From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Subject: Re: [PATCH 0/3] scsi: scsi-mq: don't hold host_busy in IO path To: Bart Van Assche , "ming.lei@redhat.com" Cc: "linux-block@vger.kernel.org" , "snitzer@redhat.com" , "hch@lst.de" , "martin.petersen@oracle.com" , "hare@suse.de" , "linux-scsi@vger.kernel.org" , "don.brace@microsemi.com" , "james.bottomley@hansenpartnership.com" , "osandov@fb.com" , "loberman@redhat.com" , "kashyap.desai@broadcom.com" References: <20180420065742.8043-1-ming.lei@redhat.com> <87619135ac0bfefa07f510170024b609e641db0c.camel@wdc.com> <607c4d2f-28e7-4152-dbf5-7fb4387e9413@kernel.dk> <647b4fe833fdf001f724cfa0803c9e3a79d78b13.camel@wdc.com> From: Jens Axboe Message-ID: Date: Fri, 27 Apr 2018 10:19:54 -0600 MIME-Version: 1.0 In-Reply-To: <647b4fe833fdf001f724cfa0803c9e3a79d78b13.camel@wdc.com> Content-Type: text/plain; charset=utf-8 List-ID: On 4/27/18 9:48 AM, Bart Van Assche wrote: > On Fri, 2018-04-27 at 09:39 -0600, Jens Axboe wrote: >> blk_mq_tagset_busy_iter(&shost->tag_set, scsi_host_check_in_flight, >> &in_flight); >> return in_flight.cnt + atomic_read(&shost->host_busy); >> >> The atomic read is basically free, once we get rid of the dirty of that >> variable on each IO. > > Hello Jens, > > What makes you think that " + atomic_read(&shost->host_busy)" is necessary? > I am not aware of any code outside the SCSI core that modifies the host_busy > member. It's a (scalable) hack to count those as well. Going forward they should be converted to just using reserved tags, of course. -- Jens Axboe