about summary refs log tree commit diff
path: root/sysdeps/unix/sysv/linux
Commit message (Collapse)AuthorAgeFilesLines
* Add adaptive elision to rwlocksAndi Kleen2014-06-139-0/+18
| | | | | | | | | | | | | | | | | | | This patch relies on the C version of the rwlocks posted earlier. With C rwlocks it is very straight forward to do adaptive elision using TSX. It is based on the infrastructure added earlier for mutexes, but uses its own elision macros. The macros are fairly general purpose and could be used for other elision purposes too. This version is much cleaner than the earlier assembler based version, and in particular implements adaptation which makes it safer. I changed the behavior slightly to not require any changes in the test suite and fully conform to all expected behaviors (generally at the cost of not eliding in various situations). In particular this means the timedlock variants are not elided. Nested trylock aborts.
* get_nprocs: Only return explictly set cache values (BZ #16996)Meador Inge2014-06-131-2/+2
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The implementation of __get_nprocs uses a stactic variable to cache the value of the current number of processors. The caching breaks when 'time (NULL) == 0': $ cat nproc.c #include <stdio.h> #include <time.h> #include <sys/time.h> int main(int argc, char *argv[]) { time_t t; struct timeval tv = {0, 0}; printf("settimeofday({0, 0}, NULL) = %d\n", settimeofday(&tv, NULL)); t = time(NULL); printf("Time: %d, CPUs: %d\n", (unsigned int)t, get_nprocs()); return 0; } $ gcc -O3 nproc.c $ ./a.out settimeofday({0, 0}, NULL) = -1 Time: 1401311578, CPUs: 4 $ sudo ./a.out settimeofday({0, 0}, NULL) = 0 Time: 0, CPUs: 0 The problem is with the condition used to check whether a cached value should be returned or not: static int cached_result; static time_t timestamp; time_t now = time (NULL); time_t prev = timestamp; atomic_read_barrier (); if (now == prev) return cached_result; This patch fixes the problem by ensuring that 'cached_result' has been set at least once before returning it.
* Move SPARC public headers out of nptl/Roland McGrath2014-06-121-0/+99
|
* SPARC: Convert fork.c to arch-fork.hRoland McGrath2014-06-121-0/+27
|
* Get rid of nptl/sysdeps/pthread/ subdirectoryRoland McGrath2014-06-122-0/+60
|
* Consolidate NPTL sigprocmask.Roland McGrath2014-06-121-1/+2
|
* Move SH code out of nptl/ subdirectory.Roland McGrath2014-06-1223-0/+6138
|
* tile: Consolidate NPTL/non versions of vforkChris Metcalf2014-06-113-6/+2
|
* Tile: Define TLS_DEFINE_INIT_TPRoland McGrath2014-06-111-24/+0
|
* Alpha: Define TLS_DEFINE_INIT_TPRoland McGrath2014-06-111-22/+0
|
* PowerPC: Define TLS_DEFINE_INIT_TPRoland McGrath2014-06-111-24/+0
|
* AArch64: Define TLS_DEFINE_INIT_TPRoland McGrath2014-06-111-23/+0
|
* tile: move sysdeps/unix/sysv/linux/tile nptl files.Chris Metcalf2014-06-1053-269/+5
|
* hppa: Remove lowlevellock.c.Bernard Ogden2014-06-091-126/+0
| | | | | | The hppa port has no need of a custom lowlevellock.c, it should use the generic version which is updated and correct. This similarly fixes bug 15119 for hppa.
* m68k: Define TLS_DEFINE_INIT_TPRoland McGrath2014-06-091-24/+0
|
* HPPA: Define TLS_DEFINE_INIT_TPRoland McGrath2014-06-091-22/+0
|
* MIPS: Define TLS_DEFINE_INIT_TPRoland McGrath2014-06-091-23/+0
|
* ARM: Define TLS_DEFINE_INIT_TPRoland McGrath2014-06-091-22/+0
|
* Start cleaning up TLS initial value for pthread_create.Roland McGrath2014-06-091-48/+0
|
* Fix memory overrun in getifaddrs_internal. Fixes bug 15698.Ondřej Bílka2014-06-051-2/+2
|
* aarch64: Remove nptl/vfork.SRichard Henderson2014-06-031-37/+0
|
* aarch64: Consolidate NPTL/non versions of vforkRichard Henderson2014-06-033-44/+74
| | | | | At the same time, incorporate the 0 -> 0x80000000 mapping of the pid expected by raise.c.
* aarch64: Consolidate NPTL/non versions of cloneRichard Henderson2014-06-032-53/+24
| | | | | | At the same time, rely on non-clobbered registers across syscall so that we eliminate the stack frame that we previously ignored in the unwind info.
* SH: Consolidate NPTL/non versions of vforkRoland McGrath2014-06-022-29/+33
|
* aarch64: Remove PSEUDO_RETRichard Henderson2014-05-301-20/+3
|
* aarch64: Fix error return from __ioctlRichard Henderson2014-05-301-6/+7
| | | | | Forgotten in ca3cfa40c16ef34c74951a07a57cfcbcd58898b1 where the error branch was moved from PSEUDO_RET.
* m68k: Convert fork.c to arch-fork.hRoland McGrath2014-05-301-3/+2
|
* aarch64: Rely on syscalls preserving registersRichard Henderson2014-05-291-24/+12
| | | | | | | | In several cases we've had asm routines rely on syscalls not clobbering call-clobbered registers, and that's now deemed ABI. So take advantage of this in the INLINE_SYSCALL path as well. Shrinks libc.so by about 1k.
* aarch64: Use tpidr_el0 rather than __errno_location in librtRichard Henderson2014-05-292-15/+5
|
* aarch64: Use tpidr_el0 rather than __read_tp in librtRichard Henderson2014-05-291-11/+3
|
* aarch64: Improve syscall-cancel stack frameRichard Henderson2014-05-291-86/+26
| | | | One push instead of N; use stp/ldp to halve the insns.
* aarch64: Pass regno parameter to SINGLE_THREAD_PRichard Henderson2014-05-291-10/+8
| | | | Use cbz instead of cmp+b.eq.
* aarch64: Share code in syscall-cancel.hRichard Henderson2014-05-291-14/+14
| | | | Fold nocancel and error handling paths.
* aarch64: Tabify sysdep-cancel.hRichard Henderson2014-05-291-51/+51
|
* Avoid stat/fstat in statvfs/fstatvfs (BZ #15132)Eric Wong2014-05-295-27/+18
| | | | | Delay the use of stat/fstat until stat data is required. When the kernel returns ST_VALID, stat data is not used by __internal_statvfs.
* SH: Consolidate NPTL/non versions of cloneRoland McGrath2014-05-281-6/+1
|
* tile: Convert fork.c to arch-fork.hRoland McGrath2014-05-281-4/+2
|
* IA64: Convert fork.c to arch-fork.hRoland McGrath2014-05-271-0/+29
|
* SH: Convert fork.c to arch-fork.hRoland McGrath2014-05-271-0/+28
|
* PowerPC: Move remaining nptl/sysdeps/unix/sysv/linux/powerpc/ files.Adhemerval Zanella2014-05-2613-0/+944
|
* PowerPC: Move NPTL ABI headers to sysdeps.Adhemerval Zanella2014-05-262-0/+266
|
* PowerPC: Move powerpc64 timer_*.c out of nptl/Adhemerval Zanella2014-05-266-0/+13
|
* PowerPC: Consolidate NPTL/non versions of cloneAdhemerval Zanella2014-05-262-27/+2
|
* PowerPC: Consolidate NPTL/non versions of vforkAdhemerval Zanella2014-05-262-0/+59
|
* arm,aarch64: Remove SINGLE_THREAD_P_PICRichard Henderson2014-05-252-2/+0
| | | | | | | | | | This macro was removed by 2005-11-16 Daniel Jacobowitz <dan@codesourcery.com> but not applied to the (still separate) eabi port so necro'd when the eabi port superceded the old abi. It was thence copied into the new AArch64 port.
* aarch64: Tidy syscall error checkRichard Henderson2014-05-252-10/+13
| | | | | Move the error branch from the PSEUDO_RET macro to the PSEUDO macro. This is in line with other architectures, and will enable further improvments.
* aarch64: Remove DOARGS/UNDOARGS macrosRichard Henderson2014-05-252-24/+1
| | | | While they do something for AArch32, they're useless for AArch64.
* aarch64: Fix DO_CALL block commentRichard Henderson2014-05-251-1/+1
|
* alpha: Do non-default symbols in pt-vfork.SRichard Henderson2014-05-252-6/+14
|
* alpha: Move remaining files out of sysdeps/unix/sysv/linux/alpha/nptl/Richard Henderson2014-05-2332-15/+16
|
generated by cgit-pink 1.4.1 (git 2.36.1) at 2024-09-18 21:10:55 +0000