From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Date: Wed, 19 Sep 2018 16:07:31 +0800 From: Ming Lei To: "jianchao.wang" Cc: Ming Lei , Jens Axboe , linux-block , Alan Stern , Christoph Hellwig , Bart Van Assche , Hannes Reinecke , Johannes Thumshirn , Adrian Hunter , James Bottomley , "Martin K. Petersen" , Linux SCSI List Subject: Re: [PATCH V3 11/17] SCSI: track pending admin commands Message-ID: <20180919080730.GC23172@ming.t460p> References: <5c13533d-b1ea-616e-07d0-99d59cf65f8a@oracle.com> <20180917113506.GA28349@ming.t460p> <1f91bfaf-7b9e-881d-7455-6f04bbb37823@oracle.com> <20180918073957.GB4805@ming.t460p> <20180918075544.GD4805@ming.t460p> <8d9049f4-a133-d725-d144-769b2d285660@oracle.com> <20180918121515.GA14172@ming.t460p> <47528f73-334a-16ef-5ec2-b3872b4eeeb7@oracle.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: <47528f73-334a-16ef-5ec2-b3872b4eeeb7@oracle.com> List-ID: On Wed, Sep 19, 2018 at 11:52:21AM +0800, jianchao.wang wrote: > Hi Ming > > On 09/18/2018 08:15 PM, Ming Lei wrote: > > The adminq request can be quisced in the scsi_device level, just as normal IO > > request, because the scsi_device instance is same with normal IO. > > Yes, the scsi_queue_rq->scsi_prep_state_check could gate things out of. > > > > > But in blk-mq level, this admin queue won't be quiesced or frozen at all. > > But the scsi_queue_rq/scsi_request_fn will be invoked, right ? > > This is different from the original implementation. For legacy path, as you mentioned, scsi_prep_fn() will work, then scsi_request_fn() won't be called. For blk-mq, scsi_queue_rq() is always called, and that is a bit different, but the result is fine given admin request is submitted much less frequently. Thanks, Ming