Joseph Myers [Wed, 7 Sep 2016 16:40:09 +0000 (16:40 +0000)]
Add femode_t functions.
TS 18661-1 defines a type femode_t to represent the set of dynamic
floating-point control modes (such as the rounding mode and trap
enablement modes), and functions fegetmode and fesetmode to manipulate
those modes (without affecting other state such as the raised
exception flags) and a corresponding macro FE_DFL_MODE.
This patch series implements those interfaces for glibc. This first
patch adds the architecture-independent pieces, the x86 and x86_64
implementations, and the <bits/fenv.h> and ABI baseline updates for
all architectures so glibc keeps building and passing the ABI tests on
all architectures. Subsequent patches add the fegetmode and fesetmode
implementations for other architectures.
femode_t is generally an integer type - the same type as fenv_t, or as
the single element of fenv_t where fenv_t is a structure containing a
single integer (or the single relevant element, where it has elements
for both status and control registers) - except where architecture
properties or consistency with the fenv_t implementation indicate
otherwise. FE_DFL_MODE follows FE_DFL_ENV in whether it's a magic
pointer value (-1 cast to const femode_t *), a value that can be
distinguished from valid pointers by its high bits but otherwise
contains a representation of the desired register contents, or a
pointer to a constant variable (the powerpc case; __fe_dfl_mode is
added as an exported constant object, an alias to __fe_dfl_env).
Note that where architectures (that share a register between control
and status bits) gain definitions of new floating-point control or
status bits in future, the implementations of fesetmode for those
architectures may need updating (depending on whether the new bits are
control or status bits and what the implementation does with
previously unknown bits), just like existing implementations of
<fenv.h> functions that take care not to touch reserved bits may need
updating when the set of reserved bits changes. (As any new bits are
outside the scope of ISO C, that's just a quality-of-implementation
issue for supporting them, not a conformance issue.)
As with fenv_t, femode_t should properly include any software DFP
rounding mode (and for both fenv_t and femode_t I'd consider that
fragment of DFP support appropriate for inclusion in glibc even in the
absence of the rest of libdfp; hardware DFP rounding modes should
already be included if the definitions of which bits are status /
control bits are correct).
Tested for x86_64, x86, mips64 (hard float, and soft float to test the
fallback version), arm (hard float) and powerpc (hard float, soft
float and e500). Other architecture versions are untested.
* math/fegetmode.c: New file.
* math/fesetmode.c: Likewise.
* sysdeps/i386/fpu/fegetmode.c: Likewise.
* sysdeps/i386/fpu/fesetmode.c: Likewise.
* sysdeps/x86_64/fpu/fegetmode.c: Likewise.
* sysdeps/x86_64/fpu/fesetmode.c: Likewise.
* math/fenv.h: Update comment on inclusion of <bits/fenv.h>.
[__GLIBC_USE (IEC_60559_BFP_EXT)] (fegetmode): New function
declaration.
[__GLIBC_USE (IEC_60559_BFP_EXT)] (fesetmode): Likewise.
* bits/fenv.h [__GLIBC_USE (IEC_60559_BFP_EXT)] (femode_t): New
typedef.
[__GLIBC_USE (IEC_60559_BFP_EXT)] (FE_DFL_MODE): New macro.
* sysdeps/aarch64/bits/fenv.h [__GLIBC_USE (IEC_60559_BFP_EXT)]
(femode_t): New typedef.
[__GLIBC_USE (IEC_60559_BFP_EXT)] (FE_DFL_MODE): New macro.
* sysdeps/alpha/fpu/bits/fenv.h [__GLIBC_USE (IEC_60559_BFP_EXT)]
(femode_t): New typedef.
[__GLIBC_USE (IEC_60559_BFP_EXT)] (FE_DFL_MODE): New macro.
* sysdeps/arm/bits/fenv.h [__GLIBC_USE (IEC_60559_BFP_EXT)]
(femode_t): New typedef.
[__GLIBC_USE (IEC_60559_BFP_EXT)] (FE_DFL_MODE): New macro.
* sysdeps/hppa/fpu/bits/fenv.h [__GLIBC_USE (IEC_60559_BFP_EXT)]
(femode_t): New typedef.
[__GLIBC_USE (IEC_60559_BFP_EXT)] (FE_DFL_MODE): New macro.
* sysdeps/ia64/bits/fenv.h [__GLIBC_USE (IEC_60559_BFP_EXT)]
(femode_t): New typedef.
[__GLIBC_USE (IEC_60559_BFP_EXT)] (FE_DFL_MODE): New macro.
* sysdeps/m68k/fpu/bits/fenv.h [__GLIBC_USE (IEC_60559_BFP_EXT)]
(femode_t): New typedef.
[__GLIBC_USE (IEC_60559_BFP_EXT)] (FE_DFL_MODE): New macro.
* sysdeps/microblaze/bits/fenv.h [__GLIBC_USE (IEC_60559_BFP_EXT)]
(femode_t): New typedef.
[__GLIBC_USE (IEC_60559_BFP_EXT)] (FE_DFL_MODE): New macro.
* sysdeps/mips/bits/fenv.h [__GLIBC_USE (IEC_60559_BFP_EXT)]
(femode_t): New typedef.
[__GLIBC_USE (IEC_60559_BFP_EXT)] (FE_DFL_MODE): New macro.
* sysdeps/nios2/bits/fenv.h [__GLIBC_USE (IEC_60559_BFP_EXT)]
(femode_t): New typedef.
[__GLIBC_USE (IEC_60559_BFP_EXT)] (FE_DFL_MODE): New macro.
* sysdeps/powerpc/bits/fenv.h [__GLIBC_USE (IEC_60559_BFP_EXT)]
(femode_t): New typedef.
[__GLIBC_USE (IEC_60559_BFP_EXT)] (__fe_dfl_mode): New variable
declaration.
[__GLIBC_USE (IEC_60559_BFP_EXT)] (FE_DFL_MODE): New macro.
* sysdeps/s390/fpu/bits/fenv.h [__GLIBC_USE (IEC_60559_BFP_EXT)]
(femode_t): New typedef.
[__GLIBC_USE (IEC_60559_BFP_EXT)] (FE_DFL_MODE): New macro.
* sysdeps/sh/bits/fenv.h [__GLIBC_USE (IEC_60559_BFP_EXT)]
(femode_t): New typedef.
[__GLIBC_USE (IEC_60559_BFP_EXT)] (FE_DFL_MODE): New macro.
* sysdeps/sparc/fpu/bits/fenv.h [__GLIBC_USE (IEC_60559_BFP_EXT)]
(femode_t): New typedef.
[__GLIBC_USE (IEC_60559_BFP_EXT)] (FE_DFL_MODE): New macro.
* sysdeps/tile/bits/fenv.h [__GLIBC_USE (IEC_60559_BFP_EXT)]
(femode_t): New typedef.
[__GLIBC_USE (IEC_60559_BFP_EXT)] (FE_DFL_MODE): New macro.
* sysdeps/x86/fpu/bits/fenv.h [__GLIBC_USE (IEC_60559_BFP_EXT)]
(femode_t): New typedef.
[__GLIBC_USE (IEC_60559_BFP_EXT)] (FE_DFL_MODE): New macro.
* manual/arith.texi (FE_DFL_MODE): Document macro.
(fegetmode): Document function.
(fesetmode): Likewise.
* math/Versions (fegetmode): New libm symbol at version
GLIBC_2.25.
(fesetmode): Likewise.
* math/Makefile (libm-support): Add fegetmode and fesetmode.
(tests): Add test-femode and test-femode-traps.
* math/test-femode-traps.c: New file.
* math/test-femode.c: Likewise.
* sysdeps/powerpc/fpu/fenv_const.c (__fe_dfl_mode): Declare as
alias for __fe_dfl_env.
* sysdeps/powerpc/nofpu/fenv_const.c (__fe_dfl_mode): Likewise.
* sysdeps/powerpc/powerpc32/e500/nofpu/fenv_const.c
(__fe_dfl_mode): Likewise.
* sysdeps/powerpc/Versions (__fe_dfl_mode): New libm symbol at
version GLIBC_2.25.
* 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.
There is transition penalty when SSE instructions are mixed with 256-bit
AVX or 512-bit AVX512 load instructions. Since _dl_runtime_resolve_avx
and _dl_runtime_profile_avx512 save/restore 256-bit YMM/512-bit ZMM
registers, there is transition penalty when SSE instructions are used
with lazy binding on AVX and AVX512 processors.
To avoid SSE transition penalty, if only the lower 128 bits of the first
8 vector registers are non-zero, we can preserve %xmm0 - %xmm7 registers
with the zero upper bits.
For AVX and AVX512 processors which support XGETBV with ECX == 1, we can
use XGETBV with ECX == 1 to check if the upper 128 bits of YMM registers
or the upper 256 bits of ZMM registers are zero. We can restore only the
non-zero portion of vector registers with AVX/AVX512 load instructions
which will zero-extend upper bits of vector registers.
This patch adds _dl_runtime_resolve_sse_vex which saves and restores
XMM registers with 128-bit AVX store/load instructions. It is used to
preserve YMM/ZMM registers when only the lower 128 bits are non-zero.
_dl_runtime_resolve_avx_opt and _dl_runtime_resolve_avx512_opt are added
and used on AVX/AVX512 processors supporting XGETBV with ECX == 1 so
that we store and load only the non-zero portion of vector registers.
This avoids SSE transition penalty caused by _dl_runtime_resolve_avx and
_dl_runtime_profile_avx512 when only the lower 128 bits of vector
registers are used.
_dl_runtime_resolve_avx_slow is added and used for AVX processors which
don't support XGETBV with ECX == 1. Since there is no SSE transition
penalty on AVX512 processors which don't support XGETBV with ECX == 1,
_dl_runtime_resolve_avx512_slow isn't provided.
[BZ #20495]
[BZ #20508]
* sysdeps/x86/cpu-features.c (init_cpu_features): For Intel
processors, set Use_dl_runtime_resolve_slow and set
Use_dl_runtime_resolve_opt if XGETBV suports ECX == 1.
* sysdeps/x86/cpu-features.h (bit_arch_Use_dl_runtime_resolve_opt):
New.
(bit_arch_Use_dl_runtime_resolve_slow): Likewise.
(index_arch_Use_dl_runtime_resolve_opt): Likewise.
(index_arch_Use_dl_runtime_resolve_slow): Likewise.
* sysdeps/x86_64/dl-machine.h (elf_machine_runtime_setup): Use
_dl_runtime_resolve_avx512_opt and _dl_runtime_resolve_avx_opt
if Use_dl_runtime_resolve_opt is set. Use
_dl_runtime_resolve_slow if Use_dl_runtime_resolve_slow is set.
* sysdeps/x86_64/dl-trampoline.S: Include <cpu-features.h>.
(_dl_runtime_resolve_opt): New. Defined for AVX and AVX512.
(_dl_runtime_resolve): Add one for _dl_runtime_resolve_sse_vex.
* sysdeps/x86_64/dl-trampoline.h (_dl_runtime_resolve_avx_slow):
New.
(_dl_runtime_resolve_opt): Likewise.
(_dl_runtime_profile): Define only if _dl_runtime_profile is
defined.
Stefan Liebler [Tue, 6 Sep 2016 12:44:15 +0000 (14:44 +0200)]
S390: Support PLT and GOT references in check-localplt.
on s390x the test elf/check-localplt is failing after recent commits:
"elf: Do not use memalign for TCB/TLS blocks allocation [BZ #17730]"
"elf: Avoid using memalign for TLS allocations [BZ #17730]"
"elf: dl-minimal malloc needs to respect fundamental alignment"
due to "Missing required PLT reference: ld.so: __libc_memalign".
After the commits __libc_memalign is only called in elf/dl-minimal.c in
malloc() function in ld.so and gcc -O2/-O3 leads to R_390_GLOB_DAT
instead of R_390_JMP_SLOT. __libc_memalign is called via
function-pointer loaded from GOT instead of calling via a plt-stub. In
this case there is the R_390_GLOB_DAT relocation in section .rela.dyn
instead of R_390_JMP_SLOT in .rela.plt.
This patch marks ld.so: __libc_memalign with R_390_GLOB_DAT in
localplt.data to allow both relocations.
If build with -fno-optimize-sibling-calls or on s390(31bit) a
R_390_JMP_SLOT is generated.
ChangeLog:
* sysdeps/unix/sysv/linux/s390/localplt.data: Mark
ld.so: __libc_memalign with "+ RELA R_390_GLOB_DAT".
Historically perl includes the current directory in the module search
path. Over the time this has been considered as a security issue and
the recent vulnerabilities [1] made people to reconsider this behaviour.
It is almost sure that this will be removed in the future [2], possibly
for the 5.26 release, although this is not yet firmly decided.
Debian has decided to backport the patches [3], so the perl binary in
unstable do not have '.' in @INC anymore.
This behaviour is used in the conform perl scripts to include the
GlibcConform module. This patch fixes that by calling perl with '-I.'.
This is not a security issue in this case as make ensures that the
current directory is $(srcdir)/conform/ when the scripts are called.
Passing the full path would do exactly the same.
The commit b632bdd3 moved the setting of the DF_1_NODELETE flag earlier
in the dl_open_worker function. However when calling dlopen with both
RTLD_NODELETE and RTLD_NOLOAD, the pointer returned by _dl_map_object is
NULL. This condition is checked just after setting the flag, while it
should be done before. Fix that.
Changelog:
[BZ #19810]
* elf/dl-open.c (dl_open_worker): Set DF_1_NODELETE flag later.
* elf/tst-noload.c: New test case.
* elf/Makefile (tests): Add tst-noload.
The support functions for sin and cos have a lot of identical
functionality, so inlining them gives a pretty decent jump in
functionality: ~19% in the sincos function. On SPEC2006 this
translates to about 2.1% in the tonto test.
The only code looks slightly different from do_sin but on closer
examination, should give exactly the same result. Drop it in favour
of the do_sin function call.
* sysdeps/ieee754/dbl-64/s_sin.c (__sin): Use do_sin.
Consolidate input partitioning into do_cos and do_sin
All calls to do_cos are preceded by code that partitions x into a
larger double that gives an offset into the sincos table and a smaller
double that is used in a polynomial computation. Consolidate all of
them into do_cos and do_sin to reduce code duplication.
* sysdeps/ieee754/dbl-64/s_sin.c (do_cos): Accept X and DX as input
arguments. Consolidate input partitioning from callers here.
(do_cos_slow): Likewise.
(do_sin): Likewise.
(do_sin_slow): Likewise.
(do_sincos_1): Remove the no longer necessary input partitioning.
(do_sincos_2): Likewise.
(__sin): Likewise.
(__cos): Likewise.
(slow1): Likewise.
(slow2): Likewise.
(sloww1): Likewise.
(sloww2): Likewise.
(bsloww1): Likewise.
(bsloww2): Likewise.
(cslow2): Likewise.
vfscanf: Avoid multiple reads of multi-byte character width
This avoids a race condition if the process-global locale is changed
while vfscanf is running. MB_LEN_MAX is always larger than MB_CUR_MAX,
so we might realloc earlier than necessary (but even MB_CUR_MAX could
be larger than the minimum required space).
The existing length was a bit questionable because str + MB_LEN_MAX
might point past the end of the buffer.
Paul E. Murphy [Wed, 24 Aug 2016 21:05:28 +0000 (16:05 -0500)]
Make common nextdown implementation generic.
With the exception of those machines using the ldbl-opt in
an Implies file, this is a trivial transformation.
nextdownl is not subject to the non-trivial versioning rules
of the other generated functions, so to keep things simple,
it is handled as a one-off case in ldbl-opt to preserve the
existing behavior.
Paul E. Murphy [Thu, 25 Aug 2016 16:25:33 +0000 (11:25 -0500)]
Make common fdim implementation generic.
The only difference is the usage of math_narrow_eval when
building s_fdiml.c. This should be harmless for long double,
but I did observe some code generation changes on m68k, but
lack the resources to test it.
Likewise, to more easily support overriding symbol generation,
the aliasing macros are always conditionally defined on their
absence to reduce boilerplate.
I also ran builds for i486, ppc64, sparcv9, aarch64,
s390x and observed no changes to s_fdim* objects.
Paul E. Murphy [Wed, 20 Jul 2016 20:20:51 +0000 (15:20 -0500)]
ldbl-128: Rename 'long double' to '_Float128'
Add a layer of macro indirection for long double files
which need to be built using another typename. Likewise,
add the L(num) macro used in a later patch to override
real constants.
These macros are only defined through the ldbl-128
math_ldbl.h header, thereby implicitly restricting
these macros to machines which back long double
with an IEEE binary128 format.
Likewise, appropriate changes are made for the few
files which indirectly include such ldbl-128 files.
These changes produce identical binaries for s390x,
aarch64, and ppc64.
Stefan Liebler [Wed, 31 Aug 2016 12:54:55 +0000 (14:54 +0200)]
S390: Do not set FE_INEXACT with feraiseexcept (FE_OWERFLOW|FE_UNDERFLOW).
On s390 feraiseexcept (FE_OVERFLOW|FE_UNDERFLOW) sets FE_INEXACT, too.
This patch uses z196 zarch load rounded instruction which can suppress
FE_INEXACT exception if gcc has z196 support in used configuration.
Otherwise FE_INEXACT flag is set as before. The gcc support is tested
in a new configure-check.
A comment in fsetexcptflg.c is corrected as new exceptions are not
executed with the next floating-point instruction if fpc is set with
_FPU_SETCW macro. It seems the comment was copied e.g. from
sysdeps/x86_64/fpu/fsetexcptflg.c file.
ChangeLog:
* config.h.in (HAVE_S390_MIN_Z196_ZARCH_ASM_SUPPORT):
New undefine.
* sysdeps/s390/configure.ac: Add test for z196 zarch support.
* sysdeps/s390/configure: Regenerated.
* sysdeps/s390/fpu/fraiseexcpt.c (__feraiseexcept): Use ledbra
instruction for raising over-/underflow if z196 zarch is supported
by default.
* sysdeps/s390/fpu/fsetexcptflg.c (fesetexceptflag):
Correct comment.
Use fabs(x) instead of branching on signedness of input to sin and cos
The sin and cos code is inconsistent about its use of fabs to get the
absolute value of X where in some places it conditionalizes the code
while in others it uses fabs. fabs seems to be a better candidate in
most cases because it avoids a branch. Similarly there is an attempt
to make it easier for the compiler to emit conditional assignment
instructions (like fcsel on aarch64) where it can, by isolating
conditional assignment constructs from the rest of the expression.
A further benefit of this change is to identify common constructs
across functions and consolidate them in future patches.
* sysdeps/ieee754/dbl-64/s_sin.c (do_cos_slow): Use ternary
instead of if/else.
(do_sin_slow): Likewise.
(do_sincos_1): Use fabs instead of if/else.
(do_sincos_2): Likewise.
(__sin): Likewise.
(__cos): Likewise.
(slow2): Likewise.
(sloww): Likewise.
(sloww1): Likewise. Drop argument M.
(sloww2): Use fabs instead of if/else.
(bsloww): Likewise.
(bsloww1): Likewise.
(bsloww2): Likewise.
This patch reshuffles the reduce_and_compute code so that the
structure matches other code structures of the same type elsewhere in
s_sin.c and s_sincos.c. This is the beginning of an attempt to
consolidate and reduce code duplication in functions in s_sin.c to
make it easier to read and possibly also easier for the compiler to
optimize.
* sysdeps/ieee754/dbl-64/s_sin.c (reduce_and_compute):
Consolidate switch cases 0 and 2.
Paul E. Murphy [Tue, 28 Jun 2016 19:28:04 +0000 (14:28 -0500)]
Convert remaining complex function to generated files
Convert cpow, clog, clog10, cexp, csqrt, and cproj functions
into generated templates. Note, ldbl-opt still retains
s_clog10l.c as the aliasing rules are non-trivial.
Paul E. Murphy [Fri, 1 Jul 2016 16:03:51 +0000 (11:03 -0500)]
Prepare to convert remaining _Complex functions
This patch has no function changes, except to
ensure the git history correctly tracks the
changes to convert the double version of these
functions into a templated version.
Joseph Myers [Mon, 29 Aug 2016 11:47:21 +0000 (11:47 +0000)]
Add fetestexceptflag.
TS 18661-1 defines an fetestexceptflag function to test the exception
state saved in an fexcept_t object by fegetexceptflag.
This patch implements this function for glibc. Almost all
architectures save exception state in such a way that it can be
directly ANDed with exception flag bits, so rather than having lots of
fetestexceptflag implementations that all do the same thing, the math/
implementation is made to use this generic logic (which is also OK in
the fallback case where FE_ALL_EXCEPT is zero). The only architecture
that seems to need anything different is s390.
(fegetexceptflag and fesetexceptflag use abbreviated filenames
fgetexcptflg.c and fsetexcptflg.c. Because we are no longer concerned
by 14-character filename limits, fetestexceptflag uses the obvious
filename fetestexceptflag.c.)
The NEWS entry is intended to be expanded along the lines given in
<https://sourceware.org/ml/libc-alpha/2016-08/msg00356.html> when
fegetmode and fesetmode are added.
Existing interposed mallocs do not define the glibc-internal
fork callbacks (and they should not), so statically interposed
mallocs lead to link failures because the strong reference from
fork pulls in glibc's malloc, resulting in multiple definitions
of malloc-related symbols.
ISO C forbids empty initializer braces (6.7.9 initializer-list must
contain at least one initializer). However GCC allows it, generating
a warning depending of the version.
With GCC 4.8 on ARM I noticed tst-initializers1.c fails to build with:
In file included from tst-initializers1.c:60:0:
../test-skeleton.c: In function 'delayed_exit_thread':
../test-skeleton.c:687:10: error: missing initializer for field 'tv_sec' of 'struct timespec' [-Werror=missing-field-initializers]
struct timespec remaining = {}
While with GCC 5.1 the same warning is just spilled with -pedantic.
To be safe this patch just zero initialize the struct as expected.
Florian Weimer [Fri, 26 Aug 2016 17:27:16 +0000 (19:27 +0200)]
nptl: Avoid expected SIGALRM in most tests [BZ #20432]
Before this change, several tests did not detect early deadlocks
because they used SIGALRM as the expected signal, and they ran
for the full default TIMEOUT seconds.
This commit adds a new delayed_exit function to the test skeleton,
along with several error-checking wrappers to pthread functions.
Additional error checking is introduced into several tests.
Mark Wielaard [Thu, 25 Aug 2016 21:47:19 +0000 (23:47 +0200)]
Reduce memory size of tsearch red-black tree.
A tsearch red-black tree node contains 3 pointers (key, left, right)
and 1 bit to hold the red-black flag. When allocating new nodes
this 1 bit is expanded to a full word. Causing the overhead per node
to be 3 times the key size.
We can reduce this overhead to just 2 times the key size.
malloc returns naturally aligned memory. All nodes are internally
allocated with malloc and the left/right node pointers are used
as implementation details. So we can use the low bits of the
left/right node pointers to store extra information.
Replace all direct accesses of the struct node_t node pointers and
red-black value with defines that take care of the red-black flag in
the low bit of the (left) node pointer. This reduces the size of the
nodes on 32-bit systems from 16 to 12 bytes and on 64-bit systems
from 32 to 24 bytes.
Also fix a call to CHECK_TREE so the code can be build (and tested)
with DEBUGGING defined again.
V2 changes:
- Add assert after malloc to catch any odd pointers from bad
interposed mallocs.
- Rename implementation flag to USE_MALLOC_LOW_BIT.
ChangeLog:
* misc/tsearch.c (struct node_t): Reduce to 3 pointers if
USE_MALLOC_LOW_BIT. Define pointer/value accessors.
(check_tree_recurse): Use newly defined accessors.
(check_tree): Likewise.
(maybe_split_for_insert): Likewise.
(__tfind): Likewise.
(__tdelete): Likewise.
(trecurse): Likewise.
(tdestroy_recurse): Likewise.
(__tsearch): Likewise. And add asserts for malloc alignment.
(__twalk): Cast root to node in case CHECK_TREE is defined.
Ernestas Kulik [Sun, 21 Aug 2016 11:39:36 +0000 (14:39 +0300)]
localedata: lt_LT: use hyphens in d_fmt [BZ #20497]
The standard currently in effect (LST ISO 8601:1997) mandates the use
of hyphens (as opposed to full stops, currently) in date formats. It
also matches current CLDR data (v29), Wikipedia's & Wikia's settings,
and Microsoft's Lithuanian Style Guide.
Paul E. Murphy [Fri, 1 Jul 2016 16:00:21 +0000 (11:00 -0500)]
Prepare to convert _Complex tangent functions
This patch has no function changes, except to
ensure the git history correctly tracks the
changes to convert the double version of these
functions into a templated version.
Paul E. Murphy [Fri, 1 Jul 2016 15:55:27 +0000 (10:55 -0500)]
Prepare to convert _Complex sine functions
This patch has no function changes, except to
ensure the git history correctly tracks the
changes to convert the double version of these
functions into a templated version.
Paul E. Murphy [Mon, 8 Aug 2016 20:58:28 +0000 (15:58 -0500)]
Merge common usage of mul_split function
A number of files share identical code for the
mul_split function.
This moves the duplicated function mul_split into its
own header, and refactors the fma usage into a single
selection macro. Likewise, mul_split when used by a
long double implementation is renamed mul_splitl for
clarity.
Paul E. Murphy [Fri, 1 Jul 2016 15:49:03 +0000 (10:49 -0500)]
Prepare to convert _Complex cosine functions
This patch has no function changes, except to
ensure the git history correctly tracks the
changes to convert the double version of these
functions into a templated version.
Paul E. Murphy [Thu, 28 Jul 2016 16:14:11 +0000 (11:14 -0500)]
Add tst-wcstod-round
This extends tst-strtod-round with a few trivial changes
to also test the wide character variants of strto* using
similar macros to other shared tests.
Torvald Riegel [Wed, 17 Aug 2016 11:56:11 +0000 (13:56 +0200)]
Fix incorrect double-checked locking related to _res_hconf.initialized.
_res_hconf.initialized was not suitable for use in a multi-threaded
environment due to the lack of atomics and memory barriers. Use of it was
also unnecessary because _res_hconf_init did the right thing by using
__libc_once. This patch fixes the glibc-internal uses by just calling
_res_hconf_init unconditionally, and switches to a release MO atomic store
for _res_hconf.initialized to fix the glibc side of the synchronization
problem (which will maintain backward compatibility, but cannot fix the
lack of acquire MO on any glibc-external loads).
Stefan Liebler [Thu, 18 Aug 2016 10:20:35 +0000 (12:20 +0200)]
Get rid of array-bounds warning in __kernel_rem_pio2[f] with gcc 6.1 -O3.
On s390x I get the following werror when build with gcc 6.1 (or current gcc head) and -O3:
../sysdeps/ieee754/dbl-64/k_rem_pio2.c: In function ‘__kernel_rem_pio2’:
../sysdeps/ieee754/dbl-64/k_rem_pio2.c:254:18: error: array subscript is below array bounds [-Werror=array-bounds]
for (k = 1; iq[jk - k] == 0; k++)
~~^~~~~~~~
I get the same error with sysdeps/ieee754/flt-32/k_rem_pio2f.c.
This patch adds DIAG_* macros around it.
ChangeLog:
* sysdeps/ieee754/dbl-64/k_rem_pio2.c (__kernel_rem_pio2):
Use DIAG_*_NEEDS_COMMENT macro to get rid of array-bounds warning.
* sysdeps/ieee754/flt-32/k_rem_pio2f.c (__kernel_rem_pio2f):
Likewise.
Florian Weimer [Thu, 18 Aug 2016 09:15:42 +0000 (11:15 +0200)]
argp: Do not override GCC keywords with macros [BZ #16907]
glibc provides fallback definitions already. It is not necessary to
suppress warnings for unknown attributes because GCC does this
automatically for system headers.
This commit does not sync with gnulib because gnulib has started to use
_GL_* macros in the header file, which are arguably in the gnulib
implementation space and not suitable for an installed glibc header
file.
Paul E. Murphy [Tue, 14 Jun 2016 16:44:14 +0000 (11:44 -0500)]
Support for type-generic libm function implementations libm
This defines a new classes of libm objects. The
<func>_template.c file which is used in conjunction
with the new makefile hooks to derive variants for
each type supported by the target machine.
The headers math-type-macros-TYPE.h are used to supply
macros to a common implementation of a function in
a file named FUNC_template.c and glued togethor via
a generated file matching existing naming in the
build directory.
This has the properties of preserving the existing
override mechanism and not requiring any arcane
build system twiddling. Likewise, it enables machines
to override these files without any additional work.
I have verified the built objects for ppc64, x86_64,
alpha, arm, and m68k do not change in any meaningful
way with these changes using the Fedora cross toolchains.
I have verified the x86_64 and ppc64 changes still run.
Joseph Myers [Tue, 16 Aug 2016 17:11:46 +0000 (17:11 +0000)]
Fix soft-fp extended.h unpacking (GCC bug 77265).
soft-fp unpacking for x86 "extended" fails to clear the implicit
mantissa high bit that is explicit in that format, resulting in
problems for operations that expect this bit to be clear in raw
unpacked values. Specifically, the code for this format is used only
for conversions to and from TFmode (__float128) in libgcc, where this
issue results in GCC bug 77265, extension of long double infinity to
__float128 wrongly produces a NaN.
This patch fixes this by always masking out the implicit bit on
unpacking, so that the results of unpacking meet the expectations of
the rest of the soft-fp code for a normal IEEE format.
Tested for x86_64 in libgcc in conjunction with a GCC testcase for
this issue (this code isn't used in glibc, only in libgcc).
* soft-fp/extended.h [_FP_W_TYPE_SIZE < 64] (FP_UNPACK_RAW_E):
Mask implicit bit out of unpacked value.
[_FP_W_TYPE_SIZE < 64] (FP_UNPACK_RAW_EP): Likewise.
[_FP_W_TYPE_SIZE >= 64] (FP_UNPACK_RAW_E): Likewise.
[_FP_W_TYPE_SIZE >= 64] (FP_UNPACK_RAW_EP): Likewise.
Joseph Myers [Tue, 16 Aug 2016 16:16:10 +0000 (16:16 +0000)]
Add fesetexcept.
TS 18661-1 defines an fesetexcept function for setting floating-point
exception flags without the side-effect of causing enabled traps to be
taken.
This patch series implements this function for glibc. The present
patch adds the fallback stub implementation, x86 and x86_64
implementations, documentation, tests and ABI baseline updates. The
remaining patches, some of them untested, add implementations for
other architectures. The implementations generally follow those of
the fesetexceptflag function.
As for fesetexceptflag, the approach taken for architectures where
setting flags causes enabled traps to be taken is to set the flags
(and potentially cause traps) rather than refusing to set the flags
and returning an error. Since ISO C and TS 18661 provide no way to
enable traps, this is formally in accordance with the standards.
The NEWS entry should be considered a placeholder, since this patch
series is intended to be followed by further such series adding other
TS 18661-1 features, so that the NEWS entry would end up looking more
like
* New <fenv.h> features from TS 18661-1:2014 are added to libm: the
fesetexcept, fetestexceptflag, fegetmode and fesetmode functions,
the femode_t type and the FE_DFL_MODE macro.
with hopefully more such entries for other features, rather than
having an entry for a single function in the end.
I believe we have consensus for adding TS 18661-1 interfaces as per
<https://sourceware.org/ml/libc-alpha/2016-06/msg00421.html>.
Tested for x86_64, x86, mips64 (hard float, and soft float to test the
fallback version), arm (hard float) and powerpc (hard float, soft
float and e500).
Florian Weimer [Tue, 16 Aug 2016 09:06:13 +0000 (11:06 +0200)]
nptl/tst-tls3-malloc: Force freeing of thread stacks
It turns out that due to the reduced stack size in tst-tls3 and the
(fixed) default stack cache size, allocated TLS variables are never
freed, so the test coverage for tst-tls3-malloc is less than complete.
This change increases the thread stack size for tst-tls3-malloc only,
to make sure thread stacks and TLS variables are freed.
Märt Põder [Mon, 15 Aug 2016 09:46:21 +0000 (11:46 +0200)]
locales: et_EE: locale has wrong {p,n}_cs_precedes value [BZ #20459]
According to "Requirements of information technology in Estonian
language and cultural environment" the monetary symbol should be
written after the amount number:
Joseph Myers [Fri, 12 Aug 2016 17:34:01 +0000 (17:34 +0000)]
Fix test-fexcept when "inexact" implicitly raised.
ISO C allows feraiseexcept to raise "inexact", in addition to the
requested exceptions, when requested to raise "overflow" or
"underflow". Testing on ARM and PowerPC e500 (where glibc's
feraiseexcept has this property) showed that the new test-fexcept test
failed to allow for this; this patch fixes it, by wrapping
feraiseexcept to clear FE_INEXACT if implicitly raised and not raised
before the call. (It would also be possible to do this with
fesetexcept, which always affects exactly the requested flags, but
this patch avoids making this fix depend on the fesetexcept changes.)
Tested for x86_64, x86, arm and e500.
* math/test-fexcept.c (feraiseexcept_exact): New function.
(test_set): Call feraiseexcept_exact instead of feraiseexcept.
(test_except): Likewise.
As shown by the test math/test-fexcept, the powerpc fesetexceptflag
implementation fails to clear a previously set FE_INVALID flag, when
that flag is clear in the saved exceptions and FE_INVALID is included
in the mask of flags to restore, because it fails to mask out the
sub-exceptions of FE_INVALID from the FPSCR state. This patch fixes
the masking logic accordingly.
Tested for powerpc.
[BZ #20455]
* sysdeps/powerpc/fpu/fsetexcptflg.c (__fesetexceptflag): Mask out
all FE_INVALID sub-exceptions from FPSCR when FE_INVALID specified
to be restored.
Joseph Myers [Wed, 10 Aug 2016 21:01:08 +0000 (21:01 +0000)]
Add tests for fegetexceptflag, fesetexceptflag.
I noticed that there was no meaningful test coverage for
fegetexceptflag and fesetexceptflag (one test ensures that calls to
them compile and link, but nothing to verify they work correctly).
This patch adds tests for these functions.
fesetexceptflag is meant to set the relevant exception flag bits to
the saved state without causing enabled traps to be taken. On some
architectures, it is not possible to set exception flag bits without
causing enabled traps to occur. Such architectures need to define
EXCEPTION_SET_FORCES_TRAP to 1 in their math-tests.h, as is done in
this patch for powerpc. x86 avoids needing to define this because the
traps resulting from setting exception bits don't occur until the next
floating-point operation or fwait instruction.
Tested for x86_64, x86 and powerpc. Note that test-fexcept fails for
powerpc because of a pre-existing bug in fesetexceptflag for powerpc,
which I'll fix separately.
* math/test-fexcept-traps.c: New file.
* math/test-fexcept.c: Likewise.
* math/Makefile (tests): Add test-fexcept and test-fexcept-traps.
* sysdeps/generic/math-tests.h (EXCEPTION_SET_FORCES_TRAP): New
macro.
* sysdeps/powerpc/math-tests.h [!__NO_FPRS__]
(EXCEPTION_SET_FORCES_TRAP): Likewise.
Martin Pitt [Fri, 14 Aug 2015 05:22:37 +0000 (01:22 -0400)]
locales: en_CA: update d_fmt [BZ #9842]
The date format in en_CA/LC_TIME specifies the date format as "%d/%m/%y".
However, it should be "%Y-%m-%d". This is the standard date format in
Canada as specified by the Canadian Standards Association in CSA Z234.5:1989,
which adopts the ISO 8601 standard.
Here's the web page from the National Research Council of Canada
citing ISO 8601 as the standard date/time format in Canada:
http://www.nrc-cnrc.gc.ca/eng/services/time/faq/#Q8
International Standard ISO 8601 specifies numeric representations
of date and time. The recommended full format is of the form
2001-12-31 23:59:28.73 UTC. The intent of this standard is to avoid
confusion in international communications which can arise with the
many different national notations. This format has the advantage
that it permits dates to be readily sorted in chronological order
by computer systems.
Aurelien Jarno [Fri, 5 Aug 2016 20:35:01 +0000 (22:35 +0200)]
sparc32/sparcv9: add a VIS3 version of fdim
sparc32 passes floating point values in the integer registers. VIS3
instructions gives access to the movwtos instruction to directly
transfer a value from an integer register to a floating point register.
Therefore it makes sense to provide a VIS3 version consisting in the
generic version compiled with -mvis3.
Changelog:
* math/s_fdim.c: Avoid alias renamed.
* math/s_fdimf.c: Likewise.
* sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/Makefile
[$(subdir) = math && $(have-as-vis3) = yes] (libm-sysdep_routines):
Add s_fdimf-vis3, s_fdim-vis3.
(CFLAGS-s_fdimf-vis3.c): New. Set to -Wa,-Av9d -mvis3.
(CFLAGS-s_fdim-vis3.c): Likewise.
sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_fdim-vis3.c: New file.
sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_fdim.c: Likewise.
Aurelien Jarno [Fri, 5 Aug 2016 20:35:01 +0000 (22:35 +0200)]
sparc: remove fdim sparc specific implementations
The fdim and fdimf functions on sparc do not fully follow the standard
and do not set errno to ERANGE when the result overflows. Since glibc
2.24 this causes the two following tests to fail:
Failure: fdim (max_value, -max_value): errno set to 0, expected 34 (ERANGE)
Failure: fdim_upward (max_value, -max_value): errno set to 0, expected 34 (ERANGE)
It happens that using GCC with the generic C code generates very similar
code to the sparc specific implementations. Therefore this patches
remove them. Note it might still worth adding a vis3 specific version of
fdim on sparc32/sparcv9, this is done in a following patch to ease
backporting.
Aurelien Jarno [Fri, 5 Aug 2016 20:35:01 +0000 (22:35 +0200)]
sparc: build with -mvis on sparc32/sparcv9 and sparc64
When building for sparc32/sparcv9 or sparc64, we assume that VIS
instructions are available and use them in the sparc specific assembly
code. However we do not tell GCC to use such instructions, resulting in
slightly suboptimal code.
Paul E. Murphy [Fri, 15 Jul 2016 15:57:07 +0000 (10:57 -0500)]
Improve gen-libm-test.pl LIT() application
When bootstrapping float128, this exposed a number of areas where
the L suffix is incorrectly applied to simple expressions when it
should be applied to each constant in the expression.
In order to stave off more macros in libm-test.inc, apply_lit is
made slightly more intelligent. It will now split expressions
based on space characters, and attempt to apply LIT() to each
token.
Having done this, there are numerous spacing issues within
libm-test.inc which have been fixed.
The above is problematic when the L real suffix is not the most
expressive modifier, and the compiler complains (i.e ppc64) or
silently truncates a value (i.e ppc64).
Joseph Myers [Fri, 5 Aug 2016 18:15:00 +0000 (18:15 +0000)]
Fix math.h comment about bits/mathdef.h.
math.h has a comment about definitions from <bits/mathdef.h>. This
comment is in the wrong place in math.h, far below the inclusion of
<bits/mathdef.h>. It was originally above the inclusion, but the
inclusion was moved by
1998-11-05 Ulrich Drepper <drepper@cygnus.com>
* math/math.h: Unconditionally include bits/mathdef.h. Declare
long double functions only if __NO_LONG_DOUBLE_MATH is not
defined.
[...]
without moving the comment. Furthermore, the comment refers
incorrectly to FLT_EVAL_METHOD and DECIMAL_DIG, which are actually
<float.h> macros, and INFINITY, which is in <bits/inf.h>.
This patch moves the comment back above the include it refers to and
removes the description of macros not defined by the header.
Tested for x86_64 and x86 (testsuite, and that installed stripped
shared libraries are unchanged by the patch).
* math/math.h: Move comment about <bits/mathdef.h> definitions
above inclusion of <bits/mathdef.h>. Do not mention
FLT_EVAL_METHOD, INFINITY or DECIMAL_DIG in that comment.
Joseph Myers [Thu, 4 Aug 2016 20:50:31 +0000 (20:50 +0000)]
Do not call __nan in scalb functions.
When libm functions return a NaN: if it is for NaN input, it should be
computed from that input (e.g. adding it to itself), so that payloads
are propagated and signaling NaNs quieted, while if it is for non-NaN
input, it should be produced by a computation such as
(x - x) / (x - x), which raises "invalid" at the same time as
producing an appropriate NaN, so avoiding any need for a call to
feraiseexcept.
Various libm functions, however, call __nan ("") (or __nanf or __nanl)
to determine the NaN to return, together with using feraiseexcept
(FE_INVALID) to raise the exception. sysdeps/generic/math_private.h
has an optimization for those functions with constant "" argument so
this doesn't actually involve a call to the __nan function, but it is
still not the preferred approach for producing NaNs. (The optimized
code also always uses the NAN macro, i.e. produces a default NaN for
float converted to whatever the target type is, and on some
architectures that may not be the same as the preferred default NaN
for double or long double.)
This patch fixes the scalb functions to use the conventional method of
generating NaNs and raising "invalid" with an appropriate
computation. (Most instances of this issue are in the complex
functions, where it can more readily be fixed once they have been made
type-generic and so only a third as many places need fixing. Some of
the complex functions use __nan ("") + __nan (""), where the addition
serves no purpose whatsoever.)
Tested for x86_64 and x86.
* math/e_scalb.c: Do not include <fenv.h>.
(invalid_fn): Do calculation resulting in NaN instead of raising
FE_INVALID and returning a NaN explicitly.
* math/e_scalbf.c: Do not include <fenv.h>.
(invalid_fn): Do calculation resulting in NaN instead of raising
FE_INVALID and returning a NaN explicitly.
* math/e_scalbl.c: Do not include <fenv.h>.
(invalid_fn): Do calculation resulting in NaN instead of raising
FE_INVALID and returning a NaN explicitly.
Florian Weimer [Thu, 4 Aug 2016 09:10:57 +0000 (11:10 +0200)]
x86: Use sysdep.o from libc.a in static libraries
Static libraries can use the sysdep.o copy in libc.a without
a performance penalty. This results in a visible difference
if libpthread.a is relinked into a single object file (which
is needed to support libraries which check for the presence
of certain symbols to enable threading support, which generally
fails with static linking unless libpthread.a is relinked).
git://sourceware.org / glibc.git / log