From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mx1.redhat.com ([209.132.183.28]:52216 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750995AbdLARgF (ORCPT ); Fri, 1 Dec 2017 12:36:05 -0500 Date: Sat, 2 Dec 2017 01:35:47 +0800 From: Ming Lei To: Bart Van Assche Cc: "mauricfo@linux.vnet.ibm.com" , "chenxiang66@hisilicon.com" , "hch@infradead.org" , "linux-block@vger.kernel.org" , "axboe@fb.com" , "osandov@fb.com" , "hare@suse.com" , "wenxiong@us.ibm.com" , "linuxarm@huawei.com" Subject: Re: [PATCH V2 0/2] block: fix queue freeze and cleanup Message-ID: <20171201173542.GB22528@ming.t460p> References: <20171123044800.25251-1-ming.lei@redhat.com> <20171127124147.GA16095@ming.t460p> <1512144528.2520.2.camel@sandisk.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: <1512144528.2520.2.camel@sandisk.com> Sender: linux-block-owner@vger.kernel.org List-Id: linux-block@vger.kernel.org On Fri, Dec 01, 2017 at 04:08:49PM +0000, Bart Van Assche wrote: > On Fri, 2017-12-01 at 13:36 -0200, Mauricio Faria de Oliveira wrote: > > On 11/29/2017 12:57 AM, chenxiang (M) wrote: > > > I applied this v2 patchset to kernel 4.15-rc1, running fio on a SATA > > > disk, then disable the disk with sysfs interface > > > (echo 0 > /sys/class/sas_phy/phy-1:0:1/enable), and find system is hung. > > > But with v1 patch, it doesn't > > > has this issue. Please have a check. > > > > Indeed, with this particular test-case (thanks, chenxiang) the problem > > can be recreated with PATCH v2 but _not_ with v1. > > > > For reference, I'm including the tests with v2 in this e-mail. > > The same tests have too been performed with v1, without blocked tasks. > > > > Interestingly, physical disk pulls did not hit the problem either > > on v1 or v2 (but it does not matter anymore) -- so v1 is the one. > > The test chenxiang ran does not prove that there is anything wrong with v2. > Maybe chenxiang hit the issue described in https://lkml.org/lkml/2017/9/5/381? No, if the issue is in SCSI EH, blk_drain_queue() shouldn't have moved on too. Anyway, V1/V3 is the correct thing to do. Thanks, Ming