From: hare@suse.de (Hannes Reinecke) Subject: [PATCH 0/2] Update command result handling Date: Thu, 19 Nov 2009 13:24:49 +0100 Message-ID: <20091119122449.D13B93A085@ochil.suse.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Return-path: Received: from cantor.suse.de ([195.135.220.2]:50601 "EHLO mx1.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750992AbZKSMYo (ORCPT ); Thu, 19 Nov 2009 07:24:44 -0500 Sender: linux-scsi-owner@vger.kernel.org List-Id: linux-scsi@vger.kernel.org To: James Bottomley Cc: dm-devel@redhat.com, linux-scsi@vger.kernel.org Currently we're updating the request result upon completion only for BLK_PC requests. This makes it impossible for the upper layers to reliable detect the real cause for an I/O failure. By attaching the result and the sense to all requests we can update multipathing to make some more elaborate choices on how to handle I/O errors. This also solves a potential data corruption with multipathing and persistent reservations. When queue_if_no_path is active multipath will queue any I/O failure (including those failed with RESERVATION CONFLICT) until the reservation status changes. But by then I/O might have been ongoing on the other paths, thus the delayed submission will severely corrupt your data. Signed-off-by: Hannes Reinecke