From patchwork Fri Jan 25 18:45:45 2013 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Kevin Wolf X-Patchwork-Id: 215856 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Received: from lists.gnu.org (lists.gnu.org [208.118.235.17]) (using TLSv1 with cipher AES256-SHA (256/256 bits)) (Client did not present a certificate) by ozlabs.org (Postfix) with ESMTPS id 3E2CF2C008E for ; Sat, 26 Jan 2013 07:23:47 +1100 (EST) Received: from localhost ([::1]:33376 helo=lists.gnu.org) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1TyoJE-0001OE-FO for incoming@patchwork.ozlabs.org; Fri, 25 Jan 2013 13:47:28 -0500 Received: from eggs.gnu.org ([208.118.235.92]:36466) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1TyoI9-0007Sj-TJ for qemu-devel@nongnu.org; Fri, 25 Jan 2013 13:46:24 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1TyoI7-0007lC-Kc for qemu-devel@nongnu.org; Fri, 25 Jan 2013 13:46:21 -0500 Received: from mx1.redhat.com ([209.132.183.28]:38322) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1TyoI7-0007kx-Cl for qemu-devel@nongnu.org; Fri, 25 Jan 2013 13:46:19 -0500 Received: from int-mx09.intmail.prod.int.phx2.redhat.com (int-mx09.intmail.prod.int.phx2.redhat.com [10.5.11.22]) by mx1.redhat.com (8.14.4/8.14.4) with ESMTP id r0PIkH9h008914 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Fri, 25 Jan 2013 13:46:17 -0500 Received: from dhcp-5-188.str.redhat.com (vpn1-6-44.ams2.redhat.com [10.36.6.44]) by int-mx09.intmail.prod.int.phx2.redhat.com (8.14.4/8.14.4) with ESMTP id r0PIk1dc018934; Fri, 25 Jan 2013 13:46:16 -0500 From: Kevin Wolf To: anthony@codemonkey.ws Date: Fri, 25 Jan 2013 19:45:45 +0100 Message-Id: <1359139560-15387-10-git-send-email-kwolf@redhat.com> In-Reply-To: <1359139560-15387-1-git-send-email-kwolf@redhat.com> References: <1359139560-15387-1-git-send-email-kwolf@redhat.com> X-Scanned-By: MIMEDefang 2.68 on 10.5.11.22 X-detected-operating-system: by eggs.gnu.org: GNU/Linux 3.x X-Received-From: 209.132.183.28 Cc: kwolf@redhat.com, qemu-devel@nongnu.org Subject: [Qemu-devel] [PATCH 09/24] mirror: switch mirror_iteration to AIO 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 From: Paolo Bonzini There is really no change in the behavior of the job here, since there is still a maximum of one in-flight I/O operation between the source and the target. However, this patch already introduces the AIO callbacks (which are unmodified in the next patch) and some of the logic to count in-flight operations and only complete the job when there is none. Signed-off-by: Paolo Bonzini Signed-off-by: Kevin Wolf --- block/mirror.c | 167 ++++++++++++++++++++++++++++++++++++++++++-------------- trace-events | 2 + 2 files changed, 128 insertions(+), 41 deletions(-) diff --git a/block/mirror.c b/block/mirror.c index 0fecb40..fc6b9b7 100644 --- a/block/mirror.c +++ b/block/mirror.c @@ -33,8 +33,19 @@ typedef struct MirrorBlockJob { unsigned long *cow_bitmap; HBitmapIter hbi; uint8_t *buf; + + int in_flight; + int ret; } MirrorBlockJob; +typedef struct MirrorOp { + MirrorBlockJob *s; + QEMUIOVector qiov; + struct iovec iov; + int64_t sector_num; + int nb_sectors; +} MirrorOp; + static BlockErrorAction mirror_error_action(MirrorBlockJob *s, bool read, int error) { @@ -48,15 +59,70 @@ static BlockErrorAction mirror_error_action(MirrorBlockJob *s, bool read, } } -static int coroutine_fn mirror_iteration(MirrorBlockJob *s, - BlockErrorAction *p_action) +static void mirror_iteration_done(MirrorOp *op, int ret) +{ + MirrorBlockJob *s = op->s; + int64_t chunk_num; + int nb_chunks, sectors_per_chunk; + + trace_mirror_iteration_done(s, op->sector_num, op->nb_sectors, ret); + + s->in_flight--; + sectors_per_chunk = s->granularity >> BDRV_SECTOR_BITS; + chunk_num = op->sector_num / sectors_per_chunk; + nb_chunks = op->nb_sectors / sectors_per_chunk; + if (s->cow_bitmap && ret >= 0) { + bitmap_set(s->cow_bitmap, chunk_num, nb_chunks); + } + + g_slice_free(MirrorOp, op); + qemu_coroutine_enter(s->common.co, NULL); +} + +static void mirror_write_complete(void *opaque, int ret) +{ + MirrorOp *op = opaque; + MirrorBlockJob *s = op->s; + if (ret < 0) { + BlockDriverState *source = s->common.bs; + BlockErrorAction action; + + bdrv_set_dirty(source, op->sector_num, op->nb_sectors); + action = mirror_error_action(s, false, -ret); + if (action == BDRV_ACTION_REPORT && s->ret >= 0) { + s->ret = ret; + } + } + mirror_iteration_done(op, ret); +} + +static void mirror_read_complete(void *opaque, int ret) +{ + MirrorOp *op = opaque; + MirrorBlockJob *s = op->s; + if (ret < 0) { + BlockDriverState *source = s->common.bs; + BlockErrorAction action; + + bdrv_set_dirty(source, op->sector_num, op->nb_sectors); + action = mirror_error_action(s, true, -ret); + if (action == BDRV_ACTION_REPORT && s->ret >= 0) { + s->ret = ret; + } + + mirror_iteration_done(op, ret); + return; + } + bdrv_aio_writev(s->target, op->sector_num, &op->qiov, op->nb_sectors, + mirror_write_complete, op); +} + +static void coroutine_fn mirror_iteration(MirrorBlockJob *s) { BlockDriverState *source = s->common.bs; - BlockDriverState *target = s->target; - QEMUIOVector qiov; - int ret, nb_sectors, sectors_per_chunk; + int nb_sectors, sectors_per_chunk; int64_t end, sector_num, chunk_num; - struct iovec iov; + MirrorOp *op; s->sector_num = hbitmap_iter_next(&s->hbi); if (s->sector_num < 0) { @@ -85,35 +151,30 @@ static int coroutine_fn mirror_iteration(MirrorBlockJob *s, end = s->common.len >> BDRV_SECTOR_BITS; nb_sectors = MIN(nb_sectors, end - sector_num); + + /* Allocate a MirrorOp that is used as an AIO callback. */ + op = g_slice_new(MirrorOp); + op->s = s; + op->iov.iov_base = s->buf; + op->iov.iov_len = nb_sectors * 512; + op->sector_num = sector_num; + op->nb_sectors = nb_sectors; + qemu_iovec_init_external(&op->qiov, &op->iov, 1); + bdrv_reset_dirty(source, sector_num, nb_sectors); /* Copy the dirty cluster. */ - iov.iov_base = s->buf; - iov.iov_len = nb_sectors * 512; - qemu_iovec_init_external(&qiov, &iov, 1); - + s->in_flight++; trace_mirror_one_iteration(s, sector_num, nb_sectors); - ret = bdrv_co_readv(source, sector_num, nb_sectors, &qiov); - if (ret < 0) { - *p_action = mirror_error_action(s, true, -ret); - goto fail; - } - ret = bdrv_co_writev(target, sector_num, nb_sectors, &qiov); - if (ret < 0) { - *p_action = mirror_error_action(s, false, -ret); - s->synced = false; - goto fail; - } - if (s->cow_bitmap) { - bitmap_set(s->cow_bitmap, sector_num / sectors_per_chunk, - nb_sectors / sectors_per_chunk); - } - return 0; + bdrv_aio_readv(source, sector_num, &op->qiov, nb_sectors, + mirror_read_complete, op); +} -fail: - /* Try again later. */ - bdrv_set_dirty(source, sector_num, nb_sectors); - return ret; +static void mirror_drain(MirrorBlockJob *s) +{ + while (s->in_flight > 0) { + qemu_coroutine_yield(); + } } static void coroutine_fn mirror_run(void *opaque) @@ -121,6 +182,7 @@ static void coroutine_fn mirror_run(void *opaque) MirrorBlockJob *s = opaque; BlockDriverState *bs = s->common.bs; int64_t sector_num, end, sectors_per_chunk, length; + uint64_t last_pause_ns; BlockDriverInfo bdi; char backing_filename[1024]; int ret = 0; @@ -179,23 +241,38 @@ static void coroutine_fn mirror_run(void *opaque) } bdrv_dirty_iter_init(bs, &s->hbi); + last_pause_ns = qemu_get_clock_ns(rt_clock); for (;;) { uint64_t delay_ns; int64_t cnt; bool should_complete; + if (s->ret < 0) { + ret = s->ret; + goto immediate_exit; + } + cnt = bdrv_get_dirty_count(bs); - if (cnt != 0) { - BlockErrorAction action = BDRV_ACTION_REPORT; - ret = mirror_iteration(s, &action); - if (ret < 0 && action == BDRV_ACTION_REPORT) { - goto immediate_exit; + + /* Note that even when no rate limit is applied we need to yield + * periodically with no pending I/O so that qemu_aio_flush() returns. + * We do so every SLICE_TIME nanoseconds, or when there is an error, + * or when the source is clean, whichever comes first. + */ + if (qemu_get_clock_ns(rt_clock) - last_pause_ns < SLICE_TIME && + s->common.iostatus == BLOCK_DEVICE_IO_STATUS_OK) { + if (s->in_flight > 0) { + trace_mirror_yield(s, s->in_flight, cnt); + qemu_coroutine_yield(); + continue; + } else if (cnt != 0) { + mirror_iteration(s); + continue; } - cnt = bdrv_get_dirty_count(bs); } should_complete = false; - if (cnt == 0) { + if (s->in_flight == 0 && cnt == 0) { trace_mirror_before_flush(s); ret = bdrv_flush(s->target); if (ret < 0) { @@ -246,15 +323,12 @@ static void coroutine_fn mirror_run(void *opaque) delay_ns = 0; } - /* Note that even when no rate limit is applied we need to yield - * with no pending I/O here so that bdrv_drain_all() returns. - */ block_job_sleep_ns(&s->common, rt_clock, delay_ns); if (block_job_is_cancelled(&s->common)) { break; } } else if (!should_complete) { - delay_ns = (cnt == 0 ? SLICE_TIME : 0); + delay_ns = (s->in_flight == 0 && cnt == 0 ? SLICE_TIME : 0); block_job_sleep_ns(&s->common, rt_clock, delay_ns); } else if (cnt == 0) { /* The two disks are in sync. Exit and report successful @@ -264,9 +338,20 @@ static void coroutine_fn mirror_run(void *opaque) s->common.cancelled = false; break; } + last_pause_ns = qemu_get_clock_ns(rt_clock); } immediate_exit: + if (s->in_flight > 0) { + /* We get here only if something went wrong. Either the job failed, + * or it was cancelled prematurely so that we do not guarantee that + * the target is a copy of the source. + */ + assert(ret < 0 || (!s->synced && block_job_is_cancelled(&s->common))); + mirror_drain(s); + } + + assert(s->in_flight == 0); qemu_vfree(s->buf); g_free(s->cow_bitmap); bdrv_set_dirty_tracking(bs, 0); diff --git a/trace-events b/trace-events index ffa2756..43e6b73 100644 --- a/trace-events +++ b/trace-events @@ -85,6 +85,8 @@ mirror_before_drain(void *s, int64_t cnt) "s %p dirty count %"PRId64 mirror_before_sleep(void *s, int64_t cnt, int synced) "s %p dirty count %"PRId64" synced %d" mirror_one_iteration(void *s, int64_t sector_num, int nb_sectors) "s %p sector_num %"PRId64" nb_sectors %d" mirror_cow(void *s, int64_t sector_num) "s %p sector_num %"PRId64 +mirror_iteration_done(void *s, int64_t sector_num, int nb_sectors, int ret) "s %p sector_num %"PRId64" nb_sectors %d ret %d" +mirror_yield(void *s, int64_t cnt, int in_flight) "s %p dirty count %"PRId64" in_flight %d" # blockdev.c qmp_block_job_cancel(void *job) "job %p"