From: Li Zefan <lizf@cn.fujitsu.com>
Date: Thu, 2 Apr 2009 23:57:49 +0000 (-0700)
Subject: cpuset: fix possible races in cpu/memory hotplug
X-Git-Tag: v2.6.30-rc1~349
X-Git-Url: https://asedeno.scripts.mit.edu/gitweb/?a=commitdiff_plain;h=0b4217b3fdddc4a58939720d3ed809537577d48b;p=linux.git

cpuset: fix possible races in cpu/memory hotplug

Change to cpuset->cpus_allowed and cpuset->mems_allowed should be protected
by callback_mutex, otherwise the reader may read wrong cpus/mems. This is
cpuset's lock rule.

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Cc: Paul Menage <menage@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
---

diff --git a/kernel/cpuset.c b/kernel/cpuset.c
index ee5ec386aa8b..31737957cb62 100644
--- a/kernel/cpuset.c
+++ b/kernel/cpuset.c
@@ -2070,7 +2070,9 @@ static int cpuset_track_online_cpus(struct notifier_block *unused_nb,
 	}
 
 	cgroup_lock();
+	mutex_lock(&callback_mutex);
 	cpumask_copy(top_cpuset.cpus_allowed, cpu_online_mask);
+	mutex_unlock(&callback_mutex);
 	scan_for_empty_cpusets(&top_cpuset);
 	ndoms = generate_sched_domains(&doms, &attr);
 	cgroup_unlock();
@@ -2093,11 +2095,12 @@ static int cpuset_track_online_nodes(struct notifier_block *self,
 	cgroup_lock();
 	switch (action) {
 	case MEM_ONLINE:
-		top_cpuset.mems_allowed = node_states[N_HIGH_MEMORY];
-		break;
 	case MEM_OFFLINE:
+		mutex_lock(&callback_mutex);
 		top_cpuset.mems_allowed = node_states[N_HIGH_MEMORY];
-		scan_for_empty_cpusets(&top_cpuset);
+		mutex_unlock(&callback_mutex);
+		if (action == MEM_OFFLINE)
+			scan_for_empty_cpusets(&top_cpuset);
 		break;
 	default:
 		break;