From mboxrd@z Thu Jan 1 00:00:00 1970 From: Kashyap Desai Subject: RE: [PATCH] scsi.mq:Added enable_cmd_list flags in hostt to reduce lock contention Date: Wed, 20 Aug 2014 18:16:13 +0530 Message-ID: References: <201408191819.s7JIIx12019025@palmhbs0.lsi.com> <53F47CF8.3080708@acm.org> <7932bf422121b7f7d23545abdc0065c8@mail.gmail.com> <20140820124115.GA15231@infradead.org> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Return-path: Received: from exprod7og105.obsmtp.com ([64.18.2.163]:33078 "EHLO exprod7og105.obsmtp.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754373AbaHTMqQ (ORCPT ); Wed, 20 Aug 2014 08:46:16 -0400 Received: by mail-lb0-f172.google.com with SMTP id z11so6756318lbi.17 for ; Wed, 20 Aug 2014 05:46:14 -0700 (PDT) In-Reply-To: <20140820124115.GA15231@infradead.org> Sender: linux-scsi-owner@vger.kernel.org List-Id: linux-scsi@vger.kernel.org To: Christoph Hellwig Cc: Bart Van Assche , linux-scsi@vger.kernel.org, aacraid@adaptec.com, Elliott@hp.com, jbottomley@parallels.com > -----Original Message----- > From: Christoph Hellwig [mailto:hch@infradead.org] > Sent: Wednesday, August 20, 2014 6:11 PM > To: Kashyap Desai > Cc: Bart Van Assche; linux-scsi@vger.kernel.org; aacraid@adaptec.com; > Elliott@hp.com; jbottomley@parallels.com; hch@infradead.org > Subject: Re: [PATCH] scsi.mq:Added enable_cmd_list flags in hostt to reduce > lock contention > > On Wed, Aug 20, 2014 at 06:08:37PM +0530, Kashyap Desai wrote: > > > This code is in the hot path which means that caching "enable_cmd_list" > > > in struct Scsi_Host (as is done for many other SCSI host parameters) > > > probably will (slightly) improve performance further. Otherwise this > > > patch looks fine to me. > > I will send updated patch which will cache host template field > > "enable_cmd_list" for faster access in Scsi Host. > > Thanks. It might be worth to only set in in the host in fact. Fine. I will remove host template entry and add code in aacraid and dpt_i2o drivers to set that value directly in "Scsi Host" > > Also please just remove the code about lock contention in scsi_mq_prep_fn > - the XXX really doesn't apply anymore and I think the code should be self- > explaining enough to not need a comment. I will do this. > > Otherwise the patch looks good to me.