From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Milburn Subject: Re: [PATCH] mvsas: defer non-NCQ cmds when NCQ cmds are active Date: Tue, 21 Sep 2010 19:24:46 -0500 Message-ID: <4C994CCE.3020607@redhat.com> References: <20100921191701.GA16726@dhcp-10-15-1-70.hsv.redhat.com> <1285102274.16440.132.camel@mulgrave.site> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit Return-path: Received: from mx1.redhat.com ([209.132.183.28]:26479 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751620Ab0IVAY2 (ORCPT ); Tue, 21 Sep 2010 20:24:28 -0400 In-Reply-To: <1285102274.16440.132.camel@mulgrave.site> Sender: linux-scsi-owner@vger.kernel.org List-Id: linux-scsi@vger.kernel.org To: James Bottomley Cc: linux-scsi@vger.kernel.org, jeff@garzik.org, k.skarlatos@gmail.com James Bottomley wrote: > On Tue, 2010-09-21 at 14:17 -0500, David Milburn wrote: >> Non-NCQ commands should run exclusively, ata_qc_issue warns when mixing >> non-NCQ and NCQ commands. >> >> Signed-off-by: David Milburn >> Reported-by: Konstantinos Skarlatos >> --- >> drivers/scsi/libsas/sas_ata.c | 1 + >> 1 files changed, 1 insertions(+), 0 deletions(-) >> >> diff --git a/drivers/scsi/libsas/sas_ata.c b/drivers/scsi/libsas/sas_ata.c >> index 042153c..ddbade7 100644 >> --- a/drivers/scsi/libsas/sas_ata.c >> +++ b/drivers/scsi/libsas/sas_ata.c >> @@ -347,6 +347,7 @@ static int sas_ata_scr_read(struct ata_link *link, unsigned int sc_reg_in, >> static struct ata_port_operations sas_sata_ops = { >> .phy_reset = sas_ata_phy_reset, >> .post_internal_cmd = sas_ata_post_internal, >> + .qc_defer = ata_std_qc_defer, > > What does this do on the card that doesn't need it: the aic94xx? That > one takes in a pipeline of commands and takes responsibility for getting > the issue right. Something like this throttles the pipeline higher than > the aic94xx is expecting. However, as long as it's not too much of a > performance problem, I don't mind applying it. James, For any driver using sas_sata_ops, it could effect performance for those drivers whose execute_task handles both, though I can't say I am real familiar with aic94xx. Thanks, David > > James > >