stop_machine: Fix race between stop_two_cpus() and stop_cpus()
authorRik van Riel <[email protected]>
Fri, 1 Nov 2013 14:41:46 +0000 (10:41 -0400)
committerIngo Molnar <[email protected]>
Mon, 11 Nov 2013 11:43:38 +0000 (12:43 +0100)
commit7053ea1a34fa8567cb5e3c39e04ace4c5d0fbeaa
treeed5866f5a2449e6d7b918ce92858f891c8bc1450
parent37dc6b50cee97954c4e6edcd5b1fa614b76038ee
stop_machine: Fix race between stop_two_cpus() and stop_cpus()

There is a race between stop_two_cpus, and the global stop_cpus.

It is possible for two CPUs to get their stopper functions queued
"backwards" from one another, resulting in the stopper threads
getting stuck, and the system hanging. This can happen because
queuing up stoppers is not synchronized.

This patch adds synchronization between stop_cpus (a rare operation),
and stop_two_cpus.

Reported-and-Tested-by: Prarit Bhargava <[email protected]>
Signed-off-by: Rik van Riel <[email protected]>
Signed-off-by: Peter Zijlstra <[email protected]>
Acked-by: Mel Gorman <[email protected]>
Link: http://lkml.kernel.org/r/[email protected]
Signed-off-by: Ingo Molnar <[email protected]>
kernel/stop_machine.c