about summary refs log tree commit diff
Commit message (Collapse)AuthorAgeFilesLines
...
* Add versions of wcscpy, wcschr, wcsrchr for power6/power7.Will Schmidt2012-08-228-2/+298
| | | | | | | | | | | | | | Initially based on the versions found in wcsmbs/* ; these files have been changed by hand unrolling, and adding some additional variables to allow some read-ahead to occur, which then relieves some of the wait-for-increment/wait-for-load/wait-for-compare-results pressure that was slowing down every iteration through the while-loop. For 64-bit Power7, These changes give an approx 20% throughput boost for the wcschr and wcsrchr functions; and approx 40% boost for the wcscpy function. 32-bit improvements appear to be slightly better with ~ %30 and ~ %45 respectively. Results for Power6 closely match those for power7.
* Micro-optimize critical path of strstr, strcase and memmem.Maxim Kuvyrkov2012-08-214-1/+18
|
* Use pointers for traversing arrays in strstr, strcasestr and memmem.Maxim Kuvyrkov2012-08-213-18/+55
|
* Detect EOL on-the-fly in strstr, strcasestr and memmem.Maxim Kuvyrkov2012-08-217-15/+108
|
* Optimize first-character loop of strstr, strcasestr and memmem.Maxim Kuvyrkov2012-08-212-1/+20
|
* Support static IFUNC calls irrespective of USE_MULTIARCH.Roland McGrath2012-08-213-46/+43
|
* Remove __ASSUME_FADVISE64_64_SYSCALL.Joseph Myers2012-08-2110-152/+50
|
* [Powerpc] Tune/optimize powerpc{32,64}/power7/memchr.S.Will Schmidt2012-08-213-44/+125
| | | | | | | | | | | | Assorted tweaking, twisting and tuning to squeeze a few additional cycles out of the memchr code. Changes include bypassing the shift pairs (sld,srd) when they are not required, and unrolling the small_loop that handles short and trailing strings. Per scrollpipe data measuring aligned strings for 64-bit, these changes save between five and eight cycles (9-13% overall) for short strings (<32), Longer aligned strings see slight improvement of 1-3% due to bypassing the shifts and the instruction rearranging.
* Fix typos in manual wrt syslog.Roland McGrath2012-08-202-2/+6
|
* Fix conditional on using DSOCAPS to match condition on defining it.Roland McGrath2012-08-202-1/+6
|
* Remove __ASSUME_SWAPCONTEXT_SYSCALL.Joseph Myers2012-08-205-21/+14
|
* Define __ASSUME_UTIMES for s390.Joseph Myers2012-08-202-2/+7
|
* Remove __ASSUME_MMAP2_SYSCALL.Joseph Myers2012-08-2015-185/+69
|
* S/390: Remove 32 bit getrlimit.c.Andreas Krebbel2012-08-202-1/+4
|
* Add bug number to NEWS.Joseph Myers2012-08-191-3/+3
|
* Fix last patch: Add missing DUMMY variableAndreas Jaeger2012-08-182-2/+6
|
* add attribute_hidden to __have_{sock_cloexec,pipe2,dup3Mike Frysinger2012-08-183-4/+11
| | | | | | | These internal knobs are not exposed as part of the public ABI, so mark them hidden to avoid generating relocations against them. Signed-off-by: Mike Frysinger <vapier@gentoo.org>
* split assume pipe2/dup3/sock_cloexec knobsMike Frysinger2012-08-183-6/+21
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | We can't assume sock_cloexec and pipe2 are bound together as the former defines are found in glibc only while the latter are a combo of kernel headers and glibc. So if we do a runtime detection of SOCK_CLOEXEC, but pipe2() is a stub inside of glibc, we hit a problem. For example: main() { getgrnam("portage"); if (!popen("ls", "r")) perror("popen()"); } getgrnam() will detect that the kernel supports SOCK_CLOEXEC and then set both __have_sock_cloexec and __have_pipe2 to true. But if glibc was built against older kernel headers where __NR_pipe2 does not exist, glibc will have a ENOSYS stub for it. So popen() will always fail as glibc assumes pipe2() works. While this isn't too much of an issue for some arches as they added the functionality to the kernel at the same time, not all arches are that lucky. Since the code already has dedicated names for each feature, delete the defines wiring these three features together and make each one a proper dedicated knob. We've been carrying this in Gentoo since glibc-2.9. Signed-off-by: Mike Frysinger <vapier@gentoo.org>
* update linux nice.c include path too after recent file shuffleMike Frysinger2012-08-172-1/+5
| | | | Signed-off-by: Mike Frysinger <vapier@gentoo.org>
* Quash warning in s_sincosl.Marek Polacek2012-08-172-2/+7
|
* Adjust old #include's of sysdeps/unix/time.c to point to ↵Roland McGrath2012-08-173-3/+5
| | | | sysdeps/posix/time.c instead.
* Make sysheaders available in config.make.Roland McGrath2012-08-174-0/+8
|
* Move some things from sysdeps/unix to sysdeps/posix.Roland McGrath2012-08-177-0/+17
|
* Fix rule dependency in last change.Roland McGrath2012-08-172-2/+6
|
* 2012-08-17 Jeff Law <law@redhat.com>Jeff Law2012-08-175-133/+31
| | | | | | | | | | | | * intl/Makefile (codeset_mo): New variable. ($(codeset_mo)): New target. (tst-codeset.out): Depend on that. Remove explicit rule. (tst-gettext3.out, tst-gettext5.out): Likewise. (LOCPATH-ENV, tst-codeset-ENV): New variables. (tst-gettext3-ENV, tst-gettext5-ENV): Likewise. * intl/tst-codeset.sh: Remove. * intl/tst-gettext3.sh: Likewise. * intl/tst-gettext5.sh: Likewise.
* Merge unix/inet/syscalls.list into unix/syscalls.list.Roland McGrath2012-08-173-23/+24
|
* Fix typo in last change.Roland McGrath2012-08-171-1/+1
|
* Fix getaddrinfo for [!_STATBUF_ST_NSEC] case.Roland McGrath2012-08-172-2/+42
|
* Make malloc build for no-threads configurations.Roland McGrath2012-08-174-12/+25
|
* Change type of constant to avoid a warning.Roland McGrath2012-08-172-2/+7
|
* Split sys/param.h out into common file and sysdeps bits/param.h file.Roland McGrath2012-08-178-163/+144
|
* Add a cast to silence a warning.Roland McGrath2012-08-172-2/+9
|
* Make libio compile without _IO_MTSAFE_IO.Roland McGrath2012-08-176-3/+31
|
* Clean up definition of _LIBC_REENTRANT and _IO_MTSAFE_IO.Roland McGrath2012-08-1718-61/+111
|
* Also set r->r_map when unmapping the first object in a namespace.Gary Benson2012-08-172-1/+6
| | | | | | When unmapping the first object in a namespace, the runtime linker did not update the externally visible pointer. This resulted in debuggers seeing pointers to memory that had been freed.
* Do not pollute name space with internal_*netgrent functions.Roland McGrath2012-08-163-29/+39
|
* Update longlong.h from GCC.Joseph Myers2012-08-162-27/+21
|
* Suppress -Wunused-but-set warnings in GMP code.Roland McGrath2012-08-166-15/+32
|
* Fifth argument of la_pltenter() is not constant.Carlos O'Donell2012-08-169-8/+32
| | | | | | | | | | | | The original runtime linker auditing interface described by Solaris allows the 5th argument of la_pltenter() to be modified. This patch cleans up the ldsodefs.h definitions such that the 5th argument is not constant. At one point the 5th argument *was* constant but this was changed with commit 2413fdba7a02ba8916f75d17199a6e9133a8f7b0. This patch updates alpha, ia64, mips, sh and sparc with similar changes.
* Remove __ASSUME_POSIX_TIMERS.Joseph Myers2012-08-1620-865/+326
|
* Fix dl-load.c for [!_LIBC_REENTRANT] case.Roland McGrath2012-08-152-0/+5
|
* Define __rtld_lock_initialize in stub bits/libc-lock.h file.Roland McGrath2012-08-152-0/+3
|
* Add a missing #include to dl-sym.c.Roland McGrath2012-08-152-1/+4
|
* Optimize __libc_lock_lock and __libc_lock_trylock for MIPS.Maxim Kuvyrkov2012-08-154-4/+57
|
* Add generic versions of pthread_spin_lock and pthread_spin_trylock.Maxim Kuvyrkov2012-08-1514-166/+123
|
* Suppress regcomp.c warnings in 32-bit builds.Roland McGrath2012-08-152-2/+9
|
* Define O_DSYNC and O_RSYNC in generic/4.4 bits/fcntl.h file.Roland McGrath2012-08-152-0/+7
|
* Move stub lseek.c to the right directory.Roland McGrath2012-08-152-0/+3
|
* Fix stub clock_nanosleep #include.Roland McGrath2012-08-152-3/+4
|
* Fix last sha512.c change to avoid compiler warning.Roland McGrath2012-08-152-1/+6
|