From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jeff Garzik Subject: Re: [PATCH] libata: fix use-after-free during driver unload/unplug Date: Thu, 26 May 2005 21:21:05 -0400 Message-ID: <42967601.7080003@pobox.com> References: <42962379.5000206@pobox.com> <42964099.6000207@pobox.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Return-path: Received: from mail.dvmed.net ([216.237.124.58]:11996 "EHLO mail.dvmed.net") by vger.kernel.org with ESMTP id S261269AbVE0BVL (ORCPT ); Thu, 26 May 2005 21:21:11 -0400 In-Reply-To: <42964099.6000207@pobox.com> Sender: linux-ide-owner@vger.kernel.org List-Id: linux-ide@vger.kernel.org To: Mark Lord Cc: "linux-ide@vger.kernel.org" , Linux Kernel , Jens Axboe , Al Viro Mark Lord wrote: > Hi Jeff! > > Jeff Garzik wrote: > >> >> * add ->host_stop_prewalk() hook, use it in sata_qstor.c (hi Mark). >> sata_qstor appears to require the host-stop-before-port-stop ordering >> that existed prior to applying the attached patch. > > > Mmm.. I'm a little bit rusty here, but I don't think qstor > cares about the order, so long as ports are marked with > ATA_FLAG_PORT_DISABLED before invoking port_stop(). > > I've tried to allow disabling/enabling individual ports > on-the-fly as needed, even though it never really happens > in practice. So host_stop() kills the whole chip, whereas > port_stop() I took to mean just one of the four SATA ports. qstor's ->host_stop() disables global interrupts, and I didn't know if you really wanted to do that prior to ->port_stop(). I would much prefer to eliminate the ->host_stop_prewalk() hook, and simply call ->host_stop after all ->port_stop() calls complete, if qstor doesn't need the pre-walk. Jeff