Chris Leonard [Sun, 15 Sep 2013 19:06:27 +0000 (15:06 -0400)]
Add country_car field to LC_ADDRESS.
Richard Sandiford [Fri, 13 Sep 2013 23:31:10 +0000 (23:31 +0000)]
Fix some types in localedef.
Will Newton [Fri, 13 Sep 2013 08:26:02 +0000 (09:26 +0100)]
Add CVE-2013-4332 to NEWS.
Jia Liu [Fri, 6 Sep 2013 16:01:08 +0000 (00:01 +0800)]
sunrpc/rpc/types.h: fix OS X and FreeBSD build problems
When I build arm-linux-gcc on OS X, I find glibc will get a build error
in sunrpc/rpc/types.h, so I add __APPLE_CC__ to make OS X build OK.
For FreeBSD, Add __FreeBSD__ to make it build OK, too.
URL: http://sourceware.org/ml/libc-alpha/2013-09/msg00155.html
URL: http://sourceware.org/ml/libc-alpha/2013-09/msg00217.html
URL: http://sourceware.org/ml/libc-alpha/2013-09/msg00240.html
Signed-off-by: Jia Liu <proljc@gmail.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Ondřej Bílka [Wed, 11 Sep 2013 15:07:38 +0000 (17:07 +0200)]
Faster strchr implementation.
Will Newton [Wed, 11 Sep 2013 12:07:33 +0000 (13:07 +0100)]
benchtests: Rename argument to TIMING_INIT macro.
The TIMING_INIT macro currently sets the number of loop iterations
to 1000, which limits usefulness. Make the argument a clock
resolution value and multiply by 1000 in bench-skeleton.c instead
to allow easier reuse.
ChangeLog:
2013-09-11 Will Newton <will.newton@linaro.org>
* benchtests/bench-timing.h (TIMING_INIT): Rename ITERS
parameter to RES. Remove hardcoded 1000 value.
* benchtests/bench-skeleton.c (main): Pass RES parameter
to TIMING_INIT and multiply result by 1000.
Will Newton [Wed, 11 Sep 2013 14:12:41 +0000 (15:12 +0100)]
Mention closing 15855, 15856 and 15857 in NEWS.
Adhemerval Zanella [Wed, 11 Sep 2013 12:38:57 +0000 (07:38 -0500)]
Update powerpc-fpu ULPs.
Andreas Schwab [Wed, 11 Sep 2013 09:15:45 +0000 (11:15 +0200)]
Add O_TMPFILE to <fcntl.h>
Will Newton [Fri, 16 Aug 2013 11:54:29 +0000 (12:54 +0100)]
malloc: Check for integer overflow in memalign.
A large bytes parameter to memalign could cause an integer overflow
and corrupt allocator internals. Check the overflow does not occur
before continuing with the allocation.
ChangeLog:
2013-09-11 Will Newton <will.newton@linaro.org>
[BZ #15857]
* malloc/malloc.c (__libc_memalign): Check the value of bytes
does not overflow.
Will Newton [Fri, 16 Aug 2013 10:59:37 +0000 (11:59 +0100)]
malloc: Check for integer overflow in valloc.
A large bytes parameter to valloc could cause an integer overflow
and corrupt allocator internals. Check the overflow does not occur
before continuing with the allocation.
ChangeLog:
2013-09-11 Will Newton <will.newton@linaro.org>
[BZ #15856]
* malloc/malloc.c (__libc_valloc): Check the value of bytes
does not overflow.
Will Newton [Mon, 12 Aug 2013 14:08:02 +0000 (15:08 +0100)]
malloc: Check for integer overflow in pvalloc.
A large bytes parameter to pvalloc could cause an integer overflow
and corrupt allocator internals. Check the overflow does not occur
before continuing with the allocation.
ChangeLog:
2013-09-11 Will Newton <will.newton@linaro.org>
[BZ #15855]
* malloc/malloc.c (__libc_pvalloc): Check the value of bytes
does not overflow.
Ondřej Bílka [Tue, 10 Sep 2013 17:14:51 +0000 (19:14 +0200)]
Remove DO_NOT_USE_THIS conditionals.
Allan McRae [Tue, 10 Sep 2013 04:12:10 +0000 (14:12 +1000)]
Clarify documentation cross-reference
The end of the "Parsing of Floats" subsection currently reads:
The GNU C Library also provides '_l' versions of these functions,
which take an additional argument, the locale to use in conversion.
*Note Parsing of Integers::.
Split the final note as it is unrelated to the above comment and
reference it with "See also" instead.
Allan McRae [Tue, 10 Sep 2013 04:09:27 +0000 (14:09 +1000)]
Update pt_chown sections of the manual
The pt-chown binary is discussed in the "Running make install" section
without clarification of the needed configure option. Clarify this
and simplfy the discription which is already covered in the "Configuring
and compiling" section.
Maciej W. Rozycki [Mon, 9 Sep 2013 21:36:57 +0000 (22:36 +0100)]
Fix static-binary lazy FPU context allocation
Long ago static startup did not parse the auxiliary vector and therefore
could not get at any `AT_FPUCW' tag to check whether upon FPU context
allocation the kernel would use a FPU control word setting different to
that provided by the `__fpu_control' variable. Static startup therefore
always initialized the FPU control word, forcing immediate FPU context
allocation even for binaries that otherwise never used the FPU.
As from GIT commit
f8f900ecb9096ec47f5b7bb7626e29223c69061a static
startup supports parsing the auxiliary vector, so now it can avoid
explicit initialization of the FPU control word, just as can dynamic
startup, in the usual case where the setting written to the FPU control
word would be the same as the kernel uses. This defers FPU context
allocation until the binary itself actually pokes at the FPU.
Note that the `AT_FPUCW' tag is usually absent from the auxiliary vector
in which case _FPU_DEFAULT is assumed to be the kernel default.
Allan McRae [Mon, 9 Sep 2013 12:54:07 +0000 (22:54 +1000)]
Fix typo in strcoll example
Allan McRae [Mon, 9 Sep 2013 12:52:58 +0000 (22:52 +1000)]
Fix memory leak in stdlib/isomac.c
Allan McRae [Mon, 9 Sep 2013 12:50:41 +0000 (22:50 +1000)]
Fix memory leaks in libio on allocation failure
Allan McRae [Mon, 9 Sep 2013 10:23:35 +0000 (20:23 +1000)]
Fix nesting of ifdefs in netgroupcache.c
Fixes unclosed '{' if HAVE_SENDFILE is defined (BZ #15895).
Will Newton [Wed, 7 Aug 2013 18:56:21 +0000 (19:56 +0100)]
malloc: Add realloc test.
The current tests don't test the functionality of realloc in detail.
Add a new test for realloc that exercises some of the corner cases
that are not otherwise tested.
ChangeLog:
2013-09-09 Will Newton <will.newton@linaro.org>
* malloc/Makefile: Add tst-realloc to tests.
* malloc/tst-realloc.c: New file.
Allan McRae [Mon, 9 Sep 2013 02:51:29 +0000 (12:51 +1000)]
Update to latest versions of GPL-2.0 and LGPL-2.1
Pull copies of these files directly from the GNU website:
http://www.gnu.org/licenses/gpl-2.0.txt
http://www.gnu.org/licenses/lgpl-2.1.txt
Fixes the address of the Free Software Foundation (BZ #15844). Also
includes some minor formatting changes and corrects references to
the GNU "Library" General Public License.
Joseph Myers [Sat, 7 Sep 2013 15:58:42 +0000 (15:58 +0000)]
Mention --disable-versioning removal in NEWS.
David S. Miller [Fri, 6 Sep 2013 19:25:54 +0000 (15:25 -0400)]
Update Chinese (traditional) translations.
* po/zh_TW.po: Update Chinese (traditional) translation from
translation project.
Richard Sandiford [Fri, 6 Sep 2013 17:20:45 +0000 (17:20 +0000)]
Make localedef output generation use more logical interfaces.
Chris Leonard [Fri, 6 Sep 2013 15:55:30 +0000 (11:55 -0400)]
Fix trailing whitespace
Chris Leonard [Fri, 6 Sep 2013 15:47:05 +0000 (11:47 -0400)]
ChangeLog entry for Deduplicate country_car.
Chris Leonard [Fri, 6 Sep 2013 15:31:02 +0000 (11:31 -0400)]
Deduplicate country_car.
Adhemerval Zanella [Thu, 29 Aug 2013 18:38:51 +0000 (15:38 -0300)]
benchtests: Add memrchr benchmark
Will Newton [Wed, 4 Sep 2013 08:11:58 +0000 (09:11 +0100)]
benchtests/Makefile: Run benchmark for memcpy.
The benchmark for memcpy got disabled accidentally. Re-enable it.
ChangeLog:
2013-09-06 Will Newton <will.newton@linaro.org>
* benchtests/Makefile (string-bench): Add memcpy.
Carlos O'Donell [Fri, 6 Sep 2013 05:02:30 +0000 (01:02 -0400)]
Coordinate IPv6 definitions for Linux and glibc
This change synchronizes the glibc headers with the Linux kernel
headers and arranges to coordinate the definition of structures
already defined the Linux kernel UAPI headers.
It is now safe to include glibc's netinet/in.h or Linux's linux/in6.h
in any order in a userspace application and you will get the same
ABI. The ABI is guaranteed by UAPI and glibc.
Adhemerval Zanella [Thu, 5 Sep 2013 14:32:56 +0000 (09:32 -0500)]
PowerPC: fix POWER7 memrchr for some large inputs
Adhemerval Zanella [Thu, 29 Aug 2013 18:28:00 +0000 (15:28 -0300)]
Add memrchr testcase
Mike Frysinger [Thu, 5 Sep 2013 07:02:46 +0000 (03:02 -0400)]
tst-fanotify: skip when we get back EPERM
Since fanotify_init requires CAP_SYS_ADMIN in order to work (which usually
means running as root), we need to handle that error case too.
Reported-by: Andreas Jaeger <aj@suse.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Joseph Myers [Wed, 4 Sep 2013 20:50:20 +0000 (20:50 +0000)]
Define MMAP2_PAGE_SHIFT to -1 for m68k.
Joseph Myers [Wed, 4 Sep 2013 20:47:21 +0000 (20:47 +0000)]
conformtest: Fix namespace testing.
Chris Leonard [Wed, 4 Sep 2013 15:45:05 +0000 (11:45 -0400)]
Update iso-1427.def and related occurrences.
Chris Leonard [Wed, 4 Sep 2013 15:42:53 +0000 (11:42 -0400)]
Update iso-1427.def and related occurrences.
Joseph Myers [Wed, 4 Sep 2013 15:25:42 +0000 (15:25 +0000)]
Remove --disable-versioning.
Will Newton [Mon, 2 Sep 2013 13:13:50 +0000 (14:13 +0100)]
benchtests: Switch string benchmarks to use bench-timing.h.
Switch the string benchmarks to using bench-timing.h instead
of hp-timing.h directly. This allows the string benchmarks to
be run usefully on architectures such as ARM that do not have
support for hp-timing.h.
In order to do this the tests have been changed from timing each
individual call and picking the lowest execution time recorded to
timing a number of calls and taking the mean execution time.
ChangeLog:
2013-09-04 Will Newton <will.newton@linaro.org>
* benchtests/bench-timing.h (TIMING_PRINT_MEAN): New macro.
* benchtests/bench-string.h: Include bench-timing.h instead
of including hp-timing.h directly. (INNER_LOOP_ITERS): New
define. (HP_TIMING_BEST): Delete macro. (test_init): Remove
call to HP_TIMING_DIFF_INIT.
* benchtests/bench-memccpy.c: Use bench-timing.h macros
instead of hp-timing.h macros.
* benchtests/bench-memchr.c: Likewise.
* benchtests/bench-memcmp.c: Likewise.
* benchtests/bench-memcpy.c: Likewise.
* benchtests/bench-memmem.c: Likewise.
* benchtests/bench-memmove.c: Likewise.
* benchtests/bench-memset.c: Likewise.
* benchtests/bench-rawmemchr.c: Likewise.
* benchtests/bench-strcasecmp.c: Likewise.
* benchtests/bench-strcasestr.c: Likewise.
* benchtests/bench-strcat.c: Likewise.
* benchtests/bench-strchr.c: Likewise.
* benchtests/bench-strcmp.c: Likewise.
* benchtests/bench-strcpy.c: Likewise.
* benchtests/bench-strcpy_chk.c: Likewise.
* benchtests/bench-strlen.c: Likewise.
* benchtests/bench-strncasecmp.c: Likewise.
* benchtests/bench-strncat.c: Likewise.
* benchtests/bench-strncmp.c: Likewise.
* benchtests/bench-strncpy.c: Likewise.
* benchtests/bench-strnlen.c: Likewise.
* benchtests/bench-strpbrk.c: Likewise.
* benchtests/bench-strrchr.c: Likewise.
* benchtests/bench-strspn.c: Likewise.
* benchtests/bench-strstr.c: Likewise.
Will Newton [Mon, 2 Sep 2013 13:00:45 +0000 (14:00 +0100)]
benchtests/Makefile: Use LDLIBS instead of LDFLAGS.
LDFLAGS puts the library too early in the command line if --as-needed
is being used. Use LDLIBS instead.
ChangeLog:
2013-09-04 Will Newton <will.newton@linaro.org>
* benchtests/Makefile: Use LDLIBS instead of LDFLAGS.
cjl [Wed, 4 Sep 2013 14:26:16 +0000 (10:26 -0400)]
Add country_car field to LC_ADDRESS
cjl [Wed, 4 Sep 2013 14:23:58 +0000 (10:23 -0400)]
Add country_car field to LC_ADDRESS
cjl [Wed, 4 Sep 2013 13:20:48 +0000 (09:20 -0400)]
Add country_car field to LC_ADDRESS
cjl [Wed, 4 Sep 2013 00:38:23 +0000 (20:38 -0400)]
Add FSF statement to ayc_PE locale.
cjl [Wed, 4 Sep 2013 00:34:50 +0000 (20:34 -0400)]
Add FSF statement to ayc_PE locale.
cjl [Wed, 4 Sep 2013 00:29:44 +0000 (20:29 -0400)]
Add FSF statement to ayc_PE locale.
Joseph Myers [Tue, 3 Sep 2013 15:32:54 +0000 (15:32 +0000)]
Fix lgammaf spurious underflow (bug 15427).
Ondřej Bílka [Tue, 3 Sep 2013 14:21:38 +0000 (16:21 +0200)]
Add unaligned strcmp.
Chris Leonard [Tue, 3 Sep 2013 12:51:35 +0000 (14:51 +0200)]
Update ht_HT locale
2013-09-03 Chris Leonard <cjl@sugarlabs.org>
[BZ#15886]
* locales/ht_HT: 1.1 revision of ht_HT locale.
Chris Leonard [Tue, 3 Sep 2013 09:06:28 +0000 (11:06 +0200)]
Add quz_PE locale
2013-09-03 Chris Leonard <cjl@sugarlabs.org>
[BZ#15887]
* locales/quz_PE: New locale for quz_PE.
* SUPPORTED: Add quz_PE.
Siddhesh Poyarekar [Tue, 3 Sep 2013 03:59:01 +0000 (09:29 +0530)]
Mark success return value as volatile to work around rescheduling
Resolves #15921
The test case nptl/tst-cleanup2 fails on s390x and power6 due to
instruction sheduling in gcc. This was reported in gcc:
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58034
but it was concluded that gcc is allowed to assume that the first
argument to sprintf is a character array - NULL not being a valid
character array.
Mike Frysinger [Tue, 3 Sep 2013 03:13:57 +0000 (23:13 -0400)]
hppa: add fanotify_mark
Another example of all the 64bit arches getting the definition via a
common file, but the 32bit ones all adding it by themselves and hppa
was missed.
I'm not entirely sure about the usage of GLIBC_2.19 symbols here.
We'd like to backport this so people can use it, but it means we'd
be releasing a glibc-2.17/glibc-2.18 with a GLIBC_2.19 symbol in it.
But maybe it won't be a big deal since you'd only get that 2.19 ref
if you actually used the symbol ?
There hasn't been a glibc release where hppa worked w/out a bunch of
patches, so in reality there's only two distros that matter -- Gentoo
and Debian.
Reported-by: Jeroen Roovers <jer@gentoo.org>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Mike Frysinger [Wed, 21 Aug 2013 15:59:01 +0000 (11:59 -0400)]
tst-fanotify: new simple test
Basic test for the fanotify functions.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Joseph Myers [Mon, 2 Sep 2013 14:51:24 +0000 (14:51 +0000)]
Fix spurious jnf underflows (bug 14155).
Ondřej Bílka [Mon, 2 Sep 2013 09:24:36 +0000 (11:24 +0200)]
Fix typo.
Roland McGrath [Fri, 30 Aug 2013 18:16:52 +0000 (11:16 -0700)]
Use sfi_* macros in armv6t2 strlen.
Roland McGrath [Fri, 30 Aug 2013 17:43:39 +0000 (10:43 -0700)]
Make armv6t2 strlen work in ARM mode too.
Ondřej Bílka [Fri, 30 Aug 2013 16:09:30 +0000 (18:09 +0200)]
Fix then/than typos.
Ondřej Bílka [Fri, 30 Aug 2013 12:32:07 +0000 (14:32 +0200)]
Fix typos.
Will Newton [Thu, 29 Aug 2013 19:10:26 +0000 (20:10 +0100)]
ARM: Fix clone code when built for Thumb.
The mov lr, pc instruction will lose the Thumb bit from the return address
so use blx lr instead.
ports/ChangeLog.arm:
2013-08-30 Will Newton <will.newton@linaro.org>
[BZ #15909]
* sysdeps/unix/sysv/linux/arm/clone.S (__clone): Use blx
instead of mov lr, pc.
Will Newton [Fri, 9 Aug 2013 08:36:41 +0000 (09:36 +0100)]
sysdeps/arm/armv6t2/strlen.S: strlen implementation for armv6t2.
This implementation of strlen is faster than the armv6 version for
all string lengths greater than 1 on a Cortex-A15.
ports/ChangeLog.arm:
2013-08-09 Will Newton <will.newton@linaro.org>
* sysdeps/arm/armv6t2/strlen.S: New file.
Ondřej Bílka [Fri, 30 Aug 2013 08:14:37 +0000 (10:14 +0200)]
Fix rawmemchr regression on bulldozer.
Roland McGrath [Thu, 29 Aug 2013 21:15:45 +0000 (14:15 -0700)]
Fix up ChangeLog formatting.
Ondřej Bílka [Thu, 29 Aug 2013 16:10:58 +0000 (18:10 +0200)]
Remove aix specific files.
Thomas Schwinge [Thu, 29 Aug 2013 10:06:14 +0000 (12:06 +0200)]
Hurd: Add ESUCCESS error_t value.
Samuel Thibault [Sun, 24 Jun 2007 15:45:35 +0000 (17:45 +0200)]
Hurd: Use __executable_start symbol instead of _start.
_start points to the first instruction, not to the ELF header.
__executable_start does point on the ELF header.
Thomas Schwinge [Sun, 26 May 2013 18:59:55 +0000 (20:59 +0200)]
Support ELFOSABI_GNU on all GNU systems.
Thomas Schwinge [Sun, 26 May 2013 18:58:06 +0000 (20:58 +0200)]
Use ELFOSABI_GNU instead of ELFOSABI_LINUX.
Thomas Schwinge [Thu, 23 May 2013 16:00:10 +0000 (18:00 +0200)]
[BZ #15522] strtod ("nan(N)") returning a sNaN in some cases
Thomas Schwinge [Thu, 23 May 2013 15:31:42 +0000 (17:31 +0200)]
math: Additional type conversion tests
These have helped me find and fix type conversion issues in QEMU's MIPS
hardware emulation. While certainly glibc is not the best place for such
tests, they're just an enhancement of tests already present.
Ondřej Bílka [Thu, 29 Aug 2013 07:08:54 +0000 (09:08 +0200)]
Fix typos.
Carlos O'Donell [Thu, 29 Aug 2013 04:17:33 +0000 (00:17 -0400)]
ARM: Pass dl_hwcap to IFUNC resolver.
For REL relocs pass dl_hwcap to the IFUNC resolver
as is required by the IFUNC API (bug 15905).
Siddhesh Poyarekar [Wed, 28 Aug 2013 04:14:19 +0000 (09:44 +0530)]
Use __glibc_unlikely instead of __builtin_expect (..., 0)
Mike Frysinger [Mon, 26 Aug 2013 04:21:26 +0000 (00:21 -0400)]
[BZ #15897] dlfcn: do not mark dlopen/dlclose as leaf functions
Since the dlopen funcs might invoke a constructor that calls a func
that is in the same compilation unit as the caller, we cannot mark
them as leaf funcs.
Similarly, dlclose might invoke a destructor that calls a func that
is in the same compilation unit as the caller.
URL: https://sourceware.org/bugzilla/show_bug.cgi?id=15897
Reportedy-by: Fabrice Bauzac <libnoon@gmail.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Roland McGrath [Tue, 27 Aug 2013 18:09:33 +0000 (11:09 -0700)]
Clean up h_errno declaration to use __thread unconditionally.
Roland McGrath [Tue, 27 Aug 2013 18:04:46 +0000 (11:04 -0700)]
Clean up _res declaration to use __thread unconditionally.
Roland McGrath [Tue, 27 Aug 2013 17:59:57 +0000 (10:59 -0700)]
Cope without sunrpc.
Roland McGrath [Tue, 27 Aug 2013 17:50:06 +0000 (10:50 -0700)]
Don't try to use ioctl unless [FIONREAD].
Roland McGrath [Tue, 27 Aug 2013 17:49:36 +0000 (10:49 -0700)]
Mild decrufting in resolv code.
Roland McGrath [Tue, 27 Aug 2013 17:42:52 +0000 (10:42 -0700)]
True stub __ifreq.
Roland McGrath [Tue, 27 Aug 2013 17:34:16 +0000 (10:34 -0700)]
Clean up __libc_sa_len helper.
Roland McGrath [Tue, 27 Aug 2013 17:33:15 +0000 (10:33 -0700)]
Replace generic bits/socket.h with 4.4 file.
Roland McGrath [Tue, 27 Aug 2013 17:32:30 +0000 (10:32 -0700)]
Flesh out 4.4 bits/socket.h with SOCK_CLOEXEC, SOCK_NONBLOCK.
Andreas Schwab [Wed, 17 Jul 2013 08:26:58 +0000 (10:26 +0200)]
Fix missing declaration of LC_CTYPE nonascii-case element
Andreas Schwab [Mon, 29 Jul 2013 12:58:41 +0000 (14:58 +0200)]
Fix handling LC_CTYPE nonascii-case fallback in i686 SSE4.2 and SSSE3 strcasecmp/strncasecmp
Roland McGrath [Mon, 26 Aug 2013 22:11:47 +0000 (15:11 -0700)]
Make stub lxstat64 call xstat64, like stub lxstat calls xstat.
Roland McGrath [Mon, 26 Aug 2013 22:04:02 +0000 (15:04 -0700)]
Use proper #include for xdecrypt declarations.
Siddhesh Poyarekar [Mon, 26 Aug 2013 10:12:29 +0000 (15:42 +0530)]
Initialize res_hconf in nscd
Fixes BZ #15890.
Andreas Schwab [Wed, 21 Aug 2013 08:19:01 +0000 (10:19 +0200)]
Fix dependencies for stdlib/tst-tls-atexit
Siddhesh Poyarekar [Mon, 26 Aug 2013 05:23:02 +0000 (10:53 +0530)]
Fix indentation in aicache.c
Mike Frysinger [Sun, 25 Aug 2013 20:01:52 +0000 (16:01 -0400)]
configure: add missing quotes in $build_pt_chown test
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Joseph Myers [Fri, 23 Aug 2013 19:45:38 +0000 (19:45 +0000)]
Fix cexp (NaN + i0) (bug 15532).
David S. Miller [Thu, 22 Aug 2013 23:35:51 +0000 (16:35 -0700)]
Add Ukrainian translations.
* po/uk.po: Add Ukrainian translations from translation project.
David S. Miller [Thu, 22 Aug 2013 23:30:34 +0000 (16:30 -0700)]
Update Catalan translations.
* po/ca.po: Update Catalan translation from translation project.
Maciej W. Rozycki [Thu, 22 Aug 2013 16:50:20 +0000 (17:50 +0100)]
MIPS: Correct the handling of reserved FCSR bits
Reserved bits in the Floating-Point Control and Status Register (FCSR)
should not be implicitly cleared by fedisableexcept or feenableexcept,
there is no reason to. Among these are the 8 condition codes and one of
the two bits reserved for architecture implementers (bits #22 & #21).
As to the latter, there is no reason to treat any of them as reserved
either, they should be user controllable and settable via __fpu_control
override as the user sees fit. For example in processors implemented by
MIPS Technologies, such as the 5Kf or the 24Kf, these bits are used to
change the treatment of denormalised operands and tiny results: bit #22
is Flush Override (FO) and bit #21 is Flush to Nearest (FN). They cause
non-IEEE-compliant behaviour, but some programs may have a use for such
modes of operation; the library should not obstruct such use just as it
does not for the architectural Flush to Zero (FS) bit (bit #24).
Therefore the change adjusts the reserved mask accordingly and also
documents the distinction between bits 22:21 and 20:18.
Joseph Myers [Wed, 21 Aug 2013 19:58:42 +0000 (19:58 +0000)]
Add bug 15867 to NEWS.
Joseph Myers [Wed, 21 Aug 2013 19:56:48 +0000 (19:56 +0000)]
Fix fdim handling of infinities (bug 15797).
Ondřej Bílka [Wed, 21 Aug 2013 17:48:48 +0000 (19:48 +0200)]
Fix typos.
Liubov Dmitrieva [Wed, 21 Aug 2013 14:24:44 +0000 (18:24 +0400)]
i686: Skip SSE4_2 version for strcmp, strncmp, strncase, strcasecmp
if bit_Slow_SSE4_2 is set.