summaryrefslogtreecommitdiff
path: root/src
AgeCommit message (Collapse)AuthorLines
2011-04-17optimize cancellation enable/disable codeRich Felker-4/+11
the goal is to be able to use pthread_setcancelstate internally in the implementation, whenever a function might want to use functions which are cancellation points but avoid becoming a cancellation point itself. i could have just used a separate internal function for temporarily inhibiting cancellation, but the solution in this commit is better because (1) it's one less implementation-specific detail in functions that need to use it, and (2) application code can also get the same benefit. previously, pthread_setcancelstate dependend on pthread_self, which would pull in unwanted thread setup overhead for non-threaded programs. now, it temporarily stores the state in the global libc struct if threads have not been initialized, and later moves it if needed. this way we can instead use __pthread_self, which has no dependencies and assumes that the thread register is already valid.
2011-04-17don't use pthread_once when there is no danger in raceRich Felker-2/+5
2011-04-17fix some minor issues in cancellation handling patchRich Felker-11/+19
signals were wrongly left masked, and cancellability state was not switched to disabled, during the execution of cleanup handlers.
2011-04-17overhaul pthread cancellationRich Felker-218/+235
this patch improves the correctness, simplicity, and size of cancellation-related code. modulo any small errors, it should now be completely conformant, safe, and resource-leak free. the notion of entering and exiting cancellation-point context has been completely eliminated and replaced with alternative syscall assembly code for cancellable syscalls. the assembly is responsible for setting up execution context information (stack pointer and address of the syscall instruction) which the cancellation signal handler can use to determine whether the interrupted code was in a cancellable state. these changes eliminate race conditions in the previous generation of cancellation handling code (whereby a cancellation request received just prior to the syscall would not be processed, leaving the syscall to block, potentially indefinitely), and remedy an issue where non-cancellable syscalls made from signal handlers became cancellable if the signal handler interrupted a cancellation point. x86_64 asm is untested and may need a second try to get it right.
2011-04-15remove stupid debug code in wordexpRich Felker-1/+0
2011-04-15implement wordexp. first try, may be buggy. intended to be safe.Rich Felker-0/+128
2011-04-15avoid setting errno when checking for ttyRich Felker-2/+2
setting errno here is completely valid, but some programs, notably busybox printf, assume that errno will not be set during output and treat this as an error condition. in any case, skipping it slightly reduces code size and saves time.
2011-04-14make tmpfile slightly more efficient (use unlink syscall instead of remove)Rich Felker-1/+1
2011-04-14change sem_trywait algorithm so it never has to call __wakeRich Felker-3/+2
2011-04-14cheap trick to further optimize locking normal mutexesRich Felker-2/+2
2011-04-14use a separate signal from SIGCANCEL for SIGEV_THREAD timersRich Felker-11/+29
otherwise we cannot support an application's desire to use asynchronous cancellation within the callback function. this change also slightly debloats pthread_create.c.
2011-04-13simplify cancellation point handlingRich Felker-16/+5
we take advantage of the fact that unless self->cancelpt is 1, cancellation cannot happen. so just increment it by 2 to temporarily block cancellation. this drops pthread_create.o well under 1k.
2011-04-13simplify syslog, add vsyslog interface (nonstandard)Rich Felker-31/+36
with datagram sockets, depending on fprintf not to flush the output early was very fragile; the new version simply uses a small fixed-size buffer. it could be updated to dynamic-allocate large buffers if needed, but i can't envision any admin being happy about finding 64kb-long lines in their syslog...
2011-04-13remove useless SIGPIPE protection from syslogRich Felker-9/+0
per the standard, SIGPIPE is not generated for SOCK_DGRAM.
2011-04-13fix syslog (corrected SIGPIPE blocking, and using dgram instead of stream)Rich Felker-10/+8
it actually appears the hacks to block SIGPIPE are probably not necessary, and potentially harmful. if i can confirm this, i'll remove them.
2011-04-13numerous fixes to sysv ipcRich Felker-5/+5
some of these definitions were just plain wrong, others based on outdated ancient "non-64" versions of the kernel interface. as much as possible has now been moved out of bits/* these changes break abi (the old abi for these functions was wrong), but since they were not working anyway it can hardly matter.
2011-04-13add syscall wrapper for flockRich Felker-0/+7
it should be noted that flock does not mix well with standard fcntl locking, but nonetheless some applications will attempt to use flock instead of fcntl if both exist. options to configure or small patches may be needed. debian maintainers have plenty of experience with this unfortunate situation...
2011-04-13fix bug whereby getopt_long allowed mismatch in last char of option nameRich Felker-1/+1
2011-04-13implement getgrouplist (for initgroups), formerly dummied-outv0.7.8Rich Felker-4/+16
2011-04-13fix prototypes/signature for setgroups, etc.Rich Felker-1/+1
2011-04-13implement memrchr (nonstandard) and optimize strrchr in terms of itRich Felker-4/+15
2011-04-13add ptsname_r (nonstandard) and split ptsname (standard) to separate fileRich Felker-11/+16
this eliminates the ugly static buffer in programs that use ptsname_r.
2011-04-12speed up threaded forkRich Felker-2/+1
after fork, we have a new process and the pid is equal to the tid of the new main thread. there is no need to make two separate syscalls to obtain the same number.
2011-04-12optimize ntohl etc. in terms of bswap functionsRich Felker-20/+12
we can do this without violating the namespace now that they are macros/inline functions rather than extern functions. the motivation is that gcc was generating giant, slow, horrible code for the old functions, and now generates a single byte-swapping instruction.
2011-04-12move bswap functions to static inline in byteswap.hRich Felker-16/+0
2011-04-12fix broken bswap_32Rich Felker-1/+1
2011-04-12fix printf("%.9g", 1.1) and similar not dropping trailing zerosRich Felker-1/+3
2011-04-11fix errno handling in scandir:Rich Felker-2/+1
1. saved errno was not being restored, illegally clearing errno to 0. 2. no need to backup and save errno around free; it will not touch except perhaps when the program has already invoked UB...
2011-04-11fix fputwc return valueRich Felker-1/+1
2011-04-11remove ugly warning-suppression hack from crypt - this invokes UB!Rich Felker-1/+1
2011-04-09run pthread tsd destructors when a timer thread pretends to exitRich Felker-0/+6
2011-04-09greatly improve SIGEV_THREAD timersRich Felker-15/+21
calling pthread_exit from, or pthread_cancel on, the timer callback thread will no longer destroy the timer.
2011-04-09work around a nasty bug in linux readv syscallRich Felker-3/+4
according to posix, readv "shall be equivalent to read(), except..." that it places the data into the buffers specified by the iov array. however on linux, when reading from a terminal, each iov element behaves almost like a separate read. this means that if the first iov exactly satisfied the request (e.g. a length-one read of '\n') and the second iov is nonzero length, the syscall will block again after getting the blank line from the terminal until another line is read. simply put, entering a single blank line becomes impossible. the solution, fortunately, is simple. whenever the buffer size is nonzero, reduce the length of the requested read by one byte and let the last byte go through the buffer. this way, readv will already be in the second (and last) iov, and won't re-block on the second iov.
2011-04-08better fix sysconf pthread stack minRich Felker-1/+1
2011-04-08consistency with pthread stack min in limits.hRich Felker-1/+1
2011-04-08fix broken sigsetjmp on x86_64Rich Felker-7/+9
2011-04-08workaround broken msghdr struct on 64bit linuxRich Felker-0/+21
POSIX clearly specifies the type of msg_iovlen and msg_controllen, and Linux ignores it and makes them both size_t instead. to work around this we add padding (instead of just using the wrong types like glibc does), but we also need to patch-up the struct before passing it to the kernel in case the caller did not zero-fill it. if i could trust the kernel to just ignore the upper 32 bits, this would not be necessary, but i don't think it will ignore them...
2011-04-08fix ipv6 address printing: 2001 appeared as 201, etc.Rich Felker-3/+5
2011-04-08fix broken dns response parsing code that made most ipv6 lookups failRich Felker-4/+6
2011-04-08return the requested string as the "canonical name" for numeric addressesRich Felker-0/+1
previously NULL was returned in ai_canonname, resulting in crashes in some callers. this behavior was incorrect. note however that the new behavior differs from glibc, which performs reverse dns lookups. POSIX is very clear that a reverse DNS lookup must not be performed for numeric addresses.
2011-04-07fix uninitialized variables in dns lookup codeRich Felker-2/+2
2011-04-07fix bug in TRE found by clang (typo && instead of &)Rich Felker-1/+1
2011-04-07fix misplaced *'s in string functions (harmless)Rich Felker-3/+3
2011-04-07fix broken unsigned comparison in wcstoumaxRich Felker-1/+1
2011-04-07fix breakage due to converting a return type to size_t in iconv...Rich Felker-1/+1
2011-04-06fixed crash in new rsyscall (failure to set sa_flags for signal handler)Rich Felker-0/+2
2011-04-06consistency: change all remaining syscalls to use SYS_ rather than __NR_ prefixRich Felker-40/+40
2011-04-06move rsyscall out of pthread_create moduleRich Felker-109/+133
this is something of a tradeoff, as now set*id() functions, rather than pthread_create, are what pull in the code overhead for dealing with linux's refusal to implement proper POSIX thread-vs-process semantics. my motivations are: 1. it's cleaner this way, especially cleaner to optimize out the rsyscall locking overhead from pthread_create when it's not needed. 2. it's expected that only a tiny number of core system programs will ever use set*id() functions, whereas many programs may want to use threads, and making thread overhead tiny is an incentive for "light" programs to try threads.
2011-04-06pthread exit stuff: don't bother setting errno when we won't check it.Rich Felker-2/+2
2011-04-06fix rsyscall handler: must not clobber errno from signal contextRich Felker-2/+4