From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail.linuxfoundation.org ([140.211.169.12]:58574 "EHLO mail.linuxfoundation.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1727273AbeHVCYd (ORCPT ); Tue, 21 Aug 2018 22:24:33 -0400 Date: Tue, 21 Aug 2018 16:02:24 -0700 From: Andrew Morton To: Viacheslav Dubeyko Cc: "Ernesto A." =?ISO-8859-1?Q?Fern=E1ndez?= , linux-fsdevel@vger.kernel.org, Anatoly Trosinenko Subject: Re: [PATCH 1/2] hfsplus: prevent crash on exit from failed search Message-Id: <20180821160224.de4301a45eab98b551823a6c@linux-foundation.org> In-Reply-To: <1530554497.16350.2.camel@slavad-ubuntu-14.04> References: <803590a35221fbf411b2c141419aea3233a6e990.1530294813.git.ernesto.mnd.fernandez@gmail.com> <1530554497.16350.2.camel@slavad-ubuntu-14.04> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8BIT Sender: linux-fsdevel-owner@vger.kernel.org List-ID: On Mon, 02 Jul 2018 11:01:37 -0700 Viacheslav Dubeyko wrote: > On Fri, 2018-06-29 at 15:34 -0300, Ernesto A. Fern�ndez wrote: > > The hfs_find_exit() function expects fd->bnode to be NULL after a > > search has failed. The hfs_brec_insert() function may instead set > > it to an error-valued pointer. Fix this to prevent a crash. > > > > Reported-by: Anatoly Trosinenko > > Signed-off-by: Ernesto A. Fern�ndez > > --- > > fs/hfsplus/brec.c | 7 ++++--- > > 1 file changed, 4 insertions(+), 3 deletions(-) > > > > diff --git a/fs/hfsplus/brec.c b/fs/hfsplus/brec.c > > index 808f4d8c859c..ed8eacb34452 100644 > > --- a/fs/hfsplus/brec.c > > +++ b/fs/hfsplus/brec.c > > @@ -73,9 +73,10 @@ int hfs_brec_insert(struct hfs_find_data *fd, void *entry, int entry_len) > > if (!fd->bnode) { > > if (!tree->root) > > hfs_btree_inc_height(tree); > > - fd->bnode = hfs_bnode_find(tree, tree->leaf_head); > > - if (IS_ERR(fd->bnode)) > > - return PTR_ERR(fd->bnode); > > > Are you sure that no caller is used this error code? Did you check this? > > Maybe, it makes sense to extract the error code and to show the error > message on the caller side instead of processing the simple NULL? > No response? Could we please get this wrapped up?