If a thread that is swapped out is made runnable, then the setrunnable()

routine wakes up proc0 so that proc0 can swap the thread back in.
Historically, this has been done by waking up proc0 directly from
setrunnable() itself via a wakeup().  When waking up a sleeping thread
that was swapped out (the usual case when waking proc0 since only sleeping
threads are eligible to be swapped out), this resulted in a bit of
recursion (e.g. wakeup() -> setrunnable() -> wakeup()).

With sleep queues having separate locks in 6.x and later, this caused a
spin lock LOR (sleepq lock -> sched_lock/thread lock -> sleepq lock).
An attempt was made to fix this in 7.0 by making the proc0 wakeup use
the ithread mechanism for doing the wakeup.  However, this required
grabbing proc0's thread lock to perform the wakeup.  If proc0 was asleep
elsewhere in the kernel (e.g. waiting for disk I/O), then this degenerated
into the same LOR since the thread lock would be some other sleepq lock.

Fix this by deferring the wakeup of the swapper until after the sleepq
lock held by the upper layer has been locked.  The setrunnable() routine
now returns a boolean value to indicate whether or not proc0 needs to be
woken up.  The end result is that consumers of the sleepq API such as
*sleep/wakeup, condition variables, sx locks, and lockmgr, have to wakeup
proc0 if they get a non-zero return value from sleepq_abort(),
sleepq_broadcast(), or sleepq_signal().

Discussed with:	jeff
Glanced at by:	sam
Tested by:	Jurgen Weber  jurgen - ish com au
MFC after:	2 weeks

1 files changed, 17 insertions, 6 deletions

diff --git a/sys/kern/kern_sx.c b/sys/kern/kern_sx.c
index 1f0011c..5006793 100644
--- a/sys/kern/kern_sx.c
+++ b/sys/kern/kern_sx.c
@@ -361,6 +361,7 @@ void
 _sx_downgrade(struct sx *sx, const char *file, int line)
 {
 	uintptr_t x;
+	int wakeup_swapper;
 
 	KASSERT(sx->sx_lock != SX_LOCK_DESTROYED,
 	    ("sx_downgrade() of destroyed sx @ %s:%d", file, line));
@@ -401,15 +402,19 @@ _sx_downgrade(struct sx *sx, const char *file, int line)
 	 * Preserve SX_LOCK_EXCLUSIVE_WAITERS while downgraded to a single
 	 * shared lock.  If there are any shared waiters, wake them up.
 	 */
+	wakeup_swapper = 0;
 	x = sx->sx_lock;
 	atomic_store_rel_ptr(&sx->sx_lock, SX_SHARERS_LOCK(1) |
 	    (x & SX_LOCK_EXCLUSIVE_WAITERS));
 	if (x & SX_LOCK_SHARED_WAITERS)
-		sleepq_broadcast(&sx->lock_object, SLEEPQ_SX, 0,
-		    SQ_SHARED_QUEUE);
+		wakeup_swapper = sleepq_broadcast(&sx->lock_object, SLEEPQ_SX,
+		    0, SQ_SHARED_QUEUE);
 	sleepq_release(&sx->lock_object);
 
 	LOCK_LOG_LOCK("XDOWNGRADE", &sx->lock_object, 0, 0, file, line);
+
+	if (wakeup_swapper)
+		kick_proc0();
 }
 
 /*
@@ -589,7 +594,7 @@ void
 _sx_xunlock_hard(struct sx *sx, uintptr_t tid, const char *file, int line)
 {
 	uintptr_t x;
-	int queue;
+	int queue, wakeup_swapper;
 
 	MPASS(!(sx->sx_lock & SX_LOCK_SHARED));
 
@@ -627,8 +632,11 @@ _sx_xunlock_hard(struct sx *sx, uintptr_t tid, const char *file, int line)
 		    __func__, sx, queue == SQ_SHARED_QUEUE ? "shared" :
 		    "exclusive");
 	atomic_store_rel_ptr(&sx->sx_lock, x);
-	sleepq_broadcast(&sx->lock_object, SLEEPQ_SX, 0, queue);
+	wakeup_swapper = sleepq_broadcast(&sx->lock_object, SLEEPQ_SX, 0,
+	    queue);
 	sleepq_release(&sx->lock_object);
+	if (wakeup_swapper)
+		kick_proc0();
 }
 
 /*
@@ -795,6 +803,7 @@ void
 _sx_sunlock_hard(struct sx *sx, const char *file, int line)
 {
 	uintptr_t x;
+	int wakeup_swapper;
 
 	for (;;) {
 		x = sx->sx_lock;
@@ -862,9 +871,11 @@ _sx_sunlock_hard(struct sx *sx, const char *file, int line)
 		if (LOCK_LOG_TEST(&sx->lock_object, 0))
 			CTR2(KTR_LOCK, "%s: %p waking up all thread on"
 			    "exclusive queue", __func__, sx);
-		sleepq_broadcast(&sx->lock_object, SLEEPQ_SX, 0,
-		    SQ_EXCLUSIVE_QUEUE);
+		wakeup_swapper = sleepq_broadcast(&sx->lock_object, SLEEPQ_SX,
+		    0, SQ_EXCLUSIVE_QUEUE);
 		sleepq_release(&sx->lock_object);
+		if (wakeup_swapper)
+			kick_proc0();
 		break;
 	}
 }