Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S965454AbbEFDO0 (ORCPT ); Tue, 5 May 2015 23:14:26 -0400 Received: from youngberry.canonical.com ([91.189.89.112]:38338 "EHLO youngberry.canonical.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757933AbbEFDOP (ORCPT ); Tue, 5 May 2015 23:14:15 -0400 MIME-Version: 1.0 In-Reply-To: <20150505165541.GV1971@htj.duckdns.org> References: <1430826595-5888-1-git-send-email-ming.lei@canonical.com> <1430826595-5888-3-git-send-email-ming.lei@canonical.com> <20150505135958.GO1971@htj.duckdns.org> <20150505165541.GV1971@htj.duckdns.org> Date: Wed, 6 May 2015 11:14:10 +0800 Message-ID: Subject: Re: [PATCH 2/2] block: loop: avoiding too many pending per work I/O From: Ming Lei To: Tejun Heo Cc: Jens Axboe , Linux Kernel Mailing List , "Justin M. Forbes" , Jeff Moyer , Christoph Hellwig , "v4.0" Content-Type: text/plain; charset=UTF-8 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 2151 Lines: 49 On Wed, May 6, 2015 at 12:55 AM, Tejun Heo wrote: > Hello, Ming. > > On Tue, May 05, 2015 at 10:46:10PM +0800, Ming Lei wrote: >> On Tue, May 5, 2015 at 9:59 PM, Tejun Heo wrote: >> > It's a bit weird to hard code this to 16 as this effectively becomes a >> > hidden bottleneck for concurrency. For cases where 16 isn't a good >> > value, hunting down what's going on can be painful as it's not visible >> > anywhere. I still think the right knob to control concurrency is >> > nr_requests for the loop device. You said that for linear IOs, it's >> > better to have higher nr_requests than concurrency but can you >> > elaborate why? >> >> I mean, in case of sequential IO, the IO may hit page cache a bit easier, >> so handling the IO may be quite quick, then it is often more efficient to >> handle them in one same context(such as, handle one by one from IO >> queue) than from different contexts(scheduled from different worker >> threads). And that can be made by setting a bigger nr_requests(queue_depth). > > Ah, so, it's about the queueing latency. Blocking the issuer from > get_request side for the same level of concurrency would incur a lot > longer latency before the next IO can be dispatched. The arbitrary 16 > is still bothering but for now it's fine I guess, but we need to > revisit the whole thing including WQ_HIGHPRI thing. Maybe it made > sense when we had only one thread servicing all IOs but w/ high > concurrency I don't think it's a good idea. Yes, I was thinking about it too, but concurrency can improve random I/O throughput a lot in my tests. Also I have patches to use aio/dio for loop, then one thread is enough, and both double cache and high context switch can be avoided. I will post them later for review. > > Please feel free to add > > Acked-by: Tejun Heo Thanks for your review and ack! thanks, Ming Lei -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/