mutex: preemption fixes
authorPeter Zijlstra <a.p.zijlstra@chello.nl>
Wed, 14 Jan 2009 14:36:26 +0000 (15:36 +0100)
committerIngo Molnar <mingo@elte.hu>
Wed, 14 Jan 2009 17:09:00 +0000 (18:09 +0100)
The problem is that dropping the spinlock right before schedule is a voluntary
preemption point and can cause a schedule, right after which we schedule again.

Fix this inefficiency by keeping preemption disabled until we schedule, do this
by explicity disabling preemption and providing a schedule() variant that
assumes preemption is already disabled.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

No differences found