@@ -126,6 +126,18 @@ struct priority_protection_data
};
+/* Define a possible thread state on 'joinstate' field. The value will be
+ cleared by the kernel when the thread terminates (CLONE_CHILD_CLEARTID),
+ so THREAD_STATE_EXITED must be 0. */
+enum
+ {
+ THREAD_STATE_EXITED = 0,
+ THREAD_STATE_EXITING,
+ THREAD_STATE_JOINABLE,
+ THREAD_STATE_DETACHED,
+ };
+
+
/* Thread descriptor data structure. */
struct pthread
{
@@ -168,8 +180,7 @@ struct pthread
GL (dl_stack_user) list. */
list_t list;
- /* Thread ID - which is also a 'is this thread descriptor (and
- therefore stack) used' flag. */
+ /* Thread ID set by the kernel with CLONE_PARENT_SETTID. */
pid_t tid;
/* Ununsed. */
@@ -346,15 +357,8 @@ struct pthread
hp_timing_t cpuclock_offset_ununsed;
#endif
- /* If the thread waits to join another one the ID of the latter is
- stored here.
-
- In case a thread is detached this field contains a pointer of the
- TCB if the thread itself. This is something which cannot happen
- in normal operation. */
- struct pthread *joinid;
- /* Check whether a thread is detached. */
-#define IS_DETACHED(pd) ((pd)->joinid == (pd))
+ /* The current thread state defined by the THREAD_STATE_* enumeration. */
+ unsigned int joinstate;
/* The result of the thread function. */
void *result;
@@ -31,7 +31,7 @@ extern size_t __nptl_stack_cache_maxsize attribute_hidden;
static inline bool
__nptl_stack_in_use (struct pthread *pd)
{
- return pd->tid <= 0;
+ return atomic_load_relaxed (&pd->joinstate) == THREAD_STATE_EXITED;
}
/* Remove the stack ELEM from its list. */
@@ -72,7 +72,8 @@ __pthread_cancel (pthread_t th)
{
volatile struct pthread *pd = (volatile struct pthread *) th;
- if (pd->tid == 0)
+ int state = atomic_load_acquire (&pd->joinstate);
+ if (state == THREAD_STATE_EXITED || state == THREAD_STATE_EXITING)
/* The thread has already exited on the kernel side. Its outcome
(regular exit, other cancelation) has already been
determined. */
@@ -30,7 +30,7 @@ ___pthread_clockjoin_np64 (pthread_t threadid, void **thread_return,
return EINVAL;
return __pthread_clockjoin_ex (threadid, thread_return,
- clockid, abstime, true);
+ clockid, abstime);
}
#if __TIMESIZE == 64
@@ -287,7 +287,7 @@ static int create_thread (struct pthread *pd, const struct pthread_attr *attr,
.flags = clone_flags,
.pidfd = (uintptr_t) &pd->tid,
.parent_tid = (uintptr_t) &pd->tid,
- .child_tid = (uintptr_t) &pd->tid,
+ .child_tid = (uintptr_t) &pd->joinstate,
.stack = (uintptr_t) stackaddr,
.stack_size = stacksize,
.tls = (uintptr_t) tp,
@@ -352,12 +352,14 @@ start_thread (void *arg)
and free any resource prior return to the pthread_create caller. */
setup_failed = pd->setup_failed == 1;
if (setup_failed)
- pd->joinid = NULL;
+ pd->joinstate = THREAD_STATE_JOINABLE;
/* And give it up right away. */
lll_unlock (pd->lock, LLL_PRIVATE);
if (setup_failed)
+ /* No need to clear the tid here, pthread_create() will join the
+ thread prior returning to caller. */
goto out;
}
@@ -489,6 +491,23 @@ start_thread (void *arg)
the breakpoint reports TD_THR_RUN state rather than TD_THR_ZOMBIE. */
atomic_bit_set (&pd->cancelhandling, EXITING_BIT);
+
+ /* CONCURRENCY NOTES:
+
+ Concurrent pthread_detach() will either set state to
+ THREAD_STATE_DETACHED or wait for the thread to terminate. The exiting
+ state set here is set so a pthread_join() wait until all the required
+ cleanup steps are done.
+
+ The 'prevstate' field will be used to determine who is responsible to
+ call __nptl_free_tcb below. */
+
+ unsigned int prevstate;
+ do
+ prevstate = atomic_load_relaxed (&pd->joinstate);
+ while (!atomic_compare_exchange_weak_acquire (&pd->joinstate, &prevstate,
+ THREAD_STATE_EXITING));
+
if (__glibc_unlikely (atomic_decrement_and_test (&__nptl_nthreads)))
/* This was the last thread. */
exit (0);
@@ -567,17 +586,17 @@ start_thread (void *arg)
pd->setxid_futex = 0;
}
- /* If the thread is detached free the TCB. */
- if (IS_DETACHED (pd))
- /* Free the TCB. */
+ if (prevstate == THREAD_STATE_DETACHED)
__nptl_free_tcb (pd);
+ pd->tid = 0;
+
out:
/* We cannot call '_exit' here. '_exit' will terminate the process.
The 'exit' implementation in the kernel will signal when the
process is really dead since 'clone' got passed the CLONE_CHILD_CLEARTID
- flag. The 'tid' field in the TCB will be set to zero.
+ flag. The 'joinstate' field in the TCB will be set to zero.
rseq TLS is still registered at this point. Rely on implicit
unregistration performed by the kernel on thread teardown. This is not a
@@ -689,7 +708,9 @@ __pthread_create_2_1 (pthread_t *newthread, const pthread_attr_t *attr,
/* Initialize the field for the ID of the thread which is waiting
for us. This is a self-reference in case the thread is created
detached. */
- pd->joinid = iattr->flags & ATTR_FLAG_DETACHSTATE ? pd : NULL;
+ pd->joinstate = iattr->flags & ATTR_FLAG_DETACHSTATE
+ ? THREAD_STATE_DETACHED
+ : THREAD_STATE_JOINABLE;
/* The debug events are inherited from the parent. */
pd->eventbuf = self->eventbuf;
@@ -848,10 +869,11 @@ __pthread_create_2_1 (pthread_t *newthread, const pthread_attr_t *attr,
/* Similar to pthread_join, but since thread creation has failed at
startup there is no need to handle all the steps. */
- pid_t tid;
- while ((tid = atomic_load_acquire (&pd->tid)) != 0)
- __futex_abstimed_wait_cancelable64 ((unsigned int *) &pd->tid,
- tid, 0, NULL, LLL_SHARED);
+ unsigned int state;
+ while ((state = atomic_load_acquire (&pd->joinstate))
+ != THREAD_STATE_EXITED)
+ __futex_abstimed_wait_cancelable64 (&pd->joinstate, state, 0,
+ NULL, LLL_SHARED);
}
/* State (c) or (d) and we have ownership of PD (see CONCURRENCY
@@ -25,32 +25,28 @@ ___pthread_detach (pthread_t th)
{
struct pthread *pd = (struct pthread *) th;
- /* Make sure the descriptor is valid. */
- if (INVALID_NOT_TERMINATED_TD_P (pd))
- /* Not a valid thread handle. */
- return ESRCH;
+ /* CONCURRENCY NOTES:
- int result = 0;
+ Concurrent pthread_detach will return EINVAL for the case the thread
+ is already detached (THREAD_STATE_DETACHED). POSIX states it is
+ undefined to call pthread_detach if TH refers to a non joinable thread.
- /* Mark the thread as detached. */
- if (atomic_compare_and_exchange_bool_acq (&pd->joinid, pd, NULL))
+ For the case the thread is being terminated (THREAD_STATE_EXITING),
+ pthread_detach will responsible to clean up the stack. */
+
+ unsigned int prevstate = atomic_load_relaxed (&pd->joinstate);
+ do
{
- /* There are two possibilities here. First, the thread might
- already be detached. In this case we return EINVAL.
- Otherwise there might already be a waiter. The standard does
- not mention what happens in this case. */
- if (IS_DETACHED (pd))
- result = EINVAL;
+ if (prevstate != THREAD_STATE_JOINABLE)
+ {
+ if (prevstate == THREAD_STATE_DETACHED)
+ return EINVAL;
+ return __pthread_join (th, 0);
+ }
}
- else
- /* Check whether the thread terminated meanwhile. In this case we
- will just free the TCB. */
- if ((pd->cancelhandling & EXITING_BITMASK) != 0)
- /* Note that the code in __free_tcb makes sure each thread
- control block is freed only once. */
- __nptl_free_tcb (pd);
-
- return result;
+ while (!atomic_compare_exchange_weak_acquire (&pd->joinstate, &prevstate,
+ THREAD_STATE_DETACHED));
+ return 0;
}
versioned_symbol (libc, ___pthread_detach, pthread_detach, GLIBC_2_34);
libc_hidden_ver (___pthread_detach, __pthread_detach)
@@ -52,7 +52,7 @@ __pthread_getattr_np (pthread_t thread_id, pthread_attr_t *attr)
iattr->flags = thread->flags;
/* The thread might be detached by now. */
- if (IS_DETACHED (thread))
+ if (atomic_load_acquire (&thread->joinstate) == THREAD_STATE_DETACHED)
iattr->flags |= ATTR_FLAG_DETACHSTATE;
/* This is the guardsize after adjusting it. */
@@ -22,7 +22,7 @@ int
___pthread_join (pthread_t threadid, void **thread_return)
{
return __pthread_clockjoin_ex (threadid, thread_return, 0 /* Ignored */,
- NULL, true);
+ NULL);
}
versioned_symbol (libc, ___pthread_join, pthread_join, GLIBC_2_34);
libc_hidden_ver (___pthread_join, __pthread_join)
@@ -22,110 +22,70 @@
#include <time.h>
#include <futex-internal.h>
-static void
-cleanup (void *arg)
+/* Check for a possible deadlock situation where the threads are waiting for
+ each other to finish. Note that this is a "may" error. To be 100% sure we
+ catch this error we would have to lock the data structures but it is not
+ necessary. In the unlikely case that two threads are really caught in this
+ situation they will deadlock. It is the programmer's problem to figure
+ this out. */
+static inline bool
+check_for_deadlock (struct pthread *pd)
{
- /* If we already changed the waiter ID, reset it. The call cannot
- fail for any reason but the thread not having done that yet so
- there is no reason for a loop. */
struct pthread *self = THREAD_SELF;
- atomic_compare_exchange_weak_acquire (&arg, &self, NULL);
+ return ((pd == self
+ || (atomic_load_acquire (&self->joinstate) == THREAD_STATE_DETACHED
+ && (pd->cancelhandling
+ & (CANCELING_BITMASK | CANCELED_BITMASK | EXITING_BITMASK
+ | TERMINATED_BITMASK)) == 0))
+ && !cancel_enabled_and_canceled (self->cancelhandling));
}
int
__pthread_clockjoin_ex (pthread_t threadid, void **thread_return,
clockid_t clockid,
- const struct __timespec64 *abstime, bool block)
+ const struct __timespec64 *abstime)
{
struct pthread *pd = (struct pthread *) threadid;
- /* Make sure the descriptor is valid. */
- if (INVALID_NOT_TERMINATED_TD_P (pd))
- /* Not a valid thread handle. */
- return ESRCH;
-
- /* Is the thread joinable?. */
- if (IS_DETACHED (pd))
- /* We cannot wait for the thread. */
- return EINVAL;
-
- struct pthread *self = THREAD_SELF;
- int result = 0;
-
LIBC_PROBE (pthread_join, 1, threadid);
- if ((pd == self
- || (self->joinid == pd
- && (pd->cancelhandling
- & (CANCELING_BITMASK | CANCELED_BITMASK | EXITING_BITMASK
- | TERMINATED_BITMASK)) == 0))
- && !cancel_enabled_and_canceled (self->cancelhandling))
- /* This is a deadlock situation. The threads are waiting for each
- other to finish. Note that this is a "may" error. To be 100%
- sure we catch this error we would have to lock the data
- structures but it is not necessary. In the unlikely case that
- two threads are really caught in this situation they will
- deadlock. It is the programmer's problem to figure this
- out. */
- return EDEADLK;
-
- /* Wait for the thread to finish. If it is already locked something
- is wrong. There can only be one waiter. */
- else if (__glibc_unlikely (atomic_compare_exchange_weak_acquire (&pd->joinid,
- &self,
- NULL)))
- /* There is already somebody waiting for the thread. */
- return EINVAL;
-
- /* BLOCK waits either indefinitely or based on an absolute time. POSIX also
- states a cancellation point shall occur for pthread_join, and we use the
- same rationale for posix_timedjoin_np. Both clockwait_tid and the futex
- call use the cancellable variant. */
- if (block)
+ int result = 0;
+ unsigned int state;
+ while ((state = atomic_load_acquire (&pd->joinstate))
+ != THREAD_STATE_EXITED)
{
- /* During the wait we change to asynchronous cancellation. If we
- are cancelled the thread we are waiting for must be marked as
- un-wait-ed for again. */
- pthread_cleanup_push (cleanup, &pd->joinid);
-
- /* We need acquire MO here so that we synchronize with the
- kernel's store to 0 when the clone terminates. (see above) */
- pid_t tid;
- while ((tid = atomic_load_acquire (&pd->tid)) != 0)
- {
- /* The kernel notifies a process which uses CLONE_CHILD_CLEARTID via
- futex wake-up when the clone terminates. The memory location
- contains the thread ID while the clone is running and is reset to
- zero by the kernel afterwards. The kernel up to version 3.16.3
- does not use the private futex operations for futex wake-up when
- the clone terminates. */
- int ret = __futex_abstimed_wait_cancelable64 (
- (unsigned int *) &pd->tid, tid, clockid, abstime, LLL_SHARED);
- if (ret == ETIMEDOUT || ret == EOVERFLOW)
- {
- result = ret;
- break;
- }
+ if (check_for_deadlock (pd))
+ return EDEADLK;
+
+ /* POSIX states calling pthread_join on a non joinable thread is
+ undefined. However, if PD is still in the cache we can still warn
+ the caller. */
+ if (state == THREAD_STATE_DETACHED)
+ return EINVAL;
+
+ /* pthread_join is a cancellation entrypoint and we use the same
+ rationale for pthread_timedjoin_np.
+
+ The kernel notifies a process which uses CLONE_CHILD_CLEARTID via
+ a memory zeroing and futex wake-up when the process terminates.
+ The futex operation is not private. */
+ int ret = __futex_abstimed_wait_cancelable64 (&pd->joinstate, state,
+ clockid, abstime,
+ LLL_SHARED);
+ if (ret == ETIMEDOUT || ret == EOVERFLOW)
+ {
+ result = ret;
+ break;
}
-
- pthread_cleanup_pop (0);
}
void *pd_result = pd->result;
- if (__glibc_likely (result == 0))
+ if (result == 0)
{
- /* We mark the thread as terminated and as joined. */
- pd->tid = -1;
-
- /* Store the return value if the caller is interested. */
if (thread_return != NULL)
*thread_return = pd_result;
-
- /* Free the TCB. */
__nptl_free_tcb (pd);
}
- else
- pd->joinid = NULL;
LIBC_PROBE (pthread_join_ret, 3, threadid, result, pd_result);
@@ -24,7 +24,7 @@ ___pthread_timedjoin_np64 (pthread_t threadid, void **thread_return,
const struct __timespec64 *abstime)
{
return __pthread_clockjoin_ex (threadid, thread_return,
- CLOCK_REALTIME, abstime, true);
+ CLOCK_REALTIME, abstime);
}
#if __TIMESIZE == 64
@@ -21,15 +21,17 @@
int
__pthread_tryjoin_np (pthread_t threadid, void **thread_return)
{
- /* Return right away if the thread hasn't terminated yet. */
- struct pthread *pd = (struct pthread *) threadid;
- if (pd->tid != 0)
- return EBUSY;
+ /* The joinable state (THREAD_STATE_JOINABLE) is straigthforward since the
+ thread hasn't finished yet and trying to join might block.
+ The exiting thread (THREAD_STATE_EXITING) also migth result in a blocking
+ call: a detached thread might change its state to exiting and a exiting
+ thread my take some time to exit (and thus let the kernel set the state
+ to THREAD_STATE_EXITED). */
- /* If pd->tid == 0 then lll_wait_tid will not block on futex
- operation. */
- return __pthread_clockjoin_ex (threadid, thread_return, 0 /* Ignored */,
- NULL, false);
+ struct pthread *pd = (struct pthread *) threadid;
+ return atomic_load_acquire (&pd->joinstate) != THREAD_STATE_EXITED
+ ? EBUSY
+ : __pthread_clockjoin_ex (threadid, thread_return, 0, NULL);
}
versioned_symbol (libc, __pthread_tryjoin_np, pthread_tryjoin_np, GLIBC_2_34);
@@ -72,7 +72,7 @@ __tls_init_tp (void)
list_add (&pd->list, &GL (dl_stack_user));
/* Early initialization of the TCB. */
- pd->tid = INTERNAL_SYSCALL_CALL (set_tid_address, &pd->tid);
+ pd->tid = INTERNAL_SYSCALL_CALL (set_tid_address, &pd->joinstate);
THREAD_SETMEM (pd, specific[0], &pd->specific_1stblock[0]);
THREAD_SETMEM (pd, user_stack, true);
@@ -128,4 +128,6 @@ __tls_init_tp (void)
It will be bigger than it actually is, but for unwind.c/pt-longjmp.c
purposes this is good enough. */
THREAD_SETMEM (pd, stackblock_size, (size_t) __libc_stack_end);
+
+ THREAD_SETMEM (pd, joinstate, THREAD_STATE_JOINABLE);
}
@@ -18,6 +18,7 @@
#include <atomic.h>
#include <pthreadP.h>
+#include <futex-internal.h>
_Noreturn static void
__libc_start_call_main (int (*main) (int, char **, char ** MAIN_AUXVEC_DECL),
@@ -65,6 +66,12 @@ __libc_start_call_main (int (*main) (int, char **, char ** MAIN_AUXVEC_DECL),
/* One less thread. Decrement the counter. If it is zero we
terminate the entire process. */
result = 0;
+
+ /* For the case a thread is waiting for the main thread to finish. */
+ struct pthread *self = THREAD_SELF;
+ atomic_store_release (&self->joinstate, THREAD_STATE_EXITED);
+ futex_wake (&self->joinstate, 1, FUTEX_SHARED);
+
if (! atomic_decrement_and_test (&__nptl_nthreads))
/* Not much left to do but to exit the thread, not the process. */
while (1)
@@ -243,7 +243,6 @@ libc_hidden_proto (__pthread_current_priority)
nothing. And if the test triggers the thread descriptor is
guaranteed to be invalid. */
#define INVALID_TD_P(pd) __builtin_expect ((pd)->tid <= 0, 0)
-#define INVALID_NOT_TERMINATED_TD_P(pd) __builtin_expect ((pd)->tid < 0, 0)
extern void __pthread_unwind (__pthread_unwind_buf_t *__buf)
__cleanup_fct_attribute __attribute ((__noreturn__))
@@ -521,7 +520,7 @@ libc_hidden_proto (__pthread_setcanceltype)
extern void __pthread_testcancel (void);
libc_hidden_proto (__pthread_testcancel)
extern int __pthread_clockjoin_ex (pthread_t, void **, clockid_t,
- const struct __timespec64 *, bool)
+ const struct __timespec64 *)
attribute_hidden;
extern int __pthread_sigmask (int, const sigset_t *, sigset_t *);
libc_hidden_proto (__pthread_sigmask);
@@ -20,14 +20,14 @@
#include <time.h>
#include <stdio.h>
#include <unistd.h>
-
+#include <limits.h>
#include <support/check.h>
static int
detach_thrd (void *arg)
{
- if (thrd_detach (thrd_current ()) != thrd_success)
- FAIL_EXIT1 ("thrd_detach failed");
+ thrd_sleep (&(struct timespec) { .tv_sec = INT_MAX }, NULL);
+
thrd_exit (thrd_success);
}
@@ -36,15 +36,11 @@ do_test (void)
{
thrd_t id;
- /* Create new thread. */
- if (thrd_create (&id, detach_thrd, NULL) != thrd_success)
- FAIL_EXIT1 ("thrd_create failed");
+ TEST_COMPARE (thrd_create (&id, detach_thrd, NULL), thrd_success);
- /* Give some time so the thread can finish. */
- thrd_sleep (&(struct timespec) {.tv_sec = 2}, NULL);
+ TEST_COMPARE (thrd_detach (id), thrd_success);
- if (thrd_join (id, NULL) == thrd_success)
- FAIL_EXIT1 ("thrd_join succeed where it should fail");
+ TEST_COMPARE (thrd_join (id, NULL), thrd_error);
return 0;
}