From patchwork Thu Feb 25 20:58:29 2016 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: John Snow X-Patchwork-Id: 8426591 Return-Path: X-Original-To: patchwork-qemu-devel@patchwork.kernel.org Delivered-To: patchwork-parsemail@patchwork2.web.kernel.org Received: from mail.kernel.org (mail.kernel.org [198.145.29.136]) by patchwork2.web.kernel.org (Postfix) with ESMTP id C32A7C0553 for ; Thu, 25 Feb 2016 20:59:38 +0000 (UTC) Received: from mail.kernel.org (localhost [127.0.0.1]) by mail.kernel.org (Postfix) with ESMTP id 225952024C for ; Thu, 25 Feb 2016 20:59:38 +0000 (UTC) Received: from lists.gnu.org (lists.gnu.org [208.118.235.17]) (using TLSv1 with cipher AES256-SHA (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 4CC0D2022A for ; Thu, 25 Feb 2016 20:59:37 +0000 (UTC) Received: from localhost ([::1]:45811 helo=lists.gnu.org) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1aZ30a-0002c4-QA for patchwork-qemu-devel@patchwork.kernel.org; Thu, 25 Feb 2016 15:59:36 -0500 Received: from eggs.gnu.org ([2001:4830:134:3::10]:33214) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1aZ2zg-0001To-L4 for qemu-devel@nongnu.org; Thu, 25 Feb 2016 15:58:41 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1aZ2zf-0008FN-Nl for qemu-devel@nongnu.org; Thu, 25 Feb 2016 15:58:40 -0500 Received: from mx1.redhat.com ([209.132.183.28]:45882) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1aZ2zc-0008EY-Mn; Thu, 25 Feb 2016 15:58:36 -0500 Received: from int-mx13.intmail.prod.int.phx2.redhat.com (int-mx13.intmail.prod.int.phx2.redhat.com [10.5.11.26]) by mx1.redhat.com (Postfix) with ESMTPS id 4029513982D; Thu, 25 Feb 2016 20:58:36 +0000 (UTC) Received: from scv.usersys.redhat.com (vpn-55-17.rdu2.redhat.com [10.10.55.17]) by int-mx13.intmail.prod.int.phx2.redhat.com (8.14.4/8.14.4) with ESMTP id u1PKwWDT028422; Thu, 25 Feb 2016 15:58:35 -0500 From: John Snow To: qemu-block@nongnu.org Date: Thu, 25 Feb 2016 15:58:29 -0500 Message-Id: <1456433911-24718-2-git-send-email-jsnow@redhat.com> In-Reply-To: <1456433911-24718-1-git-send-email-jsnow@redhat.com> References: <1456433911-24718-1-git-send-email-jsnow@redhat.com> X-Scanned-By: MIMEDefang 2.68 on 10.5.11.26 X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.5.16 (mx1.redhat.com [10.5.110.39]); Thu, 25 Feb 2016 20:58:36 +0000 (UTC) X-detected-operating-system: by eggs.gnu.org: GNU/Linux 2.2.x-3.x [generic] X-Received-From: 209.132.183.28 Cc: kwolf@redhat.com, famz@redhat.com, jcody@redhat.com, qemu-devel@nongnu.org, stefanha@redhat.com, John Snow Subject: [Qemu-devel] [PATCH v4 1/3] block/backup: make backup cluster size configurable 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+patchwork-qemu-devel=patchwork.kernel.org@nongnu.org Sender: qemu-devel-bounces+patchwork-qemu-devel=patchwork.kernel.org@nongnu.org X-Spam-Status: No, score=-1.9 required=5.0 tests=BAYES_00, UNPARSEABLE_RELAY autolearn=unavailable version=3.3.1 X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on mail.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP 64K might not always be appropriate, make this a runtime value. Signed-off-by: John Snow Reviewed-by: Fam Zheng --- block/backup.c | 64 +++++++++++++++++++++++++++++++++------------------------- 1 file changed, 36 insertions(+), 28 deletions(-) diff --git a/block/backup.c b/block/backup.c index 00cafdb..76addef 100644 --- a/block/backup.c +++ b/block/backup.c @@ -21,10 +21,7 @@ #include "qemu/ratelimit.h" #include "sysemu/block-backend.h" -#define BACKUP_CLUSTER_BITS 16 -#define BACKUP_CLUSTER_SIZE (1 << BACKUP_CLUSTER_BITS) -#define BACKUP_SECTORS_PER_CLUSTER (BACKUP_CLUSTER_SIZE / BDRV_SECTOR_SIZE) - +#define BACKUP_CLUSTER_SIZE_DEFAULT (1 << 16) #define SLICE_TIME 100000000ULL /* ns */ typedef struct CowRequest { @@ -46,9 +43,16 @@ typedef struct BackupBlockJob { CoRwlock flush_rwlock; uint64_t sectors_read; HBitmap *bitmap; + int64_t cluster_size; QLIST_HEAD(, CowRequest) inflight_reqs; } BackupBlockJob; +/* Size of a cluster in sectors, instead of bytes. */ +static inline int64_t cluster_size_sectors(BackupBlockJob *job) +{ + return job->cluster_size / BDRV_SECTOR_SIZE; +} + /* See if in-flight requests overlap and wait for them to complete */ static void coroutine_fn wait_for_overlapping_requests(BackupBlockJob *job, int64_t start, @@ -97,13 +101,14 @@ static int coroutine_fn backup_do_cow(BlockDriverState *bs, QEMUIOVector bounce_qiov; void *bounce_buffer = NULL; int ret = 0; + int64_t sectors_per_cluster = cluster_size_sectors(job); int64_t start, end; int n; qemu_co_rwlock_rdlock(&job->flush_rwlock); - start = sector_num / BACKUP_SECTORS_PER_CLUSTER; - end = DIV_ROUND_UP(sector_num + nb_sectors, BACKUP_SECTORS_PER_CLUSTER); + start = sector_num / sectors_per_cluster; + end = DIV_ROUND_UP(sector_num + nb_sectors, sectors_per_cluster); trace_backup_do_cow_enter(job, start, sector_num, nb_sectors); @@ -118,12 +123,12 @@ static int coroutine_fn backup_do_cow(BlockDriverState *bs, trace_backup_do_cow_process(job, start); - n = MIN(BACKUP_SECTORS_PER_CLUSTER, + n = MIN(sectors_per_cluster, job->common.len / BDRV_SECTOR_SIZE - - start * BACKUP_SECTORS_PER_CLUSTER); + start * sectors_per_cluster); if (!bounce_buffer) { - bounce_buffer = qemu_blockalign(bs, BACKUP_CLUSTER_SIZE); + bounce_buffer = qemu_blockalign(bs, job->cluster_size); } iov.iov_base = bounce_buffer; iov.iov_len = n * BDRV_SECTOR_SIZE; @@ -131,10 +136,10 @@ static int coroutine_fn backup_do_cow(BlockDriverState *bs, if (is_write_notifier) { ret = bdrv_co_readv_no_serialising(bs, - start * BACKUP_SECTORS_PER_CLUSTER, + start * sectors_per_cluster, n, &bounce_qiov); } else { - ret = bdrv_co_readv(bs, start * BACKUP_SECTORS_PER_CLUSTER, n, + ret = bdrv_co_readv(bs, start * sectors_per_cluster, n, &bounce_qiov); } if (ret < 0) { @@ -147,11 +152,11 @@ static int coroutine_fn backup_do_cow(BlockDriverState *bs, if (buffer_is_zero(iov.iov_base, iov.iov_len)) { ret = bdrv_co_write_zeroes(job->target, - start * BACKUP_SECTORS_PER_CLUSTER, + start * sectors_per_cluster, n, BDRV_REQ_MAY_UNMAP); } else { ret = bdrv_co_writev(job->target, - start * BACKUP_SECTORS_PER_CLUSTER, n, + start * sectors_per_cluster, n, &bounce_qiov); } if (ret < 0) { @@ -322,21 +327,22 @@ static int coroutine_fn backup_run_incremental(BackupBlockJob *job) int64_t cluster; int64_t end; int64_t last_cluster = -1; + int64_t sectors_per_cluster = cluster_size_sectors(job); BlockDriverState *bs = job->common.bs; HBitmapIter hbi; granularity = bdrv_dirty_bitmap_granularity(job->sync_bitmap); - clusters_per_iter = MAX((granularity / BACKUP_CLUSTER_SIZE), 1); + clusters_per_iter = MAX((granularity / job->cluster_size), 1); bdrv_dirty_iter_init(job->sync_bitmap, &hbi); /* Find the next dirty sector(s) */ while ((sector = hbitmap_iter_next(&hbi)) != -1) { - cluster = sector / BACKUP_SECTORS_PER_CLUSTER; + cluster = sector / sectors_per_cluster; /* Fake progress updates for any clusters we skipped */ if (cluster != last_cluster + 1) { job->common.offset += ((cluster - last_cluster - 1) * - BACKUP_CLUSTER_SIZE); + job->cluster_size); } for (end = cluster + clusters_per_iter; cluster < end; cluster++) { @@ -344,8 +350,8 @@ static int coroutine_fn backup_run_incremental(BackupBlockJob *job) if (yield_and_check(job)) { return ret; } - ret = backup_do_cow(bs, cluster * BACKUP_SECTORS_PER_CLUSTER, - BACKUP_SECTORS_PER_CLUSTER, &error_is_read, + ret = backup_do_cow(bs, cluster * sectors_per_cluster, + sectors_per_cluster, &error_is_read, false); if ((ret < 0) && backup_error_action(job, error_is_read, -ret) == @@ -357,17 +363,17 @@ static int coroutine_fn backup_run_incremental(BackupBlockJob *job) /* If the bitmap granularity is smaller than the backup granularity, * we need to advance the iterator pointer to the next cluster. */ - if (granularity < BACKUP_CLUSTER_SIZE) { - bdrv_set_dirty_iter(&hbi, cluster * BACKUP_SECTORS_PER_CLUSTER); + if (granularity < job->cluster_size) { + bdrv_set_dirty_iter(&hbi, cluster * sectors_per_cluster); } last_cluster = cluster - 1; } /* Play some final catchup with the progress meter */ - end = DIV_ROUND_UP(job->common.len, BACKUP_CLUSTER_SIZE); + end = DIV_ROUND_UP(job->common.len, job->cluster_size); if (last_cluster + 1 < end) { - job->common.offset += ((end - last_cluster - 1) * BACKUP_CLUSTER_SIZE); + job->common.offset += ((end - last_cluster - 1) * job->cluster_size); } return ret; @@ -384,13 +390,14 @@ static void coroutine_fn backup_run(void *opaque) .notify = backup_before_write_notify, }; int64_t start, end; + int64_t sectors_per_cluster = cluster_size_sectors(job); int ret = 0; QLIST_INIT(&job->inflight_reqs); qemu_co_rwlock_init(&job->flush_rwlock); start = 0; - end = DIV_ROUND_UP(job->common.len, BACKUP_CLUSTER_SIZE); + end = DIV_ROUND_UP(job->common.len, job->cluster_size); job->bitmap = hbitmap_alloc(end, 0); @@ -427,7 +434,7 @@ static void coroutine_fn backup_run(void *opaque) /* Check to see if these blocks are already in the * backing file. */ - for (i = 0; i < BACKUP_SECTORS_PER_CLUSTER;) { + for (i = 0; i < sectors_per_cluster;) { /* bdrv_is_allocated() only returns true/false based * on the first set of sectors it comes across that * are are all in the same state. @@ -436,8 +443,8 @@ static void coroutine_fn backup_run(void *opaque) * needed but at some point that is always the case. */ alloced = bdrv_is_allocated(bs, - start * BACKUP_SECTORS_PER_CLUSTER + i, - BACKUP_SECTORS_PER_CLUSTER - i, &n); + start * sectors_per_cluster + i, + sectors_per_cluster - i, &n); i += n; if (alloced == 1 || n == 0) { @@ -452,8 +459,8 @@ static void coroutine_fn backup_run(void *opaque) } } /* FULL sync mode we copy the whole drive. */ - ret = backup_do_cow(bs, start * BACKUP_SECTORS_PER_CLUSTER, - BACKUP_SECTORS_PER_CLUSTER, &error_is_read, false); + ret = backup_do_cow(bs, start * sectors_per_cluster, + sectors_per_cluster, &error_is_read, false); if (ret < 0) { /* Depending on error action, fail now or retry cluster */ BlockErrorAction action = @@ -571,6 +578,7 @@ void backup_start(BlockDriverState *bs, BlockDriverState *target, job->sync_mode = sync_mode; job->sync_bitmap = sync_mode == MIRROR_SYNC_MODE_INCREMENTAL ? sync_bitmap : NULL; + job->cluster_size = BACKUP_CLUSTER_SIZE_DEFAULT; job->common.len = len; job->common.co = qemu_coroutine_create(backup_run); block_job_txn_add_job(txn, &job->common);