From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: From: Bart Van Assche To: "ming.lei@redhat.com" , "jianchao.w.wang@oracle.com" CC: "linux-kernel@vger.kernel.org" , "linux-block@vger.kernel.org" , "axboe@kernel.dk" Subject: Re: [RFC] blk-mq: clean up the hctx restart Date: Thu, 2 Aug 2018 15:52:00 +0000 Message-ID: <9feaa41702ef6fcc00ce1b8aa19bbe179edf4e3f.camel@wdc.com> References: <1533009735-2221-1-git-send-email-jianchao.w.wang@oracle.com> <20180731045805.GE15701@ming.t460p> <8a3383e6-2926-6858-d8f2-671f3cb9e460@oracle.com> <20180731061616.GF15701@ming.t460p> <42371198-2a4b-1062-3564-411645ffba98@oracle.com> <20180801085841.GA27962@ming.t460p> In-Reply-To: <20180801085841.GA27962@ming.t460p> Content-Type: text/plain; charset="utf-7" MIME-Version: 1.0 List-ID: On Wed, 2018-08-01 at 16:58 +-0800, Ming Lei wrote: +AD4- On Wed, Aug 01, 2018 at 10:17:30AM +-0800, jianchao.wang wrote: +AD4- +AD4- However, due to the limits in hctx+AF8-may+AF8-queue, q+AF8-b s= till cannot get the +AD4- +AD4- tags. The RR restart also will not wake up q+AF8-a. +AD4- +AD4- This is unfair for q+AF8-a. +AD4- +AD4-=20 +AD4- +AD4- When we remove RR restart fashion, at least, the q+AF8-a will b= e waked up by +AD4- +AD4- the hctx restart. +AD4- +AD4- Is this the improvement of fairness you said in driver tag allo= cation ? +AD4-=20 +AD4- I mean the fairness is totally covered by the general tag allocation +AD4- algorithm now, which is sort of FIFO style because of waitqueue, but = RR +AD4- restart wakes up queue in the order of request queue. >>From sbitmap.h: +ACM-define SBQ+AF8-WAIT+AF8-QUEUES 8 What do you think is the effect of your patch if more than eight LUNs are active and the SCSI queue is full? Thanks, Bart.