|Age||Commit message (Collapse)||Author||Lines|
when the kernel is responsible for waking waiters on a robust mutex
whose owner died, it does not have a waiters count available and must
rely entirely on the waiter bit of the lock value.
normally, this bit is only set by newly arriving waiters, so it will
be clear if no new waiters arrived after the current owner obtained
the lock, even if there are other waiters present. leaving it clear is
desirable because it allows timed-lock operations to remove themselves
as waiters and avoid causing unnecessary futex wake syscalls. however,
for process-shared robust mutexes, we need to set the bit whenever
there are existing waiters so that the kernel will know to wake them.
for non-process-shared robust mutexes, the wake happens in userspace
and can look at the waiters count, so the bit does not need to be set
in the non-process-shared case.
when manipulating the robust list, the order of stores matters,
because the code may be asynchronously interrupted by a fatal signal
and the kernel will then access the robust list in what is essentially
an async-signal context.
previously, aliasing considerations made it seem unlikely that a
compiler could reorder the stores, but proving that they could not be
reordered incorrectly would have been extremely difficult. instead
I've opted to make all the pointers used as part of the robust list,
including those in the robust list head and in the individual mutexes,
in addition, the format of the robust list has been changed to point
back to the head at the end, rather than ending with a null pointer.
this is to match the documented kernel robust list ABI. the null
pointer, which was previously used, only worked because faults during
access terminate the robust list processing.
a robust mutex should not enter the unrecoverable status until it's
unlocked without marking it consistent. previously, flag 8 in the type
was used as an indication of unrecoverable, but only honored after
successful locking; this resulted in a race window where the
unrecoverable mutex could appear to a second thread as locked/busy
again while the first thread was in the process of observing it as
now, flag 8 is used to mean that the mutex is in the process of being
recovered, but not yet marked consistent. the flag only takes effect
in pthread_mutex_unlock, where it causes the value 0x40000000 (owner
dead flag, with old owner tid 0, an otherwise impossible state) to be
stored in the lock. subsequent lock attempts will interpret this state
per the resolution of Austin Group issue 755, the POSIX requirement
that ownership be enforced for recursive and error-checking mutexes
does not allow a random new thread to acquire ownership of an orphaned
mutex just because it happened to be assigned the same tid as the
original owner that exited with the mutex locked.
one possible fix for this issue would be to disallow the kernel thread
to terminate when it exited with mutexes held, permanently reserving
the tid against reuse. however, this does not solve the problem for
process-shared mutexes where lifetime cannot be controlled, so it was
the alternate approach I've taken is to reuse the robust mutex system
for non-robust recursive and error-checking mutexes. when a thread
exits, the kernel (or the new userspace robust-list code added in
commit b092f1c5fa9c048e12d002c7b972df5ecbe96d1d) will set the
owner-died bit for these orphaned mutexes, but since the mutex-type is
not robust, pthread_mutex_trylock will not allow a new owner to
acquire them. instead, they remain in a state of being permanently
locked, as desired.
the whole point of this locking is to prevent munmap, or mmap with
MAP_FIXED, from deallocating virtual addresses, or changing the
backing a given virtual address refers to, during certain race windows
involving self-synchronized unmapping or destruction of pthread
synchronization objects. there is no need for exclusion in the other
direction, so it suffices to take the lock momentarily and release it
before making the syscall, rather than holding it across the syscall.
the kernel always uses non-private wake when walking the robust list
when a thread or process exits, so it's not able to wake waiters
listening with the private futex flag. this problem is solved by doing
the equivalent in userspace as the last step of pthread_exit.
care is taken to remove mutexes from the robust list before unlocking
them so that the kernel will not attempt to access them again,
possibly after another thread locks them. this removal code can treat
the list as singly-linked, since no further code which would add or
remove items is able to run at this point. moreover, the pending
pointer is not needed since the mutexes being unlocked are all
process-local; in the case of asynchronous process termination, they
all cease to exist.
since a process-local robust mutex cannot come into existence without
a call to pthread_mutexattr_setrobust in the same process, the code
for userspace robust list processing is put in that source file, and
a weak alias to a dummy function is used to avoid pulling in this
bloat as part of pthread_exit in static-linked programs.
private-futex uses the virtual address of the futex int directly as
the hash key rather than requiring the kernel to resolve the address
to an underlying backing for the mapping in which it lies. for certain
usage patterns it improves performance significantly.
in many places, the code using futex __wake and __wait operations was
already passing a correct fixed zero or nonzero flag for the priv
argument, so no change was needed at the site of the call, only in the
__wake and __wait functions themselves. in other places, especially
where the process-shared attribute for a synchronization object was
not previously tracked, additional new code is needed. for mutexes,
the only place to store the flag is in the type field, so additional
bit masking logic is needed for accessing the type.
for non-process-shared condition variable broadcasts, the futex
requeue operation is unable to requeue from a private futex to a
process-shared one in the mutex structure, so requeue is simply
disabled in this case by waking all waiters.
for robust mutexes, the kernel always performs a non-private wake when
the owner dies. in order not to introduce a behavioral regression in
non-process-shared robust mutexes (when the owning thread dies), they
are simply forced to be treated as process-shared for now, giving
correct behavior at the expense of performance. this can be fixed by
adding explicit code to pthread_exit to do the right thing for
non-shared robust mutexes in userspace rather than relying on the
kernel to do it, and will be fixed in this way later.
since not all supported kernels have private futex support, the new
code detects EINVAL from the futex syscall and falls back to making
the call without the private flag. no attempt to cache the result is
made; caching it and using the cached value efficiently is somewhat
difficult, and not worth the complexity when the benefits would be
seen only on ancient kernels which have numerous other limitations and
C99 6.10.3p11 disallows such constructs
so use an #ifdef outside of the argument list of __syscall
isspace can be a bottleneck in a simple parser, inlining it
gives slightly smaller and faster code
src/locale/pleval.o already had this optimization, the size
change for other libc functions for i386 is
src/internal/intscan.o 2134 2118 -16
src/locale/dcngettext.o 1562 1552 -10
src/network/res_msend.o 1961 1940 -21
src/network/lookup_name.o 2627 2608 -19
src/network/getnameinfo.o 1814 1811 -3
src/network/lookup_serv.o 643 624 -19
src/stdio/vfscanf.o 2675 2663 -12
src/stdlib/atoll.o 117 107 -10
src/stdlib/atoi.o 95 91 -4
src/stdlib/atol.o 95 91 -4
src/time/strptime.o 1515 1503 -12
(TOTALS) 432451 432321 -130
when the dynamic loader is disabled, dlopen fails correctly but dlerror
did not return a human readable error string like it should have.
The function originates from SunOS 4.x in which the null argument
is allowed. glibc also handles this case.
the code which loads locale files was already rejecting locale names
containing slashes. however, LC_MESSAGES records a locale name even if
libc does not have a matching locale file, so that gettext or
application code can use the recorded locale name for message
translations to languages that libc does not support. this recorded
name was not being checked for slashes, meaning that such code could
potentially be tricked into directory traversal.
in addition, since the value of a locale category is sometimes used as
a pathname component by callers, the improved code rejects any value
beginning with a dot. this prevents traversal to the parent directory
via "..", use of the top-level locale directory via ".", and also
avoids "hidden" directories as a side effect.
finally, overly long locale names are now rejected (treated as an
unrecognized name and thus as an alias for C.UTF-8) rather than being
per the resolution of Austin Group issue #617, these are accepted for
XSI option in POSIX future and thus I'm treating them as standard
using an operator precedence parser the code size
became smaller and it is only slower by about %10
size of old vs new pleval.o on different archs:
(with inlined isspace added to pleval.c for now)
text data bss dec hex filename
828 0 0 828 33c pl.i386.o
1152 0 0 1152 480 pl.arm.o
1704 0 0 1704 6a8 pl.mips.o
1328 0 0 1328 530 pl.ppc.o
992 0 0 992 3e0 pl.x64.o
text data bss dec hex filename
693 0 0 693 2b5 pl.i386.o
972 0 0 972 3cc pl.arm.o
1276 0 0 1276 4fc pl.mips.o
1087 0 0 1087 43f pl.ppc.o
846 0 0 846 34e pl.x64.o
the previous implementations had several deficiencies, the most severe
of which was the inability to report unconfigured interfaces or
interfaces without ipv4 addresses. among the options discussed for
fixing this, using netlink turned out to be the one with the least
cost and most additional advantages. other improvements include:
if_nameindex now avoids duplicates in the list it produces, but still
includes legacy-style interface aliases if any are in use.
getifaddrs now reports hardware addresses and includes the scope_id
for link-local ipv6 addresses in the resulting address.
const parsing, depth accounting and failure handling was changed
a bit so the generated code is slightly smaller.
while the __mo_lookup backend can verify that the translated message
ends with a null terminator, is has no way to know nplurals and thus
no way to verify that sufficiently many null terminators are present
in the string to satisfy all plural forms. the code in dcngettext was
already attempting to avoid reading past the end of the mo file
mapping, but failed to do so because the strlen call itself could
over-read. using strnlen instead allows us to avoid the problem.
rather than just checking that the start of the string lies within the
mapping, also check that the nominal length remains within the
mapping, and that the null terminator is present at the nominal
length. this ensures that the caller, using the result as a C string,
will not read past the end of the mapping.
the nominal length is never exposed to the caller, but it's useful
internally to find where the null terminator should be without having
to restort to linear search via strnlen/memchr.
the new code in dcngettext was written by me, and the expression
evaluator by Szabolcs Nagy (nsz).
this commit replaces the stub implementations with working message
translation functions. translation units are factored so as to prevent
pulling in the legacy, non-library-safe functions which use a global
textdomain in modern code which is using the versions with an explicit
domain argument. bind_textdomain_codeset is also placed in its own
file since it should not be needed by most programs.
this implementation is still missing some features: the LANGUAGE
environment variable (for multiple fallback languages) is not honored,
and non-default plural-form rules are not supported. these issues will
be addressed in a later commit.
one notable difference from the GNU implementation is that there is no
default path for loading translation files. in principle one could be
added, but since the documented correct usage is to call the
bindtextdomain function, a default path is probably unnecessary.
for LC_MESSAGES, translation of strerror and similar literal message
functions is supported. for messages in other places (particularly the
dynamic linker) that use format strings, translation is not yet
supported. in order to make it possible and safe, such messages will
need to be refactored to separate the textual content from the format.
for LC_TIME, the day and month names and strftime-style format strings
provided by nl_langinfo are supported for translation. however there
may be limitations, as some of the original C-locale nl_langinfo
strings are non-unique and thus perhaps non-suitable as keys.
overall, the locale support activated by this commit should not be
seen as complete and polished but as a basis for beginning to test
locale functionality and implement locales.
these were removed from the standard but still offered as an extension
in langinfo.h, so nl_langinfo should support them.
due to a skipped slot and missing null terminator, the last few
strings were off by one or two slots from their item codes.
the core is based on a binary search; hash table is not used. both
native and reverse-endian mo files are supported. all offsets read
from the mapped mo file are checked against the mapping size to
prevent the possibility of reads outside the mapping.
this commit has no observable effects since there are not yet any
callers to the message translation code.
there is still no code which actually uses the loaded locale files, so
the main observable effect of this commit is that calls to setlocale
store and give back the names of the selected locales for the
remaining categories (LC_TIME, LC_COLLATE, LC_MONETARY) if a locale
file by the requested name could be loaded.
per POSIX (XBD 8.2) LC_*/LANG environment variables set to to the
empty string are supposed to be treated as if they were not set at
this function provides a way for third-party library code to use the
same logic that's used internally in libc for suppressing untrusted
input/state (e.g. the environment) when the application is running
with privleges elevated by the setuid or setgid bit or some other
mechanism. its semantics are intended to match the openbsd function by
the same name.
there was some question as to whether this function is necessary:
getauxval(AT_SECURE) was proposed as an alternative. however, this has
several drawbacks. the most obvious is that it asks programmers to be
aware of an implementation detail of ELF-based systems (the aux
vector) rather than simply the semantic predicate to be checked. and
trying to write a safe, reliable version of issetugid in terms of
getauxval is difficult. for example, early versions of the glibc
getauxval did not report ENOENT, which could lead to false negatives
if AT_SECURE was not present in the aux vector (this could probably
only happen when running on non-linux kernels under linux emulation,
since glibc does not support linux versions old enough to lack
AT_SECURE). as for musl, getauxval has always properly reported
errors, but prior to commit 7bece9c2095ee81f14b1088f6b0ba2f37fecb283,
the musl implementation did not emulate AT_SECURE if missing, which
would result in a false positive. since musl actually does partially
support kernels that lack AT_SECURE, this was problematic.
the intent is that library authors will use issetugid if its
availability is detected at build time, and only fall back to the
unreliable alternatives on systems that lack it.
patch by Brent Cook. commit message/rationale by Rich Felker.
this code path is used only on archs without the plain, non-at
syscalls, and only when the fstat syscall fails with EBADF on a valid
file descriptor. this in turn can happen only for O_PATH file
descriptors, and may not happen at all on the newer kernels needed for
supporting such archs.
with the flags argument omitted, spurious fstat failures may happen
when the argument register happens to have the AT_SYMLINK_NOFOLLOW bit
With the exception of a fenv implementation, the port is fully featured.
The port has been tested in or1ksim, the golden reference functional
simulator for OpenRISC 1000.
It passes all libc-test tests (except the math tests that
requires a fenv implementation).
The port assumes an or1k implementation that has support for
atomic instructions (l.lwa/l.swa).
Although it passes all the libc-test tests, the port is still
in an experimental state, and has yet experienced very little
this could happen on 2.4-series linux kernels that predate AT_SECURE
and possibly on other kernels that are emulating the linux syscall API
but not providing AT_SECURE in the aux vector at startup.
in principle applications should be checking errno anyway, but this
does not really work. to be secure, the caller would have to treat
ENOENT (indeterminate result) as possibly-suid and thereby disable
functionality in the typical non-suid usage case. and since glibc only
runs on kernels that provide AT_SECURE, applications written to the
glibc getauxval API might simply assume it succeeds.
this was originally added as a cheap but portable way to quell
warnings about reaching the end of a function that does not return,
but since _Exit is marked _Noreturn, it's not needed. removing it
makes the call to _Exit into a tail call and shaves off a few bytes of
code from minimal static programs.
per POSIX, the nmatch and pmatch arguments are ignored when the regex
was compiled with REG_NOSUB.
previously we detected this bug in configure and issued advice for a
workaround, but this turned out not to work. since then gcc 4.9.0 has
appeared in several distributions, and now 4.9.1 has been released
without a fix despite this being a wrong code generation bug which is
supposed to be a release-blocker, per gcc policy.
since the scope of the bug seems to affect only data objects (rather
than functions) whose definitions are overridable, and there are only
a very small number of these in musl, I am just changing them from
const to volatile for the time being. simply removing the const would
be sufficient to make gcc 4.9.1 work (the non-const case was
inadvertently fixed as part of another change in gcc), and this would
also be sufficient with 4.9.0 if we forced -O0 on the affected files
or on the whole build. however it's cleaner to just remove all the
broken compiler detection and use volatile, which will ensure that
they are never constant-folded. the quality of a non-broken compiler's
output should not be affected except for the fact that these objects
are no longer const and thus possibly add a few bytes to data/bss.
this change can be reconsidered and possibly reverted at some point in
the future when the broken gcc versions are no longer relevant.
the purpose of this logic is to avoid linking __stdio_exit unless any
stdio reads (which might require repositioning the file offset at exit
time) or writes (which might require flushing at exit time) could have
previously, exit called two wrapper functions for __stdio_exit named
__flush_on_exit and __seek_on_exit. both of these functions actually
performed both tasks (seek and flushing) by calling the underlying
__stdio_exit. in order to avoid doing this twice, an overridable data
object __towrite_used was used to cause __seek_on_exit to act as a nop
when __towrite was linked.
now, exit only makes one call, directly to __stdio_exit. this is
satisfiable by a weak dummy definition in exit.c, but the real
definition is pulled in by either __toread.c or __towrite.c through
their referencing a symbol which is defined only in __stdio_exit.c.
this was previously a no-op, somewhat intentionally, because I failed
to understand that it only has an effect when sending to the logging
facility fails and thus is not the nuisance that it would be if always
sent output to the console.
this behavior is no longer valid in general, and was never necessary.
if the LOG_PERROR option is set, output to stderr could still succeed.
also, when the LOG_CONS option is added, it will need syslog to
proceed even if opening the log socket fails.
this is a nonstandard feature, but easy and inexpensive to add. since
the corresponding macro has always been defined in our syslog.h, it
makes sense to actually support it. applications may reasonably be
using the presence of the macro to assume that the feature is
the behavior of omitting the 'header' part of the log message does not
seem to be well-documented, but matches other implementations (at
least glibc) which have this option.
based on a patch by Clément Vasseur, but simplified using %n.
errno must be saved upon vsyslog entry, otherwise its value could be
changed by some libc function before reaching the %m handler in
previously passing an empty string for name resulted in failure, as
expected, but only after spurious syscalls, and it produced confusing
errno values (and thus dlerror strings).
in addition to dlopen calls, this issue affected use of LD_PRELOAD
with trailing whitespace or colon characters.
r24 was wrongly being saved at a misaligned offset of 30 rather than
the correct offset of 40 in the jmp_buf. the exact effects of this
error have not been studied, but it's clear that the value of r24 was
lost across setjmp/longjmp and the saved values of r21 and/or r22 may
also have been corrupted.
if the order of object files in the static archive libc.a was not
respected by the linker, the old logic could wrongly cause POSIX
symbols outside of the ISO C namespace to be pulled into pure C
programs. this should not happen with well-behaved linkers, but
relying on the link order was a bad idea anyway.
files are renamed to better reflect their contents now that they don't
need names to control their order as members in the archive file.
1. failure to output a newline after the password is read
2. fd leaks via missing FD_CLOEXEC
3. fd leaks via failure-to-close when any of the standard streams are
closed at the time of the call
4. wrongful fallback to use of stdin when opening /dev/tty fails
5. wrongful use of stderr rather than /dev/tty for prompt
6. failure to report error reading password
the main motivation for this change is to remove the assumption that
the tid of the main thread is also the pid of the process. (the value
returned by the set_tid_address syscall was used to fill both fields
despite it semantically being the tid.) this is historically and
presently true on linux and unlikely to change, but it conceivably
could be false on other systems that otherwise reproduce the linux
only a few parts of the code were actually still using the cached pid.
in a couple places (aio and synccall) it was a minor optimization to
avoid a syscall. caching could be reintroduced, but lazily as part of
the public getpid function rather than at program startup, if it's
deemed important for performance later. in other places (cancellation
and pthread_kill) the pid was completely unnecessary; the tkill
syscall can be used instead of tgkill. this is actually a rather
subtle issue, since tgkill is supposedly a solution to race conditions
that can affect use of tkill. however, as documented in the commit
message for commit 7779dbd2663269b465951189b4f43e70839bc073, tgkill
does not actually solve this race; it just limits it to happening
within one process rather than between processes. we use a lock that
avoids the race in pthread_kill, and the use in the cancellation
signal handler is self-targeted and thus not subject to tid reuse
races, so both are safe regardless of which syscall (tgkill or tkill)
this change is presently non-functional since the callees do not yet
use their locale argument for anything.
this is mainly done for consistency with the ctype functions and to
declutter the src/locale directory.
the main practical purposes of this commit are to remove a huge amount
of clutter from the src/locale directory, to cut down on the length of
the $(AR) and $(LD) command lines, and to reduce the amount of space
wasted by object file headers in the static libc.a. build time may
also be reduced, though this has not been measured.
as an additional justification, if there ever were a need for the
behavior of these functions to vary by locale, it would be necessary
for the non-_l versions to call the _l versions, so that linking the
former without the latter would not be possible anyway.