summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2005-12-30need to make sure elf_find_dynamic(key) is big enough to handle 64bit hostsMike Frysinger
2005-12-30beautify utility buildingMike Frysinger
2005-12-30use errno.h not sys/errno.h as pointed out by sjhillMike Frysinger
2005-12-30Sync ftw w/ glibc, splitting ftw and ftw64Peter S. Mazinger
2005-12-29fix signed/unsigned warningsMike Frysinger
2005-12-29kill off some signedness/unused warningsMike Frysinger
2005-12-29sync whitespace with glibcMike Frysinger
2005-12-29touchup code a bit to prep for new weak systemMike Frysinger
2005-12-29beautifyMike Frysinger
2005-12-29hide the internal libc pthread_mutex funcsMike Frysinger
2005-12-29pointless whitespace touchupsMike Frysinger
2005-12-29hide some functions to cut down on relocationsMike Frysinger
2005-12-29make sure mq_timedsend() and mq_send() is always definedMike Frysinger
2005-12-29make sure mq_timedreceive() and mq_receive() is always definedMike Frysinger
2005-12-29hide __kernel functionsMike Frysinger
2005-12-28make sure we handle the (malloc(0)==NULL) case as Aubrey points out via the ↵Mike Frysinger
e-mail list
2005-12-28make sure we handle the (malloc(0)==NULL) case as Aubrey points out via the ↵Mike Frysinger
e-mail list
2005-12-28include grp.h just for funMike Frysinger
2005-12-28typecast away signed/unsigned warningsMike Frysinger
2005-12-27kill off unused warningMike Frysinger
2005-12-27typecast away a signed/unsigned warningMike Frysinger
2005-12-27cut pointless filesMike Frysinger
2005-12-27kill off minor unused warningsMike Frysinger
2005-12-27tweak prototypes to match the exported versionsMike Frysinger
2005-12-27fix salt type in setup_salt()Mike Frysinger
2005-12-27nlist should be unsigned, not signed, as it tracks # of elements in ↵Mike Frysinger
init_fini_list
2005-12-27revert debug "fix" since it isnt correct and previous version worksMike Frysinger
2005-12-27fix signed/unsigned warningMike Frysinger
2005-12-27kill off simple unused warningsMike Frysinger
2005-12-27kill off simple unused warningsMike Frysinger
2005-12-27kill off simple unused warningsMike Frysinger
2005-12-272005-12-15 Aubrey.Li <aubreylee@gmail.com> writes:Mike Frysinger
When I mounted nfs on my target, the kernel crashed. And I found it was caused by stack overflow. When I digged into it. I found the following issue. In the file "./uClibc/libc/inet/rpc/auth_unix.c" int max_nr_groups = sysconf (_SC_NGROUPS_MAX); gid_t gids[max_nr_groups]; And, NGROUPS_MAX is defined in the file "./linux-2.6.x/include/linux/limits.h" #define NGROUPS_MAX 65536 /* supplemental group IDs are available */ OK, here we can know max_nr_groups is assigned to 65536, that means a huge matrix "gids[65536] is in the function **authunix_create_default**. My method is doing it by malloc, the patch as follows.
2005-12-27kill off minor warningMike Frysinger
2005-12-27Aubrey writes:Mike Frysinger
When I mounted nfs on my target, the kernel crashed. And I found it was caused by stack overflow. When I digged into it. And I found not only "setgroups.c" but "getgroups.c" have the matrix (__kernel_gid_t kernel_groups[n]) on the stack which can be very large because "n" can be assigned to NGROUPS_MAX. And, NGROUPS_MAX is defined in the file "./linux-2.6.x/include/linux/limits.h" #define NGROUPS_MAX 65536 /* supplemental group IDs are available */ I also changed it to do malloc.
2005-12-27Jan-Benedict Glaw writes:Mike Frysinger
This patch is wrong and breaks (at least) building a static libc. The $(MISC_FNMATCH_OBJ) file still uses the .c extension, thus isn't built at all and eg. strip'ping the .c file fails, breaking the build.
2005-12-24Add new fnmatch code, for now mbstate_t and mbsrtowcs are depending on ↵Peter S. Mazinger
UCLIBC_HAS_LOCALE, making the new object about 5k, else it would be 10k
2005-12-24Move fnmatch.c to fnmatch_old.cPeter S. Mazinger
2005-12-23Build PIC objects if DOPIC is set, thx to pcjPeter S. Mazinger
2005-12-22Jie Zhang writes:Mike Frysinger
I think the loop was written for MMU case. When there is MMU, mmap () is used to allocate the stack. If one segment is already mapped, we should continue to see if we can use the next one. However, for NOMMU, malloc () is used. It's waste of CPU cycles to continue to try if it fails. Here is a new patch, which makes this change only for NOMMU.
2005-12-16Remove all trailing ; after *_alias and change 2 to use weak_alias instead ↵Peter S. Mazinger
of __attribute__ ...
2005-12-16Remove trailing ;Peter S. Mazinger
2005-12-16Remove trailing ; after *alias()Peter S. Mazinger
2005-12-16TypoPeter S. Mazinger
2005-12-16Adapt some missed prototypesPeter S. Mazinger
2005-12-16Forgot about the macros ;-(Peter S. Mazinger
2005-12-16memmove, exit are external iconv_mainPeter S. Mazinger
2005-12-16Correct C_SYMBOL_PREFIX handling for asm, supports only none or _Peter S. Mazinger
2005-12-16Hope it solves JBG's problem, why didn't it bail out here?Peter S. Mazinger
2005-12-16Macros are no good for jump relocs, hack to get rid of one introduced my new ↵Peter S. Mazinger
regex code
2005-12-16Use hidden towupper in new regexPeter S. Mazinger