about summary refs log tree commit diff
Commit message (Collapse)AuthorAgeFilesLines
* Remove __libc_pselect alias.Joseph Myers2014-10-313-2/+4
| | | | | | | | | | | | Continuing the removal of unused __libc_* function names, this patch removes the __libc_pselect alias. Tested for x86_64 that installed stripped shared libraries are unchanged by this patch. * misc/pselect.c [!__pselect] (__libc_pselect): Remove alias. * sysdeps/unix/sysv/linux/pselect.c [__NR_pselect6] (__libc_pselect): Likewise.
* Add bug 15215 to NEWS; move bug 17344 to correct version's list in NEWS.Joseph Myers2014-10-311-3/+3
|
* Move powerpc64 pread/pwrite definitions to syscalls.list (bug 14138).Joseph Myers2014-10-316-221/+10
| | | | | | | | | | | | | | | | Concluding the move of syscall definitions to syscalls.list, where the removal of support for old kernel versions has made this possible, this patch removes C definitions of pread, pread64, pwrite and pwrite64 for powerpc64. As far as I can tell, the existing syscalls.list definitions in sysdeps/unix/sysv/linux/wordsize-64/syscalls.list should suffice to produce results equivalent to what these C files do. [BZ #14138] * sysdeps/unix/sysv/linux/powerpc/powerpc64/pread.c: Remove file. * sysdeps/unix/sysv/linux/powerpc/powerpc64/pread64.c: Likewise. * sysdeps/unix/sysv/linux/powerpc/powerpc64/pwrite.c: Likewise. * sysdeps/unix/sysv/linux/powerpc/powerpc64/pwrite64.c: Likewise.
* Fix SPARC atomic_write_barrier.Torvald Riegel2014-10-314-4/+13
|
* Clean up internal ctype.h header.Roland McGrath2014-10-302-16/+22
|
* Remove __libc_readv and __libc_writev function names.Joseph Myers2014-10-307-22/+30
| | | | | | | | | | | | | | | | | | | | | | | | | | | Continuing the removal of __libc_* function names that are no longer used anywhere, this patch removes the __libc_readv and __libc_writev names. Tested for x86_64 that stripped installed shared libraries are unchanged by the patch. * include/sys/uio.h (__libc_readv): Remove declaration. (__libc_writev): Likewise. * misc/readv.c (__libc_readv): Rename to __readv. (__readv): Do not define as alias. (readv): Define as alias of __readv. * misc/writev.c (__libc_writev): Rename to __writev. (__writev): Do not define as alias. (writev): Define as alias of __writev. * sysdeps/posix/readv.c (__libc_readv): Rename to __readv. (__readv): Do not define as alias. (readv): Define unconditionally as alias of __readv. * sysdeps/posix/writev.c (__libc_writev): Rename to __writev. (__writev): Do not define as alias. (writev): Define unconditionally as alias of __writev. * sysdeps/unix/syscalls.list (readv): Do not define __libc_readv name. (writev): Do not define __libc_writev name.
* Clean up wchar_t conversion code in iconv program.Roland McGrath2014-10-302-79/+61
|
* Remove __libc_creat function name.Joseph Myers2014-10-307-9/+23
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | glibc has lots of __libc_* function names that no longer serve any purpose (are not used for any calls or exported at a public symbol version). This patch removes __libc_creat. It has the effect of creat becoming a strong symbol instead of a weak symbol in various cases, but that's fine; in shared libraries it doesn't matter at all, while for static linking the only other symbol sometimes defined in the same object is creat64, and whenever creat64 is a reserved name so is creat. Other such cases of unnecessary __libc_* symbols are expected to be dealt with in separate patches over time. Tested for x86_64 (testsuite, and that the disassembly of installed shared libraries is unchanged by the patch). * include/fcntl.h (__libc_creat): Remove declaration. * io/creat.c (__libc_creat): Rename to creat. (creat): Do not define as alias. * sysdeps/unix/sysv/linux/alpha/creat.c (creat64): Define as alias of creat instead of __libc_creat. * sysdeps/unix/sysv/linux/generic/creat.c (__libc_creat): Rename to creat. (creat): Do not define as alias. [__WORDSIZE == 64] (creat64): Define as alias of creat instead of __libc_creat. * sysdeps/unix/sysv/linux/syscalls.list (creat): Do not define __libc_creat name. * sysdeps/unix/sysv/linux/wordsize-64/syscalls.list (creat): Likewise.
* manual/llio.texi: Add Linux-specific comments for write().Carlos O'Donell2014-10-292-0/+30
| | | | | | | | | | Add Linux-specific comments about the atomicity of write() and the POSIX requirements. 2014-10-29 Carlos O'Donell <carlos@redhat.com> * manual/llio.texi: Add comments discussing why write() may be considered MT-unsafe on Linux.
* elf/dl-load.c: Use __strdup.Carlos O'Donell2014-10-282-19/+12
| | | | | | | | | | | | | | | | | | | | | | During a refactoring pass several repeated blocks of code in dl-load.c were turned into a call to a local function named local_strdup. There is no need for local_strdup, and the routines should instead call __strdup. This change does just that. We call the internal symbol __strdup because calling strdup is unsafe. The user might be using a standard that doesn't include strdup and may have defined this symbol in their application. During a static link we might reference the user defined symbol and crash if it doesn't implement a standards conforming strdup. The resulting code is simpler to understand, and makes it easier to debug. No regressions on x86_64. 2014-10-28 Carlos O'Donell <carlos@redhat.com> * dl-load.c (local_strdup): Remove. (expand_dynamic_string_token): Use __strdup. (decompose_rpath): Likewise. (_dl_map_object): Likewise.
* Don't use INTDEF/INTUSE in unwind-dw2-fde.c (bug 14132).Joseph Myers2014-10-282-15/+36
| | | | | | | | | | | | | | | | | | | | | | | | | | Continuing the removal of the obsolete INTDEF / INTUSE mechanism, this patch replaces its use in unwind-dw2-fde.c with hidden_def and hidden_proto. Tested for x86. This patch does result in code generation differences (for some reason GCC decides to partition __register_frame_info_bases after the patch). [BZ #14132] * sysdeps/generic/unwind-dw2-fde.c (__register_frame_info_bases_internal): Do not declare. (__register_frame_info_table_bases_internal): Likewise. (__deregister_frame_info_bases_internal): Likewise. (__register_frame_info_bases): Declare and use hidden_proto before definition. Use hidden_def instead of INTDEF. (__register_frame_info_table_bases): Likewise. (__deregister_frame_info_bases): Likewise. (__register_frame_info): Do not use INTUSE. (__register_frame): Likewise. (__register_frame_info_table): Likewise. (__register_frame_table): Likewise. (__deregister_frame_info): Likewise. (__deregister_frame): Likewise.
* arm: Re-enable PI futex support for ARM kernels >= 3.14.3Gratian Crisan2014-10-272-4/+16
| | | | | | | | | | | | | | | | | | | ARM linux kernels before 3.14.3 may or may not support futex_atomic_cmpxchg_inatomic depending on the kernel configuration (e.g. CONFIG_CPU_USE_DOMAINS && CONFIG_SMP configuration was not supported) Starting with 3.14.3 the linux kernel unconditionally enables support for ARM, and this re-enables the relevant __ASSUME_* macros. Tested on ARM both with kernels >= 3.14.3 and older kernels. * sysdeps/unix/sysv/linux/arm/kernel-features.h [__LINUX_KERNEL_VERSION >= 0x030E03] (__ASSUME_FUTEX_LOCK_PI): Do not undefine. [__LINUX_KERNEL_VERSION >= 0x030E03] (__ASSUME_REQUEUE_PI): Likewise. [__LINUX_KERNEL_VERSION >= 0x030E03] (__ASSUME_SET_ROBUST_LIST): Likewise.
* Move setfsgid/setfsuid definitions to syscalls.list (bug 14138).Joseph Myers2014-10-2719-77/+49
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Continuing the move of syscall definitions to syscalls.list, where the removal of support for old kernel versions has made this possible, this patch moves various definitions of setfsgid and setfsuid. Tested for x86. [BZ #14138] * sysdeps/unix/sysv/linux/arm/setfsgid.c: Remove file. * sysdeps/unix/sysv/linux/arm/setfsuid.c: Likewise. * sysdeps/unix/sysv/linux/i386/setfsgid.c: Likewise. * sysdeps/unix/sysv/linux/i386/setfsuid.c: Likewise. * sysdeps/unix/sysv/linux/m68k/setfsgid.c: Likewise. * sysdeps/unix/sysv/linux/m68k/setfsuid.c: Likewise. * sysdeps/unix/sysv/linux/s390/s390-32/setfsgid.c: Likewise. * sysdeps/unix/sysv/linux/s390/s390-32/setfsuid.c: Likewise. * sysdeps/unix/sysv/linux/sh/setfsgid.c: Likewise. * sysdeps/unix/sysv/linux/sh/setfsuid.c: Likewise. * sysdeps/unix/sysv/linux/sparc/sparc32/setfsgid.c: Likewise. * sysdeps/unix/sysv/linux/sparc/sparc32/setfsuid.c: Likewise. * sysdeps/unix/sysv/linux/arm/syscalls.list (setfsgid): Add syscall. (setfsuid): Likewise. * sysdeps/unix/sysv/linux/i386/syscalls.list (setfsgid): Likewise. (setfsuid): Likewise. * sysdeps/unix/sysv/linux/m68k/syscalls.list (setfsgid): Likewise. (setfsuid): Likewise. * sysdeps/unix/sysv/linux/s390/s390-32/syscalls.list (setfsgid): Likewise. (setfsuid): Likewise. * sysdeps/unix/sysv/linux/sh/syscalls.list (setfsgid): Likewise. (setfsuid): Likewise. * sysdeps/unix/sysv/linux/sparc/sparc32/syscalls.list (setfsgid): Likewise. (setfsuid): Likewise.
* Fix misdetected Slow_SSE4_2 cpu feature bit (bug 17501)Andreas Schwab2014-10-276-11/+21
|
* Rework compiler version check in configure.Roland McGrath2014-10-243-66/+48
|
* Prototypify htonl and htons definitions.Roland McGrath2014-10-243-4/+7
|
* This patch improves strncat performance by using strlen. Strlen has a fast C ↵Wilco Dijkstra2014-10-242-4/+6
| | | | | | | | | implementation, so this will improve performance even on targets which don't have an optimized strlen. It is about twice as fast as the original strncat in bench-strncat.
* This patch improves strcat performance by using strlen and strcpy. Strlen ↵Wilco Dijkstra2014-10-242-20/+5
| | | | | | | | has a fast C implementation, so this improves performance even on targets which don't have an optimized strlen and strcpy - it is 25% faster in bench-strcat. On targets which don't provide an optimized strcat but which do have an optimized strlen and strcpy, performance gain is > 2x.
* Call libc_fetestexcept_aarch64 from math_private.h rather than duplicating ↵Wilco Dijkstra2014-10-242-8/+7
| | | | functionality.
* Call libc_feholdexcept_aarch64 from math_private.h rather than duplicating ↵Wilco Dijkstra2014-10-242-24/+7
| | | | functionality.
* Call get_rounding_mode rather than duplicating functionality.Wilco Dijkstra2014-10-242-4/+7
|
* Cleanup feenableexcept to use the same logic as the ARM version. No ↵Wilco Dijkstra2014-10-242-6/+6
| | | | functional changes.
* Cleanup fedisableexcept to use the same logic as the ARM version. No ↵Wilco Dijkstra2014-10-242-6/+6
| | | | functional changes.
* Cleanup feclearexcept to use the same logic as the ARM version. No ↵Wilco Dijkstra2014-10-242-1/+6
| | | | functional changes.
* Cleanup fesetexceptflag to use the same logic as the ARM version. No ↵Wilco Dijkstra2014-10-242-2/+8
| | | | functional changes.
* Remove an unused include.Wilco Dijkstra2014-10-242-1/+5
|
* Move get*id and getgroups definitions to syscalls.list (bug 14138).Joseph Myers2014-10-2447-278/+130
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Continuing the move of syscall definitions to syscalls.list, where the removal of support for old kernel versions has made this possible, this patch moves various definitions of get*id functions and getgroups. The previous C definitions were because of the transition to 32-bit uids and gids. Tested for x86. [BZ #14138] * sysdeps/unix/sysv/linux/arm/getegid.c: Remove file. * sysdeps/unix/sysv/linux/arm/geteuid.c: Likewise. * sysdeps/unix/sysv/linux/arm/getgid.c: Likewise. * sysdeps/unix/sysv/linux/arm/getgroups.c: Likewise. * sysdeps/unix/sysv/linux/arm/getresgid.c: Likewise. * sysdeps/unix/sysv/linux/arm/getresuid.c: Likewise. * sysdeps/unix/sysv/linux/arm/getuid.c: Likewise. * sysdeps/unix/sysv/linux/i386/getegid.c: Likewise. * sysdeps/unix/sysv/linux/i386/geteuid.c: Likewise. * sysdeps/unix/sysv/linux/i386/getgid.c: Likewise. * sysdeps/unix/sysv/linux/i386/getgroups.c: Likewise. * sysdeps/unix/sysv/linux/i386/getresgid.c: Likewise. * sysdeps/unix/sysv/linux/i386/getresuid.c: Likewise. * sysdeps/unix/sysv/linux/i386/getuid.c: Likewise. * sysdeps/unix/sysv/linux/m68k/getegid.c: Likewise. * sysdeps/unix/sysv/linux/m68k/geteuid.c: Likewise. * sysdeps/unix/sysv/linux/m68k/getgid.c: Likewise. * sysdeps/unix/sysv/linux/m68k/getgroups.c: Likewise. * sysdeps/unix/sysv/linux/m68k/getresgid.c: Likewise. * sysdeps/unix/sysv/linux/m68k/getresuid.c: Likewise. * sysdeps/unix/sysv/linux/m68k/getuid.c: Likewise. * sysdeps/unix/sysv/linux/s390/s390-32/getegid.c: Likewise. * sysdeps/unix/sysv/linux/s390/s390-32/geteuid.c: Likewise. * sysdeps/unix/sysv/linux/s390/s390-32/getgid.c: Likewise. * sysdeps/unix/sysv/linux/s390/s390-32/getgroups.c: Likewise. * sysdeps/unix/sysv/linux/s390/s390-32/getresgid.c: Likewise. * sysdeps/unix/sysv/linux/s390/s390-32/getresuid.c: Likewise. * sysdeps/unix/sysv/linux/s390/s390-32/getuid.c: Likewise. * sysdeps/unix/sysv/linux/sh/getegid.c: Likewise. * sysdeps/unix/sysv/linux/sh/geteuid.c: Likewise. * sysdeps/unix/sysv/linux/sh/getgid.c: Likewise. * sysdeps/unix/sysv/linux/sh/getgroups.c: Likewise. * sysdeps/unix/sysv/linux/sh/getresgid.c: Likewise. * sysdeps/unix/sysv/linux/sh/getresuid.c: Likewise. * sysdeps/unix/sysv/linux/sh/getuid.c: Likewise. * sysdeps/unix/sysv/linux/sparc/sparc32/getegid.c: Likewise. * sysdeps/unix/sysv/linux/sparc/sparc32/geteuid.c: Likewise. * sysdeps/unix/sysv/linux/sparc/sparc32/getgid.c: Likewise. * sysdeps/unix/sysv/linux/sparc/sparc32/getgroups.c: Likewise. * sysdeps/unix/sysv/linux/sparc/sparc32/getuid.c: Likewise. * sysdeps/unix/sysv/linux/arm/syscalls.list (getegid): Add syscall. (geteuid): Likewise. (getgid): Likewise. (getuid): Likewise. (getresgid): Likewise. (getresuid): Likewise. (getgroups): Likewise. * sysdeps/unix/sysv/linux/i386/syscalls.list (getegid): Likewise. (geteuid): Likewise. (getgid): Likewise. (getuid): Likewise. (getresgid): Likewise. (getresuid): Likewise. (getgroups): Likewise. * sysdeps/unix/sysv/linux/m68k/syscalls.list (getegid): Likewise. (geteuid): Likewise. (getgid): Likewise. (getuid): Likewise. (getresgid): Likewise. (getresuid): Likewise. (getgroups): Likewise. * sysdeps/unix/sysv/linux/s390/s390-32/syscalls.list (getegid): Likewise. (geteuid): Likewise. (getgid): Likewise. (getuid): Likewise. (getresgid): Likewise. (getresuid): Likewise. (getgroups): Likewise. * sysdeps/unix/sysv/linux/sh/syscalls.list (getegid): Likewise. (geteuid): Likewise. (getgid): Likewise. (getuid): Likewise. (getresgid): Likewise. (getresuid): Likewise. (getgroups): Likewise. * sysdeps/unix/sysv/linux/sparc/sparc32/syscalls.list (getegid): Likewise. (geteuid): Likewise. (getgid): Likewise. (getuid): Likewise. (getgroups): Likewise.
* Move some chown / lchown / fchown definitions to syscalls.list (bug 14138).Joseph Myers2014-10-2418-280/+51
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Continuing the move of syscall definitions to syscalls.list, where the removal of support for old kernel versions has made this possible, this patch moves various definitions of chown, lchown and fchown. In most cases the need for special syscalls.list entries (rather than existing generic ones) is because these architectures use chown32, lchown32 and fchown32 as syscall names. Some architectures also have symbol versioning compatibility for older versions of chown having been equivalent to lchown. The aliases specified for s390-32 had the effect of exporting __chown@@GLIBC_2.1 (but not __chown@GLIBC_2.0) despite it not being listed in Versions files. (I'm not sure why versioned_symbol but not compat_symbol were effective like that to create such __chown exports in the absence of Versions entries.) The natural way to preserve that versioned export of __chown seems to be to add it in a Versions file, so I did so. (Maybe actually it should be a compat symbol, __chown@GLIBC_2.1, unless there's a good reason for that export, but this patch doesn't change anything there.) Tested for x86. [BZ #14138] * sysdeps/unix/sysv/linux/i386/chown.c: Remove file. * sysdeps/unix/sysv/linux/i386/fchown.c: Likewise. * sysdeps/unix/sysv/linux/i386/lchown.c: Likewise. * sysdeps/unix/sysv/linux/s390/s390-32/chown.c: Likewise. * sysdeps/unix/sysv/linux/s390/s390-32/fchown.c: Remove file. * sysdeps/unix/sysv/linux/s390/s390-32/lchown.c: Likewise. * sysdeps/unix/sysv/linux/sh/chown.c: Likewise. * sysdeps/unix/sysv/linux/sh/fchown.c: Likewise. * sysdeps/unix/sysv/linux/sh/lchown.c: Likewise. * sysdeps/unix/sysv/linux/sparc/sparc32/chown.c: Likewise. * sysdeps/unix/sysv/linux/sparc/sparc32/fchown.c: Likewise. * sysdeps/unix/sysv/linux/sparc/sparc32/lchown.c: Likewise. * sysdeps/unix/sysv/linux/s390/s390-32/Versions (GLIBC_2.1): Add __chown. * sysdeps/unix/sysv/linux/i386/syscalls.list (chown): Add syscall. (lchown): Likewise. (fchown): Likewise. * sysdeps/unix/sysv/linux/s390/s390-32/syscalls.list (chown): Likewise. (lchown): Likewise. (fchown): Likewise. * sysdeps/unix/sysv/linux/sh/syscalls.list (chown): Likewise. (lchown): Likewise. (fchown): Likewise. * sysdeps/unix/sysv/linux/sparc/sparc32/syscalls.list (chown): Likewise. (lchown): Likewise. (fchown): Likewise.
* Remove spaces.Wilco Dijkstra2014-10-245-4/+7
|
* hppa: Make __SIGRTMIN 32 (ABI break).Carlos O'Donell2014-10-233-8/+24
| | | | | | | | | | | | | | | | | | | | | | | | | | In the Linux kernel version 3.17 the signal numbers were rearranged in order to make hppa like every other arch. Previously we started __SIGRTMIN at 37, and that meant several pieces of important software, including systemd, would fail to build. To support systemd we removed SIGEMT and SIGLOST, and rearranged the others according to expected values. This is technically an ABI incompatible change, but because zero applications use SIGSTKFLT, SIGXCPU, SIGXFSZ and SIGSYS nothing broke. Nothing uses SIGEMT and SIGLOST, and they were present for HPUX compatibility which is no longer supported. Thus because nothing breaks we don't do any compatibility work here. Upstream kernel commit is 1f25df2eff5b25f52c139d3ff31bc883eee9a0ab. Signed-off-by: Carlos O'Donell <carlos@systemhalted.org> Signed-off-by: Helge Deller <deller@gmx.de> 2014-10-23 Carlos O'Donell <carlos@systemhalted.org> Helge Deller <deller@gmx.de> [BZ #17508] * sysdeps/unix/sysv/linux/hppa/bits/signum.h: Remove SIGEMT. Define SIGSTKFLT as 7. Define SIGSYS as 31. Define SIGXCPU as 12. Remove SIGLOST. Define SIGXFSZ as 30. Define __SIGRTMIN as 32.
* Don't use INTDEF for powerpc32 compat symbols (bug 14132).Joseph Myers2014-10-232-14/+16
| | | | | | | | | | | | | | | | | | | | | | | | | | Continuing the removal of the obsolete INTDEF / INTUSE mechanism, this patch removes the use of INTUSE to rename symbols in sysdeps/powerpc/powerpc32/libgcc-compat.S. As the names in question are purely internal to this particular object and not used anywhere else, it doesn't matter at all whether __*_v_glibc20 or __*_internal is used, so this patch just removes the macros in question. Tested for powerpc32 that stripped installed shared libraries are unchanged by this patch. [BZ #14132] * sysdeps/powerpc/powerpc32/libgcc-compat.S (__ashldi3_v_glibc20): Remove macro definition. (__ashrdi3_v_glibc20): Likewise. (__lshrdi3_v_glibc20): Likewise. (__cmpdi2_v_glibc20): Likewise. (__ucmpdi2_v_glibc20): Likewise. [!_SOFT_FLOAT && !__NO_FPRS__] (__fixdfdi_v_glibc20): Likewise. [!_SOFT_FLOAT && !__NO_FPRS__] (__fixsfdi_v_glibc20): Likewise. [!_SOFT_FLOAT && !__NO_FPRS__] (__fixunsdfdi_v_glibc20): Likewise. [!_SOFT_FLOAT && !__NO_FPRS__] (__fixunssfdi_v_glibc20): Likewise. [!_SOFT_FLOAT && !__NO_FPRS__] (__floatdidf_v_glibc20): Likewise. [!_SOFT_FLOAT && !__NO_FPRS__] (__floatdisf_v_glibc20): Likewise.
* Rework some nscd code not to use variable-length struct types.Roland McGrath2014-10-225-93/+117
|
* ARM: Use movw/movt more when availableRoland McGrath2014-10-229-69/+209
|
* MIPS: Avoid a dangling `vfork@GLIBC_2.0' referenceMaciej W. Rozycki2014-10-223-1/+8
| | | | | | | | | | | | | | | | | | | | This satisfies a symbol reference created with: .symver __libc_vfork, vfork@GLIBC_2.0 where `__libc_vfork' has not been defined or referenced. In this case the `vfork@GLIBC_2.0' reference is supposed to be discarded, however a bug present in GAS since forever causes an undefined symbol table entry to be created. This in turn triggers a problem in the linker that can manifest itself by link errors such as: ld: libpthread.so: invalid string offset 2765592330 >= 5154 for section `.dynstr' The GAS and linker bugs need to be resolved, but we can avoid them too by providing a `__libc_vfork' definition just like our other platforms. [BZ #17485] * sysdeps/unix/sysv/linux/mips/vfork.S (__libc_vfork): Define.
* Don't use INTDEF with __ldexpf (bug 14132).Joseph Myers2014-10-212-1/+5
| | | | | | | | | | | | Continuing the removal of the obsolete INTDEF / INTUSE mechanism, this patch removes the use of INTDEF for __ldexpf. As far as I can tell, the resulting alias is completely unused. Tested for x86_64 that stripped installed shared libraries are unchanged by this patch. [BZ #14132] * math/s_ldexpf.c (__ldexpf): Do not use INTDEF.
* Avoid local PLT reference in __nptl_main.Roland McGrath2014-10-212-1/+5
|
* Tiny refactoring in fts to eliminate a warning.Roland McGrath2014-10-202-5/+16
|
* NPTL: Clean up gratuitous Linuxism in libpthread.so entry point.Roland McGrath2014-10-202-7/+10
|
* NPTL: Add some missing #include'sRoland McGrath2014-10-206-0/+12
|
* Make internal lock-init macros return void.Roland McGrath2014-10-204-6/+15
|
* pthread_once: Add fast path and remove x86 variants.Torvald Riegel2014-10-204-378/+30
|
* pthread_once: Clean up constants.Torvald Riegel2014-10-204-10/+29
| | | | [BZ #15215] This just gives a name to the integer constants being used.
* Move readv and writev definitions to syscalls.list (bug 14138).Joseph Myers2014-10-204-105/+11
| | | | | | | | | | | | | | | | | | | | | | | | Continuing the move of syscall definitions to syscalls.list, where the removal of support for old kernel versions has made this possible, this patch moves definitions of readv and writev. The relevant syscalls.list entries were already in sysdeps/unix/syscalls.list, but to match the C files they needed to have the names __libc_readv and __libc_writev added. In fact, I don't see anything making use of those names - as far as I can tell, these functions could just be defined as __readv and __writev with aliases readv and writev. But cleaning up unnecessary aliases for functions should be a separate matter from cleaning up unnecessary C syscall wrappers. Tested for x86_64. [BZ #14138] * sysdeps/unix/syscalls.list (readv): Use __libc_readv as strong name. (writev): Use __libc_writev as strong name. * sysdeps/unix/sysv/linux/readv.c: Remove file. * sysdeps/unix/sysv/linux/writev.c: Likewise.
* Remove obsolete TLS_DEFINE_INIT_TP fallback.Roland McGrath2014-10-172-13/+2
|
* NPTL: Clean up THREAD_SYSINFO macros.Roland McGrath2014-10-176-10/+26
|
* NPTL: Conditionalize direct futex syscall uses.Roland McGrath2014-10-177-22/+62
|
* NPTL: Conditionalize more uses of SIGCANCEL and SIGSETXID.Roland McGrath2014-10-174-4/+33
|
* Fix NPTL build error when missing __NR_set_robust_list.Roland McGrath2014-10-172-0/+7
|
* Fix up incorrect formatting in last commitSiddhesh Poyarekar2014-10-172-4/+9
|
* strcoll: improve performance by removing the cache (#15884)Leonhard Holz2014-10-173-319/+39
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | this is a path that should solve bug 15884. It complains about the performance of strcoll(). It was found out that the runtime of strcoll() is actually bound to strlen which is needed for calculating the size of a cache that was installed to improve the comparison performance. The idea for this patch was that the cache is only useful in rare cases (strings of same length and same first-level-chars) and that it would be better to avoid memory allocation at all. To prove this I wrote a performance test bench-strcoll.c with test data in benchtests-strcoll.tar.gz. Also modifications in benchtests/Makefile and localedata/Makefile are necessary to make it work. After removing the cache the strcoll method showed the predicted behavior (getting slightly faster) in all but the test case for hindi word sorting. This was due the hindi text having much more equal words than the other ones. For equal strings the performance was worse since all comparison levels were run through and from the second level on the cache improved the comparison performance of the original version. Therefore I added a bytewise test via strcmp iff the first level comparison found that both strings did match because in this case it is very likely that equal strings are compared. This solved the problem with the hindi test case and improved the performance of the others. Performance comparison: glibc files -33.77% vi_VN.UTF-8 -34.12% en_US.UTF-8 -42.42% ar_SA.UTF-8 -27.49% zh_CN.UTF-8 +07.90% cs_CZ.UTF-8 -29.67% en_GB.UTF-8 -28.50% da_DK.UTF-8 -36.57% pl_PL.UTF-8 -39.31% fr_FR.UTF-8 -28.57% pt_PT.UTF-8 -22.82% el_GR.UTF-8 -26.77% ru_RU.UTF-8 -35.81% iw_IL.UTF-8 -35.34% es_ES.UTF-8 -34.46% hi_IN.UTF-8 -00.38% sv_SE.UTF-8 -36.99% hu_HU.UTF-8 -16.35% tr_TR.UTF-8 -27.80% is_IS.UTF-8 -33.24% it_IT.UTF-8 -24.39% sr_RS.UTF-8 -37.55% ja_JP.UTF-8 +02.84%
generated by cgit-pink 1.4.1 (git 2.36.1) at 2024-12-15 13:01:06 +0000