use atomic decrement rather than cas in pthread_exit thread count

now that blocking signals prevents any application code from running
while the last thread is exiting, the cas logic is no longer needed to
prevent decrementing below zero.

diff --git a/src/thread/pthread_create.c b/src/thread/pthread_create.c
index 6a37ee9be5c760a89923a394276406a027a45e22..0567f966e02d26fbfc366631af9919e20ce7378c 100644 (file)
--- a/src/thread/pthread_create.c
+++ b/src/thread/pthread_create.c
@@ -12,7 +12,6 @@ weak_alias(dummy_0, __pthread_tsd_run_dtors);
 _Noreturn void pthread_exit(void *result)
 {
        pthread_t self = pthread_self();
-       int n;
 
        self->result = result;
 
@@ -38,9 +37,7 @@ _Noreturn void pthread_exit(void *result)
         * reasons as well. */
        __syscall(SYS_rt_sigprocmask, SIG_BLOCK, SIGALL_SET, 0, _NSIG/8);
 
-       do n = libc.threads_minus_1;
-       while (n && a_cas(&libc.threads_minus_1, n, n-1)!=n);
-       if (!n) exit(0);
+       if (a_fetch_add(&libc.threads_minus_1, -1)==0) exit(0);
 
        if (self->detached && self->map_base) {
                /* Detached threads must avoid the kernel clear_child_tid