OSDN Git Service

uclinux-h8/uclibc-ng.git
15 years ago- arm eabi needs __UCLIBC_TRUNCATE64_HAS_4_ARGS__
Bernhard Reutner-Fischer [Thu, 25 Sep 2008 08:50:56 +0000 (08:50 -0000)]
- arm eabi needs __UCLIBC_TRUNCATE64_HAS_4_ARGS__
  (75_all_uClibc-arm-ftruncate64.patch from gentoo)

15 years ago- ARMV5 can use STRD and LDRD access instructions but these accesses need to be
Bernhard Reutner-Fischer [Thu, 25 Sep 2008 08:35:20 +0000 (08:35 -0000)]
- ARMV5 can use STRD and LDRD access instructions but these accesses need to be
  8 byte aligned. The dynamic linker's malloc needs to match this so structures
  become 8 byte aligned to void unaligned accesses.
  (71_all_uClibc-arm_fix_alignment.patch from gentoo)

15 years agosh: Fix SYSCALL_INST_STRx to work for SH-2[A] core by using the
Carmelo Amoroso [Thu, 25 Sep 2008 06:58:20 +0000 (06:58 -0000)]
sh: Fix SYSCALL_INST_STRx to work for SH-2[A] core by using the
__SH_SYSCALL_TRAP_BASE macro.

Signed-off-by: Carmelo Amoroso <carmelo.amoroso@st.com>
15 years agoAdd globally __stringify macro in libc-symbols.h.
Carmelo Amoroso [Thu, 25 Sep 2008 06:38:21 +0000 (06:38 -0000)]
Add globally __stringify macro in libc-symbols.h.
Remove all other duplicated definitions.

Signed-off-by: Carmelo Amoroso <carmelo.amoroso@st.com>
15 years agodelete/init module functions are not 2.4 specific
Mike Frysinger [Thu, 25 Sep 2008 01:21:34 +0000 (01:21 -0000)]
delete/init module functions are not 2.4 specific

15 years ago- only compile them if the respective syscalls are available
Bernhard Reutner-Fischer [Wed, 24 Sep 2008 16:31:29 +0000 (16:31 -0000)]
- only compile them if the respective syscalls are available

15 years ago- silence warning about possibly undefined preprocessor token
Bernhard Reutner-Fischer [Wed, 24 Sep 2008 16:15:01 +0000 (16:15 -0000)]
- silence warning about possibly undefined preprocessor token

15 years ago- Fix arguments to compare function. Closes #1081
Bernhard Reutner-Fischer [Wed, 24 Sep 2008 15:30:59 +0000 (15:30 -0000)]
- Fix arguments to compare function. Closes #1081

15 years ago- remove debug cruft
Bernhard Reutner-Fischer [Wed, 24 Sep 2008 15:19:14 +0000 (15:19 -0000)]
- remove debug cruft

15 years ago- deal with ARM specific syscalls. Closes #2894
Bernhard Reutner-Fischer [Wed, 24 Sep 2008 15:17:12 +0000 (15:17 -0000)]
- deal with ARM specific syscalls. Closes #2894

15 years ago- both getopt gnu and susv3 want {bits/,}getopt.h installed. Closes #1297
Bernhard Reutner-Fischer [Wed, 24 Sep 2008 14:45:57 +0000 (14:45 -0000)]
- both getopt gnu and susv3 want {bits/,}getopt.h installed. Closes #1297

15 years ago- do not segfault if we end up with empty tables
Bernhard Reutner-Fischer [Wed, 24 Sep 2008 11:57:47 +0000 (11:57 -0000)]
- do not segfault if we end up with empty tables
- silence some warnings while at it

15 years agoUse __SH_SYSCALL_TRAP_BASE for pthread vfork trap, as it will
Paul Mundt [Wed, 24 Sep 2008 10:49:32 +0000 (10:49 -0000)]
Use __SH_SYSCALL_TRAP_BASE for pthread vfork trap, as it will
change depending on ABI.

15 years agoldso: removed dead code.
Carmelo Amoroso [Tue, 23 Sep 2008 15:07:43 +0000 (15:07 -0000)]
ldso: removed dead code.

Signed-off-by: Carmelo Amoroso <carmelo.amoroso@st.com>
15 years agoAdd work-around for powerpc.
Carmelo Amoroso [Mon, 22 Sep 2008 13:16:55 +0000 (13:16 -0000)]
Add work-around for powerpc.

Do noy use syscall wrapper based on INTERNAL_SYSCALL.
A proper fix would require an arch specific implementation.

Signed-off-by: Carmelo Amoroso <carmelo.amoroso@st.com>
15 years agoFix up memset() argument ordering in open_memstream(). Previously
Paul Mundt [Sun, 21 Sep 2008 05:08:59 +0000 (05:08 -0000)]
Fix up memset() argument ordering in open_memstream(). Previously
parts of the buffers were not being zeroed out as expected. Reported
by Dmytro Gorbunov <dmitro.gorbunov@gmail.com>.

15 years ago- fix for r23427. Thanks to psm for mentioning this and sorry for the breakage
Bernhard Reutner-Fischer [Fri, 19 Sep 2008 19:45:45 +0000 (19:45 -0000)]
- fix for r23427. Thanks to psm for mentioning this and sorry for the breakage

15 years ago- add posix_fadvise. Not yet tested most likely broken in advise64
Bernhard Reutner-Fischer [Fri, 19 Sep 2008 19:43:57 +0000 (19:43 -0000)]
- add posix_fadvise. Not yet tested most likely broken in advise64

15 years ago- take LFS into account and add fallback via fadvise64
Bernhard Reutner-Fischer [Fri, 19 Sep 2008 10:31:02 +0000 (10:31 -0000)]
- take LFS into account and add fallback via fadvise64

15 years ago- posix_fadvise64 for i386
Bernhard Reutner-Fischer [Fri, 19 Sep 2008 09:54:32 +0000 (09:54 -0000)]
- posix_fadvise64 for i386

15 years agoUse __always_inline for _dl_gettimeofday as well.
Carmelo Amoroso [Thu, 18 Sep 2008 15:10:12 +0000 (15:10 -0000)]
Use __always_inline for _dl_gettimeofday as well.

Signed-off-by: Carmelo Amoroso <carmelo.amoroso@st.com>
15 years agoFix posix_fadvise[64] functions to return the error number in case
Carmelo Amoroso [Thu, 18 Sep 2008 15:06:24 +0000 (15:06 -0000)]
Fix posix_fadvise[64] functions to return the error number in case
of failure instead of -1 and setting errno, according to SuSv3
(IEEE Std 1003.1 2004 edition) specification.

Signed-off-by: Filippo Arcidiacono <filippo.arcidiacono@st.com>
Signed-off-by: Carmelo Amoroso <carmelo.amoroso@st.com>
15 years ago- remove non stdc path (old style definitions). No obj-code changes
Bernhard Reutner-Fischer [Thu, 18 Sep 2008 12:08:55 +0000 (12:08 -0000)]
- remove non stdc path (old style definitions). No obj-code changes

15 years ago- exec_comm_child can not return ever
Bernhard Reutner-Fischer [Thu, 18 Sep 2008 12:03:40 +0000 (12:03 -0000)]
- exec_comm_child can not return ever

15 years agoavr32: add defines and prototype function for sync_file_range syscall
Hans-Christian Egtvedt [Wed, 17 Sep 2008 11:45:37 +0000 (11:45 -0000)]
avr32: add defines and prototype function for sync_file_range syscall

This patch adds the needed defines and function prototype for the
sync_file_range syscall on AVR32 architecture.

Signed-off-by: Hans-Christian Egtvedt <hans-christian.egtvedt@atmel.com>
15 years ago- add script to generate lots of failing (random)configs
Bernhard Reutner-Fischer [Tue, 16 Sep 2008 16:07:35 +0000 (16:07 -0000)]
- add script to generate lots of failing (random)configs
  rm -f STOP ; ARCH=i386 nohup extra/scripts/randconfig.sh &
  if you want to stop it, touch STOP

15 years ago- sync_file_range(). Missed the i386 directory in previous commit
Bernhard Reutner-Fischer [Tue, 16 Sep 2008 14:18:48 +0000 (14:18 -0000)]
- sync_file_range(). Missed the i386 directory in previous commit

15 years agoAdd INTERNAL_SYSCALL macros for sh
Carmelo Amoroso [Tue, 16 Sep 2008 14:10:30 +0000 (14:10 -0000)]
Add INTERNAL_SYSCALL macros for sh

15 years ago- add sync_file_range()
Bernhard Reutner-Fischer [Tue, 16 Sep 2008 14:02:54 +0000 (14:02 -0000)]
- add sync_file_range()
  Tested successfully on i386..

15 years ago- do not attempt to rm twice
Bernhard Reutner-Fischer [Tue, 16 Sep 2008 08:19:07 +0000 (08:19 -0000)]
- do not attempt to rm twice
- clean targets are supposed to clean independently of any .config

15 years ago- commentary typo fix
Bernhard Reutner-Fischer [Mon, 15 Sep 2008 16:36:11 +0000 (16:36 -0000)]
- commentary typo fix

15 years agoMake menuconfig help bigger and fluffier.
Rob Landley [Mon, 15 Sep 2008 10:24:22 +0000 (10:24 -0000)]
Make menuconfig help bigger and fluffier.

15 years agoxtensa: Stack unwinding on Xtensa requires that the return address register always...
Chris Zankel [Sat, 13 Sep 2008 20:26:06 +0000 (20:26 -0000)]
xtensa: Stack unwinding on Xtensa requires that the return address register always be valid.  This patch fixes a part of ldso that had been using an Xtensa CALL0 instruction to get the current PC value, which temporarily clobbered the return address register.  The revised code just loads the address of a label in the code.  See my change to GCC from 2007-12-20 for more info (gcc.gnu.org/ml/gcc-patches/2007-12/msg01023.html).

15 years agoxtensa: The Xtensa port needs to define the __UCLIBC_TRUNCATE64_HAS_4_ARGS__ macro...
Chris Zankel [Sat, 13 Sep 2008 20:22:33 +0000 (20:22 -0000)]
xtensa: The Xtensa port needs to define the __UCLIBC_TRUNCATE64_HAS_4_ARGS__ macro so that a dummy argument will be inserted before the long long length argument, which is expected to be in an even/odd register pair.

-This line, and those below, will be ignored--

M    uClibc_arch_features.h

15 years agoxtensa: The Xtensa-specific strcmp.S file is testing the __UCLIBC_HAS_LOCALE__ macro...
Chris Zankel [Sat, 13 Sep 2008 19:54:56 +0000 (19:54 -0000)]
xtensa: The Xtensa-specific strcmp.S file is testing the __UCLIBC_HAS_LOCALE__ macro but it does not include <features.h>, so that macro is never defined.  This causes problems with strcoll when locale support is enabled.

15 years ago- fix typo
Bernhard Reutner-Fischer [Fri, 12 Sep 2008 09:21:07 +0000 (09:21 -0000)]
- fix typo

15 years ago- whitespace fixes
Bernhard Reutner-Fischer [Thu, 11 Sep 2008 16:42:23 +0000 (16:42 -0000)]
- whitespace fixes

15 years ago- remove gethostbyname_r relocation
Bernhard Reutner-Fischer [Thu, 11 Sep 2008 16:27:41 +0000 (16:27 -0000)]
- remove gethostbyname_r relocation

15 years ago- remove posix_openpt relocation
Bernhard Reutner-Fischer [Thu, 11 Sep 2008 16:17:43 +0000 (16:17 -0000)]
- remove posix_openpt relocation

15 years ago- fix typo
Bernhard Reutner-Fischer [Thu, 11 Sep 2008 16:12:50 +0000 (16:12 -0000)]
- fix typo

15 years ago- remove relocations for clock_getres
Bernhard Reutner-Fischer [Thu, 11 Sep 2008 15:48:00 +0000 (15:48 -0000)]
- remove relocations for clock_getres

15 years ago- the help target is phony
Bernhard Reutner-Fischer [Thu, 11 Sep 2008 15:44:23 +0000 (15:44 -0000)]
- the help target is phony

15 years ago- add some helpful help text, as advertised.
Bernhard Reutner-Fischer [Thu, 11 Sep 2008 14:52:05 +0000 (14:52 -0000)]
- add some helpful help text, as advertised.

15 years ago- honour EXTRA_CPPFLAGS
Bernhard Reutner-Fischer [Thu, 11 Sep 2008 11:29:39 +0000 (11:29 -0000)]
- honour EXTRA_CPPFLAGS
  Makes it easier to say something like make libc/sysdeps/linux/common/close.i EXTRA_CPPFLAGS=-P

15 years agoAdd optimized memcpy implementation for sh4 (from Stuart Menefy @STMicroelectronics).
Carmelo Amoroso [Tue, 9 Sep 2008 16:55:27 +0000 (16:55 -0000)]
Add optimized memcpy implementation for sh4 (from Stuart Menefy @STMicroelectronics).
This implementation is based on 'backward copying'.

Signed-off-by: Carmelo Amoroso <carmelo.amoroso@st.com>
15 years agoFix some locale multibyte tests failures ad below:
Carmelo Amoroso [Tue, 9 Sep 2008 13:01:58 +0000 (13:01 -0000)]
Fix some locale multibyte tests failures ad below:

libc/stdlib/_strtod.c   -> tst_wcstod;
libc/stdlib/stdlib.c    -> tst_mblen, tst_mbtowc, tst_wctomb;
libc/stdio/_scanf.c     -> tst_swscanf;
libc/string/strncmp.c   -> tst_wcsncmp;
libc/misc/wchar/wchar.c -> tst_mbrlen, tst_mbrtowc, tst_wcswidth.

Signed-off-by: Filippo Arcidiacono <filippo.arcidiacono@st.com>
Signed-off-by: Carmelo Amoroso <carmelo.amoroso@st.com>
15 years ago- adjust range of valid grp buffer size to match pwd buffer size
Bernhard Reutner-Fischer [Tue, 9 Sep 2008 12:58:57 +0000 (12:58 -0000)]
- adjust range of valid grp buffer size to match pwd buffer size

15 years agoFix problem in vfprintf when assertion are enabled.
Carmelo Amoroso [Tue, 9 Sep 2008 12:19:19 +0000 (12:19 -0000)]
Fix problem in vfprintf when assertion are enabled.
Here follow the output of my analysis:
__stdio_fwrite expects at least 1 bytes. fputs_unlocked(S,F) calls
fwrite_unlocked and this calls __stdio_fwrite only if bytes to be
written are > 0, otherwise simply returs 0 (that is correct).
During the parsing of format spec it could happen that __stdio_fwrite
is called passing an empty string and with assertion enabled it will abort.

Signed-off-by: Carmelo Amoroso <carmelo.amoroso@st.com>
15 years agoFill in termios_p data structure only if ioctl syscall
Carmelo Amoroso [Tue, 9 Sep 2008 12:10:17 +0000 (12:10 -0000)]
Fill in termios_p data structure only if ioctl syscall
doesn't fail (as glibc indeed does).

Signed-off-by: Carmelo Amoroso <carmelo.amoroso@st.com>
15 years agoHush compiler for extern inline warnings by using
Carmelo Amoroso [Tue, 9 Sep 2008 12:06:58 +0000 (12:06 -0000)]
Hush compiler for extern inline warnings by using
__extern_inline macro, this also makes gcc 4.3 happy.
(Taken from NPTL branch)

Signed-off-by: Khem Raj <raj.khem@gmail.com>
15 years agoRevert latest changes. Instead of applying this work-around
Carmelo Amoroso [Tue, 9 Sep 2008 12:04:00 +0000 (12:04 -0000)]
Revert latest changes. Instead of applying this work-around
a proper fix should ensure that statically linked mutli-threaded
binaries take always the proper locking implementation.

Signed-off-by: Carmelo Amoroso <carmelo.amoroso@st.com>
15 years agoUse getdents syscall if kernel provide supports for this
Carmelo Amoroso [Tue, 9 Sep 2008 12:01:37 +0000 (12:01 -0000)]
Use getdents syscall if kernel provide supports for this
instead of relying upon getdents64.

Signed-off-by: Filippo Arcidiacono <filippo.arcidiacono@st.com>
Signed-off-by: Carmelo Amoroso <carmelo.amoroso@st.com>
15 years ago- for extern inline, also take __extern_inline and __GNUC_GNU_INLINE__ into account
Bernhard Reutner-Fischer [Sat, 6 Sep 2008 16:45:07 +0000 (16:45 -0000)]
- for extern inline, also take __extern_inline and __GNUC_GNU_INLINE__ into account

15 years ago- fix indentation
Bernhard Reutner-Fischer [Sat, 6 Sep 2008 16:39:39 +0000 (16:39 -0000)]
- fix indentation

15 years agoFollow sh asm convention to indent the instruction into the delay slot.
Carmelo Amoroso [Thu, 4 Sep 2008 09:42:02 +0000 (09:42 -0000)]
Follow sh asm convention to indent the instruction into the delay slot.
(glibc community refused to apply a my patch for this :-) )

15 years agoKill off the GOT lookup for init/fini in the sh crti.S. This was
Paul Mundt [Thu, 4 Sep 2008 09:24:12 +0000 (09:24 -0000)]
Kill off the GOT lookup for init/fini in the sh crti.S.  This was
inversely conditional on __HAVE_SHARED__, which has been broken since
the time it was copied in originally. The compiler properly generates
the GOT references on its own, negating the entire purpose for the
init/fini reference.

15 years agoMove ifaddrs.h from uClibc-internal location libc/inet/ to include/, and make
Ricard Wanderlof [Thu, 28 Aug 2008 07:33:17 +0000 (07:33 -0000)]
Move ifaddrs.h from uClibc-internal location libc/inet/ to include/, and make
it conditionally included among the installed header files depending on
UCLIBC_HAS_AI_ADDRCONFIG.

15 years ago- remove a couple of duplicate includes
Bernhard Reutner-Fischer [Wed, 27 Aug 2008 15:06:41 +0000 (15:06 -0000)]
- remove a couple of duplicate includes

15 years ago- flesh out helptext of DOMULTI.
Bernhard Reutner-Fischer [Wed, 27 Aug 2008 14:59:15 +0000 (14:59 -0000)]
- flesh out helptext of DOMULTI.
  Put short, just leave it alone or keep all parts --
  alternatively go help with lto if you have time.

15 years ago- should either be defined or callers should do_t_strip manually afterwards
Bernhard Reutner-Fischer [Wed, 27 Aug 2008 14:42:52 +0000 (14:42 -0000)]
- should either be defined or callers should do_t_strip manually afterwards

15 years agoRemoved redundant includes.
Ricard Wanderlof [Wed, 27 Aug 2008 13:52:12 +0000 (13:52 -0000)]
Removed redundant includes.

15 years ago- pull check-lxdialog.sh from linux-2.6.26
Bernhard Reutner-Fischer [Wed, 27 Aug 2008 07:32:11 +0000 (07:32 -0000)]
- pull check-lxdialog.sh from linux-2.6.26

16 years agoFix up an msync() redefinition for linuxthreads + nommu systems,
Paul Mundt [Thu, 21 Aug 2008 09:01:36 +0000 (09:01 -0000)]
Fix up an msync() redefinition for linuxthreads + nommu systems,
common case depends on __ARCH_USE_MMU__, falling back on the
include/sys/mman.h:msync() stub otherwise.

16 years agoDisable -msoft-float usage on SH, gcc doesn't support it.
Paul Mundt [Thu, 21 Aug 2008 08:59:27 +0000 (08:59 -0000)]
Disable -msoft-float usage on SH, gcc doesn't support it.

16 years ago- Fix powerpc pread/pwrite (Wade Berrier)
Bernhard Reutner-Fischer [Wed, 20 Aug 2008 07:42:33 +0000 (07:42 -0000)]
- Fix powerpc pread/pwrite (Wade Berrier)
  pread/write wanted to send an 8 byte value in the syscall where only 4 bytes
  were allowed.

16 years agoFix SH ldso sequence startup.
Carmelo Amoroso [Fri, 8 Aug 2008 07:28:02 +0000 (07:28 -0000)]
Fix SH ldso sequence startup.

Pass via r4 the rtld finalizer
_dl_fini to the user application. This will be the 6^ arg of
__uClibc_main and will be registered with 'atexit'.
In this way the dynamic linker will be able to call destructors
defined within the loaded DSOs.

Signed-off-by: Carmelo Amoroso <carmelo.amoroso@st.com>
Acked-by: Paul Mundt <lethal@linux-sh.org>
Add missing 7th arg "stack_end".
add comment of undocumented usage of r4.
fix comment of expected __uClibc_main() prototype.

Signed-off-by: Yoshii Takashi <yoshii.takashi@renesas.com>
Acked-by: Paul Mundt <lethal@linux-sh.org>
16 years ago- some more asm keyword fixes (Yann E. MORIN)
Bernhard Reutner-Fischer [Wed, 6 Aug 2008 08:17:47 +0000 (08:17 -0000)]
- some more asm keyword fixes (Yann E. MORIN)

16 years agoRemove .depend when doing distclean
Carmelo Amoroso [Tue, 29 Jul 2008 09:40:59 +0000 (09:40 -0000)]
Remove .depend when doing distclean

16 years ago- trim any trailing whitespace
Bernhard Reutner-Fischer [Wed, 23 Jul 2008 11:23:36 +0000 (11:23 -0000)]
- trim any trailing whitespace

16 years ago- fix inline keyword
Bernhard Reutner-Fischer [Wed, 23 Jul 2008 11:19:00 +0000 (11:19 -0000)]
- fix inline keyword

16 years ago- fix asm and volatile keywords
Bernhard Reutner-Fischer [Wed, 23 Jul 2008 10:59:18 +0000 (10:59 -0000)]
- fix asm and volatile keywords

16 years agoavr32: Put underscores around asm, inline, etc.
Haavard Skinnemoen [Wed, 23 Jul 2008 08:05:45 +0000 (08:05 -0000)]
avr32: Put underscores around asm, inline, etc.

This fixes build breakage introduced in r22868.

16 years ago- poison non-standard keywords:
Bernhard Reutner-Fischer [Thu, 17 Jul 2008 16:54:31 +0000 (16:54 -0000)]
- poison non-standard keywords:
  Make sure that we can be built with non-C99 compilers, too.
  Use __\1__ instead.

16 years ago- pull fix for unwinding from the NPTL branch (Johannes Stezenbach)
Bernhard Reutner-Fischer [Thu, 17 Jul 2008 15:52:43 +0000 (15:52 -0000)]
- pull fix for unwinding from the NPTL branch (Johannes Stezenbach)
  Repair C89 compliance and severe whitespace-defects while at it.

16 years ago- improve UCLIBC_LINUX_SPECIFIC
Bernhard Reutner-Fischer [Mon, 14 Jul 2008 14:30:48 +0000 (14:30 -0000)]
- improve UCLIBC_LINUX_SPECIFIC
  compile-tested only, fixes
  libc/misc/statfs/fstatfs64.c:29: error: 'fstatfs' undeclared here (not in a function)

16 years ago- do not force LFS unto x86_64 folks
Bernhard Reutner-Fischer [Mon, 14 Jul 2008 14:16:31 +0000 (14:16 -0000)]
- do not force LFS unto x86_64 folks
  vapier, you added this in r12384. What was the reasoning back then?

16 years ago- remove the error counter
Bernhard Reutner-Fischer [Fri, 11 Jul 2008 18:56:17 +0000 (18:56 -0000)]
- remove the error counter

16 years ago- add error-counter, do not immediately exit if a testcase fails but print
Bernhard Reutner-Fischer [Fri, 11 Jul 2008 17:22:26 +0000 (17:22 -0000)]
- add error-counter, do not immediately exit if a testcase fails but print
  the overall error-counter at the end and exit accordingly.

16 years ago- do not test pthread if threads are not available
Bernhard Reutner-Fischer [Fri, 11 Jul 2008 16:45:19 +0000 (16:45 -0000)]
- do not test pthread if threads are not available

16 years ago- the interp is an order-only prereq of lib-so-y
Bernhard Reutner-Fischer [Fri, 11 Jul 2008 16:23:38 +0000 (16:23 -0000)]
- the interp is an order-only prereq of lib-so-y
- hint vi on the filetype while at it

16 years ago- don't test libcrypt stubs
Bernhard Reutner-Fischer [Fri, 11 Jul 2008 16:16:04 +0000 (16:16 -0000)]
- don't test libcrypt stubs

16 years ago- do not run tests that are not available
Bernhard Reutner-Fischer [Fri, 11 Jul 2008 16:02:24 +0000 (16:02 -0000)]
- do not run tests that are not available

16 years agowow. So why does test/ keep it's own copy of CFLAGS guessing etc?
Bernhard Reutner-Fischer [Fri, 11 Jul 2008 15:55:39 +0000 (15:55 -0000)]
wow. So why does test/ keep it's own copy of CFLAGS guessing etc?

16 years agocheck and test have to depend on test_compile
Bernhard Reutner-Fischer [Fri, 11 Jul 2008 15:54:48 +0000 (15:54 -0000)]
check and test have to depend on test_compile

16 years agoRemoved wrong preprocessed files
Carmelo Amoroso [Thu, 10 Jul 2008 10:40:45 +0000 (10:40 -0000)]
Removed wrong preprocessed files

16 years agoAdded missing tests for locale with multibyte encoding
Carmelo Amoroso [Thu, 10 Jul 2008 10:30:28 +0000 (10:30 -0000)]
Added missing tests for locale with multibyte encoding

16 years agoAdded several tests for locale support (8 bit and multibyte UTF-8)
Carmelo Amoroso [Wed, 9 Jul 2008 15:05:36 +0000 (15:05 -0000)]
Added several tests for locale support (8 bit and multibyte UTF-8)

Basically all tests have been taken from glibc. For testing multibyte encoding
EUC_JP parts have been commented out and added new section for UTF-8
that is the only multibyte  codeset currently supported on uCLibc.

Some tests are still failing due to unsupported/missing features, other have been
fixed.

Signed-off-by: Filippo Arcidiacono <filippo.arcidiacono@st.com>
Signed-off-by: Carmelo Amoroso <carmelo.amoroso@st.com>
16 years agosh4 don't need to cope with older RLIMIT implementation.
Carmelo Amoroso [Tue, 8 Jul 2008 10:37:05 +0000 (10:37 -0000)]
sh4 don't need to cope with older RLIMIT implementation.

Signed-off-by: Carmelo Amoroso <carmelo.amoroso@st.com>
16 years agoFix the test build system by installing headers on a local folder
Carmelo Amoroso [Tue, 8 Jul 2008 08:05:44 +0000 (08:05 -0000)]
Fix the test build system by installing headers on a local folder
instead of using internal headers.

Signed-off-by: Filippo Arcidiacono <filippo.arcidiacono@st.com>
Signed-off-by: Carmelo Amoroso <carmelo.amoroso@st.com>
16 years agoNew .config option UCLIBC_HAS_STUBS, enables fork() stub on NOMMU
Denis Vlasenko [Tue, 8 Jul 2008 02:41:21 +0000 (02:41 -0000)]
New .config option UCLIBC_HAS_STUBS, enables fork() stub on NOMMU

16 years agoInclude errno.h from kernel headers instead of using internal errno_values.h
Carmelo Amoroso [Wed, 2 Jul 2008 16:26:23 +0000 (16:26 -0000)]
Include errno.h from kernel headers instead of using internal errno_values.h
otherwie some errno will be missed (i.e. ENOKEY used in libusb)

Signed-off-by: Carmelo Amoroso <carmelo.amoroso@st.com>
16 years agoFix makefile target to run test when there are some shell script.
Carmelo Amoroso [Wed, 2 Jul 2008 15:45:57 +0000 (15:45 -0000)]
Fix makefile target to run test when there are some shell script.
SImply do not include SHEL_TESTS among RUNTIME_TESTS, because
shell script have a their own rule to be excuted.
The runtime evaluation by using the $(shell ...) command doesn't work
due to immediate expansion of shell function.
Currently only nptl tests have shell script, so this problem have been never
discovered before.

Signed-off-by: Filippo Arcidiacono <filippo.arcidiacono@st.com>
Signed-off-by: Carmelo Amoroso <carmelo.amoroso@st.com>
16 years agoSimplified check_pf() so it returns a bit vector in an unsigned int,
Ricard Wanderlof [Tue, 1 Jul 2008 12:54:49 +0000 (12:54 -0000)]
Simplified check_pf() so it returns a bit vector in an unsigned int,
instead of modifying the contents of two bools.

16 years ago#include <ifaddrs.h> should be #include "ifaddrs.h"
Denis Vlasenko [Sat, 28 Jun 2008 06:37:30 +0000 (06:37 -0000)]
#include <ifaddrs.h> should be #include "ifaddrs.h"

16 years agoadd test case for argv[0] == null
Mike Frysinger [Fri, 27 Jun 2008 19:13:06 +0000 (19:13 -0000)]
add test case for argv[0] == null

16 years agoif argv[0] is NULL, dont try to parse it and subsequently crash ... found by running...
Mike Frysinger [Fri, 27 Jun 2008 19:12:50 +0000 (19:12 -0000)]
if argv[0] is NULL, dont try to parse it and subsequently crash ... found by running execve05 ltp test on blackfin ([#4194])

16 years agoRemoved include/ifaddrs.h and put contents in libc/inet/ifaddrs.h, as we do not
Ricard Wanderlof [Fri, 27 Jun 2008 15:38:41 +0000 (15:38 -0000)]
Removed include/ifaddrs.h and put contents in libc/inet/ifaddrs.h, as we do not
need the file outside of uClibc anyway.

16 years agoFixed #if -> #if defined.
Ricard Wanderlof [Fri, 27 Jun 2008 15:29:48 +0000 (15:29 -0000)]
Fixed #if -> #if defined.
Take __UCLIBC_HAS_IPV4__ into account.

16 years agoAdded support for the AI_ADDRCONFIG flag in the hints->ai_flags parameter to getaddri...
Ricard Wanderlof [Fri, 27 Jun 2008 09:08:44 +0000 (09:08 -0000)]
Added support for the AI_ADDRCONFIG flag in the hints->ai_flags parameter to getaddrinfo(3).

16 years ago- make sure to build crti and crtn with PICFLAG
Bernhard Reutner-Fischer [Fri, 20 Jun 2008 09:12:42 +0000 (09:12 -0000)]
- make sure to build crti and crtn with PICFLAG