From mboxrd@z Thu Jan 1 00:00:00 1970 From: Evgeniy Polyakov Subject: Re: NFS/credentials leak in 2.6.29-rc1 Date: Tue, 20 Jan 2009 16:49:01 +0300 Message-ID: <20090120134900.GA22667@ioremap.net> References: <20090120114649.GA15832@ioremap.net> <1232458623.7433.4.camel@heimdal.trondhjem.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: linux-fsdevel@vger.kernel.org To: Trond Myklebust Return-path: Received: from cs-studio.ru ([195.178.208.66]:43647 "EHLO tservice.net.ru" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750838AbZATNtE (ORCPT ); Tue, 20 Jan 2009 08:49:04 -0500 Content-Disposition: inline In-Reply-To: <1232458623.7433.4.camel@heimdal.trondhjem.org> Sender: linux-fsdevel-owner@vger.kernel.org List-ID: On Tue, Jan 20, 2009 at 08:37:03AM -0500, Trond Myklebust (trond.myklebust@fys.uio.no) wrote: > This leak is occurring on the server only and not the client, is that > correct? Yes, I only observed it on the server. I think I can reproduce it here, at least I already crashed machine 3 times. It looks like it fires with random IO (can not say if read or write though), since the same overall amount of data read or written sequentially did not result in OOM condition, but leak still may be presented there if credential is attached in single readpages()... Do not know. > I don't think this is a known issue but I've forwarded your mail on to > Bruce Fields, who is the current NFS server maintainer, for further > information. Thank you. -- Evgeniy Polyakov