From: Eric Sandeen Subject: Re: [PATCH] : make sure the buffer head members are zeroed out before using them. Date: Sun, 25 Jan 2009 11:01:55 -0600 Message-ID: <497C9B03.4080008@redhat.com> References: <497C903E.5080108@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: ext4 , "Theodore Ts'o" , cmm@us.ibm.com To: Manish Katiyar Return-path: Received: from mx2.redhat.com ([66.187.237.31]:58358 "EHLO mx2.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751532AbZAYRCD (ORCPT ); Sun, 25 Jan 2009 12:02:03 -0500 In-Reply-To: Sender: linux-ext4-owner@vger.kernel.org List-ID: Manish Katiyar wrote: > On Sun, Jan 25, 2009 at 9:45 PM, Eric Sandeen wrote: >> Manish Katiyar wrote: >>> On Tue, Jan 20, 2009 at 10:36 PM, Manish Katiyar wrote: >>>> ext2_quota_read doesn't bzeroes tmp_bh before calling ext2_get_block() >>>> where we access the b_size of it. Since it is a local variable it >>>> might contain some garbage. Make sure it is filled with zero before >>>> passing. >>> Hi Ted/mingming, >>> >>> Any feedback on this ?? >> This looks ok to me, Manish. I'm curious, did you see this fail in real >> life, and if so, what'd the failure look like? > > Actually no......I realised this while going through the code. I was > also wondering why we haven't hit this till now. Since ext{3,4} don't > have this issue, the only reason I can think of is because ext2 with > quota is not very much used or somehow we are lucky. > >> With the change, the tmp_bh bh_size is 0, so maxblocks down the >> get_block path is also 0, but I guess that works out ok. > > Yes, but that is better than having a random garbage. Isn't it ? Absolutely; it just struck me as a little odd but it's exactly what __ext2_get_block does as well, I probably just need to take a closer look at what it does w/ a 0 max. -Eric > Thanks - > Manish