The M_ARENA_MAX and M_ARENA_TEST macros are defined in malloc.c as
well as malloc.h, and the former is unnecessary. This patch removes
the duplicate. Tested on x86_64 to verify that the generated code
remains unchanged barring changed line numbers to __malloc_assert.
The manual incorrectly references sbrk as the method used to grow and
shrink heaps and the fact that M_TRIM_THRESHOLD and M_TOP_PAD control
that behavior. In reality, a heap may be grown or shrunk through
multiple methods depending on whether it is the main arena (in which
case sbrk is correct) or not (in which case, there are a number of
strategies including allocating an additional heap to grow an arena
and/or 'mprotect' a region to make it available for allocation).
Remove references to sbrk so that it covers the behavior more
accurately.
* manual/memory.texi (M_TOP_PAD): Remove reference to sbrk.
(M_TRIM_THRESHOLD): Likewise.
The M_ARENA_* mallopt parameters are in wide use in production to
control the number of arenas that a long lived process creates and
hence there is no point in stating that this interface is non-public.
Document this interface and remove the obsolete comment.
The mallopt parameters manual does not mention the environment
variables that can be used to set these parameters at program startup.
Mention those environment variables for completeness.
* manual/memory.texi: Add environment variable alternatives to
setting mallopt parameters.
Carlos O'Donell [Wed, 26 Oct 2016 04:00:13 +0000 (00:00 -0400)]
Fix atomic_fetch_xor_release.
No code uses atomic_fetch_xor_release except for the upcoming
conditional variable rewrite. Therefore there is no user
visible bug here. The use of atomic_compare_and_exchange_bool_rel
is removed (since it doesn't exist anymore), and is replaced
by atomic_compare_exchange_weak_release.
We use weak_release because it provides better performance in
the loop (the weak semantic) and because the xor is release MO
(the release semantic). We don't reload expected in the loop
because atomic_compare_and_exchange_weak_release does this for
us as part of the CAS failure.
It is otherwise a fairly plain conversion that fixes building
the new condvar for 32-bit x86. Passes all regression tests
for x86.
ISO/IEC TS 18661-1 adds several functions in the strfrom family to stdlib.
This patch adds strfromd, strfromf, and strfroml. This is being done in
preparation for the new floating-point type, float128. The added functions
convert a floating-point value into a string, with configurable format.
Joseph Myers [Tue, 25 Oct 2016 15:54:16 +0000 (15:54 +0000)]
Fix cmpli usage in power6 memset.
Building glibc for powerpc64 with recent (2.27.51.20161012) binutils,
with multi-arch enabled, I get the error:
../sysdeps/powerpc/powerpc64/power6/memset.S: Assembler messages:
../sysdeps/powerpc/powerpc64/power6/memset.S:254: Error: operand out of range (5 is not between 0 and 1)
../sysdeps/powerpc/powerpc64/power6/memset.S:254: Error: operand out of range (128 is not between 0 and 31)
../sysdeps/powerpc/powerpc64/power6/memset.S:254: Error: missing operand
Indeed, cmpli is documented as a four-operand instruction, and looking
at nearby code it seems likely cmpldi was intended. This patch fixes
this powerpc64 code accordingly, and makes a corresponding change to
the powerpc32 code.
Tested for powerpc, powerpc64 and powerpc64le by Tulio Magno Quites
Machado Filho
* sysdeps/powerpc/powerpc32/power6/memset.S (memset): Use cmplwi
instead of cmpli.
* sysdeps/powerpc/powerpc64/power6/memset.S (memset): Use cmpldi
instead of cmpli.
Although conceptually correct for p{read,write}{64} offset argument passing,
sh4 implementation does not generate the correct expected code. The
__ALIGNMENT_ARG redefinition is incorrect for two reasons: 1. the
kernel-features.h header is included multiple times (since it contains no
guards) and 2. the value it redefines is also incorrect (should be '0, '
instead of empty definition).
This patch fixes it by adding another macro, SYSCALL_LL_PRW{64}, meant to be
used to pass the offset argument on p{read,write}64. It is basically the
already define SYSCALL_LL{64} plus __ALIGNMENT_ARG unless __ASSUME_PRW_DUMMY_ARG
is define. In this case an empty dummy argument is used regardless how
__ALIGNMENT_ARG is defined (sh4 case).
Checked on x86_64, i686, aarch64, armhf, and powerpc64le (basically a sanity
check). Also, John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de> and
James Clarke <jrtc27@jrtc27.com> help me check on a debian sh4 bootstrap using
2.24 plus this patch to verify it also corrected fixed the regression issue.
I also verified the generated object for a 2.24 build and master with this
patch for sh4 and both look identical.
* sysdeps/unix/sysv/linux/pread.c (__libc_pread): Use SYSCALL_LL_PRW.
* sysdeps/unix/sysv/linux/pwrite.c (__libc_pwrite): Likewise.
* sysdeps/unix/sysv/linux/pread64.c (__libc_pread64): Use
SYSCALL_LL64_PRW.
* sysdeps/unix/sysv/linux/pwrite64.c (__libc_pwrite64): Likewise.
* sysdeps/unix/sysv/linux/sh/kernel-features.h: Define
__ASSUME_PRW_DUMMY_ARG.
* sysdeps/unix/sysv/linux/sh/pread.c: Remove file.
* sysdeps/unix/sysv/linux/sh/pread64.c: Likewise.
* sysdeps/unix/sysv/linux/sh/pwrite.c: Likewise.
* sysdeps/unix/sysv/linux/sh/pwrite64.c: Likewise.
* sysdeps/unix/sysv/linux/sysdep.h: Define SYSCALL_LL_PRW and
SYSCALL_LL_PRW64 based on __ASSUME_PRW_DUMMY_ARG.
Florian Weimer [Sat, 22 Oct 2016 15:33:26 +0000 (17:33 +0200)]
math.h: Wrap C++ bits in extern "C++"
It is still common to include system header files in an extern "C"
block. This means that exiting <math.h>'s own extern "C" block
is not sufficient to get back to C++ mode. Use an extern "C++"
wrapper instead.
Joseph Myers [Fri, 21 Oct 2016 21:18:03 +0000 (21:18 +0000)]
Use -fno-builtin for sqrt benchmark.
This patch makes the sqrt benchmark use -fno-builtin, as already done
for benchmarks of ffs and ffsll, so that it actually benchmarks the
glibc function as (presumably) intended even in the presence of the
compiler inlining sqrt.
Tested for x86_64 and also used for benchmarking my ARM sqrt patch.
* benchtests/Makefile (CFLAGS-bench-sqrt.c): New variable.
Paul Murphy [Fri, 10 Jun 2016 19:50:32 +0000 (14:50 -0500)]
powerpc: Cleanup fenv_private.h
Some of the masks are wrong, and the naming is confusing.
There are two basic cases we really care about:
1. Stacking a new rounding mode when running certain
sections of code, and pausing exception handling.
2. Likewise, but discarding any exceptions which occur
while running under the new rounding mode.
libc_feholdexcept_setround_ppc_ctx has been removed as it basically
does the same thing as libc_feholdsetround_ppc_ctx but also clearing
any sticky bits. The restore behavior is what differentiates these
two cases as the SET_RESTORE_ROUND{,_NOEX} macros will either merge
or discard all exceptions occurring during scope of their usage.
Likewise, there are a number of routines to swap, replace,
or merge FP environments. This change reduces much of
the common and sometimes wrong code.
Joseph Myers [Thu, 20 Oct 2016 23:24:44 +0000 (23:24 +0000)]
Use VSQRT instruction for ARM sqrt (bug 20660).
This patch makes ARM sqrt and sqrtf use the VSQRT VFP square root
instruction when available, instead of much larger generic code for
computing square roots.
Now, GCC will normally inline sqrt calls except for negative arguments
where errno needs to be set, and because the benchtests fail to use
-fno-builtin that means no significant difference in benchmark results
for sqrt (note, however, there are lots of __ieee754_sqrt calls
internally in libm, which are *not* inlined - although some
architectures define __ieee754_sqrt in their math_private.h for that
purpose, ARM doesn't - so improving out-of-line sqrt performance is
still relevant to those other functions, if not for most ordinary
direct users of sqrt). With the benchtests changed to use
-fno-builtin for sqrt tests, typical performance results before the
change are ("max" is wildly varying in any case):
Joseph Myers [Wed, 19 Oct 2016 22:58:34 +0000 (22:58 +0000)]
Stop powerpc copysignl raising "invalid" for sNaN argument (bug 20718).
The powerpc (hard-float) implementations of copysignl, both 32-bit and
64-bit, raise spurious "invalid" exceptions when the first argument is
a signaling NaN. copysign functions should never raise exceptions
even for signaling NaNs.
The problem is the use of an fcmpu instruction to test the sign of the
high part of the long double argument. This patch fixes the functions
to use fsel instead (as used for fabsl following my fixes for a
similar bug there), or to examine the integer representation for older
32-bit processors without fsel.
Tested for powerpc64 and powerpc32 (configurations with and without
fsel used).
[BZ #20718]
* sysdeps/powerpc/powerpc32/fpu/s_copysignl.S (__copysignl): Do
not use floating-point comparisons to test sign.
* sysdeps/powerpc/powerpc64/fpu/s_copysignl.S (__copysignl):
Likewise.
Joseph Myers [Wed, 19 Oct 2016 01:49:09 +0000 (01:49 +0000)]
Add getpayload, getpayloadf, getpayloadl.
TS 18661-1 defines functions for manipulating the payloads of NaNs.
This patch implements the getpayload functions for glibc; these
extract the NaN payload (from an argument passed as a pointer, for
which corresponding libm-test support is added) and return it in the
same floating-point type. The return value of these functions is
unspecified for non-NaN arguments; the patch does the simplest thing
to implement, which is that the functions do not check whether the
argument is a NaN and just treat the relevant bits of the
representation as a payload regardless. A conversion from integer to
floating-point is used to produce the required return value, except in
the ldbl-128 case; as 128-bit integers are not supported for all
configurations using ldbl-128, the code constructs the required
floating-point representation of the return value directly instead.
The fallocate syscall might fail on Linux due missing support from
underlying filesystem (for instance some NFS versions). This patch
adds this check for fallocate tests. It also moves tst-fallocate{64}
to 'io' folder (since it is on fallocate{64} is built).
Checked on x86_64.
* sysdeps/unix/sysv/linux/Makefile [$(subdir) = math] (tests): Move
tst-fallocate{64}.
* sysdeps/unix/sysv/linux/tst-fallocate-common.c: Check for EOPNOTSUPP
on syscall return.
Carlos O'Donell [Mon, 17 Oct 2016 23:35:34 +0000 (19:35 -0400)]
Bug 20689: Fix FMA and AVX2 detection on Intel
In the Intel Architecture Instruction Set Extensions Programming
reference the recommended way to test for FMA in section
'2.2.1 Detection of FMA' is:
"Application Software must identify that hardware supports AVX as
explained in ... after that it must also detect support for FMA..."
We don't do that in glibc. We use osxsave to detect the use of xgetbv,
and after that we check for AVX and FMA orthogonally. It is conceivable
that you could have the AVX bit clear and the FMA bit in an undefined
state.
This commit fixes FMA and AVX2 detection to depend on usable AVX
as required by the recommended Intel sequences.
Joseph Myers [Mon, 17 Oct 2016 22:48:51 +0000 (22:48 +0000)]
Define HIGH_ORDER_BIT_IS_SET_FOR_SNAN to 0 or 1.
This patch moves the HIGH_ORDER_BIT_IS_SET_FOR_SNAN macro from being
defined or undefined to the preferred convention of always being
defined, to either 0 or 1, so allowing typo-proof tests with #if.
The macro is moved from math_private.h to a new header
nan-high-order-bit.h to make it easy for all architectures to define,
either through the sysdeps/generic version of the header or through
providing their own version of the header, without needing #ifndef in
the generic math_private.h to give a default definition. The move
also allows the macro to be used without needing math_private.h to be
included; the immediate motivation of this patch is to allow tests to
access this information (to know what kinds of NaNs 0 is a valid
payload for) without needing to include math_private.h. Existing
C level rather than preprocessor conditionals at all, but this patch
does not make such a change).
Tested for x86_64 and x86 (testsuite); also verified for x86_64, x86,
mips64 and powerpc that installed stripped shared libraries are
unchanged by the patch.
* sysdeps/generic/nan-high-order-bit.h: New file.
* sysdeps/hppa/nan-high-order-bit.h: Likewise.
* sysdeps/mips/nan-high-order-bit.h: Likewise.
* sysdeps/hppa/math_private.h: Remove file.
* sysdeps/mips/math_private.h (HIGH_ORDER_BIT_IS_SET_FOR_SNAN): Do
not define here.
* sysdeps/ieee754/dbl-64/s_issignaling.c: Include
<nan-high-order-bit.h>.
[HIGH_ORDER_BIT_IS_SET_FOR_SNAN]: Test with #if not #ifdef.
* sysdeps/ieee754/dbl-64/s_totalorder.c: Include
<nan-high-order-bit.h>.
[HIGH_ORDER_BIT_IS_SET_FOR_SNAN]: Test with #if not #ifdef.
* sysdeps/ieee754/dbl-64/s_totalordermag.c: Include
<nan-high-order-bit.h>.
[HIGH_ORDER_BIT_IS_SET_FOR_SNAN]: Test with #if not #ifdef.
* sysdeps/ieee754/dbl-64/wordsize-64/s_issignaling.c: Include
<nan-high-order-bit.h>.
[HIGH_ORDER_BIT_IS_SET_FOR_SNAN]: Test with #if not #ifdef.
* sysdeps/ieee754/dbl-64/wordsize-64/s_totalorder.c: Include
<nan-high-order-bit.h>.
[HIGH_ORDER_BIT_IS_SET_FOR_SNAN]: Test with #if not #ifdef.
* sysdeps/ieee754/dbl-64/wordsize-64/s_totalordermag.c: Include
<nan-high-order-bit.h>.
[HIGH_ORDER_BIT_IS_SET_FOR_SNAN]: Test with #if not #ifdef.
* sysdeps/ieee754/flt-32/s_issignalingf.c: Include
<nan-high-order-bit.h>.
[HIGH_ORDER_BIT_IS_SET_FOR_SNAN]: Test with #if not #ifdef.
* sysdeps/ieee754/flt-32/s_totalorderf.c: Include
<nan-high-order-bit.h>.
[HIGH_ORDER_BIT_IS_SET_FOR_SNAN]: Test with #if not #ifdef.
* sysdeps/ieee754/flt-32/s_totalordermagf.c: Include
<nan-high-order-bit.h>.
[HIGH_ORDER_BIT_IS_SET_FOR_SNAN]: Test with #if not #ifdef.
* sysdeps/ieee754/ldbl-128/s_issignalingl.c: Include
<nan-high-order-bit.h>.
[HIGH_ORDER_BIT_IS_SET_FOR_SNAN]: Test with #if not #ifdef.
* sysdeps/ieee754/ldbl-128/s_totalorderl.c: Include
<nan-high-order-bit.h>.
[HIGH_ORDER_BIT_IS_SET_FOR_SNAN]: Test with #if not #ifdef.
* sysdeps/ieee754/ldbl-128/s_totalordermagl.c: Include
<nan-high-order-bit.h>.
[HIGH_ORDER_BIT_IS_SET_FOR_SNAN]: Test with #if not #ifdef.
* sysdeps/ieee754/ldbl-128ibm/s_issignalingl.c: Include
<nan-high-order-bit.h>.
[HIGH_ORDER_BIT_IS_SET_FOR_SNAN]: Test with #if not #ifdef.
* sysdeps/ieee754/ldbl-128ibm/s_totalorderl.c: Include
<nan-high-order-bit.h>.
[HIGH_ORDER_BIT_IS_SET_FOR_SNAN]: Test with #if not #ifdef.
* sysdeps/ieee754/ldbl-128ibm/s_totalordermagl.c: Include
<nan-high-order-bit.h>.
[HIGH_ORDER_BIT_IS_SET_FOR_SNAN]: Test with #if not #ifdef.
* sysdeps/ieee754/ldbl-96/s_issignalingl.c: Include
<nan-high-order-bit.h>.
[HIGH_ORDER_BIT_IS_SET_FOR_SNAN]: Test with #if not #ifdef.
* sysdeps/ieee754/ldbl-96/s_totalorderl.c: Include
<nan-high-order-bit.h>.
[HIGH_ORDER_BIT_IS_SET_FOR_SNAN]: Test with #if not #ifdef.
* sysdeps/ieee754/ldbl-96/s_totalordermagl.c: Include
<nan-high-order-bit.h>.
[HIGH_ORDER_BIT_IS_SET_FOR_SNAN]: Test with #if not #ifdef.
Stefan Liebler [Mon, 17 Oct 2016 08:37:11 +0000 (10:37 +0200)]
S390: Fix fp comparison not raising FE_INVALID.
As gcc is using unordered comparison instructions which do not
raise invalid exception if any operand is quiet NAN,
FIX_COMPARE_INVALID is defined to 1.
Thus iseqsig is calling feraiseexcept as workaround.
In addition to the totalorder functions, TS 18661-1 defines
totalordermag functions, which do the same comparison but on the
absolute values of the arguments. This patch implements these
functions for glibc, including the type-generic macro in <tgmath.h>.
In general the implementations are similar to but simpler than those
for the totalorder functions.
Joseph Myers [Fri, 14 Oct 2016 00:12:56 +0000 (00:12 +0000)]
Clean up some complex functions raising FE_INVALID.
Some of the complex arithmetic functions have the following pattern:
in some piece of code, one part of the input (real or imaginary,
depending on the function) is either infinite or NaN. Part of the
result is to be set to NaN in either case, and FE_INVALID raised only
if the relevant part of the input was infinite.
In such a case, there is no actual need for the conditional on the
type of the input, since subtracting the relevant part of the input
from itself will produce a NaN, with FE_INVALID only if the relevant
part of the input was infinite. This simplifies the code, and as a
quality-of-implementation matter also improves things by propagating
NaN payloads. (Right now these functions always raise FE_INVALID for
signaling NaN arguments because of the call to fpclassify - at least
unless glibc is built with -Os - but if fpclassify moves to using
integer arithmetic in future, doing arithmetic on the NaN argument
also ensures an exception for sNaNs.)
Tested for x86_64 and x86.
* math/s_ccosh_template.c (M_DECL_FUNC (__ccosh)): Instead of
raising FE_INVALID with feraisexcept in case where part of
argument is infinite, subtract that part of argument from itself.
* math/s_cexp_template.c (M_DECL_FUNC (__cexp)): Likewise.
* math/s_csin_template.c (M_DECL_FUNC (__csin)): Likewise.
* math/s_csinh_template.c (M_DECL_FUNC (__csinh)): Likewise.
Recent binutils versions (at least 2.27) complains about libc.so
when linking sotruss-lib.so with:
libc.so:(*IND*+0x0): multiple definition of `posix_fadvise64@GLIBC_2.2'
libc.so::(.text+0xcf940): first defined here
libc.so:(*IND*+0x0): multiple definition of `posix_fadvise64'
collect2: error: ld returned 1 exit status
Dynamic symbols for libc.so shows (readelf --dyn-syms):
That is, two separate definitions at version GLIBC_2.2. The issue is
sysdeps/unix/sysv/linux/posix_fadvise64.c creates posix_fadvise64 weak_alias,
while sysdeps/unix/sysv/linux/mips/mips64/n64/posix_fadvise64.c then adds
compat_symbol / versioned_symbol calls.
The patch remove the weak_alias definition on mips64 specific version so
direct weak_alias is disabled.
Checked on mips64n64 build with binutils 2.27.51.20161012.
H.J. Lu [Wed, 12 Oct 2016 15:22:52 +0000 (08:22 -0700)]
X86: Don't assert on older Intel CPUs [BZ #20647]
Since the maximum CPUID level of older Intel CPUs is 1, change
handle_intel to return -1, instead of assert, when the maximum
CPUID level is less than 2.
[BZ #20647]
* sysdeps/x86/cacheinfo.c (handle_intel): Return -1 if the
maximum CPUID level is less than 2.
Joseph Myers [Wed, 12 Oct 2016 01:20:30 +0000 (01:20 +0000)]
Add totalorder, totalorderf, totalorderl.
TS 18661-1 defines totalorder functions implementing the totalOrder
comparison operation from IEEE 754-2008. This patch implements these
functions for glibc, including the type-generic macro in <tgmath.h>.
(The totalordermag functions will be added in a separate patch.)
The description of the totalOrder operation is complicated. However,
for IEEE interchange binary formats and the preferred quiet NaN
convention, what that complicated description means is that you
interpret the representation as a sign-magnitude integer (with -0
coming before +0) and do a <= comparison on that interpretation. For
finite values and infinities the ordering of the sign-magnitude
integers is just the same as the ordering of floating-point values, so
this extends that to all representations. (Different representations
of the same floating-point value - which includes same quantum in the
decimal case - must still be considered equal by this operation, but
that issue doesn't arise for IEEE interchange binary formats.) So the
complications are:
* When MIPS quiet NaN conventions are in use, the representation of
NaNs needs adjusting before making such an integer comparison. This
patch does this adjustment only when both arguments are NaNs, as
there's no need for it if only one is a NaN, and as long as both are
NaNs you can just flip the relevant bits without any problems from
this turning a NaN into an infinity.
* For the m68k version of ldbl-96, where the high mantissa bit is
"don't care" for infinities and NaNs, representations where it
differs must compare the same. Note: although the testcase for this
compiles, I have not actually tested on m68k.
* For ldbl-128ibm, the low part must be ignored when the high part is
NaN, and low parts of +0 and -0 must be considered the same whatever
the high part.
The new tests in libm-test.inc are the first tests there specifying
particular payloads for input NaNs. Separate tests are also added for
the ldbl-96 and ldbl-128ibm special cases where there are different
representations of the same value that must compare equal (which can't
be covered in libm-test.inc as that only specifies values, not
representations).
Tested for x86_64, x86, mips64 and powerpc.
* math/bits/mathcalls.h [__GLIBC_USE (IEC_60559_BFP_EXT)]
(totalorder): New declaration.
* math/tgmath.h [__GLIBC_USE (IEC_60559_BFP_EXT)] (totalorder):
New macro.
* math/Versions (totalorder): New libm symbol at version
GLIBC_2.25.
(totalorderf): Likewise.
(totalorderl): Likewise.
* math/Makefile (libm-calls): Add s_totalorderF.
* math/gen-libm-test.pl (parse_args): Escape quotes in test name
string.
* math/libm-test.inc (PAYLOAD_DIG): New macro.
(qnan_value_pl): Likewise.
(snan_value_pl): Likewise.
(qnan_value): Define using qnan_value_pl.
(snan_value): Define using snan_value_pl.
(struct test_ff_i_data): Add comment about which tests use this
structure.
(RUN_TEST_ff_b): New macro.
(RUN_TEST_LOOP_ff_b): Likewise.
(totalorder_test_data): New array.
(totalorder_test): New function.
(main): Call totalorder_test.
* math/test-tgmath.c (NCALLS): Increase to 122.
(F(compile_test)): Call totalorder.
(F(totalorder)): New function.
* manual/arith.texi (FP Comparison Functions): Document
totalorder, totalorderf and totalorderl.
* manual/libm-err-tab.pl: Update comment on interfaces without
ulps tabulated.
* sysdeps/ieee754/dbl-64/s_totalorder.c: New file.
* sysdeps/ieee754/dbl-64/wordsize-64/s_totalorder.c: Likewise.
* sysdeps/ieee754/flt-32/s_totalorderf.c: Likewise.
* sysdeps/ieee754/ldbl-128/s_totalorderl.c: Likewise.
* sysdeps/ieee754/ldbl-128ibm/s_totalorderl.c: Likewise.
* sysdeps/ieee754/ldbl-96/s_totalorderl.c: Likewise.
* sysdeps/ieee754/ldbl-opt/nldbl-totalorder.c: Likewise.
* sysdeps/ieee754/ldbl-opt/Makefile (libnldbl-calls): Add
totalorder.
(CFLAGS-nldbl-totalorder.c): New variable.
* sysdeps/ieee754/ldbl-128ibm/test-totalorderl-ldbl-128ibm.c: New
file.
* sysdeps/ieee754/ldbl-128ibm/Makefile [$(subdir) = math] (tests):
Add test-totalorderl-ldbl-128ibm.
* sysdeps/ieee754/ldbl-96/test-totalorderl-ldbl-96.c: New file.
* sysdeps/ieee754/ldbl-96/Makefile [$(subdir) = math] (tests): Add
test-totalorderl-ldbl-96.
* sysdeps/nacl/libm.abilist: Update.
* sysdeps/unix/sysv/linux/aarch64/libm.abilist: Likewise.
* sysdeps/unix/sysv/linux/alpha/libm.abilist: Likewise.
* sysdeps/unix/sysv/linux/arm/libm.abilist: Likewise.
* sysdeps/unix/sysv/linux/hppa/libm.abilist: Likewise.
* sysdeps/unix/sysv/linux/i386/libm.abilist: Likewise.
* sysdeps/unix/sysv/linux/ia64/libm.abilist: Likewise.
* sysdeps/unix/sysv/linux/m68k/coldfire/libm.abilist: Likewise.
* sysdeps/unix/sysv/linux/m68k/m680x0/libm.abilist: Likewise.
* sysdeps/unix/sysv/linux/microblaze/libm.abilist: Likewise.
* sysdeps/unix/sysv/linux/mips/mips32/libm.abilist: Likewise.
* sysdeps/unix/sysv/linux/mips/mips64/libm.abilist: Likewise.
* sysdeps/unix/sysv/linux/nios2/libm.abilist: Likewise.
* sysdeps/unix/sysv/linux/powerpc/powerpc32/fpu/libm.abilist:
Likewise.
* sysdeps/unix/sysv/linux/powerpc/powerpc32/nofpu/libm.abilist:
Likewise.
* sysdeps/unix/sysv/linux/powerpc/powerpc64/libm-le.abilist:
Likewise.
* sysdeps/unix/sysv/linux/powerpc/powerpc64/libm.abilist:
Likewise.
* sysdeps/unix/sysv/linux/s390/s390-32/libm.abilist: Likewise.
* sysdeps/unix/sysv/linux/s390/s390-64/libm.abilist: Likewise.
* sysdeps/unix/sysv/linux/sh/libm.abilist: Likewise.
* sysdeps/unix/sysv/linux/sparc/sparc32/libm.abilist: Likewise.
* sysdeps/unix/sysv/linux/sparc/sparc64/libm.abilist: Likewise.
* sysdeps/unix/sysv/linux/tile/tilegx/tilegx32/libm.abilist:
Likewise.
* sysdeps/unix/sysv/linux/tile/tilegx/tilegx64/libm.abilist:
Likewise.
* sysdeps/unix/sysv/linux/tile/tilepro/libm.abilist: Likewise.
* sysdeps/unix/sysv/linux/x86_64/64/libm.abilist: Likewise.
* sysdeps/unix/sysv/linux/x86_64/x32/libm.abilist: Likewise.
This patch consolidates all the sync_file_range implementation for Linux
in only one (sysdeps/unix/sysv/linux/sync_file_range.c). It also removes
the syscall from the auto-generation using assembly macros (except for
x86_64 due x32 [1]).
For current minimum supported kernel (2.6.32 for x86_64 and 3.2 for all
other architectures) either sync_file_range or sync_file_range2 is supported
and it is expected that any future Linux ABI will provide either of one
syscall. So the code path that returns ENOSYS in the case of missing
syscall is removed.
Checked on x86_64, i386, powerpc64le, aarch64, and armhf.
* sysdeps/unix/sysv/linux/Makefile (tests): Add tst-sync_file_range.
* sysdeps/unix/sysv/linux/mips/mips32/sync_file_range.c: Remove file.
* sysdeps/sysv/linux/powerpc/powerpc64/sync_file_range.c: Likewise.
* sysdeps/unix/sysv/linux/sync_file_range.c: New file.
* sysdeps/unix/sysv/linux/tst-sync_file_range.c (sync_file_range):
Consolidate all Linux implementations.
Joseph Myers [Tue, 11 Oct 2016 00:50:43 +0000 (00:50 +0000)]
Avoid M_NAN + M_NAN in complex functions.
Some libm complex functions have code that computes M_NAN + M_NAN.
This is nonsensical; it's just equivalent to M_NAN, since it's a quiet
NaN (and the comments suggesting this raises an exception are
similarly wrong). This patch changes the code just to use M_NAN (and
removes the bogus comments). (Preferably, code should either
propagate an input NaN or do a computation that raises "invalid" and
generates a default NaN at the same time. There are various cases,
however, that currently raise "invalid" even for NaN inputs; I think
those are cases where "invalid" is optional in ISO C so a change to
whether it's raised would be OK, but they would still need more
careful consideration than the cases where such issues do not arise.)
Tested for x86_64 and x86 (testsuite, and that installed stripped
shared libraries are unchanged by the patch).
* math/s_ccosh_template.c (M_DECL_FUNC (__ccosh)): Use M_NAN
instead of M_NAN + M_NAN.
* math/s_csinh_template.c (M_DECL_FUNC (__csinh)): Likewise.
Joseph Myers [Fri, 7 Oct 2016 23:28:33 +0000 (23:28 +0000)]
Make iseqsig handle excess precision.
iseqsig, like other type-generic comparison macros, should behave like
a comparison operator in not removing excess range and precision from
its arguments (see C11 F.10.11). This patch implements this by making
definitions of iseqsig appropriately conditional on
__FLT_EVAL_METHOD__ (including support for TS 18661-3 values of that
macro), with a corresponding testcase (that failed for 32-bit x86 in
the absence of the math.h changes) being added. (Of course the
definitions may need reworking when float128 support is added, just as
with other type-generic macros.)
Tested for x86_64 and x86.
* math/math.h [__GLIBC_USE (IEC_60559_BFP_EXT)] (iseqsig): Define
conditional on value of [__FLT_EVAL_METHOD__].
* math/test-iseqsig-excess-precision.c: New file.
* math/Makefile (tests): Add test-iseqsig-excess-precision.
This patch consolidates mostly of the Linux posix_fadvise{64} implementations
on sysdeps/unix/sysv/linux/posix_fadvise{64}.c. It still keeps arch-specific
files for:
* S390-32: it uses a packed structure to pass all the arguments on syscall.
It is the only supported port that implements __NR_fadvise64_64 in this
way.
* ARM: it does not implement __NR_fadvise64 (as other 32-bits ports), so
posix_fadvise calls internal posix_fadvise64 symbol.
* MIPS64 n64: it requires a different version number that other ports.
The new macro SYSCALL_LL{64} is used to handle the offset argument and
INTERNAL_SYSCALL_CALL to handle passing the correct number of expect
arguments.
The default Linux adds two new defines a port can use to control how
__NR_fadvise64_64 passes the kernel arguments:
* __ASSUME_FADVISE64_64_6ARG: the 'advise' argument is moved on second
position. This is the case of powerpc32 and arm to avoid implement
7 argument syscall.
* __ASSUME_FADVISE64_64_NO_ALIGN: for ABIs that defines
__ASSUME_ALIGNED_REGISTER_PAIRS packs the offset without the leading
'0'. This is the case of tile 32 bits.
ARM also defines __NR_fadvise64_64 as __NR_arm_fadvise64_64 (which is also
handled on arch kernel-feature.h).
This patch consolidates all the posix_fallocate{64} implementation for Linux
in only one (sysdeps/unix/sysv/linux/posix_fallocate{64}.c). It also removes
the syscall from the auto-generation using assembly macros.
The macro SYSCALL_LL{64} is used to handle the offset argument along with
the new INTERNAL_SYSCALL_CALL macro to define correct argument count for
internal INTERNAL_SYSCALL call.
Tested on x86_64, i686, x32, aarch64, ppc64le, and armhf.
* io/Makefile (tests): Add tst-posix_fallocate and
tst-posix_fallocate64.
* io/tst-posix_fallocate-common.c: New file.
* io/tst-posix_fallocate.c: Likewise.
* io/tst-posix_fallocate64.c: Likewise.
* sysdeps/unix/sysv/linux/mips/mips64/n32/posix_fallocate.c: Remove
file.
* sysdeps/unix/sysv/linux/mips/mips64/n32/posix_fallocate64.c:
Likewise.
* sysdeps/unix/sysv/linux/mips/mips64/n64/posix_fallocate.c: Likewise.
* sysdeps/unix/sysv/linux/mips/mips64/n64/posix_fallocate64.c:
Likewise.
* sysdeps/unix/sysv/linux/wordsize-64/posix_fallocate.c: Likewise.
* sysdeps/unix/sysv/linux/wordsize-64/posix_fallocate64.c: Likewise.
* sysdeps/unix/sysv/linux/posix_fallocate.c (posix_fallocate): Use
SYSCALL_LL to pass both offset and len arguments.
* sysdeps/unix/sysv/linux/posix_fallocate64.c (posix_fallocate64):
Likewise.
* sysdeps/unix/sysv/linux/x86_64/syscalls.list (pwrite64): Add
__libc_pwrite64 alias used by posix_fallocate64.
This patch consolidates all the fallocate{64} implementation for Linux
in only one (sysdeps/unix/sysv/linux/fallocate{64}.c). It also removes the
syscall from the auto-generation using assembly macros.
The new macro SYSCALL_LL{64} is used to handle the offset argument.
Checked on x86_64, x32, i386, aarch64, and ppc64le.
Florian Weimer [Fri, 7 Oct 2016 11:43:48 +0000 (13:43 +0200)]
resolv: Remove RES_USEBSTRING and its implementation [BZ #20629]
In ns_name_ntop, the NS_CMPRSFLGS check is no longer needed because
labellen (called earlier) already rejects everything which is not
a plain label (compression references and extended label types).
Stefan Liebler [Fri, 7 Oct 2016 07:56:48 +0000 (09:56 +0200)]
Use libc_ifunc macro for siglongjmp, longjmp in libpthread.
This patch uses the libc_ifunc macro to create already existing ifunc functions
longjmp_ifunc, siglongjmp_ifunc if HAVE_IFUNC is defined.
The s390 pt-longjmp.c includes the common pt-longjmp.c and uses strong_alias
to create the longjmp, siglongjmp symbols for glibc version 2.19.
ChangeLog:
* nptl/pt-longjmp.c (DEFINE_LONGJMP): Use libc_ifunc macro.
* sysdeps/unix/sysv/linux/s390/pt-longjmp.c (longjmp, siglongjmp):
Use strong_alias to create symbols for glibc verison 2.19.
Stefan Liebler [Fri, 7 Oct 2016 07:56:47 +0000 (09:56 +0200)]
Use libc_ifunc macro for clock_* symbols in librt.
This patch uses the libc_ifunc macro to create already existing ifunc functions
clock_getres, clock_gettime, clock_settime, clock_getcpuclockid and
clock_nanosleep. If HAVE_IFUNC is defined, the macro COMPAT_REDIRECT uses
the libc_ifunc macro.
Furthermore some whitespace damage is cleaned.
ChangeLog:
* rt/clock-compat.c (COMPAT_REDIRECT): Use libc_ifunc macro.
Stefan Liebler [Fri, 7 Oct 2016 07:56:47 +0000 (09:56 +0200)]
ppc: Use libc_ifunc macro for time, gettimeofday.
This patch uses the libc_ifunc_hidden macro to create already existing ifunc functions
time and gettimeofday on power. This way, the libc_hidden_def macro can be used
instead of inline assemblies.
On ppc32, the __GI_* symbols do not target the ifunc symbol and thus the
redirection construct has to be applied here.
ChangeLog:
* sysdeps/unix/sysv/linux/powerpc/gettimeofday.c (__gettimeofday):
Use libc_ifunc_hidden and libc_hidden_def macro. Redirect ifunced function
in header for using it as type for ifunc function because __GI_* symbols
for ppc32 do not target the ifunc symbols.
* sysdeps/unix/sysv/linux/powerpc/time.c (time): Likewise.
Stefan Liebler [Fri, 7 Oct 2016 07:56:47 +0000 (09:56 +0200)]
i386, x86: Use libc_ifunc macro for time, gettimeofday.
This patch uses the libc_ifunc_hidden macro to create already existing ifunc functions
time and gettimeofday on intel. This way, the libc_hidden_def macro can be used
instead of the libc_ifunc_hidden_def one which was only used here. Thus the
macro is removed from libc-symbols.h.
On i386, the __GI_* symbols do not target the ifunc symbol and thus the
redirection construct has to be applied here.
ChangeLog:
* sysdeps/unix/sysv/linux/x86/gettimeofday.c (__gettimeofday):
Use libc_ifunc_hidden macro. Use libc_hidden_def instead of
libc_ifunc_hidden_def.
* sysdeps/unix/sysv/linux/x86/time.c (time): Likewise.
* sysdeps/unix/sysv/linux/i386/gettimeofday.c (__gettimeofday):
Redirect ifunced function in header for using it as type of ifunc'ed
function. Redefine libc_hidden_def to use fallback non ifunc'ed
function for __GI_* symbol.
* sysdeps/unix/sysv/linux/i386/time.c (time): Likewise.
* include/libc-symbols.h
(libc_ifunc_hidden_def, libc_ifunc_hidden_def1): Delete macro.
Stefan Liebler [Fri, 7 Oct 2016 07:56:47 +0000 (09:56 +0200)]
s390: Refactor ifunc resolvers due to false debuginfo.
This patch adjusts the s390 specific ifunc helper macros in ifunc-resolve.h to
use the common __ifunc macro, which uses gcc attribute ifunc to get rid of the
false debuginfo. Therefore the redirection construct is applied where needed.
Perhaps in future we can switch some of the internal symbols __GI_* from the
fallback variant to the ifunc function. But this change is not
straightforward due to a segmentation fault while linking libc.so with older
binutils on s390.
ChangeLog:
[BZ #20478]
* sysdeps/s390/multiarch/ifunc-resolve.h
(s390_vx_libc_ifunc2, s390_libc_ifunc): Use __ifunc from libc-symbols.h
to create ifunc symbols.
(s390_vx_libc_ifunc_init, s390_vx_libc_ifunc_redirected
, s390_vx_libc_ifunc2_redirected, s390_libc_ifunc_init): New define.
* sysdeps/s390/multiarch/memchr.c: Redirect ifunced function in header
for using it as type for ifunc function.
* sysdeps/s390/multiarch/mempcpy.c: Likewise.
* sysdeps/s390/multiarch/rawmemchr.c: Likewise.
* sysdeps/s390/multiarch/stpcpy.c: Likewise.
* sysdeps/s390/multiarch/stpncpy.c: Likewise.
* sysdeps/s390/multiarch/strcat.c: Likewise.
* sysdeps/s390/multiarch/strchr.c: Likewise.
* sysdeps/s390/multiarch/strcmp.c: Likewise.
* sysdeps/s390/multiarch/strcpy.c: Likewise.
* sysdeps/s390/multiarch/strcspn.c: Likewise.
* sysdeps/s390/multiarch/strlen.c: Likewise.
* sysdeps/s390/multiarch/strncmp.c: Likewise.
* sysdeps/s390/multiarch/strncpy.c: Likewise.
* sysdeps/s390/multiarch/strnlen.c: Likewise.
* sysdeps/s390/multiarch/strpbrk.c: Likewise.
* sysdeps/s390/multiarch/strrchr.c: Likewise.
* sysdeps/s390/multiarch/strspn.c: Likewise.
* sysdeps/s390/multiarch/wcschr.c: Likewise.
* sysdeps/s390/multiarch/wcscmp.c: Likewise.
* sysdeps/s390/multiarch/wcspbrk.c: Likewise.
* sysdeps/s390/multiarch/wcsspn.c: Likewise.
* sysdeps/s390/multiarch/wmemchr.c: Likewise.
* sysdeps/s390/multiarch/wmemset.c: Likewise.
* sysdeps/s390/s390-32/multiarch/memcmp.c: Likewise.
* sysdeps/s390/s390-32/multiarch/memcpy.c: Likewise.
* sysdeps/s390/s390-32/multiarch/memset.c: Likewise.
* sysdeps/s390/s390-64/multiarch/memcmp.c: Likewise.
* sysdeps/s390/s390-64/multiarch/memcpy.c: Likewise.
* sysdeps/s390/s390-64/multiarch/memset.c: Likewise.
Stefan Liebler [Fri, 7 Oct 2016 07:56:46 +0000 (09:56 +0200)]
Use gcc attribute ifunc in libc_ifunc macro instead of inline assembly due to false debuginfo.
The current s390 ifunc resolver for vector optimized functions and the common
libc_ifunc macro in include/libc-symbols.h uses something like that to generate ifunc'ed functions:
extern void *__resolve___strlen(unsigned long int dl_hwcap) asm (strlen);
asm (".type strlen, %gnu_indirect_function");
The debuginfo for the ifunc-resolver function contains the DW_AT_linkage_name
field, which names the real function name "strlen". If you perform an inferior
function call to strlen in lldb, then it fails due to something like that:
"error: no matching function for call to 'strlen'
candidate function not viable: no known conversion from 'const char [6]'
to 'unsigned long' for 1st argument"
The unsigned long is the dl_hwcap argument of the resolver function.
The strlen function itself has no debufinfo.
The s390 ifunc resolver for memset & co uses something like that:
asm (".globl FUNC"
".type FUNC, @gnu_indirect_function"
".set FUNC, __resolve_FUNC");
This way the debuginfo for the ifunc-resolver function does not conain the
DW_AT_linkage_name field and the real function has no debuginfo, too.
Using this strategy for the vector optimized functions leads to some troubles
for functions like strnlen. Here we have __strnlen and a weak alias strnlen.
The __strnlen function is the ifunc function, which is realized with the asm-
statement above. The weak_alias-macro can't be used here due to undefined symbol:
gcc ../sysdeps/s390/multiarch/strnlen.c -c ...
In file included from <command-line>:0:0:
../sysdeps/s390/multiarch/strnlen.c:28:24: error: ‘strnlen’ aliased to undefined symbol ‘__strnlen’
weak_alias (__strnlen, strnlen)
^
./../include/libc-symbols.h:111:26: note: in definition of macro ‘_weak_alias’
extern __typeof (name) aliasname __attribute__ ((weak, alias (#name)));
^
../sysdeps/s390/multiarch/strnlen.c:28:1: note: in expansion of macro ‘weak_alias’
weak_alias (__strnlen, strnlen)
^
make[2]: *** [build/string/strnlen.o] Error 1
As the __strnlen function is defined with asm-statements the function name
__strnlen isn't known by gcc. But the weak alias can also be done with an
asm statement to resolve this issue:
__asm__ (".weak strnlen\n\t"
".set strnlen,__strnlen\n");
In order to use the weak_alias macro, gcc needs to know the ifunc function. The
minimum gcc to build glibc is currently 4.7, which supports attribute((ifunc)).
See https://gcc.gnu.org/onlinedocs/gcc-4.7.0/gcc/Function-Attributes.html.
It is only supported if gcc is configured with --enable-gnu-indirect-function
or gcc supports it by default for at least intel and s390x architecture.
This patch uses the old behaviour if gcc support is not available.
Usage of attribute ifunc is something like that:
__typeof (FUNC) FUNC __attribute__ ((ifunc ("__resolve_FUNC")));
Then gcc produces the same .globl, .type, .set assembler instructions like above.
And the debuginfo does not contain the DW_AT_linkage_name field and there is no
debuginfo for the real function, too.
But in order to get it work, there is also some extra work to do.
Currently, the glibc internal symbol on s390x e.g. __GI___strnlen is not the
ifunc symbol, but the fallback __strnlen_c symbol. Thus I have to omit the
libc_hidden_def macro in strnlen.c (here is the ifunc function __strnlen)
because it is already handled in strnlen-c.c (here is __strnlen_c).
Due to libc_hidden_proto (__strnlen) in string.h, compiling fails:
gcc ../sysdeps/s390/multiarch/strnlen.c -c ...
In file included from <command-line>:0:0:
../sysdeps/s390/multiarch/strnlen.c:53:24: error: ‘strnlen’ aliased to undefined symbol ‘__strnlen’
weak_alias (__strnlen, strnlen)
^
./../include/libc-symbols.h:111:26: note: in definition of macro ‘_weak_alias’
extern __typeof (name) aliasname __attribute__ ((weak, alias (#name)));
^
../sysdeps/s390/multiarch/strnlen.c:53:1: note: in expansion of macro ‘weak_alias’
weak_alias (__strnlen, strnlen)
^
make[2]: *** [build/string/strnlen.os] Error 1
I have to redirect the prototypes for __strnlen in string.h and create a copy
of the prototype for using as ifunc function:
__typeof (__redirect___strnlen) __strnlen __attribute__ ((ifunc ("__resolve_strnlen")));
weak_alias (__strnlen, strnlen)
This way there is no trouble with the internal __GI_* symbols.
Glibc builds fine with this construct and the debuginfo is "correct".
For functions without a __GI_* symbol like memccpy this redirection is not needed.
This patch adjusts the common libc_ifunc and libm_ifunc macro to use gcc
attribute ifunc. Due to this change, the macro users where the __GI_* symbol
does not target the ifunc symbol have to be prepared with the redirection
construct.
Furthermore a configure check to test gcc support is added. If it is not supported,
the old behaviour is used.
This patch also prepares the libc_ifunc macro to be useable in s390-ifunc-macro.
The s390 ifunc-resolver-functions do have an hwcaps parameter and not all
resolvers need the same initialization code. The next patch in this series
changes the s390 ifunc macros to use this common one.
ChangeLog:
* include/libc-symbols.h (__ifunc_resolver):
New macro is used by __ifunc* macros.
(__ifunc): New macro uses gcc attribute ifunc or inline assembly
depending on HAVE_GCC_IFUNC.
(libc_ifunc, libm_ifunc): Use __ifunc as base macro.
(libc_ifunc_redirected, libc_ifunc_hidden, libm_ifunc_init): New macro.
* sysdeps/powerpc/powerpc32/power4/fpu/multiarch/s_finite.c:
Redirect ifunced function in header for using as type for ifunc function.
* sysdeps/powerpc/powerpc32/power4/fpu/multiarch/s_finitef.c: Likewise.
* sysdeps/powerpc/powerpc32/power4/fpu/multiarch/s_isinf.c: Likewise.
* sysdeps/powerpc/powerpc32/power4/fpu/multiarch/s_isinff.c: Likewise.
* sysdeps/powerpc/powerpc32/power4/fpu/multiarch/s_isnan.c: Likewise.
* sysdeps/powerpc/powerpc32/power4/multiarch/memcmp.c: Likewise.
* sysdeps/powerpc/powerpc32/power4/multiarch/memcpy.c: Likewise.
* sysdeps/powerpc/powerpc32/power4/multiarch/memmove.c: Likewise.
* sysdeps/powerpc/powerpc32/power4/multiarch/mempcpy.c: Likewise.
* sysdeps/powerpc/powerpc32/power4/multiarch/memset.c: Likewise.
* sysdeps/powerpc/powerpc32/power4/multiarch/rawmemchr.c: Likewise.
* sysdeps/powerpc/powerpc32/power4/multiarch/strchr.c: Likewise.
* sysdeps/powerpc/powerpc32/power4/multiarch/strlen.c: Likewise.
* sysdeps/powerpc/powerpc32/power4/multiarch/strncmp.c: Likewise.
* sysdeps/powerpc/powerpc32/power4/multiarch/strnlen.c: Likewise.
* sysdeps/powerpc/powerpc64/fpu/multiarch/s_finite.c: Likewise.
* sysdeps/powerpc/powerpc64/fpu/multiarch/s_finitef.c: Likewise.
* sysdeps/powerpc/powerpc64/fpu/multiarch/s_isinf.c: Likewise.
* sysdeps/powerpc/powerpc64/fpu/multiarch/s_isinff.c: Likewise.
* sysdeps/powerpc/powerpc64/fpu/multiarch/s_isnan.c: Likewise.
* sysdeps/powerpc/powerpc64/multiarch/memcmp.c: Likewise.
* sysdeps/powerpc/powerpc64/multiarch/mempcpy.c: Likewise.
* sysdeps/powerpc/powerpc64/multiarch/rawmemchr.c: Likewise.
* sysdeps/powerpc/powerpc64/multiarch/stpncpy.c: Likewise.
* sysdeps/powerpc/powerpc64/multiarch/strcat.c: Likewise.
* sysdeps/powerpc/powerpc64/multiarch/strchr.c: Likewise.
* sysdeps/powerpc/powerpc64/multiarch/strcmp.c: Likewise.
* sysdeps/powerpc/powerpc64/multiarch/strcpy.c: Likewise.
* sysdeps/powerpc/powerpc64/multiarch/strncmp.c: Likewise.
* sysdeps/powerpc/powerpc64/multiarch/strncpy.c: Likewise.
* sysdeps/powerpc/powerpc64/multiarch/strnlen.c: Likewise.
* sysdeps/powerpc/powerpc64/multiarch/strrchr.c: Likewise.
* sysdeps/powerpc/powerpc64/multiarch/strstr.c: Likewise.
* sysdeps/powerpc/powerpc64/multiarch/wcschr.c: Likewise.
* sysdeps/powerpc/powerpc32/power4/fpu/multiarch/s_isnanf.c:
Add libc_hidden_def() and use libc_ifunc_hidden() macro
instead of libc_ifunc() macro.
* sysdeps/powerpc/powerpc64/fpu/multiarch/s_isnanf.c: Likewise.
* sysdeps/powerpc/powerpc64/multiarch/stpcpy.c: Likewise.
Stefan Liebler [Fri, 7 Oct 2016 07:56:46 +0000 (09:56 +0200)]
Add configure check to test if gcc supports attribute ifunc.
This patch adds a configure check to test if gcc supports attribute ifunc.
The support can either be enabled in <gcc-src>/gcc/config.gcc for one
architecture in general by setting default_gnu_indirect_function variable to yes
or by configuring gcc with --enable-gnu-indirect-function.
The next patch rewrites libc_ifunc macro to use gcc attribute ifunc instead
of inline assembly to generate the IFUNC symbols due to false debuginfo.
If gcc does not support attribute ifunc, the old approach for generating
ifunc'ed symbols is used. Then the debug-information is false. Thus it is
recommended to use a gcc with indirect function support (See notes in INSTALL).
After this patch-series these inline assemblies for ifunc-handling are not
scattered in multiple files but are used only indirect via ifunc-macros
and can simply removed in libc-symbols.h in future.
If glibc is configured with --enable-multi-arch and gcc does not support
attribute ifunc, a configure warning is dumped!
ChangeLog:
* config.h.in (HAVE_GCC_IFUNC): New undef.
* configure.ac: Add check if gcc supports attribute ifunc feature.
* configure: Regenerated.
* manual/install.texi: Add recommendation for gcc with
indirect-function support.
* INSTALL: Regenerated.
Joseph Myers [Thu, 6 Oct 2016 22:19:38 +0000 (22:19 +0000)]
Add iseqsig.
TS 18661-1 adds an iseqsig type-generic comparison macro to <math.h>.
This macro is like the == operator except that unordered operands
result in the "invalid" exception and errno being set to EDOM.
This patch implements this macro for glibc. Given the need to set
errno, this is implemented with out-of-line functions __iseqsigf,
__iseqsig and __iseqsigl (of which the last only exists at all if long
double is ABI-distinct from double, so no function aliases or compat
support are needed). The present patch ignores excess precision
issues; I intend to deal with those in a followup patch. (Like
comparison operators, type-generic comparison macros should *not*
convert operands to their semantic types but should preserve excess
range and precision, meaning that for some argument types and values
of FLT_EVAL_METHOD, an underlying function should be called for a
wider type than that of the arguments.)
The underlying functions are implemented with the type-generic
template machinery. Comparing x <= y && x >= y is sufficient in ISO C
to achieve an equality comparison with "invalid" raised for unordered
operands (and the results of those two comparisons can also be used to
tell whether errno needs to be set). However, some architectures have
GCC bugs meaning that unordered comparison instructions are used
instead of ordered ones. Thus, a mechanism is provided for
architectures to use an explicit call to feraiseexcept to raise
exceptions if required. If your architecture has such a bug you
should add a fix-fp-int-compare-invalid.h header for it, with a
comment pointing to the relevant GCC bug report; if such a GCC bug is
fixed, that header's contents should have a __GNUC_PREREQ conditional
added so that the workaround can eventually be removed for that
architecture.
Make quadrant shift a boolean in reduce_and_compute in s_sin.c
Like the previous change, make the quadrant shift a boolean to make it
clearer that we will do at most a single rotation of the quadrants to
compute the cosine from the sine function.
Check n instead of k1 to decide on sign of sin/cos result
For k1 in 1 and 3, n can only have values of 0 and 2, so checking k1 &
2 is equivalent to checking n & 2. We prefer the latter so that we
don't use k1 for anything other than selecting the quadrant in
do_sincos_1, thus dropping it completely.
The previous logic was:
"Compute sine for the value and based on the new rotated quadrant
(k1) negate the value if we're in the fourth quadrant."
With this change, the logic now is:
"Compute sine for the value and negate it if we were either (1) in
the fourth quadrant or (2) we actually wanted the cosine and were
in the third quadrant."
* sysdeps/ieee754/dbl-64/s_sin.c (do_sincos_1): Check N
instead of K1.
Make the quadrant shift K a bool in do_sincos_* functions
The do_sincos_* functions are helpers to compute sin/cos, where they
get cosine by computing sine for the next quadrant. This is decided
with the value of K passed to it, which is the amount by which to
shift the quadrant. Since we will only need the shift to be 0 or 1,
we make K a bool to make that explicit.
* sysdeps/ieee754/dbl-64/s_sin.c (do_sincos_1): Rename K to
SHIFT_QUADRANT and make it bool.
(do_sincos_2): Likewise.
(sloww): Likewise.
(sloww1): Likewise.
(__sin): Adjust calls to do_sincos_1 and do_sincos_2.
(__cos): Likewise.
Joseph Myers [Tue, 4 Oct 2016 16:17:59 +0000 (16:17 +0000)]
Fix LONG_WIDTH, ULONG_WIDTH include ordering issue.
As described in
<https://sourceware.org/ml/libc-alpha/2016-10/msg00047.html>, there is
an include ordering issue with the integer width macros in glibc's
<limits.h>, where definitions conditional on LONG_MAX do not work as
intended because when the headers are installed, this part of glibc's
<limits.h> is processed before the part of GCC's <limits.h> that will
define LONG_MAX. This patch changes the definitions just to use
__WORDSIZE for the expansion of LONG_WIDTH and ULONG_WIDTH rather than
making those definitions conditional on LONG_MAX.
Tested for x86_64 and x86.
* include/limits.h [__GLIBC_USE (IEC_60559_BFP_EXT)] (LONG_WIDTH):
Define to __WORDSIZE, not conditional on [LONG_MAX ==
0x7fffffffL].
[__GLIBC_USE (IEC_60559_BFP_EXT)] (ULONG_WIDTH): Likewise.
Fix powerpc-specific headers:
- Make it compatible to C89 by replace references to inline by __inline__.
- Get the definition of sigset_t used by
sysdeps/unix/sysv/linux/powerpc/sys/ucontext.h.
- Includes missing header file.
Joseph Myers [Mon, 3 Oct 2016 21:01:42 +0000 (21:01 +0000)]
Add TCP_REPAIR_WINDOW from Linux 4.8.
Linux 4.8 adds TCP_REPAIR_WINDOW to include/uapi/linux/tcp.h. This
patch adds it to sysdeps/gnu/netinet/tcp.h accordingly, along with
struct tcp_repair_window as requested in
<https://sourceware.org/ml/libc-alpha/2016-10/msg00019.html>.
Tested for x86_64 and x86 (testsuite, and that installed shared
libraries are unchanged by the patch).
* sysdeps/gnu/netinet/tcp.h (TCP_REPAIR_WINDOW): New macro.
(struct tcp_repair_window): New type.
libio: Multiple fixes for open_{w}memstram (BZ#18241 and BZ#20181)
This patches fixes multiples issues on open_{w}memstream reported on both
BZ#18241 and BZ#20181:
- failed fseek does not set errno.
- negative offset in fseek fails even when resulting position is
a valid one.
- a flush after write if the current write position is not at the
end of the stream currupt data.
The main fix is on seek operation for memstream (_IO_{w}str_seekoff), where
both _IO_read_ptr and _IO_read_end pointer are updated if a write operation
has occured (similar to default file operations). Also, to calculate the
offset on both read and write pointers, a temporary value is instead of
updating the argument supplied value. Negative offset are valid if resulting
internal pointer is within the range of _IO_{read,write}_base and
_IO_{read,write}_end.
Also POSIX states that a null or wide null shall be appended to the current
buffer iff a write moves the position to a value larger than the current
lenght. Current implementation appends a null or wide null regardless
of this condition. This patch fixes it by removing the 'else' condition
on _IO_{w}mem_sync.
Checked on x86_64.
[BZ #18241]
[BZ #20181]
* libio/Makefile (test): Add tst-memstream3 and tst-wmemstream3.
* libio/memstream.c (_IO_mem_sync): Only append a null byte if
write position is at the end the buffer.
* libio/wmemstream.c (_IO_wmem_sync): Likewise.
* libio/strops.c (_IO_str_switch_to_get_mode): New function.
(_IO_str_seekoff): Set correct offset from negative displacement and
set EINVAL for invalid ones.
* libio/wstrops.c (enlarge_userbuf): Use correct function to calculate
buffer length.
(_IO_wstr_switch_to_get_mode): New function.
(_IO_wstr_seekoff): Set correct offset from negative displacement and
set EINVAL for invalid ones.
* libio/tst-memstream3.c: New file.
* libio/tst-wmemstream3.c: Likewise.
* manual/examples/memstrm.c: Remove warning when priting size_t.
Joseph Myers [Fri, 30 Sep 2016 15:49:51 +0000 (15:49 +0000)]
Use __builtin_fma more in dbl-64 code.
sysdeps/ieee754/dbl-64/dla.h can use a macro DLA_FMS for more
efficient double-width operations when fused multiply-subtract is
supported. However, this macro is only defined for x86_64,
conditional on architecture-specific __FMA4__. This patch makes the
code use __builtin_fma conditional on __FP_FAST_FMA, as used elsewhere
in glibc.
Tested for x86_64, x86 and powerpc. On powerpc (where this is causing
fused operations to be used where they weren't previously) I see an
increase from 1ulp to 2ulp in the imaginary part of clog10:
testing double (without inline functions)
Failure: Test: Imaginary part of: clog10 (0x1.7a858p+0 - 0x6.d940dp-4 i)
Result:
is: -1.2237865208199886e-01 -0x1.f5435146bb61ap-4
should be: -1.2237865208199888e-01 -0x1.f5435146bb61cp-4
difference: 2.7755575615628914e-17 0x1.0000000000000p-55
ulp : 2.0000
max.ulp : 1.0000
Maximal error of real part of: clog10
is : 3 ulp
accepted: 3 ulp
Maximal error of imaginary part of: clog10
is : 2 ulp
accepted: 1 ulp
This is actually resulting from atan2 becoming *more* accurate (atan2
(-0x6.d940dp-4, 0x1.7a858p+0) should ideally be -0x1.208cd6e841554p-2
but was -0x1.208cd6e841555p-2 from a powerpc libm built before this
change, and is -0x1.208cd6e841554p-2 from a powerpc libm built after
this change). Since these functions are not expected to be correctly
rounding by glibc's accuracy goals, neither result is a problem, but
this does imply that some of this code, although designed to be
correctly rounding, is not in fact correctly rounding (possibly
because of GCC creating fused operations where the code does not
expect it, something we've only disabled for specific functions where
it was found to cause large errors). (Of course as previously
discussed I think we should remove the slow cases where an error
analysis shows this wouldn't increase the errors much above 0.5ulp;
it's only functions such as cratan2 that are expected to be correctly
rounding, not atan2.)
* sysdeps/ieee754/dbl-64/dla.h [__FP_FAST_FMA] (DLA_FMS): Define
macro to use __builtin_fma.
* sysdeps/x86_64/fpu/dla.h: Remove file.
Joseph Myers [Fri, 30 Sep 2016 15:08:08 +0000 (15:08 +0000)]
Fix ldbl-128ibm iscanonical for -mlong-double-64.
This patch fixes the ldbl-128ibm version of the iscanonical macro not
to use __iscanonicall when long double = double (-mlong-double-64).
Tested for powerpc.
* sysdeps/ieee754/ldbl-128ibm/bits/iscanonical.h
[__NO_LONG_DOUBLE_MATH] (__iscanonicall): Do not declare.
[__NO_LONG_DOUBLE_MATH] (iscanonical): Define to evaluate to 1.
Carlos O'Donell [Fri, 30 Sep 2016 01:54:31 +0000 (21:54 -0400)]
Bug 20292 - Simplify and test _dl_addr_inside_object
The function _dl_addr_inside_object is simplified by removing
the conditional 'reladdr - l->l_phdr[n].p_vaddr >= 0' which is
always true. The function is refactored into it's own object file
and a unit test added to verify the correct behaviour of the
function.
Joseph Myers [Fri, 30 Sep 2016 00:24:19 +0000 (00:24 +0000)]
Add iscanonical.
TS 18661-1 adds an iscanonical classification macro to <math.h>.
The motivation for this is decimal floating-point, where some values
have both canonical and noncanonical encodings. For IEEE binary
interchange formats, all encodings are canonical. For x86/m68k
ldbl-96, and for ldbl-128ibm, there are encodings that do not
represent any valid value of the type; although formally iscanonical
does not need to handle trap representations (and so could just always
return 1), it seems useful, and in line with the description in the TS
of "representations that are extraneous to the floating-point model"
as being non-canonical (as well as "redundant representations of some
or all of its values"), for it to detect those representations and
return 0 for them.
This patch adds iscanonical to glibc. It goes in a header
<bits/iscanonical.h>, included under appropriate conditions in
<math.h>. The default header version just evaluates the argument
(converted to its semantic type, though current GCC will probably
discard that conversion and any exceptions resulting from it) and
returns 1. ldbl-96 and ldbl-128ibm then have versions of the header
that call a function __iscanonicall for long double (the sizeof-based
tests will of course need updating for float128 support, like other
such type-generic macro implementations). The ldbl-96 version of
__iscanonicall has appropriate conditionals to reflect the differences
in the m68k version of that format (where the high mantissa bit may be
either 0 or 1 when the exponent is 0 or 0x7fff). Corresponding tests
for those formats are added as well. Other architectures do not have
any new functions added because just returning 1 is correct for all
their floating-point formats.
Tested for x86_64, x86, mips64 (to test the default macro version) and
powerpc.
* math/math.h [__GLIBC_USE (IEC_60559_BFP_EXT)]: Include
<bits/iscanonical.h>.
* bits/iscanonical.h: New file.
* math/s_iscanonicall.c: Likewise.
* math/Versions (__iscanonicall): New libm symbol at version
GLIBC_2.25.
* math/libm-test.inc (iscanonical_test_data): New array.
(iscanonical_test): New function.
(main): Call iscanonical_test.
* math/Makefile (headers): Add bits/iscanonical.h.
(type-ldouble-routines): Add s_iscanonicall.
* manual/arith.texi (Floating Point Classes): Document
iscanonical.
* manual/libm-err-tab.pl: Update comment on interfaces without
ulps tabulated.
* sysdeps/ieee754/ldbl-128ibm/bits/iscanonical.h: New file.
* sysdeps/ieee754/ldbl-128ibm/s_iscanonicall.c: Likewise.
* sysdeps/ieee754/ldbl-128ibm/test-iscanonical-ldbl-128ibm.c:
Likewise.
* sysdeps/ieee754/ldbl-128ibm/Makefile (tests): Add
test-iscanonical-ldbl-128ibm.
* sysdeps/ieee754/ldbl-96/bits/iscanonical.h: New file.
* sysdeps/ieee754/ldbl-96/s_iscanonicall.c: Likewise.
* sysdeps/ieee754/ldbl-96/test-iscanonical-ldbl-96.c: Likewise.
* sysdeps/ieee754/ldbl-96/Makefile: Likewise.
* sysdeps/unix/sysv/linux/i386/libm.abilist: Update.
* sysdeps/unix/sysv/linux/ia64/libm.abilist: Likewise.
* sysdeps/unix/sysv/linux/m68k/m680x0/libm.abilist: Likewise.
* sysdeps/unix/sysv/linux/powerpc/powerpc32/fpu/libm.abilist:
Likewise.
* sysdeps/unix/sysv/linux/powerpc/powerpc32/nofpu/libm.abilist:
Likewise.
* sysdeps/unix/sysv/linux/powerpc/powerpc64/libm-le.abilist:
Likewise.
* sysdeps/unix/sysv/linux/powerpc/powerpc64/libm.abilist:
Likewise.
* sysdeps/unix/sysv/linux/x86_64/64/libm.abilist: Likewise.
* sysdeps/unix/sysv/linux/x86_64/x32/libm.abilist: Likewise.
Use copysign instead of ternary for some sin/cos input ranges
These are remaining cases where we can deduce and conclude that the
sign of the result should be the same as the sign of the input being
checked. For example, for sin(x), the sign of the result is the same
as the result itself for x < pi. Likewise, for sine values where x
after range reduction falls into this range and its sign is preserved.