A child process created by posix_spawn reports errors to its parent via
a pipe, retrying infinitely on any write error to prevent falsely
reporting success. If the (original) parent dies before write is
attempted, there is nobody to report to, but the child will remain
stuck in the write loop forever if SIGPIPE is blocked or ignored.
Fix this by not retrying write if it fails with EPIPE.
this commit should make no codegen change for existing archs, but is a
prerequisite for new archs including riscv32. the wait4 emulation
backend provides both cancellable and non-cancellable variants because
waitpid is required to be a cancellation point, but all of our other
uses are not, and most of them cannot be.
based on patch by Stefan O'Rear.
the clone() function has been effectively unusable since it was added,
due to producing a child process with inconsistent state. in
particular, the child process's thread structure still contains the
tid, thread list pointers, thread count, and robust list for the
parent. this will cause malfunction in interfaces that attempt to use
the tid or thread list, some of which are specified to be
async-signal-safe.
this patch attempts to make clone() consistent in a _Fork-like sense.
as in _Fork, when the parent process is multi-threaded, the child
process inherits an async-signal context where it cannot call
AS-unsafe functions, but its context is now intended to be safe for
calling AS-safe functions. making clone fork-like would also be a
future option, if it turns out that this is what makes sense to
applications, but it's not done at this time because the changes would
be more invasive.
in the case where the CLONE_VM flag is used, clone is only vfork-like,
not _Fork-like. in particular, the child will see itself as having the
parent's tid, and cannot safely call any libc functions but one of the
exec family or _exit.
handling of flags and variadic arguments is also changed so that
arguments are only consumed with flags that indicate their presence,
and so that flags which produce an inconsistent state are disallowed
(reported as EINVAL). in particular, all libc functions carry a
contract that they are only callable with ABI requirements met, which
includes having a valid thread pointer to a thread structure that's
unique within the process, and whose contents are opaque and only able
to be setup internally by the implementation. the only way for an
application to use flags that violate these requirements without
executing any libc code is to perform the syscall from
application-provided asm.
apparently Linux clears the registered exit futex address on fork.
this means that, if after forking the child process becomes
multithreaded and the original thread exits, the thread list will
never be unlocked, and future attempts to use the thread list will
deadlock.
re-register the exit futex address after _Fork in the child to ensure
that it's preserved.
commit 167390f055 seems to have
overlooked the presence of a lock here, probably because it was one of
the exceptions not using LOCK() but a rwlock.
as such, it can't be added to the generic table of locks to take, so
add an explicit atfork function for the pthread keys table. the order
it is called does not particularly matter since nothing else in libc
but pthread_exit interacts with keys.
as reported by Alexey Izbyshev, there is a lock order inversion
deadlock between the malloc lock and aio maplock at MT-fork time:
_Fork attempts to take the aio maplock while fork already has the
malloc lock, but a concurrent aio operation holding the maplock may
attempt to allocate memory.
move the __aio_atfork calls in the parent from _Fork to fork, and
reorder the lock before most other locks, since nothing else depends
on aio(*). this leaves us with the possibility that the child will not
be able to obtain the read lock, if _Fork is used directly and happens
concurrent with an aio operation. however, in that case, the child
context is an async signal context that cannot call any further aio
functions, so all we need is to ensure that close does not attempt to
perform any aio cancellation. this can be achieved just by nulling out
the map pointer.
(*) even if other functions call close, they will only need a read
lock, not a write lock, and read locks being recursive ensures they
can obtain it. moreover, the number of read references held is bounded
by something like twice the number of live threads, meaning that the
read lock count cannot saturate.
ever since commit 8f11e6127f introduced
the thread list lock, this has been wrong. initially, it was wrong via
calling free from the context with the thread list lock held. commit
aa5a9d15e0 deferred the unsafe free but
added a lock, which was also unsafe. in particular, it could deadlock
if code holding freebuf_queue_lock was interrupted by a signal handler
that takes the thread list lock.
commit 4d5aa20a94 observed that there
was a lock here but failed to notice that it's invalid.
there is no easy solution to this problem with locks; any attempt at
solving it while still using locks would require the lock to be an
AS-safe one (blocking signals on each access to the dlerror buffer
list to check if there's deferred free work to be done) which would be
excessively costly, and there are also lock order considerations with
respect to how the lock would be handled at fork.
instead, just use an atomic list.
The generic vfork implementation uses clone(SIGCHLD) which has fork
semantics.
Implement vfork as clone(SIGCHLD|CLONE_VM|CLONE_VFORK, 0) instead which
has vfork semantics. (stack == 0 means sp is unchanged in the child.)
Some users rely on vfork semantics when memory overcommit is disabled
or when the vfork child runs code that synchronizes with the parent
process (non-conforming).
this makes it possible to perform actions on file actions objects with
a libc-internal lock held without creating lock order relationships
that are silently imposed on an application-provided malloc.
these functions are specified to fail with EBADF on negative fd
arguments. apart from close, they are also specified to fail if the
value exceeds OPEN_MAX, but as written it is not clear that this
imposes any requirement when OPEN_MAX is not defined, and it's
undesirable to impose a dynamic limit (via setrlimit) here since the
limit at the time of posix_spawn may be different from the limit at
the time of setting up the file actions. this may require revisiting
later.
as the outcome of Austin Group tracker issue #62, future editions of
POSIX have dropped the requirement that fork be AS-safe. this allows
but does not require implementations to synchronize fork with internal
locks and give forked children of multithreaded parents a partly or
fully unrestricted execution environment where they can continue to
use the standard library (per POSIX, they can only portably use
AS-safe functions).
up until recently, taking this allowance did not seem desirable.
however, commit 8ed2bd8bfc exposed the
extent to which applications and libraries are depending on the
ability to use malloc and other non-AS-safe interfaces in MT-forked
children, by converting latent very-low-probability catastrophic state
corruption into predictable deadlock. dealing with the fallout has
been a huge burden for users/distros.
while it looks like most of the non-portable usage in applications
could be fixed given sufficient effort, at least some of it seems to
occur in language runtimes which are exposing the ability to run
unrestricted code in the child as part of the contract with the
programmer. any attempt at fixing such contracts is not just a
technical problem but a social one, and is probably not tractable.
this patch extends the fork function to take locks for all libc
singletons in the parent, and release or reset those locks in the
child, so that when the underlying fork operation takes place, the
state protected by these locks is consistent and ready for the child
to use. locking is skipped in the case where the parent is
single-threaded so as not to interfere with legacy AS-safety property
of fork in single-threaded programs. lock order is mostly arbitrary,
but the malloc locks (including bump allocator in case it's used) must
be taken after the locks on any subsystems that might use malloc, and
non-AS-safe locks cannot be taken while the thread list lock is held,
imposing a requirement that it be taken last.
commit bd153422f2 reintroduced the bug
fixed in c21051e90c by refactoring the
__syscall_ret into _Fork where it once again runs before the atfork
handlers are called. since _Fork is a public interface that sets
errno, this can't be fixed the way it was fixed last time without
making new internal interfaces. instead, just save errno, and restore
it only on error to ensure that a value of 0 is never restored.
this change prevents the child created concurrently with abort from
seeing the SIGABRT disposition change from SIG_IGN to SIG_DFL (other
changes are not visible anyway) and prevents leaking the write end of
the child pipe to children created by fork in another thread, which
may block return of posix_spawn indefinitely if the forked child does
not exit or exec.
along with other changes, this suggests that __abort_lock should
perhaps eventually be renamed to reflect that it's becoming a broader
lock on related "process lifetime" state.
the _Fork interface is defined for future issue of POSIX as the
outcome of Austin Group issue 62, which drops the AS-safety
requirement for fork, and provides an AS-safe replacement that does
not run the registered atfork handlers.
if the multithreaded parent forked while another thread was calling
sigaction for SIGABRT or calling abort, the child could inherit a lock
state in which future calls to abort will deadlock, or in which the
disposition for SIGABRT has already been reset to SIG_DFL. this is
nonconforming since abort is AS-safe and permitted to be called
concurrently with fork or in the MT-forked child.
previously, if a file descriptor had aio operations pending in the
parent before fork, attempting to close it in the child would attempt
to cancel a thread belonging to the parent. this could deadlock, fail,
or crash the whole process of the cancellation signal handler was not
yet installed in the parent. in addition, further use of aio from the
child could malfunction or deadlock.
POSIX specifies that async io operations are not inherited by the
child on fork, so clear the entire aio fd map in the child, and take
the aio map lock (with signals blocked) across the fork so that the
lock is kept in a consistent state.
the child is single-threaded, but may still need to synchronize with
last changes made to memory by another thread in the parent, so set
need_locks to -1 whereby the next lock-taker will drop to 0 and
prevent further barriers/locking.
these are presently extensions, thus named with _np to match glibc and
other implementations that provide them; however they are likely to be
standardized in the future without the _np suffix as a result of
Austin Group issue 1208. if so, both names will be kept as aliases.
as reported by Tavian Barnes, a dup2 file action for the internal pipe
fd used by posix_spawn could cause it to remain open after execve and
allow the child to write an artificial error into it, confusing the
parent. POSIX allows internal use of file descriptors by the
implementation, with undefined behavior for poking at them, so this is
not a conformance problem, but it seems preferable to diagnose and
prevent the error when we can do so easily.
catch attempts to apply a dup2 action to the internal pipe fd and
emulate EBADF for it instead.
synccall may be called by AS-safe functions such as setuid/setgid after
fork. although fork() resets libc.threads_minus_one, causing synccall to
take the single-threaded path, synccall still takes the thread list
lock. This lock may be held by another thread if for example fork()
races with pthread_create(). After fork(), the value of the lock is
meaningless, so clear it.
maintainer's note: commit 8f11e6127f and
e4235d7067 introduced this regression.
the state protected by this lock is the linked list, which is entirely
replaced in the child path of fork (next=prev=self), so resetting it
is semantically sound.
The result is the same but takes less code.
Note that __execvpe calls getenv which calls __strchrnul so even
using static output the size of the executable won't grow.
the hard problem here is unlinking threads from a list when they exit
without creating a window of inconsistency where the kernel task for a
thread still exists and is still executing instructions in userspace,
but is not reflected in the list. the magic solution here is getting
rid of per-thread exit futex addresses (set_tid_address), and instead
using the exit futex to unlock the global thread list.
since pthread_join can no longer see the thread enter a detach_state
of EXITED (which depended on the exit futex address pointing to the
detach_state), it must now observe the unlocking of the thread list
lock before it can unmap the joined thread and return. it doesn't
actually have to take the lock. for this, a __tl_sync primitive is
offered, with a signature that will allow it to be enhanced for quick
return even under contention on the lock, if needed. for now, the
exiting thread always performs a futex wake on its detach_state. a
future change could optimize this out except when there is already a
joiner waiting.
initial/dynamic variants of detached state no longer need to be
tracked separately, since the futex address is always set to the
global list lock, not a thread-local address that could become invalid
on detached thread exit. all detached threads, however, must perform a
second sigprocmask syscall to block implementation-internal signals,
since locking the thread list with them already blocked is not
permissible.
the arch-independent C version of __unmapself no longer needs to take
a lock or setup its own futex address to release the lock, since it
must necessarily be called with the thread list lock already held,
guaranteeing exclusive access to the temporary stack.
changes to libc.threads_minus_1 no longer need to be atomic, since
they are guarded by the thread list lock. it is largely vestigial at
this point, and can be replaced with a cheaper boolean indicating
whether the process is multithreaded at some point in the future.
libc.h was intended to be a header for access to global libc state and
related interfaces, but ended up included all over the place because
it was the way to get the weak_alias macro. most of the inclusions
removed here are places where weak_alias was needed. a few were
recently introduced for hidden. some go all the way back to when
libc.h defined CANCELPT_BEGIN and _END, and all (wrongly implemented)
cancellation points had to include it.
remaining spurious users are mostly callers of the LOCK/UNLOCK macros
and files that use the LFS64 macro to define the awful *64 aliases.
in a few places, new inclusion of libc.h is added because several
internal headers no longer implicitly include libc.h.
declarations for __lockfile and __unlockfile are moved from libc.h to
stdio_impl.h so that the latter does not need libc.h. putting them in
libc.h made no sense at all, since the macros in stdio_impl.h are
needed to use them correctly anyway.
this was added so that posix_spawn and possibly other functionality
could be implemented in terms of vfork, but that turned out to be
unsafe. any such usage needs __clone with proper handling of stack
lifetime.
commits leading up to this one have moved the vast majority of
libc-internal interface declarations to appropriate internal headers,
allowing them to be type-checked and setting the stage to limit their
visibility. the ones that have not yet been moved are mostly
namespace-protected aliases for standard/public interfaces, which
exist to facilitate implementing plain C functions in terms of POSIX
functionality, or C or POSIX functionality in terms of extensions that
are not standardized. some don't quite fit this description, but are
"internally public" interfacs between subsystems of libc.
rather than create a number of newly-named headers to declare these
functions, and having to add explicit include directives for them to
every source file where they're needed, I have introduced a method of
wrapping the corresponding public headers.
parallel to the public headers in $(srcdir)/include, we now have
wrappers in $(srcdir)/src/include that come earlier in the include
path order. they include the public header they're wrapping, then add
declarations for namespace-protected versions of the same interfaces
and any "internally public" interfaces for the subsystem they
correspond to.
along these lines, the wrapper for features.h is now responsible for
the definition of the hidden, weak, and weak_alias macros. this means
source files will no longer need to include any special headers to
access these features.
over time, it is my expectation that the scope of what is "internally
public" will expand, reducing the number of source files which need to
include *_impl.h and related headers down to those which are actually
implementing the corresponding subsystems, not just using them.
previously, a common __posix_spawnx backend was used that accepted an
additional argument for the execve variant to call in the child. this
moderately bloated up the posix_spawn function, shuffling arguments
between stack and/or registers to call a 7-argument function from a
6-argument one.
instead, tuck the exec function pointer in an unused part of the
(large) pthread_spawnattr_t structure, and have posix_spawnp duplicate
the attributes and fill in a pointer to __execvpe. the net code size
change is minimal, but the weight is shifted to the "heavier" function
which already pulls in more dependencies.
as a bonus, we get rid of an external symbol (__posix_spawnx) that had
no really good place for a declaration because it shouldn't have
existed to begin with.
syscall.h was chosen as the header to declare it, since its intended
usage is alongside syscalls as a fallback for operations the direct
syscall does not support.
the value 0x7f00 (as if by _exit(127)) is specified only for the case
where the child is created but then fails to exec the shell, since
traditional fork+exec implementations do not admit reporting an error
via errno in this case without additional machinery. it's unclear
whether an implementation not subject to this failure mode needs to
emulate it; one could read the standard as requiring that. if so,
additional code will need to be added to map posix_spawn errors into
the form system is expected to return. but for now, returning -1 to
indicate an error is significantly better behavior than always
reporting failures as if the shell failed to exec after fork.
this is more extensible if we need to consider additional errors, and
more efficient as long as the compiler does not know it can cache the
result of __errno_location (a surprisingly complex issue detailed in
commit a603a75a72).
It's better to make execvp continue PATH search on ENOTDIR rather than
issuing an error. Bogus entries should not render rest of PATH invalid.
Maintainer's note: POSIX seems to require the search to continue like
this as part of XBD 8.3 Other Environment Variables. Only errors that
conclusively determine non-existence are candidates for continuing;
otherwise for consistency we have to report the error.
If the syscall fails, errno must be set correctly for the caller.
There's no guarantee that the handlers registered with pthread_atfork
won't clobber errno, so we need to ensure it gets set after they are
called.
the resolution to Austin Group issue #411 defined new semantics for
the posix_spawn dup2 file action in the (previously useless) case
where src and dest fd are equal. future issues will require the dup2
file action to remove the close-on-exec flag. without this change,
passing fds to a child with posix_spawn while avoiding fd-leak races
in a multithreaded parent required a complex dance with temporary fds.
based on patch by Petr Skocik. changes were made to preserve the
80-column formatting of the function and to remove code that became
unreachable as a result of the new functionality.
execvpe stack-allocates a buffer used to hold the full path
(combination of a PATH entry and the program name)
while searching through $PATH, so at least
NAME_MAX+PATH_MAX is needed.
The stack size can be made conditionally smaller
(the current 1024 appears appropriate)
should this larger size be burdensome in those situations.
per POSIX, EINVAL is not a mandatory error, only an optional one. but
reporting unsupported flags allows an application to fallback
gracefully when a requested feature is not supported. this is not
helpful now, but it may be in the future if additional flags are
added.
had this checking been present before, applications would have been
able to check for the newly-added POSIX_SPAWN_SETSID feature (added in
commit bb439bb171) at runtime.
nominally the low bits of the trap number on sh are the number of
syscall arguments, but they have never been used by the kernel, and
some code making syscalls does not even know the number of arguments
and needs to pass an arbitrary high number anyway.
sh3/sh4 traditionally used the trap range 16-31 for syscalls, but part
of this range overlapped with hardware exceptions/interrupts on sh2
hardware, so an incompatible range 32-47 was chosen for sh2.
using trap number 31 everywhere, since it's in the existing sh3/sh4
range and does not conflict with sh2 hardware, is a proposed
unification of the kernel syscall convention that will allow binaries
to be shared between sh2 and sh3/sh4. if this is not accepted into the
kernel, we can refit the sh2 target with runtime selection mechanisms
for the trap number, but doing so would be invasive and would entail
non-trivial overhead.
since 1.1.0, musl has nominally required a thread pointer to be setup.
most of the remaining code that was checking for its availability was
doing so for the sake of being usable by the dynamic linker. as of
commit 71f099cb7d, this is no longer
necessary; the thread pointer is now valid before any libc code
(outside of dynamic linker bootstrap functions) runs.
this commit essentially concludes "phase 3" of the "transition path
for removing lazy init of thread pointer" project that began during
the 1.1.0 release cycle.
as a result of commit 12e1e32468, kernel
processing of the robust list is only needed for process-shared
mutexes. previously the first attempt to lock any owner-tracked mutex
resulted in robust list initialization and a set_robust_list syscall.
this is no longer necessary, and since the kernel's record of the
robust list must now be cleared at thread exit time for detached
threads, optimizing it out is more worthwhile than before too.
the specification for execvp itself is unclear as to whether
encountering a file that cannot be executed due to EACCES during the
PATH search is a mandatory error condition; however, XBD 8.3's
specification of the PATH environment variable clarifies that the
search continues until a file with "appropriate execution permissions"
is found.
since it seems undesirable/erroneous to report ENOENT rather than
EACCES when an early path element has a non-executable file and all
later path elements lack any file by the requested name, the new code
stores a flag indicating that EACCES was seen and sets errno back to
EACCES in this case.
the write function is a cancellation point and accesses thread-local
state belonging to the calling thread in the parent process. since
cancellation is blocked for the duration of posix_spawn, this is
probably safe, but it's fragile and unnecessary. making the syscall
directly is just as easy and clearly safe.
the resolution of austin group issue #370 removes the requirement that
posix_spawn fail when the close file action is performed on an
already-closed fd. since there are no other meaningful errors for
close, just ignoring the return value completely is the simplest fix.
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
syscall api/abi.
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 7779dbd266, 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)
is used.
Alexey Izbyshev
Rich Felker
Pedro Falcato
Szabolcs Nagy
Samuel Holland
Frediano Ziglio
Patrick Oppenlander
Joseph C. Sible
Przemyslaw Pawelczyk
Bobby Bingham
Will Dietz