H.J. Lu [Sun, 1 Oct 2017 22:50:56 +0000 (15:50 -0700)]
Mark internal dirent functions with attribute_hidden [BZ #18822]
Mark internal dirent functions with attribute_hidden to allow direct
access within libc.so and libc.a without using GOT nor PLT. __readdir64
is hidden with libc_hidden_proto and libc_hidden_def since the exported
readdir64 is an alias of __readdir64.
H.J. Lu [Sun, 1 Oct 2017 22:08:14 +0000 (15:08 -0700)]
Mark __internal_statvfs[64] with attribute_hidden [BZ #18822]
Mark __internal_statvfs[64] with attribute_hidden to allow direct access
to them within libc.so and libc.a without using GOT nor PLT.
[BZ #18822]
* sysdeps/unix/sysv/linux/fstatvfs.c: Include "internal_statvfs.h"
instead of <sys/statvfs.h>.
(__internal_statvfs): Removed.
* sysdeps/unix/sysv/linux/fstatvfs64.c Include "internal_statvfs.h"
instead of <sys/statvfs.h>.
(__internal_statvfs64): Removed.
* sysdeps/unix/sysv/linux/internal_statvfs.c: Include
"internal_statvfs.h" instead of <sys/statvfs.h>.
* sysdeps/unix/sysv/linux/internal_statvfs.h: New file.
* sysdeps/unix/sysv/linux/statvfs.c Include "internal_statvfs.h"
instead of <sys/statvfs.h>.
(__internal_statvfs): Removed.
* sysdeps/unix/sysv/linux/statvfs64.c Include "internal_statvfs.h"
instead of <sys/statvfs.h>.
(__internal_statvfs64): Removed.
H.J. Lu [Sun, 1 Oct 2017 21:34:18 +0000 (14:34 -0700)]
Mark internal stdlib functions with attribute_hidden [BZ #18822]
Mark internal stdlib functions with attribute_hidden to allow direct
access within libc.so and libc.a without using GOT nor PLT. __realpath
is hidden with libc_hidden_proto and libc_hidden_def since the exported
realpath is an alias of __realpath.
H.J. Lu [Sun, 1 Oct 2017 21:33:26 +0000 (14:33 -0700)]
Mark 3 *_internal functions with attribute_hidden [BZ #18822]
Mark __ptsname_internal, __mktime_internal and __fopen_internal with
attribute_hidden to allow direct access to them within libc.so and
libc.a without using GOT nor PLT.
Fix return pointer save and PIC register restore in __getcontext and
__setcontext on hppa.
* sysdeps/unix/sysv/linux/hppa/getcontext.S (__getcontext): Save return
pointer in frame.
* sysdeps/unix/sysv/linux/hppa/setcontext.S (__setcontext): Likewise.
Correct offset used to restore PIC register.
H.J. Lu [Sat, 30 Sep 2017 08:39:14 +0000 (01:39 -0700)]
Add a test for non-PIE static executable
tst-tls1-static-non-pie is built with $(no-pie-ldflag) to make it a
non-PIE static executable, regardless if --enable-static-pie is used
to configure glibc.
Joseph Myers [Fri, 29 Sep 2017 23:54:33 +0000 (23:54 +0000)]
Use libm_alias_double for some dbl-64 functions.
Continuing the move of libm aliases to common macros that can create
_FloatN / _FloatNx aliases in future, this patch converts some dbl-64
functions to using libm_alias_double, thereby eliminating the need for
some ldbl-opt wrappers.
This patch deliberately limits what functions are converted so that it
can be verified by comparison of stipped binaries. Specifically, atan
and tan are excluded because they first need converting to being weak
aliases; fma is omitted as it has additional complications with
versions in other directories (removing the ldbl-opt version can
e.g. cause the ldbl-128 version to be used instead of dbl-64); and
functions that have both dbl-64/wordsize-64 and ldbl-opt versions are
excluded because ldbl-opt currently always wraps dbl-64 function
versions, so changing those will result in platforms using both
ldbl-opt and dbl-64/wordsize-64 (i.e. alpha) starting to use the
dbl-64/wordsize-64 versions of those functions (which is good, as an
optimization, but still best separated from the present patch to get
better validation).
Tested for x86_64, and tested with build-many-glibcs.py that installed
stripped shared libraries are unchanged by the patch.
* sysdeps/ieee754/dbl-64/s_asinh.c: Include <libm-alias-double.h>.
(asinh): Define using libm_alias_double.
* sysdeps/ieee754/dbl-64/s_cbrt.c: Include <libm-alias-double.h>.
(cbrt): Define using libm_alias_double.
* sysdeps/ieee754/dbl-64/s_copysign.c: Include
<libm-alias-double.h>.
(copysign): Define using libm_alias_double.
* sysdeps/ieee754/dbl-64/s_erf.c: Include <libm-alias-double.h>.
(erf): Define using libm_alias_double.
(erfc): Likewise.
* sysdeps/ieee754/dbl-64/s_expm1.c: Include <libm-alias-double.h>.
(expm1): Define using libm_alias_double.
* sysdeps/ieee754/dbl-64/s_fabs.c: Include <libm-alias-double.h>.
(fabs): Define using libm_alias_double.
* sysdeps/ieee754/dbl-64/s_fromfp.c (fromfp): Define using
libm_alias_double.
* sysdeps/ieee754/dbl-64/s_fromfp_main.c: Include
<libm-alias-double.h>.
* sysdeps/ieee754/dbl-64/s_fromfpx.c (fromfpx): Define using
libm_alias_double.
* sysdeps/ieee754/dbl-64/s_getpayload.c: Include
<libm-alias-double.h>.
(getpayload): Define using libm_alias_double.
* sysdeps/ieee754/dbl-64/s_llrint.c: Include
<libm-alias-double.h>.
(llrint): Define using libm_alias_double.
* sysdeps/ieee754/dbl-64/s_lrint.c: Include <libm-alias-double.h>.
(lrint): Define using libm_alias_double.
* sysdeps/ieee754/dbl-64/s_nextup.c: Include
<libm-alias-double.h>.
(nextup): Define using libm_alias_double.
* sysdeps/ieee754/dbl-64/s_roundeven.c: Include
<libm-alias-double.h>.
(roundeven): Define using libm_alias_double.
* sysdeps/ieee754/dbl-64/s_setpayload.c (setpayload): Define using
libm_alias_double.
* sysdeps/ieee754/dbl-64/s_setpayload_main.c: Include
<libm-alias-double.h>.
* sysdeps/ieee754/dbl-64/s_setpayloadsig.c (setpayloadsig): Define
using libm_alias_double.
* sysdeps/ieee754/dbl-64/s_sin.c: Include <libm-alias-double.h>.
(cos): Define using libm_alias_double.
(sin): Likewise.
* sysdeps/ieee754/dbl-64/s_sincos.c: Include
<libm-alias-double.h>.
(sincos): Define using libm_alias_double.
* sysdeps/ieee754/dbl-64/s_tanh.c: Include <libm-alias-double.h>.
(tanh): Define using libm_alias_double.
* sysdeps/ieee754/dbl-64/s_totalorder.c: Include
<libm-alias-double.h>.
(totalorder): Define using libm_alias_double.
* sysdeps/ieee754/dbl-64/s_totalordermag.c: Include
<libm-alias-double.h>.
(totalordermag): Define using libm_alias_double.
* sysdeps/ieee754/dbl-64/s_ufromfp.c (ufromfp): Define using
libm_alias_double.
* sysdeps/ieee754/dbl-64/s_ufromfpx.c (ufromfpx): Define using
libm_alias_double.
* sysdeps/ieee754/dbl-64/wordsize-64/s_getpayload.c: Include
<libm-alias-double.h>.
(getpayload): Define using libm_alias_double.
* sysdeps/ieee754/dbl-64/wordsize-64/s_roundeven.c: Include
<libm-alias-double.h>.
(roundeven): Define using libm_alias_double.
* sysdeps/ieee754/dbl-64/wordsize-64/s_setpayload_main.c: Include
<libm-alias-double.h>.
* sysdeps/ieee754/dbl-64/wordsize-64/s_totalorder.c: Include
<libm-alias-double.h>.
(totalorder): Define using libm_alias_double.
* sysdeps/ieee754/dbl-64/wordsize-64/s_totalordermag.c: Include
<libm-alias-double.h>.
(totalordermag): Define using libm_alias_double.
* sysdeps/ieee754/ldbl-opt/s_copysign.c (copysignl): Only define
libc compat symbol here.
* sysdeps/ieee754/ldbl-opt/s_asinh.c: Remove file.
* sysdeps/ieee754/ldbl-opt/s_cbrt.c: Likewise.
* sysdeps/ieee754/ldbl-opt/s_erf.c: Likewise.
* sysdeps/ieee754/ldbl-opt/s_expm1.c: Likewise.
* sysdeps/ieee754/ldbl-opt/s_fabs.c: Likewise.
* sysdeps/ieee754/ldbl-opt/s_llrint.c: Likewise.
* sysdeps/ieee754/ldbl-opt/s_lrint.c: Likewise.
* sysdeps/ieee754/ldbl-opt/s_sin.c: Likewise.
* sysdeps/ieee754/ldbl-opt/s_sincos.c: Likewise.
* sysdeps/ieee754/ldbl-opt/s_tanh.c: Likewise.
H.J. Lu [Fri, 29 Sep 2017 23:23:18 +0000 (16:23 -0700)]
arm: Update elf_machine_load_address for static PIE
When --enable-static-pie is used to configure glibc, we need to use
_dl_relocate_static_pie to compute load address in static PIE.
* sysdeps/arm/dl-machine.h (elf_machine_load_address): Use
_dl_relocate_static_pie instead of _dl_start to compute load
address in static PIE. Return 0 if _DYNAMIC is undefined for
static executable.
A few math functions still use __fabs(f/l) rather than fabs, which
means they won't be inlined. Rename them so they are inlined.
Also add -fno-builtin-fabsl to nofpu powerpc makefile to work around
BZ #29253.
* sysdeps/ieee754/dbl-64/e_lgamma_r.c
(__ieee754_lgamma_r): Use fabs rather than __fabs.
* sysdeps/ieee754/dbl-64/e_log10.c (__ieee754_log10): Likewise.
* sysdeps/ieee754/dbl-64/e_log2.c (__ieee754_log2): Likewise.
* sysdeps/ieee754/flt-32/e_lgammaf_r.c
(__ieee754_lgammaf_r): Use fabsf rather than __fabsf.
* sysdeps/ieee754/flt-32/e_log10f.c (__ieee754_log10f): Likewise.
* sysdeps/ieee754/flt-32/e_log2f.c (__ieee754_log2f): Likewise.
* sysdeps/ieee754/ldbl-128/e_lgammal_r.c
(__ieee754_lgammal_r): Use fabsl rather than __fabsl.
* sysdeps/ieee754/ldbl-128/e_log10l.c (__ieee754_log10l): Likewise.
* sysdeps/ieee754/ldbl-128/e_log2l.c (__ieee754_log2l): Likewise.
* sysdeps/ieee754/ldbl-128ibm/e_lgammal_r.c
(__ieee754_lgammal_r): Use fabsl rather than __fabsl.
* sysdeps/ieee754/ldbl-128ibm/e_log10l.c (__ieee754_log10l): Likewise.
* sysdeps/ieee754/ldbl-128ibm/e_log2l.c (__ieee754_log2l): Likewise.
* sysdeps/powerpc/nofpu/Makefile: Add -fno-builtin-fabsl for BZ #29253.
Szabolcs Nagy [Mon, 4 Sep 2017 16:55:33 +0000 (17:55 +0100)]
New generic powf
without wrapper on aarch64:
powf reciprocal-throughput: 4.2x faster
powf latency: 2.6x faster
old worst-case error: 1.11 ulp
new worst-case error: 0.82 ulp
aarch64 .text size: -780 bytes
aarch64 .rodata size: +144 bytes
powf(x,y) is implemented as exp2(y*log2(x)) with the same algorithms
that are used in exp2f and log2f, except that the log2f polynomial is
larger for extra precision and its output (and exp2f input) may be
scaled by a power of 2 (POWF_SCALE) to simplify the argument reduction
step of exp2 (possible when efficient round and convert toint operation
is available).
The special case handling tries to minimize the checks in the hot path.
When the input of exp2_inline is checked, int arithmetics is used as
that was faster on the tested aarch64 cores.
* math/Makefile (type-float-routines): Add e_powf_log2_data.
* sysdeps/ieee754/flt-32/e_powf.c: New implementation.
* sysdeps/ieee754/flt-32/e_powf_log2_data.c: New file.
* sysdeps/ieee754/flt-32/math_config.h (__powf_log2_data): Define.
(issignalingf_inline): Likewise.
(POWF_LOG2_TABLE_BITS): Likewise.
(POWF_LOG2_POLY_ORDER): Likewise.
(POWF_SCALE_BITS): Likewise.
(POWF_SCALE): Likewise.
* sysdeps/i386/fpu/e_powf_log2_data.c: New file.
* sysdeps/ia64/fpu/e_powf_log2_data.c: New file.
* sysdeps/m68k/m680x0/fpu/e_powf_log2_data.c: New file.
H.J. Lu [Thu, 28 Sep 2017 22:28:00 +0000 (15:28 -0700)]
x86: Allow undefined _DYNAMIC in static executable
When --enable-static-pie is used to build static PIE, _DYNAMIC is used
to compute the load address of static PIE. But _DYNAMIC is undefined
when creating static executable. This patch makes _DYNAMIC weak in PIE
libc.a so that it can be undefined.
* sysdeps/i386/dl-machine.h (elf_machine_load_address): Allow
undefined _DYNAMIC in PIE libc.a.
* sysdeps/x86_64/dl-machine.h (elf_machine_load_address):
Likewse.
Remove __signbit inlines from mathinline.h. Math.h already uses
the builtin when supported, so additional inlines are only used
on pre 4.0 GCCs. Similarly remove ancient copysign and fabs
inlines.
Simplify the C99 isgreater macros. Although some support was added
in GCC 2.97, not all targets added support until GCC 3.1. Therefore
only use the builtins in math.h from GCC 3.1 onwards, and defer to
generic macros otherwise. Improve the generic isunordered macro
to use compares rather than call fpclassify twice - this is not only
faster but also correct for signaling NaNs.
Joseph Myers [Thu, 28 Sep 2017 01:59:02 +0000 (01:59 +0000)]
Fix nearbyint arithmetic moved before feholdexcept (bug 22225).
In <https://sourceware.org/ml/libc-alpha/2013-05/msg00722.html> I
remarked on the possibility of arithmetic in various nearbyint
implementations being scheduled before feholdexcept calls, resulting
in spurious "inexact" exceptions.
I'm now actually observing this occurring in glibc built for ARM with
GCC 7 (in fact, both copies of the same addition/subtraction sequence
being combined and moved out before the conditionals and
feholdexcept/fesetenv pairs), resulting in test failures.
This patch makes the nearbyint implementations with this particular
feholdexcept / arithmetic / fesetenv pattern consistently use
math_opt_barrier on the function argument when first used in
arithmetic, and also consistently use math_force_eval before fesetenv
(the latter was generally already done, but the dbl-64/wordsize-64
implementation used math_opt_barrier instead, and as
math_opt_barrier's intended effect is through its output value being
used, such a use that doesn't use the return value is suspect).
Tested for x86_64 (--disable-multi-arch so more of these
implementations get used), and for ARM in a configuration where I saw
the problem scheduling.
[BZ #22225]
* sysdeps/ieee754/dbl-64/s_nearbyint.c (__nearbyint): Use
math_opt_barrier on argument when doing arithmetic on it.
* sysdeps/ieee754/dbl-64/wordsize-64/s_nearbyint.c (__nearbyint):
Likewise. Use math_force_eval not math_opt_barrier after
arithmetic.
* sysdeps/ieee754/flt-32/s_nearbyintf.c (__nearbyintf): Use
math_opt_barrier on argument when doing arithmetic on it.
* sysdeps/ieee754/ldbl-128/s_nearbyintl.c (__nearbyintl):
Likewise.
Samuel Thibault [Wed, 27 Sep 2017 22:19:18 +0000 (00:19 +0200)]
hurd: Fix `getifaddrs' and `freeifaddrs' symbol exposition
from `freeaddrinfo'.
`getifaddrs' and `freeifaddrs' are not in POSIX, they should not be
exposed along `freeaddrinfo' (through `__check_pf') which is POSIX.
* include/ifaddrs.h (__getifaddrs, __freeifaddrs): New declarations,
and use libc_hidden_def on them.
* inet/ifaddrs.c (__getifaddrs, __freeifaddrs): Use libc_hidden_def on
them.
* sysdeps/gnu/ifaddrs.c (__getifaddrs, __freeifaddrs): Likewise.
* inet/check_pf.c (__check_pf): Use __getifaddrs and __freeifaddrs
instead of getifaddrs and freeifaddrs.
Samuel Thibault [Wed, 27 Sep 2017 22:41:02 +0000 (00:41 +0200)]
hurd: Fix `seekdir' symbol exposition from `rewinddir'
`seekdir' is MISC || XOPEN, it should not be exposed along `rewinddir' which
is POSIX.
* include/dirent.h (__seekdir): New declaration.
* sysdeps/mach/hurd/seekdir.c (seekdir): Rename to __seekdir and
redefine as weak alias.
* sysdeps/mach/hurd/rewinddir.c (__rewinddir): Use __seekdir instead
of seekdir.
Samuel Thibault [Wed, 27 Sep 2017 22:01:40 +0000 (00:01 +0200)]
hurd: Fix `revoke' symbol exposition from `unlockpt'
`revoke' is MISC only, it should not be exposed along `unlockpt' which is
XOPEN.
* include/unistd.h (__revoke): New declaration.
* misc/revoke.c (revoke): Rename to __revoke, and redefine as weak
alias.
* sysdeps/mach/hurd/revoke.c (revoke): Likewise.
* sysdeps/unix/bsd/unlockpt.c (unlockpt): Use __revoke instead of
revoke.
Joseph Myers [Tue, 26 Sep 2017 21:21:01 +0000 (21:21 +0000)]
Fix make-syscalls.sh VDSO support for GCC 8.
sysdeps/unix/make-syscalls.sh has support, used only by x32, for
generating IFUNCs for kernel VDSO symbols. This support creates
IFUNCs by setting symbol types manually, which is bad for debug info
and does not work with current GCC mainline because it results in
errors from the checks on types of function aliases.
This patch fixes it to use the common __ifunc macro, which uses the
ifunc attribute when available and so works with GCC mainline. Note
however that the original error resulted from an indirect inclusion of
a header declaring __gettimeofday from the generated sources, and
using __ifunc now relies on such an indirect inclusion remaining as it
means use of __typeof to determine the correct types. If glibc's
headers change in such a way as to remove that indirect inclusion, it
will become necessary to change the syscalls.list syntax for VDSO
syscalls so the name of the header to include can be specified.
Tested (compilation only) with build-many-glibcs.py that this fixes
the build for x32 with GCC mainline.
* sysdeps/unix/make-syscalls.sh: Use __ifunc to define symbols
using VDSO.
Current implementation of tunables does not set arena_max and arena_test
values. Any value provided by glibc.malloc.arena_max and
glibc.malloc.arena_test parameters is ignored.
These tunables have minval value set to 1 (see elf/dl-tunables.list file)
and undefined maxval value. In that case default value (which is 0. see
scripts/gen-tunables.awk) is being used to set maxval.
For instance, generated tunable_list[] entry for arena_max is:
(gdb) p *cur
$1 = {name = 0x7ffff7df6217 "glibc.malloc.arena_max",
type = {type_code = TUNABLE_TYPE_SIZE_T, min = 1, max = 0},
val = {numval = 0, strval = 0x0}, initialized = false,
security_level = TUNABLE_SECLEVEL_SXID_IGNORE,
env_alias = 0x7ffff7df622e "MALLOC_ARENA_MAX"}
As a result, any value of glibc.malloc.arena_max is ignored by
TUNABLE_SET_VAL_IF_VALID_RANGE macro
__type min = (__cur)->type.min; <- initialized to 1
__type max = (__cur)->type.max; <- initialized to 0!
if (min == max) <- false
{
min = __default_min;
max = __default_max;
}
if ((__type) (__val) >= min && (__type) (val) <= max) <- false
{
(__cur)->val.numval = val;
(__cur)->initialized = true;
}
Assigning correct min/max values at a build time fixes a problem.
Plus, a bit of optimization: Setting of default min/max values for the
given type at a run time might be eliminated.
* elf/dl-tunables.c (do_tunable_update_val): Range checking fix.
* scripts/gen-tunables.awk: Set unspecified minval and/or maxval
values to correct default value for given type.
with 0 file size. ld.so should skip such PT_DYNAMIC segments.
Without a PT_DYNAMIC segment the loading of the shared object will
fail, and therefore ldd on such objects will also fail instead of
crashing. This provides better diagnostics for tooling that is
attempting to inspect the invalid shared objects which may just
contain debug information.
Joseph Myers [Tue, 26 Sep 2017 19:49:33 +0000 (19:49 +0000)]
Use generic __ifunc for SPARC.
glibc fails to build with GCC mainline for SPARC because of the use of
manually-created IFUNCs, which fail the tests of compatibility of
function alias types. This patch changes sparc-ifunc.h to use the
generic __ifunc in defining sparc_libm_ifunc. The generic __ifunc can
use the GCC ifunc attribute when available, so ensuring
type-correctness as well as better debug info than when setting symbol
types in asm statements.
Note that for this to fix the build with GCC mainline the GCC patch
<https://gcc.gnu.org/ml/gcc-patches/2017-09/msg01779.html>, or
building GCC with --enable-gnu-indirect-function, is also needed.
Tested (compilation only) with build-many-glibcs.py (sparc64-linux-gnu
and sparcv9-linux-gnu, with GCC 8 with the above patch, and also with
GCC 7).
* sysdeps/sparc/sparc-ifunc.h [!__ASSEMBLER__] (sparc_libm_ifunc):
Define using __ifunc.
Joseph Myers [Tue, 26 Sep 2017 16:30:46 +0000 (16:30 +0000)]
Fix ia64 executable stack default (bug 22156).
As per https://gcc.gnu.org/ml/gcc-patches/2017-09/msg01220.html ia64
defaults to non-executable stacks in the Linux kernel (furthermore,
the use of function descriptors means that trampolines for nested
function pointers never need an executable stack). glibc however
defines DEFAULT_STACK_PERMS to include PF_X for that architecture,
meaning (a) elf/check-execstack fails and (b) (from code inspection,
not tested, but this is why I think this is a user-visible bug) thread
stacks are unnecessarily mapped with execute permission. This patch
fixes the DEFAULT_STACK_PERMS definition in question.
Tested (compilation only) with build-many-glibcs.py for ia64. This
fixes the check-execstack failure.
Problem reported by Florian Weimer [1] and solution suggested by
Andreas Schwab [2]. It also set the same buffer size independent
of architecture max_align_t size.
Checked on x86_64-linux-gnu and i686-linux-gnu.
* lib/malloc/scratch_buffer.h (struct scratch_buffer):
Use an union instead of a max_align_t array for __space,
so that __space is the same size on all platforms.
* malloc/scratch_buffer_grow_preserve.c
(__libc_scratch_buffer_grow_preserve): Likewise.
posix: Add compat glob symbol to not follow dangling symbols
This patch follows commit 5554304f0 (posix: Allow glob to match dangling
symlinks [BZ #866]) by adding a compat symbol that follow previous
semantic of not following dangling symlinks and thus avoiding call
gl_lstat with GLOB_ALTDIRFUNC.
It avoids failure with old binaries that not set the alternate function
pointer for lstat (GNUmake for instance). The following scenario, for
instance, fails with current GNUmake because glibc will access unitialized
memory when calling gl_lstat:
obj/%.o: src/%.c
$(CC) $(CFLAGS) -c $< -o $@
$ make
This works as expected with the patch applied. Since it is for generic
ABI, default compat symbols are added with override for Linux due LFS.
Now we have two compat symbols for glob on Linux:
1. sysdeps/unix/sysv/linux/oldglob.c which implements glob64 with
the old dirent layout. For this implementation I also set it to
not follow dangling symlinks (which is the safest path).
2. sysdeps/unix/sysv/linux/glob{64}-lstat-compat.c which implements
the compat symbol for dangling symlinks. As for generic glob,
the implementation uses XSTAT_IS_XSTAT64 to define whether
both __glob_lstat_compat and __glob64_lstat_compat should be
different implementations. For archictures that define
XSTAT_IS_XSTAT64, __glob_lstat_compat is aliased to
__glob64_lstat_compat.
3. sysdeps/unix/sysv/linux/alpha/oldglob.c with a different glob_t
layout. As for 1. this patch changes it to not follow dangling
symlinks.
The patch also bumps _GNU_GLOB_INTERFACE_VERSION to 2 to advertise the
new semantic. On GNUmake, for instance, it will force to it use its
internal glob implementation instead and avoiding triggering the same
failure on builds against newer GLIBCs.
Checked on x86_64-linux-gnu and i686-linux-gnu. I also checked
with a build against the major ABIs required to check for the abilist.
The changes should also work on gnulib (I run gnulib-tool.py check glob
and it shown no regressions).
Paul Pluzhnikov [Tue, 26 Sep 2017 00:51:34 +0000 (17:51 -0700)]
Reduce total memory required to create all threads to 128MiB. This fixes
intermittent failure in stdlib/test-{atexit,at_quick_exit,...} tests
(Bug 22207).
Reviewed-by: Jonathan Nieder <jrnieder@gmail.com> Reviewed-by: Andreas Schwab <schwab@suse.de>
A lock is held by mempool_allocate() when CACHEABLE is true; we
must release this lock if we exit early.
Szabolcs Nagy [Tue, 12 Sep 2017 11:20:50 +0000 (12:20 +0100)]
New expf and exp2f version without SVID compat wrapper
This patch changes the expf and exp2f error handling semantics to only
set errno accoring to POSIX rules. New symbol version is introduced at
GLIBC_2.27.
The old wrappers are kept for compat symbols.
Internal calls to __expf now get the new error semantics, this seems to
only affect sysdeps/i386/fpu/s_expm1f.S where the errno-only behaviour
should be correct.
ia64 needed assembly change to have the new and compat versioned symbol
map to the same function.
Szabolcs Nagy [Wed, 6 Sep 2017 16:42:00 +0000 (17:42 +0100)]
Optimized generic expf and exp2f with wrappers
Based on new expf and exp2f code from
https://github.com/ARM-software/optimized-routines/
with wrapper on aarch64:
expf reciprocal-throughput: 2.3x faster
expf latency: 1.7x faster
without wrapper on aarch64:
expf reciprocal-throughput: 3.3x faster
expf latency: 1.7x faster
without wrapper on aarch64:
exp2f reciprocal-throughput: 2.8x faster
exp2f latency: 1.3x faster
libm.so size on aarch64:
.text size: -152 bytes
.rodata size: -1740 bytes
expf/exp2f worst case nearest rounding error: 0.502 ulp
worst case non-nearest rounding error: 1 ulp
Error checks are inline and errno setting is in separate tail called
functions, but the wrappers are kept in this patch to handle the
_LIB_VERSION==_SVID_ case. (So e.g. errno is set twice for expf calls
and once for __expf_finite calls on targets where the new code is used.)
Double precision arithmetics is used which is expected to be faster on
most targets (including soft-float) than using single precision and it
is easier to get good precision result with it.
Const data is kept in a separate translation unit which complicates
maintenance a bit, but is expected to give good code for literal loads
on most targets and allows sharing data across expf, exp2f and powf.
(This data is disabled on i386, m68k and ia64 which have their own
expf, exp2f and powf code.)
Some details may need target specific tweaks:
- best convert and round to int operation in the arg reduction may be
different across targets.
- code was optimized on fma target, optimal polynomial eval may be
different without fma.
- gcc does not always generate good code for fp bit representation
access via unions or it may be inherently slow on some targets.
The libm-test-ulps will need adjustment because..
- The argument reduction ideally uses nearest rounded rint, but that is
not efficient on most targets, so the polynomial can get evaluated on a
wider interval in non-nearest rounding mode making 1 ulp errors common
in that case.
- The polynomial is evaluated such that it may have 1 ulp error on
negative tiny inputs with upward rounding.
* math/Makefile (type-float-routines): Add math_errf and e_exp2f_data.
* sysdeps/aarch64/fpu/math_private.h (TOINT_INTRINSICS): Define.
(roundtoint, converttoint): Likewise.
* sysdeps/ieee754/flt-32/e_expf.c: New implementation.
* sysdeps/ieee754/flt-32/e_exp2f.c: New implementation.
* sysdeps/ieee754/flt-32/e_exp2f_data.c: New file.
* sysdeps/ieee754/flt-32/math_config.h: New file.
* sysdeps/ieee754/flt-32/math_errf.c: New file.
* sysdeps/ieee754/flt-32/t_exp2f.h: Remove.
* sysdeps/i386/fpu/e_exp2f_data.c: New file.
* sysdeps/i386/fpu/math_errf.c: New file.
* sysdeps/ia64/fpu/e_exp2f_data.c: New file.
* sysdeps/ia64/fpu/math_errf.c: New file.
* sysdeps/m68k/m680x0/fpu/e_exp2f_data.c: New file.
* sysdeps/m68k/m680x0/fpu/math_errf.c: New file.
Samuel Thibault [Sun, 24 Sep 2017 23:55:02 +0000 (01:55 +0200)]
hurd: Fix exposition of s/gettimeofday through timespec_s/get
conform/ISO11/time.h/linknamespace complains that using timespec_get exposes
gettimeofday.
conform/POSIX/time.h/linknamespace complains that using clock_settime
exposes settimeofday.
* sysdeps/unix/clock_gettime.c (realtime_gettime, __clock_gettime): Use
__gettimeofday instead of gettimeofday.
* sysdeps/unix/clock_settime.c (__clock_settime): Use __settimeofday
instead of settimeofday.
Samuel Thibault [Sun, 10 Sep 2017 15:41:03 +0000 (17:41 +0200)]
resolv_test.c: also cope with CONNREFUSED errors returned by recvfrom
server_thread_udp_process_one already takes care of calling sendto()
instead of xsendto to be able to ignore the case where the client has
closed the socket. Depending on the TCP/IP stack behavior, this error
could be notified later through recvfrom(), so we need to ignore it
there too.
* support/resolv_test.c (server_thread_udp_process_one): Call recvfrom
instead of xrecvfrom, and ignore ECONNREFUSED errors.
Samuel Thibault [Sun, 24 Sep 2017 20:20:10 +0000 (22:20 +0200)]
hurd: Fix bits/socket.h conformity
* sysdeps/mach/hurd/bits/socket.h: Include <bits/wordsize.h> instead
of <limits.h>
(__need_NULL): Do not define.
(__ss_aligntype): Use __WORDSIZE instead of ULONG_MAX to determine
alignment.
[!__USE_MISC] (pseudo_AF_XTP, pseudo_AF_RTIP, pseudo_AF_PIP,
CMGROUP_MAX, cmsgcred): Do not define.
(CMSG_FIRSTHDR, __cmsg_nxthdr): Use (struct cmsghdr *) 0 instead of
NULL.
* bits/socket.h: Likewise.
Samuel Thibault [Sun, 24 Sep 2017 15:54:02 +0000 (17:54 +0200)]
hurd: Make sure dl-sysdep.c defines proper symbol names
* sysdeps/mach/hurd/dl-sysdep.c (check_no_hidden): New macro.
(__open, __close, __libc_read, __libc_write, __writev, __libc_lseek64,
__mmap, __fxstat64, __xstat64, __access, __access_noerrno, __getpid,
__getcwd, __sbrk, __strtoul_internal, _exit, abort): Use check_no_hidden
to make sure that these symbols are defined.