| Age | Commit message (Collapse) | Author |
|---|
|
Signed-off-by: Natanael Copa <natanael.copa@gmail.com>
Signed-off-by: Austin Foxley <austinf@cetoncorp.com>
|
|
Conflicts:
libc/misc/utmp/utent.c
libc/sysdeps/linux/i386/bits/syscalls.h
Signed-off-by: Austin Foxley <austinf@cetoncorp.com>
|
|
so far
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
|
|
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
|
|
constraints
Example: time() doesn't push/pop two registers now, and is smaller:
<__GI_time>:
-57 push %edi
-8b 7c 24 08 mov 0x8(%esp),%edi
-53 push %ebx
-89 fb mov %edi,%ebx
+8b 4c 24 04 mov 0x4(%esp),%ecx
+87 cb xchg %ecx,%ebx
b8 0d 00 00 00 mov $0xd,%eax
cd 80 int $0x80
-5b pop %ebx
-5f pop %edi
+87 cb xchg %ecx,%ebx
c3 ret
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
|
|
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
|
|
Signed-off-by: Austin Foxley <austinf@cetoncorp.com>
|
|
Signed-off-by: Austin Foxley <austinf@cetoncorp.com>
|
|
Signed-off-by: Austin Foxley <austinf@cetoncorp.com>
|
|
Most ports have the same exact mman bit defines, so let's unify things
like the linux kernel has with the asm-generic efforts.
A few ports are left behind as they are non-trivial to merge -- the arch
maintainers can tackle it if they care.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
|
|
* RESET_PID support for clone.S
* SAVE_PID, RESTORE_PID in vfork.S
* fixup syscall assembly constraints to be a little less restrictive
allows arbitrary variables to be used as the syscall nr
Signed-off-by: Austin Foxley <austinf@cetoncorp.com>
|
|
The common define has the same behavior, just formatted differently.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
|
|
Drop the "#ifndef O_CLOEXEC" cruft, enable O_CLOEXEC in most fcntl.h
headers, and import __ASSUME_O_CLOEXEC from glibc.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
|
|
sed -i -e '/Experimentally off - /d' $(grep -rl "Experimentally off - " *)
sed -i -e '/^\/\*[[:space:]]*libc_hidden_proto(/d' $(grep -rl "libc_hidden_proto" *)
should be a nop
Signed-off-by: Bernhard Reutner-Fischer <rep.dot.nop@gmail.com>
|
|
This fixes compilation errors on hosts that turn off long double support
for C99 like powerpc32.
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Bernhard Reutner-Fischer <rep.dot.nop@gmail.com>
|
|
Avoid including akefile.commonarch in each Makefile.arch.
Include it instead from Makefile.in just after the arch specific
Makefile.arch
Signed-off-by: Carmelo Amoroso <carmelo.amoroso@st.com>
|
|
Fill out the stat structure so that the nanosecond resolution support is
always available. There is a small code size increase for a few ports
(three additional assignments in xstatconv), but otherwise everything
should remain the same.
While we're here, punt __old_kernel_stat from the few headers that still
define it as it is unused in uClibc and causes compile errors after these
nanosecond changes.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
|
|
The majority of the byteswap functions are the same across all arches, so
setup a common header to provide definitions if they don't exist. This
allows arches to override only the ones they actually want to implement
with inline assembly.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
|
|
Undo my old mistake. I added UCLIBC_INTERNAL define,
but later I realized _LIBC is doing exactly the same thing.
This change converts all usages of UCLIBC_INTERNAL to _LIBC,
removing all instances of UCLIBC_INTERNAL.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
|
|
Declare common NCS (non-constant syscall) variants and convert the existing
ports over to this.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
|
|
Unify all the common syscall defines in syscalls-common.h and scrub all
the duplicated code from relevant ports. This should also make converting
existing ports to INLINE_SYSCALL() much easier as they don't have to get
lost in all the unrelated noise, as well as creating new ports.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
|
|
The __syscall_error() function stores the errno value in the edx register
before invoking the __set_errno() macro. When using the pthread library
this macro calls thread_self() to determine the errno location, which might
clobber the edx register. The errno value must be stored in a "real"
variable so the compiler can take care of saving/restoring it if necessary.
Signed-off-by: Ingo van Lil <inguin@gmx.de>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
|
|
Make sure each arch has the same complete list to make comparing between
them easier.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
|
|
ustat: same, + remove superfluous "conversion"
which does nothing at best, loses high dev bits at worst
i386/bits/kernel_stat.h: update to reflect reality (wider dev_t)
h8300/bits/kernel_stat.h: same
arm/bits/kernel_stat.h: cosmetics to match kernel header to the letter
|
|
update example
libc/sysdeps/linux/i386/bits/mathinline.h:
improve __finite() macro, add __finitef macro
(why they aren't always macros? why aren't they arch independent?)
libm/math_private.h:
much better comments on math_opt_barrier() and math_force_eval()
libm/s_finite[f].c:
improve out-of-line __finite[f]() too (one byte less, yay...)
|
|
test/Rules.mak: don't pass target options to host build
test/math/basic-test.c: fix test build error
patching file test/unistd/errno.c: fix test build warning
|
|
with linux >= 2.6.28)
|
|
|
|
and x86_64 into x86
|
|
and google says only uclibc has it, no users at all.
text data bss dec hex filename
- 229374 1367 11280 242021 3b165 lib/libuClibc-0.9.30-svn.so
+ 229347 1367 11280 241994 3b14a lib/libuClibc-0.9.30-svn.so
|
|
Run tested on i386.
|
|
"Bounds Checking Projects... This project has been abandoned"
for four years at least.
|
|
help it. Code size is the same.
|
|
text data bss dec hex filename
- 44 0 4 48 30 libc/sysdeps/linux/i386/brk.o
+ 42 0 4 46 2e libc/sysdeps/linux/i386/brk.o
|
|
fix _NSIG for it.
better document what's going on in sigaction().
seems to not induce any actual code changes (sans mips).
|
|
|
|
|
|
|
|
|
|
|
|
In issue #5554 Michael wrote:
The implementation of sigaction on i386 for older kernels makes the system call using an inline asm element with two flaws:
1. The asm is not marked as depending on the kact structure or modifying the koact structure. Thus, GCC is free to assume these structures need not be kept consistent, allowing it to remove all initialization of kact.
2. The asm allows the signal number to be provided as a memory reference. But this allows GCC to provide a stack-relative operand, which will break because the assembler saves %ebx on the stack before using that operand.
1 didn't use to be a problem in practice because GCC 4.2.* didn't seize the optimization opportunity. GCC 4.3.2, however, optimizes out the "kact.sa_flags = act->sa_flags | SA_RESTORER;" line, so that the kernel sees garbage in sa_flags. This can result in the kernel seeing the SA_RESETHAND flag, causing erratic behaviour in signal dependent programs.
2 becomes an issue if "-fomit-frame-pointer" is provided. In uClibc-0.9.29 it isn't, uClibc-0.9.30-rc2 does use the flag by default.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Tested successfully on i386..
|
|
|
