From mboxrd@z Thu Jan 1 00:00:00 1970 From: Christoph Hellwig Subject: Re: [PATCH 4/6] virtio-blk: implement init_flush_rq Date: Sun, 7 Sep 2014 20:52:59 +0200 Message-ID: <20140907185259.GC23026@lst.de> References: <1410079162-9872-1-git-send-email-ming.lei@canonical.com> <1410079162-9872-5-git-send-email-ming.lei@canonical.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <1410079162-9872-5-git-send-email-ming.lei@canonical.com> Sender: linux-kernel-owner@vger.kernel.org To: Ming Lei Cc: Jens Axboe , linux-kernel@vger.kernel.org, linux-scsi@vger.kernel.org List-Id: linux-scsi@vger.kernel.org A couple comments not directly related to this patch, it's just a convenient vehicle for my rants :) > @@ -556,6 +555,19 @@ static int virtblk_init_request(void *data, struct request *rq, > struct virtio_blk *vblk = data; > struct virtblk_req *vbr = blk_mq_rq_to_pdu(rq); > > + vbr->req = rq; I really hate how we need these backpointers in most drivers. Given that struct request and the driver privata data are allocated together we should be able to do this with simple pointer arithmetics. > + sg_init_table(vbr->sg, vblk->sg_elems); Jens, what do you think about moving of the SG list handling to the core block layer? I'd really like to have the S/G list in struct request, and if we do that we could also take the scsi-mq code that allows small S/G lists preallocated in blk-mq and allocating larger ones at runtime there, avoiding the need to duplicate that code and the whole mempool magic it requires in drivers that want to make use of it.