From mboxrd@z Thu Jan 1 00:00:00 1970 From: =?utf-8?B?SsO2cm4=?= Engel Subject: Re: [patch 1/9] fs: libfs buffered write leak fix Date: Sat, 3 Feb 2007 17:49:47 +0000 Message-ID: <20070203174947.GA2656@lazybastard.org> References: <20070129081905.23584.97878.sendpatchset@linux.site> <20070129081914.23584.23886.sendpatchset@linux.site> <20070202155236.dae54aa2.akpm@linux-foundation.org> <20070203013316.GB27300@wotan.suse.de> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: Andrew Morton , Linux Kernel , Linux Filesystems , Linux Memory Management To: Nick Piggin Return-path: Content-Disposition: inline In-Reply-To: <20070203013316.GB27300@wotan.suse.de> Sender: linux-kernel-owner@vger.kernel.org List-Id: linux-fsdevel.vger.kernel.org On Sat, 3 February 2007 02:33:16 +0100, Nick Piggin wrote: >=20 > If doing a partial-write, simply clear the whole page and set it upto= date > (don't need to get too tricky). That sounds just like a bug I recently fixed in logfs. prepare_write() would clear the page, commit_write() would write the whole page. Bug can be reproduced with a simple testcate: echo -n foo > foo echo -n bar >> foo cat foo With the bug, the second write will replace "foo" with "\0\0\0" and cat will return "bar". Doing a read instead of clearing the page will return "foobar", as would be expected. Can you hit the same bug with your patch or did I miss something? J=C3=B6rn --=20 When people work hard for you for a pat on the back, you've got to give them that pat. -- Robert Heinlein