From patchwork Wed Nov 23 11:47:56 2011 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Stefan Hajnoczi X-Patchwork-Id: 127304 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Received: from lists.gnu.org (lists.gnu.org [140.186.70.17]) (using TLSv1 with cipher AES256-SHA (256/256 bits)) (Client did not present a certificate) by ozlabs.org (Postfix) with ESMTPS id 3868A1007D2 for ; Thu, 24 Nov 2011 00:09:27 +1100 (EST) Received: from localhost ([::1]:47233 helo=lists.gnu.org) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1RTBaM-0007os-TK for incoming@patchwork.ozlabs.org; Wed, 23 Nov 2011 07:05:54 -0500 Received: from eggs.gnu.org ([140.186.70.92]:47773) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1RTBZi-0005JT-0s for qemu-devel@nongnu.org; Wed, 23 Nov 2011 07:05:19 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1RTBK8-0002Dm-Ha for qemu-devel@nongnu.org; Wed, 23 Nov 2011 06:49:14 -0500 Received: from mtagate1.uk.ibm.com ([194.196.100.161]:37710) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1RTBK8-000241-AM for qemu-devel@nongnu.org; Wed, 23 Nov 2011 06:49:08 -0500 Received: from d06nrmr1507.portsmouth.uk.ibm.com (d06nrmr1507.portsmouth.uk.ibm.com [9.149.38.233]) by mtagate1.uk.ibm.com (8.13.1/8.13.1) with ESMTP id pANBm7tS019146 for ; Wed, 23 Nov 2011 11:48:07 GMT Received: from d06av05.portsmouth.uk.ibm.com (d06av05.portsmouth.uk.ibm.com [9.149.37.229]) by d06nrmr1507.portsmouth.uk.ibm.com (8.13.8/8.13.8/NCO v10.0) with ESMTP id pANBm6Ca2502818 for ; Wed, 23 Nov 2011 11:48:06 GMT Received: from d06av05.portsmouth.uk.ibm.com (loopback [127.0.0.1]) by d06av05.portsmouth.uk.ibm.com (8.14.4/8.13.1/NCO v10.0 AVout) with ESMTP id pANBm6TZ026596 for ; Wed, 23 Nov 2011 04:48:06 -0700 Received: from localhost (sig-9-79-13-177.uk.ibm.com [9.79.13.177]) by d06av05.portsmouth.uk.ibm.com (8.14.4/8.13.1/NCO v10.0 AVin) with ESMTP id pANBm5T5026582; Wed, 23 Nov 2011 04:48:06 -0700 From: Stefan Hajnoczi To: Date: Wed, 23 Nov 2011 11:47:56 +0000 Message-Id: <1322048878-26348-7-git-send-email-stefanha@linux.vnet.ibm.com> X-Mailer: git-send-email 1.7.7.1 In-Reply-To: <1322048878-26348-1-git-send-email-stefanha@linux.vnet.ibm.com> References: <1322048878-26348-1-git-send-email-stefanha@linux.vnet.ibm.com> X-detected-operating-system: by eggs.gnu.org: GNU/Linux 2.6, seldom 2.4 (older, 4) X-Received-From: 194.196.100.161 Cc: Kevin Wolf , Paolo Bonzini , Marcelo Tosatti , Stefan Hajnoczi Subject: [Qemu-devel] [PATCH v4 6/8] block: request overlap detection X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.14 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: qemu-devel-bounces+incoming=patchwork.ozlabs.org@nongnu.org Sender: qemu-devel-bounces+incoming=patchwork.ozlabs.org@nongnu.org Detect overlapping requests and remember to align to cluster boundaries if the image format uses them. This assumes that allocating I/O is performed in cluster granularity - which is true for qcow2, qed, etc. Signed-off-by: Stefan Hajnoczi --- block.c | 45 +++++++++++++++++++++++++++++++++++++++++++-- 1 files changed, 43 insertions(+), 2 deletions(-) diff --git a/block.c b/block.c index da7aaa2..c30c8f2 100644 --- a/block.c +++ b/block.c @@ -1133,21 +1133,62 @@ static void tracked_request_begin(BdrvTrackedRequest *req, QLIST_INSERT_HEAD(&bs->tracked_requests, req, list); } +/** + * Round a region to cluster boundaries + */ +static void round_to_clusters(BlockDriverState *bs, + int64_t sector_num, int nb_sectors, + int64_t *cluster_sector_num, + int *cluster_nb_sectors) +{ + BlockDriverInfo bdi; + + if (bdrv_get_info(bs, &bdi) < 0 || bdi.cluster_size == 0) { + *cluster_sector_num = sector_num; + *cluster_nb_sectors = nb_sectors; + } else { + int64_t c = bdi.cluster_size / BDRV_SECTOR_SIZE; + *cluster_sector_num = QEMU_ALIGN_DOWN(sector_num, c); + *cluster_nb_sectors = QEMU_ALIGN_UP(sector_num - *cluster_sector_num + + nb_sectors, c); + } +} + static bool tracked_request_overlaps(BdrvTrackedRequest *req, int64_t sector_num, int nb_sectors) { - return false; /* not yet implemented */ + /* aaaa bbbb */ + if (sector_num >= req->sector_num + req->nb_sectors) { + return false; + } + /* bbbb aaaa */ + if (req->sector_num >= sector_num + nb_sectors) { + return false; + } + return true; } static void coroutine_fn wait_for_overlapping_requests(BlockDriverState *bs, int64_t sector_num, int nb_sectors) { BdrvTrackedRequest *req; + int64_t cluster_sector_num; + int cluster_nb_sectors; bool retry; + /* If we touch the same cluster it counts as an overlap. This guarantees + * that allocating writes will be serialized and not race with each other + * for the same cluster. For example, in copy-on-read it ensures that the + * CoR read and write operations are atomic and guest writes cannot + * interleave between them. + */ + round_to_clusters(bs, sector_num, nb_sectors, + &cluster_sector_num, &cluster_nb_sectors); + do { retry = false; QLIST_FOREACH(req, &bs->tracked_requests, list) { - if (tracked_request_overlaps(req, sector_num, nb_sectors)) { + if (tracked_request_overlaps(req, cluster_sector_num, + cluster_nb_sectors)) { qemu_co_queue_wait(&req->wait_queue); retry = true; break;