Rich Felker [Mon, 23 May 2011 01:58:43 +0000 (21:58 -0400)]
fix strncat and wcsncat (double null termination)
also modify wcsncpy to use the same loop logic
Rich Felker [Mon, 23 May 2011 01:54:42 +0000 (21:54 -0400)]
fix wcsncpy writing past end of buffer
Rich Felker [Sun, 22 May 2011 16:34:33 +0000 (12:34 -0400)]
fix brk/sbrk behavior to match the real legacy functions
Rich Felker [Tue, 17 May 2011 18:13:06 +0000 (14:13 -0400)]
release notes for 0.7.10
Rich Felker [Wed, 11 May 2011 23:58:03 +0000 (19:58 -0400)]
fix the last known rounding bug in floating point printing
the observed symptom was that the code was incorrectly rounding up
1.0625 to 1.063 despite the rounding mode being round-to-nearest with
ties broken by rounding to even last place. however, the code was just
not right in many respects, and i'm surprised it worked as well as it
did. this time i tested the values that end up in the variables round,
small, and the expression round+small, and all look good.
Rich Felker [Sun, 8 May 2011 04:03:00 +0000 (00:03 -0400)]
eventfd syscall wrapper and read/write wrappers
Rich Felker [Sun, 8 May 2011 03:39:48 +0000 (23:39 -0400)]
optimize out useless default-attribute object in pthread_create
Rich Felker [Sun, 8 May 2011 03:37:10 +0000 (23:37 -0400)]
optimize compound-literal sigset_t's not to contain useless hurd bits
Rich Felker [Sun, 8 May 2011 03:23:58 +0000 (23:23 -0400)]
overhaul implementation-internal signal protections
the new approach relies on the fact that the only ways to create
sigset_t objects without invoking UB are to use the sig*set()
functions, or from the masks returned by sigprocmask, sigaction, etc.
or in the ucontext_t argument to a signal handler. thus, as long as
sigfillset and sigaddset avoid adding the "protected" signals, there
is no way the application will ever obtain a sigset_t including these
bits, and thus no need to add the overhead of checking/clearing them
when sigprocmask or sigaction is called.
note that the old code actually *failed* to remove the bits from
sa_mask when sigaction was called.
the new implementations are also significantly smaller, simpler, and
faster due to ignoring the useless "GNU HURD signals" 65-1024, which
are not used and, if there's any sanity in the world, never will be
used.
Rich Felker [Sat, 7 May 2011 01:45:48 +0000 (21:45 -0400)]
reduce some ridiculously large spin counts
these should be tweaked according to testing. offhand i know 1000 is
too low and 5000 is likely to be sufficiently high. consider trying to
add futexes to file locking, too...
Rich Felker [Sat, 7 May 2011 00:27:45 +0000 (20:27 -0400)]
remove debug code that was missed in barrier commit
Rich Felker [Sat, 7 May 2011 00:00:59 +0000 (20:00 -0400)]
completely new barrier implementation, addressing major correctness issues
the previous implementation had at least 2 problems:
1. the case where additional threads reached the barrier before the
first wave was finished leaving the barrier was untested and seemed
not to be working.
2. threads leaving the barrier continued to access memory within the
barrier object after other threads had successfully returned from
pthread_barrier_wait. this could lead to memory corruption or crashes
if the barrier object had automatic storage in one of the waiting
threads and went out of scope before all threads finished returning,
or if one thread unmapped the memory in which the barrier object
lived.
the new implementation avoids both problems by making the barrier
state essentially local to the first thread which enters the barrier
wait, and forces that thread to be the last to return.
Rich Felker [Thu, 5 May 2011 21:44:06 +0000 (17:44 -0400)]
add SA_NOMASK alias for SA_NODEFER with _GNU_SOURCE
Rich Felker [Mon, 2 May 2011 13:18:03 +0000 (09:18 -0400)]
fix fclose return status logic, again
the previous fix was incorrect, as it would prevent f->close(f) from
being called if fflush(f) failed. i believe this was the original
motivation for using | rather than ||. so now let's just use a second
statement to constrain the order of function calls, and to back to
using |.
Rich Felker [Mon, 2 May 2011 02:59:14 +0000 (22:59 -0400)]
fix undefined call order in fclose, possible lost output depending on compiler
pcc turned up this bug by calling f->close(f) before fflush(f),
resulting in lost output and error on flush.
Rich Felker [Mon, 2 May 2011 02:16:04 +0000 (22:16 -0400)]
workaround for preprocessor bug in pcc
with this patch, musl compiles and mostly works with pcc 1.0.0. a few
tests are still failing and i'm uncertain whether they are due to
portability problems in musl, or bugs in pcc, but i suspect the
latter.
Rich Felker [Mon, 2 May 2011 00:12:51 +0000 (20:12 -0400)]
minor optimization in puts: use inline putc_unlocked macro for newline
Rich Felker [Fri, 29 Apr 2011 15:14:55 +0000 (11:14 -0400)]
avoid crashing when nel==0 is passed to qsort
Rich Felker [Thu, 28 Apr 2011 03:41:48 +0000 (23:41 -0400)]
use compiler builtins for variadic macros when available
this slightly cuts down on the degree musl "fights with" gcc, but more
importantly, it fixes a critical bug when gcc inlines a variadic
function and optimizes out the variadic arguments due to noticing that
they were "not used" (by __builtin_va_arg).
we leave the old code in place if __GNUC__ >= 3 is false; it seems
like it might be necessary at least for tinycc support and perhaps if
anyone ever gets around to fixing gcc 2.95.3 enough to make it work..
Rich Felker [Wed, 27 Apr 2011 20:06:33 +0000 (16:06 -0400)]
correct variadic prototypes for execl* family
the old versions worked, but conflicted with programs which declared
their own prototypes and generated warnings with some versions of gcc.
Rich Felker [Wed, 27 Apr 2011 17:27:04 +0000 (13:27 -0400)]
replace heap sort with smoothsort implementation by Valentin Ochs
Smoothsort is an adaptive variant of heapsort. This version was
written by Valentin Ochs (apo) specifically for inclusion in musl. I
worked with him to get it working in O(1) memory usage even with giant
array element widths, and to optimize it heavily for size and speed.
It's still roughly 4 times as large as the old heap sort
implementation, but roughly 20 times faster given an almost-sorted
array of 1M elements (20 being the base-2 log of 1M), i.e. it really
does reduce O(n log n) to O(n) in the mostly-sorted case. It's still
somewhat slower than glibc's Introsort for random input, but now
considerably faster than glibc when the input is already sorted, or
mostly sorted.
Rich Felker [Wed, 27 Apr 2011 16:19:49 +0000 (12:19 -0400)]
add word-sized ctz function to atomic.h
strictly speaking this and a few other ops should be factored into
asm.h or the file should just be renamed to asm.h, but whatever. clean
it up someday.
Rich Felker [Tue, 26 Apr 2011 16:28:41 +0000 (12:28 -0400)]
function signature fix: add const qualifier to mempcpy src arg
Rich Felker [Tue, 26 Apr 2011 12:42:55 +0000 (08:42 -0400)]
typo in prototype for mempcpy
Rich Felker [Tue, 26 Apr 2011 12:41:54 +0000 (08:41 -0400)]
prototype for mempcpy
Rich Felker [Mon, 25 Apr 2011 21:49:21 +0000 (17:49 -0400)]
fix bug in ipv6 parsing that prevented parsing a lone "::"
Rich Felker [Mon, 25 Apr 2011 21:04:40 +0000 (17:04 -0400)]
ipv6 parsing code (formerly dummied-out)
Rich Felker [Mon, 25 Apr 2011 14:40:25 +0000 (10:40 -0400)]
fix 2 eof-related bugs in scanf
1. failed match of literal chars from the format string would always
return matching failure rather than input failure at eof, leading to
infinite loops in some programs.
2. unread of eof would wrongly adjust the character counts reported by
%n, yielding an off-by-one error.
Rich Felker [Sat, 23 Apr 2011 00:14:04 +0000 (20:14 -0400)]
update release notes to new format, add more
Rich Felker [Fri, 22 Apr 2011 22:48:57 +0000 (18:48 -0400)]
fix initial stack alignment in new threads on x86_64
Rich Felker [Fri, 22 Apr 2011 03:33:39 +0000 (23:33 -0400)]
update release notes in prep for 0.7.9 release
Rich Felker [Fri, 22 Apr 2011 01:52:41 +0000 (21:52 -0400)]
omit errno update path for syscalls that cannot fail
Rich Felker [Thu, 21 Apr 2011 20:57:00 +0000 (16:57 -0400)]
fix bogus return values for inet_pton
Rich Felker [Thu, 21 Apr 2011 18:27:28 +0000 (14:27 -0400)]
move wait.h macros out of bits. they do not vary.
Rich Felker [Thu, 21 Apr 2011 18:21:57 +0000 (14:21 -0400)]
include signal.h to avoid thorny __sigcontext/sigcontext issues
this is explicitly allowed by POSIX
Rich Felker [Thu, 21 Apr 2011 01:41:45 +0000 (21:41 -0400)]
fix minor bugs due to incorrect threaded-predicate semantics
some functions that should have been testing whether pthread_self()
had been called and initialized the thread pointer were instead
testing whether pthread_create() had been called and actually made the
program "threaded". while it's unlikely any mismatch would occur in
real-world problems, this could have introduced subtle bugs. now, we
store the address of the main thread's thread descriptor in the libc
structure and use its presence as a flag that the thread register is
initialized. note that after fork, the calling thread (not necessarily
the original main thread) is the new main thread.
Rich Felker [Thu, 21 Apr 2011 01:05:10 +0000 (21:05 -0400)]
workaround bug in linux dup2
the linux documentation for dup2 says it can fail with EBUSY due to a
race condition with open and dup in the kernel. shield applications
(and the rest of libc) from this nonsense by looping until it succeeds
Rich Felker [Thu, 21 Apr 2011 01:01:42 +0000 (21:01 -0400)]
properly create new session/controlling terminal in forkpty
Rich Felker [Thu, 21 Apr 2011 00:55:13 +0000 (20:55 -0400)]
implement (nonstandard) forkpty
Rich Felker [Thu, 21 Apr 2011 00:02:35 +0000 (20:02 -0400)]
disallow blank strings as service or host name
Rich Felker [Thu, 21 Apr 2011 00:00:59 +0000 (20:00 -0400)]
fix bugs in ipv4 parsing
Rich Felker [Wed, 20 Apr 2011 23:15:27 +0000 (19:15 -0400)]
fix initgroups (uninitialized count passed to getgrouplist)
Rich Felker [Wed, 20 Apr 2011 23:05:59 +0000 (19:05 -0400)]
shadow password fixes: empty fields should read as -1 not 0
Rich Felker [Wed, 20 Apr 2011 19:55:58 +0000 (15:55 -0400)]
namespace fixes for sys/mman.h
Rich Felker [Wed, 20 Apr 2011 19:36:15 +0000 (15:36 -0400)]
fix missing include in posix_madvise.c (compile error)
Rich Felker [Wed, 20 Apr 2011 19:25:28 +0000 (15:25 -0400)]
support posix_madvise (previous a stub)
the check against MADV_DONTNEED to because linux MADV_DONTNEED
semantics conflict dangerously with the POSIX semantics
Rich Felker [Wed, 20 Apr 2011 19:20:22 +0000 (15:20 -0400)]
add syscall wrappers for posix_fadvise, posix_fallocate
Rich Felker [Wed, 20 Apr 2011 03:42:56 +0000 (23:42 -0400)]
block cancellation in wordexp, handle more errors
Rich Felker [Wed, 20 Apr 2011 03:37:57 +0000 (23:37 -0400)]
avoid malloc of potentially-large string in wordexp
Rich Felker [Wed, 20 Apr 2011 03:09:14 +0000 (23:09 -0400)]
move some more code out of pthread_create.c
this also de-uglifies the dummy function aliasing a bit.
Rich Felker [Tue, 19 Apr 2011 17:16:59 +0000 (13:16 -0400)]
fix uninitialized waiters field in semaphores
Rich Felker [Tue, 19 Apr 2011 01:35:14 +0000 (21:35 -0400)]
dns lookups: protect against cancellation and fix incorrect error codes
Rich Felker [Tue, 19 Apr 2011 01:22:14 +0000 (21:22 -0400)]
avoid fd leak if opendir is cancelled when calloc has failed
Rich Felker [Tue, 19 Apr 2011 01:17:03 +0000 (21:17 -0400)]
protect ftw and nftw against cancellation
Rich Felker [Tue, 19 Apr 2011 01:11:23 +0000 (21:11 -0400)]
protect syslog against cancellation
these functions are allowed to be cancellation points, but then we
would have to install cleanup handlers to avoid termination with locks
held.
Rich Felker [Tue, 19 Apr 2011 00:50:37 +0000 (20:50 -0400)]
recheck cancellation disabled flag after syscall returns EINTR
we already checked before making the syscall, but it's possible that a
signal handler interrupted the blocking syscall and disabled
cancellation, and that this is the cause of EINTR. in this case, the
old behavior was testably wrong.
Rich Felker [Mon, 18 Apr 2011 22:42:34 +0000 (18:42 -0400)]
remove bogus extra logic for close cancellability
like all other syscalls, close should return to the caller if and only
if it successfully performed its action. it is necessary that the
application be able to determine whether the close succeeded.
Rich Felker [Sun, 17 Apr 2011 23:25:17 +0000 (19:25 -0400)]
fix typo in x86_64 cancellable syscall asm
Rich Felker [Sun, 17 Apr 2011 21:32:36 +0000 (17:32 -0400)]
minimal realpath implementation using /proc
clean and simple, but fails when the caller does not have permissions
to open the file for reading or when /proc is not available. i may
replace this with a full implementation later, possibly leaving this
version as an optimization to use when it works.
Rich Felker [Sun, 17 Apr 2011 21:09:41 +0000 (17:09 -0400)]
pthread_exit is not supposed to affect cancellability
if the exit was caused by cancellation, __cancel has already set these
flags anyway.
Rich Felker [Sun, 17 Apr 2011 21:06:05 +0000 (17:06 -0400)]
fix pthread_exit from cancellation handler
cancellation frames were not correctly popped, so this usage would not
only loop, but also reuse discarded and invalid parts of the stack.
Rich Felker [Sun, 17 Apr 2011 20:53:54 +0000 (16:53 -0400)]
clean up handling of thread/nothread mode, locking
Rich Felker [Sun, 17 Apr 2011 20:32:15 +0000 (16:32 -0400)]
debloat: use __syscall instead of syscall where possible
don't waste time (and significant code size due to function call
overhead!) setting errno when the result of a syscall does not matter
or when it can't fail.
Rich Felker [Sun, 17 Apr 2011 19:30:08 +0000 (15:30 -0400)]
fix bugs in cancellable syscall asm
x86_64 was just plain wrong in the cancel-flag-already-set path, and
crashing.
the more subtle error was not clearing the saved stack pointer before
returning to c code. this could result in the signal handler
misidentifying c code as the pre-syscall part of the asm, and acting
on cancellation at the wrong time, and thus resource leak race
conditions.
also, now __cancel (in the c code) is responsible for clearing the
saved sp in the already-cancelled branch. this means we have to use
call rather than jmp to ensure the stack pointer in the c will never
match what the asm saved.
Rich Felker [Sun, 17 Apr 2011 17:21:13 +0000 (13:21 -0400)]
optimize cancellation enable/disable code
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.
Rich Felker [Sun, 17 Apr 2011 16:15:55 +0000 (12:15 -0400)]
don't use pthread_once when there is no danger in race
Rich Felker [Sun, 17 Apr 2011 16:09:47 +0000 (12:09 -0400)]
fix some minor issues in cancellation handling patch
signals were wrongly left masked, and cancellability state was not
switched to disabled, during the execution of cleanup handlers.
Rich Felker [Sun, 17 Apr 2011 15:43:03 +0000 (11:43 -0400)]
overhaul pthread cancellation
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.
Rich Felker [Fri, 15 Apr 2011 16:07:26 +0000 (12:07 -0400)]
remove stupid debug code in wordexp
Rich Felker [Fri, 15 Apr 2011 16:06:34 +0000 (12:06 -0400)]
implement wordexp. first try, may be buggy. intended to be safe.
Rich Felker [Fri, 15 Apr 2011 16:04:13 +0000 (12:04 -0400)]
avoid setting errno when checking for tty
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.
Rich Felker [Fri, 15 Apr 2011 03:33:46 +0000 (23:33 -0400)]
document some changes in the upcoming release
Rich Felker [Fri, 15 Apr 2011 02:06:30 +0000 (22:06 -0400)]
fix O_SYNC definition, cleanup fcntl.h
Rich Felker [Fri, 15 Apr 2011 01:50:07 +0000 (21:50 -0400)]
fix FAPPEND typo on x86_64 (previously only fixed on i386)
Rich Felker [Fri, 15 Apr 2011 01:49:22 +0000 (21:49 -0400)]
fcntl.h: move macros that do not vary between archs out of bits
Rich Felker [Fri, 15 Apr 2011 01:45:26 +0000 (21:45 -0400)]
fix broken fcntl locks on x86_64
Rich Felker [Fri, 15 Apr 2011 01:43:49 +0000 (21:43 -0400)]
make tmpfile slightly more efficient (use unlink syscall instead of remove)
Rich Felker [Thu, 14 Apr 2011 23:14:42 +0000 (19:14 -0400)]
fix typo in legacy FAPPEND definition
Rich Felker [Thu, 14 Apr 2011 20:23:31 +0000 (16:23 -0400)]
add useless type fd_mask. it's in the reserved namespace.
Rich Felker [Thu, 14 Apr 2011 19:10:50 +0000 (15:10 -0400)]
change sem_trywait algorithm so it never has to call __wake
Rich Felker [Thu, 14 Apr 2011 18:39:57 +0000 (14:39 -0400)]
cheap trick to further optimize locking normal mutexes
Rich Felker [Thu, 14 Apr 2011 16:51:00 +0000 (12:51 -0400)]
use a separate signal from SIGCANCEL for SIGEV_THREAD timers
otherwise we cannot support an application's desire to use
asynchronous cancellation within the callback function. this change
also slightly debloats pthread_create.c.
Rich Felker [Thu, 14 Apr 2011 00:47:01 +0000 (20:47 -0400)]
simplify cancellation point handling
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.
Rich Felker [Wed, 13 Apr 2011 22:32:33 +0000 (18:32 -0400)]
simplify syslog, add vsyslog interface (nonstandard)
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...
Rich Felker [Wed, 13 Apr 2011 21:51:45 +0000 (17:51 -0400)]
remove useless SIGPIPE protection from syslog
per the standard, SIGPIPE is not generated for SOCK_DGRAM.
Rich Felker [Wed, 13 Apr 2011 21:24:25 +0000 (17:24 -0400)]
fix syslog (corrected SIGPIPE blocking, and using dgram instead of stream)
it actually appears the hacks to block SIGPIPE are probably not
necessary, and potentially harmful. if i can confirm this, i'll remove
them.
Rich Felker [Wed, 13 Apr 2011 20:45:43 +0000 (16:45 -0400)]
numerous fixes to sysv ipc
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.
Rich Felker [Wed, 13 Apr 2011 20:13:49 +0000 (16:13 -0400)]
add syslog.h cruft for syslogd to use...
Rich Felker [Wed, 13 Apr 2011 19:24:26 +0000 (15:24 -0400)]
add profile for getmntent_r
Rich Felker [Wed, 13 Apr 2011 18:55:26 +0000 (14:55 -0400)]
add syscall wrapper for flock
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...
Rich Felker [Wed, 13 Apr 2011 18:52:23 +0000 (14:52 -0400)]
fix bug whereby getopt_long allowed mismatch in last char of option name
Rich Felker [Wed, 13 Apr 2011 17:22:19 +0000 (13:22 -0400)]
fix typos on RLIM_NLIMITS, remove _GNU_SOURCE test for it
RLIM_* is in the reserved namespace for this header
Rich Felker [Wed, 13 Apr 2011 17:16:49 +0000 (13:16 -0400)]
fix and cleanup suseconds_t/timeval stuff (broken on 64-bit)
trash in the upper 32 bits was making the kernel sleep forever in
select on 64-bit systems.
Rich Felker [Wed, 13 Apr 2011 13:39:47 +0000 (09:39 -0400)]
implement getgrouplist (for initgroups), formerly dummied-out
Rich Felker [Wed, 13 Apr 2011 13:03:22 +0000 (09:03 -0400)]
fix prototypes/signature for setgroups, etc.
Rich Felker [Wed, 13 Apr 2011 12:45:28 +0000 (08:45 -0400)]
fix incorrect GNU sighandler_t typedef
Rich Felker [Wed, 13 Apr 2011 12:36:29 +0000 (08:36 -0400)]
implement memrchr (nonstandard) and optimize strrchr in terms of it
Rich Felker [Wed, 13 Apr 2011 12:35:32 +0000 (08:35 -0400)]
add ptsname_r (nonstandard) and split ptsname (standard) to separate file
this eliminates the ugly static buffer in programs that use ptsname_r.
Rich Felker [Tue, 12 Apr 2011 21:52:14 +0000 (17:52 -0400)]
speed up threaded fork
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.
Rich Felker [Tue, 12 Apr 2011 17:36:22 +0000 (13:36 -0400)]
more changes for upcoming 0.7.8