From mboxrd@z Thu Jan 1 00:00:00 1970 From: Alex Elder Subject: [PATCH 5/6, v2] rbd: fix leak of format 2 snapshot names Date: Fri, 26 Apr 2013 12:40:22 -0500 Message-ID: <517ABC06.7080200@inktank.com> References: <517A6D39.80000@inktank.com> <517A6DD1.2080102@inktank.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Return-path: Received: from mail-ie0-f172.google.com ([209.85.223.172]:34175 "EHLO mail-ie0-f172.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754814Ab3DZRkY (ORCPT ); Fri, 26 Apr 2013 13:40:24 -0400 Received: by mail-ie0-f172.google.com with SMTP id c12so5285054ieb.31 for ; Fri, 26 Apr 2013 10:40:24 -0700 (PDT) Received: from [172.22.22.4] (c-71-195-31-37.hsd1.mn.comcast.net. [71.195.31.37]) by mx.google.com with ESMTPSA id hi4sm4274171igc.6.2013.04.26.10.40.23 for (version=TLSv1 cipher=ECDHE-RSA-RC4-SHA bits=128/128); Fri, 26 Apr 2013 10:40:23 -0700 (PDT) In-Reply-To: <517A6DD1.2080102@inktank.com> Sender: ceph-devel-owner@vger.kernel.org List-ID: To: ceph-devel@vger.kernel.org (I found a bug in this (which masked another bug). Here is version 2. I have corrected it and updated the various branches that depended on it.) When the snapshot context for an rbd device gets updated (or the initial one is recorded) a a list of snapshot structures is created to represent them, one entry per snapshot. Each entry includes a dynamically-allocated copy of the snapshot name. Currently the name is allocated in rbd_snap_create(), as a duplicate of the passed-in name. For format 1 images, the snapshot name provided is just a pointer to an existing name. But for format 2 images, the passed-in name is already dynamically allocated, and in the the process of duplicating it here we are leaking the passed-in name. Fix this by dynamically allocating the name for format 1 snapshots also, and then stop allocating a duplicate in rbd_snap_create(). Change rbd_dev_v1_snap_info() so none of its parameters is side-effected unless it's going to return success. This is part of: http://tracker.ceph.com/issues/4803 Signed-off-by: Alex Elder --- v2: kill an errant semicolon; use loop iterator rather than "which" drivers/block/rbd.c | 30 ++++++++++++++---------------- 1 file changed, 14 insertions(+), 16 deletions(-) diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c index 916741b..c15bb3f 100644 --- a/drivers/block/rbd.c +++ b/drivers/block/rbd.c @@ -3427,46 +3427,44 @@ static struct rbd_snap *rbd_snap_create(struct rbd_device *rbd_dev, u64 snap_features) { struct rbd_snap *snap; - int ret; snap = kzalloc(sizeof (*snap), GFP_KERNEL); if (!snap) return ERR_PTR(-ENOMEM); - ret = -ENOMEM; - snap->name = kstrdup(snap_name, GFP_KERNEL); - if (!snap->name) - goto err; - + snap->name = snap_name; snap->id = snap_id; snap->size = snap_size; snap->features = snap_features; return snap; - -err: - kfree(snap->name); - kfree(snap); - - return ERR_PTR(ret); } +/* + * Returns a dynamically-allocated snapshot name if successful, or a + * pointer-coded error otherwise. + */ static char *rbd_dev_v1_snap_info(struct rbd_device *rbd_dev, u32 which, u64 *snap_size, u64 *snap_features) { char *snap_name; + int i; rbd_assert(which < rbd_dev->header.snapc->num_snaps); - *snap_size = rbd_dev->header.snap_sizes[which]; - *snap_features = 0; /* No features for v1 */ - /* Skip over names until we find the one we are looking for */ snap_name = rbd_dev->header.snap_names; - while (which--) + for (i = 0; i < which; i++) snap_name += strlen(snap_name) + 1; + snap_name = kstrdup(snap_name, GFP_KERNEL); + if (!snap_name) + return ERR_PTR(-ENOMEM); + + *snap_size = rbd_dev->header.snap_sizes[which]; + *snap_features = 0; /* No features for v1 */ + return snap_name; } -- 1.7.9.5