Rich Felker [Sat, 2 Apr 2011 03:07:03 +0000 (23:07 -0400)]
avoid over-allocation of brk on first malloc
if init_malloc returns positive (successful first init), malloc will
retry getting a chunk from the free bins rather than expanding the
heap again. also pass init_malloc a hint for the size of the initial
allocation.
Rich Felker [Sat, 2 Apr 2011 02:35:20 +0000 (22:35 -0400)]
reorganize the __libc structure for threaded performance issues
we want to keep atomically updated fields (locks and thread count) and
really anything writable far away from frequently-needed function
pointers. stuff some rarely-needed function pointers in between to
pad, hopefully up to a cache line boundary.
Rich Felker [Sat, 2 Apr 2011 02:15:03 +0000 (22:15 -0400)]
simplify setting result on thread cancellation
Rich Felker [Sat, 2 Apr 2011 02:07:59 +0000 (22:07 -0400)]
use bss instead of mmap for main thread's pthread thread-specific data
this simplifies code and removes a failure case
Rich Felker [Sat, 2 Apr 2011 01:10:01 +0000 (21:10 -0400)]
remove obsolete and useless useconds_t type
Rich Felker [Sat, 2 Apr 2011 00:58:40 +0000 (20:58 -0400)]
somehow timespec tv_nsec had the wrong type on x86_64... fixed
Rich Felker [Sat, 2 Apr 2011 00:48:02 +0000 (20:48 -0400)]
fix misspelled PTHREAD_CANCELED constant
Rich Felker [Sat, 2 Apr 2011 00:47:54 +0000 (20:47 -0400)]
document more changes
Rich Felker [Sat, 2 Apr 2011 00:36:01 +0000 (20:36 -0400)]
document changes for upcoming 0.7.5 release
Rich Felker [Fri, 1 Apr 2011 23:53:16 +0000 (19:53 -0400)]
use a_store to set cancel flag in pthread_cancel, to ensure a barrier
Rich Felker [Thu, 31 Mar 2011 23:06:22 +0000 (19:06 -0400)]
simplify pthread_key_delete
calling this function on an uninitialized key value is UB, so there is
no need to check that the table pointer was initialized.
Rich Felker [Thu, 31 Mar 2011 23:04:56 +0000 (19:04 -0400)]
greatly simplify pthread_key_create (~20% size reduction)
Rich Felker [Wed, 30 Mar 2011 18:14:26 +0000 (14:14 -0400)]
add some missing prototypes for nonstandard functions (strsep, clearenv)
Rich Felker [Wed, 30 Mar 2011 17:04:55 +0000 (13:04 -0400)]
avoid all malloc/free in timer creation/destruction
instead of allocating a userspace structure for signal-based timers,
simply use the kernel timer id. we use the fact that thread pointers
will always be zero in the low bit (actually more) to encode integer
timerid values as pointers.
also, this change ensures that the timer_destroy syscall has completed
before the library timer_destroy function returns, in case it matters.
Rich Felker [Wed, 30 Mar 2011 16:06:39 +0000 (12:06 -0400)]
optimize timer creation and possibly protect against some minor races
the major idea of this patch is not to depend on having the timer
pointer delivered to the signal handler, and instead use the thread
pointer to get the callback function address and argument. this way,
the parent thread can make the timer_create syscall while the child
thread is starting, and it should never have to block waiting for the
barrier.
Rich Felker [Wed, 30 Mar 2011 14:32:45 +0000 (10:32 -0400)]
avoid crash on stupid but allowable usage of pthread_mutex_unlock
unlocking an unlocked mutex is not UB for robust or error-checking
mutexes, so we must avoid calling __pthread_self (which might crash
due to lack of thread-register initialization) until after checking
that the mutex is locked.
Rich Felker [Wed, 30 Mar 2011 13:29:49 +0000 (09:29 -0400)]
rename __simple_malloc.c to lite_malloc.c - yes this affects behavior!
why does this affect behavior? well, the linker seems to traverse
archive files starting from its current position when resolving
symbols. since calloc.c comes alphabetically (and thus in sequence in
the archive file) between __simple_malloc.c and malloc.c, attempts to
resolve the "malloc" symbol for use by calloc.c were pulling in the
full malloc.c implementation rather than the __simple_malloc.c
implementation.
as of now, lite_malloc.c and malloc.c are adjacent in the archive and
in the correct order, so malloc.c should never be used to resolve
"malloc" unless it's already needed to resolve another symbol ("free"
or "realloc").
Rich Felker [Wed, 30 Mar 2011 13:06:00 +0000 (09:06 -0400)]
streamline mutex unlock to remove a useless branch, use a_store to unlock
this roughly halves the cost of pthread_mutex_unlock, at least for
non-robust, normal-type mutexes.
the a_store change is in preparation for future support of archs which
require a memory barrier or special atomic store operation, and also
should prevent the possibility of the compiler misordering writes.
Rich Felker [Wed, 30 Mar 2011 12:58:25 +0000 (08:58 -0400)]
cheap special-case optimization for normal mutexes
cycle-level benchmark on atom cpu showed typical pthread_mutex_lock
call dropping from ~120 cycles to ~90 cycles with this change. benefit
may vary with compiler options and version, but this optimization is
very cheap to make and should always help some.
Rich Felker [Wed, 30 Mar 2011 02:43:13 +0000 (22:43 -0400)]
reorder timer initialization so that timer_create does not depend on free
this allows small programs which only create times, but never delete
them, to use simple_malloc instead of the full malloc.
Rich Felker [Tue, 29 Mar 2011 22:30:27 +0000 (18:30 -0400)]
missing prototype for wcscoll (stub)
Rich Felker [Tue, 29 Mar 2011 19:11:25 +0000 (15:11 -0400)]
revert mutex "optimization" that turned out to be worse
Rich Felker [Tue, 29 Mar 2011 17:01:25 +0000 (13:01 -0400)]
implement POSIX timers
this implementation is superior to the glibc/nptl implementation, in
that it gives true realtime behavior. there is no risk of timer
expiration events being lost due to failed thread creation or failed
malloc, because the thread is created as time creation time, and
reused until the timer is deleted.
Rich Felker [Tue, 29 Mar 2011 16:58:22 +0000 (12:58 -0400)]
major improvements to cancellation handling
- there is no longer any risk of spoofing cancellation requests, since
the cancel flag is set in pthread_cancel rather than in the signal
handler.
- cancellation signal is no longer unblocked when running the
cancellation handlers. instead, pthread_create will cause any new
threads created from a cancellation handler to unblock their own
cancellation signal.
- various tweaks in preparation for POSIX timer support.
Rich Felker [Tue, 29 Mar 2011 14:05:57 +0000 (10:05 -0400)]
some preliminaries for adding POSIX timers
Rich Felker [Tue, 29 Mar 2011 13:00:22 +0000 (09:00 -0400)]
fix tempnam name generation, and a small bug in tmpnam on retry limit
Rich Felker [Tue, 29 Mar 2011 12:37:57 +0000 (08:37 -0400)]
make tmpfile fail after exceeding max tries.
Rich Felker [Tue, 29 Mar 2011 12:34:47 +0000 (08:34 -0400)]
fix tmpnam to generate better names, not depend on non-ISO-C symbols
Rich Felker [Tue, 29 Mar 2011 12:25:59 +0000 (08:25 -0400)]
fix messed-up errno if remove fails for a non-EISDIR reason
Rich Felker [Tue, 29 Mar 2011 12:24:28 +0000 (08:24 -0400)]
learned something new - remove is supposed to support directories on POSIX
Rich Felker [Tue, 29 Mar 2011 02:36:55 +0000 (22:36 -0400)]
revert some more spin optimizations that turned out to be pessimizations
Rich Felker [Tue, 29 Mar 2011 02:34:27 +0000 (22:34 -0400)]
fix bug from syscall overhaul: extra __syscall_ret call for 0-arg syscalls
this mainly just caused bloat, but could corrupt errno if a 0-arg
syscall ever failed.
Rich Felker [Tue, 29 Mar 2011 02:22:54 +0000 (22:22 -0400)]
fix broken spinlock due to miscompilation
actually this trick also seems to have made the uncontended case slower.
Rich Felker [Tue, 29 Mar 2011 00:43:51 +0000 (20:43 -0400)]
prototype for getpass
Rich Felker [Tue, 29 Mar 2011 00:29:08 +0000 (20:29 -0400)]
remove useless field in pthread struct (wasted a good bit of space)
Rich Felker [Mon, 28 Mar 2011 21:31:01 +0000 (17:31 -0400)]
fix getc - the classic error of trying to store EOF+0-255 in a char type..
Rich Felker [Mon, 28 Mar 2011 05:14:44 +0000 (01:14 -0400)]
major stdio overhaul, using readv/writev, plus other changes
the biggest change in this commit is that stdio now uses readv to fill
the caller's buffer and the FILE buffer with a single syscall, and
likewise writev to flush the FILE buffer and write out the caller's
buffer in a single syscall.
making this change required fundamental architectural changes to
stdio, so i also made a number of other improvements in the process:
- the implementation no longer assumes that further io will fail
following errors, and no longer blocks io when the error flag is set
(though the latter could easily be changed back if desired)
- unbuffered mode is no longer implemented as a one-byte buffer. as a
consequence, scanf unreading has to use ungetc, to the unget buffer
has been enlarged to hold at least 2 wide characters.
- the FILE structure has been rearranged to maintain the locations of
the fields that might be used in glibc getc/putc type macros, while
shrinking the structure to save some space.
- error cases for fflush, fseek, etc. should be more correct.
- library-internal macros are used for getc_unlocked and putc_unlocked
now, eliminating some ugly code duplication. __uflow and __overflow
are no longer used anywhere but these macros. switch to read or
write mode is also separated so the code can be better shared, e.g.
with ungetc.
- lots of other small things.
Rich Felker [Sat, 26 Mar 2011 02:13:57 +0000 (22:13 -0400)]
match glibc/lsb cancellation abi on i386
glibc made the ridiculous choice to use pass-by-register calling
convention for these functions, which is impossible to duplicate
directly on non-gcc compilers. instead, we use ugly asm to wrap and
convert the calling convention. presumably this works with every
compiler anyone could potentially want to use.
Rich Felker [Fri, 25 Mar 2011 20:50:49 +0000 (16:50 -0400)]
remove -Wno-pointer-sign example from dist/config.mak
Rich Felker [Fri, 25 Mar 2011 20:34:03 +0000 (16:34 -0400)]
fix all implicit conversion between signed/unsigned pointers
sadly the C language does not specify any such implicit conversion, so
this is not a matter of just fixing warnings (as gcc treats it) but
actual errors. i would like to revisit a number of these changes and
possibly revise the types used to reduce the number of casts required.
Rich Felker [Fri, 25 Mar 2011 03:16:52 +0000 (23:16 -0400)]
simplify and optimize FILE lock handling
Rich Felker [Fri, 25 Mar 2011 03:06:48 +0000 (23:06 -0400)]
prepare pthread_spin_unlock for archs that need memory barriers
Rich Felker [Fri, 25 Mar 2011 03:06:08 +0000 (23:06 -0400)]
optimize contended case for pthread_spin_trylock
Rich Felker [Fri, 25 Mar 2011 03:05:17 +0000 (23:05 -0400)]
optimize spinlock spin
Rich Felker [Fri, 25 Mar 2011 02:58:21 +0000 (22:58 -0400)]
fix non-atomicity of puts
Rich Felker [Thu, 24 Mar 2011 18:18:00 +0000 (14:18 -0400)]
overhaul cancellation to fix resource leaks and dangerous behavior with signals
this commit addresses two issues:
1. a race condition, whereby a cancellation request occurring after a
syscall returned from kernelspace but before the subsequent
CANCELPT_END would cause cancellable resource-allocating syscalls
(like open) to leak resources.
2. signal handlers invoked while the thread was blocked at a
cancellation point behaved as if asynchronous cancellation mode wer in
effect, resulting in potentially dangerous state corruption if a
cancellation request occurs.
the glibc/nptl implementation of threads shares both of these issues.
with this commit, both are fixed. however, cancellation points
encountered in a signal handler will not be acted upon if the signal
was received while the thread was already at a cancellation point.
they will of course be acted upon after the signal handler returns, so
in real-world usage where signal handlers quickly return, it should
not be a problem. it's possible to solve this problem too by having
sigaction() wrap all signal handlers with a function that uses a
pthread_cleanup handler to catch cancellation, patch up the saved
context, and return into the cancellable function that will catch and
act upon the cancellation. however that would be a lot of complexity
for minimal if any benefit...
Rich Felker [Wed, 23 Mar 2011 17:24:00 +0000 (13:24 -0400)]
very cheap double-free checks in malloc
Rich Felker [Sun, 20 Mar 2011 04:16:43 +0000 (00:16 -0400)]
global cleanup to use the new syscall interface
Rich Felker [Sun, 20 Mar 2011 03:18:34 +0000 (23:18 -0400)]
if returning errno value directly from a syscall, we need to negate it.
Rich Felker [Sun, 20 Mar 2011 02:18:53 +0000 (22:18 -0400)]
honor namespace for i386 syscall.h, even though it's not a standard header
Rich Felker [Sun, 20 Mar 2011 01:50:20 +0000 (21:50 -0400)]
fix typo in x86_64 part of syscall overhaul
Rich Felker [Sun, 20 Mar 2011 01:36:10 +0000 (21:36 -0400)]
syscall overhaul part two - unify public and internal syscall interface
with this patch, the syscallN() functions are no longer needed; a
variadic syscall() macro allows syscalls with anywhere from 0 to 6
arguments to be made with a single macro name. also, manually casting
each non-integer argument with (long) is no longer necessary; the
casts are hidden in the macros.
some source files which depended on being able to define the old macro
SYSCALL_RETURNS_ERRNO have been modified to directly use __syscall()
instead of syscall(). references to SYSCALL_SIGSET_SIZE and SYSCALL_LL
have also been changed.
x86_64 has not been tested, and may need a follow-up commit to fix any
minor bugs/oversights.
Rich Felker [Sat, 19 Mar 2011 23:05:43 +0000 (19:05 -0400)]
remove comment cruft that got left behind in x86_64 syscall.s
Rich Felker [Sat, 19 Mar 2011 22:58:32 +0000 (18:58 -0400)]
add some ioctl stuff to sys/mount.h
Rich Felker [Sat, 19 Mar 2011 22:51:42 +0000 (18:51 -0400)]
overhaul syscall interface
this commit shuffles around the location of syscall definitions so
that we can make a syscall() library function with both SYS_* and
__NR_* style syscall names available to user applications, provides
the syscall() library function, and optimizes the code that performs
the actual inline syscalls in the library itself.
previously on i386 when built as PIC (shared library), syscalls were
incurring bus lock (lock prefix) overhead at entry and exit, due to
the way the ebx register was being loaded (xchg instruction with a
memory operand). now the xchg takes place between two registers.
further cleanup to arch/$(ARCH)/syscall.h is planned.
Rich Felker [Sat, 19 Mar 2011 01:53:30 +0000 (21:53 -0400)]
some linux headers useful from user apps.
i'm still not sure whether it's a good idea to include or use any of
these, but i'll add them for now. it may make more sense to just add
official kernel headers to the include path for compiling programs
that need them.
Rich Felker [Sat, 19 Mar 2011 01:52:26 +0000 (21:52 -0400)]
various legacy and linux-specific stuff
this commit is part of an effort to make more of busybox work
out-of-the-box.
Rich Felker [Fri, 18 Mar 2011 13:59:20 +0000 (09:59 -0400)]
document changes for 0.7.1
Rich Felker [Fri, 18 Mar 2011 13:19:09 +0000 (09:19 -0400)]
implement [v]swprintf
Rich Felker [Fri, 18 Mar 2011 02:55:43 +0000 (22:55 -0400)]
implement wprintf family of functions
this implementation is extremely ugly and inefficient, but it avoids a
good deal of code duplication and bloat. it may be cleaned up later to
eliminate the remaining code duplication and some of the warts, but i
don't really care about its performance.
note that swprintf is not yet implemented.
Rich Felker [Fri, 18 Mar 2011 02:38:45 +0000 (22:38 -0400)]
fix broken wmemchr (unbounded search)
Rich Felker [Fri, 18 Mar 2011 00:41:37 +0000 (20:41 -0400)]
implement robust mutexes
some of this code should be cleaned up, e.g. using macros for some of
the bit flags, masks, etc. nonetheless, the code is believed to be
working and correct at this point.
Rich Felker [Thu, 17 Mar 2011 17:35:08 +0000 (13:35 -0400)]
avoid function call to pthread_self in mutex unlock
if the mutex was previously locked, we can assume pthread_self was
already called at the time of locking, and thus that the thread
pointer is initialized.
Rich Felker [Thu, 17 Mar 2011 17:17:15 +0000 (13:17 -0400)]
reorder mutex struct fields to make room for pointers (upcoming robust mutexes)
the layout has been chosen so that pointer slots 3 and 4 fit between
the integer slots on 32-bit archs, and come after the integer slots on
64-bit archs.
Rich Felker [Thu, 17 Mar 2011 16:21:32 +0000 (12:21 -0400)]
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.
Rich Felker [Thu, 17 Mar 2011 16:14:40 +0000 (12:14 -0400)]
optimize contended normal mutex case; add int compare-and-swap atomic
Rich Felker [Wed, 16 Mar 2011 20:49:42 +0000 (16:49 -0400)]
simplify logic, slightly optimize contended case for non-default mutex types
Rich Felker [Wed, 16 Mar 2011 20:25:00 +0000 (16:25 -0400)]
correct error returns for error-checking mutexes
Rich Felker [Wed, 16 Mar 2011 15:36:21 +0000 (11:36 -0400)]
cut out a syscall on thread creation in the case where guard size is 0
Rich Felker [Wed, 16 Mar 2011 15:35:46 +0000 (11:35 -0400)]
don't expose EAGAIN, etc. from timed futex wait to caller
Rich Felker [Wed, 16 Mar 2011 14:39:45 +0000 (10:39 -0400)]
optimize file locking: avoid cache-polluting writes to global storage
Rich Felker [Mon, 14 Mar 2011 15:49:55 +0000 (11:49 -0400)]
partially-written draft of fmemopen, still in #if 0
Rich Felker [Mon, 14 Mar 2011 15:49:17 +0000 (11:49 -0400)]
remove some old cruft from sys/types.h
Rich Felker [Sun, 13 Mar 2011 03:53:17 +0000 (22:53 -0500)]
misplaced & in times() made it fail to work, and clobber the stack
Rich Felker [Sun, 13 Mar 2011 02:55:45 +0000 (21:55 -0500)]
implement flockfile api, rework stdio locking
Rich Felker [Sun, 13 Mar 2011 02:54:19 +0000 (21:54 -0500)]
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.
Rich Felker [Fri, 11 Mar 2011 19:51:36 +0000 (14:51 -0500)]
document some additional important changes
Rich Felker [Fri, 11 Mar 2011 18:38:09 +0000 (13:38 -0500)]
update whatsnew file for release of 0.7.0
Rich Felker [Fri, 11 Mar 2011 15:02:17 +0000 (10:02 -0500)]
match dimensions so we can use all slots without invoking OOB-array-access
Rich Felker [Fri, 11 Mar 2011 14:51:54 +0000 (09:51 -0500)]
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.
Rich Felker [Fri, 11 Mar 2011 14:51:14 +0000 (09:51 -0500)]
fix pthread_attr_* implementations to match corrected prototypes
Rich Felker [Fri, 11 Mar 2011 14:50:54 +0000 (09:50 -0500)]
fix missing ENOTSUP error code
Rich Felker [Fri, 11 Mar 2011 14:46:31 +0000 (09:46 -0500)]
missing const in some pthread_attr_* prototypes
Rich Felker [Fri, 11 Mar 2011 14:46:12 +0000 (09:46 -0500)]
formatting whatsnew file
Rich Felker [Fri, 11 Mar 2011 05:48:40 +0000 (00:48 -0500)]
list major changes in preparation for release
Rich Felker [Fri, 11 Mar 2011 03:05:16 +0000 (22:05 -0500)]
fix failure behavior of sem_open when sem does not exist
Rich Felker [Fri, 11 Mar 2011 02:52:18 +0000 (21:52 -0500)]
fix some semaphore wait semantics (race condition deadlock and error checking)
Rich Felker [Fri, 11 Mar 2011 02:34:19 +0000 (21:34 -0500)]
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.
Rich Felker [Thu, 10 Mar 2011 23:31:37 +0000 (18:31 -0500)]
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.
Rich Felker [Thu, 10 Mar 2011 23:26:29 +0000 (18:26 -0500)]
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.
Rich Felker [Thu, 10 Mar 2011 16:59:39 +0000 (11:59 -0500)]
security fix: check that cancel/rsyscall signal was sent by the process itself
Rich Felker [Thu, 10 Mar 2011 16:06:50 +0000 (11:06 -0500)]
more cancellation points: tcdrain, clock_nanosleep
Rich Felker [Thu, 10 Mar 2011 16:02:29 +0000 (11:02 -0500)]
remove useless return value checks for functions that cannot fail
Rich Felker [Thu, 10 Mar 2011 16:01:11 +0000 (11:01 -0500)]
make sigsuspend a cancellation point
Rich Felker [Thu, 10 Mar 2011 15:59:50 +0000 (10:59 -0500)]
make sigtimedwait a cancellation point
Rich Felker [Thu, 10 Mar 2011 15:43:09 +0000 (10:43 -0500)]
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.
Rich Felker [Thu, 10 Mar 2011 15:26:16 +0000 (10:26 -0500)]
fix sigsuspend syscall
Rich Felker [Thu, 10 Mar 2011 15:17:29 +0000 (10:17 -0500)]
make sigaltstack work (missing macros in signal.h, error conditions)
Rich Felker [Thu, 10 Mar 2011 14:54:23 +0000 (09:54 -0500)]
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.
Rich Felker [Thu, 10 Mar 2011 01:31:06 +0000 (20:31 -0500)]
fix error handling for pthread_sigmask
it must return errno, not -1, and should reject invalud values for how.