x86: Optimize less_vec evex and avx512 memset-vec-unaligned-erms.S
No bug. This commit adds optimized cased for less_vec memset case that
uses the avx512vl/avx512bw mask store avoiding the excessive
branches. test-memset and test-wmemset are passing.
benchtests: Fix pthread-locks test to produce valid json
The benchtests json allows {function {variant}} categorization of
results whereas the pthread-locks tests had {function {variant
{subvariant}}}, which broke validation. Fix that by serializing the
subvariants as variant-subvariant. Also update the schema to
recognize the new benchmark attributes after fixing the naming
conventions.
No Bug. This commit expanding the range of tests / benchmarks for
memmove and memcpy. The test expansion is mostly in the vein of
increasing the maximum size, increasing the number of unique
alignments tested, and testing both source < destination and vice
versa. The benchmark expansaion is just to increase the number of
unique alignments. test-memcpy, test-memccpy, test-mempcpy,
test-memmove, and tst-memmove-overflow all pass.
Fangrui Song [Fri, 16 Apr 2021 18:26:39 +0000 (11:26 -0700)]
Set the retain attribute on _elf_set_element if CC supports [BZ #27492]
So that text_set_element/data_set_element/bss_set_element defined
variables will be retained by the linker.
Note: 'used' and 'retain' are orthogonal: 'used' makes sure the variable
will not be optimized out; 'retain' prevents section garbage collection
if the linker support SHF_GNU_RETAIN.
GNU ld 2.37 and LLD 13 will support -z start-stop-gc which allow C
identifier name sections to be GCed even if there are live
__start_/__stop_ references.
Without the change, there are some static linking problems, e.g.
_IO_cleanup (libio/genops.c) may be discarded by ld --gc-sections, so
stdout is not flushed on exit.
Note: GCC may warning 'retain' attribute ignored while __has_attribute(retain)
is 1 (https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99587).
x86: Update large memcpy case in memmove-vec-unaligned-erms.S
No Bug. This commit updates the large memcpy case (no overlap). The
update is to perform memcpy on either 2 or 4 contiguous pages at
once. This 1) helps to alleviate the affects of false memory aliasing
when destination and source have a close 4k alignment and 2) In most
cases and for most DRAM units is a modestly more efficient access
pattern. These changes are a clear performance improvement for
VEC_SIZE =16/32, though more ambiguous for VEC_SIZE=64. test-memcpy,
test-memccpy, test-mempcpy, test-memmove, and tst-memmove-overflow all
pass.
Matheus Castanho [Wed, 17 Mar 2021 13:14:15 +0000 (10:14 -0300)]
powerpc: Add missing registers to clobbers list for syscalls [BZ #27623]
Some registers that can be clobbered by the kernel during a syscall are not
listed on the clobbers list in sysdeps/unix/sysv/linux/powerpc/sysdep.h.
For syscalls using sc:
- XER is zeroed by the kernel on exit
For syscalls using scv:
- XER is zeroed by the kernel on exit
- Different from the sc case, most CR fields can be clobbered (according to
the ELF ABI and the Linux kernel's syscall ABI for powerpc
(linux/Documentation/powerpc/syscall64-abi.rst)
The same should apply to vsyscalls, which effectively execute a function call
but are not currently adding these registers as clobbers either.
These are likely not causing issues today, but they should be added to the
clobbers list just in case things change on the kernel side in the future.
Reported-by: Nicholas Piggin <npiggin@gmail.com> Reviewed-by: Nicholas Piggin <npiggin@gmail.com> Reviewed-by: Raphael M Zinsly <rzinsly@linux.ibm.com>
misc: syslog: Assume MSG_NOSIGNAL support (BZ #17144)
MSG_NOSIGNAL was added on POSIX 2008 and Hurd seems to support it.
The SIGPIPE handling also makes the implementation not thread-safe
(due the sigaction usage).
Now that libsupport abstract Linux possible missing support (either
due FS limitation that can't handle 64 bit timestamp or architectures
that do not handle values larger than unsigned 32 bit values) the
tests can be turned generic.
Checked on x86_64-linux-gnu and i686-linux-gnu. I also built the
tests for i686-gnu.
Another issue is that ld auditing ignores DT_BIND_NOW and thus tries to
relocate tlsdesc lazily, but that does not work in a BIND_NOW module
due to missing DT_TLSDESC_PLT. Unconditionally relocating tlsdesc at
load time fixes this bug 27721 too.
Szabolcs Nagy [Wed, 30 Dec 2020 21:52:38 +0000 (21:52 +0000)]
elf: Refactor _dl_update_slotinfo to avoid use after free
map is not valid to access here because it can be freed by a concurrent
dlclose: during tls access (via __tls_get_addr) _dl_update_slotinfo is
called without holding dlopen locks. So don't check the modid of map.
The map == 0 and map != 0 code paths can be shared (avoiding the dtv
resize in case of map == 0 is just an optimization: larger dtv than
necessary would be fine too).
the generation counter update is not needed in the failure path.
That commit ensures allocation in _dl_add_to_slotinfo happens before
the demarcation point in dlopen (it is called twice, first time is for
allocation only where dlopen can still be reverted on failure, then
second time actual dtv updates are done which then cannot fail).
Szabolcs Nagy [Thu, 31 Dec 2020 12:24:38 +0000 (12:24 +0000)]
elf: Add a DTV setup test [BZ #27136]
The test dlopens a large number of modules with TLS, they are reused
from an existing test.
The test relies on the reuse of slotinfo entries after dlclose, without
bug 27135 fixed this needs a failing dlopen. With a slotinfo list that
has non-monotone increasing generation counters, bug 27136 can trigger.
Szabolcs Nagy [Wed, 30 Dec 2020 23:40:14 +0000 (23:40 +0000)]
elf: Fix a DTV setup issue [BZ #27136]
The max modid is a valid index in the dtv, it should not be skipped.
The bug is observable if the last module has modid == 64 and its
generation is same or less than the max generation of the previous
modules. Then dtv[0].counter implies dtv[64] is initialized but
it isn't. Fixes bug 27136.
Szabolcs Nagy [Tue, 2 Feb 2021 15:02:09 +0000 (15:02 +0000)]
Remove PR_TAGGED_ADDR_ENABLE from sys/prctl.h
The value of PR_TAGGED_ADDR_ENABLE was incorrect in the installed
headers and the prctl command macros were missing that are needed
for it to be useful (PR_SET_TAGGED_ADDR_CTRL). Linux headers have
the definitions since 5.4 so it's widely available, we don't need
to repeat these definitions. The remaining definitions are from
Linux 5.10.
To build glibc with --enable-memory-tagging, Linux 5.4 headers and
binutils 2.33.1 or newer is needed.
linux: sysconf: limit _SC_MAX_ARG to 6 MiB (BZ #25305)
Since Linux 4.13, kernel limits the maximum command line arguments
length to 6 MiB [1]. Normally the limit is still quarter of the maximum
stack size but if that limit exceeds 6 MiB it's clamped down.
glibc's __sysconf implementation for Linux platform is not aware of
this limitation and for stack sizes of over 24 MiB it returns higher
ARG_MAX than Linux will actually accept. This can be verified by
executing the following application on Linux 4.13 or newer:
On affected systems the program will report ARG_MAX as 10 MiB but
despite that executing /bin/true with a bit over 6 MiB of command line
arguments will fail with E2BIG error. Expected result is that ARG_MAX
is reported as 6 MiB.
Update the __sysconf function to clamp ARG_MAX value to 6 MiB if it
would otherwise exceed it. This resolves bug #25305 which was market
WONTFIX as suggested solution was to cap ARG_MAX at 128 KiB.
As an aside and point of comparison, bionic (a libc implementation for
Android systems) decided to resolve this issue by always returning 128
KiB ignoring any potential xargs regressions [2].
On older kernels this results in returning overly conservative value
but that's a safer option than being aggressive and returning invalid
value on recent systems. It's also worth noting that at this point
all supported Linux releases have the 6 MiB barrier so only someone
running an unsupported kernel version would get incorrectly truncated
result.
Dan Raymond [Tue, 13 Apr 2021 13:26:12 +0000 (10:26 -0300)]
misc: syslog: Fix calls to openlog() with LOG_KERN facility (BZ #3604)
POSIX states for syslog [1]:
"Values of the priority argument are formed by OR'ing together a
severity-level value and an optional facility value. If no
facility value is specified, the current default facility value is
used."
So the patch fixes an existing violation of the openlog interface contract
where it is ignoring the facility argument when the value is zero
It allows the use LOG_KERN by calling openlog prior syslog usage.
Paul Eggert [Mon, 12 Apr 2021 02:06:00 +0000 (19:06 -0700)]
Improve documentation for malloc etc. (BZ#27719)
Cover key corner cases (e.g., whether errno is set) that are well
settled in glibc, fix some examples to avoid integer overflow, and
update some other dated examples (code needed for K&R C, e.g.).
* manual/charset.texi (Non-reentrant String Conversion):
* manual/filesys.texi (Symbolic Links):
* manual/memory.texi (Allocating Cleared Space):
* manual/socket.texi (Host Names):
* manual/string.texi (Concatenating Strings):
* manual/users.texi (Setting Groups):
Use reallocarray instead of realloc, to avoid integer overflow issues.
* manual/filesys.texi (Scanning Directory Content):
* manual/memory.texi (The GNU Allocator, Hooks for Malloc):
* manual/tunables.texi:
Use code font for 'malloc' instead of roman font.
(Symbolic Links): Don't assume readlink return value fits in 'int'.
* manual/memory.texi (Memory Allocation and C, Basic Allocation)
(Malloc Examples, Alloca Example):
* manual/stdio.texi (Formatted Output Functions):
* manual/string.texi (Concatenating Strings, Collation Functions):
Omit pointer casts that are needed only in ancient K&R C.
* manual/memory.texi (Basic Allocation):
Say that malloc sets errno on failure.
Say "convert" rather than "cast", since casts are no longer needed.
* manual/memory.texi (Basic Allocation):
* manual/string.texi (Concatenating Strings):
In examples, use C99 declarations after statements for brevity.
* manual/memory.texi (Malloc Examples): Add portability notes for
malloc (0), errno setting, and PTRDIFF_MAX.
(Changing Block Size): Say that realloc (p, 0) acts like
(p ? (free (p), NULL) : malloc (0)).
Add xreallocarray example, since other examples can use it.
Add portability notes for realloc (0, 0), realloc (p, 0),
PTRDIFF_MAX, and improve notes for reallocating to the same size.
(Allocating Cleared Space): Reword now-confusing discussion
about replacement, and xref "Replacing malloc".
* manual/stdio.texi (Formatted Output Functions):
Don't assume message size fits in 'int'.
* manual/string.texi (Concatenating Strings):
Fix undefined behavior involving arithmetic on a freed pointer.
linux: Normalize and return timeout on select (BZ #27651)
The commit 2433d39b697, which added time64 support to select, changed
the function to use __NR_pselect6 (or __NR_pelect6_time64) on all
architectures. However, on architectures where the symbol was
implemented with __NR_select the kernel normalizes the passed timeout
instead of return EINVAL. For instance, the input timeval
{ 0, 5000000 } is interpreted as { 5, 0 }.
And as indicated by BZ #27651, this semantic seems to be expected
and changing it results in some performance issues (most likely
the program does not check the return code and keeps issuing
select with unormalized tv_usec argument).
To avoid a different semantic depending whether which syscall the
architecture used to issue, select now always normalize the timeout
input. This is a slight change for some ABIs (for instance aarch64).
Fix SXID_ERASE behavior in setuid programs (BZ #27471)
When parse_tunables tries to erase a tunable marked as SXID_ERASE for
setuid programs, it ends up setting the envvar string iterator
incorrectly, because of which it may parse the next tunable
incorrectly. Given that currently the implementation allows malformed
and unrecognized tunables pass through, it may even allow SXID_ERASE
tunables to go through.
This change revamps the SXID_ERASE implementation so that:
- Only valid tunables are written back to the tunestr string, because
of which children of SXID programs will only inherit a clean list of
identified tunables that are not SXID_ERASE.
- Unrecognized tunables get scrubbed off from the environment and
subsequently from the child environment.
- This has the side-effect that a tunable that is not identified by
the setxid binary, will not be passed on to a non-setxid child even
if the child could have identified that tunable. This may break
applications that expect this behaviour but expecting such tunables
to cross the SXID boundary is wrong. Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Instead of passing GLIBC_TUNABLES via the environment, pass the
environment variable from parent to child. This allows us to test
multiple variables to ensure better coverage.
The test list currently only includes the case that's already being
tested. More tests will be added later. Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Add a new function support_capture_subprogram_self_sgid that spawns an
sgid child of the running program with its own image and returns the
exit code of the child process. This functionality is used by at
least three tests in the testsuite at the moment, so it makes sense to
consolidate.
There is also a new function support_subprogram_wait which should
provide simple system() like functionality that does not set up file
actions. This is useful in cases where only the return code of the
spawned subprocess is interesting.
This patch also ports tst-secure-getenv to this new function. A
subsequent patch will port other tests. This also brings an important
change to tst-secure-getenv behaviour. Now instead of succeeding, the
test fails as UNSUPPORTED if it is unable to spawn a setgid child,
which is how it should have been in the first place. Reviewed-by: Carlos O'Donell <carlos@redhat.com>
i.e. sp at setjmp was outside of the altstack range. Here we know that
longjmp is called from a signal handler on the altstack (SS_ONSTACK),
and that it jumps in the wrong direction (sp decreases), so the check
wants to ensure the jump goes to another stack.
The check is wrong when altstack_sp == setjmp_sp which can happen
when the altstack is a local buffer in the function that calls setjmp,
so the patch allows == too. This fixes bug 27709.
Note that the generic __longjmp_chk check seems to be different.
(it checks if longjmp was on the altstack but does not check setjmp,
so it would not catch incorrect longjmp use within the signal handler).
Paul Zimmermann [Fri, 2 Apr 2021 06:21:06 +0000 (08:21 +0200)]
Improve the accuracy of tgamma (BZ #26983)
With this patch, the maximal known error for tgamma is now reduced to 9 ulps
for dbl-64, for all rounding modes. Since exhaustive testing is not possible
for dbl-64, it might be that there are still cases with an error larger than
9 ulps, but all known cases are fixed (intensive tests were done to find cases
with large errors).
Tested on x86_64 and powerpc (and by Adhemerval Zanella on aarch64, arm,
s390x, sparc, and i686). Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
The simplification of tunable_set interfaces took care of
signed/unsigned conversions while setting values, but comparison with
bounds ended up being incorrect; comparing TUNABLE_SIZE_T values for
example will fail because SIZE_MAX is seen as -1.
Add comparison helpers that take tunable types into account and use
them to do comparison instead.
Update sv_SE to treate 'W' as a distinct character (Bug 25036)
The 13th edition of Svenska Akademiens ordlista lists 'W' as a
distinct letter that sorts after 'V'. We adjust the sv_SE locale
(and tests) to match this updated and "reformed" language change.
This harmonizes us with CLDR 1.5.0 (2007) for sv_SE sorting of
the letter 'W'.
No regressions on x86_64, and locale sorting tests all pass.
Co-authored-by: Carlos O'Donell <carlos@redhat.com>
Maninder Singh [Wed, 10 Jan 2018 15:17:30 +0000 (15:17 +0000)]
elf: Fix data race in _dl_name_match_p [BZ #21349]
dlopen updates libname_list by writing to lastp->next, but concurrent
reads in _dl_name_match_p were not synchronized when it was called
without holding GL(dl_load_lock), which can happen during lazy symbol
resolution.
This patch fixes the race between _dl_name_match_p reading lastp->next
and add_name_to_object writing to it. This could cause segfault on
targets with weak memory order when lastp->next->name is read, which
was observed on an arm system. Fixes bug 21349.
(Code is from Maninder Singh, comments and description is from Szabolcs
Nagy.)
Szabolcs Nagy [Thu, 11 Feb 2021 13:38:10 +0000 (13:38 +0000)]
aarch64: free tlsdesc data on dlclose [BZ #27403]
DL_UNMAP_IS_SPECIAL and DL_UNMAP were not defined. The definitions are
now copied from arm, since the same is needed on aarch64. The cleanup
of tlsdesc data is handled by the custom _dl_unmap.
Required after 9acda61d94acc "Fix the inaccuracy of j0f/j1f/y0f/y1f
[BZ #14469, #14470, #14471, #14472]" and db3f7bb558 "math: Remove
slow paths from asin and acos [BZ #15267]".
Paul Zimmermann [Thu, 1 Apr 2021 06:14:10 +0000 (08:14 +0200)]
Fix the inaccuracy of j0f/j1f/y0f/y1f [BZ #14469, #14470, #14471, #14472]
For j0f/j1f/y0f/y1f, the largest error for all binary32
inputs is reduced to at most 9 ulps for all rounding modes.
The new code is enabled only when there is a cancellation at the very end of
the j0f/j1f/y0f/y1f computation, or for very large inputs, thus should not
give any visible slowdown on average. Two different algorithms are used:
* around the first 64 zeros of j0/j1/y0/y1, approximation polynomials of
degree 3 are used, computed using the Sollya tool (https://www.sollya.org/)
* for large inputs, an asymptotic formula from [1] is used
[1] Fast and Accurate Bessel Function Computation,
John Harrison, Proceedings of Arith 19, 2009.
Inputs yielding the new largest errors are added to auto-libm-test-in,
and ulps are regenerated for various targets (thanks Adhemerval Zanella).
Tested on x86_64 with --disable-multi-arch and on powerpc64le-linux-gnu. Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
H.J. Lu [Tue, 16 Mar 2021 14:41:46 +0000 (07:41 -0700)]
x86_64: Correct THREAD_SETMEM/THREAD_SETMEM_NC for movq [BZ #27591]
config/i386/constraints.md in GCC has
(define_constraint "e"
"32-bit signed integer constant, or a symbolic reference known
to fit that range (for immediate operands in sign-extending x86-64
instructions)."
(match_operand 0 "x86_64_immediate_operand"))
Since movq takes a signed 32-bit immediate or a register source operand,
use "er", instead of "nr"/"ir", constraint for 32-bit signed integer
constant or register on movq.
Andreas Schwab [Wed, 31 Mar 2021 12:17:24 +0000 (14:17 +0200)]
powerpc64le: Use ifunc for _Float128 functions also in libc
This fixes missing definition of math functions in libc in a static link
that are no longer built for libm after commit 4898d9712b ("Avoid adding
duplicated symbols into static libraries").
Stefan Liebler [Wed, 31 Mar 2021 14:17:01 +0000 (16:17 +0200)]
S390: Allow "v" constraint for long double math_opt_barrier and math_force_eval with GCC 11.
Starting with GCC 11, long double values can also be processed in vector
registers if build with -march >= z14. Then GCC defines the
__LONG_DOUBLE_VX__ macro.
io: Check at runtime if timestamp supports nanoseconds
Now that non-LFS stat function is implemented on to on LFS, it will
use statx when available. It allows to check for nanosecond timestamp
if the kernel supports __NR_statx.
Stefan Liebler [Wed, 31 Mar 2021 08:23:16 +0000 (10:23 +0200)]
Fix conform linknamespace tests due to gnu_dev_makedev
If building on s390 / i686 with -Os, various conformance
tests are failing with e.g.
conform/ISO/assert.h/linknamespace.out:
[initial] __assert_fail -> [libc.a(assert.o)] __dcgettext -> [libc.a(dcgettext.o)] __dcigettext -> [libc.a(dcigettext.o)] __getcwd -> [libc.a(getcwd.o)] __fstatat64 -> [libc.a(fstatat64.o)] gnu_dev_makedev
The usage of gnu_dev_makedev was recently introduced by
usage of the makedev makro in commit: 5b980d4809913088729982865188b754939bcd39
linux: Use statx for MIPSn64
This patch is now linking against __gnu_dev_makedev as
also done in commit: 8b4a118222c7ed41bc653943b542915946dff1dd
Fix -Os gnu_dev_* linknamespace, localplt issues (bug 15105, bug 19463).
About a decade ago, I accidentally wrote the GPLv3 license text on the
test case when the rest of glibc source is LGPL v2.1 or later. As
original author of the test (and there are no other legally
significant changes to the test) I propose to update the license text
to be consistent with the project.
Avoid adding duplicated symbols into static libraries
Some math functions (such as __isnan*) are built into both libm and
libc because they are needed in libc. The symbol gets exported from
libc.so and not libm.so, because of which dynamic linking works fine;
the symbols are always resolved from libc.so and libm.so uses its
internal copy of the same function if needed.
When linking statically though, the libm variants get used throughout
because the symbols are exported in both archives and libm.a is
searched first.
This patch removes these duplicate objects from the libm.a archive so
that programs always link to libc in both, the static and dynamic
case. The difference this will cause is that libm uses of these
functions will start using the libc versions in the !SHARED case.
This is harmless at the moment because the objects are identical
except for their names.
Some of these duplicates could be removed from libm.so too, but I
avoided that in the interest of retaining an internal reference if at
all those functions get used within libm in future.
Reviewed-by: Paul Zimmermann <Paul.Zimmermann@inria.fr>
Samuel Thibault [Wed, 24 Mar 2021 20:27:34 +0000 (21:27 +0100)]
fork.h: replace with register-atfork.h
UNREGISTER_ATFORK is now defined for all ports in register-atfork.h, so most
previous includes of fork.h actually only need register-atfork.h now, and
cxa_finalize.c does not need an ifdef UNREGISTER_ATFORK any more.
The nptl-specific fork generation counters can then go to pthreadP.h, and
fork.h be removed.
H.J. Lu [Sun, 7 Mar 2021 17:45:23 +0000 (09:45 -0800)]
x86-64: Use ZMM16-ZMM31 in AVX512 memmove family functions
Update ifunc-memmove.h to select the function optimized with AVX512
instructions using ZMM16-ZMM31 registers to avoid RTM abort with usable
AVX512VL since VZEROUPPER isn't needed at function exit.
H.J. Lu [Sun, 7 Mar 2021 17:44:18 +0000 (09:44 -0800)]
x86-64: Use ZMM16-ZMM31 in AVX512 memset family functions
Update ifunc-memset.h/ifunc-wmemset.h to select the function optimized
with AVX512 instructions using ZMM16-ZMM31 registers to avoid RTM abort
with usable AVX512VL and AVX512BW since VZEROUPPER isn't needed at
function exit.
H.J. Lu [Tue, 23 Feb 2021 14:33:10 +0000 (06:33 -0800)]
x86: Add string/memory function tests in RTM region
At function exit, AVX optimized string/memory functions have VZEROUPPER
which triggers RTM abort. When such functions are called inside a
transactionally executing RTM region, RTM abort causes severe performance
degradation. Add tests to verify that string/memory functions won't
cause RTM abort in RTM region.
H.J. Lu [Fri, 5 Mar 2021 15:26:42 +0000 (07:26 -0800)]
x86-64: Add AVX optimized string/memory functions for RTM
Since VZEROUPPER triggers RTM abort while VZEROALL won't, select AVX
optimized string/memory functions with
xtest
jz 1f
vzeroall
ret
1:
vzeroupper
ret
at function exit on processors with usable RTM, but without 256-bit EVEX
instructions to avoid VZEROUPPER inside a transactionally executing RTM
region.
H.J. Lu [Fri, 5 Mar 2021 15:20:28 +0000 (07:20 -0800)]
x86-64: Add memcmp family functions with 256-bit EVEX
Update ifunc-memcmp.h to select the function optimized with 256-bit EVEX
instructions using YMM16-YMM31 registers to avoid RTM abort with usable
AVX512VL, AVX512BW and MOVBE since VZEROUPPER isn't needed at function
exit.
H.J. Lu [Fri, 5 Mar 2021 15:15:03 +0000 (07:15 -0800)]
x86-64: Add memset family functions with 256-bit EVEX
Update ifunc-memset.h/ifunc-wmemset.h to select the function optimized
with 256-bit EVEX instructions using YMM16-YMM31 registers to avoid RTM
abort with usable AVX512VL and AVX512BW since VZEROUPPER isn't needed at
function exit.
H.J. Lu [Fri, 5 Mar 2021 14:46:08 +0000 (06:46 -0800)]
x86-64: Add memmove family functions with 256-bit EVEX
Update ifunc-memmove.h to select the function optimized with 256-bit EVEX
instructions using YMM16-YMM31 registers to avoid RTM abort with usable
AVX512VL since VZEROUPPER isn't needed at function exit.
H.J. Lu [Fri, 5 Mar 2021 14:36:50 +0000 (06:36 -0800)]
x86-64: Add strcpy family functions with 256-bit EVEX
Update ifunc-strcpy.h to select the function optimized with 256-bit EVEX
instructions using YMM16-YMM31 registers to avoid RTM abort with usable
AVX512VL and AVX512BW since VZEROUPPER isn't needed at function exit.
H.J. Lu [Fri, 5 Mar 2021 14:24:52 +0000 (06:24 -0800)]
x86-64: Add ifunc-avx2.h functions with 256-bit EVEX
Update ifunc-avx2.h, strchr.c, strcmp.c, strncmp.c and wcsnlen.c to
select the function optimized with 256-bit EVEX instructions using
YMM16-YMM31 registers to avoid RTM abort with usable AVX512VL, AVX512BW
and BMI2 since VZEROUPPER isn't needed at function exit.
For strcmp/strncmp, prefer AVX2 strcmp/strncmp if Prefer_AVX2_STRCMP
is set.
H.J. Lu [Fri, 26 Feb 2021 13:36:59 +0000 (05:36 -0800)]
x86: Set Prefer_No_VZEROUPPER and add Prefer_AVX2_STRCMP
1. Set Prefer_No_VZEROUPPER if RTM is usable to avoid RTM abort triggered
by VZEROUPPER inside a transactionally executing RTM region.
2. Since to compare 2 32-byte strings, 256-bit EVEX strcmp requires 2
loads, 3 VPCMPs and 2 KORDs while AVX2 strcmp requires 1 load, 2 VPCMPEQs,
1 VPMINU and 1 VPMOVMSKB, AVX2 strcmp is faster than EVEX strcmp. Add
Prefer_AVX2_STRCMP to prefer AVX2 strcmp family functions.
Paul Zimmermann [Fri, 19 Mar 2021 09:09:20 +0000 (10:09 +0100)]
add workload traces for missing functions (double format)
This patch adds workload traces for all double format functions where such
files are missing. For each function, a set of 1000 random values is
generated at random using SageMath, such that the output values are
meaningful (for example avoiding too large inputs for exp10 where the
output would be +Inf). More details about the generated values are
given at the beginning of each file. Reviewed-by: Siddhesh Poyarekar <siddhesh@sourceware.org>
The tests are refactored to use a common skeleton that handles whether
the underlying filesystem supports 64 bit time, skips 64 bit time
tests when the TU only supports 32 bit, and also skip 64 bit time
tests larger than 32 unsigned int (y2106) if the system does not
support it (MIPSn64 on kernels without statx support).
Checked on x86_64-linux-gnu and i686-linux-gnu. I also checked
on a mips64el-linux-gnu with 4.1.4 and 5.10.0-4-5kc-malta kernel
to verify if the y2106 are indeed skipped.