* The spin_unlock() itself is semi-permeable and only protects
* one way (it only protects stuff inside the critical region and
* stops them from bleeding out - it would still allow subsequent
- * loads to move into the the critical region).
+ * loads to move into the critical region).
*/
void fastcall
prepare_to_wait(wait_queue_head_t *q, wait_queue_t *wait, int state)