From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mx1.redhat.com ([209.132.183.28]:49538 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1729016AbfABTLp (ORCPT ); Wed, 2 Jan 2019 14:11:45 -0500 From: Jeff Moyer To: Jens Axboe Cc: linux-fsdevel@vger.kernel.org, linux-aio@kvack.org, linux-block@vger.kernel.org, hch@lst.de, viro@zeniv.linux.org.uk Subject: Re: [PATCH 16/22] aio: add support for submission/completion rings References: <20181221192236.12866-1-axboe@kernel.dk> <20181221192236.12866-17-axboe@kernel.dk> Date: Wed, 02 Jan 2019 14:11:43 -0500 In-Reply-To: <20181221192236.12866-17-axboe@kernel.dk> (Jens Axboe's message of "Fri, 21 Dec 2018 12:22:30 -0700") Message-ID: MIME-Version: 1.0 Content-Type: text/plain Sender: linux-fsdevel-owner@vger.kernel.org List-ID: Jens Axboe writes: > The submission queue (SQ) and completion queue (CQ) rings are shared > between the application and the kernel. This eliminates the need to > copy data back and forth to submit and complete IO. We use the same > structures as the old aio interface. The SQ rings are indexes into a > struct iocb array, like we would submit through io_submit(), and the > CQ rings are struct io_event, like we would pass in (and copy back) > from io_getevents(). > > A new system call is added for this, io_ring_enter(). This system call > submits IO that is stored in the SQ ring, and/or completes IO and stores > the results in the CQ ring. Hence it's possible to both complete and > submit IO in a single system call. > > For IRQ driven IO, an application only needs to enter the kernel for > completions if it wants to wait for them to occur. It looks like the memory for the submission and completion queues is not migratable (scq rings are non-migratory! [1]). Maybe you should use the aio pseudofs for allocations, like is done with the original aio ring. I realize that would change the api... -Jeff [1] https://www.youtube.com/watch?v=liIlW-ovx0Y