The following changes since commit 79b677d658d3d35e1e776826ac4abb28cdce69b8: Merge tag 'net-pull-request' of https://github.com/jasowang/qemu into staging (2023-02-21 11:28:31 +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 0f385a2420d2c3f8ae7ed65fbe2712027664059e: block/rbd: Add support for layered encryption (2023-02-23 19:49:35 +0100) ---------------------------------------------------------------- Block layer patches - Lock the graph, part 2 (BlockDriver callbacks) - virtio-scsi: fix SCSIDevice hot unplug with IOThread - rbd: Add support for layered encryption ---------------------------------------------------------------- Emanuele Giuseppe Esposito (5): block/qed: add missing graph rdlock in qed_need_check_timer_entry block: Mark bdrv_co_flush() and callers GRAPH_RDLOCK block: Mark bdrv_co_pdiscard() and callers GRAPH_RDLOCK block: Mark bdrv_co_copy_range() GRAPH_RDLOCK block: Mark bdrv_co_is_inserted() and callers GRAPH_RDLOCK Kevin Wolf (18): block: Make bdrv_can_set_read_only() static mirror: Fix access of uninitialised fields during start block: Mark bdrv_co_truncate() and callers GRAPH_RDLOCK block: Mark bdrv_co_block_status() and callers GRAPH_RDLOCK block: Mark bdrv_co_ioctl() and callers GRAPH_RDLOCK block: Mark bdrv_co_pwrite_zeroes() and callers GRAPH_RDLOCK block: Mark read/write in block/io.c GRAPH_RDLOCK block: Mark public read/write functions GRAPH_RDLOCK block: Mark bdrv_co_pwrite_sync() and callers GRAPH_RDLOCK block: Mark bdrv_co_do_pwrite_zeroes() GRAPH_RDLOCK block: Mark preadv_snapshot/snapshot_block_status GRAPH_RDLOCK block: Mark bdrv_co_create() and callers GRAPH_RDLOCK block: Mark bdrv_co_io_(un)plug() and callers GRAPH_RDLOCK block: Mark bdrv_co_eject/lock_medium() and callers GRAPH_RDLOCK block: Mark bdrv_(un)register_buf() GRAPH_RDLOCK block: Mark bdrv_co_delete_file() and callers GRAPH_RDLOCK block: Mark bdrv_*_dirty_bitmap() and callers GRAPH_RDLOCK block: Mark bdrv_co_refresh_total_sectors() and callers GRAPH_RDLOCK Or Ozeri (3): block/rbd: Remove redundant stack variable passphrase_len block/rbd: Add luks-any encryption opening option block/rbd: Add support for layered encryption Stefan Hajnoczi (3): scsi: protect req->aiocb with AioContext lock dma-helpers: prevent dma_blk_cb() vs dma_aio_cancel() race virtio-scsi: reset SCSI devices from main loop thread qapi/block-core.json | 27 +++++- block/coroutines.h | 2 +- block/qcow2.h | 27 ++++-- block/qed.h | 45 +++++---- include/block/block-copy.h | 6 +- include/block/block-global-state.h | 14 +-- include/block/block-io.h | 110 ++++++++++++---------- include/block/block_int-common.h | 173 ++++++++++++++++++---------------- include/block/block_int-io.h | 53 ++++++----- include/block/dirty-bitmap.h | 12 +-- include/hw/virtio/virtio-scsi.h | 11 ++- include/sysemu/block-backend-io.h | 7 +- block.c | 12 ++- block/backup.c | 3 + block/blkdebug.c | 19 ++-- block/blklogwrites.c | 35 ++++--- block/blkreplay.c | 24 +++-- block/blkverify.c | 5 +- block/block-backend.c | 39 +++++--- block/block-copy.c | 32 ++++--- block/bochs.c | 2 +- block/commit.c | 5 +- block/copy-before-write.c | 33 ++++--- block/copy-on-read.c | 44 +++++---- block/create.c | 9 +- block/crypto.c | 16 ++-- block/dirty-bitmap.c | 2 + block/file-posix.c | 27 +++--- block/file-win32.c | 7 +- block/filter-compress.c | 36 ++++--- block/io.c | 108 +++++++++++++-------- block/iscsi.c | 28 +++--- block/mirror.c | 59 ++++++++---- block/parallels.c | 33 +++---- block/preallocate.c | 38 ++++---- block/qcow.c | 46 +++++---- block/qcow2-cluster.c | 17 ++-- block/qcow2.c | 136 +++++++++++++++------------ block/qed-check.c | 3 +- block/qed-table.c | 10 +- block/qed.c | 101 ++++++++++---------- block/quorum.c | 62 +++++++----- block/raw-format.c | 76 ++++++++------- block/rbd.c | 188 ++++++++++++++++++++++++++++++++++--- block/replication.c | 18 ++-- block/snapshot-access.c | 8 +- block/stream.c | 40 ++++---- block/throttle.c | 36 ++++--- block/vdi.c | 11 +-- block/vhdx.c | 18 ++-- block/vmdk.c | 132 ++++++++++++-------------- block/vpc.c | 11 +-- hw/scsi/scsi-disk.c | 23 +++-- hw/scsi/scsi-generic.c | 11 ++- hw/scsi/virtio-scsi.c | 169 ++++++++++++++++++++++++++------- qemu-img.c | 8 +- softmmu/dma-helpers.c | 12 ++- tests/unit/test-bdrv-drain.c | 20 ++-- tests/unit/test-block-iothread.c | 3 +- 59 files changed, 1355 insertions(+), 907 deletions(-) It is never called outside of block.c. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230203152202.49054-2-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-io.h | 2 -- block.c | 4 ++-- 2 files changed, 2 insertions(+), 4 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 @@ int bdrv_is_allocated_above(BlockDriverState *top, BlockDriverState *base, int coroutine_fn bdrv_co_is_zero_fast(BlockDriverState *bs, int64_t offset, int64_t bytes); -int bdrv_can_set_read_only(BlockDriverState *bs, bool read_only, - bool ignore_allow_rdw, Error **errp); int bdrv_apply_auto_read_only(BlockDriverState *bs, const char *errmsg, Error **errp); bool bdrv_is_read_only(BlockDriverState *bs); diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ bool bdrv_is_read_only(BlockDriverState *bs) return !(bs->open_flags & BDRV_O_RDWR); } -int bdrv_can_set_read_only(BlockDriverState *bs, bool read_only, - bool ignore_allow_rdw, Error **errp) +static int bdrv_can_set_read_only(BlockDriverState *bs, bool read_only, + bool ignore_allow_rdw, Error **errp) { IO_CODE(); -- 2.39.2 bdrv_mirror_top_pwritev() accesses the job object when active mirroring is enabled. It disables this code during early initialisation while s->job isn't set yet. However, s->job is still set way too early when the job object isn't fully initialised. For example, &s->ops_in_flight isn't initialised yet and the in_flight bitmap doesn't exist yet. This causes crashes when a write request comes in too early. Move the assignment of s->job to when the mirror job is actually fully initialised to make sure that the mirror_top driver doesn't access it too early. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230203152202.49054-3-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/mirror.c | 8 +++++++- 1 file changed, 7 insertions(+), 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 int coroutine_fn mirror_run(Job *job, Error **errp) { MirrorBlockJob *s = container_of(job, MirrorBlockJob, common.job); BlockDriverState *bs = s->mirror_top_bs->backing->bs; + MirrorBDSOpaque *mirror_top_opaque = s->mirror_top_bs->opaque; BlockDriverState *target_bs = blk_bs(s->target); bool need_drain = true; BlockDeviceIoStatus iostatus; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn mirror_run(Job *job, Error **errp) } } + /* + * Only now the job is fully initialised and mirror_top_bs should start + * accessing it. + */ + mirror_top_opaque->job = s; + assert(!s->dbi); s->dbi = bdrv_dirty_iter_new(s->dirty_bitmap); for (;;) { @@ -XXX,XX +XXX,XX @@ static BlockJob *mirror_start_job( if (!s) { goto fail; } - bs_opaque->job = s; /* The block job now has a reference to this node */ bdrv_unref(mirror_top_bs); -- 2.39.2 This adds GRAPH_RDLOCK annotations to declare that callers of bdrv_co_truncate() need to hold a reader lock for the graph. For some places, we know that they will hold the lock, but we don't have the GRAPH_RDLOCK annotations yet. In this case, add assume_graph_lock() with a FIXME comment. These places will be removed once everything is properly annotated. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230203152202.49054-4-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-io.h | 6 +++--- include/block/block_int-common.h | 7 ++++--- block/block-backend.c | 1 + block/crypto.c | 2 +- block/io.c | 1 + block/parallels.c | 14 ++++++++------ block/preallocate.c | 2 +- block/qcow.c | 17 ++++++++++++----- block/qcow2.c | 14 ++++++++------ block/raw-format.c | 6 +++--- block/vmdk.c | 2 ++ 11 files changed, 44 insertions(+), 28 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 @@ int coroutine_fn bdrv_co_pwrite_sync(BdrvChild *child, int64_t offset, int coroutine_fn bdrv_co_pwrite_zeroes(BdrvChild *child, int64_t offset, int64_t bytes, BdrvRequestFlags flags); -int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact, - PreallocMode prealloc, BdrvRequestFlags flags, - Error **errp); +int coroutine_fn GRAPH_RDLOCK +bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact, + PreallocMode prealloc, BdrvRequestFlags flags, Error **errp); int64_t coroutine_fn bdrv_co_nb_sectors(BlockDriverState *bs); int64_t co_wrapper_mixed bdrv_nb_sectors(BlockDriverState *bs); 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 { * If @exact is true and this function fails but would succeed * with @exact = false, it should return -ENOTSUP. */ - int coroutine_fn (*bdrv_co_truncate)(BlockDriverState *bs, int64_t offset, - bool exact, PreallocMode prealloc, - BdrvRequestFlags flags, Error **errp); + int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_truncate)( + BlockDriverState *bs, int64_t offset, bool exact, + PreallocMode prealloc, BdrvRequestFlags flags, Error **errp); + int64_t coroutine_fn (*bdrv_co_getlength)(BlockDriverState *bs); int64_t coroutine_fn (*bdrv_co_get_allocated_file_size)( BlockDriverState *bs); 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 @@ int coroutine_fn blk_co_truncate(BlockBackend *blk, int64_t offset, bool exact, Error **errp) { IO_OR_GS_CODE(); + GRAPH_RDLOCK_GUARD(); if (!blk_is_available(blk)) { error_setg(errp, "No medium inserted"); 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 @@ block_crypto_co_create_generic(BlockDriverState *bs, int64_t size, return ret; } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK block_crypto_co_truncate(BlockDriverState *bs, int64_t offset, bool exact, PreallocMode prealloc, BdrvRequestFlags flags, Error **errp) 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 @@ int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact, int64_t old_size, new_bytes; int ret; IO_CODE(); + assert_bdrv_graph_readable(); /* if bs->drv == NULL, bs is closed, so there's nothing to do here */ if (!drv) { 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 int64_t block_status(BDRVParallelsState *s, int64_t sector_num, return start_off; } -static coroutine_fn int64_t allocate_clusters(BlockDriverState *bs, - int64_t sector_num, - int nb_sectors, int *pnum) +static int64_t coroutine_fn GRAPH_RDLOCK +allocate_clusters(BlockDriverState *bs, int64_t sector_num, + int nb_sectors, int *pnum) { int ret = 0; BDRVParallelsState *s = bs->opaque; @@ -XXX,XX +XXX,XX @@ static coroutine_fn int parallels_co_writev(BlockDriverState *bs, QEMUIOVector hd_qiov; int ret = 0; + assume_graph_lock(); /* FIXME */ + qemu_iovec_init(&hd_qiov, qiov->niov); while (nb_sectors > 0) { @@ -XXX,XX +XXX,XX @@ static coroutine_fn int parallels_co_readv(BlockDriverState *bs, } -static int coroutine_fn parallels_co_check(BlockDriverState *bs, - BdrvCheckResult *res, - BdrvCheckMode fix) +static int coroutine_fn GRAPH_RDLOCK +parallels_co_check(BlockDriverState *bs, BdrvCheckResult *res, + BdrvCheckMode fix) { BDRVParallelsState *s = bs->opaque; int64_t size, prev_off, high_off; diff --git a/block/preallocate.c b/block/preallocate.c index XXXXXXX..XXXXXXX 100644 --- a/block/preallocate.c +++ b/block/preallocate.c @@ -XXX,XX +XXX,XX @@ static coroutine_fn int preallocate_co_pwritev_part(BlockDriverState *bs, flags); } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK preallocate_co_truncate(BlockDriverState *bs, int64_t offset, bool exact, PreallocMode prealloc, BdrvRequestFlags flags, Error **errp) 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 qcow_reopen_prepare(BDRVReopenState *state, * return 0 if not allocated, 1 if *result is assigned, and negative * errno on failure. */ -static int coroutine_fn get_cluster_offset(BlockDriverState *bs, - uint64_t offset, int allocate, - int compressed_size, - int n_start, int n_end, - uint64_t *result) +static int coroutine_fn GRAPH_RDLOCK +get_cluster_offset(BlockDriverState *bs, uint64_t offset, int allocate, + int compressed_size, int n_start, int n_end, + uint64_t *result) { BDRVQcowState *s = bs->opaque; int min_index, i, j, l1_index, l2_index, ret; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qcow_co_block_status(BlockDriverState *bs, int64_t n; uint64_t cluster_offset; + assume_graph_lock(); /* FIXME */ + qemu_co_mutex_lock(&s->lock); ret = get_cluster_offset(bs, offset, 0, 0, 0, 0, &cluster_offset); qemu_co_mutex_unlock(&s->lock); @@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow_co_preadv(BlockDriverState *bs, int64_t offset, uint8_t *buf; void *orig_buf; + assume_graph_lock(); /* FIXME */ + if (qiov->niov > 1) { buf = orig_buf = qemu_try_blockalign(bs, qiov->size); if (buf == NULL) { @@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow_co_pwritev(BlockDriverState *bs, int64_t offset, uint8_t *buf; void *orig_buf; + assume_graph_lock(); /* FIXME */ + s->cluster_cache_offset = -1; /* disable compressed cache */ /* We must always copy the iov when encrypting, so we @@ -XXX,XX +XXX,XX @@ qcow_co_pwritev_compressed(BlockDriverState *bs, int64_t offset, int64_t bytes, uint8_t *buf, *out_buf; uint64_t cluster_offset; + assume_graph_lock(); /* FIXME */ + buf = qemu_blockalign(bs, s->cluster_size); if (bytes != s->cluster_size) { if (bytes > s->cluster_size || 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 int qcow2_set_up_encryption(BlockDriverState *bs, * * Returns: 0 on success, -errno on failure. */ -static int coroutine_fn preallocate_co(BlockDriverState *bs, uint64_t offset, - uint64_t new_length, PreallocMode mode, - Error **errp) +static int coroutine_fn GRAPH_RDLOCK +preallocate_co(BlockDriverState *bs, uint64_t offset, uint64_t new_length, + PreallocMode mode, Error **errp) { BDRVQcow2State *s = bs->opaque; uint64_t bytes; @@ -XXX,XX +XXX,XX @@ fail: return ret; } -static int coroutine_fn qcow2_co_truncate(BlockDriverState *bs, int64_t offset, - bool exact, PreallocMode prealloc, - BdrvRequestFlags flags, Error **errp) +static int coroutine_fn GRAPH_RDLOCK +qcow2_co_truncate(BlockDriverState *bs, int64_t offset, bool exact, + PreallocMode prealloc, BdrvRequestFlags flags, Error **errp) { BDRVQcow2State *s = bs->opaque; uint64_t old_length; @@ -XXX,XX +XXX,XX @@ qcow2_co_pwritev_compressed_part(BlockDriverState *bs, AioTaskPool *aio = NULL; int ret = 0; + assume_graph_lock(); /* FIXME */ + if (has_data_file(bs)) { return -ENOTSUP; } 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 void raw_refresh_limits(BlockDriverState *bs, Error **errp) } } -static int coroutine_fn raw_co_truncate(BlockDriverState *bs, int64_t offset, - bool exact, PreallocMode prealloc, - BdrvRequestFlags flags, Error **errp) +static int coroutine_fn GRAPH_RDLOCK +raw_co_truncate(BlockDriverState *bs, int64_t offset, bool exact, + PreallocMode prealloc, BdrvRequestFlags flags, Error **errp) { BDRVRawState *s = bs->opaque; 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 int coroutine_fn vmdk_co_pwritev_compressed(BlockDriverState *bs, int64_t offset, int64_t bytes, QEMUIOVector *qiov) { + assume_graph_lock(); /* FIXME */ + if (bytes == 0) { /* The caller will write bytes 0 to signal EOF. * When receive it, we align EOF to a sector boundary. */ -- 2.39.2 This adds GRAPH_RDLOCK annotations to declare that callers of bdrv_co_block_status() need to hold a reader lock for the graph. For some places, we know that they will hold the lock, but we don't have the GRAPH_RDLOCK annotations yet. In this case, add assume_graph_lock() with a FIXME comment. These places will be removed once everything is properly annotated. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230203152202.49054-5-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/coroutines.h | 2 +- include/block/block-copy.h | 6 +++--- include/block/block-io.h | 22 +++++++++++----------- include/block/block_int-common.h | 3 ++- block/backup.c | 3 +++ block/block-backend.c | 2 ++ block/block-copy.c | 19 +++++++++++-------- block/io.c | 13 ++++++++----- block/mirror.c | 14 +++++++++----- block/qcow.c | 11 ++++------- block/quorum.c | 9 ++++----- block/stream.c | 32 ++++++++++++++++++-------------- qemu-img.c | 4 +++- tests/unit/test-block-iothread.c | 3 ++- 14 files changed, 81 insertions(+), 62 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_check(BlockDriverState *bs, BdrvCheckResult *res, BdrvCheckMode fix); int coroutine_fn GRAPH_RDLOCK bdrv_co_invalidate_cache(BlockDriverState *bs, Error **errp); -int coroutine_fn +int coroutine_fn GRAPH_RDLOCK bdrv_co_common_block_status_above(BlockDriverState *bs, BlockDriverState *base, bool include_base, diff --git a/include/block/block-copy.h b/include/block/block-copy.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block-copy.h +++ b/include/block/block-copy.h @@ -XXX,XX +XXX,XX @@ void block_copy_set_progress_meter(BlockCopyState *s, ProgressMeter *pm); void block_copy_state_free(BlockCopyState *s); void block_copy_reset(BlockCopyState *s, int64_t offset, int64_t bytes); -int64_t coroutine_fn block_copy_reset_unallocated(BlockCopyState *s, - int64_t offset, - int64_t *count); + +int64_t coroutine_fn GRAPH_RDLOCK +block_copy_reset_unallocated(BlockCopyState *s, int64_t offset, int64_t *count); int coroutine_fn block_copy(BlockCopyState *s, int64_t offset, int64_t bytes, bool ignore_ratelimit, uint64_t timeout_ns, 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 @@ int bdrv_block_status(BlockDriverState *bs, int64_t offset, int64_t bytes, int64_t *pnum, int64_t *map, BlockDriverState **file); -int coroutine_fn bdrv_co_block_status_above(BlockDriverState *bs, - BlockDriverState *base, - int64_t offset, int64_t bytes, - int64_t *pnum, int64_t *map, - BlockDriverState **file); +int coroutine_fn GRAPH_RDLOCK +bdrv_co_block_status_above(BlockDriverState *bs, BlockDriverState *base, + int64_t offset, int64_t bytes, int64_t *pnum, + int64_t *map, BlockDriverState **file); int bdrv_block_status_above(BlockDriverState *bs, BlockDriverState *base, int64_t offset, int64_t bytes, int64_t *pnum, int64_t *map, BlockDriverState **file); -int coroutine_fn bdrv_co_is_allocated(BlockDriverState *bs, int64_t offset, - int64_t bytes, int64_t *pnum); +int coroutine_fn GRAPH_RDLOCK +bdrv_co_is_allocated(BlockDriverState *bs, int64_t offset, int64_t bytes, + int64_t *pnum); int bdrv_is_allocated(BlockDriverState *bs, int64_t offset, int64_t bytes, int64_t *pnum); -int coroutine_fn bdrv_co_is_allocated_above(BlockDriverState *top, - BlockDriverState *base, - bool include_base, int64_t offset, - int64_t bytes, int64_t *pnum); +int coroutine_fn GRAPH_RDLOCK +bdrv_co_is_allocated_above(BlockDriverState *top, BlockDriverState *base, + bool include_base, int64_t offset, int64_t bytes, + int64_t *pnum); int bdrv_is_allocated_above(BlockDriverState *top, BlockDriverState *base, bool include_base, int64_t offset, int64_t bytes, int64_t *pnum); 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 { * *pnum value for the block-status cache on protocol nodes, prior * to clamping *pnum for return to its caller. */ - int coroutine_fn (*bdrv_co_block_status)(BlockDriverState *bs, + int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_block_status)( + BlockDriverState *bs, bool want_zero, int64_t offset, int64_t bytes, int64_t *pnum, int64_t *map, BlockDriverState **file); diff --git a/block/backup.c b/block/backup.c index XXXXXXX..XXXXXXX 100644 --- a/block/backup.c +++ b/block/backup.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn backup_run(Job *job, Error **errp) return -ECANCELED; } + /* rdlock protects the subsequent call to bdrv_is_allocated() */ + bdrv_graph_co_rdlock(); ret = block_copy_reset_unallocated(s->bcs, offset, &count); + bdrv_graph_co_rdunlock(); if (ret < 0) { return ret; } 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 @@ int coroutine_fn blk_co_block_status_above(BlockBackend *blk, BlockDriverState **file) { IO_CODE(); + GRAPH_RDLOCK_GUARD(); return bdrv_co_block_status_above(blk_bs(blk), base, offset, bytes, pnum, map, file); } @@ -XXX,XX +XXX,XX @@ int coroutine_fn blk_co_is_allocated_above(BlockBackend *blk, int64_t bytes, int64_t *pnum) { IO_CODE(); + GRAPH_RDLOCK_GUARD(); return bdrv_co_is_allocated_above(blk_bs(blk), base, include_base, offset, bytes, pnum); } diff --git a/block/block-copy.c b/block/block-copy.c index XXXXXXX..XXXXXXX 100644 --- a/block/block-copy.c +++ b/block/block-copy.c @@ -XXX,XX +XXX,XX @@ static coroutine_fn int block_copy_task_entry(AioTask *task) return ret; } -static coroutine_fn int block_copy_block_status(BlockCopyState *s, - int64_t offset, - int64_t bytes, int64_t *pnum) +static coroutine_fn GRAPH_RDLOCK +int block_copy_block_status(BlockCopyState *s, int64_t offset, int64_t bytes, + int64_t *pnum) { int64_t num; BlockDriverState *base; @@ -XXX,XX +XXX,XX @@ static coroutine_fn int block_copy_block_status(BlockCopyState *s, * Check if the cluster starting at offset is allocated or not. * return via pnum the number of contiguous clusters sharing this allocation. */ -static int coroutine_fn block_copy_is_cluster_allocated(BlockCopyState *s, - int64_t offset, - int64_t *pnum) +static int coroutine_fn GRAPH_RDLOCK +block_copy_is_cluster_allocated(BlockCopyState *s, int64_t offset, + int64_t *pnum) { BlockDriverState *bs = s->source->bs; int64_t count, total_count = 0; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn block_copy_is_cluster_allocated(BlockCopyState *s, assert(QEMU_IS_ALIGNED(offset, s->cluster_size)); while (true) { + /* protected in backup_run() */ ret = bdrv_co_is_allocated(bs, offset, bytes, &count); if (ret < 0) { return ret; @@ -XXX,XX +XXX,XX @@ int64_t coroutine_fn block_copy_reset_unallocated(BlockCopyState *s, * Returns 1 if dirty clusters found and successfully copied, 0 if no dirty * clusters found and -errno on failure. */ -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK block_copy_dirty_clusters(BlockCopyCallState *call_state) { BlockCopyState *s = call_state->s; @@ -XXX,XX +XXX,XX @@ void block_copy_kick(BlockCopyCallState *call_state) * it means that some I/O operation failed in context of _this_ block_copy call, * not some parallel operation. */ -static int coroutine_fn block_copy_common(BlockCopyCallState *call_state) +static int coroutine_fn GRAPH_RDLOCK +block_copy_common(BlockCopyCallState *call_state) { int ret; BlockCopyState *s = call_state->s; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn block_copy_common(BlockCopyCallState *call_state) static void coroutine_fn block_copy_async_co_entry(void *opaque) { + GRAPH_RDLOCK_GUARD(); block_copy_common(opaque); } 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 @@ int bdrv_flush_all(void) * BDRV_BLOCK_OFFSET_VALID bit is set, 'map' and 'file' (if non-NULL) are * set to the host mapping and BDS corresponding to the guest offset. */ -static int coroutine_fn bdrv_co_block_status(BlockDriverState *bs, - bool want_zero, - int64_t offset, int64_t bytes, - int64_t *pnum, int64_t *map, - BlockDriverState **file) +static int coroutine_fn GRAPH_RDLOCK +bdrv_co_block_status(BlockDriverState *bs, bool want_zero, + int64_t offset, int64_t bytes, + int64_t *pnum, int64_t *map, BlockDriverState **file) { int64_t total_size; int64_t n; /* bytes */ @@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_block_status(BlockDriverState *bs, bool has_filtered_child; assert(pnum); + assert_bdrv_graph_readable(); *pnum = 0; total_size = bdrv_getlength(bs); if (total_size < 0) { @@ -XXX,XX +XXX,XX @@ bdrv_co_common_block_status_above(BlockDriverState *bs, IO_CODE(); assert(!include_base || base); /* Can't include NULL base */ + assert_bdrv_graph_readable(); if (!depth) { depth = &dummy; @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_is_zero_fast(BlockDriverState *bs, int64_t offset, int64_t pnum = bytes; IO_CODE(); + assume_graph_lock(); /* FIXME */ + if (!bytes) { return 1; } 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) MirrorMethod mirror_method = MIRROR_METHOD_COPY; assert(!(offset % s->granularity)); - ret = bdrv_block_status_above(source, NULL, offset, - nb_chunks * s->granularity, - &io_bytes, NULL, NULL); + WITH_GRAPH_RDLOCK_GUARD() { + ret = bdrv_block_status_above(source, NULL, offset, + nb_chunks * s->granularity, + &io_bytes, NULL, NULL); + } if (ret < 0) { io_bytes = MIN(nb_chunks * s->granularity, max_io_bytes); } else if (ret & BDRV_BLOCK_DATA) { @@ -XXX,XX +XXX,XX @@ static int coroutine_fn mirror_dirty_init(MirrorBlockJob *s) return 0; } - ret = bdrv_is_allocated_above(bs, s->base_overlay, true, offset, bytes, - &count); + WITH_GRAPH_RDLOCK_GUARD() { + ret = bdrv_is_allocated_above(bs, s->base_overlay, true, offset, + bytes, &count); + } if (ret < 0) { 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 @@ get_cluster_offset(BlockDriverState *bs, uint64_t offset, int allocate, return 1; } -static int coroutine_fn qcow_co_block_status(BlockDriverState *bs, - bool want_zero, - int64_t offset, int64_t bytes, - int64_t *pnum, int64_t *map, - BlockDriverState **file) +static int coroutine_fn GRAPH_RDLOCK +qcow_co_block_status(BlockDriverState *bs, bool want_zero, + int64_t offset, int64_t bytes, int64_t *pnum, + int64_t *map, BlockDriverState **file) { BDRVQcowState *s = bs->opaque; int index_in_cluster, ret; int64_t n; uint64_t cluster_offset; - assume_graph_lock(); /* FIXME */ - qemu_co_mutex_lock(&s->lock); ret = get_cluster_offset(bs, offset, 0, 0, 0, 0, &cluster_offset); qemu_co_mutex_unlock(&s->lock); 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 void quorum_child_perm(BlockDriverState *bs, BdrvChild *c, * return BDRV_BLOCK_ZERO if *all* children agree that a certain * region contains zeroes, and BDRV_BLOCK_DATA otherwise. */ -static int coroutine_fn quorum_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 +quorum_co_block_status(BlockDriverState *bs, bool want_zero, + int64_t offset, int64_t count, + int64_t *pnum, int64_t *map, BlockDriverState **file) { BDRVQuorumState *s = bs->opaque; int i, ret; diff --git a/block/stream.c b/block/stream.c index XXXXXXX..XXXXXXX 100644 --- a/block/stream.c +++ b/block/stream.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn stream_run(Job *job, Error **errp) copy = false; - ret = bdrv_is_allocated(unfiltered_bs, offset, STREAM_CHUNK, &n); - if (ret == 1) { - /* Allocated in the top, no need to copy. */ - } else if (ret >= 0) { - /* Copy if allocated in the intermediate images. Limit to the - * known-unallocated area [offset, offset+n*BDRV_SECTOR_SIZE). */ - ret = bdrv_is_allocated_above(bdrv_cow_bs(unfiltered_bs), - s->base_overlay, true, - offset, n, &n); - /* Finish early if end of backing file has been reached */ - if (ret == 0 && n == 0) { - n = len - offset; + WITH_GRAPH_RDLOCK_GUARD() { + ret = bdrv_is_allocated(unfiltered_bs, offset, STREAM_CHUNK, &n); + if (ret == 1) { + /* Allocated in the top, no need to copy. */ + } else if (ret >= 0) { + /* + * Copy if allocated in the intermediate images. Limit to the + * known-unallocated area [offset, offset+n*BDRV_SECTOR_SIZE). + */ + ret = bdrv_is_allocated_above(bdrv_cow_bs(unfiltered_bs), + s->base_overlay, true, + offset, n, &n); + /* Finish early if end of backing file has been reached */ + if (ret == 0 && n == 0) { + n = len - offset; + } + + copy = (ret > 0); } - - copy = (ret > 0); } trace_stream_one_iteration(s, offset, n, ret); if (copy) { 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 void coroutine_fn convert_co_do_copy(void *opaque) qemu_co_mutex_unlock(&s->lock); break; } - n = convert_iteration_sectors(s, s->sector_num); + WITH_GRAPH_RDLOCK_GUARD() { + n = convert_iteration_sectors(s, s->sector_num); + } if (n < 0) { qemu_co_mutex_unlock(&s->lock); s->ret = n; diff --git a/tests/unit/test-block-iothread.c b/tests/unit/test-block-iothread.c index XXXXXXX..XXXXXXX 100644 --- a/tests/unit/test-block-iothread.c +++ b/tests/unit/test-block-iothread.c @@ -XXX,XX +XXX,XX @@ static void test_sync_op_blk_truncate(BlockBackend *blk) g_assert_cmpint(ret, ==, -EINVAL); } -static void test_sync_op_block_status(BdrvChild *c) +/* Disable TSA to make bdrv_test.bdrv_co_block_status writable */ +static void TSA_NO_TSA test_sync_op_block_status(BdrvChild *c) { int ret; int64_t n; -- 2.39.2 This adds GRAPH_RDLOCK annotations to declare that callers of bdrv_co_ioctl() need to hold a reader lock for the graph. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230203152202.49054-6-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-io.h | 3 ++- include/block/block_int-common.h | 9 +++++---- block/block-backend.c | 1 + block/io.c | 1 + block/raw-format.c | 4 ++-- 5 files changed, 11 insertions(+), 7 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 bdrv_aio_cancel(BlockAIOCB *acb); void bdrv_aio_cancel_async(BlockAIOCB *acb); /* sg packet commands */ -int coroutine_fn bdrv_co_ioctl(BlockDriverState *bs, int req, void *buf); +int coroutine_fn GRAPH_RDLOCK +bdrv_co_ioctl(BlockDriverState *bs, int req, void *buf); /* Ensure contents are flushed to disk. */ int coroutine_fn bdrv_co_flush(BlockDriverState *bs); 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 coroutine_fn (*bdrv_co_lock_medium)(BlockDriverState *bs, bool locked); /* to control generic scsi devices */ - BlockAIOCB *(*bdrv_aio_ioctl)(BlockDriverState *bs, - unsigned long int req, void *buf, + BlockAIOCB *coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_aio_ioctl)( + BlockDriverState *bs, unsigned long int req, void *buf, BlockCompletionFunc *cb, void *opaque); - int coroutine_fn (*bdrv_co_ioctl)(BlockDriverState *bs, - unsigned long int req, void *buf); + + int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_ioctl)( + BlockDriverState *bs, unsigned long int req, void *buf); /* * Returns 0 for completed check, -errno for internal errors. 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 @@ blk_co_do_ioctl(BlockBackend *blk, unsigned long int req, void *buf) IO_CODE(); blk_wait_while_drained(blk); + GRAPH_RDLOCK_GUARD(); if (!blk_is_available(blk)) { return -ENOMEDIUM; 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 @@ int coroutine_fn bdrv_co_ioctl(BlockDriverState *bs, int req, void *buf) }; BlockAIOCB *acb; IO_CODE(); + assert_bdrv_graph_readable(); bdrv_inc_in_flight(bs); if (!drv || (!drv->bdrv_aio_ioctl && !drv->bdrv_co_ioctl)) { 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 void coroutine_fn raw_co_lock_medium(BlockDriverState *bs, bool locked) bdrv_co_lock_medium(bs->file->bs, locked); } -static int coroutine_fn raw_co_ioctl(BlockDriverState *bs, - unsigned long int req, void *buf) +static int coroutine_fn GRAPH_RDLOCK +raw_co_ioctl(BlockDriverState *bs, unsigned long int req, void *buf) { BDRVRawState *s = bs->opaque; if (s->offset || s->has_size) { -- 2.39.2 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> This function is called in two different places: - timer callback, which does not take the graph rdlock. - bdrv_qed_drain_begin(), which is .bdrv_drain_begin() callback documented as function that does not take the lock. Since it calls recursive functions that traverse the graph, we need to protect them with the graph rdlock. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230203152202.49054-7-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/qed.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) 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 void coroutine_fn qed_unplug_allocating_write_reqs(BDRVQEDState *s) qemu_co_mutex_unlock(&s->table_lock); } -static void coroutine_fn qed_need_check_timer(BDRVQEDState *s) +static void coroutine_fn GRAPH_RDLOCK qed_need_check_timer(BDRVQEDState *s) { int ret; trace_qed_need_check_timer_cb(s); + assert_bdrv_graph_readable(); if (!qed_plug_allocating_write_reqs(s)) { return; @@ -XXX,XX +XXX,XX @@ static void coroutine_fn qed_need_check_timer(BDRVQEDState *s) static void coroutine_fn qed_need_check_timer_entry(void *opaque) { BDRVQEDState *s = opaque; + GRAPH_RDLOCK_GUARD(); qed_need_check_timer(opaque); bdrv_dec_in_flight(s->bs); -- 2.39.2 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> This adds GRAPH_RDLOCK annotations to declare that callers of bdrv_co_flush() need to hold a reader lock for the graph. For some places, we know that they will hold the lock, but we don't have the GRAPH_RDLOCK annotations yet. In this case, add assume_graph_lock() with a FIXME comment. These places will be removed once everything is properly annotated. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230203152202.49054-8-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/qcow2.h | 5 +++- block/qed.h | 29 +++++++++++++-------- include/block/block-io.h | 2 +- include/block/block_int-common.h | 12 +++++---- block/blkdebug.c | 2 +- block/blklogwrites.c | 21 ++++++++++----- block/blkreplay.c | 2 +- block/blkverify.c | 2 +- block/block-backend.c | 3 ++- block/copy-before-write.c | 2 +- block/file-posix.c | 4 +-- block/io.c | 7 +++++ block/mirror.c | 2 +- block/preallocate.c | 2 +- block/qed-check.c | 3 ++- block/qed-table.c | 6 ++--- block/qed.c | 44 +++++++++++++++++++------------- block/quorum.c | 2 +- block/throttle.c | 2 +- block/vmdk.c | 6 +++-- 20 files changed, 98 insertions(+), 60 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 coroutine_fn qcow2_detect_metadata_preallocation(BlockDriverState *bs); /* qcow2-cluster.c functions */ int qcow2_grow_l1_table(BlockDriverState *bs, uint64_t min_size, bool exact_size); -int coroutine_fn qcow2_shrink_l1_table(BlockDriverState *bs, uint64_t max_size); + +int coroutine_fn GRAPH_RDLOCK +qcow2_shrink_l1_table(BlockDriverState *bs, uint64_t max_size); + int qcow2_write_l1_entry(BlockDriverState *bs, int l1_index); int qcow2_encrypt_sectors(BDRVQcow2State *s, int64_t sector_num, uint8_t *buf, int nb_sectors, bool enc, Error **errp); diff --git a/block/qed.h b/block/qed.h index XXXXXXX..XXXXXXX 100644 --- a/block/qed.h +++ b/block/qed.h @@ -XXX,XX +XXX,XX @@ void qed_commit_l2_cache_entry(L2TableCache *l2_cache, CachedL2Table *l2_table); * Table I/O functions */ int coroutine_fn qed_read_l1_table_sync(BDRVQEDState *s); -int coroutine_fn qed_write_l1_table(BDRVQEDState *s, unsigned int index, - unsigned int n); -int coroutine_fn qed_write_l1_table_sync(BDRVQEDState *s, unsigned int index, - unsigned int n); + +int coroutine_fn GRAPH_RDLOCK +qed_write_l1_table(BDRVQEDState *s, unsigned int index, unsigned int n); + +int coroutine_fn GRAPH_RDLOCK +qed_write_l1_table_sync(BDRVQEDState *s, unsigned int index, unsigned int n); + int coroutine_fn qed_read_l2_table_sync(BDRVQEDState *s, QEDRequest *request, uint64_t offset); int coroutine_fn qed_read_l2_table(BDRVQEDState *s, QEDRequest *request, uint64_t offset); -int coroutine_fn qed_write_l2_table(BDRVQEDState *s, QEDRequest *request, - unsigned int index, unsigned int n, - bool flush); -int coroutine_fn qed_write_l2_table_sync(BDRVQEDState *s, QEDRequest *request, - unsigned int index, unsigned int n, - bool flush); + +int coroutine_fn GRAPH_RDLOCK +qed_write_l2_table(BDRVQEDState *s, QEDRequest *request, unsigned int index, + unsigned int n, bool flush); + +int coroutine_fn GRAPH_RDLOCK +qed_write_l2_table_sync(BDRVQEDState *s, QEDRequest *request, + unsigned int index, unsigned int n, bool flush); /** * Cluster functions @@ -XXX,XX +XXX,XX @@ int coroutine_fn qed_find_cluster(BDRVQEDState *s, QEDRequest *request, /** * Consistency check */ -int coroutine_fn qed_check(BDRVQEDState *s, BdrvCheckResult *result, bool fix); +int coroutine_fn GRAPH_RDLOCK +qed_check(BDRVQEDState *s, BdrvCheckResult *result, bool fix); + QEDTable *qed_alloc_table(BDRVQEDState *s); 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 @@ int coroutine_fn GRAPH_RDLOCK bdrv_co_ioctl(BlockDriverState *bs, int req, void *buf); /* Ensure contents are flushed to disk. */ -int coroutine_fn bdrv_co_flush(BlockDriverState *bs); +int coroutine_fn GRAPH_RDLOCK bdrv_co_flush(BlockDriverState *bs); int coroutine_fn bdrv_co_pdiscard(BdrvChild *child, int64_t offset, int64_t bytes); 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 { BlockAIOCB *(*bdrv_aio_pwritev)(BlockDriverState *bs, int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags, BlockCompletionFunc *cb, void *opaque); - BlockAIOCB *(*bdrv_aio_flush)(BlockDriverState *bs, - BlockCompletionFunc *cb, void *opaque); + BlockAIOCB * GRAPH_RDLOCK_PTR (*bdrv_aio_flush)( + BlockDriverState *bs, BlockCompletionFunc *cb, void *opaque); BlockAIOCB *(*bdrv_aio_pdiscard)(BlockDriverState *bs, int64_t offset, int bytes, BlockCompletionFunc *cb, void *opaque); @@ -XXX,XX +XXX,XX @@ struct BlockDriver { * layers, if needed. This function is needed for deterministic * synchronization of the flush finishing callback. */ - int coroutine_fn (*bdrv_co_flush)(BlockDriverState *bs); + int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_flush)(BlockDriverState *bs); /* Delete a created file. */ int coroutine_fn (*bdrv_co_delete_file)(BlockDriverState *bs, @@ -XXX,XX +XXX,XX @@ struct BlockDriver { * Flushes all data that was already written to the OS all the way down to * the disk (for example file-posix.c calls fsync()). */ - int coroutine_fn (*bdrv_co_flush_to_disk)(BlockDriverState *bs); + int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_flush_to_disk)( + BlockDriverState *bs); /* * Flushes all internal caches to the OS. The data may still sit in a * writeback cache of the host OS, but it will survive a crash of the qemu * process. */ - int coroutine_fn (*bdrv_co_flush_to_os)(BlockDriverState *bs); + int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_flush_to_os)( + BlockDriverState *bs); /* * Truncate @bs to @offset bytes using the given @prealloc mode diff --git a/block/blkdebug.c b/block/blkdebug.c index XXXXXXX..XXXXXXX 100644 --- a/block/blkdebug.c +++ b/block/blkdebug.c @@ -XXX,XX +XXX,XX @@ blkdebug_co_pwritev(BlockDriverState *bs, int64_t offset, int64_t bytes, return bdrv_co_pwritev(bs->file, offset, bytes, qiov, flags); } -static int coroutine_fn blkdebug_co_flush(BlockDriverState *bs) +static int GRAPH_RDLOCK coroutine_fn blkdebug_co_flush(BlockDriverState *bs) { int err = rule_check(bs, 0, 0, BLKDEBUG_IO_TYPE_FLUSH); diff --git a/block/blklogwrites.c b/block/blklogwrites.c index XXXXXXX..XXXXXXX 100644 --- a/block/blklogwrites.c +++ b/block/blklogwrites.c @@ -XXX,XX +XXX,XX @@ typedef struct BlkLogWritesFileReq { uint64_t bytes; int file_flags; QEMUIOVector *qiov; - int (*func)(struct BlkLogWritesFileReq *r); + int GRAPH_RDLOCK_PTR (*func)(struct BlkLogWritesFileReq *r); int file_ret; } BlkLogWritesFileReq; @@ -XXX,XX +XXX,XX @@ typedef struct { int log_ret; } BlkLogWritesLogReq; -static void coroutine_fn blk_log_writes_co_do_log(BlkLogWritesLogReq *lr) +static void coroutine_fn GRAPH_RDLOCK +blk_log_writes_co_do_log(BlkLogWritesLogReq *lr) { BDRVBlkLogWritesState *s = lr->bs->opaque; uint64_t cur_log_offset = s->cur_log_sector << s->sectorbits; @@ -XXX,XX +XXX,XX @@ static void coroutine_fn blk_log_writes_co_do_log(BlkLogWritesLogReq *lr) } } -static void coroutine_fn blk_log_writes_co_do_file(BlkLogWritesFileReq *fr) +static void coroutine_fn GRAPH_RDLOCK +blk_log_writes_co_do_file(BlkLogWritesFileReq *fr) { fr->file_ret = fr->func(fr); } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK blk_log_writes_co_log(BlockDriverState *bs, uint64_t offset, uint64_t bytes, QEMUIOVector *qiov, int flags, - int (*file_func)(BlkLogWritesFileReq *r), + int /*GRAPH_RDLOCK*/ (*file_func)(BlkLogWritesFileReq *r), uint64_t entry_flags, bool is_zero_write) { QEMUIOVector log_qiov; @@ -XXX,XX +XXX,XX @@ blk_log_writes_co_do_file_pwrite_zeroes(BlkLogWritesFileReq *fr) fr->file_flags); } -static int coroutine_fn blk_log_writes_co_do_file_flush(BlkLogWritesFileReq *fr) +static int coroutine_fn GRAPH_RDLOCK +blk_log_writes_co_do_file_flush(BlkLogWritesFileReq *fr) { return bdrv_co_flush(fr->bs->file->bs); } @@ -XXX,XX +XXX,XX @@ static int coroutine_fn blk_log_writes_co_pwritev(BlockDriverState *bs, int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags) { + assume_graph_lock(); /* FIXME */ return blk_log_writes_co_log(bs, offset, bytes, qiov, flags, blk_log_writes_co_do_file_pwritev, 0, false); } @@ -XXX,XX +XXX,XX @@ static int coroutine_fn blk_log_writes_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset, int64_t bytes, BdrvRequestFlags flags) { + assume_graph_lock(); /* FIXME */ return blk_log_writes_co_log(bs, offset, bytes, NULL, flags, blk_log_writes_co_do_file_pwrite_zeroes, 0, true); } -static int coroutine_fn blk_log_writes_co_flush_to_disk(BlockDriverState *bs) +static int coroutine_fn GRAPH_RDLOCK +blk_log_writes_co_flush_to_disk(BlockDriverState *bs) { return blk_log_writes_co_log(bs, 0, 0, NULL, 0, blk_log_writes_co_do_file_flush, @@ -XXX,XX +XXX,XX @@ static int coroutine_fn blk_log_writes_co_flush_to_disk(BlockDriverState *bs) static int coroutine_fn blk_log_writes_co_pdiscard(BlockDriverState *bs, int64_t offset, int64_t bytes) { + assume_graph_lock(); /* FIXME */ return blk_log_writes_co_log(bs, offset, bytes, NULL, 0, blk_log_writes_co_do_file_pdiscard, LOG_DISCARD_FLAG, false); diff --git a/block/blkreplay.c b/block/blkreplay.c index XXXXXXX..XXXXXXX 100644 --- a/block/blkreplay.c +++ b/block/blkreplay.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn blkreplay_co_pdiscard(BlockDriverState *bs, return ret; } -static int coroutine_fn blkreplay_co_flush(BlockDriverState *bs) +static int coroutine_fn GRAPH_RDLOCK blkreplay_co_flush(BlockDriverState *bs) { uint64_t reqid = blkreplay_next_id(); int ret = bdrv_co_flush(bs->file->bs); 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 @@ blkverify_co_pwritev(BlockDriverState *bs, int64_t offset, int64_t bytes, return blkverify_co_prwv(bs, &r, offset, bytes, qiov, qiov, flags, true); } -static int coroutine_fn blkverify_co_flush(BlockDriverState *bs) +static int coroutine_fn GRAPH_RDLOCK blkverify_co_flush(BlockDriverState *bs) { BDRVBlkverifyState *s = bs->opaque; 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 @@ int coroutine_fn blk_co_pdiscard(BlockBackend *blk, int64_t offset, /* To be called between exactly one pair of blk_inc/dec_in_flight() */ static int coroutine_fn blk_co_do_flush(BlockBackend *blk) { - blk_wait_while_drained(blk); IO_CODE(); + blk_wait_while_drained(blk); + GRAPH_RDLOCK_GUARD(); if (!blk_is_available(blk)) { return -ENOMEDIUM; diff --git a/block/copy-before-write.c b/block/copy-before-write.c index XXXXXXX..XXXXXXX 100644 --- a/block/copy-before-write.c +++ b/block/copy-before-write.c @@ -XXX,XX +XXX,XX @@ static coroutine_fn int cbw_co_pwritev(BlockDriverState *bs, return bdrv_co_pwritev(bs->file, offset, bytes, qiov, flags); } -static int coroutine_fn cbw_co_flush(BlockDriverState *bs) +static int coroutine_fn GRAPH_RDLOCK cbw_co_flush(BlockDriverState *bs) { if (!bs->file) { return 0; diff --git a/block/file-posix.c b/block/file-posix.c index XXXXXXX..XXXXXXX 100644 --- a/block/file-posix.c +++ b/block/file-posix.c @@ -XXX,XX +XXX,XX @@ static void coroutine_fn check_cache_dropped(BlockDriverState *bs, Error **errp) } #endif /* __linux__ */ -static void coroutine_fn raw_co_invalidate_cache(BlockDriverState *bs, - Error **errp) +static void coroutine_fn GRAPH_RDLOCK +raw_co_invalidate_cache(BlockDriverState *bs, Error **errp) { BDRVRawState *s = bs->opaque; int ret; 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 @@ int coroutine_fn bdrv_co_pwrite_sync(BdrvChild *child, int64_t offset, int ret; IO_CODE(); + assume_graph_lock(); /* FIXME */ + ret = bdrv_co_pwrite(child, offset, bytes, buf, flags); if (ret < 0) { return ret; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_driver_pwritev(BlockDriverState *bs, QEMUIOVector local_qiov; int ret; + assume_graph_lock(); /* FIXME */ + bdrv_check_qiov_request(offset, bytes, qiov, qiov_offset, &error_abort); if (!drv) { @@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_pwrite_zeroes(BlockDriverState *bs, int head = 0; int tail = 0; + assume_graph_lock(); /* FIXME */ + int64_t max_write_zeroes = MIN_NON_ZERO(bs->bl.max_pwrite_zeroes, INT64_MAX); int alignment = MAX(bs->bl.pwrite_zeroes_alignment, @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_flush(BlockDriverState *bs) int ret = 0; IO_CODE(); + assert_bdrv_graph_readable(); bdrv_inc_in_flight(bs); if (!bdrv_co_is_inserted(bs) || bdrv_is_read_only(bs) || 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 coroutine_fn bdrv_mirror_top_pwritev(BlockDriverState *bs, return ret; } -static int coroutine_fn bdrv_mirror_top_flush(BlockDriverState *bs) +static int coroutine_fn GRAPH_RDLOCK bdrv_mirror_top_flush(BlockDriverState *bs) { if (bs->backing == NULL) { /* we can be here after failed bdrv_append in mirror_start_job */ diff --git a/block/preallocate.c b/block/preallocate.c index XXXXXXX..XXXXXXX 100644 --- a/block/preallocate.c +++ b/block/preallocate.c @@ -XXX,XX +XXX,XX @@ preallocate_co_truncate(BlockDriverState *bs, int64_t offset, return 0; } -static int coroutine_fn preallocate_co_flush(BlockDriverState *bs) +static int coroutine_fn GRAPH_RDLOCK preallocate_co_flush(BlockDriverState *bs) { return bdrv_co_flush(bs->file->bs); } diff --git a/block/qed-check.c b/block/qed-check.c index XXXXXXX..XXXXXXX 100644 --- a/block/qed-check.c +++ b/block/qed-check.c @@ -XXX,XX +XXX,XX @@ static unsigned int qed_check_l2_table(QEDCheck *check, QEDTable *table) /** * Descend tables and check each cluster is referenced once only */ -static int coroutine_fn qed_check_l1_table(QEDCheck *check, QEDTable *table) +static int coroutine_fn GRAPH_RDLOCK +qed_check_l1_table(QEDCheck *check, QEDTable *table) { BDRVQEDState *s = check->s; unsigned int i, num_invalid_l1 = 0; diff --git a/block/qed-table.c b/block/qed-table.c index XXXXXXX..XXXXXXX 100644 --- a/block/qed-table.c +++ b/block/qed-table.c @@ -XXX,XX +XXX,XX @@ out: * * Called with table_lock held. */ -static int coroutine_fn qed_write_table(BDRVQEDState *s, uint64_t offset, - QEDTable *table, unsigned int index, - unsigned int n, bool flush) +static int coroutine_fn GRAPH_RDLOCK +qed_write_table(BDRVQEDState *s, uint64_t offset, QEDTable *table, + unsigned int index, unsigned int n, bool flush) { unsigned int sector_mask = BDRV_SECTOR_SIZE / sizeof(uint64_t) - 1; unsigned int start, end, i; 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 void bdrv_qed_init_state(BlockDriverState *bs) } /* Called with table_lock held. */ -static int coroutine_fn bdrv_qed_do_open(BlockDriverState *bs, QDict *options, - int flags, Error **errp) +static int coroutine_fn GRAPH_RDLOCK +bdrv_qed_do_open(BlockDriverState *bs, QDict *options, int flags, Error **errp) { BDRVQEDState *s = bs->opaque; QEDHeader le_header; @@ -XXX,XX +XXX,XX @@ typedef struct QEDOpenCo { int ret; } QEDOpenCo; -static void coroutine_fn bdrv_qed_open_entry(void *opaque) +static void coroutine_fn GRAPH_RDLOCK bdrv_qed_open_entry(void *opaque) { QEDOpenCo *qoc = opaque; BDRVQEDState *s = qoc->bs->opaque; @@ -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; @@ -XXX,XX +XXX,XX @@ static void coroutine_fn qed_aio_complete(QEDAIOCB *acb) * * Called with table_lock held. */ -static int coroutine_fn qed_aio_write_l1_update(QEDAIOCB *acb) +static int coroutine_fn GRAPH_RDLOCK qed_aio_write_l1_update(QEDAIOCB *acb) { BDRVQEDState *s = acb_to_s(acb); CachedL2Table *l2_table = acb->request.l2_table; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qed_aio_write_l1_update(QEDAIOCB *acb) * * Called with table_lock held. */ -static int coroutine_fn qed_aio_write_l2_update(QEDAIOCB *acb, uint64_t offset) +static int coroutine_fn GRAPH_RDLOCK +qed_aio_write_l2_update(QEDAIOCB *acb, uint64_t offset) { BDRVQEDState *s = acb_to_s(acb); bool need_alloc = acb->find_cluster_ret == QED_CLUSTER_L1; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qed_aio_write_main(QEDAIOCB *acb) * * Called with table_lock held. */ -static int coroutine_fn qed_aio_write_cow(QEDAIOCB *acb) +static int coroutine_fn GRAPH_RDLOCK qed_aio_write_cow(QEDAIOCB *acb) { BDRVQEDState *s = acb_to_s(acb); uint64_t start, len, offset; @@ -XXX,XX +XXX,XX @@ static bool qed_should_set_need_check(BDRVQEDState *s) * * Called with table_lock held. */ -static int coroutine_fn qed_aio_write_alloc(QEDAIOCB *acb, size_t len) +static int coroutine_fn GRAPH_RDLOCK +qed_aio_write_alloc(QEDAIOCB *acb, size_t len) { BDRVQEDState *s = acb_to_s(acb); int ret; @@ -XXX,XX +XXX,XX @@ out: * * Called with table_lock held. */ -static int coroutine_fn qed_aio_write_data(void *opaque, int ret, - uint64_t offset, size_t len) +static int coroutine_fn GRAPH_RDLOCK +qed_aio_write_data(void *opaque, int ret, uint64_t offset, size_t len) { QEDAIOCB *acb = opaque; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qed_aio_read_data(void *opaque, int ret, /** * Begin next I/O or complete the request */ -static int coroutine_fn qed_aio_next_io(QEDAIOCB *acb) +static int coroutine_fn GRAPH_RDLOCK qed_aio_next_io(QEDAIOCB *acb) { BDRVQEDState *s = acb_to_s(acb); uint64_t offset; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qed_aio_next_io(QEDAIOCB *acb) return ret; } -static int coroutine_fn qed_co_request(BlockDriverState *bs, int64_t sector_num, - QEMUIOVector *qiov, int nb_sectors, - int flags) +static int coroutine_fn GRAPH_RDLOCK +qed_co_request(BlockDriverState *bs, int64_t sector_num, QEMUIOVector *qiov, + int nb_sectors, int flags) { QEDAIOCB acb = { .bs = bs, @@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_qed_co_readv(BlockDriverState *bs, int64_t sector_num, int nb_sectors, QEMUIOVector *qiov) { + assume_graph_lock(); /* FIXME */ return qed_co_request(bs, sector_num, qiov, nb_sectors, 0); } @@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_qed_co_writev(BlockDriverState *bs, int64_t sector_num, int nb_sectors, QEMUIOVector *qiov, int flags) { + assume_graph_lock(); /* FIXME */ return qed_co_request(bs, sector_num, qiov, nb_sectors, QED_AIOCB_WRITE); } @@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_qed_co_pwrite_zeroes(BlockDriverState *bs, { BDRVQEDState *s = bs->opaque; + assume_graph_lock(); /* FIXME */ + /* * Zero writes start without an I/O buffer. If a buffer becomes necessary * then it will be allocated during request processing. @@ -XXX,XX +XXX,XX @@ static int bdrv_qed_change_backing_file(BlockDriverState *bs, return ret; } -static void coroutine_fn bdrv_qed_co_invalidate_cache(BlockDriverState *bs, - Error **errp) +static void coroutine_fn GRAPH_RDLOCK +bdrv_qed_co_invalidate_cache(BlockDriverState *bs, Error **errp) { BDRVQEDState *s = bs->opaque; int ret; @@ -XXX,XX +XXX,XX @@ static void coroutine_fn bdrv_qed_co_invalidate_cache(BlockDriverState *bs, } } -static int coroutine_fn bdrv_qed_co_check(BlockDriverState *bs, - BdrvCheckResult *result, - BdrvCheckMode fix) +static int coroutine_fn GRAPH_RDLOCK +bdrv_qed_co_check(BlockDriverState *bs, BdrvCheckResult *result, + BdrvCheckMode fix) { BDRVQEDState *s = bs->opaque; int ret; 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 int64_t coroutine_fn quorum_co_getlength(BlockDriverState *bs) return result; } -static coroutine_fn int quorum_co_flush(BlockDriverState *bs) +static coroutine_fn GRAPH_RDLOCK int quorum_co_flush(BlockDriverState *bs) { BDRVQuorumState *s = bs->opaque; QuorumVoteVersion *winner = NULL; diff --git a/block/throttle.c b/block/throttle.c index XXXXXXX..XXXXXXX 100644 --- a/block/throttle.c +++ b/block/throttle.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn throttle_co_pwritev_compressed(BlockDriverState *bs, BDRV_REQ_WRITE_COMPRESSED); } -static int coroutine_fn throttle_co_flush(BlockDriverState *bs) +static int coroutine_fn GRAPH_RDLOCK throttle_co_flush(BlockDriverState *bs) { return bdrv_co_flush(bs->file->bs); } 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: return ret; } -static int coroutine_fn vmdk_L2update(VmdkExtent *extent, VmdkMetaData *m_data, - uint32_t offset) +static int coroutine_fn GRAPH_RDLOCK +vmdk_L2update(VmdkExtent *extent, VmdkMetaData *m_data, uint32_t offset) { offset = cpu_to_le32(offset); /* update L2 table */ @@ -XXX,XX +XXX,XX @@ static int coroutine_fn vmdk_pwritev(BlockDriverState *bs, uint64_t offset, uint64_t bytes_done = 0; VmdkMetaData m_data; + assume_graph_lock(); /* FIXME */ + if (DIV_ROUND_UP(offset, BDRV_SECTOR_SIZE) > bs->total_sectors) { error_report("Wrong offset: offset=0x%" PRIx64 " total_sectors=0x%" PRIx64, -- 2.39.2 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> This adds GRAPH_RDLOCK annotations to declare that callers of bdrv_co_pdiscard() need to hold a reader lock for the graph. For some places, we know that they will hold the lock, but we don't have the GRAPH_RDLOCK annotations yet. In this case, add assume_graph_lock() with a FIXME comment. These places will be removed once everything is properly annotated. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230203152202.49054-9-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-io.h | 5 +++-- include/block/block_int-common.h | 15 +++++++++------ include/block/block_int-io.h | 2 +- block/blkdebug.c | 4 ++-- block/blklogwrites.c | 5 ++--- block/blkreplay.c | 4 ++-- block/block-backend.c | 1 + block/copy-before-write.c | 8 ++++---- block/copy-on-read.c | 4 ++-- block/filter-compress.c | 4 ++-- block/io.c | 2 ++ block/mirror.c | 14 +++++++++----- block/preallocate.c | 4 ++-- block/raw-format.c | 4 ++-- block/snapshot-access.c | 4 ++-- block/throttle.c | 4 ++-- 16 files changed, 47 insertions(+), 37 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 @@ bdrv_co_ioctl(BlockDriverState *bs, int req, void *buf); /* Ensure contents are flushed to disk. */ int coroutine_fn GRAPH_RDLOCK bdrv_co_flush(BlockDriverState *bs); -int coroutine_fn bdrv_co_pdiscard(BdrvChild *child, int64_t offset, - int64_t bytes); +int coroutine_fn GRAPH_RDLOCK bdrv_co_pdiscard(BdrvChild *child, int64_t offset, + int64_t bytes); + bool bdrv_can_write_zeroes_with_unmap(BlockDriverState *bs); int bdrv_block_status(BlockDriverState *bs, int64_t offset, int64_t bytes, int64_t *pnum, int64_t *map, 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 { BdrvRequestFlags flags, BlockCompletionFunc *cb, void *opaque); BlockAIOCB * GRAPH_RDLOCK_PTR (*bdrv_aio_flush)( BlockDriverState *bs, BlockCompletionFunc *cb, void *opaque); - BlockAIOCB *(*bdrv_aio_pdiscard)(BlockDriverState *bs, - int64_t offset, int bytes, + + BlockAIOCB * GRAPH_RDLOCK_PTR (*bdrv_aio_pdiscard)( + BlockDriverState *bs, int64_t offset, int bytes, BlockCompletionFunc *cb, void *opaque); int coroutine_fn (*bdrv_co_readv)(BlockDriverState *bs, @@ -XXX,XX +XXX,XX @@ struct BlockDriver { */ int coroutine_fn (*bdrv_co_pwrite_zeroes)(BlockDriverState *bs, int64_t offset, int64_t bytes, BdrvRequestFlags flags); - int coroutine_fn (*bdrv_co_pdiscard)(BlockDriverState *bs, - int64_t offset, int64_t bytes); + + int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_pdiscard)( + BlockDriverState *bs, int64_t offset, int64_t bytes); /* * Map [offset, offset + nbytes) range onto a child of @bs to copy from, @@ -XXX,XX +XXX,XX @@ struct BlockDriver { int coroutine_fn (*bdrv_co_snapshot_block_status)(BlockDriverState *bs, bool want_zero, int64_t offset, int64_t bytes, int64_t *pnum, int64_t *map, BlockDriverState **file); - int coroutine_fn (*bdrv_co_pdiscard_snapshot)(BlockDriverState *bs, - int64_t offset, int64_t bytes); + + int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_pdiscard_snapshot)( + BlockDriverState *bs, int64_t offset, int64_t bytes); /* * Invalidate any cached meta-data. diff --git a/include/block/block_int-io.h b/include/block/block_int-io.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block_int-io.h +++ b/include/block/block_int-io.h @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_preadv_snapshot(BdrvChild *child, int coroutine_fn bdrv_co_snapshot_block_status(BlockDriverState *bs, bool want_zero, int64_t offset, int64_t bytes, int64_t *pnum, int64_t *map, BlockDriverState **file); -int coroutine_fn bdrv_co_pdiscard_snapshot(BlockDriverState *bs, +int coroutine_fn GRAPH_RDLOCK bdrv_co_pdiscard_snapshot(BlockDriverState *bs, int64_t offset, int64_t bytes); diff --git a/block/blkdebug.c b/block/blkdebug.c index XXXXXXX..XXXXXXX 100644 --- a/block/blkdebug.c +++ b/block/blkdebug.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn blkdebug_co_pwrite_zeroes(BlockDriverState *bs, return bdrv_co_pwrite_zeroes(bs->file, offset, bytes, flags); } -static int coroutine_fn blkdebug_co_pdiscard(BlockDriverState *bs, - int64_t offset, int64_t bytes) +static int coroutine_fn GRAPH_RDLOCK +blkdebug_co_pdiscard(BlockDriverState *bs, int64_t offset, int64_t bytes) { uint32_t align = bs->bl.pdiscard_alignment; int err; diff --git a/block/blklogwrites.c b/block/blklogwrites.c index XXXXXXX..XXXXXXX 100644 --- a/block/blklogwrites.c +++ b/block/blklogwrites.c @@ -XXX,XX +XXX,XX @@ blk_log_writes_co_do_file_flush(BlkLogWritesFileReq *fr) return bdrv_co_flush(fr->bs->file->bs); } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK blk_log_writes_co_do_file_pdiscard(BlkLogWritesFileReq *fr) { return bdrv_co_pdiscard(fr->bs->file, fr->offset, fr->bytes); @@ -XXX,XX +XXX,XX @@ blk_log_writes_co_flush_to_disk(BlockDriverState *bs) LOG_FLUSH_FLAG, false); } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK blk_log_writes_co_pdiscard(BlockDriverState *bs, int64_t offset, int64_t bytes) { - assume_graph_lock(); /* FIXME */ return blk_log_writes_co_log(bs, offset, bytes, NULL, 0, blk_log_writes_co_do_file_pdiscard, LOG_DISCARD_FLAG, false); diff --git a/block/blkreplay.c b/block/blkreplay.c index XXXXXXX..XXXXXXX 100644 --- a/block/blkreplay.c +++ b/block/blkreplay.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn blkreplay_co_pwrite_zeroes(BlockDriverState *bs, return ret; } -static int coroutine_fn blkreplay_co_pdiscard(BlockDriverState *bs, - int64_t offset, int64_t bytes) +static int coroutine_fn GRAPH_RDLOCK +blkreplay_co_pdiscard(BlockDriverState *bs, int64_t offset, int64_t bytes) { uint64_t reqid = blkreplay_next_id(); int ret = bdrv_co_pdiscard(bs->file, offset, bytes); 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 @@ blk_co_do_pdiscard(BlockBackend *blk, int64_t offset, int64_t bytes) IO_CODE(); blk_wait_while_drained(blk); + GRAPH_RDLOCK_GUARD(); ret = blk_check_byte_request(blk, offset, bytes); if (ret < 0) { diff --git a/block/copy-before-write.c b/block/copy-before-write.c index XXXXXXX..XXXXXXX 100644 --- a/block/copy-before-write.c +++ b/block/copy-before-write.c @@ -XXX,XX +XXX,XX @@ static coroutine_fn int cbw_do_copy_before_write(BlockDriverState *bs, return 0; } -static int coroutine_fn cbw_co_pdiscard(BlockDriverState *bs, - int64_t offset, int64_t bytes) +static int coroutine_fn GRAPH_RDLOCK +cbw_co_pdiscard(BlockDriverState *bs, int64_t offset, int64_t bytes) { int ret = cbw_do_copy_before_write(bs, offset, bytes, 0); if (ret < 0) { @@ -XXX,XX +XXX,XX @@ cbw_co_snapshot_block_status(BlockDriverState *bs, return ret; } -static int coroutine_fn cbw_co_pdiscard_snapshot(BlockDriverState *bs, - int64_t offset, int64_t bytes) +static int coroutine_fn GRAPH_RDLOCK +cbw_co_pdiscard_snapshot(BlockDriverState *bs, int64_t offset, int64_t bytes) { BDRVCopyBeforeWriteState *s = bs->opaque; diff --git a/block/copy-on-read.c b/block/copy-on-read.c index XXXXXXX..XXXXXXX 100644 --- a/block/copy-on-read.c +++ b/block/copy-on-read.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn cor_co_pwrite_zeroes(BlockDriverState *bs, } -static int coroutine_fn cor_co_pdiscard(BlockDriverState *bs, - int64_t offset, int64_t bytes) +static int coroutine_fn GRAPH_RDLOCK +cor_co_pdiscard(BlockDriverState *bs, int64_t offset, int64_t bytes) { return bdrv_co_pdiscard(bs->file, offset, bytes); } diff --git a/block/filter-compress.c b/block/filter-compress.c index XXXXXXX..XXXXXXX 100644 --- a/block/filter-compress.c +++ b/block/filter-compress.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn compress_co_pwrite_zeroes(BlockDriverState *bs, } -static int coroutine_fn compress_co_pdiscard(BlockDriverState *bs, - int64_t offset, int64_t bytes) +static int coroutine_fn GRAPH_RDLOCK +compress_co_pdiscard(BlockDriverState *bs, int64_t offset, int64_t bytes) { return bdrv_co_pdiscard(bs->file, offset, bytes); } 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 @@ int coroutine_fn bdrv_co_pdiscard(BdrvChild *child, int64_t offset, int head, tail, align; BlockDriverState *bs = child->bs; IO_CODE(); + assert_bdrv_graph_readable(); if (!bs || !bs->drv || !bdrv_co_is_inserted(bs)) { return -ENOMEDIUM; @@ -XXX,XX +XXX,XX @@ bdrv_co_pdiscard_snapshot(BlockDriverState *bs, int64_t offset, int64_t bytes) BlockDriver *drv = bs->drv; int ret; IO_CODE(); + assert_bdrv_graph_readable(); if (!drv) { return -ENOMEDIUM; 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 coroutine_fn bdrv_mirror_top_preadv(BlockDriverState *bs, return bdrv_co_preadv(bs->backing, offset, bytes, qiov, flags); } -static int coroutine_fn bdrv_mirror_top_do_write(BlockDriverState *bs, - MirrorMethod method, uint64_t offset, uint64_t bytes, QEMUIOVector *qiov, - int flags) +static int coroutine_fn GRAPH_RDLOCK +bdrv_mirror_top_do_write(BlockDriverState *bs, MirrorMethod method, + uint64_t offset, uint64_t bytes, QEMUIOVector *qiov, + int flags) { MirrorOp *op = NULL; MirrorBDSOpaque *s = bs->opaque; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_mirror_top_pwritev(BlockDriverState *bs, int ret = 0; bool copy_to_target = false; + assume_graph_lock(); /* FIXME */ + if (s->job) { copy_to_target = s->job->ret >= 0 && !job_is_cancelled(&s->job->common.job) && @@ -XXX,XX +XXX,XX @@ static int coroutine_fn GRAPH_RDLOCK bdrv_mirror_top_flush(BlockDriverState *bs) static int coroutine_fn bdrv_mirror_top_pwrite_zeroes(BlockDriverState *bs, int64_t offset, int64_t bytes, BdrvRequestFlags flags) { + assume_graph_lock(); /* FIXME */ return bdrv_mirror_top_do_write(bs, MIRROR_METHOD_ZERO, offset, bytes, NULL, flags); } -static int coroutine_fn bdrv_mirror_top_pdiscard(BlockDriverState *bs, - int64_t offset, int64_t bytes) +static int coroutine_fn GRAPH_RDLOCK +bdrv_mirror_top_pdiscard(BlockDriverState *bs, int64_t offset, int64_t bytes) { return bdrv_mirror_top_do_write(bs, MIRROR_METHOD_DISCARD, offset, bytes, NULL, 0); diff --git a/block/preallocate.c b/block/preallocate.c index XXXXXXX..XXXXXXX 100644 --- a/block/preallocate.c +++ b/block/preallocate.c @@ -XXX,XX +XXX,XX @@ static coroutine_fn int preallocate_co_preadv_part( flags); } -static int coroutine_fn preallocate_co_pdiscard(BlockDriverState *bs, - int64_t offset, int64_t bytes) +static int coroutine_fn GRAPH_RDLOCK +preallocate_co_pdiscard(BlockDriverState *bs, int64_t offset, int64_t bytes) { return bdrv_co_pdiscard(bs->file, offset, bytes); } 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 int coroutine_fn raw_co_pwrite_zeroes(BlockDriverState *bs, return bdrv_co_pwrite_zeroes(bs->file, offset, bytes, flags); } -static int coroutine_fn raw_co_pdiscard(BlockDriverState *bs, - int64_t offset, int64_t bytes) +static int coroutine_fn GRAPH_RDLOCK +raw_co_pdiscard(BlockDriverState *bs, int64_t offset, int64_t bytes) { int ret; diff --git a/block/snapshot-access.c b/block/snapshot-access.c index XXXXXXX..XXXXXXX 100644 --- a/block/snapshot-access.c +++ b/block/snapshot-access.c @@ -XXX,XX +XXX,XX @@ snapshot_access_co_block_status(BlockDriverState *bs, bytes, pnum, map, file); } -static int coroutine_fn snapshot_access_co_pdiscard(BlockDriverState *bs, - int64_t offset, int64_t bytes) +static int coroutine_fn GRAPH_RDLOCK +snapshot_access_co_pdiscard(BlockDriverState *bs, int64_t offset, int64_t bytes) { return bdrv_co_pdiscard_snapshot(bs->file->bs, offset, bytes); } diff --git a/block/throttle.c b/block/throttle.c index XXXXXXX..XXXXXXX 100644 --- a/block/throttle.c +++ b/block/throttle.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn throttle_co_pwrite_zeroes(BlockDriverState *bs, return bdrv_co_pwrite_zeroes(bs->file, offset, bytes, flags); } -static int coroutine_fn throttle_co_pdiscard(BlockDriverState *bs, - int64_t offset, int64_t bytes) +static int coroutine_fn GRAPH_RDLOCK +throttle_co_pdiscard(BlockDriverState *bs, int64_t offset, int64_t bytes) { ThrottleGroupMember *tgm = bs->opaque; throttle_group_co_io_limits_intercept(tgm, bytes, true); -- 2.39.2 This adds GRAPH_RDLOCK annotations to declare that callers of bdrv_co_pwrite_zeroes() need to hold a reader lock for the graph. For some places, we know that they will hold the lock, but we don't have the GRAPH_RDLOCK annotations yet. In this case, add assume_graph_lock() with a FIXME comment. These places will be removed once everything is properly annotated. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230203152202.49054-10-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/qcow2.h | 6 ++++-- include/block/block-io.h | 9 +++++---- include/block/block_int-common.h | 5 +++-- block/blkdebug.c | 6 +++--- block/blklogwrites.c | 5 ++--- block/blkreplay.c | 5 +++-- block/block-copy.c | 13 +++++++------ block/copy-before-write.c | 5 +++-- block/copy-on-read.c | 6 +++--- block/filter-compress.c | 6 +++--- block/io.c | 12 +++++++++--- block/mirror.c | 6 +++--- block/preallocate.c | 11 +++++++---- block/qcow2.c | 30 ++++++++++++++++++------------ block/qed.c | 9 +++------ block/quorum.c | 15 ++++++++++----- block/raw-format.c | 6 +++--- block/throttle.c | 6 +++--- 18 files changed, 92 insertions(+), 69 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 @@ void qcow2_alloc_cluster_abort(BlockDriverState *bs, QCowL2Meta *m); int qcow2_cluster_discard(BlockDriverState *bs, uint64_t offset, uint64_t bytes, enum qcow2_discard_type type, bool full_discard); -int coroutine_fn qcow2_subcluster_zeroize(BlockDriverState *bs, uint64_t offset, - uint64_t bytes, int flags); + +int coroutine_fn GRAPH_RDLOCK +qcow2_subcluster_zeroize(BlockDriverState *bs, uint64_t offset, uint64_t bytes, + int flags); int qcow2_expand_zero_clusters(BlockDriverState *bs, BlockDriverAmendStatusCB *status_cb, 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 @@ int coroutine_fn bdrv_co_pwrite_sync(BdrvChild *child, int64_t offset, * function is not suitable for zeroing the entire image in a single request * because it may allocate memory for the entire region. */ -int coroutine_fn bdrv_co_pwrite_zeroes(BdrvChild *child, int64_t offset, - int64_t bytes, BdrvRequestFlags flags); +int coroutine_fn GRAPH_RDLOCK +bdrv_co_pwrite_zeroes(BdrvChild *child, int64_t offset, int64_t bytes, + BdrvRequestFlags flags); int coroutine_fn GRAPH_RDLOCK bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact, @@ -XXX,XX +XXX,XX @@ int bdrv_is_allocated_above(BlockDriverState *top, BlockDriverState *base, bool include_base, int64_t offset, int64_t bytes, int64_t *pnum); -int coroutine_fn bdrv_co_is_zero_fast(BlockDriverState *bs, int64_t offset, - int64_t bytes); +int coroutine_fn GRAPH_RDLOCK +bdrv_co_is_zero_fast(BlockDriverState *bs, int64_t offset, int64_t bytes); int bdrv_apply_auto_read_only(BlockDriverState *bs, const char *errmsg, 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 { * function pointer may be NULL or return -ENOSUP and .bdrv_co_writev() * will be called instead. */ - int coroutine_fn (*bdrv_co_pwrite_zeroes)(BlockDriverState *bs, - int64_t offset, int64_t bytes, BdrvRequestFlags flags); + int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_pwrite_zeroes)( + BlockDriverState *bs, int64_t offset, int64_t bytes, + BdrvRequestFlags flags); int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_pdiscard)( BlockDriverState *bs, int64_t offset, int64_t bytes); diff --git a/block/blkdebug.c b/block/blkdebug.c index XXXXXXX..XXXXXXX 100644 --- a/block/blkdebug.c +++ b/block/blkdebug.c @@ -XXX,XX +XXX,XX @@ static int GRAPH_RDLOCK coroutine_fn blkdebug_co_flush(BlockDriverState *bs) return bdrv_co_flush(bs->file->bs); } -static int coroutine_fn blkdebug_co_pwrite_zeroes(BlockDriverState *bs, - int64_t offset, int64_t bytes, - BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +blkdebug_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset, int64_t bytes, + BdrvRequestFlags flags) { uint32_t align = MAX(bs->bl.request_alignment, bs->bl.pwrite_zeroes_alignment); diff --git a/block/blklogwrites.c b/block/blklogwrites.c index XXXXXXX..XXXXXXX 100644 --- a/block/blklogwrites.c +++ b/block/blklogwrites.c @@ -XXX,XX +XXX,XX @@ blk_log_writes_co_do_file_pwritev(BlkLogWritesFileReq *fr) fr->qiov, fr->file_flags); } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK blk_log_writes_co_do_file_pwrite_zeroes(BlkLogWritesFileReq *fr) { return bdrv_co_pwrite_zeroes(fr->bs->file, fr->offset, fr->bytes, @@ -XXX,XX +XXX,XX @@ blk_log_writes_co_pwritev(BlockDriverState *bs, int64_t offset, int64_t bytes, blk_log_writes_co_do_file_pwritev, 0, false); } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK blk_log_writes_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset, int64_t bytes, BdrvRequestFlags flags) { - assume_graph_lock(); /* FIXME */ return blk_log_writes_co_log(bs, offset, bytes, NULL, flags, blk_log_writes_co_do_file_pwrite_zeroes, 0, true); diff --git a/block/blkreplay.c b/block/blkreplay.c index XXXXXXX..XXXXXXX 100644 --- a/block/blkreplay.c +++ b/block/blkreplay.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn blkreplay_co_pwritev(BlockDriverState *bs, return ret; } -static int coroutine_fn blkreplay_co_pwrite_zeroes(BlockDriverState *bs, - int64_t offset, int64_t bytes, BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +blkreplay_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset, int64_t bytes, + BdrvRequestFlags flags) { uint64_t reqid = blkreplay_next_id(); int ret = bdrv_co_pwrite_zeroes(bs->file, offset, bytes, flags); diff --git a/block/block-copy.c b/block/block-copy.c index XXXXXXX..XXXXXXX 100644 --- a/block/block-copy.c +++ b/block/block-copy.c @@ -XXX,XX +XXX,XX @@ static coroutine_fn int block_copy_task_run(AioTaskPool *pool, * value of @method should be used for subsequent tasks. * Returns 0 on success. */ -static int coroutine_fn block_copy_do_copy(BlockCopyState *s, - int64_t offset, int64_t bytes, - BlockCopyMethod *method, - bool *error_is_read) +static int coroutine_fn GRAPH_RDLOCK +block_copy_do_copy(BlockCopyState *s, int64_t offset, int64_t bytes, + BlockCopyMethod *method, bool *error_is_read) { int ret; int64_t nbytes = MIN(offset + bytes, s->len) - offset; @@ -XXX,XX +XXX,XX @@ static coroutine_fn int block_copy_task_entry(AioTask *task) BlockCopyMethod method = t->method; int ret; - ret = block_copy_do_copy(s, t->req.offset, t->req.bytes, &method, - &error_is_read); + WITH_GRAPH_RDLOCK_GUARD() { + ret = block_copy_do_copy(s, t->req.offset, t->req.bytes, &method, + &error_is_read); + } WITH_QEMU_LOCK_GUARD(&s->lock) { if (s->method == t->method) { diff --git a/block/copy-before-write.c b/block/copy-before-write.c index XXXXXXX..XXXXXXX 100644 --- a/block/copy-before-write.c +++ b/block/copy-before-write.c @@ -XXX,XX +XXX,XX @@ cbw_co_pdiscard(BlockDriverState *bs, int64_t offset, int64_t bytes) return bdrv_co_pdiscard(bs->file, offset, bytes); } -static int coroutine_fn cbw_co_pwrite_zeroes(BlockDriverState *bs, - int64_t offset, int64_t bytes, BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +cbw_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset, int64_t bytes, + BdrvRequestFlags flags) { int ret = cbw_do_copy_before_write(bs, offset, bytes, flags); if (ret < 0) { diff --git a/block/copy-on-read.c b/block/copy-on-read.c index XXXXXXX..XXXXXXX 100644 --- a/block/copy-on-read.c +++ b/block/copy-on-read.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn cor_co_pwritev_part(BlockDriverState *bs, } -static int coroutine_fn cor_co_pwrite_zeroes(BlockDriverState *bs, - int64_t offset, int64_t bytes, - BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +cor_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset, int64_t bytes, + BdrvRequestFlags flags) { return bdrv_co_pwrite_zeroes(bs->file, offset, bytes, flags); } diff --git a/block/filter-compress.c b/block/filter-compress.c index XXXXXXX..XXXXXXX 100644 --- a/block/filter-compress.c +++ b/block/filter-compress.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn compress_co_pwritev_part(BlockDriverState *bs, } -static int coroutine_fn compress_co_pwrite_zeroes(BlockDriverState *bs, - int64_t offset, int64_t bytes, - BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +compress_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset, int64_t bytes, + BdrvRequestFlags flags) { return bdrv_co_pwrite_zeroes(bs->file, offset, bytes, flags); } 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 @@ static int coroutine_fn bdrv_co_do_copy_on_readv(BdrvChild *child, int64_t progress = 0; bool skip_write; + assume_graph_lock(); /* FIXME */ + bdrv_check_qiov_request(offset, bytes, qiov, qiov_offset, &error_abort); if (!drv) { @@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_pwrite_zeroes(BlockDriverState *bs, bs->bl.request_alignment); int max_transfer = MIN_NON_ZERO(bs->bl.max_transfer, MAX_BOUNCE_BUFFER); + assert_bdrv_graph_readable(); bdrv_check_request(offset, bytes, &error_abort); if (!drv) { @@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_pwritev(BdrvChild *child, int64_t bytes_remaining = bytes; int max_transfer; + assume_graph_lock(); /* FIXME */ + bdrv_check_qiov_request(offset, bytes, qiov, qiov_offset, &error_abort); if (!drv) { @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_pwrite_zeroes(BdrvChild *child, int64_t offset, { IO_CODE(); trace_bdrv_co_pwrite_zeroes(child->bs, offset, bytes, flags); + assert_bdrv_graph_readable(); if (!(child->bs->open_flags & BDRV_O_UNMAP)) { flags &= ~BDRV_REQ_MAY_UNMAP; @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_is_zero_fast(BlockDriverState *bs, int64_t offset, int64_t pnum = bytes; IO_CODE(); - assume_graph_lock(); /* FIXME */ - if (!bytes) { return 1; } @@ -XXX,XX +XXX,XX @@ void bdrv_unregister_buf(BlockDriverState *bs, void *host, size_t size) } } -static int coroutine_fn bdrv_co_copy_range_internal( +static int coroutine_fn GRAPH_RDLOCK bdrv_co_copy_range_internal( BdrvChild *src, int64_t src_offset, BdrvChild *dst, int64_t dst_offset, int64_t bytes, BdrvRequestFlags read_flags, BdrvRequestFlags write_flags, @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_copy_range_from(BdrvChild *src, int64_t src_offset, BdrvRequestFlags write_flags) { IO_CODE(); + assume_graph_lock(); /* FIXME */ trace_bdrv_co_copy_range_from(src, src_offset, dst, dst_offset, bytes, read_flags, write_flags); return bdrv_co_copy_range_internal(src, src_offset, dst, dst_offset, @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_copy_range_to(BdrvChild *src, int64_t src_offset, BdrvRequestFlags write_flags) { IO_CODE(); + assume_graph_lock(); /* FIXME */ trace_bdrv_co_copy_range_to(src, src_offset, dst, dst_offset, bytes, read_flags, write_flags); return bdrv_co_copy_range_internal(src, src_offset, dst, dst_offset, 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 coroutine_fn GRAPH_RDLOCK bdrv_mirror_top_flush(BlockDriverState *bs) return bdrv_co_flush(bs->backing->bs); } -static int coroutine_fn bdrv_mirror_top_pwrite_zeroes(BlockDriverState *bs, - int64_t offset, int64_t bytes, BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +bdrv_mirror_top_pwrite_zeroes(BlockDriverState *bs, int64_t offset, + int64_t bytes, BdrvRequestFlags flags) { - assume_graph_lock(); /* FIXME */ return bdrv_mirror_top_do_write(bs, MIRROR_METHOD_ZERO, offset, bytes, NULL, flags); } diff --git a/block/preallocate.c b/block/preallocate.c index XXXXXXX..XXXXXXX 100644 --- a/block/preallocate.c +++ b/block/preallocate.c @@ -XXX,XX +XXX,XX @@ static bool has_prealloc_perms(BlockDriverState *bs) * want_merge_zero is used to merge write-zero request with preallocation in * one bdrv_co_pwrite_zeroes() call. */ -static bool coroutine_fn handle_write(BlockDriverState *bs, int64_t offset, - int64_t bytes, bool want_merge_zero) +static bool coroutine_fn GRAPH_RDLOCK +handle_write(BlockDriverState *bs, int64_t offset, int64_t bytes, + bool want_merge_zero) { BDRVPreallocateState *s = bs->opaque; int64_t end = offset + bytes; @@ -XXX,XX +XXX,XX @@ static bool coroutine_fn handle_write(BlockDriverState *bs, int64_t offset, return want_merge_zero; } -static int coroutine_fn preallocate_co_pwrite_zeroes(BlockDriverState *bs, - int64_t offset, int64_t bytes, BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +preallocate_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset, + int64_t bytes, BdrvRequestFlags flags) { bool want_merge_zero = !(flags & ~(BDRV_REQ_ZERO_WRITE | BDRV_REQ_NO_FALLBACK)); @@ -XXX,XX +XXX,XX @@ static coroutine_fn int preallocate_co_pwritev_part(BlockDriverState *bs, size_t qiov_offset, BdrvRequestFlags flags) { + assume_graph_lock(); /* FIXME */ handle_write(bs, offset, bytes, false); return bdrv_co_pwritev_part(bs->file, offset, bytes, qiov, qiov_offset, 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 bool merge_cow(uint64_t offset, unsigned bytes, * Return 1 if the COW regions read as zeroes, 0 if not, < 0 on error. * Note that returning 0 does not guarantee non-zero data. */ -static int coroutine_fn is_zero_cow(BlockDriverState *bs, QCowL2Meta *m) +static int coroutine_fn GRAPH_RDLOCK +is_zero_cow(BlockDriverState *bs, QCowL2Meta *m) { /* * This check is designed for optimization shortcut so it must be @@ -XXX,XX +XXX,XX @@ static int coroutine_fn is_zero_cow(BlockDriverState *bs, QCowL2Meta *m) m->cow_end.nb_bytes); } -static int coroutine_fn handle_alloc_space(BlockDriverState *bs, - QCowL2Meta *l2meta) +static int coroutine_fn GRAPH_RDLOCK +handle_alloc_space(BlockDriverState *bs, QCowL2Meta *l2meta) { BDRVQcow2State *s = bs->opaque; QCowL2Meta *m; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn handle_alloc_space(BlockDriverState *bs, * l2meta - if not NULL, qcow2_co_pwritev_task() will consume it. Caller must * not use it somehow after qcow2_co_pwritev_task() call */ -static coroutine_fn int qcow2_co_pwritev_task(BlockDriverState *bs, - uint64_t host_offset, - uint64_t offset, uint64_t bytes, - QEMUIOVector *qiov, - uint64_t qiov_offset, - QCowL2Meta *l2meta) +static coroutine_fn GRAPH_RDLOCK +int qcow2_co_pwritev_task(BlockDriverState *bs, uint64_t host_offset, + uint64_t offset, uint64_t bytes, QEMUIOVector *qiov, + uint64_t qiov_offset, QCowL2Meta *l2meta) { int ret; BDRVQcow2State *s = bs->opaque; @@ -XXX,XX +XXX,XX @@ out_locked: return ret; } -static coroutine_fn int qcow2_co_pwritev_task_entry(AioTask *task) +/* + * This function can count as GRAPH_RDLOCK because qcow2_co_pwritev_part() holds + * the graph lock and keeps it until this coroutine has terminated. + */ +static coroutine_fn GRAPH_RDLOCK int qcow2_co_pwritev_task_entry(AioTask *task) { Qcow2AioTask *t = container_of(task, Qcow2AioTask, task); @@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow2_co_pwritev_part( QCowL2Meta *l2meta = NULL; AioTaskPool *aio = NULL; + assume_graph_lock(); /* FIXME */ + trace_qcow2_writev_start_req(qemu_coroutine_self(), offset, bytes); while (bytes != 0 && aio_task_pool_status(aio) == 0) { @@ -XXX,XX +XXX,XX @@ static bool is_zero(BlockDriverState *bs, int64_t offset, int64_t bytes) return res >= 0 && (res & BDRV_BLOCK_ZERO) && bytes == 0; } -static coroutine_fn int qcow2_co_pwrite_zeroes(BlockDriverState *bs, - int64_t offset, int64_t bytes, BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +qcow2_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset, int64_t bytes, + BdrvRequestFlags flags) { int ret; BDRVQcow2State *s = bs->opaque; 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_writev(BlockDriverState *bs, return qed_co_request(bs, sector_num, qiov, nb_sectors, QED_AIOCB_WRITE); } -static int coroutine_fn bdrv_qed_co_pwrite_zeroes(BlockDriverState *bs, - int64_t offset, - int64_t bytes, - BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +bdrv_qed_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset, int64_t bytes, + BdrvRequestFlags flags) { BDRVQEDState *s = bs->opaque; - assume_graph_lock(); /* FIXME */ - /* * Zero writes start without an I/O buffer. If a buffer becomes necessary * then it will be allocated during request processing. 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 int coroutine_fn quorum_co_preadv(BlockDriverState *bs, return ret; } -static void coroutine_fn write_quorum_entry(void *opaque) +/* + * This function can count as GRAPH_RDLOCK because quorum_co_pwritev() holds the + * graph lock and keeps it until this coroutine has terminated. + */ +static void coroutine_fn GRAPH_RDLOCK write_quorum_entry(void *opaque) { QuorumCo *co = opaque; QuorumAIOCB *acb = co->acb; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn quorum_co_pwritev(BlockDriverState *bs, int64_t offset, QuorumAIOCB *acb = quorum_aio_get(bs, qiov, offset, bytes, flags); int i, ret; + assume_graph_lock(); /* FIXME */ + for (i = 0; i < s->num_children; i++) { Coroutine *co; QuorumCo data = { @@ -XXX,XX +XXX,XX @@ static int coroutine_fn quorum_co_pwritev(BlockDriverState *bs, int64_t offset, return ret; } -static int coroutine_fn quorum_co_pwrite_zeroes(BlockDriverState *bs, - int64_t offset, int64_t bytes, - BdrvRequestFlags flags) - +static int coroutine_fn GRAPH_RDLOCK +quorum_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset, int64_t bytes, + BdrvRequestFlags flags) { return quorum_co_pwritev(bs, offset, bytes, NULL, flags | BDRV_REQ_ZERO_WRITE); 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 int coroutine_fn raw_co_block_status(BlockDriverState *bs, return BDRV_BLOCK_RAW | BDRV_BLOCK_OFFSET_VALID; } -static int coroutine_fn raw_co_pwrite_zeroes(BlockDriverState *bs, - int64_t offset, int64_t bytes, - BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +raw_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset, int64_t bytes, + BdrvRequestFlags flags) { int ret; diff --git a/block/throttle.c b/block/throttle.c index XXXXXXX..XXXXXXX 100644 --- a/block/throttle.c +++ b/block/throttle.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn throttle_co_pwritev(BlockDriverState *bs, return bdrv_co_pwritev(bs->file, offset, bytes, qiov, flags); } -static int coroutine_fn throttle_co_pwrite_zeroes(BlockDriverState *bs, - int64_t offset, int64_t bytes, - BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +throttle_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset, int64_t bytes, + BdrvRequestFlags flags) { ThrottleGroupMember *tgm = bs->opaque; throttle_group_co_io_limits_intercept(tgm, bytes, true); -- 2.39.2 This adds GRAPH_RDLOCK annotations to declare that callers of bdrv_driver_*() need to hold a reader lock for the graph. It doesn't add the annotation to public functions yet. For some places, we know that they will hold the lock, but we don't have the GRAPH_RDLOCK annotations yet. In this case, add assume_graph_lock() with a FIXME comment. These places will be removed once everything is properly annotated. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230203152202.49054-11-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/qcow2.h | 5 ++- include/block/block_int-common.h | 40 ++++++++++--------- block/io.c | 66 +++++++++++++++----------------- block/parallels.c | 8 ++-- block/qcow.c | 20 ++++------ block/qcow2-cluster.c | 10 ++--- block/qcow2.c | 37 ++++++++++-------- block/qed.c | 14 +++---- block/quorum.c | 8 ++-- block/vmdk.c | 4 +- 10 files changed, 101 insertions(+), 111 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 coroutine_fn qcow2_alloc_compressed_cluster_offset(BlockDriverState *bs, void qcow2_parse_compressed_l2_entry(BlockDriverState *bs, uint64_t l2_entry, uint64_t *coffset, int *csize); -int coroutine_fn qcow2_alloc_cluster_link_l2(BlockDriverState *bs, - QCowL2Meta *m); +int coroutine_fn GRAPH_RDLOCK +qcow2_alloc_cluster_link_l2(BlockDriverState *bs, QCowL2Meta *m); + void qcow2_alloc_cluster_abort(BlockDriverState *bs, QCowL2Meta *m); int qcow2_cluster_discard(BlockDriverState *bs, uint64_t offset, uint64_t bytes, enum qcow2_discard_type type, 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 { Error **errp); /* aio */ - BlockAIOCB *(*bdrv_aio_preadv)(BlockDriverState *bs, + BlockAIOCB * GRAPH_RDLOCK_PTR (*bdrv_aio_preadv)(BlockDriverState *bs, int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags, BlockCompletionFunc *cb, void *opaque); - BlockAIOCB *(*bdrv_aio_pwritev)(BlockDriverState *bs, + + BlockAIOCB * GRAPH_RDLOCK_PTR (*bdrv_aio_pwritev)(BlockDriverState *bs, int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags, BlockCompletionFunc *cb, void *opaque); + BlockAIOCB * GRAPH_RDLOCK_PTR (*bdrv_aio_flush)( BlockDriverState *bs, BlockCompletionFunc *cb, void *opaque); @@ -XXX,XX +XXX,XX @@ struct BlockDriver { BlockDriverState *bs, int64_t offset, int bytes, BlockCompletionFunc *cb, void *opaque); - int coroutine_fn (*bdrv_co_readv)(BlockDriverState *bs, + int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_readv)(BlockDriverState *bs, int64_t sector_num, int nb_sectors, QEMUIOVector *qiov); /** @@ -XXX,XX +XXX,XX @@ struct BlockDriver { * * The buffer in @qiov may point directly to guest memory. */ - int coroutine_fn (*bdrv_co_preadv)(BlockDriverState *bs, + int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_preadv)(BlockDriverState *bs, int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags); - int coroutine_fn (*bdrv_co_preadv_part)(BlockDriverState *bs, - int64_t offset, int64_t bytes, + int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_preadv_part)( + BlockDriverState *bs, int64_t offset, int64_t bytes, QEMUIOVector *qiov, size_t qiov_offset, BdrvRequestFlags flags); - int coroutine_fn (*bdrv_co_writev)(BlockDriverState *bs, + int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_writev)(BlockDriverState *bs, int64_t sector_num, int nb_sectors, QEMUIOVector *qiov, int flags); /** @@ -XXX,XX +XXX,XX @@ struct BlockDriver { * * The buffer in @qiov may point directly to guest memory. */ - int coroutine_fn (*bdrv_co_pwritev)(BlockDriverState *bs, - int64_t offset, int64_t bytes, QEMUIOVector *qiov, - BdrvRequestFlags flags); - int coroutine_fn (*bdrv_co_pwritev_part)(BlockDriverState *bs, - int64_t offset, int64_t bytes, QEMUIOVector *qiov, size_t qiov_offset, + int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_pwritev)( + BlockDriverState *bs, int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags); + int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_pwritev_part)( + BlockDriverState *bs, int64_t offset, int64_t bytes, QEMUIOVector *qiov, + size_t qiov_offset, BdrvRequestFlags flags); /* * Efficiently zero a region of the disk image. Typically an image format @@ -XXX,XX +XXX,XX @@ struct BlockDriver { BlockMeasureInfo *(*bdrv_measure)(QemuOpts *opts, BlockDriverState *in_bs, Error **errp); - int coroutine_fn (*bdrv_co_pwritev_compressed)(BlockDriverState *bs, - int64_t offset, int64_t bytes, QEMUIOVector *qiov); - int coroutine_fn (*bdrv_co_pwritev_compressed_part)(BlockDriverState *bs, - int64_t offset, int64_t bytes, QEMUIOVector *qiov, - size_t qiov_offset); + int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_pwritev_compressed)( + BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov); + + int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_pwritev_compressed_part)( + 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); @@ -XXX,XX +XXX,XX @@ struct BlockDriver { BlockDriverState *bs, const char *name, Error **errp); }; -static inline bool block_driver_can_compress(BlockDriver *drv) +static inline bool TSA_NO_TSA block_driver_can_compress(BlockDriver *drv) { return drv->bdrv_co_pwritev_compressed || drv->bdrv_co_pwritev_compressed_part; 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); @@ -XXX,XX +XXX,XX @@ static void bdrv_co_io_em_complete(void *opaque, int ret) aio_co_wake(co->coroutine); } -static int coroutine_fn bdrv_driver_preadv(BlockDriverState *bs, - int64_t offset, int64_t bytes, - QEMUIOVector *qiov, - size_t qiov_offset, int flags) +static int coroutine_fn GRAPH_RDLOCK +bdrv_driver_preadv(BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, size_t qiov_offset, int flags) { BlockDriver *drv = bs->drv; int64_t sector_num; @@ -XXX,XX +XXX,XX @@ out: return ret; } -static int coroutine_fn bdrv_driver_pwritev(BlockDriverState *bs, - int64_t offset, int64_t bytes, - QEMUIOVector *qiov, - size_t qiov_offset, - BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +bdrv_driver_pwritev(BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, size_t qiov_offset, + BdrvRequestFlags flags) { BlockDriver *drv = bs->drv; bool emulate_fua = false; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_driver_pwritev(BlockDriverState *bs, QEMUIOVector local_qiov; int ret; - assume_graph_lock(); /* FIXME */ - bdrv_check_qiov_request(offset, bytes, qiov, qiov_offset, &error_abort); if (!drv) { @@ -XXX,XX +XXX,XX @@ emulate_flags: return ret; } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK bdrv_driver_pwritev_compressed(BlockDriverState *bs, int64_t offset, int64_t bytes, QEMUIOVector *qiov, size_t qiov_offset) @@ -XXX,XX +XXX,XX @@ bdrv_driver_pwritev_compressed(BlockDriverState *bs, int64_t offset, return ret; } -static int coroutine_fn bdrv_co_do_copy_on_readv(BdrvChild *child, - int64_t offset, int64_t bytes, QEMUIOVector *qiov, - size_t qiov_offset, int flags) +static int coroutine_fn GRAPH_RDLOCK +bdrv_co_do_copy_on_readv(BdrvChild *child, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, size_t qiov_offset, int flags) { BlockDriverState *bs = child->bs; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_copy_on_readv(BdrvChild *child, int64_t progress = 0; bool skip_write; - assume_graph_lock(); /* FIXME */ - bdrv_check_qiov_request(offset, bytes, qiov, qiov_offset, &error_abort); if (!drv) { @@ -XXX,XX +XXX,XX @@ err: * handles copy on read, zeroing after EOF, and fragmentation of large * reads; any other features must be implemented by the caller. */ -static int coroutine_fn bdrv_aligned_preadv(BdrvChild *child, - BdrvTrackedRequest *req, int64_t offset, int64_t bytes, - int64_t align, QEMUIOVector *qiov, size_t qiov_offset, int flags) +static int coroutine_fn GRAPH_RDLOCK +bdrv_aligned_preadv(BdrvChild *child, BdrvTrackedRequest *req, + int64_t offset, int64_t bytes, int64_t align, + QEMUIOVector *qiov, size_t qiov_offset, int flags) { BlockDriverState *bs = child->bs; int64_t total_bytes, max_bytes; @@ -XXX,XX +XXX,XX @@ static bool bdrv_init_padding(BlockDriverState *bs, return true; } -static coroutine_fn int bdrv_padding_rmw_read(BdrvChild *child, - BdrvTrackedRequest *req, - BdrvRequestPadding *pad, - bool zero_middle) +static int coroutine_fn GRAPH_RDLOCK +bdrv_padding_rmw_read(BdrvChild *child, BdrvTrackedRequest *req, + BdrvRequestPadding *pad, bool zero_middle) { QEMUIOVector local_qiov; BlockDriverState *bs = child->bs; @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_preadv_part(BdrvChild *child, int ret; IO_CODE(); + assume_graph_lock(); /* FIXME */ + trace_bdrv_co_preadv_part(bs, offset, bytes, flags); if (!bdrv_co_is_inserted(bs)) { @@ -XXX,XX +XXX,XX @@ bdrv_co_write_req_finish(BdrvChild *child, int64_t offset, int64_t bytes, * Forwards an already correctly aligned write request to the BlockDriver, * after possibly fragmenting it. */ -static int coroutine_fn bdrv_aligned_pwritev(BdrvChild *child, - BdrvTrackedRequest *req, int64_t offset, int64_t bytes, - int64_t align, QEMUIOVector *qiov, size_t qiov_offset, - BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +bdrv_aligned_pwritev(BdrvChild *child, BdrvTrackedRequest *req, + int64_t offset, int64_t bytes, int64_t align, + QEMUIOVector *qiov, size_t qiov_offset, + BdrvRequestFlags flags) { BlockDriverState *bs = child->bs; BlockDriver *drv = bs->drv; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_pwritev(BdrvChild *child, int64_t bytes_remaining = bytes; int max_transfer; - assume_graph_lock(); /* FIXME */ - bdrv_check_qiov_request(offset, bytes, qiov, qiov_offset, &error_abort); if (!drv) { @@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_pwritev(BdrvChild *child, return ret; } -static int coroutine_fn bdrv_co_do_zero_pwritev(BdrvChild *child, - int64_t offset, - int64_t bytes, - BdrvRequestFlags flags, - BdrvTrackedRequest *req) +static int coroutine_fn GRAPH_RDLOCK +bdrv_co_do_zero_pwritev(BdrvChild *child, int64_t offset, int64_t bytes, + BdrvRequestFlags flags, BdrvTrackedRequest *req) { BlockDriverState *bs = child->bs; QEMUIOVector local_qiov; @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_pwritev_part(BdrvChild *child, bool padded = false; IO_CODE(); + assume_graph_lock(); /* FIXME */ + trace_bdrv_co_pwritev_part(child->bs, offset, bytes, flags); if (!bdrv_co_is_inserted(bs)) { 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_block_status(BlockDriverState *bs, return BDRV_BLOCK_DATA | BDRV_BLOCK_OFFSET_VALID; } -static coroutine_fn int parallels_co_writev(BlockDriverState *bs, - int64_t sector_num, int nb_sectors, - QEMUIOVector *qiov, int flags) +static int coroutine_fn GRAPH_RDLOCK +parallels_co_writev(BlockDriverState *bs, int64_t sector_num, int nb_sectors, + QEMUIOVector *qiov, int flags) { BDRVParallelsState *s = bs->opaque; uint64_t bytes_done = 0; QEMUIOVector hd_qiov; int ret = 0; - assume_graph_lock(); /* FIXME */ - qemu_iovec_init(&hd_qiov, qiov->niov); while (nb_sectors > 0) { 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 void qcow_refresh_limits(BlockDriverState *bs, Error **errp) bs->bl.request_alignment = BDRV_SECTOR_SIZE; } -static coroutine_fn int qcow_co_preadv(BlockDriverState *bs, int64_t offset, - int64_t bytes, QEMUIOVector *qiov, - BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +qcow_co_preadv(BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, BdrvRequestFlags flags) { BDRVQcowState *s = bs->opaque; int offset_in_cluster; @@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow_co_preadv(BlockDriverState *bs, int64_t offset, uint8_t *buf; void *orig_buf; - assume_graph_lock(); /* FIXME */ - if (qiov->niov > 1) { buf = orig_buf = qemu_try_blockalign(bs, qiov->size); if (buf == NULL) { @@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow_co_preadv(BlockDriverState *bs, int64_t offset, return ret; } -static coroutine_fn int qcow_co_pwritev(BlockDriverState *bs, int64_t offset, - int64_t bytes, QEMUIOVector *qiov, - BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +qcow_co_pwritev(BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, BdrvRequestFlags flags) { BDRVQcowState *s = bs->opaque; int offset_in_cluster; @@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow_co_pwritev(BlockDriverState *bs, int64_t offset, uint8_t *buf; void *orig_buf; - assume_graph_lock(); /* FIXME */ - s->cluster_cache_offset = -1; /* disable compressed cache */ /* We must always copy the iov when encrypting, so we @@ -XXX,XX +XXX,XX @@ static int qcow_make_empty(BlockDriverState *bs) /* XXX: put compressed sectors first, then all the cluster aligned tables to avoid losing bytes in alignment */ -static coroutine_fn int +static int coroutine_fn GRAPH_RDLOCK qcow_co_pwritev_compressed(BlockDriverState *bs, int64_t offset, int64_t bytes, QEMUIOVector *qiov) { @@ -XXX,XX +XXX,XX @@ qcow_co_pwritev_compressed(BlockDriverState *bs, int64_t offset, int64_t bytes, uint8_t *buf, *out_buf; uint64_t cluster_offset; - assume_graph_lock(); /* FIXME */ - buf = qemu_blockalign(bs, s->cluster_size); if (bytes != s->cluster_size) { if (bytes > s->cluster_size || diff --git a/block/qcow2-cluster.c b/block/qcow2-cluster.c index XXXXXXX..XXXXXXX 100644 --- a/block/qcow2-cluster.c +++ b/block/qcow2-cluster.c @@ -XXX,XX +XXX,XX @@ static int count_contiguous_subclusters(BlockDriverState *bs, int nb_clusters, return count; } -static int coroutine_fn do_perform_cow_read(BlockDriverState *bs, - uint64_t src_cluster_offset, - unsigned offset_in_cluster, - QEMUIOVector *qiov) +static int coroutine_fn GRAPH_RDLOCK +do_perform_cow_read(BlockDriverState *bs, uint64_t src_cluster_offset, + unsigned offset_in_cluster, QEMUIOVector *qiov) { int ret; @@ -XXX,XX +XXX,XX @@ int coroutine_fn qcow2_alloc_compressed_cluster_offset(BlockDriverState *bs, return 0; } -static int coroutine_fn perform_cow(BlockDriverState *bs, QCowL2Meta *m) +static int coroutine_fn GRAPH_RDLOCK +perform_cow(BlockDriverState *bs, QCowL2Meta *m) { BDRVQcow2State *s = bs->opaque; Qcow2COWRegion *start = &m->cow_start; 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 int coroutine_fn qcow2_co_block_status(BlockDriverState *bs, return status; } -static coroutine_fn int qcow2_handle_l2meta(BlockDriverState *bs, - QCowL2Meta **pl2meta, - bool link_l2) +static int coroutine_fn GRAPH_RDLOCK +qcow2_handle_l2meta(BlockDriverState *bs, QCowL2Meta **pl2meta, bool link_l2) { int ret = 0; QCowL2Meta *l2meta = *pl2meta; @@ -XXX,XX +XXX,XX @@ static coroutine_fn GRAPH_RDLOCK int qcow2_co_pwritev_task_entry(AioTask *task) t->l2meta); } -static coroutine_fn int qcow2_co_pwritev_part( - BlockDriverState *bs, int64_t offset, int64_t bytes, - QEMUIOVector *qiov, size_t qiov_offset, BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +qcow2_co_pwritev_part(BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, size_t qiov_offset, + BdrvRequestFlags flags) { BDRVQcow2State *s = bs->opaque; int offset_in_cluster; @@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow2_co_pwritev_part( QCowL2Meta *l2meta = NULL; AioTaskPool *aio = NULL; - assume_graph_lock(); /* FIXME */ - trace_qcow2_writev_start_req(qemu_coroutine_self(), offset, bytes); while (bytes != 0 && aio_task_pool_status(aio) == 0) { @@ -XXX,XX +XXX,XX @@ qcow2_co_copy_range_to(BlockDriverState *bs, uint64_t host_offset; QCowL2Meta *l2meta = NULL; + assume_graph_lock(); /* FIXME */ assert(!bs->encrypted); qemu_co_mutex_lock(&s->lock); @@ -XXX,XX +XXX,XX @@ fail: return ret; } -static coroutine_fn int +static int coroutine_fn GRAPH_RDLOCK qcow2_co_pwritev_compressed_task(BlockDriverState *bs, uint64_t offset, uint64_t bytes, QEMUIOVector *qiov, size_t qiov_offset) @@ -XXX,XX +XXX,XX @@ fail: return ret; } -static coroutine_fn int qcow2_co_pwritev_compressed_task_entry(AioTask *task) +/* + * This function can count as GRAPH_RDLOCK because + * qcow2_co_pwritev_compressed_part() holds the graph lock and keeps it until + * this coroutine has terminated. + */ +static int coroutine_fn GRAPH_RDLOCK +qcow2_co_pwritev_compressed_task_entry(AioTask *task) { Qcow2AioTask *t = container_of(task, Qcow2AioTask, task); @@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow2_co_pwritev_compressed_task_entry(AioTask *task) * XXX: put compressed sectors first, then all the cluster aligned * tables to avoid losing bytes in alignment */ -static coroutine_fn int +static int coroutine_fn GRAPH_RDLOCK qcow2_co_pwritev_compressed_part(BlockDriverState *bs, int64_t offset, int64_t bytes, QEMUIOVector *qiov, size_t qiov_offset) @@ -XXX,XX +XXX,XX @@ qcow2_co_pwritev_compressed_part(BlockDriverState *bs, AioTaskPool *aio = NULL; int ret = 0; - assume_graph_lock(); /* FIXME */ - if (has_data_file(bs)) { return -ENOTSUP; } @@ -XXX,XX +XXX,XX @@ static int64_t qcow2_check_vmstate_request(BlockDriverState *bs, return pos; } -static coroutine_fn int qcow2_co_save_vmstate(BlockDriverState *bs, - QEMUIOVector *qiov, int64_t pos) +static int coroutine_fn GRAPH_RDLOCK +qcow2_co_save_vmstate(BlockDriverState *bs, QEMUIOVector *qiov, int64_t pos) { int64_t offset = qcow2_check_vmstate_request(bs, qiov, pos); if (offset < 0) { @@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow2_co_save_vmstate(BlockDriverState *bs, return bs->drv->bdrv_co_pwritev_part(bs, offset, qiov->size, qiov, 0, 0); } -static coroutine_fn int qcow2_co_load_vmstate(BlockDriverState *bs, - QEMUIOVector *qiov, int64_t pos) +static int coroutine_fn GRAPH_RDLOCK +qcow2_co_load_vmstate(BlockDriverState *bs, QEMUIOVector *qiov, int64_t pos) { int64_t offset = qcow2_check_vmstate_request(bs, qiov, pos); if (offset < 0) { 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 @@ qed_co_request(BlockDriverState *bs, int64_t sector_num, QEMUIOVector *qiov, return qed_aio_next_io(&acb); } -static int coroutine_fn bdrv_qed_co_readv(BlockDriverState *bs, - int64_t sector_num, int nb_sectors, - QEMUIOVector *qiov) +static int coroutine_fn GRAPH_RDLOCK +bdrv_qed_co_readv(BlockDriverState *bs, int64_t sector_num, int nb_sectors, + QEMUIOVector *qiov) { - assume_graph_lock(); /* FIXME */ return qed_co_request(bs, sector_num, qiov, nb_sectors, 0); } -static int coroutine_fn bdrv_qed_co_writev(BlockDriverState *bs, - int64_t sector_num, int nb_sectors, - QEMUIOVector *qiov, int flags) +static int coroutine_fn GRAPH_RDLOCK +bdrv_qed_co_writev(BlockDriverState *bs, int64_t sector_num, int nb_sectors, + QEMUIOVector *qiov, int flags) { - assume_graph_lock(); /* FIXME */ return qed_co_request(bs, sector_num, qiov, nb_sectors, QED_AIOCB_WRITE); } 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 void coroutine_fn GRAPH_RDLOCK write_quorum_entry(void *opaque) } } -static int coroutine_fn quorum_co_pwritev(BlockDriverState *bs, int64_t offset, - int64_t bytes, QEMUIOVector *qiov, - BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +quorum_co_pwritev(BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, BdrvRequestFlags flags) { BDRVQuorumState *s = bs->opaque; QuorumAIOCB *acb = quorum_aio_get(bs, qiov, offset, bytes, flags); int i, ret; - assume_graph_lock(); /* FIXME */ - for (i = 0; i < s->num_children; i++) { Coroutine *co; QuorumCo data = { 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 @@ vmdk_co_pwritev(BlockDriverState *bs, int64_t offset, int64_t bytes, return ret; } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK vmdk_co_pwritev_compressed(BlockDriverState *bs, int64_t offset, int64_t bytes, QEMUIOVector *qiov) { - assume_graph_lock(); /* FIXME */ - if (bytes == 0) { /* The caller will write bytes 0 to signal EOF. * When receive it, we align EOF to a sector boundary. */ -- 2.39.2 This adds GRAPH_RDLOCK annotations to declare that callers of bdrv_co_pread*/pwrite*() need to hold a reader lock for the graph. For some places, we know that they will hold the lock, but we don't have the GRAPH_RDLOCK annotations yet. In this case, add assume_graph_lock() with a FIXME comment. These places will be removed once everything is properly annotated. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230203152202.49054-12-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/qcow2.h | 9 ++--- block/qed.h | 18 +++++----- include/block/block_int-io.h | 14 ++++---- block/blkdebug.c | 4 +-- block/blklogwrites.c | 7 ++-- block/blkreplay.c | 10 +++--- block/block-backend.c | 2 ++ block/bochs.c | 2 +- block/commit.c | 5 +-- block/copy-before-write.c | 16 ++++----- block/copy-on-read.c | 26 ++++++--------- block/crypto.c | 4 +-- block/filter-compress.c | 19 +++++------ block/io.c | 7 ++-- block/mirror.c | 18 +++++----- block/parallels.c | 8 +++-- block/preallocate.c | 18 +++++----- block/qcow.c | 8 ++--- block/qcow2-cluster.c | 7 ++-- block/qcow2.c | 53 +++++++++++++++-------------- block/qed-table.c | 4 +-- block/qed.c | 31 +++++++++-------- block/quorum.c | 29 ++++++++++------ block/raw-format.c | 12 +++---- block/replication.c | 15 ++++----- block/throttle.c | 21 +++++------- block/vdi.c | 4 +-- block/vhdx.c | 11 +++--- block/vmdk.c | 65 +++++++++++++++--------------------- block/vpc.c | 4 +-- tests/unit/test-bdrv-drain.c | 20 ++++++----- 31 files changed, 233 insertions(+), 238 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_validate_table(BlockDriverState *bs, uint64_t offset, Error **errp); /* qcow2-refcount.c functions */ -int coroutine_fn qcow2_refcount_init(BlockDriverState *bs); +int coroutine_fn GRAPH_RDLOCK qcow2_refcount_init(BlockDriverState *bs); void qcow2_refcount_close(BlockDriverState *bs); int qcow2_get_refcount(BlockDriverState *bs, int64_t cluster_index, @@ -XXX,XX +XXX,XX @@ void qcow2_free_snapshots(BlockDriverState *bs); int qcow2_read_snapshots(BlockDriverState *bs, Error **errp); int qcow2_write_snapshots(BlockDriverState *bs); -int coroutine_fn qcow2_check_read_snapshot_table(BlockDriverState *bs, - BdrvCheckResult *result, - BdrvCheckMode fix); +int coroutine_fn GRAPH_RDLOCK +qcow2_check_read_snapshot_table(BlockDriverState *bs, BdrvCheckResult *result, + BdrvCheckMode fix); + int coroutine_fn qcow2_check_fix_snapshot_table(BlockDriverState *bs, BdrvCheckResult *result, BdrvCheckMode fix); diff --git a/block/qed.h b/block/qed.h index XXXXXXX..XXXXXXX 100644 --- a/block/qed.h +++ b/block/qed.h @@ -XXX,XX +XXX,XX @@ void qed_commit_l2_cache_entry(L2TableCache *l2_cache, CachedL2Table *l2_table); /** * Table I/O functions */ -int coroutine_fn qed_read_l1_table_sync(BDRVQEDState *s); +int coroutine_fn GRAPH_RDLOCK qed_read_l1_table_sync(BDRVQEDState *s); int coroutine_fn GRAPH_RDLOCK qed_write_l1_table(BDRVQEDState *s, unsigned int index, unsigned int n); @@ -XXX,XX +XXX,XX @@ qed_write_l1_table(BDRVQEDState *s, unsigned int index, unsigned int n); int coroutine_fn GRAPH_RDLOCK qed_write_l1_table_sync(BDRVQEDState *s, unsigned int index, unsigned int n); -int coroutine_fn qed_read_l2_table_sync(BDRVQEDState *s, QEDRequest *request, - uint64_t offset); -int coroutine_fn qed_read_l2_table(BDRVQEDState *s, QEDRequest *request, - uint64_t offset); +int coroutine_fn GRAPH_RDLOCK +qed_read_l2_table_sync(BDRVQEDState *s, QEDRequest *request, uint64_t offset); + +int coroutine_fn GRAPH_RDLOCK +qed_read_l2_table(BDRVQEDState *s, QEDRequest *request, uint64_t offset); int coroutine_fn GRAPH_RDLOCK qed_write_l2_table(BDRVQEDState *s, QEDRequest *request, unsigned int index, @@ -XXX,XX +XXX,XX @@ qed_write_l2_table_sync(BDRVQEDState *s, QEDRequest *request, /** * Cluster functions */ -int coroutine_fn qed_find_cluster(BDRVQEDState *s, QEDRequest *request, - uint64_t pos, size_t *len, - uint64_t *img_offset); +int coroutine_fn GRAPH_RDLOCK +qed_find_cluster(BDRVQEDState *s, QEDRequest *request, uint64_t pos, + size_t *len, uint64_t *img_offset); /** * Consistency check @@ -XXX,XX +XXX,XX @@ int coroutine_fn qed_find_cluster(BDRVQEDState *s, QEDRequest *request, int coroutine_fn GRAPH_RDLOCK qed_check(BDRVQEDState *s, BdrvCheckResult *result, bool fix); - QEDTable *qed_alloc_table(BDRVQEDState *s); /** diff --git a/include/block/block_int-io.h b/include/block/block_int-io.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block_int-io.h +++ b/include/block/block_int-io.h @@ -XXX,XX +XXX,XX @@ int coroutine_fn GRAPH_RDLOCK bdrv_co_pdiscard_snapshot(BlockDriverState *bs, int64_t offset, int64_t bytes); -int coroutine_fn bdrv_co_preadv(BdrvChild *child, +int coroutine_fn GRAPH_RDLOCK bdrv_co_preadv(BdrvChild *child, int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags); -int coroutine_fn bdrv_co_preadv_part(BdrvChild *child, +int coroutine_fn GRAPH_RDLOCK bdrv_co_preadv_part(BdrvChild *child, int64_t offset, int64_t bytes, QEMUIOVector *qiov, size_t qiov_offset, BdrvRequestFlags flags); -int coroutine_fn bdrv_co_pwritev(BdrvChild *child, +int coroutine_fn GRAPH_RDLOCK bdrv_co_pwritev(BdrvChild *child, int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags); -int coroutine_fn bdrv_co_pwritev_part(BdrvChild *child, +int coroutine_fn GRAPH_RDLOCK bdrv_co_pwritev_part(BdrvChild *child, int64_t offset, int64_t bytes, QEMUIOVector *qiov, size_t qiov_offset, BdrvRequestFlags flags); -static inline int coroutine_fn bdrv_co_pread(BdrvChild *child, +static inline int coroutine_fn GRAPH_RDLOCK bdrv_co_pread(BdrvChild *child, int64_t offset, int64_t bytes, void *buf, BdrvRequestFlags flags) { QEMUIOVector qiov = QEMU_IOVEC_INIT_BUF(qiov, buf, bytes); IO_CODE(); + assert_bdrv_graph_readable(); return bdrv_co_preadv(child, offset, bytes, &qiov, flags); } -static inline int coroutine_fn bdrv_co_pwrite(BdrvChild *child, +static inline int coroutine_fn GRAPH_RDLOCK bdrv_co_pwrite(BdrvChild *child, int64_t offset, int64_t bytes, const void *buf, BdrvRequestFlags flags) { QEMUIOVector qiov = QEMU_IOVEC_INIT_BUF(qiov, buf, bytes); IO_CODE(); + assert_bdrv_graph_readable(); return bdrv_co_pwritev(child, offset, bytes, &qiov, flags); } diff --git a/block/blkdebug.c b/block/blkdebug.c index XXXXXXX..XXXXXXX 100644 --- a/block/blkdebug.c +++ b/block/blkdebug.c @@ -XXX,XX +XXX,XX @@ static int rule_check(BlockDriverState *bs, uint64_t offset, uint64_t bytes, return -error; } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK blkdebug_co_preadv(BlockDriverState *bs, int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags) { @@ -XXX,XX +XXX,XX @@ blkdebug_co_preadv(BlockDriverState *bs, int64_t offset, int64_t bytes, return bdrv_co_preadv(bs->file, offset, bytes, qiov, flags); } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK blkdebug_co_pwritev(BlockDriverState *bs, int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags) { diff --git a/block/blklogwrites.c b/block/blklogwrites.c index XXXXXXX..XXXXXXX 100644 --- a/block/blklogwrites.c +++ b/block/blklogwrites.c @@ -XXX,XX +XXX,XX @@ static void blk_log_writes_refresh_limits(BlockDriverState *bs, Error **errp) bs->bl.request_alignment = s->sectorsize; } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK blk_log_writes_co_preadv(BlockDriverState *bs, int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags) { @@ -XXX,XX +XXX,XX @@ blk_log_writes_co_log(BlockDriverState *bs, uint64_t offset, uint64_t bytes, return fr.file_ret; } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK blk_log_writes_co_do_file_pwritev(BlkLogWritesFileReq *fr) { return bdrv_co_pwritev(fr->bs->file, fr->offset, fr->bytes, @@ -XXX,XX +XXX,XX @@ blk_log_writes_co_do_file_pdiscard(BlkLogWritesFileReq *fr) return bdrv_co_pdiscard(fr->bs->file, fr->offset, fr->bytes); } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK blk_log_writes_co_pwritev(BlockDriverState *bs, int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags) { - assume_graph_lock(); /* FIXME */ return blk_log_writes_co_log(bs, offset, bytes, qiov, flags, blk_log_writes_co_do_file_pwritev, 0, false); } diff --git a/block/blkreplay.c b/block/blkreplay.c index XXXXXXX..XXXXXXX 100644 --- a/block/blkreplay.c +++ b/block/blkreplay.c @@ -XXX,XX +XXX,XX @@ static void block_request_create(uint64_t reqid, BlockDriverState *bs, replay_block_event(req->bh, reqid); } -static int coroutine_fn blkreplay_co_preadv(BlockDriverState *bs, - int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +blkreplay_co_preadv(BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, BdrvRequestFlags flags) { uint64_t reqid = blkreplay_next_id(); int ret = bdrv_co_preadv(bs->file, offset, bytes, qiov, flags); @@ -XXX,XX +XXX,XX @@ static int coroutine_fn blkreplay_co_preadv(BlockDriverState *bs, return ret; } -static int coroutine_fn blkreplay_co_pwritev(BlockDriverState *bs, - int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +blkreplay_co_pwritev(BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, BdrvRequestFlags flags) { uint64_t reqid = blkreplay_next_id(); int ret = bdrv_co_pwritev(bs->file, offset, bytes, qiov, flags); 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 @@ blk_co_do_preadv_part(BlockBackend *blk, int64_t offset, int64_t bytes, IO_CODE(); blk_wait_while_drained(blk); + GRAPH_RDLOCK_GUARD(); /* Call blk_bs() only after waiting, the graph may have changed */ bs = blk_bs(blk); @@ -XXX,XX +XXX,XX @@ blk_co_do_pwritev_part(BlockBackend *blk, int64_t offset, int64_t bytes, IO_CODE(); blk_wait_while_drained(blk); + GRAPH_RDLOCK_GUARD(); /* Call blk_bs() only after waiting, the graph may have changed */ bs = blk_bs(blk); diff --git a/block/bochs.c b/block/bochs.c index XXXXXXX..XXXXXXX 100644 --- a/block/bochs.c +++ b/block/bochs.c @@ -XXX,XX +XXX,XX @@ static int64_t seek_to_sector(BlockDriverState *bs, int64_t sector_num) return bitmap_offset + (512 * (s->bitmap_blocks + extent_offset)); } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK bochs_co_preadv(BlockDriverState *bs, int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags) { diff --git a/block/commit.c b/block/commit.c index XXXXXXX..XXXXXXX 100644 --- a/block/commit.c +++ b/block/commit.c @@ -XXX,XX +XXX,XX @@ static const BlockJobDriver commit_job_driver = { }, }; -static int coroutine_fn bdrv_commit_top_preadv(BlockDriverState *bs, - int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +bdrv_commit_top_preadv(BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, BdrvRequestFlags flags) { return bdrv_co_preadv(bs->backing, offset, bytes, qiov, flags); } diff --git a/block/copy-before-write.c b/block/copy-before-write.c index XXXXXXX..XXXXXXX 100644 --- a/block/copy-before-write.c +++ b/block/copy-before-write.c @@ -XXX,XX +XXX,XX @@ typedef struct BDRVCopyBeforeWriteState { int snapshot_error; } BDRVCopyBeforeWriteState; -static coroutine_fn int cbw_co_preadv( - BlockDriverState *bs, int64_t offset, int64_t bytes, - QEMUIOVector *qiov, BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +cbw_co_preadv(BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, BdrvRequestFlags flags) { return bdrv_co_preadv(bs->file, offset, bytes, qiov, flags); } @@ -XXX,XX +XXX,XX @@ cbw_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset, int64_t bytes, return bdrv_co_pwrite_zeroes(bs->file, offset, bytes, flags); } -static coroutine_fn int cbw_co_pwritev(BlockDriverState *bs, - int64_t offset, - int64_t bytes, - QEMUIOVector *qiov, - BdrvRequestFlags flags) +static coroutine_fn GRAPH_RDLOCK +int cbw_co_pwritev(BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, BdrvRequestFlags flags) { int ret = cbw_do_copy_before_write(bs, offset, bytes, flags); if (ret < 0) { @@ -XXX,XX +XXX,XX @@ cbw_co_preadv_snapshot(BlockDriverState *bs, int64_t offset, int64_t bytes, BdrvChild *file; int ret; + assume_graph_lock(); /* FIXME */ + /* TODO: upgrade to async loop using AioTask */ while (bytes) { int64_t cur_bytes; diff --git a/block/copy-on-read.c b/block/copy-on-read.c index XXXXXXX..XXXXXXX 100644 --- a/block/copy-on-read.c +++ b/block/copy-on-read.c @@ -XXX,XX +XXX,XX @@ static int64_t coroutine_fn cor_co_getlength(BlockDriverState *bs) } -static int coroutine_fn cor_co_preadv_part(BlockDriverState *bs, - int64_t offset, int64_t bytes, - QEMUIOVector *qiov, - size_t qiov_offset, - BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +cor_co_preadv_part(BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, size_t qiov_offset, + BdrvRequestFlags flags) { int64_t n; int local_flags; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn cor_co_preadv_part(BlockDriverState *bs, } -static int coroutine_fn cor_co_pwritev_part(BlockDriverState *bs, - int64_t offset, - int64_t bytes, - QEMUIOVector *qiov, - size_t qiov_offset, - BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +cor_co_pwritev_part(BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, size_t qiov_offset, + BdrvRequestFlags flags) { return bdrv_co_pwritev_part(bs->file, offset, bytes, qiov, qiov_offset, flags); @@ -XXX,XX +XXX,XX @@ cor_co_pdiscard(BlockDriverState *bs, int64_t offset, int64_t bytes) } -static int coroutine_fn cor_co_pwritev_compressed(BlockDriverState *bs, - int64_t offset, - int64_t bytes, - QEMUIOVector *qiov) +static int coroutine_fn GRAPH_RDLOCK +cor_co_pwritev_compressed(BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov) { return bdrv_co_pwritev(bs->file, offset, bytes, qiov, BDRV_REQ_WRITE_COMPRESSED); 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 @@ static int block_crypto_reopen_prepare(BDRVReopenState *state, */ #define BLOCK_CRYPTO_MAX_IO_SIZE (1024 * 1024) -static coroutine_fn int +static int coroutine_fn GRAPH_RDLOCK block_crypto_co_preadv(BlockDriverState *bs, int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags) { @@ -XXX,XX +XXX,XX @@ block_crypto_co_preadv(BlockDriverState *bs, int64_t offset, int64_t bytes, } -static coroutine_fn int +static int coroutine_fn GRAPH_RDLOCK block_crypto_co_pwritev(BlockDriverState *bs, int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags) { diff --git a/block/filter-compress.c b/block/filter-compress.c index XXXXXXX..XXXXXXX 100644 --- a/block/filter-compress.c +++ b/block/filter-compress.c @@ -XXX,XX +XXX,XX @@ static int64_t coroutine_fn compress_co_getlength(BlockDriverState *bs) } -static int coroutine_fn compress_co_preadv_part(BlockDriverState *bs, - int64_t offset, int64_t bytes, - QEMUIOVector *qiov, - size_t qiov_offset, - BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +compress_co_preadv_part(BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, size_t qiov_offset, + BdrvRequestFlags flags) { return bdrv_co_preadv_part(bs->file, offset, bytes, qiov, qiov_offset, flags); } -static int coroutine_fn compress_co_pwritev_part(BlockDriverState *bs, - int64_t offset, - int64_t bytes, - QEMUIOVector *qiov, - size_t qiov_offset, - BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +compress_co_pwritev_part(BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, size_t qiov_offset, + BdrvRequestFlags flags) { return bdrv_co_pwritev_part(bs->file, offset, bytes, qiov, qiov_offset, flags | BDRV_REQ_WRITE_COMPRESSED); 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 @@ bdrv_driver_preadv(BlockDriverState *bs, int64_t offset, int64_t bytes, unsigned int nb_sectors; QEMUIOVector local_qiov; int ret; + assert_bdrv_graph_readable(); bdrv_check_qiov_request(offset, bytes, qiov, qiov_offset, &error_abort); assert(!(flags & ~bs->supported_read_flags)); @@ -XXX,XX +XXX,XX @@ bdrv_driver_pwritev(BlockDriverState *bs, int64_t offset, int64_t bytes, unsigned int nb_sectors; QEMUIOVector local_qiov; int ret; + assert_bdrv_graph_readable(); bdrv_check_qiov_request(offset, bytes, qiov, qiov_offset, &error_abort); @@ -XXX,XX +XXX,XX @@ bdrv_driver_pwritev_compressed(BlockDriverState *bs, int64_t offset, BlockDriver *drv = bs->drv; QEMUIOVector local_qiov; int ret; + assert_bdrv_graph_readable(); bdrv_check_qiov_request(offset, bytes, qiov, qiov_offset, &error_abort); @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_preadv_part(BdrvChild *child, int ret; IO_CODE(); - assume_graph_lock(); /* FIXME */ - trace_bdrv_co_preadv_part(bs, offset, bytes, flags); if (!bdrv_co_is_inserted(bs)) { @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_pwritev_part(BdrvChild *child, bool padded = false; IO_CODE(); - assume_graph_lock(); /* FIXME */ - trace_bdrv_co_pwritev_part(child->bs, offset, bytes, flags); if (!bdrv_co_is_inserted(bs)) { 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 void coroutine_fn mirror_co_read(void *opaque) op->is_in_flight = true; trace_mirror_one_iteration(s, op->offset, op->bytes); - ret = bdrv_co_preadv(s->mirror_top_bs->backing, op->offset, op->bytes, - &op->qiov, 0); + WITH_GRAPH_RDLOCK_GUARD() { + ret = bdrv_co_preadv(s->mirror_top_bs->backing, op->offset, op->bytes, + &op->qiov, 0); + } mirror_read_complete(op, ret); } @@ -XXX,XX +XXX,XX @@ static void coroutine_fn active_write_settle(MirrorOp *op) g_free(op); } -static int coroutine_fn bdrv_mirror_top_preadv(BlockDriverState *bs, - int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +bdrv_mirror_top_preadv(BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, BdrvRequestFlags flags) { return bdrv_co_preadv(bs->backing, offset, bytes, qiov, flags); } @@ -XXX,XX +XXX,XX @@ out: return ret; } -static int coroutine_fn bdrv_mirror_top_pwritev(BlockDriverState *bs, - int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +bdrv_mirror_top_pwritev(BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, BdrvRequestFlags flags) { MirrorBDSOpaque *s = bs->opaque; QEMUIOVector bounce_qiov; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_mirror_top_pwritev(BlockDriverState *bs, int ret = 0; bool copy_to_target = false; - assume_graph_lock(); /* FIXME */ - if (s->job) { copy_to_target = s->job->ret >= 0 && !job_is_cancelled(&s->job->common.job) && 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 @@ allocate_clusters(BlockDriverState *bs, int64_t sector_num, } -static coroutine_fn int parallels_co_flush_to_os(BlockDriverState *bs) +static int coroutine_fn GRAPH_RDLOCK +parallels_co_flush_to_os(BlockDriverState *bs) { BDRVParallelsState *s = bs->opaque; unsigned long size = DIV_ROUND_UP(s->header_size, s->bat_dirty_block); @@ -XXX,XX +XXX,XX @@ parallels_co_writev(BlockDriverState *bs, int64_t sector_num, int nb_sectors, return ret; } -static coroutine_fn int parallels_co_readv(BlockDriverState *bs, - int64_t sector_num, int nb_sectors, QEMUIOVector *qiov) +static int coroutine_fn GRAPH_RDLOCK +parallels_co_readv(BlockDriverState *bs, int64_t sector_num, int nb_sectors, + QEMUIOVector *qiov) { BDRVParallelsState *s = bs->opaque; uint64_t bytes_done = 0; diff --git a/block/preallocate.c b/block/preallocate.c index XXXXXXX..XXXXXXX 100644 --- a/block/preallocate.c +++ b/block/preallocate.c @@ -XXX,XX +XXX,XX @@ static void preallocate_reopen_abort(BDRVReopenState *state) state->opaque = NULL; } -static coroutine_fn int preallocate_co_preadv_part( - BlockDriverState *bs, int64_t offset, int64_t bytes, - QEMUIOVector *qiov, size_t qiov_offset, BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +preallocate_co_preadv_part(BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, size_t qiov_offset, + BdrvRequestFlags flags) { return bdrv_co_preadv_part(bs->file, offset, bytes, qiov, qiov_offset, flags); @@ -XXX,XX +XXX,XX @@ preallocate_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset, return bdrv_co_pwrite_zeroes(bs->file, offset, bytes, flags); } -static coroutine_fn int preallocate_co_pwritev_part(BlockDriverState *bs, - int64_t offset, - int64_t bytes, - QEMUIOVector *qiov, - size_t qiov_offset, - BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +preallocate_co_pwritev_part(BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, size_t qiov_offset, + BdrvRequestFlags flags) { - assume_graph_lock(); /* FIXME */ handle_write(bs, offset, bytes, false); return bdrv_co_pwritev_part(bs->file, offset, bytes, qiov, qiov_offset, 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 @@ typedef struct BDRVQcowState { static QemuOptsList qcow_create_opts; -static int coroutine_fn decompress_cluster(BlockDriverState *bs, - uint64_t cluster_offset); +static int coroutine_fn GRAPH_RDLOCK +decompress_cluster(BlockDriverState *bs, uint64_t cluster_offset); static int qcow_probe(const uint8_t *buf, int buf_size, const char *filename) { @@ -XXX,XX +XXX,XX @@ static int decompress_buffer(uint8_t *out_buf, int out_buf_size, return 0; } -static int coroutine_fn decompress_cluster(BlockDriverState *bs, - uint64_t cluster_offset) +static int coroutine_fn GRAPH_RDLOCK +decompress_cluster(BlockDriverState *bs, uint64_t cluster_offset) { BDRVQcowState *s = bs->opaque; int ret, csize; diff --git a/block/qcow2-cluster.c b/block/qcow2-cluster.c index XXXXXXX..XXXXXXX 100644 --- a/block/qcow2-cluster.c +++ b/block/qcow2-cluster.c @@ -XXX,XX +XXX,XX @@ do_perform_cow_read(BlockDriverState *bs, uint64_t src_cluster_offset, return 0; } -static int coroutine_fn do_perform_cow_write(BlockDriverState *bs, - uint64_t cluster_offset, - unsigned offset_in_cluster, - QEMUIOVector *qiov) +static int coroutine_fn GRAPH_RDLOCK +do_perform_cow_write(BlockDriverState *bs, uint64_t cluster_offset, + unsigned offset_in_cluster, QEMUIOVector *qiov) { BDRVQcow2State *s = bs->opaque; int 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 qcow2_add_check_result(BdrvCheckResult *out, } } -static int coroutine_fn qcow2_co_check_locked(BlockDriverState *bs, - BdrvCheckResult *result, - BdrvCheckMode fix) +static int coroutine_fn GRAPH_RDLOCK +qcow2_co_check_locked(BlockDriverState *bs, BdrvCheckResult *result, + BdrvCheckMode fix) { BdrvCheckResult snapshot_res = {}; BdrvCheckResult refcount_res = {}; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qcow2_co_check_locked(BlockDriverState *bs, return ret; } -static int coroutine_fn qcow2_co_check(BlockDriverState *bs, - BdrvCheckResult *result, - BdrvCheckMode fix) +static int coroutine_fn GRAPH_RDLOCK +qcow2_co_check(BlockDriverState *bs, BdrvCheckResult *result, + BdrvCheckMode fix) { BDRVQcow2State *s = bs->opaque; int ret; @@ -XXX,XX +XXX,XX @@ static int validate_compression_type(BDRVQcow2State *s, Error **errp) } /* Called with s->lock held. */ -static int coroutine_fn qcow2_do_open(BlockDriverState *bs, QDict *options, - int flags, bool open_data_file, - Error **errp) +static int coroutine_fn GRAPH_RDLOCK +qcow2_do_open(BlockDriverState *bs, QDict *options, int flags, + bool open_data_file, Error **errp) { ERRP_GUARD(); BDRVQcow2State *s = bs->opaque; @@ -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 */ + qemu_co_mutex_lock(&s->lock); qoc->ret = qcow2_do_open(qoc->bs, qoc->options, qoc->flags, true, qoc->errp); @@ -XXX,XX +XXX,XX @@ out: return ret; } -static coroutine_fn int +static int coroutine_fn GRAPH_RDLOCK qcow2_co_preadv_encrypted(BlockDriverState *bs, uint64_t host_offset, uint64_t offset, @@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow2_add_task(BlockDriverState *bs, return 0; } -static coroutine_fn int qcow2_co_preadv_task(BlockDriverState *bs, - QCow2SubclusterType subc_type, - uint64_t host_offset, - uint64_t offset, uint64_t bytes, - QEMUIOVector *qiov, - size_t qiov_offset) +static int coroutine_fn GRAPH_RDLOCK +qcow2_co_preadv_task(BlockDriverState *bs, QCow2SubclusterType subc_type, + uint64_t host_offset, uint64_t offset, uint64_t bytes, + QEMUIOVector *qiov, size_t qiov_offset) { BDRVQcow2State *s = bs->opaque; @@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow2_co_preadv_task(BlockDriverState *bs, g_assert_not_reached(); } -static coroutine_fn int qcow2_co_preadv_task_entry(AioTask *task) +/* + * This function can count as GRAPH_RDLOCK because qcow2_co_preadv_part() holds + * the graph lock and keeps it until this coroutine has terminated. + */ +static int coroutine_fn GRAPH_RDLOCK qcow2_co_preadv_task_entry(AioTask *task) { Qcow2AioTask *t = container_of(task, Qcow2AioTask, task); @@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow2_co_preadv_task_entry(AioTask *task) t->qiov, t->qiov_offset); } -static coroutine_fn int qcow2_co_preadv_part(BlockDriverState *bs, - int64_t offset, int64_t bytes, - QEMUIOVector *qiov, - size_t qiov_offset, - BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +qcow2_co_preadv_part(BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, size_t qiov_offset, + BdrvRequestFlags flags) { BDRVQcow2State *s = bs->opaque; int ret = 0; @@ -XXX,XX +XXX,XX @@ static void qcow2_close(BlockDriverState *bs) qcow2_do_close(bs, true); } -static void coroutine_fn qcow2_co_invalidate_cache(BlockDriverState *bs, - Error **errp) +static void coroutine_fn GRAPH_RDLOCK +qcow2_co_invalidate_cache(BlockDriverState *bs, Error **errp) { ERRP_GUARD(); BDRVQcow2State *s = bs->opaque; @@ -XXX,XX +XXX,XX @@ qcow2_co_pwritev_compressed_part(BlockDriverState *bs, return ret; } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK qcow2_co_preadv_compressed(BlockDriverState *bs, uint64_t l2_entry, uint64_t offset, diff --git a/block/qed-table.c b/block/qed-table.c index XXXXXXX..XXXXXXX 100644 --- a/block/qed-table.c +++ b/block/qed-table.c @@ -XXX,XX +XXX,XX @@ #include "qemu/memalign.h" /* Called with table_lock held. */ -static int coroutine_fn qed_read_table(BDRVQEDState *s, uint64_t offset, - QEDTable *table) +static int coroutine_fn GRAPH_RDLOCK +qed_read_table(BDRVQEDState *s, uint64_t offset, QEDTable *table) { unsigned int bytes = s->header.cluster_size * s->header.table_size; 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 @@ int qed_write_header_sync(BDRVQEDState *s) * * No new allocating reqs can start while this function runs. */ -static int coroutine_fn qed_write_header(BDRVQEDState *s) +static int coroutine_fn GRAPH_RDLOCK qed_write_header(BDRVQEDState *s) { /* We must write full sectors for O_DIRECT but cannot necessarily generate * the data following the header if an unrecognized compat feature is @@ -XXX,XX +XXX,XX @@ fail: return ret; } -static int coroutine_fn bdrv_qed_co_block_status(BlockDriverState *bs, - bool want_zero, - int64_t pos, int64_t bytes, - int64_t *pnum, int64_t *map, - BlockDriverState **file) +static int coroutine_fn GRAPH_RDLOCK +bdrv_qed_co_block_status(BlockDriverState *bs, bool want_zero, int64_t pos, + int64_t bytes, int64_t *pnum, int64_t *map, + BlockDriverState **file) { BDRVQEDState *s = bs->opaque; size_t len = MIN(bytes, SIZE_MAX); @@ -XXX,XX +XXX,XX @@ static BDRVQEDState *acb_to_s(QEDAIOCB *acb) * This function reads qiov->size bytes starting at pos from the backing file. * If there is no backing file then zeroes are read. */ -static int coroutine_fn qed_read_backing_file(BDRVQEDState *s, uint64_t pos, - QEMUIOVector *qiov) +static int coroutine_fn GRAPH_RDLOCK +qed_read_backing_file(BDRVQEDState *s, uint64_t pos, QEMUIOVector *qiov) { if (s->bs->backing) { BLKDBG_EVENT(s->bs->file, BLKDBG_READ_BACKING_AIO); @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qed_read_backing_file(BDRVQEDState *s, uint64_t pos, * @len: Number of bytes * @offset: Byte offset in image file */ -static int coroutine_fn qed_copy_from_backing_file(BDRVQEDState *s, - uint64_t pos, uint64_t len, - uint64_t offset) +static int coroutine_fn GRAPH_RDLOCK +qed_copy_from_backing_file(BDRVQEDState *s, uint64_t pos, uint64_t len, + uint64_t offset) { QEMUIOVector qiov; int ret; @@ -XXX,XX +XXX,XX @@ qed_aio_write_l2_update(QEDAIOCB *acb, uint64_t offset) * * Called with table_lock *not* held. */ -static int coroutine_fn qed_aio_write_main(QEDAIOCB *acb) +static int coroutine_fn GRAPH_RDLOCK qed_aio_write_main(QEDAIOCB *acb) { BDRVQEDState *s = acb_to_s(acb); uint64_t offset = acb->cur_cluster + @@ -XXX,XX +XXX,XX @@ qed_aio_write_alloc(QEDAIOCB *acb, size_t len) * * Called with table_lock held. */ -static int coroutine_fn qed_aio_write_inplace(QEDAIOCB *acb, uint64_t offset, - size_t len) +static int coroutine_fn GRAPH_RDLOCK +qed_aio_write_inplace(QEDAIOCB *acb, uint64_t offset, size_t len) { BDRVQEDState *s = acb_to_s(acb); int r; @@ -XXX,XX +XXX,XX @@ qed_aio_write_data(void *opaque, int ret, uint64_t offset, size_t len) * * Called with table_lock held. */ -static int coroutine_fn qed_aio_read_data(void *opaque, int ret, - uint64_t offset, size_t len) +static int coroutine_fn GRAPH_RDLOCK +qed_aio_read_data(void *opaque, int ret, uint64_t offset, size_t len) { QEDAIOCB *acb = opaque; BDRVQEDState *s = acb_to_s(acb); 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 void quorum_report_bad_versions(BDRVQuorumState *s, } } -static void coroutine_fn quorum_rewrite_entry(void *opaque) +/* + * This function can count as GRAPH_RDLOCK because read_quorum_children() holds + * the graph lock and keeps it until this coroutine has terminated. + */ +static void coroutine_fn GRAPH_RDLOCK quorum_rewrite_entry(void *opaque) { QuorumCo *co = opaque; QuorumAIOCB *acb = co->acb; @@ -XXX,XX +XXX,XX @@ static void coroutine_fn quorum_rewrite_entry(void *opaque) } } -static bool quorum_rewrite_bad_versions(QuorumAIOCB *acb, - QuorumVoteValue *value) +static bool coroutine_fn GRAPH_RDLOCK +quorum_rewrite_bad_versions(QuorumAIOCB *acb, QuorumVoteValue *value) { QuorumVoteVersion *version; QuorumVoteItem *item; @@ -XXX,XX +XXX,XX @@ static int quorum_vote_error(QuorumAIOCB *acb) return ret; } -static void quorum_vote(QuorumAIOCB *acb) +static void coroutine_fn GRAPH_RDLOCK quorum_vote(QuorumAIOCB *acb) { bool quorum = true; int i, j, ret; @@ -XXX,XX +XXX,XX @@ free_exit: quorum_free_vote_list(&acb->votes); } -static void coroutine_fn read_quorum_children_entry(void *opaque) +/* + * This function can count as GRAPH_RDLOCK because read_quorum_children() holds + * the graph lock and keeps it until this coroutine has terminated. + */ +static void coroutine_fn GRAPH_RDLOCK read_quorum_children_entry(void *opaque) { QuorumCo *co = opaque; QuorumAIOCB *acb = co->acb; @@ -XXX,XX +XXX,XX @@ static void coroutine_fn read_quorum_children_entry(void *opaque) } } -static int coroutine_fn read_quorum_children(QuorumAIOCB *acb) +static int coroutine_fn GRAPH_RDLOCK read_quorum_children(QuorumAIOCB *acb) { BDRVQuorumState *s = acb->bs->opaque; int i; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn read_quorum_children(QuorumAIOCB *acb) return acb->vote_ret; } -static int coroutine_fn read_fifo_child(QuorumAIOCB *acb) +static int coroutine_fn GRAPH_RDLOCK read_fifo_child(QuorumAIOCB *acb) { BDRVQuorumState *s = acb->bs->opaque; int n, ret; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn read_fifo_child(QuorumAIOCB *acb) return ret; } -static int coroutine_fn quorum_co_preadv(BlockDriverState *bs, - int64_t offset, int64_t bytes, - QEMUIOVector *qiov, - BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +quorum_co_preadv(BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, BdrvRequestFlags flags) { BDRVQuorumState *s = bs->opaque; QuorumAIOCB *acb = quorum_aio_get(bs, qiov, offset, bytes, flags); 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 inline int raw_adjust_offset(BlockDriverState *bs, int64_t *offset, return 0; } -static int coroutine_fn raw_co_preadv(BlockDriverState *bs, int64_t offset, - int64_t bytes, QEMUIOVector *qiov, - BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +raw_co_preadv(BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, BdrvRequestFlags flags) { int ret; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn raw_co_preadv(BlockDriverState *bs, int64_t offset, return bdrv_co_preadv(bs->file, offset, bytes, qiov, flags); } -static int coroutine_fn raw_co_pwritev(BlockDriverState *bs, int64_t offset, - int64_t bytes, QEMUIOVector *qiov, - BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +raw_co_pwritev(BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, BdrvRequestFlags flags) { void *buf = NULL; BlockDriver *drv; diff --git a/block/replication.c b/block/replication.c index XXXXXXX..XXXXXXX 100644 --- a/block/replication.c +++ b/block/replication.c @@ -XXX,XX +XXX,XX @@ static int replication_return_value(BDRVReplicationState *s, int ret) return ret; } -static coroutine_fn int replication_co_readv(BlockDriverState *bs, - int64_t sector_num, - int remaining_sectors, - QEMUIOVector *qiov) +static int coroutine_fn GRAPH_RDLOCK +replication_co_readv(BlockDriverState *bs, int64_t sector_num, + int remaining_sectors, QEMUIOVector *qiov) { BDRVReplicationState *s = bs->opaque; int ret; @@ -XXX,XX +XXX,XX @@ static coroutine_fn int replication_co_readv(BlockDriverState *bs, return replication_return_value(s, ret); } -static coroutine_fn int replication_co_writev(BlockDriverState *bs, - int64_t sector_num, - int remaining_sectors, - QEMUIOVector *qiov, - int flags) +static int coroutine_fn GRAPH_RDLOCK +replication_co_writev(BlockDriverState *bs, int64_t sector_num, + int remaining_sectors, QEMUIOVector *qiov, int flags) { BDRVReplicationState *s = bs->opaque; QEMUIOVector hd_qiov; diff --git a/block/throttle.c b/block/throttle.c index XXXXXXX..XXXXXXX 100644 --- a/block/throttle.c +++ b/block/throttle.c @@ -XXX,XX +XXX,XX @@ static int64_t coroutine_fn throttle_co_getlength(BlockDriverState *bs) return bdrv_co_getlength(bs->file->bs); } -static int coroutine_fn throttle_co_preadv(BlockDriverState *bs, - int64_t offset, int64_t bytes, - QEMUIOVector *qiov, - BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +throttle_co_preadv(BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, BdrvRequestFlags flags) { ThrottleGroupMember *tgm = bs->opaque; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn throttle_co_preadv(BlockDriverState *bs, return bdrv_co_preadv(bs->file, offset, bytes, qiov, flags); } -static int coroutine_fn throttle_co_pwritev(BlockDriverState *bs, - int64_t offset, int64_t bytes, - QEMUIOVector *qiov, - BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +throttle_co_pwritev(BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, BdrvRequestFlags flags) { ThrottleGroupMember *tgm = bs->opaque; throttle_group_co_io_limits_intercept(tgm, bytes, true); @@ -XXX,XX +XXX,XX @@ throttle_co_pdiscard(BlockDriverState *bs, int64_t offset, int64_t bytes) return bdrv_co_pdiscard(bs->file, offset, bytes); } -static int coroutine_fn throttle_co_pwritev_compressed(BlockDriverState *bs, - int64_t offset, - int64_t bytes, - QEMUIOVector *qiov) +static int coroutine_fn GRAPH_RDLOCK +throttle_co_pwritev_compressed(BlockDriverState *bs, int64_t offset, + int64_t bytes, QEMUIOVector *qiov) { return throttle_co_pwritev(bs, offset, bytes, qiov, BDRV_REQ_WRITE_COMPRESSED); 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_block_status(BlockDriverState *bs, (s->header.image_type == VDI_TYPE_STATIC ? BDRV_BLOCK_RECURSE : 0); } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK vdi_co_preadv(BlockDriverState *bs, int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags) { @@ -XXX,XX +XXX,XX @@ vdi_co_preadv(BlockDriverState *bs, int64_t offset, int64_t bytes, return ret; } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK vdi_co_pwritev(BlockDriverState *bs, int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags) { 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 @@ vhdx_co_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) } -static coroutine_fn int vhdx_co_readv(BlockDriverState *bs, int64_t sector_num, - int nb_sectors, QEMUIOVector *qiov) +static int coroutine_fn GRAPH_RDLOCK +vhdx_co_readv(BlockDriverState *bs, int64_t sector_num, int nb_sectors, + QEMUIOVector *qiov) { BDRVVHDXState *s = bs->opaque; int ret = 0; @@ -XXX,XX +XXX,XX @@ int vhdx_user_visible_write(BlockDriverState *bs, BDRVVHDXState *s) return ret; } -static coroutine_fn int vhdx_co_writev(BlockDriverState *bs, int64_t sector_num, - int nb_sectors, QEMUIOVector *qiov, - int flags) +static int coroutine_fn GRAPH_RDLOCK +vhdx_co_writev(BlockDriverState *bs, int64_t sector_num, int nb_sectors, + QEMUIOVector *qiov, int flags) { int ret = -ENOTSUP; BDRVVHDXState *s = bs->opaque; 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_refresh_limits(BlockDriverState *bs, Error **errp) * [@skip_start_sector, @skip_end_sector) is not copied or written, and leave * it for call to write user data in the request. */ -static int coroutine_fn get_whole_cluster(BlockDriverState *bs, - VmdkExtent *extent, - uint64_t cluster_offset, - uint64_t offset, - uint64_t skip_start_bytes, - uint64_t skip_end_bytes, - bool zeroed) +static int coroutine_fn GRAPH_RDLOCK +get_whole_cluster(BlockDriverState *bs, VmdkExtent *extent, + uint64_t cluster_offset, uint64_t offset, + uint64_t skip_start_bytes, uint64_t skip_end_bytes, + bool zeroed) { int ret = VMDK_OK; int64_t cluster_bytes; @@ -XXX,XX +XXX,XX @@ vmdk_L2update(VmdkExtent *extent, VmdkMetaData *m_data, uint32_t offset) * VMDK_UNALLOC if cluster is not mapped and @allocate is false. * VMDK_ERROR if failed. */ -static int coroutine_fn get_cluster_offset(BlockDriverState *bs, - VmdkExtent *extent, - VmdkMetaData *m_data, - uint64_t offset, - bool allocate, - uint64_t *cluster_offset, - uint64_t skip_start_bytes, - uint64_t skip_end_bytes) +static int coroutine_fn GRAPH_RDLOCK +get_cluster_offset(BlockDriverState *bs, VmdkExtent *extent, + VmdkMetaData *m_data, uint64_t offset, bool allocate, + uint64_t *cluster_offset, uint64_t skip_start_bytes, + uint64_t skip_end_bytes) { unsigned int l1_index, l2_offset, l2_index; int min_index, i, j; @@ -XXX,XX +XXX,XX @@ static inline uint64_t vmdk_find_offset_in_cluster(VmdkExtent *extent, return extent_relative_offset % cluster_size; } -static int coroutine_fn vmdk_co_block_status(BlockDriverState *bs, - bool want_zero, - int64_t offset, int64_t bytes, - int64_t *pnum, int64_t *map, - BlockDriverState **file) +static int coroutine_fn GRAPH_RDLOCK +vmdk_co_block_status(BlockDriverState *bs, bool want_zero, + int64_t offset, int64_t bytes, int64_t *pnum, + int64_t *map, BlockDriverState **file) { BDRVVmdkState *s = bs->opaque; int64_t index_in_cluster, n, ret; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn vmdk_co_block_status(BlockDriverState *bs, return ret; } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK vmdk_write_extent(VmdkExtent *extent, int64_t cluster_offset, int64_t offset_in_cluster, QEMUIOVector *qiov, uint64_t qiov_offset, uint64_t n_bytes, @@ -XXX,XX +XXX,XX @@ vmdk_write_extent(VmdkExtent *extent, int64_t cluster_offset, return ret; } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK vmdk_read_extent(VmdkExtent *extent, int64_t cluster_offset, - int64_t offset_in_cluster, QEMUIOVector *qiov, - int bytes) + int64_t offset_in_cluster, QEMUIOVector *qiov, int bytes) { int ret; int cluster_bytes, buf_bytes; @@ -XXX,XX +XXX,XX @@ vmdk_read_extent(VmdkExtent *extent, int64_t cluster_offset, return ret; } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK vmdk_co_preadv(BlockDriverState *bs, int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags) { @@ -XXX,XX +XXX,XX @@ fail: * * Returns: error code with 0 for success. */ -static int coroutine_fn vmdk_pwritev(BlockDriverState *bs, uint64_t offset, - uint64_t bytes, QEMUIOVector *qiov, - bool zeroed, bool zero_dry_run) +static int coroutine_fn GRAPH_RDLOCK +vmdk_pwritev(BlockDriverState *bs, uint64_t offset, uint64_t bytes, + QEMUIOVector *qiov, bool zeroed, bool zero_dry_run) { BDRVVmdkState *s = bs->opaque; VmdkExtent *extent = NULL; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn vmdk_pwritev(BlockDriverState *bs, uint64_t offset, uint64_t bytes_done = 0; VmdkMetaData m_data; - assume_graph_lock(); /* FIXME */ - if (DIV_ROUND_UP(offset, BDRV_SECTOR_SIZE) > bs->total_sectors) { error_report("Wrong offset: offset=0x%" PRIx64 " total_sectors=0x%" PRIx64, @@ -XXX,XX +XXX,XX @@ static int coroutine_fn vmdk_pwritev(BlockDriverState *bs, uint64_t offset, return 0; } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK vmdk_co_pwritev(BlockDriverState *bs, int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags) { @@ -XXX,XX +XXX,XX @@ vmdk_co_pwritev_compressed(BlockDriverState *bs, int64_t offset, int64_t bytes, return vmdk_co_pwritev(bs, offset, bytes, qiov, 0); } -static int coroutine_fn vmdk_co_pwrite_zeroes(BlockDriverState *bs, - int64_t offset, - int64_t bytes, - BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +vmdk_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset, int64_t bytes, + BdrvRequestFlags flags) { int ret; BDRVVmdkState *s = bs->opaque; @@ -XXX,XX +XXX,XX @@ static VmdkExtentInfo *vmdk_get_extent_info(VmdkExtent *extent) return info; } -static int coroutine_fn vmdk_co_check(BlockDriverState *bs, - BdrvCheckResult *result, - BdrvCheckMode fix) +static int coroutine_fn GRAPH_RDLOCK +vmdk_co_check(BlockDriverState *bs, BdrvCheckResult *result, BdrvCheckMode fix) { BDRVVmdkState *s = bs->opaque; VmdkExtent *extent = 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 @@ vpc_co_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) return 0; } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK vpc_co_preadv(BlockDriverState *bs, int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags) { @@ -XXX,XX +XXX,XX @@ fail: return ret; } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK vpc_co_pwritev(BlockDriverState *bs, int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags) { 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 bdrv_test_top_close(BlockDriverState *bs) } } -static int coroutine_fn bdrv_test_top_co_preadv(BlockDriverState *bs, - int64_t offset, int64_t bytes, - QEMUIOVector *qiov, - BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +bdrv_test_top_co_preadv(BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, BdrvRequestFlags flags) { BDRVTestTopState *tts = bs->opaque; return bdrv_co_preadv(tts->wait_child, offset, bytes, qiov, flags); @@ -XXX,XX +XXX,XX @@ static void coroutine_fn test_co_delete_by_drain(void *opaque) void *buffer = g_malloc(65536); QEMUIOVector qiov = QEMU_IOVEC_INIT_BUF(qiov, buffer, 65536); + GRAPH_RDLOCK_GUARD(); + /* Pretend some internal write operation from parent to child. * Important: We have to read from the child, not from the parent! * Draining works by first propagating it all up the tree to the @@ -XXX,XX +XXX,XX @@ static void bdrv_replace_test_close(BlockDriverState *bs) * Otherwise: * Set .has_read to true and return success. */ -static int coroutine_fn bdrv_replace_test_co_preadv(BlockDriverState *bs, - int64_t offset, - int64_t bytes, - QEMUIOVector *qiov, - BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +bdrv_replace_test_co_preadv(BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, BdrvRequestFlags flags) { BDRVReplaceTestState *s = bs->opaque; @@ -XXX,XX +XXX,XX @@ static void coroutine_fn bdrv_replace_test_read_entry(void *opaque) int ret; /* Queue a read request post-drain */ + bdrv_graph_co_rdlock(); ret = bdrv_replace_test_co_preadv(bs, 0, 1, &qiov, 0); + bdrv_graph_co_rdunlock(); + g_assert(ret >= 0); bdrv_dec_in_flight(bs); } -- 2.39.2 This adds GRAPH_RDLOCK annotations to declare that callers of bdrv_co_pwrite_sync() need to hold a reader lock for the graph. For some places, we know that they will hold the lock, but we don't have the GRAPH_RDLOCK annotations yet. In this case, add assume_graph_lock() with a FIXME comment. These places will be removed once everything is properly annotated. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230203152202.49054-13-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/qcow2.h | 2 +- include/block/block-io.h | 7 ++++--- block/io.c | 3 +-- 3 files changed, 6 insertions(+), 6 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_inc_refcounts_imrt(BlockDriverState *bs, BdrvCheckResult *res, int qcow2_change_refcount_order(BlockDriverState *bs, int refcount_order, BlockDriverAmendStatusCB *status_cb, void *cb_opaque, Error **errp); -int coroutine_fn qcow2_shrink_reftable(BlockDriverState *bs); +int coroutine_fn GRAPH_RDLOCK qcow2_shrink_reftable(BlockDriverState *bs); int64_t qcow2_get_last_cluster(BlockDriverState *bs, int64_t size); int coroutine_fn qcow2_detect_metadata_preallocation(BlockDriverState *bs); 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 @@ int co_wrapper_mixed_bdrv_rdlock bdrv_pwrite_sync(BdrvChild *child, int64_t offset, int64_t bytes, const void *buf, BdrvRequestFlags flags); -int coroutine_fn bdrv_co_pwrite_sync(BdrvChild *child, int64_t offset, - int64_t bytes, const void *buf, - BdrvRequestFlags flags); +int coroutine_fn GRAPH_RDLOCK +bdrv_co_pwrite_sync(BdrvChild *child, int64_t offset, int64_t bytes, + const void *buf, BdrvRequestFlags flags); + /* * Efficiently zero a region of the disk image. Note that this is a regular * I/O request like read or write and should have a reasonable size. This 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 @@ int coroutine_fn bdrv_co_pwrite_sync(BdrvChild *child, int64_t offset, { int ret; IO_CODE(); - - assume_graph_lock(); /* FIXME */ + assert_bdrv_graph_readable(); ret = bdrv_co_pwrite(child, offset, bytes, buf, flags); if (ret < 0) { -- 2.39.2 All callers are already GRAPH_RDLOCK, so just add the annotation and remove assume_graph_lock(). Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230203152202.49054-14-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/io.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) 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 @@ fail: return ret; } -static int coroutine_fn bdrv_co_do_pwrite_zeroes(BlockDriverState *bs, - int64_t offset, int64_t bytes, BdrvRequestFlags flags) +static int coroutine_fn GRAPH_RDLOCK +bdrv_co_do_pwrite_zeroes(BlockDriverState *bs, int64_t offset, int64_t bytes, + BdrvRequestFlags flags) { BlockDriver *drv = bs->drv; QEMUIOVector qiov; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_pwrite_zeroes(BlockDriverState *bs, int head = 0; int tail = 0; - assume_graph_lock(); /* FIXME */ - int64_t max_write_zeroes = MIN_NON_ZERO(bs->bl.max_pwrite_zeroes, INT64_MAX); int alignment = MAX(bs->bl.pwrite_zeroes_alignment, -- 2.39.2 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> This adds GRAPH_RDLOCK annotations to declare that callers of bdrv_co_copy_range() 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> Message-Id: <20230203152202.49054-15-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-io.h | 9 +++++---- include/block/block_int-common.h | 24 ++++++++---------------- include/block/block_int-io.h | 20 ++++++++++---------- block/block-backend.c | 2 ++ block/file-posix.c | 16 +++++++--------- block/io.c | 7 +++++-- block/iscsi.c | 28 ++++++++++++---------------- block/qcow2.c | 5 ++--- block/raw-format.c | 28 ++++++++++++---------------- qemu-img.c | 4 +++- 10 files changed, 66 insertions(+), 77 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 @@ bool co_wrapper bdrv_can_store_new_dirty_bitmap(BlockDriverState *bs, * * Returns: 0 if succeeded; negative error code if failed. **/ -int coroutine_fn bdrv_co_copy_range(BdrvChild *src, int64_t src_offset, - BdrvChild *dst, int64_t dst_offset, - int64_t bytes, BdrvRequestFlags read_flags, - BdrvRequestFlags write_flags); +int coroutine_fn GRAPH_RDLOCK +bdrv_co_copy_range(BdrvChild *src, int64_t src_offset, + BdrvChild *dst, int64_t dst_offset, + int64_t bytes, BdrvRequestFlags read_flags, + BdrvRequestFlags write_flags); /* * "I/O or GS" API functions. These functions can run without 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 { * See the comment of bdrv_co_copy_range for the parameter and return value * semantics. */ - int coroutine_fn (*bdrv_co_copy_range_from)(BlockDriverState *bs, - BdrvChild *src, - int64_t offset, - BdrvChild *dst, - int64_t dst_offset, - int64_t bytes, - BdrvRequestFlags read_flags, - BdrvRequestFlags write_flags); + int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_copy_range_from)( + BlockDriverState *bs, BdrvChild *src, int64_t offset, + BdrvChild *dst, int64_t dst_offset, int64_t bytes, + BdrvRequestFlags read_flags, BdrvRequestFlags write_flags); /* * Map [offset, offset + nbytes) range onto a child of bs to copy data to, @@ -XXX,XX +XXX,XX @@ struct BlockDriver { * See the comment of bdrv_co_copy_range for the parameter and return value * semantics. */ - int coroutine_fn (*bdrv_co_copy_range_to)(BlockDriverState *bs, - BdrvChild *src, - int64_t src_offset, - BdrvChild *dst, - int64_t dst_offset, - int64_t bytes, - BdrvRequestFlags read_flags, - BdrvRequestFlags write_flags); + int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_copy_range_to)( + BlockDriverState *bs, BdrvChild *src, int64_t src_offset, + BdrvChild *dst, int64_t dst_offset, int64_t bytes, + BdrvRequestFlags read_flags, BdrvRequestFlags write_flags); /* * Building block for bdrv_block_status[_above] and diff --git a/include/block/block_int-io.h b/include/block/block_int-io.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block_int-io.h +++ b/include/block/block_int-io.h @@ -XXX,XX +XXX,XX @@ void bdrv_dirty_bitmap_merge_internal(BdrvDirtyBitmap *dest, void bdrv_inc_in_flight(BlockDriverState *bs); void bdrv_dec_in_flight(BlockDriverState *bs); -int coroutine_fn bdrv_co_copy_range_from(BdrvChild *src, int64_t src_offset, - BdrvChild *dst, int64_t dst_offset, - int64_t bytes, - BdrvRequestFlags read_flags, - BdrvRequestFlags write_flags); -int coroutine_fn bdrv_co_copy_range_to(BdrvChild *src, int64_t src_offset, - BdrvChild *dst, int64_t dst_offset, - int64_t bytes, - BdrvRequestFlags read_flags, - BdrvRequestFlags write_flags); +int coroutine_fn GRAPH_RDLOCK +bdrv_co_copy_range_from(BdrvChild *src, int64_t src_offset, + BdrvChild *dst, int64_t dst_offset, + int64_t bytes, BdrvRequestFlags read_flags, + BdrvRequestFlags write_flags); +int coroutine_fn GRAPH_RDLOCK +bdrv_co_copy_range_to(BdrvChild *src, int64_t src_offset, + BdrvChild *dst, int64_t dst_offset, + int64_t bytes, BdrvRequestFlags read_flags, + BdrvRequestFlags write_flags); int coroutine_fn bdrv_co_refresh_total_sectors(BlockDriverState *bs, int64_t hint); 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 @@ int coroutine_fn blk_co_copy_range(BlockBackend *blk_in, int64_t off_in, if (r) { return r; } + + GRAPH_RDLOCK_GUARD(); return bdrv_co_copy_range(blk_in->root, off_in, blk_out->root, off_out, bytes, read_flags, write_flags); diff --git a/block/file-posix.c b/block/file-posix.c index XXXXXXX..XXXXXXX 100644 --- a/block/file-posix.c +++ b/block/file-posix.c @@ -XXX,XX +XXX,XX @@ static void raw_abort_perm_update(BlockDriverState *bs) raw_handle_perm_lock(bs, RAW_PL_ABORT, 0, 0, NULL); } -static int coroutine_fn raw_co_copy_range_from( +static int coroutine_fn GRAPH_RDLOCK raw_co_copy_range_from( BlockDriverState *bs, BdrvChild *src, int64_t src_offset, BdrvChild *dst, int64_t dst_offset, int64_t bytes, BdrvRequestFlags read_flags, BdrvRequestFlags write_flags) @@ -XXX,XX +XXX,XX @@ static int coroutine_fn raw_co_copy_range_from( read_flags, write_flags); } -static int coroutine_fn raw_co_copy_range_to(BlockDriverState *bs, - BdrvChild *src, - int64_t src_offset, - BdrvChild *dst, - int64_t dst_offset, - int64_t bytes, - BdrvRequestFlags read_flags, - BdrvRequestFlags write_flags) +static int coroutine_fn GRAPH_RDLOCK +raw_co_copy_range_to(BlockDriverState *bs, + BdrvChild *src, int64_t src_offset, + BdrvChild *dst, int64_t dst_offset, + int64_t bytes, BdrvRequestFlags read_flags, + BdrvRequestFlags write_flags) { RawPosixAIOData acb; BDRVRawState *s = bs->opaque; 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 @@ static int coroutine_fn GRAPH_RDLOCK bdrv_co_copy_range_internal( { BdrvTrackedRequest req; int ret; + assert_bdrv_graph_readable(); /* TODO We can support BDRV_REQ_NO_FALLBACK here */ assert(!(read_flags & BDRV_REQ_NO_FALLBACK)); @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_copy_range_from(BdrvChild *src, int64_t src_offset, BdrvRequestFlags write_flags) { IO_CODE(); - assume_graph_lock(); /* FIXME */ + assert_bdrv_graph_readable(); trace_bdrv_co_copy_range_from(src, src_offset, dst, dst_offset, bytes, read_flags, write_flags); return bdrv_co_copy_range_internal(src, src_offset, dst, dst_offset, @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_copy_range_to(BdrvChild *src, int64_t src_offset, BdrvRequestFlags write_flags) { IO_CODE(); - assume_graph_lock(); /* FIXME */ + assert_bdrv_graph_readable(); trace_bdrv_co_copy_range_to(src, src_offset, dst, dst_offset, bytes, read_flags, write_flags); return bdrv_co_copy_range_internal(src, src_offset, dst, dst_offset, @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_copy_range(BdrvChild *src, int64_t src_offset, BdrvRequestFlags write_flags) { IO_CODE(); + assert_bdrv_graph_readable(); + return bdrv_co_copy_range_from(src, src_offset, dst, dst_offset, bytes, read_flags, write_flags); diff --git a/block/iscsi.c b/block/iscsi.c index XXXXXXX..XXXXXXX 100644 --- a/block/iscsi.c +++ b/block/iscsi.c @@ -XXX,XX +XXX,XX @@ static void coroutine_fn iscsi_co_invalidate_cache(BlockDriverState *bs, iscsi_allocmap_invalidate(iscsilun); } -static int coroutine_fn iscsi_co_copy_range_from(BlockDriverState *bs, - BdrvChild *src, - int64_t src_offset, - BdrvChild *dst, - int64_t dst_offset, - int64_t bytes, - BdrvRequestFlags read_flags, - BdrvRequestFlags write_flags) +static int coroutine_fn GRAPH_RDLOCK +iscsi_co_copy_range_from(BlockDriverState *bs, + BdrvChild *src, int64_t src_offset, + BdrvChild *dst, int64_t dst_offset, + int64_t bytes, BdrvRequestFlags read_flags, + BdrvRequestFlags write_flags) { return bdrv_co_copy_range_to(src, src_offset, dst, dst_offset, bytes, read_flags, write_flags); @@ -XXX,XX +XXX,XX @@ static void iscsi_xcopy_data(struct iscsi_data *data, src_lba, dst_lba); } -static int coroutine_fn iscsi_co_copy_range_to(BlockDriverState *bs, - BdrvChild *src, - int64_t src_offset, - BdrvChild *dst, - int64_t dst_offset, - int64_t bytes, - BdrvRequestFlags read_flags, - BdrvRequestFlags write_flags) +static int coroutine_fn GRAPH_RDLOCK +iscsi_co_copy_range_to(BlockDriverState *bs, + BdrvChild *src, int64_t src_offset, + BdrvChild *dst, int64_t dst_offset, + int64_t bytes, BdrvRequestFlags read_flags, + BdrvRequestFlags write_flags) { IscsiLun *dst_lun = dst->bs->opaque; IscsiLun *src_lun; 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 coroutine_fn int qcow2_co_pdiscard(BlockDriverState *bs, return ret; } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK qcow2_co_copy_range_from(BlockDriverState *bs, BdrvChild *src, int64_t src_offset, BdrvChild *dst, int64_t dst_offset, @@ -XXX,XX +XXX,XX @@ out: return ret; } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK qcow2_co_copy_range_to(BlockDriverState *bs, BdrvChild *src, int64_t src_offset, BdrvChild *dst, int64_t dst_offset, @@ -XXX,XX +XXX,XX @@ qcow2_co_copy_range_to(BlockDriverState *bs, uint64_t host_offset; QCowL2Meta *l2meta = NULL; - assume_graph_lock(); /* FIXME */ assert(!bs->encrypted); qemu_co_mutex_lock(&s->lock); 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 int raw_probe_geometry(BlockDriverState *bs, HDGeometry *geo) return bdrv_probe_geometry(bs->file->bs, geo); } -static int coroutine_fn raw_co_copy_range_from(BlockDriverState *bs, - BdrvChild *src, - int64_t src_offset, - BdrvChild *dst, - int64_t dst_offset, - int64_t bytes, - BdrvRequestFlags read_flags, - BdrvRequestFlags write_flags) +static int coroutine_fn GRAPH_RDLOCK +raw_co_copy_range_from(BlockDriverState *bs, + BdrvChild *src, int64_t src_offset, + BdrvChild *dst, int64_t dst_offset, + int64_t bytes, BdrvRequestFlags read_flags, + BdrvRequestFlags write_flags) { int ret; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn raw_co_copy_range_from(BlockDriverState *bs, bytes, read_flags, write_flags); } -static int coroutine_fn raw_co_copy_range_to(BlockDriverState *bs, - BdrvChild *src, - int64_t src_offset, - BdrvChild *dst, - int64_t dst_offset, - int64_t bytes, - BdrvRequestFlags read_flags, - BdrvRequestFlags write_flags) +static int coroutine_fn GRAPH_RDLOCK +raw_co_copy_range_to(BlockDriverState *bs, + BdrvChild *src, int64_t src_offset, + BdrvChild *dst, int64_t dst_offset, + int64_t bytes, BdrvRequestFlags read_flags, + BdrvRequestFlags write_flags) { int ret; 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 @@ retry: if (s->ret == -EINPROGRESS) { if (copy_range) { - ret = convert_co_copy_range(s, sector_num, n); + WITH_GRAPH_RDLOCK_GUARD() { + ret = convert_co_copy_range(s, sector_num, n); + } if (ret) { s->copy_range = false; goto retry; -- 2.39.2 Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230203152202.49054-16-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block_int-common.h | 12 +++++++----- include/block/block_int-io.h | 8 ++++---- block/copy-before-write.c | 6 ++---- block/io.c | 2 ++ block/snapshot-access.c | 4 ++-- 5 files changed, 17 insertions(+), 15 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 { * - receive the snapshot's actual length (which may differ from bs's * length) */ - int coroutine_fn (*bdrv_co_preadv_snapshot)(BlockDriverState *bs, - int64_t offset, int64_t bytes, QEMUIOVector *qiov, size_t qiov_offset); - int coroutine_fn (*bdrv_co_snapshot_block_status)(BlockDriverState *bs, - bool want_zero, int64_t offset, int64_t bytes, int64_t *pnum, - int64_t *map, BlockDriverState **file); + int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_preadv_snapshot)( + BlockDriverState *bs, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, size_t qiov_offset); + + int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_snapshot_block_status)( + BlockDriverState *bs, bool want_zero, int64_t offset, int64_t bytes, + int64_t *pnum, int64_t *map, BlockDriverState **file); int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_pdiscard_snapshot)( BlockDriverState *bs, int64_t offset, int64_t bytes); diff --git a/include/block/block_int-io.h b/include/block/block_int-io.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block_int-io.h +++ b/include/block/block_int-io.h @@ -XXX,XX +XXX,XX @@ * the I/O API. */ -int coroutine_fn bdrv_co_preadv_snapshot(BdrvChild *child, +int coroutine_fn GRAPH_RDLOCK bdrv_co_preadv_snapshot(BdrvChild *child, int64_t offset, int64_t bytes, QEMUIOVector *qiov, size_t qiov_offset); -int coroutine_fn bdrv_co_snapshot_block_status(BlockDriverState *bs, - bool want_zero, int64_t offset, int64_t bytes, int64_t *pnum, - int64_t *map, BlockDriverState **file); +int coroutine_fn GRAPH_RDLOCK bdrv_co_snapshot_block_status( + BlockDriverState *bs, bool want_zero, int64_t offset, int64_t bytes, + int64_t *pnum, int64_t *map, BlockDriverState **file); int coroutine_fn GRAPH_RDLOCK bdrv_co_pdiscard_snapshot(BlockDriverState *bs, int64_t offset, int64_t bytes); diff --git a/block/copy-before-write.c b/block/copy-before-write.c index XXXXXXX..XXXXXXX 100644 --- a/block/copy-before-write.c +++ b/block/copy-before-write.c @@ -XXX,XX +XXX,XX @@ cbw_snapshot_read_unlock(BlockDriverState *bs, BlockReq *req) g_free(req); } -static coroutine_fn int +static int coroutine_fn GRAPH_RDLOCK cbw_co_preadv_snapshot(BlockDriverState *bs, int64_t offset, int64_t bytes, QEMUIOVector *qiov, size_t qiov_offset) { @@ -XXX,XX +XXX,XX @@ cbw_co_preadv_snapshot(BlockDriverState *bs, int64_t offset, int64_t bytes, BdrvChild *file; int ret; - assume_graph_lock(); /* FIXME */ - /* TODO: upgrade to async loop using AioTask */ while (bytes) { int64_t cur_bytes; @@ -XXX,XX +XXX,XX @@ cbw_co_preadv_snapshot(BlockDriverState *bs, int64_t offset, int64_t bytes, return 0; } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK cbw_co_snapshot_block_status(BlockDriverState *bs, bool want_zero, int64_t offset, int64_t bytes, int64_t *pnum, int64_t *map, 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 @@ bdrv_co_preadv_snapshot(BdrvChild *child, int64_t offset, int64_t bytes, BlockDriver *drv = bs->drv; int ret; IO_CODE(); + assert_bdrv_graph_readable(); if (!drv) { return -ENOMEDIUM; @@ -XXX,XX +XXX,XX @@ bdrv_co_snapshot_block_status(BlockDriverState *bs, BlockDriver *drv = bs->drv; int ret; IO_CODE(); + assert_bdrv_graph_readable(); if (!drv) { return -ENOMEDIUM; diff --git a/block/snapshot-access.c b/block/snapshot-access.c index XXXXXXX..XXXXXXX 100644 --- a/block/snapshot-access.c +++ b/block/snapshot-access.c @@ -XXX,XX +XXX,XX @@ #include "qemu/cutils.h" #include "block/block_int.h" -static coroutine_fn int +static int coroutine_fn GRAPH_RDLOCK snapshot_access_co_preadv_part(BlockDriverState *bs, int64_t offset, int64_t bytes, QEMUIOVector *qiov, size_t qiov_offset, @@ -XXX,XX +XXX,XX @@ snapshot_access_co_preadv_part(BlockDriverState *bs, return bdrv_co_preadv_snapshot(bs->file, offset, bytes, qiov, qiov_offset); } -static int coroutine_fn +static int coroutine_fn GRAPH_RDLOCK snapshot_access_co_block_status(BlockDriverState *bs, bool want_zero, int64_t offset, int64_t bytes, int64_t *pnum, -- 2.39.2 This adds GRAPH_RDLOCK annotations to declare that callers of bdrv_co_create() 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> Message-Id: <20230203152202.49054-17-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-global-state.h | 14 ++++--- include/block/block_int-common.h | 11 +++--- block.c | 1 + block/create.c | 9 ++++- block/crypto.c | 7 ++-- block/file-posix.c | 7 ++-- block/file-win32.c | 7 ++-- block/parallels.c | 7 ++-- block/qcow.c | 6 +-- block/qcow2.c | 7 ++-- block/qed.c | 7 ++-- block/raw-format.c | 7 ++-- block/vdi.c | 7 ++-- block/vhdx.c | 7 ++-- block/vmdk.c | 63 ++++++++++++++---------------- block/vpc.c | 7 ++-- 16 files changed, 84 insertions(+), 90 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 @@ BlockDriver *bdrv_find_protocol(const char *filename, Error **errp); BlockDriver *bdrv_find_format(const char *format_name); -int coroutine_fn bdrv_co_create(BlockDriver *drv, const char *filename, - QemuOpts *opts, Error **errp); -int co_wrapper bdrv_create(BlockDriver *drv, const char *filename, - QemuOpts *opts, Error **errp); +int coroutine_fn GRAPH_RDLOCK +bdrv_co_create(BlockDriver *drv, const char *filename, QemuOpts *opts, + Error **errp); -int coroutine_fn bdrv_co_create_file(const char *filename, QemuOpts *opts, - Error **errp); +int co_wrapper_bdrv_rdlock bdrv_create(BlockDriver *drv, const char *filename, + QemuOpts *opts, Error **errp); + +int coroutine_fn GRAPH_RDLOCK +bdrv_co_create_file(const char *filename, QemuOpts *opts, Error **errp); BlockDriverState *bdrv_new(void); int bdrv_append(BlockDriverState *bs_new, BlockDriverState *bs_top, 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 { Error **errp); void (*bdrv_close)(BlockDriverState *bs); - int coroutine_fn (*bdrv_co_create)(BlockdevCreateOptions *opts, - Error **errp); - int coroutine_fn (*bdrv_co_create_opts)(BlockDriver *drv, - const char *filename, - QemuOpts *opts, - Error **errp); + int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_create)( + BlockdevCreateOptions *opts, Error **errp); + + int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_create_opts)( + BlockDriver *drv, const char *filename, QemuOpts *opts, Error **errp); int (*bdrv_amend_options)(BlockDriverState *bs, QemuOpts *opts, 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(BlockDriver *drv, const char *filename, int ret; GLOBAL_STATE_CODE(); ERRP_GUARD(); + assert_bdrv_graph_readable(); if (!drv->bdrv_co_create_opts) { error_setg(errp, "Driver '%s' does not support image creation", diff --git a/block/create.c b/block/create.c index XXXXXXX..XXXXXXX 100644 --- a/block/create.c +++ b/block/create.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn blockdev_create_run(Job *job, Error **errp) int ret; GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD(); job_progress_set_remaining(&s->common, 1); ret = s->drv->bdrv_co_create(s->opts, errp); @@ -XXX,XX +XXX,XX @@ static const JobDriver blockdev_create_job_driver = { .run = blockdev_create_run, }; +/* Checking whether the function is present doesn't require the graph lock */ +static inline bool TSA_NO_TSA has_bdrv_co_create(BlockDriver *drv) +{ + return drv->bdrv_co_create; +} + void qmp_blockdev_create(const char *job_id, BlockdevCreateOptions *options, Error **errp) { @@ -XXX,XX +XXX,XX @@ void qmp_blockdev_create(const char *job_id, BlockdevCreateOptions *options, } /* Error out if the driver doesn't support .bdrv_co_create */ - if (!drv->bdrv_co_create) { + if (!has_bdrv_co_create(drv)) { error_setg(errp, "Driver does not support blockdev-create"); return; } 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 block_crypto_co_create_opts_luks(BlockDriver *drv, - const char *filename, - QemuOpts *opts, - Error **errp) +static int coroutine_fn GRAPH_RDLOCK +block_crypto_co_create_opts_luks(BlockDriver *drv, const char *filename, + QemuOpts *opts, Error **errp) { QCryptoBlockCreateOptions *create_opts = NULL; BlockDriverState *bs = NULL; diff --git a/block/file-posix.c b/block/file-posix.c index XXXXXXX..XXXXXXX 100644 --- a/block/file-posix.c +++ b/block/file-posix.c @@ -XXX,XX +XXX,XX @@ out: return result; } -static int coroutine_fn raw_co_create_opts(BlockDriver *drv, - const char *filename, - QemuOpts *opts, - Error **errp) +static int coroutine_fn GRAPH_RDLOCK +raw_co_create_opts(BlockDriver *drv, const char *filename, + QemuOpts *opts, Error **errp) { BlockdevCreateOptions options; int64_t total_size = 0; diff --git a/block/file-win32.c b/block/file-win32.c index XXXXXXX..XXXXXXX 100644 --- a/block/file-win32.c +++ b/block/file-win32.c @@ -XXX,XX +XXX,XX @@ static int raw_co_create(BlockdevCreateOptions *options, Error **errp) return 0; } -static int coroutine_fn raw_co_create_opts(BlockDriver *drv, - const char *filename, - QemuOpts *opts, - Error **errp) +static int coroutine_fn GRAPH_RDLOCK +raw_co_create_opts(BlockDriver *drv, const char *filename, + QemuOpts *opts, Error **errp) { BlockdevCreateOptions options; int64_t total_size = 0; 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 @@ exit: goto out; } -static int coroutine_fn parallels_co_create_opts(BlockDriver *drv, - const char *filename, - QemuOpts *opts, - Error **errp) +static int coroutine_fn GRAPH_RDLOCK +parallels_co_create_opts(BlockDriver *drv, const char *filename, + QemuOpts *opts, Error **errp) { BlockdevCreateOptions *create_options = NULL; BlockDriverState *bs = NULL; 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 @@ exit: return ret; } -static int coroutine_fn qcow_co_create_opts(BlockDriver *drv, - const char *filename, - QemuOpts *opts, Error **errp) +static int coroutine_fn GRAPH_RDLOCK +qcow_co_create_opts(BlockDriver *drv, const char *filename, + QemuOpts *opts, Error **errp) { BlockdevCreateOptions *create_options = NULL; BlockDriverState *bs = NULL; 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 @@ out: return ret; } -static int coroutine_fn qcow2_co_create_opts(BlockDriver *drv, - const char *filename, - QemuOpts *opts, - Error **errp) +static int coroutine_fn GRAPH_RDLOCK +qcow2_co_create_opts(BlockDriver *drv, const char *filename, QemuOpts *opts, + Error **errp) { BlockdevCreateOptions *create_options = NULL; QDict *qdict; 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 @@ out: return ret; } -static int coroutine_fn bdrv_qed_co_create_opts(BlockDriver *drv, - const char *filename, - QemuOpts *opts, - Error **errp) +static int coroutine_fn GRAPH_RDLOCK +bdrv_qed_co_create_opts(BlockDriver *drv, const char *filename, + QemuOpts *opts, Error **errp) { BlockdevCreateOptions *create_options = NULL; QDict *qdict; 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 int raw_has_zero_init(BlockDriverState *bs) return bdrv_has_zero_init(bs->file->bs); } -static int coroutine_fn raw_co_create_opts(BlockDriver *drv, - const char *filename, - QemuOpts *opts, - Error **errp) +static int coroutine_fn GRAPH_RDLOCK +raw_co_create_opts(BlockDriver *drv, const char *filename, + QemuOpts *opts, Error **errp) { return bdrv_co_create_file(filename, opts, errp); } 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_create(BlockdevCreateOptions *create_options, return vdi_co_do_create(create_options, DEFAULT_CLUSTER_SIZE, errp); } -static int coroutine_fn vdi_co_create_opts(BlockDriver *drv, - const char *filename, - QemuOpts *opts, - Error **errp) +static int coroutine_fn GRAPH_RDLOCK +vdi_co_create_opts(BlockDriver *drv, const char *filename, + QemuOpts *opts, Error **errp) { QDict *qdict = NULL; BlockdevCreateOptions *create_options = NULL; 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 @@ delete_and_exit: return ret; } -static int coroutine_fn vhdx_co_create_opts(BlockDriver *drv, - const char *filename, - QemuOpts *opts, - Error **errp) +static int coroutine_fn GRAPH_RDLOCK +vhdx_co_create_opts(BlockDriver *drv, const char *filename, + QemuOpts *opts, Error **errp) { BlockdevCreateOptions *create_options = NULL; QDict *qdict; 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: return ret; } -static int coroutine_fn vmdk_create_extent(const char *filename, - int64_t filesize, bool flat, - bool compress, bool zeroed_grain, - BlockBackend **pbb, - QemuOpts *opts, Error **errp) +static int coroutine_fn GRAPH_RDLOCK +vmdk_create_extent(const char *filename, int64_t filesize, bool flat, + bool compress, bool zeroed_grain, BlockBackend **pbb, + QemuOpts *opts, Error **errp) { int ret; BlockBackend *blk = NULL; @@ -XXX,XX +XXX,XX @@ static int filename_decompose(const char *filename, char *path, char *prefix, * non-split format. * idx >= 1: get the n-th extent if in a split subformat */ -typedef BlockBackend * coroutine_fn (*vmdk_create_extent_fn)(int64_t size, - int idx, - bool flat, - bool split, - bool compress, - bool zeroed_grain, - void *opaque, - Error **errp); +typedef BlockBackend * coroutine_fn /* GRAPH_RDLOCK */ + (*vmdk_create_extent_fn)(int64_t size, int idx, bool flat, bool split, + bool compress, bool zeroed_grain, void *opaque, + Error **errp); static void vmdk_desc_add_extent(GString *desc, const char *extent_line_fmt, @@ -XXX,XX +XXX,XX @@ static void vmdk_desc_add_extent(GString *desc, g_free(basename); } -static int coroutine_fn vmdk_co_do_create(int64_t size, - BlockdevVmdkSubformat subformat, - BlockdevVmdkAdapterType adapter_type, - const char *backing_file, - const char *hw_version, - const char *toolsversion, - bool compat6, - bool zeroed_grain, - vmdk_create_extent_fn extent_fn, - void *opaque, - Error **errp) +static int coroutine_fn GRAPH_RDLOCK +vmdk_co_do_create(int64_t size, + BlockdevVmdkSubformat subformat, + BlockdevVmdkAdapterType adapter_type, + const char *backing_file, + const char *hw_version, + const char *toolsversion, + bool compat6, + bool zeroed_grain, + vmdk_create_extent_fn extent_fn, + void *opaque, + Error **errp) { int extent_idx; BlockBackend *blk = NULL; @@ -XXX,XX +XXX,XX @@ typedef struct { QemuOpts *opts; } VMDKCreateOptsData; -static BlockBackend * coroutine_fn vmdk_co_create_opts_cb(int64_t size, int idx, - bool flat, bool split, bool compress, - bool zeroed_grain, void *opaque, - Error **errp) +static BlockBackend * coroutine_fn GRAPH_RDLOCK +vmdk_co_create_opts_cb(int64_t size, int idx, bool flat, bool split, + bool compress, bool zeroed_grain, void *opaque, + Error **errp) { BlockBackend *blk = NULL; BlockDriverState *bs = NULL; @@ -XXX,XX +XXX,XX @@ exit: return blk; } -static int coroutine_fn vmdk_co_create_opts(BlockDriver *drv, - const char *filename, - QemuOpts *opts, - Error **errp) +static int coroutine_fn GRAPH_RDLOCK +vmdk_co_create_opts(BlockDriver *drv, const char *filename, + QemuOpts *opts, Error **errp) { Error *local_err = NULL; char *desc = NULL; @@ -XXX,XX +XXX,XX @@ static BlockBackend * coroutine_fn vmdk_co_create_cb(int64_t size, int idx, return blk; } -static int coroutine_fn vmdk_co_create(BlockdevCreateOptions *create_options, - Error **errp) +static int coroutine_fn GRAPH_RDLOCK +vmdk_co_create(BlockdevCreateOptions *create_options, Error **errp) { BlockdevCreateOptionsVmdk *opts; 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 @@ out: return ret; } -static int coroutine_fn vpc_co_create_opts(BlockDriver *drv, - const char *filename, - QemuOpts *opts, - Error **errp) +static int coroutine_fn GRAPH_RDLOCK +vpc_co_create_opts(BlockDriver *drv, const char *filename, + QemuOpts *opts, Error **errp) { BlockdevCreateOptions *create_options = NULL; QDict *qdict; -- 2.39.2 This adds GRAPH_RDLOCK annotations to declare that callers of bdrv_co_io_plug() and bdrv_co_io_unplug() need to hold a reader lock for the graph. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230203152202.49054-18-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-io.h | 4 ++-- include/block/block_int-common.h | 5 +++-- block/block-backend.c | 2 ++ block/io.c | 2 ++ 4 files changed, 9 insertions(+), 4 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 coroutine_fn bdrv_co_leave(BlockDriverState *bs, AioContext *old_ctx); AioContext *child_of_bds_get_parent_aio_context(BdrvChild *c); -void coroutine_fn bdrv_co_io_plug(BlockDriverState *bs); -void coroutine_fn bdrv_co_io_unplug(BlockDriverState *bs); +void coroutine_fn GRAPH_RDLOCK bdrv_co_io_plug(BlockDriverState *bs); +void coroutine_fn GRAPH_RDLOCK bdrv_co_io_unplug(BlockDriverState *bs); bool coroutine_fn bdrv_co_can_store_new_dirty_bitmap(BlockDriverState *bs, const char *name, 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 { BlkdebugEvent event); /* io queue for linux-aio */ - void coroutine_fn (*bdrv_co_io_plug)(BlockDriverState *bs); - void coroutine_fn (*bdrv_co_io_unplug)(BlockDriverState *bs); + void coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_io_plug)(BlockDriverState *bs); + void coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_io_unplug)( + BlockDriverState *bs); /** * bdrv_drain_begin is called if implemented in the beginning of a 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 coroutine_fn blk_co_io_plug(BlockBackend *blk) { BlockDriverState *bs = blk_bs(blk); IO_CODE(); + GRAPH_RDLOCK_GUARD(); if (bs) { bdrv_co_io_plug(bs); @@ -XXX,XX +XXX,XX @@ void coroutine_fn blk_co_io_unplug(BlockBackend *blk) { BlockDriverState *bs = blk_bs(blk); IO_CODE(); + GRAPH_RDLOCK_GUARD(); if (bs) { bdrv_co_io_unplug(bs); 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 coroutine_fn bdrv_co_io_plug(BlockDriverState *bs) { BdrvChild *child; IO_CODE(); + assert_bdrv_graph_readable(); QLIST_FOREACH(child, &bs->children, next) { bdrv_co_io_plug(child->bs); @@ -XXX,XX +XXX,XX @@ void coroutine_fn bdrv_co_io_unplug(BlockDriverState *bs) { BdrvChild *child; IO_CODE(); + assert_bdrv_graph_readable(); assert(bs->io_plugged); if (qatomic_fetch_dec(&bs->io_plugged) == 1) { -- 2.39.2 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> This adds GRAPH_RDLOCK annotations to declare that callers of bdrv_co_is_inserted() need to hold a reader lock for the graph. blk_is_inserted() is done as a co_wrapper_mixed_bdrv_rdlock (unlike most other blk_* functions) because it is called a lot from other blk_co_*() functions that already hold the lock. These calls go through blk_is_available(), which becomes a co_wrapper_mixed_bdrv_rdlock, too, for the same reason. Functions that run in a coroutine and can call bdrv_co_is_available() directly are changed to do so, which results in better TSA coverage. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230203152202.49054-19-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-io.h | 4 ++-- include/block/block_int-common.h | 3 ++- include/sysemu/block-backend-io.h | 7 ++++--- block.c | 1 + block/block-backend.c | 25 ++++++++++++++----------- 5 files changed, 23 insertions(+), 17 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 @@ bool bdrv_is_writable(BlockDriverState *bs); bool bdrv_is_sg(BlockDriverState *bs); int bdrv_get_flags(BlockDriverState *bs); -bool coroutine_fn bdrv_co_is_inserted(BlockDriverState *bs); -bool co_wrapper bdrv_is_inserted(BlockDriverState *bs); +bool coroutine_fn GRAPH_RDLOCK bdrv_co_is_inserted(BlockDriverState *bs); +bool co_wrapper_bdrv_rdlock bdrv_is_inserted(BlockDriverState *bs); void coroutine_fn bdrv_co_lock_medium(BlockDriverState *bs, bool locked); void coroutine_fn bdrv_co_eject(BlockDriverState *bs, bool eject_flag); 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, QEMUIOVector *qiov, int64_t pos); /* removable device specific */ - bool coroutine_fn (*bdrv_co_is_inserted)(BlockDriverState *bs); + bool coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_is_inserted)( + BlockDriverState *bs); void coroutine_fn (*bdrv_co_eject)(BlockDriverState *bs, bool eject_flag); void coroutine_fn (*bdrv_co_lock_medium)(BlockDriverState *bs, bool locked); diff --git a/include/sysemu/block-backend-io.h b/include/sysemu/block-backend-io.h index XXXXXXX..XXXXXXX 100644 --- a/include/sysemu/block-backend-io.h +++ b/include/sysemu/block-backend-io.h @@ -XXX,XX +XXX,XX @@ BlockAIOCB *blk_aio_ioctl(BlockBackend *blk, unsigned long int req, void *buf, void blk_inc_in_flight(BlockBackend *blk); void blk_dec_in_flight(BlockBackend *blk); -bool coroutine_fn blk_co_is_inserted(BlockBackend *blk); -bool co_wrapper_mixed blk_is_inserted(BlockBackend *blk); +bool coroutine_fn GRAPH_RDLOCK blk_co_is_inserted(BlockBackend *blk); +bool co_wrapper_mixed_bdrv_rdlock blk_is_inserted(BlockBackend *blk); -bool blk_is_available(BlockBackend *blk); +bool coroutine_fn GRAPH_RDLOCK blk_co_is_available(BlockBackend *blk); +bool co_wrapper_mixed_bdrv_rdlock blk_is_available(BlockBackend *blk); void coroutine_fn blk_co_lock_medium(BlockBackend *blk, bool locked); void co_wrapper blk_lock_medium(BlockBackend *blk, bool locked); diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ bool coroutine_fn bdrv_co_is_inserted(BlockDriverState *bs) BlockDriver *drv = bs->drv; BdrvChild *child; IO_CODE(); + assert_bdrv_graph_readable(); if (!drv) { return false; 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_set_disable_request_queuing(BlockBackend *blk, bool disable) blk->disable_request_queuing = disable; } -static coroutine_fn int blk_check_byte_request(BlockBackend *blk, - int64_t offset, int64_t bytes) +static int coroutine_fn GRAPH_RDLOCK +blk_check_byte_request(BlockBackend *blk, int64_t offset, int64_t bytes) { int64_t len; @@ -XXX,XX +XXX,XX @@ static coroutine_fn int blk_check_byte_request(BlockBackend *blk, return -EIO; } - if (!blk_is_available(blk)) { + if (!blk_co_is_available(blk)) { return -ENOMEDIUM; } @@ -XXX,XX +XXX,XX @@ BlockAIOCB *blk_aio_pwrite_zeroes(BlockBackend *blk, int64_t offset, int64_t coroutine_fn blk_co_getlength(BlockBackend *blk) { IO_CODE(); + GRAPH_RDLOCK_GUARD(); - if (!blk_is_available(blk)) { + if (!blk_co_is_available(blk)) { return -ENOMEDIUM; } @@ -XXX,XX +XXX,XX @@ void blk_get_geometry(BlockBackend *blk, uint64_t *nb_sectors_ptr) int64_t coroutine_fn blk_co_nb_sectors(BlockBackend *blk) { IO_CODE(); + GRAPH_RDLOCK_GUARD(); - if (!blk_is_available(blk)) { + if (!blk_co_is_available(blk)) { return -ENOMEDIUM; } @@ -XXX,XX +XXX,XX @@ blk_co_do_ioctl(BlockBackend *blk, unsigned long int req, void *buf) blk_wait_while_drained(blk); GRAPH_RDLOCK_GUARD(); - if (!blk_is_available(blk)) { + if (!blk_co_is_available(blk)) { return -ENOMEDIUM; } @@ -XXX,XX +XXX,XX @@ static int coroutine_fn blk_co_do_flush(BlockBackend *blk) blk_wait_while_drained(blk); GRAPH_RDLOCK_GUARD(); - if (!blk_is_available(blk)) { + if (!blk_co_is_available(blk)) { return -ENOMEDIUM; } @@ -XXX,XX +XXX,XX @@ bool coroutine_fn blk_co_is_inserted(BlockBackend *blk) { BlockDriverState *bs = blk_bs(blk); IO_CODE(); + assert_bdrv_graph_readable(); return bs && bdrv_co_is_inserted(bs); } -bool blk_is_available(BlockBackend *blk) +bool coroutine_fn blk_co_is_available(BlockBackend *blk) { IO_CODE(); - return blk_is_inserted(blk) && !blk_dev_is_tray_open(blk); + return blk_co_is_inserted(blk) && !blk_dev_is_tray_open(blk); } void coroutine_fn blk_co_lock_medium(BlockBackend *blk, bool locked) @@ -XXX,XX +XXX,XX @@ int coroutine_fn blk_co_truncate(BlockBackend *blk, int64_t offset, bool exact, { IO_OR_GS_CODE(); GRAPH_RDLOCK_GUARD(); - if (!blk_is_available(blk)) { + if (!blk_co_is_available(blk)) { error_setg(errp, "No medium inserted"); return -ENOMEDIUM; } @@ -XXX,XX +XXX,XX @@ int coroutine_fn blk_co_copy_range(BlockBackend *blk_in, int64_t off_in, { int r; IO_CODE(); + GRAPH_RDLOCK_GUARD(); r = blk_check_byte_request(blk_in, off_in, bytes); if (r) { @@ -XXX,XX +XXX,XX @@ int coroutine_fn blk_co_copy_range(BlockBackend *blk_in, int64_t off_in, return r; } - GRAPH_RDLOCK_GUARD(); return bdrv_co_copy_range(blk_in->root, off_in, blk_out->root, off_out, bytes, read_flags, write_flags); -- 2.39.2 This adds GRAPH_RDLOCK annotations to declare that callers of bdrv_co_eject() and bdrv_co_lock_medium() 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> Message-Id: <20230203152202.49054-20-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-io.h | 7 +++++-- include/block/block_int-common.h | 6 ++++-- block.c | 2 ++ block/block-backend.c | 2 ++ block/copy-on-read.c | 6 ++++-- block/filter-compress.c | 4 ++-- block/raw-format.c | 6 ++++-- 7 files changed, 23 insertions(+), 10 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 @@ int bdrv_get_flags(BlockDriverState *bs); bool coroutine_fn GRAPH_RDLOCK bdrv_co_is_inserted(BlockDriverState *bs); bool co_wrapper_bdrv_rdlock bdrv_is_inserted(BlockDriverState *bs); -void coroutine_fn bdrv_co_lock_medium(BlockDriverState *bs, bool locked); -void coroutine_fn bdrv_co_eject(BlockDriverState *bs, bool eject_flag); +void coroutine_fn GRAPH_RDLOCK +bdrv_co_lock_medium(BlockDriverState *bs, bool locked); + +void coroutine_fn GRAPH_RDLOCK +bdrv_co_eject(BlockDriverState *bs, bool eject_flag); const char *bdrv_get_format_name(BlockDriverState *bs); 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 { /* removable device specific */ bool coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_is_inserted)( BlockDriverState *bs); - void coroutine_fn (*bdrv_co_eject)(BlockDriverState *bs, bool eject_flag); - void coroutine_fn (*bdrv_co_lock_medium)(BlockDriverState *bs, bool locked); + void coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_eject)( + BlockDriverState *bs, bool eject_flag); + void coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_lock_medium)( + BlockDriverState *bs, bool locked); /* to control generic scsi devices */ BlockAIOCB *coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_aio_ioctl)( diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ void coroutine_fn bdrv_co_eject(BlockDriverState *bs, bool eject_flag) { BlockDriver *drv = bs->drv; IO_CODE(); + assert_bdrv_graph_readable(); if (drv && drv->bdrv_co_eject) { drv->bdrv_co_eject(bs, eject_flag); @@ -XXX,XX +XXX,XX @@ void coroutine_fn bdrv_co_lock_medium(BlockDriverState *bs, bool locked) { BlockDriver *drv = bs->drv; IO_CODE(); + assert_bdrv_graph_readable(); trace_bdrv_lock_medium(bs, locked); if (drv && drv->bdrv_co_lock_medium) { 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 coroutine_fn blk_co_lock_medium(BlockBackend *blk, bool locked) { BlockDriverState *bs = blk_bs(blk); IO_CODE(); + GRAPH_RDLOCK_GUARD(); if (bs) { bdrv_co_lock_medium(bs, locked); @@ -XXX,XX +XXX,XX @@ void coroutine_fn blk_co_eject(BlockBackend *blk, bool eject_flag) BlockDriverState *bs = blk_bs(blk); char *id; IO_CODE(); + GRAPH_RDLOCK_GUARD(); if (bs) { bdrv_co_eject(bs, eject_flag); diff --git a/block/copy-on-read.c b/block/copy-on-read.c index XXXXXXX..XXXXXXX 100644 --- a/block/copy-on-read.c +++ b/block/copy-on-read.c @@ -XXX,XX +XXX,XX @@ cor_co_pwritev_compressed(BlockDriverState *bs, int64_t offset, int64_t bytes, } -static void coroutine_fn cor_co_eject(BlockDriverState *bs, bool eject_flag) +static void coroutine_fn GRAPH_RDLOCK +cor_co_eject(BlockDriverState *bs, bool eject_flag) { bdrv_co_eject(bs->file->bs, eject_flag); } -static void coroutine_fn cor_co_lock_medium(BlockDriverState *bs, bool locked) +static void coroutine_fn GRAPH_RDLOCK +cor_co_lock_medium(BlockDriverState *bs, bool locked) { bdrv_co_lock_medium(bs->file->bs, locked); } diff --git a/block/filter-compress.c b/block/filter-compress.c index XXXXXXX..XXXXXXX 100644 --- a/block/filter-compress.c +++ b/block/filter-compress.c @@ -XXX,XX +XXX,XX @@ static void compress_refresh_limits(BlockDriverState *bs, Error **errp) } -static void coroutine_fn +static void coroutine_fn GRAPH_RDLOCK compress_co_eject(BlockDriverState *bs, bool eject_flag) { bdrv_co_eject(bs->file->bs, eject_flag); } -static void coroutine_fn +static void coroutine_fn GRAPH_RDLOCK compress_co_lock_medium(BlockDriverState *bs, bool locked) { bdrv_co_lock_medium(bs->file->bs, locked); 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 @@ raw_co_truncate(BlockDriverState *bs, int64_t offset, bool exact, return bdrv_co_truncate(bs->file, offset, exact, prealloc, flags, errp); } -static void coroutine_fn raw_co_eject(BlockDriverState *bs, bool eject_flag) +static void coroutine_fn GRAPH_RDLOCK +raw_co_eject(BlockDriverState *bs, bool eject_flag) { bdrv_co_eject(bs->file->bs, eject_flag); } -static void coroutine_fn raw_co_lock_medium(BlockDriverState *bs, bool locked) +static void coroutine_fn GRAPH_RDLOCK +raw_co_lock_medium(BlockDriverState *bs, bool locked) { bdrv_co_lock_medium(bs->file->bs, locked); } -- 2.39.2 This adds GRAPH_RDLOCK annotations to declare that callers of bdrv_register_buf() and bdrv_unregister_buf() need to hold a reader lock for the graph. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230203152202.49054-21-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block_int-common.h | 7 ++++--- block/io.c | 14 ++++++++++---- 2 files changed, 14 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 { * * Returns: true on success, false on failure */ - bool (*bdrv_register_buf)(BlockDriverState *bs, void *host, size_t size, - Error **errp); - void (*bdrv_unregister_buf)(BlockDriverState *bs, void *host, size_t size); + bool GRAPH_RDLOCK_PTR (*bdrv_register_buf)( + BlockDriverState *bs, void *host, size_t size, Error **errp); + void GRAPH_RDLOCK_PTR (*bdrv_unregister_buf)( + BlockDriverState *bs, void *host, size_t size); /* * This field is modified only under the BQL, and is part of 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 coroutine_fn bdrv_co_io_unplug(BlockDriverState *bs) } /* Helper that undoes bdrv_register_buf() when it fails partway through */ -static void bdrv_register_buf_rollback(BlockDriverState *bs, - void *host, - size_t size, - BdrvChild *final_child) +static void GRAPH_RDLOCK +bdrv_register_buf_rollback(BlockDriverState *bs, void *host, size_t size, + BdrvChild *final_child) { BdrvChild *child; + GLOBAL_STATE_CODE(); + assert_bdrv_graph_readable(); + QLIST_FOREACH(child, &bs->children, next) { if (child == final_child) { break; @@ -XXX,XX +XXX,XX @@ bool bdrv_register_buf(BlockDriverState *bs, void *host, size_t size, BdrvChild *child; GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); + if (bs->drv && bs->drv->bdrv_register_buf) { if (!bs->drv->bdrv_register_buf(bs, host, size, errp)) { return false; @@ -XXX,XX +XXX,XX @@ void bdrv_unregister_buf(BlockDriverState *bs, void *host, size_t size) BdrvChild *child; GLOBAL_STATE_CODE(); + GRAPH_RDLOCK_GUARD_MAINLOOP(); + if (bs->drv && bs->drv->bdrv_unregister_buf) { bs->drv->bdrv_unregister_buf(bs, host, size); } -- 2.39.2 This adds GRAPH_RDLOCK annotations to declare that callers of bdrv_co_delete_file() need to hold a reader lock for the graph. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230203152202.49054-22-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-io.h | 8 ++++++-- include/block/block_int-common.h | 4 ++-- block.c | 1 + 3 files changed, 9 insertions(+), 4 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 co_wrapper bdrv_get_allocated_file_size(BlockDriverState *bs); BlockMeasureInfo *bdrv_measure(BlockDriver *drv, QemuOpts *opts, BlockDriverState *in_bs, Error **errp); void bdrv_get_geometry(BlockDriverState *bs, uint64_t *nb_sectors_ptr); -int coroutine_fn bdrv_co_delete_file(BlockDriverState *bs, Error **errp); -void coroutine_fn bdrv_co_delete_file_noerr(BlockDriverState *bs); + +int coroutine_fn GRAPH_RDLOCK +bdrv_co_delete_file(BlockDriverState *bs, Error **errp); + +void coroutine_fn GRAPH_RDLOCK +bdrv_co_delete_file_noerr(BlockDriverState *bs); /* async block I/O */ 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_flush)(BlockDriverState *bs); /* Delete a created file. */ - int coroutine_fn (*bdrv_co_delete_file)(BlockDriverState *bs, - Error **errp); + int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_delete_file)( + BlockDriverState *bs, Error **errp); /* * Flushes all data that was already written to the OS all the way down to 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_delete_file(BlockDriverState *bs, Error **errp) IO_CODE(); assert(bs != NULL); + assert_bdrv_graph_readable(); if (!bs->drv) { error_setg(errp, "Block node '%s' is not opened", bs->filename); -- 2.39.2 This adds GRAPH_RDLOCK annotations to declare that callers of bdrv_*_dirty_bitmap() need to hold a reader lock for the graph. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230203152202.49054-23-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-io.h | 14 ++++++-------- include/block/block_int-common.h | 6 ++++-- include/block/dirty-bitmap.h | 12 ++++++------ block/dirty-bitmap.c | 2 ++ 4 files changed, 18 insertions(+), 16 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 @@ AioContext *child_of_bds_get_parent_aio_context(BdrvChild *c); void coroutine_fn GRAPH_RDLOCK bdrv_co_io_plug(BlockDriverState *bs); void coroutine_fn GRAPH_RDLOCK bdrv_co_io_unplug(BlockDriverState *bs); -bool coroutine_fn bdrv_co_can_store_new_dirty_bitmap(BlockDriverState *bs, - const char *name, - uint32_t granularity, - Error **errp); -bool co_wrapper bdrv_can_store_new_dirty_bitmap(BlockDriverState *bs, - const char *name, - uint32_t granularity, - Error **errp); +bool coroutine_fn GRAPH_RDLOCK +bdrv_co_can_store_new_dirty_bitmap(BlockDriverState *bs, const char *name, + uint32_t granularity, Error **errp); +bool co_wrapper_bdrv_rdlock +bdrv_can_store_new_dirty_bitmap(BlockDriverState *bs, const char *name, + uint32_t granularity, 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 { void (*bdrv_drain_end)(BlockDriverState *bs); bool (*bdrv_supports_persistent_dirty_bitmap)(BlockDriverState *bs); - bool coroutine_fn (*bdrv_co_can_store_new_dirty_bitmap)( + + bool coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_can_store_new_dirty_bitmap)( BlockDriverState *bs, const char *name, uint32_t granularity, Error **errp); - int coroutine_fn (*bdrv_co_remove_persistent_dirty_bitmap)( + + int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_remove_persistent_dirty_bitmap)( BlockDriverState *bs, const char *name, Error **errp); }; diff --git a/include/block/dirty-bitmap.h b/include/block/dirty-bitmap.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/dirty-bitmap.h +++ b/include/block/dirty-bitmap.h @@ -XXX,XX +XXX,XX @@ int bdrv_dirty_bitmap_check(const BdrvDirtyBitmap *bitmap, uint32_t flags, void bdrv_release_dirty_bitmap(BdrvDirtyBitmap *bitmap); void bdrv_release_named_dirty_bitmaps(BlockDriverState *bs); -int coroutine_fn bdrv_co_remove_persistent_dirty_bitmap(BlockDriverState *bs, - const char *name, - Error **errp); -int co_wrapper bdrv_remove_persistent_dirty_bitmap(BlockDriverState *bs, - const char *name, - Error **errp); +int coroutine_fn GRAPH_RDLOCK +bdrv_co_remove_persistent_dirty_bitmap(BlockDriverState *bs, const char *name, + Error **errp); +int co_wrapper_bdrv_rdlock +bdrv_remove_persistent_dirty_bitmap(BlockDriverState *bs, const char *name, + Error **errp); void bdrv_disable_dirty_bitmap(BdrvDirtyBitmap *bitmap); void bdrv_enable_dirty_bitmap(BdrvDirtyBitmap *bitmap); diff --git a/block/dirty-bitmap.c b/block/dirty-bitmap.c index XXXXXXX..XXXXXXX 100644 --- a/block/dirty-bitmap.c +++ b/block/dirty-bitmap.c @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_remove_persistent_dirty_bitmap(BlockDriverState *bs, const char *name, Error **errp) { + assert_bdrv_graph_readable(); if (bs->drv && bs->drv->bdrv_co_remove_persistent_dirty_bitmap) { return bs->drv->bdrv_co_remove_persistent_dirty_bitmap(bs, name, errp); } @@ -XXX,XX +XXX,XX @@ bdrv_co_can_store_new_dirty_bitmap(BlockDriverState *bs, const char *name, uint32_t granularity, Error **errp) { BlockDriver *drv = bs->drv; + assert_bdrv_graph_readable(); if (!drv) { error_setg_errno(errp, ENOMEDIUM, -- 2.39.2 This adds GRAPH_RDLOCK annotations to declare that callers of bdrv_co_refresh_total_sectors() need to hold a reader lock for the graph. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230203152202.49054-24-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-io.h | 8 ++++---- include/block/block_int-common.h | 4 +++- include/block/block_int-io.h | 7 ++++--- block.c | 3 +++ block/blkdebug.c | 3 ++- block/blklogwrites.c | 3 ++- block/blkreplay.c | 3 ++- block/blkverify.c | 3 ++- block/copy-on-read.c | 2 +- block/crypto.c | 3 ++- block/filter-compress.c | 3 ++- block/mirror.c | 3 +++ block/preallocate.c | 3 ++- block/quorum.c | 3 ++- block/raw-format.c | 3 ++- block/replication.c | 3 ++- block/stream.c | 8 +++++--- block/throttle.c | 3 ++- 18 files changed, 45 insertions(+), 23 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 @@ int coroutine_fn GRAPH_RDLOCK bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact, PreallocMode prealloc, BdrvRequestFlags flags, Error **errp); -int64_t coroutine_fn bdrv_co_nb_sectors(BlockDriverState *bs); -int64_t co_wrapper_mixed bdrv_nb_sectors(BlockDriverState *bs); +int64_t coroutine_fn GRAPH_RDLOCK bdrv_co_nb_sectors(BlockDriverState *bs); +int64_t co_wrapper_mixed_bdrv_rdlock bdrv_nb_sectors(BlockDriverState *bs); -int64_t coroutine_fn bdrv_co_getlength(BlockDriverState *bs); -int64_t co_wrapper_mixed bdrv_getlength(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); 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, bool exact, PreallocMode prealloc, BdrvRequestFlags flags, Error **errp); - int64_t coroutine_fn (*bdrv_co_getlength)(BlockDriverState *bs); + int64_t coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_getlength)( + BlockDriverState *bs); + int64_t coroutine_fn (*bdrv_co_get_allocated_file_size)( BlockDriverState *bs); diff --git a/include/block/block_int-io.h b/include/block/block_int-io.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block_int-io.h +++ b/include/block/block_int-io.h @@ -XXX,XX +XXX,XX @@ bdrv_co_copy_range_to(BdrvChild *src, int64_t src_offset, int64_t bytes, BdrvRequestFlags read_flags, BdrvRequestFlags write_flags); -int coroutine_fn bdrv_co_refresh_total_sectors(BlockDriverState *bs, - int64_t hint); -int co_wrapper_mixed +int coroutine_fn GRAPH_RDLOCK +bdrv_co_refresh_total_sectors(BlockDriverState *bs, int64_t hint); + +int co_wrapper_mixed_bdrv_rdlock bdrv_refresh_total_sectors(BlockDriverState *bs, int64_t hint); BdrvChild *bdrv_cow_child(BlockDriverState *bs); 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_refresh_total_sectors(BlockDriverState *bs, { BlockDriver *drv = bs->drv; IO_CODE(); + assert_bdrv_graph_readable(); if (!drv) { return -ENOMEDIUM; @@ -XXX,XX +XXX,XX @@ int64_t coroutine_fn bdrv_co_nb_sectors(BlockDriverState *bs) { BlockDriver *drv = bs->drv; IO_CODE(); + assert_bdrv_graph_readable(); if (!drv) return -ENOMEDIUM; @@ -XXX,XX +XXX,XX @@ int64_t coroutine_fn bdrv_co_getlength(BlockDriverState *bs) { int64_t ret; IO_CODE(); + assert_bdrv_graph_readable(); ret = bdrv_co_nb_sectors(bs); if (ret < 0) { diff --git a/block/blkdebug.c b/block/blkdebug.c index XXXXXXX..XXXXXXX 100644 --- a/block/blkdebug.c +++ b/block/blkdebug.c @@ -XXX,XX +XXX,XX @@ static bool blkdebug_debug_is_suspended(BlockDriverState *bs, const char *tag) return false; } -static int64_t coroutine_fn blkdebug_co_getlength(BlockDriverState *bs) +static int64_t coroutine_fn GRAPH_RDLOCK +blkdebug_co_getlength(BlockDriverState *bs) { return bdrv_co_getlength(bs->file->bs); } diff --git a/block/blklogwrites.c b/block/blklogwrites.c index XXXXXXX..XXXXXXX 100644 --- a/block/blklogwrites.c +++ b/block/blklogwrites.c @@ -XXX,XX +XXX,XX @@ static void blk_log_writes_close(BlockDriverState *bs) s->log_file = NULL; } -static int64_t coroutine_fn blk_log_writes_co_getlength(BlockDriverState *bs) +static int64_t coroutine_fn GRAPH_RDLOCK +blk_log_writes_co_getlength(BlockDriverState *bs) { return bdrv_co_getlength(bs->file->bs); } diff --git a/block/blkreplay.c b/block/blkreplay.c index XXXXXXX..XXXXXXX 100644 --- a/block/blkreplay.c +++ b/block/blkreplay.c @@ -XXX,XX +XXX,XX @@ fail: return ret; } -static int64_t coroutine_fn blkreplay_co_getlength(BlockDriverState *bs) +static int64_t coroutine_fn GRAPH_RDLOCK +blkreplay_co_getlength(BlockDriverState *bs) { return bdrv_co_getlength(bs->file->bs); } 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 void blkverify_close(BlockDriverState *bs) s->test_file = NULL; } -static int64_t coroutine_fn blkverify_co_getlength(BlockDriverState *bs) +static int64_t coroutine_fn GRAPH_RDLOCK +blkverify_co_getlength(BlockDriverState *bs) { BDRVBlkverifyState *s = bs->opaque; diff --git a/block/copy-on-read.c b/block/copy-on-read.c index XXXXXXX..XXXXXXX 100644 --- a/block/copy-on-read.c +++ b/block/copy-on-read.c @@ -XXX,XX +XXX,XX @@ static void cor_child_perm(BlockDriverState *bs, BdrvChild *c, } -static int64_t coroutine_fn cor_co_getlength(BlockDriverState *bs) +static int64_t coroutine_fn GRAPH_RDLOCK cor_co_getlength(BlockDriverState *bs) { return bdrv_co_getlength(bs->file->bs); } 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 @@ static void block_crypto_refresh_limits(BlockDriverState *bs, Error **errp) } -static int64_t coroutine_fn block_crypto_co_getlength(BlockDriverState *bs) +static int64_t coroutine_fn GRAPH_RDLOCK +block_crypto_co_getlength(BlockDriverState *bs) { BlockCrypto *crypto = bs->opaque; int64_t len = bdrv_co_getlength(bs->file->bs); diff --git a/block/filter-compress.c b/block/filter-compress.c index XXXXXXX..XXXXXXX 100644 --- a/block/filter-compress.c +++ b/block/filter-compress.c @@ -XXX,XX +XXX,XX @@ static int compress_open(BlockDriverState *bs, QDict *options, int flags, } -static int64_t coroutine_fn compress_co_getlength(BlockDriverState *bs) +static int64_t coroutine_fn GRAPH_RDLOCK +compress_co_getlength(BlockDriverState *bs) { return bdrv_co_getlength(bs->file->bs); } 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 coroutine_fn mirror_run(Job *job, Error **errp) goto immediate_exit; } + bdrv_graph_co_rdlock(); s->bdev_length = bdrv_co_getlength(bs); + bdrv_graph_co_rdunlock(); + if (s->bdev_length < 0) { ret = s->bdev_length; goto immediate_exit; diff --git a/block/preallocate.c b/block/preallocate.c index XXXXXXX..XXXXXXX 100644 --- a/block/preallocate.c +++ b/block/preallocate.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn GRAPH_RDLOCK preallocate_co_flush(BlockDriverState *bs) return bdrv_co_flush(bs->file->bs); } -static int64_t coroutine_fn preallocate_co_getlength(BlockDriverState *bs) +static int64_t coroutine_fn GRAPH_RDLOCK +preallocate_co_getlength(BlockDriverState *bs) { int64_t ret; BDRVPreallocateState *s = bs->opaque; 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 @@ quorum_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset, int64_t bytes, flags | BDRV_REQ_ZERO_WRITE); } -static int64_t coroutine_fn quorum_co_getlength(BlockDriverState *bs) +static int64_t coroutine_fn GRAPH_RDLOCK +quorum_co_getlength(BlockDriverState *bs) { BDRVQuorumState *s = bs->opaque; int64_t result; 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 @@ raw_co_pdiscard(BlockDriverState *bs, int64_t offset, int64_t bytes) return bdrv_co_pdiscard(bs->file, offset, bytes); } -static int64_t coroutine_fn raw_co_getlength(BlockDriverState *bs) +static int64_t coroutine_fn GRAPH_RDLOCK +raw_co_getlength(BlockDriverState *bs) { int64_t len; BDRVRawState *s = bs->opaque; diff --git a/block/replication.c b/block/replication.c index XXXXXXX..XXXXXXX 100644 --- a/block/replication.c +++ b/block/replication.c @@ -XXX,XX +XXX,XX @@ static void replication_child_perm(BlockDriverState *bs, BdrvChild *c, return; } -static int64_t coroutine_fn replication_co_getlength(BlockDriverState *bs) +static int64_t coroutine_fn GRAPH_RDLOCK +replication_co_getlength(BlockDriverState *bs) { return bdrv_co_getlength(bs->file->bs); } diff --git a/block/stream.c b/block/stream.c index XXXXXXX..XXXXXXX 100644 --- a/block/stream.c +++ b/block/stream.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn stream_run(Job *job, Error **errp) return 0; } - len = bdrv_getlength(s->target_bs); - if (len < 0) { - return len; + WITH_GRAPH_RDLOCK_GUARD() { + len = bdrv_co_getlength(s->target_bs); + if (len < 0) { + return len; + } } job_progress_set_remaining(&s->common.job, len); diff --git a/block/throttle.c b/block/throttle.c index XXXXXXX..XXXXXXX 100644 --- a/block/throttle.c +++ b/block/throttle.c @@ -XXX,XX +XXX,XX @@ static void throttle_close(BlockDriverState *bs) } -static int64_t coroutine_fn throttle_co_getlength(BlockDriverState *bs) +static int64_t coroutine_fn GRAPH_RDLOCK +throttle_co_getlength(BlockDriverState *bs) { return bdrv_co_getlength(bs->file->bs); } -- 2.39.2 From: Stefan Hajnoczi <stefanha@redhat.com> If requests are being processed in the IOThread when a SCSIDevice is unplugged, scsi_device_purge_requests() -> scsi_req_cancel_async() races with I/O completion callbacks. Both threads load and store req->aiocb. This can lead to assert(r->req.aiocb == NULL) failures and undefined behavior. Protect r->req.aiocb with the AioContext lock to prevent the race. Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Message-Id: <20230221212218.1378734-2-stefanha@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- hw/scsi/scsi-disk.c | 23 ++++++++++++++++------- hw/scsi/scsi-generic.c | 11 ++++++----- 2 files changed, 22 insertions(+), 12 deletions(-) diff --git a/hw/scsi/scsi-disk.c b/hw/scsi/scsi-disk.c index XXXXXXX..XXXXXXX 100644 --- a/hw/scsi/scsi-disk.c +++ b/hw/scsi/scsi-disk.c @@ -XXX,XX +XXX,XX @@ static void scsi_aio_complete(void *opaque, int ret) SCSIDiskReq *r = (SCSIDiskReq *)opaque; SCSIDiskState *s = DO_UPCAST(SCSIDiskState, qdev, r->req.dev); + aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk)); + assert(r->req.aiocb != NULL); r->req.aiocb = NULL; - aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk)); + if (scsi_disk_req_check_error(r, ret, true)) { goto done; } @@ -XXX,XX +XXX,XX @@ static void scsi_dma_complete(void *opaque, int ret) SCSIDiskReq *r = (SCSIDiskReq *)opaque; SCSIDiskState *s = DO_UPCAST(SCSIDiskState, qdev, r->req.dev); + aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk)); + assert(r->req.aiocb != NULL); r->req.aiocb = NULL; - aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk)); if (ret < 0) { block_acct_failed(blk_get_stats(s->qdev.conf.blk), &r->acct); } else { @@ -XXX,XX +XXX,XX @@ static void scsi_read_complete(void *opaque, int ret) SCSIDiskReq *r = (SCSIDiskReq *)opaque; SCSIDiskState *s = DO_UPCAST(SCSIDiskState, qdev, r->req.dev); + aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk)); + assert(r->req.aiocb != NULL); r->req.aiocb = NULL; - aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk)); if (ret < 0) { block_acct_failed(blk_get_stats(s->qdev.conf.blk), &r->acct); } else { @@ -XXX,XX +XXX,XX @@ static void scsi_do_read_cb(void *opaque, int ret) SCSIDiskReq *r = (SCSIDiskReq *)opaque; SCSIDiskState *s = DO_UPCAST(SCSIDiskState, qdev, r->req.dev); + aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk)); + assert (r->req.aiocb != NULL); r->req.aiocb = NULL; - aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk)); if (ret < 0) { block_acct_failed(blk_get_stats(s->qdev.conf.blk), &r->acct); } else { @@ -XXX,XX +XXX,XX @@ static void scsi_write_complete(void * opaque, int ret) SCSIDiskReq *r = (SCSIDiskReq *)opaque; SCSIDiskState *s = DO_UPCAST(SCSIDiskState, qdev, r->req.dev); + aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk)); + assert (r->req.aiocb != NULL); r->req.aiocb = NULL; - aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk)); if (ret < 0) { block_acct_failed(blk_get_stats(s->qdev.conf.blk), &r->acct); } else { @@ -XXX,XX +XXX,XX @@ static void scsi_unmap_complete(void *opaque, int ret) SCSIDiskReq *r = data->r; SCSIDiskState *s = DO_UPCAST(SCSIDiskState, qdev, r->req.dev); + aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk)); + assert(r->req.aiocb != NULL); r->req.aiocb = NULL; - aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk)); if (scsi_disk_req_check_error(r, ret, true)) { scsi_req_unref(&r->req); g_free(data); @@ -XXX,XX +XXX,XX @@ static void scsi_write_same_complete(void *opaque, int ret) SCSIDiskReq *r = data->r; SCSIDiskState *s = DO_UPCAST(SCSIDiskState, qdev, r->req.dev); + aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk)); + assert(r->req.aiocb != NULL); r->req.aiocb = NULL; - aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk)); + if (scsi_disk_req_check_error(r, ret, true)) { goto done; } diff --git a/hw/scsi/scsi-generic.c b/hw/scsi/scsi-generic.c index XXXXXXX..XXXXXXX 100644 --- a/hw/scsi/scsi-generic.c +++ b/hw/scsi/scsi-generic.c @@ -XXX,XX +XXX,XX @@ static void scsi_command_complete(void *opaque, int ret) SCSIGenericReq *r = (SCSIGenericReq *)opaque; SCSIDevice *s = r->req.dev; + aio_context_acquire(blk_get_aio_context(s->conf.blk)); + assert(r->req.aiocb != NULL); r->req.aiocb = NULL; - aio_context_acquire(blk_get_aio_context(s->conf.blk)); scsi_command_complete_noio(r, ret); aio_context_release(blk_get_aio_context(s->conf.blk)); } @@ -XXX,XX +XXX,XX @@ static void scsi_read_complete(void * opaque, int ret) SCSIDevice *s = r->req.dev; int len; + aio_context_acquire(blk_get_aio_context(s->conf.blk)); + assert(r->req.aiocb != NULL); r->req.aiocb = NULL; - aio_context_acquire(blk_get_aio_context(s->conf.blk)); - if (ret || r->req.io_canceled) { scsi_command_complete_noio(r, ret); goto done; @@ -XXX,XX +XXX,XX @@ static void scsi_write_complete(void * opaque, int ret) trace_scsi_generic_write_complete(ret); + aio_context_acquire(blk_get_aio_context(s->conf.blk)); + assert(r->req.aiocb != NULL); r->req.aiocb = NULL; - aio_context_acquire(blk_get_aio_context(s->conf.blk)); - if (ret || r->req.io_canceled) { scsi_command_complete_noio(r, ret); goto done; -- 2.39.2 From: Stefan Hajnoczi <stefanha@redhat.com> dma_blk_cb() only takes the AioContext lock around ->io_func(). That means the rest of dma_blk_cb() is not protected. In particular, the DMAAIOCB field accesses happen outside the lock. There is a race when the main loop thread holds the AioContext lock and invokes scsi_device_purge_requests() -> bdrv_aio_cancel() -> dma_aio_cancel() while an IOThread executes dma_blk_cb(). The dbs->acb field determines how cancellation proceeds. If dma_aio_cancel() sees dbs->acb == NULL while dma_blk_cb() is still running, the request can be completed twice (-ECANCELED and the actual return value). The following assertion can occur with virtio-scsi when an IOThread is used: ../hw/scsi/scsi-disk.c:368: scsi_dma_complete: Assertion `r->req.aiocb != NULL' failed. Fix the race by holding the AioContext across dma_blk_cb(). Now dma_aio_cancel() under the AioContext lock will not see inconsistent/intermediate states. Cc: Paolo Bonzini <pbonzini@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Message-Id: <20230221212218.1378734-3-stefanha@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- hw/scsi/scsi-disk.c | 4 +--- softmmu/dma-helpers.c | 12 +++++++----- 2 files changed, 8 insertions(+), 8 deletions(-) diff --git a/hw/scsi/scsi-disk.c b/hw/scsi/scsi-disk.c index XXXXXXX..XXXXXXX 100644 --- a/hw/scsi/scsi-disk.c +++ b/hw/scsi/scsi-disk.c @@ -XXX,XX +XXX,XX @@ done: scsi_req_unref(&r->req); } +/* Called with AioContext lock held */ static void scsi_dma_complete(void *opaque, int ret) { SCSIDiskReq *r = (SCSIDiskReq *)opaque; SCSIDiskState *s = DO_UPCAST(SCSIDiskState, qdev, r->req.dev); - aio_context_acquire(blk_get_aio_context(s->qdev.conf.blk)); - assert(r->req.aiocb != NULL); r->req.aiocb = NULL; @@ -XXX,XX +XXX,XX @@ static void scsi_dma_complete(void *opaque, int ret) block_acct_done(blk_get_stats(s->qdev.conf.blk), &r->acct); } scsi_dma_complete_noio(r, ret); - aio_context_release(blk_get_aio_context(s->qdev.conf.blk)); } static void scsi_read_complete_noio(SCSIDiskReq *r, int ret) diff --git a/softmmu/dma-helpers.c b/softmmu/dma-helpers.c index XXXXXXX..XXXXXXX 100644 --- a/softmmu/dma-helpers.c +++ b/softmmu/dma-helpers.c @@ -XXX,XX +XXX,XX @@ static void dma_complete(DMAAIOCB *dbs, int ret) static void dma_blk_cb(void *opaque, int ret) { DMAAIOCB *dbs = (DMAAIOCB *)opaque; + AioContext *ctx = dbs->ctx; dma_addr_t cur_addr, cur_len; void *mem; trace_dma_blk_cb(dbs, ret); + aio_context_acquire(ctx); dbs->acb = NULL; dbs->offset += dbs->iov.size; if (dbs->sg_cur_index == dbs->sg->nsg || ret < 0) { dma_complete(dbs, ret); - return; + goto out; } dma_blk_unmap(dbs); @@ -XXX,XX +XXX,XX @@ static void dma_blk_cb(void *opaque, int ret) if (dbs->iov.size == 0) { trace_dma_map_wait(dbs); - dbs->bh = aio_bh_new(dbs->ctx, reschedule_dma, dbs); + dbs->bh = aio_bh_new(ctx, reschedule_dma, dbs); cpu_register_map_client(dbs->bh); - return; + goto out; } if (!QEMU_IS_ALIGNED(dbs->iov.size, dbs->align)) { @@ -XXX,XX +XXX,XX @@ static void dma_blk_cb(void *opaque, int ret) QEMU_ALIGN_DOWN(dbs->iov.size, dbs->align)); } - aio_context_acquire(dbs->ctx); dbs->acb = dbs->io_func(dbs->offset, &dbs->iov, dma_blk_cb, dbs, dbs->io_func_opaque); - aio_context_release(dbs->ctx); assert(dbs->acb); +out: + aio_context_release(ctx); } static void dma_aio_cancel(BlockAIOCB *acb) -- 2.39.2 From: Stefan Hajnoczi <stefanha@redhat.com> When an IOThread is configured, the ctrl virtqueue is processed in the IOThread. TMFs that reset SCSI devices are currently called directly from the IOThread and trigger an assertion failure in blk_drain() from the following call stack: virtio_scsi_handle_ctrl_req -> virtio_scsi_do_tmf -> device_code_reset -> scsi_disk_reset -> scsi_device_purge_requests -> blk_drain ../block/block-backend.c:1780: void blk_drain(BlockBackend *): Assertion `qemu_in_main_thread()' failed. The blk_drain() function is not designed to be called from an IOThread because it needs the Big QEMU Lock (BQL). This patch defers TMFs that reset SCSI devices to a Bottom Half (BH) that runs in the main loop thread under the BQL. This way it's safe to call blk_drain() and the assertion failure is avoided. Introduce s->tmf_bh_list for tracking TMF requests that have been deferred to the BH. When the BH runs it will grab the entire list and process all requests. Care must be taken to clear the list when the virtio-scsi device is reset or unrealized. Otherwise deferred TMF requests could execute later and lead to use-after-free or other undefined behavior. The s->resetting counter that's used by TMFs that reset SCSI devices is accessed from multiple threads. This patch makes that explicit by using atomic accessor functions. With this patch applied the counter is only modified by the main loop thread under the BQL but can be read by any thread. Reported-by: Qing Wang <qinwang@redhat.com> Cc: Paolo Bonzini <pbonzini@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Message-Id: <20230221212218.1378734-4-stefanha@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/hw/virtio/virtio-scsi.h | 11 ++- hw/scsi/virtio-scsi.c | 169 +++++++++++++++++++++++++------- 2 files changed, 143 insertions(+), 37 deletions(-) diff --git a/include/hw/virtio/virtio-scsi.h b/include/hw/virtio/virtio-scsi.h index XXXXXXX..XXXXXXX 100644 --- a/include/hw/virtio/virtio-scsi.h +++ b/include/hw/virtio/virtio-scsi.h @@ -XXX,XX +XXX,XX @@ struct VirtIOSCSICommon { VirtQueue **cmd_vqs; }; +struct VirtIOSCSIReq; + struct VirtIOSCSI { VirtIOSCSICommon parent_obj; SCSIBus bus; - int resetting; + int resetting; /* written from main loop thread, read from any thread */ bool events_dropped; + /* + * TMFs deferred to main loop BH. These fields are protected by + * virtio_scsi_acquire(). + */ + QEMUBH *tmf_bh; + QTAILQ_HEAD(, VirtIOSCSIReq) tmf_bh_list; + /* Fields for dataplane below */ AioContext *ctx; /* one iothread per virtio-scsi-pci for now */ diff --git a/hw/scsi/virtio-scsi.c b/hw/scsi/virtio-scsi.c index XXXXXXX..XXXXXXX 100644 --- a/hw/scsi/virtio-scsi.c +++ b/hw/scsi/virtio-scsi.c @@ -XXX,XX +XXX,XX @@ typedef struct VirtIOSCSIReq { QEMUSGList qsgl; QEMUIOVector resp_iov; - union { - /* Used for two-stage request submission */ - QTAILQ_ENTRY(VirtIOSCSIReq) next; + /* Used for two-stage request submission and TMFs deferred to BH */ + QTAILQ_ENTRY(VirtIOSCSIReq) next; - /* Used for cancellation of request during TMFs */ - int remaining; - }; + /* Used for cancellation of request during TMFs */ + int remaining; SCSIRequest *sreq; size_t resp_size; @@ -XXX,XX +XXX,XX @@ static inline void virtio_scsi_ctx_check(VirtIOSCSI *s, SCSIDevice *d) } } +static void virtio_scsi_do_one_tmf_bh(VirtIOSCSIReq *req) +{ + VirtIOSCSI *s = req->dev; + SCSIDevice *d = virtio_scsi_device_get(s, req->req.tmf.lun); + BusChild *kid; + int target; + + switch (req->req.tmf.subtype) { + case VIRTIO_SCSI_T_TMF_LOGICAL_UNIT_RESET: + if (!d) { + req->resp.tmf.response = VIRTIO_SCSI_S_BAD_TARGET; + goto out; + } + if (d->lun != virtio_scsi_get_lun(req->req.tmf.lun)) { + req->resp.tmf.response = VIRTIO_SCSI_S_INCORRECT_LUN; + goto out; + } + qatomic_inc(&s->resetting); + device_cold_reset(&d->qdev); + qatomic_dec(&s->resetting); + break; + + case VIRTIO_SCSI_T_TMF_I_T_NEXUS_RESET: + target = req->req.tmf.lun[1]; + qatomic_inc(&s->resetting); + + rcu_read_lock(); + QTAILQ_FOREACH_RCU(kid, &s->bus.qbus.children, sibling) { + SCSIDevice *d1 = SCSI_DEVICE(kid->child); + if (d1->channel == 0 && d1->id == target) { + device_cold_reset(&d1->qdev); + } + } + rcu_read_unlock(); + + qatomic_dec(&s->resetting); + break; + + default: + g_assert_not_reached(); + break; + } + +out: + object_unref(OBJECT(d)); + + virtio_scsi_acquire(s); + virtio_scsi_complete_req(req); + virtio_scsi_release(s); +} + +/* Some TMFs must be processed from the main loop thread */ +static void virtio_scsi_do_tmf_bh(void *opaque) +{ + VirtIOSCSI *s = opaque; + QTAILQ_HEAD(, VirtIOSCSIReq) reqs = QTAILQ_HEAD_INITIALIZER(reqs); + VirtIOSCSIReq *req; + VirtIOSCSIReq *tmp; + + GLOBAL_STATE_CODE(); + + virtio_scsi_acquire(s); + + QTAILQ_FOREACH_SAFE(req, &s->tmf_bh_list, next, tmp) { + QTAILQ_REMOVE(&s->tmf_bh_list, req, next); + QTAILQ_INSERT_TAIL(&reqs, req, next); + } + + qemu_bh_delete(s->tmf_bh); + s->tmf_bh = NULL; + + virtio_scsi_release(s); + + QTAILQ_FOREACH_SAFE(req, &reqs, next, tmp) { + QTAILQ_REMOVE(&reqs, req, next); + virtio_scsi_do_one_tmf_bh(req); + } +} + +static void virtio_scsi_reset_tmf_bh(VirtIOSCSI *s) +{ + VirtIOSCSIReq *req; + VirtIOSCSIReq *tmp; + + GLOBAL_STATE_CODE(); + + virtio_scsi_acquire(s); + + if (s->tmf_bh) { + qemu_bh_delete(s->tmf_bh); + s->tmf_bh = NULL; + } + + QTAILQ_FOREACH_SAFE(req, &s->tmf_bh_list, next, tmp) { + QTAILQ_REMOVE(&s->tmf_bh_list, req, next); + + /* SAM-6 6.3.2 Hard reset */ + req->resp.tmf.response = VIRTIO_SCSI_S_TARGET_FAILURE; + virtio_scsi_complete_req(req); + } + + virtio_scsi_release(s); +} + +static void virtio_scsi_defer_tmf_to_bh(VirtIOSCSIReq *req) +{ + VirtIOSCSI *s = req->dev; + + QTAILQ_INSERT_TAIL(&s->tmf_bh_list, req, next); + + if (!s->tmf_bh) { + s->tmf_bh = qemu_bh_new(virtio_scsi_do_tmf_bh, s); + qemu_bh_schedule(s->tmf_bh); + } +} + /* Return 0 if the request is ready to be completed and return to guest; * -EINPROGRESS if the request is submitted and will be completed later, in the * case of async cancellation. */ @@ -XXX,XX +XXX,XX @@ static int virtio_scsi_do_tmf(VirtIOSCSI *s, VirtIOSCSIReq *req) { SCSIDevice *d = virtio_scsi_device_get(s, req->req.tmf.lun); SCSIRequest *r, *next; - BusChild *kid; - int target; int ret = 0; virtio_scsi_ctx_check(s, d); @@ -XXX,XX +XXX,XX @@ static int virtio_scsi_do_tmf(VirtIOSCSI *s, VirtIOSCSIReq *req) break; case VIRTIO_SCSI_T_TMF_LOGICAL_UNIT_RESET: - if (!d) { - goto fail; - } - if (d->lun != virtio_scsi_get_lun(req->req.tmf.lun)) { - goto incorrect_lun; - } - s->resetting++; - device_cold_reset(&d->qdev); - s->resetting--; + case VIRTIO_SCSI_T_TMF_I_T_NEXUS_RESET: + virtio_scsi_defer_tmf_to_bh(req); + ret = -EINPROGRESS; break; case VIRTIO_SCSI_T_TMF_ABORT_TASK_SET: @@ -XXX,XX +XXX,XX @@ static int virtio_scsi_do_tmf(VirtIOSCSI *s, VirtIOSCSIReq *req) } break; - case VIRTIO_SCSI_T_TMF_I_T_NEXUS_RESET: - target = req->req.tmf.lun[1]; - s->resetting++; - - rcu_read_lock(); - QTAILQ_FOREACH_RCU(kid, &s->bus.qbus.children, sibling) { - SCSIDevice *d1 = SCSI_DEVICE(kid->child); - if (d1->channel == 0 && d1->id == target) { - device_cold_reset(&d1->qdev); - } - } - rcu_read_unlock(); - - s->resetting--; - break; - case VIRTIO_SCSI_T_TMF_CLEAR_ACA: default: req->resp.tmf.response = VIRTIO_SCSI_S_FUNCTION_REJECTED; @@ -XXX,XX +XXX,XX @@ static void virtio_scsi_request_cancelled(SCSIRequest *r) if (!req) { return; } - if (req->dev->resetting) { + if (qatomic_read(&req->dev->resetting)) { req->resp.cmd.response = VIRTIO_SCSI_S_RESET; } else { req->resp.cmd.response = VIRTIO_SCSI_S_ABORTED; @@ -XXX,XX +XXX,XX @@ static void virtio_scsi_reset(VirtIODevice *vdev) VirtIOSCSICommon *vs = VIRTIO_SCSI_COMMON(vdev); assert(!s->dataplane_started); - s->resetting++; + + virtio_scsi_reset_tmf_bh(s); + + qatomic_inc(&s->resetting); bus_cold_reset(BUS(&s->bus)); - s->resetting--; + qatomic_dec(&s->resetting); vs->sense_size = VIRTIO_SCSI_SENSE_DEFAULT_SIZE; vs->cdb_size = VIRTIO_SCSI_CDB_DEFAULT_SIZE; @@ -XXX,XX +XXX,XX @@ static void virtio_scsi_device_realize(DeviceState *dev, Error **errp) VirtIOSCSI *s = VIRTIO_SCSI(dev); Error *err = NULL; + QTAILQ_INIT(&s->tmf_bh_list); + virtio_scsi_common_realize(dev, virtio_scsi_handle_ctrl, virtio_scsi_handle_event, @@ -XXX,XX +XXX,XX @@ static void virtio_scsi_device_unrealize(DeviceState *dev) { VirtIOSCSI *s = VIRTIO_SCSI(dev); + virtio_scsi_reset_tmf_bh(s); + qbus_set_hotplug_handler(BUS(&s->bus), NULL); virtio_scsi_common_unrealize(dev); } -- 2.39.2 From: Or Ozeri <oro@il.ibm.com> Signed-off-by: Or Ozeri <oro@il.ibm.com> Message-Id: <20230129113120.722708-2-oro@oro.sl.cloud9.ibm.com> Reviewed-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/rbd.c | 16 ++++++---------- 1 file changed, 6 insertions(+), 10 deletions(-) diff --git a/block/rbd.c b/block/rbd.c index XXXXXXX..XXXXXXX 100644 --- a/block/rbd.c +++ b/block/rbd.c @@ -XXX,XX +XXX,XX @@ static int qemu_rbd_encryption_format(rbd_image_t image, { int r = 0; g_autofree char *passphrase = NULL; - size_t passphrase_len; rbd_encryption_format_t format; rbd_encryption_options_t opts; rbd_encryption_luks1_format_options_t luks_opts; @@ -XXX,XX +XXX,XX @@ static int qemu_rbd_encryption_format(rbd_image_t image, opts_size = sizeof(luks_opts); r = qemu_rbd_convert_luks_create_options( qapi_RbdEncryptionCreateOptionsLUKS_base(&encrypt->u.luks), - &luks_opts.alg, &passphrase, &passphrase_len, errp); + &luks_opts.alg, &passphrase, &luks_opts.passphrase_size, + errp); if (r < 0) { return r; } luks_opts.passphrase = passphrase; - luks_opts.passphrase_size = passphrase_len; break; } case RBD_IMAGE_ENCRYPTION_FORMAT_LUKS2: { @@ -XXX,XX +XXX,XX @@ static int qemu_rbd_encryption_format(rbd_image_t image, r = qemu_rbd_convert_luks_create_options( qapi_RbdEncryptionCreateOptionsLUKS2_base( &encrypt->u.luks2), - &luks2_opts.alg, &passphrase, &passphrase_len, errp); + &luks2_opts.alg, &passphrase, &luks2_opts.passphrase_size, + errp); if (r < 0) { return r; } luks2_opts.passphrase = passphrase; - luks2_opts.passphrase_size = passphrase_len; break; } default: { @@ -XXX,XX +XXX,XX @@ static int qemu_rbd_encryption_load(rbd_image_t image, { int r = 0; g_autofree char *passphrase = NULL; - size_t passphrase_len; rbd_encryption_luks1_format_options_t luks_opts; rbd_encryption_luks2_format_options_t luks2_opts; rbd_encryption_format_t format; @@ -XXX,XX +XXX,XX @@ static int qemu_rbd_encryption_load(rbd_image_t image, opts_size = sizeof(luks_opts); r = qemu_rbd_convert_luks_options( qapi_RbdEncryptionOptionsLUKS_base(&encrypt->u.luks), - &passphrase, &passphrase_len, errp); + &passphrase, &luks_opts.passphrase_size, errp); if (r < 0) { return r; } luks_opts.passphrase = passphrase; - luks_opts.passphrase_size = passphrase_len; break; } case RBD_IMAGE_ENCRYPTION_FORMAT_LUKS2: { @@ -XXX,XX +XXX,XX @@ static int qemu_rbd_encryption_load(rbd_image_t image, opts_size = sizeof(luks2_opts); r = qemu_rbd_convert_luks_options( qapi_RbdEncryptionOptionsLUKS2_base(&encrypt->u.luks2), - &passphrase, &passphrase_len, errp); + &passphrase, &luks2_opts.passphrase_size, errp); if (r < 0) { return r; } luks2_opts.passphrase = passphrase; - luks2_opts.passphrase_size = passphrase_len; break; } default: { -- 2.39.2 From: Or Ozeri <oro@il.ibm.com> Ceph RBD encryption API required specifying the encryption format for loading encryption. The supported formats were LUKS (v1) and LUKS2. Starting from Reef release, RBD also supports loading with "luks-any" format, which works for both versions of LUKS. This commit extends the qemu rbd driver API to enable qemu users to use this luks-any wildcard format. Signed-off-by: Or Ozeri <oro@il.ibm.com> Message-Id: <20230129113120.722708-3-oro@oro.sl.cloud9.ibm.com> Reviewed-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- qapi/block-core.json | 16 ++++++++++++++-- block/rbd.c | 19 +++++++++++++++++++ 2 files changed, 33 insertions(+), 2 deletions(-) diff --git a/qapi/block-core.json b/qapi/block-core.json index XXXXXXX..XXXXXXX 100644 --- a/qapi/block-core.json +++ b/qapi/block-core.json @@ -XXX,XX +XXX,XX @@ ## # @RbdImageEncryptionFormat: # +# @luks-any: Used for opening either luks or luks2 (Since 8.0) +# # Since: 6.1 ## { 'enum': 'RbdImageEncryptionFormat', - 'data': [ 'luks', 'luks2' ] } + 'data': [ 'luks', 'luks2', 'luks-any' ] } ## # @RbdEncryptionOptionsLUKSBase: @@ -XXX,XX +XXX,XX @@ 'base': 'RbdEncryptionOptionsLUKSBase', 'data': { } } +## +# @RbdEncryptionOptionsLUKSAny: +# +# Since: 8.0 +## +{ 'struct': 'RbdEncryptionOptionsLUKSAny', + 'base': 'RbdEncryptionOptionsLUKSBase', + 'data': { } } + ## # @RbdEncryptionCreateOptionsLUKS: # @@ -XXX,XX +XXX,XX @@ 'base': { 'format': 'RbdImageEncryptionFormat' }, 'discriminator': 'format', 'data': { 'luks': 'RbdEncryptionOptionsLUKS', - 'luks2': 'RbdEncryptionOptionsLUKS2' } } + 'luks2': 'RbdEncryptionOptionsLUKS2', + 'luks-any': 'RbdEncryptionOptionsLUKSAny'} } ## # @RbdEncryptionCreateOptions: diff --git a/block/rbd.c b/block/rbd.c index XXXXXXX..XXXXXXX 100644 --- a/block/rbd.c +++ b/block/rbd.c @@ -XXX,XX +XXX,XX @@ static int qemu_rbd_encryption_load(rbd_image_t image, g_autofree char *passphrase = NULL; rbd_encryption_luks1_format_options_t luks_opts; rbd_encryption_luks2_format_options_t luks2_opts; +#ifdef LIBRBD_SUPPORTS_ENCRYPTION_LOAD2 + rbd_encryption_luks_format_options_t luks_any_opts; +#endif rbd_encryption_format_t format; rbd_encryption_options_t opts; size_t opts_size; @@ -XXX,XX +XXX,XX @@ static int qemu_rbd_encryption_load(rbd_image_t image, luks2_opts.passphrase = passphrase; break; } +#ifdef LIBRBD_SUPPORTS_ENCRYPTION_LOAD2 + case RBD_IMAGE_ENCRYPTION_FORMAT_LUKS_ANY: { + memset(&luks_any_opts, 0, sizeof(luks_any_opts)); + format = RBD_ENCRYPTION_FORMAT_LUKS; + opts = &luks_any_opts; + opts_size = sizeof(luks_any_opts); + r = qemu_rbd_convert_luks_options( + qapi_RbdEncryptionOptionsLUKSAny_base(&encrypt->u.luks_any), + &passphrase, &luks_any_opts.passphrase_size, errp); + if (r < 0) { + return r; + } + luks_any_opts.passphrase = passphrase; + break; + } +#endif default: { r = -ENOTSUP; error_setg_errno( -- 2.39.2 From: Or Ozeri <oro@il.ibm.com> Starting from ceph Reef, RBD has built-in support for layered encryption, where each ancestor image (in a cloned image setting) can be possibly encrypted using a unique passphrase. A new function, rbd_encryption_load2, was added to librbd API. This new function supports an array of passphrases (via "spec" structs). This commit extends the qemu rbd driver API to use this new librbd API, in order to support this new layered encryption feature. Signed-off-by: Or Ozeri <oro@il.ibm.com> Message-Id: <20230129113120.722708-4-oro@oro.sl.cloud9.ibm.com> Reviewed-by: Ilya Dryomov <idryomov@gmail.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- qapi/block-core.json | 11 +++- block/rbd.c | 153 ++++++++++++++++++++++++++++++++++++++++++- 2 files changed, 162 insertions(+), 2 deletions(-) diff --git a/qapi/block-core.json b/qapi/block-core.json index XXXXXXX..XXXXXXX 100644 --- a/qapi/block-core.json +++ b/qapi/block-core.json @@ -XXX,XX +XXX,XX @@ ## # @RbdEncryptionOptions: # +# @format: Encryption format. +# +# @parent: Parent image encryption options (for cloned images). +# Can be left unspecified if this cloned image is encrypted +# using the same format and secret as its parent image (i.e. +# not explicitly formatted) or if its parent image is not +# encrypted. (Since 8.0) +# # Since: 6.1 ## { 'union': 'RbdEncryptionOptions', - 'base': { 'format': 'RbdImageEncryptionFormat' }, + 'base': { 'format': 'RbdImageEncryptionFormat', + '*parent': 'RbdEncryptionOptions' }, 'discriminator': 'format', 'data': { 'luks': 'RbdEncryptionOptionsLUKS', 'luks2': 'RbdEncryptionOptionsLUKS2', diff --git a/block/rbd.c b/block/rbd.c index XXXXXXX..XXXXXXX 100644 --- a/block/rbd.c +++ b/block/rbd.c @@ -XXX,XX +XXX,XX @@ static const char rbd_luks2_header_verification[ 'L', 'U', 'K', 'S', 0xBA, 0xBE, 0, 2 }; +static const char rbd_layered_luks_header_verification[ + RBD_ENCRYPTION_LUKS_HEADER_VERIFICATION_LEN] = { + 'R', 'B', 'D', 'L', 0xBA, 0xBE, 0, 1 +}; + +static const char rbd_layered_luks2_header_verification[ + RBD_ENCRYPTION_LUKS_HEADER_VERIFICATION_LEN] = { + 'R', 'B', 'D', 'L', 0xBA, 0xBE, 0, 2 +}; + typedef enum { RBD_AIO_READ, RBD_AIO_WRITE, @@ -XXX,XX +XXX,XX @@ static int qemu_rbd_encryption_load(rbd_image_t image, return 0; } + +#ifdef LIBRBD_SUPPORTS_ENCRYPTION_LOAD2 +static int qemu_rbd_encryption_load2(rbd_image_t image, + RbdEncryptionOptions *encrypt, + Error **errp) +{ + int r = 0; + int encrypt_count = 1; + int i; + RbdEncryptionOptions *curr_encrypt; + rbd_encryption_spec_t *specs; + rbd_encryption_luks1_format_options_t *luks_opts; + rbd_encryption_luks2_format_options_t *luks2_opts; + rbd_encryption_luks_format_options_t *luks_any_opts; + + /* count encryption options */ + for (curr_encrypt = encrypt->parent; curr_encrypt; + curr_encrypt = curr_encrypt->parent) { + ++encrypt_count; + } + + specs = g_new0(rbd_encryption_spec_t, encrypt_count); + + curr_encrypt = encrypt; + for (i = 0; i < encrypt_count; ++i) { + switch (curr_encrypt->format) { + case RBD_IMAGE_ENCRYPTION_FORMAT_LUKS: { + specs[i].format = RBD_ENCRYPTION_FORMAT_LUKS1; + + luks_opts = g_new0(rbd_encryption_luks1_format_options_t, 1); + specs[i].opts = luks_opts; + specs[i].opts_size = sizeof(*luks_opts); + + r = qemu_rbd_convert_luks_options( + qapi_RbdEncryptionOptionsLUKS_base( + &curr_encrypt->u.luks), + (char **)&luks_opts->passphrase, + &luks_opts->passphrase_size, + errp); + break; + } + case RBD_IMAGE_ENCRYPTION_FORMAT_LUKS2: { + specs[i].format = RBD_ENCRYPTION_FORMAT_LUKS2; + + luks2_opts = g_new0(rbd_encryption_luks2_format_options_t, 1); + specs[i].opts = luks2_opts; + specs[i].opts_size = sizeof(*luks2_opts); + + r = qemu_rbd_convert_luks_options( + qapi_RbdEncryptionOptionsLUKS2_base( + &curr_encrypt->u.luks2), + (char **)&luks2_opts->passphrase, + &luks2_opts->passphrase_size, + errp); + break; + } + case RBD_IMAGE_ENCRYPTION_FORMAT_LUKS_ANY: { + specs[i].format = RBD_ENCRYPTION_FORMAT_LUKS; + + luks_any_opts = g_new0(rbd_encryption_luks_format_options_t, 1); + specs[i].opts = luks_any_opts; + specs[i].opts_size = sizeof(*luks_any_opts); + + r = qemu_rbd_convert_luks_options( + qapi_RbdEncryptionOptionsLUKSAny_base( + &curr_encrypt->u.luks_any), + (char **)&luks_any_opts->passphrase, + &luks_any_opts->passphrase_size, + errp); + break; + } + default: { + r = -ENOTSUP; + error_setg_errno( + errp, -r, "unknown image encryption format: %u", + curr_encrypt->format); + } + } + + if (r < 0) { + goto exit; + } + + curr_encrypt = curr_encrypt->parent; + } + + r = rbd_encryption_load2(image, specs, encrypt_count); + if (r < 0) { + error_setg_errno(errp, -r, "layered encryption load fail"); + goto exit; + } + +exit: + for (i = 0; i < encrypt_count; ++i) { + if (!specs[i].opts) { + break; + } + + switch (specs[i].format) { + case RBD_ENCRYPTION_FORMAT_LUKS1: { + luks_opts = specs[i].opts; + g_free((void *)luks_opts->passphrase); + break; + } + case RBD_ENCRYPTION_FORMAT_LUKS2: { + luks2_opts = specs[i].opts; + g_free((void *)luks2_opts->passphrase); + break; + } + case RBD_ENCRYPTION_FORMAT_LUKS: { + luks_any_opts = specs[i].opts; + g_free((void *)luks_any_opts->passphrase); + break; + } + } + + g_free(specs[i].opts); + } + g_free(specs); + return r; +} +#endif #endif /* FIXME Deprecate and remove keypairs or make it available in QMP. */ @@ -XXX,XX +XXX,XX @@ static int qemu_rbd_open(BlockDriverState *bs, QDict *options, int flags, if (opts->encrypt) { #ifdef LIBRBD_SUPPORTS_ENCRYPTION - r = qemu_rbd_encryption_load(s->image, opts->encrypt, errp); + if (opts->encrypt->parent) { +#ifdef LIBRBD_SUPPORTS_ENCRYPTION_LOAD2 + r = qemu_rbd_encryption_load2(s->image, opts->encrypt, errp); +#else + r = -ENOTSUP; + error_setg(errp, "RBD library does not support layered encryption"); +#endif + } else { + r = qemu_rbd_encryption_load(s->image, opts->encrypt, errp); + } if (r < 0) { goto failed_post_open; } @@ -XXX,XX +XXX,XX @@ static ImageInfoSpecific *qemu_rbd_get_specific_info(BlockDriverState *bs, spec_info->u.rbd.data->encryption_format = RBD_IMAGE_ENCRYPTION_FORMAT_LUKS2; spec_info->u.rbd.data->has_encryption_format = true; + } else if (memcmp(buf, rbd_layered_luks_header_verification, + RBD_ENCRYPTION_LUKS_HEADER_VERIFICATION_LEN) == 0) { + spec_info->u.rbd.data->encryption_format = + RBD_IMAGE_ENCRYPTION_FORMAT_LUKS; + spec_info->u.rbd.data->has_encryption_format = true; + } else if (memcmp(buf, rbd_layered_luks2_header_verification, + RBD_ENCRYPTION_LUKS_HEADER_VERIFICATION_LEN) == 0) { + spec_info->u.rbd.data->encryption_format = + RBD_IMAGE_ENCRYPTION_FORMAT_LUKS2; + spec_info->u.rbd.data->has_encryption_format = true; } else { spec_info->u.rbd.data->has_encryption_format = false; } -- 2.39.2 The following changes since commit 239b8b0699a222fd21da1c5fdeba0a2456085a47: Merge tag 'trivial-branch-for-8.0-pull-request' of https://gitlab.com/laurent_vivier/qemu into staging (2023-01-19 15:05:29 +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 4711b0a8490827c332b3f9281f689ce9555b7fab: qemu-img: Change info key names for protocol nodes (2023-01-20 13:11:01 +0100) ---------------------------------------------------------------- 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/coroutine.h | 43 +++ 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 | 95 ++++++ tests/qemu-iotests/tests/qemu-img-close-errors.out | 23 ++ 69 files changed, 1207 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 From: Thomas Huth <thuth@redhat.com> "quorum" is required by iotest 312 - if it is not compiled into the QEMU binary, the test fails. Thus list "quorum" as required driver so that the test gets skipped in case it is not available. Signed-off-by: Thomas Huth <thuth@redhat.com> Message-Id: <20230104114601.269351-1-thuth@redhat.com> Reviewed-by: Alberto Garcia <berto@igalia.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- tests/qemu-iotests/312 | 1 + 1 file changed, 1 insertion(+) diff --git a/tests/qemu-iotests/312 b/tests/qemu-iotests/312 index XXXXXXX..XXXXXXX 100755 --- a/tests/qemu-iotests/312 +++ b/tests/qemu-iotests/312 @@ -XXX,XX +XXX,XX @@ _supported_fmt qcow2 _supported_proto file _supported_os Linux _unsupported_imgopts cluster_size data_file +_require_drivers quorum echo echo '### Create all images' # three source (quorum), one destination -- 2.38.1 From: Thomas Huth <thuth@redhat.com> In downstream RHEL builds, we do not have "blkverify" enabled, so iotest 262 is currently failing there. Thus let's list "blkverify" as required item so that the test properly gets skipped instead if "blkverify" is missing. Signed-off-by: Thomas Huth <thuth@redhat.com> Message-Id: <20230104112850.261480-1-thuth@redhat.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- tests/qemu-iotests/262 | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/tests/qemu-iotests/262 b/tests/qemu-iotests/262 index XXXXXXX..XXXXXXX 100755 --- a/tests/qemu-iotests/262 +++ b/tests/qemu-iotests/262 @@ -XXX,XX +XXX,XX @@ import iotests import os iotests.script_initialize(supported_fmts=['qcow2'], - supported_platforms=['linux']) + supported_platforms=['linux'], + required_fmts=['blkverify']) with iotests.FilePath('img') as img_path, \ iotests.FilePath('mig_fifo') as fifo, \ -- 2.38.1 From: Xiang Zheng <zhengxiang9@huawei.com> Currently we fill the VIRT_FLASH memory space with two 64MB NOR images when using persistent UEFI variables on virt board. Actually we only use a very small(non-zero) part of the memory while the rest significant large(zero) part of memory is wasted. So this patch checks the block status and only writes the non-zero part into memory. This requires pflash devices to use sparse files for backends. Signed-off-by: Xiang Zheng <zhengxiang9@huawei.com> [ kraxel: rebased to latest master ] Signed-off-by: Gerd Hoffmann <kraxel@redhat.com> Message-Id: <20221220084246.1984871-1-kraxel@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- hw/block/block.c | 36 +++++++++++++++++++++++++++++++++++- 1 file changed, 35 insertions(+), 1 deletion(-) diff --git a/hw/block/block.c b/hw/block/block.c index XXXXXXX..XXXXXXX 100644 --- a/hw/block/block.c +++ b/hw/block/block.c @@ -XXX,XX +XXX,XX @@ #include "qapi/error.h" #include "qapi/qapi-types-block.h" +/* + * Read the non-zeroes parts of @blk into @buf + * Reading all of the @blk is expensive if the zeroes parts of @blk + * is large enough. Therefore check the block status and only write + * the non-zeroes block into @buf. + * + * Return 0 on success, non-zero on error. + */ +static int blk_pread_nonzeroes(BlockBackend *blk, hwaddr size, void *buf) +{ + int ret; + int64_t bytes, offset = 0; + BlockDriverState *bs = blk_bs(blk); + + for (;;) { + bytes = MIN(size - offset, BDRV_REQUEST_MAX_SECTORS); + if (bytes <= 0) { + return 0; + } + ret = bdrv_block_status(bs, offset, bytes, &bytes, NULL, NULL); + if (ret < 0) { + return ret; + } + if (!(ret & BDRV_BLOCK_ZERO)) { + ret = bdrv_pread(bs->file, offset, bytes, + (uint8_t *) buf + offset, 0); + if (ret < 0) { + return ret; + } + } + offset += bytes; + } +} + /* * Read the entire contents of @blk into @buf. * @blk's contents must be @size bytes, and @size must be at most @@ -XXX,XX +XXX,XX @@ bool blk_check_size_and_read_all(BlockBackend *blk, void *buf, hwaddr size, * block device and read only on demand. */ assert(size <= BDRV_REQUEST_MAX_BYTES); - ret = blk_pread(blk, 0, size, buf, 0); + ret = blk_pread_nonzeroes(blk, size, buf); if (ret < 0) { error_setg_errno(errp, -ret, "can't read block backend"); return false; -- 2.38.1 From: Alberto Faria <afaria@redhat.com> Clang has a generic __annotate__ attribute that can be used by static analyzers to understand properties of functions and analyze the control flow. Furthermore, unlike TSA annotations, the __annotate__ attribute applies to function pointers as well. As a first step towards static analysis of coroutine_fn markers, attach the attribute to the marker when compiling with clang. Signed-off-by: Alberto Faria <afaria@redhat.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Message-Id: <20221216110758.559947-2-pbonzini@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/qemu/coroutine.h | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/include/qemu/coroutine.h b/include/qemu/coroutine.h index XXXXXXX..XXXXXXX 100644 --- a/include/qemu/coroutine.h +++ b/include/qemu/coroutine.h @@ -XXX,XX +XXX,XX @@ * .... * } */ +#ifdef __clang__ +#define coroutine_fn __attribute__((__annotate__("coroutine_fn"))) +#else #define coroutine_fn +#endif typedef struct Coroutine Coroutine; -- 2.38.1 From: Alberto Faria <afaria@redhat.com> Add more annotations to functions, describing valid and invalid calls from coroutine to non-coroutine context. When applied to a function, no_coroutine_fn advertises that it should not be called from coroutine_fn functions. This can be because the function blocks or, in the case of generated_co_wrapper, to enforce that coroutine_fn functions directly call the coroutine_fn that backs the generated_co_wrapper. coroutine_mixed_fn instead is for function that can be called in both coroutine and non-coroutine context, but will suspend when called in coroutine context. Annotating them is a first step towards enforcing that non-annotated functions are absolutely not going to suspend. These can be used for example with the vrc tool: # find functions that *really* cannot be called from no_coroutine_fn (vrc) load --loader clang libblock.fa.p/meson-generated_.._block_block-gen.c.o (vrc) paths [no_coroutine_fn,!coroutine_mixed_fn] bdrv_remove_persistent_dirty_bitmap bdrv_create bdrv_can_store_new_dirty_bitmap # find how coroutine_fns end up calling a mixed function (vrc) load --loader clang --force libblock.fa.p/*.c.o (vrc) paths [coroutine_fn] [!no_coroutine_fn]* [coroutine_mixed_fn] ... bdrv_pread <- vhdx_log_write <- vhdx_log_write_and_flush <- vhdx_co_writev ... Signed-off-by: Alberto Faria <afaria@redhat.com> [Rebase, add coroutine_mixed_fn. - Paolo] Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Message-Id: <20221216110758.559947-3-pbonzini@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-common.h | 11 ++++++---- include/qemu/coroutine.h | 39 ++++++++++++++++++++++++++++++++++++ 2 files changed, 46 insertions(+), 4 deletions(-) diff --git a/include/block/block-common.h b/include/block/block-common.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block-common.h +++ b/include/block/block-common.h @@ -XXX,XX +XXX,XX @@ * - co_wrapper_mixed_bdrv_rdlock are co_wrapper_mixed functions but * automatically take and release the graph rdlock when creating a new * coroutine. + * + * These functions should not be called from a coroutine_fn; instead, + * call the wrapped function directly. */ -#define co_wrapper -#define co_wrapper_mixed -#define co_wrapper_bdrv_rdlock -#define co_wrapper_mixed_bdrv_rdlock +#define co_wrapper no_coroutine_fn +#define co_wrapper_mixed no_coroutine_fn coroutine_mixed_fn +#define co_wrapper_bdrv_rdlock no_coroutine_fn +#define co_wrapper_mixed_bdrv_rdlock no_coroutine_fn coroutine_mixed_fn #include "block/dirty-bitmap.h" #include "block/blockjob.h" diff --git a/include/qemu/coroutine.h b/include/qemu/coroutine.h index XXXXXXX..XXXXXXX 100644 --- a/include/qemu/coroutine.h +++ b/include/qemu/coroutine.h @@ -XXX,XX +XXX,XX @@ #define coroutine_fn #endif +/** + * Mark a function that can suspend when executed in coroutine context, + * but can handle running in non-coroutine context too. + */ +#ifdef __clang__ +#define coroutine_mixed_fn __attribute__((__annotate__("coroutine_mixed_fn"))) +#else +#define coroutine_mixed_fn +#endif + +/** + * Mark a function that should not be called from a coroutine context. + * Usually there will be an analogous, coroutine_fn function that should + * be used instead. + * + * When the function is also marked as coroutine_mixed_fn, the function should + * only be called if the caller does not know whether it is in coroutine + * context. + * + * Functions that are only no_coroutine_fn, on the other hand, should not + * be called from within coroutines at all. This for example includes + * functions that block. + * + * In the future it would be nice to enable compiler or static checker + * support for catching such errors. This annotation is the first step + * towards this, and in the meantime it serves as documentation. + * + * For example: + * + * static void no_coroutine_fn foo(void) { + * .... + * } + */ +#ifdef __clang__ +#define no_coroutine_fn __attribute__((__annotate__("no_coroutine_fn"))) +#else +#define no_coroutine_fn +#endif + typedef struct Coroutine Coroutine; /** -- 2.38.1 From: Paolo Bonzini <pbonzini@redhat.com> qemu-io's do_co_pwrite_zeroes is reinventing the coroutine wrapper blk_pwrite_zeroes. Just use the real thing directly. Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Message-Id: <20221215130225.476477-1-pbonzini@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- qemu-io-cmds.c | 57 +++++++++----------------------------------------- 1 file changed, 10 insertions(+), 47 deletions(-) diff --git a/qemu-io-cmds.c b/qemu-io-cmds.c index XXXXXXX..XXXXXXX 100644 --- a/qemu-io-cmds.c +++ b/qemu-io-cmds.c @@ -XXX,XX +XXX,XX @@ static int do_pwrite(BlockBackend *blk, char *buf, int64_t offset, return 1; } -typedef struct { - BlockBackend *blk; - int64_t offset; - int64_t bytes; - int64_t *total; - int flags; - int ret; - bool done; -} CoWriteZeroes; - -static void coroutine_fn co_pwrite_zeroes_entry(void *opaque) -{ - CoWriteZeroes *data = opaque; - - data->ret = blk_co_pwrite_zeroes(data->blk, data->offset, data->bytes, - data->flags); - data->done = true; - if (data->ret < 0) { - *data->total = data->ret; - return; - } - - *data->total = data->bytes; -} - -static int do_co_pwrite_zeroes(BlockBackend *blk, int64_t offset, +static int do_pwrite_zeroes(BlockBackend *blk, int64_t offset, int64_t bytes, int flags, int64_t *total) { - Coroutine *co; - CoWriteZeroes data = { - .blk = blk, - .offset = offset, - .bytes = bytes, - .total = total, - .flags = flags, - .done = false, - }; - - co = qemu_coroutine_create(co_pwrite_zeroes_entry, &data); - bdrv_coroutine_enter(blk_bs(blk), co); - while (!data.done) { - aio_poll(blk_get_aio_context(blk), true); - } - if (data.ret < 0) { - return data.ret; - } else { - return 1; + int ret = blk_pwrite_zeroes(blk, offset, bytes, + flags | BDRV_REQ_ZERO_WRITE); + + if (ret < 0) { + return ret; } + *total = bytes; + return 1; } static int do_write_compressed(BlockBackend *blk, char *buf, int64_t offset, @@ -XXX,XX +XXX,XX @@ static void write_help(void) " -C, -- report statistics in a machine parsable format\n" " -q, -- quiet mode, do not show I/O statistics\n" " -u, -- with -z, allow unmapping\n" -" -z, -- write zeroes using blk_co_pwrite_zeroes\n" +" -z, -- write zeroes using blk_pwrite_zeroes\n" "\n"); } @@ -XXX,XX +XXX,XX @@ static int write_f(BlockBackend *blk, int argc, char **argv) if (bflag) { ret = do_save_vmstate(blk, buf, offset, count, &total); } else if (zflag) { - ret = do_co_pwrite_zeroes(blk, offset, count, flags, &total); + ret = do_pwrite_zeroes(blk, offset, count, flags, &total); } else if (cflag) { ret = do_write_compressed(blk, buf, offset, count, &total); } else { -- 2.38.1 From: Paolo Bonzini <pbonzini@redhat.com> It has only one caller---inline it and remove the function. Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Message-Id: <20221215130225.476477-2-pbonzini@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-io.h | 5 ----- block.c | 6 ------ block/block-backend.c | 2 +- 3 files changed, 1 insertion(+), 12 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 @@ AioContext *coroutine_fn bdrv_co_enter(BlockDriverState *bs); */ void coroutine_fn bdrv_co_leave(BlockDriverState *bs, AioContext *old_ctx); -/** - * Transfer control to @co in the aio context of @bs - */ -void bdrv_coroutine_enter(BlockDriverState *bs, Coroutine *co); - AioContext *child_of_bds_get_parent_aio_context(BdrvChild *c); void bdrv_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 @@ void coroutine_fn bdrv_co_unlock(BlockDriverState *bs) } } -void bdrv_coroutine_enter(BlockDriverState *bs, Coroutine *co) -{ - IO_CODE(); - aio_co_enter(bdrv_get_aio_context(bs), co); -} - static void bdrv_do_remove_aio_context_notifier(BdrvAioNotifier *ban) { GLOBAL_STATE_CODE(); 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 @@ static BlockAIOCB *blk_aio_prwv(BlockBackend *blk, int64_t offset, acb->has_returned = false; co = qemu_coroutine_create(co_entry, acb); - bdrv_coroutine_enter(blk_bs(blk), co); + aio_co_enter(blk_get_aio_context(blk), co); acb->has_returned = true; if (acb->rwco.ret != NOT_DONE) { -- 2.38.1 In order to write the bitmap table to the image file, it is converted to big endian. If the write fails, it is passed to clear_bitmap_table() to free all of the clusters it had allocated before. However, if we don't convert it back to native endianness first, we'll free things at a wrong offset. In practical terms, the offsets will be so high that we won't actually free any allocated clusters, but just run into an error, but in theory this can cause image corruption. Cc: qemu-stable@nongnu.org Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230112191454.169353-2-kwolf@redhat.com> Reviewed-by: Hanna Czenczek <hreitz@redhat.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/qcow2-bitmap.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/block/qcow2-bitmap.c b/block/qcow2-bitmap.c index XXXXXXX..XXXXXXX 100644 --- a/block/qcow2-bitmap.c +++ b/block/qcow2-bitmap.c @@ -XXX,XX +XXX,XX @@ static int update_header_sync(BlockDriverState *bs) return bdrv_flush(bs->file->bs); } -static inline void bitmap_table_to_be(uint64_t *bitmap_table, size_t size) +static inline void bitmap_table_bswap_be(uint64_t *bitmap_table, size_t size) { size_t i; @@ -XXX,XX +XXX,XX @@ static int store_bitmap(BlockDriverState *bs, Qcow2Bitmap *bm, Error **errp) goto fail; } - bitmap_table_to_be(tb, tb_size); + bitmap_table_bswap_be(tb, tb_size); ret = bdrv_pwrite(bs->file, tb_offset, tb_size * sizeof(tb[0]), tb, 0); if (ret < 0) { + bitmap_table_bswap_be(tb, tb_size); error_setg_errno(errp, -ret, "Failed to write bitmap '%s' to file", bm_name); goto fail; -- 2.38.1 blk_unref() can't report any errors that happen while closing the image. For example, if qcow2 hits an -ENOSPC error while writing out dirty bitmaps when it's closed, it prints error messages to stderr, but 'qemu-img commit' won't see any error return value and will therefore look successful with exit code 0. In order to fix this, manually inactivate the image first before calling blk_unref(). This already performs the operations that would be most likely to fail while closing the image, but it can still return errors. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230112191454.169353-3-kwolf@redhat.com> Reviewed-by: Hanna Czenczek <hreitz@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- qemu-img.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) 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 BlockBackend *img_open(bool image_opts, blk = img_open_file(filename, NULL, fmt, flags, writethrough, quiet, force_share); } + + if (blk) { + blk_set_force_allow_inactivate(blk); + } + return blk; } @@ -XXX,XX +XXX,XX @@ unref_backing: done: qemu_progress_end(); + /* + * Manually inactivate the image first because this way we can know whether + * an error occurred. blk_unref() doesn't tell us about failures. + */ + ret = bdrv_inactivate_all(); + if (ret < 0 && !local_err) { + error_setg_errno(&local_err, -ret, "Error while closing the image"); + } blk_unref(blk); if (local_err) { -- 2.38.1 blk_unref() can't report any errors that happen while closing the image. For example, if qcow2 hits an -ENOSPC error while writing out dirty bitmaps when it's closed, it prints error messages to stderr, but 'qemu-img bitmap' won't see any error return value and will therefore look successful with exit code 0. In order to fix this, manually inactivate the image first before calling blk_unref(). This already performs the operations that would be most likely to fail while closing the image, but it can still return errors. Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1330 Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230112191454.169353-4-kwolf@redhat.com> Reviewed-by: Hanna Czenczek <hreitz@redhat.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- qemu-img.c | 11 +++++++++++ 1 file changed, 11 insertions(+) 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 int img_bitmap(int argc, char **argv) QSIMPLEQ_HEAD(, ImgBitmapAction) actions; ImgBitmapAction *act, *act_next; const char *op; + int inactivate_ret; QSIMPLEQ_INIT(&actions); @@ -XXX,XX +XXX,XX @@ static int img_bitmap(int argc, char **argv) ret = 0; out: + /* + * Manually inactivate the images first because this way we can know whether + * an error occurred. blk_unref() doesn't tell us about failures. + */ + inactivate_ret = bdrv_inactivate_all(); + if (inactivate_ret < 0) { + error_report("Error while closing the image: %s", strerror(-inactivate_ret)); + ret = 1; + } + blk_unref(src); blk_unref(blk); qemu_opts_del(opts); -- 2.38.1 This tests that when an error happens while writing back bitmaps to the image file in qcow2_inactivate(), 'qemu-img bitmap/commit' actually return an error value in their exit code instead of making the operation look successful to scripts. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230112191454.169353-5-kwolf@redhat.com> Reviewed-by: Hanna Czenczek <hreitz@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- .../qemu-iotests/tests/qemu-img-close-errors | 95 +++++++++++++++++++ .../tests/qemu-img-close-errors.out | 23 +++++ 2 files changed, 118 insertions(+) create mode 100755 tests/qemu-iotests/tests/qemu-img-close-errors create mode 100644 tests/qemu-iotests/tests/qemu-img-close-errors.out diff --git a/tests/qemu-iotests/tests/qemu-img-close-errors b/tests/qemu-iotests/tests/qemu-img-close-errors new file mode 100755 index XXXXXXX..XXXXXXX --- /dev/null +++ b/tests/qemu-iotests/tests/qemu-img-close-errors @@ -XXX,XX +XXX,XX @@ +#!/usr/bin/env bash +# group: rw auto quick +# +# Check that errors while closing the image, in particular writing back dirty +# bitmaps, is correctly reported with a failing qemu-img exit code. +# +# 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 + +_supported_fmt qcow2 +_supported_proto generic + +size=1G + +# The error we are going to use is ENOSPC. Depending on how many bitmaps we +# create in the backing file (and therefore increase the used up space), we get +# failures in different places. With a low number, only merging the bitmap +# fails, whereas with a higher number, already 'qemu-img commit' fails. +for max_bitmap in 6 7; do + echo + echo "=== Test with $max_bitmap bitmaps ===" + + TEST_IMG="$TEST_IMG.base" _make_test_img -q $size + for i in $(seq 1 $max_bitmap); do + $QEMU_IMG bitmap --add "$TEST_IMG.base" "stale-bitmap-$i" + done + + # Simulate a block device of 128 MB by resizing the image file accordingly + # and then enforcing the size with the raw driver + truncate "$TEST_IMG.base" --size 128M + BASE_JSON='json:{ + "driver": "qcow2", + "file": { + "driver": "raw", + "size": 134217728, + "file": { + "driver": "file", + "filename":"'"$TEST_IMG.base"'" + } + } + }' + + _make_test_img -q -b "$BASE_JSON" -F $IMGFMT + $QEMU_IMG bitmap --add "$TEST_IMG" "good-bitmap" + + $QEMU_IO -c 'write 0 126m' "$TEST_IMG" | _filter_qemu_io + + $QEMU_IMG commit -d "$TEST_IMG" 2>&1 | _filter_generated_node_ids + echo "qemu-img commit exit code: ${PIPESTATUS[0]}" + + $QEMU_IMG bitmap --add "$BASE_JSON" "good-bitmap" + echo "qemu-img bitmap --add exit code: $?" + + $QEMU_IMG bitmap --merge "good-bitmap" -b "$TEST_IMG" "$BASE_JSON" \ + "good-bitmap" 2>&1 | _filter_generated_node_ids + echo "qemu-img bitmap --merge exit code: ${PIPESTATUS[0]}" +done + +# success, all done +echo "*** done" +rm -f $seq.full +status=0 + diff --git a/tests/qemu-iotests/tests/qemu-img-close-errors.out b/tests/qemu-iotests/tests/qemu-img-close-errors.out new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/tests/qemu-iotests/tests/qemu-img-close-errors.out @@ -XXX,XX +XXX,XX @@ +QA output created by qemu-img-close-errors + +=== Test with 6 bitmaps === +wrote 132120576/132120576 bytes at offset 0 +126 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) +Image committed. +qemu-img commit exit code: 0 +qemu-img bitmap --add exit code: 0 +qemu-img: Lost persistent bitmaps during inactivation of node 'NODE_NAME': Failed to write bitmap 'good-bitmap' to file: No space left on device +qemu-img: Error while closing the image: Invalid argument +qemu-img: Lost persistent bitmaps during inactivation of node 'NODE_NAME': Failed to write bitmap 'good-bitmap' to file: No space left on device +qemu-img bitmap --merge exit code: 1 + +=== Test with 7 bitmaps === +wrote 132120576/132120576 bytes at offset 0 +126 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) +qemu-img: Lost persistent bitmaps during inactivation of node 'NODE_NAME': Failed to write bitmap 'stale-bitmap-7' to file: No space left on device +qemu-img: Lost persistent bitmaps during inactivation of node 'NODE_NAME': Failed to write bitmap 'stale-bitmap-7' to file: No space left on device +qemu-img: Error while closing the image: Invalid argument +qemu-img commit exit code: 1 +qemu-img bitmap --add exit code: 0 +qemu-img bitmap --merge exit code: 0 +*** done -- 2.38.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> Just omit the various 'return' when the return type is void. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230113204212.359076-2-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- scripts/block-coroutine-wrapper.py | 20 +++++++++++++++----- 1 file changed, 15 insertions(+), 5 deletions(-) diff --git a/scripts/block-coroutine-wrapper.py b/scripts/block-coroutine-wrapper.py index XXXXXXX..XXXXXXX 100644 --- a/scripts/block-coroutine-wrapper.py +++ b/scripts/block-coroutine-wrapper.py @@ -XXX,XX +XXX,XX @@ def __init__(self, return_type: str, name: str, args: str, ctx = 'qemu_get_aio_context()' self.ctx = ctx + self.get_result = 's->ret = ' + self.ret = 'return s.ret;' + self.co_ret = 'return ' + self.return_field = self.return_type + " ret;" + if self.return_type == 'void': + self.get_result = '' + self.ret = '' + self.co_ret = '' + self.return_field = '' + def gen_list(self, format: str) -> str: return ', '.join(format.format_map(arg.__dict__) for arg in self.args) @@ -XXX,XX +XXX,XX @@ def create_mixed_wrapper(func: FuncDecl) -> str: {{ if (qemu_in_coroutine()) {{ {graph_assume_lock} - return {name}({ func.gen_list('{name}') }); + {func.co_ret}{name}({ func.gen_list('{name}') }); }} else {{ {struct_name} s = {{ .poll_state.ctx = {func.ctx}, @@ -XXX,XX +XXX,XX @@ def create_mixed_wrapper(func: FuncDecl) -> str: s.poll_state.co = qemu_coroutine_create({name}_entry, &s); bdrv_poll_co(&s.poll_state); - return s.ret; + {func.ret} }} }}""" @@ -XXX,XX +XXX,XX @@ def create_co_wrapper(func: FuncDecl) -> str: s.poll_state.co = qemu_coroutine_create({name}_entry, &s); bdrv_poll_co(&s.poll_state); - return s.ret; + {func.ret} }}""" @@ -XXX,XX +XXX,XX @@ def gen_wrapper(func: FuncDecl) -> str: typedef struct {struct_name} {{ BdrvPollCo poll_state; - {func.return_type} ret; + {func.return_field} { func.gen_block(' {decl};') } }} {struct_name}; @@ -XXX,XX +XXX,XX @@ def gen_wrapper(func: FuncDecl) -> str: {struct_name} *s = opaque; {graph_lock} - s->ret = {name}({ func.gen_list('s->{name}') }); + {func.get_result}{name}({ func.gen_list('s->{name}') }); {graph_unlock} s->poll_state.in_progress = false; -- 2.38.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> BlockDriver->bdrv_io_plug is categorized as IO callback, and it currently doesn't run in a coroutine. We should let it take a graph rdlock since the callback traverses the block nodes graph, which however is only possible in a coroutine. The only caller of this function is blk_io_plug(), therefore make blk_io_plug() a co_wrapper, so that we're always running in a coroutine where the lock can be taken. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230113204212.359076-3-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-io.h | 3 ++- include/block/block_int-common.h | 2 +- include/sysemu/block-backend-io.h | 4 +++- block/block-backend.c | 4 ++-- block/file-posix.c | 10 +++++----- block/io.c | 8 ++++---- block/nvme.c | 4 ++-- 7 files changed, 19 insertions(+), 16 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 coroutine_fn bdrv_co_leave(BlockDriverState *bs, AioContext *old_ctx); AioContext *child_of_bds_get_parent_aio_context(BdrvChild *c); -void bdrv_io_plug(BlockDriverState *bs); +void coroutine_fn bdrv_co_io_plug(BlockDriverState *bs); + void bdrv_io_unplug(BlockDriverState *bs); bool coroutine_fn bdrv_co_can_store_new_dirty_bitmap(BlockDriverState *bs, 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_debug_event)(BlockDriverState *bs, BlkdebugEvent event); /* io queue for linux-aio */ - void (*bdrv_io_plug)(BlockDriverState *bs); + void coroutine_fn (*bdrv_co_io_plug)(BlockDriverState *bs); void (*bdrv_io_unplug)(BlockDriverState *bs); /** diff --git a/include/sysemu/block-backend-io.h b/include/sysemu/block-backend-io.h index XXXXXXX..XXXXXXX 100644 --- a/include/sysemu/block-backend-io.h +++ b/include/sysemu/block-backend-io.h @@ -XXX,XX +XXX,XX @@ void blk_iostatus_set_err(BlockBackend *blk, int error); int blk_get_max_iov(BlockBackend *blk); int blk_get_max_hw_iov(BlockBackend *blk); -void blk_io_plug(BlockBackend *blk); +void coroutine_fn blk_co_io_plug(BlockBackend *blk); +void co_wrapper blk_io_plug(BlockBackend *blk); + void blk_io_unplug(BlockBackend *blk); AioContext *blk_get_aio_context(BlockBackend *blk); BlockAcctStats *blk_get_stats(BlockBackend *blk); 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_add_insert_bs_notifier(BlockBackend *blk, Notifier *notify) notifier_list_add(&blk->insert_bs_notifiers, notify); } -void blk_io_plug(BlockBackend *blk) +void coroutine_fn blk_co_io_plug(BlockBackend *blk) { BlockDriverState *bs = blk_bs(blk); IO_CODE(); if (bs) { - bdrv_io_plug(bs); + bdrv_co_io_plug(bs); } } diff --git a/block/file-posix.c b/block/file-posix.c index XXXXXXX..XXXXXXX 100644 --- a/block/file-posix.c +++ b/block/file-posix.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn raw_co_pwritev(BlockDriverState *bs, int64_t offset, return raw_co_prw(bs, offset, bytes, qiov, QEMU_AIO_WRITE); } -static void raw_aio_plug(BlockDriverState *bs) +static void coroutine_fn raw_co_io_plug(BlockDriverState *bs) { BDRVRawState __attribute__((unused)) *s = bs->opaque; #ifdef CONFIG_LINUX_AIO @@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_file = { .bdrv_co_copy_range_from = raw_co_copy_range_from, .bdrv_co_copy_range_to = raw_co_copy_range_to, .bdrv_refresh_limits = raw_refresh_limits, - .bdrv_io_plug = raw_aio_plug, + .bdrv_co_io_plug = raw_co_io_plug, .bdrv_io_unplug = raw_aio_unplug, .bdrv_attach_aio_context = raw_aio_attach_aio_context, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_host_device = { .bdrv_co_copy_range_from = raw_co_copy_range_from, .bdrv_co_copy_range_to = raw_co_copy_range_to, .bdrv_refresh_limits = raw_refresh_limits, - .bdrv_io_plug = raw_aio_plug, + .bdrv_co_io_plug = raw_co_io_plug, .bdrv_io_unplug = raw_aio_unplug, .bdrv_attach_aio_context = raw_aio_attach_aio_context, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_host_cdrom = { .bdrv_co_pwritev = raw_co_pwritev, .bdrv_co_flush_to_disk = raw_co_flush_to_disk, .bdrv_refresh_limits = raw_refresh_limits, - .bdrv_io_plug = raw_aio_plug, + .bdrv_co_io_plug = raw_co_io_plug, .bdrv_io_unplug = raw_aio_unplug, .bdrv_attach_aio_context = raw_aio_attach_aio_context, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_host_cdrom = { .bdrv_co_pwritev = raw_co_pwritev, .bdrv_co_flush_to_disk = raw_co_flush_to_disk, .bdrv_refresh_limits = raw_refresh_limits, - .bdrv_io_plug = raw_aio_plug, + .bdrv_co_io_plug = raw_co_io_plug, .bdrv_io_unplug = raw_aio_unplug, .bdrv_attach_aio_context = raw_aio_attach_aio_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 *qemu_try_blockalign0(BlockDriverState *bs, size_t size) return mem; } -void bdrv_io_plug(BlockDriverState *bs) +void coroutine_fn bdrv_co_io_plug(BlockDriverState *bs) { BdrvChild *child; IO_CODE(); QLIST_FOREACH(child, &bs->children, next) { - bdrv_io_plug(child->bs); + bdrv_co_io_plug(child->bs); } if (qatomic_fetch_inc(&bs->io_plugged) == 0) { BlockDriver *drv = bs->drv; - if (drv && drv->bdrv_io_plug) { - drv->bdrv_io_plug(bs); + if (drv && drv->bdrv_co_io_plug) { + drv->bdrv_co_io_plug(bs); } } } diff --git a/block/nvme.c b/block/nvme.c index XXXXXXX..XXXXXXX 100644 --- a/block/nvme.c +++ b/block/nvme.c @@ -XXX,XX +XXX,XX @@ static void nvme_attach_aio_context(BlockDriverState *bs, } } -static void nvme_aio_plug(BlockDriverState *bs) +static void coroutine_fn nvme_co_io_plug(BlockDriverState *bs) { BDRVNVMeState *s = bs->opaque; assert(!s->plugged); @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_nvme = { .bdrv_detach_aio_context = nvme_detach_aio_context, .bdrv_attach_aio_context = nvme_attach_aio_context, - .bdrv_io_plug = nvme_aio_plug, + .bdrv_co_io_plug = nvme_co_io_plug, .bdrv_io_unplug = nvme_aio_unplug, .bdrv_register_buf = nvme_register_buf, -- 2.38.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> BlockDriver->bdrv_io_unplug is categorized as IO callback, and it currently doesn't run in a coroutine. We should let it take a graph rdlock since the callback traverses the block nodes graph, which however is only possible in a coroutine. The only caller of this function is blk_io_unplug(), therefore make blk_io_unplug() a co_wrapper, so that we're always running in a coroutine where the lock can be taken. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230113204212.359076-4-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-io.h | 3 +-- include/block/block_int-common.h | 2 +- include/sysemu/block-backend-io.h | 4 +++- block/blkio.c | 4 ++-- block/block-backend.c | 4 ++-- block/file-posix.c | 10 +++++----- block/io.c | 8 ++++---- block/nvme.c | 4 ++-- 8 files changed, 20 insertions(+), 19 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 coroutine_fn bdrv_co_leave(BlockDriverState *bs, AioContext *old_ctx); AioContext *child_of_bds_get_parent_aio_context(BdrvChild *c); void coroutine_fn bdrv_co_io_plug(BlockDriverState *bs); - -void bdrv_io_unplug(BlockDriverState *bs); +void coroutine_fn bdrv_co_io_unplug(BlockDriverState *bs); bool coroutine_fn bdrv_co_can_store_new_dirty_bitmap(BlockDriverState *bs, const char *name, 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 { /* io queue for linux-aio */ void coroutine_fn (*bdrv_co_io_plug)(BlockDriverState *bs); - void (*bdrv_io_unplug)(BlockDriverState *bs); + void coroutine_fn (*bdrv_co_io_unplug)(BlockDriverState *bs); /** * bdrv_drain_begin is called if implemented in the beginning of a diff --git a/include/sysemu/block-backend-io.h b/include/sysemu/block-backend-io.h index XXXXXXX..XXXXXXX 100644 --- a/include/sysemu/block-backend-io.h +++ b/include/sysemu/block-backend-io.h @@ -XXX,XX +XXX,XX @@ int blk_get_max_hw_iov(BlockBackend *blk); void coroutine_fn blk_co_io_plug(BlockBackend *blk); void co_wrapper blk_io_plug(BlockBackend *blk); -void blk_io_unplug(BlockBackend *blk); +void coroutine_fn blk_co_io_unplug(BlockBackend *blk); +void co_wrapper blk_io_unplug(BlockBackend *blk); + AioContext *blk_get_aio_context(BlockBackend *blk); BlockAcctStats *blk_get_stats(BlockBackend *blk); void *blk_aio_get(const AIOCBInfo *aiocb_info, BlockBackend *blk, diff --git a/block/blkio.c b/block/blkio.c index XXXXXXX..XXXXXXX 100644 --- a/block/blkio.c +++ b/block/blkio.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn blkio_co_pwrite_zeroes(BlockDriverState *bs, return cod.ret; } -static void blkio_io_unplug(BlockDriverState *bs) +static void coroutine_fn blkio_co_io_unplug(BlockDriverState *bs) { BDRVBlkioState *s = bs->opaque; @@ -XXX,XX +XXX,XX @@ static void blkio_refresh_limits(BlockDriverState *bs, Error **errp) .bdrv_co_pwritev = blkio_co_pwritev, \ .bdrv_co_flush_to_disk = blkio_co_flush, \ .bdrv_co_pwrite_zeroes = blkio_co_pwrite_zeroes, \ - .bdrv_io_unplug = blkio_io_unplug, \ + .bdrv_co_io_unplug = blkio_co_io_unplug, \ .bdrv_refresh_limits = blkio_refresh_limits, \ .bdrv_register_buf = blkio_register_buf, \ .bdrv_unregister_buf = blkio_unregister_buf, \ 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 coroutine_fn blk_co_io_plug(BlockBackend *blk) } } -void blk_io_unplug(BlockBackend *blk) +void coroutine_fn blk_co_io_unplug(BlockBackend *blk) { BlockDriverState *bs = blk_bs(blk); IO_CODE(); if (bs) { - bdrv_io_unplug(bs); + bdrv_co_io_unplug(bs); } } diff --git a/block/file-posix.c b/block/file-posix.c index XXXXXXX..XXXXXXX 100644 --- a/block/file-posix.c +++ b/block/file-posix.c @@ -XXX,XX +XXX,XX @@ static void coroutine_fn raw_co_io_plug(BlockDriverState *bs) #endif } -static void raw_aio_unplug(BlockDriverState *bs) +static void coroutine_fn raw_co_io_unplug(BlockDriverState *bs) { BDRVRawState __attribute__((unused)) *s = bs->opaque; #ifdef CONFIG_LINUX_AIO @@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_file = { .bdrv_co_copy_range_to = raw_co_copy_range_to, .bdrv_refresh_limits = raw_refresh_limits, .bdrv_co_io_plug = raw_co_io_plug, - .bdrv_io_unplug = raw_aio_unplug, + .bdrv_co_io_unplug = raw_co_io_unplug, .bdrv_attach_aio_context = raw_aio_attach_aio_context, .bdrv_co_truncate = raw_co_truncate, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_host_device = { .bdrv_co_copy_range_to = raw_co_copy_range_to, .bdrv_refresh_limits = raw_refresh_limits, .bdrv_co_io_plug = raw_co_io_plug, - .bdrv_io_unplug = raw_aio_unplug, + .bdrv_co_io_unplug = raw_co_io_unplug, .bdrv_attach_aio_context = raw_aio_attach_aio_context, .bdrv_co_truncate = raw_co_truncate, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_host_cdrom = { .bdrv_co_flush_to_disk = raw_co_flush_to_disk, .bdrv_refresh_limits = raw_refresh_limits, .bdrv_co_io_plug = raw_co_io_plug, - .bdrv_io_unplug = raw_aio_unplug, + .bdrv_co_io_unplug = raw_co_io_unplug, .bdrv_attach_aio_context = raw_aio_attach_aio_context, .bdrv_co_truncate = raw_co_truncate, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_host_cdrom = { .bdrv_co_flush_to_disk = raw_co_flush_to_disk, .bdrv_refresh_limits = raw_refresh_limits, .bdrv_co_io_plug = raw_co_io_plug, - .bdrv_io_unplug = raw_aio_unplug, + .bdrv_co_io_unplug = raw_co_io_unplug, .bdrv_attach_aio_context = raw_aio_attach_aio_context, .bdrv_co_truncate = raw_co_truncate, 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 coroutine_fn bdrv_co_io_plug(BlockDriverState *bs) } } -void bdrv_io_unplug(BlockDriverState *bs) +void coroutine_fn bdrv_co_io_unplug(BlockDriverState *bs) { BdrvChild *child; IO_CODE(); @@ -XXX,XX +XXX,XX @@ void bdrv_io_unplug(BlockDriverState *bs) assert(bs->io_plugged); if (qatomic_fetch_dec(&bs->io_plugged) == 1) { BlockDriver *drv = bs->drv; - if (drv && drv->bdrv_io_unplug) { - drv->bdrv_io_unplug(bs); + if (drv && drv->bdrv_co_io_unplug) { + drv->bdrv_co_io_unplug(bs); } } QLIST_FOREACH(child, &bs->children, next) { - bdrv_io_unplug(child->bs); + bdrv_co_io_unplug(child->bs); } } diff --git a/block/nvme.c b/block/nvme.c index XXXXXXX..XXXXXXX 100644 --- a/block/nvme.c +++ b/block/nvme.c @@ -XXX,XX +XXX,XX @@ static void coroutine_fn nvme_co_io_plug(BlockDriverState *bs) s->plugged = true; } -static void nvme_aio_unplug(BlockDriverState *bs) +static void coroutine_fn nvme_co_io_unplug(BlockDriverState *bs) { BDRVNVMeState *s = bs->opaque; assert(s->plugged); @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_nvme = { .bdrv_attach_aio_context = nvme_attach_aio_context, .bdrv_co_io_plug = nvme_co_io_plug, - .bdrv_io_unplug = nvme_aio_unplug, + .bdrv_co_io_unplug = nvme_co_io_unplug, .bdrv_register_buf = nvme_register_buf, .bdrv_unregister_buf = nvme_unregister_buf, -- 2.38.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> bdrv_is_inserted() is categorized as an I/O function, and it currently doesn't run in a coroutine. We should let it take a graph rdlock since it traverses the block nodes graph, which however is only possible in a coroutine. Therefore turn it into a co_wrapper to move the actual function into a coroutine where the lock can be taken. At the same time, add also blk_is_inserted as co_wrapper_mixed, since it is called in both coroutine and non-coroutine contexts. Because now this function creates a new coroutine and polls, we need to take the AioContext lock where it is missing, for the only reason that internally c_w_mixed_bdrv_rdlock calls AIO_WAIT_WHILE and it expects to release the AioContext lock. Once the rwlock is ultimated and placed in every place it needs to be, we will poll using AIO_WAIT_WHILE_UNLOCKED and remove the AioContext lock. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230113204212.359076-5-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-io.h | 5 ++++- include/block/block_int-common.h | 2 +- include/sysemu/block-backend-io.h | 5 ++++- block.c | 8 ++++---- block/block-backend.c | 4 ++-- block/file-posix.c | 8 ++++---- block/io.c | 12 ++++++------ blockdev.c | 8 +++++++- 8 files changed, 32 insertions(+), 20 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 @@ bool bdrv_is_read_only(BlockDriverState *bs); bool bdrv_is_writable(BlockDriverState *bs); bool bdrv_is_sg(BlockDriverState *bs); int bdrv_get_flags(BlockDriverState *bs); -bool bdrv_is_inserted(BlockDriverState *bs); + +bool coroutine_fn bdrv_co_is_inserted(BlockDriverState *bs); +bool co_wrapper bdrv_is_inserted(BlockDriverState *bs); + void bdrv_lock_medium(BlockDriverState *bs, bool locked); void bdrv_eject(BlockDriverState *bs, bool eject_flag); const char *bdrv_get_format_name(BlockDriverState *bs); 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, QEMUIOVector *qiov, int64_t pos); /* removable device specific */ - bool (*bdrv_is_inserted)(BlockDriverState *bs); + bool coroutine_fn (*bdrv_co_is_inserted)(BlockDriverState *bs); void (*bdrv_eject)(BlockDriverState *bs, bool eject_flag); void (*bdrv_lock_medium)(BlockDriverState *bs, bool locked); diff --git a/include/sysemu/block-backend-io.h b/include/sysemu/block-backend-io.h index XXXXXXX..XXXXXXX 100644 --- a/include/sysemu/block-backend-io.h +++ b/include/sysemu/block-backend-io.h @@ -XXX,XX +XXX,XX @@ BlockAIOCB *blk_aio_ioctl(BlockBackend *blk, unsigned long int req, void *buf, void blk_inc_in_flight(BlockBackend *blk); void blk_dec_in_flight(BlockBackend *blk); -bool blk_is_inserted(BlockBackend *blk); + +bool coroutine_fn blk_co_is_inserted(BlockBackend *blk); +bool co_wrapper_mixed blk_is_inserted(BlockBackend *blk); + bool blk_is_available(BlockBackend *blk); void blk_lock_medium(BlockBackend *blk, bool locked); void blk_eject(BlockBackend *blk, bool eject_flag); diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ out: /** * Return TRUE if the media is present */ -bool bdrv_is_inserted(BlockDriverState *bs) +bool coroutine_fn bdrv_co_is_inserted(BlockDriverState *bs) { BlockDriver *drv = bs->drv; BdrvChild *child; @@ -XXX,XX +XXX,XX @@ bool bdrv_is_inserted(BlockDriverState *bs) if (!drv) { return false; } - if (drv->bdrv_is_inserted) { - return drv->bdrv_is_inserted(bs); + if (drv->bdrv_co_is_inserted) { + return drv->bdrv_co_is_inserted(bs); } QLIST_FOREACH(child, &bs->children, next) { - if (!bdrv_is_inserted(child->bs)) { + if (!bdrv_co_is_inserted(child->bs)) { return false; } } 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) bdrv_activate(bs, errp); } -bool blk_is_inserted(BlockBackend *blk) +bool coroutine_fn blk_co_is_inserted(BlockBackend *blk) { BlockDriverState *bs = blk_bs(blk); IO_CODE(); - return bs && bdrv_is_inserted(bs); + return bs && bdrv_co_is_inserted(bs); } bool blk_is_available(BlockBackend *blk) diff --git a/block/file-posix.c b/block/file-posix.c index XXXXXXX..XXXXXXX 100644 --- a/block/file-posix.c +++ b/block/file-posix.c @@ -XXX,XX +XXX,XX @@ out: return prio; } -static bool cdrom_is_inserted(BlockDriverState *bs) +static bool coroutine_fn cdrom_co_is_inserted(BlockDriverState *bs) { BDRVRawState *s = bs->opaque; int ret; @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_host_cdrom = { = raw_get_allocated_file_size, /* removable device support */ - .bdrv_is_inserted = cdrom_is_inserted, + .bdrv_co_is_inserted = cdrom_co_is_inserted, .bdrv_eject = cdrom_eject, .bdrv_lock_medium = cdrom_lock_medium, @@ -XXX,XX +XXX,XX @@ static int cdrom_reopen(BlockDriverState *bs) return 0; } -static bool cdrom_is_inserted(BlockDriverState *bs) +static bool coroutine_fn cdrom_co_is_inserted(BlockDriverState *bs) { return raw_getlength(bs) > 0; } @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_host_cdrom = { = raw_get_allocated_file_size, /* removable device support */ - .bdrv_is_inserted = cdrom_is_inserted, + .bdrv_co_is_inserted = cdrom_co_is_inserted, .bdrv_eject = cdrom_eject, .bdrv_lock_medium = cdrom_lock_medium, }; 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 @@ int coroutine_fn bdrv_co_preadv_part(BdrvChild *child, trace_bdrv_co_preadv_part(bs, offset, bytes, flags); - if (!bdrv_is_inserted(bs)) { + if (!bdrv_co_is_inserted(bs)) { return -ENOMEDIUM; } @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_pwritev_part(BdrvChild *child, trace_bdrv_co_pwritev_part(child->bs, offset, bytes, flags); - if (!bdrv_is_inserted(bs)) { + if (!bdrv_co_is_inserted(bs)) { return -ENOMEDIUM; } @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_flush(BlockDriverState *bs) bdrv_inc_in_flight(bs); - if (!bdrv_is_inserted(bs) || bdrv_is_read_only(bs) || + if (!bdrv_co_is_inserted(bs) || bdrv_is_read_only(bs) || bdrv_is_sg(bs)) { goto early_exit; } @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_pdiscard(BdrvChild *child, int64_t offset, BlockDriverState *bs = child->bs; IO_CODE(); - if (!bs || !bs->drv || !bdrv_is_inserted(bs)) { + if (!bs || !bs->drv || !bdrv_co_is_inserted(bs)) { return -ENOMEDIUM; } @@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_copy_range_internal( assert(!(read_flags & BDRV_REQ_NO_WAIT)); assert(!(write_flags & BDRV_REQ_NO_WAIT)); - if (!dst || !dst->bs || !bdrv_is_inserted(dst->bs)) { + if (!dst || !dst->bs || !bdrv_co_is_inserted(dst->bs)) { return -ENOMEDIUM; } ret = bdrv_check_request32(dst_offset, bytes, NULL, 0); @@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_copy_range_internal( return bdrv_co_pwrite_zeroes(dst, dst_offset, bytes, write_flags); } - if (!src || !src->bs || !bdrv_is_inserted(src->bs)) { + if (!src || !src->bs || !bdrv_co_is_inserted(src->bs)) { return -ENOMEDIUM; } ret = bdrv_check_request32(src_offset, bytes, NULL, 0); diff --git a/blockdev.c b/blockdev.c index XXXXXXX..XXXXXXX 100644 --- a/blockdev.c +++ b/blockdev.c @@ -XXX,XX +XXX,XX @@ fail: static BlockDriverState *qmp_get_root_bs(const char *name, Error **errp) { BlockDriverState *bs; + AioContext *aio_context; bs = bdrv_lookup_bs(name, name, errp); if (bs == NULL) { @@ -XXX,XX +XXX,XX @@ static BlockDriverState *qmp_get_root_bs(const char *name, Error **errp) return NULL; } + aio_context = bdrv_get_aio_context(bs); + aio_context_acquire(aio_context); + if (!bdrv_is_inserted(bs)) { error_setg(errp, "Device has no medium"); - return NULL; + bs = NULL; } + aio_context_release(aio_context); + return bs; } -- 2.38.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> The name is not good, not the least because we are going to convert this to a generated co_wrapper, which adds a _co infix after the first part of the name. No functional change intended. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230113204212.359076-6-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block_int-io.h | 2 +- block.c | 8 ++++---- block/io.c | 8 +++++--- 3 files changed, 10 insertions(+), 8 deletions(-) diff --git a/include/block/block_int-io.h b/include/block/block_int-io.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block_int-io.h +++ b/include/block/block_int-io.h @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_copy_range_to(BdrvChild *src, int64_t src_offset, BdrvRequestFlags read_flags, BdrvRequestFlags write_flags); -int refresh_total_sectors(BlockDriverState *bs, int64_t hint); +int bdrv_refresh_total_sectors(BlockDriverState *bs, int64_t hint); BdrvChild *bdrv_cow_child(BlockDriverState *bs); BdrvChild *bdrv_filter_child(BlockDriverState *bs); diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ static int find_image_format(BlockBackend *file, const char *filename, * Set the current 'total_sectors' value * Return 0 on success, -errno on error. */ -int refresh_total_sectors(BlockDriverState *bs, int64_t hint) +int bdrv_refresh_total_sectors(BlockDriverState *bs, int64_t hint) { BlockDriver *drv = bs->drv; IO_CODE(); @@ -XXX,XX +XXX,XX @@ static int bdrv_open_driver(BlockDriverState *bs, BlockDriver *drv, bs->supported_read_flags |= BDRV_REQ_REGISTERED_BUF; bs->supported_write_flags |= BDRV_REQ_REGISTERED_BUF; - ret = refresh_total_sectors(bs, bs->total_sectors); + ret = bdrv_refresh_total_sectors(bs, bs->total_sectors); if (ret < 0) { error_setg_errno(errp, -ret, "Could not refresh total sector count"); return ret; @@ -XXX,XX +XXX,XX @@ int64_t bdrv_nb_sectors(BlockDriverState *bs) return -ENOMEDIUM; if (drv->has_variable_length) { - int ret = refresh_total_sectors(bs, bs->total_sectors); + int ret = bdrv_refresh_total_sectors(bs, bs->total_sectors); if (ret < 0) { return ret; } @@ -XXX,XX +XXX,XX @@ int bdrv_activate(BlockDriverState *bs, Error **errp) bdrv_dirty_bitmap_skip_store(bm, false); } - ret = refresh_total_sectors(bs, bs->total_sectors); + ret = bdrv_refresh_total_sectors(bs, bs->total_sectors); if (ret < 0) { bs->open_flags |= BDRV_O_INACTIVE; error_setg_errno(errp, -ret, "Could not refresh total sector count"); 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 @@ int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact, goto out; } - ret = refresh_total_sectors(bs, offset >> BDRV_SECTOR_BITS); + ret = bdrv_refresh_total_sectors(bs, offset >> BDRV_SECTOR_BITS); if (ret < 0) { error_setg_errno(errp, -ret, "Could not refresh total sector count"); } else { offset = bs->total_sectors * BDRV_SECTOR_SIZE; } - /* It's possible that truncation succeeded but refresh_total_sectors + /* + * It's possible that truncation succeeded but bdrv_refresh_total_sectors * failed, but the latter doesn't affect how we should finish the request. - * Pass 0 as the last parameter so that dirty bitmaps etc. are handled. */ + * Pass 0 as the last parameter so that dirty bitmaps etc. are handled. + */ bdrv_co_write_req_finish(child, offset - new_bytes, new_bytes, &req, 0); out: -- 2.38.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> BlockDriver->bdrv_getlength is categorized as IO callback, and it currently doesn't run in a coroutine. We should let it take a graph rdlock since the callback traverses the block nodes graph, which however is only possible in a coroutine. Therefore turn it into a co_wrapper to move the actual function into a coroutine where the lock can be taken. Because now this function creates a new coroutine and polls, we need to take the AioContext lock where it is missing, for the only reason that internally co_wrapper calls AIO_WAIT_WHILE and it expects to release the AioContext lock. This is especially messy when a co_wrapper creates a coroutine and polls in bdrv_open_driver, because this function has so many callers in so many context that it can easily lead to deadlocks. Therefore the new rule for bdrv_open_driver is that the caller must always hold the AioContext lock of the given bs (except if it is a coroutine), because the function calls bdrv_refresh_total_sectors() which is now a co_wrapper. Once the rwlock is ultimated and placed in every place it needs to be, we will poll using AIO_WAIT_WHILE_UNLOCKED and remove the AioContext lock. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230113204212.359076-7-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-io.h | 8 ++++-- include/block/block_int-common.h | 2 +- include/block/block_int-io.h | 5 +++- include/sysemu/block-backend-io.h | 10 ++++++-- block.c | 32 +++++++++++++++++------ block/blkdebug.c | 6 ++--- block/blkio.c | 6 ++--- block/blklogwrites.c | 6 ++--- block/blkreplay.c | 6 ++--- block/blkverify.c | 6 ++--- block/block-backend.c | 10 +++++--- block/commit.c | 4 +-- block/copy-on-read.c | 6 ++--- block/crypto.c | 6 ++--- block/curl.c | 10 ++++---- block/file-posix.c | 42 +++++++++++++++---------------- block/file-win32.c | 8 +++--- block/filter-compress.c | 6 ++--- block/gluster.c | 12 ++++----- block/iscsi.c | 10 ++++---- block/mirror.c | 4 +-- block/nbd.c | 8 +++--- block/null.c | 6 ++--- block/nvme.c | 6 ++--- block/preallocate.c | 10 ++++---- block/qed.c | 4 +-- block/quorum.c | 8 +++--- block/raw-format.c | 6 ++--- block/rbd.c | 4 +-- block/replication.c | 6 ++--- block/ssh.c | 4 +-- block/throttle.c | 6 ++--- hw/scsi/scsi-disk.c | 5 ++++ tests/unit/test-block-iothread.c | 3 +++ block/meson.build | 1 + 35 files changed, 161 insertions(+), 121 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 @@ int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact, PreallocMode prealloc, BdrvRequestFlags flags, Error **errp); -int64_t bdrv_nb_sectors(BlockDriverState *bs); -int64_t bdrv_getlength(BlockDriverState *bs); +int64_t coroutine_fn bdrv_co_nb_sectors(BlockDriverState *bs); +int64_t co_wrapper_mixed bdrv_nb_sectors(BlockDriverState *bs); + +int64_t coroutine_fn bdrv_co_getlength(BlockDriverState *bs); +int64_t co_wrapper_mixed bdrv_getlength(BlockDriverState *bs); + int64_t 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 { int coroutine_fn (*bdrv_co_truncate)(BlockDriverState *bs, int64_t offset, bool exact, PreallocMode prealloc, BdrvRequestFlags flags, Error **errp); - int64_t (*bdrv_getlength)(BlockDriverState *bs); + int64_t coroutine_fn (*bdrv_co_getlength)(BlockDriverState *bs); int64_t (*bdrv_get_allocated_file_size)(BlockDriverState *bs); BlockMeasureInfo *(*bdrv_measure)(QemuOpts *opts, BlockDriverState *in_bs, Error **errp); diff --git a/include/block/block_int-io.h b/include/block/block_int-io.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block_int-io.h +++ b/include/block/block_int-io.h @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_copy_range_to(BdrvChild *src, int64_t src_offset, BdrvRequestFlags read_flags, BdrvRequestFlags write_flags); -int bdrv_refresh_total_sectors(BlockDriverState *bs, int64_t hint); +int coroutine_fn bdrv_co_refresh_total_sectors(BlockDriverState *bs, + int64_t hint); +int co_wrapper_mixed +bdrv_refresh_total_sectors(BlockDriverState *bs, int64_t hint); BdrvChild *bdrv_cow_child(BlockDriverState *bs); BdrvChild *bdrv_filter_child(BlockDriverState *bs); diff --git a/include/sysemu/block-backend-io.h b/include/sysemu/block-backend-io.h index XXXXXXX..XXXXXXX 100644 --- a/include/sysemu/block-backend-io.h +++ b/include/sysemu/block-backend-io.h @@ -XXX,XX +XXX,XX @@ bool co_wrapper_mixed blk_is_inserted(BlockBackend *blk); bool blk_is_available(BlockBackend *blk); void blk_lock_medium(BlockBackend *blk, bool locked); void blk_eject(BlockBackend *blk, bool eject_flag); -int64_t blk_getlength(BlockBackend *blk); + +int64_t coroutine_fn blk_co_getlength(BlockBackend *blk); +int64_t co_wrapper_mixed blk_getlength(BlockBackend *blk); + void blk_get_geometry(BlockBackend *blk, uint64_t *nb_sectors_ptr); -int64_t blk_nb_sectors(BlockBackend *blk); + +int64_t coroutine_fn blk_co_nb_sectors(BlockBackend *blk); +int64_t co_wrapper_mixed blk_nb_sectors(BlockBackend *blk); + void *blk_try_blockalign(BlockBackend *blk, size_t size); void *blk_blockalign(BlockBackend *blk, size_t size); bool blk_is_writable(BlockBackend *blk); diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ static int find_image_format(BlockBackend *file, const char *filename, * Set the current 'total_sectors' value * Return 0 on success, -errno on error. */ -int bdrv_refresh_total_sectors(BlockDriverState *bs, int64_t hint) +int coroutine_fn bdrv_co_refresh_total_sectors(BlockDriverState *bs, + int64_t hint) { BlockDriver *drv = bs->drv; IO_CODE(); @@ -XXX,XX +XXX,XX @@ int bdrv_refresh_total_sectors(BlockDriverState *bs, int64_t hint) return -ENOMEDIUM; } - /* Do not attempt drv->bdrv_getlength() on scsi-generic devices */ + /* Do not attempt drv->bdrv_co_getlength() on scsi-generic devices */ if (bdrv_is_sg(bs)) return 0; /* query actual device if possible, otherwise just trust the hint */ - if (drv->bdrv_getlength) { - int64_t length = drv->bdrv_getlength(bs); + if (drv->bdrv_co_getlength) { + int64_t length = drv->bdrv_co_getlength(bs); if (length < 0) { return length; } @@ -XXX,XX +XXX,XX @@ out: g_free(gen_node_name); } +/* + * The caller must always hold @bs AioContext lock, because this function calls + * 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) @@ -XXX,XX +XXX,XX @@ out: * The reference parameter may be used to specify an existing block device which * should be opened. If specified, neither options nor a filename may be given, * nor can an existing BDS be reused (that is, *pbs has to be NULL). + * + * The caller must always hold @filename AioContext lock, because this + * function eventually calls bdrv_refresh_total_sectors() which polls + * when called from non-coroutine context. */ static BlockDriverState *bdrv_open_inherit(const char *filename, const char *reference, @@ -XXX,XX +XXX,XX @@ close_and_fail: return NULL; } +/* + * The caller must always hold @filename AioContext lock, because this + * function eventually calls bdrv_refresh_total_sectors() which polls + * when called from non-coroutine context. + */ BlockDriverState *bdrv_open(const char *filename, const char *reference, QDict *options, int flags, Error **errp) { @@ -XXX,XX +XXX,XX @@ BlockMeasureInfo *bdrv_measure(BlockDriver *drv, QemuOpts *opts, /** * Return number of sectors on success, -errno on error. */ -int64_t bdrv_nb_sectors(BlockDriverState *bs) +int64_t coroutine_fn bdrv_co_nb_sectors(BlockDriverState *bs) { BlockDriver *drv = bs->drv; IO_CODE(); @@ -XXX,XX +XXX,XX @@ int64_t bdrv_nb_sectors(BlockDriverState *bs) return -ENOMEDIUM; if (drv->has_variable_length) { - int ret = bdrv_refresh_total_sectors(bs, bs->total_sectors); + int ret = bdrv_co_refresh_total_sectors(bs, bs->total_sectors); if (ret < 0) { return ret; } @@ -XXX,XX +XXX,XX @@ int64_t bdrv_nb_sectors(BlockDriverState *bs) * Return length in bytes on success, -errno on error. * The length is always a multiple of BDRV_SECTOR_SIZE. */ -int64_t bdrv_getlength(BlockDriverState *bs) +int64_t coroutine_fn bdrv_co_getlength(BlockDriverState *bs) { - int64_t ret = bdrv_nb_sectors(bs); + int64_t ret; IO_CODE(); + ret = bdrv_co_nb_sectors(bs); if (ret < 0) { return ret; } diff --git a/block/blkdebug.c b/block/blkdebug.c index XXXXXXX..XXXXXXX 100644 --- a/block/blkdebug.c +++ b/block/blkdebug.c @@ -XXX,XX +XXX,XX @@ static bool blkdebug_debug_is_suspended(BlockDriverState *bs, const char *tag) return false; } -static int64_t blkdebug_getlength(BlockDriverState *bs) +static int64_t coroutine_fn blkdebug_co_getlength(BlockDriverState *bs) { - return bdrv_getlength(bs->file->bs); + return bdrv_co_getlength(bs->file->bs); } static void blkdebug_refresh_filename(BlockDriverState *bs) @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_blkdebug = { .bdrv_reopen_prepare = blkdebug_reopen_prepare, .bdrv_child_perm = blkdebug_child_perm, - .bdrv_getlength = blkdebug_getlength, + .bdrv_co_getlength = blkdebug_co_getlength, .bdrv_refresh_filename = blkdebug_refresh_filename, .bdrv_refresh_limits = blkdebug_refresh_limits, diff --git a/block/blkio.c b/block/blkio.c index XXXXXXX..XXXXXXX 100644 --- a/block/blkio.c +++ b/block/blkio.c @@ -XXX,XX +XXX,XX @@ static void blkio_close(BlockDriverState *bs) } } -static int64_t blkio_getlength(BlockDriverState *bs) +static int64_t coroutine_fn blkio_co_getlength(BlockDriverState *bs) { BDRVBlkioState *s = bs->opaque; uint64_t capacity; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn blkio_truncate(BlockDriverState *bs, int64_t offset, return -ENOTSUP; } - current_length = blkio_getlength(bs); + current_length = blkio_co_getlength(bs); if (offset > current_length) { error_setg(errp, "Cannot grow device"); @@ -XXX,XX +XXX,XX @@ static void blkio_refresh_limits(BlockDriverState *bs, Error **errp) .instance_size = sizeof(BDRVBlkioState), \ .bdrv_file_open = blkio_file_open, \ .bdrv_close = blkio_close, \ - .bdrv_getlength = blkio_getlength, \ + .bdrv_co_getlength = blkio_co_getlength, \ .bdrv_co_truncate = blkio_truncate, \ .bdrv_get_info = blkio_get_info, \ .bdrv_attach_aio_context = blkio_attach_aio_context, \ diff --git a/block/blklogwrites.c b/block/blklogwrites.c index XXXXXXX..XXXXXXX 100644 --- a/block/blklogwrites.c +++ b/block/blklogwrites.c @@ -XXX,XX +XXX,XX @@ static void blk_log_writes_close(BlockDriverState *bs) s->log_file = NULL; } -static int64_t blk_log_writes_getlength(BlockDriverState *bs) +static int64_t coroutine_fn blk_log_writes_co_getlength(BlockDriverState *bs) { - return bdrv_getlength(bs->file->bs); + return bdrv_co_getlength(bs->file->bs); } static void blk_log_writes_child_perm(BlockDriverState *bs, BdrvChild *c, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_blk_log_writes = { .bdrv_open = blk_log_writes_open, .bdrv_close = blk_log_writes_close, - .bdrv_getlength = blk_log_writes_getlength, + .bdrv_co_getlength = blk_log_writes_co_getlength, .bdrv_child_perm = blk_log_writes_child_perm, .bdrv_refresh_limits = blk_log_writes_refresh_limits, diff --git a/block/blkreplay.c b/block/blkreplay.c index XXXXXXX..XXXXXXX 100644 --- a/block/blkreplay.c +++ b/block/blkreplay.c @@ -XXX,XX +XXX,XX @@ fail: return ret; } -static int64_t blkreplay_getlength(BlockDriverState *bs) +static int64_t coroutine_fn blkreplay_co_getlength(BlockDriverState *bs) { - return bdrv_getlength(bs->file->bs); + return bdrv_co_getlength(bs->file->bs); } /* This bh is used for synchronization of return from coroutines. @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_blkreplay = { .bdrv_open = blkreplay_open, .bdrv_child_perm = bdrv_default_perms, - .bdrv_getlength = blkreplay_getlength, + .bdrv_co_getlength = blkreplay_co_getlength, .bdrv_co_preadv = blkreplay_co_preadv, .bdrv_co_pwritev = blkreplay_co_pwritev, 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 void blkverify_close(BlockDriverState *bs) s->test_file = NULL; } -static int64_t blkverify_getlength(BlockDriverState *bs) +static int64_t coroutine_fn blkverify_co_getlength(BlockDriverState *bs) { BDRVBlkverifyState *s = bs->opaque; - return bdrv_getlength(s->test_file->bs); + return bdrv_co_getlength(s->test_file->bs); } static void coroutine_fn blkverify_do_test_req(void *opaque) @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_blkverify = { .bdrv_file_open = blkverify_open, .bdrv_close = blkverify_close, .bdrv_child_perm = bdrv_default_perms, - .bdrv_getlength = blkverify_getlength, + .bdrv_co_getlength = blkverify_co_getlength, .bdrv_refresh_filename = blkverify_refresh_filename, .bdrv_dirname = blkverify_dirname, 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 @@ BlockAIOCB *blk_aio_pwrite_zeroes(BlockBackend *blk, int64_t offset, flags | BDRV_REQ_ZERO_WRITE, cb, opaque); } -int64_t blk_getlength(BlockBackend *blk) +int64_t coroutine_fn blk_co_getlength(BlockBackend *blk) { IO_CODE(); + if (!blk_is_available(blk)) { return -ENOMEDIUM; } - return bdrv_getlength(blk_bs(blk)); + return bdrv_co_getlength(blk_bs(blk)); } void blk_get_geometry(BlockBackend *blk, uint64_t *nb_sectors_ptr) @@ -XXX,XX +XXX,XX @@ void blk_get_geometry(BlockBackend *blk, uint64_t *nb_sectors_ptr) } } -int64_t blk_nb_sectors(BlockBackend *blk) +int64_t coroutine_fn blk_co_nb_sectors(BlockBackend *blk) { IO_CODE(); + if (!blk_is_available(blk)) { return -ENOMEDIUM; } - return bdrv_nb_sectors(blk_bs(blk)); + return bdrv_co_nb_sectors(blk_bs(blk)); } BlockAIOCB *blk_aio_preadv(BlockBackend *blk, int64_t offset, diff --git a/block/commit.c b/block/commit.c index XXXXXXX..XXXXXXX 100644 --- a/block/commit.c +++ b/block/commit.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn commit_run(Job *job, Error **errp) QEMU_AUTO_VFREE void *buf = NULL; int64_t len, base_len; - len = blk_getlength(s->top); + len = blk_co_getlength(s->top); if (len < 0) { return len; } job_progress_set_remaining(&s->common.job, len); - base_len = blk_getlength(s->base); + base_len = blk_co_getlength(s->base); if (base_len < 0) { return base_len; } diff --git a/block/copy-on-read.c b/block/copy-on-read.c index XXXXXXX..XXXXXXX 100644 --- a/block/copy-on-read.c +++ b/block/copy-on-read.c @@ -XXX,XX +XXX,XX @@ static void cor_child_perm(BlockDriverState *bs, BdrvChild *c, } -static int64_t cor_getlength(BlockDriverState *bs) +static int64_t coroutine_fn cor_co_getlength(BlockDriverState *bs) { - return bdrv_getlength(bs->file->bs); + return bdrv_co_getlength(bs->file->bs); } @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_copy_on_read = { .bdrv_close = cor_close, .bdrv_child_perm = cor_child_perm, - .bdrv_getlength = cor_getlength, + .bdrv_co_getlength = cor_co_getlength, .bdrv_co_preadv_part = cor_co_preadv_part, .bdrv_co_pwritev_part = cor_co_pwritev_part, 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 @@ static void block_crypto_refresh_limits(BlockDriverState *bs, Error **errp) } -static int64_t block_crypto_getlength(BlockDriverState *bs) +static int64_t coroutine_fn block_crypto_co_getlength(BlockDriverState *bs) { BlockCrypto *crypto = bs->opaque; - int64_t len = bdrv_getlength(bs->file->bs); + int64_t len = bdrv_co_getlength(bs->file->bs); uint64_t offset = qcrypto_block_get_payload_offset(crypto->block); assert(offset < INT64_MAX); @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_crypto_luks = { .bdrv_refresh_limits = block_crypto_refresh_limits, .bdrv_co_preadv = block_crypto_co_preadv, .bdrv_co_pwritev = block_crypto_co_pwritev, - .bdrv_getlength = block_crypto_getlength, + .bdrv_co_getlength = block_crypto_co_getlength, .bdrv_measure = block_crypto_measure, .bdrv_get_info = block_crypto_get_info_luks, .bdrv_get_specific_info = block_crypto_get_specific_info_luks, diff --git a/block/curl.c b/block/curl.c index XXXXXXX..XXXXXXX 100644 --- a/block/curl.c +++ b/block/curl.c @@ -XXX,XX +XXX,XX @@ static void curl_close(BlockDriverState *bs) g_free(s->proxypassword); } -static int64_t curl_getlength(BlockDriverState *bs) +static int64_t coroutine_fn curl_co_getlength(BlockDriverState *bs) { BDRVCURLState *s = bs->opaque; return s->len; @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_http = { .bdrv_parse_filename = curl_parse_filename, .bdrv_file_open = curl_open, .bdrv_close = curl_close, - .bdrv_getlength = curl_getlength, + .bdrv_co_getlength = curl_co_getlength, .bdrv_co_preadv = curl_co_preadv, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_https = { .bdrv_parse_filename = curl_parse_filename, .bdrv_file_open = curl_open, .bdrv_close = curl_close, - .bdrv_getlength = curl_getlength, + .bdrv_co_getlength = curl_co_getlength, .bdrv_co_preadv = curl_co_preadv, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_ftp = { .bdrv_parse_filename = curl_parse_filename, .bdrv_file_open = curl_open, .bdrv_close = curl_close, - .bdrv_getlength = curl_getlength, + .bdrv_co_getlength = curl_co_getlength, .bdrv_co_preadv = curl_co_preadv, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_ftps = { .bdrv_parse_filename = curl_parse_filename, .bdrv_file_open = curl_open, .bdrv_close = curl_close, - .bdrv_getlength = curl_getlength, + .bdrv_co_getlength = curl_co_getlength, .bdrv_co_preadv = curl_co_preadv, diff --git a/block/file-posix.c b/block/file-posix.c index XXXXXXX..XXXXXXX 100644 --- a/block/file-posix.c +++ b/block/file-posix.c @@ -XXX,XX +XXX,XX @@ static int fd_open(BlockDriverState *bs) return -EIO; } -static int64_t raw_getlength(BlockDriverState *bs); +static int64_t coroutine_fn raw_co_getlength(BlockDriverState *bs); typedef struct RawPosixAIOData { BlockDriverState *bs; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn raw_co_truncate(BlockDriverState *bs, int64_t offset, } if (S_ISCHR(st.st_mode) || S_ISBLK(st.st_mode)) { - int64_t cur_length = raw_getlength(bs); + int64_t cur_length = raw_co_getlength(bs); if (offset != cur_length && exact) { error_setg(errp, "Cannot resize device files"); @@ -XXX,XX +XXX,XX @@ static int coroutine_fn raw_co_truncate(BlockDriverState *bs, int64_t offset, } #ifdef __OpenBSD__ -static int64_t raw_getlength(BlockDriverState *bs) +static int64_t coroutine_fn raw_co_getlength(BlockDriverState *bs) { BDRVRawState *s = bs->opaque; int fd = s->fd; @@ -XXX,XX +XXX,XX @@ static int64_t raw_getlength(BlockDriverState *bs) return st.st_size; } #elif defined(__NetBSD__) -static int64_t raw_getlength(BlockDriverState *bs) +static int64_t coroutine_fn raw_co_getlength(BlockDriverState *bs) { BDRVRawState *s = bs->opaque; int fd = s->fd; @@ -XXX,XX +XXX,XX @@ static int64_t raw_getlength(BlockDriverState *bs) return st.st_size; } #elif defined(__sun__) -static int64_t raw_getlength(BlockDriverState *bs) +static int64_t coroutine_fn raw_co_getlength(BlockDriverState *bs) { BDRVRawState *s = bs->opaque; struct dk_minfo minfo; @@ -XXX,XX +XXX,XX @@ static int64_t raw_getlength(BlockDriverState *bs) return size; } #elif defined(CONFIG_BSD) -static int64_t raw_getlength(BlockDriverState *bs) +static int64_t coroutine_fn raw_co_getlength(BlockDriverState *bs) { BDRVRawState *s = bs->opaque; int fd = s->fd; @@ -XXX,XX +XXX,XX @@ again: return size; } #else -static int64_t raw_getlength(BlockDriverState *bs) +static int64_t coroutine_fn raw_co_getlength(BlockDriverState *bs) { BDRVRawState *s = bs->opaque; int ret; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn raw_co_block_status(BlockDriverState *bs, * round up if necessary. */ if (!QEMU_IS_ALIGNED(*pnum, bs->bl.request_alignment)) { - int64_t file_length = raw_getlength(bs); + int64_t file_length = raw_co_getlength(bs); if (file_length > 0) { /* Ignore errors, this is just a safeguard */ assert(hole == file_length); @@ -XXX,XX +XXX,XX @@ static int coroutine_fn raw_co_block_status(BlockDriverState *bs, #if defined(__linux__) /* Verify that the file is not in the page cache */ -static void check_cache_dropped(BlockDriverState *bs, Error **errp) +static void coroutine_fn check_cache_dropped(BlockDriverState *bs, Error **errp) { const size_t window_size = 128 * 1024 * 1024; BDRVRawState *s = bs->opaque; @@ -XXX,XX +XXX,XX @@ static void check_cache_dropped(BlockDriverState *bs, Error **errp) page_size = sysconf(_SC_PAGESIZE); vec = g_malloc(DIV_ROUND_UP(window_size, page_size)); - end = raw_getlength(bs); + end = raw_co_getlength(bs); for (offset = 0; offset < end; offset += window_size) { void *new_window; @@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_file = { .bdrv_co_io_unplug = raw_co_io_unplug, .bdrv_attach_aio_context = raw_aio_attach_aio_context, - .bdrv_co_truncate = raw_co_truncate, - .bdrv_getlength = raw_getlength, + .bdrv_co_truncate = raw_co_truncate, + .bdrv_co_getlength = raw_co_getlength, .bdrv_get_info = raw_get_info, .bdrv_get_allocated_file_size = raw_get_allocated_file_size, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_host_device = { .bdrv_co_io_unplug = raw_co_io_unplug, .bdrv_attach_aio_context = raw_aio_attach_aio_context, - .bdrv_co_truncate = raw_co_truncate, - .bdrv_getlength = raw_getlength, + .bdrv_co_truncate = raw_co_truncate, + .bdrv_co_getlength = raw_co_getlength, .bdrv_get_info = raw_get_info, .bdrv_get_allocated_file_size = raw_get_allocated_file_size, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_host_cdrom = { .bdrv_co_io_unplug = raw_co_io_unplug, .bdrv_attach_aio_context = raw_aio_attach_aio_context, - .bdrv_co_truncate = raw_co_truncate, - .bdrv_getlength = raw_getlength, - .has_variable_length = true, + .bdrv_co_truncate = raw_co_truncate, + .bdrv_co_getlength = raw_co_getlength, + .has_variable_length = true, .bdrv_get_allocated_file_size = raw_get_allocated_file_size, @@ -XXX,XX +XXX,XX @@ static int cdrom_reopen(BlockDriverState *bs) static bool coroutine_fn cdrom_co_is_inserted(BlockDriverState *bs) { - return raw_getlength(bs) > 0; + return raw_co_getlength(bs) > 0; } static void cdrom_eject(BlockDriverState *bs, bool eject_flag) @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_host_cdrom = { .bdrv_co_io_unplug = raw_co_io_unplug, .bdrv_attach_aio_context = raw_aio_attach_aio_context, - .bdrv_co_truncate = raw_co_truncate, - .bdrv_getlength = raw_getlength, - .has_variable_length = true, + .bdrv_co_truncate = raw_co_truncate, + .bdrv_co_getlength = raw_co_getlength, + .has_variable_length = true, .bdrv_get_allocated_file_size = raw_get_allocated_file_size, diff --git a/block/file-win32.c b/block/file-win32.c index XXXXXXX..XXXXXXX 100644 --- a/block/file-win32.c +++ b/block/file-win32.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn raw_co_truncate(BlockDriverState *bs, int64_t offset, return 0; } -static int64_t raw_getlength(BlockDriverState *bs) +static int64_t coroutine_fn raw_co_getlength(BlockDriverState *bs) { BDRVRawState *s = bs->opaque; LARGE_INTEGER l; @@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_file = { .bdrv_aio_flush = raw_aio_flush, .bdrv_co_truncate = raw_co_truncate, - .bdrv_getlength = raw_getlength, + .bdrv_co_getlength = raw_co_getlength, .bdrv_get_allocated_file_size = raw_get_allocated_file_size, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_host_device = { .bdrv_detach_aio_context = raw_detach_aio_context, .bdrv_attach_aio_context = raw_attach_aio_context, - .bdrv_getlength = raw_getlength, - .has_variable_length = true, + .bdrv_co_getlength = raw_co_getlength, + .has_variable_length = true, .bdrv_get_allocated_file_size = raw_get_allocated_file_size, diff --git a/block/filter-compress.c b/block/filter-compress.c index XXXXXXX..XXXXXXX 100644 --- a/block/filter-compress.c +++ b/block/filter-compress.c @@ -XXX,XX +XXX,XX @@ static int compress_open(BlockDriverState *bs, QDict *options, int flags, } -static int64_t compress_getlength(BlockDriverState *bs) +static int64_t coroutine_fn compress_co_getlength(BlockDriverState *bs) { - return bdrv_getlength(bs->file->bs); + return bdrv_co_getlength(bs->file->bs); } @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_compress = { .bdrv_open = compress_open, .bdrv_child_perm = bdrv_default_perms, - .bdrv_getlength = compress_getlength, + .bdrv_co_getlength = compress_co_getlength, .bdrv_co_preadv_part = compress_co_preadv_part, .bdrv_co_pwritev_part = compress_co_pwritev_part, diff --git a/block/gluster.c b/block/gluster.c index XXXXXXX..XXXXXXX 100644 --- a/block/gluster.c +++ b/block/gluster.c @@ -XXX,XX +XXX,XX @@ static coroutine_fn int qemu_gluster_co_pdiscard(BlockDriverState *bs, } #endif -static int64_t qemu_gluster_getlength(BlockDriverState *bs) +static int64_t coroutine_fn qemu_gluster_co_getlength(BlockDriverState *bs) { BDRVGlusterState *s = bs->opaque; int64_t ret; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qemu_gluster_co_block_status(BlockDriverState *bs, * round up if necessary. */ if (!QEMU_IS_ALIGNED(*pnum, bs->bl.request_alignment)) { - int64_t file_length = qemu_gluster_getlength(bs); + int64_t file_length = qemu_gluster_co_getlength(bs); if (file_length > 0) { /* Ignore errors, this is just a safeguard */ assert(hole == file_length); @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_gluster = { .bdrv_close = qemu_gluster_close, .bdrv_co_create = qemu_gluster_co_create, .bdrv_co_create_opts = qemu_gluster_co_create_opts, - .bdrv_getlength = qemu_gluster_getlength, + .bdrv_co_getlength = qemu_gluster_co_getlength, .bdrv_get_allocated_file_size = qemu_gluster_allocated_file_size, .bdrv_co_truncate = qemu_gluster_co_truncate, .bdrv_co_readv = qemu_gluster_co_readv, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_gluster_tcp = { .bdrv_close = qemu_gluster_close, .bdrv_co_create = qemu_gluster_co_create, .bdrv_co_create_opts = qemu_gluster_co_create_opts, - .bdrv_getlength = qemu_gluster_getlength, + .bdrv_co_getlength = qemu_gluster_co_getlength, .bdrv_get_allocated_file_size = qemu_gluster_allocated_file_size, .bdrv_co_truncate = qemu_gluster_co_truncate, .bdrv_co_readv = qemu_gluster_co_readv, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_gluster_unix = { .bdrv_close = qemu_gluster_close, .bdrv_co_create = qemu_gluster_co_create, .bdrv_co_create_opts = qemu_gluster_co_create_opts, - .bdrv_getlength = qemu_gluster_getlength, + .bdrv_co_getlength = qemu_gluster_co_getlength, .bdrv_get_allocated_file_size = qemu_gluster_allocated_file_size, .bdrv_co_truncate = qemu_gluster_co_truncate, .bdrv_co_readv = qemu_gluster_co_readv, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_gluster_rdma = { .bdrv_close = qemu_gluster_close, .bdrv_co_create = qemu_gluster_co_create, .bdrv_co_create_opts = qemu_gluster_co_create_opts, - .bdrv_getlength = qemu_gluster_getlength, + .bdrv_co_getlength = qemu_gluster_co_getlength, .bdrv_get_allocated_file_size = qemu_gluster_allocated_file_size, .bdrv_co_truncate = qemu_gluster_co_truncate, .bdrv_co_readv = qemu_gluster_co_readv, diff --git a/block/iscsi.c b/block/iscsi.c index XXXXXXX..XXXXXXX 100644 --- a/block/iscsi.c +++ b/block/iscsi.c @@ -XXX,XX +XXX,XX @@ static BlockAIOCB *iscsi_aio_ioctl(BlockDriverState *bs, #endif -static int64_t -iscsi_getlength(BlockDriverState *bs) +static int64_t coroutine_fn +iscsi_co_getlength(BlockDriverState *bs) { IscsiLun *iscsilun = bs->opaque; int64_t len; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn iscsi_co_truncate(BlockDriverState *bs, int64_t offset, return -EIO; } - cur_length = iscsi_getlength(bs); + cur_length = iscsi_co_getlength(bs); if (offset != cur_length && exact) { error_setg(errp, "Cannot resize iSCSI devices"); return -ENOTSUP; @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_iscsi = { .bdrv_reopen_commit = iscsi_reopen_commit, .bdrv_co_invalidate_cache = iscsi_co_invalidate_cache, - .bdrv_getlength = iscsi_getlength, + .bdrv_co_getlength = iscsi_co_getlength, .bdrv_get_info = iscsi_get_info, .bdrv_co_truncate = iscsi_co_truncate, .bdrv_refresh_limits = iscsi_refresh_limits, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_iser = { .bdrv_reopen_commit = iscsi_reopen_commit, .bdrv_co_invalidate_cache = iscsi_co_invalidate_cache, - .bdrv_getlength = iscsi_getlength, + .bdrv_co_getlength = iscsi_co_getlength, .bdrv_get_info = iscsi_get_info, .bdrv_co_truncate = iscsi_co_truncate, .bdrv_refresh_limits = iscsi_refresh_limits, 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 coroutine_fn mirror_run(Job *job, Error **errp) goto immediate_exit; } - s->bdev_length = bdrv_getlength(bs); + s->bdev_length = bdrv_co_getlength(bs); if (s->bdev_length < 0) { ret = s->bdev_length; goto immediate_exit; } - target_length = blk_getlength(s->target); + target_length = blk_co_getlength(s->target); if (target_length < 0) { ret = target_length; goto immediate_exit; 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 @@ static int coroutine_fn nbd_co_truncate(BlockDriverState *bs, int64_t offset, return 0; } -static int64_t nbd_getlength(BlockDriverState *bs) +static int64_t coroutine_fn nbd_co_getlength(BlockDriverState *bs) { BDRVNBDState *s = bs->opaque; @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_nbd = { .bdrv_co_pdiscard = nbd_client_co_pdiscard, .bdrv_refresh_limits = nbd_refresh_limits, .bdrv_co_truncate = nbd_co_truncate, - .bdrv_getlength = nbd_getlength, + .bdrv_co_getlength = nbd_co_getlength, .bdrv_refresh_filename = nbd_refresh_filename, .bdrv_co_block_status = nbd_client_co_block_status, .bdrv_dirname = nbd_dirname, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_nbd_tcp = { .bdrv_co_pdiscard = nbd_client_co_pdiscard, .bdrv_refresh_limits = nbd_refresh_limits, .bdrv_co_truncate = nbd_co_truncate, - .bdrv_getlength = nbd_getlength, + .bdrv_co_getlength = nbd_co_getlength, .bdrv_refresh_filename = nbd_refresh_filename, .bdrv_co_block_status = nbd_client_co_block_status, .bdrv_dirname = nbd_dirname, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_nbd_unix = { .bdrv_co_pdiscard = nbd_client_co_pdiscard, .bdrv_refresh_limits = nbd_refresh_limits, .bdrv_co_truncate = nbd_co_truncate, - .bdrv_getlength = nbd_getlength, + .bdrv_co_getlength = nbd_co_getlength, .bdrv_refresh_filename = nbd_refresh_filename, .bdrv_co_block_status = nbd_client_co_block_status, .bdrv_dirname = nbd_dirname, diff --git a/block/null.c b/block/null.c index XXXXXXX..XXXXXXX 100644 --- a/block/null.c +++ b/block/null.c @@ -XXX,XX +XXX,XX @@ static int null_file_open(BlockDriverState *bs, QDict *options, int flags, return ret; } -static int64_t null_getlength(BlockDriverState *bs) +static int64_t coroutine_fn null_co_getlength(BlockDriverState *bs) { BDRVNullState *s = bs->opaque; return s->length; @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_null_co = { .bdrv_file_open = null_file_open, .bdrv_parse_filename = null_co_parse_filename, - .bdrv_getlength = null_getlength, + .bdrv_co_getlength = null_co_getlength, .bdrv_get_allocated_file_size = null_allocated_file_size, .bdrv_co_preadv = null_co_preadv, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_null_aio = { .bdrv_file_open = null_file_open, .bdrv_parse_filename = null_aio_parse_filename, - .bdrv_getlength = null_getlength, + .bdrv_co_getlength = null_co_getlength, .bdrv_get_allocated_file_size = null_allocated_file_size, .bdrv_aio_preadv = null_aio_preadv, diff --git a/block/nvme.c b/block/nvme.c index XXXXXXX..XXXXXXX 100644 --- a/block/nvme.c +++ b/block/nvme.c @@ -XXX,XX +XXX,XX @@ fail: return ret; } -static int64_t nvme_getlength(BlockDriverState *bs) +static int64_t coroutine_fn nvme_co_getlength(BlockDriverState *bs) { BDRVNVMeState *s = bs->opaque; return s->nsze << s->blkshift; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn nvme_co_truncate(BlockDriverState *bs, int64_t offset, return -ENOTSUP; } - cur_length = nvme_getlength(bs); + cur_length = nvme_co_getlength(bs); if (offset != cur_length && exact) { error_setg(errp, "Cannot resize NVMe devices"); return -ENOTSUP; @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_nvme = { .bdrv_parse_filename = nvme_parse_filename, .bdrv_file_open = nvme_file_open, .bdrv_close = nvme_close, - .bdrv_getlength = nvme_getlength, + .bdrv_co_getlength = nvme_co_getlength, .bdrv_probe_blocksizes = nvme_probe_blocksizes, .bdrv_co_truncate = nvme_co_truncate, diff --git a/block/preallocate.c b/block/preallocate.c index XXXXXXX..XXXXXXX 100644 --- a/block/preallocate.c +++ b/block/preallocate.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn preallocate_co_flush(BlockDriverState *bs) return bdrv_co_flush(bs->file->bs); } -static int64_t preallocate_getlength(BlockDriverState *bs) +static int64_t coroutine_fn preallocate_co_getlength(BlockDriverState *bs) { int64_t ret; BDRVPreallocateState *s = bs->opaque; @@ -XXX,XX +XXX,XX @@ static int64_t preallocate_getlength(BlockDriverState *bs) return s->data_end; } - ret = bdrv_getlength(bs->file->bs); + ret = bdrv_co_getlength(bs->file->bs); if (has_prealloc_perms(bs)) { s->file_end = s->zero_start = s->data_end = ret; @@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_preallocate_filter = { .format_name = "preallocate", .instance_size = sizeof(BDRVPreallocateState), - .bdrv_getlength = preallocate_getlength, - .bdrv_open = preallocate_open, - .bdrv_close = preallocate_close, + .bdrv_co_getlength = preallocate_co_getlength, + .bdrv_open = preallocate_open, + .bdrv_close = preallocate_close, .bdrv_reopen_prepare = preallocate_reopen_prepare, .bdrv_reopen_commit = preallocate_reopen_commit, 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_truncate(BlockDriverState *bs, return ret; } -static int64_t bdrv_qed_getlength(BlockDriverState *bs) +static int64_t coroutine_fn bdrv_qed_co_getlength(BlockDriverState *bs) { BDRVQEDState *s = bs->opaque; return s->header.image_size; @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_qed = { .bdrv_co_writev = bdrv_qed_co_writev, .bdrv_co_pwrite_zeroes = bdrv_qed_co_pwrite_zeroes, .bdrv_co_truncate = bdrv_qed_co_truncate, - .bdrv_getlength = bdrv_qed_getlength, + .bdrv_co_getlength = bdrv_qed_co_getlength, .bdrv_get_info = bdrv_qed_get_info, .bdrv_refresh_limits = bdrv_qed_refresh_limits, .bdrv_change_backing_file = bdrv_qed_change_backing_file, 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 int coroutine_fn quorum_co_pwrite_zeroes(BlockDriverState *bs, flags | BDRV_REQ_ZERO_WRITE); } -static int64_t quorum_getlength(BlockDriverState *bs) +static int64_t coroutine_fn quorum_co_getlength(BlockDriverState *bs) { BDRVQuorumState *s = bs->opaque; int64_t result; int i; /* check that all file have the same length */ - result = bdrv_getlength(s->children[0]->bs); + result = bdrv_co_getlength(s->children[0]->bs); if (result < 0) { return result; } for (i = 1; i < s->num_children; i++) { - int64_t value = bdrv_getlength(s->children[i]->bs); + int64_t value = bdrv_co_getlength(s->children[i]->bs); if (value < 0) { return value; } @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_quorum = { .bdrv_co_flush = quorum_co_flush, - .bdrv_getlength = quorum_getlength, + .bdrv_co_getlength = quorum_co_getlength, .bdrv_co_preadv = quorum_co_preadv, .bdrv_co_pwritev = quorum_co_pwritev, 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 int coroutine_fn raw_co_pdiscard(BlockDriverState *bs, return bdrv_co_pdiscard(bs->file, offset, bytes); } -static int64_t raw_getlength(BlockDriverState *bs) +static int64_t coroutine_fn raw_co_getlength(BlockDriverState *bs) { int64_t len; BDRVRawState *s = bs->opaque; /* Update size. It should not change unless the file was externally * modified. */ - len = bdrv_getlength(bs->file->bs); + len = bdrv_co_getlength(bs->file->bs); if (len < 0) { return len; } @@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_raw = { .bdrv_co_copy_range_from = &raw_co_copy_range_from, .bdrv_co_copy_range_to = &raw_co_copy_range_to, .bdrv_co_truncate = &raw_co_truncate, - .bdrv_getlength = &raw_getlength, + .bdrv_co_getlength = &raw_co_getlength, .is_format = true, .has_variable_length = true, .bdrv_measure = &raw_measure, diff --git a/block/rbd.c b/block/rbd.c index XXXXXXX..XXXXXXX 100644 --- a/block/rbd.c +++ b/block/rbd.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qemu_rbd_co_block_status(BlockDriverState *bs, return status; } -static int64_t qemu_rbd_getlength(BlockDriverState *bs) +static int64_t coroutine_fn qemu_rbd_co_getlength(BlockDriverState *bs) { BDRVRBDState *s = bs->opaque; int r; @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_rbd = { .bdrv_get_info = qemu_rbd_getinfo, .bdrv_get_specific_info = qemu_rbd_get_specific_info, .create_opts = &qemu_rbd_create_opts, - .bdrv_getlength = qemu_rbd_getlength, + .bdrv_co_getlength = qemu_rbd_co_getlength, .bdrv_co_truncate = qemu_rbd_co_truncate, .protocol_name = "rbd", diff --git a/block/replication.c b/block/replication.c index XXXXXXX..XXXXXXX 100644 --- a/block/replication.c +++ b/block/replication.c @@ -XXX,XX +XXX,XX @@ static void replication_child_perm(BlockDriverState *bs, BdrvChild *c, return; } -static int64_t replication_getlength(BlockDriverState *bs) +static int64_t coroutine_fn replication_co_getlength(BlockDriverState *bs) { - return bdrv_getlength(bs->file->bs); + return bdrv_co_getlength(bs->file->bs); } static int replication_get_io_status(BDRVReplicationState *s) @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_replication = { .bdrv_close = replication_close, .bdrv_child_perm = replication_child_perm, - .bdrv_getlength = replication_getlength, + .bdrv_co_getlength = replication_co_getlength, .bdrv_co_readv = replication_co_readv, .bdrv_co_writev = replication_co_writev, diff --git a/block/ssh.c b/block/ssh.c index XXXXXXX..XXXXXXX 100644 --- a/block/ssh.c +++ b/block/ssh.c @@ -XXX,XX +XXX,XX @@ static coroutine_fn int ssh_co_flush(BlockDriverState *bs) return ret; } -static int64_t ssh_getlength(BlockDriverState *bs) +static int64_t coroutine_fn ssh_co_getlength(BlockDriverState *bs) { BDRVSSHState *s = bs->opaque; int64_t length; @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_ssh = { .bdrv_has_zero_init = ssh_has_zero_init, .bdrv_co_readv = ssh_co_readv, .bdrv_co_writev = ssh_co_writev, - .bdrv_getlength = ssh_getlength, + .bdrv_co_getlength = ssh_co_getlength, .bdrv_co_truncate = ssh_co_truncate, .bdrv_co_flush_to_disk = ssh_co_flush, .bdrv_refresh_filename = ssh_refresh_filename, diff --git a/block/throttle.c b/block/throttle.c index XXXXXXX..XXXXXXX 100644 --- a/block/throttle.c +++ b/block/throttle.c @@ -XXX,XX +XXX,XX @@ static void throttle_close(BlockDriverState *bs) } -static int64_t throttle_getlength(BlockDriverState *bs) +static int64_t coroutine_fn throttle_co_getlength(BlockDriverState *bs) { - return bdrv_getlength(bs->file->bs); + return bdrv_co_getlength(bs->file->bs); } static int coroutine_fn throttle_co_preadv(BlockDriverState *bs, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_throttle = { .bdrv_child_perm = bdrv_default_perms, - .bdrv_getlength = throttle_getlength, + .bdrv_co_getlength = throttle_co_getlength, .bdrv_co_preadv = throttle_co_preadv, .bdrv_co_pwritev = throttle_co_pwritev, diff --git a/hw/scsi/scsi-disk.c b/hw/scsi/scsi-disk.c index XXXXXXX..XXXXXXX 100644 --- a/hw/scsi/scsi-disk.c +++ b/hw/scsi/scsi-disk.c @@ -XXX,XX +XXX,XX @@ static void scsi_disk_reset(DeviceState *dev) { SCSIDiskState *s = DO_UPCAST(SCSIDiskState, qdev.qdev, dev); uint64_t nb_sectors; + AioContext *ctx; scsi_device_purge_requests(&s->qdev, SENSE_CODE(RESET)); + ctx = blk_get_aio_context(s->qdev.conf.blk); + aio_context_acquire(ctx); blk_get_geometry(s->qdev.conf.blk, &nb_sectors); + aio_context_release(ctx); + nb_sectors /= s->qdev.blocksize / BDRV_SECTOR_SIZE; if (nb_sectors) { nb_sectors--; diff --git a/tests/unit/test-block-iothread.c b/tests/unit/test-block-iothread.c index XXXXXXX..XXXXXXX 100644 --- a/tests/unit/test-block-iothread.c +++ b/tests/unit/test-block-iothread.c @@ -XXX,XX +XXX,XX @@ static void test_attach_second_node(void) qdict_put_str(options, "driver", "raw"); qdict_put_str(options, "file", "base"); + aio_context_acquire(ctx); filter = bdrv_open(NULL, NULL, options, BDRV_O_RDWR, &error_abort); + aio_context_release(ctx); + g_assert(blk_get_aio_context(blk) == ctx); g_assert(bdrv_get_aio_context(bs) == ctx); g_assert(bdrv_get_aio_context(filter) == ctx); 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_gen_c = custom_target('block-gen.c', input: files( '../include/block/block-io.h', '../include/block/dirty-bitmap.h', + '../include/block/block_int-io.h', '../include/block/block-global-state.h', '../include/sysemu/block-backend-io.h', 'coroutines.h' -- 2.38.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> The only difference is that blk_ checks if the block is available, but this check is already performed above in blk_check_byte_request(). This is in preparation for the graph rdlock, which will be taken by both the callers of blk_check_byte_request() and blk_getlength(). Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230113204212.359076-8-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/block-backend.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) 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 @@ static int blk_check_byte_request(BlockBackend *blk, int64_t offset, } if (!blk->allow_write_beyond_eof) { - len = blk_getlength(blk); + len = bdrv_getlength(blk_bs(blk)); if (len < 0) { return len; } -- 2.38.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> In some places we are sure we are always running in a coroutine, therefore it's useless to call the generated_co_wrapper, instead call directly the _co_ function. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230113204212.359076-9-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/block-backend.c | 6 +++--- block/io.c | 4 ++-- block/preallocate.c | 6 +++--- block/qed.c | 2 +- 4 files changed, 9 insertions(+), 9 deletions(-) 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_set_disable_request_queuing(BlockBackend *blk, bool disable) blk->disable_request_queuing = disable; } -static int blk_check_byte_request(BlockBackend *blk, int64_t offset, - int64_t bytes) +static coroutine_fn int blk_check_byte_request(BlockBackend *blk, + int64_t offset, int64_t bytes) { int64_t len; @@ -XXX,XX +XXX,XX @@ static int blk_check_byte_request(BlockBackend *blk, int64_t offset, } if (!blk->allow_write_beyond_eof) { - len = bdrv_getlength(blk_bs(blk)); + len = bdrv_co_getlength(blk_bs(blk)); if (len < 0) { return len; } 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 @@ int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact, if (new_bytes && backing) { int64_t backing_len; - backing_len = bdrv_getlength(backing->bs); + backing_len = bdrv_co_getlength(backing->bs); if (backing_len < 0) { ret = backing_len; error_setg_errno(errp, -ret, "Could not get backing file size"); @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact, goto out; } - ret = bdrv_refresh_total_sectors(bs, offset >> BDRV_SECTOR_BITS); + ret = bdrv_co_refresh_total_sectors(bs, offset >> BDRV_SECTOR_BITS); if (ret < 0) { error_setg_errno(errp, -ret, "Could not refresh total sector count"); } else { diff --git a/block/preallocate.c b/block/preallocate.c index XXXXXXX..XXXXXXX 100644 --- a/block/preallocate.c +++ b/block/preallocate.c @@ -XXX,XX +XXX,XX @@ static bool coroutine_fn handle_write(BlockDriverState *bs, int64_t offset, } if (s->data_end < 0) { - s->data_end = bdrv_getlength(bs->file->bs); + s->data_end = bdrv_co_getlength(bs->file->bs); if (s->data_end < 0) { return false; } @@ -XXX,XX +XXX,XX @@ static bool coroutine_fn handle_write(BlockDriverState *bs, int64_t offset, } if (s->file_end < 0) { - s->file_end = bdrv_getlength(bs->file->bs); + s->file_end = bdrv_co_getlength(bs->file->bs); if (s->file_end < 0) { return false; } @@ -XXX,XX +XXX,XX @@ preallocate_co_truncate(BlockDriverState *bs, int64_t offset, if (s->data_end >= 0 && offset > s->data_end) { if (s->file_end < 0) { - s->file_end = bdrv_getlength(bs->file->bs); + s->file_end = bdrv_co_getlength(bs->file->bs); if (s->file_end < 0) { error_setg(errp, "failed to get file length"); return s->file_end; 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_do_open(BlockDriverState *bs, QDict *options, } /* Round down file size to the last cluster */ - file_size = bdrv_getlength(bs->file->bs); + file_size = bdrv_co_getlength(bs->file->bs); if (file_size < 0) { error_setg(errp, "Failed to get file length"); return file_size; -- 2.38.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> bdrv_get_allocated_file_size() is categorized as an I/O function, and it currently doesn't run in a coroutine. We should let it take a graph rdlock since it traverses the block nodes graph, which however is only possible in a coroutine. Therefore turn it into a co_wrapper to move the actual function into a coroutine where the lock can be taken. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230113204212.359076-10-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-io.h | 4 +++- include/block/block_int-common.h | 4 +++- block.c | 12 ++++++------ block/file-posix.c | 14 +++++--------- block/file-win32.c | 10 ++++------ block/gluster.c | 11 ++++++----- block/nfs.c | 4 ++-- block/null.c | 7 ++++--- block/qcow2-refcount.c | 2 +- block/vmdk.c | 9 +++++---- 10 files changed, 39 insertions(+), 38 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 co_wrapper_mixed bdrv_nb_sectors(BlockDriverState *bs); int64_t coroutine_fn bdrv_co_getlength(BlockDriverState *bs); int64_t co_wrapper_mixed bdrv_getlength(BlockDriverState *bs); -int64_t bdrv_get_allocated_file_size(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); + BlockMeasureInfo *bdrv_measure(BlockDriver *drv, QemuOpts *opts, BlockDriverState *in_bs, Error **errp); void bdrv_get_geometry(BlockDriverState *bs, uint64_t *nb_sectors_ptr); 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 { bool exact, PreallocMode prealloc, BdrvRequestFlags flags, Error **errp); int64_t coroutine_fn (*bdrv_co_getlength)(BlockDriverState *bs); - int64_t (*bdrv_get_allocated_file_size)(BlockDriverState *bs); + int64_t coroutine_fn (*bdrv_co_get_allocated_file_size)( + BlockDriverState *bs); + BlockMeasureInfo *(*bdrv_measure)(QemuOpts *opts, BlockDriverState *in_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 @@ exit: } /** - * Implementation of BlockDriver.bdrv_get_allocated_file_size() that + * Implementation of BlockDriver.bdrv_co_get_allocated_file_size() that * sums the size of all data-bearing children. (This excludes backing * children.) */ @@ -XXX,XX +XXX,XX @@ static int64_t bdrv_sum_allocated_file_size(BlockDriverState *bs) if (child->role & (BDRV_CHILD_DATA | BDRV_CHILD_METADATA | BDRV_CHILD_FILTERED)) { - child_size = bdrv_get_allocated_file_size(child->bs); + child_size = bdrv_co_get_allocated_file_size(child->bs); if (child_size < 0) { return child_size; } @@ -XXX,XX +XXX,XX @@ static int64_t bdrv_sum_allocated_file_size(BlockDriverState *bs) * Length of a allocated file in bytes. Sparse files are counted by actual * allocated space. Return < 0 if error or unknown. */ -int64_t bdrv_get_allocated_file_size(BlockDriverState *bs) +int64_t coroutine_fn bdrv_co_get_allocated_file_size(BlockDriverState *bs) { BlockDriver *drv = bs->drv; IO_CODE(); @@ -XXX,XX +XXX,XX @@ int64_t bdrv_get_allocated_file_size(BlockDriverState *bs) if (!drv) { return -ENOMEDIUM; } - if (drv->bdrv_get_allocated_file_size) { - return drv->bdrv_get_allocated_file_size(bs); + if (drv->bdrv_co_get_allocated_file_size) { + return drv->bdrv_co_get_allocated_file_size(bs); } if (drv->bdrv_file_open) { @@ -XXX,XX +XXX,XX @@ int64_t bdrv_get_allocated_file_size(BlockDriverState *bs) return -ENOTSUP; } else if (drv->is_filter) { /* Filter drivers default to the size of their filtered child */ - return bdrv_get_allocated_file_size(bdrv_filter_bs(bs)); + return bdrv_co_get_allocated_file_size(bdrv_filter_bs(bs)); } else { /* Other drivers default to summing their children's sizes */ return bdrv_sum_allocated_file_size(bs); diff --git a/block/file-posix.c b/block/file-posix.c index XXXXXXX..XXXXXXX 100644 --- a/block/file-posix.c +++ b/block/file-posix.c @@ -XXX,XX +XXX,XX @@ static int64_t coroutine_fn raw_co_getlength(BlockDriverState *bs) } #endif -static int64_t raw_get_allocated_file_size(BlockDriverState *bs) +static int64_t coroutine_fn raw_co_get_allocated_file_size(BlockDriverState *bs) { struct stat st; BDRVRawState *s = bs->opaque; @@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_file = { .bdrv_co_truncate = raw_co_truncate, .bdrv_co_getlength = raw_co_getlength, .bdrv_get_info = raw_get_info, - .bdrv_get_allocated_file_size - = raw_get_allocated_file_size, + .bdrv_co_get_allocated_file_size = raw_co_get_allocated_file_size, .bdrv_get_specific_stats = raw_get_specific_stats, .bdrv_check_perm = raw_check_perm, .bdrv_set_perm = raw_set_perm, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_host_device = { .bdrv_co_truncate = raw_co_truncate, .bdrv_co_getlength = raw_co_getlength, .bdrv_get_info = raw_get_info, - .bdrv_get_allocated_file_size - = raw_get_allocated_file_size, + .bdrv_co_get_allocated_file_size = raw_co_get_allocated_file_size, .bdrv_get_specific_stats = hdev_get_specific_stats, .bdrv_check_perm = raw_check_perm, .bdrv_set_perm = raw_set_perm, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_host_cdrom = { .bdrv_co_truncate = raw_co_truncate, .bdrv_co_getlength = raw_co_getlength, .has_variable_length = true, - .bdrv_get_allocated_file_size - = raw_get_allocated_file_size, + .bdrv_co_get_allocated_file_size = raw_co_get_allocated_file_size, /* removable device support */ .bdrv_co_is_inserted = cdrom_co_is_inserted, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_host_cdrom = { .bdrv_co_truncate = raw_co_truncate, .bdrv_co_getlength = raw_co_getlength, .has_variable_length = true, - .bdrv_get_allocated_file_size - = raw_get_allocated_file_size, + .bdrv_co_get_allocated_file_size = raw_co_get_allocated_file_size, /* removable device support */ .bdrv_co_is_inserted = cdrom_co_is_inserted, diff --git a/block/file-win32.c b/block/file-win32.c index XXXXXXX..XXXXXXX 100644 --- a/block/file-win32.c +++ b/block/file-win32.c @@ -XXX,XX +XXX,XX @@ static int64_t coroutine_fn raw_co_getlength(BlockDriverState *bs) return l.QuadPart; } -static int64_t raw_get_allocated_file_size(BlockDriverState *bs) +static int64_t coroutine_fn raw_co_get_allocated_file_size(BlockDriverState *bs) { typedef DWORD (WINAPI * get_compressed_t)(const char *filename, DWORD * high); @@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_file = { .bdrv_co_truncate = raw_co_truncate, .bdrv_co_getlength = raw_co_getlength, - .bdrv_get_allocated_file_size - = raw_get_allocated_file_size, + .bdrv_co_get_allocated_file_size + = raw_co_get_allocated_file_size, .create_opts = &raw_create_opts, }; @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_host_device = { .bdrv_co_getlength = raw_co_getlength, .has_variable_length = true, - - .bdrv_get_allocated_file_size - = raw_get_allocated_file_size, + .bdrv_co_get_allocated_file_size = raw_co_get_allocated_file_size, }; static void bdrv_file_init(void) diff --git a/block/gluster.c b/block/gluster.c index XXXXXXX..XXXXXXX 100644 --- a/block/gluster.c +++ b/block/gluster.c @@ -XXX,XX +XXX,XX @@ static int64_t coroutine_fn qemu_gluster_co_getlength(BlockDriverState *bs) } } -static int64_t qemu_gluster_allocated_file_size(BlockDriverState *bs) +static int64_t coroutine_fn +qemu_gluster_co_get_allocated_file_size(BlockDriverState *bs) { BDRVGlusterState *s = bs->opaque; struct stat st; @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_gluster = { .bdrv_co_create = qemu_gluster_co_create, .bdrv_co_create_opts = qemu_gluster_co_create_opts, .bdrv_co_getlength = qemu_gluster_co_getlength, - .bdrv_get_allocated_file_size = qemu_gluster_allocated_file_size, + .bdrv_co_get_allocated_file_size = qemu_gluster_co_get_allocated_file_size, .bdrv_co_truncate = qemu_gluster_co_truncate, .bdrv_co_readv = qemu_gluster_co_readv, .bdrv_co_writev = qemu_gluster_co_writev, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_gluster_tcp = { .bdrv_co_create = qemu_gluster_co_create, .bdrv_co_create_opts = qemu_gluster_co_create_opts, .bdrv_co_getlength = qemu_gluster_co_getlength, - .bdrv_get_allocated_file_size = qemu_gluster_allocated_file_size, + .bdrv_co_get_allocated_file_size = qemu_gluster_co_get_allocated_file_size, .bdrv_co_truncate = qemu_gluster_co_truncate, .bdrv_co_readv = qemu_gluster_co_readv, .bdrv_co_writev = qemu_gluster_co_writev, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_gluster_unix = { .bdrv_co_create = qemu_gluster_co_create, .bdrv_co_create_opts = qemu_gluster_co_create_opts, .bdrv_co_getlength = qemu_gluster_co_getlength, - .bdrv_get_allocated_file_size = qemu_gluster_allocated_file_size, + .bdrv_co_get_allocated_file_size = qemu_gluster_co_get_allocated_file_size, .bdrv_co_truncate = qemu_gluster_co_truncate, .bdrv_co_readv = qemu_gluster_co_readv, .bdrv_co_writev = qemu_gluster_co_writev, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_gluster_rdma = { .bdrv_co_create = qemu_gluster_co_create, .bdrv_co_create_opts = qemu_gluster_co_create_opts, .bdrv_co_getlength = qemu_gluster_co_getlength, - .bdrv_get_allocated_file_size = qemu_gluster_allocated_file_size, + .bdrv_co_get_allocated_file_size = qemu_gluster_co_get_allocated_file_size, .bdrv_co_truncate = qemu_gluster_co_truncate, .bdrv_co_readv = qemu_gluster_co_readv, .bdrv_co_writev = qemu_gluster_co_writev, diff --git a/block/nfs.c b/block/nfs.c index XXXXXXX..XXXXXXX 100644 --- a/block/nfs.c +++ b/block/nfs.c @@ -XXX,XX +XXX,XX @@ nfs_get_allocated_file_size_cb(int ret, struct nfs_context *nfs, void *data, bdrv_wakeup(task->bs); } -static int64_t nfs_get_allocated_file_size(BlockDriverState *bs) +static int64_t coroutine_fn nfs_co_get_allocated_file_size(BlockDriverState *bs) { NFSClient *client = bs->opaque; NFSRPC task = {0}; @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_nfs = { .bdrv_has_zero_init = nfs_has_zero_init, /* libnfs does not provide the allocated filesize of a file on win32. */ #if !defined(_WIN32) - .bdrv_get_allocated_file_size = nfs_get_allocated_file_size, + .bdrv_co_get_allocated_file_size = nfs_co_get_allocated_file_size, #endif .bdrv_co_truncate = nfs_file_co_truncate, diff --git a/block/null.c b/block/null.c index XXXXXXX..XXXXXXX 100644 --- a/block/null.c +++ b/block/null.c @@ -XXX,XX +XXX,XX @@ static void null_refresh_filename(BlockDriverState *bs) bs->drv->format_name); } -static int64_t null_allocated_file_size(BlockDriverState *bs) +static int64_t coroutine_fn +null_co_get_allocated_file_size(BlockDriverState *bs) { return 0; } @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_null_co = { .bdrv_file_open = null_file_open, .bdrv_parse_filename = null_co_parse_filename, .bdrv_co_getlength = null_co_getlength, - .bdrv_get_allocated_file_size = null_allocated_file_size, + .bdrv_co_get_allocated_file_size = null_co_get_allocated_file_size, .bdrv_co_preadv = null_co_preadv, .bdrv_co_pwritev = null_co_pwritev, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_null_aio = { .bdrv_file_open = null_file_open, .bdrv_parse_filename = null_aio_parse_filename, .bdrv_co_getlength = null_co_getlength, - .bdrv_get_allocated_file_size = null_allocated_file_size, + .bdrv_co_get_allocated_file_size = null_co_get_allocated_file_size, .bdrv_aio_preadv = null_aio_preadv, .bdrv_aio_pwritev = null_aio_pwritev, 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) return file_length; } - real_allocation = bdrv_get_allocated_file_size(bs->file->bs); + real_allocation = bdrv_co_get_allocated_file_size(bs->file->bs); if (real_allocation < 0) { return real_allocation; } 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 vmdk_get_allocated_file_size(BlockDriverState *bs) +static int64_t coroutine_fn +vmdk_co_get_allocated_file_size(BlockDriverState *bs) { int i; int64_t ret = 0; int64_t r; BDRVVmdkState *s = bs->opaque; - ret = bdrv_get_allocated_file_size(bs->file->bs); + ret = bdrv_co_get_allocated_file_size(bs->file->bs); if (ret < 0) { return ret; } @@ -XXX,XX +XXX,XX @@ static int64_t vmdk_get_allocated_file_size(BlockDriverState *bs) if (s->extents[i].file == bs->file) { continue; } - r = bdrv_get_allocated_file_size(s->extents[i].file->bs); + r = bdrv_co_get_allocated_file_size(s->extents[i].file->bs); if (r < 0) { return r; } @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_vmdk = { .bdrv_co_create_opts = vmdk_co_create_opts, .bdrv_co_create = vmdk_co_create, .bdrv_co_block_status = vmdk_co_block_status, - .bdrv_get_allocated_file_size = vmdk_get_allocated_file_size, + .bdrv_co_get_allocated_file_size = vmdk_co_get_allocated_file_size, .bdrv_has_zero_init = vmdk_has_zero_init, .bdrv_get_specific_info = vmdk_get_specific_info, .bdrv_refresh_limits = vmdk_refresh_limits, -- 2.38.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> bdrv_get_info() is categorized as an I/O function, and it currently doesn't run in a coroutine. We should let it take a graph rdlock since it traverses the block nodes graph, which however is only possible in a coroutine. Therefore turn it into a co_wrapper to move the actual function into a coroutine where the lock can be taken. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230113204212.359076-11-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-io.h | 5 ++++- include/block/block_int-common.h | 3 ++- block.c | 8 ++++---- block/blkio.c | 5 +++-- block/crypto.c | 8 ++++---- block/file-posix.c | 7 ++++--- block/io.c | 8 ++++---- block/iscsi.c | 7 ++++--- block/mirror.c | 2 +- block/qcow.c | 5 +++-- block/qcow2.c | 5 +++-- block/qed.c | 5 +++-- block/raw-format.c | 7 ++++--- block/rbd.c | 5 +++-- block/vdi.c | 7 ++++--- block/vhdx.c | 5 +++-- block/vmdk.c | 5 +++-- block/vpc.c | 5 +++-- 18 files changed, 59 insertions(+), 43 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 @@ bool bdrv_supports_compressed_writes(BlockDriverState *bs); 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 bdrv_get_info(BlockDriverState *bs, BlockDriverInfo *bdi); + +int coroutine_fn bdrv_co_get_info(BlockDriverState *bs, BlockDriverInfo *bdi); +int co_wrapper_mixed bdrv_get_info(BlockDriverState *bs, BlockDriverInfo *bdi); + ImageInfoSpecific *bdrv_get_specific_info(BlockDriverState *bs, Error **errp); BlockStatsSpecific *bdrv_get_specific_stats(BlockDriverState *bs); 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 offset, int64_t bytes, QEMUIOVector *qiov, size_t qiov_offset); - int (*bdrv_get_info)(BlockDriverState *bs, BlockDriverInfo *bdi); + int coroutine_fn (*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 @@ void bdrv_get_backing_filename(BlockDriverState *bs, pstrcpy(filename, filename_size, bs->backing_file); } -int bdrv_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) +int coroutine_fn bdrv_co_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) { int ret; BlockDriver *drv = bs->drv; @@ -XXX,XX +XXX,XX @@ int bdrv_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) if (!drv) { return -ENOMEDIUM; } - if (!drv->bdrv_get_info) { + if (!drv->bdrv_co_get_info) { BlockDriverState *filtered = bdrv_filter_bs(bs); if (filtered) { - return bdrv_get_info(filtered, bdi); + return bdrv_co_get_info(filtered, bdi); } return -ENOTSUP; } memset(bdi, 0, sizeof(*bdi)); - ret = drv->bdrv_get_info(bs, bdi); + ret = drv->bdrv_co_get_info(bs, bdi); if (ret < 0) { return ret; } diff --git a/block/blkio.c b/block/blkio.c index XXXXXXX..XXXXXXX 100644 --- a/block/blkio.c +++ b/block/blkio.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn blkio_truncate(BlockDriverState *bs, int64_t offset, return 0; } -static int blkio_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) +static int coroutine_fn +blkio_co_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) { return 0; } @@ -XXX,XX +XXX,XX @@ static void blkio_refresh_limits(BlockDriverState *bs, Error **errp) .bdrv_close = blkio_close, \ .bdrv_co_getlength = blkio_co_getlength, \ .bdrv_co_truncate = blkio_truncate, \ - .bdrv_get_info = blkio_get_info, \ + .bdrv_co_get_info = blkio_co_get_info, \ .bdrv_attach_aio_context = blkio_attach_aio_context, \ .bdrv_detach_aio_context = blkio_detach_aio_context, \ .bdrv_co_pdiscard = blkio_co_pdiscard, \ 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 block_crypto_get_info_luks(BlockDriverState *bs, - BlockDriverInfo *bdi) +static int coroutine_fn +block_crypto_co_get_info_luks(BlockDriverState *bs, BlockDriverInfo *bdi) { BlockDriverInfo subbdi; int ret; - ret = bdrv_get_info(bs->file->bs, &subbdi); + ret = bdrv_co_get_info(bs->file->bs, &subbdi); if (ret != 0) { return ret; } @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_crypto_luks = { .bdrv_co_pwritev = block_crypto_co_pwritev, .bdrv_co_getlength = block_crypto_co_getlength, .bdrv_measure = block_crypto_measure, - .bdrv_get_info = block_crypto_get_info_luks, + .bdrv_co_get_info = block_crypto_co_get_info_luks, .bdrv_get_specific_info = block_crypto_get_specific_info_luks, .bdrv_amend_options = block_crypto_amend_options_luks, .bdrv_co_amend = block_crypto_co_amend_luks, diff --git a/block/file-posix.c b/block/file-posix.c index XXXXXXX..XXXXXXX 100644 --- a/block/file-posix.c +++ b/block/file-posix.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn raw_co_pwrite_zeroes( return raw_do_pwrite_zeroes(bs, offset, bytes, flags, false); } -static int raw_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) +static int coroutine_fn +raw_co_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) { return 0; } @@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_file = { .bdrv_co_truncate = raw_co_truncate, .bdrv_co_getlength = raw_co_getlength, - .bdrv_get_info = raw_get_info, + .bdrv_co_get_info = raw_co_get_info, .bdrv_co_get_allocated_file_size = raw_co_get_allocated_file_size, .bdrv_get_specific_stats = raw_get_specific_stats, .bdrv_check_perm = raw_check_perm, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_host_device = { .bdrv_co_truncate = raw_co_truncate, .bdrv_co_getlength = raw_co_getlength, - .bdrv_get_info = raw_get_info, + .bdrv_co_get_info = raw_co_get_info, .bdrv_co_get_allocated_file_size = raw_co_get_allocated_file_size, .bdrv_get_specific_stats = hdev_get_specific_stats, .bdrv_check_perm = raw_check_perm, 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 bdrv_round_to_clusters(BlockDriverState *bs, +void coroutine_fn bdrv_round_to_clusters(BlockDriverState *bs, int64_t offset, int64_t bytes, int64_t *cluster_offset, int64_t *cluster_bytes) { BlockDriverInfo bdi; IO_CODE(); - if (bdrv_get_info(bs, &bdi) < 0 || bdi.cluster_size == 0) { + if (bdrv_co_get_info(bs, &bdi) < 0 || bdi.cluster_size == 0) { *cluster_offset = offset; *cluster_bytes = bytes; } else { @@ -XXX,XX +XXX,XX @@ void bdrv_round_to_clusters(BlockDriverState *bs, } } -static int bdrv_get_cluster_size(BlockDriverState *bs) +static coroutine_fn int bdrv_get_cluster_size(BlockDriverState *bs) { BlockDriverInfo bdi; int ret; - ret = bdrv_get_info(bs, &bdi); + ret = bdrv_co_get_info(bs, &bdi); if (ret < 0 || bdi.cluster_size == 0) { return bs->bl.request_alignment; } else { diff --git a/block/iscsi.c b/block/iscsi.c index XXXXXXX..XXXXXXX 100644 --- a/block/iscsi.c +++ b/block/iscsi.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn iscsi_co_truncate(BlockDriverState *bs, int64_t offset, return 0; } -static int iscsi_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) +static int coroutine_fn +iscsi_co_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) { IscsiLun *iscsilun = bs->opaque; bdi->cluster_size = iscsilun->cluster_size; @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_iscsi = { .bdrv_co_invalidate_cache = iscsi_co_invalidate_cache, .bdrv_co_getlength = iscsi_co_getlength, - .bdrv_get_info = iscsi_get_info, + .bdrv_co_get_info = iscsi_co_get_info, .bdrv_co_truncate = iscsi_co_truncate, .bdrv_refresh_limits = iscsi_refresh_limits, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_iser = { .bdrv_co_invalidate_cache = iscsi_co_invalidate_cache, .bdrv_co_getlength = iscsi_co_getlength, - .bdrv_get_info = iscsi_get_info, + .bdrv_co_get_info = iscsi_co_get_info, .bdrv_co_truncate = iscsi_co_truncate, .bdrv_refresh_limits = iscsi_refresh_limits, 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 coroutine_fn mirror_run(Job *job, Error **errp) */ bdrv_get_backing_filename(target_bs, backing_filename, sizeof(backing_filename)); - if (!bdrv_get_info(target_bs, &bdi) && bdi.cluster_size) { + 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; 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 @@ fail: return ret; } -static int qcow_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) +static int coroutine_fn +qcow_co_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) { BDRVQcowState *s = bs->opaque; bdi->cluster_size = s->cluster_size; @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_qcow = { .bdrv_make_empty = qcow_make_empty, .bdrv_co_pwritev_compressed = qcow_co_pwritev_compressed, - .bdrv_get_info = qcow_get_info, + .bdrv_co_get_info = qcow_co_get_info, .create_opts = &qcow_create_opts, .strong_runtime_opts = qcow_strong_runtime_opts, 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 @@ err: return NULL; } -static int qcow2_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) +static int coroutine_fn +qcow2_co_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) { BDRVQcow2State *s = bs->opaque; bdi->cluster_size = s->cluster_size; @@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_qcow2 = { .bdrv_snapshot_list = qcow2_snapshot_list, .bdrv_snapshot_load_tmp = qcow2_snapshot_load_tmp, .bdrv_measure = qcow2_measure, - .bdrv_get_info = qcow2_get_info, + .bdrv_co_get_info = qcow2_co_get_info, .bdrv_get_specific_info = qcow2_get_specific_info, .bdrv_save_vmstate = qcow2_save_vmstate, 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 int64_t coroutine_fn bdrv_qed_co_getlength(BlockDriverState *bs) return s->header.image_size; } -static int bdrv_qed_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) +static int coroutine_fn +bdrv_qed_co_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) { BDRVQEDState *s = bs->opaque; @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_qed = { .bdrv_co_pwrite_zeroes = bdrv_qed_co_pwrite_zeroes, .bdrv_co_truncate = bdrv_qed_co_truncate, .bdrv_co_getlength = bdrv_qed_co_getlength, - .bdrv_get_info = bdrv_qed_get_info, + .bdrv_co_get_info = bdrv_qed_co_get_info, .bdrv_refresh_limits = bdrv_qed_refresh_limits, .bdrv_change_backing_file = bdrv_qed_change_backing_file, .bdrv_co_invalidate_cache = bdrv_qed_co_invalidate_cache, 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 raw_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) +static int coroutine_fn +raw_co_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) { - return bdrv_get_info(bs->file->bs, bdi); + return bdrv_co_get_info(bs->file->bs, bdi); } static void raw_refresh_limits(BlockDriverState *bs, Error **errp) @@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_raw = { .is_format = true, .has_variable_length = true, .bdrv_measure = &raw_measure, - .bdrv_get_info = &raw_get_info, + .bdrv_co_get_info = &raw_co_get_info, .bdrv_refresh_limits = &raw_refresh_limits, .bdrv_probe_blocksizes = &raw_probe_blocksizes, .bdrv_probe_geometry = &raw_probe_geometry, diff --git a/block/rbd.c b/block/rbd.c index XXXXXXX..XXXXXXX 100644 --- a/block/rbd.c +++ b/block/rbd.c @@ -XXX,XX +XXX,XX @@ coroutine_fn qemu_rbd_co_pwrite_zeroes(BlockDriverState *bs, int64_t offset, } #endif -static int qemu_rbd_getinfo(BlockDriverState *bs, BlockDriverInfo *bdi) +static int coroutine_fn +qemu_rbd_co_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) { BDRVRBDState *s = bs->opaque; bdi->cluster_size = s->object_size; @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_rbd = { .bdrv_co_create = qemu_rbd_co_create, .bdrv_co_create_opts = qemu_rbd_co_create_opts, .bdrv_has_zero_init = bdrv_has_zero_init_1, - .bdrv_get_info = qemu_rbd_getinfo, + .bdrv_co_get_info = qemu_rbd_co_get_info, .bdrv_get_specific_info = qemu_rbd_get_specific_info, .create_opts = &qemu_rbd_create_opts, .bdrv_co_getlength = qemu_rbd_co_getlength, 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_check(BlockDriverState *bs, BdrvCheckResult *res, return 0; } -static int vdi_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) +static int coroutine_fn +vdi_co_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) { - /* TODO: vdi_get_info would be needed for machine snapshots. + /* TODO: vdi_co_get_info would be needed for machine snapshots. vm_state_offset is still missing. */ BDRVVdiState *s = (BDRVVdiState *)bs->opaque; logout("\n"); @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_vdi = { .bdrv_co_pwritev = vdi_co_pwritev, #endif - .bdrv_get_info = vdi_get_info, + .bdrv_co_get_info = vdi_co_get_info, .is_format = true, .create_opts = &vdi_create_opts, 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 void vhdx_block_translate(BDRVVHDXState *s, int64_t sector_num, } -static int vhdx_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) +static int coroutine_fn +vhdx_co_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) { BDRVVHDXState *s = bs->opaque; @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_vhdx = { .bdrv_co_writev = vhdx_co_writev, .bdrv_co_create = vhdx_co_create, .bdrv_co_create_opts = vhdx_co_create_opts, - .bdrv_get_info = vhdx_get_info, + .bdrv_co_get_info = vhdx_co_get_info, .bdrv_co_check = vhdx_co_check, .bdrv_has_zero_init = vhdx_has_zero_init, 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 bool vmdk_extents_type_eq(const VmdkExtent *a, const VmdkExtent *b) (a->flat || a->cluster_sectors == b->cluster_sectors); } -static int vmdk_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) +static int coroutine_fn +vmdk_co_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) { int i; BDRVVmdkState *s = bs->opaque; @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_vmdk = { .bdrv_has_zero_init = vmdk_has_zero_init, .bdrv_get_specific_info = vmdk_get_specific_info, .bdrv_refresh_limits = vmdk_refresh_limits, - .bdrv_get_info = vmdk_get_info, + .bdrv_co_get_info = vmdk_co_get_info, .bdrv_gather_child_options = vmdk_gather_child_options, .is_format = true, 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 @@ fail: return ret; } -static int vpc_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) +static int coroutine_fn +vpc_co_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) { BDRVVPCState *s = (BDRVVPCState *)bs->opaque; @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_vpc = { .bdrv_co_pwritev = vpc_co_pwritev, .bdrv_co_block_status = vpc_co_block_status, - .bdrv_get_info = vpc_get_info, + .bdrv_co_get_info = vpc_co_get_info, .is_format = true, .create_opts = &vpc_create_opts, -- 2.38.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> bdrv_eject() is categorized as an I/O function, and it currently doesn't run in a coroutine. We should let it take a graph rdlock since it traverses the block nodes graph, which however is only possible in a coroutine. The only caller of this function is blk_eject(). Therefore make blk_eject() a co_wrapper, so that it always creates a new coroutine, and then make bdrv_eject() coroutine_fn where the lock can be taken. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230113204212.359076-12-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-io.h | 3 ++- include/block/block_int-common.h | 2 +- include/sysemu/block-backend-io.h | 4 +++- block.c | 6 +++--- block/block-backend.c | 4 ++-- block/copy-on-read.c | 6 +++--- block/file-posix.c | 8 ++++---- block/filter-compress.c | 7 ++++--- block/raw-format.c | 6 +++--- 9 files changed, 25 insertions(+), 21 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 @@ bool coroutine_fn bdrv_co_is_inserted(BlockDriverState *bs); bool co_wrapper bdrv_is_inserted(BlockDriverState *bs); void bdrv_lock_medium(BlockDriverState *bs, bool locked); -void bdrv_eject(BlockDriverState *bs, bool eject_flag); +void coroutine_fn bdrv_co_eject(BlockDriverState *bs, bool eject_flag); + const char *bdrv_get_format_name(BlockDriverState *bs); bool bdrv_supports_compressed_writes(BlockDriverState *bs); 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 { /* removable device specific */ bool coroutine_fn (*bdrv_co_is_inserted)(BlockDriverState *bs); - void (*bdrv_eject)(BlockDriverState *bs, bool eject_flag); + void coroutine_fn (*bdrv_co_eject)(BlockDriverState *bs, bool eject_flag); void (*bdrv_lock_medium)(BlockDriverState *bs, bool locked); /* to control generic scsi devices */ diff --git a/include/sysemu/block-backend-io.h b/include/sysemu/block-backend-io.h index XXXXXXX..XXXXXXX 100644 --- a/include/sysemu/block-backend-io.h +++ b/include/sysemu/block-backend-io.h @@ -XXX,XX +XXX,XX @@ bool co_wrapper_mixed blk_is_inserted(BlockBackend *blk); bool blk_is_available(BlockBackend *blk); void blk_lock_medium(BlockBackend *blk, bool locked); -void blk_eject(BlockBackend *blk, bool eject_flag); + +void coroutine_fn blk_co_eject(BlockBackend *blk, bool eject_flag); +void co_wrapper blk_eject(BlockBackend *blk, bool eject_flag); int64_t coroutine_fn blk_co_getlength(BlockBackend *blk); int64_t co_wrapper_mixed blk_getlength(BlockBackend *blk); diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ bool coroutine_fn bdrv_co_is_inserted(BlockDriverState *bs) /** * If eject_flag is TRUE, eject the media. Otherwise, close the tray */ -void bdrv_eject(BlockDriverState *bs, bool eject_flag) +void coroutine_fn bdrv_co_eject(BlockDriverState *bs, bool eject_flag) { BlockDriver *drv = bs->drv; IO_CODE(); - if (drv && drv->bdrv_eject) { - drv->bdrv_eject(bs, eject_flag); + if (drv && drv->bdrv_co_eject) { + drv->bdrv_co_eject(bs, eject_flag); } } 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_lock_medium(BlockBackend *blk, bool locked) } } -void blk_eject(BlockBackend *blk, bool eject_flag) +void coroutine_fn blk_co_eject(BlockBackend *blk, bool eject_flag) { BlockDriverState *bs = blk_bs(blk); char *id; IO_CODE(); if (bs) { - bdrv_eject(bs, eject_flag); + bdrv_co_eject(bs, eject_flag); } /* Whether or not we ejected on the backend, diff --git a/block/copy-on-read.c b/block/copy-on-read.c index XXXXXXX..XXXXXXX 100644 --- a/block/copy-on-read.c +++ b/block/copy-on-read.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn cor_co_pwritev_compressed(BlockDriverState *bs, } -static void cor_eject(BlockDriverState *bs, bool eject_flag) +static void coroutine_fn cor_co_eject(BlockDriverState *bs, bool eject_flag) { - bdrv_eject(bs->file->bs, eject_flag); + bdrv_co_eject(bs->file->bs, eject_flag); } @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_copy_on_read = { .bdrv_co_pdiscard = cor_co_pdiscard, .bdrv_co_pwritev_compressed = cor_co_pwritev_compressed, - .bdrv_eject = cor_eject, + .bdrv_co_eject = cor_co_eject, .bdrv_lock_medium = cor_lock_medium, .has_variable_length = true, diff --git a/block/file-posix.c b/block/file-posix.c index XXXXXXX..XXXXXXX 100644 --- a/block/file-posix.c +++ b/block/file-posix.c @@ -XXX,XX +XXX,XX @@ static bool coroutine_fn cdrom_co_is_inserted(BlockDriverState *bs) return ret == CDS_DISC_OK; } -static void cdrom_eject(BlockDriverState *bs, bool eject_flag) +static void coroutine_fn cdrom_co_eject(BlockDriverState *bs, bool eject_flag) { BDRVRawState *s = bs->opaque; @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_host_cdrom = { /* removable device support */ .bdrv_co_is_inserted = cdrom_co_is_inserted, - .bdrv_eject = cdrom_eject, + .bdrv_co_eject = cdrom_co_eject, .bdrv_lock_medium = cdrom_lock_medium, /* generic scsi device */ @@ -XXX,XX +XXX,XX @@ static bool coroutine_fn cdrom_co_is_inserted(BlockDriverState *bs) return raw_co_getlength(bs) > 0; } -static void cdrom_eject(BlockDriverState *bs, bool eject_flag) +static void coroutine_fn cdrom_co_eject(BlockDriverState *bs, bool eject_flag) { BDRVRawState *s = bs->opaque; @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_host_cdrom = { /* removable device support */ .bdrv_co_is_inserted = cdrom_co_is_inserted, - .bdrv_eject = cdrom_eject, + .bdrv_co_eject = cdrom_co_eject, .bdrv_lock_medium = cdrom_lock_medium, }; #endif /* __FreeBSD__ */ diff --git a/block/filter-compress.c b/block/filter-compress.c index XXXXXXX..XXXXXXX 100644 --- a/block/filter-compress.c +++ b/block/filter-compress.c @@ -XXX,XX +XXX,XX @@ static void compress_refresh_limits(BlockDriverState *bs, Error **errp) } -static void compress_eject(BlockDriverState *bs, bool eject_flag) +static void coroutine_fn +compress_co_eject(BlockDriverState *bs, bool eject_flag) { - bdrv_eject(bs->file->bs, eject_flag); + bdrv_co_eject(bs->file->bs, eject_flag); } @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_compress = { .bdrv_co_pdiscard = compress_co_pdiscard, .bdrv_refresh_limits = compress_refresh_limits, - .bdrv_eject = compress_eject, + .bdrv_co_eject = compress_co_eject, .bdrv_lock_medium = compress_lock_medium, .has_variable_length = true, 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 int coroutine_fn raw_co_truncate(BlockDriverState *bs, int64_t offset, return bdrv_co_truncate(bs->file, offset, exact, prealloc, flags, errp); } -static void raw_eject(BlockDriverState *bs, bool eject_flag) +static void coroutine_fn raw_co_eject(BlockDriverState *bs, bool eject_flag) { - bdrv_eject(bs->file->bs, eject_flag); + bdrv_co_eject(bs->file->bs, eject_flag); } static void raw_lock_medium(BlockDriverState *bs, bool locked) @@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_raw = { .bdrv_refresh_limits = &raw_refresh_limits, .bdrv_probe_blocksizes = &raw_probe_blocksizes, .bdrv_probe_geometry = &raw_probe_geometry, - .bdrv_eject = &raw_eject, + .bdrv_co_eject = &raw_co_eject, .bdrv_lock_medium = &raw_lock_medium, .bdrv_co_ioctl = &raw_co_ioctl, .create_opts = &raw_create_opts, -- 2.38.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> bdrv_lock_medium() is categorized as an I/O function, and it currently doesn't run in a coroutine. We should let it take a graph rdlock since it traverses the block nodes graph, which however is only possible in a coroutine. The only caller of this function is blk_lock_medium(). Therefore make blk_lock_medium() a co_wrapper, so that it always creates a new coroutine, and then make bdrv_lock_medium() a coroutine_fn where the lock can be taken. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230113204212.359076-13-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-io.h | 2 +- include/block/block_int-common.h | 2 +- include/sysemu/block-backend-io.h | 4 +++- block.c | 6 +++--- block/block-backend.c | 4 ++-- block/copy-on-read.c | 6 +++--- block/file-posix.c | 8 ++++---- block/filter-compress.c | 7 ++++--- block/raw-format.c | 6 +++--- 9 files changed, 24 insertions(+), 21 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 @@ int bdrv_get_flags(BlockDriverState *bs); bool coroutine_fn bdrv_co_is_inserted(BlockDriverState *bs); bool co_wrapper bdrv_is_inserted(BlockDriverState *bs); -void bdrv_lock_medium(BlockDriverState *bs, bool locked); +void coroutine_fn bdrv_co_lock_medium(BlockDriverState *bs, bool locked); void coroutine_fn bdrv_co_eject(BlockDriverState *bs, bool eject_flag); const char *bdrv_get_format_name(BlockDriverState *bs); 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 { /* removable device specific */ bool coroutine_fn (*bdrv_co_is_inserted)(BlockDriverState *bs); void coroutine_fn (*bdrv_co_eject)(BlockDriverState *bs, bool eject_flag); - void (*bdrv_lock_medium)(BlockDriverState *bs, bool locked); + void coroutine_fn (*bdrv_co_lock_medium)(BlockDriverState *bs, bool locked); /* to control generic scsi devices */ BlockAIOCB *(*bdrv_aio_ioctl)(BlockDriverState *bs, diff --git a/include/sysemu/block-backend-io.h b/include/sysemu/block-backend-io.h index XXXXXXX..XXXXXXX 100644 --- a/include/sysemu/block-backend-io.h +++ b/include/sysemu/block-backend-io.h @@ -XXX,XX +XXX,XX @@ bool coroutine_fn blk_co_is_inserted(BlockBackend *blk); bool co_wrapper_mixed blk_is_inserted(BlockBackend *blk); bool blk_is_available(BlockBackend *blk); -void blk_lock_medium(BlockBackend *blk, bool locked); + +void coroutine_fn blk_co_lock_medium(BlockBackend *blk, bool locked); +void co_wrapper blk_lock_medium(BlockBackend *blk, bool locked); void coroutine_fn blk_co_eject(BlockBackend *blk, bool eject_flag); void co_wrapper blk_eject(BlockBackend *blk, bool eject_flag); diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ void coroutine_fn bdrv_co_eject(BlockDriverState *bs, bool eject_flag) * Lock or unlock the media (if it is locked, the user won't be able * to eject it manually). */ -void bdrv_lock_medium(BlockDriverState *bs, bool locked) +void coroutine_fn bdrv_co_lock_medium(BlockDriverState *bs, bool locked) { BlockDriver *drv = bs->drv; IO_CODE(); trace_bdrv_lock_medium(bs, locked); - if (drv && drv->bdrv_lock_medium) { - drv->bdrv_lock_medium(bs, locked); + if (drv && drv->bdrv_co_lock_medium) { + drv->bdrv_co_lock_medium(bs, locked); } } 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 @@ bool blk_is_available(BlockBackend *blk) return blk_is_inserted(blk) && !blk_dev_is_tray_open(blk); } -void blk_lock_medium(BlockBackend *blk, bool locked) +void coroutine_fn blk_co_lock_medium(BlockBackend *blk, bool locked) { BlockDriverState *bs = blk_bs(blk); IO_CODE(); if (bs) { - bdrv_lock_medium(bs, locked); + bdrv_co_lock_medium(bs, locked); } } diff --git a/block/copy-on-read.c b/block/copy-on-read.c index XXXXXXX..XXXXXXX 100644 --- a/block/copy-on-read.c +++ b/block/copy-on-read.c @@ -XXX,XX +XXX,XX @@ static void coroutine_fn cor_co_eject(BlockDriverState *bs, bool eject_flag) } -static void cor_lock_medium(BlockDriverState *bs, bool locked) +static void coroutine_fn cor_co_lock_medium(BlockDriverState *bs, bool locked) { - bdrv_lock_medium(bs->file->bs, locked); + bdrv_co_lock_medium(bs->file->bs, locked); } @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_copy_on_read = { .bdrv_co_pwritev_compressed = cor_co_pwritev_compressed, .bdrv_co_eject = cor_co_eject, - .bdrv_lock_medium = cor_lock_medium, + .bdrv_co_lock_medium = cor_co_lock_medium, .has_variable_length = true, .is_filter = true, diff --git a/block/file-posix.c b/block/file-posix.c index XXXXXXX..XXXXXXX 100644 --- a/block/file-posix.c +++ b/block/file-posix.c @@ -XXX,XX +XXX,XX @@ static void coroutine_fn cdrom_co_eject(BlockDriverState *bs, bool eject_flag) } } -static void cdrom_lock_medium(BlockDriverState *bs, bool locked) +static void coroutine_fn cdrom_co_lock_medium(BlockDriverState *bs, bool locked) { BDRVRawState *s = bs->opaque; @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_host_cdrom = { /* removable device support */ .bdrv_co_is_inserted = cdrom_co_is_inserted, .bdrv_co_eject = cdrom_co_eject, - .bdrv_lock_medium = cdrom_lock_medium, + .bdrv_co_lock_medium = cdrom_co_lock_medium, /* generic scsi device */ .bdrv_co_ioctl = hdev_co_ioctl, @@ -XXX,XX +XXX,XX @@ static void coroutine_fn cdrom_co_eject(BlockDriverState *bs, bool eject_flag) cdrom_reopen(bs); } -static void cdrom_lock_medium(BlockDriverState *bs, bool locked) +static void coroutine_fn cdrom_co_lock_medium(BlockDriverState *bs, bool locked) { BDRVRawState *s = bs->opaque; @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_host_cdrom = { /* removable device support */ .bdrv_co_is_inserted = cdrom_co_is_inserted, .bdrv_co_eject = cdrom_co_eject, - .bdrv_lock_medium = cdrom_lock_medium, + .bdrv_co_lock_medium = cdrom_co_lock_medium, }; #endif /* __FreeBSD__ */ diff --git a/block/filter-compress.c b/block/filter-compress.c index XXXXXXX..XXXXXXX 100644 --- a/block/filter-compress.c +++ b/block/filter-compress.c @@ -XXX,XX +XXX,XX @@ compress_co_eject(BlockDriverState *bs, bool eject_flag) } -static void compress_lock_medium(BlockDriverState *bs, bool locked) +static void coroutine_fn +compress_co_lock_medium(BlockDriverState *bs, bool locked) { - bdrv_lock_medium(bs->file->bs, locked); + bdrv_co_lock_medium(bs->file->bs, locked); } @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_compress = { .bdrv_refresh_limits = compress_refresh_limits, .bdrv_co_eject = compress_co_eject, - .bdrv_lock_medium = compress_lock_medium, + .bdrv_co_lock_medium = compress_co_lock_medium, .has_variable_length = true, .is_filter = true, 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 void coroutine_fn raw_co_eject(BlockDriverState *bs, bool eject_flag) bdrv_co_eject(bs->file->bs, eject_flag); } -static void raw_lock_medium(BlockDriverState *bs, bool locked) +static void coroutine_fn raw_co_lock_medium(BlockDriverState *bs, bool locked) { - bdrv_lock_medium(bs->file->bs, locked); + bdrv_co_lock_medium(bs->file->bs, locked); } static int coroutine_fn raw_co_ioctl(BlockDriverState *bs, @@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_raw = { .bdrv_probe_blocksizes = &raw_probe_blocksizes, .bdrv_probe_geometry = &raw_probe_geometry, .bdrv_co_eject = &raw_co_eject, - .bdrv_lock_medium = &raw_lock_medium, + .bdrv_co_lock_medium = &raw_co_lock_medium, .bdrv_co_ioctl = &raw_co_ioctl, .create_opts = &raw_create_opts, .bdrv_has_zero_init = &raw_has_zero_init, -- 2.38.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> bdrv_debug_event() is categorized as an I/O function, and it currently doesn't run in a coroutine. We should let it take a graph rdlock since it traverses the block nodes graph, which however is only possible in a coroutine. Therefore turn it into a co_wrapper_mixed to move the actual function into a coroutine where the lock can be taken. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230113204212.359076-14-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-io.h | 5 ++++- include/block/block_int-common.h | 3 ++- block.c | 6 +++--- block/blkdebug.c | 5 +++-- block/io.c | 22 +++++++++++----------- 5 files changed, 23 insertions(+), 18 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 bdrv_debug_event(BlockDriverState *bs, BlkdebugEvent event); +void coroutine_fn bdrv_co_debug_event(BlockDriverState *bs, + BlkdebugEvent event); +void co_wrapper_mixed 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 (*bdrv_debug_event)(BlockDriverState *bs, BlkdebugEvent event); + void coroutine_fn (*bdrv_co_debug_event)(BlockDriverState *bs, + BlkdebugEvent event); /* io queue for linux-aio */ void coroutine_fn (*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) return drv->bdrv_get_specific_stats(bs); } -void bdrv_debug_event(BlockDriverState *bs, BlkdebugEvent event) +void coroutine_fn bdrv_co_debug_event(BlockDriverState *bs, BlkdebugEvent event) { IO_CODE(); - if (!bs || !bs->drv || !bs->drv->bdrv_debug_event) { + if (!bs || !bs->drv || !bs->drv->bdrv_co_debug_event) { return; } - bs->drv->bdrv_debug_event(bs, event); + bs->drv->bdrv_co_debug_event(bs, event); } static BlockDriverState *bdrv_find_debug_node(BlockDriverState *bs) diff --git a/block/blkdebug.c b/block/blkdebug.c index XXXXXXX..XXXXXXX 100644 --- a/block/blkdebug.c +++ b/block/blkdebug.c @@ -XXX,XX +XXX,XX @@ static void process_rule(BlockDriverState *bs, struct BlkdebugRule *rule, } } -static void blkdebug_debug_event(BlockDriverState *bs, BlkdebugEvent event) +static void coroutine_fn +blkdebug_co_debug_event(BlockDriverState *bs, BlkdebugEvent event) { BDRVBlkdebugState *s = bs->opaque; struct BlkdebugRule *rule, *next; @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_blkdebug = { .bdrv_co_pdiscard = blkdebug_co_pdiscard, .bdrv_co_block_status = blkdebug_co_block_status, - .bdrv_debug_event = blkdebug_debug_event, + .bdrv_co_debug_event = blkdebug_co_debug_event, .bdrv_debug_breakpoint = blkdebug_debug_breakpoint, .bdrv_debug_remove_breakpoint = blkdebug_debug_remove_breakpoint, 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 @@ static int coroutine_fn bdrv_co_do_copy_on_readv(BdrvChild *child, goto err; } - bdrv_debug_event(bs, BLKDBG_COR_WRITE); + bdrv_co_debug_event(bs, BLKDBG_COR_WRITE); if (drv->bdrv_co_pwrite_zeroes && buffer_is_zero(bounce_buffer, pnum)) { /* FIXME: Should we (perhaps conditionally) be setting @@ -XXX,XX +XXX,XX @@ static coroutine_fn int bdrv_padding_rmw_read(BdrvChild *child, qemu_iovec_init_buf(&local_qiov, pad->buf, bytes); if (pad->head) { - bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_HEAD); + bdrv_co_debug_event(bs, BLKDBG_PWRITEV_RMW_HEAD); } if (pad->merge_reads && pad->tail) { - bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_TAIL); + bdrv_co_debug_event(bs, BLKDBG_PWRITEV_RMW_TAIL); } ret = bdrv_aligned_preadv(child, req, req->overlap_offset, bytes, align, &local_qiov, 0, 0); @@ -XXX,XX +XXX,XX @@ static coroutine_fn int bdrv_padding_rmw_read(BdrvChild *child, return ret; } if (pad->head) { - bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_AFTER_HEAD); + bdrv_co_debug_event(bs, BLKDBG_PWRITEV_RMW_AFTER_HEAD); } if (pad->merge_reads && pad->tail) { - bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_AFTER_TAIL); + bdrv_co_debug_event(bs, BLKDBG_PWRITEV_RMW_AFTER_TAIL); } if (pad->merge_reads) { @@ -XXX,XX +XXX,XX @@ static coroutine_fn int bdrv_padding_rmw_read(BdrvChild *child, if (pad->tail) { qemu_iovec_init_buf(&local_qiov, pad->tail_buf, align); - bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_TAIL); + bdrv_co_debug_event(bs, BLKDBG_PWRITEV_RMW_TAIL); ret = bdrv_aligned_preadv( child, req, req->overlap_offset + req->overlap_bytes - align, @@ -XXX,XX +XXX,XX @@ static coroutine_fn int bdrv_padding_rmw_read(BdrvChild *child, if (ret < 0) { return ret; } - bdrv_debug_event(bs, BLKDBG_PWRITEV_RMW_AFTER_TAIL); + bdrv_co_debug_event(bs, BLKDBG_PWRITEV_RMW_AFTER_TAIL); } zero_mem: @@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_pwritev(BdrvChild *child, if (ret < 0) { /* Do nothing, write notifier decided to fail this request */ } else if (flags & BDRV_REQ_ZERO_WRITE) { - bdrv_debug_event(bs, BLKDBG_PWRITEV_ZERO); + bdrv_co_debug_event(bs, BLKDBG_PWRITEV_ZERO); ret = bdrv_co_do_pwrite_zeroes(bs, offset, bytes, flags); } else if (flags & BDRV_REQ_WRITE_COMPRESSED) { ret = bdrv_driver_pwritev_compressed(bs, offset, bytes, qiov, qiov_offset); } else if (bytes <= max_transfer) { - bdrv_debug_event(bs, BLKDBG_PWRITEV); + bdrv_co_debug_event(bs, BLKDBG_PWRITEV); ret = bdrv_driver_pwritev(bs, offset, bytes, qiov, qiov_offset, flags); } else { - bdrv_debug_event(bs, BLKDBG_PWRITEV); + bdrv_co_debug_event(bs, BLKDBG_PWRITEV); while (bytes_remaining) { int num = MIN(bytes_remaining, max_transfer); int local_flags = flags; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_aligned_pwritev(BdrvChild *child, bytes_remaining -= num; } } - bdrv_debug_event(bs, BLKDBG_PWRITEV_DONE); + bdrv_co_debug_event(bs, BLKDBG_PWRITEV_DONE); if (ret >= 0) { ret = 0; -- 2.38.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> Since these functions always run in coroutine context, adjust their name to include "_co_", just like all other BlockDriver callbacks. No functional change intended. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20230113204212.359076-15-kwolf@redhat.com> Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block_int-common.h | 4 ++-- block/io.c | 8 ++++---- block/qcow2.c | 12 ++++++------ 3 files changed, 12 insertions(+), 12 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 { Error **errp); BlockStatsSpecific *(*bdrv_get_specific_stats)(BlockDriverState *bs); - int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_save_vmstate)( + int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_save_vmstate)( BlockDriverState *bs, QEMUIOVector *qiov, int64_t pos); - int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_load_vmstate)( + int coroutine_fn GRAPH_RDLOCK_PTR (*bdrv_co_load_vmstate)( BlockDriverState *bs, QEMUIOVector *qiov, int64_t pos); /* removable device specific */ 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 @@ bdrv_co_readv_vmstate(BlockDriverState *bs, QEMUIOVector *qiov, int64_t pos) bdrv_inc_in_flight(bs); - if (drv->bdrv_load_vmstate) { - ret = drv->bdrv_load_vmstate(bs, qiov, pos); + if (drv->bdrv_co_load_vmstate) { + ret = drv->bdrv_co_load_vmstate(bs, qiov, pos); } else if (child_bs) { ret = bdrv_co_readv_vmstate(child_bs, qiov, pos); } else { @@ -XXX,XX +XXX,XX @@ bdrv_co_writev_vmstate(BlockDriverState *bs, QEMUIOVector *qiov, int64_t pos) bdrv_inc_in_flight(bs); - if (drv->bdrv_save_vmstate) { - ret = drv->bdrv_save_vmstate(bs, qiov, pos); + if (drv->bdrv_co_save_vmstate) { + ret = drv->bdrv_co_save_vmstate(bs, qiov, pos); } else if (child_bs) { ret = bdrv_co_writev_vmstate(child_bs, qiov, pos); } else { 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 int64_t qcow2_check_vmstate_request(BlockDriverState *bs, return pos; } -static coroutine_fn int qcow2_save_vmstate(BlockDriverState *bs, - QEMUIOVector *qiov, int64_t pos) +static coroutine_fn int qcow2_co_save_vmstate(BlockDriverState *bs, + QEMUIOVector *qiov, int64_t pos) { int64_t offset = qcow2_check_vmstate_request(bs, qiov, pos); if (offset < 0) { @@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow2_save_vmstate(BlockDriverState *bs, return bs->drv->bdrv_co_pwritev_part(bs, offset, qiov->size, qiov, 0, 0); } -static coroutine_fn int qcow2_load_vmstate(BlockDriverState *bs, - QEMUIOVector *qiov, int64_t pos) +static coroutine_fn int qcow2_co_load_vmstate(BlockDriverState *bs, + QEMUIOVector *qiov, int64_t pos) { int64_t offset = qcow2_check_vmstate_request(bs, qiov, pos); if (offset < 0) { @@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_qcow2 = { .bdrv_co_get_info = qcow2_co_get_info, .bdrv_get_specific_info = qcow2_get_specific_info, - .bdrv_save_vmstate = qcow2_save_vmstate, - .bdrv_load_vmstate = qcow2_load_vmstate, + .bdrv_co_save_vmstate = qcow2_co_save_vmstate, + .bdrv_co_load_vmstate = qcow2_co_load_vmstate, .is_format = true, .supports_backing = true, -- 2.38.1 From: Philippe Mathieu-Daudé <philmd@linaro.org> The inlined nbd_readXX() functions call beXX_to_cpu(), themselves declared in <qemu/bswap.h>. This fixes when refactoring: In file included from ../../block/nbd.c:44: include/block/nbd.h: In function 'nbd_read16': include/block/nbd.h:383:12: error: implicit declaration of function 'be16_to_cpu' [-Werror=implicit-function-declaration] 383 | *val = be##bits##_to_cpu(*val); \ | ^~ include/block/nbd.h:387:1: note: in expansion of macro 'DEF_NBD_READ_N' 387 | DEF_NBD_READ_N(16) /* Defines nbd_read16(). */ | ^~~~~~~~~~~~~~ Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> Message-Id: <20221125175328.48539-1-philmd@linaro.org> Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/nbd.h | 1 + 1 file changed, 1 insertion(+) diff --git a/include/block/nbd.h b/include/block/nbd.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/nbd.h +++ b/include/block/nbd.h @@ -XXX,XX +XXX,XX @@ #include "io/channel-socket.h" #include "crypto/tlscreds.h" #include "qapi/error.h" +#include "qemu/bswap.h" extern const BlockExportDriver blk_exp_nbd; -- 2.38.1 From: Hanna Reitz <hreitz@redhat.com> When a block driver supports obtaining format-specific information, but that object only contains optional fields, it is possible that none of them are present, so that dump_qobject() (called by bdrv_image_info_specific_dump()) will not print anything. The callers of bdrv_image_info_specific_dump() put a header above this information ("Format specific information:\n"), which will look strange when there is nothing below. Modify bdrv_image_info_specific_dump() to print this header instead of its callers, and only if there is indeed something to be printed. Signed-off-by: Hanna Reitz <hreitz@redhat.com> Message-Id: <20220620162704.80987-2-hreitz@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/qapi.h | 3 ++- block/qapi.c | 41 +++++++++++++++++++++++++++++++++++++---- qemu-io-cmds.c | 4 ++-- 3 files changed, 41 insertions(+), 7 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 @@ void bdrv_query_image_info(BlockDriverState *bs, Error **errp); void bdrv_snapshot_dump(QEMUSnapshotInfo *sn); -void bdrv_image_info_specific_dump(ImageInfoSpecific *info_spec); +void bdrv_image_info_specific_dump(ImageInfoSpecific *info_spec, + const char *prefix); void bdrv_image_info_dump(ImageInfo *info); #endif 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 dump_qdict(int indentation, QDict *dict) } } -void bdrv_image_info_specific_dump(ImageInfoSpecific *info_spec) +/* + * Return whether dumping the given QObject with dump_qobject() would + * yield an empty dump, i.e. not print anything. + */ +static bool qobject_is_empty_dump(const QObject *obj) +{ + switch (qobject_type(obj)) { + case QTYPE_QNUM: + case QTYPE_QSTRING: + case QTYPE_QBOOL: + return false; + + case QTYPE_QDICT: + return qdict_size(qobject_to(QDict, obj)) == 0; + + case QTYPE_QLIST: + return qlist_empty(qobject_to(QList, obj)); + + default: + abort(); + } +} + +/** + * Dumps the given ImageInfoSpecific object in a human-readable form, + * prepending an optional prefix if the dump is not empty. + */ +void bdrv_image_info_specific_dump(ImageInfoSpecific *info_spec, + const char *prefix) { QObject *obj, *data; Visitor *v = qobject_output_visitor_new(&obj); @@ -XXX,XX +XXX,XX @@ void bdrv_image_info_specific_dump(ImageInfoSpecific *info_spec) visit_type_ImageInfoSpecific(v, NULL, &info_spec, &error_abort); visit_complete(v, &obj); data = qdict_get(qobject_to(QDict, obj), "data"); - dump_qobject(1, data); + if (!qobject_is_empty_dump(data)) { + if (prefix) { + qemu_printf("%s", prefix); + } + dump_qobject(1, data); + } qobject_unref(obj); visit_free(v); } @@ -XXX,XX +XXX,XX @@ void bdrv_image_info_dump(ImageInfo *info) } if (info->format_specific) { - qemu_printf("Format specific information:\n"); - bdrv_image_info_specific_dump(info->format_specific); + bdrv_image_info_specific_dump(info->format_specific, + "Format specific information:\n"); } } diff --git a/qemu-io-cmds.c b/qemu-io-cmds.c index XXXXXXX..XXXXXXX 100644 --- a/qemu-io-cmds.c +++ b/qemu-io-cmds.c @@ -XXX,XX +XXX,XX @@ static int info_f(BlockBackend *blk, int argc, char **argv) return -EIO; } if (spec_info) { - printf("Format specific information:\n"); - bdrv_image_info_specific_dump(spec_info); + bdrv_image_info_specific_dump(spec_info, + "Format specific information:\n"); qapi_free_ImageInfoSpecific(spec_info); } -- 2.38.1 From: Hanna Reitz <hreitz@redhat.com> Add some (optional) information that the file driver can provide for image files, namely the extent size hint. Signed-off-by: Hanna Reitz <hreitz@redhat.com> Message-Id: <20220620162704.80987-3-hreitz@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- qapi/block-core.json | 26 ++++++++++++++++++++++++-- block/file-posix.c | 30 ++++++++++++++++++++++++++++++ 2 files changed, 54 insertions(+), 2 deletions(-) diff --git a/qapi/block-core.json b/qapi/block-core.json index XXXXXXX..XXXXXXX 100644 --- a/qapi/block-core.json +++ b/qapi/block-core.json @@ -XXX,XX +XXX,XX @@ '*encryption-format': 'RbdImageEncryptionFormat' } } +## +# @ImageInfoSpecificFile: +# +# @extent-size-hint: Extent size hint (if available) +# +# Since: 8.0 +## +{ 'struct': 'ImageInfoSpecificFile', + 'data': { + '*extent-size-hint': 'size' + } } + ## # @ImageInfoSpecificKind: # # @luks: Since 2.7 # @rbd: Since 6.1 +# @file: Since 8.0 # # Since: 1.7 ## { 'enum': 'ImageInfoSpecificKind', - 'data': [ 'qcow2', 'vmdk', 'luks', 'rbd' ] } + 'data': [ 'qcow2', 'vmdk', 'luks', 'rbd', 'file' ] } ## # @ImageInfoSpecificQCow2Wrapper: @@ -XXX,XX +XXX,XX @@ { 'struct': 'ImageInfoSpecificRbdWrapper', 'data': { 'data': 'ImageInfoSpecificRbd' } } +## +# @ImageInfoSpecificFileWrapper: +# +# Since: 8.0 +## +{ 'struct': 'ImageInfoSpecificFileWrapper', + 'data': { 'data': 'ImageInfoSpecificFile' } } + ## # @ImageInfoSpecific: # @@ -XXX,XX +XXX,XX @@ 'qcow2': 'ImageInfoSpecificQCow2Wrapper', 'vmdk': 'ImageInfoSpecificVmdkWrapper', 'luks': 'ImageInfoSpecificLUKSWrapper', - 'rbd': 'ImageInfoSpecificRbdWrapper' + 'rbd': 'ImageInfoSpecificRbdWrapper', + 'file': 'ImageInfoSpecificFileWrapper' } } ## diff --git a/block/file-posix.c b/block/file-posix.c index XXXXXXX..XXXXXXX 100644 --- a/block/file-posix.c +++ b/block/file-posix.c @@ -XXX,XX +XXX,XX @@ raw_co_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) return 0; } +static ImageInfoSpecific *raw_get_specific_info(BlockDriverState *bs, + Error **errp) +{ + BDRVRawState *s = bs->opaque; + ImageInfoSpecificFile *file_info = g_new0(ImageInfoSpecificFile, 1); + ImageInfoSpecific *spec_info = g_new(ImageInfoSpecific, 1); + + *spec_info = (ImageInfoSpecific){ + .type = IMAGE_INFO_SPECIFIC_KIND_FILE, + .u.file.data = file_info, + }; + +#ifdef FS_IOC_FSGETXATTR + { + struct fsxattr attr; + int ret; + + ret = ioctl(s->fd, FS_IOC_FSGETXATTR, &attr); + if (!ret && attr.fsx_extsize != 0) { + file_info->has_extent_size_hint = true; + file_info->extent_size_hint = attr.fsx_extsize; + } + } +#endif + + return spec_info; +} + static BlockStatsSpecificFile get_blockstats_specific_file(BlockDriverState *bs) { BDRVRawState *s = bs->opaque; @@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_file = { .bdrv_co_truncate = raw_co_truncate, .bdrv_co_getlength = raw_co_getlength, .bdrv_co_get_info = raw_co_get_info, + .bdrv_get_specific_info = raw_get_specific_info, .bdrv_co_get_allocated_file_size = raw_co_get_allocated_file_size, .bdrv_get_specific_stats = raw_get_specific_stats, .bdrv_check_perm = raw_check_perm, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_host_device = { .bdrv_co_truncate = raw_co_truncate, .bdrv_co_getlength = raw_co_getlength, .bdrv_co_get_info = raw_co_get_info, + .bdrv_get_specific_info = raw_get_specific_info, .bdrv_co_get_allocated_file_size = raw_co_get_allocated_file_size, .bdrv_get_specific_stats = hdev_get_specific_stats, .bdrv_check_perm = raw_check_perm, -- 2.38.1 From: Hanna Reitz <hreitz@redhat.com> VMDK's implementation of .bdrv_get_specific_info() returns information about its extent files, ostensibly in the form of ImageInfo objects. However, it does not get this information through bdrv_query_image_info(), but fills only a select few fields with custom information that does not always match the fields' purposes. For example, @format, which is supposed to be a block driver name, is filled with the extent type, e.g. SPARSE or FLAT. In ImageInfo, @compressed shows whether the data that can be seen in the image is stored in compressed form or not. For example, a compressed qcow2 image will store compressed data in its data file, but when accessing the qcow2 node, you will see normal data. This is not how VMDK uses the @compressed field for its extent files: Instead, it signifies whether accessing the extent file will yield compressed data (which the VMDK driver then (de-)compresses). Create a new structure to represent the extent information. This allows us to clarify the fields' meanings, and it clearly shows that these are not complete ImageInfo objects. (That is, if a user wants an extent file's ImageInfo object, they will need to query it separately, and will not get it from ImageInfoSpecificVmdk.extents.) Note that this removes the last use of ['ImageInfo'] (i.e. an array of ImageInfo objects), so the QAPI generator will no longer generate ImageInfoList by default. However, we use it in qemu-img.c, so we need to create a dummy object to force the generate to create that type, similarly to DummyForceArrays in machine.json (introduced in commit 9f08c8ec73878122ad4b061ed334f0437afaaa32 ("qapi: Lazy creation of array types")). Signed-off-by: Hanna Reitz <hreitz@redhat.com> Message-Id: <20220620162704.80987-4-hreitz@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- qapi/block-core.json | 38 +++++++++++++++++++++++++++++++++++++- block/vmdk.c | 8 ++++---- 2 files changed, 41 insertions(+), 5 deletions(-) diff --git a/qapi/block-core.json b/qapi/block-core.json index XXXXXXX..XXXXXXX 100644 --- a/qapi/block-core.json +++ b/qapi/block-core.json @@ -XXX,XX +XXX,XX @@ 'create-type': 'str', 'cid': 'int', 'parent-cid': 'int', - 'extents': ['ImageInfo'] + 'extents': ['VmdkExtentInfo'] + } } + +## +# @VmdkExtentInfo: +# +# Information about a VMDK extent file +# +# @filename: Name of the extent file +# +# @format: Extent type (e.g. FLAT or SPARSE) +# +# @virtual-size: Number of bytes covered by this extent +# +# @cluster-size: Cluster size in bytes (for non-flat extents) +# +# @compressed: Whether this extent contains compressed data +# +# Since: 8.0 +## +{ 'struct': 'VmdkExtentInfo', + 'data': { + 'filename': 'str', + 'format': 'str', + 'virtual-size': 'int', + '*cluster-size': 'int', + '*compressed': 'bool' } } ## @@ -XXX,XX +XXX,XX @@ 'data': { 'device': 'str', '*id': 'str', '*name': 'str'}, 'returns': 'SnapshotInfo', 'allow-preconfig': true } + +## +# @DummyBlockCoreForceArrays: +# +# Not used by QMP; hack to let us use ImageInfoList internally +# +# Since: 8.0 +## +{ 'struct': 'DummyBlockCoreForceArrays', + 'data': { 'unused-image-info': ['ImageInfo'] } } 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 int vmdk_has_zero_init(BlockDriverState *bs) return 1; } -static ImageInfo *vmdk_get_extent_info(VmdkExtent *extent) +static VmdkExtentInfo *vmdk_get_extent_info(VmdkExtent *extent) { - ImageInfo *info = g_new0(ImageInfo, 1); + VmdkExtentInfo *info = g_new0(VmdkExtentInfo, 1); bdrv_refresh_filename(extent->file->bs); - *info = (ImageInfo){ + *info = (VmdkExtentInfo){ .filename = g_strdup(extent->file->bs->filename), .format = g_strdup(extent->type), .virtual_size = extent->sectors * BDRV_SECTOR_SIZE, @@ -XXX,XX +XXX,XX @@ static ImageInfoSpecific *vmdk_get_specific_info(BlockDriverState *bs, int i; BDRVVmdkState *s = bs->opaque; ImageInfoSpecific *spec_info = g_new0(ImageInfoSpecific, 1); - ImageInfoList **tail; + VmdkExtentInfoList **tail; *spec_info = (ImageInfoSpecific){ .type = IMAGE_INFO_SPECIFIC_KIND_VMDK, -- 2.38.1 From: Hanna Reitz <hreitz@redhat.com> ImageInfo sometimes contains flat information, and sometimes it does not. Split off a BlockNodeInfo struct, which only contains information about a single node and has no link to the backing image. We do this so we can extend BlockNodeInfo to a BlockGraphInfo struct, which has links to all child nodes, not just the backing node. It would be strange to base BlockGraphInfo on ImageInfo, because then this extended struct would have two links to the backing node (one in BlockGraphInfo as one of all the child links, and one in ImageInfo). Furthermore, it is quite common to ignore the backing-image field altogether: bdrv_query_image_info() does not set it, and bdrv_image_info_dump() does not evaluate it. That signals that we should have different structs for describing a single node and one that has a link to the backing image. Still, bdrv_query_image_info() and bdrv_image_info_dump() are not changed too much in this patch. Follow-up patches will handle them. Signed-off-by: Hanna Reitz <hreitz@redhat.com> Message-Id: <20220620162704.80987-5-hreitz@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- qapi/block-core.json | 24 +++++++++---- include/block/qapi.h | 3 ++ block/qapi.c | 86 ++++++++++++++++++++++++++++++++------------ 3 files changed, 85 insertions(+), 28 deletions(-) diff --git a/qapi/block-core.json b/qapi/block-core.json index XXXXXXX..XXXXXXX 100644 --- a/qapi/block-core.json +++ b/qapi/block-core.json @@ -XXX,XX +XXX,XX @@ } } ## -# @ImageInfo: +# @BlockNodeInfo: # # Information about a QEMU image file # @@ -XXX,XX +XXX,XX @@ # # @snapshots: list of VM snapshots # -# @backing-image: info of the backing image (since 1.6) -# # @format-specific: structure supplying additional format-specific # information (since 1.7) # -# Since: 1.3 +# Since: 8.0 ## -{ 'struct': 'ImageInfo', +{ 'struct': 'BlockNodeInfo', 'data': {'filename': 'str', 'format': 'str', '*dirty-flag': 'bool', '*actual-size': 'int', 'virtual-size': 'int', '*cluster-size': 'int', '*encrypted': 'bool', '*compressed': 'bool', '*backing-filename': 'str', '*full-backing-filename': 'str', '*backing-filename-format': 'str', '*snapshots': ['SnapshotInfo'], - '*backing-image': 'ImageInfo', '*format-specific': 'ImageInfoSpecific' } } +## +# @ImageInfo: +# +# Information about a QEMU image file, and potentially its backing image +# +# @backing-image: info of the backing image +# +# Since: 1.3 +## +{ 'struct': 'ImageInfo', + 'base': 'BlockNodeInfo', + 'data': { + '*backing-image': 'ImageInfo' + } } + ## # @ImageCheck: # 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 @@ BlockDeviceInfo *bdrv_block_device_info(BlockBackend *blk, int bdrv_query_snapshot_info_list(BlockDriverState *bs, SnapshotInfoList **p_list, Error **errp); +void bdrv_query_block_node_info(BlockDriverState *bs, + BlockNodeInfo **p_info, + Error **errp); void bdrv_query_image_info(BlockDriverState *bs, ImageInfo **p_info, Error **errp); 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 @@ int bdrv_query_snapshot_info_list(BlockDriverState *bs, } /** - * bdrv_query_image_info: - * @bs: block device to examine - * @p_info: location to store image information - * @errp: location to store error information - * - * Store "flat" image information in @p_info. - * - * "Flat" means it does *not* query backing image information, - * i.e. (*pinfo)->has_backing_image will be set to false and - * (*pinfo)->backing_image to NULL even when the image does in fact have - * a backing image. - * - * @p_info will be set only on success. On error, store error in @errp. + * Helper function for other query info functions. Store information about @bs + * in @info, setting @errp on error. */ -void bdrv_query_image_info(BlockDriverState *bs, - ImageInfo **p_info, - Error **errp) +static void bdrv_do_query_node_info(BlockDriverState *bs, + BlockNodeInfo *info, + Error **errp) { int64_t size; const char *backing_filename; BlockDriverInfo bdi; int ret; Error *err = NULL; - ImageInfo *info; aio_context_acquire(bdrv_get_aio_context(bs)); @@ -XXX,XX +XXX,XX @@ void bdrv_query_image_info(BlockDriverState *bs, bdrv_refresh_filename(bs); - info = g_new0(ImageInfo, 1); info->filename = g_strdup(bs->filename); info->format = g_strdup(bdrv_get_format_name(bs)); info->virtual_size = size; @@ -XXX,XX +XXX,XX @@ void bdrv_query_image_info(BlockDriverState *bs, info->format_specific = bdrv_get_specific_info(bs, &err); if (err) { error_propagate(errp, err); - qapi_free_ImageInfo(info); goto out; } backing_filename = bs->backing_file; @@ -XXX,XX +XXX,XX @@ void bdrv_query_image_info(BlockDriverState *bs, break; default: error_propagate(errp, err); - qapi_free_ImageInfo(info); goto out; } - *p_info = info; - out: aio_context_release(bdrv_get_aio_context(bs)); } +/** + * bdrv_query_block_node_info: + * @bs: block node to examine + * @p_info: location to store node information + * @errp: location to store error information + * + * Store image information about @bs in @p_info. + * + * @p_info will be set only on success. On error, store error in @errp. + */ +void bdrv_query_block_node_info(BlockDriverState *bs, + BlockNodeInfo **p_info, + Error **errp) +{ + BlockNodeInfo *info; + ERRP_GUARD(); + + info = g_new0(BlockNodeInfo, 1); + bdrv_do_query_node_info(bs, info, errp); + if (*errp) { + qapi_free_BlockNodeInfo(info); + return; + } + + *p_info = info; +} + +/** + * bdrv_query_image_info: + * @bs: block node to examine + * @p_info: location to store image information + * @errp: location to store error information + * + * Store "flat" image information in @p_info. + * + * "Flat" means it does *not* query backing image information, + * i.e. (*pinfo)->has_backing_image will be set to false and + * (*pinfo)->backing_image to NULL even when the image does in fact have + * a backing image. + * + * @p_info will be set only on success. On error, store error in @errp. + */ +void bdrv_query_image_info(BlockDriverState *bs, + ImageInfo **p_info, + Error **errp) +{ + ImageInfo *info; + ERRP_GUARD(); + + info = g_new0(ImageInfo, 1); + bdrv_do_query_node_info(bs, qapi_ImageInfo_base(info), errp); + if (*errp) { + qapi_free_ImageInfo(info); + return; + } + + *p_info = info; +} + /* @p_info will be set only on success. */ static void bdrv_query_info(BlockBackend *blk, BlockInfo **p_info, Error **errp) -- 2.38.1 From: Hanna Reitz <hreitz@redhat.com> qemu-img info never uses ImageInfo's backing-image field, because it opens the backing chain one by one with BDRV_O_NO_BACKING, and prints all backing chain nodes' information consecutively. Use BlockNodeInfo to make it clear that we only print information about a single node, and that we are not using the backing-image field. Notably, bdrv_image_info_dump() does not evaluate the backing-image field, so we can easily make it take a BlockNodeInfo pointer (and consequentially rename it to bdrv_node_info_dump()). It makes more sense this way, because again, the interface now makes it syntactically clear that backing-image is ignored by this function. Signed-off-by: Hanna Reitz <hreitz@redhat.com> Message-Id: <20220620162704.80987-6-hreitz@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- qapi/block-core.json | 4 +-- include/block/qapi.h | 2 +- block/monitor/block-hmp-cmds.c | 2 +- block/qapi.c | 2 +- qemu-img.c | 48 +++++++++++++++++----------------- 5 files changed, 29 insertions(+), 29 deletions(-) diff --git a/qapi/block-core.json b/qapi/block-core.json index XXXXXXX..XXXXXXX 100644 --- a/qapi/block-core.json +++ b/qapi/block-core.json @@ -XXX,XX +XXX,XX @@ ## # @DummyBlockCoreForceArrays: # -# Not used by QMP; hack to let us use ImageInfoList internally +# Not used by QMP; hack to let us use BlockNodeInfoList internally # # Since: 8.0 ## { 'struct': 'DummyBlockCoreForceArrays', - 'data': { 'unused-image-info': ['ImageInfo'] } } + 'data': { 'unused-block-node-info': ['BlockNodeInfo'] } } 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 @@ void bdrv_query_image_info(BlockDriverState *bs, void bdrv_snapshot_dump(QEMUSnapshotInfo *sn); void bdrv_image_info_specific_dump(ImageInfoSpecific *info_spec, const char *prefix); -void bdrv_image_info_dump(ImageInfo *info); +void bdrv_node_info_dump(BlockNodeInfo *info); #endif diff --git a/block/monitor/block-hmp-cmds.c b/block/monitor/block-hmp-cmds.c index XXXXXXX..XXXXXXX 100644 --- a/block/monitor/block-hmp-cmds.c +++ b/block/monitor/block-hmp-cmds.c @@ -XXX,XX +XXX,XX @@ static void print_block_info(Monitor *mon, BlockInfo *info, monitor_printf(mon, "\nImages:\n"); image_info = inserted->image; while (1) { - bdrv_image_info_dump(image_info); + bdrv_node_info_dump(qapi_ImageInfo_base(image_info)); if (image_info->backing_image) { image_info = image_info->backing_image; } else { 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 @@ void bdrv_image_info_specific_dump(ImageInfoSpecific *info_spec, visit_free(v); } -void bdrv_image_info_dump(ImageInfo *info) +void bdrv_node_info_dump(BlockNodeInfo *info) { char *size_buf, *dsize_buf; if (!info->has_actual_size) { 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 void dump_snapshots(BlockDriverState *bs) g_free(sn_tab); } -static void dump_json_image_info_list(ImageInfoList *list) +static void dump_json_block_node_info_list(BlockNodeInfoList *list) { GString *str; QObject *obj; Visitor *v = qobject_output_visitor_new(&obj); - visit_type_ImageInfoList(v, NULL, &list, &error_abort); + visit_type_BlockNodeInfoList(v, NULL, &list, &error_abort); visit_complete(v, &obj); str = qobject_to_json_pretty(obj, true); assert(str != NULL); @@ -XXX,XX +XXX,XX @@ static void dump_json_image_info_list(ImageInfoList *list) g_string_free(str, true); } -static void dump_json_image_info(ImageInfo *info) +static void dump_json_block_node_info(BlockNodeInfo *info) { GString *str; QObject *obj; Visitor *v = qobject_output_visitor_new(&obj); - visit_type_ImageInfo(v, NULL, &info, &error_abort); + visit_type_BlockNodeInfo(v, NULL, &info, &error_abort); visit_complete(v, &obj); str = qobject_to_json_pretty(obj, true); assert(str != NULL); @@ -XXX,XX +XXX,XX @@ static void dump_json_image_info(ImageInfo *info) g_string_free(str, true); } -static void dump_human_image_info_list(ImageInfoList *list) +static void dump_human_image_info_list(BlockNodeInfoList *list) { - ImageInfoList *elem; + BlockNodeInfoList *elem; bool delim = false; for (elem = list; elem; elem = elem->next) { @@ -XXX,XX +XXX,XX @@ static void dump_human_image_info_list(ImageInfoList *list) } delim = true; - bdrv_image_info_dump(elem->value); + bdrv_node_info_dump(elem->value); } } @@ -XXX,XX +XXX,XX @@ static gboolean str_equal_func(gconstpointer a, gconstpointer b) } /** - * Open an image file chain and return an ImageInfoList + * Open an image file chain and return an BlockNodeInfoList * * @filename: topmost image filename * @fmt: topmost image format (may be NULL to autodetect) * @chain: true - enumerate entire backing file chain * false - only topmost image file * - * Returns a list of ImageInfo objects or NULL if there was an error opening an - * image file. If there was an error a message will have been printed to - * stderr. + * Returns a list of BlockNodeInfo objects or NULL if there was an error + * opening an image file. If there was an error a message will have been + * printed to stderr. */ -static ImageInfoList *collect_image_info_list(bool image_opts, - const char *filename, - const char *fmt, - bool chain, bool force_share) +static BlockNodeInfoList *collect_image_info_list(bool image_opts, + const char *filename, + const char *fmt, + bool chain, bool force_share) { - ImageInfoList *head = NULL; - ImageInfoList **tail = &head; + BlockNodeInfoList *head = NULL; + BlockNodeInfoList **tail = &head; GHashTable *filenames; Error *err = NULL; @@ -XXX,XX +XXX,XX @@ static ImageInfoList *collect_image_info_list(bool image_opts, while (filename) { BlockBackend *blk; BlockDriverState *bs; - ImageInfo *info; + BlockNodeInfo *info; if (g_hash_table_lookup_extended(filenames, filename, NULL, NULL)) { error_report("Backing file '%s' creates an infinite loop.", @@ -XXX,XX +XXX,XX @@ static ImageInfoList *collect_image_info_list(bool image_opts, } bs = blk_bs(blk); - bdrv_query_image_info(bs, &info, &err); + bdrv_query_block_node_info(bs, &info, &err); if (err) { error_report_err(err); blk_unref(blk); @@ -XXX,XX +XXX,XX @@ static ImageInfoList *collect_image_info_list(bool image_opts, return head; err: - qapi_free_ImageInfoList(head); + qapi_free_BlockNodeInfoList(head); g_hash_table_destroy(filenames); return NULL; } @@ -XXX,XX +XXX,XX @@ static int img_info(int argc, char **argv) OutputFormat output_format = OFORMAT_HUMAN; bool chain = false; const char *filename, *fmt, *output; - ImageInfoList *list; + BlockNodeInfoList *list; bool image_opts = false; bool force_share = false; @@ -XXX,XX +XXX,XX @@ static int img_info(int argc, char **argv) break; case OFORMAT_JSON: if (chain) { - dump_json_image_info_list(list); + dump_json_block_node_info_list(list); } else { - dump_json_image_info(list->value); + dump_json_block_node_info(list->value); } break; } - qapi_free_ImageInfoList(list); + qapi_free_BlockNodeInfoList(list); return 0; } -- 2.38.1 From: Hanna Reitz <hreitz@redhat.com> There is no real reason why bdrv_query_image_info() should generally not recurse. The ImageInfo struct has a pointer to the backing image, so it should generally be filled, unless the caller explicitly opts out. This moves the recursing code from bdrv_block_device_info() into bdrv_query_image_info(). Signed-off-by: Hanna Reitz <hreitz@redhat.com> Message-Id: <20220620162704.80987-7-hreitz@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/qapi.h | 2 + block/qapi.c | 92 +++++++++++++++++++++++++++----------------- 2 files changed, 58 insertions(+), 36 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 @@ void bdrv_query_block_node_info(BlockDriverState *bs, Error **errp); void bdrv_query_image_info(BlockDriverState *bs, ImageInfo **p_info, + bool flat, + bool skip_implicit_filters, Error **errp); void bdrv_snapshot_dump(QEMUSnapshotInfo *sn); 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 @@ BlockDeviceInfo *bdrv_block_device_info(BlockBackend *blk, Error **errp) { ImageInfo **p_image_info; + ImageInfo *backing_info; BlockDriverState *bs0, *backing; BlockDeviceInfo *info; + ERRP_GUARD(); if (!bs->drv) { error_setg(errp, "Block device %s is ejected", bs->node_name); @@ -XXX,XX +XXX,XX @@ BlockDeviceInfo *bdrv_block_device_info(BlockBackend *blk, bs0 = bs; p_image_info = &info->image; info->backing_file_depth = 0; - while (1) { - Error *local_err = NULL; - bdrv_query_image_info(bs0, p_image_info, &local_err); - if (local_err) { - error_propagate(errp, local_err); - qapi_free_BlockDeviceInfo(info); - return NULL; - } - - /* stop gathering data for flat output */ - if (flat) { - break; - } - if (bs0->drv && bdrv_filter_or_cow_child(bs0)) { - /* - * Put any filtered child here (for backwards compatibility to when - * we put bs0->backing here, which might be any filtered child). - */ - info->backing_file_depth++; - bs0 = bdrv_filter_or_cow_bs(bs0); - p_image_info = &((*p_image_info)->backing_image); - } else { - break; - } + /* + * Skip automatically inserted nodes that the user isn't aware of for + * query-block (blk != NULL), but not for query-named-block-nodes + */ + bdrv_query_image_info(bs0, p_image_info, flat, blk != NULL, errp); + if (*errp) { + qapi_free_BlockDeviceInfo(info); + return NULL; + } - /* Skip automatically inserted nodes that the user isn't aware of for - * query-block (blk != NULL), but not for query-named-block-nodes */ - if (blk) { - bs0 = bdrv_skip_implicit_filters(bs0); - } + backing_info = info->image->backing_image; + while (backing_info) { + info->backing_file_depth++; + backing_info = backing_info->backing_image; } return info; @@ -XXX,XX +XXX,XX @@ void bdrv_query_block_node_info(BlockDriverState *bs, * bdrv_query_image_info: * @bs: block node to examine * @p_info: location to store image information + * @flat: skip backing node information + * @skip_implicit_filters: skip implicit filters in the backing chain * @errp: location to store error information * - * Store "flat" image information in @p_info. + * Store image information in @p_info, potentially recursively covering the + * backing chain. * - * "Flat" means it does *not* query backing image information, - * i.e. (*pinfo)->has_backing_image will be set to false and - * (*pinfo)->backing_image to NULL even when the image does in fact have - * a backing image. + * If @flat is true, do not query backing image information, i.e. + * (*p_info)->has_backing_image will be set to false and + * (*p_info)->backing_image to NULL even when the image does in fact have a + * backing image. + * + * If @skip_implicit_filters is true, implicit filter nodes in the backing chain + * will be skipped when querying backing image information. + * (@skip_implicit_filters is ignored when @flat is true.) * * @p_info will be set only on success. On error, store error in @errp. */ void bdrv_query_image_info(BlockDriverState *bs, ImageInfo **p_info, + bool flat, + bool skip_implicit_filters, Error **errp) { ImageInfo *info; @@ -XXX,XX +XXX,XX @@ void bdrv_query_image_info(BlockDriverState *bs, info = g_new0(ImageInfo, 1); bdrv_do_query_node_info(bs, qapi_ImageInfo_base(info), errp); if (*errp) { - qapi_free_ImageInfo(info); - return; + goto fail; + } + + if (!flat) { + BlockDriverState *backing; + + /* + * Use any filtered child here (for backwards compatibility to when + * we always took bs->backing, which might be any filtered child). + */ + backing = bdrv_filter_or_cow_bs(bs); + if (skip_implicit_filters) { + backing = bdrv_skip_implicit_filters(backing); + } + + if (backing) { + bdrv_query_image_info(backing, &info->backing_image, false, + skip_implicit_filters, errp); + if (*errp) { + goto fail; + } + } } *p_info = info; + return; + +fail: + assert(*errp); + qapi_free_ImageInfo(info); } /* @p_info will be set only on success. */ -- 2.38.1 From: Hanna Reitz <hreitz@redhat.com> Introduce a new QAPI type BlockGraphInfo and an associated bdrv_query_block_graph_info() function that recursively gathers BlockNodeInfo objects through a block graph. A follow-up patch is going to make "qemu-img info" use this to print information about all nodes that are (usually implicitly) opened for a given image file. Signed-off-by: Hanna Reitz <hreitz@redhat.com> Message-Id: <20220620162704.80987-8-hreitz@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- qapi/block-core.json | 35 ++++++++++++++++++++++++++++++++ include/block/qapi.h | 3 +++ block/qapi.c | 48 ++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 86 insertions(+) diff --git a/qapi/block-core.json b/qapi/block-core.json index XXXXXXX..XXXXXXX 100644 --- a/qapi/block-core.json +++ b/qapi/block-core.json @@ -XXX,XX +XXX,XX @@ '*backing-image': 'ImageInfo' } } +## +# @BlockChildInfo: +# +# Information about all nodes in the block graph starting at some node, +# annotated with information about that node in relation to its parent. +# +# @name: Child name of the root node in the BlockGraphInfo struct, in its role +# as the child of some undescribed parent node +# +# @info: Block graph information starting at this node +# +# Since: 8.0 +## +{ 'struct': 'BlockChildInfo', + 'data': { + 'name': 'str', + 'info': 'BlockGraphInfo' + } } + +## +# @BlockGraphInfo: +# +# Information about all nodes in a block (sub)graph in the form of BlockNodeInfo +# data. +# The base BlockNodeInfo struct contains the information for the (sub)graph's +# root node. +# +# @children: Array of links to this node's child nodes' information +# +# Since: 8.0 +## +{ 'struct': 'BlockGraphInfo', + 'base': 'BlockNodeInfo', + 'data': { 'children': ['BlockChildInfo'] } } + ## # @ImageCheck: # 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 @@ 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 bdrv_snapshot_dump(QEMUSnapshotInfo *sn); void bdrv_image_info_specific_dump(ImageInfoSpecific *info_spec, 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 @@ fail: qapi_free_ImageInfo(info); } +/** + * bdrv_query_block_graph_info: + * @bs: root node to start from + * @p_info: location to store image information + * @errp: location to store error information + * + * Store image information about the graph starting from @bs in @p_info. + * + * @p_info will be set only on success. On error, store error in @errp. + */ +void bdrv_query_block_graph_info(BlockDriverState *bs, + BlockGraphInfo **p_info, + Error **errp) +{ + BlockGraphInfo *info; + BlockChildInfoList **children_list_tail; + BdrvChild *c; + ERRP_GUARD(); + + info = g_new0(BlockGraphInfo, 1); + bdrv_do_query_node_info(bs, qapi_BlockGraphInfo_base(info), errp); + if (*errp) { + goto fail; + } + + children_list_tail = &info->children; + + QLIST_FOREACH(c, &bs->children, next) { + BlockChildInfo *c_info; + + c_info = g_new0(BlockChildInfo, 1); + QAPI_LIST_APPEND(children_list_tail, c_info); + + c_info->name = g_strdup(c->name); + bdrv_query_block_graph_info(c->bs, &c_info->info, errp); + if (*errp) { + goto fail; + } + } + + *p_info = info; + return; + +fail: + assert(*errp != NULL); + qapi_free_BlockGraphInfo(info); +} + /* @p_info will be set only on success. */ static void bdrv_query_info(BlockBackend *blk, BlockInfo **p_info, Error **errp) -- 2.38.1 From: Hanna Reitz <hreitz@redhat.com> In order to let qemu-img info present a block graph, add a parameter to bdrv_node_info_dump() and bdrv_image_info_specific_dump() so that the information of nodes below the root level can be given an indentation. Signed-off-by: Hanna Reitz <hreitz@redhat.com> Message-Id: <20220620162704.80987-9-hreitz@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/qapi.h | 5 ++-- block/monitor/block-hmp-cmds.c | 2 +- block/qapi.c | 47 +++++++++++++++++++--------------- qemu-img.c | 2 +- qemu-io-cmds.c | 3 ++- 5 files changed, 34 insertions(+), 25 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 @@ void bdrv_query_block_graph_info(BlockDriverState *bs, void bdrv_snapshot_dump(QEMUSnapshotInfo *sn); void bdrv_image_info_specific_dump(ImageInfoSpecific *info_spec, - const char *prefix); -void bdrv_node_info_dump(BlockNodeInfo *info); + const char *prefix, + int indentation); +void bdrv_node_info_dump(BlockNodeInfo *info, int indentation); #endif diff --git a/block/monitor/block-hmp-cmds.c b/block/monitor/block-hmp-cmds.c index XXXXXXX..XXXXXXX 100644 --- a/block/monitor/block-hmp-cmds.c +++ b/block/monitor/block-hmp-cmds.c @@ -XXX,XX +XXX,XX @@ static void print_block_info(Monitor *mon, BlockInfo *info, monitor_printf(mon, "\nImages:\n"); image_info = inserted->image; while (1) { - bdrv_node_info_dump(qapi_ImageInfo_base(image_info)); + bdrv_node_info_dump(qapi_ImageInfo_base(image_info), 0); if (image_info->backing_image) { image_info = image_info->backing_image; } else { 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 bool qobject_is_empty_dump(const QObject *obj) * prepending an optional prefix if the dump is not empty. */ void bdrv_image_info_specific_dump(ImageInfoSpecific *info_spec, - const char *prefix) + const char *prefix, + int indentation) { QObject *obj, *data; Visitor *v = qobject_output_visitor_new(&obj); @@ -XXX,XX +XXX,XX @@ void bdrv_image_info_specific_dump(ImageInfoSpecific *info_spec, data = qdict_get(qobject_to(QDict, obj), "data"); if (!qobject_is_empty_dump(data)) { if (prefix) { - qemu_printf("%s", prefix); + qemu_printf("%*s%s", indentation * 4, "", prefix); } - dump_qobject(1, data); + dump_qobject(indentation + 1, data); } qobject_unref(obj); visit_free(v); } -void bdrv_node_info_dump(BlockNodeInfo *info) +void bdrv_node_info_dump(BlockNodeInfo *info, int indentation) { char *size_buf, *dsize_buf; + g_autofree char *ind_s = g_strdup_printf("%*s", indentation * 4, ""); + if (!info->has_actual_size) { dsize_buf = g_strdup("unavailable"); } else { dsize_buf = size_to_str(info->actual_size); } size_buf = size_to_str(info->virtual_size); - qemu_printf("image: %s\n" - "file format: %s\n" - "virtual size: %s (%" PRId64 " bytes)\n" - "disk size: %s\n", - info->filename, info->format, size_buf, - info->virtual_size, - dsize_buf); + qemu_printf("%simage: %s\n" + "%sfile format: %s\n" + "%svirtual size: %s (%" PRId64 " bytes)\n" + "%sdisk size: %s\n", + ind_s, info->filename, + ind_s, info->format, + ind_s, size_buf, info->virtual_size, + ind_s, dsize_buf); g_free(size_buf); g_free(dsize_buf); if (info->has_encrypted && info->encrypted) { - qemu_printf("encrypted: yes\n"); + qemu_printf("%sencrypted: yes\n", ind_s); } if (info->has_cluster_size) { - qemu_printf("cluster_size: %" PRId64 "\n", - info->cluster_size); + qemu_printf("%scluster_size: %" PRId64 "\n", + ind_s, info->cluster_size); } if (info->has_dirty_flag && info->dirty_flag) { - qemu_printf("cleanly shut down: no\n"); + qemu_printf("%scleanly shut down: no\n", ind_s); } if (info->backing_filename) { - qemu_printf("backing file: %s", info->backing_filename); + qemu_printf("%sbacking file: %s", ind_s, info->backing_filename); if (!info->full_backing_filename) { qemu_printf(" (cannot determine actual path)"); } else if (strcmp(info->backing_filename, @@ -XXX,XX +XXX,XX @@ void bdrv_node_info_dump(BlockNodeInfo *info) } qemu_printf("\n"); if (info->backing_filename_format) { - qemu_printf("backing file format: %s\n", - info->backing_filename_format); + qemu_printf("%sbacking file format: %s\n", + ind_s, info->backing_filename_format); } } if (info->has_snapshots) { SnapshotInfoList *elem; - qemu_printf("Snapshot list:\n"); + qemu_printf("%sSnapshot list:\n", ind_s); + qemu_printf("%s", ind_s); bdrv_snapshot_dump(NULL); qemu_printf("\n"); @@ -XXX,XX +XXX,XX @@ void bdrv_node_info_dump(BlockNodeInfo *info) pstrcpy(sn.id_str, sizeof(sn.id_str), elem->value->id); pstrcpy(sn.name, sizeof(sn.name), elem->value->name); + qemu_printf("%s", ind_s); bdrv_snapshot_dump(&sn); qemu_printf("\n"); } @@ -XXX,XX +XXX,XX @@ void bdrv_node_info_dump(BlockNodeInfo *info) if (info->format_specific) { bdrv_image_info_specific_dump(info->format_specific, - "Format specific information:\n"); + "Format specific information:\n", + indentation); } } 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 void dump_human_image_info_list(BlockNodeInfoList *list) } delim = true; - bdrv_node_info_dump(elem->value); + bdrv_node_info_dump(elem->value, 0); } } diff --git a/qemu-io-cmds.c b/qemu-io-cmds.c index XXXXXXX..XXXXXXX 100644 --- a/qemu-io-cmds.c +++ b/qemu-io-cmds.c @@ -XXX,XX +XXX,XX @@ static int info_f(BlockBackend *blk, int argc, char **argv) } if (spec_info) { bdrv_image_info_specific_dump(spec_info, - "Format specific information:\n"); + "Format specific information:\n", + 0); qapi_free_ImageInfoSpecific(spec_info); } -- 2.38.1 From: Hanna Reitz <hreitz@redhat.com> Before we let qemu-img info print child node information, have common.filter, common.rc, and iotests.py filter it from the test output so we get as few reference output changes as possible. Signed-off-by: Hanna Reitz <hreitz@redhat.com> Message-Id: <20220620162704.80987-10-hreitz@redhat.com> Tested-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- tests/qemu-iotests/iotests.py | 18 +++++++++++++++--- tests/qemu-iotests/common.filter | 22 ++++++++++++++-------- tests/qemu-iotests/common.rc | 22 ++++++++++++++-------- 3 files changed, 43 insertions(+), 19 deletions(-) diff --git a/tests/qemu-iotests/iotests.py b/tests/qemu-iotests/iotests.py index XXXXXXX..XXXXXXX 100644 --- a/tests/qemu-iotests/iotests.py +++ b/tests/qemu-iotests/iotests.py @@ -XXX,XX +XXX,XX @@ def qemu_img_log(*args: str, check: bool = True def img_info_log(filename: str, filter_path: Optional[str] = None, use_image_opts: bool = False, extra_args: Sequence[str] = (), - check: bool = True, + check: bool = True, drop_child_info: bool = True, ) -> None: args = ['info'] if use_image_opts: @@ -XXX,XX +XXX,XX @@ def img_info_log(filename: str, filter_path: Optional[str] = None, output = qemu_img(*args, check=check).stdout if not filter_path: filter_path = filename - log(filter_img_info(output, filter_path)) + log(filter_img_info(output, filter_path, drop_child_info)) def qemu_io_wrap_args(args: Sequence[str]) -> List[str]: if '-f' in args or '--image-opts' in args: @@ -XXX,XX +XXX,XX @@ def _filter(_key, value): def filter_generated_node_ids(msg): return re.sub("#block[0-9]+", "NODE_NAME", msg) -def filter_img_info(output, filename): +def filter_img_info(output: str, filename: str, + drop_child_info: bool = True) -> str: lines = [] + drop_indented = False for line in output.split('\n'): if 'disk size' in line or 'actual-size' in line: continue + + # Drop child node info + if drop_indented: + if line.startswith(' '): + continue + drop_indented = False + if drop_child_info and "Child node '/" in line: + drop_indented = True + continue + line = line.replace(filename, 'TEST_IMG') line = filter_testfiles(line) line = line.replace(imgfmt, 'IMGFMT') diff --git a/tests/qemu-iotests/common.filter b/tests/qemu-iotests/common.filter index XXXXXXX..XXXXXXX 100644 --- a/tests/qemu-iotests/common.filter +++ b/tests/qemu-iotests/common.filter @@ -XXX,XX +XXX,XX @@ _filter_img_info() discard=0 regex_json_spec_start='^ *"format-specific": \{' + regex_json_child_start='^ *"children": \[' gsed -e "s#$REMOTE_TEST_DIR#TEST_DIR#g" \ -e "s#$IMGPROTO:$TEST_DIR#TEST_DIR#g" \ -e "s#$TEST_DIR#TEST_DIR#g" \ @@ -XXX,XX +XXX,XX @@ _filter_img_info() -e 's/\(compression type: \)\(zlib\|zstd\)/\1COMPRESSION_TYPE/' \ -e "s/uuid: [-a-f0-9]\\+/uuid: 00000000-0000-0000-0000-000000000000/" | \ while IFS='' read -r line; do - if [[ $format_specific == 1 ]]; then - discard=0 - elif [[ $line == "Format specific information:" ]]; then - discard=1 - elif [[ $line =~ $regex_json_spec_start ]]; then - discard=2 - regex_json_spec_end="^${line%%[^ ]*}\\},? *$" + if [[ $discard == 0 ]]; then + if [[ $format_specific == 0 && $line == "Format specific information:" ]]; then + discard=1 + elif [[ $line =~ "Child node '/" ]]; then + discard=1 + elif [[ $line =~ $regex_json_spec_start ]]; then + discard=2 + regex_json_end="^${line%%[^ ]*}\\},? *$" + elif [[ $line =~ $regex_json_child_start ]]; then + discard=2 + regex_json_end="^${line%%[^ ]*}\\],? *$" + fi fi if [[ $discard == 0 ]]; then echo "$line" elif [[ $discard == 1 && ! $line ]]; then echo discard=0 - elif [[ $discard == 2 && $line =~ $regex_json_spec_end ]]; then + elif [[ $discard == 2 && $line =~ $regex_json_end ]]; then discard=0 fi done diff --git a/tests/qemu-iotests/common.rc b/tests/qemu-iotests/common.rc index XXXXXXX..XXXXXXX 100644 --- a/tests/qemu-iotests/common.rc +++ b/tests/qemu-iotests/common.rc @@ -XXX,XX +XXX,XX @@ _img_info() discard=0 regex_json_spec_start='^ *"format-specific": \{' + regex_json_child_start='^ *"children": \[' $QEMU_IMG info $QEMU_IMG_EXTRA_ARGS "$@" "$TEST_IMG" 2>&1 | \ sed -e "s#$REMOTE_TEST_DIR#TEST_DIR#g" \ -e "s#$IMGPROTO:$TEST_DIR#TEST_DIR#g" \ @@ -XXX,XX +XXX,XX @@ _img_info() -e "/^disk size:/ D" \ -e "/actual-size/ D" | \ while IFS='' read -r line; do - if [[ $format_specific == 1 ]]; then - discard=0 - elif [[ $line == "Format specific information:" ]]; then - discard=1 - elif [[ $line =~ $regex_json_spec_start ]]; then - discard=2 - regex_json_spec_end="^${line%%[^ ]*}\\},? *$" + if [[ $discard == 0 ]]; then + if [[ $format_specific == 0 && $line == "Format specific information:" ]]; then + discard=1 + elif [[ $line =~ "Child node '/" ]]; then + discard=1 + elif [[ $format_specific == 0 && $line =~ $regex_json_spec_start ]]; then + discard=2 + regex_json_end="^${line%%[^ ]*}\\},? *$" + elif [[ $line =~ $regex_json_child_start ]]; then + discard=2 + regex_json_end="^${line%%[^ ]*}\\],? *$" + fi fi if [[ $discard == 0 ]]; then echo "$line" elif [[ $discard == 1 && ! $line ]]; then echo discard=0 - elif [[ $discard == 2 && $line =~ $regex_json_spec_end ]]; then + elif [[ $discard == 2 && $line =~ $regex_json_end ]]; then discard=0 fi done -- 2.38.1 From: Hanna Reitz <hreitz@redhat.com> These tests read size information (sometimes disk size, sometimes virtual size) from qemu-img info's output. Once qemu-img starts printing info about child nodes, we are going to see multiple instances of that per image, but these tests are only interested in the first one, so use "head -n 1" to get it. Signed-off-by: Hanna Reitz <hreitz@redhat.com> Message-Id: <20220620162704.80987-11-hreitz@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- tests/qemu-iotests/106 | 4 ++-- tests/qemu-iotests/214 | 6 ++++-- tests/qemu-iotests/308 | 4 ++-- 3 files changed, 8 insertions(+), 6 deletions(-) diff --git a/tests/qemu-iotests/106 b/tests/qemu-iotests/106 index XXXXXXX..XXXXXXX 100755 --- a/tests/qemu-iotests/106 +++ b/tests/qemu-iotests/106 @@ -XXX,XX +XXX,XX @@ for create_mode in off falloc full; do expected_size=$((expected_size + $GROWTH_SIZE)) fi - actual_size=$($QEMU_IMG info -f "$IMGFMT" "$TEST_IMG" | grep 'disk size') + actual_size=$($QEMU_IMG info -f "$IMGFMT" "$TEST_IMG" | grep 'disk size' | head -n 1) actual_size=$(echo "$actual_size" | sed -e 's/^[^0-9]*\([0-9]\+\).*$/\1/') # The actual size may exceed the expected size, depending on the file @@ -XXX,XX +XXX,XX @@ for growth_mode in falloc full; do _make_test_img -o "extent_size_hint=0" 2G $QEMU_IMG resize -f "$IMGFMT" --preallocation=$growth_mode "$TEST_IMG" +${GROWTH_SIZE}K - actual_size=$($QEMU_IMG info -f "$IMGFMT" "$TEST_IMG" | grep 'disk size') + actual_size=$($QEMU_IMG info -f "$IMGFMT" "$TEST_IMG" | grep 'disk size' | head -n 1) actual_size=$(echo "$actual_size" | sed -e 's/^[^0-9]*\([0-9]\+\).*$/\1/') if [ $actual_size -lt $GROWTH_SIZE ]; then diff --git a/tests/qemu-iotests/214 b/tests/qemu-iotests/214 index XXXXXXX..XXXXXXX 100755 --- a/tests/qemu-iotests/214 +++ b/tests/qemu-iotests/214 @@ -XXX,XX +XXX,XX @@ let data_size="8 * $cluster_size" $QEMU_IO -c "write -P 0xaa 0 $data_size" "$TEST_IMG" \ 2>&1 | _filter_qemu_io | _filter_testdir sizeA=$($QEMU_IMG info --output=json "$TEST_IMG" | - sed -n '/"actual-size":/ s/[^0-9]//gp') + sed -n '/"actual-size":/ s/[^0-9]//gp' | + head -n 1) _make_test_img 2M -o cluster_size=$cluster_size echo "Write compressed data:" @@ -XXX,XX +XXX,XX @@ $QEMU_IO -c "write -P 0xcc $offset $data_size" "json:{\ _filter_qemu_io | _filter_testdir sizeB=$($QEMU_IMG info --output=json "$TEST_IMG" | - sed -n '/"actual-size":/ s/[^0-9]//gp') + sed -n '/"actual-size":/ s/[^0-9]//gp' | + head -n 1) if [ $sizeA -lt $sizeB ] then diff --git a/tests/qemu-iotests/308 b/tests/qemu-iotests/308 index XXXXXXX..XXXXXXX 100755 --- a/tests/qemu-iotests/308 +++ b/tests/qemu-iotests/308 @@ -XXX,XX +XXX,XX @@ echo echo '=== Remove export ===' # Double-check that $EXT_MP appears as a non-empty file (the raw image) -$QEMU_IMG info -f raw "$EXT_MP" | grep 'virtual size' +$QEMU_IMG info -f raw "$EXT_MP" | grep 'virtual size' | head -n 1 fuse_export_del 'export-mp' # See that the file appears empty again -$QEMU_IMG info -f raw "$EXT_MP" | grep 'virtual size' +$QEMU_IMG info -f raw "$EXT_MP" | grep 'virtual size' | head -n 1 echo echo '=== Writable export ===' -- 2.38.1 From: Hanna Reitz <hreitz@redhat.com> For every node in the backing chain, collect its BlockGraphInfo struct using bdrv_query_block_graph_info(). Print all nodes' information, indenting child nodes and labelling them with a path constructed from the child names leading to the node from the root (e.g. /file/file). Note that we open each image with BDRV_O_NO_BACKING, so its backing child is omitted from this graph, and thus presented in the previous manner: By simply concatenating all images' information, separated with blank lines. This affects two iotests: - 065: Here we try to get the format node's format specific information. The pre-patch code does so by taking all lines from "Format specific information:" until an empty line. This format specific information is no longer followed by an empty line, though, but by child node information, so limit the range by "Child node '/file':". - 302: Calls qemu_img() for qemu-img info directly, which does not filter the output, so the child node information ends up in the output. Signed-off-by: Hanna Reitz <hreitz@redhat.com> Message-Id: <20220620162704.80987-12-hreitz@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- qapi/block-core.json | 4 +-- qemu-img.c | 69 ++++++++++++++++++++++++++------------ tests/qemu-iotests/065 | 2 +- tests/qemu-iotests/302.out | 5 +++ 4 files changed, 56 insertions(+), 24 deletions(-) diff --git a/qapi/block-core.json b/qapi/block-core.json index XXXXXXX..XXXXXXX 100644 --- a/qapi/block-core.json +++ b/qapi/block-core.json @@ -XXX,XX +XXX,XX @@ ## # @DummyBlockCoreForceArrays: # -# Not used by QMP; hack to let us use BlockNodeInfoList internally +# Not used by QMP; hack to let us use BlockGraphInfoList internally # # Since: 8.0 ## { 'struct': 'DummyBlockCoreForceArrays', - 'data': { 'unused-block-node-info': ['BlockNodeInfo'] } } + 'data': { 'unused-block-graph-info': ['BlockGraphInfo'] } } 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 void dump_snapshots(BlockDriverState *bs) g_free(sn_tab); } -static void dump_json_block_node_info_list(BlockNodeInfoList *list) +static void dump_json_block_graph_info_list(BlockGraphInfoList *list) { GString *str; QObject *obj; Visitor *v = qobject_output_visitor_new(&obj); - visit_type_BlockNodeInfoList(v, NULL, &list, &error_abort); + visit_type_BlockGraphInfoList(v, NULL, &list, &error_abort); visit_complete(v, &obj); str = qobject_to_json_pretty(obj, true); assert(str != NULL); @@ -XXX,XX +XXX,XX @@ static void dump_json_block_node_info_list(BlockNodeInfoList *list) g_string_free(str, true); } -static void dump_json_block_node_info(BlockNodeInfo *info) +static void dump_json_block_graph_info(BlockGraphInfo *info) { GString *str; QObject *obj; Visitor *v = qobject_output_visitor_new(&obj); - visit_type_BlockNodeInfo(v, NULL, &info, &error_abort); + visit_type_BlockGraphInfo(v, NULL, &info, &error_abort); visit_complete(v, &obj); str = qobject_to_json_pretty(obj, true); assert(str != NULL); @@ -XXX,XX +XXX,XX @@ static void dump_json_block_node_info(BlockNodeInfo *info) g_string_free(str, true); } -static void dump_human_image_info_list(BlockNodeInfoList *list) +static void dump_human_image_info(BlockGraphInfo *info, int indentation, + const char *path) { - BlockNodeInfoList *elem; + BlockChildInfoList *children_list; + + bdrv_node_info_dump(qapi_BlockGraphInfo_base(info), indentation); + + for (children_list = info->children; children_list; + children_list = children_list->next) + { + BlockChildInfo *child = children_list->value; + g_autofree char *child_path; + + printf("%*sChild node '%s%s':\n", + indentation * 4, "", path, child->name); + child_path = g_strdup_printf("%s%s/", path, child->name); + dump_human_image_info(child->info, indentation + 1, child_path); + } +} + +static void dump_human_image_info_list(BlockGraphInfoList *list) +{ + BlockGraphInfoList *elem; bool delim = false; for (elem = list; elem; elem = elem->next) { @@ -XXX,XX +XXX,XX @@ static void dump_human_image_info_list(BlockNodeInfoList *list) } delim = true; - bdrv_node_info_dump(elem->value, 0); + dump_human_image_info(elem->value, 0, "/"); } } @@ -XXX,XX +XXX,XX @@ static gboolean str_equal_func(gconstpointer a, gconstpointer b) } /** - * Open an image file chain and return an BlockNodeInfoList + * Open an image file chain and return an BlockGraphInfoList * * @filename: topmost image filename * @fmt: topmost image format (may be NULL to autodetect) @@ -XXX,XX +XXX,XX @@ static gboolean str_equal_func(gconstpointer a, gconstpointer b) * opening an image file. If there was an error a message will have been * printed to stderr. */ -static BlockNodeInfoList *collect_image_info_list(bool image_opts, - const char *filename, - const char *fmt, - bool chain, bool force_share) +static BlockGraphInfoList *collect_image_info_list(bool image_opts, + const char *filename, + const char *fmt, + bool chain, bool force_share) { - BlockNodeInfoList *head = NULL; - BlockNodeInfoList **tail = &head; + BlockGraphInfoList *head = NULL; + BlockGraphInfoList **tail = &head; GHashTable *filenames; Error *err = NULL; @@ -XXX,XX +XXX,XX @@ static BlockNodeInfoList *collect_image_info_list(bool image_opts, while (filename) { BlockBackend *blk; BlockDriverState *bs; - BlockNodeInfo *info; + BlockGraphInfo *info; if (g_hash_table_lookup_extended(filenames, filename, NULL, NULL)) { error_report("Backing file '%s' creates an infinite loop.", @@ -XXX,XX +XXX,XX @@ static BlockNodeInfoList *collect_image_info_list(bool image_opts, } bs = blk_bs(blk); - bdrv_query_block_node_info(bs, &info, &err); + /* + * Note that the returned BlockGraphInfo object will not have + * information about this image's backing node, because we have opened + * it with BDRV_O_NO_BACKING. Printing this object will therefore not + * duplicate the backing chain information that we obtain by walking + * the chain manually here. + */ + bdrv_query_block_graph_info(bs, &info, &err); if (err) { error_report_err(err); blk_unref(blk); @@ -XXX,XX +XXX,XX @@ static BlockNodeInfoList *collect_image_info_list(bool image_opts, return head; err: - qapi_free_BlockNodeInfoList(head); + qapi_free_BlockGraphInfoList(head); g_hash_table_destroy(filenames); return NULL; } @@ -XXX,XX +XXX,XX @@ static int img_info(int argc, char **argv) OutputFormat output_format = OFORMAT_HUMAN; bool chain = false; const char *filename, *fmt, *output; - BlockNodeInfoList *list; + BlockGraphInfoList *list; bool image_opts = false; bool force_share = false; @@ -XXX,XX +XXX,XX @@ static int img_info(int argc, char **argv) break; case OFORMAT_JSON: if (chain) { - dump_json_block_node_info_list(list); + dump_json_block_graph_info_list(list); } else { - dump_json_block_node_info(list->value); + dump_json_block_graph_info(list->value); } break; } - qapi_free_BlockNodeInfoList(list); + qapi_free_BlockGraphInfoList(list); return 0; } diff --git a/tests/qemu-iotests/065 b/tests/qemu-iotests/065 index XXXXXXX..XXXXXXX 100755 --- a/tests/qemu-iotests/065 +++ b/tests/qemu-iotests/065 @@ -XXX,XX +XXX,XX @@ class TestQemuImgInfo(TestImageInfoSpecific): def test_human(self): data = qemu_img('info', '--output=human', test_img).stdout.split('\n') data = data[(data.index('Format specific information:') + 1) - :data.index('')] + :data.index("Child node '/file':")] for field in data: self.assertTrue(re.match('^ {4}[^ ]', field) is not None) data = [line.strip() for line in data] diff --git a/tests/qemu-iotests/302.out b/tests/qemu-iotests/302.out index XXXXXXX..XXXXXXX 100644 --- a/tests/qemu-iotests/302.out +++ b/tests/qemu-iotests/302.out @@ -XXX,XX +XXX,XX @@ image: nbd+unix:///exp?socket=SOCK_DIR/PID-nbd-sock file format: raw virtual size: 448 KiB (458752 bytes) disk size: unavailable +Child node '/file': + image: nbd+unix:///exp?socket=SOCK_DIR/PID-nbd-sock + file format: nbd + virtual size: 448 KiB (458752 bytes) + disk size: unavailable === Converted image info === image: TEST_IMG -- 2.38.1 From: Hanna Reitz <hreitz@redhat.com> Currently, when querying a qcow2 image, qemu-img info reports something like this: image: test.qcow2 file format: qcow2 virtual size: 64 MiB (67108864 bytes) disk size: 196 KiB cluster_size: 65536 Format specific information: compat: 1.1 compression type: zlib lazy refcounts: false refcount bits: 16 corrupt: false extended l2: false Child node '/file': image: test.qcow2 file format: file virtual size: 192 KiB (197120 bytes) disk size: 196 KiB Format specific information: extent size hint: 1048576 Notably, the way the keys are named is specific for image files: The filename is shown under "image", the BDS driver under "file format", and the BDS length under "virtual size". This does not make much sense for nodes that are not actually supposed to be guest images, like the /file child node shown above. Give bdrv_node_info_dump() a @protocol parameter that gives a hint that the respective node is probably just used for data storage and does not necessarily present the data for a VM guest disk. This renames the keys so that with this patch, the output becomes: image: test.qcow2 [...] Child node '/file': filename: test.qcow2 protocol type: file file length: 192 KiB (197120 bytes) disk size: 196 KiB Format specific information: extent size hint: 1048576 (Perhaps we should also rename "Format specific information", but I could not come up with anything better that will not become problematic if we guess wrong with the protocol "heuristic".) This change affects iotest 302, which has protocol node information in its reference output. Signed-off-by: Hanna Reitz <hreitz@redhat.com> Message-Id: <20220620162704.80987-13-hreitz@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/qapi.h | 2 +- block/monitor/block-hmp-cmds.c | 2 +- block/qapi.c | 39 ++++++++++++++++++++++++++++------ qemu-img.c | 3 ++- tests/qemu-iotests/302.out | 6 +++--- 5 files changed, 39 insertions(+), 13 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 @@ void bdrv_snapshot_dump(QEMUSnapshotInfo *sn); void bdrv_image_info_specific_dump(ImageInfoSpecific *info_spec, const char *prefix, int indentation); -void bdrv_node_info_dump(BlockNodeInfo *info, int indentation); +void bdrv_node_info_dump(BlockNodeInfo *info, int indentation, bool protocol); #endif diff --git a/block/monitor/block-hmp-cmds.c b/block/monitor/block-hmp-cmds.c index XXXXXXX..XXXXXXX 100644 --- a/block/monitor/block-hmp-cmds.c +++ b/block/monitor/block-hmp-cmds.c @@ -XXX,XX +XXX,XX @@ static void print_block_info(Monitor *mon, BlockInfo *info, monitor_printf(mon, "\nImages:\n"); image_info = inserted->image; while (1) { - bdrv_node_info_dump(qapi_ImageInfo_base(image_info), 0); + bdrv_node_info_dump(qapi_ImageInfo_base(image_info), 0, false); if (image_info->backing_image) { image_info = image_info->backing_image; } else { 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 @@ void bdrv_image_info_specific_dump(ImageInfoSpecific *info_spec, visit_free(v); } -void bdrv_node_info_dump(BlockNodeInfo *info, int indentation) +/** + * Print the given @info object in human-readable form. Every field is indented + * using the given @indentation (four spaces per indentation level). + * + * When using this to print a whole block graph, @protocol can be set to true to + * signify that the given information is associated with a protocol node, i.e. + * just data storage for an image, such that the data it presents is not really + * a full VM disk. If so, several fields change name: For example, "virtual + * size" is printed as "file length". + * (Consider a qcow2 image, which is represented by a qcow2 node and a file + * node. Printing a "virtual size" for the file node does not make sense, + * because without the qcow2 node, it is not really a guest disk, so it does not + * have a "virtual size". Therefore, we call it "file length" instead.) + * + * @protocol is ignored when @indentation is 0, because we take that to mean + * that the associated node is the root node in the queried block graph, and + * thus is always to be interpreted as a standalone guest disk. + */ +void bdrv_node_info_dump(BlockNodeInfo *info, int indentation, bool protocol) { char *size_buf, *dsize_buf; g_autofree char *ind_s = g_strdup_printf("%*s", indentation * 4, ""); + if (indentation == 0) { + /* Top level, consider this a normal image */ + protocol = false; + } + if (!info->has_actual_size) { dsize_buf = g_strdup("unavailable"); } else { dsize_buf = size_to_str(info->actual_size); } size_buf = size_to_str(info->virtual_size); - qemu_printf("%simage: %s\n" - "%sfile format: %s\n" - "%svirtual size: %s (%" PRId64 " bytes)\n" + qemu_printf("%s%s: %s\n" + "%s%s: %s\n" + "%s%s: %s (%" PRId64 " bytes)\n" "%sdisk size: %s\n", - ind_s, info->filename, - ind_s, info->format, - ind_s, size_buf, info->virtual_size, + ind_s, protocol ? "filename" : "image", info->filename, + ind_s, protocol ? "protocol type" : "file format", + info->format, + ind_s, protocol ? "file length" : "virtual size", + size_buf, info->virtual_size, ind_s, dsize_buf); g_free(size_buf); g_free(dsize_buf); 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 void dump_human_image_info(BlockGraphInfo *info, int indentation, { BlockChildInfoList *children_list; - bdrv_node_info_dump(qapi_BlockGraphInfo_base(info), indentation); + bdrv_node_info_dump(qapi_BlockGraphInfo_base(info), indentation, + info->children == NULL); for (children_list = info->children; children_list; children_list = children_list->next) diff --git a/tests/qemu-iotests/302.out b/tests/qemu-iotests/302.out index XXXXXXX..XXXXXXX 100644 --- a/tests/qemu-iotests/302.out +++ b/tests/qemu-iotests/302.out @@ -XXX,XX +XXX,XX @@ file format: raw virtual size: 448 KiB (458752 bytes) disk size: unavailable Child node '/file': - image: nbd+unix:///exp?socket=SOCK_DIR/PID-nbd-sock - file format: nbd - virtual size: 448 KiB (458752 bytes) + filename: nbd+unix:///exp?socket=SOCK_DIR/PID-nbd-sock + protocol type: nbd + file length: 448 KiB (458752 bytes) disk size: unavailable === Converted image info === -- 2.38.1