The following changes since commit 16aaacb307ed607b9780c12702c44f0fe52edc7e: Merge remote-tracking branch 'remotes/cohuck/tags/s390x-20200430' into staging (2020-04-30 14:00:36 +0100) are available in the Git repository at: git://repo.or.cz/qemu/kevin.git tags/for-upstream for you to fetch changes up to eaae29ef89d498d0eac553c77b554f310a47f809: qemu-storage-daemon: Fix non-string --object properties (2020-04-30 17:51:07 +0200) ---------------------------------------------------------------- Block layer patches: - Fix resize (extending) of short overlays - nvme: introduce PMR support from NVMe 1.4 spec - qemu-storage-daemon: Fix non-string --object properties ---------------------------------------------------------------- Alberto Garcia (1): qcow2: Add incompatibility note between backing files and raw external data files Andrzej Jakowski (1): nvme: introduce PMR support from NVMe 1.4 spec Kevin Wolf (12): block: Add flags to BlockDriver.bdrv_co_truncate() block: Add flags to bdrv(_co)_truncate() block-backend: Add flags to blk_truncate() qcow2: Support BDRV_REQ_ZERO_WRITE for truncate raw-format: Support BDRV_REQ_ZERO_WRITE for truncate file-posix: Support BDRV_REQ_ZERO_WRITE for truncate block: truncate: Don't make backing file data visible iotests: Filter testfiles out in filter_img_info() iotests: Test committing to short backing file qcow2: Forward ZERO_WRITE flag for full preallocation qom: Factor out user_creatable_add_dict() qemu-storage-daemon: Fix non-string --object properties Paolo Bonzini (1): qemu-iotests: allow qcow2 external discarded clusters to contain stale data docs/interop/qcow2.txt | 3 + hw/block/nvme.h | 2 + include/block/block.h | 5 +- include/block/block_int.h | 10 +- include/block/nvme.h | 172 ++++++++++++++++++++++++++ include/qom/object_interfaces.h | 16 +++ include/sysemu/block-backend.h | 2 +- block.c | 3 +- block/block-backend.c | 4 +- block/commit.c | 4 +- block/crypto.c | 7 +- block/file-posix.c | 6 +- block/file-win32.c | 2 +- block/gluster.c | 1 + block/io.c | 43 ++++++- block/iscsi.c | 2 +- block/mirror.c | 2 +- block/nfs.c | 3 +- block/parallels.c | 6 +- block/qcow.c | 4 +- block/qcow2-cluster.c | 2 +- block/qcow2-refcount.c | 2 +- block/qcow2.c | 73 +++++++++-- block/qed.c | 3 +- block/raw-format.c | 6 +- block/rbd.c | 1 + block/sheepdog.c | 4 +- block/ssh.c | 2 +- block/vdi.c | 2 +- block/vhdx-log.c | 2 +- block/vhdx.c | 6 +- block/vmdk.c | 8 +- block/vpc.c | 2 +- blockdev.c | 2 +- hw/block/nvme.c | 109 ++++++++++++++++ qemu-img.c | 2 +- qemu-io-cmds.c | 2 +- qemu-storage-daemon.c | 4 +- qom/object_interfaces.c | 31 +++++ qom/qom-qmp-cmds.c | 24 +--- tests/test-block-iothread.c | 9 +- tests/qemu-iotests/iotests.py | 5 +- hw/block/Makefile.objs | 2 +- hw/block/trace-events | 4 + tests/qemu-iotests/244 | 10 +- tests/qemu-iotests/244.out | 9 +- tests/qemu-iotests/274 | 155 +++++++++++++++++++++++ tests/qemu-iotests/274.out | 268 ++++++++++++++++++++++++++++++++++++++++ tests/qemu-iotests/group | 1 + 49 files changed, 951 insertions(+), 96 deletions(-) create mode 100755 tests/qemu-iotests/274 create mode 100644 tests/qemu-iotests/274.out From: Alberto Garcia <berto@igalia.com> Backing files and raw external data files are mutually exclusive. The documentation of the raw external data bit (in autoclear_features) already indicates that, but we should also mention it on the other side. Suggested-by: Eric Blake <eblake@redhat.com> Signed-off-by: Alberto Garcia <berto@igalia.com> Message-Id: <20200410121816.8334-1-berto@igalia.com> Reviewed-by: Eric Blake <eblake@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- docs/interop/qcow2.txt | 3 +++ 1 file changed, 3 insertions(+) diff --git a/docs/interop/qcow2.txt b/docs/interop/qcow2.txt index XXXXXXX..XXXXXXX 100644 --- a/docs/interop/qcow2.txt +++ b/docs/interop/qcow2.txt @@ -XXX,XX +XXX,XX @@ The first cluster of a qcow2 image contains the file header: is stored (NB: The string is not null terminated). 0 if the image doesn't have a backing file. + Note: backing files are incompatible with raw external data + files (auto-clear feature bit 1). + 16 - 19: backing_file_size Length of the backing file name in bytes. Must not be longer than 1023 bytes. Undefined if the image doesn't have -- 2.25.3 From: Paolo Bonzini <pbonzini@redhat.com> Test 244 checks the expected behavior of qcow2 external data files with respect to zero and discarded clusters. Filesystems however are free to ignore discard requests, and this seems to be the case for overlayfs. Relax the tests to skip checks on the external data file for discarded areas, which implies not using qemu-img compare in the data_file_raw=on case. This fixes docker tests on RHEL8. Cc: Kevin Wolf <kwolf@redhat.com> Cc: qemu-block@nongnu.org Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Message-Id: <20200409191006.24429-1-pbonzini@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- tests/qemu-iotests/244 | 10 ++++++++-- tests/qemu-iotests/244.out | 9 ++++++--- 2 files changed, 14 insertions(+), 5 deletions(-) diff --git a/tests/qemu-iotests/244 b/tests/qemu-iotests/244 index XXXXXXX..XXXXXXX 100755 --- a/tests/qemu-iotests/244 +++ b/tests/qemu-iotests/244 @@ -XXX,XX +XXX,XX @@ $QEMU_IO -c 'read -P 0 0 1M' \ echo $QEMU_IO -c 'read -P 0 0 1M' \ -c 'read -P 0x11 1M 1M' \ - -c 'read -P 0 2M 2M' \ -c 'read -P 0x11 4M 1M' \ -c 'read -P 0 5M 1M' \ -f raw "$TEST_IMG.data" | @@ -XXX,XX +XXX,XX @@ $QEMU_IO -c 'read -P 0 0 1M' \ -f $IMGFMT "$TEST_IMG" | _filter_qemu_io +# Discarded clusters are only marked as such in the qcow2 metadata, but +# they can contain stale data in the external data file. Instead, zero +# clusters must be zeroed in the external data file too. echo -$QEMU_IMG compare "$TEST_IMG" "$TEST_IMG.data" +$QEMU_IO -c 'read -P 0 0 1M' \ + -c 'read -P 0x11 1M 1M' \ + -c 'read -P 0 3M 3M' \ + -f raw "$TEST_IMG".data | + _filter_qemu_io echo -n "qcow2 file size after I/O: " du -b $TEST_IMG | cut -f1 diff --git a/tests/qemu-iotests/244.out b/tests/qemu-iotests/244.out index XXXXXXX..XXXXXXX 100644 --- a/tests/qemu-iotests/244.out +++ b/tests/qemu-iotests/244.out @@ -XXX,XX +XXX,XX @@ read 1048576/1048576 bytes at offset 0 1 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) read 1048576/1048576 bytes at offset 1048576 1 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) -read 2097152/2097152 bytes at offset 2097152 -2 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) read 1048576/1048576 bytes at offset 4194304 1 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) read 1048576/1048576 bytes at offset 5242880 @@ -XXX,XX +XXX,XX @@ read 1048576/1048576 bytes at offset 1048576 read 4194304/4194304 bytes at offset 2097152 4 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) -Images are identical. +read 1048576/1048576 bytes at offset 0 +1 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) +read 1048576/1048576 bytes at offset 1048576 +1 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) +read 3145728/3145728 bytes at offset 3145728 +3 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) qcow2 file size after I/O: 327680 === bdrv_co_block_status test for file and offset=0 === -- 2.25.3 This adds a new BdrvRequestFlags parameter to the .bdrv_co_truncate() driver callbacks, and a supported_truncate_flags field in BlockDriverState that allows drivers to advertise support for request flags in the context of truncate. For now, we always pass 0 and no drivers declare support for any flag. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com> Reviewed-by: Alberto Garcia <berto@igalia.com> Reviewed-by: Max Reitz <mreitz@redhat.com> Message-Id: <20200424125448.63318-2-kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block_int.h | 10 +++++++++- block/crypto.c | 3 ++- block/file-posix.c | 2 +- block/file-win32.c | 2 +- block/gluster.c | 1 + block/io.c | 8 +++++++- block/iscsi.c | 2 +- block/nfs.c | 3 ++- block/qcow2.c | 2 +- block/qed.c | 1 + block/raw-format.c | 2 +- block/rbd.c | 1 + block/sheepdog.c | 4 ++-- block/ssh.c | 2 +- tests/test-block-iothread.c | 3 ++- 15 files changed, 33 insertions(+), 13 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 { */ int coroutine_fn (*bdrv_co_truncate)(BlockDriverState *bs, int64_t offset, bool exact, PreallocMode prealloc, - Error **errp); + BdrvRequestFlags flags, Error **errp); int64_t (*bdrv_getlength)(BlockDriverState *bs); bool has_variable_length; @@ -XXX,XX +XXX,XX @@ struct BlockDriverState { /* Flags honored during pwrite_zeroes (so far: BDRV_REQ_FUA, * BDRV_REQ_MAY_UNMAP, BDRV_REQ_WRITE_UNCHANGED) */ unsigned int supported_zero_flags; + /* + * Flags honoured during truncate (so far: BDRV_REQ_ZERO_WRITE). + * + * If BDRV_REQ_ZERO_WRITE is given, the truncate operation must make sure + * that any added space reads as all zeros. If this can't be guaranteed, + * the operation must fail. + */ + unsigned int supported_truncate_flags; /* the following member gives a name to every node on the bs graph. */ char node_name[32]; diff --git a/block/crypto.c b/block/crypto.c index XXXXXXX..XXXXXXX 100644 --- a/block/crypto.c +++ b/block/crypto.c @@ -XXX,XX +XXX,XX @@ static int block_crypto_co_create_generic(BlockDriverState *bs, static int coroutine_fn block_crypto_co_truncate(BlockDriverState *bs, int64_t offset, bool exact, - PreallocMode prealloc, Error **errp) + PreallocMode prealloc, BdrvRequestFlags flags, + Error **errp) { BlockCrypto *crypto = bs->opaque; uint64_t payload_offset = diff --git a/block/file-posix.c b/block/file-posix.c index XXXXXXX..XXXXXXX 100644 --- a/block/file-posix.c +++ b/block/file-posix.c @@ -XXX,XX +XXX,XX @@ raw_regular_truncate(BlockDriverState *bs, int fd, int64_t offset, static int coroutine_fn raw_co_truncate(BlockDriverState *bs, int64_t offset, bool exact, PreallocMode prealloc, - Error **errp) + BdrvRequestFlags flags, Error **errp) { BDRVRawState *s = bs->opaque; struct stat st; 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 void raw_close(BlockDriverState *bs) static int coroutine_fn raw_co_truncate(BlockDriverState *bs, int64_t offset, bool exact, PreallocMode prealloc, - Error **errp) + BdrvRequestFlags flags, Error **errp) { BDRVRawState *s = bs->opaque; LONG low, high; diff --git a/block/gluster.c b/block/gluster.c index XXXXXXX..XXXXXXX 100644 --- a/block/gluster.c +++ b/block/gluster.c @@ -XXX,XX +XXX,XX @@ static coroutine_fn int qemu_gluster_co_truncate(BlockDriverState *bs, int64_t offset, bool exact, PreallocMode prealloc, + BdrvRequestFlags flags, Error **errp) { BDRVGlusterState *s = bs->opaque; diff --git a/block/io.c b/block/io.c index XXXXXXX..XXXXXXX 100644 --- a/block/io.c +++ b/block/io.c @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact, BlockDriverState *bs = child->bs; BlockDriver *drv = bs->drv; BdrvTrackedRequest req; + BdrvRequestFlags flags = 0; int64_t old_size, new_bytes; int ret; @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact, } if (drv->bdrv_co_truncate) { - ret = drv->bdrv_co_truncate(bs, offset, exact, prealloc, errp); + if (flags & ~bs->supported_truncate_flags) { + error_setg(errp, "Block driver does not support requested flags"); + ret = -ENOTSUP; + goto out; + } + ret = drv->bdrv_co_truncate(bs, offset, exact, prealloc, flags, errp); } else if (bs->file && drv->is_filter) { ret = bdrv_co_truncate(bs->file, offset, exact, prealloc, errp); } else { diff --git a/block/iscsi.c b/block/iscsi.c index XXXXXXX..XXXXXXX 100644 --- a/block/iscsi.c +++ b/block/iscsi.c @@ -XXX,XX +XXX,XX @@ static void iscsi_reopen_commit(BDRVReopenState *reopen_state) static int coroutine_fn iscsi_co_truncate(BlockDriverState *bs, int64_t offset, bool exact, PreallocMode prealloc, - Error **errp) + BdrvRequestFlags flags, Error **errp) { IscsiLun *iscsilun = bs->opaque; int64_t cur_length; 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 int64_t nfs_get_allocated_file_size(BlockDriverState *bs) static int coroutine_fn nfs_file_co_truncate(BlockDriverState *bs, int64_t offset, bool exact, - PreallocMode prealloc, Error **errp) + PreallocMode prealloc, BdrvRequestFlags flags, + Error **errp) { NFSClient *client = bs->opaque; int ret; diff --git a/block/qcow2.c b/block/qcow2.c index XXXXXXX..XXXXXXX 100644 --- a/block/qcow2.c +++ b/block/qcow2.c @@ -XXX,XX +XXX,XX @@ fail: static int coroutine_fn qcow2_co_truncate(BlockDriverState *bs, int64_t offset, bool exact, PreallocMode prealloc, - Error **errp) + BdrvRequestFlags flags, Error **errp) { BDRVQcow2State *s = bs->opaque; uint64_t old_length; diff --git a/block/qed.c b/block/qed.c index XXXXXXX..XXXXXXX 100644 --- a/block/qed.c +++ b/block/qed.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_qed_co_truncate(BlockDriverState *bs, int64_t offset, bool exact, PreallocMode prealloc, + BdrvRequestFlags flags, Error **errp) { BDRVQEDState *s = bs->opaque; diff --git a/block/raw-format.c b/block/raw-format.c index XXXXXXX..XXXXXXX 100644 --- a/block/raw-format.c +++ b/block/raw-format.c @@ -XXX,XX +XXX,XX @@ static void raw_refresh_limits(BlockDriverState *bs, Error **errp) static int coroutine_fn raw_co_truncate(BlockDriverState *bs, int64_t offset, bool exact, PreallocMode prealloc, - Error **errp) + BdrvRequestFlags flags, Error **errp) { BDRVRawState *s = bs->opaque; diff --git a/block/rbd.c b/block/rbd.c index XXXXXXX..XXXXXXX 100644 --- a/block/rbd.c +++ b/block/rbd.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qemu_rbd_co_truncate(BlockDriverState *bs, int64_t offset, bool exact, PreallocMode prealloc, + BdrvRequestFlags flags, Error **errp) { int r; 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 @@ static int64_t sd_getlength(BlockDriverState *bs) static int coroutine_fn sd_co_truncate(BlockDriverState *bs, int64_t offset, bool exact, PreallocMode prealloc, - Error **errp) + BdrvRequestFlags flags, Error **errp) { BDRVSheepdogState *s = bs->opaque; int ret, fd; @@ -XXX,XX +XXX,XX @@ static coroutine_fn int sd_co_writev(BlockDriverState *bs, int64_t sector_num, assert(!flags); if (offset > s->inode.vdi_size) { - ret = sd_co_truncate(bs, offset, false, PREALLOC_MODE_OFF, NULL); + ret = sd_co_truncate(bs, offset, false, PREALLOC_MODE_OFF, 0, NULL); if (ret < 0) { return ret; } 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 int64_t ssh_getlength(BlockDriverState *bs) static int coroutine_fn ssh_co_truncate(BlockDriverState *bs, int64_t offset, bool exact, PreallocMode prealloc, - Error **errp) + BdrvRequestFlags flags, Error **errp) { BDRVSSHState *s = bs->opaque; diff --git a/tests/test-block-iothread.c b/tests/test-block-iothread.c index XXXXXXX..XXXXXXX 100644 --- a/tests/test-block-iothread.c +++ b/tests/test-block-iothread.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_test_co_pdiscard(BlockDriverState *bs, static int coroutine_fn bdrv_test_co_truncate(BlockDriverState *bs, int64_t offset, bool exact, - PreallocMode prealloc, Error **errp) + PreallocMode prealloc, BdrvRequestFlags flags, + Error **errp) { return 0; } -- 2.25.3 Now that block drivers can support flags for .bdrv_co_truncate, expose the parameter in the node level interfaces bdrv_co_truncate() and bdrv_truncate(). Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com> Reviewed-by: Alberto Garcia <berto@igalia.com> Reviewed-by: Max Reitz <mreitz@redhat.com> Message-Id: <20200424125448.63318-3-kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block.h | 5 +++-- block/block-backend.c | 2 +- block/crypto.c | 2 +- block/io.c | 12 +++++++----- block/parallels.c | 6 +++--- block/qcow.c | 4 ++-- block/qcow2-refcount.c | 2 +- block/qcow2.c | 15 +++++++++------ block/raw-format.c | 2 +- block/vhdx-log.c | 2 +- block/vhdx.c | 2 +- block/vmdk.c | 2 +- tests/test-block-iothread.c | 6 +++--- 13 files changed, 34 insertions(+), 28 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 @@ BlockDriverState *bdrv_find_backing_image(BlockDriverState *bs, void bdrv_refresh_filename(BlockDriverState *bs); int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact, - PreallocMode prealloc, Error **errp); + PreallocMode prealloc, BdrvRequestFlags flags, + Error **errp); int bdrv_truncate(BdrvChild *child, int64_t offset, bool exact, - PreallocMode prealloc, Error **errp); + PreallocMode prealloc, BdrvRequestFlags flags, Error **errp); int64_t bdrv_nb_sectors(BlockDriverState *bs); int64_t bdrv_getlength(BlockDriverState *bs); diff --git a/block/block-backend.c b/block/block-backend.c index XXXXXXX..XXXXXXX 100644 --- a/block/block-backend.c +++ b/block/block-backend.c @@ -XXX,XX +XXX,XX @@ int blk_truncate(BlockBackend *blk, int64_t offset, bool exact, return -ENOMEDIUM; } - return bdrv_truncate(blk->root, offset, exact, prealloc, errp); + return bdrv_truncate(blk->root, offset, exact, prealloc, 0, errp); } int blk_save_vmstate(BlockBackend *blk, const uint8_t *buf, diff --git a/block/crypto.c b/block/crypto.c index XXXXXXX..XXXXXXX 100644 --- a/block/crypto.c +++ b/block/crypto.c @@ -XXX,XX +XXX,XX @@ block_crypto_co_truncate(BlockDriverState *bs, int64_t offset, bool exact, offset += payload_offset; - return bdrv_co_truncate(bs->file, offset, exact, prealloc, errp); + return bdrv_co_truncate(bs->file, offset, exact, prealloc, 0, errp); } static void block_crypto_close(BlockDriverState *bs) diff --git a/block/io.c b/block/io.c index XXXXXXX..XXXXXXX 100644 --- a/block/io.c +++ b/block/io.c @@ -XXX,XX +XXX,XX @@ static void bdrv_parent_cb_resize(BlockDriverState *bs) * 'offset' bytes in length. */ int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact, - PreallocMode prealloc, Error **errp) + PreallocMode prealloc, BdrvRequestFlags flags, + Error **errp) { BlockDriverState *bs = child->bs; BlockDriver *drv = bs->drv; BdrvTrackedRequest req; - BdrvRequestFlags flags = 0; int64_t old_size, new_bytes; int ret; @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact, } ret = drv->bdrv_co_truncate(bs, offset, exact, prealloc, flags, errp); } else if (bs->file && drv->is_filter) { - ret = bdrv_co_truncate(bs->file, offset, exact, prealloc, errp); + ret = bdrv_co_truncate(bs->file, offset, exact, prealloc, flags, errp); } else { error_setg(errp, "Image format driver does not support resize"); ret = -ENOTSUP; @@ -XXX,XX +XXX,XX @@ typedef struct TruncateCo { int64_t offset; bool exact; PreallocMode prealloc; + BdrvRequestFlags flags; Error **errp; int ret; } TruncateCo; @@ -XXX,XX +XXX,XX @@ static void coroutine_fn bdrv_truncate_co_entry(void *opaque) { TruncateCo *tco = opaque; tco->ret = bdrv_co_truncate(tco->child, tco->offset, tco->exact, - tco->prealloc, tco->errp); + tco->prealloc, tco->flags, tco->errp); aio_wait_kick(); } int bdrv_truncate(BdrvChild *child, int64_t offset, bool exact, - PreallocMode prealloc, Error **errp) + PreallocMode prealloc, BdrvRequestFlags flags, Error **errp) { Coroutine *co; TruncateCo tco = { @@ -XXX,XX +XXX,XX @@ int bdrv_truncate(BdrvChild *child, int64_t offset, bool exact, .offset = offset, .exact = exact, .prealloc = prealloc, + .flags = flags, .errp = errp, .ret = NOT_DONE, }; diff --git a/block/parallels.c b/block/parallels.c index XXXXXXX..XXXXXXX 100644 --- a/block/parallels.c +++ b/block/parallels.c @@ -XXX,XX +XXX,XX @@ static int64_t allocate_clusters(BlockDriverState *bs, int64_t sector_num, } else { ret = bdrv_truncate(bs->file, (s->data_end + space) << BDRV_SECTOR_BITS, - false, PREALLOC_MODE_OFF, NULL); + false, PREALLOC_MODE_OFF, 0, NULL); } if (ret < 0) { return ret; @@ -XXX,XX +XXX,XX @@ static int coroutine_fn parallels_co_check(BlockDriverState *bs, * That means we have to pass exact=true. */ ret = bdrv_truncate(bs->file, res->image_end_offset, true, - PREALLOC_MODE_OFF, &local_err); + PREALLOC_MODE_OFF, 0, &local_err); if (ret < 0) { error_report_err(local_err); res->check_errors++; @@ -XXX,XX +XXX,XX @@ static void parallels_close(BlockDriverState *bs) /* errors are ignored, so we might as well pass exact=true */ bdrv_truncate(bs->file, s->data_end << BDRV_SECTOR_BITS, true, - PREALLOC_MODE_OFF, NULL); + PREALLOC_MODE_OFF, 0, NULL); } g_free(s->bat_dirty_bmap); diff --git a/block/qcow.c b/block/qcow.c index XXXXXXX..XXXXXXX 100644 --- a/block/qcow.c +++ b/block/qcow.c @@ -XXX,XX +XXX,XX @@ static int get_cluster_offset(BlockDriverState *bs, return -E2BIG; } ret = bdrv_truncate(bs->file, cluster_offset + s->cluster_size, - false, PREALLOC_MODE_OFF, NULL); + false, PREALLOC_MODE_OFF, 0, NULL); if (ret < 0) { return ret; } @@ -XXX,XX +XXX,XX @@ static int qcow_make_empty(BlockDriverState *bs) l1_length) < 0) return -1; ret = bdrv_truncate(bs->file, s->l1_table_offset + l1_length, false, - PREALLOC_MODE_OFF, NULL); + PREALLOC_MODE_OFF, 0, NULL); if (ret < 0) return ret; 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 check_refblocks(BlockDriverState *bs, BdrvCheckResult *res, } ret = bdrv_truncate(bs->file, offset + s->cluster_size, false, - PREALLOC_MODE_OFF, &local_err); + PREALLOC_MODE_OFF, 0, &local_err); if (ret < 0) { error_report_err(local_err); goto resize_fail; 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 preallocate_co(BlockDriverState *bs, uint64_t offset, mode = PREALLOC_MODE_OFF; } ret = bdrv_co_truncate(s->data_file, host_offset + cur_bytes, false, - mode, errp); + mode, 0, errp); if (ret < 0) { return ret; } @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qcow2_co_truncate(BlockDriverState *bs, int64_t offset, * always fulfilled, so there is no need to pass it on.) */ bdrv_co_truncate(bs->file, (last_cluster + 1) * s->cluster_size, - false, PREALLOC_MODE_OFF, &local_err); + false, PREALLOC_MODE_OFF, 0, &local_err); if (local_err) { warn_reportf_err(local_err, "Failed to truncate the tail of the image: "); @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qcow2_co_truncate(BlockDriverState *bs, int64_t offset, * file should be resized to the exact target size, too, * so we pass @exact here. */ - ret = bdrv_co_truncate(s->data_file, offset, exact, prealloc, errp); + ret = bdrv_co_truncate(s->data_file, offset, exact, prealloc, 0, + errp); if (ret < 0) { goto fail; } @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qcow2_co_truncate(BlockDriverState *bs, int64_t offset, new_file_size = allocation_start + nb_new_data_clusters * s->cluster_size; /* Image file grows, so @exact does not matter */ - ret = bdrv_co_truncate(bs->file, new_file_size, false, prealloc, errp); + ret = bdrv_co_truncate(bs->file, new_file_size, false, prealloc, 0, + errp); if (ret < 0) { error_prepend(errp, "Failed to resize underlying file: "); qcow2_free_clusters(bs, allocation_start, @@ -XXX,XX +XXX,XX @@ qcow2_co_pwritev_compressed_part(BlockDriverState *bs, if (len < 0) { return len; } - return bdrv_co_truncate(bs->file, len, false, PREALLOC_MODE_OFF, NULL); + return bdrv_co_truncate(bs->file, len, false, PREALLOC_MODE_OFF, 0, + NULL); } if (offset_into_cluster(s, offset)) { @@ -XXX,XX +XXX,XX @@ static int make_completely_empty(BlockDriverState *bs) } ret = bdrv_truncate(bs->file, (3 + l1_clusters) * s->cluster_size, false, - PREALLOC_MODE_OFF, &local_err); + PREALLOC_MODE_OFF, 0, &local_err); if (ret < 0) { error_report_err(local_err); goto fail; diff --git a/block/raw-format.c b/block/raw-format.c index XXXXXXX..XXXXXXX 100644 --- a/block/raw-format.c +++ b/block/raw-format.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn raw_co_truncate(BlockDriverState *bs, int64_t offset, s->size = offset; offset += s->offset; - return bdrv_co_truncate(bs->file, offset, exact, prealloc, errp); + return bdrv_co_truncate(bs->file, offset, exact, prealloc, 0, errp); } static void raw_eject(BlockDriverState *bs, bool eject_flag) diff --git a/block/vhdx-log.c b/block/vhdx-log.c index XXXXXXX..XXXXXXX 100644 --- a/block/vhdx-log.c +++ b/block/vhdx-log.c @@ -XXX,XX +XXX,XX @@ static int vhdx_log_flush(BlockDriverState *bs, BDRVVHDXState *s, goto exit; } ret = bdrv_truncate(bs->file, new_file_size, false, - PREALLOC_MODE_OFF, NULL); + PREALLOC_MODE_OFF, 0, NULL); if (ret < 0) { goto exit; } diff --git a/block/vhdx.c b/block/vhdx.c index XXXXXXX..XXXXXXX 100644 --- a/block/vhdx.c +++ b/block/vhdx.c @@ -XXX,XX +XXX,XX @@ static int vhdx_allocate_block(BlockDriverState *bs, BDRVVHDXState *s, } return bdrv_truncate(bs->file, *new_offset + s->block_size, false, - PREALLOC_MODE_OFF, NULL); + PREALLOC_MODE_OFF, 0, NULL); } /* diff --git a/block/vmdk.c b/block/vmdk.c index XXXXXXX..XXXXXXX 100644 --- a/block/vmdk.c +++ b/block/vmdk.c @@ -XXX,XX +XXX,XX @@ vmdk_co_pwritev_compressed(BlockDriverState *bs, uint64_t offset, } length = QEMU_ALIGN_UP(length, BDRV_SECTOR_SIZE); ret = bdrv_truncate(s->extents[i].file, length, false, - PREALLOC_MODE_OFF, NULL); + PREALLOC_MODE_OFF, 0, NULL); if (ret < 0) { return ret; } diff --git a/tests/test-block-iothread.c b/tests/test-block-iothread.c index XXXXXXX..XXXXXXX 100644 --- a/tests/test-block-iothread.c +++ b/tests/test-block-iothread.c @@ -XXX,XX +XXX,XX @@ static void test_sync_op_truncate(BdrvChild *c) int ret; /* Normal success path */ - ret = bdrv_truncate(c, 65536, false, PREALLOC_MODE_OFF, NULL); + ret = bdrv_truncate(c, 65536, false, PREALLOC_MODE_OFF, 0, NULL); g_assert_cmpint(ret, ==, 0); /* Early error: Negative offset */ - ret = bdrv_truncate(c, -2, false, PREALLOC_MODE_OFF, NULL); + ret = bdrv_truncate(c, -2, false, PREALLOC_MODE_OFF, 0, NULL); g_assert_cmpint(ret, ==, -EINVAL); /* Error: Read-only image */ c->bs->read_only = true; c->bs->open_flags &= ~BDRV_O_RDWR; - ret = bdrv_truncate(c, 65536, false, PREALLOC_MODE_OFF, NULL); + ret = bdrv_truncate(c, 65536, false, PREALLOC_MODE_OFF, 0, NULL); g_assert_cmpint(ret, ==, -EACCES); c->bs->read_only = false; -- 2.25.3 Now that node level interface bdrv_truncate() supports passing request flags to the block driver, expose this on the BlockBackend level, too. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com> Reviewed-by: Alberto Garcia <berto@igalia.com> Reviewed-by: Max Reitz <mreitz@redhat.com> Message-Id: <20200424125448.63318-4-kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/sysemu/block-backend.h | 2 +- block.c | 3 ++- block/block-backend.c | 4 ++-- block/commit.c | 4 ++-- block/crypto.c | 2 +- block/mirror.c | 2 +- block/qcow2.c | 4 ++-- block/qed.c | 2 +- block/vdi.c | 2 +- block/vhdx.c | 4 ++-- block/vmdk.c | 6 +++--- block/vpc.c | 2 +- blockdev.c | 2 +- qemu-img.c | 2 +- qemu-io-cmds.c | 2 +- 15 files changed, 22 insertions(+), 21 deletions(-) diff --git a/include/sysemu/block-backend.h b/include/sysemu/block-backend.h index XXXXXXX..XXXXXXX 100644 --- a/include/sysemu/block-backend.h +++ b/include/sysemu/block-backend.h @@ -XXX,XX +XXX,XX @@ int coroutine_fn blk_co_pwrite_zeroes(BlockBackend *blk, int64_t offset, int blk_pwrite_compressed(BlockBackend *blk, int64_t offset, const void *buf, int bytes); int blk_truncate(BlockBackend *blk, int64_t offset, bool exact, - PreallocMode prealloc, Error **errp); + PreallocMode prealloc, BdrvRequestFlags flags, Error **errp); int blk_pdiscard(BlockBackend *blk, int64_t offset, int bytes); int blk_save_vmstate(BlockBackend *blk, const uint8_t *buf, int64_t pos, int size); diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ static int64_t create_file_fallback_truncate(BlockBackend *blk, int64_t size; int ret; - ret = blk_truncate(blk, minimum_size, false, PREALLOC_MODE_OFF, &local_err); + ret = blk_truncate(blk, minimum_size, false, PREALLOC_MODE_OFF, 0, + &local_err); if (ret < 0 && ret != -ENOTSUP) { error_propagate(errp, local_err); return ret; diff --git a/block/block-backend.c b/block/block-backend.c index XXXXXXX..XXXXXXX 100644 --- a/block/block-backend.c +++ b/block/block-backend.c @@ -XXX,XX +XXX,XX @@ int blk_pwrite_compressed(BlockBackend *blk, int64_t offset, const void *buf, } int blk_truncate(BlockBackend *blk, int64_t offset, bool exact, - PreallocMode prealloc, Error **errp) + PreallocMode prealloc, BdrvRequestFlags flags, Error **errp) { if (!blk_is_available(blk)) { error_setg(errp, "No medium inserted"); return -ENOMEDIUM; } - return bdrv_truncate(blk->root, offset, exact, prealloc, 0, errp); + return bdrv_truncate(blk->root, offset, exact, prealloc, flags, errp); } int blk_save_vmstate(BlockBackend *blk, const uint8_t *buf, diff --git a/block/commit.c b/block/commit.c index XXXXXXX..XXXXXXX 100644 --- a/block/commit.c +++ b/block/commit.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn commit_run(Job *job, Error **errp) } if (base_len < len) { - ret = blk_truncate(s->base, len, false, PREALLOC_MODE_OFF, NULL); + ret = blk_truncate(s->base, len, false, PREALLOC_MODE_OFF, 0, NULL); if (ret) { goto out; } @@ -XXX,XX +XXX,XX @@ int bdrv_commit(BlockDriverState *bs) * grow the backing file image if possible. If not possible, * we must return an error */ if (length > backing_length) { - ret = blk_truncate(backing, length, false, PREALLOC_MODE_OFF, + ret = blk_truncate(backing, length, false, PREALLOC_MODE_OFF, 0, &local_err); if (ret < 0) { error_report_err(local_err); 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 ssize_t block_crypto_init_func(QCryptoBlock *block, * which will be used by the crypto header */ return blk_truncate(data->blk, data->size + headerlen, false, - data->prealloc, errp); + data->prealloc, 0, errp); } diff --git a/block/mirror.c b/block/mirror.c index XXXXXXX..XXXXXXX 100644 --- a/block/mirror.c +++ b/block/mirror.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn mirror_run(Job *job, Error **errp) if (s->bdev_length > base_length) { ret = blk_truncate(s->target, s->bdev_length, false, - PREALLOC_MODE_OFF, NULL); + PREALLOC_MODE_OFF, 0, NULL); if (ret < 0) { goto immediate_exit; } 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) /* Okay, now that we have a valid image, let's give it the right size */ ret = blk_truncate(blk, qcow2_opts->size, false, qcow2_opts->preallocation, - errp); + 0, errp); if (ret < 0) { error_prepend(errp, "Could not resize image: "); goto out; @@ -XXX,XX +XXX,XX @@ static int qcow2_amend_options(BlockDriverState *bs, QemuOpts *opts, * Amending image options should ensure that the image has * exactly the given new values, so pass exact=true here. */ - ret = blk_truncate(blk, new_size, true, PREALLOC_MODE_OFF, errp); + ret = blk_truncate(blk, new_size, true, PREALLOC_MODE_OFF, 0, errp); blk_unref(blk); if (ret < 0) { return ret; diff --git a/block/qed.c b/block/qed.c index XXXXXXX..XXXXXXX 100644 --- a/block/qed.c +++ b/block/qed.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn bdrv_qed_co_create(BlockdevCreateOptions *opts, * The QED format associates file length with allocation status, * so a new file (which is empty) must have a length of 0. */ - ret = blk_truncate(blk, 0, true, PREALLOC_MODE_OFF, errp); + ret = blk_truncate(blk, 0, true, PREALLOC_MODE_OFF, 0, errp); if (ret < 0) { goto out; } diff --git a/block/vdi.c b/block/vdi.c index XXXXXXX..XXXXXXX 100644 --- a/block/vdi.c +++ b/block/vdi.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn vdi_co_do_create(BlockdevCreateOptions *create_options, if (image_type == VDI_TYPE_STATIC) { ret = blk_truncate(blk, offset + blocks * block_size, false, - PREALLOC_MODE_OFF, errp); + PREALLOC_MODE_OFF, 0, errp); if (ret < 0) { error_prepend(errp, "Failed to statically allocate file"); goto exit; diff --git a/block/vhdx.c b/block/vhdx.c index XXXXXXX..XXXXXXX 100644 --- a/block/vhdx.c +++ b/block/vhdx.c @@ -XXX,XX +XXX,XX @@ static int vhdx_create_bat(BlockBackend *blk, BDRVVHDXState *s, /* All zeroes, so we can just extend the file - the end of the BAT * is the furthest thing we have written yet */ ret = blk_truncate(blk, data_file_offset, false, PREALLOC_MODE_OFF, - errp); + 0, errp); if (ret < 0) { goto exit; } } else if (type == VHDX_TYPE_FIXED) { ret = blk_truncate(blk, data_file_offset + image_size, false, - PREALLOC_MODE_OFF, errp); + PREALLOC_MODE_OFF, 0, errp); if (ret < 0) { goto exit; } diff --git a/block/vmdk.c b/block/vmdk.c index XXXXXXX..XXXXXXX 100644 --- a/block/vmdk.c +++ b/block/vmdk.c @@ -XXX,XX +XXX,XX @@ static int vmdk_init_extent(BlockBackend *blk, int gd_buf_size; if (flat) { - ret = blk_truncate(blk, filesize, false, PREALLOC_MODE_OFF, errp); + ret = blk_truncate(blk, filesize, false, PREALLOC_MODE_OFF, 0, errp); goto exit; } magic = cpu_to_be32(VMDK4_MAGIC); @@ -XXX,XX +XXX,XX @@ static int vmdk_init_extent(BlockBackend *blk, } ret = blk_truncate(blk, le64_to_cpu(header.grain_offset) << 9, false, - PREALLOC_MODE_OFF, errp); + PREALLOC_MODE_OFF, 0, errp); if (ret < 0) { goto exit; } @@ -XXX,XX +XXX,XX @@ static int coroutine_fn vmdk_co_do_create(int64_t size, /* bdrv_pwrite write padding zeros to align to sector, we don't need that * for description file */ if (desc_offset == 0) { - ret = blk_truncate(blk, desc_len, false, PREALLOC_MODE_OFF, errp); + ret = blk_truncate(blk, desc_len, false, PREALLOC_MODE_OFF, 0, errp); if (ret < 0) { goto exit; } diff --git a/block/vpc.c b/block/vpc.c index XXXXXXX..XXXXXXX 100644 --- a/block/vpc.c +++ b/block/vpc.c @@ -XXX,XX +XXX,XX @@ static int create_fixed_disk(BlockBackend *blk, uint8_t *buf, /* Add footer to total size */ total_size += HEADER_SIZE; - ret = blk_truncate(blk, total_size, false, PREALLOC_MODE_OFF, errp); + ret = blk_truncate(blk, total_size, false, PREALLOC_MODE_OFF, 0, errp); if (ret < 0) { return ret; } diff --git a/blockdev.c b/blockdev.c index XXXXXXX..XXXXXXX 100644 --- a/blockdev.c +++ b/blockdev.c @@ -XXX,XX +XXX,XX @@ void qmp_block_resize(bool has_device, const char *device, } bdrv_drained_begin(bs); - ret = blk_truncate(blk, size, false, PREALLOC_MODE_OFF, errp); + ret = blk_truncate(blk, size, false, PREALLOC_MODE_OFF, 0, errp); bdrv_drained_end(bs); out: diff --git a/qemu-img.c b/qemu-img.c index XXXXXXX..XXXXXXX 100644 --- a/qemu-img.c +++ b/qemu-img.c @@ -XXX,XX +XXX,XX @@ static int img_resize(int argc, char **argv) * resizing, so pass @exact=true. It is of no use to report * success when the image has not actually been resized. */ - ret = blk_truncate(blk, total_size, true, prealloc, &err); + ret = blk_truncate(blk, total_size, true, prealloc, 0, &err); if (!ret) { qprintf(quiet, "Image resized.\n"); } else { 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 truncate_f(BlockBackend *blk, int argc, char **argv) * exact=true. It is better to err on the "emit more errors" side * than to be overly permissive. */ - ret = blk_truncate(blk, offset, true, PREALLOC_MODE_OFF, &local_err); + ret = blk_truncate(blk, offset, true, PREALLOC_MODE_OFF, 0, &local_err); if (ret < 0) { error_report_err(local_err); return ret; -- 2.25.3 If BDRV_REQ_ZERO_WRITE is set and we're extending the image, calling qcow2_cluster_zeroize() with flags=0 does the right thing: It doesn't undo any previous preallocation, but just adds the zero flag to all relevant L2 entries. If an external data file is in use, a write_zeroes request to the data file is made instead. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20200424125448.63318-5-kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/qcow2-cluster.c | 2 +- block/qcow2.c | 34 ++++++++++++++++++++++++++++++++++ 2 files changed, 35 insertions(+), 1 deletion(-) 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 @@ int qcow2_cluster_zeroize(BlockDriverState *bs, uint64_t offset, /* Caller must pass aligned values, except at image end */ assert(QEMU_IS_ALIGNED(offset, s->cluster_size)); assert(QEMU_IS_ALIGNED(end_offset, s->cluster_size) || - end_offset == bs->total_sectors << BDRV_SECTOR_BITS); + end_offset >= bs->total_sectors << BDRV_SECTOR_BITS); /* The zero flag is only supported by version 3 and newer */ if (s->qcow_version < 3) { 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_do_open(BlockDriverState *bs, QDict *options, bs->supported_zero_flags = header.version >= 3 ? BDRV_REQ_MAY_UNMAP | BDRV_REQ_NO_FALLBACK : 0; + bs->supported_truncate_flags = BDRV_REQ_ZERO_WRITE; /* Repair image if dirty */ if (!(flags & (BDRV_O_CHECK | BDRV_O_INACTIVE)) && !bs->read_only && @@ -XXX,XX +XXX,XX @@ static int coroutine_fn qcow2_co_truncate(BlockDriverState *bs, int64_t offset, g_assert_not_reached(); } + if ((flags & BDRV_REQ_ZERO_WRITE) && offset > old_length) { + uint64_t zero_start = QEMU_ALIGN_UP(old_length, s->cluster_size); + + /* + * Use zero clusters as much as we can. qcow2_cluster_zeroize() + * requires a cluster-aligned start. The end may be unaligned if it is + * at the end of the image (which it is here). + */ + ret = qcow2_cluster_zeroize(bs, zero_start, offset - zero_start, 0); + if (ret < 0) { + error_setg_errno(errp, -ret, "Failed to zero out new clusters"); + goto fail; + } + + /* Write explicit zeros for the unaligned head */ + if (zero_start > old_length) { + uint64_t len = zero_start - old_length; + uint8_t *buf = qemu_blockalign0(bs, len); + QEMUIOVector qiov; + qemu_iovec_init_buf(&qiov, buf, len); + + qemu_co_mutex_unlock(&s->lock); + ret = qcow2_co_pwritev_part(bs, old_length, len, &qiov, 0, 0); + qemu_co_mutex_lock(&s->lock); + + qemu_vfree(buf); + if (ret < 0) { + error_setg_errno(errp, -ret, "Failed to zero out the new area"); + goto fail; + } + } + } + if (prealloc != PREALLOC_MODE_OFF) { /* Flush metadata before actually changing the image size */ ret = qcow2_write_caches(bs); -- 2.25.3 The raw format driver can simply forward the flag and let its bs->file child take care of actually providing the zeros. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com> Message-Id: <20200424125448.63318-6-kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/raw-format.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/block/raw-format.c b/block/raw-format.c index XXXXXXX..XXXXXXX 100644 --- a/block/raw-format.c +++ b/block/raw-format.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn raw_co_truncate(BlockDriverState *bs, int64_t offset, s->size = offset; offset += s->offset; - return bdrv_co_truncate(bs->file, offset, exact, prealloc, 0, errp); + return bdrv_co_truncate(bs->file, offset, exact, prealloc, flags, errp); } static void raw_eject(BlockDriverState *bs, bool eject_flag) @@ -XXX,XX +XXX,XX @@ static int raw_open(BlockDriverState *bs, QDict *options, int flags, bs->supported_zero_flags = BDRV_REQ_WRITE_UNCHANGED | ((BDRV_REQ_FUA | BDRV_REQ_MAY_UNMAP | BDRV_REQ_NO_FALLBACK) & bs->file->bs->supported_zero_flags); + bs->supported_truncate_flags = bs->file->bs->supported_truncate_flags & + BDRV_REQ_ZERO_WRITE; if (bs->probed && !bdrv_is_read_only(bs)) { bdrv_refresh_filename(bs->file->bs); -- 2.25.3 For regular files, we always get BDRV_REQ_ZERO_WRITE behaviour from the OS, so we can advertise the flag and just ignore it. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com> Reviewed-by: Alberto Garcia <berto@igalia.com> Reviewed-by: Max Reitz <mreitz@redhat.com> Message-Id: <20200424125448.63318-7-kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/file-posix.c | 4 ++++ 1 file changed, 4 insertions(+) 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_open_common(BlockDriverState *bs, QDict *options, #endif bs->supported_zero_flags = BDRV_REQ_MAY_UNMAP | BDRV_REQ_NO_FALLBACK; + if (S_ISREG(st.st_mode)) { + /* When extending regular files, we get zeros from the OS */ + bs->supported_truncate_flags = BDRV_REQ_ZERO_WRITE; + } ret = 0; fail: if (filename && (bdrv_flags & BDRV_O_TEMPORARY)) { -- 2.25.3 When extending the size of an image that has a backing file larger than its old size, make sure that the backing file data doesn't become visible in the guest, but the added area is properly zeroed out. Consider the following scenario where the overlay is shorter than its backing file: base.qcow2: AAAAAAAA overlay.qcow2: BBBB When resizing (extending) overlay.qcow2, the new blocks should not stay unallocated and make the additional As from base.qcow2 visible like before this patch, but zeros should be read. A similar case happens with the various variants of a commit job when an intermediate file is short (- for unallocated): base.qcow2: A-A-AAAA mid.qcow2: BB-B top.qcow2: C--C--C- After commit top.qcow2 to mid.qcow2, the following happens: mid.qcow2: CB-C00C0 (correct result) mid.qcow2: CB-C--C- (before this fix) Without the fix, blocks that previously read as zeros on top.qcow2 suddenly turn into A. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com> Message-Id: <20200424125448.63318-8-kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/io.c | 25 +++++++++++++++++++++++++ 1 file changed, 25 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 @@ int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact, goto out; } + /* + * If the image has a backing file that is large enough that it would + * provide data for the new area, we cannot leave it unallocated because + * then the backing file content would become visible. Instead, zero-fill + * the new area. + * + * Note that if the image has a backing file, but was opened without the + * backing file, taking care of keeping things consistent with that backing + * file is the user's responsibility. + */ + if (new_bytes && bs->backing) { + int64_t backing_len; + + backing_len = bdrv_getlength(backing_bs(bs)); + if (backing_len < 0) { + ret = backing_len; + error_setg_errno(errp, -ret, "Could not get backing file size"); + goto out; + } + + if (backing_len > old_size) { + flags |= BDRV_REQ_ZERO_WRITE; + } + } + if (drv->bdrv_co_truncate) { if (flags & ~bs->supported_truncate_flags) { error_setg(errp, "Block driver does not support requested flags"); -- 2.25.3 We want to keep TEST_IMG for the full path of the main test image, but filter_testfiles() must be called for other test images before replacing other things like the image format because the test directory path could contain the format as a substring. Insert a filter_testfiles() call between both. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com> Message-Id: <20200424125448.63318-9-kwolf@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- tests/qemu-iotests/iotests.py | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/tests/qemu-iotests/iotests.py b/tests/qemu-iotests/iotests.py index XXXXXXX..XXXXXXX 100644 --- a/tests/qemu-iotests/iotests.py +++ b/tests/qemu-iotests/iotests.py @@ -XXX,XX +XXX,XX @@ def filter_img_info(output, filename): for line in output.split('\n'): if 'disk size' in line or 'actual-size' in line: continue - line = line.replace(filename, 'TEST_IMG') \ - .replace(imgfmt, 'IMGFMT') + line = line.replace(filename, 'TEST_IMG') + line = filter_testfiles(line) + line = line.replace(imgfmt, 'IMGFMT') line = re.sub('iters: [0-9]+', 'iters: XXX', line) line = re.sub('uuid: [-a-f0-9]+', 'uuid: XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX', line) line = re.sub('cid: [0-9]+', 'cid: XXXXXXXXXX', line) -- 2.25.3 Signed-off-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <20200424125448.63318-10-kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- tests/qemu-iotests/274 | 155 +++++++++++++++++++++ tests/qemu-iotests/274.out | 268 +++++++++++++++++++++++++++++++++++++ tests/qemu-iotests/group | 1 + 3 files changed, 424 insertions(+) create mode 100755 tests/qemu-iotests/274 create mode 100644 tests/qemu-iotests/274.out diff --git a/tests/qemu-iotests/274 b/tests/qemu-iotests/274 new file mode 100755 index XXXXXXX..XXXXXXX --- /dev/null +++ b/tests/qemu-iotests/274 @@ -XXX,XX +XXX,XX @@ +#!/usr/bin/env python3 +# +# Copyright (C) 2019 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: Kevin Wolf <kwolf@redhat.com> +# +# Some tests for short backing files and short overlays + +import iotests + +iotests.verify_image_format(supported_fmts=['qcow2']) +iotests.verify_platform(['linux']) + +size_short = 1 * 1024 * 1024 +size_long = 2 * 1024 * 1024 +size_diff = size_long - size_short + +def create_chain() -> None: + iotests.qemu_img_log('create', '-f', iotests.imgfmt, base, + str(size_long)) + iotests.qemu_img_log('create', '-f', iotests.imgfmt, '-b', base, mid, + str(size_short)) + iotests.qemu_img_log('create', '-f', iotests.imgfmt, '-b', mid, top, + str(size_long)) + + iotests.qemu_io_log('-c', 'write -P 1 0 %d' % size_long, base) + +def create_vm() -> iotests.VM: + vm = iotests.VM() + vm.add_blockdev('file,filename=%s,node-name=base-file' % base) + vm.add_blockdev('%s,file=base-file,node-name=base' % iotests.imgfmt) + vm.add_blockdev('file,filename=%s,node-name=mid-file' % mid) + vm.add_blockdev('%s,file=mid-file,node-name=mid,backing=base' + % iotests.imgfmt) + vm.add_drive(top, 'backing=mid,node-name=top') + return vm + +with iotests.FilePath('base') as base, \ + iotests.FilePath('mid') as mid, \ + iotests.FilePath('top') as top: + + iotests.log('== Commit tests ==') + + create_chain() + + iotests.log('=== Check visible data ===') + + iotests.qemu_io_log('-c', 'read -P 1 0 %d' % size_short, top) + iotests.qemu_io_log('-c', 'read -P 0 %d %d' % (size_short, size_diff), top) + + iotests.log('=== Checking allocation status ===') + + iotests.qemu_io_log('-c', 'alloc 0 %d' % size_short, + '-c', 'alloc %d %d' % (size_short, size_diff), + base) + + iotests.qemu_io_log('-c', 'alloc 0 %d' % size_short, + '-c', 'alloc %d %d' % (size_short, size_diff), + mid) + + iotests.qemu_io_log('-c', 'alloc 0 %d' % size_short, + '-c', 'alloc %d %d' % (size_short, size_diff), + top) + + iotests.log('=== Checking map ===') + + iotests.qemu_img_log('map', '--output=json', base) + iotests.qemu_img_log('map', '--output=human', base) + iotests.qemu_img_log('map', '--output=json', mid) + iotests.qemu_img_log('map', '--output=human', mid) + iotests.qemu_img_log('map', '--output=json', top) + iotests.qemu_img_log('map', '--output=human', top) + + iotests.log('=== Testing qemu-img commit (top -> mid) ===') + + iotests.qemu_img_log('commit', top) + iotests.img_info_log(mid) + iotests.qemu_io_log('-c', 'read -P 1 0 %d' % size_short, mid) + iotests.qemu_io_log('-c', 'read -P 0 %d %d' % (size_short, size_diff), mid) + + iotests.log('=== Testing HMP commit (top -> mid) ===') + + create_chain() + with create_vm() as vm: + vm.launch() + vm.qmp_log('human-monitor-command', command_line='commit drive0') + + iotests.img_info_log(mid) + iotests.qemu_io_log('-c', 'read -P 1 0 %d' % size_short, mid) + iotests.qemu_io_log('-c', 'read -P 0 %d %d' % (size_short, size_diff), mid) + + iotests.log('=== Testing QMP active commit (top -> mid) ===') + + create_chain() + with create_vm() as vm: + vm.launch() + vm.qmp_log('block-commit', device='top', base_node='mid', + job_id='job0', auto_dismiss=False) + vm.run_job('job0', wait=5) + + iotests.img_info_log(mid) + iotests.qemu_io_log('-c', 'read -P 1 0 %d' % size_short, mid) + iotests.qemu_io_log('-c', 'read -P 0 %d %d' % (size_short, size_diff), mid) + + + iotests.log('== Resize tests ==') + + # Use different sizes for different allocation modes: + # + # We want to have at least one test where 32 bit truncation in the size of + # the overlapping area becomes visible. This is covered by the + # prealloc='off' case (1G to 6G is an overlap of 5G). + # + # However, we can only do this for modes that don't preallocate data + # because otherwise we might run out of space on the test host. + # + # We also want to test some unaligned combinations. + for (prealloc, base_size, top_size_old, top_size_new, off) in [ + ('off', '6G', '1G', '8G', '5G'), + ('metadata', '32G', '30G', '33G', '31G'), + ('falloc', '10M', '5M', '15M', '9M'), + ('full', '16M', '8M', '12M', '11M'), + ('off', '384k', '253k', '512k', '253k'), + ('off', '400k', '256k', '512k', '336k'), + ('off', '512k', '256k', '500k', '436k')]: + + iotests.log('=== preallocation=%s ===' % prealloc) + iotests.qemu_img_log('create', '-f', iotests.imgfmt, base, base_size) + iotests.qemu_img_log('create', '-f', iotests.imgfmt, '-b', base, top, + top_size_old) + iotests.qemu_io_log('-c', 'write -P 1 %s 64k' % off, base) + + # After this, top_size_old to base_size should be allocated/zeroed. + # + # In theory, leaving base_size to top_size_new unallocated would be + # correct, but in practice, if we zero out anything, we zero out + # everything up to top_size_new. + iotests.qemu_img_log('resize', '-f', iotests.imgfmt, + '--preallocation', prealloc, top, top_size_new) + iotests.qemu_io_log('-c', 'read -P 0 %s 64k' % off, top) + iotests.qemu_io_log('-c', 'map', top) + iotests.qemu_img_log('map', '--output=json', top) diff --git a/tests/qemu-iotests/274.out b/tests/qemu-iotests/274.out new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/tests/qemu-iotests/274.out @@ -XXX,XX +XXX,XX @@ +== Commit tests == +Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=2097152 cluster_size=65536 lazy_refcounts=off refcount_bits=16 + +Formatting 'TEST_DIR/PID-mid', fmt=qcow2 size=1048576 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16 + +Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=2097152 backing_file=TEST_DIR/PID-mid cluster_size=65536 lazy_refcounts=off refcount_bits=16 + +wrote 2097152/2097152 bytes at offset 0 +2 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) + +=== Check visible data === +read 1048576/1048576 bytes at offset 0 +1 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) + +read 1048576/1048576 bytes at offset 1048576 +1 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) + +=== Checking allocation status === +1048576/1048576 bytes allocated at offset 0 bytes +1048576/1048576 bytes allocated at offset 1 MiB + +0/1048576 bytes allocated at offset 0 bytes +0/0 bytes allocated at offset 1 MiB + +0/1048576 bytes allocated at offset 0 bytes +0/1048576 bytes allocated at offset 1 MiB + +=== Checking map === +[{ "start": 0, "length": 2097152, "depth": 0, "zero": false, "data": true, "offset": 327680}] + +Offset Length Mapped to File +0 0x200000 0x50000 TEST_DIR/PID-base + +[{ "start": 0, "length": 1048576, "depth": 1, "zero": false, "data": true, "offset": 327680}] + +Offset Length Mapped to File +0 0x100000 0x50000 TEST_DIR/PID-base + +[{ "start": 0, "length": 1048576, "depth": 2, "zero": false, "data": true, "offset": 327680}, +{ "start": 1048576, "length": 1048576, "depth": 0, "zero": true, "data": false}] + +Offset Length Mapped to File +0 0x100000 0x50000 TEST_DIR/PID-base + +=== Testing qemu-img commit (top -> mid) === +Image committed. + +image: TEST_IMG +file format: IMGFMT +virtual size: 2 MiB (2097152 bytes) +cluster_size: 65536 +backing file: TEST_DIR/PID-base +Format specific information: + compat: 1.1 + lazy refcounts: false + refcount bits: 16 + corrupt: false + +read 1048576/1048576 bytes at offset 0 +1 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) + +read 1048576/1048576 bytes at offset 1048576 +1 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) + +=== Testing HMP commit (top -> mid) === +Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=2097152 cluster_size=65536 lazy_refcounts=off refcount_bits=16 + +Formatting 'TEST_DIR/PID-mid', fmt=qcow2 size=1048576 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16 + +Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=2097152 backing_file=TEST_DIR/PID-mid cluster_size=65536 lazy_refcounts=off refcount_bits=16 + +wrote 2097152/2097152 bytes at offset 0 +2 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) + +{"execute": "human-monitor-command", "arguments": {"command-line": "commit drive0"}} +{"return": ""} +image: TEST_IMG +file format: IMGFMT +virtual size: 2 MiB (2097152 bytes) +cluster_size: 65536 +backing file: TEST_DIR/PID-base +Format specific information: + compat: 1.1 + lazy refcounts: false + refcount bits: 16 + corrupt: false + +read 1048576/1048576 bytes at offset 0 +1 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) + +read 1048576/1048576 bytes at offset 1048576 +1 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) + +=== Testing QMP active commit (top -> mid) === +Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=2097152 cluster_size=65536 lazy_refcounts=off refcount_bits=16 + +Formatting 'TEST_DIR/PID-mid', fmt=qcow2 size=1048576 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16 + +Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=2097152 backing_file=TEST_DIR/PID-mid cluster_size=65536 lazy_refcounts=off refcount_bits=16 + +wrote 2097152/2097152 bytes at offset 0 +2 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) + +{"execute": "block-commit", "arguments": {"auto-dismiss": false, "base-node": "mid", "device": "top", "job-id": "job0"}} +{"return": {}} +{"execute": "job-complete", "arguments": {"id": "job0"}} +{"return": {}} +{"data": {"device": "job0", "len": 0, "offset": 0, "speed": 0, "type": "commit"}, "event": "BLOCK_JOB_READY", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} +{"data": {"device": "job0", "len": 0, "offset": 0, "speed": 0, "type": "commit"}, "event": "BLOCK_JOB_COMPLETED", "timestamp": {"microseconds": "USECS", "seconds": "SECS"}} +{"execute": "job-dismiss", "arguments": {"id": "job0"}} +{"return": {}} +image: TEST_IMG +file format: IMGFMT +virtual size: 2 MiB (2097152 bytes) +cluster_size: 65536 +backing file: TEST_DIR/PID-base +Format specific information: + compat: 1.1 + lazy refcounts: false + refcount bits: 16 + corrupt: false + +read 1048576/1048576 bytes at offset 0 +1 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) + +read 1048576/1048576 bytes at offset 1048576 +1 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) + +== Resize tests == +=== preallocation=off === +Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=6442450944 cluster_size=65536 lazy_refcounts=off refcount_bits=16 + +Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=1073741824 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16 + +wrote 65536/65536 bytes at offset 5368709120 +64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) + +Image resized. + +read 65536/65536 bytes at offset 5368709120 +64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) + +1 GiB (0x40000000) bytes not allocated at offset 0 bytes (0x0) +7 GiB (0x1c0000000) bytes allocated at offset 1 GiB (0x40000000) + +[{ "start": 0, "length": 1073741824, "depth": 1, "zero": true, "data": false}, +{ "start": 1073741824, "length": 7516192768, "depth": 0, "zero": true, "data": false}] + +=== preallocation=metadata === +Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=34359738368 cluster_size=65536 lazy_refcounts=off refcount_bits=16 + +Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=32212254720 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16 + +wrote 65536/65536 bytes at offset 33285996544 +64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) + +Image resized. + +read 65536/65536 bytes at offset 33285996544 +64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) + +30 GiB (0x780000000) bytes not allocated at offset 0 bytes (0x0) +3 GiB (0xc0000000) bytes allocated at offset 30 GiB (0x780000000) + +[{ "start": 0, "length": 32212254720, "depth": 1, "zero": true, "data": false}, +{ "start": 32212254720, "length": 536870912, "depth": 0, "zero": true, "data": false, "offset": 327680}, +{ "start": 32749125632, "length": 536870912, "depth": 0, "zero": true, "data": false, "offset": 537264128}, +{ "start": 33285996544, "length": 536870912, "depth": 0, "zero": true, "data": false, "offset": 1074200576}, +{ "start": 33822867456, "length": 536870912, "depth": 0, "zero": true, "data": false, "offset": 1611137024}, +{ "start": 34359738368, "length": 536870912, "depth": 0, "zero": true, "data": false, "offset": 2148139008}, +{ "start": 34896609280, "length": 536870912, "depth": 0, "zero": true, "data": false, "offset": 2685075456}] + +=== preallocation=falloc === +Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=10485760 cluster_size=65536 lazy_refcounts=off refcount_bits=16 + +Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=5242880 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16 + +wrote 65536/65536 bytes at offset 9437184 +64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) + +Image resized. + +read 65536/65536 bytes at offset 9437184 +64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) + +5 MiB (0x500000) bytes not allocated at offset 0 bytes (0x0) +10 MiB (0xa00000) bytes allocated at offset 5 MiB (0x500000) + +[{ "start": 0, "length": 5242880, "depth": 1, "zero": true, "data": false}, +{ "start": 5242880, "length": 10485760, "depth": 0, "zero": true, "data": false, "offset": 327680}] + +=== preallocation=full === +Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=16777216 cluster_size=65536 lazy_refcounts=off refcount_bits=16 + +Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=8388608 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16 + +wrote 65536/65536 bytes at offset 11534336 +64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) + +Image resized. + +read 65536/65536 bytes at offset 11534336 +64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) + +8 MiB (0x800000) bytes not allocated at offset 0 bytes (0x0) +4 MiB (0x400000) bytes allocated at offset 8 MiB (0x800000) + +[{ "start": 0, "length": 8388608, "depth": 1, "zero": true, "data": false}, +{ "start": 8388608, "length": 4194304, "depth": 0, "zero": true, "data": false, "offset": 327680}] + +=== preallocation=off === +Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=393216 cluster_size=65536 lazy_refcounts=off refcount_bits=16 + +Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=259072 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16 + +wrote 65536/65536 bytes at offset 259072 +64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) + +Image resized. + +read 65536/65536 bytes at offset 259072 +64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) + +192 KiB (0x30000) bytes not allocated at offset 0 bytes (0x0) +320 KiB (0x50000) bytes allocated at offset 192 KiB (0x30000) + +[{ "start": 0, "length": 196608, "depth": 1, "zero": true, "data": false}, +{ "start": 196608, "length": 65536, "depth": 0, "zero": false, "data": true, "offset": 327680}, +{ "start": 262144, "length": 262144, "depth": 0, "zero": true, "data": false}] + +=== preallocation=off === +Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=409600 cluster_size=65536 lazy_refcounts=off refcount_bits=16 + +Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=262144 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16 + +wrote 65536/65536 bytes at offset 344064 +64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) + +Image resized. + +read 65536/65536 bytes at offset 344064 +64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) + +256 KiB (0x40000) bytes not allocated at offset 0 bytes (0x0) +256 KiB (0x40000) bytes allocated at offset 256 KiB (0x40000) + +[{ "start": 0, "length": 262144, "depth": 1, "zero": true, "data": false}, +{ "start": 262144, "length": 262144, "depth": 0, "zero": true, "data": false}] + +=== preallocation=off === +Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=524288 cluster_size=65536 lazy_refcounts=off refcount_bits=16 + +Formatting 'TEST_DIR/PID-top', fmt=qcow2 size=262144 backing_file=TEST_DIR/PID-base cluster_size=65536 lazy_refcounts=off refcount_bits=16 + +wrote 65536/65536 bytes at offset 446464 +64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) + +Image resized. + +read 65536/65536 bytes at offset 446464 +64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) + +256 KiB (0x40000) bytes not allocated at offset 0 bytes (0x0) +244 KiB (0x3d000) bytes allocated at offset 256 KiB (0x40000) + +[{ "start": 0, "length": 262144, "depth": 1, "zero": true, "data": false}, +{ "start": 262144, "length": 249856, "depth": 0, "zero": true, "data": false}] + 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 @@ 270 rw backing quick 272 rw 273 backing quick +274 rw backing 277 rw quick 279 rw backing quick 280 rw migration quick -- 2.25.3 The BDRV_REQ_ZERO_WRITE is currently implemented in a way that first the image is possibly preallocated and then the zero flag is added to all clusters. This means that a copy-on-write operation may be needed when writing to these clusters, despite having used preallocation, negating one of the major benefits of preallocation. Instead, try to forward the BDRV_REQ_ZERO_WRITE to the protocol driver, and if the protocol driver can ensure that the new area reads as zeros, we can skip setting the zero flag in the qcow2 layer. Unfortunately, the same approach doesn't work for metadata preallocation, so we'll still set the zero flag there. Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com> Message-Id: <20200424142701.67053-1-kwolf@redhat.com> Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/qcow2.c | 22 +++++++++++++++++++--- tests/qemu-iotests/274.out | 4 ++-- 2 files changed, 21 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 coroutine_fn qcow2_co_truncate(BlockDriverState *bs, int64_t offset, /* Allocate the data area */ new_file_size = allocation_start + nb_new_data_clusters * s->cluster_size; - /* Image file grows, so @exact does not matter */ - ret = bdrv_co_truncate(bs->file, new_file_size, false, prealloc, 0, - errp); + /* + * Image file grows, so @exact does not matter. + * + * If we need to zero out the new area, try first whether the protocol + * driver can already take care of this. + */ + if (flags & BDRV_REQ_ZERO_WRITE) { + ret = bdrv_co_truncate(bs->file, new_file_size, false, prealloc, + BDRV_REQ_ZERO_WRITE, NULL); + if (ret >= 0) { + flags &= ~BDRV_REQ_ZERO_WRITE; + } + } else { + ret = -1; + } + if (ret < 0) { + ret = bdrv_co_truncate(bs->file, new_file_size, false, prealloc, 0, + errp); + } if (ret < 0) { error_prepend(errp, "Failed to resize underlying file: "); qcow2_free_clusters(bs, allocation_start, diff --git a/tests/qemu-iotests/274.out b/tests/qemu-iotests/274.out index XXXXXXX..XXXXXXX 100644 --- a/tests/qemu-iotests/274.out +++ b/tests/qemu-iotests/274.out @@ -XXX,XX +XXX,XX @@ read 65536/65536 bytes at offset 9437184 10 MiB (0xa00000) bytes allocated at offset 5 MiB (0x500000) [{ "start": 0, "length": 5242880, "depth": 1, "zero": true, "data": false}, -{ "start": 5242880, "length": 10485760, "depth": 0, "zero": true, "data": false, "offset": 327680}] +{ "start": 5242880, "length": 10485760, "depth": 0, "zero": false, "data": true, "offset": 327680}] === preallocation=full === Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=16777216 cluster_size=65536 lazy_refcounts=off refcount_bits=16 @@ -XXX,XX +XXX,XX @@ read 65536/65536 bytes at offset 11534336 4 MiB (0x400000) bytes allocated at offset 8 MiB (0x800000) [{ "start": 0, "length": 8388608, "depth": 1, "zero": true, "data": false}, -{ "start": 8388608, "length": 4194304, "depth": 0, "zero": true, "data": false, "offset": 327680}] +{ "start": 8388608, "length": 4194304, "depth": 0, "zero": false, "data": true, "offset": 327680}] === preallocation=off === Formatting 'TEST_DIR/PID-base', fmt=qcow2 size=393216 cluster_size=65536 lazy_refcounts=off refcount_bits=16 -- 2.25.3 From: Andrzej Jakowski <andrzej.jakowski@linux.intel.com> This patch introduces support for PMR that has been defined as part of NVMe 1.4 spec. User can now specify a pmrdev option that should point to HostMemoryBackend. pmrdev memory region will subsequently be exposed as PCI BAR 2 in emulated NVMe device. Guest OS can perform mmio read and writes to the PMR region that will stay persistent across system reboot. Signed-off-by: Andrzej Jakowski <andrzej.jakowski@linux.intel.com> Reviewed-by: Klaus Jensen <k.jensen@samsung.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Message-Id: <20200330164656.9348-1-andrzej.jakowski@linux.intel.com> Reviewed-by: Keith Busch <kbusch@kernel.org> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- hw/block/nvme.h | 2 + include/block/nvme.h | 172 +++++++++++++++++++++++++++++++++++++++++ hw/block/nvme.c | 109 ++++++++++++++++++++++++++ hw/block/Makefile.objs | 2 +- hw/block/trace-events | 4 + 5 files changed, 288 insertions(+), 1 deletion(-) diff --git a/hw/block/nvme.h b/hw/block/nvme.h index XXXXXXX..XXXXXXX 100644 --- a/hw/block/nvme.h +++ b/hw/block/nvme.h @@ -XXX,XX +XXX,XX @@ typedef struct NvmeCtrl { uint64_t timestamp_set_qemu_clock_ms; /* QEMU clock time */ char *serial; + HostMemoryBackend *pmrdev; + NvmeNamespace *namespaces; NvmeSQueue **sq; NvmeCQueue **cq; diff --git a/include/block/nvme.h b/include/block/nvme.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/nvme.h +++ b/include/block/nvme.h @@ -XXX,XX +XXX,XX @@ typedef struct NvmeBar { uint64_t acq; uint32_t cmbloc; uint32_t cmbsz; + uint8_t padding[3520]; /* not used by QEMU */ + uint32_t pmrcap; + uint32_t pmrctl; + uint32_t pmrsts; + uint32_t pmrebs; + uint32_t pmrswtp; + uint32_t pmrmsc; } NvmeBar; enum NvmeCapShift { @@ -XXX,XX +XXX,XX @@ enum NvmeCapShift { CAP_CSS_SHIFT = 37, CAP_MPSMIN_SHIFT = 48, CAP_MPSMAX_SHIFT = 52, + CAP_PMR_SHIFT = 56, }; enum NvmeCapMask { @@ -XXX,XX +XXX,XX @@ enum NvmeCapMask { CAP_CSS_MASK = 0xff, CAP_MPSMIN_MASK = 0xf, CAP_MPSMAX_MASK = 0xf, + CAP_PMR_MASK = 0x1, }; #define NVME_CAP_MQES(cap) (((cap) >> CAP_MQES_SHIFT) & CAP_MQES_MASK) @@ -XXX,XX +XXX,XX @@ enum NvmeCapMask { << CAP_MPSMIN_SHIFT) #define NVME_CAP_SET_MPSMAX(cap, val) (cap |= (uint64_t)(val & CAP_MPSMAX_MASK)\ << CAP_MPSMAX_SHIFT) +#define NVME_CAP_SET_PMRS(cap, val) (cap |= (uint64_t)(val & CAP_PMR_MASK)\ + << CAP_PMR_SHIFT) enum NvmeCcShift { CC_EN_SHIFT = 0, @@ -XXX,XX +XXX,XX @@ enum NvmeCmbszMask { #define NVME_CMBSZ_GETSIZE(cmbsz) \ (NVME_CMBSZ_SZ(cmbsz) * (1 << (12 + 4 * NVME_CMBSZ_SZU(cmbsz)))) +enum NvmePmrcapShift { + PMRCAP_RDS_SHIFT = 3, + PMRCAP_WDS_SHIFT = 4, + PMRCAP_BIR_SHIFT = 5, + PMRCAP_PMRTU_SHIFT = 8, + PMRCAP_PMRWBM_SHIFT = 10, + PMRCAP_PMRTO_SHIFT = 16, + PMRCAP_CMSS_SHIFT = 24, +}; + +enum NvmePmrcapMask { + PMRCAP_RDS_MASK = 0x1, + PMRCAP_WDS_MASK = 0x1, + PMRCAP_BIR_MASK = 0x7, + PMRCAP_PMRTU_MASK = 0x3, + PMRCAP_PMRWBM_MASK = 0xf, + PMRCAP_PMRTO_MASK = 0xff, + PMRCAP_CMSS_MASK = 0x1, +}; + +#define NVME_PMRCAP_RDS(pmrcap) \ + ((pmrcap >> PMRCAP_RDS_SHIFT) & PMRCAP_RDS_MASK) +#define NVME_PMRCAP_WDS(pmrcap) \ + ((pmrcap >> PMRCAP_WDS_SHIFT) & PMRCAP_WDS_MASK) +#define NVME_PMRCAP_BIR(pmrcap) \ + ((pmrcap >> PMRCAP_BIR_SHIFT) & PMRCAP_BIR_MASK) +#define NVME_PMRCAP_PMRTU(pmrcap) \ + ((pmrcap >> PMRCAP_PMRTU_SHIFT) & PMRCAP_PMRTU_MASK) +#define NVME_PMRCAP_PMRWBM(pmrcap) \ + ((pmrcap >> PMRCAP_PMRWBM_SHIFT) & PMRCAP_PMRWBM_MASK) +#define NVME_PMRCAP_PMRTO(pmrcap) \ + ((pmrcap >> PMRCAP_PMRTO_SHIFT) & PMRCAP_PMRTO_MASK) +#define NVME_PMRCAP_CMSS(pmrcap) \ + ((pmrcap >> PMRCAP_CMSS_SHIFT) & PMRCAP_CMSS_MASK) + +#define NVME_PMRCAP_SET_RDS(pmrcap, val) \ + (pmrcap |= (uint64_t)(val & PMRCAP_RDS_MASK) << PMRCAP_RDS_SHIFT) +#define NVME_PMRCAP_SET_WDS(pmrcap, val) \ + (pmrcap |= (uint64_t)(val & PMRCAP_WDS_MASK) << PMRCAP_WDS_SHIFT) +#define NVME_PMRCAP_SET_BIR(pmrcap, val) \ + (pmrcap |= (uint64_t)(val & PMRCAP_BIR_MASK) << PMRCAP_BIR_SHIFT) +#define NVME_PMRCAP_SET_PMRTU(pmrcap, val) \ + (pmrcap |= (uint64_t)(val & PMRCAP_PMRTU_MASK) << PMRCAP_PMRTU_SHIFT) +#define NVME_PMRCAP_SET_PMRWBM(pmrcap, val) \ + (pmrcap |= (uint64_t)(val & PMRCAP_PMRWBM_MASK) << PMRCAP_PMRWBM_SHIFT) +#define NVME_PMRCAP_SET_PMRTO(pmrcap, val) \ + (pmrcap |= (uint64_t)(val & PMRCAP_PMRTO_MASK) << PMRCAP_PMRTO_SHIFT) +#define NVME_PMRCAP_SET_CMSS(pmrcap, val) \ + (pmrcap |= (uint64_t)(val & PMRCAP_CMSS_MASK) << PMRCAP_CMSS_SHIFT) + +enum NvmePmrctlShift { + PMRCTL_EN_SHIFT = 0, +}; + +enum NvmePmrctlMask { + PMRCTL_EN_MASK = 0x1, +}; + +#define NVME_PMRCTL_EN(pmrctl) ((pmrctl >> PMRCTL_EN_SHIFT) & PMRCTL_EN_MASK) + +#define NVME_PMRCTL_SET_EN(pmrctl, val) \ + (pmrctl |= (uint64_t)(val & PMRCTL_EN_MASK) << PMRCTL_EN_SHIFT) + +enum NvmePmrstsShift { + PMRSTS_ERR_SHIFT = 0, + PMRSTS_NRDY_SHIFT = 8, + PMRSTS_HSTS_SHIFT = 9, + PMRSTS_CBAI_SHIFT = 12, +}; + +enum NvmePmrstsMask { + PMRSTS_ERR_MASK = 0xff, + PMRSTS_NRDY_MASK = 0x1, + PMRSTS_HSTS_MASK = 0x7, + PMRSTS_CBAI_MASK = 0x1, +}; + +#define NVME_PMRSTS_ERR(pmrsts) \ + ((pmrsts >> PMRSTS_ERR_SHIFT) & PMRSTS_ERR_MASK) +#define NVME_PMRSTS_NRDY(pmrsts) \ + ((pmrsts >> PMRSTS_NRDY_SHIFT) & PMRSTS_NRDY_MASK) +#define NVME_PMRSTS_HSTS(pmrsts) \ + ((pmrsts >> PMRSTS_HSTS_SHIFT) & PMRSTS_HSTS_MASK) +#define NVME_PMRSTS_CBAI(pmrsts) \ + ((pmrsts >> PMRSTS_CBAI_SHIFT) & PMRSTS_CBAI_MASK) + +#define NVME_PMRSTS_SET_ERR(pmrsts, val) \ + (pmrsts |= (uint64_t)(val & PMRSTS_ERR_MASK) << PMRSTS_ERR_SHIFT) +#define NVME_PMRSTS_SET_NRDY(pmrsts, val) \ + (pmrsts |= (uint64_t)(val & PMRSTS_NRDY_MASK) << PMRSTS_NRDY_SHIFT) +#define NVME_PMRSTS_SET_HSTS(pmrsts, val) \ + (pmrsts |= (uint64_t)(val & PMRSTS_HSTS_MASK) << PMRSTS_HSTS_SHIFT) +#define NVME_PMRSTS_SET_CBAI(pmrsts, val) \ + (pmrsts |= (uint64_t)(val & PMRSTS_CBAI_MASK) << PMRSTS_CBAI_SHIFT) + +enum NvmePmrebsShift { + PMREBS_PMRSZU_SHIFT = 0, + PMREBS_RBB_SHIFT = 4, + PMREBS_PMRWBZ_SHIFT = 8, +}; + +enum NvmePmrebsMask { + PMREBS_PMRSZU_MASK = 0xf, + PMREBS_RBB_MASK = 0x1, + PMREBS_PMRWBZ_MASK = 0xffffff, +}; + +#define NVME_PMREBS_PMRSZU(pmrebs) \ + ((pmrebs >> PMREBS_PMRSZU_SHIFT) & PMREBS_PMRSZU_MASK) +#define NVME_PMREBS_RBB(pmrebs) \ + ((pmrebs >> PMREBS_RBB_SHIFT) & PMREBS_RBB_MASK) +#define NVME_PMREBS_PMRWBZ(pmrebs) \ + ((pmrebs >> PMREBS_PMRWBZ_SHIFT) & PMREBS_PMRWBZ_MASK) + +#define NVME_PMREBS_SET_PMRSZU(pmrebs, val) \ + (pmrebs |= (uint64_t)(val & PMREBS_PMRSZU_MASK) << PMREBS_PMRSZU_SHIFT) +#define NVME_PMREBS_SET_RBB(pmrebs, val) \ + (pmrebs |= (uint64_t)(val & PMREBS_RBB_MASK) << PMREBS_RBB_SHIFT) +#define NVME_PMREBS_SET_PMRWBZ(pmrebs, val) \ + (pmrebs |= (uint64_t)(val & PMREBS_PMRWBZ_MASK) << PMREBS_PMRWBZ_SHIFT) + +enum NvmePmrswtpShift { + PMRSWTP_PMRSWTU_SHIFT = 0, + PMRSWTP_PMRSWTV_SHIFT = 8, +}; + +enum NvmePmrswtpMask { + PMRSWTP_PMRSWTU_MASK = 0xf, + PMRSWTP_PMRSWTV_MASK = 0xffffff, +}; + +#define NVME_PMRSWTP_PMRSWTU(pmrswtp) \ + ((pmrswtp >> PMRSWTP_PMRSWTU_SHIFT) & PMRSWTP_PMRSWTU_MASK) +#define NVME_PMRSWTP_PMRSWTV(pmrswtp) \ + ((pmrswtp >> PMRSWTP_PMRSWTV_SHIFT) & PMRSWTP_PMRSWTV_MASK) + +#define NVME_PMRSWTP_SET_PMRSWTU(pmrswtp, val) \ + (pmrswtp |= (uint64_t)(val & PMRSWTP_PMRSWTU_MASK) << PMRSWTP_PMRSWTU_SHIFT) +#define NVME_PMRSWTP_SET_PMRSWTV(pmrswtp, val) \ + (pmrswtp |= (uint64_t)(val & PMRSWTP_PMRSWTV_MASK) << PMRSWTP_PMRSWTV_SHIFT) + +enum NvmePmrmscShift { + PMRMSC_CMSE_SHIFT = 1, + PMRMSC_CBA_SHIFT = 12, +}; + +enum NvmePmrmscMask { + PMRMSC_CMSE_MASK = 0x1, + PMRMSC_CBA_MASK = 0xfffffffffffff, +}; + +#define NVME_PMRMSC_CMSE(pmrmsc) \ + ((pmrmsc >> PMRMSC_CMSE_SHIFT) & PMRMSC_CMSE_MASK) +#define NVME_PMRMSC_CBA(pmrmsc) \ + ((pmrmsc >> PMRMSC_CBA_SHIFT) & PMRMSC_CBA_MASK) + +#define NVME_PMRMSC_SET_CMSE(pmrmsc, val) \ + (pmrmsc |= (uint64_t)(val & PMRMSC_CMSE_MASK) << PMRMSC_CMSE_SHIFT) +#define NVME_PMRMSC_SET_CBA(pmrmsc, val) \ + (pmrmsc |= (uint64_t)(val & PMRMSC_CBA_MASK) << PMRMSC_CBA_SHIFT) + typedef struct NvmeCmd { uint8_t opcode; uint8_t fuse; 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 @@ * -drive file=<file>,if=none,id=<drive_id> * -device nvme,drive=<drive_id>,serial=<serial>,id=<id[optional]>, \ * cmb_size_mb=<cmb_size_mb[optional]>, \ + * [pmrdev=<mem_backend_file_id>,] \ * num_queues=<N[optional]> * * Note cmb_size_mb denotes size of CMB in MB. CMB is assumed to be at * offset 0 in BAR2 and supports only WDS, RDS and SQS for now. + * + * cmb_size_mb= and pmrdev= options are mutually exclusive due to limitation + * in available BAR's. cmb_size_mb= will take precedence over pmrdev= when + * both provided. + * Enabling pmr emulation can be achieved by pointing to memory-backend-file. + * For example: + * -object memory-backend-file,id=<mem_id>,share=on,mem-path=<file_path>, \ + * size=<size> .... -device nvme,...,pmrdev=<mem_id> */ #include "qemu/osdep.h" @@ -XXX,XX +XXX,XX @@ #include "sysemu/sysemu.h" #include "qapi/error.h" #include "qapi/visitor.h" +#include "sysemu/hostmem.h" #include "sysemu/block-backend.h" +#include "exec/ram_addr.h" #include "qemu/log.h" #include "qemu/module.h" @@ -XXX,XX +XXX,XX @@ static void nvme_write_bar(NvmeCtrl *n, hwaddr offset, uint64_t data, NVME_GUEST_ERR(nvme_ub_mmiowr_cmbsz_readonly, "invalid write to read only CMBSZ, ignored"); return; + case 0xE00: /* PMRCAP */ + NVME_GUEST_ERR(nvme_ub_mmiowr_pmrcap_readonly, + "invalid write to PMRCAP register, ignored"); + return; + case 0xE04: /* TODO PMRCTL */ + break; + case 0xE08: /* PMRSTS */ + NVME_GUEST_ERR(nvme_ub_mmiowr_pmrsts_readonly, + "invalid write to PMRSTS register, ignored"); + return; + case 0xE0C: /* PMREBS */ + NVME_GUEST_ERR(nvme_ub_mmiowr_pmrebs_readonly, + "invalid write to PMREBS register, ignored"); + return; + case 0xE10: /* PMRSWTP */ + NVME_GUEST_ERR(nvme_ub_mmiowr_pmrswtp_readonly, + "invalid write to PMRSWTP register, ignored"); + return; + case 0xE14: /* TODO PMRMSC */ + break; default: NVME_GUEST_ERR(nvme_ub_mmiowr_invalid, "invalid MMIO write," @@ -XXX,XX +XXX,XX @@ static uint64_t nvme_mmio_read(void *opaque, hwaddr addr, unsigned size) } if (addr < sizeof(n->bar)) { + /* + * When PMRWBM bit 1 is set then read from + * from PMRSTS should ensure prior writes + * made it to persistent media + */ + if (addr == 0xE08 && + (NVME_PMRCAP_PMRWBM(n->bar.pmrcap) & 0x02)) { + qemu_ram_writeback(n->pmrdev->mr.ram_block, + 0, n->pmrdev->size); + } memcpy(&val, ptr + addr, size); } else { NVME_GUEST_ERR(nvme_ub_mmiord_invalid_ofs, @@ -XXX,XX +XXX,XX @@ static void nvme_realize(PCIDevice *pci_dev, Error **errp) error_setg(errp, "serial property not set"); return; } + + if (!n->cmb_size_mb && n->pmrdev) { + if (host_memory_backend_is_mapped(n->pmrdev)) { + char *path = object_get_canonical_path_component(OBJECT(n->pmrdev)); + error_setg(errp, "can't use already busy memdev: %s", path); + g_free(path); + return; + } + + if (!is_power_of_2(n->pmrdev->size)) { + error_setg(errp, "pmr backend size needs to be power of 2 in size"); + return; + } + + host_memory_backend_set_mapped(n->pmrdev, true); + } + blkconf_blocksizes(&n->conf); if (!blkconf_apply_backend_options(&n->conf, blk_is_read_only(n->conf.blk), false, errp)) { @@ -XXX,XX +XXX,XX @@ static void nvme_realize(PCIDevice *pci_dev, Error **errp) PCI_BASE_ADDRESS_SPACE_MEMORY | PCI_BASE_ADDRESS_MEM_TYPE_64 | PCI_BASE_ADDRESS_MEM_PREFETCH, &n->ctrl_mem); + } else if (n->pmrdev) { + /* Controller Capabilities register */ + NVME_CAP_SET_PMRS(n->bar.cap, 1); + + /* PMR Capabities register */ + n->bar.pmrcap = 0; + NVME_PMRCAP_SET_RDS(n->bar.pmrcap, 0); + NVME_PMRCAP_SET_WDS(n->bar.pmrcap, 0); + NVME_PMRCAP_SET_BIR(n->bar.pmrcap, 2); + NVME_PMRCAP_SET_PMRTU(n->bar.pmrcap, 0); + /* Turn on bit 1 support */ + NVME_PMRCAP_SET_PMRWBM(n->bar.pmrcap, 0x02); + NVME_PMRCAP_SET_PMRTO(n->bar.pmrcap, 0); + NVME_PMRCAP_SET_CMSS(n->bar.pmrcap, 0); + + /* PMR Control register */ + n->bar.pmrctl = 0; + NVME_PMRCTL_SET_EN(n->bar.pmrctl, 0); + + /* PMR Status register */ + n->bar.pmrsts = 0; + NVME_PMRSTS_SET_ERR(n->bar.pmrsts, 0); + NVME_PMRSTS_SET_NRDY(n->bar.pmrsts, 0); + NVME_PMRSTS_SET_HSTS(n->bar.pmrsts, 0); + NVME_PMRSTS_SET_CBAI(n->bar.pmrsts, 0); + + /* PMR Elasticity Buffer Size register */ + n->bar.pmrebs = 0; + NVME_PMREBS_SET_PMRSZU(n->bar.pmrebs, 0); + NVME_PMREBS_SET_RBB(n->bar.pmrebs, 0); + NVME_PMREBS_SET_PMRWBZ(n->bar.pmrebs, 0); + + /* PMR Sustained Write Throughput register */ + n->bar.pmrswtp = 0; + NVME_PMRSWTP_SET_PMRSWTU(n->bar.pmrswtp, 0); + NVME_PMRSWTP_SET_PMRSWTV(n->bar.pmrswtp, 0); + + /* PMR Memory Space Control register */ + n->bar.pmrmsc = 0; + NVME_PMRMSC_SET_CMSE(n->bar.pmrmsc, 0); + NVME_PMRMSC_SET_CBA(n->bar.pmrmsc, 0); + + pci_register_bar(pci_dev, NVME_PMRCAP_BIR(n->bar.pmrcap), + PCI_BASE_ADDRESS_SPACE_MEMORY | PCI_BASE_ADDRESS_MEM_TYPE_64 | + PCI_BASE_ADDRESS_MEM_PREFETCH, &n->pmrdev->mr); } for (i = 0; i < n->num_namespaces; i++) { @@ -XXX,XX +XXX,XX @@ static void nvme_exit(PCIDevice *pci_dev) if (n->cmb_size_mb) { g_free(n->cmbuf); } + + if (n->pmrdev) { + host_memory_backend_set_mapped(n->pmrdev, false); + } msix_uninit_exclusive_bar(pci_dev); } static Property nvme_props[] = { DEFINE_BLOCK_PROPERTIES(NvmeCtrl, conf), + DEFINE_PROP_LINK("pmrdev", NvmeCtrl, pmrdev, TYPE_MEMORY_BACKEND, + HostMemoryBackend *), DEFINE_PROP_STRING("serial", NvmeCtrl, serial), DEFINE_PROP_UINT32("cmb_size_mb", NvmeCtrl, cmb_size_mb, 0), DEFINE_PROP_UINT32("num_queues", NvmeCtrl, num_queues, 64), diff --git a/hw/block/Makefile.objs b/hw/block/Makefile.objs index XXXXXXX..XXXXXXX 100644 --- a/hw/block/Makefile.objs +++ b/hw/block/Makefile.objs @@ -XXX,XX +XXX,XX @@ common-obj-$(CONFIG_PFLASH_CFI02) += pflash_cfi02.o common-obj-$(CONFIG_XEN) += xen-block.o common-obj-$(CONFIG_ECC) += ecc.o common-obj-$(CONFIG_ONENAND) += onenand.o -common-obj-$(CONFIG_NVME_PCI) += nvme.o common-obj-$(CONFIG_SWIM) += swim.o common-obj-$(CONFIG_SH4) += tc58128.o obj-$(CONFIG_VIRTIO_BLK) += virtio-blk.o obj-$(CONFIG_VHOST_USER_BLK) += vhost-user-blk.o +obj-$(CONFIG_NVME_PCI) += nvme.o obj-y += dataplane/ 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 @@ nvme_ub_mmiowr_ssreset_w1c_unsupported(void) "attempted to W1C CSTS.NSSRO but CA 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_pmrcap_readonly(void) "invalid write to read only PMRCAP, ignored" +nvme_ub_mmiowr_pmrsts_readonly(void) "invalid write to read only PMRSTS, ignored" +nvme_ub_mmiowr_pmrebs_readonly(void) "invalid write to read only PMREBS, ignored" +nvme_ub_mmiowr_pmrswtp_readonly(void) "invalid write to read only PMRSWTP, 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"" -- 2.25.3 The QMP handler qmp_object_add() and the implementation of --object in qemu-storage-daemon can share most of the code. Currently, qemu-storage-daemon calls qmp_object_add(), but this is not correct because different visitors need to be used. As a first step towards a fix, make qmp_object_add() a wrapper around a new function user_creatable_add_dict() that can get an additional parameter. The handling of "props" is only required for compatibility and not required for the qemu-storage-daemon command line, so it stays in qmp_object_add(). Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/qom/object_interfaces.h | 12 ++++++++++++ qom/object_interfaces.c | 27 +++++++++++++++++++++++++++ qom/qom-qmp-cmds.c | 24 +----------------------- 3 files changed, 40 insertions(+), 23 deletions(-) diff --git a/include/qom/object_interfaces.h b/include/qom/object_interfaces.h index XXXXXXX..XXXXXXX 100644 --- a/include/qom/object_interfaces.h +++ b/include/qom/object_interfaces.h @@ -XXX,XX +XXX,XX @@ Object *user_creatable_add_type(const char *type, const char *id, const QDict *qdict, Visitor *v, Error **errp); +/** + * user_creatable_add_dict: + * @qdict: the object definition + * @errp: if an error occurs, a pointer to an area to store the error + * + * Create an instance of the user creatable object that is defined by + * @qdict. The object type is taken from the QDict key 'qom-type', its + * ID from the key 'id'. The remaining entries in @qdict are used to + * initialize the object properties. + */ +void user_creatable_add_dict(QDict *qdict, Error **errp); + /** * user_creatable_add_opts: * @opts: the object definition diff --git a/qom/object_interfaces.c b/qom/object_interfaces.c index XXXXXXX..XXXXXXX 100644 --- a/qom/object_interfaces.c +++ b/qom/object_interfaces.c @@ -XXX,XX +XXX,XX @@ #include "qapi/qmp/qerror.h" #include "qapi/qmp/qjson.h" #include "qapi/qmp/qstring.h" +#include "qapi/qobject-input-visitor.h" #include "qom/object_interfaces.h" #include "qemu/help_option.h" #include "qemu/module.h" @@ -XXX,XX +XXX,XX @@ out: return obj; } +void user_creatable_add_dict(QDict *qdict, Error **errp) +{ + Visitor *v; + Object *obj; + g_autofree char *type = NULL; + g_autofree char *id = NULL; + + type = g_strdup(qdict_get_try_str(qdict, "qom-type")); + if (!type) { + error_setg(errp, QERR_MISSING_PARAMETER, "qom-type"); + return; + } + qdict_del(qdict, "qom-type"); + + id = g_strdup(qdict_get_try_str(qdict, "id")); + if (!id) { + error_setg(errp, QERR_MISSING_PARAMETER, "id"); + return; + } + qdict_del(qdict, "id"); + + v = qobject_input_visitor_new(QOBJECT(qdict)); + obj = user_creatable_add_type(type, id, qdict, v, errp); + visit_free(v); + object_unref(obj); +} Object *user_creatable_add_opts(QemuOpts *opts, Error **errp) { diff --git a/qom/qom-qmp-cmds.c b/qom/qom-qmp-cmds.c index XXXXXXX..XXXXXXX 100644 --- a/qom/qom-qmp-cmds.c +++ b/qom/qom-qmp-cmds.c @@ -XXX,XX +XXX,XX @@ #include "qapi/qapi-commands-qom.h" #include "qapi/qmp/qdict.h" #include "qapi/qmp/qerror.h" -#include "qapi/qobject-input-visitor.h" #include "qemu/cutils.h" #include "qom/object_interfaces.h" #include "qom/qom-qobject.h" @@ -XXX,XX +XXX,XX @@ void qmp_object_add(QDict *qdict, QObject **ret_data, Error **errp) { QObject *props; QDict *pdict; - Visitor *v; - Object *obj; - g_autofree char *type = NULL; - g_autofree char *id = NULL; - - type = g_strdup(qdict_get_try_str(qdict, "qom-type")); - if (!type) { - error_setg(errp, QERR_MISSING_PARAMETER, "qom-type"); - return; - } - qdict_del(qdict, "qom-type"); - - id = g_strdup(qdict_get_try_str(qdict, "id")); - if (!id) { - error_setg(errp, QERR_MISSING_PARAMETER, "id"); - return; - } - qdict_del(qdict, "id"); props = qdict_get(qdict, "props"); if (props) { @@ -XXX,XX +XXX,XX @@ void qmp_object_add(QDict *qdict, QObject **ret_data, Error **errp) qobject_unref(pdict); } - v = qobject_input_visitor_new(QOBJECT(qdict)); - obj = user_creatable_add_type(type, id, qdict, v, errp); - visit_free(v); - object_unref(obj); + user_creatable_add_dict(qdict, errp); } void qmp_object_del(const char *id, Error **errp) -- 2.25.3 After processing the option string with the keyval parser, we get a QDict that contains only strings. This QDict must be fed to a keyval visitor which converts the strings into the right data types. qmp_object_add(), however, uses the normal QObject input visitor, which expects a QDict where all properties already have the QType that matches the data type required by the QOM object type. Change the --object implementation in qemu-storage-daemon so that it doesn't call qmp_object_add(), but calls user_creatable_add_dict() directly instead and pass it a new keyval boolean that decides which visitor must be used. Reported-by: Coiby Xu <coiby.xu@gmail.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/qom/object_interfaces.h | 6 +++++- qemu-storage-daemon.c | 4 +--- qom/object_interfaces.c | 8 ++++++-- qom/qom-qmp-cmds.c | 2 +- 4 files changed, 13 insertions(+), 7 deletions(-) diff --git a/include/qom/object_interfaces.h b/include/qom/object_interfaces.h index XXXXXXX..XXXXXXX 100644 --- a/include/qom/object_interfaces.h +++ b/include/qom/object_interfaces.h @@ -XXX,XX +XXX,XX @@ Object *user_creatable_add_type(const char *type, const char *id, /** * user_creatable_add_dict: * @qdict: the object definition + * @keyval: if true, use a keyval visitor for processing @qdict (i.e. + * assume that all @qdict values are strings); otherwise, use + * the normal QObject visitor (i.e. assume all @qdict values + * have the QType expected by the QOM object type) * @errp: if an error occurs, a pointer to an area to store the error * * Create an instance of the user creatable object that is defined by @@ -XXX,XX +XXX,XX @@ Object *user_creatable_add_type(const char *type, const char *id, * ID from the key 'id'. The remaining entries in @qdict are used to * initialize the object properties. */ -void user_creatable_add_dict(QDict *qdict, Error **errp); +void user_creatable_add_dict(QDict *qdict, bool keyval, Error **errp); /** * user_creatable_add_opts: diff --git a/qemu-storage-daemon.c b/qemu-storage-daemon.c index XXXXXXX..XXXXXXX 100644 --- a/qemu-storage-daemon.c +++ b/qemu-storage-daemon.c @@ -XXX,XX +XXX,XX @@ static void process_options(int argc, char *argv[]) QemuOpts *opts; const char *type; QDict *args; - QObject *ret_data = NULL; /* FIXME The keyval parser rejects 'help' arguments, so we must * unconditionall try QemuOpts first. */ @@ -XXX,XX +XXX,XX @@ static void process_options(int argc, char *argv[]) qemu_opts_del(opts); args = keyval_parse(optarg, "qom-type", &error_fatal); - qmp_object_add(args, &ret_data, &error_fatal); + user_creatable_add_dict(args, true, &error_fatal); qobject_unref(args); - qobject_unref(ret_data); break; } default: diff --git a/qom/object_interfaces.c b/qom/object_interfaces.c index XXXXXXX..XXXXXXX 100644 --- a/qom/object_interfaces.c +++ b/qom/object_interfaces.c @@ -XXX,XX +XXX,XX @@ out: return obj; } -void user_creatable_add_dict(QDict *qdict, Error **errp) +void user_creatable_add_dict(QDict *qdict, bool keyval, Error **errp) { Visitor *v; Object *obj; @@ -XXX,XX +XXX,XX @@ void user_creatable_add_dict(QDict *qdict, Error **errp) } qdict_del(qdict, "id"); - v = qobject_input_visitor_new(QOBJECT(qdict)); + if (keyval) { + v = qobject_input_visitor_new_keyval(QOBJECT(qdict)); + } else { + v = qobject_input_visitor_new(QOBJECT(qdict)); + } obj = user_creatable_add_type(type, id, qdict, v, errp); visit_free(v); object_unref(obj); diff --git a/qom/qom-qmp-cmds.c b/qom/qom-qmp-cmds.c index XXXXXXX..XXXXXXX 100644 --- a/qom/qom-qmp-cmds.c +++ b/qom/qom-qmp-cmds.c @@ -XXX,XX +XXX,XX @@ void qmp_object_add(QDict *qdict, QObject **ret_data, Error **errp) qobject_unref(pdict); } - user_creatable_add_dict(qdict, errp); + user_creatable_add_dict(qdict, false, errp); } void qmp_object_del(const char *id, Error **errp) -- 2.25.3 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