From mboxrd@z Thu Jan 1 00:00:00 1970 From: Christoph Hellwig Subject: Re: [REGRESSION v4.3] scsi_dh: use-after-free when removing scsi device Date: Sun, 4 Oct 2015 09:42:35 +0200 Message-ID: <20151004074235.GA13912@lst.de> References: <20150930003549.GA4857@xzibit.linux.bs1.fc.nec.co.jp> <20150930151857.GA26594@lst.de> <20151001005640.GA4821@xzibit.linux.bs1.fc.nec.co.jp> <20151001043844.GA8397@xzibit.linux.bs1.fc.nec.co.jp> <20151001052159.GA14247@lst.de> <20151001114021.GA15125@xzibit.linux.bs1.fc.nec.co.jp> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from verein.lst.de ([213.95.11.211]:49796 "EHLO newverein.lst.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751225AbbJDHmi (ORCPT ); Sun, 4 Oct 2015 03:42:38 -0400 Content-Disposition: inline In-Reply-To: <20151001114021.GA15125@xzibit.linux.bs1.fc.nec.co.jp> Sender: linux-scsi-owner@vger.kernel.org List-Id: linux-scsi@vger.kernel.org To: Junichi Nomura Cc: Christoph Hellwig , linux-scsi , Hannes Reinecke On Thu, Oct 01, 2015 at 11:40:23AM +0000, Junichi Nomura wrote: > On 10/01/15 14:21, Christoph Hellwig wrote: > > Any chance you could share all your multipath tests in a git repository > > somewhere? It seems like you're the only one actually having a good > > set of reproducable but minimalistic tests. > > Hmm, sorry I don't have a public git repository... > > I'm using a pair of KVM guest, one for iSCSI target and the other > for testing. Any chance you could share your various scripts in some way to that people doing multipath changes can run them to verify those changes? > > I could reproduce the crash using this loop, within a few minutes: > > service multipathd start > while true; do > multipath -F > iscsiadm -m node --logout > iscsiadm -m node --login > done > > It might implicitly depend on udev to do some small amount of I/O > after device uevent though. I can't reproduce this unfortunately. I suspect udev doesn't do enough stupid things on my old Debian test system.