:p
atchew
Login
The following changes since commit b2896c1b09878fd1c4b485b3662f8beecbe0fef4: Merge tag 'vfio-updates-20230509.0' of https://gitlab.com/alex.williamson/qemu into staging (2023-05-10 11:20:35 +0100) are available in the Git repository at: https://repo.or.cz/qemu/kevin.git tags/for-upstream for you to fetch changes up to 58a2e3f5c37be02dac3086b81bdda9414b931edf: block: compile out assert_bdrv_graph_readable() by default (2023-05-10 14:16:54 +0200) ---------------------------------------------------------------- Block layer patches - Graph locking, part 3 (more block drivers) - Compile out assert_bdrv_graph_readable() by default - Add configure options for vmdk, vhdx and vpc - Fix use after free in blockdev_mark_auto_del() - migration: Attempt disk reactivation in more failure scenarios - Coroutine correctness fixes ---------------------------------------------------------------- Emanuele Giuseppe Esposito (5): nbd: Mark nbd_co_do_establish_connection() and callers GRAPH_RDLOCK block: Mark bdrv_co_get_allocated_file_size() and callers GRAPH_RDLOCK block: Mark bdrv_co_get_info() and callers GRAPH_RDLOCK block: Mark bdrv_co_debug_event() GRAPH_RDLOCK block: Mark BlockDriver callbacks for amend job GRAPH_RDLOCK Eric Blake (1): migration: Attempt disk reactivation in more failure scenarios Kevin Wolf (18): block: Fix use after free in blockdev_mark_auto_del() iotests/nbd-reconnect-on-open: Fix NBD socket path qcow2: Don't call bdrv_getlength() in coroutine_fns block: Consistently call bdrv_activate() outside coroutine block: bdrv/blk_co_unref() for calls in coroutine context block: Don't call no_coroutine_fns in qmp_block_resize() iotests: Test resizing image attached to an iothread test-bdrv-drain: Don't modify the graph in coroutines graph-lock: Add GRAPH_UNLOCKED(_PTR) graph-lock: Fix GRAPH_RDLOCK_GUARD*() to be reader lock block: .bdrv_open is non-coroutine and unlocked nbd: Remove nbd_co_flush() wrapper function vhdx: Require GRAPH_RDLOCK for accessing a node's parent list mirror: Require GRAPH_RDLOCK for accessing a node's parent list block: Mark bdrv_query_bds_stats() and callers GRAPH_RDLOCK block: Mark bdrv_query_block_graph_info() and callers GRAPH_RDLOCK block: Mark bdrv_recurse_can_replace() and callers GRAPH_RDLOCK block: Mark bdrv_refresh_limits() and callers GRAPH_RDLOCK Paolo Bonzini (1): block: add missing coroutine_fn annotations Stefan Hajnoczi (2): aio-wait: avoid AioContext lock in aio_wait_bh_oneshot() block: compile out assert_bdrv_graph_readable() by default Vladimir Sementsov-Ogievskiy (1): block: add configure options for excluding vmdk, vhdx and vpc meson_options.txt | 8 ++ configure | 1 + block/coroutines.h | 5 +- block/qcow2.h | 4 +- include/block/aio-wait.h | 2 +- include/block/block-global-state.h | 19 +++- include/block/block-io.h | 23 +++-- include/block/block_int-common.h | 37 +++---- include/block/block_int-global-state.h | 4 +- include/block/graph-lock.h | 20 ++-- include/block/qapi.h | 7 +- include/sysemu/block-backend-global-state.h | 5 +- block.c | 25 ++++- block/amend.c | 8 +- block/blkverify.c | 5 +- block/block-backend.c | 10 +- block/crypto.c | 8 +- block/graph-lock.c | 3 + block/io.c | 12 +-- block/mirror.c | 18 +++- block/nbd.c | 50 +++++---- block/parallels.c | 6 +- block/qapi.c | 6 +- block/qcow.c | 6 +- block/qcow2-refcount.c | 2 +- block/qcow2.c | 48 ++++----- block/qed.c | 24 ++--- block/quorum.c | 4 +- block/raw-format.c | 2 +- block/vdi.c | 6 +- block/vhdx.c | 15 +-- block/vmdk.c | 20 ++-- block/vpc.c | 6 +- blockdev.c | 25 +++-- hw/block/dataplane/virtio-blk.c | 3 +- hw/scsi/virtio-scsi-dataplane.c | 2 - migration/migration.c | 24 +++-- qemu-img.c | 2 + tests/unit/test-bdrv-drain.c | 112 ++++++++++++++------- util/aio-wait.c | 2 +- block/meson.build | 18 +++- meson.build | 5 + scripts/meson-buildoptions.sh | 13 +++ tests/qemu-iotests/tests/iothreads-resize | 71 +++++++++++++ tests/qemu-iotests/tests/iothreads-resize.out | 11 ++ tests/qemu-iotests/tests/nbd-reconnect-on-open | 3 +- tests/qemu-iotests/tests/nbd-reconnect-on-open.out | 4 +- 47 files changed, 474 insertions(+), 240 deletions(-) create mode 100755 tests/qemu-iotests/tests/iothreads-resize create mode 100644 tests/qemu-iotests/tests/iothreads-resize.out
From: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru> Let's add --enable / --disable configure options for these formats, so that those who don't need them may not build them. Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru> Message-Id: <20230421092758.814122-1-vsementsov@yandex-team.ru> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- meson_options.txt | 6 ++++++ block/meson.build | 18 +++++++++++++----- meson.build | 3 +++ scripts/meson-buildoptions.sh | 9 +++++++++ 4 files changed, 31 insertions(+), 5 deletions(-) diff --git a/meson_options.txt b/meson_options.txt index XXXXXXX..XXXXXXX 100644 --- a/meson_options.txt +++ b/meson_options.txt @@ -XXX,XX +XXX,XX @@ option('qcow1', type: 'feature', value: 'auto', description: 'qcow1 image format support') option('vdi', type: 'feature', value: 'auto', description: 'vdi image format support') +option('vhdx', type: 'feature', value: 'auto', + description: 'vhdx image format support') +option('vmdk', type: 'feature', value: 'auto', + description: 'vmdk image format support') +option('vpc', type: 'feature', value: 'auto', + description: 'vpc image format support') option('vvfat', type: 'feature', value: 'auto', description: 'vvfat image format support') option('qed', type: 'feature', value: 'auto', diff --git a/block/meson.build b/block/meson.build index XXXXXXX..XXXXXXX 100644 --- a/block/meson.build +++ b/block/meson.build @@ -XXX,XX +XXX,XX @@ block_ss.add(files( 'snapshot-access.c', 'throttle-groups.c', 'throttle.c', - 'vhdx-endian.c', - 'vhdx-log.c', - 'vhdx.c', - 'vmdk.c', - 'vpc.c', 'write-threshold.c', ), zstd, zlib, gnutls) @@ -XXX,XX +XXX,XX @@ endif if get_option('vdi').allowed() block_ss.add(files('vdi.c')) endif +if get_option('vhdx').allowed() + block_ss.add(files( + 'vhdx-endian.c', + 'vhdx-log.c', + 'vhdx.c' + )) +endif +if get_option('vmdk').allowed() + block_ss.add(files('vmdk.c')) +endif +if get_option('vpc').allowed() + block_ss.add(files('vpc.c')) +endif if get_option('cloop').allowed() block_ss.add(files('cloop.c')) endif diff --git a/meson.build b/meson.build index XXXXXXX..XXXXXXX 100644 --- a/meson.build +++ b/meson.build @@ -XXX,XX +XXX,XX @@ if have_block summary_info += {'dmg support': get_option('dmg').allowed()} summary_info += {'qcow v1 support': get_option('qcow1').allowed()} summary_info += {'vdi support': get_option('vdi').allowed()} + summary_info += {'vhdx support': get_option('vhdx').allowed()} + summary_info += {'vmdk support': get_option('vmdk').allowed()} + summary_info += {'vpc support': get_option('vpc').allowed()} summary_info += {'vvfat support': get_option('vvfat').allowed()} summary_info += {'qed support': get_option('qed').allowed()} summary_info += {'parallels support': get_option('parallels').allowed()} diff --git a/scripts/meson-buildoptions.sh b/scripts/meson-buildoptions.sh index XXXXXXX..XXXXXXX 100644 --- a/scripts/meson-buildoptions.sh +++ b/scripts/meson-buildoptions.sh @@ -XXX,XX +XXX,XX @@ meson_options_help() { printf "%s\n" ' VDUSE block export support' printf "%s\n" ' vfio-user-server' printf "%s\n" ' vfio-user server support' + printf "%s\n" ' vhdx vhdx image format support' printf "%s\n" ' vhost-crypto vhost-user crypto backend support' printf "%s\n" ' vhost-kernel vhost kernel backend support' printf "%s\n" ' vhost-net vhost-net kernel acceleration support' @@ -XXX,XX +XXX,XX @@ meson_options_help() { printf "%s\n" ' vhost-vdpa vhost-vdpa kernel backend support' printf "%s\n" ' virglrenderer virgl rendering support' printf "%s\n" ' virtfs virtio-9p support' + printf "%s\n" ' vmdk vmdk image format support' printf "%s\n" ' vmnet vmnet.framework network backend support' printf "%s\n" ' vnc VNC server' printf "%s\n" ' vnc-jpeg JPEG lossy compression for VNC server' printf "%s\n" ' vnc-sasl SASL authentication for VNC server' + printf "%s\n" ' vpc vpc image format support' printf "%s\n" ' vte vte support for the gtk UI' printf "%s\n" ' vvfat vvfat image format support' printf "%s\n" ' whpx WHPX acceleration support' @@ -XXX,XX +XXX,XX @@ _meson_option_parse() { --disable-vduse-blk-export) printf "%s" -Dvduse_blk_export=disabled ;; --enable-vfio-user-server) printf "%s" -Dvfio_user_server=enabled ;; --disable-vfio-user-server) printf "%s" -Dvfio_user_server=disabled ;; + --enable-vhdx) printf "%s" -Dvhdx=enabled ;; + --disable-vhdx) printf "%s" -Dvhdx=disabled ;; --enable-vhost-crypto) printf "%s" -Dvhost_crypto=enabled ;; --disable-vhost-crypto) printf "%s" -Dvhost_crypto=disabled ;; --enable-vhost-kernel) printf "%s" -Dvhost_kernel=enabled ;; @@ -XXX,XX +XXX,XX @@ _meson_option_parse() { --disable-virglrenderer) printf "%s" -Dvirglrenderer=disabled ;; --enable-virtfs) printf "%s" -Dvirtfs=enabled ;; --disable-virtfs) printf "%s" -Dvirtfs=disabled ;; + --enable-vmdk) printf "%s" -Dvmdk=enabled ;; + --disable-vmdk) printf "%s" -Dvmdk=disabled ;; --enable-vmnet) printf "%s" -Dvmnet=enabled ;; --disable-vmnet) printf "%s" -Dvmnet=disabled ;; --enable-vnc) printf "%s" -Dvnc=enabled ;; @@ -XXX,XX +XXX,XX @@ _meson_option_parse() { --disable-vnc-jpeg) printf "%s" -Dvnc_jpeg=disabled ;; --enable-vnc-sasl) printf "%s" -Dvnc_sasl=enabled ;; --disable-vnc-sasl) printf "%s" -Dvnc_sasl=disabled ;; + --enable-vpc) printf "%s" -Dvpc=enabled ;; + --disable-vpc) printf "%s" -Dvpc=disabled ;; --enable-vte) printf "%s" -Dvte=enabled ;; --disable-vte) printf "%s" -Dvte=disabled ;; --enable-vvfat) printf "%s" -Dvvfat=enabled ;; -- 2.40.1
From: Paolo Bonzini <pbonzini@redhat.com> After the recent introduction of many new coroutine callbacks, a couple calls from non-coroutine_fn to coroutine_fn have sneaked in; fix them. Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Message-Id: <20230406101752.242125-1-pbonzini@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/graph-lock.h | 4 ++-- block/mirror.c | 4 ++-- 2 files changed, 4 insertions(+), 4 deletions(-) diff --git a/include/block/graph-lock.h b/include/block/graph-lock.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/graph-lock.h +++ b/include/block/graph-lock.h @@ -XXX,XX +XXX,XX @@ typedef struct GraphLockable { } GraphLockable; * unlocked. TSA_ASSERT() makes sure that the following calls know that we * hold the lock while unlocking is left unchecked. */ -static inline GraphLockable * TSA_ASSERT(graph_lock) TSA_NO_TSA +static inline GraphLockable * TSA_ASSERT(graph_lock) TSA_NO_TSA coroutine_fn graph_lockable_auto_lock(GraphLockable *x) { bdrv_graph_co_rdlock(); return x; } -static inline void TSA_NO_TSA +static inline void TSA_NO_TSA coroutine_fn graph_lockable_auto_unlock(GraphLockable *x) { bdrv_graph_co_rdunlock(); diff --git a/block/mirror.c b/block/mirror.c index XXXXXXX..XXXXXXX 100644 --- a/block/mirror.c +++ b/block/mirror.c @@ -XXX,XX +XXX,XX @@ static inline int64_t mirror_clip_bytes(MirrorBlockJob *s, /* Round offset and/or bytes to target cluster if COW is needed, and * return the offset of the adjusted tail against original. */ -static int mirror_cow_align(MirrorBlockJob *s, int64_t *offset, - uint64_t *bytes) +static int coroutine_fn mirror_cow_align(MirrorBlockJob *s, int64_t *offset, + uint64_t *bytes) { bool need_cow; int ret = 0; -- 2.40.1
From: Stefan Hajnoczi <stefanha@redhat.com> There is no need for the AioContext lock in aio_wait_bh_oneshot(). It's easy to remove the lock from existing callers and then switch from AIO_WAIT_WHILE() to AIO_WAIT_WHILE_UNLOCKED() in aio_wait_bh_oneshot(). Document that the AioContext lock should not be held across aio_wait_bh_oneshot(). Holding a lock across aio_poll() can cause deadlock so we don't want callers to do that. This is a step towards getting rid of the AioContext lock. Cc: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Message-Id: <20230404153307.458883-1-stefanha@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/aio-wait.h | 2 +- hw/block/dataplane/virtio-blk.c | 3 ++- hw/scsi/virtio-scsi-dataplane.c | 2 -- util/aio-wait.c | 2 +- 4 files changed, 4 insertions(+), 5 deletions(-) diff --git a/include/block/aio-wait.h b/include/block/aio-wait.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/aio-wait.h +++ b/include/block/aio-wait.h @@ -XXX,XX +XXX,XX @@ void aio_wait_kick(void); * * Run a BH in @ctx and wait for it to complete. * - * Must be called from the main loop thread with @ctx acquired exactly once. + * Must be called from the main loop thread without @ctx acquired. * Note that main loop event processing may occur. */ void aio_wait_bh_oneshot(AioContext *ctx, QEMUBHFunc *cb, void *opaque); diff --git a/hw/block/dataplane/virtio-blk.c b/hw/block/dataplane/virtio-blk.c index XXXXXXX..XXXXXXX 100644 --- a/hw/block/dataplane/virtio-blk.c +++ b/hw/block/dataplane/virtio-blk.c @@ -XXX,XX +XXX,XX @@ void virtio_blk_data_plane_stop(VirtIODevice *vdev) s->stopping = true; trace_virtio_blk_data_plane_stop(s); - aio_context_acquire(s->ctx); aio_wait_bh_oneshot(s->ctx, virtio_blk_data_plane_stop_bh, s); + aio_context_acquire(s->ctx); + /* Wait for virtio_blk_dma_restart_bh() and in flight I/O to complete */ blk_drain(s->conf->conf.blk); diff --git a/hw/scsi/virtio-scsi-dataplane.c b/hw/scsi/virtio-scsi-dataplane.c index XXXXXXX..XXXXXXX 100644 --- a/hw/scsi/virtio-scsi-dataplane.c +++ b/hw/scsi/virtio-scsi-dataplane.c @@ -XXX,XX +XXX,XX @@ void virtio_scsi_dataplane_stop(VirtIODevice *vdev) } s->dataplane_stopping = true; - aio_context_acquire(s->ctx); aio_wait_bh_oneshot(s->ctx, virtio_scsi_dataplane_stop_bh, s); - aio_context_release(s->ctx); blk_drain_all(); /* ensure there are no in-flight requests */ diff --git a/util/aio-wait.c b/util/aio-wait.c index XXXXXXX..XXXXXXX 100644 --- a/util/aio-wait.c +++ b/util/aio-wait.c @@ -XXX,XX +XXX,XX @@ void aio_wait_bh_oneshot(AioContext *ctx, QEMUBHFunc *cb, void *opaque) assert(qemu_get_current_aio_context() == qemu_get_aio_context()); aio_bh_schedule_oneshot(ctx, aio_wait_bh, &data); - AIO_WAIT_WHILE(ctx, !data.done); + AIO_WAIT_WHILE_UNLOCKED(NULL, !data.done); } -- 2.40.1
job_cancel_locked() drops the job list lock temporarily and it may call aio_poll(). We must assume that the list has changed after this call. Also, with unlucky timing, it can end up freeing the job during job_completed_txn_abort_locked(), making the job pointer invalid, too. For both reasons, we can't just continue at block_job_next_locked(job). Instead, start at the head of the list again after job_cancel_locked() and skip those jobs that we already cancelled (or that are completing anyway). Cc: qemu-stable@nongnu.org Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230503140142.474404-1-kwolf@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- blockdev.c | 18 ++++++++++++++---- 1 file changed, 14 insertions(+), 4 deletions(-) diff --git a/blockdev.c b/blockdev.c index XXXXXXX..XXXXXXX 100644 --- a/blockdev.c +++ b/blockdev.c @@ -XXX,XX +XXX,XX @@ void blockdev_mark_auto_del(BlockBackend *blk) JOB_LOCK_GUARD(); - for (job = block_job_next_locked(NULL); job; - job = block_job_next_locked(job)) { - if (block_job_has_bdrv(job, blk_bs(blk))) { + do { + job = block_job_next_locked(NULL); + while (job && (job->job.cancelled || + job->job.deferred_to_main_loop || + !block_job_has_bdrv(job, blk_bs(blk)))) + { + job = block_job_next_locked(job); + } + if (job) { + /* + * This drops the job lock temporarily and polls, so we need to + * restart processing the list from the start after this. + */ job_cancel_locked(&job->job, false); } - } + } while (job); dinfo->auto_del = 1; } -- 2.40.1
Socket paths need to be short to avoid failures. This is why there is a iotests.sock_dir (defaulting to /tmp) separate from the disk image base directory. Make use of it to fix failures in too deeply nested test directories. Fixes: ab7f7e67a7e7b49964109501dfcde4ec29bae60e Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230503165019.8867-1-kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- tests/qemu-iotests/tests/nbd-reconnect-on-open | 3 ++- tests/qemu-iotests/tests/nbd-reconnect-on-open.out | 4 ++-- 2 files changed, 4 insertions(+), 3 deletions(-) diff --git a/tests/qemu-iotests/tests/nbd-reconnect-on-open b/tests/qemu-iotests/tests/nbd-reconnect-on-open index XXXXXXX..XXXXXXX 100755 --- a/tests/qemu-iotests/tests/nbd-reconnect-on-open +++ b/tests/qemu-iotests/tests/nbd-reconnect-on-open @@ -XXX,XX +XXX,XX @@ from iotests import qemu_img_create, file_path, qemu_io_popen, qemu_nbd, \ iotests.script_initialize(supported_fmts=['qcow2']) -disk, nbd_sock = file_path('disk', 'nbd-sock') +disk = file_path('disk') +nbd_sock = file_path('nbd-sock', base_dir=iotests.sock_dir) def create_args(open_timeout): diff --git a/tests/qemu-iotests/tests/nbd-reconnect-on-open.out b/tests/qemu-iotests/tests/nbd-reconnect-on-open.out index XXXXXXX..XXXXXXX 100644 --- a/tests/qemu-iotests/tests/nbd-reconnect-on-open.out +++ b/tests/qemu-iotests/tests/nbd-reconnect-on-open.out @@ -XXX,XX +XXX,XX @@ read 1048576/1048576 bytes at offset 0 1 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) Check fail to connect with 0 seconds of timeout -qemu-io: can't open: Failed to connect to 'TEST_DIR/PID-nbd-sock': No such file or directory +qemu-io: can't open: Failed to connect to 'SOCK_DIR/PID-nbd-sock': No such file or directory qemu_io finished in 0..0.2 seconds, OK Check fail to connect with 1 seconds of timeout -qemu-io: can't open: Failed to connect to 'TEST_DIR/PID-nbd-sock': No such file or directory +qemu-io: can't open: Failed to connect to 'SOCK_DIR/PID-nbd-sock': No such file or directory qemu_io finished in 1..1.2 seconds, OK -- 2.40.1
From: Eric Blake <eblake@redhat.com> Commit fe904ea824 added a fail_inactivate label, which tries to reactivate disks on the source after a failure while s->state == MIGRATION_STATUS_ACTIVE, but didn't actually use the label if qemu_savevm_state_complete_precopy() failed. This failure to reactivate is also present in commit 6039dd5b1c (also covering the new s->state == MIGRATION_STATUS_DEVICE state) and 403d18ae (ensuring s->block_inactive is set more reliably). Consolidate the two labels back into one - no matter HOW migration is failed, if there is any chance we can reach vm_start() after having attempted inactivation, it is essential that we have tried to restart disks before then. This also makes the cleanup more like migrate_fd_cancel(). Suggested-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Eric Blake <eblake@redhat.com> Message-Id: <20230502205212.134680-1-eblake@redhat.com> Acked-by: Peter Xu <peterx@redhat.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- migration/migration.c | 24 ++++++++++++++---------- 1 file changed, 14 insertions(+), 10 deletions(-) diff --git a/migration/migration.c b/migration/migration.c index XXXXXXX..XXXXXXX 100644 --- a/migration/migration.c +++ b/migration/migration.c @@ -XXX,XX +XXX,XX @@ static void migration_completion(MigrationState *s) MIGRATION_STATUS_DEVICE); } if (ret >= 0) { + /* + * Inactivate disks except in COLO, and track that we + * have done so in order to remember to reactivate + * them if migration fails or is cancelled. + */ s->block_inactive = !migrate_colo(); qemu_file_set_rate_limit(s->to_dst_file, INT64_MAX); ret = qemu_savevm_state_complete_precopy(s->to_dst_file, false, @@ -XXX,XX +XXX,XX @@ static void migration_completion(MigrationState *s) rp_error = await_return_path_close_on_source(s); trace_migration_return_path_end_after(rp_error); if (rp_error) { - goto fail_invalidate; + goto fail; } } if (qemu_file_get_error(s->to_dst_file)) { trace_migration_completion_file_err(); - goto fail_invalidate; + goto fail; } if (migrate_colo() && s->state == MIGRATION_STATUS_ACTIVE) { @@ -XXX,XX +XXX,XX @@ static void migration_completion(MigrationState *s) return; -fail_invalidate: - /* If not doing postcopy, vm_start() will be called: let's regain - * control on images. - */ - if (s->state == MIGRATION_STATUS_ACTIVE || - s->state == MIGRATION_STATUS_DEVICE) { +fail: + if (s->block_inactive && (s->state == MIGRATION_STATUS_ACTIVE || + s->state == MIGRATION_STATUS_DEVICE)) { + /* + * If not doing postcopy, vm_start() will be called: let's + * regain control on images. + */ Error *local_err = NULL; qemu_mutex_lock_iothread(); bdrv_activate_all(&local_err); if (local_err) { error_report_err(local_err); - s->block_inactive = true; } else { s->block_inactive = false; } qemu_mutex_unlock_iothread(); } -fail: migrate_set_state(&s->state, current_active_state, MIGRATION_STATUS_FAILED); } -- 2.40.1
There is a bdrv_co_getlength() now, which should be used in coroutine context. This requires adding GRAPH_RDLOCK to some functions so that this still compiles with TSA because bdrv_co_getlength() is GRAPH_RDLOCK. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Message-Id: <20230504115750.54437-2-kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/qcow2.h | 4 +++- block/qcow2-refcount.c | 2 +- block/qcow2.c | 19 +++++++++---------- 3 files changed, 13 insertions(+), 12 deletions(-) diff --git a/block/qcow2.h b/block/qcow2.h index XXXXXXX..XXXXXXX 100644 --- a/block/qcow2.h +++ b/block/qcow2.h @@ -XXX,XX +XXX,XX @@ int qcow2_change_refcount_order(BlockDriverState *bs, int refcount_order, void *cb_opaque, Error **errp); int coroutine_fn GRAPH_RDLOCK qcow2_shrink_reftable(BlockDriverState *bs); int64_t coroutine_fn qcow2_get_last_cluster(BlockDriverState *bs, int64_t size); -int coroutine_fn qcow2_detect_metadata_preallocation(BlockDriverState *bs); + +int coroutine_fn GRAPH_RDLOCK +qcow2_detect_metadata_preallocation(BlockDriverState *bs); /* qcow2-cluster.c functions */ int qcow2_grow_l1_table(BlockDriverState *bs, uint64_t min_size, diff --git a/block/qcow2-refcount.c b/block/qcow2-refcount.c index XXXXXXX..XXXXXXX 100644 --- a/block/qcow2-refcount.c +++ b/block/qcow2-refcount.c @@ -XXX,XX +XXX,XX @@ int coroutine_fn qcow2_detect_metadata_preallocation(BlockDriverState *bs) qemu_co_mutex_assert_locked(&s->lock); - file_length = bdrv_getlength(bs->file->bs); + file_length = bdrv_co_getlength(bs->file->bs); if (file_length < 0) { return file_length; } diff --git a/block/qcow2.c b/block/qcow2.c index XXXXXXX..XXXXXXX 100644 --- a/block/qcow2.c +++ b/block/qcow2.c @@ -XXX,XX +XXX,XX @@ static void qcow2_join_options(QDict *options, QDict *old_options) } } -static int coroutine_fn qcow2_co_block_status(BlockDriverState *bs, - bool want_zero, - int64_t offset, int64_t count, - int64_t *pnum, int64_t *map, - BlockDriverState **file) +static int coroutine_fn GRAPH_RDLOCK +qcow2_co_block_status(BlockDriverState *bs, bool want_zero, int64_t offset, + int64_t count, int64_t *pnum, int64_t *map, + BlockDriverState **file) { BDRVQcow2State *s = bs->opaque; uint64_t host_offset; @@ -XXX,XX +XXX,XX @@ preallocate_co(BlockDriverState *bs, uint64_t offset, uint64_t new_length, * all of the allocated clusters (otherwise we get failing reads after * EOF). Extend the image to the last allocated sector. */ - file_length = bdrv_getlength(s->data_file->bs); + file_length = bdrv_co_getlength(s->data_file->bs); if (file_length < 0) { error_setg_errno(errp, -file_length, "Could not get file size"); ret = file_length; @@ -XXX,XX +XXX,XX @@ qcow2_co_copy_range_from(BlockDriverState *bs, case QCOW2_SUBCLUSTER_UNALLOCATED_PLAIN: case QCOW2_SUBCLUSTER_UNALLOCATED_ALLOC: if (bs->backing && bs->backing->bs) { - int64_t backing_length = bdrv_getlength(bs->backing->bs); + int64_t backing_length = bdrv_co_getlength(bs->backing->bs); if (src_offset >= backing_length) { cur_write_flags |= BDRV_REQ_ZERO_WRITE; } else { @@ -XXX,XX +XXX,XX @@ qcow2_co_truncate(BlockDriverState *bs, int64_t offset, bool exact, goto fail; } - old_file_size = bdrv_getlength(bs->file->bs); + old_file_size = bdrv_co_getlength(bs->file->bs); if (old_file_size < 0) { error_setg_errno(errp, -old_file_size, "Failed to inquire current file length"); @@ -XXX,XX +XXX,XX @@ qcow2_co_truncate(BlockDriverState *bs, int64_t offset, bool exact, break; } - old_file_size = bdrv_getlength(bs->file->bs); + old_file_size = bdrv_co_getlength(bs->file->bs); if (old_file_size < 0) { error_setg_errno(errp, -old_file_size, "Failed to inquire current file length"); @@ -XXX,XX +XXX,XX @@ qcow2_co_pwritev_compressed_part(BlockDriverState *bs, * align end of file to a sector boundary to ease reading with * sector based I/Os */ - int64_t len = bdrv_getlength(bs->file->bs); + int64_t len = bdrv_co_getlength(bs->file->bs); if (len < 0) { return len; } -- 2.40.1
Migration code can call bdrv_activate() in coroutine context, whereas other callers call it outside of coroutines. As it calls other code that is not supposed to run in coroutines, standardise on running outside of coroutines. This adds a no_co_wrapper to switch to the main loop before calling bdrv_activate(). Cc: qemu-stable@nongnu.org Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Message-Id: <20230504115750.54437-3-kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-global-state.h | 6 +++++- block/block-backend.c | 10 +++++++++- 2 files changed, 14 insertions(+), 2 deletions(-) diff --git a/include/block/block-global-state.h b/include/block/block-global-state.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block-global-state.h +++ b/include/block/block-global-state.h @@ -XXX,XX +XXX,XX @@ int bdrv_amend_options(BlockDriverState *bs_new, QemuOpts *opts, BlockDriverState *check_to_replace_node(BlockDriverState *parent_bs, const char *node_name, Error **errp); -int bdrv_activate(BlockDriverState *bs, Error **errp); +int no_coroutine_fn bdrv_activate(BlockDriverState *bs, Error **errp); + +int coroutine_fn no_co_wrapper +bdrv_co_activate(BlockDriverState *bs, Error **errp); + void bdrv_activate_all(Error **errp); int bdrv_inactivate_all(void); diff --git a/block/block-backend.c b/block/block-backend.c index XXXXXXX..XXXXXXX 100644 --- a/block/block-backend.c +++ b/block/block-backend.c @@ -XXX,XX +XXX,XX @@ void blk_activate(BlockBackend *blk, Error **errp) return; } - bdrv_activate(bs, errp); + /* + * Migration code can call this function in coroutine context, so leave + * coroutine context if necessary. + */ + if (qemu_in_coroutine()) { + bdrv_co_activate(bs, errp); + } else { + bdrv_activate(bs, errp); + } } bool coroutine_fn blk_co_is_inserted(BlockBackend *blk) -- 2.40.1
These functions must not be called in coroutine context, because they need write access to the graph. Cc: qemu-stable@nongnu.org Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Message-Id: <20230504115750.54437-4-kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-global-state.h | 3 ++- include/sysemu/block-backend-global-state.h | 5 ++++- block.c | 2 +- block/crypto.c | 6 +++--- block/parallels.c | 6 +++--- block/qcow.c | 6 +++--- block/qcow2.c | 14 +++++++------- block/qed.c | 6 +++--- block/vdi.c | 6 +++--- block/vhdx.c | 6 +++--- block/vmdk.c | 18 +++++++++--------- block/vpc.c | 6 +++--- 12 files changed, 44 insertions(+), 40 deletions(-) diff --git a/include/block/block-global-state.h b/include/block/block-global-state.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block-global-state.h +++ b/include/block/block-global-state.h @@ -XXX,XX +XXX,XX @@ void bdrv_img_create(const char *filename, const char *fmt, bool quiet, Error **errp); void bdrv_ref(BlockDriverState *bs); -void bdrv_unref(BlockDriverState *bs); +void no_coroutine_fn bdrv_unref(BlockDriverState *bs); +void coroutine_fn no_co_wrapper bdrv_co_unref(BlockDriverState *bs); void bdrv_unref_child(BlockDriverState *parent, BdrvChild *child); BdrvChild *bdrv_attach_child(BlockDriverState *parent_bs, BlockDriverState *child_bs, diff --git a/include/sysemu/block-backend-global-state.h b/include/sysemu/block-backend-global-state.h index XXXXXXX..XXXXXXX 100644 --- a/include/sysemu/block-backend-global-state.h +++ b/include/sysemu/block-backend-global-state.h @@ -XXX,XX +XXX,XX @@ blk_co_new_open(const char *filename, const char *reference, QDict *options, int blk_get_refcnt(BlockBackend *blk); void blk_ref(BlockBackend *blk); -void blk_unref(BlockBackend *blk); + +void no_coroutine_fn blk_unref(BlockBackend *blk); +void coroutine_fn no_co_wrapper blk_co_unref(BlockBackend *blk); + void blk_remove_all_bs(void); BlockBackend *blk_by_name(const char *name); BlockBackend *blk_next(BlockBackend *blk); diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_create_opts_simple(BlockDriver *drv, ret = 0; out: - blk_unref(blk); + blk_co_unref(blk); return ret; } diff --git a/block/crypto.c b/block/crypto.c index XXXXXXX..XXXXXXX 100644 --- a/block/crypto.c +++ b/block/crypto.c @@ -XXX,XX +XXX,XX @@ block_crypto_co_create_generic(BlockDriverState *bs, int64_t size, ret = 0; cleanup: qcrypto_block_free(crypto); - blk_unref(blk); + blk_co_unref(blk); return ret; } @@ -XXX,XX +XXX,XX @@ block_crypto_co_create_luks(BlockdevCreateOptions *create_options, Error **errp) ret = 0; fail: - bdrv_unref(bs); + bdrv_co_unref(bs); return ret; } @@ -XXX,XX +XXX,XX @@ fail: bdrv_co_delete_file_noerr(bs); } - bdrv_unref(bs); + bdrv_co_unref(bs); qapi_free_QCryptoBlockCreateOptions(create_opts); qobject_unref(cryptoopts); return ret; diff --git a/block/parallels.c b/block/parallels.c index XXXXXXX..XXXXXXX 100644 --- a/block/parallels.c +++ b/block/parallels.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn parallels_co_create(BlockdevCreateOptions* opts, ret = 0; out: - blk_unref(blk); - bdrv_unref(bs); + blk_co_unref(blk); + bdrv_co_unref(bs); return ret; exit: @@ -XXX,XX +XXX,XX @@ parallels_co_create_opts(BlockDriver *drv, const char *filename, done: qobject_unref(qdict); - bdrv_unref(bs); + bdrv_co_unref(bs); qapi_free_BlockdevCreateOptions(create_options); return ret; } diff --git a/block/qcow.c b/block/qcow.c index XXXXXXX..XXXXXXX 100644 --- a/block/qcow.c +++ b/block/qcow.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qcow_co_create(BlockdevCreateOptions *opts, g_free(tmp); ret = 0; exit: - blk_unref(qcow_blk); - bdrv_unref(bs); + blk_co_unref(qcow_blk); + bdrv_co_unref(bs); qcrypto_block_free(crypto); return ret; } @@ -XXX,XX +XXX,XX @@ qcow_co_create_opts(BlockDriver *drv, const char *filename, fail: g_free(backing_fmt); qobject_unref(qdict); - bdrv_unref(bs); + bdrv_co_unref(bs); qapi_free_BlockdevCreateOptions(create_options); return ret; } diff --git a/block/qcow2.c b/block/qcow2.c index XXXXXXX..XXXXXXX 100644 --- a/block/qcow2.c +++ b/block/qcow2.c @@ -XXX,XX +XXX,XX @@ qcow2_co_create(BlockdevCreateOptions *create_options, Error **errp) goto out; } - blk_unref(blk); + blk_co_unref(blk); blk = NULL; /* @@ -XXX,XX +XXX,XX @@ qcow2_co_create(BlockdevCreateOptions *create_options, Error **errp) } } - blk_unref(blk); + blk_co_unref(blk); blk = NULL; /* Reopen the image without BDRV_O_NO_FLUSH to flush it before returning. @@ -XXX,XX +XXX,XX @@ qcow2_co_create(BlockdevCreateOptions *create_options, Error **errp) ret = 0; out: - blk_unref(blk); - bdrv_unref(bs); - bdrv_unref(data_bs); + blk_co_unref(blk); + bdrv_co_unref(bs); + bdrv_co_unref(data_bs); return ret; } @@ -XXX,XX +XXX,XX @@ finish: } qobject_unref(qdict); - bdrv_unref(bs); - bdrv_unref(data_bs); + bdrv_co_unref(bs); + bdrv_co_unref(data_bs); qapi_free_BlockdevCreateOptions(create_options); return ret; } diff --git a/block/qed.c b/block/qed.c index XXXXXXX..XXXXXXX 100644 --- a/block/qed.c +++ b/block/qed.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_qed_co_create(BlockdevCreateOptions *opts, ret = 0; /* success */ out: g_free(l1_table); - blk_unref(blk); - bdrv_unref(bs); + blk_co_unref(blk); + bdrv_co_unref(bs); return ret; } @@ -XXX,XX +XXX,XX @@ bdrv_qed_co_create_opts(BlockDriver *drv, const char *filename, fail: qobject_unref(qdict); - bdrv_unref(bs); + bdrv_co_unref(bs); qapi_free_BlockdevCreateOptions(create_options); return ret; } diff --git a/block/vdi.c b/block/vdi.c index XXXXXXX..XXXXXXX 100644 --- a/block/vdi.c +++ b/block/vdi.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn vdi_co_do_create(BlockdevCreateOptions *create_options, ret = 0; exit: - blk_unref(blk); - bdrv_unref(bs_file); + blk_co_unref(blk); + bdrv_co_unref(bs_file); g_free(bmap); return ret; } @@ -XXX,XX +XXX,XX @@ vdi_co_create_opts(BlockDriver *drv, const char *filename, done: qobject_unref(qdict); qapi_free_BlockdevCreateOptions(create_options); - bdrv_unref(bs_file); + bdrv_co_unref(bs_file); return ret; } diff --git a/block/vhdx.c b/block/vhdx.c index XXXXXXX..XXXXXXX 100644 --- a/block/vhdx.c +++ b/block/vhdx.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn vhdx_co_create(BlockdevCreateOptions *opts, ret = 0; delete_and_exit: - blk_unref(blk); - bdrv_unref(bs); + blk_co_unref(blk); + bdrv_co_unref(bs); g_free(creator); return ret; } @@ -XXX,XX +XXX,XX @@ vhdx_co_create_opts(BlockDriver *drv, const char *filename, fail: qobject_unref(qdict); - bdrv_unref(bs); + bdrv_co_unref(bs); qapi_free_BlockdevCreateOptions(create_options); return ret; } diff --git a/block/vmdk.c b/block/vmdk.c index XXXXXXX..XXXXXXX 100644 --- a/block/vmdk.c +++ b/block/vmdk.c @@ -XXX,XX +XXX,XX @@ exit: if (pbb) { *pbb = blk; } else { - blk_unref(blk); + blk_co_unref(blk); blk = NULL; } } @@ -XXX,XX +XXX,XX @@ vmdk_co_do_create(int64_t size, if (strcmp(blk_bs(backing)->drv->format_name, "vmdk")) { error_setg(errp, "Invalid backing file format: %s. Must be vmdk", blk_bs(backing)->drv->format_name); - blk_unref(backing); + blk_co_unref(backing); ret = -EINVAL; goto exit; } ret = vmdk_read_cid(blk_bs(backing), 0, &parent_cid); - blk_unref(backing); + blk_co_unref(backing); if (ret) { error_setg(errp, "Failed to read parent CID"); goto exit; @@ -XXX,XX +XXX,XX @@ vmdk_co_do_create(int64_t size, blk_bs(extent_blk)->filename); created_size += cur_size; extent_idx++; - blk_unref(extent_blk); + blk_co_unref(extent_blk); } /* Check whether we got excess extents */ extent_blk = extent_fn(-1, extent_idx, flat, split, compress, zeroed_grain, opaque, NULL); if (extent_blk) { - blk_unref(extent_blk); + blk_co_unref(extent_blk); error_setg(errp, "List of extents contains unused extents"); ret = -EINVAL; goto exit; @@ -XXX,XX +XXX,XX @@ vmdk_co_do_create(int64_t size, ret = 0; exit: if (blk) { - blk_unref(blk); + blk_co_unref(blk); } g_free(desc); g_free(parent_desc_line); @@ -XXX,XX +XXX,XX @@ vmdk_co_create_opts_cb(int64_t size, int idx, bool flat, bool split, errp)) { goto exit; } - bdrv_unref(bs); + bdrv_co_unref(bs); exit: g_free(ext_filename); return blk; @@ -XXX,XX +XXX,XX @@ static BlockBackend * coroutine_fn vmdk_co_create_cb(int64_t size, int idx, return NULL; } blk_set_allow_write_beyond_eof(blk, true); - bdrv_unref(bs); + bdrv_co_unref(bs); if (size != -1) { ret = vmdk_init_extent(blk, size, flat, compress, zeroed_grain, errp); if (ret) { - blk_unref(blk); + blk_co_unref(blk); blk = NULL; } } diff --git a/block/vpc.c b/block/vpc.c index XXXXXXX..XXXXXXX 100644 --- a/block/vpc.c +++ b/block/vpc.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn vpc_co_create(BlockdevCreateOptions *opts, } out: - blk_unref(blk); - bdrv_unref(bs); + blk_co_unref(blk); + bdrv_co_unref(bs); return ret; } @@ -XXX,XX +XXX,XX @@ vpc_co_create_opts(BlockDriver *drv, const char *filename, fail: qobject_unref(qdict); - bdrv_unref(bs); + bdrv_co_unref(bs); qapi_free_BlockdevCreateOptions(create_options); return ret; } -- 2.40.1
This QMP handler runs in a coroutine, so it must use the corresponding no_co_wrappers instead. Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=2185688 Cc: qemu-stable@nongnu.org Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Message-Id: <20230504115750.54437-5-kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- blockdev.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/blockdev.c b/blockdev.c index XXXXXXX..XXXXXXX 100644 --- a/blockdev.c +++ b/blockdev.c @@ -XXX,XX +XXX,XX @@ void coroutine_fn qmp_block_resize(const char *device, const char *node_name, return; } - blk = blk_new_with_bs(bs, BLK_PERM_RESIZE, BLK_PERM_ALL, errp); + blk = blk_co_new_with_bs(bs, BLK_PERM_RESIZE, BLK_PERM_ALL, errp); if (!blk) { return; } @@ -XXX,XX +XXX,XX @@ void coroutine_fn qmp_block_resize(const char *device, const char *node_name, bdrv_co_lock(bs); bdrv_drained_end(bs); - blk_unref(blk); + blk_co_unref(blk); bdrv_co_unlock(bs); } -- 2.40.1
This tests that trying to resize an image with QMP block_resize doesn't hang or otherwise fail when the image is attached to a device running in an iothread. This is a regression test for the recent fix that changed qmp_block_resize, which is a coroutine based QMP handler, to avoid calling no_coroutine_fns directly. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230509134133.373408-1-kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- tests/qemu-iotests/tests/iothreads-resize | 71 +++++++++++++++++++ tests/qemu-iotests/tests/iothreads-resize.out | 11 +++ 2 files changed, 82 insertions(+) create mode 100755 tests/qemu-iotests/tests/iothreads-resize create mode 100644 tests/qemu-iotests/tests/iothreads-resize.out diff --git a/tests/qemu-iotests/tests/iothreads-resize b/tests/qemu-iotests/tests/iothreads-resize new file mode 100755 index XXXXXXX..XXXXXXX --- /dev/null +++ b/tests/qemu-iotests/tests/iothreads-resize @@ -XXX,XX +XXX,XX @@ +#!/usr/bin/env bash +# group: rw auto quick +# +# Test resizing an image that is attached to a separate iothread +# +# Copyright (C) 2023 Red Hat, Inc. +# +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 2 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see <http://www.gnu.org/licenses/>. +# + +# creator +owner=kwolf@redhat.com + +seq=`basename $0` +echo "QA output created by $seq" + +status=1 # failure is the default! + +_cleanup() +{ + _cleanup_test_img +} +trap "_cleanup; exit \$status" 0 1 2 3 15 + +# get standard environment, filters and checks +cd .. +. ./common.rc +. ./common.filter + +# Resizing images is only supported by a few block drivers +_supported_fmt raw qcow2 qed +_supported_proto file +_require_devices virtio-scsi-pci + +size=64M +_make_test_img $size + +qmp() { +cat <<EOF +{"execute":"qmp_capabilities"} +{'execute': 'block_resize', + 'arguments': {'node-name': 'img', 'size': 134217728}} +{"execute":"quit"} +EOF +} + +qmp | $QEMU -S -display none \ + -drive if=none,format=$IMGFMT,file="$TEST_IMG",node-name=img \ + -object iothread,id=t0 \ + -device virtio-scsi-pci,iothread=t0 \ + -device scsi-hd,drive=none0 \ + -qmp stdio \ + | _filter_qmp + +_img_info | _filter_img_info + +# success, all done +echo "*** done" +rm -f $seq.full +status=0 diff --git a/tests/qemu-iotests/tests/iothreads-resize.out b/tests/qemu-iotests/tests/iothreads-resize.out new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/tests/qemu-iotests/tests/iothreads-resize.out @@ -XXX,XX +XXX,XX @@ +QA output created by iothreads-resize +Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864 +QMP_VERSION +{"return": {}} +{"return": {}} +{"return": {}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false, "reason": "host-qmp-quit"}} +image: TEST_DIR/t.IMGFMT +file format: IMGFMT +virtual size: 128 MiB (134217728 bytes) +*** done -- 2.40.1
test-bdrv-drain contains a few test cases that are run both in coroutine and non-coroutine context. Running the entire code including the setup and shutdown in coroutines is incorrect because graph modifications can generally not happen in coroutines. Change the test so that creating and destroying the test nodes and BlockBackends always happens outside of coroutine context. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <20230504115750.54437-6-kwolf@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- tests/unit/test-bdrv-drain.c | 112 +++++++++++++++++++++++------------ 1 file changed, 75 insertions(+), 37 deletions(-) diff --git a/tests/unit/test-bdrv-drain.c b/tests/unit/test-bdrv-drain.c index XXXXXXX..XXXXXXX 100644 --- a/tests/unit/test-bdrv-drain.c +++ b/tests/unit/test-bdrv-drain.c @@ -XXX,XX +XXX,XX @@ static void do_drain_begin_unlocked(enum drain_type drain_type, BlockDriverState } } +static BlockBackend * no_coroutine_fn test_setup(void) +{ + BlockBackend *blk; + BlockDriverState *bs, *backing; + + blk = blk_new(qemu_get_aio_context(), BLK_PERM_ALL, BLK_PERM_ALL); + bs = bdrv_new_open_driver(&bdrv_test, "test-node", BDRV_O_RDWR, + &error_abort); + blk_insert_bs(blk, bs, &error_abort); + + backing = bdrv_new_open_driver(&bdrv_test, "backing", 0, &error_abort); + bdrv_set_backing_hd(bs, backing, &error_abort); + + bdrv_unref(backing); + bdrv_unref(bs); + + return blk; +} + static void do_drain_end_unlocked(enum drain_type drain_type, BlockDriverState *bs) { if (drain_type != BDRV_DRAIN_ALL) { @@ -XXX,XX +XXX,XX @@ static void do_drain_end_unlocked(enum drain_type drain_type, BlockDriverState * } } -static void test_drv_cb_common(enum drain_type drain_type, bool recursive) +static void test_drv_cb_common(BlockBackend *blk, enum drain_type drain_type, + bool recursive) { - BlockBackend *blk; - BlockDriverState *bs, *backing; + BlockDriverState *bs = blk_bs(blk); + BlockDriverState *backing = bs->backing->bs; BDRVTestState *s, *backing_s; BlockAIOCB *acb; int aio_ret; QEMUIOVector qiov = QEMU_IOVEC_INIT_BUF(qiov, NULL, 0); - blk = blk_new(qemu_get_aio_context(), BLK_PERM_ALL, BLK_PERM_ALL); - bs = bdrv_new_open_driver(&bdrv_test, "test-node", BDRV_O_RDWR, - &error_abort); s = bs->opaque; - blk_insert_bs(blk, bs, &error_abort); - - backing = bdrv_new_open_driver(&bdrv_test, "backing", 0, &error_abort); backing_s = backing->opaque; - bdrv_set_backing_hd(bs, backing, &error_abort); /* Simple bdrv_drain_all_begin/end pair, check that CBs are called */ g_assert_cmpint(s->drain_count, ==, 0); @@ -XXX,XX +XXX,XX @@ static void test_drv_cb_common(enum drain_type drain_type, bool recursive) g_assert_cmpint(s->drain_count, ==, 0); g_assert_cmpint(backing_s->drain_count, ==, 0); - - bdrv_unref(backing); - bdrv_unref(bs); - blk_unref(blk); } static void test_drv_cb_drain_all(void) { - test_drv_cb_common(BDRV_DRAIN_ALL, true); + BlockBackend *blk = test_setup(); + test_drv_cb_common(blk, BDRV_DRAIN_ALL, true); + blk_unref(blk); } static void test_drv_cb_drain(void) { - test_drv_cb_common(BDRV_DRAIN, false); + BlockBackend *blk = test_setup(); + test_drv_cb_common(blk, BDRV_DRAIN, false); + blk_unref(blk); +} + +static void coroutine_fn test_drv_cb_co_drain_all_entry(void) +{ + BlockBackend *blk = blk_all_next(NULL); + test_drv_cb_common(blk, BDRV_DRAIN_ALL, true); } static void test_drv_cb_co_drain_all(void) { - call_in_coroutine(test_drv_cb_drain_all); + BlockBackend *blk = test_setup(); + call_in_coroutine(test_drv_cb_co_drain_all_entry); + blk_unref(blk); } -static void test_drv_cb_co_drain(void) +static void coroutine_fn test_drv_cb_co_drain_entry(void) { - call_in_coroutine(test_drv_cb_drain); + BlockBackend *blk = blk_all_next(NULL); + test_drv_cb_common(blk, BDRV_DRAIN, false); } -static void test_quiesce_common(enum drain_type drain_type, bool recursive) +static void test_drv_cb_co_drain(void) { - BlockBackend *blk; - BlockDriverState *bs, *backing; - - blk = blk_new(qemu_get_aio_context(), BLK_PERM_ALL, BLK_PERM_ALL); - bs = bdrv_new_open_driver(&bdrv_test, "test-node", BDRV_O_RDWR, - &error_abort); - blk_insert_bs(blk, bs, &error_abort); + BlockBackend *blk = test_setup(); + call_in_coroutine(test_drv_cb_co_drain_entry); + blk_unref(blk); +} - backing = bdrv_new_open_driver(&bdrv_test, "backing", 0, &error_abort); - bdrv_set_backing_hd(bs, backing, &error_abort); +static void test_quiesce_common(BlockBackend *blk, enum drain_type drain_type, + bool recursive) +{ + BlockDriverState *bs = blk_bs(blk); + BlockDriverState *backing = bs->backing->bs; g_assert_cmpint(bs->quiesce_counter, ==, 0); g_assert_cmpint(backing->quiesce_counter, ==, 0); @@ -XXX,XX +XXX,XX @@ static void test_quiesce_common(enum drain_type drain_type, bool recursive) g_assert_cmpint(bs->quiesce_counter, ==, 0); g_assert_cmpint(backing->quiesce_counter, ==, 0); - - bdrv_unref(backing); - bdrv_unref(bs); - blk_unref(blk); } static void test_quiesce_drain_all(void) { - test_quiesce_common(BDRV_DRAIN_ALL, true); + BlockBackend *blk = test_setup(); + test_quiesce_common(blk, BDRV_DRAIN_ALL, true); + blk_unref(blk); } static void test_quiesce_drain(void) { - test_quiesce_common(BDRV_DRAIN, false); + BlockBackend *blk = test_setup(); + test_quiesce_common(blk, BDRV_DRAIN, false); + blk_unref(blk); +} + +static void coroutine_fn test_quiesce_co_drain_all_entry(void) +{ + BlockBackend *blk = blk_all_next(NULL); + test_quiesce_common(blk, BDRV_DRAIN_ALL, true); } static void test_quiesce_co_drain_all(void) { - call_in_coroutine(test_quiesce_drain_all); + BlockBackend *blk = test_setup(); + call_in_coroutine(test_quiesce_co_drain_all_entry); + blk_unref(blk); +} + +static void coroutine_fn test_quiesce_co_drain_entry(void) +{ + BlockBackend *blk = blk_all_next(NULL); + test_quiesce_common(blk, BDRV_DRAIN, false); } static void test_quiesce_co_drain(void) { - call_in_coroutine(test_quiesce_drain); + BlockBackend *blk = test_setup(); + call_in_coroutine(test_quiesce_co_drain_entry); + blk_unref(blk); } static void test_nested(void) -- 2.40.1
For some functions, it is part of their interface to be called without holding the graph lock. Add a new macro to document this. The macro expands to TSA_EXCLUDES(), which is a relatively weak check because it passes in cases where the compiler just doesn't know if the lock is held. Function pointers can't be checked at all. Therefore, its primary purpose is documentation. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Message-Id: <20230504115750.54437-7-kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/graph-lock.h | 2 ++ 1 file changed, 2 insertions(+) diff --git a/include/block/graph-lock.h b/include/block/graph-lock.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/graph-lock.h +++ b/include/block/graph-lock.h @@ -XXX,XX +XXX,XX @@ extern BdrvGraphLock graph_lock; */ #define GRAPH_WRLOCK TSA_REQUIRES(graph_lock) #define GRAPH_RDLOCK TSA_REQUIRES_SHARED(graph_lock) +#define GRAPH_UNLOCKED TSA_EXCLUDES(graph_lock) /* * TSA annotations are not part of function types, so checks are defeated when @@ -XXX,XX +XXX,XX @@ extern BdrvGraphLock graph_lock; */ #define GRAPH_RDLOCK_PTR TSA_GUARDED_BY(graph_lock) #define GRAPH_WRLOCK_PTR TSA_GUARDED_BY(graph_lock) +#define GRAPH_UNLOCKED_PTR /* * register_aiocontext: -- 2.40.1
GRAPH_RDLOCK_GUARD() and GRAPH_RDLOCK_GUARD_MAINLOOP() only take a reader lock for the graph, so the correct annotation for them to use is TSA_ASSERT_SHARED rather than TSA_ASSERT. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <20230504115750.54437-8-kwolf@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/graph-lock.h | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) diff --git a/include/block/graph-lock.h b/include/block/graph-lock.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/graph-lock.h +++ b/include/block/graph-lock.h @@ -XXX,XX +XXX,XX @@ typedef struct GraphLockable { } GraphLockable; #define GML_OBJ_() (&(GraphLockable) { }) /* - * This is not marked as TSA_ACQUIRE() because TSA doesn't understand the + * This is not marked as TSA_ACQUIRE_SHARED() because TSA doesn't understand the * cleanup attribute and would therefore complain that the graph is never - * unlocked. TSA_ASSERT() makes sure that the following calls know that we - * hold the lock while unlocking is left unchecked. + * unlocked. TSA_ASSERT_SHARED() makes sure that the following calls know that + * we hold the lock while unlocking is left unchecked. */ -static inline GraphLockable * TSA_ASSERT(graph_lock) TSA_NO_TSA coroutine_fn +static inline GraphLockable * TSA_ASSERT_SHARED(graph_lock) TSA_NO_TSA coroutine_fn graph_lockable_auto_lock(GraphLockable *x) { bdrv_graph_co_rdlock(); @@ -XXX,XX +XXX,XX @@ typedef struct GraphLockableMainloop { } GraphLockableMainloop; #define GMLML_OBJ_() (&(GraphLockableMainloop) { }) /* - * This is not marked as TSA_ACQUIRE() because TSA doesn't understand the + * This is not marked as TSA_ACQUIRE_SHARED() because TSA doesn't understand the * cleanup attribute and would therefore complain that the graph is never - * unlocked. TSA_ASSERT() makes sure that the following calls know that we - * hold the lock while unlocking is left unchecked. + * unlocked. TSA_ASSERT_SHARED() makes sure that the following calls know that + * we hold the lock while unlocking is left unchecked. */ -static inline GraphLockableMainloop * TSA_ASSERT(graph_lock) TSA_NO_TSA +static inline GraphLockableMainloop * TSA_ASSERT_SHARED(graph_lock) TSA_NO_TSA graph_lockable_auto_lock_mainloop(GraphLockableMainloop *x) { bdrv_graph_rdlock_main_loop(); -- 2.40.1
Drivers were a bit confused about whether .bdrv_open can run in a coroutine and whether or not it holds a graph lock. It cannot keep a graph lock from the caller across the whole function because it both changes the graph (requires a writer lock) and does I/O (requires a reader lock). Therefore, it should take these locks internally as needed. The functions used to be called in coroutine context during image creation. This was buggy for other reasons, and as of commit 32192301, all block drivers go through no_co_wrappers. So it is not called in coroutine context any more. Fix qcow2 and qed to work with the correct assumptions: The graph lock needs to be taken internally instead of just assuming it's already there, and the coroutine path is dead code that can be removed. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Message-Id: <20230504115750.54437-9-kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block_int-common.h | 8 ++++---- block.c | 6 +++--- block/qcow2.c | 15 ++++++--------- block/qed.c | 18 ++++++++---------- 4 files changed, 21 insertions(+), 26 deletions(-) diff --git a/include/block/block_int-common.h b/include/block/block_int-common.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block_int-common.h +++ b/include/block/block_int-common.h @@ -XXX,XX +XXX,XX @@ struct BlockDriver { void (*bdrv_reopen_abort)(BDRVReopenState *reopen_state); void (*bdrv_join_options)(QDict *options, QDict *old_options); - int (*bdrv_open)(BlockDriverState *bs, QDict *options, int flags, - Error **errp); + int GRAPH_UNLOCKED_PTR (*bdrv_open)( + BlockDriverState *bs, QDict *options, int flags, Error **errp); /* Protocol drivers should implement this instead of bdrv_open */ - int (*bdrv_file_open)(BlockDriverState *bs, QDict *options, int flags, - Error **errp); + int GRAPH_UNLOCKED_PTR (*bdrv_file_open)( + BlockDriverState *bs, QDict *options, int flags, Error **errp); void (*bdrv_close)(BlockDriverState *bs); int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_create)( diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ out: * bdrv_refresh_total_sectors() which polls when called from non-coroutine * context. */ -static int bdrv_open_driver(BlockDriverState *bs, BlockDriver *drv, - const char *node_name, QDict *options, - int open_flags, Error **errp) +static int no_coroutine_fn GRAPH_UNLOCKED +bdrv_open_driver(BlockDriverState *bs, BlockDriver *drv, const char *node_name, + QDict *options, int open_flags, Error **errp) { Error *local_err = NULL; int i, ret; diff --git a/block/qcow2.c b/block/qcow2.c index XXXXXXX..XXXXXXX 100644 --- a/block/qcow2.c +++ b/block/qcow2.c @@ -XXX,XX +XXX,XX @@ static void coroutine_fn qcow2_open_entry(void *opaque) QCow2OpenCo *qoc = opaque; BDRVQcow2State *s = qoc->bs->opaque; - assume_graph_lock(); /* FIXME */ + GRAPH_RDLOCK_GUARD(); qemu_co_mutex_lock(&s->lock); qoc->ret = qcow2_do_open(qoc->bs, qoc->options, qoc->flags, true, @@ -XXX,XX +XXX,XX @@ static int qcow2_open(BlockDriverState *bs, QDict *options, int flags, /* Initialise locks */ qemu_co_mutex_init(&s->lock); - if (qemu_in_coroutine()) { - /* From bdrv_co_create. */ - qcow2_open_entry(&qoc); - } else { - assert(qemu_get_current_aio_context() == qemu_get_aio_context()); - qemu_coroutine_enter(qemu_coroutine_create(qcow2_open_entry, &qoc)); - BDRV_POLL_WHILE(bs, qoc.ret == -EINPROGRESS); - } + assert(!qemu_in_coroutine()); + assert(qemu_get_current_aio_context() == qemu_get_aio_context()); + qemu_coroutine_enter(qemu_coroutine_create(qcow2_open_entry, &qoc)); + BDRV_POLL_WHILE(bs, qoc.ret == -EINPROGRESS); + return qoc.ret; } diff --git a/block/qed.c b/block/qed.c index XXXXXXX..XXXXXXX 100644 --- a/block/qed.c +++ b/block/qed.c @@ -XXX,XX +XXX,XX @@ typedef struct QEDOpenCo { int ret; } QEDOpenCo; -static void coroutine_fn GRAPH_RDLOCK bdrv_qed_open_entry(void *opaque) +static void coroutine_fn bdrv_qed_open_entry(void *opaque) { QEDOpenCo *qoc = opaque; BDRVQEDState *s = qoc->bs->opaque; + GRAPH_RDLOCK_GUARD(); + qemu_co_mutex_lock(&s->table_lock); qoc->ret = bdrv_qed_do_open(qoc->bs, qoc->options, qoc->flags, qoc->errp); qemu_co_mutex_unlock(&s->table_lock); @@ -XXX,XX +XXX,XX @@ static int bdrv_qed_open(BlockDriverState *bs, QDict *options, int flags, }; int ret; - assume_graph_lock(); /* FIXME */ - ret = bdrv_open_file_child(NULL, options, "file", bs, errp); if (ret < 0) { return ret; } bdrv_qed_init_state(bs); - if (qemu_in_coroutine()) { - bdrv_qed_open_entry(&qoc); - } else { - assert(qemu_get_current_aio_context() == qemu_get_aio_context()); - qemu_coroutine_enter(qemu_coroutine_create(bdrv_qed_open_entry, &qoc)); - BDRV_POLL_WHILE(bs, qoc.ret == -EINPROGRESS); - } + assert(!qemu_in_coroutine()); + assert(qemu_get_current_aio_context() == qemu_get_aio_context()); + qemu_coroutine_enter(qemu_coroutine_create(bdrv_qed_open_entry, &qoc)); + BDRV_POLL_WHILE(bs, qoc.ret == -EINPROGRESS); + return qoc.ret; } -- 2.40.1
The only thing nbd_co_flush() does is call nbd_client_co_flush(). Just use that function directly in the BlockDriver definitions and remove the wrapper. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Message-Id: <20230504115750.54437-10-kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/nbd.c | 11 +++-------- 1 file changed, 3 insertions(+), 8 deletions(-) diff --git a/block/nbd.c b/block/nbd.c index XXXXXXX..XXXXXXX 100644 --- a/block/nbd.c +++ b/block/nbd.c @@ -XXX,XX +XXX,XX @@ fail: return ret; } -static int coroutine_fn nbd_co_flush(BlockDriverState *bs) -{ - return nbd_client_co_flush(bs); -} - static void nbd_refresh_limits(BlockDriverState *bs, Error **errp) { BDRVNBDState *s = (BDRVNBDState *)bs->opaque; @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_nbd = { .bdrv_co_pwritev = nbd_client_co_pwritev, .bdrv_co_pwrite_zeroes = nbd_client_co_pwrite_zeroes, .bdrv_close = nbd_close, - .bdrv_co_flush_to_os = nbd_co_flush, + .bdrv_co_flush_to_os = nbd_client_co_flush, .bdrv_co_pdiscard = nbd_client_co_pdiscard, .bdrv_refresh_limits = nbd_refresh_limits, .bdrv_co_truncate = nbd_co_truncate, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_nbd_tcp = { .bdrv_co_pwritev = nbd_client_co_pwritev, .bdrv_co_pwrite_zeroes = nbd_client_co_pwrite_zeroes, .bdrv_close = nbd_close, - .bdrv_co_flush_to_os = nbd_co_flush, + .bdrv_co_flush_to_os = nbd_client_co_flush, .bdrv_co_pdiscard = nbd_client_co_pdiscard, .bdrv_refresh_limits = nbd_refresh_limits, .bdrv_co_truncate = nbd_co_truncate, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_nbd_unix = { .bdrv_co_pwritev = nbd_client_co_pwritev, .bdrv_co_pwrite_zeroes = nbd_client_co_pwrite_zeroes, .bdrv_close = nbd_close, - .bdrv_co_flush_to_os = nbd_co_flush, + .bdrv_co_flush_to_os = nbd_client_co_flush, .bdrv_co_pdiscard = nbd_client_co_pdiscard, .bdrv_refresh_limits = nbd_refresh_limits, .bdrv_co_truncate = nbd_co_truncate, -- 2.40.1
From: Emanuele Giuseppe Esposito <eesposit@redhat.com> This adds GRAPH_RDLOCK annotations to declare that callers of nbd_co_do_establish_connection() need to hold a reader lock for the graph. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Message-Id: <20230504115750.54437-11-kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/coroutines.h | 5 +++-- block/nbd.c | 39 +++++++++++++++++++++------------------ 2 files changed, 24 insertions(+), 20 deletions(-) diff --git a/block/coroutines.h b/block/coroutines.h index XXXXXXX..XXXXXXX 100644 --- a/block/coroutines.h +++ b/block/coroutines.h @@ -XXX,XX +XXX,XX @@ bdrv_co_readv_vmstate(BlockDriverState *bs, QEMUIOVector *qiov, int64_t pos); int coroutine_fn GRAPH_RDLOCK bdrv_co_writev_vmstate(BlockDriverState *bs, QEMUIOVector *qiov, int64_t pos); -int coroutine_fn +int coroutine_fn GRAPH_RDLOCK nbd_co_do_establish_connection(BlockDriverState *bs, bool blocking, Error **errp); @@ -XXX,XX +XXX,XX @@ bdrv_common_block_status_above(BlockDriverState *bs, int64_t *map, BlockDriverState **file, int *depth); -int co_wrapper_mixed + +int co_wrapper_mixed_bdrv_rdlock nbd_do_establish_connection(BlockDriverState *bs, bool blocking, Error **errp); #endif /* BLOCK_COROUTINES_H */ diff --git a/block/nbd.c b/block/nbd.c index XXXXXXX..XXXXXXX 100644 --- a/block/nbd.c +++ b/block/nbd.c @@ -XXX,XX +XXX,XX @@ int coroutine_fn nbd_co_do_establish_connection(BlockDriverState *bs, int ret; IO_CODE(); + assert_bdrv_graph_readable(); assert(!s->ioc); s->ioc = nbd_co_establish_connection(s->conn, &s->info, blocking, errp); @@ -XXX,XX +XXX,XX @@ static bool nbd_client_connecting(BDRVNBDState *s) } /* Called with s->requests_lock taken. */ -static coroutine_fn void nbd_reconnect_attempt(BDRVNBDState *s) +static void coroutine_fn GRAPH_RDLOCK nbd_reconnect_attempt(BDRVNBDState *s) { int ret; bool blocking = s->state == NBD_CLIENT_CONNECTING_WAIT; @@ -XXX,XX +XXX,XX @@ static coroutine_fn int nbd_receive_replies(BDRVNBDState *s, uint64_t handle) } } -static int coroutine_fn nbd_co_send_request(BlockDriverState *bs, - NBDRequest *request, - QEMUIOVector *qiov) +static int coroutine_fn GRAPH_RDLOCK +nbd_co_send_request(BlockDriverState *bs, NBDRequest *request, + QEMUIOVector *qiov) { BDRVNBDState *s = (BDRVNBDState *)bs->opaque; int rc, i = -1; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn nbd_co_receive_blockstatus_reply(BDRVNBDState *s, return iter.ret; } -static int coroutine_fn nbd_co_request(BlockDriverState *bs, NBDRequest *request, - QEMUIOVector *write_qiov) +static int coroutine_fn GRAPH_RDLOCK +nbd_co_request(BlockDriverState *bs, NBDRequest *request, + QEMUIOVector *write_qiov) { int ret, request_ret; Error *local_err = NULL; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn nbd_co_request(BlockDriverState *bs, NBDRequest *request return ret ? ret : request_ret; } -static int coroutine_fn nbd_client_co_preadv(BlockDriverState *bs, int64_t offset, - int64_t bytes, QEMUIOVector *qiov, - BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +nbd_client_co_preadv(BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, BdrvRequestFlags flags) { int ret, request_ret; Error *local_err = NULL; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn nbd_client_co_preadv(BlockDriverState *bs, int64_t offse return ret ? ret : request_ret; } -static int coroutine_fn nbd_client_co_pwritev(BlockDriverState *bs, int64_t offset, - int64_t bytes, QEMUIOVector *qiov, - BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +nbd_client_co_pwritev(BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, BdrvRequestFlags flags) { BDRVNBDState *s = (BDRVNBDState *)bs->opaque; NBDRequest request = { @@ -XXX,XX +XXX,XX @@ static int coroutine_fn nbd_client_co_pwritev(BlockDriverState *bs, int64_t offs return nbd_co_request(bs, &request, qiov); } -static int coroutine_fn nbd_client_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset, - int64_t bytes, BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +nbd_client_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset, int64_t bytes, + BdrvRequestFlags flags) { BDRVNBDState *s = (BDRVNBDState *)bs->opaque; NBDRequest request = { @@ -XXX,XX +XXX,XX @@ static int coroutine_fn nbd_client_co_pwrite_zeroes(BlockDriverState *bs, int64_ return nbd_co_request(bs, &request, NULL); } -static int coroutine_fn nbd_client_co_flush(BlockDriverState *bs) +static int coroutine_fn GRAPH_RDLOCK nbd_client_co_flush(BlockDriverState *bs) { BDRVNBDState *s = (BDRVNBDState *)bs->opaque; NBDRequest request = { .type = NBD_CMD_FLUSH }; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn nbd_client_co_flush(BlockDriverState *bs) return nbd_co_request(bs, &request, NULL); } -static int coroutine_fn nbd_client_co_pdiscard(BlockDriverState *bs, int64_t offset, - int64_t bytes) +static int coroutine_fn GRAPH_RDLOCK +nbd_client_co_pdiscard(BlockDriverState *bs, int64_t offset, int64_t bytes) { BDRVNBDState *s = (BDRVNBDState *)bs->opaque; NBDRequest request = { @@ -XXX,XX +XXX,XX @@ static int coroutine_fn nbd_client_co_pdiscard(BlockDriverState *bs, int64_t off return nbd_co_request(bs, &request, NULL); } -static int coroutine_fn nbd_client_co_block_status( +static int coroutine_fn GRAPH_RDLOCK nbd_client_co_block_status( BlockDriverState *bs, bool want_zero, int64_t offset, int64_t bytes, int64_t *pnum, int64_t *map, BlockDriverState **file) { -- 2.40.1
This adds GRAPH_RDLOCK annotations to declare that functions accessing the parent list of a node need to hold a reader lock for the graph. As it happens, they already do. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Message-Id: <20230504115750.54437-12-kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/vhdx.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/block/vhdx.c b/block/vhdx.c index XXXXXXX..XXXXXXX 100644 --- a/block/vhdx.c +++ b/block/vhdx.c @@ -XXX,XX +XXX,XX @@ exit: * There are 2 headers, and the highest sequence number will represent * the active header */ -static int vhdx_create_new_headers(BlockBackend *blk, uint64_t image_size, - uint32_t log_size) +static int coroutine_fn GRAPH_RDLOCK +vhdx_create_new_headers(BlockBackend *blk, uint64_t image_size, + uint32_t log_size) { BlockDriverState *bs = blk_bs(blk); BdrvChild *child; @@ -XXX,XX +XXX,XX @@ exit: * .---- ~ ----------- ~ ------------ ~ ---------------- ~ -----------. * 1MB */ -static int coroutine_fn vhdx_co_create(BlockdevCreateOptions *opts, - Error **errp) +static int coroutine_fn GRAPH_RDLOCK +vhdx_co_create(BlockdevCreateOptions *opts, Error **errp) { BlockdevCreateOptionsVhdx *vhdx_opts; BlockBackend *blk = NULL; -- 2.40.1
This adds GRAPH_RDLOCK annotations to declare that functions accessing the parent list of a node need to hold a reader lock for the graph. As it happens, they already do. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230504115750.54437-13-kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/mirror.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/block/mirror.c b/block/mirror.c index XXXXXXX..XXXXXXX 100644 --- a/block/mirror.c +++ b/block/mirror.c @@ -XXX,XX +XXX,XX @@ static MirrorOp *coroutine_fn active_write_prepare(MirrorBlockJob *s, return op; } -static void coroutine_fn active_write_settle(MirrorOp *op) +static void coroutine_fn GRAPH_RDLOCK active_write_settle(MirrorOp *op) { uint64_t start_chunk = op->offset / op->s->granularity; uint64_t end_chunk = DIV_ROUND_UP(op->offset + op->bytes, -- 2.40.1
From: Emanuele Giuseppe Esposito <eesposit@redhat.com> This adds GRAPH_RDLOCK annotations to declare that callers of bdrv_co_get_allocated_file_size() need to hold a reader lock for the graph. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <20230504115750.54437-14-kwolf@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-io.h | 7 +++++-- include/block/block_int-common.h | 2 +- block.c | 4 +++- block/vmdk.c | 2 +- 4 files changed, 10 insertions(+), 5 deletions(-) diff --git a/include/block/block-io.h b/include/block/block-io.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block-io.h +++ b/include/block/block-io.h @@ -XXX,XX +XXX,XX @@ int64_t coroutine_mixed_fn bdrv_nb_sectors(BlockDriverState *bs); int64_t coroutine_fn GRAPH_RDLOCK bdrv_co_getlength(BlockDriverState *bs); int64_t co_wrapper_mixed_bdrv_rdlock bdrv_getlength(BlockDriverState *bs); -int64_t coroutine_fn bdrv_co_get_allocated_file_size(BlockDriverState *bs); -int64_t co_wrapper bdrv_get_allocated_file_size(BlockDriverState *bs); +int64_t coroutine_fn GRAPH_RDLOCK +bdrv_co_get_allocated_file_size(BlockDriverState *bs); + +int64_t co_wrapper_bdrv_rdlock +bdrv_get_allocated_file_size(BlockDriverState *bs); BlockMeasureInfo *bdrv_measure(BlockDriver *drv, QemuOpts *opts, BlockDriverState *in_bs, Error **errp); diff --git a/include/block/block_int-common.h b/include/block/block_int-common.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block_int-common.h +++ b/include/block/block_int-common.h @@ -XXX,XX +XXX,XX @@ struct BlockDriver { int64_t coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_getlength)( BlockDriverState *bs); - int64_t coroutine_fn (*bdrv_co_get_allocated_file_size)( + int64_t coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_get_allocated_file_size)( BlockDriverState *bs); BlockMeasureInfo *(*bdrv_measure)(QemuOpts *opts, BlockDriverState *in_bs, diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ exit: * sums the size of all data-bearing children. (This excludes backing * children.) */ -static int64_t coroutine_fn bdrv_sum_allocated_file_size(BlockDriverState *bs) +static int64_t coroutine_fn GRAPH_RDLOCK +bdrv_sum_allocated_file_size(BlockDriverState *bs) { BdrvChild *child; int64_t child_size, sum = 0; @@ -XXX,XX +XXX,XX @@ int64_t coroutine_fn bdrv_co_get_allocated_file_size(BlockDriverState *bs) { BlockDriver *drv = bs->drv; IO_CODE(); + assert_bdrv_graph_readable(); if (!drv) { return -ENOMEDIUM; diff --git a/block/vmdk.c b/block/vmdk.c index XXXXXXX..XXXXXXX 100644 --- a/block/vmdk.c +++ b/block/vmdk.c @@ -XXX,XX +XXX,XX @@ static void vmdk_close(BlockDriverState *bs) error_free(s->migration_blocker); } -static int64_t coroutine_fn +static int64_t coroutine_fn GRAPH_RDLOCK vmdk_co_get_allocated_file_size(BlockDriverState *bs) { int i; -- 2.40.1
From: Emanuele Giuseppe Esposito <eesposit@redhat.com> This adds GRAPH_RDLOCK annotations to declare that callers of bdrv_co_get_info() need to hold a reader lock for the graph. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Message-Id: <20230504115750.54437-15-kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-io.h | 7 +++++-- include/block/block_int-common.h | 4 ++-- block.c | 2 ++ block/crypto.c | 2 +- block/io.c | 11 +++++------ block/mirror.c | 8 ++++++-- block/raw-format.c | 2 +- 7 files changed, 22 insertions(+), 14 deletions(-) diff --git a/include/block/block-io.h b/include/block/block-io.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block-io.h +++ b/include/block/block-io.h @@ -XXX,XX +XXX,XX @@ const char *bdrv_get_node_name(const BlockDriverState *bs); const char *bdrv_get_device_name(const BlockDriverState *bs); const char *bdrv_get_device_or_node_name(const BlockDriverState *bs); -int coroutine_fn bdrv_co_get_info(BlockDriverState *bs, BlockDriverInfo *bdi); -int co_wrapper_mixed bdrv_get_info(BlockDriverState *bs, BlockDriverInfo *bdi); +int coroutine_fn GRAPH_RDLOCK +bdrv_co_get_info(BlockDriverState *bs, BlockDriverInfo *bdi); + +int co_wrapper_mixed_bdrv_rdlock +bdrv_get_info(BlockDriverState *bs, BlockDriverInfo *bdi); ImageInfoSpecific *bdrv_get_specific_info(BlockDriverState *bs, Error **errp); diff --git a/include/block/block_int-common.h b/include/block/block_int-common.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block_int-common.h +++ b/include/block/block_int-common.h @@ -XXX,XX +XXX,XX @@ struct BlockDriver { BlockDriverState *bs, int64_t offset, int64_t bytes, QEMUIOVector *qiov, size_t qiov_offset); - int coroutine_fn (*bdrv_co_get_info)(BlockDriverState *bs, - BlockDriverInfo *bdi); + int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_get_info)( + BlockDriverState *bs, BlockDriverInfo *bdi); ImageInfoSpecific *(*bdrv_get_specific_info)(BlockDriverState *bs, Error **errp); diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) int ret; BlockDriver *drv = bs->drv; IO_CODE(); + assert_bdrv_graph_readable(); + /* if bs->drv == NULL, bs is closed, so there's nothing to do here */ if (!drv) { return -ENOMEDIUM; diff --git a/block/crypto.c b/block/crypto.c index XXXXXXX..XXXXXXX 100644 --- a/block/crypto.c +++ b/block/crypto.c @@ -XXX,XX +XXX,XX @@ fail: return ret; } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK block_crypto_co_get_info_luks(BlockDriverState *bs, BlockDriverInfo *bdi) { BlockDriverInfo subbdi; diff --git a/block/io.c b/block/io.c index XXXXXXX..XXXXXXX 100644 --- a/block/io.c +++ b/block/io.c @@ -XXX,XX +XXX,XX @@ BdrvTrackedRequest *coroutine_fn bdrv_co_get_self_request(BlockDriverState *bs) /** * Round a region to cluster boundaries */ -void coroutine_fn bdrv_round_to_clusters(BlockDriverState *bs, - int64_t offset, int64_t bytes, - int64_t *cluster_offset, - int64_t *cluster_bytes) +void coroutine_fn GRAPH_RDLOCK +bdrv_round_to_clusters(BlockDriverState *bs, int64_t offset, int64_t bytes, + int64_t *cluster_offset, int64_t *cluster_bytes) { BlockDriverInfo bdi; IO_CODE(); @@ -XXX,XX +XXX,XX @@ void coroutine_fn bdrv_round_to_clusters(BlockDriverState *bs, } } -static coroutine_fn int bdrv_get_cluster_size(BlockDriverState *bs) +static int coroutine_fn GRAPH_RDLOCK bdrv_get_cluster_size(BlockDriverState *bs) { BlockDriverInfo bdi; int ret; @@ -XXX,XX +XXX,XX @@ fail: return ret; } -static inline int coroutine_fn +static inline int coroutine_fn GRAPH_RDLOCK bdrv_co_write_req_prepare(BdrvChild *child, int64_t offset, int64_t bytes, BdrvTrackedRequest *req, int flags) { diff --git a/block/mirror.c b/block/mirror.c index XXXXXXX..XXXXXXX 100644 --- a/block/mirror.c +++ b/block/mirror.c @@ -XXX,XX +XXX,XX @@ static uint64_t coroutine_fn mirror_iteration(MirrorBlockJob *s) } else if (ret >= 0 && !(ret & BDRV_BLOCK_DATA)) { int64_t target_offset; int64_t target_bytes; - bdrv_round_to_clusters(blk_bs(s->target), offset, io_bytes, - &target_offset, &target_bytes); + WITH_GRAPH_RDLOCK_GUARD() { + bdrv_round_to_clusters(blk_bs(s->target), offset, io_bytes, + &target_offset, &target_bytes); + } if (target_offset == offset && target_bytes == io_bytes) { mirror_method = ret & BDRV_BLOCK_ZERO ? @@ -XXX,XX +XXX,XX @@ static int coroutine_fn mirror_run(Job *job, Error **errp) */ bdrv_get_backing_filename(target_bs, backing_filename, sizeof(backing_filename)); + bdrv_graph_co_rdlock(); if (!bdrv_co_get_info(target_bs, &bdi) && bdi.cluster_size) { s->target_cluster_size = bdi.cluster_size; } else { s->target_cluster_size = BDRV_SECTOR_SIZE; } + bdrv_graph_co_rdunlock(); if (backing_filename[0] && !bdrv_backing_chain_next(target_bs) && s->granularity < s->target_cluster_size) { s->buf_size = MAX(s->buf_size, s->target_cluster_size); diff --git a/block/raw-format.c b/block/raw-format.c index XXXXXXX..XXXXXXX 100644 --- a/block/raw-format.c +++ b/block/raw-format.c @@ -XXX,XX +XXX,XX @@ static BlockMeasureInfo *raw_measure(QemuOpts *opts, BlockDriverState *in_bs, return info; } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK raw_co_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) { return bdrv_co_get_info(bs->file->bs, bdi); -- 2.40.1
From: Emanuele Giuseppe Esposito <eesposit@redhat.com> This adds GRAPH_RDLOCK annotations to declare that callers of bdrv_co_debug_event() need to hold a reader lock for the graph. Unfortunately we cannot use a co_wrapper_bdrv_rdlock (i.e. make the coroutine wrapper a no_coroutine_fn), because the function is called (using the BLKDBG_EVENT macro) by mixed functions that run both in coroutine and non-coroutine context (for example many of the functions in qcow2-cluster.c and qcow2-refcount.c). Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Message-Id: <20230504115750.54437-16-kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-io.h | 9 +++++---- include/block/block_int-common.h | 4 ++-- block.c | 2 ++ 3 files changed, 9 insertions(+), 6 deletions(-) diff --git a/include/block/block-io.h b/include/block/block-io.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block-io.h +++ b/include/block/block-io.h @@ -XXX,XX +XXX,XX @@ void *qemu_try_blockalign0(BlockDriverState *bs, size_t size); void bdrv_enable_copy_on_read(BlockDriverState *bs); void bdrv_disable_copy_on_read(BlockDriverState *bs); -void coroutine_fn bdrv_co_debug_event(BlockDriverState *bs, - BlkdebugEvent event); -void co_wrapper_mixed bdrv_debug_event(BlockDriverState *bs, - BlkdebugEvent event); +void coroutine_fn GRAPH_RDLOCK +bdrv_co_debug_event(BlockDriverState *bs, BlkdebugEvent event); + +void co_wrapper_mixed_bdrv_rdlock +bdrv_debug_event(BlockDriverState *bs, BlkdebugEvent event); #define BLKDBG_EVENT(child, evt) \ do { \ diff --git a/include/block/block_int-common.h b/include/block/block_int-common.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block_int-common.h +++ b/include/block/block_int-common.h @@ -XXX,XX +XXX,XX @@ struct BlockDriver { int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_check)( BlockDriverState *bs, BdrvCheckResult *result, BdrvCheckMode fix); - void coroutine_fn (*bdrv_co_debug_event)(BlockDriverState *bs, - BlkdebugEvent event); + void coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_debug_event)( + BlockDriverState *bs, BlkdebugEvent event); /* io queue for linux-aio */ void coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_io_plug)(BlockDriverState *bs); diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ BlockStatsSpecific *bdrv_get_specific_stats(BlockDriverState *bs) void coroutine_fn bdrv_co_debug_event(BlockDriverState *bs, BlkdebugEvent event) { IO_CODE(); + assert_bdrv_graph_readable(); + if (!bs || !bs->drv || !bs->drv->bdrv_co_debug_event) { return; } -- 2.40.1
From: Emanuele Giuseppe Esposito <eesposit@redhat.com> This adds GRAPH_RDLOCK annotations to declare that callers of amend callbacks in BlockDriver need to hold a reader lock for the graph. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Message-Id: <20230504115750.54437-17-kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block_int-common.h | 12 ++++++------ block/amend.c | 8 +++++++- 2 files changed, 13 insertions(+), 7 deletions(-) diff --git a/include/block/block_int-common.h b/include/block/block_int-common.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block_int-common.h +++ b/include/block/block_int-common.h @@ -XXX,XX +XXX,XX @@ struct BlockDriver { * to allow driver-specific initialization code that requires * the BQL, like setting up specific permission flags. */ - int (*bdrv_amend_pre_run)(BlockDriverState *bs, Error **errp); + int GRAPH_RDLOCK_PTR (*bdrv_amend_pre_run)( + BlockDriverState *bs, Error **errp); /* * This function is invoked under BQL after .bdrv_co_amend() * to allow cleaning up what was done in .bdrv_amend_pre_run(). */ - void (*bdrv_amend_clean)(BlockDriverState *bs); + void GRAPH_RDLOCK_PTR (*bdrv_amend_clean)(BlockDriverState *bs); /* * Return true if @to_replace can be replaced by a BDS with the @@ -XXX,XX +XXX,XX @@ struct BlockDriver { int (*bdrv_probe)(const uint8_t *buf, int buf_size, const char *filename); - int coroutine_fn (*bdrv_co_amend)(BlockDriverState *bs, - BlockdevAmendOptions *opts, - bool force, - Error **errp); + int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_amend)( + BlockDriverState *bs, BlockdevAmendOptions *opts, bool force, + Error **errp); /* aio */ BlockAIOCB * GRAPH_RDLOCK_PTR (*bdrv_aio_preadv)(BlockDriverState *bs, diff --git a/block/amend.c b/block/amend.c index XXXXXXX..XXXXXXX 100644 --- a/block/amend.c +++ b/block/amend.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn blockdev_amend_run(Job *job, Error **errp) { BlockdevAmendJob *s = container_of(job, BlockdevAmendJob, common); int ret; + GRAPH_RDLOCK_GUARD(); job_progress_set_remaining(&s->common, 1); ret = s->bs->drv->bdrv_co_amend(s->bs, s->opts, s->force, errp); @@ -XXX,XX +XXX,XX @@ static int coroutine_fn blockdev_amend_run(Job *job, Error **errp) return ret; } -static int blockdev_amend_pre_run(BlockdevAmendJob *s, Error **errp) +static int GRAPH_RDLOCK +blockdev_amend_pre_run(BlockdevAmendJob *s, Error **errp) { if (s->bs->drv->bdrv_amend_pre_run) { return s->bs->drv->bdrv_amend_pre_run(s->bs, errp); @@ -XXX,XX +XXX,XX @@ static void blockdev_amend_free(Job *job) { BlockdevAmendJob *s = container_of(job, BlockdevAmendJob, common); + bdrv_graph_rdlock_main_loop(); if (s->bs->drv->bdrv_amend_clean) { s->bs->drv->bdrv_amend_clean(s->bs); } + bdrv_graph_rdunlock_main_loop(); bdrv_unref(s->bs); } @@ -XXX,XX +XXX,XX @@ void qmp_x_blockdev_amend(const char *job_id, BlockDriver *drv = bdrv_find_format(fmt); BlockDriverState *bs; + GRAPH_RDLOCK_GUARD_MAINLOOP(); + bs = bdrv_lookup_bs(NULL, node_name, errp); if (!bs) { return; -- 2.40.1
This adds GRAPH_RDLOCK annotations to declare that callers of bdrv_query_bds_stats() need to hold a reader lock for the graph because it accesses the children list of a node. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Message-Id: <20230504115750.54437-18-kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/qapi.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/block/qapi.c b/block/qapi.c index XXXXXXX..XXXXXXX 100644 --- a/block/qapi.c +++ b/block/qapi.c @@ -XXX,XX +XXX,XX @@ static void bdrv_query_blk_stats(BlockDeviceStats *ds, BlockBackend *blk) = bdrv_latency_histogram_stats(&hgram[BLOCK_ACCT_FLUSH]); } -static BlockStats *bdrv_query_bds_stats(BlockDriverState *bs, - bool blk_level) +static BlockStats * GRAPH_RDLOCK +bdrv_query_bds_stats(BlockDriverState *bs, bool blk_level) { BdrvChild *parent_child; BlockDriverState *filter_or_cow_bs; @@ -XXX,XX +XXX,XX @@ BlockStatsList *qmp_query_blockstats(bool has_query_nodes, BlockBackend *blk; BlockDriverState *bs; + GRAPH_RDLOCK_GUARD_MAINLOOP(); + /* Just to be safe if query_nodes is not always initialized */ if (has_query_nodes && query_nodes) { for (bs = bdrv_next_node(NULL); bs; bs = bdrv_next_node(bs)) { -- 2.40.1
This adds GRAPH_RDLOCK annotations to declare that callers of bdrv_query_block_graph_info() need to hold a reader lock for the graph because it accesses the children list of a node. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Message-Id: <20230504115750.54437-19-kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/qapi.h | 7 ++++--- qemu-img.c | 2 ++ 2 files changed, 6 insertions(+), 3 deletions(-) diff --git a/include/block/qapi.h b/include/block/qapi.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/qapi.h +++ b/include/block/qapi.h @@ -XXX,XX +XXX,XX @@ #ifndef BLOCK_QAPI_H #define BLOCK_QAPI_H +#include "block/graph-lock.h" #include "block/snapshot.h" #include "qapi/qapi-types-block-core.h" @@ -XXX,XX +XXX,XX @@ void bdrv_query_image_info(BlockDriverState *bs, bool flat, bool skip_implicit_filters, Error **errp); -void bdrv_query_block_graph_info(BlockDriverState *bs, - BlockGraphInfo **p_info, - Error **errp); +void GRAPH_RDLOCK +bdrv_query_block_graph_info(BlockDriverState *bs, BlockGraphInfo **p_info, + Error **errp); void bdrv_snapshot_dump(QEMUSnapshotInfo *sn); void bdrv_image_info_specific_dump(ImageInfoSpecific *info_spec, diff --git a/qemu-img.c b/qemu-img.c index XXXXXXX..XXXXXXX 100644 --- a/qemu-img.c +++ b/qemu-img.c @@ -XXX,XX +XXX,XX @@ static BlockGraphInfoList *collect_image_info_list(bool image_opts, } bs = blk_bs(blk); + GRAPH_RDLOCK_GUARD_MAINLOOP(); + /* * Note that the returned BlockGraphInfo object will not have * information about this image's backing node, because we have opened -- 2.40.1
This adds GRAPH_RDLOCK annotations to declare that callers of bdrv_recurse_can_replace() need to hold a reader lock for the graph because it accesses the children list of a node. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Message-Id: <20230504115750.54437-20-kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-global-state.h | 5 +++-- include/block/block_int-common.h | 4 ++-- include/block/block_int-global-state.h | 4 ++-- block/blkverify.c | 5 +++-- block/mirror.c | 4 ++++ block/quorum.c | 4 ++-- blockdev.c | 3 +++ 7 files changed, 19 insertions(+), 10 deletions(-) diff --git a/include/block/block-global-state.h b/include/block/block-global-state.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block-global-state.h +++ b/include/block/block-global-state.h @@ -XXX,XX +XXX,XX @@ int bdrv_amend_options(BlockDriverState *bs_new, QemuOpts *opts, Error **errp); /* check if a named node can be replaced when doing drive-mirror */ -BlockDriverState *check_to_replace_node(BlockDriverState *parent_bs, - const char *node_name, Error **errp); +BlockDriverState * GRAPH_RDLOCK +check_to_replace_node(BlockDriverState *parent_bs, const char *node_name, + Error **errp); int no_coroutine_fn bdrv_activate(BlockDriverState *bs, Error **errp); diff --git a/include/block/block_int-common.h b/include/block/block_int-common.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block_int-common.h +++ b/include/block/block_int-common.h @@ -XXX,XX +XXX,XX @@ struct BlockDriver { * same data as @bs without it affecting @bs's behavior (that is, * without it being visible to @bs's parents). */ - bool (*bdrv_recurse_can_replace)(BlockDriverState *bs, - BlockDriverState *to_replace); + bool GRAPH_RDLOCK_PTR (*bdrv_recurse_can_replace)( + BlockDriverState *bs, BlockDriverState *to_replace); int (*bdrv_probe_device)(const char *filename); diff --git a/include/block/block_int-global-state.h b/include/block/block_int-global-state.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block_int-global-state.h +++ b/include/block/block_int-global-state.h @@ -XXX,XX +XXX,XX @@ int bdrv_child_try_set_perm(BdrvChild *c, uint64_t perm, uint64_t shared, */ int bdrv_child_refresh_perms(BlockDriverState *bs, BdrvChild *c, Error **errp); -bool bdrv_recurse_can_replace(BlockDriverState *bs, - BlockDriverState *to_replace); +bool GRAPH_RDLOCK bdrv_recurse_can_replace(BlockDriverState *bs, + BlockDriverState *to_replace); /* * Default implementation for BlockDriver.bdrv_child_perm() that can diff --git a/block/blkverify.c b/block/blkverify.c index XXXXXXX..XXXXXXX 100644 --- a/block/blkverify.c +++ b/block/blkverify.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn GRAPH_RDLOCK blkverify_co_flush(BlockDriverState *bs) return bdrv_co_flush(s->test_file->bs); } -static bool blkverify_recurse_can_replace(BlockDriverState *bs, - BlockDriverState *to_replace) +static bool GRAPH_RDLOCK +blkverify_recurse_can_replace(BlockDriverState *bs, + BlockDriverState *to_replace) { BDRVBlkverifyState *s = bs->opaque; diff --git a/block/mirror.c b/block/mirror.c index XXXXXXX..XXXXXXX 100644 --- a/block/mirror.c +++ b/block/mirror.c @@ -XXX,XX +XXX,XX @@ static int mirror_exit_common(Job *job) * Cannot use check_to_replace_node() here, because that would * check for an op blocker on @to_replace, and we have our own * there. + * + * TODO Pull out the writer lock from bdrv_replace_node() to here */ + bdrv_graph_rdlock_main_loop(); if (bdrv_recurse_can_replace(src, to_replace)) { bdrv_replace_node(to_replace, target_bs, &local_err); } else { @@ -XXX,XX +XXX,XX @@ static int mirror_exit_common(Job *job) "would not lead to an abrupt change of visible data", to_replace->node_name, target_bs->node_name); } + bdrv_graph_rdunlock_main_loop(); bdrv_drained_end(target_bs); if (local_err) { error_report_err(local_err); diff --git a/block/quorum.c b/block/quorum.c index XXXXXXX..XXXXXXX 100644 --- a/block/quorum.c +++ b/block/quorum.c @@ -XXX,XX +XXX,XX @@ static coroutine_fn GRAPH_RDLOCK int quorum_co_flush(BlockDriverState *bs) return result; } -static bool quorum_recurse_can_replace(BlockDriverState *bs, - BlockDriverState *to_replace) +static bool GRAPH_RDLOCK +quorum_recurse_can_replace(BlockDriverState *bs, BlockDriverState *to_replace) { BDRVQuorumState *s = bs->opaque; int i; diff --git a/blockdev.c b/blockdev.c index XXXXXXX..XXXXXXX 100644 --- a/blockdev.c +++ b/blockdev.c @@ -XXX,XX +XXX,XX @@ static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs, BlockDriverState *unfiltered_bs; int job_flags = JOB_DEFAULT; + GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); + if (!has_speed) { speed = 0; } -- 2.40.1
This adds GRAPH_RDLOCK annotations to declare that callers of bdrv_refresh_limits() need to hold a reader lock for the graph because it accesses the children list of a node. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Message-Id: <20230504115750.54437-21-kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-global-state.h | 5 ++++- include/block/block_int-common.h | 3 ++- block.c | 9 +++++++++ block/io.c | 1 - 4 files changed, 15 insertions(+), 3 deletions(-) diff --git a/include/block/block-global-state.h b/include/block/block-global-state.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block-global-state.h +++ b/include/block/block-global-state.h @@ -XXX,XX +XXX,XX @@ int bdrv_reopen_set_read_only(BlockDriverState *bs, bool read_only, BlockDriverState *bdrv_find_backing_image(BlockDriverState *bs, const char *backing_file); void bdrv_refresh_filename(BlockDriverState *bs); -void bdrv_refresh_limits(BlockDriverState *bs, Transaction *tran, Error **errp); + +void GRAPH_RDLOCK +bdrv_refresh_limits(BlockDriverState *bs, Transaction *tran, Error **errp); + int bdrv_commit(BlockDriverState *bs); int bdrv_make_empty(BdrvChild *c, Error **errp); int bdrv_change_backing_file(BlockDriverState *bs, const char *backing_file, diff --git a/include/block/block_int-common.h b/include/block/block_int-common.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block_int-common.h +++ b/include/block/block_int-common.h @@ -XXX,XX +XXX,XX @@ struct BlockDriver { int (*bdrv_debug_resume)(BlockDriverState *bs, const char *tag); bool (*bdrv_debug_is_suspended)(BlockDriverState *bs, const char *tag); - void (*bdrv_refresh_limits)(BlockDriverState *bs, Error **errp); + void GRAPH_RDLOCK_PTR (*bdrv_refresh_limits)( + BlockDriverState *bs, Error **errp); /* * Returns 1 if newly created images are guaranteed to contain only diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ bdrv_open_driver(BlockDriverState *bs, BlockDriver *drv, const char *node_name, return ret; } + bdrv_graph_rdlock_main_loop(); bdrv_refresh_limits(bs, NULL, &local_err); + bdrv_graph_rdunlock_main_loop(); + if (local_err) { error_propagate(errp, local_err); return -EINVAL; @@ -XXX,XX +XXX,XX @@ static int bdrv_set_file_or_backing_noperm(BlockDriverState *parent_bs, } out: + bdrv_graph_rdlock_main_loop(); bdrv_refresh_limits(parent_bs, tran, NULL); + bdrv_graph_rdunlock_main_loop(); return 0; } @@ -XXX,XX +XXX,XX @@ static void bdrv_reopen_commit(BDRVReopenState *reopen_state) qdict_del(bs->explicit_options, "backing"); qdict_del(bs->options, "backing"); + bdrv_graph_rdlock_main_loop(); bdrv_refresh_limits(bs, NULL, NULL); + bdrv_graph_rdunlock_main_loop(); bdrv_refresh_total_sectors(bs, bs->total_sectors); } @@ -XXX,XX +XXX,XX @@ int bdrv_append(BlockDriverState *bs_new, BlockDriverState *bs_top, out: tran_finalize(tran, ret); + bdrv_graph_rdlock_main_loop(); bdrv_refresh_limits(bs_top, NULL, NULL); + bdrv_graph_rdunlock_main_loop(); if (new_context && old_context != new_context) { aio_context_release(new_context); diff --git a/block/io.c b/block/io.c index XXXXXXX..XXXXXXX 100644 --- a/block/io.c +++ b/block/io.c @@ -XXX,XX +XXX,XX @@ void bdrv_refresh_limits(BlockDriverState *bs, Transaction *tran, Error **errp) bool have_limits; GLOBAL_STATE_CODE(); - assume_graph_lock(); /* FIXME */ if (tran) { BdrvRefreshLimitsState *s = g_new(BdrvRefreshLimitsState, 1); -- 2.40.1
From: Stefan Hajnoczi <stefanha@redhat.com> reader_count() is a performance bottleneck because the global aio_context_list_lock mutex causes thread contention. Put this debugging assertion behind a new ./configure --enable-debug-graph-lock option and disable it by default. The --enable-debug-graph-lock option is also enabled by the more general --enable-debug option. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Message-Id: <20230501173443.153062-1-stefanha@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- meson_options.txt | 2 ++ configure | 1 + block/graph-lock.c | 3 +++ meson.build | 2 ++ scripts/meson-buildoptions.sh | 4 ++++ 5 files changed, 12 insertions(+) diff --git a/meson_options.txt b/meson_options.txt index XXXXXXX..XXXXXXX 100644 --- a/meson_options.txt +++ b/meson_options.txt @@ -XXX,XX +XXX,XX @@ option('rng_none', type: 'boolean', value: false, description: 'dummy RNG, avoid using /dev/(u)random and getrandom()') option('coroutine_pool', type: 'boolean', value: true, description: 'coroutine freelist (better performance)') +option('debug_graph_lock', type: 'boolean', value: false, + description: 'graph lock debugging support') option('debug_mutex', type: 'boolean', value: false, description: 'mutex debugging support') option('debug_stack_usage', type: 'boolean', value: false, diff --git a/configure b/configure index XXXXXXX..XXXXXXX 100755 --- a/configure +++ b/configure @@ -XXX,XX +XXX,XX @@ for opt do --enable-debug) # Enable debugging options that aren't excessively noisy debug_tcg="yes" + meson_option_parse --enable-debug-graph-lock "" meson_option_parse --enable-debug-mutex "" meson_option_add -Doptimization=0 fortify_source="no" diff --git a/block/graph-lock.c b/block/graph-lock.c index XXXXXXX..XXXXXXX 100644 --- a/block/graph-lock.c +++ b/block/graph-lock.c @@ -XXX,XX +XXX,XX @@ void bdrv_graph_rdunlock_main_loop(void) void assert_bdrv_graph_readable(void) { + /* reader_count() is slow due to aio_context_list_lock lock contention */ +#ifdef CONFIG_DEBUG_GRAPH_LOCK assert(qemu_in_main_thread() || reader_count()); +#endif } void assert_bdrv_graph_writable(void) diff --git a/meson.build b/meson.build index XXXXXXX..XXXXXXX 100644 --- a/meson.build +++ b/meson.build @@ -XXX,XX +XXX,XX @@ if get_option('debug_stack_usage') and have_coroutine_pool have_coroutine_pool = false endif config_host_data.set10('CONFIG_COROUTINE_POOL', have_coroutine_pool) +config_host_data.set('CONFIG_DEBUG_GRAPH_LOCK', get_option('debug_graph_lock')) config_host_data.set('CONFIG_DEBUG_MUTEX', get_option('debug_mutex')) config_host_data.set('CONFIG_DEBUG_STACK_USAGE', get_option('debug_stack_usage')) config_host_data.set('CONFIG_GPROF', get_option('gprof')) @@ -XXX,XX +XXX,XX @@ summary_info += {'PIE': get_option('b_pie')} summary_info += {'static build': config_host.has_key('CONFIG_STATIC')} summary_info += {'malloc trim support': has_malloc_trim} summary_info += {'membarrier': have_membarrier} +summary_info += {'debug graph lock': get_option('debug_graph_lock')} summary_info += {'debug stack usage': get_option('debug_stack_usage')} summary_info += {'mutex debugging': get_option('debug_mutex')} summary_info += {'memory allocator': get_option('malloc')} diff --git a/scripts/meson-buildoptions.sh b/scripts/meson-buildoptions.sh index XXXXXXX..XXXXXXX 100644 --- a/scripts/meson-buildoptions.sh +++ b/scripts/meson-buildoptions.sh @@ -XXX,XX +XXX,XX @@ meson_options_help() { printf "%s\n" ' QEMU' printf "%s\n" ' --enable-cfi Control-Flow Integrity (CFI)' printf "%s\n" ' --enable-cfi-debug Verbose errors in case of CFI violation' + printf "%s\n" ' --enable-debug-graph-lock' + printf "%s\n" ' graph lock debugging support' printf "%s\n" ' --enable-debug-mutex mutex debugging support' printf "%s\n" ' --enable-debug-stack-usage' printf "%s\n" ' measure coroutine stack usage' @@ -XXX,XX +XXX,XX @@ _meson_option_parse() { --datadir=*) quote_sh "-Ddatadir=$2" ;; --enable-dbus-display) printf "%s" -Ddbus_display=enabled ;; --disable-dbus-display) printf "%s" -Ddbus_display=disabled ;; + --enable-debug-graph-lock) printf "%s" -Ddebug_graph_lock=true ;; + --disable-debug-graph-lock) printf "%s" -Ddebug_graph_lock=false ;; --enable-debug-mutex) printf "%s" -Ddebug_mutex=true ;; --disable-debug-mutex) printf "%s" -Ddebug_mutex=false ;; --enable-debug-stack-usage) printf "%s" -Ddebug_stack_usage=true ;; -- 2.40.1
The following changes since commit 13356edb87506c148b163b8c7eb0695647d00c2a: Merge tag 'block-pull-request' of https://gitlab.com/stefanha/qemu into staging (2023-01-24 09:45:33 +0000) are available in the Git repository at: https://repo.or.cz/qemu/kevin.git tags/for-upstream for you to fetch changes up to d570177b50c389f379f93183155a27d44856ab46: qemu-img: Change info key names for protocol nodes (2023-02-01 16:52:33 +0100) v4: - Fixed the 'qemu-img-close-errors' test case to run only on Linux and only with the file protocol, use qemu-io instead of truncate v3: - Make the compiler happier on BSD and CentOS Stream 8 v2: - Rebased to resolve merge conflicts in coroutine.h ---------------------------------------------------------------- Block layer patches - qemu-img info: Show protocol-level information - Move more functions to coroutines - Make coroutine annotations ready for static analysis - qemu-img: Fix exit code for errors closing the image - qcow2 bitmaps: Fix theoretical corruption in error path - pflash: Only load non-zero parts of backend image to save memory - Code cleanup and test case improvements ---------------------------------------------------------------- Alberto Faria (2): coroutine: annotate coroutine_fn for libclang block: Add no_coroutine_fn and coroutine_mixed_fn marker Emanuele Giuseppe Esposito (14): block-coroutine-wrapper: support void functions block: Convert bdrv_io_plug() to co_wrapper block: Convert bdrv_io_unplug() to co_wrapper block: Convert bdrv_is_inserted() to co_wrapper block: Rename refresh_total_sectors to bdrv_refresh_total_sectors block: Convert bdrv_refresh_total_sectors() to co_wrapper_mixed block-backend: use bdrv_getlength instead of blk_getlength block: use bdrv_co_refresh_total_sectors when possible block: Convert bdrv_get_allocated_file_size() to co_wrapper block: Convert bdrv_get_info() to co_wrapper_mixed block: Convert bdrv_eject() to co_wrapper block: Convert bdrv_lock_medium() to co_wrapper block: Convert bdrv_debug_event() to co_wrapper_mixed block: Rename bdrv_load/save_vmstate() to bdrv_co_load/save_vmstate() Hanna Reitz (12): block: Improve empty format-specific info dump block/file: Add file-specific image info block/vmdk: Change extent info type block: Split BlockNodeInfo off of ImageInfo qemu-img: Use BlockNodeInfo block/qapi: Let bdrv_query_image_info() recurse block/qapi: Introduce BlockGraphInfo block/qapi: Add indentation to bdrv_node_info_dump() iotests: Filter child node information iotests/106, 214, 308: Read only one size line qemu-img: Let info print block graph qemu-img: Change info key names for protocol nodes Kevin Wolf (4): qcow2: Fix theoretical corruption in store_bitmap() error path qemu-img commit: Report errors while closing the image qemu-img bitmap: Report errors while closing the image qemu-iotests: Test qemu-img bitmap/commit exit code on error Paolo Bonzini (2): qemu-io: do not reinvent the blk_pwrite_zeroes wheel block: remove bdrv_coroutine_enter Philippe Mathieu-Daudé (1): block/nbd: Add missing <qemu/bswap.h> include Thomas Huth (2): tests/qemu-iotests/312: Mark "quorum" as required driver tests/qemu-iotests/262: Check for availability of "blkverify" first Xiang Zheng (1): pflash: Only read non-zero parts of backend image qapi/block-core.json | 123 +++++++- include/block/block-common.h | 11 +- include/block/block-io.h | 41 ++- include/block/block_int-common.h | 26 +- include/block/block_int-io.h | 5 +- include/block/nbd.h | 1 + include/block/qapi.h | 14 +- include/qemu/osdep.h | 44 +++ include/sysemu/block-backend-io.h | 31 +- block.c | 88 +++--- block/blkdebug.c | 11 +- block/blkio.c | 15 +- block/blklogwrites.c | 6 +- block/blkreplay.c | 6 +- block/blkverify.c | 6 +- block/block-backend.c | 38 +-- block/commit.c | 4 +- block/copy-on-read.c | 18 +- block/crypto.c | 14 +- block/curl.c | 10 +- block/file-posix.c | 137 +++++---- block/file-win32.c | 18 +- block/filter-compress.c | 20 +- block/gluster.c | 23 +- block/io.c | 76 ++--- block/iscsi.c | 17 +- block/mirror.c | 6 +- block/monitor/block-hmp-cmds.c | 2 +- block/nbd.c | 8 +- block/nfs.c | 4 +- block/null.c | 13 +- block/nvme.c | 14 +- block/preallocate.c | 16 +- block/qapi.c | 317 ++++++++++++++++----- block/qcow.c | 5 +- block/qcow2-bitmap.c | 5 +- block/qcow2-refcount.c | 2 +- block/qcow2.c | 17 +- block/qed.c | 11 +- block/quorum.c | 8 +- block/raw-format.c | 25 +- block/rbd.c | 9 +- block/replication.c | 6 +- block/ssh.c | 4 +- block/throttle.c | 6 +- block/vdi.c | 7 +- block/vhdx.c | 5 +- block/vmdk.c | 22 +- block/vpc.c | 5 +- blockdev.c | 8 +- hw/block/block.c | 36 ++- hw/scsi/scsi-disk.c | 5 + qemu-img.c | 100 +++++-- qemu-io-cmds.c | 62 +--- tests/unit/test-block-iothread.c | 3 + scripts/block-coroutine-wrapper.py | 20 +- tests/qemu-iotests/iotests.py | 18 +- block/meson.build | 1 + tests/qemu-iotests/065 | 2 +- tests/qemu-iotests/106 | 4 +- tests/qemu-iotests/214 | 6 +- tests/qemu-iotests/262 | 3 +- tests/qemu-iotests/302.out | 5 + tests/qemu-iotests/308 | 4 +- tests/qemu-iotests/312 | 1 + tests/qemu-iotests/common.filter | 22 +- tests/qemu-iotests/common.rc | 22 +- tests/qemu-iotests/tests/qemu-img-close-errors | 96 +++++++ tests/qemu-iotests/tests/qemu-img-close-errors.out | 23 ++ 69 files changed, 1209 insertions(+), 552 deletions(-) create mode 100755 tests/qemu-iotests/tests/qemu-img-close-errors create mode 100644 tests/qemu-iotests/tests/qemu-img-close-errors.out