The dbl-64 and flt-32 implementations of exp2 functions produce
spurious underflow exceptions. The underlying reason is the same in
both cases: the computation works as (2^a - 1)*2^b + 2^b for suitably
chosen a and b, where a has small magnitude so 2^a - 1 can be computed
with a low-degree polynomial approximation, and (2^a - 1)*2^b can
underflow even when the final result does not. This patch fixes this
by adjusting the threshold for when scaling is used to avoid
intermediate underflow so it works for any possible value of a where
the final result would not underflow.
Tested for x86_64 and x86.
[BZ #18219]
* sysdeps/ieee754/dbl-64/e_exp2.c (__ieee754_exp2): Reduce
threshold on absolute value of exponent for which scaling is used.
* sysdeps/ieee754/flt-32/e_exp2f.c (__ieee754_exp2f): Likewise.
* math/auto-libm-test-in: Add more tests of exp2.
* math/auto-libm-test-out: Regenerated.
Similar to various other bugs in this area, some expm1 implementations
do not raise the underflow exception for subnormal arguments, when the
result is tiny and inexact. This patch forces the exception in a
similar way to previous fixes.
(The issue does not apply to the ldbl-* implementations or to those
for x86 / x86_64 long double. The change to
sysdeps/ieee754/dbl-64/wordsize-64/e_cosh.c is one I missed when
previously fixing bug 16354; the bug in that implementation was
previously latent, but the expm1 fixes stopped it being latent and so
required it to be fixed to avoid spurious underflows from cosh.)
Tested for x86_64 and x86.
[BZ #16353]
* sysdeps/i386/fpu/s_expm1.S (dbl_min): New object.
(__expm1): Force underflow exception for arguments with small
absolute value.
* sysdeps/i386/fpu/s_expm1f.S (flt_min): New object.
(__expm1f): Force underflow exception for arguments with small
absolute value.
* sysdeps/ieee754/dbl-64/s_expm1.c: Include <float.h>.
(__expm1): Force underflow exception for arguments with small
absolute value.
* sysdeps/ieee754/flt-32/s_expm1f.c: Include <float.h>.
(__expm1f): Force underflow exception for arguments with small
absolute value.
* sysdeps/ieee754/dbl-64/wordsize-64/e_cosh.c (__ieee754_cosh):
Check for small arguments before calling __expm1.
* math/auto-libm-test-in: Do not mark underflow exceptions as
possibly missing for bug 16353.
* math/auto-libm-test-out: Regenerated.
In the x86 / x86_64 implementations of expm1l, when expm1l's result
should underflow to 0 (argument minus the least subnormal, in some
rounding modes), it can be a zero of the wrong sign. This patch fixes
this by returning the argument with underflow forced in that case
(this is a 1ulp error relative to the correctly rounded result of -0,
which is OK in terms of the documented accuracy goals, whereas a
result with the wrong sign never is).
Tested for x86_64 and x86.
[BZ #18569]
* sysdeps/i386/fpu/e_expl.S (IEEE754_EXPL) [USE_AS_EXPM1L]: Force
underflow and return argument in case of subnormal argument.
* sysdeps/x86_64/fpu/e_expl.S (IEEE754_EXPL) [USE_AS_EXPM1L]:
Likewise.
* math/auto-libm-test-in: Add more tests of expm1.
* math/auto-libm-test-out: Regenerated.
Similar to various other bugs in this area, the x86 and x86_64
implementations of expl / exp10l can fail to produce underflow
exceptions when the unscaled result has trailing 0 bits so the scaling
down to subnormal precision is exact. This patch fixes this by
forcing the exception in the case of tiny results.
Tested for x86_64 and x86.
[BZ #16361]
* sysdeps/i386/fpu/e_expl.S [!USE_AS_EXPM1L] (cmin): New object.
[!USE_AS_EXPM1L] (IEEE754_EXPL): Force underflow exception for
tiny results.
* sysdeps/x86_64/fpu/e_expl.S [!USE_AS_EXPM1L] (cmin): New object.
[!USE_AS_EXPM1L] (IEEE754_EXPL): Force underflow exception for
tiny results.
* math/auto-libm-test-in: Add more tests of exp and exp10. Do not
mark underflow exceptions as possibly missing for bug 16361.
* math/auto-libm-test-out: Regenerated.
Similar to various other bugs in this area, some asinh implementations
do not raise the underflow exception for subnormal arguments, when the
result is tiny and inexact. This patch forces the exception in a
similar way to previous fixes.
Tested for x86_64, x86 and mips64.
[BZ #16350]
* sysdeps/i386/fpu/s_asinh.S (__asinh): Force underflow exception
for arguments with small absolute value.
* sysdeps/i386/fpu/s_asinhf.S (__asinhf): Likewise.
* sysdeps/i386/fpu/s_asinhl.S (__asinhl): Likewise.
* sysdeps/ieee754/dbl-64/s_asinh.c: Include <float.h>.
(__asinh): Force underflow exception for arguments with small
absolute value.
* sysdeps/ieee754/flt-32/s_asinhf.c: Include <float.h>.
(__asinhf): Force underflow exception for arguments with small
absolute value.
* sysdeps/ieee754/ldbl-128/s_asinhl.c: Include <float.h>.
(__asinhl): Force underflow exception for arguments with small
absolute value.
* sysdeps/ieee754/ldbl-128ibm/s_asinhl.c: Include <float.h>.
(__asinhl): Force underflow exception for arguments with small
absolute value.
* sysdeps/ieee754/ldbl-96/s_asinhl.c: Include <float.h>.
(__asinhl): Force underflow exception for arguments with small
absolute value.
* math/auto-libm-test-in: Do not mark underflow exceptions as
possibly missing for bug 16350.
* math/auto-libm-test-out: Regenerated.
sysdeps/unix/sysv/linux/bits/in.h (as included in netinet/in.h, and
via that in netdb.h and arpa/inet.h) defines a series of MCAST_*
macros, both under __USE_MISC and then again unconditionally. These
are not POSIX macros, nor in any of the namespaces listed in POSIX as
reserved for this header, so should not be defined unconditionally.
This patch duly removes the unconditional definitions, leaving the
ones conditional on __USE_MISC.
Tested for x86_64 and x86 (testsuite, and that installed stripped
shared libraries are unchanged by the patch).
[BZ #18558]
* sysdeps/unix/sysv/linux/bits/in.h (MCAST_JOIN_GROUP): Remove
unconditional definition.
(MCAST_BLOCK_SOURCE): Likewise.
(MCAST_UNBLOCK_SOURCE): Likewise.
(MCAST_LEAVE_GROUP): Likewise.
(MCAST_JOIN_SOURCE_GROUP): Likewise.
(MCAST_LEAVE_SOURCE_GROUP): Likewise.
(MCAST_MSFILTER): Likewise.
* conform/Makefile (test-xfail-XOPEN2K/arpa/inet.h/conform):
Remove variable.
(test-xfail-XOPEN2K/netdb.h/conform): Likewise.
(test-xfail-XOPEN2K/netinet/in.h/conform): Likewise.
(test-xfail-XOPEN2K8/arpa/inet.h/conform): Likewise.
(test-xfail-XOPEN2K8/netdb.h/conform): Likewise.
(test-xfail-XOPEN2K8/netinet/in.h/conform): Likewise.
sysdeps/ieee754/ldbl-128ibm has its own versions of cprojl, ctanhl and
ctanl.
Having its own versions, where otherwise the math/ copies are
generally used for all floating-point formats, means they are liable
to get out of sync and not benefit from bug fixes to the generic
versions. The substantive differences (not arising from getting out
of sync and slightly different fixes for the same issues) are: long
double compat handling (also done in the ldbl-opt versions, so doesn't
require special versions for ldbl-128ibm); handling of LDBL_EPSILON
(conditionally undefined and redefined in other math/ implementations,
so doesn't justify a special version), and:
/* __gcc_qmul does not respect -0.0 so we need the following fixup. */
if ((__real__ res == 0.0L) && (__real__ x == 0.0L))
__real__ res = __real__ x;
if ((__real__ res == 0.0L) && (__imag__ x == 0.0L))
__imag__ res = __imag__ x;
But if that statement about __gcc_qmul was ever true for an old
version of that libgcc function, it's not the case for any GCC version
now supported to build glibc; there's explicit logic early in that
function (and similarly in __gcc_qdiv) to return an appropriately
signed zero if the product of the high parts is zero. So this patch
adds the special LDBL_EPSILON handling to the generic functions and
removes the ldbl-128ibm versions.
Tested for powerpc32 (compared test-ldouble.out before and after the
changes; there are slight changes to results for ctanl / ctanhl,
arising from divergence of the implementations, but nothing that
affects the overall nature of the issues shown by the testsuite, and
in particular nothing related to signs of zero resutls).
* math/s_ctanhl.c [LDBL_MANT_DIG == 106] (LDBL_EPSILON): Undefine
and redefine.
* math/s_ctanl.c [LDBL_MANT_DIG == 106] (LDBL_EPSILON): Undefine
and redefine.
* sysdeps/ieee754/ldbl-128ibm/s_cprojl.c: Remove file.
* sysdeps/ieee754/ldbl-128ibm/s_ctanhl.c: Likewise.
* sysdeps/ieee754/ldbl-128ibm/s_ctanl.c: Likewise.
nice (XPG3) calls getpriority and setpriority (in XPG4 but not XPG3,
i.e. UX-shaded in XPG4). This patch fixes this by making those
functions into weak aliases of __* functions and calling the __*
versions as needed.
Tested for x86_64 and x86 (testsuite, and that disassembly of
installed shared libraries is unchanged by this patch).
This completes cleaning up the unsorted linknamespace test XFAILs.
[BZ #18553]
* resource/getpriority.c (getpriority): Rename to __getpriority
and define as weak alias of __getpriority.
* resource/setpriority.c (setpriority): Rename to __setpriority
and define as weak alias of __setpriority.
* sysdeps/mach/hurd/getpriority.c (getpriority): Rename to
__getpriority and define as weak alias of __getpriority.
* sysdeps/mach/hurd/setpriority.c (setpriority): Rename to
__setpriority and define as weak alias of __setpriority.
* sysdeps/unix/syscalls.list (getpriority): Use __getpriority as
strong name.
(setpriority): Use __setpriority as strong name.
* sysdeps/unix/sysv/linux/getpriority.c (getpriority): Rename to
__getpriority and define as weak alias of __getpriority.
* include/sys/resource.h (__getpriority): Declare. Use
libc_hidden_proto.
(__setpriority): Likewise.
(getpriority): Don't use libc_hidden_proto.
(setpriority): Likewise.
* sysdeps/posix/nice.c (nice): Call __getpriority instead of
getpriority. Call __setpriority instead of setpriority.
* conform/Makefile (test-xfail-XPG3/unistd.h/linknamespace):
Remove variable.
mq_notify (in the 1996 edition of POSIX) brings in references to recv
and socket (not in POSIX until the 2001 edition). This patch fixes
this by using __recv and __socket, exporting them from libc at version
GLIBC_PRIVATE.
Tested for x86_64 and x86 (testsuite and comparison of installed
stripped shared libraries; PLT / dynamic symbol table changes render
the comparison not particularly useful for libc).
[BZ #18546]
* socket/recv.c (__recv): Use libc_hidden_def.
* socket/socket.c (__socket): Likewise.
* sysdeps/mach/hurd/recv.c (__recv): Likewise.
* sysdeps/mach/hurd/socket.c (__socket): Likewise.
* sysdeps/unix/sysv/linux/generic/recv.c (__recv): Likewise.
* sysdeps/unix/sysv/linux/recv.c (__recv): Use libc_hidden_weak.
* sysdeps/unix/sysv/linux/socket.c (__socket): Use
libc_hidden_def.
* sysdeps/unix/sysv/linux/x86_64/recv.c (__recv): Use
libc_hidden_weak.
* include/sys/socket.h (__socket): Do not use attribute_hidden.
Use libc_hidden_proto.
(__recv): Likewise.
* socket/Versions (libc): Export __recv and __socket at version
GLIBC_PRIVATE.
* sysdeps/unix/sysv/linux/mq_notify.c (helper_thread): Call __recv
instead of recv.
(init_mq_netlink): Call __socket instead of socket.
* conform/Makefile (test-xfail-POSIX/mqueue.h/linknamespace):
Remove variable.
mq_receive calls mq_timedreceive, and mq_send calls mq_timedsend. But
mq_receive and mq_send were in POSIX by 1996, while mq_timed* were
added in the 2001 edition of POSIX. This patch fixes this by making
mq_timed* into weak aliases for __mq_timed* and calling the
__mq_timed* names.
Tested for x86_64 and x86 (testsuite, and that disassembly of
installed shared libraries is unchanged by the patch).
[BZ #18545]
* rt/mq_timedreceive.c (mq_timedreceive): Rename to
__mq_timedreceive and define as alias of __mq_timedreceive. Use
hidden_weak.
* rt/mq_timedsend.c (mq_timedsend): Rename to __mq_timedsend and
define as alias of __mq_timedsend. Use hidden_weak.
* sysdeps/unix/sysv/linux/syscalls.list (mq_timedsend): Use
__mq_timedsend as strong name.
(mq_timedreceive): Use __mq_timedreceive as strong name.
* include/mqueue.h (__mq_timedsend): Declare. Use hidden_proto.
(__mq_timedreceive): Likewise.
* sysdeps/unix/sysv/linux/mq_receive.c (mq_receive): Call
__mq_timedreceive instead of mq_timedreceive.
* sysdeps/unix/sysv/linux/mq_send.c (mq_send): Call __mq_timedsend
instead of mq_timedsend.
* conform/Makefile (test-xfail-UNIX98/mqueue.h/linknamespace):
Remove variable.
The syscall wrappers mechanism automatically creates hidden aliases
for syscalls with libc_hidden_def / libc_hidden_weak. The use of
libc_hidden_* has the side-effect that for syscall wrappers in
non-libc libraries those aliases are not created. In turn, this means
that three mq_* syscalls in sysdeps/unix/sysv/linux/syscalls.list list
the __GI_* names explicitly.
The use of libc_hidden_* dates back to the original introduction of
that support in
2002-08-03 Roland McGrath <roland@redhat.com>
* sysdeps/unix/make-syscalls.sh: Generate libc_hidden_def or
libc_hidden_weak for every system call symbol defined.
(predating the non-libc syscalls in question) and I see no reason for
excluding non-libc syscalls. This patch changes the code to use
hidden_def / hidden_weak (via a wrapper syscall_hidden_def in the case
where the argument is itself a macro, so that the argument gets
expanded before concatenation with __GI_), so avoiding the need to
specify the hidden aliases explicitly in this case.
Tested for x86_64 and x86 (testsuite, and that disassembly of
installed stripped shared libraries is unchanged by the patch; the
mq_* symbols change from weak to strong, which is of no significance
and two of them will shortly change back to weak as part of a fix for
bug 18545).
* sysdeps/unix/make-syscalls.sh (emit_weak_aliases): Use
hidden_def and hidden_weak instead of libc_hidden_def and
libc_hidden_weak.
(top level): Refer to hidden_def in comment.
* sysdeps/unix/syscall-template.S (syscall_hidden_def): New
macro. Use it instead of libc_hidden_def.
* sysdeps/unix/sysv/linux/syscalls.list (mq_timedsend): Do not
specify __GI_* name explicitly.
(mq_timedreceive): Likewise.
(mq_setattr): Likewise.
mq_notify (present in POSIX by 1996) brings in references to
pthread_barrier_init and pthread_barrier_wait (new in the 2001 edition
of POSIX). This patch fixes this by making those functions into weak
aliases of __pthread_barrier_*, exporting the __pthread_barrier_*
names at version GLIBC_PRIVATE and using them in mq_notify.
Tested for x86_64 and x86 (testsuite, and comparison of installed
stripped shared libraries). Changes in addresses from dynamic symbol
table / PLT changes render most comparisons not particularly useful,
but when the addresses of subsequent code don't change there's no sign
of unexpected changes there. This patch does not remove any
linknamespace XFAILs because of other namespace issues remaining with
mqueue.h functions.
[BZ #18544]
* nptl/pthread_barrier_init.c (pthread_barrier_init): Rename to
__pthread_barrier_init and define as weak alias of
__pthread_barrier_init.
* sysdeps/sparc/nptl/pthread_barrier_init.c
(pthread_barrier_init): Likewise.
* nptl/pthread_barrier_wait.c (pthread_barrier_wait): Rename to
__pthread_barrier_wait and define as weak alias of
__pthread_barrier_wait.
* sysdeps/sparc/nptl/pthread_barrier_wait.c
(pthread_barrier_wait): Likewise.
* sysdeps/sparc/sparc32/pthread_barrier_wait.c
(pthread_barrier_wait): Likewise.
* sysdeps/unix/sysv/linux/i386/i486/pthread_barrier_wait.S
(pthread_barrier_wait): Likewise.
* sysdeps/unix/sysv/linux/x86_64/pthread_barrier_wait.S
(pthread_barrier_wait): Likewise.
* nptl/Versions (libpthread): Export __pthread_barrier_init and
__pthread_barrier_wait at version GLIBC_PRIVATE.
* include/pthread.h (__pthread_barrier_init): Declare.
(__pthread_barrier_wait): Likewise.
* sysdeps/unix/sysv/linux/mq_notify.c (notification_function):
Call __pthread_barrier_wait instead of pthread_barrier_wait.
(helper_thread): Likewise.
(init_mq_netlink): Call __pthread_barrier_init instead of
pthread_barrier_init.
The sem_* functions bring in references to tdelete, tfind, tsearch and
twalk. But the t* functions are XSI-shaded, while sem_* aren't. This
patch fixes this by using __t* instead, exporting those functions from
libc at version GLIBC_PRIVATE (since sem_* are in libpthread) and
using libc_hidden_* for the benefit of calls within libc.
Tested for x86_64 and x86 (testsuite, and comparison of disassembly of
installed stripped shared libraries). libpthread gets changes from
PLT reordering; addresses in libc change because of PLT / dynamic
symbol table changes.
[BZ #18536]
* misc/tsearch.c (__tsearch): Use libc_hidden_def.
(__tfind): Likewise.
(__tdelete): Likewise.
(__twalk): Likewise.
* misc/Versions (libc): Add __tdelete, __tfind, __tsearch and
__twalk to GLIBC_PRIVATE.
* include/search.h (__tsearch): Use libc_hidden_proto.
(__tfind): Likewise.
(__tdelete): Likewise.
(__twalk): Likewise.
* nptl/sem_close.c (sem_close): Call __twalk instead of twalk.
Call __tdelete instead of tdelete.
* nptl/sem_open.c (check_add_mapping): Call __tfind instead of
tfind. Call __tsearch instead of tsearch.
* sysdeps/sparc/sparc32/sem_open.c (check_add_mapping): Likewise.
* conform/Makefile (test-xfail-POSIX/semaphore.h/linknamespace):
Remove variable.
(test-xfail-POSIX2008/semaphore.h/linknamespace): Likewise.
Some of the cfi annotations used incorrect sign.
* sysdeps/aarch64/dl-tlsdesc.S (_dl_tlsdesc_return_lazy): Fix
cfi_adjust_cfa_offset argument.
(_dl_tlsdesc_undefweak, _dl_tlsdesc_dynamic): Likewise.
(_dl_tlsdesc_resolve_rela, _dl_tlsdesc_resolve_hold): Likewise.
Lazy TLSDESC initialization needs to be synchronized with concurrent TLS
accesses. The TLS descriptor contains a function pointer (entry) and an
argument that is accessed from the entry function. With lazy initialization
the first call to the entry function updates the entry and the argument to
their final value. A final entry function must make sure that it accesses an
initialized argument, this needs synchronization on systems with weak memory
ordering otherwise the writes of the first call can be observed out of order.
There are at least two issues with the current code:
tlsdesc.c (i386, x86_64, arm, aarch64) uses volatile memory accesses on the
write side (in the initial entry function) instead of C11 atomics.
And on systems with weak memory ordering (arm, aarch64) the read side
synchronization is missing from the final entry functions (dl-tlsdesc.S).
This patch only deals with aarch64.
* Write side:
Volatile accesses were replaced with C11 relaxed atomics, and a release
store was used for the initialization of entry so the read side can
synchronize with it.
* Read side:
TLS access generated by the compiler and an entry function code is roughly
ldr x1, [x0] // load the entry
blr x1 // call it
entryfunc:
ldr x0, [x0,#8] // load the arg
ret
Various alternatives were considered to force the ordering in the entry
function between the two loads:
(1) barrier
entryfunc:
dmb ishld
ldr x0, [x0,#8]
(2) address dependency (if the address of the second load depends on the
result of the first one the ordering is guaranteed):
entryfunc:
ldr x1,[x0]
and x1,x1,#8
orr x1,x1,#8
ldr x0,[x0,x1]
(3) load-acquire (ARMv8 instruction that is ordered before subsequent
loads and stores)
entryfunc:
ldar xzr,[x0]
ldr x0,[x0,#8]
Option (1) is the simplest but slowest (note: this runs at every TLS
access), options (2) and (3) do one extra load from [x0] (same address
loads are ordered so it happens-after the load on the call site),
option (2) clobbers x1 which is problematic because existing gcc does
not expect that, so approach (3) was chosen.
A new _dl_tlsdesc_return_lazy entry function was introduced for lazily
relocated static TLS, so non-lazy static TLS can avoid the synchronization
cost.
[BZ #18034]
* sysdeps/aarch64/dl-tlsdesc.h (_dl_tlsdesc_return_lazy): Declare.
* sysdeps/aarch64/dl-tlsdesc.S (_dl_tlsdesc_return_lazy): Define.
(_dl_tlsdesc_undefweak): Guarantee TLSDESC entry and argument load-load
ordering using ldar.
(_dl_tlsdesc_dynamic): Likewise.
(_dl_tlsdesc_return_lazy): Likewise.
* sysdeps/aarch64/tlsdesc.c (_dl_tlsdesc_resolve_rela_fixup): Use
relaxed atomics instead of volatile and synchronize with release store.
(_dl_tlsdesc_resolve_hold_fixup): Use relaxed atomics instead of
volatile.
* elf/tlsdeschtab.h (_dl_tlsdesc_resolve_early_return_p): Likewise.
Various functions in XPG4 bring in references to getlogin_r, which is
not in XPG4; this is also a bug for some older POSIX versions which
aren't yet covered by the linknamespace tests. This patch fixes this
by making getlogin_r into a weak alias for __getlogin_r and using
__getlogin_r as needed.
Tested for x86_64 and x86 (testsuite, and that disassembly of
installed stripped shared libraries is unchanged by the patch).
[BZ #18527]
* login/getlogin_r.c (getlogin_r): Rename to __getlogin_r and
define as weak alias of __getlogin_r. Use libc_hidden_weak.
* sysdeps/mach/hurd/getlogin_r.c (getlogin_r): Likewise.
* sysdeps/unix/getlogin_r.c (getlogin_r): Likewise.
* sysdeps/unix/sysv/linux/getlogin_r.c (getlogin_r): Likewise.
* include/unistd.h (__getlogin_r): Declare. Use
libc_hidden_proto.
* posix/glob.c (glob): Call __getlogin_r instead of getlogin_r.
* conform/Makefile (test-xfail-XPG3/glob.h/linknamespace): Remove
variable.
(test-xfail-XPG3/wordexp.h/linknamespace): Likewise.
(test-xfail-XPG4/glob.h/linknamespace): Likewise.
(test-xfail-XPG4/wordexp.h/linknamespace): Likewise.
aio_* bring in references to pread, which isn't in all the standards
containing aio_* (as a reference from one library to another, this is
a bug for dynamic as well as static linking). This patch fixes this
by using __libc_pread instead, exporting that function from libc at
symbol version GLIBC_PRIVATE; the code, with conditionals that may
call either __pread64 or __libc_pread, becomes exactly analogous to
that elsewhere in the same file that may call either __pwrite64 or
__libc_pwrite.
Tested for x86_64 and x86 (testsuite, and comparison of disassembly of
installed shared libraries). libc changes because of the PLT entry
for the newly exported __libc_pread; librt changes because of
assertion line numbers and PLT rearrangement; other stripped installed
shared libraries do not change.
[BZ #18519]
* posix/Versions (libc): Export __libc_pread at version
GLIBC_PRIVATE.
* sysdeps/pthread/aio_misc.c (handle_fildes_io): Call __libc_pread
instead of pread.
* conform/Makefile (test-xfail-POSIX/aio.h/linknamespace): Remove
variable.
Here is implementation of vectorized sin containing SSE, AVX,
AVX2 and AVX512 versions according to Vector ABI
<https://groups.google.com/forum/#!topic/x86-64-abi/LmppCfN1rZ4>.
* bits/libm-simd-decl-stubs.h: Added stubs for sin.
* math/bits/mathcalls.h: Added sin declaration with __MATHCALL_VEC.
* sysdeps/unix/sysv/linux/x86_64/libmvec.abilist: New versions added.
* sysdeps/x86/fpu/bits/math-vector.h: SIMD declaration for sin.
* sysdeps/x86_64/fpu/Makefile (libmvec-support): Added new files.
* sysdeps/x86_64/fpu/Versions: New versions added.
* sysdeps/x86_64/fpu/libm-test-ulps: Regenerated.
* sysdeps/x86_64/fpu/multiarch/Makefile (libmvec-sysdep_routines): Added
build of SSE, AVX2 and AVX512 IFUNC versions.
* sysdeps/x86_64/fpu/multiarch/svml_d_sin2_core.S: New file.
* sysdeps/x86_64/fpu/multiarch/svml_d_sin2_core_sse4.S: New file.
* sysdeps/x86_64/fpu/multiarch/svml_d_sin4_core.S: New file.
* sysdeps/x86_64/fpu/multiarch/svml_d_sin4_core_avx2.S: New file.
* sysdeps/x86_64/fpu/multiarch/svml_d_sin8_core.S: New file.
* sysdeps/x86_64/fpu/multiarch/svml_d_sin8_core_avx512.S: New file.
* sysdeps/x86_64/fpu/svml_d_sin2_core.S: New file.
* sysdeps/x86_64/fpu/svml_d_sin4_core.S: New file.
* sysdeps/x86_64/fpu/svml_d_sin4_core_avx.S: New file.
* sysdeps/x86_64/fpu/svml_d_sin8_core.S: New file.
* sysdeps/x86_64/fpu/svml_d_sin_data.S: New file.
* sysdeps/x86_64/fpu/svml_d_sin_data.h: New file.
* sysdeps/x86_64/fpu/test-double-vlen2-wrappers.c: Added vector sin test.
* sysdeps/x86_64/fpu/test-double-vlen2.c: Likewise.
* sysdeps/x86_64/fpu/test-double-vlen4-avx2-wrappers.c: Likewise.
* sysdeps/x86_64/fpu/test-double-vlen4-avx2.c: Likewise.
* sysdeps/x86_64/fpu/test-double-vlen4-wrappers.c: Likewise.
* sysdeps/x86_64/fpu/test-double-vlen4.c: Likewise.
* sysdeps/x86_64/fpu/test-double-vlen8-wrappers.c: Likewise.
* sysdeps/x86_64/fpu/test-double-vlen8.c: Likewise.
* NEWS: Mention addition of x86_64 vector sin.
Binutils 2.24 doesn't support some AVX512 instructions with ZMM
registers, so we need add more strict check.
* configure.ac: Added more strict check.
* configure: Regenerated.
This patch removes the vsyscall usage for x86_64 port. As indicated
by kernel code comments [1], vsyscalls are a legacy ABI and its concept
is problematic:
- It interferes with ASLR.
- It's awkward to write code that lives in kernel addresses but is
callable by userspace at fixed addresses.
- The whole concept is impossible for 32-bit compat userspace.
- UML cannot easily virtualize a vsyscall.
The VDSO is a better approach for such functionality. Tested on i686,
x86_64, and x32.
* sysdeps/unix/sysv/linux/i386/gettimeofday.c
(__gettimeofday_syscall): Remove vsyscall fallback.
* sysdeps/unix/sysv/linux/i386/time.c (__time_syscall): Likewise.
* sysdeps/unix/sysv/linux/x86/gettimeofday.c (__gettimeofday_syscall):
Add syscall fallback function.
(gettimeofday_ifunc): Use __gettimeofday_syscall as fallback mechanism
if vDSO is not present.
* sysdeps/unix/sysv/linux/x86/time.c (__time_syscall): Add syscall
fallback function.
(time_ifunc): Use __time_syscall as fallback mechanism if vDSO is not
present.
* sysdeps/unix/sysv/linux/x86_64/gettimeofday.c: Remove file.
* sysdeps/unix/sysv/linux/x86_64/time.c: Likewise.
[1] arch/x86/kernel/vsyscall_64.c
regcomp brings in references to wcscoll, which isn't in all the
standards that contain regcomp. In turn, wcscoll brings in references
to wcscmp, also not in all those standards. This patch fixes this by
making those functions into weak aliases of __wcscoll and __wcscmp and
calling those names instead as needed.
Tested for x86_64 and x86 (testsuite, and that disassembly of
installed shared libraries is unchanged by the patch).
[BZ #18497]
* wcsmbs/wcscmp.c [!WCSCMP] (WCSCMP): Define as __wcscmp instead
of wcscmp.
(wcscmp): Define as weak alias of WCSCMP.
* wcsmbs/wcscoll.c (STRCOLL): Define as __wcscoll instead of
wcscoll.
(USE_HIDDEN_DEF): Define.
[!USE_IN_EXTENDED_LOCALE_MODEL] (wcscoll): Define as weak alias of
__wcscoll. Don't use libc_hidden_weak.
* wcsmbs/wcscoll_l.c (STRCMP): Define as __wcscmp instead of
wcscmp.
* sysdeps/i386/i686/multiarch/wcscmp-c.c
[SHARED] (libc_hidden_def): Define __GI___wcscmp instead of
__GI_wcscmp.
(weak_alias): Undefine and redefine.
* sysdeps/i386/i686/multiarch/wcscmp.S (wcscmp): Rename to
__wcscmp and define as weak alias of __wcscmp.
* sysdeps/x86_64/wcscmp.S (wcscmp): Likewise.
* include/wchar.h (__wcscmp): Declare. Use libc_hidden_proto.
(__wcscoll): Likewise.
(wcscmp): Don't use libc_hidden_proto.
(wcscoll): Likewise.
* posix/regcomp.c (build_range_exp): Call __wcscoll instead of
wcscoll.
* posix/regexec.c (check_node_accept_bytes): Likewise.
* conform/Makefile (test-xfail-XPG3/regex.h/linknamespace): Remove
variable.
(test-xfail-XPG4/regex.h/linknamespace): Likewise.
(test-xfail-POSIX/regex.h/linknamespace): Likewise.
pathconf uses __statvfs64, and fpathconf uses __fstatvfs64. On
systems using sysdeps/unix/sysv/linux/wordsize-64, __statvfs64 then
brings in the strong symbol statvfs, and __fstatvfs64 brings in the
strong symbol fstatvfs, which are not in all the standards that have
pathconf and fpathconf. This patch fixes this by making those symbols
into weak aliases.
Tested for x86_64 and x86 (testsuite, and that disassembly of
installed shared libraries is unchanged by the patch).
[BZ #18507]
* sysdeps/unix/sysv/linux/fstatvfs.c (fstatvfs): Rename to
__fstatvfs and define as weak alias of __fstatvfs. Use
libc_hidden_weak.
* sysdeps/unix/sysv/linux/statvfs.c (statvs): Rename to __statvfs
and define as weak alias of __statvfs. Use libc_hidden_weak.
* sysdeps/unix/sysv/linux/wordsize-64/fstatvfs.c (__fstatvfs64):
Define as alias of __fstatvfs, not fstatvfs.
(fstatvfs64): Likewise.
* sysdeps/unix/sysv/linux/wordsize-64/statvfs.c (__statvfs64):
Define as alias of __statvfs, not statvfs.
(statvfs64): Likewise.
* conform/Makefile (test-xfail-POSIX/unistd.h/linknamespace):
Remove variable.
This patch consolidates the sched_getcpu implementations across all
arches (except tile, which requires its own). This patch removes
the powerpc, x86_64 and x32 specific files and change the default
linux one to use INLINE_VSYSCALL where possible (for ports that
implements it).
* math/Makefile: Added CFLAGS for new tests.
* math/test-float-vlen16.h: New file.
* math/test-float-vlen4.h: New file.
* math/test-float-vlen8.h: New file.
* math/test-double-vlen2.h: Fixed 2 argument macro and comment.
* sysdeps/x86_64/fpu/Makefile: Added new tests and variables.
* sysdeps/x86_64/fpu/libm-test-ulps: Regenerated.
* sysdeps/x86_64/fpu/test-float-vlen16-wrappers.c: New file.
* sysdeps/x86_64/fpu/test-float-vlen16.c: New file.
* sysdeps/x86_64/fpu/test-float-vlen4-wrappers.c: New file.
* sysdeps/x86_64/fpu/test-float-vlen4.c: New file.
* sysdeps/x86_64/fpu/test-float-vlen8-avx2-wrappers.c: New file.
* sysdeps/x86_64/fpu/test-float-vlen8-avx2.c: New file.
* sysdeps/x86_64/fpu/test-float-vlen8-wrappers.c: New file.
* sysdeps/x86_64/fpu/test-float-vlen8.c: New file.
Here is implementation of vectorized cosf containing SSE, AVX,
AVX2 and AVX512 versions according to Vector ABI
<https://groups.google.com/forum/#!topic/x86-64-abi/LmppCfN1rZ4>.
* sysdeps/x86_64/fpu/Makefile (libmvec-support): Added new files.
* sysdeps/x86_64/fpu/Versions: New versions added.
* sysdeps/x86_64/fpu/svml_s_cosf4_core.S: New file.
* sysdeps/x86_64/fpu/multiarch/svml_s_cosf4_core.S: New file.
* sysdeps/x86_64/fpu/multiarch/svml_s_cosf4_core_sse4.S: New file.
* sysdeps/x86_64/fpu/svml_s_cosf8_core_avx.S: New file.
* sysdeps/x86_64/fpu/svml_s_cosf8_core.S: New file.
* sysdeps/x86_64/fpu/multiarch/svml_s_cosf8_core.S: New file.
* sysdeps/x86_64/fpu/multiarch/svml_s_cosf8_core_avx2.S: New file.
* sysdeps/x86_64/fpu/svml_s_cosf16_core.S: New file.
* sysdeps/x86_64/fpu/multiarch/svml_s_cosf16_core.S: New file.
* sysdeps/x86_64/fpu/multiarch/svml_s_cosf16_core_avx512.S: New file.
* sysdeps/x86_64/fpu/svml_s_wrapper_impl.h: New file.
* sysdeps/x86_64/fpu/svml_s_cosf_data.S: New file.
* sysdeps/x86_64/fpu/svml_s_cosf_data.h: New file.
* sysdeps/x86_64/fpu/multiarch/Makefile (libmvec-sysdep_routines): Added
build of SSE, AVX2 and AVX512 IFUNC versions.
* sysdeps/unix/sysv/linux/x86_64/libmvec.abilist: New versions added.
* sysdeps/x86/fpu/bits/math-vector.h: Added SIMD declaration for cosf.
* NEWS: Mention addition of x86_64 vector cosf.
We test vector math functions using scalar tests infrastructure with
help of special wrappers from scalar versions to vector ones. Wrapper
implemented using platform specific vector types and placed in separate
file for compilation with architecture specific options, main part of
test has no such options. With help of system of definitions unfolding
of which is drived from test code we have wrapper called in individual
testing function instead of scalar function. Also system of definitions
includes generated during make check header math/libm-have-vector-test.h
with series of conditional definitions which help to avoid build fails
for functions having no vector versions; runtime architecture check
to prevent runtime fails of test run on inappropriate hardware.
* math/Makefile: Added rules for vector tests.
* math/gen-libm-have-vector-test.sh: Added generation of wrapper
declaration under condition.
* math/test-double-vlen2.h: New file.
* math/test-double-vlen4.h: New file.
* math/test-double-vlen8.h: New file.
* math/test-vec-loop.h: Added initialization macro.
* sysdeps/x86_64/fpu/Makefile: Added variables for vector tests.
* sysdeps/x86_64/fpu/libm-test-ulps: Regenarated.
* sysdeps/x86_64/fpu/math-tests-arch.h: New file.
* sysdeps/x86_64/fpu/test-double-vlen2-wrappers.c: New file.
* sysdeps/x86_64/fpu/test-double-vlen2.c: New file.
* sysdeps/x86_64/fpu/test-double-vlen4-avx2-wrappers.c: New file.
* sysdeps/x86_64/fpu/test-double-vlen4-avx2.c: New file.
* sysdeps/x86_64/fpu/test-double-vlen4-wrappers.c: New file.
* sysdeps/x86_64/fpu/test-double-vlen4.c: New file.
* sysdeps/x86_64/fpu/test-double-vlen8-wrappers.c: New file.
* sysdeps/x86_64/fpu/test-double-vlen8.c: New file.
Here is implementation of cos containing SSE, AVX, AVX2 and AVX512
versions according to Vector ABI which had been discussed in
<https://groups.google.com/forum/#!topic/x86-64-abi/LmppCfN1rZ4>.
Vector math library build and ABI testing enabled by default for x86_64.
* sysdeps/x86_64/fpu/Makefile: New file.
* sysdeps/x86_64/fpu/Versions: New file.
* sysdeps/x86_64/fpu/svml_d_cos_data.S: New file.
* sysdeps/x86_64/fpu/svml_d_cos_data.h: New file.
* sysdeps/x86_64/fpu/svml_d_cos2_core.S: New file.
* sysdeps/x86_64/fpu/svml_d_cos4_core.S: New file.
* sysdeps/x86_64/fpu/svml_d_cos4_core_avx.S: New file.
* sysdeps/x86_64/fpu/svml_d_cos8_core.S: New file.
* sysdeps/x86_64/fpu/svml_d_wrapper_impl.h: New file.
* sysdeps/x86_64/fpu/multiarch/svml_d_cos2_core.S: New file.
* sysdeps/x86_64/fpu/multiarch/svml_d_cos2_core_sse4.S: New file.
* sysdeps/x86_64/fpu/multiarch/svml_d_cos4_core.S: New file.
* sysdeps/x86_64/fpu/multiarch/svml_d_cos4_core_avx2.S: New file.
* sysdeps/x86_64/fpu/multiarch/svml_d_cos8_core.S: New file.
* sysdeps/x86_64/fpu/multiarch/svml_d_cos8_core_avx512.S: New file.
* sysdeps/x86_64/fpu/multiarch/Makefile (libmvec-sysdep_routines): Added
build of SSE, AVX2 and AVX512 IFUNC versions.
* sysdeps/x86/fpu/bits/math-vector.h: Added SIMD declaration for cos.
* math/bits/mathcalls.h: Added cos declaration with __MATHCALL_VEC.
* sysdeps/x86_64/configure.ac: Options for libmvec build.
* sysdeps/x86_64/configure: Regenerated.
* sysdeps/x86_64/sysdep.h (cfi_offset_rel_rsp): New macro.
* sysdeps/unix/sysv/linux/x86_64/libmvec.abilist: New file.
* manual/install.texi (Configuring and compiling): Document
--disable-mathvec.
* INSTALL: Regenerated.
* NEWS: Mention addition of libmvec and x86_64 vector cos.
Handle signed integer overflow correctly. Detect and reject O_APPEND.
Document drawbacks of emulation.
This does not completely address bug 15661, but improves the situation
somewhat.
Beginning with the upcoming 4.1 release, Linux on a subset of 32-bit
ARM hardware will provide fast user-space implementations of the
following system calls:
- gettimeofday
- clock_gettime
The kernel implementation depends on the ARMv7 Generic Timers
Extension to accelerate these system calls. So CPUs such as
Cortex-A15 and -A7 benefit, while Cortex-A9, -A8, and pre-v7 CPUs do
not. On systems where the VDSO does not provide any speedup, the
kernel prevents the relevant symbol lookups from succeeding.
On OMAP5 (Cortex-A15) gettimeofday latency decreases from ~350ns to
~120ns. On BeagleBone Black (Cortex-A8) it goes from ~650ns to
~660ns, which to my mind is an acceptable cost.
Verified that no new test failures are introduced on kernels with and
without the VDSO.
* sysdeps/unix/sysv/linux/arm/Makefile: (sysdep_routines):
Include dl-vdso.
* sysdeps/unix/sysv/linux/arm/init-first.c: New file:
Use VDSO routines for gettimeofday, clock_gettime if
available.
* sysdeps/unix/sysv/linux/arm/libc-vdso.h: New file:
Declare VDSO symbols.
* sysdeps/unix/sysv/linux/arm/sysdep.h:
[HAVE_GETTIMEOFDAY_VSYSCALL]: Define.
[HAVE_CLOCK_GETTIME_VSYSCALL]: Define.
* sysdeps/unix/sysv/linux/arm/Versions: Add
__vdso_clock_gettime.
This patch uses inline calls (through INLINE_SYSCALL macro) to define
the non-cancellable functions macros to avoid use of the
syscall_nocancel entrypoint.
Carlos noted in
<https://sourceware.org/ml/libc-alpha/2015-05/msg00680.html> that
various ports use potentially problematic short variables names in
their syscall macros, which could shadow variables with the same name
from containing scopes.
This patch fixes variables called err and ret in MIPS macros. (I left
result_var and _sys_result - separate variables in different macros,
which need separate names - alone.)
Tested for mips64 (all three ABIs) that installed stripped shared
libraries are unchanged by this patch.
* sysdeps/unix/sysv/linux/mips/mips32/sysdep.h (INLINE_SYSCALL):
Use variable name _sc_err instead of err.
[__mips16] (INTERNAL_SYSCALL_NCS): Use variable name _sc_ret
instead of ret.
* sysdeps/unix/sysv/linux/mips/mips64/n32/sysdep.h
(INLINE_SYSCALL): Use variable name _sc_err instead of err.
* sysdeps/unix/sysv/linux/mips/mips64/n64/sysdep.h
(INLINE_SYSCALL): Likewise.
Various code in glibc uses __strnlen instead of strnlen for namespace
reasons. However, __strnlen does not use libc_hidden_proto /
libc_hidden_def (as is normally done for any function defined and
called within the same library, whether or not exported from the
library and whatever namespace it is in), so the compiler does not
know that those calls are to a function within libc.
This patch uses libc_hidden_proto / libc_hidden_def with __strnlen.
On x86_64, it makes no difference to the installed stripped shared
libraries. On 32-bit x86, it causes __strnlen calls to go to the same
place as strnlen calls (the fallback strnlen implementation), rather
than through a PLT entry for the strnlen IFUNC; I'm not sure of the
logic behind when calls from within libc should use IFUNCs versus when
they should go direct to a particular function implementation, but
clearly it doesn't make sense for strnlen and __strnlen to be handled
differently in this regard.
Tested for x86_64 and x86 (testsuite, and comparison of installed
shared libraries as described above).
* string/strnlen.c [!STRNLEN] (__strnlen): Use libc_hidden_def.
* include/string.h (__strnlen): Use libc_hidden_proto.
* sysdeps/aarch64/strnlen.S (__strnlen): Use libc_hidden_def.
* sysdeps/i386/i686/multiarch/strnlen-c.c [SHARED]
(libc_hidden_def): Define __GI___strnlen as well as __GI_strnlen.
* sysdeps/powerpc/powerpc32/power4/multiarch/strnlen-power7.S
(libc_hidden_def): Undefine and redefine.
* sysdeps/powerpc/powerpc32/power4/multiarch/strnlen-ppc32.c
[SHARED] (libc_hidden_def): Define __GI___strnlen as well as
__GI_strnlen.
* sysdeps/powerpc/powerpc32/power7/strnlen.S (__strnlen): Use
libc_hidden_def.
* sysdeps/tile/tilegx/strnlen.c (__strnlen): Likewise.
