From mboxrd@z Thu Jan 1 00:00:00 1970 From: Douglas Gilbert Subject: Re: [PATCH 6/11] lpfc 8.1.1 : Fixes to error handlers Date: Tue, 29 Nov 2005 17:11:42 +1000 Message-ID: <438BFF2E.20407@torque.net> References: <9BB4DECD4CFE6D43AA8EA8D768ED51C21D7B76@xbl3.emulex.com> Reply-To: dougg@torque.net Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Return-path: Received: from zorg.st.net.au ([203.16.233.9]:61925 "EHLO borg.st.net.au") by vger.kernel.org with ESMTP id S1751338AbVK2HKh (ORCPT ); Tue, 29 Nov 2005 02:10:37 -0500 In-Reply-To: <9BB4DECD4CFE6D43AA8EA8D768ED51C21D7B76@xbl3.emulex.com> Sender: linux-scsi-owner@vger.kernel.org List-Id: linux-scsi@vger.kernel.org To: James.Smart@Emulex.Com Cc: hch@infradead.org, stefanr@s5r6.in-berlin.de, linux-scsi@vger.kernel.org James.Smart@Emulex.Com wrote: >>So you this storage vendor please provide us this test case >>as a start? > > > Basic test case: > - start load on the fc disks > - run sg_reset -b /dev/sgX or sg_reset -d /dev/sgX > > >>That beeing said I mentioned that we need to do SG-based >>resets via the >>EH thread aswell. As Emulex apparently cares about the above >>case would >>you mind implementing it? > > > Well - we would prefer the vendor test the system as it should behave > in normal circumstances rather than test-suite-derived conditions. > > Yes, we will sign up for this. However, we have several items on our > plate so we won't be on this quickly. I would hope there would be no > issue with accepting the patch as is and backing out the relatived code > as part of the eventual patch. We'll also be contacting the SG maintainer > to get his/her thoughts. While on the subject of resets, linux still has a fuzziness around a "device" reset. There should be two resets around this level: - target reset: reset a target device, its ports and all logical units within it [transport level] - logical unit reset: reset a logical unit only [this is a mandatory SCSI task management function] If changes are being made in this area, then perhaps this fuzziness should be addressed. Doug Gilbert