From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from verein.lst.de ([213.95.11.211]:44100 "EHLO newverein.lst.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756164AbcISNte (ORCPT ); Mon, 19 Sep 2016 09:49:34 -0400 Date: Mon, 19 Sep 2016 15:49:32 +0200 From: Christoph Hellwig To: Alexander Gordeev Cc: Christoph Hellwig , axboe@fb.com, tglx@linutronix.de, keith.busch@intel.com, linux-block@vger.kernel.org, linux-kernel@vger.kernel.org Subject: Re: [PATCH 10/13] blk-mq: provide a default queue mapping for PCI device Message-ID: <20160919134932.GB28585@lst.de> References: <1473862739-15032-1-git-send-email-hch@lst.de> <1473862739-15032-11-git-send-email-hch@lst.de> <20160919073313.GB31001@agordeev.lab.eng.brq.redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: <20160919073313.GB31001@agordeev.lab.eng.brq.redhat.com> Sender: linux-block-owner@vger.kernel.org List-Id: linux-block@vger.kernel.org On Mon, Sep 19, 2016 at 09:33:14AM +0200, Alexander Gordeev wrote: > Considering this code is to be used by any device - an assumption > that queue number and interrupt vector are always equal seems > questionable. I.e. what about non-contiguous MSI-Xs? It's just a generic default - see the discussion with Jens in the intro mail for the thread for anyother example that can't be handled easily. If we get enough devices for a given scheme we can add another common helper, in the worst case we'll need driver code for the scheme.