sem: Restart the POSIX sem_* calls after signals with SA_RESTART set.

Programs often do not expect an [EINTR] return from sem_wait() and POSIX
only allows it if the signal was installed without SA_RESTART. The timeout
in sem_timedwait() is absolute so it can be restarted normally.

The umtx call can be invoked with a relative timeout and in that case
[ERESTART] must be changed to [EINTR]. However, libc does not do this.

The old POSIX semaphore implementation did this correctly (before r249566),
unlike the new umtx one.

It may be desirable to avoid [EINTR] completely, which matches the pthread
functions and is explicitly permitted by POSIX. However, the kernel must
return [EINTR] at least for signals with SA_RESTART clear, otherwise pthread
cancellation will not abort a semaphore wait. In this commit, only restore
the 8.x behaviour which is also permitted by POSIX.

Discussed with:	jhb
MFC after:	1 week

2 files changed, 3 insertions, 3 deletions

diff --git a/sys/kern/kern_umtx.c b/sys/kern/kern_umtx.c
index 6026b80..74d6d1e 100644
--- a/sys/kern/kern_umtx.c
+++ b/sys/kern/kern_umtx.c
@@ -2980,7 +2980,9 @@ do_sem_wait(struct thread *td, struct _usem *sem, struct _umtx_time *timeout)
 		error = 0;
 	else {
 		umtxq_remove(uq);
-		if (error == ERESTART)
+		/* A relative timeout cannot be restarted. */
+		if (error == ERESTART && timeout != NULL &&
+		    (timeout->_flags & UMTX_ABSTIME) == 0)
 			error = EINTR;
 	}
 	umtxq_unlock(&uq->uq_key);
diff --git a/sys/kern/uipc_sem.c b/sys/kern/uipc_sem.c
index 0ea84fd..509f32e 100644
--- a/sys/kern/uipc_sem.c
+++ b/sys/kern/uipc_sem.c
@@ -846,8 +846,6 @@ kern_sem_wait(struct thread *td, semid_t id, int tryflag,
 err:
 	mtx_unlock(&sem_lock);
 	fdrop(fp, td);
-	if (error == ERESTART)
-		error = EINTR;
 	DP(("<<< kern_sem_wait leaving, pid=%d, error = %d\n",
 	    (int)td->td_proc->p_pid, error));
 	return (error);