From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-scsi-owner@vger.kernel.org>
Date: Mon, 2 Jul 2018 06:03:10 -0700
From: Christoph Hellwig <hch@infradead.org>
To: Johannes Thumshirn <jthumshirn@suse.de>
Cc: Christoph Hellwig <hch@infradead.org>,
        Ming Lei <ming.lei@redhat.com>,
        James Bottomley <James.Bottomley@HansenPartnership.com>,
        linux-scsi@vger.kernel.org,
        "Martin K . Petersen" <martin.petersen@oracle.com>,
        linux-block@vger.kernel.org,
        Douglas Gilbert <dgilbert@interlog.com>,
        Bart Van Assche <bart.vanassche@wdc.com>,
        Jens Axboe <axboe@kernel.dk>, Omar Sandoval <osandov@fb.com>
Subject: Re: [PATCH] scsi: scsi_debug: introduce module parameter of
 'use_blk_mq'
Message-ID: <20180702130310.GA19610@infradead.org>
References: <20180702050642.26265-1-ming.lei@redhat.com>
 <20180702125422.GA600@infradead.org>
 <20180702125652.jxjcig63xvljfqow@linux-x5ow.site>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20180702125652.jxjcig63xvljfqow@linux-x5ow.site>
Sender: linux-scsi-owner@vger.kernel.org
List-ID: <linux-block@vger.kernel.org>

On Mon, Jul 02, 2018 at 02:56:52PM +0200, Johannes Thumshirn wrote:
> On Mon, Jul 02, 2018 at 05:54:22AM -0700, Christoph Hellwig wrote:
> > On Mon, Jul 02, 2018 at 01:06:42PM +0800, Ming Lei wrote:
> > > With the introduced module parameter of 'use_blk_mq', it is easy
> > > to switch between 'blk_mq' and 'non_blk_mq' by reloading scsi_debug
> > > module, so that we can test scsi_mq/blk_mq related regressions easily.
> > 
> > No.  We should not make a per driver choice.
> 
> Btw, wouldn't it be time for:
> 
> diff --git a/drivers/scsi/Kconfig b/drivers/scsi/Kconfig
> index 35c909bbf8ba..bd115bab162e 100644
> --- a/drivers/scsi/Kconfig
> +++ b/drivers/scsi/Kconfig
> @@ -49,6 +49,7 @@ config SCSI_NETLINK
>  
>  config SCSI_MQ_DEFAULT
>  	bool "SCSI: use blk-mq I/O path by default"
> +	default y
>  	depends on SCSI
>  	---help---
>  	  This option enables the new blk-mq based I/O path for SCSI
> 
> again and see how we've improved in the last year and a half?

Yes, totally.  We'll just need to sort out the error handling with
scsi and mq first (for 4.18 in fact).