kernel/locking/qspinlock.c

46  * unlock the next pending (next->locked), we compress both these: {tail,
89  * The pending bit spinning loop count.
143  * clear_pending - clear the pending bit.
150 	WRITE_ONCE(lock->pending, 0);  in clear_pending()
154  * clear_pending_set_locked - take ownership and clear the pending bit.
189  * clear_pending - clear the pending bit.
200  * clear_pending_set_locked - take ownership and clear the pending bit.
242  * queued_fetch_set_pending_acquire - fetch the whole lock value and set pending
299  * (queue tail, pending bit, lock value)
306  * pending               :    (0,1,1) +--> (0,1,0)   \           |  :
330 	 * Wait for in-progress pending->locked hand-overs with a bounded  in queued_spin_lock_slowpath()
348 	 * trylock || pending  in queued_spin_lock_slowpath()
350 	 * 0,0,* -> 0,1,* -> 0,0,1 pending, trylock  in queued_spin_lock_slowpath()
357 	 * Undo and queue; our setting of PENDING might have made the  in queued_spin_lock_slowpath()
363 		/* Undo PENDING if we set it. */  in queued_spin_lock_slowpath()
371 	 * We're pending, wait for the owner to go away.  in queued_spin_lock_slowpath()
385 	 * take ownership and clear the pending bit.  in queued_spin_lock_slowpath()
394 	 * End of pending bit optimistic spinning and beginning of MCS  in queued_spin_lock_slowpath()
456 	 * pending stuff.  in queued_spin_lock_slowpath()
488 	 * we're at the head of the waitqueue, wait for the owner & pending to  in queued_spin_lock_slowpath()
521 	 * and nobody is pending, clear the tail code and grab the lock.  in queued_spin_lock_slowpath()
533 	 *       PENDING will make the uncontended transition fail.  in queued_spin_lock_slowpath()