Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S932414AbWAWA67 (ORCPT ); Sun, 22 Jan 2006 19:58:59 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S932422AbWAWA67 (ORCPT ); Sun, 22 Jan 2006 19:58:59 -0500 Received: from audible.transient.net ([216.254.12.79]:36320 "HELO audible.transient.net") by vger.kernel.org with SMTP id S932414AbWAWA66 (ORCPT ); Sun, 22 Jan 2006 19:58:58 -0500 Date: Sun, 22 Jan 2006 16:58:56 -0800 From: Jamie Heilman To: Arjan van de Ven Cc: Ariel , linux-ide@vger.kernel.org, linux-kernel@vger.kernel.org Subject: Re: memory leak in scsi_cmd_cache 2.6.15 Message-ID: <20060123005856.GB15490@fifty-fifty.audible.transient.net> Mail-Followup-To: Arjan van de Ven , Ariel , linux-ide@vger.kernel.org, linux-kernel@vger.kernel.org References: <1137917798.3328.2.camel@laptopd505.fenrus.org> <1137918044.3328.6.camel@laptopd505.fenrus.org> <1137956841.3328.36.camel@laptopd505.fenrus.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1137956841.3328.36.camel@laptopd505.fenrus.org> User-Agent: Mutt/1.5.11 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1531 Lines: 35 Arjan van de Ven wrote: > On Sun, 2006-01-22 at 13:51 -0500, Ariel wrote: > > On Sun, 22 Jan 2006, Arjan van de Ven wrote: > > > > > On Sun, 2006-01-22 at 09:16 +0100, Arjan van de Ven wrote: > > >> On Sat, 2006-01-21 at 21:13 -0500, Ariel wrote: > > >>> I have a memory leak in scsi_cmd_cache. > > > > >> does this happen without the binary nvidia driver too? (it appears > > >> you're using that). That's a good datapoint to have if so... > > > > I had the exact same nvidia driver with 2.6.12 (just recompiled) and it > > didn't happen there. > > > > But just in case I used slabtop to watch scsi_cmd_cache grow by 1.24KB per > > second (104MB per day), then I rmmoded nvidia and watched it grow by > > 1.16KB per > > please repeat this without nvidia ever being loaded. Just having a > module loaded before can already cause corruption that ripples through > later, so just unloading is not enough to get a clean result. Hrmph, yeah one of my workstations is exhibiting this too. I use the nvidia module, but I did clean reboot without it loaded and the leak was still there. So I'll add my data points at http://audible.transient.net/~jamie/k/ ... all the files starting with "scsi_cmd_cache." -- Jamie Heilman http://audible.transient.net/~jamie/ - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/