OSDN Git Service
rofl0r [Thu, 6 Dec 2012 15:57:22 +0000 (16:57 +0100)]
add more arch-specific MAP_ macros to bits/mman.h
these are also needed by qemu.
rofl0r [Thu, 6 Dec 2012 00:24:03 +0000 (01:24 +0100)]
remove MAP_32 from non-x86 archs
both kernel and glibc define it only on x86(_64).
rofl0r [Thu, 6 Dec 2012 00:20:48 +0000 (01:20 +0100)]
add MAP_NORESERVE to bits/mman.h
this is needed for qemu, and since it differs for each arch
it can't be circumvented easily by using a macro in CFLAGS.
Rich Felker [Thu, 6 Dec 2012 00:13:47 +0000 (19:13 -0500)]
remove fenv saving/loading code from setjmp/longjmp on arm
the issue is identical to the recent commit fixing the mips versions:
despite other implementations doing this, it conflicts with the
requirements of ISO C and it's a waste of time and code size.
Rich Felker [Wed, 5 Dec 2012 19:12:57 +0000 (14:12 -0500)]
fix inefficiency of math.h isless, etc. macros
previously, everything was going through an intermediate conversion to
long double, which caused the extern __fpclassifyl function to get
invoked, preventing virtually all optimizations of these operations.
with the new code, tests on constant float or double arguments compile
to a constant 0 or 1, and tests on non-constant expressions are
efficient. I may later add support for __builtin versions on compilers
that support them.
Rich Felker [Wed, 5 Dec 2012 17:52:40 +0000 (12:52 -0500)]
remove mips setjmp/longjmp code to save/restore fenv
nothing in the standard requires or even allows the fenv state to be
restored by longjmp. restoring the exception flags is not such a big
deal since it's probably valid to clobber them completely, but
restoring the rounding mode yields an observable side effect not
sanctioned by ISO C. saving/restoring it also wastes a few cycles and
16 bytes of code.
as for historical behavior, reportedly SGI IRIX did save/restore fenv,
and this is where glibc and uClibc got the behavior from. a few other
systems save/restore it too (on archs other than mips), even though
this is apparently wrong. further details are documented here:
http://www-personal.umich.edu/~williams/archive/computation/setjmp-fpmode.html
as musl aims for standards conformance rather than coddling historical
programs expecting non-conforming behavior, and as it's unlikely that
any historical programs actually depend on the incorrect behavior
(such programs would break on other archs, anyway), I'm making the
change not to save/restore fenv on mips.
Rich Felker [Wed, 5 Dec 2012 17:35:24 +0000 (12:35 -0500)]
add scsi headers scsi.h and sg.h
due to some historical oddity, these are considered libc headers
rather than kernel headers. the kernel used to provide them too, but
it seems modern kernels do not install them, so let's just do the
easiest thing and provide them. stripped-down versions provided by
John Spencer.
Rich Felker [Wed, 5 Dec 2012 05:00:42 +0000 (00:00 -0500)]
use __builtin_offsetof to implement offsetof when possible
apparently recent gcc versions have intentionally broken the
traditional definition by treating it as a non-constant expression.
the traditional definition may also be problematic for c++ programs.
Rich Felker [Wed, 5 Dec 2012 00:03:42 +0000 (19:03 -0500)]
use alternate argument syntax for restrict with lio_listio
for some reason I have not been able to determine, gcc 3.2 rejects the
array notation. this seems to be a gcc bug, but since it's easy to
work around, let's do the workaround and avoid gratuitously requiring
newer compilers.
Rich Felker [Tue, 4 Dec 2012 14:32:45 +0000 (09:32 -0500)]
fix regression in arm user.h that happened during big user.h changes
Rich Felker [Mon, 3 Dec 2012 22:02:56 +0000 (17:02 -0500)]
add _ALL_SOURCE as an alias for _GNU_SOURCE/enable-everything
reportedly this is a semi-common practice among some BSDs and a few
other systems, and will improve application compatibility.
Rich Felker [Mon, 3 Dec 2012 21:57:01 +0000 (16:57 -0500)]
feature test macros: make _GNU_SOURCE enable everything
previously, a few BSD features were enabled only by _BSD_SOURCE, not
by _GNU_SOURCE. since _BSD_SOURCE is default in the absence of other
feature test macros, this made adding _GNU_SOURCE to a project not a
purely additive feature test macro; it actually caused some features
to be suppressed.
most of the changes made by this patch actually bring musl in closer
alignment with the glibc behavior for _GNU_SOURCE. the only exceptions
are the added visibility of functions like strlcpy which were BSD-only
due to being disliked/rejected by glibc maintainers. here, I feel the
consistency of having _GNU_SOURCE mean "everything", and especially
the property of it being purely additive, are more valuable than
hiding functions which glibc does not have.
Rich Felker [Sun, 2 Dec 2012 20:18:05 +0000 (15:18 -0500)]
fix a couple issues in the inttypes.h PRI/SCN macros
most importantly, the format/scan macros for the [u]int_fast16_t and
[u]int_fast32_t types were defined incorrectly assuming these types
would match the native word/pointer size. this is incorrect on any
64-bit system; the "fast" types for 16- and 32-bit integers are simply
int.
another issue which was "only a warning" (despite being UB) is that
the choice of "l" versus "ll" was incorrect for 64-bit types on 64-bit
machines. while it would "work" to always use "long long" for 64-bit
types, we use "long" on 64-bit machines to match what glibc does and
what the ABI documents recommend. the macro definitions were probably
right in very old versions of musl, but became wrong when we aligned
most closely with the 'standard' ABI. checking UINTPTR_MAX is an easy
way to get the system wordsize without pulling in new headers.
finally, the useless __PRIPTR macro to allow the underlying type of
[u]intptr_t to vary has been removed. we are using "long" on all
targets, and thankfully this matches what glibc does, so I do not
envision ever needing to change it. thus, the "l" has just been
incorporated directly in the strings.
Rich Felker [Sat, 1 Dec 2012 02:00:18 +0000 (21:00 -0500)]
provide NSIG under _BSD_SOURCE (default) as well as _GNU_SOURCE
this fixes a regression related to the changes made to bits/signal.h
between 0.9.7 and 0.9.8 that broke some (non-portable) software.
Rich Felker [Fri, 30 Nov 2012 22:56:23 +0000 (17:56 -0500)]
fix ordering of shared library ctors with respect to libc init
previously, shared library constructors were being called before
important internal things like the environment (extern char **environ)
and hwcap flags (needed for sjlj to work right with float on arm) were
initialized in __libc_start_main. rather than trying to have to
dynamic linker make sure this stuff all gets initialized right, I've
opted to just defer calling shared library constructors until after
the main program's entry point is reached. this also fixes the order
of ctors to be the exact reverse of dtors, which is a desirable
property and possibly even mandated by some languages.
the main practical effect of this change is that shared libraries
calling getenv from ctors will no longer fail.
Rich Felker [Tue, 27 Nov 2012 14:44:30 +0000 (09:44 -0500)]
fix some restrict-qualifier mismatches in newly added interfaces
these should have little/no practical impact but they're needed for
strict conformance.
Rich Felker [Tue, 27 Nov 2012 02:01:30 +0000 (21:01 -0500)]
update readme and release notes for 0.9.8
Rich Felker [Tue, 27 Nov 2012 01:42:20 +0000 (20:42 -0500)]
remove stat member aliases from ppc-specific bits/stat.h
if these are to be supported, they belong in the main stat.h, not
repeated for each arch.
Rich Felker [Mon, 26 Nov 2012 18:00:58 +0000 (13:00 -0500)]
fix eventfd and inotify nonblock/cloexec flags to match arch values
Rich Felker [Mon, 26 Nov 2012 17:05:33 +0000 (12:05 -0500)]
fix missing limits when only _BSD_SOURCE is defined
the missing check did not affect the default profile, since it has
both _XOPEN_SOURCE and _BSD_SOURCE defined, but it did break programs
which explicitly define _BSD_SOURCE, causing it to be the only feature
test macro present.
Rich Felker [Mon, 26 Nov 2012 04:04:23 +0000 (23:04 -0500)]
fixup mcontext stuff to expost gregset_t/fpregset_t as appropriate
Rich Felker [Mon, 26 Nov 2012 03:28:18 +0000 (22:28 -0500)]
make sys/procfs.h mostly work on most archs
these structures are purely for use by trace/debug tools and tools
working with core files. the definition of fpregset_t, which was
previously here, has been removed because it was wrong; fpregset_t
should be the type used in mcontext_t, not the type used in
ptrace/core stuff.
Rich Felker [Mon, 26 Nov 2012 02:12:34 +0000 (21:12 -0500)]
remove microblaze user.h incorrectly copied from arm
Rich Felker [Mon, 26 Nov 2012 01:56:31 +0000 (20:56 -0500)]
eliminate gdb complaints about missing linux-gate.so.1
actually, the hard-coded name should be eliminated too, and replaced
by a search for the soname in the headers, but that can be done
separately later.
Rich Felker [Sat, 24 Nov 2012 01:05:43 +0000 (20:05 -0500)]
begin sys/user.h and sys/reg.h fixes for ports
aside from microblaze, these should be roughly correct for all archs
now. some misc junk macros and typedefs are missing, which should
probably be added for max compatibility with trace/debug tools.
Rich Felker [Sat, 24 Nov 2012 00:54:56 +0000 (19:54 -0500)]
fix powerpc ucontext, again..
it should now really match the kernel. some of the removed padding
corresponded to the difference between user and kernel sigset_t. the
space at the end was redundant with the uc_mcontext member and seems
to have been added as a result of misunderstanding glibc's definition
versus the kernel's.
Rich Felker [Fri, 23 Nov 2012 22:23:38 +0000 (17:23 -0500)]
fix termios constants on mips
Rich Felker [Fri, 23 Nov 2012 19:35:25 +0000 (14:35 -0500)]
sigcontext/mcontext cleanup for arch-specific bits
with these changes, the members/types of mcontext_t and related stuff
should closely match the glibc definitions. unlike glibc, however, the
definitions here avoid using typedefs as much as possible and work
directly with the underlying types, to minimize namespace pollution
from signal.h in the default (_BSD_SOURCE) profile.
this is a first step in improving compatibility with applications
which poke at context/register information -- mainly debuggers, trace
utilities, etc. additional definitions in ucontext.h and other headers
may be needed later.
if feature test macros are used to request a conforming namespace,
mcontext_t is replaced with an opaque structure of the equivalent size
and alignment; conforming programs cannot examine its contents anyway.
Rich Felker [Fri, 23 Nov 2012 19:34:59 +0000 (14:34 -0500)]
fix powerpc sigsetjmp asm to match the new jmp_buf size/offsets
Rich Felker [Fri, 23 Nov 2012 19:30:58 +0000 (14:30 -0500)]
fix powerpc setjmp/longjmp to save/restore float regs; enlarge/align jmp_buf
Rich Felker [Fri, 23 Nov 2012 18:31:16 +0000 (13:31 -0500)]
add missing startfiles for powerpc
Rich Felker [Fri, 23 Nov 2012 17:20:53 +0000 (12:20 -0500)]
fix up leftover, incorrect NSIG definitions in arch-specific signal.h
Rich Felker [Fri, 23 Nov 2012 16:27:25 +0000 (11:27 -0500)]
remove stuff that doesn't belong in powerpc 32-bit sigcontext
these fields were wrongly copied from the kernel's ppc64 struct def
Rich Felker [Wed, 21 Nov 2012 18:51:14 +0000 (13:51 -0500)]
more ppc signal.h typedef order fixes
Rich Felker [Wed, 21 Nov 2012 18:50:40 +0000 (13:50 -0500)]
fix misordered typedefs in ppc signal.h
Rich Felker [Wed, 21 Nov 2012 18:49:31 +0000 (13:49 -0500)]
ppc signal/NSIG related fix (warning fix, redefinition)
Rich Felker [Wed, 21 Nov 2012 18:41:58 +0000 (13:41 -0500)]
add back NSIG, removed from powerpc in last commit, but for all archs
unlike the previous definition, NSIG/_NSIG is supposed to be one more
than the highest signal number. adding this will allow simplifying
libc-internal code that makes signal-related syscalls, which can be
done as a later step. some apps might use it too; while this usage is
questionable, it's at least not insane.
Rich Felker [Wed, 21 Nov 2012 18:36:01 +0000 (13:36 -0500)]
make powerpc signal.h namespace-clean for posix/isoc
also handle the non-GNUC case where alignment attribute is not available
by simply omitting it. this will not cause problems except for
inclusion of mcontex_t/ucontext_t in application-defined structures,
since the natural alignment of the uc_mcontext member relative to the
start of ucontext_t is already correct. and shame on whoever designed
this for making it impossible to satisfy the ABI requirements without
GNUC extensions.
Rich Felker [Wed, 21 Nov 2012 18:29:50 +0000 (13:29 -0500)]
fix powerpc types to match abi, and some feature test issues
rofl0r [Wed, 21 Nov 2012 05:24:05 +0000 (06:24 +0100)]
fix invalid usage of mcontext_t in powerpc signal.h
rofl0r [Mon, 19 Nov 2012 05:21:48 +0000 (06:21 +0100)]
powerpc: handle syscall error in clone.
sigsetjmp: store temporaries in jmp_buf rather than on stack.
Rich Felker [Mon, 19 Nov 2012 04:15:47 +0000 (23:15 -0500)]
fix error in configure script using >/dev/null in noclobber mode
Rich Felker [Mon, 19 Nov 2012 03:57:32 +0000 (22:57 -0500)]
fix powerpc asm not to store data in volatile space below stack pointer
it's essential to decrement the stack pointer before writing to new
stack space, rather than afterwards. otherwise there is a race
condition during which asynchronous code (signals) could clobber the
data being stored.
it may be possible to optimize the code further using stwu, but I
wanted to avoid making any changes to the actual stack layout in this
commit. further improvements can be made separately if desired.
Rich Felker [Mon, 19 Nov 2012 00:58:15 +0000 (19:58 -0500)]
fix breakage from introducing bits header for sys/io.h
apparently some other archs have sys/io.h and should not break just
because they don't have the x86 port io functions. provide a blank
bits/io.h everywhere for now.
Rich Felker [Mon, 19 Nov 2012 00:31:58 +0000 (19:31 -0500)]
add port io functions to sys/io.h
based on proposal by Isaac Dunham. nonexistance of bits/io.h will
cause inclusion of sys/io.h to produce an error on archs that are not
supposed to have it. this is probably the desired behavior, but the
error message may be a bit unusual.
Rich Felker [Sun, 18 Nov 2012 21:58:05 +0000 (16:58 -0500)]
add missing const on powerpc FE_DFL_ENV
Rich Felker [Sun, 18 Nov 2012 21:31:14 +0000 (16:31 -0500)]
fenv support for ppc, untested
based on code sent to the mailing list by nsz, with minor changes.
Rich Felker [Sun, 18 Nov 2012 21:10:20 +0000 (16:10 -0500)]
fix feholdexcept -- it needs to clear exceptions after saving environment
Rich Felker [Sun, 18 Nov 2012 20:19:35 +0000 (15:19 -0500)]
Merge remote-tracking branch 'nsz/math'
rofl0r [Sun, 18 Nov 2012 04:14:40 +0000 (05:14 +0100)]
fcntl.h: O_SEARCH was missing for powerpc
put some macros that do not differ between architectures in the
main header and remove from bits.
restructure mips header so it has the same structure as the others.
Szabolcs Nagy [Sun, 18 Nov 2012 02:52:29 +0000 (03:52 +0100)]
math: use float constants in exp10f.c
use the 'f' suffix when a float constant is not representable
Szabolcs Nagy [Sun, 18 Nov 2012 02:49:16 +0000 (03:49 +0100)]
math: expl.c cleanup
raise overflow and underflow when necessary, fix various comments.
Szabolcs Nagy [Sun, 18 Nov 2012 02:42:09 +0000 (03:42 +0100)]
math: expf.c cleanup
similar to exp.c cleanup: use scalbnf, don't return excess precision,
drop some optimizatoins.
exp.c was changed to be more consistent with expf.c code.
rofl0r [Sun, 18 Nov 2012 01:38:41 +0000 (02:38 +0100)]
ppc socket.h: add forgotten cmsghdr
Rich Felker [Sun, 18 Nov 2012 01:04:36 +0000 (20:04 -0500)]
fix typo in dynamic linker path file loading code
fortunately the memory corruption could not hurt anything, but it
prevented clearing the final newline and thus prevented the last path
element from working.
Rich Felker [Sun, 18 Nov 2012 00:51:49 +0000 (19:51 -0500)]
add cleaned-up sys/mtio.h
this is mostly junk, but a few programs with tape-drive support
unconditionally include it, and it might be useful.
Rich Felker [Sat, 17 Nov 2012 23:42:16 +0000 (18:42 -0500)]
add stub versions of some missing optional pthread interfaces
priority inheritance is not yet supported, and priority protection
probably will not be supported ever unless there's serious demand for
it (it's a fairly heavy-weight feature).
per-thread cpu clocks would be nice to have, but to my knowledge linux
is still not capable of supporting them. glibc fakes them by using the
_process_ cpu-time clock and subtracting the thread creation time,
which gives seriously incorrect semantics (worse than not supporting
the feature at all), so until there's a way to do it right, it will
remain as a stub that always fails.
Szabolcs Nagy [Sat, 17 Nov 2012 22:39:39 +0000 (23:39 +0100)]
math: cleanup exp2.c exp2f.c and exp2l.c
* old code relied on sign extension on right shift
* exp2l ld64 wrapper was wrong
* use scalbn instead of bithacks
Rich Felker [Sat, 17 Nov 2012 22:25:17 +0000 (17:25 -0500)]
arg-skipping code for powerpc dynamic linker
this allows using the dynamic linker as a command to load programs.
Szabolcs Nagy [Sat, 17 Nov 2012 22:22:41 +0000 (23:22 +0100)]
math: exp.c clean up
overflow and underflow was incorrect when the result was not stored.
an optimization for the 0.5*ln2 < |x| < 1.5*ln2 domain was removed.
did various cleanups around static constants and made the comments
consistent with the code.
Rich Felker [Sat, 17 Nov 2012 03:22:34 +0000 (22:22 -0500)]
dynamic linking support for powerpc
incomplete but at least partly working. requires all files to be
compiled in the new "secure" plt model, not the old one that put plt
code in the data segment. TLS is untested but may work. invoking the
dynamic linker explicitly to load a program does not yet handle argv
correctly.
Rich Felker [Fri, 16 Nov 2012 06:19:42 +0000 (01:19 -0500)]
fix ppc stat structure definition
rofl0r [Fri, 16 Nov 2012 01:17:02 +0000 (02:17 +0100)]
fix powerpc termios.h and ioctl.h
Rich Felker [Fri, 16 Nov 2012 00:16:10 +0000 (19:16 -0500)]
disable SO_REUSEPORT in sys/socket.h
although a number is reserved for it, this option is not implemented
on Linux and does not work. defining it causes some applications to
use it, and subsequently break due to its failure.
Rich Felker [Thu, 15 Nov 2012 21:01:54 +0000 (16:01 -0500)]
ppc wchar_t is long, not int
Rich Felker [Thu, 15 Nov 2012 13:37:58 +0000 (08:37 -0500)]
Merge remote-tracking branch 'nsz/math'
Rich Felker [Thu, 15 Nov 2012 01:24:46 +0000 (20:24 -0500)]
update copyright file for recent contributions
Rich Felker [Wed, 14 Nov 2012 19:27:51 +0000 (14:27 -0500)]
fix indention with spaces in powerpc asm
Rich Felker [Wed, 14 Nov 2012 19:24:22 +0000 (14:24 -0500)]
fix powerpc atomic compare-and-swap function
previous version did not compare at all; it was just a fancy atomic
write. untested. further atomic fixes may be needed.
Rich Felker [Wed, 14 Nov 2012 19:08:33 +0000 (14:08 -0500)]
update ppc atomic code to match the endian-agnostic version on other archs
Rich Felker [Wed, 14 Nov 2012 19:04:10 +0000 (14:04 -0500)]
fix wchar limits mistakenly copied from arm
Rich Felker [Wed, 14 Nov 2012 19:01:39 +0000 (14:01 -0500)]
Merge remote-tracking branch 'ppc-port/ppc-squashed'
Szabolcs Nagy [Wed, 14 Nov 2012 08:27:54 +0000 (09:27 +0100)]
fenv: return FE_TONEAREST in dummy fegetround
Szabolcs Nagy [Wed, 14 Nov 2012 00:01:38 +0000 (01:01 +0100)]
math: ld80 invtrig cleanups
keeping only commonly used data in invtrigl
Szabolcs Nagy [Tue, 13 Nov 2012 20:54:32 +0000 (21:54 +0100)]
math: simplify hypot and hypotf using scalbn
this also fixes overflow/underflow raising and excess
precision issues (as those are handled well in scalbn)
Szabolcs Nagy [Tue, 13 Nov 2012 18:59:02 +0000 (19:59 +0100)]
math: turn off the STRICT_ASSIGN workaround by default
the volatile hack in STRICT_ASSIGN is only needed if
assignment is not respected and excess precision is kept.
gcc -fexcess-precision=standard and -ffloat-store both
respect assignment and musl use these flags by default.
i kept the macro for now so the workaround may be used
for bad compilers in the future.
rofl0r [Fri, 9 Nov 2012 22:36:55 +0000 (23:36 +0100)]
PPC port cleaned up, static linking works well now.
Richard Pennington [Fri, 9 Nov 2012 22:32:57 +0000 (23:32 +0100)]
import preliminary ppc work by rdp.
Szabolcs Nagy [Tue, 13 Nov 2012 12:34:45 +0000 (13:34 +0100)]
math: use '#pragma STDC FENV_ACCESS ON' when fenv is accessed
Szabolcs Nagy [Tue, 13 Nov 2012 09:55:35 +0000 (10:55 +0100)]
math: excess precision fix modf, modff, scalbn, scalbnf
old code was correct only if the result was stored (without the
excess precision) or musl was compiled with -ffloat-store.
now we use STRICT_ASSIGN to work around the issue.
(see note 160 in c11 section 6.8.6.4)
Szabolcs Nagy [Tue, 13 Nov 2012 09:30:40 +0000 (10:30 +0100)]
math: fix scalbn and scalbnf on overflow/underflow
old code was correct only if the result was stored (without the
excess precision) or musl was compiled with -ffloat-store.
(see note 160 in n1570.pdf section 6.8.6.4)
Szabolcs Nagy [Tue, 13 Nov 2012 09:12:07 +0000 (10:12 +0100)]
math: fix nextafter and nexttoward on maxdbl and maxflt input
old code (return x+x;) returns correct value and raises correct
flags only if the result is stored as double (or float)
Szabolcs Nagy [Tue, 13 Nov 2012 00:31:49 +0000 (01:31 +0100)]
complex: add C11 CMPLX macros and replace cpack with them
Szabolcs Nagy [Mon, 12 Nov 2012 23:49:55 +0000 (00:49 +0100)]
math: raise flags in logl.c on <= 0 arguments
Szabolcs Nagy [Mon, 12 Nov 2012 23:45:51 +0000 (00:45 +0100)]
math: fix logb*.c exceptions now that ilogb raises invalid
Szabolcs Nagy [Mon, 12 Nov 2012 23:30:06 +0000 (00:30 +0100)]
math: raise flags in log2l.c on <= 0 arguments, and fix volatile
Szabolcs Nagy [Mon, 12 Nov 2012 23:21:09 +0000 (00:21 +0100)]
math: raise exception flags in log1pl.c on <= -1 arguments
Szabolcs Nagy [Mon, 12 Nov 2012 22:58:18 +0000 (23:58 +0100)]
math: raise invalid flag in ilogb*.c on +-0, +-inf and nan
Szabolcs Nagy [Mon, 12 Nov 2012 22:41:48 +0000 (23:41 +0100)]
math: fix exception behaviour of expm1l.c on inf and nan
Szabolcs Nagy [Mon, 12 Nov 2012 22:13:28 +0000 (23:13 +0100)]
math: fix long double constants in exp10l.c
Rich Felker [Sun, 11 Nov 2012 21:08:38 +0000 (16:08 -0500)]
debloat src/thread tree but putting lots of junk in one file
POSIX includes mostly-useless attribute-get functions for each
attribute-set function, presumably out of some object-oriented
dogmatism. the get functions are not useful with the simple idiomatic
usage of attributes. there are of course possible valid uses of them
(like writing wrappers for pthread init functions that perform special
actions on the presence of certain attributes), but considering how
tiny these functions are anyway, little is lost by putting them all in
one file, and some build-time cost and archive-file-size benefits are
achieved.
Rich Felker [Sun, 11 Nov 2012 20:54:20 +0000 (15:54 -0500)]
report support of TPS option in unistd.h and sysconf
also update another newish feature in sysconf, stackaddr
Rich Felker [Sun, 11 Nov 2012 20:38:04 +0000 (15:38 -0500)]
add support for thread scheduling (POSIX TPS option)
linux's sched_* syscalls actually implement the TPS (thread
scheduling) functionality, not the PS (process scheduling)
functionality which the sched_* functions are supposed to have.
omitting support for the PS option (and having the sched_* interfaces
fail with ENOSYS rather than omitting them, since some broken software
assumes they exist) seems to be the only conforming way to do this on
linux.
Rich Felker [Sun, 11 Nov 2012 18:56:37 +0000 (13:56 -0500)]
fix clobber of edx in i386 vsyscall asm
this function does not obey the normal calling convention; like a
syscall instruction, it's expected not to clobber any registers except
the return value. clobbering edx could break callers that were reusing
the value cached in edx after the syscall returns.
Rich Felker [Fri, 9 Nov 2012 19:26:25 +0000 (14:26 -0500)]
always add memory streams to stdio open file list
per interpretation for austin group issue #626, fflush(0) and exit()
must block waiting for a lock if another thread has locked a memory
stream with flockfile. this adds some otherwise-unnecessary
synchronization cost to use of memory streams, but there was already a
synchronization cost calling malloc anyway.
previously the stream was only added to the open file list in
single-threaded programs, so that upon subsequent call to
pthread_create, locking could be turned on for the stream.
Rich Felker [Fri, 9 Nov 2012 18:49:40 +0000 (13:49 -0500)]
support ldso path files without final newline
Rich Felker [Fri, 9 Nov 2012 03:41:16 +0000 (22:41 -0500)]
change ldso path file logic to replace rather than add to search path
this change was originally intended just to avoid repeated attempts to
open a nonexistant /etc/ls-musl-$(ARCH).path file, but I realized it
also prevents the default paths from being searched when such a path
file exists. despite the potential to break existing usage, I believe
the new behavior is the right behavior, and it's better to fix it
sooner rather than later. with the old behavior, it was impossible to
inhibit search of default paths which might contain musl-incompatible
libs (or even libs from a different cpu arch, on multi-arch machines).
Rich Felker [Thu, 8 Nov 2012 22:20:50 +0000 (17:20 -0500)]
fix "configure --prefix=" and improve path/arg handling in configure
previously, empty string was treated as "use default". this is
apparently not compatible with standard configure semantics where an
empty prefix puts everything under /. the new logic should be a lot
cleaner and not suffer from such issues.
Rich Felker [Thu, 8 Nov 2012 22:04:20 +0000 (17:04 -0500)]
clean up sloppy nested inclusion from pthread_impl.h
this mirrors the stdio_impl.h cleanup. one header which is not
strictly needed, errno.h, is left in pthread_impl.h, because since
pthread functions return their error codes rather than using errno,
nearly every single pthread function needs the errno constants.
in a few places, rather than bringing in string.h to use memset, the
memset was replaced by direct assignment. this seems to generate much
better code anyway, and makes many functions which were previously
non-leaf functions into leaf functions (possibly eliminating a great
deal of bloat on some platforms where non-leaf functions require ugly
prologue and/or epilogue).
Rich Felker [Thu, 8 Nov 2012 21:39:41 +0000 (16:39 -0500)]
clean up stdio_impl.h
this header evolved to facilitate the extremely lazy practice of
omitting explicit includes of the necessary headers in individual
stdio source files; not only was this sloppy, but it also increased
build time.
now, stdio_impl.h is only including the headers it needs for its own
use; any further headers needed by source files are included directly
where needed.