]> err.no Git - linux-2.6/commit
[PATCH] cond_resched() fix
authorAndrew Morton <akpm@osdl.org>
Fri, 30 Jun 2006 08:56:00 +0000 (01:56 -0700)
committerLinus Torvalds <torvalds@g5.osdl.org>
Fri, 30 Jun 2006 18:25:38 +0000 (11:25 -0700)
commite7b384043e27bed4f23b108481b99c518dd01a01
tree52f944bf39d3a7b329f4e38d619d7949e35510a0
parent92fe15a3d24fa53e7e961c549c488d0bb642d895
[PATCH] cond_resched() fix

Fix a bug identified by Zou Nan hai <nanhai.zou@intel.com>:

If the system is in state SYSTEM_BOOTING, and need_resched() is true,
cond_resched() returns true even though it didn't reschedule.  Consequently
need_resched() remains true and JBD locks up.

Fix that by teaching cond_resched() to only return true if it really did call
schedule().

cond_resched_lock() and cond_resched_softirq() have a problem too.  If we're
in SYSTEM_BOOTING state and need_resched() is true, these functions will drop
the lock and will then try to call schedule(), but the SYSTEM_BOOTING state
will prevent schedule() from being called.  So on return, need_resched() will
still be true, but cond_resched_lock() has to return 1 to tell the caller that
the lock was dropped.  The caller will probably lock up.

Bottom line: if these functions dropped the lock, they _must_ call schedule()
to clear need_resched().   Make it so.

Also, uninline __cond_resched().  It's largeish, and slowpath.

Acked-by: Ingo Molnar <mingo@elte.hu>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
kernel/sched.c