Commit Graph

189 Commits

Author SHA1 Message Date
Rich Felker
b1c43161c2 unify lock and owner fields of mutex structure
this change is necessary to free up one slot in the mutex structure so
that we can use doubly-linked lists in the implementation of robust
mutexes.
2011-03-17 12:21:32 -04:00
Rich Felker
e914f8b7ec optimize contended normal mutex case; add int compare-and-swap atomic 2011-03-17 12:14:40 -04:00
Rich Felker
1d59f1eddb simplify logic, slightly optimize contended case for non-default mutex types 2011-03-16 16:49:42 -04:00
Rich Felker
d4f9e0b364 correct error returns for error-checking mutexes 2011-03-16 16:25:00 -04:00
Rich Felker
29fae65780 cut out a syscall on thread creation in the case where guard size is 0 2011-03-16 11:36:21 -04:00
Rich Felker
cc832d8a31 don't expose EAGAIN, etc. from timed futex wait to caller 2011-03-16 11:35:46 -04:00
Rich Felker
4d9cc0b399 optimize file locking: avoid cache-polluting writes to global storage 2011-03-16 10:39:45 -04:00
Rich Felker
9dd7d7e3f6 partially-written draft of fmemopen, still in #if 0 2011-03-14 11:49:55 -04:00
Rich Felker
b446c55b6f remove some old cruft from sys/types.h 2011-03-14 11:49:17 -04:00
Rich Felker
eff62b8a96 misplaced & in times() made it fail to work, and clobber the stack 2011-03-12 22:53:17 -05:00
Rich Felker
5eb0d33ec0 implement flockfile api, rework stdio locking 2011-03-12 21:55:45 -05:00
Rich Felker
d8d19f4d46 pthread.h needs clockid_t
actually it gets this from time.h if _POSIX_C_SOURCE or any other
feature test macros are defined, but it breaks if they're not.
2011-03-12 21:54:19 -05:00
Rich Felker
9f50b77d2b document some additional important changes 2011-03-11 14:51:36 -05:00
Rich Felker
6f05b6bbd5 update whatsnew file for release of 0.7.0 2011-03-11 13:38:09 -05:00
Rich Felker
8bb0e48889 match dimensions so we can use all slots without invoking OOB-array-access 2011-03-11 10:02:17 -05:00
Rich Felker
243b542382 implement dummy pthread_attr_[gs]etschedparam functions
for some reason these functions are not shaded by the PS/TPS option in
POSIX, so presumably they are mandatory, even though the functionality
they offer is optional. for now, provide them in case any programs
depend on their existence, but disallow any priority except the
default.
2011-03-11 09:51:54 -05:00
Rich Felker
1659aa0e62 fix pthread_attr_* implementations to match corrected prototypes 2011-03-11 09:51:14 -05:00
Rich Felker
cabf2ff349 fix missing ENOTSUP error code 2011-03-11 09:50:54 -05:00
Rich Felker
f1821fce81 missing const in some pthread_attr_* prototypes 2011-03-11 09:46:31 -05:00
Rich Felker
2676198733 formatting whatsnew file 2011-03-11 09:46:12 -05:00
Rich Felker
8d07a0350e list major changes in preparation for release 2011-03-11 00:48:40 -05:00
Rich Felker
682a0f271f fix failure behavior of sem_open when sem does not exist 2011-03-10 22:05:16 -05:00
Rich Felker
cfe581b6bc fix some semaphore wait semantics (race condition deadlock and error checking) 2011-03-10 21:52:18 -05:00
Rich Felker
81af503610 fix sem_open and sem_close to obey posix semantics
multiple opens of the same named semaphore must return the same
pointer, and only the last close can unmap it. thus the ugly global
state keeping track of mappings. the maximum number of distinct named
semaphores that can be opened is limited sufficiently small that the
linear searches take trivial time, especially compared to the syscall
overhead of these functions.
2011-03-10 21:34:19 -05:00
Rich Felker
5fcebcde6a optimize pthread termination in the non-detached case
we can avoid blocking signals by simply using a flag to mark that the
thread has exited and prevent it from getting counted in the rsyscall
signal-pingpong. this restores the original pthread create/join
throughput from before the sigprocmask call was added.
2011-03-10 18:31:37 -05:00
Rich Felker
dc54a7cbb9 fix errors in sigqueue (potential information leak, wrong behavior)
1. any padding in the siginfo struct was not necessarily zero-filled,
so it might have contained private data off the caller's stack.

2. the uid and pid must be filled in from userspace. the previous
rsyscall fix broke rsyscalls because the values were always incorrect.
2011-03-10 18:26:29 -05:00
Rich Felker
52213f7341 security fix: check that cancel/rsyscall signal was sent by the process itself 2011-03-10 11:59:39 -05:00
Rich Felker
11dbbe9fba more cancellation points: tcdrain, clock_nanosleep 2011-03-10 11:06:50 -05:00
Rich Felker
c8c4ef7d44 remove useless return value checks for functions that cannot fail 2011-03-10 11:02:29 -05:00
Rich Felker
0f8cc94d68 make sigsuspend a cancellation point 2011-03-10 11:01:11 -05:00
Rich Felker
5b9429adb7 make sigtimedwait a cancellation point 2011-03-10 10:59:50 -05:00
Rich Felker
0a949ebdf0 don't fail with EINTR in sigtimedwait
POSIX allows either behavior, but sigwait is not allowed to fail with
EINTR, so the retry loop would have to be in one or the other anyway.
2011-03-10 10:43:09 -05:00
Rich Felker
0558683d3d fix sigsuspend syscall 2011-03-10 10:26:16 -05:00
Rich Felker
6871fd773d make sigaltstack work (missing macros in signal.h, error conditions) 2011-03-10 10:17:29 -05:00
Rich Felker
1b538acb0a fix errno behavior in clock_* functions
these functions are specified inconsistent in whether they're
specified to return an error value, or return -1 and set errno.
hopefully now they all match what POSIX requires.
2011-03-10 09:54:23 -05:00
Rich Felker
500c969f05 fix error handling for pthread_sigmask
it must return errno, not -1, and should reject invalud values for how.
2011-03-09 20:31:06 -05:00
Rich Felker
3f5420bcda make fork properly initialize the main thread in the child process 2011-03-09 20:23:44 -05:00
Rich Felker
6dc05fbe19 optimize pthread initialization
the set_tid_address returns the tid (which is also the pid when called
from the initial thread) so there is no need to make a separate
syscall to get pid/tid.
2011-03-09 20:21:23 -05:00
Rich Felker
0bed7e0acf fix race condition in raise - just mask signals
a signal handler could fork after the pid/tid were read, causing the
wrong process to be signalled. i'm not sure if this is supposed to
have UB or not, but raise is async-signal-safe, so it probably is
allowed. the current solution is slightly expensive so this
implementation is likely to be changed in the future.
2011-03-09 20:07:24 -05:00
Rich Felker
370f78f2c8 fix raise semantics with threads. 2011-03-09 19:42:06 -05:00
Rich Felker
91f7db2621 fix typos in dirent.h 2011-03-08 17:34:26 -05:00
Rich Felker
ac12eb91bd fcntl.h should make availabel the mode constants from sys/stat.h
also remove some legacy nonsense that crept in.
2011-03-08 12:35:35 -05:00
Rich Felker
31e06075d5 simplify and optimize pthread_mutex_trylock 2011-03-08 12:20:10 -05:00
Rich Felker
e5dd18319b rwlock trylock functions were wrongly returning EAGAIN instead of EBUSY 2011-03-08 12:19:30 -05:00
Rich Felker
7e6be42a77 fix major breakage in pthread_once (it was always deadlocking)
the issue was a break statement that was breaking only from the
switch, not the enclosing for loop, and a failure to set the final
success state.
2011-03-08 12:08:40 -05:00
Rich Felker
4820f9268d fix and optimize non-default-type mutex behavior
problem 1: mutex type from the attribute was being ignored by
pthread_mutex_init, so recursive/errorchecking mutexes were never
being used at all.

problem 2: ownership of recursive mutexes was not being enforced at
unlock time.
2011-03-08 03:41:05 -05:00
Rich Felker
1d6b1f1592 implement pthread_mutexattr_[gs]etpshared functions 2011-03-08 02:33:37 -05:00
Rich Felker
9d5251f72b disallow cpu time clocks as condattr clock values 2011-03-08 02:32:42 -05:00
Rich Felker
5d0d6d9836 add macros for use with d_type extension field in dirent 2011-03-07 18:32:36 -05:00
Rich Felker
9fe28f703d fix off-by-one error in sem_(timed)wait (using old sem value instead of new) 2011-03-07 18:09:24 -05:00