From mboxrd@z Thu Jan 1 00:00:00 1970 From: Matthew Wilcox Subject: Re: [PATCH 2/7] mpt3sas: Remove use of DEF_SCSI_QCMD Date: Wed, 2 Apr 2014 09:20:14 -0400 Message-ID: <20140402132014.GD5667@linux.intel.com> References: <2ce8dda9284f463216f4c527d32fee497f9da3ac.1395952569.git.matthew.r.wilcox@intel.com> <20140402104115.GB6510@infradead.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from mga09.intel.com ([134.134.136.24]:5999 "EHLO mga09.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1758688AbaDBNWE (ORCPT ); Wed, 2 Apr 2014 09:22:04 -0400 Content-Disposition: inline In-Reply-To: <20140402104115.GB6510@infradead.org> Sender: linux-scsi-owner@vger.kernel.org List-Id: linux-scsi@vger.kernel.org To: Christoph Hellwig Cc: Matthew Wilcox , linux-scsi@vger.kernel.org On Wed, Apr 02, 2014 at 03:41:15AM -0700, Christoph Hellwig wrote: > On Thu, Mar 27, 2014 at 04:40:31PM -0400, Matthew Wilcox wrote: > > Removing the host_lock from the I/O submission path gives a huge > > scalability improvement. > > This looks reasonable to me, but did you do a full audit that all state > is properly synchronized in the functions called from _scsih_qcmd_lck? > > If so that should be stated in the patch description. I did one for mpt2sas back in July 2011 when I first submitted this patch. http://marc.info/?l=linux-scsi&m=130980763617589&w=2 I must confess to having not done so for this iteration. The equivalent to this patch has been in RHEL since 6.2, so I hope that gives us some amount of confidence.