbr_sb(sb, bindex))) ++ return bindex; ++ return -1; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* ++ * initialize the xinofile for the specified branch @br ++ * at the place/path where @base_file indicates. ++ * test whether another branch is on the same filesystem or not, ++ * if @do_test is true. ++ */ ++int au_xino_br(struct super_block *sb, struct au_branch *br, ino_t h_ino, ++ struct file *base_file, int do_test) ++{ ++ int err; ++ ino_t ino; ++ aufs_bindex_t bend, bindex; ++ struct au_branch *shared_br, *b; ++ struct file *file; ++ struct super_block *tgt_sb; ++ ++ shared_br = NULL; ++ bend = au_sbend(sb); ++ if (do_test) { ++ tgt_sb = br->br_mnt->mnt_sb; ++ for (bindex = 0; bindex <= bend; bindex++) { ++ b = au_sbr(sb, bindex); ++ if (tgt_sb == b->br_mnt->mnt_sb) { ++ shared_br = b; ++ break; ++ } ++ } ++ } ++ ++ if (!shared_br || !shared_br->br_xino.xi_file) { ++ struct au_xino_lock_dir ldir; ++ ++ au_xino_lock_dir(sb, base_file, &ldir); ++ /* mnt_want_write() is unnecessary here */ ++ file = au_xino_create2(base_file, NULL); ++ au_xino_unlock_dir(&ldir); ++ err = PTR_ERR(file); ++ if (IS_ERR(file)) ++ goto out; ++ br->br_xino.xi_file = file; ++ } else { ++ br->br_xino.xi_file = shared_br->br_xino.xi_file; ++ get_file(br->br_xino.xi_file); ++ } ++ ++ ino = AUFS_ROOT_INO; ++ err = au_xino_do_write(au_sbi(sb)->si_xwrite, br->br_xino.xi_file, ++ h_ino, ino); ++ if (!err) ++ return 0; /* success */ ++ ++ ++ out: ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* trucate a xino bitmap file */ ++ ++/* todo: slow */ ++static int do_xib_restore(struct super_block *sb, struct file *file, void *page) ++{ ++ int err, bit; ++ ssize_t sz; ++ unsigned long pindex; ++ loff_t pos, pend; ++ struct au_sbinfo *sbinfo; ++ au_readf_t func; ++ ino_t *ino; ++ unsigned long *p; ++ ++ err = 0; ++ sbinfo = au_sbi(sb); ++ MtxMustLock(&sbinfo->si_xib_mtx); ++ p = sbinfo->si_xib_buf; ++ func = sbinfo->si_xread; ++ pend = i_size_read(file->f_dentry->d_inode); ++ pos = 0; ++ while (pos < pend) { ++ sz = xino_fread(func, file, page, PAGE_SIZE, &pos); ++ err = sz; ++ if (unlikely(sz <= 0)) ++ goto out; ++ ++ err = 0; ++ for (ino = page; sz > 0; ino++, sz -= sizeof(ino)) { ++ if (unlikely(*ino < AUFS_FIRST_INO)) ++ continue; ++ ++ xib_calc_bit(*ino, &pindex, &bit); ++ AuDebugOn(page_bits <= bit); ++ err = xib_pindex(sb, pindex); ++ if (!err) ++ set_bit(bit, p); ++ else ++ goto out; ++ } ++ } ++ ++ out: ++ return err; ++} ++ ++static int xib_restore(struct super_block *sb) ++{ ++ int err; ++ aufs_bindex_t bindex, bend; ++ void *page; ++ ++ err = -ENOMEM; ++ page = (void *)__get_free_page(GFP_NOFS); ++ if (unlikely(!page)) ++ goto out; ++ ++ err = 0; ++ bend = au_sbend(sb); ++ for (bindex = 0; !err && bindex <= bend; bindex++) ++ if (!bindex || is_sb_shared(sb, bindex, bindex - 1) < 0) ++ err = do_xib_restore ++ (sb, au_sbr(sb, bindex)->br_xino.xi_file, page); ++ else ++ AuDbg("b%d\n", bindex); ++ free_page((unsigned long)page); ++ ++ out: ++ return err; ++} ++ ++int au_xib_trunc(struct super_block *sb) ++{ ++ int err; ++ ssize_t sz; ++ loff_t pos; ++ struct au_xino_lock_dir ldir; ++ struct au_sbinfo *sbinfo; ++ unsigned long *p; ++ struct file *file; ++ ++ SiMustWriteLock(sb); ++ ++ err = 0; ++ sbinfo = au_sbi(sb); ++ if (!au_opt_test(sbinfo->si_mntflags, XINO)) ++ goto out; ++ ++ file = sbinfo->si_xib; ++ if (i_size_read(file->f_dentry->d_inode) <= PAGE_SIZE) ++ goto out; ++ ++ au_xino_lock_dir(sb, file, &ldir); ++ /* mnt_want_write() is unnecessary here */ ++ file = au_xino_create2(sbinfo->si_xib, NULL); ++ au_xino_unlock_dir(&ldir); ++ err = PTR_ERR(file); ++ if (IS_ERR(file)) ++ goto out; ++ fput(sbinfo->si_xib); ++ sbinfo->si_xib = file; ++ ++ p = sbinfo->si_xib_buf; ++ memset(p, 0, PAGE_SIZE); ++ pos = 0; ++ sz = xino_fwrite(sbinfo->si_xwrite, sbinfo->si_xib, p, PAGE_SIZE, &pos); ++ if (unlikely(sz != PAGE_SIZE)) { ++ err = sz; ++ AuIOErr("err %d\n", err); ++ if (sz >= 0) ++ err = -EIO; ++ goto out; ++ } ++ ++ mutex_lock(&sbinfo->si_xib_mtx); ++ /* mnt_want_write() is unnecessary here */ ++ err = xib_restore(sb); ++ mutex_unlock(&sbinfo->si_xib_mtx); ++ ++out: ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* ++ * xino mount option handlers ++ */ ++static au_readf_t find_readf(struct file *h_file) ++{ ++ const struct file_operations *fop = h_file->f_op; ++ ++ if (fop) { ++ if (fop->read) ++ return fop->read; ++ if (fop->aio_read) ++ return do_sync_read; ++ } ++ return ERR_PTR(-ENOSYS); ++} ++ ++static au_writef_t find_writef(struct file *h_file) ++{ ++ const struct file_operations *fop = h_file->f_op; ++ ++ if (fop) { ++ if (fop->write) ++ return fop->write; ++ if (fop->aio_write) ++ return do_sync_write; ++ } ++ return ERR_PTR(-ENOSYS); ++} ++ ++/* xino bitmap */ ++static void xino_clear_xib(struct super_block *sb) ++{ ++ struct au_sbinfo *sbinfo; ++ ++ SiMustWriteLock(sb); ++ ++ sbinfo = au_sbi(sb); ++ sbinfo->si_xread = NULL; ++ sbinfo->si_xwrite = NULL; ++ if (sbinfo->si_xib) ++ fput(sbinfo->si_xib); ++ sbinfo->si_xib = NULL; ++ free_page((unsigned long)sbinfo->si_xib_buf); ++ sbinfo->si_xib_buf = NULL; ++} ++ ++static int au_xino_set_xib(struct super_block *sb, struct file *base) ++{ ++ int err; ++ loff_t pos; ++ struct au_sbinfo *sbinfo; ++ struct file *file; ++ ++ SiMustWriteLock(sb); ++ ++ sbinfo = au_sbi(sb); ++ file = au_xino_create2(base, sbinfo->si_xib); ++ err = PTR_ERR(file); ++ if (IS_ERR(file)) ++ goto out; ++ if (sbinfo->si_xib) ++ fput(sbinfo->si_xib); ++ sbinfo->si_xib = file; ++ sbinfo->si_xread = find_readf(file); ++ sbinfo->si_xwrite = find_writef(file); ++ ++ err = -ENOMEM; ++ if (!sbinfo->si_xib_buf) ++ sbinfo->si_xib_buf = (void *)get_zeroed_page(GFP_NOFS); ++ if (unlikely(!sbinfo->si_xib_buf)) ++ goto out_unset; ++ ++ sbinfo->si_xib_last_pindex = 0; ++ sbinfo->si_xib_next_bit = 0; ++ if (i_size_read(file->f_dentry->d_inode) < PAGE_SIZE) { ++ pos = 0; ++ err = xino_fwrite(sbinfo->si_xwrite, file, sbinfo->si_xib_buf, ++ PAGE_SIZE, &pos); ++ if (unlikely(err != PAGE_SIZE)) ++ goto out_free; ++ } ++ err = 0; ++ goto out; /* success */ ++ ++ out_free: ++ free_page((unsigned long)sbinfo->si_xib_buf); ++ sbinfo->si_xib_buf = NULL; ++ if (err >= 0) ++ err = -EIO; ++ out_unset: ++ fput(sbinfo->si_xib); ++ sbinfo->si_xib = NULL; ++ sbinfo->si_xread = NULL; ++ sbinfo->si_xwrite = NULL; ++ out: ++ return err; ++} ++ ++/* xino for each branch */ ++static void xino_clear_br(struct super_block *sb) ++{ ++ aufs_bindex_t bindex, bend; ++ struct au_branch *br; ++ ++ bend = au_sbend(sb); ++ for (bindex = 0; bindex <= bend; bindex++) { ++ br = au_sbr(sb, bindex); ++ if (!br || !br->br_xino.xi_file) ++ continue; ++ ++ fput(br->br_xino.xi_file); ++ br->br_xino.xi_file = NULL; ++ } ++} ++ ++static int au_xino_set_br(struct super_block *sb, struct file *base) ++{ ++ int err; ++ ino_t ino; ++ aufs_bindex_t bindex, bend, bshared; ++ struct { ++ struct file *old, *new; ++ } *fpair, *p; ++ struct au_branch *br; ++ struct inode *inode; ++ au_writef_t writef; ++ ++ SiMustWriteLock(sb); ++ ++ err = -ENOMEM; ++ bend = au_sbend(sb); ++ fpair = kcalloc(bend + 1, sizeof(*fpair), GFP_NOFS); ++ if (unlikely(!fpair)) ++ goto out; ++ ++ inode = sb->s_root->d_inode; ++ ino = AUFS_ROOT_INO; ++ writef = au_sbi(sb)->si_xwrite; ++ for (bindex = 0, p = fpair; bindex <= bend; bindex++, p++) { ++ br = au_sbr(sb, bindex); ++ bshared = is_sb_shared(sb, bindex, bindex - 1); ++ if (bshared >= 0) { ++ /* shared xino */ ++ *p = fpair[bshared]; ++ get_file(p->new); ++ } ++ ++ if (!p->new) { ++ /* new xino */ ++ p->old = br->br_xino.xi_file; ++ p->new = au_xino_create2(base, br->br_xino.xi_file); ++ err = PTR_ERR(p->new); ++ if (IS_ERR(p->new)) { ++ p->new = NULL; ++ goto out_pair; ++ } ++ } ++ ++ err = au_xino_do_write(writef, p->new, ++ au_h_iptr(inode, bindex)->i_ino, ino); ++ if (unlikely(err)) ++ goto out_pair; ++ } ++ ++ for (bindex = 0, p = fpair; bindex <= bend; bindex++, p++) { ++ br = au_sbr(sb, bindex); ++ if (br->br_xino.xi_file) ++ fput(br->br_xino.xi_file); ++ get_file(p->new); ++ br->br_xino.xi_file = p->new; ++ } ++ ++ out_pair: ++ for (bindex = 0, p = fpair; bindex <= bend; bindex++, p++) ++ if (p->new) ++ fput(p->new); ++ else ++ break; ++ kfree(fpair); ++ out: ++ return err; ++} ++ ++void au_xino_clr(struct super_block *sb) ++{ ++ struct au_sbinfo *sbinfo; ++ ++ au_xigen_clr(sb); ++ xino_clear_xib(sb); ++ xino_clear_br(sb); ++ sbinfo = au_sbi(sb); ++ /* lvalue, do not call au_mntflags() */ ++ au_opt_clr(sbinfo->si_mntflags, XINO); ++} ++ ++int au_xino_set(struct super_block *sb, struct au_opt_xino *xino, int remount) ++{ ++ int err, skip; ++ struct dentry *parent, *cur_parent; ++ struct qstr *dname, *cur_name; ++ struct file *cur_xino; ++ struct inode *dir; ++ struct au_sbinfo *sbinfo; ++ ++ SiMustWriteLock(sb); ++ ++ err = 0; ++ sbinfo = au_sbi(sb); ++ parent = dget_parent(xino->file->f_dentry); ++ if (remount) { ++ skip = 0; ++ dname = &xino->file->f_dentry->d_name; ++ cur_xino = sbinfo->si_xib; ++ if (cur_xino) { ++ cur_parent = dget_parent(cur_xino->f_dentry); ++ cur_name = &cur_xino->f_dentry->d_name; ++ skip = (cur_parent == parent ++ && dname->len == cur_name->len ++ && !memcmp(dname->name, cur_name->name, ++ dname->len)); ++ dput(cur_parent); ++ } ++ if (skip) ++ goto out; ++ } ++ ++ au_opt_set(sbinfo->si_mntflags, XINO); ++ dir = parent->d_inode; ++ mutex_lock_nested(&dir->i_mutex, AuLsc_I_PARENT); ++ /* mnt_want_write() is unnecessary here */ ++ err = au_xino_set_xib(sb, xino->file); ++ if (!err) ++ err = au_xigen_set(sb, xino->file); ++ if (!err) ++ err = au_xino_set_br(sb, xino->file); ++ mutex_unlock(&dir->i_mutex); ++ if (!err) ++ goto out; /* success */ ++ ++ /* reset all */ ++ AuIOErr("failed creating xino(%d).\n", err); ++ ++ out: ++ dput(parent); ++ return err; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* ++ * create a xinofile at the default place/path. ++ */ ++struct file *au_xino_def(struct super_block *sb) ++{ ++ struct file *file; ++ char *page, *p; ++ struct au_branch *br; ++ struct super_block *h_sb; ++ struct path path; ++ aufs_bindex_t bend, bindex, bwr; ++ ++ br = NULL; ++ bend = au_sbend(sb); ++ bwr = -1; ++ for (bindex = 0; bindex <= bend; bindex++) { ++ br = au_sbr(sb, bindex); ++ if (au_br_writable(br->br_perm) ++ && !au_test_fs_bad_xino(br->br_mnt->mnt_sb)) { ++ bwr = bindex; ++ break; ++ } ++ } ++ ++ if (bwr >= 0) { ++ file = ERR_PTR(-ENOMEM); ++ page = __getname(); ++ if (unlikely(!page)) ++ goto out; ++ path.mnt = br->br_mnt; ++ path.dentry = au_h_dptr(sb->s_root, bwr); ++ p = d_path(&path, page, PATH_MAX - sizeof(AUFS_XINO_FNAME)); ++ file = (void *)p; ++ if (!IS_ERR(p)) { ++ strcat(p, "/" AUFS_XINO_FNAME); ++ AuDbg("%s\n", p); ++ file = au_xino_create(sb, p, /*silent*/0); ++ if (!IS_ERR(file)) ++ au_xino_brid_set(sb, br->br_id); ++ } ++ __putname(page); ++ } else { ++ file = au_xino_create(sb, AUFS_XINO_DEFPATH, /*silent*/0); ++ if (IS_ERR(file)) ++ goto out; ++ h_sb = file->f_dentry->d_sb; ++ if (unlikely(au_test_fs_bad_xino(h_sb))) { ++ AuErr("xino doesn't support %s(%s)\n", ++ AUFS_XINO_DEFPATH, au_sbtype(h_sb)); ++ fput(file); ++ file = ERR_PTR(-EINVAL); ++ } ++ if (!IS_ERR(file)) ++ au_xino_brid_set(sb, -1); ++ } ++ ++ out: ++ return file; ++} ++ ++/* ---------------------------------------------------------------------- */ ++ ++int au_xino_path(struct seq_file *seq, struct file *file) ++{ ++ int err; ++ ++ err = au_seq_path(seq, &file->f_path); ++ if (unlikely(err < 0)) ++ goto out; ++ ++ err = 0; ++#define Deleted "\\040(deleted)" ++ seq->count -= sizeof(Deleted) - 1; ++ AuDebugOn(memcmp(seq->buf + seq->count, Deleted, ++ sizeof(Deleted) - 1)); ++#undef Deleted ++ ++ out: ++ return err; ++} +diff -Nur linux-2.6.31.4.orig/fs/Kconfig linux-2.6.31.4/fs/Kconfig +--- linux-2.6.31.4.orig/fs/Kconfig 2009-10-12 22:15:40.000000000 +0200 ++++ linux-2.6.31.4/fs/Kconfig 2009-10-18 11:26:53.000000000 +0200 +@@ -187,6 +187,7 @@ + source "fs/ufs/Kconfig" + source "fs/exofs/Kconfig" + source "fs/nilfs2/Kconfig" ++source "fs/aufs/Kconfig" + + endif # MISC_FILESYSTEMS + +diff -Nur linux-2.6.31.4.orig/fs/Makefile linux-2.6.31.4/fs/Makefile +--- linux-2.6.31.4.orig/fs/Makefile 2009-10-12 22:15:40.000000000 +0200 ++++ linux-2.6.31.4/fs/Makefile 2009-10-18 11:26:01.000000000 +0200 +@@ -85,6 +85,7 @@ + obj-$(CONFIG_HFS_FS) += hfs/ + obj-$(CONFIG_ECRYPT_FS) += ecryptfs/ + obj-$(CONFIG_VXFS_FS) += freevxfs/ ++obj-$(CONFIG_AUFS_FS) += aufs/ + obj-$(CONFIG_NFS_FS) += nfs/ + obj-$(CONFIG_EXPORTFS) += exportfs/ + obj-$(CONFIG_NFSD) += nfsd/ +diff -Nur linux-2.6.31.4.orig/fs/namei.c linux-2.6.31.4/fs/namei.c +--- linux-2.6.31.4.orig/fs/namei.c 2009-10-12 22:15:40.000000000 +0200 ++++ linux-2.6.31.4/fs/namei.c 2009-10-18 11:26:01.000000000 +0200 +@@ -337,6 +337,7 @@ + + return 0; + } ++EXPORT_SYMBOL(deny_write_access); + + /** + * path_get - get a reference to a path +@@ -1219,7 +1220,7 @@ + * needs parent already locked. Doesn't follow mounts. + * SMP-safe. + */ +-static struct dentry *lookup_hash(struct nameidata *nd) ++struct dentry *lookup_hash(struct nameidata *nd) + { + int err; + +@@ -1228,8 +1229,9 @@ + return ERR_PTR(err); + return __lookup_hash(&nd->last, nd->path.dentry, nd); + } ++EXPORT_SYMBOL(lookup_hash); + +-static int __lookup_one_len(const char *name, struct qstr *this, ++int __lookup_one_len(const char *name, struct qstr *this, + struct dentry *base, int len) + { + unsigned long hash; +@@ -1250,6 +1252,7 @@ + this->hash = end_name_hash(hash); + return 0; + } ++EXPORT_SYMBOL(__lookup_one_len); + + /** + * lookup_one_len - filesystem helper to lookup single pathname component +diff -Nur linux-2.6.31.4.orig/fs/namespace.c linux-2.6.31.4/fs/namespace.c +--- linux-2.6.31.4.orig/fs/namespace.c 2009-10-12 22:15:40.000000000 +0200 ++++ linux-2.6.31.4/fs/namespace.c 2009-10-18 11:26:01.000000000 +0200 +@@ -39,6 +39,7 @@ + + /* spinlock for vfsmount related operations, inplace of dcache_lock */ + __cacheline_aligned_in_smp DEFINE_SPINLOCK(vfsmount_lock); ++EXPORT_SYMBOL(vfsmount_lock); + + static int event; + static DEFINE_IDA(mnt_id_ida); +diff -Nur linux-2.6.31.4.orig/fs/open.c linux-2.6.31.4/fs/open.c +--- linux-2.6.31.4.orig/fs/open.c 2009-10-12 22:15:40.000000000 +0200 ++++ linux-2.6.31.4/fs/open.c 2009-10-18 11:26:01.000000000 +0200 +@@ -221,6 +221,7 @@ + mutex_unlock(&dentry->d_inode->i_mutex); + return err; + } ++EXPORT_SYMBOL(do_truncate); + + static long do_sys_truncate(const char __user *pathname, loff_t length) + { +diff -Nur linux-2.6.31.4.orig/fs/splice.c linux-2.6.31.4/fs/splice.c +--- linux-2.6.31.4.orig/fs/splice.c 2009-10-12 22:15:40.000000000 +0200 ++++ linux-2.6.31.4/fs/splice.c 2009-10-18 11:26:01.000000000 +0200 +@@ -1057,8 +1057,8 @@ + /* + * Attempt to initiate a splice from pipe to file. + */ +-static long do_splice_from(struct pipe_inode_info *pipe, struct file *out, +- loff_t *ppos, size_t len, unsigned int flags) ++long do_splice_from(struct pipe_inode_info *pipe, struct file *out, ++ loff_t *ppos, size_t len, unsigned int flags) + { + ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, + loff_t *, size_t, unsigned int); +@@ -1080,13 +1080,14 @@ + + return splice_write(pipe, out, ppos, len, flags); + } ++EXPORT_SYMBOL(do_splice_from); + + /* + * Attempt to initiate a splice from a file to a pipe. + */ +-static long do_splice_to(struct file *in, loff_t *ppos, +- struct pipe_inode_info *pipe, size_t len, +- unsigned int flags) ++long do_splice_to(struct file *in, loff_t *ppos, ++ struct pipe_inode_info *pipe, size_t len, ++ unsigned int flags) + { + ssize_t (*splice_read)(struct file *, loff_t *, + struct pipe_inode_info *, size_t, unsigned int); +@@ -1105,6 +1106,7 @@ + + return splice_read(in, ppos, pipe, len, flags); + } ++EXPORT_SYMBOL(do_splice_to); + + /** + * splice_direct_to_actor - splices data directly between two non-pipes +diff -Nur linux-2.6.31.4.orig/include/linux/aufs_type.h linux-2.6.31.4/include/linux/aufs_type.h +--- linux-2.6.31.4.orig/include/linux/aufs_type.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.31.4/include/linux/aufs_type.h 2009-10-18 11:26:01.000000000 +0200 +@@ -0,0 +1,109 @@ ++/* ++ * Copyright (C) 2005-2009 Junjiro R. Okajima ++ * ++ * This program, aufs is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License as published by ++ * the Free Software Foundation; either version 2 of the License, or ++ * (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA ++ */ ++ ++#ifndef __AUFS_TYPE_H__ ++#define __AUFS_TYPE_H__ ++ ++#include <linux/ioctl.h> ++ ++#define AUFS_VERSION "2-standalone.tree-30-20090803" ++ ++/* todo? move this to linux-2.6.19/include/magic.h */ ++#define AUFS_SUPER_MAGIC ('a' << 24 | 'u' << 16 | 'f' << 8 | 's') ++ ++/* ---------------------------------------------------------------------- */ ++ ++#ifdef CONFIG_AUFS_BRANCH_MAX_127 ++/* some environments treat 'char' as 'unsigned char' by default */ ++typedef signed char aufs_bindex_t; ++#define AUFS_BRANCH_MAX 127 ++#else ++typedef short aufs_bindex_t; ++#ifdef CONFIG_AUFS_BRANCH_MAX_511 ++#define AUFS_BRANCH_MAX 511 ++#elif defined(CONFIG_AUFS_BRANCH_MAX_1023) ++#define AUFS_BRANCH_MAX 1023 ++#elif defined(CONFIG_AUFS_BRANCH_MAX_32767) ++#define AUFS_BRANCH_MAX 32767 ++#endif ++#endif ++ ++#ifdef __KERNEL__ ++#ifndef AUFS_BRANCH_MAX ++#error unknown CONFIG_AUFS_BRANCH_MAX value ++#endif ++#endif /* __KERNEL__ */ ++ ++/* ---------------------------------------------------------------------- */ ++ ++#define AUFS_NAME "aufs" ++#define AUFS_FSTYPE AUFS_NAME ++ ++#define AUFS_ROOT_INO 2 ++#define AUFS_FIRST_INO 11 ++ ++#define AUFS_WH_PFX ".wh." ++#define AUFS_WH_PFX_LEN ((int)sizeof(AUFS_WH_PFX) - 1) ++#define AUFS_XINO_FNAME "." AUFS_NAME ".xino" ++#define AUFS_XINO_DEFPATH "/tmp/" AUFS_XINO_FNAME ++#define AUFS_XINO_TRUNC_INIT 64 /* blocks */ ++#define AUFS_XINO_TRUNC_STEP 4 /* blocks */ ++#define AUFS_DIRWH_DEF 3 ++#define AUFS_RDCACHE_DEF 10 /* seconds */ ++#define AUFS_RDBLK_DEF 512 /* bytes */ ++#define AUFS_RDHASH_DEF 32 ++#define AUFS_WKQ_NAME AUFS_NAME "d" ++#define AUFS_NWKQ_DEF 4 ++#define AUFS_MFS_SECOND_DEF 30 /* seconds */ ++#define AUFS_PLINK_WARN 100 /* number of plinks */ ++ ++#define AUFS_DIROPQ_NAME AUFS_WH_PFX ".opq" /* whiteouted doubly */ ++#define AUFS_WH_DIROPQ AUFS_WH_PFX AUFS_DIROPQ_NAME ++ ++#define AUFS_BASE_NAME AUFS_WH_PFX AUFS_NAME ++#define AUFS_PLINKDIR_NAME AUFS_WH_PFX "plnk" ++#define AUFS_ORPHDIR_NAME AUFS_WH_PFX "orph" ++ ++/* doubly whiteouted */ ++#define AUFS_WH_BASE AUFS_WH_PFX AUFS_BASE_NAME ++#define AUFS_WH_PLINKDIR AUFS_WH_PFX AUFS_PLINKDIR_NAME ++#define AUFS_WH_ORPHDIR AUFS_WH_PFX AUFS_ORPHDIR_NAME ++ ++/* branch permission */ ++#define AUFS_BRPERM_RW "rw" ++#define AUFS_BRPERM_RO "ro" ++#define AUFS_BRPERM_RR "rr" ++#define AUFS_BRPERM_WH "wh" ++#define AUFS_BRPERM_NLWH "nolwh" ++#define AUFS_BRPERM_ROWH AUFS_BRPERM_RO "+" AUFS_BRPERM_WH ++#define AUFS_BRPERM_RRWH AUFS_BRPERM_RR "+" AUFS_BRPERM_WH ++#define AUFS_BRPERM_RWNLWH AUFS_BRPERM_RW "+" AUFS_BRPERM_NLWH ++ ++/* ---------------------------------------------------------------------- */ ++ ++/* ioctl */ ++enum { ++ AuCtl_PLINK_MAINT, ++ AuCtl_PLINK_CLEAN ++}; ++ ++#define AuCtlType 'A' ++#define AUFS_CTL_PLINK_MAINT _IO(AuCtlType, AuCtl_PLINK_MAINT) ++#define AUFS_CTL_PLINK_CLEAN _IO(AuCtlType, AuCtl_PLINK_CLEAN) ++ ++#endif /* __AUFS_TYPE_H__ */ +diff -Nur linux-2.6.31.4.orig/include/linux/Kbuild linux-2.6.31.4/include/linux/Kbuild +--- linux-2.6.31.4.orig/include/linux/Kbuild 2009-10-12 22:15:40.000000000 +0200 ++++ linux-2.6.31.4/include/linux/Kbuild 2009-10-18 11:26:01.000000000 +0200 +@@ -34,6 +34,7 @@ + header-y += atmsap.h + header-y += atmsvc.h + header-y += atm_zatm.h ++header-y += aufs_type.h + header-y += auto_fs4.h + header-y += ax25.h + header-y += b1lli.h +diff -Nur linux-2.6.31.4.orig/include/linux/namei.h linux-2.6.31.4/include/linux/namei.h +--- linux-2.6.31.4.orig/include/linux/namei.h 2009-10-12 22:15:40.000000000 +0200 ++++ linux-2.6.31.4/include/linux/namei.h 2009-10-18 11:26:01.000000000 +0200 +@@ -75,6 +75,9 @@ + extern struct file *nameidata_to_filp(struct nameidata *nd, int flags); + extern void release_open_intent(struct nameidata *); + ++extern struct dentry *lookup_hash(struct nameidata *nd); ++extern int __lookup_one_len(const char *name, struct qstr *this, ++ struct dentry *base, int len); + extern struct dentry *lookup_one_len(const char *, struct dentry *, int); + extern struct dentry *lookup_one_noperm(const char *, struct dentry *); + +diff -Nur linux-2.6.31.4.orig/include/linux/splice.h linux-2.6.31.4/include/linux/splice.h +--- linux-2.6.31.4.orig/include/linux/splice.h 2009-10-12 22:15:40.000000000 +0200 ++++ linux-2.6.31.4/include/linux/splice.h 2009-10-18 11:26:01.000000000 +0200 +@@ -82,4 +82,10 @@ + extern ssize_t splice_direct_to_actor(struct file *, struct splice_desc *, + splice_direct_actor *); + ++extern long do_splice_from(struct pipe_inode_info *pipe, struct file *out, ++ loff_t *ppos, size_t len, unsigned int flags); ++extern long do_splice_to(struct file *in, loff_t *ppos, ++ struct pipe_inode_info *pipe, size_t len, ++ unsigned int flags); ++ + #endif +diff -Nur linux-2.6.31.4.orig/security/device_cgroup.c linux-2.6.31.4/security/device_cgroup.c +--- linux-2.6.31.4.orig/security/device_cgroup.c 2009-10-12 22:15:40.000000000 +0200 ++++ linux-2.6.31.4/security/device_cgroup.c 2009-10-18 11:26:01.000000000 +0200 +@@ -513,6 +513,7 @@ + + return -EPERM; + } ++EXPORT_SYMBOL(devcgroup_inode_permission); + + int devcgroup_inode_mknod(int mode, dev_t dev) + { +diff -Nur linux-2.6.31.4.orig/security/security.c linux-2.6.31.4/security/security.c +--- linux-2.6.31.4.orig/security/security.c 2009-10-12 22:15:40.000000000 +0200 ++++ linux-2.6.31.4/security/security.c 2009-10-18 11:26:01.000000000 +0200 +@@ -386,6 +386,7 @@ + return 0; + return security_ops->path_mkdir(path, dentry, mode); + } ++EXPORT_SYMBOL(security_path_mkdir); + + int security_path_rmdir(struct path *path, struct dentry *dentry) + { +@@ -393,6 +394,7 @@ + return 0; + return security_ops->path_rmdir(path, dentry); + } ++EXPORT_SYMBOL(security_path_rmdir); + + int security_path_unlink(struct path *path, struct dentry *dentry) + { +@@ -400,6 +402,7 @@ + return 0; + return security_ops->path_unlink(path, dentry); + } ++EXPORT_SYMBOL(security_path_unlink); + + int security_path_symlink(struct path *path, struct dentry *dentry, + const char *old_name) +@@ -408,6 +411,7 @@ + return 0; + return security_ops->path_symlink(path, dentry, old_name); + } ++EXPORT_SYMBOL(security_path_symlink); + + int security_path_link(struct dentry *old_dentry, struct path *new_dir, + struct dentry *new_dentry) +@@ -416,6 +420,7 @@ + return 0; + return security_ops->path_link(old_dentry, new_dir, new_dentry); + } ++EXPORT_SYMBOL(security_path_link); + + int security_path_rename(struct path *old_dir, struct dentry *old_dentry, + struct path *new_dir, struct dentry *new_dentry) +@@ -426,6 +431,7 @@ + return security_ops->path_rename(old_dir, old_dentry, new_dir, + new_dentry); + } ++EXPORT_SYMBOL(security_path_rename); + + int security_path_truncate(struct path *path, loff_t length, + unsigned int time_attrs) +@@ -434,6 +440,7 @@ + return 0; + return security_ops->path_truncate(path, length, time_attrs); + } ++EXPORT_SYMBOL(security_path_truncate); + #endif + + int security_inode_create(struct inode *dir, struct dentry *dentry, int mode) +@@ -505,6 +512,7 @@ + return 0; + return security_ops->inode_readlink(dentry); + } ++EXPORT_SYMBOL(security_inode_readlink); + + int security_inode_follow_link(struct dentry *dentry, struct nameidata *nd) + { +@@ -519,6 +527,7 @@ + return 0; + return security_ops->inode_permission(inode, mask); + } ++EXPORT_SYMBOL(security_inode_permission); + + int security_inode_setattr(struct dentry *dentry, struct iattr *attr) + { +@@ -619,6 +628,7 @@ + { + return security_ops->file_permission(file, mask); + } ++EXPORT_SYMBOL(security_file_permission); + + int security_file_alloc(struct file *file) + { diff --git a/target/linux/patches/2.6.31.4/cygwin-compat.patch b/target/linux/patches/2.6.31.4/cygwin-compat.patch new file mode 100644 index 000000000..8d087dddf --- /dev/null +++ b/target/linux/patches/2.6.31.4/cygwin-compat.patch @@ -0,0 +1,66 @@ +diff -Nur linux-2.6.30.orig/scripts/mod/file2alias.c linux-2.6.30/scripts/mod/file2alias.c +--- linux-2.6.30.orig/scripts/mod/file2alias.c 2009-06-10 05:05:27.000000000 +0200 ++++ linux-2.6.30/scripts/mod/file2alias.c 2009-06-11 09:17:10.000000000 +0200 +@@ -29,7 +29,11 @@ + + #include <ctype.h> + ++#ifdef __CYGWIN__ ++typedef __uint32_t __u32; ++#else + typedef uint32_t __u32; ++#endif + typedef uint16_t __u16; + typedef unsigned char __u8; + +diff -Nur linux-2.6.30.orig/scripts/mod/modpost.h linux-2.6.30/scripts/mod/modpost.h +--- linux-2.6.30.orig/scripts/mod/modpost.h 2009-06-10 05:05:27.000000000 +0200 ++++ linux-2.6.30/scripts/mod/modpost.h 2009-06-11 09:17:10.000000000 +0200 +@@ -9,6 +9,11 @@ + #include <unistd.h> + #include <elf.h> + ++#ifdef __CYGWIN__ ++typedef uint16_t Elf32_Section; ++typedef uint16_t Elf64_Section; ++#endif ++ + #include "elfconfig.h" + + #if KERNEL_ELFCLASS == ELFCLASS32 +@@ -19,13 +24,17 @@ + #define Elf_Addr Elf32_Addr + #define Elf_Sword Elf64_Sword + #define Elf_Section Elf32_Half ++#ifndef __CYGWIN__ + #define ELF_ST_BIND ELF32_ST_BIND + #define ELF_ST_TYPE ELF32_ST_TYPE ++#endif + + #define Elf_Rel Elf32_Rel + #define Elf_Rela Elf32_Rela ++#ifndef __CYGWIN__ + #define ELF_R_SYM ELF32_R_SYM + #define ELF_R_TYPE ELF32_R_TYPE ++#endif + #else + + #define Elf_Ehdr Elf64_Ehdr +@@ -43,6 +52,17 @@ + #define ELF_R_TYPE ELF64_R_TYPE + #endif + ++#define R_386_32 1 /* Direct 32 bit */ ++#define R_386_PC32 2 /* PC relative 32 bit */ ++ ++#define R_ARM_PC24 1 /* PC relative 26 bit branch */ ++#define R_ARM_ABS32 2 /* Direct 32 bit */ ++ ++#define R_MIPS_32 2 /* Direct 32 bit */ ++#define R_MIPS_26 4 /* Direct 26 bit shifted */ ++#define R_MIPS_HI16 5 /* High 16 bit */ ++#define R_MIPS_LO16 6 /* Low 16 bit */ ++ + /* The 64-bit MIPS ELF ABI uses an unusual reloc format. */ + typedef struct + { diff --git a/target/linux/patches/2.6.31.4/freebsd-compat.patch b/target/linux/patches/2.6.31.4/freebsd-compat.patch new file mode 100644 index 000000000..051fdc63e --- /dev/null +++ b/target/linux/patches/2.6.31.4/freebsd-compat.patch @@ -0,0 +1,11 @@ +diff -Nur linux-2.6.30.orig/arch/x86/boot/tools/build.c linux-2.6.30/arch/x86/boot/tools/build.c +--- linux-2.6.30.orig/arch/x86/boot/tools/build.c 2009-06-10 05:05:27.000000000 +0200 ++++ linux-2.6.30/arch/x86/boot/tools/build.c 2009-06-11 09:18:50.000000000 +0200 +@@ -29,7 +29,6 @@ + #include <stdarg.h> + #include <sys/types.h> + #include <sys/stat.h> +-#include <sys/sysmacros.h> + #include <unistd.h> + #include <fcntl.h> + #include <sys/mman.h> diff --git a/target/linux/patches/2.6.31.4/mtd-root.patch b/target/linux/patches/2.6.31.4/mtd-root.patch new file mode 100644 index 000000000..3576848be --- /dev/null +++ b/target/linux/patches/2.6.31.4/mtd-root.patch @@ -0,0 +1,62 @@ +diff -Nur linux-2.6.29.1.orig/drivers/mtd/Kconfig linux-2.6.29.1/drivers/mtd/Kconfig +--- linux-2.6.29.1.orig/drivers/mtd/Kconfig 2009-04-02 22:55:27.000000000 +0200 ++++ linux-2.6.29.1/drivers/mtd/Kconfig 2009-05-02 19:24:14.444062164 +0200 +@@ -53,6 +53,11 @@ + should normally be compiled as kernel modules. The modules perform + various checks and verifications when loaded. + ++config MTD_ROOTFS_ROOT_DEV ++ bool "Automatically set 'rootfs' partition to be root filesystem" ++ depends on MTD_PARTITIONS ++ default y ++ + config MTD_REDBOOT_PARTS + tristate "RedBoot partition table parsing" + depends on MTD_PARTITIONS +diff -Nur linux-2.6.29.1.orig/drivers/mtd/mtdpart.c linux-2.6.29.1/drivers/mtd/mtdpart.c +--- linux-2.6.29.1.orig/drivers/mtd/mtdpart.c 2009-04-02 22:55:27.000000000 +0200 ++++ linux-2.6.29.1/drivers/mtd/mtdpart.c 2009-05-02 19:26:39.038093851 +0200 +@@ -18,6 +18,7 @@ + #include <linux/mtd/mtd.h> + #include <linux/mtd/partitions.h> + #include <linux/mtd/compatmac.h> ++#include <linux/root_dev.h> + + /* Our partition linked list */ + static LIST_HEAD(mtd_partitions); +@@ -37,7 +38,7 @@ + * the pointer to that structure with this macro. + */ + #define PART(x) ((struct mtd_part *)(x)) +- ++#define IS_PART(mtd) (mtd->read == part_read) + + /* + * MTD methods which simply translate the effective address and pass through +@@ -502,14 +503,23 @@ + { + struct mtd_part *slave; + uint64_t cur_offset = 0; +- int i; ++ int i, j, ret; + + printk(KERN_NOTICE "Creating %d MTD partitions on \"%s\":\n", nbparts, master->name); + +- for (i = 0; i < nbparts; i++) { +- slave = add_one_partition(master, parts + i, i, cur_offset); ++ for (i = 0, j = 0; i < nbparts; i++) { ++ slave = add_one_partition(master, parts + i, j++, cur_offset); + if (!slave) + return -ENOMEM; ++ if (!strcmp(parts[i].name, "rootfs") && slave->registered) { ++#ifdef CONFIG_MTD_ROOTFS_ROOT_DEV ++ if (ROOT_DEV == 0) { ++ printk(KERN_NOTICE "mtd: partition \"rootfs\" " ++ "set to be root filesystem\n"); ++ ROOT_DEV = MKDEV(MTD_BLOCK_MAJOR, slave->mtd.index); ++ } ++#endif ++ } + cur_offset = slave->offset + slave->mtd.size; + } + diff --git a/target/linux/patches/2.6.31.4/ocf.patch b/target/linux/patches/2.6.31.4/ocf.patch new file mode 100644 index 000000000..64c5eeb0f --- /dev/null +++ b/target/linux/patches/2.6.31.4/ocf.patch @@ -0,0 +1,23653 @@ +diff -Nur linux-2.6.30.orig/crypto/Kconfig linux-2.6.30/crypto/Kconfig +--- linux-2.6.30.orig/crypto/Kconfig 2009-06-10 05:05:27.000000000 +0200 ++++ linux-2.6.30/crypto/Kconfig 2009-06-11 10:55:27.000000000 +0200 +@@ -781,3 +781,5 @@ + source "drivers/crypto/Kconfig" + + endif # if CRYPTO ++ ++source "crypto/ocf/Kconfig" +diff -Nur linux-2.6.30.orig/crypto/Makefile linux-2.6.30/crypto/Makefile +--- linux-2.6.30.orig/crypto/Makefile 2009-06-10 05:05:27.000000000 +0200 ++++ linux-2.6.30/crypto/Makefile 2009-06-11 10:55:27.000000000 +0200 +@@ -84,6 +84,8 @@ + obj-$(CONFIG_CRYPTO_ANSI_CPRNG) += ansi_cprng.o + obj-$(CONFIG_CRYPTO_TEST) += tcrypt.o + ++obj-$(CONFIG_OCF_OCF) += ocf/ ++ + # + # generic algorithms and the async_tx api + # +diff -Nur linux-2.6.30.orig/crypto/ocf/Config.in linux-2.6.30/crypto/ocf/Config.in +--- linux-2.6.30.orig/crypto/ocf/Config.in 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/Config.in 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,34 @@ ++############################################################################# ++ ++mainmenu_option next_comment ++comment 'OCF Configuration' ++tristate 'OCF (Open Cryptograhic Framework)' CONFIG_OCF_OCF ++dep_mbool ' enable fips RNG checks (fips check on RNG data before use)' \ ++ CONFIG_OCF_FIPS $CONFIG_OCF_OCF ++dep_mbool ' enable harvesting entropy for /dev/random' \ ++ CONFIG_OCF_RANDOMHARVEST $CONFIG_OCF_OCF ++dep_tristate ' cryptodev (user space support)' \ ++ CONFIG_OCF_CRYPTODEV $CONFIG_OCF_OCF ++dep_tristate ' cryptosoft (software crypto engine)' \ ++ CONFIG_OCF_CRYPTOSOFT $CONFIG_OCF_OCF ++dep_tristate ' safenet (HW crypto engine)' \ ++ CONFIG_OCF_SAFE $CONFIG_OCF_OCF ++dep_tristate ' IXP4xx (HW crypto engine)' \ ++ CONFIG_OCF_IXP4XX $CONFIG_OCF_OCF ++dep_mbool ' Enable IXP4xx HW to perform SHA1 and MD5 hashing (very slow)' \ ++ CONFIG_OCF_IXP4XX_SHA1_MD5 $CONFIG_OCF_IXP4XX ++dep_tristate ' hifn (HW crypto engine)' \ ++ CONFIG_OCF_HIFN $CONFIG_OCF_OCF ++dep_tristate ' talitos (HW crypto engine)' \ ++ CONFIG_OCF_TALITOS $CONFIG_OCF_OCF ++dep_tristate ' pasemi (HW crypto engine)' \ ++ CONFIG_OCF_PASEMI $CONFIG_OCF_OCF ++dep_tristate ' ep80579 (HW crypto engine)' \ ++ CONFIG_OCF_EP80579 $CONFIG_OCF_OCF ++dep_tristate ' ocfnull (does no crypto)' \ ++ CONFIG_OCF_OCFNULL $CONFIG_OCF_OCF ++dep_tristate ' ocf-bench (HW crypto in-kernel benchmark)' \ ++ CONFIG_OCF_BENCH $CONFIG_OCF_OCF ++endmenu ++ ++############################################################################# +diff -Nur linux-2.6.30.orig/crypto/ocf/criov.c linux-2.6.30/crypto/ocf/criov.c +--- linux-2.6.30.orig/crypto/ocf/criov.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/criov.c 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,215 @@ ++/* $OpenBSD: criov.c,v 1.9 2002/01/29 15:48:29 jason Exp $ */ ++ ++/* ++ * Linux port done by David McCullough <david_mccullough@securecomputing.com> ++ * Copyright (C) 2006-2007 David McCullough ++ * Copyright (C) 2004-2005 Intel Corporation. ++ * The license and original author are listed below. ++ * ++ * Copyright (c) 1999 Theo de Raadt ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions ++ * are met: ++ * ++ * 1. Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * 2. Redistributions in binary form must reproduce the above copyright ++ * notice, this list of conditions and the following disclaimer in the ++ * documentation and/or other materials provided with the distribution. ++ * 3. The name of the author may not be used to endorse or promote products ++ * derived from this software without specific prior written permission. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR ++ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES ++ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. ++ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, ++ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT ++ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, ++ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY ++ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT ++ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF ++ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. ++ * ++__FBSDID("$FreeBSD: src/sys/opencrypto/criov.c,v 1.5 2006/06/04 22:15:13 pjd Exp $"); ++ */ ++ ++#ifndef AUTOCONF_INCLUDED ++#include <linux/config.h> ++#endif ++#include <linux/module.h> ++#include <linux/init.h> ++#include <linux/slab.h> ++#include <linux/uio.h> ++#include <linux/skbuff.h> ++#include <linux/kernel.h> ++#include <linux/mm.h> ++#include <asm/io.h> ++ ++#include <uio.h> ++#include <cryptodev.h> ++ ++/* ++ * This macro is only for avoiding code duplication, as we need to skip ++ * given number of bytes in the same way in three functions below. ++ */ ++#define CUIO_SKIP() do { \ ++ KASSERT(off >= 0, ("%s: off %d < 0", __func__, off)); \ ++ KASSERT(len >= 0, ("%s: len %d < 0", __func__, len)); \ ++ while (off > 0) { \ ++ KASSERT(iol >= 0, ("%s: empty in skip", __func__)); \ ++ if (off < iov->iov_len) \ ++ break; \ ++ off -= iov->iov_len; \ ++ iol--; \ ++ iov++; \ ++ } \ ++} while (0) ++ ++void ++cuio_copydata(struct uio* uio, int off, int len, caddr_t cp) ++{ ++ struct iovec *iov = uio->uio_iov; ++ int iol = uio->uio_iovcnt; ++ unsigned count; ++ ++ CUIO_SKIP(); ++ while (len > 0) { ++ KASSERT(iol >= 0, ("%s: empty", __func__)); ++ count = min((int)(iov->iov_len - off), len); ++ memcpy(cp, ((caddr_t)iov->iov_base) + off, count); ++ len -= count; ++ cp += count; ++ off = 0; ++ iol--; ++ iov++; ++ } ++} ++ ++void ++cuio_copyback(struct uio* uio, int off, int len, caddr_t cp) ++{ ++ struct iovec *iov = uio->uio_iov; ++ int iol = uio->uio_iovcnt; ++ unsigned count; ++ ++ CUIO_SKIP(); ++ while (len > 0) { ++ KASSERT(iol >= 0, ("%s: empty", __func__)); ++ count = min((int)(iov->iov_len - off), len); ++ memcpy(((caddr_t)iov->iov_base) + off, cp, count); ++ len -= count; ++ cp += count; ++ off = 0; ++ iol--; ++ iov++; ++ } ++} ++ ++/* ++ * Return a pointer to iov/offset of location in iovec list. ++ */ ++struct iovec * ++cuio_getptr(struct uio *uio, int loc, int *off) ++{ ++ struct iovec *iov = uio->uio_iov; ++ int iol = uio->uio_iovcnt; ++ ++ while (loc >= 0) { ++ /* Normal end of search */ ++ if (loc < iov->iov_len) { ++ *off = loc; ++ return (iov); ++ } ++ ++ loc -= iov->iov_len; ++ if (iol == 0) { ++ if (loc == 0) { ++ /* Point at the end of valid data */ ++ *off = iov->iov_len; ++ return (iov); ++ } else ++ return (NULL); ++ } else { ++ iov++, iol--; ++ } ++ } ++ ++ return (NULL); ++} ++ ++EXPORT_SYMBOL(cuio_copyback); ++EXPORT_SYMBOL(cuio_copydata); ++EXPORT_SYMBOL(cuio_getptr); ++ ++ ++static void ++skb_copy_bits_back(struct sk_buff *skb, int offset, caddr_t cp, int len) ++{ ++ int i; ++ if (offset < skb_headlen(skb)) { ++ memcpy(skb->data + offset, cp, min_t(int, skb_headlen(skb), len)); ++ len -= skb_headlen(skb); ++ cp += skb_headlen(skb); ++ } ++ offset -= skb_headlen(skb); ++ for (i = 0; len > 0 && i < skb_shinfo(skb)->nr_frags; i++) { ++ if (offset < skb_shinfo(skb)->frags[i].size) { ++ memcpy(page_address(skb_shinfo(skb)->frags[i].page) + ++ skb_shinfo(skb)->frags[i].page_offset, ++ cp, min_t(int, skb_shinfo(skb)->frags[i].size, len)); ++ len -= skb_shinfo(skb)->frags[i].size; ++ cp += skb_shinfo(skb)->frags[i].size; ++ } ++ offset -= skb_shinfo(skb)->frags[i].size; ++ } ++} ++ ++void ++crypto_copyback(int flags, caddr_t buf, int off, int size, caddr_t in) ++{ ++ ++ if ((flags & CRYPTO_F_SKBUF) != 0) ++ skb_copy_bits_back((struct sk_buff *)buf, off, in, size); ++ else if ((flags & CRYPTO_F_IOV) != 0) ++ cuio_copyback((struct uio *)buf, off, size, in); ++ else ++ bcopy(in, buf + off, size); ++} ++ ++void ++crypto_copydata(int flags, caddr_t buf, int off, int size, caddr_t out) ++{ ++ ++ if ((flags & CRYPTO_F_SKBUF) != 0) ++ skb_copy_bits((struct sk_buff *)buf, off, out, size); ++ else if ((flags & CRYPTO_F_IOV) != 0) ++ cuio_copydata((struct uio *)buf, off, size, out); ++ else ++ bcopy(buf + off, out, size); ++} ++ ++int ++crypto_apply(int flags, caddr_t buf, int off, int len, ++ int (*f)(void *, void *, u_int), void *arg) ++{ ++#if 0 ++ int error; ++ ++ if ((flags & CRYPTO_F_SKBUF) != 0) ++ error = XXXXXX((struct mbuf *)buf, off, len, f, arg); ++ else if ((flags & CRYPTO_F_IOV) != 0) ++ error = cuio_apply((struct uio *)buf, off, len, f, arg); ++ else ++ error = (*f)(arg, buf + off, len); ++ return (error); ++#else ++ KASSERT(0, ("crypto_apply not implemented!\n")); ++#endif ++ return 0; ++} ++ ++EXPORT_SYMBOL(crypto_copyback); ++EXPORT_SYMBOL(crypto_copydata); ++EXPORT_SYMBOL(crypto_apply); ++ +diff -Nur linux-2.6.30.orig/crypto/ocf/crypto.c linux-2.6.30/crypto/ocf/crypto.c +--- linux-2.6.30.orig/crypto/ocf/crypto.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/crypto.c 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,1741 @@ ++/*- ++ * Linux port done by David McCullough <david_mccullough@securecomputing.com> ++ * Copyright (C) 2006-2007 David McCullough ++ * Copyright (C) 2004-2005 Intel Corporation. ++ * The license and original author are listed below. ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * Copyright (c) 2002-2006 Sam Leffler. All rights reserved. ++ * ++ * modification, are permitted provided that the following conditions ++ * are met: ++ * 1. Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * 2. Redistributions in binary form must reproduce the above copyright ++ * notice, this list of conditions and the following disclaimer in the ++ * documentation and/or other materials provided with the distribution. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR ++ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES ++ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. ++ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, ++ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT ++ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, ++ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY ++ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT ++ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF ++ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. ++ */ ++ ++#if 0 ++#include <sys/cdefs.h> ++__FBSDID("$FreeBSD: src/sys/opencrypto/crypto.c,v 1.27 2007/03/21 03:42:51 sam Exp $"); ++#endif ++ ++/* ++ * Cryptographic Subsystem. ++ * ++ * This code is derived from the Openbsd Cryptographic Framework (OCF) ++ * that has the copyright shown below. Very little of the original ++ * code remains. ++ */ ++/*- ++ * The author of this code is Angelos D. Keromytis (angelos@cis.upenn.edu) ++ * ++ * This code was written by Angelos D. Keromytis in Athens, Greece, in ++ * February 2000. Network Security Technologies Inc. (NSTI) kindly ++ * supported the development of this code. ++ * ++ * Copyright (c) 2000, 2001 Angelos D. Keromytis ++ * ++ * Permission to use, copy, and modify this software with or without fee ++ * is hereby granted, provided that this entire notice is included in ++ * all source code copies of any software which is or includes a copy or ++ * modification of this software. ++ * ++ * THIS SOFTWARE IS BEING PROVIDED "AS IS", WITHOUT ANY EXPRESS OR ++ * IMPLIED WARRANTY. IN PARTICULAR, NONE OF THE AUTHORS MAKES ANY ++ * REPRESENTATION OR WARRANTY OF ANY KIND CONCERNING THE ++ * MERCHANTABILITY OF THIS SOFTWARE OR ITS FITNESS FOR ANY PARTICULAR ++ * PURPOSE. ++ * ++__FBSDID("$FreeBSD: src/sys/opencrypto/crypto.c,v 1.16 2005/01/07 02:29:16 imp Exp $"); ++ */ ++ ++ ++#ifndef AUTOCONF_INCLUDED ++#include <linux/config.h> ++#endif ++#include <linux/module.h> ++#include <linux/init.h> ++#include <linux/list.h> ++#include <linux/slab.h> ++#include <linux/wait.h> ++#include <linux/sched.h> ++#include <linux/spinlock.h> ++#include <linux/version.h> ++#include <cryptodev.h> ++ ++/* ++ * keep track of whether or not we have been initialised, a big ++ * issue if we are linked into the kernel and a driver gets started before ++ * us ++ */ ++static int crypto_initted = 0; ++ ++/* ++ * Crypto drivers register themselves by allocating a slot in the ++ * crypto_drivers table with crypto_get_driverid() and then registering ++ * each algorithm they support with crypto_register() and crypto_kregister(). ++ */ ++ ++/* ++ * lock on driver table ++ * we track its state as spin_is_locked does not do anything on non-SMP boxes ++ */ ++static spinlock_t crypto_drivers_lock; ++static int crypto_drivers_locked; /* for non-SMP boxes */ ++ ++#define CRYPTO_DRIVER_LOCK() \ ++ ({ \ ++ spin_lock_irqsave(&crypto_drivers_lock, d_flags); \ ++ crypto_drivers_locked = 1; \ ++ dprintk("%s,%d: DRIVER_LOCK()\n", __FILE__, __LINE__); \ ++ }) ++#define CRYPTO_DRIVER_UNLOCK() \ ++ ({ \ ++ dprintk("%s,%d: DRIVER_UNLOCK()\n", __FILE__, __LINE__); \ ++ crypto_drivers_locked = 0; \ ++ spin_unlock_irqrestore(&crypto_drivers_lock, d_flags); \ ++ }) ++#define CRYPTO_DRIVER_ASSERT() \ ++ ({ \ ++ if (!crypto_drivers_locked) { \ ++ dprintk("%s,%d: DRIVER_ASSERT!\n", __FILE__, __LINE__); \ ++ } \ ++ }) ++ ++/* ++ * Crypto device/driver capabilities structure. ++ * ++ * Synchronization: ++ * (d) - protected by CRYPTO_DRIVER_LOCK() ++ * (q) - protected by CRYPTO_Q_LOCK() ++ * Not tagged fields are read-only. ++ */ ++struct cryptocap { ++ device_t cc_dev; /* (d) device/driver */ ++ u_int32_t cc_sessions; /* (d) # of sessions */ ++ u_int32_t cc_koperations; /* (d) # os asym operations */ ++ /* ++ * Largest possible operator length (in bits) for each type of ++ * encryption algorithm. XXX not used ++ */ ++ u_int16_t cc_max_op_len[CRYPTO_ALGORITHM_MAX + 1]; ++ u_int8_t cc_alg[CRYPTO_ALGORITHM_MAX + 1]; ++ u_int8_t cc_kalg[CRK_ALGORITHM_MAX + 1]; ++ ++ int cc_flags; /* (d) flags */ ++#define CRYPTOCAP_F_CLEANUP 0x80000000 /* needs resource cleanup */ ++ int cc_qblocked; /* (q) symmetric q blocked */ ++ int cc_kqblocked; /* (q) asymmetric q blocked */ ++}; ++static struct cryptocap *crypto_drivers = NULL; ++static int crypto_drivers_num = 0; ++ ++/* ++ * There are two queues for crypto requests; one for symmetric (e.g. ++ * cipher) operations and one for asymmetric (e.g. MOD)operations. ++ * A single mutex is used to lock access to both queues. We could ++ * have one per-queue but having one simplifies handling of block/unblock ++ * operations. ++ */ ++static int crp_sleep = 0; ++static LIST_HEAD(crp_q); /* request queues */ ++static LIST_HEAD(crp_kq); ++ ++static spinlock_t crypto_q_lock; ++ ++int crypto_all_qblocked = 0; /* protect with Q_LOCK */ ++module_param(crypto_all_qblocked, int, 0444); ++MODULE_PARM_DESC(crypto_all_qblocked, "Are all crypto queues blocked"); ++ ++int crypto_all_kqblocked = 0; /* protect with Q_LOCK */ ++module_param(crypto_all_kqblocked, int, 0444); ++MODULE_PARM_DESC(crypto_all_kqblocked, "Are all asym crypto queues blocked"); ++ ++#define CRYPTO_Q_LOCK() \ ++ ({ \ ++ spin_lock_irqsave(&crypto_q_lock, q_flags); \ ++ dprintk("%s,%d: Q_LOCK()\n", __FILE__, __LINE__); \ ++ }) ++#define CRYPTO_Q_UNLOCK() \ ++ ({ \ ++ dprintk("%s,%d: Q_UNLOCK()\n", __FILE__, __LINE__); \ ++ spin_unlock_irqrestore(&crypto_q_lock, q_flags); \ ++ }) ++ ++/* ++ * There are two queues for processing completed crypto requests; one ++ * for the symmetric and one for the asymmetric ops. We only need one ++ * but have two to avoid type futzing (cryptop vs. cryptkop). A single ++ * mutex is used to lock access to both queues. Note that this lock ++ * must be separate from the lock on request queues to insure driver ++ * callbacks don't generate lock order reversals. ++ */ ++static LIST_HEAD(crp_ret_q); /* callback queues */ ++static LIST_HEAD(crp_ret_kq); ++ ++static spinlock_t crypto_ret_q_lock; ++#define CRYPTO_RETQ_LOCK() \ ++ ({ \ ++ spin_lock_irqsave(&crypto_ret_q_lock, r_flags); \ ++ dprintk("%s,%d: RETQ_LOCK\n", __FILE__, __LINE__); \ ++ }) ++#define CRYPTO_RETQ_UNLOCK() \ ++ ({ \ ++ dprintk("%s,%d: RETQ_UNLOCK\n", __FILE__, __LINE__); \ ++ spin_unlock_irqrestore(&crypto_ret_q_lock, r_flags); \ ++ }) ++#define CRYPTO_RETQ_EMPTY() (list_empty(&crp_ret_q) && list_empty(&crp_ret_kq)) ++ ++#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,20) ++static kmem_cache_t *cryptop_zone; ++static kmem_cache_t *cryptodesc_zone; ++#else ++static struct kmem_cache *cryptop_zone; ++static struct kmem_cache *cryptodesc_zone; ++#endif ++ ++#define debug crypto_debug ++int crypto_debug = 0; ++module_param(crypto_debug, int, 0644); ++MODULE_PARM_DESC(crypto_debug, "Enable debug"); ++EXPORT_SYMBOL(crypto_debug); ++ ++/* ++ * Maximum number of outstanding crypto requests before we start ++ * failing requests. We need this to prevent DOS when too many ++ * requests are arriving for us to keep up. Otherwise we will ++ * run the system out of memory. Since crypto is slow, we are ++ * usually the bottleneck that needs to say, enough is enough. ++ * ++ * We cannot print errors when this condition occurs, we are already too ++ * slow, printing anything will just kill us ++ */ ++ ++static int crypto_q_cnt = 0; ++module_param(crypto_q_cnt, int, 0444); ++MODULE_PARM_DESC(crypto_q_cnt, ++ "Current number of outstanding crypto requests"); ++ ++static int crypto_q_max = 1000; ++module_param(crypto_q_max, int, 0644); ++MODULE_PARM_DESC(crypto_q_max, ++ "Maximum number of outstanding crypto requests"); ++ ++#define bootverbose crypto_verbose ++static int crypto_verbose = 0; ++module_param(crypto_verbose, int, 0644); ++MODULE_PARM_DESC(crypto_verbose, ++ "Enable verbose crypto startup"); ++ ++int crypto_usercrypto = 1; /* userland may do crypto reqs */ ++module_param(crypto_usercrypto, int, 0644); ++MODULE_PARM_DESC(crypto_usercrypto, ++ "Enable/disable user-mode access to crypto support"); ++ ++int crypto_userasymcrypto = 1; /* userland may do asym crypto reqs */ ++module_param(crypto_userasymcrypto, int, 0644); ++MODULE_PARM_DESC(crypto_userasymcrypto, ++ "Enable/disable user-mode access to asymmetric crypto support"); ++ ++int crypto_devallowsoft = 0; /* only use hardware crypto */ ++module_param(crypto_devallowsoft, int, 0644); ++MODULE_PARM_DESC(crypto_devallowsoft, ++ "Enable/disable use of software crypto support"); ++ ++static pid_t cryptoproc = (pid_t) -1; ++static struct completion cryptoproc_exited; ++static DECLARE_WAIT_QUEUE_HEAD(cryptoproc_wait); ++static pid_t cryptoretproc = (pid_t) -1; ++static struct completion cryptoretproc_exited; ++static DECLARE_WAIT_QUEUE_HEAD(cryptoretproc_wait); ++ ++static int crypto_proc(void *arg); ++static int crypto_ret_proc(void *arg); ++static int crypto_invoke(struct cryptocap *cap, struct cryptop *crp, int hint); ++static int crypto_kinvoke(struct cryptkop *krp, int flags); ++static void crypto_exit(void); ++static int crypto_init(void); ++ ++static struct cryptostats cryptostats; ++ ++static struct cryptocap * ++crypto_checkdriver(u_int32_t hid) ++{ ++ if (crypto_drivers == NULL) ++ return NULL; ++ return (hid >= crypto_drivers_num ? NULL : &crypto_drivers[hid]); ++} ++ ++/* ++ * Compare a driver's list of supported algorithms against another ++ * list; return non-zero if all algorithms are supported. ++ */ ++static int ++driver_suitable(const struct cryptocap *cap, const struct cryptoini *cri) ++{ ++ const struct cryptoini *cr; ++ ++ /* See if all the algorithms are supported. */ ++ for (cr = cri; cr; cr = cr->cri_next) ++ if (cap->cc_alg[cr->cri_alg] == 0) ++ return 0; ++ return 1; ++} ++ ++/* ++ * Select a driver for a new session that supports the specified ++ * algorithms and, optionally, is constrained according to the flags. ++ * The algorithm we use here is pretty stupid; just use the ++ * first driver that supports all the algorithms we need. If there ++ * are multiple drivers we choose the driver with the fewest active ++ * sessions. We prefer hardware-backed drivers to software ones. ++ * ++ * XXX We need more smarts here (in real life too, but that's ++ * XXX another story altogether). ++ */ ++static struct cryptocap * ++crypto_select_driver(const struct cryptoini *cri, int flags) ++{ ++ struct cryptocap *cap, *best; ++ int match, hid; ++ ++ CRYPTO_DRIVER_ASSERT(); ++ ++ /* ++ * Look first for hardware crypto devices if permitted. ++ */ ++ if (flags & CRYPTOCAP_F_HARDWARE) ++ match = CRYPTOCAP_F_HARDWARE; ++ else ++ match = CRYPTOCAP_F_SOFTWARE; ++ best = NULL; ++again: ++ for (hid = 0; hid < crypto_drivers_num; hid++) { ++ cap = &crypto_drivers[hid]; ++ /* ++ * If it's not initialized, is in the process of ++ * going away, or is not appropriate (hardware ++ * or software based on match), then skip. ++ */ ++ if (cap->cc_dev == NULL || ++ (cap->cc_flags & CRYPTOCAP_F_CLEANUP) || ++ (cap->cc_flags & match) == 0) ++ continue; ++ ++ /* verify all the algorithms are supported. */ ++ if (driver_suitable(cap, cri)) { ++ if (best == NULL || ++ cap->cc_sessions < best->cc_sessions) ++ best = cap; ++ } ++ } ++ if (best != NULL) ++ return best; ++ if (match == CRYPTOCAP_F_HARDWARE && (flags & CRYPTOCAP_F_SOFTWARE)) { ++ /* sort of an Algol 68-style for loop */ ++ match = CRYPTOCAP_F_SOFTWARE; ++ goto again; ++ } ++ return best; ++} ++ ++/* ++ * Create a new session. The crid argument specifies a crypto ++ * driver to use or constraints on a driver to select (hardware ++ * only, software only, either). Whatever driver is selected ++ * must be capable of the requested crypto algorithms. ++ */ ++int ++crypto_newsession(u_int64_t *sid, struct cryptoini *cri, int crid) ++{ ++ struct cryptocap *cap; ++ u_int32_t hid, lid; ++ int err; ++ unsigned long d_flags; ++ ++ CRYPTO_DRIVER_LOCK(); ++ if ((crid & (CRYPTOCAP_F_HARDWARE | CRYPTOCAP_F_SOFTWARE)) == 0) { ++ /* ++ * Use specified driver; verify it is capable. ++ */ ++ cap = crypto_checkdriver(crid); ++ if (cap != NULL && !driver_suitable(cap, cri)) ++ cap = NULL; ++ } else { ++ /* ++ * No requested driver; select based on crid flags. ++ */ ++ cap = crypto_select_driver(cri, crid); ++ /* ++ * if NULL then can't do everything in one session. ++ * XXX Fix this. We need to inject a "virtual" session ++ * XXX layer right about here. ++ */ ++ } ++ if (cap != NULL) { ++ /* Call the driver initialization routine. */ ++ hid = cap - crypto_drivers; ++ lid = hid; /* Pass the driver ID. */ ++ cap->cc_sessions++; ++ CRYPTO_DRIVER_UNLOCK(); ++ err = CRYPTODEV_NEWSESSION(cap->cc_dev, &lid, cri); ++ CRYPTO_DRIVER_LOCK(); ++ if (err == 0) { ++ (*sid) = (cap->cc_flags & 0xff000000) ++ | (hid & 0x00ffffff); ++ (*sid) <<= 32; ++ (*sid) |= (lid & 0xffffffff); ++ } else ++ cap->cc_sessions--; ++ } else ++ err = EINVAL; ++ CRYPTO_DRIVER_UNLOCK(); ++ return err; ++} ++ ++static void ++crypto_remove(struct cryptocap *cap) ++{ ++ CRYPTO_DRIVER_ASSERT(); ++ if (cap->cc_sessions == 0 && cap->cc_koperations == 0) ++ bzero(cap, sizeof(*cap)); ++} ++ ++/* ++ * Delete an existing session (or a reserved session on an unregistered ++ * driver). ++ */ ++int ++crypto_freesession(u_int64_t sid) ++{ ++ struct cryptocap *cap; ++ u_int32_t hid; ++ int err = 0; ++ unsigned long d_flags; ++ ++ dprintk("%s()\n", __FUNCTION__); ++ CRYPTO_DRIVER_LOCK(); ++ ++ if (crypto_drivers == NULL) { ++ err = EINVAL; ++ goto done; ++ } ++ ++ /* Determine two IDs. */ ++ hid = CRYPTO_SESID2HID(sid); ++ ++ if (hid >= crypto_drivers_num) { ++ dprintk("%s - INVALID DRIVER NUM %d\n", __FUNCTION__, hid); ++ err = ENOENT; ++ goto done; ++ } ++ cap = &crypto_drivers[hid]; ++ ++ if (cap->cc_dev) { ++ CRYPTO_DRIVER_UNLOCK(); ++ /* Call the driver cleanup routine, if available, unlocked. */ ++ err = CRYPTODEV_FREESESSION(cap->cc_dev, sid); ++ CRYPTO_DRIVER_LOCK(); ++ } ++ ++ if (cap->cc_sessions) ++ cap->cc_sessions--; ++ ++ if (cap->cc_flags & CRYPTOCAP_F_CLEANUP) ++ crypto_remove(cap); ++ ++done: ++ CRYPTO_DRIVER_UNLOCK(); ++ return err; ++} ++ ++/* ++ * Return an unused driver id. Used by drivers prior to registering ++ * support for the algorithms they handle. ++ */ ++int32_t ++crypto_get_driverid(device_t dev, int flags) ++{ ++ struct cryptocap *newdrv; ++ int i; ++ unsigned long d_flags; ++ ++ if ((flags & (CRYPTOCAP_F_HARDWARE | CRYPTOCAP_F_SOFTWARE)) == 0) { ++ printf("%s: no flags specified when registering driver\n", ++ device_get_nameunit(dev)); ++ return -1; ++ } ++ ++ CRYPTO_DRIVER_LOCK(); ++ ++ for (i = 0; i < crypto_drivers_num; i++) { ++ if (crypto_drivers[i].cc_dev == NULL && ++ (crypto_drivers[i].cc_flags & CRYPTOCAP_F_CLEANUP) == 0) { ++ break; ++ } ++ } ++ ++ /* Out of entries, allocate some more. */ ++ if (i == crypto_drivers_num) { ++ /* Be careful about wrap-around. */ ++ if (2 * crypto_drivers_num <= crypto_drivers_num) { ++ CRYPTO_DRIVER_UNLOCK(); ++ printk("crypto: driver count wraparound!\n"); ++ return -1; ++ } ++ ++ newdrv = kmalloc(2 * crypto_drivers_num * sizeof(struct cryptocap), ++ GFP_KERNEL); ++ if (newdrv == NULL) { ++ CRYPTO_DRIVER_UNLOCK(); ++ printk("crypto: no space to expand driver table!\n"); ++ return -1; ++ } ++ ++ memcpy(newdrv, crypto_drivers, ++ crypto_drivers_num * sizeof(struct cryptocap)); ++ memset(&newdrv[crypto_drivers_num], 0, ++ crypto_drivers_num * sizeof(struct cryptocap)); ++ ++ crypto_drivers_num *= 2; ++ ++ kfree(crypto_drivers); ++ crypto_drivers = newdrv; ++ } ++ ++ /* NB: state is zero'd on free */ ++ crypto_drivers[i].cc_sessions = 1; /* Mark */ ++ crypto_drivers[i].cc_dev = dev; ++ crypto_drivers[i].cc_flags = flags; ++ if (bootverbose) ++ printf("crypto: assign %s driver id %u, flags %u\n", ++ device_get_nameunit(dev), i, flags); ++ ++ CRYPTO_DRIVER_UNLOCK(); ++ ++ return i; ++} ++ ++/* ++ * Lookup a driver by name. We match against the full device ++ * name and unit, and against just the name. The latter gives ++ * us a simple widlcarding by device name. On success return the ++ * driver/hardware identifier; otherwise return -1. ++ */ ++int ++crypto_find_driver(const char *match) ++{ ++ int i, len = strlen(match); ++ unsigned long d_flags; ++ ++ CRYPTO_DRIVER_LOCK(); ++ for (i = 0; i < crypto_drivers_num; i++) { ++ device_t dev = crypto_drivers[i].cc_dev; ++ if (dev == NULL || ++ (crypto_drivers[i].cc_flags & CRYPTOCAP_F_CLEANUP)) ++ continue; ++ if (strncmp(match, device_get_nameunit(dev), len) == 0 || ++ strncmp(match, device_get_name(dev), len) == 0) ++ break; ++ } ++ CRYPTO_DRIVER_UNLOCK(); ++ return i < crypto_drivers_num ? i : -1; ++} ++ ++/* ++ * Return the device_t for the specified driver or NULL ++ * if the driver identifier is invalid. ++ */ ++device_t ++crypto_find_device_byhid(int hid) ++{ ++ struct cryptocap *cap = crypto_checkdriver(hid); ++ return cap != NULL ? cap->cc_dev : NULL; ++} ++ ++/* ++ * Return the device/driver capabilities. ++ */ ++int ++crypto_getcaps(int hid) ++{ ++ struct cryptocap *cap = crypto_checkdriver(hid); ++ return cap != NULL ? cap->cc_flags : 0; ++} ++ ++/* ++ * Register support for a key-related algorithm. This routine ++ * is called once for each algorithm supported a driver. ++ */ ++int ++crypto_kregister(u_int32_t driverid, int kalg, u_int32_t flags) ++{ ++ struct cryptocap *cap; ++ int err; ++ unsigned long d_flags; ++ ++ dprintk("%s()\n", __FUNCTION__); ++ CRYPTO_DRIVER_LOCK(); ++ ++ cap = crypto_checkdriver(driverid); ++ if (cap != NULL && ++ (CRK_ALGORITM_MIN <= kalg && kalg <= CRK_ALGORITHM_MAX)) { ++ /* ++ * XXX Do some performance testing to determine placing. ++ * XXX We probably need an auxiliary data structure that ++ * XXX describes relative performances. ++ */ ++ ++ cap->cc_kalg[kalg] = flags | CRYPTO_ALG_FLAG_SUPPORTED; ++ if (bootverbose) ++ printf("crypto: %s registers key alg %u flags %u\n" ++ , device_get_nameunit(cap->cc_dev) ++ , kalg ++ , flags ++ ); ++ err = 0; ++ } else ++ err = EINVAL; ++ ++ CRYPTO_DRIVER_UNLOCK(); ++ return err; ++} ++ ++/* ++ * Register support for a non-key-related algorithm. This routine ++ * is called once for each such algorithm supported by a driver. ++ */ ++int ++crypto_register(u_int32_t driverid, int alg, u_int16_t maxoplen, ++ u_int32_t flags) ++{ ++ struct cryptocap *cap; ++ int err; ++ unsigned long d_flags; ++ ++ dprintk("%s(id=0x%x, alg=%d, maxoplen=%d, flags=0x%x)\n", __FUNCTION__, ++ driverid, alg, maxoplen, flags); ++ ++ CRYPTO_DRIVER_LOCK(); ++ ++ cap = crypto_checkdriver(driverid); ++ /* NB: algorithms are in the range [1..max] */ ++ if (cap != NULL && ++ (CRYPTO_ALGORITHM_MIN <= alg && alg <= CRYPTO_ALGORITHM_MAX)) { ++ /* ++ * XXX Do some performance testing to determine placing. ++ * XXX We probably need an auxiliary data structure that ++ * XXX describes relative performances. ++ */ ++ ++ cap->cc_alg[alg] = flags | CRYPTO_ALG_FLAG_SUPPORTED; ++ cap->cc_max_op_len[alg] = maxoplen; ++ if (bootverbose) ++ printf("crypto: %s registers alg %u flags %u maxoplen %u\n" ++ , device_get_nameunit(cap->cc_dev) ++ , alg ++ , flags ++ , maxoplen ++ ); ++ cap->cc_sessions = 0; /* Unmark */ ++ err = 0; ++ } else ++ err = EINVAL; ++ ++ CRYPTO_DRIVER_UNLOCK(); ++ return err; ++} ++ ++static void ++driver_finis(struct cryptocap *cap) ++{ ++ u_int32_t ses, kops; ++ ++ CRYPTO_DRIVER_ASSERT(); ++ ++ ses = cap->cc_sessions; ++ kops = cap->cc_koperations; ++ bzero(cap, sizeof(*cap)); ++ if (ses != 0 || kops != 0) { ++ /* ++ * If there are pending sessions, ++ * just mark as invalid. ++ */ ++ cap->cc_flags |= CRYPTOCAP_F_CLEANUP; ++ cap->cc_sessions = ses; ++ cap->cc_koperations = kops; ++ } ++} ++ ++/* ++ * Unregister a crypto driver. If there are pending sessions using it, ++ * leave enough information around so that subsequent calls using those ++ * sessions will correctly detect the driver has been unregistered and ++ * reroute requests. ++ */ ++int ++crypto_unregister(u_int32_t driverid, int alg) ++{ ++ struct cryptocap *cap; ++ int i, err; ++ unsigned long d_flags; ++ ++ dprintk("%s()\n", __FUNCTION__); ++ CRYPTO_DRIVER_LOCK(); ++ ++ cap = crypto_checkdriver(driverid); ++ if (cap != NULL && ++ (CRYPTO_ALGORITHM_MIN <= alg && alg <= CRYPTO_ALGORITHM_MAX) && ++ cap->cc_alg[alg] != 0) { ++ cap->cc_alg[alg] = 0; ++ cap->cc_max_op_len[alg] = 0; ++ ++ /* Was this the last algorithm ? */ ++ for (i = 1; i <= CRYPTO_ALGORITHM_MAX; i++) ++ if (cap->cc_alg[i] != 0) ++ break; ++ ++ if (i == CRYPTO_ALGORITHM_MAX + 1) ++ driver_finis(cap); ++ err = 0; ++ } else ++ err = EINVAL; ++ CRYPTO_DRIVER_UNLOCK(); ++ return err; ++} ++ ++/* ++ * Unregister all algorithms associated with a crypto driver. ++ * If there are pending sessions using it, leave enough information ++ * around so that subsequent calls using those sessions will ++ * correctly detect the driver has been unregistered and reroute ++ * requests. ++ */ ++int ++crypto_unregister_all(u_int32_t driverid) ++{ ++ struct cryptocap *cap; ++ int err; ++ unsigned long d_flags; ++ ++ dprintk("%s()\n", __FUNCTION__); ++ CRYPTO_DRIVER_LOCK(); ++ cap = crypto_checkdriver(driverid); ++ if (cap != NULL) { ++ driver_finis(cap); ++ err = 0; ++ } else ++ err = EINVAL; ++ CRYPTO_DRIVER_UNLOCK(); ++ ++ return err; ++} ++ ++/* ++ * Clear blockage on a driver. The what parameter indicates whether ++ * the driver is now ready for cryptop's and/or cryptokop's. ++ */ ++int ++crypto_unblock(u_int32_t driverid, int what) ++{ ++ struct cryptocap *cap; ++ int err; ++ unsigned long q_flags; ++ ++ CRYPTO_Q_LOCK(); ++ cap = crypto_checkdriver(driverid); ++ if (cap != NULL) { ++ if (what & CRYPTO_SYMQ) { ++ cap->cc_qblocked = 0; ++ crypto_all_qblocked = 0; ++ } ++ if (what & CRYPTO_ASYMQ) { ++ cap->cc_kqblocked = 0; ++ crypto_all_kqblocked = 0; ++ } ++ if (crp_sleep) ++ wake_up_interruptible(&cryptoproc_wait); ++ err = 0; ++ } else ++ err = EINVAL; ++ CRYPTO_Q_UNLOCK(); //DAVIDM should this be a driver lock ++ ++ return err; ++} ++ ++/* ++ * Add a crypto request to a queue, to be processed by the kernel thread. ++ */ ++int ++crypto_dispatch(struct cryptop *crp) ++{ ++ struct cryptocap *cap; ++ int result = -1; ++ unsigned long q_flags; ++ ++ dprintk("%s()\n", __FUNCTION__); ++ ++ cryptostats.cs_ops++; ++ ++ CRYPTO_Q_LOCK(); ++ if (crypto_q_cnt >= crypto_q_max) { ++ CRYPTO_Q_UNLOCK(); ++ cryptostats.cs_drops++; ++ return ENOMEM; ++ } ++ crypto_q_cnt++; ++ ++ /* ++ * Caller marked the request to be processed immediately; dispatch ++ * it directly to the driver unless the driver is currently blocked. ++ */ ++ if ((crp->crp_flags & CRYPTO_F_BATCH) == 0) { ++ int hid = CRYPTO_SESID2HID(crp->crp_sid); ++ cap = crypto_checkdriver(hid); ++ /* Driver cannot disappear when there is an active session. */ ++ KASSERT(cap != NULL, ("%s: Driver disappeared.", __func__)); ++ if (!cap->cc_qblocked) { ++ crypto_all_qblocked = 0; ++ crypto_drivers[hid].cc_qblocked = 1; ++ CRYPTO_Q_UNLOCK(); ++ result = crypto_invoke(cap, crp, 0); ++ CRYPTO_Q_LOCK(); ++ if (result != ERESTART) ++ crypto_drivers[hid].cc_qblocked = 0; ++ } ++ } ++ if (result == ERESTART) { ++ /* ++ * The driver ran out of resources, mark the ++ * driver ``blocked'' for cryptop's and put ++ * the request back in the queue. It would ++ * best to put the request back where we got ++ * it but that's hard so for now we put it ++ * at the front. This should be ok; putting ++ * it at the end does not work. ++ */ ++ list_add(&crp->crp_next, &crp_q); ++ cryptostats.cs_blocks++; ++ } else if (result == -1) { ++ TAILQ_INSERT_TAIL(&crp_q, crp, crp_next); ++ } ++ if (crp_sleep) ++ wake_up_interruptible(&cryptoproc_wait); ++ CRYPTO_Q_UNLOCK(); ++ return 0; ++} ++ ++/* ++ * Add an asymetric crypto request to a queue, ++ * to be processed by the kernel thread. ++ */ ++int ++crypto_kdispatch(struct cryptkop *krp) ++{ ++ int error; ++ unsigned long q_flags; ++ ++ cryptostats.cs_kops++; ++ ++ error = crypto_kinvoke(krp, krp->krp_crid); ++ if (error == ERESTART) { ++ CRYPTO_Q_LOCK(); ++ TAILQ_INSERT_TAIL(&crp_kq, krp, krp_next); ++ if (crp_sleep) ++ wake_up_interruptible(&cryptoproc_wait); ++ CRYPTO_Q_UNLOCK(); ++ error = 0; ++ } ++ return error; ++} ++ ++/* ++ * Verify a driver is suitable for the specified operation. ++ */ ++static __inline int ++kdriver_suitable(const struct cryptocap *cap, const struct cryptkop *krp) ++{ ++ return (cap->cc_kalg[krp->krp_op] & CRYPTO_ALG_FLAG_SUPPORTED) != 0; ++} ++ ++/* ++ * Select a driver for an asym operation. The driver must ++ * support the necessary algorithm. The caller can constrain ++ * which device is selected with the flags parameter. The ++ * algorithm we use here is pretty stupid; just use the first ++ * driver that supports the algorithms we need. If there are ++ * multiple suitable drivers we choose the driver with the ++ * fewest active operations. We prefer hardware-backed ++ * drivers to software ones when either may be used. ++ */ ++static struct cryptocap * ++crypto_select_kdriver(const struct cryptkop *krp, int flags) ++{ ++ struct cryptocap *cap, *best, *blocked; ++ int match, hid; ++ ++ CRYPTO_DRIVER_ASSERT(); ++ ++ /* ++ * Look first for hardware crypto devices if permitted. ++ */ ++ if (flags & CRYPTOCAP_F_HARDWARE) ++ match = CRYPTOCAP_F_HARDWARE; ++ else ++ match = CRYPTOCAP_F_SOFTWARE; ++ best = NULL; ++ blocked = NULL; ++again: ++ for (hid = 0; hid < crypto_drivers_num; hid++) { ++ cap = &crypto_drivers[hid]; ++ /* ++ * If it's not initialized, is in the process of ++ * going away, or is not appropriate (hardware ++ * or software based on match), then skip. ++ */ ++ if (cap->cc_dev == NULL || ++ (cap->cc_flags & CRYPTOCAP_F_CLEANUP) || ++ (cap->cc_flags & match) == 0) ++ continue; ++ ++ /* verify all the algorithms are supported. */ ++ if (kdriver_suitable(cap, krp)) { ++ if (best == NULL || ++ cap->cc_koperations < best->cc_koperations) ++ best = cap; ++ } ++ } ++ if (best != NULL) ++ return best; ++ if (match == CRYPTOCAP_F_HARDWARE && (flags & CRYPTOCAP_F_SOFTWARE)) { ++ /* sort of an Algol 68-style for loop */ ++ match = CRYPTOCAP_F_SOFTWARE; ++ goto again; ++ } ++ return best; ++} ++ ++/* ++ * Dispatch an assymetric crypto request. ++ */ ++static int ++crypto_kinvoke(struct cryptkop *krp, int crid) ++{ ++ struct cryptocap *cap = NULL; ++ int error; ++ unsigned long d_flags; ++ ++ KASSERT(krp != NULL, ("%s: krp == NULL", __func__)); ++ KASSERT(krp->krp_callback != NULL, ++ ("%s: krp->crp_callback == NULL", __func__)); ++ ++ CRYPTO_DRIVER_LOCK(); ++ if ((crid & (CRYPTOCAP_F_HARDWARE | CRYPTOCAP_F_SOFTWARE)) == 0) { ++ cap = crypto_checkdriver(crid); ++ if (cap != NULL) { ++ /* ++ * Driver present, it must support the necessary ++ * algorithm and, if s/w drivers are excluded, ++ * it must be registered as hardware-backed. ++ */ ++ if (!kdriver_suitable(cap, krp) || ++ (!crypto_devallowsoft && ++ (cap->cc_flags & CRYPTOCAP_F_HARDWARE) == 0)) ++ cap = NULL; ++ } ++ } else { ++ /* ++ * No requested driver; select based on crid flags. ++ */ ++ if (!crypto_devallowsoft) /* NB: disallow s/w drivers */ ++ crid &= ~CRYPTOCAP_F_SOFTWARE; ++ cap = crypto_select_kdriver(krp, crid); ++ } ++ if (cap != NULL && !cap->cc_kqblocked) { ++ krp->krp_hid = cap - crypto_drivers; ++ cap->cc_koperations++; ++ CRYPTO_DRIVER_UNLOCK(); ++ error = CRYPTODEV_KPROCESS(cap->cc_dev, krp, 0); ++ CRYPTO_DRIVER_LOCK(); ++ if (error == ERESTART) { ++ cap->cc_koperations--; ++ CRYPTO_DRIVER_UNLOCK(); ++ return (error); ++ } ++ /* return the actual device used */ ++ krp->krp_crid = krp->krp_hid; ++ } else { ++ /* ++ * NB: cap is !NULL if device is blocked; in ++ * that case return ERESTART so the operation ++ * is resubmitted if possible. ++ */ ++ error = (cap == NULL) ? ENODEV : ERESTART; ++ } ++ CRYPTO_DRIVER_UNLOCK(); ++ ++ if (error) { ++ krp->krp_status = error; ++ crypto_kdone(krp); ++ } ++ return 0; ++} ++ ++ ++/* ++ * Dispatch a crypto request to the appropriate crypto devices. ++ */ ++static int ++crypto_invoke(struct cryptocap *cap, struct cryptop *crp, int hint) ++{ ++ KASSERT(crp != NULL, ("%s: crp == NULL", __func__)); ++ KASSERT(crp->crp_callback != NULL, ++ ("%s: crp->crp_callback == NULL", __func__)); ++ KASSERT(crp->crp_desc != NULL, ("%s: crp->crp_desc == NULL", __func__)); ++ ++ dprintk("%s()\n", __FUNCTION__); ++ ++#ifdef CRYPTO_TIMING ++ if (crypto_timing) ++ crypto_tstat(&cryptostats.cs_invoke, &crp->crp_tstamp); ++#endif ++ if (cap->cc_flags & CRYPTOCAP_F_CLEANUP) { ++ struct cryptodesc *crd; ++ u_int64_t nid; ++ ++ /* ++ * Driver has unregistered; migrate the session and return ++ * an error to the caller so they'll resubmit the op. ++ * ++ * XXX: What if there are more already queued requests for this ++ * session? ++ */ ++ crypto_freesession(crp->crp_sid); ++ ++ for (crd = crp->crp_desc; crd->crd_next; crd = crd->crd_next) ++ crd->CRD_INI.cri_next = &(crd->crd_next->CRD_INI); ++ ++ /* XXX propagate flags from initial session? */ ++ if (crypto_newsession(&nid, &(crp->crp_desc->CRD_INI), ++ CRYPTOCAP_F_HARDWARE | CRYPTOCAP_F_SOFTWARE) == 0) ++ crp->crp_sid = nid; ++ ++ crp->crp_etype = EAGAIN; ++ crypto_done(crp); ++ return 0; ++ } else { ++ /* ++ * Invoke the driver to process the request. ++ */ ++ return CRYPTODEV_PROCESS(cap->cc_dev, crp, hint); ++ } ++} ++ ++/* ++ * Release a set of crypto descriptors. ++ */ ++void ++crypto_freereq(struct cryptop *crp) ++{ ++ struct cryptodesc *crd; ++ ++ if (crp == NULL) ++ return; ++ ++#ifdef DIAGNOSTIC ++ { ++ struct cryptop *crp2; ++ unsigned long q_flags; ++ ++ CRYPTO_Q_LOCK(); ++ TAILQ_FOREACH(crp2, &crp_q, crp_next) { ++ KASSERT(crp2 != crp, ++ ("Freeing cryptop from the crypto queue (%p).", ++ crp)); ++ } ++ CRYPTO_Q_UNLOCK(); ++ CRYPTO_RETQ_LOCK(); ++ TAILQ_FOREACH(crp2, &crp_ret_q, crp_next) { ++ KASSERT(crp2 != crp, ++ ("Freeing cryptop from the return queue (%p).", ++ crp)); ++ } ++ CRYPTO_RETQ_UNLOCK(); ++ } ++#endif ++ ++ while ((crd = crp->crp_desc) != NULL) { ++ crp->crp_desc = crd->crd_next; ++ kmem_cache_free(cryptodesc_zone, crd); ++ } ++ kmem_cache_free(cryptop_zone, crp); ++} ++ ++/* ++ * Acquire a set of crypto descriptors. ++ */ ++struct cryptop * ++crypto_getreq(int num) ++{ ++ struct cryptodesc *crd; ++ struct cryptop *crp; ++ ++ crp = kmem_cache_alloc(cryptop_zone, SLAB_ATOMIC); ++ if (crp != NULL) { ++ memset(crp, 0, sizeof(*crp)); ++ INIT_LIST_HEAD(&crp->crp_next); ++ init_waitqueue_head(&crp->crp_waitq); ++ while (num--) { ++ crd = kmem_cache_alloc(cryptodesc_zone, SLAB_ATOMIC); ++ if (crd == NULL) { ++ crypto_freereq(crp); ++ return NULL; ++ } ++ memset(crd, 0, sizeof(*crd)); ++ crd->crd_next = crp->crp_desc; ++ crp->crp_desc = crd; ++ } ++ } ++ return crp; ++} ++ ++/* ++ * Invoke the callback on behalf of the driver. ++ */ ++void ++crypto_done(struct cryptop *crp) ++{ ++ unsigned long q_flags; ++ ++ dprintk("%s()\n", __FUNCTION__); ++ if ((crp->crp_flags & CRYPTO_F_DONE) == 0) { ++ crp->crp_flags |= CRYPTO_F_DONE; ++ CRYPTO_Q_LOCK(); ++ crypto_q_cnt--; ++ CRYPTO_Q_UNLOCK(); ++ } else ++ printk("crypto: crypto_done op already done, flags 0x%x", ++ crp->crp_flags); ++ if (crp->crp_etype != 0) ++ cryptostats.cs_errs++; ++ /* ++ * CBIMM means unconditionally do the callback immediately; ++ * CBIFSYNC means do the callback immediately only if the ++ * operation was done synchronously. Both are used to avoid ++ * doing extraneous context switches; the latter is mostly ++ * used with the software crypto driver. ++ */ ++ if ((crp->crp_flags & CRYPTO_F_CBIMM) || ++ ((crp->crp_flags & CRYPTO_F_CBIFSYNC) && ++ (CRYPTO_SESID2CAPS(crp->crp_sid) & CRYPTOCAP_F_SYNC))) { ++ /* ++ * Do the callback directly. This is ok when the ++ * callback routine does very little (e.g. the ++ * /dev/crypto callback method just does a wakeup). ++ */ ++ crp->crp_callback(crp); ++ } else { ++ unsigned long r_flags; ++ /* ++ * Normal case; queue the callback for the thread. ++ */ ++ CRYPTO_RETQ_LOCK(); ++ if (CRYPTO_RETQ_EMPTY()) ++ wake_up_interruptible(&cryptoretproc_wait);/* shared wait channel */ ++ TAILQ_INSERT_TAIL(&crp_ret_q, crp, crp_next); ++ CRYPTO_RETQ_UNLOCK(); ++ } ++} ++ ++/* ++ * Invoke the callback on behalf of the driver. ++ */ ++void ++crypto_kdone(struct cryptkop *krp) ++{ ++ struct cryptocap *cap; ++ unsigned long d_flags; ++ ++ if ((krp->krp_flags & CRYPTO_KF_DONE) != 0) ++ printk("crypto: crypto_kdone op already done, flags 0x%x", ++ krp->krp_flags); ++ krp->krp_flags |= CRYPTO_KF_DONE; ++ if (krp->krp_status != 0) ++ cryptostats.cs_kerrs++; ++ ++ CRYPTO_DRIVER_LOCK(); ++ /* XXX: What if driver is loaded in the meantime? */ ++ if (krp->krp_hid < crypto_drivers_num) { ++ cap = &crypto_drivers[krp->krp_hid]; ++ cap->cc_koperations--; ++ KASSERT(cap->cc_koperations >= 0, ("cc_koperations < 0")); ++ if (cap->cc_flags & CRYPTOCAP_F_CLEANUP) ++ crypto_remove(cap); ++ } ++ CRYPTO_DRIVER_UNLOCK(); ++ ++ /* ++ * CBIMM means unconditionally do the callback immediately; ++ * This is used to avoid doing extraneous context switches ++ */ ++ if ((krp->krp_flags & CRYPTO_KF_CBIMM)) { ++ /* ++ * Do the callback directly. This is ok when the ++ * callback routine does very little (e.g. the ++ * /dev/crypto callback method just does a wakeup). ++ */ ++ krp->krp_callback(krp); ++ } else { ++ unsigned long r_flags; ++ /* ++ * Normal case; queue the callback for the thread. ++ */ ++ CRYPTO_RETQ_LOCK(); ++ if (CRYPTO_RETQ_EMPTY()) ++ wake_up_interruptible(&cryptoretproc_wait);/* shared wait channel */ ++ TAILQ_INSERT_TAIL(&crp_ret_kq, krp, krp_next); ++ CRYPTO_RETQ_UNLOCK(); ++ } ++} ++ ++int ++crypto_getfeat(int *featp) ++{ ++ int hid, kalg, feat = 0; ++ unsigned long d_flags; ++ ++ CRYPTO_DRIVER_LOCK(); ++ for (hid = 0; hid < crypto_drivers_num; hid++) { ++ const struct cryptocap *cap = &crypto_drivers[hid]; ++ ++ if ((cap->cc_flags & CRYPTOCAP_F_SOFTWARE) && ++ !crypto_devallowsoft) { ++ continue; ++ } ++ for (kalg = 0; kalg < CRK_ALGORITHM_MAX; kalg++) ++ if (cap->cc_kalg[kalg] & CRYPTO_ALG_FLAG_SUPPORTED) ++ feat |= 1 << kalg; ++ } ++ CRYPTO_DRIVER_UNLOCK(); ++ *featp = feat; ++ return (0); ++} ++ ++/* ++ * Crypto thread, dispatches crypto requests. ++ */ ++static int ++crypto_proc(void *arg) ++{ ++ struct cryptop *crp, *submit; ++ struct cryptkop *krp, *krpp; ++ struct cryptocap *cap; ++ u_int32_t hid; ++ int result, hint; ++ unsigned long q_flags; ++ ++ ocf_daemonize("crypto"); ++ ++ CRYPTO_Q_LOCK(); ++ for (;;) { ++ /* ++ * we need to make sure we don't get into a busy loop with nothing ++ * to do, the two crypto_all_*blocked vars help us find out when ++ * we are all full and can do nothing on any driver or Q. If so we ++ * wait for an unblock. ++ */ ++ crypto_all_qblocked = !list_empty(&crp_q); ++ ++ /* ++ * Find the first element in the queue that can be ++ * processed and look-ahead to see if multiple ops ++ * are ready for the same driver. ++ */ ++ submit = NULL; ++ hint = 0; ++ list_for_each_entry(crp, &crp_q, crp_next) { ++ hid = CRYPTO_SESID2HID(crp->crp_sid); ++ cap = crypto_checkdriver(hid); ++ /* ++ * Driver cannot disappear when there is an active ++ * session. ++ */ ++ KASSERT(cap != NULL, ("%s:%u Driver disappeared.", ++ __func__, __LINE__)); ++ if (cap == NULL || cap->cc_dev == NULL) { ++ /* Op needs to be migrated, process it. */ ++ if (submit == NULL) ++ submit = crp; ++ break; ++ } ++ if (!cap->cc_qblocked) { ++ if (submit != NULL) { ++ /* ++ * We stop on finding another op, ++ * regardless whether its for the same ++ * driver or not. We could keep ++ * searching the queue but it might be ++ * better to just use a per-driver ++ * queue instead. ++ */ ++ if (CRYPTO_SESID2HID(submit->crp_sid) == hid) ++ hint = CRYPTO_HINT_MORE; ++ break; ++ } else { ++ submit = crp; ++ if ((submit->crp_flags & CRYPTO_F_BATCH) == 0) ++ break; ++ /* keep scanning for more are q'd */ ++ } ++ } ++ } ++ if (submit != NULL) { ++ hid = CRYPTO_SESID2HID(submit->crp_sid); ++ crypto_all_qblocked = 0; ++ list_del(&submit->crp_next); ++ crypto_drivers[hid].cc_qblocked = 1; ++ cap = crypto_checkdriver(hid); ++ CRYPTO_Q_UNLOCK(); ++ KASSERT(cap != NULL, ("%s:%u Driver disappeared.", ++ __func__, __LINE__)); ++ result = crypto_invoke(cap, submit, hint); ++ CRYPTO_Q_LOCK(); ++ if (result == ERESTART) { ++ /* ++ * The driver ran out of resources, mark the ++ * driver ``blocked'' for cryptop's and put ++ * the request back in the queue. It would ++ * best to put the request back where we got ++ * it but that's hard so for now we put it ++ * at the front. This should be ok; putting ++ * it at the end does not work. ++ */ ++ /* XXX validate sid again? */ ++ list_add(&submit->crp_next, &crp_q); ++ cryptostats.cs_blocks++; ++ } else ++ crypto_drivers[hid].cc_qblocked=0; ++ } ++ ++ crypto_all_kqblocked = !list_empty(&crp_kq); ++ ++ /* As above, but for key ops */ ++ krp = NULL; ++ list_for_each_entry(krpp, &crp_kq, krp_next) { ++ cap = crypto_checkdriver(krpp->krp_hid); ++ if (cap == NULL || cap->cc_dev == NULL) { ++ /* ++ * Operation needs to be migrated, invalidate ++ * the assigned device so it will reselect a ++ * new one below. Propagate the original ++ * crid selection flags if supplied. ++ */ ++ krp->krp_hid = krp->krp_crid & ++ (CRYPTOCAP_F_SOFTWARE|CRYPTOCAP_F_HARDWARE); ++ if (krp->krp_hid == 0) ++ krp->krp_hid = ++ CRYPTOCAP_F_SOFTWARE|CRYPTOCAP_F_HARDWARE; ++ break; ++ } ++ if (!cap->cc_kqblocked) { ++ krp = krpp; ++ break; ++ } ++ } ++ if (krp != NULL) { ++ crypto_all_kqblocked = 0; ++ list_del(&krp->krp_next); ++ crypto_drivers[krp->krp_hid].cc_kqblocked = 1; ++ CRYPTO_Q_UNLOCK(); ++ result = crypto_kinvoke(krp, krp->krp_hid); ++ CRYPTO_Q_LOCK(); ++ if (result == ERESTART) { ++ /* ++ * The driver ran out of resources, mark the ++ * driver ``blocked'' for cryptkop's and put ++ * the request back in the queue. It would ++ * best to put the request back where we got ++ * it but that's hard so for now we put it ++ * at the front. This should be ok; putting ++ * it at the end does not work. ++ */ ++ /* XXX validate sid again? */ ++ list_add(&krp->krp_next, &crp_kq); ++ cryptostats.cs_kblocks++; ++ } else ++ crypto_drivers[krp->krp_hid].cc_kqblocked = 0; ++ } ++ ++ if (submit == NULL && krp == NULL) { ++ /* ++ * Nothing more to be processed. Sleep until we're ++ * woken because there are more ops to process. ++ * This happens either by submission or by a driver ++ * becoming unblocked and notifying us through ++ * crypto_unblock. Note that when we wakeup we ++ * start processing each queue again from the ++ * front. It's not clear that it's important to ++ * preserve this ordering since ops may finish ++ * out of order if dispatched to different devices ++ * and some become blocked while others do not. ++ */ ++ dprintk("%s - sleeping (qe=%d qb=%d kqe=%d kqb=%d)\n", ++ __FUNCTION__, ++ list_empty(&crp_q), crypto_all_qblocked, ++ list_empty(&crp_kq), crypto_all_kqblocked); ++ CRYPTO_Q_UNLOCK(); ++ crp_sleep = 1; ++ wait_event_interruptible(cryptoproc_wait, ++ !(list_empty(&crp_q) || crypto_all_qblocked) || ++ !(list_empty(&crp_kq) || crypto_all_kqblocked) || ++ cryptoproc == (pid_t) -1); ++ crp_sleep = 0; ++ if (signal_pending (current)) { ++#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,0) ++ spin_lock_irq(¤t->sigmask_lock); ++#endif ++ flush_signals(current); ++#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,0) ++ spin_unlock_irq(¤t->sigmask_lock); ++#endif ++ } ++ CRYPTO_Q_LOCK(); ++ dprintk("%s - awake\n", __FUNCTION__); ++ if (cryptoproc == (pid_t) -1) ++ break; ++ cryptostats.cs_intrs++; ++ } ++ } ++ CRYPTO_Q_UNLOCK(); ++ complete_and_exit(&cryptoproc_exited, 0); ++} ++ ++/* ++ * Crypto returns thread, does callbacks for processed crypto requests. ++ * Callbacks are done here, rather than in the crypto drivers, because ++ * callbacks typically are expensive and would slow interrupt handling. ++ */ ++static int ++crypto_ret_proc(void *arg) ++{ ++ struct cryptop *crpt; ++ struct cryptkop *krpt; ++ unsigned long r_flags; ++ ++ ocf_daemonize("crypto_ret"); ++ ++ CRYPTO_RETQ_LOCK(); ++ for (;;) { ++ /* Harvest return q's for completed ops */ ++ crpt = NULL; ++ if (!list_empty(&crp_ret_q)) ++ crpt = list_entry(crp_ret_q.next, typeof(*crpt), crp_next); ++ if (crpt != NULL) ++ list_del(&crpt->crp_next); ++ ++ krpt = NULL; ++ if (!list_empty(&crp_ret_kq)) ++ krpt = list_entry(crp_ret_kq.next, typeof(*krpt), krp_next); ++ if (krpt != NULL) ++ list_del(&krpt->krp_next); ++ ++ if (crpt != NULL || krpt != NULL) { ++ CRYPTO_RETQ_UNLOCK(); ++ /* ++ * Run callbacks unlocked. ++ */ ++ if (crpt != NULL) ++ crpt->crp_callback(crpt); ++ if (krpt != NULL) ++ krpt->krp_callback(krpt); ++ CRYPTO_RETQ_LOCK(); ++ } else { ++ /* ++ * Nothing more to be processed. Sleep until we're ++ * woken because there are more returns to process. ++ */ ++ dprintk("%s - sleeping\n", __FUNCTION__); ++ CRYPTO_RETQ_UNLOCK(); ++ wait_event_interruptible(cryptoretproc_wait, ++ cryptoretproc == (pid_t) -1 || ++ !list_empty(&crp_ret_q) || ++ !list_empty(&crp_ret_kq)); ++ if (signal_pending (current)) { ++#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,0) ++ spin_lock_irq(¤t->sigmask_lock); ++#endif ++ flush_signals(current); ++#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,0) ++ spin_unlock_irq(¤t->sigmask_lock); ++#endif ++ } ++ CRYPTO_RETQ_LOCK(); ++ dprintk("%s - awake\n", __FUNCTION__); ++ if (cryptoretproc == (pid_t) -1) { ++ dprintk("%s - EXITING!\n", __FUNCTION__); ++ break; ++ } ++ cryptostats.cs_rets++; ++ } ++ } ++ CRYPTO_RETQ_UNLOCK(); ++ complete_and_exit(&cryptoretproc_exited, 0); ++} ++ ++ ++#if 0 /* should put this into /proc or something */ ++static void ++db_show_drivers(void) ++{ ++ int hid; ++ ++ db_printf("%12s %4s %4s %8s %2s %2s\n" ++ , "Device" ++ , "Ses" ++ , "Kops" ++ , "Flags" ++ , "QB" ++ , "KB" ++ ); ++ for (hid = 0; hid < crypto_drivers_num; hid++) { ++ const struct cryptocap *cap = &crypto_drivers[hid]; ++ if (cap->cc_dev == NULL) ++ continue; ++ db_printf("%-12s %4u %4u %08x %2u %2u\n" ++ , device_get_nameunit(cap->cc_dev) ++ , cap->cc_sessions ++ , cap->cc_koperations ++ , cap->cc_flags ++ , cap->cc_qblocked ++ , cap->cc_kqblocked ++ ); ++ } ++} ++ ++DB_SHOW_COMMAND(crypto, db_show_crypto) ++{ ++ struct cryptop *crp; ++ ++ db_show_drivers(); ++ db_printf("\n"); ++ ++ db_printf("%4s %8s %4s %4s %4s %4s %8s %8s\n", ++ "HID", "Caps", "Ilen", "Olen", "Etype", "Flags", ++ "Desc", "Callback"); ++ TAILQ_FOREACH(crp, &crp_q, crp_next) { ++ db_printf("%4u %08x %4u %4u %4u %04x %8p %8p\n" ++ , (int) CRYPTO_SESID2HID(crp->crp_sid) ++ , (int) CRYPTO_SESID2CAPS(crp->crp_sid) ++ , crp->crp_ilen, crp->crp_olen ++ , crp->crp_etype ++ , crp->crp_flags ++ , crp->crp_desc ++ , crp->crp_callback ++ ); ++ } ++ if (!TAILQ_EMPTY(&crp_ret_q)) { ++ db_printf("\n%4s %4s %4s %8s\n", ++ "HID", "Etype", "Flags", "Callback"); ++ TAILQ_FOREACH(crp, &crp_ret_q, crp_next) { ++ db_printf("%4u %4u %04x %8p\n" ++ , (int) CRYPTO_SESID2HID(crp->crp_sid) ++ , crp->crp_etype ++ , crp->crp_flags ++ , crp->crp_callback ++ ); ++ } ++ } ++} ++ ++DB_SHOW_COMMAND(kcrypto, db_show_kcrypto) ++{ ++ struct cryptkop *krp; ++ ++ db_show_drivers(); ++ db_printf("\n"); ++ ++ db_printf("%4s %5s %4s %4s %8s %4s %8s\n", ++ "Op", "Status", "#IP", "#OP", "CRID", "HID", "Callback"); ++ TAILQ_FOREACH(krp, &crp_kq, krp_next) { ++ db_printf("%4u %5u %4u %4u %08x %4u %8p\n" ++ , krp->krp_op ++ , krp->krp_status ++ , krp->krp_iparams, krp->krp_oparams ++ , krp->krp_crid, krp->krp_hid ++ , krp->krp_callback ++ ); ++ } ++ if (!TAILQ_EMPTY(&crp_ret_q)) { ++ db_printf("%4s %5s %8s %4s %8s\n", ++ "Op", "Status", "CRID", "HID", "Callback"); ++ TAILQ_FOREACH(krp, &crp_ret_kq, krp_next) { ++ db_printf("%4u %5u %08x %4u %8p\n" ++ , krp->krp_op ++ , krp->krp_status ++ , krp->krp_crid, krp->krp_hid ++ , krp->krp_callback ++ ); ++ } ++ } ++} ++#endif ++ ++ ++static int ++crypto_init(void) ++{ ++ int error; ++ ++ dprintk("%s(0x%x)\n", __FUNCTION__, (int) crypto_init); ++ ++ if (crypto_initted) ++ return 0; ++ crypto_initted = 1; ++ ++ spin_lock_init(&crypto_drivers_lock); ++ spin_lock_init(&crypto_q_lock); ++ spin_lock_init(&crypto_ret_q_lock); ++ ++ cryptop_zone = kmem_cache_create("cryptop", sizeof(struct cryptop), ++ 0, SLAB_HWCACHE_ALIGN, NULL ++#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,23) ++ , NULL ++#endif ++ ); ++ ++ cryptodesc_zone = kmem_cache_create("cryptodesc", sizeof(struct cryptodesc), ++ 0, SLAB_HWCACHE_ALIGN, NULL ++#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,23) ++ , NULL ++#endif ++ ); ++ ++ if (cryptodesc_zone == NULL || cryptop_zone == NULL) { ++ printk("crypto: crypto_init cannot setup crypto zones\n"); ++ error = ENOMEM; ++ goto bad; ++ } ++ ++ crypto_drivers_num = CRYPTO_DRIVERS_INITIAL; ++ crypto_drivers = kmalloc(crypto_drivers_num * sizeof(struct cryptocap), ++ GFP_KERNEL); ++ if (crypto_drivers == NULL) { ++ printk("crypto: crypto_init cannot setup crypto drivers\n"); ++ error = ENOMEM; ++ goto bad; ++ } ++ ++ memset(crypto_drivers, 0, crypto_drivers_num * sizeof(struct cryptocap)); ++ ++ init_completion(&cryptoproc_exited); ++ init_completion(&cryptoretproc_exited); ++ ++ cryptoproc = 0; /* to avoid race condition where proc runs first */ ++ cryptoproc = kernel_thread(crypto_proc, NULL, CLONE_FS|CLONE_FILES); ++ if (cryptoproc < 0) { ++ error = cryptoproc; ++ printk("crypto: crypto_init cannot start crypto thread; error %d", ++ error); ++ goto bad; ++ } ++ ++ cryptoretproc = 0; /* to avoid race condition where proc runs first */ ++ cryptoretproc = kernel_thread(crypto_ret_proc, NULL, CLONE_FS|CLONE_FILES); ++ if (cryptoretproc < 0) { ++ error = cryptoretproc; ++ printk("crypto: crypto_init cannot start cryptoret thread; error %d", ++ error); ++ goto bad; ++ } ++ ++ return 0; ++bad: ++ crypto_exit(); ++ return error; ++} ++ ++ ++static void ++crypto_exit(void) ++{ ++ pid_t p; ++ unsigned long d_flags; ++ ++ dprintk("%s()\n", __FUNCTION__); ++ ++ /* ++ * Terminate any crypto threads. ++ */ ++ ++ CRYPTO_DRIVER_LOCK(); ++ p = cryptoproc; ++ cryptoproc = (pid_t) -1; ++ kill_pid(p, SIGTERM, 1); ++ wake_up_interruptible(&cryptoproc_wait); ++ CRYPTO_DRIVER_UNLOCK(); ++ ++ wait_for_completion(&cryptoproc_exited); ++ ++ CRYPTO_DRIVER_LOCK(); ++ p = cryptoretproc; ++ cryptoretproc = (pid_t) -1; ++ kill_pid(p, SIGTERM, 1); ++ wake_up_interruptible(&cryptoretproc_wait); ++ CRYPTO_DRIVER_UNLOCK(); ++ ++ wait_for_completion(&cryptoretproc_exited); ++ ++ /* XXX flush queues??? */ ++ ++ /* ++ * Reclaim dynamically allocated resources. ++ */ ++ if (crypto_drivers != NULL) ++ kfree(crypto_drivers); ++ ++ if (cryptodesc_zone != NULL) ++ kmem_cache_destroy(cryptodesc_zone); ++ if (cryptop_zone != NULL) ++ kmem_cache_destroy(cryptop_zone); ++} ++ ++ ++EXPORT_SYMBOL(crypto_newsession); ++EXPORT_SYMBOL(crypto_freesession); ++EXPORT_SYMBOL(crypto_get_driverid); ++EXPORT_SYMBOL(crypto_kregister); ++EXPORT_SYMBOL(crypto_register); ++EXPORT_SYMBOL(crypto_unregister); ++EXPORT_SYMBOL(crypto_unregister_all); ++EXPORT_SYMBOL(crypto_unblock); ++EXPORT_SYMBOL(crypto_dispatch); ++EXPORT_SYMBOL(crypto_kdispatch); ++EXPORT_SYMBOL(crypto_freereq); ++EXPORT_SYMBOL(crypto_getreq); ++EXPORT_SYMBOL(crypto_done); ++EXPORT_SYMBOL(crypto_kdone); ++EXPORT_SYMBOL(crypto_getfeat); ++EXPORT_SYMBOL(crypto_userasymcrypto); ++EXPORT_SYMBOL(crypto_getcaps); ++EXPORT_SYMBOL(crypto_find_driver); ++EXPORT_SYMBOL(crypto_find_device_byhid); ++ ++module_init(crypto_init); ++module_exit(crypto_exit); ++ ++MODULE_LICENSE("BSD"); ++MODULE_AUTHOR("David McCullough <david_mccullough@securecomputing.com>"); ++MODULE_DESCRIPTION("OCF (OpenBSD Cryptographic Framework)"); +diff -Nur linux-2.6.30.orig/crypto/ocf/cryptodev.c linux-2.6.30/crypto/ocf/cryptodev.c +--- linux-2.6.30.orig/crypto/ocf/cryptodev.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/cryptodev.c 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,1048 @@ ++/* $OpenBSD: cryptodev.c,v 1.52 2002/06/19 07:22:46 deraadt Exp $ */ ++ ++/*- ++ * Linux port done by David McCullough <david_mccullough@securecomputing.com> ++ * Copyright (C) 2006-2007 David McCullough ++ * Copyright (C) 2004-2005 Intel Corporation. ++ * The license and original author are listed below. ++ * ++ * Copyright (c) 2001 Theo de Raadt ++ * Copyright (c) 2002-2006 Sam Leffler, Errno Consulting ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions ++ * are met: ++ * ++ * 1. Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * 2. Redistributions in binary form must reproduce the above copyright ++ * notice, this list of conditions and the following disclaimer in the ++ * documentation and/or other materials provided with the distribution. ++ * 3. The name of the author may not be used to endorse or promote products ++ * derived from this software without specific prior written permission. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR ++ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES ++ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. ++ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, ++ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT ++ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, ++ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY ++ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT ++ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF ++ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. ++ * ++ * Effort sponsored in part by the Defense Advanced Research Projects ++ * Agency (DARPA) and Air Force Research Laboratory, Air Force ++ * Materiel Command, USAF, under agreement number F30602-01-2-0537. ++ * ++__FBSDID("$FreeBSD: src/sys/opencrypto/cryptodev.c,v 1.34 2007/05/09 19:37:02 gnn Exp $"); ++ */ ++ ++#ifndef AUTOCONF_INCLUDED ++#include <linux/config.h> ++#endif ++#include <linux/types.h> ++#include <linux/time.h> ++#include <linux/delay.h> ++#include <linux/list.h> ++#include <linux/init.h> ++#include <linux/sched.h> ++#include <linux/unistd.h> ++#include <linux/module.h> ++#include <linux/wait.h> ++#include <linux/slab.h> ++#include <linux/fs.h> ++#include <linux/dcache.h> ++#include <linux/file.h> ++#include <linux/mount.h> ++#include <linux/miscdevice.h> ++#include <linux/version.h> ++#include <asm/uaccess.h> ++ ++#include <cryptodev.h> ++#include <uio.h> ++ ++extern asmlinkage long sys_dup(unsigned int fildes); ++ ++#define debug cryptodev_debug ++int cryptodev_debug = 0; ++module_param(cryptodev_debug, int, 0644); ++MODULE_PARM_DESC(cryptodev_debug, "Enable cryptodev debug"); ++ ++struct csession_info { ++ u_int16_t blocksize; ++ u_int16_t minkey, maxkey; ++ ++ u_int16_t keysize; ++ /* u_int16_t hashsize; */ ++ u_int16_t authsize; ++ /* u_int16_t ctxsize; */ ++}; ++ ++struct csession { ++ struct list_head list; ++ u_int64_t sid; ++ u_int32_t ses; ++ ++ wait_queue_head_t waitq; ++ ++ u_int32_t cipher; ++ ++ u_int32_t mac; ++ ++ caddr_t key; ++ int keylen; ++ u_char tmp_iv[EALG_MAX_BLOCK_LEN]; ++ ++ caddr_t mackey; ++ int mackeylen; ++ ++ struct csession_info info; ++ ++ struct iovec iovec; ++ struct uio uio; ++ int error; ++}; ++ ++struct fcrypt { ++ struct list_head csessions; ++ int sesn; ++}; ++ ++static struct csession *csefind(struct fcrypt *, u_int); ++static int csedelete(struct fcrypt *, struct csession *); ++static struct csession *cseadd(struct fcrypt *, struct csession *); ++static struct csession *csecreate(struct fcrypt *, u_int64_t, ++ struct cryptoini *crie, struct cryptoini *cria, struct csession_info *); ++static int csefree(struct csession *); ++ ++static int cryptodev_op(struct csession *, struct crypt_op *); ++static int cryptodev_key(struct crypt_kop *); ++static int cryptodev_find(struct crypt_find_op *); ++ ++static int cryptodev_cb(void *); ++static int cryptodev_open(struct inode *inode, struct file *filp); ++ ++/* ++ * Check a crypto identifier to see if it requested ++ * a valid crid and it's capabilities match. ++ */ ++static int ++checkcrid(int crid) ++{ ++ int hid = crid & ~(CRYPTOCAP_F_SOFTWARE | CRYPTOCAP_F_HARDWARE); ++ int typ = crid & (CRYPTOCAP_F_SOFTWARE | CRYPTOCAP_F_HARDWARE); ++ int caps = 0; ++ ++ /* if the user hasn't selected a driver, then just call newsession */ ++ if (hid == 0 && typ != 0) ++ return 0; ++ ++ caps = crypto_getcaps(hid); ++ ++ /* didn't find anything with capabilities */ ++ if (caps == 0) { ++ dprintk("%s: hid=%x typ=%x not matched\n", __FUNCTION__, hid, typ); ++ return EINVAL; ++ } ++ ++ /* the user didn't specify SW or HW, so the driver is ok */ ++ if (typ == 0) ++ return 0; ++ ++ /* if the type specified didn't match */ ++ if (typ != (caps & (CRYPTOCAP_F_SOFTWARE | CRYPTOCAP_F_HARDWARE))) { ++ dprintk("%s: hid=%x typ=%x caps=%x not matched\n", __FUNCTION__, ++ hid, typ, caps); ++ return EINVAL; ++ } ++ ++ return 0; ++} ++ ++static int ++cryptodev_op(struct csession *cse, struct crypt_op *cop) ++{ ++ struct cryptop *crp = NULL; ++ struct cryptodesc *crde = NULL, *crda = NULL; ++ int error = 0; ++ ++ dprintk("%s()\n", __FUNCTION__); ++ if (cop->len > CRYPTO_MAX_DATA_LEN) { ++ dprintk("%s: %d > %d\n", __FUNCTION__, cop->len, CRYPTO_MAX_DATA_LEN); ++ return (E2BIG); ++ } ++ ++ if (cse->info.blocksize && (cop->len % cse->info.blocksize) != 0) { ++ dprintk("%s: blocksize=%d len=%d\n", __FUNCTION__, cse->info.blocksize, ++ cop->len); ++ return (EINVAL); ++ } ++ ++ cse->uio.uio_iov = &cse->iovec; ++ cse->uio.uio_iovcnt = 1; ++ cse->uio.uio_offset = 0; ++#if 0 ++ cse->uio.uio_resid = cop->len; ++ cse->uio.uio_segflg = UIO_SYSSPACE; ++ cse->uio.uio_rw = UIO_WRITE; ++ cse->uio.uio_td = td; ++#endif ++ cse->uio.uio_iov[0].iov_len = cop->len; ++ if (cse->info.authsize) ++ cse->uio.uio_iov[0].iov_len += cse->info.authsize; ++ cse->uio.uio_iov[0].iov_base = kmalloc(cse->uio.uio_iov[0].iov_len, ++ GFP_KERNEL); ++ ++ if (cse->uio.uio_iov[0].iov_base == NULL) { ++ dprintk("%s: iov_base kmalloc(%d) failed\n", __FUNCTION__, ++ cse->uio.uio_iov[0].iov_len); ++ return (ENOMEM); ++ } ++ ++ crp = crypto_getreq((cse->info.blocksize != 0) + (cse->info.authsize != 0)); ++ if (crp == NULL) { ++ dprintk("%s: ENOMEM\n", __FUNCTION__); ++ error = ENOMEM; ++ goto bail; ++ } ++ ++ if (cse->info.authsize) { ++ crda = crp->crp_desc; ++ if (cse->info.blocksize) ++ crde = crda->crd_next; ++ } else { ++ if (cse->info.blocksize) ++ crde = crp->crp_desc; ++ else { ++ dprintk("%s: bad request\n", __FUNCTION__); ++ error = EINVAL; ++ goto bail; ++ } ++ } ++ ++ if ((error = copy_from_user(cse->uio.uio_iov[0].iov_base, cop->src, ++ cop->len))) { ++ dprintk("%s: bad copy\n", __FUNCTION__); ++ goto bail; ++ } ++ ++ if (crda) { ++ crda->crd_skip = 0; ++ crda->crd_len = cop->len; ++ crda->crd_inject = cop->len; ++ ++ crda->crd_alg = cse->mac; ++ crda->crd_key = cse->mackey; ++ crda->crd_klen = cse->mackeylen * 8; ++ } ++ ++ if (crde) { ++ if (cop->op == COP_ENCRYPT) ++ crde->crd_flags |= CRD_F_ENCRYPT; ++ else ++ crde->crd_flags &= ~CRD_F_ENCRYPT; ++ crde->crd_len = cop->len; ++ crde->crd_inject = 0; ++ ++ crde->crd_alg = cse->cipher; ++ crde->crd_key = cse->key; ++ crde->crd_klen = cse->keylen * 8; ++ } ++ ++ crp->crp_ilen = cse->uio.uio_iov[0].iov_len; ++ crp->crp_flags = CRYPTO_F_IOV | CRYPTO_F_CBIMM ++ | (cop->flags & COP_F_BATCH); ++ crp->crp_buf = (caddr_t)&cse->uio; ++ crp->crp_callback = (int (*) (struct cryptop *)) cryptodev_cb; ++ crp->crp_sid = cse->sid; ++ crp->crp_opaque = (void *)cse; ++ ++ if (cop->iv) { ++ if (crde == NULL) { ++ error = EINVAL; ++ dprintk("%s no crde\n", __FUNCTION__); ++ goto bail; ++ } ++ if (cse->cipher == CRYPTO_ARC4) { /* XXX use flag? */ ++ error = EINVAL; ++ dprintk("%s arc4 with IV\n", __FUNCTION__); ++ goto bail; ++ } ++ if ((error = copy_from_user(cse->tmp_iv, cop->iv, ++ cse->info.blocksize))) { ++ dprintk("%s bad iv copy\n", __FUNCTION__); ++ goto bail; ++ } ++ memcpy(crde->crd_iv, cse->tmp_iv, cse->info.blocksize); ++ crde->crd_flags |= CRD_F_IV_EXPLICIT | CRD_F_IV_PRESENT; ++ crde->crd_skip = 0; ++ } else if (cse->cipher == CRYPTO_ARC4) { /* XXX use flag? */ ++ crde->crd_skip = 0; ++ } else if (crde) { ++ crde->crd_flags |= CRD_F_IV_PRESENT; ++ crde->crd_skip = cse->info.blocksize; ++ crde->crd_len -= cse->info.blocksize; ++ } ++ ++ if (cop->mac && crda == NULL) { ++ error = EINVAL; ++ dprintk("%s no crda\n", __FUNCTION__); ++ goto bail; ++ } ++ ++ /* ++ * Let the dispatch run unlocked, then, interlock against the ++ * callback before checking if the operation completed and going ++ * to sleep. This insures drivers don't inherit our lock which ++ * results in a lock order reversal between crypto_dispatch forced ++ * entry and the crypto_done callback into us. ++ */ ++ error = crypto_dispatch(crp); ++ if (error == 0) { ++ dprintk("%s about to WAIT\n", __FUNCTION__); ++ /* ++ * we really need to wait for driver to complete to maintain ++ * state, luckily interrupts will be remembered ++ */ ++ do { ++ error = wait_event_interruptible(crp->crp_waitq, ++ ((crp->crp_flags & CRYPTO_F_DONE) != 0)); ++ /* ++ * we can't break out of this loop or we will leave behind ++ * a huge mess, however, staying here means if your driver ++ * is broken user applications can hang and not be killed. ++ * The solution, fix your driver :-) ++ */ ++ if (error) { ++ schedule(); ++ error = 0; ++ } ++ } while ((crp->crp_flags & CRYPTO_F_DONE) == 0); ++ dprintk("%s finished WAITING error=%d\n", __FUNCTION__, error); ++ } ++ ++ if (crp->crp_etype != 0) { ++ error = crp->crp_etype; ++ dprintk("%s error in crp processing\n", __FUNCTION__); ++ goto bail; ++ } ++ ++ if (cse->error) { ++ error = cse->error; ++ dprintk("%s error in cse processing\n", __FUNCTION__); ++ goto bail; ++ } ++ ++ if (cop->dst && (error = copy_to_user(cop->dst, ++ cse->uio.uio_iov[0].iov_base, cop->len))) { ++ dprintk("%s bad dst copy\n", __FUNCTION__); ++ goto bail; ++ } ++ ++ if (cop->mac && ++ (error=copy_to_user(cop->mac, ++ (caddr_t)cse->uio.uio_iov[0].iov_base + cop->len, ++ cse->info.authsize))) { ++ dprintk("%s bad mac copy\n", __FUNCTION__); ++ goto bail; ++ } ++ ++bail: ++ if (crp) ++ crypto_freereq(crp); ++ if (cse->uio.uio_iov[0].iov_base) ++ kfree(cse->uio.uio_iov[0].iov_base); ++ ++ return (error); ++} ++ ++static int ++cryptodev_cb(void *op) ++{ ++ struct cryptop *crp = (struct cryptop *) op; ++ struct csession *cse = (struct csession *)crp->crp_opaque; ++ int error; ++ ++ dprintk("%s()\n", __FUNCTION__); ++ error = crp->crp_etype; ++ if (error == EAGAIN) { ++ crp->crp_flags &= ~CRYPTO_F_DONE; ++#ifdef NOTYET ++ /* ++ * DAVIDM I am fairly sure that we should turn this into a batch ++ * request to stop bad karma/lockup, revisit ++ */ ++ crp->crp_flags |= CRYPTO_F_BATCH; ++#endif ++ return crypto_dispatch(crp); ++ } ++ if (error != 0 || (crp->crp_flags & CRYPTO_F_DONE)) { ++ cse->error = error; ++ wake_up_interruptible(&crp->crp_waitq); ++ } ++ return (0); ++} ++ ++static int ++cryptodevkey_cb(void *op) ++{ ++ struct cryptkop *krp = (struct cryptkop *) op; ++ dprintk("%s()\n", __FUNCTION__); ++ wake_up_interruptible(&krp->krp_waitq); ++ return (0); ++} ++ ++static int ++cryptodev_key(struct crypt_kop *kop) ++{ ++ struct cryptkop *krp = NULL; ++ int error = EINVAL; ++ int in, out, size, i; ++ ++ dprintk("%s()\n", __FUNCTION__); ++ if (kop->crk_iparams + kop->crk_oparams > CRK_MAXPARAM) { ++ dprintk("%s params too big\n", __FUNCTION__); ++ return (EFBIG); ++ } ++ ++ in = kop->crk_iparams; ++ out = kop->crk_oparams; ++ switch (kop->crk_op) { ++ case CRK_MOD_EXP: ++ if (in == 3 && out == 1) ++ break; ++ return (EINVAL); ++ case CRK_MOD_EXP_CRT: ++ if (in == 6 && out == 1) ++ break; ++ return (EINVAL); ++ case CRK_DSA_SIGN: ++ if (in == 5 && out == 2) ++ break; ++ return (EINVAL); ++ case CRK_DSA_VERIFY: ++ if (in == 7 && out == 0) ++ break; ++ return (EINVAL); ++ case CRK_DH_COMPUTE_KEY: ++ if (in == 3 && out == 1) ++ break; ++ return (EINVAL); ++ default: ++ return (EINVAL); ++ } ++ ++ krp = (struct cryptkop *)kmalloc(sizeof *krp, GFP_KERNEL); ++ if (!krp) ++ return (ENOMEM); ++ bzero(krp, sizeof *krp); ++ krp->krp_op = kop->crk_op; ++ krp->krp_status = kop->crk_status; ++ krp->krp_iparams = kop->crk_iparams; ++ krp->krp_oparams = kop->crk_oparams; ++ krp->krp_crid = kop->crk_crid; ++ krp->krp_status = 0; ++ krp->krp_flags = CRYPTO_KF_CBIMM; ++ krp->krp_callback = (int (*) (struct cryptkop *)) cryptodevkey_cb; ++ init_waitqueue_head(&krp->krp_waitq); ++ ++ for (i = 0; i < CRK_MAXPARAM; i++) ++ krp->krp_param[i].crp_nbits = kop->crk_param[i].crp_nbits; ++ for (i = 0; i < krp->krp_iparams + krp->krp_oparams; i++) { ++ size = (krp->krp_param[i].crp_nbits + 7) / 8; ++ if (size == 0) ++ continue; ++ krp->krp_param[i].crp_p = (caddr_t) kmalloc(size, GFP_KERNEL); ++ if (i >= krp->krp_iparams) ++ continue; ++ error = copy_from_user(krp->krp_param[i].crp_p, ++ kop->crk_param[i].crp_p, size); ++ if (error) ++ goto fail; ++ } ++ ++ error = crypto_kdispatch(krp); ++ if (error) ++ goto fail; ++ ++ do { ++ error = wait_event_interruptible(krp->krp_waitq, ++ ((krp->krp_flags & CRYPTO_KF_DONE) != 0)); ++ /* ++ * we can't break out of this loop or we will leave behind ++ * a huge mess, however, staying here means if your driver ++ * is broken user applications can hang and not be killed. ++ * The solution, fix your driver :-) ++ */ ++ if (error) { ++ schedule(); ++ error = 0; ++ } ++ } while ((krp->krp_flags & CRYPTO_KF_DONE) == 0); ++ ++ dprintk("%s finished WAITING error=%d\n", __FUNCTION__, error); ++ ++ kop->crk_crid = krp->krp_crid; /* device that did the work */ ++ if (krp->krp_status != 0) { ++ error = krp->krp_status; ++ goto fail; ++ } ++ ++ for (i = krp->krp_iparams; i < krp->krp_iparams + krp->krp_oparams; i++) { ++ size = (krp->krp_param[i].crp_nbits + 7) / 8; ++ if (size == 0) ++ continue; ++ error = copy_to_user(kop->crk_param[i].crp_p, krp->krp_param[i].crp_p, ++ size); ++ if (error) ++ goto fail; ++ } ++ ++fail: ++ if (krp) { ++ kop->crk_status = krp->krp_status; ++ for (i = 0; i < CRK_MAXPARAM; i++) { ++ if (krp->krp_param[i].crp_p) ++ kfree(krp->krp_param[i].crp_p); ++ } ++ kfree(krp); ++ } ++ return (error); ++} ++ ++static int ++cryptodev_find(struct crypt_find_op *find) ++{ ++ device_t dev; ++ ++ if (find->crid != -1) { ++ dev = crypto_find_device_byhid(find->crid); ++ if (dev == NULL) ++ return (ENOENT); ++ strlcpy(find->name, device_get_nameunit(dev), ++ sizeof(find->name)); ++ } else { ++ find->crid = crypto_find_driver(find->name); ++ if (find->crid == -1) ++ return (ENOENT); ++ } ++ return (0); ++} ++ ++static struct csession * ++csefind(struct fcrypt *fcr, u_int ses) ++{ ++ struct csession *cse; ++ ++ dprintk("%s()\n", __FUNCTION__); ++ list_for_each_entry(cse, &fcr->csessions, list) ++ if (cse->ses == ses) ++ return (cse); ++ return (NULL); ++} ++ ++static int ++csedelete(struct fcrypt *fcr, struct csession *cse_del) ++{ ++ struct csession *cse; ++ ++ dprintk("%s()\n", __FUNCTION__); ++ list_for_each_entry(cse, &fcr->csessions, list) { ++ if (cse == cse_del) { ++ list_del(&cse->list); ++ return (1); ++ } ++ } ++ return (0); ++} ++ ++static struct csession * ++cseadd(struct fcrypt *fcr, struct csession *cse) ++{ ++ dprintk("%s()\n", __FUNCTION__); ++ list_add_tail(&cse->list, &fcr->csessions); ++ cse->ses = fcr->sesn++; ++ return (cse); ++} ++ ++static struct csession * ++csecreate(struct fcrypt *fcr, u_int64_t sid, struct cryptoini *crie, ++ struct cryptoini *cria, struct csession_info *info) ++{ ++ struct csession *cse; ++ ++ dprintk("%s()\n", __FUNCTION__); ++ cse = (struct csession *) kmalloc(sizeof(struct csession), GFP_KERNEL); ++ if (cse == NULL) ++ return NULL; ++ memset(cse, 0, sizeof(struct csession)); ++ ++ INIT_LIST_HEAD(&cse->list); ++ init_waitqueue_head(&cse->waitq); ++ ++ cse->key = crie->cri_key; ++ cse->keylen = crie->cri_klen/8; ++ cse->mackey = cria->cri_key; ++ cse->mackeylen = cria->cri_klen/8; ++ cse->sid = sid; ++ cse->cipher = crie->cri_alg; ++ cse->mac = cria->cri_alg; ++ cse->info = *info; ++ cseadd(fcr, cse); ++ return (cse); ++} ++ ++static int ++csefree(struct csession *cse) ++{ ++ int error; ++ ++ dprintk("%s()\n", __FUNCTION__); ++ error = crypto_freesession(cse->sid); ++ if (cse->key) ++ kfree(cse->key); ++ if (cse->mackey) ++ kfree(cse->mackey); ++ kfree(cse); ++ return(error); ++} ++ ++static int ++cryptodev_ioctl( ++ struct inode *inode, ++ struct file *filp, ++ unsigned int cmd, ++ unsigned long arg) ++{ ++ struct cryptoini cria, crie; ++ struct fcrypt *fcr = filp->private_data; ++ struct csession *cse; ++ struct csession_info info; ++ struct session2_op sop; ++ struct crypt_op cop; ++ struct crypt_kop kop; ++ struct crypt_find_op fop; ++ u_int64_t sid; ++ u_int32_t ses; ++ int feat, fd, error = 0, crid; ++ mm_segment_t fs; ++ ++ dprintk("%s(cmd=%x arg=%lx)\n", __FUNCTION__, cmd, arg); ++ ++ switch (cmd) { ++ ++ case CRIOGET: { ++ dprintk("%s(CRIOGET)\n", __FUNCTION__); ++ fs = get_fs(); ++ set_fs(get_ds()); ++ for (fd = 0; fd < files_fdtable(current->files)->max_fds; fd++) ++ if (files_fdtable(current->files)->fd[fd] == filp) ++ break; ++ fd = sys_dup(fd); ++ set_fs(fs); ++ put_user(fd, (int *) arg); ++ return IS_ERR_VALUE(fd) ? fd : 0; ++ } ++ ++#define CIOCGSESSSTR (cmd == CIOCGSESSION ? "CIOCGSESSION" : "CIOCGSESSION2") ++ case CIOCGSESSION: ++ case CIOCGSESSION2: ++ dprintk("%s(%s)\n", __FUNCTION__, CIOCGSESSSTR); ++ memset(&crie, 0, sizeof(crie)); ++ memset(&cria, 0, sizeof(cria)); ++ memset(&info, 0, sizeof(info)); ++ memset(&sop, 0, sizeof(sop)); ++ ++ if (copy_from_user(&sop, (void*)arg, (cmd == CIOCGSESSION) ? ++ sizeof(struct session_op) : sizeof(sop))) { ++ dprintk("%s(%s) - bad copy\n", __FUNCTION__, CIOCGSESSSTR); ++ error = EFAULT; ++ goto bail; ++ } ++ ++ switch (sop.cipher) { ++ case 0: ++ dprintk("%s(%s) - no cipher\n", __FUNCTION__, CIOCGSESSSTR); ++ break; ++ case CRYPTO_NULL_CBC: ++ info.blocksize = NULL_BLOCK_LEN; ++ info.minkey = NULL_MIN_KEY_LEN; ++ info.maxkey = NULL_MAX_KEY_LEN; ++ break; ++ case CRYPTO_DES_CBC: ++ info.blocksize = DES_BLOCK_LEN; ++ info.minkey = DES_MIN_KEY_LEN; ++ info.maxkey = DES_MAX_KEY_LEN; ++ break; ++ case CRYPTO_3DES_CBC: ++ info.blocksize = DES3_BLOCK_LEN; ++ info.minkey = DES3_MIN_KEY_LEN; ++ info.maxkey = DES3_MAX_KEY_LEN; ++ break; ++ case CRYPTO_BLF_CBC: ++ info.blocksize = BLOWFISH_BLOCK_LEN; ++ info.minkey = BLOWFISH_MIN_KEY_LEN; ++ info.maxkey = BLOWFISH_MAX_KEY_LEN; ++ break; ++ case CRYPTO_CAST_CBC: ++ info.blocksize = CAST128_BLOCK_LEN; ++ info.minkey = CAST128_MIN_KEY_LEN; ++ info.maxkey = CAST128_MAX_KEY_LEN; ++ break; ++ case CRYPTO_SKIPJACK_CBC: ++ info.blocksize = SKIPJACK_BLOCK_LEN; ++ info.minkey = SKIPJACK_MIN_KEY_LEN; ++ info.maxkey = SKIPJACK_MAX_KEY_LEN; ++ break; ++ case CRYPTO_AES_CBC: ++ info.blocksize = AES_BLOCK_LEN; ++ info.minkey = AES_MIN_KEY_LEN; ++ info.maxkey = AES_MAX_KEY_LEN; ++ break; ++ case CRYPTO_ARC4: ++ info.blocksize = ARC4_BLOCK_LEN; ++ info.minkey = ARC4_MIN_KEY_LEN; ++ info.maxkey = ARC4_MAX_KEY_LEN; ++ break; ++ case CRYPTO_CAMELLIA_CBC: ++ info.blocksize = CAMELLIA_BLOCK_LEN; ++ info.minkey = CAMELLIA_MIN_KEY_LEN; ++ info.maxkey = CAMELLIA_MAX_KEY_LEN; ++ break; ++ default: ++ dprintk("%s(%s) - bad cipher\n", __FUNCTION__, CIOCGSESSSTR); ++ error = EINVAL; ++ goto bail; ++ } ++ ++ switch (sop.mac) { ++ case 0: ++ dprintk("%s(%s) - no mac\n", __FUNCTION__, CIOCGSESSSTR); ++ break; ++ case CRYPTO_NULL_HMAC: ++ info.authsize = NULL_HASH_LEN; ++ break; ++ case CRYPTO_MD5: ++ info.authsize = MD5_HASH_LEN; ++ break; ++ case CRYPTO_SHA1: ++ info.authsize = SHA1_HASH_LEN; ++ break; ++ case CRYPTO_SHA2_256: ++ info.authsize = SHA2_256_HASH_LEN; ++ break; ++ case CRYPTO_SHA2_384: ++ info.authsize = SHA2_384_HASH_LEN; ++ break; ++ case CRYPTO_SHA2_512: ++ info.authsize = SHA2_512_HASH_LEN; ++ break; ++ case CRYPTO_RIPEMD160: ++ info.authsize = RIPEMD160_HASH_LEN; ++ break; ++ case CRYPTO_MD5_HMAC: ++ info.authsize = MD5_HASH_LEN; ++ break; ++ case CRYPTO_SHA1_HMAC: ++ info.authsize = SHA1_HASH_LEN; ++ break; ++ case CRYPTO_SHA2_256_HMAC: ++ info.authsize = SHA2_256_HASH_LEN; ++ break; ++ case CRYPTO_SHA2_384_HMAC: ++ info.authsize = SHA2_384_HASH_LEN; ++ break; ++ case CRYPTO_SHA2_512_HMAC: ++ info.authsize = SHA2_512_HASH_LEN; ++ break; ++ case CRYPTO_RIPEMD160_HMAC: ++ info.authsize = RIPEMD160_HASH_LEN; ++ break; ++ default: ++ dprintk("%s(%s) - bad mac\n", __FUNCTION__, CIOCGSESSSTR); ++ error = EINVAL; ++ goto bail; ++ } ++ ++ if (info.blocksize) { ++ crie.cri_alg = sop.cipher; ++ crie.cri_klen = sop.keylen * 8; ++ if ((info.maxkey && sop.keylen > info.maxkey) || ++ sop.keylen < info.minkey) { ++ dprintk("%s(%s) - bad key\n", __FUNCTION__, CIOCGSESSSTR); ++ error = EINVAL; ++ goto bail; ++ } ++ ++ crie.cri_key = (u_int8_t *) kmalloc(crie.cri_klen/8+1, GFP_KERNEL); ++ if (copy_from_user(crie.cri_key, sop.key, ++ crie.cri_klen/8)) { ++ dprintk("%s(%s) - bad copy\n", __FUNCTION__, CIOCGSESSSTR); ++ error = EFAULT; ++ goto bail; ++ } ++ if (info.authsize) ++ crie.cri_next = &cria; ++ } ++ ++ if (info.authsize) { ++ cria.cri_alg = sop.mac; ++ cria.cri_klen = sop.mackeylen * 8; ++ if ((info.maxkey && sop.mackeylen > info.maxkey) || ++ sop.keylen < info.minkey) { ++ dprintk("%s(%s) - mackeylen %d\n", __FUNCTION__, CIOCGSESSSTR, ++ sop.mackeylen); ++ error = EINVAL; ++ goto bail; ++ } ++ ++ if (cria.cri_klen) { ++ cria.cri_key = (u_int8_t *) kmalloc(cria.cri_klen/8,GFP_KERNEL); ++ if (copy_from_user(cria.cri_key, sop.mackey, ++ cria.cri_klen / 8)) { ++ dprintk("%s(%s) - bad copy\n", __FUNCTION__, CIOCGSESSSTR); ++ error = EFAULT; ++ goto bail; ++ } ++ } ++ } ++ ++ /* NB: CIOGSESSION2 has the crid */ ++ if (cmd == CIOCGSESSION2) { ++ crid = sop.crid; ++ error = checkcrid(crid); ++ if (error) { ++ dprintk("%s(%s) - checkcrid %x\n", __FUNCTION__, ++ CIOCGSESSSTR, error); ++ goto bail; ++ } ++ } else { ++ /* allow either HW or SW to be used */ ++ crid = CRYPTOCAP_F_HARDWARE | CRYPTOCAP_F_SOFTWARE; ++ } ++ error = crypto_newsession(&sid, (info.blocksize ? &crie : &cria), crid); ++ if (error) { ++ dprintk("%s(%s) - newsession %d\n",__FUNCTION__,CIOCGSESSSTR,error); ++ goto bail; ++ } ++ ++ cse = csecreate(fcr, sid, &crie, &cria, &info); ++ if (cse == NULL) { ++ crypto_freesession(sid); ++ error = EINVAL; ++ dprintk("%s(%s) - csecreate failed\n", __FUNCTION__, CIOCGSESSSTR); ++ goto bail; ++ } ++ sop.ses = cse->ses; ++ ++ if (cmd == CIOCGSESSION2) { ++ /* return hardware/driver id */ ++ sop.crid = CRYPTO_SESID2HID(cse->sid); ++ } ++ ++ if (copy_to_user((void*)arg, &sop, (cmd == CIOCGSESSION) ? ++ sizeof(struct session_op) : sizeof(sop))) { ++ dprintk("%s(%s) - bad copy\n", __FUNCTION__, CIOCGSESSSTR); ++ error = EFAULT; ++ } ++bail: ++ if (error) { ++ dprintk("%s(%s) - bail %d\n", __FUNCTION__, CIOCGSESSSTR, error); ++ if (crie.cri_key) ++ kfree(crie.cri_key); ++ if (cria.cri_key) ++ kfree(cria.cri_key); ++ } ++ break; ++ case CIOCFSESSION: ++ dprintk("%s(CIOCFSESSION)\n", __FUNCTION__); ++ get_user(ses, (uint32_t*)arg); ++ cse = csefind(fcr, ses); ++ if (cse == NULL) { ++ error = EINVAL; ++ dprintk("%s(CIOCFSESSION) - Fail %d\n", __FUNCTION__, error); ++ break; ++ } ++ csedelete(fcr, cse); ++ error = csefree(cse); ++ break; ++ case CIOCCRYPT: ++ dprintk("%s(CIOCCRYPT)\n", __FUNCTION__); ++ if(copy_from_user(&cop, (void*)arg, sizeof(cop))) { ++ dprintk("%s(CIOCCRYPT) - bad copy\n", __FUNCTION__); ++ error = EFAULT; ++ goto bail; ++ } ++ cse = csefind(fcr, cop.ses); ++ if (cse == NULL) { ++ error = EINVAL; ++ dprintk("%s(CIOCCRYPT) - Fail %d\n", __FUNCTION__, error); ++ break; ++ } ++ error = cryptodev_op(cse, &cop); ++ if(copy_to_user((void*)arg, &cop, sizeof(cop))) { ++ dprintk("%s(CIOCCRYPT) - bad return copy\n", __FUNCTION__); ++ error = EFAULT; ++ goto bail; ++ } ++ break; ++ case CIOCKEY: ++ case CIOCKEY2: ++ dprintk("%s(CIOCKEY)\n", __FUNCTION__); ++ if (!crypto_userasymcrypto) ++ return (EPERM); /* XXX compat? */ ++ if(copy_from_user(&kop, (void*)arg, sizeof(kop))) { ++ dprintk("%s(CIOCKEY) - bad copy\n", __FUNCTION__); ++ error = EFAULT; ++ goto bail; ++ } ++ if (cmd == CIOCKEY) { ++ /* NB: crypto core enforces s/w driver use */ ++ kop.crk_crid = ++ CRYPTOCAP_F_HARDWARE | CRYPTOCAP_F_SOFTWARE; ++ } ++ error = cryptodev_key(&kop); ++ if(copy_to_user((void*)arg, &kop, sizeof(kop))) { ++ dprintk("%s(CIOCGKEY) - bad return copy\n", __FUNCTION__); ++ error = EFAULT; ++ goto bail; ++ } ++ break; ++ case CIOCASYMFEAT: ++ dprintk("%s(CIOCASYMFEAT)\n", __FUNCTION__); ++ if (!crypto_userasymcrypto) { ++ /* ++ * NB: if user asym crypto operations are ++ * not permitted return "no algorithms" ++ * so well-behaved applications will just ++ * fallback to doing them in software. ++ */ ++ feat = 0; ++ } else ++ error = crypto_getfeat(&feat); ++ if (!error) { ++ error = copy_to_user((void*)arg, &feat, sizeof(feat)); ++ } ++ break; ++ case CIOCFINDDEV: ++ if (copy_from_user(&fop, (void*)arg, sizeof(fop))) { ++ dprintk("%s(CIOCFINDDEV) - bad copy\n", __FUNCTION__); ++ error = EFAULT; ++ goto bail; ++ } ++ error = cryptodev_find(&fop); ++ if (copy_to_user((void*)arg, &fop, sizeof(fop))) { ++ dprintk("%s(CIOCFINDDEV) - bad return copy\n", __FUNCTION__); ++ error = EFAULT; ++ goto bail; ++ } ++ break; ++ default: ++ dprintk("%s(unknown ioctl 0x%x)\n", __FUNCTION__, cmd); ++ error = EINVAL; ++ break; ++ } ++ return(-error); ++} ++ ++#ifdef HAVE_UNLOCKED_IOCTL ++static long ++cryptodev_unlocked_ioctl( ++ struct file *filp, ++ unsigned int cmd, ++ unsigned long arg) ++{ ++ return cryptodev_ioctl(NULL, filp, cmd, arg); ++} ++#endif ++ ++static int ++cryptodev_open(struct inode *inode, struct file *filp) ++{ ++ struct fcrypt *fcr; ++ ++ dprintk("%s()\n", __FUNCTION__); ++ if (filp->private_data) { ++ printk("cryptodev: Private data already exists !\n"); ++ return(0); ++ } ++ ++ fcr = kmalloc(sizeof(*fcr), GFP_KERNEL); ++ if (!fcr) { ++ dprintk("%s() - malloc failed\n", __FUNCTION__); ++ return(-ENOMEM); ++ } ++ memset(fcr, 0, sizeof(*fcr)); ++ ++ INIT_LIST_HEAD(&fcr->csessions); ++ filp->private_data = fcr; ++ return(0); ++} ++ ++static int ++cryptodev_release(struct inode *inode, struct file *filp) ++{ ++ struct fcrypt *fcr = filp->private_data; ++ struct csession *cse, *tmp; ++ ++ dprintk("%s()\n", __FUNCTION__); ++ if (!filp) { ++ printk("cryptodev: No private data on release\n"); ++ return(0); ++ } ++ ++ list_for_each_entry_safe(cse, tmp, &fcr->csessions, list) { ++ list_del(&cse->list); ++ (void)csefree(cse); ++ } ++ filp->private_data = NULL; ++ kfree(fcr); ++ return(0); ++} ++ ++static struct file_operations cryptodev_fops = { ++ .owner = THIS_MODULE, ++ .open = cryptodev_open, ++ .release = cryptodev_release, ++ .ioctl = cryptodev_ioctl, ++#ifdef HAVE_UNLOCKED_IOCTL ++ .unlocked_ioctl = cryptodev_unlocked_ioctl, ++#endif ++}; ++ ++static struct miscdevice cryptodev = { ++ .minor = CRYPTODEV_MINOR, ++ .name = "crypto", ++ .fops = &cryptodev_fops, ++}; ++ ++static int __init ++cryptodev_init(void) ++{ ++ int rc; ++ ++ dprintk("%s(%p)\n", __FUNCTION__, cryptodev_init); ++ rc = misc_register(&cryptodev); ++ if (rc) { ++ printk(KERN_ERR "cryptodev: registration of /dev/crypto failed\n"); ++ return(rc); ++ } ++ ++ return(0); ++} ++ ++static void __exit ++cryptodev_exit(void) ++{ ++ dprintk("%s()\n", __FUNCTION__); ++ misc_deregister(&cryptodev); ++} ++ ++module_init(cryptodev_init); ++module_exit(cryptodev_exit); ++ ++MODULE_LICENSE("BSD"); ++MODULE_AUTHOR("David McCullough <david_mccullough@securecomputing.com>"); ++MODULE_DESCRIPTION("Cryptodev (user interface to OCF)"); +diff -Nur linux-2.6.30.orig/crypto/ocf/cryptodev.h linux-2.6.30/crypto/ocf/cryptodev.h +--- linux-2.6.30.orig/crypto/ocf/cryptodev.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/cryptodev.h 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,478 @@ ++/* $FreeBSD: src/sys/opencrypto/cryptodev.h,v 1.25 2007/05/09 19:37:02 gnn Exp $ */ ++/* $OpenBSD: cryptodev.h,v 1.31 2002/06/11 11:14:29 beck Exp $ */ ++ ++/*- ++ * Linux port done by David McCullough <david_mccullough@securecomputing.com> ++ * Copyright (C) 2006-2007 David McCullough ++ * Copyright (C) 2004-2005 Intel Corporation. ++ * The license and original author are listed below. ++ * ++ * The author of this code is Angelos D. Keromytis (angelos@cis.upenn.edu) ++ * Copyright (c) 2002-2006 Sam Leffler, Errno Consulting ++ * ++ * This code was written by Angelos D. Keromytis in Athens, Greece, in ++ * February 2000. Network Security Technologies Inc. (NSTI) kindly ++ * supported the development of this code. ++ * ++ * Copyright (c) 2000 Angelos D. Keromytis ++ * ++ * Permission to use, copy, and modify this software with or without fee ++ * is hereby granted, provided that this entire notice is included in ++ * all source code copies of any software which is or includes a copy or ++ * modification of this software. ++ * ++ * THIS SOFTWARE IS BEING PROVIDED "AS IS", WITHOUT ANY EXPRESS OR ++ * IMPLIED WARRANTY. IN PARTICULAR, NONE OF THE AUTHORS MAKES ANY ++ * REPRESENTATION OR WARRANTY OF ANY KIND CONCERNING THE ++ * MERCHANTABILITY OF THIS SOFTWARE OR ITS FITNESS FOR ANY PARTICULAR ++ * PURPOSE. ++ * ++ * Copyright (c) 2001 Theo de Raadt ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions ++ * are met: ++ * ++ * 1. Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * 2. Redistributions in binary form must reproduce the above copyright ++ * notice, this list of conditions and the following disclaimer in the ++ * documentation and/or other materials provided with the distribution. ++ * 3. The name of the author may not be used to endorse or promote products ++ * derived from this software without specific prior written permission. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR ++ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES ++ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. ++ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, ++ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT ++ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, ++ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY ++ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT ++ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF ++ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. ++ * ++ * Effort sponsored in part by the Defense Advanced Research Projects ++ * Agency (DARPA) and Air Force Research Laboratory, Air Force ++ * Materiel Command, USAF, under agreement number F30602-01-2-0537. ++ * ++ */ ++ ++#ifndef _CRYPTO_CRYPTO_H_ ++#define _CRYPTO_CRYPTO_H_ ++ ++/* Some initial values */ ++#define CRYPTO_DRIVERS_INITIAL 4 ++#define CRYPTO_SW_SESSIONS 32 ++ ++/* Hash values */ ++#define NULL_HASH_LEN 0 ++#define MD5_HASH_LEN 16 ++#define SHA1_HASH_LEN 20 ++#define RIPEMD160_HASH_LEN 20 ++#define SHA2_256_HASH_LEN 32 ++#define SHA2_384_HASH_LEN 48 ++#define SHA2_512_HASH_LEN 64 ++#define MD5_KPDK_HASH_LEN 16 ++#define SHA1_KPDK_HASH_LEN 20 ++/* Maximum hash algorithm result length */ ++#define HASH_MAX_LEN SHA2_512_HASH_LEN /* Keep this updated */ ++ ++/* HMAC values */ ++#define NULL_HMAC_BLOCK_LEN 1 ++#define MD5_HMAC_BLOCK_LEN 64 ++#define SHA1_HMAC_BLOCK_LEN 64 ++#define RIPEMD160_HMAC_BLOCK_LEN 64 ++#define SHA2_256_HMAC_BLOCK_LEN 64 ++#define SHA2_384_HMAC_BLOCK_LEN 128 ++#define SHA2_512_HMAC_BLOCK_LEN 128 ++/* Maximum HMAC block length */ ++#define HMAC_MAX_BLOCK_LEN SHA2_512_HMAC_BLOCK_LEN /* Keep this updated */ ++#define HMAC_IPAD_VAL 0x36 ++#define HMAC_OPAD_VAL 0x5C ++ ++/* Encryption algorithm block sizes */ ++#define NULL_BLOCK_LEN 1 ++#define DES_BLOCK_LEN 8 ++#define DES3_BLOCK_LEN 8 ++#define BLOWFISH_BLOCK_LEN 8 ++#define SKIPJACK_BLOCK_LEN 8 ++#define CAST128_BLOCK_LEN 8 ++#define RIJNDAEL128_BLOCK_LEN 16 ++#define AES_BLOCK_LEN RIJNDAEL128_BLOCK_LEN ++#define CAMELLIA_BLOCK_LEN 16 ++#define ARC4_BLOCK_LEN 1 ++#define EALG_MAX_BLOCK_LEN AES_BLOCK_LEN /* Keep this updated */ ++ ++/* Encryption algorithm min and max key sizes */ ++#define NULL_MIN_KEY_LEN 0 ++#define NULL_MAX_KEY_LEN 0 ++#define DES_MIN_KEY_LEN 8 ++#define DES_MAX_KEY_LEN 8 ++#define DES3_MIN_KEY_LEN 24 ++#define DES3_MAX_KEY_LEN 24 ++#define BLOWFISH_MIN_KEY_LEN 4 ++#define BLOWFISH_MAX_KEY_LEN 56 ++#define SKIPJACK_MIN_KEY_LEN 10 ++#define SKIPJACK_MAX_KEY_LEN 10 ++#define CAST128_MIN_KEY_LEN 5 ++#define CAST128_MAX_KEY_LEN 16 ++#define RIJNDAEL128_MIN_KEY_LEN 16 ++#define RIJNDAEL128_MAX_KEY_LEN 32 ++#define AES_MIN_KEY_LEN RIJNDAEL128_MIN_KEY_LEN ++#define AES_MAX_KEY_LEN RIJNDAEL128_MAX_KEY_LEN ++#define CAMELLIA_MIN_KEY_LEN 16 ++#define CAMELLIA_MAX_KEY_LEN 32 ++#define ARC4_MIN_KEY_LEN 1 ++#define ARC4_MAX_KEY_LEN 256 ++ ++/* Max size of data that can be processed */ ++#define CRYPTO_MAX_DATA_LEN 64*1024 - 1 ++ ++#define CRYPTO_ALGORITHM_MIN 1 ++#define CRYPTO_DES_CBC 1 ++#define CRYPTO_3DES_CBC 2 ++#define CRYPTO_BLF_CBC 3 ++#define CRYPTO_CAST_CBC 4 ++#define CRYPTO_SKIPJACK_CBC 5 ++#define CRYPTO_MD5_HMAC 6 ++#define CRYPTO_SHA1_HMAC 7 ++#define CRYPTO_RIPEMD160_HMAC 8 ++#define CRYPTO_MD5_KPDK 9 ++#define CRYPTO_SHA1_KPDK 10 ++#define CRYPTO_RIJNDAEL128_CBC 11 /* 128 bit blocksize */ ++#define CRYPTO_AES_CBC 11 /* 128 bit blocksize -- the same as above */ ++#define CRYPTO_ARC4 12 ++#define CRYPTO_MD5 13 ++#define CRYPTO_SHA1 14 ++#define CRYPTO_NULL_HMAC 15 ++#define CRYPTO_NULL_CBC 16 ++#define CRYPTO_DEFLATE_COMP 17 /* Deflate compression algorithm */ ++#define CRYPTO_SHA2_256_HMAC 18 ++#define CRYPTO_SHA2_384_HMAC 19 ++#define CRYPTO_SHA2_512_HMAC 20 ++#define CRYPTO_CAMELLIA_CBC 21 ++#define CRYPTO_SHA2_256 22 ++#define CRYPTO_SHA2_384 23 ++#define CRYPTO_SHA2_512 24 ++#define CRYPTO_RIPEMD160 25 ++#define CRYPTO_ALGORITHM_MAX 25 /* Keep updated - see below */ ++ ++/* Algorithm flags */ ++#define CRYPTO_ALG_FLAG_SUPPORTED 0x01 /* Algorithm is supported */ ++#define CRYPTO_ALG_FLAG_RNG_ENABLE 0x02 /* Has HW RNG for DH/DSA */ ++#define CRYPTO_ALG_FLAG_DSA_SHA 0x04 /* Can do SHA on msg */ ++ ++/* ++ * Crypto driver/device flags. They can set in the crid ++ * parameter when creating a session or submitting a key ++ * op to affect the device/driver assigned. If neither ++ * of these are specified then the crid is assumed to hold ++ * the driver id of an existing (and suitable) device that ++ * must be used to satisfy the request. ++ */ ++#define CRYPTO_FLAG_HARDWARE 0x01000000 /* hardware accelerated */ ++#define CRYPTO_FLAG_SOFTWARE 0x02000000 /* software implementation */ ++ ++/* NB: deprecated */ ++struct session_op { ++ u_int32_t cipher; /* ie. CRYPTO_DES_CBC */ ++ u_int32_t mac; /* ie. CRYPTO_MD5_HMAC */ ++ ++ u_int32_t keylen; /* cipher key */ ++ caddr_t key; ++ int mackeylen; /* mac key */ ++ caddr_t mackey; ++ ++ u_int32_t ses; /* returns: session # */ ++}; ++ ++struct session2_op { ++ u_int32_t cipher; /* ie. CRYPTO_DES_CBC */ ++ u_int32_t mac; /* ie. CRYPTO_MD5_HMAC */ ++ ++ u_int32_t keylen; /* cipher key */ ++ caddr_t key; ++ int mackeylen; /* mac key */ ++ caddr_t mackey; ++ ++ u_int32_t ses; /* returns: session # */ ++ int crid; /* driver id + flags (rw) */ ++ int pad[4]; /* for future expansion */ ++}; ++ ++struct crypt_op { ++ u_int32_t ses; ++ u_int16_t op; /* i.e. COP_ENCRYPT */ ++#define COP_NONE 0 ++#define COP_ENCRYPT 1 ++#define COP_DECRYPT 2 ++ u_int16_t flags; ++#define COP_F_BATCH 0x0008 /* Batch op if possible */ ++ u_int len; ++ caddr_t src, dst; /* become iov[] inside kernel */ ++ caddr_t mac; /* must be big enough for chosen MAC */ ++ caddr_t iv; ++}; ++ ++/* ++ * Parameters for looking up a crypto driver/device by ++ * device name or by id. The latter are returned for ++ * created sessions (crid) and completed key operations. ++ */ ++struct crypt_find_op { ++ int crid; /* driver id + flags */ ++ char name[32]; /* device/driver name */ ++}; ++ ++/* bignum parameter, in packed bytes, ... */ ++struct crparam { ++ caddr_t crp_p; ++ u_int crp_nbits; ++}; ++ ++#define CRK_MAXPARAM 8 ++ ++struct crypt_kop { ++ u_int crk_op; /* ie. CRK_MOD_EXP or other */ ++ u_int crk_status; /* return status */ ++ u_short crk_iparams; /* # of input parameters */ ++ u_short crk_oparams; /* # of output parameters */ ++ u_int crk_crid; /* NB: only used by CIOCKEY2 (rw) */ ++ struct crparam crk_param[CRK_MAXPARAM]; ++}; ++#define CRK_ALGORITM_MIN 0 ++#define CRK_MOD_EXP 0 ++#define CRK_MOD_EXP_CRT 1 ++#define CRK_DSA_SIGN 2 ++#define CRK_DSA_VERIFY 3 ++#define CRK_DH_COMPUTE_KEY 4 ++#define CRK_ALGORITHM_MAX 4 /* Keep updated - see below */ ++ ++#define CRF_MOD_EXP (1 << CRK_MOD_EXP) ++#define CRF_MOD_EXP_CRT (1 << CRK_MOD_EXP_CRT) ++#define CRF_DSA_SIGN (1 << CRK_DSA_SIGN) ++#define CRF_DSA_VERIFY (1 << CRK_DSA_VERIFY) ++#define CRF_DH_COMPUTE_KEY (1 << CRK_DH_COMPUTE_KEY) ++ ++/* ++ * done against open of /dev/crypto, to get a cloned descriptor. ++ * Please use F_SETFD against the cloned descriptor. ++ */ ++#define CRIOGET _IOWR('c', 100, u_int32_t) ++#define CRIOASYMFEAT CIOCASYMFEAT ++#define CRIOFINDDEV CIOCFINDDEV ++ ++/* the following are done against the cloned descriptor */ ++#define CIOCGSESSION _IOWR('c', 101, struct session_op) ++#define CIOCFSESSION _IOW('c', 102, u_int32_t) ++#define CIOCCRYPT _IOWR('c', 103, struct crypt_op) ++#define CIOCKEY _IOWR('c', 104, struct crypt_kop) ++#define CIOCASYMFEAT _IOR('c', 105, u_int32_t) ++#define CIOCGSESSION2 _IOWR('c', 106, struct session2_op) ++#define CIOCKEY2 _IOWR('c', 107, struct crypt_kop) ++#define CIOCFINDDEV _IOWR('c', 108, struct crypt_find_op) ++ ++struct cryptotstat { ++ struct timespec acc; /* total accumulated time */ ++ struct timespec min; /* min time */ ++ struct timespec max; /* max time */ ++ u_int32_t count; /* number of observations */ ++}; ++ ++struct cryptostats { ++ u_int32_t cs_ops; /* symmetric crypto ops submitted */ ++ u_int32_t cs_errs; /* symmetric crypto ops that failed */ ++ u_int32_t cs_kops; /* asymetric/key ops submitted */ ++ u_int32_t cs_kerrs; /* asymetric/key ops that failed */ ++ u_int32_t cs_intrs; /* crypto swi thread activations */ ++ u_int32_t cs_rets; /* crypto return thread activations */ ++ u_int32_t cs_blocks; /* symmetric op driver block */ ++ u_int32_t cs_kblocks; /* symmetric op driver block */ ++ /* ++ * When CRYPTO_TIMING is defined at compile time and the ++ * sysctl debug.crypto is set to 1, the crypto system will ++ * accumulate statistics about how long it takes to process ++ * crypto requests at various points during processing. ++ */ ++ struct cryptotstat cs_invoke; /* crypto_dipsatch -> crypto_invoke */ ++ struct cryptotstat cs_done; /* crypto_invoke -> crypto_done */ ++ struct cryptotstat cs_cb; /* crypto_done -> callback */ ++ struct cryptotstat cs_finis; /* callback -> callback return */ ++ ++ u_int32_t cs_drops; /* crypto ops dropped due to congestion */ ++}; ++ ++#ifdef __KERNEL__ ++ ++/* Standard initialization structure beginning */ ++struct cryptoini { ++ int cri_alg; /* Algorithm to use */ ++ int cri_klen; /* Key length, in bits */ ++ int cri_mlen; /* Number of bytes we want from the ++ entire hash. 0 means all. */ ++ caddr_t cri_key; /* key to use */ ++ u_int8_t cri_iv[EALG_MAX_BLOCK_LEN]; /* IV to use */ ++ struct cryptoini *cri_next; ++}; ++ ++/* Describe boundaries of a single crypto operation */ ++struct cryptodesc { ++ int crd_skip; /* How many bytes to ignore from start */ ++ int crd_len; /* How many bytes to process */ ++ int crd_inject; /* Where to inject results, if applicable */ ++ int crd_flags; ++ ++#define CRD_F_ENCRYPT 0x01 /* Set when doing encryption */ ++#define CRD_F_IV_PRESENT 0x02 /* When encrypting, IV is already in ++ place, so don't copy. */ ++#define CRD_F_IV_EXPLICIT 0x04 /* IV explicitly provided */ ++#define CRD_F_DSA_SHA_NEEDED 0x08 /* Compute SHA-1 of buffer for DSA */ ++#define CRD_F_KEY_EXPLICIT 0x10 /* Key explicitly provided */ ++#define CRD_F_COMP 0x0f /* Set when doing compression */ ++ ++ struct cryptoini CRD_INI; /* Initialization/context data */ ++#define crd_iv CRD_INI.cri_iv ++#define crd_key CRD_INI.cri_key ++#define crd_alg CRD_INI.cri_alg ++#define crd_klen CRD_INI.cri_klen ++ ++ struct cryptodesc *crd_next; ++}; ++ ++/* Structure describing complete operation */ ++struct cryptop { ++ struct list_head crp_next; ++ wait_queue_head_t crp_waitq; ++ ++ u_int64_t crp_sid; /* Session ID */ ++ int crp_ilen; /* Input data total length */ ++ int crp_olen; /* Result total length */ ++ ++ int crp_etype; /* ++ * Error type (zero means no error). ++ * All error codes except EAGAIN ++ * indicate possible data corruption (as in, ++ * the data have been touched). On all ++ * errors, the crp_sid may have changed ++ * (reset to a new one), so the caller ++ * should always check and use the new ++ * value on future requests. ++ */ ++ int crp_flags; ++ ++#define CRYPTO_F_SKBUF 0x0001 /* Input/output are skbuf chains */ ++#define CRYPTO_F_IOV 0x0002 /* Input/output are uio */ ++#define CRYPTO_F_REL 0x0004 /* Must return data in same place */ ++#define CRYPTO_F_BATCH 0x0008 /* Batch op if possible */ ++#define CRYPTO_F_CBIMM 0x0010 /* Do callback immediately */ ++#define CRYPTO_F_DONE 0x0020 /* Operation completed */ ++#define CRYPTO_F_CBIFSYNC 0x0040 /* Do CBIMM if op is synchronous */ ++ ++ caddr_t crp_buf; /* Data to be processed */ ++ caddr_t crp_opaque; /* Opaque pointer, passed along */ ++ struct cryptodesc *crp_desc; /* Linked list of processing descriptors */ ++ ++ int (*crp_callback)(struct cryptop *); /* Callback function */ ++}; ++ ++#define CRYPTO_BUF_CONTIG 0x0 ++#define CRYPTO_BUF_IOV 0x1 ++#define CRYPTO_BUF_SKBUF 0x2 ++ ++#define CRYPTO_OP_DECRYPT 0x0 ++#define CRYPTO_OP_ENCRYPT 0x1 ++ ++/* ++ * Hints passed to process methods. ++ */ ++#define CRYPTO_HINT_MORE 0x1 /* more ops coming shortly */ ++ ++struct cryptkop { ++ struct list_head krp_next; ++ wait_queue_head_t krp_waitq; ++ ++ int krp_flags; ++#define CRYPTO_KF_DONE 0x0001 /* Operation completed */ ++#define CRYPTO_KF_CBIMM 0x0002 /* Do callback immediately */ ++ ++ u_int krp_op; /* ie. CRK_MOD_EXP or other */ ++ u_int krp_status; /* return status */ ++ u_short krp_iparams; /* # of input parameters */ ++ u_short krp_oparams; /* # of output parameters */ ++ u_int krp_crid; /* desired device, etc. */ ++ u_int32_t krp_hid; ++ struct crparam krp_param[CRK_MAXPARAM]; /* kvm */ ++ int (*krp_callback)(struct cryptkop *); ++}; ++ ++#include <ocf-compat.h> ++ ++/* ++ * Session ids are 64 bits. The lower 32 bits contain a "local id" which ++ * is a driver-private session identifier. The upper 32 bits contain a ++ * "hardware id" used by the core crypto code to identify the driver and ++ * a copy of the driver's capabilities that can be used by client code to ++ * optimize operation. ++ */ ++#define CRYPTO_SESID2HID(_sid) (((_sid) >> 32) & 0x00ffffff) ++#define CRYPTO_SESID2CAPS(_sid) (((_sid) >> 32) & 0xff000000) ++#define CRYPTO_SESID2LID(_sid) (((u_int32_t) (_sid)) & 0xffffffff) ++ ++extern int crypto_newsession(u_int64_t *sid, struct cryptoini *cri, int hard); ++extern int crypto_freesession(u_int64_t sid); ++#define CRYPTOCAP_F_HARDWARE CRYPTO_FLAG_HARDWARE ++#define CRYPTOCAP_F_SOFTWARE CRYPTO_FLAG_SOFTWARE ++#define CRYPTOCAP_F_SYNC 0x04000000 /* operates synchronously */ ++extern int32_t crypto_get_driverid(device_t dev, int flags); ++extern int crypto_find_driver(const char *); ++extern device_t crypto_find_device_byhid(int hid); ++extern int crypto_getcaps(int hid); ++extern int crypto_register(u_int32_t driverid, int alg, u_int16_t maxoplen, ++ u_int32_t flags); ++extern int crypto_kregister(u_int32_t, int, u_int32_t); ++extern int crypto_unregister(u_int32_t driverid, int alg); ++extern int crypto_unregister_all(u_int32_t driverid); ++extern int crypto_dispatch(struct cryptop *crp); ++extern int crypto_kdispatch(struct cryptkop *); ++#define CRYPTO_SYMQ 0x1 ++#define CRYPTO_ASYMQ 0x2 ++extern int crypto_unblock(u_int32_t, int); ++extern void crypto_done(struct cryptop *crp); ++extern void crypto_kdone(struct cryptkop *); ++extern int crypto_getfeat(int *); ++ ++extern void crypto_freereq(struct cryptop *crp); ++extern struct cryptop *crypto_getreq(int num); ++ ++extern int crypto_usercrypto; /* userland may do crypto requests */ ++extern int crypto_userasymcrypto; /* userland may do asym crypto reqs */ ++extern int crypto_devallowsoft; /* only use hardware crypto */ ++ ++/* ++ * random number support, crypto_unregister_all will unregister ++ */ ++extern int crypto_rregister(u_int32_t driverid, ++ int (*read_random)(void *arg, u_int32_t *buf, int len), void *arg); ++extern int crypto_runregister_all(u_int32_t driverid); ++ ++/* ++ * Crypto-related utility routines used mainly by drivers. ++ * ++ * XXX these don't really belong here; but for now they're ++ * kept apart from the rest of the system. ++ */ ++struct uio; ++extern void cuio_copydata(struct uio* uio, int off, int len, caddr_t cp); ++extern void cuio_copyback(struct uio* uio, int off, int len, caddr_t cp); ++extern struct iovec *cuio_getptr(struct uio *uio, int loc, int *off); ++ ++extern void crypto_copyback(int flags, caddr_t buf, int off, int size, ++ caddr_t in); ++extern void crypto_copydata(int flags, caddr_t buf, int off, int size, ++ caddr_t out); ++extern int crypto_apply(int flags, caddr_t buf, int off, int len, ++ int (*f)(void *, void *, u_int), void *arg); ++ ++#endif /* __KERNEL__ */ ++#endif /* _CRYPTO_CRYPTO_H_ */ +diff -Nur linux-2.6.30.orig/crypto/ocf/cryptosoft.c linux-2.6.30/crypto/ocf/cryptosoft.c +--- linux-2.6.30.orig/crypto/ocf/cryptosoft.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/cryptosoft.c 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,898 @@ ++/* ++ * An OCF module that uses the linux kernel cryptoapi, based on the ++ * original cryptosoft for BSD by Angelos D. Keromytis (angelos@cis.upenn.edu) ++ * but is mostly unrecognisable, ++ * ++ * Written by David McCullough <david_mccullough@securecomputing.com> ++ * Copyright (C) 2004-2007 David McCullough ++ * Copyright (C) 2004-2005 Intel Corporation. ++ * ++ * LICENSE TERMS ++ * ++ * The free distribution and use of this software in both source and binary ++ * form is allowed (with or without changes) provided that: ++ * ++ * 1. distributions of this source code include the above copyright ++ * notice, this list of conditions and the following disclaimer; ++ * ++ * 2. distributions in binary form include the above copyright ++ * notice, this list of conditions and the following disclaimer ++ * in the documentation and/or other associated materials; ++ * ++ * 3. the copyright holder's name is not used to endorse products ++ * built using this software without specific written permission. ++ * ++ * ALTERNATIVELY, provided that this notice is retained in full, this product ++ * may be distributed under the terms of the GNU General Public License (GPL), ++ * in which case the provisions of the GPL apply INSTEAD OF those given above. ++ * ++ * DISCLAIMER ++ * ++ * This software is provided 'as is' with no explicit or implied warranties ++ * in respect of its properties, including, but not limited to, correctness ++ * and/or fitness for purpose. ++ * --------------------------------------------------------------------------- ++ */ ++ ++#ifndef AUTOCONF_INCLUDED ++#include <linux/config.h> ++#endif ++#include <linux/module.h> ++#include <linux/init.h> ++#include <linux/list.h> ++#include <linux/slab.h> ++#include <linux/sched.h> ++#include <linux/wait.h> ++#include <linux/crypto.h> ++#include <linux/mm.h> ++#include <linux/skbuff.h> ++#include <linux/random.h> ++#include <linux/scatterlist.h> ++ ++#include <cryptodev.h> ++#include <uio.h> ++ ++struct { ++ softc_device_decl sc_dev; ++} swcr_softc; ++ ++#define offset_in_page(p) ((unsigned long)(p) & ~PAGE_MASK) ++ ++/* Software session entry */ ++ ++#define SW_TYPE_CIPHER 0 ++#define SW_TYPE_HMAC 1 ++#define SW_TYPE_AUTH2 2 ++#define SW_TYPE_HASH 3 ++#define SW_TYPE_COMP 4 ++#define SW_TYPE_BLKCIPHER 5 ++ ++struct swcr_data { ++ int sw_type; ++ int sw_alg; ++ struct crypto_tfm *sw_tfm; ++ union { ++ struct { ++ char *sw_key; ++ int sw_klen; ++ int sw_mlen; ++ } hmac; ++ void *sw_comp_buf; ++ } u; ++ struct swcr_data *sw_next; ++}; ++ ++#ifndef CRYPTO_TFM_MODE_CBC ++/* ++ * As of linux-2.6.21 this is no longer defined, and presumably no longer ++ * needed to be passed into the crypto core code. ++ */ ++#define CRYPTO_TFM_MODE_CBC 0 ++#define CRYPTO_TFM_MODE_ECB 0 ++#endif ++ ++#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,19) ++ /* ++ * Linux 2.6.19 introduced a new Crypto API, setup macro's to convert new ++ * API into old API. ++ */ ++ ++ /* Symmetric/Block Cipher */ ++ struct blkcipher_desc ++ { ++ struct crypto_tfm *tfm; ++ void *info; ++ }; ++ #define ecb(X) #X ++ #define cbc(X) #X ++ #define crypto_has_blkcipher(X, Y, Z) crypto_alg_available(X, 0) ++ #define crypto_blkcipher_cast(X) X ++ #define crypto_blkcipher_tfm(X) X ++ #define crypto_alloc_blkcipher(X, Y, Z) crypto_alloc_tfm(X, mode) ++ #define crypto_blkcipher_ivsize(X) crypto_tfm_alg_ivsize(X) ++ #define crypto_blkcipher_blocksize(X) crypto_tfm_alg_blocksize(X) ++ #define crypto_blkcipher_setkey(X, Y, Z) crypto_cipher_setkey(X, Y, Z) ++ #define crypto_blkcipher_encrypt_iv(W, X, Y, Z) \ ++ crypto_cipher_encrypt_iv((W)->tfm, X, Y, Z, (u8 *)((W)->info)) ++ #define crypto_blkcipher_decrypt_iv(W, X, Y, Z) \ ++ crypto_cipher_decrypt_iv((W)->tfm, X, Y, Z, (u8 *)((W)->info)) ++ ++ /* Hash/HMAC/Digest */ ++ struct hash_desc ++ { ++ struct crypto_tfm *tfm; ++ }; ++ #define hmac(X) #X ++ #define crypto_has_hash(X, Y, Z) crypto_alg_available(X, 0) ++ #define crypto_hash_cast(X) X ++ #define crypto_hash_tfm(X) X ++ #define crypto_alloc_hash(X, Y, Z) crypto_alloc_tfm(X, mode) ++ #define crypto_hash_digestsize(X) crypto_tfm_alg_digestsize(X) ++ #define crypto_hash_digest(W, X, Y, Z) \ ++ crypto_digest_digest((W)->tfm, X, sg_num, Z) ++ ++ /* Asymmetric Cipher */ ++ #define crypto_has_cipher(X, Y, Z) crypto_alg_available(X, 0) ++ ++ /* Compression */ ++ #define crypto_has_comp(X, Y, Z) crypto_alg_available(X, 0) ++ #define crypto_comp_tfm(X) X ++ #define crypto_comp_cast(X) X ++ #define crypto_alloc_comp(X, Y, Z) crypto_alloc_tfm(X, mode) ++#else ++ #define ecb(X) "ecb(" #X ")" ++ #define cbc(X) "cbc(" #X ")" ++ #define hmac(X) "hmac(" #X ")" ++#endif /* if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,19) */ ++ ++struct crypto_details ++{ ++ char *alg_name; ++ int mode; ++ int sw_type; ++}; ++ ++/* ++ * This needs to be kept updated with CRYPTO_xxx list (cryptodev.h). ++ * If the Algorithm is not supported, then insert a {NULL, 0, 0} entry. ++ * ++ * IMPORTANT: The index to the array IS CRYPTO_xxx. ++ */ ++static struct crypto_details crypto_details[CRYPTO_ALGORITHM_MAX + 1] = { ++ { NULL, 0, 0 }, ++ /* CRYPTO_xxx index starts at 1 */ ++ { cbc(des), CRYPTO_TFM_MODE_CBC, SW_TYPE_BLKCIPHER }, ++ { cbc(des3_ede), CRYPTO_TFM_MODE_CBC, SW_TYPE_BLKCIPHER }, ++ { cbc(blowfish), CRYPTO_TFM_MODE_CBC, SW_TYPE_BLKCIPHER }, ++ { cbc(cast5), CRYPTO_TFM_MODE_CBC, SW_TYPE_BLKCIPHER }, ++ { cbc(skipjack), CRYPTO_TFM_MODE_CBC, SW_TYPE_BLKCIPHER }, ++ { hmac(md5), 0, SW_TYPE_HMAC }, ++ { hmac(sha1), 0, SW_TYPE_HMAC }, ++ { hmac(ripemd160), 0, SW_TYPE_HMAC }, ++ { "md5-kpdk??", 0, SW_TYPE_HASH }, ++ { "sha1-kpdk??", 0, SW_TYPE_HASH }, ++ { cbc(aes), CRYPTO_TFM_MODE_CBC, SW_TYPE_BLKCIPHER }, ++ { ecb(arc4), CRYPTO_TFM_MODE_ECB, SW_TYPE_BLKCIPHER }, ++ { "md5", 0, SW_TYPE_HASH }, ++ { "sha1", 0, SW_TYPE_HASH }, ++ { hmac(digest_null), 0, SW_TYPE_HMAC }, ++ { cbc(cipher_null), CRYPTO_TFM_MODE_CBC, SW_TYPE_BLKCIPHER }, ++ { "deflate", 0, SW_TYPE_COMP }, ++ { hmac(sha256), 0, SW_TYPE_HMAC }, ++ { hmac(sha384), 0, SW_TYPE_HMAC }, ++ { hmac(sha512), 0, SW_TYPE_HMAC }, ++ { cbc(camellia), CRYPTO_TFM_MODE_CBC, SW_TYPE_BLKCIPHER }, ++ { "sha256", 0, SW_TYPE_HASH }, ++ { "sha384", 0, SW_TYPE_HASH }, ++ { "sha512", 0, SW_TYPE_HASH }, ++ { "ripemd160", 0, SW_TYPE_HASH }, ++}; ++ ++int32_t swcr_id = -1; ++module_param(swcr_id, int, 0444); ++MODULE_PARM_DESC(swcr_id, "Read-Only OCF ID for cryptosoft driver"); ++ ++int swcr_fail_if_compression_grows = 1; ++module_param(swcr_fail_if_compression_grows, int, 0644); ++MODULE_PARM_DESC(swcr_fail_if_compression_grows, ++ "Treat compression that results in more data as a failure"); ++ ++static struct swcr_data **swcr_sessions = NULL; ++static u_int32_t swcr_sesnum = 0; ++ ++static int swcr_process(device_t, struct cryptop *, int); ++static int swcr_newsession(device_t, u_int32_t *, struct cryptoini *); ++static int swcr_freesession(device_t, u_int64_t); ++ ++static device_method_t swcr_methods = { ++ /* crypto device methods */ ++ DEVMETHOD(cryptodev_newsession, swcr_newsession), ++ DEVMETHOD(cryptodev_freesession,swcr_freesession), ++ DEVMETHOD(cryptodev_process, swcr_process), ++}; ++ ++#define debug swcr_debug ++int swcr_debug = 0; ++module_param(swcr_debug, int, 0644); ++MODULE_PARM_DESC(swcr_debug, "Enable debug"); ++ ++/* ++ * Generate a new software session. ++ */ ++static int ++swcr_newsession(device_t dev, u_int32_t *sid, struct cryptoini *cri) ++{ ++ struct swcr_data **swd; ++ u_int32_t i; ++ int error; ++ char *algo; ++ int mode, sw_type; ++ ++ dprintk("%s()\n", __FUNCTION__); ++ if (sid == NULL || cri == NULL) { ++ dprintk("%s,%d - EINVAL\n", __FILE__, __LINE__); ++ return EINVAL; ++ } ++ ++ if (swcr_sessions) { ++ for (i = 1; i < swcr_sesnum; i++) ++ if (swcr_sessions[i] == NULL) ++ break; ++ } else ++ i = 1; /* NB: to silence compiler warning */ ++ ++ if (swcr_sessions == NULL || i == swcr_sesnum) { ++ if (swcr_sessions == NULL) { ++ i = 1; /* We leave swcr_sessions[0] empty */ ++ swcr_sesnum = CRYPTO_SW_SESSIONS; ++ } else ++ swcr_sesnum *= 2; ++ ++ swd = kmalloc(swcr_sesnum * sizeof(struct swcr_data *), SLAB_ATOMIC); ++ if (swd == NULL) { ++ /* Reset session number */ ++ if (swcr_sesnum == CRYPTO_SW_SESSIONS) ++ swcr_sesnum = 0; ++ else ++ swcr_sesnum /= 2; ++ dprintk("%s,%d: ENOBUFS\n", __FILE__, __LINE__); ++ return ENOBUFS; ++ } ++ memset(swd, 0, swcr_sesnum * sizeof(struct swcr_data *)); ++ ++ /* Copy existing sessions */ ++ if (swcr_sessions) { ++ memcpy(swd, swcr_sessions, ++ (swcr_sesnum / 2) * sizeof(struct swcr_data *)); ++ kfree(swcr_sessions); ++ } ++ ++ swcr_sessions = swd; ++ } ++ ++ swd = &swcr_sessions[i]; ++ *sid = i; ++ ++ while (cri) { ++ *swd = (struct swcr_data *) kmalloc(sizeof(struct swcr_data), ++ SLAB_ATOMIC); ++ if (*swd == NULL) { ++ swcr_freesession(NULL, i); ++ dprintk("%s,%d: ENOBUFS\n", __FILE__, __LINE__); ++ return ENOBUFS; ++ } ++ memset(*swd, 0, sizeof(struct swcr_data)); ++ ++ if (cri->cri_alg > CRYPTO_ALGORITHM_MAX) { ++ printk("cryptosoft: Unknown algorithm 0x%x\n", cri->cri_alg); ++ swcr_freesession(NULL, i); ++ return EINVAL; ++ } ++ ++ algo = crypto_details[cri->cri_alg].alg_name; ++ if (!algo || !*algo) { ++ printk("cryptosoft: Unsupported algorithm 0x%x\n", cri->cri_alg); ++ swcr_freesession(NULL, i); ++ return EINVAL; ++ } ++ ++ mode = crypto_details[cri->cri_alg].mode; ++ sw_type = crypto_details[cri->cri_alg].sw_type; ++ ++ /* Algorithm specific configuration */ ++ switch (cri->cri_alg) { ++ case CRYPTO_NULL_CBC: ++ cri->cri_klen = 0; /* make it work with crypto API */ ++ break; ++ default: ++ break; ++ } ++ ++ if (sw_type == SW_TYPE_BLKCIPHER) { ++ dprintk("%s crypto_alloc_blkcipher(%s, 0x%x)\n", __FUNCTION__, ++ algo, mode); ++ ++ (*swd)->sw_tfm = crypto_blkcipher_tfm( ++ crypto_alloc_blkcipher(algo, 0, ++ CRYPTO_ALG_ASYNC)); ++ if (!(*swd)->sw_tfm) { ++ dprintk("cryptosoft: crypto_alloc_blkcipher failed(%s,0x%x)\n", ++ algo,mode); ++ swcr_freesession(NULL, i); ++ return EINVAL; ++ } ++ ++ if (debug) { ++ dprintk("%s key:cri->cri_klen=%d,(cri->cri_klen + 7)/8=%d", ++ __FUNCTION__,cri->cri_klen,(cri->cri_klen + 7)/8); ++ for (i = 0; i < (cri->cri_klen + 7) / 8; i++) ++ { ++ dprintk("%s0x%x", (i % 8) ? " " : "\n ",cri->cri_key[i]); ++ } ++ dprintk("\n"); ++ } ++ error = crypto_blkcipher_setkey( ++ crypto_blkcipher_cast((*swd)->sw_tfm), cri->cri_key, ++ (cri->cri_klen + 7) / 8); ++ if (error) { ++ printk("cryptosoft: setkey failed %d (crt_flags=0x%x)\n", error, ++ (*swd)->sw_tfm->crt_flags); ++ swcr_freesession(NULL, i); ++ return error; ++ } ++ } else if (sw_type == SW_TYPE_HMAC || sw_type == SW_TYPE_HASH) { ++ dprintk("%s crypto_alloc_hash(%s, 0x%x)\n", __FUNCTION__, ++ algo, mode); ++ ++ (*swd)->sw_tfm = crypto_hash_tfm( ++ crypto_alloc_hash(algo, 0, CRYPTO_ALG_ASYNC)); ++ ++ if (!(*swd)->sw_tfm) { ++ dprintk("cryptosoft: crypto_alloc_hash failed(%s,0x%x)\n", ++ algo, mode); ++ swcr_freesession(NULL, i); ++ return EINVAL; ++ } ++ ++ (*swd)->u.hmac.sw_klen = (cri->cri_klen + 7) / 8; ++ (*swd)->u.hmac.sw_key = (char *)kmalloc((*swd)->u.hmac.sw_klen, ++ SLAB_ATOMIC); ++ if ((*swd)->u.hmac.sw_key == NULL) { ++ swcr_freesession(NULL, i); ++ dprintk("%s,%d: ENOBUFS\n", __FILE__, __LINE__); ++ return ENOBUFS; ++ } ++ memcpy((*swd)->u.hmac.sw_key, cri->cri_key, (*swd)->u.hmac.sw_klen); ++ if (cri->cri_mlen) { ++ (*swd)->u.hmac.sw_mlen = cri->cri_mlen; ++ } else { ++ (*swd)->u.hmac.sw_mlen = ++ crypto_hash_digestsize( ++ crypto_hash_cast((*swd)->sw_tfm)); ++ } ++ } else if (sw_type == SW_TYPE_COMP) { ++ (*swd)->sw_tfm = crypto_comp_tfm( ++ crypto_alloc_comp(algo, 0, CRYPTO_ALG_ASYNC)); ++ if (!(*swd)->sw_tfm) { ++ dprintk("cryptosoft: crypto_alloc_comp failed(%s,0x%x)\n", ++ algo, mode); ++ swcr_freesession(NULL, i); ++ return EINVAL; ++ } ++ (*swd)->u.sw_comp_buf = kmalloc(CRYPTO_MAX_DATA_LEN, SLAB_ATOMIC); ++ if ((*swd)->u.sw_comp_buf == NULL) { ++ swcr_freesession(NULL, i); ++ dprintk("%s,%d: ENOBUFS\n", __FILE__, __LINE__); ++ return ENOBUFS; ++ } ++ } else { ++ printk("cryptosoft: Unhandled sw_type %d\n", sw_type); ++ swcr_freesession(NULL, i); ++ return EINVAL; ++ } ++ ++ (*swd)->sw_alg = cri->cri_alg; ++ (*swd)->sw_type = sw_type; ++ ++ cri = cri->cri_next; ++ swd = &((*swd)->sw_next); ++ } ++ return 0; ++} ++ ++/* ++ * Free a session. ++ */ ++static int ++swcr_freesession(device_t dev, u_int64_t tid) ++{ ++ struct swcr_data *swd; ++ u_int32_t sid = CRYPTO_SESID2LID(tid); ++ ++ dprintk("%s()\n", __FUNCTION__); ++ if (sid > swcr_sesnum || swcr_sessions == NULL || ++ swcr_sessions[sid] == NULL) { ++ dprintk("%s,%d: EINVAL\n", __FILE__, __LINE__); ++ return(EINVAL); ++ } ++ ++ /* Silently accept and return */ ++ if (sid == 0) ++ return(0); ++ ++ while ((swd = swcr_sessions[sid]) != NULL) { ++ swcr_sessions[sid] = swd->sw_next; ++ if (swd->sw_tfm) ++ crypto_free_tfm(swd->sw_tfm); ++ if (swd->sw_type == SW_TYPE_COMP) { ++ if (swd->u.sw_comp_buf) ++ kfree(swd->u.sw_comp_buf); ++ } else { ++ if (swd->u.hmac.sw_key) ++ kfree(swd->u.hmac.sw_key); ++ } ++ kfree(swd); ++ } ++ return 0; ++} ++ ++/* ++ * Process a software request. ++ */ ++static int ++swcr_process(device_t dev, struct cryptop *crp, int hint) ++{ ++ struct cryptodesc *crd; ++ struct swcr_data *sw; ++ u_int32_t lid; ++#define SCATTERLIST_MAX 16 ++ struct scatterlist sg[SCATTERLIST_MAX]; ++ int sg_num, sg_len, skip; ++ struct sk_buff *skb = NULL; ++ struct uio *uiop = NULL; ++ ++ dprintk("%s()\n", __FUNCTION__); ++ /* Sanity check */ ++ if (crp == NULL) { ++ dprintk("%s,%d: EINVAL\n", __FILE__, __LINE__); ++ return EINVAL; ++ } ++ ++ crp->crp_etype = 0; ++ ++ if (crp->crp_desc == NULL || crp->crp_buf == NULL) { ++ dprintk("%s,%d: EINVAL\n", __FILE__, __LINE__); ++ crp->crp_etype = EINVAL; ++ goto done; ++ } ++ ++ lid = crp->crp_sid & 0xffffffff; ++ if (lid >= swcr_sesnum || lid == 0 || swcr_sessions == NULL || ++ swcr_sessions[lid] == NULL) { ++ crp->crp_etype = ENOENT; ++ dprintk("%s,%d: ENOENT\n", __FILE__, __LINE__); ++ goto done; ++ } ++ ++ /* ++ * do some error checking outside of the loop for SKB and IOV processing ++ * this leaves us with valid skb or uiop pointers for later ++ */ ++ if (crp->crp_flags & CRYPTO_F_SKBUF) { ++ skb = (struct sk_buff *) crp->crp_buf; ++ if (skb_shinfo(skb)->nr_frags >= SCATTERLIST_MAX) { ++ printk("%s,%d: %d nr_frags > SCATTERLIST_MAX", __FILE__, __LINE__, ++ skb_shinfo(skb)->nr_frags); ++ goto done; ++ } ++ } else if (crp->crp_flags & CRYPTO_F_IOV) { ++ uiop = (struct uio *) crp->crp_buf; ++ if (uiop->uio_iovcnt > SCATTERLIST_MAX) { ++ printk("%s,%d: %d uio_iovcnt > SCATTERLIST_MAX", __FILE__, __LINE__, ++ uiop->uio_iovcnt); ++ goto done; ++ } ++ } ++ ++ /* Go through crypto descriptors, processing as we go */ ++ for (crd = crp->crp_desc; crd; crd = crd->crd_next) { ++ /* ++ * Find the crypto context. ++ * ++ * XXX Note that the logic here prevents us from having ++ * XXX the same algorithm multiple times in a session ++ * XXX (or rather, we can but it won't give us the right ++ * XXX results). To do that, we'd need some way of differentiating ++ * XXX between the various instances of an algorithm (so we can ++ * XXX locate the correct crypto context). ++ */ ++ for (sw = swcr_sessions[lid]; sw && sw->sw_alg != crd->crd_alg; ++ sw = sw->sw_next) ++ ; ++ ++ /* No such context ? */ ++ if (sw == NULL) { ++ crp->crp_etype = EINVAL; ++ dprintk("%s,%d: EINVAL\n", __FILE__, __LINE__); ++ goto done; ++ } ++ ++ skip = crd->crd_skip; ++ ++ /* ++ * setup the SG list skip from the start of the buffer ++ */ ++ memset(sg, 0, sizeof(sg)); ++ if (crp->crp_flags & CRYPTO_F_SKBUF) { ++ int i, len; ++ ++ sg_num = 0; ++ sg_len = 0; ++ ++ if (skip < skb_headlen(skb)) { ++ len = skb_headlen(skb) - skip; ++ if (len + sg_len > crd->crd_len) ++ len = crd->crd_len - sg_len; ++ sg_set_page(&sg[sg_num], ++ virt_to_page(skb->data + skip), len, ++ offset_in_page(skb->data + skip)); ++ sg_len += len; ++ sg_num++; ++ skip = 0; ++ } else ++ skip -= skb_headlen(skb); ++ ++ for (i = 0; sg_len < crd->crd_len && ++ i < skb_shinfo(skb)->nr_frags && ++ sg_num < SCATTERLIST_MAX; i++) { ++ if (skip < skb_shinfo(skb)->frags[i].size) { ++ len = skb_shinfo(skb)->frags[i].size - skip; ++ if (len + sg_len > crd->crd_len) ++ len = crd->crd_len - sg_len; ++ sg_set_page(&sg[sg_num], ++ skb_shinfo(skb)->frags[i].page, ++ len, ++ skb_shinfo(skb)->frags[i].page_offset + skip); ++ sg_len += len; ++ sg_num++; ++ skip = 0; ++ } else ++ skip -= skb_shinfo(skb)->frags[i].size; ++ } ++ } else if (crp->crp_flags & CRYPTO_F_IOV) { ++ int len; ++ ++ sg_len = 0; ++ for (sg_num = 0; sg_len <= crd->crd_len && ++ sg_num < uiop->uio_iovcnt && ++ sg_num < SCATTERLIST_MAX; sg_num++) { ++ if (skip <= uiop->uio_iov[sg_num].iov_len) { ++ len = uiop->uio_iov[sg_num].iov_len - skip; ++ if (len + sg_len > crd->crd_len) ++ len = crd->crd_len - sg_len; ++ sg_set_page(&sg[sg_num], ++ virt_to_page(uiop->uio_iov[sg_num].iov_base+skip), ++ len, ++ offset_in_page(uiop->uio_iov[sg_num].iov_base+skip)); ++ sg_len += len; ++ skip = 0; ++ } else ++ skip -= uiop->uio_iov[sg_num].iov_len; ++ } ++ } else { ++ sg_len = (crp->crp_ilen - skip); ++ if (sg_len > crd->crd_len) ++ sg_len = crd->crd_len; ++ sg_set_page(&sg[0], virt_to_page(crp->crp_buf + skip), ++ sg_len, offset_in_page(crp->crp_buf + skip)); ++ sg_num = 1; ++ } ++ ++ ++ switch (sw->sw_type) { ++ case SW_TYPE_BLKCIPHER: { ++ unsigned char iv[EALG_MAX_BLOCK_LEN]; ++ unsigned char *ivp = iv; ++ int ivsize = ++ crypto_blkcipher_ivsize(crypto_blkcipher_cast(sw->sw_tfm)); ++ struct blkcipher_desc desc; ++ ++ if (sg_len < crypto_blkcipher_blocksize( ++ crypto_blkcipher_cast(sw->sw_tfm))) { ++ crp->crp_etype = EINVAL; ++ dprintk("%s,%d: EINVAL len %d < %d\n", __FILE__, __LINE__, ++ sg_len, crypto_blkcipher_blocksize( ++ crypto_blkcipher_cast(sw->sw_tfm))); ++ goto done; ++ } ++ ++ if (ivsize > sizeof(iv)) { ++ crp->crp_etype = EINVAL; ++ dprintk("%s,%d: EINVAL\n", __FILE__, __LINE__); ++ goto done; ++ } ++ ++ if (crd->crd_flags & CRD_F_KEY_EXPLICIT) { ++ int i, error; ++ ++ if (debug) { ++ dprintk("%s key:", __FUNCTION__); ++ for (i = 0; i < (crd->crd_klen + 7) / 8; i++) ++ dprintk("%s0x%x", (i % 8) ? " " : "\n ", ++ crd->crd_key[i]); ++ dprintk("\n"); ++ } ++ error = crypto_blkcipher_setkey( ++ crypto_blkcipher_cast(sw->sw_tfm), crd->crd_key, ++ (crd->crd_klen + 7) / 8); ++ if (error) { ++ dprintk("cryptosoft: setkey failed %d (crt_flags=0x%x)\n", ++ error, sw->sw_tfm->crt_flags); ++ crp->crp_etype = -error; ++ } ++ } ++ ++ memset(&desc, 0, sizeof(desc)); ++ desc.tfm = crypto_blkcipher_cast(sw->sw_tfm); ++ ++ if (crd->crd_flags & CRD_F_ENCRYPT) { /* encrypt */ ++ ++ if (crd->crd_flags & CRD_F_IV_EXPLICIT) { ++ ivp = crd->crd_iv; ++ } else { ++ get_random_bytes(ivp, ivsize); ++ } ++ /* ++ * do we have to copy the IV back to the buffer ? ++ */ ++ if ((crd->crd_flags & CRD_F_IV_PRESENT) == 0) { ++ crypto_copyback(crp->crp_flags, crp->crp_buf, ++ crd->crd_inject, ivsize, (caddr_t)ivp); ++ } ++ desc.info = ivp; ++ crypto_blkcipher_encrypt_iv(&desc, sg, sg, sg_len); ++ ++ } else { /*decrypt */ ++ ++ if (crd->crd_flags & CRD_F_IV_EXPLICIT) { ++ ivp = crd->crd_iv; ++ } else { ++ crypto_copydata(crp->crp_flags, crp->crp_buf, ++ crd->crd_inject, ivsize, (caddr_t)ivp); ++ } ++ desc.info = ivp; ++ crypto_blkcipher_decrypt_iv(&desc, sg, sg, sg_len); ++ } ++ } break; ++ case SW_TYPE_HMAC: ++ case SW_TYPE_HASH: ++ { ++ char result[HASH_MAX_LEN]; ++ struct hash_desc desc; ++ ++ /* check we have room for the result */ ++ if (crp->crp_ilen - crd->crd_inject < sw->u.hmac.sw_mlen) { ++ dprintk( ++ "cryptosoft: EINVAL crp_ilen=%d, len=%d, inject=%d digestsize=%d\n", ++ crp->crp_ilen, crd->crd_skip + sg_len, crd->crd_inject, ++ sw->u.hmac.sw_mlen); ++ crp->crp_etype = EINVAL; ++ goto done; ++ } ++ ++ memset(&desc, 0, sizeof(desc)); ++ desc.tfm = crypto_hash_cast(sw->sw_tfm); ++ ++ memset(result, 0, sizeof(result)); ++ ++ if (sw->sw_type == SW_TYPE_HMAC) { ++#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,19) ++ crypto_hmac(sw->sw_tfm, sw->u.hmac.sw_key, &sw->u.hmac.sw_klen, ++ sg, sg_num, result); ++#else ++ crypto_hash_setkey(desc.tfm, sw->u.hmac.sw_key, ++ sw->u.hmac.sw_klen); ++ crypto_hash_digest(&desc, sg, sg_len, result); ++#endif /* #if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,19) */ ++ ++ } else { /* SW_TYPE_HASH */ ++ crypto_hash_digest(&desc, sg, sg_len, result); ++ } ++ ++ crypto_copyback(crp->crp_flags, crp->crp_buf, ++ crd->crd_inject, sw->u.hmac.sw_mlen, result); ++ } ++ break; ++ ++ case SW_TYPE_COMP: { ++ void *ibuf = NULL; ++ void *obuf = sw->u.sw_comp_buf; ++ int ilen = sg_len, olen = CRYPTO_MAX_DATA_LEN; ++ int ret = 0; ++ ++ /* ++ * we need to use an additional copy if there is more than one ++ * input chunk since the kernel comp routines do not handle ++ * SG yet. Otherwise we just use the input buffer as is. ++ * Rather than allocate another buffer we just split the tmp ++ * buffer we already have. ++ * Perhaps we should just use zlib directly ? ++ */ ++ if (sg_num > 1) { ++ int blk; ++ ++ ibuf = obuf; ++ for (blk = 0; blk < sg_num; blk++) { ++ memcpy(obuf, sg_virt(&sg[blk]), ++ sg[blk].length); ++ obuf += sg[blk].length; ++ } ++ olen -= sg_len; ++ } else ++ ibuf = sg_virt(&sg[0]); ++ ++ if (crd->crd_flags & CRD_F_ENCRYPT) { /* compress */ ++ ret = crypto_comp_compress(crypto_comp_cast(sw->sw_tfm), ++ ibuf, ilen, obuf, &olen); ++ if (!ret && olen > crd->crd_len) { ++ dprintk("cryptosoft: ERANGE compress %d into %d\n", ++ crd->crd_len, olen); ++ if (swcr_fail_if_compression_grows) ++ ret = ERANGE; ++ } ++ } else { /* decompress */ ++ ret = crypto_comp_decompress(crypto_comp_cast(sw->sw_tfm), ++ ibuf, ilen, obuf, &olen); ++ if (!ret && (olen + crd->crd_inject) > crp->crp_olen) { ++ dprintk("cryptosoft: ETOOSMALL decompress %d into %d, " ++ "space for %d,at offset %d\n", ++ crd->crd_len, olen, crp->crp_olen, crd->crd_inject); ++ ret = ETOOSMALL; ++ } ++ } ++ if (ret) ++ dprintk("%s,%d: ret = %d\n", __FILE__, __LINE__, ret); ++ ++ /* ++ * on success copy result back, ++ * linux crpyto API returns -errno, we need to fix that ++ */ ++ crp->crp_etype = ret < 0 ? -ret : ret; ++ if (ret == 0) { ++ /* copy back the result and return it's size */ ++ crypto_copyback(crp->crp_flags, crp->crp_buf, ++ crd->crd_inject, olen, obuf); ++ crp->crp_olen = olen; ++ } ++ ++ ++ } break; ++ ++ default: ++ /* Unknown/unsupported algorithm */ ++ dprintk("%s,%d: EINVAL\n", __FILE__, __LINE__); ++ crp->crp_etype = EINVAL; ++ goto done; ++ } ++ } ++ ++done: ++ crypto_done(crp); ++ return 0; ++} ++ ++static int ++cryptosoft_init(void) ++{ ++ int i, sw_type, mode; ++ char *algo; ++ ++ dprintk("%s(%p)\n", __FUNCTION__, cryptosoft_init); ++ ++ softc_device_init(&swcr_softc, "cryptosoft", 0, swcr_methods); ++ ++ swcr_id = crypto_get_driverid(softc_get_device(&swcr_softc), ++ CRYPTOCAP_F_SOFTWARE | CRYPTOCAP_F_SYNC); ++ if (swcr_id < 0) { ++ printk("Software crypto device cannot initialize!"); ++ return -ENODEV; ++ } ++ ++#define REGISTER(alg) \ ++ crypto_register(swcr_id, alg, 0,0); ++ ++ for (i = CRYPTO_ALGORITHM_MIN; i <= CRYPTO_ALGORITHM_MAX; ++i) ++ { ++ ++ algo = crypto_details[i].alg_name; ++ if (!algo || !*algo) ++ { ++ dprintk("%s:Algorithm %d not supported\n", __FUNCTION__, i); ++ continue; ++ } ++ ++ mode = crypto_details[i].mode; ++ sw_type = crypto_details[i].sw_type; ++ ++ switch (sw_type) ++ { ++ case SW_TYPE_CIPHER: ++ if (crypto_has_cipher(algo, 0, CRYPTO_ALG_ASYNC)) ++ { ++ REGISTER(i); ++ } ++ else ++ { ++ dprintk("%s:CIPHER algorithm %d:'%s' not supported\n", ++ __FUNCTION__, i, algo); ++ } ++ break; ++ case SW_TYPE_HMAC: ++ if (crypto_has_hash(algo, 0, CRYPTO_ALG_ASYNC)) ++ { ++ REGISTER(i); ++ } ++ else ++ { ++ dprintk("%s:HMAC algorithm %d:'%s' not supported\n", ++ __FUNCTION__, i, algo); ++ } ++ break; ++ case SW_TYPE_HASH: ++ if (crypto_has_hash(algo, 0, CRYPTO_ALG_ASYNC)) ++ { ++ REGISTER(i); ++ } ++ else ++ { ++ dprintk("%s:HASH algorithm %d:'%s' not supported\n", ++ __FUNCTION__, i, algo); ++ } ++ break; ++ case SW_TYPE_COMP: ++ if (crypto_has_comp(algo, 0, CRYPTO_ALG_ASYNC)) ++ { ++ REGISTER(i); ++ } ++ else ++ { ++ dprintk("%s:COMP algorithm %d:'%s' not supported\n", ++ __FUNCTION__, i, algo); ++ } ++ break; ++ case SW_TYPE_BLKCIPHER: ++ if (crypto_has_blkcipher(algo, 0, CRYPTO_ALG_ASYNC)) ++ { ++ REGISTER(i); ++ } ++ else ++ { ++ dprintk("%s:BLKCIPHER algorithm %d:'%s' not supported\n", ++ __FUNCTION__, i, algo); ++ } ++ break; ++ default: ++ dprintk( ++ "%s:Algorithm Type %d not supported (algorithm %d:'%s')\n", ++ __FUNCTION__, sw_type, i, algo); ++ break; ++ } ++ } ++ ++ return(0); ++} ++ ++static void ++cryptosoft_exit(void) ++{ ++ dprintk("%s()\n", __FUNCTION__); ++ crypto_unregister_all(swcr_id); ++ swcr_id = -1; ++} ++ ++module_init(cryptosoft_init); ++module_exit(cryptosoft_exit); ++ ++MODULE_LICENSE("Dual BSD/GPL"); ++MODULE_AUTHOR("David McCullough <david_mccullough@securecomputing.com>"); ++MODULE_DESCRIPTION("Cryptosoft (OCF module for kernel crypto)"); +diff -Nur linux-2.6.30.orig/crypto/ocf/ep80579/icp_asym.c linux-2.6.30/crypto/ocf/ep80579/icp_asym.c +--- linux-2.6.30.orig/crypto/ocf/ep80579/icp_asym.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/ep80579/icp_asym.c 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,1375 @@ ++/*************************************************************************** ++ * ++ * This file is provided under a dual BSD/GPLv2 license. When using or ++ * redistributing this file, you may do so under either license. ++ * ++ * GPL LICENSE SUMMARY ++ * ++ * Copyright(c) 2007,2008 Intel Corporation. All rights reserved. ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of version 2 of the GNU General Public License as ++ * published by the Free Software Foundation. ++ * ++ * This program is distributed in the hope that it will be useful, but ++ * WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ * General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. ++ * The full GNU General Public License is included in this distribution ++ * in the file called LICENSE.GPL. ++ * ++ * Contact Information: ++ * Intel Corporation ++ * ++ * BSD LICENSE ++ * ++ * Copyright(c) 2007,2008 Intel Corporation. All rights reserved. ++ * All rights reserved. ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions ++ * are met: ++ * ++ * * Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * * Redistributions in binary form must reproduce the above copyright ++ * notice, this list of conditions and the following disclaimer in ++ * the documentation and/or other materials provided with the ++ * distribution. ++ * * Neither the name of Intel Corporation nor the names of its ++ * contributors may be used to endorse or promote products derived ++ * from this software without specific prior written permission. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS ++ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT ++ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR ++ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT ++ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, ++ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT ++ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, ++ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY ++ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT ++ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE ++ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. ++ * ++ * ++ * version: Security.L.1.0.130 ++ * ++ ***************************************************************************/ ++ ++#include "icp_ocf.h" ++ ++/*The following define values (containing the word 'INDEX') are used to find ++the index of each input buffer of the crypto_kop struct (see OCF cryptodev.h). ++These values were found through analysis of the OCF OpenSSL patch. If the ++calling program uses different input buffer positions, these defines will have ++to be changed.*/ ++ ++/*DIFFIE HELLMAN buffer index values*/ ++#define ICP_DH_KRP_PARAM_PRIME_INDEX (0) ++#define ICP_DH_KRP_PARAM_BASE_INDEX (1) ++#define ICP_DH_KRP_PARAM_PRIVATE_VALUE_INDEX (2) ++#define ICP_DH_KRP_PARAM_RESULT_INDEX (3) ++ ++/*MOD EXP buffer index values*/ ++#define ICP_MOD_EXP_KRP_PARAM_BASE_INDEX (0) ++#define ICP_MOD_EXP_KRP_PARAM_EXPONENT_INDEX (1) ++#define ICP_MOD_EXP_KRP_PARAM_MODULUS_INDEX (2) ++#define ICP_MOD_EXP_KRP_PARAM_RESULT_INDEX (3) ++ ++#define SINGLE_BYTE_VALUE (4) ++ ++/*MOD EXP CRT buffer index values*/ ++#define ICP_MOD_EXP_CRT_KRP_PARAM_PRIME_P_INDEX (0) ++#define ICP_MOD_EXP_CRT_KRP_PARAM_PRIME_Q_INDEX (1) ++#define ICP_MOD_EXP_CRT_KRP_PARAM_I_INDEX (2) ++#define ICP_MOD_EXP_CRT_KRP_PARAM_EXPONENT_DP_INDEX (3) ++#define ICP_MOD_EXP_CRT_KRP_PARAM_EXPONENT_DQ_INDEX (4) ++#define ICP_MOD_EXP_CRT_KRP_PARAM_COEFF_QINV_INDEX (5) ++#define ICP_MOD_EXP_CRT_KRP_PARAM_RESULT_INDEX (6) ++ ++/*DSA sign buffer index values*/ ++#define ICP_DSA_SIGN_KRP_PARAM_DGST_INDEX (0) ++#define ICP_DSA_SIGN_KRP_PARAM_PRIME_P_INDEX (1) ++#define ICP_DSA_SIGN_KRP_PARAM_PRIME_Q_INDEX (2) ++#define ICP_DSA_SIGN_KRP_PARAM_G_INDEX (3) ++#define ICP_DSA_SIGN_KRP_PARAM_X_INDEX (4) ++#define ICP_DSA_SIGN_KRP_PARAM_R_RESULT_INDEX (5) ++#define ICP_DSA_SIGN_KRP_PARAM_S_RESULT_INDEX (6) ++ ++/*DSA verify buffer index values*/ ++#define ICP_DSA_VERIFY_KRP_PARAM_DGST_INDEX (0) ++#define ICP_DSA_VERIFY_KRP_PARAM_PRIME_P_INDEX (1) ++#define ICP_DSA_VERIFY_KRP_PARAM_PRIME_Q_INDEX (2) ++#define ICP_DSA_VERIFY_KRP_PARAM_G_INDEX (3) ++#define ICP_DSA_VERIFY_KRP_PARAM_PUBKEY_INDEX (4) ++#define ICP_DSA_VERIFY_KRP_PARAM_SIG_R_INDEX (5) ++#define ICP_DSA_VERIFY_KRP_PARAM_SIG_S_INDEX (6) ++ ++/*DSA sign prime Q vs random number K size check values*/ ++#define DONT_RUN_LESS_THAN_CHECK (0) ++#define FAIL_A_IS_GREATER_THAN_B (1) ++#define FAIL_A_IS_EQUAL_TO_B (1) ++#define SUCCESS_A_IS_LESS_THAN_B (0) ++#define DSA_SIGN_RAND_GEN_VAL_CHECK_MAX_ITERATIONS (500) ++ ++/* We need to set a cryptokp success value just in case it is set or allocated ++ and not set to zero outside of this module */ ++#define CRYPTO_OP_SUCCESS (0) ++ ++static int icp_ocfDrvDHComputeKey(struct cryptkop *krp); ++ ++static int icp_ocfDrvModExp(struct cryptkop *krp); ++ ++static int icp_ocfDrvModExpCRT(struct cryptkop *krp); ++ ++static int ++icp_ocfDrvCheckALessThanB(CpaFlatBuffer * pK, CpaFlatBuffer * pQ, int *doCheck); ++ ++static int icp_ocfDrvDsaSign(struct cryptkop *krp); ++ ++static int icp_ocfDrvDsaVerify(struct cryptkop *krp); ++ ++static void ++icp_ocfDrvDhP1CallBack(void *callbackTag, ++ CpaStatus status, ++ void *pOpData, CpaFlatBuffer * pLocalOctetStringPV); ++ ++static void ++icp_ocfDrvModExpCallBack(void *callbackTag, ++ CpaStatus status, ++ void *pOpData, CpaFlatBuffer * pResult); ++ ++static void ++icp_ocfDrvModExpCRTCallBack(void *callbackTag, ++ CpaStatus status, ++ void *pOpData, CpaFlatBuffer * pOutputData); ++ ++static void ++icp_ocfDrvDsaVerifyCallBack(void *callbackTag, ++ CpaStatus status, ++ void *pOpData, CpaBoolean verifyStatus); ++ ++static void ++icp_ocfDrvDsaRSSignCallBack(void *callbackTag, ++ CpaStatus status, ++ void *pOpData, ++ CpaBoolean protocolStatus, ++ CpaFlatBuffer * pR, CpaFlatBuffer * pS); ++ ++/* Name : icp_ocfDrvPkeProcess ++ * ++ * Description : This function will choose which PKE process to follow ++ * based on the input arguments ++ */ ++int icp_ocfDrvPkeProcess(device_t dev, struct cryptkop *krp, int hint) ++{ ++ CpaStatus lacStatus = CPA_STATUS_SUCCESS; ++ ++ if (NULL == krp) { ++ DPRINTK("%s(): Invalid input parameters, cryptkop = %p\n", ++ __FUNCTION__, krp); ++ return EINVAL; ++ } ++ ++ if (CPA_TRUE == atomic_read(&icp_ocfDrvIsExiting)) { ++ krp->krp_status = ECANCELED; ++ return ECANCELED; ++ } ++ ++ switch (krp->krp_op) { ++ case CRK_DH_COMPUTE_KEY: ++ DPRINTK("%s() doing DH_COMPUTE_KEY\n", __FUNCTION__); ++ lacStatus = icp_ocfDrvDHComputeKey(krp); ++ if (CPA_STATUS_SUCCESS != lacStatus) { ++ EPRINTK("%s(): icp_ocfDrvDHComputeKey failed " ++ "(%d).\n", __FUNCTION__, lacStatus); ++ krp->krp_status = ECANCELED; ++ return ECANCELED; ++ } ++ ++ break; ++ ++ case CRK_MOD_EXP: ++ DPRINTK("%s() doing MOD_EXP \n", __FUNCTION__); ++ lacStatus = icp_ocfDrvModExp(krp); ++ if (CPA_STATUS_SUCCESS != lacStatus) { ++ EPRINTK("%s(): icp_ocfDrvModExp failed (%d).\n", ++ __FUNCTION__, lacStatus); ++ krp->krp_status = ECANCELED; ++ return ECANCELED; ++ } ++ ++ break; ++ ++ case CRK_MOD_EXP_CRT: ++ DPRINTK("%s() doing MOD_EXP_CRT \n", __FUNCTION__); ++ lacStatus = icp_ocfDrvModExpCRT(krp); ++ if (CPA_STATUS_SUCCESS != lacStatus) { ++ EPRINTK("%s(): icp_ocfDrvModExpCRT " ++ "failed (%d).\n", __FUNCTION__, lacStatus); ++ krp->krp_status = ECANCELED; ++ return ECANCELED; ++ } ++ ++ break; ++ ++ case CRK_DSA_SIGN: ++ DPRINTK("%s() doing DSA_SIGN \n", __FUNCTION__); ++ lacStatus = icp_ocfDrvDsaSign(krp); ++ if (CPA_STATUS_SUCCESS != lacStatus) { ++ EPRINTK("%s(): icp_ocfDrvDsaSign " ++ "failed (%d).\n", __FUNCTION__, lacStatus); ++ krp->krp_status = ECANCELED; ++ return ECANCELED; ++ } ++ ++ break; ++ ++ case CRK_DSA_VERIFY: ++ DPRINTK("%s() doing DSA_VERIFY \n", __FUNCTION__); ++ lacStatus = icp_ocfDrvDsaVerify(krp); ++ if (CPA_STATUS_SUCCESS != lacStatus) { ++ EPRINTK("%s(): icp_ocfDrvDsaVerify " ++ "failed (%d).\n", __FUNCTION__, lacStatus); ++ krp->krp_status = ECANCELED; ++ return ECANCELED; ++ } ++ ++ break; ++ ++ default: ++ EPRINTK("%s(): Asymettric function not " ++ "supported (%d).\n", __FUNCTION__, krp->krp_op); ++ krp->krp_status = EOPNOTSUPP; ++ return EOPNOTSUPP; ++ } ++ ++ return ICP_OCF_DRV_STATUS_SUCCESS; ++} ++ ++/* Name : icp_ocfDrvSwapBytes ++ * ++ * Description : This function is used to swap the byte order of a buffer. ++ * It has been seen that in general we are passed little endian byte order ++ * buffers, but LAC only accepts big endian byte order buffers. ++ */ ++static void inline ++icp_ocfDrvSwapBytes(u_int8_t * num, u_int32_t buff_len_bytes) ++{ ++ ++ int i; ++ u_int8_t *end_ptr; ++ u_int8_t hold_val; ++ ++ end_ptr = num + (buff_len_bytes - 1); ++ buff_len_bytes = buff_len_bytes >> 1; ++ for (i = 0; i < buff_len_bytes; i++) { ++ hold_val = *num; ++ *num = *end_ptr; ++ num++; ++ *end_ptr = hold_val; ++ end_ptr--; ++ } ++} ++ ++/* Name : icp_ocfDrvDHComputeKey ++ * ++ * Description : This function will map Diffie Hellman calls from OCF ++ * to the LAC API. OCF uses this function for Diffie Hellman Phase1 and ++ * Phase2. LAC has a separate Diffie Hellman Phase2 call, however both phases ++ * break down to a modular exponentiation. ++ */ ++static int icp_ocfDrvDHComputeKey(struct cryptkop *krp) ++{ ++ CpaStatus lacStatus = CPA_STATUS_SUCCESS; ++ void *callbackTag = NULL; ++ CpaCyDhPhase1KeyGenOpData *pPhase1OpData = NULL; ++ CpaFlatBuffer *pLocalOctetStringPV = NULL; ++ uint32_t dh_prime_len_bytes = 0, dh_prime_len_bits = 0; ++ ++ /* Input checks - check prime is a multiple of 8 bits to allow for ++ allocation later */ ++ dh_prime_len_bits = ++ (krp->krp_param[ICP_DH_KRP_PARAM_PRIME_INDEX].crp_nbits); ++ ++ /* LAC can reject prime lengths based on prime key sizes, we just ++ need to make sure we can allocate space for the base and ++ exponent buffers correctly */ ++ if ((dh_prime_len_bits % NUM_BITS_IN_BYTE) != 0) { ++ APRINTK("%s(): Warning Prime number buffer size is not a " ++ "multiple of 8 bits\n", __FUNCTION__); ++ } ++ ++ /* Result storage space should be the same size as the prime as this ++ value can take up the same amount of storage space */ ++ if (dh_prime_len_bits != ++ krp->krp_param[ICP_DH_KRP_PARAM_RESULT_INDEX].crp_nbits) { ++ DPRINTK("%s(): Return Buffer must be the same size " ++ "as the Prime buffer\n", __FUNCTION__); ++ krp->krp_status = EINVAL; ++ return EINVAL; ++ } ++ /* Switch to size in bytes */ ++ BITS_TO_BYTES(dh_prime_len_bytes, dh_prime_len_bits); ++ ++ callbackTag = krp; ++ ++ pPhase1OpData = kmem_cache_zalloc(drvDH_zone, GFP_KERNEL); ++ if (NULL == pPhase1OpData) { ++ APRINTK("%s():Failed to get memory for key gen data\n", ++ __FUNCTION__); ++ krp->krp_status = ENOMEM; ++ return ENOMEM; ++ } ++ ++ pLocalOctetStringPV = kmem_cache_zalloc(drvFlatBuffer_zone, GFP_KERNEL); ++ if (NULL == pLocalOctetStringPV) { ++ APRINTK("%s():Failed to get memory for pLocalOctetStringPV\n", ++ __FUNCTION__); ++ kmem_cache_free(drvDH_zone, pPhase1OpData); ++ krp->krp_status = ENOMEM; ++ return ENOMEM; ++ } ++ ++ /* Link parameters */ ++ pPhase1OpData->primeP.pData = ++ krp->krp_param[ICP_DH_KRP_PARAM_PRIME_INDEX].crp_p; ++ ++ pPhase1OpData->primeP.dataLenInBytes = dh_prime_len_bytes; ++ ++ icp_ocfDrvSwapBytes(pPhase1OpData->primeP.pData, dh_prime_len_bytes); ++ ++ pPhase1OpData->baseG.pData = ++ krp->krp_param[ICP_DH_KRP_PARAM_BASE_INDEX].crp_p; ++ ++ BITS_TO_BYTES(pPhase1OpData->baseG.dataLenInBytes, ++ krp->krp_param[ICP_DH_KRP_PARAM_BASE_INDEX].crp_nbits); ++ ++ icp_ocfDrvSwapBytes(pPhase1OpData->baseG.pData, ++ pPhase1OpData->baseG.dataLenInBytes); ++ ++ pPhase1OpData->privateValueX.pData = ++ krp->krp_param[ICP_DH_KRP_PARAM_PRIVATE_VALUE_INDEX].crp_p; ++ ++ BITS_TO_BYTES(pPhase1OpData->privateValueX.dataLenInBytes, ++ krp->krp_param[ICP_DH_KRP_PARAM_PRIVATE_VALUE_INDEX]. ++ crp_nbits); ++ ++ icp_ocfDrvSwapBytes(pPhase1OpData->privateValueX.pData, ++ pPhase1OpData->privateValueX.dataLenInBytes); ++ ++ /* Output parameters */ ++ pLocalOctetStringPV->pData = ++ krp->krp_param[ICP_DH_KRP_PARAM_RESULT_INDEX].crp_p; ++ ++ BITS_TO_BYTES(pLocalOctetStringPV->dataLenInBytes, ++ krp->krp_param[ICP_DH_KRP_PARAM_RESULT_INDEX].crp_nbits); ++ ++ lacStatus = cpaCyDhKeyGenPhase1(CPA_INSTANCE_HANDLE_SINGLE, ++ icp_ocfDrvDhP1CallBack, ++ callbackTag, pPhase1OpData, ++ pLocalOctetStringPV); ++ ++ if (CPA_STATUS_SUCCESS != lacStatus) { ++ EPRINTK("%s(): DH Phase 1 Key Gen failed (%d).\n", ++ __FUNCTION__, lacStatus); ++ icp_ocfDrvFreeFlatBuffer(pLocalOctetStringPV); ++ kmem_cache_free(drvDH_zone, pPhase1OpData); ++ } ++ ++ return lacStatus; ++} ++ ++/* Name : icp_ocfDrvModExp ++ * ++ * Description : This function will map ordinary Modular Exponentiation calls ++ * from OCF to the LAC API. ++ * ++ */ ++static int icp_ocfDrvModExp(struct cryptkop *krp) ++{ ++ CpaStatus lacStatus = CPA_STATUS_SUCCESS; ++ void *callbackTag = NULL; ++ CpaCyLnModExpOpData *pModExpOpData = NULL; ++ CpaFlatBuffer *pResult = NULL; ++ ++ if ((krp->krp_param[ICP_MOD_EXP_KRP_PARAM_MODULUS_INDEX].crp_nbits % ++ NUM_BITS_IN_BYTE) != 0) { ++ DPRINTK("%s(): Warning - modulus buffer size (%d) is not a " ++ "multiple of 8 bits\n", __FUNCTION__, ++ krp->krp_param[ICP_MOD_EXP_KRP_PARAM_MODULUS_INDEX]. ++ crp_nbits); ++ } ++ ++ /* Result storage space should be the same size as the prime as this ++ value can take up the same amount of storage space */ ++ if (krp->krp_param[ICP_MOD_EXP_KRP_PARAM_MODULUS_INDEX].crp_nbits > ++ krp->krp_param[ICP_MOD_EXP_KRP_PARAM_RESULT_INDEX].crp_nbits) { ++ APRINTK("%s(): Return Buffer size must be the same or" ++ " greater than the Modulus buffer\n", __FUNCTION__); ++ krp->krp_status = EINVAL; ++ return EINVAL; ++ } ++ ++ callbackTag = krp; ++ ++ pModExpOpData = kmem_cache_zalloc(drvLnModExp_zone, GFP_KERNEL); ++ if (NULL == pModExpOpData) { ++ APRINTK("%s():Failed to get memory for key gen data\n", ++ __FUNCTION__); ++ krp->krp_status = ENOMEM; ++ return ENOMEM; ++ } ++ ++ pResult = kmem_cache_zalloc(drvFlatBuffer_zone, GFP_KERNEL); ++ if (NULL == pResult) { ++ APRINTK("%s():Failed to get memory for ModExp result\n", ++ __FUNCTION__); ++ kmem_cache_free(drvLnModExp_zone, pModExpOpData); ++ krp->krp_status = ENOMEM; ++ return ENOMEM; ++ } ++ ++ /* Link parameters */ ++ pModExpOpData->modulus.pData = ++ krp->krp_param[ICP_MOD_EXP_KRP_PARAM_MODULUS_INDEX].crp_p; ++ BITS_TO_BYTES(pModExpOpData->modulus.dataLenInBytes, ++ krp->krp_param[ICP_MOD_EXP_KRP_PARAM_MODULUS_INDEX]. ++ crp_nbits); ++ ++ icp_ocfDrvSwapBytes(pModExpOpData->modulus.pData, ++ pModExpOpData->modulus.dataLenInBytes); ++ ++ /*OCF patch to Openswan Pluto regularly sends the base value as 2 ++ bits in size. In this case, it has been found it is better to ++ use the base size memory space as the input buffer (if the number ++ is in bits is less than a byte, the number of bits is the input ++ value) */ ++ if (krp->krp_param[ICP_MOD_EXP_KRP_PARAM_BASE_INDEX].crp_nbits < ++ NUM_BITS_IN_BYTE) { ++ DPRINTK("%s : base is small (%d)\n", __FUNCTION__, krp-> ++ krp_param[ICP_MOD_EXP_KRP_PARAM_BASE_INDEX].crp_nbits); ++ pModExpOpData->base.dataLenInBytes = SINGLE_BYTE_VALUE; ++ pModExpOpData->base.pData = ++ (uint8_t *) & (krp-> ++ krp_param[ICP_MOD_EXP_KRP_PARAM_BASE_INDEX]. ++ crp_nbits); ++ *((uint32_t *) pModExpOpData->base.pData) = ++ htonl(*((uint32_t *) pModExpOpData->base.pData)); ++ ++ } else { ++ ++ DPRINTK("%s : base is big (%d)\n", __FUNCTION__, krp-> ++ krp_param[ICP_MOD_EXP_KRP_PARAM_BASE_INDEX].crp_nbits); ++ pModExpOpData->base.pData = ++ krp->krp_param[ICP_MOD_EXP_KRP_PARAM_BASE_INDEX].crp_p; ++ BITS_TO_BYTES(pModExpOpData->base.dataLenInBytes, ++ krp->krp_param[ICP_MOD_EXP_KRP_PARAM_BASE_INDEX]. ++ crp_nbits); ++ icp_ocfDrvSwapBytes(pModExpOpData->base.pData, ++ pModExpOpData->base.dataLenInBytes); ++ } ++ ++ pModExpOpData->exponent.pData = ++ krp->krp_param[ICP_MOD_EXP_KRP_PARAM_EXPONENT_INDEX].crp_p; ++ BITS_TO_BYTES(pModExpOpData->exponent.dataLenInBytes, ++ krp->krp_param[ICP_MOD_EXP_KRP_PARAM_EXPONENT_INDEX]. ++ crp_nbits); ++ ++ icp_ocfDrvSwapBytes(pModExpOpData->exponent.pData, ++ pModExpOpData->exponent.dataLenInBytes); ++ /* Output parameters */ ++ pResult->pData = ++ krp->krp_param[ICP_MOD_EXP_KRP_PARAM_RESULT_INDEX].crp_p, ++ BITS_TO_BYTES(pResult->dataLenInBytes, ++ krp->krp_param[ICP_MOD_EXP_KRP_PARAM_RESULT_INDEX]. ++ crp_nbits); ++ ++ lacStatus = cpaCyLnModExp(CPA_INSTANCE_HANDLE_SINGLE, ++ icp_ocfDrvModExpCallBack, ++ callbackTag, pModExpOpData, pResult); ++ ++ if (CPA_STATUS_SUCCESS != lacStatus) { ++ EPRINTK("%s(): Mod Exp Operation failed (%d).\n", ++ __FUNCTION__, lacStatus); ++ krp->krp_status = ECANCELED; ++ icp_ocfDrvFreeFlatBuffer(pResult); ++ kmem_cache_free(drvLnModExp_zone, pModExpOpData); ++ } ++ ++ return lacStatus; ++} ++ ++/* Name : icp_ocfDrvModExpCRT ++ * ++ * Description : This function will map ordinary Modular Exponentiation Chinese ++ * Remainder Theorem implementaion calls from OCF to the LAC API. ++ * ++ * Note : Mod Exp CRT for this driver is accelerated through LAC RSA type 2 ++ * decrypt operation. Therefore P and Q input values must always be prime ++ * numbers. Although basic primality checks are done in LAC, it is up to the ++ * user to do any correct prime number checking before passing the inputs. ++ */ ++ ++static int icp_ocfDrvModExpCRT(struct cryptkop *krp) ++{ ++ CpaStatus lacStatus = CPA_STATUS_SUCCESS; ++ CpaCyRsaDecryptOpData *rsaDecryptOpData = NULL; ++ void *callbackTag = NULL; ++ CpaFlatBuffer *pOutputData = NULL; ++ ++ /*Parameter input checks are all done by LAC, no need to repeat ++ them here. */ ++ callbackTag = krp; ++ ++ rsaDecryptOpData = kmem_cache_zalloc(drvRSADecrypt_zone, GFP_KERNEL); ++ if (NULL == rsaDecryptOpData) { ++ APRINTK("%s():Failed to get memory" ++ " for MOD EXP CRT Op data struct\n", __FUNCTION__); ++ krp->krp_status = ENOMEM; ++ return ENOMEM; ++ } ++ ++ rsaDecryptOpData->pRecipientPrivateKey ++ = kmem_cache_zalloc(drvRSAPrivateKey_zone, GFP_KERNEL); ++ if (NULL == rsaDecryptOpData->pRecipientPrivateKey) { ++ APRINTK("%s():Failed to get memory for MOD EXP CRT" ++ " private key values struct\n", __FUNCTION__); ++ kmem_cache_free(drvRSADecrypt_zone, rsaDecryptOpData); ++ krp->krp_status = ENOMEM; ++ return ENOMEM; ++ } ++ ++ rsaDecryptOpData->pRecipientPrivateKey-> ++ version = CPA_CY_RSA_VERSION_TWO_PRIME; ++ rsaDecryptOpData->pRecipientPrivateKey-> ++ privateKeyRepType = CPA_CY_RSA_PRIVATE_KEY_REP_TYPE_2; ++ ++ pOutputData = kmem_cache_zalloc(drvFlatBuffer_zone, GFP_KERNEL); ++ if (NULL == pOutputData) { ++ APRINTK("%s():Failed to get memory" ++ " for MOD EXP CRT output data\n", __FUNCTION__); ++ kmem_cache_free(drvRSAPrivateKey_zone, ++ rsaDecryptOpData->pRecipientPrivateKey); ++ kmem_cache_free(drvRSADecrypt_zone, rsaDecryptOpData); ++ krp->krp_status = ENOMEM; ++ return ENOMEM; ++ } ++ ++ rsaDecryptOpData->pRecipientPrivateKey-> ++ version = CPA_CY_RSA_VERSION_TWO_PRIME; ++ rsaDecryptOpData->pRecipientPrivateKey-> ++ privateKeyRepType = CPA_CY_RSA_PRIVATE_KEY_REP_TYPE_2; ++ ++ /* Link parameters */ ++ rsaDecryptOpData->inputData.pData = ++ krp->krp_param[ICP_MOD_EXP_CRT_KRP_PARAM_I_INDEX].crp_p; ++ BITS_TO_BYTES(rsaDecryptOpData->inputData.dataLenInBytes, ++ krp->krp_param[ICP_MOD_EXP_CRT_KRP_PARAM_I_INDEX]. ++ crp_nbits); ++ ++ icp_ocfDrvSwapBytes(rsaDecryptOpData->inputData.pData, ++ rsaDecryptOpData->inputData.dataLenInBytes); ++ ++ rsaDecryptOpData->pRecipientPrivateKey->privateKeyRep2.prime1P.pData = ++ krp->krp_param[ICP_MOD_EXP_CRT_KRP_PARAM_PRIME_P_INDEX].crp_p; ++ BITS_TO_BYTES(rsaDecryptOpData->pRecipientPrivateKey->privateKeyRep2. ++ prime1P.dataLenInBytes, ++ krp->krp_param[ICP_MOD_EXP_CRT_KRP_PARAM_PRIME_P_INDEX]. ++ crp_nbits); ++ ++ icp_ocfDrvSwapBytes(rsaDecryptOpData->pRecipientPrivateKey-> ++ privateKeyRep2.prime1P.pData, ++ rsaDecryptOpData->pRecipientPrivateKey-> ++ privateKeyRep2.prime1P.dataLenInBytes); ++ ++ rsaDecryptOpData->pRecipientPrivateKey->privateKeyRep2.prime2Q.pData = ++ krp->krp_param[ICP_MOD_EXP_CRT_KRP_PARAM_PRIME_Q_INDEX].crp_p; ++ BITS_TO_BYTES(rsaDecryptOpData->pRecipientPrivateKey->privateKeyRep2. ++ prime2Q.dataLenInBytes, ++ krp->krp_param[ICP_MOD_EXP_CRT_KRP_PARAM_PRIME_Q_INDEX]. ++ crp_nbits); ++ ++ icp_ocfDrvSwapBytes(rsaDecryptOpData->pRecipientPrivateKey-> ++ privateKeyRep2.prime2Q.pData, ++ rsaDecryptOpData->pRecipientPrivateKey-> ++ privateKeyRep2.prime2Q.dataLenInBytes); ++ ++ rsaDecryptOpData->pRecipientPrivateKey-> ++ privateKeyRep2.exponent1Dp.pData = ++ krp->krp_param[ICP_MOD_EXP_CRT_KRP_PARAM_EXPONENT_DP_INDEX].crp_p; ++ BITS_TO_BYTES(rsaDecryptOpData->pRecipientPrivateKey->privateKeyRep2. ++ exponent1Dp.dataLenInBytes, ++ krp-> ++ krp_param[ICP_MOD_EXP_CRT_KRP_PARAM_EXPONENT_DP_INDEX]. ++ crp_nbits); ++ ++ icp_ocfDrvSwapBytes(rsaDecryptOpData->pRecipientPrivateKey-> ++ privateKeyRep2.exponent1Dp.pData, ++ rsaDecryptOpData->pRecipientPrivateKey-> ++ privateKeyRep2.exponent1Dp.dataLenInBytes); ++ ++ rsaDecryptOpData->pRecipientPrivateKey-> ++ privateKeyRep2.exponent2Dq.pData = ++ krp->krp_param[ICP_MOD_EXP_CRT_KRP_PARAM_EXPONENT_DQ_INDEX].crp_p; ++ BITS_TO_BYTES(rsaDecryptOpData->pRecipientPrivateKey-> ++ privateKeyRep2.exponent2Dq.dataLenInBytes, ++ krp-> ++ krp_param[ICP_MOD_EXP_CRT_KRP_PARAM_EXPONENT_DQ_INDEX]. ++ crp_nbits); ++ ++ icp_ocfDrvSwapBytes(rsaDecryptOpData->pRecipientPrivateKey-> ++ privateKeyRep2.exponent2Dq.pData, ++ rsaDecryptOpData->pRecipientPrivateKey-> ++ privateKeyRep2.exponent2Dq.dataLenInBytes); ++ ++ rsaDecryptOpData->pRecipientPrivateKey-> ++ privateKeyRep2.coefficientQInv.pData = ++ krp->krp_param[ICP_MOD_EXP_CRT_KRP_PARAM_COEFF_QINV_INDEX].crp_p; ++ BITS_TO_BYTES(rsaDecryptOpData->pRecipientPrivateKey-> ++ privateKeyRep2.coefficientQInv.dataLenInBytes, ++ krp-> ++ krp_param[ICP_MOD_EXP_CRT_KRP_PARAM_COEFF_QINV_INDEX]. ++ crp_nbits); ++ ++ icp_ocfDrvSwapBytes(rsaDecryptOpData->pRecipientPrivateKey-> ++ privateKeyRep2.coefficientQInv.pData, ++ rsaDecryptOpData->pRecipientPrivateKey-> ++ privateKeyRep2.coefficientQInv.dataLenInBytes); ++ ++ /* Output Parameter */ ++ pOutputData->pData = ++ krp->krp_param[ICP_MOD_EXP_CRT_KRP_PARAM_RESULT_INDEX].crp_p; ++ BITS_TO_BYTES(pOutputData->dataLenInBytes, ++ krp->krp_param[ICP_MOD_EXP_CRT_KRP_PARAM_RESULT_INDEX]. ++ crp_nbits); ++ ++ lacStatus = cpaCyRsaDecrypt(CPA_INSTANCE_HANDLE_SINGLE, ++ icp_ocfDrvModExpCRTCallBack, ++ callbackTag, rsaDecryptOpData, pOutputData); ++ ++ if (CPA_STATUS_SUCCESS != lacStatus) { ++ EPRINTK("%s(): Mod Exp CRT Operation failed (%d).\n", ++ __FUNCTION__, lacStatus); ++ krp->krp_status = ECANCELED; ++ icp_ocfDrvFreeFlatBuffer(pOutputData); ++ kmem_cache_free(drvRSAPrivateKey_zone, ++ rsaDecryptOpData->pRecipientPrivateKey); ++ kmem_cache_free(drvRSADecrypt_zone, rsaDecryptOpData); ++ } ++ ++ return lacStatus; ++} ++ ++/* Name : icp_ocfDrvCheckALessThanB ++ * ++ * Description : This function will check whether the first argument is less ++ * than the second. It is used to check whether the DSA RS sign Random K ++ * value is less than the Prime Q value (as defined in the specification) ++ * ++ */ ++static int ++icp_ocfDrvCheckALessThanB(CpaFlatBuffer * pK, CpaFlatBuffer * pQ, int *doCheck) ++{ ++ ++ uint8_t *MSB_K = pK->pData; ++ uint8_t *MSB_Q = pQ->pData; ++ uint32_t buffer_lengths_in_bytes = pQ->dataLenInBytes; ++ ++ if (DONT_RUN_LESS_THAN_CHECK == *doCheck) { ++ return FAIL_A_IS_GREATER_THAN_B; ++ } ++ ++/*Check MSBs ++if A == B, check next MSB ++if A > B, return A_IS_GREATER_THAN_B ++if A < B, return A_IS_LESS_THAN_B (success) ++*/ ++ while (*MSB_K == *MSB_Q) { ++ MSB_K++; ++ MSB_Q++; ++ ++ buffer_lengths_in_bytes--; ++ if (0 == buffer_lengths_in_bytes) { ++ DPRINTK("%s() Buffers have equal value!!\n", ++ __FUNCTION__); ++ return FAIL_A_IS_EQUAL_TO_B; ++ } ++ ++ } ++ ++ if (*MSB_K < *MSB_Q) { ++ return SUCCESS_A_IS_LESS_THAN_B; ++ } else { ++ return FAIL_A_IS_GREATER_THAN_B; ++ } ++ ++} ++ ++/* Name : icp_ocfDrvDsaSign ++ * ++ * Description : This function will map DSA RS Sign from OCF to the LAC API. ++ * ++ * NOTE: From looking at OCF patch to OpenSSL and even the number of input ++ * parameters, OCF expects us to generate the random seed value. This value ++ * is generated and passed to LAC, however the number is discared in the ++ * callback and not returned to the user. ++ */ ++static int icp_ocfDrvDsaSign(struct cryptkop *krp) ++{ ++ CpaStatus lacStatus = CPA_STATUS_SUCCESS; ++ CpaCyDsaRSSignOpData *dsaRsSignOpData = NULL; ++ void *callbackTag = NULL; ++ CpaCyRandGenOpData randGenOpData; ++ int primeQSizeInBytes = 0; ++ int doCheck = 0; ++ CpaFlatBuffer randData; ++ CpaBoolean protocolStatus = CPA_FALSE; ++ CpaFlatBuffer *pR = NULL; ++ CpaFlatBuffer *pS = NULL; ++ ++ callbackTag = krp; ++ ++ BITS_TO_BYTES(primeQSizeInBytes, ++ krp->krp_param[ICP_DSA_SIGN_KRP_PARAM_PRIME_Q_INDEX]. ++ crp_nbits); ++ ++ if (DSA_RS_SIGN_PRIMEQ_SIZE_IN_BYTES != primeQSizeInBytes) { ++ APRINTK("%s(): DSA PRIME Q size not equal to the " ++ "FIPS defined 20bytes, = %d\n", ++ __FUNCTION__, primeQSizeInBytes); ++ krp->krp_status = EDOM; ++ return EDOM; ++ } ++ ++ dsaRsSignOpData = kmem_cache_zalloc(drvDSARSSign_zone, GFP_KERNEL); ++ if (NULL == dsaRsSignOpData) { ++ APRINTK("%s():Failed to get memory" ++ " for DSA RS Sign Op data struct\n", __FUNCTION__); ++ krp->krp_status = ENOMEM; ++ return ENOMEM; ++ } ++ ++ dsaRsSignOpData->K.pData = ++ kmem_cache_alloc(drvDSARSSignKValue_zone, GFP_ATOMIC); ++ ++ if (NULL == dsaRsSignOpData->K.pData) { ++ APRINTK("%s():Failed to get memory" ++ " for DSA RS Sign Op Random value\n", __FUNCTION__); ++ kmem_cache_free(drvDSARSSign_zone, dsaRsSignOpData); ++ krp->krp_status = ENOMEM; ++ return ENOMEM; ++ } ++ ++ pR = kmem_cache_zalloc(drvFlatBuffer_zone, GFP_KERNEL); ++ if (NULL == pR) { ++ APRINTK("%s():Failed to get memory" ++ " for DSA signature R\n", __FUNCTION__); ++ kmem_cache_free(drvDSARSSignKValue_zone, ++ dsaRsSignOpData->K.pData); ++ kmem_cache_free(drvDSARSSign_zone, dsaRsSignOpData); ++ krp->krp_status = ENOMEM; ++ return ENOMEM; ++ } ++ ++ pS = kmem_cache_zalloc(drvFlatBuffer_zone, GFP_KERNEL); ++ if (NULL == pS) { ++ APRINTK("%s():Failed to get memory" ++ " for DSA signature S\n", __FUNCTION__); ++ icp_ocfDrvFreeFlatBuffer(pR); ++ kmem_cache_free(drvDSARSSignKValue_zone, ++ dsaRsSignOpData->K.pData); ++ kmem_cache_free(drvDSARSSign_zone, dsaRsSignOpData); ++ krp->krp_status = ENOMEM; ++ return ENOMEM; ++ } ++ ++ /*link prime number parameter for ease of processing */ ++ dsaRsSignOpData->P.pData = ++ krp->krp_param[ICP_DSA_SIGN_KRP_PARAM_PRIME_P_INDEX].crp_p; ++ BITS_TO_BYTES(dsaRsSignOpData->P.dataLenInBytes, ++ krp->krp_param[ICP_DSA_SIGN_KRP_PARAM_PRIME_P_INDEX]. ++ crp_nbits); ++ ++ icp_ocfDrvSwapBytes(dsaRsSignOpData->P.pData, ++ dsaRsSignOpData->P.dataLenInBytes); ++ ++ dsaRsSignOpData->Q.pData = ++ krp->krp_param[ICP_DSA_SIGN_KRP_PARAM_PRIME_Q_INDEX].crp_p; ++ BITS_TO_BYTES(dsaRsSignOpData->Q.dataLenInBytes, ++ krp->krp_param[ICP_DSA_SIGN_KRP_PARAM_PRIME_Q_INDEX]. ++ crp_nbits); ++ ++ icp_ocfDrvSwapBytes(dsaRsSignOpData->Q.pData, ++ dsaRsSignOpData->Q.dataLenInBytes); ++ ++ /*generate random number with equal buffer size to Prime value Q, ++ but value less than Q */ ++ dsaRsSignOpData->K.dataLenInBytes = dsaRsSignOpData->Q.dataLenInBytes; ++ ++ randGenOpData.generateBits = CPA_TRUE; ++ randGenOpData.lenInBytes = dsaRsSignOpData->K.dataLenInBytes; ++ ++ icp_ocfDrvPtrAndLenToFlatBuffer(dsaRsSignOpData->K.pData, ++ dsaRsSignOpData->K.dataLenInBytes, ++ &randData); ++ ++ doCheck = 0; ++ while (icp_ocfDrvCheckALessThanB(&(dsaRsSignOpData->K), ++ &(dsaRsSignOpData->Q), &doCheck)) { ++ ++ if (CPA_STATUS_SUCCESS ++ != cpaCyRandGen(CPA_INSTANCE_HANDLE_SINGLE, ++ NULL, NULL, &randGenOpData, &randData)) { ++ APRINTK("%s(): ERROR - Failed to generate DSA RS Sign K" ++ "value\n", __FUNCTION__); ++ icp_ocfDrvFreeFlatBuffer(pS); ++ icp_ocfDrvFreeFlatBuffer(pR); ++ kmem_cache_free(drvDSARSSignKValue_zone, ++ dsaRsSignOpData->K.pData); ++ kmem_cache_free(drvDSARSSign_zone, dsaRsSignOpData); ++ krp->krp_status = EAGAIN; ++ return EAGAIN; ++ } ++ ++ doCheck++; ++ if (DSA_SIGN_RAND_GEN_VAL_CHECK_MAX_ITERATIONS == doCheck) { ++ APRINTK("%s(): ERROR - Failed to find DSA RS Sign K " ++ "value less than Q value\n", __FUNCTION__); ++ icp_ocfDrvFreeFlatBuffer(pS); ++ icp_ocfDrvFreeFlatBuffer(pR); ++ kmem_cache_free(drvDSARSSignKValue_zone, ++ dsaRsSignOpData->K.pData); ++ kmem_cache_free(drvDSARSSign_zone, dsaRsSignOpData); ++ krp->krp_status = EAGAIN; ++ return EAGAIN; ++ } ++ ++ } ++ /*Rand Data - no need to swap bytes for pK */ ++ ++ /* Link parameters */ ++ dsaRsSignOpData->G.pData = ++ krp->krp_param[ICP_DSA_SIGN_KRP_PARAM_G_INDEX].crp_p; ++ BITS_TO_BYTES(dsaRsSignOpData->G.dataLenInBytes, ++ krp->krp_param[ICP_DSA_SIGN_KRP_PARAM_G_INDEX].crp_nbits); ++ ++ icp_ocfDrvSwapBytes(dsaRsSignOpData->G.pData, ++ dsaRsSignOpData->G.dataLenInBytes); ++ ++ dsaRsSignOpData->X.pData = ++ krp->krp_param[ICP_DSA_SIGN_KRP_PARAM_X_INDEX].crp_p; ++ BITS_TO_BYTES(dsaRsSignOpData->X.dataLenInBytes, ++ krp->krp_param[ICP_DSA_SIGN_KRP_PARAM_X_INDEX].crp_nbits); ++ icp_ocfDrvSwapBytes(dsaRsSignOpData->X.pData, ++ dsaRsSignOpData->X.dataLenInBytes); ++ ++ dsaRsSignOpData->M.pData = ++ krp->krp_param[ICP_DSA_SIGN_KRP_PARAM_DGST_INDEX].crp_p; ++ BITS_TO_BYTES(dsaRsSignOpData->M.dataLenInBytes, ++ krp->krp_param[ICP_DSA_SIGN_KRP_PARAM_DGST_INDEX]. ++ crp_nbits); ++ icp_ocfDrvSwapBytes(dsaRsSignOpData->M.pData, ++ dsaRsSignOpData->M.dataLenInBytes); ++ ++ /* Output Parameters */ ++ pS->pData = krp->krp_param[ICP_DSA_SIGN_KRP_PARAM_S_RESULT_INDEX].crp_p; ++ BITS_TO_BYTES(pS->dataLenInBytes, ++ krp->krp_param[ICP_DSA_SIGN_KRP_PARAM_S_RESULT_INDEX]. ++ crp_nbits); ++ ++ pR->pData = krp->krp_param[ICP_DSA_SIGN_KRP_PARAM_R_RESULT_INDEX].crp_p; ++ BITS_TO_BYTES(pR->dataLenInBytes, ++ krp->krp_param[ICP_DSA_SIGN_KRP_PARAM_R_RESULT_INDEX]. ++ crp_nbits); ++ ++ lacStatus = cpaCyDsaSignRS(CPA_INSTANCE_HANDLE_SINGLE, ++ icp_ocfDrvDsaRSSignCallBack, ++ callbackTag, dsaRsSignOpData, ++ &protocolStatus, pR, pS); ++ ++ if (CPA_STATUS_SUCCESS != lacStatus) { ++ EPRINTK("%s(): DSA RS Sign Operation failed (%d).\n", ++ __FUNCTION__, lacStatus); ++ krp->krp_status = ECANCELED; ++ icp_ocfDrvFreeFlatBuffer(pS); ++ icp_ocfDrvFreeFlatBuffer(pR); ++ kmem_cache_free(drvDSARSSignKValue_zone, ++ dsaRsSignOpData->K.pData); ++ kmem_cache_free(drvDSARSSign_zone, dsaRsSignOpData); ++ } ++ ++ return lacStatus; ++} ++ ++/* Name : icp_ocfDrvDsaVerify ++ * ++ * Description : This function will map DSA RS Verify from OCF to the LAC API. ++ * ++ */ ++static int icp_ocfDrvDsaVerify(struct cryptkop *krp) ++{ ++ CpaStatus lacStatus = CPA_STATUS_SUCCESS; ++ CpaCyDsaVerifyOpData *dsaVerifyOpData = NULL; ++ void *callbackTag = NULL; ++ CpaBoolean verifyStatus = CPA_FALSE; ++ ++ callbackTag = krp; ++ ++ dsaVerifyOpData = kmem_cache_zalloc(drvDSAVerify_zone, GFP_KERNEL); ++ if (NULL == dsaVerifyOpData) { ++ APRINTK("%s():Failed to get memory" ++ " for DSA Verify Op data struct\n", __FUNCTION__); ++ krp->krp_status = ENOMEM; ++ return ENOMEM; ++ } ++ ++ /* Link parameters */ ++ dsaVerifyOpData->P.pData = ++ krp->krp_param[ICP_DSA_VERIFY_KRP_PARAM_PRIME_P_INDEX].crp_p; ++ BITS_TO_BYTES(dsaVerifyOpData->P.dataLenInBytes, ++ krp->krp_param[ICP_DSA_VERIFY_KRP_PARAM_PRIME_P_INDEX]. ++ crp_nbits); ++ icp_ocfDrvSwapBytes(dsaVerifyOpData->P.pData, ++ dsaVerifyOpData->P.dataLenInBytes); ++ ++ dsaVerifyOpData->Q.pData = ++ krp->krp_param[ICP_DSA_VERIFY_KRP_PARAM_PRIME_Q_INDEX].crp_p; ++ BITS_TO_BYTES(dsaVerifyOpData->Q.dataLenInBytes, ++ krp->krp_param[ICP_DSA_VERIFY_KRP_PARAM_PRIME_Q_INDEX]. ++ crp_nbits); ++ icp_ocfDrvSwapBytes(dsaVerifyOpData->Q.pData, ++ dsaVerifyOpData->Q.dataLenInBytes); ++ ++ dsaVerifyOpData->G.pData = ++ krp->krp_param[ICP_DSA_VERIFY_KRP_PARAM_G_INDEX].crp_p; ++ BITS_TO_BYTES(dsaVerifyOpData->G.dataLenInBytes, ++ krp->krp_param[ICP_DSA_VERIFY_KRP_PARAM_G_INDEX]. ++ crp_nbits); ++ icp_ocfDrvSwapBytes(dsaVerifyOpData->G.pData, ++ dsaVerifyOpData->G.dataLenInBytes); ++ ++ dsaVerifyOpData->Y.pData = ++ krp->krp_param[ICP_DSA_VERIFY_KRP_PARAM_PUBKEY_INDEX].crp_p; ++ BITS_TO_BYTES(dsaVerifyOpData->Y.dataLenInBytes, ++ krp->krp_param[ICP_DSA_VERIFY_KRP_PARAM_PUBKEY_INDEX]. ++ crp_nbits); ++ icp_ocfDrvSwapBytes(dsaVerifyOpData->Y.pData, ++ dsaVerifyOpData->Y.dataLenInBytes); ++ ++ dsaVerifyOpData->M.pData = ++ krp->krp_param[ICP_DSA_VERIFY_KRP_PARAM_DGST_INDEX].crp_p; ++ BITS_TO_BYTES(dsaVerifyOpData->M.dataLenInBytes, ++ krp->krp_param[ICP_DSA_VERIFY_KRP_PARAM_DGST_INDEX]. ++ crp_nbits); ++ icp_ocfDrvSwapBytes(dsaVerifyOpData->M.pData, ++ dsaVerifyOpData->M.dataLenInBytes); ++ ++ dsaVerifyOpData->R.pData = ++ krp->krp_param[ICP_DSA_VERIFY_KRP_PARAM_SIG_R_INDEX].crp_p; ++ BITS_TO_BYTES(dsaVerifyOpData->R.dataLenInBytes, ++ krp->krp_param[ICP_DSA_VERIFY_KRP_PARAM_SIG_R_INDEX]. ++ crp_nbits); ++ icp_ocfDrvSwapBytes(dsaVerifyOpData->R.pData, ++ dsaVerifyOpData->R.dataLenInBytes); ++ ++ dsaVerifyOpData->S.pData = ++ krp->krp_param[ICP_DSA_VERIFY_KRP_PARAM_SIG_S_INDEX].crp_p; ++ BITS_TO_BYTES(dsaVerifyOpData->S.dataLenInBytes, ++ krp->krp_param[ICP_DSA_VERIFY_KRP_PARAM_SIG_S_INDEX]. ++ crp_nbits); ++ icp_ocfDrvSwapBytes(dsaVerifyOpData->S.pData, ++ dsaVerifyOpData->S.dataLenInBytes); ++ ++ lacStatus = cpaCyDsaVerify(CPA_INSTANCE_HANDLE_SINGLE, ++ icp_ocfDrvDsaVerifyCallBack, ++ callbackTag, dsaVerifyOpData, &verifyStatus); ++ ++ if (CPA_STATUS_SUCCESS != lacStatus) { ++ EPRINTK("%s(): DSA Verify Operation failed (%d).\n", ++ __FUNCTION__, lacStatus); ++ kmem_cache_free(drvDSAVerify_zone, dsaVerifyOpData); ++ krp->krp_status = ECANCELED; ++ } ++ ++ return lacStatus; ++} ++ ++/* Name : icp_ocfDrvReadRandom ++ * ++ * Description : This function will map RNG functionality calls from OCF ++ * to the LAC API. ++ */ ++int icp_ocfDrvReadRandom(void *arg, uint32_t * buf, int maxwords) ++{ ++ CpaStatus lacStatus = CPA_STATUS_SUCCESS; ++ CpaCyRandGenOpData randGenOpData; ++ CpaFlatBuffer randData; ++ ++ if (NULL == buf) { ++ APRINTK("%s(): Invalid input parameters\n", __FUNCTION__); ++ return EINVAL; ++ } ++ ++ /* maxwords here is number of integers to generate data for */ ++ randGenOpData.generateBits = CPA_TRUE; ++ ++ randGenOpData.lenInBytes = maxwords * sizeof(uint32_t); ++ ++ icp_ocfDrvPtrAndLenToFlatBuffer((Cpa8U *) buf, ++ randGenOpData.lenInBytes, &randData); ++ ++ lacStatus = cpaCyRandGen(CPA_INSTANCE_HANDLE_SINGLE, ++ NULL, NULL, &randGenOpData, &randData); ++ if (CPA_STATUS_SUCCESS != lacStatus) { ++ EPRINTK("%s(): icp_LacSymRandGen failed (%d). \n", ++ __FUNCTION__, lacStatus); ++ return RETURN_RAND_NUM_GEN_FAILED; ++ } ++ ++ return randGenOpData.lenInBytes / sizeof(uint32_t); ++} ++ ++/* Name : icp_ocfDrvDhP1Callback ++ * ++ * Description : When this function returns it signifies that the LAC ++ * component has completed the DH operation. ++ */ ++static void ++icp_ocfDrvDhP1CallBack(void *callbackTag, ++ CpaStatus status, ++ void *pOpData, CpaFlatBuffer * pLocalOctetStringPV) ++{ ++ struct cryptkop *krp = NULL; ++ CpaCyDhPhase1KeyGenOpData *pPhase1OpData = NULL; ++ ++ if (NULL == callbackTag) { ++ DPRINTK("%s(): Invalid input parameters - " ++ "callbackTag data is NULL\n", __FUNCTION__); ++ return; ++ } ++ krp = (struct cryptkop *)callbackTag; ++ ++ if (NULL == pOpData) { ++ DPRINTK("%s(): Invalid input parameters - " ++ "Operation Data is NULL\n", __FUNCTION__); ++ krp->krp_status = ECANCELED; ++ crypto_kdone(krp); ++ return; ++ } ++ pPhase1OpData = (CpaCyDhPhase1KeyGenOpData *) pOpData; ++ ++ if (NULL == pLocalOctetStringPV) { ++ DPRINTK("%s(): Invalid input parameters - " ++ "pLocalOctetStringPV Data is NULL\n", __FUNCTION__); ++ memset(pPhase1OpData, 0, sizeof(CpaCyDhPhase1KeyGenOpData)); ++ kmem_cache_free(drvDH_zone, pPhase1OpData); ++ krp->krp_status = ECANCELED; ++ crypto_kdone(krp); ++ return; ++ } ++ ++ if (CPA_STATUS_SUCCESS == status) { ++ krp->krp_status = CRYPTO_OP_SUCCESS; ++ } else { ++ APRINTK("%s(): Diffie Hellman Phase1 Key Gen failed - " ++ "Operation Status = %d\n", __FUNCTION__, status); ++ krp->krp_status = ECANCELED; ++ } ++ ++ icp_ocfDrvSwapBytes(pLocalOctetStringPV->pData, ++ pLocalOctetStringPV->dataLenInBytes); ++ ++ icp_ocfDrvFreeFlatBuffer(pLocalOctetStringPV); ++ memset(pPhase1OpData, 0, sizeof(CpaCyDhPhase1KeyGenOpData)); ++ kmem_cache_free(drvDH_zone, pPhase1OpData); ++ ++ crypto_kdone(krp); ++ ++ return; ++} ++ ++/* Name : icp_ocfDrvModExpCallBack ++ * ++ * Description : When this function returns it signifies that the LAC ++ * component has completed the Mod Exp operation. ++ */ ++static void ++icp_ocfDrvModExpCallBack(void *callbackTag, ++ CpaStatus status, ++ void *pOpdata, CpaFlatBuffer * pResult) ++{ ++ struct cryptkop *krp = NULL; ++ CpaCyLnModExpOpData *pLnModExpOpData = NULL; ++ ++ if (NULL == callbackTag) { ++ DPRINTK("%s(): Invalid input parameters - " ++ "callbackTag data is NULL\n", __FUNCTION__); ++ return; ++ } ++ krp = (struct cryptkop *)callbackTag; ++ ++ if (NULL == pOpdata) { ++ DPRINTK("%s(): Invalid Mod Exp input parameters - " ++ "Operation Data is NULL\n", __FUNCTION__); ++ krp->krp_status = ECANCELED; ++ crypto_kdone(krp); ++ return; ++ } ++ pLnModExpOpData = (CpaCyLnModExpOpData *) pOpdata; ++ ++ if (NULL == pResult) { ++ DPRINTK("%s(): Invalid input parameters - " ++ "pResult data is NULL\n", __FUNCTION__); ++ krp->krp_status = ECANCELED; ++ memset(pLnModExpOpData, 0, sizeof(CpaCyLnModExpOpData)); ++ kmem_cache_free(drvLnModExp_zone, pLnModExpOpData); ++ crypto_kdone(krp); ++ return; ++ } ++ ++ if (CPA_STATUS_SUCCESS == status) { ++ krp->krp_status = CRYPTO_OP_SUCCESS; ++ } else { ++ APRINTK("%s(): LAC Mod Exp Operation failed - " ++ "Operation Status = %d\n", __FUNCTION__, status); ++ krp->krp_status = ECANCELED; ++ } ++ ++ icp_ocfDrvSwapBytes(pResult->pData, pResult->dataLenInBytes); ++ ++ /*switch base size value back to original */ ++ if (pLnModExpOpData->base.pData == ++ (uint8_t *) & (krp-> ++ krp_param[ICP_MOD_EXP_KRP_PARAM_BASE_INDEX]. ++ crp_nbits)) { ++ *((uint32_t *) pLnModExpOpData->base.pData) = ++ ntohl(*((uint32_t *) pLnModExpOpData->base.pData)); ++ } ++ icp_ocfDrvFreeFlatBuffer(pResult); ++ memset(pLnModExpOpData, 0, sizeof(CpaCyLnModExpOpData)); ++ kmem_cache_free(drvLnModExp_zone, pLnModExpOpData); ++ ++ crypto_kdone(krp); ++ ++ return; ++ ++} ++ ++/* Name : icp_ocfDrvModExpCRTCallBack ++ * ++ * Description : When this function returns it signifies that the LAC ++ * component has completed the Mod Exp CRT operation. ++ */ ++static void ++icp_ocfDrvModExpCRTCallBack(void *callbackTag, ++ CpaStatus status, ++ void *pOpData, CpaFlatBuffer * pOutputData) ++{ ++ struct cryptkop *krp = NULL; ++ CpaCyRsaDecryptOpData *pDecryptData = NULL; ++ ++ if (NULL == callbackTag) { ++ DPRINTK("%s(): Invalid input parameters - " ++ "callbackTag data is NULL\n", __FUNCTION__); ++ return; ++ } ++ ++ krp = (struct cryptkop *)callbackTag; ++ ++ if (NULL == pOpData) { ++ DPRINTK("%s(): Invalid input parameters - " ++ "Operation Data is NULL\n", __FUNCTION__); ++ krp->krp_status = ECANCELED; ++ crypto_kdone(krp); ++ return; ++ } ++ pDecryptData = (CpaCyRsaDecryptOpData *) pOpData; ++ ++ if (NULL == pOutputData) { ++ DPRINTK("%s(): Invalid input parameter - " ++ "pOutputData is NULL\n", __FUNCTION__); ++ memset(pDecryptData->pRecipientPrivateKey, 0, ++ sizeof(CpaCyRsaPrivateKey)); ++ kmem_cache_free(drvRSAPrivateKey_zone, ++ pDecryptData->pRecipientPrivateKey); ++ memset(pDecryptData, 0, sizeof(CpaCyRsaDecryptOpData)); ++ kmem_cache_free(drvRSADecrypt_zone, pDecryptData); ++ krp->krp_status = ECANCELED; ++ crypto_kdone(krp); ++ return; ++ } ++ ++ if (CPA_STATUS_SUCCESS == status) { ++ krp->krp_status = CRYPTO_OP_SUCCESS; ++ } else { ++ APRINTK("%s(): LAC Mod Exp CRT operation failed - " ++ "Operation Status = %d\n", __FUNCTION__, status); ++ krp->krp_status = ECANCELED; ++ } ++ ++ icp_ocfDrvSwapBytes(pOutputData->pData, pOutputData->dataLenInBytes); ++ ++ icp_ocfDrvFreeFlatBuffer(pOutputData); ++ memset(pDecryptData->pRecipientPrivateKey, 0, ++ sizeof(CpaCyRsaPrivateKey)); ++ kmem_cache_free(drvRSAPrivateKey_zone, ++ pDecryptData->pRecipientPrivateKey); ++ memset(pDecryptData, 0, sizeof(CpaCyRsaDecryptOpData)); ++ kmem_cache_free(drvRSADecrypt_zone, pDecryptData); ++ ++ crypto_kdone(krp); ++ ++ return; ++} ++ ++/* Name : icp_ocfDrvDsaRSSignCallBack ++ * ++ * Description : When this function returns it signifies that the LAC ++ * component has completed the DSA RS sign operation. ++ */ ++static void ++icp_ocfDrvDsaRSSignCallBack(void *callbackTag, ++ CpaStatus status, ++ void *pOpData, ++ CpaBoolean protocolStatus, ++ CpaFlatBuffer * pR, CpaFlatBuffer * pS) ++{ ++ struct cryptkop *krp = NULL; ++ CpaCyDsaRSSignOpData *pSignData = NULL; ++ ++ if (NULL == callbackTag) { ++ DPRINTK("%s(): Invalid input parameters - " ++ "callbackTag data is NULL\n", __FUNCTION__); ++ return; ++ } ++ ++ krp = (struct cryptkop *)callbackTag; ++ ++ if (NULL == pOpData) { ++ DPRINTK("%s(): Invalid input parameters - " ++ "Operation Data is NULL\n", __FUNCTION__); ++ krp->krp_status = ECANCELED; ++ crypto_kdone(krp); ++ return; ++ } ++ pSignData = (CpaCyDsaRSSignOpData *) pOpData; ++ ++ if (NULL == pR) { ++ DPRINTK("%s(): Invalid input parameter - " ++ "pR sign is NULL\n", __FUNCTION__); ++ icp_ocfDrvFreeFlatBuffer(pS); ++ kmem_cache_free(drvDSARSSign_zone, pSignData); ++ krp->krp_status = ECANCELED; ++ crypto_kdone(krp); ++ return; ++ } ++ ++ if (NULL == pS) { ++ DPRINTK("%s(): Invalid input parameter - " ++ "pS sign is NULL\n", __FUNCTION__); ++ icp_ocfDrvFreeFlatBuffer(pR); ++ kmem_cache_free(drvDSARSSign_zone, pSignData); ++ krp->krp_status = ECANCELED; ++ crypto_kdone(krp); ++ return; ++ } ++ ++ if (CPA_STATUS_SUCCESS != status) { ++ APRINTK("%s(): LAC DSA RS Sign operation failed - " ++ "Operation Status = %d\n", __FUNCTION__, status); ++ krp->krp_status = ECANCELED; ++ } else { ++ krp->krp_status = CRYPTO_OP_SUCCESS; ++ ++ if (CPA_TRUE != protocolStatus) { ++ DPRINTK("%s(): LAC DSA RS Sign operation failed due " ++ "to protocol error\n", __FUNCTION__); ++ krp->krp_status = EIO; ++ } ++ } ++ ++ /* Swap bytes only when the callback status is successful and ++ protocolStatus is set to true */ ++ if (CPA_STATUS_SUCCESS == status && CPA_TRUE == protocolStatus) { ++ icp_ocfDrvSwapBytes(pR->pData, pR->dataLenInBytes); ++ icp_ocfDrvSwapBytes(pS->pData, pS->dataLenInBytes); ++ } ++ ++ icp_ocfDrvFreeFlatBuffer(pR); ++ icp_ocfDrvFreeFlatBuffer(pS); ++ memset(pSignData->K.pData, 0, pSignData->K.dataLenInBytes); ++ kmem_cache_free(drvDSARSSignKValue_zone, pSignData->K.pData); ++ memset(pSignData, 0, sizeof(CpaCyDsaRSSignOpData)); ++ kmem_cache_free(drvDSARSSign_zone, pSignData); ++ crypto_kdone(krp); ++ ++ return; ++} ++ ++/* Name : icp_ocfDrvDsaVerifyCallback ++ * ++ * Description : When this function returns it signifies that the LAC ++ * component has completed the DSA Verify operation. ++ */ ++static void ++icp_ocfDrvDsaVerifyCallBack(void *callbackTag, ++ CpaStatus status, ++ void *pOpData, CpaBoolean verifyStatus) ++{ ++ ++ struct cryptkop *krp = NULL; ++ CpaCyDsaVerifyOpData *pVerData = NULL; ++ ++ if (NULL == callbackTag) { ++ DPRINTK("%s(): Invalid input parameters - " ++ "callbackTag data is NULL\n", __FUNCTION__); ++ return; ++ } ++ ++ krp = (struct cryptkop *)callbackTag; ++ ++ if (NULL == pOpData) { ++ DPRINTK("%s(): Invalid input parameters - " ++ "Operation Data is NULL\n", __FUNCTION__); ++ krp->krp_status = ECANCELED; ++ crypto_kdone(krp); ++ return; ++ } ++ pVerData = (CpaCyDsaVerifyOpData *) pOpData; ++ ++ if (CPA_STATUS_SUCCESS != status) { ++ APRINTK("%s(): LAC DSA Verify operation failed - " ++ "Operation Status = %d\n", __FUNCTION__, status); ++ krp->krp_status = ECANCELED; ++ } else { ++ krp->krp_status = CRYPTO_OP_SUCCESS; ++ ++ if (CPA_TRUE != verifyStatus) { ++ DPRINTK("%s(): DSA signature invalid\n", __FUNCTION__); ++ krp->krp_status = EIO; ++ } ++ } ++ ++ /* Swap bytes only when the callback status is successful and ++ verifyStatus is set to true */ ++ /*Just swapping back the key values for now. Possibly all ++ swapped buffers need to be reverted */ ++ if (CPA_STATUS_SUCCESS == status && CPA_TRUE == verifyStatus) { ++ icp_ocfDrvSwapBytes(pVerData->R.pData, ++ pVerData->R.dataLenInBytes); ++ icp_ocfDrvSwapBytes(pVerData->S.pData, ++ pVerData->S.dataLenInBytes); ++ } ++ ++ memset(pVerData, 0, sizeof(CpaCyDsaVerifyOpData)); ++ kmem_cache_free(drvDSAVerify_zone, pVerData); ++ crypto_kdone(krp); ++ ++ return; ++} +diff -Nur linux-2.6.30.orig/crypto/ocf/ep80579/icp_common.c linux-2.6.30/crypto/ocf/ep80579/icp_common.c +--- linux-2.6.30.orig/crypto/ocf/ep80579/icp_common.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/ep80579/icp_common.c 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,891 @@ ++/*************************************************************************** ++ * ++ * This file is provided under a dual BSD/GPLv2 license. When using or ++ * redistributing this file, you may do so under either license. ++ * ++ * GPL LICENSE SUMMARY ++ * ++ * Copyright(c) 2007,2008 Intel Corporation. All rights reserved. ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of version 2 of the GNU General Public License as ++ * published by the Free Software Foundation. ++ * ++ * This program is distributed in the hope that it will be useful, but ++ * WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ * General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. ++ * The full GNU General Public License is included in this distribution ++ * in the file called LICENSE.GPL. ++ * ++ * Contact Information: ++ * Intel Corporation ++ * ++ * BSD LICENSE ++ * ++ * Copyright(c) 2007,2008 Intel Corporation. All rights reserved. ++ * All rights reserved. ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions ++ * are met: ++ * ++ * * Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * * Redistributions in binary form must reproduce the above copyright ++ * notice, this list of conditions and the following disclaimer in ++ * the documentation and/or other materials provided with the ++ * distribution. ++ * * Neither the name of Intel Corporation nor the names of its ++ * contributors may be used to endorse or promote products derived ++ * from this software without specific prior written permission. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS ++ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT ++ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR ++ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT ++ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, ++ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT ++ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, ++ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY ++ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT ++ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE ++ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. ++ * ++ * ++ * version: Security.L.1.0.130 ++ * ++ ***************************************************************************/ ++ ++/* ++ * An OCF module that uses Intel® QuickAssist Integrated Accelerator to do the ++ * crypto. ++ * ++ * This driver requires the ICP Access Library that is available from Intel in ++ * order to operate. ++ */ ++ ++#include "icp_ocf.h" ++ ++#define ICP_OCF_COMP_NAME "ICP_OCF" ++#define ICP_OCF_VER_MAIN (2) ++#define ICP_OCF_VER_MJR (0) ++#define ICP_OCF_VER_MNR (0) ++ ++#define MAX_DEREG_RETRIES (100) ++#define DEFAULT_DEREG_RETRIES (10) ++#define DEFAULT_DEREG_DELAY_IN_JIFFIES (10) ++ ++/* This defines the maximum number of sessions possible between OCF ++ and the OCF Tolapai Driver. If set to zero, there is no limit. */ ++#define DEFAULT_OCF_TO_DRV_MAX_SESSION_COUNT (0) ++#define NUM_SUPPORTED_CAPABILITIES (21) ++ ++/*Slabs zones*/ ++struct kmem_cache *drvSessionData_zone = NULL; ++struct kmem_cache *drvOpData_zone = NULL; ++struct kmem_cache *drvDH_zone = NULL; ++struct kmem_cache *drvLnModExp_zone = NULL; ++struct kmem_cache *drvRSADecrypt_zone = NULL; ++struct kmem_cache *drvRSAPrivateKey_zone = NULL; ++struct kmem_cache *drvDSARSSign_zone = NULL; ++struct kmem_cache *drvDSARSSignKValue_zone = NULL; ++struct kmem_cache *drvDSAVerify_zone = NULL; ++ ++/*Slab zones for flatbuffers and bufferlist*/ ++struct kmem_cache *drvFlatBuffer_zone = NULL; ++ ++static int icp_ocfDrvInit(void); ++static void icp_ocfDrvExit(void); ++static void icp_ocfDrvFreeCaches(void); ++static void icp_ocfDrvDeferedFreeLacSessionProcess(void *arg); ++ ++int32_t icp_ocfDrvDriverId = INVALID_DRIVER_ID; ++ ++/* Module parameter - gives the number of times LAC deregistration shall be ++ re-tried */ ++int num_dereg_retries = DEFAULT_DEREG_RETRIES; ++ ++/* Module parameter - gives the delay time in jiffies before a LAC session ++ shall be attempted to be deregistered again */ ++int dereg_retry_delay_in_jiffies = DEFAULT_DEREG_DELAY_IN_JIFFIES; ++ ++/* Module parameter - gives the maximum number of sessions possible between ++ OCF and the OCF Tolapai Driver. If set to zero, there is no limit.*/ ++int max_sessions = DEFAULT_OCF_TO_DRV_MAX_SESSION_COUNT; ++ ++/* This is set when the module is removed from the system, no further ++ processing can take place if this is set */ ++atomic_t icp_ocfDrvIsExiting = ATOMIC_INIT(0); ++ ++/* This is used to show how many lac sessions were not deregistered*/ ++atomic_t lac_session_failed_dereg_count = ATOMIC_INIT(0); ++ ++/* This is used to track the number of registered sessions between OCF and ++ * and the OCF Tolapai driver, when max_session is set to value other than ++ * zero. This ensures that the max_session set for the OCF and the driver ++ * is equal to the LAC registered sessions */ ++atomic_t num_ocf_to_drv_registered_sessions = ATOMIC_INIT(0); ++ ++/* Head of linked list used to store session data */ ++struct list_head icp_ocfDrvGlobalSymListHead; ++struct list_head icp_ocfDrvGlobalSymListHead_FreeMemList; ++ ++spinlock_t icp_ocfDrvSymSessInfoListSpinlock = SPIN_LOCK_UNLOCKED; ++rwlock_t icp_kmem_cache_destroy_alloc_lock = RW_LOCK_UNLOCKED; ++ ++struct workqueue_struct *icp_ocfDrvFreeLacSessionWorkQ; ++ ++struct icp_drvBuffListInfo defBuffListInfo; ++ ++static struct { ++ softc_device_decl sc_dev; ++} icpDev; ++ ++static device_method_t icp_methods = { ++ /* crypto device methods */ ++ DEVMETHOD(cryptodev_newsession, icp_ocfDrvNewSession), ++ DEVMETHOD(cryptodev_freesession, icp_ocfDrvFreeLACSession), ++ DEVMETHOD(cryptodev_process, icp_ocfDrvSymProcess), ++ DEVMETHOD(cryptodev_kprocess, icp_ocfDrvPkeProcess), ++}; ++ ++module_param(num_dereg_retries, int, S_IRUGO); ++module_param(dereg_retry_delay_in_jiffies, int, S_IRUGO); ++module_param(max_sessions, int, S_IRUGO); ++ ++MODULE_PARM_DESC(num_dereg_retries, ++ "Number of times to retry LAC Sym Session Deregistration. " ++ "Default 10, Max 100"); ++MODULE_PARM_DESC(dereg_retry_delay_in_jiffies, "Delay in jiffies " ++ "(added to a schedule() function call) before a LAC Sym " ++ "Session Dereg is retried. Default 10"); ++MODULE_PARM_DESC(max_sessions, "This sets the maximum number of sessions " ++ "between OCF and this driver. If this value is set to zero, " ++ "max session count checking is disabled. Default is zero(0)"); ++ ++/* Name : icp_ocfDrvInit ++ * ++ * Description : This function will register all the symmetric and asymmetric ++ * functionality that will be accelerated by the hardware. It will also ++ * get a unique driver ID from the OCF and initialise all slab caches ++ */ ++static int __init icp_ocfDrvInit(void) ++{ ++ int ocfStatus = 0; ++ ++ IPRINTK("=== %s ver %d.%d.%d ===\n", ICP_OCF_COMP_NAME, ++ ICP_OCF_VER_MAIN, ICP_OCF_VER_MJR, ICP_OCF_VER_MNR); ++ ++ if (MAX_DEREG_RETRIES < num_dereg_retries) { ++ EPRINTK("Session deregistration retry count set to greater " ++ "than %d", MAX_DEREG_RETRIES); ++ return -1; ++ } ++ ++ /* Initialize and Start the Cryptographic component */ ++ if (CPA_STATUS_SUCCESS != ++ cpaCyStartInstance(CPA_INSTANCE_HANDLE_SINGLE)) { ++ EPRINTK("Failed to initialize and start the instance " ++ "of the Cryptographic component.\n"); ++ return -1; ++ } ++ ++ /* Set the default size of BufferList to allocate */ ++ memset(&defBuffListInfo, 0, sizeof(struct icp_drvBuffListInfo)); ++ if (ICP_OCF_DRV_STATUS_SUCCESS != ++ icp_ocfDrvBufferListMemInfo(ICP_OCF_DRV_DEFAULT_BUFFLIST_ARRAYS, ++ &defBuffListInfo)) { ++ EPRINTK("Failed to get bufferlist memory info.\n"); ++ return -1; ++ } ++ ++ /*Register OCF Tolapai Driver with OCF */ ++ memset(&icpDev, 0, sizeof(icpDev)); ++ softc_device_init(&icpDev, "icp", 0, icp_methods); ++ ++ icp_ocfDrvDriverId = crypto_get_driverid(softc_get_device(&icpDev), ++ CRYPTOCAP_F_HARDWARE); ++ ++ if (icp_ocfDrvDriverId < 0) { ++ EPRINTK("%s : ICP driver failed to register with OCF!\n", ++ __FUNCTION__); ++ return -ENODEV; ++ } ++ ++ /*Create all the slab caches used by the OCF Tolapai Driver */ ++ drvSessionData_zone = ++ ICP_CACHE_CREATE("ICP Session Data", struct icp_drvSessionData); ++ ICP_CACHE_NULL_CHECK(drvSessionData_zone); ++ ++ /* ++ * Allocation of the OpData includes the allocation space for meta data. ++ * The memory after the opData structure is reserved for this meta data. ++ */ ++ drvOpData_zone = ++ kmem_cache_create("ICP Op Data", sizeof(struct icp_drvOpData) + ++ defBuffListInfo.metaSize ,0, SLAB_HWCACHE_ALIGN, NULL, NULL); ++ ++ ++ ICP_CACHE_NULL_CHECK(drvOpData_zone); ++ ++ drvDH_zone = ICP_CACHE_CREATE("ICP DH data", CpaCyDhPhase1KeyGenOpData); ++ ICP_CACHE_NULL_CHECK(drvDH_zone); ++ ++ drvLnModExp_zone = ++ ICP_CACHE_CREATE("ICP ModExp data", CpaCyLnModExpOpData); ++ ICP_CACHE_NULL_CHECK(drvLnModExp_zone); ++ ++ drvRSADecrypt_zone = ++ ICP_CACHE_CREATE("ICP RSA decrypt data", CpaCyRsaDecryptOpData); ++ ICP_CACHE_NULL_CHECK(drvRSADecrypt_zone); ++ ++ drvRSAPrivateKey_zone = ++ ICP_CACHE_CREATE("ICP RSA private key data", CpaCyRsaPrivateKey); ++ ICP_CACHE_NULL_CHECK(drvRSAPrivateKey_zone); ++ ++ drvDSARSSign_zone = ++ ICP_CACHE_CREATE("ICP DSA Sign", CpaCyDsaRSSignOpData); ++ ICP_CACHE_NULL_CHECK(drvDSARSSign_zone); ++ ++ /*too awkward to use a macro here */ ++ drvDSARSSignKValue_zone = ++ kmem_cache_create("ICP DSA Sign Rand Val", ++ DSA_RS_SIGN_PRIMEQ_SIZE_IN_BYTES, 0, ++ SLAB_HWCACHE_ALIGN, NULL, NULL); ++ ICP_CACHE_NULL_CHECK(drvDSARSSignKValue_zone); ++ ++ drvDSAVerify_zone = ++ ICP_CACHE_CREATE("ICP DSA Verify", CpaCyDsaVerifyOpData); ++ ICP_CACHE_NULL_CHECK(drvDSAVerify_zone); ++ ++ drvFlatBuffer_zone = ++ ICP_CACHE_CREATE("ICP Flat Buffers", CpaFlatBuffer); ++ ICP_CACHE_NULL_CHECK(drvFlatBuffer_zone); ++ ++ /* Register the ICP symmetric crypto support. */ ++ ICP_REGISTER_SYM_FUNCTIONALITY_WITH_OCF(CRYPTO_NULL_CBC); ++ ICP_REGISTER_SYM_FUNCTIONALITY_WITH_OCF(CRYPTO_DES_CBC); ++ ICP_REGISTER_SYM_FUNCTIONALITY_WITH_OCF(CRYPTO_3DES_CBC); ++ ICP_REGISTER_SYM_FUNCTIONALITY_WITH_OCF(CRYPTO_AES_CBC); ++ ICP_REGISTER_SYM_FUNCTIONALITY_WITH_OCF(CRYPTO_ARC4); ++ ICP_REGISTER_SYM_FUNCTIONALITY_WITH_OCF(CRYPTO_MD5); ++ ICP_REGISTER_SYM_FUNCTIONALITY_WITH_OCF(CRYPTO_MD5_HMAC); ++ ICP_REGISTER_SYM_FUNCTIONALITY_WITH_OCF(CRYPTO_SHA1); ++ ICP_REGISTER_SYM_FUNCTIONALITY_WITH_OCF(CRYPTO_SHA1_HMAC); ++ ICP_REGISTER_SYM_FUNCTIONALITY_WITH_OCF(CRYPTO_SHA2_256); ++ ICP_REGISTER_SYM_FUNCTIONALITY_WITH_OCF(CRYPTO_SHA2_256_HMAC); ++ ICP_REGISTER_SYM_FUNCTIONALITY_WITH_OCF(CRYPTO_SHA2_384); ++ ICP_REGISTER_SYM_FUNCTIONALITY_WITH_OCF(CRYPTO_SHA2_384_HMAC); ++ ICP_REGISTER_SYM_FUNCTIONALITY_WITH_OCF(CRYPTO_SHA2_512); ++ ICP_REGISTER_SYM_FUNCTIONALITY_WITH_OCF(CRYPTO_SHA2_512_HMAC); ++ ++ /* Register the ICP asymmetric algorithm support */ ++ ICP_REGISTER_ASYM_FUNCTIONALITY_WITH_OCF(CRK_DH_COMPUTE_KEY); ++ ICP_REGISTER_ASYM_FUNCTIONALITY_WITH_OCF(CRK_MOD_EXP); ++ ICP_REGISTER_ASYM_FUNCTIONALITY_WITH_OCF(CRK_MOD_EXP_CRT); ++ ICP_REGISTER_ASYM_FUNCTIONALITY_WITH_OCF(CRK_DSA_SIGN); ++ ICP_REGISTER_ASYM_FUNCTIONALITY_WITH_OCF(CRK_DSA_VERIFY); ++ ++ /* Register the ICP random number generator support */ ++ if (OCF_REGISTRATION_STATUS_SUCCESS == ++ crypto_rregister(icp_ocfDrvDriverId, icp_ocfDrvReadRandom, NULL)) { ++ ocfStatus++; ++ } ++ ++ if (OCF_ZERO_FUNCTIONALITY_REGISTERED == ocfStatus) { ++ DPRINTK("%s: Failed to register any device capabilities\n", ++ __FUNCTION__); ++ icp_ocfDrvFreeCaches(); ++ icp_ocfDrvDriverId = INVALID_DRIVER_ID; ++ return -ECANCELED; ++ } ++ ++ DPRINTK("%s: Registered %d of %d device capabilities\n", ++ __FUNCTION__, ocfStatus, NUM_SUPPORTED_CAPABILITIES); ++ ++/*Session data linked list used during module exit*/ ++ INIT_LIST_HEAD(&icp_ocfDrvGlobalSymListHead); ++ INIT_LIST_HEAD(&icp_ocfDrvGlobalSymListHead_FreeMemList); ++ ++ icp_ocfDrvFreeLacSessionWorkQ = ++ create_singlethread_workqueue("ocfLacDeregWorkQueue"); ++ ++ return 0; ++} ++ ++/* Name : icp_ocfDrvExit ++ * ++ * Description : This function will deregister all the symmetric sessions ++ * registered with the LAC component. It will also deregister all symmetric ++ * and asymmetric functionality that can be accelerated by the hardware via OCF ++ * and random number generation if it is enabled. ++ */ ++static void icp_ocfDrvExit(void) ++{ ++ CpaStatus lacStatus = CPA_STATUS_SUCCESS; ++ struct icp_drvSessionData *sessionData = NULL; ++ struct icp_drvSessionData *tempSessionData = NULL; ++ int i, remaining_delay_time_in_jiffies = 0; ++ /* There is a possibility of a process or new session command being */ ++ /* sent before this variable is incremented. The aim of this variable */ ++ /* is to stop a loop of calls creating a deadlock situation which */ ++ /* would prevent the driver from exiting. */ ++ ++ atomic_inc(&icp_ocfDrvIsExiting); ++ ++ /*Existing sessions will be routed to another driver after these calls */ ++ crypto_unregister_all(icp_ocfDrvDriverId); ++ crypto_runregister_all(icp_ocfDrvDriverId); ++ ++ /*If any sessions are waiting to be deregistered, do that. This also ++ flushes the work queue */ ++ destroy_workqueue(icp_ocfDrvFreeLacSessionWorkQ); ++ ++ /*ENTER CRITICAL SECTION */ ++ spin_lock_bh(&icp_ocfDrvSymSessInfoListSpinlock); ++ list_for_each_entry_safe(tempSessionData, sessionData, ++ &icp_ocfDrvGlobalSymListHead, listNode) { ++ for (i = 0; i < num_dereg_retries; i++) { ++ /*No harm if bad input - LAC will handle error cases */ ++ if (ICP_SESSION_RUNNING == tempSessionData->inUse) { ++ lacStatus = ++ cpaCySymRemoveSession ++ (CPA_INSTANCE_HANDLE_SINGLE, ++ tempSessionData->sessHandle); ++ if (CPA_STATUS_SUCCESS == lacStatus) { ++ /* Succesfully deregistered */ ++ break; ++ } else if (CPA_STATUS_RETRY != lacStatus) { ++ atomic_inc ++ (&lac_session_failed_dereg_count); ++ break; ++ } ++ ++ /*schedule_timout returns the time left for completion if ++ * this task is set to TASK_INTERRUPTIBLE */ ++ remaining_delay_time_in_jiffies = ++ dereg_retry_delay_in_jiffies; ++ while (0 > remaining_delay_time_in_jiffies) { ++ remaining_delay_time_in_jiffies = ++ schedule_timeout ++ (remaining_delay_time_in_jiffies); ++ } ++ ++ DPRINTK ++ ("%s(): Retry %d to deregistrate the session\n", ++ __FUNCTION__, i); ++ } ++ } ++ ++ /*remove from current list */ ++ list_del(&(tempSessionData->listNode)); ++ /*add to free mem linked list */ ++ list_add(&(tempSessionData->listNode), ++ &icp_ocfDrvGlobalSymListHead_FreeMemList); ++ ++ } ++ ++ /*EXIT CRITICAL SECTION */ ++ spin_unlock_bh(&icp_ocfDrvSymSessInfoListSpinlock); ++ ++ /*set back to initial values */ ++ sessionData = NULL; ++ /*still have a reference in our list! */ ++ tempSessionData = NULL; ++ /*free memory */ ++ list_for_each_entry_safe(tempSessionData, sessionData, ++ &icp_ocfDrvGlobalSymListHead_FreeMemList, ++ listNode) { ++ ++ list_del(&(tempSessionData->listNode)); ++ /* Free allocated CpaCySymSessionCtx */ ++ if (NULL != tempSessionData->sessHandle) { ++ kfree(tempSessionData->sessHandle); ++ } ++ memset(tempSessionData, 0, sizeof(struct icp_drvSessionData)); ++ kmem_cache_free(drvSessionData_zone, tempSessionData); ++ } ++ ++ if (0 != atomic_read(&lac_session_failed_dereg_count)) { ++ DPRINTK("%s(): %d LAC sessions were not deregistered " ++ "correctly. This is not a clean exit! \n", ++ __FUNCTION__, ++ atomic_read(&lac_session_failed_dereg_count)); ++ } ++ ++ icp_ocfDrvFreeCaches(); ++ icp_ocfDrvDriverId = INVALID_DRIVER_ID; ++ ++ /* Shutdown the Cryptographic component */ ++ lacStatus = cpaCyStopInstance(CPA_INSTANCE_HANDLE_SINGLE); ++ if (CPA_STATUS_SUCCESS != lacStatus) { ++ DPRINTK("%s(): Failed to stop instance of the " ++ "Cryptographic component.(status == %d)\n", ++ __FUNCTION__, lacStatus); ++ } ++ ++} ++ ++/* Name : icp_ocfDrvFreeCaches ++ * ++ * Description : This function deregisters all slab caches ++ */ ++static void icp_ocfDrvFreeCaches(void) ++{ ++ if (atomic_read(&icp_ocfDrvIsExiting) != CPA_TRUE) { ++ atomic_set(&icp_ocfDrvIsExiting, 1); ++ } ++ ++ /*Sym Zones */ ++ ICP_CACHE_DESTROY(drvSessionData_zone); ++ ICP_CACHE_DESTROY(drvOpData_zone); ++ ++ /*Asym zones */ ++ ICP_CACHE_DESTROY(drvDH_zone); ++ ICP_CACHE_DESTROY(drvLnModExp_zone); ++ ICP_CACHE_DESTROY(drvRSADecrypt_zone); ++ ICP_CACHE_DESTROY(drvRSAPrivateKey_zone); ++ ICP_CACHE_DESTROY(drvDSARSSignKValue_zone); ++ ICP_CACHE_DESTROY(drvDSARSSign_zone); ++ ICP_CACHE_DESTROY(drvDSAVerify_zone); ++ ++ /*FlatBuffer and BufferList Zones */ ++ ICP_CACHE_DESTROY(drvFlatBuffer_zone); ++ ++} ++ ++/* Name : icp_ocfDrvDeregRetry ++ * ++ * Description : This function will try to farm the session deregistration ++ * off to a work queue. If it fails, nothing more can be done and it ++ * returns an error ++ */ ++ ++int icp_ocfDrvDeregRetry(CpaCySymSessionCtx sessionToDeregister) ++{ ++ struct icp_ocfDrvFreeLacSession *workstore = NULL; ++ ++ DPRINTK("%s(): Retry - Deregistering session (%p)\n", ++ __FUNCTION__, sessionToDeregister); ++ ++ /*make sure the session is not available to be allocated during this ++ process */ ++ atomic_inc(&lac_session_failed_dereg_count); ++ ++ /*Farm off to work queue */ ++ workstore = ++ kmalloc(sizeof(struct icp_ocfDrvFreeLacSession), GFP_ATOMIC); ++ if (NULL == workstore) { ++ DPRINTK("%s(): unable to free session - no memory available " ++ "for work queue\n", __FUNCTION__); ++ return ENOMEM; ++ } ++ ++ workstore->sessionToDeregister = sessionToDeregister; ++ ++ INIT_WORK(&(workstore->work), icp_ocfDrvDeferedFreeLacSessionProcess, ++ workstore); ++ queue_work(icp_ocfDrvFreeLacSessionWorkQ, &(workstore->work)); ++ ++ return ICP_OCF_DRV_STATUS_SUCCESS; ++ ++} ++ ++/* Name : icp_ocfDrvDeferedFreeLacSessionProcess ++ * ++ * Description : This function will retry (module input parameter) ++ * 'num_dereg_retries' times to deregister any symmetric session that recieves a ++ * CPA_STATUS_RETRY message from the LAC component. This function is run in ++ * Thread context because it is called from a worker thread ++ */ ++static void icp_ocfDrvDeferedFreeLacSessionProcess(void *arg) ++{ ++ struct icp_ocfDrvFreeLacSession *workstore = NULL; ++ CpaCySymSessionCtx sessionToDeregister = NULL; ++ int i = 0; ++ int remaining_delay_time_in_jiffies = 0; ++ CpaStatus lacStatus = CPA_STATUS_SUCCESS; ++ ++ workstore = (struct icp_ocfDrvFreeLacSession *)arg; ++ if (NULL == workstore) { ++ DPRINTK("%s() function called with null parameter \n", ++ __FUNCTION__); ++ return; ++ } ++ ++ sessionToDeregister = workstore->sessionToDeregister; ++ kfree(workstore); ++ ++ /*if exiting, give deregistration one more blast only */ ++ if (atomic_read(&icp_ocfDrvIsExiting) == CPA_TRUE) { ++ lacStatus = cpaCySymRemoveSession(CPA_INSTANCE_HANDLE_SINGLE, ++ sessionToDeregister); ++ ++ if (lacStatus != CPA_STATUS_SUCCESS) { ++ DPRINTK("%s() Failed to Dereg LAC session %p " ++ "during module exit\n", __FUNCTION__, ++ sessionToDeregister); ++ return; ++ } ++ ++ atomic_dec(&lac_session_failed_dereg_count); ++ return; ++ } ++ ++ for (i = 0; i <= num_dereg_retries; i++) { ++ lacStatus = cpaCySymRemoveSession(CPA_INSTANCE_HANDLE_SINGLE, ++ sessionToDeregister); ++ ++ if (lacStatus == CPA_STATUS_SUCCESS) { ++ atomic_dec(&lac_session_failed_dereg_count); ++ return; ++ } ++ if (lacStatus != CPA_STATUS_RETRY) { ++ DPRINTK("%s() Failed to deregister session - lacStatus " ++ " = %d", __FUNCTION__, lacStatus); ++ break; ++ } ++ ++ /*schedule_timout returns the time left for completion if this ++ task is set to TASK_INTERRUPTIBLE */ ++ remaining_delay_time_in_jiffies = dereg_retry_delay_in_jiffies; ++ while (0 > remaining_delay_time_in_jiffies) { ++ remaining_delay_time_in_jiffies = ++ schedule_timeout(remaining_delay_time_in_jiffies); ++ } ++ ++ } ++ ++ DPRINTK("%s(): Unable to deregister session\n", __FUNCTION__); ++ DPRINTK("%s(): Number of unavailable LAC sessions = %d\n", __FUNCTION__, ++ atomic_read(&lac_session_failed_dereg_count)); ++} ++ ++/* Name : icp_ocfDrvPtrAndLenToFlatBuffer ++ * ++ * Description : This function converts a "pointer and length" buffer ++ * structure to Fredericksburg Flat Buffer (CpaFlatBuffer) format. ++ * ++ * This function assumes that the data passed in are valid. ++ */ ++inline void ++icp_ocfDrvPtrAndLenToFlatBuffer(void *pData, uint32_t len, ++ CpaFlatBuffer * pFlatBuffer) ++{ ++ pFlatBuffer->pData = pData; ++ pFlatBuffer->dataLenInBytes = len; ++} ++ ++/* Name : icp_ocfDrvSingleSkBuffToFlatBuffer ++ * ++ * Description : This function converts a single socket buffer (sk_buff) ++ * structure to a Fredericksburg Flat Buffer (CpaFlatBuffer) format. ++ * ++ * This function assumes that the data passed in are valid. ++ */ ++static inline void ++icp_ocfDrvSingleSkBuffToFlatBuffer(struct sk_buff *pSkb, ++ CpaFlatBuffer * pFlatBuffer) ++{ ++ pFlatBuffer->pData = pSkb->data; ++ pFlatBuffer->dataLenInBytes = skb_headlen(pSkb); ++} ++ ++/* Name : icp_ocfDrvSkBuffToBufferList ++ * ++ * Description : This function converts a socket buffer (sk_buff) structure to ++ * Fredericksburg Scatter/Gather (CpaBufferList) buffer format. ++ * ++ * This function assumes that the bufferlist has been allocated with the correct ++ * number of buffer arrays. ++ * ++ */ ++inline int ++icp_ocfDrvSkBuffToBufferList(struct sk_buff *pSkb, CpaBufferList * bufferList) ++{ ++ CpaFlatBuffer *curFlatBuffer = NULL; ++ char *skbuffPageAddr = NULL; ++ struct sk_buff *pCurFrag = NULL; ++ struct skb_shared_info *pShInfo = NULL; ++ uint32_t page_offset = 0, i = 0; ++ ++ DPRINTK("%s(): Entry Point\n", __FUNCTION__); ++ ++ /* ++ * In all cases, the first skb needs to be translated to FlatBuffer. ++ * Perform a buffer translation for the first skbuff ++ */ ++ curFlatBuffer = bufferList->pBuffers; ++ icp_ocfDrvSingleSkBuffToFlatBuffer(pSkb, curFlatBuffer); ++ ++ /* Set the userData to point to the original sk_buff */ ++ bufferList->pUserData = (void *)pSkb; ++ ++ /* We now know we'll have at least one element in the SGL */ ++ bufferList->numBuffers = 1; ++ ++ if (0 == skb_is_nonlinear(pSkb)) { ++ /* Is a linear buffer - therefore it's a single skbuff */ ++ DPRINTK("%s(): Exit Point\n", __FUNCTION__); ++ return ICP_OCF_DRV_STATUS_SUCCESS; ++ } ++ ++ curFlatBuffer++; ++ pShInfo = skb_shinfo(pSkb); ++ if (pShInfo->frag_list != NULL && pShInfo->nr_frags != 0) { ++ EPRINTK("%s():" ++ "Translation for a combination of frag_list " ++ "and frags[] array not supported!\n", __FUNCTION__); ++ return ICP_OCF_DRV_STATUS_FAIL; ++ } else if (pShInfo->frag_list != NULL) { ++ /* ++ * Non linear skbuff supported through frag_list ++ * Perform translation for each fragment (sk_buff) ++ * in the frag_list of the first sk_buff. ++ */ ++ for (pCurFrag = pShInfo->frag_list; ++ pCurFrag != NULL; pCurFrag = pCurFrag->next) { ++ icp_ocfDrvSingleSkBuffToFlatBuffer(pCurFrag, ++ curFlatBuffer); ++ curFlatBuffer++; ++ bufferList->numBuffers++; ++ } ++ } else if (pShInfo->nr_frags != 0) { ++ /* ++ * Perform translation for each fragment in frags array ++ * and add to the BufferList ++ */ ++ for (i = 0; i < pShInfo->nr_frags; i++) { ++ /* Get the page address and offset of this frag */ ++ skbuffPageAddr = (char *)pShInfo->frags[i].page; ++ page_offset = pShInfo->frags[i].page_offset; ++ ++ /* Convert a pointer and length to a flat buffer */ ++ icp_ocfDrvPtrAndLenToFlatBuffer(skbuffPageAddr + ++ page_offset, ++ pShInfo->frags[i].size, ++ curFlatBuffer); ++ curFlatBuffer++; ++ bufferList->numBuffers++; ++ } ++ } else { ++ EPRINTK("%s():" "Could not recognize skbuff fragments!\n", ++ __FUNCTION__); ++ return ICP_OCF_DRV_STATUS_FAIL; ++ } ++ ++ DPRINTK("%s(): Exit Point\n", __FUNCTION__); ++ return ICP_OCF_DRV_STATUS_SUCCESS; ++} ++ ++/* Name : icp_ocfDrvBufferListToSkBuff ++ * ++ * Description : This function converts a Fredericksburg Scatter/Gather ++ * (CpaBufferList) buffer format to socket buffer structure. ++ */ ++inline int ++icp_ocfDrvBufferListToSkBuff(CpaBufferList * bufferList, struct sk_buff **skb) ++{ ++ DPRINTK("%s(): Entry Point\n", __FUNCTION__); ++ ++ /* Retrieve the orignal skbuff */ ++ *skb = (struct sk_buff *)bufferList->pUserData; ++ if (NULL == *skb) { ++ EPRINTK("%s():" ++ "Error on converting from a BufferList. " ++ "The BufferList does not contain an sk_buff.\n", ++ __FUNCTION__); ++ return ICP_OCF_DRV_STATUS_FAIL; ++ } ++ DPRINTK("%s(): Exit Point\n", __FUNCTION__); ++ return ICP_OCF_DRV_STATUS_SUCCESS; ++} ++ ++/* Name : icp_ocfDrvPtrAndLenToBufferList ++ * ++ * Description : This function converts a "pointer and length" buffer ++ * structure to Fredericksburg Scatter/Gather Buffer (CpaBufferList) format. ++ * ++ * This function assumes that the data passed in are valid. ++ */ ++inline void ++icp_ocfDrvPtrAndLenToBufferList(void *pDataIn, uint32_t length, ++ CpaBufferList * pBufferList) ++{ ++ pBufferList->numBuffers = 1; ++ pBufferList->pBuffers->pData = pDataIn; ++ pBufferList->pBuffers->dataLenInBytes = length; ++} ++ ++/* Name : icp_ocfDrvBufferListToPtrAndLen ++ * ++ * Description : This function converts Fredericksburg Scatter/Gather Buffer ++ * (CpaBufferList) format to a "pointer and length" buffer structure. ++ * ++ * This function assumes that the data passed in are valid. ++ */ ++inline void ++icp_ocfDrvBufferListToPtrAndLen(CpaBufferList * pBufferList, ++ void **ppDataOut, uint32_t * pLength) ++{ ++ *ppDataOut = pBufferList->pBuffers->pData; ++ *pLength = pBufferList->pBuffers->dataLenInBytes; ++} ++ ++/* Name : icp_ocfDrvBufferListMemInfo ++ * ++ * Description : This function will set the number of flat buffers in ++ * bufferlist, the size of memory to allocate for the pPrivateMetaData ++ * member of the CpaBufferList. ++ */ ++int ++icp_ocfDrvBufferListMemInfo(uint16_t numBuffers, ++ struct icp_drvBuffListInfo *buffListInfo) ++{ ++ buffListInfo->numBuffers = numBuffers; ++ ++ if (CPA_STATUS_SUCCESS != ++ cpaCyBufferListGetMetaSize(CPA_INSTANCE_HANDLE_SINGLE, ++ buffListInfo->numBuffers, ++ &(buffListInfo->metaSize))) { ++ EPRINTK("%s() Failed to get buffer list meta size.\n", ++ __FUNCTION__); ++ return ICP_OCF_DRV_STATUS_FAIL; ++ } ++ ++ return ICP_OCF_DRV_STATUS_SUCCESS; ++} ++ ++/* Name : icp_ocfDrvGetSkBuffFrags ++ * ++ * Description : This function will determine the number of ++ * fragments in a socket buffer(sk_buff). ++ */ ++inline uint16_t icp_ocfDrvGetSkBuffFrags(struct sk_buff * pSkb) ++{ ++ uint16_t numFrags = 0; ++ struct sk_buff *pCurFrag = NULL; ++ struct skb_shared_info *pShInfo = NULL; ++ ++ if (NULL == pSkb) ++ return 0; ++ ++ numFrags = 1; ++ if (0 == skb_is_nonlinear(pSkb)) { ++ /* Linear buffer - it's a single skbuff */ ++ return numFrags; ++ } ++ ++ pShInfo = skb_shinfo(pSkb); ++ if (NULL != pShInfo->frag_list && 0 != pShInfo->nr_frags) { ++ EPRINTK("%s(): Combination of frag_list " ++ "and frags[] array not supported!\n", __FUNCTION__); ++ return 0; ++ } else if (0 != pShInfo->nr_frags) { ++ numFrags += pShInfo->nr_frags; ++ return numFrags; ++ } else if (NULL != pShInfo->frag_list) { ++ for (pCurFrag = pShInfo->frag_list; ++ pCurFrag != NULL; pCurFrag = pCurFrag->next) { ++ numFrags++; ++ } ++ return numFrags; ++ } else { ++ return 0; ++ } ++} ++ ++/* Name : icp_ocfDrvFreeFlatBuffer ++ * ++ * Description : This function will deallocate flat buffer. ++ */ ++inline void icp_ocfDrvFreeFlatBuffer(CpaFlatBuffer * pFlatBuffer) ++{ ++ if (pFlatBuffer != NULL) { ++ memset(pFlatBuffer, 0, sizeof(CpaFlatBuffer)); ++ kmem_cache_free(drvFlatBuffer_zone, pFlatBuffer); ++ } ++} ++ ++/* Name : icp_ocfDrvAllocMetaData ++ * ++ * Description : This function will allocate memory for the ++ * pPrivateMetaData member of CpaBufferList. ++ */ ++inline int ++icp_ocfDrvAllocMetaData(CpaBufferList * pBufferList, ++ const struct icp_drvOpData *pOpData) ++{ ++ Cpa32U metaSize = 0; ++ ++ if (pBufferList->numBuffers <= ICP_OCF_DRV_DEFAULT_BUFFLIST_ARRAYS){ ++ void *pOpDataStartAddr = (void *)pOpData; ++ ++ if (0 == defBuffListInfo.metaSize) { ++ pBufferList->pPrivateMetaData = NULL; ++ return ICP_OCF_DRV_STATUS_SUCCESS; ++ } ++ /* ++ * The meta data allocation has been included as part of the ++ * op data. It has been pre-allocated in memory just after the ++ * icp_drvOpData structure. ++ */ ++ pBufferList->pPrivateMetaData = pOpDataStartAddr + ++ sizeof(struct icp_drvOpData); ++ } else { ++ if (CPA_STATUS_SUCCESS != ++ cpaCyBufferListGetMetaSize(CPA_INSTANCE_HANDLE_SINGLE, ++ pBufferList->numBuffers, ++ &metaSize)) { ++ EPRINTK("%s() Failed to get buffer list meta size.\n", ++ __FUNCTION__); ++ return ICP_OCF_DRV_STATUS_FAIL; ++ } ++ ++ if (0 == metaSize) { ++ pBufferList->pPrivateMetaData = NULL; ++ return ICP_OCF_DRV_STATUS_SUCCESS; ++ } ++ ++ pBufferList->pPrivateMetaData = kmalloc(metaSize, GFP_ATOMIC); ++ } ++ if (NULL == pBufferList->pPrivateMetaData) { ++ EPRINTK("%s() Failed to allocate pPrivateMetaData.\n", ++ __FUNCTION__); ++ return ICP_OCF_DRV_STATUS_FAIL; ++ } ++ ++ return ICP_OCF_DRV_STATUS_SUCCESS; ++} ++ ++/* Name : icp_ocfDrvFreeMetaData ++ * ++ * Description : This function will deallocate pPrivateMetaData memory. ++ */ ++inline void icp_ocfDrvFreeMetaData(CpaBufferList * pBufferList) ++{ ++ if (NULL == pBufferList->pPrivateMetaData) { ++ return; ++ } ++ ++ /* ++ * Only free the meta data if the BufferList has more than ++ * ICP_OCF_DRV_DEFAULT_BUFFLIST_ARRAYS number of buffers. ++ * Otherwise, the meta data shall be freed when the icp_drvOpData is ++ * freed. ++ */ ++ if (ICP_OCF_DRV_DEFAULT_BUFFLIST_ARRAYS < pBufferList->numBuffers){ ++ kfree(pBufferList->pPrivateMetaData); ++ } ++} ++ ++module_init(icp_ocfDrvInit); ++module_exit(icp_ocfDrvExit); ++MODULE_LICENSE("Dual BSD/GPL"); ++MODULE_AUTHOR("Intel"); ++MODULE_DESCRIPTION("OCF Driver for Intel Quick Assist crypto acceleration"); +diff -Nur linux-2.6.30.orig/crypto/ocf/ep80579/icp_ocf.h linux-2.6.30/crypto/ocf/ep80579/icp_ocf.h +--- linux-2.6.30.orig/crypto/ocf/ep80579/icp_ocf.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/ep80579/icp_ocf.h 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,363 @@ ++/*************************************************************************** ++ * ++ * This file is provided under a dual BSD/GPLv2 license. When using or ++ * redistributing this file, you may do so under either license. ++ * ++ * GPL LICENSE SUMMARY ++ * ++ * Copyright(c) 2007,2008 Intel Corporation. All rights reserved. ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of version 2 of the GNU General Public License as ++ * published by the Free Software Foundation. ++ * ++ * This program is distributed in the hope that it will be useful, but ++ * WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ * General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. ++ * The full GNU General Public License is included in this distribution ++ * in the file called LICENSE.GPL. ++ * ++ * Contact Information: ++ * Intel Corporation ++ * ++ * BSD LICENSE ++ * ++ * Copyright(c) 2007,2008 Intel Corporation. All rights reserved. ++ * All rights reserved. ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions ++ * are met: ++ * ++ * * Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * * Redistributions in binary form must reproduce the above copyright ++ * notice, this list of conditions and the following disclaimer in ++ * the documentation and/or other materials provided with the ++ * distribution. ++ * * Neither the name of Intel Corporation nor the names of its ++ * contributors may be used to endorse or promote products derived ++ * from this software without specific prior written permission. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS ++ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT ++ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR ++ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT ++ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, ++ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT ++ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, ++ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY ++ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT ++ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE ++ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. ++ * ++ * ++ * version: Security.L.1.0.130 ++ * ++ ***************************************************************************/ ++ ++/* ++ * OCF drv driver header file for the Intel ICP processor. ++ */ ++ ++#ifndef ICP_OCF_H ++#define ICP_OCF_H ++ ++#include <linux/crypto.h> ++#include <linux/delay.h> ++#include <linux/skbuff.h> ++ ++#include "cryptodev.h" ++#include "uio.h" ++ ++#include "cpa.h" ++#include "cpa_cy_im.h" ++#include "cpa_cy_sym.h" ++#include "cpa_cy_rand.h" ++#include "cpa_cy_dh.h" ++#include "cpa_cy_rsa.h" ++#include "cpa_cy_ln.h" ++#include "cpa_cy_common.h" ++#include "cpa_cy_dsa.h" ++ ++#define NUM_BITS_IN_BYTE (8) ++#define NUM_BITS_IN_BYTE_MINUS_ONE (NUM_BITS_IN_BYTE -1) ++#define INVALID_DRIVER_ID (-1) ++#define RETURN_RAND_NUM_GEN_FAILED (-1) ++ ++/*This is define means only one operation can be chained to another ++(resulting in one chain of two operations)*/ ++#define MAX_NUM_OF_CHAINED_OPS (1) ++/*This is the max block cipher initialisation vector*/ ++#define MAX_IV_LEN_IN_BYTES (20) ++/*This is used to check whether the OCF to this driver session limit has ++ been disabled*/ ++#define NO_OCF_TO_DRV_MAX_SESSIONS (0) ++ ++/*OCF values mapped here*/ ++#define ICP_SHA1_DIGEST_SIZE_IN_BYTES (SHA1_HASH_LEN) ++#define ICP_SHA256_DIGEST_SIZE_IN_BYTES (SHA2_256_HASH_LEN) ++#define ICP_SHA384_DIGEST_SIZE_IN_BYTES (SHA2_384_HASH_LEN) ++#define ICP_SHA512_DIGEST_SIZE_IN_BYTES (SHA2_512_HASH_LEN) ++#define ICP_MD5_DIGEST_SIZE_IN_BYTES (MD5_HASH_LEN) ++#define ARC4_COUNTER_LEN (ARC4_BLOCK_LEN) ++ ++#define OCF_REGISTRATION_STATUS_SUCCESS (0) ++#define OCF_ZERO_FUNCTIONALITY_REGISTERED (0) ++#define ICP_OCF_DRV_NO_CRYPTO_PROCESS_ERROR (0) ++#define ICP_OCF_DRV_STATUS_SUCCESS (0) ++#define ICP_OCF_DRV_STATUS_FAIL (1) ++ ++/*Turn on/off debug options*/ ++#define ICP_OCF_PRINT_DEBUG_MESSAGES (0) ++#define ICP_OCF_PRINT_KERN_ALERT (1) ++#define ICP_OCF_PRINT_KERN_ERRS (1) ++ ++/*DSA Prime Q size in bytes (as defined in the standard) */ ++#define DSA_RS_SIGN_PRIMEQ_SIZE_IN_BYTES (20) ++ ++/*MACRO DEFINITIONS*/ ++ ++#define BITS_TO_BYTES(bytes, bits) \ ++ bytes = (bits + NUM_BITS_IN_BYTE_MINUS_ONE) / NUM_BITS_IN_BYTE ++ ++#define ICP_CACHE_CREATE(cache_ID, cache_name) \ ++ kmem_cache_create(cache_ID, sizeof(cache_name),0, \ ++ SLAB_HWCACHE_ALIGN, NULL, NULL); ++ ++#define ICP_CACHE_NULL_CHECK(slab_zone) \ ++{ \ ++ if(NULL == slab_zone){ \ ++ icp_ocfDrvFreeCaches(); \ ++ EPRINTK("%s() line %d: Not enough memory!\n", \ ++ __FUNCTION__, __LINE__); \ ++ return ENOMEM; \ ++ } \ ++} ++ ++#define ICP_CACHE_DESTROY(slab_zone) \ ++{ \ ++ if(NULL != slab_zone){ \ ++ kmem_cache_destroy(slab_zone); \ ++ slab_zone = NULL; \ ++ } \ ++} ++ ++#define ICP_REGISTER_SYM_FUNCTIONALITY_WITH_OCF(alg) \ ++{ \ ++ if(OCF_REGISTRATION_STATUS_SUCCESS == \ ++ crypto_register(icp_ocfDrvDriverId, \ ++ alg, \ ++ 0, \ ++ 0)) { \ ++ ocfStatus++; \ ++ } \ ++} ++ ++#define ICP_REGISTER_ASYM_FUNCTIONALITY_WITH_OCF(alg) \ ++{ \ ++ if(OCF_REGISTRATION_STATUS_SUCCESS == \ ++ crypto_kregister(icp_ocfDrvDriverId, \ ++ alg, \ ++ 0)){ \ ++ ocfStatus++; \ ++ } \ ++} ++ ++#if ICP_OCF_PRINT_DEBUG_MESSAGES == 1 ++#define DPRINTK(args...) \ ++{ \ ++ printk(args); \ ++} ++ ++#else //ICP_OCF_PRINT_DEBUG_MESSAGES == 1 ++ ++#define DPRINTK(args...) ++ ++#endif //ICP_OCF_PRINT_DEBUG_MESSAGES == 1 ++ ++#if ICP_OCF_PRINT_KERN_ALERT == 1 ++#define APRINTK(args...) \ ++{ \ ++ printk(KERN_ALERT args); \ ++} ++ ++#else //ICP_OCF_PRINT_KERN_ALERT == 1 ++ ++#define APRINTK(args...) ++ ++#endif //ICP_OCF_PRINT_KERN_ALERT == 1 ++ ++#if ICP_OCF_PRINT_KERN_ERRS == 1 ++#define EPRINTK(args...) \ ++{ \ ++ printk(KERN_ERR args); \ ++} ++ ++#else //ICP_OCF_PRINT_KERN_ERRS == 1 ++ ++#define EPRINTK(args...) ++ ++#endif //ICP_OCF_PRINT_KERN_ERRS == 1 ++ ++#define IPRINTK(args...) \ ++{ \ ++ printk(KERN_INFO args); \ ++} ++ ++/*END OF MACRO DEFINITIONS*/ ++ ++typedef enum { ++ ICP_OCF_DRV_ALG_CIPHER = 0, ++ ICP_OCF_DRV_ALG_HASH ++} icp_ocf_drv_alg_type_t; ++ ++/* These are all defined in icp_common.c */ ++extern atomic_t lac_session_failed_dereg_count; ++extern atomic_t icp_ocfDrvIsExiting; ++extern atomic_t num_ocf_to_drv_registered_sessions; ++ ++/*These are use inputs used in icp_sym.c and icp_common.c ++ They are instantiated in icp_common.c*/ ++extern int max_sessions; ++ ++extern int32_t icp_ocfDrvDriverId; ++extern struct list_head icp_ocfDrvGlobalSymListHead; ++extern struct list_head icp_ocfDrvGlobalSymListHead_FreeMemList; ++extern struct workqueue_struct *icp_ocfDrvFreeLacSessionWorkQ; ++extern spinlock_t icp_ocfDrvSymSessInfoListSpinlock; ++extern rwlock_t icp_kmem_cache_destroy_alloc_lock; ++ ++/*Slab zones for symettric functionality, instantiated in icp_common.c*/ ++extern struct kmem_cache *drvSessionData_zone; ++extern struct kmem_cache *drvOpData_zone; ++ ++/*Slabs zones for asymettric functionality, instantiated in icp_common.c*/ ++extern struct kmem_cache *drvDH_zone; ++extern struct kmem_cache *drvLnModExp_zone; ++extern struct kmem_cache *drvRSADecrypt_zone; ++extern struct kmem_cache *drvRSAPrivateKey_zone; ++extern struct kmem_cache *drvDSARSSign_zone; ++extern struct kmem_cache *drvDSARSSignKValue_zone; ++extern struct kmem_cache *drvDSAVerify_zone; ++ ++/*Slab zones for flatbuffers and bufferlist*/ ++extern struct kmem_cache *drvFlatBuffer_zone; ++ ++#define ICP_OCF_DRV_DEFAULT_BUFFLIST_ARRAYS (16) ++ ++struct icp_drvBuffListInfo { ++ Cpa16U numBuffers; ++ Cpa32U metaSize; ++ Cpa32U metaOffset; ++ Cpa32U buffListSize; ++}; ++extern struct icp_drvBuffListInfo defBuffListInfo; ++ ++/* ++* This struct is used to keep a reference to the relevant node in the list ++* of sessionData structs, to the buffer type required by OCF and to the OCF ++* provided crp struct that needs to be returned. All this info is needed in ++* the callback function. ++* ++* IV can sometimes be stored in non-contiguous memory (e.g. skbuff ++* linked/frag list, therefore a contiguous memory space for the IV data must be ++* created and passed to LAC ++* ++*/ ++struct icp_drvOpData { ++ CpaCySymOpData lacOpData; ++ uint32_t digestSizeInBytes; ++ struct cryptop *crp; ++ uint8_t bufferType; ++ uint8_t ivData[MAX_IV_LEN_IN_BYTES]; ++ uint16_t numBufferListArray; ++ CpaBufferList srcBuffer; ++ CpaFlatBuffer bufferListArray[ICP_OCF_DRV_DEFAULT_BUFFLIST_ARRAYS]; ++ CpaBoolean verifyResult; ++}; ++/*Values used to derisk chances of performs being called against ++deregistered sessions (for which the slab page has been reclaimed) ++This is not a fix - since page frames are reclaimed from a slab, one cannot ++rely on that memory not being re-used by another app.*/ ++typedef enum { ++ ICP_SESSION_INITIALISED = 0x5C5C5C, ++ ICP_SESSION_RUNNING = 0x005C00, ++ ICP_SESSION_DEREGISTERED = 0xC5C5C5 ++} usage_derisk; ++ ++/* ++This is the OCF<->OCF_DRV session object: ++ ++1.The first member is a listNode. These session objects are added to a linked ++ list in order to make it easier to remove them all at session exit time. ++2.The second member is used to give the session object state and derisk the ++ possibility of OCF batch calls executing against a deregistered session (as ++ described above). ++3.The third member is a LAC<->OCF_DRV session handle (initialised with the first ++ perform request for that session). ++4.The fourth is the LAC session context. All the parameters for this structure ++ are only known when the first perform request for this session occurs. That is ++ why the OCF Tolapai Driver only registers a new LAC session at perform time ++*/ ++struct icp_drvSessionData { ++ struct list_head listNode; ++ usage_derisk inUse; ++ CpaCySymSessionCtx sessHandle; ++ CpaCySymSessionSetupData lacSessCtx; ++}; ++ ++/* This struct is required for deferred session ++ deregistration as a work queue function can ++ only have one argument*/ ++struct icp_ocfDrvFreeLacSession { ++ CpaCySymSessionCtx sessionToDeregister; ++ struct work_struct work; ++}; ++ ++int icp_ocfDrvNewSession(device_t dev, uint32_t * sild, struct cryptoini *cri); ++ ++int icp_ocfDrvFreeLACSession(device_t dev, uint64_t sid); ++ ++int icp_ocfDrvSymProcess(device_t dev, struct cryptop *crp, int hint); ++ ++int icp_ocfDrvPkeProcess(device_t dev, struct cryptkop *krp, int hint); ++ ++int icp_ocfDrvReadRandom(void *arg, uint32_t * buf, int maxwords); ++ ++int icp_ocfDrvDeregRetry(CpaCySymSessionCtx sessionToDeregister); ++ ++int icp_ocfDrvSkBuffToBufferList(struct sk_buff *skb, ++ CpaBufferList * bufferList); ++ ++int icp_ocfDrvBufferListToSkBuff(CpaBufferList * bufferList, ++ struct sk_buff **skb); ++ ++void icp_ocfDrvPtrAndLenToFlatBuffer(void *pData, uint32_t len, ++ CpaFlatBuffer * pFlatBuffer); ++ ++void icp_ocfDrvPtrAndLenToBufferList(void *pDataIn, uint32_t length, ++ CpaBufferList * pBufferList); ++ ++void icp_ocfDrvBufferListToPtrAndLen(CpaBufferList * pBufferList, ++ void **ppDataOut, uint32_t * pLength); ++ ++int icp_ocfDrvBufferListMemInfo(uint16_t numBuffers, ++ struct icp_drvBuffListInfo *buffListInfo); ++ ++uint16_t icp_ocfDrvGetSkBuffFrags(struct sk_buff *pSkb); ++ ++void icp_ocfDrvFreeFlatBuffer(CpaFlatBuffer * pFlatBuffer); ++ ++int icp_ocfDrvAllocMetaData(CpaBufferList * pBufferList, ++ const struct icp_drvOpData *pOpData); ++ ++void icp_ocfDrvFreeMetaData(CpaBufferList * pBufferList); ++ ++#endif ++/* ICP_OCF_H */ +diff -Nur linux-2.6.30.orig/crypto/ocf/ep80579/icp_sym.c linux-2.6.30/crypto/ocf/ep80579/icp_sym.c +--- linux-2.6.30.orig/crypto/ocf/ep80579/icp_sym.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/ep80579/icp_sym.c 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,1382 @@ ++/*************************************************************************** ++ * ++ * This file is provided under a dual BSD/GPLv2 license. When using or ++ * redistributing this file, you may do so under either license. ++ * ++ * GPL LICENSE SUMMARY ++ * ++ * Copyright(c) 2007,2008 Intel Corporation. All rights reserved. ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of version 2 of the GNU General Public License as ++ * published by the Free Software Foundation. ++ * ++ * This program is distributed in the hope that it will be useful, but ++ * WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++ * General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. ++ * The full GNU General Public License is included in this distribution ++ * in the file called LICENSE.GPL. ++ * ++ * Contact Information: ++ * Intel Corporation ++ * ++ * BSD LICENSE ++ * ++ * Copyright(c) 2007,2008 Intel Corporation. All rights reserved. ++ * All rights reserved. ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions ++ * are met: ++ * ++ * * Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * * Redistributions in binary form must reproduce the above copyright ++ * notice, this list of conditions and the following disclaimer in ++ * the documentation and/or other materials provided with the ++ * distribution. ++ * * Neither the name of Intel Corporation nor the names of its ++ * contributors may be used to endorse or promote products derived ++ * from this software without specific prior written permission. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS ++ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT ++ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR ++ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT ++ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, ++ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT ++ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, ++ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY ++ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT ++ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE ++ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. ++ * ++ * ++ * version: Security.L.1.0.130 ++ * ++ ***************************************************************************/ ++/* ++ * An OCF module that uses the API for Intel® QuickAssist Technology to do the ++ * cryptography. ++ * ++ * This driver requires the ICP Access Library that is available from Intel in ++ * order to operate. ++ */ ++ ++#include "icp_ocf.h" ++ ++/*This is the call back function for all symmetric cryptographic processes. ++ Its main functionality is to free driver crypto operation structure and to ++ call back to OCF*/ ++static void ++icp_ocfDrvSymCallBack(void *callbackTag, ++ CpaStatus status, ++ const CpaCySymOp operationType, ++ void *pOpData, ++ CpaBufferList * pDstBuffer, CpaBoolean verifyResult); ++ ++/*This function is used to extract crypto processing information from the OCF ++ inputs, so as that it may be passed onto LAC*/ ++static int ++icp_ocfDrvProcessDataSetup(struct icp_drvOpData *drvOpData, ++ struct cryptodesc *crp_desc); ++ ++/*This function checks whether the crp_desc argument pertains to a digest or a ++ cipher operation*/ ++static int icp_ocfDrvAlgCheck(struct cryptodesc *crp_desc); ++ ++/*This function copies all the passed in session context information and stores ++ it in a LAC context structure*/ ++static int ++icp_ocfDrvAlgorithmSetup(struct cryptoini *cri, ++ CpaCySymSessionSetupData * lacSessCtx); ++ ++/*This top level function is used to find a pointer to where a digest is ++ stored/needs to be inserted. */ ++static uint8_t *icp_ocfDrvDigestPointerFind(struct icp_drvOpData *drvOpData, ++ struct cryptodesc *crp_desc); ++ ++/*This function is called when a digest pointer has to be found within a ++ SKBUFF.*/ ++static inline uint8_t *icp_ocfDrvSkbuffDigestPointerFind(struct icp_drvOpData ++ *drvOpData, ++ int offsetInBytes, ++ uint32_t ++ digestSizeInBytes); ++ ++/*The following two functions are called if the SKBUFF digest pointer is not ++ positioned in the linear portion of the buffer (i.e. it is in a linked SKBUFF ++ or page fragment).*/ ++/*This function takes care of the page fragment case.*/ ++static inline uint8_t *icp_ocfDrvDigestSkbNRFragsCheck(struct sk_buff *skb, ++ struct skb_shared_info ++ *skb_shared, ++ int offsetInBytes, ++ uint32_t ++ digestSizeInBytes); ++ ++/*This function takes care of the linked list case.*/ ++static inline uint8_t *icp_ocfDrvDigestSkbFragListCheck(struct sk_buff *skb, ++ struct skb_shared_info ++ *skb_shared, ++ int offsetInBytes, ++ uint32_t ++ digestSizeInBytes); ++ ++/*This function is used to free an OCF->OCF_DRV session object*/ ++static void icp_ocfDrvFreeOCFSession(struct icp_drvSessionData *sessionData); ++ ++/*max IOV buffs supported in a UIO structure*/ ++#define NUM_IOV_SUPPORTED (1) ++ ++/* Name : icp_ocfDrvSymCallBack ++ * ++ * Description : When this function returns it signifies that the LAC ++ * component has completed the relevant symmetric operation. ++ * ++ * Notes : The callbackTag is a pointer to an icp_drvOpData. This memory ++ * object was passed to LAC for the cryptographic processing and contains all ++ * the relevant information for cleaning up buffer handles etc. so that the ++ * OCF Tolapai Driver portion of this crypto operation can be fully completed. ++ */ ++static void ++icp_ocfDrvSymCallBack(void *callbackTag, ++ CpaStatus status, ++ const CpaCySymOp operationType, ++ void *pOpData, ++ CpaBufferList * pDstBuffer, CpaBoolean verifyResult) ++{ ++ struct cryptop *crp = NULL; ++ struct icp_drvOpData *temp_drvOpData = ++ (struct icp_drvOpData *)callbackTag; ++ uint64_t *tempBasePtr = NULL; ++ uint32_t tempLen = 0; ++ ++ if (NULL == temp_drvOpData) { ++ DPRINTK("%s(): The callback from the LAC component" ++ " has failed due to Null userOpaque data" ++ "(status == %d).\n", __FUNCTION__, status); ++ DPRINTK("%s(): Unable to call OCF back! \n", __FUNCTION__); ++ return; ++ } ++ ++ crp = temp_drvOpData->crp; ++ crp->crp_etype = ICP_OCF_DRV_NO_CRYPTO_PROCESS_ERROR; ++ ++ if (NULL == pOpData) { ++ DPRINTK("%s(): The callback from the LAC component" ++ " has failed due to Null Symmetric Op data" ++ "(status == %d).\n", __FUNCTION__, status); ++ crp->crp_etype = ECANCELED; ++ crypto_done(crp); ++ return; ++ } ++ ++ if (NULL == pDstBuffer) { ++ DPRINTK("%s(): The callback from the LAC component" ++ " has failed due to Null Dst Bufferlist data" ++ "(status == %d).\n", __FUNCTION__, status); ++ crp->crp_etype = ECANCELED; ++ crypto_done(crp); ++ return; ++ } ++ ++ if (CPA_STATUS_SUCCESS == status) { ++ ++ if (temp_drvOpData->bufferType == CRYPTO_F_SKBUF) { ++ if (ICP_OCF_DRV_STATUS_SUCCESS != ++ icp_ocfDrvBufferListToSkBuff(pDstBuffer, ++ (struct sk_buff **) ++ &(crp->crp_buf))) { ++ EPRINTK("%s(): BufferList to SkBuff " ++ "conversion error.\n", __FUNCTION__); ++ crp->crp_etype = EPERM; ++ } ++ } else { ++ icp_ocfDrvBufferListToPtrAndLen(pDstBuffer, ++ (void **)&tempBasePtr, ++ &tempLen); ++ crp->crp_olen = (int)tempLen; ++ } ++ ++ } else { ++ DPRINTK("%s(): The callback from the LAC component has failed" ++ "(status == %d).\n", __FUNCTION__, status); ++ ++ crp->crp_etype = ECANCELED; ++ } ++ ++ if (temp_drvOpData->numBufferListArray > ++ ICP_OCF_DRV_DEFAULT_BUFFLIST_ARRAYS) { ++ kfree(pDstBuffer->pBuffers); ++ } ++ icp_ocfDrvFreeMetaData(pDstBuffer); ++ kmem_cache_free(drvOpData_zone, temp_drvOpData); ++ ++ /* Invoke the OCF callback function */ ++ crypto_done(crp); ++ ++ return; ++} ++ ++/* Name : icp_ocfDrvNewSession ++ * ++ * Description : This function will create a new Driver<->OCF session ++ * ++ * Notes : LAC session registration happens during the first perform call. ++ * That is the first time we know all information about a given session. ++ */ ++int icp_ocfDrvNewSession(device_t dev, uint32_t * sid, struct cryptoini *cri) ++{ ++ struct icp_drvSessionData *sessionData = NULL; ++ uint32_t delete_session = 0; ++ ++ /* The SID passed in should be our driver ID. We can return the */ ++ /* local ID (LID) which is a unique identifier which we can use */ ++ /* to differentiate between the encrypt/decrypt LAC session handles */ ++ if (NULL == sid) { ++ EPRINTK("%s(): Invalid input parameters - NULL sid.\n", ++ __FUNCTION__); ++ return EINVAL; ++ } ++ ++ if (NULL == cri) { ++ EPRINTK("%s(): Invalid input parameters - NULL cryptoini.\n", ++ __FUNCTION__); ++ return EINVAL; ++ } ++ ++ if (icp_ocfDrvDriverId != *sid) { ++ EPRINTK("%s(): Invalid input parameters - bad driver ID\n", ++ __FUNCTION__); ++ EPRINTK("\t sid = 0x08%p \n \t cri = 0x08%p \n", sid, cri); ++ return EINVAL; ++ } ++ ++ sessionData = kmem_cache_zalloc(drvSessionData_zone, GFP_ATOMIC); ++ if (NULL == sessionData) { ++ DPRINTK("%s():No memory for Session Data\n", __FUNCTION__); ++ return ENOMEM; ++ } ++ ++ /*ENTER CRITICAL SECTION */ ++ spin_lock_bh(&icp_ocfDrvSymSessInfoListSpinlock); ++ /*put this check in the spinlock so no new sessions can be added to the ++ linked list when we are exiting */ ++ if (CPA_TRUE == atomic_read(&icp_ocfDrvIsExiting)) { ++ delete_session++; ++ ++ } else if (NO_OCF_TO_DRV_MAX_SESSIONS != max_sessions) { ++ if (atomic_read(&num_ocf_to_drv_registered_sessions) >= ++ (max_sessions - ++ atomic_read(&lac_session_failed_dereg_count))) { ++ delete_session++; ++ } else { ++ atomic_inc(&num_ocf_to_drv_registered_sessions); ++ /* Add to session data linked list */ ++ list_add(&(sessionData->listNode), ++ &icp_ocfDrvGlobalSymListHead); ++ } ++ ++ } else if (NO_OCF_TO_DRV_MAX_SESSIONS == max_sessions) { ++ list_add(&(sessionData->listNode), ++ &icp_ocfDrvGlobalSymListHead); ++ } ++ ++ sessionData->inUse = ICP_SESSION_INITIALISED; ++ ++ /*EXIT CRITICAL SECTION */ ++ spin_unlock_bh(&icp_ocfDrvSymSessInfoListSpinlock); ++ ++ if (delete_session) { ++ DPRINTK("%s():No Session handles available\n", __FUNCTION__); ++ kmem_cache_free(drvSessionData_zone, sessionData); ++ return EPERM; ++ } ++ ++ if (ICP_OCF_DRV_STATUS_SUCCESS != ++ icp_ocfDrvAlgorithmSetup(cri, &(sessionData->lacSessCtx))) { ++ DPRINTK("%s():algorithm not supported\n", __FUNCTION__); ++ icp_ocfDrvFreeOCFSession(sessionData); ++ return EINVAL; ++ } ++ ++ if (cri->cri_next) { ++ if (cri->cri_next->cri_next != NULL) { ++ DPRINTK("%s():only two chained algorithms supported\n", ++ __FUNCTION__); ++ icp_ocfDrvFreeOCFSession(sessionData); ++ return EPERM; ++ } ++ ++ if (ICP_OCF_DRV_STATUS_SUCCESS != ++ icp_ocfDrvAlgorithmSetup(cri->cri_next, ++ &(sessionData->lacSessCtx))) { ++ DPRINTK("%s():second algorithm not supported\n", ++ __FUNCTION__); ++ icp_ocfDrvFreeOCFSession(sessionData); ++ return EINVAL; ++ } ++ ++ sessionData->lacSessCtx.symOperation = ++ CPA_CY_SYM_OP_ALGORITHM_CHAINING; ++ } ++ ++ *sid = (uint32_t) sessionData; ++ ++ return ICP_OCF_DRV_STATUS_SUCCESS; ++} ++ ++/* Name : icp_ocfDrvAlgorithmSetup ++ * ++ * Description : This function builds the session context data from the ++ * information supplied through OCF. Algorithm chain order and whether the ++ * session is Encrypt/Decrypt can only be found out at perform time however, so ++ * the session is registered with LAC at that time. ++ */ ++static int ++icp_ocfDrvAlgorithmSetup(struct cryptoini *cri, ++ CpaCySymSessionSetupData * lacSessCtx) ++{ ++ ++ lacSessCtx->sessionPriority = CPA_CY_PRIORITY_NORMAL; ++ ++ switch (cri->cri_alg) { ++ ++ case CRYPTO_NULL_CBC: ++ DPRINTK("%s(): NULL CBC\n", __FUNCTION__); ++ lacSessCtx->symOperation = CPA_CY_SYM_OP_CIPHER; ++ lacSessCtx->cipherSetupData.cipherAlgorithm = ++ CPA_CY_SYM_CIPHER_NULL; ++ lacSessCtx->cipherSetupData.cipherKeyLenInBytes = ++ cri->cri_klen / NUM_BITS_IN_BYTE; ++ lacSessCtx->cipherSetupData.pCipherKey = cri->cri_key; ++ break; ++ ++ case CRYPTO_DES_CBC: ++ DPRINTK("%s(): DES CBC\n", __FUNCTION__); ++ lacSessCtx->symOperation = CPA_CY_SYM_OP_CIPHER; ++ lacSessCtx->cipherSetupData.cipherAlgorithm = ++ CPA_CY_SYM_CIPHER_DES_CBC; ++ lacSessCtx->cipherSetupData.cipherKeyLenInBytes = ++ cri->cri_klen / NUM_BITS_IN_BYTE; ++ lacSessCtx->cipherSetupData.pCipherKey = cri->cri_key; ++ break; ++ ++ case CRYPTO_3DES_CBC: ++ DPRINTK("%s(): 3DES CBC\n", __FUNCTION__); ++ lacSessCtx->symOperation = CPA_CY_SYM_OP_CIPHER; ++ lacSessCtx->cipherSetupData.cipherAlgorithm = ++ CPA_CY_SYM_CIPHER_3DES_CBC; ++ lacSessCtx->cipherSetupData.cipherKeyLenInBytes = ++ cri->cri_klen / NUM_BITS_IN_BYTE; ++ lacSessCtx->cipherSetupData.pCipherKey = cri->cri_key; ++ break; ++ ++ case CRYPTO_AES_CBC: ++ DPRINTK("%s(): AES CBC\n", __FUNCTION__); ++ lacSessCtx->symOperation = CPA_CY_SYM_OP_CIPHER; ++ lacSessCtx->cipherSetupData.cipherAlgorithm = ++ CPA_CY_SYM_CIPHER_AES_CBC; ++ lacSessCtx->cipherSetupData.cipherKeyLenInBytes = ++ cri->cri_klen / NUM_BITS_IN_BYTE; ++ lacSessCtx->cipherSetupData.pCipherKey = cri->cri_key; ++ break; ++ ++ case CRYPTO_ARC4: ++ DPRINTK("%s(): ARC4\n", __FUNCTION__); ++ lacSessCtx->symOperation = CPA_CY_SYM_OP_CIPHER; ++ lacSessCtx->cipherSetupData.cipherAlgorithm = ++ CPA_CY_SYM_CIPHER_ARC4; ++ lacSessCtx->cipherSetupData.cipherKeyLenInBytes = ++ cri->cri_klen / NUM_BITS_IN_BYTE; ++ lacSessCtx->cipherSetupData.pCipherKey = cri->cri_key; ++ break; ++ ++ case CRYPTO_SHA1: ++ DPRINTK("%s(): SHA1\n", __FUNCTION__); ++ lacSessCtx->symOperation = CPA_CY_SYM_OP_HASH; ++ lacSessCtx->hashSetupData.hashAlgorithm = CPA_CY_SYM_HASH_SHA1; ++ lacSessCtx->hashSetupData.hashMode = CPA_CY_SYM_HASH_MODE_PLAIN; ++ lacSessCtx->hashSetupData.digestResultLenInBytes = ++ (cri->cri_mlen ? ++ cri->cri_mlen : ICP_SHA1_DIGEST_SIZE_IN_BYTES); ++ ++ break; ++ ++ case CRYPTO_SHA1_HMAC: ++ DPRINTK("%s(): SHA1_HMAC\n", __FUNCTION__); ++ lacSessCtx->symOperation = CPA_CY_SYM_OP_HASH; ++ lacSessCtx->hashSetupData.hashAlgorithm = CPA_CY_SYM_HASH_SHA1; ++ lacSessCtx->hashSetupData.hashMode = CPA_CY_SYM_HASH_MODE_AUTH; ++ lacSessCtx->hashSetupData.digestResultLenInBytes = ++ (cri->cri_mlen ? ++ cri->cri_mlen : ICP_SHA1_DIGEST_SIZE_IN_BYTES); ++ lacSessCtx->hashSetupData.authModeSetupData.authKey = ++ cri->cri_key; ++ lacSessCtx->hashSetupData.authModeSetupData.authKeyLenInBytes = ++ cri->cri_klen / NUM_BITS_IN_BYTE; ++ lacSessCtx->hashSetupData.authModeSetupData.aadLenInBytes = 0; ++ ++ break; ++ ++ case CRYPTO_SHA2_256: ++ DPRINTK("%s(): SHA256\n", __FUNCTION__); ++ lacSessCtx->symOperation = CPA_CY_SYM_OP_HASH; ++ lacSessCtx->hashSetupData.hashAlgorithm = ++ CPA_CY_SYM_HASH_SHA256; ++ lacSessCtx->hashSetupData.hashMode = CPA_CY_SYM_HASH_MODE_PLAIN; ++ lacSessCtx->hashSetupData.digestResultLenInBytes = ++ (cri->cri_mlen ? ++ cri->cri_mlen : ICP_SHA256_DIGEST_SIZE_IN_BYTES); ++ ++ break; ++ ++ case CRYPTO_SHA2_256_HMAC: ++ DPRINTK("%s(): SHA256_HMAC\n", __FUNCTION__); ++ lacSessCtx->symOperation = CPA_CY_SYM_OP_HASH; ++ lacSessCtx->hashSetupData.hashAlgorithm = ++ CPA_CY_SYM_HASH_SHA256; ++ lacSessCtx->hashSetupData.hashMode = CPA_CY_SYM_HASH_MODE_AUTH; ++ lacSessCtx->hashSetupData.digestResultLenInBytes = ++ (cri->cri_mlen ? ++ cri->cri_mlen : ICP_SHA256_DIGEST_SIZE_IN_BYTES); ++ lacSessCtx->hashSetupData.authModeSetupData.authKey = ++ cri->cri_key; ++ lacSessCtx->hashSetupData.authModeSetupData.authKeyLenInBytes = ++ cri->cri_klen / NUM_BITS_IN_BYTE; ++ lacSessCtx->hashSetupData.authModeSetupData.aadLenInBytes = 0; ++ ++ break; ++ ++ case CRYPTO_SHA2_384: ++ DPRINTK("%s(): SHA384\n", __FUNCTION__); ++ lacSessCtx->symOperation = CPA_CY_SYM_OP_HASH; ++ lacSessCtx->hashSetupData.hashAlgorithm = ++ CPA_CY_SYM_HASH_SHA384; ++ lacSessCtx->hashSetupData.hashMode = CPA_CY_SYM_HASH_MODE_PLAIN; ++ lacSessCtx->hashSetupData.digestResultLenInBytes = ++ (cri->cri_mlen ? ++ cri->cri_mlen : ICP_SHA384_DIGEST_SIZE_IN_BYTES); ++ ++ break; ++ ++ case CRYPTO_SHA2_384_HMAC: ++ DPRINTK("%s(): SHA384_HMAC\n", __FUNCTION__); ++ lacSessCtx->symOperation = CPA_CY_SYM_OP_HASH; ++ lacSessCtx->hashSetupData.hashAlgorithm = ++ CPA_CY_SYM_HASH_SHA384; ++ lacSessCtx->hashSetupData.hashMode = CPA_CY_SYM_HASH_MODE_AUTH; ++ lacSessCtx->hashSetupData.digestResultLenInBytes = ++ (cri->cri_mlen ? ++ cri->cri_mlen : ICP_SHA384_DIGEST_SIZE_IN_BYTES); ++ lacSessCtx->hashSetupData.authModeSetupData.authKey = ++ cri->cri_key; ++ lacSessCtx->hashSetupData.authModeSetupData.authKeyLenInBytes = ++ cri->cri_klen / NUM_BITS_IN_BYTE; ++ lacSessCtx->hashSetupData.authModeSetupData.aadLenInBytes = 0; ++ ++ break; ++ ++ case CRYPTO_SHA2_512: ++ DPRINTK("%s(): SHA512\n", __FUNCTION__); ++ lacSessCtx->symOperation = CPA_CY_SYM_OP_HASH; ++ lacSessCtx->hashSetupData.hashAlgorithm = ++ CPA_CY_SYM_HASH_SHA512; ++ lacSessCtx->hashSetupData.hashMode = CPA_CY_SYM_HASH_MODE_PLAIN; ++ lacSessCtx->hashSetupData.digestResultLenInBytes = ++ (cri->cri_mlen ? ++ cri->cri_mlen : ICP_SHA512_DIGEST_SIZE_IN_BYTES); ++ ++ break; ++ ++ case CRYPTO_SHA2_512_HMAC: ++ DPRINTK("%s(): SHA512_HMAC\n", __FUNCTION__); ++ lacSessCtx->symOperation = CPA_CY_SYM_OP_HASH; ++ lacSessCtx->hashSetupData.hashAlgorithm = ++ CPA_CY_SYM_HASH_SHA512; ++ lacSessCtx->hashSetupData.hashMode = CPA_CY_SYM_HASH_MODE_AUTH; ++ lacSessCtx->hashSetupData.digestResultLenInBytes = ++ (cri->cri_mlen ? ++ cri->cri_mlen : ICP_SHA512_DIGEST_SIZE_IN_BYTES); ++ lacSessCtx->hashSetupData.authModeSetupData.authKey = ++ cri->cri_key; ++ lacSessCtx->hashSetupData.authModeSetupData.authKeyLenInBytes = ++ cri->cri_klen / NUM_BITS_IN_BYTE; ++ lacSessCtx->hashSetupData.authModeSetupData.aadLenInBytes = 0; ++ ++ break; ++ ++ case CRYPTO_MD5: ++ DPRINTK("%s(): MD5\n", __FUNCTION__); ++ lacSessCtx->symOperation = CPA_CY_SYM_OP_HASH; ++ lacSessCtx->hashSetupData.hashAlgorithm = CPA_CY_SYM_HASH_MD5; ++ lacSessCtx->hashSetupData.hashMode = CPA_CY_SYM_HASH_MODE_PLAIN; ++ lacSessCtx->hashSetupData.digestResultLenInBytes = ++ (cri->cri_mlen ? ++ cri->cri_mlen : ICP_MD5_DIGEST_SIZE_IN_BYTES); ++ ++ break; ++ ++ case CRYPTO_MD5_HMAC: ++ DPRINTK("%s(): MD5_HMAC\n", __FUNCTION__); ++ lacSessCtx->symOperation = CPA_CY_SYM_OP_HASH; ++ lacSessCtx->hashSetupData.hashAlgorithm = CPA_CY_SYM_HASH_MD5; ++ lacSessCtx->hashSetupData.hashMode = CPA_CY_SYM_HASH_MODE_AUTH; ++ lacSessCtx->hashSetupData.digestResultLenInBytes = ++ (cri->cri_mlen ? ++ cri->cri_mlen : ICP_MD5_DIGEST_SIZE_IN_BYTES); ++ lacSessCtx->hashSetupData.authModeSetupData.authKey = ++ cri->cri_key; ++ lacSessCtx->hashSetupData.authModeSetupData.authKeyLenInBytes = ++ cri->cri_klen / NUM_BITS_IN_BYTE; ++ lacSessCtx->hashSetupData.authModeSetupData.aadLenInBytes = 0; ++ ++ break; ++ ++ default: ++ DPRINTK("%s(): ALG Setup FAIL\n", __FUNCTION__); ++ return ICP_OCF_DRV_STATUS_FAIL; ++ } ++ ++ return ICP_OCF_DRV_STATUS_SUCCESS; ++} ++ ++/* Name : icp_ocfDrvFreeOCFSession ++ * ++ * Description : This function deletes all existing Session data representing ++ * the Cryptographic session established between OCF and this driver. This ++ * also includes freeing the memory allocated for the session context. The ++ * session object is also removed from the session linked list. ++ */ ++static void icp_ocfDrvFreeOCFSession(struct icp_drvSessionData *sessionData) ++{ ++ ++ sessionData->inUse = ICP_SESSION_DEREGISTERED; ++ ++ /*ENTER CRITICAL SECTION */ ++ spin_lock_bh(&icp_ocfDrvSymSessInfoListSpinlock); ++ ++ if (CPA_TRUE == atomic_read(&icp_ocfDrvIsExiting)) { ++ /*If the Driver is exiting, allow that process to ++ handle any deletions */ ++ /*EXIT CRITICAL SECTION */ ++ spin_unlock_bh(&icp_ocfDrvSymSessInfoListSpinlock); ++ return; ++ } ++ ++ atomic_dec(&num_ocf_to_drv_registered_sessions); ++ ++ list_del(&(sessionData->listNode)); ++ ++ /*EXIT CRITICAL SECTION */ ++ spin_unlock_bh(&icp_ocfDrvSymSessInfoListSpinlock); ++ ++ if (NULL != sessionData->sessHandle) { ++ kfree(sessionData->sessHandle); ++ } ++ kmem_cache_free(drvSessionData_zone, sessionData); ++} ++ ++/* Name : icp_ocfDrvFreeLACSession ++ * ++ * Description : This attempts to deregister a LAC session. If it fails, the ++ * deregistation retry function is called. ++ */ ++int icp_ocfDrvFreeLACSession(device_t dev, uint64_t sid) ++{ ++ CpaCySymSessionCtx sessionToDeregister = NULL; ++ struct icp_drvSessionData *sessionData = NULL; ++ CpaStatus lacStatus = CPA_STATUS_SUCCESS; ++ int retval = 0; ++ ++ sessionData = (struct icp_drvSessionData *)CRYPTO_SESID2LID(sid); ++ if (NULL == sessionData) { ++ EPRINTK("%s(): OCF Free session called with Null Session ID.\n", ++ __FUNCTION__); ++ return EINVAL; ++ } ++ ++ sessionToDeregister = sessionData->sessHandle; ++ ++ if (ICP_SESSION_INITIALISED == sessionData->inUse) { ++ DPRINTK("%s() Session not registered with LAC\n", __FUNCTION__); ++ } else if (NULL == sessionData->sessHandle) { ++ EPRINTK ++ ("%s(): OCF Free session called with Null Session Handle.\n", ++ __FUNCTION__); ++ return EINVAL; ++ } else { ++ lacStatus = cpaCySymRemoveSession(CPA_INSTANCE_HANDLE_SINGLE, ++ sessionToDeregister); ++ if (CPA_STATUS_RETRY == lacStatus) { ++ if (ICP_OCF_DRV_STATUS_SUCCESS != ++ icp_ocfDrvDeregRetry(&sessionToDeregister)) { ++ /* the retry function increments the ++ dereg failed count */ ++ DPRINTK("%s(): LAC failed to deregister the " ++ "session. (localSessionId= %p)\n", ++ __FUNCTION__, sessionToDeregister); ++ retval = EPERM; ++ } ++ ++ } else if (CPA_STATUS_SUCCESS != lacStatus) { ++ DPRINTK("%s(): LAC failed to deregister the session. " ++ "localSessionId= %p, lacStatus = %d\n", ++ __FUNCTION__, sessionToDeregister, lacStatus); ++ atomic_inc(&lac_session_failed_dereg_count); ++ retval = EPERM; ++ } ++ } ++ ++ icp_ocfDrvFreeOCFSession(sessionData); ++ return retval; ++ ++} ++ ++/* Name : icp_ocfDrvAlgCheck ++ * ++ * Description : This function checks whether the cryptodesc argument pertains ++ * to a sym or hash function ++ */ ++static int icp_ocfDrvAlgCheck(struct cryptodesc *crp_desc) ++{ ++ ++ if (crp_desc->crd_alg == CRYPTO_3DES_CBC || ++ crp_desc->crd_alg == CRYPTO_AES_CBC || ++ crp_desc->crd_alg == CRYPTO_DES_CBC || ++ crp_desc->crd_alg == CRYPTO_NULL_CBC || ++ crp_desc->crd_alg == CRYPTO_ARC4) { ++ return ICP_OCF_DRV_ALG_CIPHER; ++ } ++ ++ return ICP_OCF_DRV_ALG_HASH; ++} ++ ++/* Name : icp_ocfDrvSymProcess ++ * ++ * Description : This function will map symmetric functionality calls from OCF ++ * to the LAC API. It will also allocate memory to store the session context. ++ * ++ * Notes: If it is the first perform call for a given session, then a LAC ++ * session is registered. After the session is registered, no checks as ++ * to whether session paramaters have changed (e.g. alg chain order) are ++ * done. ++ */ ++int icp_ocfDrvSymProcess(device_t dev, struct cryptop *crp, int hint) ++{ ++ struct icp_drvSessionData *sessionData = NULL; ++ struct icp_drvOpData *drvOpData = NULL; ++ CpaStatus lacStatus = CPA_STATUS_SUCCESS; ++ Cpa32U sessionCtxSizeInBytes = 0; ++ uint16_t numBufferListArray = 0; ++ ++ if (NULL == crp) { ++ DPRINTK("%s(): Invalid input parameters, cryptop is NULL\n", ++ __FUNCTION__); ++ return EINVAL; ++ } ++ ++ if (NULL == crp->crp_desc) { ++ DPRINTK("%s(): Invalid input parameters, no crp_desc attached " ++ "to crp\n", __FUNCTION__); ++ crp->crp_etype = EINVAL; ++ return EINVAL; ++ } ++ ++ if (NULL == crp->crp_buf) { ++ DPRINTK("%s(): Invalid input parameters, no buffer attached " ++ "to crp\n", __FUNCTION__); ++ crp->crp_etype = EINVAL; ++ return EINVAL; ++ } ++ ++ if (CPA_TRUE == atomic_read(&icp_ocfDrvIsExiting)) { ++ crp->crp_etype = EFAULT; ++ return EFAULT; ++ } ++ ++ sessionData = (struct icp_drvSessionData *) ++ (CRYPTO_SESID2LID(crp->crp_sid)); ++ if (NULL == sessionData) { ++ DPRINTK("%s(): Invalid input parameters, Null Session ID \n", ++ __FUNCTION__); ++ crp->crp_etype = EINVAL; ++ return EINVAL; ++ } ++ ++/*If we get a request against a deregisted session, cancel operation*/ ++ if (ICP_SESSION_DEREGISTERED == sessionData->inUse) { ++ DPRINTK("%s(): Session ID %d was deregistered \n", ++ __FUNCTION__, (int)(CRYPTO_SESID2LID(crp->crp_sid))); ++ crp->crp_etype = EFAULT; ++ return EFAULT; ++ } ++ ++/*If none of the session states are set, then the session structure was either ++ not initialised properly or we are reading from a freed memory area (possible ++ due to OCF batch mode not removing queued requests against deregistered ++ sessions*/ ++ if (ICP_SESSION_INITIALISED != sessionData->inUse && ++ ICP_SESSION_RUNNING != sessionData->inUse) { ++ DPRINTK("%s(): Session - ID %d - not properly initialised or " ++ "memory freed back to the kernel \n", ++ __FUNCTION__, (int)(CRYPTO_SESID2LID(crp->crp_sid))); ++ crp->crp_etype = EINVAL; ++ return EINVAL; ++ } ++ ++ /*For the below checks, remember error checking is already done in LAC. ++ We're not validating inputs subsequent to registration */ ++ if (sessionData->inUse == ICP_SESSION_INITIALISED) { ++ DPRINTK("%s(): Initialising session\n", __FUNCTION__); ++ ++ if (NULL != crp->crp_desc->crd_next) { ++ if (ICP_OCF_DRV_ALG_CIPHER == ++ icp_ocfDrvAlgCheck(crp->crp_desc)) { ++ ++ sessionData->lacSessCtx.algChainOrder = ++ CPA_CY_SYM_ALG_CHAIN_ORDER_CIPHER_THEN_HASH; ++ ++ if (crp->crp_desc->crd_flags & CRD_F_ENCRYPT) { ++ sessionData->lacSessCtx.cipherSetupData. ++ cipherDirection = ++ CPA_CY_SYM_CIPHER_DIRECTION_ENCRYPT; ++ } else { ++ sessionData->lacSessCtx.cipherSetupData. ++ cipherDirection = ++ CPA_CY_SYM_CIPHER_DIRECTION_DECRYPT; ++ } ++ } else { ++ sessionData->lacSessCtx.algChainOrder = ++ CPA_CY_SYM_ALG_CHAIN_ORDER_HASH_THEN_CIPHER; ++ ++ if (crp->crp_desc->crd_next->crd_flags & ++ CRD_F_ENCRYPT) { ++ sessionData->lacSessCtx.cipherSetupData. ++ cipherDirection = ++ CPA_CY_SYM_CIPHER_DIRECTION_ENCRYPT; ++ } else { ++ sessionData->lacSessCtx.cipherSetupData. ++ cipherDirection = ++ CPA_CY_SYM_CIPHER_DIRECTION_DECRYPT; ++ } ++ ++ } ++ ++ } else if (ICP_OCF_DRV_ALG_CIPHER == ++ icp_ocfDrvAlgCheck(crp->crp_desc)) { ++ if (crp->crp_desc->crd_flags & CRD_F_ENCRYPT) { ++ sessionData->lacSessCtx.cipherSetupData. ++ cipherDirection = ++ CPA_CY_SYM_CIPHER_DIRECTION_ENCRYPT; ++ } else { ++ sessionData->lacSessCtx.cipherSetupData. ++ cipherDirection = ++ CPA_CY_SYM_CIPHER_DIRECTION_DECRYPT; ++ } ++ ++ } ++ ++ /*No action required for standalone Auth here */ ++ ++ /* Allocate memory for SymSessionCtx before the Session Registration */ ++ lacStatus = ++ cpaCySymSessionCtxGetSize(CPA_INSTANCE_HANDLE_SINGLE, ++ &(sessionData->lacSessCtx), ++ &sessionCtxSizeInBytes); ++ if (CPA_STATUS_SUCCESS != lacStatus) { ++ EPRINTK("%s(): cpaCySymSessionCtxGetSize failed - %d\n", ++ __FUNCTION__, lacStatus); ++ return EINVAL; ++ } ++ sessionData->sessHandle = ++ kmalloc(sessionCtxSizeInBytes, GFP_ATOMIC); ++ if (NULL == sessionData->sessHandle) { ++ EPRINTK ++ ("%s(): Failed to get memory for SymSessionCtx\n", ++ __FUNCTION__); ++ return ENOMEM; ++ } ++ ++ lacStatus = cpaCySymInitSession(CPA_INSTANCE_HANDLE_SINGLE, ++ icp_ocfDrvSymCallBack, ++ &(sessionData->lacSessCtx), ++ sessionData->sessHandle); ++ ++ if (CPA_STATUS_SUCCESS != lacStatus) { ++ EPRINTK("%s(): cpaCySymInitSession failed -%d \n", ++ __FUNCTION__, lacStatus); ++ return EFAULT; ++ } ++ ++ sessionData->inUse = ICP_SESSION_RUNNING; ++ } ++ ++ drvOpData = kmem_cache_zalloc(drvOpData_zone, GFP_ATOMIC); ++ if (NULL == drvOpData) { ++ EPRINTK("%s():Failed to get memory for drvOpData\n", ++ __FUNCTION__); ++ crp->crp_etype = ENOMEM; ++ return ENOMEM; ++ } ++ ++ drvOpData->lacOpData.pSessionCtx = sessionData->sessHandle; ++ drvOpData->digestSizeInBytes = sessionData->lacSessCtx.hashSetupData. ++ digestResultLenInBytes; ++ drvOpData->crp = crp; ++ ++ /* Set the default buffer list array memory allocation */ ++ drvOpData->srcBuffer.pBuffers = drvOpData->bufferListArray; ++ drvOpData->numBufferListArray = ICP_OCF_DRV_DEFAULT_BUFFLIST_ARRAYS; ++ ++ /* ++ * Allocate buffer list array memory allocation if the ++ * data fragment is more than the default allocation ++ */ ++ if (crp->crp_flags & CRYPTO_F_SKBUF) { ++ numBufferListArray = icp_ocfDrvGetSkBuffFrags((struct sk_buff *) ++ crp->crp_buf); ++ if (ICP_OCF_DRV_DEFAULT_BUFFLIST_ARRAYS < numBufferListArray) { ++ DPRINTK("%s() numBufferListArray more than default\n", ++ __FUNCTION__); ++ drvOpData->srcBuffer.pBuffers = NULL; ++ drvOpData->srcBuffer.pBuffers = ++ kmalloc(numBufferListArray * ++ sizeof(CpaFlatBuffer), GFP_ATOMIC); ++ if (NULL == drvOpData->srcBuffer.pBuffers) { ++ EPRINTK("%s() Failed to get memory for " ++ "pBuffers\n", __FUNCTION__); ++ kmem_cache_free(drvOpData_zone, drvOpData); ++ crp->crp_etype = ENOMEM; ++ return ENOMEM; ++ } ++ drvOpData->numBufferListArray = numBufferListArray; ++ } ++ } ++ ++ /* ++ * Check the type of buffer structure we got and convert it into ++ * CpaBufferList format. ++ */ ++ if (crp->crp_flags & CRYPTO_F_SKBUF) { ++ if (ICP_OCF_DRV_STATUS_SUCCESS != ++ icp_ocfDrvSkBuffToBufferList((struct sk_buff *)crp->crp_buf, ++ &(drvOpData->srcBuffer))) { ++ EPRINTK("%s():Failed to translate from SK_BUF " ++ "to bufferlist\n", __FUNCTION__); ++ crp->crp_etype = EINVAL; ++ goto err; ++ } ++ ++ drvOpData->bufferType = CRYPTO_F_SKBUF; ++ } else if (crp->crp_flags & CRYPTO_F_IOV) { ++ /* OCF only supports IOV of one entry. */ ++ if (NUM_IOV_SUPPORTED == ++ ((struct uio *)(crp->crp_buf))->uio_iovcnt) { ++ ++ icp_ocfDrvPtrAndLenToBufferList(((struct uio *)(crp-> ++ crp_buf))-> ++ uio_iov[0].iov_base, ++ ((struct uio *)(crp-> ++ crp_buf))-> ++ uio_iov[0].iov_len, ++ &(drvOpData-> ++ srcBuffer)); ++ ++ drvOpData->bufferType = CRYPTO_F_IOV; ++ ++ } else { ++ DPRINTK("%s():Unable to handle IOVs with lengths of " ++ "greater than one!\n", __FUNCTION__); ++ crp->crp_etype = EINVAL; ++ goto err; ++ } ++ ++ } else { ++ icp_ocfDrvPtrAndLenToBufferList(crp->crp_buf, ++ crp->crp_ilen, ++ &(drvOpData->srcBuffer)); ++ ++ drvOpData->bufferType = CRYPTO_BUF_CONTIG; ++ } ++ ++ if (ICP_OCF_DRV_STATUS_SUCCESS != ++ icp_ocfDrvProcessDataSetup(drvOpData, drvOpData->crp->crp_desc)) { ++ crp->crp_etype = EINVAL; ++ goto err; ++ } ++ ++ if (drvOpData->crp->crp_desc->crd_next != NULL) { ++ if (icp_ocfDrvProcessDataSetup(drvOpData, drvOpData->crp-> ++ crp_desc->crd_next)) { ++ crp->crp_etype = EINVAL; ++ goto err; ++ } ++ ++ } ++ ++ /* Allocate srcBuffer's private meta data */ ++ if (ICP_OCF_DRV_STATUS_SUCCESS != ++ icp_ocfDrvAllocMetaData(&(drvOpData->srcBuffer), drvOpData)) { ++ EPRINTK("%s() icp_ocfDrvAllocMetaData failed\n", __FUNCTION__); ++ memset(&(drvOpData->lacOpData), 0, sizeof(CpaCySymOpData)); ++ crp->crp_etype = EINVAL; ++ goto err; ++ } ++ ++ /* Perform "in-place" crypto operation */ ++ lacStatus = cpaCySymPerformOp(CPA_INSTANCE_HANDLE_SINGLE, ++ (void *)drvOpData, ++ &(drvOpData->lacOpData), ++ &(drvOpData->srcBuffer), ++ &(drvOpData->srcBuffer), ++ &(drvOpData->verifyResult)); ++ if (CPA_STATUS_RETRY == lacStatus) { ++ DPRINTK("%s(): cpaCySymPerformOp retry, lacStatus = %d\n", ++ __FUNCTION__, lacStatus); ++ memset(&(drvOpData->lacOpData), 0, sizeof(CpaCySymOpData)); ++ crp->crp_etype = EINVAL; ++ goto err; ++ } ++ if (CPA_STATUS_SUCCESS != lacStatus) { ++ EPRINTK("%s(): cpaCySymPerformOp failed, lacStatus = %d\n", ++ __FUNCTION__, lacStatus); ++ memset(&(drvOpData->lacOpData), 0, sizeof(CpaCySymOpData)); ++ crp->crp_etype = EINVAL; ++ goto err; ++ } ++ ++ return 0; //OCF success status value ++ ++ err: ++ if (drvOpData->numBufferListArray > ICP_OCF_DRV_DEFAULT_BUFFLIST_ARRAYS) { ++ kfree(drvOpData->srcBuffer.pBuffers); ++ } ++ icp_ocfDrvFreeMetaData(&(drvOpData->srcBuffer)); ++ kmem_cache_free(drvOpData_zone, drvOpData); ++ ++ return crp->crp_etype; ++} ++ ++/* Name : icp_ocfDrvProcessDataSetup ++ * ++ * Description : This function will setup all the cryptographic operation data ++ * that is required by LAC to execute the operation. ++ */ ++static int icp_ocfDrvProcessDataSetup(struct icp_drvOpData *drvOpData, ++ struct cryptodesc *crp_desc) ++{ ++ CpaCyRandGenOpData randGenOpData; ++ CpaFlatBuffer randData; ++ ++ drvOpData->lacOpData.packetType = CPA_CY_SYM_PACKET_TYPE_FULL; ++ ++ /* Convert from the cryptop to the ICP LAC crypto parameters */ ++ switch (crp_desc->crd_alg) { ++ case CRYPTO_NULL_CBC: ++ drvOpData->lacOpData. ++ cryptoStartSrcOffsetInBytes = crp_desc->crd_skip; ++ drvOpData->lacOpData. ++ messageLenToCipherInBytes = crp_desc->crd_len; ++ drvOpData->verifyResult = CPA_FALSE; ++ drvOpData->lacOpData.ivLenInBytes = NULL_BLOCK_LEN; ++ break; ++ case CRYPTO_DES_CBC: ++ drvOpData->lacOpData. ++ cryptoStartSrcOffsetInBytes = crp_desc->crd_skip; ++ drvOpData->lacOpData. ++ messageLenToCipherInBytes = crp_desc->crd_len; ++ drvOpData->verifyResult = CPA_FALSE; ++ drvOpData->lacOpData.ivLenInBytes = DES_BLOCK_LEN; ++ break; ++ case CRYPTO_3DES_CBC: ++ drvOpData->lacOpData. ++ cryptoStartSrcOffsetInBytes = crp_desc->crd_skip; ++ drvOpData->lacOpData. ++ messageLenToCipherInBytes = crp_desc->crd_len; ++ drvOpData->verifyResult = CPA_FALSE; ++ drvOpData->lacOpData.ivLenInBytes = DES3_BLOCK_LEN; ++ break; ++ case CRYPTO_ARC4: ++ drvOpData->lacOpData. ++ cryptoStartSrcOffsetInBytes = crp_desc->crd_skip; ++ drvOpData->lacOpData. ++ messageLenToCipherInBytes = crp_desc->crd_len; ++ drvOpData->verifyResult = CPA_FALSE; ++ drvOpData->lacOpData.ivLenInBytes = ARC4_COUNTER_LEN; ++ break; ++ case CRYPTO_AES_CBC: ++ drvOpData->lacOpData. ++ cryptoStartSrcOffsetInBytes = crp_desc->crd_skip; ++ drvOpData->lacOpData. ++ messageLenToCipherInBytes = crp_desc->crd_len; ++ drvOpData->verifyResult = CPA_FALSE; ++ drvOpData->lacOpData.ivLenInBytes = RIJNDAEL128_BLOCK_LEN; ++ break; ++ case CRYPTO_SHA1: ++ case CRYPTO_SHA1_HMAC: ++ case CRYPTO_SHA2_256: ++ case CRYPTO_SHA2_256_HMAC: ++ case CRYPTO_SHA2_384: ++ case CRYPTO_SHA2_384_HMAC: ++ case CRYPTO_SHA2_512: ++ case CRYPTO_SHA2_512_HMAC: ++ case CRYPTO_MD5: ++ case CRYPTO_MD5_HMAC: ++ drvOpData->lacOpData. ++ hashStartSrcOffsetInBytes = crp_desc->crd_skip; ++ drvOpData->lacOpData. ++ messageLenToHashInBytes = crp_desc->crd_len; ++ drvOpData->lacOpData. ++ pDigestResult = ++ icp_ocfDrvDigestPointerFind(drvOpData, crp_desc); ++ ++ if (NULL == drvOpData->lacOpData.pDigestResult) { ++ DPRINTK("%s(): ERROR - could not calculate " ++ "Digest Result memory address\n", __FUNCTION__); ++ return ICP_OCF_DRV_STATUS_FAIL; ++ } ++ ++ drvOpData->lacOpData.digestVerify = CPA_FALSE; ++ break; ++ default: ++ DPRINTK("%s(): Crypto process error - algorithm not " ++ "found \n", __FUNCTION__); ++ return ICP_OCF_DRV_STATUS_FAIL; ++ } ++ ++ /* Figure out what the IV is supposed to be */ ++ if ((crp_desc->crd_alg == CRYPTO_DES_CBC) || ++ (crp_desc->crd_alg == CRYPTO_3DES_CBC) || ++ (crp_desc->crd_alg == CRYPTO_AES_CBC)) { ++ /*ARC4 doesn't use an IV */ ++ if (crp_desc->crd_flags & CRD_F_IV_EXPLICIT) { ++ /* Explicit IV provided to OCF */ ++ drvOpData->lacOpData.pIv = crp_desc->crd_iv; ++ } else { ++ /* IV is not explicitly provided to OCF */ ++ ++ /* Point the LAC OP Data IV pointer to our allocated ++ storage location for this session. */ ++ drvOpData->lacOpData.pIv = drvOpData->ivData; ++ ++ if ((crp_desc->crd_flags & CRD_F_ENCRYPT) && ++ ((crp_desc->crd_flags & CRD_F_IV_PRESENT) == 0)) { ++ ++ /* Encrypting - need to create IV */ ++ randGenOpData.generateBits = CPA_TRUE; ++ randGenOpData.lenInBytes = MAX_IV_LEN_IN_BYTES; ++ ++ icp_ocfDrvPtrAndLenToFlatBuffer((Cpa8U *) ++ drvOpData-> ++ ivData, ++ MAX_IV_LEN_IN_BYTES, ++ &randData); ++ ++ if (CPA_STATUS_SUCCESS != ++ cpaCyRandGen(CPA_INSTANCE_HANDLE_SINGLE, ++ NULL, NULL, ++ &randGenOpData, &randData)) { ++ DPRINTK("%s(): ERROR - Failed to" ++ " generate" ++ " Initialisation Vector\n", ++ __FUNCTION__); ++ return ICP_OCF_DRV_STATUS_FAIL; ++ } ++ ++ crypto_copyback(drvOpData->crp-> ++ crp_flags, ++ drvOpData->crp->crp_buf, ++ crp_desc->crd_inject, ++ drvOpData->lacOpData. ++ ivLenInBytes, ++ (caddr_t) (drvOpData->lacOpData. ++ pIv)); ++ } else { ++ /* Reading IV from buffer */ ++ crypto_copydata(drvOpData->crp-> ++ crp_flags, ++ drvOpData->crp->crp_buf, ++ crp_desc->crd_inject, ++ drvOpData->lacOpData. ++ ivLenInBytes, ++ (caddr_t) (drvOpData->lacOpData. ++ pIv)); ++ } ++ ++ } ++ ++ } ++ ++ return ICP_OCF_DRV_STATUS_SUCCESS; ++} ++ ++/* Name : icp_ocfDrvDigestPointerFind ++ * ++ * Description : This function is used to find the memory address of where the ++ * digest information shall be stored in. Input buffer types are an skbuff, iov ++ * or flat buffer. The address is found using the buffer data start address and ++ * an offset. ++ * ++ * Note: In the case of a linux skbuff, the digest address may exist within ++ * a memory space linked to from the start buffer. These linked memory spaces ++ * must be traversed by the data length offset in order to find the digest start ++ * address. Whether there is enough space for the digest must also be checked. ++ */ ++ ++static uint8_t *icp_ocfDrvDigestPointerFind(struct icp_drvOpData *drvOpData, ++ struct cryptodesc *crp_desc) ++{ ++ ++ int offsetInBytes = crp_desc->crd_inject; ++ uint32_t digestSizeInBytes = drvOpData->digestSizeInBytes; ++ uint8_t *flat_buffer_base = NULL; ++ int flat_buffer_length = 0; ++ struct sk_buff *skb; ++ ++ if (drvOpData->crp->crp_flags & CRYPTO_F_SKBUF) { ++ /*check if enough overall space to store hash */ ++ skb = (struct sk_buff *)(drvOpData->crp->crp_buf); ++ ++ if (skb->len < (offsetInBytes + digestSizeInBytes)) { ++ DPRINTK("%s() Not enough space for Digest" ++ " payload after the offset (%d), " ++ "digest size (%d) \n", __FUNCTION__, ++ offsetInBytes, digestSizeInBytes); ++ return NULL; ++ } ++ ++ return icp_ocfDrvSkbuffDigestPointerFind(drvOpData, ++ offsetInBytes, ++ digestSizeInBytes); ++ ++ } else { ++ /* IOV or flat buffer */ ++ if (drvOpData->crp->crp_flags & CRYPTO_F_IOV) { ++ /*single IOV check has already been done */ ++ flat_buffer_base = ((struct uio *) ++ (drvOpData->crp->crp_buf))-> ++ uio_iov[0].iov_base; ++ flat_buffer_length = ((struct uio *) ++ (drvOpData->crp->crp_buf))-> ++ uio_iov[0].iov_len; ++ } else { ++ flat_buffer_base = (uint8_t *) drvOpData->crp->crp_buf; ++ flat_buffer_length = drvOpData->crp->crp_ilen; ++ } ++ ++ if (flat_buffer_length < (offsetInBytes + digestSizeInBytes)) { ++ DPRINTK("%s() Not enough space for Digest " ++ "(IOV/Flat Buffer) \n", __FUNCTION__); ++ return NULL; ++ } else { ++ return (uint8_t *) (flat_buffer_base + offsetInBytes); ++ } ++ } ++ DPRINTK("%s() Should not reach this point\n", __FUNCTION__); ++ return NULL; ++} ++ ++/* Name : icp_ocfDrvSkbuffDigestPointerFind ++ * ++ * Description : This function is used by icp_ocfDrvDigestPointerFind to process ++ * the non-linear portion of the skbuff if the fragmentation type is a linked ++ * list (frag_list is not NULL in the skb_shared_info structure) ++ */ ++static inline uint8_t *icp_ocfDrvSkbuffDigestPointerFind(struct icp_drvOpData ++ *drvOpData, ++ int offsetInBytes, ++ uint32_t ++ digestSizeInBytes) ++{ ++ ++ struct sk_buff *skb = NULL; ++ struct skb_shared_info *skb_shared = NULL; ++ ++ uint32_t skbuffisnonlinear = 0; ++ ++ uint32_t skbheadlen = 0; ++ ++ skb = (struct sk_buff *)(drvOpData->crp->crp_buf); ++ skbuffisnonlinear = skb_is_nonlinear(skb); ++ ++ skbheadlen = skb_headlen(skb); ++ ++ /*Linear skb checks */ ++ if (skbheadlen > offsetInBytes) { ++ ++ if (skbheadlen >= (offsetInBytes + digestSizeInBytes)) { ++ return (uint8_t *) (skb->data + offsetInBytes); ++ } else { ++ DPRINTK("%s() Auth payload stretches " ++ "accross contiguous memory\n", __FUNCTION__); ++ return NULL; ++ } ++ } else { ++ if (skbuffisnonlinear) { ++ offsetInBytes -= skbheadlen; ++ } else { ++ DPRINTK("%s() Offset outside of buffer boundaries\n", ++ __FUNCTION__); ++ return NULL; ++ } ++ } ++ ++ /*Non Linear checks */ ++ skb_shared = (struct skb_shared_info *)(skb->end); ++ if (unlikely(NULL == skb_shared)) { ++ DPRINTK("%s() skbuff shared info stucture is NULL! \n", ++ __FUNCTION__); ++ return NULL; ++ } else if ((0 != skb_shared->nr_frags) && ++ (skb_shared->frag_list != NULL)) { ++ DPRINTK("%s() skbuff nr_frags AND " ++ "frag_list not supported \n", __FUNCTION__); ++ return NULL; ++ } ++ ++ /*TCP segmentation more likely than IP fragmentation */ ++ if (likely(0 != skb_shared->nr_frags)) { ++ return icp_ocfDrvDigestSkbNRFragsCheck(skb, skb_shared, ++ offsetInBytes, ++ digestSizeInBytes); ++ } else if (skb_shared->frag_list != NULL) { ++ return icp_ocfDrvDigestSkbFragListCheck(skb, skb_shared, ++ offsetInBytes, ++ digestSizeInBytes); ++ } else { ++ DPRINTK("%s() skbuff is non-linear but does not show any " ++ "linked data\n", __FUNCTION__); ++ return NULL; ++ } ++ ++} ++ ++/* Name : icp_ocfDrvDigestSkbNRFragsCheck ++ * ++ * Description : This function is used by icp_ocfDrvSkbuffDigestPointerFind to ++ * process the non-linear portion of the skbuff, if the fragmentation type is ++ * page fragments ++ */ ++static inline uint8_t *icp_ocfDrvDigestSkbNRFragsCheck(struct sk_buff *skb, ++ struct skb_shared_info ++ *skb_shared, ++ int offsetInBytes, ++ uint32_t ++ digestSizeInBytes) ++{ ++ int i = 0; ++ /*nr_frags starts from 1 */ ++ if (MAX_SKB_FRAGS < skb_shared->nr_frags) { ++ DPRINTK("%s error processing skbuff " ++ "page frame -- MAX FRAGS exceeded \n", __FUNCTION__); ++ return NULL; ++ } ++ ++ for (i = 0; i < skb_shared->nr_frags; i++) { ++ ++ if (offsetInBytes >= skb_shared->frags[i].size) { ++ /*offset still greater than data position */ ++ offsetInBytes -= skb_shared->frags[i].size; ++ } else { ++ /* found the page containing start of hash */ ++ ++ if (NULL == skb_shared->frags[i].page) { ++ DPRINTK("%s() Linked page is NULL!\n", ++ __FUNCTION__); ++ return NULL; ++ } ++ ++ if (offsetInBytes + digestSizeInBytes > ++ skb_shared->frags[i].size) { ++ DPRINTK("%s() Auth payload stretches accross " ++ "contiguous memory\n", __FUNCTION__); ++ return NULL; ++ } else { ++ return (uint8_t *) (skb_shared->frags[i].page + ++ skb_shared->frags[i]. ++ page_offset + ++ offsetInBytes); ++ } ++ } ++ /*only possible if internal page sizes are set wrong */ ++ if (offsetInBytes < 0) { ++ DPRINTK("%s error processing skbuff page frame " ++ "-- offset calculation \n", __FUNCTION__); ++ return NULL; ++ } ++ } ++ /*only possible if internal page sizes are set wrong */ ++ DPRINTK("%s error processing skbuff page frame " ++ "-- ran out of page fragments, remaining offset = %d \n", ++ __FUNCTION__, offsetInBytes); ++ return NULL; ++ ++} ++ ++/* Name : icp_ocfDrvDigestSkbFragListCheck ++ * ++ * Description : This function is used by icp_ocfDrvSkbuffDigestPointerFind to ++ * process the non-linear portion of the skbuff, if the fragmentation type is ++ * a linked list ++ * ++ */ ++static inline uint8_t *icp_ocfDrvDigestSkbFragListCheck(struct sk_buff *skb, ++ struct skb_shared_info ++ *skb_shared, ++ int offsetInBytes, ++ uint32_t ++ digestSizeInBytes) ++{ ++ ++ struct sk_buff *skb_list = skb_shared->frag_list; ++ /*check added for readability */ ++ if (NULL == skb_list) { ++ DPRINTK("%s error processing skbuff " ++ "-- no more list! \n", __FUNCTION__); ++ return NULL; ++ } ++ ++ for (; skb_list; skb_list = skb_list->next) { ++ if (NULL == skb_list) { ++ DPRINTK("%s error processing skbuff " ++ "-- no more list! \n", __FUNCTION__); ++ return NULL; ++ } ++ ++ if (offsetInBytes >= skb_list->len) { ++ offsetInBytes -= skb_list->len; ++ ++ } else { ++ if (offsetInBytes + digestSizeInBytes > skb_list->len) { ++ DPRINTK("%s() Auth payload stretches accross " ++ "contiguous memory\n", __FUNCTION__); ++ return NULL; ++ } else { ++ return (uint8_t *) ++ (skb_list->data + offsetInBytes); ++ } ++ ++ } ++ ++ /*This check is only needed if internal skb_list length values ++ are set wrong. */ ++ if (0 > offsetInBytes) { ++ DPRINTK("%s() error processing skbuff object -- offset " ++ "calculation \n", __FUNCTION__); ++ return NULL; ++ } ++ ++ } ++ ++ /*catch all for unusual for-loop exit. ++ This code should never be reached */ ++ DPRINTK("%s() Catch-All hit! Process error.\n", __FUNCTION__); ++ return NULL; ++} +diff -Nur linux-2.6.30.orig/crypto/ocf/ep80579/Makefile linux-2.6.30/crypto/ocf/ep80579/Makefile +--- linux-2.6.30.orig/crypto/ocf/ep80579/Makefile 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/ep80579/Makefile 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,107 @@ ++######################################################################### ++# ++# Targets supported ++# all - builds everything and installs ++# install - identical to all ++# depend - build dependencies ++# clean - clears derived objects except the .depend files ++# distclean- clears all derived objects and the .depend file ++# ++# @par ++# This file is provided under a dual BSD/GPLv2 license. When using or ++# redistributing this file, you may do so under either license. ++# ++# GPL LICENSE SUMMARY ++# ++# Copyright(c) 2007,2008 Intel Corporation. All rights reserved. ++# ++# This program is free software; you can redistribute it and/or modify ++# it under the terms of version 2 of the GNU General Public License as ++# published by the Free Software Foundation. ++# ++# This program is distributed in the hope that it will be useful, but ++# WITHOUT ANY WARRANTY; without even the implied warranty of ++# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ++# General Public License for more details. ++# ++# You should have received a copy of the GNU General Public License ++# along with this program; if not, write to the Free Software ++# Foundation, Inc., 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. ++# The full GNU General Public License is included in this distribution ++# in the file called LICENSE.GPL. ++# ++# Contact Information: ++# Intel Corporation ++# ++# BSD LICENSE ++# ++# Copyright(c) 2007,2008 Intel Corporation. All rights reserved. ++# All rights reserved. ++# ++# Redistribution and use in source and binary forms, with or without ++# modification, are permitted provided that the following conditions ++# are met: ++# ++# * Redistributions of source code must retain the above copyright ++# notice, this list of conditions and the following disclaimer. ++# * Redistributions in binary form must reproduce the above copyright ++# notice, this list of conditions and the following disclaimer in ++# the documentation and/or other materials provided with the ++# distribution. ++# * Neither the name of Intel Corporation nor the names of its ++# contributors may be used to endorse or promote products derived ++# from this software without specific prior written permission. ++# ++# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS ++# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT ++# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR ++# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT ++# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, ++# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT ++# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, ++# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY ++# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT ++# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE ++# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. ++# ++# ++# version: Security.L.1.0.130 ++############################################################################ ++ ++ ++####################Common variables and definitions######################## ++ ++# Ensure The ENV_DIR environmental var is defined. ++ifndef ICP_ENV_DIR ++$(error ICP_ENV_DIR is undefined. Please set the path to your environment makefile \ ++ "-> setenv ICP_ENV_DIR <path>") ++endif ++ ++#Add your project environment Makefile ++include $(ICP_ENV_DIR)/environment.mk ++ ++#include the makefile with all the default and common Make variable definitions ++include $(ICP_BUILDSYSTEM_PATH)/build_files/common.mk ++ ++#Add the name for the executable, Library or Module output definitions ++OUTPUT_NAME= icp_ocf ++ ++# List of Source Files to be compiled ++SOURCES= icp_common.c icp_sym.c icp_asym.c ++ ++#common includes between all supported OSes ++INCLUDES= -I $(ICP_API_DIR) -I$(ICP_LAC_API) \ ++-I$(ICP_OCF_SRC_DIR) ++ ++# The location of the os level makefile needs to be changed. ++include $(ICP_ENV_DIR)/$(ICP_OS)_$(ICP_OS_LEVEL).mk ++ ++# On the line directly below list the outputs you wish to build for, ++# e.g "lib_static lib_shared exe module" as show below ++install: module ++ ++###################Include rules makefiles######################## ++include $(ICP_BUILDSYSTEM_PATH)/build_files/rules.mk ++###################End of Rules inclusion######################### ++ ++ +diff -Nur linux-2.6.30.orig/crypto/ocf/hifn/hifn7751.c linux-2.6.30/crypto/ocf/hifn/hifn7751.c +--- linux-2.6.30.orig/crypto/ocf/hifn/hifn7751.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/hifn/hifn7751.c 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,2970 @@ ++/* $OpenBSD: hifn7751.c,v 1.120 2002/05/17 00:33:34 deraadt Exp $ */ ++ ++/*- ++ * Invertex AEON / Hifn 7751 driver ++ * Copyright (c) 1999 Invertex Inc. All rights reserved. ++ * Copyright (c) 1999 Theo de Raadt ++ * Copyright (c) 2000-2001 Network Security Technologies, Inc. ++ * http://www.netsec.net ++ * Copyright (c) 2003 Hifn Inc. ++ * ++ * This driver is based on a previous driver by Invertex, for which they ++ * requested: Please send any comments, feedback, bug-fixes, or feature ++ * requests to software@invertex.com. ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions ++ * are met: ++ * ++ * 1. Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * 2. Redistributions in binary form must reproduce the above copyright ++ * notice, this list of conditions and the following disclaimer in the ++ * documentation and/or other materials provided with the distribution. ++ * 3. The name of the author may not be used to endorse or promote products ++ * derived from this software without specific prior written permission. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR ++ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES ++ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. ++ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, ++ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT ++ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, ++ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY ++ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT ++ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF ++ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. ++ * ++ * Effort sponsored in part by the Defense Advanced Research Projects ++ * Agency (DARPA) and Air Force Research Laboratory, Air Force ++ * Materiel Command, USAF, under agreement number F30602-01-2-0537. ++ * ++ * ++__FBSDID("$FreeBSD: src/sys/dev/hifn/hifn7751.c,v 1.40 2007/03/21 03:42:49 sam Exp $"); ++ */ ++ ++/* ++ * Driver for various Hifn encryption processors. ++ */ ++#ifndef AUTOCONF_INCLUDED ++#include <linux/config.h> ++#endif ++#include <linux/module.h> ++#include <linux/init.h> ++#include <linux/list.h> ++#include <linux/slab.h> ++#include <linux/wait.h> ++#include <linux/sched.h> ++#include <linux/pci.h> ++#include <linux/delay.h> ++#include <linux/interrupt.h> ++#include <linux/spinlock.h> ++#include <linux/random.h> ++#include <linux/version.h> ++#include <linux/skbuff.h> ++#include <asm/io.h> ++ ++#include <cryptodev.h> ++#include <uio.h> ++#include <hifn/hifn7751reg.h> ++#include <hifn/hifn7751var.h> ++ ++#if 1 ++#define DPRINTF(a...) if (hifn_debug) { \ ++ printk("%s: ", sc ? \ ++ device_get_nameunit(sc->sc_dev) : "hifn"); \ ++ printk(a); \ ++ } else ++#else ++#define DPRINTF(a...) ++#endif ++ ++static inline int ++pci_get_revid(struct pci_dev *dev) ++{ ++ u8 rid = 0; ++ pci_read_config_byte(dev, PCI_REVISION_ID, &rid); ++ return rid; ++} ++ ++static struct hifn_stats hifnstats; ++ ++#define debug hifn_debug ++int hifn_debug = 0; ++module_param(hifn_debug, int, 0644); ++MODULE_PARM_DESC(hifn_debug, "Enable debug"); ++ ++int hifn_maxbatch = 1; ++module_param(hifn_maxbatch, int, 0644); ++MODULE_PARM_DESC(hifn_maxbatch, "max ops to batch w/o interrupt"); ++ ++#ifdef MODULE_PARM ++char *hifn_pllconfig = NULL; ++MODULE_PARM(hifn_pllconfig, "s"); ++#else ++char hifn_pllconfig[32]; /* This setting is RO after loading */ ++module_param_string(hifn_pllconfig, hifn_pllconfig, 32, 0444); ++#endif ++MODULE_PARM_DESC(hifn_pllconfig, "PLL config, ie., pci66, ext33, ..."); ++ ++#ifdef HIFN_VULCANDEV ++#include <sys/conf.h> ++#include <sys/uio.h> ++ ++static struct cdevsw vulcanpk_cdevsw; /* forward declaration */ ++#endif ++ ++/* ++ * Prototypes and count for the pci_device structure ++ */ ++static int hifn_probe(struct pci_dev *dev, const struct pci_device_id *ent); ++static void hifn_remove(struct pci_dev *dev); ++ ++static int hifn_newsession(device_t, u_int32_t *, struct cryptoini *); ++static int hifn_freesession(device_t, u_int64_t); ++static int hifn_process(device_t, struct cryptop *, int); ++ ++static device_method_t hifn_methods = { ++ /* crypto device methods */ ++ DEVMETHOD(cryptodev_newsession, hifn_newsession), ++ DEVMETHOD(cryptodev_freesession,hifn_freesession), ++ DEVMETHOD(cryptodev_process, hifn_process), ++}; ++ ++static void hifn_reset_board(struct hifn_softc *, int); ++static void hifn_reset_puc(struct hifn_softc *); ++static void hifn_puc_wait(struct hifn_softc *); ++static int hifn_enable_crypto(struct hifn_softc *); ++static void hifn_set_retry(struct hifn_softc *sc); ++static void hifn_init_dma(struct hifn_softc *); ++static void hifn_init_pci_registers(struct hifn_softc *); ++static int hifn_sramsize(struct hifn_softc *); ++static int hifn_dramsize(struct hifn_softc *); ++static int hifn_ramtype(struct hifn_softc *); ++static void hifn_sessions(struct hifn_softc *); ++#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,19) ++static irqreturn_t hifn_intr(int irq, void *arg); ++#else ++static irqreturn_t hifn_intr(int irq, void *arg, struct pt_regs *regs); ++#endif ++static u_int hifn_write_command(struct hifn_command *, u_int8_t *); ++static u_int32_t hifn_next_signature(u_int32_t a, u_int cnt); ++static void hifn_callback(struct hifn_softc *, struct hifn_command *, u_int8_t *); ++static int hifn_crypto(struct hifn_softc *, struct hifn_command *, struct cryptop *, int); ++static int hifn_readramaddr(struct hifn_softc *, int, u_int8_t *); ++static int hifn_writeramaddr(struct hifn_softc *, int, u_int8_t *); ++static int hifn_dmamap_load_src(struct hifn_softc *, struct hifn_command *); ++static int hifn_dmamap_load_dst(struct hifn_softc *, struct hifn_command *); ++static int hifn_init_pubrng(struct hifn_softc *); ++static void hifn_tick(unsigned long arg); ++static void hifn_abort(struct hifn_softc *); ++static void hifn_alloc_slot(struct hifn_softc *, int *, int *, int *, int *); ++ ++static void hifn_write_reg_0(struct hifn_softc *, bus_size_t, u_int32_t); ++static void hifn_write_reg_1(struct hifn_softc *, bus_size_t, u_int32_t); ++ ++#ifdef CONFIG_OCF_RANDOMHARVEST ++static int hifn_read_random(void *arg, u_int32_t *buf, int len); ++#endif ++ ++#define HIFN_MAX_CHIPS 8 ++static struct hifn_softc *hifn_chip_idx[HIFN_MAX_CHIPS]; ++ ++static __inline u_int32_t ++READ_REG_0(struct hifn_softc *sc, bus_size_t reg) ++{ ++ u_int32_t v = readl(sc->sc_bar0 + reg); ++ sc->sc_bar0_lastreg = (bus_size_t) -1; ++ return (v); ++} ++#define WRITE_REG_0(sc, reg, val) hifn_write_reg_0(sc, reg, val) ++ ++static __inline u_int32_t ++READ_REG_1(struct hifn_softc *sc, bus_size_t reg) ++{ ++ u_int32_t v = readl(sc->sc_bar1 + reg); ++ sc->sc_bar1_lastreg = (bus_size_t) -1; ++ return (v); ++} ++#define WRITE_REG_1(sc, reg, val) hifn_write_reg_1(sc, reg, val) ++ ++/* ++ * map in a given buffer (great on some arches :-) ++ */ ++ ++static int ++pci_map_uio(struct hifn_softc *sc, struct hifn_operand *buf, struct uio *uio) ++{ ++ struct iovec *iov = uio->uio_iov; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ buf->mapsize = 0; ++ for (buf->nsegs = 0; buf->nsegs < uio->uio_iovcnt; ) { ++ buf->segs[buf->nsegs].ds_addr = pci_map_single(sc->sc_pcidev, ++ iov->iov_base, iov->iov_len, ++ PCI_DMA_BIDIRECTIONAL); ++ buf->segs[buf->nsegs].ds_len = iov->iov_len; ++ buf->mapsize += iov->iov_len; ++ iov++; ++ buf->nsegs++; ++ } ++ /* identify this buffer by the first segment */ ++ buf->map = (void *) buf->segs[0].ds_addr; ++ return(0); ++} ++ ++/* ++ * map in a given sk_buff ++ */ ++ ++static int ++pci_map_skb(struct hifn_softc *sc,struct hifn_operand *buf,struct sk_buff *skb) ++{ ++ int i; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ buf->mapsize = 0; ++ ++ buf->segs[0].ds_addr = pci_map_single(sc->sc_pcidev, ++ skb->data, skb_headlen(skb), PCI_DMA_BIDIRECTIONAL); ++ buf->segs[0].ds_len = skb_headlen(skb); ++ buf->mapsize += buf->segs[0].ds_len; ++ ++ buf->nsegs = 1; ++ ++ for (i = 0; i < skb_shinfo(skb)->nr_frags; ) { ++ buf->segs[buf->nsegs].ds_len = skb_shinfo(skb)->frags[i].size; ++ buf->segs[buf->nsegs].ds_addr = pci_map_single(sc->sc_pcidev, ++ page_address(skb_shinfo(skb)->frags[i].page) + ++ skb_shinfo(skb)->frags[i].page_offset, ++ buf->segs[buf->nsegs].ds_len, PCI_DMA_BIDIRECTIONAL); ++ buf->mapsize += buf->segs[buf->nsegs].ds_len; ++ buf->nsegs++; ++ } ++ ++ /* identify this buffer by the first segment */ ++ buf->map = (void *) buf->segs[0].ds_addr; ++ return(0); ++} ++ ++/* ++ * map in a given contiguous buffer ++ */ ++ ++static int ++pci_map_buf(struct hifn_softc *sc,struct hifn_operand *buf, void *b, int len) ++{ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ buf->mapsize = 0; ++ buf->segs[0].ds_addr = pci_map_single(sc->sc_pcidev, ++ b, len, PCI_DMA_BIDIRECTIONAL); ++ buf->segs[0].ds_len = len; ++ buf->mapsize += buf->segs[0].ds_len; ++ buf->nsegs = 1; ++ ++ /* identify this buffer by the first segment */ ++ buf->map = (void *) buf->segs[0].ds_addr; ++ return(0); ++} ++ ++#if 0 /* not needed at this time */ ++static void ++pci_sync_iov(struct hifn_softc *sc, struct hifn_operand *buf) ++{ ++ int i; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ for (i = 0; i < buf->nsegs; i++) ++ pci_dma_sync_single_for_cpu(sc->sc_pcidev, buf->segs[i].ds_addr, ++ buf->segs[i].ds_len, PCI_DMA_BIDIRECTIONAL); ++} ++#endif ++ ++static void ++pci_unmap_buf(struct hifn_softc *sc, struct hifn_operand *buf) ++{ ++ int i; ++ DPRINTF("%s()\n", __FUNCTION__); ++ for (i = 0; i < buf->nsegs; i++) { ++ pci_unmap_single(sc->sc_pcidev, buf->segs[i].ds_addr, ++ buf->segs[i].ds_len, PCI_DMA_BIDIRECTIONAL); ++ buf->segs[i].ds_addr = 0; ++ buf->segs[i].ds_len = 0; ++ } ++ buf->nsegs = 0; ++ buf->mapsize = 0; ++ buf->map = 0; ++} ++ ++static const char* ++hifn_partname(struct hifn_softc *sc) ++{ ++ /* XXX sprintf numbers when not decoded */ ++ switch (pci_get_vendor(sc->sc_pcidev)) { ++ case PCI_VENDOR_HIFN: ++ switch (pci_get_device(sc->sc_pcidev)) { ++ case PCI_PRODUCT_HIFN_6500: return "Hifn 6500"; ++ case PCI_PRODUCT_HIFN_7751: return "Hifn 7751"; ++ case PCI_PRODUCT_HIFN_7811: return "Hifn 7811"; ++ case PCI_PRODUCT_HIFN_7951: return "Hifn 7951"; ++ case PCI_PRODUCT_HIFN_7955: return "Hifn 7955"; ++ case PCI_PRODUCT_HIFN_7956: return "Hifn 7956"; ++ } ++ return "Hifn unknown-part"; ++ case PCI_VENDOR_INVERTEX: ++ switch (pci_get_device(sc->sc_pcidev)) { ++ case PCI_PRODUCT_INVERTEX_AEON: return "Invertex AEON"; ++ } ++ return "Invertex unknown-part"; ++ case PCI_VENDOR_NETSEC: ++ switch (pci_get_device(sc->sc_pcidev)) { ++ case PCI_PRODUCT_NETSEC_7751: return "NetSec 7751"; ++ } ++ return "NetSec unknown-part"; ++ } ++ return "Unknown-vendor unknown-part"; ++} ++ ++static u_int ++checkmaxmin(struct pci_dev *dev, const char *what, u_int v, u_int min, u_int max) ++{ ++ struct hifn_softc *sc = pci_get_drvdata(dev); ++ if (v > max) { ++ device_printf(sc->sc_dev, "Warning, %s %u out of range, " ++ "using max %u\n", what, v, max); ++ v = max; ++ } else if (v < min) { ++ device_printf(sc->sc_dev, "Warning, %s %u out of range, " ++ "using min %u\n", what, v, min); ++ v = min; ++ } ++ return v; ++} ++ ++/* ++ * Select PLL configuration for 795x parts. This is complicated in ++ * that we cannot determine the optimal parameters without user input. ++ * The reference clock is derived from an external clock through a ++ * multiplier. The external clock is either the host bus (i.e. PCI) ++ * or an external clock generator. When using the PCI bus we assume ++ * the clock is either 33 or 66 MHz; for an external source we cannot ++ * tell the speed. ++ * ++ * PLL configuration is done with a string: "pci" for PCI bus, or "ext" ++ * for an external source, followed by the frequency. We calculate ++ * the appropriate multiplier and PLL register contents accordingly. ++ * When no configuration is given we default to "pci66" since that ++ * always will allow the card to work. If a card is using the PCI ++ * bus clock and in a 33MHz slot then it will be operating at half ++ * speed until the correct information is provided. ++ * ++ * We use a default setting of "ext66" because according to Mike Ham ++ * of HiFn, almost every board in existence has an external crystal ++ * populated at 66Mhz. Using PCI can be a problem on modern motherboards, ++ * because PCI33 can have clocks from 0 to 33Mhz, and some have ++ * non-PCI-compliant spread-spectrum clocks, which can confuse the pll. ++ */ ++static void ++hifn_getpllconfig(struct pci_dev *dev, u_int *pll) ++{ ++ const char *pllspec = hifn_pllconfig; ++ u_int freq, mul, fl, fh; ++ u_int32_t pllconfig; ++ char *nxt; ++ ++ if (pllspec == NULL) ++ pllspec = "ext66"; ++ fl = 33, fh = 66; ++ pllconfig = 0; ++ if (strncmp(pllspec, "ext", 3) == 0) { ++ pllspec += 3; ++ pllconfig |= HIFN_PLL_REF_SEL; ++ switch (pci_get_device(dev)) { ++ case PCI_PRODUCT_HIFN_7955: ++ case PCI_PRODUCT_HIFN_7956: ++ fl = 20, fh = 100; ++ break; ++#ifdef notyet ++ case PCI_PRODUCT_HIFN_7954: ++ fl = 20, fh = 66; ++ break; ++#endif ++ } ++ } else if (strncmp(pllspec, "pci", 3) == 0) ++ pllspec += 3; ++ freq = strtoul(pllspec, &nxt, 10); ++ if (nxt == pllspec) ++ freq = 66; ++ else ++ freq = checkmaxmin(dev, "frequency", freq, fl, fh); ++ /* ++ * Calculate multiplier. We target a Fck of 266 MHz, ++ * allowing only even values, possibly rounded down. ++ * Multipliers > 8 must set the charge pump current. ++ */ ++ mul = checkmaxmin(dev, "PLL divisor", (266 / freq) &~ 1, 2, 12); ++ pllconfig |= (mul / 2 - 1) << HIFN_PLL_ND_SHIFT; ++ if (mul > 8) ++ pllconfig |= HIFN_PLL_IS; ++ *pll = pllconfig; ++} ++ ++/* ++ * Attach an interface that successfully probed. ++ */ ++static int ++hifn_probe(struct pci_dev *dev, const struct pci_device_id *ent) ++{ ++ struct hifn_softc *sc = NULL; ++ char rbase; ++ u_int16_t ena, rev; ++ int rseg, rc; ++ unsigned long mem_start, mem_len; ++ static int num_chips = 0; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ if (pci_enable_device(dev) < 0) ++ return(-ENODEV); ++ ++ if (pci_set_mwi(dev)) ++ return(-ENODEV); ++ ++ if (!dev->irq) { ++ printk("hifn: found device with no IRQ assigned. check BIOS settings!"); ++ pci_disable_device(dev); ++ return(-ENODEV); ++ } ++ ++ sc = (struct hifn_softc *) kmalloc(sizeof(*sc), GFP_KERNEL); ++ if (!sc) ++ return(-ENOMEM); ++ memset(sc, 0, sizeof(*sc)); ++ ++ softc_device_init(sc, "hifn", num_chips, hifn_methods); ++ ++ sc->sc_pcidev = dev; ++ sc->sc_irq = -1; ++ sc->sc_cid = -1; ++ sc->sc_num = num_chips++; ++ if (sc->sc_num < HIFN_MAX_CHIPS) ++ hifn_chip_idx[sc->sc_num] = sc; ++ ++ pci_set_drvdata(sc->sc_pcidev, sc); ++ ++ spin_lock_init(&sc->sc_mtx); ++ ++ /* XXX handle power management */ ++ ++ /* ++ * The 7951 and 795x have a random number generator and ++ * public key support; note this. ++ */ ++ if (pci_get_vendor(dev) == PCI_VENDOR_HIFN && ++ (pci_get_device(dev) == PCI_PRODUCT_HIFN_7951 || ++ pci_get_device(dev) == PCI_PRODUCT_HIFN_7955 || ++ pci_get_device(dev) == PCI_PRODUCT_HIFN_7956)) ++ sc->sc_flags = HIFN_HAS_RNG | HIFN_HAS_PUBLIC; ++ /* ++ * The 7811 has a random number generator and ++ * we also note it's identity 'cuz of some quirks. ++ */ ++ if (pci_get_vendor(dev) == PCI_VENDOR_HIFN && ++ pci_get_device(dev) == PCI_PRODUCT_HIFN_7811) ++ sc->sc_flags |= HIFN_IS_7811 | HIFN_HAS_RNG; ++ ++ /* ++ * The 795x parts support AES. ++ */ ++ if (pci_get_vendor(dev) == PCI_VENDOR_HIFN && ++ (pci_get_device(dev) == PCI_PRODUCT_HIFN_7955 || ++ pci_get_device(dev) == PCI_PRODUCT_HIFN_7956)) { ++ sc->sc_flags |= HIFN_IS_7956 | HIFN_HAS_AES; ++ /* ++ * Select PLL configuration. This depends on the ++ * bus and board design and must be manually configured ++ * if the default setting is unacceptable. ++ */ ++ hifn_getpllconfig(dev, &sc->sc_pllconfig); ++ } ++ ++ /* ++ * Setup PCI resources. Note that we record the bus ++ * tag and handle for each register mapping, this is ++ * used by the READ_REG_0, WRITE_REG_0, READ_REG_1, ++ * and WRITE_REG_1 macros throughout the driver. ++ */ ++ mem_start = pci_resource_start(sc->sc_pcidev, 0); ++ mem_len = pci_resource_len(sc->sc_pcidev, 0); ++ sc->sc_bar0 = (ocf_iomem_t) ioremap(mem_start, mem_len); ++ if (!sc->sc_bar0) { ++ device_printf(sc->sc_dev, "cannot map bar%d register space\n", 0); ++ goto fail; ++ } ++ sc->sc_bar0_lastreg = (bus_size_t) -1; ++ ++ mem_start = pci_resource_start(sc->sc_pcidev, 1); ++ mem_len = pci_resource_len(sc->sc_pcidev, 1); ++ sc->sc_bar1 = (ocf_iomem_t) ioremap(mem_start, mem_len); ++ if (!sc->sc_bar1) { ++ device_printf(sc->sc_dev, "cannot map bar%d register space\n", 1); ++ goto fail; ++ } ++ sc->sc_bar1_lastreg = (bus_size_t) -1; ++ ++ /* fix up the bus size */ ++ if (pci_set_dma_mask(dev, DMA_32BIT_MASK)) { ++ device_printf(sc->sc_dev, "No usable DMA configuration, aborting.\n"); ++ goto fail; ++ } ++ if (pci_set_consistent_dma_mask(dev, DMA_32BIT_MASK)) { ++ device_printf(sc->sc_dev, ++ "No usable consistent DMA configuration, aborting.\n"); ++ goto fail; ++ } ++ ++ hifn_set_retry(sc); ++ ++ /* ++ * Setup the area where the Hifn DMA's descriptors ++ * and associated data structures. ++ */ ++ sc->sc_dma = (struct hifn_dma *) pci_alloc_consistent(dev, ++ sizeof(*sc->sc_dma), ++ &sc->sc_dma_physaddr); ++ if (!sc->sc_dma) { ++ device_printf(sc->sc_dev, "cannot alloc sc_dma\n"); ++ goto fail; ++ } ++ bzero(sc->sc_dma, sizeof(*sc->sc_dma)); ++ ++ /* ++ * Reset the board and do the ``secret handshake'' ++ * to enable the crypto support. Then complete the ++ * initialization procedure by setting up the interrupt ++ * and hooking in to the system crypto support so we'll ++ * get used for system services like the crypto device, ++ * IPsec, RNG device, etc. ++ */ ++ hifn_reset_board(sc, 0); ++ ++ if (hifn_enable_crypto(sc) != 0) { ++ device_printf(sc->sc_dev, "crypto enabling failed\n"); ++ goto fail; ++ } ++ hifn_reset_puc(sc); ++ ++ hifn_init_dma(sc); ++ hifn_init_pci_registers(sc); ++ ++ pci_set_master(sc->sc_pcidev); ++ ++ /* XXX can't dynamically determine ram type for 795x; force dram */ ++ if (sc->sc_flags & HIFN_IS_7956) ++ sc->sc_drammodel = 1; ++ else if (hifn_ramtype(sc)) ++ goto fail; ++ ++ if (sc->sc_drammodel == 0) ++ hifn_sramsize(sc); ++ else ++ hifn_dramsize(sc); ++ ++ /* ++ * Workaround for NetSec 7751 rev A: half ram size because two ++ * of the address lines were left floating ++ */ ++ if (pci_get_vendor(dev) == PCI_VENDOR_NETSEC && ++ pci_get_device(dev) == PCI_PRODUCT_NETSEC_7751 && ++ pci_get_revid(dev) == 0x61) /*XXX???*/ ++ sc->sc_ramsize >>= 1; ++ ++ /* ++ * Arrange the interrupt line. ++ */ ++ rc = request_irq(dev->irq, hifn_intr, IRQF_SHARED, "hifn", sc); ++ if (rc) { ++ device_printf(sc->sc_dev, "could not map interrupt: %d\n", rc); ++ goto fail; ++ } ++ sc->sc_irq = dev->irq; ++ ++ hifn_sessions(sc); ++ ++ /* ++ * NB: Keep only the low 16 bits; this masks the chip id ++ * from the 7951. ++ */ ++ rev = READ_REG_1(sc, HIFN_1_REVID) & 0xffff; ++ ++ rseg = sc->sc_ramsize / 1024; ++ rbase = 'K'; ++ if (sc->sc_ramsize >= (1024 * 1024)) { ++ rbase = 'M'; ++ rseg /= 1024; ++ } ++ device_printf(sc->sc_dev, "%s, rev %u, %d%cB %cram", ++ hifn_partname(sc), rev, ++ rseg, rbase, sc->sc_drammodel ? 'd' : 's'); ++ if (sc->sc_flags & HIFN_IS_7956) ++ printf(", pll=0x%x<%s clk, %ux mult>", ++ sc->sc_pllconfig, ++ sc->sc_pllconfig & HIFN_PLL_REF_SEL ? "ext" : "pci", ++ 2 + 2*((sc->sc_pllconfig & HIFN_PLL_ND) >> 11)); ++ printf("\n"); ++ ++ sc->sc_cid = crypto_get_driverid(softc_get_device(sc),CRYPTOCAP_F_HARDWARE); ++ if (sc->sc_cid < 0) { ++ device_printf(sc->sc_dev, "could not get crypto driver id\n"); ++ goto fail; ++ } ++ ++ WRITE_REG_0(sc, HIFN_0_PUCNFG, ++ READ_REG_0(sc, HIFN_0_PUCNFG) | HIFN_PUCNFG_CHIPID); ++ ena = READ_REG_0(sc, HIFN_0_PUSTAT) & HIFN_PUSTAT_CHIPENA; ++ ++ switch (ena) { ++ case HIFN_PUSTAT_ENA_2: ++ crypto_register(sc->sc_cid, CRYPTO_3DES_CBC, 0, 0); ++ crypto_register(sc->sc_cid, CRYPTO_ARC4, 0, 0); ++ if (sc->sc_flags & HIFN_HAS_AES) ++ crypto_register(sc->sc_cid, CRYPTO_AES_CBC, 0, 0); ++ /*FALLTHROUGH*/ ++ case HIFN_PUSTAT_ENA_1: ++ crypto_register(sc->sc_cid, CRYPTO_MD5, 0, 0); ++ crypto_register(sc->sc_cid, CRYPTO_SHA1, 0, 0); ++ crypto_register(sc->sc_cid, CRYPTO_MD5_HMAC, 0, 0); ++ crypto_register(sc->sc_cid, CRYPTO_SHA1_HMAC, 0, 0); ++ crypto_register(sc->sc_cid, CRYPTO_DES_CBC, 0, 0); ++ break; ++ } ++ ++ if (sc->sc_flags & (HIFN_HAS_PUBLIC | HIFN_HAS_RNG)) ++ hifn_init_pubrng(sc); ++ ++ init_timer(&sc->sc_tickto); ++ sc->sc_tickto.function = hifn_tick; ++ sc->sc_tickto.data = (unsigned long) sc->sc_num; ++ mod_timer(&sc->sc_tickto, jiffies + HZ); ++ ++ return (0); ++ ++fail: ++ if (sc->sc_cid >= 0) ++ crypto_unregister_all(sc->sc_cid); ++ if (sc->sc_irq != -1) ++ free_irq(sc->sc_irq, sc); ++ if (sc->sc_dma) { ++ /* Turn off DMA polling */ ++ WRITE_REG_1(sc, HIFN_1_DMA_CNFG, HIFN_DMACNFG_MSTRESET | ++ HIFN_DMACNFG_DMARESET | HIFN_DMACNFG_MODE); ++ ++ pci_free_consistent(sc->sc_pcidev, ++ sizeof(*sc->sc_dma), ++ sc->sc_dma, sc->sc_dma_physaddr); ++ } ++ kfree(sc); ++ return (-ENXIO); ++} ++ ++/* ++ * Detach an interface that successfully probed. ++ */ ++static void ++hifn_remove(struct pci_dev *dev) ++{ ++ struct hifn_softc *sc = pci_get_drvdata(dev); ++ unsigned long l_flags; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ KASSERT(sc != NULL, ("hifn_detach: null software carrier!")); ++ ++ /* disable interrupts */ ++ HIFN_LOCK(sc); ++ WRITE_REG_1(sc, HIFN_1_DMA_IER, 0); ++ HIFN_UNLOCK(sc); ++ ++ /*XXX other resources */ ++ del_timer_sync(&sc->sc_tickto); ++ ++ /* Turn off DMA polling */ ++ WRITE_REG_1(sc, HIFN_1_DMA_CNFG, HIFN_DMACNFG_MSTRESET | ++ HIFN_DMACNFG_DMARESET | HIFN_DMACNFG_MODE); ++ ++ crypto_unregister_all(sc->sc_cid); ++ ++ free_irq(sc->sc_irq, sc); ++ ++ pci_free_consistent(sc->sc_pcidev, sizeof(*sc->sc_dma), ++ sc->sc_dma, sc->sc_dma_physaddr); ++} ++ ++ ++static int ++hifn_init_pubrng(struct hifn_softc *sc) ++{ ++ int i; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ if ((sc->sc_flags & HIFN_IS_7811) == 0) { ++ /* Reset 7951 public key/rng engine */ ++ WRITE_REG_1(sc, HIFN_1_PUB_RESET, ++ READ_REG_1(sc, HIFN_1_PUB_RESET) | HIFN_PUBRST_RESET); ++ ++ for (i = 0; i < 100; i++) { ++ DELAY(1000); ++ if ((READ_REG_1(sc, HIFN_1_PUB_RESET) & ++ HIFN_PUBRST_RESET) == 0) ++ break; ++ } ++ ++ if (i == 100) { ++ device_printf(sc->sc_dev, "public key init failed\n"); ++ return (1); ++ } ++ } ++ ++ /* Enable the rng, if available */ ++#ifdef CONFIG_OCF_RANDOMHARVEST ++ if (sc->sc_flags & HIFN_HAS_RNG) { ++ if (sc->sc_flags & HIFN_IS_7811) { ++ u_int32_t r; ++ r = READ_REG_1(sc, HIFN_1_7811_RNGENA); ++ if (r & HIFN_7811_RNGENA_ENA) { ++ r &= ~HIFN_7811_RNGENA_ENA; ++ WRITE_REG_1(sc, HIFN_1_7811_RNGENA, r); ++ } ++ WRITE_REG_1(sc, HIFN_1_7811_RNGCFG, ++ HIFN_7811_RNGCFG_DEFL); ++ r |= HIFN_7811_RNGENA_ENA; ++ WRITE_REG_1(sc, HIFN_1_7811_RNGENA, r); ++ } else ++ WRITE_REG_1(sc, HIFN_1_RNG_CONFIG, ++ READ_REG_1(sc, HIFN_1_RNG_CONFIG) | ++ HIFN_RNGCFG_ENA); ++ ++ sc->sc_rngfirst = 1; ++ crypto_rregister(sc->sc_cid, hifn_read_random, sc); ++ } ++#endif ++ ++ /* Enable public key engine, if available */ ++ if (sc->sc_flags & HIFN_HAS_PUBLIC) { ++ WRITE_REG_1(sc, HIFN_1_PUB_IEN, HIFN_PUBIEN_DONE); ++ sc->sc_dmaier |= HIFN_DMAIER_PUBDONE; ++ WRITE_REG_1(sc, HIFN_1_DMA_IER, sc->sc_dmaier); ++#ifdef HIFN_VULCANDEV ++ sc->sc_pkdev = make_dev(&vulcanpk_cdevsw, 0, ++ UID_ROOT, GID_WHEEL, 0666, ++ "vulcanpk"); ++ sc->sc_pkdev->si_drv1 = sc; ++#endif ++ } ++ ++ return (0); ++} ++ ++#ifdef CONFIG_OCF_RANDOMHARVEST ++static int ++hifn_read_random(void *arg, u_int32_t *buf, int len) ++{ ++ struct hifn_softc *sc = (struct hifn_softc *) arg; ++ u_int32_t sts; ++ int i, rc = 0; ++ ++ if (len <= 0) ++ return rc; ++ ++ if (sc->sc_flags & HIFN_IS_7811) { ++ /* ONLY VALID ON 7811!!!! */ ++ for (i = 0; i < 5; i++) { ++ sts = READ_REG_1(sc, HIFN_1_7811_RNGSTS); ++ if (sts & HIFN_7811_RNGSTS_UFL) { ++ device_printf(sc->sc_dev, ++ "RNG underflow: disabling\n"); ++ /* DAVIDM perhaps return -1 */ ++ break; ++ } ++ if ((sts & HIFN_7811_RNGSTS_RDY) == 0) ++ break; ++ ++ /* ++ * There are at least two words in the RNG FIFO ++ * at this point. ++ */ ++ if (rc < len) ++ buf[rc++] = READ_REG_1(sc, HIFN_1_7811_RNGDAT); ++ if (rc < len) ++ buf[rc++] = READ_REG_1(sc, HIFN_1_7811_RNGDAT); ++ } ++ } else ++ buf[rc++] = READ_REG_1(sc, HIFN_1_RNG_DATA); ++ ++ /* NB: discard first data read */ ++ if (sc->sc_rngfirst) { ++ sc->sc_rngfirst = 0; ++ rc = 0; ++ } ++ ++ return(rc); ++} ++#endif /* CONFIG_OCF_RANDOMHARVEST */ ++ ++static void ++hifn_puc_wait(struct hifn_softc *sc) ++{ ++ int i; ++ int reg = HIFN_0_PUCTRL; ++ ++ if (sc->sc_flags & HIFN_IS_7956) { ++ reg = HIFN_0_PUCTRL2; ++ } ++ ++ for (i = 5000; i > 0; i--) { ++ DELAY(1); ++ if (!(READ_REG_0(sc, reg) & HIFN_PUCTRL_RESET)) ++ break; ++ } ++ if (!i) ++ device_printf(sc->sc_dev, "proc unit did not reset(0x%x)\n", ++ READ_REG_0(sc, HIFN_0_PUCTRL)); ++} ++ ++/* ++ * Reset the processing unit. ++ */ ++static void ++hifn_reset_puc(struct hifn_softc *sc) ++{ ++ /* Reset processing unit */ ++ int reg = HIFN_0_PUCTRL; ++ ++ if (sc->sc_flags & HIFN_IS_7956) { ++ reg = HIFN_0_PUCTRL2; ++ } ++ WRITE_REG_0(sc, reg, HIFN_PUCTRL_DMAENA); ++ ++ hifn_puc_wait(sc); ++} ++ ++/* ++ * Set the Retry and TRDY registers; note that we set them to ++ * zero because the 7811 locks up when forced to retry (section ++ * 3.6 of "Specification Update SU-0014-04". Not clear if we ++ * should do this for all Hifn parts, but it doesn't seem to hurt. ++ */ ++static void ++hifn_set_retry(struct hifn_softc *sc) ++{ ++ DPRINTF("%s()\n", __FUNCTION__); ++ /* NB: RETRY only responds to 8-bit reads/writes */ ++ pci_write_config_byte(sc->sc_pcidev, HIFN_RETRY_TIMEOUT, 0); ++ pci_write_config_dword(sc->sc_pcidev, HIFN_TRDY_TIMEOUT, 0); ++} ++ ++/* ++ * Resets the board. Values in the regesters are left as is ++ * from the reset (i.e. initial values are assigned elsewhere). ++ */ ++static void ++hifn_reset_board(struct hifn_softc *sc, int full) ++{ ++ u_int32_t reg; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ /* ++ * Set polling in the DMA configuration register to zero. 0x7 avoids ++ * resetting the board and zeros out the other fields. ++ */ ++ WRITE_REG_1(sc, HIFN_1_DMA_CNFG, HIFN_DMACNFG_MSTRESET | ++ HIFN_DMACNFG_DMARESET | HIFN_DMACNFG_MODE); ++ ++ /* ++ * Now that polling has been disabled, we have to wait 1 ms ++ * before resetting the board. ++ */ ++ DELAY(1000); ++ ++ /* Reset the DMA unit */ ++ if (full) { ++ WRITE_REG_1(sc, HIFN_1_DMA_CNFG, HIFN_DMACNFG_MODE); ++ DELAY(1000); ++ } else { ++ WRITE_REG_1(sc, HIFN_1_DMA_CNFG, ++ HIFN_DMACNFG_MODE | HIFN_DMACNFG_MSTRESET); ++ hifn_reset_puc(sc); ++ } ++ ++ KASSERT(sc->sc_dma != NULL, ("hifn_reset_board: null DMA tag!")); ++ bzero(sc->sc_dma, sizeof(*sc->sc_dma)); ++ ++ /* Bring dma unit out of reset */ ++ WRITE_REG_1(sc, HIFN_1_DMA_CNFG, HIFN_DMACNFG_MSTRESET | ++ HIFN_DMACNFG_DMARESET | HIFN_DMACNFG_MODE); ++ ++ hifn_puc_wait(sc); ++ hifn_set_retry(sc); ++ ++ if (sc->sc_flags & HIFN_IS_7811) { ++ for (reg = 0; reg < 1000; reg++) { ++ if (READ_REG_1(sc, HIFN_1_7811_MIPSRST) & ++ HIFN_MIPSRST_CRAMINIT) ++ break; ++ DELAY(1000); ++ } ++ if (reg == 1000) ++ device_printf(sc->sc_dev, ": cram init timeout\n"); ++ } else { ++ /* set up DMA configuration register #2 */ ++ /* turn off all PK and BAR0 swaps */ ++ WRITE_REG_1(sc, HIFN_1_DMA_CNFG2, ++ (3 << HIFN_DMACNFG2_INIT_WRITE_BURST_SHIFT)| ++ (3 << HIFN_DMACNFG2_INIT_READ_BURST_SHIFT)| ++ (2 << HIFN_DMACNFG2_TGT_WRITE_BURST_SHIFT)| ++ (2 << HIFN_DMACNFG2_TGT_READ_BURST_SHIFT)); ++ } ++} ++ ++static u_int32_t ++hifn_next_signature(u_int32_t a, u_int cnt) ++{ ++ int i; ++ u_int32_t v; ++ ++ for (i = 0; i < cnt; i++) { ++ ++ /* get the parity */ ++ v = a & 0x80080125; ++ v ^= v >> 16; ++ v ^= v >> 8; ++ v ^= v >> 4; ++ v ^= v >> 2; ++ v ^= v >> 1; ++ ++ a = (v & 1) ^ (a << 1); ++ } ++ ++ return a; ++} ++ ++ ++/* ++ * Checks to see if crypto is already enabled. If crypto isn't enable, ++ * "hifn_enable_crypto" is called to enable it. The check is important, ++ * as enabling crypto twice will lock the board. ++ */ ++static int ++hifn_enable_crypto(struct hifn_softc *sc) ++{ ++ u_int32_t dmacfg, ramcfg, encl, addr, i; ++ char offtbl[] = { 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, ++ 0x00, 0x00, 0x00, 0x00 }; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ ramcfg = READ_REG_0(sc, HIFN_0_PUCNFG); ++ dmacfg = READ_REG_1(sc, HIFN_1_DMA_CNFG); ++ ++ /* ++ * The RAM config register's encrypt level bit needs to be set before ++ * every read performed on the encryption level register. ++ */ ++ WRITE_REG_0(sc, HIFN_0_PUCNFG, ramcfg | HIFN_PUCNFG_CHIPID); ++ ++ encl = READ_REG_0(sc, HIFN_0_PUSTAT) & HIFN_PUSTAT_CHIPENA; ++ ++ /* ++ * Make sure we don't re-unlock. Two unlocks kills chip until the ++ * next reboot. ++ */ ++ if (encl == HIFN_PUSTAT_ENA_1 || encl == HIFN_PUSTAT_ENA_2) { ++#ifdef HIFN_DEBUG ++ if (hifn_debug) ++ device_printf(sc->sc_dev, ++ "Strong crypto already enabled!\n"); ++#endif ++ goto report; ++ } ++ ++ if (encl != 0 && encl != HIFN_PUSTAT_ENA_0) { ++#ifdef HIFN_DEBUG ++ if (hifn_debug) ++ device_printf(sc->sc_dev, ++ "Unknown encryption level 0x%x\n", encl); ++#endif ++ return 1; ++ } ++ ++ WRITE_REG_1(sc, HIFN_1_DMA_CNFG, HIFN_DMACNFG_UNLOCK | ++ HIFN_DMACNFG_MSTRESET | HIFN_DMACNFG_DMARESET | HIFN_DMACNFG_MODE); ++ DELAY(1000); ++ addr = READ_REG_1(sc, HIFN_UNLOCK_SECRET1); ++ DELAY(1000); ++ WRITE_REG_1(sc, HIFN_UNLOCK_SECRET2, 0); ++ DELAY(1000); ++ ++ for (i = 0; i <= 12; i++) { ++ addr = hifn_next_signature(addr, offtbl[i] + 0x101); ++ WRITE_REG_1(sc, HIFN_UNLOCK_SECRET2, addr); ++ ++ DELAY(1000); ++ } ++ ++ WRITE_REG_0(sc, HIFN_0_PUCNFG, ramcfg | HIFN_PUCNFG_CHIPID); ++ encl = READ_REG_0(sc, HIFN_0_PUSTAT) & HIFN_PUSTAT_CHIPENA; ++ ++#ifdef HIFN_DEBUG ++ if (hifn_debug) { ++ if (encl != HIFN_PUSTAT_ENA_1 && encl != HIFN_PUSTAT_ENA_2) ++ device_printf(sc->sc_dev, "Engine is permanently " ++ "locked until next system reset!\n"); ++ else ++ device_printf(sc->sc_dev, "Engine enabled " ++ "successfully!\n"); ++ } ++#endif ++ ++report: ++ WRITE_REG_0(sc, HIFN_0_PUCNFG, ramcfg); ++ WRITE_REG_1(sc, HIFN_1_DMA_CNFG, dmacfg); ++ ++ switch (encl) { ++ case HIFN_PUSTAT_ENA_1: ++ case HIFN_PUSTAT_ENA_2: ++ break; ++ case HIFN_PUSTAT_ENA_0: ++ default: ++ device_printf(sc->sc_dev, "disabled\n"); ++ break; ++ } ++ ++ return 0; ++} ++ ++/* ++ * Give initial values to the registers listed in the "Register Space" ++ * section of the HIFN Software Development reference manual. ++ */ ++static void ++hifn_init_pci_registers(struct hifn_softc *sc) ++{ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ /* write fixed values needed by the Initialization registers */ ++ WRITE_REG_0(sc, HIFN_0_PUCTRL, HIFN_PUCTRL_DMAENA); ++ WRITE_REG_0(sc, HIFN_0_FIFOCNFG, HIFN_FIFOCNFG_THRESHOLD); ++ WRITE_REG_0(sc, HIFN_0_PUIER, HIFN_PUIER_DSTOVER); ++ ++ /* write all 4 ring address registers */ ++ WRITE_REG_1(sc, HIFN_1_DMA_CRAR, sc->sc_dma_physaddr + ++ offsetof(struct hifn_dma, cmdr[0])); ++ WRITE_REG_1(sc, HIFN_1_DMA_SRAR, sc->sc_dma_physaddr + ++ offsetof(struct hifn_dma, srcr[0])); ++ WRITE_REG_1(sc, HIFN_1_DMA_DRAR, sc->sc_dma_physaddr + ++ offsetof(struct hifn_dma, dstr[0])); ++ WRITE_REG_1(sc, HIFN_1_DMA_RRAR, sc->sc_dma_physaddr + ++ offsetof(struct hifn_dma, resr[0])); ++ ++ DELAY(2000); ++ ++ /* write status register */ ++ WRITE_REG_1(sc, HIFN_1_DMA_CSR, ++ HIFN_DMACSR_D_CTRL_DIS | HIFN_DMACSR_R_CTRL_DIS | ++ HIFN_DMACSR_S_CTRL_DIS | HIFN_DMACSR_C_CTRL_DIS | ++ HIFN_DMACSR_D_ABORT | HIFN_DMACSR_D_DONE | HIFN_DMACSR_D_LAST | ++ HIFN_DMACSR_D_WAIT | HIFN_DMACSR_D_OVER | ++ HIFN_DMACSR_R_ABORT | HIFN_DMACSR_R_DONE | HIFN_DMACSR_R_LAST | ++ HIFN_DMACSR_R_WAIT | HIFN_DMACSR_R_OVER | ++ HIFN_DMACSR_S_ABORT | HIFN_DMACSR_S_DONE | HIFN_DMACSR_S_LAST | ++ HIFN_DMACSR_S_WAIT | ++ HIFN_DMACSR_C_ABORT | HIFN_DMACSR_C_DONE | HIFN_DMACSR_C_LAST | ++ HIFN_DMACSR_C_WAIT | ++ HIFN_DMACSR_ENGINE | ++ ((sc->sc_flags & HIFN_HAS_PUBLIC) ? ++ HIFN_DMACSR_PUBDONE : 0) | ++ ((sc->sc_flags & HIFN_IS_7811) ? ++ HIFN_DMACSR_ILLW | HIFN_DMACSR_ILLR : 0)); ++ ++ sc->sc_d_busy = sc->sc_r_busy = sc->sc_s_busy = sc->sc_c_busy = 0; ++ sc->sc_dmaier |= HIFN_DMAIER_R_DONE | HIFN_DMAIER_C_ABORT | ++ HIFN_DMAIER_D_OVER | HIFN_DMAIER_R_OVER | ++ HIFN_DMAIER_S_ABORT | HIFN_DMAIER_D_ABORT | HIFN_DMAIER_R_ABORT | ++ ((sc->sc_flags & HIFN_IS_7811) ? ++ HIFN_DMAIER_ILLW | HIFN_DMAIER_ILLR : 0); ++ sc->sc_dmaier &= ~HIFN_DMAIER_C_WAIT; ++ WRITE_REG_1(sc, HIFN_1_DMA_IER, sc->sc_dmaier); ++ ++ ++ if (sc->sc_flags & HIFN_IS_7956) { ++ u_int32_t pll; ++ ++ WRITE_REG_0(sc, HIFN_0_PUCNFG, HIFN_PUCNFG_COMPSING | ++ HIFN_PUCNFG_TCALLPHASES | ++ HIFN_PUCNFG_TCDRVTOTEM | HIFN_PUCNFG_BUS32); ++ ++ /* turn off the clocks and insure bypass is set */ ++ pll = READ_REG_1(sc, HIFN_1_PLL); ++ pll = (pll &~ (HIFN_PLL_PK_CLK_SEL | HIFN_PLL_PE_CLK_SEL)) ++ | HIFN_PLL_BP | HIFN_PLL_MBSET; ++ WRITE_REG_1(sc, HIFN_1_PLL, pll); ++ DELAY(10*1000); /* 10ms */ ++ ++ /* change configuration */ ++ pll = (pll &~ HIFN_PLL_CONFIG) | sc->sc_pllconfig; ++ WRITE_REG_1(sc, HIFN_1_PLL, pll); ++ DELAY(10*1000); /* 10ms */ ++ ++ /* disable bypass */ ++ pll &= ~HIFN_PLL_BP; ++ WRITE_REG_1(sc, HIFN_1_PLL, pll); ++ /* enable clocks with new configuration */ ++ pll |= HIFN_PLL_PK_CLK_SEL | HIFN_PLL_PE_CLK_SEL; ++ WRITE_REG_1(sc, HIFN_1_PLL, pll); ++ } else { ++ WRITE_REG_0(sc, HIFN_0_PUCNFG, HIFN_PUCNFG_COMPSING | ++ HIFN_PUCNFG_DRFR_128 | HIFN_PUCNFG_TCALLPHASES | ++ HIFN_PUCNFG_TCDRVTOTEM | HIFN_PUCNFG_BUS32 | ++ (sc->sc_drammodel ? HIFN_PUCNFG_DRAM : HIFN_PUCNFG_SRAM)); ++ } ++ ++ WRITE_REG_0(sc, HIFN_0_PUISR, HIFN_PUISR_DSTOVER); ++ WRITE_REG_1(sc, HIFN_1_DMA_CNFG, HIFN_DMACNFG_MSTRESET | ++ HIFN_DMACNFG_DMARESET | HIFN_DMACNFG_MODE | HIFN_DMACNFG_LAST | ++ ((HIFN_POLL_FREQUENCY << 16 ) & HIFN_DMACNFG_POLLFREQ) | ++ ((HIFN_POLL_SCALAR << 8) & HIFN_DMACNFG_POLLINVAL)); ++} ++ ++/* ++ * The maximum number of sessions supported by the card ++ * is dependent on the amount of context ram, which ++ * encryption algorithms are enabled, and how compression ++ * is configured. This should be configured before this ++ * routine is called. ++ */ ++static void ++hifn_sessions(struct hifn_softc *sc) ++{ ++ u_int32_t pucnfg; ++ int ctxsize; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ pucnfg = READ_REG_0(sc, HIFN_0_PUCNFG); ++ ++ if (pucnfg & HIFN_PUCNFG_COMPSING) { ++ if (pucnfg & HIFN_PUCNFG_ENCCNFG) ++ ctxsize = 128; ++ else ++ ctxsize = 512; ++ /* ++ * 7955/7956 has internal context memory of 32K ++ */ ++ if (sc->sc_flags & HIFN_IS_7956) ++ sc->sc_maxses = 32768 / ctxsize; ++ else ++ sc->sc_maxses = 1 + ++ ((sc->sc_ramsize - 32768) / ctxsize); ++ } else ++ sc->sc_maxses = sc->sc_ramsize / 16384; ++ ++ if (sc->sc_maxses > 2048) ++ sc->sc_maxses = 2048; ++} ++ ++/* ++ * Determine ram type (sram or dram). Board should be just out of a reset ++ * state when this is called. ++ */ ++static int ++hifn_ramtype(struct hifn_softc *sc) ++{ ++ u_int8_t data[8], dataexpect[8]; ++ int i; ++ ++ for (i = 0; i < sizeof(data); i++) ++ data[i] = dataexpect[i] = 0x55; ++ if (hifn_writeramaddr(sc, 0, data)) ++ return (-1); ++ if (hifn_readramaddr(sc, 0, data)) ++ return (-1); ++ if (bcmp(data, dataexpect, sizeof(data)) != 0) { ++ sc->sc_drammodel = 1; ++ return (0); ++ } ++ ++ for (i = 0; i < sizeof(data); i++) ++ data[i] = dataexpect[i] = 0xaa; ++ if (hifn_writeramaddr(sc, 0, data)) ++ return (-1); ++ if (hifn_readramaddr(sc, 0, data)) ++ return (-1); ++ if (bcmp(data, dataexpect, sizeof(data)) != 0) { ++ sc->sc_drammodel = 1; ++ return (0); ++ } ++ ++ return (0); ++} ++ ++#define HIFN_SRAM_MAX (32 << 20) ++#define HIFN_SRAM_STEP_SIZE 16384 ++#define HIFN_SRAM_GRANULARITY (HIFN_SRAM_MAX / HIFN_SRAM_STEP_SIZE) ++ ++static int ++hifn_sramsize(struct hifn_softc *sc) ++{ ++ u_int32_t a; ++ u_int8_t data[8]; ++ u_int8_t dataexpect[sizeof(data)]; ++ int32_t i; ++ ++ for (i = 0; i < sizeof(data); i++) ++ data[i] = dataexpect[i] = i ^ 0x5a; ++ ++ for (i = HIFN_SRAM_GRANULARITY - 1; i >= 0; i--) { ++ a = i * HIFN_SRAM_STEP_SIZE; ++ bcopy(&i, data, sizeof(i)); ++ hifn_writeramaddr(sc, a, data); ++ } ++ ++ for (i = 0; i < HIFN_SRAM_GRANULARITY; i++) { ++ a = i * HIFN_SRAM_STEP_SIZE; ++ bcopy(&i, dataexpect, sizeof(i)); ++ if (hifn_readramaddr(sc, a, data) < 0) ++ return (0); ++ if (bcmp(data, dataexpect, sizeof(data)) != 0) ++ return (0); ++ sc->sc_ramsize = a + HIFN_SRAM_STEP_SIZE; ++ } ++ ++ return (0); ++} ++ ++/* ++ * XXX For dram boards, one should really try all of the ++ * HIFN_PUCNFG_DSZ_*'s. This just assumes that PUCNFG ++ * is already set up correctly. ++ */ ++static int ++hifn_dramsize(struct hifn_softc *sc) ++{ ++ u_int32_t cnfg; ++ ++ if (sc->sc_flags & HIFN_IS_7956) { ++ /* ++ * 7955/7956 have a fixed internal ram of only 32K. ++ */ ++ sc->sc_ramsize = 32768; ++ } else { ++ cnfg = READ_REG_0(sc, HIFN_0_PUCNFG) & ++ HIFN_PUCNFG_DRAMMASK; ++ sc->sc_ramsize = 1 << ((cnfg >> 13) + 18); ++ } ++ return (0); ++} ++ ++static void ++hifn_alloc_slot(struct hifn_softc *sc, int *cmdp, int *srcp, int *dstp, int *resp) ++{ ++ struct hifn_dma *dma = sc->sc_dma; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ if (dma->cmdi == HIFN_D_CMD_RSIZE) { ++ dma->cmdi = 0; ++ dma->cmdr[HIFN_D_CMD_RSIZE].l = htole32(HIFN_D_JUMP|HIFN_D_MASKDONEIRQ); ++ wmb(); ++ dma->cmdr[HIFN_D_CMD_RSIZE].l |= htole32(HIFN_D_VALID); ++ HIFN_CMDR_SYNC(sc, HIFN_D_CMD_RSIZE, ++ BUS_DMASYNC_PREWRITE | BUS_DMASYNC_PREREAD); ++ } ++ *cmdp = dma->cmdi++; ++ dma->cmdk = dma->cmdi; ++ ++ if (dma->srci == HIFN_D_SRC_RSIZE) { ++ dma->srci = 0; ++ dma->srcr[HIFN_D_SRC_RSIZE].l = htole32(HIFN_D_JUMP|HIFN_D_MASKDONEIRQ); ++ wmb(); ++ dma->srcr[HIFN_D_SRC_RSIZE].l |= htole32(HIFN_D_VALID); ++ HIFN_SRCR_SYNC(sc, HIFN_D_SRC_RSIZE, ++ BUS_DMASYNC_PREWRITE | BUS_DMASYNC_PREREAD); ++ } ++ *srcp = dma->srci++; ++ dma->srck = dma->srci; ++ ++ if (dma->dsti == HIFN_D_DST_RSIZE) { ++ dma->dsti = 0; ++ dma->dstr[HIFN_D_DST_RSIZE].l = htole32(HIFN_D_JUMP|HIFN_D_MASKDONEIRQ); ++ wmb(); ++ dma->dstr[HIFN_D_DST_RSIZE].l |= htole32(HIFN_D_VALID); ++ HIFN_DSTR_SYNC(sc, HIFN_D_DST_RSIZE, ++ BUS_DMASYNC_PREWRITE | BUS_DMASYNC_PREREAD); ++ } ++ *dstp = dma->dsti++; ++ dma->dstk = dma->dsti; ++ ++ if (dma->resi == HIFN_D_RES_RSIZE) { ++ dma->resi = 0; ++ dma->resr[HIFN_D_RES_RSIZE].l = htole32(HIFN_D_JUMP|HIFN_D_MASKDONEIRQ); ++ wmb(); ++ dma->resr[HIFN_D_RES_RSIZE].l |= htole32(HIFN_D_VALID); ++ HIFN_RESR_SYNC(sc, HIFN_D_RES_RSIZE, ++ BUS_DMASYNC_PREWRITE | BUS_DMASYNC_PREREAD); ++ } ++ *resp = dma->resi++; ++ dma->resk = dma->resi; ++} ++ ++static int ++hifn_writeramaddr(struct hifn_softc *sc, int addr, u_int8_t *data) ++{ ++ struct hifn_dma *dma = sc->sc_dma; ++ hifn_base_command_t wc; ++ const u_int32_t masks = HIFN_D_VALID | HIFN_D_LAST | HIFN_D_MASKDONEIRQ; ++ int r, cmdi, resi, srci, dsti; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ wc.masks = htole16(3 << 13); ++ wc.session_num = htole16(addr >> 14); ++ wc.total_source_count = htole16(8); ++ wc.total_dest_count = htole16(addr & 0x3fff); ++ ++ hifn_alloc_slot(sc, &cmdi, &srci, &dsti, &resi); ++ ++ WRITE_REG_1(sc, HIFN_1_DMA_CSR, ++ HIFN_DMACSR_C_CTRL_ENA | HIFN_DMACSR_S_CTRL_ENA | ++ HIFN_DMACSR_D_CTRL_ENA | HIFN_DMACSR_R_CTRL_ENA); ++ ++ /* build write command */ ++ bzero(dma->command_bufs[cmdi], HIFN_MAX_COMMAND); ++ *(hifn_base_command_t *)dma->command_bufs[cmdi] = wc; ++ bcopy(data, &dma->test_src, sizeof(dma->test_src)); ++ ++ dma->srcr[srci].p = htole32(sc->sc_dma_physaddr ++ + offsetof(struct hifn_dma, test_src)); ++ dma->dstr[dsti].p = htole32(sc->sc_dma_physaddr ++ + offsetof(struct hifn_dma, test_dst)); ++ ++ dma->cmdr[cmdi].l = htole32(16 | masks); ++ dma->srcr[srci].l = htole32(8 | masks); ++ dma->dstr[dsti].l = htole32(4 | masks); ++ dma->resr[resi].l = htole32(4 | masks); ++ ++ for (r = 10000; r >= 0; r--) { ++ DELAY(10); ++ if ((dma->resr[resi].l & htole32(HIFN_D_VALID)) == 0) ++ break; ++ } ++ if (r == 0) { ++ device_printf(sc->sc_dev, "writeramaddr -- " ++ "result[%d](addr %d) still valid\n", resi, addr); ++ r = -1; ++ return (-1); ++ } else ++ r = 0; ++ ++ WRITE_REG_1(sc, HIFN_1_DMA_CSR, ++ HIFN_DMACSR_C_CTRL_DIS | HIFN_DMACSR_S_CTRL_DIS | ++ HIFN_DMACSR_D_CTRL_DIS | HIFN_DMACSR_R_CTRL_DIS); ++ ++ return (r); ++} ++ ++static int ++hifn_readramaddr(struct hifn_softc *sc, int addr, u_int8_t *data) ++{ ++ struct hifn_dma *dma = sc->sc_dma; ++ hifn_base_command_t rc; ++ const u_int32_t masks = HIFN_D_VALID | HIFN_D_LAST | HIFN_D_MASKDONEIRQ; ++ int r, cmdi, srci, dsti, resi; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ rc.masks = htole16(2 << 13); ++ rc.session_num = htole16(addr >> 14); ++ rc.total_source_count = htole16(addr & 0x3fff); ++ rc.total_dest_count = htole16(8); ++ ++ hifn_alloc_slot(sc, &cmdi, &srci, &dsti, &resi); ++ ++ WRITE_REG_1(sc, HIFN_1_DMA_CSR, ++ HIFN_DMACSR_C_CTRL_ENA | HIFN_DMACSR_S_CTRL_ENA | ++ HIFN_DMACSR_D_CTRL_ENA | HIFN_DMACSR_R_CTRL_ENA); ++ ++ bzero(dma->command_bufs[cmdi], HIFN_MAX_COMMAND); ++ *(hifn_base_command_t *)dma->command_bufs[cmdi] = rc; ++ ++ dma->srcr[srci].p = htole32(sc->sc_dma_physaddr + ++ offsetof(struct hifn_dma, test_src)); ++ dma->test_src = 0; ++ dma->dstr[dsti].p = htole32(sc->sc_dma_physaddr + ++ offsetof(struct hifn_dma, test_dst)); ++ dma->test_dst = 0; ++ dma->cmdr[cmdi].l = htole32(8 | masks); ++ dma->srcr[srci].l = htole32(8 | masks); ++ dma->dstr[dsti].l = htole32(8 | masks); ++ dma->resr[resi].l = htole32(HIFN_MAX_RESULT | masks); ++ ++ for (r = 10000; r >= 0; r--) { ++ DELAY(10); ++ if ((dma->resr[resi].l & htole32(HIFN_D_VALID)) == 0) ++ break; ++ } ++ if (r == 0) { ++ device_printf(sc->sc_dev, "readramaddr -- " ++ "result[%d](addr %d) still valid\n", resi, addr); ++ r = -1; ++ } else { ++ r = 0; ++ bcopy(&dma->test_dst, data, sizeof(dma->test_dst)); ++ } ++ ++ WRITE_REG_1(sc, HIFN_1_DMA_CSR, ++ HIFN_DMACSR_C_CTRL_DIS | HIFN_DMACSR_S_CTRL_DIS | ++ HIFN_DMACSR_D_CTRL_DIS | HIFN_DMACSR_R_CTRL_DIS); ++ ++ return (r); ++} ++ ++/* ++ * Initialize the descriptor rings. ++ */ ++static void ++hifn_init_dma(struct hifn_softc *sc) ++{ ++ struct hifn_dma *dma = sc->sc_dma; ++ int i; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ hifn_set_retry(sc); ++ ++ /* initialize static pointer values */ ++ for (i = 0; i < HIFN_D_CMD_RSIZE; i++) ++ dma->cmdr[i].p = htole32(sc->sc_dma_physaddr + ++ offsetof(struct hifn_dma, command_bufs[i][0])); ++ for (i = 0; i < HIFN_D_RES_RSIZE; i++) ++ dma->resr[i].p = htole32(sc->sc_dma_physaddr + ++ offsetof(struct hifn_dma, result_bufs[i][0])); ++ ++ dma->cmdr[HIFN_D_CMD_RSIZE].p = ++ htole32(sc->sc_dma_physaddr + offsetof(struct hifn_dma, cmdr[0])); ++ dma->srcr[HIFN_D_SRC_RSIZE].p = ++ htole32(sc->sc_dma_physaddr + offsetof(struct hifn_dma, srcr[0])); ++ dma->dstr[HIFN_D_DST_RSIZE].p = ++ htole32(sc->sc_dma_physaddr + offsetof(struct hifn_dma, dstr[0])); ++ dma->resr[HIFN_D_RES_RSIZE].p = ++ htole32(sc->sc_dma_physaddr + offsetof(struct hifn_dma, resr[0])); ++ ++ dma->cmdu = dma->srcu = dma->dstu = dma->resu = 0; ++ dma->cmdi = dma->srci = dma->dsti = dma->resi = 0; ++ dma->cmdk = dma->srck = dma->dstk = dma->resk = 0; ++} ++ ++/* ++ * Writes out the raw command buffer space. Returns the ++ * command buffer size. ++ */ ++static u_int ++hifn_write_command(struct hifn_command *cmd, u_int8_t *buf) ++{ ++ struct hifn_softc *sc = NULL; ++ u_int8_t *buf_pos; ++ hifn_base_command_t *base_cmd; ++ hifn_mac_command_t *mac_cmd; ++ hifn_crypt_command_t *cry_cmd; ++ int using_mac, using_crypt, len, ivlen; ++ u_int32_t dlen, slen; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ buf_pos = buf; ++ using_mac = cmd->base_masks & HIFN_BASE_CMD_MAC; ++ using_crypt = cmd->base_masks & HIFN_BASE_CMD_CRYPT; ++ ++ base_cmd = (hifn_base_command_t *)buf_pos; ++ base_cmd->masks = htole16(cmd->base_masks); ++ slen = cmd->src_mapsize; ++ if (cmd->sloplen) ++ dlen = cmd->dst_mapsize - cmd->sloplen + sizeof(u_int32_t); ++ else ++ dlen = cmd->dst_mapsize; ++ base_cmd->total_source_count = htole16(slen & HIFN_BASE_CMD_LENMASK_LO); ++ base_cmd->total_dest_count = htole16(dlen & HIFN_BASE_CMD_LENMASK_LO); ++ dlen >>= 16; ++ slen >>= 16; ++ base_cmd->session_num = htole16( ++ ((slen << HIFN_BASE_CMD_SRCLEN_S) & HIFN_BASE_CMD_SRCLEN_M) | ++ ((dlen << HIFN_BASE_CMD_DSTLEN_S) & HIFN_BASE_CMD_DSTLEN_M)); ++ buf_pos += sizeof(hifn_base_command_t); ++ ++ if (using_mac) { ++ mac_cmd = (hifn_mac_command_t *)buf_pos; ++ dlen = cmd->maccrd->crd_len; ++ mac_cmd->source_count = htole16(dlen & 0xffff); ++ dlen >>= 16; ++ mac_cmd->masks = htole16(cmd->mac_masks | ++ ((dlen << HIFN_MAC_CMD_SRCLEN_S) & HIFN_MAC_CMD_SRCLEN_M)); ++ mac_cmd->header_skip = htole16(cmd->maccrd->crd_skip); ++ mac_cmd->reserved = 0; ++ buf_pos += sizeof(hifn_mac_command_t); ++ } ++ ++ if (using_crypt) { ++ cry_cmd = (hifn_crypt_command_t *)buf_pos; ++ dlen = cmd->enccrd->crd_len; ++ cry_cmd->source_count = htole16(dlen & 0xffff); ++ dlen >>= 16; ++ cry_cmd->masks = htole16(cmd->cry_masks | ++ ((dlen << HIFN_CRYPT_CMD_SRCLEN_S) & HIFN_CRYPT_CMD_SRCLEN_M)); ++ cry_cmd->header_skip = htole16(cmd->enccrd->crd_skip); ++ cry_cmd->reserved = 0; ++ buf_pos += sizeof(hifn_crypt_command_t); ++ } ++ ++ if (using_mac && cmd->mac_masks & HIFN_MAC_CMD_NEW_KEY) { ++ bcopy(cmd->mac, buf_pos, HIFN_MAC_KEY_LENGTH); ++ buf_pos += HIFN_MAC_KEY_LENGTH; ++ } ++ ++ if (using_crypt && cmd->cry_masks & HIFN_CRYPT_CMD_NEW_KEY) { ++ switch (cmd->cry_masks & HIFN_CRYPT_CMD_ALG_MASK) { ++ case HIFN_CRYPT_CMD_ALG_3DES: ++ bcopy(cmd->ck, buf_pos, HIFN_3DES_KEY_LENGTH); ++ buf_pos += HIFN_3DES_KEY_LENGTH; ++ break; ++ case HIFN_CRYPT_CMD_ALG_DES: ++ bcopy(cmd->ck, buf_pos, HIFN_DES_KEY_LENGTH); ++ buf_pos += HIFN_DES_KEY_LENGTH; ++ break; ++ case HIFN_CRYPT_CMD_ALG_RC4: ++ len = 256; ++ do { ++ int clen; ++ ++ clen = MIN(cmd->cklen, len); ++ bcopy(cmd->ck, buf_pos, clen); ++ len -= clen; ++ buf_pos += clen; ++ } while (len > 0); ++ bzero(buf_pos, 4); ++ buf_pos += 4; ++ break; ++ case HIFN_CRYPT_CMD_ALG_AES: ++ /* ++ * AES keys are variable 128, 192 and ++ * 256 bits (16, 24 and 32 bytes). ++ */ ++ bcopy(cmd->ck, buf_pos, cmd->cklen); ++ buf_pos += cmd->cklen; ++ break; ++ } ++ } ++ ++ if (using_crypt && cmd->cry_masks & HIFN_CRYPT_CMD_NEW_IV) { ++ switch (cmd->cry_masks & HIFN_CRYPT_CMD_ALG_MASK) { ++ case HIFN_CRYPT_CMD_ALG_AES: ++ ivlen = HIFN_AES_IV_LENGTH; ++ break; ++ default: ++ ivlen = HIFN_IV_LENGTH; ++ break; ++ } ++ bcopy(cmd->iv, buf_pos, ivlen); ++ buf_pos += ivlen; ++ } ++ ++ if ((cmd->base_masks & (HIFN_BASE_CMD_MAC|HIFN_BASE_CMD_CRYPT)) == 0) { ++ bzero(buf_pos, 8); ++ buf_pos += 8; ++ } ++ ++ return (buf_pos - buf); ++} ++ ++static int ++hifn_dmamap_aligned(struct hifn_operand *op) ++{ ++ struct hifn_softc *sc = NULL; ++ int i; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ for (i = 0; i < op->nsegs; i++) { ++ if (op->segs[i].ds_addr & 3) ++ return (0); ++ if ((i != (op->nsegs - 1)) && (op->segs[i].ds_len & 3)) ++ return (0); ++ } ++ return (1); ++} ++ ++static __inline int ++hifn_dmamap_dstwrap(struct hifn_softc *sc, int idx) ++{ ++ struct hifn_dma *dma = sc->sc_dma; ++ ++ if (++idx == HIFN_D_DST_RSIZE) { ++ dma->dstr[idx].l = htole32(HIFN_D_VALID | HIFN_D_JUMP | ++ HIFN_D_MASKDONEIRQ); ++ HIFN_DSTR_SYNC(sc, idx, ++ BUS_DMASYNC_PREREAD | BUS_DMASYNC_PREWRITE); ++ idx = 0; ++ } ++ return (idx); ++} ++ ++static int ++hifn_dmamap_load_dst(struct hifn_softc *sc, struct hifn_command *cmd) ++{ ++ struct hifn_dma *dma = sc->sc_dma; ++ struct hifn_operand *dst = &cmd->dst; ++ u_int32_t p, l; ++ int idx, used = 0, i; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ idx = dma->dsti; ++ for (i = 0; i < dst->nsegs - 1; i++) { ++ dma->dstr[idx].p = htole32(dst->segs[i].ds_addr); ++ dma->dstr[idx].l = htole32(HIFN_D_MASKDONEIRQ | dst->segs[i].ds_len); ++ wmb(); ++ dma->dstr[idx].l |= htole32(HIFN_D_VALID); ++ HIFN_DSTR_SYNC(sc, idx, ++ BUS_DMASYNC_PREREAD | BUS_DMASYNC_PREWRITE); ++ used++; ++ ++ idx = hifn_dmamap_dstwrap(sc, idx); ++ } ++ ++ if (cmd->sloplen == 0) { ++ p = dst->segs[i].ds_addr; ++ l = HIFN_D_MASKDONEIRQ | HIFN_D_LAST | ++ dst->segs[i].ds_len; ++ } else { ++ p = sc->sc_dma_physaddr + ++ offsetof(struct hifn_dma, slop[cmd->slopidx]); ++ l = HIFN_D_MASKDONEIRQ | HIFN_D_LAST | ++ sizeof(u_int32_t); ++ ++ if ((dst->segs[i].ds_len - cmd->sloplen) != 0) { ++ dma->dstr[idx].p = htole32(dst->segs[i].ds_addr); ++ dma->dstr[idx].l = htole32(HIFN_D_MASKDONEIRQ | ++ (dst->segs[i].ds_len - cmd->sloplen)); ++ wmb(); ++ dma->dstr[idx].l |= htole32(HIFN_D_VALID); ++ HIFN_DSTR_SYNC(sc, idx, ++ BUS_DMASYNC_PREREAD | BUS_DMASYNC_PREWRITE); ++ used++; ++ ++ idx = hifn_dmamap_dstwrap(sc, idx); ++ } ++ } ++ dma->dstr[idx].p = htole32(p); ++ dma->dstr[idx].l = htole32(l); ++ wmb(); ++ dma->dstr[idx].l |= htole32(HIFN_D_VALID); ++ HIFN_DSTR_SYNC(sc, idx, BUS_DMASYNC_PREREAD | BUS_DMASYNC_PREWRITE); ++ used++; ++ ++ idx = hifn_dmamap_dstwrap(sc, idx); ++ ++ dma->dsti = idx; ++ dma->dstu += used; ++ return (idx); ++} ++ ++static __inline int ++hifn_dmamap_srcwrap(struct hifn_softc *sc, int idx) ++{ ++ struct hifn_dma *dma = sc->sc_dma; ++ ++ if (++idx == HIFN_D_SRC_RSIZE) { ++ dma->srcr[idx].l = htole32(HIFN_D_VALID | ++ HIFN_D_JUMP | HIFN_D_MASKDONEIRQ); ++ HIFN_SRCR_SYNC(sc, HIFN_D_SRC_RSIZE, ++ BUS_DMASYNC_PREWRITE | BUS_DMASYNC_PREREAD); ++ idx = 0; ++ } ++ return (idx); ++} ++ ++static int ++hifn_dmamap_load_src(struct hifn_softc *sc, struct hifn_command *cmd) ++{ ++ struct hifn_dma *dma = sc->sc_dma; ++ struct hifn_operand *src = &cmd->src; ++ int idx, i; ++ u_int32_t last = 0; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ idx = dma->srci; ++ for (i = 0; i < src->nsegs; i++) { ++ if (i == src->nsegs - 1) ++ last = HIFN_D_LAST; ++ ++ dma->srcr[idx].p = htole32(src->segs[i].ds_addr); ++ dma->srcr[idx].l = htole32(src->segs[i].ds_len | ++ HIFN_D_MASKDONEIRQ | last); ++ wmb(); ++ dma->srcr[idx].l |= htole32(HIFN_D_VALID); ++ HIFN_SRCR_SYNC(sc, idx, ++ BUS_DMASYNC_PREWRITE | BUS_DMASYNC_PREREAD); ++ ++ idx = hifn_dmamap_srcwrap(sc, idx); ++ } ++ dma->srci = idx; ++ dma->srcu += src->nsegs; ++ return (idx); ++} ++ ++ ++static int ++hifn_crypto( ++ struct hifn_softc *sc, ++ struct hifn_command *cmd, ++ struct cryptop *crp, ++ int hint) ++{ ++ struct hifn_dma *dma = sc->sc_dma; ++ u_int32_t cmdlen, csr; ++ int cmdi, resi, err = 0; ++ unsigned long l_flags; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ /* ++ * need 1 cmd, and 1 res ++ * ++ * NB: check this first since it's easy. ++ */ ++ HIFN_LOCK(sc); ++ if ((dma->cmdu + 1) > HIFN_D_CMD_RSIZE || ++ (dma->resu + 1) > HIFN_D_RES_RSIZE) { ++#ifdef HIFN_DEBUG ++ if (hifn_debug) { ++ device_printf(sc->sc_dev, ++ "cmd/result exhaustion, cmdu %u resu %u\n", ++ dma->cmdu, dma->resu); ++ } ++#endif ++ hifnstats.hst_nomem_cr++; ++ sc->sc_needwakeup |= CRYPTO_SYMQ; ++ HIFN_UNLOCK(sc); ++ return (ERESTART); ++ } ++ ++ if (crp->crp_flags & CRYPTO_F_SKBUF) { ++ if (pci_map_skb(sc, &cmd->src, cmd->src_skb)) { ++ hifnstats.hst_nomem_load++; ++ err = ENOMEM; ++ goto err_srcmap1; ++ } ++ } else if (crp->crp_flags & CRYPTO_F_IOV) { ++ if (pci_map_uio(sc, &cmd->src, cmd->src_io)) { ++ hifnstats.hst_nomem_load++; ++ err = ENOMEM; ++ goto err_srcmap1; ++ } ++ } else { ++ if (pci_map_buf(sc, &cmd->src, cmd->src_buf, crp->crp_ilen)) { ++ hifnstats.hst_nomem_load++; ++ err = ENOMEM; ++ goto err_srcmap1; ++ } ++ } ++ ++ if (hifn_dmamap_aligned(&cmd->src)) { ++ cmd->sloplen = cmd->src_mapsize & 3; ++ cmd->dst = cmd->src; ++ } else { ++ if (crp->crp_flags & CRYPTO_F_IOV) { ++ DPRINTF("%s,%d: %s - EINVAL\n",__FILE__,__LINE__,__FUNCTION__); ++ err = EINVAL; ++ goto err_srcmap; ++ } else if (crp->crp_flags & CRYPTO_F_SKBUF) { ++#ifdef NOTYET ++ int totlen, len; ++ struct mbuf *m, *m0, *mlast; ++ ++ KASSERT(cmd->dst_m == cmd->src_m, ++ ("hifn_crypto: dst_m initialized improperly")); ++ hifnstats.hst_unaligned++; ++ /* ++ * Source is not aligned on a longword boundary. ++ * Copy the data to insure alignment. If we fail ++ * to allocate mbufs or clusters while doing this ++ * we return ERESTART so the operation is requeued ++ * at the crypto later, but only if there are ++ * ops already posted to the hardware; otherwise we ++ * have no guarantee that we'll be re-entered. ++ */ ++ totlen = cmd->src_mapsize; ++ if (cmd->src_m->m_flags & M_PKTHDR) { ++ len = MHLEN; ++ MGETHDR(m0, M_DONTWAIT, MT_DATA); ++ if (m0 && !m_dup_pkthdr(m0, cmd->src_m, M_DONTWAIT)) { ++ m_free(m0); ++ m0 = NULL; ++ } ++ } else { ++ len = MLEN; ++ MGET(m0, M_DONTWAIT, MT_DATA); ++ } ++ if (m0 == NULL) { ++ hifnstats.hst_nomem_mbuf++; ++ err = dma->cmdu ? ERESTART : ENOMEM; ++ goto err_srcmap; ++ } ++ if (totlen >= MINCLSIZE) { ++ MCLGET(m0, M_DONTWAIT); ++ if ((m0->m_flags & M_EXT) == 0) { ++ hifnstats.hst_nomem_mcl++; ++ err = dma->cmdu ? ERESTART : ENOMEM; ++ m_freem(m0); ++ goto err_srcmap; ++ } ++ len = MCLBYTES; ++ } ++ totlen -= len; ++ m0->m_pkthdr.len = m0->m_len = len; ++ mlast = m0; ++ ++ while (totlen > 0) { ++ MGET(m, M_DONTWAIT, MT_DATA); ++ if (m == NULL) { ++ hifnstats.hst_nomem_mbuf++; ++ err = dma->cmdu ? ERESTART : ENOMEM; ++ m_freem(m0); ++ goto err_srcmap; ++ } ++ len = MLEN; ++ if (totlen >= MINCLSIZE) { ++ MCLGET(m, M_DONTWAIT); ++ if ((m->m_flags & M_EXT) == 0) { ++ hifnstats.hst_nomem_mcl++; ++ err = dma->cmdu ? ERESTART : ENOMEM; ++ mlast->m_next = m; ++ m_freem(m0); ++ goto err_srcmap; ++ } ++ len = MCLBYTES; ++ } ++ ++ m->m_len = len; ++ m0->m_pkthdr.len += len; ++ totlen -= len; ++ ++ mlast->m_next = m; ++ mlast = m; ++ } ++ cmd->dst_m = m0; ++#else ++ device_printf(sc->sc_dev, ++ "%s,%d: CRYPTO_F_SKBUF unaligned not implemented\n", ++ __FILE__, __LINE__); ++ err = EINVAL; ++ goto err_srcmap; ++#endif ++ } else { ++ device_printf(sc->sc_dev, ++ "%s,%d: unaligned contig buffers not implemented\n", ++ __FILE__, __LINE__); ++ err = EINVAL; ++ goto err_srcmap; ++ } ++ } ++ ++ if (cmd->dst_map == NULL) { ++ if (crp->crp_flags & CRYPTO_F_SKBUF) { ++ if (pci_map_skb(sc, &cmd->dst, cmd->dst_skb)) { ++ hifnstats.hst_nomem_map++; ++ err = ENOMEM; ++ goto err_dstmap1; ++ } ++ } else if (crp->crp_flags & CRYPTO_F_IOV) { ++ if (pci_map_uio(sc, &cmd->dst, cmd->dst_io)) { ++ hifnstats.hst_nomem_load++; ++ err = ENOMEM; ++ goto err_dstmap1; ++ } ++ } else { ++ if (pci_map_buf(sc, &cmd->dst, cmd->dst_buf, crp->crp_ilen)) { ++ hifnstats.hst_nomem_load++; ++ err = ENOMEM; ++ goto err_dstmap1; ++ } ++ } ++ } ++ ++#ifdef HIFN_DEBUG ++ if (hifn_debug) { ++ device_printf(sc->sc_dev, ++ "Entering cmd: stat %8x ien %8x u %d/%d/%d/%d n %d/%d\n", ++ READ_REG_1(sc, HIFN_1_DMA_CSR), ++ READ_REG_1(sc, HIFN_1_DMA_IER), ++ dma->cmdu, dma->srcu, dma->dstu, dma->resu, ++ cmd->src_nsegs, cmd->dst_nsegs); ++ } ++#endif ++ ++#if 0 ++ if (cmd->src_map == cmd->dst_map) { ++ bus_dmamap_sync(sc->sc_dmat, cmd->src_map, ++ BUS_DMASYNC_PREWRITE|BUS_DMASYNC_PREREAD); ++ } else { ++ bus_dmamap_sync(sc->sc_dmat, cmd->src_map, ++ BUS_DMASYNC_PREWRITE); ++ bus_dmamap_sync(sc->sc_dmat, cmd->dst_map, ++ BUS_DMASYNC_PREREAD); ++ } ++#endif ++ ++ /* ++ * need N src, and N dst ++ */ ++ if ((dma->srcu + cmd->src_nsegs) > HIFN_D_SRC_RSIZE || ++ (dma->dstu + cmd->dst_nsegs + 1) > HIFN_D_DST_RSIZE) { ++#ifdef HIFN_DEBUG ++ if (hifn_debug) { ++ device_printf(sc->sc_dev, ++ "src/dst exhaustion, srcu %u+%u dstu %u+%u\n", ++ dma->srcu, cmd->src_nsegs, ++ dma->dstu, cmd->dst_nsegs); ++ } ++#endif ++ hifnstats.hst_nomem_sd++; ++ err = ERESTART; ++ goto err_dstmap; ++ } ++ ++ if (dma->cmdi == HIFN_D_CMD_RSIZE) { ++ dma->cmdi = 0; ++ dma->cmdr[HIFN_D_CMD_RSIZE].l = htole32(HIFN_D_JUMP|HIFN_D_MASKDONEIRQ); ++ wmb(); ++ dma->cmdr[HIFN_D_CMD_RSIZE].l |= htole32(HIFN_D_VALID); ++ HIFN_CMDR_SYNC(sc, HIFN_D_CMD_RSIZE, ++ BUS_DMASYNC_PREWRITE | BUS_DMASYNC_PREREAD); ++ } ++ cmdi = dma->cmdi++; ++ cmdlen = hifn_write_command(cmd, dma->command_bufs[cmdi]); ++ HIFN_CMD_SYNC(sc, cmdi, BUS_DMASYNC_PREWRITE); ++ ++ /* .p for command/result already set */ ++ dma->cmdr[cmdi].l = htole32(cmdlen | HIFN_D_LAST | ++ HIFN_D_MASKDONEIRQ); ++ wmb(); ++ dma->cmdr[cmdi].l |= htole32(HIFN_D_VALID); ++ HIFN_CMDR_SYNC(sc, cmdi, ++ BUS_DMASYNC_PREWRITE | BUS_DMASYNC_PREREAD); ++ dma->cmdu++; ++ ++ /* ++ * We don't worry about missing an interrupt (which a "command wait" ++ * interrupt salvages us from), unless there is more than one command ++ * in the queue. ++ */ ++ if (dma->cmdu > 1) { ++ sc->sc_dmaier |= HIFN_DMAIER_C_WAIT; ++ WRITE_REG_1(sc, HIFN_1_DMA_IER, sc->sc_dmaier); ++ } ++ ++ hifnstats.hst_ipackets++; ++ hifnstats.hst_ibytes += cmd->src_mapsize; ++ ++ hifn_dmamap_load_src(sc, cmd); ++ ++ /* ++ * Unlike other descriptors, we don't mask done interrupt from ++ * result descriptor. ++ */ ++#ifdef HIFN_DEBUG ++ if (hifn_debug) ++ device_printf(sc->sc_dev, "load res\n"); ++#endif ++ if (dma->resi == HIFN_D_RES_RSIZE) { ++ dma->resi = 0; ++ dma->resr[HIFN_D_RES_RSIZE].l = htole32(HIFN_D_JUMP|HIFN_D_MASKDONEIRQ); ++ wmb(); ++ dma->resr[HIFN_D_RES_RSIZE].l |= htole32(HIFN_D_VALID); ++ HIFN_RESR_SYNC(sc, HIFN_D_RES_RSIZE, ++ BUS_DMASYNC_PREREAD | BUS_DMASYNC_PREWRITE); ++ } ++ resi = dma->resi++; ++ KASSERT(dma->hifn_commands[resi] == NULL, ++ ("hifn_crypto: command slot %u busy", resi)); ++ dma->hifn_commands[resi] = cmd; ++ HIFN_RES_SYNC(sc, resi, BUS_DMASYNC_PREREAD); ++ if ((hint & CRYPTO_HINT_MORE) && sc->sc_curbatch < hifn_maxbatch) { ++ dma->resr[resi].l = htole32(HIFN_MAX_RESULT | ++ HIFN_D_LAST | HIFN_D_MASKDONEIRQ); ++ wmb(); ++ dma->resr[resi].l |= htole32(HIFN_D_VALID); ++ sc->sc_curbatch++; ++ if (sc->sc_curbatch > hifnstats.hst_maxbatch) ++ hifnstats.hst_maxbatch = sc->sc_curbatch; ++ hifnstats.hst_totbatch++; ++ } else { ++ dma->resr[resi].l = htole32(HIFN_MAX_RESULT | HIFN_D_LAST); ++ wmb(); ++ dma->resr[resi].l |= htole32(HIFN_D_VALID); ++ sc->sc_curbatch = 0; ++ } ++ HIFN_RESR_SYNC(sc, resi, ++ BUS_DMASYNC_PREREAD | BUS_DMASYNC_PREWRITE); ++ dma->resu++; ++ ++ if (cmd->sloplen) ++ cmd->slopidx = resi; ++ ++ hifn_dmamap_load_dst(sc, cmd); ++ ++ csr = 0; ++ if (sc->sc_c_busy == 0) { ++ csr |= HIFN_DMACSR_C_CTRL_ENA; ++ sc->sc_c_busy = 1; ++ } ++ if (sc->sc_s_busy == 0) { ++ csr |= HIFN_DMACSR_S_CTRL_ENA; ++ sc->sc_s_busy = 1; ++ } ++ if (sc->sc_r_busy == 0) { ++ csr |= HIFN_DMACSR_R_CTRL_ENA; ++ sc->sc_r_busy = 1; ++ } ++ if (sc->sc_d_busy == 0) { ++ csr |= HIFN_DMACSR_D_CTRL_ENA; ++ sc->sc_d_busy = 1; ++ } ++ if (csr) ++ WRITE_REG_1(sc, HIFN_1_DMA_CSR, csr); ++ ++#ifdef HIFN_DEBUG ++ if (hifn_debug) { ++ device_printf(sc->sc_dev, "command: stat %8x ier %8x\n", ++ READ_REG_1(sc, HIFN_1_DMA_CSR), ++ READ_REG_1(sc, HIFN_1_DMA_IER)); ++ } ++#endif ++ ++ sc->sc_active = 5; ++ HIFN_UNLOCK(sc); ++ KASSERT(err == 0, ("hifn_crypto: success with error %u", err)); ++ return (err); /* success */ ++ ++err_dstmap: ++ if (cmd->src_map != cmd->dst_map) ++ pci_unmap_buf(sc, &cmd->dst); ++err_dstmap1: ++err_srcmap: ++ if (crp->crp_flags & CRYPTO_F_SKBUF) { ++ if (cmd->src_skb != cmd->dst_skb) ++#ifdef NOTYET ++ m_freem(cmd->dst_m); ++#else ++ device_printf(sc->sc_dev, ++ "%s,%d: CRYPTO_F_SKBUF src != dst not implemented\n", ++ __FILE__, __LINE__); ++#endif ++ } ++ pci_unmap_buf(sc, &cmd->src); ++err_srcmap1: ++ HIFN_UNLOCK(sc); ++ return (err); ++} ++ ++static void ++hifn_tick(unsigned long arg) ++{ ++ struct hifn_softc *sc; ++ unsigned long l_flags; ++ ++ if (arg >= HIFN_MAX_CHIPS) ++ return; ++ sc = hifn_chip_idx[arg]; ++ if (!sc) ++ return; ++ ++ HIFN_LOCK(sc); ++ if (sc->sc_active == 0) { ++ struct hifn_dma *dma = sc->sc_dma; ++ u_int32_t r = 0; ++ ++ if (dma->cmdu == 0 && sc->sc_c_busy) { ++ sc->sc_c_busy = 0; ++ r |= HIFN_DMACSR_C_CTRL_DIS; ++ } ++ if (dma->srcu == 0 && sc->sc_s_busy) { ++ sc->sc_s_busy = 0; ++ r |= HIFN_DMACSR_S_CTRL_DIS; ++ } ++ if (dma->dstu == 0 && sc->sc_d_busy) { ++ sc->sc_d_busy = 0; ++ r |= HIFN_DMACSR_D_CTRL_DIS; ++ } ++ if (dma->resu == 0 && sc->sc_r_busy) { ++ sc->sc_r_busy = 0; ++ r |= HIFN_DMACSR_R_CTRL_DIS; ++ } ++ if (r) ++ WRITE_REG_1(sc, HIFN_1_DMA_CSR, r); ++ } else ++ sc->sc_active--; ++ HIFN_UNLOCK(sc); ++ mod_timer(&sc->sc_tickto, jiffies + HZ); ++} ++ ++static irqreturn_t ++#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,19) ++hifn_intr(int irq, void *arg) ++#else ++hifn_intr(int irq, void *arg, struct pt_regs *regs) ++#endif ++{ ++ struct hifn_softc *sc = arg; ++ struct hifn_dma *dma; ++ u_int32_t dmacsr, restart; ++ int i, u; ++ unsigned long l_flags; ++ ++ dmacsr = READ_REG_1(sc, HIFN_1_DMA_CSR); ++ ++ /* Nothing in the DMA unit interrupted */ ++ if ((dmacsr & sc->sc_dmaier) == 0) ++ return IRQ_NONE; ++ ++ HIFN_LOCK(sc); ++ ++ dma = sc->sc_dma; ++ ++#ifdef HIFN_DEBUG ++ if (hifn_debug) { ++ device_printf(sc->sc_dev, ++ "irq: stat %08x ien %08x damier %08x i %d/%d/%d/%d k %d/%d/%d/%d u %d/%d/%d/%d\n", ++ dmacsr, READ_REG_1(sc, HIFN_1_DMA_IER), sc->sc_dmaier, ++ dma->cmdi, dma->srci, dma->dsti, dma->resi, ++ dma->cmdk, dma->srck, dma->dstk, dma->resk, ++ dma->cmdu, dma->srcu, dma->dstu, dma->resu); ++ } ++#endif ++ ++ WRITE_REG_1(sc, HIFN_1_DMA_CSR, dmacsr & sc->sc_dmaier); ++ ++ if ((sc->sc_flags & HIFN_HAS_PUBLIC) && ++ (dmacsr & HIFN_DMACSR_PUBDONE)) ++ WRITE_REG_1(sc, HIFN_1_PUB_STATUS, ++ READ_REG_1(sc, HIFN_1_PUB_STATUS) | HIFN_PUBSTS_DONE); ++ ++ restart = dmacsr & (HIFN_DMACSR_D_OVER | HIFN_DMACSR_R_OVER); ++ if (restart) ++ device_printf(sc->sc_dev, "overrun %x\n", dmacsr); ++ ++ if (sc->sc_flags & HIFN_IS_7811) { ++ if (dmacsr & HIFN_DMACSR_ILLR) ++ device_printf(sc->sc_dev, "illegal read\n"); ++ if (dmacsr & HIFN_DMACSR_ILLW) ++ device_printf(sc->sc_dev, "illegal write\n"); ++ } ++ ++ restart = dmacsr & (HIFN_DMACSR_C_ABORT | HIFN_DMACSR_S_ABORT | ++ HIFN_DMACSR_D_ABORT | HIFN_DMACSR_R_ABORT); ++ if (restart) { ++ device_printf(sc->sc_dev, "abort, resetting.\n"); ++ hifnstats.hst_abort++; ++ hifn_abort(sc); ++ HIFN_UNLOCK(sc); ++ return IRQ_HANDLED; ++ } ++ ++ if ((dmacsr & HIFN_DMACSR_C_WAIT) && (dma->cmdu == 0)) { ++ /* ++ * If no slots to process and we receive a "waiting on ++ * command" interrupt, we disable the "waiting on command" ++ * (by clearing it). ++ */ ++ sc->sc_dmaier &= ~HIFN_DMAIER_C_WAIT; ++ WRITE_REG_1(sc, HIFN_1_DMA_IER, sc->sc_dmaier); ++ } ++ ++ /* clear the rings */ ++ i = dma->resk; u = dma->resu; ++ while (u != 0) { ++ HIFN_RESR_SYNC(sc, i, ++ BUS_DMASYNC_POSTREAD | BUS_DMASYNC_POSTWRITE); ++ if (dma->resr[i].l & htole32(HIFN_D_VALID)) { ++ HIFN_RESR_SYNC(sc, i, ++ BUS_DMASYNC_PREREAD | BUS_DMASYNC_PREWRITE); ++ break; ++ } ++ ++ if (i != HIFN_D_RES_RSIZE) { ++ struct hifn_command *cmd; ++ u_int8_t *macbuf = NULL; ++ ++ HIFN_RES_SYNC(sc, i, BUS_DMASYNC_POSTREAD); ++ cmd = dma->hifn_commands[i]; ++ KASSERT(cmd != NULL, ++ ("hifn_intr: null command slot %u", i)); ++ dma->hifn_commands[i] = NULL; ++ ++ if (cmd->base_masks & HIFN_BASE_CMD_MAC) { ++ macbuf = dma->result_bufs[i]; ++ macbuf += 12; ++ } ++ ++ hifn_callback(sc, cmd, macbuf); ++ hifnstats.hst_opackets++; ++ u--; ++ } ++ ++ if (++i == (HIFN_D_RES_RSIZE + 1)) ++ i = 0; ++ } ++ dma->resk = i; dma->resu = u; ++ ++ i = dma->srck; u = dma->srcu; ++ while (u != 0) { ++ if (i == HIFN_D_SRC_RSIZE) ++ i = 0; ++ HIFN_SRCR_SYNC(sc, i, ++ BUS_DMASYNC_POSTREAD | BUS_DMASYNC_POSTWRITE); ++ if (dma->srcr[i].l & htole32(HIFN_D_VALID)) { ++ HIFN_SRCR_SYNC(sc, i, ++ BUS_DMASYNC_PREREAD | BUS_DMASYNC_PREWRITE); ++ break; ++ } ++ i++, u--; ++ } ++ dma->srck = i; dma->srcu = u; ++ ++ i = dma->cmdk; u = dma->cmdu; ++ while (u != 0) { ++ HIFN_CMDR_SYNC(sc, i, ++ BUS_DMASYNC_POSTREAD | BUS_DMASYNC_POSTWRITE); ++ if (dma->cmdr[i].l & htole32(HIFN_D_VALID)) { ++ HIFN_CMDR_SYNC(sc, i, ++ BUS_DMASYNC_PREREAD | BUS_DMASYNC_PREWRITE); ++ break; ++ } ++ if (i != HIFN_D_CMD_RSIZE) { ++ u--; ++ HIFN_CMD_SYNC(sc, i, BUS_DMASYNC_POSTWRITE); ++ } ++ if (++i == (HIFN_D_CMD_RSIZE + 1)) ++ i = 0; ++ } ++ dma->cmdk = i; dma->cmdu = u; ++ ++ HIFN_UNLOCK(sc); ++ ++ if (sc->sc_needwakeup) { /* XXX check high watermark */ ++ int wakeup = sc->sc_needwakeup & (CRYPTO_SYMQ|CRYPTO_ASYMQ); ++#ifdef HIFN_DEBUG ++ if (hifn_debug) ++ device_printf(sc->sc_dev, ++ "wakeup crypto (%x) u %d/%d/%d/%d\n", ++ sc->sc_needwakeup, ++ dma->cmdu, dma->srcu, dma->dstu, dma->resu); ++#endif ++ sc->sc_needwakeup &= ~wakeup; ++ crypto_unblock(sc->sc_cid, wakeup); ++ } ++ ++ return IRQ_HANDLED; ++} ++ ++/* ++ * Allocate a new 'session' and return an encoded session id. 'sidp' ++ * contains our registration id, and should contain an encoded session ++ * id on successful allocation. ++ */ ++static int ++hifn_newsession(device_t dev, u_int32_t *sidp, struct cryptoini *cri) ++{ ++ struct hifn_softc *sc = device_get_softc(dev); ++ struct cryptoini *c; ++ int mac = 0, cry = 0, sesn; ++ struct hifn_session *ses = NULL; ++ unsigned long l_flags; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ KASSERT(sc != NULL, ("hifn_newsession: null softc")); ++ if (sidp == NULL || cri == NULL || sc == NULL) { ++ DPRINTF("%s,%d: %s - EINVAL\n", __FILE__, __LINE__, __FUNCTION__); ++ return (EINVAL); ++ } ++ ++ HIFN_LOCK(sc); ++ if (sc->sc_sessions == NULL) { ++ ses = sc->sc_sessions = (struct hifn_session *)kmalloc(sizeof(*ses), ++ SLAB_ATOMIC); ++ if (ses == NULL) { ++ HIFN_UNLOCK(sc); ++ return (ENOMEM); ++ } ++ sesn = 0; ++ sc->sc_nsessions = 1; ++ } else { ++ for (sesn = 0; sesn < sc->sc_nsessions; sesn++) { ++ if (!sc->sc_sessions[sesn].hs_used) { ++ ses = &sc->sc_sessions[sesn]; ++ break; ++ } ++ } ++ ++ if (ses == NULL) { ++ sesn = sc->sc_nsessions; ++ ses = (struct hifn_session *)kmalloc((sesn + 1) * sizeof(*ses), ++ SLAB_ATOMIC); ++ if (ses == NULL) { ++ HIFN_UNLOCK(sc); ++ return (ENOMEM); ++ } ++ bcopy(sc->sc_sessions, ses, sesn * sizeof(*ses)); ++ bzero(sc->sc_sessions, sesn * sizeof(*ses)); ++ kfree(sc->sc_sessions); ++ sc->sc_sessions = ses; ++ ses = &sc->sc_sessions[sesn]; ++ sc->sc_nsessions++; ++ } ++ } ++ HIFN_UNLOCK(sc); ++ ++ bzero(ses, sizeof(*ses)); ++ ses->hs_used = 1; ++ ++ for (c = cri; c != NULL; c = c->cri_next) { ++ switch (c->cri_alg) { ++ case CRYPTO_MD5: ++ case CRYPTO_SHA1: ++ case CRYPTO_MD5_HMAC: ++ case CRYPTO_SHA1_HMAC: ++ if (mac) { ++ DPRINTF("%s,%d: %s - EINVAL\n",__FILE__,__LINE__,__FUNCTION__); ++ return (EINVAL); ++ } ++ mac = 1; ++ ses->hs_mlen = c->cri_mlen; ++ if (ses->hs_mlen == 0) { ++ switch (c->cri_alg) { ++ case CRYPTO_MD5: ++ case CRYPTO_MD5_HMAC: ++ ses->hs_mlen = 16; ++ break; ++ case CRYPTO_SHA1: ++ case CRYPTO_SHA1_HMAC: ++ ses->hs_mlen = 20; ++ break; ++ } ++ } ++ break; ++ case CRYPTO_DES_CBC: ++ case CRYPTO_3DES_CBC: ++ case CRYPTO_AES_CBC: ++ /* XXX this may read fewer, does it matter? */ ++ read_random(ses->hs_iv, ++ c->cri_alg == CRYPTO_AES_CBC ? ++ HIFN_AES_IV_LENGTH : HIFN_IV_LENGTH); ++ /*FALLTHROUGH*/ ++ case CRYPTO_ARC4: ++ if (cry) { ++ DPRINTF("%s,%d: %s - EINVAL\n",__FILE__,__LINE__,__FUNCTION__); ++ return (EINVAL); ++ } ++ cry = 1; ++ break; ++ default: ++ DPRINTF("%s,%d: %s - EINVAL\n",__FILE__,__LINE__,__FUNCTION__); ++ return (EINVAL); ++ } ++ } ++ if (mac == 0 && cry == 0) { ++ DPRINTF("%s,%d: %s - EINVAL\n",__FILE__,__LINE__,__FUNCTION__); ++ return (EINVAL); ++ } ++ ++ *sidp = HIFN_SID(device_get_unit(sc->sc_dev), sesn); ++ ++ return (0); ++} ++ ++/* ++ * Deallocate a session. ++ * XXX this routine should run a zero'd mac/encrypt key into context ram. ++ * XXX to blow away any keys already stored there. ++ */ ++static int ++hifn_freesession(device_t dev, u_int64_t tid) ++{ ++ struct hifn_softc *sc = device_get_softc(dev); ++ int session, error; ++ u_int32_t sid = CRYPTO_SESID2LID(tid); ++ unsigned long l_flags; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ KASSERT(sc != NULL, ("hifn_freesession: null softc")); ++ if (sc == NULL) { ++ DPRINTF("%s,%d: %s - EINVAL\n",__FILE__,__LINE__,__FUNCTION__); ++ return (EINVAL); ++ } ++ ++ HIFN_LOCK(sc); ++ session = HIFN_SESSION(sid); ++ if (session < sc->sc_nsessions) { ++ bzero(&sc->sc_sessions[session], sizeof(struct hifn_session)); ++ error = 0; ++ } else { ++ DPRINTF("%s,%d: %s - EINVAL\n",__FILE__,__LINE__,__FUNCTION__); ++ error = EINVAL; ++ } ++ HIFN_UNLOCK(sc); ++ ++ return (error); ++} ++ ++static int ++hifn_process(device_t dev, struct cryptop *crp, int hint) ++{ ++ struct hifn_softc *sc = device_get_softc(dev); ++ struct hifn_command *cmd = NULL; ++ int session, err, ivlen; ++ struct cryptodesc *crd1, *crd2, *maccrd, *enccrd; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ if (crp == NULL || crp->crp_callback == NULL) { ++ hifnstats.hst_invalid++; ++ DPRINTF("%s,%d: %s - EINVAL\n",__FILE__,__LINE__,__FUNCTION__); ++ return (EINVAL); ++ } ++ session = HIFN_SESSION(crp->crp_sid); ++ ++ if (sc == NULL || session >= sc->sc_nsessions) { ++ DPRINTF("%s,%d: %s - EINVAL\n",__FILE__,__LINE__,__FUNCTION__); ++ err = EINVAL; ++ goto errout; ++ } ++ ++ cmd = kmalloc(sizeof(struct hifn_command), SLAB_ATOMIC); ++ if (cmd == NULL) { ++ hifnstats.hst_nomem++; ++ err = ENOMEM; ++ goto errout; ++ } ++ memset(cmd, 0, sizeof(*cmd)); ++ ++ if (crp->crp_flags & CRYPTO_F_SKBUF) { ++ cmd->src_skb = (struct sk_buff *)crp->crp_buf; ++ cmd->dst_skb = (struct sk_buff *)crp->crp_buf; ++ } else if (crp->crp_flags & CRYPTO_F_IOV) { ++ cmd->src_io = (struct uio *)crp->crp_buf; ++ cmd->dst_io = (struct uio *)crp->crp_buf; ++ } else { ++ cmd->src_buf = crp->crp_buf; ++ cmd->dst_buf = crp->crp_buf; ++ } ++ ++ crd1 = crp->crp_desc; ++ if (crd1 == NULL) { ++ DPRINTF("%s,%d: %s - EINVAL\n",__FILE__,__LINE__,__FUNCTION__); ++ err = EINVAL; ++ goto errout; ++ } ++ crd2 = crd1->crd_next; ++ ++ if (crd2 == NULL) { ++ if (crd1->crd_alg == CRYPTO_MD5_HMAC || ++ crd1->crd_alg == CRYPTO_SHA1_HMAC || ++ crd1->crd_alg == CRYPTO_SHA1 || ++ crd1->crd_alg == CRYPTO_MD5) { ++ maccrd = crd1; ++ enccrd = NULL; ++ } else if (crd1->crd_alg == CRYPTO_DES_CBC || ++ crd1->crd_alg == CRYPTO_3DES_CBC || ++ crd1->crd_alg == CRYPTO_AES_CBC || ++ crd1->crd_alg == CRYPTO_ARC4) { ++ if ((crd1->crd_flags & CRD_F_ENCRYPT) == 0) ++ cmd->base_masks |= HIFN_BASE_CMD_DECODE; ++ maccrd = NULL; ++ enccrd = crd1; ++ } else { ++ DPRINTF("%s,%d: %s - EINVAL\n",__FILE__,__LINE__,__FUNCTION__); ++ err = EINVAL; ++ goto errout; ++ } ++ } else { ++ if ((crd1->crd_alg == CRYPTO_MD5_HMAC || ++ crd1->crd_alg == CRYPTO_SHA1_HMAC || ++ crd1->crd_alg == CRYPTO_MD5 || ++ crd1->crd_alg == CRYPTO_SHA1) && ++ (crd2->crd_alg == CRYPTO_DES_CBC || ++ crd2->crd_alg == CRYPTO_3DES_CBC || ++ crd2->crd_alg == CRYPTO_AES_CBC || ++ crd2->crd_alg == CRYPTO_ARC4) && ++ ((crd2->crd_flags & CRD_F_ENCRYPT) == 0)) { ++ cmd->base_masks = HIFN_BASE_CMD_DECODE; ++ maccrd = crd1; ++ enccrd = crd2; ++ } else if ((crd1->crd_alg == CRYPTO_DES_CBC || ++ crd1->crd_alg == CRYPTO_ARC4 || ++ crd1->crd_alg == CRYPTO_3DES_CBC || ++ crd1->crd_alg == CRYPTO_AES_CBC) && ++ (crd2->crd_alg == CRYPTO_MD5_HMAC || ++ crd2->crd_alg == CRYPTO_SHA1_HMAC || ++ crd2->crd_alg == CRYPTO_MD5 || ++ crd2->crd_alg == CRYPTO_SHA1) && ++ (crd1->crd_flags & CRD_F_ENCRYPT)) { ++ enccrd = crd1; ++ maccrd = crd2; ++ } else { ++ /* ++ * We cannot order the 7751 as requested ++ */ ++ DPRINTF("%s,%d: %s %d,%d,%d - EINVAL\n",__FILE__,__LINE__,__FUNCTION__, crd1->crd_alg, crd2->crd_alg, crd1->crd_flags & CRD_F_ENCRYPT); ++ err = EINVAL; ++ goto errout; ++ } ++ } ++ ++ if (enccrd) { ++ cmd->enccrd = enccrd; ++ cmd->base_masks |= HIFN_BASE_CMD_CRYPT; ++ switch (enccrd->crd_alg) { ++ case CRYPTO_ARC4: ++ cmd->cry_masks |= HIFN_CRYPT_CMD_ALG_RC4; ++ break; ++ case CRYPTO_DES_CBC: ++ cmd->cry_masks |= HIFN_CRYPT_CMD_ALG_DES | ++ HIFN_CRYPT_CMD_MODE_CBC | ++ HIFN_CRYPT_CMD_NEW_IV; ++ break; ++ case CRYPTO_3DES_CBC: ++ cmd->cry_masks |= HIFN_CRYPT_CMD_ALG_3DES | ++ HIFN_CRYPT_CMD_MODE_CBC | ++ HIFN_CRYPT_CMD_NEW_IV; ++ break; ++ case CRYPTO_AES_CBC: ++ cmd->cry_masks |= HIFN_CRYPT_CMD_ALG_AES | ++ HIFN_CRYPT_CMD_MODE_CBC | ++ HIFN_CRYPT_CMD_NEW_IV; ++ break; ++ default: ++ DPRINTF("%s,%d: %s - EINVAL\n",__FILE__,__LINE__,__FUNCTION__); ++ err = EINVAL; ++ goto errout; ++ } ++ if (enccrd->crd_alg != CRYPTO_ARC4) { ++ ivlen = ((enccrd->crd_alg == CRYPTO_AES_CBC) ? ++ HIFN_AES_IV_LENGTH : HIFN_IV_LENGTH); ++ if (enccrd->crd_flags & CRD_F_ENCRYPT) { ++ if (enccrd->crd_flags & CRD_F_IV_EXPLICIT) ++ bcopy(enccrd->crd_iv, cmd->iv, ivlen); ++ else ++ bcopy(sc->sc_sessions[session].hs_iv, ++ cmd->iv, ivlen); ++ ++ if ((enccrd->crd_flags & CRD_F_IV_PRESENT) ++ == 0) { ++ crypto_copyback(crp->crp_flags, ++ crp->crp_buf, enccrd->crd_inject, ++ ivlen, cmd->iv); ++ } ++ } else { ++ if (enccrd->crd_flags & CRD_F_IV_EXPLICIT) ++ bcopy(enccrd->crd_iv, cmd->iv, ivlen); ++ else { ++ crypto_copydata(crp->crp_flags, ++ crp->crp_buf, enccrd->crd_inject, ++ ivlen, cmd->iv); ++ } ++ } ++ } ++ ++ if (enccrd->crd_flags & CRD_F_KEY_EXPLICIT) ++ cmd->cry_masks |= HIFN_CRYPT_CMD_NEW_KEY; ++ cmd->ck = enccrd->crd_key; ++ cmd->cklen = enccrd->crd_klen >> 3; ++ cmd->cry_masks |= HIFN_CRYPT_CMD_NEW_KEY; ++ ++ /* ++ * Need to specify the size for the AES key in the masks. ++ */ ++ if ((cmd->cry_masks & HIFN_CRYPT_CMD_ALG_MASK) == ++ HIFN_CRYPT_CMD_ALG_AES) { ++ switch (cmd->cklen) { ++ case 16: ++ cmd->cry_masks |= HIFN_CRYPT_CMD_KSZ_128; ++ break; ++ case 24: ++ cmd->cry_masks |= HIFN_CRYPT_CMD_KSZ_192; ++ break; ++ case 32: ++ cmd->cry_masks |= HIFN_CRYPT_CMD_KSZ_256; ++ break; ++ default: ++ DPRINTF("%s,%d: %s - EINVAL\n",__FILE__,__LINE__,__FUNCTION__); ++ err = EINVAL; ++ goto errout; ++ } ++ } ++ } ++ ++ if (maccrd) { ++ cmd->maccrd = maccrd; ++ cmd->base_masks |= HIFN_BASE_CMD_MAC; ++ ++ switch (maccrd->crd_alg) { ++ case CRYPTO_MD5: ++ cmd->mac_masks |= HIFN_MAC_CMD_ALG_MD5 | ++ HIFN_MAC_CMD_RESULT | HIFN_MAC_CMD_MODE_HASH | ++ HIFN_MAC_CMD_POS_IPSEC; ++ break; ++ case CRYPTO_MD5_HMAC: ++ cmd->mac_masks |= HIFN_MAC_CMD_ALG_MD5 | ++ HIFN_MAC_CMD_RESULT | HIFN_MAC_CMD_MODE_HMAC | ++ HIFN_MAC_CMD_POS_IPSEC | HIFN_MAC_CMD_TRUNC; ++ break; ++ case CRYPTO_SHA1: ++ cmd->mac_masks |= HIFN_MAC_CMD_ALG_SHA1 | ++ HIFN_MAC_CMD_RESULT | HIFN_MAC_CMD_MODE_HASH | ++ HIFN_MAC_CMD_POS_IPSEC; ++ break; ++ case CRYPTO_SHA1_HMAC: ++ cmd->mac_masks |= HIFN_MAC_CMD_ALG_SHA1 | ++ HIFN_MAC_CMD_RESULT | HIFN_MAC_CMD_MODE_HMAC | ++ HIFN_MAC_CMD_POS_IPSEC | HIFN_MAC_CMD_TRUNC; ++ break; ++ } ++ ++ if (maccrd->crd_alg == CRYPTO_SHA1_HMAC || ++ maccrd->crd_alg == CRYPTO_MD5_HMAC) { ++ cmd->mac_masks |= HIFN_MAC_CMD_NEW_KEY; ++ bcopy(maccrd->crd_key, cmd->mac, maccrd->crd_klen >> 3); ++ bzero(cmd->mac + (maccrd->crd_klen >> 3), ++ HIFN_MAC_KEY_LENGTH - (maccrd->crd_klen >> 3)); ++ } ++ } ++ ++ cmd->crp = crp; ++ cmd->session_num = session; ++ cmd->softc = sc; ++ ++ err = hifn_crypto(sc, cmd, crp, hint); ++ if (!err) { ++ return 0; ++ } else if (err == ERESTART) { ++ /* ++ * There weren't enough resources to dispatch the request ++ * to the part. Notify the caller so they'll requeue this ++ * request and resubmit it again soon. ++ */ ++#ifdef HIFN_DEBUG ++ if (hifn_debug) ++ device_printf(sc->sc_dev, "requeue request\n"); ++#endif ++ kfree(cmd); ++ sc->sc_needwakeup |= CRYPTO_SYMQ; ++ return (err); ++ } ++ ++errout: ++ if (cmd != NULL) ++ kfree(cmd); ++ if (err == EINVAL) ++ hifnstats.hst_invalid++; ++ else ++ hifnstats.hst_nomem++; ++ crp->crp_etype = err; ++ crypto_done(crp); ++ return (err); ++} ++ ++static void ++hifn_abort(struct hifn_softc *sc) ++{ ++ struct hifn_dma *dma = sc->sc_dma; ++ struct hifn_command *cmd; ++ struct cryptop *crp; ++ int i, u; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ i = dma->resk; u = dma->resu; ++ while (u != 0) { ++ cmd = dma->hifn_commands[i]; ++ KASSERT(cmd != NULL, ("hifn_abort: null command slot %u", i)); ++ dma->hifn_commands[i] = NULL; ++ crp = cmd->crp; ++ ++ if ((dma->resr[i].l & htole32(HIFN_D_VALID)) == 0) { ++ /* Salvage what we can. */ ++ u_int8_t *macbuf; ++ ++ if (cmd->base_masks & HIFN_BASE_CMD_MAC) { ++ macbuf = dma->result_bufs[i]; ++ macbuf += 12; ++ } else ++ macbuf = NULL; ++ hifnstats.hst_opackets++; ++ hifn_callback(sc, cmd, macbuf); ++ } else { ++#if 0 ++ if (cmd->src_map == cmd->dst_map) { ++ bus_dmamap_sync(sc->sc_dmat, cmd->src_map, ++ BUS_DMASYNC_POSTREAD|BUS_DMASYNC_POSTWRITE); ++ } else { ++ bus_dmamap_sync(sc->sc_dmat, cmd->src_map, ++ BUS_DMASYNC_POSTWRITE); ++ bus_dmamap_sync(sc->sc_dmat, cmd->dst_map, ++ BUS_DMASYNC_POSTREAD); ++ } ++#endif ++ ++ if (cmd->src_skb != cmd->dst_skb) { ++#ifdef NOTYET ++ m_freem(cmd->src_m); ++ crp->crp_buf = (caddr_t)cmd->dst_m; ++#else ++ device_printf(sc->sc_dev, ++ "%s,%d: CRYPTO_F_SKBUF src != dst not implemented\n", ++ __FILE__, __LINE__); ++#endif ++ } ++ ++ /* non-shared buffers cannot be restarted */ ++ if (cmd->src_map != cmd->dst_map) { ++ /* ++ * XXX should be EAGAIN, delayed until ++ * after the reset. ++ */ ++ crp->crp_etype = ENOMEM; ++ pci_unmap_buf(sc, &cmd->dst); ++ } else ++ crp->crp_etype = ENOMEM; ++ ++ pci_unmap_buf(sc, &cmd->src); ++ ++ kfree(cmd); ++ if (crp->crp_etype != EAGAIN) ++ crypto_done(crp); ++ } ++ ++ if (++i == HIFN_D_RES_RSIZE) ++ i = 0; ++ u--; ++ } ++ dma->resk = i; dma->resu = u; ++ ++ hifn_reset_board(sc, 1); ++ hifn_init_dma(sc); ++ hifn_init_pci_registers(sc); ++} ++ ++static void ++hifn_callback(struct hifn_softc *sc, struct hifn_command *cmd, u_int8_t *macbuf) ++{ ++ struct hifn_dma *dma = sc->sc_dma; ++ struct cryptop *crp = cmd->crp; ++ struct cryptodesc *crd; ++ int i, u, ivlen; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++#if 0 ++ if (cmd->src_map == cmd->dst_map) { ++ bus_dmamap_sync(sc->sc_dmat, cmd->src_map, ++ BUS_DMASYNC_POSTWRITE | BUS_DMASYNC_POSTREAD); ++ } else { ++ bus_dmamap_sync(sc->sc_dmat, cmd->src_map, ++ BUS_DMASYNC_POSTWRITE); ++ bus_dmamap_sync(sc->sc_dmat, cmd->dst_map, ++ BUS_DMASYNC_POSTREAD); ++ } ++#endif ++ ++ if (crp->crp_flags & CRYPTO_F_SKBUF) { ++ if (cmd->src_skb != cmd->dst_skb) { ++#ifdef NOTYET ++ crp->crp_buf = (caddr_t)cmd->dst_m; ++ totlen = cmd->src_mapsize; ++ for (m = cmd->dst_m; m != NULL; m = m->m_next) { ++ if (totlen < m->m_len) { ++ m->m_len = totlen; ++ totlen = 0; ++ } else ++ totlen -= m->m_len; ++ } ++ cmd->dst_m->m_pkthdr.len = cmd->src_m->m_pkthdr.len; ++ m_freem(cmd->src_m); ++#else ++ device_printf(sc->sc_dev, ++ "%s,%d: CRYPTO_F_SKBUF src != dst not implemented\n", ++ __FILE__, __LINE__); ++#endif ++ } ++ } ++ ++ if (cmd->sloplen != 0) { ++ crypto_copyback(crp->crp_flags, crp->crp_buf, ++ cmd->src_mapsize - cmd->sloplen, cmd->sloplen, ++ (caddr_t)&dma->slop[cmd->slopidx]); ++ } ++ ++ i = dma->dstk; u = dma->dstu; ++ while (u != 0) { ++ if (i == HIFN_D_DST_RSIZE) ++ i = 0; ++#if 0 ++ bus_dmamap_sync(sc->sc_dmat, sc->sc_dmamap, ++ BUS_DMASYNC_POSTREAD | BUS_DMASYNC_POSTWRITE); ++#endif ++ if (dma->dstr[i].l & htole32(HIFN_D_VALID)) { ++#if 0 ++ bus_dmamap_sync(sc->sc_dmat, sc->sc_dmamap, ++ BUS_DMASYNC_PREREAD | BUS_DMASYNC_PREWRITE); ++#endif ++ break; ++ } ++ i++, u--; ++ } ++ dma->dstk = i; dma->dstu = u; ++ ++ hifnstats.hst_obytes += cmd->dst_mapsize; ++ ++ if ((cmd->base_masks & (HIFN_BASE_CMD_CRYPT | HIFN_BASE_CMD_DECODE)) == ++ HIFN_BASE_CMD_CRYPT) { ++ for (crd = crp->crp_desc; crd; crd = crd->crd_next) { ++ if (crd->crd_alg != CRYPTO_DES_CBC && ++ crd->crd_alg != CRYPTO_3DES_CBC && ++ crd->crd_alg != CRYPTO_AES_CBC) ++ continue; ++ ivlen = ((crd->crd_alg == CRYPTO_AES_CBC) ? ++ HIFN_AES_IV_LENGTH : HIFN_IV_LENGTH); ++ crypto_copydata(crp->crp_flags, crp->crp_buf, ++ crd->crd_skip + crd->crd_len - ivlen, ivlen, ++ cmd->softc->sc_sessions[cmd->session_num].hs_iv); ++ break; ++ } ++ } ++ ++ if (macbuf != NULL) { ++ for (crd = crp->crp_desc; crd; crd = crd->crd_next) { ++ int len; ++ ++ if (crd->crd_alg != CRYPTO_MD5 && ++ crd->crd_alg != CRYPTO_SHA1 && ++ crd->crd_alg != CRYPTO_MD5_HMAC && ++ crd->crd_alg != CRYPTO_SHA1_HMAC) { ++ continue; ++ } ++ len = cmd->softc->sc_sessions[cmd->session_num].hs_mlen; ++ crypto_copyback(crp->crp_flags, crp->crp_buf, ++ crd->crd_inject, len, macbuf); ++ break; ++ } ++ } ++ ++ if (cmd->src_map != cmd->dst_map) ++ pci_unmap_buf(sc, &cmd->dst); ++ pci_unmap_buf(sc, &cmd->src); ++ kfree(cmd); ++ crypto_done(crp); ++} ++ ++/* ++ * 7811 PB3 rev/2 parts lock-up on burst writes to Group 0 ++ * and Group 1 registers; avoid conditions that could create ++ * burst writes by doing a read in between the writes. ++ * ++ * NB: The read we interpose is always to the same register; ++ * we do this because reading from an arbitrary (e.g. last) ++ * register may not always work. ++ */ ++static void ++hifn_write_reg_0(struct hifn_softc *sc, bus_size_t reg, u_int32_t val) ++{ ++ if (sc->sc_flags & HIFN_IS_7811) { ++ if (sc->sc_bar0_lastreg == reg - 4) ++ readl(sc->sc_bar0 + HIFN_0_PUCNFG); ++ sc->sc_bar0_lastreg = reg; ++ } ++ writel(val, sc->sc_bar0 + reg); ++} ++ ++static void ++hifn_write_reg_1(struct hifn_softc *sc, bus_size_t reg, u_int32_t val) ++{ ++ if (sc->sc_flags & HIFN_IS_7811) { ++ if (sc->sc_bar1_lastreg == reg - 4) ++ readl(sc->sc_bar1 + HIFN_1_REVID); ++ sc->sc_bar1_lastreg = reg; ++ } ++ writel(val, sc->sc_bar1 + reg); ++} ++ ++ ++static struct pci_device_id hifn_pci_tbl[] = { ++ { PCI_VENDOR_HIFN, PCI_PRODUCT_HIFN_7951, ++ PCI_ANY_ID, PCI_ANY_ID, 0, 0, }, ++ { PCI_VENDOR_HIFN, PCI_PRODUCT_HIFN_7955, ++ PCI_ANY_ID, PCI_ANY_ID, 0, 0, }, ++ { PCI_VENDOR_HIFN, PCI_PRODUCT_HIFN_7956, ++ PCI_ANY_ID, PCI_ANY_ID, 0, 0, }, ++ { PCI_VENDOR_NETSEC, PCI_PRODUCT_NETSEC_7751, ++ PCI_ANY_ID, PCI_ANY_ID, 0, 0, }, ++ { PCI_VENDOR_INVERTEX, PCI_PRODUCT_INVERTEX_AEON, ++ PCI_ANY_ID, PCI_ANY_ID, 0, 0, }, ++ { PCI_VENDOR_HIFN, PCI_PRODUCT_HIFN_7811, ++ PCI_ANY_ID, PCI_ANY_ID, 0, 0, }, ++ /* ++ * Other vendors share this PCI ID as well, such as ++ * http://www.powercrypt.com, and obviously they also ++ * use the same key. ++ */ ++ { PCI_VENDOR_HIFN, PCI_PRODUCT_HIFN_7751, ++ PCI_ANY_ID, PCI_ANY_ID, 0, 0, }, ++ { 0, 0, 0, 0, 0, 0, } ++}; ++MODULE_DEVICE_TABLE(pci, hifn_pci_tbl); ++ ++static struct pci_driver hifn_driver = { ++ .name = "hifn", ++ .id_table = hifn_pci_tbl, ++ .probe = hifn_probe, ++ .remove = hifn_remove, ++ /* add PM stuff here one day */ ++}; ++ ++static int __init hifn_init (void) ++{ ++ struct hifn_softc *sc = NULL; ++ int rc; ++ ++ DPRINTF("%s(%p)\n", __FUNCTION__, hifn_init); ++ ++ rc = pci_register_driver(&hifn_driver); ++ pci_register_driver_compat(&hifn_driver, rc); ++ ++ return rc; ++} ++ ++static void __exit hifn_exit (void) ++{ ++ pci_unregister_driver(&hifn_driver); ++} ++ ++module_init(hifn_init); ++module_exit(hifn_exit); ++ ++MODULE_LICENSE("BSD"); ++MODULE_AUTHOR("David McCullough <david_mccullough@securecomputing.com>"); ++MODULE_DESCRIPTION("OCF driver for hifn PCI crypto devices"); +diff -Nur linux-2.6.30.orig/crypto/ocf/hifn/hifn7751reg.h linux-2.6.30/crypto/ocf/hifn/hifn7751reg.h +--- linux-2.6.30.orig/crypto/ocf/hifn/hifn7751reg.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/hifn/hifn7751reg.h 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,540 @@ ++/* $FreeBSD: src/sys/dev/hifn/hifn7751reg.h,v 1.7 2007/03/21 03:42:49 sam Exp $ */ ++/* $OpenBSD: hifn7751reg.h,v 1.35 2002/04/08 17:49:42 jason Exp $ */ ++ ++/*- ++ * Invertex AEON / Hifn 7751 driver ++ * Copyright (c) 1999 Invertex Inc. All rights reserved. ++ * Copyright (c) 1999 Theo de Raadt ++ * Copyright (c) 2000-2001 Network Security Technologies, Inc. ++ * http://www.netsec.net ++ * ++ * Please send any comments, feedback, bug-fixes, or feature requests to ++ * software@invertex.com. ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions ++ * are met: ++ * ++ * 1. Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * 2. Redistributions in binary form must reproduce the above copyright ++ * notice, this list of conditions and the following disclaimer in the ++ * documentation and/or other materials provided with the distribution. ++ * 3. The name of the author may not be used to endorse or promote products ++ * derived from this software without specific prior written permission. ++ * ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR ++ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES ++ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. ++ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, ++ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT ++ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, ++ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY ++ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT ++ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF ++ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. ++ * ++ * Effort sponsored in part by the Defense Advanced Research Projects ++ * Agency (DARPA) and Air Force Research Laboratory, Air Force ++ * Materiel Command, USAF, under agreement number F30602-01-2-0537. ++ * ++ */ ++#ifndef __HIFN_H__ ++#define __HIFN_H__ ++ ++/* ++ * Some PCI configuration space offset defines. The names were made ++ * identical to the names used by the Linux kernel. ++ */ ++#define HIFN_BAR0 PCIR_BAR(0) /* PUC register map */ ++#define HIFN_BAR1 PCIR_BAR(1) /* DMA register map */ ++#define HIFN_TRDY_TIMEOUT 0x40 ++#define HIFN_RETRY_TIMEOUT 0x41 ++ ++/* ++ * PCI vendor and device identifiers ++ * (the names are preserved from their OpenBSD source). ++ */ ++#define PCI_VENDOR_HIFN 0x13a3 /* Hifn */ ++#define PCI_PRODUCT_HIFN_7751 0x0005 /* 7751 */ ++#define PCI_PRODUCT_HIFN_6500 0x0006 /* 6500 */ ++#define PCI_PRODUCT_HIFN_7811 0x0007 /* 7811 */ ++#define PCI_PRODUCT_HIFN_7855 0x001f /* 7855 */ ++#define PCI_PRODUCT_HIFN_7951 0x0012 /* 7951 */ ++#define PCI_PRODUCT_HIFN_7955 0x0020 /* 7954/7955 */ ++#define PCI_PRODUCT_HIFN_7956 0x001d /* 7956 */ ++ ++#define PCI_VENDOR_INVERTEX 0x14e1 /* Invertex */ ++#define PCI_PRODUCT_INVERTEX_AEON 0x0005 /* AEON */ ++ ++#define PCI_VENDOR_NETSEC 0x1660 /* NetSec */ ++#define PCI_PRODUCT_NETSEC_7751 0x7751 /* 7751 */ ++ ++/* ++ * The values below should multiple of 4 -- and be large enough to handle ++ * any command the driver implements. ++ * ++ * MAX_COMMAND = base command + mac command + encrypt command + ++ * mac-key + rc4-key ++ * MAX_RESULT = base result + mac result + mac + encrypt result ++ * ++ * ++ */ ++#define HIFN_MAX_COMMAND (8 + 8 + 8 + 64 + 260) ++#define HIFN_MAX_RESULT (8 + 4 + 20 + 4) ++ ++/* ++ * hifn_desc_t ++ * ++ * Holds an individual descriptor for any of the rings. ++ */ ++typedef struct hifn_desc { ++ volatile u_int32_t l; /* length and status bits */ ++ volatile u_int32_t p; ++} hifn_desc_t; ++ ++/* ++ * Masks for the "length" field of struct hifn_desc. ++ */ ++#define HIFN_D_LENGTH 0x0000ffff /* length bit mask */ ++#define HIFN_D_MASKDONEIRQ 0x02000000 /* mask the done interrupt */ ++#define HIFN_D_DESTOVER 0x04000000 /* destination overflow */ ++#define HIFN_D_OVER 0x08000000 /* overflow */ ++#define HIFN_D_LAST 0x20000000 /* last descriptor in chain */ ++#define HIFN_D_JUMP 0x40000000 /* jump descriptor */ ++#define HIFN_D_VALID 0x80000000 /* valid bit */ ++ ++ ++/* ++ * Processing Unit Registers (offset from BASEREG0) ++ */ ++#define HIFN_0_PUDATA 0x00 /* Processing Unit Data */ ++#define HIFN_0_PUCTRL 0x04 /* Processing Unit Control */ ++#define HIFN_0_PUISR 0x08 /* Processing Unit Interrupt Status */ ++#define HIFN_0_PUCNFG 0x0c /* Processing Unit Configuration */ ++#define HIFN_0_PUIER 0x10 /* Processing Unit Interrupt Enable */ ++#define HIFN_0_PUSTAT 0x14 /* Processing Unit Status/Chip ID */ ++#define HIFN_0_FIFOSTAT 0x18 /* FIFO Status */ ++#define HIFN_0_FIFOCNFG 0x1c /* FIFO Configuration */ ++#define HIFN_0_PUCTRL2 0x28 /* Processing Unit Control (2nd map) */ ++#define HIFN_0_MUTE1 0x80 ++#define HIFN_0_MUTE2 0x90 ++#define HIFN_0_SPACESIZE 0x100 /* Register space size */ ++ ++/* Processing Unit Control Register (HIFN_0_PUCTRL) */ ++#define HIFN_PUCTRL_CLRSRCFIFO 0x0010 /* clear source fifo */ ++#define HIFN_PUCTRL_STOP 0x0008 /* stop pu */ ++#define HIFN_PUCTRL_LOCKRAM 0x0004 /* lock ram */ ++#define HIFN_PUCTRL_DMAENA 0x0002 /* enable dma */ ++#define HIFN_PUCTRL_RESET 0x0001 /* Reset processing unit */ ++ ++/* Processing Unit Interrupt Status Register (HIFN_0_PUISR) */ ++#define HIFN_PUISR_CMDINVAL 0x8000 /* Invalid command interrupt */ ++#define HIFN_PUISR_DATAERR 0x4000 /* Data error interrupt */ ++#define HIFN_PUISR_SRCFIFO 0x2000 /* Source FIFO ready interrupt */ ++#define HIFN_PUISR_DSTFIFO 0x1000 /* Destination FIFO ready interrupt */ ++#define HIFN_PUISR_DSTOVER 0x0200 /* Destination overrun interrupt */ ++#define HIFN_PUISR_SRCCMD 0x0080 /* Source command interrupt */ ++#define HIFN_PUISR_SRCCTX 0x0040 /* Source context interrupt */ ++#define HIFN_PUISR_SRCDATA 0x0020 /* Source data interrupt */ ++#define HIFN_PUISR_DSTDATA 0x0010 /* Destination data interrupt */ ++#define HIFN_PUISR_DSTRESULT 0x0004 /* Destination result interrupt */ ++ ++/* Processing Unit Configuration Register (HIFN_0_PUCNFG) */ ++#define HIFN_PUCNFG_DRAMMASK 0xe000 /* DRAM size mask */ ++#define HIFN_PUCNFG_DSZ_256K 0x0000 /* 256k dram */ ++#define HIFN_PUCNFG_DSZ_512K 0x2000 /* 512k dram */ ++#define HIFN_PUCNFG_DSZ_1M 0x4000 /* 1m dram */ ++#define HIFN_PUCNFG_DSZ_2M 0x6000 /* 2m dram */ ++#define HIFN_PUCNFG_DSZ_4M 0x8000 /* 4m dram */ ++#define HIFN_PUCNFG_DSZ_8M 0xa000 /* 8m dram */ ++#define HIFN_PUNCFG_DSZ_16M 0xc000 /* 16m dram */ ++#define HIFN_PUCNFG_DSZ_32M 0xe000 /* 32m dram */ ++#define HIFN_PUCNFG_DRAMREFRESH 0x1800 /* DRAM refresh rate mask */ ++#define HIFN_PUCNFG_DRFR_512 0x0000 /* 512 divisor of ECLK */ ++#define HIFN_PUCNFG_DRFR_256 0x0800 /* 256 divisor of ECLK */ ++#define HIFN_PUCNFG_DRFR_128 0x1000 /* 128 divisor of ECLK */ ++#define HIFN_PUCNFG_TCALLPHASES 0x0200 /* your guess is as good as mine... */ ++#define HIFN_PUCNFG_TCDRVTOTEM 0x0100 /* your guess is as good as mine... */ ++#define HIFN_PUCNFG_BIGENDIAN 0x0080 /* DMA big endian mode */ ++#define HIFN_PUCNFG_BUS32 0x0040 /* Bus width 32bits */ ++#define HIFN_PUCNFG_BUS16 0x0000 /* Bus width 16 bits */ ++#define HIFN_PUCNFG_CHIPID 0x0020 /* Allow chipid from PUSTAT */ ++#define HIFN_PUCNFG_DRAM 0x0010 /* Context RAM is DRAM */ ++#define HIFN_PUCNFG_SRAM 0x0000 /* Context RAM is SRAM */ ++#define HIFN_PUCNFG_COMPSING 0x0004 /* Enable single compression context */ ++#define HIFN_PUCNFG_ENCCNFG 0x0002 /* Encryption configuration */ ++ ++/* Processing Unit Interrupt Enable Register (HIFN_0_PUIER) */ ++#define HIFN_PUIER_CMDINVAL 0x8000 /* Invalid command interrupt */ ++#define HIFN_PUIER_DATAERR 0x4000 /* Data error interrupt */ ++#define HIFN_PUIER_SRCFIFO 0x2000 /* Source FIFO ready interrupt */ ++#define HIFN_PUIER_DSTFIFO 0x1000 /* Destination FIFO ready interrupt */ ++#define HIFN_PUIER_DSTOVER 0x0200 /* Destination overrun interrupt */ ++#define HIFN_PUIER_SRCCMD 0x0080 /* Source command interrupt */ ++#define HIFN_PUIER_SRCCTX 0x0040 /* Source context interrupt */ ++#define HIFN_PUIER_SRCDATA 0x0020 /* Source data interrupt */ ++#define HIFN_PUIER_DSTDATA 0x0010 /* Destination data interrupt */ ++#define HIFN_PUIER_DSTRESULT 0x0004 /* Destination result interrupt */ ++ ++/* Processing Unit Status Register/Chip ID (HIFN_0_PUSTAT) */ ++#define HIFN_PUSTAT_CMDINVAL 0x8000 /* Invalid command interrupt */ ++#define HIFN_PUSTAT_DATAERR 0x4000 /* Data error interrupt */ ++#define HIFN_PUSTAT_SRCFIFO 0x2000 /* Source FIFO ready interrupt */ ++#define HIFN_PUSTAT_DSTFIFO 0x1000 /* Destination FIFO ready interrupt */ ++#define HIFN_PUSTAT_DSTOVER 0x0200 /* Destination overrun interrupt */ ++#define HIFN_PUSTAT_SRCCMD 0x0080 /* Source command interrupt */ ++#define HIFN_PUSTAT_SRCCTX 0x0040 /* Source context interrupt */ ++#define HIFN_PUSTAT_SRCDATA 0x0020 /* Source data interrupt */ ++#define HIFN_PUSTAT_DSTDATA 0x0010 /* Destination data interrupt */ ++#define HIFN_PUSTAT_DSTRESULT 0x0004 /* Destination result interrupt */ ++#define HIFN_PUSTAT_CHIPREV 0x00ff /* Chip revision mask */ ++#define HIFN_PUSTAT_CHIPENA 0xff00 /* Chip enabled mask */ ++#define HIFN_PUSTAT_ENA_2 0x1100 /* Level 2 enabled */ ++#define HIFN_PUSTAT_ENA_1 0x1000 /* Level 1 enabled */ ++#define HIFN_PUSTAT_ENA_0 0x3000 /* Level 0 enabled */ ++#define HIFN_PUSTAT_REV_2 0x0020 /* 7751 PT6/2 */ ++#define HIFN_PUSTAT_REV_3 0x0030 /* 7751 PT6/3 */ ++ ++/* FIFO Status Register (HIFN_0_FIFOSTAT) */ ++#define HIFN_FIFOSTAT_SRC 0x7f00 /* Source FIFO available */ ++#define HIFN_FIFOSTAT_DST 0x007f /* Destination FIFO available */ ++ ++/* FIFO Configuration Register (HIFN_0_FIFOCNFG) */ ++#define HIFN_FIFOCNFG_THRESHOLD 0x0400 /* must be written as this value */ ++ ++/* ++ * DMA Interface Registers (offset from BASEREG1) ++ */ ++#define HIFN_1_DMA_CRAR 0x0c /* DMA Command Ring Address */ ++#define HIFN_1_DMA_SRAR 0x1c /* DMA Source Ring Address */ ++#define HIFN_1_DMA_RRAR 0x2c /* DMA Result Ring Address */ ++#define HIFN_1_DMA_DRAR 0x3c /* DMA Destination Ring Address */ ++#define HIFN_1_DMA_CSR 0x40 /* DMA Status and Control */ ++#define HIFN_1_DMA_IER 0x44 /* DMA Interrupt Enable */ ++#define HIFN_1_DMA_CNFG 0x48 /* DMA Configuration */ ++#define HIFN_1_PLL 0x4c /* 7955/7956: PLL config */ ++#define HIFN_1_7811_RNGENA 0x60 /* 7811: rng enable */ ++#define HIFN_1_7811_RNGCFG 0x64 /* 7811: rng config */ ++#define HIFN_1_7811_RNGDAT 0x68 /* 7811: rng data */ ++#define HIFN_1_7811_RNGSTS 0x6c /* 7811: rng status */ ++#define HIFN_1_DMA_CNFG2 0x6c /* 7955/7956: dma config #2 */ ++#define HIFN_1_7811_MIPSRST 0x94 /* 7811: MIPS reset */ ++#define HIFN_1_REVID 0x98 /* Revision ID */ ++ ++#define HIFN_1_PUB_RESET 0x204 /* Public/RNG Reset */ ++#define HIFN_1_PUB_BASE 0x300 /* Public Base Address */ ++#define HIFN_1_PUB_OPLEN 0x304 /* 7951-compat Public Operand Length */ ++#define HIFN_1_PUB_OP 0x308 /* 7951-compat Public Operand */ ++#define HIFN_1_PUB_STATUS 0x30c /* 7951-compat Public Status */ ++#define HIFN_1_PUB_IEN 0x310 /* Public Interrupt enable */ ++#define HIFN_1_RNG_CONFIG 0x314 /* RNG config */ ++#define HIFN_1_RNG_DATA 0x318 /* RNG data */ ++#define HIFN_1_PUB_MODE 0x320 /* PK mode */ ++#define HIFN_1_PUB_FIFO_OPLEN 0x380 /* first element of oplen fifo */ ++#define HIFN_1_PUB_FIFO_OP 0x384 /* first element of op fifo */ ++#define HIFN_1_PUB_MEM 0x400 /* start of Public key memory */ ++#define HIFN_1_PUB_MEMEND 0xbff /* end of Public key memory */ ++ ++/* DMA Status and Control Register (HIFN_1_DMA_CSR) */ ++#define HIFN_DMACSR_D_CTRLMASK 0xc0000000 /* Destinition Ring Control */ ++#define HIFN_DMACSR_D_CTRL_NOP 0x00000000 /* Dest. Control: no-op */ ++#define HIFN_DMACSR_D_CTRL_DIS 0x40000000 /* Dest. Control: disable */ ++#define HIFN_DMACSR_D_CTRL_ENA 0x80000000 /* Dest. Control: enable */ ++#define HIFN_DMACSR_D_ABORT 0x20000000 /* Destinition Ring PCIAbort */ ++#define HIFN_DMACSR_D_DONE 0x10000000 /* Destinition Ring Done */ ++#define HIFN_DMACSR_D_LAST 0x08000000 /* Destinition Ring Last */ ++#define HIFN_DMACSR_D_WAIT 0x04000000 /* Destinition Ring Waiting */ ++#define HIFN_DMACSR_D_OVER 0x02000000 /* Destinition Ring Overflow */ ++#define HIFN_DMACSR_R_CTRL 0x00c00000 /* Result Ring Control */ ++#define HIFN_DMACSR_R_CTRL_NOP 0x00000000 /* Result Control: no-op */ ++#define HIFN_DMACSR_R_CTRL_DIS 0x00400000 /* Result Control: disable */ ++#define HIFN_DMACSR_R_CTRL_ENA 0x00800000 /* Result Control: enable */ ++#define HIFN_DMACSR_R_ABORT 0x00200000 /* Result Ring PCI Abort */ ++#define HIFN_DMACSR_R_DONE 0x00100000 /* Result Ring Done */ ++#define HIFN_DMACSR_R_LAST 0x00080000 /* Result Ring Last */ ++#define HIFN_DMACSR_R_WAIT 0x00040000 /* Result Ring Waiting */ ++#define HIFN_DMACSR_R_OVER 0x00020000 /* Result Ring Overflow */ ++#define HIFN_DMACSR_S_CTRL 0x0000c000 /* Source Ring Control */ ++#define HIFN_DMACSR_S_CTRL_NOP 0x00000000 /* Source Control: no-op */ ++#define HIFN_DMACSR_S_CTRL_DIS 0x00004000 /* Source Control: disable */ ++#define HIFN_DMACSR_S_CTRL_ENA 0x00008000 /* Source Control: enable */ ++#define HIFN_DMACSR_S_ABORT 0x00002000 /* Source Ring PCI Abort */ ++#define HIFN_DMACSR_S_DONE 0x00001000 /* Source Ring Done */ ++#define HIFN_DMACSR_S_LAST 0x00000800 /* Source Ring Last */ ++#define HIFN_DMACSR_S_WAIT 0x00000400 /* Source Ring Waiting */ ++#define HIFN_DMACSR_ILLW 0x00000200 /* Illegal write (7811 only) */ ++#define HIFN_DMACSR_ILLR 0x00000100 /* Illegal read (7811 only) */ ++#define HIFN_DMACSR_C_CTRL 0x000000c0 /* Command Ring Control */ ++#define HIFN_DMACSR_C_CTRL_NOP 0x00000000 /* Command Control: no-op */ ++#define HIFN_DMACSR_C_CTRL_DIS 0x00000040 /* Command Control: disable */ ++#define HIFN_DMACSR_C_CTRL_ENA 0x00000080 /* Command Control: enable */ ++#define HIFN_DMACSR_C_ABORT 0x00000020 /* Command Ring PCI Abort */ ++#define HIFN_DMACSR_C_DONE 0x00000010 /* Command Ring Done */ ++#define HIFN_DMACSR_C_LAST 0x00000008 /* Command Ring Last */ ++#define HIFN_DMACSR_C_WAIT 0x00000004 /* Command Ring Waiting */ ++#define HIFN_DMACSR_PUBDONE 0x00000002 /* Public op done (7951 only) */ ++#define HIFN_DMACSR_ENGINE 0x00000001 /* Command Ring Engine IRQ */ ++ ++/* DMA Interrupt Enable Register (HIFN_1_DMA_IER) */ ++#define HIFN_DMAIER_D_ABORT 0x20000000 /* Destination Ring PCIAbort */ ++#define HIFN_DMAIER_D_DONE 0x10000000 /* Destination Ring Done */ ++#define HIFN_DMAIER_D_LAST 0x08000000 /* Destination Ring Last */ ++#define HIFN_DMAIER_D_WAIT 0x04000000 /* Destination Ring Waiting */ ++#define HIFN_DMAIER_D_OVER 0x02000000 /* Destination Ring Overflow */ ++#define HIFN_DMAIER_R_ABORT 0x00200000 /* Result Ring PCI Abort */ ++#define HIFN_DMAIER_R_DONE 0x00100000 /* Result Ring Done */ ++#define HIFN_DMAIER_R_LAST 0x00080000 /* Result Ring Last */ ++#define HIFN_DMAIER_R_WAIT 0x00040000 /* Result Ring Waiting */ ++#define HIFN_DMAIER_R_OVER 0x00020000 /* Result Ring Overflow */ ++#define HIFN_DMAIER_S_ABORT 0x00002000 /* Source Ring PCI Abort */ ++#define HIFN_DMAIER_S_DONE 0x00001000 /* Source Ring Done */ ++#define HIFN_DMAIER_S_LAST 0x00000800 /* Source Ring Last */ ++#define HIFN_DMAIER_S_WAIT 0x00000400 /* Source Ring Waiting */ ++#define HIFN_DMAIER_ILLW 0x00000200 /* Illegal write (7811 only) */ ++#define HIFN_DMAIER_ILLR 0x00000100 /* Illegal read (7811 only) */ ++#define HIFN_DMAIER_C_ABORT 0x00000020 /* Command Ring PCI Abort */ ++#define HIFN_DMAIER_C_DONE 0x00000010 /* Command Ring Done */ ++#define HIFN_DMAIER_C_LAST 0x00000008 /* Command Ring Last */ ++#define HIFN_DMAIER_C_WAIT 0x00000004 /* Command Ring Waiting */ ++#define HIFN_DMAIER_PUBDONE 0x00000002 /* public op done (7951 only) */ ++#define HIFN_DMAIER_ENGINE 0x00000001 /* Engine IRQ */ ++ ++/* DMA Configuration Register (HIFN_1_DMA_CNFG) */ ++#define HIFN_DMACNFG_BIGENDIAN 0x10000000 /* big endian mode */ ++#define HIFN_DMACNFG_POLLFREQ 0x00ff0000 /* Poll frequency mask */ ++#define HIFN_DMACNFG_UNLOCK 0x00000800 ++#define HIFN_DMACNFG_POLLINVAL 0x00000700 /* Invalid Poll Scalar */ ++#define HIFN_DMACNFG_LAST 0x00000010 /* Host control LAST bit */ ++#define HIFN_DMACNFG_MODE 0x00000004 /* DMA mode */ ++#define HIFN_DMACNFG_DMARESET 0x00000002 /* DMA Reset # */ ++#define HIFN_DMACNFG_MSTRESET 0x00000001 /* Master Reset # */ ++ ++/* DMA Configuration Register (HIFN_1_DMA_CNFG2) */ ++#define HIFN_DMACNFG2_PKSWAP32 (1 << 19) /* swap the OPLEN/OP reg */ ++#define HIFN_DMACNFG2_PKSWAP8 (1 << 18) /* swap the bits of OPLEN/OP */ ++#define HIFN_DMACNFG2_BAR0_SWAP32 (1<<17) /* swap the bytes of BAR0 */ ++#define HIFN_DMACNFG2_BAR1_SWAP8 (1<<16) /* swap the bits of BAR0 */ ++#define HIFN_DMACNFG2_INIT_WRITE_BURST_SHIFT 12 ++#define HIFN_DMACNFG2_INIT_READ_BURST_SHIFT 8 ++#define HIFN_DMACNFG2_TGT_WRITE_BURST_SHIFT 4 ++#define HIFN_DMACNFG2_TGT_READ_BURST_SHIFT 0 ++ ++/* 7811 RNG Enable Register (HIFN_1_7811_RNGENA) */ ++#define HIFN_7811_RNGENA_ENA 0x00000001 /* enable RNG */ ++ ++/* 7811 RNG Config Register (HIFN_1_7811_RNGCFG) */ ++#define HIFN_7811_RNGCFG_PRE1 0x00000f00 /* first prescalar */ ++#define HIFN_7811_RNGCFG_OPRE 0x00000080 /* output prescalar */ ++#define HIFN_7811_RNGCFG_DEFL 0x00000f80 /* 2 words/ 1/100 sec */ ++ ++/* 7811 RNG Status Register (HIFN_1_7811_RNGSTS) */ ++#define HIFN_7811_RNGSTS_RDY 0x00004000 /* two numbers in FIFO */ ++#define HIFN_7811_RNGSTS_UFL 0x00001000 /* rng underflow */ ++ ++/* 7811 MIPS Reset Register (HIFN_1_7811_MIPSRST) */ ++#define HIFN_MIPSRST_BAR2SIZE 0xffff0000 /* sdram size */ ++#define HIFN_MIPSRST_GPRAMINIT 0x00008000 /* gpram can be accessed */ ++#define HIFN_MIPSRST_CRAMINIT 0x00004000 /* ctxram can be accessed */ ++#define HIFN_MIPSRST_LED2 0x00000400 /* external LED2 */ ++#define HIFN_MIPSRST_LED1 0x00000200 /* external LED1 */ ++#define HIFN_MIPSRST_LED0 0x00000100 /* external LED0 */ ++#define HIFN_MIPSRST_MIPSDIS 0x00000004 /* disable MIPS */ ++#define HIFN_MIPSRST_MIPSRST 0x00000002 /* warm reset MIPS */ ++#define HIFN_MIPSRST_MIPSCOLD 0x00000001 /* cold reset MIPS */ ++ ++/* Public key reset register (HIFN_1_PUB_RESET) */ ++#define HIFN_PUBRST_RESET 0x00000001 /* reset public/rng unit */ ++ ++/* Public operation register (HIFN_1_PUB_OP) */ ++#define HIFN_PUBOP_AOFFSET 0x0000003e /* A offset */ ++#define HIFN_PUBOP_BOFFSET 0x00000fc0 /* B offset */ ++#define HIFN_PUBOP_MOFFSET 0x0003f000 /* M offset */ ++#define HIFN_PUBOP_OP_MASK 0x003c0000 /* Opcode: */ ++#define HIFN_PUBOP_OP_NOP 0x00000000 /* NOP */ ++#define HIFN_PUBOP_OP_ADD 0x00040000 /* ADD */ ++#define HIFN_PUBOP_OP_ADDC 0x00080000 /* ADD w/carry */ ++#define HIFN_PUBOP_OP_SUB 0x000c0000 /* SUB */ ++#define HIFN_PUBOP_OP_SUBC 0x00100000 /* SUB w/carry */ ++#define HIFN_PUBOP_OP_MODADD 0x00140000 /* Modular ADD */ ++#define HIFN_PUBOP_OP_MODSUB 0x00180000 /* Modular SUB */ ++#define HIFN_PUBOP_OP_INCA 0x001c0000 /* INC A */ ++#define HIFN_PUBOP_OP_DECA 0x00200000 /* DEC A */ ++#define HIFN_PUBOP_OP_MULT 0x00240000 /* MULT */ ++#define HIFN_PUBOP_OP_MODMULT 0x00280000 /* Modular MULT */ ++#define HIFN_PUBOP_OP_MODRED 0x002c0000 /* Modular Red */ ++#define HIFN_PUBOP_OP_MODEXP 0x00300000 /* Modular Exp */ ++ ++/* Public operand length register (HIFN_1_PUB_OPLEN) */ ++#define HIFN_PUBOPLEN_MODLEN 0x0000007f ++#define HIFN_PUBOPLEN_EXPLEN 0x0003ff80 ++#define HIFN_PUBOPLEN_REDLEN 0x003c0000 ++ ++/* Public status register (HIFN_1_PUB_STATUS) */ ++#define HIFN_PUBSTS_DONE 0x00000001 /* operation done */ ++#define HIFN_PUBSTS_CARRY 0x00000002 /* carry */ ++#define HIFN_PUBSTS_FIFO_EMPTY 0x00000100 /* fifo empty */ ++#define HIFN_PUBSTS_FIFO_FULL 0x00000200 /* fifo full */ ++#define HIFN_PUBSTS_FIFO_OVFL 0x00000400 /* fifo overflow */ ++#define HIFN_PUBSTS_FIFO_WRITE 0x000f0000 /* fifo write */ ++#define HIFN_PUBSTS_FIFO_READ 0x0f000000 /* fifo read */ ++ ++/* Public interrupt enable register (HIFN_1_PUB_IEN) */ ++#define HIFN_PUBIEN_DONE 0x00000001 /* operation done interrupt */ ++ ++/* Random number generator config register (HIFN_1_RNG_CONFIG) */ ++#define HIFN_RNGCFG_ENA 0x00000001 /* enable rng */ ++ ++/* ++ * Register offsets in register set 1 ++ */ ++ ++#define HIFN_UNLOCK_SECRET1 0xf4 ++#define HIFN_UNLOCK_SECRET2 0xfc ++ ++/* ++ * PLL config register ++ * ++ * This register is present only on 7954/7955/7956 parts. It must be ++ * programmed according to the bus interface method used by the h/w. ++ * Note that the parts require a stable clock. Since the PCI clock ++ * may vary the reference clock must usually be used. To avoid ++ * overclocking the core logic, setup must be done carefully, refer ++ * to the driver for details. The exact multiplier required varies ++ * by part and system configuration; refer to the Hifn documentation. ++ */ ++#define HIFN_PLL_REF_SEL 0x00000001 /* REF/HBI clk selection */ ++#define HIFN_PLL_BP 0x00000002 /* bypass (used during setup) */ ++/* bit 2 reserved */ ++#define HIFN_PLL_PK_CLK_SEL 0x00000008 /* public key clk select */ ++#define HIFN_PLL_PE_CLK_SEL 0x00000010 /* packet engine clk select */ ++/* bits 5-9 reserved */ ++#define HIFN_PLL_MBSET 0x00000400 /* must be set to 1 */ ++#define HIFN_PLL_ND 0x00003800 /* Fpll_ref multiplier select */ ++#define HIFN_PLL_ND_SHIFT 11 ++#define HIFN_PLL_ND_2 0x00000000 /* 2x */ ++#define HIFN_PLL_ND_4 0x00000800 /* 4x */ ++#define HIFN_PLL_ND_6 0x00001000 /* 6x */ ++#define HIFN_PLL_ND_8 0x00001800 /* 8x */ ++#define HIFN_PLL_ND_10 0x00002000 /* 10x */ ++#define HIFN_PLL_ND_12 0x00002800 /* 12x */ ++/* bits 14-15 reserved */ ++#define HIFN_PLL_IS 0x00010000 /* charge pump current select */ ++/* bits 17-31 reserved */ ++ ++/* ++ * Board configuration specifies only these bits. ++ */ ++#define HIFN_PLL_CONFIG (HIFN_PLL_IS|HIFN_PLL_ND|HIFN_PLL_REF_SEL) ++ ++/* ++ * Public Key Engine Mode Register ++ */ ++#define HIFN_PKMODE_HOSTINVERT (1 << 0) /* HOST INVERT */ ++#define HIFN_PKMODE_ENHANCED (1 << 1) /* Enable enhanced mode */ ++ ++ ++/********************************************************************* ++ * Structs for board commands ++ * ++ *********************************************************************/ ++ ++/* ++ * Structure to help build up the command data structure. ++ */ ++typedef struct hifn_base_command { ++ volatile u_int16_t masks; ++ volatile u_int16_t session_num; ++ volatile u_int16_t total_source_count; ++ volatile u_int16_t total_dest_count; ++} hifn_base_command_t; ++ ++#define HIFN_BASE_CMD_MAC 0x0400 ++#define HIFN_BASE_CMD_CRYPT 0x0800 ++#define HIFN_BASE_CMD_DECODE 0x2000 ++#define HIFN_BASE_CMD_SRCLEN_M 0xc000 ++#define HIFN_BASE_CMD_SRCLEN_S 14 ++#define HIFN_BASE_CMD_DSTLEN_M 0x3000 ++#define HIFN_BASE_CMD_DSTLEN_S 12 ++#define HIFN_BASE_CMD_LENMASK_HI 0x30000 ++#define HIFN_BASE_CMD_LENMASK_LO 0x0ffff ++ ++/* ++ * Structure to help build up the command data structure. ++ */ ++typedef struct hifn_crypt_command { ++ volatile u_int16_t masks; ++ volatile u_int16_t header_skip; ++ volatile u_int16_t source_count; ++ volatile u_int16_t reserved; ++} hifn_crypt_command_t; ++ ++#define HIFN_CRYPT_CMD_ALG_MASK 0x0003 /* algorithm: */ ++#define HIFN_CRYPT_CMD_ALG_DES 0x0000 /* DES */ ++#define HIFN_CRYPT_CMD_ALG_3DES 0x0001 /* 3DES */ ++#define HIFN_CRYPT_CMD_ALG_RC4 0x0002 /* RC4 */ ++#define HIFN_CRYPT_CMD_ALG_AES 0x0003 /* AES */ ++#define HIFN_CRYPT_CMD_MODE_MASK 0x0018 /* Encrypt mode: */ ++#define HIFN_CRYPT_CMD_MODE_ECB 0x0000 /* ECB */ ++#define HIFN_CRYPT_CMD_MODE_CBC 0x0008 /* CBC */ ++#define HIFN_CRYPT_CMD_MODE_CFB 0x0010 /* CFB */ ++#define HIFN_CRYPT_CMD_MODE_OFB 0x0018 /* OFB */ ++#define HIFN_CRYPT_CMD_CLR_CTX 0x0040 /* clear context */ ++#define HIFN_CRYPT_CMD_NEW_KEY 0x0800 /* expect new key */ ++#define HIFN_CRYPT_CMD_NEW_IV 0x1000 /* expect new iv */ ++ ++#define HIFN_CRYPT_CMD_SRCLEN_M 0xc000 ++#define HIFN_CRYPT_CMD_SRCLEN_S 14 ++ ++#define HIFN_CRYPT_CMD_KSZ_MASK 0x0600 /* AES key size: */ ++#define HIFN_CRYPT_CMD_KSZ_128 0x0000 /* 128 bit */ ++#define HIFN_CRYPT_CMD_KSZ_192 0x0200 /* 192 bit */ ++#define HIFN_CRYPT_CMD_KSZ_256 0x0400 /* 256 bit */ ++ ++/* ++ * Structure to help build up the command data structure. ++ */ ++typedef struct hifn_mac_command { ++ volatile u_int16_t masks; ++ volatile u_int16_t header_skip; ++ volatile u_int16_t source_count; ++ volatile u_int16_t reserved; ++} hifn_mac_command_t; ++ ++#define HIFN_MAC_CMD_ALG_MASK 0x0001 ++#define HIFN_MAC_CMD_ALG_SHA1 0x0000 ++#define HIFN_MAC_CMD_ALG_MD5 0x0001 ++#define HIFN_MAC_CMD_MODE_MASK 0x000c ++#define HIFN_MAC_CMD_MODE_HMAC 0x0000 ++#define HIFN_MAC_CMD_MODE_SSL_MAC 0x0004 ++#define HIFN_MAC_CMD_MODE_HASH 0x0008 ++#define HIFN_MAC_CMD_MODE_FULL 0x0004 ++#define HIFN_MAC_CMD_TRUNC 0x0010 ++#define HIFN_MAC_CMD_RESULT 0x0020 ++#define HIFN_MAC_CMD_APPEND 0x0040 ++#define HIFN_MAC_CMD_SRCLEN_M 0xc000 ++#define HIFN_MAC_CMD_SRCLEN_S 14 ++ ++/* ++ * MAC POS IPsec initiates authentication after encryption on encodes ++ * and before decryption on decodes. ++ */ ++#define HIFN_MAC_CMD_POS_IPSEC 0x0200 ++#define HIFN_MAC_CMD_NEW_KEY 0x0800 ++ ++/* ++ * The poll frequency and poll scalar defines are unshifted values used ++ * to set fields in the DMA Configuration Register. ++ */ ++#ifndef HIFN_POLL_FREQUENCY ++#define HIFN_POLL_FREQUENCY 0x1 ++#endif ++ ++#ifndef HIFN_POLL_SCALAR ++#define HIFN_POLL_SCALAR 0x0 ++#endif ++ ++#define HIFN_MAX_SEGLEN 0xffff /* maximum dma segment len */ ++#define HIFN_MAX_DMALEN 0x3ffff /* maximum dma length */ ++#endif /* __HIFN_H__ */ +diff -Nur linux-2.6.30.orig/crypto/ocf/hifn/hifn7751var.h linux-2.6.30/crypto/ocf/hifn/hifn7751var.h +--- linux-2.6.30.orig/crypto/ocf/hifn/hifn7751var.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/hifn/hifn7751var.h 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,369 @@ ++/* $FreeBSD: src/sys/dev/hifn/hifn7751var.h,v 1.9 2007/03/21 03:42:49 sam Exp $ */ ++/* $OpenBSD: hifn7751var.h,v 1.42 2002/04/08 17:49:42 jason Exp $ */ ++ ++/*- ++ * Invertex AEON / Hifn 7751 driver ++ * Copyright (c) 1999 Invertex Inc. All rights reserved. ++ * Copyright (c) 1999 Theo de Raadt ++ * Copyright (c) 2000-2001 Network Security Technologies, Inc. ++ * http://www.netsec.net ++ * ++ * Please send any comments, feedback, bug-fixes, or feature requests to ++ * software@invertex.com. ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions ++ * are met: ++ * ++ * 1. Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * 2. Redistributions in binary form must reproduce the above copyright ++ * notice, this list of conditions and the following disclaimer in the ++ * documentation and/or other materials provided with the distribution. ++ * 3. The name of the author may not be used to endorse or promote products ++ * derived from this software without specific prior written permission. ++ * ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR ++ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES ++ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. ++ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, ++ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT ++ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, ++ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY ++ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT ++ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF ++ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. ++ * ++ * Effort sponsored in part by the Defense Advanced Research Projects ++ * Agency (DARPA) and Air Force Research Laboratory, Air Force ++ * Materiel Command, USAF, under agreement number F30602-01-2-0537. ++ * ++ */ ++ ++#ifndef __HIFN7751VAR_H__ ++#define __HIFN7751VAR_H__ ++ ++#ifdef __KERNEL__ ++ ++/* ++ * Some configurable values for the driver. By default command+result ++ * descriptor rings are the same size. The src+dst descriptor rings ++ * are sized at 3.5x the number of potential commands. Slower parts ++ * (e.g. 7951) tend to run out of src descriptors; faster parts (7811) ++ * src+cmd/result descriptors. It's not clear that increasing the size ++ * of the descriptor rings helps performance significantly as other ++ * factors tend to come into play (e.g. copying misaligned packets). ++ */ ++#define HIFN_D_CMD_RSIZE 24 /* command descriptors */ ++#define HIFN_D_SRC_RSIZE ((HIFN_D_CMD_RSIZE * 7) / 2) /* source descriptors */ ++#define HIFN_D_RES_RSIZE HIFN_D_CMD_RSIZE /* result descriptors */ ++#define HIFN_D_DST_RSIZE HIFN_D_SRC_RSIZE /* destination descriptors */ ++ ++/* ++ * Length values for cryptography ++ */ ++#define HIFN_DES_KEY_LENGTH 8 ++#define HIFN_3DES_KEY_LENGTH 24 ++#define HIFN_MAX_CRYPT_KEY_LENGTH HIFN_3DES_KEY_LENGTH ++#define HIFN_IV_LENGTH 8 ++#define HIFN_AES_IV_LENGTH 16 ++#define HIFN_MAX_IV_LENGTH HIFN_AES_IV_LENGTH ++ ++/* ++ * Length values for authentication ++ */ ++#define HIFN_MAC_KEY_LENGTH 64 ++#define HIFN_MD5_LENGTH 16 ++#define HIFN_SHA1_LENGTH 20 ++#define HIFN_MAC_TRUNC_LENGTH 12 ++ ++#define MAX_SCATTER 64 ++ ++/* ++ * Data structure to hold all 4 rings and any other ring related data. ++ */ ++struct hifn_dma { ++ /* ++ * Descriptor rings. We add +1 to the size to accomidate the ++ * jump descriptor. ++ */ ++ struct hifn_desc cmdr[HIFN_D_CMD_RSIZE+1]; ++ struct hifn_desc srcr[HIFN_D_SRC_RSIZE+1]; ++ struct hifn_desc dstr[HIFN_D_DST_RSIZE+1]; ++ struct hifn_desc resr[HIFN_D_RES_RSIZE+1]; ++ ++ struct hifn_command *hifn_commands[HIFN_D_RES_RSIZE]; ++ ++ u_char command_bufs[HIFN_D_CMD_RSIZE][HIFN_MAX_COMMAND]; ++ u_char result_bufs[HIFN_D_CMD_RSIZE][HIFN_MAX_RESULT]; ++ u_int32_t slop[HIFN_D_CMD_RSIZE]; ++ ++ u_int64_t test_src, test_dst; ++ ++ /* ++ * Our current positions for insertion and removal from the desriptor ++ * rings. ++ */ ++ int cmdi, srci, dsti, resi; ++ volatile int cmdu, srcu, dstu, resu; ++ int cmdk, srck, dstk, resk; ++}; ++ ++struct hifn_session { ++ int hs_used; ++ int hs_mlen; ++ u_int8_t hs_iv[HIFN_MAX_IV_LENGTH]; ++}; ++ ++#define HIFN_RING_SYNC(sc, r, i, f) \ ++ /* DAVIDM bus_dmamap_sync((sc)->sc_dmat, (sc)->sc_dmamap, (f)) */ ++ ++#define HIFN_CMDR_SYNC(sc, i, f) HIFN_RING_SYNC((sc), cmdr, (i), (f)) ++#define HIFN_RESR_SYNC(sc, i, f) HIFN_RING_SYNC((sc), resr, (i), (f)) ++#define HIFN_SRCR_SYNC(sc, i, f) HIFN_RING_SYNC((sc), srcr, (i), (f)) ++#define HIFN_DSTR_SYNC(sc, i, f) HIFN_RING_SYNC((sc), dstr, (i), (f)) ++ ++#define HIFN_CMD_SYNC(sc, i, f) \ ++ /* DAVIDM bus_dmamap_sync((sc)->sc_dmat, (sc)->sc_dmamap, (f)) */ ++ ++#define HIFN_RES_SYNC(sc, i, f) \ ++ /* DAVIDM bus_dmamap_sync((sc)->sc_dmat, (sc)->sc_dmamap, (f)) */ ++ ++typedef int bus_size_t; ++ ++/* ++ * Holds data specific to a single HIFN board. ++ */ ++struct hifn_softc { ++ softc_device_decl sc_dev; ++ ++ struct pci_dev *sc_pcidev; /* PCI device pointer */ ++ spinlock_t sc_mtx; /* per-instance lock */ ++ ++ int sc_num; /* for multiple devs */ ++ ++ ocf_iomem_t sc_bar0; ++ bus_size_t sc_bar0_lastreg;/* bar0 last reg written */ ++ ocf_iomem_t sc_bar1; ++ bus_size_t sc_bar1_lastreg;/* bar1 last reg written */ ++ ++ int sc_irq; ++ ++ u_int32_t sc_dmaier; ++ u_int32_t sc_drammodel; /* 1=dram, 0=sram */ ++ u_int32_t sc_pllconfig; /* 7954/7955/7956 PLL config */ ++ ++ struct hifn_dma *sc_dma; ++ dma_addr_t sc_dma_physaddr;/* physical address of sc_dma */ ++ ++ int sc_dmansegs; ++ int32_t sc_cid; ++ int sc_maxses; ++ int sc_nsessions; ++ struct hifn_session *sc_sessions; ++ int sc_ramsize; ++ int sc_flags; ++#define HIFN_HAS_RNG 0x1 /* includes random number generator */ ++#define HIFN_HAS_PUBLIC 0x2 /* includes public key support */ ++#define HIFN_HAS_AES 0x4 /* includes AES support */ ++#define HIFN_IS_7811 0x8 /* Hifn 7811 part */ ++#define HIFN_IS_7956 0x10 /* Hifn 7956/7955 don't have SDRAM */ ++ ++ struct timer_list sc_tickto; /* for managing DMA */ ++ ++ int sc_rngfirst; ++ int sc_rnghz; /* RNG polling frequency */ ++ ++ int sc_c_busy; /* command ring busy */ ++ int sc_s_busy; /* source data ring busy */ ++ int sc_d_busy; /* destination data ring busy */ ++ int sc_r_busy; /* result ring busy */ ++ int sc_active; /* for initial countdown */ ++ int sc_needwakeup; /* ops q'd wating on resources */ ++ int sc_curbatch; /* # ops submitted w/o int */ ++ int sc_suspended; ++#ifdef HIFN_VULCANDEV ++ struct cdev *sc_pkdev; ++#endif ++}; ++ ++#define HIFN_LOCK(_sc) spin_lock_irqsave(&(_sc)->sc_mtx, l_flags) ++#define HIFN_UNLOCK(_sc) spin_unlock_irqrestore(&(_sc)->sc_mtx, l_flags) ++ ++/* ++ * hifn_command_t ++ * ++ * This is the control structure used to pass commands to hifn_encrypt(). ++ * ++ * flags ++ * ----- ++ * Flags is the bitwise "or" values for command configuration. A single ++ * encrypt direction needs to be set: ++ * ++ * HIFN_ENCODE or HIFN_DECODE ++ * ++ * To use cryptography, a single crypto algorithm must be included: ++ * ++ * HIFN_CRYPT_3DES or HIFN_CRYPT_DES ++ * ++ * To use authentication is used, a single MAC algorithm must be included: ++ * ++ * HIFN_MAC_MD5 or HIFN_MAC_SHA1 ++ * ++ * By default MD5 uses a 16 byte hash and SHA-1 uses a 20 byte hash. ++ * If the value below is set, hash values are truncated or assumed ++ * truncated to 12 bytes: ++ * ++ * HIFN_MAC_TRUNC ++ * ++ * Keys for encryption and authentication can be sent as part of a command, ++ * or the last key value used with a particular session can be retrieved ++ * and used again if either of these flags are not specified. ++ * ++ * HIFN_CRYPT_NEW_KEY, HIFN_MAC_NEW_KEY ++ * ++ * session_num ++ * ----------- ++ * A number between 0 and 2048 (for DRAM models) or a number between ++ * 0 and 768 (for SRAM models). Those who don't want to use session ++ * numbers should leave value at zero and send a new crypt key and/or ++ * new MAC key on every command. If you use session numbers and ++ * don't send a key with a command, the last key sent for that same ++ * session number will be used. ++ * ++ * Warning: Using session numbers and multiboard at the same time ++ * is currently broken. ++ * ++ * mbuf ++ * ---- ++ * Either fill in the mbuf pointer and npa=0 or ++ * fill packp[] and packl[] and set npa to > 0 ++ * ++ * mac_header_skip ++ * --------------- ++ * The number of bytes of the source_buf that are skipped over before ++ * authentication begins. This must be a number between 0 and 2^16-1 ++ * and can be used by IPsec implementers to skip over IP headers. ++ * *** Value ignored if authentication not used *** ++ * ++ * crypt_header_skip ++ * ----------------- ++ * The number of bytes of the source_buf that are skipped over before ++ * the cryptographic operation begins. This must be a number between 0 ++ * and 2^16-1. For IPsec, this number will always be 8 bytes larger ++ * than the auth_header_skip (to skip over the ESP header). ++ * *** Value ignored if cryptography not used *** ++ * ++ */ ++struct hifn_operand { ++ union { ++ struct sk_buff *skb; ++ struct uio *io; ++ unsigned char *buf; ++ } u; ++ void *map; ++ bus_size_t mapsize; ++ int nsegs; ++ struct { ++ dma_addr_t ds_addr; ++ int ds_len; ++ } segs[MAX_SCATTER]; ++}; ++ ++struct hifn_command { ++ u_int16_t session_num; ++ u_int16_t base_masks, cry_masks, mac_masks; ++ u_int8_t iv[HIFN_MAX_IV_LENGTH], *ck, mac[HIFN_MAC_KEY_LENGTH]; ++ int cklen; ++ int sloplen, slopidx; ++ ++ struct hifn_operand src; ++ struct hifn_operand dst; ++ ++ struct hifn_softc *softc; ++ struct cryptop *crp; ++ struct cryptodesc *enccrd, *maccrd; ++}; ++ ++#define src_skb src.u.skb ++#define src_io src.u.io ++#define src_map src.map ++#define src_mapsize src.mapsize ++#define src_segs src.segs ++#define src_nsegs src.nsegs ++#define src_buf src.u.buf ++ ++#define dst_skb dst.u.skb ++#define dst_io dst.u.io ++#define dst_map dst.map ++#define dst_mapsize dst.mapsize ++#define dst_segs dst.segs ++#define dst_nsegs dst.nsegs ++#define dst_buf dst.u.buf ++ ++/* ++ * Return values for hifn_crypto() ++ */ ++#define HIFN_CRYPTO_SUCCESS 0 ++#define HIFN_CRYPTO_BAD_INPUT (-1) ++#define HIFN_CRYPTO_RINGS_FULL (-2) ++ ++/************************************************************************** ++ * ++ * Function: hifn_crypto ++ * ++ * Purpose: Called by external drivers to begin an encryption on the ++ * HIFN board. ++ * ++ * Blocking/Non-blocking Issues ++ * ============================ ++ * The driver cannot block in hifn_crypto (no calls to tsleep) currently. ++ * hifn_crypto() returns HIFN_CRYPTO_RINGS_FULL if there is not enough ++ * room in any of the rings for the request to proceed. ++ * ++ * Return Values ++ * ============= ++ * 0 for success, negative values on error ++ * ++ * Defines for negative error codes are: ++ * ++ * HIFN_CRYPTO_BAD_INPUT : The passed in command had invalid settings. ++ * HIFN_CRYPTO_RINGS_FULL : All DMA rings were full and non-blocking ++ * behaviour was requested. ++ * ++ *************************************************************************/ ++ ++/* ++ * Convert back and forth from 'sid' to 'card' and 'session' ++ */ ++#define HIFN_CARD(sid) (((sid) & 0xf0000000) >> 28) ++#define HIFN_SESSION(sid) ((sid) & 0x000007ff) ++#define HIFN_SID(crd,ses) (((crd) << 28) | ((ses) & 0x7ff)) ++ ++#endif /* _KERNEL */ ++ ++struct hifn_stats { ++ u_int64_t hst_ibytes; ++ u_int64_t hst_obytes; ++ u_int32_t hst_ipackets; ++ u_int32_t hst_opackets; ++ u_int32_t hst_invalid; ++ u_int32_t hst_nomem; /* malloc or one of hst_nomem_* */ ++ u_int32_t hst_abort; ++ u_int32_t hst_noirq; /* IRQ for no reason */ ++ u_int32_t hst_totbatch; /* ops submitted w/o interrupt */ ++ u_int32_t hst_maxbatch; /* max ops submitted together */ ++ u_int32_t hst_unaligned; /* unaligned src caused copy */ ++ /* ++ * The following divides hst_nomem into more specific buckets. ++ */ ++ u_int32_t hst_nomem_map; /* bus_dmamap_create failed */ ++ u_int32_t hst_nomem_load; /* bus_dmamap_load_* failed */ ++ u_int32_t hst_nomem_mbuf; /* MGET* failed */ ++ u_int32_t hst_nomem_mcl; /* MCLGET* failed */ ++ u_int32_t hst_nomem_cr; /* out of command/result descriptor */ ++ u_int32_t hst_nomem_sd; /* out of src/dst descriptors */ ++}; ++ ++#endif /* __HIFN7751VAR_H__ */ +diff -Nur linux-2.6.30.orig/crypto/ocf/hifn/hifnHIPP.c linux-2.6.30/crypto/ocf/hifn/hifnHIPP.c +--- linux-2.6.30.orig/crypto/ocf/hifn/hifnHIPP.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/hifn/hifnHIPP.c 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,420 @@ ++/*- ++ * Driver for Hifn HIPP-I/II chipset ++ * Copyright (c) 2006 Michael Richardson <mcr@xelerance.com> ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions ++ * are met: ++ * ++ * 1. Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * 2. Redistributions in binary form must reproduce the above copyright ++ * notice, this list of conditions and the following disclaimer in the ++ * documentation and/or other materials provided with the distribution. ++ * 3. The name of the author may not be used to endorse or promote products ++ * derived from this software without specific prior written permission. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR ++ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES ++ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. ++ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, ++ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT ++ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, ++ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY ++ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT ++ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF ++ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. ++ * ++ * Effort sponsored by Hifn Inc. ++ * ++ */ ++ ++/* ++ * Driver for various Hifn encryption processors. ++ */ ++#ifndef AUTOCONF_INCLUDED ++#include <linux/config.h> ++#endif ++#include <linux/module.h> ++#include <linux/init.h> ++#include <linux/list.h> ++#include <linux/slab.h> ++#include <linux/wait.h> ++#include <linux/sched.h> ++#include <linux/pci.h> ++#include <linux/delay.h> ++#include <linux/interrupt.h> ++#include <linux/spinlock.h> ++#include <linux/random.h> ++#include <linux/version.h> ++#include <linux/skbuff.h> ++#include <linux/uio.h> ++#include <linux/sysfs.h> ++#include <linux/miscdevice.h> ++#include <asm/io.h> ++ ++#include <cryptodev.h> ++ ++#include "hifnHIPPreg.h" ++#include "hifnHIPPvar.h" ++ ++#if 1 ++#define DPRINTF(a...) if (hipp_debug) { \ ++ printk("%s: ", sc ? \ ++ device_get_nameunit(sc->sc_dev) : "hifn"); \ ++ printk(a); \ ++ } else ++#else ++#define DPRINTF(a...) ++#endif ++ ++typedef int bus_size_t; ++ ++static inline int ++pci_get_revid(struct pci_dev *dev) ++{ ++ u8 rid = 0; ++ pci_read_config_byte(dev, PCI_REVISION_ID, &rid); ++ return rid; ++} ++ ++#define debug hipp_debug ++int hipp_debug = 0; ++module_param(hipp_debug, int, 0644); ++MODULE_PARM_DESC(hipp_debug, "Enable debug"); ++ ++int hipp_maxbatch = 1; ++module_param(hipp_maxbatch, int, 0644); ++MODULE_PARM_DESC(hipp_maxbatch, "max ops to batch w/o interrupt"); ++ ++static int hipp_probe(struct pci_dev *dev, const struct pci_device_id *ent); ++static void hipp_remove(struct pci_dev *dev); ++#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,19) ++static irqreturn_t hipp_intr(int irq, void *arg); ++#else ++static irqreturn_t hipp_intr(int irq, void *arg, struct pt_regs *regs); ++#endif ++ ++static int hipp_num_chips = 0; ++static struct hipp_softc *hipp_chip_idx[HIPP_MAX_CHIPS]; ++ ++static int hipp_newsession(device_t, u_int32_t *, struct cryptoini *); ++static int hipp_freesession(device_t, u_int64_t); ++static int hipp_process(device_t, struct cryptop *, int); ++ ++static device_method_t hipp_methods = { ++ /* crypto device methods */ ++ DEVMETHOD(cryptodev_newsession, hipp_newsession), ++ DEVMETHOD(cryptodev_freesession,hipp_freesession), ++ DEVMETHOD(cryptodev_process, hipp_process), ++}; ++ ++static __inline u_int32_t ++READ_REG(struct hipp_softc *sc, unsigned int barno, bus_size_t reg) ++{ ++ u_int32_t v = readl(sc->sc_bar[barno] + reg); ++ //sc->sc_bar0_lastreg = (bus_size_t) -1; ++ return (v); ++} ++static __inline void ++WRITE_REG(struct hipp_softc *sc, unsigned int barno, bus_size_t reg, u_int32_t val) ++{ ++ writel(val, sc->sc_bar[barno] + reg); ++} ++ ++#define READ_REG_0(sc, reg) READ_REG(sc, 0, reg) ++#define WRITE_REG_0(sc, reg, val) WRITE_REG(sc,0, reg, val) ++#define READ_REG_1(sc, reg) READ_REG(sc, 1, reg) ++#define WRITE_REG_1(sc, reg, val) WRITE_REG(sc,1, reg, val) ++ ++static int ++hipp_newsession(device_t dev, u_int32_t *sidp, struct cryptoini *cri) ++{ ++ return EINVAL; ++} ++ ++static int ++hipp_freesession(device_t dev, u_int64_t tid) ++{ ++ return EINVAL; ++} ++ ++static int ++hipp_process(device_t dev, struct cryptop *crp, int hint) ++{ ++ return EINVAL; ++} ++ ++static const char* ++hipp_partname(struct hipp_softc *sc, char buf[128], size_t blen) ++{ ++ char *n = NULL; ++ ++ switch (pci_get_vendor(sc->sc_pcidev)) { ++ case PCI_VENDOR_HIFN: ++ switch (pci_get_device(sc->sc_pcidev)) { ++ case PCI_PRODUCT_HIFN_7855: n = "Hifn 7855"; ++ case PCI_PRODUCT_HIFN_8155: n = "Hifn 8155"; ++ case PCI_PRODUCT_HIFN_6500: n = "Hifn 6500"; ++ } ++ } ++ ++ if(n==NULL) { ++ snprintf(buf, blen, "VID=%02x,PID=%02x", ++ pci_get_vendor(sc->sc_pcidev), ++ pci_get_device(sc->sc_pcidev)); ++ } else { ++ buf[0]='\0'; ++ strncat(buf, n, blen); ++ } ++ return buf; ++} ++ ++struct hipp_fs_entry { ++ struct attribute attr; ++ /* other stuff */ ++}; ++ ++ ++static ssize_t ++cryptoid_show(struct device *dev, ++ struct device_attribute *attr, ++ char *buf) ++{ ++ struct hipp_softc *sc; ++ ++ sc = pci_get_drvdata(to_pci_dev (dev)); ++ return sprintf (buf, "%d\n", sc->sc_cid); ++} ++ ++struct device_attribute hipp_dev_cryptoid = __ATTR_RO(cryptoid); ++ ++/* ++ * Attach an interface that successfully probed. ++ */ ++static int ++hipp_probe(struct pci_dev *dev, const struct pci_device_id *ent) ++{ ++ struct hipp_softc *sc = NULL; ++ int i; ++ //char rbase; ++ //u_int16_t ena; ++ int rev; ++ //int rseg; ++ int rc; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ if (pci_enable_device(dev) < 0) ++ return(-ENODEV); ++ ++ if (pci_set_mwi(dev)) ++ return(-ENODEV); ++ ++ if (!dev->irq) { ++ printk("hifn: found device with no IRQ assigned. check BIOS settings!"); ++ pci_disable_device(dev); ++ return(-ENODEV); ++ } ++ ++ sc = (struct hipp_softc *) kmalloc(sizeof(*sc), GFP_KERNEL); ++ if (!sc) ++ return(-ENOMEM); ++ memset(sc, 0, sizeof(*sc)); ++ ++ softc_device_init(sc, "hifn-hipp", hipp_num_chips, hipp_methods); ++ ++ sc->sc_pcidev = dev; ++ sc->sc_irq = -1; ++ sc->sc_cid = -1; ++ sc->sc_num = hipp_num_chips++; ++ ++ if (sc->sc_num < HIPP_MAX_CHIPS) ++ hipp_chip_idx[sc->sc_num] = sc; ++ ++ pci_set_drvdata(sc->sc_pcidev, sc); ++ ++ spin_lock_init(&sc->sc_mtx); ++ ++ /* ++ * Setup PCI resources. ++ * The READ_REG_0, WRITE_REG_0, READ_REG_1, ++ * and WRITE_REG_1 macros throughout the driver are used ++ * to permit better debugging. ++ */ ++ for(i=0; i<4; i++) { ++ unsigned long mem_start, mem_len; ++ mem_start = pci_resource_start(sc->sc_pcidev, i); ++ mem_len = pci_resource_len(sc->sc_pcidev, i); ++ sc->sc_barphy[i] = (caddr_t)mem_start; ++ sc->sc_bar[i] = (ocf_iomem_t) ioremap(mem_start, mem_len); ++ if (!sc->sc_bar[i]) { ++ device_printf(sc->sc_dev, "cannot map bar%d register space\n", i); ++ goto fail; ++ } ++ } ++ ++ //hipp_reset_board(sc, 0); ++ pci_set_master(sc->sc_pcidev); ++ ++ /* ++ * Arrange the interrupt line. ++ */ ++ rc = request_irq(dev->irq, hipp_intr, IRQF_SHARED, "hifn", sc); ++ if (rc) { ++ device_printf(sc->sc_dev, "could not map interrupt: %d\n", rc); ++ goto fail; ++ } ++ sc->sc_irq = dev->irq; ++ ++ rev = READ_REG_1(sc, HIPP_1_REVID) & 0xffff; ++ ++ { ++ char b[32]; ++ device_printf(sc->sc_dev, "%s, rev %u", ++ hipp_partname(sc, b, sizeof(b)), rev); ++ } ++ ++#if 0 ++ if (sc->sc_flags & HIFN_IS_7956) ++ printf(", pll=0x%x<%s clk, %ux mult>", ++ sc->sc_pllconfig, ++ sc->sc_pllconfig & HIFN_PLL_REF_SEL ? "ext" : "pci", ++ 2 + 2*((sc->sc_pllconfig & HIFN_PLL_ND) >> 11)); ++#endif ++ printf("\n"); ++ ++ sc->sc_cid = crypto_get_driverid(softc_get_device(sc),CRYPTOCAP_F_HARDWARE); ++ if (sc->sc_cid < 0) { ++ device_printf(sc->sc_dev, "could not get crypto driver id\n"); ++ goto fail; ++ } ++ ++#if 0 /* cannot work with a non-GPL module */ ++ /* make a sysfs entry to let the world know what entry we got */ ++ sysfs_create_file(&sc->sc_pcidev->dev.kobj, &hipp_dev_cryptoid.attr); ++#endif ++ ++#if 0 ++ init_timer(&sc->sc_tickto); ++ sc->sc_tickto.function = hifn_tick; ++ sc->sc_tickto.data = (unsigned long) sc->sc_num; ++ mod_timer(&sc->sc_tickto, jiffies + HZ); ++#endif ++ ++#if 0 /* no code here yet ?? */ ++ crypto_register(sc->sc_cid, CRYPTO_3DES_CBC, 0, 0); ++#endif ++ ++ return (0); ++ ++fail: ++ if (sc->sc_cid >= 0) ++ crypto_unregister_all(sc->sc_cid); ++ if (sc->sc_irq != -1) ++ free_irq(sc->sc_irq, sc); ++ ++#if 0 ++ if (sc->sc_dma) { ++ /* Turn off DMA polling */ ++ WRITE_REG_1(sc, HIFN_1_DMA_CNFG, HIFN_DMACNFG_MSTRESET | ++ HIFN_DMACNFG_DMARESET | HIFN_DMACNFG_MODE); ++ ++ pci_free_consistent(sc->sc_pcidev, ++ sizeof(*sc->sc_dma), ++ sc->sc_dma, sc->sc_dma_physaddr); ++ } ++#endif ++ kfree(sc); ++ return (-ENXIO); ++} ++ ++/* ++ * Detach an interface that successfully probed. ++ */ ++static void ++hipp_remove(struct pci_dev *dev) ++{ ++ struct hipp_softc *sc = pci_get_drvdata(dev); ++ unsigned long l_flags; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ /* disable interrupts */ ++ HIPP_LOCK(sc); ++ ++#if 0 ++ WRITE_REG_1(sc, HIFN_1_DMA_IER, 0); ++ HIFN_UNLOCK(sc); ++ ++ /*XXX other resources */ ++ del_timer_sync(&sc->sc_tickto); ++ ++ /* Turn off DMA polling */ ++ WRITE_REG_1(sc, HIFN_1_DMA_CNFG, HIFN_DMACNFG_MSTRESET | ++ HIFN_DMACNFG_DMARESET | HIFN_DMACNFG_MODE); ++#endif ++ ++ crypto_unregister_all(sc->sc_cid); ++ ++ free_irq(sc->sc_irq, sc); ++ ++#if 0 ++ pci_free_consistent(sc->sc_pcidev, sizeof(*sc->sc_dma), ++ sc->sc_dma, sc->sc_dma_physaddr); ++#endif ++} ++ ++#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,19) ++static irqreturn_t hipp_intr(int irq, void *arg) ++#else ++static irqreturn_t hipp_intr(int irq, void *arg, struct pt_regs *regs) ++#endif ++{ ++ struct hipp_softc *sc = arg; ++ ++ sc = sc; /* shut up compiler */ ++ ++ return IRQ_HANDLED; ++} ++ ++static struct pci_device_id hipp_pci_tbl[] = { ++ { PCI_VENDOR_HIFN, PCI_PRODUCT_HIFN_7855, ++ PCI_ANY_ID, PCI_ANY_ID, 0, 0, }, ++ { PCI_VENDOR_HIFN, PCI_PRODUCT_HIFN_8155, ++ PCI_ANY_ID, PCI_ANY_ID, 0, 0, }, ++}; ++MODULE_DEVICE_TABLE(pci, hipp_pci_tbl); ++ ++static struct pci_driver hipp_driver = { ++ .name = "hipp", ++ .id_table = hipp_pci_tbl, ++ .probe = hipp_probe, ++ .remove = hipp_remove, ++ /* add PM stuff here one day */ ++}; ++ ++static int __init hipp_init (void) ++{ ++ struct hipp_softc *sc = NULL; ++ int rc; ++ ++ DPRINTF("%s(%p)\n", __FUNCTION__, hipp_init); ++ ++ rc = pci_register_driver(&hipp_driver); ++ pci_register_driver_compat(&hipp_driver, rc); ++ ++ return rc; ++} ++ ++static void __exit hipp_exit (void) ++{ ++ pci_unregister_driver(&hipp_driver); ++} ++ ++module_init(hipp_init); ++module_exit(hipp_exit); ++ ++MODULE_LICENSE("BSD"); ++MODULE_AUTHOR("Michael Richardson <mcr@xelerance.com>"); ++MODULE_DESCRIPTION("OCF driver for hifn HIPP-I/II PCI crypto devices"); +diff -Nur linux-2.6.30.orig/crypto/ocf/hifn/hifnHIPPreg.h linux-2.6.30/crypto/ocf/hifn/hifnHIPPreg.h +--- linux-2.6.30.orig/crypto/ocf/hifn/hifnHIPPreg.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/hifn/hifnHIPPreg.h 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,46 @@ ++/*- ++ * Hifn HIPP-I/HIPP-II (7855/8155) driver. ++ * Copyright (c) 2006 Michael Richardson <mcr@xelerance.com> ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions ++ * are met: ++ * ++ * 1. Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * 2. Redistributions in binary form must reproduce the above copyright ++ * notice, this list of conditions and the following disclaimer in the ++ * documentation and/or other materials provided with the distribution. ++ * 3. The name of the author may not be used to endorse or promote products ++ * derived from this software without specific prior written permission. ++ * ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR ++ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES ++ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. ++ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, ++ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT ++ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, ++ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY ++ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT ++ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF ++ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. ++ * ++ * Effort sponsored by Hifn inc. ++ * ++ */ ++ ++#ifndef __HIFNHIPP_H__ ++#define __HIFNHIPP_H__ ++ ++/* ++ * PCI vendor and device identifiers ++ */ ++#define PCI_VENDOR_HIFN 0x13a3 /* Hifn */ ++#define PCI_PRODUCT_HIFN_6500 0x0006 /* 6500 */ ++#define PCI_PRODUCT_HIFN_7855 0x001f /* 7855 */ ++#define PCI_PRODUCT_HIFN_8155 0x999 /* XXX 8155 */ ++ ++#define HIPP_1_REVID 0x01 /* BOGUS */ ++ ++#endif /* __HIPP_H__ */ +diff -Nur linux-2.6.30.orig/crypto/ocf/hifn/hifnHIPPvar.h linux-2.6.30/crypto/ocf/hifn/hifnHIPPvar.h +--- linux-2.6.30.orig/crypto/ocf/hifn/hifnHIPPvar.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/hifn/hifnHIPPvar.h 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,93 @@ ++/* ++ * Hifn HIPP-I/HIPP-II (7855/8155) driver. ++ * Copyright (c) 2006 Michael Richardson <mcr@xelerance.com> * ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions ++ * are met: ++ * ++ * 1. Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * 2. Redistributions in binary form must reproduce the above copyright ++ * notice, this list of conditions and the following disclaimer in the ++ * documentation and/or other materials provided with the distribution. ++ * 3. The name of the author may not be used to endorse or promote products ++ * derived from this software without specific prior written permission. ++ * ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR ++ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES ++ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. ++ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, ++ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT ++ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, ++ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY ++ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT ++ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF ++ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. ++ * ++ * Effort sponsored by Hifn inc. ++ * ++ */ ++ ++#ifndef __HIFNHIPPVAR_H__ ++#define __HIFNHIPPVAR_H__ ++ ++#define HIPP_MAX_CHIPS 8 ++ ++/* ++ * Holds data specific to a single Hifn HIPP-I board. ++ */ ++struct hipp_softc { ++ softc_device_decl sc_dev; ++ ++ struct pci_dev *sc_pcidev; /* device backpointer */ ++ ocf_iomem_t sc_bar[5]; ++ caddr_t sc_barphy[5]; /* physical address */ ++ int sc_num; /* for multiple devs */ ++ spinlock_t sc_mtx; /* per-instance lock */ ++ int32_t sc_cid; ++ int sc_irq; ++ ++#if 0 ++ ++ u_int32_t sc_dmaier; ++ u_int32_t sc_drammodel; /* 1=dram, 0=sram */ ++ u_int32_t sc_pllconfig; /* 7954/7955/7956 PLL config */ ++ ++ struct hifn_dma *sc_dma; ++ dma_addr_t sc_dma_physaddr;/* physical address of sc_dma */ ++ ++ int sc_dmansegs; ++ int sc_maxses; ++ int sc_nsessions; ++ struct hifn_session *sc_sessions; ++ int sc_ramsize; ++ int sc_flags; ++#define HIFN_HAS_RNG 0x1 /* includes random number generator */ ++#define HIFN_HAS_PUBLIC 0x2 /* includes public key support */ ++#define HIFN_HAS_AES 0x4 /* includes AES support */ ++#define HIFN_IS_7811 0x8 /* Hifn 7811 part */ ++#define HIFN_IS_7956 0x10 /* Hifn 7956/7955 don't have SDRAM */ ++ ++ struct timer_list sc_tickto; /* for managing DMA */ ++ ++ int sc_rngfirst; ++ int sc_rnghz; /* RNG polling frequency */ ++ ++ int sc_c_busy; /* command ring busy */ ++ int sc_s_busy; /* source data ring busy */ ++ int sc_d_busy; /* destination data ring busy */ ++ int sc_r_busy; /* result ring busy */ ++ int sc_active; /* for initial countdown */ ++ int sc_needwakeup; /* ops q'd wating on resources */ ++ int sc_curbatch; /* # ops submitted w/o int */ ++ int sc_suspended; ++ struct miscdevice sc_miscdev; ++#endif ++}; ++ ++#define HIPP_LOCK(_sc) spin_lock_irqsave(&(_sc)->sc_mtx, l_flags) ++#define HIPP_UNLOCK(_sc) spin_unlock_irqrestore(&(_sc)->sc_mtx, l_flags) ++ ++#endif /* __HIFNHIPPVAR_H__ */ +diff -Nur linux-2.6.30.orig/crypto/ocf/hifn/Makefile linux-2.6.30/crypto/ocf/hifn/Makefile +--- linux-2.6.30.orig/crypto/ocf/hifn/Makefile 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/hifn/Makefile 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,13 @@ ++# for SGlinux builds ++-include $(ROOTDIR)/modules/.config ++ ++obj-$(CONFIG_OCF_HIFN) += hifn7751.o ++obj-$(CONFIG_OCF_HIFNHIPP) += hifnHIPP.o ++ ++obj ?= . ++EXTRA_CFLAGS += -I$(obj)/.. -I$(obj)/ ++ ++ifdef TOPDIR ++-include $(TOPDIR)/Rules.make ++endif ++ +diff -Nur linux-2.6.30.orig/crypto/ocf/ixp4xx/ixp4xx.c linux-2.6.30/crypto/ocf/ixp4xx/ixp4xx.c +--- linux-2.6.30.orig/crypto/ocf/ixp4xx/ixp4xx.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/ixp4xx/ixp4xx.c 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,1328 @@ ++/* ++ * An OCF module that uses Intels IXP CryptACC API to do the crypto. ++ * This driver requires the IXP400 Access Library that is available ++ * from Intel in order to operate (or compile). ++ * ++ * Written by David McCullough <david_mccullough@securecomputing.com> ++ * Copyright (C) 2006-2007 David McCullough ++ * Copyright (C) 2004-2005 Intel Corporation. ++ * ++ * LICENSE TERMS ++ * ++ * The free distribution and use of this software in both source and binary ++ * form is allowed (with or without changes) provided that: ++ * ++ * 1. distributions of this source code include the above copyright ++ * notice, this list of conditions and the following disclaimer; ++ * ++ * 2. distributions in binary form include the above copyright ++ * notice, this list of conditions and the following disclaimer ++ * in the documentation and/or other associated materials; ++ * ++ * 3. the copyright holder's name is not used to endorse products ++ * built using this software without specific written permission. ++ * ++ * ALTERNATIVELY, provided that this notice is retained in full, this product ++ * may be distributed under the terms of the GNU General Public License (GPL), ++ * in which case the provisions of the GPL apply INSTEAD OF those given above. ++ * ++ * DISCLAIMER ++ * ++ * This software is provided 'as is' with no explicit or implied warranties ++ * in respect of its properties, including, but not limited to, correctness ++ * and/or fitness for purpose. ++ */ ++ ++#ifndef AUTOCONF_INCLUDED ++#include <linux/config.h> ++#endif ++#include <linux/module.h> ++#include <linux/init.h> ++#include <linux/list.h> ++#include <linux/slab.h> ++#include <linux/sched.h> ++#include <linux/wait.h> ++#include <linux/crypto.h> ++#include <linux/interrupt.h> ++#include <asm/scatterlist.h> ++ ++#include <IxTypes.h> ++#include <IxOsBuffMgt.h> ++#include <IxNpeDl.h> ++#include <IxCryptoAcc.h> ++#include <IxQMgr.h> ++#include <IxOsServices.h> ++#include <IxOsCacheMMU.h> ++ ++#include <cryptodev.h> ++#include <uio.h> ++ ++#ifndef IX_MBUF_PRIV ++#define IX_MBUF_PRIV(x) ((x)->priv) ++#endif ++ ++struct ixp_data; ++ ++struct ixp_q { ++ struct list_head ixp_q_list; ++ struct ixp_data *ixp_q_data; ++ struct cryptop *ixp_q_crp; ++ struct cryptodesc *ixp_q_ccrd; ++ struct cryptodesc *ixp_q_acrd; ++ IX_MBUF ixp_q_mbuf; ++ UINT8 *ixp_hash_dest; /* Location for hash in client buffer */ ++ UINT8 *ixp_hash_src; /* Location of hash in internal buffer */ ++ unsigned char ixp_q_iv_data[IX_CRYPTO_ACC_MAX_CIPHER_IV_LENGTH]; ++ unsigned char *ixp_q_iv; ++}; ++ ++struct ixp_data { ++ int ixp_registered; /* is the context registered */ ++ int ixp_crd_flags; /* detect direction changes */ ++ ++ int ixp_cipher_alg; ++ int ixp_auth_alg; ++ ++ UINT32 ixp_ctx_id; ++ UINT32 ixp_hash_key_id; /* used when hashing */ ++ IxCryptoAccCtx ixp_ctx; ++ IX_MBUF ixp_pri_mbuf; ++ IX_MBUF ixp_sec_mbuf; ++ ++ struct work_struct ixp_pending_work; ++ struct work_struct ixp_registration_work; ++ struct list_head ixp_q; /* unprocessed requests */ ++}; ++ ++#ifdef __ixp46X ++ ++#define MAX_IOP_SIZE 64 /* words */ ++#define MAX_OOP_SIZE 128 ++ ++#define MAX_PARAMS 3 ++ ++struct ixp_pkq { ++ struct list_head pkq_list; ++ struct cryptkop *pkq_krp; ++ ++ IxCryptoAccPkeEauInOperands pkq_op; ++ IxCryptoAccPkeEauOpResult pkq_result; ++ ++ UINT32 pkq_ibuf0[MAX_IOP_SIZE]; ++ UINT32 pkq_ibuf1[MAX_IOP_SIZE]; ++ UINT32 pkq_ibuf2[MAX_IOP_SIZE]; ++ UINT32 pkq_obuf[MAX_OOP_SIZE]; ++}; ++ ++static LIST_HEAD(ixp_pkq); /* current PK wait list */ ++static struct ixp_pkq *ixp_pk_cur; ++static spinlock_t ixp_pkq_lock; ++ ++#endif /* __ixp46X */ ++ ++static int ixp_blocked = 0; ++ ++static int32_t ixp_id = -1; ++static struct ixp_data **ixp_sessions = NULL; ++static u_int32_t ixp_sesnum = 0; ++ ++static int ixp_process(device_t, struct cryptop *, int); ++static int ixp_newsession(device_t, u_int32_t *, struct cryptoini *); ++static int ixp_freesession(device_t, u_int64_t); ++#ifdef __ixp46X ++static int ixp_kprocess(device_t, struct cryptkop *krp, int hint); ++#endif ++ ++#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,20) ++static kmem_cache_t *qcache; ++#else ++static struct kmem_cache *qcache; ++#endif ++ ++#define debug ixp_debug ++static int ixp_debug = 0; ++module_param(ixp_debug, int, 0644); ++MODULE_PARM_DESC(ixp_debug, "Enable debug"); ++ ++static int ixp_init_crypto = 1; ++module_param(ixp_init_crypto, int, 0444); /* RO after load/boot */ ++MODULE_PARM_DESC(ixp_init_crypto, "Call ixCryptoAccInit (default is 1)"); ++ ++static void ixp_process_pending(void *arg); ++static void ixp_registration(void *arg); ++#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,20) ++static void ixp_process_pending_wq(struct work_struct *work); ++static void ixp_registration_wq(struct work_struct *work); ++#endif ++ ++/* ++ * dummy device structure ++ */ ++ ++static struct { ++ softc_device_decl sc_dev; ++} ixpdev; ++ ++static device_method_t ixp_methods = { ++ /* crypto device methods */ ++ DEVMETHOD(cryptodev_newsession, ixp_newsession), ++ DEVMETHOD(cryptodev_freesession,ixp_freesession), ++ DEVMETHOD(cryptodev_process, ixp_process), ++#ifdef __ixp46X ++ DEVMETHOD(cryptodev_kprocess, ixp_kprocess), ++#endif ++}; ++ ++/* ++ * Generate a new software session. ++ */ ++static int ++ixp_newsession(device_t dev, u_int32_t *sid, struct cryptoini *cri) ++{ ++ struct ixp_data *ixp; ++ u_int32_t i; ++#define AUTH_LEN(cri, def) \ ++ (cri->cri_mlen ? cri->cri_mlen : (def)) ++ ++ dprintk("%s():alg %d\n", __FUNCTION__,cri->cri_alg); ++ if (sid == NULL || cri == NULL) { ++ dprintk("%s,%d - EINVAL\n", __FILE__, __LINE__); ++ return EINVAL; ++ } ++ ++ if (ixp_sessions) { ++ for (i = 1; i < ixp_sesnum; i++) ++ if (ixp_sessions[i] == NULL) ++ break; ++ } else ++ i = 1; /* NB: to silence compiler warning */ ++ ++ if (ixp_sessions == NULL || i == ixp_sesnum) { ++ struct ixp_data **ixpd; ++ ++ if (ixp_sessions == NULL) { ++ i = 1; /* We leave ixp_sessions[0] empty */ ++ ixp_sesnum = CRYPTO_SW_SESSIONS; ++ } else ++ ixp_sesnum *= 2; ++ ++ ixpd = kmalloc(ixp_sesnum * sizeof(struct ixp_data *), SLAB_ATOMIC); ++ if (ixpd == NULL) { ++ /* Reset session number */ ++ if (ixp_sesnum == CRYPTO_SW_SESSIONS) ++ ixp_sesnum = 0; ++ else ++ ixp_sesnum /= 2; ++ dprintk("%s,%d: ENOBUFS\n", __FILE__, __LINE__); ++ return ENOBUFS; ++ } ++ memset(ixpd, 0, ixp_sesnum * sizeof(struct ixp_data *)); ++ ++ /* Copy existing sessions */ ++ if (ixp_sessions) { ++ memcpy(ixpd, ixp_sessions, ++ (ixp_sesnum / 2) * sizeof(struct ixp_data *)); ++ kfree(ixp_sessions); ++ } ++ ++ ixp_sessions = ixpd; ++ } ++ ++ ixp_sessions[i] = (struct ixp_data *) kmalloc(sizeof(struct ixp_data), ++ SLAB_ATOMIC); ++ if (ixp_sessions[i] == NULL) { ++ ixp_freesession(NULL, i); ++ dprintk("%s,%d: EINVAL\n", __FILE__, __LINE__); ++ return ENOBUFS; ++ } ++ ++ *sid = i; ++ ++ ixp = ixp_sessions[i]; ++ memset(ixp, 0, sizeof(*ixp)); ++ ++ ixp->ixp_cipher_alg = -1; ++ ixp->ixp_auth_alg = -1; ++ ixp->ixp_ctx_id = -1; ++ INIT_LIST_HEAD(&ixp->ixp_q); ++ ++ ixp->ixp_ctx.useDifferentSrcAndDestMbufs = 0; ++ ++ while (cri) { ++ switch (cri->cri_alg) { ++ case CRYPTO_DES_CBC: ++ ixp->ixp_cipher_alg = cri->cri_alg; ++ ixp->ixp_ctx.cipherCtx.cipherAlgo = IX_CRYPTO_ACC_CIPHER_DES; ++ ixp->ixp_ctx.cipherCtx.cipherMode = IX_CRYPTO_ACC_MODE_CBC; ++ ixp->ixp_ctx.cipherCtx.cipherKeyLen = (cri->cri_klen + 7) / 8; ++ ixp->ixp_ctx.cipherCtx.cipherBlockLen = IX_CRYPTO_ACC_DES_BLOCK_64; ++ ixp->ixp_ctx.cipherCtx.cipherInitialVectorLen = ++ IX_CRYPTO_ACC_DES_IV_64; ++ memcpy(ixp->ixp_ctx.cipherCtx.key.cipherKey, ++ cri->cri_key, (cri->cri_klen + 7) / 8); ++ break; ++ ++ case CRYPTO_3DES_CBC: ++ ixp->ixp_cipher_alg = cri->cri_alg; ++ ixp->ixp_ctx.cipherCtx.cipherAlgo = IX_CRYPTO_ACC_CIPHER_3DES; ++ ixp->ixp_ctx.cipherCtx.cipherMode = IX_CRYPTO_ACC_MODE_CBC; ++ ixp->ixp_ctx.cipherCtx.cipherKeyLen = (cri->cri_klen + 7) / 8; ++ ixp->ixp_ctx.cipherCtx.cipherBlockLen = IX_CRYPTO_ACC_DES_BLOCK_64; ++ ixp->ixp_ctx.cipherCtx.cipherInitialVectorLen = ++ IX_CRYPTO_ACC_DES_IV_64; ++ memcpy(ixp->ixp_ctx.cipherCtx.key.cipherKey, ++ cri->cri_key, (cri->cri_klen + 7) / 8); ++ break; ++ ++ case CRYPTO_RIJNDAEL128_CBC: ++ ixp->ixp_cipher_alg = cri->cri_alg; ++ ixp->ixp_ctx.cipherCtx.cipherAlgo = IX_CRYPTO_ACC_CIPHER_AES; ++ ixp->ixp_ctx.cipherCtx.cipherMode = IX_CRYPTO_ACC_MODE_CBC; ++ ixp->ixp_ctx.cipherCtx.cipherKeyLen = (cri->cri_klen + 7) / 8; ++ ixp->ixp_ctx.cipherCtx.cipherBlockLen = 16; ++ ixp->ixp_ctx.cipherCtx.cipherInitialVectorLen = 16; ++ memcpy(ixp->ixp_ctx.cipherCtx.key.cipherKey, ++ cri->cri_key, (cri->cri_klen + 7) / 8); ++ break; ++ ++ case CRYPTO_MD5: ++ case CRYPTO_MD5_HMAC: ++ ixp->ixp_auth_alg = cri->cri_alg; ++ ixp->ixp_ctx.authCtx.authAlgo = IX_CRYPTO_ACC_AUTH_MD5; ++ ixp->ixp_ctx.authCtx.authDigestLen = AUTH_LEN(cri, MD5_HASH_LEN); ++ ixp->ixp_ctx.authCtx.aadLen = 0; ++ /* Only MD5_HMAC needs a key */ ++ if (cri->cri_alg == CRYPTO_MD5_HMAC) { ++ ixp->ixp_ctx.authCtx.authKeyLen = (cri->cri_klen + 7) / 8; ++ if (ixp->ixp_ctx.authCtx.authKeyLen > ++ sizeof(ixp->ixp_ctx.authCtx.key.authKey)) { ++ printk( ++ "ixp4xx: Invalid key length for MD5_HMAC - %d bits\n", ++ cri->cri_klen); ++ ixp_freesession(NULL, i); ++ return EINVAL; ++ } ++ memcpy(ixp->ixp_ctx.authCtx.key.authKey, ++ cri->cri_key, (cri->cri_klen + 7) / 8); ++ } ++ break; ++ ++ case CRYPTO_SHA1: ++ case CRYPTO_SHA1_HMAC: ++ ixp->ixp_auth_alg = cri->cri_alg; ++ ixp->ixp_ctx.authCtx.authAlgo = IX_CRYPTO_ACC_AUTH_SHA1; ++ ixp->ixp_ctx.authCtx.authDigestLen = AUTH_LEN(cri, SHA1_HASH_LEN); ++ ixp->ixp_ctx.authCtx.aadLen = 0; ++ /* Only SHA1_HMAC needs a key */ ++ if (cri->cri_alg == CRYPTO_SHA1_HMAC) { ++ ixp->ixp_ctx.authCtx.authKeyLen = (cri->cri_klen + 7) / 8; ++ if (ixp->ixp_ctx.authCtx.authKeyLen > ++ sizeof(ixp->ixp_ctx.authCtx.key.authKey)) { ++ printk( ++ "ixp4xx: Invalid key length for SHA1_HMAC - %d bits\n", ++ cri->cri_klen); ++ ixp_freesession(NULL, i); ++ return EINVAL; ++ } ++ memcpy(ixp->ixp_ctx.authCtx.key.authKey, ++ cri->cri_key, (cri->cri_klen + 7) / 8); ++ } ++ break; ++ ++ default: ++ printk("ixp: unknown algo 0x%x\n", cri->cri_alg); ++ ixp_freesession(NULL, i); ++ return EINVAL; ++ } ++ cri = cri->cri_next; ++ } ++ ++#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,20) ++ INIT_WORK(&ixp->ixp_pending_work, ixp_process_pending_wq); ++ INIT_WORK(&ixp->ixp_registration_work, ixp_registration_wq); ++#else ++ INIT_WORK(&ixp->ixp_pending_work, ixp_process_pending, ixp); ++ INIT_WORK(&ixp->ixp_registration_work, ixp_registration, ixp); ++#endif ++ ++ return 0; ++} ++ ++ ++/* ++ * Free a session. ++ */ ++static int ++ixp_freesession(device_t dev, u_int64_t tid) ++{ ++ u_int32_t sid = CRYPTO_SESID2LID(tid); ++ ++ dprintk("%s()\n", __FUNCTION__); ++ if (sid > ixp_sesnum || ixp_sessions == NULL || ++ ixp_sessions[sid] == NULL) { ++ dprintk("%s,%d: EINVAL\n", __FILE__, __LINE__); ++ return EINVAL; ++ } ++ ++ /* Silently accept and return */ ++ if (sid == 0) ++ return 0; ++ ++ if (ixp_sessions[sid]) { ++ if (ixp_sessions[sid]->ixp_ctx_id != -1) { ++ ixCryptoAccCtxUnregister(ixp_sessions[sid]->ixp_ctx_id); ++ ixp_sessions[sid]->ixp_ctx_id = -1; ++ } ++ ++ flush_scheduled_work(); ++ ++ kfree(ixp_sessions[sid]); ++ } ++ ixp_sessions[sid] = NULL; ++ if (ixp_blocked) { ++ ixp_blocked = 0; ++ crypto_unblock(ixp_id, CRYPTO_SYMQ); ++ } ++ return 0; ++} ++ ++ ++/* ++ * callback for when hash processing is complete ++ */ ++ ++static void ++ixp_hash_perform_cb( ++ UINT32 hash_key_id, ++ IX_MBUF *bufp, ++ IxCryptoAccStatus status) ++{ ++ struct ixp_q *q; ++ ++ dprintk("%s(%u, %p, 0x%x)\n", __FUNCTION__, hash_key_id, bufp, status); ++ ++ if (bufp == NULL) { ++ printk("ixp: NULL buf in %s\n", __FUNCTION__); ++ return; ++ } ++ ++ q = IX_MBUF_PRIV(bufp); ++ if (q == NULL) { ++ printk("ixp: NULL priv in %s\n", __FUNCTION__); ++ return; ++ } ++ ++ if (status == IX_CRYPTO_ACC_STATUS_SUCCESS) { ++ /* On success, need to copy hash back into original client buffer */ ++ memcpy(q->ixp_hash_dest, q->ixp_hash_src, ++ (q->ixp_q_data->ixp_auth_alg == CRYPTO_SHA1) ? ++ SHA1_HASH_LEN : MD5_HASH_LEN); ++ } ++ else { ++ printk("ixp: hash perform failed status=%d\n", status); ++ q->ixp_q_crp->crp_etype = EINVAL; ++ } ++ ++ /* Free internal buffer used for hashing */ ++ kfree(IX_MBUF_MDATA(&q->ixp_q_mbuf)); ++ ++ crypto_done(q->ixp_q_crp); ++ kmem_cache_free(qcache, q); ++} ++ ++/* ++ * setup a request and perform it ++ */ ++static void ++ixp_q_process(struct ixp_q *q) ++{ ++ IxCryptoAccStatus status; ++ struct ixp_data *ixp = q->ixp_q_data; ++ int auth_off = 0; ++ int auth_len = 0; ++ int crypt_off = 0; ++ int crypt_len = 0; ++ int icv_off = 0; ++ char *crypt_func; ++ ++ dprintk("%s(%p)\n", __FUNCTION__, q); ++ ++ if (q->ixp_q_ccrd) { ++ if (q->ixp_q_ccrd->crd_flags & CRD_F_IV_EXPLICIT) { ++ q->ixp_q_iv = q->ixp_q_ccrd->crd_iv; ++ } else { ++ q->ixp_q_iv = q->ixp_q_iv_data; ++ crypto_copydata(q->ixp_q_crp->crp_flags, q->ixp_q_crp->crp_buf, ++ q->ixp_q_ccrd->crd_inject, ++ ixp->ixp_ctx.cipherCtx.cipherInitialVectorLen, ++ (caddr_t) q->ixp_q_iv); ++ } ++ ++ if (q->ixp_q_acrd) { ++ auth_off = q->ixp_q_acrd->crd_skip; ++ auth_len = q->ixp_q_acrd->crd_len; ++ icv_off = q->ixp_q_acrd->crd_inject; ++ } ++ ++ crypt_off = q->ixp_q_ccrd->crd_skip; ++ crypt_len = q->ixp_q_ccrd->crd_len; ++ } else { /* if (q->ixp_q_acrd) */ ++ auth_off = q->ixp_q_acrd->crd_skip; ++ auth_len = q->ixp_q_acrd->crd_len; ++ icv_off = q->ixp_q_acrd->crd_inject; ++ } ++ ++ if (q->ixp_q_crp->crp_flags & CRYPTO_F_SKBUF) { ++ struct sk_buff *skb = (struct sk_buff *) q->ixp_q_crp->crp_buf; ++ if (skb_shinfo(skb)->nr_frags) { ++ /* ++ * DAVIDM fix this limitation one day by using ++ * a buffer pool and chaining, it is not currently ++ * needed for current user/kernel space acceleration ++ */ ++ printk("ixp: Cannot handle fragmented skb's yet !\n"); ++ q->ixp_q_crp->crp_etype = ENOENT; ++ goto done; ++ } ++ IX_MBUF_MLEN(&q->ixp_q_mbuf) = ++ IX_MBUF_PKT_LEN(&q->ixp_q_mbuf) = skb->len; ++ IX_MBUF_MDATA(&q->ixp_q_mbuf) = skb->data; ++ } else if (q->ixp_q_crp->crp_flags & CRYPTO_F_IOV) { ++ struct uio *uiop = (struct uio *) q->ixp_q_crp->crp_buf; ++ if (uiop->uio_iovcnt != 1) { ++ /* ++ * DAVIDM fix this limitation one day by using ++ * a buffer pool and chaining, it is not currently ++ * needed for current user/kernel space acceleration ++ */ ++ printk("ixp: Cannot handle more than 1 iovec yet !\n"); ++ q->ixp_q_crp->crp_etype = ENOENT; ++ goto done; ++ } ++ IX_MBUF_MLEN(&q->ixp_q_mbuf) = ++ IX_MBUF_PKT_LEN(&q->ixp_q_mbuf) = uiop->uio_iov[0].iov_len; ++ IX_MBUF_MDATA(&q->ixp_q_mbuf) = uiop->uio_iov[0].iov_base; ++ } else /* contig buffer */ { ++ IX_MBUF_MLEN(&q->ixp_q_mbuf) = ++ IX_MBUF_PKT_LEN(&q->ixp_q_mbuf) = q->ixp_q_crp->crp_ilen; ++ IX_MBUF_MDATA(&q->ixp_q_mbuf) = q->ixp_q_crp->crp_buf; ++ } ++ ++ IX_MBUF_PRIV(&q->ixp_q_mbuf) = q; ++ ++ if (ixp->ixp_auth_alg == CRYPTO_SHA1 || ixp->ixp_auth_alg == CRYPTO_MD5) { ++ /* ++ * For SHA1 and MD5 hash, need to create an internal buffer that is big ++ * enough to hold the original data + the appropriate padding for the ++ * hash algorithm. ++ */ ++ UINT8 *tbuf = NULL; ++ ++ IX_MBUF_MLEN(&q->ixp_q_mbuf) = IX_MBUF_PKT_LEN(&q->ixp_q_mbuf) = ++ ((IX_MBUF_MLEN(&q->ixp_q_mbuf) * 8) + 72 + 511) / 8; ++ tbuf = kmalloc(IX_MBUF_MLEN(&q->ixp_q_mbuf), SLAB_ATOMIC); ++ ++ if (IX_MBUF_MDATA(&q->ixp_q_mbuf) == NULL) { ++ printk("ixp: kmalloc(%u, SLAB_ATOMIC) failed\n", ++ IX_MBUF_MLEN(&q->ixp_q_mbuf)); ++ q->ixp_q_crp->crp_etype = ENOMEM; ++ goto done; ++ } ++ memcpy(tbuf, &(IX_MBUF_MDATA(&q->ixp_q_mbuf))[auth_off], auth_len); ++ ++ /* Set location in client buffer to copy hash into */ ++ q->ixp_hash_dest = ++ &(IX_MBUF_MDATA(&q->ixp_q_mbuf))[auth_off + auth_len]; ++ ++ IX_MBUF_MDATA(&q->ixp_q_mbuf) = tbuf; ++ ++ /* Set location in internal buffer for where hash starts */ ++ q->ixp_hash_src = &(IX_MBUF_MDATA(&q->ixp_q_mbuf))[auth_len]; ++ ++ crypt_func = "ixCryptoAccHashPerform"; ++ status = ixCryptoAccHashPerform(ixp->ixp_ctx.authCtx.authAlgo, ++ &q->ixp_q_mbuf, ixp_hash_perform_cb, 0, auth_len, auth_len, ++ &ixp->ixp_hash_key_id); ++ } ++ else { ++ crypt_func = "ixCryptoAccAuthCryptPerform"; ++ status = ixCryptoAccAuthCryptPerform(ixp->ixp_ctx_id, &q->ixp_q_mbuf, ++ NULL, auth_off, auth_len, crypt_off, crypt_len, icv_off, ++ q->ixp_q_iv); ++ } ++ ++ if (IX_CRYPTO_ACC_STATUS_SUCCESS == status) ++ return; ++ ++ if (IX_CRYPTO_ACC_STATUS_QUEUE_FULL == status) { ++ q->ixp_q_crp->crp_etype = ENOMEM; ++ goto done; ++ } ++ ++ printk("ixp: %s failed %u\n", crypt_func, status); ++ q->ixp_q_crp->crp_etype = EINVAL; ++ ++done: ++ crypto_done(q->ixp_q_crp); ++ kmem_cache_free(qcache, q); ++} ++ ++ ++/* ++ * because we cannot process the Q from the Register callback ++ * we do it here on a task Q. ++ */ ++ ++static void ++ixp_process_pending(void *arg) ++{ ++ struct ixp_data *ixp = arg; ++ struct ixp_q *q = NULL; ++ ++ dprintk("%s(%p)\n", __FUNCTION__, arg); ++ ++ if (!ixp) ++ return; ++ ++ while (!list_empty(&ixp->ixp_q)) { ++ q = list_entry(ixp->ixp_q.next, struct ixp_q, ixp_q_list); ++ list_del(&q->ixp_q_list); ++ ixp_q_process(q); ++ } ++} ++ ++#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,20) ++static void ++ixp_process_pending_wq(struct work_struct *work) ++{ ++ struct ixp_data *ixp = container_of(work, struct ixp_data, ++ ixp_pending_work); ++ ixp_process_pending(ixp); ++} ++#endif ++ ++/* ++ * callback for when context registration is complete ++ */ ++ ++static void ++ixp_register_cb(UINT32 ctx_id, IX_MBUF *bufp, IxCryptoAccStatus status) ++{ ++ int i; ++ struct ixp_data *ixp; ++ struct ixp_q *q; ++ ++ dprintk("%s(%d, %p, %d)\n", __FUNCTION__, ctx_id, bufp, status); ++ ++ /* ++ * free any buffer passed in to this routine ++ */ ++ if (bufp) { ++ IX_MBUF_MLEN(bufp) = IX_MBUF_PKT_LEN(bufp) = 0; ++ kfree(IX_MBUF_MDATA(bufp)); ++ IX_MBUF_MDATA(bufp) = NULL; ++ } ++ ++ for (i = 0; i < ixp_sesnum; i++) { ++ ixp = ixp_sessions[i]; ++ if (ixp && ixp->ixp_ctx_id == ctx_id) ++ break; ++ } ++ if (i >= ixp_sesnum) { ++ printk("ixp: invalid context id %d\n", ctx_id); ++ return; ++ } ++ ++ if (IX_CRYPTO_ACC_STATUS_WAIT == status) { ++ /* this is normal to free the first of two buffers */ ++ dprintk("ixp: register not finished yet.\n"); ++ return; ++ } ++ ++ if (IX_CRYPTO_ACC_STATUS_SUCCESS != status) { ++ printk("ixp: register failed 0x%x\n", status); ++ while (!list_empty(&ixp->ixp_q)) { ++ q = list_entry(ixp->ixp_q.next, struct ixp_q, ixp_q_list); ++ list_del(&q->ixp_q_list); ++ q->ixp_q_crp->crp_etype = EINVAL; ++ crypto_done(q->ixp_q_crp); ++ kmem_cache_free(qcache, q); ++ } ++ return; ++ } ++ ++ /* ++ * we are now registered, we cannot start processing the Q here ++ * or we get strange errors with AES (DES/3DES seem to be ok). ++ */ ++ ixp->ixp_registered = 1; ++ schedule_work(&ixp->ixp_pending_work); ++} ++ ++ ++/* ++ * callback for when data processing is complete ++ */ ++ ++static void ++ixp_perform_cb( ++ UINT32 ctx_id, ++ IX_MBUF *sbufp, ++ IX_MBUF *dbufp, ++ IxCryptoAccStatus status) ++{ ++ struct ixp_q *q; ++ ++ dprintk("%s(%d, %p, %p, 0x%x)\n", __FUNCTION__, ctx_id, sbufp, ++ dbufp, status); ++ ++ if (sbufp == NULL) { ++ printk("ixp: NULL sbuf in ixp_perform_cb\n"); ++ return; ++ } ++ ++ q = IX_MBUF_PRIV(sbufp); ++ if (q == NULL) { ++ printk("ixp: NULL priv in ixp_perform_cb\n"); ++ return; ++ } ++ ++ if (status != IX_CRYPTO_ACC_STATUS_SUCCESS) { ++ printk("ixp: perform failed status=%d\n", status); ++ q->ixp_q_crp->crp_etype = EINVAL; ++ } ++ ++ crypto_done(q->ixp_q_crp); ++ kmem_cache_free(qcache, q); ++} ++ ++ ++/* ++ * registration is not callable at IRQ time, so we defer ++ * to a task queue, this routines completes the registration for us ++ * when the task queue runs ++ * ++ * Unfortunately this means we cannot tell OCF that the driver is blocked, ++ * we do that on the next request. ++ */ ++ ++static void ++ixp_registration(void *arg) ++{ ++ struct ixp_data *ixp = arg; ++ struct ixp_q *q = NULL; ++ IX_MBUF *pri = NULL, *sec = NULL; ++ int status = IX_CRYPTO_ACC_STATUS_SUCCESS; ++ ++ if (!ixp) { ++ printk("ixp: ixp_registration with no arg\n"); ++ return; ++ } ++ ++ if (ixp->ixp_ctx_id != -1) { ++ ixCryptoAccCtxUnregister(ixp->ixp_ctx_id); ++ ixp->ixp_ctx_id = -1; ++ } ++ ++ if (list_empty(&ixp->ixp_q)) { ++ printk("ixp: ixp_registration with no Q\n"); ++ return; ++ } ++ ++ /* ++ * setup the primary and secondary buffers ++ */ ++ q = list_entry(ixp->ixp_q.next, struct ixp_q, ixp_q_list); ++ if (q->ixp_q_acrd) { ++ pri = &ixp->ixp_pri_mbuf; ++ sec = &ixp->ixp_sec_mbuf; ++ IX_MBUF_MLEN(pri) = IX_MBUF_PKT_LEN(pri) = 128; ++ IX_MBUF_MDATA(pri) = (unsigned char *) kmalloc(128, SLAB_ATOMIC); ++ IX_MBUF_MLEN(sec) = IX_MBUF_PKT_LEN(sec) = 128; ++ IX_MBUF_MDATA(sec) = (unsigned char *) kmalloc(128, SLAB_ATOMIC); ++ } ++ ++ /* Only need to register if a crypt op or HMAC op */ ++ if (!(ixp->ixp_auth_alg == CRYPTO_SHA1 || ++ ixp->ixp_auth_alg == CRYPTO_MD5)) { ++ status = ixCryptoAccCtxRegister( ++ &ixp->ixp_ctx, ++ pri, sec, ++ ixp_register_cb, ++ ixp_perform_cb, ++ &ixp->ixp_ctx_id); ++ } ++ else { ++ /* Otherwise we start processing pending q */ ++ schedule_work(&ixp->ixp_pending_work); ++ } ++ ++ if (IX_CRYPTO_ACC_STATUS_SUCCESS == status) ++ return; ++ ++ if (IX_CRYPTO_ACC_STATUS_EXCEED_MAX_TUNNELS == status) { ++ printk("ixp: ixCryptoAccCtxRegister failed (out of tunnels)\n"); ++ ixp_blocked = 1; ++ /* perhaps we should return EGAIN on queued ops ? */ ++ return; ++ } ++ ++ printk("ixp: ixCryptoAccCtxRegister failed %d\n", status); ++ ixp->ixp_ctx_id = -1; ++ ++ /* ++ * everything waiting is toasted ++ */ ++ while (!list_empty(&ixp->ixp_q)) { ++ q = list_entry(ixp->ixp_q.next, struct ixp_q, ixp_q_list); ++ list_del(&q->ixp_q_list); ++ q->ixp_q_crp->crp_etype = ENOENT; ++ crypto_done(q->ixp_q_crp); ++ kmem_cache_free(qcache, q); ++ } ++} ++ ++#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,20) ++static void ++ixp_registration_wq(struct work_struct *work) ++{ ++ struct ixp_data *ixp = container_of(work, struct ixp_data, ++ ixp_registration_work); ++ ixp_registration(ixp); ++} ++#endif ++ ++/* ++ * Process a request. ++ */ ++static int ++ixp_process(device_t dev, struct cryptop *crp, int hint) ++{ ++ struct ixp_data *ixp; ++ unsigned int lid; ++ struct ixp_q *q = NULL; ++ int status; ++ ++ dprintk("%s()\n", __FUNCTION__); ++ ++ /* Sanity check */ ++ if (crp == NULL) { ++ dprintk("%s,%d: EINVAL\n", __FILE__, __LINE__); ++ return EINVAL; ++ } ++ ++ crp->crp_etype = 0; ++ ++ if (ixp_blocked) ++ return ERESTART; ++ ++ if (crp->crp_desc == NULL || crp->crp_buf == NULL) { ++ dprintk("%s,%d: EINVAL\n", __FILE__, __LINE__); ++ crp->crp_etype = EINVAL; ++ goto done; ++ } ++ ++ /* ++ * find the session we are using ++ */ ++ ++ lid = crp->crp_sid & 0xffffffff; ++ if (lid >= ixp_sesnum || lid == 0 || ixp_sessions == NULL || ++ ixp_sessions[lid] == NULL) { ++ crp->crp_etype = ENOENT; ++ dprintk("%s,%d: ENOENT\n", __FILE__, __LINE__); ++ goto done; ++ } ++ ixp = ixp_sessions[lid]; ++ ++ /* ++ * setup a new request ready for queuing ++ */ ++ q = kmem_cache_alloc(qcache, SLAB_ATOMIC); ++ if (q == NULL) { ++ dprintk("%s,%d: ENOMEM\n", __FILE__, __LINE__); ++ crp->crp_etype = ENOMEM; ++ goto done; ++ } ++ /* ++ * save some cycles by only zeroing the important bits ++ */ ++ memset(&q->ixp_q_mbuf, 0, sizeof(q->ixp_q_mbuf)); ++ q->ixp_q_ccrd = NULL; ++ q->ixp_q_acrd = NULL; ++ q->ixp_q_crp = crp; ++ q->ixp_q_data = ixp; ++ ++ /* ++ * point the cipher and auth descriptors appropriately ++ * check that we have something to do ++ */ ++ if (crp->crp_desc->crd_alg == ixp->ixp_cipher_alg) ++ q->ixp_q_ccrd = crp->crp_desc; ++ else if (crp->crp_desc->crd_alg == ixp->ixp_auth_alg) ++ q->ixp_q_acrd = crp->crp_desc; ++ else { ++ crp->crp_etype = ENOENT; ++ dprintk("%s,%d: bad desc match: ENOENT\n", __FILE__, __LINE__); ++ goto done; ++ } ++ if (crp->crp_desc->crd_next) { ++ if (crp->crp_desc->crd_next->crd_alg == ixp->ixp_cipher_alg) ++ q->ixp_q_ccrd = crp->crp_desc->crd_next; ++ else if (crp->crp_desc->crd_next->crd_alg == ixp->ixp_auth_alg) ++ q->ixp_q_acrd = crp->crp_desc->crd_next; ++ else { ++ crp->crp_etype = ENOENT; ++ dprintk("%s,%d: bad desc match: ENOENT\n", __FILE__, __LINE__); ++ goto done; ++ } ++ } ++ ++ /* ++ * If there is a direction change for this context then we mark it as ++ * unregistered and re-register is for the new direction. This is not ++ * a very expensive operation and currently only tends to happen when ++ * user-space application are doing benchmarks ++ * ++ * DM - we should be checking for pending requests before unregistering. ++ */ ++ if (q->ixp_q_ccrd && ixp->ixp_registered && ++ ixp->ixp_crd_flags != (q->ixp_q_ccrd->crd_flags & CRD_F_ENCRYPT)) { ++ dprintk("%s - detected direction change on session\n", __FUNCTION__); ++ ixp->ixp_registered = 0; ++ } ++ ++ /* ++ * if we are registered, call straight into the perform code ++ */ ++ if (ixp->ixp_registered) { ++ ixp_q_process(q); ++ return 0; ++ } ++ ++ /* ++ * the only part of the context not set in newsession is the direction ++ * dependent parts ++ */ ++ if (q->ixp_q_ccrd) { ++ ixp->ixp_crd_flags = (q->ixp_q_ccrd->crd_flags & CRD_F_ENCRYPT); ++ if (q->ixp_q_ccrd->crd_flags & CRD_F_ENCRYPT) { ++ ixp->ixp_ctx.operation = q->ixp_q_acrd ? ++ IX_CRYPTO_ACC_OP_ENCRYPT_AUTH : IX_CRYPTO_ACC_OP_ENCRYPT; ++ } else { ++ ixp->ixp_ctx.operation = q->ixp_q_acrd ? ++ IX_CRYPTO_ACC_OP_AUTH_DECRYPT : IX_CRYPTO_ACC_OP_DECRYPT; ++ } ++ } else { ++ /* q->ixp_q_acrd must be set if we are here */ ++ ixp->ixp_ctx.operation = IX_CRYPTO_ACC_OP_AUTH_CALC; ++ } ++ ++ status = list_empty(&ixp->ixp_q); ++ list_add_tail(&q->ixp_q_list, &ixp->ixp_q); ++ if (status) ++ schedule_work(&ixp->ixp_registration_work); ++ return 0; ++ ++done: ++ if (q) ++ kmem_cache_free(qcache, q); ++ crypto_done(crp); ++ return 0; ++} ++ ++ ++#ifdef __ixp46X ++/* ++ * key processing support for the ixp465 ++ */ ++ ++ ++/* ++ * copy a BN (LE) into a buffer (BE) an fill out the op appropriately ++ * assume zeroed and only copy bits that are significant ++ */ ++ ++static int ++ixp_copy_ibuf(struct crparam *p, IxCryptoAccPkeEauOperand *op, UINT32 *buf) ++{ ++ unsigned char *src = (unsigned char *) p->crp_p; ++ unsigned char *dst; ++ int len, bits = p->crp_nbits; ++ ++ dprintk("%s()\n", __FUNCTION__); ++ ++ if (bits > MAX_IOP_SIZE * sizeof(UINT32) * 8) { ++ dprintk("%s - ibuf too big (%d > %d)\n", __FUNCTION__, ++ bits, MAX_IOP_SIZE * sizeof(UINT32) * 8); ++ return -1; ++ } ++ ++ len = (bits + 31) / 32; /* the number UINT32's needed */ ++ ++ dst = (unsigned char *) &buf[len]; ++ dst--; ++ ++ while (bits > 0) { ++ *dst-- = *src++; ++ bits -= 8; ++ } ++ ++#if 0 /* no need to zero remaining bits as it is done during request alloc */ ++ while (dst > (unsigned char *) buf) ++ *dst-- = '\0'; ++#endif ++ ++ op->pData = buf; ++ op->dataLen = len; ++ return 0; ++} ++ ++/* ++ * copy out the result, be as forgiving as we can about small output buffers ++ */ ++ ++static int ++ixp_copy_obuf(struct crparam *p, IxCryptoAccPkeEauOpResult *op, UINT32 *buf) ++{ ++ unsigned char *dst = (unsigned char *) p->crp_p; ++ unsigned char *src = (unsigned char *) buf; ++ int len, z, bits = p->crp_nbits; ++ ++ dprintk("%s()\n", __FUNCTION__); ++ ++ len = op->dataLen * sizeof(UINT32); ++ ++ /* skip leading zeroes to be small buffer friendly */ ++ z = 0; ++ while (z < len && src[z] == '\0') ++ z++; ++ ++ src += len; ++ src--; ++ len -= z; ++ ++ while (len > 0 && bits > 0) { ++ *dst++ = *src--; ++ len--; ++ bits -= 8; ++ } ++ ++ while (bits > 0) { ++ *dst++ = '\0'; ++ bits -= 8; ++ } ++ ++ if (len > 0) { ++ dprintk("%s - obuf is %d (z=%d, ob=%d) bytes too small\n", ++ __FUNCTION__, len, z, p->crp_nbits / 8); ++ return -1; ++ } ++ ++ return 0; ++} ++ ++ ++/* ++ * the parameter offsets for exp_mod ++ */ ++ ++#define IXP_PARAM_BASE 0 ++#define IXP_PARAM_EXP 1 ++#define IXP_PARAM_MOD 2 ++#define IXP_PARAM_RES 3 ++ ++/* ++ * key processing complete callback, is also used to start processing ++ * by passing a NULL for pResult ++ */ ++ ++static void ++ixp_kperform_cb( ++ IxCryptoAccPkeEauOperation operation, ++ IxCryptoAccPkeEauOpResult *pResult, ++ BOOL carryOrBorrow, ++ IxCryptoAccStatus status) ++{ ++ struct ixp_pkq *q, *tmp; ++ unsigned long flags; ++ ++ dprintk("%s(0x%x, %p, %d, 0x%x)\n", __FUNCTION__, operation, pResult, ++ carryOrBorrow, status); ++ ++ /* handle a completed request */ ++ if (pResult) { ++ if (ixp_pk_cur && &ixp_pk_cur->pkq_result == pResult) { ++ q = ixp_pk_cur; ++ if (status != IX_CRYPTO_ACC_STATUS_SUCCESS) { ++ dprintk("%s() - op failed 0x%x\n", __FUNCTION__, status); ++ q->pkq_krp->krp_status = ERANGE; /* could do better */ ++ } else { ++ /* copy out the result */ ++ if (ixp_copy_obuf(&q->pkq_krp->krp_param[IXP_PARAM_RES], ++ &q->pkq_result, q->pkq_obuf)) ++ q->pkq_krp->krp_status = ERANGE; ++ } ++ crypto_kdone(q->pkq_krp); ++ kfree(q); ++ ixp_pk_cur = NULL; ++ } else ++ printk("%s - callback with invalid result pointer\n", __FUNCTION__); ++ } ++ ++ spin_lock_irqsave(&ixp_pkq_lock, flags); ++ if (ixp_pk_cur || list_empty(&ixp_pkq)) { ++ spin_unlock_irqrestore(&ixp_pkq_lock, flags); ++ return; ++ } ++ ++ list_for_each_entry_safe(q, tmp, &ixp_pkq, pkq_list) { ++ ++ list_del(&q->pkq_list); ++ ixp_pk_cur = q; ++ ++ spin_unlock_irqrestore(&ixp_pkq_lock, flags); ++ ++ status = ixCryptoAccPkeEauPerform( ++ IX_CRYPTO_ACC_OP_EAU_MOD_EXP, ++ &q->pkq_op, ++ ixp_kperform_cb, ++ &q->pkq_result); ++ ++ if (status == IX_CRYPTO_ACC_STATUS_SUCCESS) { ++ dprintk("%s() - ixCryptoAccPkeEauPerform SUCCESS\n", __FUNCTION__); ++ return; /* callback will return here for callback */ ++ } else if (status == IX_CRYPTO_ACC_STATUS_RETRY) { ++ printk("%s() - ixCryptoAccPkeEauPerform RETRY\n", __FUNCTION__); ++ } else { ++ printk("%s() - ixCryptoAccPkeEauPerform failed %d\n", ++ __FUNCTION__, status); ++ } ++ q->pkq_krp->krp_status = ERANGE; /* could do better */ ++ crypto_kdone(q->pkq_krp); ++ kfree(q); ++ spin_lock_irqsave(&ixp_pkq_lock, flags); ++ } ++ spin_unlock_irqrestore(&ixp_pkq_lock, flags); ++} ++ ++ ++static int ++ixp_kprocess(device_t dev, struct cryptkop *krp, int hint) ++{ ++ struct ixp_pkq *q; ++ int rc = 0; ++ unsigned long flags; ++ ++ dprintk("%s l1=%d l2=%d l3=%d l4=%d\n", __FUNCTION__, ++ krp->krp_param[IXP_PARAM_BASE].crp_nbits, ++ krp->krp_param[IXP_PARAM_EXP].crp_nbits, ++ krp->krp_param[IXP_PARAM_MOD].crp_nbits, ++ krp->krp_param[IXP_PARAM_RES].crp_nbits); ++ ++ ++ if (krp->krp_op != CRK_MOD_EXP) { ++ krp->krp_status = EOPNOTSUPP; ++ goto err; ++ } ++ ++ q = (struct ixp_pkq *) kmalloc(sizeof(*q), GFP_KERNEL); ++ if (q == NULL) { ++ krp->krp_status = ENOMEM; ++ goto err; ++ } ++ ++ /* ++ * The PKE engine does not appear to zero the output buffer ++ * appropriately, so we need to do it all here. ++ */ ++ memset(q, 0, sizeof(*q)); ++ ++ q->pkq_krp = krp; ++ INIT_LIST_HEAD(&q->pkq_list); ++ ++ if (ixp_copy_ibuf(&krp->krp_param[IXP_PARAM_BASE], &q->pkq_op.modExpOpr.M, ++ q->pkq_ibuf0)) ++ rc = 1; ++ if (!rc && ixp_copy_ibuf(&krp->krp_param[IXP_PARAM_EXP], ++ &q->pkq_op.modExpOpr.e, q->pkq_ibuf1)) ++ rc = 2; ++ if (!rc && ixp_copy_ibuf(&krp->krp_param[IXP_PARAM_MOD], ++ &q->pkq_op.modExpOpr.N, q->pkq_ibuf2)) ++ rc = 3; ++ ++ if (rc) { ++ kfree(q); ++ krp->krp_status = ERANGE; ++ goto err; ++ } ++ ++ q->pkq_result.pData = q->pkq_obuf; ++ q->pkq_result.dataLen = ++ (krp->krp_param[IXP_PARAM_RES].crp_nbits + 31) / 32; ++ ++ spin_lock_irqsave(&ixp_pkq_lock, flags); ++ list_add_tail(&q->pkq_list, &ixp_pkq); ++ spin_unlock_irqrestore(&ixp_pkq_lock, flags); ++ ++ if (!ixp_pk_cur) ++ ixp_kperform_cb(0, NULL, 0, 0); ++ return (0); ++ ++err: ++ crypto_kdone(krp); ++ return (0); ++} ++ ++ ++ ++#ifdef CONFIG_OCF_RANDOMHARVEST ++/* ++ * We run the random number generator output through SHA so that it ++ * is FIPS compliant. ++ */ ++ ++static volatile int sha_done = 0; ++static unsigned char sha_digest[20]; ++ ++static void ++ixp_hash_cb(UINT8 *digest, IxCryptoAccStatus status) ++{ ++ dprintk("%s(%p, %d)\n", __FUNCTION__, digest, status); ++ if (sha_digest != digest) ++ printk("digest error\n"); ++ if (IX_CRYPTO_ACC_STATUS_SUCCESS == status) ++ sha_done = 1; ++ else ++ sha_done = -status; ++} ++ ++static int ++ixp_read_random(void *arg, u_int32_t *buf, int maxwords) ++{ ++ IxCryptoAccStatus status; ++ int i, n, rc; ++ ++ dprintk("%s(%p, %d)\n", __FUNCTION__, buf, maxwords); ++ memset(buf, 0, maxwords * sizeof(*buf)); ++ status = ixCryptoAccPkePseudoRandomNumberGet(maxwords, buf); ++ if (status != IX_CRYPTO_ACC_STATUS_SUCCESS) { ++ dprintk("%s: ixCryptoAccPkePseudoRandomNumberGet failed %d\n", ++ __FUNCTION__, status); ++ return 0; ++ } ++ ++ /* ++ * run the random data through SHA to make it look more random ++ */ ++ ++ n = sizeof(sha_digest); /* process digest bytes at a time */ ++ ++ rc = 0; ++ for (i = 0; i < maxwords; i += n / sizeof(*buf)) { ++ if ((maxwords - i) * sizeof(*buf) < n) ++ n = (maxwords - i) * sizeof(*buf); ++ sha_done = 0; ++ status = ixCryptoAccPkeHashPerform(IX_CRYPTO_ACC_AUTH_SHA1, ++ (UINT8 *) &buf[i], n, ixp_hash_cb, sha_digest); ++ if (status != IX_CRYPTO_ACC_STATUS_SUCCESS) { ++ dprintk("ixCryptoAccPkeHashPerform failed %d\n", status); ++ return -EIO; ++ } ++ while (!sha_done) ++ schedule(); ++ if (sha_done < 0) { ++ dprintk("ixCryptoAccPkeHashPerform failed CB %d\n", -sha_done); ++ return 0; ++ } ++ memcpy(&buf[i], sha_digest, n); ++ rc += n / sizeof(*buf);; ++ } ++ ++ return rc; ++} ++#endif /* CONFIG_OCF_RANDOMHARVEST */ ++ ++#endif /* __ixp46X */ ++ ++ ++ ++/* ++ * our driver startup and shutdown routines ++ */ ++ ++static int ++ixp_init(void) ++{ ++ dprintk("%s(%p)\n", __FUNCTION__, ixp_init); ++ ++ if (ixp_init_crypto && ixCryptoAccInit() != IX_CRYPTO_ACC_STATUS_SUCCESS) ++ printk("ixCryptoAccInit failed, assuming already initialised!\n"); ++ ++ qcache = kmem_cache_create("ixp4xx_q", sizeof(struct ixp_q), 0, ++ SLAB_HWCACHE_ALIGN, NULL ++#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,23) ++ , NULL ++#endif ++ ); ++ if (!qcache) { ++ printk("failed to create Qcache\n"); ++ return -ENOENT; ++ } ++ ++ memset(&ixpdev, 0, sizeof(ixpdev)); ++ softc_device_init(&ixpdev, "ixp4xx", 0, ixp_methods); ++ ++ ixp_id = crypto_get_driverid(softc_get_device(&ixpdev), ++ CRYPTOCAP_F_HARDWARE); ++ if (ixp_id < 0) ++ panic("IXP/OCF crypto device cannot initialize!"); ++ ++#define REGISTER(alg) \ ++ crypto_register(ixp_id,alg,0,0) ++ ++ REGISTER(CRYPTO_DES_CBC); ++ REGISTER(CRYPTO_3DES_CBC); ++ REGISTER(CRYPTO_RIJNDAEL128_CBC); ++#ifdef CONFIG_OCF_IXP4XX_SHA1_MD5 ++ REGISTER(CRYPTO_MD5); ++ REGISTER(CRYPTO_SHA1); ++#endif ++ REGISTER(CRYPTO_MD5_HMAC); ++ REGISTER(CRYPTO_SHA1_HMAC); ++#undef REGISTER ++ ++#ifdef __ixp46X ++ spin_lock_init(&ixp_pkq_lock); ++ /* ++ * we do not enable the go fast options here as they can potentially ++ * allow timing based attacks ++ * ++ * http://www.openssl.org/news/secadv_20030219.txt ++ */ ++ ixCryptoAccPkeEauExpConfig(0, 0); ++ crypto_kregister(ixp_id, CRK_MOD_EXP, 0); ++#ifdef CONFIG_OCF_RANDOMHARVEST ++ crypto_rregister(ixp_id, ixp_read_random, NULL); ++#endif ++#endif ++ ++ return 0; ++} ++ ++static void ++ixp_exit(void) ++{ ++ dprintk("%s()\n", __FUNCTION__); ++ crypto_unregister_all(ixp_id); ++ ixp_id = -1; ++ kmem_cache_destroy(qcache); ++ qcache = NULL; ++} ++ ++module_init(ixp_init); ++module_exit(ixp_exit); ++ ++MODULE_LICENSE("Dual BSD/GPL"); ++MODULE_AUTHOR("David McCullough <dmccullough@cyberguard.com>"); ++MODULE_DESCRIPTION("ixp (OCF module for IXP4xx crypto)"); +diff -Nur linux-2.6.30.orig/crypto/ocf/ixp4xx/Makefile linux-2.6.30/crypto/ocf/ixp4xx/Makefile +--- linux-2.6.30.orig/crypto/ocf/ixp4xx/Makefile 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/ixp4xx/Makefile 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,104 @@ ++# for SGlinux builds ++-include $(ROOTDIR)/modules/.config ++ ++# ++# You will need to point this at your Intel ixp425 includes, this portion ++# of the Makefile only really works under SGLinux with the appropriate libs ++# installed. They can be downloaded from http://www.snapgear.org/ ++# ++ifeq ($(CONFIG_CPU_IXP46X),y) ++IXPLATFORM = ixp46X ++else ++ifeq ($(CONFIG_CPU_IXP43X),y) ++IXPLATFORM = ixp43X ++else ++IXPLATFORM = ixp42X ++endif ++endif ++ ++ifdef CONFIG_IXP400_LIB_2_4 ++IX_XSCALE_SW = $(ROOTDIR)/modules/ixp425/ixp400-2.4/ixp400_xscale_sw ++OSAL_DIR = $(ROOTDIR)/modules/ixp425/ixp400-2.4/ixp_osal ++endif ++ifdef CONFIG_IXP400_LIB_2_1 ++IX_XSCALE_SW = $(ROOTDIR)/modules/ixp425/ixp400-2.1/ixp400_xscale_sw ++OSAL_DIR = $(ROOTDIR)/modules/ixp425/ixp400-2.1/ixp_osal ++endif ++ifdef CONFIG_IXP400_LIB_2_0 ++IX_XSCALE_SW = $(ROOTDIR)/modules/ixp425/ixp400-2.0/ixp400_xscale_sw ++OSAL_DIR = $(ROOTDIR)/modules/ixp425/ixp400-2.0/ixp_osal ++endif ++ifdef IX_XSCALE_SW ++ifdef CONFIG_IXP400_LIB_2_4 ++IXP_CFLAGS = \ ++ -I$(ROOTDIR)/. \ ++ -I$(IX_XSCALE_SW)/src/include \ ++ -I$(OSAL_DIR)/common/include/ \ ++ -I$(OSAL_DIR)/common/include/modules/ \ ++ -I$(OSAL_DIR)/common/include/modules/ddk/ \ ++ -I$(OSAL_DIR)/common/include/modules/bufferMgt/ \ ++ -I$(OSAL_DIR)/common/include/modules/ioMem/ \ ++ -I$(OSAL_DIR)/common/os/linux/include/ \ ++ -I$(OSAL_DIR)/common/os/linux/include/core/ \ ++ -I$(OSAL_DIR)/common/os/linux/include/modules/ \ ++ -I$(OSAL_DIR)/common/os/linux/include/modules/ddk/ \ ++ -I$(OSAL_DIR)/common/os/linux/include/modules/bufferMgt/ \ ++ -I$(OSAL_DIR)/common/os/linux/include/modules/ioMem/ \ ++ -I$(OSAL_DIR)/platforms/$(IXPLATFORM)/include/ \ ++ -I$(OSAL_DIR)/platforms/$(IXPLATFORM)/os/linux/include/ \ ++ -DENABLE_IOMEM -DENABLE_BUFFERMGT -DENABLE_DDK \ ++ -DUSE_IXP4XX_CRYPTO ++else ++IXP_CFLAGS = \ ++ -I$(ROOTDIR)/. \ ++ -I$(IX_XSCALE_SW)/src/include \ ++ -I$(OSAL_DIR)/ \ ++ -I$(OSAL_DIR)/os/linux/include/ \ ++ -I$(OSAL_DIR)/os/linux/include/modules/ \ ++ -I$(OSAL_DIR)/os/linux/include/modules/ioMem/ \ ++ -I$(OSAL_DIR)/os/linux/include/modules/bufferMgt/ \ ++ -I$(OSAL_DIR)/os/linux/include/core/ \ ++ -I$(OSAL_DIR)/os/linux/include/platforms/ \ ++ -I$(OSAL_DIR)/os/linux/include/platforms/ixp400/ \ ++ -I$(OSAL_DIR)/os/linux/include/platforms/ixp400/ixp425 \ ++ -I$(OSAL_DIR)/os/linux/include/platforms/ixp400/ixp465 \ ++ -I$(OSAL_DIR)/os/linux/include/core/ \ ++ -I$(OSAL_DIR)/include/ \ ++ -I$(OSAL_DIR)/include/modules/ \ ++ -I$(OSAL_DIR)/include/modules/bufferMgt/ \ ++ -I$(OSAL_DIR)/include/modules/ioMem/ \ ++ -I$(OSAL_DIR)/include/platforms/ \ ++ -I$(OSAL_DIR)/include/platforms/ixp400/ \ ++ -DUSE_IXP4XX_CRYPTO ++endif ++endif ++ifdef CONFIG_IXP400_LIB_1_4 ++IXP_CFLAGS = \ ++ -I$(ROOTDIR)/. \ ++ -I$(ROOTDIR)/modules/ixp425/ixp400-1.4/ixp400_xscale_sw/src/include \ ++ -I$(ROOTDIR)/modules/ixp425/ixp400-1.4/ixp400_xscale_sw/src/linux \ ++ -DUSE_IXP4XX_CRYPTO ++endif ++ifndef IXPDIR ++IXPDIR = ixp-version-is-not-supported ++endif ++ ++ifeq ($(CONFIG_CPU_IXP46X),y) ++IXP_CFLAGS += -D__ixp46X ++else ++ifeq ($(CONFIG_CPU_IXP43X),y) ++IXP_CFLAGS += -D__ixp43X ++else ++IXP_CFLAGS += -D__ixp42X ++endif ++endif ++ ++obj-$(CONFIG_OCF_IXP4XX) += ixp4xx.o ++ ++obj ?= . ++EXTRA_CFLAGS += $(IXP_CFLAGS) -I$(obj)/.. -I$(obj)/. ++ ++ifdef TOPDIR ++-include $(TOPDIR)/Rules.make ++endif ++ +diff -Nur linux-2.6.30.orig/crypto/ocf/Kconfig linux-2.6.30/crypto/ocf/Kconfig +--- linux-2.6.30.orig/crypto/ocf/Kconfig 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/Kconfig 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,101 @@ ++menu "OCF Configuration" ++ ++config OCF_OCF ++ tristate "OCF (Open Cryptograhic Framework)" ++ help ++ A linux port of the OpenBSD/FreeBSD crypto framework. ++ ++config OCF_RANDOMHARVEST ++ bool "crypto random --- harvest entropy for /dev/random" ++ depends on OCF_OCF ++ help ++ Includes code to harvest random numbers from devices that support it. ++ ++config OCF_FIPS ++ bool "enable fips RNG checks" ++ depends on OCF_OCF && OCF_RANDOMHARVEST ++ help ++ Run all RNG provided data through a fips check before ++ adding it /dev/random's entropy pool. ++ ++config OCF_CRYPTODEV ++ tristate "cryptodev (user space support)" ++ depends on OCF_OCF ++ help ++ The user space API to access crypto hardware. ++ ++config OCF_CRYPTOSOFT ++ tristate "cryptosoft (software crypto engine)" ++ depends on OCF_OCF ++ help ++ A software driver for the OCF framework that uses ++ the kernel CryptoAPI. ++ ++config OCF_SAFE ++ tristate "safenet (HW crypto engine)" ++ depends on OCF_OCF ++ help ++ A driver for a number of the safenet Excel crypto accelerators. ++ Currently tested and working on the 1141 and 1741. ++ ++config OCF_IXP4XX ++ tristate "IXP4xx (HW crypto engine)" ++ depends on OCF_OCF ++ help ++ XScale IXP4xx crypto accelerator driver. Requires the ++ Intel Access library. ++ ++config OCF_IXP4XX_SHA1_MD5 ++ bool "IXP4xx SHA1 and MD5 Hashing" ++ depends on OCF_IXP4XX ++ help ++ Allows the IXP4xx crypto accelerator to perform SHA1 and MD5 hashing. ++ Note: this is MUCH slower than using cryptosoft (software crypto engine). ++ ++config OCF_HIFN ++ tristate "hifn (HW crypto engine)" ++ depends on OCF_OCF ++ help ++ OCF driver for various HIFN based crypto accelerators. ++ (7951, 7955, 7956, 7751, 7811) ++ ++config OCF_HIFNHIPP ++ tristate "Hifn HIPP (HW packet crypto engine)" ++ depends on OCF_OCF ++ help ++ OCF driver for various HIFN (HIPP) based crypto accelerators ++ (7855) ++ ++config OCF_TALITOS ++ tristate "talitos (HW crypto engine)" ++ depends on OCF_OCF ++ help ++ OCF driver for Freescale's security engine (SEC/talitos). ++ ++config OCF_PASEMI ++ tristate "pasemi (HW crypto engine)" ++ depends on OCF_OCF && PPC_PASEMI ++ help ++ OCF driver for the PA Semi PWRficient DMA Engine ++ ++config OCF_EP80579 ++ tristate "ep80579 (HW crypto engine)" ++ depends on OCF_OCF ++ help ++ OCF driver for the Intel EP80579 Integrated Processor Product Line. ++ ++config OCF_OCFNULL ++ tristate "ocfnull (fake crypto engine)" ++ depends on OCF_OCF ++ help ++ OCF driver for measuring ipsec overheads (does no crypto) ++ ++config OCF_BENCH ++ tristate "ocf-bench (HW crypto in-kernel benchmark)" ++ depends on OCF_OCF ++ help ++ A very simple encryption test for the in-kernel interface ++ of OCF. Also includes code to benchmark the IXP Access library ++ for comparison. ++ ++endmenu +diff -Nur linux-2.6.30.orig/crypto/ocf/Makefile linux-2.6.30/crypto/ocf/Makefile +--- linux-2.6.30.orig/crypto/ocf/Makefile 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/Makefile 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,121 @@ ++# for SGlinux builds ++-include $(ROOTDIR)/modules/.config ++ ++OCF_OBJS = crypto.o criov.o ++ ++ifdef CONFIG_OCF_RANDOMHARVEST ++ OCF_OBJS += random.o ++endif ++ ++ifdef CONFIG_OCF_FIPS ++ OCF_OBJS += rndtest.o ++endif ++ ++# Add in autoconf.h to get #defines for CONFIG_xxx ++AUTOCONF_H=$(ROOTDIR)/modules/autoconf.h ++ifeq ($(AUTOCONF_H), $(wildcard $(AUTOCONF_H))) ++ EXTRA_CFLAGS += -include $(AUTOCONF_H) ++ export EXTRA_CFLAGS ++endif ++ ++ifndef obj ++ obj ?= . ++ _obj = subdir ++ mod-subdirs := safe hifn ixp4xx talitos ocfnull ++ export-objs += crypto.o criov.o random.o ++ list-multi += ocf.o ++ _slash := ++else ++ _obj = obj ++ _slash := / ++endif ++ ++EXTRA_CFLAGS += -I$(obj)/. ++ ++obj-$(CONFIG_OCF_OCF) += ocf.o ++obj-$(CONFIG_OCF_CRYPTODEV) += cryptodev.o ++obj-$(CONFIG_OCF_CRYPTOSOFT) += cryptosoft.o ++obj-$(CONFIG_OCF_BENCH) += ocf-bench.o ++ ++$(_obj)-$(CONFIG_OCF_SAFE) += safe$(_slash) ++$(_obj)-$(CONFIG_OCF_HIFN) += hifn$(_slash) ++$(_obj)-$(CONFIG_OCF_IXP4XX) += ixp4xx$(_slash) ++$(_obj)-$(CONFIG_OCF_TALITOS) += talitos$(_slash) ++$(_obj)-$(CONFIG_OCF_PASEMI) += pasemi$(_slash) ++$(_obj)-$(CONFIG_OCF_EP80579) += ep80579$(_slash) ++$(_obj)-$(CONFIG_OCF_OCFNULL) += ocfnull$(_slash) ++ ++ocf-objs := $(OCF_OBJS) ++ ++$(list-multi) dummy1: $(ocf-objs) ++ $(LD) -r -o $@ $(ocf-objs) ++ ++.PHONY: ++clean: ++ rm -f *.o *.ko .*.o.flags .*.ko.cmd .*.o.cmd .*.mod.o.cmd *.mod.c ++ rm -f */*.o */*.ko */.*.o.cmd */.*.ko.cmd */.*.mod.o.cmd */*.mod.c */.*.o.flags ++ ++ifdef TOPDIR ++-include $(TOPDIR)/Rules.make ++endif ++ ++# ++# release gen targets ++# ++ ++.PHONY: patch ++patch: ++ REL=`date +%Y%m%d`; \ ++ patch=ocf-linux-$$REL.patch; \ ++ patch24=ocf-linux-24-$$REL.patch; \ ++ patch26=ocf-linux-26-$$REL.patch; \ ++ ( \ ++ find . -name Makefile; \ ++ find . -name Config.in; \ ++ find . -name Kconfig; \ ++ find . -name README; \ ++ find . -name '*.[ch]' | grep -v '.mod.c'; \ ++ ) | while read t; do \ ++ diff -Nau /dev/null $$t | sed 's?^+++ \./?+++ linux/crypto/ocf/?'; \ ++ done > $$patch; \ ++ cat patches/linux-2.4.35-ocf.patch $$patch > $$patch24; \ ++ cat patches/linux-2.6.26-ocf.patch $$patch > $$patch26 ++ ++.PHONY: tarball ++tarball: ++ REL=`date +%Y%m%d`; RELDIR=/tmp/ocf-linux-$$REL; \ ++ CURDIR=`pwd`; \ ++ rm -rf /tmp/ocf-linux-$$REL*; \ ++ mkdir -p $$RELDIR/tools; \ ++ cp README* $$RELDIR; \ ++ cp patches/openss*.patch $$RELDIR; \ ++ cp patches/crypto-tools.patch $$RELDIR; \ ++ cp tools/[!C]* $$RELDIR/tools; \ ++ cd ..; \ ++ tar cvf $$RELDIR/ocf-linux.tar \ ++ --exclude=CVS \ ++ --exclude=.* \ ++ --exclude=*.o \ ++ --exclude=*.ko \ ++ --exclude=*.mod.* \ ++ --exclude=README* \ ++ --exclude=ocf-*.patch \ ++ --exclude=ocf/patches/openss*.patch \ ++ --exclude=ocf/patches/crypto-tools.patch \ ++ --exclude=ocf/tools \ ++ ocf; \ ++ gzip -9 $$RELDIR/ocf-linux.tar; \ ++ cd /tmp; \ ++ tar cvf ocf-linux-$$REL.tar ocf-linux-$$REL; \ ++ gzip -9 ocf-linux-$$REL.tar; \ ++ cd $$CURDIR/../../user; \ ++ rm -rf /tmp/crypto-tools-$$REL*; \ ++ tar cvf /tmp/crypto-tools-$$REL.tar \ ++ --exclude=CVS \ ++ --exclude=.* \ ++ --exclude=*.o \ ++ --exclude=cryptotest \ ++ --exclude=cryptokeytest \ ++ crypto-tools; \ ++ gzip -9 /tmp/crypto-tools-$$REL.tar ++ +diff -Nur linux-2.6.30.orig/crypto/ocf/ocf-bench.c linux-2.6.30/crypto/ocf/ocf-bench.c +--- linux-2.6.30.orig/crypto/ocf/ocf-bench.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/ocf-bench.c 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,436 @@ ++/* ++ * A loadable module that benchmarks the OCF crypto speed from kernel space. ++ * ++ * Copyright (C) 2004-2007 David McCullough <david_mccullough@securecomputing.com> ++ * ++ * LICENSE TERMS ++ * ++ * The free distribution and use of this software in both source and binary ++ * form is allowed (with or without changes) provided that: ++ * ++ * 1. distributions of this source code include the above copyright ++ * notice, this list of conditions and the following disclaimer; ++ * ++ * 2. distributions in binary form include the above copyright ++ * notice, this list of conditions and the following disclaimer ++ * in the documentation and/or other associated materials; ++ * ++ * 3. the copyright holder's name is not used to endorse products ++ * built using this software without specific written permission. ++ * ++ * ALTERNATIVELY, provided that this notice is retained in full, this product ++ * may be distributed under the terms of the GNU General Public License (GPL), ++ * in which case the provisions of the GPL apply INSTEAD OF those given above. ++ * ++ * DISCLAIMER ++ * ++ * This software is provided 'as is' with no explicit or implied warranties ++ * in respect of its properties, including, but not limited to, correctness ++ * and/or fitness for purpose. ++ */ ++ ++ ++#ifndef AUTOCONF_INCLUDED ++#include <linux/config.h> ++#endif ++#include <linux/module.h> ++#include <linux/init.h> ++#include <linux/list.h> ++#include <linux/slab.h> ++#include <linux/wait.h> ++#include <linux/sched.h> ++#include <linux/spinlock.h> ++#include <linux/version.h> ++#include <linux/interrupt.h> ++#include <cryptodev.h> ++ ++#ifdef I_HAVE_AN_XSCALE_WITH_INTEL_SDK ++#define BENCH_IXP_ACCESS_LIB 1 ++#endif ++#ifdef BENCH_IXP_ACCESS_LIB ++#include <IxTypes.h> ++#include <IxOsBuffMgt.h> ++#include <IxNpeDl.h> ++#include <IxCryptoAcc.h> ++#include <IxQMgr.h> ++#include <IxOsServices.h> ++#include <IxOsCacheMMU.h> ++#endif ++ ++/* ++ * support for access lib version 1.4 ++ */ ++#ifndef IX_MBUF_PRIV ++#define IX_MBUF_PRIV(x) ((x)->priv) ++#endif ++ ++/* ++ * the number of simultaneously active requests ++ */ ++static int request_q_len = 20; ++module_param(request_q_len, int, 0); ++MODULE_PARM_DESC(request_q_len, "Number of outstanding requests"); ++/* ++ * how many requests we want to have processed ++ */ ++static int request_num = 1024; ++module_param(request_num, int, 0); ++MODULE_PARM_DESC(request_num, "run for at least this many requests"); ++/* ++ * the size of each request ++ */ ++static int request_size = 1500; ++module_param(request_size, int, 0); ++MODULE_PARM_DESC(request_size, "size of each request"); ++ ++/* ++ * a structure for each request ++ */ ++typedef struct { ++ struct work_struct work; ++#ifdef BENCH_IXP_ACCESS_LIB ++ IX_MBUF mbuf; ++#endif ++ unsigned char *buffer; ++} request_t; ++ ++static request_t *requests; ++ ++static int outstanding; ++static int total; ++ ++/*************************************************************************/ ++/* ++ * OCF benchmark routines ++ */ ++ ++static uint64_t ocf_cryptoid; ++static int ocf_init(void); ++static int ocf_cb(struct cryptop *crp); ++static void ocf_request(void *arg); ++#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,20) ++static void ocf_request_wq(struct work_struct *work); ++#endif ++ ++static int ++ocf_init(void) ++{ ++ int error; ++ struct cryptoini crie, cria; ++ struct cryptodesc crda, crde; ++ ++ memset(&crie, 0, sizeof(crie)); ++ memset(&cria, 0, sizeof(cria)); ++ memset(&crde, 0, sizeof(crde)); ++ memset(&crda, 0, sizeof(crda)); ++ ++ cria.cri_alg = CRYPTO_SHA1_HMAC; ++ cria.cri_klen = 20 * 8; ++ cria.cri_key = "0123456789abcdefghij"; ++ ++ crie.cri_alg = CRYPTO_3DES_CBC; ++ crie.cri_klen = 24 * 8; ++ crie.cri_key = "0123456789abcdefghijklmn"; ++ ++ crie.cri_next = &cria; ++ ++ error = crypto_newsession(&ocf_cryptoid, &crie, 0); ++ if (error) { ++ printk("crypto_newsession failed %d\n", error); ++ return -1; ++ } ++ return 0; ++} ++ ++static int ++ocf_cb(struct cryptop *crp) ++{ ++ request_t *r = (request_t *) crp->crp_opaque; ++ ++ if (crp->crp_etype) ++ printk("Error in OCF processing: %d\n", crp->crp_etype); ++ total++; ++ crypto_freereq(crp); ++ crp = NULL; ++ ++ if (total > request_num) { ++ outstanding--; ++ return 0; ++ } ++ ++#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,20) ++ INIT_WORK(&r->work, ocf_request_wq); ++#else ++ INIT_WORK(&r->work, ocf_request, r); ++#endif ++ schedule_work(&r->work); ++ return 0; ++} ++ ++ ++static void ++ocf_request(void *arg) ++{ ++ request_t *r = arg; ++ struct cryptop *crp = crypto_getreq(2); ++ struct cryptodesc *crde, *crda; ++ ++ if (!crp) { ++ outstanding--; ++ return; ++ } ++ ++ crde = crp->crp_desc; ++ crda = crde->crd_next; ++ ++ crda->crd_skip = 0; ++ crda->crd_flags = 0; ++ crda->crd_len = request_size; ++ crda->crd_inject = request_size; ++ crda->crd_alg = CRYPTO_SHA1_HMAC; ++ crda->crd_key = "0123456789abcdefghij"; ++ crda->crd_klen = 20 * 8; ++ ++ crde->crd_skip = 0; ++ crde->crd_flags = CRD_F_IV_EXPLICIT | CRD_F_ENCRYPT; ++ crde->crd_len = request_size; ++ crde->crd_inject = request_size; ++ crde->crd_alg = CRYPTO_3DES_CBC; ++ crde->crd_key = "0123456789abcdefghijklmn"; ++ crde->crd_klen = 24 * 8; ++ ++ crp->crp_ilen = request_size + 64; ++ crp->crp_flags = CRYPTO_F_CBIMM; ++ crp->crp_buf = (caddr_t) r->buffer; ++ crp->crp_callback = ocf_cb; ++ crp->crp_sid = ocf_cryptoid; ++ crp->crp_opaque = (caddr_t) r; ++ crypto_dispatch(crp); ++} ++ ++#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,20) ++static void ++ocf_request_wq(struct work_struct *work) ++{ ++ request_t *r = container_of(work, request_t, work); ++ ocf_request(r); ++} ++#endif ++ ++/*************************************************************************/ ++#ifdef BENCH_IXP_ACCESS_LIB ++/*************************************************************************/ ++/* ++ * CryptoAcc benchmark routines ++ */ ++ ++static IxCryptoAccCtx ixp_ctx; ++static UINT32 ixp_ctx_id; ++static IX_MBUF ixp_pri; ++static IX_MBUF ixp_sec; ++static int ixp_registered = 0; ++ ++static void ixp_register_cb(UINT32 ctx_id, IX_MBUF *bufp, ++ IxCryptoAccStatus status); ++static void ixp_perform_cb(UINT32 ctx_id, IX_MBUF *sbufp, IX_MBUF *dbufp, ++ IxCryptoAccStatus status); ++static void ixp_request(void *arg); ++#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,20) ++static void ixp_request_wq(struct work_struct *work); ++#endif ++ ++static int ++ixp_init(void) ++{ ++ IxCryptoAccStatus status; ++ ++ ixp_ctx.cipherCtx.cipherAlgo = IX_CRYPTO_ACC_CIPHER_3DES; ++ ixp_ctx.cipherCtx.cipherMode = IX_CRYPTO_ACC_MODE_CBC; ++ ixp_ctx.cipherCtx.cipherKeyLen = 24; ++ ixp_ctx.cipherCtx.cipherBlockLen = IX_CRYPTO_ACC_DES_BLOCK_64; ++ ixp_ctx.cipherCtx.cipherInitialVectorLen = IX_CRYPTO_ACC_DES_IV_64; ++ memcpy(ixp_ctx.cipherCtx.key.cipherKey, "0123456789abcdefghijklmn", 24); ++ ++ ixp_ctx.authCtx.authAlgo = IX_CRYPTO_ACC_AUTH_SHA1; ++ ixp_ctx.authCtx.authDigestLen = 12; ++ ixp_ctx.authCtx.aadLen = 0; ++ ixp_ctx.authCtx.authKeyLen = 20; ++ memcpy(ixp_ctx.authCtx.key.authKey, "0123456789abcdefghij", 20); ++ ++ ixp_ctx.useDifferentSrcAndDestMbufs = 0; ++ ixp_ctx.operation = IX_CRYPTO_ACC_OP_ENCRYPT_AUTH ; ++ ++ IX_MBUF_MLEN(&ixp_pri) = IX_MBUF_PKT_LEN(&ixp_pri) = 128; ++ IX_MBUF_MDATA(&ixp_pri) = (unsigned char *) kmalloc(128, SLAB_ATOMIC); ++ IX_MBUF_MLEN(&ixp_sec) = IX_MBUF_PKT_LEN(&ixp_sec) = 128; ++ IX_MBUF_MDATA(&ixp_sec) = (unsigned char *) kmalloc(128, SLAB_ATOMIC); ++ ++ status = ixCryptoAccCtxRegister(&ixp_ctx, &ixp_pri, &ixp_sec, ++ ixp_register_cb, ixp_perform_cb, &ixp_ctx_id); ++ ++ if (IX_CRYPTO_ACC_STATUS_SUCCESS == status) { ++ while (!ixp_registered) ++ schedule(); ++ return ixp_registered < 0 ? -1 : 0; ++ } ++ ++ printk("ixp: ixCryptoAccCtxRegister failed %d\n", status); ++ return -1; ++} ++ ++static void ++ixp_register_cb(UINT32 ctx_id, IX_MBUF *bufp, IxCryptoAccStatus status) ++{ ++ if (bufp) { ++ IX_MBUF_MLEN(bufp) = IX_MBUF_PKT_LEN(bufp) = 0; ++ kfree(IX_MBUF_MDATA(bufp)); ++ IX_MBUF_MDATA(bufp) = NULL; ++ } ++ ++ if (IX_CRYPTO_ACC_STATUS_WAIT == status) ++ return; ++ if (IX_CRYPTO_ACC_STATUS_SUCCESS == status) ++ ixp_registered = 1; ++ else ++ ixp_registered = -1; ++} ++ ++static void ++ixp_perform_cb( ++ UINT32 ctx_id, ++ IX_MBUF *sbufp, ++ IX_MBUF *dbufp, ++ IxCryptoAccStatus status) ++{ ++ request_t *r = NULL; ++ ++ total++; ++ if (total > request_num) { ++ outstanding--; ++ return; ++ } ++ ++ if (!sbufp || !(r = IX_MBUF_PRIV(sbufp))) { ++ printk("crappo %p %p\n", sbufp, r); ++ outstanding--; ++ return; ++ } ++ ++#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,20) ++ INIT_WORK(&r->work, ixp_request_wq); ++#else ++ INIT_WORK(&r->work, ixp_request, r); ++#endif ++ schedule_work(&r->work); ++} ++ ++static void ++ixp_request(void *arg) ++{ ++ request_t *r = arg; ++ IxCryptoAccStatus status; ++ ++ memset(&r->mbuf, 0, sizeof(r->mbuf)); ++ IX_MBUF_MLEN(&r->mbuf) = IX_MBUF_PKT_LEN(&r->mbuf) = request_size + 64; ++ IX_MBUF_MDATA(&r->mbuf) = r->buffer; ++ IX_MBUF_PRIV(&r->mbuf) = r; ++ status = ixCryptoAccAuthCryptPerform(ixp_ctx_id, &r->mbuf, NULL, ++ 0, request_size, 0, request_size, request_size, r->buffer); ++ if (IX_CRYPTO_ACC_STATUS_SUCCESS != status) { ++ printk("status1 = %d\n", status); ++ outstanding--; ++ return; ++ } ++ return; ++} ++ ++#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,20) ++static void ++ixp_request_wq(struct work_struct *work) ++{ ++ request_t *r = container_of(work, request_t, work); ++ ixp_request(r); ++} ++#endif ++ ++/*************************************************************************/ ++#endif /* BENCH_IXP_ACCESS_LIB */ ++/*************************************************************************/ ++ ++int ++ocfbench_init(void) ++{ ++ int i, jstart, jstop; ++ ++ printk("Crypto Speed tests\n"); ++ ++ requests = kmalloc(sizeof(request_t) * request_q_len, GFP_KERNEL); ++ if (!requests) { ++ printk("malloc failed\n"); ++ return -EINVAL; ++ } ++ ++ for (i = 0; i < request_q_len; i++) { ++ /* +64 for return data */ ++ requests[i].buffer = kmalloc(request_size + 128, GFP_DMA); ++ if (!requests[i].buffer) { ++ printk("malloc failed\n"); ++ return -EINVAL; ++ } ++ memset(requests[i].buffer, '0' + i, request_size + 128); ++ } ++ ++ /* ++ * OCF benchmark ++ */ ++ printk("OCF: testing ...\n"); ++ ocf_init(); ++ total = outstanding = 0; ++ jstart = jiffies; ++ for (i = 0; i < request_q_len; i++) { ++ outstanding++; ++ ocf_request(&requests[i]); ++ } ++ while (outstanding > 0) ++ schedule(); ++ jstop = jiffies; ++ ++ printk("OCF: %d requests of %d bytes in %d jiffies\n", total, request_size, ++ jstop - jstart); ++ ++#ifdef BENCH_IXP_ACCESS_LIB ++ /* ++ * IXP benchmark ++ */ ++ printk("IXP: testing ...\n"); ++ ixp_init(); ++ total = outstanding = 0; ++ jstart = jiffies; ++ for (i = 0; i < request_q_len; i++) { ++ outstanding++; ++ ixp_request(&requests[i]); ++ } ++ while (outstanding > 0) ++ schedule(); ++ jstop = jiffies; ++ ++ printk("IXP: %d requests of %d bytes in %d jiffies\n", total, request_size, ++ jstop - jstart); ++#endif /* BENCH_IXP_ACCESS_LIB */ ++ ++ for (i = 0; i < request_q_len; i++) ++ kfree(requests[i].buffer); ++ kfree(requests); ++ return -EINVAL; /* always fail to load so it can be re-run quickly ;-) */ ++} ++ ++static void __exit ocfbench_exit(void) ++{ ++} ++ ++module_init(ocfbench_init); ++module_exit(ocfbench_exit); ++ ++MODULE_LICENSE("BSD"); ++MODULE_AUTHOR("David McCullough <david_mccullough@securecomputing.com>"); ++MODULE_DESCRIPTION("Benchmark various in-kernel crypto speeds"); +diff -Nur linux-2.6.30.orig/crypto/ocf/ocf-compat.h linux-2.6.30/crypto/ocf/ocf-compat.h +--- linux-2.6.30.orig/crypto/ocf/ocf-compat.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/ocf-compat.h 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,270 @@ ++#ifndef _BSD_COMPAT_H_ ++#define _BSD_COMPAT_H_ 1 ++/****************************************************************************/ ++/* ++ * Provide compat routines for older linux kernels and BSD kernels ++ * ++ * Written by David McCullough <david_mccullough@securecomputing.com> ++ * Copyright (C) 2007 David McCullough <david_mccullough@securecomputing.com> ++ * ++ * LICENSE TERMS ++ * ++ * The free distribution and use of this software in both source and binary ++ * form is allowed (with or without changes) provided that: ++ * ++ * 1. distributions of this source code include the above copyright ++ * notice, this list of conditions and the following disclaimer; ++ * ++ * 2. distributions in binary form include the above copyright ++ * notice, this list of conditions and the following disclaimer ++ * in the documentation and/or other associated materials; ++ * ++ * 3. the copyright holder's name is not used to endorse products ++ * built using this software without specific written permission. ++ * ++ * ALTERNATIVELY, provided that this notice is retained in full, this file ++ * may be distributed under the terms of the GNU General Public License (GPL), ++ * in which case the provisions of the GPL apply INSTEAD OF those given above. ++ * ++ * DISCLAIMER ++ * ++ * This software is provided 'as is' with no explicit or implied warranties ++ * in respect of its properties, including, but not limited to, correctness ++ * and/or fitness for purpose. ++ */ ++/****************************************************************************/ ++#ifdef __KERNEL__ ++/* ++ * fake some BSD driver interface stuff specifically for OCF use ++ */ ++ ++typedef struct ocf_device *device_t; ++ ++typedef struct { ++ int (*cryptodev_newsession)(device_t dev, u_int32_t *sidp, struct cryptoini *cri); ++ int (*cryptodev_freesession)(device_t dev, u_int64_t tid); ++ int (*cryptodev_process)(device_t dev, struct cryptop *crp, int hint); ++ int (*cryptodev_kprocess)(device_t dev, struct cryptkop *krp, int hint); ++} device_method_t; ++#define DEVMETHOD(id, func) id: func ++ ++struct ocf_device { ++ char name[32]; /* the driver name */ ++ char nameunit[32]; /* the driver name + HW instance */ ++ int unit; ++ device_method_t methods; ++ void *softc; ++}; ++ ++#define CRYPTODEV_NEWSESSION(dev, sid, cri) \ ++ ((*(dev)->methods.cryptodev_newsession)(dev,sid,cri)) ++#define CRYPTODEV_FREESESSION(dev, sid) \ ++ ((*(dev)->methods.cryptodev_freesession)(dev, sid)) ++#define CRYPTODEV_PROCESS(dev, crp, hint) \ ++ ((*(dev)->methods.cryptodev_process)(dev, crp, hint)) ++#define CRYPTODEV_KPROCESS(dev, krp, hint) \ ++ ((*(dev)->methods.cryptodev_kprocess)(dev, krp, hint)) ++ ++#define device_get_name(dev) ((dev)->name) ++#define device_get_nameunit(dev) ((dev)->nameunit) ++#define device_get_unit(dev) ((dev)->unit) ++#define device_get_softc(dev) ((dev)->softc) ++ ++#define softc_device_decl \ ++ struct ocf_device _device; \ ++ device_t ++ ++#define softc_device_init(_sc, _name, _unit, _methods) \ ++ if (1) {\ ++ strncpy((_sc)->_device.name, _name, sizeof((_sc)->_device.name) - 1); \ ++ snprintf((_sc)->_device.nameunit, sizeof((_sc)->_device.name), "%s%d", _name, _unit); \ ++ (_sc)->_device.unit = _unit; \ ++ (_sc)->_device.methods = _methods; \ ++ (_sc)->_device.softc = (void *) _sc; \ ++ *(device_t *)((softc_get_device(_sc))+1) = &(_sc)->_device; \ ++ } else ++ ++#define softc_get_device(_sc) (&(_sc)->_device) ++ ++/* ++ * iomem support for 2.4 and 2.6 kernels ++ */ ++#include <linux/version.h> ++#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,0) ++#define ocf_iomem_t unsigned long ++ ++/* ++ * implement simple workqueue like support for older kernels ++ */ ++ ++#include <linux/tqueue.h> ++ ++#define work_struct tq_struct ++ ++#define INIT_WORK(wp, fp, ap) \ ++ do { \ ++ (wp)->sync = 0; \ ++ (wp)->routine = (fp); \ ++ (wp)->data = (ap); \ ++ } while (0) ++ ++#define schedule_work(wp) \ ++ do { \ ++ queue_task((wp), &tq_immediate); \ ++ mark_bh(IMMEDIATE_BH); \ ++ } while (0) ++ ++#define flush_scheduled_work() run_task_queue(&tq_immediate) ++ ++#else ++#define ocf_iomem_t void __iomem * ++ ++#include <linux/workqueue.h> ++ ++#endif ++ ++#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,26) ++#include <linux/fdtable.h> ++#elif LINUX_VERSION_CODE < KERNEL_VERSION(2,6,11) ++#define files_fdtable(files) (files) ++#endif ++ ++#ifdef MODULE_PARM ++#undef module_param /* just in case */ ++#define module_param(a,b,c) MODULE_PARM(a,"i") ++#endif ++ ++#define bzero(s,l) memset(s,0,l) ++#define bcopy(s,d,l) memcpy(d,s,l) ++#define bcmp(x, y, l) memcmp(x,y,l) ++ ++#define MIN(x,y) ((x) < (y) ? (x) : (y)) ++ ++#define device_printf(dev, a...) ({ \ ++ printk("%s: ", device_get_nameunit(dev)); printk(a); \ ++ }) ++ ++#undef printf ++#define printf(fmt...) printk(fmt) ++ ++#define KASSERT(c,p) if (!(c)) { printk p ; } else ++ ++#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,0) ++#define ocf_daemonize(str) \ ++ daemonize(); \ ++ spin_lock_irq(¤t->sigmask_lock); \ ++ sigemptyset(¤t->blocked); \ ++ recalc_sigpending(current); \ ++ spin_unlock_irq(¤t->sigmask_lock); \ ++ sprintf(current->comm, str); ++#else ++#define ocf_daemonize(str) daemonize(str); ++#endif ++ ++#define TAILQ_INSERT_TAIL(q,d,m) list_add_tail(&(d)->m, (q)) ++#define TAILQ_EMPTY(q) list_empty(q) ++#define TAILQ_FOREACH(v, q, m) list_for_each_entry(v, q, m) ++ ++#define read_random(p,l) get_random_bytes(p,l) ++ ++#define DELAY(x) ((x) > 2000 ? mdelay((x)/1000) : udelay(x)) ++#define strtoul simple_strtoul ++ ++#define pci_get_vendor(dev) ((dev)->vendor) ++#define pci_get_device(dev) ((dev)->device) ++ ++#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,0) ++#define pci_set_consistent_dma_mask(dev, mask) (0) ++#endif ++#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,10) ++#define pci_dma_sync_single_for_cpu pci_dma_sync_single ++#endif ++ ++#ifndef DMA_32BIT_MASK ++#define DMA_32BIT_MASK 0x00000000ffffffffULL ++#endif ++ ++#define htole32(x) cpu_to_le32(x) ++#define htobe32(x) cpu_to_be32(x) ++#define htole16(x) cpu_to_le16(x) ++#define htobe16(x) cpu_to_be16(x) ++ ++/* older kernels don't have these */ ++ ++#ifndef IRQ_NONE ++#define IRQ_NONE ++#define IRQ_HANDLED ++#define irqreturn_t void ++#endif ++#ifndef IRQF_SHARED ++#define IRQF_SHARED SA_SHIRQ ++#endif ++ ++#if LINUX_VERSION_CODE < KERNEL_VERSION(2,5,0) ++# define strlcpy(dest,src,len) \ ++ ({strncpy(dest,src,(len)-1); ((char *)dest)[(len)-1] = '\0'; }) ++#endif ++ ++#ifndef MAX_ERRNO ++#define MAX_ERRNO 4095 ++#endif ++#ifndef IS_ERR_VALUE ++#define IS_ERR_VALUE(x) ((unsigned long)(x) >= (unsigned long)-MAX_ERRNO) ++#endif ++ ++/* ++ * common debug for all ++ */ ++#if 1 ++#define dprintk(a...) do { if (debug) printk(a); } while(0) ++#else ++#define dprintk(a...) ++#endif ++ ++#ifndef SLAB_ATOMIC ++/* Changed in 2.6.20, must use GFP_ATOMIC now */ ++#define SLAB_ATOMIC GFP_ATOMIC ++#endif ++ ++/* ++ * need some additional support for older kernels */ ++#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,2) ++#define pci_register_driver_compat(driver, rc) \ ++ do { \ ++ if ((rc) > 0) { \ ++ (rc) = 0; \ ++ } else if (rc == 0) { \ ++ (rc) = -ENODEV; \ ++ } else { \ ++ pci_unregister_driver(driver); \ ++ } \ ++ } while (0) ++#elif LINUX_VERSION_CODE < KERNEL_VERSION(2,6,10) ++#define pci_register_driver_compat(driver,rc) ((rc) = (rc) < 0 ? (rc) : 0) ++#else ++#define pci_register_driver_compat(driver,rc) ++#endif ++ ++#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,24) ++ ++#include <asm/scatterlist.h> ++ ++static inline void sg_set_page(struct scatterlist *sg, struct page *page, ++ unsigned int len, unsigned int offset) ++{ ++ sg->page = page; ++ sg->offset = offset; ++ sg->length = len; ++} ++ ++static inline void *sg_virt(struct scatterlist *sg) ++{ ++ return page_address(sg->page) + sg->offset; ++} ++ ++#endif ++ ++#endif /* __KERNEL__ */ ++ ++/****************************************************************************/ ++#endif /* _BSD_COMPAT_H_ */ +diff -Nur linux-2.6.30.orig/crypto/ocf/ocfnull/Makefile linux-2.6.30/crypto/ocf/ocfnull/Makefile +--- linux-2.6.30.orig/crypto/ocf/ocfnull/Makefile 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/ocfnull/Makefile 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,12 @@ ++# for SGlinux builds ++-include $(ROOTDIR)/modules/.config ++ ++obj-$(CONFIG_OCF_OCFNULL) += ocfnull.o ++ ++obj ?= . ++EXTRA_CFLAGS += -I$(obj)/.. ++ ++ifdef TOPDIR ++-include $(TOPDIR)/Rules.make ++endif ++ +diff -Nur linux-2.6.30.orig/crypto/ocf/ocfnull/ocfnull.c linux-2.6.30/crypto/ocf/ocfnull/ocfnull.c +--- linux-2.6.30.orig/crypto/ocf/ocfnull/ocfnull.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/ocfnull/ocfnull.c 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,203 @@ ++/* ++ * An OCF module for determining the cost of crypto versus the cost of ++ * IPSec processing outside of OCF. This modules gives us the effect of ++ * zero cost encryption, of course you will need to run it at both ends ++ * since it does no crypto at all. ++ * ++ * Written by David McCullough <david_mccullough@securecomputing.com> ++ * Copyright (C) 2006-2007 David McCullough ++ * ++ * LICENSE TERMS ++ * ++ * The free distribution and use of this software in both source and binary ++ * form is allowed (with or without changes) provided that: ++ * ++ * 1. distributions of this source code include the above copyright ++ * notice, this list of conditions and the following disclaimer; ++ * ++ * 2. distributions in binary form include the above copyright ++ * notice, this list of conditions and the following disclaimer ++ * in the documentation and/or other associated materials; ++ * ++ * 3. the copyright holder's name is not used to endorse products ++ * built using this software without specific written permission. ++ * ++ * ALTERNATIVELY, provided that this notice is retained in full, this product ++ * may be distributed under the terms of the GNU General Public License (GPL), ++ * in which case the provisions of the GPL apply INSTEAD OF those given above. ++ * ++ * DISCLAIMER ++ * ++ * This software is provided 'as is' with no explicit or implied warranties ++ * in respect of its properties, including, but not limited to, correctness ++ * and/or fitness for purpose. ++ */ ++ ++#ifndef AUTOCONF_INCLUDED ++#include <linux/config.h> ++#endif ++#include <linux/module.h> ++#include <linux/init.h> ++#include <linux/list.h> ++#include <linux/slab.h> ++#include <linux/sched.h> ++#include <linux/wait.h> ++#include <linux/crypto.h> ++#include <linux/interrupt.h> ++ ++#include <cryptodev.h> ++#include <uio.h> ++ ++static int32_t null_id = -1; ++static u_int32_t null_sesnum = 0; ++ ++static int null_process(device_t, struct cryptop *, int); ++static int null_newsession(device_t, u_int32_t *, struct cryptoini *); ++static int null_freesession(device_t, u_int64_t); ++ ++#define debug ocfnull_debug ++int ocfnull_debug = 0; ++module_param(ocfnull_debug, int, 0644); ++MODULE_PARM_DESC(ocfnull_debug, "Enable debug"); ++ ++/* ++ * dummy device structure ++ */ ++ ++static struct { ++ softc_device_decl sc_dev; ++} nulldev; ++ ++static device_method_t null_methods = { ++ /* crypto device methods */ ++ DEVMETHOD(cryptodev_newsession, null_newsession), ++ DEVMETHOD(cryptodev_freesession,null_freesession), ++ DEVMETHOD(cryptodev_process, null_process), ++}; ++ ++/* ++ * Generate a new software session. ++ */ ++static int ++null_newsession(device_t arg, u_int32_t *sid, struct cryptoini *cri) ++{ ++ dprintk("%s()\n", __FUNCTION__); ++ if (sid == NULL || cri == NULL) { ++ dprintk("%s,%d - EINVAL\n", __FILE__, __LINE__); ++ return EINVAL; ++ } ++ ++ if (null_sesnum == 0) ++ null_sesnum++; ++ *sid = null_sesnum++; ++ return 0; ++} ++ ++ ++/* ++ * Free a session. ++ */ ++static int ++null_freesession(device_t arg, u_int64_t tid) ++{ ++ u_int32_t sid = CRYPTO_SESID2LID(tid); ++ ++ dprintk("%s()\n", __FUNCTION__); ++ if (sid > null_sesnum) { ++ dprintk("%s,%d: EINVAL\n", __FILE__, __LINE__); ++ return EINVAL; ++ } ++ ++ /* Silently accept and return */ ++ if (sid == 0) ++ return 0; ++ return 0; ++} ++ ++ ++/* ++ * Process a request. ++ */ ++static int ++null_process(device_t arg, struct cryptop *crp, int hint) ++{ ++ unsigned int lid; ++ ++ dprintk("%s()\n", __FUNCTION__); ++ ++ /* Sanity check */ ++ if (crp == NULL) { ++ dprintk("%s,%d: EINVAL\n", __FILE__, __LINE__); ++ return EINVAL; ++ } ++ ++ crp->crp_etype = 0; ++ ++ if (crp->crp_desc == NULL || crp->crp_buf == NULL) { ++ dprintk("%s,%d: EINVAL\n", __FILE__, __LINE__); ++ crp->crp_etype = EINVAL; ++ goto done; ++ } ++ ++ /* ++ * find the session we are using ++ */ ++ ++ lid = crp->crp_sid & 0xffffffff; ++ if (lid >= null_sesnum || lid == 0) { ++ crp->crp_etype = ENOENT; ++ dprintk("%s,%d: ENOENT\n", __FILE__, __LINE__); ++ goto done; ++ } ++ ++done: ++ crypto_done(crp); ++ return 0; ++} ++ ++ ++/* ++ * our driver startup and shutdown routines ++ */ ++ ++static int ++null_init(void) ++{ ++ dprintk("%s(%p)\n", __FUNCTION__, null_init); ++ ++ memset(&nulldev, 0, sizeof(nulldev)); ++ softc_device_init(&nulldev, "ocfnull", 0, null_methods); ++ ++ null_id = crypto_get_driverid(softc_get_device(&nulldev), ++ CRYPTOCAP_F_HARDWARE); ++ if (null_id < 0) ++ panic("ocfnull: crypto device cannot initialize!"); ++ ++#define REGISTER(alg) \ ++ crypto_register(null_id,alg,0,0) ++ REGISTER(CRYPTO_DES_CBC); ++ REGISTER(CRYPTO_3DES_CBC); ++ REGISTER(CRYPTO_RIJNDAEL128_CBC); ++ REGISTER(CRYPTO_MD5); ++ REGISTER(CRYPTO_SHA1); ++ REGISTER(CRYPTO_MD5_HMAC); ++ REGISTER(CRYPTO_SHA1_HMAC); ++#undef REGISTER ++ ++ return 0; ++} ++ ++static void ++null_exit(void) ++{ ++ dprintk("%s()\n", __FUNCTION__); ++ crypto_unregister_all(null_id); ++ null_id = -1; ++} ++ ++module_init(null_init); ++module_exit(null_exit); ++ ++MODULE_LICENSE("Dual BSD/GPL"); ++MODULE_AUTHOR("David McCullough <david_mccullough@securecomputing.com>"); ++MODULE_DESCRIPTION("ocfnull - claims a lot but does nothing"); +diff -Nur linux-2.6.30.orig/crypto/ocf/pasemi/Makefile linux-2.6.30/crypto/ocf/pasemi/Makefile +--- linux-2.6.30.orig/crypto/ocf/pasemi/Makefile 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/pasemi/Makefile 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,12 @@ ++# for SGlinux builds ++-include $(ROOTDIR)/modules/.config ++ ++obj-$(CONFIG_OCF_PASEMI) += pasemi.o ++ ++obj ?= . ++EXTRA_CFLAGS += -I$(obj)/.. -I$(obj)/ ++ ++ifdef TOPDIR ++-include $(TOPDIR)/Rules.make ++endif ++ +diff -Nur linux-2.6.30.orig/crypto/ocf/pasemi/pasemi.c linux-2.6.30/crypto/ocf/pasemi/pasemi.c +--- linux-2.6.30.orig/crypto/ocf/pasemi/pasemi.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/pasemi/pasemi.c 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,1009 @@ ++/* ++ * Copyright (C) 2007 PA Semi, Inc ++ * ++ * Driver for the PA Semi PWRficient DMA Crypto Engine ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License version 2 as ++ * published by the Free Software Foundation. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA ++ */ ++ ++#ifndef AUTOCONF_INCLUDED ++#include <linux/config.h> ++#endif ++#include <linux/module.h> ++#include <linux/init.h> ++#include <linux/interrupt.h> ++#include <linux/timer.h> ++#include <linux/random.h> ++#include <linux/skbuff.h> ++#include <asm/scatterlist.h> ++#include <linux/moduleparam.h> ++#include <linux/pci.h> ++#include <cryptodev.h> ++#include <uio.h> ++#include "pasemi_fnu.h" ++ ++#define DRV_NAME "pasemi" ++ ++#define TIMER_INTERVAL 1000 ++ ++static void __devexit pasemi_dma_remove(struct pci_dev *pdev); ++static struct pasdma_status volatile * dma_status; ++ ++static int debug; ++module_param(debug, int, 0644); ++MODULE_PARM_DESC(debug, "Enable debug"); ++ ++static void pasemi_desc_start(struct pasemi_desc *desc, u64 hdr) ++{ ++ desc->postop = 0; ++ desc->quad[0] = hdr; ++ desc->quad_cnt = 1; ++ desc->size = 1; ++} ++ ++static void pasemi_desc_build(struct pasemi_desc *desc, u64 val) ++{ ++ desc->quad[desc->quad_cnt++] = val; ++ desc->size = (desc->quad_cnt + 1) / 2; ++} ++ ++static void pasemi_desc_hdr(struct pasemi_desc *desc, u64 hdr) ++{ ++ desc->quad[0] |= hdr; ++} ++ ++static int pasemi_desc_size(struct pasemi_desc *desc) ++{ ++ return desc->size; ++} ++ ++static void pasemi_ring_add_desc( ++ struct pasemi_fnu_txring *ring, ++ struct pasemi_desc *desc, ++ struct cryptop *crp) { ++ int i; ++ int ring_index = 2 * (ring->next_to_fill & (TX_RING_SIZE-1)); ++ ++ TX_DESC_INFO(ring, ring->next_to_fill).desc_size = desc->size; ++ TX_DESC_INFO(ring, ring->next_to_fill).desc_postop = desc->postop; ++ TX_DESC_INFO(ring, ring->next_to_fill).cf_crp = crp; ++ ++ for (i = 0; i < desc->quad_cnt; i += 2) { ++ ring_index = 2 * (ring->next_to_fill & (TX_RING_SIZE-1)); ++ ring->desc[ring_index] = desc->quad[i]; ++ ring->desc[ring_index + 1] = desc->quad[i + 1]; ++ ring->next_to_fill++; ++ } ++ ++ if (desc->quad_cnt & 1) ++ ring->desc[ring_index + 1] = 0; ++} ++ ++static void pasemi_ring_incr(struct pasemi_softc *sc, int chan_index, int incr) ++{ ++ out_le32(sc->dma_regs + PAS_DMA_TXCHAN_INCR(sc->base_chan + chan_index), ++ incr); ++} ++ ++/* ++ * Generate a new software session. ++ */ ++static int ++pasemi_newsession(device_t dev, u_int32_t *sidp, struct cryptoini *cri) ++{ ++ struct cryptoini *c, *encini = NULL, *macini = NULL; ++ struct pasemi_softc *sc = device_get_softc(dev); ++ struct pasemi_session *ses = NULL, **sespp; ++ int sesn, blksz = 0; ++ u64 ccmd = 0; ++ unsigned long flags; ++ struct pasemi_desc init_desc; ++ struct pasemi_fnu_txring *txring; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ if (sidp == NULL || cri == NULL || sc == NULL) { ++ DPRINTF("%s,%d - EINVAL\n", __FILE__, __LINE__); ++ return -EINVAL; ++ } ++ for (c = cri; c != NULL; c = c->cri_next) { ++ if (ALG_IS_SIG(c->cri_alg)) { ++ if (macini) ++ return -EINVAL; ++ macini = c; ++ } else if (ALG_IS_CIPHER(c->cri_alg)) { ++ if (encini) ++ return -EINVAL; ++ encini = c; ++ } else { ++ DPRINTF("UNKNOWN c->cri_alg %d\n", c->cri_alg); ++ return -EINVAL; ++ } ++ } ++ if (encini == NULL && macini == NULL) ++ return -EINVAL; ++ if (encini) { ++ /* validate key length */ ++ switch (encini->cri_alg) { ++ case CRYPTO_DES_CBC: ++ if (encini->cri_klen != 64) ++ return -EINVAL; ++ ccmd = DMA_CALGO_DES; ++ break; ++ case CRYPTO_3DES_CBC: ++ if (encini->cri_klen != 192) ++ return -EINVAL; ++ ccmd = DMA_CALGO_3DES; ++ break; ++ case CRYPTO_AES_CBC: ++ if (encini->cri_klen != 128 && ++ encini->cri_klen != 192 && ++ encini->cri_klen != 256) ++ return -EINVAL; ++ ccmd = DMA_CALGO_AES; ++ break; ++ case CRYPTO_ARC4: ++ if (encini->cri_klen != 128) ++ return -EINVAL; ++ ccmd = DMA_CALGO_ARC; ++ break; ++ default: ++ DPRINTF("UNKNOWN encini->cri_alg %d\n", ++ encini->cri_alg); ++ return -EINVAL; ++ } ++ } ++ ++ if (macini) { ++ switch (macini->cri_alg) { ++ case CRYPTO_MD5: ++ case CRYPTO_MD5_HMAC: ++ blksz = 16; ++ break; ++ case CRYPTO_SHA1: ++ case CRYPTO_SHA1_HMAC: ++ blksz = 20; ++ break; ++ default: ++ DPRINTF("UNKNOWN macini->cri_alg %d\n", ++ macini->cri_alg); ++ return -EINVAL; ++ } ++ if (((macini->cri_klen + 7) / 8) > blksz) { ++ DPRINTF("key length %d bigger than blksize %d not supported\n", ++ ((macini->cri_klen + 7) / 8), blksz); ++ return -EINVAL; ++ } ++ } ++ ++ for (sesn = 0; sesn < sc->sc_nsessions; sesn++) { ++ if (sc->sc_sessions[sesn] == NULL) { ++ sc->sc_sessions[sesn] = (struct pasemi_session *) ++ kzalloc(sizeof(struct pasemi_session), GFP_ATOMIC); ++ ses = sc->sc_sessions[sesn]; ++ break; ++ } else if (sc->sc_sessions[sesn]->used == 0) { ++ ses = sc->sc_sessions[sesn]; ++ break; ++ } ++ } ++ ++ if (ses == NULL) { ++ sespp = (struct pasemi_session **) ++ kzalloc(sc->sc_nsessions * 2 * ++ sizeof(struct pasemi_session *), GFP_ATOMIC); ++ if (sespp == NULL) ++ return -ENOMEM; ++ memcpy(sespp, sc->sc_sessions, ++ sc->sc_nsessions * sizeof(struct pasemi_session *)); ++ kfree(sc->sc_sessions); ++ sc->sc_sessions = sespp; ++ sesn = sc->sc_nsessions; ++ ses = sc->sc_sessions[sesn] = (struct pasemi_session *) ++ kzalloc(sizeof(struct pasemi_session), GFP_ATOMIC); ++ if (ses == NULL) ++ return -ENOMEM; ++ sc->sc_nsessions *= 2; ++ } ++ ++ ses->used = 1; ++ ++ ses->dma_addr = pci_map_single(sc->dma_pdev, (void *) ses->civ, ++ sizeof(struct pasemi_session), DMA_TO_DEVICE); ++ ++ /* enter the channel scheduler */ ++ spin_lock_irqsave(&sc->sc_chnlock, flags); ++ ++ /* ARC4 has to be processed by the even channel */ ++ if (encini && (encini->cri_alg == CRYPTO_ARC4)) ++ ses->chan = sc->sc_lastchn & ~1; ++ else ++ ses->chan = sc->sc_lastchn; ++ sc->sc_lastchn = (sc->sc_lastchn + 1) % sc->sc_num_channels; ++ ++ spin_unlock_irqrestore(&sc->sc_chnlock, flags); ++ ++ txring = &sc->tx[ses->chan]; ++ ++ if (encini) { ++ ses->ccmd = ccmd; ++ ++ /* get an IV */ ++ /* XXX may read fewer than requested */ ++ get_random_bytes(ses->civ, sizeof(ses->civ)); ++ ++ ses->keysz = (encini->cri_klen - 63) / 64; ++ memcpy(ses->key, encini->cri_key, (ses->keysz + 1) * 8); ++ ++ pasemi_desc_start(&init_desc, ++ XCT_CTRL_HDR(ses->chan, (encini && macini) ? 0x68 : 0x40, DMA_FN_CIV0)); ++ pasemi_desc_build(&init_desc, ++ XCT_FUN_SRC_PTR((encini && macini) ? 0x68 : 0x40, ses->dma_addr)); ++ } ++ if (macini) { ++ if (macini->cri_alg == CRYPTO_MD5_HMAC || ++ macini->cri_alg == CRYPTO_SHA1_HMAC) ++ memcpy(ses->hkey, macini->cri_key, blksz); ++ else { ++ /* Load initialization constants(RFC 1321, 3174) */ ++ ses->hiv[0] = 0x67452301efcdab89ULL; ++ ses->hiv[1] = 0x98badcfe10325476ULL; ++ ses->hiv[2] = 0xc3d2e1f000000000ULL; ++ } ++ ses->hseq = 0ULL; ++ } ++ ++ spin_lock_irqsave(&txring->fill_lock, flags); ++ ++ if (((txring->next_to_fill + pasemi_desc_size(&init_desc)) - ++ txring->next_to_clean) > TX_RING_SIZE) { ++ spin_unlock_irqrestore(&txring->fill_lock, flags); ++ return ERESTART; ++ } ++ ++ if (encini) { ++ pasemi_ring_add_desc(txring, &init_desc, NULL); ++ pasemi_ring_incr(sc, ses->chan, ++ pasemi_desc_size(&init_desc)); ++ } ++ ++ txring->sesn = sesn; ++ spin_unlock_irqrestore(&txring->fill_lock, flags); ++ ++ *sidp = PASEMI_SID(sesn); ++ return 0; ++} ++ ++/* ++ * Deallocate a session. ++ */ ++static int ++pasemi_freesession(device_t dev, u_int64_t tid) ++{ ++ struct pasemi_softc *sc = device_get_softc(dev); ++ int session; ++ u_int32_t sid = ((u_int32_t) tid) & 0xffffffff; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ if (sc == NULL) ++ return -EINVAL; ++ session = PASEMI_SESSION(sid); ++ if (session >= sc->sc_nsessions || !sc->sc_sessions[session]) ++ return -EINVAL; ++ ++ pci_unmap_single(sc->dma_pdev, ++ sc->sc_sessions[session]->dma_addr, ++ sizeof(struct pasemi_session), DMA_TO_DEVICE); ++ memset(sc->sc_sessions[session], 0, ++ sizeof(struct pasemi_session)); ++ ++ return 0; ++} ++ ++static int ++pasemi_process(device_t dev, struct cryptop *crp, int hint) ++{ ++ ++ int err = 0, ivsize, srclen = 0, reinit = 0, reinit_size = 0, chsel; ++ struct pasemi_softc *sc = device_get_softc(dev); ++ struct cryptodesc *crd1, *crd2, *maccrd, *enccrd; ++ caddr_t ivp; ++ struct pasemi_desc init_desc, work_desc; ++ struct pasemi_session *ses; ++ struct sk_buff *skb; ++ struct uio *uiop; ++ unsigned long flags; ++ struct pasemi_fnu_txring *txring; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ if (crp == NULL || crp->crp_callback == NULL || sc == NULL) ++ return -EINVAL; ++ ++ crp->crp_etype = 0; ++ if (PASEMI_SESSION(crp->crp_sid) >= sc->sc_nsessions) ++ return -EINVAL; ++ ++ ses = sc->sc_sessions[PASEMI_SESSION(crp->crp_sid)]; ++ ++ crd1 = crp->crp_desc; ++ if (crd1 == NULL) { ++ err = -EINVAL; ++ goto errout; ++ } ++ crd2 = crd1->crd_next; ++ ++ if (ALG_IS_SIG(crd1->crd_alg)) { ++ maccrd = crd1; ++ if (crd2 == NULL) ++ enccrd = NULL; ++ else if (ALG_IS_CIPHER(crd2->crd_alg) && ++ (crd2->crd_flags & CRD_F_ENCRYPT) == 0) ++ enccrd = crd2; ++ else ++ goto erralg; ++ } else if (ALG_IS_CIPHER(crd1->crd_alg)) { ++ enccrd = crd1; ++ if (crd2 == NULL) ++ maccrd = NULL; ++ else if (ALG_IS_SIG(crd2->crd_alg) && ++ (crd1->crd_flags & CRD_F_ENCRYPT)) ++ maccrd = crd2; ++ else ++ goto erralg; ++ } else ++ goto erralg; ++ ++ chsel = ses->chan; ++ ++ txring = &sc->tx[chsel]; ++ ++ if (enccrd && !maccrd) { ++ if (enccrd->crd_alg == CRYPTO_ARC4) ++ reinit = 1; ++ reinit_size = 0x40; ++ srclen = crp->crp_ilen; ++ ++ pasemi_desc_start(&work_desc, XCT_FUN_O | XCT_FUN_I ++ | XCT_FUN_FUN(chsel)); ++ if (enccrd->crd_flags & CRD_F_ENCRYPT) ++ pasemi_desc_hdr(&work_desc, XCT_FUN_CRM_ENC); ++ else ++ pasemi_desc_hdr(&work_desc, XCT_FUN_CRM_DEC); ++ } else if (enccrd && maccrd) { ++ if (enccrd->crd_alg == CRYPTO_ARC4) ++ reinit = 1; ++ reinit_size = 0x68; ++ ++ if (enccrd->crd_flags & CRD_F_ENCRYPT) { ++ /* Encrypt -> Authenticate */ ++ pasemi_desc_start(&work_desc, XCT_FUN_O | XCT_FUN_I | XCT_FUN_CRM_ENC_SIG ++ | XCT_FUN_A | XCT_FUN_FUN(chsel)); ++ srclen = maccrd->crd_skip + maccrd->crd_len; ++ } else { ++ /* Authenticate -> Decrypt */ ++ pasemi_desc_start(&work_desc, XCT_FUN_O | XCT_FUN_I | XCT_FUN_CRM_SIG_DEC ++ | XCT_FUN_24BRES | XCT_FUN_FUN(chsel)); ++ pasemi_desc_build(&work_desc, 0); ++ pasemi_desc_build(&work_desc, 0); ++ pasemi_desc_build(&work_desc, 0); ++ work_desc.postop = PASEMI_CHECK_SIG; ++ srclen = crp->crp_ilen; ++ } ++ ++ pasemi_desc_hdr(&work_desc, XCT_FUN_SHL(maccrd->crd_skip / 4)); ++ pasemi_desc_hdr(&work_desc, XCT_FUN_CHL(enccrd->crd_skip - maccrd->crd_skip)); ++ } else if (!enccrd && maccrd) { ++ srclen = maccrd->crd_len; ++ ++ pasemi_desc_start(&init_desc, ++ XCT_CTRL_HDR(chsel, 0x58, DMA_FN_HKEY0)); ++ pasemi_desc_build(&init_desc, ++ XCT_FUN_SRC_PTR(0x58, ((struct pasemi_session *)ses->dma_addr)->hkey)); ++ ++ pasemi_desc_start(&work_desc, XCT_FUN_O | XCT_FUN_I | XCT_FUN_CRM_SIG ++ | XCT_FUN_A | XCT_FUN_FUN(chsel)); ++ } ++ ++ if (enccrd) { ++ switch (enccrd->crd_alg) { ++ case CRYPTO_3DES_CBC: ++ pasemi_desc_hdr(&work_desc, XCT_FUN_ALG_3DES | ++ XCT_FUN_BCM_CBC); ++ ivsize = sizeof(u64); ++ break; ++ case CRYPTO_DES_CBC: ++ pasemi_desc_hdr(&work_desc, XCT_FUN_ALG_DES | ++ XCT_FUN_BCM_CBC); ++ ivsize = sizeof(u64); ++ break; ++ case CRYPTO_AES_CBC: ++ pasemi_desc_hdr(&work_desc, XCT_FUN_ALG_AES | ++ XCT_FUN_BCM_CBC); ++ ivsize = 2 * sizeof(u64); ++ break; ++ case CRYPTO_ARC4: ++ pasemi_desc_hdr(&work_desc, XCT_FUN_ALG_ARC); ++ ivsize = 0; ++ break; ++ default: ++ printk(DRV_NAME ": unimplemented enccrd->crd_alg %d\n", ++ enccrd->crd_alg); ++ err = -EINVAL; ++ goto errout; ++ } ++ ++ ivp = (ivsize == sizeof(u64)) ? (caddr_t) &ses->civ[1] : (caddr_t) &ses->civ[0]; ++ if (enccrd->crd_flags & CRD_F_ENCRYPT) { ++ if (enccrd->crd_flags & CRD_F_IV_EXPLICIT) ++ memcpy(ivp, enccrd->crd_iv, ivsize); ++ /* If IV is not present in the buffer already, it has to be copied there */ ++ if ((enccrd->crd_flags & CRD_F_IV_PRESENT) == 0) ++ crypto_copyback(crp->crp_flags, crp->crp_buf, ++ enccrd->crd_inject, ivsize, ivp); ++ } else { ++ if (enccrd->crd_flags & CRD_F_IV_EXPLICIT) ++ /* IV is provided expicitly in descriptor */ ++ memcpy(ivp, enccrd->crd_iv, ivsize); ++ else ++ /* IV is provided in the packet */ ++ crypto_copydata(crp->crp_flags, crp->crp_buf, ++ enccrd->crd_inject, ivsize, ++ ivp); ++ } ++ } ++ ++ if (maccrd) { ++ switch (maccrd->crd_alg) { ++ case CRYPTO_MD5: ++ pasemi_desc_hdr(&work_desc, XCT_FUN_SIG_MD5 | ++ XCT_FUN_HSZ((crp->crp_ilen - maccrd->crd_inject) / 4)); ++ break; ++ case CRYPTO_SHA1: ++ pasemi_desc_hdr(&work_desc, XCT_FUN_SIG_SHA1 | ++ XCT_FUN_HSZ((crp->crp_ilen - maccrd->crd_inject) / 4)); ++ break; ++ case CRYPTO_MD5_HMAC: ++ pasemi_desc_hdr(&work_desc, XCT_FUN_SIG_HMAC_MD5 | ++ XCT_FUN_HSZ((crp->crp_ilen - maccrd->crd_inject) / 4)); ++ break; ++ case CRYPTO_SHA1_HMAC: ++ pasemi_desc_hdr(&work_desc, XCT_FUN_SIG_HMAC_SHA1 | ++ XCT_FUN_HSZ((crp->crp_ilen - maccrd->crd_inject) / 4)); ++ break; ++ default: ++ printk(DRV_NAME ": unimplemented maccrd->crd_alg %d\n", ++ maccrd->crd_alg); ++ err = -EINVAL; ++ goto errout; ++ } ++ } ++ ++ if (crp->crp_flags & CRYPTO_F_SKBUF) { ++ /* using SKB buffers */ ++ skb = (struct sk_buff *)crp->crp_buf; ++ if (skb_shinfo(skb)->nr_frags) { ++ printk(DRV_NAME ": skb frags unimplemented\n"); ++ err = -EINVAL; ++ goto errout; ++ } ++ pasemi_desc_build( ++ &work_desc, ++ XCT_FUN_DST_PTR(skb->len, pci_map_single( ++ sc->dma_pdev, skb->data, ++ skb->len, DMA_TO_DEVICE))); ++ pasemi_desc_build( ++ &work_desc, ++ XCT_FUN_SRC_PTR( ++ srclen, pci_map_single( ++ sc->dma_pdev, skb->data, ++ srclen, DMA_TO_DEVICE))); ++ pasemi_desc_hdr(&work_desc, XCT_FUN_LLEN(srclen)); ++ } else if (crp->crp_flags & CRYPTO_F_IOV) { ++ /* using IOV buffers */ ++ uiop = (struct uio *)crp->crp_buf; ++ if (uiop->uio_iovcnt > 1) { ++ printk(DRV_NAME ": iov frags unimplemented\n"); ++ err = -EINVAL; ++ goto errout; ++ } ++ ++ /* crp_olen is never set; always use crp_ilen */ ++ pasemi_desc_build( ++ &work_desc, ++ XCT_FUN_DST_PTR(crp->crp_ilen, pci_map_single( ++ sc->dma_pdev, ++ uiop->uio_iov->iov_base, ++ crp->crp_ilen, DMA_TO_DEVICE))); ++ pasemi_desc_hdr(&work_desc, XCT_FUN_LLEN(srclen)); ++ ++ pasemi_desc_build( ++ &work_desc, ++ XCT_FUN_SRC_PTR(srclen, pci_map_single( ++ sc->dma_pdev, ++ uiop->uio_iov->iov_base, ++ srclen, DMA_TO_DEVICE))); ++ } else { ++ /* using contig buffers */ ++ pasemi_desc_build( ++ &work_desc, ++ XCT_FUN_DST_PTR(crp->crp_ilen, pci_map_single( ++ sc->dma_pdev, ++ crp->crp_buf, ++ crp->crp_ilen, DMA_TO_DEVICE))); ++ pasemi_desc_build( ++ &work_desc, ++ XCT_FUN_SRC_PTR(srclen, pci_map_single( ++ sc->dma_pdev, ++ crp->crp_buf, srclen, ++ DMA_TO_DEVICE))); ++ pasemi_desc_hdr(&work_desc, XCT_FUN_LLEN(srclen)); ++ } ++ ++ spin_lock_irqsave(&txring->fill_lock, flags); ++ ++ if (txring->sesn != PASEMI_SESSION(crp->crp_sid)) { ++ txring->sesn = PASEMI_SESSION(crp->crp_sid); ++ reinit = 1; ++ } ++ ++ if (enccrd) { ++ pasemi_desc_start(&init_desc, ++ XCT_CTRL_HDR(chsel, reinit ? reinit_size : 0x10, DMA_FN_CIV0)); ++ pasemi_desc_build(&init_desc, ++ XCT_FUN_SRC_PTR(reinit ? reinit_size : 0x10, ses->dma_addr)); ++ } ++ ++ if (((txring->next_to_fill + pasemi_desc_size(&init_desc) + ++ pasemi_desc_size(&work_desc)) - ++ txring->next_to_clean) > TX_RING_SIZE) { ++ spin_unlock_irqrestore(&txring->fill_lock, flags); ++ err = ERESTART; ++ goto errout; ++ } ++ ++ pasemi_ring_add_desc(txring, &init_desc, NULL); ++ pasemi_ring_add_desc(txring, &work_desc, crp); ++ ++ pasemi_ring_incr(sc, chsel, ++ pasemi_desc_size(&init_desc) + ++ pasemi_desc_size(&work_desc)); ++ ++ spin_unlock_irqrestore(&txring->fill_lock, flags); ++ ++ mod_timer(&txring->crypto_timer, jiffies + TIMER_INTERVAL); ++ ++ return 0; ++ ++erralg: ++ printk(DRV_NAME ": unsupported algorithm or algorithm order alg1 %d alg2 %d\n", ++ crd1->crd_alg, crd2->crd_alg); ++ err = -EINVAL; ++ ++errout: ++ if (err != ERESTART) { ++ crp->crp_etype = err; ++ crypto_done(crp); ++ } ++ return err; ++} ++ ++static int pasemi_clean_tx(struct pasemi_softc *sc, int chan) ++{ ++ int i, j, ring_idx; ++ struct pasemi_fnu_txring *ring = &sc->tx[chan]; ++ u16 delta_cnt; ++ int flags, loops = 10; ++ int desc_size; ++ struct cryptop *crp; ++ ++ spin_lock_irqsave(&ring->clean_lock, flags); ++ ++ while ((delta_cnt = (dma_status->tx_sta[sc->base_chan + chan] ++ & PAS_STATUS_PCNT_M) - ring->total_pktcnt) ++ && loops--) { ++ ++ for (i = 0; i < delta_cnt; i++) { ++ desc_size = TX_DESC_INFO(ring, ring->next_to_clean).desc_size; ++ crp = TX_DESC_INFO(ring, ring->next_to_clean).cf_crp; ++ if (crp) { ++ ring_idx = 2 * (ring->next_to_clean & (TX_RING_SIZE-1)); ++ if (TX_DESC_INFO(ring, ring->next_to_clean).desc_postop & PASEMI_CHECK_SIG) { ++ /* Need to make sure signature matched, ++ * if not - return error */ ++ if (!(ring->desc[ring_idx + 1] & (1ULL << 63))) ++ crp->crp_etype = -EINVAL; ++ } ++ crypto_done(TX_DESC_INFO(ring, ++ ring->next_to_clean).cf_crp); ++ TX_DESC_INFO(ring, ring->next_to_clean).cf_crp = NULL; ++ pci_unmap_single( ++ sc->dma_pdev, ++ XCT_PTR_ADDR_LEN(ring->desc[ring_idx + 1]), ++ PCI_DMA_TODEVICE); ++ ++ ring->desc[ring_idx] = ring->desc[ring_idx + 1] = 0; ++ ++ ring->next_to_clean++; ++ for (j = 1; j < desc_size; j++) { ++ ring_idx = 2 * ++ (ring->next_to_clean & ++ (TX_RING_SIZE-1)); ++ pci_unmap_single( ++ sc->dma_pdev, ++ XCT_PTR_ADDR_LEN(ring->desc[ring_idx]), ++ PCI_DMA_TODEVICE); ++ if (ring->desc[ring_idx + 1]) ++ pci_unmap_single( ++ sc->dma_pdev, ++ XCT_PTR_ADDR_LEN( ++ ring->desc[ ++ ring_idx + 1]), ++ PCI_DMA_TODEVICE); ++ ring->desc[ring_idx] = ++ ring->desc[ring_idx + 1] = 0; ++ ring->next_to_clean++; ++ } ++ } else { ++ for (j = 0; j < desc_size; j++) { ++ ring_idx = 2 * (ring->next_to_clean & (TX_RING_SIZE-1)); ++ ring->desc[ring_idx] = ++ ring->desc[ring_idx + 1] = 0; ++ ring->next_to_clean++; ++ } ++ } ++ } ++ ++ ring->total_pktcnt += delta_cnt; ++ } ++ spin_unlock_irqrestore(&ring->clean_lock, flags); ++ ++ return 0; ++} ++ ++static void sweepup_tx(struct pasemi_softc *sc) ++{ ++ int i; ++ ++ for (i = 0; i < sc->sc_num_channels; i++) ++ pasemi_clean_tx(sc, i); ++} ++ ++static irqreturn_t pasemi_intr(int irq, void *arg, struct pt_regs *regs) ++{ ++ struct pasemi_softc *sc = arg; ++ unsigned int reg; ++ int chan = irq - sc->base_irq; ++ int chan_index = sc->base_chan + chan; ++ u64 stat = dma_status->tx_sta[chan_index]; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ if (!(stat & PAS_STATUS_CAUSE_M)) ++ return IRQ_NONE; ++ ++ pasemi_clean_tx(sc, chan); ++ ++ stat = dma_status->tx_sta[chan_index]; ++ ++ reg = PAS_IOB_DMA_TXCH_RESET_PINTC | ++ PAS_IOB_DMA_TXCH_RESET_PCNT(sc->tx[chan].total_pktcnt); ++ ++ if (stat & PAS_STATUS_SOFT) ++ reg |= PAS_IOB_DMA_RXCH_RESET_SINTC; ++ ++ out_le32(sc->iob_regs + PAS_IOB_DMA_TXCH_RESET(chan_index), reg); ++ ++ ++ return IRQ_HANDLED; ++} ++ ++static int pasemi_dma_setup_tx_resources(struct pasemi_softc *sc, int chan) ++{ ++ u32 val; ++ int chan_index = chan + sc->base_chan; ++ int ret; ++ struct pasemi_fnu_txring *ring; ++ ++ ring = &sc->tx[chan]; ++ ++ spin_lock_init(&ring->fill_lock); ++ spin_lock_init(&ring->clean_lock); ++ ++ ring->desc_info = kzalloc(sizeof(struct pasemi_desc_info) * ++ TX_RING_SIZE, GFP_KERNEL); ++ if (!ring->desc_info) ++ return -ENOMEM; ++ ++ /* Allocate descriptors */ ++ ring->desc = dma_alloc_coherent(&sc->dma_pdev->dev, ++ TX_RING_SIZE * ++ 2 * sizeof(u64), ++ &ring->dma, GFP_KERNEL); ++ if (!ring->desc) ++ return -ENOMEM; ++ ++ memset((void *) ring->desc, 0, TX_RING_SIZE * 2 * sizeof(u64)); ++ ++ out_le32(sc->iob_regs + PAS_IOB_DMA_TXCH_RESET(chan_index), 0x30); ++ ++ ring->total_pktcnt = 0; ++ ++ out_le32(sc->dma_regs + PAS_DMA_TXCHAN_BASEL(chan_index), ++ PAS_DMA_TXCHAN_BASEL_BRBL(ring->dma)); ++ ++ val = PAS_DMA_TXCHAN_BASEU_BRBH(ring->dma >> 32); ++ val |= PAS_DMA_TXCHAN_BASEU_SIZ(TX_RING_SIZE >> 2); ++ ++ out_le32(sc->dma_regs + PAS_DMA_TXCHAN_BASEU(chan_index), val); ++ ++ out_le32(sc->dma_regs + PAS_DMA_TXCHAN_CFG(chan_index), ++ PAS_DMA_TXCHAN_CFG_TY_FUNC | ++ PAS_DMA_TXCHAN_CFG_TATTR(chan) | ++ PAS_DMA_TXCHAN_CFG_WT(2)); ++ ++ /* enable tx channel */ ++ out_le32(sc->dma_regs + ++ PAS_DMA_TXCHAN_TCMDSTA(chan_index), ++ PAS_DMA_TXCHAN_TCMDSTA_EN); ++ ++ out_le32(sc->iob_regs + PAS_IOB_DMA_TXCH_CFG(chan_index), ++ PAS_IOB_DMA_TXCH_CFG_CNTTH(1000)); ++ ++ ring->next_to_fill = 0; ++ ring->next_to_clean = 0; ++ ++ snprintf(ring->irq_name, sizeof(ring->irq_name), ++ "%s%d", "crypto", chan); ++ ++ ring->irq = irq_create_mapping(NULL, sc->base_irq + chan); ++ ret = request_irq(ring->irq, (irq_handler_t) ++ pasemi_intr, IRQF_DISABLED, ring->irq_name, sc); ++ if (ret) { ++ printk(KERN_ERR DRV_NAME ": failed to hook irq %d ret %d\n", ++ ring->irq, ret); ++ ring->irq = -1; ++ return ret; ++ } ++ ++ setup_timer(&ring->crypto_timer, (void *) sweepup_tx, (unsigned long) sc); ++ ++ return 0; ++} ++ ++static device_method_t pasemi_methods = { ++ /* crypto device methods */ ++ DEVMETHOD(cryptodev_newsession, pasemi_newsession), ++ DEVMETHOD(cryptodev_freesession, pasemi_freesession), ++ DEVMETHOD(cryptodev_process, pasemi_process), ++}; ++ ++/* Set up the crypto device structure, private data, ++ * and anything else we need before we start */ ++ ++static int __devinit ++pasemi_dma_probe(struct pci_dev *pdev, const struct pci_device_id *ent) ++{ ++ struct pasemi_softc *sc; ++ int ret, i; ++ ++ DPRINTF(KERN_ERR "%s()\n", __FUNCTION__); ++ ++ sc = kzalloc(sizeof(*sc), GFP_KERNEL); ++ if (!sc) ++ return -ENOMEM; ++ ++ softc_device_init(sc, DRV_NAME, 1, pasemi_methods); ++ ++ pci_set_drvdata(pdev, sc); ++ ++ spin_lock_init(&sc->sc_chnlock); ++ ++ sc->sc_sessions = (struct pasemi_session **) ++ kzalloc(PASEMI_INITIAL_SESSIONS * ++ sizeof(struct pasemi_session *), GFP_ATOMIC); ++ if (sc->sc_sessions == NULL) { ++ ret = -ENOMEM; ++ goto out; ++ } ++ ++ sc->sc_nsessions = PASEMI_INITIAL_SESSIONS; ++ sc->sc_lastchn = 0; ++ sc->base_irq = pdev->irq + 6; ++ sc->base_chan = 6; ++ sc->sc_cid = -1; ++ sc->dma_pdev = pdev; ++ ++ sc->iob_pdev = pci_get_device(PCI_VENDOR_ID_PASEMI, 0xa001, NULL); ++ if (!sc->iob_pdev) { ++ dev_err(&pdev->dev, "Can't find I/O Bridge\n"); ++ ret = -ENODEV; ++ goto out; ++ } ++ ++ /* This is hardcoded and ugly, but we have some firmware versions ++ * who don't provide the register space in the device tree. Luckily ++ * they are at well-known locations so we can just do the math here. ++ */ ++ sc->dma_regs = ++ ioremap(0xe0000000 + (sc->dma_pdev->devfn << 12), 0x2000); ++ sc->iob_regs = ++ ioremap(0xe0000000 + (sc->iob_pdev->devfn << 12), 0x2000); ++ if (!sc->dma_regs || !sc->iob_regs) { ++ dev_err(&pdev->dev, "Can't map registers\n"); ++ ret = -ENODEV; ++ goto out; ++ } ++ ++ dma_status = __ioremap(0xfd800000, 0x1000, 0); ++ if (!dma_status) { ++ ret = -ENODEV; ++ dev_err(&pdev->dev, "Can't map dmastatus space\n"); ++ goto out; ++ } ++ ++ sc->tx = (struct pasemi_fnu_txring *) ++ kzalloc(sizeof(struct pasemi_fnu_txring) ++ * 8, GFP_KERNEL); ++ if (!sc->tx) { ++ ret = -ENOMEM; ++ goto out; ++ } ++ ++ /* Initialize the h/w */ ++ out_le32(sc->dma_regs + PAS_DMA_COM_CFG, ++ (in_le32(sc->dma_regs + PAS_DMA_COM_CFG) | ++ PAS_DMA_COM_CFG_FWF)); ++ out_le32(sc->dma_regs + PAS_DMA_COM_TXCMD, PAS_DMA_COM_TXCMD_EN); ++ ++ for (i = 0; i < PASEMI_FNU_CHANNELS; i++) { ++ sc->sc_num_channels++; ++ ret = pasemi_dma_setup_tx_resources(sc, i); ++ if (ret) ++ goto out; ++ } ++ ++ sc->sc_cid = crypto_get_driverid(softc_get_device(sc), ++ CRYPTOCAP_F_HARDWARE); ++ if (sc->sc_cid < 0) { ++ printk(KERN_ERR DRV_NAME ": could not get crypto driver id\n"); ++ ret = -ENXIO; ++ goto out; ++ } ++ ++ /* register algorithms with the framework */ ++ printk(DRV_NAME ":"); ++ ++ crypto_register(sc->sc_cid, CRYPTO_DES_CBC, 0, 0); ++ crypto_register(sc->sc_cid, CRYPTO_3DES_CBC, 0, 0); ++ crypto_register(sc->sc_cid, CRYPTO_AES_CBC, 0, 0); ++ crypto_register(sc->sc_cid, CRYPTO_ARC4, 0, 0); ++ crypto_register(sc->sc_cid, CRYPTO_SHA1, 0, 0); ++ crypto_register(sc->sc_cid, CRYPTO_MD5, 0, 0); ++ crypto_register(sc->sc_cid, CRYPTO_SHA1_HMAC, 0, 0); ++ crypto_register(sc->sc_cid, CRYPTO_MD5_HMAC, 0, 0); ++ ++ return 0; ++ ++out: ++ pasemi_dma_remove(pdev); ++ return ret; ++} ++ ++#define MAX_RETRIES 5000 ++ ++static void pasemi_free_tx_resources(struct pasemi_softc *sc, int chan) ++{ ++ struct pasemi_fnu_txring *ring = &sc->tx[chan]; ++ int chan_index = chan + sc->base_chan; ++ int retries; ++ u32 stat; ++ ++ /* Stop the channel */ ++ out_le32(sc->dma_regs + ++ PAS_DMA_TXCHAN_TCMDSTA(chan_index), ++ PAS_DMA_TXCHAN_TCMDSTA_ST); ++ ++ for (retries = 0; retries < MAX_RETRIES; retries++) { ++ stat = in_le32(sc->dma_regs + ++ PAS_DMA_TXCHAN_TCMDSTA(chan_index)); ++ if (!(stat & PAS_DMA_TXCHAN_TCMDSTA_ACT)) ++ break; ++ cond_resched(); ++ } ++ ++ if (stat & PAS_DMA_TXCHAN_TCMDSTA_ACT) ++ dev_err(&sc->dma_pdev->dev, "Failed to stop tx channel %d\n", ++ chan_index); ++ ++ /* Disable the channel */ ++ out_le32(sc->dma_regs + ++ PAS_DMA_TXCHAN_TCMDSTA(chan_index), ++ 0); ++ ++ if (ring->desc_info) ++ kfree((void *) ring->desc_info); ++ if (ring->desc) ++ dma_free_coherent(&sc->dma_pdev->dev, ++ TX_RING_SIZE * ++ 2 * sizeof(u64), ++ (void *) ring->desc, ring->dma); ++ if (ring->irq != -1) ++ free_irq(ring->irq, sc); ++ ++ del_timer(&ring->crypto_timer); ++} ++ ++static void __devexit pasemi_dma_remove(struct pci_dev *pdev) ++{ ++ struct pasemi_softc *sc = pci_get_drvdata(pdev); ++ int i; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ if (sc->sc_cid >= 0) { ++ crypto_unregister_all(sc->sc_cid); ++ } ++ ++ if (sc->tx) { ++ for (i = 0; i < sc->sc_num_channels; i++) ++ pasemi_free_tx_resources(sc, i); ++ ++ kfree(sc->tx); ++ } ++ if (sc->sc_sessions) { ++ for (i = 0; i < sc->sc_nsessions; i++) ++ kfree(sc->sc_sessions[i]); ++ kfree(sc->sc_sessions); ++ } ++ if (sc->iob_pdev) ++ pci_dev_put(sc->iob_pdev); ++ if (sc->dma_regs) ++ iounmap(sc->dma_regs); ++ if (sc->iob_regs) ++ iounmap(sc->iob_regs); ++ kfree(sc); ++} ++ ++static struct pci_device_id pasemi_dma_pci_tbl[] = { ++ { PCI_DEVICE(PCI_VENDOR_ID_PASEMI, 0xa007) }, ++}; ++ ++MODULE_DEVICE_TABLE(pci, pasemi_dma_pci_tbl); ++ ++static struct pci_driver pasemi_dma_driver = { ++ .name = "pasemi_dma", ++ .id_table = pasemi_dma_pci_tbl, ++ .probe = pasemi_dma_probe, ++ .remove = __devexit_p(pasemi_dma_remove), ++}; ++ ++static void __exit pasemi_dma_cleanup_module(void) ++{ ++ pci_unregister_driver(&pasemi_dma_driver); ++ __iounmap(dma_status); ++ dma_status = NULL; ++} ++ ++int pasemi_dma_init_module(void) ++{ ++ return pci_register_driver(&pasemi_dma_driver); ++} ++ ++module_init(pasemi_dma_init_module); ++module_exit(pasemi_dma_cleanup_module); ++ ++MODULE_LICENSE("Dual BSD/GPL"); ++MODULE_AUTHOR("Egor Martovetsky egor@pasemi.com"); ++MODULE_DESCRIPTION("OCF driver for PA Semi PWRficient DMA Crypto Engine"); +diff -Nur linux-2.6.30.orig/crypto/ocf/pasemi/pasemi_fnu.h linux-2.6.30/crypto/ocf/pasemi/pasemi_fnu.h +--- linux-2.6.30.orig/crypto/ocf/pasemi/pasemi_fnu.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/pasemi/pasemi_fnu.h 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,410 @@ ++/* ++ * Copyright (C) 2007 PA Semi, Inc ++ * ++ * Driver for the PA Semi PWRficient DMA Crypto Engine, soft state and ++ * hardware register layouts. ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License version 2 as ++ * published by the Free Software Foundation. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ * ++ * You should have received a copy of the GNU General Public License ++ * along with this program; if not, write to the Free Software ++ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA ++ */ ++ ++#ifndef PASEMI_FNU_H ++#define PASEMI_FNU_H ++ ++#include <linux/spinlock.h> ++ ++#define PASEMI_SESSION(sid) ((sid) & 0xffffffff) ++#define PASEMI_SID(sesn) ((sesn) & 0xffffffff) ++#define DPRINTF(a...) if (debug) { printk(DRV_NAME ": " a); } ++ ++/* Must be a power of two */ ++#define RX_RING_SIZE 512 ++#define TX_RING_SIZE 512 ++#define TX_DESC(ring, num) ((ring)->desc[2 * (num & (TX_RING_SIZE-1))]) ++#define TX_DESC_INFO(ring, num) ((ring)->desc_info[(num) & (TX_RING_SIZE-1)]) ++#define MAX_DESC_SIZE 8 ++#define PASEMI_INITIAL_SESSIONS 10 ++#define PASEMI_FNU_CHANNELS 8 ++ ++/* DMA descriptor */ ++struct pasemi_desc { ++ u64 quad[2*MAX_DESC_SIZE]; ++ int quad_cnt; ++ int size; ++ int postop; ++}; ++ ++/* ++ * Holds per descriptor data ++ */ ++struct pasemi_desc_info { ++ int desc_size; ++ int desc_postop; ++#define PASEMI_CHECK_SIG 0x1 ++ ++ struct cryptop *cf_crp; ++}; ++ ++/* ++ * Holds per channel data ++ */ ++struct pasemi_fnu_txring { ++ volatile u64 *desc; ++ volatile struct ++ pasemi_desc_info *desc_info; ++ dma_addr_t dma; ++ struct timer_list crypto_timer; ++ spinlock_t fill_lock; ++ spinlock_t clean_lock; ++ unsigned int next_to_fill; ++ unsigned int next_to_clean; ++ u16 total_pktcnt; ++ int irq; ++ int sesn; ++ char irq_name[10]; ++}; ++ ++/* ++ * Holds data specific to a single pasemi device. ++ */ ++struct pasemi_softc { ++ softc_device_decl sc_cdev; ++ struct pci_dev *dma_pdev; /* device backpointer */ ++ struct pci_dev *iob_pdev; /* device backpointer */ ++ void __iomem *dma_regs; ++ void __iomem *iob_regs; ++ int base_irq; ++ int base_chan; ++ int32_t sc_cid; /* crypto tag */ ++ int sc_nsessions; ++ struct pasemi_session **sc_sessions; ++ int sc_num_channels;/* number of crypto channels */ ++ ++ /* pointer to the array of txring datastructures, one txring per channel */ ++ struct pasemi_fnu_txring *tx; ++ ++ /* ++ * mutual exclusion for the channel scheduler ++ */ ++ spinlock_t sc_chnlock; ++ /* last channel used, for now use round-robin to allocate channels */ ++ int sc_lastchn; ++}; ++ ++struct pasemi_session { ++ u64 civ[2]; ++ u64 keysz; ++ u64 key[4]; ++ u64 ccmd; ++ u64 hkey[4]; ++ u64 hseq; ++ u64 giv[2]; ++ u64 hiv[4]; ++ ++ int used; ++ dma_addr_t dma_addr; ++ int chan; ++}; ++ ++/* status register layout in IOB region, at 0xfd800000 */ ++struct pasdma_status { ++ u64 rx_sta[64]; ++ u64 tx_sta[20]; ++}; ++ ++#define ALG_IS_CIPHER(alg) ((alg == CRYPTO_DES_CBC) || \ ++ (alg == CRYPTO_3DES_CBC) || \ ++ (alg == CRYPTO_AES_CBC) || \ ++ (alg == CRYPTO_ARC4) || \ ++ (alg == CRYPTO_NULL_CBC)) ++ ++#define ALG_IS_SIG(alg) ((alg == CRYPTO_MD5) || \ ++ (alg == CRYPTO_MD5_HMAC) || \ ++ (alg == CRYPTO_SHA1) || \ ++ (alg == CRYPTO_SHA1_HMAC) || \ ++ (alg == CRYPTO_NULL_HMAC)) ++ ++enum { ++ PAS_DMA_COM_TXCMD = 0x100, /* Transmit Command Register */ ++ PAS_DMA_COM_TXSTA = 0x104, /* Transmit Status Register */ ++ PAS_DMA_COM_RXCMD = 0x108, /* Receive Command Register */ ++ PAS_DMA_COM_RXSTA = 0x10c, /* Receive Status Register */ ++ PAS_DMA_COM_CFG = 0x114, /* DMA Configuration Register */ ++}; ++ ++/* All these registers live in the PCI configuration space for the DMA PCI ++ * device. Use the normal PCI config access functions for them. ++ */ ++ ++#define PAS_DMA_COM_CFG_FWF 0x18000000 ++ ++#define PAS_DMA_COM_TXCMD_EN 0x00000001 /* enable */ ++#define PAS_DMA_COM_TXSTA_ACT 0x00000001 /* active */ ++#define PAS_DMA_COM_RXCMD_EN 0x00000001 /* enable */ ++#define PAS_DMA_COM_RXSTA_ACT 0x00000001 /* active */ ++ ++#define _PAS_DMA_TXCHAN_STRIDE 0x20 /* Size per channel */ ++#define _PAS_DMA_TXCHAN_TCMDSTA 0x300 /* Command / Status */ ++#define _PAS_DMA_TXCHAN_CFG 0x304 /* Configuration */ ++#define _PAS_DMA_TXCHAN_DSCRBU 0x308 /* Descriptor BU Allocation */ ++#define _PAS_DMA_TXCHAN_INCR 0x310 /* Descriptor increment */ ++#define _PAS_DMA_TXCHAN_CNT 0x314 /* Descriptor count/offset */ ++#define _PAS_DMA_TXCHAN_BASEL 0x318 /* Descriptor ring base (low) */ ++#define _PAS_DMA_TXCHAN_BASEU 0x31c /* (high) */ ++#define PAS_DMA_TXCHAN_TCMDSTA(c) (0x300+(c)*_PAS_DMA_TXCHAN_STRIDE) ++#define PAS_DMA_TXCHAN_TCMDSTA_EN 0x00000001 /* Enabled */ ++#define PAS_DMA_TXCHAN_TCMDSTA_ST 0x00000002 /* Stop interface */ ++#define PAS_DMA_TXCHAN_TCMDSTA_ACT 0x00010000 /* Active */ ++#define PAS_DMA_TXCHAN_CFG(c) (0x304+(c)*_PAS_DMA_TXCHAN_STRIDE) ++#define PAS_DMA_TXCHAN_CFG_TY_FUNC 0x00000002 /* Type = interface */ ++#define PAS_DMA_TXCHAN_CFG_TY_IFACE 0x00000000 /* Type = interface */ ++#define PAS_DMA_TXCHAN_CFG_TATTR_M 0x0000003c ++#define PAS_DMA_TXCHAN_CFG_TATTR_S 2 ++#define PAS_DMA_TXCHAN_CFG_TATTR(x) (((x) << PAS_DMA_TXCHAN_CFG_TATTR_S) & \ ++ PAS_DMA_TXCHAN_CFG_TATTR_M) ++#define PAS_DMA_TXCHAN_CFG_WT_M 0x000001c0 ++#define PAS_DMA_TXCHAN_CFG_WT_S 6 ++#define PAS_DMA_TXCHAN_CFG_WT(x) (((x) << PAS_DMA_TXCHAN_CFG_WT_S) & \ ++ PAS_DMA_TXCHAN_CFG_WT_M) ++#define PAS_DMA_TXCHAN_CFG_LPSQ_FAST 0x00000400 ++#define PAS_DMA_TXCHAN_CFG_LPDQ_FAST 0x00000800 ++#define PAS_DMA_TXCHAN_CFG_CF 0x00001000 /* Clean first line */ ++#define PAS_DMA_TXCHAN_CFG_CL 0x00002000 /* Clean last line */ ++#define PAS_DMA_TXCHAN_CFG_UP 0x00004000 /* update tx descr when sent */ ++#define PAS_DMA_TXCHAN_INCR(c) (0x310+(c)*_PAS_DMA_TXCHAN_STRIDE) ++#define PAS_DMA_TXCHAN_BASEL(c) (0x318+(c)*_PAS_DMA_TXCHAN_STRIDE) ++#define PAS_DMA_TXCHAN_BASEL_BRBL_M 0xffffffc0 ++#define PAS_DMA_TXCHAN_BASEL_BRBL_S 0 ++#define PAS_DMA_TXCHAN_BASEL_BRBL(x) (((x) << PAS_DMA_TXCHAN_BASEL_BRBL_S) & \ ++ PAS_DMA_TXCHAN_BASEL_BRBL_M) ++#define PAS_DMA_TXCHAN_BASEU(c) (0x31c+(c)*_PAS_DMA_TXCHAN_STRIDE) ++#define PAS_DMA_TXCHAN_BASEU_BRBH_M 0x00000fff ++#define PAS_DMA_TXCHAN_BASEU_BRBH_S 0 ++#define PAS_DMA_TXCHAN_BASEU_BRBH(x) (((x) << PAS_DMA_TXCHAN_BASEU_BRBH_S) & \ ++ PAS_DMA_TXCHAN_BASEU_BRBH_M) ++/* # of cache lines worth of buffer ring */ ++#define PAS_DMA_TXCHAN_BASEU_SIZ_M 0x3fff0000 ++#define PAS_DMA_TXCHAN_BASEU_SIZ_S 16 /* 0 = 16K */ ++#define PAS_DMA_TXCHAN_BASEU_SIZ(x) (((x) << PAS_DMA_TXCHAN_BASEU_SIZ_S) & \ ++ PAS_DMA_TXCHAN_BASEU_SIZ_M) ++ ++#define PAS_STATUS_PCNT_M 0x000000000000ffffull ++#define PAS_STATUS_PCNT_S 0 ++#define PAS_STATUS_DCNT_M 0x00000000ffff0000ull ++#define PAS_STATUS_DCNT_S 16 ++#define PAS_STATUS_BPCNT_M 0x0000ffff00000000ull ++#define PAS_STATUS_BPCNT_S 32 ++#define PAS_STATUS_CAUSE_M 0xf000000000000000ull ++#define PAS_STATUS_TIMER 0x1000000000000000ull ++#define PAS_STATUS_ERROR 0x2000000000000000ull ++#define PAS_STATUS_SOFT 0x4000000000000000ull ++#define PAS_STATUS_INT 0x8000000000000000ull ++ ++#define PAS_IOB_DMA_RXCH_CFG(i) (0x1100 + (i)*4) ++#define PAS_IOB_DMA_RXCH_CFG_CNTTH_M 0x00000fff ++#define PAS_IOB_DMA_RXCH_CFG_CNTTH_S 0 ++#define PAS_IOB_DMA_RXCH_CFG_CNTTH(x) (((x) << PAS_IOB_DMA_RXCH_CFG_CNTTH_S) & \ ++ PAS_IOB_DMA_RXCH_CFG_CNTTH_M) ++#define PAS_IOB_DMA_TXCH_CFG(i) (0x1200 + (i)*4) ++#define PAS_IOB_DMA_TXCH_CFG_CNTTH_M 0x00000fff ++#define PAS_IOB_DMA_TXCH_CFG_CNTTH_S 0 ++#define PAS_IOB_DMA_TXCH_CFG_CNTTH(x) (((x) << PAS_IOB_DMA_TXCH_CFG_CNTTH_S) & \ ++ PAS_IOB_DMA_TXCH_CFG_CNTTH_M) ++#define PAS_IOB_DMA_RXCH_STAT(i) (0x1300 + (i)*4) ++#define PAS_IOB_DMA_RXCH_STAT_INTGEN 0x00001000 ++#define PAS_IOB_DMA_RXCH_STAT_CNTDEL_M 0x00000fff ++#define PAS_IOB_DMA_RXCH_STAT_CNTDEL_S 0 ++#define PAS_IOB_DMA_RXCH_STAT_CNTDEL(x) (((x) << PAS_IOB_DMA_RXCH_STAT_CNTDEL_S) &\ ++ PAS_IOB_DMA_RXCH_STAT_CNTDEL_M) ++#define PAS_IOB_DMA_TXCH_STAT(i) (0x1400 + (i)*4) ++#define PAS_IOB_DMA_TXCH_STAT_INTGEN 0x00001000 ++#define PAS_IOB_DMA_TXCH_STAT_CNTDEL_M 0x00000fff ++#define PAS_IOB_DMA_TXCH_STAT_CNTDEL_S 0 ++#define PAS_IOB_DMA_TXCH_STAT_CNTDEL(x) (((x) << PAS_IOB_DMA_TXCH_STAT_CNTDEL_S) &\ ++ PAS_IOB_DMA_TXCH_STAT_CNTDEL_M) ++#define PAS_IOB_DMA_RXCH_RESET(i) (0x1500 + (i)*4) ++#define PAS_IOB_DMA_RXCH_RESET_PCNT_M 0xffff0000 ++#define PAS_IOB_DMA_RXCH_RESET_PCNT_S 16 ++#define PAS_IOB_DMA_RXCH_RESET_PCNT(x) (((x) << PAS_IOB_DMA_RXCH_RESET_PCNT_S) & \ ++ PAS_IOB_DMA_RXCH_RESET_PCNT_M) ++#define PAS_IOB_DMA_RXCH_RESET_PCNTRST 0x00000020 ++#define PAS_IOB_DMA_RXCH_RESET_DCNTRST 0x00000010 ++#define PAS_IOB_DMA_RXCH_RESET_TINTC 0x00000008 ++#define PAS_IOB_DMA_RXCH_RESET_DINTC 0x00000004 ++#define PAS_IOB_DMA_RXCH_RESET_SINTC 0x00000002 ++#define PAS_IOB_DMA_RXCH_RESET_PINTC 0x00000001 ++#define PAS_IOB_DMA_TXCH_RESET(i) (0x1600 + (i)*4) ++#define PAS_IOB_DMA_TXCH_RESET_PCNT_M 0xffff0000 ++#define PAS_IOB_DMA_TXCH_RESET_PCNT_S 16 ++#define PAS_IOB_DMA_TXCH_RESET_PCNT(x) (((x) << PAS_IOB_DMA_TXCH_RESET_PCNT_S) & \ ++ PAS_IOB_DMA_TXCH_RESET_PCNT_M) ++#define PAS_IOB_DMA_TXCH_RESET_PCNTRST 0x00000020 ++#define PAS_IOB_DMA_TXCH_RESET_DCNTRST 0x00000010 ++#define PAS_IOB_DMA_TXCH_RESET_TINTC 0x00000008 ++#define PAS_IOB_DMA_TXCH_RESET_DINTC 0x00000004 ++#define PAS_IOB_DMA_TXCH_RESET_SINTC 0x00000002 ++#define PAS_IOB_DMA_TXCH_RESET_PINTC 0x00000001 ++ ++#define PAS_IOB_DMA_COM_TIMEOUTCFG 0x1700 ++#define PAS_IOB_DMA_COM_TIMEOUTCFG_TCNT_M 0x00ffffff ++#define PAS_IOB_DMA_COM_TIMEOUTCFG_TCNT_S 0 ++#define PAS_IOB_DMA_COM_TIMEOUTCFG_TCNT(x) (((x) << PAS_IOB_DMA_COM_TIMEOUTCFG_TCNT_S) & \ ++ PAS_IOB_DMA_COM_TIMEOUTCFG_TCNT_M) ++ ++/* Transmit descriptor fields */ ++#define XCT_MACTX_T 0x8000000000000000ull ++#define XCT_MACTX_ST 0x4000000000000000ull ++#define XCT_MACTX_NORES 0x0000000000000000ull ++#define XCT_MACTX_8BRES 0x1000000000000000ull ++#define XCT_MACTX_24BRES 0x2000000000000000ull ++#define XCT_MACTX_40BRES 0x3000000000000000ull ++#define XCT_MACTX_I 0x0800000000000000ull ++#define XCT_MACTX_O 0x0400000000000000ull ++#define XCT_MACTX_E 0x0200000000000000ull ++#define XCT_MACTX_VLAN_M 0x0180000000000000ull ++#define XCT_MACTX_VLAN_NOP 0x0000000000000000ull ++#define XCT_MACTX_VLAN_REMOVE 0x0080000000000000ull ++#define XCT_MACTX_VLAN_INSERT 0x0100000000000000ull ++#define XCT_MACTX_VLAN_REPLACE 0x0180000000000000ull ++#define XCT_MACTX_CRC_M 0x0060000000000000ull ++#define XCT_MACTX_CRC_NOP 0x0000000000000000ull ++#define XCT_MACTX_CRC_INSERT 0x0020000000000000ull ++#define XCT_MACTX_CRC_PAD 0x0040000000000000ull ++#define XCT_MACTX_CRC_REPLACE 0x0060000000000000ull ++#define XCT_MACTX_SS 0x0010000000000000ull ++#define XCT_MACTX_LLEN_M 0x00007fff00000000ull ++#define XCT_MACTX_LLEN_S 32ull ++#define XCT_MACTX_LLEN(x) ((((long)(x)) << XCT_MACTX_LLEN_S) & \ ++ XCT_MACTX_LLEN_M) ++#define XCT_MACTX_IPH_M 0x00000000f8000000ull ++#define XCT_MACTX_IPH_S 27ull ++#define XCT_MACTX_IPH(x) ((((long)(x)) << XCT_MACTX_IPH_S) & \ ++ XCT_MACTX_IPH_M) ++#define XCT_MACTX_IPO_M 0x0000000007c00000ull ++#define XCT_MACTX_IPO_S 22ull ++#define XCT_MACTX_IPO(x) ((((long)(x)) << XCT_MACTX_IPO_S) & \ ++ XCT_MACTX_IPO_M) ++#define XCT_MACTX_CSUM_M 0x0000000000000060ull ++#define XCT_MACTX_CSUM_NOP 0x0000000000000000ull ++#define XCT_MACTX_CSUM_TCP 0x0000000000000040ull ++#define XCT_MACTX_CSUM_UDP 0x0000000000000060ull ++#define XCT_MACTX_V6 0x0000000000000010ull ++#define XCT_MACTX_C 0x0000000000000004ull ++#define XCT_MACTX_AL2 0x0000000000000002ull ++ ++#define XCT_PTR_T 0x8000000000000000ull ++#define XCT_PTR_LEN_M 0x7ffff00000000000ull ++#define XCT_PTR_LEN_S 44 ++#define XCT_PTR_LEN(x) ((((long)(x)) << XCT_PTR_LEN_S) & \ ++ XCT_PTR_LEN_M) ++#define XCT_PTR_ADDR_M 0x00000fffffffffffull ++#define XCT_PTR_ADDR_S 0 ++#define XCT_PTR_ADDR(x) ((((long)(x)) << XCT_PTR_ADDR_S) & \ ++ XCT_PTR_ADDR_M) ++ ++/* Function descriptor fields */ ++#define XCT_FUN_T 0x8000000000000000ull ++#define XCT_FUN_ST 0x4000000000000000ull ++#define XCT_FUN_NORES 0x0000000000000000ull ++#define XCT_FUN_8BRES 0x1000000000000000ull ++#define XCT_FUN_24BRES 0x2000000000000000ull ++#define XCT_FUN_40BRES 0x3000000000000000ull ++#define XCT_FUN_I 0x0800000000000000ull ++#define XCT_FUN_O 0x0400000000000000ull ++#define XCT_FUN_E 0x0200000000000000ull ++#define XCT_FUN_FUN_S 54 ++#define XCT_FUN_FUN_M 0x01c0000000000000ull ++#define XCT_FUN_FUN(num) ((((long)(num)) << XCT_FUN_FUN_S) & \ ++ XCT_FUN_FUN_M) ++#define XCT_FUN_CRM_NOP 0x0000000000000000ull ++#define XCT_FUN_CRM_SIG 0x0008000000000000ull ++#define XCT_FUN_CRM_ENC 0x0010000000000000ull ++#define XCT_FUN_CRM_DEC 0x0018000000000000ull ++#define XCT_FUN_CRM_SIG_ENC 0x0020000000000000ull ++#define XCT_FUN_CRM_ENC_SIG 0x0028000000000000ull ++#define XCT_FUN_CRM_SIG_DEC 0x0030000000000000ull ++#define XCT_FUN_CRM_DEC_SIG 0x0038000000000000ull ++#define XCT_FUN_LLEN_M 0x0007ffff00000000ull ++#define XCT_FUN_LLEN_S 32ULL ++#define XCT_FUN_LLEN(x) ((((long)(x)) << XCT_FUN_LLEN_S) & \ ++ XCT_FUN_LLEN_M) ++#define XCT_FUN_SHL_M 0x00000000f8000000ull ++#define XCT_FUN_SHL_S 27ull ++#define XCT_FUN_SHL(x) ((((long)(x)) << XCT_FUN_SHL_S) & \ ++ XCT_FUN_SHL_M) ++#define XCT_FUN_CHL_M 0x0000000007c00000ull ++#define XCT_FUN_CHL_S 22ull ++#define XCT_FUN_CHL(x) ((((long)(x)) << XCT_FUN_CHL_S) & \ ++ XCT_FUN_CHL_M) ++#define XCT_FUN_HSZ_M 0x00000000003c0000ull ++#define XCT_FUN_HSZ_S 18ull ++#define XCT_FUN_HSZ(x) ((((long)(x)) << XCT_FUN_HSZ_S) & \ ++ XCT_FUN_HSZ_M) ++#define XCT_FUN_ALG_DES 0x0000000000000000ull ++#define XCT_FUN_ALG_3DES 0x0000000000008000ull ++#define XCT_FUN_ALG_AES 0x0000000000010000ull ++#define XCT_FUN_ALG_ARC 0x0000000000018000ull ++#define XCT_FUN_ALG_KASUMI 0x0000000000020000ull ++#define XCT_FUN_BCM_ECB 0x0000000000000000ull ++#define XCT_FUN_BCM_CBC 0x0000000000001000ull ++#define XCT_FUN_BCM_CFB 0x0000000000002000ull ++#define XCT_FUN_BCM_OFB 0x0000000000003000ull ++#define XCT_FUN_BCM_CNT 0x0000000000003800ull ++#define XCT_FUN_BCM_KAS_F8 0x0000000000002800ull ++#define XCT_FUN_BCM_KAS_F9 0x0000000000001800ull ++#define XCT_FUN_BCP_NO_PAD 0x0000000000000000ull ++#define XCT_FUN_BCP_ZRO 0x0000000000000200ull ++#define XCT_FUN_BCP_PL 0x0000000000000400ull ++#define XCT_FUN_BCP_INCR 0x0000000000000600ull ++#define XCT_FUN_SIG_MD5 (0ull << 4) ++#define XCT_FUN_SIG_SHA1 (2ull << 4) ++#define XCT_FUN_SIG_HMAC_MD5 (8ull << 4) ++#define XCT_FUN_SIG_HMAC_SHA1 (10ull << 4) ++#define XCT_FUN_A 0x0000000000000008ull ++#define XCT_FUN_C 0x0000000000000004ull ++#define XCT_FUN_AL2 0x0000000000000002ull ++#define XCT_FUN_SE 0x0000000000000001ull ++ ++#define XCT_FUN_SRC_PTR(len, addr) (XCT_PTR_LEN(len) | XCT_PTR_ADDR(addr)) ++#define XCT_FUN_DST_PTR(len, addr) (XCT_FUN_SRC_PTR(len, addr) | \ ++ 0x8000000000000000ull) ++ ++#define XCT_CTRL_HDR_FUN_NUM_M 0x01c0000000000000ull ++#define XCT_CTRL_HDR_FUN_NUM_S 54 ++#define XCT_CTRL_HDR_LEN_M 0x0007ffff00000000ull ++#define XCT_CTRL_HDR_LEN_S 32 ++#define XCT_CTRL_HDR_REG_M 0x00000000000000ffull ++#define XCT_CTRL_HDR_REG_S 0 ++ ++#define XCT_CTRL_HDR(funcN,len,reg) (0x9400000000000000ull | \ ++ ((((long)(funcN)) << XCT_CTRL_HDR_FUN_NUM_S) \ ++ & XCT_CTRL_HDR_FUN_NUM_M) | \ ++ ((((long)(len)) << \ ++ XCT_CTRL_HDR_LEN_S) & XCT_CTRL_HDR_LEN_M) | \ ++ ((((long)(reg)) << \ ++ XCT_CTRL_HDR_REG_S) & XCT_CTRL_HDR_REG_M)) ++ ++/* Function config command options */ ++#define DMA_CALGO_DES 0x00 ++#define DMA_CALGO_3DES 0x01 ++#define DMA_CALGO_AES 0x02 ++#define DMA_CALGO_ARC 0x03 ++ ++#define DMA_FN_CIV0 0x02 ++#define DMA_FN_CIV1 0x03 ++#define DMA_FN_HKEY0 0x0a ++ ++#define XCT_PTR_ADDR_LEN(ptr) ((ptr) & XCT_PTR_ADDR_M), \ ++ (((ptr) & XCT_PTR_LEN_M) >> XCT_PTR_LEN_S) ++ ++#endif /* PASEMI_FNU_H */ +diff -Nur linux-2.6.30.orig/crypto/ocf/random.c linux-2.6.30/crypto/ocf/random.c +--- linux-2.6.30.orig/crypto/ocf/random.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/random.c 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,317 @@ ++/* ++ * A system independant way of adding entropy to the kernels pool ++ * this way the drivers can focus on the real work and we can take ++ * care of pushing it to the appropriate place in the kernel. ++ * ++ * This should be fast and callable from timers/interrupts ++ * ++ * Written by David McCullough <david_mccullough@securecomputing.com> ++ * Copyright (C) 2006-2007 David McCullough ++ * Copyright (C) 2004-2005 Intel Corporation. ++ * ++ * LICENSE TERMS ++ * ++ * The free distribution and use of this software in both source and binary ++ * form is allowed (with or without changes) provided that: ++ * ++ * 1. distributions of this source code include the above copyright ++ * notice, this list of conditions and the following disclaimer; ++ * ++ * 2. distributions in binary form include the above copyright ++ * notice, this list of conditions and the following disclaimer ++ * in the documentation and/or other associated materials; ++ * ++ * 3. the copyright holder's name is not used to endorse products ++ * built using this software without specific written permission. ++ * ++ * ALTERNATIVELY, provided that this notice is retained in full, this product ++ * may be distributed under the terms of the GNU General Public License (GPL), ++ * in which case the provisions of the GPL apply INSTEAD OF those given above. ++ * ++ * DISCLAIMER ++ * ++ * This software is provided 'as is' with no explicit or implied warranties ++ * in respect of its properties, including, but not limited to, correctness ++ * and/or fitness for purpose. ++ */ ++ ++#ifndef AUTOCONF_INCLUDED ++#include <linux/config.h> ++#endif ++#include <linux/module.h> ++#include <linux/init.h> ++#include <linux/list.h> ++#include <linux/slab.h> ++#include <linux/wait.h> ++#include <linux/sched.h> ++#include <linux/spinlock.h> ++#include <linux/version.h> ++#include <linux/unistd.h> ++#include <linux/poll.h> ++#include <linux/random.h> ++#include <cryptodev.h> ++ ++#ifdef CONFIG_OCF_FIPS ++#include "rndtest.h" ++#endif ++ ++#ifndef HAS_RANDOM_INPUT_WAIT ++#error "Please do not enable OCF_RANDOMHARVEST unless you have applied patches" ++#endif ++ ++/* ++ * a hack to access the debug levels from the crypto driver ++ */ ++extern int crypto_debug; ++#define debug crypto_debug ++ ++/* ++ * a list of all registered random providers ++ */ ++static LIST_HEAD(random_ops); ++static int started = 0; ++static int initted = 0; ++ ++struct random_op { ++ struct list_head random_list; ++ u_int32_t driverid; ++ int (*read_random)(void *arg, u_int32_t *buf, int len); ++ void *arg; ++}; ++ ++static int random_proc(void *arg); ++ ++static pid_t randomproc = (pid_t) -1; ++static spinlock_t random_lock; ++ ++/* ++ * just init the spin locks ++ */ ++static int ++crypto_random_init(void) ++{ ++ spin_lock_init(&random_lock); ++ initted = 1; ++ return(0); ++} ++ ++/* ++ * Add the given random reader to our list (if not present) ++ * and start the thread (if not already started) ++ * ++ * we have to assume that driver id is ok for now ++ */ ++int ++crypto_rregister( ++ u_int32_t driverid, ++ int (*read_random)(void *arg, u_int32_t *buf, int len), ++ void *arg) ++{ ++ unsigned long flags; ++ int ret = 0; ++ struct random_op *rops, *tmp; ++ ++ dprintk("%s,%d: %s(0x%x, %p, %p)\n", __FILE__, __LINE__, ++ __FUNCTION__, driverid, read_random, arg); ++ ++ if (!initted) ++ crypto_random_init(); ++ ++#if 0 ++ struct cryptocap *cap; ++ ++ cap = crypto_checkdriver(driverid); ++ if (!cap) ++ return EINVAL; ++#endif ++ ++ list_for_each_entry_safe(rops, tmp, &random_ops, random_list) { ++ if (rops->driverid == driverid && rops->read_random == read_random) ++ return EEXIST; ++ } ++ ++ rops = (struct random_op *) kmalloc(sizeof(*rops), GFP_KERNEL); ++ if (!rops) ++ return ENOMEM; ++ ++ rops->driverid = driverid; ++ rops->read_random = read_random; ++ rops->arg = arg; ++ ++ spin_lock_irqsave(&random_lock, flags); ++ list_add_tail(&rops->random_list, &random_ops); ++ if (!started) { ++ randomproc = kernel_thread(random_proc, NULL, CLONE_FS|CLONE_FILES); ++ if (randomproc < 0) { ++ ret = randomproc; ++ printk("crypto: crypto_rregister cannot start random thread; " ++ "error %d", ret); ++ } else ++ started = 1; ++ } ++ spin_unlock_irqrestore(&random_lock, flags); ++ ++ return ret; ++} ++EXPORT_SYMBOL(crypto_rregister); ++ ++int ++crypto_runregister_all(u_int32_t driverid) ++{ ++ struct random_op *rops, *tmp; ++ unsigned long flags; ++ ++ dprintk("%s,%d: %s(0x%x)\n", __FILE__, __LINE__, __FUNCTION__, driverid); ++ ++ list_for_each_entry_safe(rops, tmp, &random_ops, random_list) { ++ if (rops->driverid == driverid) { ++ list_del(&rops->random_list); ++ kfree(rops); ++ } ++ } ++ ++ spin_lock_irqsave(&random_lock, flags); ++ if (list_empty(&random_ops) && started) ++ kill_pid(randomproc, SIGKILL, 1); ++ spin_unlock_irqrestore(&random_lock, flags); ++ return(0); ++} ++EXPORT_SYMBOL(crypto_runregister_all); ++ ++/* ++ * while we can add entropy to random.c continue to read random data from ++ * the drivers and push it to random. ++ */ ++static int ++random_proc(void *arg) ++{ ++ int n; ++ int wantcnt; ++ int bufcnt = 0; ++ int retval = 0; ++ int *buf = NULL; ++ ++#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,0) ++ daemonize(); ++ spin_lock_irq(¤t->sigmask_lock); ++ sigemptyset(¤t->blocked); ++ recalc_sigpending(current); ++ spin_unlock_irq(¤t->sigmask_lock); ++ sprintf(current->comm, "ocf-random"); ++#else ++ daemonize("ocf-random"); ++ allow_signal(SIGKILL); ++#endif ++ ++ (void) get_fs(); ++ set_fs(get_ds()); ++ ++#ifdef CONFIG_OCF_FIPS ++#define NUM_INT (RNDTEST_NBYTES/sizeof(int)) ++#else ++#define NUM_INT 32 ++#endif ++ ++ /* ++ * some devices can transferr their RNG data direct into memory, ++ * so make sure it is device friendly ++ */ ++ buf = kmalloc(NUM_INT * sizeof(int), GFP_DMA); ++ if (NULL == buf) { ++ printk("crypto: RNG could not allocate memory\n"); ++ retval = -ENOMEM; ++ goto bad_alloc; ++ } ++ ++ wantcnt = NUM_INT; /* start by adding some entropy */ ++ ++ /* ++ * its possible due to errors or driver removal that we no longer ++ * have anything to do, if so exit or we will consume all the CPU ++ * doing nothing ++ */ ++ while (!list_empty(&random_ops)) { ++ struct random_op *rops, *tmp; ++ ++#ifdef CONFIG_OCF_FIPS ++ if (wantcnt) ++ wantcnt = NUM_INT; /* FIPs mode can do 20000 bits or none */ ++#endif ++ ++ /* see if we can get enough entropy to make the world ++ * a better place. ++ */ ++ while (bufcnt < wantcnt && bufcnt < NUM_INT) { ++ list_for_each_entry_safe(rops, tmp, &random_ops, random_list) { ++ ++ n = (*rops->read_random)(rops->arg, &buf[bufcnt], ++ NUM_INT - bufcnt); ++ ++ /* on failure remove the random number generator */ ++ if (n == -1) { ++ list_del(&rops->random_list); ++ printk("crypto: RNG (driverid=0x%x) failed, disabling\n", ++ rops->driverid); ++ kfree(rops); ++ } else if (n > 0) ++ bufcnt += n; ++ } ++ /* give up CPU for a bit, just in case as this is a loop */ ++ schedule(); ++ } ++ ++ ++#ifdef CONFIG_OCF_FIPS ++ if (bufcnt > 0 && rndtest_buf((unsigned char *) &buf[0])) { ++ dprintk("crypto: buffer had fips errors, discarding\n"); ++ bufcnt = 0; ++ } ++#endif ++ ++ /* ++ * if we have a certified buffer, we can send some data ++ * to /dev/random and move along ++ */ ++ if (bufcnt > 0) { ++ /* add what we have */ ++ random_input_words(buf, bufcnt, bufcnt*sizeof(int)*8); ++ bufcnt = 0; ++ } ++ ++ /* give up CPU for a bit so we don't hog while filling */ ++ schedule(); ++ ++ /* wait for needing more */ ++ wantcnt = random_input_wait(); ++ ++ if (wantcnt <= 0) ++ wantcnt = 0; /* try to get some info again */ ++ else ++ /* round up to one word or we can loop forever */ ++ wantcnt = (wantcnt + (sizeof(int)*8)) / (sizeof(int)*8); ++ if (wantcnt > NUM_INT) { ++ wantcnt = NUM_INT; ++ } ++ ++ if (signal_pending(current)) { ++#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,0) ++ spin_lock_irq(¤t->sigmask_lock); ++#endif ++ flush_signals(current); ++#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,0) ++ spin_unlock_irq(¤t->sigmask_lock); ++#endif ++ } ++ } ++ ++ kfree(buf); ++ ++bad_alloc: ++ spin_lock_irq(&random_lock); ++ randomproc = (pid_t) -1; ++ started = 0; ++ spin_unlock_irq(&random_lock); ++ ++ return retval; ++} ++ +diff -Nur linux-2.6.30.orig/crypto/ocf/README linux-2.6.30/crypto/ocf/README +--- linux-2.6.30.orig/crypto/ocf/README 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/README 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,167 @@ ++README - ocf-linux-20071215 ++--------------------------- ++ ++This README provides instructions for getting ocf-linux compiled and ++operating in a generic linux environment. For other information you ++might like to visit the home page for this project: ++ ++ http://ocf-linux.sourceforge.net/ ++ ++Adding OCF to linux ++------------------- ++ ++ Not much in this file for now, just some notes. I usually build ++ the ocf support as modules but it can be built into the kernel as ++ well. To use it: ++ ++ * mknod /dev/crypto c 10 70 ++ ++ * to add OCF to your kernel source, you have two options. Apply ++ the kernel specific patch: ++ ++ cd linux-2.4*; gunzip < ocf-linux-24-XXXXXXXX.patch.gz | patch -p1 ++ cd linux-2.6*; gunzip < ocf-linux-26-XXXXXXXX.patch.gz | patch -p1 ++ ++ if you do one of the above, then you can proceed to the next step, ++ or you can do the above process by hand with using the patches against ++ linux-2.4.35 and 2.6.23 to include the ocf code under crypto/ocf. ++ Here's how to add it: ++ ++ for 2.4.35 (and later) ++ ++ cd linux-2.4.35/crypto ++ tar xvzf ocf-linux.tar.gz ++ cd .. ++ patch -p1 < crypto/ocf/patches/linux-2.4.35-ocf.patch ++ ++ for 2.6.23 (and later), find the kernel patch specific (or nearest) ++ to your kernel versions and then: ++ ++ cd linux-2.6.NN/crypto ++ tar xvzf ocf-linux.tar.gz ++ cd .. ++ patch -p1 < crypto/ocf/patches/linux-2.6.NN-ocf.patch ++ ++ It should be easy to take this patch and apply it to other more ++ recent versions of the kernels. The same patches should also work ++ relatively easily on kernels as old as 2.6.11 and 2.4.18. ++ ++ * under 2.4 if you are on a non-x86 platform, you may need to: ++ ++ cp linux-2.X.x/include/asm-i386/kmap_types.h linux-2.X.x/include/asm-YYY ++ ++ so that you can build the kernel crypto support needed for the cryptosoft ++ driver. ++ ++ * For simplicity you should enable all the crypto support in your kernel ++ except for the test driver. Likewise for the OCF options. Do not ++ enable OCF crypto drivers for HW that you do not have (for example ++ ixp4xx will not compile on non-Xscale systems). ++ ++ * make sure that cryptodev.h (from ocf-linux.tar.gz) is installed as ++ crypto/cryptodev.h in an include directory that is used for building ++ applications for your platform. For example on a host system that ++ might be: ++ ++ /usr/include/crypto/cryptodev.h ++ ++ * patch your openssl-0.9.8i code with the openssl-0.9.8i.patch. ++ (NOTE: there is no longer a need to patch ssh). The patch is against: ++ openssl-0_9_8e ++ ++ If you need a patch for an older version of openssl, you should look ++ to older OCF releases. This patch is unlikely to work on older ++ openssl versions. ++ ++ openssl-0.9.8i.patch ++ - enables --with-cryptodev for non BSD systems ++ - adds -cpu option to openssl speed for calculating CPU load ++ under linux ++ - fixes null pointer in openssl speed multi thread output. ++ - fixes test keys to work with linux crypto's more stringent ++ key checking. ++ - adds MD5/SHA acceleration (Ronen Shitrit), only enabled ++ with the --with-cryptodev-digests option ++ - fixes bug in engine code caching. ++ ++ * build crypto-tools-XXXXXXXX.tar.gz if you want to try some of the BSD ++ tools for testing OCF (ie., cryptotest). ++ ++How to load the OCF drivers ++--------------------------- ++ ++ First insert the base modules: ++ ++ insmod ocf ++ insmod cryptodev ++ ++ You can then install the software OCF driver with: ++ ++ insmod cryptosoft ++ ++ and one or more of the OCF HW drivers with: ++ ++ insmod safe ++ insmod hifn7751 ++ insmod ixp4xx ++ ... ++ ++ all the drivers take a debug option to enable verbose debug so that ++ you can see what is going on. For debug you load them as: ++ ++ insmod ocf crypto_debug=1 ++ insmod cryptodev cryptodev_debug=1 ++ insmod cryptosoft swcr_debug=1 ++ ++ You may load more than one OCF crypto driver but then there is no guarantee ++ as to which will be used. ++ ++ You can also enable debug at run time on 2.6 systems with the following: ++ ++ echo 1 > /sys/module/ocf/parameters/crypto_debug ++ echo 1 > /sys/module/cryptodev/parameters/cryptodev_debug ++ echo 1 > /sys/module/cryptosoft/parameters/swcr_debug ++ echo 1 > /sys/module/hifn7751/parameters/hifn_debug ++ echo 1 > /sys/module/safe/parameters/safe_debug ++ echo 1 > /sys/module/ixp4xx/parameters/ixp_debug ++ ... ++ ++Testing the OCF support ++----------------------- ++ ++ run "cryptotest", it should do a short test for a couple of ++ des packets. If it does everything is working. ++ ++ If this works, then ssh will use the driver when invoked as: ++ ++ ssh -c 3des username@host ++ ++ to see for sure that it is operating, enable debug as defined above. ++ ++ To get a better idea of performance run: ++ ++ cryptotest 100 4096 ++ ++ There are more options to cryptotest, see the help. ++ ++ It is also possible to use openssl to test the speed of the crypto ++ drivers. ++ ++ openssl speed -evp des -engine cryptodev -elapsed ++ openssl speed -evp des3 -engine cryptodev -elapsed ++ openssl speed -evp aes128 -engine cryptodev -elapsed ++ ++ and multiple threads (10) with: ++ ++ openssl speed -evp des -engine cryptodev -elapsed -multi 10 ++ openssl speed -evp des3 -engine cryptodev -elapsed -multi 10 ++ openssl speed -evp aes128 -engine cryptodev -elapsed -multi 10 ++ ++ for public key testing you can try: ++ ++ cryptokeytest ++ openssl speed -engine cryptodev rsa -elapsed ++ openssl speed -engine cryptodev dsa -elapsed ++ ++David McCullough ++david_mccullough@securecomputing.com +diff -Nur linux-2.6.30.orig/crypto/ocf/rndtest.c linux-2.6.30/crypto/ocf/rndtest.c +--- linux-2.6.30.orig/crypto/ocf/rndtest.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/rndtest.c 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,300 @@ ++/* $OpenBSD$ */ ++ ++/* ++ * OCF/Linux port done by David McCullough <david_mccullough@securecomputing.com> ++ * Copyright (C) 2006-2007 David McCullough ++ * Copyright (C) 2004-2005 Intel Corporation. ++ * The license and original author are listed below. ++ * ++ * Copyright (c) 2002 Jason L. Wright (jason@thought.net) ++ * All rights reserved. ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions ++ * are met: ++ * 1. Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * 2. Redistributions in binary form must reproduce the above copyright ++ * notice, this list of conditions and the following disclaimer in the ++ * documentation and/or other materials provided with the distribution. ++ * 3. All advertising materials mentioning features or use of this software ++ * must display the following acknowledgement: ++ * This product includes software developed by Jason L. Wright ++ * 4. The name of the author may not be used to endorse or promote products ++ * derived from this software without specific prior written permission. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR ++ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED ++ * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE ++ * DISCLAIMED. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, ++ * INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES ++ * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR ++ * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) ++ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, ++ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ++ * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE ++ * POSSIBILITY OF SUCH DAMAGE. ++ */ ++ ++#ifndef AUTOCONF_INCLUDED ++#include <linux/config.h> ++#endif ++#include <linux/module.h> ++#include <linux/list.h> ++#include <linux/wait.h> ++#include <linux/time.h> ++#include <linux/version.h> ++#include <linux/unistd.h> ++#include <linux/kernel.h> ++#include <linux/string.h> ++#include <linux/time.h> ++#include <cryptodev.h> ++#include "rndtest.h" ++ ++static struct rndtest_stats rndstats; ++ ++static void rndtest_test(struct rndtest_state *); ++ ++/* The tests themselves */ ++static int rndtest_monobit(struct rndtest_state *); ++static int rndtest_runs(struct rndtest_state *); ++static int rndtest_longruns(struct rndtest_state *); ++static int rndtest_chi_4(struct rndtest_state *); ++ ++static int rndtest_runs_check(struct rndtest_state *, int, int *); ++static void rndtest_runs_record(struct rndtest_state *, int, int *); ++ ++static const struct rndtest_testfunc { ++ int (*test)(struct rndtest_state *); ++} rndtest_funcs[] = { ++ { rndtest_monobit }, ++ { rndtest_runs }, ++ { rndtest_chi_4 }, ++ { rndtest_longruns }, ++}; ++ ++#define RNDTEST_NTESTS (sizeof(rndtest_funcs)/sizeof(rndtest_funcs[0])) ++ ++static void ++rndtest_test(struct rndtest_state *rsp) ++{ ++ int i, rv = 0; ++ ++ rndstats.rst_tests++; ++ for (i = 0; i < RNDTEST_NTESTS; i++) ++ rv |= (*rndtest_funcs[i].test)(rsp); ++ rsp->rs_discard = (rv != 0); ++} ++ ++ ++extern int crypto_debug; ++#define rndtest_verbose 2 ++#define rndtest_report(rsp, failure, fmt, a...) \ ++ { if (failure || crypto_debug) { printk("rng_test: " fmt "\n", a); } else; } ++ ++#define RNDTEST_MONOBIT_MINONES 9725 ++#define RNDTEST_MONOBIT_MAXONES 10275 ++ ++static int ++rndtest_monobit(struct rndtest_state *rsp) ++{ ++ int i, ones = 0, j; ++ u_int8_t r; ++ ++ for (i = 0; i < RNDTEST_NBYTES; i++) { ++ r = rsp->rs_buf[i]; ++ for (j = 0; j < 8; j++, r <<= 1) ++ if (r & 0x80) ++ ones++; ++ } ++ if (ones > RNDTEST_MONOBIT_MINONES && ++ ones < RNDTEST_MONOBIT_MAXONES) { ++ if (rndtest_verbose > 1) ++ rndtest_report(rsp, 0, "monobit pass (%d < %d < %d)", ++ RNDTEST_MONOBIT_MINONES, ones, ++ RNDTEST_MONOBIT_MAXONES); ++ return (0); ++ } else { ++ if (rndtest_verbose) ++ rndtest_report(rsp, 1, ++ "monobit failed (%d ones)", ones); ++ rndstats.rst_monobit++; ++ return (-1); ++ } ++} ++ ++#define RNDTEST_RUNS_NINTERVAL 6 ++ ++static const struct rndtest_runs_tabs { ++ u_int16_t min, max; ++} rndtest_runs_tab[] = { ++ { 2343, 2657 }, ++ { 1135, 1365 }, ++ { 542, 708 }, ++ { 251, 373 }, ++ { 111, 201 }, ++ { 111, 201 }, ++}; ++ ++static int ++rndtest_runs(struct rndtest_state *rsp) ++{ ++ int i, j, ones, zeros, rv = 0; ++ int onei[RNDTEST_RUNS_NINTERVAL], zeroi[RNDTEST_RUNS_NINTERVAL]; ++ u_int8_t c; ++ ++ bzero(onei, sizeof(onei)); ++ bzero(zeroi, sizeof(zeroi)); ++ ones = zeros = 0; ++ for (i = 0; i < RNDTEST_NBYTES; i++) { ++ c = rsp->rs_buf[i]; ++ for (j = 0; j < 8; j++, c <<= 1) { ++ if (c & 0x80) { ++ ones++; ++ rndtest_runs_record(rsp, zeros, zeroi); ++ zeros = 0; ++ } else { ++ zeros++; ++ rndtest_runs_record(rsp, ones, onei); ++ ones = 0; ++ } ++ } ++ } ++ rndtest_runs_record(rsp, ones, onei); ++ rndtest_runs_record(rsp, zeros, zeroi); ++ ++ rv |= rndtest_runs_check(rsp, 0, zeroi); ++ rv |= rndtest_runs_check(rsp, 1, onei); ++ ++ if (rv) ++ rndstats.rst_runs++; ++ ++ return (rv); ++} ++ ++static void ++rndtest_runs_record(struct rndtest_state *rsp, int len, int *intrv) ++{ ++ if (len == 0) ++ return; ++ if (len > RNDTEST_RUNS_NINTERVAL) ++ len = RNDTEST_RUNS_NINTERVAL; ++ len -= 1; ++ intrv[len]++; ++} ++ ++static int ++rndtest_runs_check(struct rndtest_state *rsp, int val, int *src) ++{ ++ int i, rv = 0; ++ ++ for (i = 0; i < RNDTEST_RUNS_NINTERVAL; i++) { ++ if (src[i] < rndtest_runs_tab[i].min || ++ src[i] > rndtest_runs_tab[i].max) { ++ rndtest_report(rsp, 1, ++ "%s interval %d failed (%d, %d-%d)", ++ val ? "ones" : "zeros", ++ i + 1, src[i], rndtest_runs_tab[i].min, ++ rndtest_runs_tab[i].max); ++ rv = -1; ++ } else { ++ rndtest_report(rsp, 0, ++ "runs pass %s interval %d (%d < %d < %d)", ++ val ? "ones" : "zeros", ++ i + 1, rndtest_runs_tab[i].min, src[i], ++ rndtest_runs_tab[i].max); ++ } ++ } ++ return (rv); ++} ++ ++static int ++rndtest_longruns(struct rndtest_state *rsp) ++{ ++ int i, j, ones = 0, zeros = 0, maxones = 0, maxzeros = 0; ++ u_int8_t c; ++ ++ for (i = 0; i < RNDTEST_NBYTES; i++) { ++ c = rsp->rs_buf[i]; ++ for (j = 0; j < 8; j++, c <<= 1) { ++ if (c & 0x80) { ++ zeros = 0; ++ ones++; ++ if (ones > maxones) ++ maxones = ones; ++ } else { ++ ones = 0; ++ zeros++; ++ if (zeros > maxzeros) ++ maxzeros = zeros; ++ } ++ } ++ } ++ ++ if (maxones < 26 && maxzeros < 26) { ++ rndtest_report(rsp, 0, "longruns pass (%d ones, %d zeros)", ++ maxones, maxzeros); ++ return (0); ++ } else { ++ rndtest_report(rsp, 1, "longruns fail (%d ones, %d zeros)", ++ maxones, maxzeros); ++ rndstats.rst_longruns++; ++ return (-1); ++ } ++} ++ ++/* ++ * chi^2 test over 4 bits: (this is called the poker test in FIPS 140-2, ++ * but it is really the chi^2 test over 4 bits (the poker test as described ++ * by Knuth vol 2 is something different, and I take him as authoritative ++ * on nomenclature over NIST). ++ */ ++#define RNDTEST_CHI4_K 16 ++#define RNDTEST_CHI4_K_MASK (RNDTEST_CHI4_K - 1) ++ ++/* ++ * The unnormalized values are used so that we don't have to worry about ++ * fractional precision. The "real" value is found by: ++ * (V - 1562500) * (16 / 5000) = Vn (where V is the unnormalized value) ++ */ ++#define RNDTEST_CHI4_VMIN 1563181 /* 2.1792 */ ++#define RNDTEST_CHI4_VMAX 1576929 /* 46.1728 */ ++ ++static int ++rndtest_chi_4(struct rndtest_state *rsp) ++{ ++ unsigned int freq[RNDTEST_CHI4_K], i, sum; ++ ++ for (i = 0; i < RNDTEST_CHI4_K; i++) ++ freq[i] = 0; ++ ++ /* Get number of occurances of each 4 bit pattern */ ++ for (i = 0; i < RNDTEST_NBYTES; i++) { ++ freq[(rsp->rs_buf[i] >> 4) & RNDTEST_CHI4_K_MASK]++; ++ freq[(rsp->rs_buf[i] >> 0) & RNDTEST_CHI4_K_MASK]++; ++ } ++ ++ for (i = 0, sum = 0; i < RNDTEST_CHI4_K; i++) ++ sum += freq[i] * freq[i]; ++ ++ if (sum >= 1563181 && sum <= 1576929) { ++ rndtest_report(rsp, 0, "chi^2(4): pass (sum %u)", sum); ++ return (0); ++ } else { ++ rndtest_report(rsp, 1, "chi^2(4): failed (sum %u)", sum); ++ rndstats.rst_chi++; ++ return (-1); ++ } ++} ++ ++int ++rndtest_buf(unsigned char *buf) ++{ ++ struct rndtest_state rsp; ++ ++ memset(&rsp, 0, sizeof(rsp)); ++ rsp.rs_buf = buf; ++ rndtest_test(&rsp); ++ return(rsp.rs_discard); ++} ++ +diff -Nur linux-2.6.30.orig/crypto/ocf/rndtest.h linux-2.6.30/crypto/ocf/rndtest.h +--- linux-2.6.30.orig/crypto/ocf/rndtest.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/rndtest.h 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,54 @@ ++/* $FreeBSD: src/sys/dev/rndtest/rndtest.h,v 1.1 2003/03/11 22:54:44 sam Exp $ */ ++/* $OpenBSD$ */ ++ ++/* ++ * Copyright (c) 2002 Jason L. Wright (jason@thought.net) ++ * All rights reserved. ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions ++ * are met: ++ * 1. Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * 2. Redistributions in binary form must reproduce the above copyright ++ * notice, this list of conditions and the following disclaimer in the ++ * documentation and/or other materials provided with the distribution. ++ * 3. All advertising materials mentioning features or use of this software ++ * must display the following acknowledgement: ++ * This product includes software developed by Jason L. Wright ++ * 4. The name of the author may not be used to endorse or promote products ++ * derived from this software without specific prior written permission. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR ++ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED ++ * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE ++ * DISCLAIMED. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, ++ * INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES ++ * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR ++ * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) ++ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, ++ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ++ * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE ++ * POSSIBILITY OF SUCH DAMAGE. ++ */ ++ ++ ++/* Some of the tests depend on these values */ ++#define RNDTEST_NBYTES 2500 ++#define RNDTEST_NBITS (8 * RNDTEST_NBYTES) ++ ++struct rndtest_state { ++ int rs_discard; /* discard/accept random data */ ++ u_int8_t *rs_buf; ++}; ++ ++struct rndtest_stats { ++ u_int32_t rst_discard; /* number of bytes discarded */ ++ u_int32_t rst_tests; /* number of test runs */ ++ u_int32_t rst_monobit; /* monobit test failures */ ++ u_int32_t rst_runs; /* 0/1 runs failures */ ++ u_int32_t rst_longruns; /* longruns failures */ ++ u_int32_t rst_chi; /* chi^2 failures */ ++}; ++ ++extern int rndtest_buf(unsigned char *buf); +diff -Nur linux-2.6.30.orig/crypto/ocf/safe/Makefile linux-2.6.30/crypto/ocf/safe/Makefile +--- linux-2.6.30.orig/crypto/ocf/safe/Makefile 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/safe/Makefile 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,12 @@ ++# for SGlinux builds ++-include $(ROOTDIR)/modules/.config ++ ++obj-$(CONFIG_OCF_SAFE) += safe.o ++ ++obj ?= . ++EXTRA_CFLAGS += -I$(obj)/.. -I$(obj)/ ++ ++ifdef TOPDIR ++-include $(TOPDIR)/Rules.make ++endif ++ +diff -Nur linux-2.6.30.orig/crypto/ocf/safe/md5.c linux-2.6.30/crypto/ocf/safe/md5.c +--- linux-2.6.30.orig/crypto/ocf/safe/md5.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/safe/md5.c 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,308 @@ ++/* $KAME: md5.c,v 1.5 2000/11/08 06:13:08 itojun Exp $ */ ++/* ++ * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. ++ * All rights reserved. ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions ++ * are met: ++ * 1. Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * 2. Redistributions in binary form must reproduce the above copyright ++ * notice, this list of conditions and the following disclaimer in the ++ * documentation and/or other materials provided with the distribution. ++ * 3. Neither the name of the project nor the names of its contributors ++ * may be used to endorse or promote products derived from this software ++ * without specific prior written permission. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND ++ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE ++ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ++ * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE ++ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL ++ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS ++ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) ++ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT ++ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY ++ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF ++ * SUCH DAMAGE. ++ */ ++ ++#if 0 ++#include <sys/cdefs.h> ++__FBSDID("$FreeBSD: src/sys/crypto/md5.c,v 1.9 2004/01/27 19:49:19 des Exp $"); ++ ++#include <sys/types.h> ++#include <sys/cdefs.h> ++#include <sys/time.h> ++#include <sys/systm.h> ++#include <crypto/md5.h> ++#endif ++ ++#define SHIFT(X, s) (((X) << (s)) | ((X) >> (32 - (s)))) ++ ++#define F(X, Y, Z) (((X) & (Y)) | ((~X) & (Z))) ++#define G(X, Y, Z) (((X) & (Z)) | ((Y) & (~Z))) ++#define H(X, Y, Z) ((X) ^ (Y) ^ (Z)) ++#define I(X, Y, Z) ((Y) ^ ((X) | (~Z))) ++ ++#define ROUND1(a, b, c, d, k, s, i) { \ ++ (a) = (a) + F((b), (c), (d)) + X[(k)] + T[(i)]; \ ++ (a) = SHIFT((a), (s)); \ ++ (a) = (b) + (a); \ ++} ++ ++#define ROUND2(a, b, c, d, k, s, i) { \ ++ (a) = (a) + G((b), (c), (d)) + X[(k)] + T[(i)]; \ ++ (a) = SHIFT((a), (s)); \ ++ (a) = (b) + (a); \ ++} ++ ++#define ROUND3(a, b, c, d, k, s, i) { \ ++ (a) = (a) + H((b), (c), (d)) + X[(k)] + T[(i)]; \ ++ (a) = SHIFT((a), (s)); \ ++ (a) = (b) + (a); \ ++} ++ ++#define ROUND4(a, b, c, d, k, s, i) { \ ++ (a) = (a) + I((b), (c), (d)) + X[(k)] + T[(i)]; \ ++ (a) = SHIFT((a), (s)); \ ++ (a) = (b) + (a); \ ++} ++ ++#define Sa 7 ++#define Sb 12 ++#define Sc 17 ++#define Sd 22 ++ ++#define Se 5 ++#define Sf 9 ++#define Sg 14 ++#define Sh 20 ++ ++#define Si 4 ++#define Sj 11 ++#define Sk 16 ++#define Sl 23 ++ ++#define Sm 6 ++#define Sn 10 ++#define So 15 ++#define Sp 21 ++ ++#define MD5_A0 0x67452301 ++#define MD5_B0 0xefcdab89 ++#define MD5_C0 0x98badcfe ++#define MD5_D0 0x10325476 ++ ++/* Integer part of 4294967296 times abs(sin(i)), where i is in radians. */ ++static const u_int32_t T[65] = { ++ 0, ++ 0xd76aa478, 0xe8c7b756, 0x242070db, 0xc1bdceee, ++ 0xf57c0faf, 0x4787c62a, 0xa8304613, 0xfd469501, ++ 0x698098d8, 0x8b44f7af, 0xffff5bb1, 0x895cd7be, ++ 0x6b901122, 0xfd987193, 0xa679438e, 0x49b40821, ++ ++ 0xf61e2562, 0xc040b340, 0x265e5a51, 0xe9b6c7aa, ++ 0xd62f105d, 0x2441453, 0xd8a1e681, 0xe7d3fbc8, ++ 0x21e1cde6, 0xc33707d6, 0xf4d50d87, 0x455a14ed, ++ 0xa9e3e905, 0xfcefa3f8, 0x676f02d9, 0x8d2a4c8a, ++ ++ 0xfffa3942, 0x8771f681, 0x6d9d6122, 0xfde5380c, ++ 0xa4beea44, 0x4bdecfa9, 0xf6bb4b60, 0xbebfbc70, ++ 0x289b7ec6, 0xeaa127fa, 0xd4ef3085, 0x4881d05, ++ 0xd9d4d039, 0xe6db99e5, 0x1fa27cf8, 0xc4ac5665, ++ ++ 0xf4292244, 0x432aff97, 0xab9423a7, 0xfc93a039, ++ 0x655b59c3, 0x8f0ccc92, 0xffeff47d, 0x85845dd1, ++ 0x6fa87e4f, 0xfe2ce6e0, 0xa3014314, 0x4e0811a1, ++ 0xf7537e82, 0xbd3af235, 0x2ad7d2bb, 0xeb86d391, ++}; ++ ++static const u_int8_t md5_paddat[MD5_BUFLEN] = { ++ 0x80, 0, 0, 0, 0, 0, 0, 0, ++ 0, 0, 0, 0, 0, 0, 0, 0, ++ 0, 0, 0, 0, 0, 0, 0, 0, ++ 0, 0, 0, 0, 0, 0, 0, 0, ++ 0, 0, 0, 0, 0, 0, 0, 0, ++ 0, 0, 0, 0, 0, 0, 0, 0, ++ 0, 0, 0, 0, 0, 0, 0, 0, ++ 0, 0, 0, 0, 0, 0, 0, 0, ++}; ++ ++static void md5_calc(u_int8_t *, md5_ctxt *); ++ ++void md5_init(ctxt) ++ md5_ctxt *ctxt; ++{ ++ ctxt->md5_n = 0; ++ ctxt->md5_i = 0; ++ ctxt->md5_sta = MD5_A0; ++ ctxt->md5_stb = MD5_B0; ++ ctxt->md5_stc = MD5_C0; ++ ctxt->md5_std = MD5_D0; ++ bzero(ctxt->md5_buf, sizeof(ctxt->md5_buf)); ++} ++ ++void md5_loop(ctxt, input, len) ++ md5_ctxt *ctxt; ++ u_int8_t *input; ++ u_int len; /* number of bytes */ ++{ ++ u_int gap, i; ++ ++ ctxt->md5_n += len * 8; /* byte to bit */ ++ gap = MD5_BUFLEN - ctxt->md5_i; ++ ++ if (len >= gap) { ++ bcopy((void *)input, (void *)(ctxt->md5_buf + ctxt->md5_i), ++ gap); ++ md5_calc(ctxt->md5_buf, ctxt); ++ ++ for (i = gap; i + MD5_BUFLEN <= len; i += MD5_BUFLEN) { ++ md5_calc((u_int8_t *)(input + i), ctxt); ++ } ++ ++ ctxt->md5_i = len - i; ++ bcopy((void *)(input + i), (void *)ctxt->md5_buf, ctxt->md5_i); ++ } else { ++ bcopy((void *)input, (void *)(ctxt->md5_buf + ctxt->md5_i), ++ len); ++ ctxt->md5_i += len; ++ } ++} ++ ++void md5_pad(ctxt) ++ md5_ctxt *ctxt; ++{ ++ u_int gap; ++ ++ /* Don't count up padding. Keep md5_n. */ ++ gap = MD5_BUFLEN - ctxt->md5_i; ++ if (gap > 8) { ++ bcopy(md5_paddat, ++ (void *)(ctxt->md5_buf + ctxt->md5_i), ++ gap - sizeof(ctxt->md5_n)); ++ } else { ++ /* including gap == 8 */ ++ bcopy(md5_paddat, (void *)(ctxt->md5_buf + ctxt->md5_i), ++ gap); ++ md5_calc(ctxt->md5_buf, ctxt); ++ bcopy((md5_paddat + gap), ++ (void *)ctxt->md5_buf, ++ MD5_BUFLEN - sizeof(ctxt->md5_n)); ++ } ++ ++ /* 8 byte word */ ++#if BYTE_ORDER == LITTLE_ENDIAN ++ bcopy(&ctxt->md5_n8[0], &ctxt->md5_buf[56], 8); ++#endif ++#if BYTE_ORDER == BIG_ENDIAN ++ ctxt->md5_buf[56] = ctxt->md5_n8[7]; ++ ctxt->md5_buf[57] = ctxt->md5_n8[6]; ++ ctxt->md5_buf[58] = ctxt->md5_n8[5]; ++ ctxt->md5_buf[59] = ctxt->md5_n8[4]; ++ ctxt->md5_buf[60] = ctxt->md5_n8[3]; ++ ctxt->md5_buf[61] = ctxt->md5_n8[2]; ++ ctxt->md5_buf[62] = ctxt->md5_n8[1]; ++ ctxt->md5_buf[63] = ctxt->md5_n8[0]; ++#endif ++ ++ md5_calc(ctxt->md5_buf, ctxt); ++} ++ ++void md5_result(digest, ctxt) ++ u_int8_t *digest; ++ md5_ctxt *ctxt; ++{ ++ /* 4 byte words */ ++#if BYTE_ORDER == LITTLE_ENDIAN ++ bcopy(&ctxt->md5_st8[0], digest, 16); ++#endif ++#if BYTE_ORDER == BIG_ENDIAN ++ digest[ 0] = ctxt->md5_st8[ 3]; digest[ 1] = ctxt->md5_st8[ 2]; ++ digest[ 2] = ctxt->md5_st8[ 1]; digest[ 3] = ctxt->md5_st8[ 0]; ++ digest[ 4] = ctxt->md5_st8[ 7]; digest[ 5] = ctxt->md5_st8[ 6]; ++ digest[ 6] = ctxt->md5_st8[ 5]; digest[ 7] = ctxt->md5_st8[ 4]; ++ digest[ 8] = ctxt->md5_st8[11]; digest[ 9] = ctxt->md5_st8[10]; ++ digest[10] = ctxt->md5_st8[ 9]; digest[11] = ctxt->md5_st8[ 8]; ++ digest[12] = ctxt->md5_st8[15]; digest[13] = ctxt->md5_st8[14]; ++ digest[14] = ctxt->md5_st8[13]; digest[15] = ctxt->md5_st8[12]; ++#endif ++} ++ ++static void md5_calc(b64, ctxt) ++ u_int8_t *b64; ++ md5_ctxt *ctxt; ++{ ++ u_int32_t A = ctxt->md5_sta; ++ u_int32_t B = ctxt->md5_stb; ++ u_int32_t C = ctxt->md5_stc; ++ u_int32_t D = ctxt->md5_std; ++#if BYTE_ORDER == LITTLE_ENDIAN ++ u_int32_t *X = (u_int32_t *)b64; ++#endif ++#if BYTE_ORDER == BIG_ENDIAN ++ /* 4 byte words */ ++ /* what a brute force but fast! */ ++ u_int32_t X[16]; ++ u_int8_t *y = (u_int8_t *)X; ++ y[ 0] = b64[ 3]; y[ 1] = b64[ 2]; y[ 2] = b64[ 1]; y[ 3] = b64[ 0]; ++ y[ 4] = b64[ 7]; y[ 5] = b64[ 6]; y[ 6] = b64[ 5]; y[ 7] = b64[ 4]; ++ y[ 8] = b64[11]; y[ 9] = b64[10]; y[10] = b64[ 9]; y[11] = b64[ 8]; ++ y[12] = b64[15]; y[13] = b64[14]; y[14] = b64[13]; y[15] = b64[12]; ++ y[16] = b64[19]; y[17] = b64[18]; y[18] = b64[17]; y[19] = b64[16]; ++ y[20] = b64[23]; y[21] = b64[22]; y[22] = b64[21]; y[23] = b64[20]; ++ y[24] = b64[27]; y[25] = b64[26]; y[26] = b64[25]; y[27] = b64[24]; ++ y[28] = b64[31]; y[29] = b64[30]; y[30] = b64[29]; y[31] = b64[28]; ++ y[32] = b64[35]; y[33] = b64[34]; y[34] = b64[33]; y[35] = b64[32]; ++ y[36] = b64[39]; y[37] = b64[38]; y[38] = b64[37]; y[39] = b64[36]; ++ y[40] = b64[43]; y[41] = b64[42]; y[42] = b64[41]; y[43] = b64[40]; ++ y[44] = b64[47]; y[45] = b64[46]; y[46] = b64[45]; y[47] = b64[44]; ++ y[48] = b64[51]; y[49] = b64[50]; y[50] = b64[49]; y[51] = b64[48]; ++ y[52] = b64[55]; y[53] = b64[54]; y[54] = b64[53]; y[55] = b64[52]; ++ y[56] = b64[59]; y[57] = b64[58]; y[58] = b64[57]; y[59] = b64[56]; ++ y[60] = b64[63]; y[61] = b64[62]; y[62] = b64[61]; y[63] = b64[60]; ++#endif ++ ++ ROUND1(A, B, C, D, 0, Sa, 1); ROUND1(D, A, B, C, 1, Sb, 2); ++ ROUND1(C, D, A, B, 2, Sc, 3); ROUND1(B, C, D, A, 3, Sd, 4); ++ ROUND1(A, B, C, D, 4, Sa, 5); ROUND1(D, A, B, C, 5, Sb, 6); ++ ROUND1(C, D, A, B, 6, Sc, 7); ROUND1(B, C, D, A, 7, Sd, 8); ++ ROUND1(A, B, C, D, 8, Sa, 9); ROUND1(D, A, B, C, 9, Sb, 10); ++ ROUND1(C, D, A, B, 10, Sc, 11); ROUND1(B, C, D, A, 11, Sd, 12); ++ ROUND1(A, B, C, D, 12, Sa, 13); ROUND1(D, A, B, C, 13, Sb, 14); ++ ROUND1(C, D, A, B, 14, Sc, 15); ROUND1(B, C, D, A, 15, Sd, 16); ++ ++ ROUND2(A, B, C, D, 1, Se, 17); ROUND2(D, A, B, C, 6, Sf, 18); ++ ROUND2(C, D, A, B, 11, Sg, 19); ROUND2(B, C, D, A, 0, Sh, 20); ++ ROUND2(A, B, C, D, 5, Se, 21); ROUND2(D, A, B, C, 10, Sf, 22); ++ ROUND2(C, D, A, B, 15, Sg, 23); ROUND2(B, C, D, A, 4, Sh, 24); ++ ROUND2(A, B, C, D, 9, Se, 25); ROUND2(D, A, B, C, 14, Sf, 26); ++ ROUND2(C, D, A, B, 3, Sg, 27); ROUND2(B, C, D, A, 8, Sh, 28); ++ ROUND2(A, B, C, D, 13, Se, 29); ROUND2(D, A, B, C, 2, Sf, 30); ++ ROUND2(C, D, A, B, 7, Sg, 31); ROUND2(B, C, D, A, 12, Sh, 32); ++ ++ ROUND3(A, B, C, D, 5, Si, 33); ROUND3(D, A, B, C, 8, Sj, 34); ++ ROUND3(C, D, A, B, 11, Sk, 35); ROUND3(B, C, D, A, 14, Sl, 36); ++ ROUND3(A, B, C, D, 1, Si, 37); ROUND3(D, A, B, C, 4, Sj, 38); ++ ROUND3(C, D, A, B, 7, Sk, 39); ROUND3(B, C, D, A, 10, Sl, 40); ++ ROUND3(A, B, C, D, 13, Si, 41); ROUND3(D, A, B, C, 0, Sj, 42); ++ ROUND3(C, D, A, B, 3, Sk, 43); ROUND3(B, C, D, A, 6, Sl, 44); ++ ROUND3(A, B, C, D, 9, Si, 45); ROUND3(D, A, B, C, 12, Sj, 46); ++ ROUND3(C, D, A, B, 15, Sk, 47); ROUND3(B, C, D, A, 2, Sl, 48); ++ ++ ROUND4(A, B, C, D, 0, Sm, 49); ROUND4(D, A, B, C, 7, Sn, 50); ++ ROUND4(C, D, A, B, 14, So, 51); ROUND4(B, C, D, A, 5, Sp, 52); ++ ROUND4(A, B, C, D, 12, Sm, 53); ROUND4(D, A, B, C, 3, Sn, 54); ++ ROUND4(C, D, A, B, 10, So, 55); ROUND4(B, C, D, A, 1, Sp, 56); ++ ROUND4(A, B, C, D, 8, Sm, 57); ROUND4(D, A, B, C, 15, Sn, 58); ++ ROUND4(C, D, A, B, 6, So, 59); ROUND4(B, C, D, A, 13, Sp, 60); ++ ROUND4(A, B, C, D, 4, Sm, 61); ROUND4(D, A, B, C, 11, Sn, 62); ++ ROUND4(C, D, A, B, 2, So, 63); ROUND4(B, C, D, A, 9, Sp, 64); ++ ++ ctxt->md5_sta += A; ++ ctxt->md5_stb += B; ++ ctxt->md5_stc += C; ++ ctxt->md5_std += D; ++} +diff -Nur linux-2.6.30.orig/crypto/ocf/safe/md5.h linux-2.6.30/crypto/ocf/safe/md5.h +--- linux-2.6.30.orig/crypto/ocf/safe/md5.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/safe/md5.h 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,76 @@ ++/* $FreeBSD: src/sys/crypto/md5.h,v 1.4 2002/03/20 05:13:50 alfred Exp $ */ ++/* $KAME: md5.h,v 1.4 2000/03/27 04:36:22 sumikawa Exp $ */ ++ ++/* ++ * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. ++ * All rights reserved. ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions ++ * are met: ++ * 1. Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * 2. Redistributions in binary form must reproduce the above copyright ++ * notice, this list of conditions and the following disclaimer in the ++ * documentation and/or other materials provided with the distribution. ++ * 3. Neither the name of the project nor the names of its contributors ++ * may be used to endorse or promote products derived from this software ++ * without specific prior written permission. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND ++ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE ++ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ++ * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE ++ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL ++ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS ++ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) ++ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT ++ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY ++ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF ++ * SUCH DAMAGE. ++ */ ++ ++#ifndef _NETINET6_MD5_H_ ++#define _NETINET6_MD5_H_ ++ ++#define MD5_BUFLEN 64 ++ ++typedef struct { ++ union { ++ u_int32_t md5_state32[4]; ++ u_int8_t md5_state8[16]; ++ } md5_st; ++ ++#define md5_sta md5_st.md5_state32[0] ++#define md5_stb md5_st.md5_state32[1] ++#define md5_stc md5_st.md5_state32[2] ++#define md5_std md5_st.md5_state32[3] ++#define md5_st8 md5_st.md5_state8 ++ ++ union { ++ u_int64_t md5_count64; ++ u_int8_t md5_count8[8]; ++ } md5_count; ++#define md5_n md5_count.md5_count64 ++#define md5_n8 md5_count.md5_count8 ++ ++ u_int md5_i; ++ u_int8_t md5_buf[MD5_BUFLEN]; ++} md5_ctxt; ++ ++extern void md5_init(md5_ctxt *); ++extern void md5_loop(md5_ctxt *, u_int8_t *, u_int); ++extern void md5_pad(md5_ctxt *); ++extern void md5_result(u_int8_t *, md5_ctxt *); ++ ++/* compatibility */ ++#define MD5_CTX md5_ctxt ++#define MD5Init(x) md5_init((x)) ++#define MD5Update(x, y, z) md5_loop((x), (y), (z)) ++#define MD5Final(x, y) \ ++do { \ ++ md5_pad((y)); \ ++ md5_result((x), (y)); \ ++} while (0) ++ ++#endif /* ! _NETINET6_MD5_H_*/ +diff -Nur linux-2.6.30.orig/crypto/ocf/safe/safe.c linux-2.6.30/crypto/ocf/safe/safe.c +--- linux-2.6.30.orig/crypto/ocf/safe/safe.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/safe/safe.c 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,2288 @@ ++/*- ++ * Linux port done by David McCullough <david_mccullough@securecomputing.com> ++ * Copyright (C) 2004-2007 David McCullough ++ * The license and original author are listed below. ++ * ++ * Copyright (c) 2003 Sam Leffler, Errno Consulting ++ * Copyright (c) 2003 Global Technology Associates, Inc. ++ * All rights reserved. ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions ++ * are met: ++ * 1. Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * 2. Redistributions in binary form must reproduce the above copyright ++ * notice, this list of conditions and the following disclaimer in the ++ * documentation and/or other materials provided with the distribution. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND ++ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE ++ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ++ * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE ++ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL ++ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS ++ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) ++ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT ++ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY ++ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF ++ * SUCH DAMAGE. ++ * ++__FBSDID("$FreeBSD: src/sys/dev/safe/safe.c,v 1.18 2007/03/21 03:42:50 sam Exp $"); ++ */ ++ ++#ifndef AUTOCONF_INCLUDED ++#include <linux/config.h> ++#endif ++#include <linux/module.h> ++#include <linux/kernel.h> ++#include <linux/init.h> ++#include <linux/list.h> ++#include <linux/slab.h> ++#include <linux/wait.h> ++#include <linux/sched.h> ++#include <linux/pci.h> ++#include <linux/delay.h> ++#include <linux/interrupt.h> ++#include <linux/spinlock.h> ++#include <linux/random.h> ++#include <linux/version.h> ++#include <linux/skbuff.h> ++#include <asm/io.h> ++ ++/* ++ * SafeNet SafeXcel-1141 hardware crypto accelerator ++ */ ++ ++#include <cryptodev.h> ++#include <uio.h> ++#include <safe/safereg.h> ++#include <safe/safevar.h> ++ ++#if 1 ++#define DPRINTF(a) do { \ ++ if (debug) { \ ++ printk("%s: ", sc ? \ ++ device_get_nameunit(sc->sc_dev) : "safe"); \ ++ printk a; \ ++ } \ ++ } while (0) ++#else ++#define DPRINTF(a) ++#endif ++ ++/* ++ * until we find a cleaner way, include the BSD md5/sha1 code ++ * here ++ */ ++#define HMAC_HACK 1 ++#ifdef HMAC_HACK ++#define LITTLE_ENDIAN 1234 ++#define BIG_ENDIAN 4321 ++#ifdef __LITTLE_ENDIAN ++#define BYTE_ORDER LITTLE_ENDIAN ++#endif ++#ifdef __BIG_ENDIAN ++#define BYTE_ORDER BIG_ENDIAN ++#endif ++#include <safe/md5.h> ++#include <safe/md5.c> ++#include <safe/sha1.h> ++#include <safe/sha1.c> ++ ++u_int8_t hmac_ipad_buffer[64] = { ++ 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, ++ 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, ++ 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, ++ 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, ++ 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, ++ 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, ++ 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, ++ 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36 ++}; ++ ++u_int8_t hmac_opad_buffer[64] = { ++ 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, ++ 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, ++ 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, ++ 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, ++ 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, ++ 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, ++ 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, ++ 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C ++}; ++#endif /* HMAC_HACK */ ++ ++/* add proc entry for this */ ++struct safe_stats safestats; ++ ++#define debug safe_debug ++int safe_debug = 0; ++module_param(safe_debug, int, 0644); ++MODULE_PARM_DESC(safe_debug, "Enable debug"); ++ ++static void safe_callback(struct safe_softc *, struct safe_ringentry *); ++static void safe_feed(struct safe_softc *, struct safe_ringentry *); ++#if defined(CONFIG_OCF_RANDOMHARVEST) && !defined(SAFE_NO_RNG) ++static void safe_rng_init(struct safe_softc *); ++int safe_rngbufsize = 8; /* 32 bytes each read */ ++module_param(safe_rngbufsize, int, 0644); ++MODULE_PARM_DESC(safe_rngbufsize, "RNG polling buffer size (32-bit words)"); ++int safe_rngmaxalarm = 8; /* max alarms before reset */ ++module_param(safe_rngmaxalarm, int, 0644); ++MODULE_PARM_DESC(safe_rngmaxalarm, "RNG max alarms before reset"); ++#endif /* SAFE_NO_RNG */ ++ ++static void safe_totalreset(struct safe_softc *sc); ++static int safe_dmamap_aligned(struct safe_softc *sc, const struct safe_operand *op); ++static int safe_dmamap_uniform(struct safe_softc *sc, const struct safe_operand *op); ++static int safe_free_entry(struct safe_softc *sc, struct safe_ringentry *re); ++static int safe_kprocess(device_t dev, struct cryptkop *krp, int hint); ++static int safe_kstart(struct safe_softc *sc); ++static int safe_ksigbits(struct safe_softc *sc, struct crparam *cr); ++static void safe_kfeed(struct safe_softc *sc); ++static void safe_kpoll(unsigned long arg); ++static void safe_kload_reg(struct safe_softc *sc, u_int32_t off, ++ u_int32_t len, struct crparam *n); ++ ++static int safe_newsession(device_t, u_int32_t *, struct cryptoini *); ++static int safe_freesession(device_t, u_int64_t); ++static int safe_process(device_t, struct cryptop *, int); ++ ++static device_method_t safe_methods = { ++ /* crypto device methods */ ++ DEVMETHOD(cryptodev_newsession, safe_newsession), ++ DEVMETHOD(cryptodev_freesession,safe_freesession), ++ DEVMETHOD(cryptodev_process, safe_process), ++ DEVMETHOD(cryptodev_kprocess, safe_kprocess), ++}; ++ ++#define READ_REG(sc,r) readl((sc)->sc_base_addr + (r)) ++#define WRITE_REG(sc,r,val) writel((val), (sc)->sc_base_addr + (r)) ++ ++#define SAFE_MAX_CHIPS 8 ++static struct safe_softc *safe_chip_idx[SAFE_MAX_CHIPS]; ++ ++/* ++ * split our buffers up into safe DMAable byte fragments to avoid lockup ++ * bug in 1141 HW on rev 1.0. ++ */ ++ ++static int ++pci_map_linear( ++ struct safe_softc *sc, ++ struct safe_operand *buf, ++ void *addr, ++ int len) ++{ ++ dma_addr_t tmp; ++ int chunk, tlen = len; ++ ++ tmp = pci_map_single(sc->sc_pcidev, addr, len, PCI_DMA_BIDIRECTIONAL); ++ ++ buf->mapsize += len; ++ while (len > 0) { ++ chunk = (len > sc->sc_max_dsize) ? sc->sc_max_dsize : len; ++ buf->segs[buf->nsegs].ds_addr = tmp; ++ buf->segs[buf->nsegs].ds_len = chunk; ++ buf->segs[buf->nsegs].ds_tlen = tlen; ++ buf->nsegs++; ++ tmp += chunk; ++ len -= chunk; ++ tlen = 0; ++ } ++ return 0; ++} ++ ++/* ++ * map in a given uio buffer (great on some arches :-) ++ */ ++ ++static int ++pci_map_uio(struct safe_softc *sc, struct safe_operand *buf, struct uio *uio) ++{ ++ struct iovec *iov = uio->uio_iov; ++ int n; ++ ++ DPRINTF(("%s()\n", __FUNCTION__)); ++ ++ buf->mapsize = 0; ++ buf->nsegs = 0; ++ ++ for (n = 0; n < uio->uio_iovcnt; n++) { ++ pci_map_linear(sc, buf, iov->iov_base, iov->iov_len); ++ iov++; ++ } ++ ++ /* identify this buffer by the first segment */ ++ buf->map = (void *) buf->segs[0].ds_addr; ++ return(0); ++} ++ ++/* ++ * map in a given sk_buff ++ */ ++ ++static int ++pci_map_skb(struct safe_softc *sc,struct safe_operand *buf,struct sk_buff *skb) ++{ ++ int i; ++ ++ DPRINTF(("%s()\n", __FUNCTION__)); ++ ++ buf->mapsize = 0; ++ buf->nsegs = 0; ++ ++ pci_map_linear(sc, buf, skb->data, skb_headlen(skb)); ++ ++ for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) { ++ pci_map_linear(sc, buf, ++ page_address(skb_shinfo(skb)->frags[i].page) + ++ skb_shinfo(skb)->frags[i].page_offset, ++ skb_shinfo(skb)->frags[i].size); ++ } ++ ++ /* identify this buffer by the first segment */ ++ buf->map = (void *) buf->segs[0].ds_addr; ++ return(0); ++} ++ ++ ++#if 0 /* not needed at this time */ ++static void ++pci_sync_operand(struct safe_softc *sc, struct safe_operand *buf) ++{ ++ int i; ++ ++ DPRINTF(("%s()\n", __FUNCTION__)); ++ for (i = 0; i < buf->nsegs; i++) ++ pci_dma_sync_single_for_cpu(sc->sc_pcidev, buf->segs[i].ds_addr, ++ buf->segs[i].ds_len, PCI_DMA_BIDIRECTIONAL); ++} ++#endif ++ ++static void ++pci_unmap_operand(struct safe_softc *sc, struct safe_operand *buf) ++{ ++ int i; ++ DPRINTF(("%s()\n", __FUNCTION__)); ++ for (i = 0; i < buf->nsegs; i++) { ++ if (buf->segs[i].ds_tlen) { ++ DPRINTF(("%s - unmap %d 0x%x %d\n", __FUNCTION__, i, buf->segs[i].ds_addr, buf->segs[i].ds_tlen)); ++ pci_unmap_single(sc->sc_pcidev, buf->segs[i].ds_addr, ++ buf->segs[i].ds_tlen, PCI_DMA_BIDIRECTIONAL); ++ DPRINTF(("%s - unmap %d 0x%x %d done\n", __FUNCTION__, i, buf->segs[i].ds_addr, buf->segs[i].ds_tlen)); ++ } ++ buf->segs[i].ds_addr = 0; ++ buf->segs[i].ds_len = 0; ++ buf->segs[i].ds_tlen = 0; ++ } ++ buf->nsegs = 0; ++ buf->mapsize = 0; ++ buf->map = 0; ++} ++ ++ ++/* ++ * SafeXcel Interrupt routine ++ */ ++static irqreturn_t ++#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,19) ++safe_intr(int irq, void *arg) ++#else ++safe_intr(int irq, void *arg, struct pt_regs *regs) ++#endif ++{ ++ struct safe_softc *sc = arg; ++ int stat; ++ unsigned long flags; ++ ++ stat = READ_REG(sc, SAFE_HM_STAT); ++ ++ DPRINTF(("%s(stat=0x%x)\n", __FUNCTION__, stat)); ++ ++ if (stat == 0) /* shared irq, not for us */ ++ return IRQ_NONE; ++ ++ WRITE_REG(sc, SAFE_HI_CLR, stat); /* IACK */ ++ ++ if ((stat & SAFE_INT_PE_DDONE)) { ++ /* ++ * Descriptor(s) done; scan the ring and ++ * process completed operations. ++ */ ++ spin_lock_irqsave(&sc->sc_ringmtx, flags); ++ while (sc->sc_back != sc->sc_front) { ++ struct safe_ringentry *re = sc->sc_back; ++ ++#ifdef SAFE_DEBUG ++ if (debug) { ++ safe_dump_ringstate(sc, __func__); ++ safe_dump_request(sc, __func__, re); ++ } ++#endif ++ /* ++ * safe_process marks ring entries that were allocated ++ * but not used with a csr of zero. This insures the ++ * ring front pointer never needs to be set backwards ++ * in the event that an entry is allocated but not used ++ * because of a setup error. ++ */ ++ DPRINTF(("%s re->re_desc.d_csr=0x%x\n", __FUNCTION__, re->re_desc.d_csr)); ++ if (re->re_desc.d_csr != 0) { ++ if (!SAFE_PE_CSR_IS_DONE(re->re_desc.d_csr)) { ++ DPRINTF(("%s !CSR_IS_DONE\n", __FUNCTION__)); ++ break; ++ } ++ if (!SAFE_PE_LEN_IS_DONE(re->re_desc.d_len)) { ++ DPRINTF(("%s !LEN_IS_DONE\n", __FUNCTION__)); ++ break; ++ } ++ sc->sc_nqchip--; ++ safe_callback(sc, re); ++ } ++ if (++(sc->sc_back) == sc->sc_ringtop) ++ sc->sc_back = sc->sc_ring; ++ } ++ spin_unlock_irqrestore(&sc->sc_ringmtx, flags); ++ } ++ ++ /* ++ * Check to see if we got any DMA Error ++ */ ++ if (stat & SAFE_INT_PE_ERROR) { ++ printk("%s: dmaerr dmastat %08x\n", device_get_nameunit(sc->sc_dev), ++ (int)READ_REG(sc, SAFE_PE_DMASTAT)); ++ safestats.st_dmaerr++; ++ safe_totalreset(sc); ++#if 0 ++ safe_feed(sc); ++#endif ++ } ++ ++ if (sc->sc_needwakeup) { /* XXX check high watermark */ ++ int wakeup = sc->sc_needwakeup & (CRYPTO_SYMQ|CRYPTO_ASYMQ); ++ DPRINTF(("%s: wakeup crypto %x\n", __func__, ++ sc->sc_needwakeup)); ++ sc->sc_needwakeup &= ~wakeup; ++ crypto_unblock(sc->sc_cid, wakeup); ++ } ++ ++ return IRQ_HANDLED; ++} ++ ++/* ++ * safe_feed() - post a request to chip ++ */ ++static void ++safe_feed(struct safe_softc *sc, struct safe_ringentry *re) ++{ ++ DPRINTF(("%s()\n", __FUNCTION__)); ++#ifdef SAFE_DEBUG ++ if (debug) { ++ safe_dump_ringstate(sc, __func__); ++ safe_dump_request(sc, __func__, re); ++ } ++#endif ++ sc->sc_nqchip++; ++ if (sc->sc_nqchip > safestats.st_maxqchip) ++ safestats.st_maxqchip = sc->sc_nqchip; ++ /* poke h/w to check descriptor ring, any value can be written */ ++ WRITE_REG(sc, SAFE_HI_RD_DESCR, 0); ++} ++ ++#define N(a) (sizeof(a) / sizeof (a[0])) ++static void ++safe_setup_enckey(struct safe_session *ses, caddr_t key) ++{ ++ int i; ++ ++ bcopy(key, ses->ses_key, ses->ses_klen / 8); ++ ++ /* PE is little-endian, insure proper byte order */ ++ for (i = 0; i < N(ses->ses_key); i++) ++ ses->ses_key[i] = htole32(ses->ses_key[i]); ++} ++ ++static void ++safe_setup_mackey(struct safe_session *ses, int algo, caddr_t key, int klen) ++{ ++#ifdef HMAC_HACK ++ MD5_CTX md5ctx; ++ SHA1_CTX sha1ctx; ++ int i; ++ ++ ++ for (i = 0; i < klen; i++) ++ key[i] ^= HMAC_IPAD_VAL; ++ ++ if (algo == CRYPTO_MD5_HMAC) { ++ MD5Init(&md5ctx); ++ MD5Update(&md5ctx, key, klen); ++ MD5Update(&md5ctx, hmac_ipad_buffer, MD5_HMAC_BLOCK_LEN - klen); ++ bcopy(md5ctx.md5_st8, ses->ses_hminner, sizeof(md5ctx.md5_st8)); ++ } else { ++ SHA1Init(&sha1ctx); ++ SHA1Update(&sha1ctx, key, klen); ++ SHA1Update(&sha1ctx, hmac_ipad_buffer, ++ SHA1_HMAC_BLOCK_LEN - klen); ++ bcopy(sha1ctx.h.b32, ses->ses_hminner, sizeof(sha1ctx.h.b32)); ++ } ++ ++ for (i = 0; i < klen; i++) ++ key[i] ^= (HMAC_IPAD_VAL ^ HMAC_OPAD_VAL); ++ ++ if (algo == CRYPTO_MD5_HMAC) { ++ MD5Init(&md5ctx); ++ MD5Update(&md5ctx, key, klen); ++ MD5Update(&md5ctx, hmac_opad_buffer, MD5_HMAC_BLOCK_LEN - klen); ++ bcopy(md5ctx.md5_st8, ses->ses_hmouter, sizeof(md5ctx.md5_st8)); ++ } else { ++ SHA1Init(&sha1ctx); ++ SHA1Update(&sha1ctx, key, klen); ++ SHA1Update(&sha1ctx, hmac_opad_buffer, ++ SHA1_HMAC_BLOCK_LEN - klen); ++ bcopy(sha1ctx.h.b32, ses->ses_hmouter, sizeof(sha1ctx.h.b32)); ++ } ++ ++ for (i = 0; i < klen; i++) ++ key[i] ^= HMAC_OPAD_VAL; ++ ++#if 0 ++ /* ++ * this code prevents SHA working on a BE host, ++ * so it is obviously wrong. I think the byte ++ * swap setup we do with the chip fixes this for us ++ */ ++ ++ /* PE is little-endian, insure proper byte order */ ++ for (i = 0; i < N(ses->ses_hminner); i++) { ++ ses->ses_hminner[i] = htole32(ses->ses_hminner[i]); ++ ses->ses_hmouter[i] = htole32(ses->ses_hmouter[i]); ++ } ++#endif ++#else /* HMAC_HACK */ ++ printk("safe: md5/sha not implemented\n"); ++#endif /* HMAC_HACK */ ++} ++#undef N ++ ++/* ++ * Allocate a new 'session' and return an encoded session id. 'sidp' ++ * contains our registration id, and should contain an encoded session ++ * id on successful allocation. ++ */ ++static int ++safe_newsession(device_t dev, u_int32_t *sidp, struct cryptoini *cri) ++{ ++ struct safe_softc *sc = device_get_softc(dev); ++ struct cryptoini *c, *encini = NULL, *macini = NULL; ++ struct safe_session *ses = NULL; ++ int sesn; ++ ++ DPRINTF(("%s()\n", __FUNCTION__)); ++ ++ if (sidp == NULL || cri == NULL || sc == NULL) ++ return (EINVAL); ++ ++ for (c = cri; c != NULL; c = c->cri_next) { ++ if (c->cri_alg == CRYPTO_MD5_HMAC || ++ c->cri_alg == CRYPTO_SHA1_HMAC || ++ c->cri_alg == CRYPTO_NULL_HMAC) { ++ if (macini) ++ return (EINVAL); ++ macini = c; ++ } else if (c->cri_alg == CRYPTO_DES_CBC || ++ c->cri_alg == CRYPTO_3DES_CBC || ++ c->cri_alg == CRYPTO_AES_CBC || ++ c->cri_alg == CRYPTO_NULL_CBC) { ++ if (encini) ++ return (EINVAL); ++ encini = c; ++ } else ++ return (EINVAL); ++ } ++ if (encini == NULL && macini == NULL) ++ return (EINVAL); ++ if (encini) { /* validate key length */ ++ switch (encini->cri_alg) { ++ case CRYPTO_DES_CBC: ++ if (encini->cri_klen != 64) ++ return (EINVAL); ++ break; ++ case CRYPTO_3DES_CBC: ++ if (encini->cri_klen != 192) ++ return (EINVAL); ++ break; ++ case CRYPTO_AES_CBC: ++ if (encini->cri_klen != 128 && ++ encini->cri_klen != 192 && ++ encini->cri_klen != 256) ++ return (EINVAL); ++ break; ++ } ++ } ++ ++ if (sc->sc_sessions == NULL) { ++ ses = sc->sc_sessions = (struct safe_session *) ++ kmalloc(sizeof(struct safe_session), SLAB_ATOMIC); ++ if (ses == NULL) ++ return (ENOMEM); ++ memset(ses, 0, sizeof(struct safe_session)); ++ sesn = 0; ++ sc->sc_nsessions = 1; ++ } else { ++ for (sesn = 0; sesn < sc->sc_nsessions; sesn++) { ++ if (sc->sc_sessions[sesn].ses_used == 0) { ++ ses = &sc->sc_sessions[sesn]; ++ break; ++ } ++ } ++ ++ if (ses == NULL) { ++ sesn = sc->sc_nsessions; ++ ses = (struct safe_session *) ++ kmalloc((sesn + 1) * sizeof(struct safe_session), SLAB_ATOMIC); ++ if (ses == NULL) ++ return (ENOMEM); ++ memset(ses, 0, (sesn + 1) * sizeof(struct safe_session)); ++ bcopy(sc->sc_sessions, ses, sesn * ++ sizeof(struct safe_session)); ++ bzero(sc->sc_sessions, sesn * ++ sizeof(struct safe_session)); ++ kfree(sc->sc_sessions); ++ sc->sc_sessions = ses; ++ ses = &sc->sc_sessions[sesn]; ++ sc->sc_nsessions++; ++ } ++ } ++ ++ bzero(ses, sizeof(struct safe_session)); ++ ses->ses_used = 1; ++ ++ if (encini) { ++ /* get an IV */ ++ /* XXX may read fewer than requested */ ++ read_random(ses->ses_iv, sizeof(ses->ses_iv)); ++ ++ ses->ses_klen = encini->cri_klen; ++ if (encini->cri_key != NULL) ++ safe_setup_enckey(ses, encini->cri_key); ++ } ++ ++ if (macini) { ++ ses->ses_mlen = macini->cri_mlen; ++ if (ses->ses_mlen == 0) { ++ if (macini->cri_alg == CRYPTO_MD5_HMAC) ++ ses->ses_mlen = MD5_HASH_LEN; ++ else ++ ses->ses_mlen = SHA1_HASH_LEN; ++ } ++ ++ if (macini->cri_key != NULL) { ++ safe_setup_mackey(ses, macini->cri_alg, macini->cri_key, ++ macini->cri_klen / 8); ++ } ++ } ++ ++ *sidp = SAFE_SID(device_get_unit(sc->sc_dev), sesn); ++ return (0); ++} ++ ++/* ++ * Deallocate a session. ++ */ ++static int ++safe_freesession(device_t dev, u_int64_t tid) ++{ ++ struct safe_softc *sc = device_get_softc(dev); ++ int session, ret; ++ u_int32_t sid = ((u_int32_t) tid) & 0xffffffff; ++ ++ DPRINTF(("%s()\n", __FUNCTION__)); ++ ++ if (sc == NULL) ++ return (EINVAL); ++ ++ session = SAFE_SESSION(sid); ++ if (session < sc->sc_nsessions) { ++ bzero(&sc->sc_sessions[session], sizeof(sc->sc_sessions[session])); ++ ret = 0; ++ } else ++ ret = EINVAL; ++ return (ret); ++} ++ ++ ++static int ++safe_process(device_t dev, struct cryptop *crp, int hint) ++{ ++ struct safe_softc *sc = device_get_softc(dev); ++ int err = 0, i, nicealign, uniform; ++ struct cryptodesc *crd1, *crd2, *maccrd, *enccrd; ++ int bypass, oplen, ivsize; ++ caddr_t iv; ++ int16_t coffset; ++ struct safe_session *ses; ++ struct safe_ringentry *re; ++ struct safe_sarec *sa; ++ struct safe_pdesc *pd; ++ u_int32_t cmd0, cmd1, staterec; ++ unsigned long flags; ++ ++ DPRINTF(("%s()\n", __FUNCTION__)); ++ ++ if (crp == NULL || crp->crp_callback == NULL || sc == NULL) { ++ safestats.st_invalid++; ++ return (EINVAL); ++ } ++ if (SAFE_SESSION(crp->crp_sid) >= sc->sc_nsessions) { ++ safestats.st_badsession++; ++ return (EINVAL); ++ } ++ ++ spin_lock_irqsave(&sc->sc_ringmtx, flags); ++ if (sc->sc_front == sc->sc_back && sc->sc_nqchip != 0) { ++ safestats.st_ringfull++; ++ sc->sc_needwakeup |= CRYPTO_SYMQ; ++ spin_unlock_irqrestore(&sc->sc_ringmtx, flags); ++ return (ERESTART); ++ } ++ re = sc->sc_front; ++ ++ staterec = re->re_sa.sa_staterec; /* save */ ++ /* NB: zero everything but the PE descriptor */ ++ bzero(&re->re_sa, sizeof(struct safe_ringentry) - sizeof(re->re_desc)); ++ re->re_sa.sa_staterec = staterec; /* restore */ ++ ++ re->re_crp = crp; ++ re->re_sesn = SAFE_SESSION(crp->crp_sid); ++ ++ re->re_src.nsegs = 0; ++ re->re_dst.nsegs = 0; ++ ++ if (crp->crp_flags & CRYPTO_F_SKBUF) { ++ re->re_src_skb = (struct sk_buff *)crp->crp_buf; ++ re->re_dst_skb = (struct sk_buff *)crp->crp_buf; ++ } else if (crp->crp_flags & CRYPTO_F_IOV) { ++ re->re_src_io = (struct uio *)crp->crp_buf; ++ re->re_dst_io = (struct uio *)crp->crp_buf; ++ } else { ++ safestats.st_badflags++; ++ err = EINVAL; ++ goto errout; /* XXX we don't handle contiguous blocks! */ ++ } ++ ++ sa = &re->re_sa; ++ ses = &sc->sc_sessions[re->re_sesn]; ++ ++ crd1 = crp->crp_desc; ++ if (crd1 == NULL) { ++ safestats.st_nodesc++; ++ err = EINVAL; ++ goto errout; ++ } ++ crd2 = crd1->crd_next; ++ ++ cmd0 = SAFE_SA_CMD0_BASIC; /* basic group operation */ ++ cmd1 = 0; ++ if (crd2 == NULL) { ++ if (crd1->crd_alg == CRYPTO_MD5_HMAC || ++ crd1->crd_alg == CRYPTO_SHA1_HMAC || ++ crd1->crd_alg == CRYPTO_NULL_HMAC) { ++ maccrd = crd1; ++ enccrd = NULL; ++ cmd0 |= SAFE_SA_CMD0_OP_HASH; ++ } else if (crd1->crd_alg == CRYPTO_DES_CBC || ++ crd1->crd_alg == CRYPTO_3DES_CBC || ++ crd1->crd_alg == CRYPTO_AES_CBC || ++ crd1->crd_alg == CRYPTO_NULL_CBC) { ++ maccrd = NULL; ++ enccrd = crd1; ++ cmd0 |= SAFE_SA_CMD0_OP_CRYPT; ++ } else { ++ safestats.st_badalg++; ++ err = EINVAL; ++ goto errout; ++ } ++ } else { ++ if ((crd1->crd_alg == CRYPTO_MD5_HMAC || ++ crd1->crd_alg == CRYPTO_SHA1_HMAC || ++ crd1->crd_alg == CRYPTO_NULL_HMAC) && ++ (crd2->crd_alg == CRYPTO_DES_CBC || ++ crd2->crd_alg == CRYPTO_3DES_CBC || ++ crd2->crd_alg == CRYPTO_AES_CBC || ++ crd2->crd_alg == CRYPTO_NULL_CBC) && ++ ((crd2->crd_flags & CRD_F_ENCRYPT) == 0)) { ++ maccrd = crd1; ++ enccrd = crd2; ++ } else if ((crd1->crd_alg == CRYPTO_DES_CBC || ++ crd1->crd_alg == CRYPTO_3DES_CBC || ++ crd1->crd_alg == CRYPTO_AES_CBC || ++ crd1->crd_alg == CRYPTO_NULL_CBC) && ++ (crd2->crd_alg == CRYPTO_MD5_HMAC || ++ crd2->crd_alg == CRYPTO_SHA1_HMAC || ++ crd2->crd_alg == CRYPTO_NULL_HMAC) && ++ (crd1->crd_flags & CRD_F_ENCRYPT)) { ++ enccrd = crd1; ++ maccrd = crd2; ++ } else { ++ safestats.st_badalg++; ++ err = EINVAL; ++ goto errout; ++ } ++ cmd0 |= SAFE_SA_CMD0_OP_BOTH; ++ } ++ ++ if (enccrd) { ++ if (enccrd->crd_flags & CRD_F_KEY_EXPLICIT) ++ safe_setup_enckey(ses, enccrd->crd_key); ++ ++ if (enccrd->crd_alg == CRYPTO_DES_CBC) { ++ cmd0 |= SAFE_SA_CMD0_DES; ++ cmd1 |= SAFE_SA_CMD1_CBC; ++ ivsize = 2*sizeof(u_int32_t); ++ } else if (enccrd->crd_alg == CRYPTO_3DES_CBC) { ++ cmd0 |= SAFE_SA_CMD0_3DES; ++ cmd1 |= SAFE_SA_CMD1_CBC; ++ ivsize = 2*sizeof(u_int32_t); ++ } else if (enccrd->crd_alg == CRYPTO_AES_CBC) { ++ cmd0 |= SAFE_SA_CMD0_AES; ++ cmd1 |= SAFE_SA_CMD1_CBC; ++ if (ses->ses_klen == 128) ++ cmd1 |= SAFE_SA_CMD1_AES128; ++ else if (ses->ses_klen == 192) ++ cmd1 |= SAFE_SA_CMD1_AES192; ++ else ++ cmd1 |= SAFE_SA_CMD1_AES256; ++ ivsize = 4*sizeof(u_int32_t); ++ } else { ++ cmd0 |= SAFE_SA_CMD0_CRYPT_NULL; ++ ivsize = 0; ++ } ++ ++ /* ++ * Setup encrypt/decrypt state. When using basic ops ++ * we can't use an inline IV because hash/crypt offset ++ * must be from the end of the IV to the start of the ++ * crypt data and this leaves out the preceding header ++ * from the hash calculation. Instead we place the IV ++ * in the state record and set the hash/crypt offset to ++ * copy both the header+IV. ++ */ ++ if (enccrd->crd_flags & CRD_F_ENCRYPT) { ++ cmd0 |= SAFE_SA_CMD0_OUTBOUND; ++ ++ if (enccrd->crd_flags & CRD_F_IV_EXPLICIT) ++ iv = enccrd->crd_iv; ++ else ++ iv = (caddr_t) ses->ses_iv; ++ if ((enccrd->crd_flags & CRD_F_IV_PRESENT) == 0) { ++ crypto_copyback(crp->crp_flags, crp->crp_buf, ++ enccrd->crd_inject, ivsize, iv); ++ } ++ bcopy(iv, re->re_sastate.sa_saved_iv, ivsize); ++ /* make iv LE */ ++ for (i = 0; i < ivsize/sizeof(re->re_sastate.sa_saved_iv[0]); i++) ++ re->re_sastate.sa_saved_iv[i] = ++ cpu_to_le32(re->re_sastate.sa_saved_iv[i]); ++ cmd0 |= SAFE_SA_CMD0_IVLD_STATE | SAFE_SA_CMD0_SAVEIV; ++ re->re_flags |= SAFE_QFLAGS_COPYOUTIV; ++ } else { ++ cmd0 |= SAFE_SA_CMD0_INBOUND; ++ ++ if (enccrd->crd_flags & CRD_F_IV_EXPLICIT) { ++ bcopy(enccrd->crd_iv, ++ re->re_sastate.sa_saved_iv, ivsize); ++ } else { ++ crypto_copydata(crp->crp_flags, crp->crp_buf, ++ enccrd->crd_inject, ivsize, ++ (caddr_t)re->re_sastate.sa_saved_iv); ++ } ++ /* make iv LE */ ++ for (i = 0; i < ivsize/sizeof(re->re_sastate.sa_saved_iv[0]); i++) ++ re->re_sastate.sa_saved_iv[i] = ++ cpu_to_le32(re->re_sastate.sa_saved_iv[i]); ++ cmd0 |= SAFE_SA_CMD0_IVLD_STATE; ++ } ++ /* ++ * For basic encryption use the zero pad algorithm. ++ * This pads results to an 8-byte boundary and ++ * suppresses padding verification for inbound (i.e. ++ * decrypt) operations. ++ * ++ * NB: Not sure if the 8-byte pad boundary is a problem. ++ */ ++ cmd0 |= SAFE_SA_CMD0_PAD_ZERO; ++ ++ /* XXX assert key bufs have the same size */ ++ bcopy(ses->ses_key, sa->sa_key, sizeof(sa->sa_key)); ++ } ++ ++ if (maccrd) { ++ if (maccrd->crd_flags & CRD_F_KEY_EXPLICIT) { ++ safe_setup_mackey(ses, maccrd->crd_alg, ++ maccrd->crd_key, maccrd->crd_klen / 8); ++ } ++ ++ if (maccrd->crd_alg == CRYPTO_MD5_HMAC) { ++ cmd0 |= SAFE_SA_CMD0_MD5; ++ cmd1 |= SAFE_SA_CMD1_HMAC; /* NB: enable HMAC */ ++ } else if (maccrd->crd_alg == CRYPTO_SHA1_HMAC) { ++ cmd0 |= SAFE_SA_CMD0_SHA1; ++ cmd1 |= SAFE_SA_CMD1_HMAC; /* NB: enable HMAC */ ++ } else { ++ cmd0 |= SAFE_SA_CMD0_HASH_NULL; ++ } ++ /* ++ * Digest data is loaded from the SA and the hash ++ * result is saved to the state block where we ++ * retrieve it for return to the caller. ++ */ ++ /* XXX assert digest bufs have the same size */ ++ bcopy(ses->ses_hminner, sa->sa_indigest, ++ sizeof(sa->sa_indigest)); ++ bcopy(ses->ses_hmouter, sa->sa_outdigest, ++ sizeof(sa->sa_outdigest)); ++ ++ cmd0 |= SAFE_SA_CMD0_HSLD_SA | SAFE_SA_CMD0_SAVEHASH; ++ re->re_flags |= SAFE_QFLAGS_COPYOUTICV; ++ } ++ ++ if (enccrd && maccrd) { ++ /* ++ * The offset from hash data to the start of ++ * crypt data is the difference in the skips. ++ */ ++ bypass = maccrd->crd_skip; ++ coffset = enccrd->crd_skip - maccrd->crd_skip; ++ if (coffset < 0) { ++ DPRINTF(("%s: hash does not precede crypt; " ++ "mac skip %u enc skip %u\n", ++ __func__, maccrd->crd_skip, enccrd->crd_skip)); ++ safestats.st_skipmismatch++; ++ err = EINVAL; ++ goto errout; ++ } ++ oplen = enccrd->crd_skip + enccrd->crd_len; ++ if (maccrd->crd_skip + maccrd->crd_len != oplen) { ++ DPRINTF(("%s: hash amount %u != crypt amount %u\n", ++ __func__, maccrd->crd_skip + maccrd->crd_len, ++ oplen)); ++ safestats.st_lenmismatch++; ++ err = EINVAL; ++ goto errout; ++ } ++#ifdef SAFE_DEBUG ++ if (debug) { ++ printf("mac: skip %d, len %d, inject %d\n", ++ maccrd->crd_skip, maccrd->crd_len, ++ maccrd->crd_inject); ++ printf("enc: skip %d, len %d, inject %d\n", ++ enccrd->crd_skip, enccrd->crd_len, ++ enccrd->crd_inject); ++ printf("bypass %d coffset %d oplen %d\n", ++ bypass, coffset, oplen); ++ } ++#endif ++ if (coffset & 3) { /* offset must be 32-bit aligned */ ++ DPRINTF(("%s: coffset %u misaligned\n", ++ __func__, coffset)); ++ safestats.st_coffmisaligned++; ++ err = EINVAL; ++ goto errout; ++ } ++ coffset >>= 2; ++ if (coffset > 255) { /* offset must be <256 dwords */ ++ DPRINTF(("%s: coffset %u too big\n", ++ __func__, coffset)); ++ safestats.st_cofftoobig++; ++ err = EINVAL; ++ goto errout; ++ } ++ /* ++ * Tell the hardware to copy the header to the output. ++ * The header is defined as the data from the end of ++ * the bypass to the start of data to be encrypted. ++ * Typically this is the inline IV. Note that you need ++ * to do this even if src+dst are the same; it appears ++ * that w/o this bit the crypted data is written ++ * immediately after the bypass data. ++ */ ++ cmd1 |= SAFE_SA_CMD1_HDRCOPY; ++ /* ++ * Disable IP header mutable bit handling. This is ++ * needed to get correct HMAC calculations. ++ */ ++ cmd1 |= SAFE_SA_CMD1_MUTABLE; ++ } else { ++ if (enccrd) { ++ bypass = enccrd->crd_skip; ++ oplen = bypass + enccrd->crd_len; ++ } else { ++ bypass = maccrd->crd_skip; ++ oplen = bypass + maccrd->crd_len; ++ } ++ coffset = 0; ++ } ++ /* XXX verify multiple of 4 when using s/g */ ++ if (bypass > 96) { /* bypass offset must be <= 96 bytes */ ++ DPRINTF(("%s: bypass %u too big\n", __func__, bypass)); ++ safestats.st_bypasstoobig++; ++ err = EINVAL; ++ goto errout; ++ } ++ ++ if (crp->crp_flags & CRYPTO_F_SKBUF) { ++ if (pci_map_skb(sc, &re->re_src, re->re_src_skb)) { ++ safestats.st_noload++; ++ err = ENOMEM; ++ goto errout; ++ } ++ } else if (crp->crp_flags & CRYPTO_F_IOV) { ++ if (pci_map_uio(sc, &re->re_src, re->re_src_io)) { ++ safestats.st_noload++; ++ err = ENOMEM; ++ goto errout; ++ } ++ } ++ nicealign = safe_dmamap_aligned(sc, &re->re_src); ++ uniform = safe_dmamap_uniform(sc, &re->re_src); ++ ++ DPRINTF(("src nicealign %u uniform %u nsegs %u\n", ++ nicealign, uniform, re->re_src.nsegs)); ++ if (re->re_src.nsegs > 1) { ++ re->re_desc.d_src = sc->sc_spalloc.dma_paddr + ++ ((caddr_t) sc->sc_spfree - (caddr_t) sc->sc_spring); ++ for (i = 0; i < re->re_src_nsegs; i++) { ++ /* NB: no need to check if there's space */ ++ pd = sc->sc_spfree; ++ if (++(sc->sc_spfree) == sc->sc_springtop) ++ sc->sc_spfree = sc->sc_spring; ++ ++ KASSERT((pd->pd_flags&3) == 0 || ++ (pd->pd_flags&3) == SAFE_PD_DONE, ++ ("bogus source particle descriptor; flags %x", ++ pd->pd_flags)); ++ pd->pd_addr = re->re_src_segs[i].ds_addr; ++ pd->pd_size = re->re_src_segs[i].ds_len; ++ pd->pd_flags = SAFE_PD_READY; ++ } ++ cmd0 |= SAFE_SA_CMD0_IGATHER; ++ } else { ++ /* ++ * No need for gather, reference the operand directly. ++ */ ++ re->re_desc.d_src = re->re_src_segs[0].ds_addr; ++ } ++ ++ if (enccrd == NULL && maccrd != NULL) { ++ /* ++ * Hash op; no destination needed. ++ */ ++ } else { ++ if (crp->crp_flags & (CRYPTO_F_IOV|CRYPTO_F_SKBUF)) { ++ if (!nicealign) { ++ safestats.st_iovmisaligned++; ++ err = EINVAL; ++ goto errout; ++ } ++ if (uniform != 1) { ++ device_printf(sc->sc_dev, "!uniform source\n"); ++ if (!uniform) { ++ /* ++ * There's no way to handle the DMA ++ * requirements with this uio. We ++ * could create a separate DMA area for ++ * the result and then copy it back, ++ * but for now we just bail and return ++ * an error. Note that uio requests ++ * > SAFE_MAX_DSIZE are handled because ++ * the DMA map and segment list for the ++ * destination wil result in a ++ * destination particle list that does ++ * the necessary scatter DMA. ++ */ ++ safestats.st_iovnotuniform++; ++ err = EINVAL; ++ goto errout; ++ } ++ } else ++ re->re_dst = re->re_src; ++ } else { ++ safestats.st_badflags++; ++ err = EINVAL; ++ goto errout; ++ } ++ ++ if (re->re_dst.nsegs > 1) { ++ re->re_desc.d_dst = sc->sc_dpalloc.dma_paddr + ++ ((caddr_t) sc->sc_dpfree - (caddr_t) sc->sc_dpring); ++ for (i = 0; i < re->re_dst_nsegs; i++) { ++ pd = sc->sc_dpfree; ++ KASSERT((pd->pd_flags&3) == 0 || ++ (pd->pd_flags&3) == SAFE_PD_DONE, ++ ("bogus dest particle descriptor; flags %x", ++ pd->pd_flags)); ++ if (++(sc->sc_dpfree) == sc->sc_dpringtop) ++ sc->sc_dpfree = sc->sc_dpring; ++ pd->pd_addr = re->re_dst_segs[i].ds_addr; ++ pd->pd_flags = SAFE_PD_READY; ++ } ++ cmd0 |= SAFE_SA_CMD0_OSCATTER; ++ } else { ++ /* ++ * No need for scatter, reference the operand directly. ++ */ ++ re->re_desc.d_dst = re->re_dst_segs[0].ds_addr; ++ } ++ } ++ ++ /* ++ * All done with setup; fillin the SA command words ++ * and the packet engine descriptor. The operation ++ * is now ready for submission to the hardware. ++ */ ++ sa->sa_cmd0 = cmd0 | SAFE_SA_CMD0_IPCI | SAFE_SA_CMD0_OPCI; ++ sa->sa_cmd1 = cmd1 ++ | (coffset << SAFE_SA_CMD1_OFFSET_S) ++ | SAFE_SA_CMD1_SAREV1 /* Rev 1 SA data structure */ ++ | SAFE_SA_CMD1_SRPCI ++ ; ++ /* ++ * NB: the order of writes is important here. In case the ++ * chip is scanning the ring because of an outstanding request ++ * it might nab this one too. In that case we need to make ++ * sure the setup is complete before we write the length ++ * field of the descriptor as it signals the descriptor is ++ * ready for processing. ++ */ ++ re->re_desc.d_csr = SAFE_PE_CSR_READY | SAFE_PE_CSR_SAPCI; ++ if (maccrd) ++ re->re_desc.d_csr |= SAFE_PE_CSR_LOADSA | SAFE_PE_CSR_HASHFINAL; ++ wmb(); ++ re->re_desc.d_len = oplen ++ | SAFE_PE_LEN_READY ++ | (bypass << SAFE_PE_LEN_BYPASS_S) ++ ; ++ ++ safestats.st_ipackets++; ++ safestats.st_ibytes += oplen; ++ ++ if (++(sc->sc_front) == sc->sc_ringtop) ++ sc->sc_front = sc->sc_ring; ++ ++ /* XXX honor batching */ ++ safe_feed(sc, re); ++ spin_unlock_irqrestore(&sc->sc_ringmtx, flags); ++ return (0); ++ ++errout: ++ if (re->re_src.map != re->re_dst.map) ++ pci_unmap_operand(sc, &re->re_dst); ++ if (re->re_src.map) ++ pci_unmap_operand(sc, &re->re_src); ++ spin_unlock_irqrestore(&sc->sc_ringmtx, flags); ++ if (err != ERESTART) { ++ crp->crp_etype = err; ++ crypto_done(crp); ++ } else { ++ sc->sc_needwakeup |= CRYPTO_SYMQ; ++ } ++ return (err); ++} ++ ++static void ++safe_callback(struct safe_softc *sc, struct safe_ringentry *re) ++{ ++ struct cryptop *crp = (struct cryptop *)re->re_crp; ++ struct cryptodesc *crd; ++ ++ DPRINTF(("%s()\n", __FUNCTION__)); ++ ++ safestats.st_opackets++; ++ safestats.st_obytes += re->re_dst.mapsize; ++ ++ if (re->re_desc.d_csr & SAFE_PE_CSR_STATUS) { ++ device_printf(sc->sc_dev, "csr 0x%x cmd0 0x%x cmd1 0x%x\n", ++ re->re_desc.d_csr, ++ re->re_sa.sa_cmd0, re->re_sa.sa_cmd1); ++ safestats.st_peoperr++; ++ crp->crp_etype = EIO; /* something more meaningful? */ ++ } ++ ++ if (re->re_dst.map != NULL && re->re_dst.map != re->re_src.map) ++ pci_unmap_operand(sc, &re->re_dst); ++ pci_unmap_operand(sc, &re->re_src); ++ ++ /* ++ * If result was written to a differet mbuf chain, swap ++ * it in as the return value and reclaim the original. ++ */ ++ if ((crp->crp_flags & CRYPTO_F_SKBUF) && re->re_src_skb != re->re_dst_skb) { ++ device_printf(sc->sc_dev, "no CRYPTO_F_SKBUF swapping support\n"); ++ /* kfree_skb(skb) */ ++ /* crp->crp_buf = (caddr_t)re->re_dst_skb */ ++ return; ++ } ++ ++ if (re->re_flags & SAFE_QFLAGS_COPYOUTIV) { ++ /* copy out IV for future use */ ++ for (crd = crp->crp_desc; crd; crd = crd->crd_next) { ++ int i; ++ int ivsize; ++ ++ if (crd->crd_alg == CRYPTO_DES_CBC || ++ crd->crd_alg == CRYPTO_3DES_CBC) { ++ ivsize = 2*sizeof(u_int32_t); ++ } else if (crd->crd_alg == CRYPTO_AES_CBC) { ++ ivsize = 4*sizeof(u_int32_t); ++ } else ++ continue; ++ crypto_copydata(crp->crp_flags, crp->crp_buf, ++ crd->crd_skip + crd->crd_len - ivsize, ivsize, ++ (caddr_t)sc->sc_sessions[re->re_sesn].ses_iv); ++ for (i = 0; ++ i < ivsize/sizeof(sc->sc_sessions[re->re_sesn].ses_iv[0]); ++ i++) ++ sc->sc_sessions[re->re_sesn].ses_iv[i] = ++ cpu_to_le32(sc->sc_sessions[re->re_sesn].ses_iv[i]); ++ break; ++ } ++ } ++ ++ if (re->re_flags & SAFE_QFLAGS_COPYOUTICV) { ++ /* copy out ICV result */ ++ for (crd = crp->crp_desc; crd; crd = crd->crd_next) { ++ if (!(crd->crd_alg == CRYPTO_MD5_HMAC || ++ crd->crd_alg == CRYPTO_SHA1_HMAC || ++ crd->crd_alg == CRYPTO_NULL_HMAC)) ++ continue; ++ if (crd->crd_alg == CRYPTO_SHA1_HMAC) { ++ /* ++ * SHA-1 ICV's are byte-swapped; fix 'em up ++ * before copy them to their destination. ++ */ ++ re->re_sastate.sa_saved_indigest[0] = ++ cpu_to_be32(re->re_sastate.sa_saved_indigest[0]); ++ re->re_sastate.sa_saved_indigest[1] = ++ cpu_to_be32(re->re_sastate.sa_saved_indigest[1]); ++ re->re_sastate.sa_saved_indigest[2] = ++ cpu_to_be32(re->re_sastate.sa_saved_indigest[2]); ++ } else { ++ re->re_sastate.sa_saved_indigest[0] = ++ cpu_to_le32(re->re_sastate.sa_saved_indigest[0]); ++ re->re_sastate.sa_saved_indigest[1] = ++ cpu_to_le32(re->re_sastate.sa_saved_indigest[1]); ++ re->re_sastate.sa_saved_indigest[2] = ++ cpu_to_le32(re->re_sastate.sa_saved_indigest[2]); ++ } ++ crypto_copyback(crp->crp_flags, crp->crp_buf, ++ crd->crd_inject, ++ sc->sc_sessions[re->re_sesn].ses_mlen, ++ (caddr_t)re->re_sastate.sa_saved_indigest); ++ break; ++ } ++ } ++ crypto_done(crp); ++} ++ ++ ++#if defined(CONFIG_OCF_RANDOMHARVEST) && !defined(SAFE_NO_RNG) ++#define SAFE_RNG_MAXWAIT 1000 ++ ++static void ++safe_rng_init(struct safe_softc *sc) ++{ ++ u_int32_t w, v; ++ int i; ++ ++ DPRINTF(("%s()\n", __FUNCTION__)); ++ ++ WRITE_REG(sc, SAFE_RNG_CTRL, 0); ++ /* use default value according to the manual */ ++ WRITE_REG(sc, SAFE_RNG_CNFG, 0x834); /* magic from SafeNet */ ++ WRITE_REG(sc, SAFE_RNG_ALM_CNT, 0); ++ ++ /* ++ * There is a bug in rev 1.0 of the 1140 that when the RNG ++ * is brought out of reset the ready status flag does not ++ * work until the RNG has finished its internal initialization. ++ * ++ * So in order to determine the device is through its ++ * initialization we must read the data register, using the ++ * status reg in the read in case it is initialized. Then read ++ * the data register until it changes from the first read. ++ * Once it changes read the data register until it changes ++ * again. At this time the RNG is considered initialized. ++ * This could take between 750ms - 1000ms in time. ++ */ ++ i = 0; ++ w = READ_REG(sc, SAFE_RNG_OUT); ++ do { ++ v = READ_REG(sc, SAFE_RNG_OUT); ++ if (v != w) { ++ w = v; ++ break; ++ } ++ DELAY(10); ++ } while (++i < SAFE_RNG_MAXWAIT); ++ ++ /* Wait Until data changes again */ ++ i = 0; ++ do { ++ v = READ_REG(sc, SAFE_RNG_OUT); ++ if (v != w) ++ break; ++ DELAY(10); ++ } while (++i < SAFE_RNG_MAXWAIT); ++} ++ ++static __inline void ++safe_rng_disable_short_cycle(struct safe_softc *sc) ++{ ++ DPRINTF(("%s()\n", __FUNCTION__)); ++ ++ WRITE_REG(sc, SAFE_RNG_CTRL, ++ READ_REG(sc, SAFE_RNG_CTRL) &~ SAFE_RNG_CTRL_SHORTEN); ++} ++ ++static __inline void ++safe_rng_enable_short_cycle(struct safe_softc *sc) ++{ ++ DPRINTF(("%s()\n", __FUNCTION__)); ++ ++ WRITE_REG(sc, SAFE_RNG_CTRL, ++ READ_REG(sc, SAFE_RNG_CTRL) | SAFE_RNG_CTRL_SHORTEN); ++} ++ ++static __inline u_int32_t ++safe_rng_read(struct safe_softc *sc) ++{ ++ int i; ++ ++ i = 0; ++ while (READ_REG(sc, SAFE_RNG_STAT) != 0 && ++i < SAFE_RNG_MAXWAIT) ++ ; ++ return READ_REG(sc, SAFE_RNG_OUT); ++} ++ ++static int ++safe_read_random(void *arg, u_int32_t *buf, int maxwords) ++{ ++ struct safe_softc *sc = (struct safe_softc *) arg; ++ int i, rc; ++ ++ DPRINTF(("%s()\n", __FUNCTION__)); ++ ++ safestats.st_rng++; ++ /* ++ * Fetch the next block of data. ++ */ ++ if (maxwords > safe_rngbufsize) ++ maxwords = safe_rngbufsize; ++ if (maxwords > SAFE_RNG_MAXBUFSIZ) ++ maxwords = SAFE_RNG_MAXBUFSIZ; ++retry: ++ /* read as much as we can */ ++ for (rc = 0; rc < maxwords; rc++) { ++ if (READ_REG(sc, SAFE_RNG_STAT) != 0) ++ break; ++ buf[rc] = READ_REG(sc, SAFE_RNG_OUT); ++ } ++ if (rc == 0) ++ return 0; ++ /* ++ * Check the comparator alarm count and reset the h/w if ++ * it exceeds our threshold. This guards against the ++ * hardware oscillators resonating with external signals. ++ */ ++ if (READ_REG(sc, SAFE_RNG_ALM_CNT) > safe_rngmaxalarm) { ++ u_int32_t freq_inc, w; ++ ++ DPRINTF(("%s: alarm count %u exceeds threshold %u\n", __func__, ++ (unsigned)READ_REG(sc, SAFE_RNG_ALM_CNT), safe_rngmaxalarm)); ++ safestats.st_rngalarm++; ++ safe_rng_enable_short_cycle(sc); ++ freq_inc = 18; ++ for (i = 0; i < 64; i++) { ++ w = READ_REG(sc, SAFE_RNG_CNFG); ++ freq_inc = ((w + freq_inc) & 0x3fL); ++ w = ((w & ~0x3fL) | freq_inc); ++ WRITE_REG(sc, SAFE_RNG_CNFG, w); ++ ++ WRITE_REG(sc, SAFE_RNG_ALM_CNT, 0); ++ ++ (void) safe_rng_read(sc); ++ DELAY(25); ++ ++ if (READ_REG(sc, SAFE_RNG_ALM_CNT) == 0) { ++ safe_rng_disable_short_cycle(sc); ++ goto retry; ++ } ++ freq_inc = 1; ++ } ++ safe_rng_disable_short_cycle(sc); ++ } else ++ WRITE_REG(sc, SAFE_RNG_ALM_CNT, 0); ++ ++ return(rc); ++} ++#endif /* defined(CONFIG_OCF_RANDOMHARVEST) && !defined(SAFE_NO_RNG) */ ++ ++ ++/* ++ * Resets the board. Values in the regesters are left as is ++ * from the reset (i.e. initial values are assigned elsewhere). ++ */ ++static void ++safe_reset_board(struct safe_softc *sc) ++{ ++ u_int32_t v; ++ /* ++ * Reset the device. The manual says no delay ++ * is needed between marking and clearing reset. ++ */ ++ DPRINTF(("%s()\n", __FUNCTION__)); ++ ++ v = READ_REG(sc, SAFE_PE_DMACFG) &~ ++ (SAFE_PE_DMACFG_PERESET | SAFE_PE_DMACFG_PDRRESET | ++ SAFE_PE_DMACFG_SGRESET); ++ WRITE_REG(sc, SAFE_PE_DMACFG, v ++ | SAFE_PE_DMACFG_PERESET ++ | SAFE_PE_DMACFG_PDRRESET ++ | SAFE_PE_DMACFG_SGRESET); ++ WRITE_REG(sc, SAFE_PE_DMACFG, v); ++} ++ ++/* ++ * Initialize registers we need to touch only once. ++ */ ++static void ++safe_init_board(struct safe_softc *sc) ++{ ++ u_int32_t v, dwords; ++ ++ DPRINTF(("%s()\n", __FUNCTION__)); ++ ++ v = READ_REG(sc, SAFE_PE_DMACFG); ++ v &=~ ( SAFE_PE_DMACFG_PEMODE ++ | SAFE_PE_DMACFG_FSENA /* failsafe enable */ ++ | SAFE_PE_DMACFG_GPRPCI /* gather ring on PCI */ ++ | SAFE_PE_DMACFG_SPRPCI /* scatter ring on PCI */ ++ | SAFE_PE_DMACFG_ESDESC /* endian-swap descriptors */ ++ | SAFE_PE_DMACFG_ESPDESC /* endian-swap part. desc's */ ++ | SAFE_PE_DMACFG_ESSA /* endian-swap SA's */ ++ | SAFE_PE_DMACFG_ESPACKET /* swap the packet data */ ++ ); ++ v |= SAFE_PE_DMACFG_FSENA /* failsafe enable */ ++ | SAFE_PE_DMACFG_GPRPCI /* gather ring on PCI */ ++ | SAFE_PE_DMACFG_SPRPCI /* scatter ring on PCI */ ++ | SAFE_PE_DMACFG_ESDESC /* endian-swap descriptors */ ++ | SAFE_PE_DMACFG_ESPDESC /* endian-swap part. desc's */ ++ | SAFE_PE_DMACFG_ESSA /* endian-swap SA's */ ++#if 0 ++ | SAFE_PE_DMACFG_ESPACKET /* swap the packet data */ ++#endif ++ ; ++ WRITE_REG(sc, SAFE_PE_DMACFG, v); ++ ++#ifdef __BIG_ENDIAN ++ /* tell the safenet that we are 4321 and not 1234 */ ++ WRITE_REG(sc, SAFE_ENDIAN, 0xe4e41b1b); ++#endif ++ ++ if (sc->sc_chiprev == SAFE_REV(1,0)) { ++ /* ++ * Avoid large PCI DMA transfers. Rev 1.0 has a bug where ++ * "target mode transfers" done while the chip is DMA'ing ++ * >1020 bytes cause the hardware to lockup. To avoid this ++ * we reduce the max PCI transfer size and use small source ++ * particle descriptors (<= 256 bytes). ++ */ ++ WRITE_REG(sc, SAFE_DMA_CFG, 256); ++ device_printf(sc->sc_dev, ++ "Reduce max DMA size to %u words for rev %u.%u WAR\n", ++ (unsigned) ((READ_REG(sc, SAFE_DMA_CFG)>>2) & 0xff), ++ (unsigned) SAFE_REV_MAJ(sc->sc_chiprev), ++ (unsigned) SAFE_REV_MIN(sc->sc_chiprev)); ++ sc->sc_max_dsize = 256; ++ } else { ++ sc->sc_max_dsize = SAFE_MAX_DSIZE; ++ } ++ ++ /* NB: operands+results are overlaid */ ++ WRITE_REG(sc, SAFE_PE_PDRBASE, sc->sc_ringalloc.dma_paddr); ++ WRITE_REG(sc, SAFE_PE_RDRBASE, sc->sc_ringalloc.dma_paddr); ++ /* ++ * Configure ring entry size and number of items in the ring. ++ */ ++ KASSERT((sizeof(struct safe_ringentry) % sizeof(u_int32_t)) == 0, ++ ("PE ring entry not 32-bit aligned!")); ++ dwords = sizeof(struct safe_ringentry) / sizeof(u_int32_t); ++ WRITE_REG(sc, SAFE_PE_RINGCFG, ++ (dwords << SAFE_PE_RINGCFG_OFFSET_S) | SAFE_MAX_NQUEUE); ++ WRITE_REG(sc, SAFE_PE_RINGPOLL, 0); /* disable polling */ ++ ++ WRITE_REG(sc, SAFE_PE_GRNGBASE, sc->sc_spalloc.dma_paddr); ++ WRITE_REG(sc, SAFE_PE_SRNGBASE, sc->sc_dpalloc.dma_paddr); ++ WRITE_REG(sc, SAFE_PE_PARTSIZE, ++ (SAFE_TOTAL_DPART<<16) | SAFE_TOTAL_SPART); ++ /* ++ * NB: destination particles are fixed size. We use ++ * an mbuf cluster and require all results go to ++ * clusters or smaller. ++ */ ++ WRITE_REG(sc, SAFE_PE_PARTCFG, sc->sc_max_dsize); ++ ++ /* it's now safe to enable PE mode, do it */ ++ WRITE_REG(sc, SAFE_PE_DMACFG, v | SAFE_PE_DMACFG_PEMODE); ++ ++ /* ++ * Configure hardware to use level-triggered interrupts and ++ * to interrupt after each descriptor is processed. ++ */ ++ WRITE_REG(sc, SAFE_HI_CFG, SAFE_HI_CFG_LEVEL); ++ WRITE_REG(sc, SAFE_HI_CLR, 0xffffffff); ++ WRITE_REG(sc, SAFE_HI_DESC_CNT, 1); ++ WRITE_REG(sc, SAFE_HI_MASK, SAFE_INT_PE_DDONE | SAFE_INT_PE_ERROR); ++} ++ ++ ++/* ++ * Clean up after a chip crash. ++ * It is assumed that the caller in splimp() ++ */ ++static void ++safe_cleanchip(struct safe_softc *sc) ++{ ++ DPRINTF(("%s()\n", __FUNCTION__)); ++ ++ if (sc->sc_nqchip != 0) { ++ struct safe_ringentry *re = sc->sc_back; ++ ++ while (re != sc->sc_front) { ++ if (re->re_desc.d_csr != 0) ++ safe_free_entry(sc, re); ++ if (++re == sc->sc_ringtop) ++ re = sc->sc_ring; ++ } ++ sc->sc_back = re; ++ sc->sc_nqchip = 0; ++ } ++} ++ ++/* ++ * free a safe_q ++ * It is assumed that the caller is within splimp(). ++ */ ++static int ++safe_free_entry(struct safe_softc *sc, struct safe_ringentry *re) ++{ ++ struct cryptop *crp; ++ ++ DPRINTF(("%s()\n", __FUNCTION__)); ++ ++ /* ++ * Free header MCR ++ */ ++ if ((re->re_dst_skb != NULL) && (re->re_src_skb != re->re_dst_skb)) ++#ifdef NOTYET ++ m_freem(re->re_dst_m); ++#else ++ printk("%s,%d: SKB not supported\n", __FILE__, __LINE__); ++#endif ++ ++ crp = (struct cryptop *)re->re_crp; ++ ++ re->re_desc.d_csr = 0; ++ ++ crp->crp_etype = EFAULT; ++ crypto_done(crp); ++ return(0); ++} ++ ++/* ++ * Routine to reset the chip and clean up. ++ * It is assumed that the caller is in splimp() ++ */ ++static void ++safe_totalreset(struct safe_softc *sc) ++{ ++ DPRINTF(("%s()\n", __FUNCTION__)); ++ ++ safe_reset_board(sc); ++ safe_init_board(sc); ++ safe_cleanchip(sc); ++} ++ ++/* ++ * Is the operand suitable aligned for direct DMA. Each ++ * segment must be aligned on a 32-bit boundary and all ++ * but the last segment must be a multiple of 4 bytes. ++ */ ++static int ++safe_dmamap_aligned(struct safe_softc *sc, const struct safe_operand *op) ++{ ++ int i; ++ ++ DPRINTF(("%s()\n", __FUNCTION__)); ++ ++ for (i = 0; i < op->nsegs; i++) { ++ if (op->segs[i].ds_addr & 3) ++ return (0); ++ if (i != (op->nsegs - 1) && (op->segs[i].ds_len & 3)) ++ return (0); ++ } ++ return (1); ++} ++ ++/* ++ * Is the operand suitable for direct DMA as the destination ++ * of an operation. The hardware requires that each ``particle'' ++ * but the last in an operation result have the same size. We ++ * fix that size at SAFE_MAX_DSIZE bytes. This routine returns ++ * 0 if some segment is not a multiple of of this size, 1 if all ++ * segments are exactly this size, or 2 if segments are at worst ++ * a multple of this size. ++ */ ++static int ++safe_dmamap_uniform(struct safe_softc *sc, const struct safe_operand *op) ++{ ++ int result = 1; ++ ++ DPRINTF(("%s()\n", __FUNCTION__)); ++ ++ if (op->nsegs > 0) { ++ int i; ++ ++ for (i = 0; i < op->nsegs-1; i++) { ++ if (op->segs[i].ds_len % sc->sc_max_dsize) ++ return (0); ++ if (op->segs[i].ds_len != sc->sc_max_dsize) ++ result = 2; ++ } ++ } ++ return (result); ++} ++ ++static int ++safe_kprocess(device_t dev, struct cryptkop *krp, int hint) ++{ ++ struct safe_softc *sc = device_get_softc(dev); ++ struct safe_pkq *q; ++ unsigned long flags; ++ ++ DPRINTF(("%s()\n", __FUNCTION__)); ++ ++ if (sc == NULL) { ++ krp->krp_status = EINVAL; ++ goto err; ++ } ++ ++ if (krp->krp_op != CRK_MOD_EXP) { ++ krp->krp_status = EOPNOTSUPP; ++ goto err; ++ } ++ ++ q = (struct safe_pkq *) kmalloc(sizeof(*q), GFP_KERNEL); ++ if (q == NULL) { ++ krp->krp_status = ENOMEM; ++ goto err; ++ } ++ memset(q, 0, sizeof(*q)); ++ q->pkq_krp = krp; ++ INIT_LIST_HEAD(&q->pkq_list); ++ ++ spin_lock_irqsave(&sc->sc_pkmtx, flags); ++ list_add_tail(&q->pkq_list, &sc->sc_pkq); ++ safe_kfeed(sc); ++ spin_unlock_irqrestore(&sc->sc_pkmtx, flags); ++ return (0); ++ ++err: ++ crypto_kdone(krp); ++ return (0); ++} ++ ++#define SAFE_CRK_PARAM_BASE 0 ++#define SAFE_CRK_PARAM_EXP 1 ++#define SAFE_CRK_PARAM_MOD 2 ++ ++static int ++safe_kstart(struct safe_softc *sc) ++{ ++ struct cryptkop *krp = sc->sc_pkq_cur->pkq_krp; ++ int exp_bits, mod_bits, base_bits; ++ u_int32_t op, a_off, b_off, c_off, d_off; ++ ++ DPRINTF(("%s()\n", __FUNCTION__)); ++ ++ if (krp->krp_iparams < 3 || krp->krp_oparams != 1) { ++ krp->krp_status = EINVAL; ++ return (1); ++ } ++ ++ base_bits = safe_ksigbits(sc, &krp->krp_param[SAFE_CRK_PARAM_BASE]); ++ if (base_bits > 2048) ++ goto too_big; ++ if (base_bits <= 0) /* 5. base not zero */ ++ goto too_small; ++ ++ exp_bits = safe_ksigbits(sc, &krp->krp_param[SAFE_CRK_PARAM_EXP]); ++ if (exp_bits > 2048) ++ goto too_big; ++ if (exp_bits <= 0) /* 1. exponent word length > 0 */ ++ goto too_small; /* 4. exponent not zero */ ++ ++ mod_bits = safe_ksigbits(sc, &krp->krp_param[SAFE_CRK_PARAM_MOD]); ++ if (mod_bits > 2048) ++ goto too_big; ++ if (mod_bits <= 32) /* 2. modulus word length > 1 */ ++ goto too_small; /* 8. MSW of modulus != zero */ ++ if (mod_bits < exp_bits) /* 3 modulus len >= exponent len */ ++ goto too_small; ++ if ((krp->krp_param[SAFE_CRK_PARAM_MOD].crp_p[0] & 1) == 0) ++ goto bad_domain; /* 6. modulus is odd */ ++ if (mod_bits > krp->krp_param[krp->krp_iparams].crp_nbits) ++ goto too_small; /* make sure result will fit */ ++ ++ /* 7. modulus > base */ ++ if (mod_bits < base_bits) ++ goto too_small; ++ if (mod_bits == base_bits) { ++ u_int8_t *basep, *modp; ++ int i; ++ ++ basep = krp->krp_param[SAFE_CRK_PARAM_BASE].crp_p + ++ ((base_bits + 7) / 8) - 1; ++ modp = krp->krp_param[SAFE_CRK_PARAM_MOD].crp_p + ++ ((mod_bits + 7) / 8) - 1; ++ ++ for (i = 0; i < (mod_bits + 7) / 8; i++, basep--, modp--) { ++ if (*modp < *basep) ++ goto too_small; ++ if (*modp > *basep) ++ break; ++ } ++ } ++ ++ /* And on the 9th step, he rested. */ ++ ++ WRITE_REG(sc, SAFE_PK_A_LEN, (exp_bits + 31) / 32); ++ WRITE_REG(sc, SAFE_PK_B_LEN, (mod_bits + 31) / 32); ++ if (mod_bits > 1024) { ++ op = SAFE_PK_FUNC_EXP4; ++ a_off = 0x000; ++ b_off = 0x100; ++ c_off = 0x200; ++ d_off = 0x300; ++ } else { ++ op = SAFE_PK_FUNC_EXP16; ++ a_off = 0x000; ++ b_off = 0x080; ++ c_off = 0x100; ++ d_off = 0x180; ++ } ++ sc->sc_pk_reslen = b_off - a_off; ++ sc->sc_pk_resoff = d_off; ++ ++ /* A is exponent, B is modulus, C is base, D is result */ ++ safe_kload_reg(sc, a_off, b_off - a_off, ++ &krp->krp_param[SAFE_CRK_PARAM_EXP]); ++ WRITE_REG(sc, SAFE_PK_A_ADDR, a_off >> 2); ++ safe_kload_reg(sc, b_off, b_off - a_off, ++ &krp->krp_param[SAFE_CRK_PARAM_MOD]); ++ WRITE_REG(sc, SAFE_PK_B_ADDR, b_off >> 2); ++ safe_kload_reg(sc, c_off, b_off - a_off, ++ &krp->krp_param[SAFE_CRK_PARAM_BASE]); ++ WRITE_REG(sc, SAFE_PK_C_ADDR, c_off >> 2); ++ WRITE_REG(sc, SAFE_PK_D_ADDR, d_off >> 2); ++ ++ WRITE_REG(sc, SAFE_PK_FUNC, op | SAFE_PK_FUNC_RUN); ++ ++ return (0); ++ ++too_big: ++ krp->krp_status = E2BIG; ++ return (1); ++too_small: ++ krp->krp_status = ERANGE; ++ return (1); ++bad_domain: ++ krp->krp_status = EDOM; ++ return (1); ++} ++ ++static int ++safe_ksigbits(struct safe_softc *sc, struct crparam *cr) ++{ ++ u_int plen = (cr->crp_nbits + 7) / 8; ++ int i, sig = plen * 8; ++ u_int8_t c, *p = cr->crp_p; ++ ++ DPRINTF(("%s()\n", __FUNCTION__)); ++ ++ for (i = plen - 1; i >= 0; i--) { ++ c = p[i]; ++ if (c != 0) { ++ while ((c & 0x80) == 0) { ++ sig--; ++ c <<= 1; ++ } ++ break; ++ } ++ sig -= 8; ++ } ++ return (sig); ++} ++ ++static void ++safe_kfeed(struct safe_softc *sc) ++{ ++ struct safe_pkq *q, *tmp; ++ ++ DPRINTF(("%s()\n", __FUNCTION__)); ++ ++ if (list_empty(&sc->sc_pkq) && sc->sc_pkq_cur == NULL) ++ return; ++ if (sc->sc_pkq_cur != NULL) ++ return; ++ list_for_each_entry_safe(q, tmp, &sc->sc_pkq, pkq_list) { ++ sc->sc_pkq_cur = q; ++ list_del(&q->pkq_list); ++ if (safe_kstart(sc) != 0) { ++ crypto_kdone(q->pkq_krp); ++ kfree(q); ++ sc->sc_pkq_cur = NULL; ++ } else { ++ /* op started, start polling */ ++ mod_timer(&sc->sc_pkto, jiffies + 1); ++ break; ++ } ++ } ++} ++ ++static void ++safe_kpoll(unsigned long arg) ++{ ++ struct safe_softc *sc = NULL; ++ struct safe_pkq *q; ++ struct crparam *res; ++ int i; ++ u_int32_t buf[64]; ++ unsigned long flags; ++ ++ DPRINTF(("%s()\n", __FUNCTION__)); ++ ++ if (arg >= SAFE_MAX_CHIPS) ++ return; ++ sc = safe_chip_idx[arg]; ++ if (!sc) { ++ DPRINTF(("%s() - bad callback\n", __FUNCTION__)); ++ return; ++ } ++ ++ spin_lock_irqsave(&sc->sc_pkmtx, flags); ++ if (sc->sc_pkq_cur == NULL) ++ goto out; ++ if (READ_REG(sc, SAFE_PK_FUNC) & SAFE_PK_FUNC_RUN) { ++ /* still running, check back later */ ++ mod_timer(&sc->sc_pkto, jiffies + 1); ++ goto out; ++ } ++ ++ q = sc->sc_pkq_cur; ++ res = &q->pkq_krp->krp_param[q->pkq_krp->krp_iparams]; ++ bzero(buf, sizeof(buf)); ++ bzero(res->crp_p, (res->crp_nbits + 7) / 8); ++ for (i = 0; i < sc->sc_pk_reslen >> 2; i++) ++ buf[i] = le32_to_cpu(READ_REG(sc, SAFE_PK_RAM_START + ++ sc->sc_pk_resoff + (i << 2))); ++ bcopy(buf, res->crp_p, (res->crp_nbits + 7) / 8); ++ /* ++ * reduce the bits that need copying if possible ++ */ ++ res->crp_nbits = min(res->crp_nbits,sc->sc_pk_reslen * 8); ++ res->crp_nbits = safe_ksigbits(sc, res); ++ ++ for (i = SAFE_PK_RAM_START; i < SAFE_PK_RAM_END; i += 4) ++ WRITE_REG(sc, i, 0); ++ ++ crypto_kdone(q->pkq_krp); ++ kfree(q); ++ sc->sc_pkq_cur = NULL; ++ ++ safe_kfeed(sc); ++out: ++ spin_unlock_irqrestore(&sc->sc_pkmtx, flags); ++} ++ ++static void ++safe_kload_reg(struct safe_softc *sc, u_int32_t off, u_int32_t len, ++ struct crparam *n) ++{ ++ u_int32_t buf[64], i; ++ ++ DPRINTF(("%s()\n", __FUNCTION__)); ++ ++ bzero(buf, sizeof(buf)); ++ bcopy(n->crp_p, buf, (n->crp_nbits + 7) / 8); ++ ++ for (i = 0; i < len >> 2; i++) ++ WRITE_REG(sc, SAFE_PK_RAM_START + off + (i << 2), ++ cpu_to_le32(buf[i])); ++} ++ ++#ifdef SAFE_DEBUG ++static void ++safe_dump_dmastatus(struct safe_softc *sc, const char *tag) ++{ ++ printf("%s: ENDIAN 0x%x SRC 0x%x DST 0x%x STAT 0x%x\n" ++ , tag ++ , READ_REG(sc, SAFE_DMA_ENDIAN) ++ , READ_REG(sc, SAFE_DMA_SRCADDR) ++ , READ_REG(sc, SAFE_DMA_DSTADDR) ++ , READ_REG(sc, SAFE_DMA_STAT) ++ ); ++} ++ ++static void ++safe_dump_intrstate(struct safe_softc *sc, const char *tag) ++{ ++ printf("%s: HI_CFG 0x%x HI_MASK 0x%x HI_DESC_CNT 0x%x HU_STAT 0x%x HM_STAT 0x%x\n" ++ , tag ++ , READ_REG(sc, SAFE_HI_CFG) ++ , READ_REG(sc, SAFE_HI_MASK) ++ , READ_REG(sc, SAFE_HI_DESC_CNT) ++ , READ_REG(sc, SAFE_HU_STAT) ++ , READ_REG(sc, SAFE_HM_STAT) ++ ); ++} ++ ++static void ++safe_dump_ringstate(struct safe_softc *sc, const char *tag) ++{ ++ u_int32_t estat = READ_REG(sc, SAFE_PE_ERNGSTAT); ++ ++ /* NB: assume caller has lock on ring */ ++ printf("%s: ERNGSTAT %x (next %u) back %lu front %lu\n", ++ tag, ++ estat, (estat >> SAFE_PE_ERNGSTAT_NEXT_S), ++ (unsigned long)(sc->sc_back - sc->sc_ring), ++ (unsigned long)(sc->sc_front - sc->sc_ring)); ++} ++ ++static void ++safe_dump_request(struct safe_softc *sc, const char* tag, struct safe_ringentry *re) ++{ ++ int ix, nsegs; ++ ++ ix = re - sc->sc_ring; ++ printf("%s: %p (%u): csr %x src %x dst %x sa %x len %x\n" ++ , tag ++ , re, ix ++ , re->re_desc.d_csr ++ , re->re_desc.d_src ++ , re->re_desc.d_dst ++ , re->re_desc.d_sa ++ , re->re_desc.d_len ++ ); ++ if (re->re_src.nsegs > 1) { ++ ix = (re->re_desc.d_src - sc->sc_spalloc.dma_paddr) / ++ sizeof(struct safe_pdesc); ++ for (nsegs = re->re_src.nsegs; nsegs; nsegs--) { ++ printf(" spd[%u] %p: %p size %u flags %x" ++ , ix, &sc->sc_spring[ix] ++ , (caddr_t)(uintptr_t) sc->sc_spring[ix].pd_addr ++ , sc->sc_spring[ix].pd_size ++ , sc->sc_spring[ix].pd_flags ++ ); ++ if (sc->sc_spring[ix].pd_size == 0) ++ printf(" (zero!)"); ++ printf("\n"); ++ if (++ix == SAFE_TOTAL_SPART) ++ ix = 0; ++ } ++ } ++ if (re->re_dst.nsegs > 1) { ++ ix = (re->re_desc.d_dst - sc->sc_dpalloc.dma_paddr) / ++ sizeof(struct safe_pdesc); ++ for (nsegs = re->re_dst.nsegs; nsegs; nsegs--) { ++ printf(" dpd[%u] %p: %p flags %x\n" ++ , ix, &sc->sc_dpring[ix] ++ , (caddr_t)(uintptr_t) sc->sc_dpring[ix].pd_addr ++ , sc->sc_dpring[ix].pd_flags ++ ); ++ if (++ix == SAFE_TOTAL_DPART) ++ ix = 0; ++ } ++ } ++ printf("sa: cmd0 %08x cmd1 %08x staterec %x\n", ++ re->re_sa.sa_cmd0, re->re_sa.sa_cmd1, re->re_sa.sa_staterec); ++ printf("sa: key %x %x %x %x %x %x %x %x\n" ++ , re->re_sa.sa_key[0] ++ , re->re_sa.sa_key[1] ++ , re->re_sa.sa_key[2] ++ , re->re_sa.sa_key[3] ++ , re->re_sa.sa_key[4] ++ , re->re_sa.sa_key[5] ++ , re->re_sa.sa_key[6] ++ , re->re_sa.sa_key[7] ++ ); ++ printf("sa: indigest %x %x %x %x %x\n" ++ , re->re_sa.sa_indigest[0] ++ , re->re_sa.sa_indigest[1] ++ , re->re_sa.sa_indigest[2] ++ , re->re_sa.sa_indigest[3] ++ , re->re_sa.sa_indigest[4] ++ ); ++ printf("sa: outdigest %x %x %x %x %x\n" ++ , re->re_sa.sa_outdigest[0] ++ , re->re_sa.sa_outdigest[1] ++ , re->re_sa.sa_outdigest[2] ++ , re->re_sa.sa_outdigest[3] ++ , re->re_sa.sa_outdigest[4] ++ ); ++ printf("sr: iv %x %x %x %x\n" ++ , re->re_sastate.sa_saved_iv[0] ++ , re->re_sastate.sa_saved_iv[1] ++ , re->re_sastate.sa_saved_iv[2] ++ , re->re_sastate.sa_saved_iv[3] ++ ); ++ printf("sr: hashbc %u indigest %x %x %x %x %x\n" ++ , re->re_sastate.sa_saved_hashbc ++ , re->re_sastate.sa_saved_indigest[0] ++ , re->re_sastate.sa_saved_indigest[1] ++ , re->re_sastate.sa_saved_indigest[2] ++ , re->re_sastate.sa_saved_indigest[3] ++ , re->re_sastate.sa_saved_indigest[4] ++ ); ++} ++ ++static void ++safe_dump_ring(struct safe_softc *sc, const char *tag) ++{ ++ unsigned long flags; ++ ++ spin_lock_irqsave(&sc->sc_ringmtx, flags); ++ printf("\nSafeNet Ring State:\n"); ++ safe_dump_intrstate(sc, tag); ++ safe_dump_dmastatus(sc, tag); ++ safe_dump_ringstate(sc, tag); ++ if (sc->sc_nqchip) { ++ struct safe_ringentry *re = sc->sc_back; ++ do { ++ safe_dump_request(sc, tag, re); ++ if (++re == sc->sc_ringtop) ++ re = sc->sc_ring; ++ } while (re != sc->sc_front); ++ } ++ spin_unlock_irqrestore(&sc->sc_ringmtx, flags); ++} ++#endif /* SAFE_DEBUG */ ++ ++ ++static int safe_probe(struct pci_dev *dev, const struct pci_device_id *ent) ++{ ++ struct safe_softc *sc = NULL; ++ u32 mem_start, mem_len, cmd; ++ int i, rc, devinfo; ++ dma_addr_t raddr; ++ static int num_chips = 0; ++ ++ DPRINTF(("%s()\n", __FUNCTION__)); ++ ++ if (pci_enable_device(dev) < 0) ++ return(-ENODEV); ++ ++ if (!dev->irq) { ++ printk("safe: found device with no IRQ assigned. check BIOS settings!"); ++ pci_disable_device(dev); ++ return(-ENODEV); ++ } ++ ++ if (pci_set_mwi(dev)) { ++ printk("safe: pci_set_mwi failed!"); ++ return(-ENODEV); ++ } ++ ++ sc = (struct safe_softc *) kmalloc(sizeof(*sc), GFP_KERNEL); ++ if (!sc) ++ return(-ENOMEM); ++ memset(sc, 0, sizeof(*sc)); ++ ++ softc_device_init(sc, "safe", num_chips, safe_methods); ++ ++ sc->sc_irq = -1; ++ sc->sc_cid = -1; ++ sc->sc_pcidev = dev; ++ if (num_chips < SAFE_MAX_CHIPS) { ++ safe_chip_idx[device_get_unit(sc->sc_dev)] = sc; ++ num_chips++; ++ } ++ ++ INIT_LIST_HEAD(&sc->sc_pkq); ++ spin_lock_init(&sc->sc_pkmtx); ++ ++ pci_set_drvdata(sc->sc_pcidev, sc); ++ ++ /* we read its hardware registers as memory */ ++ mem_start = pci_resource_start(sc->sc_pcidev, 0); ++ mem_len = pci_resource_len(sc->sc_pcidev, 0); ++ ++ sc->sc_base_addr = (ocf_iomem_t) ioremap(mem_start, mem_len); ++ if (!sc->sc_base_addr) { ++ device_printf(sc->sc_dev, "failed to ioremap 0x%x-0x%x\n", ++ mem_start, mem_start + mem_len - 1); ++ goto out; ++ } ++ ++ /* fix up the bus size */ ++ if (pci_set_dma_mask(sc->sc_pcidev, DMA_32BIT_MASK)) { ++ device_printf(sc->sc_dev, "No usable DMA configuration, aborting.\n"); ++ goto out; ++ } ++ if (pci_set_consistent_dma_mask(sc->sc_pcidev, DMA_32BIT_MASK)) { ++ device_printf(sc->sc_dev, "No usable consistent DMA configuration, aborting.\n"); ++ goto out; ++ } ++ ++ pci_set_master(sc->sc_pcidev); ++ ++ pci_read_config_dword(sc->sc_pcidev, PCI_COMMAND, &cmd); ++ ++ if (!(cmd & PCI_COMMAND_MEMORY)) { ++ device_printf(sc->sc_dev, "failed to enable memory mapping\n"); ++ goto out; ++ } ++ ++ if (!(cmd & PCI_COMMAND_MASTER)) { ++ device_printf(sc->sc_dev, "failed to enable bus mastering\n"); ++ goto out; ++ } ++ ++ rc = request_irq(dev->irq, safe_intr, IRQF_SHARED, "safe", sc); ++ if (rc) { ++ device_printf(sc->sc_dev, "failed to hook irq %d\n", sc->sc_irq); ++ goto out; ++ } ++ sc->sc_irq = dev->irq; ++ ++ sc->sc_chiprev = READ_REG(sc, SAFE_DEVINFO) & ++ (SAFE_DEVINFO_REV_MAJ | SAFE_DEVINFO_REV_MIN); ++ ++ /* ++ * Allocate packet engine descriptors. ++ */ ++ sc->sc_ringalloc.dma_vaddr = pci_alloc_consistent(sc->sc_pcidev, ++ SAFE_MAX_NQUEUE * sizeof (struct safe_ringentry), ++ &sc->sc_ringalloc.dma_paddr); ++ if (!sc->sc_ringalloc.dma_vaddr) { ++ device_printf(sc->sc_dev, "cannot allocate PE descriptor ring\n"); ++ goto out; ++ } ++ ++ /* ++ * Hookup the static portion of all our data structures. ++ */ ++ sc->sc_ring = (struct safe_ringentry *) sc->sc_ringalloc.dma_vaddr; ++ sc->sc_ringtop = sc->sc_ring + SAFE_MAX_NQUEUE; ++ sc->sc_front = sc->sc_ring; ++ sc->sc_back = sc->sc_ring; ++ raddr = sc->sc_ringalloc.dma_paddr; ++ bzero(sc->sc_ring, SAFE_MAX_NQUEUE * sizeof(struct safe_ringentry)); ++ for (i = 0; i < SAFE_MAX_NQUEUE; i++) { ++ struct safe_ringentry *re = &sc->sc_ring[i]; ++ ++ re->re_desc.d_sa = raddr + ++ offsetof(struct safe_ringentry, re_sa); ++ re->re_sa.sa_staterec = raddr + ++ offsetof(struct safe_ringentry, re_sastate); ++ ++ raddr += sizeof (struct safe_ringentry); ++ } ++ spin_lock_init(&sc->sc_ringmtx); ++ ++ /* ++ * Allocate scatter and gather particle descriptors. ++ */ ++ sc->sc_spalloc.dma_vaddr = pci_alloc_consistent(sc->sc_pcidev, ++ SAFE_TOTAL_SPART * sizeof (struct safe_pdesc), ++ &sc->sc_spalloc.dma_paddr); ++ if (!sc->sc_spalloc.dma_vaddr) { ++ device_printf(sc->sc_dev, "cannot allocate source particle descriptor ring\n"); ++ goto out; ++ } ++ sc->sc_spring = (struct safe_pdesc *) sc->sc_spalloc.dma_vaddr; ++ sc->sc_springtop = sc->sc_spring + SAFE_TOTAL_SPART; ++ sc->sc_spfree = sc->sc_spring; ++ bzero(sc->sc_spring, SAFE_TOTAL_SPART * sizeof(struct safe_pdesc)); ++ ++ sc->sc_dpalloc.dma_vaddr = pci_alloc_consistent(sc->sc_pcidev, ++ SAFE_TOTAL_DPART * sizeof (struct safe_pdesc), ++ &sc->sc_dpalloc.dma_paddr); ++ if (!sc->sc_dpalloc.dma_vaddr) { ++ device_printf(sc->sc_dev, "cannot allocate destination particle descriptor ring\n"); ++ goto out; ++ } ++ sc->sc_dpring = (struct safe_pdesc *) sc->sc_dpalloc.dma_vaddr; ++ sc->sc_dpringtop = sc->sc_dpring + SAFE_TOTAL_DPART; ++ sc->sc_dpfree = sc->sc_dpring; ++ bzero(sc->sc_dpring, SAFE_TOTAL_DPART * sizeof(struct safe_pdesc)); ++ ++ sc->sc_cid = crypto_get_driverid(softc_get_device(sc), CRYPTOCAP_F_HARDWARE); ++ if (sc->sc_cid < 0) { ++ device_printf(sc->sc_dev, "could not get crypto driver id\n"); ++ goto out; ++ } ++ ++ printf("%s:", device_get_nameunit(sc->sc_dev)); ++ ++ devinfo = READ_REG(sc, SAFE_DEVINFO); ++ if (devinfo & SAFE_DEVINFO_RNG) { ++ sc->sc_flags |= SAFE_FLAGS_RNG; ++ printf(" rng"); ++ } ++ if (devinfo & SAFE_DEVINFO_PKEY) { ++ printf(" key"); ++ sc->sc_flags |= SAFE_FLAGS_KEY; ++ crypto_kregister(sc->sc_cid, CRK_MOD_EXP, 0); ++#if 0 ++ crypto_kregister(sc->sc_cid, CRK_MOD_EXP_CRT, 0); ++#endif ++ init_timer(&sc->sc_pkto); ++ sc->sc_pkto.function = safe_kpoll; ++ sc->sc_pkto.data = (unsigned long) device_get_unit(sc->sc_dev); ++ } ++ if (devinfo & SAFE_DEVINFO_DES) { ++ printf(" des/3des"); ++ crypto_register(sc->sc_cid, CRYPTO_3DES_CBC, 0, 0); ++ crypto_register(sc->sc_cid, CRYPTO_DES_CBC, 0, 0); ++ } ++ if (devinfo & SAFE_DEVINFO_AES) { ++ printf(" aes"); ++ crypto_register(sc->sc_cid, CRYPTO_AES_CBC, 0, 0); ++ } ++ if (devinfo & SAFE_DEVINFO_MD5) { ++ printf(" md5"); ++ crypto_register(sc->sc_cid, CRYPTO_MD5_HMAC, 0, 0); ++ } ++ if (devinfo & SAFE_DEVINFO_SHA1) { ++ printf(" sha1"); ++ crypto_register(sc->sc_cid, CRYPTO_SHA1_HMAC, 0, 0); ++ } ++ printf(" null"); ++ crypto_register(sc->sc_cid, CRYPTO_NULL_CBC, 0, 0); ++ crypto_register(sc->sc_cid, CRYPTO_NULL_HMAC, 0, 0); ++ /* XXX other supported algorithms */ ++ printf("\n"); ++ ++ safe_reset_board(sc); /* reset h/w */ ++ safe_init_board(sc); /* init h/w */ ++ ++#if defined(CONFIG_OCF_RANDOMHARVEST) && !defined(SAFE_NO_RNG) ++ if (sc->sc_flags & SAFE_FLAGS_RNG) { ++ safe_rng_init(sc); ++ crypto_rregister(sc->sc_cid, safe_read_random, sc); ++ } ++#endif /* SAFE_NO_RNG */ ++ ++ return (0); ++ ++out: ++ if (sc->sc_cid >= 0) ++ crypto_unregister_all(sc->sc_cid); ++ if (sc->sc_irq != -1) ++ free_irq(sc->sc_irq, sc); ++ if (sc->sc_ringalloc.dma_vaddr) ++ pci_free_consistent(sc->sc_pcidev, ++ SAFE_MAX_NQUEUE * sizeof (struct safe_ringentry), ++ sc->sc_ringalloc.dma_vaddr, sc->sc_ringalloc.dma_paddr); ++ if (sc->sc_spalloc.dma_vaddr) ++ pci_free_consistent(sc->sc_pcidev, ++ SAFE_TOTAL_DPART * sizeof (struct safe_pdesc), ++ sc->sc_spalloc.dma_vaddr, sc->sc_spalloc.dma_paddr); ++ if (sc->sc_dpalloc.dma_vaddr) ++ pci_free_consistent(sc->sc_pcidev, ++ SAFE_TOTAL_DPART * sizeof (struct safe_pdesc), ++ sc->sc_dpalloc.dma_vaddr, sc->sc_dpalloc.dma_paddr); ++ kfree(sc); ++ return(-ENODEV); ++} ++ ++static void safe_remove(struct pci_dev *dev) ++{ ++ struct safe_softc *sc = pci_get_drvdata(dev); ++ ++ DPRINTF(("%s()\n", __FUNCTION__)); ++ ++ /* XXX wait/abort active ops */ ++ ++ WRITE_REG(sc, SAFE_HI_MASK, 0); /* disable interrupts */ ++ ++ del_timer_sync(&sc->sc_pkto); ++ ++ crypto_unregister_all(sc->sc_cid); ++ ++ safe_cleanchip(sc); ++ ++ if (sc->sc_irq != -1) ++ free_irq(sc->sc_irq, sc); ++ if (sc->sc_ringalloc.dma_vaddr) ++ pci_free_consistent(sc->sc_pcidev, ++ SAFE_MAX_NQUEUE * sizeof (struct safe_ringentry), ++ sc->sc_ringalloc.dma_vaddr, sc->sc_ringalloc.dma_paddr); ++ if (sc->sc_spalloc.dma_vaddr) ++ pci_free_consistent(sc->sc_pcidev, ++ SAFE_TOTAL_DPART * sizeof (struct safe_pdesc), ++ sc->sc_spalloc.dma_vaddr, sc->sc_spalloc.dma_paddr); ++ if (sc->sc_dpalloc.dma_vaddr) ++ pci_free_consistent(sc->sc_pcidev, ++ SAFE_TOTAL_DPART * sizeof (struct safe_pdesc), ++ sc->sc_dpalloc.dma_vaddr, sc->sc_dpalloc.dma_paddr); ++ sc->sc_irq = -1; ++ sc->sc_ringalloc.dma_vaddr = NULL; ++ sc->sc_spalloc.dma_vaddr = NULL; ++ sc->sc_dpalloc.dma_vaddr = NULL; ++} ++ ++static struct pci_device_id safe_pci_tbl[] = { ++ { PCI_VENDOR_SAFENET, PCI_PRODUCT_SAFEXCEL, ++ PCI_ANY_ID, PCI_ANY_ID, 0, 0, }, ++ { }, ++}; ++MODULE_DEVICE_TABLE(pci, safe_pci_tbl); ++ ++static struct pci_driver safe_driver = { ++ .name = "safe", ++ .id_table = safe_pci_tbl, ++ .probe = safe_probe, ++ .remove = safe_remove, ++ /* add PM stuff here one day */ ++}; ++ ++static int __init safe_init (void) ++{ ++ struct safe_softc *sc = NULL; ++ int rc; ++ ++ DPRINTF(("%s(%p)\n", __FUNCTION__, safe_init)); ++ ++ rc = pci_register_driver(&safe_driver); ++ pci_register_driver_compat(&safe_driver, rc); ++ ++ return rc; ++} ++ ++static void __exit safe_exit (void) ++{ ++ pci_unregister_driver(&safe_driver); ++} ++ ++module_init(safe_init); ++module_exit(safe_exit); ++ ++MODULE_LICENSE("BSD"); ++MODULE_AUTHOR("David McCullough <david_mccullough@securecomputing.com>"); ++MODULE_DESCRIPTION("OCF driver for safenet PCI crypto devices"); +diff -Nur linux-2.6.30.orig/crypto/ocf/safe/safereg.h linux-2.6.30/crypto/ocf/safe/safereg.h +--- linux-2.6.30.orig/crypto/ocf/safe/safereg.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/safe/safereg.h 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,421 @@ ++/*- ++ * Copyright (c) 2003 Sam Leffler, Errno Consulting ++ * Copyright (c) 2003 Global Technology Associates, Inc. ++ * All rights reserved. ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions ++ * are met: ++ * 1. Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * 2. Redistributions in binary form must reproduce the above copyright ++ * notice, this list of conditions and the following disclaimer in the ++ * documentation and/or other materials provided with the distribution. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND ++ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE ++ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ++ * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE ++ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL ++ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS ++ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) ++ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT ++ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY ++ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF ++ * SUCH DAMAGE. ++ * ++ * $FreeBSD: src/sys/dev/safe/safereg.h,v 1.1 2003/07/21 21:46:07 sam Exp $ ++ */ ++#ifndef _SAFE_SAFEREG_H_ ++#define _SAFE_SAFEREG_H_ ++ ++/* ++ * Register definitions for SafeNet SafeXcel-1141 crypto device. ++ * Definitions from revision 1.3 (Nov 6 2002) of the User's Manual. ++ */ ++ ++#define BS_BAR 0x10 /* DMA base address register */ ++#define BS_TRDY_TIMEOUT 0x40 /* TRDY timeout */ ++#define BS_RETRY_TIMEOUT 0x41 /* DMA retry timeout */ ++ ++#define PCI_VENDOR_SAFENET 0x16ae /* SafeNet, Inc. */ ++ ++/* SafeNet */ ++#define PCI_PRODUCT_SAFEXCEL 0x1141 /* 1141 */ ++ ++#define SAFE_PE_CSR 0x0000 /* Packet Enginge Ctrl/Status */ ++#define SAFE_PE_SRC 0x0004 /* Packet Engine Source */ ++#define SAFE_PE_DST 0x0008 /* Packet Engine Destination */ ++#define SAFE_PE_SA 0x000c /* Packet Engine SA */ ++#define SAFE_PE_LEN 0x0010 /* Packet Engine Length */ ++#define SAFE_PE_DMACFG 0x0040 /* Packet Engine DMA Configuration */ ++#define SAFE_PE_DMASTAT 0x0044 /* Packet Engine DMA Status */ ++#define SAFE_PE_PDRBASE 0x0048 /* Packet Engine Descriptor Ring Base */ ++#define SAFE_PE_RDRBASE 0x004c /* Packet Engine Result Ring Base */ ++#define SAFE_PE_RINGCFG 0x0050 /* Packet Engine Ring Configuration */ ++#define SAFE_PE_RINGPOLL 0x0054 /* Packet Engine Ring Poll */ ++#define SAFE_PE_IRNGSTAT 0x0058 /* Packet Engine Internal Ring Status */ ++#define SAFE_PE_ERNGSTAT 0x005c /* Packet Engine External Ring Status */ ++#define SAFE_PE_IOTHRESH 0x0060 /* Packet Engine I/O Threshold */ ++#define SAFE_PE_GRNGBASE 0x0064 /* Packet Engine Gather Ring Base */ ++#define SAFE_PE_SRNGBASE 0x0068 /* Packet Engine Scatter Ring Base */ ++#define SAFE_PE_PARTSIZE 0x006c /* Packet Engine Particlar Ring Size */ ++#define SAFE_PE_PARTCFG 0x0070 /* Packet Engine Particle Ring Config */ ++#define SAFE_CRYPTO_CTRL 0x0080 /* Crypto Control */ ++#define SAFE_DEVID 0x0084 /* Device ID */ ++#define SAFE_DEVINFO 0x0088 /* Device Info */ ++#define SAFE_HU_STAT 0x00a0 /* Host Unmasked Status */ ++#define SAFE_HM_STAT 0x00a4 /* Host Masked Status (read-only) */ ++#define SAFE_HI_CLR 0x00a4 /* Host Clear Interrupt (write-only) */ ++#define SAFE_HI_MASK 0x00a8 /* Host Mask Control */ ++#define SAFE_HI_CFG 0x00ac /* Interrupt Configuration */ ++#define SAFE_HI_RD_DESCR 0x00b4 /* Force Descriptor Read */ ++#define SAFE_HI_DESC_CNT 0x00b8 /* Host Descriptor Done Count */ ++#define SAFE_DMA_ENDIAN 0x00c0 /* Master Endian Status */ ++#define SAFE_DMA_SRCADDR 0x00c4 /* DMA Source Address Status */ ++#define SAFE_DMA_DSTADDR 0x00c8 /* DMA Destination Address Status */ ++#define SAFE_DMA_STAT 0x00cc /* DMA Current Status */ ++#define SAFE_DMA_CFG 0x00d4 /* DMA Configuration/Status */ ++#define SAFE_ENDIAN 0x00e0 /* Endian Configuration */ ++#define SAFE_PK_A_ADDR 0x0800 /* Public Key A Address */ ++#define SAFE_PK_B_ADDR 0x0804 /* Public Key B Address */ ++#define SAFE_PK_C_ADDR 0x0808 /* Public Key C Address */ ++#define SAFE_PK_D_ADDR 0x080c /* Public Key D Address */ ++#define SAFE_PK_A_LEN 0x0810 /* Public Key A Length */ ++#define SAFE_PK_B_LEN 0x0814 /* Public Key B Length */ ++#define SAFE_PK_SHIFT 0x0818 /* Public Key Shift */ ++#define SAFE_PK_FUNC 0x081c /* Public Key Function */ ++#define SAFE_PK_RAM_START 0x1000 /* Public Key RAM start address */ ++#define SAFE_PK_RAM_END 0x1fff /* Public Key RAM end address */ ++ ++#define SAFE_RNG_OUT 0x0100 /* RNG Output */ ++#define SAFE_RNG_STAT 0x0104 /* RNG Status */ ++#define SAFE_RNG_CTRL 0x0108 /* RNG Control */ ++#define SAFE_RNG_A 0x010c /* RNG A */ ++#define SAFE_RNG_B 0x0110 /* RNG B */ ++#define SAFE_RNG_X_LO 0x0114 /* RNG X [31:0] */ ++#define SAFE_RNG_X_MID 0x0118 /* RNG X [63:32] */ ++#define SAFE_RNG_X_HI 0x011c /* RNG X [80:64] */ ++#define SAFE_RNG_X_CNTR 0x0120 /* RNG Counter */ ++#define SAFE_RNG_ALM_CNT 0x0124 /* RNG Alarm Count */ ++#define SAFE_RNG_CNFG 0x0128 /* RNG Configuration */ ++#define SAFE_RNG_LFSR1_LO 0x012c /* RNG LFSR1 [31:0] */ ++#define SAFE_RNG_LFSR1_HI 0x0130 /* RNG LFSR1 [47:32] */ ++#define SAFE_RNG_LFSR2_LO 0x0134 /* RNG LFSR1 [31:0] */ ++#define SAFE_RNG_LFSR2_HI 0x0138 /* RNG LFSR1 [47:32] */ ++ ++#define SAFE_PE_CSR_READY 0x00000001 /* ready for processing */ ++#define SAFE_PE_CSR_DONE 0x00000002 /* h/w completed processing */ ++#define SAFE_PE_CSR_LOADSA 0x00000004 /* load SA digests */ ++#define SAFE_PE_CSR_HASHFINAL 0x00000010 /* do hash pad & write result */ ++#define SAFE_PE_CSR_SABUSID 0x000000c0 /* bus id for SA */ ++#define SAFE_PE_CSR_SAPCI 0x00000040 /* PCI bus id for SA */ ++#define SAFE_PE_CSR_NXTHDR 0x0000ff00 /* next hdr value for IPsec */ ++#define SAFE_PE_CSR_FPAD 0x0000ff00 /* fixed pad for basic ops */ ++#define SAFE_PE_CSR_STATUS 0x00ff0000 /* operation result status */ ++#define SAFE_PE_CSR_AUTH_FAIL 0x00010000 /* ICV mismatch (inbound) */ ++#define SAFE_PE_CSR_PAD_FAIL 0x00020000 /* pad verify fail (inbound) */ ++#define SAFE_PE_CSR_SEQ_FAIL 0x00040000 /* sequence number (inbound) */ ++#define SAFE_PE_CSR_XERROR 0x00080000 /* extended error follows */ ++#define SAFE_PE_CSR_XECODE 0x00f00000 /* extended error code */ ++#define SAFE_PE_CSR_XECODE_S 20 ++#define SAFE_PE_CSR_XECODE_BADCMD 0 /* invalid command */ ++#define SAFE_PE_CSR_XECODE_BADALG 1 /* invalid algorithm */ ++#define SAFE_PE_CSR_XECODE_ALGDIS 2 /* algorithm disabled */ ++#define SAFE_PE_CSR_XECODE_ZEROLEN 3 /* zero packet length */ ++#define SAFE_PE_CSR_XECODE_DMAERR 4 /* bus DMA error */ ++#define SAFE_PE_CSR_XECODE_PIPEABORT 5 /* secondary bus DMA error */ ++#define SAFE_PE_CSR_XECODE_BADSPI 6 /* IPsec SPI mismatch */ ++#define SAFE_PE_CSR_XECODE_TIMEOUT 10 /* failsafe timeout */ ++#define SAFE_PE_CSR_PAD 0xff000000 /* ESP padding control/status */ ++#define SAFE_PE_CSR_PAD_MIN 0x00000000 /* minimum IPsec padding */ ++#define SAFE_PE_CSR_PAD_16 0x08000000 /* pad to 16-byte boundary */ ++#define SAFE_PE_CSR_PAD_32 0x10000000 /* pad to 32-byte boundary */ ++#define SAFE_PE_CSR_PAD_64 0x20000000 /* pad to 64-byte boundary */ ++#define SAFE_PE_CSR_PAD_128 0x40000000 /* pad to 128-byte boundary */ ++#define SAFE_PE_CSR_PAD_256 0x80000000 /* pad to 256-byte boundary */ ++ ++/* ++ * Check the CSR to see if the PE has returned ownership to ++ * the host. Note that before processing a descriptor this ++ * must be done followed by a check of the SAFE_PE_LEN register ++ * status bits to avoid premature processing of a descriptor ++ * on its way back to the host. ++ */ ++#define SAFE_PE_CSR_IS_DONE(_csr) \ ++ (((_csr) & (SAFE_PE_CSR_READY | SAFE_PE_CSR_DONE)) == SAFE_PE_CSR_DONE) ++ ++#define SAFE_PE_LEN_LENGTH 0x000fffff /* total length (bytes) */ ++#define SAFE_PE_LEN_READY 0x00400000 /* ready for processing */ ++#define SAFE_PE_LEN_DONE 0x00800000 /* h/w completed processing */ ++#define SAFE_PE_LEN_BYPASS 0xff000000 /* bypass offset (bytes) */ ++#define SAFE_PE_LEN_BYPASS_S 24 ++ ++#define SAFE_PE_LEN_IS_DONE(_len) \ ++ (((_len) & (SAFE_PE_LEN_READY | SAFE_PE_LEN_DONE)) == SAFE_PE_LEN_DONE) ++ ++/* NB: these apply to HU_STAT, HM_STAT, HI_CLR, and HI_MASK */ ++#define SAFE_INT_PE_CDONE 0x00000002 /* PE context done */ ++#define SAFE_INT_PE_DDONE 0x00000008 /* PE descriptor done */ ++#define SAFE_INT_PE_ERROR 0x00000010 /* PE error */ ++#define SAFE_INT_PE_ODONE 0x00000020 /* PE operation done */ ++ ++#define SAFE_HI_CFG_PULSE 0x00000001 /* use pulse interrupt */ ++#define SAFE_HI_CFG_LEVEL 0x00000000 /* use level interrupt */ ++#define SAFE_HI_CFG_AUTOCLR 0x00000002 /* auto-clear pulse interrupt */ ++ ++#define SAFE_ENDIAN_PASS 0x000000e4 /* straight pass-thru */ ++#define SAFE_ENDIAN_SWAB 0x0000001b /* swap bytes in 32-bit word */ ++ ++#define SAFE_PE_DMACFG_PERESET 0x00000001 /* reset packet engine */ ++#define SAFE_PE_DMACFG_PDRRESET 0x00000002 /* reset PDR counters/ptrs */ ++#define SAFE_PE_DMACFG_SGRESET 0x00000004 /* reset scatter/gather cache */ ++#define SAFE_PE_DMACFG_FSENA 0x00000008 /* enable failsafe reset */ ++#define SAFE_PE_DMACFG_PEMODE 0x00000100 /* packet engine mode */ ++#define SAFE_PE_DMACFG_SAPREC 0x00000200 /* SA precedes packet */ ++#define SAFE_PE_DMACFG_PKFOLL 0x00000400 /* packet follows descriptor */ ++#define SAFE_PE_DMACFG_GPRBID 0x00003000 /* gather particle ring busid */ ++#define SAFE_PE_DMACFG_GPRPCI 0x00001000 /* PCI gather particle ring */ ++#define SAFE_PE_DMACFG_SPRBID 0x0000c000 /* scatter part. ring busid */ ++#define SAFE_PE_DMACFG_SPRPCI 0x00004000 /* PCI scatter part. ring */ ++#define SAFE_PE_DMACFG_ESDESC 0x00010000 /* endian swap descriptors */ ++#define SAFE_PE_DMACFG_ESSA 0x00020000 /* endian swap SA data */ ++#define SAFE_PE_DMACFG_ESPACKET 0x00040000 /* endian swap packet data */ ++#define SAFE_PE_DMACFG_ESPDESC 0x00080000 /* endian swap particle desc. */ ++#define SAFE_PE_DMACFG_NOPDRUP 0x00100000 /* supp. PDR ownership update */ ++#define SAFE_PD_EDMACFG_PCIMODE 0x01000000 /* PCI target mode */ ++ ++#define SAFE_PE_DMASTAT_PEIDONE 0x00000001 /* PE core input done */ ++#define SAFE_PE_DMASTAT_PEODONE 0x00000002 /* PE core output done */ ++#define SAFE_PE_DMASTAT_ENCDONE 0x00000004 /* encryption done */ ++#define SAFE_PE_DMASTAT_IHDONE 0x00000008 /* inner hash done */ ++#define SAFE_PE_DMASTAT_OHDONE 0x00000010 /* outer hash (HMAC) done */ ++#define SAFE_PE_DMASTAT_PADFLT 0x00000020 /* crypto pad fault */ ++#define SAFE_PE_DMASTAT_ICVFLT 0x00000040 /* ICV fault */ ++#define SAFE_PE_DMASTAT_SPIMIS 0x00000080 /* SPI mismatch */ ++#define SAFE_PE_DMASTAT_CRYPTO 0x00000100 /* crypto engine timeout */ ++#define SAFE_PE_DMASTAT_CQACT 0x00000200 /* command queue active */ ++#define SAFE_PE_DMASTAT_IRACT 0x00000400 /* input request active */ ++#define SAFE_PE_DMASTAT_ORACT 0x00000800 /* output request active */ ++#define SAFE_PE_DMASTAT_PEISIZE 0x003ff000 /* PE input size:32-bit words */ ++#define SAFE_PE_DMASTAT_PEOSIZE 0xffc00000 /* PE out. size:32-bit words */ ++ ++#define SAFE_PE_RINGCFG_SIZE 0x000003ff /* ring size (descriptors) */ ++#define SAFE_PE_RINGCFG_OFFSET 0xffff0000 /* offset btw desc's (dwords) */ ++#define SAFE_PE_RINGCFG_OFFSET_S 16 ++ ++#define SAFE_PE_RINGPOLL_POLL 0x00000fff /* polling frequency/divisor */ ++#define SAFE_PE_RINGPOLL_RETRY 0x03ff0000 /* polling frequency/divisor */ ++#define SAFE_PE_RINGPOLL_CONT 0x80000000 /* continuously poll */ ++ ++#define SAFE_PE_IRNGSTAT_CQAVAIL 0x00000001 /* command queue available */ ++ ++#define SAFE_PE_ERNGSTAT_NEXT 0x03ff0000 /* index of next packet desc. */ ++#define SAFE_PE_ERNGSTAT_NEXT_S 16 ++ ++#define SAFE_PE_IOTHRESH_INPUT 0x000003ff /* input threshold (dwords) */ ++#define SAFE_PE_IOTHRESH_OUTPUT 0x03ff0000 /* output threshold (dwords) */ ++ ++#define SAFE_PE_PARTCFG_SIZE 0x0000ffff /* scatter particle size */ ++#define SAFE_PE_PARTCFG_GBURST 0x00030000 /* gather particle burst */ ++#define SAFE_PE_PARTCFG_GBURST_2 0x00000000 ++#define SAFE_PE_PARTCFG_GBURST_4 0x00010000 ++#define SAFE_PE_PARTCFG_GBURST_8 0x00020000 ++#define SAFE_PE_PARTCFG_GBURST_16 0x00030000 ++#define SAFE_PE_PARTCFG_SBURST 0x000c0000 /* scatter particle burst */ ++#define SAFE_PE_PARTCFG_SBURST_2 0x00000000 ++#define SAFE_PE_PARTCFG_SBURST_4 0x00040000 ++#define SAFE_PE_PARTCFG_SBURST_8 0x00080000 ++#define SAFE_PE_PARTCFG_SBURST_16 0x000c0000 ++ ++#define SAFE_PE_PARTSIZE_SCAT 0xffff0000 /* scatter particle ring size */ ++#define SAFE_PE_PARTSIZE_GATH 0x0000ffff /* gather particle ring size */ ++ ++#define SAFE_CRYPTO_CTRL_3DES 0x00000001 /* enable 3DES support */ ++#define SAFE_CRYPTO_CTRL_PKEY 0x00010000 /* enable public key support */ ++#define SAFE_CRYPTO_CTRL_RNG 0x00020000 /* enable RNG support */ ++ ++#define SAFE_DEVINFO_REV_MIN 0x0000000f /* minor rev for chip */ ++#define SAFE_DEVINFO_REV_MAJ 0x000000f0 /* major rev for chip */ ++#define SAFE_DEVINFO_REV_MAJ_S 4 ++#define SAFE_DEVINFO_DES 0x00000100 /* DES/3DES support present */ ++#define SAFE_DEVINFO_ARC4 0x00000200 /* ARC4 support present */ ++#define SAFE_DEVINFO_AES 0x00000400 /* AES support present */ ++#define SAFE_DEVINFO_MD5 0x00001000 /* MD5 support present */ ++#define SAFE_DEVINFO_SHA1 0x00002000 /* SHA-1 support present */ ++#define SAFE_DEVINFO_RIPEMD 0x00004000 /* RIPEMD support present */ ++#define SAFE_DEVINFO_DEFLATE 0x00010000 /* Deflate support present */ ++#define SAFE_DEVINFO_SARAM 0x00100000 /* on-chip SA RAM present */ ++#define SAFE_DEVINFO_EMIBUS 0x00200000 /* EMI bus present */ ++#define SAFE_DEVINFO_PKEY 0x00400000 /* public key support present */ ++#define SAFE_DEVINFO_RNG 0x00800000 /* RNG present */ ++ ++#define SAFE_REV(_maj, _min) (((_maj) << SAFE_DEVINFO_REV_MAJ_S) | (_min)) ++#define SAFE_REV_MAJ(_chiprev) \ ++ (((_chiprev) & SAFE_DEVINFO_REV_MAJ) >> SAFE_DEVINFO_REV_MAJ_S) ++#define SAFE_REV_MIN(_chiprev) ((_chiprev) & SAFE_DEVINFO_REV_MIN) ++ ++#define SAFE_PK_FUNC_MULT 0x00000001 /* Multiply function */ ++#define SAFE_PK_FUNC_SQUARE 0x00000004 /* Square function */ ++#define SAFE_PK_FUNC_ADD 0x00000010 /* Add function */ ++#define SAFE_PK_FUNC_SUB 0x00000020 /* Subtract function */ ++#define SAFE_PK_FUNC_LSHIFT 0x00000040 /* Left-shift function */ ++#define SAFE_PK_FUNC_RSHIFT 0x00000080 /* Right-shift function */ ++#define SAFE_PK_FUNC_DIV 0x00000100 /* Divide function */ ++#define SAFE_PK_FUNC_CMP 0x00000400 /* Compare function */ ++#define SAFE_PK_FUNC_COPY 0x00000800 /* Copy function */ ++#define SAFE_PK_FUNC_EXP16 0x00002000 /* Exponentiate (4-bit ACT) */ ++#define SAFE_PK_FUNC_EXP4 0x00004000 /* Exponentiate (2-bit ACT) */ ++#define SAFE_PK_FUNC_RUN 0x00008000 /* start/status */ ++ ++#define SAFE_RNG_STAT_BUSY 0x00000001 /* busy, data not valid */ ++ ++#define SAFE_RNG_CTRL_PRE_LFSR 0x00000001 /* enable output pre-LFSR */ ++#define SAFE_RNG_CTRL_TST_MODE 0x00000002 /* enable test mode */ ++#define SAFE_RNG_CTRL_TST_RUN 0x00000004 /* start test state machine */ ++#define SAFE_RNG_CTRL_ENA_RING1 0x00000008 /* test entropy oscillator #1 */ ++#define SAFE_RNG_CTRL_ENA_RING2 0x00000010 /* test entropy oscillator #2 */ ++#define SAFE_RNG_CTRL_DIS_ALARM 0x00000020 /* disable RNG alarm reports */ ++#define SAFE_RNG_CTRL_TST_CLOCK 0x00000040 /* enable test clock */ ++#define SAFE_RNG_CTRL_SHORTEN 0x00000080 /* shorten state timers */ ++#define SAFE_RNG_CTRL_TST_ALARM 0x00000100 /* simulate alarm state */ ++#define SAFE_RNG_CTRL_RST_LFSR 0x00000200 /* reset LFSR */ ++ ++/* ++ * Packet engine descriptor. Note that d_csr is a copy of the ++ * SAFE_PE_CSR register and all definitions apply, and d_len ++ * is a copy of the SAFE_PE_LEN register and all definitions apply. ++ * d_src and d_len may point directly to contiguous data or to a ++ * list of ``particle descriptors'' when using scatter/gather i/o. ++ */ ++struct safe_desc { ++ u_int32_t d_csr; /* per-packet control/status */ ++ u_int32_t d_src; /* source address */ ++ u_int32_t d_dst; /* destination address */ ++ u_int32_t d_sa; /* SA address */ ++ u_int32_t d_len; /* length, bypass, status */ ++}; ++ ++/* ++ * Scatter/Gather particle descriptor. ++ * ++ * NB: scatter descriptors do not specify a size; this is fixed ++ * by the setting of the SAFE_PE_PARTCFG register. ++ */ ++struct safe_pdesc { ++ u_int32_t pd_addr; /* particle address */ ++#ifdef __BIG_ENDIAN ++ u_int16_t pd_flags; /* control word */ ++ u_int16_t pd_size; /* particle size (bytes) */ ++#else ++ u_int16_t pd_flags; /* control word */ ++ u_int16_t pd_size; /* particle size (bytes) */ ++#endif ++}; ++ ++#define SAFE_PD_READY 0x0001 /* ready for processing */ ++#define SAFE_PD_DONE 0x0002 /* h/w completed processing */ ++ ++/* ++ * Security Association (SA) Record (Rev 1). One of these is ++ * required for each operation processed by the packet engine. ++ */ ++struct safe_sarec { ++ u_int32_t sa_cmd0; ++ u_int32_t sa_cmd1; ++ u_int32_t sa_resv0; ++ u_int32_t sa_resv1; ++ u_int32_t sa_key[8]; /* DES/3DES/AES key */ ++ u_int32_t sa_indigest[5]; /* inner digest */ ++ u_int32_t sa_outdigest[5]; /* outer digest */ ++ u_int32_t sa_spi; /* SPI */ ++ u_int32_t sa_seqnum; /* sequence number */ ++ u_int32_t sa_seqmask[2]; /* sequence number mask */ ++ u_int32_t sa_resv2; ++ u_int32_t sa_staterec; /* address of state record */ ++ u_int32_t sa_resv3[2]; ++ u_int32_t sa_samgmt0; /* SA management field 0 */ ++ u_int32_t sa_samgmt1; /* SA management field 0 */ ++}; ++ ++#define SAFE_SA_CMD0_OP 0x00000007 /* operation code */ ++#define SAFE_SA_CMD0_OP_CRYPT 0x00000000 /* encrypt/decrypt (basic) */ ++#define SAFE_SA_CMD0_OP_BOTH 0x00000001 /* encrypt-hash/hash-decrypto */ ++#define SAFE_SA_CMD0_OP_HASH 0x00000003 /* hash (outbound-only) */ ++#define SAFE_SA_CMD0_OP_ESP 0x00000000 /* ESP in/out (proto) */ ++#define SAFE_SA_CMD0_OP_AH 0x00000001 /* AH in/out (proto) */ ++#define SAFE_SA_CMD0_INBOUND 0x00000008 /* inbound operation */ ++#define SAFE_SA_CMD0_OUTBOUND 0x00000000 /* outbound operation */ ++#define SAFE_SA_CMD0_GROUP 0x00000030 /* operation group */ ++#define SAFE_SA_CMD0_BASIC 0x00000000 /* basic operation */ ++#define SAFE_SA_CMD0_PROTO 0x00000010 /* protocol/packet operation */ ++#define SAFE_SA_CMD0_BUNDLE 0x00000020 /* bundled operation (resvd) */ ++#define SAFE_SA_CMD0_PAD 0x000000c0 /* crypto pad method */ ++#define SAFE_SA_CMD0_PAD_IPSEC 0x00000000 /* IPsec padding */ ++#define SAFE_SA_CMD0_PAD_PKCS7 0x00000040 /* PKCS#7 padding */ ++#define SAFE_SA_CMD0_PAD_CONS 0x00000080 /* constant padding */ ++#define SAFE_SA_CMD0_PAD_ZERO 0x000000c0 /* zero padding */ ++#define SAFE_SA_CMD0_CRYPT_ALG 0x00000f00 /* symmetric crypto algorithm */ ++#define SAFE_SA_CMD0_DES 0x00000000 /* DES crypto algorithm */ ++#define SAFE_SA_CMD0_3DES 0x00000100 /* 3DES crypto algorithm */ ++#define SAFE_SA_CMD0_AES 0x00000300 /* AES crypto algorithm */ ++#define SAFE_SA_CMD0_CRYPT_NULL 0x00000f00 /* null crypto algorithm */ ++#define SAFE_SA_CMD0_HASH_ALG 0x0000f000 /* hash algorithm */ ++#define SAFE_SA_CMD0_MD5 0x00000000 /* MD5 hash algorithm */ ++#define SAFE_SA_CMD0_SHA1 0x00001000 /* SHA-1 hash algorithm */ ++#define SAFE_SA_CMD0_HASH_NULL 0x0000f000 /* null hash algorithm */ ++#define SAFE_SA_CMD0_HDR_PROC 0x00080000 /* header processing */ ++#define SAFE_SA_CMD0_IBUSID 0x00300000 /* input bus id */ ++#define SAFE_SA_CMD0_IPCI 0x00100000 /* PCI input bus id */ ++#define SAFE_SA_CMD0_OBUSID 0x00c00000 /* output bus id */ ++#define SAFE_SA_CMD0_OPCI 0x00400000 /* PCI output bus id */ ++#define SAFE_SA_CMD0_IVLD 0x03000000 /* IV loading */ ++#define SAFE_SA_CMD0_IVLD_NONE 0x00000000 /* IV no load (reuse) */ ++#define SAFE_SA_CMD0_IVLD_IBUF 0x01000000 /* IV load from input buffer */ ++#define SAFE_SA_CMD0_IVLD_STATE 0x02000000 /* IV load from state */ ++#define SAFE_SA_CMD0_HSLD 0x0c000000 /* hash state loading */ ++#define SAFE_SA_CMD0_HSLD_SA 0x00000000 /* hash state load from SA */ ++#define SAFE_SA_CMD0_HSLD_STATE 0x08000000 /* hash state load from state */ ++#define SAFE_SA_CMD0_HSLD_NONE 0x0c000000 /* hash state no load */ ++#define SAFE_SA_CMD0_SAVEIV 0x10000000 /* save IV */ ++#define SAFE_SA_CMD0_SAVEHASH 0x20000000 /* save hash state */ ++#define SAFE_SA_CMD0_IGATHER 0x40000000 /* input gather */ ++#define SAFE_SA_CMD0_OSCATTER 0x80000000 /* output scatter */ ++ ++#define SAFE_SA_CMD1_HDRCOPY 0x00000002 /* copy header to output */ ++#define SAFE_SA_CMD1_PAYCOPY 0x00000004 /* copy payload to output */ ++#define SAFE_SA_CMD1_PADCOPY 0x00000008 /* copy pad to output */ ++#define SAFE_SA_CMD1_IPV4 0x00000000 /* IPv4 protocol */ ++#define SAFE_SA_CMD1_IPV6 0x00000010 /* IPv6 protocol */ ++#define SAFE_SA_CMD1_MUTABLE 0x00000020 /* mutable bit processing */ ++#define SAFE_SA_CMD1_SRBUSID 0x000000c0 /* state record bus id */ ++#define SAFE_SA_CMD1_SRPCI 0x00000040 /* state record from PCI */ ++#define SAFE_SA_CMD1_CRMODE 0x00000300 /* crypto mode */ ++#define SAFE_SA_CMD1_ECB 0x00000000 /* ECB crypto mode */ ++#define SAFE_SA_CMD1_CBC 0x00000100 /* CBC crypto mode */ ++#define SAFE_SA_CMD1_OFB 0x00000200 /* OFB crypto mode */ ++#define SAFE_SA_CMD1_CFB 0x00000300 /* CFB crypto mode */ ++#define SAFE_SA_CMD1_CRFEEDBACK 0x00000c00 /* crypto feedback mode */ ++#define SAFE_SA_CMD1_64BIT 0x00000000 /* 64-bit crypto feedback */ ++#define SAFE_SA_CMD1_8BIT 0x00000400 /* 8-bit crypto feedback */ ++#define SAFE_SA_CMD1_1BIT 0x00000800 /* 1-bit crypto feedback */ ++#define SAFE_SA_CMD1_128BIT 0x00000c00 /* 128-bit crypto feedback */ ++#define SAFE_SA_CMD1_OPTIONS 0x00001000 /* HMAC/options mutable bit */ ++#define SAFE_SA_CMD1_HMAC SAFE_SA_CMD1_OPTIONS ++#define SAFE_SA_CMD1_SAREV1 0x00008000 /* SA Revision 1 */ ++#define SAFE_SA_CMD1_OFFSET 0x00ff0000 /* hash/crypto offset(dwords) */ ++#define SAFE_SA_CMD1_OFFSET_S 16 ++#define SAFE_SA_CMD1_AESKEYLEN 0x0f000000 /* AES key length */ ++#define SAFE_SA_CMD1_AES128 0x02000000 /* 128-bit AES key */ ++#define SAFE_SA_CMD1_AES192 0x03000000 /* 192-bit AES key */ ++#define SAFE_SA_CMD1_AES256 0x04000000 /* 256-bit AES key */ ++ ++/* ++ * Security Associate State Record (Rev 1). ++ */ ++struct safe_sastate { ++ u_int32_t sa_saved_iv[4]; /* saved IV (DES/3DES/AES) */ ++ u_int32_t sa_saved_hashbc; /* saved hash byte count */ ++ u_int32_t sa_saved_indigest[5]; /* saved inner digest */ ++}; ++#endif /* _SAFE_SAFEREG_H_ */ +diff -Nur linux-2.6.30.orig/crypto/ocf/safe/safevar.h linux-2.6.30/crypto/ocf/safe/safevar.h +--- linux-2.6.30.orig/crypto/ocf/safe/safevar.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/safe/safevar.h 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,230 @@ ++/*- ++ * The linux port of this code done by David McCullough ++ * Copyright (C) 2004-2007 David McCullough <david_mccullough@securecomputing.com> ++ * The license and original author are listed below. ++ * ++ * Copyright (c) 2003 Sam Leffler, Errno Consulting ++ * Copyright (c) 2003 Global Technology Associates, Inc. ++ * All rights reserved. ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions ++ * are met: ++ * 1. Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * 2. Redistributions in binary form must reproduce the above copyright ++ * notice, this list of conditions and the following disclaimer in the ++ * documentation and/or other materials provided with the distribution. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND ++ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE ++ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ++ * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE ++ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL ++ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS ++ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) ++ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT ++ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY ++ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF ++ * SUCH DAMAGE. ++ * ++ * $FreeBSD: src/sys/dev/safe/safevar.h,v 1.2 2006/05/17 18:34:26 pjd Exp $ ++ */ ++#ifndef _SAFE_SAFEVAR_H_ ++#define _SAFE_SAFEVAR_H_ ++ ++/* Maximum queue length */ ++#ifndef SAFE_MAX_NQUEUE ++#define SAFE_MAX_NQUEUE 60 ++#endif ++ ++#define SAFE_MAX_PART 64 /* Maximum scatter/gather depth */ ++#define SAFE_DMA_BOUNDARY 0 /* No boundary for source DMA ops */ ++#define SAFE_MAX_DSIZE 2048 /* MCLBYTES Fixed scatter particle size */ ++#define SAFE_MAX_SSIZE 0x0ffff /* Maximum gather particle size */ ++#define SAFE_MAX_DMA 0xfffff /* Maximum PE operand size (20 bits) */ ++/* total src+dst particle descriptors */ ++#define SAFE_TOTAL_DPART (SAFE_MAX_NQUEUE * SAFE_MAX_PART) ++#define SAFE_TOTAL_SPART (SAFE_MAX_NQUEUE * SAFE_MAX_PART) ++ ++#define SAFE_RNG_MAXBUFSIZ 128 /* 32-bit words */ ++ ++#define SAFE_CARD(sid) (((sid) & 0xf0000000) >> 28) ++#define SAFE_SESSION(sid) ( (sid) & 0x0fffffff) ++#define SAFE_SID(crd, sesn) (((crd) << 28) | ((sesn) & 0x0fffffff)) ++ ++#define SAFE_DEF_RTY 0xff /* PCI Retry Timeout */ ++#define SAFE_DEF_TOUT 0xff /* PCI TRDY Timeout */ ++#define SAFE_DEF_CACHELINE 0x01 /* Cache Line setting */ ++ ++#ifdef __KERNEL__ ++/* ++ * State associated with the allocation of each chunk ++ * of memory setup for DMA. ++ */ ++struct safe_dma_alloc { ++ dma_addr_t dma_paddr; ++ void *dma_vaddr; ++}; ++ ++/* ++ * Cryptographic operand state. One of these exists for each ++ * source and destination operand passed in from the crypto ++ * subsystem. When possible source and destination operands ++ * refer to the same memory. More often they are distinct. ++ * We track the virtual address of each operand as well as ++ * where each is mapped for DMA. ++ */ ++struct safe_operand { ++ union { ++ struct sk_buff *skb; ++ struct uio *io; ++ } u; ++ void *map; ++ int mapsize; /* total number of bytes in segs */ ++ struct { ++ dma_addr_t ds_addr; ++ int ds_len; ++ int ds_tlen; ++ } segs[SAFE_MAX_PART]; ++ int nsegs; ++}; ++ ++/* ++ * Packet engine ring entry and cryptographic operation state. ++ * The packet engine requires a ring of descriptors that contain ++ * pointers to various cryptographic state. However the ring ++ * configuration register allows you to specify an arbitrary size ++ * for ring entries. We use this feature to collect most of the ++ * state for each cryptographic request into one spot. Other than ++ * ring entries only the ``particle descriptors'' (scatter/gather ++ * lists) and the actual operand data are kept separate. The ++ * particle descriptors must also be organized in rings. The ++ * operand data can be located aribtrarily (modulo alignment constraints). ++ * ++ * Note that the descriptor ring is mapped onto the PCI bus so ++ * the hardware can DMA data. This means the entire ring must be ++ * contiguous. ++ */ ++struct safe_ringentry { ++ struct safe_desc re_desc; /* command descriptor */ ++ struct safe_sarec re_sa; /* SA record */ ++ struct safe_sastate re_sastate; /* SA state record */ ++ ++ struct cryptop *re_crp; /* crypto operation */ ++ ++ struct safe_operand re_src; /* source operand */ ++ struct safe_operand re_dst; /* destination operand */ ++ ++ int re_sesn; /* crypto session ID */ ++ int re_flags; ++#define SAFE_QFLAGS_COPYOUTIV 0x1 /* copy back on completion */ ++#define SAFE_QFLAGS_COPYOUTICV 0x2 /* copy back on completion */ ++}; ++ ++#define re_src_skb re_src.u.skb ++#define re_src_io re_src.u.io ++#define re_src_map re_src.map ++#define re_src_nsegs re_src.nsegs ++#define re_src_segs re_src.segs ++#define re_src_mapsize re_src.mapsize ++ ++#define re_dst_skb re_dst.u.skb ++#define re_dst_io re_dst.u.io ++#define re_dst_map re_dst.map ++#define re_dst_nsegs re_dst.nsegs ++#define re_dst_segs re_dst.segs ++#define re_dst_mapsize re_dst.mapsize ++ ++struct rndstate_test; ++ ++struct safe_session { ++ u_int32_t ses_used; ++ u_int32_t ses_klen; /* key length in bits */ ++ u_int32_t ses_key[8]; /* DES/3DES/AES key */ ++ u_int32_t ses_mlen; /* hmac length in bytes */ ++ u_int32_t ses_hminner[5]; /* hmac inner state */ ++ u_int32_t ses_hmouter[5]; /* hmac outer state */ ++ u_int32_t ses_iv[4]; /* DES/3DES/AES iv */ ++}; ++ ++struct safe_pkq { ++ struct list_head pkq_list; ++ struct cryptkop *pkq_krp; ++}; ++ ++struct safe_softc { ++ softc_device_decl sc_dev; ++ u32 sc_irq; ++ ++ struct pci_dev *sc_pcidev; ++ ocf_iomem_t sc_base_addr; ++ ++ u_int sc_chiprev; /* major/minor chip revision */ ++ int sc_flags; /* device specific flags */ ++#define SAFE_FLAGS_KEY 0x01 /* has key accelerator */ ++#define SAFE_FLAGS_RNG 0x02 /* hardware rng */ ++ int sc_suspended; ++ int sc_needwakeup; /* notify crypto layer */ ++ int32_t sc_cid; /* crypto tag */ ++ ++ struct safe_dma_alloc sc_ringalloc; /* PE ring allocation state */ ++ struct safe_ringentry *sc_ring; /* PE ring */ ++ struct safe_ringentry *sc_ringtop; /* PE ring top */ ++ struct safe_ringentry *sc_front; /* next free entry */ ++ struct safe_ringentry *sc_back; /* next pending entry */ ++ int sc_nqchip; /* # passed to chip */ ++ spinlock_t sc_ringmtx; /* PE ring lock */ ++ struct safe_pdesc *sc_spring; /* src particle ring */ ++ struct safe_pdesc *sc_springtop; /* src particle ring top */ ++ struct safe_pdesc *sc_spfree; /* next free src particle */ ++ struct safe_dma_alloc sc_spalloc; /* src particle ring state */ ++ struct safe_pdesc *sc_dpring; /* dest particle ring */ ++ struct safe_pdesc *sc_dpringtop; /* dest particle ring top */ ++ struct safe_pdesc *sc_dpfree; /* next free dest particle */ ++ struct safe_dma_alloc sc_dpalloc; /* dst particle ring state */ ++ int sc_nsessions; /* # of sessions */ ++ struct safe_session *sc_sessions; /* sessions */ ++ ++ struct timer_list sc_pkto; /* PK polling */ ++ spinlock_t sc_pkmtx; /* PK lock */ ++ struct list_head sc_pkq; /* queue of PK requests */ ++ struct safe_pkq *sc_pkq_cur; /* current processing request */ ++ u_int32_t sc_pk_reslen, sc_pk_resoff; ++ ++ int sc_max_dsize; /* maximum safe DMA size */ ++}; ++#endif /* __KERNEL__ */ ++ ++struct safe_stats { ++ u_int64_t st_ibytes; ++ u_int64_t st_obytes; ++ u_int32_t st_ipackets; ++ u_int32_t st_opackets; ++ u_int32_t st_invalid; /* invalid argument */ ++ u_int32_t st_badsession; /* invalid session id */ ++ u_int32_t st_badflags; /* flags indicate !(mbuf | uio) */ ++ u_int32_t st_nodesc; /* op submitted w/o descriptors */ ++ u_int32_t st_badalg; /* unsupported algorithm */ ++ u_int32_t st_ringfull; /* PE descriptor ring full */ ++ u_int32_t st_peoperr; /* PE marked error */ ++ u_int32_t st_dmaerr; /* PE DMA error */ ++ u_int32_t st_bypasstoobig; /* bypass > 96 bytes */ ++ u_int32_t st_skipmismatch; /* enc part begins before auth part */ ++ u_int32_t st_lenmismatch; /* enc length different auth length */ ++ u_int32_t st_coffmisaligned; /* crypto offset not 32-bit aligned */ ++ u_int32_t st_cofftoobig; /* crypto offset > 255 words */ ++ u_int32_t st_iovmisaligned; /* iov op not aligned */ ++ u_int32_t st_iovnotuniform; /* iov op not suitable */ ++ u_int32_t st_unaligned; /* unaligned src caused copy */ ++ u_int32_t st_notuniform; /* non-uniform src caused copy */ ++ u_int32_t st_nomap; /* bus_dmamap_create failed */ ++ u_int32_t st_noload; /* bus_dmamap_load_* failed */ ++ u_int32_t st_nombuf; /* MGET* failed */ ++ u_int32_t st_nomcl; /* MCLGET* failed */ ++ u_int32_t st_maxqchip; /* max mcr1 ops out for processing */ ++ u_int32_t st_rng; /* RNG requests */ ++ u_int32_t st_rngalarm; /* RNG alarm requests */ ++ u_int32_t st_noicvcopy; /* ICV data copies suppressed */ ++}; ++#endif /* _SAFE_SAFEVAR_H_ */ +diff -Nur linux-2.6.30.orig/crypto/ocf/safe/sha1.c linux-2.6.30/crypto/ocf/safe/sha1.c +--- linux-2.6.30.orig/crypto/ocf/safe/sha1.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/safe/sha1.c 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,279 @@ ++/* $KAME: sha1.c,v 1.5 2000/11/08 06:13:08 itojun Exp $ */ ++/* ++ * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. ++ * All rights reserved. ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions ++ * are met: ++ * 1. Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * 2. Redistributions in binary form must reproduce the above copyright ++ * notice, this list of conditions and the following disclaimer in the ++ * documentation and/or other materials provided with the distribution. ++ * 3. Neither the name of the project nor the names of its contributors ++ * may be used to endorse or promote products derived from this software ++ * without specific prior written permission. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND ++ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE ++ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ++ * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE ++ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL ++ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS ++ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) ++ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT ++ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY ++ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF ++ * SUCH DAMAGE. ++ */ ++ ++/* ++ * FIPS pub 180-1: Secure Hash Algorithm (SHA-1) ++ * based on: http://csrc.nist.gov/fips/fip180-1.txt ++ * implemented by Jun-ichiro itojun Itoh <itojun@itojun.org> ++ */ ++ ++#if 0 ++#include <sys/cdefs.h> ++__FBSDID("$FreeBSD: src/sys/crypto/sha1.c,v 1.9 2003/06/10 21:36:57 obrien Exp $"); ++ ++#include <sys/types.h> ++#include <sys/cdefs.h> ++#include <sys/time.h> ++#include <sys/systm.h> ++ ++#include <crypto/sha1.h> ++#endif ++ ++/* sanity check */ ++#if BYTE_ORDER != BIG_ENDIAN ++# if BYTE_ORDER != LITTLE_ENDIAN ++# define unsupported 1 ++# endif ++#endif ++ ++#ifndef unsupported ++ ++/* constant table */ ++static u_int32_t _K[] = { 0x5a827999, 0x6ed9eba1, 0x8f1bbcdc, 0xca62c1d6 }; ++#define K(t) _K[(t) / 20] ++ ++#define F0(b, c, d) (((b) & (c)) | ((~(b)) & (d))) ++#define F1(b, c, d) (((b) ^ (c)) ^ (d)) ++#define F2(b, c, d) (((b) & (c)) | ((b) & (d)) | ((c) & (d))) ++#define F3(b, c, d) (((b) ^ (c)) ^ (d)) ++ ++#define S(n, x) (((x) << (n)) | ((x) >> (32 - n))) ++ ++#undef H ++#define H(n) (ctxt->h.b32[(n)]) ++#define COUNT (ctxt->count) ++#define BCOUNT (ctxt->c.b64[0] / 8) ++#define W(n) (ctxt->m.b32[(n)]) ++ ++#define PUTBYTE(x) { \ ++ ctxt->m.b8[(COUNT % 64)] = (x); \ ++ COUNT++; \ ++ COUNT %= 64; \ ++ ctxt->c.b64[0] += 8; \ ++ if (COUNT % 64 == 0) \ ++ sha1_step(ctxt); \ ++ } ++ ++#define PUTPAD(x) { \ ++ ctxt->m.b8[(COUNT % 64)] = (x); \ ++ COUNT++; \ ++ COUNT %= 64; \ ++ if (COUNT % 64 == 0) \ ++ sha1_step(ctxt); \ ++ } ++ ++static void sha1_step(struct sha1_ctxt *); ++ ++static void ++sha1_step(ctxt) ++ struct sha1_ctxt *ctxt; ++{ ++ u_int32_t a, b, c, d, e; ++ size_t t, s; ++ u_int32_t tmp; ++ ++#if BYTE_ORDER == LITTLE_ENDIAN ++ struct sha1_ctxt tctxt; ++ bcopy(&ctxt->m.b8[0], &tctxt.m.b8[0], 64); ++ ctxt->m.b8[0] = tctxt.m.b8[3]; ctxt->m.b8[1] = tctxt.m.b8[2]; ++ ctxt->m.b8[2] = tctxt.m.b8[1]; ctxt->m.b8[3] = tctxt.m.b8[0]; ++ ctxt->m.b8[4] = tctxt.m.b8[7]; ctxt->m.b8[5] = tctxt.m.b8[6]; ++ ctxt->m.b8[6] = tctxt.m.b8[5]; ctxt->m.b8[7] = tctxt.m.b8[4]; ++ ctxt->m.b8[8] = tctxt.m.b8[11]; ctxt->m.b8[9] = tctxt.m.b8[10]; ++ ctxt->m.b8[10] = tctxt.m.b8[9]; ctxt->m.b8[11] = tctxt.m.b8[8]; ++ ctxt->m.b8[12] = tctxt.m.b8[15]; ctxt->m.b8[13] = tctxt.m.b8[14]; ++ ctxt->m.b8[14] = tctxt.m.b8[13]; ctxt->m.b8[15] = tctxt.m.b8[12]; ++ ctxt->m.b8[16] = tctxt.m.b8[19]; ctxt->m.b8[17] = tctxt.m.b8[18]; ++ ctxt->m.b8[18] = tctxt.m.b8[17]; ctxt->m.b8[19] = tctxt.m.b8[16]; ++ ctxt->m.b8[20] = tctxt.m.b8[23]; ctxt->m.b8[21] = tctxt.m.b8[22]; ++ ctxt->m.b8[22] = tctxt.m.b8[21]; ctxt->m.b8[23] = tctxt.m.b8[20]; ++ ctxt->m.b8[24] = tctxt.m.b8[27]; ctxt->m.b8[25] = tctxt.m.b8[26]; ++ ctxt->m.b8[26] = tctxt.m.b8[25]; ctxt->m.b8[27] = tctxt.m.b8[24]; ++ ctxt->m.b8[28] = tctxt.m.b8[31]; ctxt->m.b8[29] = tctxt.m.b8[30]; ++ ctxt->m.b8[30] = tctxt.m.b8[29]; ctxt->m.b8[31] = tctxt.m.b8[28]; ++ ctxt->m.b8[32] = tctxt.m.b8[35]; ctxt->m.b8[33] = tctxt.m.b8[34]; ++ ctxt->m.b8[34] = tctxt.m.b8[33]; ctxt->m.b8[35] = tctxt.m.b8[32]; ++ ctxt->m.b8[36] = tctxt.m.b8[39]; ctxt->m.b8[37] = tctxt.m.b8[38]; ++ ctxt->m.b8[38] = tctxt.m.b8[37]; ctxt->m.b8[39] = tctxt.m.b8[36]; ++ ctxt->m.b8[40] = tctxt.m.b8[43]; ctxt->m.b8[41] = tctxt.m.b8[42]; ++ ctxt->m.b8[42] = tctxt.m.b8[41]; ctxt->m.b8[43] = tctxt.m.b8[40]; ++ ctxt->m.b8[44] = tctxt.m.b8[47]; ctxt->m.b8[45] = tctxt.m.b8[46]; ++ ctxt->m.b8[46] = tctxt.m.b8[45]; ctxt->m.b8[47] = tctxt.m.b8[44]; ++ ctxt->m.b8[48] = tctxt.m.b8[51]; ctxt->m.b8[49] = tctxt.m.b8[50]; ++ ctxt->m.b8[50] = tctxt.m.b8[49]; ctxt->m.b8[51] = tctxt.m.b8[48]; ++ ctxt->m.b8[52] = tctxt.m.b8[55]; ctxt->m.b8[53] = tctxt.m.b8[54]; ++ ctxt->m.b8[54] = tctxt.m.b8[53]; ctxt->m.b8[55] = tctxt.m.b8[52]; ++ ctxt->m.b8[56] = tctxt.m.b8[59]; ctxt->m.b8[57] = tctxt.m.b8[58]; ++ ctxt->m.b8[58] = tctxt.m.b8[57]; ctxt->m.b8[59] = tctxt.m.b8[56]; ++ ctxt->m.b8[60] = tctxt.m.b8[63]; ctxt->m.b8[61] = tctxt.m.b8[62]; ++ ctxt->m.b8[62] = tctxt.m.b8[61]; ctxt->m.b8[63] = tctxt.m.b8[60]; ++#endif ++ ++ a = H(0); b = H(1); c = H(2); d = H(3); e = H(4); ++ ++ for (t = 0; t < 20; t++) { ++ s = t & 0x0f; ++ if (t >= 16) { ++ W(s) = S(1, W((s+13) & 0x0f) ^ W((s+8) & 0x0f) ^ W((s+2) & 0x0f) ^ W(s)); ++ } ++ tmp = S(5, a) + F0(b, c, d) + e + W(s) + K(t); ++ e = d; d = c; c = S(30, b); b = a; a = tmp; ++ } ++ for (t = 20; t < 40; t++) { ++ s = t & 0x0f; ++ W(s) = S(1, W((s+13) & 0x0f) ^ W((s+8) & 0x0f) ^ W((s+2) & 0x0f) ^ W(s)); ++ tmp = S(5, a) + F1(b, c, d) + e + W(s) + K(t); ++ e = d; d = c; c = S(30, b); b = a; a = tmp; ++ } ++ for (t = 40; t < 60; t++) { ++ s = t & 0x0f; ++ W(s) = S(1, W((s+13) & 0x0f) ^ W((s+8) & 0x0f) ^ W((s+2) & 0x0f) ^ W(s)); ++ tmp = S(5, a) + F2(b, c, d) + e + W(s) + K(t); ++ e = d; d = c; c = S(30, b); b = a; a = tmp; ++ } ++ for (t = 60; t < 80; t++) { ++ s = t & 0x0f; ++ W(s) = S(1, W((s+13) & 0x0f) ^ W((s+8) & 0x0f) ^ W((s+2) & 0x0f) ^ W(s)); ++ tmp = S(5, a) + F3(b, c, d) + e + W(s) + K(t); ++ e = d; d = c; c = S(30, b); b = a; a = tmp; ++ } ++ ++ H(0) = H(0) + a; ++ H(1) = H(1) + b; ++ H(2) = H(2) + c; ++ H(3) = H(3) + d; ++ H(4) = H(4) + e; ++ ++ bzero(&ctxt->m.b8[0], 64); ++} ++ ++/*------------------------------------------------------------*/ ++ ++void ++sha1_init(ctxt) ++ struct sha1_ctxt *ctxt; ++{ ++ bzero(ctxt, sizeof(struct sha1_ctxt)); ++ H(0) = 0x67452301; ++ H(1) = 0xefcdab89; ++ H(2) = 0x98badcfe; ++ H(3) = 0x10325476; ++ H(4) = 0xc3d2e1f0; ++} ++ ++void ++sha1_pad(ctxt) ++ struct sha1_ctxt *ctxt; ++{ ++ size_t padlen; /*pad length in bytes*/ ++ size_t padstart; ++ ++ PUTPAD(0x80); ++ ++ padstart = COUNT % 64; ++ padlen = 64 - padstart; ++ if (padlen < 8) { ++ bzero(&ctxt->m.b8[padstart], padlen); ++ COUNT += padlen; ++ COUNT %= 64; ++ sha1_step(ctxt); ++ padstart = COUNT % 64; /* should be 0 */ ++ padlen = 64 - padstart; /* should be 64 */ ++ } ++ bzero(&ctxt->m.b8[padstart], padlen - 8); ++ COUNT += (padlen - 8); ++ COUNT %= 64; ++#if BYTE_ORDER == BIG_ENDIAN ++ PUTPAD(ctxt->c.b8[0]); PUTPAD(ctxt->c.b8[1]); ++ PUTPAD(ctxt->c.b8[2]); PUTPAD(ctxt->c.b8[3]); ++ PUTPAD(ctxt->c.b8[4]); PUTPAD(ctxt->c.b8[5]); ++ PUTPAD(ctxt->c.b8[6]); PUTPAD(ctxt->c.b8[7]); ++#else ++ PUTPAD(ctxt->c.b8[7]); PUTPAD(ctxt->c.b8[6]); ++ PUTPAD(ctxt->c.b8[5]); PUTPAD(ctxt->c.b8[4]); ++ PUTPAD(ctxt->c.b8[3]); PUTPAD(ctxt->c.b8[2]); ++ PUTPAD(ctxt->c.b8[1]); PUTPAD(ctxt->c.b8[0]); ++#endif ++} ++ ++void ++sha1_loop(ctxt, input, len) ++ struct sha1_ctxt *ctxt; ++ const u_int8_t *input; ++ size_t len; ++{ ++ size_t gaplen; ++ size_t gapstart; ++ size_t off; ++ size_t copysiz; ++ ++ off = 0; ++ ++ while (off < len) { ++ gapstart = COUNT % 64; ++ gaplen = 64 - gapstart; ++ ++ copysiz = (gaplen < len - off) ? gaplen : len - off; ++ bcopy(&input[off], &ctxt->m.b8[gapstart], copysiz); ++ COUNT += copysiz; ++ COUNT %= 64; ++ ctxt->c.b64[0] += copysiz * 8; ++ if (COUNT % 64 == 0) ++ sha1_step(ctxt); ++ off += copysiz; ++ } ++} ++ ++void ++sha1_result(ctxt, digest0) ++ struct sha1_ctxt *ctxt; ++ caddr_t digest0; ++{ ++ u_int8_t *digest; ++ ++ digest = (u_int8_t *)digest0; ++ sha1_pad(ctxt); ++#if BYTE_ORDER == BIG_ENDIAN ++ bcopy(&ctxt->h.b8[0], digest, 20); ++#else ++ digest[0] = ctxt->h.b8[3]; digest[1] = ctxt->h.b8[2]; ++ digest[2] = ctxt->h.b8[1]; digest[3] = ctxt->h.b8[0]; ++ digest[4] = ctxt->h.b8[7]; digest[5] = ctxt->h.b8[6]; ++ digest[6] = ctxt->h.b8[5]; digest[7] = ctxt->h.b8[4]; ++ digest[8] = ctxt->h.b8[11]; digest[9] = ctxt->h.b8[10]; ++ digest[10] = ctxt->h.b8[9]; digest[11] = ctxt->h.b8[8]; ++ digest[12] = ctxt->h.b8[15]; digest[13] = ctxt->h.b8[14]; ++ digest[14] = ctxt->h.b8[13]; digest[15] = ctxt->h.b8[12]; ++ digest[16] = ctxt->h.b8[19]; digest[17] = ctxt->h.b8[18]; ++ digest[18] = ctxt->h.b8[17]; digest[19] = ctxt->h.b8[16]; ++#endif ++} ++ ++#endif /*unsupported*/ +diff -Nur linux-2.6.30.orig/crypto/ocf/safe/sha1.h linux-2.6.30/crypto/ocf/safe/sha1.h +--- linux-2.6.30.orig/crypto/ocf/safe/sha1.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/safe/sha1.h 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,72 @@ ++/* $FreeBSD: src/sys/crypto/sha1.h,v 1.8 2002/03/20 05:13:50 alfred Exp $ */ ++/* $KAME: sha1.h,v 1.5 2000/03/27 04:36:23 sumikawa Exp $ */ ++ ++/* ++ * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. ++ * All rights reserved. ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions ++ * are met: ++ * 1. Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * 2. Redistributions in binary form must reproduce the above copyright ++ * notice, this list of conditions and the following disclaimer in the ++ * documentation and/or other materials provided with the distribution. ++ * 3. Neither the name of the project nor the names of its contributors ++ * may be used to endorse or promote products derived from this software ++ * without specific prior written permission. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND ++ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE ++ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ++ * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE ++ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL ++ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS ++ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) ++ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT ++ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY ++ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF ++ * SUCH DAMAGE. ++ */ ++/* ++ * FIPS pub 180-1: Secure Hash Algorithm (SHA-1) ++ * based on: http://csrc.nist.gov/fips/fip180-1.txt ++ * implemented by Jun-ichiro itojun Itoh <itojun@itojun.org> ++ */ ++ ++#ifndef _NETINET6_SHA1_H_ ++#define _NETINET6_SHA1_H_ ++ ++struct sha1_ctxt { ++ union { ++ u_int8_t b8[20]; ++ u_int32_t b32[5]; ++ } h; ++ union { ++ u_int8_t b8[8]; ++ u_int64_t b64[1]; ++ } c; ++ union { ++ u_int8_t b8[64]; ++ u_int32_t b32[16]; ++ } m; ++ u_int8_t count; ++}; ++ ++#ifdef __KERNEL__ ++extern void sha1_init(struct sha1_ctxt *); ++extern void sha1_pad(struct sha1_ctxt *); ++extern void sha1_loop(struct sha1_ctxt *, const u_int8_t *, size_t); ++extern void sha1_result(struct sha1_ctxt *, caddr_t); ++ ++/* compatibilty with other SHA1 source codes */ ++typedef struct sha1_ctxt SHA1_CTX; ++#define SHA1Init(x) sha1_init((x)) ++#define SHA1Update(x, y, z) sha1_loop((x), (y), (z)) ++#define SHA1Final(x, y) sha1_result((y), (x)) ++#endif /* __KERNEL__ */ ++ ++#define SHA1_RESULTLEN (160/8) ++ ++#endif /*_NETINET6_SHA1_H_*/ +diff -Nur linux-2.6.30.orig/crypto/ocf/talitos/Makefile linux-2.6.30/crypto/ocf/talitos/Makefile +--- linux-2.6.30.orig/crypto/ocf/talitos/Makefile 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/talitos/Makefile 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,12 @@ ++# for SGlinux builds ++-include $(ROOTDIR)/modules/.config ++ ++obj-$(CONFIG_OCF_TALITOS) += talitos.o ++ ++obj ?= . ++EXTRA_CFLAGS += -I$(obj)/.. -I$(obj)/ ++ ++ifdef TOPDIR ++-include $(TOPDIR)/Rules.make ++endif ++ +diff -Nur linux-2.6.30.orig/crypto/ocf/talitos/talitos.c linux-2.6.30/crypto/ocf/talitos/talitos.c +--- linux-2.6.30.orig/crypto/ocf/talitos/talitos.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/talitos/talitos.c 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,1359 @@ ++/* ++ * crypto/ocf/talitos/talitos.c ++ * ++ * An OCF-Linux module that uses Freescale's SEC to do the crypto. ++ * Based on crypto/ocf/hifn and crypto/ocf/safe OCF drivers ++ * ++ * Copyright (c) 2006 Freescale Semiconductor, Inc. ++ * ++ * This code written by Kim A. B. Phillips <kim.phillips@freescale.com> ++ * some code copied from files with the following: ++ * Copyright (C) 2004-2007 David McCullough <david_mccullough@securecomputing.com ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions ++ * are met: ++ * ++ * 1. Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * 2. Redistributions in binary form must reproduce the above copyright ++ * notice, this list of conditions and the following disclaimer in the ++ * documentation and/or other materials provided with the distribution. ++ * 3. The name of the author may not be used to endorse or promote products ++ * derived from this software without specific prior written permission. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR ++ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES ++ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. ++ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, ++ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT ++ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, ++ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY ++ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT ++ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF ++ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. ++ * ++ * --------------------------------------------------------------------------- ++ * ++ * NOTES: ++ * ++ * The Freescale SEC (also known as 'talitos') resides on the ++ * internal bus, and runs asynchronous to the processor core. It has ++ * a wide gamut of cryptographic acceleration features, including single- ++ * pass IPsec (also known as algorithm chaining). To properly utilize ++ * all of the SEC's performance enhancing features, further reworking ++ * of higher level code (framework, applications) will be necessary. ++ * ++ * The following table shows which SEC version is present in which devices: ++ * ++ * Devices SEC version ++ * ++ * 8272, 8248 SEC 1.0 ++ * 885, 875 SEC 1.2 ++ * 8555E, 8541E SEC 2.0 ++ * 8349E SEC 2.01 ++ * 8548E SEC 2.1 ++ * ++ * The following table shows the features offered by each SEC version: ++ * ++ * Max. chan- ++ * version Bus I/F Clock nels DEU AESU AFEU MDEU PKEU RNG KEU ++ * ++ * SEC 1.0 internal 64b 100MHz 4 1 1 1 1 1 1 0 ++ * SEC 1.2 internal 32b 66MHz 1 1 1 0 1 0 0 0 ++ * SEC 2.0 internal 64b 166MHz 4 1 1 1 1 1 1 0 ++ * SEC 2.01 internal 64b 166MHz 4 1 1 1 1 1 1 0 ++ * SEC 2.1 internal 64b 333MHz 4 1 1 1 1 1 1 1 ++ * ++ * Each execution unit in the SEC has two modes of execution; channel and ++ * slave/debug. This driver employs the channel infrastructure in the ++ * device for convenience. Only the RNG is directly accessed due to the ++ * convenience of its random fifo pool. The relationship between the ++ * channels and execution units is depicted in the following diagram: ++ * ++ * ------- ------------ ++ * ---| ch0 |---| | ++ * ------- | | ++ * | |------+-------+-------+-------+------------ ++ * ------- | | | | | | | ++ * ---| ch1 |---| | | | | | | ++ * ------- | | ------ ------ ------ ------ ------ ++ * |controller| |DEU | |AESU| |MDEU| |PKEU| ... |RNG | ++ * ------- | | ------ ------ ------ ------ ------ ++ * ---| ch2 |---| | | | | | | ++ * ------- | | | | | | | ++ * | |------+-------+-------+-------+------------ ++ * ------- | | ++ * ---| ch3 |---| | ++ * ------- ------------ ++ * ++ * Channel ch0 may drive an aes operation to the aes unit (AESU), ++ * and, at the same time, ch1 may drive a message digest operation ++ * to the mdeu. Each channel has an input descriptor FIFO, and the ++ * FIFO can contain, e.g. on the 8541E, up to 24 entries, before a ++ * a buffer overrun error is triggered. The controller is responsible ++ * for fetching the data from descriptor pointers, and passing the ++ * data to the appropriate EUs. The controller also writes the ++ * cryptographic operation's result to memory. The SEC notifies ++ * completion by triggering an interrupt and/or setting the 1st byte ++ * of the hdr field to 0xff. ++ * ++ * TODO: ++ * o support more algorithms ++ * o support more versions of the SEC ++ * o add support for linux 2.4 ++ * o scatter-gather (sg) support ++ * o add support for public key ops (PKEU) ++ * o add statistics ++ */ ++ ++#ifndef AUTOCONF_INCLUDED ++#include <linux/config.h> ++#endif ++#include <linux/module.h> ++#include <linux/init.h> ++#include <linux/interrupt.h> ++#include <linux/spinlock.h> ++#include <linux/random.h> ++#include <linux/skbuff.h> ++#include <asm/scatterlist.h> ++#include <linux/dma-mapping.h> /* dma_map_single() */ ++#include <linux/moduleparam.h> ++ ++#include <linux/version.h> ++#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,15) ++#include <linux/platform_device.h> ++#endif ++ ++#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,19) ++#include <linux/of_platform.h> ++#endif ++ ++#include <cryptodev.h> ++#include <uio.h> ++ ++#define DRV_NAME "talitos" ++ ++#include "talitos_dev.h" ++#include "talitos_soft.h" ++ ++#define read_random(p,l) get_random_bytes(p,l) ++ ++const char talitos_driver_name[] = "Talitos OCF"; ++const char talitos_driver_version[] = "0.2"; ++ ++static int talitos_newsession(device_t dev, u_int32_t *sidp, ++ struct cryptoini *cri); ++static int talitos_freesession(device_t dev, u_int64_t tid); ++static int talitos_process(device_t dev, struct cryptop *crp, int hint); ++static void dump_talitos_status(struct talitos_softc *sc); ++static int talitos_submit(struct talitos_softc *sc, struct talitos_desc *td, ++ int chsel); ++static void talitos_doneprocessing(struct talitos_softc *sc); ++static void talitos_init_device(struct talitos_softc *sc); ++static void talitos_reset_device_master(struct talitos_softc *sc); ++static void talitos_reset_device(struct talitos_softc *sc); ++static void talitos_errorprocessing(struct talitos_softc *sc); ++#ifdef CONFIG_PPC_MERGE ++static int talitos_probe(struct of_device *ofdev, const struct of_device_id *match); ++static int talitos_remove(struct of_device *ofdev); ++#else ++static int talitos_probe(struct platform_device *pdev); ++static int talitos_remove(struct platform_device *pdev); ++#endif ++#ifdef CONFIG_OCF_RANDOMHARVEST ++static int talitos_read_random(void *arg, u_int32_t *buf, int maxwords); ++static void talitos_rng_init(struct talitos_softc *sc); ++#endif ++ ++static device_method_t talitos_methods = { ++ /* crypto device methods */ ++ DEVMETHOD(cryptodev_newsession, talitos_newsession), ++ DEVMETHOD(cryptodev_freesession,talitos_freesession), ++ DEVMETHOD(cryptodev_process, talitos_process), ++}; ++ ++#define debug talitos_debug ++int talitos_debug = 0; ++module_param(talitos_debug, int, 0644); ++MODULE_PARM_DESC(talitos_debug, "Enable debug"); ++ ++static inline void talitos_write(volatile unsigned *addr, u32 val) ++{ ++ out_be32(addr, val); ++} ++ ++static inline u32 talitos_read(volatile unsigned *addr) ++{ ++ u32 val; ++ val = in_be32(addr); ++ return val; ++} ++ ++static void dump_talitos_status(struct talitos_softc *sc) ++{ ++ unsigned int v, v_hi, i, *ptr; ++ v = talitos_read(sc->sc_base_addr + TALITOS_MCR); ++ v_hi = talitos_read(sc->sc_base_addr + TALITOS_MCR_HI); ++ printk(KERN_INFO "%s: MCR 0x%08x_%08x\n", ++ device_get_nameunit(sc->sc_cdev), v, v_hi); ++ v = talitos_read(sc->sc_base_addr + TALITOS_IMR); ++ v_hi = talitos_read(sc->sc_base_addr + TALITOS_IMR_HI); ++ printk(KERN_INFO "%s: IMR 0x%08x_%08x\n", ++ device_get_nameunit(sc->sc_cdev), v, v_hi); ++ v = talitos_read(sc->sc_base_addr + TALITOS_ISR); ++ v_hi = talitos_read(sc->sc_base_addr + TALITOS_ISR_HI); ++ printk(KERN_INFO "%s: ISR 0x%08x_%08x\n", ++ device_get_nameunit(sc->sc_cdev), v, v_hi); ++ for (i = 0; i < sc->sc_num_channels; i++) { ++ v = talitos_read(sc->sc_base_addr + i*TALITOS_CH_OFFSET + ++ TALITOS_CH_CDPR); ++ v_hi = talitos_read(sc->sc_base_addr + i*TALITOS_CH_OFFSET + ++ TALITOS_CH_CDPR_HI); ++ printk(KERN_INFO "%s: CDPR ch%d 0x%08x_%08x\n", ++ device_get_nameunit(sc->sc_cdev), i, v, v_hi); ++ } ++ for (i = 0; i < sc->sc_num_channels; i++) { ++ v = talitos_read(sc->sc_base_addr + i*TALITOS_CH_OFFSET + ++ TALITOS_CH_CCPSR); ++ v_hi = talitos_read(sc->sc_base_addr + i*TALITOS_CH_OFFSET + ++ TALITOS_CH_CCPSR_HI); ++ printk(KERN_INFO "%s: CCPSR ch%d 0x%08x_%08x\n", ++ device_get_nameunit(sc->sc_cdev), i, v, v_hi); ++ } ++ ptr = sc->sc_base_addr + TALITOS_CH_DESCBUF; ++ for (i = 0; i < 16; i++) { ++ v = talitos_read(ptr++); v_hi = talitos_read(ptr++); ++ printk(KERN_INFO "%s: DESCBUF ch0 0x%08x_%08x (tdp%02d)\n", ++ device_get_nameunit(sc->sc_cdev), v, v_hi, i); ++ } ++ return; ++} ++ ++ ++#ifdef CONFIG_OCF_RANDOMHARVEST ++/* ++ * pull random numbers off the RNG FIFO, not exceeding amount available ++ */ ++static int ++talitos_read_random(void *arg, u_int32_t *buf, int maxwords) ++{ ++ struct talitos_softc *sc = (struct talitos_softc *) arg; ++ int rc; ++ u_int32_t v; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ /* check for things like FIFO underflow */ ++ v = talitos_read(sc->sc_base_addr + TALITOS_RNGISR_HI); ++ if (unlikely(v)) { ++ printk(KERN_ERR "%s: RNGISR_HI error %08x\n", ++ device_get_nameunit(sc->sc_cdev), v); ++ return 0; ++ } ++ /* ++ * OFL is number of available 64-bit words, ++ * shift and convert to a 32-bit word count ++ */ ++ v = talitos_read(sc->sc_base_addr + TALITOS_RNGSR_HI); ++ v = (v & TALITOS_RNGSR_HI_OFL) >> (16 - 1); ++ if (maxwords > v) ++ maxwords = v; ++ for (rc = 0; rc < maxwords; rc++) { ++ buf[rc] = talitos_read(sc->sc_base_addr + ++ TALITOS_RNG_FIFO + rc*sizeof(u_int32_t)); ++ } ++ if (maxwords & 1) { ++ /* ++ * RNG will complain with an AE in the RNGISR ++ * if we don't complete the pairs of 32-bit reads ++ * to its 64-bit register based FIFO ++ */ ++ v = talitos_read(sc->sc_base_addr + ++ TALITOS_RNG_FIFO + rc*sizeof(u_int32_t)); ++ } ++ ++ return rc; ++} ++ ++static void ++talitos_rng_init(struct talitos_softc *sc) ++{ ++ u_int32_t v; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ /* reset RNG EU */ ++ v = talitos_read(sc->sc_base_addr + TALITOS_RNGRCR_HI); ++ v |= TALITOS_RNGRCR_HI_SR; ++ talitos_write(sc->sc_base_addr + TALITOS_RNGRCR_HI, v); ++ while ((talitos_read(sc->sc_base_addr + TALITOS_RNGSR_HI) ++ & TALITOS_RNGSR_HI_RD) == 0) ++ cpu_relax(); ++ /* ++ * we tell the RNG to start filling the RNG FIFO ++ * by writing the RNGDSR ++ */ ++ v = talitos_read(sc->sc_base_addr + TALITOS_RNGDSR_HI); ++ talitos_write(sc->sc_base_addr + TALITOS_RNGDSR_HI, v); ++ /* ++ * 64 bits of data will be pushed onto the FIFO every ++ * 256 SEC cycles until the FIFO is full. The RNG then ++ * attempts to keep the FIFO full. ++ */ ++ v = talitos_read(sc->sc_base_addr + TALITOS_RNGISR_HI); ++ if (v) { ++ printk(KERN_ERR "%s: RNGISR_HI error %08x\n", ++ device_get_nameunit(sc->sc_cdev), v); ++ return; ++ } ++ /* ++ * n.b. we need to add a FIPS test here - if the RNG is going ++ * to fail, it's going to fail at reset time ++ */ ++ return; ++} ++#endif /* CONFIG_OCF_RANDOMHARVEST */ ++ ++/* ++ * Generate a new software session. ++ */ ++static int ++talitos_newsession(device_t dev, u_int32_t *sidp, struct cryptoini *cri) ++{ ++ struct cryptoini *c, *encini = NULL, *macini = NULL; ++ struct talitos_softc *sc = device_get_softc(dev); ++ struct talitos_session *ses = NULL; ++ int sesn; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ if (sidp == NULL || cri == NULL || sc == NULL) { ++ DPRINTF("%s,%d - EINVAL\n", __FILE__, __LINE__); ++ return EINVAL; ++ } ++ for (c = cri; c != NULL; c = c->cri_next) { ++ if (c->cri_alg == CRYPTO_MD5 || ++ c->cri_alg == CRYPTO_MD5_HMAC || ++ c->cri_alg == CRYPTO_SHA1 || ++ c->cri_alg == CRYPTO_SHA1_HMAC || ++ c->cri_alg == CRYPTO_NULL_HMAC) { ++ if (macini) ++ return EINVAL; ++ macini = c; ++ } else if (c->cri_alg == CRYPTO_DES_CBC || ++ c->cri_alg == CRYPTO_3DES_CBC || ++ c->cri_alg == CRYPTO_AES_CBC || ++ c->cri_alg == CRYPTO_NULL_CBC) { ++ if (encini) ++ return EINVAL; ++ encini = c; ++ } else { ++ DPRINTF("UNKNOWN c->cri_alg %d\n", encini->cri_alg); ++ return EINVAL; ++ } ++ } ++ if (encini == NULL && macini == NULL) ++ return EINVAL; ++ if (encini) { ++ /* validate key length */ ++ switch (encini->cri_alg) { ++ case CRYPTO_DES_CBC: ++ if (encini->cri_klen != 64) ++ return EINVAL; ++ break; ++ case CRYPTO_3DES_CBC: ++ if (encini->cri_klen != 192) { ++ return EINVAL; ++ } ++ break; ++ case CRYPTO_AES_CBC: ++ if (encini->cri_klen != 128 && ++ encini->cri_klen != 192 && ++ encini->cri_klen != 256) ++ return EINVAL; ++ break; ++ default: ++ DPRINTF("UNKNOWN encini->cri_alg %d\n", ++ encini->cri_alg); ++ return EINVAL; ++ } ++ } ++ ++ if (sc->sc_sessions == NULL) { ++ ses = sc->sc_sessions = (struct talitos_session *) ++ kmalloc(sizeof(struct talitos_session), SLAB_ATOMIC); ++ if (ses == NULL) ++ return ENOMEM; ++ memset(ses, 0, sizeof(struct talitos_session)); ++ sesn = 0; ++ sc->sc_nsessions = 1; ++ } else { ++ for (sesn = 0; sesn < sc->sc_nsessions; sesn++) { ++ if (sc->sc_sessions[sesn].ses_used == 0) { ++ ses = &sc->sc_sessions[sesn]; ++ break; ++ } ++ } ++ ++ if (ses == NULL) { ++ /* allocating session */ ++ sesn = sc->sc_nsessions; ++ ses = (struct talitos_session *) kmalloc( ++ (sesn + 1) * sizeof(struct talitos_session), ++ SLAB_ATOMIC); ++ if (ses == NULL) ++ return ENOMEM; ++ memset(ses, 0, ++ (sesn + 1) * sizeof(struct talitos_session)); ++ memcpy(ses, sc->sc_sessions, ++ sesn * sizeof(struct talitos_session)); ++ memset(sc->sc_sessions, 0, ++ sesn * sizeof(struct talitos_session)); ++ kfree(sc->sc_sessions); ++ sc->sc_sessions = ses; ++ ses = &sc->sc_sessions[sesn]; ++ sc->sc_nsessions++; ++ } ++ } ++ ++ ses->ses_used = 1; ++ ++ if (encini) { ++ /* get an IV */ ++ /* XXX may read fewer than requested */ ++ read_random(ses->ses_iv, sizeof(ses->ses_iv)); ++ ++ ses->ses_klen = (encini->cri_klen + 7) / 8; ++ memcpy(ses->ses_key, encini->cri_key, ses->ses_klen); ++ if (macini) { ++ /* doing hash on top of cipher */ ++ ses->ses_hmac_len = (macini->cri_klen + 7) / 8; ++ memcpy(ses->ses_hmac, macini->cri_key, ++ ses->ses_hmac_len); ++ } ++ } else if (macini) { ++ /* doing hash */ ++ ses->ses_klen = (macini->cri_klen + 7) / 8; ++ memcpy(ses->ses_key, macini->cri_key, ses->ses_klen); ++ } ++ ++ /* back compat way of determining MSC result len */ ++ if (macini) { ++ ses->ses_mlen = macini->cri_mlen; ++ if (ses->ses_mlen == 0) { ++ if (macini->cri_alg == CRYPTO_MD5_HMAC) ++ ses->ses_mlen = MD5_HASH_LEN; ++ else ++ ses->ses_mlen = SHA1_HASH_LEN; ++ } ++ } ++ ++ /* really should make up a template td here, ++ * and only fill things like i/o and direction in process() */ ++ ++ /* assign session ID */ ++ *sidp = TALITOS_SID(sc->sc_num, sesn); ++ return 0; ++} ++ ++/* ++ * Deallocate a session. ++ */ ++static int ++talitos_freesession(device_t dev, u_int64_t tid) ++{ ++ struct talitos_softc *sc = device_get_softc(dev); ++ int session, ret; ++ u_int32_t sid = ((u_int32_t) tid) & 0xffffffff; ++ ++ if (sc == NULL) ++ return EINVAL; ++ session = TALITOS_SESSION(sid); ++ if (session < sc->sc_nsessions) { ++ memset(&sc->sc_sessions[session], 0, ++ sizeof(sc->sc_sessions[session])); ++ ret = 0; ++ } else ++ ret = EINVAL; ++ return ret; ++} ++ ++/* ++ * launch device processing - it will come back with done notification ++ * in the form of an interrupt and/or HDR_DONE_BITS in header ++ */ ++static int ++talitos_submit( ++ struct talitos_softc *sc, ++ struct talitos_desc *td, ++ int chsel) ++{ ++ u_int32_t v; ++ ++ v = dma_map_single(NULL, td, sizeof(*td), DMA_TO_DEVICE); ++ talitos_write(sc->sc_base_addr + ++ chsel*TALITOS_CH_OFFSET + TALITOS_CH_FF, 0); ++ talitos_write(sc->sc_base_addr + ++ chsel*TALITOS_CH_OFFSET + TALITOS_CH_FF_HI, v); ++ return 0; ++} ++ ++static int ++talitos_process(device_t dev, struct cryptop *crp, int hint) ++{ ++ int i, err = 0, ivsize; ++ struct talitos_softc *sc = device_get_softc(dev); ++ struct cryptodesc *crd1, *crd2, *maccrd, *enccrd; ++ caddr_t iv; ++ struct talitos_session *ses; ++ struct talitos_desc *td; ++ unsigned long flags; ++ /* descriptor mappings */ ++ int hmac_key, hmac_data, cipher_iv, cipher_key, ++ in_fifo, out_fifo, cipher_iv_out; ++ static int chsel = -1; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ if (crp == NULL || crp->crp_callback == NULL || sc == NULL) { ++ return EINVAL; ++ } ++ crp->crp_etype = 0; ++ if (TALITOS_SESSION(crp->crp_sid) >= sc->sc_nsessions) { ++ return EINVAL; ++ } ++ ++ ses = &sc->sc_sessions[TALITOS_SESSION(crp->crp_sid)]; ++ ++ /* enter the channel scheduler */ ++ spin_lock_irqsave(&sc->sc_chnfifolock[sc->sc_num_channels], flags); ++ ++ /* reuse channel that already had/has requests for the required EU */ ++ for (i = 0; i < sc->sc_num_channels; i++) { ++ if (sc->sc_chnlastalg[i] == crp->crp_desc->crd_alg) ++ break; ++ } ++ if (i == sc->sc_num_channels) { ++ /* ++ * haven't seen this algo the last sc_num_channels or more ++ * use round robin in this case ++ * nb: sc->sc_num_channels must be power of 2 ++ */ ++ chsel = (chsel + 1) & (sc->sc_num_channels - 1); ++ } else { ++ /* ++ * matches channel with same target execution unit; ++ * use same channel in this case ++ */ ++ chsel = i; ++ } ++ sc->sc_chnlastalg[chsel] = crp->crp_desc->crd_alg; ++ ++ /* release the channel scheduler lock */ ++ spin_unlock_irqrestore(&sc->sc_chnfifolock[sc->sc_num_channels], flags); ++ ++ /* acquire the selected channel fifo lock */ ++ spin_lock_irqsave(&sc->sc_chnfifolock[chsel], flags); ++ ++ /* find and reserve next available descriptor-cryptop pair */ ++ for (i = 0; i < sc->sc_chfifo_len; i++) { ++ if (sc->sc_chnfifo[chsel][i].cf_desc.hdr == 0) { ++ /* ++ * ensure correct descriptor formation by ++ * avoiding inadvertently setting "optional" entries ++ * e.g. not using "optional" dptr2 for MD/HMAC descs ++ */ ++ memset(&sc->sc_chnfifo[chsel][i].cf_desc, ++ 0, sizeof(*td)); ++ /* reserve it with done notification request bit */ ++ sc->sc_chnfifo[chsel][i].cf_desc.hdr |= ++ TALITOS_DONE_NOTIFY; ++ break; ++ } ++ } ++ spin_unlock_irqrestore(&sc->sc_chnfifolock[chsel], flags); ++ ++ if (i == sc->sc_chfifo_len) { ++ /* fifo full */ ++ err = ERESTART; ++ goto errout; ++ } ++ ++ td = &sc->sc_chnfifo[chsel][i].cf_desc; ++ sc->sc_chnfifo[chsel][i].cf_crp = crp; ++ ++ crd1 = crp->crp_desc; ++ if (crd1 == NULL) { ++ err = EINVAL; ++ goto errout; ++ } ++ crd2 = crd1->crd_next; ++ /* prevent compiler warning */ ++ hmac_key = 0; ++ hmac_data = 0; ++ if (crd2 == NULL) { ++ td->hdr |= TD_TYPE_COMMON_NONSNOOP_NO_AFEU; ++ /* assign descriptor dword ptr mappings for this desc. type */ ++ cipher_iv = 1; ++ cipher_key = 2; ++ in_fifo = 3; ++ cipher_iv_out = 5; ++ if (crd1->crd_alg == CRYPTO_MD5_HMAC || ++ crd1->crd_alg == CRYPTO_SHA1_HMAC || ++ crd1->crd_alg == CRYPTO_SHA1 || ++ crd1->crd_alg == CRYPTO_MD5) { ++ out_fifo = 5; ++ maccrd = crd1; ++ enccrd = NULL; ++ } else if (crd1->crd_alg == CRYPTO_DES_CBC || ++ crd1->crd_alg == CRYPTO_3DES_CBC || ++ crd1->crd_alg == CRYPTO_AES_CBC || ++ crd1->crd_alg == CRYPTO_ARC4) { ++ out_fifo = 4; ++ maccrd = NULL; ++ enccrd = crd1; ++ } else { ++ DPRINTF("UNKNOWN crd1->crd_alg %d\n", crd1->crd_alg); ++ err = EINVAL; ++ goto errout; ++ } ++ } else { ++ if (sc->sc_desc_types & TALITOS_HAS_DT_IPSEC_ESP) { ++ td->hdr |= TD_TYPE_IPSEC_ESP; ++ } else { ++ DPRINTF("unimplemented: multiple descriptor ipsec\n"); ++ err = EINVAL; ++ goto errout; ++ } ++ /* assign descriptor dword ptr mappings for this desc. type */ ++ hmac_key = 0; ++ hmac_data = 1; ++ cipher_iv = 2; ++ cipher_key = 3; ++ in_fifo = 4; ++ out_fifo = 5; ++ cipher_iv_out = 6; ++ if ((crd1->crd_alg == CRYPTO_MD5_HMAC || ++ crd1->crd_alg == CRYPTO_SHA1_HMAC || ++ crd1->crd_alg == CRYPTO_MD5 || ++ crd1->crd_alg == CRYPTO_SHA1) && ++ (crd2->crd_alg == CRYPTO_DES_CBC || ++ crd2->crd_alg == CRYPTO_3DES_CBC || ++ crd2->crd_alg == CRYPTO_AES_CBC || ++ crd2->crd_alg == CRYPTO_ARC4) && ++ ((crd2->crd_flags & CRD_F_ENCRYPT) == 0)) { ++ maccrd = crd1; ++ enccrd = crd2; ++ } else if ((crd1->crd_alg == CRYPTO_DES_CBC || ++ crd1->crd_alg == CRYPTO_ARC4 || ++ crd1->crd_alg == CRYPTO_3DES_CBC || ++ crd1->crd_alg == CRYPTO_AES_CBC) && ++ (crd2->crd_alg == CRYPTO_MD5_HMAC || ++ crd2->crd_alg == CRYPTO_SHA1_HMAC || ++ crd2->crd_alg == CRYPTO_MD5 || ++ crd2->crd_alg == CRYPTO_SHA1) && ++ (crd1->crd_flags & CRD_F_ENCRYPT)) { ++ enccrd = crd1; ++ maccrd = crd2; ++ } else { ++ /* We cannot order the SEC as requested */ ++ printk("%s: cannot do the order\n", ++ device_get_nameunit(sc->sc_cdev)); ++ err = EINVAL; ++ goto errout; ++ } ++ } ++ /* assign in_fifo and out_fifo based on input/output struct type */ ++ if (crp->crp_flags & CRYPTO_F_SKBUF) { ++ /* using SKB buffers */ ++ struct sk_buff *skb = (struct sk_buff *)crp->crp_buf; ++ if (skb_shinfo(skb)->nr_frags) { ++ printk("%s: skb frags unimplemented\n", ++ device_get_nameunit(sc->sc_cdev)); ++ err = EINVAL; ++ goto errout; ++ } ++ td->ptr[in_fifo].ptr = dma_map_single(NULL, skb->data, ++ skb->len, DMA_TO_DEVICE); ++ td->ptr[in_fifo].len = skb->len; ++ td->ptr[out_fifo].ptr = dma_map_single(NULL, skb->data, ++ skb->len, DMA_TO_DEVICE); ++ td->ptr[out_fifo].len = skb->len; ++ td->ptr[hmac_data].ptr = dma_map_single(NULL, skb->data, ++ skb->len, DMA_TO_DEVICE); ++ } else if (crp->crp_flags & CRYPTO_F_IOV) { ++ /* using IOV buffers */ ++ struct uio *uiop = (struct uio *)crp->crp_buf; ++ if (uiop->uio_iovcnt > 1) { ++ printk("%s: iov frags unimplemented\n", ++ device_get_nameunit(sc->sc_cdev)); ++ err = EINVAL; ++ goto errout; ++ } ++ td->ptr[in_fifo].ptr = dma_map_single(NULL, ++ uiop->uio_iov->iov_base, crp->crp_ilen, DMA_TO_DEVICE); ++ td->ptr[in_fifo].len = crp->crp_ilen; ++ /* crp_olen is never set; always use crp_ilen */ ++ td->ptr[out_fifo].ptr = dma_map_single(NULL, ++ uiop->uio_iov->iov_base, ++ crp->crp_ilen, DMA_TO_DEVICE); ++ td->ptr[out_fifo].len = crp->crp_ilen; ++ } else { ++ /* using contig buffers */ ++ td->ptr[in_fifo].ptr = dma_map_single(NULL, ++ crp->crp_buf, crp->crp_ilen, DMA_TO_DEVICE); ++ td->ptr[in_fifo].len = crp->crp_ilen; ++ td->ptr[out_fifo].ptr = dma_map_single(NULL, ++ crp->crp_buf, crp->crp_ilen, DMA_TO_DEVICE); ++ td->ptr[out_fifo].len = crp->crp_ilen; ++ } ++ if (enccrd) { ++ switch (enccrd->crd_alg) { ++ case CRYPTO_3DES_CBC: ++ td->hdr |= TALITOS_MODE0_DEU_3DES; ++ /* FALLTHROUGH */ ++ case CRYPTO_DES_CBC: ++ td->hdr |= TALITOS_SEL0_DEU ++ | TALITOS_MODE0_DEU_CBC; ++ if (enccrd->crd_flags & CRD_F_ENCRYPT) ++ td->hdr |= TALITOS_MODE0_DEU_ENC; ++ ivsize = 2*sizeof(u_int32_t); ++ DPRINTF("%cDES ses %d ch %d len %d\n", ++ (td->hdr & TALITOS_MODE0_DEU_3DES)?'3':'1', ++ (u32)TALITOS_SESSION(crp->crp_sid), ++ chsel, td->ptr[in_fifo].len); ++ break; ++ case CRYPTO_AES_CBC: ++ td->hdr |= TALITOS_SEL0_AESU ++ | TALITOS_MODE0_AESU_CBC; ++ if (enccrd->crd_flags & CRD_F_ENCRYPT) ++ td->hdr |= TALITOS_MODE0_AESU_ENC; ++ ivsize = 4*sizeof(u_int32_t); ++ DPRINTF("AES ses %d ch %d len %d\n", ++ (u32)TALITOS_SESSION(crp->crp_sid), ++ chsel, td->ptr[in_fifo].len); ++ break; ++ default: ++ printk("%s: unimplemented enccrd->crd_alg %d\n", ++ device_get_nameunit(sc->sc_cdev), enccrd->crd_alg); ++ err = EINVAL; ++ goto errout; ++ } ++ /* ++ * Setup encrypt/decrypt state. When using basic ops ++ * we can't use an inline IV because hash/crypt offset ++ * must be from the end of the IV to the start of the ++ * crypt data and this leaves out the preceding header ++ * from the hash calculation. Instead we place the IV ++ * in the state record and set the hash/crypt offset to ++ * copy both the header+IV. ++ */ ++ if (enccrd->crd_flags & CRD_F_ENCRYPT) { ++ td->hdr |= TALITOS_DIR_OUTBOUND; ++ if (enccrd->crd_flags & CRD_F_IV_EXPLICIT) ++ iv = enccrd->crd_iv; ++ else ++ iv = (caddr_t) ses->ses_iv; ++ if ((enccrd->crd_flags & CRD_F_IV_PRESENT) == 0) { ++ crypto_copyback(crp->crp_flags, crp->crp_buf, ++ enccrd->crd_inject, ivsize, iv); ++ } ++ } else { ++ td->hdr |= TALITOS_DIR_INBOUND; ++ if (enccrd->crd_flags & CRD_F_IV_EXPLICIT) { ++ iv = enccrd->crd_iv; ++ bcopy(enccrd->crd_iv, iv, ivsize); ++ } else { ++ iv = (caddr_t) ses->ses_iv; ++ crypto_copydata(crp->crp_flags, crp->crp_buf, ++ enccrd->crd_inject, ivsize, iv); ++ } ++ } ++ td->ptr[cipher_iv].ptr = dma_map_single(NULL, iv, ivsize, ++ DMA_TO_DEVICE); ++ td->ptr[cipher_iv].len = ivsize; ++ /* ++ * we don't need the cipher iv out length/pointer ++ * field to do ESP IPsec. Therefore we set the len field as 0, ++ * which tells the SEC not to do anything with this len/ptr ++ * field. Previously, when length/pointer as pointing to iv, ++ * it gave us corruption of packets. ++ */ ++ td->ptr[cipher_iv_out].len = 0; ++ } ++ if (enccrd && maccrd) { ++ /* this is ipsec only for now */ ++ td->hdr |= TALITOS_SEL1_MDEU ++ | TALITOS_MODE1_MDEU_INIT ++ | TALITOS_MODE1_MDEU_PAD; ++ switch (maccrd->crd_alg) { ++ case CRYPTO_MD5: ++ td->hdr |= TALITOS_MODE1_MDEU_MD5; ++ break; ++ case CRYPTO_MD5_HMAC: ++ td->hdr |= TALITOS_MODE1_MDEU_MD5_HMAC; ++ break; ++ case CRYPTO_SHA1: ++ td->hdr |= TALITOS_MODE1_MDEU_SHA1; ++ break; ++ case CRYPTO_SHA1_HMAC: ++ td->hdr |= TALITOS_MODE1_MDEU_SHA1_HMAC; ++ break; ++ default: ++ /* We cannot order the SEC as requested */ ++ printk("%s: cannot do the order\n", ++ device_get_nameunit(sc->sc_cdev)); ++ err = EINVAL; ++ goto errout; ++ } ++ if ((maccrd->crd_alg == CRYPTO_MD5_HMAC) || ++ (maccrd->crd_alg == CRYPTO_SHA1_HMAC)) { ++ /* ++ * The offset from hash data to the start of ++ * crypt data is the difference in the skips. ++ */ ++ /* ipsec only for now */ ++ td->ptr[hmac_key].ptr = dma_map_single(NULL, ++ ses->ses_hmac, ses->ses_hmac_len, DMA_TO_DEVICE); ++ td->ptr[hmac_key].len = ses->ses_hmac_len; ++ td->ptr[in_fifo].ptr += enccrd->crd_skip; ++ td->ptr[in_fifo].len = enccrd->crd_len; ++ td->ptr[out_fifo].ptr += enccrd->crd_skip; ++ td->ptr[out_fifo].len = enccrd->crd_len; ++ /* bytes of HMAC to postpend to ciphertext */ ++ td->ptr[out_fifo].extent = ses->ses_mlen; ++ td->ptr[hmac_data].ptr += maccrd->crd_skip; ++ td->ptr[hmac_data].len = enccrd->crd_skip - maccrd->crd_skip; ++ } ++ if (enccrd->crd_flags & CRD_F_KEY_EXPLICIT) { ++ printk("%s: CRD_F_KEY_EXPLICIT unimplemented\n", ++ device_get_nameunit(sc->sc_cdev)); ++ } ++ } ++ if (!enccrd && maccrd) { ++ /* single MD5 or SHA */ ++ td->hdr |= TALITOS_SEL0_MDEU ++ | TALITOS_MODE0_MDEU_INIT ++ | TALITOS_MODE0_MDEU_PAD; ++ switch (maccrd->crd_alg) { ++ case CRYPTO_MD5: ++ td->hdr |= TALITOS_MODE0_MDEU_MD5; ++ DPRINTF("MD5 ses %d ch %d len %d\n", ++ (u32)TALITOS_SESSION(crp->crp_sid), ++ chsel, td->ptr[in_fifo].len); ++ break; ++ case CRYPTO_MD5_HMAC: ++ td->hdr |= TALITOS_MODE0_MDEU_MD5_HMAC; ++ break; ++ case CRYPTO_SHA1: ++ td->hdr |= TALITOS_MODE0_MDEU_SHA1; ++ DPRINTF("SHA1 ses %d ch %d len %d\n", ++ (u32)TALITOS_SESSION(crp->crp_sid), ++ chsel, td->ptr[in_fifo].len); ++ break; ++ case CRYPTO_SHA1_HMAC: ++ td->hdr |= TALITOS_MODE0_MDEU_SHA1_HMAC; ++ break; ++ default: ++ /* We cannot order the SEC as requested */ ++ DPRINTF("cannot do the order\n"); ++ err = EINVAL; ++ goto errout; ++ } ++ ++ if (crp->crp_flags & CRYPTO_F_IOV) ++ td->ptr[out_fifo].ptr += maccrd->crd_inject; ++ ++ if ((maccrd->crd_alg == CRYPTO_MD5_HMAC) || ++ (maccrd->crd_alg == CRYPTO_SHA1_HMAC)) { ++ td->ptr[hmac_key].ptr = dma_map_single(NULL, ++ ses->ses_hmac, ses->ses_hmac_len, ++ DMA_TO_DEVICE); ++ td->ptr[hmac_key].len = ses->ses_hmac_len; ++ } ++ } ++ else { ++ /* using process key (session data has duplicate) */ ++ td->ptr[cipher_key].ptr = dma_map_single(NULL, ++ enccrd->crd_key, (enccrd->crd_klen + 7) / 8, ++ DMA_TO_DEVICE); ++ td->ptr[cipher_key].len = (enccrd->crd_klen + 7) / 8; ++ } ++ /* descriptor complete - GO! */ ++ return talitos_submit(sc, td, chsel); ++ ++errout: ++ if (err != ERESTART) { ++ crp->crp_etype = err; ++ crypto_done(crp); ++ } ++ return err; ++} ++ ++/* go through all channels descriptors, notifying OCF what has ++ * _and_hasn't_ successfully completed and reset the device ++ * (otherwise it's up to decoding desc hdrs!) ++ */ ++static void talitos_errorprocessing(struct talitos_softc *sc) ++{ ++ unsigned long flags; ++ int i, j; ++ ++ /* disable further scheduling until under control */ ++ spin_lock_irqsave(&sc->sc_chnfifolock[sc->sc_num_channels], flags); ++ ++ if (debug) dump_talitos_status(sc); ++ /* go through descriptors, try and salvage those successfully done, ++ * and EIO those that weren't ++ */ ++ for (i = 0; i < sc->sc_num_channels; i++) { ++ spin_lock_irqsave(&sc->sc_chnfifolock[i], flags); ++ for (j = 0; j < sc->sc_chfifo_len; j++) { ++ if (sc->sc_chnfifo[i][j].cf_desc.hdr) { ++ if ((sc->sc_chnfifo[i][j].cf_desc.hdr ++ & TALITOS_HDR_DONE_BITS) ++ != TALITOS_HDR_DONE_BITS) { ++ /* this one didn't finish */ ++ /* signify in crp->etype */ ++ sc->sc_chnfifo[i][j].cf_crp->crp_etype ++ = EIO; ++ } ++ } else ++ continue; /* free entry */ ++ /* either way, notify ocf */ ++ crypto_done(sc->sc_chnfifo[i][j].cf_crp); ++ /* and tag it available again ++ * ++ * memset to ensure correct descriptor formation by ++ * avoiding inadvertently setting "optional" entries ++ * e.g. not using "optional" dptr2 MD/HMAC processing ++ */ ++ memset(&sc->sc_chnfifo[i][j].cf_desc, ++ 0, sizeof(struct talitos_desc)); ++ } ++ spin_unlock_irqrestore(&sc->sc_chnfifolock[i], flags); ++ } ++ /* reset and initialize the SEC h/w device */ ++ talitos_reset_device(sc); ++ talitos_init_device(sc); ++#ifdef CONFIG_OCF_RANDOMHARVEST ++ if (sc->sc_exec_units & TALITOS_HAS_EU_RNG) ++ talitos_rng_init(sc); ++#endif ++ ++ /* Okay. Stand by. */ ++ spin_unlock_irqrestore(&sc->sc_chnfifolock[sc->sc_num_channels], flags); ++ ++ return; ++} ++ ++/* go through all channels descriptors, notifying OCF what's been done */ ++static void talitos_doneprocessing(struct talitos_softc *sc) ++{ ++ unsigned long flags; ++ int i, j; ++ ++ /* go through descriptors looking for done bits */ ++ for (i = 0; i < sc->sc_num_channels; i++) { ++ spin_lock_irqsave(&sc->sc_chnfifolock[i], flags); ++ for (j = 0; j < sc->sc_chfifo_len; j++) { ++ /* descriptor has done bits set? */ ++ if ((sc->sc_chnfifo[i][j].cf_desc.hdr ++ & TALITOS_HDR_DONE_BITS) ++ == TALITOS_HDR_DONE_BITS) { ++ /* notify ocf */ ++ crypto_done(sc->sc_chnfifo[i][j].cf_crp); ++ /* and tag it available again ++ * ++ * memset to ensure correct descriptor formation by ++ * avoiding inadvertently setting "optional" entries ++ * e.g. not using "optional" dptr2 MD/HMAC processing ++ */ ++ memset(&sc->sc_chnfifo[i][j].cf_desc, ++ 0, sizeof(struct talitos_desc)); ++ } ++ } ++ spin_unlock_irqrestore(&sc->sc_chnfifolock[i], flags); ++ } ++ return; ++} ++ ++static irqreturn_t ++#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,19) ++talitos_intr(int irq, void *arg) ++#else ++talitos_intr(int irq, void *arg, struct pt_regs *regs) ++#endif ++{ ++ struct talitos_softc *sc = arg; ++ u_int32_t v, v_hi; ++ ++ /* ack */ ++ v = talitos_read(sc->sc_base_addr + TALITOS_ISR); ++ v_hi = talitos_read(sc->sc_base_addr + TALITOS_ISR_HI); ++ talitos_write(sc->sc_base_addr + TALITOS_ICR, v); ++ talitos_write(sc->sc_base_addr + TALITOS_ICR_HI, v_hi); ++ ++ if (unlikely(v & TALITOS_ISR_ERROR)) { ++ /* Okay, Houston, we've had a problem here. */ ++ printk(KERN_DEBUG "%s: got error interrupt - ISR 0x%08x_%08x\n", ++ device_get_nameunit(sc->sc_cdev), v, v_hi); ++ talitos_errorprocessing(sc); ++ } else ++ if (likely(v & TALITOS_ISR_DONE)) { ++ talitos_doneprocessing(sc); ++ } ++ return IRQ_HANDLED; ++} ++ ++/* ++ * Initialize registers we need to touch only once. ++ */ ++static void ++talitos_init_device(struct talitos_softc *sc) ++{ ++ u_int32_t v; ++ int i; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ /* init all channels */ ++ for (i = 0; i < sc->sc_num_channels; i++) { ++ v = talitos_read(sc->sc_base_addr + ++ i*TALITOS_CH_OFFSET + TALITOS_CH_CCCR_HI); ++ v |= TALITOS_CH_CCCR_HI_CDWE ++ | TALITOS_CH_CCCR_HI_CDIE; /* invoke interrupt if done */ ++ talitos_write(sc->sc_base_addr + ++ i*TALITOS_CH_OFFSET + TALITOS_CH_CCCR_HI, v); ++ } ++ /* enable all interrupts */ ++ v = talitos_read(sc->sc_base_addr + TALITOS_IMR); ++ v |= TALITOS_IMR_ALL; ++ talitos_write(sc->sc_base_addr + TALITOS_IMR, v); ++ v = talitos_read(sc->sc_base_addr + TALITOS_IMR_HI); ++ v |= TALITOS_IMR_HI_ERRONLY; ++ talitos_write(sc->sc_base_addr + TALITOS_IMR_HI, v); ++ return; ++} ++ ++/* ++ * set the master reset bit on the device. ++ */ ++static void ++talitos_reset_device_master(struct talitos_softc *sc) ++{ ++ u_int32_t v; ++ ++ /* Reset the device by writing 1 to MCR:SWR and waiting 'til cleared */ ++ v = talitos_read(sc->sc_base_addr + TALITOS_MCR); ++ talitos_write(sc->sc_base_addr + TALITOS_MCR, v | TALITOS_MCR_SWR); ++ ++ while (talitos_read(sc->sc_base_addr + TALITOS_MCR) & TALITOS_MCR_SWR) ++ cpu_relax(); ++ ++ return; ++} ++ ++/* ++ * Resets the device. Values in the registers are left as is ++ * from the reset (i.e. initial values are assigned elsewhere). ++ */ ++static void ++talitos_reset_device(struct talitos_softc *sc) ++{ ++ u_int32_t v; ++ int i; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ /* ++ * Master reset ++ * errata documentation: warning: certain SEC interrupts ++ * are not fully cleared by writing the MCR:SWR bit, ++ * set bit twice to completely reset ++ */ ++ talitos_reset_device_master(sc); /* once */ ++ talitos_reset_device_master(sc); /* and once again */ ++ ++ /* reset all channels */ ++ for (i = 0; i < sc->sc_num_channels; i++) { ++ v = talitos_read(sc->sc_base_addr + i*TALITOS_CH_OFFSET + ++ TALITOS_CH_CCCR); ++ talitos_write(sc->sc_base_addr + i*TALITOS_CH_OFFSET + ++ TALITOS_CH_CCCR, v | TALITOS_CH_CCCR_RESET); ++ } ++} ++ ++/* Set up the crypto device structure, private data, ++ * and anything else we need before we start */ ++#ifdef CONFIG_PPC_MERGE ++static int talitos_probe(struct of_device *ofdev, const struct of_device_id *match) ++#else ++static int talitos_probe(struct platform_device *pdev) ++#endif ++{ ++ struct talitos_softc *sc = NULL; ++ struct resource *r; ++#ifdef CONFIG_PPC_MERGE ++ struct device *device = &ofdev->dev; ++ struct device_node *np = ofdev->node; ++ const unsigned int *prop; ++ int err; ++ struct resource res; ++#endif ++ static int num_chips = 0; ++ int rc; ++ int i; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ ++ sc = (struct talitos_softc *) kmalloc(sizeof(*sc), GFP_KERNEL); ++ if (!sc) ++ return -ENOMEM; ++ memset(sc, 0, sizeof(*sc)); ++ ++ softc_device_init(sc, DRV_NAME, num_chips, talitos_methods); ++ ++ sc->sc_irq = -1; ++ sc->sc_cid = -1; ++#ifndef CONFIG_PPC_MERGE ++ sc->sc_dev = pdev; ++#endif ++ sc->sc_num = num_chips++; ++ ++#ifdef CONFIG_PPC_MERGE ++ dev_set_drvdata(device, sc); ++#else ++ platform_set_drvdata(sc->sc_dev, sc); ++#endif ++ ++ /* get the irq line */ ++#ifdef CONFIG_PPC_MERGE ++ err = of_address_to_resource(np, 0, &res); ++ if (err) ++ return -EINVAL; ++ r = &res; ++ ++ sc->sc_irq = irq_of_parse_and_map(np, 0); ++#else ++ /* get a pointer to the register memory */ ++ r = platform_get_resource(pdev, IORESOURCE_MEM, 0); ++ ++ sc->sc_irq = platform_get_irq(pdev, 0); ++#endif ++ rc = request_irq(sc->sc_irq, talitos_intr, 0, ++ device_get_nameunit(sc->sc_cdev), sc); ++ if (rc) { ++ printk(KERN_ERR "%s: failed to hook irq %d\n", ++ device_get_nameunit(sc->sc_cdev), sc->sc_irq); ++ sc->sc_irq = -1; ++ goto out; ++ } ++ ++ sc->sc_base_addr = (ocf_iomem_t) ioremap(r->start, (r->end - r->start)); ++ if (!sc->sc_base_addr) { ++ printk(KERN_ERR "%s: failed to ioremap\n", ++ device_get_nameunit(sc->sc_cdev)); ++ goto out; ++ } ++ ++ /* figure out our SEC's properties and capabilities */ ++ sc->sc_chiprev = (u64)talitos_read(sc->sc_base_addr + TALITOS_ID) << 32 ++ | talitos_read(sc->sc_base_addr + TALITOS_ID_HI); ++ DPRINTF("sec id 0x%llx\n", sc->sc_chiprev); ++ ++#ifdef CONFIG_PPC_MERGE ++ /* get SEC properties from device tree, defaulting to SEC 2.0 */ ++ ++ prop = of_get_property(np, "num-channels", NULL); ++ sc->sc_num_channels = prop ? *prop : TALITOS_NCHANNELS_SEC_2_0; ++ ++ prop = of_get_property(np, "channel-fifo-len", NULL); ++ sc->sc_chfifo_len = prop ? *prop : TALITOS_CHFIFOLEN_SEC_2_0; ++ ++ prop = of_get_property(np, "exec-units-mask", NULL); ++ sc->sc_exec_units = prop ? *prop : TALITOS_HAS_EUS_SEC_2_0; ++ ++ prop = of_get_property(np, "descriptor-types-mask", NULL); ++ sc->sc_desc_types = prop ? *prop : TALITOS_HAS_DESCTYPES_SEC_2_0; ++#else ++ /* bulk should go away with openfirmware flat device tree support */ ++ if (sc->sc_chiprev & TALITOS_ID_SEC_2_0) { ++ sc->sc_num_channels = TALITOS_NCHANNELS_SEC_2_0; ++ sc->sc_chfifo_len = TALITOS_CHFIFOLEN_SEC_2_0; ++ sc->sc_exec_units = TALITOS_HAS_EUS_SEC_2_0; ++ sc->sc_desc_types = TALITOS_HAS_DESCTYPES_SEC_2_0; ++ } else { ++ printk(KERN_ERR "%s: failed to id device\n", ++ device_get_nameunit(sc->sc_cdev)); ++ goto out; ++ } ++#endif ++ ++ /* + 1 is for the meta-channel lock used by the channel scheduler */ ++ sc->sc_chnfifolock = (spinlock_t *) kmalloc( ++ (sc->sc_num_channels + 1) * sizeof(spinlock_t), GFP_KERNEL); ++ if (!sc->sc_chnfifolock) ++ goto out; ++ for (i = 0; i < sc->sc_num_channels + 1; i++) { ++ spin_lock_init(&sc->sc_chnfifolock[i]); ++ } ++ ++ sc->sc_chnlastalg = (int *) kmalloc( ++ sc->sc_num_channels * sizeof(int), GFP_KERNEL); ++ if (!sc->sc_chnlastalg) ++ goto out; ++ memset(sc->sc_chnlastalg, 0, sc->sc_num_channels * sizeof(int)); ++ ++ sc->sc_chnfifo = (struct desc_cryptop_pair **) kmalloc( ++ sc->sc_num_channels * sizeof(struct desc_cryptop_pair *), ++ GFP_KERNEL); ++ if (!sc->sc_chnfifo) ++ goto out; ++ for (i = 0; i < sc->sc_num_channels; i++) { ++ sc->sc_chnfifo[i] = (struct desc_cryptop_pair *) kmalloc( ++ sc->sc_chfifo_len * sizeof(struct desc_cryptop_pair), ++ GFP_KERNEL); ++ if (!sc->sc_chnfifo[i]) ++ goto out; ++ memset(sc->sc_chnfifo[i], 0, ++ sc->sc_chfifo_len * sizeof(struct desc_cryptop_pair)); ++ } ++ ++ /* reset and initialize the SEC h/w device */ ++ talitos_reset_device(sc); ++ talitos_init_device(sc); ++ ++ sc->sc_cid = crypto_get_driverid(softc_get_device(sc),CRYPTOCAP_F_HARDWARE); ++ if (sc->sc_cid < 0) { ++ printk(KERN_ERR "%s: could not get crypto driver id\n", ++ device_get_nameunit(sc->sc_cdev)); ++ goto out; ++ } ++ ++ /* register algorithms with the framework */ ++ printk("%s:", device_get_nameunit(sc->sc_cdev)); ++ ++ if (sc->sc_exec_units & TALITOS_HAS_EU_RNG) { ++ printk(" rng"); ++#ifdef CONFIG_OCF_RANDOMHARVEST ++ talitos_rng_init(sc); ++ crypto_rregister(sc->sc_cid, talitos_read_random, sc); ++#endif ++ } ++ if (sc->sc_exec_units & TALITOS_HAS_EU_DEU) { ++ printk(" des/3des"); ++ crypto_register(sc->sc_cid, CRYPTO_3DES_CBC, 0, 0); ++ crypto_register(sc->sc_cid, CRYPTO_DES_CBC, 0, 0); ++ } ++ if (sc->sc_exec_units & TALITOS_HAS_EU_AESU) { ++ printk(" aes"); ++ crypto_register(sc->sc_cid, CRYPTO_AES_CBC, 0, 0); ++ } ++ if (sc->sc_exec_units & TALITOS_HAS_EU_MDEU) { ++ printk(" md5"); ++ crypto_register(sc->sc_cid, CRYPTO_MD5, 0, 0); ++ /* HMAC support only with IPsec for now */ ++ crypto_register(sc->sc_cid, CRYPTO_MD5_HMAC, 0, 0); ++ printk(" sha1"); ++ crypto_register(sc->sc_cid, CRYPTO_SHA1, 0, 0); ++ /* HMAC support only with IPsec for now */ ++ crypto_register(sc->sc_cid, CRYPTO_SHA1_HMAC, 0, 0); ++ } ++ printk("\n"); ++ return 0; ++ ++out: ++#ifndef CONFIG_PPC_MERGE ++ talitos_remove(pdev); ++#endif ++ return -ENOMEM; ++} ++ ++#ifdef CONFIG_PPC_MERGE ++static int talitos_remove(struct of_device *ofdev) ++#else ++static int talitos_remove(struct platform_device *pdev) ++#endif ++{ ++#ifdef CONFIG_PPC_MERGE ++ struct talitos_softc *sc = dev_get_drvdata(&ofdev->dev); ++#else ++ struct talitos_softc *sc = platform_get_drvdata(pdev); ++#endif ++ int i; ++ ++ DPRINTF("%s()\n", __FUNCTION__); ++ if (sc->sc_cid >= 0) ++ crypto_unregister_all(sc->sc_cid); ++ if (sc->sc_chnfifo) { ++ for (i = 0; i < sc->sc_num_channels; i++) ++ if (sc->sc_chnfifo[i]) ++ kfree(sc->sc_chnfifo[i]); ++ kfree(sc->sc_chnfifo); ++ } ++ if (sc->sc_chnlastalg) ++ kfree(sc->sc_chnlastalg); ++ if (sc->sc_chnfifolock) ++ kfree(sc->sc_chnfifolock); ++ if (sc->sc_irq != -1) ++ free_irq(sc->sc_irq, sc); ++ if (sc->sc_base_addr) ++ iounmap((void *) sc->sc_base_addr); ++ kfree(sc); ++ return 0; ++} ++ ++#ifdef CONFIG_PPC_MERGE ++static struct of_device_id talitos_match[] = { ++ { ++ .type = "crypto", ++ .compatible = "talitos", ++ }, ++ {}, ++}; ++ ++MODULE_DEVICE_TABLE(of, talitos_match); ++ ++static struct of_platform_driver talitos_driver = { ++ .name = DRV_NAME, ++ .match_table = talitos_match, ++ .probe = talitos_probe, ++ .remove = talitos_remove, ++}; ++ ++static int __init talitos_init(void) ++{ ++ return of_register_platform_driver(&talitos_driver); ++} ++ ++static void __exit talitos_exit(void) ++{ ++ of_unregister_platform_driver(&talitos_driver); ++} ++#else ++/* Structure for a platform device driver */ ++static struct platform_driver talitos_driver = { ++ .probe = talitos_probe, ++ .remove = talitos_remove, ++ .driver = { ++ .name = "fsl-sec2", ++ } ++}; ++ ++static int __init talitos_init(void) ++{ ++ return platform_driver_register(&talitos_driver); ++} ++ ++static void __exit talitos_exit(void) ++{ ++ platform_driver_unregister(&talitos_driver); ++} ++#endif ++ ++module_init(talitos_init); ++module_exit(talitos_exit); ++ ++MODULE_LICENSE("Dual BSD/GPL"); ++MODULE_AUTHOR("kim.phillips@freescale.com"); ++MODULE_DESCRIPTION("OCF driver for Freescale SEC (talitos)"); +diff -Nur linux-2.6.30.orig/crypto/ocf/talitos/talitos_dev.h linux-2.6.30/crypto/ocf/talitos/talitos_dev.h +--- linux-2.6.30.orig/crypto/ocf/talitos/talitos_dev.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/talitos/talitos_dev.h 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,277 @@ ++/* ++ * Freescale SEC (talitos) device dependent data structures ++ * ++ * Copyright (c) 2006 Freescale Semiconductor, Inc. ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions ++ * are met: ++ * ++ * 1. Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * 2. Redistributions in binary form must reproduce the above copyright ++ * notice, this list of conditions and the following disclaimer in the ++ * documentation and/or other materials provided with the distribution. ++ * 3. The name of the author may not be used to endorse or promote products ++ * derived from this software without specific prior written permission. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR ++ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES ++ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. ++ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, ++ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT ++ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, ++ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY ++ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT ++ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF ++ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. ++ * ++ */ ++ ++/* device ID register values */ ++#define TALITOS_ID_SEC_2_0 0x40 ++#define TALITOS_ID_SEC_2_1 0x40 /* cross ref with IP block revision reg */ ++ ++/* ++ * following num_channels, channel-fifo-depth, exec-unit-mask, and ++ * descriptor-types-mask are for forward-compatibility with openfirmware ++ * flat device trees ++ */ ++ ++/* ++ * num_channels : the number of channels available in each SEC version. ++ */ ++ ++/* n.b. this driver requires these values be a power of 2 */ ++#define TALITOS_NCHANNELS_SEC_1_0 4 ++#define TALITOS_NCHANNELS_SEC_1_2 1 ++#define TALITOS_NCHANNELS_SEC_2_0 4 ++#define TALITOS_NCHANNELS_SEC_2_01 4 ++#define TALITOS_NCHANNELS_SEC_2_1 4 ++#define TALITOS_NCHANNELS_SEC_2_4 4 ++ ++/* ++ * channel-fifo-depth : The number of descriptor ++ * pointers a channel fetch fifo can hold. ++ */ ++#define TALITOS_CHFIFOLEN_SEC_1_0 1 ++#define TALITOS_CHFIFOLEN_SEC_1_2 1 ++#define TALITOS_CHFIFOLEN_SEC_2_0 24 ++#define TALITOS_CHFIFOLEN_SEC_2_01 24 ++#define TALITOS_CHFIFOLEN_SEC_2_1 24 ++#define TALITOS_CHFIFOLEN_SEC_2_4 24 ++ ++/* ++ * exec-unit-mask : The bitmask representing what Execution Units (EUs) ++ * are available. EU information should be encoded following the SEC's ++ * EU_SEL0 bitfield documentation, i.e. as follows: ++ * ++ * bit 31 = set if SEC permits no-EU selection (should be always set) ++ * bit 30 = set if SEC has the ARC4 EU (AFEU) ++ * bit 29 = set if SEC has the des/3des EU (DEU) ++ * bit 28 = set if SEC has the message digest EU (MDEU) ++ * bit 27 = set if SEC has the random number generator EU (RNG) ++ * bit 26 = set if SEC has the public key EU (PKEU) ++ * bit 25 = set if SEC has the aes EU (AESU) ++ * bit 24 = set if SEC has the Kasumi EU (KEU) ++ * ++ */ ++#define TALITOS_HAS_EU_NONE (1<<0) ++#define TALITOS_HAS_EU_AFEU (1<<1) ++#define TALITOS_HAS_EU_DEU (1<<2) ++#define TALITOS_HAS_EU_MDEU (1<<3) ++#define TALITOS_HAS_EU_RNG (1<<4) ++#define TALITOS_HAS_EU_PKEU (1<<5) ++#define TALITOS_HAS_EU_AESU (1<<6) ++#define TALITOS_HAS_EU_KEU (1<<7) ++ ++/* the corresponding masks for each SEC version */ ++#define TALITOS_HAS_EUS_SEC_1_0 0x7f ++#define TALITOS_HAS_EUS_SEC_1_2 0x4d ++#define TALITOS_HAS_EUS_SEC_2_0 0x7f ++#define TALITOS_HAS_EUS_SEC_2_01 0x7f ++#define TALITOS_HAS_EUS_SEC_2_1 0xff ++#define TALITOS_HAS_EUS_SEC_2_4 0x7f ++ ++/* ++ * descriptor-types-mask : The bitmask representing what descriptors ++ * are available. Descriptor type information should be encoded ++ * following the SEC's Descriptor Header Dword DESC_TYPE field ++ * documentation, i.e. as follows: ++ * ++ * bit 0 = set if SEC supports the aesu_ctr_nonsnoop desc. type ++ * bit 1 = set if SEC supports the ipsec_esp descriptor type ++ * bit 2 = set if SEC supports the common_nonsnoop desc. type ++ * bit 3 = set if SEC supports the 802.11i AES ccmp desc. type ++ * bit 4 = set if SEC supports the hmac_snoop_no_afeu desc. type ++ * bit 5 = set if SEC supports the srtp descriptor type ++ * bit 6 = set if SEC supports the non_hmac_snoop_no_afeu desc.type ++ * bit 7 = set if SEC supports the pkeu_assemble descriptor type ++ * bit 8 = set if SEC supports the aesu_key_expand_output desc.type ++ * bit 9 = set if SEC supports the pkeu_ptmul descriptor type ++ * bit 10 = set if SEC supports the common_nonsnoop_afeu desc. type ++ * bit 11 = set if SEC supports the pkeu_ptadd_dbl descriptor type ++ * ++ * ..and so on and so forth. ++ */ ++#define TALITOS_HAS_DT_AESU_CTR_NONSNOOP (1<<0) ++#define TALITOS_HAS_DT_IPSEC_ESP (1<<1) ++#define TALITOS_HAS_DT_COMMON_NONSNOOP (1<<2) ++ ++/* the corresponding masks for each SEC version */ ++#define TALITOS_HAS_DESCTYPES_SEC_2_0 0x01010ebf ++#define TALITOS_HAS_DESCTYPES_SEC_2_1 0x012b0ebf ++ ++/* ++ * a TALITOS_xxx_HI address points to the low data bits (32-63) of the register ++ */ ++ ++/* global register offset addresses */ ++#define TALITOS_ID 0x1020 ++#define TALITOS_ID_HI 0x1024 ++#define TALITOS_MCR 0x1030 /* master control register */ ++#define TALITOS_MCR_HI 0x1038 /* master control register */ ++#define TALITOS_MCR_SWR 0x1 ++#define TALITOS_IMR 0x1008 /* interrupt mask register */ ++#define TALITOS_IMR_ALL 0x00010fff /* enable all interrupts mask */ ++#define TALITOS_IMR_ERRONLY 0x00010aaa /* enable error interrupts */ ++#define TALITOS_IMR_HI 0x100C /* interrupt mask register */ ++#define TALITOS_IMR_HI_ALL 0x00323333 /* enable all interrupts mask */ ++#define TALITOS_IMR_HI_ERRONLY 0x00222222 /* enable error interrupts */ ++#define TALITOS_ISR 0x1010 /* interrupt status register */ ++#define TALITOS_ISR_ERROR 0x00010faa /* errors mask */ ++#define TALITOS_ISR_DONE 0x00000055 /* channel(s) done mask */ ++#define TALITOS_ISR_HI 0x1014 /* interrupt status register */ ++#define TALITOS_ICR 0x1018 /* interrupt clear register */ ++#define TALITOS_ICR_HI 0x101C /* interrupt clear register */ ++ ++/* channel register address stride */ ++#define TALITOS_CH_OFFSET 0x100 ++ ++/* channel register offset addresses and bits */ ++#define TALITOS_CH_CCCR 0x1108 /* Crypto-Channel Config Register */ ++#define TALITOS_CH_CCCR_RESET 0x1 /* Channel Reset bit */ ++#define TALITOS_CH_CCCR_HI 0x110c /* Crypto-Channel Config Register */ ++#define TALITOS_CH_CCCR_HI_CDWE 0x10 /* Channel done writeback enable bit */ ++#define TALITOS_CH_CCCR_HI_NT 0x4 /* Notification type bit */ ++#define TALITOS_CH_CCCR_HI_CDIE 0x2 /* Channel Done Interrupt Enable bit */ ++#define TALITOS_CH_CCPSR 0x1110 /* Crypto-Channel Pointer Status Reg */ ++#define TALITOS_CH_CCPSR_HI 0x1114 /* Crypto-Channel Pointer Status Reg */ ++#define TALITOS_CH_FF 0x1148 /* Fetch FIFO */ ++#define TALITOS_CH_FF_HI 0x114c /* Fetch FIFO's FETCH_ADRS */ ++#define TALITOS_CH_CDPR 0x1140 /* Crypto-Channel Pointer Status Reg */ ++#define TALITOS_CH_CDPR_HI 0x1144 /* Crypto-Channel Pointer Status Reg */ ++#define TALITOS_CH_DESCBUF 0x1180 /* (thru 11bf) Crypto-Channel ++ * Descriptor Buffer (debug) */ ++ ++/* execution unit register offset addresses and bits */ ++#define TALITOS_DEUSR 0x2028 /* DEU status register */ ++#define TALITOS_DEUSR_HI 0x202c /* DEU status register */ ++#define TALITOS_DEUISR 0x2030 /* DEU interrupt status register */ ++#define TALITOS_DEUISR_HI 0x2034 /* DEU interrupt status register */ ++#define TALITOS_DEUICR 0x2038 /* DEU interrupt control register */ ++#define TALITOS_DEUICR_HI 0x203c /* DEU interrupt control register */ ++#define TALITOS_AESUISR 0x4030 /* AESU interrupt status register */ ++#define TALITOS_AESUISR_HI 0x4034 /* AESU interrupt status register */ ++#define TALITOS_AESUICR 0x4038 /* AESU interrupt control register */ ++#define TALITOS_AESUICR_HI 0x403c /* AESU interrupt control register */ ++#define TALITOS_MDEUISR 0x6030 /* MDEU interrupt status register */ ++#define TALITOS_MDEUISR_HI 0x6034 /* MDEU interrupt status register */ ++#define TALITOS_RNGSR 0xa028 /* RNG status register */ ++#define TALITOS_RNGSR_HI 0xa02c /* RNG status register */ ++#define TALITOS_RNGSR_HI_RD 0x1 /* RNG Reset done */ ++#define TALITOS_RNGSR_HI_OFL 0xff0000/* number of dwords in RNG output FIFO*/ ++#define TALITOS_RNGDSR 0xa010 /* RNG data size register */ ++#define TALITOS_RNGDSR_HI 0xa014 /* RNG data size register */ ++#define TALITOS_RNG_FIFO 0xa800 /* RNG FIFO - pool of random numbers */ ++#define TALITOS_RNGISR 0xa030 /* RNG Interrupt status register */ ++#define TALITOS_RNGISR_HI 0xa034 /* RNG Interrupt status register */ ++#define TALITOS_RNGRCR 0xa018 /* RNG Reset control register */ ++#define TALITOS_RNGRCR_HI 0xa01c /* RNG Reset control register */ ++#define TALITOS_RNGRCR_HI_SR 0x1 /* RNG RNGRCR:Software Reset */ ++ ++/* descriptor pointer entry */ ++struct talitos_desc_ptr { ++ u16 len; /* length */ ++ u8 extent; /* jump (to s/g link table) and extent */ ++ u8 res; /* reserved */ ++ u32 ptr; /* pointer */ ++}; ++ ++/* descriptor */ ++struct talitos_desc { ++ u32 hdr; /* header */ ++ u32 res; /* reserved */ ++ struct talitos_desc_ptr ptr[7]; /* ptr/len pair array */ ++}; ++ ++/* talitos descriptor header (hdr) bits */ ++ ++/* primary execution unit select */ ++#define TALITOS_SEL0_AFEU 0x10000000 ++#define TALITOS_SEL0_DEU 0x20000000 ++#define TALITOS_SEL0_MDEU 0x30000000 ++#define TALITOS_SEL0_RNG 0x40000000 ++#define TALITOS_SEL0_PKEU 0x50000000 ++#define TALITOS_SEL0_AESU 0x60000000 ++ ++/* primary execution unit mode (MODE0) and derivatives */ ++#define TALITOS_MODE0_AESU_CBC 0x00200000 ++#define TALITOS_MODE0_AESU_ENC 0x00100000 ++#define TALITOS_MODE0_DEU_CBC 0x00400000 ++#define TALITOS_MODE0_DEU_3DES 0x00200000 ++#define TALITOS_MODE0_DEU_ENC 0x00100000 ++#define TALITOS_MODE0_MDEU_INIT 0x01000000 /* init starting regs */ ++#define TALITOS_MODE0_MDEU_HMAC 0x00800000 ++#define TALITOS_MODE0_MDEU_PAD 0x00400000 /* PD */ ++#define TALITOS_MODE0_MDEU_MD5 0x00200000 ++#define TALITOS_MODE0_MDEU_SHA256 0x00100000 ++#define TALITOS_MODE0_MDEU_SHA1 0x00000000 /* SHA-160 */ ++#define TALITOS_MODE0_MDEU_MD5_HMAC \ ++ (TALITOS_MODE0_MDEU_MD5 | TALITOS_MODE0_MDEU_HMAC) ++#define TALITOS_MODE0_MDEU_SHA256_HMAC \ ++ (TALITOS_MODE0_MDEU_SHA256 | TALITOS_MODE0_MDEU_HMAC) ++#define TALITOS_MODE0_MDEU_SHA1_HMAC \ ++ (TALITOS_MODE0_MDEU_SHA1 | TALITOS_MODE0_MDEU_HMAC) ++ ++/* secondary execution unit select (SEL1) */ ++/* it's MDEU or nothing */ ++#define TALITOS_SEL1_MDEU 0x00030000 ++ ++/* secondary execution unit mode (MODE1) and derivatives */ ++#define TALITOS_MODE1_MDEU_INIT 0x00001000 /* init starting regs */ ++#define TALITOS_MODE1_MDEU_HMAC 0x00000800 ++#define TALITOS_MODE1_MDEU_PAD 0x00000400 /* PD */ ++#define TALITOS_MODE1_MDEU_MD5 0x00000200 ++#define TALITOS_MODE1_MDEU_SHA256 0x00000100 ++#define TALITOS_MODE1_MDEU_SHA1 0x00000000 /* SHA-160 */ ++#define TALITOS_MODE1_MDEU_MD5_HMAC \ ++ (TALITOS_MODE1_MDEU_MD5 | TALITOS_MODE1_MDEU_HMAC) ++#define TALITOS_MODE1_MDEU_SHA256_HMAC \ ++ (TALITOS_MODE1_MDEU_SHA256 | TALITOS_MODE1_MDEU_HMAC) ++#define TALITOS_MODE1_MDEU_SHA1_HMAC \ ++ (TALITOS_MODE1_MDEU_SHA1 | TALITOS_MODE1_MDEU_HMAC) ++ ++/* direction of overall data flow (DIR) */ ++#define TALITOS_DIR_OUTBOUND 0x00000000 ++#define TALITOS_DIR_INBOUND 0x00000002 ++ ++/* done notification (DN) */ ++#define TALITOS_DONE_NOTIFY 0x00000001 ++ ++/* descriptor types */ ++/* odd numbers here are valid on SEC2 and greater only (e.g. ipsec_esp) */ ++#define TD_TYPE_AESU_CTR_NONSNOOP (0 << 3) ++#define TD_TYPE_IPSEC_ESP (1 << 3) ++#define TD_TYPE_COMMON_NONSNOOP_NO_AFEU (2 << 3) ++#define TD_TYPE_HMAC_SNOOP_NO_AFEU (4 << 3) ++ ++#define TALITOS_HDR_DONE_BITS 0xff000000 ++ ++#define DPRINTF(a...) do { \ ++ if (debug) { \ ++ printk("%s: ", sc ? \ ++ device_get_nameunit(sc->sc_cdev) : "talitos"); \ ++ printk(a); \ ++ } \ ++ } while (0) +diff -Nur linux-2.6.30.orig/crypto/ocf/talitos/talitos_soft.h linux-2.6.30/crypto/ocf/talitos/talitos_soft.h +--- linux-2.6.30.orig/crypto/ocf/talitos/talitos_soft.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/talitos/talitos_soft.h 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,77 @@ ++/* ++ * Freescale SEC data structures for integration with ocf-linux ++ * ++ * Copyright (c) 2006 Freescale Semiconductor, Inc. ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions ++ * are met: ++ * ++ * 1. Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * 2. Redistributions in binary form must reproduce the above copyright ++ * notice, this list of conditions and the following disclaimer in the ++ * documentation and/or other materials provided with the distribution. ++ * 3. The name of the author may not be used to endorse or promote products ++ * derived from this software without specific prior written permission. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR ++ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES ++ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. ++ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, ++ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT ++ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, ++ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY ++ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT ++ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF ++ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. ++ */ ++ ++/* ++ * paired descriptor and associated crypto operation ++ */ ++struct desc_cryptop_pair { ++ struct talitos_desc cf_desc; /* descriptor ptr */ ++ struct cryptop *cf_crp; /* cryptop ptr */ ++}; ++ ++/* ++ * Holds data specific to a single talitos device. ++ */ ++struct talitos_softc { ++ softc_device_decl sc_cdev; ++ struct platform_device *sc_dev; /* device backpointer */ ++ ocf_iomem_t sc_base_addr; ++ int sc_irq; ++ int sc_num; /* if we have multiple chips */ ++ int32_t sc_cid; /* crypto tag */ ++ u64 sc_chiprev; /* major/minor chip revision */ ++ int sc_nsessions; ++ struct talitos_session *sc_sessions; ++ int sc_num_channels;/* number of crypto channels */ ++ int sc_chfifo_len; /* channel fetch fifo len */ ++ int sc_exec_units; /* execution units mask */ ++ int sc_desc_types; /* descriptor types mask */ ++ /* ++ * mutual exclusion for intra-channel resources, e.g. fetch fifos ++ * the last entry is a meta-channel lock used by the channel scheduler ++ */ ++ spinlock_t *sc_chnfifolock; ++ /* sc_chnlastalgo contains last algorithm for that channel */ ++ int *sc_chnlastalg; ++ /* sc_chnfifo holds pending descriptor--crypto operation pairs */ ++ struct desc_cryptop_pair **sc_chnfifo; ++}; ++ ++struct talitos_session { ++ u_int32_t ses_used; ++ u_int32_t ses_klen; /* key length in bits */ ++ u_int32_t ses_key[8]; /* DES/3DES/AES key */ ++ u_int32_t ses_hmac[5]; /* hmac inner state */ ++ u_int32_t ses_hmac_len; /* hmac length */ ++ u_int32_t ses_iv[4]; /* DES/3DES/AES iv */ ++ u_int32_t ses_mlen; /* desired hash result len (12=ipsec or 16) */ ++}; ++ ++#define TALITOS_SESSION(sid) ((sid) & 0x0fffffff) ++#define TALITOS_SID(crd, sesn) (((crd) << 28) | ((sesn) & 0x0fffffff)) +diff -Nur linux-2.6.30.orig/crypto/ocf/uio.h linux-2.6.30/crypto/ocf/uio.h +--- linux-2.6.30.orig/crypto/ocf/uio.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/crypto/ocf/uio.h 2009-06-11 10:55:27.000000000 +0200 +@@ -0,0 +1,54 @@ ++#ifndef _OCF_UIO_H_ ++#define _OCF_UIO_H_ ++ ++#include <linux/uio.h> ++ ++/* ++ * The linux uio.h doesn't have all we need. To be fully api compatible ++ * with the BSD cryptodev, we need to keep this around. Perhaps this can ++ * be moved back into the linux/uio.h ++ * ++ * Linux port done by David McCullough <david_mccullough@securecomputing.com> ++ * Copyright (C) 2006-2007 David McCullough ++ * Copyright (C) 2004-2005 Intel Corporation. ++ * ++ * LICENSE TERMS ++ * ++ * The free distribution and use of this software in both source and binary ++ * form is allowed (with or without changes) provided that: ++ * ++ * 1. distributions of this source code include the above copyright ++ * notice, this list of conditions and the following disclaimer; ++ * ++ * 2. distributions in binary form include the above copyright ++ * notice, this list of conditions and the following disclaimer ++ * in the documentation and/or other associated materials; ++ * ++ * 3. the copyright holder's name is not used to endorse products ++ * built using this software without specific written permission. ++ * ++ * ALTERNATIVELY, provided that this notice is retained in full, this product ++ * may be distributed under the terms of the GNU General Public License (GPL), ++ * in which case the provisions of the GPL apply INSTEAD OF those given above. ++ * ++ * DISCLAIMER ++ * ++ * This software is provided 'as is' with no explicit or implied warranties ++ * in respect of its properties, including, but not limited to, correctness ++ * and/or fitness for purpose. ++ * --------------------------------------------------------------------------- ++ */ ++ ++struct uio { ++ struct iovec *uio_iov; ++ int uio_iovcnt; ++ off_t uio_offset; ++ int uio_resid; ++#if 0 ++ enum uio_seg uio_segflg; ++ enum uio_rw uio_rw; ++ struct thread *uio_td; ++#endif ++}; ++ ++#endif +diff -Nur linux-2.6.30.orig/drivers/char/random.c linux-2.6.30/drivers/char/random.c +--- linux-2.6.30.orig/drivers/char/random.c 2009-06-10 05:05:27.000000000 +0200 ++++ linux-2.6.30/drivers/char/random.c 2009-06-11 10:55:27.000000000 +0200 +@@ -129,6 +129,9 @@ + * unsigned int value); + * void add_interrupt_randomness(int irq); + * ++ * void random_input_words(__u32 *buf, size_t wordcount, int ent_count) ++ * int random_input_wait(void); ++ * + * add_input_randomness() uses the input layer interrupt timing, as well as + * the event type information from the hardware. + * +@@ -140,6 +143,13 @@ + * a better measure, since the timing of the disk interrupts are more + * unpredictable. + * ++ * random_input_words() just provides a raw block of entropy to the input ++ * pool, such as from a hardware entropy generator. ++ * ++ * random_input_wait() suspends the caller until such time as the ++ * entropy pool falls below the write threshold, and returns a count of how ++ * much entropy (in bits) is needed to sustain the pool. ++ * + * All of these routines try to estimate how many bits of randomness a + * particular randomness source. They do this by keeping track of the + * first and second order deltas of the event timings. +@@ -712,6 +722,61 @@ + } + #endif + ++/* ++ * random_input_words - add bulk entropy to pool ++ * ++ * @buf: buffer to add ++ * @wordcount: number of __u32 words to add ++ * @ent_count: total amount of entropy (in bits) to credit ++ * ++ * this provides bulk input of entropy to the input pool ++ * ++ */ ++void random_input_words(__u32 *buf, size_t wordcount, int ent_count) ++{ ++ mix_pool_bytes(&input_pool, buf, wordcount*4); ++ ++ credit_entropy_bits(&input_pool, ent_count); ++ ++ DEBUG_ENT("crediting %d bits => %d\n", ++ ent_count, input_pool.entropy_count); ++ /* ++ * Wake up waiting processes if we have enough ++ * entropy. ++ */ ++ if (input_pool.entropy_count >= random_read_wakeup_thresh) ++ wake_up_interruptible(&random_read_wait); ++} ++EXPORT_SYMBOL(random_input_words); ++ ++/* ++ * random_input_wait - wait until random needs entropy ++ * ++ * this function sleeps until the /dev/random subsystem actually ++ * needs more entropy, and then return the amount of entropy ++ * that it would be nice to have added to the system. ++ */ ++int random_input_wait(void) ++{ ++ int count; ++ ++ wait_event_interruptible(random_write_wait, ++ input_pool.entropy_count < random_write_wakeup_thresh); ++ ++ count = random_write_wakeup_thresh - input_pool.entropy_count; ++ ++ /* likely we got woken up due to a signal */ ++ if (count <= 0) count = random_read_wakeup_thresh; ++ ++ DEBUG_ENT("requesting %d bits from input_wait()er %d<%d\n", ++ count, ++ input_pool.entropy_count, random_write_wakeup_thresh); ++ ++ return count; ++} ++EXPORT_SYMBOL(random_input_wait); ++ ++ + #define EXTRACT_SIZE 10 + + /********************************************************************* +diff -Nur linux-2.6.30.orig/fs/fcntl.c linux-2.6.30/fs/fcntl.c +--- linux-2.6.30.orig/fs/fcntl.c 2009-06-10 05:05:27.000000000 +0200 ++++ linux-2.6.30/fs/fcntl.c 2009-06-11 10:55:27.000000000 +0200 +@@ -142,6 +142,7 @@ + } + return ret; + } ++EXPORT_SYMBOL(sys_dup); + + #define SETFL_MASK (O_APPEND | O_NONBLOCK | O_NDELAY | O_DIRECT | O_NOATIME) + +diff -Nur linux-2.6.30.orig/include/linux/miscdevice.h linux-2.6.30/include/linux/miscdevice.h +--- linux-2.6.30.orig/include/linux/miscdevice.h 2009-06-10 05:05:27.000000000 +0200 ++++ linux-2.6.30/include/linux/miscdevice.h 2009-06-11 10:55:27.000000000 +0200 +@@ -12,6 +12,7 @@ + #define APOLLO_MOUSE_MINOR 7 + #define PC110PAD_MINOR 9 + /*#define ADB_MOUSE_MINOR 10 FIXME OBSOLETE */ ++#define CRYPTODEV_MINOR 70 /* /dev/crypto */ + #define WATCHDOG_MINOR 130 /* Watchdog timer */ + #define TEMP_MINOR 131 /* Temperature Sensor */ + #define RTC_MINOR 135 +diff -Nur linux-2.6.30.orig/include/linux/random.h linux-2.6.30/include/linux/random.h +--- linux-2.6.30.orig/include/linux/random.h 2009-06-10 05:05:27.000000000 +0200 ++++ linux-2.6.30/include/linux/random.h 2009-06-11 10:55:27.000000000 +0200 +@@ -34,6 +34,30 @@ + /* Clear the entropy pool and associated counters. (Superuser only.) */ + #define RNDCLEARPOOL _IO( 'R', 0x06 ) + ++#ifdef CONFIG_FIPS_RNG ++ ++/* Size of seed value - equal to AES blocksize */ ++#define AES_BLOCK_SIZE_BYTES 16 ++#define SEED_SIZE_BYTES AES_BLOCK_SIZE_BYTES ++/* Size of AES key */ ++#define KEY_SIZE_BYTES 16 ++ ++/* ioctl() structure used by FIPS 140-2 Tests */ ++struct rand_fips_test { ++ unsigned char key[KEY_SIZE_BYTES]; /* Input */ ++ unsigned char datetime[SEED_SIZE_BYTES]; /* Input */ ++ unsigned char seed[SEED_SIZE_BYTES]; /* Input */ ++ unsigned char result[SEED_SIZE_BYTES]; /* Output */ ++}; ++ ++/* FIPS 140-2 RNG Variable Seed Test. (Superuser only.) */ ++#define RNDFIPSVST _IOWR('R', 0x10, struct rand_fips_test) ++ ++/* FIPS 140-2 RNG Monte Carlo Test. (Superuser only.) */ ++#define RNDFIPSMCT _IOWR('R', 0x11, struct rand_fips_test) ++ ++#endif /* #ifdef CONFIG_FIPS_RNG */ ++ + struct rand_pool_info { + int entropy_count; + int buf_size; +@@ -50,6 +74,10 @@ + unsigned int value); + extern void add_interrupt_randomness(int irq); + ++extern void random_input_words(__u32 *buf, size_t wordcount, int ent_count); ++extern int random_input_wait(void); ++#define HAS_RANDOM_INPUT_WAIT 1 ++ + extern void get_random_bytes(void *buf, int nbytes); + void generate_random_uuid(unsigned char uuid_out[16]); + diff --git a/target/linux/patches/2.6.31.4/startup.patch b/target/linux/patches/2.6.31.4/startup.patch new file mode 100644 index 000000000..25302f8d9 --- /dev/null +++ b/target/linux/patches/2.6.31.4/startup.patch @@ -0,0 +1,20 @@ +diff -Nur linux-2.6.30.5.orig/init/main.c linux-2.6.30.5/init/main.c +--- linux-2.6.30.5.orig/init/main.c 2009-08-16 23:19:38.000000000 +0200 ++++ linux-2.6.30.5/init/main.c 2009-08-22 20:26:39.000000000 +0200 +@@ -811,7 +811,7 @@ + numa_default_policy(); + + if (sys_open((const char __user *) "/dev/console", O_RDWR, 0) < 0) +- printk(KERN_WARNING "Warning: unable to open an initial console.\n"); ++ printk(KERN_WARNING "Starting Linux (build with OpenADK).\n"); + + (void) sys_dup(0); + (void) sys_dup(0); +@@ -835,6 +835,7 @@ + printk(KERN_WARNING "Failed to execute %s. Attempting " + "defaults...\n", execute_command); + } ++ run_init_process("/init"); + run_init_process("/sbin/init"); + run_init_process("/etc/init"); + run_init_process("/bin/init"); diff --git a/target/linux/patches/2.6.31.4/swconfig.patch b/target/linux/patches/2.6.31.4/swconfig.patch new file mode 100644 index 000000000..3297bb116 --- /dev/null +++ b/target/linux/patches/2.6.31.4/swconfig.patch @@ -0,0 +1,1075 @@ +diff -Nur linux-2.6.30.orig/drivers/net/phy/Kconfig linux-2.6.30/drivers/net/phy/Kconfig +--- linux-2.6.30.orig/drivers/net/phy/Kconfig 2009-06-10 05:05:27.000000000 +0200 ++++ linux-2.6.30/drivers/net/phy/Kconfig 2009-06-11 09:22:50.000000000 +0200 +@@ -13,6 +13,12 @@ + + if PHYLIB + ++config SWCONFIG ++ tristate "Switch configuration API" ++ ---help--- ++ Switch configuration API using netlink. This allows ++ you to configure the VLAN features of certain switches. ++ + comment "MII PHY device drivers" + + config MARVELL_PHY +diff -Nur linux-2.6.30.orig/drivers/net/phy/Makefile linux-2.6.30/drivers/net/phy/Makefile +--- linux-2.6.30.orig/drivers/net/phy/Makefile 2009-06-10 05:05:27.000000000 +0200 ++++ linux-2.6.30/drivers/net/phy/Makefile 2009-06-11 09:22:50.000000000 +0200 +@@ -3,6 +3,7 @@ + libphy-objs := phy.o phy_device.o mdio_bus.o + + obj-$(CONFIG_PHYLIB) += libphy.o ++obj-$(CONFIG_SWCONFIG) += swconfig.o + obj-$(CONFIG_MARVELL_PHY) += marvell.o + obj-$(CONFIG_DAVICOM_PHY) += davicom.o + obj-$(CONFIG_CICADA_PHY) += cicada.o +diff -Nur linux-2.6.30.orig/drivers/net/phy/swconfig.c linux-2.6.30/drivers/net/phy/swconfig.c +--- linux-2.6.30.orig/drivers/net/phy/swconfig.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/drivers/net/phy/swconfig.c 2009-06-11 09:22:50.000000000 +0200 +@@ -0,0 +1,872 @@ ++/* ++ * swconfig.c: Switch configuration API ++ * ++ * Copyright (C) 2008 Felix Fietkau <nbd@openwrt.org> ++ * ++ * This program is free software; you can redistribute it and/or ++ * modify it under the terms of the GNU General Public License ++ * as published by the Free Software Foundation; either version 2 ++ * of the License, or (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ */ ++ ++#include <linux/types.h> ++#include <linux/module.h> ++#include <linux/init.h> ++#include <linux/list.h> ++#include <linux/if.h> ++#include <linux/if_ether.h> ++#include <linux/capability.h> ++#include <linux/skbuff.h> ++#include <linux/switch.h> ++ ++//#define DEBUG 1 ++#ifdef DEBUG ++#define DPRINTF(format, ...) printk("%s: " format, __func__, ##__VA_ARGS__) ++#else ++#define DPRINTF(...) do {} while(0) ++#endif ++ ++MODULE_AUTHOR("Felix Fietkau <nbd@openwrt.org>"); ++MODULE_LICENSE("GPL"); ++ ++static int swdev_id = 0; ++static struct list_head swdevs; ++static spinlock_t swdevs_lock = SPIN_LOCK_UNLOCKED; ++struct swconfig_callback; ++ ++struct swconfig_callback ++{ ++ struct sk_buff *msg; ++ struct genlmsghdr *hdr; ++ struct genl_info *info; ++ int cmd; ++ ++ /* callback for filling in the message data */ ++ int (*fill)(struct swconfig_callback *cb, void *arg); ++ ++ /* callback for closing the message before sending it */ ++ int (*close)(struct swconfig_callback *cb, void *arg); ++ ++ struct nlattr *nest[4]; ++ int args[4]; ++}; ++ ++/* defaults */ ++ ++static int ++swconfig_get_vlan_ports(struct switch_dev *dev, struct switch_attr *attr, struct switch_val *val) ++{ ++ int ret; ++ if (val->port_vlan >= dev->vlans) ++ return -EINVAL; ++ ++ if (!dev->get_vlan_ports) ++ return -EOPNOTSUPP; ++ ++ ret = dev->get_vlan_ports(dev, val); ++ printk("SET PORTS %d\n", val->len); ++ return ret; ++} ++ ++static int ++swconfig_set_vlan_ports(struct switch_dev *dev, struct switch_attr *attr, struct switch_val *val) ++{ ++ int i; ++ ++ if (val->port_vlan >= dev->vlans) ++ return -EINVAL; ++ ++ /* validate ports */ ++ if (val->len > dev->ports) ++ return -EINVAL; ++ ++ for (i = 0; i < val->len; i++) { ++ if (val->value.ports[i].id >= dev->ports) ++ return -EINVAL; ++ } ++ ++ if (!dev->set_vlan_ports) ++ return -EOPNOTSUPP; ++ ++ printk("SET PORTS %d\n", val->len); ++ return dev->set_vlan_ports(dev, val); ++} ++ ++static int ++swconfig_apply_config(struct switch_dev *dev, struct switch_attr *attr, struct switch_val *val) ++{ ++ /* don't complain if not supported by the switch driver */ ++ if (!dev->apply_config) ++ return 0; ++ ++ return dev->apply_config(dev); ++} ++ ++ ++enum global_defaults { ++ GLOBAL_APPLY, ++}; ++ ++enum vlan_defaults { ++ VLAN_PORTS, ++}; ++ ++enum port_defaults { ++ PORT_LINK, ++}; ++ ++static struct switch_attr default_global[] = { ++ [GLOBAL_APPLY] = { ++ .type = SWITCH_TYPE_NOVAL, ++ .name = "apply", ++ .description = "Activate changes in the hardware", ++ .set = swconfig_apply_config, ++ } ++}; ++ ++static struct switch_attr default_port[] = { ++ [PORT_LINK] = { ++ .type = SWITCH_TYPE_INT, ++ .name = "link", ++ .description = "Current link speed", ++ } ++}; ++ ++static struct switch_attr default_vlan[] = { ++ [VLAN_PORTS] = { ++ .type = SWITCH_TYPE_PORTS, ++ .name = "ports", ++ .description = "VLAN port mapping", ++ .set = swconfig_set_vlan_ports, ++ .get = swconfig_get_vlan_ports, ++ }, ++}; ++ ++ ++static void swconfig_defaults_init(struct switch_dev *dev) ++{ ++ dev->def_global = 0; ++ dev->def_vlan = 0; ++ dev->def_port = 0; ++ ++ if (dev->get_vlan_ports || dev->set_vlan_ports) ++ set_bit(VLAN_PORTS, &dev->def_vlan); ++ ++ /* always present, can be no-op */ ++ set_bit(GLOBAL_APPLY, &dev->def_global); ++} ++ ++ ++static struct genl_family switch_fam = { ++ .id = GENL_ID_GENERATE, ++ .name = "switch", ++ .hdrsize = 0, ++ .version = 1, ++ .maxattr = SWITCH_ATTR_MAX, ++}; ++ ++static const struct nla_policy switch_policy[SWITCH_ATTR_MAX+1] = { ++ [SWITCH_ATTR_ID] = { .type = NLA_U32 }, ++ [SWITCH_ATTR_OP_ID] = { .type = NLA_U32 }, ++ [SWITCH_ATTR_OP_PORT] = { .type = NLA_U32 }, ++ [SWITCH_ATTR_OP_VLAN] = { .type = NLA_U32 }, ++ [SWITCH_ATTR_OP_VALUE_INT] = { .type = NLA_U32 }, ++ [SWITCH_ATTR_OP_VALUE_STR] = { .type = NLA_NUL_STRING }, ++ [SWITCH_ATTR_OP_VALUE_PORTS] = { .type = NLA_NESTED }, ++ [SWITCH_ATTR_TYPE] = { .type = NLA_U32 }, ++}; ++ ++static const struct nla_policy port_policy[SWITCH_PORT_ATTR_MAX+1] = { ++ [SWITCH_PORT_ID] = { .type = NLA_U32 }, ++ [SWITCH_PORT_FLAG_TAGGED] = { .type = NLA_FLAG }, ++}; ++ ++static inline void ++swconfig_lock(void) ++{ ++ spin_lock(&swdevs_lock); ++} ++ ++static inline void ++swconfig_unlock(void) ++{ ++ spin_unlock(&swdevs_lock); ++} ++ ++static struct switch_dev * ++swconfig_get_dev(struct genl_info *info) ++{ ++ struct switch_dev *dev = NULL; ++ struct switch_dev *p; ++ int id; ++ ++ if (!info->attrs[SWITCH_ATTR_ID]) ++ goto done; ++ ++ id = nla_get_u32(info->attrs[SWITCH_ATTR_ID]); ++ swconfig_lock(); ++ list_for_each_entry(p, &swdevs, dev_list) { ++ if (id != p->id) ++ continue; ++ ++ dev = p; ++ break; ++ } ++ if (dev) ++ spin_lock(&dev->lock); ++ else ++ DPRINTF("device %d not found\n", id); ++ swconfig_unlock(); ++done: ++ return dev; ++} ++ ++static inline void ++swconfig_put_dev(struct switch_dev *dev) ++{ ++ spin_unlock(&dev->lock); ++} ++ ++static int ++swconfig_dump_attr(struct swconfig_callback *cb, void *arg) ++{ ++ struct switch_attr *op = arg; ++ struct genl_info *info = cb->info; ++ struct sk_buff *msg = cb->msg; ++ int id = cb->args[0]; ++ void *hdr; ++ ++ hdr = genlmsg_put(msg, info->snd_pid, info->snd_seq, &switch_fam, ++ NLM_F_MULTI, SWITCH_CMD_NEW_ATTR); ++ if (IS_ERR(hdr)) ++ return -1; ++ ++ NLA_PUT_U32(msg, SWITCH_ATTR_OP_ID, id); ++ NLA_PUT_U32(msg, SWITCH_ATTR_OP_TYPE, op->type); ++ NLA_PUT_STRING(msg, SWITCH_ATTR_OP_NAME, op->name); ++ if (op->description) ++ NLA_PUT_STRING(msg, SWITCH_ATTR_OP_DESCRIPTION, ++ op->description); ++ ++ return genlmsg_end(msg, hdr); ++nla_put_failure: ++ genlmsg_cancel(msg, hdr); ++ return -EMSGSIZE; ++} ++ ++/* spread multipart messages across multiple message buffers */ ++static int ++swconfig_send_multipart(struct swconfig_callback *cb, void *arg) ++{ ++ struct genl_info *info = cb->info; ++ int restart = 0; ++ int err; ++ ++ do { ++ if (!cb->msg) { ++ cb->msg = nlmsg_new(NLMSG_GOODSIZE, GFP_KERNEL); ++ if (cb->msg == NULL) ++ goto error; ++ } ++ ++ if (!(cb->fill(cb, arg) < 0)) ++ break; ++ ++ /* fill failed, check if this was already the second attempt */ ++ if (restart) ++ goto error; ++ ++ /* try again in a new message, send the current one */ ++ restart = 1; ++ if (cb->close) { ++ if (cb->close(cb, arg) < 0) ++ goto error; ++ } ++ err = genlmsg_unicast(cb->msg, info->snd_pid); ++ cb->msg = NULL; ++ if (err < 0) ++ goto error; ++ ++ } while (restart); ++ ++ return 0; ++ ++error: ++ if (cb->msg) ++ nlmsg_free(cb->msg); ++ return -1; ++} ++ ++static int ++swconfig_list_attrs(struct sk_buff *skb, struct genl_info *info) ++{ ++ struct genlmsghdr *hdr = nlmsg_data(info->nlhdr); ++ const struct switch_attrlist *alist; ++ struct switch_dev *dev; ++ struct swconfig_callback cb; ++ int err = -EINVAL; ++ int i; ++ ++ /* defaults */ ++ struct switch_attr *def_list; ++ unsigned long *def_active; ++ int n_def; ++ ++ dev = swconfig_get_dev(info); ++ if (!dev) ++ return -EINVAL; ++ ++ switch(hdr->cmd) { ++ case SWITCH_CMD_LIST_GLOBAL: ++ alist = &dev->attr_global; ++ def_list = default_global; ++ def_active = &dev->def_global; ++ n_def = ARRAY_SIZE(default_global); ++ break; ++ case SWITCH_CMD_LIST_VLAN: ++ alist = &dev->attr_vlan; ++ def_list = default_vlan; ++ def_active = &dev->def_vlan; ++ n_def = ARRAY_SIZE(default_vlan); ++ break; ++ case SWITCH_CMD_LIST_PORT: ++ alist = &dev->attr_port; ++ def_list = default_port; ++ def_active = &dev->def_port; ++ n_def = ARRAY_SIZE(default_port); ++ break; ++ default: ++ WARN_ON(1); ++ goto out; ++ } ++ ++ memset(&cb, 0, sizeof(cb)); ++ cb.info = info; ++ cb.fill = swconfig_dump_attr; ++ for (i = 0; i < alist->n_attr; i++) { ++ if (alist->attr[i].disabled) ++ continue; ++ cb.args[0] = i; ++ err = swconfig_send_multipart(&cb, &alist->attr[i]); ++ if (err < 0) ++ goto error; ++ } ++ ++ /* defaults */ ++ for (i = 0; i < n_def; i++) { ++ if (!test_bit(i, def_active)) ++ continue; ++ cb.args[0] = SWITCH_ATTR_DEFAULTS_OFFSET + i; ++ err = swconfig_send_multipart(&cb, &def_list[i]); ++ if (err < 0) ++ goto error; ++ } ++ swconfig_put_dev(dev); ++ ++ if (!cb.msg) ++ return 0; ++ ++ return genlmsg_unicast(cb.msg, info->snd_pid); ++ ++error: ++ if (cb.msg) ++ nlmsg_free(cb.msg); ++out: ++ swconfig_put_dev(dev); ++ return err; ++} ++ ++static struct switch_attr * ++swconfig_lookup_attr(struct switch_dev *dev, struct genl_info *info, ++ struct switch_val *val) ++{ ++ struct genlmsghdr *hdr = nlmsg_data(info->nlhdr); ++ const struct switch_attrlist *alist; ++ struct switch_attr *attr = NULL; ++ int attr_id; ++ ++ /* defaults */ ++ struct switch_attr *def_list; ++ unsigned long *def_active; ++ int n_def; ++ ++ if (!info->attrs[SWITCH_ATTR_OP_ID]) ++ goto done; ++ ++ switch(hdr->cmd) { ++ case SWITCH_CMD_SET_GLOBAL: ++ case SWITCH_CMD_GET_GLOBAL: ++ alist = &dev->attr_global; ++ def_list = default_global; ++ def_active = &dev->def_global; ++ n_def = ARRAY_SIZE(default_global); ++ break; ++ case SWITCH_CMD_SET_VLAN: ++ case SWITCH_CMD_GET_VLAN: ++ alist = &dev->attr_vlan; ++ def_list = default_vlan; ++ def_active = &dev->def_vlan; ++ n_def = ARRAY_SIZE(default_vlan); ++ if (!info->attrs[SWITCH_ATTR_OP_VLAN]) ++ goto done; ++ val->port_vlan = nla_get_u32(info->attrs[SWITCH_ATTR_OP_VLAN]); ++ break; ++ case SWITCH_CMD_SET_PORT: ++ case SWITCH_CMD_GET_PORT: ++ alist = &dev->attr_port; ++ def_list = default_port; ++ def_active = &dev->def_port; ++ n_def = ARRAY_SIZE(default_port); ++ if (!info->attrs[SWITCH_ATTR_OP_PORT]) ++ goto done; ++ val->port_vlan = nla_get_u32(info->attrs[SWITCH_ATTR_OP_PORT]); ++ break; ++ default: ++ WARN_ON(1); ++ goto done; ++ } ++ ++ if (!alist) ++ goto done; ++ ++ attr_id = nla_get_u32(info->attrs[SWITCH_ATTR_OP_ID]); ++ if (attr_id >= SWITCH_ATTR_DEFAULTS_OFFSET) { ++ attr_id -= SWITCH_ATTR_DEFAULTS_OFFSET; ++ if (attr_id >= n_def) ++ goto done; ++ if (!test_bit(attr_id, def_active)) ++ goto done; ++ attr = &def_list[attr_id]; ++ } else { ++ if (attr_id >= alist->n_attr) ++ goto done; ++ attr = &alist->attr[attr_id]; ++ } ++ ++ if (attr->disabled) ++ attr = NULL; ++ ++done: ++ if (!attr) ++ DPRINTF("attribute lookup failed\n"); ++ val->attr = attr; ++ return attr; ++} ++ ++static int ++swconfig_parse_ports(struct sk_buff *msg, struct nlattr *head, ++ struct switch_val *val, int max) ++{ ++ struct nlattr *nla; ++ int rem; ++ ++ val->len = 0; ++ nla_for_each_nested(nla, head, rem) { ++ struct nlattr *tb[SWITCH_PORT_ATTR_MAX+1]; ++ struct switch_port *port = &val->value.ports[val->len]; ++ ++ if (val->len >= max) ++ return -EINVAL; ++ ++ if (nla_parse_nested(tb, SWITCH_PORT_ATTR_MAX, nla, ++ port_policy)) ++ return -EINVAL; ++ ++ if (!tb[SWITCH_PORT_ID]) ++ return -EINVAL; ++ ++ port->id = nla_get_u32(tb[SWITCH_PORT_ID]); ++ if (tb[SWITCH_PORT_FLAG_TAGGED]) ++ port->flags |= (1 << SWITCH_PORT_FLAG_TAGGED); ++ val->len++; ++ } ++ ++ return 0; ++} ++ ++static int ++swconfig_set_attr(struct sk_buff *skb, struct genl_info *info) ++{ ++ struct switch_attr *attr; ++ struct switch_dev *dev; ++ struct switch_val val; ++ int err = -EINVAL; ++ ++ dev = swconfig_get_dev(info); ++ if (!dev) ++ return -EINVAL; ++ ++ memset(&val, 0, sizeof(val)); ++ attr = swconfig_lookup_attr(dev, info, &val); ++ if (!attr || !attr->set) ++ goto error; ++ ++ val.attr = attr; ++ switch(attr->type) { ++ case SWITCH_TYPE_NOVAL: ++ break; ++ case SWITCH_TYPE_INT: ++ if (!info->attrs[SWITCH_ATTR_OP_VALUE_INT]) ++ goto error; ++ val.value.i = ++ nla_get_u32(info->attrs[SWITCH_ATTR_OP_VALUE_INT]); ++ break; ++ case SWITCH_TYPE_STRING: ++ if (!info->attrs[SWITCH_ATTR_OP_VALUE_STR]) ++ goto error; ++ val.value.s = ++ nla_data(info->attrs[SWITCH_ATTR_OP_VALUE_STR]); ++ break; ++ case SWITCH_TYPE_PORTS: ++ val.value.ports = dev->portbuf; ++ memset(dev->portbuf, 0, ++ sizeof(struct switch_port) * dev->ports); ++ ++ /* TODO: implement multipart? */ ++ if (info->attrs[SWITCH_ATTR_OP_VALUE_PORTS]) { ++ err = swconfig_parse_ports(skb, ++ info->attrs[SWITCH_ATTR_OP_VALUE_PORTS], &val, dev->ports); ++ if (err < 0) ++ goto error; ++ } else { ++ val.len = 0; ++ err = 0; ++ } ++ break; ++ default: ++ goto error; ++ } ++ ++ err = attr->set(dev, attr, &val); ++error: ++ swconfig_put_dev(dev); ++ return err; ++} ++ ++static int ++swconfig_close_portlist(struct swconfig_callback *cb, void *arg) ++{ ++ if (cb->nest[0]) ++ nla_nest_end(cb->msg, cb->nest[0]); ++ return 0; ++} ++ ++static int ++swconfig_send_port(struct swconfig_callback *cb, void *arg) ++{ ++ const struct switch_port *port = arg; ++ struct nlattr *p = NULL; ++ ++ if (!cb->nest[0]) { ++ cb->nest[0] = nla_nest_start(cb->msg, cb->cmd); ++ if (!cb->nest[0]) ++ return -1; ++ } ++ ++ p = nla_nest_start(cb->msg, SWITCH_ATTR_PORT); ++ if (!p) ++ goto error; ++ ++ NLA_PUT_U32(cb->msg, SWITCH_PORT_ID, port->id); ++ if (port->flags & (1 << SWITCH_PORT_FLAG_TAGGED)) ++ NLA_PUT_FLAG(cb->msg, SWITCH_PORT_FLAG_TAGGED); ++ ++ nla_nest_end(cb->msg, p); ++ return 0; ++ ++nla_put_failure: ++ nla_nest_cancel(cb->msg, p); ++error: ++ nla_nest_cancel(cb->msg, cb->nest[0]); ++ return -1; ++} ++ ++static int ++swconfig_send_ports(struct sk_buff **msg, struct genl_info *info, int attr, ++ const struct switch_val *val) ++{ ++ struct swconfig_callback cb; ++ int err = 0; ++ int i; ++ ++ if (!val->value.ports) ++ return -EINVAL; ++ ++ memset(&cb, 0, sizeof(cb)); ++ cb.cmd = attr; ++ cb.msg = *msg; ++ cb.info = info; ++ cb.fill = swconfig_send_port; ++ cb.close = swconfig_close_portlist; ++ ++ cb.nest[0] = nla_nest_start(cb.msg, cb.cmd); ++ for (i = 0; i < val->len; i++) { ++ err = swconfig_send_multipart(&cb, &val->value.ports[i]); ++ if (err) ++ goto done; ++ } ++ err = val->len; ++ swconfig_close_portlist(&cb, NULL); ++ *msg = cb.msg; ++ ++done: ++ return err; ++} ++ ++static int ++swconfig_get_attr(struct sk_buff *skb, struct genl_info *info) ++{ ++ struct genlmsghdr *hdr = nlmsg_data(info->nlhdr); ++ struct switch_attr *attr; ++ struct switch_dev *dev; ++ struct sk_buff *msg = NULL; ++ struct switch_val val; ++ int err = -EINVAL; ++ int cmd = hdr->cmd; ++ ++ dev = swconfig_get_dev(info); ++ if (!dev) ++ return -EINVAL; ++ ++ memset(&val, 0, sizeof(val)); ++ attr = swconfig_lookup_attr(dev, info, &val); ++ if (!attr || !attr->get) ++ goto error_dev; ++ ++ if (attr->type == SWITCH_TYPE_PORTS) { ++ val.value.ports = dev->portbuf; ++ memset(dev->portbuf, 0, ++ sizeof(struct switch_port) * dev->ports); ++ } ++ ++ err = attr->get(dev, attr, &val); ++ if (err) ++ goto error; ++ ++ msg = nlmsg_new(NLMSG_GOODSIZE, GFP_KERNEL); ++ if (!msg) ++ goto error; ++ ++ hdr = genlmsg_put(msg, info->snd_pid, info->snd_seq, &switch_fam, ++ 0, cmd); ++ if (IS_ERR(hdr)) ++ goto nla_put_failure; ++ ++ switch(attr->type) { ++ case SWITCH_TYPE_INT: ++ NLA_PUT_U32(msg, SWITCH_ATTR_OP_VALUE_INT, val.value.i); ++ break; ++ case SWITCH_TYPE_STRING: ++ NLA_PUT_STRING(msg, SWITCH_ATTR_OP_VALUE_STR, val.value.s); ++ break; ++ case SWITCH_TYPE_PORTS: ++ err = swconfig_send_ports(&msg, info, ++ SWITCH_ATTR_OP_VALUE_PORTS, &val); ++ if (err < 0) ++ goto nla_put_failure; ++ break; ++ default: ++ DPRINTF("invalid type in attribute\n"); ++ err = -EINVAL; ++ goto error; ++ } ++ err = genlmsg_end(msg, hdr); ++ if (err < 0) ++ goto nla_put_failure; ++ ++ swconfig_put_dev(dev); ++ return genlmsg_unicast(msg, info->snd_pid); ++ ++nla_put_failure: ++ if (msg) ++ nlmsg_free(msg); ++error_dev: ++ swconfig_put_dev(dev); ++error: ++ if (!err) ++ err = -ENOMEM; ++ return err; ++} ++ ++static int ++swconfig_send_switch(struct sk_buff *msg, u32 pid, u32 seq, int flags, ++ const struct switch_dev *dev) ++{ ++ void *hdr; ++ ++ hdr = genlmsg_put(msg, pid, seq, &switch_fam, flags, ++ SWITCH_CMD_NEW_ATTR); ++ if (IS_ERR(hdr)) ++ return -1; ++ ++ NLA_PUT_U32(msg, SWITCH_ATTR_ID, dev->id); ++ NLA_PUT_STRING(msg, SWITCH_ATTR_NAME, dev->name); ++ NLA_PUT_STRING(msg, SWITCH_ATTR_DEV_NAME, dev->devname); ++ NLA_PUT_U32(msg, SWITCH_ATTR_VLANS, dev->vlans); ++ NLA_PUT_U32(msg, SWITCH_ATTR_PORTS, dev->ports); ++ ++ return genlmsg_end(msg, hdr); ++nla_put_failure: ++ genlmsg_cancel(msg, hdr); ++ return -EMSGSIZE; ++} ++ ++static int swconfig_dump_switches(struct sk_buff *skb, ++ struct netlink_callback *cb) ++{ ++ struct switch_dev *dev; ++ int start = cb->args[0]; ++ int idx = 0; ++ ++ swconfig_lock(); ++ list_for_each_entry(dev, &swdevs, dev_list) { ++ if (++idx <= start) ++ continue; ++ if (swconfig_send_switch(skb, NETLINK_CB(cb->skb).pid, ++ cb->nlh->nlmsg_seq, NLM_F_MULTI, ++ dev) < 0) ++ break; ++ } ++ swconfig_unlock(); ++ cb->args[0] = idx; ++ ++ return skb->len; ++} ++ ++static int ++swconfig_done(struct netlink_callback *cb) ++{ ++ return 0; ++} ++ ++static struct genl_ops swconfig_ops[] = { ++ { ++ .cmd = SWITCH_CMD_LIST_GLOBAL, ++ .doit = swconfig_list_attrs, ++ .policy = switch_policy, ++ }, ++ { ++ .cmd = SWITCH_CMD_LIST_VLAN, ++ .doit = swconfig_list_attrs, ++ .policy = switch_policy, ++ }, ++ { ++ .cmd = SWITCH_CMD_LIST_PORT, ++ .doit = swconfig_list_attrs, ++ .policy = switch_policy, ++ }, ++ { ++ .cmd = SWITCH_CMD_GET_GLOBAL, ++ .doit = swconfig_get_attr, ++ .policy = switch_policy, ++ }, ++ { ++ .cmd = SWITCH_CMD_GET_VLAN, ++ .doit = swconfig_get_attr, ++ .policy = switch_policy, ++ }, ++ { ++ .cmd = SWITCH_CMD_GET_PORT, ++ .doit = swconfig_get_attr, ++ .policy = switch_policy, ++ }, ++ { ++ .cmd = SWITCH_CMD_SET_GLOBAL, ++ .doit = swconfig_set_attr, ++ .policy = switch_policy, ++ }, ++ { ++ .cmd = SWITCH_CMD_SET_VLAN, ++ .doit = swconfig_set_attr, ++ .policy = switch_policy, ++ }, ++ { ++ .cmd = SWITCH_CMD_SET_PORT, ++ .doit = swconfig_set_attr, ++ .policy = switch_policy, ++ }, ++ { ++ .cmd = SWITCH_CMD_GET_SWITCH, ++ .dumpit = swconfig_dump_switches, ++ .policy = switch_policy, ++ .done = swconfig_done, ++ } ++}; ++ ++int ++register_switch(struct switch_dev *dev, struct net_device *netdev) ++{ ++ INIT_LIST_HEAD(&dev->dev_list); ++ if (netdev) { ++ dev->netdev = netdev; ++ if (!dev->devname) ++ dev->devname = netdev->name; ++ } ++ BUG_ON(!dev->devname); ++ ++ if (dev->ports > 0) { ++ dev->portbuf = kzalloc(sizeof(struct switch_port) * dev->ports, ++ GFP_KERNEL); ++ if (!dev->portbuf) ++ return -ENOMEM; ++ } ++ dev->id = ++swdev_id; ++ swconfig_defaults_init(dev); ++ spin_lock_init(&dev->lock); ++ swconfig_lock(); ++ list_add(&dev->dev_list, &swdevs); ++ swconfig_unlock(); ++ ++ return 0; ++} ++EXPORT_SYMBOL_GPL(register_switch); ++ ++void ++unregister_switch(struct switch_dev *dev) ++{ ++ kfree(dev->portbuf); ++ spin_lock(&dev->lock); ++ swconfig_lock(); ++ list_del(&dev->dev_list); ++ swconfig_unlock(); ++} ++EXPORT_SYMBOL_GPL(unregister_switch); ++ ++ ++static int __init ++swconfig_init(void) ++{ ++ int i, err; ++ ++ INIT_LIST_HEAD(&swdevs); ++ err = genl_register_family(&switch_fam); ++ if (err) ++ return err; ++ ++ for (i = 0; i < ARRAY_SIZE(swconfig_ops); i++) { ++ err = genl_register_ops(&switch_fam, &swconfig_ops[i]); ++ if (err) ++ goto unregister; ++ } ++ ++ return 0; ++ ++unregister: ++ genl_unregister_family(&switch_fam); ++ return err; ++} ++ ++static void __exit ++swconfig_exit(void) ++{ ++ genl_unregister_family(&switch_fam); ++} ++ ++module_init(swconfig_init); ++module_exit(swconfig_exit); ++ +diff -Nur linux-2.6.30.orig/include/linux/switch.h linux-2.6.30/include/linux/switch.h +--- linux-2.6.30.orig/include/linux/switch.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/include/linux/switch.h 2009-06-11 09:22:50.000000000 +0200 +@@ -0,0 +1,168 @@ ++/* ++ * switch.h: Switch configuration API ++ * ++ * Copyright (C) 2008 Felix Fietkau <nbd@openwrt.org> ++ * ++ * This program is free software; you can redistribute it and/or ++ * modify it under the terms of the GNU General Public License ++ * as published by the Free Software Foundation; either version 2 ++ * of the License, or (at your option) any later version. ++ * ++ * This program is distributed in the hope that it will be useful, ++ * but WITHOUT ANY WARRANTY; without even the implied warranty of ++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ++ * GNU General Public License for more details. ++ */ ++ ++#ifndef __LINUX_SWITCH_H ++#define __LINUX_SWITCH_H ++ ++#include <linux/types.h> ++#include <linux/netdevice.h> ++#include <linux/netlink.h> ++#include <linux/genetlink.h> ++#ifndef __KERNEL__ ++#include <netlink/netlink.h> ++#include <netlink/genl/genl.h> ++#include <netlink/genl/ctrl.h> ++#else ++#include <net/genetlink.h> ++#endif ++ ++/* main attributes */ ++enum { ++ SWITCH_ATTR_UNSPEC, ++ /* global */ ++ SWITCH_ATTR_TYPE, ++ /* device */ ++ SWITCH_ATTR_ID, ++ SWITCH_ATTR_NAME, ++ SWITCH_ATTR_DEV_NAME, ++ SWITCH_ATTR_VLANS, ++ SWITCH_ATTR_PORTS, ++ /* attributes */ ++ SWITCH_ATTR_OP_ID, ++ SWITCH_ATTR_OP_TYPE, ++ SWITCH_ATTR_OP_NAME, ++ SWITCH_ATTR_OP_PORT, ++ SWITCH_ATTR_OP_VLAN, ++ SWITCH_ATTR_OP_VALUE_INT, ++ SWITCH_ATTR_OP_VALUE_STR, ++ SWITCH_ATTR_OP_VALUE_PORTS, ++ SWITCH_ATTR_OP_DESCRIPTION, ++ /* port lists */ ++ SWITCH_ATTR_PORT, ++ SWITCH_ATTR_MAX ++}; ++ ++/* commands */ ++enum { ++ SWITCH_CMD_UNSPEC, ++ SWITCH_CMD_GET_SWITCH, ++ SWITCH_CMD_NEW_ATTR, ++ SWITCH_CMD_LIST_GLOBAL, ++ SWITCH_CMD_GET_GLOBAL, ++ SWITCH_CMD_SET_GLOBAL, ++ SWITCH_CMD_LIST_PORT, ++ SWITCH_CMD_GET_PORT, ++ SWITCH_CMD_SET_PORT, ++ SWITCH_CMD_LIST_VLAN, ++ SWITCH_CMD_GET_VLAN, ++ SWITCH_CMD_SET_VLAN ++}; ++ ++/* data types */ ++enum switch_val_type { ++ SWITCH_TYPE_UNSPEC, ++ SWITCH_TYPE_INT, ++ SWITCH_TYPE_STRING, ++ SWITCH_TYPE_PORTS, ++ SWITCH_TYPE_NOVAL, ++}; ++ ++/* port nested attributes */ ++enum { ++ SWITCH_PORT_UNSPEC, ++ SWITCH_PORT_ID, ++ SWITCH_PORT_FLAG_TAGGED, ++ SWITCH_PORT_ATTR_MAX ++}; ++ ++#define SWITCH_ATTR_DEFAULTS_OFFSET 0x1000 ++ ++#ifdef __KERNEL__ ++ ++struct switch_dev; ++struct switch_op; ++struct switch_val; ++struct switch_attr; ++struct switch_attrlist; ++ ++int register_switch(struct switch_dev *dev, struct net_device *netdev); ++void unregister_switch(struct switch_dev *dev); ++ ++struct switch_attrlist { ++ /* filled in by the driver */ ++ int n_attr; ++ struct switch_attr *attr; ++}; ++ ++ ++struct switch_dev { ++ int id; ++ void *priv; ++ const char *name; ++ ++ /* NB: either devname or netdev must be set */ ++ const char *devname; ++ struct net_device *netdev; ++ ++ int ports; ++ int vlans; ++ int cpu_port; ++ struct switch_attrlist attr_global, attr_port, attr_vlan; ++ ++ spinlock_t lock; ++ struct switch_port *portbuf; ++ struct list_head dev_list; ++ unsigned long def_global, def_port, def_vlan; ++ ++ int (*get_vlan_ports)(struct switch_dev *dev, struct switch_val *val); ++ int (*set_vlan_ports)(struct switch_dev *dev, struct switch_val *val); ++ int (*apply_config)(struct switch_dev *dev); ++}; ++ ++struct switch_port { ++ u32 id; ++ u32 flags; ++}; ++ ++struct switch_val { ++ struct switch_attr *attr; ++ int port_vlan; ++ int len; ++ union { ++ const char *s; ++ u32 i; ++ struct switch_port *ports; ++ } value; ++}; ++ ++struct switch_attr { ++ int disabled; ++ int type; ++ const char *name; ++ const char *description; ++ ++ int (*set)(struct switch_dev *dev, struct switch_attr *attr, struct switch_val *val); ++ int (*get)(struct switch_dev *dev, struct switch_attr *attr, struct switch_val *val); ++ ++ /* for driver internal use */ ++ int id; ++ int ofs; ++ int max; ++}; ++ ++#endif ++ ++#endif diff --git a/target/linux/patches/2.6.31.4/yaffs2.patch b/target/linux/patches/2.6.31.4/yaffs2.patch new file mode 100644 index 000000000..a19ab9c84 --- /dev/null +++ b/target/linux/patches/2.6.31.4/yaffs2.patch @@ -0,0 +1,15066 @@ +diff -Nur linux-2.6.30.orig/fs/Kconfig linux-2.6.30/fs/Kconfig +--- linux-2.6.30.orig/fs/Kconfig 2009-06-10 05:05:27.000000000 +0200 ++++ linux-2.6.30/fs/Kconfig 2009-06-11 09:21:04.000000000 +0200 +@@ -162,6 +162,10 @@ + source "fs/befs/Kconfig" + source "fs/bfs/Kconfig" + source "fs/efs/Kconfig" ++ ++# Patched by YAFFS ++source "fs/yaffs2/Kconfig" ++ + source "fs/jffs2/Kconfig" + # UBIFS File system configuration + source "fs/ubifs/Kconfig" +diff -Nur linux-2.6.30.orig/fs/Makefile linux-2.6.30/fs/Makefile +--- linux-2.6.30.orig/fs/Makefile 2009-06-10 05:05:27.000000000 +0200 ++++ linux-2.6.30/fs/Makefile 2009-06-11 09:21:31.000000000 +0200 +@@ -124,3 +124,4 @@ + obj-$(CONFIG_BTRFS_FS) += btrfs/ + obj-$(CONFIG_GFS2_FS) += gfs2/ + obj-$(CONFIG_EXOFS_FS) += exofs/ ++obj-$(CONFIG_YAFFS_FS) += yaffs2/ +diff -Nur linux-2.6.30.orig/fs/Makefile.pre.yaffs linux-2.6.30/fs/Makefile.pre.yaffs +--- linux-2.6.30.orig/fs/Makefile.pre.yaffs 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/Makefile.pre.yaffs 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,126 @@ ++# ++# Makefile for the Linux filesystems. ++# ++# 14 Sep 2000, Christoph Hellwig <hch@infradead.org> ++# Rewritten to use lists instead of if-statements. ++# ++ ++obj-y := open.o read_write.o file_table.o super.o \ ++ char_dev.o stat.o exec.o pipe.o namei.o fcntl.o \ ++ ioctl.o readdir.o select.o fifo.o dcache.o inode.o \ ++ attr.o bad_inode.o file.o filesystems.o namespace.o \ ++ seq_file.o xattr.o libfs.o fs-writeback.o \ ++ pnode.o drop_caches.o splice.o sync.o utimes.o \ ++ stack.o ++ ++ifeq ($(CONFIG_BLOCK),y) ++obj-y += buffer.o bio.o block_dev.o direct-io.o mpage.o ioprio.o ++else ++obj-y += no-block.o ++endif ++ ++obj-$(CONFIG_BLK_DEV_INTEGRITY) += bio-integrity.o ++obj-y += notify/ ++obj-$(CONFIG_EPOLL) += eventpoll.o ++obj-$(CONFIG_ANON_INODES) += anon_inodes.o ++obj-$(CONFIG_SIGNALFD) += signalfd.o ++obj-$(CONFIG_TIMERFD) += timerfd.o ++obj-$(CONFIG_EVENTFD) += eventfd.o ++obj-$(CONFIG_AIO) += aio.o ++obj-$(CONFIG_FILE_LOCKING) += locks.o ++obj-$(CONFIG_COMPAT) += compat.o compat_ioctl.o ++ ++nfsd-$(CONFIG_NFSD) := nfsctl.o ++obj-y += $(nfsd-y) $(nfsd-m) ++ ++obj-$(CONFIG_BINFMT_AOUT) += binfmt_aout.o ++obj-$(CONFIG_BINFMT_EM86) += binfmt_em86.o ++obj-$(CONFIG_BINFMT_MISC) += binfmt_misc.o ++ ++# binfmt_script is always there ++obj-y += binfmt_script.o ++ ++obj-$(CONFIG_BINFMT_ELF) += binfmt_elf.o ++obj-$(CONFIG_COMPAT_BINFMT_ELF) += compat_binfmt_elf.o ++obj-$(CONFIG_BINFMT_ELF_FDPIC) += binfmt_elf_fdpic.o ++obj-$(CONFIG_BINFMT_SOM) += binfmt_som.o ++obj-$(CONFIG_BINFMT_FLAT) += binfmt_flat.o ++ ++obj-$(CONFIG_FS_MBCACHE) += mbcache.o ++obj-$(CONFIG_FS_POSIX_ACL) += posix_acl.o xattr_acl.o ++obj-$(CONFIG_NFS_COMMON) += nfs_common/ ++obj-$(CONFIG_GENERIC_ACL) += generic_acl.o ++ ++obj-$(CONFIG_QUOTA) += dquot.o ++obj-$(CONFIG_QFMT_V1) += quota_v1.o ++obj-$(CONFIG_QFMT_V2) += quota_v2.o ++obj-$(CONFIG_QUOTA_TREE) += quota_tree.o ++obj-$(CONFIG_QUOTACTL) += quota.o ++ ++obj-$(CONFIG_PROC_FS) += proc/ ++obj-y += partitions/ ++obj-$(CONFIG_SYSFS) += sysfs/ ++obj-$(CONFIG_CONFIGFS_FS) += configfs/ ++obj-y += devpts/ ++ ++obj-$(CONFIG_PROFILING) += dcookies.o ++obj-$(CONFIG_DLM) += dlm/ ++ ++# Do not add any filesystems before this line ++obj-$(CONFIG_REISERFS_FS) += reiserfs/ ++obj-$(CONFIG_EXT3_FS) += ext3/ # Before ext2 so root fs can be ext3 ++obj-$(CONFIG_EXT2_FS) += ext2/ ++# We place ext4 after ext2 so plain ext2 root fs's are mounted using ext2 ++# unless explicitly requested by rootfstype ++obj-$(CONFIG_EXT4_FS) += ext4/ ++obj-$(CONFIG_JBD) += jbd/ ++obj-$(CONFIG_JBD2) += jbd2/ ++obj-$(CONFIG_CRAMFS) += cramfs/ ++obj-$(CONFIG_SQUASHFS) += squashfs/ ++obj-y += ramfs/ ++obj-$(CONFIG_HUGETLBFS) += hugetlbfs/ ++obj-$(CONFIG_CODA_FS) += coda/ ++obj-$(CONFIG_MINIX_FS) += minix/ ++obj-$(CONFIG_FAT_FS) += fat/ ++obj-$(CONFIG_BFS_FS) += bfs/ ++obj-$(CONFIG_ISO9660_FS) += isofs/ ++obj-$(CONFIG_HFSPLUS_FS) += hfsplus/ # Before hfs to find wrapped HFS+ ++obj-$(CONFIG_HFS_FS) += hfs/ ++obj-$(CONFIG_ECRYPT_FS) += ecryptfs/ ++obj-$(CONFIG_VXFS_FS) += freevxfs/ ++obj-$(CONFIG_NFS_FS) += nfs/ ++obj-$(CONFIG_EXPORTFS) += exportfs/ ++obj-$(CONFIG_NFSD) += nfsd/ ++obj-$(CONFIG_LOCKD) += lockd/ ++obj-$(CONFIG_NLS) += nls/ ++obj-$(CONFIG_SYSV_FS) += sysv/ ++obj-$(CONFIG_SMB_FS) += smbfs/ ++obj-$(CONFIG_CIFS) += cifs/ ++obj-$(CONFIG_NCP_FS) += ncpfs/ ++obj-$(CONFIG_HPFS_FS) += hpfs/ ++obj-$(CONFIG_NTFS_FS) += ntfs/ ++obj-$(CONFIG_UFS_FS) += ufs/ ++obj-$(CONFIG_EFS_FS) += efs/ ++obj-$(CONFIG_JFFS2_FS) += jffs2/ ++obj-$(CONFIG_UBIFS_FS) += ubifs/ ++obj-$(CONFIG_AFFS_FS) += affs/ ++obj-$(CONFIG_ROMFS_FS) += romfs/ ++obj-$(CONFIG_QNX4FS_FS) += qnx4/ ++obj-$(CONFIG_AUTOFS_FS) += autofs/ ++obj-$(CONFIG_AUTOFS4_FS) += autofs4/ ++obj-$(CONFIG_ADFS_FS) += adfs/ ++obj-$(CONFIG_FUSE_FS) += fuse/ ++obj-$(CONFIG_UDF_FS) += udf/ ++obj-$(CONFIG_SUN_OPENPROMFS) += openpromfs/ ++obj-$(CONFIG_OMFS_FS) += omfs/ ++obj-$(CONFIG_JFS_FS) += jfs/ ++obj-$(CONFIG_XFS_FS) += xfs/ ++obj-$(CONFIG_9P_FS) += 9p/ ++obj-$(CONFIG_AFS_FS) += afs/ ++obj-$(CONFIG_BEFS_FS) += befs/ ++obj-$(CONFIG_HOSTFS) += hostfs/ ++obj-$(CONFIG_HPPFS) += hppfs/ ++obj-$(CONFIG_DEBUG_FS) += debugfs/ ++obj-$(CONFIG_OCFS2_FS) += ocfs2/ ++obj-$(CONFIG_BTRFS_FS) += btrfs/ ++obj-$(CONFIG_GFS2_FS) += gfs2/ +diff -Nur linux-2.6.30.orig/fs/yaffs2/devextras.h linux-2.6.30/fs/yaffs2/devextras.h +--- linux-2.6.30.orig/fs/yaffs2/devextras.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/devextras.h 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,196 @@ ++/* ++ * YAFFS: Yet another Flash File System . A NAND-flash specific file system. ++ * ++ * Copyright (C) 2002-2007 Aleph One Ltd. ++ * for Toby Churchill Ltd and Brightstar Engineering ++ * ++ * Created by Charles Manning <charles@aleph1.co.uk> ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU Lesser General Public License version 2.1 as ++ * published by the Free Software Foundation. ++ * ++ * Note: Only YAFFS headers are LGPL, YAFFS C code is covered by GPL. ++ */ ++ ++/* ++ * This file is just holds extra declarations of macros that would normally ++ * be providesd in the Linux kernel. These macros have been written from ++ * scratch but are functionally equivalent to the Linux ones. ++ * ++ */ ++ ++#ifndef __EXTRAS_H__ ++#define __EXTRAS_H__ ++ ++ ++#if !(defined __KERNEL__) ++ ++/* Definition of types */ ++typedef unsigned char __u8; ++typedef unsigned short __u16; ++typedef unsigned __u32; ++ ++#endif ++ ++/* ++ * This is a simple doubly linked list implementation that matches the ++ * way the Linux kernel doubly linked list implementation works. ++ */ ++ ++struct ylist_head { ++ struct ylist_head *next; /* next in chain */ ++ struct ylist_head *prev; /* previous in chain */ ++}; ++ ++ ++/* Initialise a static list */ ++#define YLIST_HEAD(name) \ ++struct ylist_head name = { &(name), &(name)} ++ ++ ++ ++/* Initialise a list head to an empty list */ ++#define YINIT_LIST_HEAD(p) \ ++do { \ ++ (p)->next = (p);\ ++ (p)->prev = (p); \ ++} while (0) ++ ++ ++/* Add an element to a list */ ++static __inline__ void ylist_add(struct ylist_head *newEntry, ++ struct ylist_head *list) ++{ ++ struct ylist_head *listNext = list->next; ++ ++ list->next = newEntry; ++ newEntry->prev = list; ++ newEntry->next = listNext; ++ listNext->prev = newEntry; ++ ++} ++ ++static __inline__ void ylist_add_tail(struct ylist_head *newEntry, ++ struct ylist_head *list) ++{ ++ struct ylist_head *listPrev = list->prev; ++ ++ list->prev = newEntry; ++ newEntry->next = list; ++ newEntry->prev = listPrev; ++ listPrev->next = newEntry; ++ ++} ++ ++ ++/* Take an element out of its current list, with or without ++ * reinitialising the links.of the entry*/ ++static __inline__ void ylist_del(struct ylist_head *entry) ++{ ++ struct ylist_head *listNext = entry->next; ++ struct ylist_head *listPrev = entry->prev; ++ ++ listNext->prev = listPrev; ++ listPrev->next = listNext; ++ ++} ++ ++static __inline__ void ylist_del_init(struct ylist_head *entry) ++{ ++ ylist_del(entry); ++ entry->next = entry->prev = entry; ++} ++ ++ ++/* Test if the list is empty */ ++static __inline__ int ylist_empty(struct ylist_head *entry) ++{ ++ return (entry->next == entry); ++} ++ ++ ++/* ylist_entry takes a pointer to a list entry and offsets it to that ++ * we can find a pointer to the object it is embedded in. ++ */ ++ ++ ++#define ylist_entry(entry, type, member) \ ++ ((type *)((char *)(entry)-(unsigned long)(&((type *)NULL)->member))) ++ ++ ++/* ylist_for_each and list_for_each_safe iterate over lists. ++ * ylist_for_each_safe uses temporary storage to make the list delete safe ++ */ ++ ++#define ylist_for_each(itervar, list) \ ++ for (itervar = (list)->next; itervar != (list); itervar = itervar->next) ++ ++#define ylist_for_each_safe(itervar, saveVar, list) \ ++ for (itervar = (list)->next, saveVar = (list)->next->next; \ ++ itervar != (list); itervar = saveVar, saveVar = saveVar->next) ++ ++ ++#if !(defined __KERNEL__) ++ ++ ++#ifndef WIN32 ++#include <sys/stat.h> ++#endif ++ ++ ++#ifdef CONFIG_YAFFS_PROVIDE_DEFS ++/* File types */ ++ ++ ++#define DT_UNKNOWN 0 ++#define DT_FIFO 1 ++#define DT_CHR 2 ++#define DT_DIR 4 ++#define DT_BLK 6 ++#define DT_REG 8 ++#define DT_LNK 10 ++#define DT_SOCK 12 ++#define DT_WHT 14 ++ ++ ++#ifndef WIN32 ++#include <sys/stat.h> ++#endif ++ ++/* ++ * Attribute flags. These should be or-ed together to figure out what ++ * has been changed! ++ */ ++#define ATTR_MODE 1 ++#define ATTR_UID 2 ++#define ATTR_GID 4 ++#define ATTR_SIZE 8 ++#define ATTR_ATIME 16 ++#define ATTR_MTIME 32 ++#define ATTR_CTIME 64 ++ ++struct iattr { ++ unsigned int ia_valid; ++ unsigned ia_mode; ++ unsigned ia_uid; ++ unsigned ia_gid; ++ unsigned ia_size; ++ unsigned ia_atime; ++ unsigned ia_mtime; ++ unsigned ia_ctime; ++ unsigned int ia_attr_flags; ++}; ++ ++#endif ++ ++#else ++ ++#include <linux/types.h> ++#include <linux/fs.h> ++#include <linux/stat.h> ++ ++#endif ++ ++ ++#endif +diff -Nur linux-2.6.30.orig/fs/yaffs2/Kconfig linux-2.6.30/fs/yaffs2/Kconfig +--- linux-2.6.30.orig/fs/yaffs2/Kconfig 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/Kconfig 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,156 @@ ++# ++# YAFFS file system configurations ++# ++ ++config YAFFS_FS ++ tristate "YAFFS2 file system support" ++ default n ++ depends on MTD_BLOCK ++ select YAFFS_YAFFS1 ++ select YAFFS_YAFFS2 ++ help ++ YAFFS2, or Yet Another Flash Filing System, is a filing system ++ optimised for NAND Flash chips. ++ ++ To compile the YAFFS2 file system support as a module, choose M ++ here: the module will be called yaffs2. ++ ++ If unsure, say N. ++ ++ Further information on YAFFS2 is available at ++ <http://www.aleph1.co.uk/yaffs/>. ++ ++config YAFFS_YAFFS1 ++ bool "512 byte / page devices" ++ depends on YAFFS_FS ++ default y ++ help ++ Enable YAFFS1 support -- yaffs for 512 byte / page devices ++ ++ Not needed for 2K-page devices. ++ ++ If unsure, say Y. ++ ++config YAFFS_9BYTE_TAGS ++ bool "Use older-style on-NAND data format with pageStatus byte" ++ depends on YAFFS_YAFFS1 ++ default n ++ help ++ ++ Older-style on-NAND data format has a "pageStatus" byte to record ++ chunk/page state. This byte is zero when the page is discarded. ++ Choose this option if you have existing on-NAND data using this ++ format that you need to continue to support. New data written ++ also uses the older-style format. Note: Use of this option ++ generally requires that MTD's oob layout be adjusted to use the ++ older-style format. See notes on tags formats and MTD versions ++ in yaffs_mtdif1.c. ++ ++ If unsure, say N. ++ ++config YAFFS_DOES_ECC ++ bool "Lets Yaffs do its own ECC" ++ depends on YAFFS_FS && YAFFS_YAFFS1 && !YAFFS_9BYTE_TAGS ++ default n ++ help ++ This enables Yaffs to use its own ECC functions instead of using ++ the ones from the generic MTD-NAND driver. ++ ++ If unsure, say N. ++ ++config YAFFS_ECC_WRONG_ORDER ++ bool "Use the same ecc byte order as Steven Hill's nand_ecc.c" ++ depends on YAFFS_FS && YAFFS_DOES_ECC && !YAFFS_9BYTE_TAGS ++ default n ++ help ++ This makes yaffs_ecc.c use the same ecc byte order as Steven ++ Hill's nand_ecc.c. If not set, then you get the same ecc byte ++ order as SmartMedia. ++ ++ If unsure, say N. ++ ++config YAFFS_YAFFS2 ++ bool "2048 byte (or larger) / page devices" ++ depends on YAFFS_FS ++ default y ++ help ++ Enable YAFFS2 support -- yaffs for >= 2K bytes per page devices ++ ++ If unsure, say Y. ++ ++config YAFFS_AUTO_YAFFS2 ++ bool "Autoselect yaffs2 format" ++ depends on YAFFS_YAFFS2 ++ default y ++ help ++ Without this, you need to explicitely use yaffs2 as the file ++ system type. With this, you can say "yaffs" and yaffs or yaffs2 ++ will be used depending on the device page size (yaffs on ++ 512-byte page devices, yaffs2 on 2K page devices). ++ ++ If unsure, say Y. ++ ++config YAFFS_DISABLE_LAZY_LOAD ++ bool "Disable lazy loading" ++ depends on YAFFS_YAFFS2 ++ default n ++ help ++ "Lazy loading" defers loading file details until they are ++ required. This saves mount time, but makes the first look-up ++ a bit longer. ++ ++ Lazy loading will only happen if enabled by this option being 'n' ++ and if the appropriate tags are available, else yaffs2 will ++ automatically fall back to immediate loading and do the right ++ thing. ++ ++ Lazy laoding will be required by checkpointing. ++ ++ Setting this to 'y' will disable lazy loading. ++ ++ If unsure, say N. ++ ++ ++config YAFFS_DISABLE_WIDE_TNODES ++ bool "Turn off wide tnodes" ++ depends on YAFFS_FS ++ default n ++ help ++ Wide tnodes are only used for NAND arrays >=32MB for 512-byte ++ page devices and >=128MB for 2k page devices. They use slightly ++ more RAM but are faster since they eliminate chunk group ++ searching. ++ ++ Setting this to 'y' will force tnode width to 16 bits and save ++ memory but make large arrays slower. ++ ++ If unsure, say N. ++ ++config YAFFS_ALWAYS_CHECK_CHUNK_ERASED ++ bool "Force chunk erase check" ++ depends on YAFFS_FS ++ default n ++ help ++ Normally YAFFS only checks chunks before writing until an erased ++ chunk is found. This helps to detect any partially written ++ chunks that might have happened due to power loss. ++ ++ Enabling this forces on the test that chunks are erased in flash ++ before writing to them. This takes more time but is potentially ++ a bit more secure. ++ ++ Suggest setting Y during development and ironing out driver ++ issues etc. Suggest setting to N if you want faster writing. ++ ++ If unsure, say Y. ++ ++config YAFFS_SHORT_NAMES_IN_RAM ++ bool "Cache short names in RAM" ++ depends on YAFFS_FS ++ default y ++ help ++ If this config is set, then short names are stored with the ++ yaffs_Object. This costs an extra 16 bytes of RAM per object, ++ but makes look-ups faster. ++ ++ If unsure, say Y. +diff -Nur linux-2.6.30.orig/fs/yaffs2/Makefile linux-2.6.30/fs/yaffs2/Makefile +--- linux-2.6.30.orig/fs/yaffs2/Makefile 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/Makefile 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,10 @@ ++# ++# Makefile for the linux YAFFS filesystem routines. ++# ++ ++obj-$(CONFIG_YAFFS_FS) += yaffs.o ++ ++yaffs-y := yaffs_ecc.o yaffs_fs.o yaffs_guts.o yaffs_checkptrw.o ++yaffs-y += yaffs_packedtags1.o yaffs_packedtags2.o yaffs_nand.o yaffs_qsort.o ++yaffs-y += yaffs_tagscompat.o yaffs_tagsvalidity.o ++yaffs-y += yaffs_mtdif.o yaffs_mtdif1.o yaffs_mtdif2.o +diff -Nur linux-2.6.30.orig/fs/yaffs2/moduleconfig.h linux-2.6.30/fs/yaffs2/moduleconfig.h +--- linux-2.6.30.orig/fs/yaffs2/moduleconfig.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/moduleconfig.h 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,65 @@ ++/* ++ * YAFFS: Yet another Flash File System . A NAND-flash specific file system. ++ * ++ * Copyright (C) 2002-2007 Aleph One Ltd. ++ * for Toby Churchill Ltd and Brightstar Engineering ++ * ++ * Created by Martin Fouts <Martin.Fouts@palmsource.com> ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU Lesser General Public License version 2.1 as ++ * published by the Free Software Foundation. ++ * ++ * Note: Only YAFFS headers are LGPL, YAFFS C code is covered by GPL. ++ */ ++ ++#ifndef __YAFFS_CONFIG_H__ ++#define __YAFFS_CONFIG_H__ ++ ++#ifdef YAFFS_OUT_OF_TREE ++ ++/* DO NOT UNSET THESE THREE. YAFFS2 will not compile if you do. */ ++#define CONFIG_YAFFS_FS ++#define CONFIG_YAFFS_YAFFS1 ++#define CONFIG_YAFFS_YAFFS2 ++ ++/* These options are independent of each other. Select those that matter. */ ++ ++/* Default: Not selected */ ++/* Meaning: Yaffs does its own ECC, rather than using MTD ECC */ ++/* #define CONFIG_YAFFS_DOES_ECC */ ++ ++/* Default: Not selected */ ++/* Meaning: ECC byte order is 'wrong'. Only meaningful if */ ++/* CONFIG_YAFFS_DOES_ECC is set */ ++/* #define CONFIG_YAFFS_ECC_WRONG_ORDER */ ++ ++/* Default: Selected */ ++/* Meaning: Disables testing whether chunks are erased before writing to them*/ ++#define CONFIG_YAFFS_DISABLE_CHUNK_ERASED_CHECK ++ ++/* Default: Selected */ ++/* Meaning: Cache short names, taking more RAM, but faster look-ups */ ++#define CONFIG_YAFFS_SHORT_NAMES_IN_RAM ++ ++/* Default: 10 */ ++/* Meaning: set the count of blocks to reserve for checkpointing */ ++#define CONFIG_YAFFS_CHECKPOINT_RESERVED_BLOCKS 10 ++ ++/* ++Older-style on-NAND data format has a "pageStatus" byte to record ++chunk/page state. This byte is zeroed when the page is discarded. ++Choose this option if you have existing on-NAND data in this format ++that you need to continue to support. New data written also uses the ++older-style format. ++Note: Use of this option generally requires that MTD's oob layout be ++adjusted to use the older-style format. See notes on tags formats and ++MTD versions in yaffs_mtdif1.c. ++*/ ++/* Default: Not selected */ ++/* Meaning: Use older-style on-NAND data format with pageStatus byte */ ++/* #define CONFIG_YAFFS_9BYTE_TAGS */ ++ ++#endif /* YAFFS_OUT_OF_TREE */ ++ ++#endif /* __YAFFS_CONFIG_H__ */ +diff -Nur linux-2.6.30.orig/fs/yaffs2/yaffs_checkptrw.c linux-2.6.30/fs/yaffs2/yaffs_checkptrw.c +--- linux-2.6.30.orig/fs/yaffs2/yaffs_checkptrw.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/yaffs_checkptrw.c 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,394 @@ ++/* ++ * YAFFS: Yet Another Flash File System. A NAND-flash specific file system. ++ * ++ * Copyright (C) 2002-2007 Aleph One Ltd. ++ * for Toby Churchill Ltd and Brightstar Engineering ++ * ++ * Created by Charles Manning <charles@aleph1.co.uk> ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License version 2 as ++ * published by the Free Software Foundation. ++ */ ++ ++const char *yaffs_checkptrw_c_version = ++ "$Id: yaffs_checkptrw.c,v 1.18 2009-03-06 17:20:49 wookey Exp $"; ++ ++ ++#include "yaffs_checkptrw.h" ++#include "yaffs_getblockinfo.h" ++ ++static int yaffs_CheckpointSpaceOk(yaffs_Device *dev) ++{ ++ int blocksAvailable = dev->nErasedBlocks - dev->nReservedBlocks; ++ ++ T(YAFFS_TRACE_CHECKPOINT, ++ (TSTR("checkpt blocks available = %d" TENDSTR), ++ blocksAvailable)); ++ ++ return (blocksAvailable <= 0) ? 0 : 1; ++} ++ ++ ++static int yaffs_CheckpointErase(yaffs_Device *dev) ++{ ++ int i; ++ ++ if (!dev->eraseBlockInNAND) ++ return 0; ++ T(YAFFS_TRACE_CHECKPOINT, (TSTR("checking blocks %d to %d"TENDSTR), ++ dev->internalStartBlock, dev->internalEndBlock)); ++ ++ for (i = dev->internalStartBlock; i <= dev->internalEndBlock; i++) { ++ yaffs_BlockInfo *bi = yaffs_GetBlockInfo(dev, i); ++ if (bi->blockState == YAFFS_BLOCK_STATE_CHECKPOINT) { ++ T(YAFFS_TRACE_CHECKPOINT, (TSTR("erasing checkpt block %d"TENDSTR), i)); ++ if (dev->eraseBlockInNAND(dev, i - dev->blockOffset /* realign */)) { ++ bi->blockState = YAFFS_BLOCK_STATE_EMPTY; ++ dev->nErasedBlocks++; ++ dev->nFreeChunks += dev->nChunksPerBlock; ++ } else { ++ dev->markNANDBlockBad(dev, i); ++ bi->blockState = YAFFS_BLOCK_STATE_DEAD; ++ } ++ } ++ } ++ ++ dev->blocksInCheckpoint = 0; ++ ++ return 1; ++} ++ ++ ++static void yaffs_CheckpointFindNextErasedBlock(yaffs_Device *dev) ++{ ++ int i; ++ int blocksAvailable = dev->nErasedBlocks - dev->nReservedBlocks; ++ T(YAFFS_TRACE_CHECKPOINT, ++ (TSTR("allocating checkpt block: erased %d reserved %d avail %d next %d "TENDSTR), ++ dev->nErasedBlocks, dev->nReservedBlocks, blocksAvailable, dev->checkpointNextBlock)); ++ ++ if (dev->checkpointNextBlock >= 0 && ++ dev->checkpointNextBlock <= dev->internalEndBlock && ++ blocksAvailable > 0) { ++ ++ for (i = dev->checkpointNextBlock; i <= dev->internalEndBlock; i++) { ++ yaffs_BlockInfo *bi = yaffs_GetBlockInfo(dev, i); ++ if (bi->blockState == YAFFS_BLOCK_STATE_EMPTY) { ++ dev->checkpointNextBlock = i + 1; ++ dev->checkpointCurrentBlock = i; ++ T(YAFFS_TRACE_CHECKPOINT, (TSTR("allocating checkpt block %d"TENDSTR), i)); ++ return; ++ } ++ } ++ } ++ T(YAFFS_TRACE_CHECKPOINT, (TSTR("out of checkpt blocks"TENDSTR))); ++ ++ dev->checkpointNextBlock = -1; ++ dev->checkpointCurrentBlock = -1; ++} ++ ++static void yaffs_CheckpointFindNextCheckpointBlock(yaffs_Device *dev) ++{ ++ int i; ++ yaffs_ExtendedTags tags; ++ ++ T(YAFFS_TRACE_CHECKPOINT, (TSTR("find next checkpt block: start: blocks %d next %d" TENDSTR), ++ dev->blocksInCheckpoint, dev->checkpointNextBlock)); ++ ++ if (dev->blocksInCheckpoint < dev->checkpointMaxBlocks) ++ for (i = dev->checkpointNextBlock; i <= dev->internalEndBlock; i++) { ++ int chunk = i * dev->nChunksPerBlock; ++ int realignedChunk = chunk - dev->chunkOffset; ++ ++ dev->readChunkWithTagsFromNAND(dev, realignedChunk, ++ NULL, &tags); ++ T(YAFFS_TRACE_CHECKPOINT, (TSTR("find next checkpt block: search: block %d oid %d seq %d eccr %d" TENDSTR), ++ i, tags.objectId, tags.sequenceNumber, tags.eccResult)); ++ ++ if (tags.sequenceNumber == YAFFS_SEQUENCE_CHECKPOINT_DATA) { ++ /* Right kind of block */ ++ dev->checkpointNextBlock = tags.objectId; ++ dev->checkpointCurrentBlock = i; ++ dev->checkpointBlockList[dev->blocksInCheckpoint] = i; ++ dev->blocksInCheckpoint++; ++ T(YAFFS_TRACE_CHECKPOINT, (TSTR("found checkpt block %d"TENDSTR), i)); ++ return; ++ } ++ } ++ ++ T(YAFFS_TRACE_CHECKPOINT, (TSTR("found no more checkpt blocks"TENDSTR))); ++ ++ dev->checkpointNextBlock = -1; ++ dev->checkpointCurrentBlock = -1; ++} ++ ++ ++int yaffs_CheckpointOpen(yaffs_Device *dev, int forWriting) ++{ ++ ++ /* Got the functions we need? */ ++ if (!dev->writeChunkWithTagsToNAND || ++ !dev->readChunkWithTagsFromNAND || ++ !dev->eraseBlockInNAND || ++ !dev->markNANDBlockBad) ++ return 0; ++ ++ if (forWriting && !yaffs_CheckpointSpaceOk(dev)) ++ return 0; ++ ++ if (!dev->checkpointBuffer) ++ dev->checkpointBuffer = YMALLOC_DMA(dev->totalBytesPerChunk); ++ if (!dev->checkpointBuffer) ++ return 0; ++ ++ ++ dev->checkpointPageSequence = 0; ++ ++ dev->checkpointOpenForWrite = forWriting; ++ ++ dev->checkpointByteCount = 0; ++ dev->checkpointSum = 0; ++ dev->checkpointXor = 0; ++ dev->checkpointCurrentBlock = -1; ++ dev->checkpointCurrentChunk = -1; ++ dev->checkpointNextBlock = dev->internalStartBlock; ++ ++ /* Erase all the blocks in the checkpoint area */ ++ if (forWriting) { ++ memset(dev->checkpointBuffer, 0, dev->nDataBytesPerChunk); ++ dev->checkpointByteOffset = 0; ++ return yaffs_CheckpointErase(dev); ++ } else { ++ int i; ++ /* Set to a value that will kick off a read */ ++ dev->checkpointByteOffset = dev->nDataBytesPerChunk; ++ /* A checkpoint block list of 1 checkpoint block per 16 block is (hopefully) ++ * going to be way more than we need */ ++ dev->blocksInCheckpoint = 0; ++ dev->checkpointMaxBlocks = (dev->internalEndBlock - dev->internalStartBlock)/16 + 2; ++ dev->checkpointBlockList = YMALLOC(sizeof(int) * dev->checkpointMaxBlocks); ++ for (i = 0; i < dev->checkpointMaxBlocks; i++) ++ dev->checkpointBlockList[i] = -1; ++ } ++ ++ return 1; ++} ++ ++int yaffs_GetCheckpointSum(yaffs_Device *dev, __u32 *sum) ++{ ++ __u32 compositeSum; ++ compositeSum = (dev->checkpointSum << 8) | (dev->checkpointXor & 0xFF); ++ *sum = compositeSum; ++ return 1; ++} ++ ++static int yaffs_CheckpointFlushBuffer(yaffs_Device *dev) ++{ ++ int chunk; ++ int realignedChunk; ++ ++ yaffs_ExtendedTags tags; ++ ++ if (dev->checkpointCurrentBlock < 0) { ++ yaffs_CheckpointFindNextErasedBlock(dev); ++ dev->checkpointCurrentChunk = 0; ++ } ++ ++ if (dev->checkpointCurrentBlock < 0) ++ return 0; ++ ++ tags.chunkDeleted = 0; ++ tags.objectId = dev->checkpointNextBlock; /* Hint to next place to look */ ++ tags.chunkId = dev->checkpointPageSequence + 1; ++ tags.sequenceNumber = YAFFS_SEQUENCE_CHECKPOINT_DATA; ++ tags.byteCount = dev->nDataBytesPerChunk; ++ if (dev->checkpointCurrentChunk == 0) { ++ /* First chunk we write for the block? Set block state to ++ checkpoint */ ++ yaffs_BlockInfo *bi = yaffs_GetBlockInfo(dev, dev->checkpointCurrentBlock); ++ bi->blockState = YAFFS_BLOCK_STATE_CHECKPOINT; ++ dev->blocksInCheckpoint++; ++ } ++ ++ chunk = dev->checkpointCurrentBlock * dev->nChunksPerBlock + dev->checkpointCurrentChunk; ++ ++ ++ T(YAFFS_TRACE_CHECKPOINT, (TSTR("checkpoint wite buffer nand %d(%d:%d) objid %d chId %d" TENDSTR), ++ chunk, dev->checkpointCurrentBlock, dev->checkpointCurrentChunk, tags.objectId, tags.chunkId)); ++ ++ realignedChunk = chunk - dev->chunkOffset; ++ ++ dev->writeChunkWithTagsToNAND(dev, realignedChunk, ++ dev->checkpointBuffer, &tags); ++ dev->checkpointByteOffset = 0; ++ dev->checkpointPageSequence++; ++ dev->checkpointCurrentChunk++; ++ if (dev->checkpointCurrentChunk >= dev->nChunksPerBlock) { ++ dev->checkpointCurrentChunk = 0; ++ dev->checkpointCurrentBlock = -1; ++ } ++ memset(dev->checkpointBuffer, 0, dev->nDataBytesPerChunk); ++ ++ return 1; ++} ++ ++ ++int yaffs_CheckpointWrite(yaffs_Device *dev, const void *data, int nBytes) ++{ ++ int i = 0; ++ int ok = 1; ++ ++ ++ __u8 * dataBytes = (__u8 *)data; ++ ++ ++ ++ if (!dev->checkpointBuffer) ++ return 0; ++ ++ if (!dev->checkpointOpenForWrite) ++ return -1; ++ ++ while (i < nBytes && ok) { ++ dev->checkpointBuffer[dev->checkpointByteOffset] = *dataBytes; ++ dev->checkpointSum += *dataBytes; ++ dev->checkpointXor ^= *dataBytes; ++ ++ dev->checkpointByteOffset++; ++ i++; ++ dataBytes++; ++ dev->checkpointByteCount++; ++ ++ ++ if (dev->checkpointByteOffset < 0 || ++ dev->checkpointByteOffset >= dev->nDataBytesPerChunk) ++ ok = yaffs_CheckpointFlushBuffer(dev); ++ } ++ ++ return i; ++} ++ ++int yaffs_CheckpointRead(yaffs_Device *dev, void *data, int nBytes) ++{ ++ int i = 0; ++ int ok = 1; ++ yaffs_ExtendedTags tags; ++ ++ ++ int chunk; ++ int realignedChunk; ++ ++ __u8 *dataBytes = (__u8 *)data; ++ ++ if (!dev->checkpointBuffer) ++ return 0; ++ ++ if (dev->checkpointOpenForWrite) ++ return -1; ++ ++ while (i < nBytes && ok) { ++ ++ ++ if (dev->checkpointByteOffset < 0 || ++ dev->checkpointByteOffset >= dev->nDataBytesPerChunk) { ++ ++ if (dev->checkpointCurrentBlock < 0) { ++ yaffs_CheckpointFindNextCheckpointBlock(dev); ++ dev->checkpointCurrentChunk = 0; ++ } ++ ++ if (dev->checkpointCurrentBlock < 0) ++ ok = 0; ++ else { ++ chunk = dev->checkpointCurrentBlock * ++ dev->nChunksPerBlock + ++ dev->checkpointCurrentChunk; ++ ++ realignedChunk = chunk - dev->chunkOffset; ++ ++ /* read in the next chunk */ ++ /* printf("read checkpoint page %d\n",dev->checkpointPage); */ ++ dev->readChunkWithTagsFromNAND(dev, ++ realignedChunk, ++ dev->checkpointBuffer, ++ &tags); ++ ++ if (tags.chunkId != (dev->checkpointPageSequence + 1) || ++ tags.eccResult > YAFFS_ECC_RESULT_FIXED || ++ tags.sequenceNumber != YAFFS_SEQUENCE_CHECKPOINT_DATA) ++ ok = 0; ++ ++ dev->checkpointByteOffset = 0; ++ dev->checkpointPageSequence++; ++ dev->checkpointCurrentChunk++; ++ ++ if (dev->checkpointCurrentChunk >= dev->nChunksPerBlock) ++ dev->checkpointCurrentBlock = -1; ++ } ++ } ++ ++ if (ok) { ++ *dataBytes = dev->checkpointBuffer[dev->checkpointByteOffset]; ++ dev->checkpointSum += *dataBytes; ++ dev->checkpointXor ^= *dataBytes; ++ dev->checkpointByteOffset++; ++ i++; ++ dataBytes++; ++ dev->checkpointByteCount++; ++ } ++ } ++ ++ return i; ++} ++ ++int yaffs_CheckpointClose(yaffs_Device *dev) ++{ ++ ++ if (dev->checkpointOpenForWrite) { ++ if (dev->checkpointByteOffset != 0) ++ yaffs_CheckpointFlushBuffer(dev); ++ } else { ++ int i; ++ for (i = 0; i < dev->blocksInCheckpoint && dev->checkpointBlockList[i] >= 0; i++) { ++ yaffs_BlockInfo *bi = yaffs_GetBlockInfo(dev, dev->checkpointBlockList[i]); ++ if (bi->blockState == YAFFS_BLOCK_STATE_EMPTY) ++ bi->blockState = YAFFS_BLOCK_STATE_CHECKPOINT; ++ else { ++ /* Todo this looks odd... */ ++ } ++ } ++ YFREE(dev->checkpointBlockList); ++ dev->checkpointBlockList = NULL; ++ } ++ ++ dev->nFreeChunks -= dev->blocksInCheckpoint * dev->nChunksPerBlock; ++ dev->nErasedBlocks -= dev->blocksInCheckpoint; ++ ++ ++ T(YAFFS_TRACE_CHECKPOINT, (TSTR("checkpoint byte count %d" TENDSTR), ++ dev->checkpointByteCount)); ++ ++ if (dev->checkpointBuffer) { ++ /* free the buffer */ ++ YFREE(dev->checkpointBuffer); ++ dev->checkpointBuffer = NULL; ++ return 1; ++ } else ++ return 0; ++} ++ ++int yaffs_CheckpointInvalidateStream(yaffs_Device *dev) ++{ ++ /* Erase the first checksum block */ ++ ++ T(YAFFS_TRACE_CHECKPOINT, (TSTR("checkpoint invalidate"TENDSTR))); ++ ++ if (!yaffs_CheckpointSpaceOk(dev)) ++ return 0; ++ ++ return yaffs_CheckpointErase(dev); ++} ++ ++ ++ +diff -Nur linux-2.6.30.orig/fs/yaffs2/yaffs_checkptrw.h linux-2.6.30/fs/yaffs2/yaffs_checkptrw.h +--- linux-2.6.30.orig/fs/yaffs2/yaffs_checkptrw.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/yaffs_checkptrw.h 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,35 @@ ++/* ++ * YAFFS: Yet another Flash File System . A NAND-flash specific file system. ++ * ++ * Copyright (C) 2002-2007 Aleph One Ltd. ++ * for Toby Churchill Ltd and Brightstar Engineering ++ * ++ * Created by Charles Manning <charles@aleph1.co.uk> ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU Lesser General Public License version 2.1 as ++ * published by the Free Software Foundation. ++ * ++ * Note: Only YAFFS headers are LGPL, YAFFS C code is covered by GPL. ++ */ ++ ++#ifndef __YAFFS_CHECKPTRW_H__ ++#define __YAFFS_CHECKPTRW_H__ ++ ++#include "yaffs_guts.h" ++ ++int yaffs_CheckpointOpen(yaffs_Device *dev, int forWriting); ++ ++int yaffs_CheckpointWrite(yaffs_Device *dev, const void *data, int nBytes); ++ ++int yaffs_CheckpointRead(yaffs_Device *dev, void *data, int nBytes); ++ ++int yaffs_GetCheckpointSum(yaffs_Device *dev, __u32 *sum); ++ ++int yaffs_CheckpointClose(yaffs_Device *dev); ++ ++int yaffs_CheckpointInvalidateStream(yaffs_Device *dev); ++ ++ ++#endif ++ +diff -Nur linux-2.6.30.orig/fs/yaffs2/yaffs_ecc.c linux-2.6.30/fs/yaffs2/yaffs_ecc.c +--- linux-2.6.30.orig/fs/yaffs2/yaffs_ecc.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/yaffs_ecc.c 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,326 @@ ++/* ++ * YAFFS: Yet Another Flash File System. A NAND-flash specific file system. ++ * ++ * Copyright (C) 2002-2007 Aleph One Ltd. ++ * for Toby Churchill Ltd and Brightstar Engineering ++ * ++ * Created by Charles Manning <charles@aleph1.co.uk> ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License version 2 as ++ * published by the Free Software Foundation. ++ */ ++ ++/* ++ * This code implements the ECC algorithm used in SmartMedia. ++ * ++ * The ECC comprises 22 bits of parity information and is stuffed into 3 bytes. ++ * The two unused bit are set to 1. ++ * The ECC can correct single bit errors in a 256-byte page of data. Thus, two such ECC ++ * blocks are used on a 512-byte NAND page. ++ * ++ */ ++ ++/* Table generated by gen-ecc.c ++ * Using a table means we do not have to calculate p1..p4 and p1'..p4' ++ * for each byte of data. These are instead provided in a table in bits7..2. ++ * Bit 0 of each entry indicates whether the entry has an odd or even parity, and therefore ++ * this bytes influence on the line parity. ++ */ ++ ++const char *yaffs_ecc_c_version = ++ "$Id: yaffs_ecc.c,v 1.11 2009-03-06 17:20:50 wookey Exp $"; ++ ++#include "yportenv.h" ++ ++#include "yaffs_ecc.h" ++ ++static const unsigned char column_parity_table[] = { ++ 0x00, 0x55, 0x59, 0x0c, 0x65, 0x30, 0x3c, 0x69, ++ 0x69, 0x3c, 0x30, 0x65, 0x0c, 0x59, 0x55, 0x00, ++ 0x95, 0xc0, 0xcc, 0x99, 0xf0, 0xa5, 0xa9, 0xfc, ++ 0xfc, 0xa9, 0xa5, 0xf0, 0x99, 0xcc, 0xc0, 0x95, ++ 0x99, 0xcc, 0xc0, 0x95, 0xfc, 0xa9, 0xa5, 0xf0, ++ 0xf0, 0xa5, 0xa9, 0xfc, 0x95, 0xc0, 0xcc, 0x99, ++ 0x0c, 0x59, 0x55, 0x00, 0x69, 0x3c, 0x30, 0x65, ++ 0x65, 0x30, 0x3c, 0x69, 0x00, 0x55, 0x59, 0x0c, ++ 0xa5, 0xf0, 0xfc, 0xa9, 0xc0, 0x95, 0x99, 0xcc, ++ 0xcc, 0x99, 0x95, 0xc0, 0xa9, 0xfc, 0xf0, 0xa5, ++ 0x30, 0x65, 0x69, 0x3c, 0x55, 0x00, 0x0c, 0x59, ++ 0x59, 0x0c, 0x00, 0x55, 0x3c, 0x69, 0x65, 0x30, ++ 0x3c, 0x69, 0x65, 0x30, 0x59, 0x0c, 0x00, 0x55, ++ 0x55, 0x00, 0x0c, 0x59, 0x30, 0x65, 0x69, 0x3c, ++ 0xa9, 0xfc, 0xf0, 0xa5, 0xcc, 0x99, 0x95, 0xc0, ++ 0xc0, 0x95, 0x99, 0xcc, 0xa5, 0xf0, 0xfc, 0xa9, ++ 0xa9, 0xfc, 0xf0, 0xa5, 0xcc, 0x99, 0x95, 0xc0, ++ 0xc0, 0x95, 0x99, 0xcc, 0xa5, 0xf0, 0xfc, 0xa9, ++ 0x3c, 0x69, 0x65, 0x30, 0x59, 0x0c, 0x00, 0x55, ++ 0x55, 0x00, 0x0c, 0x59, 0x30, 0x65, 0x69, 0x3c, ++ 0x30, 0x65, 0x69, 0x3c, 0x55, 0x00, 0x0c, 0x59, ++ 0x59, 0x0c, 0x00, 0x55, 0x3c, 0x69, 0x65, 0x30, ++ 0xa5, 0xf0, 0xfc, 0xa9, 0xc0, 0x95, 0x99, 0xcc, ++ 0xcc, 0x99, 0x95, 0xc0, 0xa9, 0xfc, 0xf0, 0xa5, ++ 0x0c, 0x59, 0x55, 0x00, 0x69, 0x3c, 0x30, 0x65, ++ 0x65, 0x30, 0x3c, 0x69, 0x00, 0x55, 0x59, 0x0c, ++ 0x99, 0xcc, 0xc0, 0x95, 0xfc, 0xa9, 0xa5, 0xf0, ++ 0xf0, 0xa5, 0xa9, 0xfc, 0x95, 0xc0, 0xcc, 0x99, ++ 0x95, 0xc0, 0xcc, 0x99, 0xf0, 0xa5, 0xa9, 0xfc, ++ 0xfc, 0xa9, 0xa5, 0xf0, 0x99, 0xcc, 0xc0, 0x95, ++ 0x00, 0x55, 0x59, 0x0c, 0x65, 0x30, 0x3c, 0x69, ++ 0x69, 0x3c, 0x30, 0x65, 0x0c, 0x59, 0x55, 0x00, ++}; ++ ++/* Count the bits in an unsigned char or a U32 */ ++ ++static int yaffs_CountBits(unsigned char x) ++{ ++ int r = 0; ++ while (x) { ++ if (x & 1) ++ r++; ++ x >>= 1; ++ } ++ return r; ++} ++ ++static int yaffs_CountBits32(unsigned x) ++{ ++ int r = 0; ++ while (x) { ++ if (x & 1) ++ r++; ++ x >>= 1; ++ } ++ return r; ++} ++ ++/* Calculate the ECC for a 256-byte block of data */ ++void yaffs_ECCCalculate(const unsigned char *data, unsigned char *ecc) ++{ ++ unsigned int i; ++ ++ unsigned char col_parity = 0; ++ unsigned char line_parity = 0; ++ unsigned char line_parity_prime = 0; ++ unsigned char t; ++ unsigned char b; ++ ++ for (i = 0; i < 256; i++) { ++ b = column_parity_table[*data++]; ++ col_parity ^= b; ++ ++ if (b & 0x01) { /* odd number of bits in the byte */ ++ line_parity ^= i; ++ line_parity_prime ^= ~i; ++ } ++ } ++ ++ ecc[2] = (~col_parity) | 0x03; ++ ++ t = 0; ++ if (line_parity & 0x80) ++ t |= 0x80; ++ if (line_parity_prime & 0x80) ++ t |= 0x40; ++ if (line_parity & 0x40) ++ t |= 0x20; ++ if (line_parity_prime & 0x40) ++ t |= 0x10; ++ if (line_parity & 0x20) ++ t |= 0x08; ++ if (line_parity_prime & 0x20) ++ t |= 0x04; ++ if (line_parity & 0x10) ++ t |= 0x02; ++ if (line_parity_prime & 0x10) ++ t |= 0x01; ++ ecc[1] = ~t; ++ ++ t = 0; ++ if (line_parity & 0x08) ++ t |= 0x80; ++ if (line_parity_prime & 0x08) ++ t |= 0x40; ++ if (line_parity & 0x04) ++ t |= 0x20; ++ if (line_parity_prime & 0x04) ++ t |= 0x10; ++ if (line_parity & 0x02) ++ t |= 0x08; ++ if (line_parity_prime & 0x02) ++ t |= 0x04; ++ if (line_parity & 0x01) ++ t |= 0x02; ++ if (line_parity_prime & 0x01) ++ t |= 0x01; ++ ecc[0] = ~t; ++ ++#ifdef CONFIG_YAFFS_ECC_WRONG_ORDER ++ /* Swap the bytes into the wrong order */ ++ t = ecc[0]; ++ ecc[0] = ecc[1]; ++ ecc[1] = t; ++#endif ++} ++ ++ ++/* Correct the ECC on a 256 byte block of data */ ++ ++int yaffs_ECCCorrect(unsigned char *data, unsigned char *read_ecc, ++ const unsigned char *test_ecc) ++{ ++ unsigned char d0, d1, d2; /* deltas */ ++ ++ d0 = read_ecc[0] ^ test_ecc[0]; ++ d1 = read_ecc[1] ^ test_ecc[1]; ++ d2 = read_ecc[2] ^ test_ecc[2]; ++ ++ if ((d0 | d1 | d2) == 0) ++ return 0; /* no error */ ++ ++ if (((d0 ^ (d0 >> 1)) & 0x55) == 0x55 && ++ ((d1 ^ (d1 >> 1)) & 0x55) == 0x55 && ++ ((d2 ^ (d2 >> 1)) & 0x54) == 0x54) { ++ /* Single bit (recoverable) error in data */ ++ ++ unsigned byte; ++ unsigned bit; ++ ++#ifdef CONFIG_YAFFS_ECC_WRONG_ORDER ++ /* swap the bytes to correct for the wrong order */ ++ unsigned char t; ++ ++ t = d0; ++ d0 = d1; ++ d1 = t; ++#endif ++ ++ bit = byte = 0; ++ ++ if (d1 & 0x80) ++ byte |= 0x80; ++ if (d1 & 0x20) ++ byte |= 0x40; ++ if (d1 & 0x08) ++ byte |= 0x20; ++ if (d1 & 0x02) ++ byte |= 0x10; ++ if (d0 & 0x80) ++ byte |= 0x08; ++ if (d0 & 0x20) ++ byte |= 0x04; ++ if (d0 & 0x08) ++ byte |= 0x02; ++ if (d0 & 0x02) ++ byte |= 0x01; ++ ++ if (d2 & 0x80) ++ bit |= 0x04; ++ if (d2 & 0x20) ++ bit |= 0x02; ++ if (d2 & 0x08) ++ bit |= 0x01; ++ ++ data[byte] ^= (1 << bit); ++ ++ return 1; /* Corrected the error */ ++ } ++ ++ if ((yaffs_CountBits(d0) + ++ yaffs_CountBits(d1) + ++ yaffs_CountBits(d2)) == 1) { ++ /* Reccoverable error in ecc */ ++ ++ read_ecc[0] = test_ecc[0]; ++ read_ecc[1] = test_ecc[1]; ++ read_ecc[2] = test_ecc[2]; ++ ++ return 1; /* Corrected the error */ ++ } ++ ++ /* Unrecoverable error */ ++ ++ return -1; ++ ++} ++ ++ ++/* ++ * ECCxxxOther does ECC calcs on arbitrary n bytes of data ++ */ ++void yaffs_ECCCalculateOther(const unsigned char *data, unsigned nBytes, ++ yaffs_ECCOther *eccOther) ++{ ++ unsigned int i; ++ ++ unsigned char col_parity = 0; ++ unsigned line_parity = 0; ++ unsigned line_parity_prime = 0; ++ unsigned char b; ++ ++ for (i = 0; i < nBytes; i++) { ++ b = column_parity_table[*data++]; ++ col_parity ^= b; ++ ++ if (b & 0x01) { ++ /* odd number of bits in the byte */ ++ line_parity ^= i; ++ line_parity_prime ^= ~i; ++ } ++ ++ } ++ ++ eccOther->colParity = (col_parity >> 2) & 0x3f; ++ eccOther->lineParity = line_parity; ++ eccOther->lineParityPrime = line_parity_prime; ++} ++ ++int yaffs_ECCCorrectOther(unsigned char *data, unsigned nBytes, ++ yaffs_ECCOther *read_ecc, ++ const yaffs_ECCOther *test_ecc) ++{ ++ unsigned char cDelta; /* column parity delta */ ++ unsigned lDelta; /* line parity delta */ ++ unsigned lDeltaPrime; /* line parity delta */ ++ unsigned bit; ++ ++ cDelta = read_ecc->colParity ^ test_ecc->colParity; ++ lDelta = read_ecc->lineParity ^ test_ecc->lineParity; ++ lDeltaPrime = read_ecc->lineParityPrime ^ test_ecc->lineParityPrime; ++ ++ if ((cDelta | lDelta | lDeltaPrime) == 0) ++ return 0; /* no error */ ++ ++ if (lDelta == ~lDeltaPrime && ++ (((cDelta ^ (cDelta >> 1)) & 0x15) == 0x15)) { ++ /* Single bit (recoverable) error in data */ ++ ++ bit = 0; ++ ++ if (cDelta & 0x20) ++ bit |= 0x04; ++ if (cDelta & 0x08) ++ bit |= 0x02; ++ if (cDelta & 0x02) ++ bit |= 0x01; ++ ++ if (lDelta >= nBytes) ++ return -1; ++ ++ data[lDelta] ^= (1 << bit); ++ ++ return 1; /* corrected */ ++ } ++ ++ if ((yaffs_CountBits32(lDelta) + yaffs_CountBits32(lDeltaPrime) + ++ yaffs_CountBits(cDelta)) == 1) { ++ /* Reccoverable error in ecc */ ++ ++ *read_ecc = *test_ecc; ++ return 1; /* corrected */ ++ } ++ ++ /* Unrecoverable error */ ++ ++ return -1; ++} +diff -Nur linux-2.6.30.orig/fs/yaffs2/yaffs_ecc.h linux-2.6.30/fs/yaffs2/yaffs_ecc.h +--- linux-2.6.30.orig/fs/yaffs2/yaffs_ecc.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/yaffs_ecc.h 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,44 @@ ++/* ++ * YAFFS: Yet another Flash File System . A NAND-flash specific file system. ++ * ++ * Copyright (C) 2002-2007 Aleph One Ltd. ++ * for Toby Churchill Ltd and Brightstar Engineering ++ * ++ * Created by Charles Manning <charles@aleph1.co.uk> ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU Lesser General Public License version 2.1 as ++ * published by the Free Software Foundation. ++ * ++ * Note: Only YAFFS headers are LGPL, YAFFS C code is covered by GPL. ++ */ ++ ++/* ++ * This code implements the ECC algorithm used in SmartMedia. ++ * ++ * The ECC comprises 22 bits of parity information and is stuffed into 3 bytes. ++ * The two unused bit are set to 1. ++ * The ECC can correct single bit errors in a 256-byte page of data. Thus, two such ECC ++ * blocks are used on a 512-byte NAND page. ++ * ++ */ ++ ++#ifndef __YAFFS_ECC_H__ ++#define __YAFFS_ECC_H__ ++ ++typedef struct { ++ unsigned char colParity; ++ unsigned lineParity; ++ unsigned lineParityPrime; ++} yaffs_ECCOther; ++ ++void yaffs_ECCCalculate(const unsigned char *data, unsigned char *ecc); ++int yaffs_ECCCorrect(unsigned char *data, unsigned char *read_ecc, ++ const unsigned char *test_ecc); ++ ++void yaffs_ECCCalculateOther(const unsigned char *data, unsigned nBytes, ++ yaffs_ECCOther *ecc); ++int yaffs_ECCCorrectOther(unsigned char *data, unsigned nBytes, ++ yaffs_ECCOther *read_ecc, ++ const yaffs_ECCOther *test_ecc); ++#endif +diff -Nur linux-2.6.30.orig/fs/yaffs2/yaffs_fs.c linux-2.6.30/fs/yaffs2/yaffs_fs.c +--- linux-2.6.30.orig/fs/yaffs2/yaffs_fs.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/yaffs_fs.c 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,2529 @@ ++/* ++ * YAFFS: Yet Another Flash File System. A NAND-flash specific file system. ++ * ++ * Copyright (C) 2002-2009 Aleph One Ltd. ++ * for Toby Churchill Ltd and Brightstar Engineering ++ * ++ * Created by Charles Manning <charles@aleph1.co.uk> ++ * Acknowledgements: ++ * Luc van OostenRyck for numerous patches. ++ * Nick Bane for numerous patches. ++ * Nick Bane for 2.5/2.6 integration. ++ * Andras Toth for mknod rdev issue. ++ * Michael Fischer for finding the problem with inode inconsistency. ++ * Some code bodily lifted from JFFS ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License version 2 as ++ * published by the Free Software Foundation. ++ */ ++ ++/* ++ * ++ * This is the file system front-end to YAFFS that hooks it up to ++ * the VFS. ++ * ++ * Special notes: ++ * >> 2.4: sb->u.generic_sbp points to the yaffs_Device associated with ++ * this superblock ++ * >> 2.6: sb->s_fs_info points to the yaffs_Device associated with this ++ * superblock ++ * >> inode->u.generic_ip points to the associated yaffs_Object. ++ */ ++ ++const char *yaffs_fs_c_version = ++ "$Id: yaffs_fs.c,v 1.79 2009-03-17 01:12:00 wookey Exp $"; ++extern const char *yaffs_guts_c_version; ++ ++#include <linux/version.h> ++#if (LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 19)) ++#include <linux/config.h> ++#endif ++#include <linux/kernel.h> ++#include <linux/module.h> ++#include <linux/slab.h> ++#include <linux/init.h> ++#include <linux/fs.h> ++#include <linux/proc_fs.h> ++#include <linux/smp_lock.h> ++#include <linux/pagemap.h> ++#include <linux/mtd/mtd.h> ++#include <linux/interrupt.h> ++#include <linux/string.h> ++#include <linux/ctype.h> ++ ++#include "asm/div64.h" ++ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 5, 0)) ++ ++#include <linux/statfs.h> /* Added NCB 15-8-2003 */ ++#include <linux/statfs.h> ++#define UnlockPage(p) unlock_page(p) ++#define Page_Uptodate(page) test_bit(PG_uptodate, &(page)->flags) ++ ++/* FIXME: use sb->s_id instead ? */ ++#define yaffs_devname(sb, buf) bdevname(sb->s_bdev, buf) ++ ++#else ++ ++#include <linux/locks.h> ++#define BDEVNAME_SIZE 0 ++#define yaffs_devname(sb, buf) kdevname(sb->s_dev) ++ ++#if (LINUX_VERSION_CODE < KERNEL_VERSION(2, 5, 0)) ++/* added NCB 26/5/2006 for 2.4.25-vrs2-tcl1 kernel */ ++#define __user ++#endif ++ ++#endif ++ ++#if (LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 26)) ++#define YPROC_ROOT (&proc_root) ++#else ++#define YPROC_ROOT NULL ++#endif ++ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 6, 17)) ++#define WRITE_SIZE_STR "writesize" ++#define WRITE_SIZE(mtd) ((mtd)->writesize) ++#else ++#define WRITE_SIZE_STR "oobblock" ++#define WRITE_SIZE(mtd) ((mtd)->oobblock) ++#endif ++ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 6, 27)) ++#define YAFFS_USE_WRITE_BEGIN_END 1 ++#else ++#define YAFFS_USE_WRITE_BEGIN_END 0 ++#endif ++ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 6, 28)) ++static uint32_t YCALCBLOCKS(uint64_t partition_size, uint32_t block_size) ++{ ++ uint64_t result = partition_size; ++ do_div(result, block_size); ++ return (uint32_t)result; ++} ++#else ++#define YCALCBLOCKS(s, b) ((s)/(b)) ++#endif ++ ++#include <linux/uaccess.h> ++ ++#include "yportenv.h" ++#include "yaffs_guts.h" ++ ++#include <linux/mtd/mtd.h> ++#include "yaffs_mtdif.h" ++#include "yaffs_mtdif1.h" ++#include "yaffs_mtdif2.h" ++ ++unsigned int yaffs_traceMask = YAFFS_TRACE_BAD_BLOCKS; ++unsigned int yaffs_wr_attempts = YAFFS_WR_ATTEMPTS; ++unsigned int yaffs_auto_checkpoint = 1; ++ ++/* Module Parameters */ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 5, 0)) ++module_param(yaffs_traceMask, uint, 0644); ++module_param(yaffs_wr_attempts, uint, 0644); ++module_param(yaffs_auto_checkpoint, uint, 0644); ++#else ++MODULE_PARM(yaffs_traceMask, "i"); ++MODULE_PARM(yaffs_wr_attempts, "i"); ++MODULE_PARM(yaffs_auto_checkpoint, "i"); ++#endif ++ ++#if (LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 25)) ++/* use iget and read_inode */ ++#define Y_IGET(sb, inum) iget((sb), (inum)) ++static void yaffs_read_inode(struct inode *inode); ++ ++#else ++/* Call local equivalent */ ++#define YAFFS_USE_OWN_IGET ++#define Y_IGET(sb, inum) yaffs_iget((sb), (inum)) ++ ++static struct inode *yaffs_iget(struct super_block *sb, unsigned long ino); ++#endif ++ ++/*#define T(x) printk x */ ++ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 6, 18)) ++#define yaffs_InodeToObjectLV(iptr) ((iptr)->i_private) ++#else ++#define yaffs_InodeToObjectLV(iptr) ((iptr)->u.generic_ip) ++#endif ++ ++#define yaffs_InodeToObject(iptr) ((yaffs_Object *)(yaffs_InodeToObjectLV(iptr))) ++#define yaffs_DentryToObject(dptr) yaffs_InodeToObject((dptr)->d_inode) ++ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 5, 0)) ++#define yaffs_SuperToDevice(sb) ((yaffs_Device *)sb->s_fs_info) ++#else ++#define yaffs_SuperToDevice(sb) ((yaffs_Device *)sb->u.generic_sbp) ++#endif ++ ++static void yaffs_put_super(struct super_block *sb); ++ ++static ssize_t yaffs_file_write(struct file *f, const char *buf, size_t n, ++ loff_t *pos); ++static ssize_t yaffs_hold_space(struct file *f); ++static void yaffs_release_space(struct file *f); ++ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 6, 17)) ++static int yaffs_file_flush(struct file *file, fl_owner_t id); ++#else ++static int yaffs_file_flush(struct file *file); ++#endif ++ ++static int yaffs_sync_object(struct file *file, struct dentry *dentry, ++ int datasync); ++ ++static int yaffs_readdir(struct file *f, void *dirent, filldir_t filldir); ++ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 5, 0)) ++static int yaffs_create(struct inode *dir, struct dentry *dentry, int mode, ++ struct nameidata *n); ++static struct dentry *yaffs_lookup(struct inode *dir, struct dentry *dentry, ++ struct nameidata *n); ++#else ++static int yaffs_create(struct inode *dir, struct dentry *dentry, int mode); ++static struct dentry *yaffs_lookup(struct inode *dir, struct dentry *dentry); ++#endif ++static int yaffs_link(struct dentry *old_dentry, struct inode *dir, ++ struct dentry *dentry); ++static int yaffs_unlink(struct inode *dir, struct dentry *dentry); ++static int yaffs_symlink(struct inode *dir, struct dentry *dentry, ++ const char *symname); ++static int yaffs_mkdir(struct inode *dir, struct dentry *dentry, int mode); ++ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 5, 0)) ++static int yaffs_mknod(struct inode *dir, struct dentry *dentry, int mode, ++ dev_t dev); ++#else ++static int yaffs_mknod(struct inode *dir, struct dentry *dentry, int mode, ++ int dev); ++#endif ++static int yaffs_rename(struct inode *old_dir, struct dentry *old_dentry, ++ struct inode *new_dir, struct dentry *new_dentry); ++static int yaffs_setattr(struct dentry *dentry, struct iattr *attr); ++ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 6, 17)) ++static int yaffs_sync_fs(struct super_block *sb, int wait); ++static void yaffs_write_super(struct super_block *sb); ++#else ++static int yaffs_sync_fs(struct super_block *sb); ++static int yaffs_write_super(struct super_block *sb); ++#endif ++ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 6, 17)) ++static int yaffs_statfs(struct dentry *dentry, struct kstatfs *buf); ++#elif (LINUX_VERSION_CODE > KERNEL_VERSION(2, 5, 0)) ++static int yaffs_statfs(struct super_block *sb, struct kstatfs *buf); ++#else ++static int yaffs_statfs(struct super_block *sb, struct statfs *buf); ++#endif ++ ++#ifdef YAFFS_HAS_PUT_INODE ++static void yaffs_put_inode(struct inode *inode); ++#endif ++ ++static void yaffs_delete_inode(struct inode *); ++static void yaffs_clear_inode(struct inode *); ++ ++static int yaffs_readpage(struct file *file, struct page *page); ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 5, 0)) ++static int yaffs_writepage(struct page *page, struct writeback_control *wbc); ++#else ++static int yaffs_writepage(struct page *page); ++#endif ++ ++ ++#if (YAFFS_USE_WRITE_BEGIN_END != 0) ++static int yaffs_write_begin(struct file *filp, struct address_space *mapping, ++ loff_t pos, unsigned len, unsigned flags, ++ struct page **pagep, void **fsdata); ++static int yaffs_write_end(struct file *filp, struct address_space *mapping, ++ loff_t pos, unsigned len, unsigned copied, ++ struct page *pg, void *fsdadata); ++#else ++static int yaffs_prepare_write(struct file *f, struct page *pg, ++ unsigned offset, unsigned to); ++static int yaffs_commit_write(struct file *f, struct page *pg, unsigned offset, ++ unsigned to); ++ ++#endif ++ ++static int yaffs_readlink(struct dentry *dentry, char __user *buffer, ++ int buflen); ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 6, 13)) ++static void *yaffs_follow_link(struct dentry *dentry, struct nameidata *nd); ++#else ++static int yaffs_follow_link(struct dentry *dentry, struct nameidata *nd); ++#endif ++ ++static struct address_space_operations yaffs_file_address_operations = { ++ .readpage = yaffs_readpage, ++ .writepage = yaffs_writepage, ++#if (YAFFS_USE_WRITE_BEGIN_END > 0) ++ .write_begin = yaffs_write_begin, ++ .write_end = yaffs_write_end, ++#else ++ .prepare_write = yaffs_prepare_write, ++ .commit_write = yaffs_commit_write, ++#endif ++}; ++ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 6, 22)) ++static const struct file_operations yaffs_file_operations = { ++ .read = do_sync_read, ++ .write = do_sync_write, ++ .aio_read = generic_file_aio_read, ++ .aio_write = generic_file_aio_write, ++ .mmap = generic_file_mmap, ++ .flush = yaffs_file_flush, ++ .fsync = yaffs_sync_object, ++ .splice_read = generic_file_splice_read, ++ .splice_write = generic_file_splice_write, ++ .llseek = generic_file_llseek, ++}; ++ ++#elif (LINUX_VERSION_CODE > KERNEL_VERSION(2, 6, 18)) ++ ++static const struct file_operations yaffs_file_operations = { ++ .read = do_sync_read, ++ .write = do_sync_write, ++ .aio_read = generic_file_aio_read, ++ .aio_write = generic_file_aio_write, ++ .mmap = generic_file_mmap, ++ .flush = yaffs_file_flush, ++ .fsync = yaffs_sync_object, ++ .sendfile = generic_file_sendfile, ++}; ++ ++#else ++ ++static const struct file_operations yaffs_file_operations = { ++ .read = generic_file_read, ++ .write = generic_file_write, ++ .mmap = generic_file_mmap, ++ .flush = yaffs_file_flush, ++ .fsync = yaffs_sync_object, ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 5, 0)) ++ .sendfile = generic_file_sendfile, ++#endif ++}; ++#endif ++ ++static const struct inode_operations yaffs_file_inode_operations = { ++ .setattr = yaffs_setattr, ++}; ++ ++static const struct inode_operations yaffs_symlink_inode_operations = { ++ .readlink = yaffs_readlink, ++ .follow_link = yaffs_follow_link, ++ .setattr = yaffs_setattr, ++}; ++ ++static const struct inode_operations yaffs_dir_inode_operations = { ++ .create = yaffs_create, ++ .lookup = yaffs_lookup, ++ .link = yaffs_link, ++ .unlink = yaffs_unlink, ++ .symlink = yaffs_symlink, ++ .mkdir = yaffs_mkdir, ++ .rmdir = yaffs_unlink, ++ .mknod = yaffs_mknod, ++ .rename = yaffs_rename, ++ .setattr = yaffs_setattr, ++}; ++ ++static const struct file_operations yaffs_dir_operations = { ++ .read = generic_read_dir, ++ .readdir = yaffs_readdir, ++ .fsync = yaffs_sync_object, ++}; ++ ++static const struct super_operations yaffs_super_ops = { ++ .statfs = yaffs_statfs, ++ ++#ifndef YAFFS_USE_OWN_IGET ++ .read_inode = yaffs_read_inode, ++#endif ++#ifdef YAFFS_HAS_PUT_INODE ++ .put_inode = yaffs_put_inode, ++#endif ++ .put_super = yaffs_put_super, ++ .delete_inode = yaffs_delete_inode, ++ .clear_inode = yaffs_clear_inode, ++ .sync_fs = yaffs_sync_fs, ++ .write_super = yaffs_write_super, ++}; ++ ++static void yaffs_GrossLock(yaffs_Device *dev) ++{ ++ T(YAFFS_TRACE_OS, ("yaffs locking %p\n", current)); ++ down(&dev->grossLock); ++ T(YAFFS_TRACE_OS, ("yaffs locked %p\n", current)); ++} ++ ++static void yaffs_GrossUnlock(yaffs_Device *dev) ++{ ++ T(YAFFS_TRACE_OS, ("yaffs unlocking %p\n", current)); ++ up(&dev->grossLock); ++} ++ ++static int yaffs_readlink(struct dentry *dentry, char __user *buffer, ++ int buflen) ++{ ++ unsigned char *alias; ++ int ret; ++ ++ yaffs_Device *dev = yaffs_DentryToObject(dentry)->myDev; ++ ++ yaffs_GrossLock(dev); ++ ++ alias = yaffs_GetSymlinkAlias(yaffs_DentryToObject(dentry)); ++ ++ yaffs_GrossUnlock(dev); ++ ++ if (!alias) ++ return -ENOMEM; ++ ++ ret = vfs_readlink(dentry, buffer, buflen, alias); ++ kfree(alias); ++ return ret; ++} ++ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 6, 13)) ++static void *yaffs_follow_link(struct dentry *dentry, struct nameidata *nd) ++#else ++static int yaffs_follow_link(struct dentry *dentry, struct nameidata *nd) ++#endif ++{ ++ unsigned char *alias; ++ int ret; ++ yaffs_Device *dev = yaffs_DentryToObject(dentry)->myDev; ++ ++ yaffs_GrossLock(dev); ++ ++ alias = yaffs_GetSymlinkAlias(yaffs_DentryToObject(dentry)); ++ ++ yaffs_GrossUnlock(dev); ++ ++ if (!alias) { ++ ret = -ENOMEM; ++ goto out; ++ } ++ ++ ret = vfs_follow_link(nd, alias); ++ kfree(alias); ++out: ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 6, 13)) ++ return ERR_PTR(ret); ++#else ++ return ret; ++#endif ++} ++ ++struct inode *yaffs_get_inode(struct super_block *sb, int mode, int dev, ++ yaffs_Object *obj); ++ ++/* ++ * Lookup is used to find objects in the fs ++ */ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 5, 0)) ++ ++static struct dentry *yaffs_lookup(struct inode *dir, struct dentry *dentry, ++ struct nameidata *n) ++#else ++static struct dentry *yaffs_lookup(struct inode *dir, struct dentry *dentry) ++#endif ++{ ++ yaffs_Object *obj; ++ struct inode *inode = NULL; /* NCB 2.5/2.6 needs NULL here */ ++ ++ yaffs_Device *dev = yaffs_InodeToObject(dir)->myDev; ++ ++ yaffs_GrossLock(dev); ++ ++ T(YAFFS_TRACE_OS, ++ ("yaffs_lookup for %d:%s\n", ++ yaffs_InodeToObject(dir)->objectId, dentry->d_name.name)); ++ ++ obj = yaffs_FindObjectByName(yaffs_InodeToObject(dir), ++ dentry->d_name.name); ++ ++ obj = yaffs_GetEquivalentObject(obj); /* in case it was a hardlink */ ++ ++ /* Can't hold gross lock when calling yaffs_get_inode() */ ++ yaffs_GrossUnlock(dev); ++ ++ if (obj) { ++ T(YAFFS_TRACE_OS, ++ ("yaffs_lookup found %d\n", obj->objectId)); ++ ++ inode = yaffs_get_inode(dir->i_sb, obj->yst_mode, 0, obj); ++ ++ if (inode) { ++ T(YAFFS_TRACE_OS, ++ ("yaffs_loookup dentry \n")); ++/* #if 0 asserted by NCB for 2.5/6 compatability - falls through to ++ * d_add even if NULL inode */ ++#if 0 ++ /*dget(dentry); // try to solve directory bug */ ++ d_add(dentry, inode); ++ ++ /* return dentry; */ ++ return NULL; ++#endif ++ } ++ ++ } else { ++ T(YAFFS_TRACE_OS, ("yaffs_lookup not found\n")); ++ ++ } ++ ++/* added NCB for 2.5/6 compatability - forces add even if inode is ++ * NULL which creates dentry hash */ ++ d_add(dentry, inode); ++ ++ return NULL; ++} ++ ++ ++#ifdef YAFFS_HAS_PUT_INODE ++ ++/* For now put inode is just for debugging ++ * Put inode is called when the inode **structure** is put. ++ */ ++static void yaffs_put_inode(struct inode *inode) ++{ ++ T(YAFFS_TRACE_OS, ++ ("yaffs_put_inode: ino %d, count %d\n", (int)inode->i_ino, ++ atomic_read(&inode->i_count))); ++ ++} ++#endif ++ ++/* clear is called to tell the fs to release any per-inode data it holds */ ++static void yaffs_clear_inode(struct inode *inode) ++{ ++ yaffs_Object *obj; ++ yaffs_Device *dev; ++ ++ obj = yaffs_InodeToObject(inode); ++ ++ T(YAFFS_TRACE_OS, ++ ("yaffs_clear_inode: ino %d, count %d %s\n", (int)inode->i_ino, ++ atomic_read(&inode->i_count), ++ obj ? "object exists" : "null object")); ++ ++ if (obj) { ++ dev = obj->myDev; ++ yaffs_GrossLock(dev); ++ ++ /* Clear the association between the inode and ++ * the yaffs_Object. ++ */ ++ obj->myInode = NULL; ++ yaffs_InodeToObjectLV(inode) = NULL; ++ ++ /* If the object freeing was deferred, then the real ++ * free happens now. ++ * This should fix the inode inconsistency problem. ++ */ ++ ++ yaffs_HandleDeferedFree(obj); ++ ++ yaffs_GrossUnlock(dev); ++ } ++ ++} ++ ++/* delete is called when the link count is zero and the inode ++ * is put (ie. nobody wants to know about it anymore, time to ++ * delete the file). ++ * NB Must call clear_inode() ++ */ ++static void yaffs_delete_inode(struct inode *inode) ++{ ++ yaffs_Object *obj = yaffs_InodeToObject(inode); ++ yaffs_Device *dev; ++ ++ T(YAFFS_TRACE_OS, ++ ("yaffs_delete_inode: ino %d, count %d %s\n", (int)inode->i_ino, ++ atomic_read(&inode->i_count), ++ obj ? "object exists" : "null object")); ++ ++ if (obj) { ++ dev = obj->myDev; ++ yaffs_GrossLock(dev); ++ yaffs_DeleteObject(obj); ++ yaffs_GrossUnlock(dev); ++ } ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 6, 13)) ++ truncate_inode_pages(&inode->i_data, 0); ++#endif ++ clear_inode(inode); ++} ++ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 6, 17)) ++static int yaffs_file_flush(struct file *file, fl_owner_t id) ++#else ++static int yaffs_file_flush(struct file *file) ++#endif ++{ ++ yaffs_Object *obj = yaffs_DentryToObject(file->f_dentry); ++ ++ yaffs_Device *dev = obj->myDev; ++ ++ T(YAFFS_TRACE_OS, ++ ("yaffs_file_flush object %d (%s)\n", obj->objectId, ++ obj->dirty ? "dirty" : "clean")); ++ ++ yaffs_GrossLock(dev); ++ ++ yaffs_FlushFile(obj, 1); ++ ++ yaffs_GrossUnlock(dev); ++ ++ return 0; ++} ++ ++static int yaffs_readpage_nolock(struct file *f, struct page *pg) ++{ ++ /* Lifted from jffs2 */ ++ ++ yaffs_Object *obj; ++ unsigned char *pg_buf; ++ int ret; ++ ++ yaffs_Device *dev; ++ ++ T(YAFFS_TRACE_OS, ("yaffs_readpage at %08x, size %08x\n", ++ (unsigned)(pg->index << PAGE_CACHE_SHIFT), ++ (unsigned)PAGE_CACHE_SIZE)); ++ ++ obj = yaffs_DentryToObject(f->f_dentry); ++ ++ dev = obj->myDev; ++ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 5, 0)) ++ BUG_ON(!PageLocked(pg)); ++#else ++ if (!PageLocked(pg)) ++ PAGE_BUG(pg); ++#endif ++ ++ pg_buf = kmap(pg); ++ /* FIXME: Can kmap fail? */ ++ ++ yaffs_GrossLock(dev); ++ ++ ret = yaffs_ReadDataFromFile(obj, pg_buf, ++ pg->index << PAGE_CACHE_SHIFT, ++ PAGE_CACHE_SIZE); ++ ++ yaffs_GrossUnlock(dev); ++ ++ if (ret >= 0) ++ ret = 0; ++ ++ if (ret) { ++ ClearPageUptodate(pg); ++ SetPageError(pg); ++ } else { ++ SetPageUptodate(pg); ++ ClearPageError(pg); ++ } ++ ++ flush_dcache_page(pg); ++ kunmap(pg); ++ ++ T(YAFFS_TRACE_OS, ("yaffs_readpage done\n")); ++ return ret; ++} ++ ++static int yaffs_readpage_unlock(struct file *f, struct page *pg) ++{ ++ int ret = yaffs_readpage_nolock(f, pg); ++ UnlockPage(pg); ++ return ret; ++} ++ ++static int yaffs_readpage(struct file *f, struct page *pg) ++{ ++ return yaffs_readpage_unlock(f, pg); ++} ++ ++/* writepage inspired by/stolen from smbfs */ ++ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 5, 0)) ++static int yaffs_writepage(struct page *page, struct writeback_control *wbc) ++#else ++static int yaffs_writepage(struct page *page) ++#endif ++{ ++ struct address_space *mapping = page->mapping; ++ loff_t offset = (loff_t) page->index << PAGE_CACHE_SHIFT; ++ struct inode *inode; ++ unsigned long end_index; ++ char *buffer; ++ yaffs_Object *obj; ++ int nWritten = 0; ++ unsigned nBytes; ++ ++ if (!mapping) ++ BUG(); ++ inode = mapping->host; ++ if (!inode) ++ BUG(); ++ ++ if (offset > inode->i_size) { ++ T(YAFFS_TRACE_OS, ++ ("yaffs_writepage at %08x, inode size = %08x!!!\n", ++ (unsigned)(page->index << PAGE_CACHE_SHIFT), ++ (unsigned)inode->i_size)); ++ T(YAFFS_TRACE_OS, ++ (" -> don't care!!\n")); ++ unlock_page(page); ++ return 0; ++ } ++ ++ end_index = inode->i_size >> PAGE_CACHE_SHIFT; ++ ++ /* easy case */ ++ if (page->index < end_index) ++ nBytes = PAGE_CACHE_SIZE; ++ else ++ nBytes = inode->i_size & (PAGE_CACHE_SIZE - 1); ++ ++ get_page(page); ++ ++ buffer = kmap(page); ++ ++ obj = yaffs_InodeToObject(inode); ++ yaffs_GrossLock(obj->myDev); ++ ++ T(YAFFS_TRACE_OS, ++ ("yaffs_writepage at %08x, size %08x\n", ++ (unsigned)(page->index << PAGE_CACHE_SHIFT), nBytes)); ++ T(YAFFS_TRACE_OS, ++ ("writepag0: obj = %05x, ino = %05x\n", ++ (int)obj->variant.fileVariant.fileSize, (int)inode->i_size)); ++ ++ nWritten = yaffs_WriteDataToFile(obj, buffer, ++ page->index << PAGE_CACHE_SHIFT, nBytes, 0); ++ ++ T(YAFFS_TRACE_OS, ++ ("writepag1: obj = %05x, ino = %05x\n", ++ (int)obj->variant.fileVariant.fileSize, (int)inode->i_size)); ++ ++ yaffs_GrossUnlock(obj->myDev); ++ ++ kunmap(page); ++ SetPageUptodate(page); ++ UnlockPage(page); ++ put_page(page); ++ ++ return (nWritten == nBytes) ? 0 : -ENOSPC; ++} ++ ++ ++#if (YAFFS_USE_WRITE_BEGIN_END > 0) ++static int yaffs_write_begin(struct file *filp, struct address_space *mapping, ++ loff_t pos, unsigned len, unsigned flags, ++ struct page **pagep, void **fsdata) ++{ ++ struct page *pg = NULL; ++ pgoff_t index = pos >> PAGE_CACHE_SHIFT; ++ uint32_t offset = pos & (PAGE_CACHE_SIZE - 1); ++ uint32_t to = offset + len; ++ ++ int ret = 0; ++ int space_held = 0; ++ ++ T(YAFFS_TRACE_OS, ("start yaffs_write_begin\n")); ++ /* Get a page */ ++#if LINUX_VERSION_CODE > KERNEL_VERSION(2, 6, 28) ++ pg = grab_cache_page_write_begin(mapping, index, flags); ++#else ++ pg = __grab_cache_page(mapping, index); ++#endif ++ ++ *pagep = pg; ++ if (!pg) { ++ ret = -ENOMEM; ++ goto out; ++ } ++ /* Get fs space */ ++ space_held = yaffs_hold_space(filp); ++ ++ if (!space_held) { ++ ret = -ENOSPC; ++ goto out; ++ } ++ ++ /* Update page if required */ ++ ++ if (!Page_Uptodate(pg) && (offset || to < PAGE_CACHE_SIZE)) ++ ret = yaffs_readpage_nolock(filp, pg); ++ ++ if (ret) ++ goto out; ++ ++ /* Happy path return */ ++ T(YAFFS_TRACE_OS, ("end yaffs_write_begin - ok\n")); ++ ++ return 0; ++ ++out: ++ T(YAFFS_TRACE_OS, ("end yaffs_write_begin fail returning %d\n", ret)); ++ if (space_held) ++ yaffs_release_space(filp); ++ if (pg) { ++ unlock_page(pg); ++ page_cache_release(pg); ++ } ++ return ret; ++} ++ ++#else ++ ++static int yaffs_prepare_write(struct file *f, struct page *pg, ++ unsigned offset, unsigned to) ++{ ++ T(YAFFS_TRACE_OS, ("yaffs_prepair_write\n")); ++ ++ if (!Page_Uptodate(pg) && (offset || to < PAGE_CACHE_SIZE)) ++ return yaffs_readpage_nolock(f, pg); ++ return 0; ++} ++#endif ++ ++#if (YAFFS_USE_WRITE_BEGIN_END > 0) ++static int yaffs_write_end(struct file *filp, struct address_space *mapping, ++ loff_t pos, unsigned len, unsigned copied, ++ struct page *pg, void *fsdadata) ++{ ++ int ret = 0; ++ void *addr, *kva; ++ uint32_t offset_into_page = pos & (PAGE_CACHE_SIZE - 1); ++ ++ kva = kmap(pg); ++ addr = kva + offset_into_page; ++ ++ T(YAFFS_TRACE_OS, ++ ("yaffs_write_end addr %x pos %x nBytes %d\n", ++ (unsigned) addr, ++ (int)pos, copied)); ++ ++ ret = yaffs_file_write(filp, addr, copied, &pos); ++ ++ if (ret != copied) { ++ T(YAFFS_TRACE_OS, ++ ("yaffs_write_end not same size ret %d copied %d\n", ++ ret, copied)); ++ SetPageError(pg); ++ ClearPageUptodate(pg); ++ } else { ++ SetPageUptodate(pg); ++ } ++ ++ kunmap(pg); ++ ++ yaffs_release_space(filp); ++ unlock_page(pg); ++ page_cache_release(pg); ++ return ret; ++} ++#else ++ ++static int yaffs_commit_write(struct file *f, struct page *pg, unsigned offset, ++ unsigned to) ++{ ++ void *addr, *kva; ++ ++ loff_t pos = (((loff_t) pg->index) << PAGE_CACHE_SHIFT) + offset; ++ int nBytes = to - offset; ++ int nWritten; ++ ++ unsigned spos = pos; ++ unsigned saddr; ++ ++ kva = kmap(pg); ++ addr = kva + offset; ++ ++ saddr = (unsigned) addr; ++ ++ T(YAFFS_TRACE_OS, ++ ("yaffs_commit_write addr %x pos %x nBytes %d\n", ++ saddr, spos, nBytes)); ++ ++ nWritten = yaffs_file_write(f, addr, nBytes, &pos); ++ ++ if (nWritten != nBytes) { ++ T(YAFFS_TRACE_OS, ++ ("yaffs_commit_write not same size nWritten %d nBytes %d\n", ++ nWritten, nBytes)); ++ SetPageError(pg); ++ ClearPageUptodate(pg); ++ } else { ++ SetPageUptodate(pg); ++ } ++ ++ kunmap(pg); ++ ++ T(YAFFS_TRACE_OS, ++ ("yaffs_commit_write returning %d\n", ++ nWritten == nBytes ? 0 : nWritten)); ++ ++ return nWritten == nBytes ? 0 : nWritten; ++} ++#endif ++ ++ ++static void yaffs_FillInodeFromObject(struct inode *inode, yaffs_Object *obj) ++{ ++ if (inode && obj) { ++ ++ ++ /* Check mode against the variant type and attempt to repair if broken. */ ++ __u32 mode = obj->yst_mode; ++ switch (obj->variantType) { ++ case YAFFS_OBJECT_TYPE_FILE: ++ if (!S_ISREG(mode)) { ++ obj->yst_mode &= ~S_IFMT; ++ obj->yst_mode |= S_IFREG; ++ } ++ ++ break; ++ case YAFFS_OBJECT_TYPE_SYMLINK: ++ if (!S_ISLNK(mode)) { ++ obj->yst_mode &= ~S_IFMT; ++ obj->yst_mode |= S_IFLNK; ++ } ++ ++ break; ++ case YAFFS_OBJECT_TYPE_DIRECTORY: ++ if (!S_ISDIR(mode)) { ++ obj->yst_mode &= ~S_IFMT; ++ obj->yst_mode |= S_IFDIR; ++ } ++ ++ break; ++ case YAFFS_OBJECT_TYPE_UNKNOWN: ++ case YAFFS_OBJECT_TYPE_HARDLINK: ++ case YAFFS_OBJECT_TYPE_SPECIAL: ++ default: ++ /* TODO? */ ++ break; ++ } ++ ++ inode->i_flags |= S_NOATIME; ++ ++ inode->i_ino = obj->objectId; ++ inode->i_mode = obj->yst_mode; ++ inode->i_uid = obj->yst_uid; ++ inode->i_gid = obj->yst_gid; ++#if (LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 19)) ++ inode->i_blksize = inode->i_sb->s_blocksize; ++#endif ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 5, 0)) ++ ++ inode->i_rdev = old_decode_dev(obj->yst_rdev); ++ inode->i_atime.tv_sec = (time_t) (obj->yst_atime); ++ inode->i_atime.tv_nsec = 0; ++ inode->i_mtime.tv_sec = (time_t) obj->yst_mtime; ++ inode->i_mtime.tv_nsec = 0; ++ inode->i_ctime.tv_sec = (time_t) obj->yst_ctime; ++ inode->i_ctime.tv_nsec = 0; ++#else ++ inode->i_rdev = obj->yst_rdev; ++ inode->i_atime = obj->yst_atime; ++ inode->i_mtime = obj->yst_mtime; ++ inode->i_ctime = obj->yst_ctime; ++#endif ++ inode->i_size = yaffs_GetObjectFileLength(obj); ++ inode->i_blocks = (inode->i_size + 511) >> 9; ++ ++ inode->i_nlink = yaffs_GetObjectLinkCount(obj); ++ ++ T(YAFFS_TRACE_OS, ++ ("yaffs_FillInode mode %x uid %d gid %d size %d count %d\n", ++ inode->i_mode, inode->i_uid, inode->i_gid, ++ (int)inode->i_size, atomic_read(&inode->i_count))); ++ ++ switch (obj->yst_mode & S_IFMT) { ++ default: /* fifo, device or socket */ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 5, 0)) ++ init_special_inode(inode, obj->yst_mode, ++ old_decode_dev(obj->yst_rdev)); ++#else ++ init_special_inode(inode, obj->yst_mode, ++ (dev_t) (obj->yst_rdev)); ++#endif ++ break; ++ case S_IFREG: /* file */ ++ inode->i_op = &yaffs_file_inode_operations; ++ inode->i_fop = &yaffs_file_operations; ++ inode->i_mapping->a_ops = ++ &yaffs_file_address_operations; ++ break; ++ case S_IFDIR: /* directory */ ++ inode->i_op = &yaffs_dir_inode_operations; ++ inode->i_fop = &yaffs_dir_operations; ++ break; ++ case S_IFLNK: /* symlink */ ++ inode->i_op = &yaffs_symlink_inode_operations; ++ break; ++ } ++ ++ yaffs_InodeToObjectLV(inode) = obj; ++ ++ obj->myInode = inode; ++ ++ } else { ++ T(YAFFS_TRACE_OS, ++ ("yaffs_FileInode invalid parameters\n")); ++ } ++ ++} ++ ++struct inode *yaffs_get_inode(struct super_block *sb, int mode, int dev, ++ yaffs_Object *obj) ++{ ++ struct inode *inode; ++ ++ if (!sb) { ++ T(YAFFS_TRACE_OS, ++ ("yaffs_get_inode for NULL super_block!!\n")); ++ return NULL; ++ ++ } ++ ++ if (!obj) { ++ T(YAFFS_TRACE_OS, ++ ("yaffs_get_inode for NULL object!!\n")); ++ return NULL; ++ ++ } ++ ++ T(YAFFS_TRACE_OS, ++ ("yaffs_get_inode for object %d\n", obj->objectId)); ++ ++ inode = Y_IGET(sb, obj->objectId); ++ if (IS_ERR(inode)) ++ return NULL; ++ ++ /* NB Side effect: iget calls back to yaffs_read_inode(). */ ++ /* iget also increments the inode's i_count */ ++ /* NB You can't be holding grossLock or deadlock will happen! */ ++ ++ return inode; ++} ++ ++static ssize_t yaffs_file_write(struct file *f, const char *buf, size_t n, ++ loff_t *pos) ++{ ++ yaffs_Object *obj; ++ int nWritten, ipos; ++ struct inode *inode; ++ yaffs_Device *dev; ++ ++ obj = yaffs_DentryToObject(f->f_dentry); ++ ++ dev = obj->myDev; ++ ++ yaffs_GrossLock(dev); ++ ++ inode = f->f_dentry->d_inode; ++ ++ if (!S_ISBLK(inode->i_mode) && f->f_flags & O_APPEND) ++ ipos = inode->i_size; ++ else ++ ipos = *pos; ++ ++ if (!obj) ++ T(YAFFS_TRACE_OS, ++ ("yaffs_file_write: hey obj is null!\n")); ++ else ++ T(YAFFS_TRACE_OS, ++ ("yaffs_file_write about to write writing %zu bytes" ++ "to object %d at %d\n", ++ n, obj->objectId, ipos)); ++ ++ nWritten = yaffs_WriteDataToFile(obj, buf, ipos, n, 0); ++ ++ T(YAFFS_TRACE_OS, ++ ("yaffs_file_write writing %zu bytes, %d written at %d\n", ++ n, nWritten, ipos)); ++ ++ if (nWritten > 0) { ++ ipos += nWritten; ++ *pos = ipos; ++ if (ipos > inode->i_size) { ++ inode->i_size = ipos; ++ inode->i_blocks = (ipos + 511) >> 9; ++ ++ T(YAFFS_TRACE_OS, ++ ("yaffs_file_write size updated to %d bytes, " ++ "%d blocks\n", ++ ipos, (int)(inode->i_blocks))); ++ } ++ ++ } ++ yaffs_GrossUnlock(dev); ++ return nWritten == 0 ? -ENOSPC : nWritten; ++} ++ ++/* Space holding and freeing is done to ensure we have space available for write_begin/end */ ++/* For now we just assume few parallel writes and check against a small number. */ ++/* Todo: need to do this with a counter to handle parallel reads better */ ++ ++static ssize_t yaffs_hold_space(struct file *f) ++{ ++ yaffs_Object *obj; ++ yaffs_Device *dev; ++ ++ int nFreeChunks; ++ ++ ++ obj = yaffs_DentryToObject(f->f_dentry); ++ ++ dev = obj->myDev; ++ ++ yaffs_GrossLock(dev); ++ ++ nFreeChunks = yaffs_GetNumberOfFreeChunks(dev); ++ ++ yaffs_GrossUnlock(dev); ++ ++ return (nFreeChunks > 20) ? 1 : 0; ++} ++ ++static void yaffs_release_space(struct file *f) ++{ ++ yaffs_Object *obj; ++ yaffs_Device *dev; ++ ++ ++ obj = yaffs_DentryToObject(f->f_dentry); ++ ++ dev = obj->myDev; ++ ++ yaffs_GrossLock(dev); ++ ++ ++ yaffs_GrossUnlock(dev); ++} ++ ++static int yaffs_readdir(struct file *f, void *dirent, filldir_t filldir) ++{ ++ yaffs_Object *obj; ++ yaffs_Device *dev; ++ struct inode *inode = f->f_dentry->d_inode; ++ unsigned long offset, curoffs; ++ struct ylist_head *i; ++ yaffs_Object *l; ++ ++ char name[YAFFS_MAX_NAME_LENGTH + 1]; ++ ++ obj = yaffs_DentryToObject(f->f_dentry); ++ dev = obj->myDev; ++ ++ yaffs_GrossLock(dev); ++ ++ offset = f->f_pos; ++ ++ T(YAFFS_TRACE_OS, ("yaffs_readdir: starting at %d\n", (int)offset)); ++ ++ if (offset == 0) { ++ T(YAFFS_TRACE_OS, ++ ("yaffs_readdir: entry . ino %d \n", ++ (int)inode->i_ino)); ++ if (filldir(dirent, ".", 1, offset, inode->i_ino, DT_DIR) < 0) ++ goto out; ++ offset++; ++ f->f_pos++; ++ } ++ if (offset == 1) { ++ T(YAFFS_TRACE_OS, ++ ("yaffs_readdir: entry .. ino %d \n", ++ (int)f->f_dentry->d_parent->d_inode->i_ino)); ++ if (filldir(dirent, "..", 2, offset, ++ f->f_dentry->d_parent->d_inode->i_ino, DT_DIR) < 0) ++ goto out; ++ offset++; ++ f->f_pos++; ++ } ++ ++ curoffs = 1; ++ ++ /* If the directory has changed since the open or last call to ++ readdir, rewind to after the 2 canned entries. */ ++ ++ if (f->f_version != inode->i_version) { ++ offset = 2; ++ f->f_pos = offset; ++ f->f_version = inode->i_version; ++ } ++ ++ ylist_for_each(i, &obj->variant.directoryVariant.children) { ++ curoffs++; ++ if (curoffs >= offset) { ++ l = ylist_entry(i, yaffs_Object, siblings); ++ ++ yaffs_GetObjectName(l, name, ++ YAFFS_MAX_NAME_LENGTH + 1); ++ T(YAFFS_TRACE_OS, ++ ("yaffs_readdir: %s inode %d\n", name, ++ yaffs_GetObjectInode(l))); ++ ++ if (filldir(dirent, ++ name, ++ strlen(name), ++ offset, ++ yaffs_GetObjectInode(l), ++ yaffs_GetObjectType(l)) < 0) ++ goto up_and_out; ++ ++ offset++; ++ f->f_pos++; ++ } ++ } ++ ++up_and_out: ++out: ++ yaffs_GrossUnlock(dev); ++ ++ return 0; ++} ++ ++/* ++ * File creation. Allocate an inode, and we're done.. ++ */ ++ ++#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 29) ++#define YCRED(x) x ++#else ++#define YCRED(x) (x->cred) ++#endif ++ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 5, 0)) ++static int yaffs_mknod(struct inode *dir, struct dentry *dentry, int mode, ++ dev_t rdev) ++#else ++static int yaffs_mknod(struct inode *dir, struct dentry *dentry, int mode, ++ int rdev) ++#endif ++{ ++ struct inode *inode; ++ ++ yaffs_Object *obj = NULL; ++ yaffs_Device *dev; ++ ++ yaffs_Object *parent = yaffs_InodeToObject(dir); ++ ++ int error = -ENOSPC; ++ uid_t uid = YCRED(current)->fsuid; ++ gid_t gid = (dir->i_mode & S_ISGID) ? dir->i_gid : YCRED(current)->fsgid; ++ ++ if ((dir->i_mode & S_ISGID) && S_ISDIR(mode)) ++ mode |= S_ISGID; ++ ++ if (parent) { ++ T(YAFFS_TRACE_OS, ++ ("yaffs_mknod: parent object %d type %d\n", ++ parent->objectId, parent->variantType)); ++ } else { ++ T(YAFFS_TRACE_OS, ++ ("yaffs_mknod: could not get parent object\n")); ++ return -EPERM; ++ } ++ ++ T(YAFFS_TRACE_OS, ("yaffs_mknod: making oject for %s, " ++ "mode %x dev %x\n", ++ dentry->d_name.name, mode, rdev)); ++ ++ dev = parent->myDev; ++ ++ yaffs_GrossLock(dev); ++ ++ switch (mode & S_IFMT) { ++ default: ++ /* Special (socket, fifo, device...) */ ++ T(YAFFS_TRACE_OS, ("yaffs_mknod: making special\n")); ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 5, 0)) ++ obj = yaffs_MknodSpecial(parent, dentry->d_name.name, mode, uid, ++ gid, old_encode_dev(rdev)); ++#else ++ obj = yaffs_MknodSpecial(parent, dentry->d_name.name, mode, uid, ++ gid, rdev); ++#endif ++ break; ++ case S_IFREG: /* file */ ++ T(YAFFS_TRACE_OS, ("yaffs_mknod: making file\n")); ++ obj = yaffs_MknodFile(parent, dentry->d_name.name, mode, uid, ++ gid); ++ break; ++ case S_IFDIR: /* directory */ ++ T(YAFFS_TRACE_OS, ++ ("yaffs_mknod: making directory\n")); ++ obj = yaffs_MknodDirectory(parent, dentry->d_name.name, mode, ++ uid, gid); ++ break; ++ case S_IFLNK: /* symlink */ ++ T(YAFFS_TRACE_OS, ("yaffs_mknod: making symlink\n")); ++ obj = NULL; /* Do we ever get here? */ ++ break; ++ } ++ ++ /* Can not call yaffs_get_inode() with gross lock held */ ++ yaffs_GrossUnlock(dev); ++ ++ if (obj) { ++ inode = yaffs_get_inode(dir->i_sb, mode, rdev, obj); ++ d_instantiate(dentry, inode); ++ T(YAFFS_TRACE_OS, ++ ("yaffs_mknod created object %d count = %d\n", ++ obj->objectId, atomic_read(&inode->i_count))); ++ error = 0; ++ } else { ++ T(YAFFS_TRACE_OS, ++ ("yaffs_mknod failed making object\n")); ++ error = -ENOMEM; ++ } ++ ++ return error; ++} ++ ++static int yaffs_mkdir(struct inode *dir, struct dentry *dentry, int mode) ++{ ++ int retVal; ++ T(YAFFS_TRACE_OS, ("yaffs_mkdir\n")); ++ retVal = yaffs_mknod(dir, dentry, mode | S_IFDIR, 0); ++ return retVal; ++} ++ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 5, 0)) ++static int yaffs_create(struct inode *dir, struct dentry *dentry, int mode, ++ struct nameidata *n) ++#else ++static int yaffs_create(struct inode *dir, struct dentry *dentry, int mode) ++#endif ++{ ++ T(YAFFS_TRACE_OS, ("yaffs_create\n")); ++ return yaffs_mknod(dir, dentry, mode | S_IFREG, 0); ++} ++ ++static int yaffs_unlink(struct inode *dir, struct dentry *dentry) ++{ ++ int retVal; ++ ++ yaffs_Device *dev; ++ ++ T(YAFFS_TRACE_OS, ++ ("yaffs_unlink %d:%s\n", (int)(dir->i_ino), ++ dentry->d_name.name)); ++ ++ dev = yaffs_InodeToObject(dir)->myDev; ++ ++ yaffs_GrossLock(dev); ++ ++ retVal = yaffs_Unlink(yaffs_InodeToObject(dir), dentry->d_name.name); ++ ++ if (retVal == YAFFS_OK) { ++ dentry->d_inode->i_nlink--; ++ dir->i_version++; ++ yaffs_GrossUnlock(dev); ++ mark_inode_dirty(dentry->d_inode); ++ return 0; ++ } ++ yaffs_GrossUnlock(dev); ++ return -ENOTEMPTY; ++} ++ ++/* ++ * Create a link... ++ */ ++static int yaffs_link(struct dentry *old_dentry, struct inode *dir, ++ struct dentry *dentry) ++{ ++ struct inode *inode = old_dentry->d_inode; ++ yaffs_Object *obj = NULL; ++ yaffs_Object *link = NULL; ++ yaffs_Device *dev; ++ ++ T(YAFFS_TRACE_OS, ("yaffs_link\n")); ++ ++ obj = yaffs_InodeToObject(inode); ++ dev = obj->myDev; ++ ++ yaffs_GrossLock(dev); ++ ++ if (!S_ISDIR(inode->i_mode)) /* Don't link directories */ ++ link = yaffs_Link(yaffs_InodeToObject(dir), dentry->d_name.name, ++ obj); ++ ++ if (link) { ++ old_dentry->d_inode->i_nlink = yaffs_GetObjectLinkCount(obj); ++ d_instantiate(dentry, old_dentry->d_inode); ++ atomic_inc(&old_dentry->d_inode->i_count); ++ T(YAFFS_TRACE_OS, ++ ("yaffs_link link count %d i_count %d\n", ++ old_dentry->d_inode->i_nlink, ++ atomic_read(&old_dentry->d_inode->i_count))); ++ } ++ ++ yaffs_GrossUnlock(dev); ++ ++ if (link) ++ return 0; ++ ++ return -EPERM; ++} ++ ++static int yaffs_symlink(struct inode *dir, struct dentry *dentry, ++ const char *symname) ++{ ++ yaffs_Object *obj; ++ yaffs_Device *dev; ++ uid_t uid = YCRED(current)->fsuid; ++ gid_t gid = (dir->i_mode & S_ISGID) ? dir->i_gid : YCRED(current)->fsgid; ++ ++ T(YAFFS_TRACE_OS, ("yaffs_symlink\n")); ++ ++ dev = yaffs_InodeToObject(dir)->myDev; ++ yaffs_GrossLock(dev); ++ obj = yaffs_MknodSymLink(yaffs_InodeToObject(dir), dentry->d_name.name, ++ S_IFLNK | S_IRWXUGO, uid, gid, symname); ++ yaffs_GrossUnlock(dev); ++ ++ if (obj) { ++ struct inode *inode; ++ ++ inode = yaffs_get_inode(dir->i_sb, obj->yst_mode, 0, obj); ++ d_instantiate(dentry, inode); ++ T(YAFFS_TRACE_OS, ("symlink created OK\n")); ++ return 0; ++ } else { ++ T(YAFFS_TRACE_OS, ("symlink not created\n")); ++ } ++ ++ return -ENOMEM; ++} ++ ++static int yaffs_sync_object(struct file *file, struct dentry *dentry, ++ int datasync) ++{ ++ ++ yaffs_Object *obj; ++ yaffs_Device *dev; ++ ++ obj = yaffs_DentryToObject(dentry); ++ ++ dev = obj->myDev; ++ ++ T(YAFFS_TRACE_OS, ("yaffs_sync_object\n")); ++ yaffs_GrossLock(dev); ++ yaffs_FlushFile(obj, 1); ++ yaffs_GrossUnlock(dev); ++ return 0; ++} ++ ++/* ++ * The VFS layer already does all the dentry stuff for rename. ++ * ++ * NB: POSIX says you can rename an object over an old object of the same name ++ */ ++static int yaffs_rename(struct inode *old_dir, struct dentry *old_dentry, ++ struct inode *new_dir, struct dentry *new_dentry) ++{ ++ yaffs_Device *dev; ++ int retVal = YAFFS_FAIL; ++ yaffs_Object *target; ++ ++ T(YAFFS_TRACE_OS, ("yaffs_rename\n")); ++ dev = yaffs_InodeToObject(old_dir)->myDev; ++ ++ yaffs_GrossLock(dev); ++ ++ /* Check if the target is an existing directory that is not empty. */ ++ target = yaffs_FindObjectByName(yaffs_InodeToObject(new_dir), ++ new_dentry->d_name.name); ++ ++ ++ ++ if (target && target->variantType == YAFFS_OBJECT_TYPE_DIRECTORY && ++ !ylist_empty(&target->variant.directoryVariant.children)) { ++ ++ T(YAFFS_TRACE_OS, ("target is non-empty dir\n")); ++ ++ retVal = YAFFS_FAIL; ++ } else { ++ /* Now does unlinking internally using shadowing mechanism */ ++ T(YAFFS_TRACE_OS, ("calling yaffs_RenameObject\n")); ++ ++ retVal = yaffs_RenameObject(yaffs_InodeToObject(old_dir), ++ old_dentry->d_name.name, ++ yaffs_InodeToObject(new_dir), ++ new_dentry->d_name.name); ++ } ++ yaffs_GrossUnlock(dev); ++ ++ if (retVal == YAFFS_OK) { ++ if (target) { ++ new_dentry->d_inode->i_nlink--; ++ mark_inode_dirty(new_dentry->d_inode); ++ } ++ ++ return 0; ++ } else { ++ return -ENOTEMPTY; ++ } ++} ++ ++static int yaffs_setattr(struct dentry *dentry, struct iattr *attr) ++{ ++ struct inode *inode = dentry->d_inode; ++ int error; ++ yaffs_Device *dev; ++ ++ T(YAFFS_TRACE_OS, ++ ("yaffs_setattr of object %d\n", ++ yaffs_InodeToObject(inode)->objectId)); ++ ++ error = inode_change_ok(inode, attr); ++ if (error == 0) { ++ dev = yaffs_InodeToObject(inode)->myDev; ++ yaffs_GrossLock(dev); ++ if (yaffs_SetAttributes(yaffs_InodeToObject(inode), attr) == ++ YAFFS_OK) { ++ error = 0; ++ } else { ++ error = -EPERM; ++ } ++ yaffs_GrossUnlock(dev); ++ if (!error) ++ error = inode_setattr(inode, attr); ++ } ++ return error; ++} ++ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 6, 17)) ++static int yaffs_statfs(struct dentry *dentry, struct kstatfs *buf) ++{ ++ yaffs_Device *dev = yaffs_DentryToObject(dentry)->myDev; ++ struct super_block *sb = dentry->d_sb; ++#elif (LINUX_VERSION_CODE > KERNEL_VERSION(2, 5, 0)) ++static int yaffs_statfs(struct super_block *sb, struct kstatfs *buf) ++{ ++ yaffs_Device *dev = yaffs_SuperToDevice(sb); ++#else ++static int yaffs_statfs(struct super_block *sb, struct statfs *buf) ++{ ++ yaffs_Device *dev = yaffs_SuperToDevice(sb); ++#endif ++ ++ T(YAFFS_TRACE_OS, ("yaffs_statfs\n")); ++ ++ yaffs_GrossLock(dev); ++ ++ buf->f_type = YAFFS_MAGIC; ++ buf->f_bsize = sb->s_blocksize; ++ buf->f_namelen = 255; ++ ++ if (dev->nDataBytesPerChunk & (dev->nDataBytesPerChunk - 1)) { ++ /* Do this if chunk size is not a power of 2 */ ++ ++ uint64_t bytesInDev; ++ uint64_t bytesFree; ++ ++ bytesInDev = ((uint64_t)((dev->endBlock - dev->startBlock + 1))) * ++ ((uint64_t)(dev->nChunksPerBlock * dev->nDataBytesPerChunk)); ++ ++ do_div(bytesInDev, sb->s_blocksize); /* bytesInDev becomes the number of blocks */ ++ buf->f_blocks = bytesInDev; ++ ++ bytesFree = ((uint64_t)(yaffs_GetNumberOfFreeChunks(dev))) * ++ ((uint64_t)(dev->nDataBytesPerChunk)); ++ ++ do_div(bytesFree, sb->s_blocksize); ++ ++ buf->f_bfree = bytesFree; ++ ++ } else if (sb->s_blocksize > dev->nDataBytesPerChunk) { ++ ++ buf->f_blocks = ++ (dev->endBlock - dev->startBlock + 1) * ++ dev->nChunksPerBlock / ++ (sb->s_blocksize / dev->nDataBytesPerChunk); ++ buf->f_bfree = ++ yaffs_GetNumberOfFreeChunks(dev) / ++ (sb->s_blocksize / dev->nDataBytesPerChunk); ++ } else { ++ buf->f_blocks = ++ (dev->endBlock - dev->startBlock + 1) * ++ dev->nChunksPerBlock * ++ (dev->nDataBytesPerChunk / sb->s_blocksize); ++ ++ buf->f_bfree = ++ yaffs_GetNumberOfFreeChunks(dev) * ++ (dev->nDataBytesPerChunk / sb->s_blocksize); ++ } ++ ++ buf->f_files = 0; ++ buf->f_ffree = 0; ++ buf->f_bavail = buf->f_bfree; ++ ++ yaffs_GrossUnlock(dev); ++ return 0; ++} ++ ++ ++static int yaffs_do_sync_fs(struct super_block *sb) ++{ ++ ++ yaffs_Device *dev = yaffs_SuperToDevice(sb); ++ T(YAFFS_TRACE_OS, ("yaffs_do_sync_fs\n")); ++ ++ if (sb->s_dirt) { ++ yaffs_GrossLock(dev); ++ ++ if (dev) { ++ yaffs_FlushEntireDeviceCache(dev); ++ yaffs_CheckpointSave(dev); ++ } ++ ++ yaffs_GrossUnlock(dev); ++ ++ sb->s_dirt = 0; ++ } ++ return 0; ++} ++ ++ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 6, 17)) ++static void yaffs_write_super(struct super_block *sb) ++#else ++static int yaffs_write_super(struct super_block *sb) ++#endif ++{ ++ ++ T(YAFFS_TRACE_OS, ("yaffs_write_super\n")); ++ if (yaffs_auto_checkpoint >= 2) ++ yaffs_do_sync_fs(sb); ++#if (LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 18)) ++ return 0; ++#endif ++} ++ ++ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 6, 17)) ++static int yaffs_sync_fs(struct super_block *sb, int wait) ++#else ++static int yaffs_sync_fs(struct super_block *sb) ++#endif ++{ ++ T(YAFFS_TRACE_OS, ("yaffs_sync_fs\n")); ++ ++ if (yaffs_auto_checkpoint >= 1) ++ yaffs_do_sync_fs(sb); ++ ++ return 0; ++} ++ ++#ifdef YAFFS_USE_OWN_IGET ++ ++static struct inode *yaffs_iget(struct super_block *sb, unsigned long ino) ++{ ++ struct inode *inode; ++ yaffs_Object *obj; ++ yaffs_Device *dev = yaffs_SuperToDevice(sb); ++ ++ T(YAFFS_TRACE_OS, ++ ("yaffs_iget for %lu\n", ino)); ++ ++ inode = iget_locked(sb, ino); ++ if (!inode) ++ return ERR_PTR(-ENOMEM); ++ if (!(inode->i_state & I_NEW)) ++ return inode; ++ ++ /* NB This is called as a side effect of other functions, but ++ * we had to release the lock to prevent deadlocks, so ++ * need to lock again. ++ */ ++ ++ yaffs_GrossLock(dev); ++ ++ obj = yaffs_FindObjectByNumber(dev, inode->i_ino); ++ ++ yaffs_FillInodeFromObject(inode, obj); ++ ++ yaffs_GrossUnlock(dev); ++ ++ unlock_new_inode(inode); ++ return inode; ++} ++ ++#else ++ ++static void yaffs_read_inode(struct inode *inode) ++{ ++ /* NB This is called as a side effect of other functions, but ++ * we had to release the lock to prevent deadlocks, so ++ * need to lock again. ++ */ ++ ++ yaffs_Object *obj; ++ yaffs_Device *dev = yaffs_SuperToDevice(inode->i_sb); ++ ++ T(YAFFS_TRACE_OS, ++ ("yaffs_read_inode for %d\n", (int)inode->i_ino)); ++ ++ yaffs_GrossLock(dev); ++ ++ obj = yaffs_FindObjectByNumber(dev, inode->i_ino); ++ ++ yaffs_FillInodeFromObject(inode, obj); ++ ++ yaffs_GrossUnlock(dev); ++} ++ ++#endif ++ ++static YLIST_HEAD(yaffs_dev_list); ++ ++#if 0 /* not used */ ++static int yaffs_remount_fs(struct super_block *sb, int *flags, char *data) ++{ ++ yaffs_Device *dev = yaffs_SuperToDevice(sb); ++ ++ if (*flags & MS_RDONLY) { ++ struct mtd_info *mtd = yaffs_SuperToDevice(sb)->genericDevice; ++ ++ T(YAFFS_TRACE_OS, ++ ("yaffs_remount_fs: %s: RO\n", dev->name)); ++ ++ yaffs_GrossLock(dev); ++ ++ yaffs_FlushEntireDeviceCache(dev); ++ ++ yaffs_CheckpointSave(dev); ++ ++ if (mtd->sync) ++ mtd->sync(mtd); ++ ++ yaffs_GrossUnlock(dev); ++ } else { ++ T(YAFFS_TRACE_OS, ++ ("yaffs_remount_fs: %s: RW\n", dev->name)); ++ } ++ ++ return 0; ++} ++#endif ++ ++static void yaffs_put_super(struct super_block *sb) ++{ ++ yaffs_Device *dev = yaffs_SuperToDevice(sb); ++ ++ T(YAFFS_TRACE_OS, ("yaffs_put_super\n")); ++ ++ yaffs_GrossLock(dev); ++ ++ yaffs_FlushEntireDeviceCache(dev); ++ ++ yaffs_CheckpointSave(dev); ++ ++ if (dev->putSuperFunc) ++ dev->putSuperFunc(sb); ++ ++ yaffs_Deinitialise(dev); ++ ++ yaffs_GrossUnlock(dev); ++ ++ /* we assume this is protected by lock_kernel() in mount/umount */ ++ ylist_del(&dev->devList); ++ ++ if (dev->spareBuffer) { ++ YFREE(dev->spareBuffer); ++ dev->spareBuffer = NULL; ++ } ++ ++ kfree(dev); ++} ++ ++ ++static void yaffs_MTDPutSuper(struct super_block *sb) ++{ ++ struct mtd_info *mtd = yaffs_SuperToDevice(sb)->genericDevice; ++ ++ if (mtd->sync) ++ mtd->sync(mtd); ++ ++ put_mtd_device(mtd); ++} ++ ++ ++static void yaffs_MarkSuperBlockDirty(void *vsb) ++{ ++ struct super_block *sb = (struct super_block *)vsb; ++ ++ T(YAFFS_TRACE_OS, ("yaffs_MarkSuperBlockDirty() sb = %p\n", sb)); ++ if (sb) ++ sb->s_dirt = 1; ++} ++ ++typedef struct { ++ int inband_tags; ++ int skip_checkpoint_read; ++ int skip_checkpoint_write; ++ int no_cache; ++} yaffs_options; ++ ++#define MAX_OPT_LEN 20 ++static int yaffs_parse_options(yaffs_options *options, const char *options_str) ++{ ++ char cur_opt[MAX_OPT_LEN + 1]; ++ int p; ++ int error = 0; ++ ++ /* Parse through the options which is a comma seperated list */ ++ ++ while (options_str && *options_str && !error) { ++ memset(cur_opt, 0, MAX_OPT_LEN + 1); ++ p = 0; ++ ++ while (*options_str && *options_str != ',') { ++ if (p < MAX_OPT_LEN) { ++ cur_opt[p] = *options_str; ++ p++; ++ } ++ options_str++; ++ } ++ ++ if (!strcmp(cur_opt, "inband-tags")) ++ options->inband_tags = 1; ++ else if (!strcmp(cur_opt, "no-cache")) ++ options->no_cache = 1; ++ else if (!strcmp(cur_opt, "no-checkpoint-read")) ++ options->skip_checkpoint_read = 1; ++ else if (!strcmp(cur_opt, "no-checkpoint-write")) ++ options->skip_checkpoint_write = 1; ++ else if (!strcmp(cur_opt, "no-checkpoint")) { ++ options->skip_checkpoint_read = 1; ++ options->skip_checkpoint_write = 1; ++ } else { ++ printk(KERN_INFO "yaffs: Bad mount option \"%s\"\n", ++ cur_opt); ++ error = 1; ++ } ++ } ++ ++ return error; ++} ++ ++static struct super_block *yaffs_internal_read_super(int yaffsVersion, ++ struct super_block *sb, ++ void *data, int silent) ++{ ++ int nBlocks; ++ struct inode *inode = NULL; ++ struct dentry *root; ++ yaffs_Device *dev = 0; ++ char devname_buf[BDEVNAME_SIZE + 1]; ++ struct mtd_info *mtd; ++ int err; ++ char *data_str = (char *)data; ++ ++ yaffs_options options; ++ ++ sb->s_magic = YAFFS_MAGIC; ++ sb->s_op = &yaffs_super_ops; ++ sb->s_flags |= MS_NOATIME; ++ ++ if (!sb) ++ printk(KERN_INFO "yaffs: sb is NULL\n"); ++ else if (!sb->s_dev) ++ printk(KERN_INFO "yaffs: sb->s_dev is NULL\n"); ++ else if (!yaffs_devname(sb, devname_buf)) ++ printk(KERN_INFO "yaffs: devname is NULL\n"); ++ else ++ printk(KERN_INFO "yaffs: dev is %d name is \"%s\"\n", ++ sb->s_dev, ++ yaffs_devname(sb, devname_buf)); ++ ++ if (!data_str) ++ data_str = ""; ++ ++ printk(KERN_INFO "yaffs: passed flags \"%s\"\n", data_str); ++ ++ memset(&options, 0, sizeof(options)); ++ ++ if (yaffs_parse_options(&options, data_str)) { ++ /* Option parsing failed */ ++ return NULL; ++ } ++ ++ ++ sb->s_blocksize = PAGE_CACHE_SIZE; ++ sb->s_blocksize_bits = PAGE_CACHE_SHIFT; ++ T(YAFFS_TRACE_OS, ("yaffs_read_super: Using yaffs%d\n", yaffsVersion)); ++ T(YAFFS_TRACE_OS, ++ ("yaffs_read_super: block size %d\n", (int)(sb->s_blocksize))); ++ ++#ifdef CONFIG_YAFFS_DISABLE_WRITE_VERIFY ++ T(YAFFS_TRACE_OS, ++ ("yaffs: Write verification disabled. All guarantees " ++ "null and void\n")); ++#endif ++ ++ T(YAFFS_TRACE_ALWAYS, ("yaffs: Attempting MTD mount on %u.%u, " ++ "\"%s\"\n", ++ MAJOR(sb->s_dev), MINOR(sb->s_dev), ++ yaffs_devname(sb, devname_buf))); ++ ++ /* Check it's an mtd device..... */ ++ if (MAJOR(sb->s_dev) != MTD_BLOCK_MAJOR) ++ return NULL; /* This isn't an mtd device */ ++ ++ /* Get the device */ ++ mtd = get_mtd_device(NULL, MINOR(sb->s_dev)); ++ if (!mtd) { ++ T(YAFFS_TRACE_ALWAYS, ++ ("yaffs: MTD device #%u doesn't appear to exist\n", ++ MINOR(sb->s_dev))); ++ return NULL; ++ } ++ /* Check it's NAND */ ++ if (mtd->type != MTD_NANDFLASH) { ++ T(YAFFS_TRACE_ALWAYS, ++ ("yaffs: MTD device is not NAND it's type %d\n", mtd->type)); ++ return NULL; ++ } ++ ++ T(YAFFS_TRACE_OS, (" erase %p\n", mtd->erase)); ++ T(YAFFS_TRACE_OS, (" read %p\n", mtd->read)); ++ T(YAFFS_TRACE_OS, (" write %p\n", mtd->write)); ++ T(YAFFS_TRACE_OS, (" readoob %p\n", mtd->read_oob)); ++ T(YAFFS_TRACE_OS, (" writeoob %p\n", mtd->write_oob)); ++ T(YAFFS_TRACE_OS, (" block_isbad %p\n", mtd->block_isbad)); ++ T(YAFFS_TRACE_OS, (" block_markbad %p\n", mtd->block_markbad)); ++ T(YAFFS_TRACE_OS, (" %s %d\n", WRITE_SIZE_STR, WRITE_SIZE(mtd))); ++ T(YAFFS_TRACE_OS, (" oobsize %d\n", mtd->oobsize)); ++ T(YAFFS_TRACE_OS, (" erasesize %d\n", mtd->erasesize)); ++#if LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 29) ++ T(YAFFS_TRACE_OS, (" size %u\n", mtd->size)); ++#else ++ T(YAFFS_TRACE_OS, (" size %lld\n", mtd->size)); ++#endif ++ ++#ifdef CONFIG_YAFFS_AUTO_YAFFS2 ++ ++ if (yaffsVersion == 1 && WRITE_SIZE(mtd) >= 2048) { ++ T(YAFFS_TRACE_ALWAYS, ("yaffs: auto selecting yaffs2\n")); ++ yaffsVersion = 2; ++ } ++ ++ /* Added NCB 26/5/2006 for completeness */ ++ if (yaffsVersion == 2 && !options.inband_tags && WRITE_SIZE(mtd) == 512) { ++ T(YAFFS_TRACE_ALWAYS, ("yaffs: auto selecting yaffs1\n")); ++ yaffsVersion = 1; ++ } ++ ++#endif ++ ++ if (yaffsVersion == 2) { ++ /* Check for version 2 style functions */ ++ if (!mtd->erase || ++ !mtd->block_isbad || ++ !mtd->block_markbad || ++ !mtd->read || ++ !mtd->write || ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 6, 17)) ++ !mtd->read_oob || !mtd->write_oob) { ++#else ++ !mtd->write_ecc || ++ !mtd->read_ecc || !mtd->read_oob || !mtd->write_oob) { ++#endif ++ T(YAFFS_TRACE_ALWAYS, ++ ("yaffs: MTD device does not support required " ++ "functions\n"));; ++ return NULL; ++ } ++ ++ if ((WRITE_SIZE(mtd) < YAFFS_MIN_YAFFS2_CHUNK_SIZE || ++ mtd->oobsize < YAFFS_MIN_YAFFS2_SPARE_SIZE) && ++ !options.inband_tags) { ++ T(YAFFS_TRACE_ALWAYS, ++ ("yaffs: MTD device does not have the " ++ "right page sizes\n")); ++ return NULL; ++ } ++ } else { ++ /* Check for V1 style functions */ ++ if (!mtd->erase || ++ !mtd->read || ++ !mtd->write || ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 6, 17)) ++ !mtd->read_oob || !mtd->write_oob) { ++#else ++ !mtd->write_ecc || ++ !mtd->read_ecc || !mtd->read_oob || !mtd->write_oob) { ++#endif ++ T(YAFFS_TRACE_ALWAYS, ++ ("yaffs: MTD device does not support required " ++ "functions\n"));; ++ return NULL; ++ } ++ ++ if (WRITE_SIZE(mtd) < YAFFS_BYTES_PER_CHUNK || ++ mtd->oobsize != YAFFS_BYTES_PER_SPARE) { ++ T(YAFFS_TRACE_ALWAYS, ++ ("yaffs: MTD device does not support have the " ++ "right page sizes\n")); ++ return NULL; ++ } ++ } ++ ++ /* OK, so if we got here, we have an MTD that's NAND and looks ++ * like it has the right capabilities ++ * Set the yaffs_Device up for mtd ++ */ ++ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 5, 0)) ++ sb->s_fs_info = dev = kmalloc(sizeof(yaffs_Device), GFP_KERNEL); ++#else ++ sb->u.generic_sbp = dev = kmalloc(sizeof(yaffs_Device), GFP_KERNEL); ++#endif ++ if (!dev) { ++ /* Deep shit could not allocate device structure */ ++ T(YAFFS_TRACE_ALWAYS, ++ ("yaffs_read_super: Failed trying to allocate " ++ "yaffs_Device. \n")); ++ return NULL; ++ } ++ ++ memset(dev, 0, sizeof(yaffs_Device)); ++ dev->genericDevice = mtd; ++ dev->name = mtd->name; ++ ++ /* Set up the memory size parameters.... */ ++ ++ nBlocks = YCALCBLOCKS(mtd->size, (YAFFS_CHUNKS_PER_BLOCK * YAFFS_BYTES_PER_CHUNK)); ++ ++ dev->startBlock = 0; ++ dev->endBlock = nBlocks - 1; ++ dev->nChunksPerBlock = YAFFS_CHUNKS_PER_BLOCK; ++ dev->totalBytesPerChunk = YAFFS_BYTES_PER_CHUNK; ++ dev->nReservedBlocks = 5; ++ dev->nShortOpCaches = (options.no_cache) ? 0 : 10; ++ dev->inbandTags = options.inband_tags; ++ ++ /* ... and the functions. */ ++ if (yaffsVersion == 2) { ++ dev->writeChunkWithTagsToNAND = ++ nandmtd2_WriteChunkWithTagsToNAND; ++ dev->readChunkWithTagsFromNAND = ++ nandmtd2_ReadChunkWithTagsFromNAND; ++ dev->markNANDBlockBad = nandmtd2_MarkNANDBlockBad; ++ dev->queryNANDBlock = nandmtd2_QueryNANDBlock; ++ dev->spareBuffer = YMALLOC(mtd->oobsize); ++ dev->isYaffs2 = 1; ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 6, 17)) ++ dev->totalBytesPerChunk = mtd->writesize; ++ dev->nChunksPerBlock = mtd->erasesize / mtd->writesize; ++#else ++ dev->totalBytesPerChunk = mtd->oobblock; ++ dev->nChunksPerBlock = mtd->erasesize / mtd->oobblock; ++#endif ++ nBlocks = YCALCBLOCKS(mtd->size, mtd->erasesize); ++ ++ dev->startBlock = 0; ++ dev->endBlock = nBlocks - 1; ++ } else { ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 6, 17)) ++ /* use the MTD interface in yaffs_mtdif1.c */ ++ dev->writeChunkWithTagsToNAND = ++ nandmtd1_WriteChunkWithTagsToNAND; ++ dev->readChunkWithTagsFromNAND = ++ nandmtd1_ReadChunkWithTagsFromNAND; ++ dev->markNANDBlockBad = nandmtd1_MarkNANDBlockBad; ++ dev->queryNANDBlock = nandmtd1_QueryNANDBlock; ++#else ++ dev->writeChunkToNAND = nandmtd_WriteChunkToNAND; ++ dev->readChunkFromNAND = nandmtd_ReadChunkFromNAND; ++#endif ++ dev->isYaffs2 = 0; ++ } ++ /* ... and common functions */ ++ dev->eraseBlockInNAND = nandmtd_EraseBlockInNAND; ++ dev->initialiseNAND = nandmtd_InitialiseNAND; ++ ++ dev->putSuperFunc = yaffs_MTDPutSuper; ++ ++ dev->superBlock = (void *)sb; ++ dev->markSuperBlockDirty = yaffs_MarkSuperBlockDirty; ++ ++ ++#ifndef CONFIG_YAFFS_DOES_ECC ++ dev->useNANDECC = 1; ++#endif ++ ++#ifdef CONFIG_YAFFS_DISABLE_WIDE_TNODES ++ dev->wideTnodesDisabled = 1; ++#endif ++ ++ dev->skipCheckpointRead = options.skip_checkpoint_read; ++ dev->skipCheckpointWrite = options.skip_checkpoint_write; ++ ++ /* we assume this is protected by lock_kernel() in mount/umount */ ++ ylist_add_tail(&dev->devList, &yaffs_dev_list); ++ ++ init_MUTEX(&dev->grossLock); ++ ++ yaffs_GrossLock(dev); ++ ++ err = yaffs_GutsInitialise(dev); ++ ++ T(YAFFS_TRACE_OS, ++ ("yaffs_read_super: guts initialised %s\n", ++ (err == YAFFS_OK) ? "OK" : "FAILED")); ++ ++ /* Release lock before yaffs_get_inode() */ ++ yaffs_GrossUnlock(dev); ++ ++ /* Create root inode */ ++ if (err == YAFFS_OK) ++ inode = yaffs_get_inode(sb, S_IFDIR | 0755, 0, ++ yaffs_Root(dev)); ++ ++ if (!inode) ++ return NULL; ++ ++ inode->i_op = &yaffs_dir_inode_operations; ++ inode->i_fop = &yaffs_dir_operations; ++ ++ T(YAFFS_TRACE_OS, ("yaffs_read_super: got root inode\n")); ++ ++ root = d_alloc_root(inode); ++ ++ T(YAFFS_TRACE_OS, ("yaffs_read_super: d_alloc_root done\n")); ++ ++ if (!root) { ++ iput(inode); ++ return NULL; ++ } ++ sb->s_root = root; ++ sb->s_dirt = !dev->isCheckpointed; ++ T(YAFFS_TRACE_ALWAYS, ++ ("yaffs_read_super: isCheckpointed %d\n", dev->isCheckpointed)); ++ ++ T(YAFFS_TRACE_OS, ("yaffs_read_super: done\n")); ++ return sb; ++} ++ ++ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 5, 0)) ++static int yaffs_internal_read_super_mtd(struct super_block *sb, void *data, ++ int silent) ++{ ++ return yaffs_internal_read_super(1, sb, data, silent) ? 0 : -EINVAL; ++} ++ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 6, 17)) ++static int yaffs_read_super(struct file_system_type *fs, ++ int flags, const char *dev_name, ++ void *data, struct vfsmount *mnt) ++{ ++ ++ return get_sb_bdev(fs, flags, dev_name, data, ++ yaffs_internal_read_super_mtd, mnt); ++} ++#else ++static struct super_block *yaffs_read_super(struct file_system_type *fs, ++ int flags, const char *dev_name, ++ void *data) ++{ ++ ++ return get_sb_bdev(fs, flags, dev_name, data, ++ yaffs_internal_read_super_mtd); ++} ++#endif ++ ++static struct file_system_type yaffs_fs_type = { ++ .owner = THIS_MODULE, ++ .name = "yaffs", ++ .get_sb = yaffs_read_super, ++ .kill_sb = kill_block_super, ++ .fs_flags = FS_REQUIRES_DEV, ++}; ++#else ++static struct super_block *yaffs_read_super(struct super_block *sb, void *data, ++ int silent) ++{ ++ return yaffs_internal_read_super(1, sb, data, silent); ++} ++ ++static DECLARE_FSTYPE(yaffs_fs_type, "yaffs", yaffs_read_super, ++ FS_REQUIRES_DEV); ++#endif ++ ++ ++#ifdef CONFIG_YAFFS_YAFFS2 ++ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 5, 0)) ++static int yaffs2_internal_read_super_mtd(struct super_block *sb, void *data, ++ int silent) ++{ ++ return yaffs_internal_read_super(2, sb, data, silent) ? 0 : -EINVAL; ++} ++ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 6, 17)) ++static int yaffs2_read_super(struct file_system_type *fs, ++ int flags, const char *dev_name, void *data, ++ struct vfsmount *mnt) ++{ ++ return get_sb_bdev(fs, flags, dev_name, data, ++ yaffs2_internal_read_super_mtd, mnt); ++} ++#else ++static struct super_block *yaffs2_read_super(struct file_system_type *fs, ++ int flags, const char *dev_name, ++ void *data) ++{ ++ ++ return get_sb_bdev(fs, flags, dev_name, data, ++ yaffs2_internal_read_super_mtd); ++} ++#endif ++ ++static struct file_system_type yaffs2_fs_type = { ++ .owner = THIS_MODULE, ++ .name = "yaffs2", ++ .get_sb = yaffs2_read_super, ++ .kill_sb = kill_block_super, ++ .fs_flags = FS_REQUIRES_DEV, ++}; ++#else ++static struct super_block *yaffs2_read_super(struct super_block *sb, ++ void *data, int silent) ++{ ++ return yaffs_internal_read_super(2, sb, data, silent); ++} ++ ++static DECLARE_FSTYPE(yaffs2_fs_type, "yaffs2", yaffs2_read_super, ++ FS_REQUIRES_DEV); ++#endif ++ ++#endif /* CONFIG_YAFFS_YAFFS2 */ ++ ++static struct proc_dir_entry *my_proc_entry; ++ ++static char *yaffs_dump_dev(char *buf, yaffs_Device * dev) ++{ ++ buf += sprintf(buf, "startBlock......... %d\n", dev->startBlock); ++ buf += sprintf(buf, "endBlock........... %d\n", dev->endBlock); ++ buf += sprintf(buf, "totalBytesPerChunk. %d\n", dev->totalBytesPerChunk); ++ buf += sprintf(buf, "nDataBytesPerChunk. %d\n", dev->nDataBytesPerChunk); ++ buf += sprintf(buf, "chunkGroupBits..... %d\n", dev->chunkGroupBits); ++ buf += sprintf(buf, "chunkGroupSize..... %d\n", dev->chunkGroupSize); ++ buf += sprintf(buf, "nErasedBlocks...... %d\n", dev->nErasedBlocks); ++ buf += sprintf(buf, "nReservedBlocks.... %d\n", dev->nReservedBlocks); ++ buf += sprintf(buf, "blocksInCheckpoint. %d\n", dev->blocksInCheckpoint); ++ buf += sprintf(buf, "nTnodesCreated..... %d\n", dev->nTnodesCreated); ++ buf += sprintf(buf, "nFreeTnodes........ %d\n", dev->nFreeTnodes); ++ buf += sprintf(buf, "nObjectsCreated.... %d\n", dev->nObjectsCreated); ++ buf += sprintf(buf, "nFreeObjects....... %d\n", dev->nFreeObjects); ++ buf += sprintf(buf, "nFreeChunks........ %d\n", dev->nFreeChunks); ++ buf += sprintf(buf, "nPageWrites........ %d\n", dev->nPageWrites); ++ buf += sprintf(buf, "nPageReads......... %d\n", dev->nPageReads); ++ buf += sprintf(buf, "nBlockErasures..... %d\n", dev->nBlockErasures); ++ buf += sprintf(buf, "nGCCopies.......... %d\n", dev->nGCCopies); ++ buf += sprintf(buf, "garbageCollections. %d\n", dev->garbageCollections); ++ buf += sprintf(buf, "passiveGCs......... %d\n", ++ dev->passiveGarbageCollections); ++ buf += sprintf(buf, "nRetriedWrites..... %d\n", dev->nRetriedWrites); ++ buf += sprintf(buf, "nShortOpCaches..... %d\n", dev->nShortOpCaches); ++ buf += sprintf(buf, "nRetireBlocks...... %d\n", dev->nRetiredBlocks); ++ buf += sprintf(buf, "eccFixed........... %d\n", dev->eccFixed); ++ buf += sprintf(buf, "eccUnfixed......... %d\n", dev->eccUnfixed); ++ buf += sprintf(buf, "tagsEccFixed....... %d\n", dev->tagsEccFixed); ++ buf += sprintf(buf, "tagsEccUnfixed..... %d\n", dev->tagsEccUnfixed); ++ buf += sprintf(buf, "cacheHits.......... %d\n", dev->cacheHits); ++ buf += sprintf(buf, "nDeletedFiles...... %d\n", dev->nDeletedFiles); ++ buf += sprintf(buf, "nUnlinkedFiles..... %d\n", dev->nUnlinkedFiles); ++ buf += ++ sprintf(buf, "nBackgroudDeletions %d\n", dev->nBackgroundDeletions); ++ buf += sprintf(buf, "useNANDECC......... %d\n", dev->useNANDECC); ++ buf += sprintf(buf, "isYaffs2........... %d\n", dev->isYaffs2); ++ buf += sprintf(buf, "inbandTags......... %d\n", dev->inbandTags); ++ ++ return buf; ++} ++ ++static int yaffs_proc_read(char *page, ++ char **start, ++ off_t offset, int count, int *eof, void *data) ++{ ++ struct ylist_head *item; ++ char *buf = page; ++ int step = offset; ++ int n = 0; ++ ++ /* Get proc_file_read() to step 'offset' by one on each sucessive call. ++ * We use 'offset' (*ppos) to indicate where we are in devList. ++ * This also assumes the user has posted a read buffer large ++ * enough to hold the complete output; but that's life in /proc. ++ */ ++ ++ *(int *)start = 1; ++ ++ /* Print header first */ ++ if (step == 0) { ++ buf += sprintf(buf, "YAFFS built:" __DATE__ " " __TIME__ ++ "\n%s\n%s\n", yaffs_fs_c_version, ++ yaffs_guts_c_version); ++ } ++ ++ /* hold lock_kernel while traversing yaffs_dev_list */ ++ lock_kernel(); ++ ++ /* Locate and print the Nth entry. Order N-squared but N is small. */ ++ ylist_for_each(item, &yaffs_dev_list) { ++ yaffs_Device *dev = ylist_entry(item, yaffs_Device, devList); ++ if (n < step) { ++ n++; ++ continue; ++ } ++ buf += sprintf(buf, "\nDevice %d \"%s\"\n", n, dev->name); ++ buf = yaffs_dump_dev(buf, dev); ++ break; ++ } ++ unlock_kernel(); ++ ++ return buf - page < count ? buf - page : count; ++} ++ ++/** ++ * Set the verbosity of the warnings and error messages. ++ * ++ * Note that the names can only be a..z or _ with the current code. ++ */ ++ ++static struct { ++ char *mask_name; ++ unsigned mask_bitfield; ++} mask_flags[] = { ++ {"allocate", YAFFS_TRACE_ALLOCATE}, ++ {"always", YAFFS_TRACE_ALWAYS}, ++ {"bad_blocks", YAFFS_TRACE_BAD_BLOCKS}, ++ {"buffers", YAFFS_TRACE_BUFFERS}, ++ {"bug", YAFFS_TRACE_BUG}, ++ {"checkpt", YAFFS_TRACE_CHECKPOINT}, ++ {"deletion", YAFFS_TRACE_DELETION}, ++ {"erase", YAFFS_TRACE_ERASE}, ++ {"error", YAFFS_TRACE_ERROR}, ++ {"gc_detail", YAFFS_TRACE_GC_DETAIL}, ++ {"gc", YAFFS_TRACE_GC}, ++ {"mtd", YAFFS_TRACE_MTD}, ++ {"nandaccess", YAFFS_TRACE_NANDACCESS}, ++ {"os", YAFFS_TRACE_OS}, ++ {"scan_debug", YAFFS_TRACE_SCAN_DEBUG}, ++ {"scan", YAFFS_TRACE_SCAN}, ++ {"tracing", YAFFS_TRACE_TRACING}, ++ ++ {"verify", YAFFS_TRACE_VERIFY}, ++ {"verify_nand", YAFFS_TRACE_VERIFY_NAND}, ++ {"verify_full", YAFFS_TRACE_VERIFY_FULL}, ++ {"verify_all", YAFFS_TRACE_VERIFY_ALL}, ++ ++ {"write", YAFFS_TRACE_WRITE}, ++ {"all", 0xffffffff}, ++ {"none", 0}, ++ {NULL, 0}, ++}; ++ ++#define MAX_MASK_NAME_LENGTH 40 ++static int yaffs_proc_write(struct file *file, const char *buf, ++ unsigned long count, void *data) ++{ ++ unsigned rg = 0, mask_bitfield; ++ char *end; ++ char *mask_name; ++ const char *x; ++ char substring[MAX_MASK_NAME_LENGTH + 1]; ++ int i; ++ int done = 0; ++ int add, len = 0; ++ int pos = 0; ++ ++ rg = yaffs_traceMask; ++ ++ while (!done && (pos < count)) { ++ done = 1; ++ while ((pos < count) && isspace(buf[pos])) ++ pos++; ++ ++ switch (buf[pos]) { ++ case '+': ++ case '-': ++ case '=': ++ add = buf[pos]; ++ pos++; ++ break; ++ ++ default: ++ add = ' '; ++ break; ++ } ++ mask_name = NULL; ++ ++ mask_bitfield = simple_strtoul(buf + pos, &end, 0); ++ ++ if (end > buf + pos) { ++ mask_name = "numeral"; ++ len = end - (buf + pos); ++ pos += len; ++ done = 0; ++ } else { ++ for (x = buf + pos, i = 0; ++ (*x == '_' || (*x >= 'a' && *x <= 'z')) && ++ i < MAX_MASK_NAME_LENGTH; x++, i++, pos++) ++ substring[i] = *x; ++ substring[i] = '\0'; ++ ++ for (i = 0; mask_flags[i].mask_name != NULL; i++) { ++ if (strcmp(substring, mask_flags[i].mask_name) == 0) { ++ mask_name = mask_flags[i].mask_name; ++ mask_bitfield = mask_flags[i].mask_bitfield; ++ done = 0; ++ break; ++ } ++ } ++ } ++ ++ if (mask_name != NULL) { ++ done = 0; ++ switch (add) { ++ case '-': ++ rg &= ~mask_bitfield; ++ break; ++ case '+': ++ rg |= mask_bitfield; ++ break; ++ case '=': ++ rg = mask_bitfield; ++ break; ++ default: ++ rg |= mask_bitfield; ++ break; ++ } ++ } ++ } ++ ++ yaffs_traceMask = rg | YAFFS_TRACE_ALWAYS; ++ ++ printk(KERN_DEBUG "new trace = 0x%08X\n", yaffs_traceMask); ++ ++ if (rg & YAFFS_TRACE_ALWAYS) { ++ for (i = 0; mask_flags[i].mask_name != NULL; i++) { ++ char flag; ++ flag = ((rg & mask_flags[i].mask_bitfield) == mask_flags[i].mask_bitfield) ? '+' : '-'; ++ printk(KERN_DEBUG "%c%s\n", flag, mask_flags[i].mask_name); ++ } ++ } ++ ++ return count; ++} ++ ++/* Stuff to handle installation of file systems */ ++struct file_system_to_install { ++ struct file_system_type *fst; ++ int installed; ++}; ++ ++static struct file_system_to_install fs_to_install[] = { ++ {&yaffs_fs_type, 0}, ++ {&yaffs2_fs_type, 0}, ++ {NULL, 0} ++}; ++ ++static int __init init_yaffs_fs(void) ++{ ++ int error = 0; ++ struct file_system_to_install *fsinst; ++ ++ T(YAFFS_TRACE_ALWAYS, ++ ("yaffs " __DATE__ " " __TIME__ " Installing. \n")); ++ ++ /* Install the proc_fs entry */ ++ my_proc_entry = create_proc_entry("yaffs", ++ S_IRUGO | S_IFREG, ++ YPROC_ROOT); ++ ++ if (my_proc_entry) { ++ my_proc_entry->write_proc = yaffs_proc_write; ++ my_proc_entry->read_proc = yaffs_proc_read; ++ my_proc_entry->data = NULL; ++ } else ++ return -ENOMEM; ++ ++ /* Now add the file system entries */ ++ ++ fsinst = fs_to_install; ++ ++ while (fsinst->fst && !error) { ++ error = register_filesystem(fsinst->fst); ++ if (!error) ++ fsinst->installed = 1; ++ fsinst++; ++ } ++ ++ /* Any errors? uninstall */ ++ if (error) { ++ fsinst = fs_to_install; ++ ++ while (fsinst->fst) { ++ if (fsinst->installed) { ++ unregister_filesystem(fsinst->fst); ++ fsinst->installed = 0; ++ } ++ fsinst++; ++ } ++ } ++ ++ return error; ++} ++ ++static void __exit exit_yaffs_fs(void) ++{ ++ ++ struct file_system_to_install *fsinst; ++ ++ T(YAFFS_TRACE_ALWAYS, ("yaffs " __DATE__ " " __TIME__ ++ " removing. \n")); ++ ++ remove_proc_entry("yaffs", YPROC_ROOT); ++ ++ fsinst = fs_to_install; ++ ++ while (fsinst->fst) { ++ if (fsinst->installed) { ++ unregister_filesystem(fsinst->fst); ++ fsinst->installed = 0; ++ } ++ fsinst++; ++ } ++} ++ ++module_init(init_yaffs_fs) ++module_exit(exit_yaffs_fs) ++ ++MODULE_DESCRIPTION("YAFFS2 - a NAND specific flash file system"); ++MODULE_AUTHOR("Charles Manning, Aleph One Ltd., 2002-2006"); ++MODULE_LICENSE("GPL"); +diff -Nur linux-2.6.30.orig/fs/yaffs2/yaffs_getblockinfo.h linux-2.6.30/fs/yaffs2/yaffs_getblockinfo.h +--- linux-2.6.30.orig/fs/yaffs2/yaffs_getblockinfo.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/yaffs_getblockinfo.h 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,34 @@ ++/* ++ * YAFFS: Yet another Flash File System . A NAND-flash specific file system. ++ * ++ * Copyright (C) 2002-2007 Aleph One Ltd. ++ * for Toby Churchill Ltd and Brightstar Engineering ++ * ++ * Created by Charles Manning <charles@aleph1.co.uk> ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU Lesser General Public License version 2.1 as ++ * published by the Free Software Foundation. ++ * ++ * Note: Only YAFFS headers are LGPL, YAFFS C code is covered by GPL. ++ */ ++ ++#ifndef __YAFFS_GETBLOCKINFO_H__ ++#define __YAFFS_GETBLOCKINFO_H__ ++ ++#include "yaffs_guts.h" ++ ++/* Function to manipulate block info */ ++static Y_INLINE yaffs_BlockInfo *yaffs_GetBlockInfo(yaffs_Device * dev, int blk) ++{ ++ if (blk < dev->internalStartBlock || blk > dev->internalEndBlock) { ++ T(YAFFS_TRACE_ERROR, ++ (TSTR ++ ("**>> yaffs: getBlockInfo block %d is not valid" TENDSTR), ++ blk)); ++ YBUG(); ++ } ++ return &dev->blockInfo[blk - dev->internalStartBlock]; ++} ++ ++#endif +diff -Nur linux-2.6.30.orig/fs/yaffs2/yaffs_guts.c linux-2.6.30/fs/yaffs2/yaffs_guts.c +--- linux-2.6.30.orig/fs/yaffs2/yaffs_guts.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/yaffs_guts.c 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,7552 @@ ++/* ++ * YAFFS: Yet Another Flash File System. A NAND-flash specific file system. ++ * ++ * Copyright (C) 2002-2007 Aleph One Ltd. ++ * for Toby Churchill Ltd and Brightstar Engineering ++ * ++ * Created by Charles Manning <charles@aleph1.co.uk> ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License version 2 as ++ * published by the Free Software Foundation. ++ */ ++ ++const char *yaffs_guts_c_version = ++ "$Id: yaffs_guts.c,v 1.82 2009-03-09 04:24:17 charles Exp $"; ++ ++#include "yportenv.h" ++ ++#include "yaffsinterface.h" ++#include "yaffs_guts.h" ++#include "yaffs_tagsvalidity.h" ++#include "yaffs_getblockinfo.h" ++ ++#include "yaffs_tagscompat.h" ++#ifndef CONFIG_YAFFS_USE_OWN_SORT ++#include "yaffs_qsort.h" ++#endif ++#include "yaffs_nand.h" ++ ++#include "yaffs_checkptrw.h" ++ ++#include "yaffs_nand.h" ++#include "yaffs_packedtags2.h" ++ ++ ++#define YAFFS_PASSIVE_GC_CHUNKS 2 ++ ++#include "yaffs_ecc.h" ++ ++ ++/* Robustification (if it ever comes about...) */ ++static void yaffs_RetireBlock(yaffs_Device *dev, int blockInNAND); ++static void yaffs_HandleWriteChunkError(yaffs_Device *dev, int chunkInNAND, ++ int erasedOk); ++static void yaffs_HandleWriteChunkOk(yaffs_Device *dev, int chunkInNAND, ++ const __u8 *data, ++ const yaffs_ExtendedTags *tags); ++static void yaffs_HandleUpdateChunk(yaffs_Device *dev, int chunkInNAND, ++ const yaffs_ExtendedTags *tags); ++ ++/* Other local prototypes */ ++static int yaffs_UnlinkObject(yaffs_Object *obj); ++static int yaffs_ObjectHasCachedWriteData(yaffs_Object *obj); ++ ++static void yaffs_HardlinkFixup(yaffs_Device *dev, yaffs_Object *hardList); ++ ++static int yaffs_WriteNewChunkWithTagsToNAND(yaffs_Device *dev, ++ const __u8 *buffer, ++ yaffs_ExtendedTags *tags, ++ int useReserve); ++static int yaffs_PutChunkIntoFile(yaffs_Object *in, int chunkInInode, ++ int chunkInNAND, int inScan); ++ ++static yaffs_Object *yaffs_CreateNewObject(yaffs_Device *dev, int number, ++ yaffs_ObjectType type); ++static void yaffs_AddObjectToDirectory(yaffs_Object *directory, ++ yaffs_Object *obj); ++static int yaffs_UpdateObjectHeader(yaffs_Object *in, const YCHAR *name, ++ int force, int isShrink, int shadows); ++static void yaffs_RemoveObjectFromDirectory(yaffs_Object *obj); ++static int yaffs_CheckStructures(void); ++static int yaffs_DeleteWorker(yaffs_Object *in, yaffs_Tnode *tn, __u32 level, ++ int chunkOffset, int *limit); ++static int yaffs_DoGenericObjectDeletion(yaffs_Object *in); ++ ++static yaffs_BlockInfo *yaffs_GetBlockInfo(yaffs_Device *dev, int blockNo); ++ ++ ++static int yaffs_CheckChunkErased(struct yaffs_DeviceStruct *dev, ++ int chunkInNAND); ++ ++static int yaffs_UnlinkWorker(yaffs_Object *obj); ++ ++static int yaffs_TagsMatch(const yaffs_ExtendedTags *tags, int objectId, ++ int chunkInObject); ++ ++static int yaffs_AllocateChunk(yaffs_Device *dev, int useReserve, ++ yaffs_BlockInfo **blockUsedPtr); ++ ++static void yaffs_VerifyFreeChunks(yaffs_Device *dev); ++ ++static void yaffs_CheckObjectDetailsLoaded(yaffs_Object *in); ++ ++static void yaffs_VerifyDirectory(yaffs_Object *directory); ++#ifdef YAFFS_PARANOID ++static int yaffs_CheckFileSanity(yaffs_Object *in); ++#else ++#define yaffs_CheckFileSanity(in) ++#endif ++ ++static void yaffs_InvalidateWholeChunkCache(yaffs_Object *in); ++static void yaffs_InvalidateChunkCache(yaffs_Object *object, int chunkId); ++ ++static void yaffs_InvalidateCheckpoint(yaffs_Device *dev); ++ ++static int yaffs_FindChunkInFile(yaffs_Object *in, int chunkInInode, ++ yaffs_ExtendedTags *tags); ++ ++static __u32 yaffs_GetChunkGroupBase(yaffs_Device *dev, yaffs_Tnode *tn, ++ unsigned pos); ++static yaffs_Tnode *yaffs_FindLevel0Tnode(yaffs_Device *dev, ++ yaffs_FileStructure *fStruct, ++ __u32 chunkId); ++ ++ ++/* Function to calculate chunk and offset */ ++ ++static void yaffs_AddrToChunk(yaffs_Device *dev, loff_t addr, int *chunkOut, ++ __u32 *offsetOut) ++{ ++ int chunk; ++ __u32 offset; ++ ++ chunk = (__u32)(addr >> dev->chunkShift); ++ ++ if (dev->chunkDiv == 1) { ++ /* easy power of 2 case */ ++ offset = (__u32)(addr & dev->chunkMask); ++ } else { ++ /* Non power-of-2 case */ ++ ++ loff_t chunkBase; ++ ++ chunk /= dev->chunkDiv; ++ ++ chunkBase = ((loff_t)chunk) * dev->nDataBytesPerChunk; ++ offset = (__u32)(addr - chunkBase); ++ } ++ ++ *chunkOut = chunk; ++ *offsetOut = offset; ++} ++ ++/* Function to return the number of shifts for a power of 2 greater than or ++ * equal to the given number ++ * Note we don't try to cater for all possible numbers and this does not have to ++ * be hellishly efficient. ++ */ ++ ++static __u32 ShiftsGE(__u32 x) ++{ ++ int extraBits; ++ int nShifts; ++ ++ nShifts = extraBits = 0; ++ ++ while (x > 1) { ++ if (x & 1) ++ extraBits++; ++ x >>= 1; ++ nShifts++; ++ } ++ ++ if (extraBits) ++ nShifts++; ++ ++ return nShifts; ++} ++ ++/* Function to return the number of shifts to get a 1 in bit 0 ++ */ ++ ++static __u32 Shifts(__u32 x) ++{ ++ int nShifts; ++ ++ nShifts = 0; ++ ++ if (!x) ++ return 0; ++ ++ while (!(x&1)) { ++ x >>= 1; ++ nShifts++; ++ } ++ ++ return nShifts; ++} ++ ++ ++ ++/* ++ * Temporary buffer manipulations. ++ */ ++ ++static int yaffs_InitialiseTempBuffers(yaffs_Device *dev) ++{ ++ int i; ++ __u8 *buf = (__u8 *)1; ++ ++ memset(dev->tempBuffer, 0, sizeof(dev->tempBuffer)); ++ ++ for (i = 0; buf && i < YAFFS_N_TEMP_BUFFERS; i++) { ++ dev->tempBuffer[i].line = 0; /* not in use */ ++ dev->tempBuffer[i].buffer = buf = ++ YMALLOC_DMA(dev->totalBytesPerChunk); ++ } ++ ++ return buf ? YAFFS_OK : YAFFS_FAIL; ++} ++ ++__u8 *yaffs_GetTempBuffer(yaffs_Device *dev, int lineNo) ++{ ++ int i, j; ++ ++ dev->tempInUse++; ++ if (dev->tempInUse > dev->maxTemp) ++ dev->maxTemp = dev->tempInUse; ++ ++ for (i = 0; i < YAFFS_N_TEMP_BUFFERS; i++) { ++ if (dev->tempBuffer[i].line == 0) { ++ dev->tempBuffer[i].line = lineNo; ++ if ((i + 1) > dev->maxTemp) { ++ dev->maxTemp = i + 1; ++ for (j = 0; j <= i; j++) ++ dev->tempBuffer[j].maxLine = ++ dev->tempBuffer[j].line; ++ } ++ ++ return dev->tempBuffer[i].buffer; ++ } ++ } ++ ++ T(YAFFS_TRACE_BUFFERS, ++ (TSTR("Out of temp buffers at line %d, other held by lines:"), ++ lineNo)); ++ for (i = 0; i < YAFFS_N_TEMP_BUFFERS; i++) ++ T(YAFFS_TRACE_BUFFERS, (TSTR(" %d "), dev->tempBuffer[i].line)); ++ ++ T(YAFFS_TRACE_BUFFERS, (TSTR(" " TENDSTR))); ++ ++ /* ++ * If we got here then we have to allocate an unmanaged one ++ * This is not good. ++ */ ++ ++ dev->unmanagedTempAllocations++; ++ return YMALLOC(dev->nDataBytesPerChunk); ++ ++} ++ ++void yaffs_ReleaseTempBuffer(yaffs_Device *dev, __u8 *buffer, ++ int lineNo) ++{ ++ int i; ++ ++ dev->tempInUse--; ++ ++ for (i = 0; i < YAFFS_N_TEMP_BUFFERS; i++) { ++ if (dev->tempBuffer[i].buffer == buffer) { ++ dev->tempBuffer[i].line = 0; ++ return; ++ } ++ } ++ ++ if (buffer) { ++ /* assume it is an unmanaged one. */ ++ T(YAFFS_TRACE_BUFFERS, ++ (TSTR("Releasing unmanaged temp buffer in line %d" TENDSTR), ++ lineNo)); ++ YFREE(buffer); ++ dev->unmanagedTempDeallocations++; ++ } ++ ++} ++ ++/* ++ * Determine if we have a managed buffer. ++ */ ++int yaffs_IsManagedTempBuffer(yaffs_Device *dev, const __u8 *buffer) ++{ ++ int i; ++ ++ for (i = 0; i < YAFFS_N_TEMP_BUFFERS; i++) { ++ if (dev->tempBuffer[i].buffer == buffer) ++ return 1; ++ } ++ ++ for (i = 0; i < dev->nShortOpCaches; i++) { ++ if (dev->srCache[i].data == buffer) ++ return 1; ++ } ++ ++ if (buffer == dev->checkpointBuffer) ++ return 1; ++ ++ T(YAFFS_TRACE_ALWAYS, ++ (TSTR("yaffs: unmaged buffer detected.\n" TENDSTR))); ++ return 0; ++} ++ ++ ++ ++/* ++ * Chunk bitmap manipulations ++ */ ++ ++static Y_INLINE __u8 *yaffs_BlockBits(yaffs_Device *dev, int blk) ++{ ++ if (blk < dev->internalStartBlock || blk > dev->internalEndBlock) { ++ T(YAFFS_TRACE_ERROR, ++ (TSTR("**>> yaffs: BlockBits block %d is not valid" TENDSTR), ++ blk)); ++ YBUG(); ++ } ++ return dev->chunkBits + ++ (dev->chunkBitmapStride * (blk - dev->internalStartBlock)); ++} ++ ++static Y_INLINE void yaffs_VerifyChunkBitId(yaffs_Device *dev, int blk, int chunk) ++{ ++ if (blk < dev->internalStartBlock || blk > dev->internalEndBlock || ++ chunk < 0 || chunk >= dev->nChunksPerBlock) { ++ T(YAFFS_TRACE_ERROR, ++ (TSTR("**>> yaffs: Chunk Id (%d:%d) invalid"TENDSTR), ++ blk, chunk)); ++ YBUG(); ++ } ++} ++ ++static Y_INLINE void yaffs_ClearChunkBits(yaffs_Device *dev, int blk) ++{ ++ __u8 *blkBits = yaffs_BlockBits(dev, blk); ++ ++ memset(blkBits, 0, dev->chunkBitmapStride); ++} ++ ++static Y_INLINE void yaffs_ClearChunkBit(yaffs_Device *dev, int blk, int chunk) ++{ ++ __u8 *blkBits = yaffs_BlockBits(dev, blk); ++ ++ yaffs_VerifyChunkBitId(dev, blk, chunk); ++ ++ blkBits[chunk / 8] &= ~(1 << (chunk & 7)); ++} ++ ++static Y_INLINE void yaffs_SetChunkBit(yaffs_Device *dev, int blk, int chunk) ++{ ++ __u8 *blkBits = yaffs_BlockBits(dev, blk); ++ ++ yaffs_VerifyChunkBitId(dev, blk, chunk); ++ ++ blkBits[chunk / 8] |= (1 << (chunk & 7)); ++} ++ ++static Y_INLINE int yaffs_CheckChunkBit(yaffs_Device *dev, int blk, int chunk) ++{ ++ __u8 *blkBits = yaffs_BlockBits(dev, blk); ++ yaffs_VerifyChunkBitId(dev, blk, chunk); ++ ++ return (blkBits[chunk / 8] & (1 << (chunk & 7))) ? 1 : 0; ++} ++ ++static Y_INLINE int yaffs_StillSomeChunkBits(yaffs_Device *dev, int blk) ++{ ++ __u8 *blkBits = yaffs_BlockBits(dev, blk); ++ int i; ++ for (i = 0; i < dev->chunkBitmapStride; i++) { ++ if (*blkBits) ++ return 1; ++ blkBits++; ++ } ++ return 0; ++} ++ ++static int yaffs_CountChunkBits(yaffs_Device *dev, int blk) ++{ ++ __u8 *blkBits = yaffs_BlockBits(dev, blk); ++ int i; ++ int n = 0; ++ for (i = 0; i < dev->chunkBitmapStride; i++) { ++ __u8 x = *blkBits; ++ while (x) { ++ if (x & 1) ++ n++; ++ x >>= 1; ++ } ++ ++ blkBits++; ++ } ++ return n; ++} ++ ++/* ++ * Verification code ++ */ ++ ++static int yaffs_SkipVerification(yaffs_Device *dev) ++{ ++ return !(yaffs_traceMask & (YAFFS_TRACE_VERIFY | YAFFS_TRACE_VERIFY_FULL)); ++} ++ ++static int yaffs_SkipFullVerification(yaffs_Device *dev) ++{ ++ return !(yaffs_traceMask & (YAFFS_TRACE_VERIFY_FULL)); ++} ++ ++static int yaffs_SkipNANDVerification(yaffs_Device *dev) ++{ ++ return !(yaffs_traceMask & (YAFFS_TRACE_VERIFY_NAND)); ++} ++ ++static const char *blockStateName[] = { ++"Unknown", ++"Needs scanning", ++"Scanning", ++"Empty", ++"Allocating", ++"Full", ++"Dirty", ++"Checkpoint", ++"Collecting", ++"Dead" ++}; ++ ++static void yaffs_VerifyBlock(yaffs_Device *dev, yaffs_BlockInfo *bi, int n) ++{ ++ int actuallyUsed; ++ int inUse; ++ ++ if (yaffs_SkipVerification(dev)) ++ return; ++ ++ /* Report illegal runtime states */ ++ if (bi->blockState >= YAFFS_NUMBER_OF_BLOCK_STATES) ++ T(YAFFS_TRACE_VERIFY, (TSTR("Block %d has undefined state %d"TENDSTR), n, bi->blockState)); ++ ++ switch (bi->blockState) { ++ case YAFFS_BLOCK_STATE_UNKNOWN: ++ case YAFFS_BLOCK_STATE_SCANNING: ++ case YAFFS_BLOCK_STATE_NEEDS_SCANNING: ++ T(YAFFS_TRACE_VERIFY, (TSTR("Block %d has bad run-state %s"TENDSTR), ++ n, blockStateName[bi->blockState])); ++ } ++ ++ /* Check pages in use and soft deletions are legal */ ++ ++ actuallyUsed = bi->pagesInUse - bi->softDeletions; ++ ++ if (bi->pagesInUse < 0 || bi->pagesInUse > dev->nChunksPerBlock || ++ bi->softDeletions < 0 || bi->softDeletions > dev->nChunksPerBlock || ++ actuallyUsed < 0 || actuallyUsed > dev->nChunksPerBlock) ++ T(YAFFS_TRACE_VERIFY, (TSTR("Block %d has illegal values pagesInUsed %d softDeletions %d"TENDSTR), ++ n, bi->pagesInUse, bi->softDeletions)); ++ ++ ++ /* Check chunk bitmap legal */ ++ inUse = yaffs_CountChunkBits(dev, n); ++ if (inUse != bi->pagesInUse) ++ T(YAFFS_TRACE_VERIFY, (TSTR("Block %d has inconsistent values pagesInUse %d counted chunk bits %d"TENDSTR), ++ n, bi->pagesInUse, inUse)); ++ ++ /* Check that the sequence number is valid. ++ * Ten million is legal, but is very unlikely ++ */ ++ if (dev->isYaffs2 && ++ (bi->blockState == YAFFS_BLOCK_STATE_ALLOCATING || bi->blockState == YAFFS_BLOCK_STATE_FULL) && ++ (bi->sequenceNumber < YAFFS_LOWEST_SEQUENCE_NUMBER || bi->sequenceNumber > 10000000)) ++ T(YAFFS_TRACE_VERIFY, (TSTR("Block %d has suspect sequence number of %d"TENDSTR), ++ n, bi->sequenceNumber)); ++} ++ ++static void yaffs_VerifyCollectedBlock(yaffs_Device *dev, yaffs_BlockInfo *bi, ++ int n) ++{ ++ yaffs_VerifyBlock(dev, bi, n); ++ ++ /* After collection the block should be in the erased state */ ++ /* This will need to change if we do partial gc */ ++ ++ if (bi->blockState != YAFFS_BLOCK_STATE_COLLECTING && ++ bi->blockState != YAFFS_BLOCK_STATE_EMPTY) { ++ T(YAFFS_TRACE_ERROR, (TSTR("Block %d is in state %d after gc, should be erased"TENDSTR), ++ n, bi->blockState)); ++ } ++} ++ ++static void yaffs_VerifyBlocks(yaffs_Device *dev) ++{ ++ int i; ++ int nBlocksPerState[YAFFS_NUMBER_OF_BLOCK_STATES]; ++ int nIllegalBlockStates = 0; ++ ++ if (yaffs_SkipVerification(dev)) ++ return; ++ ++ memset(nBlocksPerState, 0, sizeof(nBlocksPerState)); ++ ++ for (i = dev->internalStartBlock; i <= dev->internalEndBlock; i++) { ++ yaffs_BlockInfo *bi = yaffs_GetBlockInfo(dev, i); ++ yaffs_VerifyBlock(dev, bi, i); ++ ++ if (bi->blockState < YAFFS_NUMBER_OF_BLOCK_STATES) ++ nBlocksPerState[bi->blockState]++; ++ else ++ nIllegalBlockStates++; ++ } ++ ++ T(YAFFS_TRACE_VERIFY, (TSTR(""TENDSTR))); ++ T(YAFFS_TRACE_VERIFY, (TSTR("Block summary"TENDSTR))); ++ ++ T(YAFFS_TRACE_VERIFY, (TSTR("%d blocks have illegal states"TENDSTR), nIllegalBlockStates)); ++ if (nBlocksPerState[YAFFS_BLOCK_STATE_ALLOCATING] > 1) ++ T(YAFFS_TRACE_VERIFY, (TSTR("Too many allocating blocks"TENDSTR))); ++ ++ for (i = 0; i < YAFFS_NUMBER_OF_BLOCK_STATES; i++) ++ T(YAFFS_TRACE_VERIFY, ++ (TSTR("%s %d blocks"TENDSTR), ++ blockStateName[i], nBlocksPerState[i])); ++ ++ if (dev->blocksInCheckpoint != nBlocksPerState[YAFFS_BLOCK_STATE_CHECKPOINT]) ++ T(YAFFS_TRACE_VERIFY, ++ (TSTR("Checkpoint block count wrong dev %d count %d"TENDSTR), ++ dev->blocksInCheckpoint, nBlocksPerState[YAFFS_BLOCK_STATE_CHECKPOINT])); ++ ++ if (dev->nErasedBlocks != nBlocksPerState[YAFFS_BLOCK_STATE_EMPTY]) ++ T(YAFFS_TRACE_VERIFY, ++ (TSTR("Erased block count wrong dev %d count %d"TENDSTR), ++ dev->nErasedBlocks, nBlocksPerState[YAFFS_BLOCK_STATE_EMPTY])); ++ ++ if (nBlocksPerState[YAFFS_BLOCK_STATE_COLLECTING] > 1) ++ T(YAFFS_TRACE_VERIFY, ++ (TSTR("Too many collecting blocks %d (max is 1)"TENDSTR), ++ nBlocksPerState[YAFFS_BLOCK_STATE_COLLECTING])); ++ ++ T(YAFFS_TRACE_VERIFY, (TSTR(""TENDSTR))); ++ ++} ++ ++/* ++ * Verify the object header. oh must be valid, but obj and tags may be NULL in which ++ * case those tests will not be performed. ++ */ ++static void yaffs_VerifyObjectHeader(yaffs_Object *obj, yaffs_ObjectHeader *oh, yaffs_ExtendedTags *tags, int parentCheck) ++{ ++ if (obj && yaffs_SkipVerification(obj->myDev)) ++ return; ++ ++ if (!(tags && obj && oh)) { ++ T(YAFFS_TRACE_VERIFY, ++ (TSTR("Verifying object header tags %x obj %x oh %x"TENDSTR), ++ (__u32)tags, (__u32)obj, (__u32)oh)); ++ return; ++ } ++ ++ if (oh->type <= YAFFS_OBJECT_TYPE_UNKNOWN || ++ oh->type > YAFFS_OBJECT_TYPE_MAX) ++ T(YAFFS_TRACE_VERIFY, ++ (TSTR("Obj %d header type is illegal value 0x%x"TENDSTR), ++ tags->objectId, oh->type)); ++ ++ if (tags->objectId != obj->objectId) ++ T(YAFFS_TRACE_VERIFY, ++ (TSTR("Obj %d header mismatch objectId %d"TENDSTR), ++ tags->objectId, obj->objectId)); ++ ++ ++ /* ++ * Check that the object's parent ids match if parentCheck requested. ++ * ++ * Tests do not apply to the root object. ++ */ ++ ++ if (parentCheck && tags->objectId > 1 && !obj->parent) ++ T(YAFFS_TRACE_VERIFY, ++ (TSTR("Obj %d header mismatch parentId %d obj->parent is NULL"TENDSTR), ++ tags->objectId, oh->parentObjectId)); ++ ++ if (parentCheck && obj->parent && ++ oh->parentObjectId != obj->parent->objectId && ++ (oh->parentObjectId != YAFFS_OBJECTID_UNLINKED || ++ obj->parent->objectId != YAFFS_OBJECTID_DELETED)) ++ T(YAFFS_TRACE_VERIFY, ++ (TSTR("Obj %d header mismatch parentId %d parentObjectId %d"TENDSTR), ++ tags->objectId, oh->parentObjectId, obj->parent->objectId)); ++ ++ if (tags->objectId > 1 && oh->name[0] == 0) /* Null name */ ++ T(YAFFS_TRACE_VERIFY, ++ (TSTR("Obj %d header name is NULL"TENDSTR), ++ obj->objectId)); ++ ++ if (tags->objectId > 1 && ((__u8)(oh->name[0])) == 0xff) /* Trashed name */ ++ T(YAFFS_TRACE_VERIFY, ++ (TSTR("Obj %d header name is 0xFF"TENDSTR), ++ obj->objectId)); ++} ++ ++ ++ ++static int yaffs_VerifyTnodeWorker(yaffs_Object *obj, yaffs_Tnode *tn, ++ __u32 level, int chunkOffset) ++{ ++ int i; ++ yaffs_Device *dev = obj->myDev; ++ int ok = 1; ++ ++ if (tn) { ++ if (level > 0) { ++ ++ for (i = 0; i < YAFFS_NTNODES_INTERNAL && ok; i++) { ++ if (tn->internal[i]) { ++ ok = yaffs_VerifyTnodeWorker(obj, ++ tn->internal[i], ++ level - 1, ++ (chunkOffset<<YAFFS_TNODES_INTERNAL_BITS) + i); ++ } ++ } ++ } else if (level == 0) { ++ yaffs_ExtendedTags tags; ++ __u32 objectId = obj->objectId; ++ ++ chunkOffset <<= YAFFS_TNODES_LEVEL0_BITS; ++ ++ for (i = 0; i < YAFFS_NTNODES_LEVEL0; i++) { ++ __u32 theChunk = yaffs_GetChunkGroupBase(dev, tn, i); ++ ++ if (theChunk > 0) { ++ /* T(~0,(TSTR("verifying (%d:%d) %d"TENDSTR),tags.objectId,tags.chunkId,theChunk)); */ ++ yaffs_ReadChunkWithTagsFromNAND(dev, theChunk, NULL, &tags); ++ if (tags.objectId != objectId || tags.chunkId != chunkOffset) { ++ T(~0, (TSTR("Object %d chunkId %d NAND mismatch chunk %d tags (%d:%d)"TENDSTR), ++ objectId, chunkOffset, theChunk, ++ tags.objectId, tags.chunkId)); ++ } ++ } ++ chunkOffset++; ++ } ++ } ++ } ++ ++ return ok; ++ ++} ++ ++ ++static void yaffs_VerifyFile(yaffs_Object *obj) ++{ ++ int requiredTallness; ++ int actualTallness; ++ __u32 lastChunk; ++ __u32 x; ++ __u32 i; ++ yaffs_Device *dev; ++ yaffs_ExtendedTags tags; ++ yaffs_Tnode *tn; ++ __u32 objectId; ++ ++ if (!obj) ++ return; ++ ++ if (yaffs_SkipVerification(obj->myDev)) ++ return; ++ ++ dev = obj->myDev; ++ objectId = obj->objectId; ++ ++ /* Check file size is consistent with tnode depth */ ++ lastChunk = obj->variant.fileVariant.fileSize / dev->nDataBytesPerChunk + 1; ++ x = lastChunk >> YAFFS_TNODES_LEVEL0_BITS; ++ requiredTallness = 0; ++ while (x > 0) { ++ x >>= YAFFS_TNODES_INTERNAL_BITS; ++ requiredTallness++; ++ } ++ ++ actualTallness = obj->variant.fileVariant.topLevel; ++ ++ if (requiredTallness > actualTallness) ++ T(YAFFS_TRACE_VERIFY, ++ (TSTR("Obj %d had tnode tallness %d, needs to be %d"TENDSTR), ++ obj->objectId, actualTallness, requiredTallness)); ++ ++ ++ /* Check that the chunks in the tnode tree are all correct. ++ * We do this by scanning through the tnode tree and ++ * checking the tags for every chunk match. ++ */ ++ ++ if (yaffs_SkipNANDVerification(dev)) ++ return; ++ ++ for (i = 1; i <= lastChunk; i++) { ++ tn = yaffs_FindLevel0Tnode(dev, &obj->variant.fileVariant, i); ++ ++ if (tn) { ++ __u32 theChunk = yaffs_GetChunkGroupBase(dev, tn, i); ++ if (theChunk > 0) { ++ /* T(~0,(TSTR("verifying (%d:%d) %d"TENDSTR),objectId,i,theChunk)); */ ++ yaffs_ReadChunkWithTagsFromNAND(dev, theChunk, NULL, &tags); ++ if (tags.objectId != objectId || tags.chunkId != i) { ++ T(~0, (TSTR("Object %d chunkId %d NAND mismatch chunk %d tags (%d:%d)"TENDSTR), ++ objectId, i, theChunk, ++ tags.objectId, tags.chunkId)); ++ } ++ } ++ } ++ } ++} ++ ++ ++static void yaffs_VerifyHardLink(yaffs_Object *obj) ++{ ++ if (obj && yaffs_SkipVerification(obj->myDev)) ++ return; ++ ++ /* Verify sane equivalent object */ ++} ++ ++static void yaffs_VerifySymlink(yaffs_Object *obj) ++{ ++ if (obj && yaffs_SkipVerification(obj->myDev)) ++ return; ++ ++ /* Verify symlink string */ ++} ++ ++static void yaffs_VerifySpecial(yaffs_Object *obj) ++{ ++ if (obj && yaffs_SkipVerification(obj->myDev)) ++ return; ++} ++ ++static void yaffs_VerifyObject(yaffs_Object *obj) ++{ ++ yaffs_Device *dev; ++ ++ __u32 chunkMin; ++ __u32 chunkMax; ++ ++ __u32 chunkIdOk; ++ __u32 chunkInRange; ++ __u32 chunkShouldNotBeDeleted; ++ __u32 chunkValid; ++ ++ if (!obj) ++ return; ++ ++ if (obj->beingCreated) ++ return; ++ ++ dev = obj->myDev; ++ ++ if (yaffs_SkipVerification(dev)) ++ return; ++ ++ /* Check sane object header chunk */ ++ ++ chunkMin = dev->internalStartBlock * dev->nChunksPerBlock; ++ chunkMax = (dev->internalEndBlock+1) * dev->nChunksPerBlock - 1; ++ ++ chunkInRange = (((unsigned)(obj->hdrChunk)) >= chunkMin && ((unsigned)(obj->hdrChunk)) <= chunkMax); ++ chunkIdOk = chunkInRange || obj->hdrChunk == 0; ++ chunkValid = chunkInRange && ++ yaffs_CheckChunkBit(dev, ++ obj->hdrChunk / dev->nChunksPerBlock, ++ obj->hdrChunk % dev->nChunksPerBlock); ++ chunkShouldNotBeDeleted = chunkInRange && !chunkValid; ++ ++ if (!obj->fake && ++ (!chunkIdOk || chunkShouldNotBeDeleted)) { ++ T(YAFFS_TRACE_VERIFY, ++ (TSTR("Obj %d has chunkId %d %s %s"TENDSTR), ++ obj->objectId, obj->hdrChunk, ++ chunkIdOk ? "" : ",out of range", ++ chunkShouldNotBeDeleted ? ",marked as deleted" : "")); ++ } ++ ++ if (chunkValid && !yaffs_SkipNANDVerification(dev)) { ++ yaffs_ExtendedTags tags; ++ yaffs_ObjectHeader *oh; ++ __u8 *buffer = yaffs_GetTempBuffer(dev, __LINE__); ++ ++ oh = (yaffs_ObjectHeader *)buffer; ++ ++ yaffs_ReadChunkWithTagsFromNAND(dev, obj->hdrChunk, buffer, ++ &tags); ++ ++ yaffs_VerifyObjectHeader(obj, oh, &tags, 1); ++ ++ yaffs_ReleaseTempBuffer(dev, buffer, __LINE__); ++ } ++ ++ /* Verify it has a parent */ ++ if (obj && !obj->fake && ++ (!obj->parent || obj->parent->myDev != dev)) { ++ T(YAFFS_TRACE_VERIFY, ++ (TSTR("Obj %d has parent pointer %p which does not look like an object"TENDSTR), ++ obj->objectId, obj->parent)); ++ } ++ ++ /* Verify parent is a directory */ ++ if (obj->parent && obj->parent->variantType != YAFFS_OBJECT_TYPE_DIRECTORY) { ++ T(YAFFS_TRACE_VERIFY, ++ (TSTR("Obj %d's parent is not a directory (type %d)"TENDSTR), ++ obj->objectId, obj->parent->variantType)); ++ } ++ ++ switch (obj->variantType) { ++ case YAFFS_OBJECT_TYPE_FILE: ++ yaffs_VerifyFile(obj); ++ break; ++ case YAFFS_OBJECT_TYPE_SYMLINK: ++ yaffs_VerifySymlink(obj); ++ break; ++ case YAFFS_OBJECT_TYPE_DIRECTORY: ++ yaffs_VerifyDirectory(obj); ++ break; ++ case YAFFS_OBJECT_TYPE_HARDLINK: ++ yaffs_VerifyHardLink(obj); ++ break; ++ case YAFFS_OBJECT_TYPE_SPECIAL: ++ yaffs_VerifySpecial(obj); ++ break; ++ case YAFFS_OBJECT_TYPE_UNKNOWN: ++ default: ++ T(YAFFS_TRACE_VERIFY, ++ (TSTR("Obj %d has illegaltype %d"TENDSTR), ++ obj->objectId, obj->variantType)); ++ break; ++ } ++} ++ ++static void yaffs_VerifyObjects(yaffs_Device *dev) ++{ ++ yaffs_Object *obj; ++ int i; ++ struct ylist_head *lh; ++ ++ if (yaffs_SkipVerification(dev)) ++ return; ++ ++ /* Iterate through the objects in each hash entry */ ++ ++ for (i = 0; i < YAFFS_NOBJECT_BUCKETS; i++) { ++ ylist_for_each(lh, &dev->objectBucket[i].list) { ++ if (lh) { ++ obj = ylist_entry(lh, yaffs_Object, hashLink); ++ yaffs_VerifyObject(obj); ++ } ++ } ++ } ++} ++ ++ ++/* ++ * Simple hash function. Needs to have a reasonable spread ++ */ ++ ++static Y_INLINE int yaffs_HashFunction(int n) ++{ ++ n = abs(n); ++ return n % YAFFS_NOBJECT_BUCKETS; ++} ++ ++/* ++ * Access functions to useful fake objects. ++ * Note that root might have a presence in NAND if permissions are set. ++ */ ++ ++yaffs_Object *yaffs_Root(yaffs_Device *dev) ++{ ++ return dev->rootDir; ++} ++ ++yaffs_Object *yaffs_LostNFound(yaffs_Device *dev) ++{ ++ return dev->lostNFoundDir; ++} ++ ++ ++/* ++ * Erased NAND checking functions ++ */ ++ ++int yaffs_CheckFF(__u8 *buffer, int nBytes) ++{ ++ /* Horrible, slow implementation */ ++ while (nBytes--) { ++ if (*buffer != 0xFF) ++ return 0; ++ buffer++; ++ } ++ return 1; ++} ++ ++static int yaffs_CheckChunkErased(struct yaffs_DeviceStruct *dev, ++ int chunkInNAND) ++{ ++ int retval = YAFFS_OK; ++ __u8 *data = yaffs_GetTempBuffer(dev, __LINE__); ++ yaffs_ExtendedTags tags; ++ int result; ++ ++ result = yaffs_ReadChunkWithTagsFromNAND(dev, chunkInNAND, data, &tags); ++ ++ if (tags.eccResult > YAFFS_ECC_RESULT_NO_ERROR) ++ retval = YAFFS_FAIL; ++ ++ if (!yaffs_CheckFF(data, dev->nDataBytesPerChunk) || tags.chunkUsed) { ++ T(YAFFS_TRACE_NANDACCESS, ++ (TSTR("Chunk %d not erased" TENDSTR), chunkInNAND)); ++ retval = YAFFS_FAIL; ++ } ++ ++ yaffs_ReleaseTempBuffer(dev, data, __LINE__); ++ ++ return retval; ++ ++} ++ ++static int yaffs_WriteNewChunkWithTagsToNAND(struct yaffs_DeviceStruct *dev, ++ const __u8 *data, ++ yaffs_ExtendedTags *tags, ++ int useReserve) ++{ ++ int attempts = 0; ++ int writeOk = 0; ++ int chunk; ++ ++ yaffs_InvalidateCheckpoint(dev); ++ ++ do { ++ yaffs_BlockInfo *bi = 0; ++ int erasedOk = 0; ++ ++ chunk = yaffs_AllocateChunk(dev, useReserve, &bi); ++ if (chunk < 0) { ++ /* no space */ ++ break; ++ } ++ ++ /* First check this chunk is erased, if it needs ++ * checking. The checking policy (unless forced ++ * always on) is as follows: ++ * ++ * Check the first page we try to write in a block. ++ * If the check passes then we don't need to check any ++ * more. If the check fails, we check again... ++ * If the block has been erased, we don't need to check. ++ * ++ * However, if the block has been prioritised for gc, ++ * then we think there might be something odd about ++ * this block and stop using it. ++ * ++ * Rationale: We should only ever see chunks that have ++ * not been erased if there was a partially written ++ * chunk due to power loss. This checking policy should ++ * catch that case with very few checks and thus save a ++ * lot of checks that are most likely not needed. ++ */ ++ if (bi->gcPrioritise) { ++ yaffs_DeleteChunk(dev, chunk, 1, __LINE__); ++ /* try another chunk */ ++ continue; ++ } ++ ++ /* let's give it a try */ ++ attempts++; ++ ++#ifdef CONFIG_YAFFS_ALWAYS_CHECK_CHUNK_ERASED ++ bi->skipErasedCheck = 0; ++#endif ++ if (!bi->skipErasedCheck) { ++ erasedOk = yaffs_CheckChunkErased(dev, chunk); ++ if (erasedOk != YAFFS_OK) { ++ T(YAFFS_TRACE_ERROR, ++ (TSTR("**>> yaffs chunk %d was not erased" ++ TENDSTR), chunk)); ++ ++ /* try another chunk */ ++ continue; ++ } ++ bi->skipErasedCheck = 1; ++ } ++ ++ writeOk = yaffs_WriteChunkWithTagsToNAND(dev, chunk, ++ data, tags); ++ if (writeOk != YAFFS_OK) { ++ yaffs_HandleWriteChunkError(dev, chunk, erasedOk); ++ /* try another chunk */ ++ continue; ++ } ++ ++ /* Copy the data into the robustification buffer */ ++ yaffs_HandleWriteChunkOk(dev, chunk, data, tags); ++ ++ } while (writeOk != YAFFS_OK && ++ (yaffs_wr_attempts <= 0 || attempts <= yaffs_wr_attempts)); ++ ++ if (!writeOk) ++ chunk = -1; ++ ++ if (attempts > 1) { ++ T(YAFFS_TRACE_ERROR, ++ (TSTR("**>> yaffs write required %d attempts" TENDSTR), ++ attempts)); ++ ++ dev->nRetriedWrites += (attempts - 1); ++ } ++ ++ return chunk; ++} ++ ++/* ++ * Block retiring for handling a broken block. ++ */ ++ ++static void yaffs_RetireBlock(yaffs_Device *dev, int blockInNAND) ++{ ++ yaffs_BlockInfo *bi = yaffs_GetBlockInfo(dev, blockInNAND); ++ ++ yaffs_InvalidateCheckpoint(dev); ++ ++ if (yaffs_MarkBlockBad(dev, blockInNAND) != YAFFS_OK) { ++ if (yaffs_EraseBlockInNAND(dev, blockInNAND) != YAFFS_OK) { ++ T(YAFFS_TRACE_ALWAYS, (TSTR( ++ "yaffs: Failed to mark bad and erase block %d" ++ TENDSTR), blockInNAND)); ++ } else { ++ yaffs_ExtendedTags tags; ++ int chunkId = blockInNAND * dev->nChunksPerBlock; ++ ++ __u8 *buffer = yaffs_GetTempBuffer(dev, __LINE__); ++ ++ memset(buffer, 0xff, dev->nDataBytesPerChunk); ++ yaffs_InitialiseTags(&tags); ++ tags.sequenceNumber = YAFFS_SEQUENCE_BAD_BLOCK; ++ if (dev->writeChunkWithTagsToNAND(dev, chunkId - ++ dev->chunkOffset, buffer, &tags) != YAFFS_OK) ++ T(YAFFS_TRACE_ALWAYS, (TSTR("yaffs: Failed to " ++ TCONT("write bad block marker to block %d") ++ TENDSTR), blockInNAND)); ++ ++ yaffs_ReleaseTempBuffer(dev, buffer, __LINE__); ++ } ++ } ++ ++ bi->blockState = YAFFS_BLOCK_STATE_DEAD; ++ bi->gcPrioritise = 0; ++ bi->needsRetiring = 0; ++ ++ dev->nRetiredBlocks++; ++} ++ ++/* ++ * Functions for robustisizing TODO ++ * ++ */ ++ ++static void yaffs_HandleWriteChunkOk(yaffs_Device *dev, int chunkInNAND, ++ const __u8 *data, ++ const yaffs_ExtendedTags *tags) ++{ ++} ++ ++static void yaffs_HandleUpdateChunk(yaffs_Device *dev, int chunkInNAND, ++ const yaffs_ExtendedTags *tags) ++{ ++} ++ ++void yaffs_HandleChunkError(yaffs_Device *dev, yaffs_BlockInfo *bi) ++{ ++ if (!bi->gcPrioritise) { ++ bi->gcPrioritise = 1; ++ dev->hasPendingPrioritisedGCs = 1; ++ bi->chunkErrorStrikes++; ++ ++ if (bi->chunkErrorStrikes > 3) { ++ bi->needsRetiring = 1; /* Too many stikes, so retire this */ ++ T(YAFFS_TRACE_ALWAYS, (TSTR("yaffs: Block struck out" TENDSTR))); ++ ++ } ++ } ++} ++ ++static void yaffs_HandleWriteChunkError(yaffs_Device *dev, int chunkInNAND, ++ int erasedOk) ++{ ++ int blockInNAND = chunkInNAND / dev->nChunksPerBlock; ++ yaffs_BlockInfo *bi = yaffs_GetBlockInfo(dev, blockInNAND); ++ ++ yaffs_HandleChunkError(dev, bi); ++ ++ if (erasedOk) { ++ /* Was an actual write failure, so mark the block for retirement */ ++ bi->needsRetiring = 1; ++ T(YAFFS_TRACE_ERROR | YAFFS_TRACE_BAD_BLOCKS, ++ (TSTR("**>> Block %d needs retiring" TENDSTR), blockInNAND)); ++ } ++ ++ /* Delete the chunk */ ++ yaffs_DeleteChunk(dev, chunkInNAND, 1, __LINE__); ++} ++ ++ ++/*---------------- Name handling functions ------------*/ ++ ++static __u16 yaffs_CalcNameSum(const YCHAR *name) ++{ ++ __u16 sum = 0; ++ __u16 i = 1; ++ ++ const YUCHAR *bname = (const YUCHAR *) name; ++ if (bname) { ++ while ((*bname) && (i < (YAFFS_MAX_NAME_LENGTH/2))) { ++ ++#ifdef CONFIG_YAFFS_CASE_INSENSITIVE ++ sum += yaffs_toupper(*bname) * i; ++#else ++ sum += (*bname) * i; ++#endif ++ i++; ++ bname++; ++ } ++ } ++ return sum; ++} ++ ++static void yaffs_SetObjectName(yaffs_Object *obj, const YCHAR *name) ++{ ++#ifdef CONFIG_YAFFS_SHORT_NAMES_IN_RAM ++ memset(obj->shortName, 0, sizeof(YCHAR) * (YAFFS_SHORT_NAME_LENGTH+1)); ++ if (name && yaffs_strlen(name) <= YAFFS_SHORT_NAME_LENGTH) ++ yaffs_strcpy(obj->shortName, name); ++ else ++ obj->shortName[0] = _Y('\0'); ++#endif ++ obj->sum = yaffs_CalcNameSum(name); ++} ++ ++/*-------------------- TNODES ------------------- ++ ++ * List of spare tnodes ++ * The list is hooked together using the first pointer ++ * in the tnode. ++ */ ++ ++/* yaffs_CreateTnodes creates a bunch more tnodes and ++ * adds them to the tnode free list. ++ * Don't use this function directly ++ */ ++ ++static int yaffs_CreateTnodes(yaffs_Device *dev, int nTnodes) ++{ ++ int i; ++ int tnodeSize; ++ yaffs_Tnode *newTnodes; ++ __u8 *mem; ++ yaffs_Tnode *curr; ++ yaffs_Tnode *next; ++ yaffs_TnodeList *tnl; ++ ++ if (nTnodes < 1) ++ return YAFFS_OK; ++ ++ /* Calculate the tnode size in bytes for variable width tnode support. ++ * Must be a multiple of 32-bits */ ++ tnodeSize = (dev->tnodeWidth * YAFFS_NTNODES_LEVEL0)/8; ++ ++ if (tnodeSize < sizeof(yaffs_Tnode)) ++ tnodeSize = sizeof(yaffs_Tnode); ++ ++ /* make these things */ ++ ++ newTnodes = YMALLOC(nTnodes * tnodeSize); ++ mem = (__u8 *)newTnodes; ++ ++ if (!newTnodes) { ++ T(YAFFS_TRACE_ERROR, ++ (TSTR("yaffs: Could not allocate Tnodes" TENDSTR))); ++ return YAFFS_FAIL; ++ } ++ ++ /* Hook them into the free list */ ++#if 0 ++ for (i = 0; i < nTnodes - 1; i++) { ++ newTnodes[i].internal[0] = &newTnodes[i + 1]; ++#ifdef CONFIG_YAFFS_TNODE_LIST_DEBUG ++ newTnodes[i].internal[YAFFS_NTNODES_INTERNAL] = (void *)1; ++#endif ++ } ++ ++ newTnodes[nTnodes - 1].internal[0] = dev->freeTnodes; ++#ifdef CONFIG_YAFFS_TNODE_LIST_DEBUG ++ newTnodes[nTnodes - 1].internal[YAFFS_NTNODES_INTERNAL] = (void *)1; ++#endif ++ dev->freeTnodes = newTnodes; ++#else ++ /* New hookup for wide tnodes */ ++ for (i = 0; i < nTnodes - 1; i++) { ++ curr = (yaffs_Tnode *) &mem[i * tnodeSize]; ++ next = (yaffs_Tnode *) &mem[(i+1) * tnodeSize]; ++ curr->internal[0] = next; ++ } ++ ++ curr = (yaffs_Tnode *) &mem[(nTnodes - 1) * tnodeSize]; ++ curr->internal[0] = dev->freeTnodes; ++ dev->freeTnodes = (yaffs_Tnode *)mem; ++ ++#endif ++ ++ ++ dev->nFreeTnodes += nTnodes; ++ dev->nTnodesCreated += nTnodes; ++ ++ /* Now add this bunch of tnodes to a list for freeing up. ++ * NB If we can't add this to the management list it isn't fatal ++ * but it just means we can't free this bunch of tnodes later. ++ */ ++ ++ tnl = YMALLOC(sizeof(yaffs_TnodeList)); ++ if (!tnl) { ++ T(YAFFS_TRACE_ERROR, ++ (TSTR ++ ("yaffs: Could not add tnodes to management list" TENDSTR))); ++ return YAFFS_FAIL; ++ } else { ++ tnl->tnodes = newTnodes; ++ tnl->next = dev->allocatedTnodeList; ++ dev->allocatedTnodeList = tnl; ++ } ++ ++ T(YAFFS_TRACE_ALLOCATE, (TSTR("yaffs: Tnodes added" TENDSTR))); ++ ++ return YAFFS_OK; ++} ++ ++/* GetTnode gets us a clean tnode. Tries to make allocate more if we run out */ ++ ++static yaffs_Tnode *yaffs_GetTnodeRaw(yaffs_Device *dev) ++{ ++ yaffs_Tnode *tn = NULL; ++ ++ /* If there are none left make more */ ++ if (!dev->freeTnodes) ++ yaffs_CreateTnodes(dev, YAFFS_ALLOCATION_NTNODES); ++ ++ if (dev->freeTnodes) { ++ tn = dev->freeTnodes; ++#ifdef CONFIG_YAFFS_TNODE_LIST_DEBUG ++ if (tn->internal[YAFFS_NTNODES_INTERNAL] != (void *)1) { ++ /* Hoosterman, this thing looks like it isn't in the list */ ++ T(YAFFS_TRACE_ALWAYS, ++ (TSTR("yaffs: Tnode list bug 1" TENDSTR))); ++ } ++#endif ++ dev->freeTnodes = dev->freeTnodes->internal[0]; ++ dev->nFreeTnodes--; ++ } ++ ++ dev->nCheckpointBlocksRequired = 0; /* force recalculation*/ ++ ++ return tn; ++} ++ ++static yaffs_Tnode *yaffs_GetTnode(yaffs_Device *dev) ++{ ++ yaffs_Tnode *tn = yaffs_GetTnodeRaw(dev); ++ int tnodeSize = (dev->tnodeWidth * YAFFS_NTNODES_LEVEL0)/8; ++ ++ if (tnodeSize < sizeof(yaffs_Tnode)) ++ tnodeSize = sizeof(yaffs_Tnode); ++ ++ if (tn) ++ memset(tn, 0, tnodeSize); ++ ++ return tn; ++} ++ ++/* FreeTnode frees up a tnode and puts it back on the free list */ ++static void yaffs_FreeTnode(yaffs_Device *dev, yaffs_Tnode *tn) ++{ ++ if (tn) { ++#ifdef CONFIG_YAFFS_TNODE_LIST_DEBUG ++ if (tn->internal[YAFFS_NTNODES_INTERNAL] != 0) { ++ /* Hoosterman, this thing looks like it is already in the list */ ++ T(YAFFS_TRACE_ALWAYS, ++ (TSTR("yaffs: Tnode list bug 2" TENDSTR))); ++ } ++ tn->internal[YAFFS_NTNODES_INTERNAL] = (void *)1; ++#endif ++ tn->internal[0] = dev->freeTnodes; ++ dev->freeTnodes = tn; ++ dev->nFreeTnodes++; ++ } ++ dev->nCheckpointBlocksRequired = 0; /* force recalculation*/ ++} ++ ++static void yaffs_DeinitialiseTnodes(yaffs_Device *dev) ++{ ++ /* Free the list of allocated tnodes */ ++ yaffs_TnodeList *tmp; ++ ++ while (dev->allocatedTnodeList) { ++ tmp = dev->allocatedTnodeList->next; ++ ++ YFREE(dev->allocatedTnodeList->tnodes); ++ YFREE(dev->allocatedTnodeList); ++ dev->allocatedTnodeList = tmp; ++ ++ } ++ ++ dev->freeTnodes = NULL; ++ dev->nFreeTnodes = 0; ++} ++ ++static void yaffs_InitialiseTnodes(yaffs_Device *dev) ++{ ++ dev->allocatedTnodeList = NULL; ++ dev->freeTnodes = NULL; ++ dev->nFreeTnodes = 0; ++ dev->nTnodesCreated = 0; ++} ++ ++ ++void yaffs_PutLevel0Tnode(yaffs_Device *dev, yaffs_Tnode *tn, unsigned pos, ++ unsigned val) ++{ ++ __u32 *map = (__u32 *)tn; ++ __u32 bitInMap; ++ __u32 bitInWord; ++ __u32 wordInMap; ++ __u32 mask; ++ ++ pos &= YAFFS_TNODES_LEVEL0_MASK; ++ val >>= dev->chunkGroupBits; ++ ++ bitInMap = pos * dev->tnodeWidth; ++ wordInMap = bitInMap / 32; ++ bitInWord = bitInMap & (32 - 1); ++ ++ mask = dev->tnodeMask << bitInWord; ++ ++ map[wordInMap] &= ~mask; ++ map[wordInMap] |= (mask & (val << bitInWord)); ++ ++ if (dev->tnodeWidth > (32 - bitInWord)) { ++ bitInWord = (32 - bitInWord); ++ wordInMap++;; ++ mask = dev->tnodeMask >> (/*dev->tnodeWidth -*/ bitInWord); ++ map[wordInMap] &= ~mask; ++ map[wordInMap] |= (mask & (val >> bitInWord)); ++ } ++} ++ ++static __u32 yaffs_GetChunkGroupBase(yaffs_Device *dev, yaffs_Tnode *tn, ++ unsigned pos) ++{ ++ __u32 *map = (__u32 *)tn; ++ __u32 bitInMap; ++ __u32 bitInWord; ++ __u32 wordInMap; ++ __u32 val; ++ ++ pos &= YAFFS_TNODES_LEVEL0_MASK; ++ ++ bitInMap = pos * dev->tnodeWidth; ++ wordInMap = bitInMap / 32; ++ bitInWord = bitInMap & (32 - 1); ++ ++ val = map[wordInMap] >> bitInWord; ++ ++ if (dev->tnodeWidth > (32 - bitInWord)) { ++ bitInWord = (32 - bitInWord); ++ wordInMap++;; ++ val |= (map[wordInMap] << bitInWord); ++ } ++ ++ val &= dev->tnodeMask; ++ val <<= dev->chunkGroupBits; ++ ++ return val; ++} ++ ++/* ------------------- End of individual tnode manipulation -----------------*/ ++ ++/* ---------Functions to manipulate the look-up tree (made up of tnodes) ------ ++ * The look up tree is represented by the top tnode and the number of topLevel ++ * in the tree. 0 means only the level 0 tnode is in the tree. ++ */ ++ ++/* FindLevel0Tnode finds the level 0 tnode, if one exists. */ ++static yaffs_Tnode *yaffs_FindLevel0Tnode(yaffs_Device *dev, ++ yaffs_FileStructure *fStruct, ++ __u32 chunkId) ++{ ++ yaffs_Tnode *tn = fStruct->top; ++ __u32 i; ++ int requiredTallness; ++ int level = fStruct->topLevel; ++ ++ /* Check sane level and chunk Id */ ++ if (level < 0 || level > YAFFS_TNODES_MAX_LEVEL) ++ return NULL; ++ ++ if (chunkId > YAFFS_MAX_CHUNK_ID) ++ return NULL; ++ ++ /* First check we're tall enough (ie enough topLevel) */ ++ ++ i = chunkId >> YAFFS_TNODES_LEVEL0_BITS; ++ requiredTallness = 0; ++ while (i) { ++ i >>= YAFFS_TNODES_INTERNAL_BITS; ++ requiredTallness++; ++ } ++ ++ if (requiredTallness > fStruct->topLevel) ++ return NULL; /* Not tall enough, so we can't find it */ ++ ++ /* Traverse down to level 0 */ ++ while (level > 0 && tn) { ++ tn = tn->internal[(chunkId >> ++ (YAFFS_TNODES_LEVEL0_BITS + ++ (level - 1) * ++ YAFFS_TNODES_INTERNAL_BITS)) & ++ YAFFS_TNODES_INTERNAL_MASK]; ++ level--; ++ } ++ ++ return tn; ++} ++ ++/* AddOrFindLevel0Tnode finds the level 0 tnode if it exists, otherwise first expands the tree. ++ * This happens in two steps: ++ * 1. If the tree isn't tall enough, then make it taller. ++ * 2. Scan down the tree towards the level 0 tnode adding tnodes if required. ++ * ++ * Used when modifying the tree. ++ * ++ * If the tn argument is NULL, then a fresh tnode will be added otherwise the specified tn will ++ * be plugged into the ttree. ++ */ ++ ++static yaffs_Tnode *yaffs_AddOrFindLevel0Tnode(yaffs_Device *dev, ++ yaffs_FileStructure *fStruct, ++ __u32 chunkId, ++ yaffs_Tnode *passedTn) ++{ ++ int requiredTallness; ++ int i; ++ int l; ++ yaffs_Tnode *tn; ++ ++ __u32 x; ++ ++ ++ /* Check sane level and page Id */ ++ if (fStruct->topLevel < 0 || fStruct->topLevel > YAFFS_TNODES_MAX_LEVEL) ++ return NULL; ++ ++ if (chunkId > YAFFS_MAX_CHUNK_ID) ++ return NULL; ++ ++ /* First check we're tall enough (ie enough topLevel) */ ++ ++ x = chunkId >> YAFFS_TNODES_LEVEL0_BITS; ++ requiredTallness = 0; ++ while (x) { ++ x >>= YAFFS_TNODES_INTERNAL_BITS; ++ requiredTallness++; ++ } ++ ++ ++ if (requiredTallness > fStruct->topLevel) { ++ /* Not tall enough, gotta make the tree taller */ ++ for (i = fStruct->topLevel; i < requiredTallness; i++) { ++ ++ tn = yaffs_GetTnode(dev); ++ ++ if (tn) { ++ tn->internal[0] = fStruct->top; ++ fStruct->top = tn; ++ } else { ++ T(YAFFS_TRACE_ERROR, ++ (TSTR("yaffs: no more tnodes" TENDSTR))); ++ } ++ } ++ ++ fStruct->topLevel = requiredTallness; ++ } ++ ++ /* Traverse down to level 0, adding anything we need */ ++ ++ l = fStruct->topLevel; ++ tn = fStruct->top; ++ ++ if (l > 0) { ++ while (l > 0 && tn) { ++ x = (chunkId >> ++ (YAFFS_TNODES_LEVEL0_BITS + ++ (l - 1) * YAFFS_TNODES_INTERNAL_BITS)) & ++ YAFFS_TNODES_INTERNAL_MASK; ++ ++ ++ if ((l > 1) && !tn->internal[x]) { ++ /* Add missing non-level-zero tnode */ ++ tn->internal[x] = yaffs_GetTnode(dev); ++ ++ } else if (l == 1) { ++ /* Looking from level 1 at level 0 */ ++ if (passedTn) { ++ /* If we already have one, then release it.*/ ++ if (tn->internal[x]) ++ yaffs_FreeTnode(dev, tn->internal[x]); ++ tn->internal[x] = passedTn; ++ ++ } else if (!tn->internal[x]) { ++ /* Don't have one, none passed in */ ++ tn->internal[x] = yaffs_GetTnode(dev); ++ } ++ } ++ ++ tn = tn->internal[x]; ++ l--; ++ } ++ } else { ++ /* top is level 0 */ ++ if (passedTn) { ++ memcpy(tn, passedTn, (dev->tnodeWidth * YAFFS_NTNODES_LEVEL0)/8); ++ yaffs_FreeTnode(dev, passedTn); ++ } ++ } ++ ++ return tn; ++} ++ ++static int yaffs_FindChunkInGroup(yaffs_Device *dev, int theChunk, ++ yaffs_ExtendedTags *tags, int objectId, ++ int chunkInInode) ++{ ++ int j; ++ ++ for (j = 0; theChunk && j < dev->chunkGroupSize; j++) { ++ if (yaffs_CheckChunkBit(dev, theChunk / dev->nChunksPerBlock, ++ theChunk % dev->nChunksPerBlock)) { ++ yaffs_ReadChunkWithTagsFromNAND(dev, theChunk, NULL, ++ tags); ++ if (yaffs_TagsMatch(tags, objectId, chunkInInode)) { ++ /* found it; */ ++ return theChunk; ++ } ++ } ++ theChunk++; ++ } ++ return -1; ++} ++ ++ ++/* DeleteWorker scans backwards through the tnode tree and deletes all the ++ * chunks and tnodes in the file ++ * Returns 1 if the tree was deleted. ++ * Returns 0 if it stopped early due to hitting the limit and the delete is incomplete. ++ */ ++ ++static int yaffs_DeleteWorker(yaffs_Object *in, yaffs_Tnode *tn, __u32 level, ++ int chunkOffset, int *limit) ++{ ++ int i; ++ int chunkInInode; ++ int theChunk; ++ yaffs_ExtendedTags tags; ++ int foundChunk; ++ yaffs_Device *dev = in->myDev; ++ ++ int allDone = 1; ++ ++ if (tn) { ++ if (level > 0) { ++ for (i = YAFFS_NTNODES_INTERNAL - 1; allDone && i >= 0; ++ i--) { ++ if (tn->internal[i]) { ++ if (limit && (*limit) < 0) { ++ allDone = 0; ++ } else { ++ allDone = ++ yaffs_DeleteWorker(in, ++ tn-> ++ internal ++ [i], ++ level - ++ 1, ++ (chunkOffset ++ << ++ YAFFS_TNODES_INTERNAL_BITS) ++ + i, ++ limit); ++ } ++ if (allDone) { ++ yaffs_FreeTnode(dev, ++ tn-> ++ internal[i]); ++ tn->internal[i] = NULL; ++ } ++ } ++ } ++ return (allDone) ? 1 : 0; ++ } else if (level == 0) { ++ int hitLimit = 0; ++ ++ for (i = YAFFS_NTNODES_LEVEL0 - 1; i >= 0 && !hitLimit; ++ i--) { ++ theChunk = yaffs_GetChunkGroupBase(dev, tn, i); ++ if (theChunk) { ++ ++ chunkInInode = (chunkOffset << ++ YAFFS_TNODES_LEVEL0_BITS) + i; ++ ++ foundChunk = ++ yaffs_FindChunkInGroup(dev, ++ theChunk, ++ &tags, ++ in->objectId, ++ chunkInInode); ++ ++ if (foundChunk > 0) { ++ yaffs_DeleteChunk(dev, ++ foundChunk, 1, ++ __LINE__); ++ in->nDataChunks--; ++ if (limit) { ++ *limit = *limit - 1; ++ if (*limit <= 0) ++ hitLimit = 1; ++ } ++ ++ } ++ ++ yaffs_PutLevel0Tnode(dev, tn, i, 0); ++ } ++ ++ } ++ return (i < 0) ? 1 : 0; ++ ++ } ++ ++ } ++ ++ return 1; ++ ++} ++ ++static void yaffs_SoftDeleteChunk(yaffs_Device *dev, int chunk) ++{ ++ yaffs_BlockInfo *theBlock; ++ ++ T(YAFFS_TRACE_DELETION, (TSTR("soft delete chunk %d" TENDSTR), chunk)); ++ ++ theBlock = yaffs_GetBlockInfo(dev, chunk / dev->nChunksPerBlock); ++ if (theBlock) { ++ theBlock->softDeletions++; ++ dev->nFreeChunks++; ++ } ++} ++ ++/* SoftDeleteWorker scans backwards through the tnode tree and soft deletes all the chunks in the file. ++ * All soft deleting does is increment the block's softdelete count and pulls the chunk out ++ * of the tnode. ++ * Thus, essentially this is the same as DeleteWorker except that the chunks are soft deleted. ++ */ ++ ++static int yaffs_SoftDeleteWorker(yaffs_Object *in, yaffs_Tnode *tn, ++ __u32 level, int chunkOffset) ++{ ++ int i; ++ int theChunk; ++ int allDone = 1; ++ yaffs_Device *dev = in->myDev; ++ ++ if (tn) { ++ if (level > 0) { ++ ++ for (i = YAFFS_NTNODES_INTERNAL - 1; allDone && i >= 0; ++ i--) { ++ if (tn->internal[i]) { ++ allDone = ++ yaffs_SoftDeleteWorker(in, ++ tn-> ++ internal[i], ++ level - 1, ++ (chunkOffset ++ << ++ YAFFS_TNODES_INTERNAL_BITS) ++ + i); ++ if (allDone) { ++ yaffs_FreeTnode(dev, ++ tn-> ++ internal[i]); ++ tn->internal[i] = NULL; ++ } else { ++ /* Hoosterman... how could this happen? */ ++ } ++ } ++ } ++ return (allDone) ? 1 : 0; ++ } else if (level == 0) { ++ ++ for (i = YAFFS_NTNODES_LEVEL0 - 1; i >= 0; i--) { ++ theChunk = yaffs_GetChunkGroupBase(dev, tn, i); ++ if (theChunk) { ++ /* Note this does not find the real chunk, only the chunk group. ++ * We make an assumption that a chunk group is not larger than ++ * a block. ++ */ ++ yaffs_SoftDeleteChunk(dev, theChunk); ++ yaffs_PutLevel0Tnode(dev, tn, i, 0); ++ } ++ ++ } ++ return 1; ++ ++ } ++ ++ } ++ ++ return 1; ++ ++} ++ ++static void yaffs_SoftDeleteFile(yaffs_Object *obj) ++{ ++ if (obj->deleted && ++ obj->variantType == YAFFS_OBJECT_TYPE_FILE && !obj->softDeleted) { ++ if (obj->nDataChunks <= 0) { ++ /* Empty file with no duplicate object headers, just delete it immediately */ ++ yaffs_FreeTnode(obj->myDev, ++ obj->variant.fileVariant.top); ++ obj->variant.fileVariant.top = NULL; ++ T(YAFFS_TRACE_TRACING, ++ (TSTR("yaffs: Deleting empty file %d" TENDSTR), ++ obj->objectId)); ++ yaffs_DoGenericObjectDeletion(obj); ++ } else { ++ yaffs_SoftDeleteWorker(obj, ++ obj->variant.fileVariant.top, ++ obj->variant.fileVariant. ++ topLevel, 0); ++ obj->softDeleted = 1; ++ } ++ } ++} ++ ++/* Pruning removes any part of the file structure tree that is beyond the ++ * bounds of the file (ie that does not point to chunks). ++ * ++ * A file should only get pruned when its size is reduced. ++ * ++ * Before pruning, the chunks must be pulled from the tree and the ++ * level 0 tnode entries must be zeroed out. ++ * Could also use this for file deletion, but that's probably better handled ++ * by a special case. ++ */ ++ ++static yaffs_Tnode *yaffs_PruneWorker(yaffs_Device *dev, yaffs_Tnode *tn, ++ __u32 level, int del0) ++{ ++ int i; ++ int hasData; ++ ++ if (tn) { ++ hasData = 0; ++ ++ for (i = 0; i < YAFFS_NTNODES_INTERNAL; i++) { ++ if (tn->internal[i] && level > 0) { ++ tn->internal[i] = ++ yaffs_PruneWorker(dev, tn->internal[i], ++ level - 1, ++ (i == 0) ? del0 : 1); ++ } ++ ++ if (tn->internal[i]) ++ hasData++; ++ } ++ ++ if (hasData == 0 && del0) { ++ /* Free and return NULL */ ++ ++ yaffs_FreeTnode(dev, tn); ++ tn = NULL; ++ } ++ ++ } ++ ++ return tn; ++ ++} ++ ++static int yaffs_PruneFileStructure(yaffs_Device *dev, ++ yaffs_FileStructure *fStruct) ++{ ++ int i; ++ int hasData; ++ int done = 0; ++ yaffs_Tnode *tn; ++ ++ if (fStruct->topLevel > 0) { ++ fStruct->top = ++ yaffs_PruneWorker(dev, fStruct->top, fStruct->topLevel, 0); ++ ++ /* Now we have a tree with all the non-zero branches NULL but the height ++ * is the same as it was. ++ * Let's see if we can trim internal tnodes to shorten the tree. ++ * We can do this if only the 0th element in the tnode is in use ++ * (ie all the non-zero are NULL) ++ */ ++ ++ while (fStruct->topLevel && !done) { ++ tn = fStruct->top; ++ ++ hasData = 0; ++ for (i = 1; i < YAFFS_NTNODES_INTERNAL; i++) { ++ if (tn->internal[i]) ++ hasData++; ++ } ++ ++ if (!hasData) { ++ fStruct->top = tn->internal[0]; ++ fStruct->topLevel--; ++ yaffs_FreeTnode(dev, tn); ++ } else { ++ done = 1; ++ } ++ } ++ } ++ ++ return YAFFS_OK; ++} ++ ++/*-------------------- End of File Structure functions.-------------------*/ ++ ++/* yaffs_CreateFreeObjects creates a bunch more objects and ++ * adds them to the object free list. ++ */ ++static int yaffs_CreateFreeObjects(yaffs_Device *dev, int nObjects) ++{ ++ int i; ++ yaffs_Object *newObjects; ++ yaffs_ObjectList *list; ++ ++ if (nObjects < 1) ++ return YAFFS_OK; ++ ++ /* make these things */ ++ newObjects = YMALLOC(nObjects * sizeof(yaffs_Object)); ++ list = YMALLOC(sizeof(yaffs_ObjectList)); ++ ++ if (!newObjects || !list) { ++ if (newObjects) ++ YFREE(newObjects); ++ if (list) ++ YFREE(list); ++ T(YAFFS_TRACE_ALLOCATE, ++ (TSTR("yaffs: Could not allocate more objects" TENDSTR))); ++ return YAFFS_FAIL; ++ } ++ ++ /* Hook them into the free list */ ++ for (i = 0; i < nObjects - 1; i++) { ++ newObjects[i].siblings.next = ++ (struct ylist_head *)(&newObjects[i + 1]); ++ } ++ ++ newObjects[nObjects - 1].siblings.next = (void *)dev->freeObjects; ++ dev->freeObjects = newObjects; ++ dev->nFreeObjects += nObjects; ++ dev->nObjectsCreated += nObjects; ++ ++ /* Now add this bunch of Objects to a list for freeing up. */ ++ ++ list->objects = newObjects; ++ list->next = dev->allocatedObjectList; ++ dev->allocatedObjectList = list; ++ ++ return YAFFS_OK; ++} ++ ++ ++/* AllocateEmptyObject gets us a clean Object. Tries to make allocate more if we run out */ ++static yaffs_Object *yaffs_AllocateEmptyObject(yaffs_Device *dev) ++{ ++ yaffs_Object *tn = NULL; ++ ++#ifdef VALGRIND_TEST ++ tn = YMALLOC(sizeof(yaffs_Object)); ++#else ++ /* If there are none left make more */ ++ if (!dev->freeObjects) ++ yaffs_CreateFreeObjects(dev, YAFFS_ALLOCATION_NOBJECTS); ++ ++ if (dev->freeObjects) { ++ tn = dev->freeObjects; ++ dev->freeObjects = ++ (yaffs_Object *) (dev->freeObjects->siblings.next); ++ dev->nFreeObjects--; ++ } ++#endif ++ if (tn) { ++ /* Now sweeten it up... */ ++ ++ memset(tn, 0, sizeof(yaffs_Object)); ++ tn->beingCreated = 1; ++ ++ tn->myDev = dev; ++ tn->hdrChunk = 0; ++ tn->variantType = YAFFS_OBJECT_TYPE_UNKNOWN; ++ YINIT_LIST_HEAD(&(tn->hardLinks)); ++ YINIT_LIST_HEAD(&(tn->hashLink)); ++ YINIT_LIST_HEAD(&tn->siblings); ++ ++ ++ /* Now make the directory sane */ ++ if (dev->rootDir) { ++ tn->parent = dev->rootDir; ++ ylist_add(&(tn->siblings), &dev->rootDir->variant.directoryVariant.children); ++ } ++ ++ /* Add it to the lost and found directory. ++ * NB Can't put root or lostNFound in lostNFound so ++ * check if lostNFound exists first ++ */ ++ if (dev->lostNFoundDir) ++ yaffs_AddObjectToDirectory(dev->lostNFoundDir, tn); ++ ++ tn->beingCreated = 0; ++ } ++ ++ dev->nCheckpointBlocksRequired = 0; /* force recalculation*/ ++ ++ return tn; ++} ++ ++static yaffs_Object *yaffs_CreateFakeDirectory(yaffs_Device *dev, int number, ++ __u32 mode) ++{ ++ ++ yaffs_Object *obj = ++ yaffs_CreateNewObject(dev, number, YAFFS_OBJECT_TYPE_DIRECTORY); ++ if (obj) { ++ obj->fake = 1; /* it is fake so it might have no NAND presence... */ ++ obj->renameAllowed = 0; /* ... and we're not allowed to rename it... */ ++ obj->unlinkAllowed = 0; /* ... or unlink it */ ++ obj->deleted = 0; ++ obj->unlinked = 0; ++ obj->yst_mode = mode; ++ obj->myDev = dev; ++ obj->hdrChunk = 0; /* Not a valid chunk. */ ++ } ++ ++ return obj; ++ ++} ++ ++static void yaffs_UnhashObject(yaffs_Object *tn) ++{ ++ int bucket; ++ yaffs_Device *dev = tn->myDev; ++ ++ /* If it is still linked into the bucket list, free from the list */ ++ if (!ylist_empty(&tn->hashLink)) { ++ ylist_del_init(&tn->hashLink); ++ bucket = yaffs_HashFunction(tn->objectId); ++ dev->objectBucket[bucket].count--; ++ } ++} ++ ++/* FreeObject frees up a Object and puts it back on the free list */ ++static void yaffs_FreeObject(yaffs_Object *tn) ++{ ++ yaffs_Device *dev = tn->myDev; ++ ++#ifdef __KERNEL__ ++ T(YAFFS_TRACE_OS, (TSTR("FreeObject %p inode %p"TENDSTR), tn, tn->myInode)); ++#endif ++ ++ if (tn->parent) ++ YBUG(); ++ if (!ylist_empty(&tn->siblings)) ++ YBUG(); ++ ++ ++#ifdef __KERNEL__ ++ if (tn->myInode) { ++ /* We're still hooked up to a cached inode. ++ * Don't delete now, but mark for later deletion ++ */ ++ tn->deferedFree = 1; ++ return; ++ } ++#endif ++ ++ yaffs_UnhashObject(tn); ++ ++#ifdef VALGRIND_TEST ++ YFREE(tn); ++#else ++ /* Link into the free list. */ ++ tn->siblings.next = (struct ylist_head *)(dev->freeObjects); ++ dev->freeObjects = tn; ++ dev->nFreeObjects++; ++#endif ++ dev->nCheckpointBlocksRequired = 0; /* force recalculation*/ ++} ++ ++#ifdef __KERNEL__ ++ ++void yaffs_HandleDeferedFree(yaffs_Object *obj) ++{ ++ if (obj->deferedFree) ++ yaffs_FreeObject(obj); ++} ++ ++#endif ++ ++static void yaffs_DeinitialiseObjects(yaffs_Device *dev) ++{ ++ /* Free the list of allocated Objects */ ++ ++ yaffs_ObjectList *tmp; ++ ++ while (dev->allocatedObjectList) { ++ tmp = dev->allocatedObjectList->next; ++ YFREE(dev->allocatedObjectList->objects); ++ YFREE(dev->allocatedObjectList); ++ ++ dev->allocatedObjectList = tmp; ++ } ++ ++ dev->freeObjects = NULL; ++ dev->nFreeObjects = 0; ++} ++ ++static void yaffs_InitialiseObjects(yaffs_Device *dev) ++{ ++ int i; ++ ++ dev->allocatedObjectList = NULL; ++ dev->freeObjects = NULL; ++ dev->nFreeObjects = 0; ++ ++ for (i = 0; i < YAFFS_NOBJECT_BUCKETS; i++) { ++ YINIT_LIST_HEAD(&dev->objectBucket[i].list); ++ dev->objectBucket[i].count = 0; ++ } ++} ++ ++static int yaffs_FindNiceObjectBucket(yaffs_Device *dev) ++{ ++ static int x; ++ int i; ++ int l = 999; ++ int lowest = 999999; ++ ++ /* First let's see if we can find one that's empty. */ ++ ++ for (i = 0; i < 10 && lowest > 0; i++) { ++ x++; ++ x %= YAFFS_NOBJECT_BUCKETS; ++ if (dev->objectBucket[x].count < lowest) { ++ lowest = dev->objectBucket[x].count; ++ l = x; ++ } ++ ++ } ++ ++ /* If we didn't find an empty list, then try ++ * looking a bit further for a short one ++ */ ++ ++ for (i = 0; i < 10 && lowest > 3; i++) { ++ x++; ++ x %= YAFFS_NOBJECT_BUCKETS; ++ if (dev->objectBucket[x].count < lowest) { ++ lowest = dev->objectBucket[x].count; ++ l = x; ++ } ++ ++ } ++ ++ return l; ++} ++ ++static int yaffs_CreateNewObjectNumber(yaffs_Device *dev) ++{ ++ int bucket = yaffs_FindNiceObjectBucket(dev); ++ ++ /* Now find an object value that has not already been taken ++ * by scanning the list. ++ */ ++ ++ int found = 0; ++ struct ylist_head *i; ++ ++ __u32 n = (__u32) bucket; ++ ++ /* yaffs_CheckObjectHashSanity(); */ ++ ++ while (!found) { ++ found = 1; ++ n += YAFFS_NOBJECT_BUCKETS; ++ if (1 || dev->objectBucket[bucket].count > 0) { ++ ylist_for_each(i, &dev->objectBucket[bucket].list) { ++ /* If there is already one in the list */ ++ if (i && ylist_entry(i, yaffs_Object, ++ hashLink)->objectId == n) { ++ found = 0; ++ } ++ } ++ } ++ } ++ ++ return n; ++} ++ ++static void yaffs_HashObject(yaffs_Object *in) ++{ ++ int bucket = yaffs_HashFunction(in->objectId); ++ yaffs_Device *dev = in->myDev; ++ ++ ylist_add(&in->hashLink, &dev->objectBucket[bucket].list); ++ dev->objectBucket[bucket].count++; ++} ++ ++yaffs_Object *yaffs_FindObjectByNumber(yaffs_Device *dev, __u32 number) ++{ ++ int bucket = yaffs_HashFunction(number); ++ struct ylist_head *i; ++ yaffs_Object *in; ++ ++ ylist_for_each(i, &dev->objectBucket[bucket].list) { ++ /* Look if it is in the list */ ++ if (i) { ++ in = ylist_entry(i, yaffs_Object, hashLink); ++ if (in->objectId == number) { ++#ifdef __KERNEL__ ++ /* Don't tell the VFS about this one if it is defered free */ ++ if (in->deferedFree) ++ return NULL; ++#endif ++ ++ return in; ++ } ++ } ++ } ++ ++ return NULL; ++} ++ ++yaffs_Object *yaffs_CreateNewObject(yaffs_Device *dev, int number, ++ yaffs_ObjectType type) ++{ ++ yaffs_Object *theObject; ++ yaffs_Tnode *tn = NULL; ++ ++ if (number < 0) ++ number = yaffs_CreateNewObjectNumber(dev); ++ ++ theObject = yaffs_AllocateEmptyObject(dev); ++ if (!theObject) ++ return NULL; ++ ++ if (type == YAFFS_OBJECT_TYPE_FILE) { ++ tn = yaffs_GetTnode(dev); ++ if (!tn) { ++ yaffs_FreeObject(theObject); ++ return NULL; ++ } ++ } ++ ++ if (theObject) { ++ theObject->fake = 0; ++ theObject->renameAllowed = 1; ++ theObject->unlinkAllowed = 1; ++ theObject->objectId = number; ++ yaffs_HashObject(theObject); ++ theObject->variantType = type; ++#ifdef CONFIG_YAFFS_WINCE ++ yfsd_WinFileTimeNow(theObject->win_atime); ++ theObject->win_ctime[0] = theObject->win_mtime[0] = ++ theObject->win_atime[0]; ++ theObject->win_ctime[1] = theObject->win_mtime[1] = ++ theObject->win_atime[1]; ++ ++#else ++ ++ theObject->yst_atime = theObject->yst_mtime = ++ theObject->yst_ctime = Y_CURRENT_TIME; ++#endif ++ switch (type) { ++ case YAFFS_OBJECT_TYPE_FILE: ++ theObject->variant.fileVariant.fileSize = 0; ++ theObject->variant.fileVariant.scannedFileSize = 0; ++ theObject->variant.fileVariant.shrinkSize = 0xFFFFFFFF; /* max __u32 */ ++ theObject->variant.fileVariant.topLevel = 0; ++ theObject->variant.fileVariant.top = tn; ++ break; ++ case YAFFS_OBJECT_TYPE_DIRECTORY: ++ YINIT_LIST_HEAD(&theObject->variant.directoryVariant. ++ children); ++ break; ++ case YAFFS_OBJECT_TYPE_SYMLINK: ++ case YAFFS_OBJECT_TYPE_HARDLINK: ++ case YAFFS_OBJECT_TYPE_SPECIAL: ++ /* No action required */ ++ break; ++ case YAFFS_OBJECT_TYPE_UNKNOWN: ++ /* todo this should not happen */ ++ break; ++ } ++ } ++ ++ return theObject; ++} ++ ++static yaffs_Object *yaffs_FindOrCreateObjectByNumber(yaffs_Device *dev, ++ int number, ++ yaffs_ObjectType type) ++{ ++ yaffs_Object *theObject = NULL; ++ ++ if (number > 0) ++ theObject = yaffs_FindObjectByNumber(dev, number); ++ ++ if (!theObject) ++ theObject = yaffs_CreateNewObject(dev, number, type); ++ ++ return theObject; ++ ++} ++ ++ ++static YCHAR *yaffs_CloneString(const YCHAR *str) ++{ ++ YCHAR *newStr = NULL; ++ ++ if (str && *str) { ++ newStr = YMALLOC((yaffs_strlen(str) + 1) * sizeof(YCHAR)); ++ if (newStr) ++ yaffs_strcpy(newStr, str); ++ } ++ ++ return newStr; ++ ++} ++ ++/* ++ * Mknod (create) a new object. ++ * equivalentObject only has meaning for a hard link; ++ * aliasString only has meaning for a sumlink. ++ * rdev only has meaning for devices (a subset of special objects) ++ */ ++ ++static yaffs_Object *yaffs_MknodObject(yaffs_ObjectType type, ++ yaffs_Object *parent, ++ const YCHAR *name, ++ __u32 mode, ++ __u32 uid, ++ __u32 gid, ++ yaffs_Object *equivalentObject, ++ const YCHAR *aliasString, __u32 rdev) ++{ ++ yaffs_Object *in; ++ YCHAR *str = NULL; ++ ++ yaffs_Device *dev = parent->myDev; ++ ++ /* Check if the entry exists. If it does then fail the call since we don't want a dup.*/ ++ if (yaffs_FindObjectByName(parent, name)) ++ return NULL; ++ ++ in = yaffs_CreateNewObject(dev, -1, type); ++ ++ if (!in) ++ return YAFFS_FAIL; ++ ++ if (type == YAFFS_OBJECT_TYPE_SYMLINK) { ++ str = yaffs_CloneString(aliasString); ++ if (!str) { ++ yaffs_FreeObject(in); ++ return NULL; ++ } ++ } ++ ++ ++ ++ if (in) { ++ in->hdrChunk = 0; ++ in->valid = 1; ++ in->variantType = type; ++ ++ in->yst_mode = mode; ++ ++#ifdef CONFIG_YAFFS_WINCE ++ yfsd_WinFileTimeNow(in->win_atime); ++ in->win_ctime[0] = in->win_mtime[0] = in->win_atime[0]; ++ in->win_ctime[1] = in->win_mtime[1] = in->win_atime[1]; ++ ++#else ++ in->yst_atime = in->yst_mtime = in->yst_ctime = Y_CURRENT_TIME; ++ ++ in->yst_rdev = rdev; ++ in->yst_uid = uid; ++ in->yst_gid = gid; ++#endif ++ in->nDataChunks = 0; ++ ++ yaffs_SetObjectName(in, name); ++ in->dirty = 1; ++ ++ yaffs_AddObjectToDirectory(parent, in); ++ ++ in->myDev = parent->myDev; ++ ++ switch (type) { ++ case YAFFS_OBJECT_TYPE_SYMLINK: ++ in->variant.symLinkVariant.alias = str; ++ break; ++ case YAFFS_OBJECT_TYPE_HARDLINK: ++ in->variant.hardLinkVariant.equivalentObject = ++ equivalentObject; ++ in->variant.hardLinkVariant.equivalentObjectId = ++ equivalentObject->objectId; ++ ylist_add(&in->hardLinks, &equivalentObject->hardLinks); ++ break; ++ case YAFFS_OBJECT_TYPE_FILE: ++ case YAFFS_OBJECT_TYPE_DIRECTORY: ++ case YAFFS_OBJECT_TYPE_SPECIAL: ++ case YAFFS_OBJECT_TYPE_UNKNOWN: ++ /* do nothing */ ++ break; ++ } ++ ++ if (yaffs_UpdateObjectHeader(in, name, 0, 0, 0) < 0) { ++ /* Could not create the object header, fail the creation */ ++ yaffs_DeleteObject(in); ++ in = NULL; ++ } ++ ++ } ++ ++ return in; ++} ++ ++yaffs_Object *yaffs_MknodFile(yaffs_Object *parent, const YCHAR *name, ++ __u32 mode, __u32 uid, __u32 gid) ++{ ++ return yaffs_MknodObject(YAFFS_OBJECT_TYPE_FILE, parent, name, mode, ++ uid, gid, NULL, NULL, 0); ++} ++ ++yaffs_Object *yaffs_MknodDirectory(yaffs_Object *parent, const YCHAR *name, ++ __u32 mode, __u32 uid, __u32 gid) ++{ ++ return yaffs_MknodObject(YAFFS_OBJECT_TYPE_DIRECTORY, parent, name, ++ mode, uid, gid, NULL, NULL, 0); ++} ++ ++yaffs_Object *yaffs_MknodSpecial(yaffs_Object *parent, const YCHAR *name, ++ __u32 mode, __u32 uid, __u32 gid, __u32 rdev) ++{ ++ return yaffs_MknodObject(YAFFS_OBJECT_TYPE_SPECIAL, parent, name, mode, ++ uid, gid, NULL, NULL, rdev); ++} ++ ++yaffs_Object *yaffs_MknodSymLink(yaffs_Object *parent, const YCHAR *name, ++ __u32 mode, __u32 uid, __u32 gid, ++ const YCHAR *alias) ++{ ++ return yaffs_MknodObject(YAFFS_OBJECT_TYPE_SYMLINK, parent, name, mode, ++ uid, gid, NULL, alias, 0); ++} ++ ++/* yaffs_Link returns the object id of the equivalent object.*/ ++yaffs_Object *yaffs_Link(yaffs_Object *parent, const YCHAR *name, ++ yaffs_Object *equivalentObject) ++{ ++ /* Get the real object in case we were fed a hard link as an equivalent object */ ++ equivalentObject = yaffs_GetEquivalentObject(equivalentObject); ++ ++ if (yaffs_MknodObject ++ (YAFFS_OBJECT_TYPE_HARDLINK, parent, name, 0, 0, 0, ++ equivalentObject, NULL, 0)) { ++ return equivalentObject; ++ } else { ++ return NULL; ++ } ++ ++} ++ ++static int yaffs_ChangeObjectName(yaffs_Object *obj, yaffs_Object *newDir, ++ const YCHAR *newName, int force, int shadows) ++{ ++ int unlinkOp; ++ int deleteOp; ++ ++ yaffs_Object *existingTarget; ++ ++ if (newDir == NULL) ++ newDir = obj->parent; /* use the old directory */ ++ ++ if (newDir->variantType != YAFFS_OBJECT_TYPE_DIRECTORY) { ++ T(YAFFS_TRACE_ALWAYS, ++ (TSTR ++ ("tragedy: yaffs_ChangeObjectName: newDir is not a directory" ++ TENDSTR))); ++ YBUG(); ++ } ++ ++ /* TODO: Do we need this different handling for YAFFS2 and YAFFS1?? */ ++ if (obj->myDev->isYaffs2) ++ unlinkOp = (newDir == obj->myDev->unlinkedDir); ++ else ++ unlinkOp = (newDir == obj->myDev->unlinkedDir ++ && obj->variantType == YAFFS_OBJECT_TYPE_FILE); ++ ++ deleteOp = (newDir == obj->myDev->deletedDir); ++ ++ existingTarget = yaffs_FindObjectByName(newDir, newName); ++ ++ /* If the object is a file going into the unlinked directory, ++ * then it is OK to just stuff it in since duplicate names are allowed. ++ * else only proceed if the new name does not exist and if we're putting ++ * it into a directory. ++ */ ++ if ((unlinkOp || ++ deleteOp || ++ force || ++ (shadows > 0) || ++ !existingTarget) && ++ newDir->variantType == YAFFS_OBJECT_TYPE_DIRECTORY) { ++ yaffs_SetObjectName(obj, newName); ++ obj->dirty = 1; ++ ++ yaffs_AddObjectToDirectory(newDir, obj); ++ ++ if (unlinkOp) ++ obj->unlinked = 1; ++ ++ /* If it is a deletion then we mark it as a shrink for gc purposes. */ ++ if (yaffs_UpdateObjectHeader(obj, newName, 0, deleteOp, shadows) >= 0) ++ return YAFFS_OK; ++ } ++ ++ return YAFFS_FAIL; ++} ++ ++int yaffs_RenameObject(yaffs_Object *oldDir, const YCHAR *oldName, ++ yaffs_Object *newDir, const YCHAR *newName) ++{ ++ yaffs_Object *obj = NULL; ++ yaffs_Object *existingTarget = NULL; ++ int force = 0; ++ ++ ++ if (!oldDir || oldDir->variantType != YAFFS_OBJECT_TYPE_DIRECTORY) ++ YBUG(); ++ if (!newDir || newDir->variantType != YAFFS_OBJECT_TYPE_DIRECTORY) ++ YBUG(); ++ ++#ifdef CONFIG_YAFFS_CASE_INSENSITIVE ++ /* Special case for case insemsitive systems (eg. WinCE). ++ * While look-up is case insensitive, the name isn't. ++ * Therefore we might want to change x.txt to X.txt ++ */ ++ if (oldDir == newDir && yaffs_strcmp(oldName, newName) == 0) ++ force = 1; ++#endif ++ ++ else if (yaffs_strlen(newName) > YAFFS_MAX_NAME_LENGTH) ++ /* ENAMETOOLONG */ ++ return YAFFS_FAIL; ++ ++ obj = yaffs_FindObjectByName(oldDir, oldName); ++ ++ if (obj && obj->renameAllowed) { ++ ++ /* Now do the handling for an existing target, if there is one */ ++ ++ existingTarget = yaffs_FindObjectByName(newDir, newName); ++ if (existingTarget && ++ existingTarget->variantType == YAFFS_OBJECT_TYPE_DIRECTORY && ++ !ylist_empty(&existingTarget->variant.directoryVariant.children)) { ++ /* There is a target that is a non-empty directory, so we fail */ ++ return YAFFS_FAIL; /* EEXIST or ENOTEMPTY */ ++ } else if (existingTarget && existingTarget != obj) { ++ /* Nuke the target first, using shadowing, ++ * but only if it isn't the same object ++ */ ++ yaffs_ChangeObjectName(obj, newDir, newName, force, ++ existingTarget->objectId); ++ yaffs_UnlinkObject(existingTarget); ++ } ++ ++ return yaffs_ChangeObjectName(obj, newDir, newName, 1, 0); ++ } ++ return YAFFS_FAIL; ++} ++ ++/*------------------------- Block Management and Page Allocation ----------------*/ ++ ++static int yaffs_InitialiseBlocks(yaffs_Device *dev) ++{ ++ int nBlocks = dev->internalEndBlock - dev->internalStartBlock + 1; ++ ++ dev->blockInfo = NULL; ++ dev->chunkBits = NULL; ++ ++ dev->allocationBlock = -1; /* force it to get a new one */ ++ ++ /* If the first allocation strategy fails, thry the alternate one */ ++ dev->blockInfo = YMALLOC(nBlocks * sizeof(yaffs_BlockInfo)); ++ if (!dev->blockInfo) { ++ dev->blockInfo = YMALLOC_ALT(nBlocks * sizeof(yaffs_BlockInfo)); ++ dev->blockInfoAlt = 1; ++ } else ++ dev->blockInfoAlt = 0; ++ ++ if (dev->blockInfo) { ++ /* Set up dynamic blockinfo stuff. */ ++ dev->chunkBitmapStride = (dev->nChunksPerBlock + 7) / 8; /* round up bytes */ ++ dev->chunkBits = YMALLOC(dev->chunkBitmapStride * nBlocks); ++ if (!dev->chunkBits) { ++ dev->chunkBits = YMALLOC_ALT(dev->chunkBitmapStride * nBlocks); ++ dev->chunkBitsAlt = 1; ++ } else ++ dev->chunkBitsAlt = 0; ++ } ++ ++ if (dev->blockInfo && dev->chunkBits) { ++ memset(dev->blockInfo, 0, nBlocks * sizeof(yaffs_BlockInfo)); ++ memset(dev->chunkBits, 0, dev->chunkBitmapStride * nBlocks); ++ return YAFFS_OK; ++ } ++ ++ return YAFFS_FAIL; ++} ++ ++static void yaffs_DeinitialiseBlocks(yaffs_Device *dev) ++{ ++ if (dev->blockInfoAlt && dev->blockInfo) ++ YFREE_ALT(dev->blockInfo); ++ else if (dev->blockInfo) ++ YFREE(dev->blockInfo); ++ ++ dev->blockInfoAlt = 0; ++ ++ dev->blockInfo = NULL; ++ ++ if (dev->chunkBitsAlt && dev->chunkBits) ++ YFREE_ALT(dev->chunkBits); ++ else if (dev->chunkBits) ++ YFREE(dev->chunkBits); ++ dev->chunkBitsAlt = 0; ++ dev->chunkBits = NULL; ++} ++ ++static int yaffs_BlockNotDisqualifiedFromGC(yaffs_Device *dev, ++ yaffs_BlockInfo *bi) ++{ ++ int i; ++ __u32 seq; ++ yaffs_BlockInfo *b; ++ ++ if (!dev->isYaffs2) ++ return 1; /* disqualification only applies to yaffs2. */ ++ ++ if (!bi->hasShrinkHeader) ++ return 1; /* can gc */ ++ ++ /* Find the oldest dirty sequence number if we don't know it and save it ++ * so we don't have to keep recomputing it. ++ */ ++ if (!dev->oldestDirtySequence) { ++ seq = dev->sequenceNumber; ++ ++ for (i = dev->internalStartBlock; i <= dev->internalEndBlock; ++ i++) { ++ b = yaffs_GetBlockInfo(dev, i); ++ if (b->blockState == YAFFS_BLOCK_STATE_FULL && ++ (b->pagesInUse - b->softDeletions) < ++ dev->nChunksPerBlock && b->sequenceNumber < seq) { ++ seq = b->sequenceNumber; ++ } ++ } ++ dev->oldestDirtySequence = seq; ++ } ++ ++ /* Can't do gc of this block if there are any blocks older than this one that have ++ * discarded pages. ++ */ ++ return (bi->sequenceNumber <= dev->oldestDirtySequence); ++} ++ ++/* FindDiretiestBlock is used to select the dirtiest block (or close enough) ++ * for garbage collection. ++ */ ++ ++static int yaffs_FindBlockForGarbageCollection(yaffs_Device *dev, ++ int aggressive) ++{ ++ int b = dev->currentDirtyChecker; ++ ++ int i; ++ int iterations; ++ int dirtiest = -1; ++ int pagesInUse = 0; ++ int prioritised = 0; ++ yaffs_BlockInfo *bi; ++ int pendingPrioritisedExist = 0; ++ ++ /* First let's see if we need to grab a prioritised block */ ++ if (dev->hasPendingPrioritisedGCs) { ++ for (i = dev->internalStartBlock; i < dev->internalEndBlock && !prioritised; i++) { ++ ++ bi = yaffs_GetBlockInfo(dev, i); ++ /* yaffs_VerifyBlock(dev,bi,i); */ ++ ++ if (bi->gcPrioritise) { ++ pendingPrioritisedExist = 1; ++ if (bi->blockState == YAFFS_BLOCK_STATE_FULL && ++ yaffs_BlockNotDisqualifiedFromGC(dev, bi)) { ++ pagesInUse = (bi->pagesInUse - bi->softDeletions); ++ dirtiest = i; ++ prioritised = 1; ++ aggressive = 1; /* Fool the non-aggressive skip logiv below */ ++ } ++ } ++ } ++ ++ if (!pendingPrioritisedExist) /* None found, so we can clear this */ ++ dev->hasPendingPrioritisedGCs = 0; ++ } ++ ++ /* If we're doing aggressive GC then we are happy to take a less-dirty block, and ++ * search harder. ++ * else (we're doing a leasurely gc), then we only bother to do this if the ++ * block has only a few pages in use. ++ */ ++ ++ dev->nonAggressiveSkip--; ++ ++ if (!aggressive && (dev->nonAggressiveSkip > 0)) ++ return -1; ++ ++ if (!prioritised) ++ pagesInUse = ++ (aggressive) ? dev->nChunksPerBlock : YAFFS_PASSIVE_GC_CHUNKS + 1; ++ ++ if (aggressive) ++ iterations = ++ dev->internalEndBlock - dev->internalStartBlock + 1; ++ else { ++ iterations = ++ dev->internalEndBlock - dev->internalStartBlock + 1; ++ iterations = iterations / 16; ++ if (iterations > 200) ++ iterations = 200; ++ } ++ ++ for (i = 0; i <= iterations && pagesInUse > 0 && !prioritised; i++) { ++ b++; ++ if (b < dev->internalStartBlock || b > dev->internalEndBlock) ++ b = dev->internalStartBlock; ++ ++ if (b < dev->internalStartBlock || b > dev->internalEndBlock) { ++ T(YAFFS_TRACE_ERROR, ++ (TSTR("**>> Block %d is not valid" TENDSTR), b)); ++ YBUG(); ++ } ++ ++ bi = yaffs_GetBlockInfo(dev, b); ++ ++ if (bi->blockState == YAFFS_BLOCK_STATE_FULL && ++ (bi->pagesInUse - bi->softDeletions) < pagesInUse && ++ yaffs_BlockNotDisqualifiedFromGC(dev, bi)) { ++ dirtiest = b; ++ pagesInUse = (bi->pagesInUse - bi->softDeletions); ++ } ++ } ++ ++ dev->currentDirtyChecker = b; ++ ++ if (dirtiest > 0) { ++ T(YAFFS_TRACE_GC, ++ (TSTR("GC Selected block %d with %d free, prioritised:%d" TENDSTR), dirtiest, ++ dev->nChunksPerBlock - pagesInUse, prioritised)); ++ } ++ ++ dev->oldestDirtySequence = 0; ++ ++ if (dirtiest > 0) ++ dev->nonAggressiveSkip = 4; ++ ++ return dirtiest; ++} ++ ++static void yaffs_BlockBecameDirty(yaffs_Device *dev, int blockNo) ++{ ++ yaffs_BlockInfo *bi = yaffs_GetBlockInfo(dev, blockNo); ++ ++ int erasedOk = 0; ++ ++ /* If the block is still healthy erase it and mark as clean. ++ * If the block has had a data failure, then retire it. ++ */ ++ ++ T(YAFFS_TRACE_GC | YAFFS_TRACE_ERASE, ++ (TSTR("yaffs_BlockBecameDirty block %d state %d %s"TENDSTR), ++ blockNo, bi->blockState, (bi->needsRetiring) ? "needs retiring" : "")); ++ ++ bi->blockState = YAFFS_BLOCK_STATE_DIRTY; ++ ++ if (!bi->needsRetiring) { ++ yaffs_InvalidateCheckpoint(dev); ++ erasedOk = yaffs_EraseBlockInNAND(dev, blockNo); ++ if (!erasedOk) { ++ dev->nErasureFailures++; ++ T(YAFFS_TRACE_ERROR | YAFFS_TRACE_BAD_BLOCKS, ++ (TSTR("**>> Erasure failed %d" TENDSTR), blockNo)); ++ } ++ } ++ ++ if (erasedOk && ++ ((yaffs_traceMask & YAFFS_TRACE_ERASE) || !yaffs_SkipVerification(dev))) { ++ int i; ++ for (i = 0; i < dev->nChunksPerBlock; i++) { ++ if (!yaffs_CheckChunkErased ++ (dev, blockNo * dev->nChunksPerBlock + i)) { ++ T(YAFFS_TRACE_ERROR, ++ (TSTR ++ (">>Block %d erasure supposedly OK, but chunk %d not erased" ++ TENDSTR), blockNo, i)); ++ } ++ } ++ } ++ ++ if (erasedOk) { ++ /* Clean it up... */ ++ bi->blockState = YAFFS_BLOCK_STATE_EMPTY; ++ dev->nErasedBlocks++; ++ bi->pagesInUse = 0; ++ bi->softDeletions = 0; ++ bi->hasShrinkHeader = 0; ++ bi->skipErasedCheck = 1; /* This is clean, so no need to check */ ++ bi->gcPrioritise = 0; ++ yaffs_ClearChunkBits(dev, blockNo); ++ ++ T(YAFFS_TRACE_ERASE, ++ (TSTR("Erased block %d" TENDSTR), blockNo)); ++ } else { ++ dev->nFreeChunks -= dev->nChunksPerBlock; /* We lost a block of free space */ ++ ++ yaffs_RetireBlock(dev, blockNo); ++ T(YAFFS_TRACE_ERROR | YAFFS_TRACE_BAD_BLOCKS, ++ (TSTR("**>> Block %d retired" TENDSTR), blockNo)); ++ } ++} ++ ++static int yaffs_FindBlockForAllocation(yaffs_Device *dev) ++{ ++ int i; ++ ++ yaffs_BlockInfo *bi; ++ ++ if (dev->nErasedBlocks < 1) { ++ /* Hoosterman we've got a problem. ++ * Can't get space to gc ++ */ ++ T(YAFFS_TRACE_ERROR, ++ (TSTR("yaffs tragedy: no more erased blocks" TENDSTR))); ++ ++ return -1; ++ } ++ ++ /* Find an empty block. */ ++ ++ for (i = dev->internalStartBlock; i <= dev->internalEndBlock; i++) { ++ dev->allocationBlockFinder++; ++ if (dev->allocationBlockFinder < dev->internalStartBlock ++ || dev->allocationBlockFinder > dev->internalEndBlock) { ++ dev->allocationBlockFinder = dev->internalStartBlock; ++ } ++ ++ bi = yaffs_GetBlockInfo(dev, dev->allocationBlockFinder); ++ ++ if (bi->blockState == YAFFS_BLOCK_STATE_EMPTY) { ++ bi->blockState = YAFFS_BLOCK_STATE_ALLOCATING; ++ dev->sequenceNumber++; ++ bi->sequenceNumber = dev->sequenceNumber; ++ dev->nErasedBlocks--; ++ T(YAFFS_TRACE_ALLOCATE, ++ (TSTR("Allocated block %d, seq %d, %d left" TENDSTR), ++ dev->allocationBlockFinder, dev->sequenceNumber, ++ dev->nErasedBlocks)); ++ return dev->allocationBlockFinder; ++ } ++ } ++ ++ T(YAFFS_TRACE_ALWAYS, ++ (TSTR ++ ("yaffs tragedy: no more erased blocks, but there should have been %d" ++ TENDSTR), dev->nErasedBlocks)); ++ ++ return -1; ++} ++ ++ ++ ++static int yaffs_CalcCheckpointBlocksRequired(yaffs_Device *dev) ++{ ++ if (!dev->nCheckpointBlocksRequired && ++ dev->isYaffs2) { ++ /* Not a valid value so recalculate */ ++ int nBytes = 0; ++ int nBlocks; ++ int devBlocks = (dev->endBlock - dev->startBlock + 1); ++ int tnodeSize; ++ ++ tnodeSize = (dev->tnodeWidth * YAFFS_NTNODES_LEVEL0)/8; ++ ++ if (tnodeSize < sizeof(yaffs_Tnode)) ++ tnodeSize = sizeof(yaffs_Tnode); ++ ++ nBytes += sizeof(yaffs_CheckpointValidity); ++ nBytes += sizeof(yaffs_CheckpointDevice); ++ nBytes += devBlocks * sizeof(yaffs_BlockInfo); ++ nBytes += devBlocks * dev->chunkBitmapStride; ++ nBytes += (sizeof(yaffs_CheckpointObject) + sizeof(__u32)) * (dev->nObjectsCreated - dev->nFreeObjects); ++ nBytes += (tnodeSize + sizeof(__u32)) * (dev->nTnodesCreated - dev->nFreeTnodes); ++ nBytes += sizeof(yaffs_CheckpointValidity); ++ nBytes += sizeof(__u32); /* checksum*/ ++ ++ /* Round up and add 2 blocks to allow for some bad blocks, so add 3 */ ++ ++ nBlocks = (nBytes/(dev->nDataBytesPerChunk * dev->nChunksPerBlock)) + 3; ++ ++ dev->nCheckpointBlocksRequired = nBlocks; ++ } ++ ++ return dev->nCheckpointBlocksRequired; ++} ++ ++/* ++ * Check if there's space to allocate... ++ * Thinks.... do we need top make this ths same as yaffs_GetFreeChunks()? ++ */ ++static int yaffs_CheckSpaceForAllocation(yaffs_Device *dev) ++{ ++ int reservedChunks; ++ int reservedBlocks = dev->nReservedBlocks; ++ int checkpointBlocks; ++ ++ if (dev->isYaffs2) { ++ checkpointBlocks = yaffs_CalcCheckpointBlocksRequired(dev) - ++ dev->blocksInCheckpoint; ++ if (checkpointBlocks < 0) ++ checkpointBlocks = 0; ++ } else { ++ checkpointBlocks = 0; ++ } ++ ++ reservedChunks = ((reservedBlocks + checkpointBlocks) * dev->nChunksPerBlock); ++ ++ return (dev->nFreeChunks > reservedChunks); ++} ++ ++static int yaffs_AllocateChunk(yaffs_Device *dev, int useReserve, ++ yaffs_BlockInfo **blockUsedPtr) ++{ ++ int retVal; ++ yaffs_BlockInfo *bi; ++ ++ if (dev->allocationBlock < 0) { ++ /* Get next block to allocate off */ ++ dev->allocationBlock = yaffs_FindBlockForAllocation(dev); ++ dev->allocationPage = 0; ++ } ++ ++ if (!useReserve && !yaffs_CheckSpaceForAllocation(dev)) { ++ /* Not enough space to allocate unless we're allowed to use the reserve. */ ++ return -1; ++ } ++ ++ if (dev->nErasedBlocks < dev->nReservedBlocks ++ && dev->allocationPage == 0) { ++ T(YAFFS_TRACE_ALLOCATE, (TSTR("Allocating reserve" TENDSTR))); ++ } ++ ++ /* Next page please.... */ ++ if (dev->allocationBlock >= 0) { ++ bi = yaffs_GetBlockInfo(dev, dev->allocationBlock); ++ ++ retVal = (dev->allocationBlock * dev->nChunksPerBlock) + ++ dev->allocationPage; ++ bi->pagesInUse++; ++ yaffs_SetChunkBit(dev, dev->allocationBlock, ++ dev->allocationPage); ++ ++ dev->allocationPage++; ++ ++ dev->nFreeChunks--; ++ ++ /* If the block is full set the state to full */ ++ if (dev->allocationPage >= dev->nChunksPerBlock) { ++ bi->blockState = YAFFS_BLOCK_STATE_FULL; ++ dev->allocationBlock = -1; ++ } ++ ++ if (blockUsedPtr) ++ *blockUsedPtr = bi; ++ ++ return retVal; ++ } ++ ++ T(YAFFS_TRACE_ERROR, ++ (TSTR("!!!!!!!!! Allocator out !!!!!!!!!!!!!!!!!" TENDSTR))); ++ ++ return -1; ++} ++ ++static int yaffs_GetErasedChunks(yaffs_Device *dev) ++{ ++ int n; ++ ++ n = dev->nErasedBlocks * dev->nChunksPerBlock; ++ ++ if (dev->allocationBlock > 0) ++ n += (dev->nChunksPerBlock - dev->allocationPage); ++ ++ return n; ++ ++} ++ ++static int yaffs_GarbageCollectBlock(yaffs_Device *dev, int block, ++ int wholeBlock) ++{ ++ int oldChunk; ++ int newChunk; ++ int markNAND; ++ int retVal = YAFFS_OK; ++ int cleanups = 0; ++ int i; ++ int isCheckpointBlock; ++ int matchingChunk; ++ int maxCopies; ++ ++ int chunksBefore = yaffs_GetErasedChunks(dev); ++ int chunksAfter; ++ ++ yaffs_ExtendedTags tags; ++ ++ yaffs_BlockInfo *bi = yaffs_GetBlockInfo(dev, block); ++ ++ yaffs_Object *object; ++ ++ isCheckpointBlock = (bi->blockState == YAFFS_BLOCK_STATE_CHECKPOINT); ++ ++ bi->blockState = YAFFS_BLOCK_STATE_COLLECTING; ++ ++ T(YAFFS_TRACE_TRACING, ++ (TSTR("Collecting block %d, in use %d, shrink %d, wholeBlock %d" TENDSTR), ++ block, ++ bi->pagesInUse, ++ bi->hasShrinkHeader, ++ wholeBlock)); ++ ++ /*yaffs_VerifyFreeChunks(dev); */ ++ ++ bi->hasShrinkHeader = 0; /* clear the flag so that the block can erase */ ++ ++ /* Take off the number of soft deleted entries because ++ * they're going to get really deleted during GC. ++ */ ++ dev->nFreeChunks -= bi->softDeletions; ++ ++ dev->isDoingGC = 1; ++ ++ if (isCheckpointBlock || ++ !yaffs_StillSomeChunkBits(dev, block)) { ++ T(YAFFS_TRACE_TRACING, ++ (TSTR ++ ("Collecting block %d that has no chunks in use" TENDSTR), ++ block)); ++ yaffs_BlockBecameDirty(dev, block); ++ } else { ++ ++ __u8 *buffer = yaffs_GetTempBuffer(dev, __LINE__); ++ ++ yaffs_VerifyBlock(dev, bi, block); ++ ++ maxCopies = (wholeBlock) ? dev->nChunksPerBlock : 10; ++ oldChunk = block * dev->nChunksPerBlock + dev->gcChunk; ++ ++ for (/* init already done */; ++ retVal == YAFFS_OK && ++ dev->gcChunk < dev->nChunksPerBlock && ++ (bi->blockState == YAFFS_BLOCK_STATE_COLLECTING) && ++ maxCopies > 0; ++ dev->gcChunk++, oldChunk++) { ++ if (yaffs_CheckChunkBit(dev, block, dev->gcChunk)) { ++ ++ /* This page is in use and might need to be copied off */ ++ ++ maxCopies--; ++ ++ markNAND = 1; ++ ++ yaffs_InitialiseTags(&tags); ++ ++ yaffs_ReadChunkWithTagsFromNAND(dev, oldChunk, ++ buffer, &tags); ++ ++ object = ++ yaffs_FindObjectByNumber(dev, ++ tags.objectId); ++ ++ T(YAFFS_TRACE_GC_DETAIL, ++ (TSTR ++ ("Collecting chunk in block %d, %d %d %d " TENDSTR), ++ dev->gcChunk, tags.objectId, tags.chunkId, ++ tags.byteCount)); ++ ++ if (object && !yaffs_SkipVerification(dev)) { ++ if (tags.chunkId == 0) ++ matchingChunk = object->hdrChunk; ++ else if (object->softDeleted) ++ matchingChunk = oldChunk; /* Defeat the test */ ++ else ++ matchingChunk = yaffs_FindChunkInFile(object, tags.chunkId, NULL); ++ ++ if (oldChunk != matchingChunk) ++ T(YAFFS_TRACE_ERROR, ++ (TSTR("gc: page in gc mismatch: %d %d %d %d"TENDSTR), ++ oldChunk, matchingChunk, tags.objectId, tags.chunkId)); ++ ++ } ++ ++ if (!object) { ++ T(YAFFS_TRACE_ERROR, ++ (TSTR ++ ("page %d in gc has no object: %d %d %d " ++ TENDSTR), oldChunk, ++ tags.objectId, tags.chunkId, tags.byteCount)); ++ } ++ ++ if (object && ++ object->deleted && ++ object->softDeleted && ++ tags.chunkId != 0) { ++ /* Data chunk in a soft deleted file, throw it away ++ * It's a soft deleted data chunk, ++ * No need to copy this, just forget about it and ++ * fix up the object. ++ */ ++ ++ object->nDataChunks--; ++ ++ if (object->nDataChunks <= 0) { ++ /* remeber to clean up the object */ ++ dev->gcCleanupList[cleanups] = ++ tags.objectId; ++ cleanups++; ++ } ++ markNAND = 0; ++ } else if (0) { ++ /* Todo object && object->deleted && object->nDataChunks == 0 */ ++ /* Deleted object header with no data chunks. ++ * Can be discarded and the file deleted. ++ */ ++ object->hdrChunk = 0; ++ yaffs_FreeTnode(object->myDev, ++ object->variant. ++ fileVariant.top); ++ object->variant.fileVariant.top = NULL; ++ yaffs_DoGenericObjectDeletion(object); ++ ++ } else if (object) { ++ /* It's either a data chunk in a live file or ++ * an ObjectHeader, so we're interested in it. ++ * NB Need to keep the ObjectHeaders of deleted files ++ * until the whole file has been deleted off ++ */ ++ tags.serialNumber++; ++ ++ dev->nGCCopies++; ++ ++ if (tags.chunkId == 0) { ++ /* It is an object Id, ++ * We need to nuke the shrinkheader flags first ++ * We no longer want the shrinkHeader flag since its work is done ++ * and if it is left in place it will mess up scanning. ++ */ ++ ++ yaffs_ObjectHeader *oh; ++ oh = (yaffs_ObjectHeader *)buffer; ++ oh->isShrink = 0; ++ tags.extraIsShrinkHeader = 0; ++ ++ yaffs_VerifyObjectHeader(object, oh, &tags, 1); ++ } ++ ++ newChunk = ++ yaffs_WriteNewChunkWithTagsToNAND(dev, buffer, &tags, 1); ++ ++ if (newChunk < 0) { ++ retVal = YAFFS_FAIL; ++ } else { ++ ++ /* Ok, now fix up the Tnodes etc. */ ++ ++ if (tags.chunkId == 0) { ++ /* It's a header */ ++ object->hdrChunk = newChunk; ++ object->serial = tags.serialNumber; ++ } else { ++ /* It's a data chunk */ ++ yaffs_PutChunkIntoFile ++ (object, ++ tags.chunkId, ++ newChunk, 0); ++ } ++ } ++ } ++ ++ if (retVal == YAFFS_OK) ++ yaffs_DeleteChunk(dev, oldChunk, markNAND, __LINE__); ++ ++ } ++ } ++ ++ yaffs_ReleaseTempBuffer(dev, buffer, __LINE__); ++ ++ ++ /* Do any required cleanups */ ++ for (i = 0; i < cleanups; i++) { ++ /* Time to delete the file too */ ++ object = ++ yaffs_FindObjectByNumber(dev, ++ dev->gcCleanupList[i]); ++ if (object) { ++ yaffs_FreeTnode(dev, ++ object->variant.fileVariant. ++ top); ++ object->variant.fileVariant.top = NULL; ++ T(YAFFS_TRACE_GC, ++ (TSTR ++ ("yaffs: About to finally delete object %d" ++ TENDSTR), object->objectId)); ++ yaffs_DoGenericObjectDeletion(object); ++ object->myDev->nDeletedFiles--; ++ } ++ ++ } ++ ++ } ++ ++ yaffs_VerifyCollectedBlock(dev, bi, block); ++ ++ chunksAfter = yaffs_GetErasedChunks(dev); ++ if (chunksBefore >= chunksAfter) { ++ T(YAFFS_TRACE_GC, ++ (TSTR ++ ("gc did not increase free chunks before %d after %d" ++ TENDSTR), chunksBefore, chunksAfter)); ++ } ++ ++ /* If the gc completed then clear the current gcBlock so that we find another. */ ++ if (bi->blockState != YAFFS_BLOCK_STATE_COLLECTING) { ++ dev->gcBlock = -1; ++ dev->gcChunk = 0; ++ } ++ ++ dev->isDoingGC = 0; ++ ++ return retVal; ++} ++ ++/* New garbage collector ++ * If we're very low on erased blocks then we do aggressive garbage collection ++ * otherwise we do "leasurely" garbage collection. ++ * Aggressive gc looks further (whole array) and will accept less dirty blocks. ++ * Passive gc only inspects smaller areas and will only accept more dirty blocks. ++ * ++ * The idea is to help clear out space in a more spread-out manner. ++ * Dunno if it really does anything useful. ++ */ ++static int yaffs_CheckGarbageCollection(yaffs_Device *dev) ++{ ++ int block; ++ int aggressive; ++ int gcOk = YAFFS_OK; ++ int maxTries = 0; ++ ++ int checkpointBlockAdjust; ++ ++ if (dev->isDoingGC) { ++ /* Bail out so we don't get recursive gc */ ++ return YAFFS_OK; ++ } ++ ++ /* This loop should pass the first time. ++ * We'll only see looping here if the erase of the collected block fails. ++ */ ++ ++ do { ++ maxTries++; ++ ++ checkpointBlockAdjust = yaffs_CalcCheckpointBlocksRequired(dev) - dev->blocksInCheckpoint; ++ if (checkpointBlockAdjust < 0) ++ checkpointBlockAdjust = 0; ++ ++ if (dev->nErasedBlocks < (dev->nReservedBlocks + checkpointBlockAdjust + 2)) { ++ /* We need a block soon...*/ ++ aggressive = 1; ++ } else { ++ /* We're in no hurry */ ++ aggressive = 0; ++ } ++ ++ if (dev->gcBlock <= 0) { ++ dev->gcBlock = yaffs_FindBlockForGarbageCollection(dev, aggressive); ++ dev->gcChunk = 0; ++ } ++ ++ block = dev->gcBlock; ++ ++ if (block > 0) { ++ dev->garbageCollections++; ++ if (!aggressive) ++ dev->passiveGarbageCollections++; ++ ++ T(YAFFS_TRACE_GC, ++ (TSTR ++ ("yaffs: GC erasedBlocks %d aggressive %d" TENDSTR), ++ dev->nErasedBlocks, aggressive)); ++ ++ gcOk = yaffs_GarbageCollectBlock(dev, block, aggressive); ++ } ++ ++ if (dev->nErasedBlocks < (dev->nReservedBlocks) && block > 0) { ++ T(YAFFS_TRACE_GC, ++ (TSTR ++ ("yaffs: GC !!!no reclaim!!! erasedBlocks %d after try %d block %d" ++ TENDSTR), dev->nErasedBlocks, maxTries, block)); ++ } ++ } while ((dev->nErasedBlocks < dev->nReservedBlocks) && ++ (block > 0) && ++ (maxTries < 2)); ++ ++ return aggressive ? gcOk : YAFFS_OK; ++} ++ ++/*------------------------- TAGS --------------------------------*/ ++ ++static int yaffs_TagsMatch(const yaffs_ExtendedTags *tags, int objectId, ++ int chunkInObject) ++{ ++ return (tags->chunkId == chunkInObject && ++ tags->objectId == objectId && !tags->chunkDeleted) ? 1 : 0; ++ ++} ++ ++ ++/*-------------------- Data file manipulation -----------------*/ ++ ++static int yaffs_FindChunkInFile(yaffs_Object *in, int chunkInInode, ++ yaffs_ExtendedTags *tags) ++{ ++ /*Get the Tnode, then get the level 0 offset chunk offset */ ++ yaffs_Tnode *tn; ++ int theChunk = -1; ++ yaffs_ExtendedTags localTags; ++ int retVal = -1; ++ ++ yaffs_Device *dev = in->myDev; ++ ++ if (!tags) { ++ /* Passed a NULL, so use our own tags space */ ++ tags = &localTags; ++ } ++ ++ tn = yaffs_FindLevel0Tnode(dev, &in->variant.fileVariant, chunkInInode); ++ ++ if (tn) { ++ theChunk = yaffs_GetChunkGroupBase(dev, tn, chunkInInode); ++ ++ retVal = ++ yaffs_FindChunkInGroup(dev, theChunk, tags, in->objectId, ++ chunkInInode); ++ } ++ return retVal; ++} ++ ++static int yaffs_FindAndDeleteChunkInFile(yaffs_Object *in, int chunkInInode, ++ yaffs_ExtendedTags *tags) ++{ ++ /* Get the Tnode, then get the level 0 offset chunk offset */ ++ yaffs_Tnode *tn; ++ int theChunk = -1; ++ yaffs_ExtendedTags localTags; ++ ++ yaffs_Device *dev = in->myDev; ++ int retVal = -1; ++ ++ if (!tags) { ++ /* Passed a NULL, so use our own tags space */ ++ tags = &localTags; ++ } ++ ++ tn = yaffs_FindLevel0Tnode(dev, &in->variant.fileVariant, chunkInInode); ++ ++ if (tn) { ++ ++ theChunk = yaffs_GetChunkGroupBase(dev, tn, chunkInInode); ++ ++ retVal = ++ yaffs_FindChunkInGroup(dev, theChunk, tags, in->objectId, ++ chunkInInode); ++ ++ /* Delete the entry in the filestructure (if found) */ ++ if (retVal != -1) ++ yaffs_PutLevel0Tnode(dev, tn, chunkInInode, 0); ++ } ++ ++ return retVal; ++} ++ ++#ifdef YAFFS_PARANOID ++ ++static int yaffs_CheckFileSanity(yaffs_Object *in) ++{ ++ int chunk; ++ int nChunks; ++ int fSize; ++ int failed = 0; ++ int objId; ++ yaffs_Tnode *tn; ++ yaffs_Tags localTags; ++ yaffs_Tags *tags = &localTags; ++ int theChunk; ++ int chunkDeleted; ++ ++ if (in->variantType != YAFFS_OBJECT_TYPE_FILE) ++ return YAFFS_FAIL; ++ ++ objId = in->objectId; ++ fSize = in->variant.fileVariant.fileSize; ++ nChunks = ++ (fSize + in->myDev->nDataBytesPerChunk - 1) / in->myDev->nDataBytesPerChunk; ++ ++ for (chunk = 1; chunk <= nChunks; chunk++) { ++ tn = yaffs_FindLevel0Tnode(in->myDev, &in->variant.fileVariant, ++ chunk); ++ ++ if (tn) { ++ ++ theChunk = yaffs_GetChunkGroupBase(dev, tn, chunk); ++ ++ if (yaffs_CheckChunkBits ++ (dev, theChunk / dev->nChunksPerBlock, ++ theChunk % dev->nChunksPerBlock)) { ++ ++ yaffs_ReadChunkTagsFromNAND(in->myDev, theChunk, ++ tags, ++ &chunkDeleted); ++ if (yaffs_TagsMatch ++ (tags, in->objectId, chunk, chunkDeleted)) { ++ /* found it; */ ++ ++ } ++ } else { ++ ++ failed = 1; ++ } ++ ++ } else { ++ /* T(("No level 0 found for %d\n", chunk)); */ ++ } ++ } ++ ++ return failed ? YAFFS_FAIL : YAFFS_OK; ++} ++ ++#endif ++ ++static int yaffs_PutChunkIntoFile(yaffs_Object *in, int chunkInInode, ++ int chunkInNAND, int inScan) ++{ ++ /* NB inScan is zero unless scanning. ++ * For forward scanning, inScan is > 0; ++ * for backward scanning inScan is < 0 ++ */ ++ ++ yaffs_Tnode *tn; ++ yaffs_Device *dev = in->myDev; ++ int existingChunk; ++ yaffs_ExtendedTags existingTags; ++ yaffs_ExtendedTags newTags; ++ unsigned existingSerial, newSerial; ++ ++ if (in->variantType != YAFFS_OBJECT_TYPE_FILE) { ++ /* Just ignore an attempt at putting a chunk into a non-file during scanning ++ * If it is not during Scanning then something went wrong! ++ */ ++ if (!inScan) { ++ T(YAFFS_TRACE_ERROR, ++ (TSTR ++ ("yaffs tragedy:attempt to put data chunk into a non-file" ++ TENDSTR))); ++ YBUG(); ++ } ++ ++ yaffs_DeleteChunk(dev, chunkInNAND, 1, __LINE__); ++ return YAFFS_OK; ++ } ++ ++ tn = yaffs_AddOrFindLevel0Tnode(dev, ++ &in->variant.fileVariant, ++ chunkInInode, ++ NULL); ++ if (!tn) ++ return YAFFS_FAIL; ++ ++ existingChunk = yaffs_GetChunkGroupBase(dev, tn, chunkInInode); ++ ++ if (inScan != 0) { ++ /* If we're scanning then we need to test for duplicates ++ * NB This does not need to be efficient since it should only ever ++ * happen when the power fails during a write, then only one ++ * chunk should ever be affected. ++ * ++ * Correction for YAFFS2: This could happen quite a lot and we need to think about efficiency! TODO ++ * Update: For backward scanning we don't need to re-read tags so this is quite cheap. ++ */ ++ ++ if (existingChunk > 0) { ++ /* NB Right now existing chunk will not be real chunkId if the device >= 32MB ++ * thus we have to do a FindChunkInFile to get the real chunk id. ++ * ++ * We have a duplicate now we need to decide which one to use: ++ * ++ * Backwards scanning YAFFS2: The old one is what we use, dump the new one. ++ * Forward scanning YAFFS2: The new one is what we use, dump the old one. ++ * YAFFS1: Get both sets of tags and compare serial numbers. ++ */ ++ ++ if (inScan > 0) { ++ /* Only do this for forward scanning */ ++ yaffs_ReadChunkWithTagsFromNAND(dev, ++ chunkInNAND, ++ NULL, &newTags); ++ ++ /* Do a proper find */ ++ existingChunk = ++ yaffs_FindChunkInFile(in, chunkInInode, ++ &existingTags); ++ } ++ ++ if (existingChunk <= 0) { ++ /*Hoosterman - how did this happen? */ ++ ++ T(YAFFS_TRACE_ERROR, ++ (TSTR ++ ("yaffs tragedy: existing chunk < 0 in scan" ++ TENDSTR))); ++ ++ } ++ ++ /* NB The deleted flags should be false, otherwise the chunks will ++ * not be loaded during a scan ++ */ ++ ++ if (inScan > 0) { ++ newSerial = newTags.serialNumber; ++ existingSerial = existingTags.serialNumber; ++ } ++ ++ if ((inScan > 0) && ++ (in->myDev->isYaffs2 || ++ existingChunk <= 0 || ++ ((existingSerial + 1) & 3) == newSerial)) { ++ /* Forward scanning. ++ * Use new ++ * Delete the old one and drop through to update the tnode ++ */ ++ yaffs_DeleteChunk(dev, existingChunk, 1, ++ __LINE__); ++ } else { ++ /* Backward scanning or we want to use the existing one ++ * Use existing. ++ * Delete the new one and return early so that the tnode isn't changed ++ */ ++ yaffs_DeleteChunk(dev, chunkInNAND, 1, ++ __LINE__); ++ return YAFFS_OK; ++ } ++ } ++ ++ } ++ ++ if (existingChunk == 0) ++ in->nDataChunks++; ++ ++ yaffs_PutLevel0Tnode(dev, tn, chunkInInode, chunkInNAND); ++ ++ return YAFFS_OK; ++} ++ ++static int yaffs_ReadChunkDataFromObject(yaffs_Object *in, int chunkInInode, ++ __u8 *buffer) ++{ ++ int chunkInNAND = yaffs_FindChunkInFile(in, chunkInInode, NULL); ++ ++ if (chunkInNAND >= 0) ++ return yaffs_ReadChunkWithTagsFromNAND(in->myDev, chunkInNAND, ++ buffer, NULL); ++ else { ++ T(YAFFS_TRACE_NANDACCESS, ++ (TSTR("Chunk %d not found zero instead" TENDSTR), ++ chunkInNAND)); ++ /* get sane (zero) data if you read a hole */ ++ memset(buffer, 0, in->myDev->nDataBytesPerChunk); ++ return 0; ++ } ++ ++} ++ ++void yaffs_DeleteChunk(yaffs_Device *dev, int chunkId, int markNAND, int lyn) ++{ ++ int block; ++ int page; ++ yaffs_ExtendedTags tags; ++ yaffs_BlockInfo *bi; ++ ++ if (chunkId <= 0) ++ return; ++ ++ dev->nDeletions++; ++ block = chunkId / dev->nChunksPerBlock; ++ page = chunkId % dev->nChunksPerBlock; ++ ++ ++ if (!yaffs_CheckChunkBit(dev, block, page)) ++ T(YAFFS_TRACE_VERIFY, ++ (TSTR("Deleting invalid chunk %d"TENDSTR), ++ chunkId)); ++ ++ bi = yaffs_GetBlockInfo(dev, block); ++ ++ T(YAFFS_TRACE_DELETION, ++ (TSTR("line %d delete of chunk %d" TENDSTR), lyn, chunkId)); ++ ++ if (markNAND && ++ bi->blockState != YAFFS_BLOCK_STATE_COLLECTING && !dev->isYaffs2) { ++ ++ yaffs_InitialiseTags(&tags); ++ ++ tags.chunkDeleted = 1; ++ ++ yaffs_WriteChunkWithTagsToNAND(dev, chunkId, NULL, &tags); ++ yaffs_HandleUpdateChunk(dev, chunkId, &tags); ++ } else { ++ dev->nUnmarkedDeletions++; ++ } ++ ++ /* Pull out of the management area. ++ * If the whole block became dirty, this will kick off an erasure. ++ */ ++ if (bi->blockState == YAFFS_BLOCK_STATE_ALLOCATING || ++ bi->blockState == YAFFS_BLOCK_STATE_FULL || ++ bi->blockState == YAFFS_BLOCK_STATE_NEEDS_SCANNING || ++ bi->blockState == YAFFS_BLOCK_STATE_COLLECTING) { ++ dev->nFreeChunks++; ++ ++ yaffs_ClearChunkBit(dev, block, page); ++ ++ bi->pagesInUse--; ++ ++ if (bi->pagesInUse == 0 && ++ !bi->hasShrinkHeader && ++ bi->blockState != YAFFS_BLOCK_STATE_ALLOCATING && ++ bi->blockState != YAFFS_BLOCK_STATE_NEEDS_SCANNING) { ++ yaffs_BlockBecameDirty(dev, block); ++ } ++ ++ } ++ ++} ++ ++static int yaffs_WriteChunkDataToObject(yaffs_Object *in, int chunkInInode, ++ const __u8 *buffer, int nBytes, ++ int useReserve) ++{ ++ /* Find old chunk Need to do this to get serial number ++ * Write new one and patch into tree. ++ * Invalidate old tags. ++ */ ++ ++ int prevChunkId; ++ yaffs_ExtendedTags prevTags; ++ ++ int newChunkId; ++ yaffs_ExtendedTags newTags; ++ ++ yaffs_Device *dev = in->myDev; ++ ++ yaffs_CheckGarbageCollection(dev); ++ ++ /* Get the previous chunk at this location in the file if it exists */ ++ prevChunkId = yaffs_FindChunkInFile(in, chunkInInode, &prevTags); ++ ++ /* Set up new tags */ ++ yaffs_InitialiseTags(&newTags); ++ ++ newTags.chunkId = chunkInInode; ++ newTags.objectId = in->objectId; ++ newTags.serialNumber = ++ (prevChunkId >= 0) ? prevTags.serialNumber + 1 : 1; ++ newTags.byteCount = nBytes; ++ ++ if (nBytes < 1 || nBytes > dev->totalBytesPerChunk) { ++ T(YAFFS_TRACE_ERROR, ++ (TSTR("Writing %d bytes to chunk!!!!!!!!!" TENDSTR), nBytes)); ++ YBUG(); ++ } ++ ++ newChunkId = ++ yaffs_WriteNewChunkWithTagsToNAND(dev, buffer, &newTags, ++ useReserve); ++ ++ if (newChunkId >= 0) { ++ yaffs_PutChunkIntoFile(in, chunkInInode, newChunkId, 0); ++ ++ if (prevChunkId >= 0) ++ yaffs_DeleteChunk(dev, prevChunkId, 1, __LINE__); ++ ++ yaffs_CheckFileSanity(in); ++ } ++ return newChunkId; ++ ++} ++ ++/* UpdateObjectHeader updates the header on NAND for an object. ++ * If name is not NULL, then that new name is used. ++ */ ++int yaffs_UpdateObjectHeader(yaffs_Object *in, const YCHAR *name, int force, ++ int isShrink, int shadows) ++{ ++ ++ yaffs_BlockInfo *bi; ++ ++ yaffs_Device *dev = in->myDev; ++ ++ int prevChunkId; ++ int retVal = 0; ++ int result = 0; ++ ++ int newChunkId; ++ yaffs_ExtendedTags newTags; ++ yaffs_ExtendedTags oldTags; ++ ++ __u8 *buffer = NULL; ++ YCHAR oldName[YAFFS_MAX_NAME_LENGTH + 1]; ++ ++ yaffs_ObjectHeader *oh = NULL; ++ ++ yaffs_strcpy(oldName, _Y("silly old name")); ++ ++ ++ if (!in->fake || ++ in == dev->rootDir || /* The rootDir should also be saved */ ++ force) { ++ ++ yaffs_CheckGarbageCollection(dev); ++ yaffs_CheckObjectDetailsLoaded(in); ++ ++ buffer = yaffs_GetTempBuffer(in->myDev, __LINE__); ++ oh = (yaffs_ObjectHeader *) buffer; ++ ++ prevChunkId = in->hdrChunk; ++ ++ if (prevChunkId > 0) { ++ result = yaffs_ReadChunkWithTagsFromNAND(dev, prevChunkId, ++ buffer, &oldTags); ++ ++ yaffs_VerifyObjectHeader(in, oh, &oldTags, 0); ++ ++ memcpy(oldName, oh->name, sizeof(oh->name)); ++ } ++ ++ memset(buffer, 0xFF, dev->nDataBytesPerChunk); ++ ++ oh->type = in->variantType; ++ oh->yst_mode = in->yst_mode; ++ oh->shadowsObject = oh->inbandShadowsObject = shadows; ++ ++#ifdef CONFIG_YAFFS_WINCE ++ oh->win_atime[0] = in->win_atime[0]; ++ oh->win_ctime[0] = in->win_ctime[0]; ++ oh->win_mtime[0] = in->win_mtime[0]; ++ oh->win_atime[1] = in->win_atime[1]; ++ oh->win_ctime[1] = in->win_ctime[1]; ++ oh->win_mtime[1] = in->win_mtime[1]; ++#else ++ oh->yst_uid = in->yst_uid; ++ oh->yst_gid = in->yst_gid; ++ oh->yst_atime = in->yst_atime; ++ oh->yst_mtime = in->yst_mtime; ++ oh->yst_ctime = in->yst_ctime; ++ oh->yst_rdev = in->yst_rdev; ++#endif ++ if (in->parent) ++ oh->parentObjectId = in->parent->objectId; ++ else ++ oh->parentObjectId = 0; ++ ++ if (name && *name) { ++ memset(oh->name, 0, sizeof(oh->name)); ++ yaffs_strncpy(oh->name, name, YAFFS_MAX_NAME_LENGTH); ++ } else if (prevChunkId >= 0) ++ memcpy(oh->name, oldName, sizeof(oh->name)); ++ else ++ memset(oh->name, 0, sizeof(oh->name)); ++ ++ oh->isShrink = isShrink; ++ ++ switch (in->variantType) { ++ case YAFFS_OBJECT_TYPE_UNKNOWN: ++ /* Should not happen */ ++ break; ++ case YAFFS_OBJECT_TYPE_FILE: ++ oh->fileSize = ++ (oh->parentObjectId == YAFFS_OBJECTID_DELETED ++ || oh->parentObjectId == ++ YAFFS_OBJECTID_UNLINKED) ? 0 : in->variant. ++ fileVariant.fileSize; ++ break; ++ case YAFFS_OBJECT_TYPE_HARDLINK: ++ oh->equivalentObjectId = ++ in->variant.hardLinkVariant.equivalentObjectId; ++ break; ++ case YAFFS_OBJECT_TYPE_SPECIAL: ++ /* Do nothing */ ++ break; ++ case YAFFS_OBJECT_TYPE_DIRECTORY: ++ /* Do nothing */ ++ break; ++ case YAFFS_OBJECT_TYPE_SYMLINK: ++ yaffs_strncpy(oh->alias, ++ in->variant.symLinkVariant.alias, ++ YAFFS_MAX_ALIAS_LENGTH); ++ oh->alias[YAFFS_MAX_ALIAS_LENGTH] = 0; ++ break; ++ } ++ ++ /* Tags */ ++ yaffs_InitialiseTags(&newTags); ++ in->serial++; ++ newTags.chunkId = 0; ++ newTags.objectId = in->objectId; ++ newTags.serialNumber = in->serial; ++ ++ /* Add extra info for file header */ ++ ++ newTags.extraHeaderInfoAvailable = 1; ++ newTags.extraParentObjectId = oh->parentObjectId; ++ newTags.extraFileLength = oh->fileSize; ++ newTags.extraIsShrinkHeader = oh->isShrink; ++ newTags.extraEquivalentObjectId = oh->equivalentObjectId; ++ newTags.extraShadows = (oh->shadowsObject > 0) ? 1 : 0; ++ newTags.extraObjectType = in->variantType; ++ ++ yaffs_VerifyObjectHeader(in, oh, &newTags, 1); ++ ++ /* Create new chunk in NAND */ ++ newChunkId = ++ yaffs_WriteNewChunkWithTagsToNAND(dev, buffer, &newTags, ++ (prevChunkId >= 0) ? 1 : 0); ++ ++ if (newChunkId >= 0) { ++ ++ in->hdrChunk = newChunkId; ++ ++ if (prevChunkId >= 0) { ++ yaffs_DeleteChunk(dev, prevChunkId, 1, ++ __LINE__); ++ } ++ ++ if (!yaffs_ObjectHasCachedWriteData(in)) ++ in->dirty = 0; ++ ++ /* If this was a shrink, then mark the block that the chunk lives on */ ++ if (isShrink) { ++ bi = yaffs_GetBlockInfo(in->myDev, ++ newChunkId / in->myDev->nChunksPerBlock); ++ bi->hasShrinkHeader = 1; ++ } ++ ++ } ++ ++ retVal = newChunkId; ++ ++ } ++ ++ if (buffer) ++ yaffs_ReleaseTempBuffer(dev, buffer, __LINE__); ++ ++ return retVal; ++} ++ ++/*------------------------ Short Operations Cache ---------------------------------------- ++ * In many situations where there is no high level buffering (eg WinCE) a lot of ++ * reads might be short sequential reads, and a lot of writes may be short ++ * sequential writes. eg. scanning/writing a jpeg file. ++ * In these cases, a short read/write cache can provide a huge perfomance benefit ++ * with dumb-as-a-rock code. ++ * In Linux, the page cache provides read buffering aand the short op cache provides write ++ * buffering. ++ * ++ * There are a limited number (~10) of cache chunks per device so that we don't ++ * need a very intelligent search. ++ */ ++ ++static int yaffs_ObjectHasCachedWriteData(yaffs_Object *obj) ++{ ++ yaffs_Device *dev = obj->myDev; ++ int i; ++ yaffs_ChunkCache *cache; ++ int nCaches = obj->myDev->nShortOpCaches; ++ ++ for (i = 0; i < nCaches; i++) { ++ cache = &dev->srCache[i]; ++ if (cache->object == obj && ++ cache->dirty) ++ return 1; ++ } ++ ++ return 0; ++} ++ ++ ++static void yaffs_FlushFilesChunkCache(yaffs_Object *obj) ++{ ++ yaffs_Device *dev = obj->myDev; ++ int lowest = -99; /* Stop compiler whining. */ ++ int i; ++ yaffs_ChunkCache *cache; ++ int chunkWritten = 0; ++ int nCaches = obj->myDev->nShortOpCaches; ++ ++ if (nCaches > 0) { ++ do { ++ cache = NULL; ++ ++ /* Find the dirty cache for this object with the lowest chunk id. */ ++ for (i = 0; i < nCaches; i++) { ++ if (dev->srCache[i].object == obj && ++ dev->srCache[i].dirty) { ++ if (!cache ++ || dev->srCache[i].chunkId < ++ lowest) { ++ cache = &dev->srCache[i]; ++ lowest = cache->chunkId; ++ } ++ } ++ } ++ ++ if (cache && !cache->locked) { ++ /* Write it out and free it up */ ++ ++ chunkWritten = ++ yaffs_WriteChunkDataToObject(cache->object, ++ cache->chunkId, ++ cache->data, ++ cache->nBytes, ++ 1); ++ cache->dirty = 0; ++ cache->object = NULL; ++ } ++ ++ } while (cache && chunkWritten > 0); ++ ++ if (cache) { ++ /* Hoosterman, disk full while writing cache out. */ ++ T(YAFFS_TRACE_ERROR, ++ (TSTR("yaffs tragedy: no space during cache write" TENDSTR))); ++ ++ } ++ } ++ ++} ++ ++/*yaffs_FlushEntireDeviceCache(dev) ++ * ++ * ++ */ ++ ++void yaffs_FlushEntireDeviceCache(yaffs_Device *dev) ++{ ++ yaffs_Object *obj; ++ int nCaches = dev->nShortOpCaches; ++ int i; ++ ++ /* Find a dirty object in the cache and flush it... ++ * until there are no further dirty objects. ++ */ ++ do { ++ obj = NULL; ++ for (i = 0; i < nCaches && !obj; i++) { ++ if (dev->srCache[i].object && ++ dev->srCache[i].dirty) ++ obj = dev->srCache[i].object; ++ ++ } ++ if (obj) ++ yaffs_FlushFilesChunkCache(obj); ++ ++ } while (obj); ++ ++} ++ ++ ++/* Grab us a cache chunk for use. ++ * First look for an empty one. ++ * Then look for the least recently used non-dirty one. ++ * Then look for the least recently used dirty one...., flush and look again. ++ */ ++static yaffs_ChunkCache *yaffs_GrabChunkCacheWorker(yaffs_Device *dev) ++{ ++ int i; ++ ++ if (dev->nShortOpCaches > 0) { ++ for (i = 0; i < dev->nShortOpCaches; i++) { ++ if (!dev->srCache[i].object) ++ return &dev->srCache[i]; ++ } ++ } ++ ++ return NULL; ++} ++ ++static yaffs_ChunkCache *yaffs_GrabChunkCache(yaffs_Device *dev) ++{ ++ yaffs_ChunkCache *cache; ++ yaffs_Object *theObj; ++ int usage; ++ int i; ++ int pushout; ++ ++ if (dev->nShortOpCaches > 0) { ++ /* Try find a non-dirty one... */ ++ ++ cache = yaffs_GrabChunkCacheWorker(dev); ++ ++ if (!cache) { ++ /* They were all dirty, find the last recently used object and flush ++ * its cache, then find again. ++ * NB what's here is not very accurate, we actually flush the object ++ * the last recently used page. ++ */ ++ ++ /* With locking we can't assume we can use entry zero */ ++ ++ theObj = NULL; ++ usage = -1; ++ cache = NULL; ++ pushout = -1; ++ ++ for (i = 0; i < dev->nShortOpCaches; i++) { ++ if (dev->srCache[i].object && ++ !dev->srCache[i].locked && ++ (dev->srCache[i].lastUse < usage || !cache)) { ++ usage = dev->srCache[i].lastUse; ++ theObj = dev->srCache[i].object; ++ cache = &dev->srCache[i]; ++ pushout = i; ++ } ++ } ++ ++ if (!cache || cache->dirty) { ++ /* Flush and try again */ ++ yaffs_FlushFilesChunkCache(theObj); ++ cache = yaffs_GrabChunkCacheWorker(dev); ++ } ++ ++ } ++ return cache; ++ } else ++ return NULL; ++ ++} ++ ++/* Find a cached chunk */ ++static yaffs_ChunkCache *yaffs_FindChunkCache(const yaffs_Object *obj, ++ int chunkId) ++{ ++ yaffs_Device *dev = obj->myDev; ++ int i; ++ if (dev->nShortOpCaches > 0) { ++ for (i = 0; i < dev->nShortOpCaches; i++) { ++ if (dev->srCache[i].object == obj && ++ dev->srCache[i].chunkId == chunkId) { ++ dev->cacheHits++; ++ ++ return &dev->srCache[i]; ++ } ++ } ++ } ++ return NULL; ++} ++ ++/* Mark the chunk for the least recently used algorithym */ ++static void yaffs_UseChunkCache(yaffs_Device *dev, yaffs_ChunkCache *cache, ++ int isAWrite) ++{ ++ ++ if (dev->nShortOpCaches > 0) { ++ if (dev->srLastUse < 0 || dev->srLastUse > 100000000) { ++ /* Reset the cache usages */ ++ int i; ++ for (i = 1; i < dev->nShortOpCaches; i++) ++ dev->srCache[i].lastUse = 0; ++ ++ dev->srLastUse = 0; ++ } ++ ++ dev->srLastUse++; ++ ++ cache->lastUse = dev->srLastUse; ++ ++ if (isAWrite) ++ cache->dirty = 1; ++ } ++} ++ ++/* Invalidate a single cache page. ++ * Do this when a whole page gets written, ++ * ie the short cache for this page is no longer valid. ++ */ ++static void yaffs_InvalidateChunkCache(yaffs_Object *object, int chunkId) ++{ ++ if (object->myDev->nShortOpCaches > 0) { ++ yaffs_ChunkCache *cache = yaffs_FindChunkCache(object, chunkId); ++ ++ if (cache) ++ cache->object = NULL; ++ } ++} ++ ++/* Invalidate all the cache pages associated with this object ++ * Do this whenever ther file is deleted or resized. ++ */ ++static void yaffs_InvalidateWholeChunkCache(yaffs_Object *in) ++{ ++ int i; ++ yaffs_Device *dev = in->myDev; ++ ++ if (dev->nShortOpCaches > 0) { ++ /* Invalidate it. */ ++ for (i = 0; i < dev->nShortOpCaches; i++) { ++ if (dev->srCache[i].object == in) ++ dev->srCache[i].object = NULL; ++ } ++ } ++} ++ ++/*--------------------- Checkpointing --------------------*/ ++ ++ ++static int yaffs_WriteCheckpointValidityMarker(yaffs_Device *dev, int head) ++{ ++ yaffs_CheckpointValidity cp; ++ ++ memset(&cp, 0, sizeof(cp)); ++ ++ cp.structType = sizeof(cp); ++ cp.magic = YAFFS_MAGIC; ++ cp.version = YAFFS_CHECKPOINT_VERSION; ++ cp.head = (head) ? 1 : 0; ++ ++ return (yaffs_CheckpointWrite(dev, &cp, sizeof(cp)) == sizeof(cp)) ? ++ 1 : 0; ++} ++ ++static int yaffs_ReadCheckpointValidityMarker(yaffs_Device *dev, int head) ++{ ++ yaffs_CheckpointValidity cp; ++ int ok; ++ ++ ok = (yaffs_CheckpointRead(dev, &cp, sizeof(cp)) == sizeof(cp)); ++ ++ if (ok) ++ ok = (cp.structType == sizeof(cp)) && ++ (cp.magic == YAFFS_MAGIC) && ++ (cp.version == YAFFS_CHECKPOINT_VERSION) && ++ (cp.head == ((head) ? 1 : 0)); ++ return ok ? 1 : 0; ++} ++ ++static void yaffs_DeviceToCheckpointDevice(yaffs_CheckpointDevice *cp, ++ yaffs_Device *dev) ++{ ++ cp->nErasedBlocks = dev->nErasedBlocks; ++ cp->allocationBlock = dev->allocationBlock; ++ cp->allocationPage = dev->allocationPage; ++ cp->nFreeChunks = dev->nFreeChunks; ++ ++ cp->nDeletedFiles = dev->nDeletedFiles; ++ cp->nUnlinkedFiles = dev->nUnlinkedFiles; ++ cp->nBackgroundDeletions = dev->nBackgroundDeletions; ++ cp->sequenceNumber = dev->sequenceNumber; ++ cp->oldestDirtySequence = dev->oldestDirtySequence; ++ ++} ++ ++static void yaffs_CheckpointDeviceToDevice(yaffs_Device *dev, ++ yaffs_CheckpointDevice *cp) ++{ ++ dev->nErasedBlocks = cp->nErasedBlocks; ++ dev->allocationBlock = cp->allocationBlock; ++ dev->allocationPage = cp->allocationPage; ++ dev->nFreeChunks = cp->nFreeChunks; ++ ++ dev->nDeletedFiles = cp->nDeletedFiles; ++ dev->nUnlinkedFiles = cp->nUnlinkedFiles; ++ dev->nBackgroundDeletions = cp->nBackgroundDeletions; ++ dev->sequenceNumber = cp->sequenceNumber; ++ dev->oldestDirtySequence = cp->oldestDirtySequence; ++} ++ ++ ++static int yaffs_WriteCheckpointDevice(yaffs_Device *dev) ++{ ++ yaffs_CheckpointDevice cp; ++ __u32 nBytes; ++ __u32 nBlocks = (dev->internalEndBlock - dev->internalStartBlock + 1); ++ ++ int ok; ++ ++ /* Write device runtime values*/ ++ yaffs_DeviceToCheckpointDevice(&cp, dev); ++ cp.structType = sizeof(cp); ++ ++ ok = (yaffs_CheckpointWrite(dev, &cp, sizeof(cp)) == sizeof(cp)); ++ ++ /* Write block info */ ++ if (ok) { ++ nBytes = nBlocks * sizeof(yaffs_BlockInfo); ++ ok = (yaffs_CheckpointWrite(dev, dev->blockInfo, nBytes) == nBytes); ++ } ++ ++ /* Write chunk bits */ ++ if (ok) { ++ nBytes = nBlocks * dev->chunkBitmapStride; ++ ok = (yaffs_CheckpointWrite(dev, dev->chunkBits, nBytes) == nBytes); ++ } ++ return ok ? 1 : 0; ++ ++} ++ ++static int yaffs_ReadCheckpointDevice(yaffs_Device *dev) ++{ ++ yaffs_CheckpointDevice cp; ++ __u32 nBytes; ++ __u32 nBlocks = (dev->internalEndBlock - dev->internalStartBlock + 1); ++ ++ int ok; ++ ++ ok = (yaffs_CheckpointRead(dev, &cp, sizeof(cp)) == sizeof(cp)); ++ if (!ok) ++ return 0; ++ ++ if (cp.structType != sizeof(cp)) ++ return 0; ++ ++ ++ yaffs_CheckpointDeviceToDevice(dev, &cp); ++ ++ nBytes = nBlocks * sizeof(yaffs_BlockInfo); ++ ++ ok = (yaffs_CheckpointRead(dev, dev->blockInfo, nBytes) == nBytes); ++ ++ if (!ok) ++ return 0; ++ nBytes = nBlocks * dev->chunkBitmapStride; ++ ++ ok = (yaffs_CheckpointRead(dev, dev->chunkBits, nBytes) == nBytes); ++ ++ return ok ? 1 : 0; ++} ++ ++static void yaffs_ObjectToCheckpointObject(yaffs_CheckpointObject *cp, ++ yaffs_Object *obj) ++{ ++ ++ cp->objectId = obj->objectId; ++ cp->parentId = (obj->parent) ? obj->parent->objectId : 0; ++ cp->hdrChunk = obj->hdrChunk; ++ cp->variantType = obj->variantType; ++ cp->deleted = obj->deleted; ++ cp->softDeleted = obj->softDeleted; ++ cp->unlinked = obj->unlinked; ++ cp->fake = obj->fake; ++ cp->renameAllowed = obj->renameAllowed; ++ cp->unlinkAllowed = obj->unlinkAllowed; ++ cp->serial = obj->serial; ++ cp->nDataChunks = obj->nDataChunks; ++ ++ if (obj->variantType == YAFFS_OBJECT_TYPE_FILE) ++ cp->fileSizeOrEquivalentObjectId = obj->variant.fileVariant.fileSize; ++ else if (obj->variantType == YAFFS_OBJECT_TYPE_HARDLINK) ++ cp->fileSizeOrEquivalentObjectId = obj->variant.hardLinkVariant.equivalentObjectId; ++} ++ ++static int yaffs_CheckpointObjectToObject(yaffs_Object *obj, yaffs_CheckpointObject *cp) ++{ ++ ++ yaffs_Object *parent; ++ ++ if (obj->variantType != cp->variantType) { ++ T(YAFFS_TRACE_ERROR, (TSTR("Checkpoint read object %d type %d " ++ TCONT("chunk %d does not match existing object type %d") ++ TENDSTR), cp->objectId, cp->variantType, cp->hdrChunk, ++ obj->variantType)); ++ return 0; ++ } ++ ++ obj->objectId = cp->objectId; ++ ++ if (cp->parentId) ++ parent = yaffs_FindOrCreateObjectByNumber( ++ obj->myDev, ++ cp->parentId, ++ YAFFS_OBJECT_TYPE_DIRECTORY); ++ else ++ parent = NULL; ++ ++ if (parent) { ++ if (parent->variantType != YAFFS_OBJECT_TYPE_DIRECTORY) { ++ T(YAFFS_TRACE_ALWAYS, (TSTR("Checkpoint read object %d parent %d type %d" ++ TCONT(" chunk %d Parent type, %d, not directory") ++ TENDSTR), ++ cp->objectId, cp->parentId, cp->variantType, ++ cp->hdrChunk, parent->variantType)); ++ return 0; ++ } ++ yaffs_AddObjectToDirectory(parent, obj); ++ } ++ ++ obj->hdrChunk = cp->hdrChunk; ++ obj->variantType = cp->variantType; ++ obj->deleted = cp->deleted; ++ obj->softDeleted = cp->softDeleted; ++ obj->unlinked = cp->unlinked; ++ obj->fake = cp->fake; ++ obj->renameAllowed = cp->renameAllowed; ++ obj->unlinkAllowed = cp->unlinkAllowed; ++ obj->serial = cp->serial; ++ obj->nDataChunks = cp->nDataChunks; ++ ++ if (obj->variantType == YAFFS_OBJECT_TYPE_FILE) ++ obj->variant.fileVariant.fileSize = cp->fileSizeOrEquivalentObjectId; ++ else if (obj->variantType == YAFFS_OBJECT_TYPE_HARDLINK) ++ obj->variant.hardLinkVariant.equivalentObjectId = cp->fileSizeOrEquivalentObjectId; ++ ++ if (obj->hdrChunk > 0) ++ obj->lazyLoaded = 1; ++ return 1; ++} ++ ++ ++ ++static int yaffs_CheckpointTnodeWorker(yaffs_Object *in, yaffs_Tnode *tn, ++ __u32 level, int chunkOffset) ++{ ++ int i; ++ yaffs_Device *dev = in->myDev; ++ int ok = 1; ++ int tnodeSize = (dev->tnodeWidth * YAFFS_NTNODES_LEVEL0)/8; ++ ++ if (tnodeSize < sizeof(yaffs_Tnode)) ++ tnodeSize = sizeof(yaffs_Tnode); ++ ++ ++ if (tn) { ++ if (level > 0) { ++ ++ for (i = 0; i < YAFFS_NTNODES_INTERNAL && ok; i++) { ++ if (tn->internal[i]) { ++ ok = yaffs_CheckpointTnodeWorker(in, ++ tn->internal[i], ++ level - 1, ++ (chunkOffset<<YAFFS_TNODES_INTERNAL_BITS) + i); ++ } ++ } ++ } else if (level == 0) { ++ __u32 baseOffset = chunkOffset << YAFFS_TNODES_LEVEL0_BITS; ++ ok = (yaffs_CheckpointWrite(dev, &baseOffset, sizeof(baseOffset)) == sizeof(baseOffset)); ++ if (ok) ++ ok = (yaffs_CheckpointWrite(dev, tn, tnodeSize) == tnodeSize); ++ } ++ } ++ ++ return ok; ++ ++} ++ ++static int yaffs_WriteCheckpointTnodes(yaffs_Object *obj) ++{ ++ __u32 endMarker = ~0; ++ int ok = 1; ++ ++ if (obj->variantType == YAFFS_OBJECT_TYPE_FILE) { ++ ok = yaffs_CheckpointTnodeWorker(obj, ++ obj->variant.fileVariant.top, ++ obj->variant.fileVariant.topLevel, ++ 0); ++ if (ok) ++ ok = (yaffs_CheckpointWrite(obj->myDev, &endMarker, sizeof(endMarker)) == ++ sizeof(endMarker)); ++ } ++ ++ return ok ? 1 : 0; ++} ++ ++static int yaffs_ReadCheckpointTnodes(yaffs_Object *obj) ++{ ++ __u32 baseChunk; ++ int ok = 1; ++ yaffs_Device *dev = obj->myDev; ++ yaffs_FileStructure *fileStructPtr = &obj->variant.fileVariant; ++ yaffs_Tnode *tn; ++ int nread = 0; ++ int tnodeSize = (dev->tnodeWidth * YAFFS_NTNODES_LEVEL0)/8; ++ ++ if (tnodeSize < sizeof(yaffs_Tnode)) ++ tnodeSize = sizeof(yaffs_Tnode); ++ ++ ok = (yaffs_CheckpointRead(dev, &baseChunk, sizeof(baseChunk)) == sizeof(baseChunk)); ++ ++ while (ok && (~baseChunk)) { ++ nread++; ++ /* Read level 0 tnode */ ++ ++ ++ tn = yaffs_GetTnodeRaw(dev); ++ if (tn) ++ ok = (yaffs_CheckpointRead(dev, tn, tnodeSize) == tnodeSize); ++ else ++ ok = 0; ++ ++ if (tn && ok) ++ ok = yaffs_AddOrFindLevel0Tnode(dev, ++ fileStructPtr, ++ baseChunk, ++ tn) ? 1 : 0; ++ ++ if (ok) ++ ok = (yaffs_CheckpointRead(dev, &baseChunk, sizeof(baseChunk)) == sizeof(baseChunk)); ++ ++ } ++ ++ T(YAFFS_TRACE_CHECKPOINT, ( ++ TSTR("Checkpoint read tnodes %d records, last %d. ok %d" TENDSTR), ++ nread, baseChunk, ok)); ++ ++ return ok ? 1 : 0; ++} ++ ++ ++static int yaffs_WriteCheckpointObjects(yaffs_Device *dev) ++{ ++ yaffs_Object *obj; ++ yaffs_CheckpointObject cp; ++ int i; ++ int ok = 1; ++ struct ylist_head *lh; ++ ++ ++ /* Iterate through the objects in each hash entry, ++ * dumping them to the checkpointing stream. ++ */ ++ ++ for (i = 0; ok && i < YAFFS_NOBJECT_BUCKETS; i++) { ++ ylist_for_each(lh, &dev->objectBucket[i].list) { ++ if (lh) { ++ obj = ylist_entry(lh, yaffs_Object, hashLink); ++ if (!obj->deferedFree) { ++ yaffs_ObjectToCheckpointObject(&cp, obj); ++ cp.structType = sizeof(cp); ++ ++ T(YAFFS_TRACE_CHECKPOINT, ( ++ TSTR("Checkpoint write object %d parent %d type %d chunk %d obj addr %x" TENDSTR), ++ cp.objectId, cp.parentId, cp.variantType, cp.hdrChunk, (unsigned) obj)); ++ ++ ok = (yaffs_CheckpointWrite(dev, &cp, sizeof(cp)) == sizeof(cp)); ++ ++ if (ok && obj->variantType == YAFFS_OBJECT_TYPE_FILE) ++ ok = yaffs_WriteCheckpointTnodes(obj); ++ } ++ } ++ } ++ } ++ ++ /* Dump end of list */ ++ memset(&cp, 0xFF, sizeof(yaffs_CheckpointObject)); ++ cp.structType = sizeof(cp); ++ ++ if (ok) ++ ok = (yaffs_CheckpointWrite(dev, &cp, sizeof(cp)) == sizeof(cp)); ++ ++ return ok ? 1 : 0; ++} ++ ++static int yaffs_ReadCheckpointObjects(yaffs_Device *dev) ++{ ++ yaffs_Object *obj; ++ yaffs_CheckpointObject cp; ++ int ok = 1; ++ int done = 0; ++ yaffs_Object *hardList = NULL; ++ ++ while (ok && !done) { ++ ok = (yaffs_CheckpointRead(dev, &cp, sizeof(cp)) == sizeof(cp)); ++ if (cp.structType != sizeof(cp)) { ++ T(YAFFS_TRACE_CHECKPOINT, (TSTR("struct size %d instead of %d ok %d"TENDSTR), ++ cp.structType, sizeof(cp), ok)); ++ ok = 0; ++ } ++ ++ T(YAFFS_TRACE_CHECKPOINT, (TSTR("Checkpoint read object %d parent %d type %d chunk %d " TENDSTR), ++ cp.objectId, cp.parentId, cp.variantType, cp.hdrChunk)); ++ ++ if (ok && cp.objectId == ~0) ++ done = 1; ++ else if (ok) { ++ obj = yaffs_FindOrCreateObjectByNumber(dev, cp.objectId, cp.variantType); ++ if (obj) { ++ ok = yaffs_CheckpointObjectToObject(obj, &cp); ++ if (!ok) ++ break; ++ if (obj->variantType == YAFFS_OBJECT_TYPE_FILE) { ++ ok = yaffs_ReadCheckpointTnodes(obj); ++ } else if (obj->variantType == YAFFS_OBJECT_TYPE_HARDLINK) { ++ obj->hardLinks.next = ++ (struct ylist_head *) hardList; ++ hardList = obj; ++ } ++ } else ++ ok = 0; ++ } ++ } ++ ++ if (ok) ++ yaffs_HardlinkFixup(dev, hardList); ++ ++ return ok ? 1 : 0; ++} ++ ++static int yaffs_WriteCheckpointSum(yaffs_Device *dev) ++{ ++ __u32 checkpointSum; ++ int ok; ++ ++ yaffs_GetCheckpointSum(dev, &checkpointSum); ++ ++ ok = (yaffs_CheckpointWrite(dev, &checkpointSum, sizeof(checkpointSum)) == sizeof(checkpointSum)); ++ ++ if (!ok) ++ return 0; ++ ++ return 1; ++} ++ ++static int yaffs_ReadCheckpointSum(yaffs_Device *dev) ++{ ++ __u32 checkpointSum0; ++ __u32 checkpointSum1; ++ int ok; ++ ++ yaffs_GetCheckpointSum(dev, &checkpointSum0); ++ ++ ok = (yaffs_CheckpointRead(dev, &checkpointSum1, sizeof(checkpointSum1)) == sizeof(checkpointSum1)); ++ ++ if (!ok) ++ return 0; ++ ++ if (checkpointSum0 != checkpointSum1) ++ return 0; ++ ++ return 1; ++} ++ ++ ++static int yaffs_WriteCheckpointData(yaffs_Device *dev) ++{ ++ int ok = 1; ++ ++ if (dev->skipCheckpointWrite || !dev->isYaffs2) { ++ T(YAFFS_TRACE_CHECKPOINT, (TSTR("skipping checkpoint write" TENDSTR))); ++ ok = 0; ++ } ++ ++ if (ok) ++ ok = yaffs_CheckpointOpen(dev, 1); ++ ++ if (ok) { ++ T(YAFFS_TRACE_CHECKPOINT, (TSTR("write checkpoint validity" TENDSTR))); ++ ok = yaffs_WriteCheckpointValidityMarker(dev, 1); ++ } ++ if (ok) { ++ T(YAFFS_TRACE_CHECKPOINT, (TSTR("write checkpoint device" TENDSTR))); ++ ok = yaffs_WriteCheckpointDevice(dev); ++ } ++ if (ok) { ++ T(YAFFS_TRACE_CHECKPOINT, (TSTR("write checkpoint objects" TENDSTR))); ++ ok = yaffs_WriteCheckpointObjects(dev); ++ } ++ if (ok) { ++ T(YAFFS_TRACE_CHECKPOINT, (TSTR("write checkpoint validity" TENDSTR))); ++ ok = yaffs_WriteCheckpointValidityMarker(dev, 0); ++ } ++ ++ if (ok) ++ ok = yaffs_WriteCheckpointSum(dev); ++ ++ if (!yaffs_CheckpointClose(dev)) ++ ok = 0; ++ ++ if (ok) ++ dev->isCheckpointed = 1; ++ else ++ dev->isCheckpointed = 0; ++ ++ return dev->isCheckpointed; ++} ++ ++static int yaffs_ReadCheckpointData(yaffs_Device *dev) ++{ ++ int ok = 1; ++ ++ if (dev->skipCheckpointRead || !dev->isYaffs2) { ++ T(YAFFS_TRACE_CHECKPOINT, (TSTR("skipping checkpoint read" TENDSTR))); ++ ok = 0; ++ } ++ ++ if (ok) ++ ok = yaffs_CheckpointOpen(dev, 0); /* open for read */ ++ ++ if (ok) { ++ T(YAFFS_TRACE_CHECKPOINT, (TSTR("read checkpoint validity" TENDSTR))); ++ ok = yaffs_ReadCheckpointValidityMarker(dev, 1); ++ } ++ if (ok) { ++ T(YAFFS_TRACE_CHECKPOINT, (TSTR("read checkpoint device" TENDSTR))); ++ ok = yaffs_ReadCheckpointDevice(dev); ++ } ++ if (ok) { ++ T(YAFFS_TRACE_CHECKPOINT, (TSTR("read checkpoint objects" TENDSTR))); ++ ok = yaffs_ReadCheckpointObjects(dev); ++ } ++ if (ok) { ++ T(YAFFS_TRACE_CHECKPOINT, (TSTR("read checkpoint validity" TENDSTR))); ++ ok = yaffs_ReadCheckpointValidityMarker(dev, 0); ++ } ++ ++ if (ok) { ++ ok = yaffs_ReadCheckpointSum(dev); ++ T(YAFFS_TRACE_CHECKPOINT, (TSTR("read checkpoint checksum %d" TENDSTR), ok)); ++ } ++ ++ if (!yaffs_CheckpointClose(dev)) ++ ok = 0; ++ ++ if (ok) ++ dev->isCheckpointed = 1; ++ else ++ dev->isCheckpointed = 0; ++ ++ return ok ? 1 : 0; ++ ++} ++ ++static void yaffs_InvalidateCheckpoint(yaffs_Device *dev) ++{ ++ if (dev->isCheckpointed || ++ dev->blocksInCheckpoint > 0) { ++ dev->isCheckpointed = 0; ++ yaffs_CheckpointInvalidateStream(dev); ++ if (dev->superBlock && dev->markSuperBlockDirty) ++ dev->markSuperBlockDirty(dev->superBlock); ++ } ++} ++ ++ ++int yaffs_CheckpointSave(yaffs_Device *dev) ++{ ++ ++ T(YAFFS_TRACE_CHECKPOINT, (TSTR("save entry: isCheckpointed %d"TENDSTR), dev->isCheckpointed)); ++ ++ yaffs_VerifyObjects(dev); ++ yaffs_VerifyBlocks(dev); ++ yaffs_VerifyFreeChunks(dev); ++ ++ if (!dev->isCheckpointed) { ++ yaffs_InvalidateCheckpoint(dev); ++ yaffs_WriteCheckpointData(dev); ++ } ++ ++ T(YAFFS_TRACE_ALWAYS, (TSTR("save exit: isCheckpointed %d"TENDSTR), dev->isCheckpointed)); ++ ++ return dev->isCheckpointed; ++} ++ ++int yaffs_CheckpointRestore(yaffs_Device *dev) ++{ ++ int retval; ++ T(YAFFS_TRACE_CHECKPOINT, (TSTR("restore entry: isCheckpointed %d"TENDSTR), dev->isCheckpointed)); ++ ++ retval = yaffs_ReadCheckpointData(dev); ++ ++ if (dev->isCheckpointed) { ++ yaffs_VerifyObjects(dev); ++ yaffs_VerifyBlocks(dev); ++ yaffs_VerifyFreeChunks(dev); ++ } ++ ++ T(YAFFS_TRACE_CHECKPOINT, (TSTR("restore exit: isCheckpointed %d"TENDSTR), dev->isCheckpointed)); ++ ++ return retval; ++} ++ ++/*--------------------- File read/write ------------------------ ++ * Read and write have very similar structures. ++ * In general the read/write has three parts to it ++ * An incomplete chunk to start with (if the read/write is not chunk-aligned) ++ * Some complete chunks ++ * An incomplete chunk to end off with ++ * ++ * Curve-balls: the first chunk might also be the last chunk. ++ */ ++ ++int yaffs_ReadDataFromFile(yaffs_Object *in, __u8 *buffer, loff_t offset, ++ int nBytes) ++{ ++ ++ int chunk; ++ __u32 start; ++ int nToCopy; ++ int n = nBytes; ++ int nDone = 0; ++ yaffs_ChunkCache *cache; ++ ++ yaffs_Device *dev; ++ ++ dev = in->myDev; ++ ++ while (n > 0) { ++ /* chunk = offset / dev->nDataBytesPerChunk + 1; */ ++ /* start = offset % dev->nDataBytesPerChunk; */ ++ yaffs_AddrToChunk(dev, offset, &chunk, &start); ++ chunk++; ++ ++ /* OK now check for the curveball where the start and end are in ++ * the same chunk. ++ */ ++ if ((start + n) < dev->nDataBytesPerChunk) ++ nToCopy = n; ++ else ++ nToCopy = dev->nDataBytesPerChunk - start; ++ ++ cache = yaffs_FindChunkCache(in, chunk); ++ ++ /* If the chunk is already in the cache or it is less than a whole chunk ++ * or we're using inband tags then use the cache (if there is caching) ++ * else bypass the cache. ++ */ ++ if (cache || nToCopy != dev->nDataBytesPerChunk || dev->inbandTags) { ++ if (dev->nShortOpCaches > 0) { ++ ++ /* If we can't find the data in the cache, then load it up. */ ++ ++ if (!cache) { ++ cache = yaffs_GrabChunkCache(in->myDev); ++ cache->object = in; ++ cache->chunkId = chunk; ++ cache->dirty = 0; ++ cache->locked = 0; ++ yaffs_ReadChunkDataFromObject(in, chunk, ++ cache-> ++ data); ++ cache->nBytes = 0; ++ } ++ ++ yaffs_UseChunkCache(dev, cache, 0); ++ ++ cache->locked = 1; ++ ++ ++ memcpy(buffer, &cache->data[start], nToCopy); ++ ++ cache->locked = 0; ++ } else { ++ /* Read into the local buffer then copy..*/ ++ ++ __u8 *localBuffer = ++ yaffs_GetTempBuffer(dev, __LINE__); ++ yaffs_ReadChunkDataFromObject(in, chunk, ++ localBuffer); ++ ++ memcpy(buffer, &localBuffer[start], nToCopy); ++ ++ ++ yaffs_ReleaseTempBuffer(dev, localBuffer, ++ __LINE__); ++ } ++ ++ } else { ++ ++ /* A full chunk. Read directly into the supplied buffer. */ ++ yaffs_ReadChunkDataFromObject(in, chunk, buffer); ++ ++ } ++ ++ n -= nToCopy; ++ offset += nToCopy; ++ buffer += nToCopy; ++ nDone += nToCopy; ++ ++ } ++ ++ return nDone; ++} ++ ++int yaffs_WriteDataToFile(yaffs_Object *in, const __u8 *buffer, loff_t offset, ++ int nBytes, int writeThrough) ++{ ++ ++ int chunk; ++ __u32 start; ++ int nToCopy; ++ int n = nBytes; ++ int nDone = 0; ++ int nToWriteBack; ++ int startOfWrite = offset; ++ int chunkWritten = 0; ++ __u32 nBytesRead; ++ __u32 chunkStart; ++ ++ yaffs_Device *dev; ++ ++ dev = in->myDev; ++ ++ while (n > 0 && chunkWritten >= 0) { ++ /* chunk = offset / dev->nDataBytesPerChunk + 1; */ ++ /* start = offset % dev->nDataBytesPerChunk; */ ++ yaffs_AddrToChunk(dev, offset, &chunk, &start); ++ ++ if (chunk * dev->nDataBytesPerChunk + start != offset || ++ start >= dev->nDataBytesPerChunk) { ++ T(YAFFS_TRACE_ERROR, ( ++ TSTR("AddrToChunk of offset %d gives chunk %d start %d" ++ TENDSTR), ++ (int)offset, chunk, start)); ++ } ++ chunk++; ++ ++ /* OK now check for the curveball where the start and end are in ++ * the same chunk. ++ */ ++ ++ if ((start + n) < dev->nDataBytesPerChunk) { ++ nToCopy = n; ++ ++ /* Now folks, to calculate how many bytes to write back.... ++ * If we're overwriting and not writing to then end of file then ++ * we need to write back as much as was there before. ++ */ ++ ++ chunkStart = ((chunk - 1) * dev->nDataBytesPerChunk); ++ ++ if (chunkStart > in->variant.fileVariant.fileSize) ++ nBytesRead = 0; /* Past end of file */ ++ else ++ nBytesRead = in->variant.fileVariant.fileSize - chunkStart; ++ ++ if (nBytesRead > dev->nDataBytesPerChunk) ++ nBytesRead = dev->nDataBytesPerChunk; ++ ++ nToWriteBack = ++ (nBytesRead > ++ (start + n)) ? nBytesRead : (start + n); ++ ++ if (nToWriteBack < 0 || nToWriteBack > dev->nDataBytesPerChunk) ++ YBUG(); ++ ++ } else { ++ nToCopy = dev->nDataBytesPerChunk - start; ++ nToWriteBack = dev->nDataBytesPerChunk; ++ } ++ ++ if (nToCopy != dev->nDataBytesPerChunk || dev->inbandTags) { ++ /* An incomplete start or end chunk (or maybe both start and end chunk), ++ * or we're using inband tags, so we want to use the cache buffers. ++ */ ++ if (dev->nShortOpCaches > 0) { ++ yaffs_ChunkCache *cache; ++ /* If we can't find the data in the cache, then load the cache */ ++ cache = yaffs_FindChunkCache(in, chunk); ++ ++ if (!cache ++ && yaffs_CheckSpaceForAllocation(in-> ++ myDev)) { ++ cache = yaffs_GrabChunkCache(in->myDev); ++ cache->object = in; ++ cache->chunkId = chunk; ++ cache->dirty = 0; ++ cache->locked = 0; ++ yaffs_ReadChunkDataFromObject(in, chunk, ++ cache-> ++ data); ++ } else if (cache && ++ !cache->dirty && ++ !yaffs_CheckSpaceForAllocation(in->myDev)) { ++ /* Drop the cache if it was a read cache item and ++ * no space check has been made for it. ++ */ ++ cache = NULL; ++ } ++ ++ if (cache) { ++ yaffs_UseChunkCache(dev, cache, 1); ++ cache->locked = 1; ++ ++ ++ memcpy(&cache->data[start], buffer, ++ nToCopy); ++ ++ ++ cache->locked = 0; ++ cache->nBytes = nToWriteBack; ++ ++ if (writeThrough) { ++ chunkWritten = ++ yaffs_WriteChunkDataToObject ++ (cache->object, ++ cache->chunkId, ++ cache->data, cache->nBytes, ++ 1); ++ cache->dirty = 0; ++ } ++ ++ } else { ++ chunkWritten = -1; /* fail the write */ ++ } ++ } else { ++ /* An incomplete start or end chunk (or maybe both start and end chunk) ++ * Read into the local buffer then copy, then copy over and write back. ++ */ ++ ++ __u8 *localBuffer = ++ yaffs_GetTempBuffer(dev, __LINE__); ++ ++ yaffs_ReadChunkDataFromObject(in, chunk, ++ localBuffer); ++ ++ ++ ++ memcpy(&localBuffer[start], buffer, nToCopy); ++ ++ chunkWritten = ++ yaffs_WriteChunkDataToObject(in, chunk, ++ localBuffer, ++ nToWriteBack, ++ 0); ++ ++ yaffs_ReleaseTempBuffer(dev, localBuffer, ++ __LINE__); ++ ++ } ++ ++ } else { ++ /* A full chunk. Write directly from the supplied buffer. */ ++ ++ ++ ++ chunkWritten = ++ yaffs_WriteChunkDataToObject(in, chunk, buffer, ++ dev->nDataBytesPerChunk, ++ 0); ++ ++ /* Since we've overwritten the cached data, we better invalidate it. */ ++ yaffs_InvalidateChunkCache(in, chunk); ++ } ++ ++ if (chunkWritten >= 0) { ++ n -= nToCopy; ++ offset += nToCopy; ++ buffer += nToCopy; ++ nDone += nToCopy; ++ } ++ ++ } ++ ++ /* Update file object */ ++ ++ if ((startOfWrite + nDone) > in->variant.fileVariant.fileSize) ++ in->variant.fileVariant.fileSize = (startOfWrite + nDone); ++ ++ in->dirty = 1; ++ ++ return nDone; ++} ++ ++ ++/* ---------------------- File resizing stuff ------------------ */ ++ ++static void yaffs_PruneResizedChunks(yaffs_Object *in, int newSize) ++{ ++ ++ yaffs_Device *dev = in->myDev; ++ int oldFileSize = in->variant.fileVariant.fileSize; ++ ++ int lastDel = 1 + (oldFileSize - 1) / dev->nDataBytesPerChunk; ++ ++ int startDel = 1 + (newSize + dev->nDataBytesPerChunk - 1) / ++ dev->nDataBytesPerChunk; ++ int i; ++ int chunkId; ++ ++ /* Delete backwards so that we don't end up with holes if ++ * power is lost part-way through the operation. ++ */ ++ for (i = lastDel; i >= startDel; i--) { ++ /* NB this could be optimised somewhat, ++ * eg. could retrieve the tags and write them without ++ * using yaffs_DeleteChunk ++ */ ++ ++ chunkId = yaffs_FindAndDeleteChunkInFile(in, i, NULL); ++ if (chunkId > 0) { ++ if (chunkId < ++ (dev->internalStartBlock * dev->nChunksPerBlock) ++ || chunkId >= ++ ((dev->internalEndBlock + ++ 1) * dev->nChunksPerBlock)) { ++ T(YAFFS_TRACE_ALWAYS, ++ (TSTR("Found daft chunkId %d for %d" TENDSTR), ++ chunkId, i)); ++ } else { ++ in->nDataChunks--; ++ yaffs_DeleteChunk(dev, chunkId, 1, __LINE__); ++ } ++ } ++ } ++ ++} ++ ++int yaffs_ResizeFile(yaffs_Object *in, loff_t newSize) ++{ ++ ++ int oldFileSize = in->variant.fileVariant.fileSize; ++ __u32 newSizeOfPartialChunk; ++ int newFullChunks; ++ ++ yaffs_Device *dev = in->myDev; ++ ++ yaffs_AddrToChunk(dev, newSize, &newFullChunks, &newSizeOfPartialChunk); ++ ++ yaffs_FlushFilesChunkCache(in); ++ yaffs_InvalidateWholeChunkCache(in); ++ ++ yaffs_CheckGarbageCollection(dev); ++ ++ if (in->variantType != YAFFS_OBJECT_TYPE_FILE) ++ return YAFFS_FAIL; ++ ++ if (newSize == oldFileSize) ++ return YAFFS_OK; ++ ++ if (newSize < oldFileSize) { ++ ++ yaffs_PruneResizedChunks(in, newSize); ++ ++ if (newSizeOfPartialChunk != 0) { ++ int lastChunk = 1 + newFullChunks; ++ ++ __u8 *localBuffer = yaffs_GetTempBuffer(dev, __LINE__); ++ ++ /* Got to read and rewrite the last chunk with its new size and zero pad */ ++ yaffs_ReadChunkDataFromObject(in, lastChunk, ++ localBuffer); ++ ++ memset(localBuffer + newSizeOfPartialChunk, 0, ++ dev->nDataBytesPerChunk - newSizeOfPartialChunk); ++ ++ yaffs_WriteChunkDataToObject(in, lastChunk, localBuffer, ++ newSizeOfPartialChunk, 1); ++ ++ yaffs_ReleaseTempBuffer(dev, localBuffer, __LINE__); ++ } ++ ++ in->variant.fileVariant.fileSize = newSize; ++ ++ yaffs_PruneFileStructure(dev, &in->variant.fileVariant); ++ } else { ++ /* newsSize > oldFileSize */ ++ in->variant.fileVariant.fileSize = newSize; ++ } ++ ++ ++ /* Write a new object header. ++ * show we've shrunk the file, if need be ++ * Do this only if the file is not in the deleted directories. ++ */ ++ if (in->parent && ++ in->parent->objectId != YAFFS_OBJECTID_UNLINKED && ++ in->parent->objectId != YAFFS_OBJECTID_DELETED) ++ yaffs_UpdateObjectHeader(in, NULL, 0, ++ (newSize < oldFileSize) ? 1 : 0, 0); ++ ++ return YAFFS_OK; ++} ++ ++loff_t yaffs_GetFileSize(yaffs_Object *obj) ++{ ++ obj = yaffs_GetEquivalentObject(obj); ++ ++ switch (obj->variantType) { ++ case YAFFS_OBJECT_TYPE_FILE: ++ return obj->variant.fileVariant.fileSize; ++ case YAFFS_OBJECT_TYPE_SYMLINK: ++ return yaffs_strlen(obj->variant.symLinkVariant.alias); ++ default: ++ return 0; ++ } ++} ++ ++ ++ ++int yaffs_FlushFile(yaffs_Object *in, int updateTime) ++{ ++ int retVal; ++ if (in->dirty) { ++ yaffs_FlushFilesChunkCache(in); ++ if (updateTime) { ++#ifdef CONFIG_YAFFS_WINCE ++ yfsd_WinFileTimeNow(in->win_mtime); ++#else ++ ++ in->yst_mtime = Y_CURRENT_TIME; ++ ++#endif ++ } ++ ++ retVal = (yaffs_UpdateObjectHeader(in, NULL, 0, 0, 0) >= ++ 0) ? YAFFS_OK : YAFFS_FAIL; ++ } else { ++ retVal = YAFFS_OK; ++ } ++ ++ return retVal; ++ ++} ++ ++static int yaffs_DoGenericObjectDeletion(yaffs_Object *in) ++{ ++ ++ /* First off, invalidate the file's data in the cache, without flushing. */ ++ yaffs_InvalidateWholeChunkCache(in); ++ ++ if (in->myDev->isYaffs2 && (in->parent != in->myDev->deletedDir)) { ++ /* Move to the unlinked directory so we have a record that it was deleted. */ ++ yaffs_ChangeObjectName(in, in->myDev->deletedDir, _Y("deleted"), 0, 0); ++ ++ } ++ ++ yaffs_RemoveObjectFromDirectory(in); ++ yaffs_DeleteChunk(in->myDev, in->hdrChunk, 1, __LINE__); ++ in->hdrChunk = 0; ++ ++ yaffs_FreeObject(in); ++ return YAFFS_OK; ++ ++} ++ ++/* yaffs_DeleteFile deletes the whole file data ++ * and the inode associated with the file. ++ * It does not delete the links associated with the file. ++ */ ++static int yaffs_UnlinkFileIfNeeded(yaffs_Object *in) ++{ ++ ++ int retVal; ++ int immediateDeletion = 0; ++ ++#ifdef __KERNEL__ ++ if (!in->myInode) ++ immediateDeletion = 1; ++#else ++ if (in->inUse <= 0) ++ immediateDeletion = 1; ++#endif ++ ++ if (immediateDeletion) { ++ retVal = ++ yaffs_ChangeObjectName(in, in->myDev->deletedDir, ++ _Y("deleted"), 0, 0); ++ T(YAFFS_TRACE_TRACING, ++ (TSTR("yaffs: immediate deletion of file %d" TENDSTR), ++ in->objectId)); ++ in->deleted = 1; ++ in->myDev->nDeletedFiles++; ++ if (1 || in->myDev->isYaffs2) ++ yaffs_ResizeFile(in, 0); ++ yaffs_SoftDeleteFile(in); ++ } else { ++ retVal = ++ yaffs_ChangeObjectName(in, in->myDev->unlinkedDir, ++ _Y("unlinked"), 0, 0); ++ } ++ ++ ++ return retVal; ++} ++ ++int yaffs_DeleteFile(yaffs_Object *in) ++{ ++ int retVal = YAFFS_OK; ++ int deleted = in->deleted; ++ ++ yaffs_ResizeFile(in, 0); ++ ++ if (in->nDataChunks > 0) { ++ /* Use soft deletion if there is data in the file. ++ * That won't be the case if it has been resized to zero. ++ */ ++ if (!in->unlinked) ++ retVal = yaffs_UnlinkFileIfNeeded(in); ++ ++ if (retVal == YAFFS_OK && in->unlinked && !in->deleted) { ++ in->deleted = 1; ++ deleted = 1; ++ in->myDev->nDeletedFiles++; ++ yaffs_SoftDeleteFile(in); ++ } ++ return deleted ? YAFFS_OK : YAFFS_FAIL; ++ } else { ++ /* The file has no data chunks so we toss it immediately */ ++ yaffs_FreeTnode(in->myDev, in->variant.fileVariant.top); ++ in->variant.fileVariant.top = NULL; ++ yaffs_DoGenericObjectDeletion(in); ++ ++ return YAFFS_OK; ++ } ++} ++ ++static int yaffs_DeleteDirectory(yaffs_Object *in) ++{ ++ /* First check that the directory is empty. */ ++ if (ylist_empty(&in->variant.directoryVariant.children)) ++ return yaffs_DoGenericObjectDeletion(in); ++ ++ return YAFFS_FAIL; ++ ++} ++ ++static int yaffs_DeleteSymLink(yaffs_Object *in) ++{ ++ YFREE(in->variant.symLinkVariant.alias); ++ ++ return yaffs_DoGenericObjectDeletion(in); ++} ++ ++static int yaffs_DeleteHardLink(yaffs_Object *in) ++{ ++ /* remove this hardlink from the list assocaited with the equivalent ++ * object ++ */ ++ ylist_del_init(&in->hardLinks); ++ return yaffs_DoGenericObjectDeletion(in); ++} ++ ++int yaffs_DeleteObject(yaffs_Object *obj) ++{ ++int retVal = -1; ++ switch (obj->variantType) { ++ case YAFFS_OBJECT_TYPE_FILE: ++ retVal = yaffs_DeleteFile(obj); ++ break; ++ case YAFFS_OBJECT_TYPE_DIRECTORY: ++ return yaffs_DeleteDirectory(obj); ++ break; ++ case YAFFS_OBJECT_TYPE_SYMLINK: ++ retVal = yaffs_DeleteSymLink(obj); ++ break; ++ case YAFFS_OBJECT_TYPE_HARDLINK: ++ retVal = yaffs_DeleteHardLink(obj); ++ break; ++ case YAFFS_OBJECT_TYPE_SPECIAL: ++ retVal = yaffs_DoGenericObjectDeletion(obj); ++ break; ++ case YAFFS_OBJECT_TYPE_UNKNOWN: ++ retVal = 0; ++ break; /* should not happen. */ ++ } ++ ++ return retVal; ++} ++ ++static int yaffs_UnlinkWorker(yaffs_Object *obj) ++{ ++ ++ int immediateDeletion = 0; ++ ++#ifdef __KERNEL__ ++ if (!obj->myInode) ++ immediateDeletion = 1; ++#else ++ if (obj->inUse <= 0) ++ immediateDeletion = 1; ++#endif ++ ++ if (obj->variantType == YAFFS_OBJECT_TYPE_HARDLINK) { ++ return yaffs_DeleteHardLink(obj); ++ } else if (!ylist_empty(&obj->hardLinks)) { ++ /* Curve ball: We're unlinking an object that has a hardlink. ++ * ++ * This problem arises because we are not strictly following ++ * The Linux link/inode model. ++ * ++ * We can't really delete the object. ++ * Instead, we do the following: ++ * - Select a hardlink. ++ * - Unhook it from the hard links ++ * - Unhook it from its parent directory (so that the rename can work) ++ * - Rename the object to the hardlink's name. ++ * - Delete the hardlink ++ */ ++ ++ yaffs_Object *hl; ++ int retVal; ++ YCHAR name[YAFFS_MAX_NAME_LENGTH + 1]; ++ ++ hl = ylist_entry(obj->hardLinks.next, yaffs_Object, hardLinks); ++ ++ ylist_del_init(&hl->hardLinks); ++ ylist_del_init(&hl->siblings); ++ ++ yaffs_GetObjectName(hl, name, YAFFS_MAX_NAME_LENGTH + 1); ++ ++ retVal = yaffs_ChangeObjectName(obj, hl->parent, name, 0, 0); ++ ++ if (retVal == YAFFS_OK) ++ retVal = yaffs_DoGenericObjectDeletion(hl); ++ ++ return retVal; ++ ++ } else if (immediateDeletion) { ++ switch (obj->variantType) { ++ case YAFFS_OBJECT_TYPE_FILE: ++ return yaffs_DeleteFile(obj); ++ break; ++ case YAFFS_OBJECT_TYPE_DIRECTORY: ++ return yaffs_DeleteDirectory(obj); ++ break; ++ case YAFFS_OBJECT_TYPE_SYMLINK: ++ return yaffs_DeleteSymLink(obj); ++ break; ++ case YAFFS_OBJECT_TYPE_SPECIAL: ++ return yaffs_DoGenericObjectDeletion(obj); ++ break; ++ case YAFFS_OBJECT_TYPE_HARDLINK: ++ case YAFFS_OBJECT_TYPE_UNKNOWN: ++ default: ++ return YAFFS_FAIL; ++ } ++ } else ++ return yaffs_ChangeObjectName(obj, obj->myDev->unlinkedDir, ++ _Y("unlinked"), 0, 0); ++} ++ ++ ++static int yaffs_UnlinkObject(yaffs_Object *obj) ++{ ++ ++ if (obj && obj->unlinkAllowed) ++ return yaffs_UnlinkWorker(obj); ++ ++ return YAFFS_FAIL; ++ ++} ++int yaffs_Unlink(yaffs_Object *dir, const YCHAR *name) ++{ ++ yaffs_Object *obj; ++ ++ obj = yaffs_FindObjectByName(dir, name); ++ return yaffs_UnlinkObject(obj); ++} ++ ++/*----------------------- Initialisation Scanning ---------------------- */ ++ ++static void yaffs_HandleShadowedObject(yaffs_Device *dev, int objId, ++ int backwardScanning) ++{ ++ yaffs_Object *obj; ++ ++ if (!backwardScanning) { ++ /* Handle YAFFS1 forward scanning case ++ * For YAFFS1 we always do the deletion ++ */ ++ ++ } else { ++ /* Handle YAFFS2 case (backward scanning) ++ * If the shadowed object exists then ignore. ++ */ ++ if (yaffs_FindObjectByNumber(dev, objId)) ++ return; ++ } ++ ++ /* Let's create it (if it does not exist) assuming it is a file so that it can do shrinking etc. ++ * We put it in unlinked dir to be cleaned up after the scanning ++ */ ++ obj = ++ yaffs_FindOrCreateObjectByNumber(dev, objId, ++ YAFFS_OBJECT_TYPE_FILE); ++ if (!obj) ++ return; ++ yaffs_AddObjectToDirectory(dev->unlinkedDir, obj); ++ obj->variant.fileVariant.shrinkSize = 0; ++ obj->valid = 1; /* So that we don't read any other info for this file */ ++ ++} ++ ++typedef struct { ++ int seq; ++ int block; ++} yaffs_BlockIndex; ++ ++ ++static void yaffs_HardlinkFixup(yaffs_Device *dev, yaffs_Object *hardList) ++{ ++ yaffs_Object *hl; ++ yaffs_Object *in; ++ ++ while (hardList) { ++ hl = hardList; ++ hardList = (yaffs_Object *) (hardList->hardLinks.next); ++ ++ in = yaffs_FindObjectByNumber(dev, ++ hl->variant.hardLinkVariant. ++ equivalentObjectId); ++ ++ if (in) { ++ /* Add the hardlink pointers */ ++ hl->variant.hardLinkVariant.equivalentObject = in; ++ ylist_add(&hl->hardLinks, &in->hardLinks); ++ } else { ++ /* Todo Need to report/handle this better. ++ * Got a problem... hardlink to a non-existant object ++ */ ++ hl->variant.hardLinkVariant.equivalentObject = NULL; ++ YINIT_LIST_HEAD(&hl->hardLinks); ++ ++ } ++ } ++} ++ ++ ++ ++ ++ ++static int ybicmp(const void *a, const void *b) ++{ ++ register int aseq = ((yaffs_BlockIndex *)a)->seq; ++ register int bseq = ((yaffs_BlockIndex *)b)->seq; ++ register int ablock = ((yaffs_BlockIndex *)a)->block; ++ register int bblock = ((yaffs_BlockIndex *)b)->block; ++ if (aseq == bseq) ++ return ablock - bblock; ++ else ++ return aseq - bseq; ++} ++ ++ ++struct yaffs_ShadowFixerStruct { ++ int objectId; ++ int shadowedId; ++ struct yaffs_ShadowFixerStruct *next; ++}; ++ ++ ++static void yaffs_StripDeletedObjects(yaffs_Device *dev) ++{ ++ /* ++ * Sort out state of unlinked and deleted objects after scanning. ++ */ ++ struct ylist_head *i; ++ struct ylist_head *n; ++ yaffs_Object *l; ++ ++ /* Soft delete all the unlinked files */ ++ ylist_for_each_safe(i, n, ++ &dev->unlinkedDir->variant.directoryVariant.children) { ++ if (i) { ++ l = ylist_entry(i, yaffs_Object, siblings); ++ yaffs_DeleteObject(l); ++ } ++ } ++ ++ ylist_for_each_safe(i, n, ++ &dev->deletedDir->variant.directoryVariant.children) { ++ if (i) { ++ l = ylist_entry(i, yaffs_Object, siblings); ++ yaffs_DeleteObject(l); ++ } ++ } ++ ++} ++ ++static int yaffs_Scan(yaffs_Device *dev) ++{ ++ yaffs_ExtendedTags tags; ++ int blk; ++ int blockIterator; ++ int startIterator; ++ int endIterator; ++ int result; ++ ++ int chunk; ++ int c; ++ int deleted; ++ yaffs_BlockState state; ++ yaffs_Object *hardList = NULL; ++ yaffs_BlockInfo *bi; ++ __u32 sequenceNumber; ++ yaffs_ObjectHeader *oh; ++ yaffs_Object *in; ++ yaffs_Object *parent; ++ ++ int alloc_failed = 0; ++ ++ struct yaffs_ShadowFixerStruct *shadowFixerList = NULL; ++ ++ ++ __u8 *chunkData; ++ ++ ++ ++ T(YAFFS_TRACE_SCAN, ++ (TSTR("yaffs_Scan starts intstartblk %d intendblk %d..." TENDSTR), ++ dev->internalStartBlock, dev->internalEndBlock)); ++ ++ chunkData = yaffs_GetTempBuffer(dev, __LINE__); ++ ++ dev->sequenceNumber = YAFFS_LOWEST_SEQUENCE_NUMBER; ++ ++ /* Scan all the blocks to determine their state */ ++ for (blk = dev->internalStartBlock; blk <= dev->internalEndBlock; blk++) { ++ bi = yaffs_GetBlockInfo(dev, blk); ++ yaffs_ClearChunkBits(dev, blk); ++ bi->pagesInUse = 0; ++ bi->softDeletions = 0; ++ ++ yaffs_QueryInitialBlockState(dev, blk, &state, &sequenceNumber); ++ ++ bi->blockState = state; ++ bi->sequenceNumber = sequenceNumber; ++ ++ if (bi->sequenceNumber == YAFFS_SEQUENCE_BAD_BLOCK) ++ bi->blockState = state = YAFFS_BLOCK_STATE_DEAD; ++ ++ T(YAFFS_TRACE_SCAN_DEBUG, ++ (TSTR("Block scanning block %d state %d seq %d" TENDSTR), blk, ++ state, sequenceNumber)); ++ ++ if (state == YAFFS_BLOCK_STATE_DEAD) { ++ T(YAFFS_TRACE_BAD_BLOCKS, ++ (TSTR("block %d is bad" TENDSTR), blk)); ++ } else if (state == YAFFS_BLOCK_STATE_EMPTY) { ++ T(YAFFS_TRACE_SCAN_DEBUG, ++ (TSTR("Block empty " TENDSTR))); ++ dev->nErasedBlocks++; ++ dev->nFreeChunks += dev->nChunksPerBlock; ++ } ++ } ++ ++ startIterator = dev->internalStartBlock; ++ endIterator = dev->internalEndBlock; ++ ++ /* For each block.... */ ++ for (blockIterator = startIterator; !alloc_failed && blockIterator <= endIterator; ++ blockIterator++) { ++ ++ YYIELD(); ++ ++ YYIELD(); ++ ++ blk = blockIterator; ++ ++ bi = yaffs_GetBlockInfo(dev, blk); ++ state = bi->blockState; ++ ++ deleted = 0; ++ ++ /* For each chunk in each block that needs scanning....*/ ++ for (c = 0; !alloc_failed && c < dev->nChunksPerBlock && ++ state == YAFFS_BLOCK_STATE_NEEDS_SCANNING; c++) { ++ /* Read the tags and decide what to do */ ++ chunk = blk * dev->nChunksPerBlock + c; ++ ++ result = yaffs_ReadChunkWithTagsFromNAND(dev, chunk, NULL, ++ &tags); ++ ++ /* Let's have a good look at this chunk... */ ++ ++ if (tags.eccResult == YAFFS_ECC_RESULT_UNFIXED || tags.chunkDeleted) { ++ /* YAFFS1 only... ++ * A deleted chunk ++ */ ++ deleted++; ++ dev->nFreeChunks++; ++ /*T((" %d %d deleted\n",blk,c)); */ ++ } else if (!tags.chunkUsed) { ++ /* An unassigned chunk in the block ++ * This means that either the block is empty or ++ * this is the one being allocated from ++ */ ++ ++ if (c == 0) { ++ /* We're looking at the first chunk in the block so the block is unused */ ++ state = YAFFS_BLOCK_STATE_EMPTY; ++ dev->nErasedBlocks++; ++ } else { ++ /* this is the block being allocated from */ ++ T(YAFFS_TRACE_SCAN, ++ (TSTR ++ (" Allocating from %d %d" TENDSTR), ++ blk, c)); ++ state = YAFFS_BLOCK_STATE_ALLOCATING; ++ dev->allocationBlock = blk; ++ dev->allocationPage = c; ++ dev->allocationBlockFinder = blk; ++ /* Set it to here to encourage the allocator to go forth from here. */ ++ ++ } ++ ++ dev->nFreeChunks += (dev->nChunksPerBlock - c); ++ } else if (tags.chunkId > 0) { ++ /* chunkId > 0 so it is a data chunk... */ ++ unsigned int endpos; ++ ++ yaffs_SetChunkBit(dev, blk, c); ++ bi->pagesInUse++; ++ ++ in = yaffs_FindOrCreateObjectByNumber(dev, ++ tags. ++ objectId, ++ YAFFS_OBJECT_TYPE_FILE); ++ /* PutChunkIntoFile checks for a clash (two data chunks with ++ * the same chunkId). ++ */ ++ ++ if (!in) ++ alloc_failed = 1; ++ ++ if (in) { ++ if (!yaffs_PutChunkIntoFile(in, tags.chunkId, chunk, 1)) ++ alloc_failed = 1; ++ } ++ ++ endpos = ++ (tags.chunkId - 1) * dev->nDataBytesPerChunk + ++ tags.byteCount; ++ if (in && ++ in->variantType == YAFFS_OBJECT_TYPE_FILE ++ && in->variant.fileVariant.scannedFileSize < ++ endpos) { ++ in->variant.fileVariant. ++ scannedFileSize = endpos; ++ if (!dev->useHeaderFileSize) { ++ in->variant.fileVariant. ++ fileSize = ++ in->variant.fileVariant. ++ scannedFileSize; ++ } ++ ++ } ++ /* T((" %d %d data %d %d\n",blk,c,tags.objectId,tags.chunkId)); */ ++ } else { ++ /* chunkId == 0, so it is an ObjectHeader. ++ * Thus, we read in the object header and make the object ++ */ ++ yaffs_SetChunkBit(dev, blk, c); ++ bi->pagesInUse++; ++ ++ result = yaffs_ReadChunkWithTagsFromNAND(dev, chunk, ++ chunkData, ++ NULL); ++ ++ oh = (yaffs_ObjectHeader *) chunkData; ++ ++ in = yaffs_FindObjectByNumber(dev, ++ tags.objectId); ++ if (in && in->variantType != oh->type) { ++ /* This should not happen, but somehow ++ * Wev'e ended up with an objectId that has been reused but not yet ++ * deleted, and worse still it has changed type. Delete the old object. ++ */ ++ ++ yaffs_DeleteObject(in); ++ ++ in = 0; ++ } ++ ++ in = yaffs_FindOrCreateObjectByNumber(dev, ++ tags. ++ objectId, ++ oh->type); ++ ++ if (!in) ++ alloc_failed = 1; ++ ++ if (in && oh->shadowsObject > 0) { ++ ++ struct yaffs_ShadowFixerStruct *fixer; ++ fixer = YMALLOC(sizeof(struct yaffs_ShadowFixerStruct)); ++ if (fixer) { ++ fixer->next = shadowFixerList; ++ shadowFixerList = fixer; ++ fixer->objectId = tags.objectId; ++ fixer->shadowedId = oh->shadowsObject; ++ } ++ ++ } ++ ++ if (in && in->valid) { ++ /* We have already filled this one. We have a duplicate and need to resolve it. */ ++ ++ unsigned existingSerial = in->serial; ++ unsigned newSerial = tags.serialNumber; ++ ++ if (((existingSerial + 1) & 3) == newSerial) { ++ /* Use new one - destroy the exisiting one */ ++ yaffs_DeleteChunk(dev, ++ in->hdrChunk, ++ 1, __LINE__); ++ in->valid = 0; ++ } else { ++ /* Use existing - destroy this one. */ ++ yaffs_DeleteChunk(dev, chunk, 1, ++ __LINE__); ++ } ++ } ++ ++ if (in && !in->valid && ++ (tags.objectId == YAFFS_OBJECTID_ROOT || ++ tags.objectId == YAFFS_OBJECTID_LOSTNFOUND)) { ++ /* We only load some info, don't fiddle with directory structure */ ++ in->valid = 1; ++ in->variantType = oh->type; ++ ++ in->yst_mode = oh->yst_mode; ++#ifdef CONFIG_YAFFS_WINCE ++ in->win_atime[0] = oh->win_atime[0]; ++ in->win_ctime[0] = oh->win_ctime[0]; ++ in->win_mtime[0] = oh->win_mtime[0]; ++ in->win_atime[1] = oh->win_atime[1]; ++ in->win_ctime[1] = oh->win_ctime[1]; ++ in->win_mtime[1] = oh->win_mtime[1]; ++#else ++ in->yst_uid = oh->yst_uid; ++ in->yst_gid = oh->yst_gid; ++ in->yst_atime = oh->yst_atime; ++ in->yst_mtime = oh->yst_mtime; ++ in->yst_ctime = oh->yst_ctime; ++ in->yst_rdev = oh->yst_rdev; ++#endif ++ in->hdrChunk = chunk; ++ in->serial = tags.serialNumber; ++ ++ } else if (in && !in->valid) { ++ /* we need to load this info */ ++ ++ in->valid = 1; ++ in->variantType = oh->type; ++ ++ in->yst_mode = oh->yst_mode; ++#ifdef CONFIG_YAFFS_WINCE ++ in->win_atime[0] = oh->win_atime[0]; ++ in->win_ctime[0] = oh->win_ctime[0]; ++ in->win_mtime[0] = oh->win_mtime[0]; ++ in->win_atime[1] = oh->win_atime[1]; ++ in->win_ctime[1] = oh->win_ctime[1]; ++ in->win_mtime[1] = oh->win_mtime[1]; ++#else ++ in->yst_uid = oh->yst_uid; ++ in->yst_gid = oh->yst_gid; ++ in->yst_atime = oh->yst_atime; ++ in->yst_mtime = oh->yst_mtime; ++ in->yst_ctime = oh->yst_ctime; ++ in->yst_rdev = oh->yst_rdev; ++#endif ++ in->hdrChunk = chunk; ++ in->serial = tags.serialNumber; ++ ++ yaffs_SetObjectName(in, oh->name); ++ in->dirty = 0; ++ ++ /* directory stuff... ++ * hook up to parent ++ */ ++ ++ parent = ++ yaffs_FindOrCreateObjectByNumber ++ (dev, oh->parentObjectId, ++ YAFFS_OBJECT_TYPE_DIRECTORY); ++ if (!parent) ++ alloc_failed = 1; ++ if (parent && parent->variantType == ++ YAFFS_OBJECT_TYPE_UNKNOWN) { ++ /* Set up as a directory */ ++ parent->variantType = ++ YAFFS_OBJECT_TYPE_DIRECTORY; ++ YINIT_LIST_HEAD(&parent->variant. ++ directoryVariant. ++ children); ++ } else if (!parent || parent->variantType != ++ YAFFS_OBJECT_TYPE_DIRECTORY) { ++ /* Hoosterman, another problem.... ++ * We're trying to use a non-directory as a directory ++ */ ++ ++ T(YAFFS_TRACE_ERROR, ++ (TSTR ++ ("yaffs tragedy: attempting to use non-directory as a directory in scan. Put in lost+found." ++ TENDSTR))); ++ parent = dev->lostNFoundDir; ++ } ++ ++ yaffs_AddObjectToDirectory(parent, in); ++ ++ if (0 && (parent == dev->deletedDir || ++ parent == dev->unlinkedDir)) { ++ in->deleted = 1; /* If it is unlinked at start up then it wants deleting */ ++ dev->nDeletedFiles++; ++ } ++ /* Note re hardlinks. ++ * Since we might scan a hardlink before its equivalent object is scanned ++ * we put them all in a list. ++ * After scanning is complete, we should have all the objects, so we run through this ++ * list and fix up all the chains. ++ */ ++ ++ switch (in->variantType) { ++ case YAFFS_OBJECT_TYPE_UNKNOWN: ++ /* Todo got a problem */ ++ break; ++ case YAFFS_OBJECT_TYPE_FILE: ++ if (dev->useHeaderFileSize) ++ ++ in->variant.fileVariant. ++ fileSize = ++ oh->fileSize; ++ ++ break; ++ case YAFFS_OBJECT_TYPE_HARDLINK: ++ in->variant.hardLinkVariant. ++ equivalentObjectId = ++ oh->equivalentObjectId; ++ in->hardLinks.next = ++ (struct ylist_head *) ++ hardList; ++ hardList = in; ++ break; ++ case YAFFS_OBJECT_TYPE_DIRECTORY: ++ /* Do nothing */ ++ break; ++ case YAFFS_OBJECT_TYPE_SPECIAL: ++ /* Do nothing */ ++ break; ++ case YAFFS_OBJECT_TYPE_SYMLINK: ++ in->variant.symLinkVariant.alias = ++ yaffs_CloneString(oh->alias); ++ if (!in->variant.symLinkVariant.alias) ++ alloc_failed = 1; ++ break; ++ } ++ ++/* ++ if (parent == dev->deletedDir) { ++ yaffs_DestroyObject(in); ++ bi->hasShrinkHeader = 1; ++ } ++*/ ++ } ++ } ++ } ++ ++ if (state == YAFFS_BLOCK_STATE_NEEDS_SCANNING) { ++ /* If we got this far while scanning, then the block is fully allocated.*/ ++ state = YAFFS_BLOCK_STATE_FULL; ++ } ++ ++ bi->blockState = state; ++ ++ /* Now let's see if it was dirty */ ++ if (bi->pagesInUse == 0 && ++ !bi->hasShrinkHeader && ++ bi->blockState == YAFFS_BLOCK_STATE_FULL) { ++ yaffs_BlockBecameDirty(dev, blk); ++ } ++ ++ } ++ ++ ++ /* Ok, we've done all the scanning. ++ * Fix up the hard link chains. ++ * We should now have scanned all the objects, now it's time to add these ++ * hardlinks. ++ */ ++ ++ yaffs_HardlinkFixup(dev, hardList); ++ ++ /* Fix up any shadowed objects */ ++ { ++ struct yaffs_ShadowFixerStruct *fixer; ++ yaffs_Object *obj; ++ ++ while (shadowFixerList) { ++ fixer = shadowFixerList; ++ shadowFixerList = fixer->next; ++ /* Complete the rename transaction by deleting the shadowed object ++ * then setting the object header to unshadowed. ++ */ ++ obj = yaffs_FindObjectByNumber(dev, fixer->shadowedId); ++ if (obj) ++ yaffs_DeleteObject(obj); ++ ++ obj = yaffs_FindObjectByNumber(dev, fixer->objectId); ++ ++ if (obj) ++ yaffs_UpdateObjectHeader(obj, NULL, 1, 0, 0); ++ ++ YFREE(fixer); ++ } ++ } ++ ++ yaffs_ReleaseTempBuffer(dev, chunkData, __LINE__); ++ ++ if (alloc_failed) ++ return YAFFS_FAIL; ++ ++ T(YAFFS_TRACE_SCAN, (TSTR("yaffs_Scan ends" TENDSTR))); ++ ++ ++ return YAFFS_OK; ++} ++ ++static void yaffs_CheckObjectDetailsLoaded(yaffs_Object *in) ++{ ++ __u8 *chunkData; ++ yaffs_ObjectHeader *oh; ++ yaffs_Device *dev; ++ yaffs_ExtendedTags tags; ++ int result; ++ int alloc_failed = 0; ++ ++ if (!in) ++ return; ++ ++ dev = in->myDev; ++ ++#if 0 ++ T(YAFFS_TRACE_SCAN, (TSTR("details for object %d %s loaded" TENDSTR), ++ in->objectId, ++ in->lazyLoaded ? "not yet" : "already")); ++#endif ++ ++ if (in->lazyLoaded && in->hdrChunk > 0) { ++ in->lazyLoaded = 0; ++ chunkData = yaffs_GetTempBuffer(dev, __LINE__); ++ ++ result = yaffs_ReadChunkWithTagsFromNAND(dev, in->hdrChunk, chunkData, &tags); ++ oh = (yaffs_ObjectHeader *) chunkData; ++ ++ in->yst_mode = oh->yst_mode; ++#ifdef CONFIG_YAFFS_WINCE ++ in->win_atime[0] = oh->win_atime[0]; ++ in->win_ctime[0] = oh->win_ctime[0]; ++ in->win_mtime[0] = oh->win_mtime[0]; ++ in->win_atime[1] = oh->win_atime[1]; ++ in->win_ctime[1] = oh->win_ctime[1]; ++ in->win_mtime[1] = oh->win_mtime[1]; ++#else ++ in->yst_uid = oh->yst_uid; ++ in->yst_gid = oh->yst_gid; ++ in->yst_atime = oh->yst_atime; ++ in->yst_mtime = oh->yst_mtime; ++ in->yst_ctime = oh->yst_ctime; ++ in->yst_rdev = oh->yst_rdev; ++ ++#endif ++ yaffs_SetObjectName(in, oh->name); ++ ++ if (in->variantType == YAFFS_OBJECT_TYPE_SYMLINK) { ++ in->variant.symLinkVariant.alias = ++ yaffs_CloneString(oh->alias); ++ if (!in->variant.symLinkVariant.alias) ++ alloc_failed = 1; /* Not returned to caller */ ++ } ++ ++ yaffs_ReleaseTempBuffer(dev, chunkData, __LINE__); ++ } ++} ++ ++static int yaffs_ScanBackwards(yaffs_Device *dev) ++{ ++ yaffs_ExtendedTags tags; ++ int blk; ++ int blockIterator; ++ int startIterator; ++ int endIterator; ++ int nBlocksToScan = 0; ++ ++ int chunk; ++ int result; ++ int c; ++ int deleted; ++ yaffs_BlockState state; ++ yaffs_Object *hardList = NULL; ++ yaffs_BlockInfo *bi; ++ __u32 sequenceNumber; ++ yaffs_ObjectHeader *oh; ++ yaffs_Object *in; ++ yaffs_Object *parent; ++ int nBlocks = dev->internalEndBlock - dev->internalStartBlock + 1; ++ int itsUnlinked; ++ __u8 *chunkData; ++ ++ int fileSize; ++ int isShrink; ++ int foundChunksInBlock; ++ int equivalentObjectId; ++ int alloc_failed = 0; ++ ++ ++ yaffs_BlockIndex *blockIndex = NULL; ++ int altBlockIndex = 0; ++ ++ if (!dev->isYaffs2) { ++ T(YAFFS_TRACE_SCAN, ++ (TSTR("yaffs_ScanBackwards is only for YAFFS2!" TENDSTR))); ++ return YAFFS_FAIL; ++ } ++ ++ T(YAFFS_TRACE_SCAN, ++ (TSTR ++ ("yaffs_ScanBackwards starts intstartblk %d intendblk %d..." ++ TENDSTR), dev->internalStartBlock, dev->internalEndBlock)); ++ ++ ++ dev->sequenceNumber = YAFFS_LOWEST_SEQUENCE_NUMBER; ++ ++ blockIndex = YMALLOC(nBlocks * sizeof(yaffs_BlockIndex)); ++ ++ if (!blockIndex) { ++ blockIndex = YMALLOC_ALT(nBlocks * sizeof(yaffs_BlockIndex)); ++ altBlockIndex = 1; ++ } ++ ++ if (!blockIndex) { ++ T(YAFFS_TRACE_SCAN, ++ (TSTR("yaffs_Scan() could not allocate block index!" TENDSTR))); ++ return YAFFS_FAIL; ++ } ++ ++ dev->blocksInCheckpoint = 0; ++ ++ chunkData = yaffs_GetTempBuffer(dev, __LINE__); ++ ++ /* Scan all the blocks to determine their state */ ++ for (blk = dev->internalStartBlock; blk <= dev->internalEndBlock; blk++) { ++ bi = yaffs_GetBlockInfo(dev, blk); ++ yaffs_ClearChunkBits(dev, blk); ++ bi->pagesInUse = 0; ++ bi->softDeletions = 0; ++ ++ yaffs_QueryInitialBlockState(dev, blk, &state, &sequenceNumber); ++ ++ bi->blockState = state; ++ bi->sequenceNumber = sequenceNumber; ++ ++ if (bi->sequenceNumber == YAFFS_SEQUENCE_CHECKPOINT_DATA) ++ bi->blockState = state = YAFFS_BLOCK_STATE_CHECKPOINT; ++ if (bi->sequenceNumber == YAFFS_SEQUENCE_BAD_BLOCK) ++ bi->blockState = state = YAFFS_BLOCK_STATE_DEAD; ++ ++ T(YAFFS_TRACE_SCAN_DEBUG, ++ (TSTR("Block scanning block %d state %d seq %d" TENDSTR), blk, ++ state, sequenceNumber)); ++ ++ ++ if (state == YAFFS_BLOCK_STATE_CHECKPOINT) { ++ dev->blocksInCheckpoint++; ++ ++ } else if (state == YAFFS_BLOCK_STATE_DEAD) { ++ T(YAFFS_TRACE_BAD_BLOCKS, ++ (TSTR("block %d is bad" TENDSTR), blk)); ++ } else if (state == YAFFS_BLOCK_STATE_EMPTY) { ++ T(YAFFS_TRACE_SCAN_DEBUG, ++ (TSTR("Block empty " TENDSTR))); ++ dev->nErasedBlocks++; ++ dev->nFreeChunks += dev->nChunksPerBlock; ++ } else if (state == YAFFS_BLOCK_STATE_NEEDS_SCANNING) { ++ ++ /* Determine the highest sequence number */ ++ if (sequenceNumber >= YAFFS_LOWEST_SEQUENCE_NUMBER && ++ sequenceNumber < YAFFS_HIGHEST_SEQUENCE_NUMBER) { ++ ++ blockIndex[nBlocksToScan].seq = sequenceNumber; ++ blockIndex[nBlocksToScan].block = blk; ++ ++ nBlocksToScan++; ++ ++ if (sequenceNumber >= dev->sequenceNumber) ++ dev->sequenceNumber = sequenceNumber; ++ } else { ++ /* TODO: Nasty sequence number! */ ++ T(YAFFS_TRACE_SCAN, ++ (TSTR ++ ("Block scanning block %d has bad sequence number %d" ++ TENDSTR), blk, sequenceNumber)); ++ ++ } ++ } ++ } ++ ++ T(YAFFS_TRACE_SCAN, ++ (TSTR("%d blocks to be sorted..." TENDSTR), nBlocksToScan)); ++ ++ ++ ++ YYIELD(); ++ ++ /* Sort the blocks */ ++#ifndef CONFIG_YAFFS_USE_OWN_SORT ++ { ++ /* Use qsort now. */ ++ yaffs_qsort(blockIndex, nBlocksToScan, sizeof(yaffs_BlockIndex), ybicmp); ++ } ++#else ++ { ++ /* Dungy old bubble sort... */ ++ ++ yaffs_BlockIndex temp; ++ int i; ++ int j; ++ ++ for (i = 0; i < nBlocksToScan; i++) ++ for (j = i + 1; j < nBlocksToScan; j++) ++ if (blockIndex[i].seq > blockIndex[j].seq) { ++ temp = blockIndex[j]; ++ blockIndex[j] = blockIndex[i]; ++ blockIndex[i] = temp; ++ } ++ } ++#endif ++ ++ YYIELD(); ++ ++ T(YAFFS_TRACE_SCAN, (TSTR("...done" TENDSTR))); ++ ++ /* Now scan the blocks looking at the data. */ ++ startIterator = 0; ++ endIterator = nBlocksToScan - 1; ++ T(YAFFS_TRACE_SCAN_DEBUG, ++ (TSTR("%d blocks to be scanned" TENDSTR), nBlocksToScan)); ++ ++ /* For each block.... backwards */ ++ for (blockIterator = endIterator; !alloc_failed && blockIterator >= startIterator; ++ blockIterator--) { ++ /* Cooperative multitasking! This loop can run for so ++ long that watchdog timers expire. */ ++ YYIELD(); ++ ++ /* get the block to scan in the correct order */ ++ blk = blockIndex[blockIterator].block; ++ ++ bi = yaffs_GetBlockInfo(dev, blk); ++ ++ ++ state = bi->blockState; ++ ++ deleted = 0; ++ ++ /* For each chunk in each block that needs scanning.... */ ++ foundChunksInBlock = 0; ++ for (c = dev->nChunksPerBlock - 1; ++ !alloc_failed && c >= 0 && ++ (state == YAFFS_BLOCK_STATE_NEEDS_SCANNING || ++ state == YAFFS_BLOCK_STATE_ALLOCATING); c--) { ++ /* Scan backwards... ++ * Read the tags and decide what to do ++ */ ++ ++ chunk = blk * dev->nChunksPerBlock + c; ++ ++ result = yaffs_ReadChunkWithTagsFromNAND(dev, chunk, NULL, ++ &tags); ++ ++ /* Let's have a good look at this chunk... */ ++ ++ if (!tags.chunkUsed) { ++ /* An unassigned chunk in the block. ++ * If there are used chunks after this one, then ++ * it is a chunk that was skipped due to failing the erased ++ * check. Just skip it so that it can be deleted. ++ * But, more typically, We get here when this is an unallocated ++ * chunk and his means that either the block is empty or ++ * this is the one being allocated from ++ */ ++ ++ if (foundChunksInBlock) { ++ /* This is a chunk that was skipped due to failing the erased check */ ++ } else if (c == 0) { ++ /* We're looking at the first chunk in the block so the block is unused */ ++ state = YAFFS_BLOCK_STATE_EMPTY; ++ dev->nErasedBlocks++; ++ } else { ++ if (state == YAFFS_BLOCK_STATE_NEEDS_SCANNING || ++ state == YAFFS_BLOCK_STATE_ALLOCATING) { ++ if (dev->sequenceNumber == bi->sequenceNumber) { ++ /* this is the block being allocated from */ ++ ++ T(YAFFS_TRACE_SCAN, ++ (TSTR ++ (" Allocating from %d %d" ++ TENDSTR), blk, c)); ++ ++ state = YAFFS_BLOCK_STATE_ALLOCATING; ++ dev->allocationBlock = blk; ++ dev->allocationPage = c; ++ dev->allocationBlockFinder = blk; ++ } else { ++ /* This is a partially written block that is not ++ * the current allocation block. This block must have ++ * had a write failure, so set up for retirement. ++ */ ++ ++ /* bi->needsRetiring = 1; ??? TODO */ ++ bi->gcPrioritise = 1; ++ ++ T(YAFFS_TRACE_ALWAYS, ++ (TSTR("Partially written block %d detected" TENDSTR), ++ blk)); ++ } ++ } ++ } ++ ++ dev->nFreeChunks++; ++ ++ } else if (tags.eccResult == YAFFS_ECC_RESULT_UNFIXED) { ++ T(YAFFS_TRACE_SCAN, ++ (TSTR(" Unfixed ECC in chunk(%d:%d), chunk ignored"TENDSTR), ++ blk, c)); ++ ++ dev->nFreeChunks++; ++ ++ } else if (tags.chunkId > 0) { ++ /* chunkId > 0 so it is a data chunk... */ ++ unsigned int endpos; ++ __u32 chunkBase = ++ (tags.chunkId - 1) * dev->nDataBytesPerChunk; ++ ++ foundChunksInBlock = 1; ++ ++ ++ yaffs_SetChunkBit(dev, blk, c); ++ bi->pagesInUse++; ++ ++ in = yaffs_FindOrCreateObjectByNumber(dev, ++ tags. ++ objectId, ++ YAFFS_OBJECT_TYPE_FILE); ++ if (!in) { ++ /* Out of memory */ ++ alloc_failed = 1; ++ } ++ ++ if (in && ++ in->variantType == YAFFS_OBJECT_TYPE_FILE ++ && chunkBase < ++ in->variant.fileVariant.shrinkSize) { ++ /* This has not been invalidated by a resize */ ++ if (!yaffs_PutChunkIntoFile(in, tags.chunkId, ++ chunk, -1)) { ++ alloc_failed = 1; ++ } ++ ++ /* File size is calculated by looking at the data chunks if we have not ++ * seen an object header yet. Stop this practice once we find an object header. ++ */ ++ endpos = ++ (tags.chunkId - ++ 1) * dev->nDataBytesPerChunk + ++ tags.byteCount; ++ ++ if (!in->valid && /* have not got an object header yet */ ++ in->variant.fileVariant. ++ scannedFileSize < endpos) { ++ in->variant.fileVariant. ++ scannedFileSize = endpos; ++ in->variant.fileVariant. ++ fileSize = ++ in->variant.fileVariant. ++ scannedFileSize; ++ } ++ ++ } else if (in) { ++ /* This chunk has been invalidated by a resize, so delete */ ++ yaffs_DeleteChunk(dev, chunk, 1, __LINE__); ++ ++ } ++ } else { ++ /* chunkId == 0, so it is an ObjectHeader. ++ * Thus, we read in the object header and make the object ++ */ ++ foundChunksInBlock = 1; ++ ++ yaffs_SetChunkBit(dev, blk, c); ++ bi->pagesInUse++; ++ ++ oh = NULL; ++ in = NULL; ++ ++ if (tags.extraHeaderInfoAvailable) { ++ in = yaffs_FindOrCreateObjectByNumber ++ (dev, tags.objectId, ++ tags.extraObjectType); ++ if (!in) ++ alloc_failed = 1; ++ } ++ ++ if (!in || ++#ifdef CONFIG_YAFFS_DISABLE_LAZY_LOAD ++ !in->valid || ++#endif ++ tags.extraShadows || ++ (!in->valid && ++ (tags.objectId == YAFFS_OBJECTID_ROOT || ++ tags.objectId == YAFFS_OBJECTID_LOSTNFOUND))) { ++ ++ /* If we don't have valid info then we need to read the chunk ++ * TODO In future we can probably defer reading the chunk and ++ * living with invalid data until needed. ++ */ ++ ++ result = yaffs_ReadChunkWithTagsFromNAND(dev, ++ chunk, ++ chunkData, ++ NULL); ++ ++ oh = (yaffs_ObjectHeader *) chunkData; ++ ++ if (dev->inbandTags) { ++ /* Fix up the header if they got corrupted by inband tags */ ++ oh->shadowsObject = oh->inbandShadowsObject; ++ oh->isShrink = oh->inbandIsShrink; ++ } ++ ++ if (!in) { ++ in = yaffs_FindOrCreateObjectByNumber(dev, tags.objectId, oh->type); ++ if (!in) ++ alloc_failed = 1; ++ } ++ ++ } ++ ++ if (!in) { ++ /* TODO Hoosterman we have a problem! */ ++ T(YAFFS_TRACE_ERROR, ++ (TSTR ++ ("yaffs tragedy: Could not make object for object %d at chunk %d during scan" ++ TENDSTR), tags.objectId, chunk)); ++ continue; ++ } ++ ++ if (in->valid) { ++ /* We have already filled this one. ++ * We have a duplicate that will be discarded, but ++ * we first have to suck out resize info if it is a file. ++ */ ++ ++ if ((in->variantType == YAFFS_OBJECT_TYPE_FILE) && ++ ((oh && ++ oh->type == YAFFS_OBJECT_TYPE_FILE) || ++ (tags.extraHeaderInfoAvailable && ++ tags.extraObjectType == YAFFS_OBJECT_TYPE_FILE))) { ++ __u32 thisSize = ++ (oh) ? oh->fileSize : tags. ++ extraFileLength; ++ __u32 parentObjectId = ++ (oh) ? oh-> ++ parentObjectId : tags. ++ extraParentObjectId; ++ ++ ++ isShrink = ++ (oh) ? oh->isShrink : tags. ++ extraIsShrinkHeader; ++ ++ /* If it is deleted (unlinked at start also means deleted) ++ * we treat the file size as being zeroed at this point. ++ */ ++ if (parentObjectId == ++ YAFFS_OBJECTID_DELETED ++ || parentObjectId == ++ YAFFS_OBJECTID_UNLINKED) { ++ thisSize = 0; ++ isShrink = 1; ++ } ++ ++ if (isShrink && ++ in->variant.fileVariant. ++ shrinkSize > thisSize) { ++ in->variant.fileVariant. ++ shrinkSize = ++ thisSize; ++ } ++ ++ if (isShrink) ++ bi->hasShrinkHeader = 1; ++ ++ } ++ /* Use existing - destroy this one. */ ++ yaffs_DeleteChunk(dev, chunk, 1, __LINE__); ++ ++ } ++ ++ if (!in->valid && in->variantType != ++ (oh ? oh->type : tags.extraObjectType)) ++ T(YAFFS_TRACE_ERROR, ( ++ TSTR("yaffs tragedy: Bad object type, " ++ TCONT("%d != %d, for object %d at chunk ") ++ TCONT("%d during scan") ++ TENDSTR), oh ? ++ oh->type : tags.extraObjectType, ++ in->variantType, tags.objectId, ++ chunk)); ++ ++ if (!in->valid && ++ (tags.objectId == YAFFS_OBJECTID_ROOT || ++ tags.objectId == ++ YAFFS_OBJECTID_LOSTNFOUND)) { ++ /* We only load some info, don't fiddle with directory structure */ ++ in->valid = 1; ++ ++ if (oh) { ++ in->variantType = oh->type; ++ ++ in->yst_mode = oh->yst_mode; ++#ifdef CONFIG_YAFFS_WINCE ++ in->win_atime[0] = oh->win_atime[0]; ++ in->win_ctime[0] = oh->win_ctime[0]; ++ in->win_mtime[0] = oh->win_mtime[0]; ++ in->win_atime[1] = oh->win_atime[1]; ++ in->win_ctime[1] = oh->win_ctime[1]; ++ in->win_mtime[1] = oh->win_mtime[1]; ++#else ++ in->yst_uid = oh->yst_uid; ++ in->yst_gid = oh->yst_gid; ++ in->yst_atime = oh->yst_atime; ++ in->yst_mtime = oh->yst_mtime; ++ in->yst_ctime = oh->yst_ctime; ++ in->yst_rdev = oh->yst_rdev; ++ ++#endif ++ } else { ++ in->variantType = tags.extraObjectType; ++ in->lazyLoaded = 1; ++ } ++ ++ in->hdrChunk = chunk; ++ ++ } else if (!in->valid) { ++ /* we need to load this info */ ++ ++ in->valid = 1; ++ in->hdrChunk = chunk; ++ ++ if (oh) { ++ in->variantType = oh->type; ++ ++ in->yst_mode = oh->yst_mode; ++#ifdef CONFIG_YAFFS_WINCE ++ in->win_atime[0] = oh->win_atime[0]; ++ in->win_ctime[0] = oh->win_ctime[0]; ++ in->win_mtime[0] = oh->win_mtime[0]; ++ in->win_atime[1] = oh->win_atime[1]; ++ in->win_ctime[1] = oh->win_ctime[1]; ++ in->win_mtime[1] = oh->win_mtime[1]; ++#else ++ in->yst_uid = oh->yst_uid; ++ in->yst_gid = oh->yst_gid; ++ in->yst_atime = oh->yst_atime; ++ in->yst_mtime = oh->yst_mtime; ++ in->yst_ctime = oh->yst_ctime; ++ in->yst_rdev = oh->yst_rdev; ++#endif ++ ++ if (oh->shadowsObject > 0) ++ yaffs_HandleShadowedObject(dev, ++ oh-> ++ shadowsObject, ++ 1); ++ ++ ++ yaffs_SetObjectName(in, oh->name); ++ parent = ++ yaffs_FindOrCreateObjectByNumber ++ (dev, oh->parentObjectId, ++ YAFFS_OBJECT_TYPE_DIRECTORY); ++ ++ fileSize = oh->fileSize; ++ isShrink = oh->isShrink; ++ equivalentObjectId = oh->equivalentObjectId; ++ ++ } else { ++ in->variantType = tags.extraObjectType; ++ parent = ++ yaffs_FindOrCreateObjectByNumber ++ (dev, tags.extraParentObjectId, ++ YAFFS_OBJECT_TYPE_DIRECTORY); ++ fileSize = tags.extraFileLength; ++ isShrink = tags.extraIsShrinkHeader; ++ equivalentObjectId = tags.extraEquivalentObjectId; ++ in->lazyLoaded = 1; ++ ++ } ++ in->dirty = 0; ++ ++ if (!parent) ++ alloc_failed = 1; ++ ++ /* directory stuff... ++ * hook up to parent ++ */ ++ ++ if (parent && parent->variantType == ++ YAFFS_OBJECT_TYPE_UNKNOWN) { ++ /* Set up as a directory */ ++ parent->variantType = ++ YAFFS_OBJECT_TYPE_DIRECTORY; ++ YINIT_LIST_HEAD(&parent->variant. ++ directoryVariant. ++ children); ++ } else if (!parent || parent->variantType != ++ YAFFS_OBJECT_TYPE_DIRECTORY) { ++ /* Hoosterman, another problem.... ++ * We're trying to use a non-directory as a directory ++ */ ++ ++ T(YAFFS_TRACE_ERROR, ++ (TSTR ++ ("yaffs tragedy: attempting to use non-directory as a directory in scan. Put in lost+found." ++ TENDSTR))); ++ parent = dev->lostNFoundDir; ++ } ++ ++ yaffs_AddObjectToDirectory(parent, in); ++ ++ itsUnlinked = (parent == dev->deletedDir) || ++ (parent == dev->unlinkedDir); ++ ++ if (isShrink) { ++ /* Mark the block as having a shrinkHeader */ ++ bi->hasShrinkHeader = 1; ++ } ++ ++ /* Note re hardlinks. ++ * Since we might scan a hardlink before its equivalent object is scanned ++ * we put them all in a list. ++ * After scanning is complete, we should have all the objects, so we run ++ * through this list and fix up all the chains. ++ */ ++ ++ switch (in->variantType) { ++ case YAFFS_OBJECT_TYPE_UNKNOWN: ++ /* Todo got a problem */ ++ break; ++ case YAFFS_OBJECT_TYPE_FILE: ++ ++ if (in->variant.fileVariant. ++ scannedFileSize < fileSize) { ++ /* This covers the case where the file size is greater ++ * than where the data is ++ * This will happen if the file is resized to be larger ++ * than its current data extents. ++ */ ++ in->variant.fileVariant.fileSize = fileSize; ++ in->variant.fileVariant.scannedFileSize = ++ in->variant.fileVariant.fileSize; ++ } ++ ++ if (isShrink && ++ in->variant.fileVariant.shrinkSize > fileSize) { ++ in->variant.fileVariant.shrinkSize = fileSize; ++ } ++ ++ break; ++ case YAFFS_OBJECT_TYPE_HARDLINK: ++ if (!itsUnlinked) { ++ in->variant.hardLinkVariant.equivalentObjectId = ++ equivalentObjectId; ++ in->hardLinks.next = ++ (struct ylist_head *) hardList; ++ hardList = in; ++ } ++ break; ++ case YAFFS_OBJECT_TYPE_DIRECTORY: ++ /* Do nothing */ ++ break; ++ case YAFFS_OBJECT_TYPE_SPECIAL: ++ /* Do nothing */ ++ break; ++ case YAFFS_OBJECT_TYPE_SYMLINK: ++ if (oh) { ++ in->variant.symLinkVariant.alias = ++ yaffs_CloneString(oh->alias); ++ if (!in->variant.symLinkVariant.alias) ++ alloc_failed = 1; ++ } ++ break; ++ } ++ ++ } ++ ++ } ++ ++ } /* End of scanning for each chunk */ ++ ++ if (state == YAFFS_BLOCK_STATE_NEEDS_SCANNING) { ++ /* If we got this far while scanning, then the block is fully allocated. */ ++ state = YAFFS_BLOCK_STATE_FULL; ++ } ++ ++ bi->blockState = state; ++ ++ /* Now let's see if it was dirty */ ++ if (bi->pagesInUse == 0 && ++ !bi->hasShrinkHeader && ++ bi->blockState == YAFFS_BLOCK_STATE_FULL) { ++ yaffs_BlockBecameDirty(dev, blk); ++ } ++ ++ } ++ ++ if (altBlockIndex) ++ YFREE_ALT(blockIndex); ++ else ++ YFREE(blockIndex); ++ ++ /* Ok, we've done all the scanning. ++ * Fix up the hard link chains. ++ * We should now have scanned all the objects, now it's time to add these ++ * hardlinks. ++ */ ++ yaffs_HardlinkFixup(dev, hardList); ++ ++ ++ yaffs_ReleaseTempBuffer(dev, chunkData, __LINE__); ++ ++ if (alloc_failed) ++ return YAFFS_FAIL; ++ ++ T(YAFFS_TRACE_SCAN, (TSTR("yaffs_ScanBackwards ends" TENDSTR))); ++ ++ return YAFFS_OK; ++} ++ ++/*------------------------------ Directory Functions ----------------------------- */ ++ ++static void yaffs_VerifyObjectInDirectory(yaffs_Object *obj) ++{ ++ struct ylist_head *lh; ++ yaffs_Object *listObj; ++ ++ int count = 0; ++ ++ if (!obj) { ++ T(YAFFS_TRACE_ALWAYS, (TSTR("No object to verify" TENDSTR))); ++ YBUG(); ++ return; ++ } ++ ++ if (yaffs_SkipVerification(obj->myDev)) ++ return; ++ ++ if (!obj->parent) { ++ T(YAFFS_TRACE_ALWAYS, (TSTR("Object does not have parent" TENDSTR))); ++ YBUG(); ++ return; ++ } ++ ++ if (obj->parent->variantType != YAFFS_OBJECT_TYPE_DIRECTORY) { ++ T(YAFFS_TRACE_ALWAYS, (TSTR("Parent is not directory" TENDSTR))); ++ YBUG(); ++ } ++ ++ /* Iterate through the objects in each hash entry */ ++ ++ ylist_for_each(lh, &obj->parent->variant.directoryVariant.children) { ++ if (lh) { ++ listObj = ylist_entry(lh, yaffs_Object, siblings); ++ yaffs_VerifyObject(listObj); ++ if (obj == listObj) ++ count++; ++ } ++ } ++ ++ if (count != 1) { ++ T(YAFFS_TRACE_ALWAYS, (TSTR("Object in directory %d times" TENDSTR), count)); ++ YBUG(); ++ } ++} ++ ++static void yaffs_VerifyDirectory(yaffs_Object *directory) ++{ ++ struct ylist_head *lh; ++ yaffs_Object *listObj; ++ ++ if (!directory) { ++ YBUG(); ++ return; ++ } ++ ++ if (yaffs_SkipFullVerification(directory->myDev)) ++ return; ++ ++ if (directory->variantType != YAFFS_OBJECT_TYPE_DIRECTORY) { ++ T(YAFFS_TRACE_ALWAYS, (TSTR("Directory has wrong type: %d" TENDSTR), directory->variantType)); ++ YBUG(); ++ } ++ ++ /* Iterate through the objects in each hash entry */ ++ ++ ylist_for_each(lh, &directory->variant.directoryVariant.children) { ++ if (lh) { ++ listObj = ylist_entry(lh, yaffs_Object, siblings); ++ if (listObj->parent != directory) { ++ T(YAFFS_TRACE_ALWAYS, (TSTR("Object in directory list has wrong parent %p" TENDSTR), listObj->parent)); ++ YBUG(); ++ } ++ yaffs_VerifyObjectInDirectory(listObj); ++ } ++ } ++} ++ ++ ++static void yaffs_RemoveObjectFromDirectory(yaffs_Object *obj) ++{ ++ yaffs_Device *dev = obj->myDev; ++ yaffs_Object *parent; ++ ++ yaffs_VerifyObjectInDirectory(obj); ++ parent = obj->parent; ++ ++ yaffs_VerifyDirectory(parent); ++ ++ if (dev && dev->removeObjectCallback) ++ dev->removeObjectCallback(obj); ++ ++ ++ ylist_del_init(&obj->siblings); ++ obj->parent = NULL; ++ ++ yaffs_VerifyDirectory(parent); ++} ++ ++ ++static void yaffs_AddObjectToDirectory(yaffs_Object *directory, ++ yaffs_Object *obj) ++{ ++ if (!directory) { ++ T(YAFFS_TRACE_ALWAYS, ++ (TSTR ++ ("tragedy: Trying to add an object to a null pointer directory" ++ TENDSTR))); ++ YBUG(); ++ return; ++ } ++ if (directory->variantType != YAFFS_OBJECT_TYPE_DIRECTORY) { ++ T(YAFFS_TRACE_ALWAYS, ++ (TSTR ++ ("tragedy: Trying to add an object to a non-directory" ++ TENDSTR))); ++ YBUG(); ++ } ++ ++ if (obj->siblings.prev == NULL) { ++ /* Not initialised */ ++ YBUG(); ++ } ++ ++ ++ yaffs_VerifyDirectory(directory); ++ ++ yaffs_RemoveObjectFromDirectory(obj); ++ ++ ++ /* Now add it */ ++ ylist_add(&obj->siblings, &directory->variant.directoryVariant.children); ++ obj->parent = directory; ++ ++ if (directory == obj->myDev->unlinkedDir ++ || directory == obj->myDev->deletedDir) { ++ obj->unlinked = 1; ++ obj->myDev->nUnlinkedFiles++; ++ obj->renameAllowed = 0; ++ } ++ ++ yaffs_VerifyDirectory(directory); ++ yaffs_VerifyObjectInDirectory(obj); ++} ++ ++yaffs_Object *yaffs_FindObjectByName(yaffs_Object *directory, ++ const YCHAR *name) ++{ ++ int sum; ++ ++ struct ylist_head *i; ++ YCHAR buffer[YAFFS_MAX_NAME_LENGTH + 1]; ++ ++ yaffs_Object *l; ++ ++ if (!name) ++ return NULL; ++ ++ if (!directory) { ++ T(YAFFS_TRACE_ALWAYS, ++ (TSTR ++ ("tragedy: yaffs_FindObjectByName: null pointer directory" ++ TENDSTR))); ++ YBUG(); ++ return NULL; ++ } ++ if (directory->variantType != YAFFS_OBJECT_TYPE_DIRECTORY) { ++ T(YAFFS_TRACE_ALWAYS, ++ (TSTR ++ ("tragedy: yaffs_FindObjectByName: non-directory" TENDSTR))); ++ YBUG(); ++ } ++ ++ sum = yaffs_CalcNameSum(name); ++ ++ ylist_for_each(i, &directory->variant.directoryVariant.children) { ++ if (i) { ++ l = ylist_entry(i, yaffs_Object, siblings); ++ ++ if (l->parent != directory) ++ YBUG(); ++ ++ yaffs_CheckObjectDetailsLoaded(l); ++ ++ /* Special case for lost-n-found */ ++ if (l->objectId == YAFFS_OBJECTID_LOSTNFOUND) { ++ if (yaffs_strcmp(name, YAFFS_LOSTNFOUND_NAME) == 0) ++ return l; ++ } else if (yaffs_SumCompare(l->sum, sum) || l->hdrChunk <= 0) { ++ /* LostnFound chunk called Objxxx ++ * Do a real check ++ */ ++ yaffs_GetObjectName(l, buffer, ++ YAFFS_MAX_NAME_LENGTH); ++ if (yaffs_strncmp(name, buffer, YAFFS_MAX_NAME_LENGTH) == 0) ++ return l; ++ } ++ } ++ } ++ ++ return NULL; ++} ++ ++ ++#if 0 ++int yaffs_ApplyToDirectoryChildren(yaffs_Object *theDir, ++ int (*fn) (yaffs_Object *)) ++{ ++ struct ylist_head *i; ++ yaffs_Object *l; ++ ++ if (!theDir) { ++ T(YAFFS_TRACE_ALWAYS, ++ (TSTR ++ ("tragedy: yaffs_FindObjectByName: null pointer directory" ++ TENDSTR))); ++ YBUG(); ++ return YAFFS_FAIL; ++ } ++ if (theDir->variantType != YAFFS_OBJECT_TYPE_DIRECTORY) { ++ T(YAFFS_TRACE_ALWAYS, ++ (TSTR ++ ("tragedy: yaffs_FindObjectByName: non-directory" TENDSTR))); ++ YBUG(); ++ return YAFFS_FAIL; ++ } ++ ++ ylist_for_each(i, &theDir->variant.directoryVariant.children) { ++ if (i) { ++ l = ylist_entry(i, yaffs_Object, siblings); ++ if (l && !fn(l)) ++ return YAFFS_FAIL; ++ } ++ } ++ ++ return YAFFS_OK; ++ ++} ++#endif ++ ++/* GetEquivalentObject dereferences any hard links to get to the ++ * actual object. ++ */ ++ ++yaffs_Object *yaffs_GetEquivalentObject(yaffs_Object *obj) ++{ ++ if (obj && obj->variantType == YAFFS_OBJECT_TYPE_HARDLINK) { ++ /* We want the object id of the equivalent object, not this one */ ++ obj = obj->variant.hardLinkVariant.equivalentObject; ++ yaffs_CheckObjectDetailsLoaded(obj); ++ } ++ return obj; ++} ++ ++int yaffs_GetObjectName(yaffs_Object *obj, YCHAR *name, int buffSize) ++{ ++ memset(name, 0, buffSize * sizeof(YCHAR)); ++ ++ yaffs_CheckObjectDetailsLoaded(obj); ++ ++ if (obj->objectId == YAFFS_OBJECTID_LOSTNFOUND) { ++ yaffs_strncpy(name, YAFFS_LOSTNFOUND_NAME, buffSize - 1); ++ } else if (obj->hdrChunk <= 0) { ++ YCHAR locName[20]; ++ YCHAR numString[20]; ++ YCHAR *x = &numString[19]; ++ unsigned v = obj->objectId; ++ numString[19] = 0; ++ while (v > 0) { ++ x--; ++ *x = '0' + (v % 10); ++ v /= 10; ++ } ++ /* make up a name */ ++ yaffs_strcpy(locName, YAFFS_LOSTNFOUND_PREFIX); ++ yaffs_strcat(locName, x); ++ yaffs_strncpy(name, locName, buffSize - 1); ++ ++ } ++#ifdef CONFIG_YAFFS_SHORT_NAMES_IN_RAM ++ else if (obj->shortName[0]) ++ yaffs_strcpy(name, obj->shortName); ++#endif ++ else { ++ int result; ++ __u8 *buffer = yaffs_GetTempBuffer(obj->myDev, __LINE__); ++ ++ yaffs_ObjectHeader *oh = (yaffs_ObjectHeader *) buffer; ++ ++ memset(buffer, 0, obj->myDev->nDataBytesPerChunk); ++ ++ if (obj->hdrChunk > 0) { ++ result = yaffs_ReadChunkWithTagsFromNAND(obj->myDev, ++ obj->hdrChunk, buffer, ++ NULL); ++ } ++ yaffs_strncpy(name, oh->name, buffSize - 1); ++ ++ yaffs_ReleaseTempBuffer(obj->myDev, buffer, __LINE__); ++ } ++ ++ return yaffs_strlen(name); ++} ++ ++int yaffs_GetObjectFileLength(yaffs_Object *obj) ++{ ++ /* Dereference any hard linking */ ++ obj = yaffs_GetEquivalentObject(obj); ++ ++ if (obj->variantType == YAFFS_OBJECT_TYPE_FILE) ++ return obj->variant.fileVariant.fileSize; ++ if (obj->variantType == YAFFS_OBJECT_TYPE_SYMLINK) ++ return yaffs_strlen(obj->variant.symLinkVariant.alias); ++ else { ++ /* Only a directory should drop through to here */ ++ return obj->myDev->nDataBytesPerChunk; ++ } ++} ++ ++int yaffs_GetObjectLinkCount(yaffs_Object *obj) ++{ ++ int count = 0; ++ struct ylist_head *i; ++ ++ if (!obj->unlinked) ++ count++; /* the object itself */ ++ ++ ylist_for_each(i, &obj->hardLinks) ++ count++; /* add the hard links; */ ++ ++ return count; ++} ++ ++int yaffs_GetObjectInode(yaffs_Object *obj) ++{ ++ obj = yaffs_GetEquivalentObject(obj); ++ ++ return obj->objectId; ++} ++ ++unsigned yaffs_GetObjectType(yaffs_Object *obj) ++{ ++ obj = yaffs_GetEquivalentObject(obj); ++ ++ switch (obj->variantType) { ++ case YAFFS_OBJECT_TYPE_FILE: ++ return DT_REG; ++ break; ++ case YAFFS_OBJECT_TYPE_DIRECTORY: ++ return DT_DIR; ++ break; ++ case YAFFS_OBJECT_TYPE_SYMLINK: ++ return DT_LNK; ++ break; ++ case YAFFS_OBJECT_TYPE_HARDLINK: ++ return DT_REG; ++ break; ++ case YAFFS_OBJECT_TYPE_SPECIAL: ++ if (S_ISFIFO(obj->yst_mode)) ++ return DT_FIFO; ++ if (S_ISCHR(obj->yst_mode)) ++ return DT_CHR; ++ if (S_ISBLK(obj->yst_mode)) ++ return DT_BLK; ++ if (S_ISSOCK(obj->yst_mode)) ++ return DT_SOCK; ++ default: ++ return DT_REG; ++ break; ++ } ++} ++ ++YCHAR *yaffs_GetSymlinkAlias(yaffs_Object *obj) ++{ ++ obj = yaffs_GetEquivalentObject(obj); ++ if (obj->variantType == YAFFS_OBJECT_TYPE_SYMLINK) ++ return yaffs_CloneString(obj->variant.symLinkVariant.alias); ++ else ++ return yaffs_CloneString(_Y("")); ++} ++ ++#ifndef CONFIG_YAFFS_WINCE ++ ++int yaffs_SetAttributes(yaffs_Object *obj, struct iattr *attr) ++{ ++ unsigned int valid = attr->ia_valid; ++ ++ if (valid & ATTR_MODE) ++ obj->yst_mode = attr->ia_mode; ++ if (valid & ATTR_UID) ++ obj->yst_uid = attr->ia_uid; ++ if (valid & ATTR_GID) ++ obj->yst_gid = attr->ia_gid; ++ ++ if (valid & ATTR_ATIME) ++ obj->yst_atime = Y_TIME_CONVERT(attr->ia_atime); ++ if (valid & ATTR_CTIME) ++ obj->yst_ctime = Y_TIME_CONVERT(attr->ia_ctime); ++ if (valid & ATTR_MTIME) ++ obj->yst_mtime = Y_TIME_CONVERT(attr->ia_mtime); ++ ++ if (valid & ATTR_SIZE) ++ yaffs_ResizeFile(obj, attr->ia_size); ++ ++ yaffs_UpdateObjectHeader(obj, NULL, 1, 0, 0); ++ ++ return YAFFS_OK; ++ ++} ++int yaffs_GetAttributes(yaffs_Object *obj, struct iattr *attr) ++{ ++ unsigned int valid = 0; ++ ++ attr->ia_mode = obj->yst_mode; ++ valid |= ATTR_MODE; ++ attr->ia_uid = obj->yst_uid; ++ valid |= ATTR_UID; ++ attr->ia_gid = obj->yst_gid; ++ valid |= ATTR_GID; ++ ++ Y_TIME_CONVERT(attr->ia_atime) = obj->yst_atime; ++ valid |= ATTR_ATIME; ++ Y_TIME_CONVERT(attr->ia_ctime) = obj->yst_ctime; ++ valid |= ATTR_CTIME; ++ Y_TIME_CONVERT(attr->ia_mtime) = obj->yst_mtime; ++ valid |= ATTR_MTIME; ++ ++ attr->ia_size = yaffs_GetFileSize(obj); ++ valid |= ATTR_SIZE; ++ ++ attr->ia_valid = valid; ++ ++ return YAFFS_OK; ++} ++ ++#endif ++ ++#if 0 ++int yaffs_DumpObject(yaffs_Object *obj) ++{ ++ YCHAR name[257]; ++ ++ yaffs_GetObjectName(obj, name, 256); ++ ++ T(YAFFS_TRACE_ALWAYS, ++ (TSTR ++ ("Object %d, inode %d \"%s\"\n dirty %d valid %d serial %d sum %d" ++ " chunk %d type %d size %d\n" ++ TENDSTR), obj->objectId, yaffs_GetObjectInode(obj), name, ++ obj->dirty, obj->valid, obj->serial, obj->sum, obj->hdrChunk, ++ yaffs_GetObjectType(obj), yaffs_GetObjectFileLength(obj))); ++ ++ return YAFFS_OK; ++} ++#endif ++ ++/*---------------------------- Initialisation code -------------------------------------- */ ++ ++static int yaffs_CheckDevFunctions(const yaffs_Device *dev) ++{ ++ ++ /* Common functions, gotta have */ ++ if (!dev->eraseBlockInNAND || !dev->initialiseNAND) ++ return 0; ++ ++#ifdef CONFIG_YAFFS_YAFFS2 ++ ++ /* Can use the "with tags" style interface for yaffs1 or yaffs2 */ ++ if (dev->writeChunkWithTagsToNAND && ++ dev->readChunkWithTagsFromNAND && ++ !dev->writeChunkToNAND && ++ !dev->readChunkFromNAND && ++ dev->markNANDBlockBad && dev->queryNANDBlock) ++ return 1; ++#endif ++ ++ /* Can use the "spare" style interface for yaffs1 */ ++ if (!dev->isYaffs2 && ++ !dev->writeChunkWithTagsToNAND && ++ !dev->readChunkWithTagsFromNAND && ++ dev->writeChunkToNAND && ++ dev->readChunkFromNAND && ++ !dev->markNANDBlockBad && !dev->queryNANDBlock) ++ return 1; ++ ++ return 0; /* bad */ ++} ++ ++ ++static int yaffs_CreateInitialDirectories(yaffs_Device *dev) ++{ ++ /* Initialise the unlinked, deleted, root and lost and found directories */ ++ ++ dev->lostNFoundDir = dev->rootDir = NULL; ++ dev->unlinkedDir = dev->deletedDir = NULL; ++ ++ dev->unlinkedDir = ++ yaffs_CreateFakeDirectory(dev, YAFFS_OBJECTID_UNLINKED, S_IFDIR); ++ ++ dev->deletedDir = ++ yaffs_CreateFakeDirectory(dev, YAFFS_OBJECTID_DELETED, S_IFDIR); ++ ++ dev->rootDir = ++ yaffs_CreateFakeDirectory(dev, YAFFS_OBJECTID_ROOT, ++ YAFFS_ROOT_MODE | S_IFDIR); ++ dev->lostNFoundDir = ++ yaffs_CreateFakeDirectory(dev, YAFFS_OBJECTID_LOSTNFOUND, ++ YAFFS_LOSTNFOUND_MODE | S_IFDIR); ++ ++ if (dev->lostNFoundDir && dev->rootDir && dev->unlinkedDir && dev->deletedDir) { ++ yaffs_AddObjectToDirectory(dev->rootDir, dev->lostNFoundDir); ++ return YAFFS_OK; ++ } ++ ++ return YAFFS_FAIL; ++} ++ ++int yaffs_GutsInitialise(yaffs_Device *dev) ++{ ++ int init_failed = 0; ++ unsigned x; ++ int bits; ++ ++ T(YAFFS_TRACE_TRACING, (TSTR("yaffs: yaffs_GutsInitialise()" TENDSTR))); ++ ++ /* Check stuff that must be set */ ++ ++ if (!dev) { ++ T(YAFFS_TRACE_ALWAYS, (TSTR("yaffs: Need a device" TENDSTR))); ++ return YAFFS_FAIL; ++ } ++ ++ dev->internalStartBlock = dev->startBlock; ++ dev->internalEndBlock = dev->endBlock; ++ dev->blockOffset = 0; ++ dev->chunkOffset = 0; ++ dev->nFreeChunks = 0; ++ ++ dev->gcBlock = -1; ++ ++ if (dev->startBlock == 0) { ++ dev->internalStartBlock = dev->startBlock + 1; ++ dev->internalEndBlock = dev->endBlock + 1; ++ dev->blockOffset = 1; ++ dev->chunkOffset = dev->nChunksPerBlock; ++ } ++ ++ /* Check geometry parameters. */ ++ ++ if ((!dev->inbandTags && dev->isYaffs2 && dev->totalBytesPerChunk < 1024) || ++ (!dev->isYaffs2 && dev->totalBytesPerChunk < 512) || ++ (dev->inbandTags && !dev->isYaffs2) || ++ dev->nChunksPerBlock < 2 || ++ dev->nReservedBlocks < 2 || ++ dev->internalStartBlock <= 0 || ++ dev->internalEndBlock <= 0 || ++ dev->internalEndBlock <= (dev->internalStartBlock + dev->nReservedBlocks + 2)) { /* otherwise it is too small */ ++ T(YAFFS_TRACE_ALWAYS, ++ (TSTR ++ ("yaffs: NAND geometry problems: chunk size %d, type is yaffs%s, inbandTags %d " ++ TENDSTR), dev->totalBytesPerChunk, dev->isYaffs2 ? "2" : "", dev->inbandTags)); ++ return YAFFS_FAIL; ++ } ++ ++ if (yaffs_InitialiseNAND(dev) != YAFFS_OK) { ++ T(YAFFS_TRACE_ALWAYS, ++ (TSTR("yaffs: InitialiseNAND failed" TENDSTR))); ++ return YAFFS_FAIL; ++ } ++ ++ /* Sort out space for inband tags, if required */ ++ if (dev->inbandTags) ++ dev->nDataBytesPerChunk = dev->totalBytesPerChunk - sizeof(yaffs_PackedTags2TagsPart); ++ else ++ dev->nDataBytesPerChunk = dev->totalBytesPerChunk; ++ ++ /* Got the right mix of functions? */ ++ if (!yaffs_CheckDevFunctions(dev)) { ++ /* Function missing */ ++ T(YAFFS_TRACE_ALWAYS, ++ (TSTR ++ ("yaffs: device function(s) missing or wrong\n" TENDSTR))); ++ ++ return YAFFS_FAIL; ++ } ++ ++ /* This is really a compilation check. */ ++ if (!yaffs_CheckStructures()) { ++ T(YAFFS_TRACE_ALWAYS, ++ (TSTR("yaffs_CheckStructures failed\n" TENDSTR))); ++ return YAFFS_FAIL; ++ } ++ ++ if (dev->isMounted) { ++ T(YAFFS_TRACE_ALWAYS, ++ (TSTR("yaffs: device already mounted\n" TENDSTR))); ++ return YAFFS_FAIL; ++ } ++ ++ /* Finished with most checks. One or two more checks happen later on too. */ ++ ++ dev->isMounted = 1; ++ ++ /* OK now calculate a few things for the device */ ++ ++ /* ++ * Calculate all the chunk size manipulation numbers: ++ */ ++ x = dev->nDataBytesPerChunk; ++ /* We always use dev->chunkShift and dev->chunkDiv */ ++ dev->chunkShift = Shifts(x); ++ x >>= dev->chunkShift; ++ dev->chunkDiv = x; ++ /* We only use chunk mask if chunkDiv is 1 */ ++ dev->chunkMask = (1<<dev->chunkShift) - 1; ++ ++ /* ++ * Calculate chunkGroupBits. ++ * We need to find the next power of 2 > than internalEndBlock ++ */ ++ ++ x = dev->nChunksPerBlock * (dev->internalEndBlock + 1); ++ ++ bits = ShiftsGE(x); ++ ++ /* Set up tnode width if wide tnodes are enabled. */ ++ if (!dev->wideTnodesDisabled) { ++ /* bits must be even so that we end up with 32-bit words */ ++ if (bits & 1) ++ bits++; ++ if (bits < 16) ++ dev->tnodeWidth = 16; ++ else ++ dev->tnodeWidth = bits; ++ } else ++ dev->tnodeWidth = 16; ++ ++ dev->tnodeMask = (1<<dev->tnodeWidth)-1; ++ ++ /* Level0 Tnodes are 16 bits or wider (if wide tnodes are enabled), ++ * so if the bitwidth of the ++ * chunk range we're using is greater than 16 we need ++ * to figure out chunk shift and chunkGroupSize ++ */ ++ ++ if (bits <= dev->tnodeWidth) ++ dev->chunkGroupBits = 0; ++ else ++ dev->chunkGroupBits = bits - dev->tnodeWidth; ++ ++ ++ dev->chunkGroupSize = 1 << dev->chunkGroupBits; ++ ++ if (dev->nChunksPerBlock < dev->chunkGroupSize) { ++ /* We have a problem because the soft delete won't work if ++ * the chunk group size > chunks per block. ++ * This can be remedied by using larger "virtual blocks". ++ */ ++ T(YAFFS_TRACE_ALWAYS, ++ (TSTR("yaffs: chunk group too large\n" TENDSTR))); ++ ++ return YAFFS_FAIL; ++ } ++ ++ /* OK, we've finished verifying the device, lets continue with initialisation */ ++ ++ /* More device initialisation */ ++ dev->garbageCollections = 0; ++ dev->passiveGarbageCollections = 0; ++ dev->currentDirtyChecker = 0; ++ dev->bufferedBlock = -1; ++ dev->doingBufferedBlockRewrite = 0; ++ dev->nDeletedFiles = 0; ++ dev->nBackgroundDeletions = 0; ++ dev->nUnlinkedFiles = 0; ++ dev->eccFixed = 0; ++ dev->eccUnfixed = 0; ++ dev->tagsEccFixed = 0; ++ dev->tagsEccUnfixed = 0; ++ dev->nErasureFailures = 0; ++ dev->nErasedBlocks = 0; ++ dev->isDoingGC = 0; ++ dev->hasPendingPrioritisedGCs = 1; /* Assume the worst for now, will get fixed on first GC */ ++ ++ /* Initialise temporary buffers and caches. */ ++ if (!yaffs_InitialiseTempBuffers(dev)) ++ init_failed = 1; ++ ++ dev->srCache = NULL; ++ dev->gcCleanupList = NULL; ++ ++ ++ if (!init_failed && ++ dev->nShortOpCaches > 0) { ++ int i; ++ void *buf; ++ int srCacheBytes = dev->nShortOpCaches * sizeof(yaffs_ChunkCache); ++ ++ if (dev->nShortOpCaches > YAFFS_MAX_SHORT_OP_CACHES) ++ dev->nShortOpCaches = YAFFS_MAX_SHORT_OP_CACHES; ++ ++ dev->srCache = YMALLOC(srCacheBytes); ++ ++ buf = (__u8 *) dev->srCache; ++ ++ if (dev->srCache) ++ memset(dev->srCache, 0, srCacheBytes); ++ ++ for (i = 0; i < dev->nShortOpCaches && buf; i++) { ++ dev->srCache[i].object = NULL; ++ dev->srCache[i].lastUse = 0; ++ dev->srCache[i].dirty = 0; ++ dev->srCache[i].data = buf = YMALLOC_DMA(dev->totalBytesPerChunk); ++ } ++ if (!buf) ++ init_failed = 1; ++ ++ dev->srLastUse = 0; ++ } ++ ++ dev->cacheHits = 0; ++ ++ if (!init_failed) { ++ dev->gcCleanupList = YMALLOC(dev->nChunksPerBlock * sizeof(__u32)); ++ if (!dev->gcCleanupList) ++ init_failed = 1; ++ } ++ ++ if (dev->isYaffs2) ++ dev->useHeaderFileSize = 1; ++ ++ if (!init_failed && !yaffs_InitialiseBlocks(dev)) ++ init_failed = 1; ++ ++ yaffs_InitialiseTnodes(dev); ++ yaffs_InitialiseObjects(dev); ++ ++ if (!init_failed && !yaffs_CreateInitialDirectories(dev)) ++ init_failed = 1; ++ ++ ++ if (!init_failed) { ++ /* Now scan the flash. */ ++ if (dev->isYaffs2) { ++ if (yaffs_CheckpointRestore(dev)) { ++ yaffs_CheckObjectDetailsLoaded(dev->rootDir); ++ T(YAFFS_TRACE_ALWAYS, ++ (TSTR("yaffs: restored from checkpoint" TENDSTR))); ++ } else { ++ ++ /* Clean up the mess caused by an aborted checkpoint load ++ * and scan backwards. ++ */ ++ yaffs_DeinitialiseBlocks(dev); ++ yaffs_DeinitialiseTnodes(dev); ++ yaffs_DeinitialiseObjects(dev); ++ ++ ++ dev->nErasedBlocks = 0; ++ dev->nFreeChunks = 0; ++ dev->allocationBlock = -1; ++ dev->allocationPage = -1; ++ dev->nDeletedFiles = 0; ++ dev->nUnlinkedFiles = 0; ++ dev->nBackgroundDeletions = 0; ++ dev->oldestDirtySequence = 0; ++ ++ if (!init_failed && !yaffs_InitialiseBlocks(dev)) ++ init_failed = 1; ++ ++ yaffs_InitialiseTnodes(dev); ++ yaffs_InitialiseObjects(dev); ++ ++ if (!init_failed && !yaffs_CreateInitialDirectories(dev)) ++ init_failed = 1; ++ ++ if (!init_failed && !yaffs_ScanBackwards(dev)) ++ init_failed = 1; ++ } ++ } else if (!yaffs_Scan(dev)) ++ init_failed = 1; ++ ++ yaffs_StripDeletedObjects(dev); ++ } ++ ++ if (init_failed) { ++ /* Clean up the mess */ ++ T(YAFFS_TRACE_TRACING, ++ (TSTR("yaffs: yaffs_GutsInitialise() aborted.\n" TENDSTR))); ++ ++ yaffs_Deinitialise(dev); ++ return YAFFS_FAIL; ++ } ++ ++ /* Zero out stats */ ++ dev->nPageReads = 0; ++ dev->nPageWrites = 0; ++ dev->nBlockErasures = 0; ++ dev->nGCCopies = 0; ++ dev->nRetriedWrites = 0; ++ ++ dev->nRetiredBlocks = 0; ++ ++ yaffs_VerifyFreeChunks(dev); ++ yaffs_VerifyBlocks(dev); ++ ++ ++ T(YAFFS_TRACE_TRACING, ++ (TSTR("yaffs: yaffs_GutsInitialise() done.\n" TENDSTR))); ++ return YAFFS_OK; ++ ++} ++ ++void yaffs_Deinitialise(yaffs_Device *dev) ++{ ++ if (dev->isMounted) { ++ int i; ++ ++ yaffs_DeinitialiseBlocks(dev); ++ yaffs_DeinitialiseTnodes(dev); ++ yaffs_DeinitialiseObjects(dev); ++ if (dev->nShortOpCaches > 0 && ++ dev->srCache) { ++ ++ for (i = 0; i < dev->nShortOpCaches; i++) { ++ if (dev->srCache[i].data) ++ YFREE(dev->srCache[i].data); ++ dev->srCache[i].data = NULL; ++ } ++ ++ YFREE(dev->srCache); ++ dev->srCache = NULL; ++ } ++ ++ YFREE(dev->gcCleanupList); ++ ++ for (i = 0; i < YAFFS_N_TEMP_BUFFERS; i++) ++ YFREE(dev->tempBuffer[i].buffer); ++ ++ dev->isMounted = 0; ++ ++ if (dev->deinitialiseNAND) ++ dev->deinitialiseNAND(dev); ++ } ++} ++ ++static int yaffs_CountFreeChunks(yaffs_Device *dev) ++{ ++ int nFree; ++ int b; ++ ++ yaffs_BlockInfo *blk; ++ ++ for (nFree = 0, b = dev->internalStartBlock; b <= dev->internalEndBlock; ++ b++) { ++ blk = yaffs_GetBlockInfo(dev, b); ++ ++ switch (blk->blockState) { ++ case YAFFS_BLOCK_STATE_EMPTY: ++ case YAFFS_BLOCK_STATE_ALLOCATING: ++ case YAFFS_BLOCK_STATE_COLLECTING: ++ case YAFFS_BLOCK_STATE_FULL: ++ nFree += ++ (dev->nChunksPerBlock - blk->pagesInUse + ++ blk->softDeletions); ++ break; ++ default: ++ break; ++ } ++ } ++ ++ return nFree; ++} ++ ++int yaffs_GetNumberOfFreeChunks(yaffs_Device *dev) ++{ ++ /* This is what we report to the outside world */ ++ ++ int nFree; ++ int nDirtyCacheChunks; ++ int blocksForCheckpoint; ++ int i; ++ ++#if 1 ++ nFree = dev->nFreeChunks; ++#else ++ nFree = yaffs_CountFreeChunks(dev); ++#endif ++ ++ nFree += dev->nDeletedFiles; ++ ++ /* Now count the number of dirty chunks in the cache and subtract those */ ++ ++ for (nDirtyCacheChunks = 0, i = 0; i < dev->nShortOpCaches; i++) { ++ if (dev->srCache[i].dirty) ++ nDirtyCacheChunks++; ++ } ++ ++ nFree -= nDirtyCacheChunks; ++ ++ nFree -= ((dev->nReservedBlocks + 1) * dev->nChunksPerBlock); ++ ++ /* Now we figure out how much to reserve for the checkpoint and report that... */ ++ blocksForCheckpoint = yaffs_CalcCheckpointBlocksRequired(dev) - dev->blocksInCheckpoint; ++ if (blocksForCheckpoint < 0) ++ blocksForCheckpoint = 0; ++ ++ nFree -= (blocksForCheckpoint * dev->nChunksPerBlock); ++ ++ if (nFree < 0) ++ nFree = 0; ++ ++ return nFree; ++ ++} ++ ++static int yaffs_freeVerificationFailures; ++ ++static void yaffs_VerifyFreeChunks(yaffs_Device *dev) ++{ ++ int counted; ++ int difference; ++ ++ if (yaffs_SkipVerification(dev)) ++ return; ++ ++ counted = yaffs_CountFreeChunks(dev); ++ ++ difference = dev->nFreeChunks - counted; ++ ++ if (difference) { ++ T(YAFFS_TRACE_ALWAYS, ++ (TSTR("Freechunks verification failure %d %d %d" TENDSTR), ++ dev->nFreeChunks, counted, difference)); ++ yaffs_freeVerificationFailures++; ++ } ++} ++ ++/*---------------------------------------- YAFFS test code ----------------------*/ ++ ++#define yaffs_CheckStruct(structure, syze, name) \ ++ do { \ ++ if (sizeof(structure) != syze) { \ ++ T(YAFFS_TRACE_ALWAYS, (TSTR("%s should be %d but is %d\n" TENDSTR),\ ++ name, syze, sizeof(structure))); \ ++ return YAFFS_FAIL; \ ++ } \ ++ } while (0) ++ ++static int yaffs_CheckStructures(void) ++{ ++/* yaffs_CheckStruct(yaffs_Tags,8,"yaffs_Tags"); */ ++/* yaffs_CheckStruct(yaffs_TagsUnion,8,"yaffs_TagsUnion"); */ ++/* yaffs_CheckStruct(yaffs_Spare,16,"yaffs_Spare"); */ ++#ifndef CONFIG_YAFFS_TNODE_LIST_DEBUG ++ yaffs_CheckStruct(yaffs_Tnode, 2 * YAFFS_NTNODES_LEVEL0, "yaffs_Tnode"); ++#endif ++#ifndef CONFIG_YAFFS_WINCE ++ yaffs_CheckStruct(yaffs_ObjectHeader, 512, "yaffs_ObjectHeader"); ++#endif ++ return YAFFS_OK; ++} +diff -Nur linux-2.6.30.orig/fs/yaffs2/yaffs_guts.h linux-2.6.30/fs/yaffs2/yaffs_guts.h +--- linux-2.6.30.orig/fs/yaffs2/yaffs_guts.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/yaffs_guts.h 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,904 @@ ++/* ++ * YAFFS: Yet another Flash File System . A NAND-flash specific file system. ++ * ++ * Copyright (C) 2002-2007 Aleph One Ltd. ++ * for Toby Churchill Ltd and Brightstar Engineering ++ * ++ * Created by Charles Manning <charles@aleph1.co.uk> ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU Lesser General Public License version 2.1 as ++ * published by the Free Software Foundation. ++ * ++ * Note: Only YAFFS headers are LGPL, YAFFS C code is covered by GPL. ++ */ ++ ++#ifndef __YAFFS_GUTS_H__ ++#define __YAFFS_GUTS_H__ ++ ++#include "devextras.h" ++#include "yportenv.h" ++ ++#define YAFFS_OK 1 ++#define YAFFS_FAIL 0 ++ ++/* Give us a Y=0x59, ++ * Give us an A=0x41, ++ * Give us an FF=0xFF ++ * Give us an S=0x53 ++ * And what have we got... ++ */ ++#define YAFFS_MAGIC 0x5941FF53 ++ ++#define YAFFS_NTNODES_LEVEL0 16 ++#define YAFFS_TNODES_LEVEL0_BITS 4 ++#define YAFFS_TNODES_LEVEL0_MASK 0xf ++ ++#define YAFFS_NTNODES_INTERNAL (YAFFS_NTNODES_LEVEL0 / 2) ++#define YAFFS_TNODES_INTERNAL_BITS (YAFFS_TNODES_LEVEL0_BITS - 1) ++#define YAFFS_TNODES_INTERNAL_MASK 0x7 ++#define YAFFS_TNODES_MAX_LEVEL 6 ++ ++#ifndef CONFIG_YAFFS_NO_YAFFS1 ++#define YAFFS_BYTES_PER_SPARE 16 ++#define YAFFS_BYTES_PER_CHUNK 512 ++#define YAFFS_CHUNK_SIZE_SHIFT 9 ++#define YAFFS_CHUNKS_PER_BLOCK 32 ++#define YAFFS_BYTES_PER_BLOCK (YAFFS_CHUNKS_PER_BLOCK*YAFFS_BYTES_PER_CHUNK) ++#endif ++ ++#define YAFFS_MIN_YAFFS2_CHUNK_SIZE 1024 ++#define YAFFS_MIN_YAFFS2_SPARE_SIZE 32 ++ ++#define YAFFS_MAX_CHUNK_ID 0x000FFFFF ++ ++#define YAFFS_UNUSED_OBJECT_ID 0x0003FFFF ++ ++#define YAFFS_ALLOCATION_NOBJECTS 100 ++#define YAFFS_ALLOCATION_NTNODES 100 ++#define YAFFS_ALLOCATION_NLINKS 100 ++ ++#define YAFFS_NOBJECT_BUCKETS 256 ++ ++ ++#define YAFFS_OBJECT_SPACE 0x40000 ++ ++#define YAFFS_CHECKPOINT_VERSION 3 ++ ++#ifdef CONFIG_YAFFS_UNICODE ++#define YAFFS_MAX_NAME_LENGTH 127 ++#define YAFFS_MAX_ALIAS_LENGTH 79 ++#else ++#define YAFFS_MAX_NAME_LENGTH 255 ++#define YAFFS_MAX_ALIAS_LENGTH 159 ++#endif ++ ++#define YAFFS_SHORT_NAME_LENGTH 15 ++ ++/* Some special object ids for pseudo objects */ ++#define YAFFS_OBJECTID_ROOT 1 ++#define YAFFS_OBJECTID_LOSTNFOUND 2 ++#define YAFFS_OBJECTID_UNLINKED 3 ++#define YAFFS_OBJECTID_DELETED 4 ++ ++/* Sseudo object ids for checkpointing */ ++#define YAFFS_OBJECTID_SB_HEADER 0x10 ++#define YAFFS_OBJECTID_CHECKPOINT_DATA 0x20 ++#define YAFFS_SEQUENCE_CHECKPOINT_DATA 0x21 ++ ++/* */ ++ ++#define YAFFS_MAX_SHORT_OP_CACHES 20 ++ ++#define YAFFS_N_TEMP_BUFFERS 6 ++ ++/* We limit the number attempts at sucessfully saving a chunk of data. ++ * Small-page devices have 32 pages per block; large-page devices have 64. ++ * Default to something in the order of 5 to 10 blocks worth of chunks. ++ */ ++#define YAFFS_WR_ATTEMPTS (5*64) ++ ++/* Sequence numbers are used in YAFFS2 to determine block allocation order. ++ * The range is limited slightly to help distinguish bad numbers from good. ++ * This also allows us to perhaps in the future use special numbers for ++ * special purposes. ++ * EFFFFF00 allows the allocation of 8 blocks per second (~1Mbytes) for 15 years, ++ * and is a larger number than the lifetime of a 2GB device. ++ */ ++#define YAFFS_LOWEST_SEQUENCE_NUMBER 0x00001000 ++#define YAFFS_HIGHEST_SEQUENCE_NUMBER 0xEFFFFF00 ++ ++/* Special sequence number for bad block that failed to be marked bad */ ++#define YAFFS_SEQUENCE_BAD_BLOCK 0xFFFF0000 ++ ++/* ChunkCache is used for short read/write operations.*/ ++typedef struct { ++ struct yaffs_ObjectStruct *object; ++ int chunkId; ++ int lastUse; ++ int dirty; ++ int nBytes; /* Only valid if the cache is dirty */ ++ int locked; /* Can't push out or flush while locked. */ ++#ifdef CONFIG_YAFFS_YAFFS2 ++ __u8 *data; ++#else ++ __u8 data[YAFFS_BYTES_PER_CHUNK]; ++#endif ++} yaffs_ChunkCache; ++ ++ ++ ++/* Tags structures in RAM ++ * NB This uses bitfield. Bitfields should not straddle a u32 boundary otherwise ++ * the structure size will get blown out. ++ */ ++ ++#ifndef CONFIG_YAFFS_NO_YAFFS1 ++typedef struct { ++ unsigned chunkId:20; ++ unsigned serialNumber:2; ++ unsigned byteCountLSB:10; ++ unsigned objectId:18; ++ unsigned ecc:12; ++ unsigned byteCountMSB:2; ++} yaffs_Tags; ++ ++typedef union { ++ yaffs_Tags asTags; ++ __u8 asBytes[8]; ++} yaffs_TagsUnion; ++ ++#endif ++ ++/* Stuff used for extended tags in YAFFS2 */ ++ ++typedef enum { ++ YAFFS_ECC_RESULT_UNKNOWN, ++ YAFFS_ECC_RESULT_NO_ERROR, ++ YAFFS_ECC_RESULT_FIXED, ++ YAFFS_ECC_RESULT_UNFIXED ++} yaffs_ECCResult; ++ ++typedef enum { ++ YAFFS_OBJECT_TYPE_UNKNOWN, ++ YAFFS_OBJECT_TYPE_FILE, ++ YAFFS_OBJECT_TYPE_SYMLINK, ++ YAFFS_OBJECT_TYPE_DIRECTORY, ++ YAFFS_OBJECT_TYPE_HARDLINK, ++ YAFFS_OBJECT_TYPE_SPECIAL ++} yaffs_ObjectType; ++ ++#define YAFFS_OBJECT_TYPE_MAX YAFFS_OBJECT_TYPE_SPECIAL ++ ++typedef struct { ++ ++ unsigned validMarker0; ++ unsigned chunkUsed; /* Status of the chunk: used or unused */ ++ unsigned objectId; /* If 0 then this is not part of an object (unused) */ ++ unsigned chunkId; /* If 0 then this is a header, else a data chunk */ ++ unsigned byteCount; /* Only valid for data chunks */ ++ ++ /* The following stuff only has meaning when we read */ ++ yaffs_ECCResult eccResult; ++ unsigned blockBad; ++ ++ /* YAFFS 1 stuff */ ++ unsigned chunkDeleted; /* The chunk is marked deleted */ ++ unsigned serialNumber; /* Yaffs1 2-bit serial number */ ++ ++ /* YAFFS2 stuff */ ++ unsigned sequenceNumber; /* The sequence number of this block */ ++ ++ /* Extra info if this is an object header (YAFFS2 only) */ ++ ++ unsigned extraHeaderInfoAvailable; /* There is extra info available if this is not zero */ ++ unsigned extraParentObjectId; /* The parent object */ ++ unsigned extraIsShrinkHeader; /* Is it a shrink header? */ ++ unsigned extraShadows; /* Does this shadow another object? */ ++ ++ yaffs_ObjectType extraObjectType; /* What object type? */ ++ ++ unsigned extraFileLength; /* Length if it is a file */ ++ unsigned extraEquivalentObjectId; /* Equivalent object Id if it is a hard link */ ++ ++ unsigned validMarker1; ++ ++} yaffs_ExtendedTags; ++ ++/* Spare structure for YAFFS1 */ ++typedef struct { ++ __u8 tagByte0; ++ __u8 tagByte1; ++ __u8 tagByte2; ++ __u8 tagByte3; ++ __u8 pageStatus; /* set to 0 to delete the chunk */ ++ __u8 blockStatus; ++ __u8 tagByte4; ++ __u8 tagByte5; ++ __u8 ecc1[3]; ++ __u8 tagByte6; ++ __u8 tagByte7; ++ __u8 ecc2[3]; ++} yaffs_Spare; ++ ++/*Special structure for passing through to mtd */ ++struct yaffs_NANDSpare { ++ yaffs_Spare spare; ++ int eccres1; ++ int eccres2; ++}; ++ ++/* Block data in RAM */ ++ ++typedef enum { ++ YAFFS_BLOCK_STATE_UNKNOWN = 0, ++ ++ YAFFS_BLOCK_STATE_SCANNING, ++ YAFFS_BLOCK_STATE_NEEDS_SCANNING, ++ /* The block might have something on it (ie it is allocating or full, perhaps empty) ++ * but it needs to be scanned to determine its true state. ++ * This state is only valid during yaffs_Scan. ++ * NB We tolerate empty because the pre-scanner might be incapable of deciding ++ * However, if this state is returned on a YAFFS2 device, then we expect a sequence number ++ */ ++ ++ YAFFS_BLOCK_STATE_EMPTY, ++ /* This block is empty */ ++ ++ YAFFS_BLOCK_STATE_ALLOCATING, ++ /* This block is partially allocated. ++ * At least one page holds valid data. ++ * This is the one currently being used for page ++ * allocation. Should never be more than one of these ++ */ ++ ++ YAFFS_BLOCK_STATE_FULL, ++ /* All the pages in this block have been allocated. ++ */ ++ ++ YAFFS_BLOCK_STATE_DIRTY, ++ /* All pages have been allocated and deleted. ++ * Erase me, reuse me. ++ */ ++ ++ YAFFS_BLOCK_STATE_CHECKPOINT, ++ /* This block is assigned to holding checkpoint data. ++ */ ++ ++ YAFFS_BLOCK_STATE_COLLECTING, ++ /* This block is being garbage collected */ ++ ++ YAFFS_BLOCK_STATE_DEAD ++ /* This block has failed and is not in use */ ++} yaffs_BlockState; ++ ++#define YAFFS_NUMBER_OF_BLOCK_STATES (YAFFS_BLOCK_STATE_DEAD + 1) ++ ++ ++typedef struct { ++ ++ int softDeletions:10; /* number of soft deleted pages */ ++ int pagesInUse:10; /* number of pages in use */ ++ unsigned blockState:4; /* One of the above block states. NB use unsigned because enum is sometimes an int */ ++ __u32 needsRetiring:1; /* Data has failed on this block, need to get valid data off */ ++ /* and retire the block. */ ++ __u32 skipErasedCheck:1; /* If this is set we can skip the erased check on this block */ ++ __u32 gcPrioritise:1; /* An ECC check or blank check has failed on this block. ++ It should be prioritised for GC */ ++ __u32 chunkErrorStrikes:3; /* How many times we've had ecc etc failures on this block and tried to reuse it */ ++ ++#ifdef CONFIG_YAFFS_YAFFS2 ++ __u32 hasShrinkHeader:1; /* This block has at least one shrink object header */ ++ __u32 sequenceNumber; /* block sequence number for yaffs2 */ ++#endif ++ ++} yaffs_BlockInfo; ++ ++/* -------------------------- Object structure -------------------------------*/ ++/* This is the object structure as stored on NAND */ ++ ++typedef struct { ++ yaffs_ObjectType type; ++ ++ /* Apply to everything */ ++ int parentObjectId; ++ __u16 sum__NoLongerUsed; /* checksum of name. No longer used */ ++ YCHAR name[YAFFS_MAX_NAME_LENGTH + 1]; ++ ++ /* The following apply to directories, files, symlinks - not hard links */ ++ __u32 yst_mode; /* protection */ ++ ++#ifdef CONFIG_YAFFS_WINCE ++ __u32 notForWinCE[5]; ++#else ++ __u32 yst_uid; ++ __u32 yst_gid; ++ __u32 yst_atime; ++ __u32 yst_mtime; ++ __u32 yst_ctime; ++#endif ++ ++ /* File size applies to files only */ ++ int fileSize; ++ ++ /* Equivalent object id applies to hard links only. */ ++ int equivalentObjectId; ++ ++ /* Alias is for symlinks only. */ ++ YCHAR alias[YAFFS_MAX_ALIAS_LENGTH + 1]; ++ ++ __u32 yst_rdev; /* device stuff for block and char devices (major/min) */ ++ ++#ifdef CONFIG_YAFFS_WINCE ++ __u32 win_ctime[2]; ++ __u32 win_atime[2]; ++ __u32 win_mtime[2]; ++#else ++ __u32 roomToGrow[6]; ++ ++#endif ++ __u32 inbandShadowsObject; ++ __u32 inbandIsShrink; ++ ++ __u32 reservedSpace[2]; ++ int shadowsObject; /* This object header shadows the specified object if > 0 */ ++ ++ /* isShrink applies to object headers written when we shrink the file (ie resize) */ ++ __u32 isShrink; ++ ++} yaffs_ObjectHeader; ++ ++/*--------------------------- Tnode -------------------------- */ ++ ++union yaffs_Tnode_union { ++#ifdef CONFIG_YAFFS_TNODE_LIST_DEBUG ++ union yaffs_Tnode_union *internal[YAFFS_NTNODES_INTERNAL + 1]; ++#else ++ union yaffs_Tnode_union *internal[YAFFS_NTNODES_INTERNAL]; ++#endif ++/* __u16 level0[YAFFS_NTNODES_LEVEL0]; */ ++ ++}; ++ ++typedef union yaffs_Tnode_union yaffs_Tnode; ++ ++struct yaffs_TnodeList_struct { ++ struct yaffs_TnodeList_struct *next; ++ yaffs_Tnode *tnodes; ++}; ++ ++typedef struct yaffs_TnodeList_struct yaffs_TnodeList; ++ ++/*------------------------ Object -----------------------------*/ ++/* An object can be one of: ++ * - a directory (no data, has children links ++ * - a regular file (data.... not prunes :->). ++ * - a symlink [symbolic link] (the alias). ++ * - a hard link ++ */ ++ ++typedef struct { ++ __u32 fileSize; ++ __u32 scannedFileSize; ++ __u32 shrinkSize; ++ int topLevel; ++ yaffs_Tnode *top; ++} yaffs_FileStructure; ++ ++typedef struct { ++ struct ylist_head children; /* list of child links */ ++} yaffs_DirectoryStructure; ++ ++typedef struct { ++ YCHAR *alias; ++} yaffs_SymLinkStructure; ++ ++typedef struct { ++ struct yaffs_ObjectStruct *equivalentObject; ++ __u32 equivalentObjectId; ++} yaffs_HardLinkStructure; ++ ++typedef union { ++ yaffs_FileStructure fileVariant; ++ yaffs_DirectoryStructure directoryVariant; ++ yaffs_SymLinkStructure symLinkVariant; ++ yaffs_HardLinkStructure hardLinkVariant; ++} yaffs_ObjectVariant; ++ ++struct yaffs_ObjectStruct { ++ __u8 deleted:1; /* This should only apply to unlinked files. */ ++ __u8 softDeleted:1; /* it has also been soft deleted */ ++ __u8 unlinked:1; /* An unlinked file. The file should be in the unlinked directory.*/ ++ __u8 fake:1; /* A fake object has no presence on NAND. */ ++ __u8 renameAllowed:1; /* Some objects are not allowed to be renamed. */ ++ __u8 unlinkAllowed:1; ++ __u8 dirty:1; /* the object needs to be written to flash */ ++ __u8 valid:1; /* When the file system is being loaded up, this ++ * object might be created before the data ++ * is available (ie. file data records appear before the header). ++ */ ++ __u8 lazyLoaded:1; /* This object has been lazy loaded and is missing some detail */ ++ ++ __u8 deferedFree:1; /* For Linux kernel. Object is removed from NAND, but is ++ * still in the inode cache. Free of object is defered. ++ * until the inode is released. ++ */ ++ __u8 beingCreated:1; /* This object is still being created so skip some checks. */ ++ ++ __u8 serial; /* serial number of chunk in NAND. Cached here */ ++ __u16 sum; /* sum of the name to speed searching */ ++ ++ struct yaffs_DeviceStruct *myDev; /* The device I'm on */ ++ ++ struct ylist_head hashLink; /* list of objects in this hash bucket */ ++ ++ struct ylist_head hardLinks; /* all the equivalent hard linked objects */ ++ ++ /* directory structure stuff */ ++ /* also used for linking up the free list */ ++ struct yaffs_ObjectStruct *parent; ++ struct ylist_head siblings; ++ ++ /* Where's my object header in NAND? */ ++ int hdrChunk; ++ ++ int nDataChunks; /* Number of data chunks attached to the file. */ ++ ++ __u32 objectId; /* the object id value */ ++ ++ __u32 yst_mode; ++ ++#ifdef CONFIG_YAFFS_SHORT_NAMES_IN_RAM ++ YCHAR shortName[YAFFS_SHORT_NAME_LENGTH + 1]; ++#endif ++ ++#ifndef __KERNEL__ ++ __u32 inUse; ++#endif ++ ++#ifdef CONFIG_YAFFS_WINCE ++ __u32 win_ctime[2]; ++ __u32 win_mtime[2]; ++ __u32 win_atime[2]; ++#else ++ __u32 yst_uid; ++ __u32 yst_gid; ++ __u32 yst_atime; ++ __u32 yst_mtime; ++ __u32 yst_ctime; ++#endif ++ ++ __u32 yst_rdev; ++ ++#ifdef __KERNEL__ ++ struct inode *myInode; ++ ++#endif ++ ++ yaffs_ObjectType variantType; ++ ++ yaffs_ObjectVariant variant; ++ ++}; ++ ++typedef struct yaffs_ObjectStruct yaffs_Object; ++ ++struct yaffs_ObjectList_struct { ++ yaffs_Object *objects; ++ struct yaffs_ObjectList_struct *next; ++}; ++ ++typedef struct yaffs_ObjectList_struct yaffs_ObjectList; ++ ++typedef struct { ++ struct ylist_head list; ++ int count; ++} yaffs_ObjectBucket; ++ ++ ++/* yaffs_CheckpointObject holds the definition of an object as dumped ++ * by checkpointing. ++ */ ++ ++typedef struct { ++ int structType; ++ __u32 objectId; ++ __u32 parentId; ++ int hdrChunk; ++ yaffs_ObjectType variantType:3; ++ __u8 deleted:1; ++ __u8 softDeleted:1; ++ __u8 unlinked:1; ++ __u8 fake:1; ++ __u8 renameAllowed:1; ++ __u8 unlinkAllowed:1; ++ __u8 serial; ++ ++ int nDataChunks; ++ __u32 fileSizeOrEquivalentObjectId; ++} yaffs_CheckpointObject; ++ ++/*--------------------- Temporary buffers ---------------- ++ * ++ * These are chunk-sized working buffers. Each device has a few ++ */ ++ ++typedef struct { ++ __u8 *buffer; ++ int line; /* track from whence this buffer was allocated */ ++ int maxLine; ++} yaffs_TempBuffer; ++ ++/*----------------- Device ---------------------------------*/ ++ ++struct yaffs_DeviceStruct { ++ struct ylist_head devList; ++ const char *name; ++ ++ /* Entry parameters set up way early. Yaffs sets up the rest.*/ ++ int nDataBytesPerChunk; /* Should be a power of 2 >= 512 */ ++ int nChunksPerBlock; /* does not need to be a power of 2 */ ++ int spareBytesPerChunk; /* spare area size */ ++ int startBlock; /* Start block we're allowed to use */ ++ int endBlock; /* End block we're allowed to use */ ++ int nReservedBlocks; /* We want this tuneable so that we can reduce */ ++ /* reserved blocks on NOR and RAM. */ ++ ++ ++ /* Stuff used by the shared space checkpointing mechanism */ ++ /* If this value is zero, then this mechanism is disabled */ ++ ++/* int nCheckpointReservedBlocks; */ /* Blocks to reserve for checkpoint data */ ++ ++ ++ int nShortOpCaches; /* If <= 0, then short op caching is disabled, else ++ * the number of short op caches (don't use too many) ++ */ ++ ++ int useHeaderFileSize; /* Flag to determine if we should use file sizes from the header */ ++ ++ int useNANDECC; /* Flag to decide whether or not to use NANDECC */ ++ ++ void *genericDevice; /* Pointer to device context ++ * On an mtd this holds the mtd pointer. ++ */ ++ void *superBlock; ++ ++ /* NAND access functions (Must be set before calling YAFFS)*/ ++ ++ int (*writeChunkToNAND) (struct yaffs_DeviceStruct *dev, ++ int chunkInNAND, const __u8 *data, ++ const yaffs_Spare *spare); ++ int (*readChunkFromNAND) (struct yaffs_DeviceStruct *dev, ++ int chunkInNAND, __u8 *data, ++ yaffs_Spare *spare); ++ int (*eraseBlockInNAND) (struct yaffs_DeviceStruct *dev, ++ int blockInNAND); ++ int (*initialiseNAND) (struct yaffs_DeviceStruct *dev); ++ int (*deinitialiseNAND) (struct yaffs_DeviceStruct *dev); ++ ++#ifdef CONFIG_YAFFS_YAFFS2 ++ int (*writeChunkWithTagsToNAND) (struct yaffs_DeviceStruct *dev, ++ int chunkInNAND, const __u8 *data, ++ const yaffs_ExtendedTags *tags); ++ int (*readChunkWithTagsFromNAND) (struct yaffs_DeviceStruct *dev, ++ int chunkInNAND, __u8 *data, ++ yaffs_ExtendedTags *tags); ++ int (*markNANDBlockBad) (struct yaffs_DeviceStruct *dev, int blockNo); ++ int (*queryNANDBlock) (struct yaffs_DeviceStruct *dev, int blockNo, ++ yaffs_BlockState *state, __u32 *sequenceNumber); ++#endif ++ ++ int isYaffs2; ++ ++ /* The removeObjectCallback function must be supplied by OS flavours that ++ * need it. The Linux kernel does not use this, but yaffs direct does use ++ * it to implement the faster readdir ++ */ ++ void (*removeObjectCallback)(struct yaffs_ObjectStruct *obj); ++ ++ /* Callback to mark the superblock dirsty */ ++ void (*markSuperBlockDirty)(void *superblock); ++ ++ int wideTnodesDisabled; /* Set to disable wide tnodes */ ++ ++ YCHAR *pathDividers; /* String of legal path dividers */ ++ ++ ++ /* End of stuff that must be set before initialisation. */ ++ ++ /* Checkpoint control. Can be set before or after initialisation */ ++ __u8 skipCheckpointRead; ++ __u8 skipCheckpointWrite; ++ ++ /* Runtime parameters. Set up by YAFFS. */ ++ ++ __u16 chunkGroupBits; /* 0 for devices <= 32MB. else log2(nchunks) - 16 */ ++ __u16 chunkGroupSize; /* == 2^^chunkGroupBits */ ++ ++ /* Stuff to support wide tnodes */ ++ __u32 tnodeWidth; ++ __u32 tnodeMask; ++ ++ /* Stuff for figuring out file offset to chunk conversions */ ++ __u32 chunkShift; /* Shift value */ ++ __u32 chunkDiv; /* Divisor after shifting: 1 for power-of-2 sizes */ ++ __u32 chunkMask; /* Mask to use for power-of-2 case */ ++ ++ /* Stuff to handle inband tags */ ++ int inbandTags; ++ __u32 totalBytesPerChunk; ++ ++#ifdef __KERNEL__ ++ ++ struct semaphore sem; /* Semaphore for waiting on erasure.*/ ++ struct semaphore grossLock; /* Gross locking semaphore */ ++ __u8 *spareBuffer; /* For mtdif2 use. Don't know the size of the buffer ++ * at compile time so we have to allocate it. ++ */ ++ void (*putSuperFunc) (struct super_block *sb); ++#endif ++ ++ int isMounted; ++ ++ int isCheckpointed; ++ ++ ++ /* Stuff to support block offsetting to support start block zero */ ++ int internalStartBlock; ++ int internalEndBlock; ++ int blockOffset; ++ int chunkOffset; ++ ++ ++ /* Runtime checkpointing stuff */ ++ int checkpointPageSequence; /* running sequence number of checkpoint pages */ ++ int checkpointByteCount; ++ int checkpointByteOffset; ++ __u8 *checkpointBuffer; ++ int checkpointOpenForWrite; ++ int blocksInCheckpoint; ++ int checkpointCurrentChunk; ++ int checkpointCurrentBlock; ++ int checkpointNextBlock; ++ int *checkpointBlockList; ++ int checkpointMaxBlocks; ++ __u32 checkpointSum; ++ __u32 checkpointXor; ++ ++ int nCheckpointBlocksRequired; /* Number of blocks needed to store current checkpoint set */ ++ ++ /* Block Info */ ++ yaffs_BlockInfo *blockInfo; ++ __u8 *chunkBits; /* bitmap of chunks in use */ ++ unsigned blockInfoAlt:1; /* was allocated using alternative strategy */ ++ unsigned chunkBitsAlt:1; /* was allocated using alternative strategy */ ++ int chunkBitmapStride; /* Number of bytes of chunkBits per block. ++ * Must be consistent with nChunksPerBlock. ++ */ ++ ++ int nErasedBlocks; ++ int allocationBlock; /* Current block being allocated off */ ++ __u32 allocationPage; ++ int allocationBlockFinder; /* Used to search for next allocation block */ ++ ++ /* Runtime state */ ++ int nTnodesCreated; ++ yaffs_Tnode *freeTnodes; ++ int nFreeTnodes; ++ yaffs_TnodeList *allocatedTnodeList; ++ ++ int isDoingGC; ++ int gcBlock; ++ int gcChunk; ++ ++ int nObjectsCreated; ++ yaffs_Object *freeObjects; ++ int nFreeObjects; ++ ++ int nHardLinks; ++ ++ yaffs_ObjectList *allocatedObjectList; ++ ++ yaffs_ObjectBucket objectBucket[YAFFS_NOBJECT_BUCKETS]; ++ ++ int nFreeChunks; ++ ++ int currentDirtyChecker; /* Used to find current dirtiest block */ ++ ++ __u32 *gcCleanupList; /* objects to delete at the end of a GC. */ ++ int nonAggressiveSkip; /* GC state/mode */ ++ ++ /* Statistcs */ ++ int nPageWrites; ++ int nPageReads; ++ int nBlockErasures; ++ int nErasureFailures; ++ int nGCCopies; ++ int garbageCollections; ++ int passiveGarbageCollections; ++ int nRetriedWrites; ++ int nRetiredBlocks; ++ int eccFixed; ++ int eccUnfixed; ++ int tagsEccFixed; ++ int tagsEccUnfixed; ++ int nDeletions; ++ int nUnmarkedDeletions; ++ ++ int hasPendingPrioritisedGCs; /* We think this device might have pending prioritised gcs */ ++ ++ /* Special directories */ ++ yaffs_Object *rootDir; ++ yaffs_Object *lostNFoundDir; ++ ++ /* Buffer areas for storing data to recover from write failures TODO ++ * __u8 bufferedData[YAFFS_CHUNKS_PER_BLOCK][YAFFS_BYTES_PER_CHUNK]; ++ * yaffs_Spare bufferedSpare[YAFFS_CHUNKS_PER_BLOCK]; ++ */ ++ ++ int bufferedBlock; /* Which block is buffered here? */ ++ int doingBufferedBlockRewrite; ++ ++ yaffs_ChunkCache *srCache; ++ int srLastUse; ++ ++ int cacheHits; ++ ++ /* Stuff for background deletion and unlinked files.*/ ++ yaffs_Object *unlinkedDir; /* Directory where unlinked and deleted files live. */ ++ yaffs_Object *deletedDir; /* Directory where deleted objects are sent to disappear. */ ++ yaffs_Object *unlinkedDeletion; /* Current file being background deleted.*/ ++ int nDeletedFiles; /* Count of files awaiting deletion;*/ ++ int nUnlinkedFiles; /* Count of unlinked files. */ ++ int nBackgroundDeletions; /* Count of background deletions. */ ++ ++ ++ /* Temporary buffer management */ ++ yaffs_TempBuffer tempBuffer[YAFFS_N_TEMP_BUFFERS]; ++ int maxTemp; ++ int tempInUse; ++ int unmanagedTempAllocations; ++ int unmanagedTempDeallocations; ++ ++ /* yaffs2 runtime stuff */ ++ unsigned sequenceNumber; /* Sequence number of currently allocating block */ ++ unsigned oldestDirtySequence; ++ ++}; ++ ++typedef struct yaffs_DeviceStruct yaffs_Device; ++ ++/* The static layout of block usage etc is stored in the super block header */ ++typedef struct { ++ int StructType; ++ int version; ++ int checkpointStartBlock; ++ int checkpointEndBlock; ++ int startBlock; ++ int endBlock; ++ int rfu[100]; ++} yaffs_SuperBlockHeader; ++ ++/* The CheckpointDevice structure holds the device information that changes at runtime and ++ * must be preserved over unmount/mount cycles. ++ */ ++typedef struct { ++ int structType; ++ int nErasedBlocks; ++ int allocationBlock; /* Current block being allocated off */ ++ __u32 allocationPage; ++ int nFreeChunks; ++ ++ int nDeletedFiles; /* Count of files awaiting deletion;*/ ++ int nUnlinkedFiles; /* Count of unlinked files. */ ++ int nBackgroundDeletions; /* Count of background deletions. */ ++ ++ /* yaffs2 runtime stuff */ ++ unsigned sequenceNumber; /* Sequence number of currently allocating block */ ++ unsigned oldestDirtySequence; ++ ++} yaffs_CheckpointDevice; ++ ++ ++typedef struct { ++ int structType; ++ __u32 magic; ++ __u32 version; ++ __u32 head; ++} yaffs_CheckpointValidity; ++ ++ ++/*----------------------- YAFFS Functions -----------------------*/ ++ ++int yaffs_GutsInitialise(yaffs_Device *dev); ++void yaffs_Deinitialise(yaffs_Device *dev); ++ ++int yaffs_GetNumberOfFreeChunks(yaffs_Device *dev); ++ ++int yaffs_RenameObject(yaffs_Object *oldDir, const YCHAR *oldName, ++ yaffs_Object *newDir, const YCHAR *newName); ++ ++int yaffs_Unlink(yaffs_Object *dir, const YCHAR *name); ++int yaffs_DeleteObject(yaffs_Object *obj); ++ ++int yaffs_GetObjectName(yaffs_Object *obj, YCHAR *name, int buffSize); ++int yaffs_GetObjectFileLength(yaffs_Object *obj); ++int yaffs_GetObjectInode(yaffs_Object *obj); ++unsigned yaffs_GetObjectType(yaffs_Object *obj); ++int yaffs_GetObjectLinkCount(yaffs_Object *obj); ++ ++int yaffs_SetAttributes(yaffs_Object *obj, struct iattr *attr); ++int yaffs_GetAttributes(yaffs_Object *obj, struct iattr *attr); ++ ++/* File operations */ ++int yaffs_ReadDataFromFile(yaffs_Object *obj, __u8 *buffer, loff_t offset, ++ int nBytes); ++int yaffs_WriteDataToFile(yaffs_Object *obj, const __u8 *buffer, loff_t offset, ++ int nBytes, int writeThrough); ++int yaffs_ResizeFile(yaffs_Object *obj, loff_t newSize); ++ ++yaffs_Object *yaffs_MknodFile(yaffs_Object *parent, const YCHAR *name, ++ __u32 mode, __u32 uid, __u32 gid); ++int yaffs_FlushFile(yaffs_Object *obj, int updateTime); ++ ++/* Flushing and checkpointing */ ++void yaffs_FlushEntireDeviceCache(yaffs_Device *dev); ++ ++int yaffs_CheckpointSave(yaffs_Device *dev); ++int yaffs_CheckpointRestore(yaffs_Device *dev); ++ ++/* Directory operations */ ++yaffs_Object *yaffs_MknodDirectory(yaffs_Object *parent, const YCHAR *name, ++ __u32 mode, __u32 uid, __u32 gid); ++yaffs_Object *yaffs_FindObjectByName(yaffs_Object *theDir, const YCHAR *name); ++int yaffs_ApplyToDirectoryChildren(yaffs_Object *theDir, ++ int (*fn) (yaffs_Object *)); ++ ++yaffs_Object *yaffs_FindObjectByNumber(yaffs_Device *dev, __u32 number); ++ ++/* Link operations */ ++yaffs_Object *yaffs_Link(yaffs_Object *parent, const YCHAR *name, ++ yaffs_Object *equivalentObject); ++ ++yaffs_Object *yaffs_GetEquivalentObject(yaffs_Object *obj); ++ ++/* Symlink operations */ ++yaffs_Object *yaffs_MknodSymLink(yaffs_Object *parent, const YCHAR *name, ++ __u32 mode, __u32 uid, __u32 gid, ++ const YCHAR *alias); ++YCHAR *yaffs_GetSymlinkAlias(yaffs_Object *obj); ++ ++/* Special inodes (fifos, sockets and devices) */ ++yaffs_Object *yaffs_MknodSpecial(yaffs_Object *parent, const YCHAR *name, ++ __u32 mode, __u32 uid, __u32 gid, __u32 rdev); ++ ++/* Special directories */ ++yaffs_Object *yaffs_Root(yaffs_Device *dev); ++yaffs_Object *yaffs_LostNFound(yaffs_Device *dev); ++ ++#ifdef CONFIG_YAFFS_WINCE ++/* CONFIG_YAFFS_WINCE special stuff */ ++void yfsd_WinFileTimeNow(__u32 target[2]); ++#endif ++ ++#ifdef __KERNEL__ ++ ++void yaffs_HandleDeferedFree(yaffs_Object *obj); ++#endif ++ ++/* Debug dump */ ++int yaffs_DumpObject(yaffs_Object *obj); ++ ++void yaffs_GutsTest(yaffs_Device *dev); ++ ++/* A few useful functions */ ++void yaffs_InitialiseTags(yaffs_ExtendedTags *tags); ++void yaffs_DeleteChunk(yaffs_Device *dev, int chunkId, int markNAND, int lyn); ++int yaffs_CheckFF(__u8 *buffer, int nBytes); ++void yaffs_HandleChunkError(yaffs_Device *dev, yaffs_BlockInfo *bi); ++ ++__u8 *yaffs_GetTempBuffer(yaffs_Device *dev, int lineNo); ++void yaffs_ReleaseTempBuffer(yaffs_Device *dev, __u8 *buffer, int lineNo); ++ ++#endif +diff -Nur linux-2.6.30.orig/fs/yaffs2/yaffsinterface.h linux-2.6.30/fs/yaffs2/yaffsinterface.h +--- linux-2.6.30.orig/fs/yaffs2/yaffsinterface.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/yaffsinterface.h 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,21 @@ ++/* ++ * YAFFS: Yet another Flash File System . A NAND-flash specific file system. ++ * ++ * Copyright (C) 2002-2007 Aleph One Ltd. ++ * for Toby Churchill Ltd and Brightstar Engineering ++ * ++ * Created by Charles Manning <charles@aleph1.co.uk> ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU Lesser General Public License version 2.1 as ++ * published by the Free Software Foundation. ++ * ++ * Note: Only YAFFS headers are LGPL, YAFFS C code is covered by GPL. ++ */ ++ ++#ifndef __YAFFSINTERFACE_H__ ++#define __YAFFSINTERFACE_H__ ++ ++int yaffs_Initialise(unsigned nBlocks); ++ ++#endif +diff -Nur linux-2.6.30.orig/fs/yaffs2/yaffs_mtdif1.c linux-2.6.30/fs/yaffs2/yaffs_mtdif1.c +--- linux-2.6.30.orig/fs/yaffs2/yaffs_mtdif1.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/yaffs_mtdif1.c 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,365 @@ ++/* ++ * YAFFS: Yet another FFS. A NAND-flash specific file system. ++ * yaffs_mtdif1.c NAND mtd interface functions for small-page NAND. ++ * ++ * Copyright (C) 2002 Aleph One Ltd. ++ * for Toby Churchill Ltd and Brightstar Engineering ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License version 2 as ++ * published by the Free Software Foundation. ++ */ ++ ++/* ++ * This module provides the interface between yaffs_nand.c and the ++ * MTD API. This version is used when the MTD interface supports the ++ * 'mtd_oob_ops' style calls to read_oob and write_oob, circa 2.6.17, ++ * and we have small-page NAND device. ++ * ++ * These functions are invoked via function pointers in yaffs_nand.c. ++ * This replaces functionality provided by functions in yaffs_mtdif.c ++ * and the yaffs_TagsCompatability functions in yaffs_tagscompat.c that are ++ * called in yaffs_mtdif.c when the function pointers are NULL. ++ * We assume the MTD layer is performing ECC (useNANDECC is true). ++ */ ++ ++#include "yportenv.h" ++#include "yaffs_guts.h" ++#include "yaffs_packedtags1.h" ++#include "yaffs_tagscompat.h" /* for yaffs_CalcTagsECC */ ++ ++#include "linux/kernel.h" ++#include "linux/version.h" ++#include "linux/types.h" ++#include "linux/mtd/mtd.h" ++ ++/* Don't compile this module if we don't have MTD's mtd_oob_ops interface */ ++#if (MTD_VERSION_CODE > MTD_VERSION(2, 6, 17)) ++ ++const char *yaffs_mtdif1_c_version = "$Id: yaffs_mtdif1.c,v 1.10 2009-03-09 07:41:10 charles Exp $"; ++ ++#ifndef CONFIG_YAFFS_9BYTE_TAGS ++# define YTAG1_SIZE 8 ++#else ++# define YTAG1_SIZE 9 ++#endif ++ ++#if 0 ++/* Use the following nand_ecclayout with MTD when using ++ * CONFIG_YAFFS_9BYTE_TAGS and the older on-NAND tags layout. ++ * If you have existing Yaffs images and the byte order differs from this, ++ * adjust 'oobfree' to match your existing Yaffs data. ++ * ++ * This nand_ecclayout scatters/gathers to/from the old-yaffs layout with the ++ * pageStatus byte (at NAND spare offset 4) scattered/gathered from/to ++ * the 9th byte. ++ * ++ * Old-style on-NAND format: T0,T1,T2,T3,P,B,T4,T5,E0,E1,E2,T6,T7,E3,E4,E5 ++ * We have/need PackedTags1 plus pageStatus: T0,T1,T2,T3,T4,T5,T6,T7,P ++ * where Tn are the tag bytes, En are MTD's ECC bytes, P is the pageStatus ++ * byte and B is the small-page bad-block indicator byte. ++ */ ++static struct nand_ecclayout nand_oob_16 = { ++ .eccbytes = 6, ++ .eccpos = { 8, 9, 10, 13, 14, 15 }, ++ .oobavail = 9, ++ .oobfree = { { 0, 4 }, { 6, 2 }, { 11, 2 }, { 4, 1 } } ++}; ++#endif ++ ++/* Write a chunk (page) of data to NAND. ++ * ++ * Caller always provides ExtendedTags data which are converted to a more ++ * compact (packed) form for storage in NAND. A mini-ECC runs over the ++ * contents of the tags meta-data; used to valid the tags when read. ++ * ++ * - Pack ExtendedTags to PackedTags1 form ++ * - Compute mini-ECC for PackedTags1 ++ * - Write data and packed tags to NAND. ++ * ++ * Note: Due to the use of the PackedTags1 meta-data which does not include ++ * a full sequence number (as found in the larger PackedTags2 form) it is ++ * necessary for Yaffs to re-write a chunk/page (just once) to mark it as ++ * discarded and dirty. This is not ideal: newer NAND parts are supposed ++ * to be written just once. When Yaffs performs this operation, this ++ * function is called with a NULL data pointer -- calling MTD write_oob ++ * without data is valid usage (2.6.17). ++ * ++ * Any underlying MTD error results in YAFFS_FAIL. ++ * Returns YAFFS_OK or YAFFS_FAIL. ++ */ ++int nandmtd1_WriteChunkWithTagsToNAND(yaffs_Device *dev, ++ int chunkInNAND, const __u8 *data, const yaffs_ExtendedTags *etags) ++{ ++ struct mtd_info *mtd = dev->genericDevice; ++ int chunkBytes = dev->nDataBytesPerChunk; ++ loff_t addr = ((loff_t)chunkInNAND) * chunkBytes; ++ struct mtd_oob_ops ops; ++ yaffs_PackedTags1 pt1; ++ int retval; ++ ++ /* we assume that PackedTags1 and yaffs_Tags are compatible */ ++ compile_time_assertion(sizeof(yaffs_PackedTags1) == 12); ++ compile_time_assertion(sizeof(yaffs_Tags) == 8); ++ ++ dev->nPageWrites++; ++ ++ yaffs_PackTags1(&pt1, etags); ++ yaffs_CalcTagsECC((yaffs_Tags *)&pt1); ++ ++ /* When deleting a chunk, the upper layer provides only skeletal ++ * etags, one with chunkDeleted set. However, we need to update the ++ * tags, not erase them completely. So we use the NAND write property ++ * that only zeroed-bits stick and set tag bytes to all-ones and ++ * zero just the (not) deleted bit. ++ */ ++#ifndef CONFIG_YAFFS_9BYTE_TAGS ++ if (etags->chunkDeleted) { ++ memset(&pt1, 0xff, 8); ++ /* clear delete status bit to indicate deleted */ ++ pt1.deleted = 0; ++ } ++#else ++ ((__u8 *)&pt1)[8] = 0xff; ++ if (etags->chunkDeleted) { ++ memset(&pt1, 0xff, 8); ++ /* zero pageStatus byte to indicate deleted */ ++ ((__u8 *)&pt1)[8] = 0; ++ } ++#endif ++ ++ memset(&ops, 0, sizeof(ops)); ++ ops.mode = MTD_OOB_AUTO; ++ ops.len = (data) ? chunkBytes : 0; ++ ops.ooblen = YTAG1_SIZE; ++ ops.datbuf = (__u8 *)data; ++ ops.oobbuf = (__u8 *)&pt1; ++ ++ retval = mtd->write_oob(mtd, addr, &ops); ++ if (retval) { ++ yaffs_trace(YAFFS_TRACE_MTD, ++ "write_oob failed, chunk %d, mtd error %d\n", ++ chunkInNAND, retval); ++ } ++ return retval ? YAFFS_FAIL : YAFFS_OK; ++} ++ ++/* Return with empty ExtendedTags but add eccResult. ++ */ ++static int rettags(yaffs_ExtendedTags *etags, int eccResult, int retval) ++{ ++ if (etags) { ++ memset(etags, 0, sizeof(*etags)); ++ etags->eccResult = eccResult; ++ } ++ return retval; ++} ++ ++/* Read a chunk (page) from NAND. ++ * ++ * Caller expects ExtendedTags data to be usable even on error; that is, ++ * all members except eccResult and blockBad are zeroed. ++ * ++ * - Check ECC results for data (if applicable) ++ * - Check for blank/erased block (return empty ExtendedTags if blank) ++ * - Check the PackedTags1 mini-ECC (correct if necessary/possible) ++ * - Convert PackedTags1 to ExtendedTags ++ * - Update eccResult and blockBad members to refect state. ++ * ++ * Returns YAFFS_OK or YAFFS_FAIL. ++ */ ++int nandmtd1_ReadChunkWithTagsFromNAND(yaffs_Device *dev, ++ int chunkInNAND, __u8 *data, yaffs_ExtendedTags *etags) ++{ ++ struct mtd_info *mtd = dev->genericDevice; ++ int chunkBytes = dev->nDataBytesPerChunk; ++ loff_t addr = ((loff_t)chunkInNAND) * chunkBytes; ++ int eccres = YAFFS_ECC_RESULT_NO_ERROR; ++ struct mtd_oob_ops ops; ++ yaffs_PackedTags1 pt1; ++ int retval; ++ int deleted; ++ ++ dev->nPageReads++; ++ ++ memset(&ops, 0, sizeof(ops)); ++ ops.mode = MTD_OOB_AUTO; ++ ops.len = (data) ? chunkBytes : 0; ++ ops.ooblen = YTAG1_SIZE; ++ ops.datbuf = data; ++ ops.oobbuf = (__u8 *)&pt1; ++ ++#if (MTD_VERSION_CODE < MTD_VERSION(2, 6, 20)) ++ /* In MTD 2.6.18 to 2.6.19 nand_base.c:nand_do_read_oob() has a bug; ++ * help it out with ops.len = ops.ooblen when ops.datbuf == NULL. ++ */ ++ ops.len = (ops.datbuf) ? ops.len : ops.ooblen; ++#endif ++ /* Read page and oob using MTD. ++ * Check status and determine ECC result. ++ */ ++ retval = mtd->read_oob(mtd, addr, &ops); ++ if (retval) { ++ yaffs_trace(YAFFS_TRACE_MTD, ++ "read_oob failed, chunk %d, mtd error %d\n", ++ chunkInNAND, retval); ++ } ++ ++ switch (retval) { ++ case 0: ++ /* no error */ ++ break; ++ ++ case -EUCLEAN: ++ /* MTD's ECC fixed the data */ ++ eccres = YAFFS_ECC_RESULT_FIXED; ++ dev->eccFixed++; ++ break; ++ ++ case -EBADMSG: ++ /* MTD's ECC could not fix the data */ ++ dev->eccUnfixed++; ++ /* fall into... */ ++ default: ++ rettags(etags, YAFFS_ECC_RESULT_UNFIXED, 0); ++ etags->blockBad = (mtd->block_isbad)(mtd, addr); ++ return YAFFS_FAIL; ++ } ++ ++ /* Check for a blank/erased chunk. ++ */ ++ if (yaffs_CheckFF((__u8 *)&pt1, 8)) { ++ /* when blank, upper layers want eccResult to be <= NO_ERROR */ ++ return rettags(etags, YAFFS_ECC_RESULT_NO_ERROR, YAFFS_OK); ++ } ++ ++#ifndef CONFIG_YAFFS_9BYTE_TAGS ++ /* Read deleted status (bit) then return it to it's non-deleted ++ * state before performing tags mini-ECC check. pt1.deleted is ++ * inverted. ++ */ ++ deleted = !pt1.deleted; ++ pt1.deleted = 1; ++#else ++ deleted = (yaffs_CountBits(((__u8 *)&pt1)[8]) < 7); ++#endif ++ ++ /* Check the packed tags mini-ECC and correct if necessary/possible. ++ */ ++ retval = yaffs_CheckECCOnTags((yaffs_Tags *)&pt1); ++ switch (retval) { ++ case 0: ++ /* no tags error, use MTD result */ ++ break; ++ case 1: ++ /* recovered tags-ECC error */ ++ dev->tagsEccFixed++; ++ if (eccres == YAFFS_ECC_RESULT_NO_ERROR) ++ eccres = YAFFS_ECC_RESULT_FIXED; ++ break; ++ default: ++ /* unrecovered tags-ECC error */ ++ dev->tagsEccUnfixed++; ++ return rettags(etags, YAFFS_ECC_RESULT_UNFIXED, YAFFS_FAIL); ++ } ++ ++ /* Unpack the tags to extended form and set ECC result. ++ * [set shouldBeFF just to keep yaffs_UnpackTags1 happy] ++ */ ++ pt1.shouldBeFF = 0xFFFFFFFF; ++ yaffs_UnpackTags1(etags, &pt1); ++ etags->eccResult = eccres; ++ ++ /* Set deleted state */ ++ etags->chunkDeleted = deleted; ++ return YAFFS_OK; ++} ++ ++/* Mark a block bad. ++ * ++ * This is a persistant state. ++ * Use of this function should be rare. ++ * ++ * Returns YAFFS_OK or YAFFS_FAIL. ++ */ ++int nandmtd1_MarkNANDBlockBad(struct yaffs_DeviceStruct *dev, int blockNo) ++{ ++ struct mtd_info *mtd = dev->genericDevice; ++ int blocksize = dev->nChunksPerBlock * dev->nDataBytesPerChunk; ++ int retval; ++ ++ yaffs_trace(YAFFS_TRACE_BAD_BLOCKS, "marking block %d bad\n", blockNo); ++ ++ retval = mtd->block_markbad(mtd, (loff_t)blocksize * blockNo); ++ return (retval) ? YAFFS_FAIL : YAFFS_OK; ++} ++ ++/* Check any MTD prerequists. ++ * ++ * Returns YAFFS_OK or YAFFS_FAIL. ++ */ ++static int nandmtd1_TestPrerequists(struct mtd_info *mtd) ++{ ++ /* 2.6.18 has mtd->ecclayout->oobavail */ ++ /* 2.6.21 has mtd->ecclayout->oobavail and mtd->oobavail */ ++ int oobavail = mtd->ecclayout->oobavail; ++ ++ if (oobavail < YTAG1_SIZE) { ++ yaffs_trace(YAFFS_TRACE_ERROR, ++ "mtd device has only %d bytes for tags, need %d\n", ++ oobavail, YTAG1_SIZE); ++ return YAFFS_FAIL; ++ } ++ return YAFFS_OK; ++} ++ ++/* Query for the current state of a specific block. ++ * ++ * Examine the tags of the first chunk of the block and return the state: ++ * - YAFFS_BLOCK_STATE_DEAD, the block is marked bad ++ * - YAFFS_BLOCK_STATE_NEEDS_SCANNING, the block is in use ++ * - YAFFS_BLOCK_STATE_EMPTY, the block is clean ++ * ++ * Always returns YAFFS_OK. ++ */ ++int nandmtd1_QueryNANDBlock(struct yaffs_DeviceStruct *dev, int blockNo, ++ yaffs_BlockState *pState, __u32 *pSequenceNumber) ++{ ++ struct mtd_info *mtd = dev->genericDevice; ++ int chunkNo = blockNo * dev->nChunksPerBlock; ++ loff_t addr = (loff_t)chunkNo * dev->nDataBytesPerChunk; ++ yaffs_ExtendedTags etags; ++ int state = YAFFS_BLOCK_STATE_DEAD; ++ int seqnum = 0; ++ int retval; ++ ++ /* We don't yet have a good place to test for MTD config prerequists. ++ * Do it here as we are called during the initial scan. ++ */ ++ if (nandmtd1_TestPrerequists(mtd) != YAFFS_OK) ++ return YAFFS_FAIL; ++ ++ retval = nandmtd1_ReadChunkWithTagsFromNAND(dev, chunkNo, NULL, &etags); ++ etags.blockBad = (mtd->block_isbad)(mtd, addr); ++ if (etags.blockBad) { ++ yaffs_trace(YAFFS_TRACE_BAD_BLOCKS, ++ "block %d is marked bad\n", blockNo); ++ state = YAFFS_BLOCK_STATE_DEAD; ++ } else if (etags.eccResult != YAFFS_ECC_RESULT_NO_ERROR) { ++ /* bad tags, need to look more closely */ ++ state = YAFFS_BLOCK_STATE_NEEDS_SCANNING; ++ } else if (etags.chunkUsed) { ++ state = YAFFS_BLOCK_STATE_NEEDS_SCANNING; ++ seqnum = etags.sequenceNumber; ++ } else { ++ state = YAFFS_BLOCK_STATE_EMPTY; ++ } ++ ++ *pState = state; ++ *pSequenceNumber = seqnum; ++ ++ /* query always succeeds */ ++ return YAFFS_OK; ++} ++ ++#endif /*MTD_VERSION*/ +diff -Nur linux-2.6.30.orig/fs/yaffs2/yaffs_mtdif1.h linux-2.6.30/fs/yaffs2/yaffs_mtdif1.h +--- linux-2.6.30.orig/fs/yaffs2/yaffs_mtdif1.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/yaffs_mtdif1.h 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,28 @@ ++/* ++ * YAFFS: Yet another Flash File System. A NAND-flash specific file system. ++ * ++ * Copyright (C) 2002-2007 Aleph One Ltd. ++ * for Toby Churchill Ltd and Brightstar Engineering ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU Lesser General Public License version 2.1 as ++ * published by the Free Software Foundation. ++ * ++ * Note: Only YAFFS headers are LGPL, YAFFS C code is covered by GPL. ++ */ ++ ++#ifndef __YAFFS_MTDIF1_H__ ++#define __YAFFS_MTDIF1_H__ ++ ++int nandmtd1_WriteChunkWithTagsToNAND(yaffs_Device *dev, int chunkInNAND, ++ const __u8 *data, const yaffs_ExtendedTags *tags); ++ ++int nandmtd1_ReadChunkWithTagsFromNAND(yaffs_Device *dev, int chunkInNAND, ++ __u8 *data, yaffs_ExtendedTags *tags); ++ ++int nandmtd1_MarkNANDBlockBad(struct yaffs_DeviceStruct *dev, int blockNo); ++ ++int nandmtd1_QueryNANDBlock(struct yaffs_DeviceStruct *dev, int blockNo, ++ yaffs_BlockState *state, __u32 *sequenceNumber); ++ ++#endif +diff -Nur linux-2.6.30.orig/fs/yaffs2/yaffs_mtdif2.c linux-2.6.30/fs/yaffs2/yaffs_mtdif2.c +--- linux-2.6.30.orig/fs/yaffs2/yaffs_mtdif2.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/yaffs_mtdif2.c 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,246 @@ ++/* ++ * YAFFS: Yet Another Flash File System. A NAND-flash specific file system. ++ * ++ * Copyright (C) 2002-2007 Aleph One Ltd. ++ * for Toby Churchill Ltd and Brightstar Engineering ++ * ++ * Created by Charles Manning <charles@aleph1.co.uk> ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License version 2 as ++ * published by the Free Software Foundation. ++ */ ++ ++/* mtd interface for YAFFS2 */ ++ ++const char *yaffs_mtdif2_c_version = ++ "$Id: yaffs_mtdif2.c,v 1.23 2009-03-06 17:20:53 wookey Exp $"; ++ ++#include "yportenv.h" ++ ++ ++#include "yaffs_mtdif2.h" ++ ++#include "linux/mtd/mtd.h" ++#include "linux/types.h" ++#include "linux/time.h" ++ ++#include "yaffs_packedtags2.h" ++ ++/* NB For use with inband tags.... ++ * We assume that the data buffer is of size totalBytersPerChunk so that we can also ++ * use it to load the tags. ++ */ ++int nandmtd2_WriteChunkWithTagsToNAND(yaffs_Device *dev, int chunkInNAND, ++ const __u8 *data, ++ const yaffs_ExtendedTags *tags) ++{ ++ struct mtd_info *mtd = (struct mtd_info *)(dev->genericDevice); ++#if (MTD_VERSION_CODE > MTD_VERSION(2, 6, 17)) ++ struct mtd_oob_ops ops; ++#else ++ size_t dummy; ++#endif ++ int retval = 0; ++ ++ loff_t addr; ++ ++ yaffs_PackedTags2 pt; ++ ++ T(YAFFS_TRACE_MTD, ++ (TSTR ++ ("nandmtd2_WriteChunkWithTagsToNAND chunk %d data %p tags %p" ++ TENDSTR), chunkInNAND, data, tags)); ++ ++ ++ addr = ((loff_t) chunkInNAND) * dev->totalBytesPerChunk; ++ ++ /* For yaffs2 writing there must be both data and tags. ++ * If we're using inband tags, then the tags are stuffed into ++ * the end of the data buffer. ++ */ ++ if (!data || !tags) ++ BUG(); ++ else if (dev->inbandTags) { ++ yaffs_PackedTags2TagsPart *pt2tp; ++ pt2tp = (yaffs_PackedTags2TagsPart *)(data + dev->nDataBytesPerChunk); ++ yaffs_PackTags2TagsPart(pt2tp, tags); ++ } else ++ yaffs_PackTags2(&pt, tags); ++ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 6, 17)) ++ ops.mode = MTD_OOB_AUTO; ++ ops.ooblen = (dev->inbandTags) ? 0 : sizeof(pt); ++ ops.len = dev->totalBytesPerChunk; ++ ops.ooboffs = 0; ++ ops.datbuf = (__u8 *)data; ++ ops.oobbuf = (dev->inbandTags) ? NULL : (void *)&pt; ++ retval = mtd->write_oob(mtd, addr, &ops); ++ ++#else ++ if (!dev->inbandTags) { ++ retval = ++ mtd->write_ecc(mtd, addr, dev->nDataBytesPerChunk, ++ &dummy, data, (__u8 *) &pt, NULL); ++ } else { ++ retval = ++ mtd->write(mtd, addr, dev->totalBytesPerChunk, &dummy, ++ data); ++ } ++#endif ++ ++ if (retval == 0) ++ return YAFFS_OK; ++ else ++ return YAFFS_FAIL; ++} ++ ++int nandmtd2_ReadChunkWithTagsFromNAND(yaffs_Device *dev, int chunkInNAND, ++ __u8 *data, yaffs_ExtendedTags *tags) ++{ ++ struct mtd_info *mtd = (struct mtd_info *)(dev->genericDevice); ++#if (MTD_VERSION_CODE > MTD_VERSION(2, 6, 17)) ++ struct mtd_oob_ops ops; ++#endif ++ size_t dummy; ++ int retval = 0; ++ int localData = 0; ++ ++ loff_t addr = ((loff_t) chunkInNAND) * dev->totalBytesPerChunk; ++ ++ yaffs_PackedTags2 pt; ++ ++ T(YAFFS_TRACE_MTD, ++ (TSTR ++ ("nandmtd2_ReadChunkWithTagsFromNAND chunk %d data %p tags %p" ++ TENDSTR), chunkInNAND, data, tags)); ++ ++ if (dev->inbandTags) { ++ ++ if (!data) { ++ localData = 1; ++ data = yaffs_GetTempBuffer(dev, __LINE__); ++ } ++ ++ ++ } ++ ++ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 6, 17)) ++ if (dev->inbandTags || (data && !tags)) ++ retval = mtd->read(mtd, addr, dev->totalBytesPerChunk, ++ &dummy, data); ++ else if (tags) { ++ ops.mode = MTD_OOB_AUTO; ++ ops.ooblen = sizeof(pt); ++ ops.len = data ? dev->nDataBytesPerChunk : sizeof(pt); ++ ops.ooboffs = 0; ++ ops.datbuf = data; ++ ops.oobbuf = dev->spareBuffer; ++ retval = mtd->read_oob(mtd, addr, &ops); ++ } ++#else ++ if (!dev->inbandTags && data && tags) { ++ ++ retval = mtd->read_ecc(mtd, addr, dev->nDataBytesPerChunk, ++ &dummy, data, dev->spareBuffer, ++ NULL); ++ } else { ++ if (data) ++ retval = ++ mtd->read(mtd, addr, dev->nDataBytesPerChunk, &dummy, ++ data); ++ if (!dev->inbandTags && tags) ++ retval = ++ mtd->read_oob(mtd, addr, mtd->oobsize, &dummy, ++ dev->spareBuffer); ++ } ++#endif ++ ++ ++ if (dev->inbandTags) { ++ if (tags) { ++ yaffs_PackedTags2TagsPart *pt2tp; ++ pt2tp = (yaffs_PackedTags2TagsPart *)&data[dev->nDataBytesPerChunk]; ++ yaffs_UnpackTags2TagsPart(tags, pt2tp); ++ } ++ } else { ++ if (tags) { ++ memcpy(&pt, dev->spareBuffer, sizeof(pt)); ++ yaffs_UnpackTags2(tags, &pt); ++ } ++ } ++ ++ if (localData) ++ yaffs_ReleaseTempBuffer(dev, data, __LINE__); ++ ++ if (tags && retval == -EBADMSG && tags->eccResult == YAFFS_ECC_RESULT_NO_ERROR) ++ tags->eccResult = YAFFS_ECC_RESULT_UNFIXED; ++ if (retval == 0) ++ return YAFFS_OK; ++ else ++ return YAFFS_FAIL; ++} ++ ++int nandmtd2_MarkNANDBlockBad(struct yaffs_DeviceStruct *dev, int blockNo) ++{ ++ struct mtd_info *mtd = (struct mtd_info *)(dev->genericDevice); ++ int retval; ++ T(YAFFS_TRACE_MTD, ++ (TSTR("nandmtd2_MarkNANDBlockBad %d" TENDSTR), blockNo)); ++ ++ retval = ++ mtd->block_markbad(mtd, ++ blockNo * dev->nChunksPerBlock * ++ dev->totalBytesPerChunk); ++ ++ if (retval == 0) ++ return YAFFS_OK; ++ else ++ return YAFFS_FAIL; ++ ++} ++ ++int nandmtd2_QueryNANDBlock(struct yaffs_DeviceStruct *dev, int blockNo, ++ yaffs_BlockState *state, __u32 *sequenceNumber) ++{ ++ struct mtd_info *mtd = (struct mtd_info *)(dev->genericDevice); ++ int retval; ++ ++ T(YAFFS_TRACE_MTD, ++ (TSTR("nandmtd2_QueryNANDBlock %d" TENDSTR), blockNo)); ++ retval = ++ mtd->block_isbad(mtd, ++ blockNo * dev->nChunksPerBlock * ++ dev->totalBytesPerChunk); ++ ++ if (retval) { ++ T(YAFFS_TRACE_MTD, (TSTR("block is bad" TENDSTR))); ++ ++ *state = YAFFS_BLOCK_STATE_DEAD; ++ *sequenceNumber = 0; ++ } else { ++ yaffs_ExtendedTags t; ++ nandmtd2_ReadChunkWithTagsFromNAND(dev, ++ blockNo * ++ dev->nChunksPerBlock, NULL, ++ &t); ++ ++ if (t.chunkUsed) { ++ *sequenceNumber = t.sequenceNumber; ++ *state = YAFFS_BLOCK_STATE_NEEDS_SCANNING; ++ } else { ++ *sequenceNumber = 0; ++ *state = YAFFS_BLOCK_STATE_EMPTY; ++ } ++ } ++ T(YAFFS_TRACE_MTD, ++ (TSTR("block is bad seq %d state %d" TENDSTR), *sequenceNumber, ++ *state)); ++ ++ if (retval == 0) ++ return YAFFS_OK; ++ else ++ return YAFFS_FAIL; ++} ++ +diff -Nur linux-2.6.30.orig/fs/yaffs2/yaffs_mtdif2.h linux-2.6.30/fs/yaffs2/yaffs_mtdif2.h +--- linux-2.6.30.orig/fs/yaffs2/yaffs_mtdif2.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/yaffs_mtdif2.h 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,29 @@ ++/* ++ * YAFFS: Yet another Flash File System . A NAND-flash specific file system. ++ * ++ * Copyright (C) 2002-2007 Aleph One Ltd. ++ * for Toby Churchill Ltd and Brightstar Engineering ++ * ++ * Created by Charles Manning <charles@aleph1.co.uk> ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU Lesser General Public License version 2.1 as ++ * published by the Free Software Foundation. ++ * ++ * Note: Only YAFFS headers are LGPL, YAFFS C code is covered by GPL. ++ */ ++ ++#ifndef __YAFFS_MTDIF2_H__ ++#define __YAFFS_MTDIF2_H__ ++ ++#include "yaffs_guts.h" ++int nandmtd2_WriteChunkWithTagsToNAND(yaffs_Device *dev, int chunkInNAND, ++ const __u8 *data, ++ const yaffs_ExtendedTags *tags); ++int nandmtd2_ReadChunkWithTagsFromNAND(yaffs_Device *dev, int chunkInNAND, ++ __u8 *data, yaffs_ExtendedTags *tags); ++int nandmtd2_MarkNANDBlockBad(struct yaffs_DeviceStruct *dev, int blockNo); ++int nandmtd2_QueryNANDBlock(struct yaffs_DeviceStruct *dev, int blockNo, ++ yaffs_BlockState *state, __u32 *sequenceNumber); ++ ++#endif +diff -Nur linux-2.6.30.orig/fs/yaffs2/yaffs_mtdif.c linux-2.6.30/fs/yaffs2/yaffs_mtdif.c +--- linux-2.6.30.orig/fs/yaffs2/yaffs_mtdif.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/yaffs_mtdif.c 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,241 @@ ++/* ++ * YAFFS: Yet Another Flash File System. A NAND-flash specific file system. ++ * ++ * Copyright (C) 2002-2007 Aleph One Ltd. ++ * for Toby Churchill Ltd and Brightstar Engineering ++ * ++ * Created by Charles Manning <charles@aleph1.co.uk> ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License version 2 as ++ * published by the Free Software Foundation. ++ */ ++ ++const char *yaffs_mtdif_c_version = ++ "$Id: yaffs_mtdif.c,v 1.22 2009-03-06 17:20:51 wookey Exp $"; ++ ++#include "yportenv.h" ++ ++ ++#include "yaffs_mtdif.h" ++ ++#include "linux/mtd/mtd.h" ++#include "linux/types.h" ++#include "linux/time.h" ++#include "linux/mtd/nand.h" ++ ++#if (MTD_VERSION_CODE < MTD_VERSION(2, 6, 18)) ++static struct nand_oobinfo yaffs_oobinfo = { ++ .useecc = 1, ++ .eccbytes = 6, ++ .eccpos = {8, 9, 10, 13, 14, 15} ++}; ++ ++static struct nand_oobinfo yaffs_noeccinfo = { ++ .useecc = 0, ++}; ++#endif ++ ++#if (MTD_VERSION_CODE > MTD_VERSION(2, 6, 17)) ++static inline void translate_spare2oob(const yaffs_Spare *spare, __u8 *oob) ++{ ++ oob[0] = spare->tagByte0; ++ oob[1] = spare->tagByte1; ++ oob[2] = spare->tagByte2; ++ oob[3] = spare->tagByte3; ++ oob[4] = spare->tagByte4; ++ oob[5] = spare->tagByte5 & 0x3f; ++ oob[5] |= spare->blockStatus == 'Y' ? 0 : 0x80; ++ oob[5] |= spare->pageStatus == 0 ? 0 : 0x40; ++ oob[6] = spare->tagByte6; ++ oob[7] = spare->tagByte7; ++} ++ ++static inline void translate_oob2spare(yaffs_Spare *spare, __u8 *oob) ++{ ++ struct yaffs_NANDSpare *nspare = (struct yaffs_NANDSpare *)spare; ++ spare->tagByte0 = oob[0]; ++ spare->tagByte1 = oob[1]; ++ spare->tagByte2 = oob[2]; ++ spare->tagByte3 = oob[3]; ++ spare->tagByte4 = oob[4]; ++ spare->tagByte5 = oob[5] == 0xff ? 0xff : oob[5] & 0x3f; ++ spare->blockStatus = oob[5] & 0x80 ? 0xff : 'Y'; ++ spare->pageStatus = oob[5] & 0x40 ? 0xff : 0; ++ spare->ecc1[0] = spare->ecc1[1] = spare->ecc1[2] = 0xff; ++ spare->tagByte6 = oob[6]; ++ spare->tagByte7 = oob[7]; ++ spare->ecc2[0] = spare->ecc2[1] = spare->ecc2[2] = 0xff; ++ ++ nspare->eccres1 = nspare->eccres2 = 0; /* FIXME */ ++} ++#endif ++ ++int nandmtd_WriteChunkToNAND(yaffs_Device *dev, int chunkInNAND, ++ const __u8 *data, const yaffs_Spare *spare) ++{ ++ struct mtd_info *mtd = (struct mtd_info *)(dev->genericDevice); ++#if (MTD_VERSION_CODE > MTD_VERSION(2, 6, 17)) ++ struct mtd_oob_ops ops; ++#endif ++ size_t dummy; ++ int retval = 0; ++ ++ loff_t addr = ((loff_t) chunkInNAND) * dev->nDataBytesPerChunk; ++#if (MTD_VERSION_CODE > MTD_VERSION(2, 6, 17)) ++ __u8 spareAsBytes[8]; /* OOB */ ++ ++ if (data && !spare) ++ retval = mtd->write(mtd, addr, dev->nDataBytesPerChunk, ++ &dummy, data); ++ else if (spare) { ++ if (dev->useNANDECC) { ++ translate_spare2oob(spare, spareAsBytes); ++ ops.mode = MTD_OOB_AUTO; ++ ops.ooblen = 8; /* temp hack */ ++ } else { ++ ops.mode = MTD_OOB_RAW; ++ ops.ooblen = YAFFS_BYTES_PER_SPARE; ++ } ++ ops.len = data ? dev->nDataBytesPerChunk : ops.ooblen; ++ ops.datbuf = (u8 *)data; ++ ops.ooboffs = 0; ++ ops.oobbuf = spareAsBytes; ++ retval = mtd->write_oob(mtd, addr, &ops); ++ } ++#else ++ __u8 *spareAsBytes = (__u8 *) spare; ++ ++ if (data && spare) { ++ if (dev->useNANDECC) ++ retval = ++ mtd->write_ecc(mtd, addr, dev->nDataBytesPerChunk, ++ &dummy, data, spareAsBytes, ++ &yaffs_oobinfo); ++ else ++ retval = ++ mtd->write_ecc(mtd, addr, dev->nDataBytesPerChunk, ++ &dummy, data, spareAsBytes, ++ &yaffs_noeccinfo); ++ } else { ++ if (data) ++ retval = ++ mtd->write(mtd, addr, dev->nDataBytesPerChunk, &dummy, ++ data); ++ if (spare) ++ retval = ++ mtd->write_oob(mtd, addr, YAFFS_BYTES_PER_SPARE, ++ &dummy, spareAsBytes); ++ } ++#endif ++ ++ if (retval == 0) ++ return YAFFS_OK; ++ else ++ return YAFFS_FAIL; ++} ++ ++int nandmtd_ReadChunkFromNAND(yaffs_Device *dev, int chunkInNAND, __u8 *data, ++ yaffs_Spare *spare) ++{ ++ struct mtd_info *mtd = (struct mtd_info *)(dev->genericDevice); ++#if (MTD_VERSION_CODE > MTD_VERSION(2, 6, 17)) ++ struct mtd_oob_ops ops; ++#endif ++ size_t dummy; ++ int retval = 0; ++ ++ loff_t addr = ((loff_t) chunkInNAND) * dev->nDataBytesPerChunk; ++#if (MTD_VERSION_CODE > MTD_VERSION(2, 6, 17)) ++ __u8 spareAsBytes[8]; /* OOB */ ++ ++ if (data && !spare) ++ retval = mtd->read(mtd, addr, dev->nDataBytesPerChunk, ++ &dummy, data); ++ else if (spare) { ++ if (dev->useNANDECC) { ++ ops.mode = MTD_OOB_AUTO; ++ ops.ooblen = 8; /* temp hack */ ++ } else { ++ ops.mode = MTD_OOB_RAW; ++ ops.ooblen = YAFFS_BYTES_PER_SPARE; ++ } ++ ops.len = data ? dev->nDataBytesPerChunk : ops.ooblen; ++ ops.datbuf = data; ++ ops.ooboffs = 0; ++ ops.oobbuf = spareAsBytes; ++ retval = mtd->read_oob(mtd, addr, &ops); ++ if (dev->useNANDECC) ++ translate_oob2spare(spare, spareAsBytes); ++ } ++#else ++ __u8 *spareAsBytes = (__u8 *) spare; ++ ++ if (data && spare) { ++ if (dev->useNANDECC) { ++ /* Careful, this call adds 2 ints */ ++ /* to the end of the spare data. Calling function */ ++ /* should allocate enough memory for spare, */ ++ /* i.e. [YAFFS_BYTES_PER_SPARE+2*sizeof(int)]. */ ++ retval = ++ mtd->read_ecc(mtd, addr, dev->nDataBytesPerChunk, ++ &dummy, data, spareAsBytes, ++ &yaffs_oobinfo); ++ } else { ++ retval = ++ mtd->read_ecc(mtd, addr, dev->nDataBytesPerChunk, ++ &dummy, data, spareAsBytes, ++ &yaffs_noeccinfo); ++ } ++ } else { ++ if (data) ++ retval = ++ mtd->read(mtd, addr, dev->nDataBytesPerChunk, &dummy, ++ data); ++ if (spare) ++ retval = ++ mtd->read_oob(mtd, addr, YAFFS_BYTES_PER_SPARE, ++ &dummy, spareAsBytes); ++ } ++#endif ++ ++ if (retval == 0) ++ return YAFFS_OK; ++ else ++ return YAFFS_FAIL; ++} ++ ++int nandmtd_EraseBlockInNAND(yaffs_Device *dev, int blockNumber) ++{ ++ struct mtd_info *mtd = (struct mtd_info *)(dev->genericDevice); ++ __u32 addr = ++ ((loff_t) blockNumber) * dev->nDataBytesPerChunk ++ * dev->nChunksPerBlock; ++ struct erase_info ei; ++ int retval = 0; ++ ++ ei.mtd = mtd; ++ ei.addr = addr; ++ ei.len = dev->nDataBytesPerChunk * dev->nChunksPerBlock; ++ ei.time = 1000; ++ ei.retries = 2; ++ ei.callback = NULL; ++ ei.priv = (u_long) dev; ++ ++ /* Todo finish off the ei if required */ ++ ++ sema_init(&dev->sem, 0); ++ ++ retval = mtd->erase(mtd, &ei); ++ ++ if (retval == 0) ++ return YAFFS_OK; ++ else ++ return YAFFS_FAIL; ++} ++ ++int nandmtd_InitialiseNAND(yaffs_Device *dev) ++{ ++ return YAFFS_OK; ++} ++ +diff -Nur linux-2.6.30.orig/fs/yaffs2/yaffs_mtdif.h linux-2.6.30/fs/yaffs2/yaffs_mtdif.h +--- linux-2.6.30.orig/fs/yaffs2/yaffs_mtdif.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/yaffs_mtdif.h 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,32 @@ ++/* ++ * YAFFS: Yet another Flash File System . A NAND-flash specific file system. ++ * ++ * Copyright (C) 2002-2007 Aleph One Ltd. ++ * for Toby Churchill Ltd and Brightstar Engineering ++ * ++ * Created by Charles Manning <charles@aleph1.co.uk> ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU Lesser General Public License version 2.1 as ++ * published by the Free Software Foundation. ++ * ++ * Note: Only YAFFS headers are LGPL, YAFFS C code is covered by GPL. ++ */ ++ ++#ifndef __YAFFS_MTDIF_H__ ++#define __YAFFS_MTDIF_H__ ++ ++#include "yaffs_guts.h" ++ ++#if (MTD_VERSION_CODE < MTD_VERSION(2, 6, 18)) ++extern struct nand_oobinfo yaffs_oobinfo; ++extern struct nand_oobinfo yaffs_noeccinfo; ++#endif ++ ++int nandmtd_WriteChunkToNAND(yaffs_Device *dev, int chunkInNAND, ++ const __u8 *data, const yaffs_Spare *spare); ++int nandmtd_ReadChunkFromNAND(yaffs_Device *dev, int chunkInNAND, __u8 *data, ++ yaffs_Spare *spare); ++int nandmtd_EraseBlockInNAND(yaffs_Device *dev, int blockNumber); ++int nandmtd_InitialiseNAND(yaffs_Device *dev); ++#endif +diff -Nur linux-2.6.30.orig/fs/yaffs2/yaffs_nand.c linux-2.6.30/fs/yaffs2/yaffs_nand.c +--- linux-2.6.30.orig/fs/yaffs2/yaffs_nand.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/yaffs_nand.c 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,135 @@ ++/* ++ * YAFFS: Yet Another Flash File System. A NAND-flash specific file system. ++ * ++ * Copyright (C) 2002-2007 Aleph One Ltd. ++ * for Toby Churchill Ltd and Brightstar Engineering ++ * ++ * Created by Charles Manning <charles@aleph1.co.uk> ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License version 2 as ++ * published by the Free Software Foundation. ++ */ ++ ++const char *yaffs_nand_c_version = ++ "$Id: yaffs_nand.c,v 1.10 2009-03-06 17:20:54 wookey Exp $"; ++ ++#include "yaffs_nand.h" ++#include "yaffs_tagscompat.h" ++#include "yaffs_tagsvalidity.h" ++ ++#include "yaffs_getblockinfo.h" ++ ++int yaffs_ReadChunkWithTagsFromNAND(yaffs_Device *dev, int chunkInNAND, ++ __u8 *buffer, ++ yaffs_ExtendedTags *tags) ++{ ++ int result; ++ yaffs_ExtendedTags localTags; ++ ++ int realignedChunkInNAND = chunkInNAND - dev->chunkOffset; ++ ++ /* If there are no tags provided, use local tags to get prioritised gc working */ ++ if (!tags) ++ tags = &localTags; ++ ++ if (dev->readChunkWithTagsFromNAND) ++ result = dev->readChunkWithTagsFromNAND(dev, realignedChunkInNAND, buffer, ++ tags); ++ else ++ result = yaffs_TagsCompatabilityReadChunkWithTagsFromNAND(dev, ++ realignedChunkInNAND, ++ buffer, ++ tags); ++ if (tags && ++ tags->eccResult > YAFFS_ECC_RESULT_NO_ERROR) { ++ ++ yaffs_BlockInfo *bi = yaffs_GetBlockInfo(dev, chunkInNAND/dev->nChunksPerBlock); ++ yaffs_HandleChunkError(dev, bi); ++ } ++ ++ return result; ++} ++ ++int yaffs_WriteChunkWithTagsToNAND(yaffs_Device *dev, ++ int chunkInNAND, ++ const __u8 *buffer, ++ yaffs_ExtendedTags *tags) ++{ ++ chunkInNAND -= dev->chunkOffset; ++ ++ ++ if (tags) { ++ tags->sequenceNumber = dev->sequenceNumber; ++ tags->chunkUsed = 1; ++ if (!yaffs_ValidateTags(tags)) { ++ T(YAFFS_TRACE_ERROR, ++ (TSTR("Writing uninitialised tags" TENDSTR))); ++ YBUG(); ++ } ++ T(YAFFS_TRACE_WRITE, ++ (TSTR("Writing chunk %d tags %d %d" TENDSTR), chunkInNAND, ++ tags->objectId, tags->chunkId)); ++ } else { ++ T(YAFFS_TRACE_ERROR, (TSTR("Writing with no tags" TENDSTR))); ++ YBUG(); ++ } ++ ++ if (dev->writeChunkWithTagsToNAND) ++ return dev->writeChunkWithTagsToNAND(dev, chunkInNAND, buffer, ++ tags); ++ else ++ return yaffs_TagsCompatabilityWriteChunkWithTagsToNAND(dev, ++ chunkInNAND, ++ buffer, ++ tags); ++} ++ ++int yaffs_MarkBlockBad(yaffs_Device *dev, int blockNo) ++{ ++ blockNo -= dev->blockOffset; ++ ++; ++ if (dev->markNANDBlockBad) ++ return dev->markNANDBlockBad(dev, blockNo); ++ else ++ return yaffs_TagsCompatabilityMarkNANDBlockBad(dev, blockNo); ++} ++ ++int yaffs_QueryInitialBlockState(yaffs_Device *dev, ++ int blockNo, ++ yaffs_BlockState *state, ++ __u32 *sequenceNumber) ++{ ++ blockNo -= dev->blockOffset; ++ ++ if (dev->queryNANDBlock) ++ return dev->queryNANDBlock(dev, blockNo, state, sequenceNumber); ++ else ++ return yaffs_TagsCompatabilityQueryNANDBlock(dev, blockNo, ++ state, ++ sequenceNumber); ++} ++ ++ ++int yaffs_EraseBlockInNAND(struct yaffs_DeviceStruct *dev, ++ int blockInNAND) ++{ ++ int result; ++ ++ blockInNAND -= dev->blockOffset; ++ ++ ++ dev->nBlockErasures++; ++ result = dev->eraseBlockInNAND(dev, blockInNAND); ++ ++ return result; ++} ++ ++int yaffs_InitialiseNAND(struct yaffs_DeviceStruct *dev) ++{ ++ return dev->initialiseNAND(dev); ++} ++ ++ ++ +diff -Nur linux-2.6.30.orig/fs/yaffs2/yaffs_nandemul2k.h linux-2.6.30/fs/yaffs2/yaffs_nandemul2k.h +--- linux-2.6.30.orig/fs/yaffs2/yaffs_nandemul2k.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/yaffs_nandemul2k.h 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,39 @@ ++/* ++ * YAFFS: Yet another Flash File System . A NAND-flash specific file system. ++ * ++ * Copyright (C) 2002-2007 Aleph One Ltd. ++ * for Toby Churchill Ltd and Brightstar Engineering ++ * ++ * Created by Charles Manning <charles@aleph1.co.uk> ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU Lesser General Public License version 2.1 as ++ * published by the Free Software Foundation. ++ * ++ * Note: Only YAFFS headers are LGPL, YAFFS C code is covered by GPL. ++ */ ++ ++/* Interface to emulated NAND functions (2k page size) */ ++ ++#ifndef __YAFFS_NANDEMUL2K_H__ ++#define __YAFFS_NANDEMUL2K_H__ ++ ++#include "yaffs_guts.h" ++ ++int nandemul2k_WriteChunkWithTagsToNAND(struct yaffs_DeviceStruct *dev, ++ int chunkInNAND, const __u8 *data, ++ const yaffs_ExtendedTags *tags); ++int nandemul2k_ReadChunkWithTagsFromNAND(struct yaffs_DeviceStruct *dev, ++ int chunkInNAND, __u8 *data, ++ yaffs_ExtendedTags *tags); ++int nandemul2k_MarkNANDBlockBad(struct yaffs_DeviceStruct *dev, int blockNo); ++int nandemul2k_QueryNANDBlock(struct yaffs_DeviceStruct *dev, int blockNo, ++ yaffs_BlockState *state, __u32 *sequenceNumber); ++int nandemul2k_EraseBlockInNAND(struct yaffs_DeviceStruct *dev, ++ int blockInNAND); ++int nandemul2k_InitialiseNAND(struct yaffs_DeviceStruct *dev); ++int nandemul2k_GetBytesPerChunk(void); ++int nandemul2k_GetChunksPerBlock(void); ++int nandemul2k_GetNumberOfBlocks(void); ++ ++#endif +diff -Nur linux-2.6.30.orig/fs/yaffs2/yaffs_nand.h linux-2.6.30/fs/yaffs2/yaffs_nand.h +--- linux-2.6.30.orig/fs/yaffs2/yaffs_nand.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/yaffs_nand.h 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,44 @@ ++/* ++ * YAFFS: Yet another Flash File System . A NAND-flash specific file system. ++ * ++ * Copyright (C) 2002-2007 Aleph One Ltd. ++ * for Toby Churchill Ltd and Brightstar Engineering ++ * ++ * Created by Charles Manning <charles@aleph1.co.uk> ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU Lesser General Public License version 2.1 as ++ * published by the Free Software Foundation. ++ * ++ * Note: Only YAFFS headers are LGPL, YAFFS C code is covered by GPL. ++ */ ++ ++#ifndef __YAFFS_NAND_H__ ++#define __YAFFS_NAND_H__ ++#include "yaffs_guts.h" ++ ++ ++ ++int yaffs_ReadChunkWithTagsFromNAND(yaffs_Device *dev, int chunkInNAND, ++ __u8 *buffer, ++ yaffs_ExtendedTags *tags); ++ ++int yaffs_WriteChunkWithTagsToNAND(yaffs_Device *dev, ++ int chunkInNAND, ++ const __u8 *buffer, ++ yaffs_ExtendedTags *tags); ++ ++int yaffs_MarkBlockBad(yaffs_Device *dev, int blockNo); ++ ++int yaffs_QueryInitialBlockState(yaffs_Device *dev, ++ int blockNo, ++ yaffs_BlockState *state, ++ unsigned *sequenceNumber); ++ ++int yaffs_EraseBlockInNAND(struct yaffs_DeviceStruct *dev, ++ int blockInNAND); ++ ++int yaffs_InitialiseNAND(struct yaffs_DeviceStruct *dev); ++ ++#endif ++ +diff -Nur linux-2.6.30.orig/fs/yaffs2/yaffs_packedtags1.c linux-2.6.30/fs/yaffs2/yaffs_packedtags1.c +--- linux-2.6.30.orig/fs/yaffs2/yaffs_packedtags1.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/yaffs_packedtags1.c 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,50 @@ ++/* ++ * YAFFS: Yet Another Flash File System. A NAND-flash specific file system. ++ * ++ * Copyright (C) 2002-2007 Aleph One Ltd. ++ * for Toby Churchill Ltd and Brightstar Engineering ++ * ++ * Created by Charles Manning <charles@aleph1.co.uk> ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License version 2 as ++ * published by the Free Software Foundation. ++ */ ++ ++#include "yaffs_packedtags1.h" ++#include "yportenv.h" ++ ++void yaffs_PackTags1(yaffs_PackedTags1 *pt, const yaffs_ExtendedTags *t) ++{ ++ pt->chunkId = t->chunkId; ++ pt->serialNumber = t->serialNumber; ++ pt->byteCount = t->byteCount; ++ pt->objectId = t->objectId; ++ pt->ecc = 0; ++ pt->deleted = (t->chunkDeleted) ? 0 : 1; ++ pt->unusedStuff = 0; ++ pt->shouldBeFF = 0xFFFFFFFF; ++ ++} ++ ++void yaffs_UnpackTags1(yaffs_ExtendedTags *t, const yaffs_PackedTags1 *pt) ++{ ++ static const __u8 allFF[] = ++ { 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, ++0xff }; ++ ++ if (memcmp(allFF, pt, sizeof(yaffs_PackedTags1))) { ++ t->blockBad = 0; ++ if (pt->shouldBeFF != 0xFFFFFFFF) ++ t->blockBad = 1; ++ t->chunkUsed = 1; ++ t->objectId = pt->objectId; ++ t->chunkId = pt->chunkId; ++ t->byteCount = pt->byteCount; ++ t->eccResult = YAFFS_ECC_RESULT_NO_ERROR; ++ t->chunkDeleted = (pt->deleted) ? 0 : 1; ++ t->serialNumber = pt->serialNumber; ++ } else { ++ memset(t, 0, sizeof(yaffs_ExtendedTags)); ++ } ++} +diff -Nur linux-2.6.30.orig/fs/yaffs2/yaffs_packedtags1.h linux-2.6.30/fs/yaffs2/yaffs_packedtags1.h +--- linux-2.6.30.orig/fs/yaffs2/yaffs_packedtags1.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/yaffs_packedtags1.h 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,37 @@ ++/* ++ * YAFFS: Yet another Flash File System . A NAND-flash specific file system. ++ * ++ * Copyright (C) 2002-2007 Aleph One Ltd. ++ * for Toby Churchill Ltd and Brightstar Engineering ++ * ++ * Created by Charles Manning <charles@aleph1.co.uk> ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU Lesser General Public License version 2.1 as ++ * published by the Free Software Foundation. ++ * ++ * Note: Only YAFFS headers are LGPL, YAFFS C code is covered by GPL. ++ */ ++ ++/* This is used to pack YAFFS1 tags, not YAFFS2 tags. */ ++ ++#ifndef __YAFFS_PACKEDTAGS1_H__ ++#define __YAFFS_PACKEDTAGS1_H__ ++ ++#include "yaffs_guts.h" ++ ++typedef struct { ++ unsigned chunkId:20; ++ unsigned serialNumber:2; ++ unsigned byteCount:10; ++ unsigned objectId:18; ++ unsigned ecc:12; ++ unsigned deleted:1; ++ unsigned unusedStuff:1; ++ unsigned shouldBeFF; ++ ++} yaffs_PackedTags1; ++ ++void yaffs_PackTags1(yaffs_PackedTags1 *pt, const yaffs_ExtendedTags *t); ++void yaffs_UnpackTags1(yaffs_ExtendedTags *t, const yaffs_PackedTags1 *pt); ++#endif +diff -Nur linux-2.6.30.orig/fs/yaffs2/yaffs_packedtags2.c linux-2.6.30/fs/yaffs2/yaffs_packedtags2.c +--- linux-2.6.30.orig/fs/yaffs2/yaffs_packedtags2.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/yaffs_packedtags2.c 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,206 @@ ++/* ++ * YAFFS: Yet Another Flash File System. A NAND-flash specific file system. ++ * ++ * Copyright (C) 2002-2007 Aleph One Ltd. ++ * for Toby Churchill Ltd and Brightstar Engineering ++ * ++ * Created by Charles Manning <charles@aleph1.co.uk> ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License version 2 as ++ * published by the Free Software Foundation. ++ */ ++ ++#include "yaffs_packedtags2.h" ++#include "yportenv.h" ++#include "yaffs_tagsvalidity.h" ++ ++/* This code packs a set of extended tags into a binary structure for ++ * NAND storage ++ */ ++ ++/* Some of the information is "extra" struff which can be packed in to ++ * speed scanning ++ * This is defined by having the EXTRA_HEADER_INFO_FLAG set. ++ */ ++ ++/* Extra flags applied to chunkId */ ++ ++#define EXTRA_HEADER_INFO_FLAG 0x80000000 ++#define EXTRA_SHRINK_FLAG 0x40000000 ++#define EXTRA_SHADOWS_FLAG 0x20000000 ++#define EXTRA_SPARE_FLAGS 0x10000000 ++ ++#define ALL_EXTRA_FLAGS 0xF0000000 ++ ++/* Also, the top 4 bits of the object Id are set to the object type. */ ++#define EXTRA_OBJECT_TYPE_SHIFT (28) ++#define EXTRA_OBJECT_TYPE_MASK ((0x0F) << EXTRA_OBJECT_TYPE_SHIFT) ++ ++ ++static void yaffs_DumpPackedTags2TagsPart(const yaffs_PackedTags2TagsPart *ptt) ++{ ++ T(YAFFS_TRACE_MTD, ++ (TSTR("packed tags obj %d chunk %d byte %d seq %d" TENDSTR), ++ ptt->objectId, ptt->chunkId, ptt->byteCount, ++ ptt->sequenceNumber)); ++} ++static void yaffs_DumpPackedTags2(const yaffs_PackedTags2 *pt) ++{ ++ yaffs_DumpPackedTags2TagsPart(&pt->t); ++} ++ ++static void yaffs_DumpTags2(const yaffs_ExtendedTags *t) ++{ ++ T(YAFFS_TRACE_MTD, ++ (TSTR ++ ("ext.tags eccres %d blkbad %d chused %d obj %d chunk%d byte %d del %d ser %d seq %d" ++ TENDSTR), t->eccResult, t->blockBad, t->chunkUsed, t->objectId, ++ t->chunkId, t->byteCount, t->chunkDeleted, t->serialNumber, ++ t->sequenceNumber)); ++ ++} ++ ++void yaffs_PackTags2TagsPart(yaffs_PackedTags2TagsPart *ptt, ++ const yaffs_ExtendedTags *t) ++{ ++ ptt->chunkId = t->chunkId; ++ ptt->sequenceNumber = t->sequenceNumber; ++ ptt->byteCount = t->byteCount; ++ ptt->objectId = t->objectId; ++ ++ if (t->chunkId == 0 && t->extraHeaderInfoAvailable) { ++ /* Store the extra header info instead */ ++ /* We save the parent object in the chunkId */ ++ ptt->chunkId = EXTRA_HEADER_INFO_FLAG ++ | t->extraParentObjectId; ++ if (t->extraIsShrinkHeader) ++ ptt->chunkId |= EXTRA_SHRINK_FLAG; ++ if (t->extraShadows) ++ ptt->chunkId |= EXTRA_SHADOWS_FLAG; ++ ++ ptt->objectId &= ~EXTRA_OBJECT_TYPE_MASK; ++ ptt->objectId |= ++ (t->extraObjectType << EXTRA_OBJECT_TYPE_SHIFT); ++ ++ if (t->extraObjectType == YAFFS_OBJECT_TYPE_HARDLINK) ++ ptt->byteCount = t->extraEquivalentObjectId; ++ else if (t->extraObjectType == YAFFS_OBJECT_TYPE_FILE) ++ ptt->byteCount = t->extraFileLength; ++ else ++ ptt->byteCount = 0; ++ } ++ ++ yaffs_DumpPackedTags2TagsPart(ptt); ++ yaffs_DumpTags2(t); ++} ++ ++ ++void yaffs_PackTags2(yaffs_PackedTags2 *pt, const yaffs_ExtendedTags *t) ++{ ++ yaffs_PackTags2TagsPart(&pt->t, t); ++ ++#ifndef YAFFS_IGNORE_TAGS_ECC ++ { ++ yaffs_ECCCalculateOther((unsigned char *)&pt->t, ++ sizeof(yaffs_PackedTags2TagsPart), ++ &pt->ecc); ++ } ++#endif ++} ++ ++ ++void yaffs_UnpackTags2TagsPart(yaffs_ExtendedTags *t, ++ yaffs_PackedTags2TagsPart *ptt) ++{ ++ ++ memset(t, 0, sizeof(yaffs_ExtendedTags)); ++ ++ yaffs_InitialiseTags(t); ++ ++ if (ptt->sequenceNumber != 0xFFFFFFFF) { ++ t->blockBad = 0; ++ t->chunkUsed = 1; ++ t->objectId = ptt->objectId; ++ t->chunkId = ptt->chunkId; ++ t->byteCount = ptt->byteCount; ++ t->chunkDeleted = 0; ++ t->serialNumber = 0; ++ t->sequenceNumber = ptt->sequenceNumber; ++ ++ /* Do extra header info stuff */ ++ ++ if (ptt->chunkId & EXTRA_HEADER_INFO_FLAG) { ++ t->chunkId = 0; ++ t->byteCount = 0; ++ ++ t->extraHeaderInfoAvailable = 1; ++ t->extraParentObjectId = ++ ptt->chunkId & (~(ALL_EXTRA_FLAGS)); ++ t->extraIsShrinkHeader = ++ (ptt->chunkId & EXTRA_SHRINK_FLAG) ? 1 : 0; ++ t->extraShadows = ++ (ptt->chunkId & EXTRA_SHADOWS_FLAG) ? 1 : 0; ++ t->extraObjectType = ++ ptt->objectId >> EXTRA_OBJECT_TYPE_SHIFT; ++ t->objectId &= ~EXTRA_OBJECT_TYPE_MASK; ++ ++ if (t->extraObjectType == YAFFS_OBJECT_TYPE_HARDLINK) ++ t->extraEquivalentObjectId = ptt->byteCount; ++ else ++ t->extraFileLength = ptt->byteCount; ++ } ++ } ++ ++ yaffs_DumpPackedTags2TagsPart(ptt); ++ yaffs_DumpTags2(t); ++ ++} ++ ++ ++void yaffs_UnpackTags2(yaffs_ExtendedTags *t, yaffs_PackedTags2 *pt) ++{ ++ ++ yaffs_ECCResult eccResult = YAFFS_ECC_RESULT_NO_ERROR; ++ ++ if (pt->t.sequenceNumber != 0xFFFFFFFF) { ++ /* Page is in use */ ++#ifndef YAFFS_IGNORE_TAGS_ECC ++ { ++ yaffs_ECCOther ecc; ++ int result; ++ yaffs_ECCCalculateOther((unsigned char *)&pt->t, ++ sizeof ++ (yaffs_PackedTags2TagsPart), ++ &ecc); ++ result = ++ yaffs_ECCCorrectOther((unsigned char *)&pt->t, ++ sizeof ++ (yaffs_PackedTags2TagsPart), ++ &pt->ecc, &ecc); ++ switch (result) { ++ case 0: ++ eccResult = YAFFS_ECC_RESULT_NO_ERROR; ++ break; ++ case 1: ++ eccResult = YAFFS_ECC_RESULT_FIXED; ++ break; ++ case -1: ++ eccResult = YAFFS_ECC_RESULT_UNFIXED; ++ break; ++ default: ++ eccResult = YAFFS_ECC_RESULT_UNKNOWN; ++ } ++ } ++#endif ++ } ++ ++ yaffs_UnpackTags2TagsPart(t, &pt->t); ++ ++ t->eccResult = eccResult; ++ ++ yaffs_DumpPackedTags2(pt); ++ yaffs_DumpTags2(t); ++ ++} ++ +diff -Nur linux-2.6.30.orig/fs/yaffs2/yaffs_packedtags2.h linux-2.6.30/fs/yaffs2/yaffs_packedtags2.h +--- linux-2.6.30.orig/fs/yaffs2/yaffs_packedtags2.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/yaffs_packedtags2.h 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,43 @@ ++/* ++ * YAFFS: Yet another Flash File System . A NAND-flash specific file system. ++ * ++ * Copyright (C) 2002-2007 Aleph One Ltd. ++ * for Toby Churchill Ltd and Brightstar Engineering ++ * ++ * Created by Charles Manning <charles@aleph1.co.uk> ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU Lesser General Public License version 2.1 as ++ * published by the Free Software Foundation. ++ * ++ * Note: Only YAFFS headers are LGPL, YAFFS C code is covered by GPL. ++ */ ++ ++/* This is used to pack YAFFS2 tags, not YAFFS1tags. */ ++ ++#ifndef __YAFFS_PACKEDTAGS2_H__ ++#define __YAFFS_PACKEDTAGS2_H__ ++ ++#include "yaffs_guts.h" ++#include "yaffs_ecc.h" ++ ++typedef struct { ++ unsigned sequenceNumber; ++ unsigned objectId; ++ unsigned chunkId; ++ unsigned byteCount; ++} yaffs_PackedTags2TagsPart; ++ ++typedef struct { ++ yaffs_PackedTags2TagsPart t; ++ yaffs_ECCOther ecc; ++} yaffs_PackedTags2; ++ ++/* Full packed tags with ECC, used for oob tags */ ++void yaffs_PackTags2(yaffs_PackedTags2 *pt, const yaffs_ExtendedTags *t); ++void yaffs_UnpackTags2(yaffs_ExtendedTags *t, yaffs_PackedTags2 *pt); ++ ++/* Only the tags part (no ECC for use with inband tags */ ++void yaffs_PackTags2TagsPart(yaffs_PackedTags2TagsPart *pt, const yaffs_ExtendedTags *t); ++void yaffs_UnpackTags2TagsPart(yaffs_ExtendedTags *t, yaffs_PackedTags2TagsPart *pt); ++#endif +diff -Nur linux-2.6.30.orig/fs/yaffs2/yaffs_qsort.c linux-2.6.30/fs/yaffs2/yaffs_qsort.c +--- linux-2.6.30.orig/fs/yaffs2/yaffs_qsort.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/yaffs_qsort.c 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,163 @@ ++/* ++ * Copyright (c) 1992, 1993 ++ * The Regents of the University of California. All rights reserved. ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions ++ * are met: ++ * 1. Redistributions of source code must retain the above copyright ++ * notice, this list of conditions and the following disclaimer. ++ * 2. Redistributions in binary form must reproduce the above copyright ++ * notice, this list of conditions and the following disclaimer in the ++ * documentation and/or other materials provided with the distribution. ++ * 3. Neither the name of the University nor the names of its contributors ++ * may be used to endorse or promote products derived from this software ++ * without specific prior written permission. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND ++ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE ++ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ++ * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE ++ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL ++ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS ++ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) ++ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT ++ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY ++ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF ++ * SUCH DAMAGE. ++ */ ++ ++#include "yportenv.h" ++/* #include <linux/string.h> */ ++ ++/* ++ * Qsort routine from Bentley & McIlroy's "Engineering a Sort Function". ++ */ ++#define swapcode(TYPE, parmi, parmj, n) do { \ ++ long i = (n) / sizeof (TYPE); \ ++ register TYPE *pi = (TYPE *) (parmi); \ ++ register TYPE *pj = (TYPE *) (parmj); \ ++ do { \ ++ register TYPE t = *pi; \ ++ *pi++ = *pj; \ ++ *pj++ = t; \ ++ } while (--i > 0); \ ++} while (0) ++ ++#define SWAPINIT(a, es) swaptype = ((char *)a - (char *)0) % sizeof(long) || \ ++ es % sizeof(long) ? 2 : es == sizeof(long) ? 0 : 1; ++ ++static __inline void ++swapfunc(char *a, char *b, int n, int swaptype) ++{ ++ if (swaptype <= 1) ++ swapcode(long, a, b, n); ++ else ++ swapcode(char, a, b, n); ++} ++ ++#define yswap(a, b) do { \ ++ if (swaptype == 0) { \ ++ long t = *(long *)(a); \ ++ *(long *)(a) = *(long *)(b); \ ++ *(long *)(b) = t; \ ++ } else \ ++ swapfunc(a, b, es, swaptype); \ ++} while (0) ++ ++#define vecswap(a, b, n) if ((n) > 0) swapfunc(a, b, n, swaptype) ++ ++static __inline char * ++med3(char *a, char *b, char *c, int (*cmp)(const void *, const void *)) ++{ ++ return cmp(a, b) < 0 ? ++ (cmp(b, c) < 0 ? b : (cmp(a, c) < 0 ? c : a)) ++ : (cmp(b, c) > 0 ? b : (cmp(a, c) < 0 ? a : c)); ++} ++ ++#ifndef min ++#define min(a, b) (((a) < (b)) ? (a) : (b)) ++#endif ++ ++void ++yaffs_qsort(void *aa, size_t n, size_t es, ++ int (*cmp)(const void *, const void *)) ++{ ++ char *pa, *pb, *pc, *pd, *pl, *pm, *pn; ++ int d, r, swaptype, swap_cnt; ++ register char *a = aa; ++ ++loop: SWAPINIT(a, es); ++ swap_cnt = 0; ++ if (n < 7) { ++ for (pm = (char *)a + es; pm < (char *) a + n * es; pm += es) ++ for (pl = pm; pl > (char *) a && cmp(pl - es, pl) > 0; ++ pl -= es) ++ yswap(pl, pl - es); ++ return; ++ } ++ pm = (char *)a + (n / 2) * es; ++ if (n > 7) { ++ pl = (char *)a; ++ pn = (char *)a + (n - 1) * es; ++ if (n > 40) { ++ d = (n / 8) * es; ++ pl = med3(pl, pl + d, pl + 2 * d, cmp); ++ pm = med3(pm - d, pm, pm + d, cmp); ++ pn = med3(pn - 2 * d, pn - d, pn, cmp); ++ } ++ pm = med3(pl, pm, pn, cmp); ++ } ++ yswap(a, pm); ++ pa = pb = (char *)a + es; ++ ++ pc = pd = (char *)a + (n - 1) * es; ++ for (;;) { ++ while (pb <= pc && (r = cmp(pb, a)) <= 0) { ++ if (r == 0) { ++ swap_cnt = 1; ++ yswap(pa, pb); ++ pa += es; ++ } ++ pb += es; ++ } ++ while (pb <= pc && (r = cmp(pc, a)) >= 0) { ++ if (r == 0) { ++ swap_cnt = 1; ++ yswap(pc, pd); ++ pd -= es; ++ } ++ pc -= es; ++ } ++ if (pb > pc) ++ break; ++ yswap(pb, pc); ++ swap_cnt = 1; ++ pb += es; ++ pc -= es; ++ } ++ if (swap_cnt == 0) { /* Switch to insertion sort */ ++ for (pm = (char *) a + es; pm < (char *) a + n * es; pm += es) ++ for (pl = pm; pl > (char *) a && cmp(pl - es, pl) > 0; ++ pl -= es) ++ yswap(pl, pl - es); ++ return; ++ } ++ ++ pn = (char *)a + n * es; ++ r = min(pa - (char *)a, pb - pa); ++ vecswap(a, pb - r, r); ++ r = min((long)(pd - pc), (long)(pn - pd - es)); ++ vecswap(pb, pn - r, r); ++ r = pb - pa; ++ if (r > es) ++ yaffs_qsort(a, r / es, es, cmp); ++ r = pd - pc; ++ if (r > es) { ++ /* Iterate rather than recurse to save stack space */ ++ a = pn - r; ++ n = r / es; ++ goto loop; ++ } ++/* yaffs_qsort(pn - r, r / es, es, cmp);*/ ++} +diff -Nur linux-2.6.30.orig/fs/yaffs2/yaffs_qsort.h linux-2.6.30/fs/yaffs2/yaffs_qsort.h +--- linux-2.6.30.orig/fs/yaffs2/yaffs_qsort.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/yaffs_qsort.h 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,23 @@ ++/* ++ * YAFFS: Yet another Flash File System . A NAND-flash specific file system. ++ * ++ * Copyright (C) 2002-2007 Aleph One Ltd. ++ * for Toby Churchill Ltd and Brightstar Engineering ++ * ++ * Created by Charles Manning <charles@aleph1.co.uk> ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU Lesser General Public License version 2.1 as ++ * published by the Free Software Foundation. ++ * ++ * Note: Only YAFFS headers are LGPL, YAFFS C code is covered by GPL. ++ */ ++ ++ ++#ifndef __YAFFS_QSORT_H__ ++#define __YAFFS_QSORT_H__ ++ ++extern void yaffs_qsort(void *const base, size_t total_elems, size_t size, ++ int (*cmp)(const void *, const void *)); ++ ++#endif +diff -Nur linux-2.6.30.orig/fs/yaffs2/yaffs_tagscompat.c linux-2.6.30/fs/yaffs2/yaffs_tagscompat.c +--- linux-2.6.30.orig/fs/yaffs2/yaffs_tagscompat.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/yaffs_tagscompat.c 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,541 @@ ++/* ++ * YAFFS: Yet Another Flash File System. A NAND-flash specific file system. ++ * ++ * Copyright (C) 2002-2007 Aleph One Ltd. ++ * for Toby Churchill Ltd and Brightstar Engineering ++ * ++ * Created by Charles Manning <charles@aleph1.co.uk> ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License version 2 as ++ * published by the Free Software Foundation. ++ */ ++ ++#include "yaffs_guts.h" ++#include "yaffs_tagscompat.h" ++#include "yaffs_ecc.h" ++#include "yaffs_getblockinfo.h" ++ ++static void yaffs_HandleReadDataError(yaffs_Device *dev, int chunkInNAND); ++#ifdef NOTYET ++static void yaffs_CheckWrittenBlock(yaffs_Device *dev, int chunkInNAND); ++static void yaffs_HandleWriteChunkOk(yaffs_Device *dev, int chunkInNAND, ++ const __u8 *data, ++ const yaffs_Spare *spare); ++static void yaffs_HandleUpdateChunk(yaffs_Device *dev, int chunkInNAND, ++ const yaffs_Spare *spare); ++static void yaffs_HandleWriteChunkError(yaffs_Device *dev, int chunkInNAND); ++#endif ++ ++static const char yaffs_countBitsTable[256] = { ++ 0, 1, 1, 2, 1, 2, 2, 3, 1, 2, 2, 3, 2, 3, 3, 4, ++ 1, 2, 2, 3, 2, 3, 3, 4, 2, 3, 3, 4, 3, 4, 4, 5, ++ 1, 2, 2, 3, 2, 3, 3, 4, 2, 3, 3, 4, 3, 4, 4, 5, ++ 2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6, ++ 1, 2, 2, 3, 2, 3, 3, 4, 2, 3, 3, 4, 3, 4, 4, 5, ++ 2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6, ++ 2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6, ++ 3, 4, 4, 5, 4, 5, 5, 6, 4, 5, 5, 6, 5, 6, 6, 7, ++ 1, 2, 2, 3, 2, 3, 3, 4, 2, 3, 3, 4, 3, 4, 4, 5, ++ 2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6, ++ 2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6, ++ 3, 4, 4, 5, 4, 5, 5, 6, 4, 5, 5, 6, 5, 6, 6, 7, ++ 2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6, ++ 3, 4, 4, 5, 4, 5, 5, 6, 4, 5, 5, 6, 5, 6, 6, 7, ++ 3, 4, 4, 5, 4, 5, 5, 6, 4, 5, 5, 6, 5, 6, 6, 7, ++ 4, 5, 5, 6, 5, 6, 6, 7, 5, 6, 6, 7, 6, 7, 7, 8 ++}; ++ ++int yaffs_CountBits(__u8 x) ++{ ++ int retVal; ++ retVal = yaffs_countBitsTable[x]; ++ return retVal; ++} ++ ++/********** Tags ECC calculations *********/ ++ ++void yaffs_CalcECC(const __u8 *data, yaffs_Spare *spare) ++{ ++ yaffs_ECCCalculate(data, spare->ecc1); ++ yaffs_ECCCalculate(&data[256], spare->ecc2); ++} ++ ++void yaffs_CalcTagsECC(yaffs_Tags *tags) ++{ ++ /* Calculate an ecc */ ++ ++ unsigned char *b = ((yaffs_TagsUnion *) tags)->asBytes; ++ unsigned i, j; ++ unsigned ecc = 0; ++ unsigned bit = 0; ++ ++ tags->ecc = 0; ++ ++ for (i = 0; i < 8; i++) { ++ for (j = 1; j & 0xff; j <<= 1) { ++ bit++; ++ if (b[i] & j) ++ ecc ^= bit; ++ } ++ } ++ ++ tags->ecc = ecc; ++ ++} ++ ++int yaffs_CheckECCOnTags(yaffs_Tags *tags) ++{ ++ unsigned ecc = tags->ecc; ++ ++ yaffs_CalcTagsECC(tags); ++ ++ ecc ^= tags->ecc; ++ ++ if (ecc && ecc <= 64) { ++ /* TODO: Handle the failure better. Retire? */ ++ unsigned char *b = ((yaffs_TagsUnion *) tags)->asBytes; ++ ++ ecc--; ++ ++ b[ecc / 8] ^= (1 << (ecc & 7)); ++ ++ /* Now recvalc the ecc */ ++ yaffs_CalcTagsECC(tags); ++ ++ return 1; /* recovered error */ ++ } else if (ecc) { ++ /* Wierd ecc failure value */ ++ /* TODO Need to do somethiong here */ ++ return -1; /* unrecovered error */ ++ } ++ ++ return 0; ++} ++ ++/********** Tags **********/ ++ ++static void yaffs_LoadTagsIntoSpare(yaffs_Spare *sparePtr, ++ yaffs_Tags *tagsPtr) ++{ ++ yaffs_TagsUnion *tu = (yaffs_TagsUnion *) tagsPtr; ++ ++ yaffs_CalcTagsECC(tagsPtr); ++ ++ sparePtr->tagByte0 = tu->asBytes[0]; ++ sparePtr->tagByte1 = tu->asBytes[1]; ++ sparePtr->tagByte2 = tu->asBytes[2]; ++ sparePtr->tagByte3 = tu->asBytes[3]; ++ sparePtr->tagByte4 = tu->asBytes[4]; ++ sparePtr->tagByte5 = tu->asBytes[5]; ++ sparePtr->tagByte6 = tu->asBytes[6]; ++ sparePtr->tagByte7 = tu->asBytes[7]; ++} ++ ++static void yaffs_GetTagsFromSpare(yaffs_Device *dev, yaffs_Spare *sparePtr, ++ yaffs_Tags *tagsPtr) ++{ ++ yaffs_TagsUnion *tu = (yaffs_TagsUnion *) tagsPtr; ++ int result; ++ ++ tu->asBytes[0] = sparePtr->tagByte0; ++ tu->asBytes[1] = sparePtr->tagByte1; ++ tu->asBytes[2] = sparePtr->tagByte2; ++ tu->asBytes[3] = sparePtr->tagByte3; ++ tu->asBytes[4] = sparePtr->tagByte4; ++ tu->asBytes[5] = sparePtr->tagByte5; ++ tu->asBytes[6] = sparePtr->tagByte6; ++ tu->asBytes[7] = sparePtr->tagByte7; ++ ++ result = yaffs_CheckECCOnTags(tagsPtr); ++ if (result > 0) ++ dev->tagsEccFixed++; ++ else if (result < 0) ++ dev->tagsEccUnfixed++; ++} ++ ++static void yaffs_SpareInitialise(yaffs_Spare *spare) ++{ ++ memset(spare, 0xFF, sizeof(yaffs_Spare)); ++} ++ ++static int yaffs_WriteChunkToNAND(struct yaffs_DeviceStruct *dev, ++ int chunkInNAND, const __u8 *data, ++ yaffs_Spare *spare) ++{ ++ if (chunkInNAND < dev->startBlock * dev->nChunksPerBlock) { ++ T(YAFFS_TRACE_ERROR, ++ (TSTR("**>> yaffs chunk %d is not valid" TENDSTR), ++ chunkInNAND)); ++ return YAFFS_FAIL; ++ } ++ ++ dev->nPageWrites++; ++ return dev->writeChunkToNAND(dev, chunkInNAND, data, spare); ++} ++ ++static int yaffs_ReadChunkFromNAND(struct yaffs_DeviceStruct *dev, ++ int chunkInNAND, ++ __u8 *data, ++ yaffs_Spare *spare, ++ yaffs_ECCResult *eccResult, ++ int doErrorCorrection) ++{ ++ int retVal; ++ yaffs_Spare localSpare; ++ ++ dev->nPageReads++; ++ ++ if (!spare && data) { ++ /* If we don't have a real spare, then we use a local one. */ ++ /* Need this for the calculation of the ecc */ ++ spare = &localSpare; ++ } ++ ++ if (!dev->useNANDECC) { ++ retVal = dev->readChunkFromNAND(dev, chunkInNAND, data, spare); ++ if (data && doErrorCorrection) { ++ /* Do ECC correction */ ++ /* Todo handle any errors */ ++ int eccResult1, eccResult2; ++ __u8 calcEcc[3]; ++ ++ yaffs_ECCCalculate(data, calcEcc); ++ eccResult1 = ++ yaffs_ECCCorrect(data, spare->ecc1, calcEcc); ++ yaffs_ECCCalculate(&data[256], calcEcc); ++ eccResult2 = ++ yaffs_ECCCorrect(&data[256], spare->ecc2, calcEcc); ++ ++ if (eccResult1 > 0) { ++ T(YAFFS_TRACE_ERROR, ++ (TSTR ++ ("**>>yaffs ecc error fix performed on chunk %d:0" ++ TENDSTR), chunkInNAND)); ++ dev->eccFixed++; ++ } else if (eccResult1 < 0) { ++ T(YAFFS_TRACE_ERROR, ++ (TSTR ++ ("**>>yaffs ecc error unfixed on chunk %d:0" ++ TENDSTR), chunkInNAND)); ++ dev->eccUnfixed++; ++ } ++ ++ if (eccResult2 > 0) { ++ T(YAFFS_TRACE_ERROR, ++ (TSTR ++ ("**>>yaffs ecc error fix performed on chunk %d:1" ++ TENDSTR), chunkInNAND)); ++ dev->eccFixed++; ++ } else if (eccResult2 < 0) { ++ T(YAFFS_TRACE_ERROR, ++ (TSTR ++ ("**>>yaffs ecc error unfixed on chunk %d:1" ++ TENDSTR), chunkInNAND)); ++ dev->eccUnfixed++; ++ } ++ ++ if (eccResult1 || eccResult2) { ++ /* We had a data problem on this page */ ++ yaffs_HandleReadDataError(dev, chunkInNAND); ++ } ++ ++ if (eccResult1 < 0 || eccResult2 < 0) ++ *eccResult = YAFFS_ECC_RESULT_UNFIXED; ++ else if (eccResult1 > 0 || eccResult2 > 0) ++ *eccResult = YAFFS_ECC_RESULT_FIXED; ++ else ++ *eccResult = YAFFS_ECC_RESULT_NO_ERROR; ++ } ++ } else { ++ /* Must allocate enough memory for spare+2*sizeof(int) */ ++ /* for ecc results from device. */ ++ struct yaffs_NANDSpare nspare; ++ ++ memset(&nspare, 0, sizeof(nspare)); ++ ++ retVal = dev->readChunkFromNAND(dev, chunkInNAND, data, ++ (yaffs_Spare *) &nspare); ++ memcpy(spare, &nspare, sizeof(yaffs_Spare)); ++ if (data && doErrorCorrection) { ++ if (nspare.eccres1 > 0) { ++ T(YAFFS_TRACE_ERROR, ++ (TSTR ++ ("**>>mtd ecc error fix performed on chunk %d:0" ++ TENDSTR), chunkInNAND)); ++ } else if (nspare.eccres1 < 0) { ++ T(YAFFS_TRACE_ERROR, ++ (TSTR ++ ("**>>mtd ecc error unfixed on chunk %d:0" ++ TENDSTR), chunkInNAND)); ++ } ++ ++ if (nspare.eccres2 > 0) { ++ T(YAFFS_TRACE_ERROR, ++ (TSTR ++ ("**>>mtd ecc error fix performed on chunk %d:1" ++ TENDSTR), chunkInNAND)); ++ } else if (nspare.eccres2 < 0) { ++ T(YAFFS_TRACE_ERROR, ++ (TSTR ++ ("**>>mtd ecc error unfixed on chunk %d:1" ++ TENDSTR), chunkInNAND)); ++ } ++ ++ if (nspare.eccres1 || nspare.eccres2) { ++ /* We had a data problem on this page */ ++ yaffs_HandleReadDataError(dev, chunkInNAND); ++ } ++ ++ if (nspare.eccres1 < 0 || nspare.eccres2 < 0) ++ *eccResult = YAFFS_ECC_RESULT_UNFIXED; ++ else if (nspare.eccres1 > 0 || nspare.eccres2 > 0) ++ *eccResult = YAFFS_ECC_RESULT_FIXED; ++ else ++ *eccResult = YAFFS_ECC_RESULT_NO_ERROR; ++ ++ } ++ } ++ return retVal; ++} ++ ++#ifdef NOTYET ++static int yaffs_CheckChunkErased(struct yaffs_DeviceStruct *dev, ++ int chunkInNAND) ++{ ++ static int init; ++ static __u8 cmpbuf[YAFFS_BYTES_PER_CHUNK]; ++ static __u8 data[YAFFS_BYTES_PER_CHUNK]; ++ /* Might as well always allocate the larger size for */ ++ /* dev->useNANDECC == true; */ ++ static __u8 spare[sizeof(struct yaffs_NANDSpare)]; ++ ++ dev->readChunkFromNAND(dev, chunkInNAND, data, (yaffs_Spare *) spare); ++ ++ if (!init) { ++ memset(cmpbuf, 0xff, YAFFS_BYTES_PER_CHUNK); ++ init = 1; ++ } ++ ++ if (memcmp(cmpbuf, data, YAFFS_BYTES_PER_CHUNK)) ++ return YAFFS_FAIL; ++ if (memcmp(cmpbuf, spare, 16)) ++ return YAFFS_FAIL; ++ ++ return YAFFS_OK; ++ ++} ++#endif ++ ++/* ++ * Functions for robustisizing ++ */ ++ ++static void yaffs_HandleReadDataError(yaffs_Device *dev, int chunkInNAND) ++{ ++ int blockInNAND = chunkInNAND / dev->nChunksPerBlock; ++ ++ /* Mark the block for retirement */ ++ yaffs_GetBlockInfo(dev, blockInNAND + dev->blockOffset)->needsRetiring = 1; ++ T(YAFFS_TRACE_ERROR | YAFFS_TRACE_BAD_BLOCKS, ++ (TSTR("**>>Block %d marked for retirement" TENDSTR), blockInNAND)); ++ ++ /* TODO: ++ * Just do a garbage collection on the affected block ++ * then retire the block ++ * NB recursion ++ */ ++} ++ ++#ifdef NOTYET ++static void yaffs_CheckWrittenBlock(yaffs_Device *dev, int chunkInNAND) ++{ ++} ++ ++static void yaffs_HandleWriteChunkOk(yaffs_Device *dev, int chunkInNAND, ++ const __u8 *data, ++ const yaffs_Spare *spare) ++{ ++} ++ ++static void yaffs_HandleUpdateChunk(yaffs_Device *dev, int chunkInNAND, ++ const yaffs_Spare *spare) ++{ ++} ++ ++static void yaffs_HandleWriteChunkError(yaffs_Device *dev, int chunkInNAND) ++{ ++ int blockInNAND = chunkInNAND / dev->nChunksPerBlock; ++ ++ /* Mark the block for retirement */ ++ yaffs_GetBlockInfo(dev, blockInNAND)->needsRetiring = 1; ++ /* Delete the chunk */ ++ yaffs_DeleteChunk(dev, chunkInNAND, 1, __LINE__); ++} ++ ++static int yaffs_VerifyCompare(const __u8 *d0, const __u8 *d1, ++ const yaffs_Spare *s0, const yaffs_Spare *s1) ++{ ++ ++ if (memcmp(d0, d1, YAFFS_BYTES_PER_CHUNK) != 0 || ++ s0->tagByte0 != s1->tagByte0 || ++ s0->tagByte1 != s1->tagByte1 || ++ s0->tagByte2 != s1->tagByte2 || ++ s0->tagByte3 != s1->tagByte3 || ++ s0->tagByte4 != s1->tagByte4 || ++ s0->tagByte5 != s1->tagByte5 || ++ s0->tagByte6 != s1->tagByte6 || ++ s0->tagByte7 != s1->tagByte7 || ++ s0->ecc1[0] != s1->ecc1[0] || ++ s0->ecc1[1] != s1->ecc1[1] || ++ s0->ecc1[2] != s1->ecc1[2] || ++ s0->ecc2[0] != s1->ecc2[0] || ++ s0->ecc2[1] != s1->ecc2[1] || s0->ecc2[2] != s1->ecc2[2]) { ++ return 0; ++ } ++ ++ return 1; ++} ++#endif /* NOTYET */ ++ ++int yaffs_TagsCompatabilityWriteChunkWithTagsToNAND(yaffs_Device *dev, ++ int chunkInNAND, ++ const __u8 *data, ++ const yaffs_ExtendedTags *eTags) ++{ ++ yaffs_Spare spare; ++ yaffs_Tags tags; ++ ++ yaffs_SpareInitialise(&spare); ++ ++ if (eTags->chunkDeleted) ++ spare.pageStatus = 0; ++ else { ++ tags.objectId = eTags->objectId; ++ tags.chunkId = eTags->chunkId; ++ ++ tags.byteCountLSB = eTags->byteCount & 0x3ff; ++ ++ if (dev->nDataBytesPerChunk >= 1024) ++ tags.byteCountMSB = (eTags->byteCount >> 10) & 3; ++ else ++ tags.byteCountMSB = 3; ++ ++ ++ tags.serialNumber = eTags->serialNumber; ++ ++ if (!dev->useNANDECC && data) ++ yaffs_CalcECC(data, &spare); ++ ++ yaffs_LoadTagsIntoSpare(&spare, &tags); ++ ++ } ++ ++ return yaffs_WriteChunkToNAND(dev, chunkInNAND, data, &spare); ++} ++ ++int yaffs_TagsCompatabilityReadChunkWithTagsFromNAND(yaffs_Device *dev, ++ int chunkInNAND, ++ __u8 *data, ++ yaffs_ExtendedTags *eTags) ++{ ++ ++ yaffs_Spare spare; ++ yaffs_Tags tags; ++ yaffs_ECCResult eccResult = YAFFS_ECC_RESULT_UNKNOWN; ++ ++ static yaffs_Spare spareFF; ++ static int init; ++ ++ if (!init) { ++ memset(&spareFF, 0xFF, sizeof(spareFF)); ++ init = 1; ++ } ++ ++ if (yaffs_ReadChunkFromNAND ++ (dev, chunkInNAND, data, &spare, &eccResult, 1)) { ++ /* eTags may be NULL */ ++ if (eTags) { ++ ++ int deleted = ++ (yaffs_CountBits(spare.pageStatus) < 7) ? 1 : 0; ++ ++ eTags->chunkDeleted = deleted; ++ eTags->eccResult = eccResult; ++ eTags->blockBad = 0; /* We're reading it */ ++ /* therefore it is not a bad block */ ++ eTags->chunkUsed = ++ (memcmp(&spareFF, &spare, sizeof(spareFF)) != ++ 0) ? 1 : 0; ++ ++ if (eTags->chunkUsed) { ++ yaffs_GetTagsFromSpare(dev, &spare, &tags); ++ ++ eTags->objectId = tags.objectId; ++ eTags->chunkId = tags.chunkId; ++ eTags->byteCount = tags.byteCountLSB; ++ ++ if (dev->nDataBytesPerChunk >= 1024) ++ eTags->byteCount |= (((unsigned) tags.byteCountMSB) << 10); ++ ++ eTags->serialNumber = tags.serialNumber; ++ } ++ } ++ ++ return YAFFS_OK; ++ } else { ++ return YAFFS_FAIL; ++ } ++} ++ ++int yaffs_TagsCompatabilityMarkNANDBlockBad(struct yaffs_DeviceStruct *dev, ++ int blockInNAND) ++{ ++ ++ yaffs_Spare spare; ++ ++ memset(&spare, 0xff, sizeof(yaffs_Spare)); ++ ++ spare.blockStatus = 'Y'; ++ ++ yaffs_WriteChunkToNAND(dev, blockInNAND * dev->nChunksPerBlock, NULL, ++ &spare); ++ yaffs_WriteChunkToNAND(dev, blockInNAND * dev->nChunksPerBlock + 1, ++ NULL, &spare); ++ ++ return YAFFS_OK; ++ ++} ++ ++int yaffs_TagsCompatabilityQueryNANDBlock(struct yaffs_DeviceStruct *dev, ++ int blockNo, ++ yaffs_BlockState *state, ++ __u32 *sequenceNumber) ++{ ++ ++ yaffs_Spare spare0, spare1; ++ static yaffs_Spare spareFF; ++ static int init; ++ yaffs_ECCResult dummy; ++ ++ if (!init) { ++ memset(&spareFF, 0xFF, sizeof(spareFF)); ++ init = 1; ++ } ++ ++ *sequenceNumber = 0; ++ ++ yaffs_ReadChunkFromNAND(dev, blockNo * dev->nChunksPerBlock, NULL, ++ &spare0, &dummy, 1); ++ yaffs_ReadChunkFromNAND(dev, blockNo * dev->nChunksPerBlock + 1, NULL, ++ &spare1, &dummy, 1); ++ ++ if (yaffs_CountBits(spare0.blockStatus & spare1.blockStatus) < 7) ++ *state = YAFFS_BLOCK_STATE_DEAD; ++ else if (memcmp(&spareFF, &spare0, sizeof(spareFF)) == 0) ++ *state = YAFFS_BLOCK_STATE_EMPTY; ++ else ++ *state = YAFFS_BLOCK_STATE_NEEDS_SCANNING; ++ ++ return YAFFS_OK; ++} +diff -Nur linux-2.6.30.orig/fs/yaffs2/yaffs_tagscompat.h linux-2.6.30/fs/yaffs2/yaffs_tagscompat.h +--- linux-2.6.30.orig/fs/yaffs2/yaffs_tagscompat.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/yaffs_tagscompat.h 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,39 @@ ++/* ++ * YAFFS: Yet another Flash File System . A NAND-flash specific file system. ++ * ++ * Copyright (C) 2002-2007 Aleph One Ltd. ++ * for Toby Churchill Ltd and Brightstar Engineering ++ * ++ * Created by Charles Manning <charles@aleph1.co.uk> ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU Lesser General Public License version 2.1 as ++ * published by the Free Software Foundation. ++ * ++ * Note: Only YAFFS headers are LGPL, YAFFS C code is covered by GPL. ++ */ ++ ++#ifndef __YAFFS_TAGSCOMPAT_H__ ++#define __YAFFS_TAGSCOMPAT_H__ ++ ++#include "yaffs_guts.h" ++int yaffs_TagsCompatabilityWriteChunkWithTagsToNAND(yaffs_Device *dev, ++ int chunkInNAND, ++ const __u8 *data, ++ const yaffs_ExtendedTags *tags); ++int yaffs_TagsCompatabilityReadChunkWithTagsFromNAND(yaffs_Device *dev, ++ int chunkInNAND, ++ __u8 *data, ++ yaffs_ExtendedTags *tags); ++int yaffs_TagsCompatabilityMarkNANDBlockBad(struct yaffs_DeviceStruct *dev, ++ int blockNo); ++int yaffs_TagsCompatabilityQueryNANDBlock(struct yaffs_DeviceStruct *dev, ++ int blockNo, ++ yaffs_BlockState *state, ++ __u32 *sequenceNumber); ++ ++void yaffs_CalcTagsECC(yaffs_Tags *tags); ++int yaffs_CheckECCOnTags(yaffs_Tags *tags); ++int yaffs_CountBits(__u8 byte); ++ ++#endif +diff -Nur linux-2.6.30.orig/fs/yaffs2/yaffs_tagsvalidity.c linux-2.6.30/fs/yaffs2/yaffs_tagsvalidity.c +--- linux-2.6.30.orig/fs/yaffs2/yaffs_tagsvalidity.c 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/yaffs_tagsvalidity.c 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,28 @@ ++/* ++ * YAFFS: Yet Another Flash File System. A NAND-flash specific file system. ++ * ++ * Copyright (C) 2002-2007 Aleph One Ltd. ++ * for Toby Churchill Ltd and Brightstar Engineering ++ * ++ * Created by Charles Manning <charles@aleph1.co.uk> ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU General Public License version 2 as ++ * published by the Free Software Foundation. ++ */ ++ ++#include "yaffs_tagsvalidity.h" ++ ++void yaffs_InitialiseTags(yaffs_ExtendedTags *tags) ++{ ++ memset(tags, 0, sizeof(yaffs_ExtendedTags)); ++ tags->validMarker0 = 0xAAAAAAAA; ++ tags->validMarker1 = 0x55555555; ++} ++ ++int yaffs_ValidateTags(yaffs_ExtendedTags *tags) ++{ ++ return (tags->validMarker0 == 0xAAAAAAAA && ++ tags->validMarker1 == 0x55555555); ++ ++} +diff -Nur linux-2.6.30.orig/fs/yaffs2/yaffs_tagsvalidity.h linux-2.6.30/fs/yaffs2/yaffs_tagsvalidity.h +--- linux-2.6.30.orig/fs/yaffs2/yaffs_tagsvalidity.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/yaffs_tagsvalidity.h 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,24 @@ ++/* ++ * YAFFS: Yet another Flash File System . A NAND-flash specific file system. ++ * ++ * Copyright (C) 2002-2007 Aleph One Ltd. ++ * for Toby Churchill Ltd and Brightstar Engineering ++ * ++ * Created by Charles Manning <charles@aleph1.co.uk> ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU Lesser General Public License version 2.1 as ++ * published by the Free Software Foundation. ++ * ++ * Note: Only YAFFS headers are LGPL, YAFFS C code is covered by GPL. ++ */ ++ ++ ++#ifndef __YAFFS_TAGS_VALIDITY_H__ ++#define __YAFFS_TAGS_VALIDITY_H__ ++ ++#include "yaffs_guts.h" ++ ++void yaffs_InitialiseTags(yaffs_ExtendedTags *tags); ++int yaffs_ValidateTags(yaffs_ExtendedTags *tags); ++#endif +diff -Nur linux-2.6.30.orig/fs/yaffs2/yportenv.h linux-2.6.30/fs/yaffs2/yportenv.h +--- linux-2.6.30.orig/fs/yaffs2/yportenv.h 1970-01-01 01:00:00.000000000 +0100 ++++ linux-2.6.30/fs/yaffs2/yportenv.h 2009-06-11 09:21:04.000000000 +0200 +@@ -0,0 +1,203 @@ ++/* ++ * YAFFS: Yet another Flash File System . A NAND-flash specific file system. ++ * ++ * Copyright (C) 2002-2007 Aleph One Ltd. ++ * for Toby Churchill Ltd and Brightstar Engineering ++ * ++ * Created by Charles Manning <charles@aleph1.co.uk> ++ * ++ * This program is free software; you can redistribute it and/or modify ++ * it under the terms of the GNU Lesser General Public License version 2.1 as ++ * published by the Free Software Foundation. ++ * ++ * Note: Only YAFFS headers are LGPL, YAFFS C code is covered by GPL. ++ */ ++ ++ ++#ifndef __YPORTENV_H__ ++#define __YPORTENV_H__ ++ ++/* ++ * Define the MTD version in terms of Linux Kernel versions ++ * This allows yaffs to be used independantly of the kernel ++ * as well as with it. ++ */ ++ ++#define MTD_VERSION(a, b, c) (((a) << 16) + ((b) << 8) + (c)) ++ ++#if defined CONFIG_YAFFS_WINCE ++ ++#include "ywinceenv.h" ++ ++#elif defined __KERNEL__ ++ ++#include "moduleconfig.h" ++ ++/* Linux kernel */ ++ ++#include <linux/version.h> ++#define MTD_VERSION_CODE LINUX_VERSION_CODE ++ ++#if (LINUX_VERSION_CODE < KERNEL_VERSION(2, 6, 19)) ++#include <linux/config.h> ++#endif ++#include <linux/kernel.h> ++#include <linux/mm.h> ++#include <linux/sched.h> ++#include <linux/string.h> ++#include <linux/slab.h> ++#include <linux/vmalloc.h> ++ ++#define YCHAR char ++#define YUCHAR unsigned char ++#define _Y(x) x ++#define yaffs_strcat(a, b) strcat(a, b) ++#define yaffs_strcpy(a, b) strcpy(a, b) ++#define yaffs_strncpy(a, b, c) strncpy(a, b, c) ++#define yaffs_strncmp(a, b, c) strncmp(a, b, c) ++#define yaffs_strlen(s) strlen(s) ++#define yaffs_sprintf sprintf ++#define yaffs_toupper(a) toupper(a) ++ ++#define Y_INLINE inline ++ ++#define YAFFS_LOSTNFOUND_NAME "lost+found" ++#define YAFFS_LOSTNFOUND_PREFIX "obj" ++ ++/* #define YPRINTF(x) printk x */ ++#define YMALLOC(x) kmalloc(x, GFP_NOFS) ++#define YFREE(x) kfree(x) ++#define YMALLOC_ALT(x) vmalloc(x) ++#define YFREE_ALT(x) vfree(x) ++#define YMALLOC_DMA(x) YMALLOC(x) ++ ++/* KR - added for use in scan so processes aren't blocked indefinitely. */ ++#define YYIELD() schedule() ++ ++#define YAFFS_ROOT_MODE 0666 ++#define YAFFS_LOSTNFOUND_MODE 0666 ++ ++#if (LINUX_VERSION_CODE > KERNEL_VERSION(2, 5, 0)) ++#define Y_CURRENT_TIME CURRENT_TIME.tv_sec ++#define Y_TIME_CONVERT(x) (x).tv_sec ++#else ++#define Y_CURRENT_TIME CURRENT_TIME ++#define Y_TIME_CONVERT(x) (x) ++#endif ++ ++#define yaffs_SumCompare(x, y) ((x) == (y)) ++#define yaffs_strcmp(a, b) strcmp(a, b) ++ ++#define TENDSTR "\n" ++#define TSTR(x) KERN_WARNING x ++#define TCONT(x) x ++#define TOUT(p) printk p ++ ++#define yaffs_trace(mask, fmt, args...) \ ++ do { if ((mask) & (yaffs_traceMask|YAFFS_TRACE_ERROR)) \ ++ printk(KERN_WARNING "yaffs: " fmt, ## args); \ ++ } while (0) ++ ++#define compile_time_assertion(assertion) \ ++ ({ int x = __builtin_choose_expr(assertion, 0, (void)0); (void) x; }) ++ ++#elif defined CONFIG_YAFFS_DIRECT ++ ++#define MTD_VERSION_CODE MTD_VERSION(2, 6, 22) ++ ++/* Direct interface */ ++#include "ydirectenv.h" ++ ++#elif defined CONFIG_YAFFS_UTIL ++ ++/* Stuff for YAFFS utilities */ ++ ++#include "stdlib.h" ++#include "stdio.h" ++#include "string.h" ++ ++#include "devextras.h" ++ ++#define YMALLOC(x) malloc(x) ++#define YFREE(x) free(x) ++#define YMALLOC_ALT(x) malloc(x) ++#define YFREE_ALT(x) free(x) ++ ++#define YCHAR char ++#define YUCHAR unsigned char ++#define _Y(x) x ++#define yaffs_strcat(a, b) strcat(a, b) ++#define yaffs_strcpy(a, b) strcpy(a, b) ++#define yaffs_strncpy(a, b, c) strncpy(a, b, c) ++#define yaffs_strlen(s) strlen(s) ++#define yaffs_sprintf sprintf ++#define yaffs_toupper(a) toupper(a) ++ ++#define Y_INLINE inline ++ ++/* #define YINFO(s) YPRINTF(( __FILE__ " %d %s\n",__LINE__,s)) */ ++/* #define YALERT(s) YINFO(s) */ ++ ++#define TENDSTR "\n" ++#define TSTR(x) x ++#define TOUT(p) printf p ++ ++#define YAFFS_LOSTNFOUND_NAME "lost+found" ++#define YAFFS_LOSTNFOUND_PREFIX "obj" ++/* #define YPRINTF(x) printf x */ ++ ++#define YAFFS_ROOT_MODE 0666 ++#define YAFFS_LOSTNFOUND_MODE 0666 ++ ++#define yaffs_SumCompare(x, y) ((x) == (y)) ++#define yaffs_strcmp(a, b) strcmp(a, b) ++ ++#else ++/* Should have specified a configuration type */ ++#error Unknown configuration ++ ++#endif ++ ++/* see yaffs_fs.c */ ++extern unsigned int yaffs_traceMask; ++extern unsigned int yaffs_wr_attempts; ++ ++/* ++ * Tracing flags. ++ * The flags masked in YAFFS_TRACE_ALWAYS are always traced. ++ */ ++ ++#define YAFFS_TRACE_OS 0x00000002 ++#define YAFFS_TRACE_ALLOCATE 0x00000004 ++#define YAFFS_TRACE_SCAN 0x00000008 ++#define YAFFS_TRACE_BAD_BLOCKS 0x00000010 ++#define YAFFS_TRACE_ERASE 0x00000020 ++#define YAFFS_TRACE_GC 0x00000040 ++#define YAFFS_TRACE_WRITE 0x00000080 ++#define YAFFS_TRACE_TRACING 0x00000100 ++#define YAFFS_TRACE_DELETION 0x00000200 ++#define YAFFS_TRACE_BUFFERS 0x00000400 ++#define YAFFS_TRACE_NANDACCESS 0x00000800 ++#define YAFFS_TRACE_GC_DETAIL 0x00001000 ++#define YAFFS_TRACE_SCAN_DEBUG 0x00002000 ++#define YAFFS_TRACE_MTD 0x00004000 ++#define YAFFS_TRACE_CHECKPOINT 0x00008000 ++ ++#define YAFFS_TRACE_VERIFY 0x00010000 ++#define YAFFS_TRACE_VERIFY_NAND 0x00020000 ++#define YAFFS_TRACE_VERIFY_FULL 0x00040000 ++#define YAFFS_TRACE_VERIFY_ALL 0x000F0000 ++ ++ ++#define YAFFS_TRACE_ERROR 0x40000000 ++#define YAFFS_TRACE_BUG 0x80000000 ++#define YAFFS_TRACE_ALWAYS 0xF0000000 ++ ++ ++#define T(mask, p) do { if ((mask) & (yaffs_traceMask | YAFFS_TRACE_ALWAYS)) TOUT(p); } while (0) ++ ++#ifndef YBUG ++#define YBUG() do {T(YAFFS_TRACE_BUG, (TSTR("==>> yaffs bug: " __FILE__ " %d" TENDSTR), __LINE__)); } while (0) ++#endif ++ ++#endif diff --git a/target/native/device.mk b/target/native/device.mk index 8f861905d..522355bee 100644 --- a/target/native/device.mk +++ b/target/native/device.mk @@ -1,6 +1,6 @@ ARCH:= $(shell uname -m|sed -e "s/i.*86/x86/" -e "s/_64//") CPU_ARCH:= $(shell uname -m) -KERNEL_VERSION:= 2.6.30.5 +KERNEL_VERSION:= 2.6.31.4 KERNEL_RELEASE:= 1 KERNEL_MD5SUM:= be9c3a697a54ac099c910d068ff0dc03 TARGET_OPTIMIZATION:= -Os -pipe diff --git a/target/qemu-cris/device.mk b/target/qemu-cris/device.mk index 3f4b7fbde..e74ad894a 100644 --- a/target/qemu-cris/device.mk +++ b/target/qemu-cris/device.mk @@ -1,6 +1,6 @@ ARCH:= cris CPU_ARCH:= crisv32 -KERNEL_VERSION:= 2.6.30.5 +KERNEL_VERSION:= 2.6.31.4 KERNEL_RELEASE:= 1 KERNEL_MD5SUM:= be9c3a697a54ac099c910d068ff0dc03 TARGET_OPTIMIZATION:= -Os -pipe diff --git a/target/qemu-mips/device.mk b/target/qemu-mips/device.mk index fce11a237..548d13759 100644 --- a/target/qemu-mips/device.mk +++ b/target/qemu-mips/device.mk @@ -1,6 +1,6 @@ ARCH:= mips CPU_ARCH:= mips -KERNEL_VERSION:= 2.6.30.5 +KERNEL_VERSION:= 2.6.31.4 KERNEL_RELEASE:= 1 KERNEL_MD5SUM:= be9c3a697a54ac099c910d068ff0dc03 TARGET_OPTIMIZATION:= -Os -pipe diff --git a/target/qemu-mips64/device.mk b/target/qemu-mips64/device.mk index 50d5211b0..db6ae9de4 100644 --- a/target/qemu-mips64/device.mk +++ b/target/qemu-mips64/device.mk @@ -1,6 +1,6 @@ ARCH:= mips CPU_ARCH:= mips64 -KERNEL_VERSION:= 2.6.30.5 +KERNEL_VERSION:= 2.6.31.4 KERNEL_RELEASE:= 1 KERNEL_MD5SUM:= be9c3a697a54ac099c910d068ff0dc03 TARGET_OPTIMIZATION:= -Os -pipe diff --git a/target/qemu-mips64el/device.mk b/target/qemu-mips64el/device.mk index 1157f2c45..36cad3e02 100644 --- a/target/qemu-mips64el/device.mk +++ b/target/qemu-mips64el/device.mk @@ -1,6 +1,6 @@ ARCH:= mips CPU_ARCH:= mips64el -KERNEL_VERSION:= 2.6.30.5 +KERNEL_VERSION:= 2.6.31.4 KERNEL_RELEASE:= 1 KERNEL_MD5SUM:= be9c3a697a54ac099c910d068ff0dc03 TARGET_OPTIMIZATION:= -Os -pipe diff --git a/target/qemu-mipsel/device.mk b/target/qemu-mipsel/device.mk index e888d79ff..8e6bb16ce 100644 --- a/target/qemu-mipsel/device.mk +++ b/target/qemu-mipsel/device.mk @@ -1,6 +1,6 @@ ARCH:= mips CPU_ARCH:= mipsel -KERNEL_VERSION:= 2.6.30.5 +KERNEL_VERSION:= 2.6.31.4 KERNEL_RELEASE:= 1 KERNEL_MD5SUM:= be9c3a697a54ac099c910d068ff0dc03 TARGET_OPTIMIZATION:= -Os -pipe diff --git a/target/qemu-x86/device.mk b/target/qemu-x86/device.mk index e29db1f01..3f64b3462 100644 --- a/target/qemu-x86/device.mk +++ b/target/qemu-x86/device.mk @@ -1,6 +1,6 @@ ARCH:= x86 CPU_ARCH:= i586 -KERNEL_VERSION:= 2.6.30.5 +KERNEL_VERSION:= 2.6.31.4 KERNEL_RELEASE:= 1 KERNEL_MD5SUM:= be9c3a697a54ac099c910d068ff0dc03 TARGET_OPTIMIZATION:= -Os -pipe diff --git a/target/rb411/device.mk b/target/rb411/device.mk index 93fd3c2cb..042f0cf1c 100644 --- a/target/rb411/device.mk +++ b/target/rb411/device.mk @@ -1,6 +1,6 @@ ARCH:= mips CPU_ARCH:= mips -KERNEL_VERSION:= 2.6.30.5 +KERNEL_VERSION:= 2.6.31.4 KERNEL_RELEASE:= 1 KERNEL_MD5SUM:= be9c3a697a54ac099c910d068ff0dc03 TARGET_OPTIMIZATION:= -Os -pipe diff --git a/target/rb433/device.mk b/target/rb433/device.mk index 93fd3c2cb..042f0cf1c 100644 --- a/target/rb433/device.mk +++ b/target/rb433/device.mk @@ -1,6 +1,6 @@ ARCH:= mips CPU_ARCH:= mips -KERNEL_VERSION:= 2.6.30.5 +KERNEL_VERSION:= 2.6.31.4 KERNEL_RELEASE:= 1 KERNEL_MD5SUM:= be9c3a697a54ac099c910d068ff0dc03 TARGET_OPTIMIZATION:= -Os -pipe diff --git a/target/rb532/device.mk b/target/rb532/device.mk index 664e4db13..8a7aceb79 100644 --- a/target/rb532/device.mk +++ b/target/rb532/device.mk @@ -1,6 +1,6 @@ ARCH:= mips CPU_ARCH:= mipsel -KERNEL_VERSION:= 2.6.30.5 +KERNEL_VERSION:= 2.6.31.4 KERNEL_RELEASE:= 1 KERNEL_MD5SUM:= be9c3a697a54ac099c910d068ff0dc03 TARGET_OPTIMIZATION:= -Os -pipe diff --git a/target/rescue-x86/device.mk b/target/rescue-x86/device.mk index 5ea885d6c..cd606e6bd 100644 --- a/target/rescue-x86/device.mk +++ b/target/rescue-x86/device.mk @@ -1,6 +1,6 @@ ARCH:= x86 CPU_ARCH:= i686 -KERNEL_VERSION:= 2.6.30.5 +KERNEL_VERSION:= 2.6.31.4 KERNEL_RELEASE:= 1 KERNEL_MD5SUM:= be9c3a697a54ac099c910d068ff0dc03 TARGET_OPTIMIZATION:= -O2 -pipe diff --git a/target/rescue-x86_64/device.mk b/target/rescue-x86_64/device.mk index 3241868c7..244b241f3 100644 --- a/target/rescue-x86_64/device.mk +++ b/target/rescue-x86_64/device.mk @@ -1,6 +1,6 @@ ARCH:= x86 CPU_ARCH:= x86_64 -KERNEL_VERSION:= 2.6.30.5 +KERNEL_VERSION:= 2.6.31.4 KERNEL_RELEASE:= 1 KERNEL_MD5SUM:= be9c3a697a54ac099c910d068ff0dc03 TARGET_OPTIMIZATION:= -O2 -pipe diff --git a/target/shuttle/device.mk b/target/shuttle/device.mk index f9c6224ae..4c4de9b78 100644 --- a/target/shuttle/device.mk +++ b/target/shuttle/device.mk @@ -1,6 +1,6 @@ ARCH:= x86 CPU_ARCH:= x86_64 -KERNEL_VERSION:= 2.6.30.5 +KERNEL_VERSION:= 2.6.31.4 KERNEL_RELEASE:= 1 KERNEL_MD5SUM:= be9c3a697a54ac099c910d068ff0dc03 TARGET_OPTIMIZATION:= -O2 -pipe diff --git a/target/tomtom/device.mk b/target/tomtom/device.mk index 5f13b2700..407eed2f8 100644 --- a/target/tomtom/device.mk +++ b/target/tomtom/device.mk @@ -1,6 +1,6 @@ ARCH:= arm CPU_ARCH:= armel -KERNEL_VERSION:= 2.6.30.5 +KERNEL_VERSION:= 2.6.31.4 KERNEL_RELEASE:= 1 KERNEL_MD5SUM:= be9c3a697a54ac099c910d068ff0dc03 TARGET_OPTIMIZATION:= -Os -pipe diff --git a/target/wag54g/device.mk b/target/wag54g/device.mk index 664e4db13..8a7aceb79 100644 --- a/target/wag54g/device.mk +++ b/target/wag54g/device.mk @@ -1,6 +1,6 @@ ARCH:= mips CPU_ARCH:= mipsel -KERNEL_VERSION:= 2.6.30.5 +KERNEL_VERSION:= 2.6.31.4 KERNEL_RELEASE:= 1 KERNEL_MD5SUM:= be9c3a697a54ac099c910d068ff0dc03 TARGET_OPTIMIZATION:= -Os -pipe |