this
will improve performance even on targets which don't have an optimized strlen. It is about twice
as
fast as the original strncat in bench-strncat.
implementation, so this improves performance even on targets which don't have an optimized
strlen and strcpy - it is 25% faster in bench-strcat. On targets which don't provide an
optimized strcat but which do have an optimized strlen and strcpy, performance gain is > 2x.
Continuing the move of syscall definitions to syscalls.list, where the
removal of support for old kernel versions has made this possible,
this patch moves various definitions of chown, lchown and fchown.
In most cases the need for special syscalls.list entries (rather than
existing generic ones) is because these architectures use chown32,
lchown32 and fchown32 as syscall names. Some architectures also have
symbol versioning compatibility for older versions of chown having
been equivalent to lchown.
The aliases specified for s390-32 had the effect of exporting
__chown@@GLIBC_2.1 (but not __chown@GLIBC_2.0) despite it not being
listed in Versions files. (I'm not sure why versioned_symbol but not
compat_symbol were effective like that to create such __chown exports
in the absence of Versions entries.) The natural way to preserve that
versioned export of __chown seems to be to add it in a Versions file,
so I did so. (Maybe actually it should be a compat symbol,
__chown@GLIBC_2.1, unless there's a good reason for that export, but
this patch doesn't change anything there.)
Tested for x86.
[BZ #14138]
* sysdeps/unix/sysv/linux/i386/chown.c: Remove file.
* sysdeps/unix/sysv/linux/i386/fchown.c: Likewise.
* sysdeps/unix/sysv/linux/i386/lchown.c: Likewise.
* sysdeps/unix/sysv/linux/s390/s390-32/chown.c: Likewise.
* sysdeps/unix/sysv/linux/s390/s390-32/fchown.c: Remove file.
* sysdeps/unix/sysv/linux/s390/s390-32/lchown.c: Likewise.
* sysdeps/unix/sysv/linux/sh/chown.c: Likewise.
* sysdeps/unix/sysv/linux/sh/fchown.c: Likewise.
* sysdeps/unix/sysv/linux/sh/lchown.c: Likewise.
* sysdeps/unix/sysv/linux/sparc/sparc32/chown.c: Likewise.
* sysdeps/unix/sysv/linux/sparc/sparc32/fchown.c: Likewise.
* sysdeps/unix/sysv/linux/sparc/sparc32/lchown.c: Likewise.
* sysdeps/unix/sysv/linux/s390/s390-32/Versions (GLIBC_2.1): Add
__chown.
* sysdeps/unix/sysv/linux/i386/syscalls.list (chown): Add syscall.
(lchown): Likewise.
(fchown): Likewise.
* sysdeps/unix/sysv/linux/s390/s390-32/syscalls.list (chown):
Likewise.
(lchown): Likewise.
(fchown): Likewise.
* sysdeps/unix/sysv/linux/sh/syscalls.list (chown): Likewise.
(lchown): Likewise.
(fchown): Likewise.
* sysdeps/unix/sysv/linux/sparc/sparc32/syscalls.list (chown):
Likewise.
(lchown): Likewise.
(fchown): Likewise.
In the Linux kernel version 3.17 the signal numbers were rearranged in
order to make hppa like every other arch. Previously we started
__SIGRTMIN at 37, and that meant several pieces of important software,
including systemd, would fail to build. To support systemd we removed
SIGEMT and SIGLOST, and rearranged the others according to expected
values. This is technically an ABI incompatible change, but because
zero applications use SIGSTKFLT, SIGXCPU, SIGXFSZ and SIGSYS nothing
broke. Nothing uses SIGEMT and SIGLOST, and they were present for
HPUX compatibility which is no longer supported. Thus because nothing
breaks we don't do any compatibility work here.
Upstream kernel commit is 1f25df2eff5b25f52c139d3ff31bc883eee9a0ab.
Signed-off-by: Carlos O'Donell <carlos@systemhalted.org>
Signed-off-by: Helge Deller <deller@gmx.de>
2014-10-23 Carlos O'Donell <carlos@systemhalted.org>
Helge Deller <deller@gmx.de>
[BZ #17508]
* sysdeps/unix/sysv/linux/hppa/bits/signum.h: Remove SIGEMT.
Define SIGSTKFLT as 7. Define SIGSYS as 31. Define SIGXCPU as 12.
Remove SIGLOST. Define SIGXFSZ as 30. Define __SIGRTMIN as 32.
Continuing the removal of the obsolete INTDEF / INTUSE mechanism, this
patch removes the use of INTUSE to rename symbols in
sysdeps/powerpc/powerpc32/libgcc-compat.S. As the names in question
are purely internal to this particular object and not used anywhere
else, it doesn't matter at all whether __*_v_glibc20 or __*_internal
is used, so this patch just removes the macros in question.
Tested for powerpc32 that stripped installed shared libraries are
unchanged by this patch.
[BZ #14132]
* sysdeps/powerpc/powerpc32/libgcc-compat.S (__ashldi3_v_glibc20):
Remove macro definition.
(__ashrdi3_v_glibc20): Likewise.
(__lshrdi3_v_glibc20): Likewise.
(__cmpdi2_v_glibc20): Likewise.
(__ucmpdi2_v_glibc20): Likewise.
[!_SOFT_FLOAT && !__NO_FPRS__] (__fixdfdi_v_glibc20): Likewise.
[!_SOFT_FLOAT && !__NO_FPRS__] (__fixsfdi_v_glibc20): Likewise.
[!_SOFT_FLOAT && !__NO_FPRS__] (__fixunsdfdi_v_glibc20): Likewise.
[!_SOFT_FLOAT && !__NO_FPRS__] (__fixunssfdi_v_glibc20): Likewise.
[!_SOFT_FLOAT && !__NO_FPRS__] (__floatdidf_v_glibc20): Likewise.
[!_SOFT_FLOAT && !__NO_FPRS__] (__floatdisf_v_glibc20): Likewise.
This satisfies a symbol reference created with:
.symver __libc_vfork, vfork@GLIBC_2.0
where `__libc_vfork' has not been defined or referenced. In this case
the `vfork@GLIBC_2.0' reference is supposed to be discarded, however a
bug present in GAS since forever causes an undefined symbol table entry
to be created. This in turn triggers a problem in the linker that can
manifest itself by link errors such as:
ld: libpthread.so: invalid string offset 2765592330 >= 5154 for section `.dynstr'
The GAS and linker bugs need to be resolved, but we can avoid them too
by providing a `__libc_vfork' definition just like our other platforms.
[BZ #17485]
* sysdeps/unix/sysv/linux/mips/vfork.S (__libc_vfork): Define.
Continuing the removal of the obsolete INTDEF / INTUSE mechanism, this
patch removes the use of INTDEF for __ldexpf. As far as I can tell,
the resulting alias is completely unused.
Tested for x86_64 that stripped installed shared libraries are
unchanged by this patch.
[BZ #14132]
* math/s_ldexpf.c (__ldexpf): Do not use INTDEF.
Continuing the move of syscall definitions to syscalls.list, where the
removal of support for old kernel versions has made this possible,
this patch moves definitions of readv and writev.
The relevant syscalls.list entries were already in
sysdeps/unix/syscalls.list, but to match the C files they needed to
have the names __libc_readv and __libc_writev added. In fact, I don't
see anything making use of those names - as far as I can tell, these
functions could just be defined as __readv and __writev with aliases
readv and writev. But cleaning up unnecessary aliases for functions
should be a separate matter from cleaning up unnecessary C syscall
wrappers.
Tested for x86_64.
[BZ #14138]
* sysdeps/unix/syscalls.list (readv): Use __libc_readv as strong
name.
(writev): Use __libc_writev as strong name.
* sysdeps/unix/sysv/linux/readv.c: Remove file.
* sysdeps/unix/sysv/linux/writev.c: Likewise.
this is a path that should solve bug 15884. It complains about the performance
of strcoll(). It was found out that the runtime of strcoll() is actually bound
to strlen which is needed for calculating the size of a cache that was
installed to improve the comparison performance.
The idea for this patch was that the cache is only useful in rare cases
(strings of same length and same first-level-chars) and that it would be
better to avoid memory allocation at all. To prove this I wrote a performance
test bench-strcoll.c with test data in benchtests-strcoll.tar.gz. Also
modifications in benchtests/Makefile and localedata/Makefile are necessary to
make it work.
After removing the cache the strcoll method showed the predicted behavior
(getting slightly faster) in all but the test case for hindi word sorting.
This was due the hindi text having much more equal words than the other ones.
For equal strings the performance was worse since all comparison levels were
run through and from the second level on the cache improved the comparison
performance of the original version.
Therefore I added a bytewise test via strcmp iff the first level comparison
found that both strings did match because in this case it is very likely that
equal strings are compared. This solved the problem with the hindi test case
and improved the performance of the others.
Performance comparison:
glibc files -33.77%
vi_VN.UTF-8 -34.12%
en_US.UTF-8 -42.42%
ar_SA.UTF-8 -27.49%
zh_CN.UTF-8 +07.90%
cs_CZ.UTF-8 -29.67%
en_GB.UTF-8 -28.50%
da_DK.UTF-8 -36.57%
pl_PL.UTF-8 -39.31%
fr_FR.UTF-8 -28.57%
pt_PT.UTF-8 -22.82%
el_GR.UTF-8 -26.77%
ru_RU.UTF-8 -35.81%
iw_IL.UTF-8 -35.34%
es_ES.UTF-8 -34.46%
hi_IN.UTF-8 -00.38%
sv_SE.UTF-8 -36.99%
hu_HU.UTF-8 -16.35%
tr_TR.UTF-8 -27.80%
is_IS.UTF-8 -33.24%
it_IT.UTF-8 -24.39%
sr_RS.UTF-8 -37.55%
ja_JP.UTF-8 +02.84%
Continuing the series of patches to clean up conformtest expectations
for "POSIX" (1995/6) based on review of the expectations against the
standard, this patch cleans up expectations for sys/utsname.h and
sys/wait.h. Tested x86_64; a new XFAIL for sys/wait.h is added.
* conform/data/sys/utsname.h-data (*_t): Allow.
* conform/data/sys/wait.h-data [POSIX] (uid_t): Do not define.
[POSIX] (WEXITED): Do not expect constant.
[POSIX] (WSTOPPED): Likewise.
[POSIX] (WNOHANG): Likewise.
[POSIX] (WNOWAIT): Likewise.
[POSIX] (siginfo_t): Do not expect type or elements.
[POSIX] (pid_t): Do not expect type.
[POSIX] (signal.h): Do not allow header.
[POSIX] (sys/resource.h): Likewise.
[POSIX] (si_*): Do not allow pattern.
[POSIX] (W*): Likewise.
[POSIX] (P_*): Likewise.
[POSIX] (BUS_*): Likewise.
[POSIX] (CLD_*): Likewise.
[POSIX] (FPE_*): Likewise.
[POSIX] (ILL_*): Likewise.
[POSIX] (POLL_*): Likewise.
[POSIX] (SEGV_*): Likewise.
[POSIX] (SI_*): Likewise.
[POSIX] (TRAP_*): Likewise.
* conform/Makefile (test-xfail-POSIX/sys/wait.h/conform): New
variable.
The recvmsg could return 0 under some conditions and cause the
make_request function to be stuck in an infinite loop.
Thank you Jim King <jim.king@simplivity.com> for posting Paul's patch
on the list.
As far as I can tell, CANCEL-FCT-WAIVE and CANCEL-FILE-WAIVE are old
notes from the addition of cancellation support to glibc and are not
currently used by any glibc testcases or otherwise in the build
process, and it does not seem useful to me to keep them around. This
patch removes them.
Tested for x86_64.
* CANCEL-FCT-WAIVE: Remove file.
* CANCEL-FILE-WAIVE: Likewise.
Continuing the removal of the obsolete INTDEF / INTVARDEF / INTUSE
mechanism, this patch replaces its use for __libc_enable_secure with
the use of rtld_hidden_data_def and rtld_hidden_proto.
Tested for x86_64 that installed stripped shared libraries are
unchanged by the patch.
[BZ #14132]
* elf/dl-sysdep.c (__libc_enable_secure): Use rtld_hidden_data_def
instead of INTVARDEF.
(_dl_sysdep_start): Do not use INTUSE with __libc_enable_secure.
* sysdeps/mach/hurd/dl-sysdep.c (__libc_enable_secure): Use
rtld_hidden_data_def instead of INTVARDEF.
(_dl_sysdep_start): Do not use INTUSE with __libc_enable_secure.
* elf/dl-deps.c (expand_dst): Likewise.
* elf/dl-load.c (_dl_dst_count): Likewise.
(_dl_dst_substitute): Likewise.
(decompose_rpath): Likewise.
(_dl_init_paths): Likewise.
(open_path): Likewise.
(_dl_map_object): Likewise.
* elf/rtld.c (dl_main): Likewise.
(process_dl_audit): Likewise.
(process_envvars): Likewise.
* include/unistd.h [IS_IN_rtld] (__libc_enable_secure_internal):
Remove declaration.
(__libc_enable_secure): Use rtld_hidden_proto.
Continuing the addition of soft-fp features in the Linux kernel
version, this patch adds _FP_TO_INT support for rsigned == 2 (reduce
overflowing results modulo 2^rsize to fit in the destination, used for
alpha emulation).
The kernel version is buggy; it can left shift by a negative amount
when right shifting is required in an overflow case (the kernel
version also has other bugs fixed long ago in glibc; at least,
spurious exceptions converting to the most negative integer). This
version avoids that by handling overflow (other than to 0) for rsigned
== 2 along with the normal non-overflow case, which already properly
determines the direction in which to shift.
Tested for powerpc-nofpu. Some functions get slightly bigger and some
get slightly smaller, no doubt as a result of the change to where in
the macro "inexact" is raised, but I don't think those changes are
significant. Also tested for powerpc-nofpu with the relevant __fix*
functions changed to use rsigned == 2 (which is after all just as
valid as rsigned == 1 in IEEE terms), including verifying the results
and exceptions for various cases of conversions.
With these seven patches, the one remaining feature to add for the
soft-fp code to have all the features of the kernel version is
_FP_TO_INT_ROUND.
* soft-fp/op-common.h (_FP_TO_INT): Handle rsigned == 2.
As previously discussed
<https://sourceware.org/ml/libc-alpha/2013-10/msg00345.html>, it would
be desirable to be able to use the same version of the soft-fp code in
the Linux kernel as well as in glibc and libgcc (instead of an old
version in the kernel that's missing ten years of bug fixes,
performance improvements and new features), and to that end it is
useful to add to glibc's copy features in the kernel's copy, even when
they are not directly useful in glibc.
To that end, this patch adds one of those features: support for more
precise "invalid" exceptions describing the particular kind of invalid
operation. These are relevant for powerpc emulation, and are also as
described in IEEE 754-2008 as sub-exceptions.
The set of sub-exceptions here is the union of those supported on
powerpc and those from IEEE 754-2008 (the former adds a distinction
between 0/0 and Inf/Inf; the latter adds a distinction between Inf*0
from multiplication and the same from fma). This includes
sub-exceptions for sqrt, conversions to integer and comparisons that
are not supported in the kernel; I see no obvious reason for these
being missing from the kernel support, given that they are supported
on powerpc so accurate powerpc emulation should generate them.
Tested for powerpc-nofpu that the disassembly of installed shared
libraries is unchanged by this patch.
* soft-fp/soft-fp.h (FP_EX_INVALID_SNAN): New macro.
(FP_EX_INVALID_IMZ): Likewise.
(FP_EX_INVALID_IMZ_FMA): Likewise.
(FP_EX_INVALID_ISI): Likewise.
(FP_EX_INVALID_ZDZ): Likewise.
(FP_EX_INVALID_IDI): Likewise.
(FP_EX_INVALID_SQRT): Likewise.
(FP_EX_INVALID_CVI): Likewise.
(FP_EX_INVALID_VC): Likewise.
* soft-fp/op-common.h (_FP_UNPACK_CANONICAL): Specify more precise
"invalid" exceptions.
(_FP_CHECK_SIGNAN_SEMIRAW): Likewise.
(_FP_ADD_INTERNAL): Likewise.
(_FP_MUL): Likewise.
(_FP_FMA): Likewise.
(_FP_DIV): Likewise.
(_FP_CMP_CHECK_NAN): Likewise.
(_FP_SQRT): Likewise.
(_FP_TO_INT): Likewise.
(FP_EXTEND): Likewise.
This patch removes use of the obsolete INTDEF/INTUSE mechanism for
__cxa_atexit, replacing it with libc_hidden_def/libc_hidden_proto.
Tested for x86_64 that installed stripped shared libraries are
unchanged by the patch.
[BZ #14132]
* stdlib/cxa_atexit.c (__cxa_atexit): Use libc_hidden_def instead
of INTDEF.
* include/stdlib.h (__cxa_atexit_internal): Remove declaration.
(__cxa_atexit): Use libc_hidden_proto.
[!NOT_IN_libc] (__cxa_atexit): Remove macro definition.
This patch removes some stray (unused) *_internal aliases, and
function prototypes with no corresponding definitions at all, at least
some of which were missed in previous INTDEF / INTUSE removal.
Not removed in this patch: __canonicalize_directory_name_internal,
noticed in the course of preparing this patch, isn't an alias, but an
actual function in sysdeps/mach/hurd/getcwd.c - apparently unused,
however.
Tested for x86_64 that installed stripped shared libraries are
unchanged by this patch.
[BZ #14132]
* include/wctype.h [!_ISOMAC] (__iswalpha_l_internal): Remove
declaration.
[!_ISOMAC] (__iswdigit_l_internal): Likewise.
[!_ISOMAC] (__iswspace_l_internal): Likewise.
[!_ISOMAC] (__iswxdigit_l_internal): Likewise.
[!_ISOMAC] (__iswctype_internal): Likewise.
* stdio-common/siglist.c (_sys_siglist_internal): Remove alias.
* sysdeps/unix/syscalls.list (chown): Remove __chown_internal
alias.
(fcntl): Remove __fcntl_internal alias.
* sysdeps/unix/sysv/linux/hppa/syscalls.list (connect): Remove
__connect_internal alias.
* sysdeps/unix/sysv/linux/sparc/sparc64/syscalls.list (connect):
Likewise.
Continuing the addition of soft-fp features used in the Linux kernel,
this patch adds soft-fp support for FP_DENORM_ZERO (flushing input
subnormal operands to zero of the same sign).
There are some differences from the kernel version. In the kernel,
the "inexact" exception is set when flushing to zero. This does not
appear to match the documented semantics for either of the
architectures (alpha and sh) for which the kernel uses FP_DENORM_ZERO,
so this patch does not set "inexact" in this case. More operations
now use raw or semi-raw unpacking for optimization than did in the
ten-year-old soft-fp version in the kernel, so checks of
FP_DENORM_ZERO are inserted in those operations. They are also
inserted for comparisons (which already used raw unpacking in the old
version) as I believe that's the correct thing to do when input
subnormals are flushed to zero. They are *not* inserted for _FP_NEG.
(If any processors do flush input subnormals to zero for negation, or
otherwise vary from the rules implemented when FP_DENORM_ZERO is set,
further macros for sfp-machine.h to control this may need to be
added.)
Although the addition for comparisons will cause FP_EX_DENORM to be
set in this case, it still won't be set for comparisons involving
subnormals when not flushed to zero. It's quite possible that
accurate emulation of processors that have such an exception for
subnormal operands will require further changes relating to when
FP_EX_DENORM is set (in general, the support for things defined by
IEEE should be considered more reliable and mature than the support
for things outside the scope of IEEE floating point).
Although some processors also have a mode for abrupt underflow -
producing zeroes instead of output subnormals - there is no such mode
in the kernel's soft-fp, so no such mode is added to glibc's soft-fp
(although it could be if someone wanted to emulate such processor
support).
Tested for powerpc-nofpu that the disassembly of installed shared
libraries is unchanged by this patch.
* soft-fp/soft-fp.h (FP_DENORM_ZERO): New macro.
* soft-fp/op-common.h (_FP_UNPACK_CANONICAL): Check
FP_DENORM_ZERO.
(_FP_CHECK_FLUSH_ZERO): New macro.
(_FP_ADD_INTERNAL): Call _FP_CHECK_FLUSH_ZERO.
(_FP_CMP): Likewise.
(_FP_CMP_EQ): Likewise.
(_FP_TO_INT): Do not set inexact for subnormal arguments if
FP_DENORM_ZERO.
(FP_EXTEND): Call _FP_CHECK_FLUSH_ZERO.
(FP_TRUNC): Likewise.
This patch fixes a latent bug in _FP_TO_INT regarding handling of
arguments with maximum exponent (infinities and NaNs). If the maximum
exponent is below that calculated as an overflow threshold, such
values would incorrectly be treated as normal values for the purposes
of the conversion. This could not occur for any of the conversions
actually occurring in glibc, libgcc or the Linux kernel (the maximum
exponent for float is, just, big enough to ensure overflow for
unsigned __int128), but would apply if soft-fp were used for IEEE
binary16. Appropriate checks are inserted to ensure that the maximum
exponent is always treated as an overflowing exponent, and never as a
normal one.
Tested for powerpc-nofpu that the disassembly of installed shared
libraries is unchanged by this patch.
* soft-fp/op-common.h (_FP_TO_INT): Ensure maximum exponent is
treated as invalid conversion, not as normal exponent.
