svn commit: r225057 - head/sys/kern

[Attilio Rao]

Author: attilio
Date: Sun Aug 21 10:52:50 2011
New Revision: 225057
URL: http://svn.freebsd.org/changeset/base/225057

Log:
  callout_cpu_switch() allows preemption when dropping the outcoming
  callout cpu lock (and after having dropped it).
  If the newly scheduled thread wants to acquire the old queue it will
  just spin forever.
  
  Fix this by disabling preemption and interrupts entirely (because fast
  interrupt handlers may incur in the same problem too) while switching
  locks.
  
  Reported by:	hrs, Mike Tancsa <mike AT sentex DOT net>,
  		Chip Camden <sterling AT camdensoftware DOT com>
  Tested by:	hrs, Mike Tancsa <mike AT sentex DOT net>,
  		Chip Camden <sterling AT camdensoftware DOT com>,
  		Nicholas Esborn <nick AT desert DOT net>
  Approved by:	re (kib)
  MFC after:	10 days

Modified:
  head/sys/kern/kern_timeout.c

Modified: head/sys/kern/kern_timeout.c
==============================================================================
--- head/sys/kern/kern_timeout.c	Sun Aug 21 10:05:39 2011	(r225056)
+++ head/sys/kern/kern_timeout.c	Sun Aug 21 10:52:50 2011	(r225057)
@@ -269,10 +269,17 @@ callout_cpu_switch(struct callout *c, st
 	MPASS(c != NULL && cc != NULL);
 	CC_LOCK_ASSERT(cc);
 
+	/*
+	 * Avoid interrupts and preemption firing after the callout cpu
+	 * is blocked in order to avoid deadlocks as the new thread
+	 * may be willing to acquire the callout cpu lock.
+	 */
 	c->c_cpu = CPUBLOCK;
+	spinlock_enter();
 	CC_UNLOCK(cc);
 	new_cc = CC_CPU(new_cpu);
 	CC_LOCK(new_cc);
+	spinlock_exit();
 	c->c_cpu = new_cpu;
 	return (new_cc);
 }