[PATCH 4.7 02/59] apparmor: fix refcount race when finding a child profile

[Greg Kroah-Hartman]

4.7-stable review patch.  If anyone has any objections, please let me know.

------------------

From: John Johansen <john.johansen@xxxxxxxxxxxxx>

commit de7c4cc947f9f56f61520ee7edaf380434a98c8d upstream.

When finding a child profile via an rcu critical section, the profile
may be put and scheduled for deletion after the child is found but
before its refcount is incremented.

Protect against this by repeating the lookup if the profiles refcount
is 0 and is one its way to deletion.

Signed-off-by: John Johansen <john.johansen@xxxxxxxxxxxxx>
Acked-by: Seth Arnold <seth.arnold@xxxxxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>

---
 security/apparmor/policy.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

--- a/security/apparmor/policy.c
+++ b/security/apparmor/policy.c
@@ -766,7 +766,9 @@ struct aa_profile *aa_find_child(struct
 	struct aa_profile *profile;
 
 	rcu_read_lock();
-	profile = aa_get_profile(__find_child(&parent->base.profiles, name));
+	do {
+		profile = __find_child(&parent->base.profiles, name);
+	} while (profile && !aa_get_profile_not0(profile));
 	rcu_read_unlock();
 
 	/* refcount released by caller */