From mboxrd@z Thu Jan 1 00:00:00 1970 From: Mike Anderson Subject: Re: [dm-devel] [PATCH v5] dm mpath: avoid call to blk_abort_queue by default Date: Thu, 18 Nov 2010 12:39:55 -0800 Message-ID: <20101118203954.GA6241@linux.vnet.ibm.com> References: <1290107959-17167-1-git-send-email-snitzer@redhat.com> <1290110853-17541-1-git-send-email-snitzer@redhat.com> <20101118201841.GE30435@agk-dp.fab.redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from e6.ny.us.ibm.com ([32.97.182.146]:41472 "EHLO e6.ny.us.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1760322Ab0KRUj7 (ORCPT ); Thu, 18 Nov 2010 15:39:59 -0500 Received: from d01relay03.pok.ibm.com (d01relay03.pok.ibm.com [9.56.227.235]) by e6.ny.us.ibm.com (8.14.4/8.13.1) with ESMTP id oAIKf1aS032100 for ; Thu, 18 Nov 2010 15:41:01 -0500 Received: from d01av04.pok.ibm.com (d01av04.pok.ibm.com [9.56.224.64]) by d01relay03.pok.ibm.com (8.13.8/8.13.8/NCO v10.0) with ESMTP id oAIKdwqn302158 for ; Thu, 18 Nov 2010 15:39:58 -0500 Received: from d01av04.pok.ibm.com (loopback [127.0.0.1]) by d01av04.pok.ibm.com (8.14.4/8.13.1/NCO v10.0 AVout) with ESMTP id oAIKdwhN013385 for ; Thu, 18 Nov 2010 15:39:58 -0500 Content-Disposition: inline In-Reply-To: <20101118201841.GE30435@agk-dp.fab.redhat.com> Sender: linux-scsi-owner@vger.kernel.org List-Id: linux-scsi@vger.kernel.org To: Mike Snitzer , dm-devel@redhat.com, Mike Christie , linux-scsi@vger.kernel.org Alasdair G Kergon wrote: > On Thu, Nov 18, 2010 at 03:07:33PM -0500, Mike Snitzer wrote: > > It is expected that this race will be fixed in the near-term so it makes > > little since to remove all associated code. Providing control over the > > call to blk_abort_queue() facilitates continued testing and future > > flexibility to opt-in to lower latency path deactivation. Opting to > > enable this feature will emit a warning for the time being. > > Ah, but then the *feature* keyword - which is only going to be needed > temporarily until it's fixed - becomes a permanent part of the interface... > > If it's broken, let's just disable it (#define-style) until it gets fixed. > (And only if it doesn't get fixed in a reasonable amount of time, remove the > code.) I understand that if its broken that we could #define it out in dm or we could #define the code in blk_abort_queue. I believe the thought was that post addressing the race issue that others may want control. As there maybe cases where the waking up of the error_handler and a extra abort may not be wanted in all cases. -andmike -- Michael Anderson andmike@linux.vnet.ibm.com