Rich Felker [Sat, 17 Mar 2012 21:34:30 +0000 (17:34 -0400)]
Merge remote branch 'nsz/master'
nsz [Sat, 17 Mar 2012 12:46:15 +0000 (13:46 +0100)]
fix i386 fegetround and make fesetround faster
Note that the new fesetround has slightly different semantics:
Storing the floating-point environment with fnstenv makes the
next fldenv (or fldcw) "non-signaling", so unmasked and pending
exceptions does not invoke the exception handler.
(These are rare since exceptions are handled immediately and by
default all exceptions are masked anyway. But if one manually
unmasks an exception in the control word then either sets the
corresponding exception flag in the status word or the execution
of an exception raising floating-point operation gets interrupted
then it may happen).
So the old implementation did not trap in some rare cases
where the new implementation traps.
However POSIX does not specify anything like the x87 exception
handling traps and the fnstenv/fldenv pair is significantly slower
than the fnstcw/fldcw pair (new code is about 5x faster here and
it's dominated by the function call overhead).
Rich Felker [Sat, 17 Mar 2012 04:48:55 +0000 (00:48 -0400)]
don't fail on inability to install dynamic linker (e.g. if not root)
Rich Felker [Sat, 17 Mar 2012 04:02:36 +0000 (00:02 -0400)]
one more fenv availability issue: lround
Rich Felker [Sat, 17 Mar 2012 03:58:49 +0000 (23:58 -0400)]
make fma and lrint functions build without full fenv support
this is necessary to support archs where fenv is incomplete or
unavailable (presently arm). fma, fmal, and the lrint family should
work perfectly fine with this change; fmaf is slightly broken with
respect to rounding as it depends on non-default rounding modes to do
its work.
Rich Felker [Sat, 17 Mar 2012 01:20:53 +0000 (21:20 -0400)]
other side of the signgam namespace fix: use the internal name
Rich Felker [Sat, 17 Mar 2012 01:18:48 +0000 (21:18 -0400)]
make signgam a weak alias for an internal symbol
otherwise, the standard C lgamma function will clobber a symbol in the
namespace reserved for the application.
Rich Felker [Sat, 17 Mar 2012 01:16:32 +0000 (21:16 -0400)]
fix namespace issues for lgamma, etc.
standard functions cannot depend on nonstandard symbols
Rich Felker [Sat, 17 Mar 2012 01:01:34 +0000 (21:01 -0400)]
Merge remote branch 'nsz/master'
Rich Felker [Sat, 17 Mar 2012 00:53:05 +0000 (20:53 -0400)]
remove junk sincos implementations in preparation to merge nsz's real ones
Rich Felker [Sat, 17 Mar 2012 00:52:00 +0000 (20:52 -0400)]
revert COPYRIGHT file changes in preparation to merge nsz's math branch
Rich Felker [Fri, 16 Mar 2012 23:18:00 +0000 (19:18 -0400)]
update copyright status (math library and new year)
Rich Felker [Thu, 15 Mar 2012 23:56:36 +0000 (19:56 -0400)]
remove special nan handling from x86 sqrt asm
a double precision nan, when converted to extended (80-bit) precision,
will never end in 0x400, since the corresponding bits do not exist in
the original double precision value. thus there's no need to waste
time and code size on this check.
Rich Felker [Thu, 15 Mar 2012 16:16:51 +0000 (12:16 -0400)]
simplify nan check in sqrt (x86 asm); result of sqrt is never negative
nsz [Thu, 15 Mar 2012 08:29:53 +0000 (09:29 +0100)]
in math.h make lgamma_r and non-double bessel _GNU_SOURCE only
long double and float bessel functions are no longer xsi extensions
nsz [Thu, 15 Mar 2012 07:17:28 +0000 (08:17 +0100)]
efficient sincos based on sin and cos
Rich Felker [Thu, 15 Mar 2012 06:38:42 +0000 (02:38 -0400)]
implement sincosf and sincosl functions; add prototypes
presumably broken gcc may generate calls to these, and it's said that
ffmpeg makes use of sincosf.
Rich Felker [Thu, 15 Mar 2012 05:55:54 +0000 (01:55 -0400)]
avoid changing NaNs in sqrt (x86 asm) to satisfy c99 f.9 recommendation
Rich Felker [Thu, 15 Mar 2012 05:29:03 +0000 (01:29 -0400)]
correctly rounded sqrt() asm for x86 (i387)
the fsqrt opcode is correctly rounded, but only in the fpu's selected
precision mode, which is 80-bit extended precision. to get a correctly
rounded double precision output, we check for the only corner cases
where two-step rounding could give different results than one-step
(extended-precision mantissa ending in 0x400) and adjust the mantissa
slightly in the opposite direction of the rounding which the fpu
already did (reported in the c1 flag of the fpu status word).
this should have near-zero cost in the non-corner cases and at worst
very low cost.
note that in order for sqrt() to get used when compiling with gcc, the
broken, non-conformant builtin sqrt must be disabled.
Rich Felker [Wed, 14 Mar 2012 02:15:52 +0000 (22:15 -0400)]
correct rounding for i387 sqrtf function
nsz [Tue, 13 Mar 2012 20:11:46 +0000 (21:11 +0100)]
math cleanup: use 1.0f instead of 1.0F
nsz [Tue, 13 Mar 2012 19:24:23 +0000 (20:24 +0100)]
math cleanup: use 1.0f instead of (float)1.0
nsz [Tue, 13 Mar 2012 18:51:14 +0000 (19:51 +0100)]
remove libm.h includes when math.h and float.h are enough
Rich Felker [Tue, 13 Mar 2012 16:37:51 +0000 (12:37 -0400)]
fix scanf handling of "0" (followed by immediate EOF) with "%x"
other cases with %x were probably broken too.
I would actually like to go ahead and replace this code in scanf with
calls to the new __intparse framework, but for now this calls for a
quick and unobtrusive fix without the risk of breaking other things.
nsz [Tue, 13 Mar 2012 15:38:22 +0000 (16:38 +0100)]
clean up __expo2.c, use a slightly better k constant
nsz [Tue, 13 Mar 2012 15:06:43 +0000 (16:06 +0100)]
fix copyright notice for the math lib
Rich Felker [Tue, 13 Mar 2012 05:55:25 +0000 (01:55 -0400)]
implement nan, nanf, nanl
Rich Felker [Tue, 13 Mar 2012 05:17:53 +0000 (01:17 -0400)]
first commit of the new libm!
thanks to the hard work of Szabolcs Nagy (nsz), identifying the best
(from correctness and license standpoint) implementations from freebsd
and openbsd and cleaning them up! musl should now fully support c99
float and long double math functions, and has near-complete complex
math support. tgmath should also work (fully on gcc-compatible
compilers, and mostly on any c99 compiler).
based largely on commit
0376d44a890fea261506f1fc63833e7a686dca19 from
nsz's libm git repo, with some additions (dummy versions of a few
missing long double complex functions, etc.) by me.
various cleanups still need to be made, including re-adding (if
they're correct) some asm functions that were dropped.
Rich Felker [Fri, 9 Mar 2012 08:39:28 +0000 (03:39 -0500)]
add .gitignore file
I've had this around for a long time but somehow it never got
committed.
Rich Felker [Sat, 3 Mar 2012 03:35:37 +0000 (22:35 -0500)]
fix nan/infinity macros in math.h, etc.
the previous version not only failed to work in c++, but also failed
to produce constant expressions, making the macros useless as
initializers for objects of static storage duration.
gcc 3.3 and later have builtins for these, which sadly seem to be the
most "portable" solution. the alternative definitions produce
exceptions (for NAN) and compiler warnings (for INFINITY) on newer
versions of gcc.
Rich Felker [Fri, 2 Mar 2012 17:48:17 +0000 (12:48 -0500)]
fix obscure bug in strtoull reading the highest 16 possible values
Rich Felker [Fri, 2 Mar 2012 16:38:39 +0000 (11:38 -0500)]
typo in math.h c version check
Rich Felker [Fri, 2 Mar 2012 05:36:26 +0000 (00:36 -0500)]
make math.h compatibe with c89
Rich Felker [Fri, 2 Mar 2012 05:24:49 +0000 (00:24 -0500)]
remove debug cruft that was left in getdate
Rich Felker [Fri, 2 Mar 2012 05:24:17 +0000 (00:24 -0500)]
first try at implementing getdate function
Rich Felker [Fri, 2 Mar 2012 05:23:43 +0000 (00:23 -0500)]
fix bugs in strptime handling of string day/month names, literals
Rich Felker [Fri, 2 Mar 2012 04:43:31 +0000 (23:43 -0500)]
implement a64l and l64a (legacy xsi stuff)
Rich Felker [Fri, 2 Mar 2012 04:24:45 +0000 (23:24 -0500)]
add all missing wchar functions except floating point parsers
these are mostly untested and adapted directly from corresponding byte
string functions and similar.
Rich Felker [Fri, 2 Mar 2012 03:08:05 +0000 (22:08 -0500)]
support null buffer argument to getcwd, auto-allocating behavior
this is a popular extension some programs depend on, and by using a
temporary buffer and strdup rather than malloc prior to the syscall,
i've avoided the dependency on free and thus minimized the bloat cost
of supporting this feature.
Rich Felker [Thu, 1 Mar 2012 06:34:58 +0000 (01:34 -0500)]
add memory.h, bogus legacy alias for string.h
Rich Felker [Thu, 1 Mar 2012 05:21:20 +0000 (00:21 -0500)]
search internal headers first
this is necessitated by the ugly <syscall.h> just added
Rich Felker [Thu, 1 Mar 2012 03:55:08 +0000 (22:55 -0500)]
use c++-friendly initializers for pthread initializer definitions
these will also avoid obnoxious warnings with gcc -Wbraces.
Rich Felker [Thu, 1 Mar 2012 03:36:06 +0000 (22:36 -0500)]
add <syscall.h> as an alias for <sys/syscall.h>
apparently some broken stuff (libstdc++) needs this.
Rich Felker [Tue, 28 Feb 2012 20:59:01 +0000 (15:59 -0500)]
implement wcsftime function
Rich Felker [Tue, 28 Feb 2012 16:56:13 +0000 (11:56 -0500)]
release notes for 0.8.6
Rich Felker [Tue, 28 Feb 2012 15:13:35 +0000 (10:13 -0500)]
fix pthread_cleanup_pop(1) crash in non-thread-capable, static-linked programs
Rich Felker [Tue, 28 Feb 2012 01:12:06 +0000 (20:12 -0500)]
update release notes for 0.8.5
Rich Felker [Mon, 27 Feb 2012 23:51:02 +0000 (18:51 -0500)]
work around "signal loses thread pointer" issue with "approach 2"
this was discussed on the mailing list and no consensus on the
preferred solution was reached, so in anticipation of a release, i'm
just committing a minimally-invasive solution that avoids the problem
by ensuring that multi-threaded-capable programs will always have
initialized the thread pointer before any signal handler can run.
in the long term we may switch to initializing the thread pointer at
program start time whenever the program has the potential to access
any per-thread data.
Rich Felker [Sat, 25 Feb 2012 07:52:18 +0000 (02:52 -0500)]
use __attribute__((const)) on arm __pthread_self function
Rich Felker [Sat, 25 Feb 2012 04:23:47 +0000 (23:23 -0500)]
replace prototype for basename in string.h with non-prototype declaration
GNU programs may expect the GNU version of basename, which has a
different prototype (argument is const-qualified) and prototype it
themselves too. of course if they're expecting the GNU behavior for
the function, they'll still run into problems, but at least this
eliminates some compile-time failures.
Rich Felker [Sat, 25 Feb 2012 01:07:21 +0000 (20:07 -0500)]
new attempt at working around the gcc 3 visibility bug
since gcc is failing to generate the necessary ".hidden" directive in
the output asm, generate it explicitly with an __asm__ statement...
Rich Felker [Sat, 25 Feb 2012 01:02:42 +0000 (20:02 -0500)]
remove useless attribute visibility from definitions
this was a failed attempt at working around the gcc 3 visibility bug
affecting x86_64. subsequent patch will address it with an ugly but
working hack.
Rich Felker [Fri, 24 Feb 2012 02:24:56 +0000 (21:24 -0500)]
cleanup and work around visibility bug in gcc 3 that affects x86_64
in gcc 3, the visibility attribute must be placed on both the
declaration and on the definition. if it's omitted from the
definition, the compiler fails to emit the ".hidden" directive in the
assembly, and the linker will either generate textrels (if supported,
such as on i386) or refuse to link (on targets where certain types of
textrels are forbidden or impossible without further assumptions about
memory layout, such as on x86_64).
this patch also unifies the decision about when to use visibility into
libc.h and makes the visibility in the utf-8 state machine tables
based on libc.h rather than a duplicate test.
Rich Felker [Thu, 23 Feb 2012 18:08:47 +0000 (13:08 -0500)]
fix (hopefully) PTRACE_TRACEME (command 0) argument handling
Rich Felker [Thu, 23 Feb 2012 18:07:20 +0000 (13:07 -0500)]
fix for previous incorrect fix of cancellation in dns lookups
uninitialized file descriptor was being closed on return, causing
stdin to be closed in many cases.
Rich Felker [Sat, 18 Feb 2012 04:56:28 +0000 (23:56 -0500)]
fix get_current_dir_name behavior
Rich Felker [Sat, 18 Feb 2012 04:51:23 +0000 (23:51 -0500)]
remove -std=gnu99 from musl-gcc wrapper
while probably desirable, changing the default language variant is
outside the scope of the wrapper's responsibility.
Rich Felker [Sat, 18 Feb 2012 04:17:48 +0000 (23:17 -0500)]
two fixes for "make install" handling of shared libs
1. don't try to install (and thus build) shared libs when they were
disabled in config.mak
2. ensure that the path for the dynamic linker exists before
attempting to install it.
Rich Felker [Sat, 18 Feb 2012 04:10:00 +0000 (23:10 -0500)]
add get_current_dir_name function
Rich Felker [Thu, 16 Feb 2012 02:47:55 +0000 (21:47 -0500)]
add float_t and double_t to math.h
Rich Felker [Sat, 11 Feb 2012 05:06:32 +0000 (00:06 -0500)]
fix default nameserver when resolv.conf doesn't exist
Rich Felker [Sat, 11 Feb 2012 05:05:58 +0000 (00:05 -0500)]
fix illegal goto out of cleanup context in dns lookups
Rich Felker [Fri, 10 Feb 2012 02:24:56 +0000 (21:24 -0500)]
small fix for new pthread cleanup stuff
even if pthread_create/exit code is not linked, run flag needs to be
checked and cleanup function potentially run on pop. thus, move the
code to the module that's always linked when pthread_cleanup_push/pop
is used.
Rich Felker [Thu, 9 Feb 2012 07:33:08 +0000 (02:33 -0500)]
replace bad cancellation cleanup abi with a sane one
the old abi was intended to duplicate glibc's abi at the expense of
being ugly and slow, but it turns out glib was not even using that abi
except on non-gcc-compatible compilers (which it doesn't even support)
and was instead using an exceptions-in-c/unwind-based approach whose
abi we could not duplicate anyway without nasty dwarf2/unwind
integration.
the new abi is copied from a very old glibc abi, which seems to still
be supported/present in current glibc. it avoids all unwinding,
whether by sjlj or exceptions, and merely maintains a linked list of
cleanup functions to be called from the context of pthread_exit. i've
made some care to ensure that longjmp out of a cleanup function should
work, even though it is not required to.
this change breaks abi compatibility with programs which were using
pthread cancellation, which is unfortunate, but that's why i'm making
the change now rather than later. considering that most pthread
features have not been usable until recently anyway, i don't see it as
a major issue at this point.
Rich Felker [Thu, 9 Feb 2012 05:27:19 +0000 (00:27 -0500)]
document iswspace and remove wrongly-included zwsp character
Rich Felker [Thu, 9 Feb 2012 05:20:24 +0000 (00:20 -0500)]
fix typo in iswspace space list table
Rich Felker [Wed, 8 Feb 2012 03:48:58 +0000 (22:48 -0500)]
fix typo in inotify.h
Rich Felker [Wed, 8 Feb 2012 01:31:27 +0000 (20:31 -0500)]
protect against cancellation in dlopen
i'm not sure that it's "correct" for dlopen to block cancellation
when calling constructors for libraries it loads, but it sure seems
like the right thing. in any case, dlopen itself needs cancellation
blocked.
Rich Felker [Wed, 8 Feb 2012 01:29:29 +0000 (20:29 -0500)]
reduce some wasted space in dso structure
Rich Felker [Tue, 7 Feb 2012 18:10:30 +0000 (13:10 -0500)]
declare basename in string.h when _GNU_SOURCE is defined
note that it still will have the standards-conformant behavior, not
the GNU behavior. but at least this prevents broken code from ending
up with truncated pointers due to implicit declarations...
Rich Felker [Tue, 7 Feb 2012 17:08:27 +0000 (12:08 -0500)]
revert hacks for types of stdint.h integer constant macros
per 7.18.4: Each invocation of one of these macros shall expand to an
integer constant expression suitable for use in #if preprocessing
directives. The type of the expression shall have the same type as
would an expression of the corresponding type converted according to
the integer promotions. The value of the expression shall be that of
the argument.
the key phrase is "converted according to the integer promotions".
thus there is no intent or allowance that the expression have
smaller-than-int types.
Rich Felker [Tue, 7 Feb 2012 02:51:02 +0000 (21:51 -0500)]
more locale_t interfaces (string stuff) and header updates
this should be everything except for some functions where the non-_l
version isn't even implemented yet (mainly some non-ISO-C wcs*
functions).
Rich Felker [Tue, 7 Feb 2012 02:33:40 +0000 (21:33 -0500)]
fix some omissions and mistakes in locale_t interface definitions
Rich Felker [Tue, 7 Feb 2012 02:29:31 +0000 (21:29 -0500)]
add more of the locale_t interfaces, all dummied out to ignore the locale
Rich Felker [Mon, 6 Feb 2012 23:23:11 +0000 (18:23 -0500)]
x86_64 vfork implementation
untested; should work.
Rich Felker [Mon, 6 Feb 2012 22:57:29 +0000 (17:57 -0500)]
run ctors/dtors for shared objects loaded with dlopen
Rich Felker [Mon, 6 Feb 2012 19:39:09 +0000 (14:39 -0500)]
add support for init/finit (constructors and destructors)
this is mainly in hopes of supporting c++ (not yet possible for other
reasons) but will also help applications/libraries which use (and more
often, abuse) the gcc __attribute__((__constructor__)) feature in "C"
code.
x86_64 and arm versions of the new startup asm are untested and may
have minor problems.
Rich Felker [Mon, 6 Feb 2012 06:14:23 +0000 (01:14 -0500)]
add deprecated (removed from posix) [efg]cvt() functions
these have not been heavily tested, but they should work as described
in the old standards. probably broken for non-finite values...
Rich Felker [Fri, 3 Feb 2012 08:16:07 +0000 (03:16 -0500)]
include dummied-out dlopen and dlsym functions for static binaries
these don't work (or do anything at all) but at least make it possible
to static link programs that insist on "having" dynamic loading
support...as long as they don't actually need to use it.
adding real support for dlopen/dlsym with static linking is going to
be significantly more difficult...
Rich Felker [Fri, 3 Feb 2012 04:54:46 +0000 (23:54 -0500)]
improve gcc wrapper to support -nostdlib, -nostartfiles
Rich Felker [Thu, 2 Feb 2012 05:11:29 +0000 (00:11 -0500)]
make stdio open, read, and write operations cancellation points
it should be noted that only the actual underlying buffer flush and
fill operations are cancellable, not reads from or writes to the
buffer. this behavior is compatible with POSIX, which makes all
cancellation points in stdio optional, and it achieves the goal of
allowing cancellation of a thread that's "stuck" on IO (due to a
non-responsive socket/pipe peer, slow/stuck hardware, etc.) without
imposing any measurable performance cost.
Rich Felker [Thu, 2 Feb 2012 04:51:19 +0000 (23:51 -0500)]
make passwd/group functions safe against cancellation in stdio
these changes are a prerequisite to making stdio cancellable.
Rich Felker [Wed, 1 Feb 2012 19:12:19 +0000 (14:12 -0500)]
apparently gnu caddr_t is supposed to be char *, not unsigned long
this type should never be used anyway, but some old junk uses it..
Rich Felker [Wed, 1 Feb 2012 19:10:38 +0000 (14:10 -0500)]
fix utmp.h mess, try 2
Rich Felker [Tue, 31 Jan 2012 22:10:02 +0000 (17:10 -0500)]
fix previous utmp.h junk commit
Rich Felker [Tue, 31 Jan 2012 19:39:17 +0000 (14:39 -0500)]
add some missing junk in utmp.h
Rich Felker [Sun, 29 Jan 2012 16:54:18 +0000 (11:54 -0500)]
add fgetpwent (nonstandard function)
based on patch by Jeremy Huntwork
Rich Felker [Sun, 29 Jan 2012 02:09:50 +0000 (21:09 -0500)]
add linux setfs[ug]id syscall wrappers
patch by Jeremy Huntwork
Rich Felker [Thu, 26 Jan 2012 15:53:37 +0000 (10:53 -0500)]
remove useless "extern" keywords in headers
Rich Felker [Wed, 25 Jan 2012 20:30:42 +0000 (15:30 -0500)]
fix typo in FPE_FLTUND definition, signal.h
Rich Felker [Wed, 25 Jan 2012 03:51:27 +0000 (22:51 -0500)]
add MIN/MAX macros to sys/param.h
this is a nonstandard junk header anyway, so just do what apps expect..
Rich Felker [Wed, 25 Jan 2012 00:50:44 +0000 (19:50 -0500)]
add legacy futimes and lutimes functions
based on patch by sh4rm4. these functions are deprecated; futimens and
utimensat should be used instead in new programs.
Rich Felker [Tue, 24 Jan 2012 05:22:27 +0000 (00:22 -0500)]
make gcc wrapper support -shared correctly
it was previously attempting to link start files as part of shared
objects. this is definitely wrong and depending on the platform and
linker could range from just adding extraneous junk to introducing
textrels to making linking fail entirely.
Rich Felker [Tue, 24 Jan 2012 00:51:34 +0000 (19:51 -0500)]
make glob mark symlinks-to-directories with the GLOB_MARK flag
POSIX is unclear on whether it should, but all historical
implementations seem to behave this way, and it seems more useful to
applications.
Rich Felker [Mon, 23 Jan 2012 23:32:40 +0000 (18:32 -0500)]
fix broken copy relocations from dynamic linker cleanup
this issue affected programs which use global variables exported by
non-libc libraries.
Rich Felker [Mon, 23 Jan 2012 07:02:59 +0000 (02:02 -0500)]
dynamic linker support for PIE binaries (position-independent main program)
even with this change, PIE will not work yet due to deficiencies in
the crt1.o startup code.
Rich Felker [Mon, 23 Jan 2012 05:57:38 +0000 (00:57 -0500)]
cleanup dynamic linker, removing some code duplication
Rich Felker [Mon, 23 Jan 2012 05:37:45 +0000 (00:37 -0500)]
update release notes for 0.8.4
Rich Felker [Sun, 22 Jan 2012 22:19:37 +0000 (17:19 -0500)]
fix cancellation failure in single-threaded programs
even a single-threaded program can be cancellable, e.g. if it's called
pthread_cancel(pthread_self()). the correct predicate to check is not
whether multiple threads have been invoked, but whether pthread_self
has been invoked.
Rich Felker [Sun, 22 Jan 2012 20:49:42 +0000 (15:49 -0500)]
support GLOB_PERIOD flag (GNU extension) to glob function
patch by sh4rm4