From: Li Zefan <lizefan@huawei.com>
Date: Thu, 27 Feb 2014 10:19:36 +0000 (+0800)
Subject: cpuset: fix a race condition in __cpuset_node_allowed_softwall()
X-Git-Tag: v3.13.7~79
X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=9a59929c894645615835deb478f99de178530b22;p=thirdparty%2Fkernel%2Fstable.git

cpuset: fix a race condition in __cpuset_node_allowed_softwall()

commit 99afb0fd5f05aac467ffa85c36778fec4396209b upstream.

It's not safe to access task's cpuset after releasing task_lock().
Holding callback_mutex won't help.

Signed-off-by: Li Zefan <lizefan@huawei.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---

diff --git a/kernel/cpuset.c b/kernel/cpuset.c
index bd1dfa2c7d890..5ae9f950e024a 100644
--- a/kernel/cpuset.c
+++ b/kernel/cpuset.c
@@ -2507,9 +2507,9 @@ int __cpuset_node_allowed_softwall(int node, gfp_t gfp_mask)
 
 	task_lock(current);
 	cs = nearest_hardwall_ancestor(task_cs(current));
+	allowed = node_isset(node, cs->mems_allowed);
 	task_unlock(current);
 
-	allowed = node_isset(node, cs->mems_allowed);
 	mutex_unlock(&callback_mutex);
 	return allowed;
 }