The following changes since commit d9bbfea646e86426d549bd612cd9f91e49aa50c2: Merge remote-tracking branch 'remotes/riscv/tags/riscv-qemu-upstream-v8.2' into staging (2018-03-09 10:58:57 +0000) are available in the git repository at: git://repo.or.cz/qemu/kevin.git tags/for-upstream for you to fetch changes up to a1be5921e35dcf84ce9e3c9a5c3029cea530a60b: Merge remote-tracking branch 'mreitz/tags/pull-block-2018-03-09' into queue-block (2018-03-09 16:09:06 +0100) ---------------------------------------------------------------- Block layer patches ---------------------------------------------------------------- Alberto Garcia (8): qcow2: Generalize validate_table_offset() into qcow2_validate_table() qcow2: Check L1 table offset in qcow2_snapshot_load_tmp() qcow2: Check L1 table parameters in qcow2_expand_zero_clusters() qcow2: Check snapshot L1 tables in qcow2_check_metadata_overlap() qcow2: Check snapshot L1 table in qcow2_snapshot_goto() qcow2: Check snapshot L1 table in qcow2_snapshot_delete() qcow2: Make qemu-img check detect corrupted L1 tables in snapshots iotests: Tweak 030 in order to trigger a race condition with parallel jobs Daniel P. Berrangé (1): block: implement the bdrv_reopen_prepare helper for LUKS driver Eric Blake (1): iotests: Mark all tests executable Fam Zheng (2): iotests: Test creating overlay when guest running iotests: Skip test for ENOMEM error Kevin Wolf (38): block/qapi: Introduce BlockdevCreateOptions block/qapi: Add qcow2 create options to schema qcow2: Rename qcow2_co_create2() to qcow2_co_create() qcow2: Let qcow2_create() handle protocol layer qcow2: Pass BlockdevCreateOptions to qcow2_co_create() qcow2: Use BlockdevRef in qcow2_co_create() qcow2: Use QCryptoBlockCreateOptions in qcow2_co_create() qcow2: Handle full/falloc preallocation in qcow2_co_create() util: Add qemu_opts_to_qdict_filtered() test-qemu-opts: Test qemu_opts_append() test-qemu-opts: Test qemu_opts_to_qdict_filtered() qdict: Introduce qdict_rename_keys() qcow2: Use visitor for options in qcow2_create() block: Make bdrv_is_whitelisted() public block: x-blockdev-create QMP command file-posix: Support .bdrv_co_create file-win32: Support .bdrv_co_create gluster: Support .bdrv_co_create rbd: Fix use after free in qemu_rbd_set_keypairs() error path rbd: Factor out qemu_rbd_connect() rbd: Remove non-schema options from runtime_opts rbd: Pass BlockdevOptionsRbd to qemu_rbd_connect() rbd: Support .bdrv_co_create rbd: Assign s->snap/image_name in qemu_rbd_open() rbd: Use qemu_rbd_connect() in qemu_rbd_do_create() nfs: Use QAPI options in nfs_client_open() nfs: Support .bdrv_co_create sheepdog: QAPIfy "redundancy" create option sheepdog: Support .bdrv_co_create ssh: Use QAPI BlockdevOptionsSsh object ssh: QAPIfy host-key-check option ssh: Pass BlockdevOptionsSsh to connect_to_ssh() ssh: Support .bdrv_co_create file-posix: Fix no-op bdrv_truncate() with falloc preallocation block: Fail bdrv_truncate() with negative size qemu-iotests: Test qcow2 over file image creation with QMP qemu-iotests: Test ssh image creation over QMP Merge remote-tracking branch 'mreitz/tags/pull-block-2018-03-09' into queue-block Paolo Bonzini (6): qcow2: introduce qcow2_write_caches and qcow2_flush_caches qcow2: fix flushing after dirty bitmap metadata writes qcow2: make qcow2_do_open a coroutine_fn qed: make bdrv_qed_do_open a coroutine_fn block: convert bdrv_invalidate_cache callback to coroutine_fn block: convert bdrv_check callback to coroutine_fn Stefan Hajnoczi (1): qemu-iotests: fix 203 migration completion race qapi/block-core.json | 326 +++++++++++++++++++++++++- block/qcow2.h | 12 +- include/block/block.h | 2 + include/block/block_int.h | 13 +- include/qapi/qmp/qdict.h | 6 + include/qemu/option.h | 2 + block.c | 138 +++++++++++- block/create.c | 76 +++++++ block/crypto.c | 7 + block/file-posix.c | 93 +++++--- block/file-win32.c | 47 +++- block/gluster.c | 135 +++++++---- block/iscsi.c | 6 +- block/nfs.c | 244 +++++++++----------- block/parallels.c | 17 +- block/qcow2-bitmap.c | 4 +- block/qcow2-cluster.c | 24 +- block/qcow2-refcount.c | 52 ++++- block/qcow2-snapshot.c | 24 +- block/qcow2.c | 552 ++++++++++++++++++++++++++++----------------- block/qed-check.c | 1 + block/qed-table.c | 26 +-- block/qed.c | 66 ++++-- block/rbd.c | 403 +++++++++++++++++---------------- block/sheepdog.c | 321 ++++++++++++++++++-------- block/ssh.c | 290 +++++++++++++----------- block/vdi.c | 6 +- block/vhdx.c | 7 +- block/vmdk.c | 7 +- qobject/qdict.c | 34 +++ tests/check-qdict.c | 129 +++++++++++ tests/test-qemu-opts.c | 253 +++++++++++++++++++++ util/qemu-option.c | 42 +++- block/Makefile.objs | 2 +- tests/qemu-iotests/030 | 52 ++++- tests/qemu-iotests/030.out | 4 +- tests/qemu-iotests/049.out | 8 +- tests/qemu-iotests/059 | 5 +- tests/qemu-iotests/080 | 22 +- tests/qemu-iotests/080.out | 58 ++++- tests/qemu-iotests/096 | 0 tests/qemu-iotests/112.out | 4 +- tests/qemu-iotests/124 | 0 tests/qemu-iotests/129 | 0 tests/qemu-iotests/132 | 0 tests/qemu-iotests/136 | 0 tests/qemu-iotests/139 | 0 tests/qemu-iotests/148 | 0 tests/qemu-iotests/152 | 0 tests/qemu-iotests/153 | 8 +- tests/qemu-iotests/153.out | 7 +- tests/qemu-iotests/163 | 0 tests/qemu-iotests/203 | 15 +- tests/qemu-iotests/203.out | 5 + tests/qemu-iotests/205 | 0 tests/qemu-iotests/206 | 436 +++++++++++++++++++++++++++++++++++ tests/qemu-iotests/206.out | 209 +++++++++++++++++ tests/qemu-iotests/207 | 261 +++++++++++++++++++++ tests/qemu-iotests/207.out | 75 ++++++ tests/qemu-iotests/group | 2 + 60 files changed, 3574 insertions(+), 964 deletions(-) create mode 100644 block/create.c mode change 100644 => 100755 tests/qemu-iotests/096 mode change 100644 => 100755 tests/qemu-iotests/124 mode change 100644 => 100755 tests/qemu-iotests/129 mode change 100644 => 100755 tests/qemu-iotests/132 mode change 100644 => 100755 tests/qemu-iotests/136 mode change 100644 => 100755 tests/qemu-iotests/139 mode change 100644 => 100755 tests/qemu-iotests/148 mode change 100644 => 100755 tests/qemu-iotests/152 mode change 100644 => 100755 tests/qemu-iotests/163 mode change 100644 => 100755 tests/qemu-iotests/205 create mode 100755 tests/qemu-iotests/206 create mode 100644 tests/qemu-iotests/206.out create mode 100755 tests/qemu-iotests/207 create mode 100644 tests/qemu-iotests/207.out From: "Daniel P. Berrange" <berrange@redhat.com> If the bdrv_reopen_prepare helper isn't provided, the qemu-img commit command fails to re-open the base layer after committing changes into it. Provide a no-op implementation for the LUKS driver, since there is not any custom work that needs doing to re-open it. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/crypto.c | 7 +++++++ 1 file changed, 7 insertions(+) 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_close(BlockDriverState *bs) qcrypto_block_free(crypto->block); } +static int block_crypto_reopen_prepare(BDRVReopenState *state, + BlockReopenQueue *queue, Error **errp) +{ + /* nothing needs checking */ + return 0; +} /* * 1 MB bounce buffer gives good performance / memory tradeoff @@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_crypto_luks = { .bdrv_truncate = block_crypto_truncate, .create_opts = &block_crypto_create_opts_luks, + .bdrv_reopen_prepare = block_crypto_reopen_prepare, .bdrv_refresh_limits = block_crypto_refresh_limits, .bdrv_co_preadv = block_crypto_co_preadv, .bdrv_co_pwritev = block_crypto_co_pwritev, -- 2.13.6 From: Paolo Bonzini <pbonzini@redhat.com> They will be used to avoid recursively taking s->lock during bdrv_open or bdrv_check. Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Message-Id: <1516279431-30424-7-git-send-email-pbonzini@redhat.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/qcow2.h | 2 ++ block/qcow2-refcount.c | 28 ++++++++++++++++++++++++++++ block/qcow2.c | 20 ++++---------------- 3 files changed, 34 insertions(+), 16 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_free_any_clusters(BlockDriverState *bs, uint64_t l2_entry, int qcow2_update_snapshot_refcount(BlockDriverState *bs, int64_t l1_table_offset, int l1_size, int addend); +int coroutine_fn qcow2_flush_caches(BlockDriverState *bs); +int coroutine_fn qcow2_write_caches(BlockDriverState *bs); int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res, BdrvCheckMode fix); 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 @@ void qcow2_free_any_clusters(BlockDriverState *bs, uint64_t l2_entry, } } +int coroutine_fn qcow2_write_caches(BlockDriverState *bs) +{ + BDRVQcow2State *s = bs->opaque; + int ret; + ret = qcow2_cache_write(bs, s->l2_table_cache); + if (ret < 0) { + return ret; + } + + if (qcow2_need_accurate_refcounts(s)) { + ret = qcow2_cache_write(bs, s->refcount_block_cache); + if (ret < 0) { + return ret; + } + } + + return 0; +} + +int coroutine_fn qcow2_flush_caches(BlockDriverState *bs) +{ + int ret = qcow2_write_caches(bs); + if (ret < 0) { + return ret; + } + + return bdrv_flush(bs->file->bs); +} /*********************************************************/ /* snapshots and image creation */ 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_mark_clean(BlockDriverState *bs) s->incompatible_features &= ~QCOW2_INCOMPAT_DIRTY; - ret = bdrv_flush(bs); + ret = qcow2_flush_caches(bs); if (ret < 0) { return ret; } @@ -XXX,XX +XXX,XX @@ int qcow2_mark_consistent(BlockDriverState *bs) BDRVQcow2State *s = bs->opaque; if (s->incompatible_features & QCOW2_INCOMPAT_CORRUPT) { - int ret = bdrv_flush(bs); + int ret = qcow2_flush_caches(bs); if (ret < 0) { return ret; } @@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow2_co_flush_to_os(BlockDriverState *bs) int ret; qemu_co_mutex_lock(&s->lock); - ret = qcow2_cache_write(bs, s->l2_table_cache); - if (ret < 0) { - qemu_co_mutex_unlock(&s->lock); - return ret; - } - - if (qcow2_need_accurate_refcounts(s)) { - ret = qcow2_cache_write(bs, s->refcount_block_cache); - if (ret < 0) { - qemu_co_mutex_unlock(&s->lock); - return ret; - } - } + ret = qcow2_write_caches(bs); qemu_co_mutex_unlock(&s->lock); - return 0; + return ret; } static BlockMeasureInfo *qcow2_measure(QemuOpts *opts, BlockDriverState *in_bs, -- 2.13.6 From: Paolo Bonzini <pbonzini@redhat.com> update_header_sync itself does not need to flush the caches to disk. The only paths that allocate clusters are: - bitmap_list_store with in_place=false, called by update_ext_header_and_dir - store_bitmap_data, called by store_bitmap - store_bitmap, called by qcow2_store_persistent_dirty_bitmaps and followed by update_ext_header_and_dir So in the end the central place where we need to flush the caches is update_ext_header_and_dir. Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/qcow2-bitmap.c | 4 ++-- 1 file changed, 2 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 ret; } - return bdrv_flush(bs); + return bdrv_flush(bs->file->bs); } static inline void bitmap_table_to_be(uint64_t *bitmap_table, size_t size) @@ -XXX,XX +XXX,XX @@ static int update_ext_header_and_dir(BlockDriverState *bs, return ret; } - ret = bdrv_flush(bs->file->bs); + ret = qcow2_flush_caches(bs); if (ret < 0) { goto fail; } -- 2.13.6 From: Paolo Bonzini <pbonzini@redhat.com> It is called from qcow2_invalidate_cache in coroutine context (incoming migration runs in a coroutine), so it's cleaner if metadata is always loaded from a coroutine. Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Message-Id: <1516279431-30424-4-git-send-email-pbonzini@redhat.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/qcow2.c | 46 +++++++++++++++++++++++++++++++++++++++++----- 1 file changed, 41 insertions(+), 5 deletions(-) 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_update_options(BlockDriverState *bs, QDict *options, return ret; } -static int qcow2_do_open(BlockDriverState *bs, QDict *options, int flags, - Error **errp) +/* Called with s->lock held. */ +static int coroutine_fn qcow2_do_open(BlockDriverState *bs, QDict *options, + int flags, Error **errp) { BDRVQcow2State *s = bs->opaque; unsigned int len, i; @@ -XXX,XX +XXX,XX @@ static int qcow2_do_open(BlockDriverState *bs, QDict *options, int flags, } } - /* Initialise locks */ - qemu_co_mutex_init(&s->lock); bs->supported_zero_flags = header.version >= 3 ? BDRV_REQ_MAY_UNMAP : 0; /* Repair image if dirty */ @@ -XXX,XX +XXX,XX @@ static int qcow2_do_open(BlockDriverState *bs, QDict *options, int flags, return ret; } +typedef struct QCow2OpenCo { + BlockDriverState *bs; + QDict *options; + int flags; + Error **errp; + int ret; +} QCow2OpenCo; + +static void coroutine_fn qcow2_open_entry(void *opaque) +{ + QCow2OpenCo *qoc = opaque; + BDRVQcow2State *s = qoc->bs->opaque; + + qemu_co_mutex_lock(&s->lock); + qoc->ret = qcow2_do_open(qoc->bs, qoc->options, qoc->flags, qoc->errp); + qemu_co_mutex_unlock(&s->lock); +} + static int qcow2_open(BlockDriverState *bs, QDict *options, int flags, Error **errp) { + BDRVQcow2State *s = bs->opaque; + QCow2OpenCo qoc = { + .bs = bs, + .options = options, + .flags = flags, + .errp = errp, + .ret = -EINPROGRESS + }; + bs->file = bdrv_open_child(NULL, options, "file", bs, &child_file, false, errp); if (!bs->file) { return -EINVAL; } - return qcow2_do_open(bs, options, flags, errp); + /* Initialise locks */ + qemu_co_mutex_init(&s->lock); + + if (qemu_in_coroutine()) { + /* From bdrv_co_create. */ + qcow2_open_entry(&qoc); + } else { + qemu_coroutine_enter(qemu_coroutine_create(qcow2_open_entry, &qoc)); + BDRV_POLL_WHILE(bs, qoc.ret == -EINPROGRESS); + } + return qoc.ret; } static void qcow2_refresh_limits(BlockDriverState *bs, Error **errp) -- 2.13.6 From: Paolo Bonzini <pbonzini@redhat.com> It is called from qed_invalidate_cache in coroutine context (incoming migration runs in a coroutine), so it's cleaner if metadata is always loaded from a coroutine. Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Message-Id: <1516279431-30424-5-git-send-email-pbonzini@redhat.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/qed.c | 40 +++++++++++++++++++++++++++++++++++++--- 1 file changed, 37 insertions(+), 3 deletions(-) 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) qemu_co_queue_init(&s->allocating_write_reqs); } -static int bdrv_qed_do_open(BlockDriverState *bs, QDict *options, int flags, - Error **errp) +/* Called with table_lock held. */ +static int coroutine_fn bdrv_qed_do_open(BlockDriverState *bs, QDict *options, + int flags, Error **errp) { BDRVQEDState *s = bs->opaque; QEDHeader le_header; @@ -XXX,XX +XXX,XX @@ out: return ret; } +typedef struct QEDOpenCo { + BlockDriverState *bs; + QDict *options; + int flags; + Error **errp; + int ret; +} QEDOpenCo; + +static void coroutine_fn bdrv_qed_open_entry(void *opaque) +{ + QEDOpenCo *qoc = opaque; + BDRVQEDState *s = qoc->bs->opaque; + + qemu_co_mutex_lock(&s->table_lock); + qoc->ret = bdrv_qed_do_open(qoc->bs, qoc->options, qoc->flags, qoc->errp); + qemu_co_mutex_unlock(&s->table_lock); +} + static int bdrv_qed_open(BlockDriverState *bs, QDict *options, int flags, Error **errp) { + QEDOpenCo qoc = { + .bs = bs, + .options = options, + .flags = flags, + .errp = errp, + .ret = -EINPROGRESS + }; + bs->file = bdrv_open_child(NULL, options, "file", bs, &child_file, false, errp); if (!bs->file) { @@ -XXX,XX +XXX,XX @@ static int bdrv_qed_open(BlockDriverState *bs, QDict *options, int flags, } bdrv_qed_init_state(bs); - return bdrv_qed_do_open(bs, options, flags, errp); + if (qemu_in_coroutine()) { + bdrv_qed_open_entry(&qoc); + } else { + qemu_coroutine_enter(qemu_coroutine_create(bdrv_qed_open_entry, &qoc)); + BDRV_POLL_WHILE(bs, qoc.ret == -EINPROGRESS); + } + BDRV_POLL_WHILE(bs, qoc.ret == -EINPROGRESS); + return qoc.ret; } static void bdrv_qed_refresh_limits(BlockDriverState *bs, Error **errp) -- 2.13.6 From: Paolo Bonzini <pbonzini@redhat.com> QED's bdrv_invalidate_cache implementation would like to reuse functions that acquire/release the metadata locks. Call it from coroutine context to simplify the logic. Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Message-Id: <1516279431-30424-6-git-send-email-pbonzini@redhat.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block_int.h | 3 ++- block.c | 41 +++++++++++++++++++++++++++++++++++++---- block/iscsi.c | 6 +++--- block/nfs.c | 6 +++--- block/qcow2.c | 7 +++++-- block/qed.c | 13 +++++-------- block/rbd.c | 6 +++--- 7 files changed, 58 insertions(+), 24 deletions(-) diff --git a/include/block/block_int.h b/include/block/block_int.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block_int.h +++ b/include/block/block_int.h @@ -XXX,XX +XXX,XX @@ struct BlockDriver { /* * Invalidate any cached meta-data. */ - void (*bdrv_invalidate_cache)(BlockDriverState *bs, Error **errp); + void coroutine_fn (*bdrv_co_invalidate_cache)(BlockDriverState *bs, + Error **errp); int (*bdrv_inactivate)(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 bdrv_init_with_whitelist(void) bdrv_init(); } -void bdrv_invalidate_cache(BlockDriverState *bs, Error **errp) +static void coroutine_fn bdrv_co_invalidate_cache(BlockDriverState *bs, + Error **errp) { BdrvChild *child, *parent; uint64_t perm, shared_perm; @@ -XXX,XX +XXX,XX @@ void bdrv_invalidate_cache(BlockDriverState *bs, Error **errp) } QLIST_FOREACH(child, &bs->children, next) { - bdrv_invalidate_cache(child->bs, &local_err); + bdrv_co_invalidate_cache(child->bs, &local_err); if (local_err) { error_propagate(errp, local_err); return; @@ -XXX,XX +XXX,XX @@ void bdrv_invalidate_cache(BlockDriverState *bs, Error **errp) } bdrv_set_perm(bs, perm, shared_perm); - if (bs->drv->bdrv_invalidate_cache) { - bs->drv->bdrv_invalidate_cache(bs, &local_err); + if (bs->drv->bdrv_co_invalidate_cache) { + bs->drv->bdrv_co_invalidate_cache(bs, &local_err); if (local_err) { bs->open_flags |= BDRV_O_INACTIVE; error_propagate(errp, local_err); @@ -XXX,XX +XXX,XX @@ void bdrv_invalidate_cache(BlockDriverState *bs, Error **errp) } } +typedef struct InvalidateCacheCo { + BlockDriverState *bs; + Error **errp; + bool done; +} InvalidateCacheCo; + +static void coroutine_fn bdrv_invalidate_cache_co_entry(void *opaque) +{ + InvalidateCacheCo *ico = opaque; + bdrv_co_invalidate_cache(ico->bs, ico->errp); + ico->done = true; +} + +void bdrv_invalidate_cache(BlockDriverState *bs, Error **errp) +{ + Coroutine *co; + InvalidateCacheCo ico = { + .bs = bs, + .done = false, + .errp = errp + }; + + if (qemu_in_coroutine()) { + /* Fast-path if already in coroutine context */ + bdrv_invalidate_cache_co_entry(&ico); + } else { + co = qemu_coroutine_create(bdrv_invalidate_cache_co_entry, &ico); + qemu_coroutine_enter(co); + BDRV_POLL_WHILE(bs, !ico.done); + } +} + void bdrv_invalidate_cache_all(Error **errp) { BlockDriverState *bs; 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 iscsi_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) return 0; } -static void iscsi_invalidate_cache(BlockDriverState *bs, - Error **errp) +static void coroutine_fn iscsi_co_invalidate_cache(BlockDriverState *bs, + Error **errp) { IscsiLun *iscsilun = bs->opaque; iscsi_allocmap_invalidate(iscsilun); @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_iscsi = { .create_opts = &iscsi_create_opts, .bdrv_reopen_prepare = iscsi_reopen_prepare, .bdrv_reopen_commit = iscsi_reopen_commit, - .bdrv_invalidate_cache = iscsi_invalidate_cache, + .bdrv_co_invalidate_cache = iscsi_co_invalidate_cache, .bdrv_getlength = iscsi_getlength, .bdrv_get_info = iscsi_get_info, 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 @@ static void nfs_refresh_filename(BlockDriverState *bs, QDict *options) } #ifdef LIBNFS_FEATURE_PAGECACHE -static void nfs_invalidate_cache(BlockDriverState *bs, - Error **errp) +static void coroutine_fn nfs_co_invalidate_cache(BlockDriverState *bs, + Error **errp) { NFSClient *client = bs->opaque; nfs_pagecache_invalidate(client->context, client->fh); @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_nfs = { .bdrv_refresh_filename = nfs_refresh_filename, #ifdef LIBNFS_FEATURE_PAGECACHE - .bdrv_invalidate_cache = nfs_invalidate_cache, + .bdrv_co_invalidate_cache = nfs_co_invalidate_cache, #endif }; 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_close(BlockDriverState *bs) qcow2_free_snapshots(bs); } -static void qcow2_invalidate_cache(BlockDriverState *bs, Error **errp) +static void coroutine_fn qcow2_co_invalidate_cache(BlockDriverState *bs, + Error **errp) { BDRVQcow2State *s = bs->opaque; int flags = s->flags; @@ -XXX,XX +XXX,XX @@ static void qcow2_invalidate_cache(BlockDriverState *bs, Error **errp) options = qdict_clone_shallow(bs->options); flags &= ~BDRV_O_INACTIVE; + qemu_co_mutex_lock(&s->lock); ret = qcow2_do_open(bs, options, flags, &local_err); + qemu_co_mutex_unlock(&s->lock); QDECREF(options); if (local_err) { error_propagate(errp, local_err); @@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_qcow2 = { .bdrv_change_backing_file = qcow2_change_backing_file, .bdrv_refresh_limits = qcow2_refresh_limits, - .bdrv_invalidate_cache = qcow2_invalidate_cache, + .bdrv_co_invalidate_cache = qcow2_co_invalidate_cache, .bdrv_inactivate = qcow2_inactivate, .create_opts = &qcow2_create_opts, 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 bdrv_qed_change_backing_file(BlockDriverState *bs, return ret; } -static void bdrv_qed_invalidate_cache(BlockDriverState *bs, Error **errp) +static void coroutine_fn bdrv_qed_co_invalidate_cache(BlockDriverState *bs, + Error **errp) { BDRVQEDState *s = bs->opaque; Error *local_err = NULL; @@ -XXX,XX +XXX,XX @@ static void bdrv_qed_invalidate_cache(BlockDriverState *bs, Error **errp) bdrv_qed_close(bs); bdrv_qed_init_state(bs); - if (qemu_in_coroutine()) { - qemu_co_mutex_lock(&s->table_lock); - } + qemu_co_mutex_lock(&s->table_lock); ret = bdrv_qed_do_open(bs, NULL, bs->open_flags, &local_err); - if (qemu_in_coroutine()) { - qemu_co_mutex_unlock(&s->table_lock); - } + qemu_co_mutex_unlock(&s->table_lock); if (local_err) { error_propagate(errp, local_err); error_prepend(errp, "Could not reopen qed layer: "); @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_qed = { .bdrv_get_info = bdrv_qed_get_info, .bdrv_refresh_limits = bdrv_qed_refresh_limits, .bdrv_change_backing_file = bdrv_qed_change_backing_file, - .bdrv_invalidate_cache = bdrv_qed_invalidate_cache, + .bdrv_co_invalidate_cache = bdrv_qed_co_invalidate_cache, .bdrv_check = bdrv_qed_check, .bdrv_detach_aio_context = bdrv_qed_detach_aio_context, .bdrv_attach_aio_context = bdrv_qed_attach_aio_context, 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 BlockAIOCB *qemu_rbd_aio_pdiscard(BlockDriverState *bs, #endif #ifdef LIBRBD_SUPPORTS_INVALIDATE -static void qemu_rbd_invalidate_cache(BlockDriverState *bs, - Error **errp) +static void coroutine_fn qemu_rbd_co_invalidate_cache(BlockDriverState *bs, + Error **errp) { BDRVRBDState *s = bs->opaque; int r = rbd_invalidate_cache(s->image); @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_rbd = { .bdrv_snapshot_list = qemu_rbd_snap_list, .bdrv_snapshot_goto = qemu_rbd_snap_rollback, #ifdef LIBRBD_SUPPORTS_INVALIDATE - .bdrv_invalidate_cache = qemu_rbd_invalidate_cache, + .bdrv_co_invalidate_cache = qemu_rbd_co_invalidate_cache, #endif }; -- 2.13.6 From: Paolo Bonzini <pbonzini@redhat.com> Suggested-by: Kevin Wolf <kwolf@redhat.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Message-Id: <1516279431-30424-8-git-send-email-pbonzini@redhat.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block_int.h | 5 +++-- block.c | 43 ++++++++++++++++++++++++++++++++++++++++--- block/parallels.c | 17 +++++++++++------ block/qcow2.c | 23 +++++++++++++++++++---- block/qed-check.c | 1 + block/qed-table.c | 26 +++++++++----------------- block/qed.c | 13 +++++++++---- block/vdi.c | 6 +++--- block/vhdx.c | 7 ++++--- block/vmdk.c | 7 ++++--- 10 files changed, 103 insertions(+), 45 deletions(-) diff --git a/include/block/block_int.h b/include/block/block_int.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block_int.h +++ b/include/block/block_int.h @@ -XXX,XX +XXX,XX @@ struct BlockDriver { * Returns 0 for completed check, -errno for internal errors. * The check results are stored in result. */ - int (*bdrv_check)(BlockDriverState *bs, BdrvCheckResult *result, - BdrvCheckMode fix); + int coroutine_fn (*bdrv_co_check)(BlockDriverState *bs, + BdrvCheckResult *result, + BdrvCheckMode fix); int (*bdrv_amend_options)(BlockDriverState *bs, QemuOpts *opts, BlockDriverAmendStatusCB *status_cb, diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ static void bdrv_delete(BlockDriverState *bs) * free of errors) or -errno when an internal error occurred. The results of the * check are stored in res. */ -int bdrv_check(BlockDriverState *bs, BdrvCheckResult *res, BdrvCheckMode fix) +static int coroutine_fn bdrv_co_check(BlockDriverState *bs, + BdrvCheckResult *res, BdrvCheckMode fix) { if (bs->drv == NULL) { return -ENOMEDIUM; } - if (bs->drv->bdrv_check == NULL) { + if (bs->drv->bdrv_co_check == NULL) { return -ENOTSUP; } memset(res, 0, sizeof(*res)); - return bs->drv->bdrv_check(bs, res, fix); + return bs->drv->bdrv_co_check(bs, res, fix); +} + +typedef struct CheckCo { + BlockDriverState *bs; + BdrvCheckResult *res; + BdrvCheckMode fix; + int ret; +} CheckCo; + +static void bdrv_check_co_entry(void *opaque) +{ + CheckCo *cco = opaque; + cco->ret = bdrv_co_check(cco->bs, cco->res, cco->fix); +} + +int bdrv_check(BlockDriverState *bs, + BdrvCheckResult *res, BdrvCheckMode fix) +{ + Coroutine *co; + CheckCo cco = { + .bs = bs, + .res = res, + .ret = -EINPROGRESS, + .fix = fix, + }; + + if (qemu_in_coroutine()) { + /* Fast-path if already in coroutine context */ + bdrv_check_co_entry(&cco); + } else { + co = qemu_coroutine_create(bdrv_check_co_entry, &cco); + qemu_coroutine_enter(co); + BDRV_POLL_WHILE(bs, cco.ret == -EINPROGRESS); + } + + return cco.ret; } /* diff --git a/block/parallels.c b/block/parallels.c index XXXXXXX..XXXXXXX 100644 --- a/block/parallels.c +++ b/block/parallels.c @@ -XXX,XX +XXX,XX @@ static coroutine_fn int parallels_co_readv(BlockDriverState *bs, } -static int parallels_check(BlockDriverState *bs, BdrvCheckResult *res, - BdrvCheckMode fix) +static int coroutine_fn parallels_co_check(BlockDriverState *bs, + BdrvCheckResult *res, + BdrvCheckMode fix) { BDRVParallelsState *s = bs->opaque; int64_t size, prev_off, high_off; @@ -XXX,XX +XXX,XX @@ static int parallels_check(BlockDriverState *bs, BdrvCheckResult *res, return size; } + qemu_co_mutex_lock(&s->lock); if (s->header_unclean) { fprintf(stderr, "%s image was not closed correctly\n", fix & BDRV_FIX_ERRORS ? "Repairing" : "ERROR"); @@ -XXX,XX +XXX,XX @@ static int parallels_check(BlockDriverState *bs, BdrvCheckResult *res, prev_off = off; } + ret = 0; if (flush_bat) { ret = bdrv_pwrite_sync(bs->file, 0, s->header, s->header_size); if (ret < 0) { res->check_errors++; - return ret; + goto out; } } @@ -XXX,XX +XXX,XX @@ static int parallels_check(BlockDriverState *bs, BdrvCheckResult *res, if (ret < 0) { error_report_err(local_err); res->check_errors++; - return ret; + goto out; } res->leaks_fixed += count; } } - return 0; +out: + qemu_co_mutex_unlock(&s->lock); + return ret; } @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_parallels = { .bdrv_co_writev = parallels_co_writev, .supports_backing = true, .bdrv_co_create_opts = parallels_co_create_opts, - .bdrv_check = parallels_check, + .bdrv_co_check = parallels_co_check, .create_opts = &parallels_create_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 @@ int qcow2_mark_consistent(BlockDriverState *bs) return 0; } -static int qcow2_check(BlockDriverState *bs, BdrvCheckResult *result, - BdrvCheckMode fix) +static int coroutine_fn qcow2_co_check_locked(BlockDriverState *bs, + BdrvCheckResult *result, + BdrvCheckMode fix) { int ret = qcow2_check_refcounts(bs, result, fix); if (ret < 0) { @@ -XXX,XX +XXX,XX @@ static int qcow2_check(BlockDriverState *bs, BdrvCheckResult *result, return ret; } +static int coroutine_fn qcow2_co_check(BlockDriverState *bs, + BdrvCheckResult *result, + BdrvCheckMode fix) +{ + BDRVQcow2State *s = bs->opaque; + int ret; + + qemu_co_mutex_lock(&s->lock); + ret = qcow2_co_check_locked(bs, result, fix); + qemu_co_mutex_unlock(&s->lock); + return ret; +} + static int validate_table_offset(BlockDriverState *bs, uint64_t offset, uint64_t entries, size_t entry_len) { @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qcow2_do_open(BlockDriverState *bs, QDict *options, (s->incompatible_features & QCOW2_INCOMPAT_DIRTY)) { BdrvCheckResult result = {0}; - ret = qcow2_check(bs, &result, BDRV_FIX_ERRORS | BDRV_FIX_LEAKS); + ret = qcow2_co_check_locked(bs, &result, + BDRV_FIX_ERRORS | BDRV_FIX_LEAKS); if (ret < 0 || result.check_errors) { if (ret >= 0) { ret = -EIO; @@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_qcow2 = { .bdrv_inactivate = qcow2_inactivate, .create_opts = &qcow2_create_opts, - .bdrv_check = qcow2_check, + .bdrv_co_check = qcow2_co_check, .bdrv_amend_options = qcow2_amend_options, .bdrv_detach_aio_context = qcow2_detach_aio_context, 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 void qed_check_mark_clean(BDRVQEDState *s, BdrvCheckResult *result) qed_write_header_sync(s); } +/* Called with table_lock held. */ int qed_check(BDRVQEDState *s, BdrvCheckResult *result, bool fix) { QEDCheck check = { 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 "qed.h" #include "qemu/bswap.h" -/* Called either from qed_check or with table_lock held. */ +/* Called with table_lock held. */ static int qed_read_table(BDRVQEDState *s, uint64_t offset, QEDTable *table) { QEMUIOVector qiov; @@ -XXX,XX +XXX,XX @@ static int qed_read_table(BDRVQEDState *s, uint64_t offset, QEDTable *table) trace_qed_read_table(s, offset, table); - if (qemu_in_coroutine()) { - qemu_co_mutex_unlock(&s->table_lock); - } + qemu_co_mutex_unlock(&s->table_lock); ret = bdrv_preadv(s->bs->file, offset, &qiov); - if (qemu_in_coroutine()) { - qemu_co_mutex_lock(&s->table_lock); - } + qemu_co_mutex_lock(&s->table_lock); if (ret < 0) { goto out; } @@ -XXX,XX +XXX,XX @@ out: * @n: Number of elements * @flush: Whether or not to sync to disk * - * Called either from qed_check or with table_lock held. + * Called with table_lock held. */ static int qed_write_table(BDRVQEDState *s, uint64_t offset, QEDTable *table, unsigned int index, unsigned int n, bool flush) @@ -XXX,XX +XXX,XX @@ static int qed_write_table(BDRVQEDState *s, uint64_t offset, QEDTable *table, /* Adjust for offset into table */ offset += start * sizeof(uint64_t); - if (qemu_in_coroutine()) { - qemu_co_mutex_unlock(&s->table_lock); - } + qemu_co_mutex_unlock(&s->table_lock); ret = bdrv_pwritev(s->bs->file, offset, &qiov); - if (qemu_in_coroutine()) { - qemu_co_mutex_lock(&s->table_lock); - } + qemu_co_mutex_lock(&s->table_lock); trace_qed_write_table_cb(s, table, flush, ret); if (ret < 0) { goto out; @@ -XXX,XX +XXX,XX @@ int qed_read_l1_table_sync(BDRVQEDState *s) return qed_read_table(s, s->header.l1_table_offset, s->l1_table); } -/* Called either from qed_check or with table_lock held. */ +/* Called with table_lock held. */ int qed_write_l1_table(BDRVQEDState *s, unsigned int index, unsigned int n) { BLKDBG_EVENT(s->bs->file, BLKDBG_L1_UPDATE); @@ -XXX,XX +XXX,XX @@ int qed_write_l1_table_sync(BDRVQEDState *s, unsigned int index, return qed_write_l1_table(s, index, n); } -/* Called either from qed_check or with table_lock held. */ +/* Called with table_lock held. */ int qed_read_l2_table(BDRVQEDState *s, QEDRequest *request, uint64_t offset) { int ret; @@ -XXX,XX +XXX,XX @@ int qed_read_l2_table_sync(BDRVQEDState *s, QEDRequest *request, uint64_t offset return qed_read_l2_table(s, request, offset); } -/* Called either from qed_check or with table_lock held. */ +/* Called with table_lock held. */ int qed_write_l2_table(BDRVQEDState *s, QEDRequest *request, unsigned int index, unsigned int n, bool flush) { 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 bdrv_qed_co_invalidate_cache(BlockDriverState *bs, } } -static int bdrv_qed_check(BlockDriverState *bs, BdrvCheckResult *result, - BdrvCheckMode fix) +static int bdrv_qed_co_check(BlockDriverState *bs, BdrvCheckResult *result, + BdrvCheckMode fix) { BDRVQEDState *s = bs->opaque; + int ret; - return qed_check(s, result, !!fix); + qemu_co_mutex_lock(&s->table_lock); + ret = qed_check(s, result, !!fix); + qemu_co_mutex_unlock(&s->table_lock); + + return ret; } static QemuOptsList qed_create_opts = { @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_qed = { .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, - .bdrv_check = bdrv_qed_check, + .bdrv_co_check = bdrv_qed_co_check, .bdrv_detach_aio_context = bdrv_qed_detach_aio_context, .bdrv_attach_aio_context = bdrv_qed_attach_aio_context, .bdrv_co_drain_begin = bdrv_qed_co_drain_begin, 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 void vdi_header_print(VdiHeader *header) } #endif -static int vdi_check(BlockDriverState *bs, BdrvCheckResult *res, - BdrvCheckMode fix) +static int coroutine_fn vdi_co_check(BlockDriverState *bs, BdrvCheckResult *res, + BdrvCheckMode fix) { /* TODO: additional checks possible. */ BDRVVdiState *s = (BDRVVdiState *)bs->opaque; @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_vdi = { .bdrv_get_info = vdi_get_info, .create_opts = &vdi_create_opts, - .bdrv_check = vdi_check, + .bdrv_co_check = vdi_co_check, }; static void bdrv_vdi_init(void) diff --git a/block/vhdx.c b/block/vhdx.c index XXXXXXX..XXXXXXX 100644 --- a/block/vhdx.c +++ b/block/vhdx.c @@ -XXX,XX +XXX,XX @@ exit: * r/w and any log has already been replayed, so there is nothing (currently) * for us to do here */ -static int vhdx_check(BlockDriverState *bs, BdrvCheckResult *result, - BdrvCheckMode fix) +static int coroutine_fn vhdx_co_check(BlockDriverState *bs, + BdrvCheckResult *result, + BdrvCheckMode fix) { BDRVVHDXState *s = bs->opaque; @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_vhdx = { .bdrv_co_writev = vhdx_co_writev, .bdrv_co_create_opts = vhdx_co_create_opts, .bdrv_get_info = vhdx_get_info, - .bdrv_check = vhdx_check, + .bdrv_co_check = vhdx_co_check, .bdrv_has_zero_init = bdrv_has_zero_init_1, .create_opts = &vhdx_create_opts, 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 ImageInfo *vmdk_get_extent_info(VmdkExtent *extent) return info; } -static int vmdk_check(BlockDriverState *bs, BdrvCheckResult *result, - BdrvCheckMode fix) +static int coroutine_fn vmdk_co_check(BlockDriverState *bs, + BdrvCheckResult *result, + BdrvCheckMode fix) { BDRVVmdkState *s = bs->opaque; VmdkExtent *extent = NULL; @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_vmdk = { .instance_size = sizeof(BDRVVmdkState), .bdrv_probe = vmdk_probe, .bdrv_open = vmdk_open, - .bdrv_check = vmdk_check, + .bdrv_co_check = vmdk_co_check, .bdrv_reopen_prepare = vmdk_reopen_prepare, .bdrv_child_perm = bdrv_format_default_perms, .bdrv_co_preadv = vmdk_co_preadv, -- 2.13.6 From: Alberto Garcia <berto@igalia.com> This function checks that the offset and size of a table are valid. While the offset checks are fine, the size check is too generic, since it only verifies that the total size in bytes fits in a 64-bit integer. In practice all tables used in qcow2 have much smaller size limits, so the size needs to be checked again for each table using its actual limit. This patch generalizes this function by allowing the caller to specify the maximum size for that table. In addition to that it allows passing an Error variable. The function is also renamed and made public since we're going to use it in other parts of the code. Signed-off-by: Alberto Garcia <berto@igalia.com> Reviewed-by: Eric Blake <eblake@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/qcow2.h | 10 +++--- block/qcow2.c | 77 ++++++++++++++++++---------------------------- tests/qemu-iotests/080.out | 16 +++++----- 3 files changed, 43 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 @@ static inline int64_t qcow2_vm_state_offset(BDRVQcow2State *s) return (int64_t)s->l1_vm_state_index << (s->cluster_bits + s->l2_bits); } -static inline uint64_t qcow2_max_refcount_clusters(BDRVQcow2State *s) -{ - return QCOW_MAX_REFTABLE_SIZE >> s->cluster_bits; -} - static inline QCow2ClusterType qcow2_get_cluster_type(uint64_t l2_entry) { if (l2_entry & QCOW_OFLAG_COMPRESSED) { @@ -XXX,XX +XXX,XX @@ void qcow2_signal_corruption(BlockDriverState *bs, bool fatal, int64_t offset, int64_t size, const char *message_format, ...) GCC_FMT_ATTR(5, 6); +int qcow2_validate_table(BlockDriverState *bs, uint64_t offset, + uint64_t entries, size_t entry_len, + int64_t max_size_bytes, const char *table_name, + Error **errp); + /* qcow2-refcount.c functions */ int qcow2_refcount_init(BlockDriverState *bs); void qcow2_refcount_close(BlockDriverState *bs); 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_check(BlockDriverState *bs, return ret; } -static int validate_table_offset(BlockDriverState *bs, uint64_t offset, - uint64_t entries, size_t entry_len) +int qcow2_validate_table(BlockDriverState *bs, uint64_t offset, + uint64_t entries, size_t entry_len, + int64_t max_size_bytes, const char *table_name, + Error **errp) { BDRVQcow2State *s = bs->opaque; - uint64_t size; - /* Use signed INT64_MAX as the maximum even for uint64_t header fields, - * because values will be passed to qemu functions taking int64_t. */ - if (entries > INT64_MAX / entry_len) { - return -EINVAL; - } - - size = entries * entry_len; - - if (INT64_MAX - size < offset) { - return -EINVAL; + if (entries > max_size_bytes / entry_len) { + error_setg(errp, "%s too large", table_name); + return -EFBIG; } - /* Tables must be cluster aligned */ - if (offset_into_cluster(s, offset) != 0) { + /* Use signed INT64_MAX as the maximum even for uint64_t header fields, + * because values will be passed to qemu functions taking int64_t. */ + if ((INT64_MAX - entries * entry_len < offset) || + (offset_into_cluster(s, offset) != 0)) { + error_setg(errp, "%s offset invalid", table_name); return -EINVAL; } @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qcow2_do_open(BlockDriverState *bs, QDict *options, s->refcount_table_size = header.refcount_table_clusters << (s->cluster_bits - 3); - if (header.refcount_table_clusters > qcow2_max_refcount_clusters(s)) { - error_setg(errp, "Reference count table too large"); - ret = -EINVAL; - goto fail; - } - if (header.refcount_table_clusters == 0 && !(flags & BDRV_O_CHECK)) { error_setg(errp, "Image does not contain a reference count table"); ret = -EINVAL; goto fail; } - ret = validate_table_offset(bs, s->refcount_table_offset, - s->refcount_table_size, sizeof(uint64_t)); + ret = qcow2_validate_table(bs, s->refcount_table_offset, + header.refcount_table_clusters, + s->cluster_size, QCOW_MAX_REFTABLE_SIZE, + "Reference count table", errp); if (ret < 0) { - error_setg(errp, "Invalid reference count table offset"); goto fail; } - /* Snapshot table offset/length */ - if (header.nb_snapshots > QCOW_MAX_SNAPSHOTS) { - error_setg(errp, "Too many snapshots"); - ret = -EINVAL; - goto fail; - } - - ret = validate_table_offset(bs, header.snapshots_offset, - header.nb_snapshots, - sizeof(QCowSnapshotHeader)); + /* The total size in bytes of the snapshot table is checked in + * qcow2_read_snapshots() because the size of each snapshot is + * variable and we don't know it yet. + * Here we only check the offset and number of snapshots. */ + ret = qcow2_validate_table(bs, header.snapshots_offset, + header.nb_snapshots, + sizeof(QCowSnapshotHeader), + sizeof(QCowSnapshotHeader) * QCOW_MAX_SNAPSHOTS, + "Snapshot table", errp); if (ret < 0) { - error_setg(errp, "Invalid snapshot table offset"); goto fail; } /* read the level 1 table */ - if (header.l1_size > QCOW_MAX_L1_SIZE / sizeof(uint64_t)) { - error_setg(errp, "Active L1 table too large"); - ret = -EFBIG; + ret = qcow2_validate_table(bs, header.l1_table_offset, + header.l1_size, sizeof(uint64_t), + QCOW_MAX_L1_SIZE, "Active L1 table", errp); + if (ret < 0) { goto fail; } s->l1_size = header.l1_size; + s->l1_table_offset = header.l1_table_offset; l1_vm_state_index = size_to_l1(s, header.size); if (l1_vm_state_index > INT_MAX) { @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qcow2_do_open(BlockDriverState *bs, QDict *options, goto fail; } - ret = validate_table_offset(bs, header.l1_table_offset, - header.l1_size, sizeof(uint64_t)); - if (ret < 0) { - error_setg(errp, "Invalid L1 table offset"); - goto fail; - } - s->l1_table_offset = header.l1_table_offset; - - if (s->l1_size > 0) { s->l1_table = qemu_try_blockalign(bs->file->bs, ROUND_UP(s->l1_size * sizeof(uint64_t), 512)); diff --git a/tests/qemu-iotests/080.out b/tests/qemu-iotests/080.out index XXXXXXX..XXXXXXX 100644 --- a/tests/qemu-iotests/080.out +++ b/tests/qemu-iotests/080.out @@ -XXX,XX +XXX,XX @@ can't open device TEST_DIR/t.qcow2: Reference count table too large == Misaligned refcount table == Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864 -can't open device TEST_DIR/t.qcow2: Invalid reference count table offset +can't open device TEST_DIR/t.qcow2: Reference count table offset invalid == Huge refcount offset == Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864 -can't open device TEST_DIR/t.qcow2: Invalid reference count table offset +can't open device TEST_DIR/t.qcow2: Reference count table offset invalid == Invalid snapshot table == Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864 -can't open device TEST_DIR/t.qcow2: Too many snapshots -can't open device TEST_DIR/t.qcow2: Too many snapshots -can't open device TEST_DIR/t.qcow2: Invalid snapshot table offset -can't open device TEST_DIR/t.qcow2: Invalid snapshot table offset +can't open device TEST_DIR/t.qcow2: Snapshot table too large +can't open device TEST_DIR/t.qcow2: Snapshot table too large +can't open device TEST_DIR/t.qcow2: Snapshot table offset invalid +can't open device TEST_DIR/t.qcow2: Snapshot table offset invalid == Hitting snapshot table size limit == Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864 @@ -XXX,XX +XXX,XX @@ read 512/512 bytes at offset 0 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864 can't open device TEST_DIR/t.qcow2: Active L1 table too large can't open device TEST_DIR/t.qcow2: Active L1 table too large -can't open device TEST_DIR/t.qcow2: Invalid L1 table offset -can't open device TEST_DIR/t.qcow2: Invalid L1 table offset +can't open device TEST_DIR/t.qcow2: Active L1 table offset invalid +can't open device TEST_DIR/t.qcow2: Active L1 table offset invalid == Invalid L1 table (with internal snapshot in the image) == Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864 -- 2.13.6 From: Alberto Garcia <berto@igalia.com> This function checks that the size of a snapshot's L1 table is not too large, but it doesn't validate the offset. We now have a function to take care of this, so let's use it. Signed-off-by: Alberto Garcia <berto@igalia.com> Reviewed-by: Eric Blake <eblake@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/qcow2-snapshot.c | 8 +++++--- tests/qemu-iotests/080 | 10 +++++++++- tests/qemu-iotests/080.out | 8 +++++++- 3 files changed, 21 insertions(+), 5 deletions(-) diff --git a/block/qcow2-snapshot.c b/block/qcow2-snapshot.c index XXXXXXX..XXXXXXX 100644 --- a/block/qcow2-snapshot.c +++ b/block/qcow2-snapshot.c @@ -XXX,XX +XXX,XX @@ int qcow2_snapshot_load_tmp(BlockDriverState *bs, sn = &s->snapshots[snapshot_index]; /* Allocate and read in the snapshot's L1 table */ - if (sn->l1_size > QCOW_MAX_L1_SIZE / sizeof(uint64_t)) { - error_setg(errp, "Snapshot L1 table too large"); - return -EFBIG; + ret = qcow2_validate_table(bs, sn->l1_table_offset, sn->l1_size, + sizeof(uint64_t), QCOW_MAX_L1_SIZE, + "Snapshot L1 table", errp); + if (ret < 0) { + return ret; } new_l1_bytes = sn->l1_size * sizeof(uint64_t); new_l1_table = qemu_try_blockalign(bs->file->bs, diff --git a/tests/qemu-iotests/080 b/tests/qemu-iotests/080 index XXXXXXX..XXXXXXX 100755 --- a/tests/qemu-iotests/080 +++ b/tests/qemu-iotests/080 @@ -XXX,XX +XXX,XX @@ poke_file "$TEST_IMG" "$offset_l2_table_0" "\x80\x00\x00\xff\xff\xff\x00\x00" { $QEMU_IMG snapshot -c test $TEST_IMG; } 2>&1 | _filter_qemu_io | _filter_testdir echo -echo "== Invalid snapshot L1 table ==" +echo "== Invalid snapshot L1 table offset ==" +_make_test_img 64M +{ $QEMU_IO -c "write 0 512" $TEST_IMG; } 2>&1 | _filter_qemu_io | _filter_testdir +{ $QEMU_IMG snapshot -c test $TEST_IMG; } 2>&1 | _filter_testdir +poke_file "$TEST_IMG" "$offset_snap1_l1_offset" "\x00\x00\x00\x00\x00\x40\x02\x00" +{ $QEMU_IMG convert -s test $TEST_IMG $TEST_IMG.snap; } 2>&1 | _filter_testdir + +echo +echo "== Invalid snapshot L1 table size ==" _make_test_img 64M { $QEMU_IO -c "write 0 512" $TEST_IMG; } 2>&1 | _filter_qemu_io | _filter_testdir { $QEMU_IMG snapshot -c test $TEST_IMG; } 2>&1 | _filter_testdir diff --git a/tests/qemu-iotests/080.out b/tests/qemu-iotests/080.out index XXXXXXX..XXXXXXX 100644 --- a/tests/qemu-iotests/080.out +++ b/tests/qemu-iotests/080.out @@ -XXX,XX +XXX,XX @@ wrote 512/512 bytes at offset 0 qemu-img: Could not create snapshot 'test': -27 (File too large) qemu-img: Could not create snapshot 'test': -11 (Resource temporarily unavailable) -== Invalid snapshot L1 table == +== Invalid snapshot L1 table offset == +Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864 +wrote 512/512 bytes at offset 0 +512 bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) +qemu-img: Failed to load snapshot: Snapshot L1 table offset invalid + +== Invalid snapshot L1 table size == Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864 wrote 512/512 bytes at offset 0 512 bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) -- 2.13.6 From: Alberto Garcia <berto@igalia.com> This function iterates over all snapshots of a qcow2 file in order to expand all zero clusters, but it does not validate the snapshots' L1 tables first. We now have a function to take care of this, so let's use it. We can also take the opportunity to replace the sector-based bdrv_read() with bdrv_pread(). Cc: Eric Blake <eblake@redhat.com> Signed-off-by: Alberto Garcia <berto@igalia.com> Reviewed-by: Eric Blake <eblake@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/qcow2-cluster.c | 24 +++++++++++++++++------- tests/qemu-iotests/080 | 2 ++ tests/qemu-iotests/080.out | 4 ++++ 3 files changed, 23 insertions(+), 7 deletions(-) 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 @@ #include "qemu/osdep.h" #include <zlib.h> +#include "qapi/error.h" #include "qemu-common.h" #include "block/block_int.h" #include "block/qcow2.h" @@ -XXX,XX +XXX,XX @@ int qcow2_expand_zero_clusters(BlockDriverState *bs, } for (i = 0; i < s->nb_snapshots; i++) { - int l1_sectors = DIV_ROUND_UP(s->snapshots[i].l1_size * - sizeof(uint64_t), BDRV_SECTOR_SIZE); + int l1_size2; + uint64_t *new_l1_table; + Error *local_err = NULL; + + ret = qcow2_validate_table(bs, s->snapshots[i].l1_table_offset, + s->snapshots[i].l1_size, sizeof(uint64_t), + QCOW_MAX_L1_SIZE, "Snapshot L1 table", + &local_err); + if (ret < 0) { + error_report_err(local_err); + goto fail; + } - uint64_t *new_l1_table = - g_try_realloc(l1_table, l1_sectors * BDRV_SECTOR_SIZE); + l1_size2 = s->snapshots[i].l1_size * sizeof(uint64_t); + new_l1_table = g_try_realloc(l1_table, l1_size2); if (!new_l1_table) { ret = -ENOMEM; @@ -XXX,XX +XXX,XX @@ int qcow2_expand_zero_clusters(BlockDriverState *bs, l1_table = new_l1_table; - ret = bdrv_read(bs->file, - s->snapshots[i].l1_table_offset / BDRV_SECTOR_SIZE, - (void *)l1_table, l1_sectors); + ret = bdrv_pread(bs->file, s->snapshots[i].l1_table_offset, + l1_table, l1_size2); if (ret < 0) { goto fail; } diff --git a/tests/qemu-iotests/080 b/tests/qemu-iotests/080 index XXXXXXX..XXXXXXX 100755 --- a/tests/qemu-iotests/080 +++ b/tests/qemu-iotests/080 @@ -XXX,XX +XXX,XX @@ _make_test_img 64M { $QEMU_IMG snapshot -c test $TEST_IMG; } 2>&1 | _filter_testdir poke_file "$TEST_IMG" "$offset_snap1_l1_offset" "\x00\x00\x00\x00\x00\x40\x02\x00" { $QEMU_IMG convert -s test $TEST_IMG $TEST_IMG.snap; } 2>&1 | _filter_testdir +{ $QEMU_IMG amend -o compat=0.10 $TEST_IMG; } 2>&1 | _filter_testdir echo echo "== Invalid snapshot L1 table size ==" @@ -XXX,XX +XXX,XX @@ _make_test_img 64M { $QEMU_IMG snapshot -c test $TEST_IMG; } 2>&1 | _filter_testdir poke_file "$TEST_IMG" "$offset_snap1_l1_size" "\x10\x00\x00\x00" { $QEMU_IMG convert -s test $TEST_IMG $TEST_IMG.snap; } 2>&1 | _filter_testdir +{ $QEMU_IMG amend -o compat=0.10 $TEST_IMG; } 2>&1 | _filter_testdir # success, all done echo "*** done" diff --git a/tests/qemu-iotests/080.out b/tests/qemu-iotests/080.out index XXXXXXX..XXXXXXX 100644 --- a/tests/qemu-iotests/080.out +++ b/tests/qemu-iotests/080.out @@ -XXX,XX +XXX,XX @@ Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864 wrote 512/512 bytes at offset 0 512 bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) qemu-img: Failed to load snapshot: Snapshot L1 table offset invalid +qemu-img: Snapshot L1 table offset invalid +qemu-img: Error while amending options: Invalid argument == Invalid snapshot L1 table size == Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864 wrote 512/512 bytes at offset 0 512 bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) qemu-img: Failed to load snapshot: Snapshot L1 table too large +qemu-img: Snapshot L1 table too large +qemu-img: Error while amending options: File too large *** done -- 2.13.6 From: Alberto Garcia <berto@igalia.com> The inactive-l2 overlap check iterates uses the L1 tables from all snapshots, but it does not validate them first. We now have a function to take care of this, so let's use it. Signed-off-by: Alberto Garcia <berto@igalia.com> Reviewed-by: Eric Blake <eblake@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/qcow2-refcount.c | 10 +++++++++- tests/qemu-iotests/080 | 4 ++++ tests/qemu-iotests/080.out | 4 ++++ 3 files changed, 17 insertions(+), 1 deletion(-) 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 qcow2_check_metadata_overlap(BlockDriverState *bs, int ign, int64_t offset, uint64_t l1_ofs = s->snapshots[i].l1_table_offset; uint32_t l1_sz = s->snapshots[i].l1_size; uint64_t l1_sz2 = l1_sz * sizeof(uint64_t); - uint64_t *l1 = g_try_malloc(l1_sz2); + uint64_t *l1; int ret; + ret = qcow2_validate_table(bs, l1_ofs, l1_sz, sizeof(uint64_t), + QCOW_MAX_L1_SIZE, "", NULL); + if (ret < 0) { + return ret; + } + + l1 = g_try_malloc(l1_sz2); + if (l1_sz2 && l1 == NULL) { return -ENOMEM; } diff --git a/tests/qemu-iotests/080 b/tests/qemu-iotests/080 index XXXXXXX..XXXXXXX 100755 --- a/tests/qemu-iotests/080 +++ b/tests/qemu-iotests/080 @@ -XXX,XX +XXX,XX @@ _make_test_img 64M poke_file "$TEST_IMG" "$offset_snap1_l1_offset" "\x00\x00\x00\x00\x00\x40\x02\x00" { $QEMU_IMG convert -s test $TEST_IMG $TEST_IMG.snap; } 2>&1 | _filter_testdir { $QEMU_IMG amend -o compat=0.10 $TEST_IMG; } 2>&1 | _filter_testdir +{ $QEMU_IO -c "open -o overlap-check.inactive-l2=on $TEST_IMG" \ + -c 'write 0 4k'; } 2>&1 | _filter_qemu_io | _filter_testdir echo echo "== Invalid snapshot L1 table size ==" @@ -XXX,XX +XXX,XX @@ _make_test_img 64M poke_file "$TEST_IMG" "$offset_snap1_l1_size" "\x10\x00\x00\x00" { $QEMU_IMG convert -s test $TEST_IMG $TEST_IMG.snap; } 2>&1 | _filter_testdir { $QEMU_IMG amend -o compat=0.10 $TEST_IMG; } 2>&1 | _filter_testdir +{ $QEMU_IO -c "open -o overlap-check.inactive-l2=on $TEST_IMG" \ + -c 'write 0 4k'; } 2>&1 | _filter_qemu_io | _filter_testdir # success, all done echo "*** done" diff --git a/tests/qemu-iotests/080.out b/tests/qemu-iotests/080.out index XXXXXXX..XXXXXXX 100644 --- a/tests/qemu-iotests/080.out +++ b/tests/qemu-iotests/080.out @@ -XXX,XX +XXX,XX @@ wrote 512/512 bytes at offset 0 qemu-img: Failed to load snapshot: Snapshot L1 table offset invalid qemu-img: Snapshot L1 table offset invalid qemu-img: Error while amending options: Invalid argument +Failed to flush the refcount block cache: Invalid argument +write failed: Invalid argument == Invalid snapshot L1 table size == Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864 @@ -XXX,XX +XXX,XX @@ wrote 512/512 bytes at offset 0 qemu-img: Failed to load snapshot: Snapshot L1 table too large qemu-img: Snapshot L1 table too large qemu-img: Error while amending options: File too large +Failed to flush the refcount block cache: File too large +write failed: File too large *** done -- 2.13.6 From: Alberto Garcia <berto@igalia.com> This function copies a snapshot's L1 table into the active one without validating it first. We now have a function to take care of this, so let's use it. Cc: Eric Blake <eblake@redhat.com> Signed-off-by: Alberto Garcia <berto@igalia.com> Reviewed-by: Eric Blake <eblake@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/qcow2-snapshot.c | 9 +++++++++ tests/qemu-iotests/080 | 2 ++ tests/qemu-iotests/080.out | 4 ++++ 3 files changed, 15 insertions(+) diff --git a/block/qcow2-snapshot.c b/block/qcow2-snapshot.c index XXXXXXX..XXXXXXX 100644 --- a/block/qcow2-snapshot.c +++ b/block/qcow2-snapshot.c @@ -XXX,XX +XXX,XX @@ int qcow2_snapshot_goto(BlockDriverState *bs, const char *snapshot_id) { BDRVQcow2State *s = bs->opaque; QCowSnapshot *sn; + Error *local_err = NULL; int i, snapshot_index; int cur_l1_bytes, sn_l1_bytes; int ret; @@ -XXX,XX +XXX,XX @@ int qcow2_snapshot_goto(BlockDriverState *bs, const char *snapshot_id) } sn = &s->snapshots[snapshot_index]; + ret = qcow2_validate_table(bs, sn->l1_table_offset, sn->l1_size, + sizeof(uint64_t), QCOW_MAX_L1_SIZE, + "Snapshot L1 table", &local_err); + if (ret < 0) { + error_report_err(local_err); + goto fail; + } + if (sn->disk_size != bs->total_sectors * BDRV_SECTOR_SIZE) { error_report("qcow2: Loading snapshots with different disk " "size is not implemented"); diff --git a/tests/qemu-iotests/080 b/tests/qemu-iotests/080 index XXXXXXX..XXXXXXX 100755 --- a/tests/qemu-iotests/080 +++ b/tests/qemu-iotests/080 @@ -XXX,XX +XXX,XX @@ poke_file "$TEST_IMG" "$offset_snap1_l1_offset" "\x00\x00\x00\x00\x00\x40\x02\x0 { $QEMU_IMG amend -o compat=0.10 $TEST_IMG; } 2>&1 | _filter_testdir { $QEMU_IO -c "open -o overlap-check.inactive-l2=on $TEST_IMG" \ -c 'write 0 4k'; } 2>&1 | _filter_qemu_io | _filter_testdir +{ $QEMU_IMG snapshot -a test $TEST_IMG; } 2>&1 | _filter_testdir echo echo "== Invalid snapshot L1 table size ==" @@ -XXX,XX +XXX,XX @@ poke_file "$TEST_IMG" "$offset_snap1_l1_size" "\x10\x00\x00\x00" { $QEMU_IMG amend -o compat=0.10 $TEST_IMG; } 2>&1 | _filter_testdir { $QEMU_IO -c "open -o overlap-check.inactive-l2=on $TEST_IMG" \ -c 'write 0 4k'; } 2>&1 | _filter_qemu_io | _filter_testdir +{ $QEMU_IMG snapshot -a test $TEST_IMG; } 2>&1 | _filter_testdir # success, all done echo "*** done" diff --git a/tests/qemu-iotests/080.out b/tests/qemu-iotests/080.out index XXXXXXX..XXXXXXX 100644 --- a/tests/qemu-iotests/080.out +++ b/tests/qemu-iotests/080.out @@ -XXX,XX +XXX,XX @@ qemu-img: Snapshot L1 table offset invalid qemu-img: Error while amending options: Invalid argument Failed to flush the refcount block cache: Invalid argument write failed: Invalid argument +qemu-img: Snapshot L1 table offset invalid +qemu-img: Could not apply snapshot 'test': Failed to load snapshot: Invalid argument == Invalid snapshot L1 table size == Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864 @@ -XXX,XX +XXX,XX @@ qemu-img: Snapshot L1 table too large qemu-img: Error while amending options: File too large Failed to flush the refcount block cache: File too large write failed: File too large +qemu-img: Snapshot L1 table too large +qemu-img: Could not apply snapshot 'test': Failed to load snapshot: File too large *** done -- 2.13.6 From: Alberto Garcia <berto@igalia.com> This function deletes a snapshot from disk, removing its entry from the snapshot table, freeing its L1 table and decreasing the refcounts of all clusters. The L1 table offset and size are however not validated. If we use invalid values in this function we'll probably corrupt the image even more, so we should return an error instead. We now have a function to take care of this, so let's use it. Signed-off-by: Alberto Garcia <berto@igalia.com> Reviewed-by: Eric Blake <eblake@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/qcow2-snapshot.c | 7 +++++++ tests/qemu-iotests/080 | 2 ++ tests/qemu-iotests/080.out | 2 ++ 3 files changed, 11 insertions(+) diff --git a/block/qcow2-snapshot.c b/block/qcow2-snapshot.c index XXXXXXX..XXXXXXX 100644 --- a/block/qcow2-snapshot.c +++ b/block/qcow2-snapshot.c @@ -XXX,XX +XXX,XX @@ int qcow2_snapshot_delete(BlockDriverState *bs, } sn = s->snapshots[snapshot_index]; + ret = qcow2_validate_table(bs, sn.l1_table_offset, sn.l1_size, + sizeof(uint64_t), QCOW_MAX_L1_SIZE, + "Snapshot L1 table", errp); + if (ret < 0) { + return ret; + } + /* Remove it from the snapshot list */ memmove(s->snapshots + snapshot_index, s->snapshots + snapshot_index + 1, diff --git a/tests/qemu-iotests/080 b/tests/qemu-iotests/080 index XXXXXXX..XXXXXXX 100755 --- a/tests/qemu-iotests/080 +++ b/tests/qemu-iotests/080 @@ -XXX,XX +XXX,XX @@ poke_file "$TEST_IMG" "$offset_snap1_l1_offset" "\x00\x00\x00\x00\x00\x40\x02\x0 { $QEMU_IO -c "open -o overlap-check.inactive-l2=on $TEST_IMG" \ -c 'write 0 4k'; } 2>&1 | _filter_qemu_io | _filter_testdir { $QEMU_IMG snapshot -a test $TEST_IMG; } 2>&1 | _filter_testdir +{ $QEMU_IMG snapshot -d test $TEST_IMG; } 2>&1 | _filter_testdir echo echo "== Invalid snapshot L1 table size ==" @@ -XXX,XX +XXX,XX @@ poke_file "$TEST_IMG" "$offset_snap1_l1_size" "\x10\x00\x00\x00" { $QEMU_IO -c "open -o overlap-check.inactive-l2=on $TEST_IMG" \ -c 'write 0 4k'; } 2>&1 | _filter_qemu_io | _filter_testdir { $QEMU_IMG snapshot -a test $TEST_IMG; } 2>&1 | _filter_testdir +{ $QEMU_IMG snapshot -d test $TEST_IMG; } 2>&1 | _filter_testdir # success, all done echo "*** done" diff --git a/tests/qemu-iotests/080.out b/tests/qemu-iotests/080.out index XXXXXXX..XXXXXXX 100644 --- a/tests/qemu-iotests/080.out +++ b/tests/qemu-iotests/080.out @@ -XXX,XX +XXX,XX @@ Failed to flush the refcount block cache: Invalid argument write failed: Invalid argument qemu-img: Snapshot L1 table offset invalid qemu-img: Could not apply snapshot 'test': Failed to load snapshot: Invalid argument +qemu-img: Could not delete snapshot 'test': Snapshot L1 table offset invalid == Invalid snapshot L1 table size == Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864 @@ -XXX,XX +XXX,XX @@ Failed to flush the refcount block cache: File too large write failed: File too large qemu-img: Snapshot L1 table too large qemu-img: Could not apply snapshot 'test': Failed to load snapshot: File too large +qemu-img: Could not delete snapshot 'test': Snapshot L1 table too large *** done -- 2.13.6 From: Alberto Garcia <berto@igalia.com> 'qemu-img check' cannot detect if a snapshot's L1 table is corrupted. This patch checks the table's offset and size and reports corruption if the values are not valid. This patch doesn't add code to fix that corruption yet, only to detect and report it. Signed-off-by: Alberto Garcia <berto@igalia.com> Reviewed-by: Eric Blake <eblake@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/qcow2-refcount.c | 14 ++++++++++++++ tests/qemu-iotests/080 | 2 ++ tests/qemu-iotests/080.out | 20 ++++++++++++++++++++ 3 files changed, 36 insertions(+) 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 @@ static int calculate_refcounts(BlockDriverState *bs, BdrvCheckResult *res, /* snapshots */ for (i = 0; i < s->nb_snapshots; i++) { sn = s->snapshots + i; + if (offset_into_cluster(s, sn->l1_table_offset)) { + fprintf(stderr, "ERROR snapshot %s (%s) l1_offset=%#" PRIx64 ": " + "L1 table is not cluster aligned; snapshot table entry " + "corrupted\n", sn->id_str, sn->name, sn->l1_table_offset); + res->corruptions++; + continue; + } + if (sn->l1_size > QCOW_MAX_L1_SIZE / sizeof(uint64_t)) { + fprintf(stderr, "ERROR snapshot %s (%s) l1_size=%#" PRIx32 ": " + "L1 table is too large; snapshot table entry corrupted\n", + sn->id_str, sn->name, sn->l1_size); + res->corruptions++; + continue; + } ret = check_refcounts_l1(bs, res, refcount_table, nb_clusters, sn->l1_table_offset, sn->l1_size, 0, fix); if (ret < 0) { diff --git a/tests/qemu-iotests/080 b/tests/qemu-iotests/080 index XXXXXXX..XXXXXXX 100755 --- a/tests/qemu-iotests/080 +++ b/tests/qemu-iotests/080 @@ -XXX,XX +XXX,XX @@ poke_file "$TEST_IMG" "$offset_snap1_l1_offset" "\x00\x00\x00\x00\x00\x40\x02\x0 -c 'write 0 4k'; } 2>&1 | _filter_qemu_io | _filter_testdir { $QEMU_IMG snapshot -a test $TEST_IMG; } 2>&1 | _filter_testdir { $QEMU_IMG snapshot -d test $TEST_IMG; } 2>&1 | _filter_testdir +_check_test_img echo echo "== Invalid snapshot L1 table size ==" @@ -XXX,XX +XXX,XX @@ poke_file "$TEST_IMG" "$offset_snap1_l1_size" "\x10\x00\x00\x00" -c 'write 0 4k'; } 2>&1 | _filter_qemu_io | _filter_testdir { $QEMU_IMG snapshot -a test $TEST_IMG; } 2>&1 | _filter_testdir { $QEMU_IMG snapshot -d test $TEST_IMG; } 2>&1 | _filter_testdir +_check_test_img # success, all done echo "*** done" diff --git a/tests/qemu-iotests/080.out b/tests/qemu-iotests/080.out index XXXXXXX..XXXXXXX 100644 --- a/tests/qemu-iotests/080.out +++ b/tests/qemu-iotests/080.out @@ -XXX,XX +XXX,XX @@ write failed: Invalid argument qemu-img: Snapshot L1 table offset invalid qemu-img: Could not apply snapshot 'test': Failed to load snapshot: Invalid argument qemu-img: Could not delete snapshot 'test': Snapshot L1 table offset invalid +ERROR snapshot 1 (test) l1_offset=0x400200: L1 table is not cluster aligned; snapshot table entry corrupted +Leaked cluster 4 refcount=2 reference=1 +Leaked cluster 5 refcount=2 reference=1 +Leaked cluster 6 refcount=1 reference=0 + +1 errors were found on the image. +Data may be corrupted, or further writes to the image may corrupt it. + +3 leaked clusters were found on the image. +This means waste of disk space, but no harm to data. == Invalid snapshot L1 table size == Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864 @@ -XXX,XX +XXX,XX @@ write failed: File too large qemu-img: Snapshot L1 table too large qemu-img: Could not apply snapshot 'test': Failed to load snapshot: File too large qemu-img: Could not delete snapshot 'test': Snapshot L1 table too large +ERROR snapshot 1 (test) l1_size=0x10000000: L1 table is too large; snapshot table entry corrupted +Leaked cluster 4 refcount=2 reference=1 +Leaked cluster 5 refcount=2 reference=1 +Leaked cluster 6 refcount=1 reference=0 + +1 errors were found on the image. +Data may be corrupted, or further writes to the image may corrupt it. + +3 leaked clusters were found on the image. +This means waste of disk space, but no harm to data. *** done -- 2.13.6 This creates a BlockdevCreateOptions union type that will contain all of the options for image creation. We'll start out with an empty struct type BlockdevCreateNotSupported for all drivers. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> --- qapi/block-core.json | 62 ++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 62 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 @@ { 'command': 'blockdev-del', 'data': { 'node-name': 'str' } } ## +# @BlockdevCreateNotSupported: +# +# This is used for all drivers that don't support creating images. +# +# Since: 2.12 +## +{ 'struct': 'BlockdevCreateNotSupported', 'data': {}} + +## +# @BlockdevCreateOptions: +# +# Options for creating an image format on a given node. +# +# @driver block driver to create the image format +# +# Since: 2.12 +## +{ 'union': 'BlockdevCreateOptions', + 'base': { + 'driver': 'BlockdevDriver' }, + 'discriminator': 'driver', + 'data': { + 'blkdebug': 'BlockdevCreateNotSupported', + 'blkverify': 'BlockdevCreateNotSupported', + 'bochs': 'BlockdevCreateNotSupported', + 'cloop': 'BlockdevCreateNotSupported', + 'dmg': 'BlockdevCreateNotSupported', + 'file': 'BlockdevCreateNotSupported', + 'ftp': 'BlockdevCreateNotSupported', + 'ftps': 'BlockdevCreateNotSupported', + 'gluster': 'BlockdevCreateNotSupported', + 'host_cdrom': 'BlockdevCreateNotSupported', + 'host_device': 'BlockdevCreateNotSupported', + 'http': 'BlockdevCreateNotSupported', + 'https': 'BlockdevCreateNotSupported', + 'iscsi': 'BlockdevCreateNotSupported', + 'luks': 'BlockdevCreateNotSupported', + 'nbd': 'BlockdevCreateNotSupported', + 'nfs': 'BlockdevCreateNotSupported', + 'null-aio': 'BlockdevCreateNotSupported', + 'null-co': 'BlockdevCreateNotSupported', + 'nvme': 'BlockdevCreateNotSupported', + 'parallels': 'BlockdevCreateNotSupported', + 'qcow2': 'BlockdevCreateNotSupported', + 'qcow': 'BlockdevCreateNotSupported', + 'qed': 'BlockdevCreateNotSupported', + 'quorum': 'BlockdevCreateNotSupported', + 'raw': 'BlockdevCreateNotSupported', + 'rbd': 'BlockdevCreateNotSupported', + 'replication': 'BlockdevCreateNotSupported', + 'sheepdog': 'BlockdevCreateNotSupported', + 'ssh': 'BlockdevCreateNotSupported', + 'throttle': 'BlockdevCreateNotSupported', + 'vdi': 'BlockdevCreateNotSupported', + 'vhdx': 'BlockdevCreateNotSupported', + 'vmdk': 'BlockdevCreateNotSupported', + 'vpc': 'BlockdevCreateNotSupported', + 'vvfat': 'BlockdevCreateNotSupported', + 'vxhs': 'BlockdevCreateNotSupported' + } } + +## # @blockdev-open-tray: # # Opens a block device's tray. If there is a block driver state tree inserted as -- 2.13.6 Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> --- qapi/block-core.json | 45 ++++++++++++++++++++++++++++++++++++++++++++- 1 file changed, 44 insertions(+), 1 deletion(-) 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 @@ { 'command': 'blockdev-del', 'data': { 'node-name': 'str' } } ## +# @BlockdevQcow2Version: +# +# @v2: The original QCOW2 format as introduced in qemu 0.10 (version 2) +# @v3: The extended QCOW2 format as introduced in qemu 1.1 (version 3) +# +# Since: 2.12 +## +{ 'enum': 'BlockdevQcow2Version', + 'data': [ 'v2', 'v3' ] } + + +## +# @BlockdevCreateOptionsQcow2: +# +# Driver specific image creation options for qcow2. +# +# @file Node to create the image format on +# @size Size of the virtual disk in bytes +# @version Compatibility level (default: v3) +# @backing-file File name of the backing file if a backing file +# should be used +# @backing-fmt Name of the block driver to use for the backing file +# @encrypt Encryption options if the image should be encrypted +# @cluster-size qcow2 cluster size in bytes (default: 65536) +# @preallocation Preallocation mode for the new image (default: off) +# @lazy-refcounts True if refcounts may be updated lazily (default: off) +# @refcount-bits Width of reference counts in bits (default: 16) +# +# Since: 2.12 +## +{ 'struct': 'BlockdevCreateOptionsQcow2', + 'data': { 'file': 'BlockdevRef', + 'size': 'size', + '*version': 'BlockdevQcow2Version', + '*backing-file': 'str', + '*backing-fmt': 'BlockdevDriver', + '*encrypt': 'QCryptoBlockCreateOptions', + '*cluster-size': 'size', + '*preallocation': 'PreallocMode', + '*lazy-refcounts': 'bool', + '*refcount-bits': 'int' } } + +## # @BlockdevCreateNotSupported: # # This is used for all drivers that don't support creating images. @@ -XXX,XX +XXX,XX @@ 'null-co': 'BlockdevCreateNotSupported', 'nvme': 'BlockdevCreateNotSupported', 'parallels': 'BlockdevCreateNotSupported', - 'qcow2': 'BlockdevCreateNotSupported', + 'qcow2': 'BlockdevCreateOptionsQcow2', 'qcow': 'BlockdevCreateNotSupported', 'qed': 'BlockdevCreateNotSupported', 'quorum': 'BlockdevCreateNotSupported', -- 2.13.6 The functions originally known as qcow2_create() and qcow2_create2() are now called qcow2_co_create_opts() and qcow2_co_create(), which matches the names of the BlockDriver callbacks that they will implement at the end of this patch series. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> --- block/qcow2.c | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) 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 uint64_t qcow2_opt_get_refcount_bits_del(QemuOpts *opts, int version, } static int coroutine_fn -qcow2_co_create2(const char *filename, int64_t total_size, - const char *backing_file, const char *backing_format, - int flags, size_t cluster_size, PreallocMode prealloc, - QemuOpts *opts, int version, int refcount_order, - const char *encryptfmt, Error **errp) +qcow2_co_create(const char *filename, int64_t total_size, + const char *backing_file, const char *backing_format, + int flags, size_t cluster_size, PreallocMode prealloc, + QemuOpts *opts, int version, int refcount_order, + const char *encryptfmt, Error **errp) { QDict *options; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qcow2_co_create_opts(const char *filename, QemuOpts *opt refcount_order = ctz32(refcount_bits); - ret = qcow2_co_create2(filename, size, backing_file, backing_fmt, flags, - cluster_size, prealloc, opts, version, refcount_order, - encryptfmt, &local_err); + ret = qcow2_co_create(filename, size, backing_file, backing_fmt, flags, + cluster_size, prealloc, opts, version, refcount_order, + encryptfmt, &local_err); error_propagate(errp, local_err); finish: -- 2.13.6 Currently, qcow2_create() only parses the QemuOpts and then calls qcow2_co_create() for the actual image creation, which includes both the creation of the actual file on the file system and writing a valid empty qcow2 image into that file. The plan is that qcow2_co_create() becomes the function that implements the functionality for a future 'blockdev-create' QMP command, which only creates the qcow2 layer on an already opened file node. This is a first step towards that goal: Let's move out anything that deals with the protocol layer from qcow2_co_create() into qcow2_create(). This means that qcow2_co_create() doesn't need a file name any more. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> --- block/qcow2.c | 64 +++++++++++++++++++++++++++++++++++------------------------ 1 file changed, 38 insertions(+), 26 deletions(-) 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 uint64_t qcow2_opt_get_refcount_bits_del(QemuOpts *opts, int version, } static int coroutine_fn -qcow2_co_create(const char *filename, int64_t total_size, +qcow2_co_create(BlockDriverState *bs, int64_t total_size, const char *backing_file, const char *backing_format, int flags, size_t cluster_size, PreallocMode prealloc, QemuOpts *opts, int version, int refcount_order, @@ -XXX,XX +XXX,XX @@ qcow2_co_create(const char *filename, int64_t total_size, Error *local_err = NULL; int ret; - if (prealloc == PREALLOC_MODE_FULL || prealloc == PREALLOC_MODE_FALLOC) { - int64_t prealloc_size = - qcow2_calc_prealloc_size(total_size, cluster_size, refcount_order); - qemu_opt_set_number(opts, BLOCK_OPT_SIZE, prealloc_size, &error_abort); - qemu_opt_set(opts, BLOCK_OPT_PREALLOC, PreallocMode_str(prealloc), - &error_abort); - } - - ret = bdrv_create_file(filename, opts, &local_err); + blk = blk_new(BLK_PERM_WRITE | BLK_PERM_RESIZE, BLK_PERM_ALL); + ret = blk_insert_bs(blk, bs, errp); if (ret < 0) { - error_propagate(errp, local_err); - return ret; - } - - blk = blk_new_open(filename, NULL, NULL, - BDRV_O_RDWR | BDRV_O_RESIZE | BDRV_O_PROTOCOL, - &local_err); - if (blk == NULL) { - error_propagate(errp, local_err); - return -EIO; + goto out; } - blk_set_allow_write_beyond_eof(blk, true); /* Write the header */ @@ -XXX,XX +XXX,XX @@ qcow2_co_create(const char *filename, int64_t total_size, */ options = qdict_new(); qdict_put_str(options, "driver", "qcow2"); - blk = blk_new_open(filename, NULL, options, + qdict_put_str(options, "file", bs->node_name); + blk = blk_new_open(NULL, NULL, options, BDRV_O_RDWR | BDRV_O_RESIZE | BDRV_O_NO_FLUSH, &local_err); if (blk == NULL) { @@ -XXX,XX +XXX,XX @@ qcow2_co_create(const char *filename, int64_t total_size, */ options = qdict_new(); qdict_put_str(options, "driver", "qcow2"); - blk = blk_new_open(filename, NULL, options, + qdict_put_str(options, "file", bs->node_name); + blk = blk_new_open(NULL, NULL, options, BDRV_O_RDWR | BDRV_O_NO_BACKING | BDRV_O_NO_IO, &local_err); if (blk == NULL) { @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qcow2_co_create_opts(const char *filename, QemuOpts *opt uint64_t refcount_bits; int refcount_order; char *encryptfmt = NULL; + BlockDriverState *bs = NULL; Error *local_err = NULL; int ret; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qcow2_co_create_opts(const char *filename, QemuOpts *opt refcount_order = ctz32(refcount_bits); - ret = qcow2_co_create(filename, size, backing_file, backing_fmt, flags, + /* Create and open the file (protocol layer) */ + if (prealloc == PREALLOC_MODE_FULL || prealloc == PREALLOC_MODE_FALLOC) { + int64_t prealloc_size = + qcow2_calc_prealloc_size(size, cluster_size, refcount_order); + qemu_opt_set_number(opts, BLOCK_OPT_SIZE, prealloc_size, &error_abort); + qemu_opt_set(opts, BLOCK_OPT_PREALLOC, PreallocMode_str(prealloc), + &error_abort); + } + + ret = bdrv_create_file(filename, opts, errp); + if (ret < 0) { + goto finish; + } + + bs = bdrv_open(filename, NULL, NULL, + BDRV_O_RDWR | BDRV_O_RESIZE | BDRV_O_PROTOCOL, errp); + if (bs == NULL) { + ret = -EIO; + goto finish; + } + + /* Create the qcow2 image (format layer) */ + ret = qcow2_co_create(bs, size, backing_file, backing_fmt, flags, cluster_size, prealloc, opts, version, refcount_order, - encryptfmt, &local_err); - error_propagate(errp, local_err); + encryptfmt, errp); + if (ret < 0) { + goto finish; + } finish: + bdrv_unref(bs); + g_free(backing_file); g_free(backing_fmt); g_free(encryptfmt); -- 2.13.6 All of the simple options are now passed to qcow2_co_create() in a BlockdevCreateOptions object. Still missing: node-name and the encryption options. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> --- block/qcow2.c | 189 ++++++++++++++++++++++++++++++++++++++++++++++------------ 1 file changed, 151 insertions(+), 38 deletions(-) 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_calc_prealloc_size(int64_t total_size, return meta_size + aligned_total_size; } -static size_t qcow2_opt_get_cluster_size_del(QemuOpts *opts, Error **errp) +static bool validate_cluster_size(size_t cluster_size, Error **errp) { - size_t cluster_size; - int cluster_bits; - - cluster_size = qemu_opt_get_size_del(opts, BLOCK_OPT_CLUSTER_SIZE, - DEFAULT_CLUSTER_SIZE); - cluster_bits = ctz32(cluster_size); + int cluster_bits = ctz32(cluster_size); if (cluster_bits < MIN_CLUSTER_BITS || cluster_bits > MAX_CLUSTER_BITS || (1 << cluster_bits) != cluster_size) { error_setg(errp, "Cluster size must be a power of two between %d and " "%dk", 1 << MIN_CLUSTER_BITS, 1 << (MAX_CLUSTER_BITS - 10)); + return false; + } + return true; +} + +static size_t qcow2_opt_get_cluster_size_del(QemuOpts *opts, Error **errp) +{ + size_t cluster_size; + + cluster_size = qemu_opt_get_size_del(opts, BLOCK_OPT_CLUSTER_SIZE, + DEFAULT_CLUSTER_SIZE); + if (!validate_cluster_size(cluster_size, errp)) { return 0; } return cluster_size; @@ -XXX,XX +XXX,XX @@ static uint64_t qcow2_opt_get_refcount_bits_del(QemuOpts *opts, int version, } static int coroutine_fn -qcow2_co_create(BlockDriverState *bs, int64_t total_size, - const char *backing_file, const char *backing_format, - int flags, size_t cluster_size, PreallocMode prealloc, - QemuOpts *opts, int version, int refcount_order, - const char *encryptfmt, Error **errp) +qcow2_co_create(BlockDriverState *bs, BlockdevCreateOptions *create_options, + QemuOpts *opts, const char *encryptfmt, Error **errp) { + BlockdevCreateOptionsQcow2 *qcow2_opts; QDict *options; /* @@ -XXX,XX +XXX,XX @@ qcow2_co_create(BlockDriverState *bs, int64_t total_size, */ BlockBackend *blk; QCowHeader *header; + size_t cluster_size; + int version; + int refcount_order; uint64_t* refcount_table; Error *local_err = NULL; int ret; + /* Validate options and set default values */ + assert(create_options->driver == BLOCKDEV_DRIVER_QCOW2); + qcow2_opts = &create_options->u.qcow2; + + if (!QEMU_IS_ALIGNED(qcow2_opts->size, BDRV_SECTOR_SIZE)) { + error_setg(errp, "Image size must be a multiple of 512 bytes"); + ret = -EINVAL; + goto out; + } + + if (qcow2_opts->has_version) { + switch (qcow2_opts->version) { + case BLOCKDEV_QCOW2_VERSION_V2: + version = 2; + break; + case BLOCKDEV_QCOW2_VERSION_V3: + version = 3; + break; + default: + g_assert_not_reached(); + } + } else { + version = 3; + } + + if (qcow2_opts->has_cluster_size) { + cluster_size = qcow2_opts->cluster_size; + } else { + cluster_size = DEFAULT_CLUSTER_SIZE; + } + + if (!validate_cluster_size(cluster_size, errp)) { + return -EINVAL; + } + + if (!qcow2_opts->has_preallocation) { + qcow2_opts->preallocation = PREALLOC_MODE_OFF; + } + if (qcow2_opts->has_backing_file && + qcow2_opts->preallocation != PREALLOC_MODE_OFF) + { + error_setg(errp, "Backing file and preallocation cannot be used at " + "the same time"); + return -EINVAL; + } + if (qcow2_opts->has_backing_fmt && !qcow2_opts->has_backing_file) { + error_setg(errp, "Backing format cannot be used without backing file"); + return -EINVAL; + } + + if (!qcow2_opts->has_lazy_refcounts) { + qcow2_opts->lazy_refcounts = false; + } + if (version < 3 && qcow2_opts->lazy_refcounts) { + error_setg(errp, "Lazy refcounts only supported with compatibility " + "level 1.1 and above (use compat=1.1 or greater)"); + return -EINVAL; + } + + if (!qcow2_opts->has_refcount_bits) { + qcow2_opts->refcount_bits = 16; + } + if (qcow2_opts->refcount_bits > 64 || + !is_power_of_2(qcow2_opts->refcount_bits)) + { + error_setg(errp, "Refcount width must be a power of two and may not " + "exceed 64 bits"); + return -EINVAL; + } + if (version < 3 && qcow2_opts->refcount_bits != 16) { + error_setg(errp, "Different refcount widths than 16 bits require " + "compatibility level 1.1 or above (use compat=1.1 or " + "greater)"); + return -EINVAL; + } + refcount_order = ctz32(qcow2_opts->refcount_bits); + + + /* Create BlockBackend to write to the image */ blk = blk_new(BLK_PERM_WRITE | BLK_PERM_RESIZE, BLK_PERM_ALL); ret = blk_insert_bs(blk, bs, errp); if (ret < 0) { @@ -XXX,XX +XXX,XX @@ qcow2_co_create(BlockDriverState *bs, int64_t total_size, /* We'll update this to correct value later */ header->crypt_method = cpu_to_be32(QCOW_CRYPT_NONE); - if (flags & BLOCK_FLAG_LAZY_REFCOUNTS) { + if (qcow2_opts->lazy_refcounts) { header->compatible_features |= cpu_to_be64(QCOW2_COMPAT_LAZY_REFCOUNTS); } @@ -XXX,XX +XXX,XX @@ qcow2_co_create(BlockDriverState *bs, int64_t total_size, } /* Okay, now that we have a valid image, let's give it the right size */ - ret = blk_truncate(blk, total_size, PREALLOC_MODE_OFF, errp); + ret = blk_truncate(blk, qcow2_opts->size, PREALLOC_MODE_OFF, errp); if (ret < 0) { error_prepend(errp, "Could not resize image: "); goto out; } /* Want a backing file? There you go.*/ - if (backing_file) { - ret = bdrv_change_backing_file(blk_bs(blk), backing_file, backing_format); + if (qcow2_opts->has_backing_file) { + const char *backing_format = NULL; + + if (qcow2_opts->has_backing_fmt) { + backing_format = BlockdevDriver_str(qcow2_opts->backing_fmt); + } + + ret = bdrv_change_backing_file(blk_bs(blk), qcow2_opts->backing_file, + backing_format); if (ret < 0) { error_setg_errno(errp, -ret, "Could not assign backing file '%s' " - "with format '%s'", backing_file, backing_format); + "with format '%s'", qcow2_opts->backing_file, + backing_format); goto out; } } @@ -XXX,XX +XXX,XX @@ qcow2_co_create(BlockDriverState *bs, int64_t total_size, } /* And if we're supposed to preallocate metadata, do that now */ - if (prealloc != PREALLOC_MODE_OFF) { - ret = preallocate(blk_bs(blk), 0, total_size); + if (qcow2_opts->preallocation != PREALLOC_MODE_OFF) { + ret = preallocate(blk_bs(blk), 0, qcow2_opts->size); if (ret < 0) { error_setg_errno(errp, -ret, "Could not preallocate metadata"); goto out; @@ -XXX,XX +XXX,XX @@ out: static int coroutine_fn qcow2_co_create_opts(const char *filename, QemuOpts *opts, Error **errp) { + BlockdevCreateOptions create_options; char *backing_file = NULL; char *backing_fmt = NULL; + BlockdevDriver backing_drv; char *buf = NULL; uint64_t size = 0; int flags = 0; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qcow2_co_create_opts(const char *filename, QemuOpts *opt PreallocMode prealloc; int version; uint64_t refcount_bits; - int refcount_order; char *encryptfmt = NULL; BlockDriverState *bs = NULL; Error *local_err = NULL; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qcow2_co_create_opts(const char *filename, QemuOpts *opt BDRV_SECTOR_SIZE); backing_file = qemu_opt_get_del(opts, BLOCK_OPT_BACKING_FILE); backing_fmt = qemu_opt_get_del(opts, BLOCK_OPT_BACKING_FMT); + backing_drv = qapi_enum_parse(&BlockdevDriver_lookup, backing_fmt, + 0, &local_err); + if (local_err) { + error_propagate(errp, local_err); + ret = -EINVAL; + goto finish; + } encryptfmt = qemu_opt_get_del(opts, BLOCK_OPT_ENCRYPT_FORMAT); if (encryptfmt) { if (qemu_opt_get(opts, BLOCK_OPT_ENCRYPT)) { @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qcow2_co_create_opts(const char *filename, QemuOpts *opt flags |= BLOCK_FLAG_LAZY_REFCOUNTS; } - if (backing_file && prealloc != PREALLOC_MODE_OFF) { - error_setg(errp, "Backing file and preallocation cannot be used at " - "the same time"); - ret = -EINVAL; - goto finish; - } - - if (version < 3 && (flags & BLOCK_FLAG_LAZY_REFCOUNTS)) { - error_setg(errp, "Lazy refcounts only supported with compatibility " - "level 1.1 and above (use compat=1.1 or greater)"); - ret = -EINVAL; - goto finish; - } - refcount_bits = qcow2_opt_get_refcount_bits_del(opts, version, &local_err); if (local_err) { error_propagate(errp, local_err); @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qcow2_co_create_opts(const char *filename, QemuOpts *opt goto finish; } - refcount_order = ctz32(refcount_bits); /* Create and open the file (protocol layer) */ if (prealloc == PREALLOC_MODE_FULL || prealloc == PREALLOC_MODE_FALLOC) { + int refcount_order = ctz32(refcount_bits); int64_t prealloc_size = qcow2_calc_prealloc_size(size, cluster_size, refcount_order); qemu_opt_set_number(opts, BLOCK_OPT_SIZE, prealloc_size, &error_abort); @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qcow2_co_create_opts(const char *filename, QemuOpts *opt } /* Create the qcow2 image (format layer) */ - ret = qcow2_co_create(bs, size, backing_file, backing_fmt, flags, - cluster_size, prealloc, opts, version, refcount_order, - encryptfmt, errp); + create_options = (BlockdevCreateOptions) { + .driver = BLOCKDEV_DRIVER_QCOW2, + .u.qcow2 = { + .file = &(BlockdevRef) { + .type = QTYPE_QSTRING, + .u.reference = bs->node_name, + }, + .size = size, + .has_version = true, + .version = version == 2 + ? BLOCKDEV_QCOW2_VERSION_V2 + : BLOCKDEV_QCOW2_VERSION_V3, + .has_backing_file = (backing_file != NULL), + .backing_file = backing_file, + .has_backing_fmt = (backing_fmt != NULL), + .backing_fmt = backing_drv, + .has_cluster_size = true, + .cluster_size = cluster_size, + .has_preallocation = true, + .preallocation = prealloc, + .has_lazy_refcounts = true, + .lazy_refcounts = (flags & BLOCK_FLAG_LAZY_REFCOUNTS), + .has_refcount_bits = true, + .refcount_bits = refcount_bits, + }, + }; + ret = qcow2_co_create(bs, &create_options, opts, encryptfmt, errp); if (ret < 0) { goto finish; } -- 2.13.6 Instead of passing a separate BlockDriverState* into qcow2_co_create(), make use of the BlockdevRef that is included in BlockdevCreateOptions. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> --- include/block/block.h | 1 + block.c | 47 +++++++++++++++++++++++++++++++++++++++++++++++ block/qcow2.c | 39 +++++++++++++++++++++++++-------------- 3 files changed, 73 insertions(+), 14 deletions(-) diff --git a/include/block/block.h b/include/block/block.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block.h +++ b/include/block/block.h @@ -XXX,XX +XXX,XX @@ BdrvChild *bdrv_open_child(const char *filename, BlockDriverState* parent, const BdrvChildRole *child_role, bool allow_none, Error **errp); +BlockDriverState *bdrv_open_blockdev_ref(BlockdevRef *ref, Error **errp); void bdrv_set_backing_hd(BlockDriverState *bs, BlockDriverState *backing_hd, Error **errp); int bdrv_open_backing_file(BlockDriverState *bs, QDict *parent_options, diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ #include "qapi/qmp/qdict.h" #include "qapi/qmp/qjson.h" #include "qapi/qmp/qstring.h" +#include "qapi/qobject-output-visitor.h" +#include "qapi/qapi-visit-block-core.h" #include "sysemu/block-backend.h" #include "sysemu/sysemu.h" #include "qemu/notify.h" @@ -XXX,XX +XXX,XX @@ BdrvChild *bdrv_open_child(const char *filename, return c; } +/* TODO Future callers may need to specify parent/child_role in order for + * option inheritance to work. Existing callers use it for the root node. */ +BlockDriverState *bdrv_open_blockdev_ref(BlockdevRef *ref, Error **errp) +{ + BlockDriverState *bs = NULL; + Error *local_err = NULL; + QObject *obj = NULL; + QDict *qdict = NULL; + const char *reference = NULL; + Visitor *v = NULL; + + if (ref->type == QTYPE_QSTRING) { + reference = ref->u.reference; + } else { + BlockdevOptions *options = &ref->u.definition; + assert(ref->type == QTYPE_QDICT); + + v = qobject_output_visitor_new(&obj); + visit_type_BlockdevOptions(v, NULL, &options, &local_err); + if (local_err) { + error_propagate(errp, local_err); + goto fail; + } + visit_complete(v, &obj); + + qdict = qobject_to_qdict(obj); + qdict_flatten(qdict); + + /* bdrv_open_inherit() defaults to the values in bdrv_flags (for + * compatibility with other callers) rather than what we want as the + * real defaults. Apply the defaults here instead. */ + qdict_set_default_str(qdict, BDRV_OPT_CACHE_DIRECT, "off"); + qdict_set_default_str(qdict, BDRV_OPT_CACHE_NO_FLUSH, "off"); + qdict_set_default_str(qdict, BDRV_OPT_READ_ONLY, "off"); + } + + bs = bdrv_open_inherit(NULL, reference, qdict, 0, NULL, NULL, errp); + obj = NULL; + +fail: + qobject_decref(obj); + visit_free(v); + return bs; +} + static BlockDriverState *bdrv_append_temp_snapshot(BlockDriverState *bs, int flags, QDict *snapshot_options, 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 uint64_t qcow2_opt_get_refcount_bits_del(QemuOpts *opts, int version, } static int coroutine_fn -qcow2_co_create(BlockDriverState *bs, BlockdevCreateOptions *create_options, - QemuOpts *opts, const char *encryptfmt, Error **errp) +qcow2_co_create(BlockdevCreateOptions *create_options, QemuOpts *opts, + const char *encryptfmt, Error **errp) { BlockdevCreateOptionsQcow2 *qcow2_opts; QDict *options; @@ -XXX,XX +XXX,XX @@ qcow2_co_create(BlockDriverState *bs, BlockdevCreateOptions *create_options, * 2 GB for 64k clusters, and we don't want to have a 2 GB initial file * size for any qcow2 image. */ - BlockBackend *blk; + BlockBackend *blk = NULL; + BlockDriverState *bs = NULL; QCowHeader *header; size_t cluster_size; int version; @@ -XXX,XX +XXX,XX @@ qcow2_co_create(BlockDriverState *bs, BlockdevCreateOptions *create_options, Error *local_err = NULL; int ret; - /* Validate options and set default values */ assert(create_options->driver == BLOCKDEV_DRIVER_QCOW2); qcow2_opts = &create_options->u.qcow2; + bs = bdrv_open_blockdev_ref(qcow2_opts->file, errp); + if (bs == NULL) { + return -EIO; + } + + /* Validate options and set default values */ if (!QEMU_IS_ALIGNED(qcow2_opts->size, BDRV_SECTOR_SIZE)) { error_setg(errp, "Image size must be a multiple of 512 bytes"); ret = -EINVAL; @@ -XXX,XX +XXX,XX @@ qcow2_co_create(BlockDriverState *bs, BlockdevCreateOptions *create_options, } if (!validate_cluster_size(cluster_size, errp)) { - return -EINVAL; + ret = -EINVAL; + goto out; } if (!qcow2_opts->has_preallocation) { @@ -XXX,XX +XXX,XX @@ qcow2_co_create(BlockDriverState *bs, BlockdevCreateOptions *create_options, { error_setg(errp, "Backing file and preallocation cannot be used at " "the same time"); - return -EINVAL; + ret = -EINVAL; + goto out; } if (qcow2_opts->has_backing_fmt && !qcow2_opts->has_backing_file) { error_setg(errp, "Backing format cannot be used without backing file"); - return -EINVAL; + ret = -EINVAL; + goto out; } if (!qcow2_opts->has_lazy_refcounts) { @@ -XXX,XX +XXX,XX @@ qcow2_co_create(BlockDriverState *bs, BlockdevCreateOptions *create_options, if (version < 3 && qcow2_opts->lazy_refcounts) { error_setg(errp, "Lazy refcounts only supported with compatibility " "level 1.1 and above (use compat=1.1 or greater)"); - return -EINVAL; + ret = -EINVAL; + goto out; } if (!qcow2_opts->has_refcount_bits) { @@ -XXX,XX +XXX,XX @@ qcow2_co_create(BlockDriverState *bs, BlockdevCreateOptions *create_options, { error_setg(errp, "Refcount width must be a power of two and may not " "exceed 64 bits"); - return -EINVAL; + ret = -EINVAL; + goto out; } if (version < 3 && qcow2_opts->refcount_bits != 16) { error_setg(errp, "Different refcount widths than 16 bits require " "compatibility level 1.1 or above (use compat=1.1 or " "greater)"); - return -EINVAL; + ret = -EINVAL; + goto out; } refcount_order = ctz32(qcow2_opts->refcount_bits); @@ -XXX,XX +XXX,XX @@ qcow2_co_create(BlockDriverState *bs, BlockdevCreateOptions *create_options, ret = 0; out: - if (blk) { - blk_unref(blk); - } + blk_unref(blk); + bdrv_unref(bs); return ret; } @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qcow2_co_create_opts(const char *filename, QemuOpts *opt .refcount_bits = refcount_bits, }, }; - ret = qcow2_co_create(bs, &create_options, opts, encryptfmt, errp); + ret = qcow2_co_create(&create_options, opts, encryptfmt, errp); if (ret < 0) { goto finish; } -- 2.13.6 Instead of passing the encryption format name and the QemuOpts down, use the QCryptoBlockCreateOptions contained in BlockdevCreateOptions. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> --- block/qcow2.c | 62 +++++++++++++++++++++++++++++++++++++++++++---------------- 1 file changed, 45 insertions(+), 17 deletions(-) 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_crypt_method_from_format(const char *encryptfmt) } } -static int qcow2_set_up_encryption(BlockDriverState *bs, const char *encryptfmt, - QemuOpts *opts, Error **errp) +static QCryptoBlockCreateOptions * +qcow2_parse_encryption(const char *encryptfmt, QemuOpts *opts, Error **errp) { - BDRVQcow2State *s = bs->opaque; QCryptoBlockCreateOptions *cryptoopts = NULL; - QCryptoBlock *crypto = NULL; - int ret = -EINVAL; QDict *options, *encryptopts; int fmt; @@ -XXX,XX +XXX,XX @@ static int qcow2_set_up_encryption(BlockDriverState *bs, const char *encryptfmt, error_setg(errp, "Unknown encryption format '%s'", encryptfmt); break; } - if (!cryptoopts) { - ret = -EINVAL; - goto out; + + QDECREF(encryptopts); + return cryptoopts; +} + +static int qcow2_set_up_encryption(BlockDriverState *bs, + QCryptoBlockCreateOptions *cryptoopts, + Error **errp) +{ + BDRVQcow2State *s = bs->opaque; + QCryptoBlock *crypto = NULL; + int fmt, ret; + + switch (cryptoopts->format) { + case Q_CRYPTO_BLOCK_FORMAT_LUKS: + fmt = QCOW_CRYPT_LUKS; + break; + case Q_CRYPTO_BLOCK_FORMAT_QCOW: + fmt = QCOW_CRYPT_AES; + break; + default: + error_setg(errp, "Crypto format not supported in qcow2"); + return -EINVAL; } + s->crypt_method_header = fmt; crypto = qcrypto_block_create(cryptoopts, "encrypt.", @@ -XXX,XX +XXX,XX @@ static int qcow2_set_up_encryption(BlockDriverState *bs, const char *encryptfmt, qcow2_crypto_hdr_write_func, bs, errp); if (!crypto) { - ret = -EINVAL; - goto out; + return -EINVAL; } ret = qcow2_update_header(bs); @@ -XXX,XX +XXX,XX @@ static int qcow2_set_up_encryption(BlockDriverState *bs, const char *encryptfmt, goto out; } + ret = 0; out: - QDECREF(encryptopts); qcrypto_block_free(crypto); - qapi_free_QCryptoBlockCreateOptions(cryptoopts); return ret; } @@ -XXX,XX +XXX,XX @@ static uint64_t qcow2_opt_get_refcount_bits_del(QemuOpts *opts, int version, } static int coroutine_fn -qcow2_co_create(BlockdevCreateOptions *create_options, QemuOpts *opts, - const char *encryptfmt, Error **errp) +qcow2_co_create(BlockdevCreateOptions *create_options, Error **errp) { BlockdevCreateOptionsQcow2 *qcow2_opts; QDict *options; @@ -XXX,XX +XXX,XX @@ qcow2_co_create(BlockdevCreateOptions *create_options, QemuOpts *opts, } /* Want encryption? There you go. */ - if (encryptfmt) { - ret = qcow2_set_up_encryption(blk_bs(blk), encryptfmt, opts, errp); + if (qcow2_opts->has_encrypt) { + ret = qcow2_set_up_encryption(blk_bs(blk), qcow2_opts->encrypt, errp); if (ret < 0) { goto out; } @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qcow2_co_create_opts(const char *filename, QemuOpts *opt int version; uint64_t refcount_bits; char *encryptfmt = NULL; + QCryptoBlockCreateOptions *cryptoopts = NULL; BlockDriverState *bs = NULL; Error *local_err = NULL; int ret; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qcow2_co_create_opts(const char *filename, QemuOpts *opt ret = -EINVAL; goto finish; } + encryptfmt = qemu_opt_get_del(opts, BLOCK_OPT_ENCRYPT_FORMAT); if (encryptfmt) { if (qemu_opt_get(opts, BLOCK_OPT_ENCRYPT)) { @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qcow2_co_create_opts(const char *filename, QemuOpts *opt } else if (qemu_opt_get_bool_del(opts, BLOCK_OPT_ENCRYPT, false)) { encryptfmt = g_strdup("aes"); } + if (encryptfmt) { + cryptoopts = qcow2_parse_encryption(encryptfmt, opts, errp); + if (cryptoopts == NULL) { + ret = -EINVAL; + goto finish; + } + } + cluster_size = qcow2_opt_get_cluster_size_del(opts, &local_err); if (local_err) { error_propagate(errp, local_err); @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qcow2_co_create_opts(const char *filename, QemuOpts *opt .backing_file = backing_file, .has_backing_fmt = (backing_fmt != NULL), .backing_fmt = backing_drv, + .has_encrypt = (encryptfmt != NULL), + .encrypt = cryptoopts, .has_cluster_size = true, .cluster_size = cluster_size, .has_preallocation = true, @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qcow2_co_create_opts(const char *filename, QemuOpts *opt .refcount_bits = refcount_bits, }, }; - ret = qcow2_co_create(&create_options, opts, encryptfmt, errp); + ret = qcow2_co_create(&create_options, errp); if (ret < 0) { goto finish; } @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qcow2_co_create_opts(const char *filename, QemuOpts *opt finish: bdrv_unref(bs); + qapi_free_QCryptoBlockCreateOptions(cryptoopts); g_free(backing_file); g_free(backing_fmt); g_free(encryptfmt); -- 2.13.6 Once qcow2_co_create() can be called directly on an already existing node, we must provide the 'full' and 'falloc' preallocation modes outside of creating the image on the protocol layer. Fortunately, we have preallocated truncate now which can provide this functionality. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> --- block/qcow2.c | 28 +++++++++++++++++++--------- 1 file changed, 19 insertions(+), 9 deletions(-) diff --git a/block/qcow2.c b/block/qcow2.c index XXXXXXX..XXXXXXX 100644 --- a/block/qcow2.c +++ b/block/qcow2.c @@ -XXX,XX +XXX,XX @@ qcow2_co_create(BlockdevCreateOptions *create_options, Error **errp) } blk_set_allow_write_beyond_eof(blk, true); + /* Clear the protocol layer and preallocate it if necessary */ + ret = blk_truncate(blk, 0, PREALLOC_MODE_OFF, errp); + if (ret < 0) { + goto out; + } + + if (qcow2_opts->preallocation == PREALLOC_MODE_FULL || + qcow2_opts->preallocation == PREALLOC_MODE_FALLOC) + { + int64_t prealloc_size = + qcow2_calc_prealloc_size(qcow2_opts->size, cluster_size, + refcount_order); + + ret = blk_truncate(blk, prealloc_size, qcow2_opts->preallocation, errp); + if (ret < 0) { + goto out; + } + } + /* Write the header */ QEMU_BUILD_BUG_ON((1 << MIN_CLUSTER_BITS) < sizeof(*header)); header = g_malloc0(cluster_size); @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qcow2_co_create_opts(const char *filename, QemuOpts *opt /* Create and open the file (protocol layer) */ - if (prealloc == PREALLOC_MODE_FULL || prealloc == PREALLOC_MODE_FALLOC) { - int refcount_order = ctz32(refcount_bits); - int64_t prealloc_size = - qcow2_calc_prealloc_size(size, cluster_size, refcount_order); - qemu_opt_set_number(opts, BLOCK_OPT_SIZE, prealloc_size, &error_abort); - qemu_opt_set(opts, BLOCK_OPT_PREALLOC, PreallocMode_str(prealloc), - &error_abort); - } - ret = bdrv_create_file(filename, opts, errp); if (ret < 0) { goto finish; -- 2.13.6 This allows, given a QemuOpts for a QemuOptsList that was merged from multiple QemuOptsList, to only consider those options that exist in one specific list. Block drivers need this to separate format-layer create options from protocol-level options. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> --- include/qemu/option.h | 2 ++ util/qemu-option.c | 42 +++++++++++++++++++++++++++++++++++++----- 2 files changed, 39 insertions(+), 5 deletions(-) diff --git a/include/qemu/option.h b/include/qemu/option.h index XXXXXXX..XXXXXXX 100644 --- a/include/qemu/option.h +++ b/include/qemu/option.h @@ -XXX,XX +XXX,XX @@ void qemu_opts_set_defaults(QemuOptsList *list, const char *params, int permit_abbrev); QemuOpts *qemu_opts_from_qdict(QemuOptsList *list, const QDict *qdict, Error **errp); +QDict *qemu_opts_to_qdict_filtered(QemuOpts *opts, QDict *qdict, + QemuOptsList *list, bool del); QDict *qemu_opts_to_qdict(QemuOpts *opts, QDict *qdict); void qemu_opts_absorb_qdict(QemuOpts *opts, QDict *qdict, Error **errp); diff --git a/util/qemu-option.c b/util/qemu-option.c index XXXXXXX..XXXXXXX 100644 --- a/util/qemu-option.c +++ b/util/qemu-option.c @@ -XXX,XX +XXX,XX @@ void qemu_opts_absorb_qdict(QemuOpts *opts, QDict *qdict, Error **errp) } /* - * Convert from QemuOpts to QDict. - * The QDict values are of type QString. + * Convert from QemuOpts to QDict. The QDict values are of type QString. + * + * If @list is given, only add those options to the QDict that are contained in + * the list. If @del is true, any options added to the QDict are removed from + * the QemuOpts, otherwise they remain there. + * + * If two options in @opts have the same name, they are processed in order + * so that the last one wins (consistent with the reverse iteration in + * qemu_opt_find()), but all of them are deleted if @del is true. + * * TODO We'll want to use types appropriate for opt->desc->type, but * this is enough for now. */ -QDict *qemu_opts_to_qdict(QemuOpts *opts, QDict *qdict) +QDict *qemu_opts_to_qdict_filtered(QemuOpts *opts, QDict *qdict, + QemuOptsList *list, bool del) { - QemuOpt *opt; + QemuOpt *opt, *next; if (!qdict) { qdict = qdict_new(); @@ -XXX,XX +XXX,XX @@ QDict *qemu_opts_to_qdict(QemuOpts *opts, QDict *qdict) if (opts->id) { qdict_put_str(qdict, "id", opts->id); } - QTAILQ_FOREACH(opt, &opts->head, next) { + QTAILQ_FOREACH_SAFE(opt, &opts->head, next, next) { + if (list) { + QemuOptDesc *desc; + bool found = false; + for (desc = list->desc; desc->name; desc++) { + if (!strcmp(desc->name, opt->name)) { + found = true; + break; + } + } + if (!found) { + continue; + } + } qdict_put_str(qdict, opt->name, opt->str); + if (del) { + qemu_opt_del(opt); + } } return qdict; } +/* Copy all options in a QemuOpts to the given QDict. See + * qemu_opts_to_qdict_filtered() for details. */ +QDict *qemu_opts_to_qdict(QemuOpts *opts, QDict *qdict) +{ + return qemu_opts_to_qdict_filtered(opts, qdict, NULL, false); +} + /* Validate parsed opts against descriptions where no * descriptions were provided in the QemuOptsList. */ -- 2.13.6 Basic test for merging two QemuOptsLists. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> --- tests/test-qemu-opts.c | 128 +++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) diff --git a/tests/test-qemu-opts.c b/tests/test-qemu-opts.c index XXXXXXX..XXXXXXX 100644 --- a/tests/test-qemu-opts.c +++ b/tests/test-qemu-opts.c @@ -XXX,XX +XXX,XX @@ static QemuOptsList opts_list_01 = { { .name = "str1", .type = QEMU_OPT_STRING, + .help = "Help texts are preserved in qemu_opts_append", + .def_value_str = "default", },{ .name = "str2", .type = QEMU_OPT_STRING, @@ -XXX,XX +XXX,XX @@ static QemuOptsList opts_list_01 = { },{ .name = "number1", .type = QEMU_OPT_NUMBER, + .help = "Having help texts only for some options is okay", },{ .name = "number2", .type = QEMU_OPT_NUMBER, @@ -XXX,XX +XXX,XX @@ static void test_opts_parse_size(void) qemu_opts_reset(&opts_list_02); } +static void append_verify_list_01(QemuOptDesc *desc, bool with_overlapping) +{ + int i = 0; + + if (with_overlapping) { + g_assert_cmpstr(desc[i].name, ==, "str1"); + g_assert_cmpint(desc[i].type, ==, QEMU_OPT_STRING); + g_assert_cmpstr(desc[i].help, ==, + "Help texts are preserved in qemu_opts_append"); + g_assert_cmpstr(desc[i].def_value_str, ==, "default"); + i++; + + g_assert_cmpstr(desc[i].name, ==, "str2"); + g_assert_cmpint(desc[i].type, ==, QEMU_OPT_STRING); + g_assert_cmpstr(desc[i].help, ==, NULL); + g_assert_cmpstr(desc[i].def_value_str, ==, NULL); + i++; + } + + g_assert_cmpstr(desc[i].name, ==, "str3"); + g_assert_cmpint(desc[i].type, ==, QEMU_OPT_STRING); + g_assert_cmpstr(desc[i].help, ==, NULL); + g_assert_cmpstr(desc[i].def_value_str, ==, NULL); + i++; + + g_assert_cmpstr(desc[i].name, ==, "number1"); + g_assert_cmpint(desc[i].type, ==, QEMU_OPT_NUMBER); + g_assert_cmpstr(desc[i].help, ==, + "Having help texts only for some options is okay"); + g_assert_cmpstr(desc[i].def_value_str, ==, NULL); + i++; + + g_assert_cmpstr(desc[i].name, ==, "number2"); + g_assert_cmpint(desc[i].type, ==, QEMU_OPT_NUMBER); + g_assert_cmpstr(desc[i].help, ==, NULL); + g_assert_cmpstr(desc[i].def_value_str, ==, NULL); + i++; + + g_assert_cmpstr(desc[i].name, ==, NULL); +} + +static void append_verify_list_02(QemuOptDesc *desc) +{ + int i = 0; + + g_assert_cmpstr(desc[i].name, ==, "str1"); + g_assert_cmpint(desc[i].type, ==, QEMU_OPT_STRING); + g_assert_cmpstr(desc[i].help, ==, NULL); + g_assert_cmpstr(desc[i].def_value_str, ==, NULL); + i++; + + g_assert_cmpstr(desc[i].name, ==, "str2"); + g_assert_cmpint(desc[i].type, ==, QEMU_OPT_STRING); + g_assert_cmpstr(desc[i].help, ==, NULL); + g_assert_cmpstr(desc[i].def_value_str, ==, NULL); + i++; + + g_assert_cmpstr(desc[i].name, ==, "bool1"); + g_assert_cmpint(desc[i].type, ==, QEMU_OPT_BOOL); + g_assert_cmpstr(desc[i].help, ==, NULL); + g_assert_cmpstr(desc[i].def_value_str, ==, NULL); + i++; + + g_assert_cmpstr(desc[i].name, ==, "bool2"); + g_assert_cmpint(desc[i].type, ==, QEMU_OPT_BOOL); + g_assert_cmpstr(desc[i].help, ==, NULL); + g_assert_cmpstr(desc[i].def_value_str, ==, NULL); + i++; + + g_assert_cmpstr(desc[i].name, ==, "size1"); + g_assert_cmpint(desc[i].type, ==, QEMU_OPT_SIZE); + g_assert_cmpstr(desc[i].help, ==, NULL); + g_assert_cmpstr(desc[i].def_value_str, ==, NULL); + i++; + + g_assert_cmpstr(desc[i].name, ==, "size2"); + g_assert_cmpint(desc[i].type, ==, QEMU_OPT_SIZE); + g_assert_cmpstr(desc[i].help, ==, NULL); + g_assert_cmpstr(desc[i].def_value_str, ==, NULL); + i++; + + g_assert_cmpstr(desc[i].name, ==, "size3"); + g_assert_cmpint(desc[i].type, ==, QEMU_OPT_SIZE); + g_assert_cmpstr(desc[i].help, ==, NULL); + g_assert_cmpstr(desc[i].def_value_str, ==, NULL); +} + +static void test_opts_append_to_null(void) +{ + QemuOptsList *merged; + + merged = qemu_opts_append(NULL, &opts_list_01); + g_assert(merged != &opts_list_01); + + g_assert_cmpstr(merged->name, ==, NULL); + g_assert_cmpstr(merged->implied_opt_name, ==, NULL); + g_assert_false(merged->merge_lists); + + append_verify_list_01(merged->desc, true); + + qemu_opts_free(merged); +} + +static void test_opts_append(void) +{ + QemuOptsList *first, *merged; + + first = qemu_opts_append(NULL, &opts_list_02); + merged = qemu_opts_append(first, &opts_list_01); + g_assert(first != &opts_list_02); + g_assert(merged != &opts_list_01); + + g_assert_cmpstr(merged->name, ==, NULL); + g_assert_cmpstr(merged->implied_opt_name, ==, NULL); + g_assert_false(merged->merge_lists); + + append_verify_list_02(&merged->desc[0]); + append_verify_list_01(&merged->desc[7], false); + + qemu_opts_free(merged); +} + + int main(int argc, char *argv[]) { register_opts(); @@ -XXX,XX +XXX,XX @@ int main(int argc, char *argv[]) g_test_add_func("/qemu-opts/opts_parse/bool", test_opts_parse_bool); g_test_add_func("/qemu-opts/opts_parse/number", test_opts_parse_number); g_test_add_func("/qemu-opts/opts_parse/size", test_opts_parse_size); + g_test_add_func("/qemu-opts/append_to_null", test_opts_append_to_null); + g_test_add_func("/qemu-opts/append", test_opts_append); g_test_run(); return 0; } -- 2.13.6 Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> --- tests/test-qemu-opts.c | 125 +++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 125 insertions(+) diff --git a/tests/test-qemu-opts.c b/tests/test-qemu-opts.c index XXXXXXX..XXXXXXX 100644 --- a/tests/test-qemu-opts.c +++ b/tests/test-qemu-opts.c @@ -XXX,XX +XXX,XX @@ #include "qemu/osdep.h" #include "qemu/cutils.h" #include "qemu/option.h" +#include "qemu/option_int.h" #include "qapi/error.h" #include "qapi/qmp/qdict.h" #include "qapi/qmp/qstring.h" @@ -XXX,XX +XXX,XX @@ static void test_opts_append(void) qemu_opts_free(merged); } +static void test_opts_to_qdict_basic(void) +{ + QemuOpts *opts; + QDict *dict; + + opts = qemu_opts_parse(&opts_list_01, "str1=foo,str2=,str3=bar,number1=42", + false, &error_abort); + g_assert(opts != NULL); + + dict = qemu_opts_to_qdict(opts, NULL); + g_assert(dict != NULL); + + g_assert_cmpstr(qdict_get_str(dict, "str1"), ==, "foo"); + g_assert_cmpstr(qdict_get_str(dict, "str2"), ==, ""); + g_assert_cmpstr(qdict_get_str(dict, "str3"), ==, "bar"); + g_assert_cmpstr(qdict_get_str(dict, "number1"), ==, "42"); + g_assert_false(qdict_haskey(dict, "number2")); + + QDECREF(dict); + qemu_opts_del(opts); +} + +static void test_opts_to_qdict_filtered(void) +{ + QemuOptsList *first, *merged; + QemuOpts *opts; + QDict *dict; + + first = qemu_opts_append(NULL, &opts_list_02); + merged = qemu_opts_append(first, &opts_list_01); + + opts = qemu_opts_parse(merged, + "str1=foo,str2=,str3=bar,bool1=off,number1=42", + false, &error_abort); + g_assert(opts != NULL); + + /* Convert to QDict without deleting from opts */ + dict = qemu_opts_to_qdict_filtered(opts, NULL, &opts_list_01, false); + g_assert(dict != NULL); + g_assert_cmpstr(qdict_get_str(dict, "str1"), ==, "foo"); + g_assert_cmpstr(qdict_get_str(dict, "str2"), ==, ""); + g_assert_cmpstr(qdict_get_str(dict, "str3"), ==, "bar"); + g_assert_cmpstr(qdict_get_str(dict, "number1"), ==, "42"); + g_assert_false(qdict_haskey(dict, "number2")); + g_assert_false(qdict_haskey(dict, "bool1")); + QDECREF(dict); + + dict = qemu_opts_to_qdict_filtered(opts, NULL, &opts_list_02, false); + g_assert(dict != NULL); + g_assert_cmpstr(qdict_get_str(dict, "str1"), ==, "foo"); + g_assert_cmpstr(qdict_get_str(dict, "str2"), ==, ""); + g_assert_cmpstr(qdict_get_str(dict, "bool1"), ==, "off"); + g_assert_false(qdict_haskey(dict, "str3")); + g_assert_false(qdict_haskey(dict, "number1")); + g_assert_false(qdict_haskey(dict, "number2")); + QDECREF(dict); + + /* Now delete converted options from opts */ + dict = qemu_opts_to_qdict_filtered(opts, NULL, &opts_list_01, true); + g_assert(dict != NULL); + g_assert_cmpstr(qdict_get_str(dict, "str1"), ==, "foo"); + g_assert_cmpstr(qdict_get_str(dict, "str2"), ==, ""); + g_assert_cmpstr(qdict_get_str(dict, "str3"), ==, "bar"); + g_assert_cmpstr(qdict_get_str(dict, "number1"), ==, "42"); + g_assert_false(qdict_haskey(dict, "number2")); + g_assert_false(qdict_haskey(dict, "bool1")); + QDECREF(dict); + + dict = qemu_opts_to_qdict_filtered(opts, NULL, &opts_list_02, true); + g_assert(dict != NULL); + g_assert_cmpstr(qdict_get_str(dict, "bool1"), ==, "off"); + g_assert_false(qdict_haskey(dict, "str1")); + g_assert_false(qdict_haskey(dict, "str2")); + g_assert_false(qdict_haskey(dict, "str3")); + g_assert_false(qdict_haskey(dict, "number1")); + g_assert_false(qdict_haskey(dict, "number2")); + QDECREF(dict); + + g_assert_true(QTAILQ_EMPTY(&opts->head)); + + qemu_opts_del(opts); + qemu_opts_free(merged); +} + +static void test_opts_to_qdict_duplicates(void) +{ + QemuOpts *opts; + QemuOpt *opt; + QDict *dict; + + opts = qemu_opts_parse(&opts_list_03, "foo=a,foo=b", false, &error_abort); + g_assert(opts != NULL); + + /* Verify that opts has two options with the same name */ + opt = QTAILQ_FIRST(&opts->head); + g_assert_cmpstr(opt->name, ==, "foo"); + g_assert_cmpstr(opt->str , ==, "a"); + + opt = QTAILQ_NEXT(opt, next); + g_assert_cmpstr(opt->name, ==, "foo"); + g_assert_cmpstr(opt->str , ==, "b"); + + opt = QTAILQ_NEXT(opt, next); + g_assert(opt == NULL); + + /* In the conversion to QDict, the last one wins */ + dict = qemu_opts_to_qdict(opts, NULL); + g_assert(dict != NULL); + g_assert_cmpstr(qdict_get_str(dict, "foo"), ==, "b"); + QDECREF(dict); + + /* The last one still wins if entries are deleted, and both are deleted */ + dict = qemu_opts_to_qdict_filtered(opts, NULL, NULL, true); + g_assert(dict != NULL); + g_assert_cmpstr(qdict_get_str(dict, "foo"), ==, "b"); + QDECREF(dict); + + g_assert_true(QTAILQ_EMPTY(&opts->head)); + + qemu_opts_del(opts); +} int main(int argc, char *argv[]) { @@ -XXX,XX +XXX,XX @@ int main(int argc, char *argv[]) g_test_add_func("/qemu-opts/opts_parse/size", test_opts_parse_size); g_test_add_func("/qemu-opts/append_to_null", test_opts_append_to_null); g_test_add_func("/qemu-opts/append", test_opts_append); + g_test_add_func("/qemu-opts/to_qdict/basic", test_opts_to_qdict_basic); + g_test_add_func("/qemu-opts/to_qdict/filtered", test_opts_to_qdict_filtered); + g_test_add_func("/qemu-opts/to_qdict/duplicates", test_opts_to_qdict_duplicates); g_test_run(); return 0; } -- 2.13.6 A few block drivers will need to rename .bdrv_create options for their QAPIfication, so let's have a helper function for that. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> --- include/qapi/qmp/qdict.h | 6 +++ qobject/qdict.c | 34 +++++++++++++ tests/check-qdict.c | 129 +++++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 169 insertions(+) diff --git a/include/qapi/qmp/qdict.h b/include/qapi/qmp/qdict.h index XXXXXXX..XXXXXXX 100644 --- a/include/qapi/qmp/qdict.h +++ b/include/qapi/qmp/qdict.h @@ -XXX,XX +XXX,XX @@ QObject *qdict_crumple(const QDict *src, Error **errp); void qdict_join(QDict *dest, QDict *src, bool overwrite); +typedef struct QDictRenames { + const char *from; + const char *to; +} QDictRenames; +bool qdict_rename_keys(QDict *qdict, const QDictRenames *renames, Error **errp); + #endif /* QDICT_H */ diff --git a/qobject/qdict.c b/qobject/qdict.c index XXXXXXX..XXXXXXX 100644 --- a/qobject/qdict.c +++ b/qobject/qdict.c @@ -XXX,XX +XXX,XX @@ void qdict_join(QDict *dest, QDict *src, bool overwrite) entry = next; } } + +/** + * qdict_rename_keys(): Rename keys in qdict according to the replacements + * specified in the array renames. The array must be terminated by an entry + * with from = NULL. + * + * The renames are performed individually in the order of the array, so entries + * may be renamed multiple times and may or may not conflict depending on the + * order of the renames array. + * + * Returns true for success, false in error cases. + */ +bool qdict_rename_keys(QDict *qdict, const QDictRenames *renames, Error **errp) +{ + QObject *qobj; + + while (renames->from) { + if (qdict_haskey(qdict, renames->from)) { + if (qdict_haskey(qdict, renames->to)) { + error_setg(errp, "'%s' and its alias '%s' can't be used at the " + "same time", renames->to, renames->from); + return false; + } + + qobj = qdict_get(qdict, renames->from); + qobject_incref(qobj); + qdict_put_obj(qdict, renames->to, qobj); + qdict_del(qdict, renames->from); + } + + renames++; + } + return true; +} diff --git a/tests/check-qdict.c b/tests/check-qdict.c index XXXXXXX..XXXXXXX 100644 --- a/tests/check-qdict.c +++ b/tests/check-qdict.c @@ -XXX,XX +XXX,XX @@ static void qdict_crumple_test_empty(void) QDECREF(dst); } +static int qdict_count_entries(QDict *dict) +{ + const QDictEntry *e; + int count = 0; + + for (e = qdict_first(dict); e; e = qdict_next(dict, e)) { + count++; + } + + return count; +} + +static void qdict_rename_keys_test(void) +{ + QDict *dict = qdict_new(); + QDict *copy; + QDictRenames *renames; + Error *local_err = NULL; + + qdict_put_str(dict, "abc", "foo"); + qdict_put_str(dict, "abcdef", "bar"); + qdict_put_int(dict, "number", 42); + qdict_put_bool(dict, "flag", true); + qdict_put_null(dict, "nothing"); + + /* Empty rename list */ + renames = (QDictRenames[]) { + { NULL, "this can be anything" } + }; + copy = qdict_clone_shallow(dict); + qdict_rename_keys(copy, renames, &error_abort); + + g_assert_cmpstr(qdict_get_str(copy, "abc"), ==, "foo"); + g_assert_cmpstr(qdict_get_str(copy, "abcdef"), ==, "bar"); + g_assert_cmpint(qdict_get_int(copy, "number"), ==, 42); + g_assert_cmpint(qdict_get_bool(copy, "flag"), ==, true); + g_assert(qobject_type(qdict_get(copy, "nothing")) == QTYPE_QNULL); + g_assert_cmpint(qdict_count_entries(copy), ==, 5); + + QDECREF(copy); + + /* Simple rename of all entries */ + renames = (QDictRenames[]) { + { "abc", "str1" }, + { "abcdef", "str2" }, + { "number", "int" }, + { "flag", "bool" }, + { "nothing", "null" }, + { NULL , NULL } + }; + copy = qdict_clone_shallow(dict); + qdict_rename_keys(copy, renames, &error_abort); + + g_assert(!qdict_haskey(copy, "abc")); + g_assert(!qdict_haskey(copy, "abcdef")); + g_assert(!qdict_haskey(copy, "number")); + g_assert(!qdict_haskey(copy, "flag")); + g_assert(!qdict_haskey(copy, "nothing")); + + g_assert_cmpstr(qdict_get_str(copy, "str1"), ==, "foo"); + g_assert_cmpstr(qdict_get_str(copy, "str2"), ==, "bar"); + g_assert_cmpint(qdict_get_int(copy, "int"), ==, 42); + g_assert_cmpint(qdict_get_bool(copy, "bool"), ==, true); + g_assert(qobject_type(qdict_get(copy, "null")) == QTYPE_QNULL); + g_assert_cmpint(qdict_count_entries(copy), ==, 5); + + QDECREF(copy); + + /* Renames are processed top to bottom */ + renames = (QDictRenames[]) { + { "abc", "tmp" }, + { "abcdef", "abc" }, + { "number", "abcdef" }, + { "flag", "number" }, + { "nothing", "flag" }, + { "tmp", "nothing" }, + { NULL , NULL } + }; + copy = qdict_clone_shallow(dict); + qdict_rename_keys(copy, renames, &error_abort); + + g_assert_cmpstr(qdict_get_str(copy, "nothing"), ==, "foo"); + g_assert_cmpstr(qdict_get_str(copy, "abc"), ==, "bar"); + g_assert_cmpint(qdict_get_int(copy, "abcdef"), ==, 42); + g_assert_cmpint(qdict_get_bool(copy, "number"), ==, true); + g_assert(qobject_type(qdict_get(copy, "flag")) == QTYPE_QNULL); + g_assert(!qdict_haskey(copy, "tmp")); + g_assert_cmpint(qdict_count_entries(copy), ==, 5); + + QDECREF(copy); + + /* Conflicting rename */ + renames = (QDictRenames[]) { + { "abcdef", "abc" }, + { NULL , NULL } + }; + copy = qdict_clone_shallow(dict); + qdict_rename_keys(copy, renames, &local_err); + + g_assert(local_err != NULL); + error_free(local_err); + local_err = NULL; + + g_assert_cmpstr(qdict_get_str(copy, "abc"), ==, "foo"); + g_assert_cmpstr(qdict_get_str(copy, "abcdef"), ==, "bar"); + g_assert_cmpint(qdict_get_int(copy, "number"), ==, 42); + g_assert_cmpint(qdict_get_bool(copy, "flag"), ==, true); + g_assert(qobject_type(qdict_get(copy, "nothing")) == QTYPE_QNULL); + g_assert_cmpint(qdict_count_entries(copy), ==, 5); + + QDECREF(copy); + + /* Renames in an empty dict */ + renames = (QDictRenames[]) { + { "abcdef", "abc" }, + { NULL , NULL } + }; + + QDECREF(dict); + dict = qdict_new(); + + qdict_rename_keys(dict, renames, &error_abort); + g_assert(qdict_first(dict) == NULL); + + QDECREF(dict); +} + static void qdict_crumple_test_bad_inputs(void) { QDict *src; @@ -XXX,XX +XXX,XX @@ int main(int argc, char **argv) g_test_add_func("/public/crumple/bad_inputs", qdict_crumple_test_bad_inputs); + g_test_add_func("/public/rename_keys", qdict_rename_keys_test); + /* The Big one */ if (g_test_slow()) { g_test_add_func("/stress/test", qdict_stress_test); -- 2.13.6 Instead of manually creating the BlockdevCreateOptions object, use a visitor to parse the given options into the QAPI object. This involves translation from the old command line syntax to the syntax mandated by the QAPI schema. Option names are still checked against qcow2_create_opts, so only the old option names are allowed on the command line, even if they are translated in qcow2_create(). In contrast, new option values are optionally recognised besides the old values: 'compat' accepts 'v2'/'v3' as an alias for '0.10'/'1.1', and 'encrypt.format' accepts 'qcow' as an alias for 'aes' now. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> --- block/qcow2.c | 218 ++++++++++++++++----------------------------- tests/qemu-iotests/049.out | 8 +- tests/qemu-iotests/112.out | 4 +- 3 files changed, 84 insertions(+), 146 deletions(-) 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 @@ #include "qemu/option_int.h" #include "qemu/cutils.h" #include "qemu/bswap.h" -#include "qapi/opts-visitor.h" +#include "qapi/qobject-input-visitor.h" +#include "qapi/qapi-visit-block-core.h" #include "block/crypto.h" /* @@ -XXX,XX +XXX,XX @@ static int qcow2_crypt_method_from_format(const char *encryptfmt) } } -static QCryptoBlockCreateOptions * -qcow2_parse_encryption(const char *encryptfmt, QemuOpts *opts, Error **errp) -{ - QCryptoBlockCreateOptions *cryptoopts = NULL; - QDict *options, *encryptopts; - int fmt; - - options = qemu_opts_to_qdict(opts, NULL); - qdict_extract_subqdict(options, &encryptopts, "encrypt."); - QDECREF(options); - - fmt = qcow2_crypt_method_from_format(encryptfmt); - - switch (fmt) { - case QCOW_CRYPT_LUKS: - cryptoopts = block_crypto_create_opts_init( - Q_CRYPTO_BLOCK_FORMAT_LUKS, encryptopts, errp); - break; - case QCOW_CRYPT_AES: - cryptoopts = block_crypto_create_opts_init( - Q_CRYPTO_BLOCK_FORMAT_QCOW, encryptopts, errp); - break; - default: - error_setg(errp, "Unknown encryption format '%s'", encryptfmt); - break; - } - - QDECREF(encryptopts); - return cryptoopts; -} - static int qcow2_set_up_encryption(BlockDriverState *bs, QCryptoBlockCreateOptions *cryptoopts, Error **errp) @@ -XXX,XX +XXX,XX @@ qcow2_co_create(BlockdevCreateOptions *create_options, Error **errp) } if (version < 3 && qcow2_opts->lazy_refcounts) { error_setg(errp, "Lazy refcounts only supported with compatibility " - "level 1.1 and above (use compat=1.1 or greater)"); + "level 1.1 and above (use version=v3 or greater)"); ret = -EINVAL; goto out; } @@ -XXX,XX +XXX,XX @@ qcow2_co_create(BlockdevCreateOptions *create_options, Error **errp) } if (version < 3 && qcow2_opts->refcount_bits != 16) { error_setg(errp, "Different refcount widths than 16 bits require " - "compatibility level 1.1 or above (use compat=1.1 or " + "compatibility level 1.1 or above (use version=v3 or " "greater)"); ret = -EINVAL; goto out; @@ -XXX,XX +XXX,XX @@ out: static int coroutine_fn qcow2_co_create_opts(const char *filename, QemuOpts *opts, Error **errp) { - BlockdevCreateOptions create_options; - char *backing_file = NULL; - char *backing_fmt = NULL; - BlockdevDriver backing_drv; - char *buf = NULL; - uint64_t size = 0; - int flags = 0; - size_t cluster_size = DEFAULT_CLUSTER_SIZE; - PreallocMode prealloc; - int version; - uint64_t refcount_bits; - char *encryptfmt = NULL; - QCryptoBlockCreateOptions *cryptoopts = NULL; + BlockdevCreateOptions *create_options = NULL; + QDict *qdict = NULL; + QObject *qobj; + Visitor *v; BlockDriverState *bs = NULL; Error *local_err = NULL; + const char *val; int ret; - /* Read out options */ - size = ROUND_UP(qemu_opt_get_size_del(opts, BLOCK_OPT_SIZE, 0), - BDRV_SECTOR_SIZE); - backing_file = qemu_opt_get_del(opts, BLOCK_OPT_BACKING_FILE); - backing_fmt = qemu_opt_get_del(opts, BLOCK_OPT_BACKING_FMT); - backing_drv = qapi_enum_parse(&BlockdevDriver_lookup, backing_fmt, - 0, &local_err); - if (local_err) { - error_propagate(errp, local_err); + /* Only the keyval visitor supports the dotted syntax needed for + * encryption, so go through a QDict before getting a QAPI type. Ignore + * options meant for the protocol layer so that the visitor doesn't + * complain. */ + qdict = qemu_opts_to_qdict_filtered(opts, NULL, bdrv_qcow2.create_opts, + true); + + /* Handle encryption options */ + val = qdict_get_try_str(qdict, BLOCK_OPT_ENCRYPT); + if (val && !strcmp(val, "on")) { + qdict_put_str(qdict, BLOCK_OPT_ENCRYPT, "qcow"); + } else if (val && !strcmp(val, "off")) { + qdict_del(qdict, BLOCK_OPT_ENCRYPT); + } + + val = qdict_get_try_str(qdict, BLOCK_OPT_ENCRYPT_FORMAT); + if (val && !strcmp(val, "aes")) { + qdict_put_str(qdict, BLOCK_OPT_ENCRYPT_FORMAT, "qcow"); + } + + /* Convert compat=0.10/1.1 into compat=v2/v3, to be renamed into + * version=v2/v3 below. */ + val = qdict_get_try_str(qdict, BLOCK_OPT_COMPAT_LEVEL); + if (val && !strcmp(val, "0.10")) { + qdict_put_str(qdict, BLOCK_OPT_COMPAT_LEVEL, "v2"); + } else if (val && !strcmp(val, "1.1")) { + qdict_put_str(qdict, BLOCK_OPT_COMPAT_LEVEL, "v3"); + } + + /* Change legacy command line options into QMP ones */ + static const QDictRenames opt_renames[] = { + { BLOCK_OPT_BACKING_FILE, "backing-file" }, + { BLOCK_OPT_BACKING_FMT, "backing-fmt" }, + { BLOCK_OPT_CLUSTER_SIZE, "cluster-size" }, + { BLOCK_OPT_LAZY_REFCOUNTS, "lazy-refcounts" }, + { BLOCK_OPT_REFCOUNT_BITS, "refcount-bits" }, + { BLOCK_OPT_ENCRYPT, BLOCK_OPT_ENCRYPT_FORMAT }, + { BLOCK_OPT_COMPAT_LEVEL, "version" }, + { NULL, NULL }, + }; + + if (!qdict_rename_keys(qdict, opt_renames, errp)) { ret = -EINVAL; goto finish; } - encryptfmt = qemu_opt_get_del(opts, BLOCK_OPT_ENCRYPT_FORMAT); - if (encryptfmt) { - if (qemu_opt_get(opts, BLOCK_OPT_ENCRYPT)) { - error_setg(errp, "Options " BLOCK_OPT_ENCRYPT " and " - BLOCK_OPT_ENCRYPT_FORMAT " are mutually exclusive"); - ret = -EINVAL; - goto finish; - } - } else if (qemu_opt_get_bool_del(opts, BLOCK_OPT_ENCRYPT, false)) { - encryptfmt = g_strdup("aes"); - } - if (encryptfmt) { - cryptoopts = qcow2_parse_encryption(encryptfmt, opts, errp); - if (cryptoopts == NULL) { - ret = -EINVAL; - goto finish; - } - } - - cluster_size = qcow2_opt_get_cluster_size_del(opts, &local_err); - if (local_err) { - error_propagate(errp, local_err); - ret = -EINVAL; + /* Create and open the file (protocol layer) */ + ret = bdrv_create_file(filename, opts, errp); + if (ret < 0) { goto finish; } - buf = qemu_opt_get_del(opts, BLOCK_OPT_PREALLOC); - prealloc = qapi_enum_parse(&PreallocMode_lookup, buf, - PREALLOC_MODE_OFF, &local_err); - if (local_err) { - error_propagate(errp, local_err); - ret = -EINVAL; + + bs = bdrv_open(filename, NULL, NULL, + BDRV_O_RDWR | BDRV_O_RESIZE | BDRV_O_PROTOCOL, errp); + if (bs == NULL) { + ret = -EIO; goto finish; } - version = qcow2_opt_get_version_del(opts, &local_err); - if (local_err) { - error_propagate(errp, local_err); + /* Set 'driver' and 'node' options */ + qdict_put_str(qdict, "driver", "qcow2"); + qdict_put_str(qdict, "file", bs->node_name); + + /* Now get the QAPI type BlockdevCreateOptions */ + qobj = qdict_crumple(qdict, errp); + QDECREF(qdict); + qdict = qobject_to_qdict(qobj); + if (qdict == NULL) { ret = -EINVAL; goto finish; } - if (qemu_opt_get_bool_del(opts, BLOCK_OPT_LAZY_REFCOUNTS, false)) { - flags |= BLOCK_FLAG_LAZY_REFCOUNTS; - } + v = qobject_input_visitor_new_keyval(QOBJECT(qdict)); + visit_type_BlockdevCreateOptions(v, NULL, &create_options, &local_err); + visit_free(v); - refcount_bits = qcow2_opt_get_refcount_bits_del(opts, version, &local_err); if (local_err) { error_propagate(errp, local_err); ret = -EINVAL; goto finish; } - - /* Create and open the file (protocol layer) */ - ret = bdrv_create_file(filename, opts, errp); - if (ret < 0) { - goto finish; - } - - bs = bdrv_open(filename, NULL, NULL, - BDRV_O_RDWR | BDRV_O_RESIZE | BDRV_O_PROTOCOL, errp); - if (bs == NULL) { - ret = -EIO; - goto finish; - } + /* Silently round up size */ + create_options->u.qcow2.size = ROUND_UP(create_options->u.qcow2.size, + BDRV_SECTOR_SIZE); /* Create the qcow2 image (format layer) */ - create_options = (BlockdevCreateOptions) { - .driver = BLOCKDEV_DRIVER_QCOW2, - .u.qcow2 = { - .file = &(BlockdevRef) { - .type = QTYPE_QSTRING, - .u.reference = bs->node_name, - }, - .size = size, - .has_version = true, - .version = version == 2 - ? BLOCKDEV_QCOW2_VERSION_V2 - : BLOCKDEV_QCOW2_VERSION_V3, - .has_backing_file = (backing_file != NULL), - .backing_file = backing_file, - .has_backing_fmt = (backing_fmt != NULL), - .backing_fmt = backing_drv, - .has_encrypt = (encryptfmt != NULL), - .encrypt = cryptoopts, - .has_cluster_size = true, - .cluster_size = cluster_size, - .has_preallocation = true, - .preallocation = prealloc, - .has_lazy_refcounts = true, - .lazy_refcounts = (flags & BLOCK_FLAG_LAZY_REFCOUNTS), - .has_refcount_bits = true, - .refcount_bits = refcount_bits, - }, - }; - ret = qcow2_co_create(&create_options, errp); + ret = qcow2_co_create(create_options, errp); if (ret < 0) { goto finish; } + ret = 0; finish: + QDECREF(qdict); bdrv_unref(bs); - - qapi_free_QCryptoBlockCreateOptions(cryptoopts); - g_free(backing_file); - g_free(backing_fmt); - g_free(encryptfmt); - g_free(buf); + qapi_free_BlockdevCreateOptions(create_options); return ret; } diff --git a/tests/qemu-iotests/049.out b/tests/qemu-iotests/049.out index XXXXXXX..XXXXXXX 100644 --- a/tests/qemu-iotests/049.out +++ b/tests/qemu-iotests/049.out @@ -XXX,XX +XXX,XX @@ qemu-img create -f qcow2 -o compat=1.1 TEST_DIR/t.qcow2 64M Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=1.1 cluster_size=65536 lazy_refcounts=off refcount_bits=16 qemu-img create -f qcow2 -o compat=0.42 TEST_DIR/t.qcow2 64M -qemu-img: TEST_DIR/t.qcow2: Invalid compatibility level: '0.42' +qemu-img: TEST_DIR/t.qcow2: Invalid parameter '0.42' Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=0.42 cluster_size=65536 lazy_refcounts=off refcount_bits=16 qemu-img create -f qcow2 -o compat=foobar TEST_DIR/t.qcow2 64M -qemu-img: TEST_DIR/t.qcow2: Invalid compatibility level: 'foobar' +qemu-img: TEST_DIR/t.qcow2: Invalid parameter 'foobar' Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=foobar cluster_size=65536 lazy_refcounts=off refcount_bits=16 == Check preallocation option == @@ -XXX,XX +XXX,XX @@ qemu-img create -f qcow2 -o preallocation=metadata TEST_DIR/t.qcow2 64M Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=65536 preallocation=metadata lazy_refcounts=off refcount_bits=16 qemu-img create -f qcow2 -o preallocation=1234 TEST_DIR/t.qcow2 64M -qemu-img: TEST_DIR/t.qcow2: invalid parameter value: 1234 +qemu-img: TEST_DIR/t.qcow2: Invalid parameter '1234' Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 cluster_size=65536 preallocation=1234 lazy_refcounts=off refcount_bits=16 == Check encryption option == @@ -XXX,XX +XXX,XX @@ qemu-img create -f qcow2 -o compat=0.10,lazy_refcounts=off TEST_DIR/t.qcow2 64M Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=0.10 cluster_size=65536 lazy_refcounts=off refcount_bits=16 qemu-img create -f qcow2 -o compat=0.10,lazy_refcounts=on TEST_DIR/t.qcow2 64M -qemu-img: TEST_DIR/t.qcow2: Lazy refcounts only supported with compatibility level 1.1 and above (use compat=1.1 or greater) +qemu-img: TEST_DIR/t.qcow2: Lazy refcounts only supported with compatibility level 1.1 and above (use version=v3 or greater) Formatting 'TEST_DIR/t.qcow2', fmt=qcow2 size=67108864 compat=0.10 cluster_size=65536 lazy_refcounts=on refcount_bits=16 *** done diff --git a/tests/qemu-iotests/112.out b/tests/qemu-iotests/112.out index XXXXXXX..XXXXXXX 100644 --- a/tests/qemu-iotests/112.out +++ b/tests/qemu-iotests/112.out @@ -XXX,XX +XXX,XX @@ refcount bits: 16 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864 refcount bits: 16 -qemu-img: TEST_DIR/t.IMGFMT: Different refcount widths than 16 bits require compatibility level 1.1 or above (use or greater) +qemu-img: TEST_DIR/t.IMGFMT: Different refcount widths than 16 bits require compatibility level 1.1 or above (use version=v3 or greater) Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864 -qemu-img: TEST_DIR/t.IMGFMT: Different refcount widths than 16 bits require compatibility level 1.1 or above (use or greater) +qemu-img: TEST_DIR/t.IMGFMT: Different refcount widths than 16 bits require compatibility level 1.1 or above (use version=v3 or greater) Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864 === Snapshot limit on refcount_bits=1 === -- 2.13.6 We'll use a separate source file for image creation, and we need to check there whether the requested driver is whitelisted. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> --- include/block/block.h | 1 + block.c | 2 +- 2 files changed, 2 insertions(+), 1 deletion(-) diff --git a/include/block/block.h b/include/block/block.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block.h +++ b/include/block/block.h @@ -XXX,XX +XXX,XX @@ char *bdrv_perm_names(uint64_t perm); void bdrv_init(void); void bdrv_init_with_whitelist(void); bool bdrv_uses_whitelist(void); +int bdrv_is_whitelisted(BlockDriver *drv, bool read_only); BlockDriver *bdrv_find_protocol(const char *filename, bool allow_protocol_prefix, Error **errp); diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ BlockDriver *bdrv_find_format(const char *format_name) return bdrv_do_find_format(format_name); } -static int bdrv_is_whitelisted(BlockDriver *drv, bool read_only) +int bdrv_is_whitelisted(BlockDriver *drv, bool read_only) { static const char *whitelist_rw[] = { CONFIG_BDRV_RW_WHITELIST -- 2.13.6 This adds a synchronous x-blockdev-create QMP command that can create qcow2 images on a given node name. We don't want to block while creating an image, so this is not the final interface in all aspects, but BlockdevCreateOptionsQcow2 and .bdrv_co_create() are what they actually might look like in the end. In any case, this should be good enough to test whether we interpret BlockdevCreateOptions as we should. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> --- qapi/block-core.json | 12 ++++++++ include/block/block_int.h | 5 +++- block/create.c | 76 +++++++++++++++++++++++++++++++++++++++++++++++ block/qcow2.c | 1 + block/Makefile.objs | 2 +- 5 files changed, 94 insertions(+), 2 deletions(-) create mode 100644 block/create.c 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 @@ } } ## +# @x-blockdev-create: +# +# Create an image format on a given node. +# TODO Replace with something asynchronous (block job?) +# +# Since: 2.12 +## +{ 'command': 'x-blockdev-create', + 'data': 'BlockdevCreateOptions', + 'boxed': true } + +## # @blockdev-open-tray: # # Opens a block device's tray. If there is a block driver state tree inserted as diff --git a/include/block/block_int.h b/include/block/block_int.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block_int.h +++ b/include/block/block_int.h @@ -XXX,XX +XXX,XX @@ struct BlockDriver { int (*bdrv_file_open)(BlockDriverState *bs, QDict *options, int flags, Error **errp); void (*bdrv_close)(BlockDriverState *bs); - int coroutine_fn (*bdrv_co_create_opts)(const char *filename, QemuOpts *opts, + int coroutine_fn (*bdrv_co_create)(BlockdevCreateOptions *opts, Error **errp); + int coroutine_fn (*bdrv_co_create_opts)(const char *filename, + QemuOpts *opts, + Error **errp); int (*bdrv_make_empty)(BlockDriverState *bs); void (*bdrv_refresh_filename)(BlockDriverState *bs, QDict *options); diff --git a/block/create.c b/block/create.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/block/create.c @@ -XXX,XX +XXX,XX @@ +/* + * Block layer code related to image creation + * + * Copyright (c) 2018 Kevin Wolf <kwolf@redhat.com> + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to deal + * in the Software without restriction, including without limitation the rights + * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell + * copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN + * THE SOFTWARE. + */ + +#include "qemu/osdep.h" +#include "block/block_int.h" +#include "qapi/qapi-commands-block-core.h" +#include "qapi/error.h" + +typedef struct BlockdevCreateCo { + BlockDriver *drv; + BlockdevCreateOptions *opts; + int ret; + Error **errp; +} BlockdevCreateCo; + +static void coroutine_fn bdrv_co_create_co_entry(void *opaque) +{ + BlockdevCreateCo *cco = opaque; + cco->ret = cco->drv->bdrv_co_create(cco->opts, cco->errp); +} + +void qmp_x_blockdev_create(BlockdevCreateOptions *options, Error **errp) +{ + const char *fmt = BlockdevDriver_str(options->driver); + BlockDriver *drv = bdrv_find_format(fmt); + Coroutine *co; + BlockdevCreateCo cco; + + /* If the driver is in the schema, we know that it exists. But it may not + * be whitelisted. */ + assert(drv); + if (bdrv_uses_whitelist() && !bdrv_is_whitelisted(drv, false)) { + error_setg(errp, "Driver is not whitelisted"); + return; + } + + /* Call callback if it exists */ + if (!drv->bdrv_co_create) { + error_setg(errp, "Driver does not support blockdev-create"); + return; + } + + cco = (BlockdevCreateCo) { + .drv = drv, + .opts = options, + .ret = -EINPROGRESS, + .errp = errp, + }; + + co = qemu_coroutine_create(bdrv_co_create_co_entry, &cco); + qemu_coroutine_enter(co); + while (cco.ret == -EINPROGRESS) { + aio_poll(qemu_get_aio_context(), true); + } +} 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 @@ BlockDriver bdrv_qcow2 = { .bdrv_join_options = qcow2_join_options, .bdrv_child_perm = bdrv_format_default_perms, .bdrv_co_create_opts = qcow2_co_create_opts, + .bdrv_co_create = qcow2_co_create, .bdrv_has_zero_init = bdrv_has_zero_init_1, .bdrv_co_block_status = qcow2_co_block_status, diff --git a/block/Makefile.objs b/block/Makefile.objs index XXXXXXX..XXXXXXX 100644 --- a/block/Makefile.objs +++ b/block/Makefile.objs @@ -XXX,XX +XXX,XX @@ block-obj-y += block-backend.o snapshot.o qapi.o block-obj-$(CONFIG_WIN32) += file-win32.o win32-aio.o block-obj-$(CONFIG_POSIX) += file-posix.o block-obj-$(CONFIG_LINUX_AIO) += linux-aio.o -block-obj-y += null.o mirror.o commit.o io.o +block-obj-y += null.o mirror.o commit.o io.o create.o block-obj-y += throttle-groups.o block-obj-$(CONFIG_LINUX) += nvme.o -- 2.13.6 This adds the .bdrv_co_create driver callback to file, which enables image creation over QMP. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> --- qapi/block-core.json | 20 ++++++++++++- block/file-posix.c | 79 +++++++++++++++++++++++++++++++++++++--------------- 2 files changed, 75 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 @@ { 'command': 'blockdev-del', 'data': { 'node-name': 'str' } } ## +# @BlockdevCreateOptionsFile: +# +# Driver specific image creation options for file. +# +# @filename Filename for the new image file +# @size Size of the virtual disk in bytes +# @preallocation Preallocation mode for the new image (default: off) +# @nocow Turn off copy-on-write (valid only on btrfs; default: off) +# +# Since: 2.12 +## +{ 'struct': 'BlockdevCreateOptionsFile', + 'data': { 'filename': 'str', + 'size': 'size', + '*preallocation': 'PreallocMode', + '*nocow': 'bool' } } + +## # @BlockdevQcow2Version: # # @v2: The original QCOW2 format as introduced in qemu 0.10 (version 2) @@ -XXX,XX +XXX,XX @@ 'bochs': 'BlockdevCreateNotSupported', 'cloop': 'BlockdevCreateNotSupported', 'dmg': 'BlockdevCreateNotSupported', - 'file': 'BlockdevCreateNotSupported', + 'file': 'BlockdevCreateOptionsFile', 'ftp': 'BlockdevCreateNotSupported', 'ftps': 'BlockdevCreateNotSupported', 'gluster': 'BlockdevCreateNotSupported', 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 raw_get_allocated_file_size(BlockDriverState *bs) return (int64_t)st.st_blocks * 512; } -static int coroutine_fn raw_co_create_opts(const char *filename, QemuOpts *opts, - Error **errp) +static int raw_co_create(BlockdevCreateOptions *options, Error **errp) { + BlockdevCreateOptionsFile *file_opts; int fd; int result = 0; - int64_t total_size = 0; - bool nocow = false; - PreallocMode prealloc; - char *buf = NULL; - Error *local_err = NULL; - strstart(filename, "file:", &filename); + /* Validate options and set default values */ + assert(options->driver == BLOCKDEV_DRIVER_FILE); + file_opts = &options->u.file; - /* Read out options */ - total_size = ROUND_UP(qemu_opt_get_size_del(opts, BLOCK_OPT_SIZE, 0), - BDRV_SECTOR_SIZE); - nocow = qemu_opt_get_bool(opts, BLOCK_OPT_NOCOW, false); - buf = qemu_opt_get_del(opts, BLOCK_OPT_PREALLOC); - prealloc = qapi_enum_parse(&PreallocMode_lookup, buf, - PREALLOC_MODE_OFF, &local_err); - g_free(buf); - if (local_err) { - error_propagate(errp, local_err); - result = -EINVAL; - goto out; + if (!file_opts->has_nocow) { + file_opts->nocow = false; + } + if (!file_opts->has_preallocation) { + file_opts->preallocation = PREALLOC_MODE_OFF; } - fd = qemu_open(filename, O_RDWR | O_CREAT | O_TRUNC | O_BINARY, + /* Create file */ + fd = qemu_open(file_opts->filename, O_RDWR | O_CREAT | O_TRUNC | O_BINARY, 0644); if (fd < 0) { result = -errno; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn raw_co_create_opts(const char *filename, QemuOpts *opts, goto out; } - if (nocow) { + if (file_opts->nocow) { #ifdef __linux__ /* Set NOCOW flag to solve performance issue on fs like btrfs. * This is an optimisation. The FS_IOC_SETFLAGS ioctl return value @@ -XXX,XX +XXX,XX @@ static int coroutine_fn raw_co_create_opts(const char *filename, QemuOpts *opts, #endif } - result = raw_regular_truncate(fd, total_size, prealloc, errp); + result = raw_regular_truncate(fd, file_opts->size, file_opts->preallocation, + errp); if (result < 0) { goto out_close; } @@ -XXX,XX +XXX,XX @@ out: return result; } +static int coroutine_fn raw_co_create_opts(const char *filename, QemuOpts *opts, + Error **errp) +{ + BlockdevCreateOptions options; + int64_t total_size = 0; + bool nocow = false; + PreallocMode prealloc; + char *buf = NULL; + Error *local_err = NULL; + + /* Skip file: protocol prefix */ + strstart(filename, "file:", &filename); + + /* Read out options */ + total_size = ROUND_UP(qemu_opt_get_size_del(opts, BLOCK_OPT_SIZE, 0), + BDRV_SECTOR_SIZE); + nocow = qemu_opt_get_bool(opts, BLOCK_OPT_NOCOW, false); + buf = qemu_opt_get_del(opts, BLOCK_OPT_PREALLOC); + prealloc = qapi_enum_parse(&PreallocMode_lookup, buf, + PREALLOC_MODE_OFF, &local_err); + g_free(buf); + if (local_err) { + error_propagate(errp, local_err); + return -EINVAL; + } + + options = (BlockdevCreateOptions) { + .driver = BLOCKDEV_DRIVER_FILE, + .u.file = { + .filename = (char *) filename, + .size = total_size, + .has_preallocation = true, + .preallocation = prealloc, + .has_nocow = true, + .nocow = nocow, + }, + }; + return raw_co_create(&options, errp); +} + /* * Find allocation range in @bs around offset @start. * May change underlying file descriptor's file offset. @@ -XXX,XX +XXX,XX @@ BlockDriver bdrv_file = { .bdrv_reopen_commit = raw_reopen_commit, .bdrv_reopen_abort = raw_reopen_abort, .bdrv_close = raw_close, + .bdrv_co_create = raw_co_create, .bdrv_co_create_opts = raw_co_create_opts, .bdrv_has_zero_init = bdrv_has_zero_init_1, .bdrv_co_block_status = raw_co_block_status, -- 2.13.6 This adds the .bdrv_co_create driver callback to file-win32, which enables image creation over QMP. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> --- block/file-win32.c | 47 ++++++++++++++++++++++++++++++++++++++--------- 1 file changed, 38 insertions(+), 9 deletions(-) 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 raw_get_allocated_file_size(BlockDriverState *bs) return st.st_size; } -static int coroutine_fn raw_co_create_opts(const char *filename, QemuOpts *opts, - Error **errp) +static int raw_co_create(BlockdevCreateOptions *options, Error **errp) { + BlockdevCreateOptionsFile *file_opts; int fd; - int64_t total_size = 0; - strstart(filename, "file:", &filename); + assert(options->driver == BLOCKDEV_DRIVER_FILE); + file_opts = &options->u.file; - /* Read out options */ - total_size = ROUND_UP(qemu_opt_get_size_del(opts, BLOCK_OPT_SIZE, 0), - BDRV_SECTOR_SIZE); + if (file_opts->has_preallocation) { + error_setg(errp, "Preallocation is not supported on Windows"); + return -EINVAL; + } + if (file_opts->has_nocow) { + error_setg(errp, "nocow is not supported on Windows"); + return -EINVAL; + } - fd = qemu_open(filename, O_WRONLY | O_CREAT | O_TRUNC | O_BINARY, + fd = qemu_open(file_opts->filename, O_WRONLY | O_CREAT | O_TRUNC | O_BINARY, 0644); if (fd < 0) { error_setg_errno(errp, errno, "Could not create file"); return -EIO; } set_sparse(fd); - ftruncate(fd, total_size); + ftruncate(fd, file_opts->size); qemu_close(fd); + return 0; } +static int coroutine_fn raw_co_create_opts(const char *filename, QemuOpts *opts, + Error **errp) +{ + BlockdevCreateOptions options; + int64_t total_size = 0; + + strstart(filename, "file:", &filename); + + /* Read out options */ + total_size = ROUND_UP(qemu_opt_get_size_del(opts, BLOCK_OPT_SIZE, 0), + BDRV_SECTOR_SIZE); + + options = (BlockdevCreateOptions) { + .driver = BLOCKDEV_DRIVER_FILE, + .u.file = { + .filename = (char *) filename, + .size = total_size, + .has_preallocation = false, + .has_nocow = false, + }, + }; + return raw_co_create(&options, errp); +} static QemuOptsList raw_create_opts = { .name = "raw-create-opts", -- 2.13.6 This adds the .bdrv_co_create driver callback to gluster, which enables image creation over QMP. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> --- qapi/block-core.json | 18 ++++++- block/gluster.c | 135 ++++++++++++++++++++++++++++++++++----------------- 2 files changed, 108 insertions(+), 45 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 @@ '*nocow': 'bool' } } ## +# @BlockdevCreateOptionsGluster: +# +# Driver specific image creation options for gluster. +# +# @location Where to store the new image file +# @size Size of the virtual disk in bytes +# @preallocation Preallocation mode for the new image (default: off) +# +# Since: 2.12 +## +{ 'struct': 'BlockdevCreateOptionsGluster', + 'data': { 'location': 'BlockdevOptionsGluster', + 'size': 'size', + '*preallocation': 'PreallocMode' } } + +## # @BlockdevQcow2Version: # # @v2: The original QCOW2 format as introduced in qemu 0.10 (version 2) @@ -XXX,XX +XXX,XX @@ 'file': 'BlockdevCreateOptionsFile', 'ftp': 'BlockdevCreateNotSupported', 'ftps': 'BlockdevCreateNotSupported', - 'gluster': 'BlockdevCreateNotSupported', + 'gluster': 'BlockdevCreateOptionsGluster', 'host_cdrom': 'BlockdevCreateNotSupported', 'host_device': 'BlockdevCreateNotSupported', 'http': 'BlockdevCreateNotSupported', 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 @@ out: return -errno; } -static struct glfs *qemu_gluster_init(BlockdevOptionsGluster *gconf, - const char *filename, - QDict *options, Error **errp) +/* Converts options given in @filename and the @options QDict into the QAPI + * object @gconf. */ +static int qemu_gluster_parse(BlockdevOptionsGluster *gconf, + const char *filename, + QDict *options, Error **errp) { int ret; if (filename) { @@ -XXX,XX +XXX,XX @@ static struct glfs *qemu_gluster_init(BlockdevOptionsGluster *gconf, "[host[:port]]volume/path[?socket=...]" "[,file.debug=N]" "[,file.logfile=/path/filename.log]\n"); - errno = -ret; - return NULL; + return ret; } } else { ret = qemu_gluster_parse_json(gconf, options, errp); @@ -XXX,XX +XXX,XX @@ static struct glfs *qemu_gluster_init(BlockdevOptionsGluster *gconf, "file.server.1.transport=unix," "file.server.1.socket=/var/run/glusterd.socket ..." "\n"); - errno = -ret; - return NULL; + return ret; } + } + return 0; +} + +static struct glfs *qemu_gluster_init(BlockdevOptionsGluster *gconf, + const char *filename, + QDict *options, Error **errp) +{ + int ret; + + ret = qemu_gluster_parse(gconf, filename, options, errp); + if (ret < 0) { + errno = -ret; + return NULL; } return qemu_gluster_glfs_init(gconf, errp); @@ -XXX,XX +XXX,XX @@ static int qemu_gluster_do_truncate(struct glfs_fd *fd, int64_t offset, return 0; } +static int qemu_gluster_co_create(BlockdevCreateOptions *options, + Error **errp) +{ + BlockdevCreateOptionsGluster *opts = &options->u.gluster; + struct glfs *glfs; + struct glfs_fd *fd = NULL; + int ret = 0; + + assert(options->driver == BLOCKDEV_DRIVER_GLUSTER); + + glfs = qemu_gluster_glfs_init(opts->location, errp); + if (!glfs) { + ret = -errno; + goto out; + } + + fd = glfs_creat(glfs, opts->location->path, + O_WRONLY | O_CREAT | O_TRUNC | O_BINARY, S_IRUSR | S_IWUSR); + if (!fd) { + ret = -errno; + goto out; + } + + ret = qemu_gluster_do_truncate(fd, opts->size, opts->preallocation, errp); + +out: + if (fd) { + if (glfs_close(fd) != 0 && ret == 0) { + ret = -errno; + } + } + glfs_clear_preopened(glfs); + return ret; +} + static int coroutine_fn qemu_gluster_co_create_opts(const char *filename, QemuOpts *opts, Error **errp) { + BlockdevCreateOptions *options; + BlockdevCreateOptionsGluster *gopts; BlockdevOptionsGluster *gconf; - struct glfs *glfs; - struct glfs_fd *fd = NULL; - int ret = 0; - PreallocMode prealloc; - int64_t total_size = 0; char *tmp = NULL; Error *local_err = NULL; + int ret; + + options = g_new0(BlockdevCreateOptions, 1); + options->driver = BLOCKDEV_DRIVER_GLUSTER; + gopts = &options->u.gluster; gconf = g_new0(BlockdevOptionsGluster, 1); + gopts->location = gconf; + + gopts->size = ROUND_UP(qemu_opt_get_size_del(opts, BLOCK_OPT_SIZE, 0), + BDRV_SECTOR_SIZE); + + tmp = qemu_opt_get_del(opts, BLOCK_OPT_PREALLOC); + gopts->preallocation = qapi_enum_parse(&PreallocMode_lookup, tmp, + PREALLOC_MODE_OFF, &local_err); + g_free(tmp); + if (local_err) { + error_propagate(errp, local_err); + ret = -EINVAL; + goto fail; + } + gconf->debug = qemu_opt_get_number_del(opts, GLUSTER_OPT_DEBUG, GLUSTER_DEBUG_DEFAULT); if (gconf->debug < 0) { @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qemu_gluster_co_create_opts(const char *filename, } gconf->has_logfile = true; - glfs = qemu_gluster_init(gconf, filename, NULL, errp); - if (!glfs) { - ret = -errno; - goto out; - } - - total_size = ROUND_UP(qemu_opt_get_size_del(opts, BLOCK_OPT_SIZE, 0), - BDRV_SECTOR_SIZE); - - tmp = qemu_opt_get_del(opts, BLOCK_OPT_PREALLOC); - prealloc = qapi_enum_parse(&PreallocMode_lookup, tmp, PREALLOC_MODE_OFF, - &local_err); - g_free(tmp); - if (local_err) { - error_propagate(errp, local_err); - ret = -EINVAL; - goto out; + ret = qemu_gluster_parse(gconf, filename, NULL, errp); + if (ret < 0) { + goto fail; } - fd = glfs_creat(glfs, gconf->path, - O_WRONLY | O_CREAT | O_TRUNC | O_BINARY, S_IRUSR | S_IWUSR); - if (!fd) { - ret = -errno; - goto out; + ret = qemu_gluster_co_create(options, errp); + if (ret < 0) { + goto fail; } - ret = qemu_gluster_do_truncate(fd, total_size, prealloc, errp); - -out: - if (fd) { - if (glfs_close(fd) != 0 && ret == 0) { - ret = -errno; - } - } - qapi_free_BlockdevOptionsGluster(gconf); - glfs_clear_preopened(glfs); + ret = 0; +fail: + qapi_free_BlockdevCreateOptions(options); return ret; } @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_gluster = { .bdrv_reopen_commit = qemu_gluster_reopen_commit, .bdrv_reopen_abort = qemu_gluster_reopen_abort, .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_get_allocated_file_size = qemu_gluster_allocated_file_size, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_gluster_tcp = { .bdrv_reopen_commit = qemu_gluster_reopen_commit, .bdrv_reopen_abort = qemu_gluster_reopen_abort, .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_get_allocated_file_size = qemu_gluster_allocated_file_size, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_gluster_unix = { .bdrv_reopen_commit = qemu_gluster_reopen_commit, .bdrv_reopen_abort = qemu_gluster_reopen_abort, .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_get_allocated_file_size = qemu_gluster_allocated_file_size, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_gluster_rdma = { .bdrv_reopen_commit = qemu_gluster_reopen_commit, .bdrv_reopen_abort = qemu_gluster_reopen_abort, .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_get_allocated_file_size = qemu_gluster_allocated_file_size, -- 2.13.6 If we want to include the invalid option name in the error message, we can't free the string earlier than that. Cc: qemu-stable@nongnu.org Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> --- block/rbd.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) 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_set_keypairs(rados_t cluster, const char *keypairs_json, key = qstring_get_str(name); ret = rados_conf_set(cluster, key, qstring_get_str(value)); - QDECREF(name); QDECREF(value); if (ret < 0) { error_setg_errno(errp, -ret, "invalid conf option %s", key); + QDECREF(name); ret = -EINVAL; break; } + QDECREF(name); } QDECREF(keypairs); -- 2.13.6 The code to establish an RBD connection is duplicated between open and create. In order to be able to share the code, factor out the code from qemu_rbd_open() as a first step. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> --- block/rbd.c | 100 ++++++++++++++++++++++++++++++++++++------------------------ 1 file changed, 60 insertions(+), 40 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 @@ out: return rados_str; } -static int qemu_rbd_open(BlockDriverState *bs, QDict *options, int flags, - Error **errp) +static int qemu_rbd_connect(rados_t *cluster, rados_ioctx_t *io_ctx, + char **s_snap, char **s_image_name, + QDict *options, bool cache, Error **errp) { - BDRVRBDState *s = bs->opaque; - const char *pool, *snap, *conf, *user, *image_name, *keypairs; - const char *secretid, *filename; QemuOpts *opts; - Error *local_err = NULL; char *mon_host = NULL; + const char *pool, *snap, *conf, *user, *image_name, *keypairs; + const char *secretid; + Error *local_err = NULL; int r; - /* If we are given a filename, parse the filename, with precedence given to - * filename encoded options */ - filename = qdict_get_try_str(options, "filename"); - if (filename) { - warn_report("'filename' option specified. " - "This is an unsupported option, and may be deprecated " - "in the future"); - qemu_rbd_parse_filename(filename, options, &local_err); - if (local_err) { - r = -EINVAL; - error_propagate(errp, local_err); - goto exit; - } - } - opts = qemu_opts_create(&runtime_opts, NULL, 0, &error_abort); qemu_opts_absorb_qdict(opts, options, &local_err); if (local_err) { @@ -XXX,XX +XXX,XX @@ static int qemu_rbd_open(BlockDriverState *bs, QDict *options, int flags, goto failed_opts; } - r = rados_create(&s->cluster, user); + r = rados_create(cluster, user); if (r < 0) { error_setg_errno(errp, -r, "error initializing"); goto failed_opts; } - s->snap = g_strdup(snap); - s->image_name = g_strdup(image_name); + *s_snap = g_strdup(snap); + *s_image_name = g_strdup(image_name); /* try default location when conf=NULL, but ignore failure */ - r = rados_conf_read_file(s->cluster, conf); + r = rados_conf_read_file(*cluster, conf); if (conf && r < 0) { error_setg_errno(errp, -r, "error reading conf file %s", conf); goto failed_shutdown; } - r = qemu_rbd_set_keypairs(s->cluster, keypairs, errp); + r = qemu_rbd_set_keypairs(*cluster, keypairs, errp); if (r < 0) { goto failed_shutdown; } if (mon_host) { - r = rados_conf_set(s->cluster, "mon_host", mon_host); + r = rados_conf_set(*cluster, "mon_host", mon_host); if (r < 0) { goto failed_shutdown; } } - if (qemu_rbd_set_auth(s->cluster, secretid, errp) < 0) { + if (qemu_rbd_set_auth(*cluster, secretid, errp) < 0) { r = -EIO; goto failed_shutdown; } @@ -XXX,XX +XXX,XX @@ static int qemu_rbd_open(BlockDriverState *bs, QDict *options, int flags, * librbd defaults to no caching. If write through caching cannot * be set up, fall back to no caching. */ - if (flags & BDRV_O_NOCACHE) { - rados_conf_set(s->cluster, "rbd_cache", "false"); + if (cache) { + rados_conf_set(*cluster, "rbd_cache", "true"); } else { - rados_conf_set(s->cluster, "rbd_cache", "true"); + rados_conf_set(*cluster, "rbd_cache", "false"); } - r = rados_connect(s->cluster); + r = rados_connect(*cluster); if (r < 0) { error_setg_errno(errp, -r, "error connecting"); goto failed_shutdown; } - r = rados_ioctx_create(s->cluster, pool, &s->io_ctx); + r = rados_ioctx_create(*cluster, pool, io_ctx); if (r < 0) { error_setg_errno(errp, -r, "error opening pool %s", pool); goto failed_shutdown; } + qemu_opts_del(opts); + return 0; + +failed_shutdown: + rados_shutdown(*cluster); + g_free(*s_snap); + g_free(*s_image_name); +failed_opts: + qemu_opts_del(opts); + g_free(mon_host); + return r; +} + +static int qemu_rbd_open(BlockDriverState *bs, QDict *options, int flags, + Error **errp) +{ + BDRVRBDState *s = bs->opaque; + Error *local_err = NULL; + const char *filename; + int r; + + /* If we are given a filename, parse the filename, with precedence given to + * filename encoded options */ + filename = qdict_get_try_str(options, "filename"); + if (filename) { + warn_report("'filename' option specified. " + "This is an unsupported option, and may be deprecated " + "in the future"); + qemu_rbd_parse_filename(filename, options, &local_err); + if (local_err) { + error_propagate(errp, local_err); + return -EINVAL; + } + } + + r = qemu_rbd_connect(&s->cluster, &s->io_ctx, &s->snap, &s->image_name, + options, !(flags & BDRV_O_NOCACHE), errp); + if (r < 0) { + return r; + } + /* rbd_open is always r/w */ r = rbd_open(s->io_ctx, s->image_name, &s->image, s->snap); if (r < 0) { @@ -XXX,XX +XXX,XX @@ static int qemu_rbd_open(BlockDriverState *bs, QDict *options, int flags, } } - qemu_opts_del(opts); return 0; failed_open: rados_ioctx_destroy(s->io_ctx); -failed_shutdown: - rados_shutdown(s->cluster); g_free(s->snap); g_free(s->image_name); -failed_opts: - qemu_opts_del(opts); - g_free(mon_host); -exit: + rados_shutdown(s->cluster); return r; } -- 2.13.6 Instead of the QemuOpts in qemu_rbd_connect(), we want to use QAPI objects. As a preparation, fetch those options directly from the QDict that .bdrv_open() supports in the rbd driver and that are not in the schema. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> --- block/rbd.c | 55 ++++++++++++++++++++++++------------------------------- 1 file changed, 24 insertions(+), 31 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 QemuOptsList runtime_opts = { /* * server.* extracted manually, see qemu_rbd_mon_host() */ - { - .name = "password-secret", - .type = QEMU_OPT_STRING, - .help = "ID of secret providing the password", - }, - - /* - * Keys for qemu_rbd_parse_filename(), not in the QAPI schema - */ - { - /* - * HACK: name starts with '=' so that qemu_opts_parse() - * can't set it - */ - .name = "=keyvalue-pairs", - .type = QEMU_OPT_STRING, - .help = "Legacy rados key/value option parameters", - }, - { - .name = "filename", - .type = QEMU_OPT_STRING, - }, { /* end of list */ } }, }; @@ -XXX,XX +XXX,XX @@ out: static int qemu_rbd_connect(rados_t *cluster, rados_ioctx_t *io_ctx, char **s_snap, char **s_image_name, - QDict *options, bool cache, Error **errp) + QDict *options, bool cache, + const char *keypairs, const char *secretid, + Error **errp) { QemuOpts *opts; char *mon_host = NULL; - const char *pool, *snap, *conf, *user, *image_name, *keypairs; - const char *secretid; + const char *pool, *snap, *conf, *user, *image_name; Error *local_err = NULL; int r; @@ -XXX,XX +XXX,XX @@ static int qemu_rbd_connect(rados_t *cluster, rados_ioctx_t *io_ctx, goto failed_opts; } - secretid = qemu_opt_get(opts, "password-secret"); - pool = qemu_opt_get(opts, "pool"); conf = qemu_opt_get(opts, "conf"); snap = qemu_opt_get(opts, "snapshot"); user = qemu_opt_get(opts, "user"); image_name = qemu_opt_get(opts, "image"); - keypairs = qemu_opt_get(opts, "=keyvalue-pairs"); if (!pool || !image_name) { error_setg(errp, "Parameters 'pool' and 'image' are required"); @@ -XXX,XX +XXX,XX @@ static int qemu_rbd_open(BlockDriverState *bs, QDict *options, int flags, BDRVRBDState *s = bs->opaque; Error *local_err = NULL; const char *filename; + char *keypairs, *secretid; int r; /* If we are given a filename, parse the filename, with precedence given to @@ -XXX,XX +XXX,XX @@ static int qemu_rbd_open(BlockDriverState *bs, QDict *options, int flags, "This is an unsupported option, and may be deprecated " "in the future"); qemu_rbd_parse_filename(filename, options, &local_err); + qdict_del(options, "filename"); if (local_err) { error_propagate(errp, local_err); return -EINVAL; } } + keypairs = g_strdup(qdict_get_try_str(options, "=keyvalue-pairs")); + if (keypairs) { + qdict_del(options, "=keyvalue-pairs"); + } + + secretid = g_strdup(qdict_get_try_str(options, "password-secret")); + if (secretid) { + qdict_del(options, "password-secret"); + } + r = qemu_rbd_connect(&s->cluster, &s->io_ctx, &s->snap, &s->image_name, - options, !(flags & BDRV_O_NOCACHE), errp); + options, !(flags & BDRV_O_NOCACHE), keypairs, secretid, + errp); if (r < 0) { - return r; + goto out; } /* rbd_open is always r/w */ @@ -XXX,XX +XXX,XX @@ static int qemu_rbd_open(BlockDriverState *bs, QDict *options, int flags, } } - return 0; + r = 0; + goto out; failed_open: rados_ioctx_destroy(s->io_ctx); g_free(s->snap); g_free(s->image_name); rados_shutdown(s->cluster); +out: + g_free(keypairs); + g_free(secretid); return r; } -- 2.13.6 With the conversion to a QAPI options object, the function is now prepared to be used in a .bdrv_co_create implementation. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> --- block/rbd.c | 115 +++++++++++++++++++++++++++++------------------------------- 1 file changed, 55 insertions(+), 60 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 @@ #include "qapi/qmp/qdict.h" #include "qapi/qmp/qjson.h" #include "qapi/qmp/qlist.h" +#include "qapi/qobject-input-visitor.h" +#include "qapi/qapi-visit-block-core.h" /* * When specifying the image filename use: @@ -XXX,XX +XXX,XX @@ static void qemu_rbd_complete_aio(RADOSCB *rcb) qemu_aio_unref(acb); } -static char *qemu_rbd_mon_host(QDict *options, Error **errp) +static char *qemu_rbd_mon_host(BlockdevOptionsRbd *opts, Error **errp) { - const char **vals = g_new(const char *, qdict_size(options) + 1); - char keybuf[32]; + const char **vals; const char *host, *port; char *rados_str; - int i; - - for (i = 0;; i++) { - sprintf(keybuf, "server.%d.host", i); - host = qdict_get_try_str(options, keybuf); - qdict_del(options, keybuf); - sprintf(keybuf, "server.%d.port", i); - port = qdict_get_try_str(options, keybuf); - qdict_del(options, keybuf); - if (!host && !port) { - break; - } - if (!host) { - error_setg(errp, "Parameter server.%d.host is missing", i); - rados_str = NULL; - goto out; - } + InetSocketAddressBaseList *p; + int i, cnt; + + if (!opts->has_server) { + return NULL; + } + + for (cnt = 0, p = opts->server; p; p = p->next) { + cnt++; + } + + vals = g_new(const char *, cnt + 1); + + for (i = 0, p = opts->server; p; p = p->next, i++) { + host = p->value->host; + port = p->value->port; if (strchr(host, ':')) { - vals[i] = port ? g_strdup_printf("[%s]:%s", host, port) - : g_strdup_printf("[%s]", host); + vals[i] = g_strdup_printf("[%s]:%s", host, port); } else { - vals[i] = port ? g_strdup_printf("%s:%s", host, port) - : g_strdup(host); + vals[i] = g_strdup_printf("%s:%s", host, port); } } vals[i] = NULL; rados_str = i ? g_strjoinv(";", (char **)vals) : NULL; -out: g_strfreev((char **)vals); return rados_str; } static int qemu_rbd_connect(rados_t *cluster, rados_ioctx_t *io_ctx, char **s_snap, char **s_image_name, - QDict *options, bool cache, + BlockdevOptionsRbd *opts, bool cache, const char *keypairs, const char *secretid, Error **errp) { - QemuOpts *opts; char *mon_host = NULL; - const char *pool, *snap, *conf, *user, *image_name; Error *local_err = NULL; int r; - opts = qemu_opts_create(&runtime_opts, NULL, 0, &error_abort); - qemu_opts_absorb_qdict(opts, options, &local_err); - if (local_err) { - error_propagate(errp, local_err); - r = -EINVAL; - goto failed_opts; - } - - mon_host = qemu_rbd_mon_host(options, &local_err); + mon_host = qemu_rbd_mon_host(opts, &local_err); if (local_err) { error_propagate(errp, local_err); r = -EINVAL; goto failed_opts; } - pool = qemu_opt_get(opts, "pool"); - conf = qemu_opt_get(opts, "conf"); - snap = qemu_opt_get(opts, "snapshot"); - user = qemu_opt_get(opts, "user"); - image_name = qemu_opt_get(opts, "image"); - - if (!pool || !image_name) { - error_setg(errp, "Parameters 'pool' and 'image' are required"); - r = -EINVAL; - goto failed_opts; - } - - r = rados_create(cluster, user); + r = rados_create(cluster, opts->user); if (r < 0) { error_setg_errno(errp, -r, "error initializing"); goto failed_opts; } - *s_snap = g_strdup(snap); - *s_image_name = g_strdup(image_name); + *s_snap = g_strdup(opts->snapshot); + *s_image_name = g_strdup(opts->image); /* try default location when conf=NULL, but ignore failure */ - r = rados_conf_read_file(*cluster, conf); - if (conf && r < 0) { - error_setg_errno(errp, -r, "error reading conf file %s", conf); + r = rados_conf_read_file(*cluster, opts->conf); + if (opts->has_conf && r < 0) { + error_setg_errno(errp, -r, "error reading conf file %s", opts->conf); goto failed_shutdown; } @@ -XXX,XX +XXX,XX @@ static int qemu_rbd_connect(rados_t *cluster, rados_ioctx_t *io_ctx, goto failed_shutdown; } - r = rados_ioctx_create(*cluster, pool, io_ctx); + r = rados_ioctx_create(*cluster, opts->pool, io_ctx); if (r < 0) { - error_setg_errno(errp, -r, "error opening pool %s", pool); + error_setg_errno(errp, -r, "error opening pool %s", opts->pool); goto failed_shutdown; } - qemu_opts_del(opts); return 0; failed_shutdown: @@ -XXX,XX +XXX,XX @@ failed_shutdown: g_free(*s_snap); g_free(*s_image_name); failed_opts: - qemu_opts_del(opts); g_free(mon_host); return r; } @@ -XXX,XX +XXX,XX @@ static int qemu_rbd_open(BlockDriverState *bs, QDict *options, int flags, Error **errp) { BDRVRBDState *s = bs->opaque; + BlockdevOptionsRbd *opts = NULL; + Visitor *v; + QObject *crumpled = NULL; Error *local_err = NULL; const char *filename; char *keypairs, *secretid; @@ -XXX,XX +XXX,XX @@ static int qemu_rbd_open(BlockDriverState *bs, QDict *options, int flags, qdict_del(options, "password-secret"); } + /* Convert the remaining options into a QAPI object */ + crumpled = qdict_crumple(options, errp); + if (crumpled == NULL) { + r = -EINVAL; + goto out; + } + + v = qobject_input_visitor_new_keyval(crumpled); + visit_type_BlockdevOptionsRbd(v, NULL, &opts, &local_err); + visit_free(v); + qobject_decref(crumpled); + + if (local_err) { + error_propagate(errp, local_err); + r = -EINVAL; + goto out; + } + r = qemu_rbd_connect(&s->cluster, &s->io_ctx, &s->snap, &s->image_name, - options, !(flags & BDRV_O_NOCACHE), keypairs, secretid, + opts, !(flags & BDRV_O_NOCACHE), keypairs, secretid, errp); if (r < 0) { goto out; @@ -XXX,XX +XXX,XX @@ failed_open: g_free(s->image_name); rados_shutdown(s->cluster); out: + qapi_free_BlockdevOptionsRbd(opts); g_free(keypairs); g_free(secretid); return r; -- 2.13.6 This adds the .bdrv_co_create driver callback to rbd, which enables image creation over QMP. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> --- qapi/block-core.json | 19 ++++++- block/rbd.c | 150 ++++++++++++++++++++++++++++++++++----------------- 2 files changed, 118 insertions(+), 51 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 @@ '*refcount-bits': 'int' } } ## +# @BlockdevCreateOptionsRbd: +# +# Driver specific image creation options for rbd/Ceph. +# +# @location Where to store the new image file. This location cannot +# point to a snapshot. +# @size Size of the virtual disk in bytes +# @cluster-size RBD object size +# +# Since: 2.12 +## +{ 'struct': 'BlockdevCreateOptionsRbd', + 'data': { 'location': 'BlockdevOptionsRbd', + 'size': 'size', + '*cluster-size' : 'size' } } + +## # @BlockdevCreateNotSupported: # # This is used for all drivers that don't support creating images. @@ -XXX,XX +XXX,XX @@ 'qed': 'BlockdevCreateNotSupported', 'quorum': 'BlockdevCreateNotSupported', 'raw': 'BlockdevCreateNotSupported', - 'rbd': 'BlockdevCreateNotSupported', + 'rbd': 'BlockdevCreateOptionsRbd', 'replication': 'BlockdevCreateNotSupported', 'sheepdog': 'BlockdevCreateNotSupported', 'ssh': 'BlockdevCreateNotSupported', 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 QemuOptsList runtime_opts = { }, }; -static int coroutine_fn qemu_rbd_co_create_opts(const char *filename, - QemuOpts *opts, - Error **errp) +/* FIXME Deprecate and remove keypairs or make it available in QMP. + * password_secret should eventually be configurable in opts->location. Support + * for it in .bdrv_open will make it work here as well. */ +static int qemu_rbd_do_create(BlockdevCreateOptions *options, + const char *keypairs, const char *password_secret, + Error **errp) { - Error *local_err = NULL; - int64_t bytes = 0; - int64_t objsize; - int obj_order = 0; - const char *pool, *image_name, *conf, *user, *keypairs; - const char *secretid; + BlockdevCreateOptionsRbd *opts = &options->u.rbd; rados_t cluster; rados_ioctx_t io_ctx; - QDict *options = NULL; - int ret = 0; + int obj_order = 0; + int ret; + + assert(options->driver == BLOCKDEV_DRIVER_RBD); + if (opts->location->has_snapshot) { + error_setg(errp, "Can't use snapshot name for image creation"); + return -EINVAL; + } - secretid = qemu_opt_get(opts, "password-secret"); + /* TODO Remove the limitation */ + if (opts->location->has_server) { + error_setg(errp, "Can't specify server for image creation"); + return -EINVAL; + } - /* Read out options */ - bytes = ROUND_UP(qemu_opt_get_size_del(opts, BLOCK_OPT_SIZE, 0), - BDRV_SECTOR_SIZE); - objsize = qemu_opt_get_size_del(opts, BLOCK_OPT_CLUSTER_SIZE, 0); - if (objsize) { + if (opts->has_cluster_size) { + int64_t objsize = opts->cluster_size; if ((objsize - 1) & objsize) { /* not a power of 2? */ error_setg(errp, "obj size needs to be power of 2"); - ret = -EINVAL; - goto exit; + return -EINVAL; } if (objsize < 4096) { error_setg(errp, "obj size too small"); - ret = -EINVAL; - goto exit; + return -EINVAL; } obj_order = ctz32(objsize); } - options = qdict_new(); - qemu_rbd_parse_filename(filename, options, &local_err); - if (local_err) { - ret = -EINVAL; - error_propagate(errp, local_err); - goto exit; - } - - /* - * Caution: while qdict_get_try_str() is fine, getting non-string - * types would require more care. When @options come from -blockdev - * or blockdev_add, its members are typed according to the QAPI - * schema, but when they come from -drive, they're all QString. - */ - pool = qdict_get_try_str(options, "pool"); - conf = qdict_get_try_str(options, "conf"); - user = qdict_get_try_str(options, "user"); - image_name = qdict_get_try_str(options, "image"); - keypairs = qdict_get_try_str(options, "=keyvalue-pairs"); - - ret = rados_create(&cluster, user); + ret = rados_create(&cluster, opts->location->user); if (ret < 0) { error_setg_errno(errp, -ret, "error initializing"); - goto exit; + return ret; } /* try default location when conf=NULL, but ignore failure */ - ret = rados_conf_read_file(cluster, conf); - if (conf && ret < 0) { - error_setg_errno(errp, -ret, "error reading conf file %s", conf); + ret = rados_conf_read_file(cluster, opts->location->conf); + if (opts->location->conf && ret < 0) { + error_setg_errno(errp, -ret, "error reading conf file %s", + opts->location->conf); ret = -EIO; goto shutdown; } @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qemu_rbd_co_create_opts(const char *filename, goto shutdown; } - if (qemu_rbd_set_auth(cluster, secretid, errp) < 0) { + if (qemu_rbd_set_auth(cluster, password_secret, errp) < 0) { ret = -EIO; goto shutdown; } @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qemu_rbd_co_create_opts(const char *filename, goto shutdown; } - ret = rados_ioctx_create(cluster, pool, &io_ctx); + ret = rados_ioctx_create(cluster, opts->location->pool, &io_ctx); if (ret < 0) { - error_setg_errno(errp, -ret, "error opening pool %s", pool); + error_setg_errno(errp, -ret, "error opening pool %s", + opts->location->pool); goto shutdown; } - ret = rbd_create(io_ctx, image_name, bytes, &obj_order); + ret = rbd_create(io_ctx, opts->location->image, opts->size, &obj_order); if (ret < 0) { error_setg_errno(errp, -ret, "error rbd create"); } rados_ioctx_destroy(io_ctx); + ret = 0; shutdown: rados_shutdown(cluster); + return ret; +} + +static int qemu_rbd_co_create(BlockdevCreateOptions *options, Error **errp) +{ + return qemu_rbd_do_create(options, NULL, NULL, errp); +} + +static int coroutine_fn qemu_rbd_co_create_opts(const char *filename, + QemuOpts *opts, + Error **errp) +{ + BlockdevCreateOptions *create_options; + BlockdevCreateOptionsRbd *rbd_opts; + BlockdevOptionsRbd *loc; + Error *local_err = NULL; + const char *keypairs, *password_secret; + QDict *options = NULL; + int ret = 0; + + create_options = g_new0(BlockdevCreateOptions, 1); + create_options->driver = BLOCKDEV_DRIVER_RBD; + rbd_opts = &create_options->u.rbd; + + rbd_opts->location = g_new0(BlockdevOptionsRbd, 1); + + password_secret = qemu_opt_get(opts, "password-secret"); + + /* Read out options */ + rbd_opts->size = ROUND_UP(qemu_opt_get_size_del(opts, BLOCK_OPT_SIZE, 0), + BDRV_SECTOR_SIZE); + rbd_opts->cluster_size = qemu_opt_get_size_del(opts, + BLOCK_OPT_CLUSTER_SIZE, 0); + rbd_opts->has_cluster_size = (rbd_opts->cluster_size != 0); + + options = qdict_new(); + qemu_rbd_parse_filename(filename, options, &local_err); + if (local_err) { + ret = -EINVAL; + error_propagate(errp, local_err); + goto exit; + } + + /* + * Caution: while qdict_get_try_str() is fine, getting non-string + * types would require more care. When @options come from -blockdev + * or blockdev_add, its members are typed according to the QAPI + * schema, but when they come from -drive, they're all QString. + */ + loc = rbd_opts->location; + loc->pool = g_strdup(qdict_get_try_str(options, "pool")); + loc->conf = g_strdup(qdict_get_try_str(options, "conf")); + loc->has_conf = !!loc->conf; + loc->user = g_strdup(qdict_get_try_str(options, "user")); + loc->has_user = !!loc->user; + loc->image = g_strdup(qdict_get_try_str(options, "image")); + keypairs = qdict_get_try_str(options, "=keyvalue-pairs"); + + ret = qemu_rbd_do_create(create_options, keypairs, password_secret, errp); + if (ret < 0) { + goto exit; + } exit: QDECREF(options); + qapi_free_BlockdevCreateOptions(create_options); return ret; } @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_rbd = { .bdrv_file_open = qemu_rbd_open, .bdrv_close = qemu_rbd_close, .bdrv_reopen_prepare = qemu_rbd_reopen_prepare, + .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, -- 2.13.6 Now that the options are already available in qemu_rbd_open() and not only parsed in qemu_rbd_connect(), we can assign s->snap and s->image_name there instead of passing the fields by reference to qemu_rbd_connect(). Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> --- block/rbd.c | 14 +++++--------- 1 file changed, 5 insertions(+), 9 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 char *qemu_rbd_mon_host(BlockdevOptionsRbd *opts, Error **errp) } static int qemu_rbd_connect(rados_t *cluster, rados_ioctx_t *io_ctx, - char **s_snap, char **s_image_name, BlockdevOptionsRbd *opts, bool cache, const char *keypairs, const char *secretid, Error **errp) @@ -XXX,XX +XXX,XX @@ static int qemu_rbd_connect(rados_t *cluster, rados_ioctx_t *io_ctx, goto failed_opts; } - *s_snap = g_strdup(opts->snapshot); - *s_image_name = g_strdup(opts->image); - /* try default location when conf=NULL, but ignore failure */ r = rados_conf_read_file(*cluster, opts->conf); if (opts->has_conf && r < 0) { @@ -XXX,XX +XXX,XX @@ static int qemu_rbd_connect(rados_t *cluster, rados_ioctx_t *io_ctx, failed_shutdown: rados_shutdown(*cluster); - g_free(*s_snap); - g_free(*s_image_name); failed_opts: g_free(mon_host); return r; @@ -XXX,XX +XXX,XX @@ static int qemu_rbd_open(BlockDriverState *bs, QDict *options, int flags, goto out; } - r = qemu_rbd_connect(&s->cluster, &s->io_ctx, &s->snap, &s->image_name, - opts, !(flags & BDRV_O_NOCACHE), keypairs, secretid, - errp); + r = qemu_rbd_connect(&s->cluster, &s->io_ctx, opts, + !(flags & BDRV_O_NOCACHE), keypairs, secretid, errp); if (r < 0) { goto out; } + s->snap = g_strdup(opts->snapshot); + s->image_name = g_strdup(opts->image); + /* rbd_open is always r/w */ r = rbd_open(s->io_ctx, s->image_name, &s->image, s->snap); if (r < 0) { -- 2.13.6 This is almost exactly the same code. The differences are that qemu_rbd_connect() supports BlockdevOptionsRbd.server and that the cache mode is set explicitly. Supporting 'server' is a welcome new feature for image creation. Caching is disabled by default, so leave it that way. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> --- block/rbd.c | 54 ++++++++++-------------------------------------------- 1 file changed, 10 insertions(+), 44 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 @@ typedef struct BDRVRBDState { char *snap; } BDRVRBDState; +static int qemu_rbd_connect(rados_t *cluster, rados_ioctx_t *io_ctx, + BlockdevOptionsRbd *opts, bool cache, + const char *keypairs, const char *secretid, + Error **errp); + static char *qemu_rbd_next_tok(char *src, char delim, char **p) { char *end; @@ -XXX,XX +XXX,XX @@ static int qemu_rbd_do_create(BlockdevCreateOptions *options, return -EINVAL; } - /* TODO Remove the limitation */ - if (opts->location->has_server) { - error_setg(errp, "Can't specify server for image creation"); - return -EINVAL; - } - if (opts->has_cluster_size) { int64_t objsize = opts->cluster_size; if ((objsize - 1) & objsize) { /* not a power of 2? */ @@ -XXX,XX +XXX,XX @@ static int qemu_rbd_do_create(BlockdevCreateOptions *options, obj_order = ctz32(objsize); } - ret = rados_create(&cluster, opts->location->user); + ret = qemu_rbd_connect(&cluster, &io_ctx, opts->location, false, keypairs, + password_secret, errp); if (ret < 0) { - error_setg_errno(errp, -ret, "error initializing"); return ret; } - /* try default location when conf=NULL, but ignore failure */ - ret = rados_conf_read_file(cluster, opts->location->conf); - if (opts->location->conf && ret < 0) { - error_setg_errno(errp, -ret, "error reading conf file %s", - opts->location->conf); - ret = -EIO; - goto shutdown; - } - - ret = qemu_rbd_set_keypairs(cluster, keypairs, errp); - if (ret < 0) { - ret = -EIO; - goto shutdown; - } - - if (qemu_rbd_set_auth(cluster, password_secret, errp) < 0) { - ret = -EIO; - goto shutdown; - } - - ret = rados_connect(cluster); - if (ret < 0) { - error_setg_errno(errp, -ret, "error connecting"); - goto shutdown; - } - - ret = rados_ioctx_create(cluster, opts->location->pool, &io_ctx); - if (ret < 0) { - error_setg_errno(errp, -ret, "error opening pool %s", - opts->location->pool); - goto shutdown; - } - ret = rbd_create(io_ctx, opts->location->image, opts->size, &obj_order); if (ret < 0) { error_setg_errno(errp, -ret, "error rbd create"); + goto out; } - rados_ioctx_destroy(io_ctx); - ret = 0; -shutdown: +out: + rados_ioctx_destroy(io_ctx); rados_shutdown(cluster); return ret; } -- 2.13.6 Using the QAPI visitor to turn all options into QAPI BlockdevOptionsNfs simplifies the code a lot. It will also be useful for implementing the QAPI based .bdrv_co_create callback. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> --- block/nfs.c | 176 ++++++++++++++++++------------------------------------------ 1 file changed, 53 insertions(+), 123 deletions(-) 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 @@ static int coroutine_fn nfs_co_flush(BlockDriverState *bs) return task.ret; } -static QemuOptsList runtime_opts = { - .name = "nfs", - .head = QTAILQ_HEAD_INITIALIZER(runtime_opts.head), - .desc = { - { - .name = "path", - .type = QEMU_OPT_STRING, - .help = "Path of the image on the host", - }, - { - .name = "user", - .type = QEMU_OPT_NUMBER, - .help = "UID value to use when talking to the server", - }, - { - .name = "group", - .type = QEMU_OPT_NUMBER, - .help = "GID value to use when talking to the server", - }, - { - .name = "tcp-syn-count", - .type = QEMU_OPT_NUMBER, - .help = "Number of SYNs to send during the session establish", - }, - { - .name = "readahead-size", - .type = QEMU_OPT_NUMBER, - .help = "Set the readahead size in bytes", - }, - { - .name = "page-cache-size", - .type = QEMU_OPT_NUMBER, - .help = "Set the pagecache size in bytes", - }, - { - .name = "debug", - .type = QEMU_OPT_NUMBER, - .help = "Set the NFS debug level (max 2)", - }, - { /* end of list */ } - }, -}; - static void nfs_detach_aio_context(BlockDriverState *bs) { NFSClient *client = bs->opaque; @@ -XXX,XX +XXX,XX @@ static void nfs_file_close(BlockDriverState *bs) nfs_client_close(client); } -static NFSServer *nfs_config(QDict *options, Error **errp) -{ - NFSServer *server = NULL; - QDict *addr = NULL; - QObject *crumpled_addr = NULL; - Visitor *iv = NULL; - Error *local_error = NULL; - - qdict_extract_subqdict(options, &addr, "server."); - if (!qdict_size(addr)) { - error_setg(errp, "NFS server address missing"); - goto out; - } - - crumpled_addr = qdict_crumple(addr, errp); - if (!crumpled_addr) { - goto out; - } - - /* - * Caution: this works only because all scalar members of - * NFSServer are QString in @crumpled_addr. The visitor expects - * @crumpled_addr to be typed according to the QAPI schema. It - * is when @options come from -blockdev or blockdev_add. But when - * they come from -drive, they're all QString. - */ - iv = qobject_input_visitor_new(crumpled_addr); - visit_type_NFSServer(iv, NULL, &server, &local_error); - if (local_error) { - error_propagate(errp, local_error); - goto out; - } - -out: - QDECREF(addr); - qobject_decref(crumpled_addr); - visit_free(iv); - return server; -} - - -static int64_t nfs_client_open(NFSClient *client, QDict *options, +static int64_t nfs_client_open(NFSClient *client, BlockdevOptionsNfs *opts, int flags, int open_flags, Error **errp) { int64_t ret = -EINVAL; - QemuOpts *opts = NULL; - Error *local_err = NULL; struct stat st; char *file = NULL, *strp = NULL; qemu_mutex_init(&client->mutex); - opts = qemu_opts_create(&runtime_opts, NULL, 0, &error_abort); - qemu_opts_absorb_qdict(opts, options, &local_err); - if (local_err) { - error_propagate(errp, local_err); - ret = -EINVAL; - goto fail; - } - client->path = g_strdup(qemu_opt_get(opts, "path")); - if (!client->path) { - ret = -EINVAL; - error_setg(errp, "No path was specified"); - goto fail; - } + client->path = g_strdup(opts->path); strp = strrchr(client->path, '/'); if (strp == NULL) { @@ -XXX,XX +XXX,XX @@ static int64_t nfs_client_open(NFSClient *client, QDict *options, file = g_strdup(strp); *strp = 0; - /* Pop the config into our state object, Exit if invalid */ - client->server = nfs_config(options, errp); - if (!client->server) { - ret = -EINVAL; - goto fail; - } + /* Steal the NFSServer object from opts; set the original pointer to NULL + * to avoid use after free and double free. */ + client->server = opts->server; + opts->server = NULL; client->context = nfs_init_context(); if (client->context == NULL) { @@ -XXX,XX +XXX,XX @@ static int64_t nfs_client_open(NFSClient *client, QDict *options, goto fail; } - if (qemu_opt_get(opts, "user")) { - client->uid = qemu_opt_get_number(opts, "user", 0); + if (opts->has_user) { + client->uid = opts->user; nfs_set_uid(client->context, client->uid); } - if (qemu_opt_get(opts, "group")) { - client->gid = qemu_opt_get_number(opts, "group", 0); + if (opts->has_group) { + client->gid = opts->group; nfs_set_gid(client->context, client->gid); } - if (qemu_opt_get(opts, "tcp-syn-count")) { - client->tcp_syncnt = qemu_opt_get_number(opts, "tcp-syn-count", 0); + if (opts->has_tcp_syn_count) { + client->tcp_syncnt = opts->tcp_syn_count; nfs_set_tcp_syncnt(client->context, client->tcp_syncnt); } #ifdef LIBNFS_FEATURE_READAHEAD - if (qemu_opt_get(opts, "readahead-size")) { + if (opts->has_readahead_size) { if (open_flags & BDRV_O_NOCACHE) { error_setg(errp, "Cannot enable NFS readahead " "if cache.direct = on"); goto fail; } - client->readahead = qemu_opt_get_number(opts, "readahead-size", 0); + client->readahead = opts->readahead_size; if (client->readahead > QEMU_NFS_MAX_READAHEAD_SIZE) { warn_report("Truncating NFS readahead size to %d", QEMU_NFS_MAX_READAHEAD_SIZE); @@ -XXX,XX +XXX,XX @@ static int64_t nfs_client_open(NFSClient *client, QDict *options, #endif #ifdef LIBNFS_FEATURE_PAGECACHE - if (qemu_opt_get(opts, "page-cache-size")) { + if (opts->has_page_cache_size) { if (open_flags & BDRV_O_NOCACHE) { error_setg(errp, "Cannot enable NFS pagecache " "if cache.direct = on"); goto fail; } - client->pagecache = qemu_opt_get_number(opts, "page-cache-size", 0); + client->pagecache = opts->page_cache_size; if (client->pagecache > QEMU_NFS_MAX_PAGECACHE_SIZE) { warn_report("Truncating NFS pagecache size to %d pages", QEMU_NFS_MAX_PAGECACHE_SIZE); @@ -XXX,XX +XXX,XX @@ static int64_t nfs_client_open(NFSClient *client, QDict *options, #endif #ifdef LIBNFS_FEATURE_DEBUG - if (qemu_opt_get(opts, "debug")) { - client->debug = qemu_opt_get_number(opts, "debug", 0); + if (opts->has_debug) { + client->debug = opts->debug; /* limit the maximum debug level to avoid potential flooding * of our log files. */ if (client->debug > QEMU_NFS_MAX_DEBUG_LEVEL) { @@ -XXX,XX +XXX,XX @@ static int64_t nfs_client_open(NFSClient *client, QDict *options, fail: nfs_client_close(client); out: - qemu_opts_del(opts); g_free(file); return ret; } +static int64_t nfs_client_open_qdict(NFSClient *client, QDict *options, + int flags, int open_flags, Error **errp) +{ + BlockdevOptionsNfs *opts = NULL; + QObject *crumpled = NULL; + Visitor *v; + Error *local_err = NULL; + int ret; + + crumpled = qdict_crumple(options, errp); + if (crumpled == NULL) { + return -EINVAL; + } + + v = qobject_input_visitor_new_keyval(crumpled); + visit_type_BlockdevOptionsNfs(v, NULL, &opts, &local_err); + visit_free(v); + + if (local_err) { + error_propagate(errp, local_err); + ret = -EINVAL; + goto fail; + } + + ret = nfs_client_open(client, opts, flags, open_flags, errp); +fail: + qobject_decref(crumpled); + qapi_free_BlockdevOptionsNfs(opts); + return ret; +} + static int nfs_file_open(BlockDriverState *bs, QDict *options, int flags, Error **errp) { NFSClient *client = bs->opaque; @@ -XXX,XX +XXX,XX @@ static int nfs_file_open(BlockDriverState *bs, QDict *options, int flags, client->aio_context = bdrv_get_aio_context(bs); - ret = nfs_client_open(client, options, - (flags & BDRV_O_RDWR) ? O_RDWR : O_RDONLY, - bs->open_flags, errp); + ret = nfs_client_open_qdict(client, options, + (flags & BDRV_O_RDWR) ? O_RDWR : O_RDONLY, + bs->open_flags, errp); if (ret < 0) { return ret; } @@ -XXX,XX +XXX,XX @@ static int coroutine_fn nfs_file_co_create_opts(const char *url, QemuOpts *opts, goto out; } - ret = nfs_client_open(client, options, O_CREAT, 0, errp); + ret = nfs_client_open_qdict(client, options, O_CREAT, 0, errp); if (ret < 0) { goto out; } -- 2.13.6 This adds the .bdrv_co_create driver callback to nfs, which enables image creation over QMP. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> --- qapi/block-core.json | 16 ++++++++++- block/nfs.c | 76 +++++++++++++++++++++++++++++++++++++++++----------- 2 files changed, 75 insertions(+), 17 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 @@ '*preallocation': 'PreallocMode' } } ## +# @BlockdevCreateOptionsNfs: +# +# Driver specific image creation options for NFS. +# +# @location Where to store the new image file +# @size Size of the virtual disk in bytes +# +# Since: 2.12 +## +{ 'struct': 'BlockdevCreateOptionsNfs', + 'data': { 'location': 'BlockdevOptionsNfs', + 'size': 'size' } } + +## # @BlockdevQcow2Version: # # @v2: The original QCOW2 format as introduced in qemu 0.10 (version 2) @@ -XXX,XX +XXX,XX @@ 'iscsi': 'BlockdevCreateNotSupported', 'luks': 'BlockdevCreateNotSupported', 'nbd': 'BlockdevCreateNotSupported', - 'nfs': 'BlockdevCreateNotSupported', + 'nfs': 'BlockdevCreateOptionsNfs', 'null-aio': 'BlockdevCreateNotSupported', 'null-co': 'BlockdevCreateNotSupported', 'nvme': 'BlockdevCreateNotSupported', 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 @@ out: return ret; } -static int64_t nfs_client_open_qdict(NFSClient *client, QDict *options, - int flags, int open_flags, Error **errp) +static BlockdevOptionsNfs *nfs_options_qdict_to_qapi(QDict *options, + Error **errp) { BlockdevOptionsNfs *opts = NULL; QObject *crumpled = NULL; Visitor *v; Error *local_err = NULL; - int ret; crumpled = qdict_crumple(options, errp); if (crumpled == NULL) { - return -EINVAL; + return NULL; } v = qobject_input_visitor_new_keyval(crumpled); visit_type_BlockdevOptionsNfs(v, NULL, &opts, &local_err); visit_free(v); + qobject_decref(crumpled); if (local_err) { - error_propagate(errp, local_err); + return NULL; + } + + return opts; +} + +static int64_t nfs_client_open_qdict(NFSClient *client, QDict *options, + int flags, int open_flags, Error **errp) +{ + BlockdevOptionsNfs *opts; + int ret; + + opts = nfs_options_qdict_to_qapi(options, errp); + if (opts == NULL) { ret = -EINVAL; goto fail; } ret = nfs_client_open(client, opts, flags, open_flags, errp); fail: - qobject_decref(crumpled); qapi_free_BlockdevOptionsNfs(opts); return ret; } @@ -XXX,XX +XXX,XX @@ static QemuOptsList nfs_create_opts = { } }; -static int coroutine_fn nfs_file_co_create_opts(const char *url, QemuOpts *opts, - Error **errp) +static int nfs_file_co_create(BlockdevCreateOptions *options, Error **errp) { - int64_t ret, total_size; + BlockdevCreateOptionsNfs *opts = &options->u.nfs; NFSClient *client = g_new0(NFSClient, 1); - QDict *options = NULL; + int ret; + + assert(options->driver == BLOCKDEV_DRIVER_NFS); client->aio_context = qemu_get_aio_context(); + ret = nfs_client_open(client, opts->location, O_CREAT, 0, errp); + if (ret < 0) { + goto out; + } + ret = nfs_ftruncate(client->context, client->fh, opts->size); + nfs_client_close(client); + +out: + g_free(client); + return ret; +} + +static int coroutine_fn nfs_file_co_create_opts(const char *url, QemuOpts *opts, + Error **errp) +{ + BlockdevCreateOptions *create_options; + BlockdevCreateOptionsNfs *nfs_opts; + QDict *options; + int ret; + + create_options = g_new0(BlockdevCreateOptions, 1); + create_options->driver = BLOCKDEV_DRIVER_NFS; + nfs_opts = &create_options->u.nfs; + /* Read out options */ - total_size = ROUND_UP(qemu_opt_get_size_del(opts, BLOCK_OPT_SIZE, 0), - BDRV_SECTOR_SIZE); + nfs_opts->size = ROUND_UP(qemu_opt_get_size_del(opts, BLOCK_OPT_SIZE, 0), + BDRV_SECTOR_SIZE); options = qdict_new(); ret = nfs_parse_uri(url, options, errp); @@ -XXX,XX +XXX,XX @@ static int coroutine_fn nfs_file_co_create_opts(const char *url, QemuOpts *opts, goto out; } - ret = nfs_client_open_qdict(client, options, O_CREAT, 0, errp); + nfs_opts->location = nfs_options_qdict_to_qapi(options, errp); + if (nfs_opts->location == NULL) { + ret = -EINVAL; + goto out; + } + + ret = nfs_file_co_create(create_options, errp); if (ret < 0) { goto out; } - ret = nfs_ftruncate(client->context, client->fh, total_size); - nfs_client_close(client); + + ret = 0; out: QDECREF(options); - g_free(client); + qapi_free_BlockdevCreateOptions(create_options); return ret; } @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_nfs = { .bdrv_file_open = nfs_file_open, .bdrv_close = nfs_file_close, + .bdrv_co_create = nfs_file_co_create, .bdrv_co_create_opts = nfs_file_co_create_opts, .bdrv_reopen_prepare = nfs_reopen_prepare, -- 2.13.6 The "redundancy" option for Sheepdog image creation is currently a string that can encode one or two integers depending on its format, which at the same time implicitly selects a mode. This patch turns it into a QAPI union and converts the string into such a QAPI object before interpreting the values. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> --- qapi/block-core.json | 45 +++++++++++++++++++++++++ block/sheepdog.c | 94 +++++++++++++++++++++++++++++++++++++--------------- 2 files changed, 112 insertions(+), 27 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 @@ '*cluster-size' : 'size' } } ## +# @SheepdogRedundancyType: +# +# @full Create a fully replicated vdi with x copies +# @erasure-coded Create an erasure coded vdi with x data strips and +# y parity strips +# +# Since: 2.12 +## +{ 'enum': 'SheepdogRedundancyType', + 'data': [ 'full', 'erasure-coded' ] } + +## +# @SheepdogRedundancyFull: +# +# @copies Number of copies to use (between 1 and 31) +# +# Since: 2.12 +## +{ 'struct': 'SheepdogRedundancyFull', + 'data': { 'copies': 'int' }} + +## +# @SheepdogRedundancyErasureCoded: +# +# @data-strips Number of data strips to use (one of {2,4,8,16}) +# @parity-strips Number of parity strips to use (between 1 and 15) +# +# Since: 2.12 +## +{ 'struct': 'SheepdogRedundancyErasureCoded', + 'data': { 'data-strips': 'int', + 'parity-strips': 'int' }} + +## +# @SheepdogRedundancy: +# +# Since: 2.12 +## +{ 'union': 'SheepdogRedundancy', + 'base': { 'type': 'SheepdogRedundancyType' }, + 'discriminator': 'type', + 'data': { 'full': 'SheepdogRedundancyFull', + 'erasure-coded': 'SheepdogRedundancyErasureCoded' } } + +## # @BlockdevCreateNotSupported: # # This is used for all drivers that don't support creating images. diff --git a/block/sheepdog.c b/block/sheepdog.c index XXXXXXX..XXXXXXX 100644 --- a/block/sheepdog.c +++ b/block/sheepdog.c @@ -XXX,XX +XXX,XX @@ out_with_err_set: return ret; } +static int parse_redundancy(BDRVSheepdogState *s, SheepdogRedundancy *opt) +{ + struct SheepdogInode *inode = &s->inode; + + switch (opt->type) { + case SHEEPDOG_REDUNDANCY_TYPE_FULL: + if (opt->u.full.copies > SD_MAX_COPIES || opt->u.full.copies < 1) { + return -EINVAL; + } + inode->copy_policy = 0; + inode->nr_copies = opt->u.full.copies; + return 0; + + case SHEEPDOG_REDUNDANCY_TYPE_ERASURE_CODED: + { + int64_t copy = opt->u.erasure_coded.data_strips; + int64_t parity = opt->u.erasure_coded.parity_strips; + + if (copy != 2 && copy != 4 && copy != 8 && copy != 16) { + return -EINVAL; + } + + if (parity >= SD_EC_MAX_STRIP || parity < 1) { + return -EINVAL; + } + + /* + * 4 bits for parity and 4 bits for data. + * We have to compress upper data bits because it can't represent 16 + */ + inode->copy_policy = ((copy / 2) << 4) + parity; + inode->nr_copies = copy + parity; + return 0; + } + + default: + g_assert_not_reached(); + } + + return -EINVAL; +} + /* * Sheepdog support two kinds of redundancy, full replication and erasure * coding. @@ -XXX,XX +XXX,XX @@ out_with_err_set: * # create a erasure coded vdi with x data strips and y parity strips * -o redundancy=x:y (x must be one of {2,4,8,16} and 1 <= y < SD_EC_MAX_STRIP) */ -static int parse_redundancy(BDRVSheepdogState *s, const char *opt) +static int parse_redundancy_str(BDRVSheepdogState *s, const char *opt) { - struct SheepdogInode *inode = &s->inode; + struct SheepdogRedundancy redundancy; const char *n1, *n2; long copy, parity; char p[10]; + int ret; pstrcpy(p, sizeof(p), opt); n1 = strtok(p, ":"); @@ -XXX,XX +XXX,XX @@ static int parse_redundancy(BDRVSheepdogState *s, const char *opt) return -EINVAL; } - copy = strtol(n1, NULL, 10); - /* FIXME fix error checking by switching to qemu_strtol() */ - if (copy > SD_MAX_COPIES || copy < 1) { - return -EINVAL; - } - if (!n2) { - inode->copy_policy = 0; - inode->nr_copies = copy; - return 0; + ret = qemu_strtol(n1, NULL, 10, &copy); + if (ret < 0) { + return ret; } - if (copy != 2 && copy != 4 && copy != 8 && copy != 16) { - return -EINVAL; - } + if (!n2) { + redundancy = (SheepdogRedundancy) { + .type = SHEEPDOG_REDUNDANCY_TYPE_FULL, + .u.full.copies = copy, + }; + } else { + ret = qemu_strtol(n2, NULL, 10, &parity); + if (ret < 0) { + return ret; + } - parity = strtol(n2, NULL, 10); - /* FIXME fix error checking by switching to qemu_strtol() */ - if (parity >= SD_EC_MAX_STRIP || parity < 1) { - return -EINVAL; + redundancy = (SheepdogRedundancy) { + .type = SHEEPDOG_REDUNDANCY_TYPE_ERASURE_CODED, + .u.erasure_coded = { + .data_strips = copy, + .parity_strips = parity, + }, + }; } - /* - * 4 bits for parity and 4 bits for data. - * We have to compress upper data bits because it can't represent 16 - */ - inode->copy_policy = ((copy / 2) << 4) + parity; - inode->nr_copies = copy + parity; - - return 0; + return parse_redundancy(s, &redundancy); } static int parse_block_size_shift(BDRVSheepdogState *s, QemuOpts *opt) @@ -XXX,XX +XXX,XX @@ static int coroutine_fn sd_co_create_opts(const char *filename, QemuOpts *opts, g_free(buf); buf = qemu_opt_get_del(opts, BLOCK_OPT_REDUNDANCY); if (buf) { - ret = parse_redundancy(s, buf); + ret = parse_redundancy_str(s, buf); if (ret < 0) { error_setg(errp, "Invalid redundancy mode: '%s'", buf); goto out; -- 2.13.6 This adds the .bdrv_co_create driver callback to sheepdog, which enables image creation over QMP. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> --- qapi/block-core.json | 24 ++++- block/sheepdog.c | 243 +++++++++++++++++++++++++++++++++++---------------- 2 files changed, 192 insertions(+), 75 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 @@ 'erasure-coded': 'SheepdogRedundancyErasureCoded' } } ## +# @BlockdevCreateOptionsSheepdog: +# +# Driver specific image creation options for Sheepdog. +# +# @location Where to store the new image file +# @size Size of the virtual disk in bytes +# @backing-file File name of a base image +# @preallocation Preallocation mode (allowed values: off, full) +# @redundancy Redundancy of the image +# @object-size Object size of the image +# +# Since: 2.12 +## +{ 'struct': 'BlockdevCreateOptionsSheepdog', + 'data': { 'location': 'BlockdevOptionsSheepdog', + 'size': 'size', + '*backing-file': 'str', + '*preallocation': 'PreallocMode', + '*redundancy': 'SheepdogRedundancy', + '*object-size': 'size' } } + +## # @BlockdevCreateNotSupported: # # This is used for all drivers that don't support creating images. @@ -XXX,XX +XXX,XX @@ 'raw': 'BlockdevCreateNotSupported', 'rbd': 'BlockdevCreateOptionsRbd', 'replication': 'BlockdevCreateNotSupported', - 'sheepdog': 'BlockdevCreateNotSupported', + 'sheepdog': 'BlockdevCreateOptionsSheepdog', 'ssh': 'BlockdevCreateNotSupported', 'throttle': 'BlockdevCreateNotSupported', 'vdi': 'BlockdevCreateNotSupported', diff --git a/block/sheepdog.c b/block/sheepdog.c index XXXXXXX..XXXXXXX 100644 --- a/block/sheepdog.c +++ b/block/sheepdog.c @@ -XXX,XX +XXX,XX @@ #include "qemu/osdep.h" #include "qapi/error.h" #include "qapi/qapi-visit-sockets.h" +#include "qapi/qapi-visit-block-core.h" #include "qapi/qmp/qdict.h" #include "qapi/qobject-input-visitor.h" +#include "qapi/qobject-output-visitor.h" #include "qemu/uri.h" #include "qemu/error-report.h" #include "qemu/option.h" @@ -XXX,XX +XXX,XX @@ static void sd_aio_setup(SheepdogAIOCB *acb, BDRVSheepdogState *s, qemu_co_mutex_unlock(&s->queue_lock); } -static SocketAddress *sd_socket_address(const char *path, - const char *host, const char *port) -{ - SocketAddress *addr = g_new0(SocketAddress, 1); - - if (path) { - addr->type = SOCKET_ADDRESS_TYPE_UNIX; - addr->u.q_unix.path = g_strdup(path); - } else { - addr->type = SOCKET_ADDRESS_TYPE_INET; - addr->u.inet.host = g_strdup(host ?: SD_DEFAULT_ADDR); - addr->u.inet.port = g_strdup(port ?: stringify(SD_DEFAULT_PORT)); - } - - return addr; -} - static SocketAddress *sd_server_config(QDict *options, Error **errp) { QDict *server = NULL; @@ -XXX,XX +XXX,XX @@ out_with_err_set: return ret; } +static int sd_create_prealloc(BlockdevOptionsSheepdog *location, int64_t size, + Error **errp) +{ + BlockDriverState *bs; + Visitor *v; + QObject *obj = NULL; + QDict *qdict; + Error *local_err = NULL; + int ret; + + v = qobject_output_visitor_new(&obj); + visit_type_BlockdevOptionsSheepdog(v, NULL, &location, &local_err); + visit_free(v); + + if (local_err) { + error_propagate(errp, local_err); + qobject_decref(obj); + return -EINVAL; + } + + qdict = qobject_to_qdict(obj); + qdict_flatten(qdict); + + qdict_put_str(qdict, "driver", "sheepdog"); + + bs = bdrv_open(NULL, NULL, qdict, BDRV_O_PROTOCOL | BDRV_O_RDWR, errp); + if (bs == NULL) { + ret = -EIO; + goto fail; + } + + ret = sd_prealloc(bs, 0, size, errp); +fail: + bdrv_unref(bs); + QDECREF(qdict); + return ret; +} + static int parse_redundancy(BDRVSheepdogState *s, SheepdogRedundancy *opt) { struct SheepdogInode *inode = &s->inode; @@ -XXX,XX +XXX,XX @@ static int parse_redundancy(BDRVSheepdogState *s, SheepdogRedundancy *opt) * # create a erasure coded vdi with x data strips and y parity strips * -o redundancy=x:y (x must be one of {2,4,8,16} and 1 <= y < SD_EC_MAX_STRIP) */ -static int parse_redundancy_str(BDRVSheepdogState *s, const char *opt) +static SheepdogRedundancy *parse_redundancy_str(const char *opt) { - struct SheepdogRedundancy redundancy; + SheepdogRedundancy *redundancy; const char *n1, *n2; long copy, parity; char p[10]; @@ -XXX,XX +XXX,XX @@ static int parse_redundancy_str(BDRVSheepdogState *s, const char *opt) n2 = strtok(NULL, ":"); if (!n1) { - return -EINVAL; + return NULL; } ret = qemu_strtol(n1, NULL, 10, &copy); if (ret < 0) { - return ret; + return NULL; } + redundancy = g_new0(SheepdogRedundancy, 1); if (!n2) { - redundancy = (SheepdogRedundancy) { + *redundancy = (SheepdogRedundancy) { .type = SHEEPDOG_REDUNDANCY_TYPE_FULL, .u.full.copies = copy, }; } else { ret = qemu_strtol(n2, NULL, 10, &parity); if (ret < 0) { - return ret; + return NULL; } - redundancy = (SheepdogRedundancy) { + *redundancy = (SheepdogRedundancy) { .type = SHEEPDOG_REDUNDANCY_TYPE_ERASURE_CODED, .u.erasure_coded = { .data_strips = copy, @@ -XXX,XX +XXX,XX @@ static int parse_redundancy_str(BDRVSheepdogState *s, const char *opt) }; } - return parse_redundancy(s, &redundancy); + return redundancy; } -static int parse_block_size_shift(BDRVSheepdogState *s, QemuOpts *opt) +static int parse_block_size_shift(BDRVSheepdogState *s, + BlockdevCreateOptionsSheepdog *opts) { struct SheepdogInode *inode = &s->inode; uint64_t object_size; int obj_order; - object_size = qemu_opt_get_size_del(opt, BLOCK_OPT_OBJECT_SIZE, 0); - if (object_size) { + if (opts->has_object_size) { + object_size = opts->object_size; + if ((object_size - 1) & object_size) { /* not a power of 2? */ return -EINVAL; } @@ -XXX,XX +XXX,XX @@ static int parse_block_size_shift(BDRVSheepdogState *s, QemuOpts *opt) return 0; } -static int coroutine_fn sd_co_create_opts(const char *filename, QemuOpts *opts, - Error **errp) +static int sd_co_create(BlockdevCreateOptions *options, Error **errp) { - Error *err = NULL; + BlockdevCreateOptionsSheepdog *opts = &options->u.sheepdog; int ret = 0; uint32_t vid = 0; char *backing_file = NULL; char *buf = NULL; BDRVSheepdogState *s; - SheepdogConfig cfg; uint64_t max_vdi_size; bool prealloc = false; + assert(options->driver == BLOCKDEV_DRIVER_SHEEPDOG); + s = g_new0(BDRVSheepdogState, 1); - if (strstr(filename, "://")) { - sd_parse_uri(&cfg, filename, &err); - } else { - parse_vdiname(&cfg, filename, &err); - } - if (err) { - error_propagate(errp, err); + /* Steal SocketAddress from QAPI, set NULL to prevent double free */ + s->addr = opts->location->server; + opts->location->server = NULL; + + if (strlen(opts->location->vdi) >= sizeof(s->name)) { + error_setg(errp, "'vdi' string too long"); + ret = -EINVAL; goto out; } + pstrcpy(s->name, sizeof(s->name), opts->location->vdi); - buf = cfg.port ? g_strdup_printf("%d", cfg.port) : NULL; - s->addr = sd_socket_address(cfg.path, cfg.host, buf); - g_free(buf); - strcpy(s->name, cfg.vdi); - sd_config_done(&cfg); + s->inode.vdi_size = opts->size; + backing_file = opts->backing_file; - s->inode.vdi_size = ROUND_UP(qemu_opt_get_size_del(opts, BLOCK_OPT_SIZE, 0), - BDRV_SECTOR_SIZE); - backing_file = qemu_opt_get_del(opts, BLOCK_OPT_BACKING_FILE); - buf = qemu_opt_get_del(opts, BLOCK_OPT_PREALLOC); - if (!buf || !strcmp(buf, "off")) { + if (!opts->has_preallocation) { + opts->preallocation = PREALLOC_MODE_OFF; + } + switch (opts->preallocation) { + case PREALLOC_MODE_OFF: prealloc = false; - } else if (!strcmp(buf, "full")) { + break; + case PREALLOC_MODE_FULL: prealloc = true; - } else { - error_setg(errp, "Invalid preallocation mode: '%s'", buf); + break; + default: + error_setg(errp, "Preallocation mode not supported for Sheepdog"); ret = -EINVAL; goto out; } - g_free(buf); - buf = qemu_opt_get_del(opts, BLOCK_OPT_REDUNDANCY); - if (buf) { - ret = parse_redundancy_str(s, buf); + if (opts->has_redundancy) { + ret = parse_redundancy(s, opts->redundancy); if (ret < 0) { - error_setg(errp, "Invalid redundancy mode: '%s'", buf); + error_setg(errp, "Invalid redundancy mode"); goto out; } } @@ -XXX,XX +XXX,XX @@ static int coroutine_fn sd_co_create_opts(const char *filename, QemuOpts *opts, goto out; } - if (backing_file) { + if (opts->has_backing_file) { BlockBackend *blk; BDRVSheepdogState *base; BlockDriver *drv; /* Currently, only Sheepdog backing image is supported. */ - drv = bdrv_find_protocol(backing_file, true, NULL); + drv = bdrv_find_protocol(opts->backing_file, true, NULL); if (!drv || strcmp(drv->protocol_name, "sheepdog") != 0) { error_setg(errp, "backing_file must be a sheepdog image"); ret = -EINVAL; goto out; } - blk = blk_new_open(backing_file, NULL, NULL, + blk = blk_new_open(opts->backing_file, NULL, NULL, BDRV_O_PROTOCOL, errp); if (blk == NULL) { ret = -EIO; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn sd_co_create_opts(const char *filename, QemuOpts *opts, } if (prealloc) { - BlockDriverState *bs; - QDict *opts; - - opts = qdict_new(); - qdict_put_str(opts, "driver", "sheepdog"); - bs = bdrv_open(filename, NULL, opts, BDRV_O_PROTOCOL | BDRV_O_RDWR, - errp); - if (!bs) { - goto out; - } - - ret = sd_prealloc(bs, 0, s->inode.vdi_size, errp); - - bdrv_unref(bs); + ret = sd_create_prealloc(opts->location, opts->size, errp); } out: g_free(backing_file); g_free(buf); + g_free(s->addr); g_free(s); return ret; } +static int coroutine_fn sd_co_create_opts(const char *filename, QemuOpts *opts, + Error **errp) +{ + BlockdevCreateOptions *create_options = NULL; + QDict *qdict, *location_qdict; + QObject *crumpled; + Visitor *v; + const char *redundancy; + Error *local_err = NULL; + int ret; + + redundancy = qemu_opt_get_del(opts, BLOCK_OPT_REDUNDANCY); + + qdict = qemu_opts_to_qdict(opts, NULL); + qdict_put_str(qdict, "driver", "sheepdog"); + + location_qdict = qdict_new(); + qdict_put(qdict, "location", location_qdict); + + sd_parse_filename(filename, location_qdict, &local_err); + if (local_err) { + error_propagate(errp, local_err); + ret = -EINVAL; + goto fail; + } + + qdict_flatten(qdict); + + /* Change legacy command line options into QMP ones */ + static const QDictRenames opt_renames[] = { + { BLOCK_OPT_BACKING_FILE, "backing-file" }, + { BLOCK_OPT_OBJECT_SIZE, "object-size" }, + { NULL, NULL }, + }; + + if (!qdict_rename_keys(qdict, opt_renames, errp)) { + ret = -EINVAL; + goto fail; + } + + /* Get the QAPI object */ + crumpled = qdict_crumple(qdict, errp); + if (crumpled == NULL) { + ret = -EINVAL; + goto fail; + } + + v = qobject_input_visitor_new_keyval(crumpled); + visit_type_BlockdevCreateOptions(v, NULL, &create_options, &local_err); + visit_free(v); + qobject_decref(crumpled); + + if (local_err) { + error_propagate(errp, local_err); + ret = -EINVAL; + goto fail; + } + + assert(create_options->driver == BLOCKDEV_DRIVER_SHEEPDOG); + create_options->u.sheepdog.size = + ROUND_UP(create_options->u.sheepdog.size, BDRV_SECTOR_SIZE); + + if (redundancy) { + create_options->u.sheepdog.has_redundancy = true; + create_options->u.sheepdog.redundancy = + parse_redundancy_str(redundancy); + if (create_options->u.sheepdog.redundancy == NULL) { + error_setg(errp, "Invalid redundancy mode"); + ret = -EINVAL; + goto fail; + } + } + + ret = sd_co_create(create_options, errp); +fail: + qapi_free_BlockdevCreateOptions(create_options); + QDECREF(qdict); + return ret; +} + static void sd_close(BlockDriverState *bs) { Error *local_err = NULL; @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_sheepdog = { .bdrv_reopen_commit = sd_reopen_commit, .bdrv_reopen_abort = sd_reopen_abort, .bdrv_close = sd_close, + .bdrv_co_create = sd_co_create, .bdrv_co_create_opts = sd_co_create_opts, .bdrv_has_zero_init = bdrv_has_zero_init_1, .bdrv_getlength = sd_getlength, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_sheepdog_tcp = { .bdrv_reopen_commit = sd_reopen_commit, .bdrv_reopen_abort = sd_reopen_abort, .bdrv_close = sd_close, + .bdrv_co_create = sd_co_create, .bdrv_co_create_opts = sd_co_create_opts, .bdrv_has_zero_init = bdrv_has_zero_init_1, .bdrv_getlength = sd_getlength, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_sheepdog_unix = { .bdrv_reopen_commit = sd_reopen_commit, .bdrv_reopen_abort = sd_reopen_abort, .bdrv_close = sd_close, + .bdrv_co_create = sd_co_create, .bdrv_co_create_opts = sd_co_create_opts, .bdrv_has_zero_init = bdrv_has_zero_init_1, .bdrv_getlength = sd_getlength, -- 2.13.6 Create a BlockdevOptionsSsh object in connect_to_ssh() and take the options from there. 'host_key_check' is still processed separately because it's not in the schema yet. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> --- block/ssh.c | 137 +++++++++++++++++++++++++++--------------------------------- 1 file changed, 62 insertions(+), 75 deletions(-) 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 @@ #include "qemu/sockets.h" #include "qemu/uri.h" #include "qapi/qapi-visit-sockets.h" +#include "qapi/qapi-visit-block-core.h" #include "qapi/qmp/qdict.h" #include "qapi/qmp/qstring.h" #include "qapi/qobject-input-visitor.h" @@ -XXX,XX +XXX,XX @@ static QemuOptsList ssh_runtime_opts = { .type = QEMU_OPT_NUMBER, .help = "Port to connect to", }, - { - .name = "path", - .type = QEMU_OPT_STRING, - .help = "Path of the image on the host", - }, - { - .name = "user", - .type = QEMU_OPT_STRING, - .help = "User as which to connect", - }, - { - .name = "host_key_check", - .type = QEMU_OPT_STRING, - .help = "Defines how and what to check the host key against", - }, { /* end of list */ } }, }; @@ -XXX,XX +XXX,XX @@ static bool ssh_process_legacy_socket_options(QDict *output_opts, return true; } -static InetSocketAddress *ssh_config(QDict *options, Error **errp) +static BlockdevOptionsSsh *ssh_parse_options(QDict *options, Error **errp) { - InetSocketAddress *inet = NULL; - QDict *addr = NULL; - QObject *crumpled_addr = NULL; - Visitor *iv = NULL; - Error *local_error = NULL; - - qdict_extract_subqdict(options, &addr, "server."); - if (!qdict_size(addr)) { - error_setg(errp, "SSH server address missing"); - goto out; + BlockdevOptionsSsh *result = NULL; + QemuOpts *opts = NULL; + Error *local_err = NULL; + QObject *crumpled; + const QDictEntry *e; + Visitor *v; + + /* Translate legacy options */ + opts = qemu_opts_create(&ssh_runtime_opts, NULL, 0, &error_abort); + qemu_opts_absorb_qdict(opts, options, &local_err); + if (local_err) { + error_propagate(errp, local_err); + goto fail; } - crumpled_addr = qdict_crumple(addr, errp); - if (!crumpled_addr) { - goto out; + if (!ssh_process_legacy_socket_options(options, opts, errp)) { + goto fail; + } + + /* Create the QAPI object */ + crumpled = qdict_crumple(options, errp); + if (crumpled == NULL) { + goto fail; } /* @@ -XXX,XX +XXX,XX @@ static InetSocketAddress *ssh_config(QDict *options, Error **errp) * but when they come from -drive, they're all QString. The * visitor expects the former. */ - iv = qobject_input_visitor_new(crumpled_addr); - visit_type_InetSocketAddress(iv, NULL, &inet, &local_error); - if (local_error) { - error_propagate(errp, local_error); - goto out; + v = qobject_input_visitor_new(crumpled); + visit_type_BlockdevOptionsSsh(v, NULL, &result, &local_err); + visit_free(v); + qobject_decref(crumpled); + + if (local_err) { + error_propagate(errp, local_err); + goto fail; } -out: - QDECREF(addr); - qobject_decref(crumpled_addr); - visit_free(iv); - return inet; + /* Remove the processed options from the QDict (the visitor processes + * _all_ options in the QDict) */ + while ((e = qdict_first(options))) { + qdict_del(options, e->key); + } + +fail: + qemu_opts_del(opts); + return result; } static int connect_to_ssh(BDRVSSHState *s, QDict *options, int ssh_flags, int creat_mode, Error **errp) { + BlockdevOptionsSsh *opts; int r, ret; - QemuOpts *opts = NULL; - Error *local_err = NULL; - const char *user, *path, *host_key_check; + const char *user, *host_key_check; long port = 0; - opts = qemu_opts_create(&ssh_runtime_opts, NULL, 0, &error_abort); - qemu_opts_absorb_qdict(opts, options, &local_err); - if (local_err) { - ret = -EINVAL; - error_propagate(errp, local_err); - goto err; - } - - if (!ssh_process_legacy_socket_options(options, opts, errp)) { - ret = -EINVAL; - goto err; + host_key_check = qdict_get_try_str(options, "host_key_check"); + if (!host_key_check) { + host_key_check = "yes"; + } else { + qdict_del(options, "host_key_check"); } - path = qemu_opt_get(opts, "path"); - if (!path) { - ret = -EINVAL; - error_setg(errp, "No path was specified"); - goto err; + opts = ssh_parse_options(options, errp); + if (opts == NULL) { + return -EINVAL; } - user = qemu_opt_get(opts, "user"); - if (!user) { + if (opts->has_user) { + user = opts->user; + } else { user = g_get_user_name(); if (!user) { error_setg_errno(errp, errno, "Can't get user name"); @@ -XXX,XX +XXX,XX @@ static int connect_to_ssh(BDRVSSHState *s, QDict *options, } } - host_key_check = qemu_opt_get(opts, "host_key_check"); - if (!host_key_check) { - host_key_check = "yes"; - } - /* Pop the config into our state object, Exit if invalid */ - s->inet = ssh_config(options, errp); - if (!s->inet) { - ret = -EINVAL; - goto err; - } + s->inet = opts->server; + opts->server = NULL; if (qemu_strtol(s->inet->port, NULL, 10, &port) < 0) { error_setg(errp, "Use only numeric port value"); @@ -XXX,XX +XXX,XX @@ static int connect_to_ssh(BDRVSSHState *s, QDict *options, /* Open the remote file. */ DPRINTF("opening file %s flags=0x%x creat_mode=0%o", - path, ssh_flags, creat_mode); - s->sftp_handle = libssh2_sftp_open(s->sftp, path, ssh_flags, creat_mode); + opts->path, ssh_flags, creat_mode); + s->sftp_handle = libssh2_sftp_open(s->sftp, opts->path, ssh_flags, + creat_mode); if (!s->sftp_handle) { - session_error_setg(errp, s, "failed to open remote file '%s'", path); + session_error_setg(errp, s, "failed to open remote file '%s'", + opts->path); ret = -EINVAL; goto err; } - qemu_opts_del(opts); + qapi_free_BlockdevOptionsSsh(opts); r = libssh2_sftp_fstat(s->sftp_handle, &s->attrs); if (r < 0) { @@ -XXX,XX +XXX,XX @@ static int connect_to_ssh(BDRVSSHState *s, QDict *options, } s->session = NULL; - qemu_opts_del(opts); + qapi_free_BlockdevOptionsSsh(opts); return ret; } -- 2.13.6 This makes the host-key-check option available in blockdev-add. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> --- qapi/block-core.json | 63 +++++++++++++++++++++++++++++++++++-- block/ssh.c | 88 +++++++++++++++++++++++++++++++++------------------- 2 files changed, 117 insertions(+), 34 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 @@ '*encrypt': 'BlockdevQcow2Encryption' } } ## +# @SshHostKeyCheckMode: +# +# @none Don't check the host key at all +# @hash Compare the host key with a given hash +# @known_hosts Check the host key against the known_hosts file +# +# Since: 2.12 +## +{ 'enum': 'SshHostKeyCheckMode', + 'data': [ 'none', 'hash', 'known_hosts' ] } + +## +# @SshHostKeyCheckHashType: +# +# @md5 The given hash is an md5 hash +# @sha1 The given hash is an sha1 hash +# +# Since: 2.12 +## +{ 'enum': 'SshHostKeyCheckHashType', + 'data': [ 'md5', 'sha1' ] } + +## +# @SshHostKeyHash: +# +# @type The hash algorithm used for the hash +# @hash The expected hash value +# +# Since: 2.12 +## +{ 'struct': 'SshHostKeyHash', + 'data': { 'type': 'SshHostKeyCheckHashType', + 'hash': 'str' }} + +## +# @SshHostKeyDummy: +# +# For those union branches that don't need additional fields. +# +# Since: 2.12 +## +{ 'struct': 'SshHostKeyDummy', + 'data': {} } + +## +# @SshHostKeyCheck: +# +# Since: 2.12 +## +{ 'union': 'SshHostKeyCheck', + 'base': { 'mode': 'SshHostKeyCheckMode' }, + 'discriminator': 'mode', + 'data': { 'none': 'SshHostKeyDummy', + 'hash': 'SshHostKeyHash', + 'known_hosts': 'SshHostKeyDummy' } } + +## # @BlockdevOptionsSsh: # # @server: host address @@ -XXX,XX +XXX,XX @@ # @user: user as which to connect, defaults to current # local user name # -# TODO: Expose the host_key_check option in QMP +# @host-key-check: Defines how and what to check the host key against +# (default: known_hosts) # # Since: 2.9 ## { 'struct': 'BlockdevOptionsSsh', 'data': { 'server': 'InetSocketAddress', 'path': 'str', - '*user': 'str' } } + '*user': 'str', + '*host-key-check': 'SshHostKeyCheck' } } ## 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 @@ check_host_key_hash(BDRVSSHState *s, const char *hash, } static int check_host_key(BDRVSSHState *s, const char *host, int port, - const char *host_key_check, Error **errp) + SshHostKeyCheck *hkc, Error **errp) { - /* host_key_check=no */ - if (strcmp(host_key_check, "no") == 0) { - return 0; - } + SshHostKeyCheckMode mode; - /* host_key_check=md5:xx:yy:zz:... */ - if (strncmp(host_key_check, "md5:", 4) == 0) { - return check_host_key_hash(s, &host_key_check[4], - LIBSSH2_HOSTKEY_HASH_MD5, 16, errp); - } - - /* host_key_check=sha1:xx:yy:zz:... */ - if (strncmp(host_key_check, "sha1:", 5) == 0) { - return check_host_key_hash(s, &host_key_check[5], - LIBSSH2_HOSTKEY_HASH_SHA1, 20, errp); + if (hkc) { + mode = hkc->mode; + } else { + mode = SSH_HOST_KEY_CHECK_MODE_KNOWN_HOSTS; } - /* host_key_check=yes */ - if (strcmp(host_key_check, "yes") == 0) { + switch (mode) { + case SSH_HOST_KEY_CHECK_MODE_NONE: + return 0; + case SSH_HOST_KEY_CHECK_MODE_HASH: + if (hkc->u.hash.type == SSH_HOST_KEY_CHECK_HASH_TYPE_MD5) { + return check_host_key_hash(s, hkc->u.hash.hash, + LIBSSH2_HOSTKEY_HASH_MD5, 16, errp); + } else if (hkc->u.hash.type == SSH_HOST_KEY_CHECK_HASH_TYPE_SHA1) { + return check_host_key_hash(s, hkc->u.hash.hash, + LIBSSH2_HOSTKEY_HASH_SHA1, 20, errp); + } + g_assert_not_reached(); + break; + case SSH_HOST_KEY_CHECK_MODE_KNOWN_HOSTS: return check_host_key_knownhosts(s, host, port, errp); + default: + g_assert_not_reached(); } - error_setg(errp, "unknown host_key_check setting (%s)", host_key_check); return -EINVAL; } @@ -XXX,XX +XXX,XX @@ static QemuOptsList ssh_runtime_opts = { .type = QEMU_OPT_NUMBER, .help = "Port to connect to", }, + { + .name = "host_key_check", + .type = QEMU_OPT_STRING, + .help = "Defines how and what to check the host key against", + }, { /* end of list */ } }, }; -static bool ssh_process_legacy_socket_options(QDict *output_opts, - QemuOpts *legacy_opts, - Error **errp) +static bool ssh_process_legacy_options(QDict *output_opts, + QemuOpts *legacy_opts, + Error **errp) { const char *host = qemu_opt_get(legacy_opts, "host"); const char *port = qemu_opt_get(legacy_opts, "port"); + const char *host_key_check = qemu_opt_get(legacy_opts, "host_key_check"); if (!host && port) { error_setg(errp, "port may not be used without host"); @@ -XXX,XX +XXX,XX @@ static bool ssh_process_legacy_socket_options(QDict *output_opts, qdict_put_str(output_opts, "server.port", port ?: stringify(22)); } + if (host_key_check) { + if (strcmp(host_key_check, "no") == 0) { + qdict_put_str(output_opts, "host-key-check.mode", "none"); + } else if (strncmp(host_key_check, "md5:", 4) == 0) { + qdict_put_str(output_opts, "host-key-check.mode", "hash"); + qdict_put_str(output_opts, "host-key-check.type", "md5"); + qdict_put_str(output_opts, "host-key-check.hash", + &host_key_check[4]); + } else if (strncmp(host_key_check, "sha1:", 5) == 0) { + qdict_put_str(output_opts, "host-key-check.mode", "hash"); + qdict_put_str(output_opts, "host-key-check.type", "sha1"); + qdict_put_str(output_opts, "host-key-check.hash", + &host_key_check[5]); + } else if (strcmp(host_key_check, "yes") == 0) { + qdict_put_str(output_opts, "host-key-check.mode", "known_hosts"); + } else { + error_setg(errp, "unknown host_key_check setting (%s)", + host_key_check); + return false; + } + } + return true; } @@ -XXX,XX +XXX,XX @@ static BlockdevOptionsSsh *ssh_parse_options(QDict *options, Error **errp) goto fail; } - if (!ssh_process_legacy_socket_options(options, opts, errp)) { + if (!ssh_process_legacy_options(options, opts, errp)) { goto fail; } @@ -XXX,XX +XXX,XX @@ static int connect_to_ssh(BDRVSSHState *s, QDict *options, { BlockdevOptionsSsh *opts; int r, ret; - const char *user, *host_key_check; + const char *user; long port = 0; - host_key_check = qdict_get_try_str(options, "host_key_check"); - if (!host_key_check) { - host_key_check = "yes"; - } else { - qdict_del(options, "host_key_check"); - } - opts = ssh_parse_options(options, errp); if (opts == NULL) { return -EINVAL; @@ -XXX,XX +XXX,XX @@ static int connect_to_ssh(BDRVSSHState *s, QDict *options, } /* Check the remote host's key against known_hosts. */ - ret = check_host_key(s, s->inet->host, port, host_key_check, - errp); + ret = check_host_key(s, s->inet->host, port, opts->host_key_check, errp); if (ret < 0) { goto err; } -- 2.13.6 Move the parsing of the QDict options up to the callers, in preparation for the .bdrv_co_create implementation that directly gets a QAPI type. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> --- block/ssh.c | 34 +++++++++++++++++++++------------- 1 file changed, 21 insertions(+), 13 deletions(-) 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 @@ fail: return result; } -static int connect_to_ssh(BDRVSSHState *s, QDict *options, +static int connect_to_ssh(BDRVSSHState *s, BlockdevOptionsSsh *opts, int ssh_flags, int creat_mode, Error **errp) { - BlockdevOptionsSsh *opts; int r, ret; const char *user; long port = 0; - opts = ssh_parse_options(options, errp); - if (opts == NULL) { - return -EINVAL; - } - if (opts->has_user) { user = opts->user; } else { @@ -XXX,XX +XXX,XX @@ static int connect_to_ssh(BDRVSSHState *s, QDict *options, goto err; } - qapi_free_BlockdevOptionsSsh(opts); - r = libssh2_sftp_fstat(s->sftp_handle, &s->attrs); if (r < 0) { sftp_error_setg(errp, s, "failed to read file attributes"); @@ -XXX,XX +XXX,XX @@ static int connect_to_ssh(BDRVSSHState *s, QDict *options, } s->session = NULL; - qapi_free_BlockdevOptionsSsh(opts); - return ret; } @@ -XXX,XX +XXX,XX @@ static int ssh_file_open(BlockDriverState *bs, QDict *options, int bdrv_flags, Error **errp) { BDRVSSHState *s = bs->opaque; + BlockdevOptionsSsh *opts; int ret; int ssh_flags; @@ -XXX,XX +XXX,XX @@ static int ssh_file_open(BlockDriverState *bs, QDict *options, int bdrv_flags, ssh_flags |= LIBSSH2_FXF_WRITE; } + opts = ssh_parse_options(options, errp); + if (opts == NULL) { + return -EINVAL; + } + /* Start up SSH. */ - ret = connect_to_ssh(s, options, ssh_flags, 0, errp); + ret = connect_to_ssh(s, opts, ssh_flags, 0, errp); if (ret < 0) { goto err; } @@ -XXX,XX +XXX,XX @@ static int ssh_file_open(BlockDriverState *bs, QDict *options, int bdrv_flags, /* Go non-blocking. */ libssh2_session_set_blocking(s->session, 0); + qapi_free_BlockdevOptionsSsh(opts); + return 0; err: @@ -XXX,XX +XXX,XX @@ static int ssh_file_open(BlockDriverState *bs, QDict *options, int bdrv_flags, } s->sock = -1; + qapi_free_BlockdevOptionsSsh(opts); + return ret; } @@ -XXX,XX +XXX,XX @@ static int coroutine_fn ssh_co_create_opts(const char *filename, QemuOpts *opts, int r, ret; int64_t total_size = 0; QDict *uri_options = NULL; + BlockdevOptionsSsh *ssh_opts = NULL; BDRVSSHState s; ssh_state_init(&s); @@ -XXX,XX +XXX,XX @@ static int coroutine_fn ssh_co_create_opts(const char *filename, QemuOpts *opts, goto out; } - r = connect_to_ssh(&s, uri_options, + ssh_opts = ssh_parse_options(uri_options, errp); + if (ssh_opts == NULL) { + ret = -EINVAL; + goto out; + } + + r = connect_to_ssh(&s, ssh_opts, LIBSSH2_FXF_READ|LIBSSH2_FXF_WRITE| LIBSSH2_FXF_CREAT|LIBSSH2_FXF_TRUNC, 0644, errp); @@ -XXX,XX +XXX,XX @@ static int coroutine_fn ssh_co_create_opts(const char *filename, QemuOpts *opts, if (uri_options != NULL) { QDECREF(uri_options); } + qapi_free_BlockdevOptionsSsh(ssh_opts); return ret; } -- 2.13.6 This adds the .bdrv_co_create driver callback to ssh, which enables image creation over QMP. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> --- qapi/block-core.json | 16 +++++++++- block/ssh.c | 83 ++++++++++++++++++++++++++++++---------------------- 2 files changed, 63 insertions(+), 36 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 @@ '*object-size': 'size' } } ## +# @BlockdevCreateOptionsSsh: +# +# Driver specific image creation options for SSH. +# +# @location Where to store the new image file +# @size Size of the virtual disk in bytes +# +# Since: 2.12 +## +{ 'struct': 'BlockdevCreateOptionsSsh', + 'data': { 'location': 'BlockdevOptionsSsh', + 'size': 'size' } } + +## # @BlockdevCreateNotSupported: # # This is used for all drivers that don't support creating images. @@ -XXX,XX +XXX,XX @@ 'rbd': 'BlockdevCreateOptionsRbd', 'replication': 'BlockdevCreateNotSupported', 'sheepdog': 'BlockdevCreateOptionsSheepdog', - 'ssh': 'BlockdevCreateNotSupported', + 'ssh': 'BlockdevCreateOptionsSsh', 'throttle': 'BlockdevCreateNotSupported', 'vdi': 'BlockdevCreateNotSupported', 'vhdx': 'BlockdevCreateNotSupported', 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 QemuOptsList ssh_create_opts = { } }; +static int ssh_co_create(BlockdevCreateOptions *options, Error **errp) +{ + BlockdevCreateOptionsSsh *opts = &options->u.ssh; + BDRVSSHState s; + int ret; + + assert(options->driver == BLOCKDEV_DRIVER_SSH); + + ssh_state_init(&s); + + ret = connect_to_ssh(&s, opts->location, + LIBSSH2_FXF_READ|LIBSSH2_FXF_WRITE| + LIBSSH2_FXF_CREAT|LIBSSH2_FXF_TRUNC, + 0644, errp); + if (ret < 0) { + goto fail; + } + + if (opts->size > 0) { + ret = ssh_grow_file(&s, opts->size, errp); + if (ret < 0) { + goto fail; + } + } + + ret = 0; +fail: + ssh_state_free(&s); + return ret; +} + static int coroutine_fn ssh_co_create_opts(const char *filename, QemuOpts *opts, Error **errp) { - int r, ret; - int64_t total_size = 0; + BlockdevCreateOptions *create_options; + BlockdevCreateOptionsSsh *ssh_opts; + int ret; QDict *uri_options = NULL; - BlockdevOptionsSsh *ssh_opts = NULL; - BDRVSSHState s; - ssh_state_init(&s); + create_options = g_new0(BlockdevCreateOptions, 1); + create_options->driver = BLOCKDEV_DRIVER_SSH; + ssh_opts = &create_options->u.ssh; /* Get desired file size. */ - total_size = ROUND_UP(qemu_opt_get_size_del(opts, BLOCK_OPT_SIZE, 0), - BDRV_SECTOR_SIZE); - DPRINTF("total_size=%" PRIi64, total_size); + ssh_opts->size = ROUND_UP(qemu_opt_get_size_del(opts, BLOCK_OPT_SIZE, 0), + BDRV_SECTOR_SIZE); + DPRINTF("total_size=%" PRIi64, ssh_opts->size); uri_options = qdict_new(); - r = parse_uri(filename, uri_options, errp); - if (r < 0) { - ret = r; + ret = parse_uri(filename, uri_options, errp); + if (ret < 0) { goto out; } - ssh_opts = ssh_parse_options(uri_options, errp); - if (ssh_opts == NULL) { + ssh_opts->location = ssh_parse_options(uri_options, errp); + if (ssh_opts->location == NULL) { ret = -EINVAL; goto out; } - r = connect_to_ssh(&s, ssh_opts, - LIBSSH2_FXF_READ|LIBSSH2_FXF_WRITE| - LIBSSH2_FXF_CREAT|LIBSSH2_FXF_TRUNC, - 0644, errp); - if (r < 0) { - ret = r; - goto out; - } - - if (total_size > 0) { - ret = ssh_grow_file(&s, total_size, errp); - if (ret < 0) { - goto out; - } - } - - ret = 0; + ret = ssh_co_create(create_options, errp); out: - ssh_state_free(&s); - if (uri_options != NULL) { - QDECREF(uri_options); - } - qapi_free_BlockdevOptionsSsh(ssh_opts); + QDECREF(uri_options); + qapi_free_BlockdevCreateOptions(create_options); return ret; } @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_ssh = { .instance_size = sizeof(BDRVSSHState), .bdrv_parse_filename = ssh_parse_filename, .bdrv_file_open = ssh_file_open, + .bdrv_co_create = ssh_co_create, .bdrv_co_create_opts = ssh_co_create_opts, .bdrv_close = ssh_close, .bdrv_has_zero_init = ssh_has_zero_init, -- 2.13.6 If bdrv_truncate() is called, but the requested size is the same as before, don't call posix_fallocate(), which returns -EINVAL for length zero and would therefore make bdrv_truncate() fail. The problem can be triggered by creating a zero-sized raw image with 'falloc' preallocation mode. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> --- block/file-posix.c | 14 +++++++++----- 1 file changed, 9 insertions(+), 5 deletions(-) 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 raw_regular_truncate(int fd, int64_t offset, PreallocMode prealloc, * file systems that do not support fallocate(), trying to check if a * block is allocated before allocating it, so don't do that here. */ - result = -posix_fallocate(fd, current_length, offset - current_length); - if (result != 0) { - /* posix_fallocate() doesn't set errno. */ - error_setg_errno(errp, -result, - "Could not preallocate new data"); + if (offset != current_length) { + result = -posix_fallocate(fd, current_length, offset - current_length); + if (result != 0) { + /* posix_fallocate() doesn't set errno. */ + error_setg_errno(errp, -result, + "Could not preallocate new data"); + } + } else { + result = 0; } goto out; #endif -- 2.13.6 Most callers have their own checks, but something like this should also be checked centrally. As it happens, x-blockdev-create can pass negative image sizes to format drivers (because there is no QAPI type that would reject negative numbers) and triggers the check added by this patch. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> --- block.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ int bdrv_truncate(BdrvChild *child, int64_t offset, PreallocMode prealloc, error_setg(errp, "No medium inserted"); return -ENOMEDIUM; } + if (offset < 0) { + error_setg(errp, "Image size cannot be negative"); + return -EINVAL; + } + if (!drv->bdrv_truncate) { if (bs->file && drv->is_filter) { return bdrv_truncate(bs->file, offset, prealloc, errp); -- 2.13.6 Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> --- tests/qemu-iotests/206 | 436 +++++++++++++++++++++++++++++++++++++++++++++ tests/qemu-iotests/206.out | 209 ++++++++++++++++++++++ tests/qemu-iotests/group | 1 + 3 files changed, 646 insertions(+) create mode 100755 tests/qemu-iotests/206 create mode 100644 tests/qemu-iotests/206.out diff --git a/tests/qemu-iotests/206 b/tests/qemu-iotests/206 new file mode 100755 index XXXXXXX..XXXXXXX --- /dev/null +++ b/tests/qemu-iotests/206 @@ -XXX,XX +XXX,XX @@ +#!/bin/bash +# +# Test qcow2 and file image creation +# +# Copyright (C) 2018 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 $seq" + +here=`pwd` +status=1 # failure is the default! + +# get standard environment, filters and checks +. ./common.rc +. ./common.filter + +_supported_fmt qcow2 +_supported_proto file +_supported_os Linux + +function do_run_qemu() +{ + echo Testing: "$@" + $QEMU -nographic -qmp stdio -serial none "$@" + echo +} + +function run_qemu() +{ + do_run_qemu "$@" 2>&1 | _filter_testdir | _filter_qmp \ + | _filter_qemu | _filter_imgfmt \ + | _filter_actual_image_size +} + +echo +echo "=== Successful image creation (defaults) ===" +echo + +size=$((128 * 1024 * 1024)) + +run_qemu <<EOF +{ "execute": "qmp_capabilities" } +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "file", + "filename": "$TEST_IMG", + "size": 0 + } +} +{ "execute": "blockdev-add", + "arguments": { + "driver": "file", + "node-name": "imgfile", + "filename": "$TEST_IMG" + } +} +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "$IMGFMT", + "file": "imgfile", + "size": $size + } +} +{ "execute": "quit" } +EOF + +_img_info --format-specific + +echo +echo "=== Successful image creation (inline blockdev-add, explicit defaults) ===" +echo + +# Choose a different size to show that we got a new image +size=$((64 * 1024 * 1024)) + +run_qemu <<EOF +{ "execute": "qmp_capabilities" } +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "file", + "filename": "$TEST_IMG", + "size": 0, + "preallocation": "off", + "nocow": false + } +} +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "$IMGFMT", + "file": { + "driver": "file", + "filename": "$TEST_IMG" + }, + "size": $size, + "version": "v3", + "cluster-size": 65536, + "preallocation": "off", + "lazy-refcounts": false, + "refcount-bits": 16 + } +} +{ "execute": "quit" } +EOF + +_img_info --format-specific + +echo +echo "=== Successful image creation (v3 non-default options) ===" +echo + +# Choose a different size to show that we got a new image +size=$((32 * 1024 * 1024)) + +run_qemu <<EOF +{ "execute": "qmp_capabilities" } +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "file", + "filename": "$TEST_IMG", + "size": 0, + "preallocation": "falloc", + "nocow": true + } +} +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "$IMGFMT", + "file": { + "driver": "file", + "filename": "$TEST_IMG" + }, + "size": $size, + "version": "v3", + "cluster-size": 2097152, + "preallocation": "metadata", + "lazy-refcounts": true, + "refcount-bits": 1 + } +} +{ "execute": "quit" } +EOF + +_img_info --format-specific + +echo +echo "=== Successful image creation (v2 non-default options) ===" +echo + +mv $TEST_IMG $TEST_IMG.base + +run_qemu <<EOF +{ "execute": "qmp_capabilities" } +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "file", + "filename": "$TEST_IMG", + "size": 0 + } +} +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "$IMGFMT", + "file": { + "driver": "file", + "filename": "$TEST_IMG" + }, + "size": $size, + "backing-file": "$TEST_IMG.base", + "backing-fmt": "qcow2", + "version": "v2", + "cluster-size": 512 + } +} +{ "execute": "quit" } +EOF + +_img_info --format-specific + +echo +echo "=== Successful image creation (encrypted) ===" +echo + +run_qemu -object secret,id=keysec0,data="foo" <<EOF +{ "execute": "qmp_capabilities" } +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "$IMGFMT", + "file": { + "driver": "file", + "filename": "$TEST_IMG" + }, + "size": $size, + "encrypt": { + "format": "luks", + "key-secret": "keysec0", + "cipher-alg": "twofish-128", + "cipher-mode": "ctr", + "ivgen-alg": "plain64", + "ivgen-hash-alg": "md5", + "hash-alg": "sha1", + "iter-time": 10 + } + } +} +{ "execute": "quit" } +EOF + +_img_info --format-specific | _filter_img_info --format-specific + +echo +echo "=== Invalid BlockdevRef ===" +echo + +run_qemu <<EOF +{ "execute": "qmp_capabilities" } +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "$IMGFMT", + "file": "this doesn't exist", + "size": $size + } +} +{ "execute": "quit" } +EOF + + +echo +echo "=== Invalid sizes ===" +echo + +# TODO Negative image sizes aren't handled correctly, but this is a problem +# with QAPI's implementation of the 'size' type and affects other commands as +# well. Once this is fixed, we may want to add a test case here. + +# 1. Misaligned image size +# 2. 2^64 - 512 +# 3. 2^63 = 8 EB (qemu-img enforces image sizes less than this) +# 4. 2^63 - 512 (generally valid, but qcow2 can't handle images this size) + +run_qemu -blockdev driver=file,filename="$TEST_IMG",node-name=node0 <<EOF +{ "execute": "qmp_capabilities" } +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "$IMGFMT", + "file": "node0", + "size": 1234 + } +} +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "$IMGFMT", + "file": "node0", + "size": 18446744073709551104 + } +} +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "$IMGFMT", + "file": "node0", + "size": 9223372036854775808 + } +} +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "$IMGFMT", + "file": "node0", + "size": 9223372036854775296 + } +} +{ "execute": "quit" } +EOF + +echo +echo "=== Invalid version ===" +echo + +run_qemu -blockdev driver=file,filename="$TEST_IMG",node-name=node0 <<EOF +{ "execute": "qmp_capabilities" } +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "$IMGFMT", + "file": "node0", + "size": 67108864, + "version": "v1" + } +} +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "$IMGFMT", + "file": "node0", + "size": 67108864, + "version": "v2", + "lazy-refcounts": true + } +} +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "$IMGFMT", + "file": "node0", + "size": 67108864, + "version": "v2", + "refcount-bits": 8 + } +} +{ "execute": "quit" } +EOF + +echo +echo "=== Invalid backing file options ===" +echo + +run_qemu -blockdev driver=file,filename="$TEST_IMG",node-name=node0 <<EOF +{ "execute": "qmp_capabilities" } +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "$IMGFMT", + "file": "node0", + "size": 67108864, + "backing-file": "/dev/null", + "preallocation": "full" + } +} +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "$IMGFMT", + "file": "node0", + "size": 67108864, + "backing-fmt": "$IMGFMT" + } +} +{ "execute": "quit" } +EOF + +echo +echo "=== Invalid cluster size ===" +echo + +run_qemu -blockdev driver=file,filename="$TEST_IMG",node-name=node0 <<EOF +{ "execute": "qmp_capabilities" } +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "$IMGFMT", + "file": "node0", + "size": 67108864, + "cluster-size": 1234 + } +} +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "$IMGFMT", + "file": "node0", + "size": 67108864, + "cluster-size": 128 + } +} +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "$IMGFMT", + "file": "node0", + "size": 67108864, + "cluster-size": 4194304 + } +} +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "$IMGFMT", + "file": "node0", + "size": 67108864, + "cluster-size": 0 + } +} +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "$IMGFMT", + "file": "node0", + "size": 281474976710656, + "cluster-size": 512 + } +} +{ "execute": "quit" } +EOF + +echo +echo "=== Invalid refcount width ===" +echo + +run_qemu -blockdev driver=file,filename="$TEST_IMG",node-name=node0 <<EOF +{ "execute": "qmp_capabilities" } +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "$IMGFMT", + "file": "node0", + "size": 67108864, + "refcount-bits": 128 + } +} +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "$IMGFMT", + "file": "node0", + "size": 67108864, + "refcount-bits": 0 + } +} +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "$IMGFMT", + "file": "node0", + "size": 67108864, + "refcount-bits": 7 + } +} +{ "execute": "quit" } +EOF + +# success, all done +echo "*** done" +rm -f $seq.full +status=0 diff --git a/tests/qemu-iotests/206.out b/tests/qemu-iotests/206.out new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/tests/qemu-iotests/206.out @@ -XXX,XX +XXX,XX @@ +QA output created by 206 + +=== Successful image creation (defaults) === + +Testing: +QMP_VERSION +{"return": {}} +{"return": {}} +{"return": {}} +{"return": {}} +{"return": {}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}} + +image: TEST_DIR/t.IMGFMT +file format: IMGFMT +virtual size: 128M (134217728 bytes) +cluster_size: 65536 +Format specific information: + compat: 1.1 + lazy refcounts: false + refcount bits: 16 + corrupt: false + +=== Successful image creation (inline blockdev-add, explicit defaults) === + +Testing: +QMP_VERSION +{"return": {}} +{"return": {}} +{"return": {}} +{"return": {}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}} + +image: TEST_DIR/t.IMGFMT +file format: IMGFMT +virtual size: 64M (67108864 bytes) +cluster_size: 65536 +Format specific information: + compat: 1.1 + lazy refcounts: false + refcount bits: 16 + corrupt: false + +=== Successful image creation (v3 non-default options) === + +Testing: +QMP_VERSION +{"return": {}} +{"return": {}} +{"return": {}} +{"return": {}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}} + +image: TEST_DIR/t.IMGFMT +file format: IMGFMT +virtual size: 32M (33554432 bytes) +cluster_size: 2097152 +Format specific information: + compat: 1.1 + lazy refcounts: true + refcount bits: 1 + corrupt: false + +=== Successful image creation (v2 non-default options) === + +Testing: +QMP_VERSION +{"return": {}} +{"return": {}} +{"return": {}} +{"return": {}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}} + +image: TEST_DIR/t.IMGFMT +file format: IMGFMT +virtual size: 32M (33554432 bytes) +cluster_size: 512 +backing file: TEST_DIR/t.IMGFMT.base +backing file format: IMGFMT +Format specific information: + compat: 0.10 + refcount bits: 16 + +=== Successful image creation (encrypted) === + +Testing: -object secret,id=keysec0,data=foo +QMP_VERSION +{"return": {}} +{"return": {}} +{"return": {}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}} + +image: TEST_DIR/t.IMGFMT +file format: IMGFMT +virtual size: 32M (33554432 bytes) +Format specific information: + compat: 1.1 + lazy refcounts: false + refcount bits: 16 + encrypt: + ivgen alg: plain64 + hash alg: sha1 + cipher alg: twofish-128 + uuid: 00000000-0000-0000-0000-000000000000 + format: luks + cipher mode: ctr + slots: + [0]: + active: true + iters: 1024 + key offset: 4096 + stripes: 4000 + [1]: + active: false + key offset: 69632 + [2]: + active: false + key offset: 135168 + [3]: + active: false + key offset: 200704 + [4]: + active: false + key offset: 266240 + [5]: + active: false + key offset: 331776 + [6]: + active: false + key offset: 397312 + [7]: + active: false + key offset: 462848 + payload offset: 528384 + master key iters: 1024 + corrupt: false + +=== Invalid BlockdevRef === + +Testing: +QMP_VERSION +{"return": {}} +{"error": {"class": "GenericError", "desc": "Cannot find device=this doesn't exist nor node_name=this doesn't exist"}} +{"return": {}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}} + + +=== Invalid sizes === + +Testing: -blockdev driver=file,filename=TEST_DIR/t.IMGFMT,node-name=node0 +QMP_VERSION +{"return": {}} +{"error": {"class": "GenericError", "desc": "Image size must be a multiple of 512 bytes"}} +{"error": {"class": "GenericError", "desc": "Could not resize image: Image size cannot be negative"}} +{"error": {"class": "GenericError", "desc": "Could not resize image: Image size cannot be negative"}} +{"error": {"class": "GenericError", "desc": "Could not resize image: Failed to grow the L1 table: File too large"}} +{"return": {}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}} + + +=== Invalid version === + +Testing: -blockdev driver=file,filename=TEST_DIR/t.IMGFMT,node-name=node0 +QMP_VERSION +{"return": {}} +{"error": {"class": "GenericError", "desc": "Invalid parameter 'v1'"}} +{"error": {"class": "GenericError", "desc": "Lazy refcounts only supported with compatibility level 1.1 and above (use version=v3 or greater)"}} +{"error": {"class": "GenericError", "desc": "Different refcount widths than 16 bits require compatibility level 1.1 or above (use version=v3 or greater)"}} +{"return": {}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}} + + +=== Invalid backing file options === + +Testing: -blockdev driver=file,filename=TEST_DIR/t.IMGFMT,node-name=node0 +QMP_VERSION +{"return": {}} +{"error": {"class": "GenericError", "desc": "Backing file and preallocation cannot be used at the same time"}} +{"error": {"class": "GenericError", "desc": "Backing format cannot be used without backing file"}} +{"return": {}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}} + + +=== Invalid cluster size === + +Testing: -blockdev driver=file,filename=TEST_DIR/t.IMGFMT,node-name=node0 +QMP_VERSION +{"return": {}} +{"error": {"class": "GenericError", "desc": "Cluster size must be a power of two between 512 and 2048k"}} +{"error": {"class": "GenericError", "desc": "Cluster size must be a power of two between 512 and 2048k"}} +{"error": {"class": "GenericError", "desc": "Cluster size must be a power of two between 512 and 2048k"}} +{"error": {"class": "GenericError", "desc": "Cluster size must be a power of two between 512 and 2048k"}} +{"error": {"class": "GenericError", "desc": "Could not resize image: Failed to grow the L1 table: File too large"}} +{"return": {}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}} + + +=== Invalid refcount width === + +Testing: -blockdev driver=file,filename=TEST_DIR/t.IMGFMT,node-name=node0 +QMP_VERSION +{"return": {}} +{"error": {"class": "GenericError", "desc": "Refcount width must be a power of two and may not exceed 64 bits"}} +{"error": {"class": "GenericError", "desc": "Refcount width must be a power of two and may not exceed 64 bits"}} +{"error": {"class": "GenericError", "desc": "Refcount width must be a power of two and may not exceed 64 bits"}} +{"return": {}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}} + +*** done diff --git a/tests/qemu-iotests/group b/tests/qemu-iotests/group index XXXXXXX..XXXXXXX 100644 --- a/tests/qemu-iotests/group +++ b/tests/qemu-iotests/group @@ -XXX,XX +XXX,XX @@ 203 rw auto 204 rw auto quick 205 rw auto quick +206 rw auto -- 2.13.6 Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> --- tests/qemu-iotests/207 | 261 +++++++++++++++++++++++++++++++++++++++++++++ tests/qemu-iotests/207.out | 75 +++++++++++++ tests/qemu-iotests/group | 1 + 3 files changed, 337 insertions(+) create mode 100755 tests/qemu-iotests/207 create mode 100644 tests/qemu-iotests/207.out diff --git a/tests/qemu-iotests/207 b/tests/qemu-iotests/207 new file mode 100755 index XXXXXXX..XXXXXXX --- /dev/null +++ b/tests/qemu-iotests/207 @@ -XXX,XX +XXX,XX @@ +#!/bin/bash +# +# Test ssh image creation +# +# Copyright (C) 2018 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 $seq" + +here=`pwd` +status=1 # failure is the default! + +# get standard environment, filters and checks +. ./common.rc +. ./common.filter + +_supported_fmt raw +_supported_proto ssh +_supported_os Linux + +function do_run_qemu() +{ + echo Testing: "$@" + $QEMU -nographic -qmp stdio -serial none "$@" + echo +} + +function run_qemu() +{ + do_run_qemu "$@" 2>&1 | _filter_testdir | _filter_qmp \ + | _filter_qemu | _filter_imgfmt \ + | _filter_actual_image_size +} + +echo +echo "=== Successful image creation (defaults) ===" +echo + +run_qemu <<EOF +{ "execute": "qmp_capabilities" } +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "ssh", + "location": { + "path": "$TEST_IMG_FILE", + "server": { + "host": "127.0.0.1", + "port": "22" + } + }, + "size": 4194304 + } +} +{ "execute": "quit" } +EOF + +_img_info | _filter_img_info +echo +TEST_IMG=$TEST_IMG_FILE _img_info | _filter_img_info + +echo +echo "=== Test host-key-check options ===" +echo + +run_qemu <<EOF +{ "execute": "qmp_capabilities" } +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "ssh", + "location": { + "path": "$TEST_IMG_FILE", + "server": { + "host": "127.0.0.1", + "port": "22" + }, + "host-key-check": { + "mode": "none" + } + }, + "size": 8388608 + } +} +{ "execute": "quit" } +EOF + +_img_info | _filter_img_info + +run_qemu <<EOF +{ "execute": "qmp_capabilities" } +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "ssh", + "location": { + "path": "$TEST_IMG_FILE", + "server": { + "host": "127.0.0.1", + "port": "22" + }, + "host-key-check": { + "mode": "known_hosts" + } + }, + "size": 4194304 + } +} +{ "execute": "quit" } +EOF + +_img_info | _filter_img_info + + +key=$(ssh-keyscan -t rsa 127.0.0.1 2>/dev/null | grep -v "\\^#" | + cut -d" " -f3 | base64 -d | md5sum -b | cut -d" " -f1) + +run_qemu <<EOF +{ "execute": "qmp_capabilities" } +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "ssh", + "location": { + "path": "$TEST_IMG_FILE", + "server": { + "host": "127.0.0.1", + "port": "22" + }, + "host-key-check": { + "mode": "hash", + "type": "md5", + "hash": "wrong" + } + }, + "size": 8388608 + } +} +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "ssh", + "location": { + "path": "$TEST_IMG_FILE", + "server": { + "host": "127.0.0.1", + "port": "22" + }, + "host-key-check": { + "mode": "hash", + "type": "md5", + "hash": "$key" + } + }, + "size": 8388608 + } +} +{ "execute": "quit" } +EOF + +_img_info | _filter_img_info + + +key=$(ssh-keyscan -t rsa 127.0.0.1 2>/dev/null | grep -v "\\^#" | + cut -d" " -f3 | base64 -d | sha1sum -b | cut -d" " -f1) + +run_qemu <<EOF +{ "execute": "qmp_capabilities" } +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "ssh", + "location": { + "path": "$TEST_IMG_FILE", + "server": { + "host": "127.0.0.1", + "port": "22" + }, + "host-key-check": { + "mode": "hash", + "type": "sha1", + "hash": "wrong" + } + }, + "size": 4194304 + } +} +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "ssh", + "location": { + "path": "$TEST_IMG_FILE", + "server": { + "host": "127.0.0.1", + "port": "22" + }, + "host-key-check": { + "mode": "hash", + "type": "sha1", + "hash": "$key" + } + }, + "size": 4194304 + } +} +{ "execute": "quit" } +EOF + +_img_info | _filter_img_info + +echo +echo "=== Invalid path and user ===" +echo + +run_qemu <<EOF +{ "execute": "qmp_capabilities" } +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "ssh", + "location": { + "path": "/this/is/not/an/existing/path", + "server": { + "host": "127.0.0.1", + "port": "22" + } + }, + "size": 4194304 + } +} +{ "execute": "x-blockdev-create", + "arguments": { + "driver": "ssh", + "location": { + "path": "$TEST_IMG_FILE", + "user": "invalid user", + "server": { + "host": "127.0.0.1", + "port": "22" + } + }, + "size": 4194304 + } +} +{ "execute": "quit" } +EOF + +# success, all done +echo "*** done" +rm -f $seq.full +status=0 diff --git a/tests/qemu-iotests/207.out b/tests/qemu-iotests/207.out new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/tests/qemu-iotests/207.out @@ -XXX,XX +XXX,XX @@ +QA output created by 207 + +=== Successful image creation (defaults) === + +Testing: +QMP_VERSION +{"return": {}} +{"return": {}} +{"return": {}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}} + +image: json:{"driver": "IMGFMT", "file": {"server.host": "127.0.0.1", "server.port": "22", "driver": "ssh", "path": "TEST_DIR/t.IMGFMT"}} +file format: IMGFMT +virtual size: 4.0M (4194304 bytes) + +image: TEST_DIR/t.IMGFMT +file format: IMGFMT +virtual size: 4.0M (4194304 bytes) + +=== Test host-key-check options === + +Testing: +QMP_VERSION +{"return": {}} +{"return": {}} +{"return": {}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}} + +image: json:{"driver": "IMGFMT", "file": {"server.host": "127.0.0.1", "server.port": "22", "driver": "ssh", "path": "TEST_DIR/t.IMGFMT"}} +file format: IMGFMT +virtual size: 8.0M (8388608 bytes) +Testing: +QMP_VERSION +{"return": {}} +{"return": {}} +{"return": {}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}} + +image: json:{"driver": "IMGFMT", "file": {"server.host": "127.0.0.1", "server.port": "22", "driver": "ssh", "path": "TEST_DIR/t.IMGFMT"}} +file format: IMGFMT +virtual size: 4.0M (4194304 bytes) +Testing: +QMP_VERSION +{"return": {}} +{"error": {"class": "GenericError", "desc": "remote host key does not match host_key_check 'wrong'"}} +{"return": {}} +{"return": {}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}} + +image: json:{"driver": "IMGFMT", "file": {"server.host": "127.0.0.1", "server.port": "22", "driver": "ssh", "path": "TEST_DIR/t.IMGFMT"}} +file format: IMGFMT +virtual size: 8.0M (8388608 bytes) +Testing: +QMP_VERSION +{"return": {}} +{"error": {"class": "GenericError", "desc": "remote host key does not match host_key_check 'wrong'"}} +{"return": {}} +{"return": {}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}} + +image: json:{"driver": "IMGFMT", "file": {"server.host": "127.0.0.1", "server.port": "22", "driver": "ssh", "path": "TEST_DIR/t.IMGFMT"}} +file format: IMGFMT +virtual size: 4.0M (4194304 bytes) + +=== Invalid path and user === + +Testing: +QMP_VERSION +{"return": {}} +{"error": {"class": "GenericError", "desc": "failed to open remote file '/this/is/not/an/existing/path': Failed opening remote file (libssh2 error code: -31)"}} +{"error": {"class": "GenericError", "desc": "failed to authenticate using publickey authentication and the identities held by your ssh-agent"}} +{"return": {}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false}} + +*** done diff --git a/tests/qemu-iotests/group b/tests/qemu-iotests/group index XXXXXXX..XXXXXXX 100644 --- a/tests/qemu-iotests/group +++ b/tests/qemu-iotests/group @@ -XXX,XX +XXX,XX @@ 204 rw auto quick 205 rw auto quick 206 rw auto +207 rw auto -- 2.13.6 From: Fam Zheng <famz@redhat.com> Signed-off-by: Fam Zheng <famz@redhat.com> Message-id: 20171225025107.23985-1-famz@redhat.com Reviewed-by: Eric Blake <eblake@redhat.com> Signed-off-by: Max Reitz <mreitz@redhat.com> --- tests/qemu-iotests/153 | 8 +++++--- tests/qemu-iotests/153.out | 7 ++++--- 2 files changed, 9 insertions(+), 6 deletions(-) diff --git a/tests/qemu-iotests/153 b/tests/qemu-iotests/153 index XXXXXXX..XXXXXXX 100755 --- a/tests/qemu-iotests/153 +++ b/tests/qemu-iotests/153 @@ -XXX,XX +XXX,XX @@ _cleanup() { _cleanup_test_img rm -f "${TEST_IMG}.base" + rm -f "${TEST_IMG}.overlay" rm -f "${TEST_IMG}.convert" rm -f "${TEST_IMG}.a" rm -f "${TEST_IMG}.b" @@ -XXX,XX +XXX,XX @@ rm -f "${TEST_IMG}.lnk" &>/dev/null ln -s ${TEST_IMG} "${TEST_IMG}.lnk" || echo "Failed to create link" _run_qemu_with_images "${TEST_IMG}.lnk" "${TEST_IMG}" -echo -echo "== Closing an image should unlock it ==" _launch_qemu _send_qemu_cmd $QEMU_HANDLE \ @@ -XXX,XX +XXX,XX @@ _send_qemu_cmd $QEMU_HANDLE \ _run_cmd $QEMU_IO "${TEST_IMG}" -c 'write 0 512' -echo "Closing drive" +echo "Creating overlay with qemu-img when the guest is running should be allowed" +_run_cmd $QEMU_IMG create -f $IMGFMT -b "${TEST_IMG}" "${TEST_IMG}.overlay" + +echo "== Closing an image should unlock it ==" _send_qemu_cmd $QEMU_HANDLE \ "{ 'execute': 'human-monitor-command', 'arguments': { 'command-line': 'drive_del d0' } }" \ diff --git a/tests/qemu-iotests/153.out b/tests/qemu-iotests/153.out index XXXXXXX..XXXXXXX 100644 --- a/tests/qemu-iotests/153.out +++ b/tests/qemu-iotests/153.out @@ -XXX,XX +XXX,XX @@ Is another process using the image? == Symbolic link == QEMU_PROG: -drive if=none,file=TEST_DIR/t.qcow2: Failed to get "write" lock Is another process using the image? - -== Closing an image should unlock it == {"return": {}} Adding drive _qemu_io_wrapper TEST_DIR/t.qcow2 -c write 0 512 can't open device TEST_DIR/t.qcow2: Failed to get "write" lock Is another process using the image? -Closing drive +Creating overlay with qemu-img when the guest is running should be allowed + +_qemu_img_wrapper create -f qcow2 -b TEST_DIR/t.qcow2 TEST_DIR/t.qcow2.overlay +== Closing an image should unlock it == _qemu_io_wrapper TEST_DIR/t.qcow2 -c write 0 512 Adding two and closing one -- 2.13.6 From: Fam Zheng <famz@redhat.com> The AFL image is to exercise the code validating image size, which doesn't work on 32 bit or when out of memory (there is a large allocation before the interesting point). So check that and skip the test, instead of faking the result. Signed-off-by: Fam Zheng <famz@redhat.com> Message-id: 20180301011413.11531-1-famz@redhat.com Reviewed-by: Eric Blake <eblake@redhat.com> Signed-off-by: Max Reitz <mreitz@redhat.com> --- tests/qemu-iotests/059 | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/tests/qemu-iotests/059 b/tests/qemu-iotests/059 index XXXXXXX..XXXXXXX 100755 --- a/tests/qemu-iotests/059 +++ b/tests/qemu-iotests/059 @@ -XXX,XX +XXX,XX @@ done echo echo "=== Testing afl image with a very large capacity ===" _use_sample_img afl9.vmdk.bz2 -# The sed makes this test pass on machines with little RAM -# (and also with 32 bit builds) -_img_info | sed -e 's/Cannot allocate memory/Invalid argument/' +_img_info | grep -q 'Cannot allocate memory' && _notrun "Insufficent memory, skipped test" +_img_info _cleanup_test_img # success, all done -- 2.13.6 From: Alberto Garcia <berto@igalia.com> This patch tweaks TestParallelOps in iotest 030 so it allocates data in smaller regions (256KB/512KB instead of 512KB/1MB) and the block-stream job in test_stream_commit() only needs to copy data that is at the very end of the image. This way when the block-stream job is awakened it will finish right away without any chance of being stopped by block_job_sleep_ns(). This triggers the bug that was fixed by 3d5d319e1221082974711af1d09d82f and 1a63a907507fbbcfaee3f622907ec24 and is therefore a more useful test case for parallel block jobs. After this patch the aforementiond bug can also be reproduced with the test_stream_parallel() test case. Since with this change the stream job in test_stream_commit() finishes early, this patch introduces a similar test case where both jobs are slowed down so they can actually run in parallel. Signed-off-by: Alberto Garcia <berto@igalia.com> Cc: John Snow <jsnow@redhat.com> Message-id: 20180306130121.30243-1-berto@igalia.com Signed-off-by: Max Reitz <mreitz@redhat.com> --- tests/qemu-iotests/030 | 52 ++++++++++++++++++++++++++++++++++++++-------- tests/qemu-iotests/030.out | 4 ++-- 2 files changed, 45 insertions(+), 11 deletions(-) diff --git a/tests/qemu-iotests/030 b/tests/qemu-iotests/030 index XXXXXXX..XXXXXXX 100755 --- a/tests/qemu-iotests/030 +++ b/tests/qemu-iotests/030 @@ -XXX,XX +XXX,XX @@ class TestSingleDrive(iotests.QMPTestCase): class TestParallelOps(iotests.QMPTestCase): num_ops = 4 # Number of parallel block-stream operations num_imgs = num_ops * 2 + 1 - image_len = num_ops * 1024 * 1024 + image_len = num_ops * 512 * 1024 imgs = [] def setUp(self): @@ -XXX,XX +XXX,XX @@ class TestParallelOps(iotests.QMPTestCase): '-o', 'backing_file=%s' % self.imgs[i-1], self.imgs[i]) # Put data into the images we are copying data from - for i in range(self.num_imgs / 2): - img_index = i * 2 + 1 - # Alternate between 512k and 1M. + odd_img_indexes = [x for x in reversed(range(self.num_imgs)) if x % 2 == 1] + for i in range(len(odd_img_indexes)): + # Alternate between 256KB and 512KB. # This way jobs will not finish in the same order they were created - num_kb = 512 + 512 * (i % 2) + num_kb = 256 + 256 * (i % 2) qemu_io('-f', iotests.imgfmt, - '-c', 'write -P %d %d %d' % (i, i*1024*1024, num_kb * 1024), - self.imgs[img_index]) + '-c', 'write -P 0xFF %dk %dk' % (i * 512, num_kb), + self.imgs[odd_img_indexes[i]]) # Attach the drive to the VM self.vm = iotests.VM() @@ -XXX,XX +XXX,XX @@ class TestParallelOps(iotests.QMPTestCase): self.wait_until_completed(drive='commit-drive0') # Test a block-stream and a block-commit job in parallel - def test_stream_commit(self): + # Here the stream job is supposed to finish quickly in order to reproduce + # the scenario that triggers the bug fixed in 3d5d319e1221 and 1a63a907507 + def test_stream_commit_1(self): self.assertLessEqual(8, self.num_imgs) self.assert_no_active_block_jobs() # Stream from node0 into node2 - result = self.vm.qmp('block-stream', device='node2', job_id='node2') + result = self.vm.qmp('block-stream', device='node2', base_node='node0', job_id='node2') self.assert_qmp(result, 'return', {}) # Commit from the active layer into node3 @@ -XXX,XX +XXX,XX @@ class TestParallelOps(iotests.QMPTestCase): self.assert_no_active_block_jobs() + # This is similar to test_stream_commit_1 but both jobs are slowed + # down so they can run in parallel for a little while. + def test_stream_commit_2(self): + self.assertLessEqual(8, self.num_imgs) + self.assert_no_active_block_jobs() + + # Stream from node0 into node4 + result = self.vm.qmp('block-stream', device='node4', base_node='node0', job_id='node4', speed=1024*1024) + self.assert_qmp(result, 'return', {}) + + # Commit from the active layer into node5 + result = self.vm.qmp('block-commit', device='drive0', base=self.imgs[5], speed=1024*1024) + self.assert_qmp(result, 'return', {}) + + # Wait for all jobs to be finished. + pending_jobs = ['node4', 'drive0'] + while len(pending_jobs) > 0: + for event in self.vm.get_qmp_events(wait=True): + if event['event'] == 'BLOCK_JOB_COMPLETED': + node_name = self.dictpath(event, 'data/device') + self.assertTrue(node_name in pending_jobs) + self.assert_qmp_absent(event, 'data/error') + pending_jobs.remove(node_name) + if event['event'] == 'BLOCK_JOB_READY': + self.assert_qmp(event, 'data/device', 'drive0') + self.assert_qmp(event, 'data/type', 'commit') + self.assert_qmp_absent(event, 'data/error') + self.assertTrue('drive0' in pending_jobs) + self.vm.qmp('block-job-complete', device='drive0') + + self.assert_no_active_block_jobs() + # Test the base_node parameter def test_stream_base_node_name(self): self.assert_no_active_block_jobs() diff --git a/tests/qemu-iotests/030.out b/tests/qemu-iotests/030.out index XXXXXXX..XXXXXXX 100644 --- a/tests/qemu-iotests/030.out +++ b/tests/qemu-iotests/030.out @@ -XXX,XX +XXX,XX @@ -....................... +........................ ---------------------------------------------------------------------- -Ran 23 tests +Ran 24 tests OK -- 2.13.6 From: Stefan Hajnoczi <stefanha@redhat.com> There is a race between the test's 'query-migrate' QMP command after the QMP 'STOP' event and completing the migration: The test case invokes 'query-migrate' upon receiving 'STOP'. At this point the migration thread may still be in the process of completing. Therefore 'query-migrate' can return 'status': 'active' for a brief window of time instead of 'status': 'completed'. This results in qemu-iotests 203 hanging. Solve the race by enabling the 'events' migration capability, which causes QEMU to emit migration-specific QMP events that do not suffer from this race condition. Wait for the QMP 'MIGRATION' event with 'status': 'completed'. Reported-by: Max Reitz <mreitz@redhat.com> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Message-id: 20180305155926.25858-1-stefanha@redhat.com Reviewed-by: Max Reitz <mreitz@redhat.com> Signed-off-by: Max Reitz <mreitz@redhat.com> --- tests/qemu-iotests/203 | 15 +++++++++++---- tests/qemu-iotests/203.out | 5 +++++ 2 files changed, 16 insertions(+), 4 deletions(-) diff --git a/tests/qemu-iotests/203 b/tests/qemu-iotests/203 index XXXXXXX..XXXXXXX 100755 --- a/tests/qemu-iotests/203 +++ b/tests/qemu-iotests/203 @@ -XXX,XX +XXX,XX @@ with iotests.FilePath('disk0.img') as disk0_img_path, \ node_name='drive1-node', iothread='iothread0', force=True)) + iotests.log('Enabling migration QMP events...') + iotests.log(vm.qmp('migrate-set-capabilities', capabilities=[ + { + 'capability': 'events', + 'state': True + } + ])) + iotests.log('Starting migration...') iotests.log(vm.qmp('migrate', uri='exec:cat >/dev/null')) while True: - vm.get_qmp_event(wait=60.0) - result = vm.qmp('query-migrate') - status = result.get('return', {}).get('status', None) - if status == 'completed': + event = vm.event_wait('MIGRATION') + iotests.log(event, filters=[iotests.filter_qmp_event]) + if event['data']['status'] == 'completed': break diff --git a/tests/qemu-iotests/203.out b/tests/qemu-iotests/203.out index XXXXXXX..XXXXXXX 100644 --- a/tests/qemu-iotests/203.out +++ b/tests/qemu-iotests/203.out @@ -XXX,XX +XXX,XX @@ Launching VM... Setting IOThreads... {u'return': {}} {u'return': {}} +Enabling migration QMP events... +{u'return': {}} Starting migration... {u'return': {}} +{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'setup'}, u'event': u'MIGRATION'} +{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'active'}, u'event': u'MIGRATION'} +{u'timestamp': {u'seconds': 'SECS', u'microseconds': 'USECS'}, u'data': {u'status': u'completed'}, u'event': u'MIGRATION'} -- 2.13.6 The following changes since commit 281f327487c9c9b1599f93c589a408bbf4a651b8: Merge remote-tracking branch 'remotes/vivier/tags/m68k-for-2.12-pull-request' into staging (2017-12-22 00:11:36 +0000) are available in the git repository at: git://repo.or.cz/qemu/kevin.git tags/for-upstream for you to fetch changes up to 1a63a907507fbbcfaee3f622907ec244b7eabda8: block: Keep nodes drained between reopen_queue/multiple (2017-12-22 15:05:32 +0100) ---------------------------------------------------------------- Block layer patches ---------------------------------------------------------------- Doug Gale (1): nvme: Add tracing Edgar Kaziakhmedov (1): qcow2: get rid of qcow2_backing_read1 routine Fam Zheng (2): block: Open backing image in force share mode for size probe block: Remove unused bdrv_requests_pending John Snow (1): iotests: fix 197 for vpc Kevin Wolf (27): block: Formats don't need CONSISTENT_READ with NO_IO block: Make bdrv_drain_invoke() recursive block: Call .drain_begin only once in bdrv_drain_all_begin() test-bdrv-drain: Test BlockDriver callbacks for drain block: bdrv_drain_recurse(): Remove unused begin parameter block: Don't wait for requests in bdrv_drain*_end() block: Unify order in drain functions block: Don't acquire AioContext in hmp_qemu_io() block: Document that x-blockdev-change breaks quorum children list block: Assert drain_all is only called from main AioContext block: Make bdrv_drain() driver callbacks non-recursive test-bdrv-drain: Test callback for bdrv_drain test-bdrv-drain: Test bs->quiesce_counter blockjob: Pause job on draining any job BDS test-bdrv-drain: Test drain vs. block jobs block: Don't block_job_pause_all() in bdrv_drain_all() block: Nested drain_end must still call callbacks test-bdrv-drain: Test nested drain sections block: Don't notify parents in drain call chain block: Add bdrv_subtree_drained_begin/end() test-bdrv-drain: Tests for bdrv_subtree_drain test-bdrv-drain: Test behaviour in coroutine context test-bdrv-drain: Recursive draining with multiple parents block: Allow graph changes in subtree drained section test-bdrv-drain: Test graph changes in drained section commit: Simplify reopen of base block: Keep nodes drained between reopen_queue/multiple Thomas Huth (3): block: Remove the obsolete -drive boot=on|off parameter block: Remove the deprecated -hdachs option block: Mention -drive cyls/heads/secs/trans/serial/addr in deprecation chapter qapi/block-core.json | 4 + block/qcow2.h | 3 - include/block/block.h | 15 +- include/block/block_int.h | 6 +- block.c | 75 ++++- block/commit.c | 8 +- block/io.c | 164 +++++++--- block/qcow2.c | 51 +-- block/replication.c | 6 + blockdev.c | 11 - blockjob.c | 22 +- hmp.c | 6 - hw/block/nvme.c | 349 +++++++++++++++++---- qemu-io-cmds.c | 3 + tests/test-bdrv-drain.c | 651 +++++++++++++++++++++++++++++++++++++++ vl.c | 86 +----- hw/block/trace-events | 93 ++++++ qemu-doc.texi | 29 +- qemu-options.hx | 19 +- tests/Makefile.include | 2 + tests/qemu-iotests/197 | 4 + tests/qemu-iotests/common.filter | 3 +- 22 files changed, 1294 insertions(+), 316 deletions(-) create mode 100644 tests/test-bdrv-drain.c Commit 1f4ad7d fixed 'qemu-img info' for raw images that are currently in use as a mirror target. It is not enough for image formats, though, as these still unconditionally request BLK_PERM_CONSISTENT_READ. As this permission is geared towards whether the guest-visible data is consistent, and has no impact on whether the metadata is sane, and 'qemu-img info' does not read guest-visible data (except for the raw format), it makes sense to not require BLK_PERM_CONSISTENT_READ if there is not going to be any guest I/O performed, regardless of image format. Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ void bdrv_format_default_perms(BlockDriverState *bs, BdrvChild *c, assert(role == &child_backing || role == &child_file); if (!backing) { + int flags = bdrv_reopen_get_flags(reopen_queue, bs); + /* Apart from the modifications below, the same permissions are * forwarded and left alone as for filters */ bdrv_filter_default_perms(bs, c, role, reopen_queue, perm, shared, @@ -XXX,XX +XXX,XX @@ void bdrv_format_default_perms(BlockDriverState *bs, BdrvChild *c, /* bs->file always needs to be consistent because of the metadata. We * can never allow other users to resize or write to it. */ - perm |= BLK_PERM_CONSISTENT_READ; + if (!(flags & BDRV_O_NO_IO)) { + perm |= BLK_PERM_CONSISTENT_READ; + } shared &= ~(BLK_PERM_WRITE | BLK_PERM_RESIZE); } else { /* We want consistent read from backing files if the parent needs it. -- 2.13.6 From: John Snow <jsnow@redhat.com> VPC has some difficulty creating geometries of particular size. However, we can indeed force it to use a literal one, so let's do that for the sake of test 197, which is testing some specific offsets. Signed-off-by: John Snow <jsnow@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Lukáš Doktor <ldoktor@redhat.com> --- tests/qemu-iotests/197 | 4 ++++ tests/qemu-iotests/common.filter | 3 ++- 2 files changed, 6 insertions(+), 1 deletion(-) diff --git a/tests/qemu-iotests/197 b/tests/qemu-iotests/197 index XXXXXXX..XXXXXXX 100755 --- a/tests/qemu-iotests/197 +++ b/tests/qemu-iotests/197 @@ -XXX,XX +XXX,XX @@ echo '=== Copy-on-read ===' echo # Prep the images +# VPC rounds image sizes to a specific geometry, force a specific size. +if [ "$IMGFMT" = "vpc" ]; then + IMGOPTS=$(_optstr_add "$IMGOPTS" "force_size") +fi _make_test_img 4G $QEMU_IO -c "write -P 55 3G 1k" "$TEST_IMG" | _filter_qemu_io IMGPROTO=file IMGFMT=qcow2 IMGOPTS= TEST_IMG_FILE="$TEST_WRAP" \ 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_create() -e "s# log_size=[0-9]\\+##g" \ -e "s# refcount_bits=[0-9]\\+##g" \ -e "s# key-secret=[a-zA-Z0-9]\\+##g" \ - -e "s# iter-time=[0-9]\\+##g" + -e "s# iter-time=[0-9]\\+##g" \ + -e "s# force_size=\\(on\\|off\\)##g" } _filter_img_info() -- 2.13.6 This change separates bdrv_drain_invoke(), which calls the BlockDriver drain callbacks, from bdrv_drain_recurse(). Instead, the function performs its own recursion now. One reason for this is that bdrv_drain_recurse() can be called multiple times by bdrv_drain_all_begin(), but the callbacks may only be called once. The separation is necessary to fix this bug. The other reason is that we intend to go to a model where we call all driver callbacks first, and only then start polling. This is not fully achieved yet with this patch, as bdrv_drain_invoke() contains a BDRV_POLL_WHILE() loop for the block driver callbacks, which can still call callbacks for any unrelated event. It's a step in this direction anyway. Cc: qemu-stable@nongnu.org Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> --- block/io.c | 14 +++++++++++--- 1 file changed, 11 insertions(+), 3 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 @@ static void coroutine_fn bdrv_drain_invoke_entry(void *opaque) bdrv_wakeup(bs); } +/* Recursively call BlockDriver.bdrv_co_drain_begin/end callbacks */ static void bdrv_drain_invoke(BlockDriverState *bs, bool begin) { + BdrvChild *child, *tmp; BdrvCoDrainData data = { .bs = bs, .done = false, .begin = begin}; if (!bs->drv || (begin && !bs->drv->bdrv_co_drain_begin) || @@ -XXX,XX +XXX,XX @@ static void bdrv_drain_invoke(BlockDriverState *bs, bool begin) data.co = qemu_coroutine_create(bdrv_drain_invoke_entry, &data); bdrv_coroutine_enter(bs, data.co); BDRV_POLL_WHILE(bs, !data.done); + + QLIST_FOREACH_SAFE(child, &bs->children, next, tmp) { + bdrv_drain_invoke(child->bs, begin); + } } static bool bdrv_drain_recurse(BlockDriverState *bs, bool begin) @@ -XXX,XX +XXX,XX @@ static bool bdrv_drain_recurse(BlockDriverState *bs, bool begin) BdrvChild *child, *tmp; bool waited; - /* Ensure any pending metadata writes are submitted to bs->file. */ - bdrv_drain_invoke(bs, begin); - /* Wait for drained requests to finish */ waited = BDRV_POLL_WHILE(bs, atomic_read(&bs->in_flight) > 0); @@ -XXX,XX +XXX,XX @@ void bdrv_drained_begin(BlockDriverState *bs) bdrv_parent_drained_begin(bs); } + bdrv_drain_invoke(bs, true); bdrv_drain_recurse(bs, true); } @@ -XXX,XX +XXX,XX @@ void bdrv_drained_end(BlockDriverState *bs) } bdrv_parent_drained_end(bs); + bdrv_drain_invoke(bs, false); bdrv_drain_recurse(bs, false); aio_enable_external(bdrv_get_aio_context(bs)); } @@ -XXX,XX +XXX,XX @@ void bdrv_drain_all_begin(void) aio_context_acquire(aio_context); for (bs = bdrv_first(&it); bs; bs = bdrv_next(&it)) { if (aio_context == bdrv_get_aio_context(bs)) { + /* FIXME Calling this multiple times is wrong */ + bdrv_drain_invoke(bs, true); waited |= bdrv_drain_recurse(bs, true); } } @@ -XXX,XX +XXX,XX @@ void bdrv_drain_all_end(void) aio_context_acquire(aio_context); aio_enable_external(aio_context); bdrv_parent_drained_end(bs); + bdrv_drain_invoke(bs, false); bdrv_drain_recurse(bs, false); aio_context_release(aio_context); } -- 2.13.6 bdrv_drain_all_begin() used to call the .bdrv_co_drain_begin() driver callback inside its polling loop. This means that how many times it got called for each node depended on long it had to poll the event loop. This is obviously not right and results in nodes that stay drained even after bdrv_drain_all_end(), which calls .bdrv_co_drain_begin() once per node. Fix bdrv_drain_all_begin() to call the callback only once, too. Cc: qemu-stable@nongnu.org Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> --- block/io.c | 3 +-- 1 file changed, 1 insertion(+), 2 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 @@ void bdrv_drain_all_begin(void) aio_context_acquire(aio_context); bdrv_parent_drained_begin(bs); aio_disable_external(aio_context); + bdrv_drain_invoke(bs, true); aio_context_release(aio_context); if (!g_slist_find(aio_ctxs, aio_context)) { @@ -XXX,XX +XXX,XX @@ void bdrv_drain_all_begin(void) aio_context_acquire(aio_context); for (bs = bdrv_first(&it); bs; bs = bdrv_next(&it)) { if (aio_context == bdrv_get_aio_context(bs)) { - /* FIXME Calling this multiple times is wrong */ - bdrv_drain_invoke(bs, true); waited |= bdrv_drain_recurse(bs, true); } } -- 2.13.6 This adds a test case that the BlockDriver callbacks for drain are called in bdrv_drained_all_begin/end(), and that both of them are called exactly once. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> --- tests/test-bdrv-drain.c | 137 ++++++++++++++++++++++++++++++++++++++++++++++++ tests/Makefile.include | 2 + 2 files changed, 139 insertions(+) create mode 100644 tests/test-bdrv-drain.c diff --git a/tests/test-bdrv-drain.c b/tests/test-bdrv-drain.c new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/tests/test-bdrv-drain.c @@ -XXX,XX +XXX,XX @@ +/* + * Block node draining tests + * + * Copyright (c) 2017 Kevin Wolf <kwolf@redhat.com> + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to deal + * in the Software without restriction, including without limitation the rights + * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell + * copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN + * THE SOFTWARE. + */ + +#include "qemu/osdep.h" +#include "block/block.h" +#include "sysemu/block-backend.h" +#include "qapi/error.h" + +typedef struct BDRVTestState { + int drain_count; +} BDRVTestState; + +static void coroutine_fn bdrv_test_co_drain_begin(BlockDriverState *bs) +{ + BDRVTestState *s = bs->opaque; + s->drain_count++; +} + +static void coroutine_fn bdrv_test_co_drain_end(BlockDriverState *bs) +{ + BDRVTestState *s = bs->opaque; + s->drain_count--; +} + +static void bdrv_test_close(BlockDriverState *bs) +{ + BDRVTestState *s = bs->opaque; + g_assert_cmpint(s->drain_count, >, 0); +} + +static int coroutine_fn bdrv_test_co_preadv(BlockDriverState *bs, + uint64_t offset, uint64_t bytes, + QEMUIOVector *qiov, int flags) +{ + /* We want this request to stay until the polling loop in drain waits for + * it to complete. We need to sleep a while as bdrv_drain_invoke() comes + * first and polls its result, too, but it shouldn't accidentally complete + * this request yet. */ + qemu_co_sleep_ns(QEMU_CLOCK_REALTIME, 100000); + + return 0; +} + +static BlockDriver bdrv_test = { + .format_name = "test", + .instance_size = sizeof(BDRVTestState), + + .bdrv_close = bdrv_test_close, + .bdrv_co_preadv = bdrv_test_co_preadv, + + .bdrv_co_drain_begin = bdrv_test_co_drain_begin, + .bdrv_co_drain_end = bdrv_test_co_drain_end, +}; + +static void aio_ret_cb(void *opaque, int ret) +{ + int *aio_ret = opaque; + *aio_ret = ret; +} + +static void test_drv_cb_drain_all(void) +{ + BlockBackend *blk; + BlockDriverState *bs; + BDRVTestState *s; + BlockAIOCB *acb; + int aio_ret; + + QEMUIOVector qiov; + struct iovec iov = { + .iov_base = NULL, + .iov_len = 0, + }; + qemu_iovec_init_external(&qiov, &iov, 1); + + blk = blk_new(BLK_PERM_ALL, BLK_PERM_ALL); + bs = bdrv_new_open_driver(&bdrv_test, "test-node", BDRV_O_RDWR, + &error_abort); + s = bs->opaque; + blk_insert_bs(blk, bs, &error_abort); + + /* Simple bdrv_drain_all_begin/end pair, check that CBs are called */ + g_assert_cmpint(s->drain_count, ==, 0); + bdrv_drain_all_begin(); + g_assert_cmpint(s->drain_count, ==, 1); + bdrv_drain_all_end(); + g_assert_cmpint(s->drain_count, ==, 0); + + /* Now do the same while a request is pending */ + aio_ret = -EINPROGRESS; + acb = blk_aio_preadv(blk, 0, &qiov, 0, aio_ret_cb, &aio_ret); + g_assert(acb != NULL); + g_assert_cmpint(aio_ret, ==, -EINPROGRESS); + + g_assert_cmpint(s->drain_count, ==, 0); + bdrv_drain_all_begin(); + g_assert_cmpint(aio_ret, ==, 0); + g_assert_cmpint(s->drain_count, ==, 1); + bdrv_drain_all_end(); + g_assert_cmpint(s->drain_count, ==, 0); + + bdrv_unref(bs); + blk_unref(blk); +} + +int main(int argc, char **argv) +{ + bdrv_init(); + qemu_init_main_loop(&error_abort); + + g_test_init(&argc, &argv, NULL); + + g_test_add_func("/bdrv-drain/driver-cb/drain_all", test_drv_cb_drain_all); + + return g_test_run(); +} diff --git a/tests/Makefile.include b/tests/Makefile.include index XXXXXXX..XXXXXXX 100644 --- a/tests/Makefile.include +++ b/tests/Makefile.include @@ -XXX,XX +XXX,XX @@ gcov-files-test-thread-pool-y = thread-pool.c gcov-files-test-hbitmap-y = util/hbitmap.c check-unit-y += tests/test-hbitmap$(EXESUF) gcov-files-test-hbitmap-y = blockjob.c +check-unit-y += tests/test-bdrv-drain$(EXESUF) check-unit-y += tests/test-blockjob$(EXESUF) check-unit-y += tests/test-blockjob-txn$(EXESUF) check-unit-y += tests/test-x86-cpuid$(EXESUF) @@ -XXX,XX +XXX,XX @@ tests/test-coroutine$(EXESUF): tests/test-coroutine.o $(test-block-obj-y) tests/test-aio$(EXESUF): tests/test-aio.o $(test-block-obj-y) tests/test-aio-multithread$(EXESUF): tests/test-aio-multithread.o $(test-block-obj-y) tests/test-throttle$(EXESUF): tests/test-throttle.o $(test-block-obj-y) +tests/test-bdrv-drain$(EXESUF): tests/test-bdrv-drain.o $(test-block-obj-y) $(test-util-obj-y) tests/test-blockjob$(EXESUF): tests/test-blockjob.o $(test-block-obj-y) $(test-util-obj-y) tests/test-blockjob-txn$(EXESUF): tests/test-blockjob-txn.o $(test-block-obj-y) $(test-util-obj-y) tests/test-thread-pool$(EXESUF): tests/test-thread-pool.o $(test-block-obj-y) -- 2.13.6 Now that the bdrv_drain_invoke() calls are pulled up to the callers of bdrv_drain_recurse(), the 'begin' parameter isn't needed any more. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> --- block/io.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 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 @@ static void bdrv_drain_invoke(BlockDriverState *bs, bool begin) } } -static bool bdrv_drain_recurse(BlockDriverState *bs, bool begin) +static bool bdrv_drain_recurse(BlockDriverState *bs) { BdrvChild *child, *tmp; bool waited; @@ -XXX,XX +XXX,XX @@ static bool bdrv_drain_recurse(BlockDriverState *bs, bool begin) */ bdrv_ref(bs); } - waited |= bdrv_drain_recurse(bs, begin); + waited |= bdrv_drain_recurse(bs); if (in_main_loop) { bdrv_unref(bs); } @@ -XXX,XX +XXX,XX @@ void bdrv_drained_begin(BlockDriverState *bs) } bdrv_drain_invoke(bs, true); - bdrv_drain_recurse(bs, true); + bdrv_drain_recurse(bs); } void bdrv_drained_end(BlockDriverState *bs) @@ -XXX,XX +XXX,XX @@ void bdrv_drained_end(BlockDriverState *bs) bdrv_parent_drained_end(bs); bdrv_drain_invoke(bs, false); - bdrv_drain_recurse(bs, false); + bdrv_drain_recurse(bs); aio_enable_external(bdrv_get_aio_context(bs)); } @@ -XXX,XX +XXX,XX @@ void bdrv_drain_all_begin(void) aio_context_acquire(aio_context); for (bs = bdrv_first(&it); bs; bs = bdrv_next(&it)) { if (aio_context == bdrv_get_aio_context(bs)) { - waited |= bdrv_drain_recurse(bs, true); + waited |= bdrv_drain_recurse(bs); } } aio_context_release(aio_context); @@ -XXX,XX +XXX,XX @@ void bdrv_drain_all_end(void) aio_enable_external(aio_context); bdrv_parent_drained_end(bs); bdrv_drain_invoke(bs, false); - bdrv_drain_recurse(bs, false); + bdrv_drain_recurse(bs); aio_context_release(aio_context); } -- 2.13.6 The device is drained, so there is no point in waiting for requests at the end of the drained section. Remove the bdrv_drain_recurse() calls there. The bdrv_drain_recurse() calls were introduced in commit 481cad48e5e in order to call the .bdrv_co_drain_end() driver callback. This is now done by a separate bdrv_drain_invoke() call. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> --- block/io.c | 2 -- 1 file changed, 2 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 @@ void bdrv_drained_end(BlockDriverState *bs) bdrv_parent_drained_end(bs); bdrv_drain_invoke(bs, false); - bdrv_drain_recurse(bs); aio_enable_external(bdrv_get_aio_context(bs)); } @@ -XXX,XX +XXX,XX @@ void bdrv_drain_all_end(void) aio_enable_external(aio_context); bdrv_parent_drained_end(bs); bdrv_drain_invoke(bs, false); - bdrv_drain_recurse(bs); aio_context_release(aio_context); } -- 2.13.6 Drain requests are propagated to child nodes, parent nodes and directly to the AioContext. The order in which this happened was different between all combinations of drain/drain_all and begin/end. The correct order is to keep children only drained when their parents are also drained. This means that at the start of a drained section, the AioContext needs to be drained first, the parents second and only then the children. The correct order for the end of a drained section is the opposite. This patch changes the three other functions to follow the example of bdrv_drained_begin(), which is the only one that got it right. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> --- block/io.c | 12 ++++++++---- 1 file changed, 8 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 @@ void bdrv_drained_begin(BlockDriverState *bs) return; } + /* Stop things in parent-to-child order */ if (atomic_fetch_inc(&bs->quiesce_counter) == 0) { aio_disable_external(bdrv_get_aio_context(bs)); bdrv_parent_drained_begin(bs); @@ -XXX,XX +XXX,XX @@ void bdrv_drained_end(BlockDriverState *bs) return; } - bdrv_parent_drained_end(bs); + /* Re-enable things in child-to-parent order */ bdrv_drain_invoke(bs, false); + bdrv_parent_drained_end(bs); aio_enable_external(bdrv_get_aio_context(bs)); } @@ -XXX,XX +XXX,XX @@ void bdrv_drain_all_begin(void) for (bs = bdrv_first(&it); bs; bs = bdrv_next(&it)) { AioContext *aio_context = bdrv_get_aio_context(bs); + /* Stop things in parent-to-child order */ aio_context_acquire(aio_context); - bdrv_parent_drained_begin(bs); aio_disable_external(aio_context); + bdrv_parent_drained_begin(bs); bdrv_drain_invoke(bs, true); aio_context_release(aio_context); @@ -XXX,XX +XXX,XX @@ void bdrv_drain_all_end(void) for (bs = bdrv_first(&it); bs; bs = bdrv_next(&it)) { AioContext *aio_context = bdrv_get_aio_context(bs); + /* Re-enable things in child-to-parent order */ aio_context_acquire(aio_context); - aio_enable_external(aio_context); - bdrv_parent_drained_end(bs); bdrv_drain_invoke(bs, false); + bdrv_parent_drained_end(bs); + aio_enable_external(aio_context); aio_context_release(aio_context); } -- 2.13.6 Commit 15afd94a047 added code to acquire and release the AioContext in qemuio_command(). This means that the lock is taken twice now in the call path from hmp_qemu_io(). This causes BDRV_POLL_WHILE() to hang for any requests issued to nodes in a non-mainloop AioContext. Dropping the first locking from hmp_qemu_io() fixes the problem. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> --- hmp.c | 6 ------ 1 file changed, 6 deletions(-) diff --git a/hmp.c b/hmp.c index XXXXXXX..XXXXXXX 100644 --- a/hmp.c +++ b/hmp.c @@ -XXX,XX +XXX,XX @@ void hmp_qemu_io(Monitor *mon, const QDict *qdict) { BlockBackend *blk; BlockBackend *local_blk = NULL; - AioContext *aio_context; const char* device = qdict_get_str(qdict, "device"); const char* command = qdict_get_str(qdict, "command"); Error *err = NULL; @@ -XXX,XX +XXX,XX @@ void hmp_qemu_io(Monitor *mon, const QDict *qdict) } } - aio_context = blk_get_aio_context(blk); - aio_context_acquire(aio_context); - /* * Notably absent: Proper permission management. This is sad, but it seems * almost impossible to achieve without changing the semantics and thereby @@ -XXX,XX +XXX,XX @@ void hmp_qemu_io(Monitor *mon, const QDict *qdict) */ qemuio_command(blk, command); - aio_context_release(aio_context); - fail: blk_unref(local_blk); hmp_handle_error(mon, &err); -- 2.13.6 From: Edgar Kaziakhmedov <edgar.kaziakhmedov@virtuozzo.com> Since bdrv_co_preadv does all neccessary checks including reading after the end of the backing file, avoid duplication of verification before bdrv_co_preadv call. Signed-off-by: Edgar Kaziakhmedov <edgar.kaziakhmedov@virtuozzo.com> Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com> Reviewed-by: Eric Blake <eblake@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/qcow2.h | 3 --- block/qcow2.c | 51 ++++++++------------------------------------------- 2 files changed, 8 insertions(+), 46 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 @@ uint32_t offset_to_reftable_index(BDRVQcow2State *s, uint64_t offset) } /* qcow2.c functions */ -int qcow2_backing_read1(BlockDriverState *bs, QEMUIOVector *qiov, - int64_t sector_num, int nb_sectors); - int64_t qcow2_refcount_metadata_size(int64_t clusters, size_t cluster_size, int refcount_order, bool generous_increase, uint64_t *refblock_count); 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 coroutine_fn qcow2_co_get_block_status(BlockDriverState *bs, return status; } -/* handle reading after the end of the backing file */ -int qcow2_backing_read1(BlockDriverState *bs, QEMUIOVector *qiov, - int64_t offset, int bytes) -{ - uint64_t bs_size = bs->total_sectors * BDRV_SECTOR_SIZE; - int n1; - - if ((offset + bytes) <= bs_size) { - return bytes; - } - - if (offset >= bs_size) { - n1 = 0; - } else { - n1 = bs_size - offset; - } - - qemu_iovec_memset(qiov, n1, 0, bytes - n1); - - return n1; -} - static coroutine_fn int qcow2_co_preadv(BlockDriverState *bs, uint64_t offset, uint64_t bytes, QEMUIOVector *qiov, int flags) { BDRVQcow2State *s = bs->opaque; - int offset_in_cluster, n1; + int offset_in_cluster; int ret; unsigned int cur_bytes; /* number of bytes in current iteration */ uint64_t cluster_offset = 0; @@ -XXX,XX +XXX,XX @@ static coroutine_fn int qcow2_co_preadv(BlockDriverState *bs, uint64_t offset, case QCOW2_CLUSTER_UNALLOCATED: if (bs->backing) { - /* read from the base image */ - n1 = qcow2_backing_read1(bs->backing->bs, &hd_qiov, - offset, cur_bytes); - if (n1 > 0) { - QEMUIOVector local_qiov; - - qemu_iovec_init(&local_qiov, hd_qiov.niov); - qemu_iovec_concat(&local_qiov, &hd_qiov, 0, n1); - - BLKDBG_EVENT(bs->file, BLKDBG_READ_BACKING_AIO); - qemu_co_mutex_unlock(&s->lock); - ret = bdrv_co_preadv(bs->backing, offset, n1, - &local_qiov, 0); - qemu_co_mutex_lock(&s->lock); - - qemu_iovec_destroy(&local_qiov); - - if (ret < 0) { - goto fail; - } + BLKDBG_EVENT(bs->file, BLKDBG_READ_BACKING_AIO); + qemu_co_mutex_unlock(&s->lock); + ret = bdrv_co_preadv(bs->backing, offset, cur_bytes, + &hd_qiov, 0); + qemu_co_mutex_lock(&s->lock); + if (ret < 0) { + goto fail; } } else { /* Note: in this case, no need to wait */ -- 2.13.6 Removing a quorum child node with x-blockdev-change results in a quorum driver state that cannot be recreated with create options because it would require a list with gaps. This causes trouble in at least .bdrv_refresh_filename(). Document this problem so that we won't accidentally mark the command stable without having addressed it. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Alberto Garcia <berto@igalia.com> --- qapi/block-core.json | 4 ++++ 1 file changed, 4 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 @@ # does not support all kinds of operations, all kinds of children, nor # all block drivers. # +# FIXME Removing children from a quorum node means introducing gaps in the +# child indices. This cannot be represented in the 'children' list of +# BlockdevOptionsQuorum, as returned by .bdrv_refresh_filename(). +# # Warning: The data in a new quorum child MUST be consistent with that of # the rest of the array. # -- 2.13.6 From: Doug Gale <doug16k@gmail.com> Add trace output for commands, errors, and undefined behavior. Add guest error log output for undefined behavior. Report invalid undefined accesses to MMIO. Annotate unlikely error checks with unlikely. Signed-off-by: Doug Gale <doug16k@gmail.com> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- hw/block/nvme.c | 349 ++++++++++++++++++++++++++++++++++++++++++-------- hw/block/trace-events | 93 ++++++++++++++ 2 files changed, 390 insertions(+), 52 deletions(-) diff --git a/hw/block/nvme.c b/hw/block/nvme.c index XXXXXXX..XXXXXXX 100644 --- a/hw/block/nvme.c +++ b/hw/block/nvme.c @@ -XXX,XX +XXX,XX @@ #include "qapi/visitor.h" #include "sysemu/block-backend.h" +#include "qemu/log.h" +#include "trace.h" #include "nvme.h" +#define NVME_GUEST_ERR(trace, fmt, ...) \ + do { \ + (trace_##trace)(__VA_ARGS__); \ + qemu_log_mask(LOG_GUEST_ERROR, #trace \ + " in %s: " fmt "\n", __func__, ## __VA_ARGS__); \ + } while (0) + static void nvme_process_sq(void *opaque); static void nvme_addr_read(NvmeCtrl *n, hwaddr addr, void *buf, int size) @@ -XXX,XX +XXX,XX @@ static void nvme_isr_notify(NvmeCtrl *n, NvmeCQueue *cq) { if (cq->irq_enabled) { if (msix_enabled(&(n->parent_obj))) { + trace_nvme_irq_msix(cq->vector); msix_notify(&(n->parent_obj), cq->vector); } else { + trace_nvme_irq_pin(); pci_irq_pulse(&n->parent_obj); } + } else { + trace_nvme_irq_masked(); } } @@ -XXX,XX +XXX,XX @@ static uint16_t nvme_map_prp(QEMUSGList *qsg, QEMUIOVector *iov, uint64_t prp1, trans_len = MIN(len, trans_len); int num_prps = (len >> n->page_bits) + 1; - if (!prp1) { + if (unlikely(!prp1)) { + trace_nvme_err_invalid_prp(); return NVME_INVALID_FIELD | NVME_DNR; } else if (n->cmbsz && prp1 >= n->ctrl_mem.addr && prp1 < n->ctrl_mem.addr + int128_get64(n->ctrl_mem.size)) { @@ -XXX,XX +XXX,XX @@ static uint16_t nvme_map_prp(QEMUSGList *qsg, QEMUIOVector *iov, uint64_t prp1, } len -= trans_len; if (len) { - if (!prp2) { + if (unlikely(!prp2)) { + trace_nvme_err_invalid_prp2_missing(); goto unmap; } if (len > n->page_size) { @@ -XXX,XX +XXX,XX @@ static uint16_t nvme_map_prp(QEMUSGList *qsg, QEMUIOVector *iov, uint64_t prp1, uint64_t prp_ent = le64_to_cpu(prp_list[i]); if (i == n->max_prp_ents - 1 && len > n->page_size) { - if (!prp_ent || prp_ent & (n->page_size - 1)) { + if (unlikely(!prp_ent || prp_ent & (n->page_size - 1))) { + trace_nvme_err_invalid_prplist_ent(prp_ent); goto unmap; } @@ -XXX,XX +XXX,XX @@ static uint16_t nvme_map_prp(QEMUSGList *qsg, QEMUIOVector *iov, uint64_t prp1, prp_ent = le64_to_cpu(prp_list[i]); } - if (!prp_ent || prp_ent & (n->page_size - 1)) { + if (unlikely(!prp_ent || prp_ent & (n->page_size - 1))) { + trace_nvme_err_invalid_prplist_ent(prp_ent); goto unmap; } @@ -XXX,XX +XXX,XX @@ static uint16_t nvme_map_prp(QEMUSGList *qsg, QEMUIOVector *iov, uint64_t prp1, i++; } } else { - if (prp2 & (n->page_size - 1)) { + if (unlikely(prp2 & (n->page_size - 1))) { + trace_nvme_err_invalid_prp2_align(prp2); goto unmap; } if (qsg->nsg) { @@ -XXX,XX +XXX,XX @@ static uint16_t nvme_dma_read_prp(NvmeCtrl *n, uint8_t *ptr, uint32_t len, QEMUIOVector iov; uint16_t status = NVME_SUCCESS; + trace_nvme_dma_read(prp1, prp2); + if (nvme_map_prp(&qsg, &iov, prp1, prp2, len, n)) { return NVME_INVALID_FIELD | NVME_DNR; } if (qsg.nsg > 0) { - if (dma_buf_read(ptr, len, &qsg)) { + if (unlikely(dma_buf_read(ptr, len, &qsg))) { + trace_nvme_err_invalid_dma(); status = NVME_INVALID_FIELD | NVME_DNR; } qemu_sglist_destroy(&qsg); } else { - if (qemu_iovec_to_buf(&iov, 0, ptr, len) != len) { + if (unlikely(qemu_iovec_to_buf(&iov, 0, ptr, len) != len)) { + trace_nvme_err_invalid_dma(); status = NVME_INVALID_FIELD | NVME_DNR; } qemu_iovec_destroy(&iov); @@ -XXX,XX +XXX,XX @@ static uint16_t nvme_write_zeros(NvmeCtrl *n, NvmeNamespace *ns, NvmeCmd *cmd, uint64_t aio_slba = slba << (data_shift - BDRV_SECTOR_BITS); uint32_t aio_nlb = nlb << (data_shift - BDRV_SECTOR_BITS); - if (slba + nlb > ns->id_ns.nsze) { + if (unlikely(slba + nlb > ns->id_ns.nsze)) { + trace_nvme_err_invalid_lba_range(slba, nlb, ns->id_ns.nsze); return NVME_LBA_RANGE | NVME_DNR; } @@ -XXX,XX +XXX,XX @@ static uint16_t nvme_rw(NvmeCtrl *n, NvmeNamespace *ns, NvmeCmd *cmd, int is_write = rw->opcode == NVME_CMD_WRITE ? 1 : 0; enum BlockAcctType acct = is_write ? BLOCK_ACCT_WRITE : BLOCK_ACCT_READ; - if ((slba + nlb) > ns->id_ns.nsze) { + trace_nvme_rw(is_write ? "write" : "read", nlb, data_size, slba); + + if (unlikely((slba + nlb) > ns->id_ns.nsze)) { block_acct_invalid(blk_get_stats(n->conf.blk), acct); + trace_nvme_err_invalid_lba_range(slba, nlb, ns->id_ns.nsze); return NVME_LBA_RANGE | NVME_DNR; } @@ -XXX,XX +XXX,XX @@ static uint16_t nvme_io_cmd(NvmeCtrl *n, NvmeCmd *cmd, NvmeRequest *req) NvmeNamespace *ns; uint32_t nsid = le32_to_cpu(cmd->nsid); - if (nsid == 0 || nsid > n->num_namespaces) { + if (unlikely(nsid == 0 || nsid > n->num_namespaces)) { + trace_nvme_err_invalid_ns(nsid, n->num_namespaces); return NVME_INVALID_NSID | NVME_DNR; } @@ -XXX,XX +XXX,XX @@ static uint16_t nvme_io_cmd(NvmeCtrl *n, NvmeCmd *cmd, NvmeRequest *req) case NVME_CMD_READ: return nvme_rw(n, ns, cmd, req); default: + trace_nvme_err_invalid_opc(cmd->opcode); return NVME_INVALID_OPCODE | NVME_DNR; } } @@ -XXX,XX +XXX,XX @@ static uint16_t nvme_del_sq(NvmeCtrl *n, NvmeCmd *cmd) NvmeCQueue *cq; uint16_t qid = le16_to_cpu(c->qid); - if (!qid || nvme_check_sqid(n, qid)) { + if (unlikely(!qid || nvme_check_sqid(n, qid))) { + trace_nvme_err_invalid_del_sq(qid); return NVME_INVALID_QID | NVME_DNR; } + trace_nvme_del_sq(qid); + sq = n->sq[qid]; while (!QTAILQ_EMPTY(&sq->out_req_list)) { req = QTAILQ_FIRST(&sq->out_req_list); @@ -XXX,XX +XXX,XX @@ static uint16_t nvme_create_sq(NvmeCtrl *n, NvmeCmd *cmd) uint16_t qflags = le16_to_cpu(c->sq_flags); uint64_t prp1 = le64_to_cpu(c->prp1); - if (!cqid || nvme_check_cqid(n, cqid)) { + trace_nvme_create_sq(prp1, sqid, cqid, qsize, qflags); + + if (unlikely(!cqid || nvme_check_cqid(n, cqid))) { + trace_nvme_err_invalid_create_sq_cqid(cqid); return NVME_INVALID_CQID | NVME_DNR; } - if (!sqid || !nvme_check_sqid(n, sqid)) { + if (unlikely(!sqid || !nvme_check_sqid(n, sqid))) { + trace_nvme_err_invalid_create_sq_sqid(sqid); return NVME_INVALID_QID | NVME_DNR; } - if (!qsize || qsize > NVME_CAP_MQES(n->bar.cap)) { + if (unlikely(!qsize || qsize > NVME_CAP_MQES(n->bar.cap))) { + trace_nvme_err_invalid_create_sq_size(qsize); return NVME_MAX_QSIZE_EXCEEDED | NVME_DNR; } - if (!prp1 || prp1 & (n->page_size - 1)) { + if (unlikely(!prp1 || prp1 & (n->page_size - 1))) { + trace_nvme_err_invalid_create_sq_addr(prp1); return NVME_INVALID_FIELD | NVME_DNR; } - if (!(NVME_SQ_FLAGS_PC(qflags))) { + if (unlikely(!(NVME_SQ_FLAGS_PC(qflags)))) { + trace_nvme_err_invalid_create_sq_qflags(NVME_SQ_FLAGS_PC(qflags)); return NVME_INVALID_FIELD | NVME_DNR; } sq = g_malloc0(sizeof(*sq)); @@ -XXX,XX +XXX,XX @@ static uint16_t nvme_del_cq(NvmeCtrl *n, NvmeCmd *cmd) NvmeCQueue *cq; uint16_t qid = le16_to_cpu(c->qid); - if (!qid || nvme_check_cqid(n, qid)) { + if (unlikely(!qid || nvme_check_cqid(n, qid))) { + trace_nvme_err_invalid_del_cq_cqid(qid); return NVME_INVALID_CQID | NVME_DNR; } cq = n->cq[qid]; - if (!QTAILQ_EMPTY(&cq->sq_list)) { + if (unlikely(!QTAILQ_EMPTY(&cq->sq_list))) { + trace_nvme_err_invalid_del_cq_notempty(qid); return NVME_INVALID_QUEUE_DEL; } + trace_nvme_del_cq(qid); nvme_free_cq(cq, n); return NVME_SUCCESS; } @@ -XXX,XX +XXX,XX @@ static uint16_t nvme_create_cq(NvmeCtrl *n, NvmeCmd *cmd) uint16_t qflags = le16_to_cpu(c->cq_flags); uint64_t prp1 = le64_to_cpu(c->prp1); - if (!cqid || !nvme_check_cqid(n, cqid)) { + trace_nvme_create_cq(prp1, cqid, vector, qsize, qflags, + NVME_CQ_FLAGS_IEN(qflags) != 0); + + if (unlikely(!cqid || !nvme_check_cqid(n, cqid))) { + trace_nvme_err_invalid_create_cq_cqid(cqid); return NVME_INVALID_CQID | NVME_DNR; } - if (!qsize || qsize > NVME_CAP_MQES(n->bar.cap)) { + if (unlikely(!qsize || qsize > NVME_CAP_MQES(n->bar.cap))) { + trace_nvme_err_invalid_create_cq_size(qsize); return NVME_MAX_QSIZE_EXCEEDED | NVME_DNR; } - if (!prp1) { + if (unlikely(!prp1)) { + trace_nvme_err_invalid_create_cq_addr(prp1); return NVME_INVALID_FIELD | NVME_DNR; } - if (vector > n->num_queues) { + if (unlikely(vector > n->num_queues)) { + trace_nvme_err_invalid_create_cq_vector(vector); return NVME_INVALID_IRQ_VECTOR | NVME_DNR; } - if (!(NVME_CQ_FLAGS_PC(qflags))) { + if (unlikely(!(NVME_CQ_FLAGS_PC(qflags)))) { + trace_nvme_err_invalid_create_cq_qflags(NVME_CQ_FLAGS_PC(qflags)); return NVME_INVALID_FIELD | NVME_DNR; } @@ -XXX,XX +XXX,XX @@ static uint16_t nvme_identify_ctrl(NvmeCtrl *n, NvmeIdentify *c) uint64_t prp1 = le64_to_cpu(c->prp1); uint64_t prp2 = le64_to_cpu(c->prp2); + trace_nvme_identify_ctrl(); + return nvme_dma_read_prp(n, (uint8_t *)&n->id_ctrl, sizeof(n->id_ctrl), prp1, prp2); } @@ -XXX,XX +XXX,XX @@ static uint16_t nvme_identify_ns(NvmeCtrl *n, NvmeIdentify *c) uint64_t prp1 = le64_to_cpu(c->prp1); uint64_t prp2 = le64_to_cpu(c->prp2); - if (nsid == 0 || nsid > n->num_namespaces) { + trace_nvme_identify_ns(nsid); + + if (unlikely(nsid == 0 || nsid > n->num_namespaces)) { + trace_nvme_err_invalid_ns(nsid, n->num_namespaces); return NVME_INVALID_NSID | NVME_DNR; } ns = &n->namespaces[nsid - 1]; + return nvme_dma_read_prp(n, (uint8_t *)&ns->id_ns, sizeof(ns->id_ns), prp1, prp2); } @@ -XXX,XX +XXX,XX @@ static uint16_t nvme_identify_nslist(NvmeCtrl *n, NvmeIdentify *c) uint16_t ret; int i, j = 0; + trace_nvme_identify_nslist(min_nsid); + list = g_malloc0(data_len); for (i = 0; i < n->num_namespaces; i++) { if (i < min_nsid) { @@ -XXX,XX +XXX,XX @@ static uint16_t nvme_identify(NvmeCtrl *n, NvmeCmd *cmd) case 0x02: return nvme_identify_nslist(n, c); default: + trace_nvme_err_invalid_identify_cns(le32_to_cpu(c->cns)); return NVME_INVALID_FIELD | NVME_DNR; } } @@ -XXX,XX +XXX,XX @@ static uint16_t nvme_get_feature(NvmeCtrl *n, NvmeCmd *cmd, NvmeRequest *req) switch (dw10) { case NVME_VOLATILE_WRITE_CACHE: result = blk_enable_write_cache(n->conf.blk); + trace_nvme_getfeat_vwcache(result ? "enabled" : "disabled"); break; case NVME_NUMBER_OF_QUEUES: result = cpu_to_le32((n->num_queues - 2) | ((n->num_queues - 2) << 16)); + trace_nvme_getfeat_numq(result); break; default: + trace_nvme_err_invalid_getfeat(dw10); return NVME_INVALID_FIELD | NVME_DNR; } @@ -XXX,XX +XXX,XX @@ static uint16_t nvme_set_feature(NvmeCtrl *n, NvmeCmd *cmd, NvmeRequest *req) blk_set_enable_write_cache(n->conf.blk, dw11 & 1); break; case NVME_NUMBER_OF_QUEUES: + trace_nvme_setfeat_numq((dw11 & 0xFFFF) + 1, + ((dw11 >> 16) & 0xFFFF) + 1, + n->num_queues - 1, n->num_queues - 1); req->cqe.result = cpu_to_le32((n->num_queues - 2) | ((n->num_queues - 2) << 16)); break; default: + trace_nvme_err_invalid_setfeat(dw10); return NVME_INVALID_FIELD | NVME_DNR; } return NVME_SUCCESS; @@ -XXX,XX +XXX,XX @@ static uint16_t nvme_admin_cmd(NvmeCtrl *n, NvmeCmd *cmd, NvmeRequest *req) case NVME_ADM_CMD_GET_FEATURES: return nvme_get_feature(n, cmd, req); default: + trace_nvme_err_invalid_admin_opc(cmd->opcode); return NVME_INVALID_OPCODE | NVME_DNR; } } @@ -XXX,XX +XXX,XX @@ static int nvme_start_ctrl(NvmeCtrl *n) uint32_t page_bits = NVME_CC_MPS(n->bar.cc) + 12; uint32_t page_size = 1 << page_bits; - if (n->cq[0] || n->sq[0] || !n->bar.asq || !n->bar.acq || - n->bar.asq & (page_size - 1) || n->bar.acq & (page_size - 1) || - NVME_CC_MPS(n->bar.cc) < NVME_CAP_MPSMIN(n->bar.cap) || - NVME_CC_MPS(n->bar.cc) > NVME_CAP_MPSMAX(n->bar.cap) || - NVME_CC_IOCQES(n->bar.cc) < NVME_CTRL_CQES_MIN(n->id_ctrl.cqes) || - NVME_CC_IOCQES(n->bar.cc) > NVME_CTRL_CQES_MAX(n->id_ctrl.cqes) || - NVME_CC_IOSQES(n->bar.cc) < NVME_CTRL_SQES_MIN(n->id_ctrl.sqes) || - NVME_CC_IOSQES(n->bar.cc) > NVME_CTRL_SQES_MAX(n->id_ctrl.sqes) || - !NVME_AQA_ASQS(n->bar.aqa) || !NVME_AQA_ACQS(n->bar.aqa)) { + if (unlikely(n->cq[0])) { + trace_nvme_err_startfail_cq(); + return -1; + } + if (unlikely(n->sq[0])) { + trace_nvme_err_startfail_sq(); + return -1; + } + if (unlikely(!n->bar.asq)) { + trace_nvme_err_startfail_nbarasq(); + return -1; + } + if (unlikely(!n->bar.acq)) { + trace_nvme_err_startfail_nbaracq(); + return -1; + } + if (unlikely(n->bar.asq & (page_size - 1))) { + trace_nvme_err_startfail_asq_misaligned(n->bar.asq); + return -1; + } + if (unlikely(n->bar.acq & (page_size - 1))) { + trace_nvme_err_startfail_acq_misaligned(n->bar.acq); + return -1; + } + if (unlikely(NVME_CC_MPS(n->bar.cc) < + NVME_CAP_MPSMIN(n->bar.cap))) { + trace_nvme_err_startfail_page_too_small( + NVME_CC_MPS(n->bar.cc), + NVME_CAP_MPSMIN(n->bar.cap)); + return -1; + } + if (unlikely(NVME_CC_MPS(n->bar.cc) > + NVME_CAP_MPSMAX(n->bar.cap))) { + trace_nvme_err_startfail_page_too_large( + NVME_CC_MPS(n->bar.cc), + NVME_CAP_MPSMAX(n->bar.cap)); + return -1; + } + if (unlikely(NVME_CC_IOCQES(n->bar.cc) < + NVME_CTRL_CQES_MIN(n->id_ctrl.cqes))) { + trace_nvme_err_startfail_cqent_too_small( + NVME_CC_IOCQES(n->bar.cc), + NVME_CTRL_CQES_MIN(n->bar.cap)); + return -1; + } + if (unlikely(NVME_CC_IOCQES(n->bar.cc) > + NVME_CTRL_CQES_MAX(n->id_ctrl.cqes))) { + trace_nvme_err_startfail_cqent_too_large( + NVME_CC_IOCQES(n->bar.cc), + NVME_CTRL_CQES_MAX(n->bar.cap)); + return -1; + } + if (unlikely(NVME_CC_IOSQES(n->bar.cc) < + NVME_CTRL_SQES_MIN(n->id_ctrl.sqes))) { + trace_nvme_err_startfail_sqent_too_small( + NVME_CC_IOSQES(n->bar.cc), + NVME_CTRL_SQES_MIN(n->bar.cap)); + return -1; + } + if (unlikely(NVME_CC_IOSQES(n->bar.cc) > + NVME_CTRL_SQES_MAX(n->id_ctrl.sqes))) { + trace_nvme_err_startfail_sqent_too_large( + NVME_CC_IOSQES(n->bar.cc), + NVME_CTRL_SQES_MAX(n->bar.cap)); + return -1; + } + if (unlikely(!NVME_AQA_ASQS(n->bar.aqa))) { + trace_nvme_err_startfail_asqent_sz_zero(); + return -1; + } + if (unlikely(!NVME_AQA_ACQS(n->bar.aqa))) { + trace_nvme_err_startfail_acqent_sz_zero(); return -1; } @@ -XXX,XX +XXX,XX @@ static int nvme_start_ctrl(NvmeCtrl *n) static void nvme_write_bar(NvmeCtrl *n, hwaddr offset, uint64_t data, unsigned size) { + if (unlikely(offset & (sizeof(uint32_t) - 1))) { + NVME_GUEST_ERR(nvme_ub_mmiowr_misaligned32, + "MMIO write not 32-bit aligned," + " offset=0x%"PRIx64"", offset); + /* should be ignored, fall through for now */ + } + + if (unlikely(size < sizeof(uint32_t))) { + NVME_GUEST_ERR(nvme_ub_mmiowr_toosmall, + "MMIO write smaller than 32-bits," + " offset=0x%"PRIx64", size=%u", + offset, size); + /* should be ignored, fall through for now */ + } + switch (offset) { - case 0xc: + case 0xc: /* INTMS */ + if (unlikely(msix_enabled(&(n->parent_obj)))) { + NVME_GUEST_ERR(nvme_ub_mmiowr_intmask_with_msix, + "undefined access to interrupt mask set" + " when MSI-X is enabled"); + /* should be ignored, fall through for now */ + } n->bar.intms |= data & 0xffffffff; n->bar.intmc = n->bar.intms; + trace_nvme_mmio_intm_set(data & 0xffffffff, + n->bar.intmc); break; - case 0x10: + case 0x10: /* INTMC */ + if (unlikely(msix_enabled(&(n->parent_obj)))) { + NVME_GUEST_ERR(nvme_ub_mmiowr_intmask_with_msix, + "undefined access to interrupt mask clr" + " when MSI-X is enabled"); + /* should be ignored, fall through for now */ + } n->bar.intms &= ~(data & 0xffffffff); n->bar.intmc = n->bar.intms; + trace_nvme_mmio_intm_clr(data & 0xffffffff, + n->bar.intmc); break; - case 0x14: + case 0x14: /* CC */ + trace_nvme_mmio_cfg(data & 0xffffffff); /* Windows first sends data, then sends enable bit */ if (!NVME_CC_EN(data) && !NVME_CC_EN(n->bar.cc) && !NVME_CC_SHN(data) && !NVME_CC_SHN(n->bar.cc)) @@ -XXX,XX +XXX,XX @@ static void nvme_write_bar(NvmeCtrl *n, hwaddr offset, uint64_t data, if (NVME_CC_EN(data) && !NVME_CC_EN(n->bar.cc)) { n->bar.cc = data; - if (nvme_start_ctrl(n)) { + if (unlikely(nvme_start_ctrl(n))) { + trace_nvme_err_startfail(); n->bar.csts = NVME_CSTS_FAILED; } else { + trace_nvme_mmio_start_success(); n->bar.csts = NVME_CSTS_READY; } } else if (!NVME_CC_EN(data) && NVME_CC_EN(n->bar.cc)) { + trace_nvme_mmio_stopped(); nvme_clear_ctrl(n); n->bar.csts &= ~NVME_CSTS_READY; } if (NVME_CC_SHN(data) && !(NVME_CC_SHN(n->bar.cc))) { - nvme_clear_ctrl(n); - n->bar.cc = data; - n->bar.csts |= NVME_CSTS_SHST_COMPLETE; + trace_nvme_mmio_shutdown_set(); + nvme_clear_ctrl(n); + n->bar.cc = data; + n->bar.csts |= NVME_CSTS_SHST_COMPLETE; } else if (!NVME_CC_SHN(data) && NVME_CC_SHN(n->bar.cc)) { - n->bar.csts &= ~NVME_CSTS_SHST_COMPLETE; - n->bar.cc = data; + trace_nvme_mmio_shutdown_cleared(); + n->bar.csts &= ~NVME_CSTS_SHST_COMPLETE; + n->bar.cc = data; + } + break; + case 0x1C: /* CSTS */ + if (data & (1 << 4)) { + NVME_GUEST_ERR(nvme_ub_mmiowr_ssreset_w1c_unsupported, + "attempted to W1C CSTS.NSSRO" + " but CAP.NSSRS is zero (not supported)"); + } else if (data != 0) { + NVME_GUEST_ERR(nvme_ub_mmiowr_ro_csts, + "attempted to set a read only bit" + " of controller status"); + } + break; + case 0x20: /* NSSR */ + if (data == 0x4E564D65) { + trace_nvme_ub_mmiowr_ssreset_unsupported(); + } else { + /* The spec says that writes of other values have no effect */ + return; } break; - case 0x24: + case 0x24: /* AQA */ n->bar.aqa = data & 0xffffffff; + trace_nvme_mmio_aqattr(data & 0xffffffff); break; - case 0x28: + case 0x28: /* ASQ */ n->bar.asq = data; + trace_nvme_mmio_asqaddr(data); break; - case 0x2c: + case 0x2c: /* ASQ hi */ n->bar.asq |= data << 32; + trace_nvme_mmio_asqaddr_hi(data, n->bar.asq); break; - case 0x30: + case 0x30: /* ACQ */ + trace_nvme_mmio_acqaddr(data); n->bar.acq = data; break; - case 0x34: + case 0x34: /* ACQ hi */ n->bar.acq |= data << 32; + trace_nvme_mmio_acqaddr_hi(data, n->bar.acq); break; + case 0x38: /* CMBLOC */ + NVME_GUEST_ERR(nvme_ub_mmiowr_cmbloc_reserved, + "invalid write to reserved CMBLOC" + " when CMBSZ is zero, ignored"); + return; + case 0x3C: /* CMBSZ */ + NVME_GUEST_ERR(nvme_ub_mmiowr_cmbsz_readonly, + "invalid write to read only CMBSZ, ignored"); + return; default: + NVME_GUEST_ERR(nvme_ub_mmiowr_invalid, + "invalid MMIO write," + " offset=0x%"PRIx64", data=%"PRIx64"", + offset, data); break; } } @@ -XXX,XX +XXX,XX @@ static uint64_t nvme_mmio_read(void *opaque, hwaddr addr, unsigned size) uint8_t *ptr = (uint8_t *)&n->bar; uint64_t val = 0; + if (unlikely(addr & (sizeof(uint32_t) - 1))) { + NVME_GUEST_ERR(nvme_ub_mmiord_misaligned32, + "MMIO read not 32-bit aligned," + " offset=0x%"PRIx64"", addr); + /* should RAZ, fall through for now */ + } else if (unlikely(size < sizeof(uint32_t))) { + NVME_GUEST_ERR(nvme_ub_mmiord_toosmall, + "MMIO read smaller than 32-bits," + " offset=0x%"PRIx64"", addr); + /* should RAZ, fall through for now */ + } + if (addr < sizeof(n->bar)) { memcpy(&val, ptr + addr, size); + } else { + NVME_GUEST_ERR(nvme_ub_mmiord_invalid_ofs, + "MMIO read beyond last register," + " offset=0x%"PRIx64", returning 0", addr); } + return val; } @@ -XXX,XX +XXX,XX @@ static void nvme_process_db(NvmeCtrl *n, hwaddr addr, int val) { uint32_t qid; - if (addr & ((1 << 2) - 1)) { + if (unlikely(addr & ((1 << 2) - 1))) { + NVME_GUEST_ERR(nvme_ub_db_wr_misaligned, + "doorbell write not 32-bit aligned," + " offset=0x%"PRIx64", ignoring", addr); return; } if (((addr - 0x1000) >> 2) & 1) { + /* Completion queue doorbell write */ + uint16_t new_head = val & 0xffff; int start_sqs; NvmeCQueue *cq; qid = (addr - (0x1000 + (1 << 2))) >> 3; - if (nvme_check_cqid(n, qid)) { + if (unlikely(nvme_check_cqid(n, qid))) { + NVME_GUEST_ERR(nvme_ub_db_wr_invalid_cq, + "completion queue doorbell write" + " for nonexistent queue," + " sqid=%"PRIu32", ignoring", qid); return; } cq = n->cq[qid]; - if (new_head >= cq->size) { + if (unlikely(new_head >= cq->size)) { + NVME_GUEST_ERR(nvme_ub_db_wr_invalid_cqhead, + "completion queue doorbell write value" + " beyond queue size, sqid=%"PRIu32"," + " new_head=%"PRIu16", ignoring", + qid, new_head); return; } @@ -XXX,XX +XXX,XX @@ static void nvme_process_db(NvmeCtrl *n, hwaddr addr, int val) nvme_isr_notify(n, cq); } } else { + /* Submission queue doorbell write */ + uint16_t new_tail = val & 0xffff; NvmeSQueue *sq; qid = (addr - 0x1000) >> 3; - if (nvme_check_sqid(n, qid)) { + if (unlikely(nvme_check_sqid(n, qid))) { + NVME_GUEST_ERR(nvme_ub_db_wr_invalid_sq, + "submission queue doorbell write" + " for nonexistent queue," + " sqid=%"PRIu32", ignoring", qid); return; } sq = n->sq[qid]; - if (new_tail >= sq->size) { + if (unlikely(new_tail >= sq->size)) { + NVME_GUEST_ERR(nvme_ub_db_wr_invalid_sqtail, + "submission queue doorbell write value" + " beyond queue size, sqid=%"PRIu32"," + " new_tail=%"PRIu16", ignoring", + qid, new_tail); return; } diff --git a/hw/block/trace-events b/hw/block/trace-events index XXXXXXX..XXXXXXX 100644 --- a/hw/block/trace-events +++ b/hw/block/trace-events @@ -XXX,XX +XXX,XX @@ virtio_blk_submit_multireq(void *vdev, void *mrb, int start, int num_reqs, uint6 hd_geometry_lchs_guess(void *blk, int cyls, int heads, int secs) "blk %p LCHS %d %d %d" hd_geometry_guess(void *blk, uint32_t cyls, uint32_t heads, uint32_t secs, int trans) "blk %p CHS %u %u %u trans %d" +# hw/block/nvme.c +# nvme traces for successful events +nvme_irq_msix(uint32_t vector) "raising MSI-X IRQ vector %u" +nvme_irq_pin(void) "pulsing IRQ pin" +nvme_irq_masked(void) "IRQ is masked" +nvme_dma_read(uint64_t prp1, uint64_t prp2) "DMA read, prp1=0x%"PRIx64" prp2=0x%"PRIx64"" +nvme_rw(char const *verb, uint32_t blk_count, uint64_t byte_count, uint64_t lba) "%s %"PRIu32" blocks (%"PRIu64" bytes) from LBA %"PRIu64"" +nvme_create_sq(uint64_t addr, uint16_t sqid, uint16_t cqid, uint16_t qsize, uint16_t qflags) "create submission queue, addr=0x%"PRIx64", sqid=%"PRIu16", cqid=%"PRIu16", qsize=%"PRIu16", qflags=%"PRIu16"" +nvme_create_cq(uint64_t addr, uint16_t cqid, uint16_t vector, uint16_t size, uint16_t qflags, int ien) "create completion queue, addr=0x%"PRIx64", cqid=%"PRIu16", vector=%"PRIu16", qsize=%"PRIu16", qflags=%"PRIu16", ien=%d" +nvme_del_sq(uint16_t qid) "deleting submission queue sqid=%"PRIu16"" +nvme_del_cq(uint16_t cqid) "deleted completion queue, sqid=%"PRIu16"" +nvme_identify_ctrl(void) "identify controller" +nvme_identify_ns(uint16_t ns) "identify namespace, nsid=%"PRIu16"" +nvme_identify_nslist(uint16_t ns) "identify namespace list, nsid=%"PRIu16"" +nvme_getfeat_vwcache(char const* result) "get feature volatile write cache, result=%s" +nvme_getfeat_numq(int result) "get feature number of queues, result=%d" +nvme_setfeat_numq(int reqcq, int reqsq, int gotcq, int gotsq) "requested cq_count=%d sq_count=%d, responding with cq_count=%d sq_count=%d" +nvme_mmio_intm_set(uint64_t data, uint64_t new_mask) "wrote MMIO, interrupt mask set, data=0x%"PRIx64", new_mask=0x%"PRIx64"" +nvme_mmio_intm_clr(uint64_t data, uint64_t new_mask) "wrote MMIO, interrupt mask clr, data=0x%"PRIx64", new_mask=0x%"PRIx64"" +nvme_mmio_cfg(uint64_t data) "wrote MMIO, config controller config=0x%"PRIx64"" +nvme_mmio_aqattr(uint64_t data) "wrote MMIO, admin queue attributes=0x%"PRIx64"" +nvme_mmio_asqaddr(uint64_t data) "wrote MMIO, admin submission queue address=0x%"PRIx64"" +nvme_mmio_acqaddr(uint64_t data) "wrote MMIO, admin completion queue address=0x%"PRIx64"" +nvme_mmio_asqaddr_hi(uint64_t data, uint64_t new_addr) "wrote MMIO, admin submission queue high half=0x%"PRIx64", new_address=0x%"PRIx64"" +nvme_mmio_acqaddr_hi(uint64_t data, uint64_t new_addr) "wrote MMIO, admin completion queue high half=0x%"PRIx64", new_address=0x%"PRIx64"" +nvme_mmio_start_success(void) "setting controller enable bit succeeded" +nvme_mmio_stopped(void) "cleared controller enable bit" +nvme_mmio_shutdown_set(void) "shutdown bit set" +nvme_mmio_shutdown_cleared(void) "shutdown bit cleared" + +# nvme traces for error conditions +nvme_err_invalid_dma(void) "PRP/SGL is too small for transfer size" +nvme_err_invalid_prplist_ent(uint64_t prplist) "PRP list entry is null or not page aligned: 0x%"PRIx64"" +nvme_err_invalid_prp2_align(uint64_t prp2) "PRP2 is not page aligned: 0x%"PRIx64"" +nvme_err_invalid_prp2_missing(void) "PRP2 is null and more data to be transferred" +nvme_err_invalid_field(void) "invalid field" +nvme_err_invalid_prp(void) "invalid PRP" +nvme_err_invalid_sgl(void) "invalid SGL" +nvme_err_invalid_ns(uint32_t ns, uint32_t limit) "invalid namespace %u not within 1-%u" +nvme_err_invalid_opc(uint8_t opc) "invalid opcode 0x%"PRIx8"" +nvme_err_invalid_admin_opc(uint8_t opc) "invalid admin opcode 0x%"PRIx8"" +nvme_err_invalid_lba_range(uint64_t start, uint64_t len, uint64_t limit) "Invalid LBA start=%"PRIu64" len=%"PRIu64" limit=%"PRIu64"" +nvme_err_invalid_del_sq(uint16_t qid) "invalid submission queue deletion, sid=%"PRIu16"" +nvme_err_invalid_create_sq_cqid(uint16_t cqid) "failed creating submission queue, invalid cqid=%"PRIu16"" +nvme_err_invalid_create_sq_sqid(uint16_t sqid) "failed creating submission queue, invalid sqid=%"PRIu16"" +nvme_err_invalid_create_sq_size(uint16_t qsize) "failed creating submission queue, invalid qsize=%"PRIu16"" +nvme_err_invalid_create_sq_addr(uint64_t addr) "failed creating submission queue, addr=0x%"PRIx64"" +nvme_err_invalid_create_sq_qflags(uint16_t qflags) "failed creating submission queue, qflags=%"PRIu16"" +nvme_err_invalid_del_cq_cqid(uint16_t cqid) "failed deleting completion queue, cqid=%"PRIu16"" +nvme_err_invalid_del_cq_notempty(uint16_t cqid) "failed deleting completion queue, it is not empty, cqid=%"PRIu16"" +nvme_err_invalid_create_cq_cqid(uint16_t cqid) "failed creating completion queue, cqid=%"PRIu16"" +nvme_err_invalid_create_cq_size(uint16_t size) "failed creating completion queue, size=%"PRIu16"" +nvme_err_invalid_create_cq_addr(uint64_t addr) "failed creating completion queue, addr=0x%"PRIx64"" +nvme_err_invalid_create_cq_vector(uint16_t vector) "failed creating completion queue, vector=%"PRIu16"" +nvme_err_invalid_create_cq_qflags(uint16_t qflags) "failed creating completion queue, qflags=%"PRIu16"" +nvme_err_invalid_identify_cns(uint16_t cns) "identify, invalid cns=0x%"PRIx16"" +nvme_err_invalid_getfeat(int dw10) "invalid get features, dw10=0x%"PRIx32"" +nvme_err_invalid_setfeat(uint32_t dw10) "invalid set features, dw10=0x%"PRIx32"" +nvme_err_startfail_cq(void) "nvme_start_ctrl failed because there are non-admin completion queues" +nvme_err_startfail_sq(void) "nvme_start_ctrl failed because there are non-admin submission queues" +nvme_err_startfail_nbarasq(void) "nvme_start_ctrl failed because the admin submission queue address is null" +nvme_err_startfail_nbaracq(void) "nvme_start_ctrl failed because the admin completion queue address is null" +nvme_err_startfail_asq_misaligned(uint64_t addr) "nvme_start_ctrl failed because the admin submission queue address is misaligned: 0x%"PRIx64"" +nvme_err_startfail_acq_misaligned(uint64_t addr) "nvme_start_ctrl failed because the admin completion queue address is misaligned: 0x%"PRIx64"" +nvme_err_startfail_page_too_small(uint8_t log2ps, uint8_t maxlog2ps) "nvme_start_ctrl failed because the page size is too small: log2size=%u, min=%u" +nvme_err_startfail_page_too_large(uint8_t log2ps, uint8_t maxlog2ps) "nvme_start_ctrl failed because the page size is too large: log2size=%u, max=%u" +nvme_err_startfail_cqent_too_small(uint8_t log2ps, uint8_t maxlog2ps) "nvme_start_ctrl failed because the completion queue entry size is too small: log2size=%u, min=%u" +nvme_err_startfail_cqent_too_large(uint8_t log2ps, uint8_t maxlog2ps) "nvme_start_ctrl failed because the completion queue entry size is too large: log2size=%u, max=%u" +nvme_err_startfail_sqent_too_small(uint8_t log2ps, uint8_t maxlog2ps) "nvme_start_ctrl failed because the submission queue entry size is too small: log2size=%u, min=%u" +nvme_err_startfail_sqent_too_large(uint8_t log2ps, uint8_t maxlog2ps) "nvme_start_ctrl failed because the submission queue entry size is too large: log2size=%u, max=%u" +nvme_err_startfail_asqent_sz_zero(void) "nvme_start_ctrl failed because the admin submission queue size is zero" +nvme_err_startfail_acqent_sz_zero(void) "nvme_start_ctrl failed because the admin completion queue size is zero" +nvme_err_startfail(void) "setting controller enable bit failed" + +# Traces for undefined behavior +nvme_ub_mmiowr_misaligned32(uint64_t offset) "MMIO write not 32-bit aligned, offset=0x%"PRIx64"" +nvme_ub_mmiowr_toosmall(uint64_t offset, unsigned size) "MMIO write smaller than 32 bits, offset=0x%"PRIx64", size=%u" +nvme_ub_mmiowr_intmask_with_msix(void) "undefined access to interrupt mask set when MSI-X is enabled" +nvme_ub_mmiowr_ro_csts(void) "attempted to set a read only bit of controller status" +nvme_ub_mmiowr_ssreset_w1c_unsupported(void) "attempted to W1C CSTS.NSSRO but CAP.NSSRS is zero (not supported)" +nvme_ub_mmiowr_ssreset_unsupported(void) "attempted NVM subsystem reset but CAP.NSSRS is zero (not supported)" +nvme_ub_mmiowr_cmbloc_reserved(void) "invalid write to reserved CMBLOC when CMBSZ is zero, ignored" +nvme_ub_mmiowr_cmbsz_readonly(void) "invalid write to read only CMBSZ, ignored" +nvme_ub_mmiowr_invalid(uint64_t offset, uint64_t data) "invalid MMIO write, offset=0x%"PRIx64", data=0x%"PRIx64"" +nvme_ub_mmiord_misaligned32(uint64_t offset) "MMIO read not 32-bit aligned, offset=0x%"PRIx64"" +nvme_ub_mmiord_toosmall(uint64_t offset) "MMIO read smaller than 32-bits, offset=0x%"PRIx64"" +nvme_ub_mmiord_invalid_ofs(uint64_t offset) "MMIO read beyond last register, offset=0x%"PRIx64", returning 0" +nvme_ub_db_wr_misaligned(uint64_t offset) "doorbell write not 32-bit aligned, offset=0x%"PRIx64", ignoring" +nvme_ub_db_wr_invalid_cq(uint32_t qid) "completion queue doorbell write for nonexistent queue, cqid=%"PRIu32", ignoring" +nvme_ub_db_wr_invalid_cqhead(uint32_t qid, uint16_t new_head) "completion queue doorbell write value beyond queue size, cqid=%"PRIu32", new_head=%"PRIu16", ignoring" +nvme_ub_db_wr_invalid_sq(uint32_t qid) "submission queue doorbell write for nonexistent queue, sqid=%"PRIu32", ignoring" +nvme_ub_db_wr_invalid_sqtail(uint32_t qid, uint16_t new_tail) "submission queue doorbell write value beyond queue size, sqid=%"PRIu32", new_head=%"PRIu16", ignoring" + # hw/block/xen_disk.c xen_disk_alloc(char *name) "%s" xen_disk_init(char *name) "%s" -- 2.13.6 From: Fam Zheng <famz@redhat.com> Management tools create overlays of running guests with qemu-img: $ qemu-img create -b /image/in/use.qcow2 -f qcow2 /overlay/image.qcow2 but this doesn't work anymore due to image locking: qemu-img: /overlay/image.qcow2: Failed to get shared "write" lock Is another process using the image? Could not open backing image to determine size. Use the force share option to allow this use case again. Cc: qemu-stable@nongnu.org Signed-off-by: Fam Zheng <famz@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ void bdrv_img_create(const char *filename, const char *fmt, back_flags = flags; back_flags &= ~(BDRV_O_RDWR | BDRV_O_SNAPSHOT | BDRV_O_NO_BACKING); + backing_options = qdict_new(); if (backing_fmt) { - backing_options = qdict_new(); qdict_put_str(backing_options, "driver", backing_fmt); } + qdict_put_bool(backing_options, BDRV_OPT_FORCE_SHARE, true); bs = bdrv_open(full_backing, NULL, backing_options, back_flags, &local_err); -- 2.13.6 From: Thomas Huth <thuth@redhat.com> It's not working anymore since QEMU v1.3.0 - time to remove it now. Signed-off-by: Thomas Huth <thuth@redhat.com> Reviewed-by: John Snow <jsnow@redhat.com> Reviewed-by: Markus Armbruster <armbru@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- blockdev.c | 11 ----------- qemu-doc.texi | 6 ------ 2 files changed, 17 deletions(-) diff --git a/blockdev.c b/blockdev.c index XXXXXXX..XXXXXXX 100644 --- a/blockdev.c +++ b/blockdev.c @@ -XXX,XX +XXX,XX @@ QemuOptsList qemu_legacy_drive_opts = { .type = QEMU_OPT_STRING, .help = "chs translation (auto, lba, none)", },{ - .name = "boot", - .type = QEMU_OPT_BOOL, - .help = "(deprecated, ignored)", - },{ .name = "addr", .type = QEMU_OPT_STRING, .help = "pci address (virtio only)", @@ -XXX,XX +XXX,XX @@ DriveInfo *drive_new(QemuOpts *all_opts, BlockInterfaceType block_default_type) goto fail; } - /* Deprecated option boot=[on|off] */ - if (qemu_opt_get(legacy_opts, "boot") != NULL) { - fprintf(stderr, "qemu-kvm: boot=on|off is deprecated and will be " - "ignored. Future versions will reject this parameter. Please " - "update your scripts.\n"); - } - /* Other deprecated options */ if (!qtest_enabled()) { for (i = 0; i < ARRAY_SIZE(deprecated); i++) { diff --git a/qemu-doc.texi b/qemu-doc.texi index XXXXXXX..XXXXXXX 100644 --- a/qemu-doc.texi +++ b/qemu-doc.texi @@ -XXX,XX +XXX,XX @@ deprecated. @section System emulator command line arguments -@subsection -drive boot=on|off (since 1.3.0) - -The ``boot=on|off'' option to the ``-drive'' argument is -ignored. Applications should use the ``bootindex=N'' parameter -to set an absolute ordering between devices instead. - @subsection -tdf (since 1.3.0) The ``-tdf'' argument is ignored. The behaviour implemented -- 2.13.6 From: Thomas Huth <thuth@redhat.com> It's been marked as deprecated since QEMU v2.10.0, and so far nobody complained that we should keep it, so let's remove this legacy option now to simplify the code quite a bit. Signed-off-by: Thomas Huth <thuth@redhat.com> Reviewed-by: John Snow <jsnow@redhat.com> Reviewed-by: Markus Armbruster <armbru@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- vl.c | 86 ++------------------------------------------------------- qemu-doc.texi | 8 ------ qemu-options.hx | 19 ++----------- 3 files changed, 4 insertions(+), 109 deletions(-) diff --git a/vl.c b/vl.c index XXXXXXX..XXXXXXX 100644 --- a/vl.c +++ b/vl.c @@ -XXX,XX +XXX,XX @@ int main(int argc, char **argv, char **envp) const char *boot_order = NULL; const char *boot_once = NULL; DisplayState *ds; - int cyls, heads, secs, translation; QemuOpts *opts, *machine_opts; - QemuOpts *hda_opts = NULL, *icount_opts = NULL, *accel_opts = NULL; + QemuOpts *icount_opts = NULL, *accel_opts = NULL; QemuOptsList *olist; int optind; const char *optarg; @@ -XXX,XX +XXX,XX @@ int main(int argc, char **argv, char **envp) cpu_model = NULL; snapshot = 0; - cyls = heads = secs = 0; - translation = BIOS_ATA_TRANSLATION_AUTO; nb_nics = 0; @@ -XXX,XX +XXX,XX @@ int main(int argc, char **argv, char **envp) if (optind >= argc) break; if (argv[optind][0] != '-') { - hda_opts = drive_add(IF_DEFAULT, 0, argv[optind++], HD_OPTS); + drive_add(IF_DEFAULT, 0, argv[optind++], HD_OPTS); } else { const QEMUOption *popt; @@ -XXX,XX +XXX,XX @@ int main(int argc, char **argv, char **envp) cpu_model = optarg; break; case QEMU_OPTION_hda: - { - char buf[256]; - if (cyls == 0) - snprintf(buf, sizeof(buf), "%s", HD_OPTS); - else - snprintf(buf, sizeof(buf), - "%s,cyls=%d,heads=%d,secs=%d%s", - HD_OPTS , cyls, heads, secs, - translation == BIOS_ATA_TRANSLATION_LBA ? - ",trans=lba" : - translation == BIOS_ATA_TRANSLATION_NONE ? - ",trans=none" : ""); - drive_add(IF_DEFAULT, 0, optarg, buf); - break; - } case QEMU_OPTION_hdb: case QEMU_OPTION_hdc: case QEMU_OPTION_hdd: @@ -XXX,XX +XXX,XX @@ int main(int argc, char **argv, char **envp) case QEMU_OPTION_snapshot: snapshot = 1; break; - case QEMU_OPTION_hdachs: - { - const char *p; - p = optarg; - cyls = strtol(p, (char **)&p, 0); - if (cyls < 1 || cyls > 16383) - goto chs_fail; - if (*p != ',') - goto chs_fail; - p++; - heads = strtol(p, (char **)&p, 0); - if (heads < 1 || heads > 16) - goto chs_fail; - if (*p != ',') - goto chs_fail; - p++; - secs = strtol(p, (char **)&p, 0); - if (secs < 1 || secs > 63) - goto chs_fail; - if (*p == ',') { - p++; - if (!strcmp(p, "large")) { - translation = BIOS_ATA_TRANSLATION_LARGE; - } else if (!strcmp(p, "rechs")) { - translation = BIOS_ATA_TRANSLATION_RECHS; - } else if (!strcmp(p, "none")) { - translation = BIOS_ATA_TRANSLATION_NONE; - } else if (!strcmp(p, "lba")) { - translation = BIOS_ATA_TRANSLATION_LBA; - } else if (!strcmp(p, "auto")) { - translation = BIOS_ATA_TRANSLATION_AUTO; - } else { - goto chs_fail; - } - } else if (*p != '\0') { - chs_fail: - error_report("invalid physical CHS format"); - exit(1); - } - if (hda_opts != NULL) { - qemu_opt_set_number(hda_opts, "cyls", cyls, - &error_abort); - qemu_opt_set_number(hda_opts, "heads", heads, - &error_abort); - qemu_opt_set_number(hda_opts, "secs", secs, - &error_abort); - if (translation == BIOS_ATA_TRANSLATION_LARGE) { - qemu_opt_set(hda_opts, "trans", "large", - &error_abort); - } else if (translation == BIOS_ATA_TRANSLATION_RECHS) { - qemu_opt_set(hda_opts, "trans", "rechs", - &error_abort); - } else if (translation == BIOS_ATA_TRANSLATION_LBA) { - qemu_opt_set(hda_opts, "trans", "lba", - &error_abort); - } else if (translation == BIOS_ATA_TRANSLATION_NONE) { - qemu_opt_set(hda_opts, "trans", "none", - &error_abort); - } - } - } - error_report("'-hdachs' is deprecated, please use '-device" - " ide-hd,cyls=c,heads=h,secs=s,...' instead"); - break; case QEMU_OPTION_numa: opts = qemu_opts_parse_noisily(qemu_find_opts("numa"), optarg, true); diff --git a/qemu-doc.texi b/qemu-doc.texi index XXXXXXX..XXXXXXX 100644 --- a/qemu-doc.texi +++ b/qemu-doc.texi @@ -XXX,XX +XXX,XX @@ The ``--net dump'' argument is now replaced with the ``-object filter-dump'' argument which works in combination with the modern ``-netdev`` backends instead. -@subsection -hdachs (since 2.10.0) - -The ``-hdachs'' argument is now a synonym for setting -the ``cyls'', ``heads'', ``secs'', and ``trans'' properties -on the ``ide-hd'' device using the ``-device'' argument. -The new syntax allows different settings to be provided -per disk. - @subsection -usbdevice (since 2.10.0) The ``-usbdevice DEV'' argument is now a synonym for setting diff --git a/qemu-options.hx b/qemu-options.hx index XXXXXXX..XXXXXXX 100644 --- a/qemu-options.hx +++ b/qemu-options.hx @@ -XXX,XX +XXX,XX @@ of available connectors of a given interface type. @item media=@var{media} This option defines the type of the media: disk or cdrom. @item cyls=@var{c},heads=@var{h},secs=@var{s}[,trans=@var{t}] -These options have the same definition as they have in @option{-hdachs}. -These parameters are deprecated, use the corresponding parameters +Force disk physical geometry and the optional BIOS translation (trans=none or +lba). These parameters are deprecated, use the corresponding parameters of @code{-device} instead. @item snapshot=@var{snapshot} @var{snapshot} is "on" or "off" and controls snapshot mode for the given drive @@ -XXX,XX +XXX,XX @@ the raw disk image you use is not written back. You can however force the write back by pressing @key{C-a s} (@pxref{disk_images}). ETEXI -DEF("hdachs", HAS_ARG, QEMU_OPTION_hdachs, \ - "-hdachs c,h,s[,t]\n" \ - " force hard disk 0 physical geometry and the optional BIOS\n" \ - " translation (t=none or lba) (usually QEMU can guess them)\n", - QEMU_ARCH_ALL) -STEXI -@item -hdachs @var{c},@var{h},@var{s},[,@var{t}] -@findex -hdachs -Force hard disk 0 physical geometry (1 <= @var{c} <= 16383, 1 <= -@var{h} <= 16, 1 <= @var{s} <= 63) and optionally force the BIOS -translation mode (@var{t}=none, lba or auto). Usually QEMU can guess -all those parameters. This option is deprecated, please use -@code{-device ide-hd,cyls=c,heads=h,secs=s,...} instead. -ETEXI - DEF("fsdev", HAS_ARG, QEMU_OPTION_fsdev, "-fsdev fsdriver,id=id[,path=path,][security_model={mapped-xattr|mapped-file|passthrough|none}]\n" " [,writeout=immediate][,readonly][,socket=socket|sock_fd=sock_fd][,fmode=fmode][,dmode=dmode]\n" -- 2.13.6 From: Thomas Huth <thuth@redhat.com> Looks like we forgot to announce the deprecation of these options in the corresponding chapter of the qemu-doc text, so let's do that now. Signed-off-by: Thomas Huth <thuth@redhat.com> Reviewed-by: John Snow <jsnow@redhat.com> Reviewed-by: Markus Armbruster <armbru@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- qemu-doc.texi | 15 +++++++++++++++ 1 file changed, 15 insertions(+) diff --git a/qemu-doc.texi b/qemu-doc.texi index XXXXXXX..XXXXXXX 100644 --- a/qemu-doc.texi +++ b/qemu-doc.texi @@ -XXX,XX +XXX,XX @@ longer be directly supported in QEMU. The ``-drive if=scsi'' argument is replaced by the the ``-device BUS-TYPE'' argument combined with ``-drive if=none''. +@subsection -drive cyls=...,heads=...,secs=...,trans=... (since 2.10.0) + +The drive geometry arguments are replaced by the the geometry arguments +that can be specified with the ``-device'' parameter. + +@subsection -drive serial=... (since 2.10.0) + +The drive serial argument is replaced by the the serial argument +that can be specified with the ``-device'' parameter. + +@subsection -drive addr=... (since 2.10.0) + +The drive addr argument is replaced by the the addr argument +that can be specified with the ``-device'' parameter. + @subsection -net dump (since 2.10.0) The ``--net dump'' argument is now replaced with the -- 2.13.6 From: Fam Zheng <famz@redhat.com> Signed-off-by: Fam Zheng <famz@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block_int.h | 1 - block/io.c | 18 ------------------ 2 files changed, 19 deletions(-) diff --git a/include/block/block_int.h b/include/block/block_int.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block_int.h +++ b/include/block/block_int.h @@ -XXX,XX +XXX,XX @@ bool blk_dev_is_tray_open(BlockBackend *blk); bool blk_dev_is_medium_locked(BlockBackend *blk); void bdrv_set_dirty(BlockDriverState *bs, int64_t offset, int64_t bytes); -bool bdrv_requests_pending(BlockDriverState *bs); void bdrv_clear_dirty_bitmap(BdrvDirtyBitmap *bitmap, HBitmap **out); void bdrv_undo_clear_dirty_bitmap(BdrvDirtyBitmap *bitmap, HBitmap *in); 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_disable_copy_on_read(BlockDriverState *bs) assert(old >= 1); } -/* Check if any requests are in-flight (including throttled requests) */ -bool bdrv_requests_pending(BlockDriverState *bs) -{ - BdrvChild *child; - - if (atomic_read(&bs->in_flight)) { - return true; - } - - QLIST_FOREACH(child, &bs->children, next) { - if (bdrv_requests_pending(child->bs)) { - return true; - } - } - - return false; -} - typedef struct { Coroutine *co; BlockDriverState *bs; -- 2.13.6 Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Fam Zheng <famz@redhat.com> --- block/io.c | 6 ++++++ 1 file changed, 6 insertions(+) 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_drain_all_begin(void) BdrvNextIterator it; GSList *aio_ctxs = NULL, *ctx; + /* BDRV_POLL_WHILE() for a node can only be called from its own I/O thread + * or the main loop AioContext. We potentially use BDRV_POLL_WHILE() on + * nodes in several different AioContexts, so make sure we're in the main + * context. */ + assert(qemu_get_current_aio_context() == qemu_get_aio_context()); + block_job_pause_all(); for (bs = bdrv_first(&it); bs; bs = bdrv_next(&it)) { -- 2.13.6 bdrv_drained_begin() doesn't increase bs->quiesce_counter recursively and also doesn't notify other parent nodes of children, which both means that the child nodes are not actually drained, and bdrv_drained_begin() is providing useful functionality only on a single node. To keep things consistent, we also shouldn't call the block driver callbacks recursively. A proper recursive drain version that provides an actually working drained section for child nodes will be introduced later. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Fam Zheng <famz@redhat.com> --- block/io.c | 16 +++++++++------- 1 file changed, 9 insertions(+), 7 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 @@ static void coroutine_fn bdrv_drain_invoke_entry(void *opaque) } /* Recursively call BlockDriver.bdrv_co_drain_begin/end callbacks */ -static void bdrv_drain_invoke(BlockDriverState *bs, bool begin) +static void bdrv_drain_invoke(BlockDriverState *bs, bool begin, bool recursive) { BdrvChild *child, *tmp; BdrvCoDrainData data = { .bs = bs, .done = false, .begin = begin}; @@ -XXX,XX +XXX,XX @@ static void bdrv_drain_invoke(BlockDriverState *bs, bool begin) bdrv_coroutine_enter(bs, data.co); BDRV_POLL_WHILE(bs, !data.done); - QLIST_FOREACH_SAFE(child, &bs->children, next, tmp) { - bdrv_drain_invoke(child->bs, begin); + if (recursive) { + QLIST_FOREACH_SAFE(child, &bs->children, next, tmp) { + bdrv_drain_invoke(child->bs, begin, true); + } } } @@ -XXX,XX +XXX,XX @@ void bdrv_drained_begin(BlockDriverState *bs) bdrv_parent_drained_begin(bs); } - bdrv_drain_invoke(bs, true); + bdrv_drain_invoke(bs, true, false); bdrv_drain_recurse(bs); } @@ -XXX,XX +XXX,XX @@ void bdrv_drained_end(BlockDriverState *bs) } /* Re-enable things in child-to-parent order */ - bdrv_drain_invoke(bs, false); + bdrv_drain_invoke(bs, false, false); bdrv_parent_drained_end(bs); aio_enable_external(bdrv_get_aio_context(bs)); } @@ -XXX,XX +XXX,XX @@ void bdrv_drain_all_begin(void) aio_context_acquire(aio_context); aio_disable_external(aio_context); bdrv_parent_drained_begin(bs); - bdrv_drain_invoke(bs, true); + bdrv_drain_invoke(bs, true, true); aio_context_release(aio_context); if (!g_slist_find(aio_ctxs, aio_context)) { @@ -XXX,XX +XXX,XX @@ void bdrv_drain_all_end(void) /* Re-enable things in child-to-parent order */ aio_context_acquire(aio_context); - bdrv_drain_invoke(bs, false); + bdrv_drain_invoke(bs, false, true); bdrv_parent_drained_end(bs); aio_enable_external(aio_context); aio_context_release(aio_context); -- 2.13.6 The existing test is for bdrv_drain_all_begin/end() only. Generalise the test case so that it can be run for the other variants as well. At the moment this is only bdrv_drain_begin/end(), but in a while, we'll add another one. Also, add a backing file to the test node to test whether the operations work recursively. Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- tests/test-bdrv-drain.c | 69 ++++++++++++++++++++++++++++++++++++++++++++----- 1 file changed, 62 insertions(+), 7 deletions(-) diff --git a/tests/test-bdrv-drain.c b/tests/test-bdrv-drain.c index XXXXXXX..XXXXXXX 100644 --- a/tests/test-bdrv-drain.c +++ b/tests/test-bdrv-drain.c @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_test = { .bdrv_co_drain_begin = bdrv_test_co_drain_begin, .bdrv_co_drain_end = bdrv_test_co_drain_end, + + .bdrv_child_perm = bdrv_format_default_perms, }; static void aio_ret_cb(void *opaque, int ret) @@ -XXX,XX +XXX,XX @@ static void aio_ret_cb(void *opaque, int ret) *aio_ret = ret; } -static void test_drv_cb_drain_all(void) +enum drain_type { + BDRV_DRAIN_ALL, + BDRV_DRAIN, +}; + +static void do_drain_begin(enum drain_type drain_type, BlockDriverState *bs) +{ + switch (drain_type) { + case BDRV_DRAIN_ALL: bdrv_drain_all_begin(); break; + case BDRV_DRAIN: bdrv_drained_begin(bs); break; + default: g_assert_not_reached(); + } +} + +static void do_drain_end(enum drain_type drain_type, BlockDriverState *bs) +{ + switch (drain_type) { + case BDRV_DRAIN_ALL: bdrv_drain_all_end(); break; + case BDRV_DRAIN: bdrv_drained_end(bs); break; + default: g_assert_not_reached(); + } +} + +static void test_drv_cb_common(enum drain_type drain_type, bool recursive) { BlockBackend *blk; - BlockDriverState *bs; - BDRVTestState *s; + BlockDriverState *bs, *backing; + BDRVTestState *s, *backing_s; BlockAIOCB *acb; int aio_ret; @@ -XXX,XX +XXX,XX @@ static void test_drv_cb_drain_all(void) s = bs->opaque; blk_insert_bs(blk, bs, &error_abort); + backing = bdrv_new_open_driver(&bdrv_test, "backing", 0, &error_abort); + backing_s = backing->opaque; + bdrv_set_backing_hd(bs, backing, &error_abort); + /* Simple bdrv_drain_all_begin/end pair, check that CBs are called */ g_assert_cmpint(s->drain_count, ==, 0); - bdrv_drain_all_begin(); + g_assert_cmpint(backing_s->drain_count, ==, 0); + + do_drain_begin(drain_type, bs); + g_assert_cmpint(s->drain_count, ==, 1); - bdrv_drain_all_end(); + g_assert_cmpint(backing_s->drain_count, ==, !!recursive); + + do_drain_end(drain_type, bs); + g_assert_cmpint(s->drain_count, ==, 0); + g_assert_cmpint(backing_s->drain_count, ==, 0); /* Now do the same while a request is pending */ aio_ret = -EINPROGRESS; @@ -XXX,XX +XXX,XX @@ static void test_drv_cb_drain_all(void) g_assert_cmpint(aio_ret, ==, -EINPROGRESS); g_assert_cmpint(s->drain_count, ==, 0); - bdrv_drain_all_begin(); + g_assert_cmpint(backing_s->drain_count, ==, 0); + + do_drain_begin(drain_type, bs); + g_assert_cmpint(aio_ret, ==, 0); g_assert_cmpint(s->drain_count, ==, 1); - bdrv_drain_all_end(); + g_assert_cmpint(backing_s->drain_count, ==, !!recursive); + + do_drain_end(drain_type, bs); + g_assert_cmpint(s->drain_count, ==, 0); + g_assert_cmpint(backing_s->drain_count, ==, 0); + bdrv_unref(backing); bdrv_unref(bs); blk_unref(blk); } +static void test_drv_cb_drain_all(void) +{ + test_drv_cb_common(BDRV_DRAIN_ALL, true); +} + +static void test_drv_cb_drain(void) +{ + test_drv_cb_common(BDRV_DRAIN, false); +} + int main(int argc, char **argv) { bdrv_init(); @@ -XXX,XX +XXX,XX @@ int main(int argc, char **argv) g_test_init(&argc, &argv, NULL); g_test_add_func("/bdrv-drain/driver-cb/drain_all", test_drv_cb_drain_all); + g_test_add_func("/bdrv-drain/driver-cb/drain", test_drv_cb_drain); return g_test_run(); } -- 2.13.6 This is currently only working correctly for bdrv_drain(), not for bdrv_drain_all(). Leave a comment for the drain_all case, we'll address it later. Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- tests/test-bdrv-drain.c | 45 +++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 45 insertions(+) diff --git a/tests/test-bdrv-drain.c b/tests/test-bdrv-drain.c index XXXXXXX..XXXXXXX 100644 --- a/tests/test-bdrv-drain.c +++ b/tests/test-bdrv-drain.c @@ -XXX,XX +XXX,XX @@ static void test_drv_cb_drain(void) test_drv_cb_common(BDRV_DRAIN, false); } +static void test_quiesce_common(enum drain_type drain_type, bool recursive) +{ + BlockBackend *blk; + BlockDriverState *bs, *backing; + + blk = blk_new(BLK_PERM_ALL, BLK_PERM_ALL); + bs = bdrv_new_open_driver(&bdrv_test, "test-node", BDRV_O_RDWR, + &error_abort); + blk_insert_bs(blk, bs, &error_abort); + + backing = bdrv_new_open_driver(&bdrv_test, "backing", 0, &error_abort); + bdrv_set_backing_hd(bs, backing, &error_abort); + + g_assert_cmpint(bs->quiesce_counter, ==, 0); + g_assert_cmpint(backing->quiesce_counter, ==, 0); + + do_drain_begin(drain_type, bs); + + g_assert_cmpint(bs->quiesce_counter, ==, 1); + g_assert_cmpint(backing->quiesce_counter, ==, !!recursive); + + do_drain_end(drain_type, bs); + + g_assert_cmpint(bs->quiesce_counter, ==, 0); + g_assert_cmpint(backing->quiesce_counter, ==, 0); + + bdrv_unref(backing); + bdrv_unref(bs); + blk_unref(blk); +} + +static void test_quiesce_drain_all(void) +{ + // XXX drain_all doesn't quiesce + //test_quiesce_common(BDRV_DRAIN_ALL, true); +} + +static void test_quiesce_drain(void) +{ + test_quiesce_common(BDRV_DRAIN, false); +} + int main(int argc, char **argv) { bdrv_init(); @@ -XXX,XX +XXX,XX @@ int main(int argc, char **argv) g_test_add_func("/bdrv-drain/driver-cb/drain_all", test_drv_cb_drain_all); g_test_add_func("/bdrv-drain/driver-cb/drain", test_drv_cb_drain); + g_test_add_func("/bdrv-drain/quiesce/drain_all", test_quiesce_drain_all); + g_test_add_func("/bdrv-drain/quiesce/drain", test_quiesce_drain); + return g_test_run(); } -- 2.13.6 Block jobs already paused themselves when their main BlockBackend entered a drained section. This is not good enough: We also want to pause a block job and may not submit new requests if, for example, the mirror target node should be drained. This implements .drained_begin/end callbacks in child_job in order to consider all block nodes related to the job, and removes the BlockBackend callbacks which are unnecessary now because the root of the job main BlockBackend is always referenced with a child_job, too. Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- blockjob.c | 22 +++++++++------------- 1 file changed, 9 insertions(+), 13 deletions(-) diff --git a/blockjob.c b/blockjob.c index XXXXXXX..XXXXXXX 100644 --- a/blockjob.c +++ b/blockjob.c @@ -XXX,XX +XXX,XX @@ static char *child_job_get_parent_desc(BdrvChild *c) job->id); } -static const BdrvChildRole child_job = { - .get_parent_desc = child_job_get_parent_desc, - .stay_at_node = true, -}; - -static void block_job_drained_begin(void *opaque) +static void child_job_drained_begin(BdrvChild *c) { - BlockJob *job = opaque; + BlockJob *job = c->opaque; block_job_pause(job); } -static void block_job_drained_end(void *opaque) +static void child_job_drained_end(BdrvChild *c) { - BlockJob *job = opaque; + BlockJob *job = c->opaque; block_job_resume(job); } -static const BlockDevOps block_job_dev_ops = { - .drained_begin = block_job_drained_begin, - .drained_end = block_job_drained_end, +static const BdrvChildRole child_job = { + .get_parent_desc = child_job_get_parent_desc, + .drained_begin = child_job_drained_begin, + .drained_end = child_job_drained_end, + .stay_at_node = true, }; void block_job_remove_all_bdrv(BlockJob *job) @@ -XXX,XX +XXX,XX @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver, block_job_add_bdrv(job, "main node", bs, 0, BLK_PERM_ALL, &error_abort); bs->job = job; - blk_set_dev_ops(blk, &block_job_dev_ops, job); bdrv_op_unblock(bs, BLOCK_OP_TYPE_DATAPLANE, job->blocker); QLIST_INSERT_HEAD(&block_jobs, job, job_list); -- 2.13.6 Block jobs must be paused if any of the involved nodes are drained. Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- tests/test-bdrv-drain.c | 121 ++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 121 insertions(+) diff --git a/tests/test-bdrv-drain.c b/tests/test-bdrv-drain.c index XXXXXXX..XXXXXXX 100644 --- a/tests/test-bdrv-drain.c +++ b/tests/test-bdrv-drain.c @@ -XXX,XX +XXX,XX @@ #include "qemu/osdep.h" #include "block/block.h" +#include "block/blockjob_int.h" #include "sysemu/block-backend.h" #include "qapi/error.h" @@ -XXX,XX +XXX,XX @@ static void test_quiesce_drain(void) test_quiesce_common(BDRV_DRAIN, false); } + +typedef struct TestBlockJob { + BlockJob common; + bool should_complete; +} TestBlockJob; + +static void test_job_completed(BlockJob *job, void *opaque) +{ + block_job_completed(job, 0); +} + +static void coroutine_fn test_job_start(void *opaque) +{ + TestBlockJob *s = opaque; + + while (!s->should_complete) { + block_job_sleep_ns(&s->common, 100000); + } + + block_job_defer_to_main_loop(&s->common, test_job_completed, NULL); +} + +static void test_job_complete(BlockJob *job, Error **errp) +{ + TestBlockJob *s = container_of(job, TestBlockJob, common); + s->should_complete = true; +} + +BlockJobDriver test_job_driver = { + .instance_size = sizeof(TestBlockJob), + .start = test_job_start, + .complete = test_job_complete, +}; + +static void test_blockjob_common(enum drain_type drain_type) +{ + BlockBackend *blk_src, *blk_target; + BlockDriverState *src, *target; + BlockJob *job; + int ret; + + src = bdrv_new_open_driver(&bdrv_test, "source", BDRV_O_RDWR, + &error_abort); + blk_src = blk_new(BLK_PERM_ALL, BLK_PERM_ALL); + blk_insert_bs(blk_src, src, &error_abort); + + target = bdrv_new_open_driver(&bdrv_test, "target", BDRV_O_RDWR, + &error_abort); + blk_target = blk_new(BLK_PERM_ALL, BLK_PERM_ALL); + blk_insert_bs(blk_target, target, &error_abort); + + job = block_job_create("job0", &test_job_driver, src, 0, BLK_PERM_ALL, 0, + 0, NULL, NULL, &error_abort); + block_job_add_bdrv(job, "target", target, 0, BLK_PERM_ALL, &error_abort); + block_job_start(job); + + g_assert_cmpint(job->pause_count, ==, 0); + g_assert_false(job->paused); + g_assert_false(job->busy); /* We're in block_job_sleep_ns() */ + + do_drain_begin(drain_type, src); + + if (drain_type == BDRV_DRAIN_ALL) { + /* bdrv_drain_all() drains both src and target, and involves an + * additional block_job_pause_all() */ + g_assert_cmpint(job->pause_count, ==, 3); + } else { + g_assert_cmpint(job->pause_count, ==, 1); + } + /* XXX We don't wait until the job is actually paused. Is this okay? */ + /* g_assert_true(job->paused); */ + g_assert_false(job->busy); /* The job is paused */ + + do_drain_end(drain_type, src); + + g_assert_cmpint(job->pause_count, ==, 0); + g_assert_false(job->paused); + g_assert_false(job->busy); /* We're in block_job_sleep_ns() */ + + do_drain_begin(drain_type, target); + + if (drain_type == BDRV_DRAIN_ALL) { + /* bdrv_drain_all() drains both src and target, and involves an + * additional block_job_pause_all() */ + g_assert_cmpint(job->pause_count, ==, 3); + } else { + g_assert_cmpint(job->pause_count, ==, 1); + } + /* XXX We don't wait until the job is actually paused. Is this okay? */ + /* g_assert_true(job->paused); */ + g_assert_false(job->busy); /* The job is paused */ + + do_drain_end(drain_type, target); + + g_assert_cmpint(job->pause_count, ==, 0); + g_assert_false(job->paused); + g_assert_false(job->busy); /* We're in block_job_sleep_ns() */ + + ret = block_job_complete_sync(job, &error_abort); + g_assert_cmpint(ret, ==, 0); + + blk_unref(blk_src); + blk_unref(blk_target); + bdrv_unref(src); + bdrv_unref(target); +} + +static void test_blockjob_drain_all(void) +{ + test_blockjob_common(BDRV_DRAIN_ALL); +} + +static void test_blockjob_drain(void) +{ + test_blockjob_common(BDRV_DRAIN); +} + int main(int argc, char **argv) { bdrv_init(); @@ -XXX,XX +XXX,XX @@ int main(int argc, char **argv) g_test_add_func("/bdrv-drain/quiesce/drain_all", test_quiesce_drain_all); g_test_add_func("/bdrv-drain/quiesce/drain", test_quiesce_drain); + g_test_add_func("/bdrv-drain/blockjob/drain_all", test_blockjob_drain_all); + g_test_add_func("/bdrv-drain/blockjob/drain", test_blockjob_drain); + return g_test_run(); } -- 2.13.6 Block jobs are already paused using the BdrvChildRole drain callbacks, so we don't need an additional block_job_pause_all() call. Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/io.c | 4 ---- tests/test-bdrv-drain.c | 10 ++++------ 2 files changed, 4 insertions(+), 10 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 @@ void bdrv_drain_all_begin(void) * context. */ assert(qemu_get_current_aio_context() == qemu_get_aio_context()); - block_job_pause_all(); - for (bs = bdrv_first(&it); bs; bs = bdrv_next(&it)) { AioContext *aio_context = bdrv_get_aio_context(bs); @@ -XXX,XX +XXX,XX @@ void bdrv_drain_all_end(void) aio_enable_external(aio_context); aio_context_release(aio_context); } - - block_job_resume_all(); } void bdrv_drain_all(void) diff --git a/tests/test-bdrv-drain.c b/tests/test-bdrv-drain.c index XXXXXXX..XXXXXXX 100644 --- a/tests/test-bdrv-drain.c +++ b/tests/test-bdrv-drain.c @@ -XXX,XX +XXX,XX @@ static void test_blockjob_common(enum drain_type drain_type) do_drain_begin(drain_type, src); if (drain_type == BDRV_DRAIN_ALL) { - /* bdrv_drain_all() drains both src and target, and involves an - * additional block_job_pause_all() */ - g_assert_cmpint(job->pause_count, ==, 3); + /* bdrv_drain_all() drains both src and target */ + g_assert_cmpint(job->pause_count, ==, 2); } else { g_assert_cmpint(job->pause_count, ==, 1); } @@ -XXX,XX +XXX,XX @@ static void test_blockjob_common(enum drain_type drain_type) do_drain_begin(drain_type, target); if (drain_type == BDRV_DRAIN_ALL) { - /* bdrv_drain_all() drains both src and target, and involves an - * additional block_job_pause_all() */ - g_assert_cmpint(job->pause_count, ==, 3); + /* bdrv_drain_all() drains both src and target */ + g_assert_cmpint(job->pause_count, ==, 2); } else { g_assert_cmpint(job->pause_count, ==, 1); } -- 2.13.6 bdrv_do_drained_begin() restricts the call of parent callbacks and aio_disable_external() to the outermost drain section, but the block driver callbacks are always called. bdrv_do_drained_end() must match this behaviour, otherwise nodes stay drained even if begin/end calls were balanced. Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/io.c | 12 +++++++----- 1 file changed, 7 insertions(+), 5 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 @@ void bdrv_drained_begin(BlockDriverState *bs) void bdrv_drained_end(BlockDriverState *bs) { + int old_quiesce_counter; + if (qemu_in_coroutine()) { bdrv_co_yield_to_drain(bs, false); return; } assert(bs->quiesce_counter > 0); - if (atomic_fetch_dec(&bs->quiesce_counter) > 1) { - return; - } + old_quiesce_counter = atomic_fetch_dec(&bs->quiesce_counter); /* Re-enable things in child-to-parent order */ bdrv_drain_invoke(bs, false, false); - bdrv_parent_drained_end(bs); - aio_enable_external(bdrv_get_aio_context(bs)); + if (old_quiesce_counter == 1) { + bdrv_parent_drained_end(bs); + aio_enable_external(bdrv_get_aio_context(bs)); + } } /* -- 2.13.6 Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- tests/test-bdrv-drain.c | 57 +++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 57 insertions(+) diff --git a/tests/test-bdrv-drain.c b/tests/test-bdrv-drain.c index XXXXXXX..XXXXXXX 100644 --- a/tests/test-bdrv-drain.c +++ b/tests/test-bdrv-drain.c @@ -XXX,XX +XXX,XX @@ static void aio_ret_cb(void *opaque, int ret) enum drain_type { BDRV_DRAIN_ALL, BDRV_DRAIN, + DRAIN_TYPE_MAX, }; static void do_drain_begin(enum drain_type drain_type, BlockDriverState *bs) @@ -XXX,XX +XXX,XX @@ static void test_quiesce_drain(void) test_quiesce_common(BDRV_DRAIN, false); } +static void test_nested(void) +{ + BlockBackend *blk; + BlockDriverState *bs, *backing; + BDRVTestState *s, *backing_s; + enum drain_type outer, inner; + + blk = blk_new(BLK_PERM_ALL, BLK_PERM_ALL); + bs = bdrv_new_open_driver(&bdrv_test, "test-node", BDRV_O_RDWR, + &error_abort); + s = bs->opaque; + blk_insert_bs(blk, bs, &error_abort); + + backing = bdrv_new_open_driver(&bdrv_test, "backing", 0, &error_abort); + backing_s = backing->opaque; + bdrv_set_backing_hd(bs, backing, &error_abort); + + for (outer = 0; outer < DRAIN_TYPE_MAX; outer++) { + for (inner = 0; inner < DRAIN_TYPE_MAX; inner++) { + /* XXX bdrv_drain_all() doesn't increase the quiesce_counter */ + int bs_quiesce = (outer != BDRV_DRAIN_ALL) + + (inner != BDRV_DRAIN_ALL); + int backing_quiesce = 0; + int backing_cb_cnt = (outer != BDRV_DRAIN) + + (inner != BDRV_DRAIN); + + g_assert_cmpint(bs->quiesce_counter, ==, 0); + g_assert_cmpint(backing->quiesce_counter, ==, 0); + g_assert_cmpint(s->drain_count, ==, 0); + g_assert_cmpint(backing_s->drain_count, ==, 0); + + do_drain_begin(outer, bs); + do_drain_begin(inner, bs); + + g_assert_cmpint(bs->quiesce_counter, ==, bs_quiesce); + g_assert_cmpint(backing->quiesce_counter, ==, backing_quiesce); + g_assert_cmpint(s->drain_count, ==, 2); + g_assert_cmpint(backing_s->drain_count, ==, backing_cb_cnt); + + do_drain_end(inner, bs); + do_drain_end(outer, bs); + + g_assert_cmpint(bs->quiesce_counter, ==, 0); + g_assert_cmpint(backing->quiesce_counter, ==, 0); + g_assert_cmpint(s->drain_count, ==, 0); + g_assert_cmpint(backing_s->drain_count, ==, 0); + } + } + + bdrv_unref(backing); + bdrv_unref(bs); + blk_unref(blk); +} + typedef struct TestBlockJob { BlockJob common; @@ -XXX,XX +XXX,XX @@ int main(int argc, char **argv) g_test_add_func("/bdrv-drain/quiesce/drain_all", test_quiesce_drain_all); g_test_add_func("/bdrv-drain/quiesce/drain", test_quiesce_drain); + g_test_add_func("/bdrv-drain/nested", test_nested); + g_test_add_func("/bdrv-drain/blockjob/drain_all", test_blockjob_drain_all); g_test_add_func("/bdrv-drain/blockjob/drain", test_blockjob_drain); -- 2.13.6 This is in preparation for subtree drains, i.e. drained sections that affect not only a single node, but recursively all child nodes, too. Calling the parent callbacks for drain is pointless when we just came from that parent node recursively and leads to multiple increases of bs->quiesce_counter in a single drain call. Don't do it. In order for this to work correctly, the parent callback must be called for every bdrv_drain_begin/end() call, not only for the outermost one: If we have a node N with two parents A and B, recursive draining of A should cause the quiesce_counter of B to increase because its child N is drained independently of B. If now B is recursively drained, too, A must increase its quiesce_counter because N is drained independently of A only now, even if N is going from quiesce_counter 1 to 2. Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block.h | 4 ++-- block.c | 13 +++++++++---- block/io.c | 47 ++++++++++++++++++++++++++++++++++------------- 3 files changed, 45 insertions(+), 19 deletions(-) diff --git a/include/block/block.h b/include/block/block.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block.h +++ b/include/block/block.h @@ -XXX,XX +XXX,XX @@ void bdrv_io_unplug(BlockDriverState *bs); * Begin a quiesced section of all users of @bs. This is part of * bdrv_drained_begin. */ -void bdrv_parent_drained_begin(BlockDriverState *bs); +void bdrv_parent_drained_begin(BlockDriverState *bs, BdrvChild *ignore); /** * bdrv_parent_drained_end: @@ -XXX,XX +XXX,XX @@ void bdrv_parent_drained_begin(BlockDriverState *bs); * End a quiesced section of all users of @bs. This is part of * bdrv_drained_end. */ -void bdrv_parent_drained_end(BlockDriverState *bs); +void bdrv_parent_drained_end(BlockDriverState *bs, BdrvChild *ignore); /** * bdrv_drained_begin: diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ static void bdrv_replace_child_noperm(BdrvChild *child, BlockDriverState *new_bs) { BlockDriverState *old_bs = child->bs; + int i; if (old_bs && new_bs) { assert(bdrv_get_aio_context(old_bs) == bdrv_get_aio_context(new_bs)); } if (old_bs) { if (old_bs->quiesce_counter && child->role->drained_end) { - child->role->drained_end(child); + for (i = 0; i < old_bs->quiesce_counter; i++) { + child->role->drained_end(child); + } } if (child->role->detach) { child->role->detach(child); @@ -XXX,XX +XXX,XX @@ static void bdrv_replace_child_noperm(BdrvChild *child, if (new_bs) { QLIST_INSERT_HEAD(&new_bs->parents, child, next_parent); if (new_bs->quiesce_counter && child->role->drained_begin) { - child->role->drained_begin(child); + for (i = 0; i < new_bs->quiesce_counter; i++) { + child->role->drained_begin(child); + } } if (child->role->attach) { @@ -XXX,XX +XXX,XX @@ void bdrv_set_aio_context(BlockDriverState *bs, AioContext *new_context) AioContext *ctx = bdrv_get_aio_context(bs); aio_disable_external(ctx); - bdrv_parent_drained_begin(bs); + bdrv_parent_drained_begin(bs, NULL); bdrv_drain(bs); /* ensure there are no in-flight requests */ while (aio_poll(ctx, false)) { @@ -XXX,XX +XXX,XX @@ void bdrv_set_aio_context(BlockDriverState *bs, AioContext *new_context) */ aio_context_acquire(new_context); bdrv_attach_aio_context(bs, new_context); - bdrv_parent_drained_end(bs); + bdrv_parent_drained_end(bs, NULL); aio_enable_external(ctx); aio_context_release(new_context); } diff --git a/block/io.c b/block/io.c index XXXXXXX..XXXXXXX 100644 --- a/block/io.c +++ b/block/io.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_co_do_pwrite_zeroes(BlockDriverState *bs, int64_t offset, int bytes, BdrvRequestFlags flags); -void bdrv_parent_drained_begin(BlockDriverState *bs) +void bdrv_parent_drained_begin(BlockDriverState *bs, BdrvChild *ignore) { BdrvChild *c, *next; QLIST_FOREACH_SAFE(c, &bs->parents, next_parent, next) { + if (c == ignore) { + continue; + } if (c->role->drained_begin) { c->role->drained_begin(c); } } } -void bdrv_parent_drained_end(BlockDriverState *bs) +void bdrv_parent_drained_end(BlockDriverState *bs, BdrvChild *ignore) { BdrvChild *c, *next; QLIST_FOREACH_SAFE(c, &bs->parents, next_parent, next) { + if (c == ignore) { + continue; + } if (c->role->drained_end) { c->role->drained_end(c); } @@ -XXX,XX +XXX,XX @@ typedef struct { BlockDriverState *bs; bool done; bool begin; + BdrvChild *parent; } BdrvCoDrainData; static void coroutine_fn bdrv_drain_invoke_entry(void *opaque) @@ -XXX,XX +XXX,XX @@ static bool bdrv_drain_recurse(BlockDriverState *bs) return waited; } +static void bdrv_do_drained_begin(BlockDriverState *bs, BdrvChild *parent); +static void bdrv_do_drained_end(BlockDriverState *bs, BdrvChild *parent); + static void bdrv_co_drain_bh_cb(void *opaque) { BdrvCoDrainData *data = opaque; @@ -XXX,XX +XXX,XX @@ static void bdrv_co_drain_bh_cb(void *opaque) bdrv_dec_in_flight(bs); if (data->begin) { - bdrv_drained_begin(bs); + bdrv_do_drained_begin(bs, data->parent); } else { - bdrv_drained_end(bs); + bdrv_do_drained_end(bs, data->parent); } data->done = true; @@ -XXX,XX +XXX,XX @@ static void bdrv_co_drain_bh_cb(void *opaque) } static void coroutine_fn bdrv_co_yield_to_drain(BlockDriverState *bs, - bool begin) + bool begin, BdrvChild *parent) { BdrvCoDrainData data; @@ -XXX,XX +XXX,XX @@ static void coroutine_fn bdrv_co_yield_to_drain(BlockDriverState *bs, .bs = bs, .done = false, .begin = begin, + .parent = parent, }; bdrv_inc_in_flight(bs); aio_bh_schedule_oneshot(bdrv_get_aio_context(bs), @@ -XXX,XX +XXX,XX @@ static void coroutine_fn bdrv_co_yield_to_drain(BlockDriverState *bs, assert(data.done); } -void bdrv_drained_begin(BlockDriverState *bs) +static void bdrv_do_drained_begin(BlockDriverState *bs, BdrvChild *parent) { if (qemu_in_coroutine()) { - bdrv_co_yield_to_drain(bs, true); + bdrv_co_yield_to_drain(bs, true, parent); return; } /* Stop things in parent-to-child order */ if (atomic_fetch_inc(&bs->quiesce_counter) == 0) { aio_disable_external(bdrv_get_aio_context(bs)); - bdrv_parent_drained_begin(bs); } + bdrv_parent_drained_begin(bs, parent); bdrv_drain_invoke(bs, true, false); bdrv_drain_recurse(bs); } -void bdrv_drained_end(BlockDriverState *bs) +void bdrv_drained_begin(BlockDriverState *bs) +{ + bdrv_do_drained_begin(bs, NULL); +} + +static void bdrv_do_drained_end(BlockDriverState *bs, BdrvChild *parent) { int old_quiesce_counter; if (qemu_in_coroutine()) { - bdrv_co_yield_to_drain(bs, false); + bdrv_co_yield_to_drain(bs, false, parent); return; } assert(bs->quiesce_counter > 0); @@ -XXX,XX +XXX,XX @@ void bdrv_drained_end(BlockDriverState *bs) /* Re-enable things in child-to-parent order */ bdrv_drain_invoke(bs, false, false); + bdrv_parent_drained_end(bs, parent); if (old_quiesce_counter == 1) { - bdrv_parent_drained_end(bs); aio_enable_external(bdrv_get_aio_context(bs)); } } +void bdrv_drained_end(BlockDriverState *bs) +{ + bdrv_do_drained_end(bs, NULL); +} + /* * Wait for pending requests to complete on a single BlockDriverState subtree, * and suspend block driver's internal I/O until next request arrives. @@ -XXX,XX +XXX,XX @@ void bdrv_drain_all_begin(void) /* Stop things in parent-to-child order */ aio_context_acquire(aio_context); aio_disable_external(aio_context); - bdrv_parent_drained_begin(bs); + bdrv_parent_drained_begin(bs, NULL); bdrv_drain_invoke(bs, true, true); aio_context_release(aio_context); @@ -XXX,XX +XXX,XX @@ void bdrv_drain_all_end(void) /* Re-enable things in child-to-parent order */ aio_context_acquire(aio_context); bdrv_drain_invoke(bs, false, true); - bdrv_parent_drained_end(bs); + bdrv_parent_drained_end(bs, NULL); aio_enable_external(aio_context); aio_context_release(aio_context); } -- 2.13.6 bdrv_drained_begin() waits for the completion of requests in the whole subtree, but it only actually keeps its immediate bs parameter quiesced until bdrv_drained_end(). Add a version that keeps the whole subtree drained. As of this commit, graph changes cannot be allowed during a subtree drained section, but this will be fixed soon. Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block.h | 13 +++++++++++++ block/io.c | 54 ++++++++++++++++++++++++++++++++++++++++----------- 2 files changed, 56 insertions(+), 11 deletions(-) diff --git a/include/block/block.h b/include/block/block.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block.h +++ b/include/block/block.h @@ -XXX,XX +XXX,XX @@ void bdrv_parent_drained_end(BlockDriverState *bs, BdrvChild *ignore); void bdrv_drained_begin(BlockDriverState *bs); /** + * Like bdrv_drained_begin, but recursively begins a quiesced section for + * exclusive access to all child nodes as well. + * + * Graph changes are not allowed during a subtree drain section. + */ +void bdrv_subtree_drained_begin(BlockDriverState *bs); + +/** * bdrv_drained_end: * * End a quiescent section started by bdrv_drained_begin(). */ void bdrv_drained_end(BlockDriverState *bs); +/** + * End a quiescent section started by bdrv_subtree_drained_begin(). + */ +void bdrv_subtree_drained_end(BlockDriverState *bs); + void bdrv_add_child(BlockDriverState *parent, BlockDriverState *child, Error **errp); void bdrv_del_child(BlockDriverState *parent, BdrvChild *child, 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 @@ typedef struct { BlockDriverState *bs; bool done; bool begin; + bool recursive; BdrvChild *parent; } BdrvCoDrainData; @@ -XXX,XX +XXX,XX @@ static bool bdrv_drain_recurse(BlockDriverState *bs) return waited; } -static void bdrv_do_drained_begin(BlockDriverState *bs, BdrvChild *parent); -static void bdrv_do_drained_end(BlockDriverState *bs, BdrvChild *parent); +static void bdrv_do_drained_begin(BlockDriverState *bs, bool recursive, + BdrvChild *parent); +static void bdrv_do_drained_end(BlockDriverState *bs, bool recursive, + BdrvChild *parent); static void bdrv_co_drain_bh_cb(void *opaque) { @@ -XXX,XX +XXX,XX @@ static void bdrv_co_drain_bh_cb(void *opaque) bdrv_dec_in_flight(bs); if (data->begin) { - bdrv_do_drained_begin(bs, data->parent); + bdrv_do_drained_begin(bs, data->recursive, data->parent); } else { - bdrv_do_drained_end(bs, data->parent); + bdrv_do_drained_end(bs, data->recursive, data->parent); } data->done = true; @@ -XXX,XX +XXX,XX @@ static void bdrv_co_drain_bh_cb(void *opaque) } static void coroutine_fn bdrv_co_yield_to_drain(BlockDriverState *bs, - bool begin, BdrvChild *parent) + bool begin, bool recursive, + BdrvChild *parent) { BdrvCoDrainData data; @@ -XXX,XX +XXX,XX @@ static void coroutine_fn bdrv_co_yield_to_drain(BlockDriverState *bs, .bs = bs, .done = false, .begin = begin, + .recursive = recursive, .parent = parent, }; bdrv_inc_in_flight(bs); @@ -XXX,XX +XXX,XX @@ static void coroutine_fn bdrv_co_yield_to_drain(BlockDriverState *bs, assert(data.done); } -static void bdrv_do_drained_begin(BlockDriverState *bs, BdrvChild *parent) +static void bdrv_do_drained_begin(BlockDriverState *bs, bool recursive, + BdrvChild *parent) { + BdrvChild *child, *next; + if (qemu_in_coroutine()) { - bdrv_co_yield_to_drain(bs, true, parent); + bdrv_co_yield_to_drain(bs, true, recursive, parent); return; } @@ -XXX,XX +XXX,XX @@ static void bdrv_do_drained_begin(BlockDriverState *bs, BdrvChild *parent) bdrv_parent_drained_begin(bs, parent); bdrv_drain_invoke(bs, true, false); bdrv_drain_recurse(bs); + + if (recursive) { + QLIST_FOREACH_SAFE(child, &bs->children, next, next) { + bdrv_do_drained_begin(child->bs, true, child); + } + } } void bdrv_drained_begin(BlockDriverState *bs) { - bdrv_do_drained_begin(bs, NULL); + bdrv_do_drained_begin(bs, false, NULL); +} + +void bdrv_subtree_drained_begin(BlockDriverState *bs) +{ + bdrv_do_drained_begin(bs, true, NULL); } -static void bdrv_do_drained_end(BlockDriverState *bs, BdrvChild *parent) +static void bdrv_do_drained_end(BlockDriverState *bs, bool recursive, + BdrvChild *parent) { + BdrvChild *child, *next; int old_quiesce_counter; if (qemu_in_coroutine()) { - bdrv_co_yield_to_drain(bs, false, parent); + bdrv_co_yield_to_drain(bs, false, recursive, parent); return; } assert(bs->quiesce_counter > 0); @@ -XXX,XX +XXX,XX @@ static void bdrv_do_drained_end(BlockDriverState *bs, BdrvChild *parent) if (old_quiesce_counter == 1) { aio_enable_external(bdrv_get_aio_context(bs)); } + + if (recursive) { + QLIST_FOREACH_SAFE(child, &bs->children, next, next) { + bdrv_do_drained_end(child->bs, true, child); + } + } } void bdrv_drained_end(BlockDriverState *bs) { - bdrv_do_drained_end(bs, NULL); + bdrv_do_drained_end(bs, false, NULL); +} + +void bdrv_subtree_drained_end(BlockDriverState *bs) +{ + bdrv_do_drained_end(bs, true, NULL); } /* -- 2.13.6 Add a subtree drain version to the existing test cases. Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- tests/test-bdrv-drain.c | 27 ++++++++++++++++++++++++++- 1 file changed, 26 insertions(+), 1 deletion(-) diff --git a/tests/test-bdrv-drain.c b/tests/test-bdrv-drain.c index XXXXXXX..XXXXXXX 100644 --- a/tests/test-bdrv-drain.c +++ b/tests/test-bdrv-drain.c @@ -XXX,XX +XXX,XX @@ static void aio_ret_cb(void *opaque, int ret) enum drain_type { BDRV_DRAIN_ALL, BDRV_DRAIN, + BDRV_SUBTREE_DRAIN, DRAIN_TYPE_MAX, }; @@ -XXX,XX +XXX,XX @@ static void do_drain_begin(enum drain_type drain_type, BlockDriverState *bs) switch (drain_type) { case BDRV_DRAIN_ALL: bdrv_drain_all_begin(); break; case BDRV_DRAIN: bdrv_drained_begin(bs); break; + case BDRV_SUBTREE_DRAIN: bdrv_subtree_drained_begin(bs); break; default: g_assert_not_reached(); } } @@ -XXX,XX +XXX,XX @@ static void do_drain_end(enum drain_type drain_type, BlockDriverState *bs) switch (drain_type) { case BDRV_DRAIN_ALL: bdrv_drain_all_end(); break; case BDRV_DRAIN: bdrv_drained_end(bs); break; + case BDRV_SUBTREE_DRAIN: bdrv_subtree_drained_end(bs); break; default: g_assert_not_reached(); } } @@ -XXX,XX +XXX,XX @@ static void test_drv_cb_drain(void) test_drv_cb_common(BDRV_DRAIN, false); } +static void test_drv_cb_drain_subtree(void) +{ + test_drv_cb_common(BDRV_SUBTREE_DRAIN, true); +} + static void test_quiesce_common(enum drain_type drain_type, bool recursive) { BlockBackend *blk; @@ -XXX,XX +XXX,XX @@ static void test_quiesce_drain(void) test_quiesce_common(BDRV_DRAIN, false); } +static void test_quiesce_drain_subtree(void) +{ + test_quiesce_common(BDRV_SUBTREE_DRAIN, true); +} + static void test_nested(void) { BlockBackend *blk; @@ -XXX,XX +XXX,XX @@ static void test_nested(void) /* XXX bdrv_drain_all() doesn't increase the quiesce_counter */ int bs_quiesce = (outer != BDRV_DRAIN_ALL) + (inner != BDRV_DRAIN_ALL); - int backing_quiesce = 0; + int backing_quiesce = (outer == BDRV_SUBTREE_DRAIN) + + (inner == BDRV_SUBTREE_DRAIN); int backing_cb_cnt = (outer != BDRV_DRAIN) + (inner != BDRV_DRAIN); @@ -XXX,XX +XXX,XX @@ static void test_blockjob_drain(void) test_blockjob_common(BDRV_DRAIN); } +static void test_blockjob_drain_subtree(void) +{ + test_blockjob_common(BDRV_SUBTREE_DRAIN); +} + int main(int argc, char **argv) { bdrv_init(); @@ -XXX,XX +XXX,XX @@ int main(int argc, char **argv) g_test_add_func("/bdrv-drain/driver-cb/drain_all", test_drv_cb_drain_all); g_test_add_func("/bdrv-drain/driver-cb/drain", test_drv_cb_drain); + g_test_add_func("/bdrv-drain/driver-cb/drain_subtree", + test_drv_cb_drain_subtree); g_test_add_func("/bdrv-drain/quiesce/drain_all", test_quiesce_drain_all); g_test_add_func("/bdrv-drain/quiesce/drain", test_quiesce_drain); + g_test_add_func("/bdrv-drain/quiesce/drain_subtree", + test_quiesce_drain_subtree); g_test_add_func("/bdrv-drain/nested", test_nested); g_test_add_func("/bdrv-drain/blockjob/drain_all", test_blockjob_drain_all); g_test_add_func("/bdrv-drain/blockjob/drain", test_blockjob_drain); + g_test_add_func("/bdrv-drain/blockjob/drain_subtree", + test_blockjob_drain_subtree); return g_test_run(); } -- 2.13.6 If bdrv_do_drained_begin/end() are called in coroutine context, they first use a BH to get out of the coroutine context. Call some existing tests again from a coroutine to cover this code path. Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- tests/test-bdrv-drain.c | 59 +++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 59 insertions(+) diff --git a/tests/test-bdrv-drain.c b/tests/test-bdrv-drain.c index XXXXXXX..XXXXXXX 100644 --- a/tests/test-bdrv-drain.c +++ b/tests/test-bdrv-drain.c @@ -XXX,XX +XXX,XX @@ static void aio_ret_cb(void *opaque, int ret) *aio_ret = ret; } +typedef struct CallInCoroutineData { + void (*entry)(void); + bool done; +} CallInCoroutineData; + +static coroutine_fn void call_in_coroutine_entry(void *opaque) +{ + CallInCoroutineData *data = opaque; + + data->entry(); + data->done = true; +} + +static void call_in_coroutine(void (*entry)(void)) +{ + Coroutine *co; + CallInCoroutineData data = { + .entry = entry, + .done = false, + }; + + co = qemu_coroutine_create(call_in_coroutine_entry, &data); + qemu_coroutine_enter(co); + while (!data.done) { + aio_poll(qemu_get_aio_context(), true); + } +} + enum drain_type { BDRV_DRAIN_ALL, BDRV_DRAIN, @@ -XXX,XX +XXX,XX @@ static void test_drv_cb_drain_subtree(void) test_drv_cb_common(BDRV_SUBTREE_DRAIN, true); } +static void test_drv_cb_co_drain(void) +{ + call_in_coroutine(test_drv_cb_drain); +} + +static void test_drv_cb_co_drain_subtree(void) +{ + call_in_coroutine(test_drv_cb_drain_subtree); +} + static void test_quiesce_common(enum drain_type drain_type, bool recursive) { BlockBackend *blk; @@ -XXX,XX +XXX,XX @@ static void test_quiesce_drain_subtree(void) test_quiesce_common(BDRV_SUBTREE_DRAIN, true); } +static void test_quiesce_co_drain(void) +{ + call_in_coroutine(test_quiesce_drain); +} + +static void test_quiesce_co_drain_subtree(void) +{ + call_in_coroutine(test_quiesce_drain_subtree); +} + static void test_nested(void) { BlockBackend *blk; @@ -XXX,XX +XXX,XX @@ int main(int argc, char **argv) g_test_add_func("/bdrv-drain/driver-cb/drain_subtree", test_drv_cb_drain_subtree); + // XXX bdrv_drain_all() doesn't work in coroutine context + g_test_add_func("/bdrv-drain/driver-cb/co/drain", test_drv_cb_co_drain); + g_test_add_func("/bdrv-drain/driver-cb/co/drain_subtree", + test_drv_cb_co_drain_subtree); + + g_test_add_func("/bdrv-drain/quiesce/drain_all", test_quiesce_drain_all); g_test_add_func("/bdrv-drain/quiesce/drain", test_quiesce_drain); g_test_add_func("/bdrv-drain/quiesce/drain_subtree", test_quiesce_drain_subtree); + // XXX bdrv_drain_all() doesn't work in coroutine context + g_test_add_func("/bdrv-drain/quiesce/co/drain", test_quiesce_co_drain); + g_test_add_func("/bdrv-drain/quiesce/co/drain_subtree", + test_quiesce_co_drain_subtree); + g_test_add_func("/bdrv-drain/nested", test_nested); g_test_add_func("/bdrv-drain/blockjob/drain_all", test_blockjob_drain_all); -- 2.13.6 Test that drain sections are correctly propagated through the graph. Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- tests/test-bdrv-drain.c | 74 +++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 74 insertions(+) diff --git a/tests/test-bdrv-drain.c b/tests/test-bdrv-drain.c index XXXXXXX..XXXXXXX 100644 --- a/tests/test-bdrv-drain.c +++ b/tests/test-bdrv-drain.c @@ -XXX,XX +XXX,XX @@ static void test_nested(void) blk_unref(blk); } +static void test_multiparent(void) +{ + BlockBackend *blk_a, *blk_b; + BlockDriverState *bs_a, *bs_b, *backing; + BDRVTestState *a_s, *b_s, *backing_s; + + blk_a = blk_new(BLK_PERM_ALL, BLK_PERM_ALL); + bs_a = bdrv_new_open_driver(&bdrv_test, "test-node-a", BDRV_O_RDWR, + &error_abort); + a_s = bs_a->opaque; + blk_insert_bs(blk_a, bs_a, &error_abort); + + blk_b = blk_new(BLK_PERM_ALL, BLK_PERM_ALL); + bs_b = bdrv_new_open_driver(&bdrv_test, "test-node-b", BDRV_O_RDWR, + &error_abort); + b_s = bs_b->opaque; + blk_insert_bs(blk_b, bs_b, &error_abort); + + backing = bdrv_new_open_driver(&bdrv_test, "backing", 0, &error_abort); + backing_s = backing->opaque; + bdrv_set_backing_hd(bs_a, backing, &error_abort); + bdrv_set_backing_hd(bs_b, backing, &error_abort); + + g_assert_cmpint(bs_a->quiesce_counter, ==, 0); + g_assert_cmpint(bs_b->quiesce_counter, ==, 0); + g_assert_cmpint(backing->quiesce_counter, ==, 0); + g_assert_cmpint(a_s->drain_count, ==, 0); + g_assert_cmpint(b_s->drain_count, ==, 0); + g_assert_cmpint(backing_s->drain_count, ==, 0); + + do_drain_begin(BDRV_SUBTREE_DRAIN, bs_a); + + g_assert_cmpint(bs_a->quiesce_counter, ==, 1); + g_assert_cmpint(bs_b->quiesce_counter, ==, 1); + g_assert_cmpint(backing->quiesce_counter, ==, 1); + g_assert_cmpint(a_s->drain_count, ==, 1); + g_assert_cmpint(b_s->drain_count, ==, 1); + g_assert_cmpint(backing_s->drain_count, ==, 1); + + do_drain_begin(BDRV_SUBTREE_DRAIN, bs_b); + + g_assert_cmpint(bs_a->quiesce_counter, ==, 2); + g_assert_cmpint(bs_b->quiesce_counter, ==, 2); + g_assert_cmpint(backing->quiesce_counter, ==, 2); + g_assert_cmpint(a_s->drain_count, ==, 2); + g_assert_cmpint(b_s->drain_count, ==, 2); + g_assert_cmpint(backing_s->drain_count, ==, 2); + + do_drain_end(BDRV_SUBTREE_DRAIN, bs_b); + + g_assert_cmpint(bs_a->quiesce_counter, ==, 1); + g_assert_cmpint(bs_b->quiesce_counter, ==, 1); + g_assert_cmpint(backing->quiesce_counter, ==, 1); + g_assert_cmpint(a_s->drain_count, ==, 1); + g_assert_cmpint(b_s->drain_count, ==, 1); + g_assert_cmpint(backing_s->drain_count, ==, 1); + + do_drain_end(BDRV_SUBTREE_DRAIN, bs_a); + + g_assert_cmpint(bs_a->quiesce_counter, ==, 0); + g_assert_cmpint(bs_b->quiesce_counter, ==, 0); + g_assert_cmpint(backing->quiesce_counter, ==, 0); + g_assert_cmpint(a_s->drain_count, ==, 0); + g_assert_cmpint(b_s->drain_count, ==, 0); + g_assert_cmpint(backing_s->drain_count, ==, 0); + + bdrv_unref(backing); + bdrv_unref(bs_a); + bdrv_unref(bs_b); + blk_unref(blk_a); + blk_unref(blk_b); +} + typedef struct TestBlockJob { BlockJob common; @@ -XXX,XX +XXX,XX @@ int main(int argc, char **argv) test_quiesce_co_drain_subtree); g_test_add_func("/bdrv-drain/nested", test_nested); + g_test_add_func("/bdrv-drain/multiparent", test_multiparent); g_test_add_func("/bdrv-drain/blockjob/drain_all", test_blockjob_drain_all); g_test_add_func("/bdrv-drain/blockjob/drain", test_blockjob_drain); -- 2.13.6 We need to remember how many of the drain sections in which a node is were recursive (i.e. subtree drain rather than node drain), so that they can be correctly applied when children are added or removed during the drained section. With this change, it is safe to modify the graph even inside a bdrv_subtree_drained_begin/end() section. Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block.h | 2 -- include/block/block_int.h | 5 +++++ block.c | 32 +++++++++++++++++++++++++++++--- block/io.c | 28 ++++++++++++++++++++++++---- 4 files changed, 58 insertions(+), 9 deletions(-) diff --git a/include/block/block.h b/include/block/block.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block.h +++ b/include/block/block.h @@ -XXX,XX +XXX,XX @@ void bdrv_drained_begin(BlockDriverState *bs); /** * Like bdrv_drained_begin, but recursively begins a quiesced section for * exclusive access to all child nodes as well. - * - * Graph changes are not allowed during a subtree drain section. */ void bdrv_subtree_drained_begin(BlockDriverState *bs); diff --git a/include/block/block_int.h b/include/block/block_int.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block_int.h +++ b/include/block/block_int.h @@ -XXX,XX +XXX,XX @@ struct BlockDriverState { /* Accessed with atomic ops. */ int quiesce_counter; + int recursive_quiesce_counter; + unsigned int write_gen; /* Current data generation */ /* Protected by reqs_lock. */ @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_pwritev(BdrvChild *child, int64_t offset, unsigned int bytes, QEMUIOVector *qiov, BdrvRequestFlags flags); +void bdrv_apply_subtree_drain(BdrvChild *child, BlockDriverState *new_parent); +void bdrv_unapply_subtree_drain(BdrvChild *child, BlockDriverState *old_parent); + int get_tmp_filename(char *filename, int size); BlockDriver *bdrv_probe_all(const uint8_t *buf, int buf_size, const char *filename); diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ static void bdrv_child_cb_drained_end(BdrvChild *child) bdrv_drained_end(bs); } +static void bdrv_child_cb_attach(BdrvChild *child) +{ + BlockDriverState *bs = child->opaque; + bdrv_apply_subtree_drain(child, bs); +} + +static void bdrv_child_cb_detach(BdrvChild *child) +{ + BlockDriverState *bs = child->opaque; + bdrv_unapply_subtree_drain(child, bs); +} + static int bdrv_child_cb_inactivate(BdrvChild *child) { BlockDriverState *bs = child->opaque; @@ -XXX,XX +XXX,XX @@ const BdrvChildRole child_file = { .inherit_options = bdrv_inherited_options, .drained_begin = bdrv_child_cb_drained_begin, .drained_end = bdrv_child_cb_drained_end, + .attach = bdrv_child_cb_attach, + .detach = bdrv_child_cb_detach, .inactivate = bdrv_child_cb_inactivate, }; @@ -XXX,XX +XXX,XX @@ const BdrvChildRole child_format = { .inherit_options = bdrv_inherited_fmt_options, .drained_begin = bdrv_child_cb_drained_begin, .drained_end = bdrv_child_cb_drained_end, + .attach = bdrv_child_cb_attach, + .detach = bdrv_child_cb_detach, .inactivate = bdrv_child_cb_inactivate, }; @@ -XXX,XX +XXX,XX @@ static void bdrv_backing_attach(BdrvChild *c) parent->backing_blocker); bdrv_op_unblock(backing_hd, BLOCK_OP_TYPE_BACKUP_TARGET, parent->backing_blocker); + + bdrv_child_cb_attach(c); } static void bdrv_backing_detach(BdrvChild *c) @@ -XXX,XX +XXX,XX @@ static void bdrv_backing_detach(BdrvChild *c) bdrv_op_unblock_all(c->bs, parent->backing_blocker); error_free(parent->backing_blocker); parent->backing_blocker = NULL; + + bdrv_child_cb_detach(c); } /* @@ -XXX,XX +XXX,XX @@ static void bdrv_replace_child_noperm(BdrvChild *child, assert(bdrv_get_aio_context(old_bs) == bdrv_get_aio_context(new_bs)); } if (old_bs) { + /* Detach first so that the recursive drain sections coming from @child + * are already gone and we only end the drain sections that came from + * elsewhere. */ + if (child->role->detach) { + child->role->detach(child); + } if (old_bs->quiesce_counter && child->role->drained_end) { for (i = 0; i < old_bs->quiesce_counter; i++) { child->role->drained_end(child); } } - if (child->role->detach) { - child->role->detach(child); - } QLIST_REMOVE(child, next_parent); } @@ -XXX,XX +XXX,XX @@ static void bdrv_replace_child_noperm(BdrvChild *child, } } + /* Attach only after starting new drained sections, so that recursive + * drain sections coming from @child don't get an extra .drained_begin + * callback. */ if (child->role->attach) { child->role->attach(child); } 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 void coroutine_fn bdrv_co_yield_to_drain(BlockDriverState *bs, assert(data.done); } -static void bdrv_do_drained_begin(BlockDriverState *bs, bool recursive, - BdrvChild *parent) +void bdrv_do_drained_begin(BlockDriverState *bs, bool recursive, + BdrvChild *parent) { BdrvChild *child, *next; @@ -XXX,XX +XXX,XX @@ static void bdrv_do_drained_begin(BlockDriverState *bs, bool recursive, bdrv_drain_recurse(bs); if (recursive) { + bs->recursive_quiesce_counter++; QLIST_FOREACH_SAFE(child, &bs->children, next, next) { bdrv_do_drained_begin(child->bs, true, child); } @@ -XXX,XX +XXX,XX @@ void bdrv_subtree_drained_begin(BlockDriverState *bs) bdrv_do_drained_begin(bs, true, NULL); } -static void bdrv_do_drained_end(BlockDriverState *bs, bool recursive, - BdrvChild *parent) +void bdrv_do_drained_end(BlockDriverState *bs, bool recursive, + BdrvChild *parent) { BdrvChild *child, *next; int old_quiesce_counter; @@ -XXX,XX +XXX,XX @@ static void bdrv_do_drained_end(BlockDriverState *bs, bool recursive, } if (recursive) { + bs->recursive_quiesce_counter--; QLIST_FOREACH_SAFE(child, &bs->children, next, next) { bdrv_do_drained_end(child->bs, true, child); } @@ -XXX,XX +XXX,XX @@ void bdrv_subtree_drained_end(BlockDriverState *bs) bdrv_do_drained_end(bs, true, NULL); } +void bdrv_apply_subtree_drain(BdrvChild *child, BlockDriverState *new_parent) +{ + int i; + + for (i = 0; i < new_parent->recursive_quiesce_counter; i++) { + bdrv_do_drained_begin(child->bs, true, child); + } +} + +void bdrv_unapply_subtree_drain(BdrvChild *child, BlockDriverState *old_parent) +{ + int i; + + for (i = 0; i < old_parent->recursive_quiesce_counter; i++) { + bdrv_do_drained_end(child->bs, true, child); + } +} + /* * Wait for pending requests to complete on a single BlockDriverState subtree, * and suspend block driver's internal I/O until next request arrives. -- 2.13.6 Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- tests/test-bdrv-drain.c | 80 +++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 80 insertions(+) diff --git a/tests/test-bdrv-drain.c b/tests/test-bdrv-drain.c index XXXXXXX..XXXXXXX 100644 --- a/tests/test-bdrv-drain.c +++ b/tests/test-bdrv-drain.c @@ -XXX,XX +XXX,XX @@ static void test_multiparent(void) blk_unref(blk_b); } +static void test_graph_change(void) +{ + BlockBackend *blk_a, *blk_b; + BlockDriverState *bs_a, *bs_b, *backing; + BDRVTestState *a_s, *b_s, *backing_s; + + blk_a = blk_new(BLK_PERM_ALL, BLK_PERM_ALL); + bs_a = bdrv_new_open_driver(&bdrv_test, "test-node-a", BDRV_O_RDWR, + &error_abort); + a_s = bs_a->opaque; + blk_insert_bs(blk_a, bs_a, &error_abort); + + blk_b = blk_new(BLK_PERM_ALL, BLK_PERM_ALL); + bs_b = bdrv_new_open_driver(&bdrv_test, "test-node-b", BDRV_O_RDWR, + &error_abort); + b_s = bs_b->opaque; + blk_insert_bs(blk_b, bs_b, &error_abort); + + backing = bdrv_new_open_driver(&bdrv_test, "backing", 0, &error_abort); + backing_s = backing->opaque; + bdrv_set_backing_hd(bs_a, backing, &error_abort); + + g_assert_cmpint(bs_a->quiesce_counter, ==, 0); + g_assert_cmpint(bs_b->quiesce_counter, ==, 0); + g_assert_cmpint(backing->quiesce_counter, ==, 0); + g_assert_cmpint(a_s->drain_count, ==, 0); + g_assert_cmpint(b_s->drain_count, ==, 0); + g_assert_cmpint(backing_s->drain_count, ==, 0); + + do_drain_begin(BDRV_SUBTREE_DRAIN, bs_a); + do_drain_begin(BDRV_SUBTREE_DRAIN, bs_a); + do_drain_begin(BDRV_SUBTREE_DRAIN, bs_a); + do_drain_begin(BDRV_SUBTREE_DRAIN, bs_b); + do_drain_begin(BDRV_SUBTREE_DRAIN, bs_b); + + bdrv_set_backing_hd(bs_b, backing, &error_abort); + g_assert_cmpint(bs_a->quiesce_counter, ==, 5); + g_assert_cmpint(bs_b->quiesce_counter, ==, 5); + g_assert_cmpint(backing->quiesce_counter, ==, 5); + g_assert_cmpint(a_s->drain_count, ==, 5); + g_assert_cmpint(b_s->drain_count, ==, 5); + g_assert_cmpint(backing_s->drain_count, ==, 5); + + bdrv_set_backing_hd(bs_b, NULL, &error_abort); + g_assert_cmpint(bs_a->quiesce_counter, ==, 3); + g_assert_cmpint(bs_b->quiesce_counter, ==, 2); + g_assert_cmpint(backing->quiesce_counter, ==, 3); + g_assert_cmpint(a_s->drain_count, ==, 3); + g_assert_cmpint(b_s->drain_count, ==, 2); + g_assert_cmpint(backing_s->drain_count, ==, 3); + + bdrv_set_backing_hd(bs_b, backing, &error_abort); + g_assert_cmpint(bs_a->quiesce_counter, ==, 5); + g_assert_cmpint(bs_b->quiesce_counter, ==, 5); + g_assert_cmpint(backing->quiesce_counter, ==, 5); + g_assert_cmpint(a_s->drain_count, ==, 5); + g_assert_cmpint(b_s->drain_count, ==, 5); + g_assert_cmpint(backing_s->drain_count, ==, 5); + + do_drain_end(BDRV_SUBTREE_DRAIN, bs_b); + do_drain_end(BDRV_SUBTREE_DRAIN, bs_b); + do_drain_end(BDRV_SUBTREE_DRAIN, bs_a); + do_drain_end(BDRV_SUBTREE_DRAIN, bs_a); + do_drain_end(BDRV_SUBTREE_DRAIN, bs_a); + + g_assert_cmpint(bs_a->quiesce_counter, ==, 0); + g_assert_cmpint(bs_b->quiesce_counter, ==, 0); + g_assert_cmpint(backing->quiesce_counter, ==, 0); + g_assert_cmpint(a_s->drain_count, ==, 0); + g_assert_cmpint(b_s->drain_count, ==, 0); + g_assert_cmpint(backing_s->drain_count, ==, 0); + + bdrv_unref(backing); + bdrv_unref(bs_a); + bdrv_unref(bs_b); + blk_unref(blk_a); + blk_unref(blk_b); +} + typedef struct TestBlockJob { BlockJob common; @@ -XXX,XX +XXX,XX @@ int main(int argc, char **argv) g_test_add_func("/bdrv-drain/nested", test_nested); g_test_add_func("/bdrv-drain/multiparent", test_multiparent); + g_test_add_func("/bdrv-drain/graph-change", test_graph_change); g_test_add_func("/bdrv-drain/blockjob/drain_all", test_blockjob_drain_all); g_test_add_func("/bdrv-drain/blockjob/drain", test_blockjob_drain); -- 2.13.6 Since commit bde70715, base is the only node that is reopened in commit_start(). This means that the code, which still involves an explicit BlockReopenQueue, can now be simplified by using bdrv_reopen(). Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Fam Zheng <famz@redhat.com> --- block/commit.c | 8 +------- 1 file changed, 1 insertion(+), 7 deletions(-) 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 @@ void commit_start(const char *job_id, BlockDriverState *bs, const char *filter_node_name, Error **errp) { CommitBlockJob *s; - BlockReopenQueue *reopen_queue = NULL; int orig_base_flags; BlockDriverState *iter; BlockDriverState *commit_top_bs = NULL; @@ -XXX,XX +XXX,XX @@ void commit_start(const char *job_id, BlockDriverState *bs, /* convert base to r/w, if necessary */ orig_base_flags = bdrv_get_flags(base); if (!(orig_base_flags & BDRV_O_RDWR)) { - reopen_queue = bdrv_reopen_queue(reopen_queue, base, NULL, - orig_base_flags | BDRV_O_RDWR); - } - - if (reopen_queue) { - bdrv_reopen_multiple(bdrv_get_aio_context(bs), reopen_queue, &local_err); + bdrv_reopen(base, orig_base_flags | BDRV_O_RDWR, &local_err); if (local_err != NULL) { error_propagate(errp, local_err); goto fail; -- 2.13.6 The bdrv_reopen*() implementation doesn't like it if the graph is changed between queuing nodes for reopen and actually reopening them (one of the reasons is that queuing can be recursive). So instead of draining the device only in bdrv_reopen_multiple(), require that callers already drained all affected nodes, and assert this in bdrv_reopen_queue(). Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Fam Zheng <famz@redhat.com> --- block.c | 23 ++++++++++++++++------- block/replication.c | 6 ++++++ qemu-io-cmds.c | 3 +++ 3 files changed, 25 insertions(+), 7 deletions(-) diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ BlockDriverState *bdrv_open(const char *filename, const char *reference, * returns a pointer to bs_queue, which is either the newly allocated * bs_queue, or the existing bs_queue being used. * + * bs must be drained between bdrv_reopen_queue() and bdrv_reopen_multiple(). */ static BlockReopenQueue *bdrv_reopen_queue_child(BlockReopenQueue *bs_queue, BlockDriverState *bs, @@ -XXX,XX +XXX,XX @@ static BlockReopenQueue *bdrv_reopen_queue_child(BlockReopenQueue *bs_queue, BdrvChild *child; QDict *old_options, *explicit_options; + /* Make sure that the caller remembered to use a drained section. This is + * important to avoid graph changes between the recursive queuing here and + * bdrv_reopen_multiple(). */ + assert(bs->quiesce_counter > 0); + if (bs_queue == NULL) { bs_queue = g_new0(BlockReopenQueue, 1); QSIMPLEQ_INIT(bs_queue); @@ -XXX,XX +XXX,XX @@ BlockReopenQueue *bdrv_reopen_queue(BlockReopenQueue *bs_queue, * If all devices prepare successfully, then the changes are committed * to all devices. * + * All affected nodes must be drained between bdrv_reopen_queue() and + * bdrv_reopen_multiple(). */ int bdrv_reopen_multiple(AioContext *ctx, BlockReopenQueue *bs_queue, Error **errp) { @@ -XXX,XX +XXX,XX @@ int bdrv_reopen_multiple(AioContext *ctx, BlockReopenQueue *bs_queue, Error **er assert(bs_queue != NULL); - aio_context_release(ctx); - bdrv_drain_all_begin(); - aio_context_acquire(ctx); - QSIMPLEQ_FOREACH(bs_entry, bs_queue, entry) { + assert(bs_entry->state.bs->quiesce_counter > 0); if (bdrv_reopen_prepare(&bs_entry->state, bs_queue, &local_err)) { error_propagate(errp, local_err); goto cleanup; @@ -XXX,XX +XXX,XX @@ cleanup: } g_free(bs_queue); - bdrv_drain_all_end(); - return ret; } @@ -XXX,XX +XXX,XX @@ int bdrv_reopen(BlockDriverState *bs, int bdrv_flags, Error **errp) { int ret = -1; Error *local_err = NULL; - BlockReopenQueue *queue = bdrv_reopen_queue(NULL, bs, NULL, bdrv_flags); + BlockReopenQueue *queue; + bdrv_subtree_drained_begin(bs); + + queue = bdrv_reopen_queue(NULL, bs, NULL, bdrv_flags); ret = bdrv_reopen_multiple(bdrv_get_aio_context(bs), queue, &local_err); if (local_err != NULL) { error_propagate(errp, local_err); } + + bdrv_subtree_drained_end(bs); + return ret; } 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 reopen_backing_file(BlockDriverState *bs, bool writable, new_secondary_flags = s->orig_secondary_flags; } + bdrv_subtree_drained_begin(s->hidden_disk->bs); + bdrv_subtree_drained_begin(s->secondary_disk->bs); + if (orig_hidden_flags != new_hidden_flags) { reopen_queue = bdrv_reopen_queue(reopen_queue, s->hidden_disk->bs, NULL, new_hidden_flags); @@ -XXX,XX +XXX,XX @@ static void reopen_backing_file(BlockDriverState *bs, bool writable, reopen_queue, &local_err); error_propagate(errp, local_err); } + + bdrv_subtree_drained_end(s->hidden_disk->bs); + bdrv_subtree_drained_end(s->secondary_disk->bs); } static void backup_job_cleanup(BlockDriverState *bs) 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 reopen_f(BlockBackend *blk, int argc, char **argv) opts = qopts ? qemu_opts_to_qdict(qopts, NULL) : NULL; qemu_opts_reset(&reopen_opts); + bdrv_subtree_drained_begin(bs); brq = bdrv_reopen_queue(NULL, bs, opts, flags); bdrv_reopen_multiple(bdrv_get_aio_context(bs), brq, &local_err); + bdrv_subtree_drained_end(bs); + if (local_err) { error_report_err(local_err); } else { -- 2.13.6