From f9f65a4af0ea8162d2339a979350f891942e3037 Mon Sep 17 00:00:00 2001 From: Fam Zheng Date: Fri, 5 May 2017 11:23:36 +0800 Subject: [PATCH 01/23] docker: Run tests with current user MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit We've used --add-current-user to create a user in the image, use it to run tests, because root has too much priviledge, and can surprise test cases. Signed-off-by: Fam Zheng Message-Id: <20170505032340.26467-2-famz@redhat.com> Reviewed-by: Philippe Mathieu-Daudé Reviewed-by: Alex Bennée Reviewed-by: Paolo Bonzini Signed-off-by: Fam Zheng --- tests/docker/Makefile.include | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/docker/Makefile.include b/tests/docker/Makefile.include index 03eda37bf4..0ed8c3d323 100644 --- a/tests/docker/Makefile.include +++ b/tests/docker/Makefile.include @@ -126,7 +126,7 @@ docker-run: docker-qemu-src " COPYING $(EXECUTABLE) to $(IMAGE)")) $(call quiet-command, \ $(SRC_PATH)/tests/docker/docker.py run \ - -t \ + $(if $(NOUSER),,-u $(shell id -u)) -t \ $(if $V,,--rm) \ $(if $(DEBUG),-i,--net=none) \ -e TARGET_LIST=$(TARGET_LIST) \ From 73a27bbb698202243d2092d802079dd6a0be7325 Mon Sep 17 00:00:00 2001 From: Fam Zheng Date: Fri, 5 May 2017 11:23:37 +0800 Subject: [PATCH 02/23] docker: Add bzip2 and hostname to fedora image MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit It is used by qemu-iotests. Signed-off-by: Fam Zheng Message-Id: <20170505032340.26467-3-famz@redhat.com> Reviewed-by: Philippe Mathieu-Daudé Reviewed-by: Alex Bennée Reviewed-by: Paolo Bonzini Signed-off-by: Fam Zheng --- tests/docker/dockerfiles/fedora.docker | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/docker/dockerfiles/fedora.docker b/tests/docker/dockerfiles/fedora.docker index c4f80ad3d8..39f6b5852d 100644 --- a/tests/docker/dockerfiles/fedora.docker +++ b/tests/docker/dockerfiles/fedora.docker @@ -1,6 +1,6 @@ FROM fedora:latest ENV PACKAGES \ - ccache git tar PyYAML sparse flex bison python2 \ + ccache git tar PyYAML sparse flex bison python2 bzip2 hostname \ glib2-devel pixman-devel zlib-devel SDL-devel libfdt-devel \ gcc gcc-c++ clang make perl which bc findutils \ mingw32-pixman mingw32-glib2 mingw32-gmp mingw32-SDL mingw32-pkg-config \ From 80c58a5b1b1ff1bc447f64c8aec981d5c6226cec Mon Sep 17 00:00:00 2001 From: Fam Zheng Date: Fri, 5 May 2017 11:23:39 +0800 Subject: [PATCH 03/23] docker: Add libaio to fedora image MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: Fam Zheng Message-Id: <20170505032340.26467-5-famz@redhat.com> Reviewed-by: Philippe Mathieu-Daudé Reviewed-by: Alex Bennée Reviewed-by: Paolo Bonzini Signed-off-by: Fam Zheng --- tests/docker/dockerfiles/fedora.docker | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/docker/dockerfiles/fedora.docker b/tests/docker/dockerfiles/fedora.docker index 39f6b5852d..4eaa8ed2a5 100644 --- a/tests/docker/dockerfiles/fedora.docker +++ b/tests/docker/dockerfiles/fedora.docker @@ -2,7 +2,7 @@ FROM fedora:latest ENV PACKAGES \ ccache git tar PyYAML sparse flex bison python2 bzip2 hostname \ glib2-devel pixman-devel zlib-devel SDL-devel libfdt-devel \ - gcc gcc-c++ clang make perl which bc findutils \ + gcc gcc-c++ clang make perl which bc findutils libaio-devel \ mingw32-pixman mingw32-glib2 mingw32-gmp mingw32-SDL mingw32-pkg-config \ mingw32-gtk2 mingw32-gtk3 mingw32-gnutls mingw32-nettle mingw32-libtasn1 \ mingw32-libjpeg-turbo mingw32-libpng mingw32-curl mingw32-libssh2 \ From 79f24568e5e70892b9bf4712f88b26f52255077f Mon Sep 17 00:00:00 2001 From: Fam Zheng Date: Wed, 24 May 2017 08:52:06 +0800 Subject: [PATCH 04/23] docker: Add flex and bison to centos6 image MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Currently there are warnings about flex and bison being missing when building in the centos6 image: make[1]: flex: Command not found BISON dtc-parser.tab.c make[1]: bison: Command not found Add them. Reported-by: Thomas Huth Signed-off-by: Fam Zheng Message-Id: <20170524005206.31916-1-famz@redhat.com> Reviewed-by: Alex Bennée Reviewed-by: Philippe Mathieu-Daudé Signed-off-by: Fam Zheng --- tests/docker/dockerfiles/centos6.docker | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/docker/dockerfiles/centos6.docker b/tests/docker/dockerfiles/centos6.docker index 34e0d3b91e..17a4d24d54 100644 --- a/tests/docker/dockerfiles/centos6.docker +++ b/tests/docker/dockerfiles/centos6.docker @@ -1,7 +1,7 @@ FROM centos:6 RUN yum install -y epel-release ENV PACKAGES libfdt-devel ccache \ - tar git make gcc g++ \ + tar git make gcc g++ flex bison \ zlib-devel glib2-devel SDL-devel pixman-devel \ epel-release RUN yum install -y $PACKAGES From d3faa13e5ffda779dcd58f39e8745370adb05d67 Mon Sep 17 00:00:00 2001 From: Paolo Bonzini Date: Mon, 5 Jun 2017 14:38:50 +0200 Subject: [PATCH 05/23] block: access copy_on_read with atomic ops Reviewed-by: Stefan Hajnoczi Signed-off-by: Paolo Bonzini Message-Id: <20170605123908.18777-2-pbonzini@redhat.com> Signed-off-by: Fam Zheng --- block.c | 6 ++++-- block/io.c | 8 ++++---- blockdev.c | 2 +- include/block/block_int.h | 11 ++++++----- 4 files changed, 15 insertions(+), 12 deletions(-) diff --git a/block.c b/block.c index fa1d06d846..af6366bcef 100644 --- a/block.c +++ b/block.c @@ -1300,7 +1300,9 @@ static int bdrv_open_common(BlockDriverState *bs, BlockBackend *file, goto fail_opts; } - assert(bs->copy_on_read == 0); /* bdrv_new() and bdrv_close() make it so */ + /* bdrv_new() and bdrv_close() make it so */ + assert(atomic_read(&bs->copy_on_read) == 0); + if (bs->open_flags & BDRV_O_COPY_ON_READ) { if (!bs->read_only) { bdrv_enable_copy_on_read(bs); @@ -3063,7 +3065,7 @@ static void bdrv_close(BlockDriverState *bs) g_free(bs->opaque); bs->opaque = NULL; - bs->copy_on_read = 0; + atomic_set(&bs->copy_on_read, 0); bs->backing_file[0] = '\0'; bs->backing_format[0] = '\0'; bs->total_sectors = 0; diff --git a/block/io.c b/block/io.c index ed31810c0a..98c690f306 100644 --- a/block/io.c +++ b/block/io.c @@ -130,13 +130,13 @@ void bdrv_refresh_limits(BlockDriverState *bs, Error **errp) */ void bdrv_enable_copy_on_read(BlockDriverState *bs) { - bs->copy_on_read++; + atomic_inc(&bs->copy_on_read); } void bdrv_disable_copy_on_read(BlockDriverState *bs) { - assert(bs->copy_on_read > 0); - bs->copy_on_read--; + int old = atomic_fetch_dec(&bs->copy_on_read); + assert(old >= 1); } /* Check if any requests are in-flight (including throttled requests) */ @@ -1144,7 +1144,7 @@ int coroutine_fn bdrv_co_preadv(BdrvChild *child, bdrv_inc_in_flight(bs); /* Don't do copy-on-read if we read data before write operation */ - if (bs->copy_on_read && !(flags & BDRV_REQ_NO_SERIALISING)) { + if (atomic_read(&bs->copy_on_read) && !(flags & BDRV_REQ_NO_SERIALISING)) { flags |= BDRV_REQ_COPY_ON_READ; } diff --git a/blockdev.c b/blockdev.c index 6472548186..ceea976b03 100644 --- a/blockdev.c +++ b/blockdev.c @@ -1791,7 +1791,7 @@ static void external_snapshot_commit(BlkActionState *common) /* We don't need (or want) to use the transactional * bdrv_reopen_multiple() across all the entries at once, because we * don't want to abort all of them if one of them fails the reopen */ - if (!state->old_bs->copy_on_read) { + if (!atomic_read(&state->old_bs->copy_on_read)) { bdrv_reopen(state->old_bs, state->old_bs->open_flags & ~BDRV_O_RDWR, NULL); } diff --git a/include/block/block_int.h b/include/block/block_int.h index cb78c4fa82..49f2ebb95a 100644 --- a/include/block/block_int.h +++ b/include/block/block_int.h @@ -595,11 +595,6 @@ struct BlockDriverState { /* Protected by AioContext lock */ - /* If true, copy read backing sectors into image. Can be >1 if more - * than one client has requested copy-on-read. - */ - int copy_on_read; - /* If we are reading a disk image, give its size in sectors. * Generally read-only; it is written to by load_snapshot and * save_snaphost, but the block layer is quiescent during those. @@ -633,6 +628,12 @@ struct BlockDriverState { QLIST_HEAD(, BdrvDirtyBitmap) dirty_bitmaps; + /* If true, copy read backing sectors into image. Can be >1 if more + * than one client has requested copy-on-read. Accessed with atomic + * ops. + */ + int copy_on_read; + /* do we need to tell the quest if we have a volatile write cache? */ int enable_write_cache; From 414c2ec3585408a5c7b458664db180e91480bb03 Mon Sep 17 00:00:00 2001 From: Paolo Bonzini Date: Mon, 5 Jun 2017 14:38:51 +0200 Subject: [PATCH 06/23] block: access quiesce_counter with atomic ops Reviewed-by: Stefan Hajnoczi Reviewed-by: Alberto Garcia Signed-off-by: Paolo Bonzini Message-Id: <20170605123908.18777-3-pbonzini@redhat.com> Signed-off-by: Fam Zheng --- block/io.c | 4 ++-- include/block/block_int.h | 1 + 2 files changed, 3 insertions(+), 2 deletions(-) diff --git a/block/io.c b/block/io.c index 98c690f306..70643df494 100644 --- a/block/io.c +++ b/block/io.c @@ -241,7 +241,7 @@ void bdrv_drained_begin(BlockDriverState *bs) return; } - if (!bs->quiesce_counter++) { + if (atomic_fetch_inc(&bs->quiesce_counter) == 0) { aio_disable_external(bdrv_get_aio_context(bs)); bdrv_parent_drained_begin(bs); } @@ -252,7 +252,7 @@ void bdrv_drained_begin(BlockDriverState *bs) void bdrv_drained_end(BlockDriverState *bs) { assert(bs->quiesce_counter > 0); - if (--bs->quiesce_counter > 0) { + if (atomic_fetch_dec(&bs->quiesce_counter) > 1) { return; } diff --git a/include/block/block_int.h b/include/block/block_int.h index 49f2ebb95a..1824e0e135 100644 --- a/include/block/block_int.h +++ b/include/block/block_int.h @@ -637,6 +637,7 @@ struct BlockDriverState { /* do we need to tell the quest if we have a volatile write cache? */ int enable_write_cache; + /* Accessed with atomic ops. */ int quiesce_counter; }; From d993b85804b7ec099d4e1d377161ac8af398d855 Mon Sep 17 00:00:00 2001 From: Paolo Bonzini Date: Mon, 5 Jun 2017 14:38:52 +0200 Subject: [PATCH 07/23] block: access io_limits_disabled with atomic ops Reviewed-by: Stefan Hajnoczi Reviewed-by: Alberto Garcia Signed-off-by: Paolo Bonzini Message-Id: <20170605123908.18777-4-pbonzini@redhat.com> Signed-off-by: Fam Zheng --- block/block-backend.c | 4 ++-- block/throttle-groups.c | 2 +- include/sysemu/block-backend.h | 3 ++- 3 files changed, 5 insertions(+), 4 deletions(-) diff --git a/block/block-backend.c b/block/block-backend.c index 7d7f3697d1..69d0e11466 100644 --- a/block/block-backend.c +++ b/block/block-backend.c @@ -1953,7 +1953,7 @@ static void blk_root_drained_begin(BdrvChild *child) /* Note that blk->root may not be accessible here yet if we are just * attaching to a BlockDriverState that is drained. Use child instead. */ - if (blk->public.io_limits_disabled++ == 0) { + if (atomic_fetch_inc(&blk->public.io_limits_disabled) == 0) { throttle_group_restart_blk(blk); } } @@ -1964,7 +1964,7 @@ static void blk_root_drained_end(BdrvChild *child) assert(blk->quiesce_counter); assert(blk->public.io_limits_disabled); - --blk->public.io_limits_disabled; + atomic_dec(&blk->public.io_limits_disabled); if (--blk->quiesce_counter == 0) { if (blk->dev_ops && blk->dev_ops->drained_end) { diff --git a/block/throttle-groups.c b/block/throttle-groups.c index b73e7a800b..69bfbd44d9 100644 --- a/block/throttle-groups.c +++ b/block/throttle-groups.c @@ -240,7 +240,7 @@ static bool throttle_group_schedule_timer(BlockBackend *blk, bool is_write) ThrottleGroup *tg = container_of(ts, ThrottleGroup, ts); bool must_wait; - if (blkp->io_limits_disabled) { + if (atomic_read(&blkp->io_limits_disabled)) { return false; } diff --git a/include/sysemu/block-backend.h b/include/sysemu/block-backend.h index 840ad6134c..24b63d6b03 100644 --- a/include/sysemu/block-backend.h +++ b/include/sysemu/block-backend.h @@ -80,7 +80,8 @@ typedef struct BlockBackendPublic { CoQueue throttled_reqs[2]; /* Nonzero if the I/O limits are currently being ignored; generally - * it is zero. */ + * it is zero. Accessed with atomic operations. + */ unsigned int io_limits_disabled; /* The following fields are protected by the ThrottleGroup lock. From 20fc71b25cfd8102b7f12a2b44133894ad90040a Mon Sep 17 00:00:00 2001 From: Paolo Bonzini Date: Mon, 5 Jun 2017 14:38:53 +0200 Subject: [PATCH 08/23] block: access serialising_in_flight with atomic ops Reviewed-by: Stefan Hajnoczi Signed-off-by: Paolo Bonzini Message-Id: <20170605123908.18777-5-pbonzini@redhat.com> Signed-off-by: Fam Zheng --- block/io.c | 6 +++--- include/block/block_int.h | 10 ++++++---- 2 files changed, 9 insertions(+), 7 deletions(-) diff --git a/block/io.c b/block/io.c index 70643df494..d76202b452 100644 --- a/block/io.c +++ b/block/io.c @@ -375,7 +375,7 @@ void bdrv_drain_all(void) static void tracked_request_end(BdrvTrackedRequest *req) { if (req->serialising) { - req->bs->serialising_in_flight--; + atomic_dec(&req->bs->serialising_in_flight); } QLIST_REMOVE(req, list); @@ -414,7 +414,7 @@ static void mark_request_serialising(BdrvTrackedRequest *req, uint64_t align) - overlap_offset; if (!req->serialising) { - req->bs->serialising_in_flight++; + atomic_inc(&req->bs->serialising_in_flight); req->serialising = true; } @@ -519,7 +519,7 @@ static bool coroutine_fn wait_serialising_requests(BdrvTrackedRequest *self) bool retry; bool waited = false; - if (!bs->serialising_in_flight) { + if (!atomic_read(&bs->serialising_in_flight)) { return false; } diff --git a/include/block/block_int.h b/include/block/block_int.h index 1824e0e135..39be34af5c 100644 --- a/include/block/block_int.h +++ b/include/block/block_int.h @@ -604,10 +604,6 @@ struct BlockDriverState { /* Callback before write request is processed */ NotifierWithReturnList before_write_notifiers; - /* number of in-flight requests; overall and serialising */ - unsigned int in_flight; - unsigned int serialising_in_flight; - bool wakeup; /* Offset after the highest byte written to */ @@ -634,6 +630,12 @@ struct BlockDriverState { */ int copy_on_read; + /* number of in-flight requests; overall and serialising. + * Accessed with atomic ops. + */ + unsigned int in_flight; + unsigned int serialising_in_flight; + /* do we need to tell the quest if we have a volatile write cache? */ int enable_write_cache; From e2a6ae7fe57c17199624e4d47826ec46ca57d546 Mon Sep 17 00:00:00 2001 From: Paolo Bonzini Date: Mon, 5 Jun 2017 14:38:54 +0200 Subject: [PATCH 09/23] block: access wakeup with atomic ops Reviewed-by: Stefan Hajnoczi Signed-off-by: Paolo Bonzini Message-Id: <20170605123908.18777-6-pbonzini@redhat.com> Signed-off-by: Fam Zheng --- block/io.c | 3 ++- block/nfs.c | 4 +++- block/sheepdog.c | 3 ++- include/block/block.h | 5 +++-- include/block/block_int.h | 7 +++++-- 5 files changed, 15 insertions(+), 7 deletions(-) diff --git a/block/io.c b/block/io.c index d76202b452..4a598299a9 100644 --- a/block/io.c +++ b/block/io.c @@ -501,7 +501,8 @@ static void dummy_bh_cb(void *opaque) void bdrv_wakeup(BlockDriverState *bs) { - if (bs->wakeup) { + /* The barrier (or an atomic op) is in the caller. */ + if (atomic_read(&bs->wakeup)) { aio_bh_schedule_oneshot(qemu_get_aio_context(), dummy_bh_cb, NULL); } } diff --git a/block/nfs.c b/block/nfs.c index 848b2c0bb0..18c87d2f25 100644 --- a/block/nfs.c +++ b/block/nfs.c @@ -730,7 +730,9 @@ nfs_get_allocated_file_size_cb(int ret, struct nfs_context *nfs, void *data, if (task->ret < 0) { error_report("NFS Error: %s", nfs_get_error(nfs)); } - task->complete = 1; + + /* Set task->complete before reading bs->wakeup. */ + atomic_mb_set(&task->complete, 1); bdrv_wakeup(task->bs); } diff --git a/block/sheepdog.c b/block/sheepdog.c index a18315a1ca..5ebf5d9fbb 100644 --- a/block/sheepdog.c +++ b/block/sheepdog.c @@ -698,7 +698,8 @@ out: srco->co = NULL; srco->ret = ret; - srco->finished = true; + /* Set srco->finished before reading bs->wakeup. */ + atomic_mb_set(&srco->finished, true); if (srco->bs) { bdrv_wakeup(srco->bs); } diff --git a/include/block/block.h b/include/block/block.h index 9b355e92d8..a4f09df95a 100644 --- a/include/block/block.h +++ b/include/block/block.h @@ -402,7 +402,8 @@ void bdrv_drain_all(void); * block_job_defer_to_main_loop for how to do it). \ */ \ assert(!bs_->wakeup); \ - bs_->wakeup = true; \ + /* Set bs->wakeup before evaluating cond. */ \ + atomic_mb_set(&bs_->wakeup, true); \ while (busy_) { \ if ((cond)) { \ waited_ = busy_ = true; \ @@ -414,7 +415,7 @@ void bdrv_drain_all(void); waited_ |= busy_; \ } \ } \ - bs_->wakeup = false; \ + atomic_set(&bs_->wakeup, false); \ } \ waited_; }) diff --git a/include/block/block_int.h b/include/block/block_int.h index 39be34af5c..cf544b7f31 100644 --- a/include/block/block_int.h +++ b/include/block/block_int.h @@ -604,8 +604,6 @@ struct BlockDriverState { /* Callback before write request is processed */ NotifierWithReturnList before_write_notifiers; - bool wakeup; - /* Offset after the highest byte written to */ uint64_t wr_highest_offset; @@ -636,6 +634,11 @@ struct BlockDriverState { unsigned int in_flight; unsigned int serialising_in_flight; + /* Internal to BDRV_POLL_WHILE and bdrv_wakeup. Accessed with atomic + * ops. + */ + bool wakeup; + /* do we need to tell the quest if we have a volatile write cache? */ int enable_write_cache; From 850d54a2a9cf16f589f4aa40272515294671633f Mon Sep 17 00:00:00 2001 From: Paolo Bonzini Date: Mon, 5 Jun 2017 14:38:55 +0200 Subject: [PATCH 10/23] block: access io_plugged with atomic ops Reviewed-by: Stefan Hajnoczi Signed-off-by: Paolo Bonzini Message-Id: <20170605123908.18777-7-pbonzini@redhat.com> Signed-off-by: Fam Zheng --- block/io.c | 4 ++-- include/block/block_int.h | 8 +++++--- 2 files changed, 7 insertions(+), 5 deletions(-) diff --git a/block/io.c b/block/io.c index 4a598299a9..bb1c9c5432 100644 --- a/block/io.c +++ b/block/io.c @@ -2645,7 +2645,7 @@ void bdrv_io_plug(BlockDriverState *bs) bdrv_io_plug(child->bs); } - if (bs->io_plugged++ == 0) { + if (atomic_fetch_inc(&bs->io_plugged) == 0) { BlockDriver *drv = bs->drv; if (drv && drv->bdrv_io_plug) { drv->bdrv_io_plug(bs); @@ -2658,7 +2658,7 @@ void bdrv_io_unplug(BlockDriverState *bs) BdrvChild *child; assert(bs->io_plugged); - if (--bs->io_plugged == 0) { + if (atomic_fetch_dec(&bs->io_plugged) == 1) { BlockDriver *drv = bs->drv; if (drv && drv->bdrv_io_unplug) { drv->bdrv_io_unplug(bs); diff --git a/include/block/block_int.h b/include/block/block_int.h index cf544b7f31..091b5d361c 100644 --- a/include/block/block_int.h +++ b/include/block/block_int.h @@ -611,9 +611,6 @@ struct BlockDriverState { uint64_t write_threshold_offset; NotifierWithReturn write_threshold_notifier; - /* counter for nested bdrv_io_plug */ - unsigned io_plugged; - QLIST_HEAD(, BdrvTrackedRequest) tracked_requests; CoQueue flush_queue; /* Serializing flush queue */ bool active_flush_req; /* Flush request in flight? */ @@ -639,6 +636,11 @@ struct BlockDriverState { */ bool wakeup; + /* counter for nested bdrv_io_plug. + * Accessed with atomic ops. + */ + unsigned io_plugged; + /* do we need to tell the quest if we have a volatile write cache? */ int enable_write_cache; From 7258ed930c78e7b85a064c7a0aff5d415efb8fbb Mon Sep 17 00:00:00 2001 From: Paolo Bonzini Date: Mon, 5 Jun 2017 14:38:56 +0200 Subject: [PATCH 11/23] throttle-groups: only start one coroutine from drained_begin Starting all waiting coroutines from bdrv_drain_all is unnecessary; throttle_group_co_io_limits_intercept calls schedule_next_request as soon as the coroutine restarts, which in turn will restart the next request if possible. If we only start the first request and let the coroutines dance from there the code is simpler and there is more reuse between throttle_group_config, throttle_group_restart_blk and timer_cb. The next patch will benefit from this. We also stop accessing from throttle_group_restart_blk the blkp->throttled_reqs CoQueues even when there was no attached throttling group. This worked but is not pretty. The only thing that can interrupt the dance is the QEMU_CLOCK_VIRTUAL timer when switching from one block device to the next, because the timer is set to "now + 1" but QEMU_CLOCK_VIRTUAL might not be running. Set that timer to point in the present ("now") rather than the future and things work. Reviewed-by: Alberto Garcia Reviewed-by: Stefan Hajnoczi Signed-off-by: Paolo Bonzini Message-Id: <20170605123908.18777-8-pbonzini@redhat.com> Signed-off-by: Fam Zheng --- block/throttle-groups.c | 45 +++++++++++++++++++++++------------------ 1 file changed, 25 insertions(+), 20 deletions(-) diff --git a/block/throttle-groups.c b/block/throttle-groups.c index 69bfbd44d9..85169ecfb0 100644 --- a/block/throttle-groups.c +++ b/block/throttle-groups.c @@ -292,7 +292,7 @@ static void schedule_next_request(BlockBackend *blk, bool is_write) } else { ThrottleTimers *tt = &blk_get_public(token)->throttle_timers; int64_t now = qemu_clock_get_ns(tt->clock_type); - timer_mod(tt->timers[is_write], now + 1); + timer_mod(tt->timers[is_write], now); tg->any_timer_armed[is_write] = true; } tg->tokens[is_write] = token; @@ -340,15 +340,32 @@ void coroutine_fn throttle_group_co_io_limits_intercept(BlockBackend *blk, qemu_mutex_unlock(&tg->lock); } +static void throttle_group_restart_queue(BlockBackend *blk, bool is_write) +{ + BlockBackendPublic *blkp = blk_get_public(blk); + ThrottleGroup *tg = container_of(blkp->throttle_state, ThrottleGroup, ts); + bool empty_queue; + + aio_context_acquire(blk_get_aio_context(blk)); + empty_queue = !qemu_co_enter_next(&blkp->throttled_reqs[is_write]); + aio_context_release(blk_get_aio_context(blk)); + + /* If the request queue was empty then we have to take care of + * scheduling the next one */ + if (empty_queue) { + qemu_mutex_lock(&tg->lock); + schedule_next_request(blk, is_write); + qemu_mutex_unlock(&tg->lock); + } +} + void throttle_group_restart_blk(BlockBackend *blk) { BlockBackendPublic *blkp = blk_get_public(blk); - int i; - for (i = 0; i < 2; i++) { - while (qemu_co_enter_next(&blkp->throttled_reqs[i])) { - ; - } + if (blkp->throttle_state) { + throttle_group_restart_queue(blk, 0); + throttle_group_restart_queue(blk, 1); } } @@ -376,8 +393,7 @@ void throttle_group_config(BlockBackend *blk, ThrottleConfig *cfg) throttle_config(ts, tt, cfg); qemu_mutex_unlock(&tg->lock); - qemu_co_enter_next(&blkp->throttled_reqs[0]); - qemu_co_enter_next(&blkp->throttled_reqs[1]); + throttle_group_restart_blk(blk); } /* Get the throttle configuration from a particular group. Similar to @@ -408,7 +424,6 @@ static void timer_cb(BlockBackend *blk, bool is_write) BlockBackendPublic *blkp = blk_get_public(blk); ThrottleState *ts = blkp->throttle_state; ThrottleGroup *tg = container_of(ts, ThrottleGroup, ts); - bool empty_queue; /* The timer has just been fired, so we can update the flag */ qemu_mutex_lock(&tg->lock); @@ -416,17 +431,7 @@ static void timer_cb(BlockBackend *blk, bool is_write) qemu_mutex_unlock(&tg->lock); /* Run the request that was waiting for this timer */ - aio_context_acquire(blk_get_aio_context(blk)); - empty_queue = !qemu_co_enter_next(&blkp->throttled_reqs[is_write]); - aio_context_release(blk_get_aio_context(blk)); - - /* If the request queue was empty then we have to take care of - * scheduling the next one */ - if (empty_queue) { - qemu_mutex_lock(&tg->lock); - schedule_next_request(blk, is_write); - qemu_mutex_unlock(&tg->lock); - } + throttle_group_restart_queue(blk, is_write); } static void read_timer_cb(void *opaque) From 3b170dc86732539ebc5927880097d32d4f11df8c Mon Sep 17 00:00:00 2001 From: Paolo Bonzini Date: Mon, 5 Jun 2017 14:38:57 +0200 Subject: [PATCH 12/23] throttle-groups: do not use qemu_co_enter_next Prepare for removing this function; always restart throttled requests from coroutine context. This will matter when restarting throttled requests will have to acquire a CoMutex. Reviewed-by: Alberto Garcia Reviewed-by: Stefan Hajnoczi Signed-off-by: Paolo Bonzini Message-Id: <20170605123908.18777-9-pbonzini@redhat.com> Signed-off-by: Fam Zheng --- block/throttle-groups.c | 42 ++++++++++++++++++++++++++++++++++++----- 1 file changed, 37 insertions(+), 5 deletions(-) diff --git a/block/throttle-groups.c b/block/throttle-groups.c index 85169ecfb0..8bf1031efa 100644 --- a/block/throttle-groups.c +++ b/block/throttle-groups.c @@ -260,6 +260,20 @@ static bool throttle_group_schedule_timer(BlockBackend *blk, bool is_write) return must_wait; } +/* Start the next pending I/O request for a BlockBackend. Return whether + * any request was actually pending. + * + * @blk: the current BlockBackend + * @is_write: the type of operation (read/write) + */ +static bool coroutine_fn throttle_group_co_restart_queue(BlockBackend *blk, + bool is_write) +{ + BlockBackendPublic *blkp = blk_get_public(blk); + + return qemu_co_queue_next(&blkp->throttled_reqs[is_write]); +} + /* Look for the next pending I/O request and schedule it. * * This assumes that tg->lock is held. @@ -287,7 +301,7 @@ static void schedule_next_request(BlockBackend *blk, bool is_write) if (!must_wait) { /* Give preference to requests from the current blk */ if (qemu_in_coroutine() && - qemu_co_queue_next(&blkp->throttled_reqs[is_write])) { + throttle_group_co_restart_queue(blk, is_write)) { token = blk; } else { ThrottleTimers *tt = &blk_get_public(token)->throttle_timers; @@ -340,15 +354,21 @@ void coroutine_fn throttle_group_co_io_limits_intercept(BlockBackend *blk, qemu_mutex_unlock(&tg->lock); } -static void throttle_group_restart_queue(BlockBackend *blk, bool is_write) +typedef struct { + BlockBackend *blk; + bool is_write; +} RestartData; + +static void coroutine_fn throttle_group_restart_queue_entry(void *opaque) { + RestartData *data = opaque; + BlockBackend *blk = data->blk; + bool is_write = data->is_write; BlockBackendPublic *blkp = blk_get_public(blk); ThrottleGroup *tg = container_of(blkp->throttle_state, ThrottleGroup, ts); bool empty_queue; - aio_context_acquire(blk_get_aio_context(blk)); - empty_queue = !qemu_co_enter_next(&blkp->throttled_reqs[is_write]); - aio_context_release(blk_get_aio_context(blk)); + empty_queue = !throttle_group_co_restart_queue(blk, is_write); /* If the request queue was empty then we have to take care of * scheduling the next one */ @@ -359,6 +379,18 @@ static void throttle_group_restart_queue(BlockBackend *blk, bool is_write) } } +static void throttle_group_restart_queue(BlockBackend *blk, bool is_write) +{ + Coroutine *co; + RestartData rd = { + .blk = blk, + .is_write = is_write + }; + + co = qemu_coroutine_create(throttle_group_restart_queue_entry, &rd); + aio_co_enter(blk_get_aio_context(blk), co); +} + void throttle_group_restart_blk(BlockBackend *blk) { BlockBackendPublic *blkp = blk_get_public(blk); From 93001e9d87ae40e7569eb0ece0eb86d9c9582e41 Mon Sep 17 00:00:00 2001 From: Paolo Bonzini Date: Mon, 5 Jun 2017 14:38:58 +0200 Subject: [PATCH 13/23] throttle-groups: protect throttled requests with a CoMutex Another possibility is to use tg->lock, which we're holding anyway in both schedule_next_request and throttle_group_co_io_limits_intercept. This would require open-coding the CoQueue however, so I've chosen this alternative. Reviewed-by: Stefan Hajnoczi Signed-off-by: Paolo Bonzini Message-Id: <20170605123908.18777-10-pbonzini@redhat.com> Signed-off-by: Fam Zheng --- block/block-backend.c | 1 + block/throttle-groups.c | 12 ++++++++++-- include/sysemu/block-backend.h | 7 ++----- 3 files changed, 13 insertions(+), 7 deletions(-) diff --git a/block/block-backend.c b/block/block-backend.c index 69d0e11466..738882dd2e 100644 --- a/block/block-backend.c +++ b/block/block-backend.c @@ -216,6 +216,7 @@ BlockBackend *blk_new(uint64_t perm, uint64_t shared_perm) blk->shared_perm = shared_perm; blk_set_enable_write_cache(blk, true); + qemu_co_mutex_init(&blk->public.throttled_reqs_lock); qemu_co_queue_init(&blk->public.throttled_reqs[0]); qemu_co_queue_init(&blk->public.throttled_reqs[1]); diff --git a/block/throttle-groups.c b/block/throttle-groups.c index 8bf1031efa..a181cb1dee 100644 --- a/block/throttle-groups.c +++ b/block/throttle-groups.c @@ -270,8 +270,13 @@ static bool coroutine_fn throttle_group_co_restart_queue(BlockBackend *blk, bool is_write) { BlockBackendPublic *blkp = blk_get_public(blk); + bool ret; - return qemu_co_queue_next(&blkp->throttled_reqs[is_write]); + qemu_co_mutex_lock(&blkp->throttled_reqs_lock); + ret = qemu_co_queue_next(&blkp->throttled_reqs[is_write]); + qemu_co_mutex_unlock(&blkp->throttled_reqs_lock); + + return ret; } /* Look for the next pending I/O request and schedule it. @@ -340,7 +345,10 @@ void coroutine_fn throttle_group_co_io_limits_intercept(BlockBackend *blk, if (must_wait || blkp->pending_reqs[is_write]) { blkp->pending_reqs[is_write]++; qemu_mutex_unlock(&tg->lock); - qemu_co_queue_wait(&blkp->throttled_reqs[is_write], NULL); + qemu_co_mutex_lock(&blkp->throttled_reqs_lock); + qemu_co_queue_wait(&blkp->throttled_reqs[is_write], + &blkp->throttled_reqs_lock); + qemu_co_mutex_unlock(&blkp->throttled_reqs_lock); qemu_mutex_lock(&tg->lock); blkp->pending_reqs[is_write]--; } diff --git a/include/sysemu/block-backend.h b/include/sysemu/block-backend.h index 24b63d6b03..999eb2333a 100644 --- a/include/sysemu/block-backend.h +++ b/include/sysemu/block-backend.h @@ -72,11 +72,8 @@ typedef struct BlockDevOps { * fields that must be public. This is in particular for QLIST_ENTRY() and * friends so that BlockBackends can be kept in lists outside block-backend.c */ typedef struct BlockBackendPublic { - /* I/O throttling has its own locking, but also some fields are - * protected by the AioContext lock. - */ - - /* Protected by AioContext lock. */ + /* throttled_reqs_lock protects the CoQueues for throttled requests. */ + CoMutex throttled_reqs_lock; CoQueue throttled_reqs[2]; /* Nonzero if the I/O limits are currently being ignored; generally From ae2d489c341ee98f2daad819a3812c6199481848 Mon Sep 17 00:00:00 2001 From: Paolo Bonzini Date: Mon, 5 Jun 2017 14:38:59 +0200 Subject: [PATCH 14/23] util: add stats64 module This module provides fast paths for 64-bit atomic operations on machines that only have 32-bit atomic access. Reviewed-by: Stefan Hajnoczi Signed-off-by: Paolo Bonzini Signed-off-by: Fam Zheng Message-Id: <20170605123908.18777-11-pbonzini@redhat.com> Signed-off-by: Fam Zheng --- include/qemu/stats64.h | 193 +++++++++++++++++++++++++++++++++++++++++ util/Makefile.objs | 1 + util/stats64.c | 137 +++++++++++++++++++++++++++++ 3 files changed, 331 insertions(+) create mode 100644 include/qemu/stats64.h create mode 100644 util/stats64.c diff --git a/include/qemu/stats64.h b/include/qemu/stats64.h new file mode 100644 index 0000000000..4a357b3e9d --- /dev/null +++ b/include/qemu/stats64.h @@ -0,0 +1,193 @@ +/* + * Atomic operations on 64-bit quantities. + * + * Copyright (C) 2017 Red Hat, Inc. + * + * Author: Paolo Bonzini + * + * This work is licensed under the terms of the GNU GPL, version 2 or later. + * See the COPYING file in the top-level directory. + */ + +#ifndef QEMU_STATS64_H +#define QEMU_STATS64_H 1 + +#include "qemu/atomic.h" + +/* This provides atomic operations on 64-bit type, using a reader-writer + * spinlock on architectures that do not have 64-bit accesses. Even on + * those architectures, it tries hard not to take the lock. + */ + +typedef struct Stat64 { +#ifdef CONFIG_ATOMIC64 + uint64_t value; +#else + uint32_t low, high; + uint32_t lock; +#endif +} Stat64; + +#ifdef CONFIG_ATOMIC64 +static inline void stat64_init(Stat64 *s, uint64_t value) +{ + /* This is not guaranteed to be atomic! */ + *s = (Stat64) { value }; +} + +static inline uint64_t stat64_get(const Stat64 *s) +{ + return atomic_read__nocheck(&s->value); +} + +static inline void stat64_add(Stat64 *s, uint64_t value) +{ + atomic_add(&s->value, value); +} + +static inline void stat64_min(Stat64 *s, uint64_t value) +{ + uint64_t orig = atomic_read__nocheck(&s->value); + while (orig > value) { + orig = atomic_cmpxchg__nocheck(&s->value, orig, value); + } +} + +static inline void stat64_max(Stat64 *s, uint64_t value) +{ + uint64_t orig = atomic_read__nocheck(&s->value); + while (orig < value) { + orig = atomic_cmpxchg__nocheck(&s->value, orig, value); + } +} +#else +uint64_t stat64_get(const Stat64 *s); +bool stat64_min_slow(Stat64 *s, uint64_t value); +bool stat64_max_slow(Stat64 *s, uint64_t value); +bool stat64_add32_carry(Stat64 *s, uint32_t low, uint32_t high); + +static inline void stat64_init(Stat64 *s, uint64_t value) +{ + /* This is not guaranteed to be atomic! */ + *s = (Stat64) { .low = value, .high = value >> 32, .lock = 0 }; +} + +static inline void stat64_add(Stat64 *s, uint64_t value) +{ + uint32_t low, high; + high = value >> 32; + low = (uint32_t) value; + if (!low) { + if (high) { + atomic_add(&s->high, high); + } + return; + } + + for (;;) { + uint32_t orig = s->low; + uint32_t result = orig + low; + uint32_t old; + + if (result < low || high) { + /* If the high part is affected, take the lock. */ + if (stat64_add32_carry(s, low, high)) { + return; + } + continue; + } + + /* No carry, try with a 32-bit cmpxchg. The result is independent of + * the high 32 bits, so it can race just fine with stat64_add32_carry + * and even stat64_get! + */ + old = atomic_cmpxchg(&s->low, orig, result); + if (orig == old) { + return; + } + } +} + +static inline void stat64_min(Stat64 *s, uint64_t value) +{ + uint32_t low, high; + uint32_t orig_low, orig_high; + + high = value >> 32; + low = (uint32_t) value; + do { + orig_high = atomic_read(&s->high); + if (orig_high < high) { + return; + } + + if (orig_high == high) { + /* High 32 bits are equal. Read low after high, otherwise we + * can get a false positive (e.g. 0x1235,0x0000 changes to + * 0x1234,0x8000 and we read it as 0x1234,0x0000). Pairs with + * the write barrier in stat64_min_slow. + */ + smp_rmb(); + orig_low = atomic_read(&s->low); + if (orig_low <= low) { + return; + } + + /* See if we were lucky and a writer raced against us. The + * barrier is theoretically unnecessary, but if we remove it + * we may miss being lucky. + */ + smp_rmb(); + orig_high = atomic_read(&s->high); + if (orig_high < high) { + return; + } + } + + /* If the value changes in any way, we have to take the lock. */ + } while (!stat64_min_slow(s, value)); +} + +static inline void stat64_max(Stat64 *s, uint64_t value) +{ + uint32_t low, high; + uint32_t orig_low, orig_high; + + high = value >> 32; + low = (uint32_t) value; + do { + orig_high = atomic_read(&s->high); + if (orig_high > high) { + return; + } + + if (orig_high == high) { + /* High 32 bits are equal. Read low after high, otherwise we + * can get a false positive (e.g. 0x1234,0x8000 changes to + * 0x1235,0x0000 and we read it as 0x1235,0x8000). Pairs with + * the write barrier in stat64_max_slow. + */ + smp_rmb(); + orig_low = atomic_read(&s->low); + if (orig_low >= low) { + return; + } + + /* See if we were lucky and a writer raced against us. The + * barrier is theoretically unnecessary, but if we remove it + * we may miss being lucky. + */ + smp_rmb(); + orig_high = atomic_read(&s->high); + if (orig_high > high) { + return; + } + } + + /* If the value changes in any way, we have to take the lock. */ + } while (!stat64_max_slow(s, value)); +} + +#endif + +#endif diff --git a/util/Makefile.objs b/util/Makefile.objs index c6205ebf86..8a333d3dd7 100644 --- a/util/Makefile.objs +++ b/util/Makefile.objs @@ -42,4 +42,5 @@ util-obj-y += log.o util-obj-y += qdist.o util-obj-y += qht.o util-obj-y += range.o +util-obj-y += stats64.o util-obj-y += systemd.o diff --git a/util/stats64.c b/util/stats64.c new file mode 100644 index 0000000000..9968fcceac --- /dev/null +++ b/util/stats64.c @@ -0,0 +1,137 @@ +/* + * Atomic operations on 64-bit quantities. + * + * Copyright (C) 2017 Red Hat, Inc. + * + * Author: Paolo Bonzini + * + * This work is licensed under the terms of the GNU GPL, version 2 or later. + * See the COPYING file in the top-level directory. + */ + +#include "qemu/osdep.h" +#include "qemu/atomic.h" +#include "qemu/stats64.h" +#include "qemu/processor.h" + +#ifndef CONFIG_ATOMIC64 +static inline void stat64_rdlock(Stat64 *s) +{ + /* Keep out incoming writers to avoid them starving us. */ + atomic_add(&s->lock, 2); + + /* If there is a concurrent writer, wait for it. */ + while (atomic_read(&s->lock) & 1) { + cpu_relax(); + } +} + +static inline void stat64_rdunlock(Stat64 *s) +{ + atomic_sub(&s->lock, 2); +} + +static inline bool stat64_wrtrylock(Stat64 *s) +{ + return atomic_cmpxchg(&s->lock, 0, 1) == 0; +} + +static inline void stat64_wrunlock(Stat64 *s) +{ + atomic_dec(&s->lock); +} + +uint64_t stat64_get(const Stat64 *s) +{ + uint32_t high, low; + + stat64_rdlock((Stat64 *)s); + + /* 64-bit writes always take the lock, so we can read in + * any order. + */ + high = atomic_read(&s->high); + low = atomic_read(&s->low); + stat64_rdunlock((Stat64 *)s); + + return ((uint64_t)high << 32) | low; +} + +bool stat64_add32_carry(Stat64 *s, uint32_t low, uint32_t high) +{ + uint32_t old; + + if (!stat64_wrtrylock(s)) { + cpu_relax(); + return false; + } + + /* 64-bit reads always take the lock, so they don't care about the + * order of our update. By updating s->low first, we can check + * whether we have to carry into s->high. + */ + old = atomic_fetch_add(&s->low, low); + high += (old + low) < old; + atomic_add(&s->high, high); + stat64_wrunlock(s); + return true; +} + +bool stat64_min_slow(Stat64 *s, uint64_t value) +{ + uint32_t high, low; + uint64_t orig; + + if (!stat64_wrtrylock(s)) { + cpu_relax(); + return false; + } + + high = atomic_read(&s->high); + low = atomic_read(&s->low); + + orig = ((uint64_t)high << 32) | low; + if (orig < value) { + /* We have to set low before high, just like stat64_min reads + * high before low. The value may become higher temporarily, but + * stat64_get does not notice (it takes the lock) and the only ill + * effect on stat64_min is that the slow path may be triggered + * unnecessarily. + */ + atomic_set(&s->low, (uint32_t)value); + smp_wmb(); + atomic_set(&s->high, value >> 32); + } + stat64_wrunlock(s); + return true; +} + +bool stat64_max_slow(Stat64 *s, uint64_t value) +{ + uint32_t high, low; + uint64_t orig; + + if (!stat64_wrtrylock(s)) { + cpu_relax(); + return false; + } + + high = atomic_read(&s->high); + low = atomic_read(&s->low); + + orig = ((uint64_t)high << 32) | low; + if (orig > value) { + /* We have to set low before high, just like stat64_max reads + * high before low. The value may become lower temporarily, but + * stat64_get does not notice (it takes the lock) and the only ill + * effect on stat64_max is that the slow path may be triggered + * unnecessarily. + */ + atomic_set(&s->low, (uint32_t)value); + smp_wmb(); + atomic_set(&s->high, value >> 32); + } + stat64_wrunlock(s); + return true; +} +#endif From f7946da274cf451d66e5207f21286fea14c4795d Mon Sep 17 00:00:00 2001 From: Paolo Bonzini Date: Mon, 5 Jun 2017 14:39:00 +0200 Subject: [PATCH 15/23] block: use Stat64 for wr_highest_offset Reviewed-by: Stefan Hajnoczi Signed-off-by: Paolo Bonzini Message-Id: <20170605123908.18777-12-pbonzini@redhat.com> Signed-off-by: Fam Zheng --- block/io.c | 4 +--- block/qapi.c | 2 +- include/block/block_int.h | 7 ++++--- 3 files changed, 6 insertions(+), 7 deletions(-) diff --git a/block/io.c b/block/io.c index bb1c9c5432..bc69b4cecd 100644 --- a/block/io.c +++ b/block/io.c @@ -1405,9 +1405,7 @@ static int coroutine_fn bdrv_aligned_pwritev(BdrvChild *child, ++bs->write_gen; bdrv_set_dirty(bs, start_sector, end_sector - start_sector); - if (bs->wr_highest_offset < offset + bytes) { - bs->wr_highest_offset = offset + bytes; - } + stat64_max(&bs->wr_highest_offset, offset + bytes); if (ret >= 0) { bs->total_sectors = MAX(bs->total_sectors, end_sector); diff --git a/block/qapi.c b/block/qapi.c index a40922ea26..14b60ae66c 100644 --- a/block/qapi.c +++ b/block/qapi.c @@ -441,7 +441,7 @@ static BlockStats *bdrv_query_bds_stats(const BlockDriverState *bs, s->node_name = g_strdup(bdrv_get_node_name(bs)); } - s->stats->wr_highest_offset = bs->wr_highest_offset; + s->stats->wr_highest_offset = stat64_get(&bs->wr_highest_offset); if (bs->file) { s->has_parent = true; diff --git a/include/block/block_int.h b/include/block/block_int.h index 091b5d361c..28f8b1e526 100644 --- a/include/block/block_int.h +++ b/include/block/block_int.h @@ -29,6 +29,7 @@ #include "qemu/option.h" #include "qemu/queue.h" #include "qemu/coroutine.h" +#include "qemu/stats64.h" #include "qemu/timer.h" #include "qapi-types.h" #include "qemu/hbitmap.h" @@ -604,9 +605,6 @@ struct BlockDriverState { /* Callback before write request is processed */ NotifierWithReturnList before_write_notifiers; - /* Offset after the highest byte written to */ - uint64_t wr_highest_offset; - /* threshold limit for writes, in bytes. "High water mark". */ uint64_t write_threshold_offset; NotifierWithReturn write_threshold_notifier; @@ -619,6 +617,9 @@ struct BlockDriverState { QLIST_HEAD(, BdrvDirtyBitmap) dirty_bitmaps; + /* Offset after the highest byte written to */ + Stat64 wr_highest_offset; + /* If true, copy read backing sectors into image. Can be >1 if more * than one client has requested copy-on-read. Accessed with atomic * ops. From 47fec59941a7182380bc8dde3faf17cfb05b770f Mon Sep 17 00:00:00 2001 From: Paolo Bonzini Date: Mon, 5 Jun 2017 14:39:01 +0200 Subject: [PATCH 16/23] block: access write_gen with atomics Reviewed-by: Stefan Hajnoczi Signed-off-by: Paolo Bonzini Message-Id: <20170605123908.18777-13-pbonzini@redhat.com> Signed-off-by: Fam Zheng --- block.c | 2 +- block/io.c | 6 +++--- include/block/block_int.h | 2 +- 3 files changed, 5 insertions(+), 5 deletions(-) diff --git a/block.c b/block.c index af6366bcef..361005cba9 100644 --- a/block.c +++ b/block.c @@ -3424,7 +3424,7 @@ int bdrv_truncate(BdrvChild *child, int64_t offset, Error **errp) ret = refresh_total_sectors(bs, offset >> BDRV_SECTOR_BITS); bdrv_dirty_bitmap_truncate(bs); bdrv_parent_cb_resize(bs); - ++bs->write_gen; + atomic_inc(&bs->write_gen); } return ret; } diff --git a/block/io.c b/block/io.c index bc69b4cecd..036f5a4870 100644 --- a/block/io.c +++ b/block/io.c @@ -1402,7 +1402,7 @@ static int coroutine_fn bdrv_aligned_pwritev(BdrvChild *child, } bdrv_debug_event(bs, BLKDBG_PWRITEV_DONE); - ++bs->write_gen; + atomic_inc(&bs->write_gen); bdrv_set_dirty(bs, start_sector, end_sector - start_sector); stat64_max(&bs->wr_highest_offset, offset + bytes); @@ -2291,7 +2291,7 @@ int coroutine_fn bdrv_co_flush(BlockDriverState *bs) goto early_exit; } - current_gen = bs->write_gen; + current_gen = atomic_read(&bs->write_gen); /* Wait until any previous flushes are completed */ while (bs->active_flush_req) { @@ -2516,7 +2516,7 @@ int coroutine_fn bdrv_co_pdiscard(BlockDriverState *bs, int64_t offset, } ret = 0; out: - ++bs->write_gen; + atomic_inc(&bs->write_gen); bdrv_set_dirty(bs, req.offset >> BDRV_SECTOR_BITS, req.bytes >> BDRV_SECTOR_BITS); tracked_request_end(&req); diff --git a/include/block/block_int.h b/include/block/block_int.h index 28f8b1e526..b7d0dfb0b9 100644 --- a/include/block/block_int.h +++ b/include/block/block_int.h @@ -612,7 +612,6 @@ struct BlockDriverState { QLIST_HEAD(, BdrvTrackedRequest) tracked_requests; CoQueue flush_queue; /* Serializing flush queue */ bool active_flush_req; /* Flush request in flight? */ - unsigned int write_gen; /* Current data generation */ unsigned int flushed_gen; /* Flushed write generation */ QLIST_HEAD(, BdrvDirtyBitmap) dirty_bitmaps; @@ -647,6 +646,7 @@ struct BlockDriverState { /* Accessed with atomic ops. */ int quiesce_counter; + unsigned int write_gen; /* Current data generation */ }; struct BlockBackendRootState { From 3783fa3dd3d0104e9780bde073f66f1e37f1ce1c Mon Sep 17 00:00:00 2001 From: Paolo Bonzini Date: Mon, 5 Jun 2017 14:39:02 +0200 Subject: [PATCH 17/23] block: protect tracked_requests and flush_queue with reqs_lock Reviewed-by: Stefan Hajnoczi Signed-off-by: Paolo Bonzini Message-Id: <20170605123908.18777-14-pbonzini@redhat.com> Signed-off-by: Fam Zheng --- block.c | 1 + block/io.c | 16 ++++++++++++++-- include/block/block_int.h | 14 +++++++++----- 3 files changed, 24 insertions(+), 7 deletions(-) diff --git a/block.c b/block.c index 361005cba9..a5cbd4540f 100644 --- a/block.c +++ b/block.c @@ -320,6 +320,7 @@ BlockDriverState *bdrv_new(void) QLIST_INIT(&bs->op_blockers[i]); } notifier_with_return_list_init(&bs->before_write_notifiers); + qemu_co_mutex_init(&bs->reqs_lock); bs->refcnt = 1; bs->aio_context = qemu_get_aio_context(); diff --git a/block/io.c b/block/io.c index 036f5a4870..91611ffb2a 100644 --- a/block/io.c +++ b/block/io.c @@ -378,8 +378,10 @@ static void tracked_request_end(BdrvTrackedRequest *req) atomic_dec(&req->bs->serialising_in_flight); } + qemu_co_mutex_lock(&req->bs->reqs_lock); QLIST_REMOVE(req, list); qemu_co_queue_restart_all(&req->wait_queue); + qemu_co_mutex_unlock(&req->bs->reqs_lock); } /** @@ -404,7 +406,9 @@ static void tracked_request_begin(BdrvTrackedRequest *req, qemu_co_queue_init(&req->wait_queue); + qemu_co_mutex_lock(&bs->reqs_lock); QLIST_INSERT_HEAD(&bs->tracked_requests, req, list); + qemu_co_mutex_unlock(&bs->reqs_lock); } static void mark_request_serialising(BdrvTrackedRequest *req, uint64_t align) @@ -526,6 +530,7 @@ static bool coroutine_fn wait_serialising_requests(BdrvTrackedRequest *self) do { retry = false; + qemu_co_mutex_lock(&bs->reqs_lock); QLIST_FOREACH(req, &bs->tracked_requests, list) { if (req == self || (!req->serialising && !self->serialising)) { continue; @@ -544,7 +549,7 @@ static bool coroutine_fn wait_serialising_requests(BdrvTrackedRequest *self) * (instead of producing a deadlock in the former case). */ if (!req->waiting_for) { self->waiting_for = req; - qemu_co_queue_wait(&req->wait_queue, NULL); + qemu_co_queue_wait(&req->wait_queue, &bs->reqs_lock); self->waiting_for = NULL; retry = true; waited = true; @@ -552,6 +557,7 @@ static bool coroutine_fn wait_serialising_requests(BdrvTrackedRequest *self) } } } + qemu_co_mutex_unlock(&bs->reqs_lock); } while (retry); return waited; @@ -2291,14 +2297,17 @@ int coroutine_fn bdrv_co_flush(BlockDriverState *bs) goto early_exit; } + qemu_co_mutex_lock(&bs->reqs_lock); current_gen = atomic_read(&bs->write_gen); /* Wait until any previous flushes are completed */ while (bs->active_flush_req) { - qemu_co_queue_wait(&bs->flush_queue, NULL); + qemu_co_queue_wait(&bs->flush_queue, &bs->reqs_lock); } + /* Flushes reach this point in nondecreasing current_gen order. */ bs->active_flush_req = true; + qemu_co_mutex_unlock(&bs->reqs_lock); /* Write back all layers by calling one driver function */ if (bs->drv->bdrv_co_flush) { @@ -2370,9 +2379,12 @@ out: if (ret == 0) { bs->flushed_gen = current_gen; } + + qemu_co_mutex_lock(&bs->reqs_lock); bs->active_flush_req = false; /* Return value is ignored - it's ok if wait queue is empty */ qemu_co_queue_next(&bs->flush_queue); + qemu_co_mutex_unlock(&bs->reqs_lock); early_exit: bdrv_dec_in_flight(bs); diff --git a/include/block/block_int.h b/include/block/block_int.h index b7d0dfb0b9..ae74df97ec 100644 --- a/include/block/block_int.h +++ b/include/block/block_int.h @@ -609,11 +609,6 @@ struct BlockDriverState { uint64_t write_threshold_offset; NotifierWithReturn write_threshold_notifier; - QLIST_HEAD(, BdrvTrackedRequest) tracked_requests; - CoQueue flush_queue; /* Serializing flush queue */ - bool active_flush_req; /* Flush request in flight? */ - unsigned int flushed_gen; /* Flushed write generation */ - QLIST_HEAD(, BdrvDirtyBitmap) dirty_bitmaps; /* Offset after the highest byte written to */ @@ -647,6 +642,15 @@ struct BlockDriverState { /* Accessed with atomic ops. */ int quiesce_counter; unsigned int write_gen; /* Current data generation */ + + /* Protected by reqs_lock. */ + CoMutex reqs_lock; + QLIST_HEAD(, BdrvTrackedRequest) tracked_requests; + CoQueue flush_queue; /* Serializing flush queue */ + bool active_flush_req; /* Flush request in flight? */ + + /* Only read/written by whoever has set active_flush_req to true. */ + unsigned int flushed_gen; /* Flushed write generation */ }; struct BlockBackendRootState { From 2119882c7eb7e2c612b24fc0c8d86f5887d6f1c3 Mon Sep 17 00:00:00 2001 From: Paolo Bonzini Date: Mon, 5 Jun 2017 14:39:03 +0200 Subject: [PATCH 18/23] block: introduce dirty_bitmap_mutex It protects only the list of dirty bitmaps; in the next patch we will also protect their content. Reviewed-by: Stefan Hajnoczi Signed-off-by: Paolo Bonzini Message-Id: <20170605123908.18777-15-pbonzini@redhat.com> Signed-off-by: Fam Zheng --- block.c | 1 + block/dirty-bitmap.c | 44 ++++++++++++++++++++++++++++++++++++++- block/mirror.c | 3 ++- blockdev.c | 44 +++++++-------------------------------- include/block/block_int.h | 5 +++++ migration/block.c | 6 ------ 6 files changed, 58 insertions(+), 45 deletions(-) diff --git a/block.c b/block.c index a5cbd4540f..694396281b 100644 --- a/block.c +++ b/block.c @@ -321,6 +321,7 @@ BlockDriverState *bdrv_new(void) } notifier_with_return_list_init(&bs->before_write_notifiers); qemu_co_mutex_init(&bs->reqs_lock); + qemu_mutex_init(&bs->dirty_bitmap_mutex); bs->refcnt = 1; bs->aio_context = qemu_get_aio_context(); diff --git a/block/dirty-bitmap.c b/block/dirty-bitmap.c index 519737c8d3..fa78109365 100644 --- a/block/dirty-bitmap.c +++ b/block/dirty-bitmap.c @@ -52,6 +52,17 @@ struct BdrvDirtyBitmapIter { BdrvDirtyBitmap *bitmap; }; +static inline void bdrv_dirty_bitmaps_lock(BlockDriverState *bs) +{ + qemu_mutex_lock(&bs->dirty_bitmap_mutex); +} + +static inline void bdrv_dirty_bitmaps_unlock(BlockDriverState *bs) +{ + qemu_mutex_unlock(&bs->dirty_bitmap_mutex); +} + +/* Called with BQL or dirty_bitmap lock taken. */ BdrvDirtyBitmap *bdrv_find_dirty_bitmap(BlockDriverState *bs, const char *name) { BdrvDirtyBitmap *bm; @@ -65,6 +76,7 @@ BdrvDirtyBitmap *bdrv_find_dirty_bitmap(BlockDriverState *bs, const char *name) return NULL; } +/* Called with BQL taken. */ void bdrv_dirty_bitmap_make_anon(BdrvDirtyBitmap *bitmap) { assert(!bdrv_dirty_bitmap_frozen(bitmap)); @@ -72,6 +84,7 @@ void bdrv_dirty_bitmap_make_anon(BdrvDirtyBitmap *bitmap) bitmap->name = NULL; } +/* Called with BQL taken. */ BdrvDirtyBitmap *bdrv_create_dirty_bitmap(BlockDriverState *bs, uint32_t granularity, const char *name, @@ -100,7 +113,9 @@ BdrvDirtyBitmap *bdrv_create_dirty_bitmap(BlockDriverState *bs, bitmap->size = bitmap_size; bitmap->name = g_strdup(name); bitmap->disabled = false; + bdrv_dirty_bitmaps_lock(bs); QLIST_INSERT_HEAD(&bs->dirty_bitmaps, bitmap, list); + bdrv_dirty_bitmaps_unlock(bs); return bitmap; } @@ -164,16 +179,19 @@ const char *bdrv_dirty_bitmap_name(const BdrvDirtyBitmap *bitmap) return bitmap->name; } +/* Called with BQL taken. */ bool bdrv_dirty_bitmap_frozen(BdrvDirtyBitmap *bitmap) { return bitmap->successor; } +/* Called with BQL taken. */ bool bdrv_dirty_bitmap_enabled(BdrvDirtyBitmap *bitmap) { return !(bitmap->disabled || bitmap->successor); } +/* Called with BQL taken. */ DirtyBitmapStatus bdrv_dirty_bitmap_status(BdrvDirtyBitmap *bitmap) { if (bdrv_dirty_bitmap_frozen(bitmap)) { @@ -188,6 +206,7 @@ DirtyBitmapStatus bdrv_dirty_bitmap_status(BdrvDirtyBitmap *bitmap) /** * Create a successor bitmap destined to replace this bitmap after an operation. * Requires that the bitmap is not frozen and has no successor. + * Called with BQL taken. */ int bdrv_dirty_bitmap_create_successor(BlockDriverState *bs, BdrvDirtyBitmap *bitmap, Error **errp) @@ -220,6 +239,7 @@ int bdrv_dirty_bitmap_create_successor(BlockDriverState *bs, /** * For a bitmap with a successor, yield our name to the successor, * delete the old bitmap, and return a handle to the new bitmap. + * Called with BQL taken. */ BdrvDirtyBitmap *bdrv_dirty_bitmap_abdicate(BlockDriverState *bs, BdrvDirtyBitmap *bitmap, @@ -247,6 +267,7 @@ BdrvDirtyBitmap *bdrv_dirty_bitmap_abdicate(BlockDriverState *bs, * In cases of failure where we can no longer safely delete the parent, * we may wish to re-join the parent and child/successor. * The merged parent will be un-frozen, but not explicitly re-enabled. + * Called with BQL taken. */ BdrvDirtyBitmap *bdrv_reclaim_dirty_bitmap(BlockDriverState *bs, BdrvDirtyBitmap *parent, @@ -271,25 +292,30 @@ BdrvDirtyBitmap *bdrv_reclaim_dirty_bitmap(BlockDriverState *bs, /** * Truncates _all_ bitmaps attached to a BDS. + * Called with BQL taken. */ void bdrv_dirty_bitmap_truncate(BlockDriverState *bs) { BdrvDirtyBitmap *bitmap; uint64_t size = bdrv_nb_sectors(bs); + bdrv_dirty_bitmaps_lock(bs); QLIST_FOREACH(bitmap, &bs->dirty_bitmaps, list) { assert(!bdrv_dirty_bitmap_frozen(bitmap)); assert(!bitmap->active_iterators); hbitmap_truncate(bitmap->bitmap, size); bitmap->size = size; } + bdrv_dirty_bitmaps_unlock(bs); } +/* Called with BQL taken. */ static void bdrv_do_release_matching_dirty_bitmap(BlockDriverState *bs, BdrvDirtyBitmap *bitmap, bool only_named) { BdrvDirtyBitmap *bm, *next; + bdrv_dirty_bitmaps_lock(bs); QLIST_FOREACH_SAFE(bm, &bs->dirty_bitmaps, list, next) { if ((!bitmap || bm == bitmap) && (!only_named || bm->name)) { assert(!bm->active_iterators); @@ -301,15 +327,19 @@ static void bdrv_do_release_matching_dirty_bitmap(BlockDriverState *bs, g_free(bm); if (bitmap) { - return; + goto out; } } } if (bitmap) { abort(); } + +out: + bdrv_dirty_bitmaps_unlock(bs); } +/* Called with BQL taken. */ void bdrv_release_dirty_bitmap(BlockDriverState *bs, BdrvDirtyBitmap *bitmap) { bdrv_do_release_matching_dirty_bitmap(bs, bitmap, false); @@ -318,18 +348,21 @@ void bdrv_release_dirty_bitmap(BlockDriverState *bs, BdrvDirtyBitmap *bitmap) /** * Release all named dirty bitmaps attached to a BDS (for use in bdrv_close()). * There must not be any frozen bitmaps attached. + * Called with BQL taken. */ void bdrv_release_named_dirty_bitmaps(BlockDriverState *bs) { bdrv_do_release_matching_dirty_bitmap(bs, NULL, true); } +/* Called with BQL taken. */ void bdrv_disable_dirty_bitmap(BdrvDirtyBitmap *bitmap) { assert(!bdrv_dirty_bitmap_frozen(bitmap)); bitmap->disabled = true; } +/* Called with BQL taken. */ void bdrv_enable_dirty_bitmap(BdrvDirtyBitmap *bitmap) { assert(!bdrv_dirty_bitmap_frozen(bitmap)); @@ -342,6 +375,7 @@ BlockDirtyInfoList *bdrv_query_dirty_bitmaps(BlockDriverState *bs) BlockDirtyInfoList *list = NULL; BlockDirtyInfoList **plist = &list; + bdrv_dirty_bitmaps_lock(bs); QLIST_FOREACH(bm, &bs->dirty_bitmaps, list) { BlockDirtyInfo *info = g_new0(BlockDirtyInfo, 1); BlockDirtyInfoList *entry = g_new0(BlockDirtyInfoList, 1); @@ -354,6 +388,7 @@ BlockDirtyInfoList *bdrv_query_dirty_bitmaps(BlockDriverState *bs) *plist = entry; plist = &entry->next; } + bdrv_dirty_bitmaps_unlock(bs); return list; } @@ -508,12 +543,19 @@ void bdrv_set_dirty(BlockDriverState *bs, int64_t cur_sector, int64_t nr_sectors) { BdrvDirtyBitmap *bitmap; + + if (QLIST_EMPTY(&bs->dirty_bitmaps)) { + return; + } + + bdrv_dirty_bitmaps_lock(bs); QLIST_FOREACH(bitmap, &bs->dirty_bitmaps, list) { if (!bdrv_dirty_bitmap_enabled(bitmap)) { continue; } hbitmap_set(bitmap->bitmap, cur_sector, nr_sectors); } + bdrv_dirty_bitmaps_unlock(bs); } /** diff --git a/block/mirror.c b/block/mirror.c index a2a970301c..88ae882c46 100644 --- a/block/mirror.c +++ b/block/mirror.c @@ -506,6 +506,8 @@ static void mirror_exit(BlockJob *job, void *opaque) BlockDriverState *mirror_top_bs = s->mirror_top_bs; Error *local_err = NULL; + bdrv_release_dirty_bitmap(src, s->dirty_bitmap); + /* Make sure that the source BDS doesn't go away before we called * block_job_completed(). */ bdrv_ref(src); @@ -904,7 +906,6 @@ immediate_exit: g_free(s->cow_bitmap); g_free(s->in_flight_bitmap); bdrv_dirty_iter_free(s->dbi); - bdrv_release_dirty_bitmap(bs, s->dirty_bitmap); data = g_malloc(sizeof(*data)); data->ret = ret; diff --git a/blockdev.c b/blockdev.c index ceea976b03..39f289ba68 100644 --- a/blockdev.c +++ b/blockdev.c @@ -1362,12 +1362,10 @@ out_aio_context: static BdrvDirtyBitmap *block_dirty_bitmap_lookup(const char *node, const char *name, BlockDriverState **pbs, - AioContext **paio, Error **errp) { BlockDriverState *bs; BdrvDirtyBitmap *bitmap; - AioContext *aio_context; if (!node) { error_setg(errp, "Node cannot be NULL"); @@ -1383,29 +1381,17 @@ static BdrvDirtyBitmap *block_dirty_bitmap_lookup(const char *node, return NULL; } - aio_context = bdrv_get_aio_context(bs); - aio_context_acquire(aio_context); - bitmap = bdrv_find_dirty_bitmap(bs, name); if (!bitmap) { error_setg(errp, "Dirty bitmap '%s' not found", name); - goto fail; + return NULL; } if (pbs) { *pbs = bs; } - if (paio) { - *paio = aio_context; - } else { - aio_context_release(aio_context); - } return bitmap; - - fail: - aio_context_release(aio_context); - return NULL; } /* New and old BlockDriverState structs for atomic group operations */ @@ -2025,7 +2011,6 @@ static void block_dirty_bitmap_clear_prepare(BlkActionState *common, state->bitmap = block_dirty_bitmap_lookup(action->node, action->name, &state->bs, - &state->aio_context, errp); if (!state->bitmap) { return; @@ -2733,7 +2718,6 @@ void qmp_block_dirty_bitmap_add(const char *node, const char *name, bool has_granularity, uint32_t granularity, Error **errp) { - AioContext *aio_context; BlockDriverState *bs; if (!name || name[0] == '\0') { @@ -2746,14 +2730,11 @@ void qmp_block_dirty_bitmap_add(const char *node, const char *name, return; } - aio_context = bdrv_get_aio_context(bs); - aio_context_acquire(aio_context); - if (has_granularity) { if (granularity < 512 || !is_power_of_2(granularity)) { error_setg(errp, "Granularity must be power of 2 " "and at least 512"); - goto out; + return; } } else { /* Default to cluster size, if available: */ @@ -2761,19 +2742,15 @@ void qmp_block_dirty_bitmap_add(const char *node, const char *name, } bdrv_create_dirty_bitmap(bs, granularity, name, errp); - - out: - aio_context_release(aio_context); } void qmp_block_dirty_bitmap_remove(const char *node, const char *name, Error **errp) { - AioContext *aio_context; BlockDriverState *bs; BdrvDirtyBitmap *bitmap; - bitmap = block_dirty_bitmap_lookup(node, name, &bs, &aio_context, errp); + bitmap = block_dirty_bitmap_lookup(node, name, &bs, errp); if (!bitmap || !bs) { return; } @@ -2782,13 +2759,10 @@ void qmp_block_dirty_bitmap_remove(const char *node, const char *name, error_setg(errp, "Bitmap '%s' is currently frozen and cannot be removed", name); - goto out; + return; } bdrv_dirty_bitmap_make_anon(bitmap); bdrv_release_dirty_bitmap(bs, bitmap); - - out: - aio_context_release(aio_context); } /** @@ -2798,11 +2772,10 @@ void qmp_block_dirty_bitmap_remove(const char *node, const char *name, void qmp_block_dirty_bitmap_clear(const char *node, const char *name, Error **errp) { - AioContext *aio_context; BdrvDirtyBitmap *bitmap; BlockDriverState *bs; - bitmap = block_dirty_bitmap_lookup(node, name, &bs, &aio_context, errp); + bitmap = block_dirty_bitmap_lookup(node, name, &bs, errp); if (!bitmap || !bs) { return; } @@ -2811,18 +2784,15 @@ void qmp_block_dirty_bitmap_clear(const char *node, const char *name, error_setg(errp, "Bitmap '%s' is currently frozen and cannot be modified", name); - goto out; + return; } else if (!bdrv_dirty_bitmap_enabled(bitmap)) { error_setg(errp, "Bitmap '%s' is currently disabled and cannot be cleared", name); - goto out; + return; } bdrv_clear_dirty_bitmap(bitmap, NULL); - - out: - aio_context_release(aio_context); } void hmp_drive_del(Monitor *mon, const QDict *qdict) diff --git a/include/block/block_int.h b/include/block/block_int.h index ae74df97ec..21cb65bd60 100644 --- a/include/block/block_int.h +++ b/include/block/block_int.h @@ -609,6 +609,11 @@ struct BlockDriverState { uint64_t write_threshold_offset; NotifierWithReturn write_threshold_notifier; + /* Writing to the list requires the BQL _and_ the dirty_bitmap_mutex. + * Reading from the list can be done with either the BQL or the + * dirty_bitmap_mutex. Modifying a bitmap requires the AioContext + * lock. */ + QemuMutex dirty_bitmap_mutex; QLIST_HEAD(, BdrvDirtyBitmap) dirty_bitmaps; /* Offset after the highest byte written to */ diff --git a/migration/block.c b/migration/block.c index 3aae5a375e..9e2092681e 100644 --- a/migration/block.c +++ b/migration/block.c @@ -341,10 +341,8 @@ static int set_dirty_tracking(void) int ret; QSIMPLEQ_FOREACH(bmds, &block_mig_state.bmds_list, entry) { - aio_context_acquire(blk_get_aio_context(bmds->blk)); bmds->dirty_bitmap = bdrv_create_dirty_bitmap(blk_bs(bmds->blk), BLOCK_SIZE, NULL, NULL); - aio_context_release(blk_get_aio_context(bmds->blk)); if (!bmds->dirty_bitmap) { ret = -errno; goto fail; @@ -355,9 +353,7 @@ static int set_dirty_tracking(void) fail: QSIMPLEQ_FOREACH(bmds, &block_mig_state.bmds_list, entry) { if (bmds->dirty_bitmap) { - aio_context_acquire(blk_get_aio_context(bmds->blk)); bdrv_release_dirty_bitmap(blk_bs(bmds->blk), bmds->dirty_bitmap); - aio_context_release(blk_get_aio_context(bmds->blk)); } } return ret; @@ -370,9 +366,7 @@ static void unset_dirty_tracking(void) BlkMigDevState *bmds; QSIMPLEQ_FOREACH(bmds, &block_mig_state.bmds_list, entry) { - aio_context_acquire(blk_get_aio_context(bmds->blk)); bdrv_release_dirty_bitmap(blk_bs(bmds->blk), bmds->dirty_bitmap); - aio_context_release(blk_get_aio_context(bmds->blk)); } } From c0bad49946287ef218e871608b753b9991f6e54c Mon Sep 17 00:00:00 2001 From: Paolo Bonzini Date: Mon, 5 Jun 2017 14:39:04 +0200 Subject: [PATCH 19/23] migration/block: reset dirty bitmap before reading Any data that is returned by read may be stale already, the bitmap has to be cleared before issuing the read. Reviewed-by: Stefan Hajnoczi Signed-off-by: Paolo Bonzini Message-Id: <20170605123908.18777-16-pbonzini@redhat.com> Signed-off-by: Fam Zheng --- migration/block.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/migration/block.c b/migration/block.c index 9e2092681e..423877bb40 100644 --- a/migration/block.c +++ b/migration/block.c @@ -532,6 +532,8 @@ static int mig_save_device_dirty(QEMUFile *f, BlkMigDevState *bmds, } else { nr_sectors = BDRV_SECTORS_PER_DIRTY_CHUNK; } + bdrv_reset_dirty_bitmap(bmds->dirty_bitmap, sector, nr_sectors); + blk = g_new(BlkMigBlock, 1); blk->buf = g_malloc(BLOCK_SIZE); blk->bmds = bmds; @@ -564,7 +566,6 @@ static int mig_save_device_dirty(QEMUFile *f, BlkMigDevState *bmds, g_free(blk); } - bdrv_reset_dirty_bitmap(bmds->dirty_bitmap, sector, nr_sectors); sector += nr_sectors; bmds->cur_dirty = sector; From b64bd51efa9bbf30df1b2f91477d2805678d0b93 Mon Sep 17 00:00:00 2001 From: Paolo Bonzini Date: Mon, 5 Jun 2017 14:39:05 +0200 Subject: [PATCH 20/23] block: protect modification of dirty bitmaps with a mutex Reviewed-by: Stefan Hajnoczi Signed-off-by: Paolo Bonzini Message-Id: <20170605123908.18777-17-pbonzini@redhat.com> Signed-off-by: Fam Zheng --- block/dirty-bitmap.c | 70 +++++++++++++++++++++++++++++++----- block/mirror.c | 11 ++++-- include/block/block_int.h | 4 +-- include/block/dirty-bitmap.h | 25 +++++++++---- migration/block.c | 10 +++--- 5 files changed, 95 insertions(+), 25 deletions(-) diff --git a/block/dirty-bitmap.c b/block/dirty-bitmap.c index fa78109365..a04c6e4154 100644 --- a/block/dirty-bitmap.c +++ b/block/dirty-bitmap.c @@ -37,6 +37,7 @@ * or enabled. A frozen bitmap can only abdicate() or reclaim(). */ struct BdrvDirtyBitmap { + QemuMutex *mutex; HBitmap *bitmap; /* Dirty sector bitmap implementation */ HBitmap *meta; /* Meta dirty bitmap */ BdrvDirtyBitmap *successor; /* Anonymous child; implies frozen status */ @@ -62,6 +63,16 @@ static inline void bdrv_dirty_bitmaps_unlock(BlockDriverState *bs) qemu_mutex_unlock(&bs->dirty_bitmap_mutex); } +void bdrv_dirty_bitmap_lock(BdrvDirtyBitmap *bitmap) +{ + qemu_mutex_lock(bitmap->mutex); +} + +void bdrv_dirty_bitmap_unlock(BdrvDirtyBitmap *bitmap) +{ + qemu_mutex_unlock(bitmap->mutex); +} + /* Called with BQL or dirty_bitmap lock taken. */ BdrvDirtyBitmap *bdrv_find_dirty_bitmap(BlockDriverState *bs, const char *name) { @@ -109,6 +120,7 @@ BdrvDirtyBitmap *bdrv_create_dirty_bitmap(BlockDriverState *bs, return NULL; } bitmap = g_new0(BdrvDirtyBitmap, 1); + bitmap->mutex = &bs->dirty_bitmap_mutex; bitmap->bitmap = hbitmap_alloc(bitmap_size, ctz32(sector_granularity)); bitmap->size = bitmap_size; bitmap->name = g_strdup(name); @@ -134,20 +146,24 @@ void bdrv_create_meta_dirty_bitmap(BdrvDirtyBitmap *bitmap, int chunk_size) { assert(!bitmap->meta); + qemu_mutex_lock(bitmap->mutex); bitmap->meta = hbitmap_create_meta(bitmap->bitmap, chunk_size * BITS_PER_BYTE); + qemu_mutex_unlock(bitmap->mutex); } void bdrv_release_meta_dirty_bitmap(BdrvDirtyBitmap *bitmap) { assert(bitmap->meta); + qemu_mutex_lock(bitmap->mutex); hbitmap_free_meta(bitmap->bitmap); bitmap->meta = NULL; + qemu_mutex_unlock(bitmap->mutex); } -int bdrv_dirty_bitmap_get_meta(BlockDriverState *bs, - BdrvDirtyBitmap *bitmap, int64_t sector, - int nb_sectors) +int bdrv_dirty_bitmap_get_meta_locked(BlockDriverState *bs, + BdrvDirtyBitmap *bitmap, int64_t sector, + int nb_sectors) { uint64_t i; int sectors_per_bit = 1 << hbitmap_granularity(bitmap->meta); @@ -162,11 +178,26 @@ int bdrv_dirty_bitmap_get_meta(BlockDriverState *bs, return false; } +int bdrv_dirty_bitmap_get_meta(BlockDriverState *bs, + BdrvDirtyBitmap *bitmap, int64_t sector, + int nb_sectors) +{ + bool dirty; + + qemu_mutex_lock(bitmap->mutex); + dirty = bdrv_dirty_bitmap_get_meta_locked(bs, bitmap, sector, nb_sectors); + qemu_mutex_unlock(bitmap->mutex); + + return dirty; +} + void bdrv_dirty_bitmap_reset_meta(BlockDriverState *bs, BdrvDirtyBitmap *bitmap, int64_t sector, int nb_sectors) { + qemu_mutex_lock(bitmap->mutex); hbitmap_reset(bitmap->meta, sector, nb_sectors); + qemu_mutex_unlock(bitmap->mutex); } int64_t bdrv_dirty_bitmap_size(const BdrvDirtyBitmap *bitmap) @@ -393,8 +424,9 @@ BlockDirtyInfoList *bdrv_query_dirty_bitmaps(BlockDriverState *bs) return list; } -int bdrv_get_dirty(BlockDriverState *bs, BdrvDirtyBitmap *bitmap, - int64_t sector) +/* Called within bdrv_dirty_bitmap_lock..unlock */ +int bdrv_get_dirty_locked(BlockDriverState *bs, BdrvDirtyBitmap *bitmap, + int64_t sector) { if (bitmap) { return hbitmap_get(bitmap->bitmap, sector); @@ -467,23 +499,42 @@ int64_t bdrv_dirty_iter_next(BdrvDirtyBitmapIter *iter) return hbitmap_iter_next(&iter->hbi); } -void bdrv_set_dirty_bitmap(BdrvDirtyBitmap *bitmap, - int64_t cur_sector, int64_t nr_sectors) +/* Called within bdrv_dirty_bitmap_lock..unlock */ +void bdrv_set_dirty_bitmap_locked(BdrvDirtyBitmap *bitmap, + int64_t cur_sector, int64_t nr_sectors) { assert(bdrv_dirty_bitmap_enabled(bitmap)); hbitmap_set(bitmap->bitmap, cur_sector, nr_sectors); } -void bdrv_reset_dirty_bitmap(BdrvDirtyBitmap *bitmap, - int64_t cur_sector, int64_t nr_sectors) +void bdrv_set_dirty_bitmap(BdrvDirtyBitmap *bitmap, + int64_t cur_sector, int64_t nr_sectors) +{ + bdrv_dirty_bitmap_lock(bitmap); + bdrv_set_dirty_bitmap_locked(bitmap, cur_sector, nr_sectors); + bdrv_dirty_bitmap_unlock(bitmap); +} + +/* Called within bdrv_dirty_bitmap_lock..unlock */ +void bdrv_reset_dirty_bitmap_locked(BdrvDirtyBitmap *bitmap, + int64_t cur_sector, int64_t nr_sectors) { assert(bdrv_dirty_bitmap_enabled(bitmap)); hbitmap_reset(bitmap->bitmap, cur_sector, nr_sectors); } +void bdrv_reset_dirty_bitmap(BdrvDirtyBitmap *bitmap, + int64_t cur_sector, int64_t nr_sectors) +{ + bdrv_dirty_bitmap_lock(bitmap); + bdrv_reset_dirty_bitmap_locked(bitmap, cur_sector, nr_sectors); + bdrv_dirty_bitmap_unlock(bitmap); +} + void bdrv_clear_dirty_bitmap(BdrvDirtyBitmap *bitmap, HBitmap **out) { assert(bdrv_dirty_bitmap_enabled(bitmap)); + bdrv_dirty_bitmap_lock(bitmap); if (!out) { hbitmap_reset_all(bitmap->bitmap); } else { @@ -492,6 +543,7 @@ void bdrv_clear_dirty_bitmap(BdrvDirtyBitmap *bitmap, HBitmap **out) hbitmap_granularity(backup)); *out = backup; } + bdrv_dirty_bitmap_unlock(bitmap); } void bdrv_undo_clear_dirty_bitmap(BdrvDirtyBitmap *bitmap, HBitmap *in) diff --git a/block/mirror.c b/block/mirror.c index 88ae882c46..19afcc6f1a 100644 --- a/block/mirror.c +++ b/block/mirror.c @@ -342,6 +342,7 @@ static uint64_t coroutine_fn mirror_iteration(MirrorBlockJob *s) int max_io_sectors = MAX((s->buf_size >> BDRV_SECTOR_BITS) / MAX_IN_FLIGHT, MAX_IO_SECTORS); + bdrv_dirty_bitmap_lock(s->dirty_bitmap); sector_num = bdrv_dirty_iter_next(s->dbi); if (sector_num < 0) { bdrv_set_dirty_iter(s->dbi, 0); @@ -349,6 +350,7 @@ static uint64_t coroutine_fn mirror_iteration(MirrorBlockJob *s) trace_mirror_restart_iter(s, bdrv_get_dirty_count(s->dirty_bitmap)); assert(sector_num >= 0); } + bdrv_dirty_bitmap_unlock(s->dirty_bitmap); first_chunk = sector_num / sectors_per_chunk; while (test_bit(first_chunk, s->in_flight_bitmap)) { @@ -360,12 +362,13 @@ static uint64_t coroutine_fn mirror_iteration(MirrorBlockJob *s) /* Find the number of consective dirty chunks following the first dirty * one, and wait for in flight requests in them. */ + bdrv_dirty_bitmap_lock(s->dirty_bitmap); while (nb_chunks * sectors_per_chunk < (s->buf_size >> BDRV_SECTOR_BITS)) { int64_t next_dirty; int64_t next_sector = sector_num + nb_chunks * sectors_per_chunk; int64_t next_chunk = next_sector / sectors_per_chunk; if (next_sector >= end || - !bdrv_get_dirty(source, s->dirty_bitmap, next_sector)) { + !bdrv_get_dirty_locked(source, s->dirty_bitmap, next_sector)) { break; } if (test_bit(next_chunk, s->in_flight_bitmap)) { @@ -386,8 +389,10 @@ static uint64_t coroutine_fn mirror_iteration(MirrorBlockJob *s) * calling bdrv_get_block_status_above could yield - if some blocks are * marked dirty in this window, we need to know. */ - bdrv_reset_dirty_bitmap(s->dirty_bitmap, sector_num, - nb_chunks * sectors_per_chunk); + bdrv_reset_dirty_bitmap_locked(s->dirty_bitmap, sector_num, + nb_chunks * sectors_per_chunk); + bdrv_dirty_bitmap_unlock(s->dirty_bitmap); + bitmap_set(s->in_flight_bitmap, sector_num / sectors_per_chunk, nb_chunks); while (nb_chunks > 0 && sector_num < end) { int64_t ret; diff --git a/include/block/block_int.h b/include/block/block_int.h index 21cb65bd60..748970055e 100644 --- a/include/block/block_int.h +++ b/include/block/block_int.h @@ -611,8 +611,8 @@ struct BlockDriverState { /* Writing to the list requires the BQL _and_ the dirty_bitmap_mutex. * Reading from the list can be done with either the BQL or the - * dirty_bitmap_mutex. Modifying a bitmap requires the AioContext - * lock. */ + * dirty_bitmap_mutex. Modifying a bitmap only requires + * dirty_bitmap_mutex. */ QemuMutex dirty_bitmap_mutex; QLIST_HEAD(, BdrvDirtyBitmap) dirty_bitmaps; diff --git a/include/block/dirty-bitmap.h b/include/block/dirty-bitmap.h index 9dea14ba03..ad6558af56 100644 --- a/include/block/dirty-bitmap.h +++ b/include/block/dirty-bitmap.h @@ -36,8 +36,6 @@ bool bdrv_dirty_bitmap_frozen(BdrvDirtyBitmap *bitmap); const char *bdrv_dirty_bitmap_name(const BdrvDirtyBitmap *bitmap); int64_t bdrv_dirty_bitmap_size(const BdrvDirtyBitmap *bitmap); DirtyBitmapStatus bdrv_dirty_bitmap_status(BdrvDirtyBitmap *bitmap); -int bdrv_get_dirty(BlockDriverState *bs, BdrvDirtyBitmap *bitmap, - int64_t sector); void bdrv_set_dirty_bitmap(BdrvDirtyBitmap *bitmap, int64_t cur_sector, int64_t nr_sectors); void bdrv_reset_dirty_bitmap(BdrvDirtyBitmap *bitmap, @@ -45,6 +43,9 @@ void bdrv_reset_dirty_bitmap(BdrvDirtyBitmap *bitmap, int bdrv_dirty_bitmap_get_meta(BlockDriverState *bs, BdrvDirtyBitmap *bitmap, int64_t sector, int nb_sectors); +int bdrv_dirty_bitmap_get_meta_locked(BlockDriverState *bs, + BdrvDirtyBitmap *bitmap, int64_t sector, + int nb_sectors); void bdrv_dirty_bitmap_reset_meta(BlockDriverState *bs, BdrvDirtyBitmap *bitmap, int64_t sector, int nb_sectors); @@ -52,11 +53,6 @@ BdrvDirtyBitmapIter *bdrv_dirty_meta_iter_new(BdrvDirtyBitmap *bitmap); BdrvDirtyBitmapIter *bdrv_dirty_iter_new(BdrvDirtyBitmap *bitmap, uint64_t first_sector); void bdrv_dirty_iter_free(BdrvDirtyBitmapIter *iter); -int64_t bdrv_dirty_iter_next(BdrvDirtyBitmapIter *iter); -void bdrv_set_dirty_iter(BdrvDirtyBitmapIter *hbi, int64_t sector_num); -int64_t bdrv_get_dirty_count(BdrvDirtyBitmap *bitmap); -int64_t bdrv_get_meta_dirty_count(BdrvDirtyBitmap *bitmap); -void bdrv_dirty_bitmap_truncate(BlockDriverState *bs); uint64_t bdrv_dirty_bitmap_serialization_size(const BdrvDirtyBitmap *bitmap, uint64_t start, uint64_t count); @@ -72,4 +68,19 @@ void bdrv_dirty_bitmap_deserialize_zeroes(BdrvDirtyBitmap *bitmap, bool finish); void bdrv_dirty_bitmap_deserialize_finish(BdrvDirtyBitmap *bitmap); +/* Functions that require manual locking. */ +void bdrv_dirty_bitmap_lock(BdrvDirtyBitmap *bitmap); +void bdrv_dirty_bitmap_unlock(BdrvDirtyBitmap *bitmap); +int bdrv_get_dirty_locked(BlockDriverState *bs, BdrvDirtyBitmap *bitmap, + int64_t sector); +void bdrv_set_dirty_bitmap_locked(BdrvDirtyBitmap *bitmap, + int64_t cur_sector, int64_t nr_sectors); +void bdrv_reset_dirty_bitmap_locked(BdrvDirtyBitmap *bitmap, + int64_t cur_sector, int64_t nr_sectors); +int64_t bdrv_dirty_iter_next(BdrvDirtyBitmapIter *iter); +void bdrv_set_dirty_iter(BdrvDirtyBitmapIter *hbi, int64_t sector_num); +int64_t bdrv_get_dirty_count(BdrvDirtyBitmap *bitmap); +int64_t bdrv_get_meta_dirty_count(BdrvDirtyBitmap *bitmap); +void bdrv_dirty_bitmap_truncate(BlockDriverState *bs); + #endif diff --git a/migration/block.c b/migration/block.c index 423877bb40..7674ae1078 100644 --- a/migration/block.c +++ b/migration/block.c @@ -525,14 +525,15 @@ static int mig_save_device_dirty(QEMUFile *f, BlkMigDevState *bmds, } else { blk_mig_unlock(); } - if (bdrv_get_dirty(bs, bmds->dirty_bitmap, sector)) { - + bdrv_dirty_bitmap_lock(bmds->dirty_bitmap); + if (bdrv_get_dirty_locked(bs, bmds->dirty_bitmap, sector)) { if (total_sectors - sector < BDRV_SECTORS_PER_DIRTY_CHUNK) { nr_sectors = total_sectors - sector; } else { nr_sectors = BDRV_SECTORS_PER_DIRTY_CHUNK; } - bdrv_reset_dirty_bitmap(bmds->dirty_bitmap, sector, nr_sectors); + bdrv_reset_dirty_bitmap_locked(bmds->dirty_bitmap, sector, nr_sectors); + bdrv_dirty_bitmap_unlock(bmds->dirty_bitmap); blk = g_new(BlkMigBlock, 1); blk->buf = g_malloc(BLOCK_SIZE); @@ -568,9 +569,10 @@ static int mig_save_device_dirty(QEMUFile *f, BlkMigDevState *bmds, sector += nr_sectors; bmds->cur_dirty = sector; - break; } + + bdrv_dirty_bitmap_unlock(bmds->dirty_bitmap); sector += BDRV_SECTORS_PER_DIRTY_CHUNK; bmds->cur_dirty = sector; } From 39c1b4254e6200519d747f6cb8c888004e2446e2 Mon Sep 17 00:00:00 2001 From: Paolo Bonzini Date: Mon, 5 Jun 2017 14:39:06 +0200 Subject: [PATCH 21/23] block: introduce block_account_one_io This is the common code to account operations that produced actual I/O. Reviewed-by: Alberto Garcia Reviewed-by: Stefan Hajnoczi Signed-off-by: Paolo Bonzini Message-Id: <20170605123908.18777-18-pbonzini@redhat.com> Signed-off-by: Fam Zheng --- block/accounting.c | 51 ++++++++++++++++++++-------------------------- 1 file changed, 22 insertions(+), 29 deletions(-) diff --git a/block/accounting.c b/block/accounting.c index 3f457c4e73..a279e0b124 100644 --- a/block/accounting.c +++ b/block/accounting.c @@ -86,7 +86,8 @@ void block_acct_start(BlockAcctStats *stats, BlockAcctCookie *cookie, cookie->type = type; } -void block_acct_done(BlockAcctStats *stats, BlockAcctCookie *cookie) +static void block_account_one_io(BlockAcctStats *stats, BlockAcctCookie *cookie, + bool failed) { BlockAcctTimedStats *s; int64_t time_ns = qemu_clock_get_ns(clock_type); @@ -98,31 +99,14 @@ void block_acct_done(BlockAcctStats *stats, BlockAcctCookie *cookie) assert(cookie->type < BLOCK_MAX_IOTYPE); - stats->nr_bytes[cookie->type] += cookie->bytes; - stats->nr_ops[cookie->type]++; - stats->total_time_ns[cookie->type] += latency_ns; - stats->last_access_time_ns = time_ns; - - QSLIST_FOREACH(s, &stats->intervals, entries) { - timed_average_account(&s->latency[cookie->type], latency_ns); + if (failed) { + stats->failed_ops[cookie->type]++; + } else { + stats->nr_bytes[cookie->type] += cookie->bytes; + stats->nr_ops[cookie->type]++; } -} - -void block_acct_failed(BlockAcctStats *stats, BlockAcctCookie *cookie) -{ - assert(cookie->type < BLOCK_MAX_IOTYPE); - - stats->failed_ops[cookie->type]++; - - if (stats->account_failed) { - BlockAcctTimedStats *s; - int64_t time_ns = qemu_clock_get_ns(clock_type); - int64_t latency_ns = time_ns - cookie->start_time_ns; - - if (qtest_enabled()) { - latency_ns = qtest_latency_ns; - } + if (!failed || stats->account_failed) { stats->total_time_ns[cookie->type] += latency_ns; stats->last_access_time_ns = time_ns; @@ -132,15 +116,24 @@ void block_acct_failed(BlockAcctStats *stats, BlockAcctCookie *cookie) } } +void block_acct_done(BlockAcctStats *stats, BlockAcctCookie *cookie) +{ + block_account_one_io(stats, cookie, false); +} + +void block_acct_failed(BlockAcctStats *stats, BlockAcctCookie *cookie) +{ + block_account_one_io(stats, cookie, true); +} + void block_acct_invalid(BlockAcctStats *stats, enum BlockAcctType type) { assert(type < BLOCK_MAX_IOTYPE); - /* block_acct_done() and block_acct_failed() update - * total_time_ns[], but this one does not. The reason is that - * invalid requests are accounted during their submission, - * therefore there's no actual I/O involved. */ - + /* block_account_one_io() updates total_time_ns[], but this one does + * not. The reason is that invalid requests are accounted during their + * submission, therefore there's no actual I/O involved. + */ stats->invalid_ops[type]++; if (stats->account_invalid) { From 9caa6f3dbe20f2c506df6698386fce941fc6238a Mon Sep 17 00:00:00 2001 From: Paolo Bonzini Date: Mon, 5 Jun 2017 14:39:07 +0200 Subject: [PATCH 22/23] block: split BlockAcctStats creation and setup block_acct_destroy is called unconditionally in blk_delete, but there is no BlockAcctStats function that is called unconditionally in blk_new. Split block_acct_init in two, so that it will be possible to create a QemuMutex in block_acct_init and destroy it in block_acct_cleanup. Cc: Alberto Garcia Signed-off-by: Paolo Bonzini Message-Id: <20170605123908.18777-19-pbonzini@redhat.com> Reviewed-by: Alberto Garcia Signed-off-by: Fam Zheng --- block/accounting.c | 13 ++++++++----- block/block-backend.c | 1 + blockdev.c | 2 +- include/block/accounting.h | 3 ++- 4 files changed, 12 insertions(+), 7 deletions(-) diff --git a/block/accounting.c b/block/accounting.c index a279e0b124..ce6dbf7760 100644 --- a/block/accounting.c +++ b/block/accounting.c @@ -32,17 +32,20 @@ static QEMUClockType clock_type = QEMU_CLOCK_REALTIME; static const int qtest_latency_ns = NANOSECONDS_PER_SECOND / 1000; -void block_acct_init(BlockAcctStats *stats, bool account_invalid, - bool account_failed) +void block_acct_init(BlockAcctStats *stats) { - stats->account_invalid = account_invalid; - stats->account_failed = account_failed; - if (qtest_enabled()) { clock_type = QEMU_CLOCK_VIRTUAL; } } +void block_acct_setup(BlockAcctStats *stats, bool account_invalid, + bool account_failed) +{ + stats->account_invalid = account_invalid; + stats->account_failed = account_failed; +} + void block_acct_cleanup(BlockAcctStats *stats) { BlockAcctTimedStats *s, *next; diff --git a/block/block-backend.c b/block/block-backend.c index 738882dd2e..a2bbae90b1 100644 --- a/block/block-backend.c +++ b/block/block-backend.c @@ -219,6 +219,7 @@ BlockBackend *blk_new(uint64_t perm, uint64_t shared_perm) qemu_co_mutex_init(&blk->public.throttled_reqs_lock); qemu_co_queue_init(&blk->public.throttled_reqs[0]); qemu_co_queue_init(&blk->public.throttled_reqs[1]); + block_acct_init(&blk->stats); notifier_list_init(&blk->remove_bs_notifiers); notifier_list_init(&blk->insert_bs_notifiers); diff --git a/blockdev.c b/blockdev.c index 39f289ba68..39d6b9712b 100644 --- a/blockdev.c +++ b/blockdev.c @@ -595,7 +595,7 @@ static BlockBackend *blockdev_init(const char *file, QDict *bs_opts, autostart = 0; } - block_acct_init(blk_get_stats(blk), account_invalid, account_failed); + block_acct_setup(blk_get_stats(blk), account_invalid, account_failed); if (!parse_stats_intervals(blk_get_stats(blk), interval_list, errp)) { blk_unref(blk); diff --git a/include/block/accounting.h b/include/block/accounting.h index 20891639d5..55cb06fdb6 100644 --- a/include/block/accounting.h +++ b/include/block/accounting.h @@ -61,7 +61,8 @@ typedef struct BlockAcctCookie { enum BlockAcctType type; } BlockAcctCookie; -void block_acct_init(BlockAcctStats *stats, bool account_invalid, +void block_acct_init(BlockAcctStats *stats); +void block_acct_setup(BlockAcctStats *stats, bool account_invalid, bool account_failed); void block_acct_cleanup(BlockAcctStats *stats); void block_acct_add_interval(BlockAcctStats *stats, unsigned interval_length); From 5b50bf77ce6e773b04a303a2912876c9a1bfca43 Mon Sep 17 00:00:00 2001 From: Paolo Bonzini Date: Mon, 5 Jun 2017 14:39:08 +0200 Subject: [PATCH 23/23] block: make accounting thread-safe I'm not trying too hard yet. Later, with multiqueue support, this may cause mutex contention or cacheline bouncing. Cc: Alberto Garcia Reviewed-by: Stefan Hajnoczi Signed-off-by: Paolo Bonzini Message-Id: <20170605123908.18777-20-pbonzini@redhat.com> Reviewed-by: Alberto Garcia Signed-off-by: Fam Zheng --- block/accounting.c | 16 ++++++++++++++++ include/block/accounting.h | 8 ++++++-- 2 files changed, 22 insertions(+), 2 deletions(-) diff --git a/block/accounting.c b/block/accounting.c index ce6dbf7760..87ef5bbfaa 100644 --- a/block/accounting.c +++ b/block/accounting.c @@ -34,6 +34,7 @@ static const int qtest_latency_ns = NANOSECONDS_PER_SECOND / 1000; void block_acct_init(BlockAcctStats *stats) { + qemu_mutex_init(&stats->lock); if (qtest_enabled()) { clock_type = QEMU_CLOCK_VIRTUAL; } @@ -52,6 +53,7 @@ void block_acct_cleanup(BlockAcctStats *stats) QSLIST_FOREACH_SAFE(s, &stats->intervals, entries, next) { g_free(s); } + qemu_mutex_destroy(&stats->lock); } void block_acct_add_interval(BlockAcctStats *stats, unsigned interval_length) @@ -61,12 +63,15 @@ void block_acct_add_interval(BlockAcctStats *stats, unsigned interval_length) s = g_new0(BlockAcctTimedStats, 1); s->interval_length = interval_length; + s->stats = stats; + qemu_mutex_lock(&stats->lock); QSLIST_INSERT_HEAD(&stats->intervals, s, entries); for (i = 0; i < BLOCK_MAX_IOTYPE; i++) { timed_average_init(&s->latency[i], clock_type, (uint64_t) interval_length * NANOSECONDS_PER_SECOND); } + qemu_mutex_unlock(&stats->lock); } BlockAcctTimedStats *block_acct_interval_next(BlockAcctStats *stats, @@ -102,6 +107,8 @@ static void block_account_one_io(BlockAcctStats *stats, BlockAcctCookie *cookie, assert(cookie->type < BLOCK_MAX_IOTYPE); + qemu_mutex_lock(&stats->lock); + if (failed) { stats->failed_ops[cookie->type]++; } else { @@ -117,6 +124,8 @@ static void block_account_one_io(BlockAcctStats *stats, BlockAcctCookie *cookie, timed_average_account(&s->latency[cookie->type], latency_ns); } } + + qemu_mutex_unlock(&stats->lock); } void block_acct_done(BlockAcctStats *stats, BlockAcctCookie *cookie) @@ -137,18 +146,23 @@ void block_acct_invalid(BlockAcctStats *stats, enum BlockAcctType type) * not. The reason is that invalid requests are accounted during their * submission, therefore there's no actual I/O involved. */ + qemu_mutex_lock(&stats->lock); stats->invalid_ops[type]++; if (stats->account_invalid) { stats->last_access_time_ns = qemu_clock_get_ns(clock_type); } + qemu_mutex_unlock(&stats->lock); } void block_acct_merge_done(BlockAcctStats *stats, enum BlockAcctType type, int num_requests) { assert(type < BLOCK_MAX_IOTYPE); + + qemu_mutex_lock(&stats->lock); stats->merged[type] += num_requests; + qemu_mutex_unlock(&stats->lock); } int64_t block_acct_idle_time_ns(BlockAcctStats *stats) @@ -163,7 +177,9 @@ double block_acct_queue_depth(BlockAcctTimedStats *stats, assert(type < BLOCK_MAX_IOTYPE); + qemu_mutex_lock(&stats->stats->lock); sum = timed_average_sum(&stats->latency[type], &elapsed); + qemu_mutex_unlock(&stats->stats->lock); return (double) sum / elapsed; } diff --git a/include/block/accounting.h b/include/block/accounting.h index 55cb06fdb6..b833d26d6c 100644 --- a/include/block/accounting.h +++ b/include/block/accounting.h @@ -26,8 +26,10 @@ #define BLOCK_ACCOUNTING_H #include "qemu/timed-average.h" +#include "qemu/thread.h" typedef struct BlockAcctTimedStats BlockAcctTimedStats; +typedef struct BlockAcctStats BlockAcctStats; enum BlockAcctType { BLOCK_ACCT_READ, @@ -37,12 +39,14 @@ enum BlockAcctType { }; struct BlockAcctTimedStats { + BlockAcctStats *stats; TimedAverage latency[BLOCK_MAX_IOTYPE]; unsigned interval_length; /* in seconds */ QSLIST_ENTRY(BlockAcctTimedStats) entries; }; -typedef struct BlockAcctStats { +struct BlockAcctStats { + QemuMutex lock; uint64_t nr_bytes[BLOCK_MAX_IOTYPE]; uint64_t nr_ops[BLOCK_MAX_IOTYPE]; uint64_t invalid_ops[BLOCK_MAX_IOTYPE]; @@ -53,7 +57,7 @@ typedef struct BlockAcctStats { QSLIST_HEAD(, BlockAcctTimedStats) intervals; bool account_invalid; bool account_failed; -} BlockAcctStats; +}; typedef struct BlockAcctCookie { int64_t bytes;