external/glibc.git
5 years agoLinux: Add oddly-named arm syscalls to syscall-names.list
Florian Weimer [Fri, 31 May 2019 11:52:16 +0000 (13:52 +0200)]
Linux: Add oddly-named arm syscalls to syscall-names.list

<asm/unistd.h> on arm defines the following macros:

#define __ARM_NR_breakpoint             (__ARM_NR_BASE+1)
#define __ARM_NR_cacheflush             (__ARM_NR_BASE+2)
#define __ARM_NR_usr26                  (__ARM_NR_BASE+3)
#define __ARM_NR_usr32                  (__ARM_NR_BASE+4)
#define __ARM_NR_set_tls                (__ARM_NR_BASE+5)
#define __ARM_NR_get_tls                (__ARM_NR_BASE+6)

These do not follow the regular __NR_* naming convention and
have so far been ignored by the syscall-names.list consistency
checks.  This commit adds these names to the file, preparing
for the availability of these names in the regular __NR_*
namespace.

5 years agopowerpc: Fix build failures with current GCC
Gabriel F. T. Gomes [Mon, 27 May 2019 18:21:22 +0000 (15:21 -0300)]
powerpc: Fix build failures with current GCC

Since GCC commit 271500 (svn), also known as the following commit on the
git mirror:

commit 61edec870f9fdfb5df3fa4e40f28cbaede28a5b1
Author: amodra <amodra@138bc75d-0d04-0410-961f-82ee72b054a4>
Date:   Wed May 22 04:34:26 2019 +0000

    [RS6000] Don't pass -many to the assembler

glibc builds are failing when an assembly implementation does not
declare the correct '.machine' directive, or when no such directive is
declared at all.  For example, when a POWER6 instruction is used, but
'.machine power6' is not declared, the assembler will fail with an error
similar to the following:

    ../sysdeps/powerpc/powerpc64/power8/strcmp.S: Assembler messages:
     24 ../sysdeps/powerpc/powerpc64/power8/strcmp.S:55: Error: unrecognized opcode: `cmpb'

This patch adds '.machine powerN' directives where none existed, as well
as it updates '.machine power7' directives on POWER8 files, because the
minimum binutils version required to build glibc (binutils 2.25) now
provides this machine version.  It also adds '-many' to the assembler
command used to build tst-set_ppr.c.

Tested for powerpc, powerpc64, and powerpc64le, as well as with
build-many-glibcs.py for powerpc targets.

Reviewed-by: Tulio Magno Quites Machado Filho <tuliom@linux.ibm.com>
5 years agoRemove unused get_clockfreq files
Adhemerval Zanella [Wed, 29 May 2019 13:26:30 +0000 (10:26 -0300)]
Remove unused get_clockfreq files

The patch 6e8ba7fd574f meant to remove the all get_clockfreq.c.  This
patch removes the missing files for sparcv9 and x86_64.

Checked against a build to x86_64-linux-gnu and sparcv9-linux-gnu.

* sysdeps/unix/sysv/linux/sparc/sparc32/sparcv9/get_clockfreq.c:
Remove file.
* sysdeps/unix/sysv/linux/x86_64/get_clockfreq.c: Likewise.

5 years agopowerpc: generic nearbyint/nearbyintf
Adhemerval Zanella [Fri, 15 Mar 2019 20:40:51 +0000 (20:40 +0000)]
powerpc: generic nearbyint/nearbyintf

This patches consolidates all the powerpc nearbyint{f} implementations
on the generic sysdeps/powerpc/fpu/s_nearbyint{f}.

* sysdeps/powerpc/fpu/round_to_integer.h (set_fenv_mode): Add
NEARBYINT handling.
* sysdeps/powerpc/fpu/s_nearbyint.c: New file.
* sysdeps/powerpc/fpu/s_nearbyintf.c: Likewise.
* sysdeps/powerpc/powerpc32/fpu/s_nearbyint.S: Remove file.
* sysdeps/powerpc/powerpc32/fpu/s_nearbyintf.S: Likewise.
* sysdeps/powerpc/powerpc64/fpu/s_nearbyint.S: Likewise.
* sysdeps/powerpc/powerpc64/fpu/s_nearbyintf.S: Likewise.

5 years agott_RU: Add lang_name [BZ #24370]
mansayk [Sat, 16 Mar 2019 08:46:39 +0000 (11:46 +0300)]
tt_RU: Add lang_name [BZ #24370]

This commit adds a lang_name according to CLDR-35.1.

[BZ #24370]
* localedata/locales/tt_RU (lang_name): Add from CLDR-35.1.

5 years agott_RU: Fix orthographic mistakes in mon and abmon sections [BZ #24369]
mansayk [Sat, 16 Mar 2019 08:33:30 +0000 (11:33 +0300)]
tt_RU: Fix orthographic mistakes in mon and abmon sections [BZ #24369]

This commit fixes some errors and converts all month names to lowercase.
The content is synchronized with CLDR-35.1 now but trailing dots are
removed from abmon values in order to maintain consistency with the
previous values and with many other locales which do the same.

[BZ #24369]
* localedata/locales/tt_RU (mon): Update from CLDR-35.1, fix errors.
(abmon): Likewise, but remove the trailing dots.

5 years agoAdd IGMP_MRDISC_ADV from Linux 5.1 to netinet/igmp.h.
Joseph Myers [Tue, 28 May 2019 12:01:12 +0000 (12:01 +0000)]
Add IGMP_MRDISC_ADV from Linux 5.1 to netinet/igmp.h.

This patch adds the IGMP_MRDISC_ADV macro from Linux 5.1 to
netinet/igmp.h.

Tested for x86_64.

* inet/netinet/igmp.h (IGMP_MRDISC_ADV): New macro.

5 years agonptl: Add comment to __pthread_get_minstack about external users
Florian Weimer [Mon, 27 May 2019 10:57:45 +0000 (12:57 +0200)]
nptl: Add comment to __pthread_get_minstack about external users

5 years agonss_dns: Check for proper A/AAAA address alignment
Florian Weimer [Fri, 24 May 2019 20:14:04 +0000 (22:14 +0200)]
nss_dns: Check for proper A/AAAA address alignment

Reviewed-by: Carlos O'Donell <carlos@redhat.com>
5 years agoAdd F_SEAL_FUTURE_WRITE from Linux 5.1 to bits/fcntl-linux.h.
Joseph Myers [Thu, 23 May 2019 13:20:48 +0000 (13:20 +0000)]
Add F_SEAL_FUTURE_WRITE from Linux 5.1 to bits/fcntl-linux.h.

This patch adds the new F_SEAL_FUTURE_WRITE constant from Linux 5.1 to
bits/fcntl-linux.h.

Tested for x86_64.

* sysdeps/unix/sysv/linux/bits/fcntl-linux.h [__USE_GNU]
(F_SEAL_FUTURE_WRITE): New macro.

5 years agoelf: Add tst-ldconfig-bad-aux-cache test [BZ #18093]
Alexandra Hájková [Thu, 9 May 2019 11:51:40 +0000 (13:51 +0200)]
elf: Add tst-ldconfig-bad-aux-cache test [BZ #18093]

This test corrupts /var/cache/ldconfig/aux-cache and executes ldconfig
to check it will not segfault using the corrupted aux_cache. The test
uses the test-in-container framework. Verified no regressions on
x86_64.

5 years agoAdd ChangeLog entry for previous commit.
Zack Weinberg [Wed, 22 May 2019 19:09:32 +0000 (15:09 -0400)]
Add ChangeLog entry for previous commit.

5 years agoRemove support for PowerPC SPE extension (powerpc*-*-*gnuspe*).
Zack Weinberg [Thu, 16 May 2019 17:34:27 +0000 (13:34 -0400)]
Remove support for PowerPC SPE extension (powerpc*-*-*gnuspe*).

GCC 9 dropped support for the SPE extensions to PowerPC, which means
powerpc*-*-*gnuspe* configurations are no longer buildable with that
compiler.  This ISA extension was peculiar to the “e500” line of
embedded PowerPC chips, which, as far as I can tell, are no longer
being manufactured, so I think we should follow suit.

This patch was developed by grepping for “e500”, “__SPE__”, and
“__NO_FPRS__”, and may not eliminate every vestige of SPE support.
Most uses of __NO_FPRS__ are left alone, as they are relevant to
normal embedded PowerPC with soft-float.

        * sysdeps/powerpc/preconfigure: Error out on powerpc-*-*gnuspe*
        host type.
        * scripts/build-many-glibcs.py: Remove powerpc-*-linux-gnuspe
        and powerpc-*-linux-gnuspe-e500v1 from list of build configurations.

        * sysdeps/powerpc/powerpc32/e500: Recursively delete.
        * sysdeps/unix/sysv/linux/powerpc/powerpc32/e500: Recursively delete.
        * sysdeps/unix/sysv/linux/powerpc/powerpc32/nofpu/context-e500.h:
        Delete.

        * sysdeps/powerpc/fpu_control.h: Remove SPE variant.
        Issue an #error if used with a compiler in SPE-float mode.
        * sysdeps/powerpc/powerpc32/__longjmp_common.S
        * sysdeps/powerpc/powerpc32/setjmp_common.S
        * sysdeps/unix/sysv/linux/powerpc/powerpc32/getcontext-common.S
        * sysdeps/unix/sysv/linux/powerpc/powerpc32/nofpu/getcontext.S
        * sysdeps/unix/sysv/linux/powerpc/powerpc32/nofpu/setcontext.S
        * sysdeps/unix/sysv/linux/powerpc/powerpc32/nofpu/swapcontext.S
        * sysdeps/unix/sysv/linux/powerpc/powerpc32/setcontext-common.S
        * sysdeps/unix/sysv/linux/powerpc/powerpc32/swapcontext-common.S:
        Remove code to preserve SPE register state.

        * sysdeps/unix/sysv/linux/powerpc/elision-lock.c
        * sysdeps/unix/sysv/linux/powerpc/elision-trylock.c
        * sysdeps/unix/sysv/linux/powerpc/elision-unlock.c
        Remove __SPE__ ifndefs.

5 years agoImprove string benchtest timing
Wilco Dijkstra [Tue, 21 May 2019 14:19:06 +0000 (15:19 +0100)]
Improve string benchtest timing

Improve string benchtest timing.  Many tests run for 0.01s which is way too
short to give accurate results.  Other tests take over 40 seconds which is
way too long.  Significantly increase the iterations of the short running
tests.  Reduce number of alignment variations in the long running memcpy walk
tests so they take less than 5 seconds.

As a result most tests take at least 0.1s and all finish within 5 seconds.

* benchtests/bench-memcpy-random.c (do_one_test): Use medium iterations.
* benchtests/bench-memcpy-walk.c (test_main): Reduce alignment tests.
* benchtests/bench-memmem.c (do_one_test): Use small iterations.
* benchtests/bench-memmove-walk.c (test_main): Reduce alignment tests.
* benchtests/bench-memset-walk.c (test_main): Reduce alignment tests.
* benchtests/bench-strcasestr.c (do_one_test): Use small iterations.
* benchtests/bench-string.h (INNER_LOOP_ITERS): Increase iterations.
(INNER_LOOP_ITERS_MEDIUM): New define.
(INNER_LOOP_ITERS_SMALL): New define.
* benchtests/bench-strpbrk.c (do_one_test): Use medium iterations.
* benchtests/bench-strsep.c (do_one_test): Use small iterations.
* benchtests/bench-strspn.c (do_one_test): Use medium iterations.
* benchtests/bench-strstr.c (do_one_test): Use small iterations.
* benchtests/bench-strtok.c (do_one_test): Use small iterations.

5 years agosysvipc: Add missing bit of semtimedop s390 consolidation
Adhemerval Zanella [Tue, 21 May 2019 13:40:31 +0000 (10:40 -0300)]
sysvipc: Add missing bit of semtimedop s390 consolidation

This patch add the missing SEMTIMEDOP_IPC_ARGS definions on powerpc
and sparc ipc_priv.h.

Checked on powerpc64le-linux-gnu and with a build for sparc64-linux-gnu.

* sysdeps/unix/sysv/linux/powerpc/ipc_priv.h (SEMTIMEDOP_IPC_ARGS):
New define.
* sysdeps/unix/sysv/linux/sparc/sparc64/ipc_priv.h
(SEMTIMEDOP_IPC_ARGS): Likewise.

5 years agowcsmbs: Fix data race in __wcsmbs_clone_conv [BZ #24584]
Florian Weimer [Tue, 21 May 2019 08:19:46 +0000 (10:19 +0200)]
wcsmbs: Fix data race in __wcsmbs_clone_conv [BZ #24584]

This also adds an overflow check and documents the synchronization
requirement in <gconv.h>.

5 years agolibio: Fix gconv-related memory leak [BZ #24583]
Florian Weimer [Tue, 21 May 2019 08:34:21 +0000 (10:34 +0200)]
libio: Fix gconv-related memory leak [BZ #24583]

struct gconv_fcts for the C locale is statically allocated,
and __gconv_close_transform deallocates the steps object.
Therefore this commit introduces __wcsmbs_close_conv to avoid
freeing the statically allocated steps objects.

5 years agolibio: Remove codecvt vtable [BZ #24588]
Florian Weimer [Mon, 20 May 2019 19:54:57 +0000 (21:54 +0200)]
libio: Remove codecvt vtable [BZ #24588]

The codecvt vtable is not a real vtable because it also contains the
conversion state data.  Furthermore, wide stream support was added to
GCC 3.0, after a C++ ABI bump, so there is no compatibility
requirement with libstdc++.

This change removes several unmangled function pointers which could
be used with a corrupted FILE object to redirect execution.  (libio
vtable verification did not cover the codecvt vtable.)

Reviewed-by: Yann Droneaud <ydroneaud@opteya.com>
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
5 years agosupport: Expose sbindir as support_sbindir_prefix
Florian Weimer [Mon, 20 May 2019 19:08:40 +0000 (21:08 +0200)]
support: Expose sbindir as support_sbindir_prefix

5 years agosupport: Add missing EOL terminators on timespec
Mike Crowe [Mon, 20 May 2019 17:56:48 +0000 (14:56 -0300)]
support: Add missing EOL terminators on timespec

The original implementations of test_timespec_before_impl and
test_timespec_equal_or_after in 519839965197291924895a3988804e325035beee
were missing the backslash required for a newline.

Checked on x86_64-linux-gnu.

* support/timespec.c: Add backslash to correct newline in failure
message.

Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
5 years agosupport: Correct confusing comment
Mike Crowe [Mon, 20 May 2019 17:55:33 +0000 (14:55 -0300)]
support: Correct confusing comment

* support/timespec.h: Correct confusing comment.

Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
5 years agosysvipc: Consolidate semtimedop s390
Adhemerval Zanella [Wed, 15 May 2019 13:23:49 +0000 (10:23 -0300)]
sysvipc: Consolidate semtimedop s390

This patch consolidates the s390-32 semtimedop implementation by defining
a arch-specific SEMTIMEDOP_IPC_ARGS to rearrange the arguments expected
by s390 Linux kABI.  The idea is to avoid have multiples semtimedop
implementation changes for Linux v5.1 change to enable wire-up sysvipc
support.

Checked with a s390-linux-gnu and s390x-linux-gnu and checking that
resulting semtimedop objects did not change.

* sysdeps/unix/sysv/linux/ipc_priv.h (SEMTIMEDOP_IPC_ARGS): New
define.
* sysdpes/unix/sysv/linux/s390/ipc_priv.h: New file.
* sysdeps/unix/sysv/linux/s390/semtimedop.c: Remove file.
* sysdeps/unix/sysv/linux/semtimedop.c (semtimedop): Use
SEMTIMEDOP_IPC_ARGS for calls with __NR_ipc.

5 years agosysvipc: Fix compat msgctl (BZ#24570)
Adhemerval Zanella [Tue, 14 May 2019 18:01:32 +0000 (15:01 -0300)]
sysvipc: Fix compat msgctl (BZ#24570)

The __IPC64 flags is meant to be used to enable the new sysv struct
format when the architectures supports it (ARCH_WANT_IPC_PARSE_VERSION
config flag on Linux kernel).

This currently issue only affects alpha.

[BZ #24570]
* sysdeps/unix/sysv/linux/msgctl.c (__old_msgctl): Remove __IPC_64
usage.

5 years agoAdd NT_ARM_PACA_KEYS and NT_ARM_PACG_KEYS from Linux 5.1 to elf.h.
Joseph Myers [Mon, 20 May 2019 11:51:58 +0000 (11:51 +0000)]
Add NT_ARM_PACA_KEYS and NT_ARM_PACG_KEYS from Linux 5.1 to elf.h.

This patch adds the new NT_ARM_PACA_KEYS and NT_ARM_PACG_KEYS from
Linux 5.1 to glibc's elf.h.

Tested for x86_64.

* elf/elf.h (NT_ARM_PACA_KEYS): New macro.
(NT_ARM_PACG_KEYS): Likewise.

5 years agoSmall tcache improvements
Wilco Dijkstra [Fri, 17 May 2019 17:16:20 +0000 (18:16 +0100)]
Small tcache improvements

Change the tcache->counts[] entries to uint16_t - this removes
the limit set by char and allows a larger tcache.  Remove a few
redundant asserts.

bench-malloc-thread with 4 threads is ~15% faster on Cortex-A72.

Reviewed-by: DJ Delorie <dj@redhat.com>
* malloc/malloc.c (MAX_TCACHE_COUNT): Increase to UINT16_MAX.
(tcache_put): Remove redundant assert.
(tcache_get): Remove redundant asserts.
(__libc_malloc): Check tcache count is not zero.
* manual/tunables.texi (glibc.malloc.tcache_count): Update maximum.

5 years agomanual: Document O_DIRECTORY
Florian Weimer [Fri, 17 May 2019 09:14:52 +0000 (11:14 +0200)]
manual: Document O_DIRECTORY

5 years agoUpdate kernel-features.h files for Linux 5.1.
Joseph Myers [Thu, 16 May 2019 20:03:39 +0000 (20:03 +0000)]
Update kernel-features.h files for Linux 5.1.

Linux 5.1 adds missing syscalls to the syscall table for many Linux
kernel architectures.  This patch updates the kernel-features.h
headers accordingly.  __ASSUME_DIRECT_SYSVIPC_SYSCALLS is not updated
because of the differences between new and old syscalls described in
<https://sourceware.org/ml/libc-alpha/2019-05/msg00235.html>.  The
statfs64 structure used by alpha matches what the new kernel syscalls
use.

Tested with build-many-glibcs.py.

* sysdeps/unix/sysv/linux/alpha/kernel-features.h
(__ASSUME_STATFS64): Only undefine if [__LINUX_KERNEL_VERSION <
0x050100].
* sysdeps/unix/sysv/linux/ia64/kernel-features.h (__ASSUME_STATX):
Likewise.
* sysdeps/unix/sysv/linux/sh/kernel-features.h
(__ASSUME_STATX): Likewise.

5 years agonss_nis, nss_nisplus: Remove RES_USE_INET6 handling
Florian Weimer [Thu, 16 May 2019 15:23:23 +0000 (17:23 +0200)]
nss_nis, nss_nisplus: Remove RES_USE_INET6 handling

Since commit 3f8b44be0a658266adff5ece1e4bc3ce097a5dbe ("resolv:
Remove support for RES_USE_INET6 and the inet6 option"),
res_use_inet6 () always evaluates to false.

5 years agonss_files: Remove RES_USE_INET6 from hosts processing
Florian Weimer [Thu, 16 May 2019 13:11:23 +0000 (15:11 +0200)]
nss_files: Remove RES_USE_INET6 from hosts processing

Since commit 3f8b44be0a658266adff5ece1e4bc3ce097a5dbe ("resolv:
Remove support for RES_USE_INET6 and the inet6 option"),
res_use_inet6 () always evaluates to false.

5 years agosupport: Report NULL blobs explicitly in TEST_COMPARE
Florian Weimer [Thu, 16 May 2019 12:50:15 +0000 (14:50 +0200)]
support: Report NULL blobs explicitly in TEST_COMPARE

Provide an explicit diagnostic if the length is positive, and
do not just crash with a null pointer dereference.  Null pointers
are only valid if the length is zero, so this can only happen with
a faulty test.

5 years agodlfcn: Guard __dlerror_main_freeres with __libc_once_get (once) [BZ# 24476]
Mark Wielaard [Wed, 15 May 2019 15:14:01 +0000 (17:14 +0200)]
dlfcn: Guard __dlerror_main_freeres with __libc_once_get (once) [BZ# 24476]

dlerror.c (__dlerror_main_freeres) will try to free resources which only
have been initialized when init () has been called. That function is
called when resources are needed using __libc_once (once, init) where
once is a __libc_once_define (static, once) in the dlerror.c file.
Trying to free those resources if init () hasn't been called will
produce errors under valgrind memcheck. So guard the freeing of those
resources using __libc_once_get (once) and make sure we have a valid
key. Also add a similar guard to __dlerror ().

* dlfcn/dlerror.c (__dlerror_main_freeres): Guard using
__libc_once_get (once) and static_bug == NULL.
(__dlerror): Check we have a valid key, set result to static_buf
otherwise.

Reviewed-by: Carlos O'Donell <carlos@redhat.com>
5 years agoAdd missing Changelog entry
DJ Delorie [Wed, 15 May 2019 16:11:12 +0000 (12:11 -0400)]
Add missing Changelog entry

For 5b06f538c5aee0389ed034f60d90a8884d6d54de

5 years agoFix crash in _IO_wfile_sync (bug 20568)
Andreas Schwab [Tue, 14 May 2019 15:14:59 +0000 (17:14 +0200)]
Fix crash in _IO_wfile_sync (bug 20568)

When computing the length of the converted part of the stdio buffer, use
the number of consumed wide characters, not the (negative) distance to the
end of the wide buffer.

5 years agonss: Turn __nss_database_lookup into a compatibility symbol
Florian Weimer [Wed, 15 May 2019 11:51:35 +0000 (13:51 +0200)]
nss: Turn __nss_database_lookup into a compatibility symbol

The function uses the internal service_user type, so it is not
really usable from the outside of glibc.  Rename the function
to __nss_database_lookup2 for internal use, and change
__nss_database_lookup to always indicate failure to the caller.

__nss_next already was a compatibility symbol.  The new
implementation always fails and no longer calls __nss_next2.

unscd, the alternative nscd implementation, does not use
__nss_database_lookup, so it is not affected by this change.

5 years agosupport: Add support_install_rootsbindir
Alexandra Hájková [Mon, 13 May 2019 17:31:53 +0000 (19:31 +0200)]
support: Add support_install_rootsbindir

Reviewed by: Adhemerval Zanella <adhemerval.zanella@linaro.org>

5 years agoiconv: Remove public declaration of __gconv_transliterate
Florian Weimer [Wed, 15 May 2019 05:16:47 +0000 (07:16 +0200)]
iconv: Remove public declaration of __gconv_transliterate

Commit ba7b4d294b01870ce3497971e9d07ee261cdc540 ("Complete the
removal of __gconv_translit_find") added a declaration of the
GLIBC_PRIVATE function, __gconv_transliterate, to the installed
header <gconv.h>.  It should have been added to the internal
<gconv_int.h> header.

Reviewed-by: Carlos O'Donell <carlos@redhat.com>
5 years agoLinux: Add the tgkill function
Florian Weimer [Tue, 14 May 2019 19:35:09 +0000 (21:35 +0200)]
Linux: Add the tgkill function

The tgkill function is sometimes used in crash handlers.

<bits/signal_ext.h> follows the same approach as <bits/unistd_ext.h>
(which was added for the gettid system call wrapper).

Reviewed-by: Carlos O'Donell <carlos@redhat.com>
5 years agomanual: Adjust twalk_r documentation.
Carlos O'Donell [Tue, 14 May 2019 19:33:02 +0000 (15:33 -0400)]
manual: Adjust twalk_r documentation.

Reviewed-by: Florian Weimer <fweimer@redhat.com>
5 years agoelf: Fix tst-pldd for non-default --prefix and/or --bindir (BZ#24544)
Adhemerval Zanella [Mon, 13 May 2019 16:13:46 +0000 (13:13 -0300)]
elf: Fix tst-pldd for non-default --prefix and/or --bindir (BZ#24544)

Use a new libsupport support_bindir_prefix instead of a hardcoded
/usr/bin to create the pldd path on container directory.

Checked on x86_64-linux-gnu with default and non-default --prefix and
--bindir paths, as well with --enable-hardcoded-path-in-tests.

[BZ #24544]
* elf/tst-pldd.c (do_test): Use support_bindir_prefix instead of
pre-defined value.

Reviewed-by: DJ Delorie <dj@redhat.com>
5 years agosupport: Export bindir path on support_path
Adhemerval Zanella [Mon, 13 May 2019 16:57:37 +0000 (13:57 -0300)]
support: Export bindir path on support_path

Checked on x86_64-linux-gnu.

* support/Makefile (CFLAGS-support_paths.c): Add -DBINDIR_PATH.
* support/support.h (support_bindir_prefix): New variable.
* support/support_paths.c [BINDIR_PATH] (support_bindir_prefix):

Reviewed-by: DJ Delorie <dj@redhat.com>
5 years agoMake --bindir effective
Adhemerval Zanella [Mon, 13 May 2019 16:54:49 +0000 (13:54 -0300)]
Make --bindir effective

This allows sets a path using --bindir.  Checked on x86_64-linux-gnu
with a non-default --bindir and checked resulting installed binaries
(pldd for instance).

* config.make.in (bindir): New variable.

Reviewed-by: DJ Delorie <dj@redhat.com>
5 years agox86: Remove arch-specific low level lock implementation
Adhemerval Zanella [Thu, 21 Feb 2019 16:33:54 +0000 (13:33 -0300)]
x86: Remove arch-specific low level lock implementation

This patch removes the arch-specific x86 assembly implementation for
low level locking and consolidate both 64 bits and 32 bits in a
single implementation.

Different than other architectures, x86 lll_trylock, lll_lock, and
lll_unlock implements a single-thread optimization to avoid atomic
operation, using cmpxchgl instead.  This patch implements by using
the new single-thread.h definitions in a generic way, although using
the previous semantic.

The lll_cond_trylock, lll_cond_lock, and lll_timedlock just use
atomic operations plus calls to lll_lock_wait*.

For __lll_lock_wait_private and __lll_lock_wait the generic implemtation
there is no indication that assembly implementation is required
performance-wise.

Checked on x86_64-linux-gnu and i686-linux-gnu.

* sysdeps/nptl/lowlevellock.h (__lll_trylock): New macro.
(lll_trylock): Call __lll_trylock.
* sysdeps/unix/sysv/linux/i386/libc-lowlevellock.S: Remove file.
* sysdeps/unix/sysv/linux/i386/lll_timedlock_wait.c: Likewise.
* sysdeps/unix/sysv/linux/i386/lowlevellock.S: Likewise.
* sysdeps/unix/sysv/linux/i386/lowlevellock.h: Likewise.
* sysdeps/unix/sysv/linux/x86_64/libc-lowlevellock.S: Likewise.
* sysdeps/unix/sysv/linux/x86_64/lll_timedlock_wait.c: Likewise.
* sysdeps/unix/sysv/linux/x86_64/lowlevellock.S: Likewise.
* sysdeps/unix/sysv/linux/x86_64/lowlevellock.h: Likewise.
* sysdeps/unix/sysv/linux/x86/lowlevellock.h: New file.
* sysdeps/unix/sysv/linux/x86_64/cancellation.S: Include
lowlevellock-futex.h.

5 years agoAssume LLL_LOCK_INITIALIZER is 0
Adhemerval Zanella [Fri, 22 Feb 2019 15:10:26 +0000 (12:10 -0300)]
Assume LLL_LOCK_INITIALIZER is 0

Since hppa is not an outlier anymore regarding LLL_LOCK_INITIALIZER value,
we can now assume it 0 for all architectures.

Checked on a build for all major ABIs.

* nptl/nptl-init.c (__pthread_initialize_minimal_internal): Remove
initialization for LLL_LOCK_INITIALIZER different than 0.
* nptl/old_pthread_cond_broadcast.c (__pthread_cond_broadcast_2_0):
Assume LLL_LOCK_INITIALIZER being 0.
* nptl/old_pthread_cond_signal.c (__pthread_cond_signal_2_0): Likewise.
* nptl/old_pthread_cond_timedwait.c (__pthread_cond_timedwait_2_0):
Likewise.
* nptl/old_pthread_cond_wait.c (__pthread_cond_wait_2_0): Likewise.
* sysdeps/nptl/libc-lockP.h (__libc_lock_define_initialized): Likewise.

5 years agoSmall optimization for lowlevellock
Adhemerval Zanella [Fri, 22 Feb 2019 13:27:08 +0000 (10:27 -0300)]
Small optimization for lowlevellock

This patch optimizes both __lll_lock_wait_private and __lll_lock_wait
by issuing only one lll_futex_wait.  Since it is defined as an inlined
syscall and inlined syscalls are defined using inlined assembly the
compiler usually can not see both calls are equal and optimize
accordingly.

On aarch64 the resulting binary is change from:

0000000000000060 <__lll_lock_wait>:
  60:   2a0103e5        mov     w5, w1
  64:   b9400001        ldr     w1, [x0]
  68:   aa0003e4        mov     x4, x0
  6c:   7100083f        cmp     w1, #0x2
  70:   540000e1        b.ne    8c <__lll_lock_wait+0x2c>  // b.any
  74:   521900a1        eor     w1, w5, #0x80
  78:   d2800042        mov     x2, #0x2                        // #2
  7c:   93407c21        sxtw    x1, w1
  80:   d2800003        mov     x3, #0x0                        // #0
  84:   d2800c48        mov     x8, #0x62                       // #98
  88:   d4000001        svc     #0x0
  8c:   521900a5        eor     w5, w5, #0x80
  90:   52800046        mov     w6, #0x2                        // #2
  94:   93407ca5        sxtw    x5, w5
  98:   14000008        b       b8 <__lll_lock_wait+0x58>
  9c:   d503201f        nop
  a0:   aa0403e0        mov     x0, x4
  a4:   aa0503e1        mov     x1, x5
  a8:   d2800042        mov     x2, #0x2                        // #2
  ac:   d2800003        mov     x3, #0x0                        // #0
  b0:   d2800c48        mov     x8, #0x62                       // #98
  b4:   d4000001        svc     #0x0
  b8:   885ffc80        ldaxr   w0, [x4]
  bc:   88017c86        stxr    w1, w6, [x4]
  c0:   35ffffc1        cbnz    w1, b8 <__lll_lock_wait+0x58>
  c4:   35fffee0        cbnz    w0, a0 <__lll_lock_wait+0x40>
  c8:   d65f03c0        ret

To:

0000000000000048 <__lll_lock_wait>:
  48:   aa0003e4        mov     x4, x0
  4c:   2a0103e5        mov     w5, w1
  50:   b9400000        ldr     w0, [x0]
  54:   7100081f        cmp     w0, #0x2
  58:   540000c0        b.eq    70 <__lll_lock_wait+0x28>  // b.none
  5c:   52800041        mov     w1, #0x2                        // #2
  60:   885ffc80        ldaxr   w0, [x4]
  64:   88027c81        stxr    w2, w1, [x4]
  68:   35ffffc2        cbnz    w2, 60 <__lll_lock_wait+0x18>
  6c:   34000120        cbz     w0, 90 <__lll_lock_wait+0x48>
  70:   521900a1        eor     w1, w5, #0x80
  74:   aa0403e0        mov     x0, x4
  78:   93407c21        sxtw    x1, w1
  7c:   d2800042        mov     x2, #0x2                        // #2
  80:   d2800003        mov     x3, #0x0                        // #0
  84:   d2800c48        mov     x8, #0x62                       // #98
  88:   d4000001        svc     #0x0
  8c:   17fffff4        b       5c <__lll_lock_wait+0x14>
  90:   d65f03c0        ret

I see similar changes on powerpc and other architectures.  It also aligns
with x86_64 implementation by adding the systemtap probes.

Checker on aarch64-linux-gnu.

* nptl/lowlevellock.c (__lll_lock_wait, __lll_lock_wait_private):
Optimize futex call and add systemtap probe.

5 years agoAdd single-thread.h header
Adhemerval Zanella [Thu, 21 Feb 2019 17:33:55 +0000 (14:33 -0300)]
Add single-thread.h header

This patch move the single-thread syscall optimization defintions from
syscall-cancel.h to new header file single-thread.h and also move the
cancellation definitions from pthreadP.h to syscall-cancel.h.

The idea is just simplify the inclusion of both syscall-cancel.h and
single-thread.h (without the requirement of including all pthreadP.h
defintions).

No semantic changes expected, checked on a build for all major ABIs.

* nptl/pthreadP.h (CANCEL_ASYNC, CANCEL_RESET, LIBC_CANCEL_ASYNC,
LIBC_CANCEL_RESET, __libc_enable_asynccancel,
__libc_disable_asynccancel, __librt_enable_asynccancel,
__libc_disable_asynccancel, __librt_enable_asynccancel,
__librt_disable_asynccancel): Move to ...
* sysdeps/unix/sysv/linux/sysdep-cancel.h: ... here.
(SINGLE_THREAD_P, RTLD_SINGLE_THREAD_P): Move to ...
* sysdeps/unix/sysv/linux/single-thread.h: ... here.
* sysdeps/generic/single-thread.h: New file.
* sysdeps/unix/sysdep.h: Include single-thread.h.
* sysdeps/unix/sysv/linux/futex-internal.h: Include sysdep-cancel.h.
* sysdeps/unix/sysv/linux/lowlevellock-futex.h: Likewise.

5 years agoBug 24535: Update to Unicode 12.1.0
Mike FABIAN [Wed, 8 May 2019 05:54:15 +0000 (07:54 +0200)]
Bug 24535: Update to Unicode 12.1.0

Unicode 12.1.0 Support: Character encoding, character type info, and
transliteration tables are all updated to Unicode 12.1.0, using
the generator scripts contributed by Mike FABIAN (Red Hat).

Some info about the number of characters added or changed:

Total added characters in newly generated CHARMAP: 1
added: <U32FF>     /xe3/x8b/xbf SQUARE ERA NAME REIWA
Total added characters in newly generated WIDTH: 1
added: <U32FF> 2 : eaw=W category=So bidi=L   name=SQUARE ERA NAME REIWA
graph: Added 1 characters in new ctype which were not in old ctype
graph: Added: ㋿ U+32FF SQUARE ERA NAME REIWA
print: Added 1 characters in new ctype which were not in old ctype
print: Added: ㋿ U+32FF SQUARE ERA NAME REIWA
punct: Added 1 characters in new ctype which were not in old ctype
punct: Added: ㋿ U+32FF SQUARE ERA NAME REIWA

5 years agoFix tcache count maximum (BZ #24531)
Wilco Dijkstra [Fri, 10 May 2019 15:38:21 +0000 (16:38 +0100)]
Fix tcache count maximum (BZ #24531)

The tcache counts[] array is a char, which has a very small range and thus
may overflow.  When setting tcache_count tunable, there is no overflow check.
However the tunable must not be larger than the maximum value of the tcache
counts[] array, otherwise it can overflow when filling the tcache.

[BZ #24531]
* malloc/malloc.c (MAX_TCACHE_COUNT): New define.
(do_set_tcache_count): Only update if count is small enough.
* manual/tunables.texi (glibc.malloc.tcache_count): Document max value.

5 years agosem_close: Use __twalk_r
Florian Weimer [Fri, 10 May 2019 12:17:26 +0000 (14:17 +0200)]
sem_close: Use __twalk_r

5 years agosupport: Fix timespec printf
Adhemerval Zanella [Fri, 10 May 2019 12:35:18 +0000 (09:35 -0300)]
support: Fix timespec printf

The patch print timespec members as intmax_t instead of long int.
It avoid the -Werror=format= build issue on x32:

  timespec.c: In function 'test_timespec_before_impl':
  timespec.c:32:23: error: format '%ld' expects argument of type 'long int',
  but argument 4 has type '__time_t' {aka 'const long long int'} [-Werror=format=]

Checked on x86_64-linux-gnu-x32, x86_64-linux-gnu, and i686-linux-gnu.

* support/timespec.c (test_timespec_before_impl,
test_timespec_equal_or_after_impl): print timespec member as intmax_t
insted of long int.

Reviewed-by: Florian Weimer <fweimer@redhat.com>
5 years agonptl/tst-abstime: Use libsupport
Mike Crowe [Thu, 9 May 2019 17:17:23 +0000 (14:17 -0300)]
nptl/tst-abstime: Use libsupport

Checked on x86_64-linux-gnu and i686-linux-gnu.

* nptl/tst-abstime.c: Use libsupport.

Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
5 years agonptl: Convert some rwlock tests to use libsupport
Mike Crowe [Thu, 9 May 2019 17:14:32 +0000 (14:14 -0300)]
nptl: Convert some rwlock tests to use libsupport

Checked on x86_64-linux-gnu and i686-linux-gnu.

* nptl/tst-rwlock6.c: Use libsupport. This also happens to fix a
small bug where only tv.tv_usec was checked which could cause an
erroneous pass if pthread_rwlock_timedrdlock incorrectly took more
than a second.

* nptl/tst-rwlock7.c, nptl/tst-rwlock9.c, nptl/tst-rwlock14.c: Use
libsupport.

Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
5 years agonptl: Use recent additions to libsupport in tst-sem5
Mike Crowe [Thu, 9 May 2019 17:06:06 +0000 (14:06 -0300)]
nptl: Use recent additions to libsupport in tst-sem5

Checked on x86_64-linux-gnu and i686-linux-gnu.

* nptl/tst-sem5.c(do_test): Use xclock_gettime, timespec_add and
TEST_TIMESPEC_NOW_OR_AFTER from libsupport.

Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
5 years agonptl: Convert tst-cond11.c to use libsupport
Mike Crowe [Thu, 9 May 2019 17:22:25 +0000 (14:22 -0300)]
nptl: Convert tst-cond11.c to use libsupport

Checked on x86_64-linux-gnu and i686-linux-gnu.

* nptl/tst-cond11.c: Use libsupport.

Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
5 years agosupport: Add timespec.h
Mike Crowe [Thu, 9 May 2019 17:19:21 +0000 (14:19 -0300)]
support: Add timespec.h

It adds useful functions for tests that use struct timespec.

Checked on x86_64-linux-gnu and i686-linux-gnu.

* support/timespec.h: New file.  Provide timespec helper functions
along with macros in the style of those in check.h.
* support/timespec.c: New file.  Implement check functions declared
in support/timespec.h.
* support/timespec-add.c: New file from gnulib containing
timespec_add implementation that handles overflow.
* support/timespec-sub.c: New file from gnulib containing
timespec_sub implementation that handles overflow.
* support/README: Mention timespec.h.

Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
5 years agoMove nptl/tst-eintr1 to xtests
Szabolcs Nagy [Wed, 8 May 2019 13:50:47 +0000 (14:50 +0100)]
Move nptl/tst-eintr1 to xtests

Don't run nptl/tst-eintr1 by normal make check because it can spuriously
break testing on various linux kernels. (Currently this affects the
aarch64 glibc buildbot machine which regularly fails and loses test
results.)

[BZ #24537]
* nptl/Makefile: Move tst-eintr1 to xtests.

5 years agopowerpc: trunc/truncf refactor
Adhemerval Zanella [Fri, 15 Mar 2019 20:37:23 +0000 (20:37 +0000)]
powerpc: trunc/truncf refactor

This patches consolidates all the powerpc trunc{f} implementations on
the generic sysdeps/powerpc/fpu/s_trunc{f}.  The generic implementation
uses either the compiler builts for ISA 2.03+ (which generates the
frim instruction) or a generic implementation which uses FP only
operations.

The IFUNC organization for powerpc64 is also change to be enabled only
for powerpc64 and not for powerpc64le (since minium ISA of 2.08 does not
require the fallback generic implementation).

Checked on powerpc-linux-gnu (built without --with-cpu, with
--with-cpu=power4 and with --with-cpu=power5+ and --disable-multi-arch),
powerpc64-linux-gnu (built without --with-cp and with --with-cpu=power5+
and --disable-multi-arch).

* sysdeps/powerpc/fpu/trunc_to_integer.h (set_fenv_mode): Add
 TRUNC handling.
(round_mode): Add definition for TRUNC.
* sysdeps/powerpc/fpu/s_trunc.c: New file.
* sysdeps/powerpc/fpu/s_truncf.c: New file.
* sysdeps/powerpc/powerpc32/fpu/s_trunc.S: Remove file.
* sysdeps/powerpc/powerpc32/fpu/s_truncf.S: Likewise.
* sysdep/powerpc/powepc32/power4/fpu/multiarch/s_trunc-power5+.S:
Likewise.
* sysdep/powerpc/powepc32/power4/fpu/multiarch/s_trunc-ppc32.S:
Likewise.
* sysdep/powerpc/powepc32/power4/fpu/multiarch/s_truncf-power5+.S:
Likewise.
* sysdep/powerpc/powepc32/power4/fpu/multiarch/s_truncf-ppc32.S:
Likewise.
* sysdep/powerpc/powepc32/power4/fpu/multiarch/s_trunc-power5+.c: New
file.
* sysdep/powerpc/powepc32/power4/fpu/multiarch/s_trunc-ppc32.c:
Likewise.
* sysdep/powerpc/powepc32/power4/fpu/multiarch/s_truncf-power5+.c:
Likewise.
* sysdep/powerpc/powepc32/power4/fpu/multiarch/s_truncf-ppc32.c:
Likewise.
* sysdep/powerpc/powerpc32/power5+/fpu/s_trunc.S: Remove file.
* sysdep/powerpc/powerpc32/power5+/fpu/s_truncf.S: Likewise.
* sysdep/powerpc/powerpc64/be/fpu/multiarch/Makefile
(libm-sysdep_routines): Add s_trunc-power5+, s_trunc-ppc64,
s_truncf-power5+, and s_truncf-ppc64.
(CFLAGS-s_trunc-power5+.c, CFLAGS-s_truncf-power5+.c): New rule.
* sysdep/powerpc/powercp64/be/fpu/multiarch/s_trunc-power5+.c: New
file.
* sysdep/powerpc/powercp64/be/fpu/multiarch/s_trunc-ppc64.c: Likewise.
* sysdeps/powerpc/powerpc64/fpu/multiarch/s_trunc.c: Move to ...
* sysdeps/powerpc/powerpc64/be/fpu/multiarch/s_trunc.c: ... here.
* sysdep/powerpc/powercp64/be/fpu/multiarch/s_truncf-power5+.c: New
file.
* sysdep/powerpc/powercp64/be/fpu/multiarch/s_truncf-ppc64.c:
Likewise.
* sysdeps/powerpc/powerpc64/fpu/multiarch/s_truncf.c: Move to ...
* sysdeps/powerpc/powerpc64/be/fpu/multiarch/s_truncf.c: ... here.
* sysdeps/powerpc/powerpc64/fpu/multiarch/Makefile
(libm-sysdep_routines): Remove s_trunc-power5+, s_trunc-ppc64,
s_truncf-power5+, and s_truncf-ppc64.
* sysdep/powerpc/powerpc64/fpu/multiarch/s_trunc-power5+.S: Remove
file.
* sysdep/powerpc/powerpc64/fpu/multiarch/s_trunc-ppc64.S: Likewise.
* sysdep/powerpc/powerpc64/fpu/multiarch/s_truncf-power5+.S:
Likewise.
* sysdep/powerpc/powerpc64/fpu/multiarch/s_truncf-ppc64.S: Likewise.
* sysdeps/powerpc/powerpc64/fpu/s_trunc.S: Likewise.
* sysdeps/powerpc/powerpc64/fpu/s_truncf.S: Likewise.
* sysdep/powerpc/powerpc64/power5+/fpu/s_trunc.S: Likewise.
* sysdep/powerpc/powerpc64/power5+/fpu/s_truncf.S: Likewise.

Reviewed-by: Gabriel F. T. Gomes <gabriel@inconstante.eti.br>
5 years agopowerpc: round/roundf refactor
Adhemerval Zanella [Fri, 8 Mar 2019 15:20:08 +0000 (15:20 +0000)]
powerpc: round/roundf refactor

This patches consolidates all the powerpc round{f} implementations on
the generic sysdeps/powerpc/fpu/s_round{f}.  The generic implementation
uses either the compiler builts for ISA 2.03+ (which generates the
frim instruction) or a generic implementation which uses FP only
operations.

The IFUNC organization for powerpc64 is also change to be enabled only
for powerpc64 and not for powerpc64le (since minium ISA of 2.08 does not
require the fallback generic implementation).

Checked on powerpc-linux-gnu (built without --with-cpu, with
--with-cpu=power4 and with --with-cpu=power5+ and --disable-multi-arch),
powerpc64-linux-gnu (built without --with-cp and with --with-cpu=power5+
and --disable-multi-arch).

* sysdeps/powerpc/fpu/round_to_integer.h (set_fenv_mode): Add
ROUND handling.
(round_mode): Add definition for ROUND.
(round_to_integer_float): Likewise.
* sysdeps/powerpc/fpu/s_round.c: New file.
* sysdeps/powerpc/fpu/s_roundf.c: New file.
* sysdeps/powerpc/powerpc32/fpu/s_round.S: Remove file.
* sysdeps/powerpc/powerpc32/fpu/s_roundf.S: Likewise.
* sysdep/powerpc/powepc32/power4/fpu/multiarch/s_round-power5+.S:
Likewise.
* sysdep/powerpc/powepc32/power4/fpu/multiarch/s_round-ppc32.S:
Likewise.
* sysdep/powerpc/powepc32/power4/fpu/multiarch/s_roundf-power5+.S:
Likewise.
* sysdep/powerpc/powepc32/power4/fpu/multiarch/s_roundf-ppc32.S:
Likewise.
* sysdep/powerpc/powepc32/power4/fpu/multiarch/s_round-power5+.c: New
file.
* sysdep/powerpc/powepc32/power4/fpu/multiarch/s_round-ppc32.c:
Likewise.
* sysdep/powerpc/powepc32/power4/fpu/multiarch/s_roundf-power5+.c:
Likewise.
* sysdep/powerpc/powepc32/power4/fpu/multiarch/s_roundf-ppc32.c:
Likewise.
* sysdep/powerpc/powerpc32/power5+/fpu/s_round.S: Remove file.
* sysdep/powerpc/powerpc32/power5+/fpu/s_roundf.S: Likewise.
* sysdep/powerpc/powerpc64/be/fpu/multiarch/Makefile
(libm-sysdep_routines): Add s_round-power5+, s_round-ppc64,
s_roundf-power5+, and s_roundf-ppc64.
(CFLAGS-s_round-power5+.c, CFLAGS-s_roundf-power5+.c): New rule.
* sysdep/powerpc/powercp64/be/fpu/multiarch/s_round-power5+.c: New
file.
* sysdep/powerpc/powercp64/be/fpu/multiarch/s_round-ppc64.c: Likewise.
* sysdeps/powerpc/powerpc64/fpu/multiarch/s_round.c: Move to ...
* sysdeps/powerpc/powerpc64/be/fpu/multiarch/s_round.c: ... here.
* sysdep/powerpc/powercp64/be/fpu/multiarch/s_roundf-power5+.c: New
file.
* sysdep/powerpc/powercp64/be/fpu/multiarch/s_roundf-ppc64.c:
Likewise.
* sysdeps/powerpc/powerpc64/fpu/multiarch/s_roundf.c: Move to ...
* sysdeps/powerpc/powerpc64/be/fpu/multiarch/s_roundf.c: ... here.
* sysdeps/powerpc/powerpc64/fpu/multiarch/Makefile
(libm-sysdep_routines): Remove s_round-power5+, s_round-ppc64,
s_roundf-power5+, and s_roundf-ppc64.
* sysdep/powerpc/powerpc64/fpu/multiarch/s_round-power5+.S: Remove
file.
* sysdep/powerpc/powerpc64/fpu/multiarch/s_round-ppc64.S: Likewise.
* sysdep/powerpc/powerpc64/fpu/multiarch/s_roundf-power5+.S:
Likewise.
* sysdep/powerpc/powerpc64/fpu/multiarch/s_roundf-ppc64.S: Likewise.
* sysdeps/powerpc/powerpc64/fpu/s_round.S: Likewise.
* sysdeps/powerpc/powerpc64/fpu/s_roundf.S: Likewise.
* sysdep/powerpc/powerpc64/power5+/fpu/s_round.S: Likewise.
* sysdep/powerpc/powerpc64/power5+/fpu/s_roundf.S: Likewise.

Reviewed-by: Gabriel F. T. Gomes <gabriel@inconstante.eti.br>
5 years agopowerpc: floor/floorf refactor
Adhemerval Zanella [Fri, 8 Mar 2019 15:04:02 +0000 (15:04 +0000)]
powerpc: floor/floorf refactor

This patches consolidates all the powerpc floor{f} implementations on
the generic sysdeps/powerpc/fpu/s_floor{f}.  The generic implementation
uses either the compiler builts for ISA 2.03+ (which generates the
frim instruction) or a generic implementation which uses FP only
operations.

The IFUNC organization for powerpc64 is also change to be enabled only
for powerpc64 and not for powerpc64le (since minium ISA of 2.08 does not
require the fallback generic implementation).

Checked on powerpc-linux-gnu (built without --with-cpu, with
--with-cpu=power4 and with --with-cpu=power5+ and --disable-multi-arch),
powerpc64-linux-gnu (built without --with-cp and with --with-cpu=power5+
and --disable-multi-arch).

* sysdeps/powerpc/fpu/round_to_integer.h (set_fenv_mode):
Add FLOOR option.
(round_mode): Add definition for FLOOR.
* sysdeps/powerpc/fpu/s_floor.c: New file.
* sysdeps/powerpc/fpu/s_floorf.c: Likewise.
* sysdeps/powerpc/powerpc32/fpu/s_floor.S: Remove file.
* sysdeps/powerpc/powerpc32/fpu/s_floorf.S: Likewise.
* sysdeps/powerpc/powerpc32/power4/fpu/multiarch/s_floor-power5+.S:
Remove file.
* sysdeps/powerpc/powerpc32/power4/fpu/multiarch/s_floor-ppc32.S:
Likewise
* sysdeps/powerpc/powerpc32/power4/fpu/multiarch/s_floorf-power5+.S:
Likewise.
* sysdeps/powerpc/powerpc32/power4/fpu/multiarch/s_floorf-ppc32.S:
Likewise.
* sysdeps/powerpc/powerpc32/power4/fpu/multiarch/s_floor-power5+.c:
New file.
* sysdeps/powerpc/powerpc32/power4/fpu/multiarch/s_floor-ppc32.c:
Likewise.
* sysdeps/powerpc/powerpc32/power4/fpu/multiarch/s_floorf-power5+.c:
Likewise.
* sysdeps/powerpc/powerpc32/power4/fpu/multiarch/s_floorf-ppc32.c:
Likewise.
* sysdeps/powerpc/powerpc32/power5+/fpu/s_floor.S: Remove file.
* sysdeps/powerpc/powerpc32/power5+/fpu/s_floorf.S: Remove file.
* sysdeps/powerpc/powerpc64/be/fpu/multiarch/Makefile
(libm-sysdep_routines): Add s_floor-power5+, s_floor-ppc64,
s_floorf-power5+, and s_floorf-ppc64.
(CFLAGS-s_floor-power5+.c, CFLAGS-s_floorf-power5+.c): New rule.
* sysdep/powerpc/powerpc64/be/fpu/multiarch/s_floor-power5+.c: New
file.
* sysdep/powerpc/powerpc64/be/fpu/multiarch/s_floor-ppc64.c: Likewise.
* sysdeps/powerpc/powerpc64/fpu/multiarch/s_floor.c: Move to ...
* sysdeps/powerpc/powerpc64/be/fpu/multiarch/s_floor.c: ... here.
* sysdep/powerpc/powerpc64/be/fpu/multiarch/s_floorf-power5+.c: New
file.
* sysdep/powerpc/powerpc64/be/fpu/multiarch/s_floorf-ppc64.c:
Likewise.
* sysdeps/powerpc/powerpc64/fpu/multiarch/s_floorf.c: Move to ...
* sysdeps/powerpc/powerpc64/be/fpu/multiarch/s_floorf.c: ... here.
* sysdeps/powerpc/powerpc64/fpu/multiarch/Makefile
(libm-sysdep_routines): Remove s_floor-power5+, s_floor-ppc64,
s_floorf-power5+, and s_floorf-ppc64.
* sysdep/powerpc/powerpc64/fpu/multiarch/s_floor-power5+.S: Remove
file.
* sysdeps/powerpc/powerpc64/fpu/multiarch/s_floor-ppc64.S: Remove
file.
* sysdeps/powerpc/powerpc64/fpu/multiarch/s_floorf-power5+.S:
Likewise.
* sysdeps/powerpc/powerpc64/fpu/multiarch/s_floorf-ppc64.S:
Likewise.
* sysdeps/powerpc/powerpc64/fpu/s_floor.S: Likewise.
* sysdeps/powerpc/powerpc64/fpu/s_floorf.S: Likewise.
* sysdeps/powerpc/powerpc64/power5+/fpu/s_floor.S: Likewise.
* sysdeps/powerpc/powerpc64/power5+/fpu/s_floorf.S: Likewise.

Reviewed-by: Gabriel F. T. Gomes <gabriel@inconstante.eti.br>
5 years agosupport: Add xclock_gettime
Mike Crowe [Wed, 8 May 2019 14:58:17 +0000 (16:58 +0200)]
support: Add xclock_gettime

* support/xclock_gettime.c (xclock_gettime): New file. Provide
clock_gettime wrapper for use in tests that fails the test rather
than returning failure.

* support/xtime.h: New file to declare xclock_gettime.

* support/Makefile: Add xclock_gettime.c.

* support/README: Mention xtime.h.

5 years agomalloc/tst-mallocfork2: Use process-shared barriers
Florian Weimer [Wed, 8 May 2019 13:29:13 +0000 (15:29 +0200)]
malloc/tst-mallocfork2: Use process-shared barriers

This synchronization method has a lower overhead and makes
it more likely that the signal arrives during one of the critical
functions.

Also test for fork deadlocks explicitly.

Reviewed-by: Szabolcs Nagy <szabolcs.nagy@arm.com>
5 years agoUpdate syscall-names.list for Linux 5.1.
Joseph Myers [Tue, 7 May 2019 23:57:26 +0000 (23:57 +0000)]
Update syscall-names.list for Linux 5.1.

This patch updates syscall-names.list for Linux 5.1 (which has many
new syscalls, mainly but not entirely ones for 64-bit time).

Tested with build-many-glibcs.py (before the revert of the move to
Linux 5.1 there; verified there were no tst-syscall-list failures).

* sysdeps/unix/sysv/linux/syscall-names.list: Update kernel
version to 5.1.
(clock_adjtime64) New syscall.
(clock_getres_time64) Likewise.
(clock_gettime64) Likewise.
(clock_nanosleep_time64) Likewise.
(clock_settime64) Likewise.
(futex_time64) Likewise.
(io_pgetevents_time64) Likewise.
(io_uring_enter) Likewise.
(io_uring_register) Likewise.
(io_uring_setup) Likewise.
(mq_timedreceive_time64) Likewise.
(mq_timedsend_time64) Likewise.
(pidfd_send_signal) Likewise.
(ppoll_time64) Likewise.
(pselect6_time64) Likewise.
(recvmmsg_time64) Likewise.
(rt_sigtimedwait_time64) Likewise.
(sched_rr_get_interval_time64) Likewise.
(semtimedop_time64) Likewise.
(timer_gettime64) Likewise.
(timer_settime64) Likewise.
(timerfd_gettime64) Likewise.
(timerfd_settime64) Likewise.
(utimensat_time64) Likewise.

5 years agoRevert "Use Linux 5.1 in build-many-glibcs.py."
Joseph Myers [Tue, 7 May 2019 22:45:51 +0000 (22:45 +0000)]
Revert "Use Linux 5.1 in build-many-glibcs.py."

This reverts commit c2b11710fb4a2e8d337ae8f042724143c5ccf173.

Linux 5.1 headers are not in fact usable for glibc testing, because
"[PATCH] uapi: avoid namespace conflict in linux/posix_types.h"
<https://lore.kernel.org/lkml/20190319165123.3967889-1-arnd@arndb.de/>
did not get merged for 5.1 and so many conform/ tests fail.

5 years agoUse Linux 5.1 in build-many-glibcs.py.
Joseph Myers [Tue, 7 May 2019 14:49:05 +0000 (14:49 +0000)]
Use Linux 5.1 in build-many-glibcs.py.

* scripts/build-many-glibcs.py (Context.checkout): Default Linux
version to 5.1.

5 years agoUse GCC 9 in build-many-glibcs.py.
Joseph Myers [Tue, 7 May 2019 14:46:11 +0000 (14:46 +0000)]
Use GCC 9 in build-many-glibcs.py.

* scripts/build-many-glibcs.py (Context.checkout): Default GCC
version to 9 branch.

5 years agoaarch64: thunderx2 memmove performance improvements
Anton Youdkevitch [Fri, 3 May 2019 18:01:34 +0000 (11:01 -0700)]
aarch64: thunderx2 memmove performance improvements

The performance improvement is about 20%-30% for
larger cases and about 1%-5% for smaller cases.

Used SIMD load/store instead of GPR for large
overlapping forward moves.

Reused existing memcpy implementation for smaller
or overlapping backward moves.

Fixed the existing memcpy implementation to allow it
to deal with the overlapping case.

Simplified loop tails in the memcpy implementation -
use branchless overlapping sequence of fixed length
load/stores instead of branching depending on the
size.

A cleanup/optimization converting str's to stp's.

Added __memmove_thunderx2 to the list of the
available implementations.

5 years agomisc/tst-tsearch: Additional explicit error checking
Florian Weimer [Fri, 3 May 2019 07:22:33 +0000 (09:22 +0200)]
misc/tst-tsearch: Additional explicit error checking

This avoids an undefined variable warning with certain GCC versions.

5 years agoAdd missing bug number on CL entry for BZ#24506 (b2af6fb2ed239)
Adhemerval Zanella [Thu, 2 May 2019 11:46:20 +0000 (08:46 -0300)]
Add missing bug number on CL entry for BZ#24506 (b2af6fb2ed239)

5 years agoelf: Fix elf/tst-pldd with --enable-hardcoded-path-in-tests (BZ#24506)
Adhemerval Zanella [Tue, 30 Apr 2019 19:11:57 +0000 (16:11 -0300)]
elf: Fix elf/tst-pldd with --enable-hardcoded-path-in-tests (BZ#24506)

The elf/tst-pldd (added by 1a4c27355e146 to fix BZ#18035) test does
not expect the hardcoded paths that are output by pldd when the test
is built with --enable-hardcoded-path-in-tests.  Instead of showing
the ABI installed library names for loader and libc (such as
ld-linux-x86-64.so.2 and libc.so.6 for x86_64), pldd shows the default
built ld.so and libc.so.

It makes the tests fail with an invalid expected loader/libc name.

This patch fixes the elf-pldd test by adding the canonical ld.so and
libc.so names in the expected list of possible outputs when parsing
the result output from pldd.  The test now handles both default
build and --enable-hardcoded-path-in-tests option.

Checked on x86_64-linux-gnu (built with and without
--enable-hardcoded-path-in-tests) and i686-linux-gnu.

* elf/tst-pldd.c (in_str_list): New function.
(do_test): Add default names for ld and libc as one option.

Reviewed-by: Carlos O'Donell <carlos@redhat.com>
5 years agomisc: Add twalk_r function
Florian Weimer [Thu, 2 May 2019 09:42:51 +0000 (11:42 +0200)]
misc: Add twalk_r function

The twalk function is very difficult to use in a multi-threaded
program because there is no way to pass external state to the
iterator function.

Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
5 years agoMake mktime etc. compatible with __time64_t
Paul Eggert [Mon, 18 Mar 2019 21:14:15 +0000 (14:14 -0700)]
Make mktime etc. compatible with __time64_t

Keep these functions compatible with Gnulib while adding
__time64_t support.  The basic idea is to move private API
declarations from include/time.h to time/mktime-internal.h, since
the former file cannot easily be shared with Gnulib whereas the
latter can.
Also, do some other minor cleanup while in the neighborhood.
* include/time.h: Include stdbool.h, time/mktime-internal.h.
(__mktime_internal): Move this prototype to time/mktime-internal.h,
since Gnulib needs it.
(__localtime64_r, __gmtime64_r) [__TIMESIZE == 64]:
Move these macros to time/mktime-internal.h, since Gnulib needs them.
(__mktime64, __timegm64) [__TIMESIZE != 64]: New prototypes.
(in_time_t_range): New static function.
* posix/bits/types.h (__time64_t) [__TIMESIZE == 64 && !defined __LIBC]:
Do not define as a macro in this case, so that portable code is
less tempted to use __time64_t.
* time/mktime-internal.h: Rewrite so that it does both glibc
and Gnulib work.  Include time.h if not _LIBC.
(mktime_offset_t) [!_LIBC]: Define for gnulib.
(__time64_t, __gmtime64_r, __localtime64_r, __mktime64, __timegm64)
[!_LIBC || __TIMESIZE == 64]: New macros, mostly moved here
from include/time.h.
(__gmtime_r, __localtime_r, __mktime_internal) [!_LIBC]:
New macros, taken from GNulib.
(__mktime_internal): New prototype, moved here from include/time.h.
* time/mktime.c (mktime_min, mktime_max, convert_time)
(ranged_convert, __mktime_internal, __mktime64):
* time/timegm.c (__timegm64):
Use __time64_t, not time_t.
* time/mktime.c: Stop worrying about whether time_t is floating-point.
(__mktime64) [! (_LIBC && __TIMESIZE != 64)]:
Rename from mktime.
(mktime) [_LIBC && __TIMESIZE != 64]: New function.
* time/timegm.c [!_LIBC]: Include libc-config.h, not config.h,
for libc_hidden_def.
Include errno.h.
(__timegm64) [! (_LIBC && __TIMESIZE != 64)]:
Rename from timegm.
(timegm) [_LIBC && __TIMESIZE != 64]: New function.

First cut at publicizing __time64_t

5 years agoFix -O1 compilation errors with `__ddivl' and `__fdivl' [BZ #19444]
Maciej W. Rozycki [Tue, 30 Apr 2019 01:24:49 +0000 (02:24 +0100)]
Fix -O1 compilation errors with `__ddivl' and `__fdivl' [BZ #19444]

Complementing commit 4a06ceea33ec ("sysdeps/ieee754/soft-fp: ignore
maybe-uninitialized with -O [BZ #19444]") and commit 27c5e756a2a8
("sysdeps/ieee754: prevent maybe-uninitialized errors with -O [BZ
#19444]") also fix compilation errors observed at -O1 in `__ddivl' and
`__fdivl' with GCC 9 and RISC-V targets:

In file included from ../soft-fp/soft-fp.h:318,
                 from ../sysdeps/ieee754/soft-fp/s_fdivl.c:27:
../sysdeps/ieee754/soft-fp/s_fdivl.c: In function '__fdivl':
../soft-fp/op-2.h:108:9: error: 'R_f1' may be used uninitialized in this function [-Werror=maybe-uninitialized]
  108 |         : (X##_f1 << (2*_FP_W_TYPE_SIZE - (N)))) \
      |         ^
../sysdeps/ieee754/soft-fp/s_fdivl.c:37:14: note: 'R_f1' was declared here
   37 |   FP_DECL_Q (R);
      |              ^
../soft-fp/op-common.h:39:3: note: in expansion of macro '_FP_FRAC_DECL_2'
   39 |   _FP_FRAC_DECL_##wc (X)
      |   ^~~~~~~~~~~~~~
../soft-fp/quad.h:226:24: note: in expansion of macro '_FP_DECL'
  226 | # define FP_DECL_Q(X)  _FP_DECL (2, X)
      |                        ^~~~~~~~
../sysdeps/ieee754/soft-fp/s_fdivl.c:37:3: note: in expansion of macro 'FP_DECL_Q'
   37 |   FP_DECL_Q (R);
      |   ^~~~~~~~~
../soft-fp/op-2.h:109:8: error: 'R_f0' may be used uninitialized in this function [-Werror=maybe-uninitialized]
  109 |        | X##_f0) != 0));    \
      |        ^
../sysdeps/ieee754/soft-fp/s_fdivl.c:37:14: note: 'R_f0' was declared here
   37 |   FP_DECL_Q (R);
      |              ^
../soft-fp/op-common.h:39:3: note: in expansion of macro '_FP_FRAC_DECL_2'
   39 |   _FP_FRAC_DECL_##wc (X)
      |   ^~~~~~~~~~~~~~
../soft-fp/quad.h:226:24: note: in expansion of macro '_FP_DECL'
  226 | # define FP_DECL_Q(X)  _FP_DECL (2, X)
      |                        ^~~~~~~~
../sysdeps/ieee754/soft-fp/s_fdivl.c:37:3: note: in expansion of macro 'FP_DECL_Q'
   37 |   FP_DECL_Q (R);
      |   ^~~~~~~~~
In file included from ../soft-fp/soft-fp.h:318,
                 from ../sysdeps/ieee754/soft-fp/s_ddivl.c:31:
../sysdeps/ieee754/soft-fp/s_ddivl.c: In function '__ddivl':
../soft-fp/op-2.h:98:25: error: 'R_f1' may be used uninitialized in this function [-Werror=maybe-uninitialized]
   98 |        X##_f0 = (X##_f1 << (_FP_W_TYPE_SIZE - (N)) | X##_f0 >> (N) \
      |                         ^~
../sysdeps/ieee754/soft-fp/s_ddivl.c:41:14: note: 'R_f1' was declared here
   41 |   FP_DECL_Q (R);
      |              ^
../soft-fp/op-2.h:37:36: note: in definition of macro '_FP_FRAC_DECL_2'
   37 |   _FP_W_TYPE X##_f0 _FP_ZERO_INIT, X##_f1 _FP_ZERO_INIT
      |                                    ^
../soft-fp/quad.h:226:24: note: in expansion of macro '_FP_DECL'
  226 | # define FP_DECL_Q(X)  _FP_DECL (2, X)
      |                        ^~~~~~~~
../sysdeps/ieee754/soft-fp/s_ddivl.c:41:3: note: in expansion of macro 'FP_DECL_Q'
   41 |   FP_DECL_Q (R);
      |   ^~~~~~~~~
../soft-fp/op-2.h:101:17: error: 'R_f0' may be used uninitialized in this function [-Werror=maybe-uninitialized]
  101 |       : (X##_f0 << (_FP_W_TYPE_SIZE - (N))) != 0)); \
      |                 ^~
../sysdeps/ieee754/soft-fp/s_ddivl.c:41:14: note: 'R_f0' was declared here
   41 |   FP_DECL_Q (R);
      |              ^
../soft-fp/op-2.h:37:14: note: in definition of macro '_FP_FRAC_DECL_2'
   37 |   _FP_W_TYPE X##_f0 _FP_ZERO_INIT, X##_f1 _FP_ZERO_INIT
      |              ^
../soft-fp/quad.h:226:24: note: in expansion of macro '_FP_DECL'
  226 | # define FP_DECL_Q(X)  _FP_DECL (2, X)
      |                        ^~~~~~~~
../sysdeps/ieee754/soft-fp/s_ddivl.c:41:3: note: in expansion of macro 'FP_DECL_Q'
   41 |   FP_DECL_Q (R);
      |   ^~~~~~~~~
cc1: all warnings being treated as errors
make[2]: *** [.../sysd-rules:587: .../math/s_fdivl.o] Error 1
make[2]: *** Waiting for unfinished jobs....
cc1: all warnings being treated as errors
make[2]: *** [.../sysd-rules:587: .../math/s_ddivl.o] Error 1

This comes from cases in _FP_DIV that return a result described as
FP_CLS_ZERO or FP_CLS_INF and do not initialize the fractional part,
which is then operated on unconditionally in FP_TRUNC_COOKED before
being ignored by _FP_PACK_CANONICAL.

Clearly at this optimization level GCC cannot guarantee to be able to
determine that the fractional part is ultimately unused, so ignore the
error as with the earlier commits referred, letting compilation proceed.

[BZ #19444]
* sysdeps/ieee754/soft-fp/s_ddivl.c (__ddivl): Ignore errors
from `-Wmaybe-uninitialized'.
* sysdeps/ieee754/soft-fp/s_fdivl.c (__fdivl): Likewise.

5 years agopowerpc: ceil/ceilf refactor
Adhemerval Zanella [Fri, 15 Mar 2019 18:42:00 +0000 (18:42 +0000)]
powerpc: ceil/ceilf refactor

This patches consolidates all the powerpc ceil{f} implementations on
the generic sysdeps/powerpc/fpu/s_ceil{f}.  The generic implementation
uses either the compiler builts for ISA 2.03+ (which generates the frip
instruction) or a generic implementation which uses FP only operations.

It adds a generic implementation (round_to_integer.h) which is shared
with other rounding to integer routines.  The resulting code should be
similar in term os performance to previous assembly one.

The IFUNC organization for powerpc64 is also change to be enabled only
for powerpc64 and not for powerpc64le (since minium ISA of 2.08 does not
require the fallback generic implementation).

Checked on powerpc-linux-gnu (built without --with-cpu, with
--with-cpu=power4 and with --with-cpu=power5+ and --disable-multi-arch),
powerpc64-linux-gnu (built without --with-cp and with --with-cpu=power5+
and --disable-multi-arch).

* sysdeps/powerpc/fpu/fenv_libc.h (__fesetround_inline_nocheck): New
function.
* sysdeps/powerpc/fpu/round_to_integer.h: New file.
* sysdeps/powerpc/fpu/s_ceil.c: Likewise.
* sysdeps/powerpc/fpu/s_ceilf.c: Likewise.
* sysdeps/powerpc/powerpc32/fpu/s_ceil.S: Remove file.
* sysdeps/powerpc/powerpc32/fpu/s_ceilf.S: Likewise.
* sysdeps/powerpc/powerpc32/power4/fpu/multiarch/Makefile
(CFLAGS-s_ceil-power5+.c, CFLAGS-s_ceilf-power5+.c): New rule.
* sysdeps/powerpc/powerpc32/power4/fpu/multiarch/s_ceil-power5+.S:
Remove file.
* sysdeps/powerpc/powerpc32/power4/fpu/multiarch/s_ceil-ppc32.S:
Likewise.
* sysdeps/powerpc/powerpc32/power4/fpu/multiarch/s_ceilf-power5+.S:
Likewise.
* sysdeps/powerpc/powerpc32/power4/fpu/multiarch/s_ceilf-ppc32.S:
Likewise.
* sysdeps/powerpc/powerpc32/power4/fpu/multiarch/s_ceil-power5+.c:
New file.
* sysdeps/powerpc/powerpc32/power4/fpu/multiarch/s_ceil-ppc32.c:
Likewise.
* sysdeps/powerpc/powerpc32/power4/fpu/multiarch/s_ceilf-power5+.c:
Likewise.
* sysdeps/powerpc/powerpc32/power4/fpu/multiarch/s_ceilf-ppc32.c:
Likewise.
* sysdeps/powerpc/powerpc32/power5+/fpu/s_ceil.S: Remove file.
* sysdeps/powerpc/powerpc32/power5+/fpu/s_ceilf.S: Likewise.
* sysdeps/powerpc/powerpc64/be/fpu/multiarch/Makefile: New file.
* sysdeps/powerpc/powerpc64/be/fpu/multiarch/s_ceil-power5+.c:
Likewise.
* sysdeps/powerpc/powerpc64/be/fpu/multiarch/s_ceil-ppc64.c:
Likewise.
* sysdeps/powerpc/powerpc64/fpu/multiarch/s_ceil.c: Move to ...
* sysdeps/powerpc/powerpc64/be/fpu/multiarch/s_ceil.c: ... here.
* sysdeps/powerpc/powerpc64/be/fpu/multiarch/s_ceilf-power5+.c: New
file.
* sysdeps/powerpc/powerpc64/be/fpu/multiarch/s_ceilf-ppc64.c:
Likewise.
* sysdeps/powerpc/powerpc64/fpu/multiarch/s_ceilf.c: Move to ...
* sysdeps/powerpc/powerpc64/be/fpu/multiarch/s_ceilf.c: ...
* here.
* sysdeps/powerpc/powerpc64/fpu/multiarch/Makefile
(libm-sysdep_routines): Remove s_ceil-power5+, s_ceil-ppc64,
s_ceilf-power5+, and s_ceilf-ppc64.
* sysdeps/powerpc/powerpc64/fpu/multiarch/s_ceil-power5+.S: Remove
file.
* sysdeps/powerpc/powerpc64/fpu/multiarch/s_ceil-ppc64.S: Likewise.
* sysdeps/powerpc/powerpc64/fpu/multiarch/s_ceilf-power5+.S: Likewise.
* sysdeps/powerpc/powerpc64/fpu/multiarch/s_ceilf-ppc64.S: Likewise.
* sysdeps/powerpc/powerpc64/fpu/s_ceil.S: Likewise.
* sysdeps/powerpc/powerpc64/fpu/s_ceilf.S: Likewise.
* sysdeps/powerpc/powerpc64/power5+/fpu/s_ceil.S: Likewise.
* sysdeps/powerpc/powerpc64/power5+/fpu/s_ceilf.S: Likewise.

Reviewed-by: Gabriel F. T. Gomes <gabriel@inconstante.eti.br>
5 years agopowerpc: Remove power4 mpa optimization
Adhemerval Zanella [Wed, 20 Mar 2019 15:34:10 +0000 (15:34 +0000)]
powerpc: Remove power4 mpa optimization

This patch removes the POWER4 optimized mpa optimization used currently
on all powerpc targets.  In fact for newer chips, GCC generates *worse*
code than generic implementation as below.  One possibilty would to
add ifunc variants for the mpa routines (as x86_64), but it will add
complexity only for older chips (and one would need to check if
power5, power5+, and power6 do benefict from this optimization),
and only for specific implementation (since most used one such
as sin, cos, exp, pow where optimized to avoid calling the slow
multiprecision path).

* POWER9 patched
$ ./testrun.sh benchtests/bench-atan
  "atan": {
   "": {
    "duration": 5.12565e+09,
    "iterations": 1.552e+08,
    "max": 100.552,
    "min": 7.799,
    "mean": 33.0261
   },
   "144bits": {
    "duration": 5.12745e+09,
    "iterations": 825000,
    "max": 7517.17,
    "min": 6186.3,
    "mean": 6215.09
   }
  }
$ ./testrun.sh benchtests/bench-acos
  "acos": {
   "": {
    "duration": 5.21741e+09,
    "iterations": 1.269e+08,
    "max": 191.738,
    "min": 7.931,
    "mean": 41.1144
   },
   "slow": {
    "duration": 5.25999e+09,
    "iterations": 198000,
    "max": 26681.7,
    "min": 26463.6,
    "mean": 26565.6
   }
  }

* POWER9 master
$ ./testrun.sh benchtests/bench-atan
  "atan": {
   "": {
    "duration": 5.12815e+09,
    "iterations": 1.552e+08,
    "max": 134.788,
    "min": 7.803,
    "mean": 33.0422
   },
   "144bits": {
    "duration": 5.1209e+09,
    "iterations": 447000,
    "max": 11615.8,
    "min": 11301.8,
    "mean": 11456.2
   }
  }
$ ./testrun.sh benchtests/bench-acos
  "acos": {
   "": {
    "duration": 5.22272e+09,
    "iterations": 1.269e+08,
    "max": 115.981,
    "min": 7.931,
    "mean": 41.1562
   },
   "slow": {
    "duration": 5.28723e+09,
    "iterations": 96000,
    "max": 55434.1,
    "min": 54820.6,
    "mean": 55075.3
   }
  }

* POWER8 patched
$ taskset -c 16 ./testrun.sh benchtests/bench-acos
  "acos": {
   "": {
    "duration": 5.16398e+09,
    "iterations": 9.99e+07,
    "max": 174.408,
    "min": 8.645,
    "mean": 51.6915
   },
   "slow": {
    "duration": 5.16982e+09,
    "iterations": 96000,
    "max": 54830.5,
    "min": 53703.8,
    "mean": 53852.3
   }
  }
* POWER8 master
$ taskset -c 16 ./testrun.sh benchtests/bench-acos
  "acos": {
   "": {
    "duration": 5.17019e+09,
    "iterations": 9.99e+07,
    "max": 186.127,
    "min": 8.633,
    "mean": 51.7537
   },
   "slow": {
    "duration": 5.34225e+09,
    "iterations": 90000,
    "max": 60353.2,
    "min": 59155.3,
    "mean": 59358.4
   }
  }

* POWER7 patched
$ taskset -c 16 benchtests/bench-asin
  "asin": {
   "": {
    "duration": 5.15559e+09,
    "iterations": 6.5e+07,
    "max": 193.335,
    "min": 12.227,
    "mean": 79.3168
   },
   "slow": {
    "duration": 5.20538e+09,
    "iterations": 80000,
    "max": 65705.2,
    "min": 64299.4,
    "mean": 65067.3
   }
  }
* POWER7 master
$ taskset -c 16 benchtests/bench-asin
  "asin": {
   "": {
    "duration": 5.15446e+09,
    "iterations": 6.5e+07,
    "max": 184.575,
    "min": 12.226,
    "mean": 79.2994
   },
   "slow": {
    "duration": 5.20616e+09,
    "iterations": 80000,
    "max": 65705.1,
    "min": 64336.6,
    "mean": 65076.9
   }
  }

Checked on powerpc-linux-gnu (built without --with-cpu, with
--with-cpu=power4 and with --with-cpu=power5+ and --disable-multi-arch),
powerpc64-linux-gnu (built without --with-cp and with --with-cpu=power5+
and --disable-multi-arch).

* sysdeps/powerpc/power4/fpu/Makefile: Remove file.
* sysdeps/powerpc/power4/fpu/mpa-arch.h: Likewise.
* sysdeps/powerpc/power4/fpu/mpa.c: Likewise.

Reviewed-by: Gabriel F. T. Gomes <gabriel@inconstante.eti.br>
5 years ago<semaphore.h>: Add nonnull attributes
Yann Droneaud [Mon, 29 Apr 2019 08:11:35 +0000 (10:11 +0200)]
<semaphore.h>: Add nonnull attributes

Except the following functions, NPTL implementation assume sem_t
argument (or other arguments) are not NULL, so they would benefit
from having the nonnull attribute.

- sem_close(): can cope with a NULL sem_t and return -1 with error EINVAL;
- sem_destroy(): does nothing at all

* sysdeps/pthread/semaphore.h (sem_init): Add __nonnull attribute.
(sem_destroy, sem_open, sem_close, sem_unlink): Likewise.
(sem_wait, sem_timedwait, sem_trywait, sem_post): Likewise.
(sem_getvalue): Likewise.

5 years agoelf: Link sotruss-lib.so with BIND_NOW for --enable-bind-now
Florian Weimer [Fri, 26 Apr 2019 05:16:56 +0000 (07:16 +0200)]
elf: Link sotruss-lib.so with BIND_NOW for --enable-bind-now

The audit module itself can be linked with BIND_NOW; it does not
affect its functionality.

This should complete the leftovers from commit
2d6ab5df3b675e96ee587ae6a8c2ce004c6b1ba9 ("Document and fix
--enable-bind-now [BZ #21015]").

5 years agoMakeconfig: Move -Wl,-rpath-link options before library references
Florian Weimer [Fri, 26 Apr 2019 05:16:30 +0000 (07:16 +0200)]
Makeconfig: Move -Wl,-rpath-link options before library references

Previously, the -Wl,-rpath-link options came after the libraries
injected using LDLIBS-* variables on the link editor command line for
main programs.  As a result, it could happen that installed libraries
that reference glibc libraries used the installed glibc from the system
directories, instead of the glibc from the build tree.  This can lead to
link failures if the wrong version of libpthread.so.0 is used, for
instance, due to differences in the internal GLIBC_PRIVATE interfaces,
as seen with memusagestat and -lgd after commit
f9b645b4b0a10c43753296ce3fa40053fa44606a ("memusagestat: use local glibc
when linking [BZ #18465]").

The isolation is necessarily imperfect because these installed
libraries are linked against the installed glibc in the system
directories.  However, in most cases, the built glibc will be newer
than the installed glibc, and this link is permitted because of the
ABI backwards compatibility glibc provides.

5 years agoMakeconfig: Move $(CC) to +link command variables
Florian Weimer [Fri, 26 Apr 2019 05:16:47 +0000 (07:16 +0200)]
Makeconfig: Move $(CC) to +link command variables

This change is needed to add linker flags which come very early in the
command linke (before LDFLAGS) and are not applied to test programs
(only to installed programs).

5 years agoriscv: remove DL_RO_DYN_SECTION
David Abdurachmanov [Tue, 9 Apr 2019 11:25:29 +0000 (13:25 +0200)]
riscv: remove DL_RO_DYN_SECTION

While working on enabling D front-end (GDC) in GCC we noticed that
druntime was segfaulting if it is linked dynamically. This was tracked
to DL_RO_DYN_SECTION.

DL_RO_DYN_SECTION lines seem to be copied from MIPS file (which is the
only user of it), but the comment doesn't apply to RISC-V. There is no
such requirement in RISC-V ABI.

        [BZ#24484]
* sysdeps/riscv/ldsodefs.h: Remove DL_RO_DYN_SECTION as it is not
required by RISC-V ABI.

5 years agobenchtests: Enable BIND_NOW if configured with --enable-bind-now
Florian Weimer [Thu, 25 Apr 2019 08:41:52 +0000 (10:41 +0200)]
benchtests: Enable BIND_NOW if configured with --enable-bind-now

Benchmarks should reflect distribution build policies, so it makes
sense to honor the BIND_NOW configuration for them.

This commit keeps using $(+link-tests), so that the benchmarks are
linked according to the --enable-hardcoded-path-in-tests configure
option.

Reviewed-by: Carlos O'Donell <carlos@redhat.com>
5 years agoExtend BIND_NOW to installed programs with --enable-bind-now
Florian Weimer [Thu, 25 Apr 2019 08:41:43 +0000 (10:41 +0200)]
Extend BIND_NOW to installed programs with --enable-bind-now

Commit 2d6ab5df3b675e96ee587ae6a8c2ce004c6b1ba9 ("Document and fix
--enable-bind-now [BZ #21015]") extended BIND_NOW to all installed
shared objects.  This change also covers installed programs.

Reviewed-by: Carlos O'Donell <carlos@redhat.com>
5 years agoReduce benchtests time
Wilco Dijkstra [Wed, 24 Apr 2019 14:38:49 +0000 (15:38 +0100)]
Reduce benchtests time

Reduce the total time taken by benchtests.  The malloc thread test takes 4
minutes to run which is significantly more than most other tests. Reduce
this to a more reasonable 40 seconds.  The math tests take 10 seconds each,
eventhough all they do is loop on the same input.  Anything more than 1
second runtime is way overkill, so set the limit to 1 second.

* benchtests/Makefile (BENCH_DURATION): Set to 1 second.
* benchtests/bench-malloc-thread.c (BENCH_DURATION): Set to 10 seconds.

5 years agomemusagestat: use local glibc when linking [BZ #18465]
Mike Frysinger [Wed, 24 Apr 2019 11:32:22 +0000 (13:32 +0200)]
memusagestat: use local glibc when linking [BZ #18465]

The memusagestat is the only binary that has its own link line which
causes it to be linked against the existing installed C library.  It
has been this way since it was originally committed in 1999, but I
don't see any reason as to why.  Since we want all the programs we
build locally to be against the new copy of glibc, change the build
to be like all other programs.

5 years agolocale/tst-locale-locpath: Run test only for $(run-built-tests) == yes
Florian Weimer [Wed, 24 Apr 2019 05:31:29 +0000 (07:31 +0200)]
locale/tst-locale-locpath: Run test only for $(run-built-tests) == yes

5 years agoelf: Fix pldd (BZ#18035)
Adhemerval Zanella [Thu, 11 Apr 2019 21:12:00 +0000 (18:12 -0300)]
elf: Fix pldd (BZ#18035)

Since 9182aa67994 (Fix vDSO l_name for GDB's, BZ#387) the initial link_map
for executable itself and loader will have both l_name and l_libname->name
holding the same value due:

 elf/dl-object.c

 95   new->l_name = *realname ? realname : (char *) newname->name + libname_len - 1;

Since newname->name points to new->l_libname->name.

This leads to pldd to an infinite call at:

 elf/pldd-xx.c

203     again:
204       while (1)
205         {
206           ssize_t n = pread64 (memfd, tmpbuf.data, tmpbuf.length, name_offset);

228           /* Try the l_libname element.  */
229           struct E(libname_list) ln;
230           if (pread64 (memfd, &ln, sizeof (ln), m.l_libname) == sizeof (ln))
231             {
232               name_offset = ln.name;
233               goto again;
234             }

Since the value at ln.name (l_libname->name) will be the same as previously
read. The straightforward fix is just avoid the check and read the new list
entry.

I checked also against binaries issues with old loaders with fix for BZ#387,
and pldd could dump the shared objects.

Checked on x86_64-linux-gnu, i686-linux-gnu, aarch64-linux-gnu, and
powerpc64le-linux-gnu.

[BZ #18035]
* elf/Makefile (tests-container): Add tst-pldd.
* elf/pldd-xx.c: Use _Static_assert in of pldd_assert.
(E(find_maps)): Avoid use alloca, use default read file operations
instead of explicit LFS names, and fix infinite loop.
* elf/pldd.c: Explicit set _FILE_OFFSET_BITS, cleanup headers.
(get_process_info): Use _Static_assert instead of assert, use default
directory operations instead of explicit LFS names, and free some
leadek pointers.
* elf/tst-pldd.c: New file.

5 years agoRemove do_set_mallopt_check prototype
H.J. Lu [Tue, 23 Apr 2019 20:12:13 +0000 (13:12 -0700)]
Remove do_set_mallopt_check prototype

Remove do_set_mallopt_check prototype since it is unused.

* malloc/arena.c (do_set_mallopt_check): Removed.

Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Reviewed-by: DJ Delorie <dj@redhat.com>
5 years agolocale: Add LOCPATH diagnostics to the locale program
Florian Weimer [Tue, 23 Apr 2019 16:16:26 +0000 (18:16 +0200)]
locale: Add LOCPATH diagnostics to the locale program

The implementation of quote_string is based on support_quote_blob.

Reviewed-by: Carlos O'Donell <carlos@redhat.com>
5 years agoFix NEWS entry from 9bf8e29ca136
Adhemerval Zanella [Sun, 21 Apr 2019 11:39:25 +0000 (08:39 -0300)]
Fix NEWS entry from 9bf8e29ca136

* NEWS: Move memory allocation changes of BZ#23741 from 2.29
to 2.30 notes.

5 years agomalloc: make malloc fail with requests larger than PTRDIFF_MAX (BZ#23741)
Adhemerval Zanella [Tue, 18 Dec 2018 18:30:56 +0000 (16:30 -0200)]
malloc: make malloc fail with requests larger than PTRDIFF_MAX (BZ#23741)

As discussed previously on libc-alpha [1], this patch follows up the idea
and add both the __attribute_alloc_size__ on malloc functions (malloc,
calloc, realloc, reallocarray, valloc, pvalloc, and memalign) and limit
maximum requested allocation size to up PTRDIFF_MAX (taking into
consideration internal padding and alignment).

This aligns glibc with gcc expected size defined by default warning
-Walloc-size-larger-than value which warns for allocation larger than
PTRDIFF_MAX.  It also aligns with gcc expectation regarding libc and
expected size, such as described in PR#67999 [2] and previously discussed
ISO C11 issues [3] on libc-alpha.

From the RFC thread [4] and previous discussion, it seems that consensus
is only to limit such requested size for malloc functions, not the system
allocation one (mmap, sbrk, etc.).

The implementation changes checked_request2size to check for both overflow
and maximum object size up to PTRDIFF_MAX. No additional checks are done
on sysmalloc, so it can still issue mmap with values larger than
PTRDIFF_T depending on the requested size.

The __attribute_alloc_size__ is for functions that return a pointer only,
which means it cannot be applied to posix_memalign (see remarks in GCC
PR#87683 [5]). The runtimes checks to limit maximum requested allocation
size does applies to posix_memalign.

Checked on x86_64-linux-gnu and i686-linux-gnu.

[1] https://sourceware.org/ml/libc-alpha/2018-11/msg00223.html
[2] https://gcc.gnu.org/bugzilla//show_bug.cgi?id=67999
[3] https://sourceware.org/ml/libc-alpha/2011-12/msg00066.html
[4] https://sourceware.org/ml/libc-alpha/2018-11/msg00224.html
[5] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87683

[BZ #23741]
* malloc/hooks.c (malloc_check, realloc_check): Use
__builtin_add_overflow on overflow check and adapt to
checked_request2size change.
* malloc/malloc.c (__libc_malloc, __libc_realloc, _mid_memalign,
__libc_pvalloc, __libc_calloc, _int_memalign): Limit maximum
allocation size to PTRDIFF_MAX.
(REQUEST_OUT_OF_RANGE): Remove macro.
(checked_request2size): Change to inline function and limit maximum
requested size to PTRDIFF_MAX.
(__libc_malloc, __libc_realloc, _int_malloc, _int_memalign): Limit
maximum allocation size to PTRDIFF_MAX.
(_mid_memalign): Use _int_memalign call for overflow check.
(__libc_pvalloc): Use __builtin_add_overflow on overflow check.
(__libc_calloc): Use __builtin_mul_overflow for overflow check and
limit maximum requested size to PTRDIFF_MAX.
* malloc/malloc.h (malloc, calloc, realloc, reallocarray, memalign,
valloc, pvalloc): Add __attribute_alloc_size__.
* stdlib/stdlib.h (malloc, realloc, reallocarray, valloc): Likewise.
* malloc/tst-malloc-too-large.c (do_test): Add check for allocation
larger than PTRDIFF_MAX.
* malloc/tst-memalign.c (do_test): Disable -Walloc-size-larger-than=
around tests of malloc with negative sizes.
* malloc/tst-posix_memalign.c (do_test): Likewise.
* malloc/tst-pvalloc.c (do_test): Likewise.
* malloc/tst-valloc.c (do_test): Likewise.
* malloc/tst-reallocarray.c (do_test): Replace call to reallocarray
with resulting size allocation larger than PTRDIFF_MAX with
reallocarray_nowarn.
(reallocarray_nowarn): New function.
* NEWS: Mention the malloc function semantic change.

5 years agopowerpc: Fix format issue from 3a16dd780eeba602
Adhemerval Zanella [Wed, 17 Apr 2019 21:32:01 +0000 (18:32 -0300)]
powerpc: Fix format issue from 3a16dd780eeba602

* sysdeps/powerpc/fpu/s_fma.c: Fix format.
* sysdeps/powerpc/fpu/s_fmaf.c: Likewise.

5 years agopowerpc: fma using builtins
Adhemerval Zanella [Wed, 20 Mar 2019 04:14:28 +0000 (04:14 +0000)]
powerpc: fma using builtins

This patch just refactor the assembly implementation to use compiler
builtins instead.

Checked on powerpc-linux-gnu (built without --with-cpu, with
--with-cpu=power4 and with --with-cpu=power5+ and --disable-multi-arch),
powerpc64-linux-gnu (built without --with-cp and with --with-cpu=power5+
and --disable-multi-arch).

* sysdeps/powerpc/fpu/s_fma.S: Remove file.
* sysdeps/powerpc/fpu/s_fmaf.S: Likewise.
* sysdeps/powerpc/fpu/s_fma.c: New file.
* sysdeps/powerpc/fpu/s_fmaf.c: Likewise.

5 years agopowerpc: Use generic fabs{f} implementations
Adhemerval Zanella [Wed, 20 Mar 2019 04:03:22 +0000 (04:03 +0000)]
powerpc: Use generic fabs{f} implementations

Since be2e25bbd78f9fdf the generic ieee754 implementation uses
compiler builtin which generates fabs{f} for all supported targets.

Checked on powerpc-linux-gnu (built without --with-cpu, with
--with-cpu=power4 and with --with-cpu=power5+ and --disable-multi-arch),
powerpc64-linux-gnu (built without --with-cp and with --with-cpu=power5+
and --disable-multi-arch).

* sysdeps/powerpc/fpu/s_fabs.S: Remove file.
* sysdeps/powerpc/fpu/s_fabsf.S: Likewise.

5 years agomips: Remove rt_sigreturn usage on context function
Adhemerval Zanella [Fri, 15 Feb 2019 14:04:45 +0000 (12:04 -0200)]
mips: Remove rt_sigreturn usage on context function

Similar to powerpc, mips also issues rt_sigreturn for setcontext
case the v0 value saved is not the one set by setcontext or
makecontext. As for powerpc, it is intention is no really supported
since setcontext is not async-signal-safe.

Checked the context tests on mips64-linux-gnu and mips-linux-gnu.

* sysdeps/unix/sysv/linux/mips/getcontext.S (__getcontext): Remove
the magic flag store.
* sysdeps/unix/sysv/linux/mips/makecontext.S (__makecontext):
Likewise.
* sysdeps/unix/sysv/linux/mips/swapcontext.S (__swapcontext):
Likewise.
* sysdeps/unix/sysv/linux/mips/setcontext.S (__setcontext):
Remove rt_sigreturn call.

5 years agopowerpc: Remove rt_sigreturn usage on context function
Adhemerval Zanella [Fri, 15 Feb 2019 13:55:16 +0000 (11:55 -0200)]
powerpc: Remove rt_sigreturn usage on context function

As described in a recent glibc thread [1], the rt_sigreturn syscall
on setcontext and swapcontext is not used on default use and its
intention is no really supported since neither setcontext nor
swapcontext are async-signal-safe.

Checked on powerpc64-linux-gnu and powerpc-linux-gnu.

* sysdeps/unix/sysv/linux/powerpc/powerpc32/setcontext-common.S:
Remove rt_sigreturn call.
* sysdeps/unix/sysv/linux/powerpc/powerpc32/swapcontext-common.S:
Likewise.
* sysdeps/unix/sysv/linux/powerpc/powerpc64/setcontext.S: Likewie.
* sysdeps/unix/sysv/linux/powerpc/powerpc64/swapcontext.S: Likewise.

[1] https://sourceware.org/ml/libc-alpha/2019-02/msg00367.html

5 years agosupport: Add support_capture_subprogram
Adhemerval Zanella [Fri, 12 Apr 2019 20:39:53 +0000 (17:39 -0300)]
support: Add support_capture_subprogram

Its API is similar to support_capture_subprocess, but rather creates a
new process based on the input path and arguments.  Under the hoods it
uses posix_spawn to create the new process.

It also allows the use of other support_capture_* functions to check
for expected results and free the resources.

Checked on x86_64-linux-gnu.

* support/Makefile (libsupport-routines): Add support_subprocess,
xposix_spawn, xposix_spawn_file_actions_addclose, and
xposix_spawn_file_actions_adddup2.
(tst-support_capture_subprocess-ARGS): New rule.
* support/capture_subprocess.h (support_capture_subprogram): New
prototype.
* support/support_capture_subprocess.c (support_capture_subprocess):
Refactor to use support_subprocess and support_capture_poll.
(support_capture_subprogram): New function.
* support/tst-support_capture_subprocess.c (write_mode_to_str,
str_to_write_mode, test_common, parse_int, handle_restart,
do_subprocess, do_subprogram, do_multiple_tests): New functions.
(do_test): Add support_capture_subprogram tests.
* support/subprocess.h: New file.
* support/support_subprocess.c: Likewise.
* support/xposix_spawn.c: Likewise.
* support/xposix_spawn_file_actions_addclose.c: Likewise.
* support/xposix_spawn_file_actions_adddup2.c: Likewise.
* support/xspawn.h: Likewise.

Reviewed-by: Carlos O'Donell <carlos@redhat.com>
5 years agostdlib/tst-secure-getenv: handle >64 groups
Mike Gerow [Wed, 17 Apr 2019 09:45:34 +0000 (11:45 +0200)]
stdlib/tst-secure-getenv: handle >64 groups

This test would fail unnecessarily if the user running it had more than
64 groups since getgroups returns EINVAL if the size provided is less
than the number of supplementary group IDs. Instead dynamically
determine the number of supplementary groups the user has.

5 years agoChangeLog: Correct excess and deficiency of spaces
TAMUKI Shoichi [Wed, 17 Apr 2019 03:44:26 +0000 (12:44 +0900)]
ChangeLog: Correct excess and deficiency of spaces

5 years agoChangeLog: Convert leading spaces to tabs
TAMUKI Shoichi [Tue, 16 Apr 2019 01:17:24 +0000 (10:17 +0900)]
ChangeLog: Convert leading spaces to tabs

5 years agonss_dns: Do not replace root domain with empty string
Florian Weimer [Thu, 11 Apr 2019 09:37:47 +0000 (11:37 +0200)]
nss_dns: Do not replace root domain with empty string

The purpose of the bp[0] == '.' check is unclear.  Only the root domain
starts with '.'.  The empty string is accepted as a domain name in many
places, denoting the root, but using it implicitly is confusing.

5 years agoalloc_buffer: Return unqualified pointer type in alloc_buffer_next
Florian Weimer [Thu, 11 Apr 2019 07:43:16 +0000 (09:43 +0200)]
alloc_buffer: Return unqualified pointer type in alloc_buffer_next

alloc_buffer_next is useful for peeking to the remaining part of the
buffer and update it, with subsequent allocation (once the length
is known) using alloc_buffer_alloc_bytes.  This is not as robust
as the other interfaces, but it allows using alloc_buffer with
string-writing interfaces such as snprintf and ns_name_ntop.

5 years agoAdd missing ChangeLog entry for commit 10dd69e9a00
Florian Weimer [Thu, 11 Apr 2019 07:41:08 +0000 (09:41 +0200)]
Add missing ChangeLog entry for commit 10dd69e9a00