From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mx1.redhat.com ([209.132.183.28]:57962 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751013AbdGZKnr (ORCPT ); Wed, 26 Jul 2017 06:43:47 -0400 Received: from smtp.corp.redhat.com (int-mx04.intmail.prod.int.phx2.redhat.com [10.5.11.14]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id 8D83C7EA8B for ; Wed, 26 Jul 2017 10:43:47 +0000 (UTC) Date: Wed, 26 Jul 2017 18:43:45 +0800 From: Eryu Guan Subject: Re: [BUG] xfs/305 hangs 4.10-rc4 kernel Message-ID: <20170726104345.GZ9167@eguan.usersys.redhat.com> References: <20170126032950.GM1859@eguan.usersys.redhat.com> <20170126184427.GA39683@bfoster.bfoster> <20170127025219.GR1859@eguan.usersys.redhat.com> <20170130181224.GC8737@bfoster.bfoster> <20170130215952.GA11230@bfoster.bfoster> <20170204114700.GH1859@eguan.usersys.redhat.com> <20170206175908.GI57865@bfoster.bfoster> <20170213164307.GA25568@bfoster.bfoster> <20170726050952.GT9167@eguan.usersys.redhat.com> <20170726103502.GA59833@bfoster.bfoster> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20170726103502.GA59833@bfoster.bfoster> Sender: linux-xfs-owner@vger.kernel.org List-ID: List-Id: xfs To: Brian Foster Cc: linux-xfs@vger.kernel.org On Wed, Jul 26, 2017 at 06:35:03AM -0400, Brian Foster wrote: > On Wed, Jul 26, 2017 at 01:09:52PM +0800, Eryu Guan wrote: > > Hi Brian, > > > > [sorry for bringing up an old thread, but this seems relavent] > > [snip] > > > > Is xfs/305 supposed to pass with 4.13-rc2 kernel now? I know there're > > some quota patches went in XFS recently, but I'm not sure if there's any > > known leftover bug. > > > > Anyway, I hit a test hang with 4.13-rc2 kernel with xfs/305 again, test > > configuration is rmapbt XFS with 4k block size on ppc64le host. But I > > haven't tried to reproduce it yet. Just want to confirm with you first. > > > > I believe it is still possible for quotaoff to deadlock the log. It is > not shown below, but I assume you have a blocked xfs_quota process > attempting to run a quotaoff..? How big is your fs/log? I'm not sure if there's a quotaoff blocked too, that's all I have for now, this was an automated test job and it didn't print any sysrq-w output as expected. The fs is 15G but I don't know much about the log, the host host was returned to the test pool.. > > I sent a patch a while ago, but it wasn't safe because it could > potentially lead to nested transactions in the quotaoff context. We > don't have a proper fix for this as of yet. Ok, thanks! At least this isn't something totally surprising people, I'll see if I can hit it again in future test runs. Thanks, Eryu