From mboxrd@z Thu Jan 1 00:00:00 1970 From: Dan Carpenter Subject: [patch 1/2] libceph: fix error handling in handle_reply() Date: Thu, 15 Aug 2013 08:51:58 +0300 Message-ID: <20130815055158.GA23580@elgon.mountain> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: "David S. Miller" , ceph-devel@vger.kernel.org, netdev@vger.kernel.org, kernel-janitors@vger.kernel.org To: Sage Weil Return-path: Content-Disposition: inline Sender: ceph-devel-owner@vger.kernel.org List-Id: netdev.vger.kernel.org We've tried to fix the error paths in this function before, but there is still a hidden goto in the ceph_decode_need() macro which goes to the wrong place. We need to release the "req" and unlock a mutex before returning. Signed-off-by: Dan Carpenter diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c index dbc0a73..559a832 100644 --- a/net/ceph/osd_client.c +++ b/net/ceph/osd_client.c @@ -1488,14 +1488,14 @@ static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg, dout("handle_reply %p tid %llu req %p result %d\n", msg, tid, req, result); - ceph_decode_need(&p, end, 4, bad); + ceph_decode_need(&p, end, 4, bad_put); numops = ceph_decode_32(&p); if (numops > CEPH_OSD_MAX_OP) goto bad_put; if (numops != req->r_num_ops) goto bad_put; payload_len = 0; - ceph_decode_need(&p, end, numops * sizeof(struct ceph_osd_op), bad); + ceph_decode_need(&p, end, numops * sizeof(struct ceph_osd_op), bad_put); for (i = 0; i < numops; i++) { struct ceph_osd_op *op = p; int len; @@ -1513,7 +1513,7 @@ static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg, goto bad_put; } - ceph_decode_need(&p, end, 4 + numops * 4, bad); + ceph_decode_need(&p, end, 4 + numops * 4, bad_put); retry_attempt = ceph_decode_32(&p); for (i = 0; i < numops; i++) req->r_reply_op_result[i] = ceph_decode_32(&p);