linux: Add fallback for 64-bit time_t SO_{RCV,SND}TIMEO
The constant values will be changed for __TIMESIZE=64, so binaries built
with 64-bit time support might fail to work properly on old kernels.
Both {get,set}sockopt will retry the syscall with the old constant
values and the timeout value adjusted when kernel returns ENOTPROTOPT.
It also adds an internal only SO_{RCV,SND}TIMEO where
COMPAT_SO_{RCV,SND}TIMEO_OLD indicates pre 32-bit time support and
COMPAT_SO_{RCV,SND}TIMEO_NEW indicates time64 support. It allows to
refer to constant independently of the time_t ABI and kernel version
used.
Checked on x86_64-linux-gnu and i686-linux-gnu (on 5.4 and on 4.15
kernel).
Reviewed-by: Carlos O'Donell <carlos@redhat.com> Tested-by: Carlos O'Donell <carlos@redhat.com>
Reinstate gconv-modules as the default configuration file
Reinstate gconv-modules as the main file so that the configuration
files in gconv-modules.d/ become add-on configuration. With this, the
effective user visible change is that GCONV_PATH can now have
supplementary configuration in GCONV_PATH/gconv-modules.d/ in addition
to the main GCONV_PATH/gconv-modules file.
Add build option to disable usage of scv on powerpc
Commit 68ab82f56690ada86ac1e0c46bad06ba189a10ef added support for the scv
syscall ABI on powerpc. Since then systems that have kernel and processor
support started using scv. However adding the proper support for a new syscall
ABI requires changes to several other projects (e.g. qemu, valgrind, strace,
kernel), which are gradually receiving support.
Meanwhile, having a way to disable scv on glibc at build time can be useful for
distros that may encounter conflicts with projects that still do not support the
scv ABI, buying time until proper support is added.
Florian Weimer [Thu, 10 Jun 2021 08:09:51 +0000 (10:09 +0200)]
io: Fix sporadic test failures in io/tst-stat
support_stat_nanoseconds cannot restore the ctime time, and
this may lead to sporadic test failures. Therefore, probe for
nanoseconds support before the initial statx call.
nptl: Avoid async cancellation to wrongly update __nptl_nthreads (BZ #19366)
The testcase provided on BZ#19366 may update __nptl_nthreads in a wrong
order, triggering an early process exit because the thread decrement
the value twice.
The issue is once the thread exits without acting on cancellation,
it decreaments '__nptl_nthreads' and then atomically set
'cancelhandling' with EXITING_BIT (thus preventing further cancellation
handler to act). The issue happens if a SIGCANCEL is received between
checking '__ntpl_nthreads' and setting EXITING_BIT. To avoid it, the
'__nptl_nthreads' decrement is moved after EXITING_BIT.
It does fully follow the POSIX XSH 2.9.5 Thread Cancellation under
the heading Thread Cancellation Cleanup Handlers that states that
when a cancellation request is acted upon, or when a thread calls
pthread_exit(), the thread first disables cancellation by setting its
cancelability state to PTHREAD_CANCEL_DISABLE and its cancelability type
to PTHREAD_CANCEL_DEFERRED. The issue is '__pthread_enable_asynccancel'
explicit enabled assynchrnous cancellation, so an interrupted syscall
within the cancellation cleanup handlers might see an invalid cancelling
type (a possible fix might be possible with my proposed solution to
BZ#12683).
Trying to come up with a test is quite hard since it requires to
mimic the timing issue described below, however I see that the
bug report reproducer does not early exit anymore.
It consolidates the tgkill call and it is the first step of making
pthread_cancel async-signal-safe. It also fix a possible issue
where the 'struct pthread' tid is not read atomically, which might
send an invalid cancellation signal (similar to what db988e50a87f613cb6b9e98a2fc66a4848bc3546 fixed for pthread_join).
Checked on x86_64-linux-gnu and aarch64-linux-gnu.
Now that pthread_kill is provided by libc.so it is possible to
implement the generic POSIX implementation as
'pthread_kill(pthread_self(), sig)'.
For Linux implementation, pthread_kill read the targeting TID from
the TCB. For raise, this it not possible because it would make raise
fail when issue after vfork (where creates the resulting process
has a different TID from the parent, but its TCB is not updated as
for pthread_create). To make raise use pthread_kill, it is make
usable from vfork by getting the target thread id through gettid
syscall.
Checked on x86_64-linux-gnu and aarch64-linux-gnu.
Now that thread cancellation state is not accessed concurrently anymore,
it is possible to move it out the 'cancelhandling'.
The code is also simplified: CANCELLATION_P is replaced with a
internal pthread_testcancel call and the CANCELSTATE_BIT{MASK} is
removed.
With this behavior pthread_setcancelstate does not require to act on
cancellation if cancel type is asynchronous (is already handled either
by pthread_setcanceltype or by the signal handler).
Checked on x86_64-linux-gnu and aarch64-linux-gnu.
The CANCELING_BITMASK is used as an optimization to avoid sending
the signal when pthread_cancel is called in a concurrent manner.
This requires then to put both the cancellation state and type on a
shared state (cancelhandling), since 'pthread_cancel' checks whether
cancellation is enabled and asynchrnous to either cancel itself of
sending the signal.
It also requires handle the CANCELING_BITMASK on
__pthread_disable_asynccancel, however this incurs in the same issues
described on BZ#12683: the cancellation is acted upon even *after*
syscall returns with user visible side-effects.
This patch removes this optimization and simplifies the pthread
cancellation implementation: pthread_cancel now first checks if
cancellation is already pending and if not always, sends a signal
if the target is not itself. The SIGCANCEL handler is also simpified
since there is not need to setup a CAS loop.
It also allows to move both the cancellation state and mode out of
'cancelhadling' (it is done in subsequent patches).
Checked on x86_64-linux-gnu and aarch64-linux-gnu.
nptl: Install cancellation handler on pthread_cancel
Now that cancellation is not used anymore to handle thread setup
creation failure, the sighandle can be installed only when
pthread_cancel is actually used.
Checked on x86_64-linux-gnu and aarch64-linux-gnu.
nptl: Deallocate the thread stack on setup failure (BZ #19511)
To setup either the thread scheduling parameters or affinity,
pthread_create enforce synchronization on created thread to wait until
its parent either release PD ownership or send a cancellation signal if
a failure occurs.
However, cancelling the thread does not deallocate the newly created
stack since cancellation expects that a pthread_join to deallocate any
allocated thread resouces (threads stack or TLS).
This patch changes on how the thread resource is deallocate in case of
failure to be synchronous, where the creating thread will signal the
created thread to exit early so it could be joined. The creating thread
will be reponsible for the resource cleanup before returning to the
caller.
To signal the creating thread that a failure has occured, an unused
'struct pthread' member, parent_cancelhandling_unsed, now indicates
whether the setup has failed so creating thread can proper exit.
This strategy also simplifies by not using thread cancellation and
thus not running libgcc_so load in the signal handler (which is
avoided in thread cancellation since 'pthread_cancel' is the one
responsible to dlopen libgcc_s). Another advantage is since the
early exit is move to first step at thread creation, the signal
mask is not already set and thus it can not act on change ID setxid
handler.
Checked on x86_64-linux-gnu and aarch64-linux-gnu.
Stefan Liebler [Wed, 9 Jun 2021 11:19:51 +0000 (13:19 +0200)]
Fix elf/tst-tls9-static after libdl cleanups.
The testcase elf/tst-tls9-static sometimes fails with:
cannot open 'tst-tlsmod5.so': tst-tlsmod5.so: cannot open shared object file: No such file or directory
cannot open 'tst-tlsmod6.so': tst-tlsmod6.so: cannot open shared object file: No such file or directory
After recent commit 6f1c701026287f6928d3bdd1aea7359308635abe
"dlfcn: Cleanups after -ldl is no longer required"
the libdl variable is not set anymore and thus the
dependencies were missing.
iconvdata: Split out non-essential gconv module configuration
Split module configuration so that only the bare minimum charsets,
i.e. ANSI_X3.110, ISO8859-15, ISO8859-1, CP1252, UNICODE, UTF-16,
UTF-32 and UTF-7 are configured in gconv-modules.conf. The remaining
module configurations are now in gconv-modules-extra.conf.
iconvdata: Move gconv-modules configuration to gconv-modules.conf
Move all gconv-modules configuration files to gconv-modules.conf.
That is, the S390 extensions now become gconv-modules-s390.conf. Move
both configuration files into gconv-modules.d.
Now GCONV_PATH/gconv-modules is read only for backward compatibility
for third-party gconv modules directories.
gconv_conf: Read configuration files in gconv-modules.d
Read configuration files with names ending in .conf in
GCONV_PATH/gconv-modules.d to mirror configuration flexibility in
iconvconfig into the iconv program and function.
iconvconfig: Read configuration from gconv-modules.d subdirectory
In addition to GCONV_PATH/gconv-modules, also read module
configuration from *.conf files in GCONV_PATH/gconv-modules.d. This
allows a single gconv directory to have multiple sets of gconv modules
but at the same time, a single modules cache.
With this feature, one could separate the glibc supported gconv
modules into a minimal essential set (ISO-8859-*, UTF, etc.) from the
remaining modules. In future, these could be further segregated into
langpack-associated sets with their own
gconv-modules.d/someconfig.conf.
Wilco Dijkstra [Tue, 8 Jun 2021 12:33:09 +0000 (13:33 +0100)]
AArch64: Add support for roundeven[f]
Add inline assembler for the roundeven functions.
Passes GLIBC regression. Note GCC does not inline the builtin (PR100966),
so this cannot be used for now.
Naohiro Tamura [Sat, 22 May 2021 02:42:48 +0000 (02:42 +0000)]
configure: Replaced obsolete AC_TRY_COMPILE
This patch replaced obsolete AC_TRY_COMPILE to AC_COMPILE_IFELSE or
AC_PREPROC_IFELSE.
It has been confirmed that GNU 'autoconf' 2.69 suppressed obsolete
warnings, updated the following files:
- configure
- sysdeps/mach/configure
- sysdeps/mach/hurd/configure
- sysdeps/s390/configure
- sysdeps/unix/sysv/linux/configure
and didn't change the following files:
- sysdeps/ieee754/ldbl-opt/configure
- sysdeps/unix/sysv/linux/powerpc/configure
Florian Weimer [Thu, 3 Jun 2021 06:26:04 +0000 (08:26 +0200)]
dlfcn: Rework static dlopen hooks
Consolidate all hooks structures into a single one. There are
no static dlopen ABI concerns because glibc 2.34 already comes
with substantial ABI-incompatible changes in this area. (Static
dlopen requires the exact same dynamic glibc version that was used
for static linking.)
The new approach uses a pointer to the hooks structure into
_rtld_global_ro and initalizes it in __rtld_static_init. This avoids
a back-and-forth with various callback functions.
Florian Weimer [Thu, 3 Jun 2021 06:26:04 +0000 (08:26 +0200)]
dlfcn: Eliminate GLIBC_PRIVATE dependency from tststatic2
The test appears to use _dlfcn_hook@@GLIBC_PRIVATE as a way to
test dlvsym without having to know the appropriate symbol version.
With <first-versions.h>, we can use a public symbol and the symbol
version at which it was defined first.
Florian Weimer [Thu, 3 Jun 2021 06:26:04 +0000 (08:26 +0200)]
dlfcn: Cleanups after -ldl is no longer required
This commit removes the ELF constructor and internal variables from
dlfcn/dlfcn.c. The file now serves the same purpose as
nptl/libpthread-compat.c, so it is renamed to dlfcn/libdl-compat.c.
The use of libdl-shared-only-routines ensures that libdl.a is empty.
This commit adjusts the test suite not to use $(libdl). The libdl.so
symbolic link is no longer installed.
Sunil K Pandey [Mon, 31 May 2021 18:08:12 +0000 (11:08 -0700)]
Improve test coverage of strlen function
This patch covers the following conditions:
- Strings start with different alignments and end at the page boundary
with less than 64 byte length.
- Strings starts with different alignments and cross page boundary with
fixed length.
Florian Weimer [Wed, 2 Jun 2021 05:32:20 +0000 (07:32 +0200)]
dlfcn: Move dlerror into libc
The symbol was moved using scripts/move-symbol-to-libc.py.
There is a minor functionality enhancement: dlerror now sets
errno if it was set as part of the exception. (This is the result
of using %m in asprintf, to avoid the strerror PLT call.) The
previous errno value upon function return was unpredictable.
Documenting this as a feature is premature; we need to make sure
that the error codes are meaningful when they are set by the dynamic
loader.
Florian Weimer [Wed, 2 Jun 2021 05:32:19 +0000 (07:32 +0200)]
Add libc ABI extension kludge for baseline-violating libdl symbols
Some targets have a GLIBC_2.0 baseline for libdl, while using
GLIBC_2.2 for libc. This means that the generated libc.map file
does not have any version nodes for GLIBC_2.0 or GLIBC_2.1. However,
moving symbols from libdl into libc needs such version nodes.
(Future symbol moves from librt will need this as well.)
This kludge is only necessary for symbols predating GLIBC_2.2 because
the affected targets use GLIBC_2.2 as the baseline for libc. Given
the small number and fixed set of affected architectures, no generic
mechanism is implemented, and instead the map file fragment is
hard-coded in scripts/versions.mk.
The compat_symbol macro already emits the appropriate version strings,
so no adjustments are needed there.
Florian Weimer [Wed, 2 Jun 2021 05:32:19 +0000 (07:32 +0200)]
scripts/versions.awk: Add local: * to all version nodes
This requires that all exported symbol versions are listed in
Versions files. It results in more consistent behavior across
architectures because previously, symbols could be exported
via explicit versioned_symbol and compat_symbol macros if the
version node existed in some Versions file (without listing the
symbol), and it was not the base version for the library (which
already had the local: * directive).
Florian Weimer [Wed, 2 Jun 2021 05:32:19 +0000 (07:32 +0200)]
Add missing symbols to Version files
Some symbols have explicit versioned_symbol or compat_symbol markers
in the sources, but no corresponding entry in the Versions files.
This presently works because the local: * directive is only applied
to the base version.
Florian Weimer [Tue, 1 Jun 2021 15:51:41 +0000 (17:51 +0200)]
Fix use of __pthread_attr_copy in mq_notify (bug 27896)
__pthread_attr_copy can fail and does not initialize the attribute
structure in that case.
If __pthread_attr_copy is never called and there is no allocated
attribute, pthread_attr_destroy should not be called, otherwise
there is a null pointer dereference in rt/tst-mqueue6.
Joseph Myers [Tue, 1 Jun 2021 14:22:06 +0000 (14:22 +0000)]
Update floating-point feature test macro handling for C2X
ISO C2X has made some changes to the handling of feature test macros
related to features from the floating-point TSes, and to exactly what
such features are present in what headers, that require corresponding
changes in glibc.
* For the few features that were controlled by
__STDC_WANT_IEC_60559_BFP_EXT__ (and the corresponding DFP macro) in
C2X, there is now instead a new feature test macro
__STDC_WANT_IEC_60559_EXT__ covering both binary and decimal FP.
This controls CR_DECIMAL_DIG in <float.h> (provided by GCC; I
implemented support for the new feature test macro for GCC 11) and
the totalorder and payload functions in <math.h>. C2X no longer
says anything about __STDC_WANT_IEC_60559_BFP_EXT__ (so it's
appropriate for that macro to continue to enable exactly the
features from TS 18661-1).
* The SNAN macros for each floating-point type have moved to <float.h>
(and been renamed in the process). Thus, the copies in <math.h>
should only be defined for __STDC_WANT_IEC_60559_BFP_EXT__, not for
C2X.
* The fmaxmag and fminmag functions have been removed (replaced by new
functions for the new min/max operations in IEEE 754-2019). Thus
those should also only be declared for
__STDC_WANT_IEC_60559_BFP_EXT__.
* The _FloatN / _FloatNx handling for the last two points in glibc is
trickier, since __STDC_WANT_IEC_60559_TYPES_EXT__ is still in C2X
(the integration of TS 18661-3 as an Annex, that is, which hasn't
yet been merged into the C standard git repository but has been
accepted by WG14), so C2X with that macro should not declare some
things that are declared for older standards with that macro. The
approach taken here is to provide the declarations (when
__STDC_WANT_IEC_60559_TYPES_EXT__ is enabled) only when (defined
__USE_GNU || !__GLIBC_USE (ISOC2X)), so if C2X features are enabled
then those declarations (that are only in TS 18661-3 and not in C2X)
will only be provided if _GNU_SOURCE is defined as well. Thus
_GNU_SOURCE remains a superset of the TS features as well as of C2X.
Some other somewhat related changes in C2X are not addressed here.
There's an open proposal not to include the fmin and fmax functions
for the _FloatN / _FloatNx types, given the new min/max operations,
which could be handled like the previous point if adopted. And the
fromfp functions have been changed to return a result in floating type
rather than intmax_t / uintmax_t; my inclination there is to treat
that like that change of totalorder type (new symbol versions etc. for
the ABI change; old versions become compat symbols and are no longer
supported as an API).
Florian Weimer [Tue, 1 Jun 2021 14:00:48 +0000 (16:00 +0200)]
stdio-common: Remove _IO_vfwscanf
The symbol has never been exported, so no compatibility symbol is
needed. Removing this file prevents ld from creation an exported
symbol in case GLIBC_2_0 expands to a symbol version which
does not have a local: *; directive in the symbol version map file.
This patch was based on the __memcmp_power8 and the recent
__strlen_power10.
Improvements from __memcmp_power8:
1. Don't need alignment code.
On POWER10 lxvp and lxvl do not generate alignment interrupts, so
they are safe for use on caching-inhibited memory. Notice that the
comparison on the main loop will wait for both VSR to be ready.
Therefore aligning one of the input address does not improve
performance. In order to align both registers a vperm is necessary
which add too much overhead.
2. Uses new POWER10 instructions
This code uses lxvp to decrease contention on load by loading 32 bytes
per instruction.
The vextractbm is used to have a smaller tail code for calculating the
return value.
3. Performance improvement
This version has around 35% better performance on average. I saw no
performance regressions for any length or alignment.
Thanks Matheus for helping me out with some details.
Co-authored-by: Matheus Castanho <msc@linux.ibm.com> Reviewed-by: Raphael M Zinsly <rzinsly@linux.ibm.com>
nptl: Add pthread_attr_setaffinity_np failure test
It checks whether an invalid affinity mask does return an error,
similar to what sysdeps/pthread/tst-bad-schedattr.c does for
pthread_attr_setschedparam.
The 'create_thread' function is moved to pthread_create.c. It removes
the START_THREAD_DEFN and START_THREAD_SELF macros and make the
lock usage more clear (no need to cross-reference multiple files).
Florian Weimer [Thu, 27 May 2021 13:37:35 +0000 (15:37 +0200)]
nptl: Install SIGSETXID handler with SA_ONSTACK [BZ #27914]
The signal is sent to all threads, some of which may have switched
to very small stacks. If they have also installed an alternate
signal stack, SA_ONSTACK makes this work. The Go runtime needs this:
runtime: C.setuid/C.setgid smashes Go stack
<https://github.com/golang/go/issues/9400>
Doing this for SIGCANCEL is less obviously beneficial and needs further
testing.
Naohiro Tamura [Thu, 27 May 2021 07:44:12 +0000 (07:44 +0000)]
aarch64: Added optimized memset for A64FX
This patch optimizes the performance of memset for A64FX [1] which
implements ARMv8-A SVE and has L1 64KB cache per core and L2 8MB cache
per NUMA node.
The performance optimization makes use of Scalable Vector Register
with several techniques such as loop unrolling, memory access
alignment, cache zero fill and prefetch.
SVE assembler code for memset is implemented as Vector Length Agnostic
code so theoretically it can be run on any SOC which supports ARMv8-A
SVE standard.
We confirmed that all testcases have been passed by running 'make
check' and 'make xcheck' not only on A64FX but also on ThunderX2.
And also we confirmed that the SVE 512 bit vector register performance
is roughly 4 times better than Advanced SIMD 128 bit register and 8
times better than scalar 64 bit register by running 'make bench'.
[1] https://github.com/fujitsu/A64FX
Reviewed-by: Wilco Dijkstra <Wilco.Dijkstra@arm.com> Reviewed-by: Szabolcs Nagy <Szabolcs.Nagy@arm.com>
Naohiro Tamura [Thu, 27 May 2021 07:42:35 +0000 (07:42 +0000)]
aarch64: Added optimized memcpy and memmove for A64FX
This patch optimizes the performance of memcpy/memmove for A64FX [1]
which implements ARMv8-A SVE and has L1 64KB cache per core and L2 8MB
cache per NUMA node.
The performance optimization makes use of Scalable Vector Register
with several techniques such as loop unrolling, memory access
alignment, cache zero fill, and software pipelining.
SVE assembler code for memcpy/memmove is implemented as Vector Length
Agnostic code so theoretically it can be run on any SOC which supports
ARMv8-A SVE standard.
We confirmed that all testcases have been passed by running 'make
check' and 'make xcheck' not only on A64FX but also on ThunderX2.
And also we confirmed that the SVE 512 bit vector register performance
is roughly 4 times better than Advanced SIMD 128 bit register and 8
times better than scalar 64 bit register by running 'make bench'.
[1] https://github.com/fujitsu/A64FX
Reviewed-by: Wilco Dijkstra <Wilco.Dijkstra@arm.com> Reviewed-by: Szabolcs Nagy <Szabolcs.Nagy@arm.com>
Naohiro Tamura [Wed, 12 May 2021 09:27:20 +0000 (09:27 +0000)]
aarch64: define BTI_C and BTI_J macros as NOP unless HAVE_AARCH64_BTI
This patch defines BTI_C and BTI_J macros conditionally for
performance.
If HAVE_AARCH64_BTI is true, BTI_C and BTI_J are defined as HINT
instruction for ARMv8.5 BTI (Branch Target Identification).
If HAVE_AARCH64_BTI is false, both BTI_C and BTI_J are defined as
NOP.
Yang Xu [Thu, 18 Feb 2021 02:56:33 +0000 (10:56 +0800)]
tst-mallinfo2.c: Use correct multiple for total variable
Since test uses 160 multiple for malloc size, we should also use 160 multiple
for total variable instead of 16, then comparison is meaningful. So fix it.
Also change the ">" to ">=" so that the test is technically valid.
The testrun.sh --help message was missing the "container" option.
Besides just adding the missing option I decide to rewrite it in the
same layout as other tools.
Nicholas Piggin [Thu, 20 May 2021 14:00:36 +0000 (11:00 -0300)]
powerpc: Fix handling of scv return error codes [BZ #27892]
When using scv for templated ASM syscalls, current code interprets any
negative return value as error, but the only valid error codes are in
the range -4095..-1 according to the ABI.
This commit also fixes 'signal.gen.test' strace test, where the issue
was first identified.
2. Add TEST_STACK_ALIGN_INIT to TEST_STACK_ALIGN.
3. Add a common TEST_STACK_ALIGN_INIT to check 16-byte stack alignment
for both i386 and x86-64.
4. Update powerpc to use TEST_STACK_ALIGN_INIT.
Noah Goldstein [Sun, 23 May 2021 23:43:24 +0000 (19:43 -0400)]
x86: Improve memmove-vec-unaligned-erms.S
This patch changes the condition for copy 4x VEC so that if length is
exactly equal to 4 * VEC_SIZE it will use the 4x VEC case instead of
8x VEC case.
Results For Skylake memcpy-avx2-erms
size, al1 , al2 , Cur T , New T , Win , New / Cur
128 , 0 , 0 , 9.137 , 6.873 , New , 75.22
128 , 7 , 0 , 12.933 , 7.732 , New , 59.79
128 , 0 , 7 , 11.852 , 6.76 , New , 57.04
128 , 7 , 7 , 12.587 , 6.808 , New , 54.09
Results For Icelake memcpy-evex-erms
size, al1 , al2 , Cur T , New T , Win , New / Cur
128 , 0 , 0 , 9.963 , 5.416 , New , 54.36
128 , 7 , 0 , 16.467 , 8.061 , New , 48.95
128 , 0 , 7 , 14.388 , 7.644 , New , 53.13
128 , 7 , 7 , 14.546 , 7.642 , New , 52.54
Results For Tigerlake memcpy-evex-erms
size, al1 , al2 , Cur T , New T , Win , New / Cur
128 , 0 , 0 , 8.979 , 4.95 , New , 55.13
128 , 7 , 0 , 14.245 , 7.122 , New , 50.0
128 , 0 , 7 , 12.668 , 6.675 , New , 52.69
128 , 7 , 7 , 13.042 , 6.802 , New , 52.15
Results For Skylake memmove-avx2-erms
size, al1 , al2 , Cur T , New T , Win , New / Cur
128 , 0 , 32 , 6.181 , 5.691 , New , 92.07
128 , 32 , 0 , 6.165 , 5.752 , New , 93.3
128 , 0 , 7 , 13.923 , 9.37 , New , 67.3
128 , 7 , 0 , 12.049 , 10.182 , New , 84.5
Results For Icelake memmove-evex-erms
size, al1 , al2 , Cur T , New T , Win , New / Cur
128 , 0 , 32 , 5.479 , 4.889 , New , 89.23
128 , 32 , 0 , 5.127 , 4.911 , New , 95.79
128 , 0 , 7 , 18.885 , 13.547 , New , 71.73
128 , 7 , 0 , 15.565 , 14.436 , New , 92.75
Results For Tigerlake memmove-evex-erms
size, al1 , al2 , Cur T , New T , Win , New / Cur
128 , 0 , 32 , 5.275 , 4.815 , New , 91.28
128 , 32 , 0 , 5.376 , 4.565 , New , 84.91
128 , 0 , 7 , 19.426 , 14.273 , New , 73.47
128 , 7 , 0 , 15.924 , 14.951 , New , 93.89
Noah Goldstein [Sun, 23 May 2021 23:36:36 +0000 (19:36 -0400)]
Bench: Add support for choose direction of memcpy in benchtests
This patch adds support for testing memcpy with both dst > src and dst
< src. Since memcpy is implemented as memmove which has seperate
control flows for certain sizes depending on dst > src it seems like
1) information that should be provided in the benchtest output and a
variable that can be controlled for the benchmarks.
Florian Weimer [Fri, 21 May 2021 20:35:00 +0000 (22:35 +0200)]
nptl: Move pthread_create, thrd_create into libc
The symbols were moved using scripts/move-symbol-to-libc.py.
The libpthread placeholder symbols need some changes because some
symbol versions have gone away completely. But
__errno_location@@GLIBC_2.0 still exists, so the GLIBC_2.0 version
is still there.
The internal __pthread_create symbol now points to the correct
function, so the sysdeps/nptl/thrd_create.c override is no longer
necessary.
There was an issue how the hidden alias of pthread_getattr_default_np
was defined, so this commit cleans up that aspects and removes the
GLIBC_PRIVATE export altogether.
Florian Weimer [Fri, 21 May 2021 20:35:00 +0000 (22:35 +0200)]
nptl: Eliminate the __static_tls_size, __static_tls_align_m1 variables
Use the __nptl_tls_static_size_for_stack inline function instead,
and the GLRO (dl_tls_static_align) value directly.
The computation of GLRO (dl_tls_static_align) in
_dl_determine_tlsoffset ensures that the alignment is at least
TLS_TCB_ALIGN, which at least STACK_ALIGN (see allocate_stack).
Therefore, the additional rounding-up step is removed.
ALso move the initialization of the default stack size from
__pthread_initialize_minimal_internal to __pthread_early_init.
This introduces an extra system call during single-threaded startup,
but this simplifies the initialization sequence. No locking is
needed around the writes to __default_pthread_attr because the
process is single-threaded at this point.
Florian Weimer [Fri, 21 May 2021 20:35:00 +0000 (22:35 +0200)]
nptl: Perform signal initialization upon pthread_create
Install signal handlers and unblock signals before pthread_create
creates the first thread.
create_thread in sysdeps/unix/sysv/linux/createthread.c can send
SIGCANCEL to the current thread, so the SIGCANCEL handler is currently
needed even if pthread_cancel is never called. (The way timer_create
uses SIGCANCEL does not need a signal handler; both SIG_DFL and SIG_IGN
dispositions should work.)
Noah Goldstein [Thu, 20 May 2021 17:13:51 +0000 (13:13 -0400)]
x86: Improve memset-vec-unaligned-erms.S
No bug. This commit makes a few small improvements to
memset-vec-unaligned-erms.S. The changes are 1) only aligning to 64
instead of 128. Either alignment will perform equally well in a loop
and 128 just increases the odds of having to do an extra iteration
which can be significant overhead for small values. 2) Align some
targets and the loop. 3) Remove an ALU from the alignment process. 4)
Reorder the last 4x VEC so that they are stored after the loop. 5)
Move the condition for leq 8x VEC to before the alignment
process. test-memset and test-wmemset are both passing.
Signed-off-by: Noah Goldstein <goldstein.w.n@gmail.com> Reviewed-by: H.J. Lu <hjl.tools@gmail.com>
Noah Goldstein [Mon, 17 May 2021 17:57:24 +0000 (13:57 -0400)]
x86: Optimize memcmp-evex-movbe.S
No bug. This commit optimizes memcmp-evex.S. The optimizations include
adding a new vec compare path for small sizes, reorganizing the entry
control flow, removing some unnecissary ALU instructions from the main
loop, and most importantly replacing the heavy use of vpcmp + kand
logic with vpxor + vptern. test-memcmp and test-wmemcmp are both
passing.
Signed-off-by: Noah Goldstein <goldstein.w.n@gmail.com> Reviewed-by: H.J. Lu <hjl.tools@gmail.com>
Noah Goldstein [Mon, 17 May 2021 17:56:52 +0000 (13:56 -0400)]
x86: Optimize memcmp-avx2-movbe.S
No bug. This commit optimizes memcmp-avx2.S. The optimizations include
adding a new vec compare path for small sizes, reorganizing the entry
control flow, and removing some unnecissary ALU instructions from the
main loop. test-memcmp and test-wmemcmp are both passing.
Signed-off-by: Noah Goldstein <goldstein.w.n@gmail.com> Reviewed-by: H.J. Lu <hjl.tools@gmail.com>
Noah Goldstein [Mon, 17 May 2021 17:56:17 +0000 (13:56 -0400)]
x86: Expand bench-memcmp.c and test-memcmp.c
No bug. This commit adds some additional performance test cases to
bench-memcmp.c and test-memcmp.c. The new benchtests include some
medium range sizes, as well as small sizes near page cross. The new
correctness tests correspond with the new benchtests though add some
additional cases for checking the page cross logic.
Signed-off-by: Noah Goldstein <goldstein.w.n@gmail.com> Reviewed-by: H.J. Lu <hjl.tools@gmail.com>
Joseph Myers [Tue, 18 May 2021 19:47:49 +0000 (19:47 +0000)]
Do not declare asctime_r and ctime_r for C2X
ISO C2X added the asctime_r, ctime_r, gmtime_r and localtime_r
functions from POSIX. It's now removed asctime_r and ctime_r again,
reflecting that they are marked obsolescent in POSIX; update glibc's
time.h accordingly.
The same change that removed those two functions from C2X also marked
asctime and ctime as deprecated (reflecting how POSIX shows them as
obsolescent), i.e. using the [[deprecated]] attribute in the
prototypes shown in C2X. It's less clear if we should explicitly
deprecate those functions like that in the glibc headers; this patch
does nothing regarding such a deprecation (there's no normative
requirement from C2X showing the functions as deprecated).
Stefan Liebler [Mon, 17 May 2021 14:19:36 +0000 (16:19 +0200)]
Fix stringop-overflow warning in bug-regex19.c.
Starting with commit 26492c0a14966c32c43cd6ca1d0dca5e62c6cfef
"Annotate additional APIs with GCC attribute access.",
gcc emits this warning on s390x:
In function 'do_one_test',
inlined from 'do_mb_tests' at bug-regex19.c:385:11:
bug-regex19.c:271:9: error: 're_search' specified size 18446744073709551615 exceeds maximum object size 9223372036854775807 [-Werror=stringop-overflow=]
271 | res = re_search (®buf, test->string, strlen (test->string),
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
272 | test->start, strlen (test->string) - test->start, NULL);
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
In file included from ../include/regex.h:2,
from bug-regex19.c:22:
bug-regex19.c: In function 'do_mb_tests':
../posix/regex.h:554:17: note: in a call to function 're_search' declared with attribute 'read_only (2, 3)'
554 | extern regoff_t re_search (struct re_pattern_buffer *__buffer,
| ^~~~~~~~~
...
The function do_one_test is inlined into do_mb_tests on s390x (at least with
gcc 10). If do_one_test is marked with __attribute__ ((noinline)), there are
no warnings on s390x. If do_one_test is marked with
__attribute__ ((always_inline)), there are the same warnings on x86_64.
test->string points to a variable length array on stack of do_mb_tests
and the content is generated based on the passed test struct.
According to Martin Sebor, this is a false positive caused by the same bug as
the one in nss/makedb.c. It's fixed in GCC 11 and will also be available in
the next GCC 10.4 release.
elf/cache.c: Fix resource leaks identified by static analyzers
A coverity run identified a number of resource leaks in cache.c.
There are a couple of simple memory leaks where a local allocation is
not freed before function return. Then there is a mmap leak and a
file descriptor leak where a map is not unmapped in the error case and
a file descriptor remains open respectively.