From: Sage Weil Date: Tue, 31 Jul 2012 01:17:13 +0000 (-0700) Subject: libceph: avoid dropping con mutex before fault X-Git-Tag: v3.4.20~15 X-Git-Url: https://git.karo-electronics.de/?a=commitdiff_plain;h=328677c24bbae17f3f91ffce3b2351a27933741c;p=karo-tx-linux.git libceph: avoid dropping con mutex before fault (cherry picked from commit 8636ea672f0c5ab7478c42c5b6705ebd1db7eb6a) The ceph_fault() function takes the con mutex, so we should avoid dropping it before calling it. This fixes a potential race with another thread calling ceph_con_close(), or _open(), or similar (we don't reverify con->state after retaking the lock). Add annotation so that lockdep realizes we will drop the mutex before returning. Signed-off-by: Sage Weil Reviewed-by: Alex Elder Signed-off-by: Greg Kroah-Hartman --- diff --git a/net/ceph/messenger.c b/net/ceph/messenger.c index bb21e5b0fb55..c2f0ca36f4ab 100644 --- a/net/ceph/messenger.c +++ b/net/ceph/messenger.c @@ -2336,7 +2336,6 @@ done_unlocked: return; fault: - mutex_unlock(&con->mutex); ceph_fault(con); /* error/fault path */ goto done_unlocked; } @@ -2347,9 +2346,8 @@ fault: * exponential backoff */ static void ceph_fault(struct ceph_connection *con) + __releases(con->mutex) { - mutex_lock(&con->mutex); - pr_err("%s%lld %s %s\n", ENTITY_NAME(con->peer_name), ceph_pr_addr(&con->peer_addr.in_addr), con->error_msg); dout("fault %p state %lu to peer %s\n",