From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from userp1040.oracle.com ([156.151.31.81]:49078 "EHLO userp1040.oracle.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754391AbcGLRGB (ORCPT ); Tue, 12 Jul 2016 13:06:01 -0400 Date: Tue, 12 Jul 2016 10:08:45 -0700 From: Liu Bo To: Chris Mason Cc: linux-btrfs@vger.kernel.org, David Sterba Subject: Re: [PATCH] Btrfs: fix panic in balance due to EIO Message-ID: <20160712170845.GA2870@localhost.localdomain> Reply-To: bo.li.liu@oracle.com References: <1468283820-25471-1-git-send-email-bo.li.liu@oracle.com> <855a29a1-1b46-c1be-418c-de08cb484bd7@fb.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: <855a29a1-1b46-c1be-418c-de08cb484bd7@fb.com> Sender: linux-btrfs-owner@vger.kernel.org List-ID: On Tue, Jul 12, 2016 at 11:05:48AM -0400, Chris Mason wrote: > > > On 07/11/2016 08:37 PM, Liu Bo wrote: > > During build_backref_tree(), if we fail to read a btree node, > > we can eventually run into BUG_ON(cache->nr_nodes) that we put > > in backref_cache_cleanup(), meaning we have at least one > > memory leak. > > > > This frees the backref_node that we allocate at the very beginning of build_backref_tree(). > > > > Signed-off-by: Liu Bo > > --- > > fs/btrfs/relocation.c | 7 +++++++ > > 1 file changed, 7 insertions(+) > > > > diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c > > index 0477dca..f00267a 100644 > > --- a/fs/btrfs/relocation.c > > +++ b/fs/btrfs/relocation.c > > @@ -1135,6 +1135,8 @@ out: > > btrfs_free_path(path1); > > btrfs_free_path(path2); > > if (err) { > > + int orig_free = 0; > > + > > while (!list_empty(&useless)) { > > lower = list_entry(useless.next, > > struct backref_node, list); > > @@ -1171,8 +1173,13 @@ out: > > lower = list_entry(useless.next, > > struct backref_node, list); > > list_del_init(&lower->list); > > + if (lower == node) > > + orig_free = 1; > > free_backref_node(cache, lower); > > } > > + > > + if (!orig_free) > > + free_backref_node(cache, node); > > return ERR_PTR(err); > > } > > ASSERT(!node || !node->detached); > > Instead of doing the orig_free set and test > > ... > if (lower == node) > node = NULL > free_backref_node(cache, lower) > } > free_backref_node(cache, node); > return ERR_PTR(err); > > Your patch isn't wrong, but having node NULL after it was free'd makes us > less likely to make mistakes as the code changes. Golden rule indeed. Thanks, -liubo