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 6629bf78aac7e53f83fd0bcbdbe322e2302dfd1f: Merge remote-tracking branch 'remotes/pmaydell/tags/pull-target-arm-20220302' into staging (2022-03-03 14:46:48 +0000) are available in the Git repository at: https://gitlab.com/kmwolf/qemu.git tags/for-upstream for you to fetch changes up to a5df6d1db7db37636c8624bf4a9df9da645853aa: block/amend: Keep strong reference to BDS (2022-03-04 17:15:33 +0100) ---------------------------------------------------------------- Block layer patches - qemu-storage-daemon: Add --daemonize - Fix x-blockdev-amend and block node activation code which incorrectly executed code in the iothread that must run in the main thread. - Add macros for coroutine-safe TLS variables (required for correctness with LTO) - Fix crashes with concurrent I/O and bdrv_refresh_limits() - Split block APIs in global state and I/O - iotests: Don't refuse to run at all without GNU sed, just skip tests that need it ---------------------------------------------------------------- Emanuele Giuseppe Esposito (36): crypto: perform permission checks under BQL crypto: distinguish between main loop and I/O in block_crypto_amend_options_generic_luks block: introduce bdrv_activate block: rename bdrv_invalidate_cache_all, blk_invalidate_cache and test_sync_op_invalidate_cache block: move BQL logic of bdrv_co_invalidate_cache in bdrv_activate main-loop.h: introduce qemu_in_main_thread() main loop: macros to mark GS and I/O functions include/block/block: split header into I/O and global state API assertions for block global state API IO_CODE and IO_OR_GS_CODE for block I/O API block/export/fuse.c: allow writable exports to take RESIZE permission include/sysemu/block-backend: split header into I/O and global state (GS) API block/block-backend.c: assertions for block-backend IO_CODE and IO_OR_GS_CODE for block-backend I/O API block.c: assertions to the block layer permissions API include/block/block_int: split header into I/O and global state API assertions for block_int global state API IO_CODE and IO_OR_GS_CODE for block_int I/O API block: introduce assert_bdrv_graph_writable include/block/blockjob_int.h: split header into I/O and GS API GS and IO CODE macros for blockjob_int.h block.c: add assertions to static functions include/block/blockjob.h: global state API assertions for blockjob.h global state API include/sysemu/blockdev.h: global state API assertions for blockdev.h global state API include/block/snapshot: global state API + assertions block/copy-before-write.h: global state API + assertions block/coroutines: I/O and "I/O or GS" API block_int-common.h: split function pointers in BlockDriver block_int-common.h: assertions in the callers of BlockDriver function pointers block_int-common.h: split function pointers in BdrvChildClass block_int-common.h: assertions in the callers of BdrvChildClass function pointers block-backend-common.h: split function pointers in BlockDevOps job.h: split function pointers in JobDriver job.h: assertions in the callers of JobDriver function pointers Hanna Reitz (9): os-posix: Add os_set_daemonize() qsd: Add pre-init argument parsing pass qsd: Add --daemonize iotests/185: Add post-READY quit tests block: Make bdrv_refresh_limits() non-recursive iotests: Allow using QMP with the QSD iotests/graph-changes-while-io: New test block/amend: Always call .bdrv_amend_clean() block/amend: Keep strong reference to BDS Stefan Hajnoczi (4): tls: add macros for coroutine-safe TLS variables util/async: replace __thread with QEMU TLS macros rcu: use coroutine TLS macros cpus: use coroutine TLS macros for iothread_locked Thomas Huth (1): tests/qemu-iotests: Rework the checks and spots using GNU sed docs/tools/qemu-storage-daemon.rst | 7 + block/copy-before-write.h | 7 + block/coroutines.h | 81 +- include/block/block-common.h | 418 ++++++ include/block/block-global-state.h | 253 ++++ include/block/block-io.h | 368 +++++ include/block/block.h | 878 +----------- include/block/block_int-common.h | 1222 ++++++++++++++++ include/block/block_int-global-state.h | 329 +++++ include/block/block_int-io.h | 185 +++ include/block/block_int.h | 1475 +------------------- include/block/blockjob.h | 29 +- include/block/blockjob_int.h | 28 + include/block/snapshot.h | 13 +- include/qemu/coroutine-tls.h | 165 +++ include/qemu/job.h | 22 + include/qemu/main-loop.h | 42 + include/qemu/rcu.h | 7 +- include/sysemu/block-backend-common.h | 102 ++ include/sysemu/block-backend-global-state.h | 116 ++ include/sysemu/block-backend-io.h | 161 +++ include/sysemu/block-backend.h | 269 +--- include/sysemu/blockdev.h | 13 +- include/sysemu/os-posix.h | 1 + include/sysemu/os-win32.h | 8 + block.c | 321 ++++- block/amend.c | 28 + block/backup.c | 1 + block/block-backend.c | 166 ++- block/commit.c | 4 + block/copy-before-write.c | 2 + block/create.c | 2 + block/crypto.c | 68 +- block/dirty-bitmap.c | 5 + block/export/export.c | 2 +- block/export/fuse.c | 25 +- block/io.c | 75 +- block/mirror.c | 4 + block/monitor/bitmap-qmp-cmds.c | 6 + block/nbd.c | 1 + block/parallels.c | 2 +- block/snapshot.c | 28 + block/stream.c | 2 + blockdev.c | 29 + blockjob.c | 16 + hw/block/pflash_cfi01.c | 2 +- hw/nvram/spapr_nvram.c | 2 +- job.c | 10 + migration/block.c | 2 +- migration/migration.c | 14 +- migration/savevm.c | 8 +- monitor/qmp-cmds.c | 2 +- os-posix.c | 6 + softmmu/cpus.c | 14 +- softmmu/qdev-monitor.c | 2 + storage-daemon/qemu-storage-daemon.c | 58 +- stubs/iothread-lock.c | 5 + tests/unit/rcutorture.c | 10 +- tests/unit/test-block-iothread.c | 8 +- tests/unit/test-rcu-list.c | 4 +- util/async.c | 12 +- util/rcu.c | 10 +- tests/qemu-iotests/iotests.py | 32 +- block/meson.build | 7 +- tests/check-block.sh | 12 - tests/qemu-iotests/185 | 190 ++- tests/qemu-iotests/185.out | 48 + tests/qemu-iotests/271 | 2 +- tests/qemu-iotests/296 | 8 +- tests/qemu-iotests/296.out | 17 +- tests/qemu-iotests/common.filter | 65 +- tests/qemu-iotests/common.rc | 45 +- tests/qemu-iotests/tests/graph-changes-while-io | 91 ++ .../qemu-iotests/tests/graph-changes-while-io.out | 5 + 74 files changed, 4854 insertions(+), 2823 deletions(-) create mode 100644 include/block/block-common.h create mode 100644 include/block/block-global-state.h create mode 100644 include/block/block-io.h create mode 100644 include/block/block_int-common.h create mode 100644 include/block/block_int-global-state.h create mode 100644 include/block/block_int-io.h create mode 100644 include/qemu/coroutine-tls.h create mode 100644 include/sysemu/block-backend-common.h create mode 100644 include/sysemu/block-backend-global-state.h create mode 100644 include/sysemu/block-backend-io.h create mode 100755 tests/qemu-iotests/tests/graph-changes-while-io create mode 100644 tests/qemu-iotests/tests/graph-changes-while-io.out From: Emanuele Giuseppe Esposito <eesposit@redhat.com> Move the permission API calls into driver-specific callbacks that always run under BQL. In this case, bdrv_crypto_luks needs to perform permission checks before and after qcrypto_block_amend_options(). The problem is that the caller, block_crypto_amend_options_generic_luks(), can also run in I/O from .bdrv_co_amend(). This does not comply with Global State-I/O API split, as permissions API must always run under BQL. Firstly, introduce .bdrv_amend_pre_run() and .bdrv_amend_clean() callbacks. These two callbacks are guaranteed to be invoked under BQL, respectively before and after .bdrv_co_amend(). They take care of performing the permission checks in the same way as they are currently done before and after qcrypto_block_amend_options(). These callbacks are in preparation for next patch, where we delete the original permission check. Right now they just add redundant control. Then, call .bdrv_amend_pre_run() before job_start in qmp_x_blockdev_amend(), so that it will be run before the job coroutine is created and stay in the main loop. As a cleanup, use JobDriver's .clean() callback to call .bdrv_amend_clean(), and run amend-specific cleanup callbacks under BQL. After this patch, permission failures occur early in the blockdev-amend job to update a LUKS volume's keys. iotest 296 must now expect them in x-blockdev-amend's QMP reply instead of waiting for the actual job to fail later. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Message-Id: <20220209105452.1694545-2-eesposit@redhat.com> Signed-off-by: Hanna Reitz <hreitz@redhat.com> Message-Id: <20220304153729.711387-6-hreitz@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block_int.h | 14 ++++++++++++++ block/amend.c | 25 +++++++++++++++++++++++++ block/crypto.c | 33 +++++++++++++++++++++++++++++++++ tests/qemu-iotests/296 | 8 ++++++-- tests/qemu-iotests/296.out | 17 +++++------------ 5 files changed, 83 insertions(+), 14 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 { * on those children. */ bool is_format; + + /* + * This function is invoked under BQL before .bdrv_co_amend() + * (which in contrast does not necessarily run under the BQL) + * to allow driver-specific initialization code that requires + * the BQL, like setting up specific permission flags. + */ + int (*bdrv_amend_pre_run)(BlockDriverState *bs, Error **errp); + /* + * This function is invoked under BQL after .bdrv_co_amend() + * to allow cleaning up what was done in .bdrv_amend_pre_run(). + */ + void (*bdrv_amend_clean)(BlockDriverState *bs); + /* * Return true if @to_replace can be replaced by a BDS with the * same data as @bs without it affecting @bs's behavior (that is, diff --git a/block/amend.c b/block/amend.c index XXXXXXX..XXXXXXX 100644 --- a/block/amend.c +++ b/block/amend.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn blockdev_amend_run(Job *job, Error **errp) return ret; } +static int blockdev_amend_pre_run(BlockdevAmendJob *s, Error **errp) +{ + if (s->bs->drv->bdrv_amend_pre_run) { + return s->bs->drv->bdrv_amend_pre_run(s->bs, errp); + } + + return 0; +} + +static void blockdev_amend_clean(Job *job) +{ + BlockdevAmendJob *s = container_of(job, BlockdevAmendJob, common); + + if (s->bs->drv->bdrv_amend_clean) { + s->bs->drv->bdrv_amend_clean(s->bs); + } +} + static const JobDriver blockdev_amend_job_driver = { .instance_size = sizeof(BlockdevAmendJob), .job_type = JOB_TYPE_AMEND, .run = blockdev_amend_run, + .clean = blockdev_amend_clean, }; void qmp_x_blockdev_amend(const char *job_id, @@ -XXX,XX +XXX,XX @@ void qmp_x_blockdev_amend(const char *job_id, s->bs = bs, s->opts = QAPI_CLONE(BlockdevAmendOptions, options), s->force = has_force ? force : false; + + if (blockdev_amend_pre_run(s, errp)) { + job_early_fail(&s->common); + return; + } + job_start(&s->common); } 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_get_specific_info_luks(BlockDriverState *bs, Error **errp) return spec_info; } +static int +block_crypto_amend_prepare(BlockDriverState *bs, Error **errp) +{ + BlockCrypto *crypto = bs->opaque; + int ret; + + /* apply for exclusive read/write permissions to the underlying file */ + crypto->updating_keys = true; + ret = bdrv_child_refresh_perms(bs, bs->file, errp); + if (ret < 0) { + /* Well, in this case we will not be updating any keys */ + crypto->updating_keys = false; + } + return ret; +} + +static void +block_crypto_amend_cleanup(BlockDriverState *bs) +{ + BlockCrypto *crypto = bs->opaque; + Error *errp = NULL; + + /* release exclusive read/write permissions to the underlying file */ + crypto->updating_keys = false; + bdrv_child_refresh_perms(bs, bs->file, &errp); + + if (errp) { + error_report_err(errp); + } +} + static int block_crypto_amend_options_generic_luks(BlockDriverState *bs, QCryptoBlockAmendOptions *amend_options, @@ -XXX,XX +XXX,XX @@ static BlockDriver bdrv_crypto_luks = { .bdrv_get_specific_info = block_crypto_get_specific_info_luks, .bdrv_amend_options = block_crypto_amend_options_luks, .bdrv_co_amend = block_crypto_co_amend_luks, + .bdrv_amend_pre_run = block_crypto_amend_prepare, + .bdrv_amend_clean = block_crypto_amend_cleanup, .is_format = true, diff --git a/tests/qemu-iotests/296 b/tests/qemu-iotests/296 index XXXXXXX..XXXXXXX 100755 --- a/tests/qemu-iotests/296 +++ b/tests/qemu-iotests/296 @@ -XXX,XX +XXX,XX @@ class EncryptionSetupTestCase(iotests.QMPTestCase): } result = vm.qmp('x-blockdev-amend', **args) - assert result['return'] == {} - vm.run_job('job0') + iotests.log(result) + # Run the job only if it was created + event = ('JOB_STATUS_CHANGE', + {'data': {'id': 'job0', 'status': 'created'}}) + if vm.events_wait([event], timeout=0.0) is not None: + vm.run_job('job0') # test that when the image opened by two qemu processes, # neither of them can update the encryption keys diff --git a/tests/qemu-iotests/296.out b/tests/qemu-iotests/296.out index XXXXXXX..XXXXXXX 100644 --- a/tests/qemu-iotests/296.out +++ b/tests/qemu-iotests/296.out @@ -XXX,XX +XXX,XX @@ -{"execute": "job-dismiss", "arguments": {"id": "job0"}} {"return": {}} -Job failed: Failed to get shared "consistent read" lock {"execute": "job-dismiss", "arguments": {"id": "job0"}} {"return": {}} -Job failed: Failed to get shared "consistent read" lock -{"execute": "job-dismiss", "arguments": {"id": "job0"}} +{"error": {"class": "GenericError", "desc": "Failed to get shared \"consistent read\" lock"}} +{"error": {"class": "GenericError", "desc": "Failed to get shared \"consistent read\" lock"}} {"return": {}} {"execute": "job-dismiss", "arguments": {"id": "job0"}} {"return": {}} @@ -XXX,XX +XXX,XX @@ qemu-img: Failed to get shared "consistent read" lock Is another process using the image [TEST_DIR/test.img]? . -Job failed: Block node is read-only -{"execute": "job-dismiss", "arguments": {"id": "job0"}} -{"return": {}} -Job failed: Failed to get shared "consistent read" lock -{"execute": "job-dismiss", "arguments": {"id": "job0"}} -{"return": {}} -Job failed: Failed to get shared "consistent read" lock -{"execute": "job-dismiss", "arguments": {"id": "job0"}} +{"error": {"class": "GenericError", "desc": "Block node is read-only"}} +{"error": {"class": "GenericError", "desc": "Failed to get shared \"consistent read\" lock"}} +{"error": {"class": "GenericError", "desc": "Failed to get shared \"consistent read\" lock"}} {"return": {}} {"execute": "job-dismiss", "arguments": {"id": "job0"}} {"return": {}} -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> block_crypto_amend_options_generic_luks uses the block layer permission API, therefore it should be called with the BQL held. However, the same function is being called by two BlockDriver callbacks: bdrv_amend_options (under BQL) and bdrv_co_amend (I/O). The latter is I/O because it is invoked by block/amend.c's blockdev_amend_run(), a .run callback of the amend JobDriver. Therefore we want to change this function to still perform the permission check, but making sure it is done under BQL regardless of the caller context. Remove the permission check in block_crypto_amend_options_generic_luks() and: - in block_crypto_amend_options_luks() (BQL case, called by .bdrv_amend_options()), reuse helper functions block_crypto_amend_{prepare/cleanup} that take care of checking permissions. - for block_crypto_co_amend_luks() (I/O case, called by .bdrv_co_amend()), don't check for permissions but delegate .bdrv_amend_pre_run() and .bdrv_amend_clean() to do it, performing these checks before and after the job runs in its aiocontext. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Message-Id: <20220209105452.1694545-3-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/crypto.c | 35 +++++++++++++++-------------------- 1 file changed, 15 insertions(+), 20 deletions(-) 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_amend_options_generic_luks(BlockDriverState *bs, Error **errp) { BlockCrypto *crypto = bs->opaque; - int ret; assert(crypto); assert(crypto->block); - /* apply for exclusive read/write permissions to the underlying file*/ - crypto->updating_keys = true; - ret = bdrv_child_refresh_perms(bs, bs->file, errp); - if (ret) { - goto cleanup; - } - - ret = qcrypto_block_amend_options(crypto->block, - block_crypto_read_func, - block_crypto_write_func, - bs, - amend_options, - force, - errp); -cleanup: - /* release exclusive read/write permissions to the underlying file*/ - crypto->updating_keys = false; - bdrv_child_refresh_perms(bs, bs->file, errp); - return ret; + return qcrypto_block_amend_options(crypto->block, + block_crypto_read_func, + block_crypto_write_func, + bs, + amend_options, + force, + errp); } static int @@ -XXX,XX +XXX,XX @@ block_crypto_amend_options_luks(BlockDriverState *bs, if (!amend_options) { goto cleanup; } + + ret = block_crypto_amend_prepare(bs, errp); + if (ret) { + goto perm_cleanup; + } ret = block_crypto_amend_options_generic_luks(bs, amend_options, force, errp); + +perm_cleanup: + block_crypto_amend_cleanup(bs); cleanup: qapi_free_QCryptoBlockAmendOptions(amend_options); return ret; -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> This function is currently just a wrapper for bdrv_invalidate_cache(), but in future will contain the code of bdrv_co_invalidate_cache() that has to always be protected by BQL, and leave the rest in the I/O coroutine. Replace all bdrv_invalidate_cache() invokations with bdrv_activate(). Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Reviewed-by: Hanna Reitz <hreitz@redhat.com> Message-Id: <20220209105452.1694545-4-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block.h | 1 + block.c | 7 ++++++- block/block-backend.c | 2 +- block/export/export.c | 2 +- block/parallels.c | 2 +- tests/unit/test-block-iothread.c | 2 +- 6 files changed, 11 insertions(+), 5 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 @@ int bdrv_co_ioctl(BlockDriverState *bs, int req, void *buf); int generated_co_wrapper bdrv_invalidate_cache(BlockDriverState *bs, Error **errp); void bdrv_invalidate_cache_all(Error **errp); +int bdrv_activate(BlockDriverState *bs, Error **errp); int bdrv_inactivate_all(void); /* Ensure contents are flushed to disk. */ diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ void bdrv_init_with_whitelist(void) bdrv_init(); } +int bdrv_activate(BlockDriverState *bs, Error **errp) +{ + return bdrv_invalidate_cache(bs, errp); +} + int coroutine_fn bdrv_co_invalidate_cache(BlockDriverState *bs, Error **errp) { BdrvChild *child, *parent; @@ -XXX,XX +XXX,XX @@ void bdrv_invalidate_cache_all(Error **errp) int ret; aio_context_acquire(aio_context); - ret = bdrv_invalidate_cache(bs, errp); + ret = bdrv_activate(bs, errp); aio_context_release(aio_context); if (ret < 0) { bdrv_next_cleanup(&it); diff --git a/block/block-backend.c b/block/block-backend.c index XXXXXXX..XXXXXXX 100644 --- a/block/block-backend.c +++ b/block/block-backend.c @@ -XXX,XX +XXX,XX @@ void blk_invalidate_cache(BlockBackend *blk, Error **errp) return; } - bdrv_invalidate_cache(bs, errp); + bdrv_activate(bs, errp); } bool blk_is_inserted(BlockBackend *blk) diff --git a/block/export/export.c b/block/export/export.c index XXXXXXX..XXXXXXX 100644 --- a/block/export/export.c +++ b/block/export/export.c @@ -XXX,XX +XXX,XX @@ BlockExport *blk_exp_add(BlockExportOptions *export, Error **errp) * access since the export could be available before migration handover. * ctx was acquired in the caller. */ - bdrv_invalidate_cache(bs, NULL); + bdrv_activate(bs, NULL); perm = BLK_PERM_CONSISTENT_READ; if (export->writable) { diff --git a/block/parallels.c b/block/parallels.c index XXXXXXX..XXXXXXX 100644 --- a/block/parallels.c +++ b/block/parallels.c @@ -XXX,XX +XXX,XX @@ static int parallels_open(BlockDriverState *bs, QDict *options, int flags, s->bat_dirty_bmap = bitmap_new(DIV_ROUND_UP(s->header_size, s->bat_dirty_block)); - /* Disable migration until bdrv_invalidate_cache method is added */ + /* Disable migration until bdrv_activate method is added */ error_setg(&s->migration_blocker, "The Parallels format used by node '%s' " "does not support live migration", bdrv_get_device_or_node_name(bs)); diff --git a/tests/unit/test-block-iothread.c b/tests/unit/test-block-iothread.c index XXXXXXX..XXXXXXX 100644 --- a/tests/unit/test-block-iothread.c +++ b/tests/unit/test-block-iothread.c @@ -XXX,XX +XXX,XX @@ static void test_sync_op_check(BdrvChild *c) static void test_sync_op_invalidate_cache(BdrvChild *c) { /* Early success: Image is not inactive */ - bdrv_invalidate_cache(c->bs, NULL); + bdrv_activate(c->bs, NULL); } -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> Following the bdrv_activate renaming, change also the name of the respective callers. bdrv_invalidate_cache_all -> bdrv_activate_all blk_invalidate_cache -> blk_activate test_sync_op_invalidate_cache -> test_sync_op_activate No functional change intended. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Reviewed-by: Hanna Reitz <hreitz@redhat.com> Message-Id: <20220209105452.1694545-5-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block.h | 2 +- include/sysemu/block-backend.h | 2 +- block.c | 2 +- block/block-backend.c | 2 +- hw/block/pflash_cfi01.c | 2 +- hw/nvram/spapr_nvram.c | 2 +- migration/block.c | 2 +- migration/migration.c | 14 +++++++------- migration/savevm.c | 6 +++--- monitor/qmp-cmds.c | 2 +- tests/unit/test-block-iothread.c | 6 +++--- 11 files changed, 21 insertions(+), 21 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 @@ int bdrv_co_ioctl(BlockDriverState *bs, int req, void *buf); /* Invalidate any cached metadata used by image formats */ int generated_co_wrapper bdrv_invalidate_cache(BlockDriverState *bs, Error **errp); -void bdrv_invalidate_cache_all(Error **errp); int bdrv_activate(BlockDriverState *bs, Error **errp); +void bdrv_activate_all(Error **errp); int bdrv_inactivate_all(void); /* Ensure contents are flushed to disk. */ 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 @@ bool blk_is_writable(BlockBackend *blk); bool blk_is_sg(BlockBackend *blk); bool blk_enable_write_cache(BlockBackend *blk); void blk_set_enable_write_cache(BlockBackend *blk, bool wce); -void blk_invalidate_cache(BlockBackend *blk, Error **errp); +void blk_activate(BlockBackend *blk, Error **errp); bool blk_is_inserted(BlockBackend *blk); bool blk_is_available(BlockBackend *blk); void blk_lock_medium(BlockBackend *blk, bool locked); diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_invalidate_cache(BlockDriverState *bs, Error **errp) return 0; } -void bdrv_invalidate_cache_all(Error **errp) +void bdrv_activate_all(Error **errp) { BlockDriverState *bs; BdrvNextIterator it; diff --git a/block/block-backend.c b/block/block-backend.c index XXXXXXX..XXXXXXX 100644 --- a/block/block-backend.c +++ b/block/block-backend.c @@ -XXX,XX +XXX,XX @@ void blk_set_enable_write_cache(BlockBackend *blk, bool wce) blk->enable_write_cache = wce; } -void blk_invalidate_cache(BlockBackend *blk, Error **errp) +void blk_activate(BlockBackend *blk, Error **errp) { BlockDriverState *bs = blk_bs(blk); diff --git a/hw/block/pflash_cfi01.c b/hw/block/pflash_cfi01.c index XXXXXXX..XXXXXXX 100644 --- a/hw/block/pflash_cfi01.c +++ b/hw/block/pflash_cfi01.c @@ -XXX,XX +XXX,XX @@ static void postload_update_cb(void *opaque, bool running, RunState state) { PFlashCFI01 *pfl = opaque; - /* This is called after bdrv_invalidate_cache_all. */ + /* This is called after bdrv_activate_all. */ qemu_del_vm_change_state_handler(pfl->vmstate); pfl->vmstate = NULL; diff --git a/hw/nvram/spapr_nvram.c b/hw/nvram/spapr_nvram.c index XXXXXXX..XXXXXXX 100644 --- a/hw/nvram/spapr_nvram.c +++ b/hw/nvram/spapr_nvram.c @@ -XXX,XX +XXX,XX @@ static void postload_update_cb(void *opaque, bool running, RunState state) { SpaprNvram *nvram = opaque; - /* This is called after bdrv_invalidate_cache_all. */ + /* This is called after bdrv_activate_all. */ qemu_del_vm_change_state_handler(nvram->vmstate); nvram->vmstate = NULL; diff --git a/migration/block.c b/migration/block.c index XXXXXXX..XXXXXXX 100644 --- a/migration/block.c +++ b/migration/block.c @@ -XXX,XX +XXX,XX @@ static int block_load(QEMUFile *f, void *opaque, int version_id) return -EINVAL; } - blk_invalidate_cache(blk, &local_err); + blk_activate(blk, &local_err); if (local_err) { error_report_err(local_err); return -EINVAL; diff --git a/migration/migration.c b/migration/migration.c index XXXXXXX..XXXXXXX 100644 --- a/migration/migration.c +++ b/migration/migration.c @@ -XXX,XX +XXX,XX @@ static void process_incoming_migration_bh(void *opaque) if (!migrate_late_block_activate() || (autostart && (!global_state_received() || global_state_get_runstate() == RUN_STATE_RUNNING))) { - /* Make sure all file formats flush their mutable metadata. + /* Make sure all file formats throw away their mutable metadata. * If we get an error here, just don't restart the VM yet. */ - bdrv_invalidate_cache_all(&local_err); + bdrv_activate_all(&local_err); if (local_err) { error_report_err(local_err); local_err = NULL; @@ -XXX,XX +XXX,XX @@ static void process_incoming_migration_co(void *opaque) /* we get COLO info, and know if we are in COLO mode */ if (!ret && migration_incoming_colo_enabled()) { - /* Make sure all file formats flush their mutable metadata */ - bdrv_invalidate_cache_all(&local_err); + /* Make sure all file formats throw away their mutable metadata */ + bdrv_activate_all(&local_err); if (local_err) { error_report_err(local_err); goto fail; @@ -XXX,XX +XXX,XX @@ static void migrate_fd_cancel(MigrationState *s) if (s->state == MIGRATION_STATUS_CANCELLING && s->block_inactive) { Error *local_err = NULL; - bdrv_invalidate_cache_all(&local_err); + bdrv_activate_all(&local_err); if (local_err) { error_report_err(local_err); } else { @@ -XXX,XX +XXX,XX @@ fail: */ Error *local_err = NULL; - bdrv_invalidate_cache_all(&local_err); + bdrv_activate_all(&local_err); if (local_err) { error_report_err(local_err); } @@ -XXX,XX +XXX,XX @@ fail_invalidate: Error *local_err = NULL; qemu_mutex_lock_iothread(); - bdrv_invalidate_cache_all(&local_err); + bdrv_activate_all(&local_err); if (local_err) { error_report_err(local_err); } else { diff --git a/migration/savevm.c b/migration/savevm.c index XXXXXXX..XXXXXXX 100644 --- a/migration/savevm.c +++ b/migration/savevm.c @@ -XXX,XX +XXX,XX @@ int qemu_savevm_state_complete_precopy_non_iterable(QEMUFile *f, if (inactivate_disks) { /* Inactivate before sending QEMU_VM_EOF so that the - * bdrv_invalidate_cache_all() on the other end won't fail. */ + * bdrv_activate_all() on the other end won't fail. */ ret = bdrv_inactivate_all(); if (ret) { error_report("%s: bdrv_inactivate_all() failed (%d)", @@ -XXX,XX +XXX,XX @@ static void loadvm_postcopy_handle_run_bh(void *opaque) trace_loadvm_postcopy_handle_run_bh("after announce"); - /* Make sure all file formats flush their mutable metadata. + /* Make sure all file formats throw away their mutable metadata. * If we get an error here, just don't restart the VM yet. */ - bdrv_invalidate_cache_all(&local_err); + bdrv_activate_all(&local_err); if (local_err) { error_report_err(local_err); local_err = NULL; diff --git a/monitor/qmp-cmds.c b/monitor/qmp-cmds.c index XXXXXXX..XXXXXXX 100644 --- a/monitor/qmp-cmds.c +++ b/monitor/qmp-cmds.c @@ -XXX,XX +XXX,XX @@ void qmp_cont(Error **errp) * If there are no inactive block nodes (e.g. because the VM was just * paused rather than completing a migration), bdrv_inactivate_all() simply * doesn't do anything. */ - bdrv_invalidate_cache_all(&local_err); + bdrv_activate_all(&local_err); if (local_err) { error_propagate(errp, local_err); return; diff --git a/tests/unit/test-block-iothread.c b/tests/unit/test-block-iothread.c index XXXXXXX..XXXXXXX 100644 --- a/tests/unit/test-block-iothread.c +++ b/tests/unit/test-block-iothread.c @@ -XXX,XX +XXX,XX @@ static void test_sync_op_check(BdrvChild *c) g_assert_cmpint(ret, ==, -ENOTSUP); } -static void test_sync_op_invalidate_cache(BdrvChild *c) +static void test_sync_op_activate(BdrvChild *c) { /* Early success: Image is not inactive */ bdrv_activate(c->bs, NULL); @@ -XXX,XX +XXX,XX @@ const SyncOpTest sync_op_tests[] = { .name = "/sync-op/check", .fn = test_sync_op_check, }, { - .name = "/sync-op/invalidate_cache", - .fn = test_sync_op_invalidate_cache, + .name = "/sync-op/activate", + .fn = test_sync_op_activate, }, }; -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> Split bdrv_co_invalidate cache in two: the Global State (under BQL) code that takes care of permissions and running GS callbacks, and leave only the I/O code (->bdrv_co_invalidate_cache) running in the I/O coroutine. The only side effect is that bdrv_co_invalidate_cache is not recursive anymore, and so is every direct call to bdrv_invalidate_cache(). Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Message-Id: <20220209105452.1694545-6-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block.c | 37 +++++++++++++++++++++++-------------- 1 file changed, 23 insertions(+), 14 deletions(-) diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ void bdrv_init_with_whitelist(void) } int bdrv_activate(BlockDriverState *bs, Error **errp) -{ - return bdrv_invalidate_cache(bs, errp); -} - -int coroutine_fn bdrv_co_invalidate_cache(BlockDriverState *bs, Error **errp) { BdrvChild *child, *parent; Error *local_err = NULL; @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_invalidate_cache(BlockDriverState *bs, Error **errp) } QLIST_FOREACH(child, &bs->children, next) { - bdrv_co_invalidate_cache(child->bs, &local_err); + bdrv_activate(child->bs, &local_err); if (local_err) { error_propagate(errp, local_err); return -EINVAL; @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_invalidate_cache(BlockDriverState *bs, Error **errp) * Note that the required permissions of inactive images are always a * subset of the permissions required after activating the image. This * allows us to just get the permissions upfront without restricting - * drv->bdrv_invalidate_cache(). + * bdrv_co_invalidate_cache(). * * It also means that in error cases, we don't have to try and revert to * the old permissions (which is an operation that could fail, too). We can @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_invalidate_cache(BlockDriverState *bs, Error **errp) return ret; } - if (bs->drv->bdrv_co_invalidate_cache) { - bs->drv->bdrv_co_invalidate_cache(bs, &local_err); - if (local_err) { - bs->open_flags |= BDRV_O_INACTIVE; - error_propagate(errp, local_err); - return -EINVAL; - } + ret = bdrv_invalidate_cache(bs, errp); + if (ret < 0) { + bs->open_flags |= BDRV_O_INACTIVE; + return ret; } FOR_EACH_DIRTY_BITMAP(bs, bm) { @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_invalidate_cache(BlockDriverState *bs, Error **errp) return 0; } +int coroutine_fn bdrv_co_invalidate_cache(BlockDriverState *bs, Error **errp) +{ + Error *local_err = NULL; + + assert(!(bs->open_flags & BDRV_O_INACTIVE)); + + if (bs->drv->bdrv_co_invalidate_cache) { + bs->drv->bdrv_co_invalidate_cache(bs, &local_err); + if (local_err) { + error_propagate(errp, local_err); + return -EINVAL; + } + } + + return 0; +} + void bdrv_activate_all(Error **errp) { BlockDriverState *bs; -- 2.35.1 From: Stefan Hajnoczi <stefanha@redhat.com> Compiler optimizations can cache TLS values across coroutine yield points, resulting in stale values from the previous thread when a coroutine is re-entered by a new thread. Serge Guelton developed an __attribute__((noinline)) wrapper and tested it with clang and gcc. I formatted his idea according to QEMU's coding style and wrote documentation. The compiler can still optimize based on analyzing noinline code, so an asm volatile barrier with an output constraint is required to prevent unwanted optimizations. Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=1952483 Suggested-by: Serge Guelton <sguelton@redhat.com> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Message-Id: <20220222140150.27240-2-stefanha@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/qemu/coroutine-tls.h | 165 +++++++++++++++++++++++++++++++++++ 1 file changed, 165 insertions(+) create mode 100644 include/qemu/coroutine-tls.h diff --git a/include/qemu/coroutine-tls.h b/include/qemu/coroutine-tls.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/qemu/coroutine-tls.h @@ -XXX,XX +XXX,XX @@ +/* + * QEMU Thread Local Storage for coroutines + * + * Copyright Red Hat + * + * SPDX-License-Identifier: LGPL-2.1-or-later + * + * This work is licensed under the terms of the GNU LGPL, version 2.1 or later. + * See the COPYING.LIB file in the top-level directory. + * + * It is forbidden to access Thread Local Storage in coroutines because + * compiler optimizations may cause values to be cached across coroutine + * re-entry. Coroutines can run in more than one thread through the course of + * their life, leading bugs when stale TLS values from the wrong thread are + * used as a result of compiler optimization. + * + * An example is: + * + * ..code-block:: c + * :caption: A coroutine that may see the wrong TLS value + * + * static __thread AioContext *current_aio_context; + * ... + * static void coroutine_fn foo(void) + * { + * aio_notify(current_aio_context); + * qemu_coroutine_yield(); + * aio_notify(current_aio_context); // <-- may be stale after yielding! + * } + * + * This header provides macros for safely defining variables in Thread Local + * Storage: + * + * ..code-block:: c + * :caption: A coroutine that safely uses TLS + * + * QEMU_DEFINE_STATIC_CO_TLS(AioContext *, current_aio_context) + * ... + * static void coroutine_fn foo(void) + * { + * aio_notify(get_current_aio_context()); + * qemu_coroutine_yield(); + * aio_notify(get_current_aio_context()); // <-- safe + * } + */ + +#ifndef QEMU_COROUTINE_TLS_H +#define QEMU_COROUTINE_TLS_H + +/* + * To stop the compiler from caching TLS values we define accessor functions + * with __attribute__((noinline)) plus asm volatile("") to prevent + * optimizations that override noinline. + * + * The compiler can still analyze noinline code and make optimizations based on + * that knowledge, so an inline asm output operand is used to prevent + * optimizations that make assumptions about the address of the TLS variable. + * + * This is fragile and ultimately needs to be solved by a mechanism that is + * guaranteed to work by the compiler (e.g. stackless coroutines), but for now + * we use this approach to prevent issues. + */ + +/** + * QEMU_DECLARE_CO_TLS: + * @type: the variable's C type + * @var: the variable name + * + * Declare an extern variable in Thread Local Storage from a header file: + * + * .. code-block:: c + * :caption: Declaring an extern variable in Thread Local Storage + * + * QEMU_DECLARE_CO_TLS(int, my_count) + * ... + * int c = get_my_count(); + * set_my_count(c + 1); + * *get_ptr_my_count() = 0; + * + * This is a coroutine-safe replacement for the __thread keyword and is + * equivalent to the following code: + * + * .. code-block:: c + * :caption: Declaring a TLS variable using __thread + * + * extern __thread int my_count; + * ... + * int c = my_count; + * my_count = c + 1; + * *(&my_count) = 0; + */ +#define QEMU_DECLARE_CO_TLS(type, var) \ + __attribute__((noinline)) type get_##var(void); \ + __attribute__((noinline)) void set_##var(type v); \ + __attribute__((noinline)) type *get_ptr_##var(void); + +/** + * QEMU_DEFINE_CO_TLS: + * @type: the variable's C type + * @var: the variable name + * + * Define a variable in Thread Local Storage that was previously declared from + * a header file with QEMU_DECLARE_CO_TLS(): + * + * .. code-block:: c + * :caption: Defining a variable in Thread Local Storage + * + * QEMU_DEFINE_CO_TLS(int, my_count) + * + * This is a coroutine-safe replacement for the __thread keyword and is + * equivalent to the following code: + * + * .. code-block:: c + * :caption: Defining a TLS variable using __thread + * + * __thread int my_count; + */ +#define QEMU_DEFINE_CO_TLS(type, var) \ + static __thread type co_tls_##var; \ + type get_##var(void) { asm volatile(""); return co_tls_##var; } \ + void set_##var(type v) { asm volatile(""); co_tls_##var = v; } \ + type *get_ptr_##var(void) \ + { type *ptr = &co_tls_##var; asm volatile("" : "+rm" (ptr)); return ptr; } + +/** + * QEMU_DEFINE_STATIC_CO_TLS: + * @type: the variable's C type + * @var: the variable name + * + * Define a static variable in Thread Local Storage: + * + * .. code-block:: c + * :caption: Defining a static variable in Thread Local Storage + * + * QEMU_DEFINE_STATIC_CO_TLS(int, my_count) + * ... + * int c = get_my_count(); + * set_my_count(c + 1); + * *get_ptr_my_count() = 0; + * + * This is a coroutine-safe replacement for the __thread keyword and is + * equivalent to the following code: + * + * .. code-block:: c + * :caption: Defining a static TLS variable using __thread + * + * static __thread int my_count; + * ... + * int c = my_count; + * my_count = c + 1; + * *(&my_count) = 0; + */ +#define QEMU_DEFINE_STATIC_CO_TLS(type, var) \ + static __thread type co_tls_##var; \ + static __attribute__((noinline, unused)) \ + type get_##var(void) \ + { asm volatile(""); return co_tls_##var; } \ + static __attribute__((noinline, unused)) \ + void set_##var(type v) \ + { asm volatile(""); co_tls_##var = v; } \ + static __attribute__((noinline, unused)) \ + type *get_ptr_##var(void) \ + { type *ptr = &co_tls_##var; asm volatile("" : "+rm" (ptr)); return ptr; } + +#endif /* QEMU_COROUTINE_TLS_H */ -- 2.35.1 From: Stefan Hajnoczi <stefanha@redhat.com> QEMU TLS macros must be used to make TLS variables safe with coroutines. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Message-Id: <20220222140150.27240-3-stefanha@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- util/async.c | 12 +++++++----- 1 file changed, 7 insertions(+), 5 deletions(-) diff --git a/util/async.c b/util/async.c index XXXXXXX..XXXXXXX 100644 --- a/util/async.c +++ b/util/async.c @@ -XXX,XX +XXX,XX @@ #include "qemu/rcu_queue.h" #include "block/raw-aio.h" #include "qemu/coroutine_int.h" +#include "qemu/coroutine-tls.h" #include "trace.h" /***********************************************************/ @@ -XXX,XX +XXX,XX @@ void aio_context_release(AioContext *ctx) qemu_rec_mutex_unlock(&ctx->lock); } -static __thread AioContext *my_aiocontext; +QEMU_DEFINE_STATIC_CO_TLS(AioContext *, my_aiocontext) AioContext *qemu_get_current_aio_context(void) { - if (my_aiocontext) { - return my_aiocontext; + AioContext *ctx = get_my_aiocontext(); + if (ctx) { + return ctx; } if (qemu_mutex_iothread_locked()) { /* Possibly in a vCPU thread. */ @@ -XXX,XX +XXX,XX @@ AioContext *qemu_get_current_aio_context(void) void qemu_set_current_aio_context(AioContext *ctx) { - assert(!my_aiocontext); - my_aiocontext = ctx; + assert(!get_my_aiocontext()); + set_my_aiocontext(ctx); } -- 2.35.1 From: Stefan Hajnoczi <stefanha@redhat.com> RCU may be used from coroutines. Standard __thread variables cannot be used by coroutines. Use the coroutine TLS macros instead. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Message-Id: <20220222140150.27240-4-stefanha@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/qemu/rcu.h | 7 ++++--- tests/unit/rcutorture.c | 10 +++++----- tests/unit/test-rcu-list.c | 4 ++-- util/rcu.c | 10 +++++----- 4 files changed, 16 insertions(+), 15 deletions(-) diff --git a/include/qemu/rcu.h b/include/qemu/rcu.h index XXXXXXX..XXXXXXX 100644 --- a/include/qemu/rcu.h +++ b/include/qemu/rcu.h @@ -XXX,XX +XXX,XX @@ #include "qemu/atomic.h" #include "qemu/notify.h" #include "qemu/sys_membarrier.h" +#include "qemu/coroutine-tls.h" #ifdef __cplusplus extern "C" { @@ -XXX,XX +XXX,XX @@ struct rcu_reader_data { NotifierList force_rcu; }; -extern __thread struct rcu_reader_data rcu_reader; +QEMU_DECLARE_CO_TLS(struct rcu_reader_data, rcu_reader) static inline void rcu_read_lock(void) { - struct rcu_reader_data *p_rcu_reader = &rcu_reader; + struct rcu_reader_data *p_rcu_reader = get_ptr_rcu_reader(); unsigned ctr; if (p_rcu_reader->depth++ > 0) { @@ -XXX,XX +XXX,XX @@ static inline void rcu_read_lock(void) static inline void rcu_read_unlock(void) { - struct rcu_reader_data *p_rcu_reader = &rcu_reader; + struct rcu_reader_data *p_rcu_reader = get_ptr_rcu_reader(); assert(p_rcu_reader->depth != 0); if (--p_rcu_reader->depth > 0) { diff --git a/tests/unit/rcutorture.c b/tests/unit/rcutorture.c index XXXXXXX..XXXXXXX 100644 --- a/tests/unit/rcutorture.c +++ b/tests/unit/rcutorture.c @@ -XXX,XX +XXX,XX @@ static void *rcu_read_perf_test(void *arg) rcu_register_thread(); - *(struct rcu_reader_data **)arg = &rcu_reader; + *(struct rcu_reader_data **)arg = get_ptr_rcu_reader(); qatomic_inc(&nthreadsrunning); while (goflag == GOFLAG_INIT) { g_usleep(1000); @@ -XXX,XX +XXX,XX @@ static void *rcu_update_perf_test(void *arg) rcu_register_thread(); - *(struct rcu_reader_data **)arg = &rcu_reader; + *(struct rcu_reader_data **)arg = get_ptr_rcu_reader(); qatomic_inc(&nthreadsrunning); while (goflag == GOFLAG_INIT) { g_usleep(1000); @@ -XXX,XX +XXX,XX @@ static void *rcu_read_stress_test(void *arg) rcu_register_thread(); - *(struct rcu_reader_data **)arg = &rcu_reader; + *(struct rcu_reader_data **)arg = get_ptr_rcu_reader(); while (goflag == GOFLAG_INIT) { g_usleep(1000); } @@ -XXX,XX +XXX,XX @@ static void *rcu_update_stress_test(void *arg) struct rcu_stress *cp = qatomic_read(&rcu_stress_current); rcu_register_thread(); - *(struct rcu_reader_data **)arg = &rcu_reader; + *(struct rcu_reader_data **)arg = get_ptr_rcu_reader(); while (goflag == GOFLAG_INIT) { g_usleep(1000); @@ -XXX,XX +XXX,XX @@ static void *rcu_fake_update_stress_test(void *arg) { rcu_register_thread(); - *(struct rcu_reader_data **)arg = &rcu_reader; + *(struct rcu_reader_data **)arg = get_ptr_rcu_reader(); while (goflag == GOFLAG_INIT) { g_usleep(1000); } diff --git a/tests/unit/test-rcu-list.c b/tests/unit/test-rcu-list.c index XXXXXXX..XXXXXXX 100644 --- a/tests/unit/test-rcu-list.c +++ b/tests/unit/test-rcu-list.c @@ -XXX,XX +XXX,XX @@ static void *rcu_q_reader(void *arg) rcu_register_thread(); - *(struct rcu_reader_data **)arg = &rcu_reader; + *(struct rcu_reader_data **)arg = get_ptr_rcu_reader(); qatomic_inc(&nthreadsrunning); while (qatomic_read(&goflag) == GOFLAG_INIT) { g_usleep(1000); @@ -XXX,XX +XXX,XX @@ static void *rcu_q_updater(void *arg) long long n_removed_local = 0; struct list_element *el, *prev_el; - *(struct rcu_reader_data **)arg = &rcu_reader; + *(struct rcu_reader_data **)arg = get_ptr_rcu_reader(); qatomic_inc(&nthreadsrunning); while (qatomic_read(&goflag) == GOFLAG_INIT) { g_usleep(1000); diff --git a/util/rcu.c b/util/rcu.c index XXXXXXX..XXXXXXX 100644 --- a/util/rcu.c +++ b/util/rcu.c @@ -XXX,XX +XXX,XX @@ static inline int rcu_gp_ongoing(unsigned long *ctr) /* Written to only by each individual reader. Read by both the reader and the * writers. */ -__thread struct rcu_reader_data rcu_reader; +QEMU_DEFINE_CO_TLS(struct rcu_reader_data, rcu_reader) /* Protected by rcu_registry_lock. */ typedef QLIST_HEAD(, rcu_reader_data) ThreadList; @@ -XXX,XX +XXX,XX @@ void drain_call_rcu(void) void rcu_register_thread(void) { - assert(rcu_reader.ctr == 0); + assert(get_ptr_rcu_reader()->ctr == 0); qemu_mutex_lock(&rcu_registry_lock); - QLIST_INSERT_HEAD(&registry, &rcu_reader, node); + QLIST_INSERT_HEAD(&registry, get_ptr_rcu_reader(), node); qemu_mutex_unlock(&rcu_registry_lock); } void rcu_unregister_thread(void) { qemu_mutex_lock(&rcu_registry_lock); - QLIST_REMOVE(&rcu_reader, node); + QLIST_REMOVE(get_ptr_rcu_reader(), node); qemu_mutex_unlock(&rcu_registry_lock); } void rcu_add_force_rcu_notifier(Notifier *n) { qemu_mutex_lock(&rcu_registry_lock); - notifier_list_add(&rcu_reader.force_rcu, n); + notifier_list_add(&get_ptr_rcu_reader()->force_rcu, n); qemu_mutex_unlock(&rcu_registry_lock); } -- 2.35.1 From: Stefan Hajnoczi <stefanha@redhat.com> qemu_mutex_iothread_locked() may be used from coroutines. Standard __thread variables cannot be used by coroutines. Use the coroutine TLS macros instead. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Message-Id: <20220222140150.27240-5-stefanha@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- softmmu/cpus.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/softmmu/cpus.c b/softmmu/cpus.c index XXXXXXX..XXXXXXX 100644 --- a/softmmu/cpus.c +++ b/softmmu/cpus.c @@ -XXX,XX +XXX,XX @@ #include "qemu/osdep.h" #include "qemu-common.h" #include "monitor/monitor.h" +#include "qemu/coroutine-tls.h" #include "qapi/error.h" #include "qapi/qapi-commands-machine.h" #include "qapi/qapi-commands-misc.h" @@ -XXX,XX +XXX,XX @@ bool qemu_in_vcpu_thread(void) return current_cpu && qemu_cpu_is_self(current_cpu); } -static __thread bool iothread_locked = false; +QEMU_DEFINE_STATIC_CO_TLS(bool, iothread_locked) bool qemu_mutex_iothread_locked(void) { - return iothread_locked; + return get_iothread_locked(); } /* @@ -XXX,XX +XXX,XX @@ void qemu_mutex_lock_iothread_impl(const char *file, int line) g_assert(!qemu_mutex_iothread_locked()); bql_lock(&qemu_global_mutex, file, line); - iothread_locked = true; + set_iothread_locked(true); } void qemu_mutex_unlock_iothread(void) { g_assert(qemu_mutex_iothread_locked()); - iothread_locked = false; + set_iothread_locked(false); qemu_mutex_unlock(&qemu_global_mutex); } -- 2.35.1 From: Hanna Reitz <hreitz@redhat.com> The daemonizing functions in os-posix (os_daemonize() and os_setup_post()) only daemonize the process if the static `daemonize` variable is set. Right now, it can only be set by os_parse_cmd_args(). In order to use os_daemonize() and os_setup_post() from the storage daemon to have it be daemonized, we need some other way to set this `daemonize` variable, because I would rather not tap into the system emulator's arg-parsing code. Therefore, this patch adds an os_set_daemonize() function, which will return an error on os-win32 (because daemonizing is not supported there). Signed-off-by: Hanna Reitz <hreitz@redhat.com> Message-Id: <20220303164814.284974-2-hreitz@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/sysemu/os-posix.h | 1 + include/sysemu/os-win32.h | 8 ++++++++ os-posix.c | 6 ++++++ 3 files changed, 15 insertions(+) diff --git a/include/sysemu/os-posix.h b/include/sysemu/os-posix.h index XXXXXXX..XXXXXXX 100644 --- a/include/sysemu/os-posix.h +++ b/include/sysemu/os-posix.h @@ -XXX,XX +XXX,XX @@ int os_mlock(void); typedef struct timeval qemu_timeval; #define qemu_gettimeofday(tp) gettimeofday(tp, NULL) +int os_set_daemonize(bool d); bool is_daemonized(void); /** diff --git a/include/sysemu/os-win32.h b/include/sysemu/os-win32.h index XXXXXXX..XXXXXXX 100644 --- a/include/sysemu/os-win32.h +++ b/include/sysemu/os-win32.h @@ -XXX,XX +XXX,XX @@ typedef struct { } qemu_timeval; int qemu_gettimeofday(qemu_timeval *tp); +static inline int os_set_daemonize(bool d) +{ + if (d) { + return -ENOTSUP; + } + return 0; +} + static inline bool is_daemonized(void) { return false; diff --git a/os-posix.c b/os-posix.c index XXXXXXX..XXXXXXX 100644 --- a/os-posix.c +++ b/os-posix.c @@ -XXX,XX +XXX,XX @@ bool is_daemonized(void) return daemonize; } +int os_set_daemonize(bool d) +{ + daemonize = d; + return 0; +} + int os_mlock(void) { #ifdef HAVE_MLOCKALL -- 2.35.1 From: Hanna Reitz <hreitz@redhat.com> In contrast to qemu-nbd (where it is called --fork) and the system emulator, QSD does not have a --daemonize switch yet. Just like them, QSD allows setting up block devices and exports on the command line. When doing so, it is often necessary for whoever invoked the QSD to wait until these exports are fully set up. A --daemonize switch allows precisely this, by virtue of the parent process exiting once everything is set up. Note that there are alternative ways of waiting for all exports to be set up, for example: - Passing the --pidfile option and waiting until the respective file exists (but I do not know if there is a way of implementing this without a busy wait loop) - Set up some network server (e.g. on a Unix socket) and have the QSD connect to it after all arguments have been processed by appending corresponding --chardev and --monitor options to the command line, and then wait until the QSD connects Having a --daemonize option would make this simpler, though, without having to rely on additional tools (to set up a network server) or busy waiting. Implementing a --daemonize switch means having to fork the QSD process. Ideally, we should do this as early as possible: All the parent process has to do is to wait for the child process to signal completion of its set-up phase, and therefore there is basically no initialization that needs to be done before the fork. On the other hand, forking after initialization steps means having to consider how those steps (like setting up the block layer or QMP) interact with a later fork, which is often not trivial. In order to fork this early, we must scan the command line for --daemonize long before our current process_options() call. Instead of adding custom new code to do so, just reuse process_options() and give it a @pre_init_pass argument to distinguish the two passes. I believe there are some other switches but --daemonize that deserve parsing in the first pass: - --help and --version are supposed to only print some text and then immediately exit (so any initialization we do would be for naught). This changes behavior, because now "--blockdev inv-drv --help" will print a help text instead of complaining about the --blockdev argument. Note that this is similar in behavior to other tools, though: "--help" is generally immediately acted upon when finding it in the argument list, potentially before other arguments (even ones before it) are acted on. For example, "ls /does-not-exist --help" prints a help text and does not complain about ENOENT. - --pidfile does not need initialization, and is already exempted from the sequential order that process_options() claims to strictly follow (the PID file is only created after all arguments are processed, not at the time the --pidfile argument appears), so it makes sense to include it in the same category as --daemonize. - Invalid arguments should always be reported as soon as possible. (The same caveat with --help applies: That means that "--blockdev inv-drv --inv-arg" will now complain about --inv-arg, not inv-drv.) This patch does make some references to --daemonize without having implemented it yet, but that will happen in the next patch. Signed-off-by: Hanna Reitz <hreitz@redhat.com> Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com> Message-Id: <20220303164814.284974-3-hreitz@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- storage-daemon/qemu-storage-daemon.c | 43 ++++++++++++++++++++++++---- 1 file changed, 38 insertions(+), 5 deletions(-) diff --git a/storage-daemon/qemu-storage-daemon.c b/storage-daemon/qemu-storage-daemon.c index XXXXXXX..XXXXXXX 100644 --- a/storage-daemon/qemu-storage-daemon.c +++ b/storage-daemon/qemu-storage-daemon.c @@ -XXX,XX +XXX,XX @@ static int getopt_set_loc(int argc, char **argv, const char *optstring, return c; } -static void process_options(int argc, char *argv[]) +/** + * Process QSD command-line arguments. + * + * This is done in two passes: + * + * First (@pre_init_pass is true), we do a pass where all global + * arguments pertaining to the QSD process (like --help or --daemonize) + * are processed. This pass is done before most of the QEMU-specific + * initialization steps (e.g. initializing the block layer or QMP), and + * so must only process arguments that are not really QEMU-specific. + * + * Second (@pre_init_pass is false), we (sequentially) process all + * QEMU/QSD-specific arguments. Many of these arguments are effectively + * translated to QMP commands (like --blockdev for blockdev-add, or + * --export for block-export-add). + */ +static void process_options(int argc, char *argv[], bool pre_init_pass) { int c; @@ -XXX,XX +XXX,XX @@ static void process_options(int argc, char *argv[]) }; /* - * In contrast to the system emulator, options are processed in the order - * they are given on the command lines. This means that things must be - * defined first before they can be referenced in another option. + * In contrast to the system emulator, QEMU-specific options are processed + * in the order they are given on the command lines. This means that things + * must be defined first before they can be referenced in another option. */ + optind = 1; while ((c = getopt_set_loc(argc, argv, "-hT:V", long_options)) != -1) { + bool handle_option_pre_init; + + /* Should this argument be processed in the pre-init pass? */ + handle_option_pre_init = + c == '?' || + c == 'h' || + c == 'V' || + c == OPTION_PIDFILE; + + /* Process every option only in its respective pass */ + if (pre_init_pass != handle_option_pre_init) { + continue; + } + switch (c) { case '?': exit(EXIT_FAILURE); @@ -XXX,XX +XXX,XX @@ int main(int argc, char *argv[]) qemu_init_exec_dir(argv[0]); os_setup_signal_handling(); + process_options(argc, argv, true); + module_call_init(MODULE_INIT_QOM); module_call_init(MODULE_INIT_TRACE); qemu_add_opts(&qemu_trace_opts); @@ -XXX,XX +XXX,XX @@ int main(int argc, char *argv[]) qemu_set_log(LOG_TRACE); qemu_init_main_loop(&error_fatal); - process_options(argc, argv); + process_options(argc, argv, false); /* * Write the pid file after creating chardevs, exports, and NBD servers but -- 2.35.1 From: Hanna Reitz <hreitz@redhat.com> To implement this, we reuse the existing daemonizing functions from the system emulator, which mainly do the following: - Fork off a child process, and set up a pipe between parent and child - The parent process waits until the child sends a status byte over the pipe (0 means that the child was set up successfully; anything else (including errors or EOF) means that the child was not set up successfully), and then exits with an appropriate exit status - The child process enters a new session (forking off again), changes the umask, and will ignore terminal signals from then on - Once set-up is complete, the child will chdir to /, redirect all standard I/O streams to /dev/null, and tell the parent that set-up has been completed successfully In contrast to qemu-nbd's --fork implementation, during the set up phase, error messages are not piped through the parent process. qemu-nbd mainly does this to detect errors, though (while os_daemonize() has the child explicitly signal success after set up); because we do not redirect stderr after forking, error messages continue to appear on whatever the parent's stderr was (until set up is complete). Signed-off-by: Hanna Reitz <hreitz@redhat.com> Message-Id: <20220303164814.284974-4-hreitz@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- docs/tools/qemu-storage-daemon.rst | 7 +++++++ storage-daemon/qemu-storage-daemon.c | 15 +++++++++++++++ 2 files changed, 22 insertions(+) diff --git a/docs/tools/qemu-storage-daemon.rst b/docs/tools/qemu-storage-daemon.rst index XXXXXXX..XXXXXXX 100644 --- a/docs/tools/qemu-storage-daemon.rst +++ b/docs/tools/qemu-storage-daemon.rst @@ -XXX,XX +XXX,XX @@ Standard options: created but before accepting connections. The daemon has started successfully when the pid file is written and clients may begin connecting. +.. option:: --daemonize + + Daemonize the process. The parent process will exit once startup is complete + (i.e., after the pid file has been or would have been written) or failure + occurs. Its exit code reflects whether the child has started up successfully + or failed to do so. + Examples -------- Launch the daemon with QMP monitor socket ``qmp.sock`` so clients can execute diff --git a/storage-daemon/qemu-storage-daemon.c b/storage-daemon/qemu-storage-daemon.c index XXXXXXX..XXXXXXX 100644 --- a/storage-daemon/qemu-storage-daemon.c +++ b/storage-daemon/qemu-storage-daemon.c @@ -XXX,XX +XXX,XX @@ static void help(void) " --chardev <options> configure a character device backend\n" " (see the qemu(1) man page for possible options)\n" "\n" +" --daemonize daemonize the process, and have the parent exit\n" +" once startup is complete\n" +"\n" " --export [type=]nbd,id=<id>,node-name=<node-name>[,name=<export-name>]\n" " [,writable=on|off][,bitmap=<name>]\n" " export the specified block node over NBD\n" @@ -XXX,XX +XXX,XX @@ QEMU_HELP_BOTTOM "\n", enum { OPTION_BLOCKDEV = 256, OPTION_CHARDEV, + OPTION_DAEMONIZE, OPTION_EXPORT, OPTION_MONITOR, OPTION_NBD_SERVER, @@ -XXX,XX +XXX,XX @@ static void process_options(int argc, char *argv[], bool pre_init_pass) static const struct option long_options[] = { {"blockdev", required_argument, NULL, OPTION_BLOCKDEV}, {"chardev", required_argument, NULL, OPTION_CHARDEV}, + {"daemonize", no_argument, NULL, OPTION_DAEMONIZE}, {"export", required_argument, NULL, OPTION_EXPORT}, {"help", no_argument, NULL, 'h'}, {"monitor", required_argument, NULL, OPTION_MONITOR}, @@ -XXX,XX +XXX,XX @@ static void process_options(int argc, char *argv[], bool pre_init_pass) c == '?' || c == 'h' || c == 'V' || + c == OPTION_DAEMONIZE || c == OPTION_PIDFILE; /* Process every option only in its respective pass */ @@ -XXX,XX +XXX,XX @@ static void process_options(int argc, char *argv[], bool pre_init_pass) qemu_opts_del(opts); break; } + case OPTION_DAEMONIZE: + if (os_set_daemonize(true) < 0) { + error_report("--daemonize not supported in this build"); + exit(EXIT_FAILURE); + } + break; case OPTION_EXPORT: { Visitor *v; @@ -XXX,XX +XXX,XX @@ int main(int argc, char *argv[]) process_options(argc, argv, true); + os_daemonize(); + module_call_init(MODULE_INIT_QOM); module_call_init(MODULE_INIT_TRACE); qemu_add_opts(&qemu_trace_opts); @@ -XXX,XX +XXX,XX @@ int main(int argc, char *argv[]) * it. */ pid_file_init(); + os_setup_post(); while (!exit_requested) { main_loop_wait(false); -- 2.35.1 From: Hanna Reitz <hreitz@redhat.com> 185 tests quitting qemu while a block job is active. It does not specifically test quitting qemu while a mirror or active commit job is in its READY phase. Add two test cases for this, where we respectively mirror or commit to an external QSD instance, which provides a throttled block device. qemu is supposed to cancel the job so that it can quit as soon as possible instead of waiting for the job to complete (which it did before 6.2). Signed-off-by: Hanna Reitz <hreitz@redhat.com> Message-Id: <20220303164814.284974-5-hreitz@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- tests/qemu-iotests/185 | 190 ++++++++++++++++++++++++++++++++++++- tests/qemu-iotests/185.out | 48 ++++++++++ 2 files changed, 237 insertions(+), 1 deletion(-) diff --git a/tests/qemu-iotests/185 b/tests/qemu-iotests/185 index XXXXXXX..XXXXXXX 100755 --- a/tests/qemu-iotests/185 +++ b/tests/qemu-iotests/185 @@ -XXX,XX +XXX,XX @@ _cleanup() _rm_test_img "${TEST_IMG}.copy" _cleanup_test_img _cleanup_qemu + + if [ -f "$TEST_DIR/qsd.pid" ]; then + kill -SIGKILL "$(cat "$TEST_DIR/qsd.pid")" + rm -f "$TEST_DIR/qsd.pid" + fi + rm -f "$SOCK_DIR/qsd.sock" } trap "_cleanup; exit \$status" 0 1 2 3 15 @@ -XXX,XX +XXX,XX @@ _supported_fmt qcow2 _supported_proto file _supported_os Linux -size=64M +size=$((64 * 1048576)) TEST_IMG="${TEST_IMG}.base" _make_test_img $size echo @@ -XXX,XX +XXX,XX @@ wait=1 _cleanup_qemu | grep -v 'JOB_STATUS_CHANGE' _check_test_img +echo +echo === Start mirror to throttled QSD and exit qemu === +echo + +# Mirror to a throttled QSD instance (so that qemu cannot drain the +# throttling), wait for READY, then write some data to the device, +# and then quit qemu. +# (qemu should force-cancel the job and not wait for the data to be +# written to the target.) + +_make_test_img $size + +# Will be used by this and the next case +set_up_throttled_qsd() { + $QSD \ + --object throttle-group,id=thrgr,limits.bps-total=1048576 \ + --blockdev null-co,node-name=null,size=$size \ + --blockdev throttle,node-name=throttled,throttle-group=thrgr,file=null \ + --nbd-server addr.type=unix,addr.path="$SOCK_DIR/qsd.sock" \ + --export nbd,id=exp,node-name=throttled,name=target,writable=true \ + --pidfile "$TEST_DIR/qsd.pid" \ + --daemonize +} + +set_up_throttled_qsd + +# Need a virtio-blk device so that qemu-io writes will not block the monitor +_launch_qemu \ + --blockdev file,node-name=source-proto,filename="$TEST_IMG" \ + --blockdev qcow2,node-name=source-fmt,file=source-proto \ + --device virtio-blk,id=vblk,drive=source-fmt \ + --blockdev "{\"driver\": \"nbd\", + \"node-name\": \"target\", + \"server\": { + \"type\": \"unix\", + \"path\": \"$SOCK_DIR/qsd.sock\" + }, + \"export\": \"target\"}" + +h=$QEMU_HANDLE +_send_qemu_cmd $h '{"execute": "qmp_capabilities"}' 'return' + +# Use sync=top, so the first pass will not copy the whole image +_send_qemu_cmd $h \ + '{"execute": "blockdev-mirror", + "arguments": { + "job-id": "mirror", + "device": "source-fmt", + "target": "target", + "sync": "top" + }}' \ + 'return' \ + | grep -v JOB_STATUS_CHANGE # Ignore these events during creation + +# This too will be used by this and the next case +# $1: QEMU handle +# $2: Image size +wait_for_job_and_quit() { + h=$1 + size=$2 + + # List of expected events + capture_events='BLOCK_JOB_READY JOB_STATUS_CHANGE' + _wait_event $h 'BLOCK_JOB_READY' + QEMU_EVENTS= # Ignore all JOB_STATUS_CHANGE events that came before READY + + # Write something to the device for post-READY mirroring. Write it in + # blocks matching the cluster size, each spaced one block apart, so + # that the mirror job will have to spawn one request per cluster. + # Because the number of concurrent requests is limited (to 16), this + # limits the number of bytes concurrently in flight, which speeds up + # cancelling the job (in-flight requests still are waited for). + # To limit the number of bytes in flight, we could alternatively pass + # something for blockdev-mirror's @buf-size parameter, but + # block-commit does not have such a parameter, so we need to figure + # something out that works for both. + + cluster_size=65536 + step=$((cluster_size * 2)) + + echo '--- Writing data to the virtio-blk device ---' + + for ofs in $(seq 0 $step $((size - step))); do + qemu_io_cmd="qemu-io -d vblk/virtio-backend " + qemu_io_cmd+="\\\"aio_write $ofs $cluster_size\\\"" + + # Do not include these requests in the reference output + # (it's just too much) + silent=yes _send_qemu_cmd $h \ + "{\"execute\": \"human-monitor-command\", + \"arguments\": { + \"command-line\": \"$qemu_io_cmd\" + }}" \ + 'return' + done + + # Wait until the job's length is updated to reflect the write requests + + # We have written to half of the device, so this is the expected job length + final_len=$((size / 2)) + timeout=100 # unit: 0.1 seconds + while true; do + len=$( + _send_qemu_cmd $h \ + '{"execute": "query-block-jobs"}' \ + 'return.*"len": [0-9]\+' \ + | grep 'return.*"len": [0-9]\+' \ + | sed -e 's/.*"len": \([0-9]\+\).*/\1/' + ) + if [ "$len" -eq "$final_len" ]; then + break + fi + timeout=$((timeout - 1)) + if [ "$timeout" -eq 0 ]; then + echo "ERROR: Timeout waiting for job to reach len=$final_len" + break + fi + sleep 0.1 + done + + sleep 1 + + _send_qemu_cmd $h \ + '{"execute": "quit"}' \ + 'return' + + # List of expected events + capture_events='BLOCK_JOB_CANCELLED JOB_STATUS_CHANGE SHUTDOWN' + _wait_event $h 'SHUTDOWN' + QEMU_EVENTS= # Ignore all JOB_STATUS_CHANGE events that came before SHUTDOWN + _wait_event $h 'JOB_STATUS_CHANGE' # standby + _wait_event $h 'JOB_STATUS_CHANGE' # ready + _wait_event $h 'JOB_STATUS_CHANGE' # aborting + # Filter the offset (depends on when exactly `quit` was issued) + _wait_event $h 'BLOCK_JOB_CANCELLED' \ + | sed -e 's/"offset": [0-9]\+/"offset": (filtered)/' + _wait_event $h 'JOB_STATUS_CHANGE' # concluded + _wait_event $h 'JOB_STATUS_CHANGE' # null + + wait=yes _cleanup_qemu + + kill -SIGTERM "$(cat "$TEST_DIR/qsd.pid")" +} + +wait_for_job_and_quit $h $size + +echo +echo === Start active commit to throttled QSD and exit qemu === +echo + +# Same as the above, but instead of mirroring, do an active commit + +_make_test_img $size + +set_up_throttled_qsd + +_launch_qemu \ + --blockdev "{\"driver\": \"nbd\", + \"node-name\": \"target\", + \"server\": { + \"type\": \"unix\", + \"path\": \"$SOCK_DIR/qsd.sock\" + }, + \"export\": \"target\"}" \ + --blockdev file,node-name=source-proto,filename="$TEST_IMG" \ + --blockdev qcow2,node-name=source-fmt,file=source-proto,backing=target \ + --device virtio-blk,id=vblk,drive=source-fmt + +h=$QEMU_HANDLE +_send_qemu_cmd $h '{"execute": "qmp_capabilities"}' 'return' + +_send_qemu_cmd $h \ + '{"execute": "block-commit", + "arguments": { + "job-id": "commit", + "device": "source-fmt" + }}' \ + 'return' \ + | grep -v JOB_STATUS_CHANGE # Ignore these events during creation + +wait_for_job_and_quit $h $size + # success, all done echo "*** done" rm -f $seq.full diff --git a/tests/qemu-iotests/185.out b/tests/qemu-iotests/185.out index XXXXXXX..XXXXXXX 100644 --- a/tests/qemu-iotests/185.out +++ b/tests/qemu-iotests/185.out @@ -XXX,XX +XXX,XX @@ Formatting 'TEST_DIR/t.qcow2.copy', fmt=qcow2 cluster_size=65536 extended_l2=off {"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false, "reason": "host-qmp-quit"}} {"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_CANCELLED", "data": {"device": "disk", "len": 67108864, "offset": 524288, "speed": 65536, "type": "stream"}} No errors were found on the image. + +=== Start mirror to throttled QSD and exit qemu === + +Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864 +{"execute": "qmp_capabilities"} +{"return": {}} +{"execute": "blockdev-mirror", + "arguments": { + "job-id": "mirror", + "device": "source-fmt", + "target": "target", + "sync": "top" + }} +{"return": {}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_READY", "data": {"device": "mirror", "len": 0, "offset": 0, "speed": 0, "type": "mirror"}} +--- Writing data to the virtio-blk device --- +{"execute": "quit"} +{"return": {}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false, "reason": "host-qmp-quit"}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "standby", "id": "mirror"}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "ready", "id": "mirror"}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "mirror"}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_CANCELLED", "data": {"device": "mirror", "len": 33554432, "offset": (filtered), "speed": 0, "type": "mirror"}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "mirror"}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "mirror"}} + +=== Start active commit to throttled QSD and exit qemu === + +Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864 +{"execute": "qmp_capabilities"} +{"return": {}} +{"execute": "block-commit", + "arguments": { + "job-id": "commit", + "device": "source-fmt" + }} +{"return": {}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_READY", "data": {"device": "commit", "len": 0, "offset": 0, "speed": 0, "type": "commit"}} +--- Writing data to the virtio-blk device --- +{"execute": "quit"} +{"return": {}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "SHUTDOWN", "data": {"guest": false, "reason": "host-qmp-quit"}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "standby", "id": "commit"}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "ready", "id": "commit"}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "aborting", "id": "commit"}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "BLOCK_JOB_CANCELLED", "data": {"device": "commit", "len": 33554432, "offset": (filtered), "speed": 0, "type": "commit"}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "concluded", "id": "commit"}} +{"timestamp": {"seconds": TIMESTAMP, "microseconds": TIMESTAMP}, "event": "JOB_STATUS_CHANGE", "data": {"status": "null", "id": "commit"}} *** done -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> When invoked from the main loop, this function is the same as qemu_mutex_iothread_locked, and returns true if the BQL is held. When invoked from iothreads or tests, it returns true only if the current AioContext is the Main Loop. This essentially just extends qemu_mutex_iothread_locked to work also in unit tests or other users like storage-daemon, that run in the Main Loop but end up using the implementation in stubs/iothread-lock.c. Using qemu_mutex_iothread_locked in unit tests defaults to false because they use the implementation in stubs/iothread-lock, making all assertions added in next patches fail despite the AioContext is still the main loop. See the comment in the function header for more information. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Message-Id: <20220303151616.325444-2-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/qemu/main-loop.h | 24 ++++++++++++++++++++++++ softmmu/cpus.c | 5 +++++ stubs/iothread-lock.c | 5 +++++ 3 files changed, 34 insertions(+) diff --git a/include/qemu/main-loop.h b/include/qemu/main-loop.h index XXXXXXX..XXXXXXX 100644 --- a/include/qemu/main-loop.h +++ b/include/qemu/main-loop.h @@ -XXX,XX +XXX,XX @@ AioContext *iohandler_get_aio_context(void); * must always be taken outside other locks. This function helps * functions take different paths depending on whether the current * thread is running within the main loop mutex. + * + * This function should never be used in the block layer, because + * unit tests, block layer tools and qemu-storage-daemon do not + * have a BQL. + * Please instead refer to qemu_in_main_thread(). */ bool qemu_mutex_iothread_locked(void); +/** + * qemu_in_main_thread: return whether it's possible to safely access + * the global state of the block layer. + * + * Global state of the block layer is not accessible from I/O threads + * or worker threads; only from threads that "own" the default + * AioContext that qemu_get_aio_context() returns. For tests, block + * layer tools and qemu-storage-daemon there is a designated thread that + * runs the event loop for qemu_get_aio_context(), and that is the + * main thread. + * + * For emulators, however, any thread that holds the BQL can act + * as the block layer main thread; this will be any of the actual + * main thread, the vCPU threads or the RCU thread. + * + * For clarity, do not use this function outside the block layer. + */ +bool qemu_in_main_thread(void); + /** * qemu_mutex_lock_iothread: Lock the main loop mutex. * diff --git a/softmmu/cpus.c b/softmmu/cpus.c index XXXXXXX..XXXXXXX 100644 --- a/softmmu/cpus.c +++ b/softmmu/cpus.c @@ -XXX,XX +XXX,XX @@ bool qemu_mutex_iothread_locked(void) return get_iothread_locked(); } +bool qemu_in_main_thread(void) +{ + return qemu_mutex_iothread_locked(); +} + /* * The BQL is taken from so many places that it is worth profiling the * callers directly, instead of funneling them all through a single function. diff --git a/stubs/iothread-lock.c b/stubs/iothread-lock.c index XXXXXXX..XXXXXXX 100644 --- a/stubs/iothread-lock.c +++ b/stubs/iothread-lock.c @@ -XXX,XX +XXX,XX @@ bool qemu_mutex_iothread_locked(void) return false; } +bool qemu_in_main_thread(void) +{ + return qemu_get_current_aio_context() == qemu_get_aio_context(); +} + void qemu_mutex_lock_iothread_impl(const char *file, int line) { } -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> Righ now, IO_CODE and IO_OR_GS_CODE are nop, as there isn't really a way to check that a function is only called in I/O. On the other side, we can use qemu_in_main_thread() to check if we are in the main loop. The usage of macros makes easy to extend them in the future without making changes in all callers. They will also visually help understanding in which category each function is, without looking at the header. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Message-Id: <20220303151616.325444-3-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/qemu/main-loop.h | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+) diff --git a/include/qemu/main-loop.h b/include/qemu/main-loop.h index XXXXXXX..XXXXXXX 100644 --- a/include/qemu/main-loop.h +++ b/include/qemu/main-loop.h @@ -XXX,XX +XXX,XX @@ bool qemu_mutex_iothread_locked(void); */ bool qemu_in_main_thread(void); +/* Mark and check that the function is part of the global state API. */ +#define GLOBAL_STATE_CODE() \ + do { \ + assert(qemu_in_main_thread()); \ + } while (0) + +/* Mark and check that the function is part of the I/O API. */ +#define IO_CODE() \ + do { \ + /* nop */ \ + } while (0) + +/* Mark and check that the function is part of the "I/O OR GS" API. */ +#define IO_OR_GS_CODE() \ + do { \ + /* nop */ \ + } while (0) + /** * qemu_mutex_lock_iothread: Lock the main loop mutex. * -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> block.h currently contains a mix of functions: some of them run under the BQL and modify the block layer graph, others are instead thread-safe and perform I/O in iothreads. Some others can only be called by either the main loop or the iothread running the AioContext (and not other iothreads), and using them in another thread would cause deadlocks, and therefore it is not ideal to define them as I/O. It is not easy to understand which function is part of which group (I/O vs GS vs "I/O or GS"), and this patch aims to clarify it. The "GS" functions need the BQL, and often use aio_context_acquire/release and/or drain to be sure they can modify the graph safely. The I/O function are instead thread safe, and can run in any AioContext. "I/O or GS" functions run instead in the main loop or in a single iothread, and use BDRV_POLL_WHILE(). By splitting the header in two files, block-io.h and block-global-state.h we have a clearer view on what needs what kind of protection. block-common.h contains common structures shared by both headers. block.h is left there for legacy and to avoid changing all includes in all c files that use the block APIs. Assertions are added in the next patch. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Message-Id: <20220303151616.325444-4-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-common.h | 418 ++++++++++++++ include/block/block-global-state.h | 253 +++++++++ include/block/block-io.h | 367 ++++++++++++ include/block/block.h | 879 +---------------------------- block.c | 3 + block/meson.build | 7 +- 6 files changed, 1069 insertions(+), 858 deletions(-) create mode 100644 include/block/block-common.h create mode 100644 include/block/block-global-state.h create mode 100644 include/block/block-io.h diff --git a/include/block/block-common.h b/include/block/block-common.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/block/block-common.h @@ -XXX,XX +XXX,XX @@ +/* + * QEMU System Emulator block driver + * + * Copyright (c) 2003 Fabrice Bellard + * + * 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. + */ +#ifndef BLOCK_COMMON_H +#define BLOCK_COMMON_H + +#include "block/aio.h" +#include "block/aio-wait.h" +#include "qemu/iov.h" +#include "qemu/coroutine.h" +#include "block/accounting.h" +#include "block/dirty-bitmap.h" +#include "block/blockjob.h" +#include "qemu/hbitmap.h" +#include "qemu/transactions.h" + +/* + * generated_co_wrapper + * + * Function specifier, which does nothing but mark functions to be + * generated by scripts/block-coroutine-wrapper.py + * + * Read more in docs/devel/block-coroutine-wrapper.rst + */ +#define generated_co_wrapper + +/* block.c */ +typedef struct BlockDriver BlockDriver; +typedef struct BdrvChild BdrvChild; +typedef struct BdrvChildClass BdrvChildClass; + +typedef struct BlockDriverInfo { + /* in bytes, 0 if irrelevant */ + int cluster_size; + /* offset at which the VM state can be saved (0 if not possible) */ + int64_t vm_state_offset; + bool is_dirty; + /* + * True if this block driver only supports compressed writes + */ + bool needs_compressed_writes; +} BlockDriverInfo; + +typedef struct BlockFragInfo { + uint64_t allocated_clusters; + uint64_t total_clusters; + uint64_t fragmented_clusters; + uint64_t compressed_clusters; +} BlockFragInfo; + +typedef enum { + BDRV_REQ_COPY_ON_READ = 0x1, + BDRV_REQ_ZERO_WRITE = 0x2, + + /* + * The BDRV_REQ_MAY_UNMAP flag is used in write_zeroes requests to indicate + * that the block driver should unmap (discard) blocks if it is guaranteed + * that the result will read back as zeroes. The flag is only passed to the + * driver if the block device is opened with BDRV_O_UNMAP. + */ + BDRV_REQ_MAY_UNMAP = 0x4, + + BDRV_REQ_FUA = 0x10, + BDRV_REQ_WRITE_COMPRESSED = 0x20, + + /* + * Signifies that this write request will not change the visible disk + * content. + */ + BDRV_REQ_WRITE_UNCHANGED = 0x40, + + /* + * Forces request serialisation. Use only with write requests. + */ + BDRV_REQ_SERIALISING = 0x80, + + /* + * Execute the request only if the operation can be offloaded or otherwise + * be executed efficiently, but return an error instead of using a slow + * fallback. + */ + BDRV_REQ_NO_FALLBACK = 0x100, + + /* + * BDRV_REQ_PREFETCH makes sense only in the context of copy-on-read + * (i.e., together with the BDRV_REQ_COPY_ON_READ flag or when a COR + * filter is involved), in which case it signals that the COR operation + * need not read the data into memory (qiov) but only ensure they are + * copied to the top layer (i.e., that COR operation is done). + */ + BDRV_REQ_PREFETCH = 0x200, + + /* + * If we need to wait for other requests, just fail immediately. Used + * only together with BDRV_REQ_SERIALISING. + */ + BDRV_REQ_NO_WAIT = 0x400, + + /* Mask of valid flags */ + BDRV_REQ_MASK = 0x7ff, +} BdrvRequestFlags; + +#define BDRV_O_NO_SHARE 0x0001 /* don't share permissions */ +#define BDRV_O_RDWR 0x0002 +#define BDRV_O_RESIZE 0x0004 /* request permission for resizing the node */ +#define BDRV_O_SNAPSHOT 0x0008 /* open the file read only and save + writes in a snapshot */ +#define BDRV_O_TEMPORARY 0x0010 /* delete the file after use */ +#define BDRV_O_NOCACHE 0x0020 /* do not use the host page cache */ +#define BDRV_O_NATIVE_AIO 0x0080 /* use native AIO instead of the + thread pool */ +#define BDRV_O_NO_BACKING 0x0100 /* don't open the backing file */ +#define BDRV_O_NO_FLUSH 0x0200 /* disable flushing on this disk */ +#define BDRV_O_COPY_ON_READ 0x0400 /* copy read backing sectors into image */ +#define BDRV_O_INACTIVE 0x0800 /* consistency hint for migration handoff */ +#define BDRV_O_CHECK 0x1000 /* open solely for consistency check */ +#define BDRV_O_ALLOW_RDWR 0x2000 /* allow reopen to change from r/o to r/w */ +#define BDRV_O_UNMAP 0x4000 /* execute guest UNMAP/TRIM operations */ +#define BDRV_O_PROTOCOL 0x8000 /* if no block driver is explicitly given: + select an appropriate protocol driver, + ignoring the format layer */ +#define BDRV_O_NO_IO 0x10000 /* don't initialize for I/O */ +#define BDRV_O_AUTO_RDONLY 0x20000 /* degrade to read-only if opening + read-write fails */ +#define BDRV_O_IO_URING 0x40000 /* use io_uring instead of the thread pool */ + +#define BDRV_O_CACHE_MASK (BDRV_O_NOCACHE | BDRV_O_NO_FLUSH) + + +/* Option names of options parsed by the block layer */ + +#define BDRV_OPT_CACHE_WB "cache.writeback" +#define BDRV_OPT_CACHE_DIRECT "cache.direct" +#define BDRV_OPT_CACHE_NO_FLUSH "cache.no-flush" +#define BDRV_OPT_READ_ONLY "read-only" +#define BDRV_OPT_AUTO_READ_ONLY "auto-read-only" +#define BDRV_OPT_DISCARD "discard" +#define BDRV_OPT_FORCE_SHARE "force-share" + + +#define BDRV_SECTOR_BITS 9 +#define BDRV_SECTOR_SIZE (1ULL << BDRV_SECTOR_BITS) + +#define BDRV_REQUEST_MAX_SECTORS MIN_CONST(SIZE_MAX >> BDRV_SECTOR_BITS, \ + INT_MAX >> BDRV_SECTOR_BITS) +#define BDRV_REQUEST_MAX_BYTES (BDRV_REQUEST_MAX_SECTORS << BDRV_SECTOR_BITS) + +/* + * We want allow aligning requests and disk length up to any 32bit alignment + * and don't afraid of overflow. + * To achieve it, and in the same time use some pretty number as maximum disk + * size, let's define maximum "length" (a limit for any offset/bytes request and + * for disk size) to be the greatest power of 2 less than INT64_MAX. + */ +#define BDRV_MAX_ALIGNMENT (1L << 30) +#define BDRV_MAX_LENGTH (QEMU_ALIGN_DOWN(INT64_MAX, BDRV_MAX_ALIGNMENT)) + +/* + * Allocation status flags for bdrv_block_status() and friends. + * + * Public flags: + * BDRV_BLOCK_DATA: allocation for data at offset is tied to this layer + * BDRV_BLOCK_ZERO: offset reads as zero + * BDRV_BLOCK_OFFSET_VALID: an associated offset exists for accessing raw data + * BDRV_BLOCK_ALLOCATED: the content of the block is determined by this + * layer rather than any backing, set by block layer + * BDRV_BLOCK_EOF: the returned pnum covers through end of file for this + * layer, set by block layer + * + * Internal flags: + * BDRV_BLOCK_RAW: for use by passthrough drivers, such as raw, to request + * that the block layer recompute the answer from the returned + * BDS; must be accompanied by just BDRV_BLOCK_OFFSET_VALID. + * BDRV_BLOCK_RECURSE: request that the block layer will recursively search for + * zeroes in file child of current block node inside + * returned region. Only valid together with both + * BDRV_BLOCK_DATA and BDRV_BLOCK_OFFSET_VALID. Should not + * appear with BDRV_BLOCK_ZERO. + * + * If BDRV_BLOCK_OFFSET_VALID is set, the map parameter represents the + * host offset within the returned BDS that is allocated for the + * corresponding raw guest data. However, whether that offset + * actually contains data also depends on BDRV_BLOCK_DATA, as follows: + * + * DATA ZERO OFFSET_VALID + * t t t sectors read as zero, returned file is zero at offset + * t f t sectors read as valid from file at offset + * f t t sectors preallocated, read as zero, returned file not + * necessarily zero at offset + * f f t sectors preallocated but read from backing_hd, + * returned file contains garbage at offset + * t t f sectors preallocated, read as zero, unknown offset + * t f f sectors read from unknown file or offset + * f t f not allocated or unknown offset, read as zero + * f f f not allocated or unknown offset, read from backing_hd + */ +#define BDRV_BLOCK_DATA 0x01 +#define BDRV_BLOCK_ZERO 0x02 +#define BDRV_BLOCK_OFFSET_VALID 0x04 +#define BDRV_BLOCK_RAW 0x08 +#define BDRV_BLOCK_ALLOCATED 0x10 +#define BDRV_BLOCK_EOF 0x20 +#define BDRV_BLOCK_RECURSE 0x40 + +typedef QTAILQ_HEAD(BlockReopenQueue, BlockReopenQueueEntry) BlockReopenQueue; + +typedef struct BDRVReopenState { + BlockDriverState *bs; + int flags; + BlockdevDetectZeroesOptions detect_zeroes; + bool backing_missing; + BlockDriverState *old_backing_bs; /* keep pointer for permissions update */ + BlockDriverState *old_file_bs; /* keep pointer for permissions update */ + QDict *options; + QDict *explicit_options; + void *opaque; +} BDRVReopenState; + +/* + * Block operation types + */ +typedef enum BlockOpType { + BLOCK_OP_TYPE_BACKUP_SOURCE, + BLOCK_OP_TYPE_BACKUP_TARGET, + BLOCK_OP_TYPE_CHANGE, + BLOCK_OP_TYPE_COMMIT_SOURCE, + BLOCK_OP_TYPE_COMMIT_TARGET, + BLOCK_OP_TYPE_DATAPLANE, + BLOCK_OP_TYPE_DRIVE_DEL, + BLOCK_OP_TYPE_EJECT, + BLOCK_OP_TYPE_EXTERNAL_SNAPSHOT, + BLOCK_OP_TYPE_INTERNAL_SNAPSHOT, + BLOCK_OP_TYPE_INTERNAL_SNAPSHOT_DELETE, + BLOCK_OP_TYPE_MIRROR_SOURCE, + BLOCK_OP_TYPE_MIRROR_TARGET, + BLOCK_OP_TYPE_RESIZE, + BLOCK_OP_TYPE_STREAM, + BLOCK_OP_TYPE_REPLACE, + BLOCK_OP_TYPE_MAX, +} BlockOpType; + +/* Block node permission constants */ +enum { + /** + * A user that has the "permission" of consistent reads is guaranteed that + * their view of the contents of the block device is complete and + * self-consistent, representing the contents of a disk at a specific + * point. + * + * For most block devices (including their backing files) this is true, but + * the property cannot be maintained in a few situations like for + * intermediate nodes of a commit block job. + */ + BLK_PERM_CONSISTENT_READ = 0x01, + + /** This permission is required to change the visible disk contents. */ + BLK_PERM_WRITE = 0x02, + + /** + * This permission (which is weaker than BLK_PERM_WRITE) is both enough and + * required for writes to the block node when the caller promises that + * the visible disk content doesn't change. + * + * As the BLK_PERM_WRITE permission is strictly stronger, either is + * sufficient to perform an unchanging write. + */ + BLK_PERM_WRITE_UNCHANGED = 0x04, + + /** This permission is required to change the size of a block node. */ + BLK_PERM_RESIZE = 0x08, + + /** + * There was a now-removed bit BLK_PERM_GRAPH_MOD, with value of 0x10. QEMU + * 6.1 and earlier may still lock the corresponding byte in block/file-posix + * locking. So, implementing some new permission should be very careful to + * not interfere with this old unused thing. + */ + + BLK_PERM_ALL = 0x0f, + + DEFAULT_PERM_PASSTHROUGH = BLK_PERM_CONSISTENT_READ + | BLK_PERM_WRITE + | BLK_PERM_WRITE_UNCHANGED + | BLK_PERM_RESIZE, + + DEFAULT_PERM_UNCHANGED = BLK_PERM_ALL & ~DEFAULT_PERM_PASSTHROUGH, +}; + +/* + * Flags that parent nodes assign to child nodes to specify what kind of + * role(s) they take. + * + * At least one of DATA, METADATA, FILTERED, or COW must be set for + * every child. + */ +enum BdrvChildRoleBits { + /* + * This child stores data. + * Any node may have an arbitrary number of such children. + */ + BDRV_CHILD_DATA = (1 << 0), + + /* + * This child stores metadata. + * Any node may have an arbitrary number of metadata-storing + * children. + */ + BDRV_CHILD_METADATA = (1 << 1), + + /* + * A child that always presents exactly the same visible data as + * the parent, e.g. by virtue of the parent forwarding all reads + * and writes. + * This flag is mutually exclusive with DATA, METADATA, and COW. + * Any node may have at most one filtered child at a time. + */ + BDRV_CHILD_FILTERED = (1 << 2), + + /* + * Child from which to read all data that isn't allocated in the + * parent (i.e., the backing child); such data is copied to the + * parent through COW (and optionally COR). + * This field is mutually exclusive with DATA, METADATA, and + * FILTERED. + * Any node may have at most one such backing child at a time. + */ + BDRV_CHILD_COW = (1 << 3), + + /* + * The primary child. For most drivers, this is the child whose + * filename applies best to the parent node. + * Any node may have at most one primary child at a time. + */ + BDRV_CHILD_PRIMARY = (1 << 4), + + /* Useful combination of flags */ + BDRV_CHILD_IMAGE = BDRV_CHILD_DATA + | BDRV_CHILD_METADATA + | BDRV_CHILD_PRIMARY, +}; + +/* Mask of BdrvChildRoleBits values */ +typedef unsigned int BdrvChildRole; + +typedef struct BdrvCheckResult { + int corruptions; + int leaks; + int check_errors; + int corruptions_fixed; + int leaks_fixed; + int64_t image_end_offset; + BlockFragInfo bfi; +} BdrvCheckResult; + +typedef enum { + BDRV_FIX_LEAKS = 1, + BDRV_FIX_ERRORS = 2, +} BdrvCheckMode; + +typedef struct BlockSizes { + uint32_t phys; + uint32_t log; +} BlockSizes; + +typedef struct HDGeometry { + uint32_t heads; + uint32_t sectors; + uint32_t cylinders; +} HDGeometry; + +/* + * Common functions that are neither I/O nor Global State. + * + * These functions must never call any function from other categories + * (I/O, "I/O or GS", Global State) except this one, but can be invoked by + * all of them. + */ + +char *bdrv_perm_names(uint64_t perm); +uint64_t bdrv_qapi_perm_to_blk_perm(BlockPermission qapi_perm); + +void bdrv_init_with_whitelist(void); +bool bdrv_uses_whitelist(void); +int bdrv_is_whitelisted(BlockDriver *drv, bool read_only); + +int bdrv_parse_aio(const char *mode, int *flags); +int bdrv_parse_cache_mode(const char *mode, int *flags, bool *writethrough); +int bdrv_parse_discard_flags(const char *mode, int *flags); + +int path_has_protocol(const char *path); +int path_is_absolute(const char *path); +char *path_combine(const char *base_path, const char *filename); + +char *bdrv_get_full_backing_filename_from_filename(const char *backed, + const char *backing, + Error **errp); + +#endif /* BLOCK_COMMON_H */ diff --git a/include/block/block-global-state.h b/include/block/block-global-state.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/block/block-global-state.h @@ -XXX,XX +XXX,XX @@ +/* + * QEMU System Emulator block driver + * + * Copyright (c) 2003 Fabrice Bellard + * + * 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. + */ +#ifndef BLOCK_GLOBAL_STATE_H +#define BLOCK_GLOBAL_STATE_H + +#include "block-common.h" + +/* + * Global state (GS) API. These functions run under the BQL. + * + * If a function modifies the graph, it also uses drain and/or + * aio_context_acquire/release to be sure it has unique access. + * aio_context locking is needed together with BQL because of + * the thread-safe I/O API that concurrently runs and accesses + * the graph without the BQL. + * + * It is important to note that not all of these functions are + * necessarily limited to running under the BQL, but they would + * require additional auditing and many small thread-safety changes + * to move them into the I/O API. Often it's not worth doing that + * work since the APIs are only used with the BQL held at the + * moment, so they have been placed in the GS API (for now). + * + * These functions can call any function from this and other categories + * (I/O, "I/O or GS", Common), but must be invoked only by other GS APIs. + * + * All functions in this header must use the macro + * GLOBAL_STATE_CODE(); + * to catch when they are accidentally called without the BQL. + */ + +void bdrv_init(void); +BlockDriver *bdrv_find_protocol(const char *filename, + bool allow_protocol_prefix, + Error **errp); +BlockDriver *bdrv_find_format(const char *format_name); +int bdrv_create(BlockDriver *drv, const char* filename, + QemuOpts *opts, Error **errp); +int bdrv_create_file(const char *filename, QemuOpts *opts, Error **errp); + +BlockDriverState *bdrv_new(void); +int bdrv_append(BlockDriverState *bs_new, BlockDriverState *bs_top, + Error **errp); +int bdrv_replace_node(BlockDriverState *from, BlockDriverState *to, + Error **errp); +int bdrv_replace_child_bs(BdrvChild *child, BlockDriverState *new_bs, + Error **errp); +BlockDriverState *bdrv_insert_node(BlockDriverState *bs, QDict *node_options, + int flags, Error **errp); +int bdrv_drop_filter(BlockDriverState *bs, Error **errp); + +BdrvChild *bdrv_open_child(const char *filename, + QDict *options, const char *bdref_key, + BlockDriverState *parent, + const BdrvChildClass *child_class, + BdrvChildRole child_role, + bool allow_none, Error **errp); +BlockDriverState *bdrv_open_blockdev_ref(BlockdevRef *ref, Error **errp); +int bdrv_set_backing_hd(BlockDriverState *bs, BlockDriverState *backing_hd, + Error **errp); +int bdrv_open_backing_file(BlockDriverState *bs, QDict *parent_options, + const char *bdref_key, Error **errp); +BlockDriverState *bdrv_open(const char *filename, const char *reference, + QDict *options, int flags, Error **errp); +BlockDriverState *bdrv_new_open_driver_opts(BlockDriver *drv, + const char *node_name, + QDict *options, int flags, + Error **errp); +BlockDriverState *bdrv_new_open_driver(BlockDriver *drv, const char *node_name, + int flags, Error **errp); +BlockReopenQueue *bdrv_reopen_queue(BlockReopenQueue *bs_queue, + BlockDriverState *bs, QDict *options, + bool keep_old_opts); +void bdrv_reopen_queue_free(BlockReopenQueue *bs_queue); +int bdrv_reopen_multiple(BlockReopenQueue *bs_queue, Error **errp); +int bdrv_reopen(BlockDriverState *bs, QDict *opts, bool keep_old_opts, + Error **errp); +int bdrv_reopen_set_read_only(BlockDriverState *bs, bool read_only, + Error **errp); +BlockDriverState *bdrv_find_backing_image(BlockDriverState *bs, + const char *backing_file); +void bdrv_refresh_filename(BlockDriverState *bs); +void bdrv_refresh_limits(BlockDriverState *bs, Transaction *tran, Error **errp); +int bdrv_commit(BlockDriverState *bs); +int bdrv_make_empty(BdrvChild *c, Error **errp); +int bdrv_change_backing_file(BlockDriverState *bs, const char *backing_file, + const char *backing_fmt, bool warn); +void bdrv_register(BlockDriver *bdrv); +int bdrv_drop_intermediate(BlockDriverState *top, BlockDriverState *base, + const char *backing_file_str); +BlockDriverState *bdrv_find_overlay(BlockDriverState *active, + BlockDriverState *bs); +BlockDriverState *bdrv_find_base(BlockDriverState *bs); +bool bdrv_is_backing_chain_frozen(BlockDriverState *bs, BlockDriverState *base, + Error **errp); +int bdrv_freeze_backing_chain(BlockDriverState *bs, BlockDriverState *base, + Error **errp); +void bdrv_unfreeze_backing_chain(BlockDriverState *bs, BlockDriverState *base); + +/* + * The units of offset and total_work_size may be chosen arbitrarily by the + * block driver; total_work_size may change during the course of the amendment + * operation + */ +typedef void BlockDriverAmendStatusCB(BlockDriverState *bs, int64_t offset, + int64_t total_work_size, void *opaque); +int bdrv_amend_options(BlockDriverState *bs_new, QemuOpts *opts, + BlockDriverAmendStatusCB *status_cb, void *cb_opaque, + bool force, + Error **errp); + +/* check if a named node can be replaced when doing drive-mirror */ +BlockDriverState *check_to_replace_node(BlockDriverState *parent_bs, + const char *node_name, Error **errp); + +int bdrv_activate(BlockDriverState *bs, Error **errp); +void bdrv_activate_all(Error **errp); +int bdrv_inactivate_all(void); + +int bdrv_flush_all(void); +void bdrv_close_all(void); +void bdrv_drain_all_begin(void); +void bdrv_drain_all_end(void); +void bdrv_drain_all(void); + +int bdrv_has_zero_init_1(BlockDriverState *bs); +int bdrv_has_zero_init(BlockDriverState *bs); +BlockDriverState *bdrv_find_node(const char *node_name); +BlockDeviceInfoList *bdrv_named_nodes_list(bool flat, Error **errp); +XDbgBlockGraph *bdrv_get_xdbg_block_graph(Error **errp); +BlockDriverState *bdrv_lookup_bs(const char *device, + const char *node_name, + Error **errp); +bool bdrv_chain_contains(BlockDriverState *top, BlockDriverState *base); +BlockDriverState *bdrv_next_node(BlockDriverState *bs); +BlockDriverState *bdrv_next_all_states(BlockDriverState *bs); + +typedef struct BdrvNextIterator { + enum { + BDRV_NEXT_BACKEND_ROOTS, + BDRV_NEXT_MONITOR_OWNED, + } phase; + BlockBackend *blk; + BlockDriverState *bs; +} BdrvNextIterator; + +BlockDriverState *bdrv_first(BdrvNextIterator *it); +BlockDriverState *bdrv_next(BdrvNextIterator *it); +void bdrv_next_cleanup(BdrvNextIterator *it); + +BlockDriverState *bdrv_next_monitor_owned(BlockDriverState *bs); +void bdrv_iterate_format(void (*it)(void *opaque, const char *name), + void *opaque, bool read_only); +int bdrv_get_flags(BlockDriverState *bs); +char *bdrv_get_full_backing_filename(BlockDriverState *bs, Error **errp); +char *bdrv_dirname(BlockDriverState *bs, Error **errp); + +void bdrv_img_create(const char *filename, const char *fmt, + const char *base_filename, const char *base_fmt, + char *options, uint64_t img_size, int flags, + bool quiet, Error **errp); + +void bdrv_ref(BlockDriverState *bs); +void bdrv_unref(BlockDriverState *bs); +void bdrv_unref_child(BlockDriverState *parent, BdrvChild *child); +BdrvChild *bdrv_attach_child(BlockDriverState *parent_bs, + BlockDriverState *child_bs, + const char *child_name, + const BdrvChildClass *child_class, + BdrvChildRole child_role, + Error **errp); + +bool bdrv_op_is_blocked(BlockDriverState *bs, BlockOpType op, Error **errp); +void bdrv_op_block(BlockDriverState *bs, BlockOpType op, Error *reason); +void bdrv_op_unblock(BlockDriverState *bs, BlockOpType op, Error *reason); +void bdrv_op_block_all(BlockDriverState *bs, Error *reason); +void bdrv_op_unblock_all(BlockDriverState *bs, Error *reason); +bool bdrv_op_blocker_is_empty(BlockDriverState *bs); + +int bdrv_debug_breakpoint(BlockDriverState *bs, const char *event, + const char *tag); +int bdrv_debug_remove_breakpoint(BlockDriverState *bs, const char *tag); +int bdrv_debug_resume(BlockDriverState *bs, const char *tag); +bool bdrv_debug_is_suspended(BlockDriverState *bs, const char *tag); + +/** + * Locks the AioContext of @bs if it's not the current AioContext. This avoids + * double locking which could lead to deadlocks: This is a coroutine_fn, so we + * know we already own the lock of the current AioContext. + * + * May only be called in the main thread. + */ +void coroutine_fn bdrv_co_lock(BlockDriverState *bs); + +/** + * Unlocks the AioContext of @bs if it's not the current AioContext. + */ +void coroutine_fn bdrv_co_unlock(BlockDriverState *bs); + +void bdrv_set_aio_context_ignore(BlockDriverState *bs, + AioContext *new_context, GSList **ignore); +int bdrv_try_set_aio_context(BlockDriverState *bs, AioContext *ctx, + Error **errp); +int bdrv_child_try_set_aio_context(BlockDriverState *bs, AioContext *ctx, + BdrvChild *ignore_child, Error **errp); +bool bdrv_child_can_set_aio_context(BdrvChild *c, AioContext *ctx, + GSList **ignore, Error **errp); +bool bdrv_can_set_aio_context(BlockDriverState *bs, AioContext *ctx, + GSList **ignore, Error **errp); +AioContext *bdrv_child_get_parent_aio_context(BdrvChild *c); + +int bdrv_probe_blocksizes(BlockDriverState *bs, BlockSizes *bsz); +int bdrv_probe_geometry(BlockDriverState *bs, HDGeometry *geo); + +void bdrv_add_child(BlockDriverState *parent, BlockDriverState *child, + Error **errp); +void bdrv_del_child(BlockDriverState *parent, BdrvChild *child, Error **errp); + +/** + * + * bdrv_register_buf/bdrv_unregister_buf: + * + * Register/unregister a buffer for I/O. For example, VFIO drivers are + * interested to know the memory areas that would later be used for I/O, so + * that they can prepare IOMMU mapping etc., to get better performance. + */ +void bdrv_register_buf(BlockDriverState *bs, void *host, size_t size); +void bdrv_unregister_buf(BlockDriverState *bs, void *host); + +void bdrv_cancel_in_flight(BlockDriverState *bs); + +#endif /* BLOCK_GLOBAL_STATE_H */ diff --git a/include/block/block-io.h b/include/block/block-io.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/block/block-io.h @@ -XXX,XX +XXX,XX @@ +/* + * QEMU System Emulator block driver + * + * Copyright (c) 2003 Fabrice Bellard + * + * 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. + */ +#ifndef BLOCK_IO_H +#define BLOCK_IO_H + +#include "block-common.h" + +/* + * I/O API functions. These functions are thread-safe, and therefore + * can run in any thread as long as the thread has called + * aio_context_acquire/release(). + * + * These functions can only call functions from I/O and Common categories, + * but can be invoked by GS, "I/O or GS" and I/O APIs. + * + * All functions in this category must use the macro + * IO_CODE(); + * to catch when they are accidentally called by the wrong API. + */ + +int bdrv_pwrite_zeroes(BdrvChild *child, int64_t offset, + int64_t bytes, BdrvRequestFlags flags); +int bdrv_make_zero(BdrvChild *child, BdrvRequestFlags flags); +int bdrv_pread(BdrvChild *child, int64_t offset, void *buf, int64_t bytes); +int bdrv_pwrite(BdrvChild *child, int64_t offset, const void *buf, + int64_t bytes); +int bdrv_pwrite_sync(BdrvChild *child, int64_t offset, + const void *buf, int64_t bytes); +/* + * Efficiently zero a region of the disk image. Note that this is a regular + * I/O request like read or write and should have a reasonable size. This + * function is not suitable for zeroing the entire image in a single request + * because it may allocate memory for the entire region. + */ +int coroutine_fn bdrv_co_pwrite_zeroes(BdrvChild *child, int64_t offset, + int64_t bytes, BdrvRequestFlags flags); + +int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact, + PreallocMode prealloc, BdrvRequestFlags flags, + Error **errp); + +int64_t bdrv_nb_sectors(BlockDriverState *bs); +int64_t bdrv_getlength(BlockDriverState *bs); +int64_t bdrv_get_allocated_file_size(BlockDriverState *bs); +BlockMeasureInfo *bdrv_measure(BlockDriver *drv, QemuOpts *opts, + BlockDriverState *in_bs, Error **errp); +void bdrv_get_geometry(BlockDriverState *bs, uint64_t *nb_sectors_ptr); +int coroutine_fn bdrv_co_delete_file(BlockDriverState *bs, Error **errp); +void coroutine_fn bdrv_co_delete_file_noerr(BlockDriverState *bs); + + +/* async block I/O */ +void bdrv_aio_cancel(BlockAIOCB *acb); +void bdrv_aio_cancel_async(BlockAIOCB *acb); + +/* sg packet commands */ +int bdrv_co_ioctl(BlockDriverState *bs, int req, void *buf); + +/* Ensure contents are flushed to disk. */ +int coroutine_fn bdrv_co_flush(BlockDriverState *bs); + +int bdrv_co_pdiscard(BdrvChild *child, int64_t offset, int64_t bytes); +bool bdrv_can_write_zeroes_with_unmap(BlockDriverState *bs); +int bdrv_block_status(BlockDriverState *bs, int64_t offset, + int64_t bytes, int64_t *pnum, int64_t *map, + BlockDriverState **file); +int bdrv_block_status_above(BlockDriverState *bs, BlockDriverState *base, + int64_t offset, int64_t bytes, int64_t *pnum, + int64_t *map, BlockDriverState **file); +int bdrv_is_allocated(BlockDriverState *bs, int64_t offset, int64_t bytes, + int64_t *pnum); +int bdrv_is_allocated_above(BlockDriverState *top, BlockDriverState *base, + bool include_base, int64_t offset, int64_t bytes, + int64_t *pnum); +int coroutine_fn bdrv_co_is_zero_fast(BlockDriverState *bs, int64_t offset, + int64_t bytes); + +int bdrv_can_set_read_only(BlockDriverState *bs, bool read_only, + bool ignore_allow_rdw, Error **errp); +int bdrv_apply_auto_read_only(BlockDriverState *bs, const char *errmsg, + Error **errp); +bool bdrv_is_read_only(BlockDriverState *bs); +bool bdrv_is_writable(BlockDriverState *bs); +bool bdrv_is_sg(BlockDriverState *bs); +bool bdrv_is_inserted(BlockDriverState *bs); +void bdrv_lock_medium(BlockDriverState *bs, bool locked); +void bdrv_eject(BlockDriverState *bs, bool eject_flag); +const char *bdrv_get_format_name(BlockDriverState *bs); + +bool bdrv_supports_compressed_writes(BlockDriverState *bs); +const char *bdrv_get_node_name(const BlockDriverState *bs); +const char *bdrv_get_device_name(const BlockDriverState *bs); +const char *bdrv_get_device_or_node_name(const BlockDriverState *bs); +int bdrv_get_info(BlockDriverState *bs, BlockDriverInfo *bdi); +ImageInfoSpecific *bdrv_get_specific_info(BlockDriverState *bs, + Error **errp); +BlockStatsSpecific *bdrv_get_specific_stats(BlockDriverState *bs); +void bdrv_round_to_clusters(BlockDriverState *bs, + int64_t offset, int64_t bytes, + int64_t *cluster_offset, + int64_t *cluster_bytes); + +void bdrv_get_backing_filename(BlockDriverState *bs, + char *filename, int filename_size); + +int bdrv_save_vmstate(BlockDriverState *bs, const uint8_t *buf, + int64_t pos, int size); + +int bdrv_load_vmstate(BlockDriverState *bs, uint8_t *buf, + int64_t pos, int size); + +/* + * Returns the alignment in bytes that is required so that no bounce buffer + * is required throughout the stack + */ +size_t bdrv_min_mem_align(BlockDriverState *bs); +/* Returns optimal alignment in bytes for bounce buffer */ +size_t bdrv_opt_mem_align(BlockDriverState *bs); +void *qemu_blockalign(BlockDriverState *bs, size_t size); +void *qemu_blockalign0(BlockDriverState *bs, size_t size); +void *qemu_try_blockalign(BlockDriverState *bs, size_t size); +void *qemu_try_blockalign0(BlockDriverState *bs, size_t size); +bool bdrv_qiov_is_aligned(BlockDriverState *bs, QEMUIOVector *qiov); + +void bdrv_enable_copy_on_read(BlockDriverState *bs); +void bdrv_disable_copy_on_read(BlockDriverState *bs); + +void bdrv_debug_event(BlockDriverState *bs, BlkdebugEvent event); + +#define BLKDBG_EVENT(child, evt) \ + do { \ + if (child) { \ + bdrv_debug_event(child->bs, evt); \ + } \ + } while (0) + +/** + * bdrv_get_aio_context: + * + * Returns: the currently bound #AioContext + */ +AioContext *bdrv_get_aio_context(BlockDriverState *bs); + +/** + * Move the current coroutine to the AioContext of @bs and return the old + * AioContext of the coroutine. Increase bs->in_flight so that draining @bs + * will wait for the operation to proceed until the corresponding + * bdrv_co_leave(). + * + * Consequently, you can't call drain inside a bdrv_co_enter/leave() section as + * this will deadlock. + */ +AioContext *coroutine_fn bdrv_co_enter(BlockDriverState *bs); + +/** + * Ends a section started by bdrv_co_enter(). Move the current coroutine back + * to old_ctx and decrease bs->in_flight again. + */ +void coroutine_fn bdrv_co_leave(BlockDriverState *bs, AioContext *old_ctx); + +/** + * Transfer control to @co in the aio context of @bs + */ +void bdrv_coroutine_enter(BlockDriverState *bs, Coroutine *co); + +AioContext *child_of_bds_get_parent_aio_context(BdrvChild *c); + +void bdrv_io_plug(BlockDriverState *bs); +void bdrv_io_unplug(BlockDriverState *bs); + +bool bdrv_can_store_new_dirty_bitmap(BlockDriverState *bs, const char *name, + uint32_t granularity, Error **errp); + +/** + * + * bdrv_co_copy_range: + * + * Do offloaded copy between two children. If the operation is not implemented + * by the driver, or if the backend storage doesn't support it, a negative + * error code will be returned. + * + * Note: block layer doesn't emulate or fallback to a bounce buffer approach + * because usually the caller shouldn't attempt offloaded copy any more (e.g. + * calling copy_file_range(2)) after the first error, thus it should fall back + * to a read+write path in the caller level. + * + * @src: Source child to copy data from + * @src_offset: offset in @src image to read data + * @dst: Destination child to copy data to + * @dst_offset: offset in @dst image to write data + * @bytes: number of bytes to copy + * @flags: request flags. Supported flags: + * BDRV_REQ_ZERO_WRITE - treat the @src range as zero data and do zero + * write on @dst as if bdrv_co_pwrite_zeroes is + * called. Used to simplify caller code, or + * during BlockDriver.bdrv_co_copy_range_from() + * recursion. + * BDRV_REQ_NO_SERIALISING - do not serialize with other overlapping + * requests currently in flight. + * + * Returns: 0 if succeeded; negative error code if failed. + **/ +int coroutine_fn bdrv_co_copy_range(BdrvChild *src, int64_t src_offset, + BdrvChild *dst, int64_t dst_offset, + int64_t bytes, BdrvRequestFlags read_flags, + BdrvRequestFlags write_flags); + +/** + * bdrv_drained_end_no_poll: + * + * Same as bdrv_drained_end(), but do not poll for the subgraph to + * actually become unquiesced. Therefore, no graph changes will occur + * with this function. + * + * *drained_end_counter is incremented for every background operation + * that is scheduled, and will be decremented for every operation once + * it settles. The caller must poll until it reaches 0. The counter + * should be accessed using atomic operations only. + */ +void bdrv_drained_end_no_poll(BlockDriverState *bs, int *drained_end_counter); + + +/* + * "I/O or GS" API functions. These functions can run without + * the BQL, but only in one specific iothread/main loop. + * + * More specifically, these functions use BDRV_POLL_WHILE(bs), which + * requires the caller to be either in the main thread and hold + * the BlockdriverState (bs) AioContext lock, or directly in the + * home thread that runs the bs AioContext. Calling them from + * another thread in another AioContext would cause deadlocks. + * + * Therefore, these functions are not proper I/O, because they + * can't run in *any* iothreads, but only in a specific one. + * + * These functions can call any function from I/O, Common and this + * categories, but must be invoked only by other "I/O or GS" and GS APIs. + * + * All functions in this category must use the macro + * IO_OR_GS_CODE(); + * to catch when they are accidentally called by the wrong API. + */ + +#define BDRV_POLL_WHILE(bs, cond) ({ \ + BlockDriverState *bs_ = (bs); \ + AIO_WAIT_WHILE(bdrv_get_aio_context(bs_), \ + cond); }) + +void bdrv_drain(BlockDriverState *bs); +void coroutine_fn bdrv_co_drain(BlockDriverState *bs); + +int generated_co_wrapper +bdrv_truncate(BdrvChild *child, int64_t offset, bool exact, + PreallocMode prealloc, BdrvRequestFlags flags, Error **errp); + +int generated_co_wrapper bdrv_check(BlockDriverState *bs, BdrvCheckResult *res, + BdrvCheckMode fix); + +/* Invalidate any cached metadata used by image formats */ +int generated_co_wrapper bdrv_invalidate_cache(BlockDriverState *bs, + Error **errp); +int generated_co_wrapper bdrv_flush(BlockDriverState *bs); +int generated_co_wrapper bdrv_pdiscard(BdrvChild *child, int64_t offset, + int64_t bytes); +int generated_co_wrapper +bdrv_readv_vmstate(BlockDriverState *bs, QEMUIOVector *qiov, int64_t pos); +int generated_co_wrapper +bdrv_writev_vmstate(BlockDriverState *bs, QEMUIOVector *qiov, int64_t pos); + +/** + * bdrv_parent_drained_begin_single: + * + * Begin a quiesced section for the parent of @c. If @poll is true, wait for + * any pending activity to cease. + */ +void bdrv_parent_drained_begin_single(BdrvChild *c, bool poll); + +/** + * bdrv_parent_drained_end_single: + * + * End a quiesced section for the parent of @c. + * + * This polls @bs's AioContext until all scheduled sub-drained_ends + * have settled, which may result in graph changes. + */ +void bdrv_parent_drained_end_single(BdrvChild *c); + +/** + * bdrv_drain_poll: + * + * Poll for pending requests in @bs, its parents (except for @ignore_parent), + * and if @recursive is true its children as well (used for subtree drain). + * + * If @ignore_bds_parents is true, parents that are BlockDriverStates must + * ignore the drain request because they will be drained separately (used for + * drain_all). + * + * This is part of bdrv_drained_begin. + */ +bool bdrv_drain_poll(BlockDriverState *bs, bool recursive, + BdrvChild *ignore_parent, bool ignore_bds_parents); + +/** + * bdrv_drained_begin: + * + * Begin a quiesced section for exclusive access to the BDS, by disabling + * external request sources including NBD server, block jobs, and device model. + * + * This function can be recursive. + */ +void bdrv_drained_begin(BlockDriverState *bs); + +/** + * bdrv_do_drained_begin_quiesce: + * + * Quiesces a BDS like bdrv_drained_begin(), but does not wait for already + * running requests to complete. + */ +void bdrv_do_drained_begin_quiesce(BlockDriverState *bs, + BdrvChild *parent, bool ignore_bds_parents); + +/** + * Like bdrv_drained_begin, but recursively begins a quiesced section for + * exclusive access to all child nodes as well. + */ +void bdrv_subtree_drained_begin(BlockDriverState *bs); + +/** + * bdrv_drained_end: + * + * End a quiescent section started by bdrv_drained_begin(). + * + * This polls @bs's AioContext until all scheduled sub-drained_ends + * have settled. On one hand, that may result in graph changes. On + * the other, this requires that the caller either runs in the main + * loop; or that all involved nodes (@bs and all of its parents) are + * in the caller's AioContext. + */ +void bdrv_drained_end(BlockDriverState *bs); + +/** + * End a quiescent section started by bdrv_subtree_drained_begin(). + */ +void bdrv_subtree_drained_end(BlockDriverState *bs); + +#endif /* BLOCK_IO_H */ 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 @@ -#ifndef BLOCK_H -#define BLOCK_H - -#include "block/aio.h" -#include "block/aio-wait.h" -#include "qemu/iov.h" -#include "qemu/coroutine.h" -#include "block/accounting.h" -#include "block/dirty-bitmap.h" -#include "block/blockjob.h" -#include "qemu/hbitmap.h" -#include "qemu/transactions.h" - /* - * generated_co_wrapper - * - * Function specifier, which does nothing but mark functions to be - * generated by scripts/block-coroutine-wrapper.py - * - * Read more in docs/devel/block-coroutine-wrapper.rst - */ -#define generated_co_wrapper - -/* block.c */ -typedef struct BlockDriver BlockDriver; -typedef struct BdrvChild BdrvChild; -typedef struct BdrvChildClass BdrvChildClass; - -typedef struct BlockDriverInfo { - /* in bytes, 0 if irrelevant */ - int cluster_size; - /* offset at which the VM state can be saved (0 if not possible) */ - int64_t vm_state_offset; - bool is_dirty; - /* - * True if this block driver only supports compressed writes - */ - bool needs_compressed_writes; -} BlockDriverInfo; - -typedef struct BlockFragInfo { - uint64_t allocated_clusters; - uint64_t total_clusters; - uint64_t fragmented_clusters; - uint64_t compressed_clusters; -} BlockFragInfo; - -typedef enum { - BDRV_REQ_COPY_ON_READ = 0x1, - BDRV_REQ_ZERO_WRITE = 0x2, - - /* - * The BDRV_REQ_MAY_UNMAP flag is used in write_zeroes requests to indicate - * that the block driver should unmap (discard) blocks if it is guaranteed - * that the result will read back as zeroes. The flag is only passed to the - * driver if the block device is opened with BDRV_O_UNMAP. - */ - BDRV_REQ_MAY_UNMAP = 0x4, - - BDRV_REQ_FUA = 0x10, - BDRV_REQ_WRITE_COMPRESSED = 0x20, - - /* Signifies that this write request will not change the visible disk - * content. */ - BDRV_REQ_WRITE_UNCHANGED = 0x40, - - /* Forces request serialisation. Use only with write requests. */ - BDRV_REQ_SERIALISING = 0x80, - - /* Execute the request only if the operation can be offloaded or otherwise - * be executed efficiently, but return an error instead of using a slow - * fallback. */ - BDRV_REQ_NO_FALLBACK = 0x100, - - /* - * BDRV_REQ_PREFETCH makes sense only in the context of copy-on-read - * (i.e., together with the BDRV_REQ_COPY_ON_READ flag or when a COR - * filter is involved), in which case it signals that the COR operation - * need not read the data into memory (qiov) but only ensure they are - * copied to the top layer (i.e., that COR operation is done). - */ - BDRV_REQ_PREFETCH = 0x200, - - /* - * If we need to wait for other requests, just fail immediately. Used - * only together with BDRV_REQ_SERIALISING. - */ - BDRV_REQ_NO_WAIT = 0x400, - - /* Mask of valid flags */ - BDRV_REQ_MASK = 0x7ff, -} BdrvRequestFlags; - -typedef struct BlockSizes { - uint32_t phys; - uint32_t log; -} BlockSizes; - -typedef struct HDGeometry { - uint32_t heads; - uint32_t sectors; - uint32_t cylinders; -} HDGeometry; - -#define BDRV_O_NO_SHARE 0x0001 /* don't share permissions */ -#define BDRV_O_RDWR 0x0002 -#define BDRV_O_RESIZE 0x0004 /* request permission for resizing the node */ -#define BDRV_O_SNAPSHOT 0x0008 /* open the file read only and save writes in a snapshot */ -#define BDRV_O_TEMPORARY 0x0010 /* delete the file after use */ -#define BDRV_O_NOCACHE 0x0020 /* do not use the host page cache */ -#define BDRV_O_NATIVE_AIO 0x0080 /* use native AIO instead of the thread pool */ -#define BDRV_O_NO_BACKING 0x0100 /* don't open the backing file */ -#define BDRV_O_NO_FLUSH 0x0200 /* disable flushing on this disk */ -#define BDRV_O_COPY_ON_READ 0x0400 /* copy read backing sectors into image */ -#define BDRV_O_INACTIVE 0x0800 /* consistency hint for migration handoff */ -#define BDRV_O_CHECK 0x1000 /* open solely for consistency check */ -#define BDRV_O_ALLOW_RDWR 0x2000 /* allow reopen to change from r/o to r/w */ -#define BDRV_O_UNMAP 0x4000 /* execute guest UNMAP/TRIM operations */ -#define BDRV_O_PROTOCOL 0x8000 /* if no block driver is explicitly given: - select an appropriate protocol driver, - ignoring the format layer */ -#define BDRV_O_NO_IO 0x10000 /* don't initialize for I/O */ -#define BDRV_O_AUTO_RDONLY 0x20000 /* degrade to read-only if opening read-write fails */ -#define BDRV_O_IO_URING 0x40000 /* use io_uring instead of the thread pool */ - -#define BDRV_O_CACHE_MASK (BDRV_O_NOCACHE | BDRV_O_NO_FLUSH) - - -/* Option names of options parsed by the block layer */ - -#define BDRV_OPT_CACHE_WB "cache.writeback" -#define BDRV_OPT_CACHE_DIRECT "cache.direct" -#define BDRV_OPT_CACHE_NO_FLUSH "cache.no-flush" -#define BDRV_OPT_READ_ONLY "read-only" -#define BDRV_OPT_AUTO_READ_ONLY "auto-read-only" -#define BDRV_OPT_DISCARD "discard" -#define BDRV_OPT_FORCE_SHARE "force-share" - - -#define BDRV_SECTOR_BITS 9 -#define BDRV_SECTOR_SIZE (1ULL << BDRV_SECTOR_BITS) - -#define BDRV_REQUEST_MAX_SECTORS MIN_CONST(SIZE_MAX >> BDRV_SECTOR_BITS, \ - INT_MAX >> BDRV_SECTOR_BITS) -#define BDRV_REQUEST_MAX_BYTES (BDRV_REQUEST_MAX_SECTORS << BDRV_SECTOR_BITS) - -/* - * We want allow aligning requests and disk length up to any 32bit alignment - * and don't afraid of overflow. - * To achieve it, and in the same time use some pretty number as maximum disk - * size, let's define maximum "length" (a limit for any offset/bytes request and - * for disk size) to be the greatest power of 2 less than INT64_MAX. - */ -#define BDRV_MAX_ALIGNMENT (1L << 30) -#define BDRV_MAX_LENGTH (QEMU_ALIGN_DOWN(INT64_MAX, BDRV_MAX_ALIGNMENT)) - -/* - * Allocation status flags for bdrv_block_status() and friends. - * - * Public flags: - * BDRV_BLOCK_DATA: allocation for data at offset is tied to this layer - * BDRV_BLOCK_ZERO: offset reads as zero - * BDRV_BLOCK_OFFSET_VALID: an associated offset exists for accessing raw data - * BDRV_BLOCK_ALLOCATED: the content of the block is determined by this - * layer rather than any backing, set by block layer - * BDRV_BLOCK_EOF: the returned pnum covers through end of file for this - * layer, set by block layer - * - * Internal flags: - * BDRV_BLOCK_RAW: for use by passthrough drivers, such as raw, to request - * that the block layer recompute the answer from the returned - * BDS; must be accompanied by just BDRV_BLOCK_OFFSET_VALID. - * BDRV_BLOCK_RECURSE: request that the block layer will recursively search for - * zeroes in file child of current block node inside - * returned region. Only valid together with both - * BDRV_BLOCK_DATA and BDRV_BLOCK_OFFSET_VALID. Should not - * appear with BDRV_BLOCK_ZERO. - * - * If BDRV_BLOCK_OFFSET_VALID is set, the map parameter represents the - * host offset within the returned BDS that is allocated for the - * corresponding raw guest data. However, whether that offset - * actually contains data also depends on BDRV_BLOCK_DATA, as follows: - * - * DATA ZERO OFFSET_VALID - * t t t sectors read as zero, returned file is zero at offset - * t f t sectors read as valid from file at offset - * f t t sectors preallocated, read as zero, returned file not - * necessarily zero at offset - * f f t sectors preallocated but read from backing_hd, - * returned file contains garbage at offset - * t t f sectors preallocated, read as zero, unknown offset - * t f f sectors read from unknown file or offset - * f t f not allocated or unknown offset, read as zero - * f f f not allocated or unknown offset, read from backing_hd - */ -#define BDRV_BLOCK_DATA 0x01 -#define BDRV_BLOCK_ZERO 0x02 -#define BDRV_BLOCK_OFFSET_VALID 0x04 -#define BDRV_BLOCK_RAW 0x08 -#define BDRV_BLOCK_ALLOCATED 0x10 -#define BDRV_BLOCK_EOF 0x20 -#define BDRV_BLOCK_RECURSE 0x40 - -typedef QTAILQ_HEAD(BlockReopenQueue, BlockReopenQueueEntry) BlockReopenQueue; - -typedef struct BDRVReopenState { - BlockDriverState *bs; - int flags; - BlockdevDetectZeroesOptions detect_zeroes; - bool backing_missing; - BlockDriverState *old_backing_bs; /* keep pointer for permissions update */ - BlockDriverState *old_file_bs; /* keep pointer for permissions update */ - QDict *options; - QDict *explicit_options; - void *opaque; -} BDRVReopenState; - -/* - * Block operation types - */ -typedef enum BlockOpType { - BLOCK_OP_TYPE_BACKUP_SOURCE, - BLOCK_OP_TYPE_BACKUP_TARGET, - BLOCK_OP_TYPE_CHANGE, - BLOCK_OP_TYPE_COMMIT_SOURCE, - BLOCK_OP_TYPE_COMMIT_TARGET, - BLOCK_OP_TYPE_DATAPLANE, - BLOCK_OP_TYPE_DRIVE_DEL, - BLOCK_OP_TYPE_EJECT, - BLOCK_OP_TYPE_EXTERNAL_SNAPSHOT, - BLOCK_OP_TYPE_INTERNAL_SNAPSHOT, - BLOCK_OP_TYPE_INTERNAL_SNAPSHOT_DELETE, - BLOCK_OP_TYPE_MIRROR_SOURCE, - BLOCK_OP_TYPE_MIRROR_TARGET, - BLOCK_OP_TYPE_RESIZE, - BLOCK_OP_TYPE_STREAM, - BLOCK_OP_TYPE_REPLACE, - BLOCK_OP_TYPE_MAX, -} BlockOpType; - -/* Block node permission constants */ -enum { - /** - * A user that has the "permission" of consistent reads is guaranteed that - * their view of the contents of the block device is complete and - * self-consistent, representing the contents of a disk at a specific - * point. - * - * For most block devices (including their backing files) this is true, but - * the property cannot be maintained in a few situations like for - * intermediate nodes of a commit block job. - */ - BLK_PERM_CONSISTENT_READ = 0x01, - - /** This permission is required to change the visible disk contents. */ - BLK_PERM_WRITE = 0x02, - - /** - * This permission (which is weaker than BLK_PERM_WRITE) is both enough and - * required for writes to the block node when the caller promises that - * the visible disk content doesn't change. - * - * As the BLK_PERM_WRITE permission is strictly stronger, either is - * sufficient to perform an unchanging write. - */ - BLK_PERM_WRITE_UNCHANGED = 0x04, - - /** This permission is required to change the size of a block node. */ - BLK_PERM_RESIZE = 0x08, - - /** - * There was a now-removed bit BLK_PERM_GRAPH_MOD, with value of 0x10. QEMU - * 6.1 and earlier may still lock the corresponding byte in block/file-posix - * locking. So, implementing some new permission should be very careful to - * not interfere with this old unused thing. - */ - - BLK_PERM_ALL = 0x0f, - - DEFAULT_PERM_PASSTHROUGH = BLK_PERM_CONSISTENT_READ - | BLK_PERM_WRITE - | BLK_PERM_WRITE_UNCHANGED - | BLK_PERM_RESIZE, - - DEFAULT_PERM_UNCHANGED = BLK_PERM_ALL & ~DEFAULT_PERM_PASSTHROUGH, -}; - -/* - * Flags that parent nodes assign to child nodes to specify what kind of - * role(s) they take. - * - * At least one of DATA, METADATA, FILTERED, or COW must be set for - * every child. - */ -enum BdrvChildRoleBits { - /* - * This child stores data. - * Any node may have an arbitrary number of such children. - */ - BDRV_CHILD_DATA = (1 << 0), - - /* - * This child stores metadata. - * Any node may have an arbitrary number of metadata-storing - * children. - */ - BDRV_CHILD_METADATA = (1 << 1), - - /* - * A child that always presents exactly the same visible data as - * the parent, e.g. by virtue of the parent forwarding all reads - * and writes. - * This flag is mutually exclusive with DATA, METADATA, and COW. - * Any node may have at most one filtered child at a time. - */ - BDRV_CHILD_FILTERED = (1 << 2), - - /* - * Child from which to read all data that isn't allocated in the - * parent (i.e., the backing child); such data is copied to the - * parent through COW (and optionally COR). - * This field is mutually exclusive with DATA, METADATA, and - * FILTERED. - * Any node may have at most one such backing child at a time. - */ - BDRV_CHILD_COW = (1 << 3), - - /* - * The primary child. For most drivers, this is the child whose - * filename applies best to the parent node. - * Any node may have at most one primary child at a time. - */ - BDRV_CHILD_PRIMARY = (1 << 4), - - /* Useful combination of flags */ - BDRV_CHILD_IMAGE = BDRV_CHILD_DATA - | BDRV_CHILD_METADATA - | BDRV_CHILD_PRIMARY, -}; - -/* Mask of BdrvChildRoleBits values */ -typedef unsigned int BdrvChildRole; - -char *bdrv_perm_names(uint64_t perm); -uint64_t bdrv_qapi_perm_to_blk_perm(BlockPermission qapi_perm); - -void bdrv_init(void); -void bdrv_init_with_whitelist(void); -bool bdrv_uses_whitelist(void); -int bdrv_is_whitelisted(BlockDriver *drv, bool read_only); -BlockDriver *bdrv_find_protocol(const char *filename, - bool allow_protocol_prefix, - Error **errp); -BlockDriver *bdrv_find_format(const char *format_name); -int bdrv_create(BlockDriver *drv, const char* filename, - QemuOpts *opts, Error **errp); -int bdrv_create_file(const char *filename, QemuOpts *opts, Error **errp); - -BlockDriverState *bdrv_new(void); -int bdrv_append(BlockDriverState *bs_new, BlockDriverState *bs_top, - Error **errp); -int bdrv_replace_node(BlockDriverState *from, BlockDriverState *to, - Error **errp); -int bdrv_replace_child_bs(BdrvChild *child, BlockDriverState *new_bs, - Error **errp); -BlockDriverState *bdrv_insert_node(BlockDriverState *bs, QDict *node_options, - int flags, Error **errp); -int bdrv_drop_filter(BlockDriverState *bs, Error **errp); - -int bdrv_parse_aio(const char *mode, int *flags); -int bdrv_parse_cache_mode(const char *mode, int *flags, bool *writethrough); -int bdrv_parse_discard_flags(const char *mode, int *flags); -BdrvChild *bdrv_open_child(const char *filename, - QDict *options, const char *bdref_key, - BlockDriverState* parent, - const BdrvChildClass *child_class, - BdrvChildRole child_role, - bool allow_none, Error **errp); -BlockDriverState *bdrv_open_blockdev_ref(BlockdevRef *ref, Error **errp); -int bdrv_set_backing_hd(BlockDriverState *bs, BlockDriverState *backing_hd, - Error **errp); -int bdrv_open_backing_file(BlockDriverState *bs, QDict *parent_options, - const char *bdref_key, Error **errp); -BlockDriverState *bdrv_open(const char *filename, const char *reference, - QDict *options, int flags, Error **errp); -BlockDriverState *bdrv_new_open_driver_opts(BlockDriver *drv, - const char *node_name, - QDict *options, int flags, - Error **errp); -BlockDriverState *bdrv_new_open_driver(BlockDriver *drv, const char *node_name, - int flags, Error **errp); -BlockReopenQueue *bdrv_reopen_queue(BlockReopenQueue *bs_queue, - BlockDriverState *bs, QDict *options, - bool keep_old_opts); -void bdrv_reopen_queue_free(BlockReopenQueue *bs_queue); -int bdrv_reopen_multiple(BlockReopenQueue *bs_queue, Error **errp); -int bdrv_reopen(BlockDriverState *bs, QDict *opts, bool keep_old_opts, - Error **errp); -int bdrv_reopen_set_read_only(BlockDriverState *bs, bool read_only, - Error **errp); -int bdrv_pwrite_zeroes(BdrvChild *child, int64_t offset, - int64_t bytes, BdrvRequestFlags flags); -int bdrv_make_zero(BdrvChild *child, BdrvRequestFlags flags); -int bdrv_pread(BdrvChild *child, int64_t offset, void *buf, int64_t bytes); -int bdrv_pwrite(BdrvChild *child, int64_t offset, const void *buf, - int64_t bytes); -int bdrv_pwrite_sync(BdrvChild *child, int64_t offset, - const void *buf, int64_t bytes); -/* - * Efficiently zero a region of the disk image. Note that this is a regular - * I/O request like read or write and should have a reasonable size. This - * function is not suitable for zeroing the entire image in a single request - * because it may allocate memory for the entire region. - */ -int coroutine_fn bdrv_co_pwrite_zeroes(BdrvChild *child, int64_t offset, - int64_t bytes, BdrvRequestFlags flags); -BlockDriverState *bdrv_find_backing_image(BlockDriverState *bs, - const char *backing_file); -void bdrv_refresh_filename(BlockDriverState *bs); - -int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact, - PreallocMode prealloc, BdrvRequestFlags flags, - Error **errp); -int generated_co_wrapper -bdrv_truncate(BdrvChild *child, int64_t offset, bool exact, - PreallocMode prealloc, BdrvRequestFlags flags, Error **errp); - -int64_t bdrv_nb_sectors(BlockDriverState *bs); -int64_t bdrv_getlength(BlockDriverState *bs); -int64_t bdrv_get_allocated_file_size(BlockDriverState *bs); -BlockMeasureInfo *bdrv_measure(BlockDriver *drv, QemuOpts *opts, - BlockDriverState *in_bs, Error **errp); -void bdrv_get_geometry(BlockDriverState *bs, uint64_t *nb_sectors_ptr); -void bdrv_refresh_limits(BlockDriverState *bs, Transaction *tran, Error **errp); -int bdrv_commit(BlockDriverState *bs); -int bdrv_make_empty(BdrvChild *c, Error **errp); -int bdrv_change_backing_file(BlockDriverState *bs, const char *backing_file, - const char *backing_fmt, bool warn); -void bdrv_register(BlockDriver *bdrv); -int bdrv_drop_intermediate(BlockDriverState *top, BlockDriverState *base, - const char *backing_file_str); -BlockDriverState *bdrv_find_overlay(BlockDriverState *active, - BlockDriverState *bs); -BlockDriverState *bdrv_find_base(BlockDriverState *bs); -bool bdrv_is_backing_chain_frozen(BlockDriverState *bs, BlockDriverState *base, - Error **errp); -int bdrv_freeze_backing_chain(BlockDriverState *bs, BlockDriverState *base, - Error **errp); -void bdrv_unfreeze_backing_chain(BlockDriverState *bs, BlockDriverState *base); -int coroutine_fn bdrv_co_delete_file(BlockDriverState *bs, Error **errp); -void coroutine_fn bdrv_co_delete_file_noerr(BlockDriverState *bs); - - -typedef struct BdrvCheckResult { - int corruptions; - int leaks; - int check_errors; - int corruptions_fixed; - int leaks_fixed; - int64_t image_end_offset; - BlockFragInfo bfi; -} BdrvCheckResult; - -typedef enum { - BDRV_FIX_LEAKS = 1, - BDRV_FIX_ERRORS = 2, -} BdrvCheckMode; - -int generated_co_wrapper bdrv_check(BlockDriverState *bs, BdrvCheckResult *res, - BdrvCheckMode fix); - -/* The units of offset and total_work_size may be chosen arbitrarily by the - * block driver; total_work_size may change during the course of the amendment - * operation */ -typedef void BlockDriverAmendStatusCB(BlockDriverState *bs, int64_t offset, - int64_t total_work_size, void *opaque); -int bdrv_amend_options(BlockDriverState *bs_new, QemuOpts *opts, - BlockDriverAmendStatusCB *status_cb, void *cb_opaque, - bool force, - Error **errp); - -/* check if a named node can be replaced when doing drive-mirror */ -BlockDriverState *check_to_replace_node(BlockDriverState *parent_bs, - const char *node_name, Error **errp); - -/* async block I/O */ -void bdrv_aio_cancel(BlockAIOCB *acb); -void bdrv_aio_cancel_async(BlockAIOCB *acb); - -/* sg packet commands */ -int bdrv_co_ioctl(BlockDriverState *bs, int req, void *buf); - -/* Invalidate any cached metadata used by image formats */ -int generated_co_wrapper bdrv_invalidate_cache(BlockDriverState *bs, - Error **errp); -int bdrv_activate(BlockDriverState *bs, Error **errp); -void bdrv_activate_all(Error **errp); -int bdrv_inactivate_all(void); - -/* Ensure contents are flushed to disk. */ -int generated_co_wrapper bdrv_flush(BlockDriverState *bs); -int coroutine_fn bdrv_co_flush(BlockDriverState *bs); -int bdrv_flush_all(void); -void bdrv_close_all(void); -void bdrv_drain(BlockDriverState *bs); -void coroutine_fn bdrv_co_drain(BlockDriverState *bs); -void bdrv_drain_all_begin(void); -void bdrv_drain_all_end(void); -void bdrv_drain_all(void); - -#define BDRV_POLL_WHILE(bs, cond) ({ \ - BlockDriverState *bs_ = (bs); \ - AIO_WAIT_WHILE(bdrv_get_aio_context(bs_), \ - cond); }) - -int generated_co_wrapper bdrv_pdiscard(BdrvChild *child, int64_t offset, - int64_t bytes); -int bdrv_co_pdiscard(BdrvChild *child, int64_t offset, int64_t bytes); -int bdrv_has_zero_init_1(BlockDriverState *bs); -int bdrv_has_zero_init(BlockDriverState *bs); -bool bdrv_can_write_zeroes_with_unmap(BlockDriverState *bs); -int bdrv_block_status(BlockDriverState *bs, int64_t offset, - int64_t bytes, int64_t *pnum, int64_t *map, - BlockDriverState **file); -int bdrv_block_status_above(BlockDriverState *bs, BlockDriverState *base, - int64_t offset, int64_t bytes, int64_t *pnum, - int64_t *map, BlockDriverState **file); -int bdrv_is_allocated(BlockDriverState *bs, int64_t offset, int64_t bytes, - int64_t *pnum); -int bdrv_is_allocated_above(BlockDriverState *top, BlockDriverState *base, - bool include_base, int64_t offset, int64_t bytes, - int64_t *pnum); -int coroutine_fn bdrv_co_is_zero_fast(BlockDriverState *bs, int64_t offset, - int64_t bytes); - -bool bdrv_is_read_only(BlockDriverState *bs); -int bdrv_can_set_read_only(BlockDriverState *bs, bool read_only, - bool ignore_allow_rdw, Error **errp); -int bdrv_apply_auto_read_only(BlockDriverState *bs, const char *errmsg, - Error **errp); -bool bdrv_is_writable(BlockDriverState *bs); -bool bdrv_is_sg(BlockDriverState *bs); -bool bdrv_is_inserted(BlockDriverState *bs); -void bdrv_lock_medium(BlockDriverState *bs, bool locked); -void bdrv_eject(BlockDriverState *bs, bool eject_flag); -const char *bdrv_get_format_name(BlockDriverState *bs); -BlockDriverState *bdrv_find_node(const char *node_name); -BlockDeviceInfoList *bdrv_named_nodes_list(bool flat, Error **errp); -XDbgBlockGraph *bdrv_get_xdbg_block_graph(Error **errp); -BlockDriverState *bdrv_lookup_bs(const char *device, - const char *node_name, - Error **errp); -bool bdrv_chain_contains(BlockDriverState *top, BlockDriverState *base); -BlockDriverState *bdrv_next_node(BlockDriverState *bs); -BlockDriverState *bdrv_next_all_states(BlockDriverState *bs); - -typedef struct BdrvNextIterator { - enum { - BDRV_NEXT_BACKEND_ROOTS, - BDRV_NEXT_MONITOR_OWNED, - } phase; - BlockBackend *blk; - BlockDriverState *bs; -} BdrvNextIterator; - -BlockDriverState *bdrv_first(BdrvNextIterator *it); -BlockDriverState *bdrv_next(BdrvNextIterator *it); -void bdrv_next_cleanup(BdrvNextIterator *it); - -BlockDriverState *bdrv_next_monitor_owned(BlockDriverState *bs); -bool bdrv_supports_compressed_writes(BlockDriverState *bs); -void bdrv_iterate_format(void (*it)(void *opaque, const char *name), - void *opaque, bool read_only); -const char *bdrv_get_node_name(const BlockDriverState *bs); -const char *bdrv_get_device_name(const BlockDriverState *bs); -const char *bdrv_get_device_or_node_name(const BlockDriverState *bs); -int bdrv_get_flags(BlockDriverState *bs); -int bdrv_get_info(BlockDriverState *bs, BlockDriverInfo *bdi); -ImageInfoSpecific *bdrv_get_specific_info(BlockDriverState *bs, - Error **errp); -BlockStatsSpecific *bdrv_get_specific_stats(BlockDriverState *bs); -void bdrv_round_to_clusters(BlockDriverState *bs, - int64_t offset, int64_t bytes, - int64_t *cluster_offset, - int64_t *cluster_bytes); - -void bdrv_get_backing_filename(BlockDriverState *bs, - char *filename, int filename_size); -char *bdrv_get_full_backing_filename(BlockDriverState *bs, Error **errp); -char *bdrv_get_full_backing_filename_from_filename(const char *backed, - const char *backing, - Error **errp); -char *bdrv_dirname(BlockDriverState *bs, Error **errp); - -int path_has_protocol(const char *path); -int path_is_absolute(const char *path); -char *path_combine(const char *base_path, const char *filename); - -int generated_co_wrapper -bdrv_readv_vmstate(BlockDriverState *bs, QEMUIOVector *qiov, int64_t pos); -int generated_co_wrapper -bdrv_writev_vmstate(BlockDriverState *bs, QEMUIOVector *qiov, int64_t pos); -int bdrv_save_vmstate(BlockDriverState *bs, const uint8_t *buf, - int64_t pos, int size); - -int bdrv_load_vmstate(BlockDriverState *bs, uint8_t *buf, - int64_t pos, int size); - -void bdrv_img_create(const char *filename, const char *fmt, - const char *base_filename, const char *base_fmt, - char *options, uint64_t img_size, int flags, - bool quiet, Error **errp); - -/* Returns the alignment in bytes that is required so that no bounce buffer - * is required throughout the stack */ -size_t bdrv_min_mem_align(BlockDriverState *bs); -/* Returns optimal alignment in bytes for bounce buffer */ -size_t bdrv_opt_mem_align(BlockDriverState *bs); -void *qemu_blockalign(BlockDriverState *bs, size_t size); -void *qemu_blockalign0(BlockDriverState *bs, size_t size); -void *qemu_try_blockalign(BlockDriverState *bs, size_t size); -void *qemu_try_blockalign0(BlockDriverState *bs, size_t size); -bool bdrv_qiov_is_aligned(BlockDriverState *bs, QEMUIOVector *qiov); - -void bdrv_enable_copy_on_read(BlockDriverState *bs); -void bdrv_disable_copy_on_read(BlockDriverState *bs); - -void bdrv_ref(BlockDriverState *bs); -void bdrv_unref(BlockDriverState *bs); -void bdrv_unref_child(BlockDriverState *parent, BdrvChild *child); -BdrvChild *bdrv_attach_child(BlockDriverState *parent_bs, - BlockDriverState *child_bs, - const char *child_name, - const BdrvChildClass *child_class, - BdrvChildRole child_role, - Error **errp); - -bool bdrv_op_is_blocked(BlockDriverState *bs, BlockOpType op, Error **errp); -void bdrv_op_block(BlockDriverState *bs, BlockOpType op, Error *reason); -void bdrv_op_unblock(BlockDriverState *bs, BlockOpType op, Error *reason); -void bdrv_op_block_all(BlockDriverState *bs, Error *reason); -void bdrv_op_unblock_all(BlockDriverState *bs, Error *reason); -bool bdrv_op_blocker_is_empty(BlockDriverState *bs); - -#define BLKDBG_EVENT(child, evt) \ - do { \ - if (child) { \ - bdrv_debug_event(child->bs, evt); \ - } \ - } while (0) - -void bdrv_debug_event(BlockDriverState *bs, BlkdebugEvent event); - -int bdrv_debug_breakpoint(BlockDriverState *bs, const char *event, - const char *tag); -int bdrv_debug_remove_breakpoint(BlockDriverState *bs, const char *tag); -int bdrv_debug_resume(BlockDriverState *bs, const char *tag); -bool bdrv_debug_is_suspended(BlockDriverState *bs, const char *tag); - -/** - * bdrv_get_aio_context: + * QEMU System Emulator block driver * - * Returns: the currently bound #AioContext - */ -AioContext *bdrv_get_aio_context(BlockDriverState *bs); - -/** - * Move the current coroutine to the AioContext of @bs and return the old - * AioContext of the coroutine. Increase bs->in_flight so that draining @bs - * will wait for the operation to proceed until the corresponding - * bdrv_co_leave(). + * Copyright (c) 2003 Fabrice Bellard * - * Consequently, you can't call drain inside a bdrv_co_enter/leave() section as - * this will deadlock. - */ -AioContext *coroutine_fn bdrv_co_enter(BlockDriverState *bs); - -/** - * Ends a section started by bdrv_co_enter(). Move the current coroutine back - * to old_ctx and decrease bs->in_flight again. - */ -void coroutine_fn bdrv_co_leave(BlockDriverState *bs, AioContext *old_ctx); - -/** - * Locks the AioContext of @bs if it's not the current AioContext. This avoids - * double locking which could lead to deadlocks: This is a coroutine_fn, so we - * know we already own the lock of the current AioContext. + * 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: * - * May only be called in the main thread. - */ -void coroutine_fn bdrv_co_lock(BlockDriverState *bs); - -/** - * Unlocks the AioContext of @bs if it's not the current AioContext. - */ -void coroutine_fn bdrv_co_unlock(BlockDriverState *bs); - -/** - * Transfer control to @co in the aio context of @bs - */ -void bdrv_coroutine_enter(BlockDriverState *bs, Coroutine *co); - -void bdrv_set_aio_context_ignore(BlockDriverState *bs, - AioContext *new_context, GSList **ignore); -int bdrv_try_set_aio_context(BlockDriverState *bs, AioContext *ctx, - Error **errp); -int bdrv_child_try_set_aio_context(BlockDriverState *bs, AioContext *ctx, - BdrvChild *ignore_child, Error **errp); -bool bdrv_child_can_set_aio_context(BdrvChild *c, AioContext *ctx, - GSList **ignore, Error **errp); -bool bdrv_can_set_aio_context(BlockDriverState *bs, AioContext *ctx, - GSList **ignore, Error **errp); -AioContext *bdrv_child_get_parent_aio_context(BdrvChild *c); -AioContext *child_of_bds_get_parent_aio_context(BdrvChild *c); - -int bdrv_probe_blocksizes(BlockDriverState *bs, BlockSizes *bsz); -int bdrv_probe_geometry(BlockDriverState *bs, HDGeometry *geo); - -void bdrv_io_plug(BlockDriverState *bs); -void bdrv_io_unplug(BlockDriverState *bs); - -/** - * bdrv_parent_drained_begin_single: + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. * - * Begin a quiesced section for the parent of @c. If @poll is true, wait for - * any pending activity to cease. + * 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. */ -void bdrv_parent_drained_begin_single(BdrvChild *c, bool poll); - -/** - * bdrv_parent_drained_end_single: - * - * End a quiesced section for the parent of @c. - * - * This polls @bs's AioContext until all scheduled sub-drained_ends - * have settled, which may result in graph changes. - */ -void bdrv_parent_drained_end_single(BdrvChild *c); - -/** - * bdrv_drain_poll: - * - * Poll for pending requests in @bs, its parents (except for @ignore_parent), - * and if @recursive is true its children as well (used for subtree drain). - * - * If @ignore_bds_parents is true, parents that are BlockDriverStates must - * ignore the drain request because they will be drained separately (used for - * drain_all). - * - * This is part of bdrv_drained_begin. - */ -bool bdrv_drain_poll(BlockDriverState *bs, bool recursive, - BdrvChild *ignore_parent, bool ignore_bds_parents); - -/** - * bdrv_drained_begin: - * - * Begin a quiesced section for exclusive access to the BDS, by disabling - * external request sources including NBD server, block jobs, and device model. - * - * This function can be recursive. - */ -void bdrv_drained_begin(BlockDriverState *bs); - -/** - * bdrv_do_drained_begin_quiesce: - * - * Quiesces a BDS like bdrv_drained_begin(), but does not wait for already - * running requests to complete. - */ -void bdrv_do_drained_begin_quiesce(BlockDriverState *bs, - BdrvChild *parent, bool ignore_bds_parents); - -/** - * Like bdrv_drained_begin, but recursively begins a quiesced section for - * exclusive access to all child nodes as well. - */ -void bdrv_subtree_drained_begin(BlockDriverState *bs); - -/** - * bdrv_drained_end: - * - * End a quiescent section started by bdrv_drained_begin(). - * - * This polls @bs's AioContext until all scheduled sub-drained_ends - * have settled. On one hand, that may result in graph changes. On - * the other, this requires that the caller either runs in the main - * loop; or that all involved nodes (@bs and all of its parents) are - * in the caller's AioContext. - */ -void bdrv_drained_end(BlockDriverState *bs); - -/** - * bdrv_drained_end_no_poll: - * - * Same as bdrv_drained_end(), but do not poll for the subgraph to - * actually become unquiesced. Therefore, no graph changes will occur - * with this function. - * - * *drained_end_counter is incremented for every background operation - * that is scheduled, and will be decremented for every operation once - * it settles. The caller must poll until it reaches 0. The counter - * should be accessed using atomic operations only. - */ -void bdrv_drained_end_no_poll(BlockDriverState *bs, int *drained_end_counter); - -/** - * 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); - -bool bdrv_can_store_new_dirty_bitmap(BlockDriverState *bs, const char *name, - uint32_t granularity, Error **errp); -/** - * - * bdrv_register_buf/bdrv_unregister_buf: - * - * Register/unregister a buffer for I/O. For example, VFIO drivers are - * interested to know the memory areas that would later be used for I/O, so - * that they can prepare IOMMU mapping etc., to get better performance. - */ -void bdrv_register_buf(BlockDriverState *bs, void *host, size_t size); -void bdrv_unregister_buf(BlockDriverState *bs, void *host); +#ifndef BLOCK_H +#define BLOCK_H -/** - * - * bdrv_co_copy_range: - * - * Do offloaded copy between two children. If the operation is not implemented - * by the driver, or if the backend storage doesn't support it, a negative - * error code will be returned. - * - * Note: block layer doesn't emulate or fallback to a bounce buffer approach - * because usually the caller shouldn't attempt offloaded copy any more (e.g. - * calling copy_file_range(2)) after the first error, thus it should fall back - * to a read+write path in the caller level. - * - * @src: Source child to copy data from - * @src_offset: offset in @src image to read data - * @dst: Destination child to copy data to - * @dst_offset: offset in @dst image to write data - * @bytes: number of bytes to copy - * @flags: request flags. Supported flags: - * BDRV_REQ_ZERO_WRITE - treat the @src range as zero data and do zero - * write on @dst as if bdrv_co_pwrite_zeroes is - * called. Used to simplify caller code, or - * during BlockDriver.bdrv_co_copy_range_from() - * recursion. - * BDRV_REQ_NO_SERIALISING - do not serialize with other overlapping - * requests currently in flight. - * - * Returns: 0 if succeeded; negative error code if failed. - **/ -int coroutine_fn bdrv_co_copy_range(BdrvChild *src, int64_t src_offset, - BdrvChild *dst, int64_t dst_offset, - int64_t bytes, BdrvRequestFlags read_flags, - BdrvRequestFlags write_flags); +#include "block-global-state.h" +#include "block-io.h" -void bdrv_cancel_in_flight(BlockDriverState *bs); +/* DO NOT ADD ANYTHING IN HERE. USE ONE OF THE HEADERS INCLUDED ABOVE */ -#endif +#endif /* BLOCK_H */ diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ #define NOT_DONE 0x7fffffff /* used while emulated sync operation in progress */ +/* Protected by BQL */ static QTAILQ_HEAD(, BlockDriverState) graph_bdrv_states = QTAILQ_HEAD_INITIALIZER(graph_bdrv_states); +/* Protected by BQL */ static QTAILQ_HEAD(, BlockDriverState) all_bdrv_states = QTAILQ_HEAD_INITIALIZER(all_bdrv_states); +/* Protected by BQL */ static QLIST_HEAD(, BlockDriver) bdrv_drivers = QLIST_HEAD_INITIALIZER(bdrv_drivers); diff --git a/block/meson.build b/block/meson.build index XXXXXXX..XXXXXXX 100644 --- a/block/meson.build +++ b/block/meson.build @@ -XXX,XX +XXX,XX @@ block_ss.add(module_block_h) wrapper_py = find_program('../scripts/block-coroutine-wrapper.py') block_gen_c = custom_target('block-gen.c', output: 'block-gen.c', - input: files('../include/block/block.h', - 'coroutines.h'), + input: files( + '../include/block/block-io.h', + '../include/block/block-global-state.h', + 'coroutines.h' + ), command: [wrapper_py, '@OUTPUT@', '@INPUT@']) block_ss.add(block_gen_c) -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> All the global state (GS) API functions will check that qemu_in_main_thread() returns true. If not, it means that the safety of BQL cannot be guaranteed, and they need to be moved to I/O. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Message-Id: <20220303151616.325444-5-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block.c | 125 ++++++++++++++++++++++++++++++++++++++++++++++++- block/commit.c | 2 + block/io.c | 11 +++++ blockdev.c | 1 + 4 files changed, 137 insertions(+), 2 deletions(-) diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ static char *bdrv_make_absolute_filename(BlockDriverState *relative_to, char *bdrv_get_full_backing_filename(BlockDriverState *bs, Error **errp) { + GLOBAL_STATE_CODE(); return bdrv_make_absolute_filename(bs, bs->backing_file, errp); } void bdrv_register(BlockDriver *bdrv) { assert(bdrv->format_name); + GLOBAL_STATE_CODE(); QLIST_INSERT_HEAD(&bdrv_drivers, bdrv, list); } @@ -XXX,XX +XXX,XX @@ BlockDriverState *bdrv_new(void) BlockDriverState *bs; int i; + GLOBAL_STATE_CODE(); + bs = g_new0(BlockDriverState, 1); QLIST_INIT(&bs->dirty_bitmaps); for (i = 0; i < BLOCK_OP_TYPE_MAX; i++) { @@ -XXX,XX +XXX,XX @@ BlockDriver *bdrv_find_format(const char *format_name) BlockDriver *drv1; int i; + GLOBAL_STATE_CODE(); + drv1 = bdrv_do_find_format(format_name); if (drv1) { return drv1; @@ -XXX,XX +XXX,XX @@ static int bdrv_format_is_whitelisted(const char *format_name, bool read_only) int bdrv_is_whitelisted(BlockDriver *drv, bool read_only) { + GLOBAL_STATE_CODE(); return bdrv_format_is_whitelisted(drv->format_name, read_only); } @@ -XXX,XX +XXX,XX @@ int bdrv_create(BlockDriver *drv, const char* filename, { int ret; + GLOBAL_STATE_CODE(); + Coroutine *co; CreateCo cco = { .drv = drv, @@ -XXX,XX +XXX,XX @@ int bdrv_create_file(const char *filename, QemuOpts *opts, Error **errp) QDict *qdict; int ret; + GLOBAL_STATE_CODE(); + drv = bdrv_find_protocol(filename, true, errp); if (drv == NULL) { return -ENOENT; @@ -XXX,XX +XXX,XX @@ int bdrv_probe_blocksizes(BlockDriverState *bs, BlockSizes *bsz) { BlockDriver *drv = bs->drv; BlockDriverState *filtered = bdrv_filter_bs(bs); + GLOBAL_STATE_CODE(); if (drv && drv->bdrv_probe_blocksizes) { return drv->bdrv_probe_blocksizes(bs, bsz); @@ -XXX,XX +XXX,XX @@ int bdrv_probe_geometry(BlockDriverState *bs, HDGeometry *geo) { BlockDriver *drv = bs->drv; BlockDriverState *filtered = bdrv_filter_bs(bs); + GLOBAL_STATE_CODE(); if (drv && drv->bdrv_probe_geometry) { return drv->bdrv_probe_geometry(bs, geo); @@ -XXX,XX +XXX,XX @@ BlockDriver *bdrv_find_protocol(const char *filename, const char *p; int i; + GLOBAL_STATE_CODE(); /* TODO Drivers without bdrv_file_open must be specified explicitly */ /* @@ -XXX,XX +XXX,XX @@ BlockDriverState *bdrv_new_open_driver_opts(BlockDriver *drv, BlockDriverState *bs; int ret; + GLOBAL_STATE_CODE(); + bs = bdrv_new(); bs->open_flags = flags; bs->options = options ?: qdict_new(); @@ -XXX,XX +XXX,XX @@ BlockDriverState *bdrv_new_open_driver_opts(BlockDriver *drv, BlockDriverState *bdrv_new_open_driver(BlockDriver *drv, const char *node_name, int flags, Error **errp) { + GLOBAL_STATE_CODE(); return bdrv_new_open_driver_opts(drv, node_name, NULL, flags, errp); } @@ -XXX,XX +XXX,XX @@ BdrvChild *bdrv_attach_child(BlockDriverState *parent_bs, BdrvChild *child = NULL; Transaction *tran = tran_new(); + GLOBAL_STATE_CODE(); + ret = bdrv_attach_child_noperm(parent_bs, child_bs, child_name, child_class, child_role, &child, tran, errp); if (ret < 0) { @@ -XXX,XX +XXX,XX @@ void bdrv_root_unref_child(BdrvChild *child) { BlockDriverState *child_bs; + GLOBAL_STATE_CODE(); + child_bs = child->bs; bdrv_detach_child(&child); bdrv_unref(child_bs); @@ -XXX,XX +XXX,XX @@ static void bdrv_unset_inherits_from(BlockDriverState *root, BdrvChild *child, /* Callers must ensure that child->frozen is false. */ void bdrv_unref_child(BlockDriverState *parent, BdrvChild *child) { + GLOBAL_STATE_CODE(); if (child == NULL) { return; } @@ -XXX,XX +XXX,XX @@ int bdrv_set_backing_hd(BlockDriverState *bs, BlockDriverState *backing_hd, int ret; Transaction *tran = tran_new(); + GLOBAL_STATE_CODE(); bdrv_drained_begin(bs); ret = bdrv_set_backing_noperm(bs, backing_hd, tran, errp); @@ -XXX,XX +XXX,XX @@ int bdrv_open_backing_file(BlockDriverState *bs, QDict *parent_options, QDict *tmp_parent_options = NULL; Error *local_err = NULL; + GLOBAL_STATE_CODE(); + if (bs->backing != NULL) { goto free_exit; } @@ -XXX,XX +XXX,XX @@ BdrvChild *bdrv_open_child(const char *filename, { BlockDriverState *bs; + GLOBAL_STATE_CODE(); + bs = bdrv_open_child_bs(filename, options, bdref_key, parent, child_class, child_role, allow_none, errp); if (bs == NULL) { @@ -XXX,XX +XXX,XX @@ BlockDriverState *bdrv_open_blockdev_ref(BlockdevRef *ref, Error **errp) const char *reference = NULL; Visitor *v = NULL; + GLOBAL_STATE_CODE(); + if (ref->type == QTYPE_QSTRING) { reference = ref->u.reference; } else { @@ -XXX,XX +XXX,XX @@ close_and_fail: BlockDriverState *bdrv_open(const char *filename, const char *reference, QDict *options, int flags, Error **errp) { + GLOBAL_STATE_CODE(); + return bdrv_open_inherit(filename, reference, options, flags, NULL, NULL, 0, errp); } @@ -XXX,XX +XXX,XX @@ BlockReopenQueue *bdrv_reopen_queue(BlockReopenQueue *bs_queue, BlockDriverState *bs, QDict *options, bool keep_old_opts) { + GLOBAL_STATE_CODE(); + return bdrv_reopen_queue_child(bs_queue, bs, options, NULL, 0, false, NULL, 0, keep_old_opts); } void bdrv_reopen_queue_free(BlockReopenQueue *bs_queue) { + GLOBAL_STATE_CODE(); if (bs_queue) { BlockReopenQueueEntry *bs_entry, *next; QTAILQ_FOREACH_SAFE(bs_entry, bs_queue, entry, next) { @@ -XXX,XX +XXX,XX @@ int bdrv_reopen(BlockDriverState *bs, QDict *opts, bool keep_old_opts, BlockReopenQueue *queue; int ret; + GLOBAL_STATE_CODE(); + bdrv_subtree_drained_begin(bs); if (ctx != qemu_get_aio_context()) { aio_context_release(ctx); @@ -XXX,XX +XXX,XX @@ int bdrv_reopen_set_read_only(BlockDriverState *bs, bool read_only, { QDict *opts = qdict_new(); + GLOBAL_STATE_CODE(); + qdict_put_bool(opts, BDRV_OPT_READ_ONLY, read_only); return bdrv_reopen(bs, opts, true, errp); @@ -XXX,XX +XXX,XX @@ static void bdrv_close(BlockDriverState *bs) BdrvAioNotifier *ban, *ban_next; BdrvChild *child, *next; + GLOBAL_STATE_CODE(); assert(!bs->refcnt); bdrv_drained_begin(bs); /* complete I/O */ @@ -XXX,XX +XXX,XX @@ static void bdrv_close(BlockDriverState *bs) void bdrv_close_all(void) { assert(job_next(NULL) == NULL); + GLOBAL_STATE_CODE(); /* Drop references from requests still in flight, such as canceled block * jobs whose AIO context has not been polled yet */ @@ -XXX,XX +XXX,XX @@ out: int bdrv_replace_node(BlockDriverState *from, BlockDriverState *to, Error **errp) { + GLOBAL_STATE_CODE(); + return bdrv_replace_node_common(from, to, true, false, errp); } int bdrv_drop_filter(BlockDriverState *bs, Error **errp) { + GLOBAL_STATE_CODE(); + return bdrv_replace_node_common(bs, bdrv_filter_or_cow_bs(bs), true, true, errp); } @@ -XXX,XX +XXX,XX @@ int bdrv_append(BlockDriverState *bs_new, BlockDriverState *bs_top, int ret; Transaction *tran = tran_new(); + GLOBAL_STATE_CODE(); + assert(!bs_new->backing); ret = bdrv_attach_child_noperm(bs_new, bs_top, "backing", @@ -XXX,XX +XXX,XX @@ int bdrv_replace_child_bs(BdrvChild *child, BlockDriverState *new_bs, g_autoptr(GSList) refresh_list = NULL; BlockDriverState *old_bs = child->bs; + GLOBAL_STATE_CODE(); + bdrv_ref(old_bs); bdrv_drained_begin(old_bs); bdrv_drained_begin(new_bs); @@ -XXX,XX +XXX,XX @@ static void bdrv_delete(BlockDriverState *bs) { assert(bdrv_op_blocker_is_empty(bs)); assert(!bs->refcnt); + GLOBAL_STATE_CODE(); /* remove from list, if necessary */ if (bs->node_name[0] != '\0') { @@ -XXX,XX +XXX,XX @@ BlockDriverState *bdrv_insert_node(BlockDriverState *bs, QDict *options, node_name = qdict_get_try_str(options, "node-name"); + GLOBAL_STATE_CODE(); + new_node_bs = bdrv_new_open_driver_opts(drv, node_name, options, flags, errp); options = NULL; /* bdrv_new_open_driver() eats options */ @@ -XXX,XX +XXX,XX @@ int bdrv_change_backing_file(BlockDriverState *bs, const char *backing_file, BlockDriver *drv = bs->drv; int ret; + GLOBAL_STATE_CODE(); + if (!drv) { return -ENOMEDIUM; } @@ -XXX,XX +XXX,XX @@ int bdrv_change_backing_file(BlockDriverState *bs, const char *backing_file, BlockDriverState *bdrv_find_overlay(BlockDriverState *active, BlockDriverState *bs) { + + GLOBAL_STATE_CODE(); + bs = bdrv_skip_filters(bs); active = bdrv_skip_filters(active); @@ -XXX,XX +XXX,XX @@ BlockDriverState *bdrv_find_overlay(BlockDriverState *active, /* Given a BDS, searches for the base layer. */ BlockDriverState *bdrv_find_base(BlockDriverState *bs) { + GLOBAL_STATE_CODE(); + return bdrv_find_overlay(bs, NULL); } @@ -XXX,XX +XXX,XX @@ bool bdrv_is_backing_chain_frozen(BlockDriverState *bs, BlockDriverState *base, BlockDriverState *i; BdrvChild *child; + GLOBAL_STATE_CODE(); + for (i = bs; i != base; i = child_bs(child)) { child = bdrv_filter_or_cow_child(i); @@ -XXX,XX +XXX,XX @@ int bdrv_freeze_backing_chain(BlockDriverState *bs, BlockDriverState *base, BlockDriverState *i; BdrvChild *child; + GLOBAL_STATE_CODE(); + if (bdrv_is_backing_chain_frozen(bs, base, errp)) { return -EPERM; } @@ -XXX,XX +XXX,XX @@ void bdrv_unfreeze_backing_chain(BlockDriverState *bs, BlockDriverState *base) BlockDriverState *i; BdrvChild *child; + GLOBAL_STATE_CODE(); + for (i = bs; i != base; i = child_bs(child)) { child = bdrv_filter_or_cow_child(i); if (child) { @@ -XXX,XX +XXX,XX @@ int bdrv_drop_intermediate(BlockDriverState *top, BlockDriverState *base, g_autoptr(GSList) updated_children = NULL; GSList *p; + GLOBAL_STATE_CODE(); + bdrv_ref(top); bdrv_subtree_drained_begin(top); @@ -XXX,XX +XXX,XX @@ void bdrv_iterate_format(void (*it)(void *opaque, const char *name), int i; const char **formats = NULL; + GLOBAL_STATE_CODE(); + QLIST_FOREACH(drv, &bdrv_drivers, list) { if (drv->format_name) { bool found = false; @@ -XXX,XX +XXX,XX @@ BlockDriverState *bdrv_find_node(const char *node_name) BlockDriverState *bs; assert(node_name); + GLOBAL_STATE_CODE(); QTAILQ_FOREACH(bs, &graph_bdrv_states, node_list) { if (!strcmp(node_name, bs->node_name)) { @@ -XXX,XX +XXX,XX @@ BlockDeviceInfoList *bdrv_named_nodes_list(bool flat, BlockDeviceInfoList *list; BlockDriverState *bs; + GLOBAL_STATE_CODE(); + list = NULL; QTAILQ_FOREACH(bs, &graph_bdrv_states, node_list) { BlockDeviceInfo *info = bdrv_block_device_info(NULL, bs, flat, errp); @@ -XXX,XX +XXX,XX @@ XDbgBlockGraph *bdrv_get_xdbg_block_graph(Error **errp) BdrvChild *child; XDbgBlockGraphConstructor *gr = xdbg_graph_new(); + GLOBAL_STATE_CODE(); + for (blk = blk_all_next(NULL); blk; blk = blk_all_next(blk)) { char *allocated_name = NULL; const char *name = blk_name(blk); @@ -XXX,XX +XXX,XX @@ BlockDriverState *bdrv_lookup_bs(const char *device, BlockBackend *blk; BlockDriverState *bs; + GLOBAL_STATE_CODE(); + if (device) { blk = blk_by_name(device); @@ -XXX,XX +XXX,XX @@ BlockDriverState *bdrv_lookup_bs(const char *device, * return false. If either argument is NULL, return false. */ bool bdrv_chain_contains(BlockDriverState *top, BlockDriverState *base) { + + GLOBAL_STATE_CODE(); + while (top && top != base) { top = bdrv_filter_or_cow_bs(top); } @@ -XXX,XX +XXX,XX @@ bool bdrv_chain_contains(BlockDriverState *top, BlockDriverState *base) BlockDriverState *bdrv_next_node(BlockDriverState *bs) { + GLOBAL_STATE_CODE(); if (!bs) { return QTAILQ_FIRST(&graph_bdrv_states); } @@ -XXX,XX +XXX,XX @@ BlockDriverState *bdrv_next_node(BlockDriverState *bs) BlockDriverState *bdrv_next_all_states(BlockDriverState *bs) { + GLOBAL_STATE_CODE(); if (!bs) { return QTAILQ_FIRST(&all_bdrv_states); } @@ -XXX,XX +XXX,XX @@ const char *bdrv_get_device_or_node_name(const BlockDriverState *bs) int bdrv_get_flags(BlockDriverState *bs) { + GLOBAL_STATE_CODE(); return bs->open_flags; } int bdrv_has_zero_init_1(BlockDriverState *bs) { + GLOBAL_STATE_CODE(); return 1; } int bdrv_has_zero_init(BlockDriverState *bs) { BlockDriverState *filtered; + GLOBAL_STATE_CODE(); if (!bs->drv) { return 0; @@ -XXX,XX +XXX,XX @@ static BlockDriverState *bdrv_find_debug_node(BlockDriverState *bs) int bdrv_debug_breakpoint(BlockDriverState *bs, const char *event, const char *tag) { + GLOBAL_STATE_CODE(); bs = bdrv_find_debug_node(bs); if (bs) { return bs->drv->bdrv_debug_breakpoint(bs, event, tag); @@ -XXX,XX +XXX,XX @@ int bdrv_debug_breakpoint(BlockDriverState *bs, const char *event, int bdrv_debug_remove_breakpoint(BlockDriverState *bs, const char *tag) { + GLOBAL_STATE_CODE(); bs = bdrv_find_debug_node(bs); if (bs) { return bs->drv->bdrv_debug_remove_breakpoint(bs, tag); @@ -XXX,XX +XXX,XX @@ int bdrv_debug_remove_breakpoint(BlockDriverState *bs, const char *tag) int bdrv_debug_resume(BlockDriverState *bs, const char *tag) { + GLOBAL_STATE_CODE(); while (bs && (!bs->drv || !bs->drv->bdrv_debug_resume)) { bs = bdrv_primary_bs(bs); } @@ -XXX,XX +XXX,XX @@ int bdrv_debug_resume(BlockDriverState *bs, const char *tag) bool bdrv_debug_is_suspended(BlockDriverState *bs, const char *tag) { + GLOBAL_STATE_CODE(); while (bs && bs->drv && !bs->drv->bdrv_debug_is_suspended) { bs = bdrv_primary_bs(bs); } @@ -XXX,XX +XXX,XX @@ BlockDriverState *bdrv_find_backing_image(BlockDriverState *bs, BlockDriverState *retval = NULL; BlockDriverState *bs_below; + GLOBAL_STATE_CODE(); + if (!bs || !bs->drv || !backing_file) { return NULL; } @@ -XXX,XX +XXX,XX @@ int bdrv_activate(BlockDriverState *bs, Error **errp) int ret; BdrvDirtyBitmap *bm; + GLOBAL_STATE_CODE(); + if (!bs->drv) { return -ENOMEDIUM; } @@ -XXX,XX +XXX,XX @@ void bdrv_activate_all(Error **errp) BlockDriverState *bs; BdrvNextIterator it; + GLOBAL_STATE_CODE(); + for (bs = bdrv_first(&it); bs; bs = bdrv_next(&it)) { AioContext *aio_context = bdrv_get_aio_context(bs); int ret; @@ -XXX,XX +XXX,XX @@ int bdrv_inactivate_all(void) int ret = 0; GSList *aio_ctxs = NULL, *ctx; + GLOBAL_STATE_CODE(); + for (bs = bdrv_first(&it); bs; bs = bdrv_next(&it)) { AioContext *aio_context = bdrv_get_aio_context(bs); @@ -XXX,XX +XXX,XX @@ void bdrv_lock_medium(BlockDriverState *bs, bool locked) /* Get a reference to bs */ void bdrv_ref(BlockDriverState *bs) { + GLOBAL_STATE_CODE(); bs->refcnt++; } @@ -XXX,XX +XXX,XX @@ void bdrv_ref(BlockDriverState *bs) * deleted. */ void bdrv_unref(BlockDriverState *bs) { + GLOBAL_STATE_CODE(); if (!bs) { return; } @@ -XXX,XX +XXX,XX @@ struct BdrvOpBlocker { bool bdrv_op_is_blocked(BlockDriverState *bs, BlockOpType op, Error **errp) { BdrvOpBlocker *blocker; + GLOBAL_STATE_CODE(); assert((int) op >= 0 && op < BLOCK_OP_TYPE_MAX); if (!QLIST_EMPTY(&bs->op_blockers[op])) { blocker = QLIST_FIRST(&bs->op_blockers[op]); @@ -XXX,XX +XXX,XX @@ bool bdrv_op_is_blocked(BlockDriverState *bs, BlockOpType op, Error **errp) void bdrv_op_block(BlockDriverState *bs, BlockOpType op, Error *reason) { BdrvOpBlocker *blocker; + GLOBAL_STATE_CODE(); assert((int) op >= 0 && op < BLOCK_OP_TYPE_MAX); blocker = g_new0(BdrvOpBlocker, 1); @@ -XXX,XX +XXX,XX @@ void bdrv_op_block(BlockDriverState *bs, BlockOpType op, Error *reason) void bdrv_op_unblock(BlockDriverState *bs, BlockOpType op, Error *reason) { BdrvOpBlocker *blocker, *next; + GLOBAL_STATE_CODE(); assert((int) op >= 0 && op < BLOCK_OP_TYPE_MAX); QLIST_FOREACH_SAFE(blocker, &bs->op_blockers[op], list, next) { if (blocker->reason == reason) { @@ -XXX,XX +XXX,XX @@ void bdrv_op_unblock(BlockDriverState *bs, BlockOpType op, Error *reason) void bdrv_op_block_all(BlockDriverState *bs, Error *reason) { int i; + GLOBAL_STATE_CODE(); for (i = 0; i < BLOCK_OP_TYPE_MAX; i++) { bdrv_op_block(bs, i, reason); } @@ -XXX,XX +XXX,XX @@ void bdrv_op_block_all(BlockDriverState *bs, Error *reason) void bdrv_op_unblock_all(BlockDriverState *bs, Error *reason) { int i; + GLOBAL_STATE_CODE(); for (i = 0; i < BLOCK_OP_TYPE_MAX; i++) { bdrv_op_unblock(bs, i, reason); } @@ -XXX,XX +XXX,XX @@ void bdrv_op_unblock_all(BlockDriverState *bs, Error *reason) bool bdrv_op_blocker_is_empty(BlockDriverState *bs) { int i; - + GLOBAL_STATE_CODE(); for (i = 0; i < BLOCK_OP_TYPE_MAX; i++) { if (!QLIST_EMPTY(&bs->op_blockers[i])) { return false; @@ -XXX,XX +XXX,XX @@ void bdrv_img_create(const char *filename, const char *fmt, Error *local_err = NULL; int ret = 0; + GLOBAL_STATE_CODE(); + /* Find driver and parse its options */ drv = bdrv_find_format(fmt); if (!drv) { @@ -XXX,XX +XXX,XX @@ static bool bdrv_parent_can_set_aio_context(BdrvChild *c, AioContext *ctx, bool bdrv_child_can_set_aio_context(BdrvChild *c, AioContext *ctx, GSList **ignore, Error **errp) { + GLOBAL_STATE_CODE(); if (g_slist_find(*ignore, c)) { return true; } @@ -XXX,XX +XXX,XX @@ bool bdrv_can_set_aio_context(BlockDriverState *bs, AioContext *ctx, return true; } + GLOBAL_STATE_CODE(); + QLIST_FOREACH(c, &bs->parents, next_parent) { if (!bdrv_parent_can_set_aio_context(c, ctx, ignore, errp)) { return false; @@ -XXX,XX +XXX,XX @@ int bdrv_child_try_set_aio_context(BlockDriverState *bs, AioContext *ctx, GSList *ignore; bool ret; + GLOBAL_STATE_CODE(); + ignore = ignore_child ? g_slist_prepend(NULL, ignore_child) : NULL; ret = bdrv_can_set_aio_context(bs, ctx, &ignore, errp); g_slist_free(ignore); @@ -XXX,XX +XXX,XX @@ int bdrv_child_try_set_aio_context(BlockDriverState *bs, AioContext *ctx, int bdrv_try_set_aio_context(BlockDriverState *bs, AioContext *ctx, Error **errp) { + GLOBAL_STATE_CODE(); return bdrv_child_try_set_aio_context(bs, ctx, NULL, errp); } @@ -XXX,XX +XXX,XX @@ void bdrv_add_aio_context_notifier(BlockDriverState *bs, .detach_aio_context = detach_aio_context, .opaque = opaque }; + GLOBAL_STATE_CODE(); QLIST_INSERT_HEAD(&bs->aio_notifiers, ban, list); } @@ -XXX,XX +XXX,XX @@ void bdrv_remove_aio_context_notifier(BlockDriverState *bs, void *opaque) { BdrvAioNotifier *ban, *ban_next; + GLOBAL_STATE_CODE(); QLIST_FOREACH_SAFE(ban, &bs->aio_notifiers, list, ban_next) { if (ban->attached_aio_context == attached_aio_context && @@ -XXX,XX +XXX,XX @@ int bdrv_amend_options(BlockDriverState *bs, QemuOpts *opts, bool force, Error **errp) { + GLOBAL_STATE_CODE(); if (!bs->drv) { error_setg(errp, "Node is ejected"); return -ENOMEDIUM; @@ -XXX,XX +XXX,XX @@ BlockDriverState *check_to_replace_node(BlockDriverState *parent_bs, BlockDriverState *to_replace_bs = bdrv_find_node(node_name); AioContext *aio_context; + GLOBAL_STATE_CODE(); + if (!to_replace_bs) { error_setg(errp, "Failed to find node with node-name='%s'", node_name); return NULL; @@ -XXX,XX +XXX,XX @@ void bdrv_refresh_filename(BlockDriverState *bs) bool generate_json_filename; /* Whether our default implementation should fill exact_filename (false) or not (true) */ + GLOBAL_STATE_CODE(); + if (!drv) { return; } @@ -XXX,XX +XXX,XX @@ char *bdrv_dirname(BlockDriverState *bs, Error **errp) BlockDriver *drv = bs->drv; BlockDriverState *child_bs; + GLOBAL_STATE_CODE(); + if (!drv) { error_setg(errp, "Node '%s' is ejected", bs->node_name); return NULL; @@ -XXX,XX +XXX,XX @@ char *bdrv_dirname(BlockDriverState *bs, Error **errp) void bdrv_add_child(BlockDriverState *parent_bs, BlockDriverState *child_bs, Error **errp) { - + GLOBAL_STATE_CODE(); if (!parent_bs->drv || !parent_bs->drv->bdrv_add_child) { error_setg(errp, "The node %s does not support adding a child", bdrv_get_device_or_node_name(parent_bs)); @@ -XXX,XX +XXX,XX @@ void bdrv_del_child(BlockDriverState *parent_bs, BdrvChild *child, Error **errp) { BdrvChild *tmp; + GLOBAL_STATE_CODE(); if (!parent_bs->drv || !parent_bs->drv->bdrv_del_child) { error_setg(errp, "The node %s does not support removing a child", bdrv_get_device_or_node_name(parent_bs)); @@ -XXX,XX +XXX,XX @@ int bdrv_make_empty(BdrvChild *c, Error **errp) BlockDriver *drv = c->bs->drv; int ret; + GLOBAL_STATE_CODE(); assert(c->perm & (BLK_PERM_WRITE | BLK_PERM_WRITE_UNCHANGED)); if (!drv->bdrv_make_empty) { 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 @@ int bdrv_commit(BlockDriverState *bs) QEMU_AUTO_VFREE uint8_t *buf = NULL; Error *local_err = NULL; + GLOBAL_STATE_CODE(); + if (!drv) return -ENOMEDIUM; diff --git a/block/io.c b/block/io.c index XXXXXXX..XXXXXXX 100644 --- a/block/io.c +++ b/block/io.c @@ -XXX,XX +XXX,XX @@ void bdrv_refresh_limits(BlockDriverState *bs, Transaction *tran, Error **errp) BdrvChild *c; bool have_limits; + GLOBAL_STATE_CODE(); + if (tran) { BdrvRefreshLimitsState *s = g_new(BdrvRefreshLimitsState, 1); *s = (BdrvRefreshLimitsState) { @@ -XXX,XX +XXX,XX @@ static bool bdrv_drain_all_poll(void) { BlockDriverState *bs = NULL; bool result = false; + GLOBAL_STATE_CODE(); /* bdrv_drain_poll() can't make changes to the graph and we are holding the * main AioContext lock, so iterating bdrv_next_all_states() is safe. */ @@ -XXX,XX +XXX,XX @@ static bool bdrv_drain_all_poll(void) void bdrv_drain_all_begin(void) { BlockDriverState *bs = NULL; + GLOBAL_STATE_CODE(); if (qemu_in_coroutine()) { bdrv_co_yield_to_drain(NULL, true, false, NULL, true, true, NULL); @@ -XXX,XX +XXX,XX @@ void bdrv_drain_all_end(void) { BlockDriverState *bs = NULL; int drained_end_counter = 0; + GLOBAL_STATE_CODE(); /* * bdrv queue is managed by record/replay, @@ -XXX,XX +XXX,XX @@ void bdrv_drain_all_end(void) void bdrv_drain_all(void) { + GLOBAL_STATE_CODE(); bdrv_drain_all_begin(); bdrv_drain_all_end(); } @@ -XXX,XX +XXX,XX @@ int bdrv_flush_all(void) BlockDriverState *bs = NULL; int result = 0; + GLOBAL_STATE_CODE(); + /* * bdrv queue is managed by record/replay, * creating new flush request for stopping @@ -XXX,XX +XXX,XX @@ void bdrv_register_buf(BlockDriverState *bs, void *host, size_t size) { BdrvChild *child; + GLOBAL_STATE_CODE(); if (bs->drv && bs->drv->bdrv_register_buf) { bs->drv->bdrv_register_buf(bs, host, size); } @@ -XXX,XX +XXX,XX @@ void bdrv_unregister_buf(BlockDriverState *bs, void *host) { BdrvChild *child; + GLOBAL_STATE_CODE(); if (bs->drv && bs->drv->bdrv_unregister_buf) { bs->drv->bdrv_unregister_buf(bs, host); } @@ -XXX,XX +XXX,XX @@ out: void bdrv_cancel_in_flight(BlockDriverState *bs) { + GLOBAL_STATE_CODE(); if (!bs || !bs->drv) { return; } diff --git a/blockdev.c b/blockdev.c index XXXXXXX..XXXXXXX 100644 --- a/blockdev.c +++ b/blockdev.c @@ -XXX,XX +XXX,XX @@ void blockdev_close_all_bdrv_states(void) /* Iterates over the list of monitor-owned BlockDriverStates */ BlockDriverState *bdrv_next_monitor_owned(BlockDriverState *bs) { + GLOBAL_STATE_CODE(); return bs ? QTAILQ_NEXT(bs, monitor_list) : QTAILQ_FIRST(&monitor_bdrv_states); } -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> Mark all I/O functions with IO_CODE, and all "I/O OR GS" with IO_OR_GS_CODE. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Message-Id: <20220303151616.325444-6-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block-io.h | 1 + block.c | 37 +++++++++++++++++++++++++++++++++- block/dirty-bitmap.c | 1 + block/io.c | 43 ++++++++++++++++++++++++++++++++++++++-- 4 files changed, 79 insertions(+), 3 deletions(-) diff --git a/include/block/block-io.h b/include/block/block-io.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block-io.h +++ b/include/block/block-io.h @@ -XXX,XX +XXX,XX @@ void bdrv_drained_end_no_poll(BlockDriverState *bs, int *drained_end_counter); #define BDRV_POLL_WHILE(bs, cond) ({ \ BlockDriverState *bs_ = (bs); \ + IO_OR_GS_CODE(); \ AIO_WAIT_WHILE(bdrv_get_aio_context(bs_), \ cond); }) diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ size_t bdrv_opt_mem_align(BlockDriverState *bs) /* page size or 4k (hdd sector size) should be on the safe side */ return MAX(4096, qemu_real_host_page_size); } + IO_CODE(); return bs->bl.opt_mem_alignment; } @@ -XXX,XX +XXX,XX @@ size_t bdrv_min_mem_align(BlockDriverState *bs) /* page size or 4k (hdd sector size) should be on the safe side */ return MAX(4096, qemu_real_host_page_size); } + IO_CODE(); return bs->bl.min_mem_alignment; } @@ -XXX,XX +XXX,XX @@ void bdrv_parse_filename_strip_prefix(const char *filename, const char *prefix, * image is inactivated. */ bool bdrv_is_read_only(BlockDriverState *bs) { + IO_CODE(); return !(bs->open_flags & BDRV_O_RDWR); } int bdrv_can_set_read_only(BlockDriverState *bs, bool read_only, bool ignore_allow_rdw, Error **errp) { + IO_CODE(); + /* Do not set read_only if copy_on_read is enabled */ if (bs->copy_on_read && read_only) { error_setg(errp, "Can't set node '%s' to r/o with copy-on-read enabled", @@ -XXX,XX +XXX,XX @@ int bdrv_apply_auto_read_only(BlockDriverState *bs, const char *errmsg, Error **errp) { int ret = 0; + IO_CODE(); if (!(bs->open_flags & BDRV_O_RDWR)) { return 0; @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_delete_file(BlockDriverState *bs, Error **errp) Error *local_err = NULL; int ret; + IO_CODE(); assert(bs != NULL); if (!bs->drv) { @@ -XXX,XX +XXX,XX @@ void coroutine_fn bdrv_co_delete_file_noerr(BlockDriverState *bs) { Error *local_err = NULL; int ret; + IO_CODE(); if (!bs) { return; @@ -XXX,XX +XXX,XX @@ static int bdrv_child_cb_update_filename(BdrvChild *c, BlockDriverState *base, AioContext *child_of_bds_get_parent_aio_context(BdrvChild *c) { BlockDriverState *bs = c->opaque; + IO_CODE(); return bdrv_get_aio_context(bs); } @@ -XXX,XX +XXX,XX @@ const BdrvChildClass child_of_bds = { AioContext *bdrv_child_get_parent_aio_context(BdrvChild *c) { + IO_CODE(); return c->klass->get_parent_aio_context(c); } @@ -XXX,XX +XXX,XX @@ static bool bdrv_is_writable_after_reopen(BlockDriverState *bs, */ bool bdrv_is_writable(BlockDriverState *bs) { + IO_CODE(); return bdrv_is_writable_after_reopen(bs, NULL); } @@ -XXX,XX +XXX,XX @@ static int64_t bdrv_sum_allocated_file_size(BlockDriverState *bs) int64_t bdrv_get_allocated_file_size(BlockDriverState *bs) { BlockDriver *drv = bs->drv; + IO_CODE(); + if (!drv) { return -ENOMEDIUM; } @@ -XXX,XX +XXX,XX @@ int64_t bdrv_get_allocated_file_size(BlockDriverState *bs) BlockMeasureInfo *bdrv_measure(BlockDriver *drv, QemuOpts *opts, BlockDriverState *in_bs, Error **errp) { + IO_CODE(); if (!drv->bdrv_measure) { error_setg(errp, "Block driver '%s' does not support size measurement", drv->format_name); @@ -XXX,XX +XXX,XX @@ BlockMeasureInfo *bdrv_measure(BlockDriver *drv, QemuOpts *opts, int64_t bdrv_nb_sectors(BlockDriverState *bs) { BlockDriver *drv = bs->drv; + IO_CODE(); if (!drv) return -ENOMEDIUM; @@ -XXX,XX +XXX,XX @@ int64_t bdrv_nb_sectors(BlockDriverState *bs) int64_t bdrv_getlength(BlockDriverState *bs) { int64_t ret = bdrv_nb_sectors(bs); + IO_CODE(); if (ret < 0) { return ret; @@ -XXX,XX +XXX,XX @@ int64_t bdrv_getlength(BlockDriverState *bs) void bdrv_get_geometry(BlockDriverState *bs, uint64_t *nb_sectors_ptr) { int64_t nb_sectors = bdrv_nb_sectors(bs); + IO_CODE(); *nb_sectors_ptr = nb_sectors < 0 ? 0 : nb_sectors; } bool bdrv_is_sg(BlockDriverState *bs) { + IO_CODE(); return bs->sg; } @@ -XXX,XX +XXX,XX @@ bool bdrv_is_sg(BlockDriverState *bs) bool bdrv_supports_compressed_writes(BlockDriverState *bs) { BlockDriverState *filtered; + IO_CODE(); if (!bs->drv || !block_driver_can_compress(bs->drv)) { return false; @@ -XXX,XX +XXX,XX @@ bool bdrv_supports_compressed_writes(BlockDriverState *bs) const char *bdrv_get_format_name(BlockDriverState *bs) { + IO_CODE(); return bs->drv ? bs->drv->format_name : NULL; } @@ -XXX,XX +XXX,XX @@ BlockDriverState *bdrv_next_all_states(BlockDriverState *bs) const char *bdrv_get_node_name(const BlockDriverState *bs) { + IO_CODE(); return bs->node_name; } @@ -XXX,XX +XXX,XX @@ const char *bdrv_get_parent_name(const BlockDriverState *bs) /* TODO check what callers really want: bs->node_name or blk_name() */ const char *bdrv_get_device_name(const BlockDriverState *bs) { + IO_CODE(); return bdrv_get_parent_name(bs) ?: ""; } @@ -XXX,XX +XXX,XX @@ const char *bdrv_get_device_name(const BlockDriverState *bs) * absent, then this returns an empty (non-null) string. */ const char *bdrv_get_device_or_node_name(const BlockDriverState *bs) { + IO_CODE(); return bdrv_get_parent_name(bs) ?: bs->node_name; } @@ -XXX,XX +XXX,XX @@ int bdrv_has_zero_init(BlockDriverState *bs) bool bdrv_can_write_zeroes_with_unmap(BlockDriverState *bs) { + IO_CODE(); if (!(bs->open_flags & BDRV_O_UNMAP)) { return false; } @@ -XXX,XX +XXX,XX @@ bool bdrv_can_write_zeroes_with_unmap(BlockDriverState *bs) void bdrv_get_backing_filename(BlockDriverState *bs, char *filename, int filename_size) { + IO_CODE(); pstrcpy(filename, filename_size, bs->backing_file); } @@ -XXX,XX +XXX,XX @@ int bdrv_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) { int ret; BlockDriver *drv = bs->drv; + IO_CODE(); /* if bs->drv == NULL, bs is closed, so there's nothing to do here */ if (!drv) { return -ENOMEDIUM; @@ -XXX,XX +XXX,XX @@ ImageInfoSpecific *bdrv_get_specific_info(BlockDriverState *bs, Error **errp) { BlockDriver *drv = bs->drv; + IO_CODE(); if (drv && drv->bdrv_get_specific_info) { return drv->bdrv_get_specific_info(bs, errp); } @@ -XXX,XX +XXX,XX @@ ImageInfoSpecific *bdrv_get_specific_info(BlockDriverState *bs, BlockStatsSpecific *bdrv_get_specific_stats(BlockDriverState *bs) { BlockDriver *drv = bs->drv; + IO_CODE(); if (!drv || !drv->bdrv_get_specific_stats) { return NULL; } @@ -XXX,XX +XXX,XX @@ BlockStatsSpecific *bdrv_get_specific_stats(BlockDriverState *bs) void bdrv_debug_event(BlockDriverState *bs, BlkdebugEvent event) { + IO_CODE(); if (!bs || !bs->drv || !bs->drv->bdrv_debug_event) { return; } @@ -XXX,XX +XXX,XX @@ bool bdrv_is_inserted(BlockDriverState *bs) { BlockDriver *drv = bs->drv; BdrvChild *child; + IO_CODE(); if (!drv) { return false; @@ -XXX,XX +XXX,XX @@ bool bdrv_is_inserted(BlockDriverState *bs) void bdrv_eject(BlockDriverState *bs, bool eject_flag) { BlockDriver *drv = bs->drv; + IO_CODE(); if (drv && drv->bdrv_eject) { drv->bdrv_eject(bs, eject_flag); @@ -XXX,XX +XXX,XX @@ void bdrv_eject(BlockDriverState *bs, bool eject_flag) void bdrv_lock_medium(BlockDriverState *bs, bool locked) { BlockDriver *drv = bs->drv; - + IO_CODE(); trace_bdrv_lock_medium(bs, locked); if (drv && drv->bdrv_lock_medium) { @@ -XXX,XX +XXX,XX @@ out: AioContext *bdrv_get_aio_context(BlockDriverState *bs) { + IO_CODE(); return bs ? bs->aio_context : qemu_get_aio_context(); } @@ -XXX,XX +XXX,XX @@ AioContext *coroutine_fn bdrv_co_enter(BlockDriverState *bs) Coroutine *self = qemu_coroutine_self(); AioContext *old_ctx = qemu_coroutine_get_aio_context(self); AioContext *new_ctx; + IO_CODE(); /* * Increase bs->in_flight to ensure that this operation is completed before @@ -XXX,XX +XXX,XX @@ AioContext *coroutine_fn bdrv_co_enter(BlockDriverState *bs) void coroutine_fn bdrv_co_leave(BlockDriverState *bs, AioContext *old_ctx) { + IO_CODE(); aio_co_reschedule_self(old_ctx); bdrv_dec_in_flight(bs); } @@ -XXX,XX +XXX,XX @@ void coroutine_fn bdrv_co_unlock(BlockDriverState *bs) void bdrv_coroutine_enter(BlockDriverState *bs, Coroutine *co) { + IO_CODE(); aio_co_enter(bdrv_get_aio_context(bs), co); } diff --git a/block/dirty-bitmap.c b/block/dirty-bitmap.c index XXXXXXX..XXXXXXX 100644 --- a/block/dirty-bitmap.c +++ b/block/dirty-bitmap.c @@ -XXX,XX +XXX,XX @@ static void coroutine_fn bdrv_co_can_store_new_dirty_bitmap_entry(void *opaque) bool bdrv_can_store_new_dirty_bitmap(BlockDriverState *bs, const char *name, uint32_t granularity, Error **errp) { + IO_CODE(); if (qemu_in_coroutine()) { return bdrv_co_can_store_new_dirty_bitmap(bs, name, granularity, errp); } else { 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_drained_end_single_no_poll(BdrvChild *c, void bdrv_parent_drained_end_single(BdrvChild *c) { int drained_end_counter = 0; + IO_OR_GS_CODE(); bdrv_parent_drained_end_single_no_poll(c, &drained_end_counter); BDRV_POLL_WHILE(c->bs, qatomic_read(&drained_end_counter) > 0); } @@ -XXX,XX +XXX,XX @@ static bool bdrv_parent_drained_poll(BlockDriverState *bs, BdrvChild *ignore, void bdrv_parent_drained_begin_single(BdrvChild *c, bool poll) { + IO_OR_GS_CODE(); c->parent_quiesce_counter++; if (c->klass->drained_begin) { c->klass->drained_begin(c); @@ -XXX,XX +XXX,XX @@ void bdrv_refresh_limits(BlockDriverState *bs, Transaction *tran, Error **errp) */ void bdrv_enable_copy_on_read(BlockDriverState *bs) { + IO_CODE(); qatomic_inc(&bs->copy_on_read); } void bdrv_disable_copy_on_read(BlockDriverState *bs) { int old = qatomic_fetch_dec(&bs->copy_on_read); + IO_CODE(); assert(old >= 1); } @@ -XXX,XX +XXX,XX @@ bool bdrv_drain_poll(BlockDriverState *bs, bool recursive, BdrvChild *ignore_parent, bool ignore_bds_parents) { BdrvChild *child, *next; + IO_OR_GS_CODE(); if (bdrv_parent_drained_poll(bs, ignore_parent, ignore_bds_parents)) { return true; @@ -XXX,XX +XXX,XX @@ static void coroutine_fn bdrv_co_yield_to_drain(BlockDriverState *bs, void bdrv_do_drained_begin_quiesce(BlockDriverState *bs, BdrvChild *parent, bool ignore_bds_parents) { + IO_OR_GS_CODE(); assert(!qemu_in_coroutine()); /* Stop things in parent-to-child order */ @@ -XXX,XX +XXX,XX @@ static void bdrv_do_drained_begin(BlockDriverState *bs, bool recursive, void bdrv_drained_begin(BlockDriverState *bs) { + IO_OR_GS_CODE(); bdrv_do_drained_begin(bs, false, NULL, false, true); } void bdrv_subtree_drained_begin(BlockDriverState *bs) { + IO_OR_GS_CODE(); bdrv_do_drained_begin(bs, true, NULL, false, true); } @@ -XXX,XX +XXX,XX @@ static void bdrv_do_drained_end(BlockDriverState *bs, bool recursive, void bdrv_drained_end(BlockDriverState *bs) { int drained_end_counter = 0; + IO_OR_GS_CODE(); bdrv_do_drained_end(bs, false, NULL, false, &drained_end_counter); BDRV_POLL_WHILE(bs, qatomic_read(&drained_end_counter) > 0); } void bdrv_drained_end_no_poll(BlockDriverState *bs, int *drained_end_counter) { + IO_CODE(); bdrv_do_drained_end(bs, false, NULL, false, drained_end_counter); } void bdrv_subtree_drained_end(BlockDriverState *bs) { int drained_end_counter = 0; + IO_OR_GS_CODE(); bdrv_do_drained_end(bs, true, NULL, false, &drained_end_counter); BDRV_POLL_WHILE(bs, qatomic_read(&drained_end_counter) > 0); } @@ -XXX,XX +XXX,XX @@ void bdrv_unapply_subtree_drain(BdrvChild *child, BlockDriverState *old_parent) */ void coroutine_fn bdrv_co_drain(BlockDriverState *bs) { + IO_OR_GS_CODE(); assert(qemu_in_coroutine()); bdrv_drained_begin(bs); bdrv_drained_end(bs); @@ -XXX,XX +XXX,XX @@ void coroutine_fn bdrv_co_drain(BlockDriverState *bs) void bdrv_drain(BlockDriverState *bs) { + IO_OR_GS_CODE(); bdrv_drained_begin(bs); bdrv_drained_end(bs); } @@ -XXX,XX +XXX,XX @@ void bdrv_round_to_clusters(BlockDriverState *bs, int64_t *cluster_bytes) { BlockDriverInfo bdi; - + IO_CODE(); if (bdrv_get_info(bs, &bdi) < 0 || bdi.cluster_size == 0) { *cluster_offset = offset; *cluster_bytes = bytes; @@ -XXX,XX +XXX,XX @@ static int bdrv_check_request32(int64_t offset, int64_t bytes, int bdrv_pwrite_zeroes(BdrvChild *child, int64_t offset, int64_t bytes, BdrvRequestFlags flags) { + IO_CODE(); return bdrv_pwritev(child, offset, bytes, NULL, BDRV_REQ_ZERO_WRITE | flags); } @@ -XXX,XX +XXX,XX @@ int bdrv_make_zero(BdrvChild *child, BdrvRequestFlags flags) int ret; int64_t target_size, bytes, offset = 0; BlockDriverState *bs = child->bs; + IO_CODE(); target_size = bdrv_getlength(bs); if (target_size < 0) { @@ -XXX,XX +XXX,XX @@ int bdrv_pread(BdrvChild *child, int64_t offset, void *buf, int64_t bytes) { int ret; QEMUIOVector qiov = QEMU_IOVEC_INIT_BUF(qiov, buf, bytes); + IO_CODE(); if (bytes < 0) { return -EINVAL; @@ -XXX,XX +XXX,XX @@ int bdrv_pwrite(BdrvChild *child, int64_t offset, const void *buf, { int ret; QEMUIOVector qiov = QEMU_IOVEC_INIT_BUF(qiov, buf, bytes); + IO_CODE(); if (bytes < 0) { return -EINVAL; @@ -XXX,XX +XXX,XX @@ int bdrv_pwrite_sync(BdrvChild *child, int64_t offset, const void *buf, int64_t count) { int ret; + IO_CODE(); ret = bdrv_pwrite(child, offset, buf, count); if (ret < 0) { @@ -XXX,XX +XXX,XX @@ out: int coroutine_fn bdrv_co_pwrite_zeroes(BdrvChild *child, int64_t offset, int64_t bytes, BdrvRequestFlags flags) { + IO_CODE(); trace_bdrv_co_pwrite_zeroes(child->bs, offset, bytes, flags); if (!(child->bs->open_flags & BDRV_O_UNMAP)) { @@ -XXX,XX +XXX,XX @@ int bdrv_block_status_above(BlockDriverState *bs, BlockDriverState *base, int64_t offset, int64_t bytes, int64_t *pnum, int64_t *map, BlockDriverState **file) { + IO_CODE(); return bdrv_common_block_status_above(bs, base, false, true, offset, bytes, pnum, map, file, NULL); } @@ -XXX,XX +XXX,XX @@ int bdrv_block_status_above(BlockDriverState *bs, BlockDriverState *base, int bdrv_block_status(BlockDriverState *bs, int64_t offset, int64_t bytes, int64_t *pnum, int64_t *map, BlockDriverState **file) { + IO_CODE(); return bdrv_block_status_above(bs, bdrv_filter_or_cow_bs(bs), offset, bytes, pnum, map, file); } @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_is_zero_fast(BlockDriverState *bs, int64_t offset, { int ret; int64_t pnum = bytes; + IO_CODE(); if (!bytes) { return 1; @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_is_allocated(BlockDriverState *bs, int64_t offset, { int ret; int64_t dummy; + IO_CODE(); ret = bdrv_common_block_status_above(bs, bs, true, false, offset, bytes, pnum ? pnum : &dummy, NULL, @@ -XXX,XX +XXX,XX @@ int bdrv_is_allocated_above(BlockDriverState *top, int ret = bdrv_common_block_status_above(top, base, include_base, false, offset, bytes, pnum, NULL, NULL, &depth); + IO_CODE(); if (ret < 0) { return ret; } @@ -XXX,XX +XXX,XX @@ int bdrv_save_vmstate(BlockDriverState *bs, const uint8_t *buf, { QEMUIOVector qiov = QEMU_IOVEC_INIT_BUF(qiov, buf, size); int ret = bdrv_writev_vmstate(bs, &qiov, pos); + IO_CODE(); return ret < 0 ? ret : size; } @@ -XXX,XX +XXX,XX @@ int bdrv_load_vmstate(BlockDriverState *bs, uint8_t *buf, { QEMUIOVector qiov = QEMU_IOVEC_INIT_BUF(qiov, buf, size); int ret = bdrv_readv_vmstate(bs, &qiov, pos); + IO_CODE(); return ret < 0 ? ret : size; } @@ -XXX,XX +XXX,XX @@ int bdrv_load_vmstate(BlockDriverState *bs, uint8_t *buf, void bdrv_aio_cancel(BlockAIOCB *acb) { + IO_CODE(); qemu_aio_ref(acb); bdrv_aio_cancel_async(acb); while (acb->refcnt > 1) { @@ -XXX,XX +XXX,XX @@ void bdrv_aio_cancel(BlockAIOCB *acb) * In either case the completion callback must be called. */ void bdrv_aio_cancel_async(BlockAIOCB *acb) { + IO_CODE(); if (acb->aiocb_info->cancel_async) { acb->aiocb_info->cancel_async(acb); } @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_flush(BlockDriverState *bs) BdrvChild *child; int current_gen; int ret = 0; + IO_CODE(); bdrv_inc_in_flight(bs); @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_pdiscard(BdrvChild *child, int64_t offset, int64_t max_pdiscard; int head, tail, align; BlockDriverState *bs = child->bs; + IO_CODE(); if (!bs || !bs->drv || !bdrv_is_inserted(bs)) { return -ENOMEDIUM; @@ -XXX,XX +XXX,XX @@ int bdrv_co_ioctl(BlockDriverState *bs, int req, void *buf) .coroutine = qemu_coroutine_self(), }; BlockAIOCB *acb; + IO_CODE(); bdrv_inc_in_flight(bs); if (!drv || (!drv->bdrv_aio_ioctl && !drv->bdrv_co_ioctl)) { @@ -XXX,XX +XXX,XX @@ out: void *qemu_blockalign(BlockDriverState *bs, size_t size) { + IO_CODE(); return qemu_memalign(bdrv_opt_mem_align(bs), size); } void *qemu_blockalign0(BlockDriverState *bs, size_t size) { + IO_CODE(); return memset(qemu_blockalign(bs, size), 0, size); } void *qemu_try_blockalign(BlockDriverState *bs, size_t size) { size_t align = bdrv_opt_mem_align(bs); + IO_CODE(); /* Ensure that NULL is never returned on success */ assert(align > 0); @@ -XXX,XX +XXX,XX @@ void *qemu_try_blockalign(BlockDriverState *bs, size_t size) void *qemu_try_blockalign0(BlockDriverState *bs, size_t size) { void *mem = qemu_try_blockalign(bs, size); + IO_CODE(); if (mem) { memset(mem, 0, size); @@ -XXX,XX +XXX,XX @@ bool bdrv_qiov_is_aligned(BlockDriverState *bs, QEMUIOVector *qiov) { int i; size_t alignment = bdrv_min_mem_align(bs); + IO_CODE(); for (i = 0; i < qiov->niov; i++) { if ((uintptr_t) qiov->iov[i].iov_base % alignment) { @@ -XXX,XX +XXX,XX @@ bool bdrv_qiov_is_aligned(BlockDriverState *bs, QEMUIOVector *qiov) void bdrv_io_plug(BlockDriverState *bs) { BdrvChild *child; + IO_CODE(); QLIST_FOREACH(child, &bs->children, next) { bdrv_io_plug(child->bs); @@ -XXX,XX +XXX,XX @@ void bdrv_io_plug(BlockDriverState *bs) void bdrv_io_unplug(BlockDriverState *bs) { BdrvChild *child; + IO_CODE(); assert(bs->io_plugged); if (qatomic_fetch_dec(&bs->io_plugged) == 1) { @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_copy_range(BdrvChild *src, int64_t src_offset, int64_t bytes, BdrvRequestFlags read_flags, BdrvRequestFlags write_flags) { + IO_CODE(); return bdrv_co_copy_range_from(src, src_offset, dst, dst_offset, bytes, read_flags, write_flags); @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact, BdrvTrackedRequest req; int64_t old_size, new_bytes; int ret; - + IO_CODE(); /* if bs->drv == NULL, bs is closed, so there's nothing to do here */ if (!drv) { -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> Allow writable exports to get BLK_PERM_RESIZE permission from creation, in fuse_export_create(). In this way, there is no need to give the permission in fuse_do_truncate(), which might be run in an iothread. Permissions should be set only in the main thread, so in any case if an iothread tries to set RESIZE, it will be blocked. Also assert in fuse_do_truncate that if we give the RESIZE permission we can then restore the original ones. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Reviewed-by: Hanna Reitz <hreitz@redhat.com> Message-Id: <20220303151616.325444-7-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/export/fuse.c | 25 ++++++++++++++++++------- 1 file changed, 18 insertions(+), 7 deletions(-) diff --git a/block/export/fuse.c b/block/export/fuse.c index XXXXXXX..XXXXXXX 100644 --- a/block/export/fuse.c +++ b/block/export/fuse.c @@ -XXX,XX +XXX,XX @@ static int fuse_export_create(BlockExport *blk_exp, assert(blk_exp_args->type == BLOCK_EXPORT_TYPE_FUSE); - /* For growable exports, take the RESIZE permission */ - if (args->growable) { + /* For growable and writable exports, take the RESIZE permission */ + if (args->growable || blk_exp_args->writable) { uint64_t blk_perm, blk_shared_perm; blk_get_perm(exp->common.blk, &blk_perm, &blk_shared_perm); @@ -XXX,XX +XXX,XX @@ static int fuse_do_truncate(const FuseExport *exp, int64_t size, { uint64_t blk_perm, blk_shared_perm; BdrvRequestFlags truncate_flags = 0; - int ret; + bool add_resize_perm; + int ret, ret_check; + + /* Growable and writable exports have a permanent RESIZE permission */ + add_resize_perm = !exp->growable && !exp->writable; if (req_zero_write) { truncate_flags |= BDRV_REQ_ZERO_WRITE; } - /* Growable exports have a permanent RESIZE permission */ - if (!exp->growable) { + if (add_resize_perm) { + + if (!qemu_in_main_thread()) { + /* Changing permissions like below only works in the main thread */ + return -EPERM; + } + blk_get_perm(exp->common.blk, &blk_perm, &blk_shared_perm); ret = blk_set_perm(exp->common.blk, blk_perm | BLK_PERM_RESIZE, @@ -XXX,XX +XXX,XX @@ static int fuse_do_truncate(const FuseExport *exp, int64_t size, ret = blk_truncate(exp->common.blk, size, true, prealloc, truncate_flags, NULL); - if (!exp->growable) { + if (add_resize_perm) { /* Must succeed, because we are only giving up the RESIZE permission */ - blk_set_perm(exp->common.blk, blk_perm, blk_shared_perm, &error_abort); + ret_check = blk_set_perm(exp->common.blk, blk_perm, + blk_shared_perm, &error_abort); + assert(ret_check == 0); } return ret; -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> Similarly to the previous patches, split block-backend.h in block-backend-io.h and block-backend-global-state.h In addition, remove "block/block.h" include as it seems it is not necessary anymore, together with "qemu/iov.h" block-backend-common.h contains the structures shared between the two headers, and the functions that can't be categorized as I/O or global state. Assertions are added in the next patch. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Message-Id: <20220303151616.325444-8-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/sysemu/block-backend-common.h | 84 ++++++ include/sysemu/block-backend-global-state.h | 116 +++++++++ include/sysemu/block-backend-io.h | 159 ++++++++++++ include/sysemu/block-backend.h | 269 +------------------- block/block-backend.c | 9 +- 5 files changed, 368 insertions(+), 269 deletions(-) create mode 100644 include/sysemu/block-backend-common.h create mode 100644 include/sysemu/block-backend-global-state.h create mode 100644 include/sysemu/block-backend-io.h diff --git a/include/sysemu/block-backend-common.h b/include/sysemu/block-backend-common.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/sysemu/block-backend-common.h @@ -XXX,XX +XXX,XX @@ +/* + * QEMU Block backends + * + * Copyright (C) 2014-2016 Red Hat, Inc. + * + * Authors: + * Markus Armbruster <armbru@redhat.com>, + * + * This work is licensed under the terms of the GNU LGPL, version 2.1 + * or later. See the COPYING.LIB file in the top-level directory. + */ + +#ifndef BLOCK_BACKEND_COMMON_H +#define BLOCK_BACKEND_COMMON_H + +#include "qemu/iov.h" +#include "block/throttle-groups.h" + +/* + * TODO Have to include block/block.h for a bunch of block layer + * types. Unfortunately, this pulls in the whole BlockDriverState + * API, which we don't want used by many BlockBackend users. Some of + * the types belong here, and the rest should be split into a common + * header and one for the BlockDriverState API. + */ +#include "block/block.h" + +/* Callbacks for block device models */ +typedef struct BlockDevOps { + /* + * Runs when virtual media changed (monitor commands eject, change) + * Argument load is true on load and false on eject. + * Beware: doesn't run when a host device's physical media + * changes. Sure would be useful if it did. + * Device models with removable media must implement this callback. + */ + void (*change_media_cb)(void *opaque, bool load, Error **errp); + /* + * Runs when an eject request is issued from the monitor, the tray + * is closed, and the medium is locked. + * Device models that do not implement is_medium_locked will not need + * this callback. Device models that can lock the medium or tray might + * want to implement the callback and unlock the tray when "force" is + * true, even if they do not support eject requests. + */ + void (*eject_request_cb)(void *opaque, bool force); + /* + * Is the virtual tray open? + * Device models implement this only when the device has a tray. + */ + bool (*is_tray_open)(void *opaque); + /* + * Is the virtual medium locked into the device? + * Device models implement this only when device has such a lock. + */ + bool (*is_medium_locked)(void *opaque); + /* + * Runs when the size changed (e.g. monitor command block_resize) + */ + void (*resize_cb)(void *opaque); + /* + * Runs when the backend receives a drain request. + */ + void (*drained_begin)(void *opaque); + /* + * Runs when the backend's last drain request ends. + */ + void (*drained_end)(void *opaque); + /* + * Is the device still busy? + */ + bool (*drained_poll)(void *opaque); +} BlockDevOps; + +/* + * This struct is embedded in (the private) BlockBackend struct and contains + * fields that must be public. This is in particular for QLIST_ENTRY() and + * friends so that BlockBackends can be kept in lists outside block-backend.c + */ +typedef struct BlockBackendPublic { + ThrottleGroupMember throttle_group_member; +} BlockBackendPublic; + +#endif /* BLOCK_BACKEND_COMMON_H */ diff --git a/include/sysemu/block-backend-global-state.h b/include/sysemu/block-backend-global-state.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/sysemu/block-backend-global-state.h @@ -XXX,XX +XXX,XX @@ +/* + * QEMU Block backends + * + * Copyright (C) 2014-2016 Red Hat, Inc. + * + * Authors: + * Markus Armbruster <armbru@redhat.com>, + * + * This work is licensed under the terms of the GNU LGPL, version 2.1 + * or later. See the COPYING.LIB file in the top-level directory. + */ + +#ifndef BLOCK_BACKEND_GS_H +#define BLOCK_BACKEND_GS_H + +#include "block-backend-common.h" + +/* + * Global state (GS) API. These functions run under the BQL. + * + * See include/block/block-global-state.h for more information about + * the GS API. + */ + +BlockBackend *blk_new(AioContext *ctx, uint64_t perm, uint64_t shared_perm); +BlockBackend *blk_new_with_bs(BlockDriverState *bs, uint64_t perm, + uint64_t shared_perm, Error **errp); +BlockBackend *blk_new_open(const char *filename, const char *reference, + QDict *options, int flags, Error **errp); +int blk_get_refcnt(BlockBackend *blk); +void blk_ref(BlockBackend *blk); +void blk_unref(BlockBackend *blk); +void blk_remove_all_bs(void); +BlockBackend *blk_by_name(const char *name); +BlockBackend *blk_next(BlockBackend *blk); +BlockBackend *blk_all_next(BlockBackend *blk); +bool monitor_add_blk(BlockBackend *blk, const char *name, Error **errp); +void monitor_remove_blk(BlockBackend *blk); + +BlockBackendPublic *blk_get_public(BlockBackend *blk); +BlockBackend *blk_by_public(BlockBackendPublic *public); + +void blk_remove_bs(BlockBackend *blk); +int blk_insert_bs(BlockBackend *blk, BlockDriverState *bs, Error **errp); +int blk_replace_bs(BlockBackend *blk, BlockDriverState *new_bs, Error **errp); +bool bdrv_has_blk(BlockDriverState *bs); +bool bdrv_is_root_node(BlockDriverState *bs); +int blk_set_perm(BlockBackend *blk, uint64_t perm, uint64_t shared_perm, + Error **errp); +void blk_get_perm(BlockBackend *blk, uint64_t *perm, uint64_t *shared_perm); + +void blk_iostatus_enable(BlockBackend *blk); +BlockDeviceIoStatus blk_iostatus(const BlockBackend *blk); +void blk_iostatus_disable(BlockBackend *blk); +void blk_iostatus_reset(BlockBackend *blk); +int blk_attach_dev(BlockBackend *blk, DeviceState *dev); +void blk_detach_dev(BlockBackend *blk, DeviceState *dev); +DeviceState *blk_get_attached_dev(BlockBackend *blk); +BlockBackend *blk_by_dev(void *dev); +BlockBackend *blk_by_qdev_id(const char *id, Error **errp); +void blk_set_dev_ops(BlockBackend *blk, const BlockDevOps *ops, void *opaque); + +void blk_activate(BlockBackend *blk, Error **errp); + +int blk_make_zero(BlockBackend *blk, BdrvRequestFlags flags); +void blk_aio_cancel(BlockAIOCB *acb); +int blk_commit_all(void); +void blk_drain(BlockBackend *blk); +void blk_drain_all(void); +void blk_set_on_error(BlockBackend *blk, BlockdevOnError on_read_error, + BlockdevOnError on_write_error); +bool blk_supports_write_perm(BlockBackend *blk); +bool blk_is_sg(BlockBackend *blk); +void blk_set_enable_write_cache(BlockBackend *blk, bool wce); +int blk_get_flags(BlockBackend *blk); +bool blk_op_is_blocked(BlockBackend *blk, BlockOpType op, Error **errp); +void blk_op_unblock(BlockBackend *blk, BlockOpType op, Error *reason); +void blk_op_block_all(BlockBackend *blk, Error *reason); +void blk_op_unblock_all(BlockBackend *blk, Error *reason); +int blk_set_aio_context(BlockBackend *blk, AioContext *new_context, + Error **errp); +void blk_add_aio_context_notifier(BlockBackend *blk, + void (*attached_aio_context)(AioContext *new_context, void *opaque), + void (*detach_aio_context)(void *opaque), void *opaque); +void blk_remove_aio_context_notifier(BlockBackend *blk, + void (*attached_aio_context)(AioContext *, + void *), + void (*detach_aio_context)(void *), + void *opaque); +void blk_add_remove_bs_notifier(BlockBackend *blk, Notifier *notify); +void blk_add_insert_bs_notifier(BlockBackend *blk, Notifier *notify); +BlockBackendRootState *blk_get_root_state(BlockBackend *blk); +void blk_update_root_state(BlockBackend *blk); +bool blk_get_detect_zeroes_from_root_state(BlockBackend *blk); +int blk_get_open_flags_from_root_state(BlockBackend *blk); + +int blk_save_vmstate(BlockBackend *blk, const uint8_t *buf, + int64_t pos, int size); +int blk_load_vmstate(BlockBackend *blk, uint8_t *buf, int64_t pos, int size); +int blk_probe_blocksizes(BlockBackend *blk, BlockSizes *bsz); +int blk_probe_geometry(BlockBackend *blk, HDGeometry *geo); + +void blk_set_io_limits(BlockBackend *blk, ThrottleConfig *cfg); +void blk_io_limits_disable(BlockBackend *blk); +void blk_io_limits_enable(BlockBackend *blk, const char *group); +void blk_io_limits_update_group(BlockBackend *blk, const char *group); +void blk_set_force_allow_inactivate(BlockBackend *blk); + +void blk_register_buf(BlockBackend *blk, void *host, size_t size); +void blk_unregister_buf(BlockBackend *blk, void *host); + +const BdrvChild *blk_root(BlockBackend *blk); + +int blk_make_empty(BlockBackend *blk, Error **errp); + +#endif /* BLOCK_BACKEND_GS_H */ diff --git a/include/sysemu/block-backend-io.h b/include/sysemu/block-backend-io.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/sysemu/block-backend-io.h @@ -XXX,XX +XXX,XX @@ +/* + * QEMU Block backends + * + * Copyright (C) 2014-2016 Red Hat, Inc. + * + * Authors: + * Markus Armbruster <armbru@redhat.com>, + * + * This work is licensed under the terms of the GNU LGPL, version 2.1 + * or later. See the COPYING.LIB file in the top-level directory. + */ + +#ifndef BLOCK_BACKEND_IO_H +#define BLOCK_BACKEND_IO_H + +#include "block-backend-common.h" + +/* + * I/O API functions. These functions are thread-safe. + * + * See include/block/block-io.h for more information about + * the I/O API. + */ + +const char *blk_name(const BlockBackend *blk); + +BlockDriverState *blk_bs(BlockBackend *blk); + +void blk_set_allow_write_beyond_eof(BlockBackend *blk, bool allow); +void blk_set_allow_aio_context_change(BlockBackend *blk, bool allow); +void blk_set_disable_request_queuing(BlockBackend *blk, bool disable); +bool blk_iostatus_is_enabled(const BlockBackend *blk); + +char *blk_get_attached_dev_id(BlockBackend *blk); + +BlockAIOCB *blk_aio_pwrite_zeroes(BlockBackend *blk, int64_t offset, + int64_t bytes, BdrvRequestFlags flags, + BlockCompletionFunc *cb, void *opaque); + +BlockAIOCB *blk_aio_preadv(BlockBackend *blk, int64_t offset, + QEMUIOVector *qiov, BdrvRequestFlags flags, + BlockCompletionFunc *cb, void *opaque); +BlockAIOCB *blk_aio_pwritev(BlockBackend *blk, int64_t offset, + QEMUIOVector *qiov, BdrvRequestFlags flags, + BlockCompletionFunc *cb, void *opaque); +BlockAIOCB *blk_aio_flush(BlockBackend *blk, + BlockCompletionFunc *cb, void *opaque); +BlockAIOCB *blk_aio_pdiscard(BlockBackend *blk, int64_t offset, int64_t bytes, + BlockCompletionFunc *cb, void *opaque); +void blk_aio_cancel_async(BlockAIOCB *acb); +BlockAIOCB *blk_aio_ioctl(BlockBackend *blk, unsigned long int req, void *buf, + BlockCompletionFunc *cb, void *opaque); + +void blk_inc_in_flight(BlockBackend *blk); +void blk_dec_in_flight(BlockBackend *blk); +bool blk_is_inserted(BlockBackend *blk); +bool blk_is_available(BlockBackend *blk); +void blk_lock_medium(BlockBackend *blk, bool locked); +void blk_eject(BlockBackend *blk, bool eject_flag); +int64_t blk_getlength(BlockBackend *blk); +void blk_get_geometry(BlockBackend *blk, uint64_t *nb_sectors_ptr); +int64_t blk_nb_sectors(BlockBackend *blk); +void *blk_try_blockalign(BlockBackend *blk, size_t size); +void *blk_blockalign(BlockBackend *blk, size_t size); +bool blk_is_writable(BlockBackend *blk); +bool blk_enable_write_cache(BlockBackend *blk); +BlockdevOnError blk_get_on_error(BlockBackend *blk, bool is_read); +BlockErrorAction blk_get_error_action(BlockBackend *blk, bool is_read, + int error); +void blk_error_action(BlockBackend *blk, BlockErrorAction action, + bool is_read, int error); +void blk_iostatus_set_err(BlockBackend *blk, int error); +int blk_get_max_iov(BlockBackend *blk); +int blk_get_max_hw_iov(BlockBackend *blk); +void blk_set_guest_block_size(BlockBackend *blk, int align); + +void blk_io_plug(BlockBackend *blk); +void blk_io_unplug(BlockBackend *blk); +AioContext *blk_get_aio_context(BlockBackend *blk); +BlockAcctStats *blk_get_stats(BlockBackend *blk); +void *blk_aio_get(const AIOCBInfo *aiocb_info, BlockBackend *blk, + BlockCompletionFunc *cb, void *opaque); +BlockAIOCB *blk_abort_aio_request(BlockBackend *blk, + BlockCompletionFunc *cb, + void *opaque, int ret); + +uint32_t blk_get_request_alignment(BlockBackend *blk); +uint32_t blk_get_max_transfer(BlockBackend *blk); +uint64_t blk_get_max_hw_transfer(BlockBackend *blk); + +int coroutine_fn blk_co_copy_range(BlockBackend *blk_in, int64_t off_in, + BlockBackend *blk_out, int64_t off_out, + int64_t bytes, BdrvRequestFlags read_flags, + BdrvRequestFlags write_flags); + + +/* + * "I/O or GS" API functions. These functions can run without + * the BQL, but only in one specific iothread/main loop. + * + * See include/block/block-io.h for more information about + * the "I/O or GS" API. + */ + +int blk_pread(BlockBackend *blk, int64_t offset, void *buf, int bytes); +int blk_pwrite(BlockBackend *blk, int64_t offset, const void *buf, int bytes, + BdrvRequestFlags flags); +int coroutine_fn blk_co_preadv(BlockBackend *blk, int64_t offset, + int64_t bytes, QEMUIOVector *qiov, + BdrvRequestFlags flags); +int coroutine_fn blk_co_pwritev_part(BlockBackend *blk, int64_t offset, + int64_t bytes, + QEMUIOVector *qiov, size_t qiov_offset, + BdrvRequestFlags flags); +int coroutine_fn blk_co_pwritev(BlockBackend *blk, int64_t offset, + int64_t bytes, QEMUIOVector *qiov, + BdrvRequestFlags flags); + +static inline int coroutine_fn blk_co_pread(BlockBackend *blk, int64_t offset, + int64_t bytes, void *buf, + BdrvRequestFlags flags) +{ + QEMUIOVector qiov = QEMU_IOVEC_INIT_BUF(qiov, buf, bytes); + + assert(bytes <= SIZE_MAX); + + return blk_co_preadv(blk, offset, bytes, &qiov, flags); +} + +static inline int coroutine_fn blk_co_pwrite(BlockBackend *blk, int64_t offset, + int64_t bytes, void *buf, + BdrvRequestFlags flags) +{ + QEMUIOVector qiov = QEMU_IOVEC_INIT_BUF(qiov, buf, bytes); + + assert(bytes <= SIZE_MAX); + + return blk_co_pwritev(blk, offset, bytes, &qiov, flags); +} + +int coroutine_fn blk_co_pdiscard(BlockBackend *blk, int64_t offset, + int64_t bytes); + +int coroutine_fn blk_co_flush(BlockBackend *blk); +int blk_flush(BlockBackend *blk); + +int blk_ioctl(BlockBackend *blk, unsigned long int req, void *buf); + +int blk_pwrite_compressed(BlockBackend *blk, int64_t offset, const void *buf, + int64_t bytes); +int blk_pdiscard(BlockBackend *blk, int64_t offset, int64_t bytes); +int blk_pwrite_zeroes(BlockBackend *blk, int64_t offset, + int64_t bytes, BdrvRequestFlags flags); +int coroutine_fn blk_co_pwrite_zeroes(BlockBackend *blk, int64_t offset, + int64_t bytes, BdrvRequestFlags flags); +int blk_truncate(BlockBackend *blk, int64_t offset, bool exact, + PreallocMode prealloc, BdrvRequestFlags flags, Error **errp); + +#endif /* BLOCK_BACKEND_IO_H */ 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 @@ #ifndef BLOCK_BACKEND_H #define BLOCK_BACKEND_H -#include "qemu/iov.h" -#include "block/throttle-groups.h" +#include "block-backend-global-state.h" +#include "block-backend-io.h" -/* - * TODO Have to include block/block.h for a bunch of block layer - * types. Unfortunately, this pulls in the whole BlockDriverState - * API, which we don't want used by many BlockBackend users. Some of - * the types belong here, and the rest should be split into a common - * header and one for the BlockDriverState API. - */ -#include "block/block.h" - -/* Callbacks for block device models */ -typedef struct BlockDevOps { - /* - * Runs when virtual media changed (monitor commands eject, change) - * Argument load is true on load and false on eject. - * Beware: doesn't run when a host device's physical media - * changes. Sure would be useful if it did. - * Device models with removable media must implement this callback. - */ - void (*change_media_cb)(void *opaque, bool load, Error **errp); - /* - * Runs when an eject request is issued from the monitor, the tray - * is closed, and the medium is locked. - * Device models that do not implement is_medium_locked will not need - * this callback. Device models that can lock the medium or tray might - * want to implement the callback and unlock the tray when "force" is - * true, even if they do not support eject requests. - */ - void (*eject_request_cb)(void *opaque, bool force); - /* - * Is the virtual tray open? - * Device models implement this only when the device has a tray. - */ - bool (*is_tray_open)(void *opaque); - /* - * Is the virtual medium locked into the device? - * Device models implement this only when device has such a lock. - */ - bool (*is_medium_locked)(void *opaque); - /* - * Runs when the size changed (e.g. monitor command block_resize) - */ - void (*resize_cb)(void *opaque); - /* - * Runs when the backend receives a drain request. - */ - void (*drained_begin)(void *opaque); - /* - * Runs when the backend's last drain request ends. - */ - void (*drained_end)(void *opaque); - /* - * Is the device still busy? - */ - bool (*drained_poll)(void *opaque); -} BlockDevOps; - -/* This struct is embedded in (the private) BlockBackend struct and contains - * fields that must be public. This is in particular for QLIST_ENTRY() and - * friends so that BlockBackends can be kept in lists outside block-backend.c - * */ -typedef struct BlockBackendPublic { - ThrottleGroupMember throttle_group_member; -} BlockBackendPublic; - -BlockBackend *blk_new(AioContext *ctx, uint64_t perm, uint64_t shared_perm); -BlockBackend *blk_new_with_bs(BlockDriverState *bs, uint64_t perm, - uint64_t shared_perm, Error **errp); -BlockBackend *blk_new_open(const char *filename, const char *reference, - QDict *options, int flags, Error **errp); -int blk_get_refcnt(BlockBackend *blk); -void blk_ref(BlockBackend *blk); -void blk_unref(BlockBackend *blk); -void blk_remove_all_bs(void); -const char *blk_name(const BlockBackend *blk); -BlockBackend *blk_by_name(const char *name); -BlockBackend *blk_next(BlockBackend *blk); -BlockBackend *blk_all_next(BlockBackend *blk); -bool monitor_add_blk(BlockBackend *blk, const char *name, Error **errp); -void monitor_remove_blk(BlockBackend *blk); - -BlockBackendPublic *blk_get_public(BlockBackend *blk); -BlockBackend *blk_by_public(BlockBackendPublic *public); - -BlockDriverState *blk_bs(BlockBackend *blk); -void blk_remove_bs(BlockBackend *blk); -int blk_insert_bs(BlockBackend *blk, BlockDriverState *bs, Error **errp); -int blk_replace_bs(BlockBackend *blk, BlockDriverState *new_bs, Error **errp); -bool bdrv_has_blk(BlockDriverState *bs); -bool bdrv_is_root_node(BlockDriverState *bs); -int blk_set_perm(BlockBackend *blk, uint64_t perm, uint64_t shared_perm, - Error **errp); -void blk_get_perm(BlockBackend *blk, uint64_t *perm, uint64_t *shared_perm); - -void blk_set_allow_write_beyond_eof(BlockBackend *blk, bool allow); -void blk_set_allow_aio_context_change(BlockBackend *blk, bool allow); -void blk_set_disable_request_queuing(BlockBackend *blk, bool disable); -void blk_iostatus_enable(BlockBackend *blk); -bool blk_iostatus_is_enabled(const BlockBackend *blk); -BlockDeviceIoStatus blk_iostatus(const BlockBackend *blk); -void blk_iostatus_disable(BlockBackend *blk); -void blk_iostatus_reset(BlockBackend *blk); -void blk_iostatus_set_err(BlockBackend *blk, int error); -int blk_attach_dev(BlockBackend *blk, DeviceState *dev); -void blk_detach_dev(BlockBackend *blk, DeviceState *dev); -DeviceState *blk_get_attached_dev(BlockBackend *blk); -char *blk_get_attached_dev_id(BlockBackend *blk); -BlockBackend *blk_by_dev(void *dev); -BlockBackend *blk_by_qdev_id(const char *id, Error **errp); -void blk_set_dev_ops(BlockBackend *blk, const BlockDevOps *ops, void *opaque); -int coroutine_fn blk_co_preadv(BlockBackend *blk, int64_t offset, - int64_t bytes, QEMUIOVector *qiov, - BdrvRequestFlags flags); -int coroutine_fn blk_co_pwritev_part(BlockBackend *blk, int64_t offset, - int64_t bytes, - QEMUIOVector *qiov, size_t qiov_offset, - BdrvRequestFlags flags); -int coroutine_fn blk_co_pwritev(BlockBackend *blk, int64_t offset, - int64_t bytes, QEMUIOVector *qiov, - BdrvRequestFlags flags); - -static inline int coroutine_fn blk_co_pread(BlockBackend *blk, int64_t offset, - int64_t bytes, void *buf, - BdrvRequestFlags flags) -{ - QEMUIOVector qiov = QEMU_IOVEC_INIT_BUF(qiov, buf, bytes); - - assert(bytes <= SIZE_MAX); - - return blk_co_preadv(blk, offset, bytes, &qiov, flags); -} - -static inline int coroutine_fn blk_co_pwrite(BlockBackend *blk, int64_t offset, - int64_t bytes, void *buf, - BdrvRequestFlags flags) -{ - QEMUIOVector qiov = QEMU_IOVEC_INIT_BUF(qiov, buf, bytes); - - assert(bytes <= SIZE_MAX); - - return blk_co_pwritev(blk, offset, bytes, &qiov, flags); -} - -int blk_pwrite_zeroes(BlockBackend *blk, int64_t offset, - int64_t bytes, BdrvRequestFlags flags); -BlockAIOCB *blk_aio_pwrite_zeroes(BlockBackend *blk, int64_t offset, - int64_t bytes, BdrvRequestFlags flags, - BlockCompletionFunc *cb, void *opaque); -int blk_make_zero(BlockBackend *blk, BdrvRequestFlags flags); -int blk_pread(BlockBackend *blk, int64_t offset, void *buf, int bytes); -int blk_pwrite(BlockBackend *blk, int64_t offset, const void *buf, int bytes, - BdrvRequestFlags flags); -int64_t blk_getlength(BlockBackend *blk); -void blk_get_geometry(BlockBackend *blk, uint64_t *nb_sectors_ptr); -int64_t blk_nb_sectors(BlockBackend *blk); -BlockAIOCB *blk_aio_preadv(BlockBackend *blk, int64_t offset, - QEMUIOVector *qiov, BdrvRequestFlags flags, - BlockCompletionFunc *cb, void *opaque); -BlockAIOCB *blk_aio_pwritev(BlockBackend *blk, int64_t offset, - QEMUIOVector *qiov, BdrvRequestFlags flags, - BlockCompletionFunc *cb, void *opaque); -BlockAIOCB *blk_aio_flush(BlockBackend *blk, - BlockCompletionFunc *cb, void *opaque); -BlockAIOCB *blk_aio_pdiscard(BlockBackend *blk, int64_t offset, int64_t bytes, - BlockCompletionFunc *cb, void *opaque); -void blk_aio_cancel(BlockAIOCB *acb); -void blk_aio_cancel_async(BlockAIOCB *acb); -int blk_ioctl(BlockBackend *blk, unsigned long int req, void *buf); -BlockAIOCB *blk_aio_ioctl(BlockBackend *blk, unsigned long int req, void *buf, - BlockCompletionFunc *cb, void *opaque); -int coroutine_fn blk_co_pdiscard(BlockBackend *blk, int64_t offset, - int64_t bytes); -int coroutine_fn blk_co_flush(BlockBackend *blk); -int blk_flush(BlockBackend *blk); -int blk_commit_all(void); -void blk_inc_in_flight(BlockBackend *blk); -void blk_dec_in_flight(BlockBackend *blk); -void blk_drain(BlockBackend *blk); -void blk_drain_all(void); -void blk_set_on_error(BlockBackend *blk, BlockdevOnError on_read_error, - BlockdevOnError on_write_error); -BlockdevOnError blk_get_on_error(BlockBackend *blk, bool is_read); -BlockErrorAction blk_get_error_action(BlockBackend *blk, bool is_read, - int error); -void blk_error_action(BlockBackend *blk, BlockErrorAction action, - bool is_read, int error); -bool blk_supports_write_perm(BlockBackend *blk); -bool blk_is_writable(BlockBackend *blk); -bool blk_is_sg(BlockBackend *blk); -bool blk_enable_write_cache(BlockBackend *blk); -void blk_set_enable_write_cache(BlockBackend *blk, bool wce); -void blk_activate(BlockBackend *blk, Error **errp); -bool blk_is_inserted(BlockBackend *blk); -bool blk_is_available(BlockBackend *blk); -void blk_lock_medium(BlockBackend *blk, bool locked); -void blk_eject(BlockBackend *blk, bool eject_flag); -int blk_get_flags(BlockBackend *blk); -uint32_t blk_get_request_alignment(BlockBackend *blk); -uint32_t blk_get_max_transfer(BlockBackend *blk); -uint64_t blk_get_max_hw_transfer(BlockBackend *blk); -int blk_get_max_iov(BlockBackend *blk); -int blk_get_max_hw_iov(BlockBackend *blk); -void blk_set_guest_block_size(BlockBackend *blk, int align); -void *blk_try_blockalign(BlockBackend *blk, size_t size); -void *blk_blockalign(BlockBackend *blk, size_t size); -bool blk_op_is_blocked(BlockBackend *blk, BlockOpType op, Error **errp); -void blk_op_unblock(BlockBackend *blk, BlockOpType op, Error *reason); -void blk_op_block_all(BlockBackend *blk, Error *reason); -void blk_op_unblock_all(BlockBackend *blk, Error *reason); -AioContext *blk_get_aio_context(BlockBackend *blk); -int blk_set_aio_context(BlockBackend *blk, AioContext *new_context, - Error **errp); -void blk_add_aio_context_notifier(BlockBackend *blk, - void (*attached_aio_context)(AioContext *new_context, void *opaque), - void (*detach_aio_context)(void *opaque), void *opaque); -void blk_remove_aio_context_notifier(BlockBackend *blk, - void (*attached_aio_context)(AioContext *, - void *), - void (*detach_aio_context)(void *), - void *opaque); -void blk_add_remove_bs_notifier(BlockBackend *blk, Notifier *notify); -void blk_add_insert_bs_notifier(BlockBackend *blk, Notifier *notify); -void blk_io_plug(BlockBackend *blk); -void blk_io_unplug(BlockBackend *blk); -BlockAcctStats *blk_get_stats(BlockBackend *blk); -BlockBackendRootState *blk_get_root_state(BlockBackend *blk); -void blk_update_root_state(BlockBackend *blk); -bool blk_get_detect_zeroes_from_root_state(BlockBackend *blk); -int blk_get_open_flags_from_root_state(BlockBackend *blk); - -void *blk_aio_get(const AIOCBInfo *aiocb_info, BlockBackend *blk, - BlockCompletionFunc *cb, void *opaque); -int coroutine_fn blk_co_pwrite_zeroes(BlockBackend *blk, int64_t offset, - int64_t bytes, BdrvRequestFlags flags); -int blk_pwrite_compressed(BlockBackend *blk, int64_t offset, const void *buf, - int64_t bytes); -int blk_truncate(BlockBackend *blk, int64_t offset, bool exact, - PreallocMode prealloc, BdrvRequestFlags flags, Error **errp); -int blk_pdiscard(BlockBackend *blk, int64_t offset, int64_t bytes); -int blk_save_vmstate(BlockBackend *blk, const uint8_t *buf, - int64_t pos, int size); -int blk_load_vmstate(BlockBackend *blk, uint8_t *buf, int64_t pos, int size); -int blk_probe_blocksizes(BlockBackend *blk, BlockSizes *bsz); -int blk_probe_geometry(BlockBackend *blk, HDGeometry *geo); -BlockAIOCB *blk_abort_aio_request(BlockBackend *blk, - BlockCompletionFunc *cb, - void *opaque, int ret); - -void blk_set_io_limits(BlockBackend *blk, ThrottleConfig *cfg); -void blk_io_limits_disable(BlockBackend *blk); -void blk_io_limits_enable(BlockBackend *blk, const char *group); -void blk_io_limits_update_group(BlockBackend *blk, const char *group); -void blk_set_force_allow_inactivate(BlockBackend *blk); - -void blk_register_buf(BlockBackend *blk, void *host, size_t size); -void blk_unregister_buf(BlockBackend *blk, void *host); - -int coroutine_fn blk_co_copy_range(BlockBackend *blk_in, int64_t off_in, - BlockBackend *blk_out, int64_t off_out, - int64_t bytes, BdrvRequestFlags read_flags, - BdrvRequestFlags write_flags); - -const BdrvChild *blk_root(BlockBackend *blk); - -int blk_make_empty(BlockBackend *blk, Error **errp); +/* DO NOT ADD ANYTHING IN HERE. USE ONE OF THE HEADERS INCLUDED ABOVE */ #endif 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 @@ struct BlockBackend { bool allow_aio_context_change; bool allow_write_beyond_eof; + /* Protected by BQL */ NotifierList remove_bs_notifiers, insert_bs_notifiers; QLIST_HEAD(, BlockBackendAioNotifier) aio_notifiers; @@ -XXX,XX +XXX,XX @@ static const AIOCBInfo block_backend_aiocb_info = { static void drive_info_del(DriveInfo *dinfo); static BlockBackend *bdrv_first_blk(BlockDriverState *bs); -/* All BlockBackends */ +/* All BlockBackends. Protected by BQL. */ static QTAILQ_HEAD(, BlockBackend) block_backends = QTAILQ_HEAD_INITIALIZER(block_backends); -/* All BlockBackends referenced by the monitor and which are iterated through by - * blk_next() */ +/* + * All BlockBackends referenced by the monitor and which are iterated through by + * blk_next(). Protected by BQL. + */ static QTAILQ_HEAD(, BlockBackend) monitor_block_backends = QTAILQ_HEAD_INITIALIZER(monitor_block_backends); -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> All the global state (GS) API functions will check that qemu_in_main_thread() returns true. If not, it means that the safety of BQL cannot be guaranteed, and they need to be moved to I/O. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Message-Id: <20220303151616.325444-9-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/block-backend.c | 78 ++++++++++++++++++++++++++++++++++++++++++ softmmu/qdev-monitor.c | 2 ++ 2 files changed, 80 insertions(+) diff --git a/block/block-backend.c b/block/block-backend.c index XXXXXXX..XXXXXXX 100644 --- a/block/block-backend.c +++ b/block/block-backend.c @@ -XXX,XX +XXX,XX @@ static void blk_root_activate(BdrvChild *child, Error **errp) void blk_set_force_allow_inactivate(BlockBackend *blk) { + GLOBAL_STATE_CODE(); blk->force_allow_inactivate = true; } @@ -XXX,XX +XXX,XX @@ BlockBackend *blk_new(AioContext *ctx, uint64_t perm, uint64_t shared_perm) { BlockBackend *blk; + GLOBAL_STATE_CODE(); + blk = g_new0(BlockBackend, 1); blk->refcnt = 1; blk->ctx = ctx; @@ -XXX,XX +XXX,XX @@ BlockBackend *blk_new_with_bs(BlockDriverState *bs, uint64_t perm, { BlockBackend *blk = blk_new(bdrv_get_aio_context(bs), perm, shared_perm); + GLOBAL_STATE_CODE(); + if (blk_insert_bs(blk, bs, errp) < 0) { blk_unref(blk); return NULL; @@ -XXX,XX +XXX,XX @@ BlockBackend *blk_new_open(const char *filename, const char *reference, uint64_t perm = 0; uint64_t shared = BLK_PERM_ALL; + GLOBAL_STATE_CODE(); + /* * blk_new_open() is mainly used in .bdrv_create implementations and the * tools where sharing isn't a major concern because the BDS stays private @@ -XXX,XX +XXX,XX @@ static void drive_info_del(DriveInfo *dinfo) int blk_get_refcnt(BlockBackend *blk) { + GLOBAL_STATE_CODE(); return blk ? blk->refcnt : 0; } @@ -XXX,XX +XXX,XX @@ int blk_get_refcnt(BlockBackend *blk) void blk_ref(BlockBackend *blk) { assert(blk->refcnt > 0); + GLOBAL_STATE_CODE(); blk->refcnt++; } @@ -XXX,XX +XXX,XX @@ void blk_ref(BlockBackend *blk) */ void blk_unref(BlockBackend *blk) { + GLOBAL_STATE_CODE(); if (blk) { assert(blk->refcnt > 0); if (blk->refcnt > 1) { @@ -XXX,XX +XXX,XX @@ void blk_unref(BlockBackend *blk) */ BlockBackend *blk_all_next(BlockBackend *blk) { + GLOBAL_STATE_CODE(); return blk ? QTAILQ_NEXT(blk, link) : QTAILQ_FIRST(&block_backends); } @@ -XXX,XX +XXX,XX @@ void blk_remove_all_bs(void) { BlockBackend *blk = NULL; + GLOBAL_STATE_CODE(); + while ((blk = blk_all_next(blk)) != NULL) { AioContext *ctx = blk_get_aio_context(blk); @@ -XXX,XX +XXX,XX @@ void blk_remove_all_bs(void) */ BlockBackend *blk_next(BlockBackend *blk) { + GLOBAL_STATE_CODE(); return blk ? QTAILQ_NEXT(blk, monitor_link) : QTAILQ_FIRST(&monitor_block_backends); } @@ -XXX,XX +XXX,XX @@ static void bdrv_next_reset(BdrvNextIterator *it) BlockDriverState *bdrv_first(BdrvNextIterator *it) { + GLOBAL_STATE_CODE(); bdrv_next_reset(it); return bdrv_next(it); } @@ -XXX,XX +XXX,XX @@ bool monitor_add_blk(BlockBackend *blk, const char *name, Error **errp) { assert(!blk->name); assert(name && name[0]); + GLOBAL_STATE_CODE(); if (!id_wellformed(name)) { error_setg(errp, "Invalid device name"); @@ -XXX,XX +XXX,XX @@ bool monitor_add_blk(BlockBackend *blk, const char *name, Error **errp) */ void monitor_remove_blk(BlockBackend *blk) { + GLOBAL_STATE_CODE(); + if (!blk->name) { return; } @@ -XXX,XX +XXX,XX @@ BlockBackend *blk_by_name(const char *name) { BlockBackend *blk = NULL; + GLOBAL_STATE_CODE(); assert(name); while ((blk = blk_next(blk)) != NULL) { if (!strcmp(name, blk->name)) { @@ -XXX,XX +XXX,XX @@ static BlockBackend *bdrv_first_blk(BlockDriverState *bs) */ bool bdrv_has_blk(BlockDriverState *bs) { + GLOBAL_STATE_CODE(); return bdrv_first_blk(bs) != NULL; } @@ -XXX,XX +XXX,XX @@ bool bdrv_is_root_node(BlockDriverState *bs) { BdrvChild *c; + GLOBAL_STATE_CODE(); QLIST_FOREACH(c, &bs->parents, next_parent) { if (c->klass != &child_root) { return false; @@ -XXX,XX +XXX,XX @@ BlockBackend *blk_by_legacy_dinfo(DriveInfo *dinfo) */ BlockBackendPublic *blk_get_public(BlockBackend *blk) { + GLOBAL_STATE_CODE(); return &blk->public; } @@ -XXX,XX +XXX,XX @@ BlockBackendPublic *blk_get_public(BlockBackend *blk) */ BlockBackend *blk_by_public(BlockBackendPublic *public) { + GLOBAL_STATE_CODE(); return container_of(public, BlockBackend, public); } @@ -XXX,XX +XXX,XX @@ void blk_remove_bs(BlockBackend *blk) ThrottleGroupMember *tgm = &blk->public.throttle_group_member; BdrvChild *root; + GLOBAL_STATE_CODE(); + notifier_list_notify(&blk->remove_bs_notifiers, blk); if (tgm->throttle_state) { BlockDriverState *bs = blk_bs(blk); @@ -XXX,XX +XXX,XX @@ void blk_remove_bs(BlockBackend *blk) int blk_insert_bs(BlockBackend *blk, BlockDriverState *bs, Error **errp) { ThrottleGroupMember *tgm = &blk->public.throttle_group_member; + GLOBAL_STATE_CODE(); bdrv_ref(bs); blk->root = bdrv_root_attach_child(bs, "root", &child_root, BDRV_CHILD_FILTERED | BDRV_CHILD_PRIMARY, @@ -XXX,XX +XXX,XX @@ int blk_insert_bs(BlockBackend *blk, BlockDriverState *bs, Error **errp) */ int blk_replace_bs(BlockBackend *blk, BlockDriverState *new_bs, Error **errp) { + GLOBAL_STATE_CODE(); return bdrv_replace_child_bs(blk->root, new_bs, errp); } @@ -XXX,XX +XXX,XX @@ int blk_set_perm(BlockBackend *blk, uint64_t perm, uint64_t shared_perm, Error **errp) { int ret; + GLOBAL_STATE_CODE(); if (blk->root && !blk->disable_perm) { ret = bdrv_child_try_set_perm(blk->root, perm, shared_perm, errp); @@ -XXX,XX +XXX,XX @@ int blk_set_perm(BlockBackend *blk, uint64_t perm, uint64_t shared_perm, void blk_get_perm(BlockBackend *blk, uint64_t *perm, uint64_t *shared_perm) { + GLOBAL_STATE_CODE(); *perm = blk->perm; *shared_perm = blk->shared_perm; } @@ -XXX,XX +XXX,XX @@ void blk_get_perm(BlockBackend *blk, uint64_t *perm, uint64_t *shared_perm) */ int blk_attach_dev(BlockBackend *blk, DeviceState *dev) { + GLOBAL_STATE_CODE(); if (blk->dev) { return -EBUSY; } @@ -XXX,XX +XXX,XX @@ int blk_attach_dev(BlockBackend *blk, DeviceState *dev) void blk_detach_dev(BlockBackend *blk, DeviceState *dev) { assert(blk->dev == dev); + GLOBAL_STATE_CODE(); blk->dev = NULL; blk->dev_ops = NULL; blk->dev_opaque = NULL; @@ -XXX,XX +XXX,XX @@ void blk_detach_dev(BlockBackend *blk, DeviceState *dev) */ DeviceState *blk_get_attached_dev(BlockBackend *blk) { + GLOBAL_STATE_CODE(); return blk->dev; } @@ -XXX,XX +XXX,XX @@ BlockBackend *blk_by_dev(void *dev) { BlockBackend *blk = NULL; + GLOBAL_STATE_CODE(); + assert(dev != NULL); while ((blk = blk_all_next(blk)) != NULL) { if (blk->dev == dev) { @@ -XXX,XX +XXX,XX @@ BlockBackend *blk_by_dev(void *dev) void blk_set_dev_ops(BlockBackend *blk, const BlockDevOps *ops, void *opaque) { + GLOBAL_STATE_CODE(); blk->dev_ops = ops; blk->dev_opaque = opaque; @@ -XXX,XX +XXX,XX @@ void blk_set_dev_ops(BlockBackend *blk, const BlockDevOps *ops, */ void blk_dev_change_media_cb(BlockBackend *blk, bool load, Error **errp) { + GLOBAL_STATE_CODE(); if (blk->dev_ops && blk->dev_ops->change_media_cb) { bool tray_was_open, tray_is_open; Error *local_err = NULL; @@ -XXX,XX +XXX,XX @@ static void blk_root_resize(BdrvChild *child) void blk_iostatus_enable(BlockBackend *blk) { + GLOBAL_STATE_CODE(); blk->iostatus_enabled = true; blk->iostatus = BLOCK_DEVICE_IO_STATUS_OK; } @@ -XXX,XX +XXX,XX @@ bool blk_iostatus_is_enabled(const BlockBackend *blk) BlockDeviceIoStatus blk_iostatus(const BlockBackend *blk) { + GLOBAL_STATE_CODE(); return blk->iostatus; } void blk_iostatus_disable(BlockBackend *blk) { + GLOBAL_STATE_CODE(); blk->iostatus_enabled = false; } void blk_iostatus_reset(BlockBackend *blk) { + GLOBAL_STATE_CODE(); if (blk_iostatus_is_enabled(blk)) { blk->iostatus = BLOCK_DEVICE_IO_STATUS_OK; } @@ -XXX,XX +XXX,XX @@ int blk_pwrite_zeroes(BlockBackend *blk, int64_t offset, int blk_make_zero(BlockBackend *blk, BdrvRequestFlags flags) { + GLOBAL_STATE_CODE(); return bdrv_make_zero(blk->root, flags); } @@ -XXX,XX +XXX,XX @@ BlockAIOCB *blk_aio_pwritev(BlockBackend *blk, int64_t offset, void blk_aio_cancel(BlockAIOCB *acb) { + GLOBAL_STATE_CODE(); bdrv_aio_cancel(acb); } @@ -XXX,XX +XXX,XX @@ int blk_flush(BlockBackend *blk) void blk_drain(BlockBackend *blk) { BlockDriverState *bs = blk_bs(blk); + GLOBAL_STATE_CODE(); if (bs) { bdrv_ref(bs); @@ -XXX,XX +XXX,XX @@ void blk_drain_all(void) { BlockBackend *blk = NULL; + GLOBAL_STATE_CODE(); + bdrv_drain_all_begin(); while ((blk = blk_all_next(blk)) != NULL) { @@ -XXX,XX +XXX,XX @@ void blk_drain_all(void) void blk_set_on_error(BlockBackend *blk, BlockdevOnError on_read_error, BlockdevOnError on_write_error) { + GLOBAL_STATE_CODE(); blk->on_read_error = on_read_error; blk->on_write_error = on_write_error; } @@ -XXX,XX +XXX,XX @@ void blk_error_action(BlockBackend *blk, BlockErrorAction action, bool blk_supports_write_perm(BlockBackend *blk) { BlockDriverState *bs = blk_bs(blk); + GLOBAL_STATE_CODE(); if (bs) { return !bdrv_is_read_only(bs); @@ -XXX,XX +XXX,XX @@ bool blk_is_writable(BlockBackend *blk) bool blk_is_sg(BlockBackend *blk) { BlockDriverState *bs = blk_bs(blk); + GLOBAL_STATE_CODE(); if (!bs) { return false; @@ -XXX,XX +XXX,XX @@ bool blk_enable_write_cache(BlockBackend *blk) void blk_set_enable_write_cache(BlockBackend *blk, bool wce) { + GLOBAL_STATE_CODE(); blk->enable_write_cache = wce; } void blk_activate(BlockBackend *blk, Error **errp) { BlockDriverState *bs = blk_bs(blk); + GLOBAL_STATE_CODE(); if (!bs) { error_setg(errp, "Device '%s' has no medium", blk->name); @@ -XXX,XX +XXX,XX @@ void blk_eject(BlockBackend *blk, bool eject_flag) int blk_get_flags(BlockBackend *blk) { BlockDriverState *bs = blk_bs(blk); + GLOBAL_STATE_CODE(); if (bs) { return bdrv_get_flags(bs); @@ -XXX,XX +XXX,XX @@ void *blk_blockalign(BlockBackend *blk, size_t size) bool blk_op_is_blocked(BlockBackend *blk, BlockOpType op, Error **errp) { BlockDriverState *bs = blk_bs(blk); + GLOBAL_STATE_CODE(); if (!bs) { return false; @@ -XXX,XX +XXX,XX @@ bool blk_op_is_blocked(BlockBackend *blk, BlockOpType op, Error **errp) void blk_op_unblock(BlockBackend *blk, BlockOpType op, Error *reason) { BlockDriverState *bs = blk_bs(blk); + GLOBAL_STATE_CODE(); if (bs) { bdrv_op_unblock(bs, op, reason); @@ -XXX,XX +XXX,XX @@ void blk_op_unblock(BlockBackend *blk, BlockOpType op, Error *reason) void blk_op_block_all(BlockBackend *blk, Error *reason) { BlockDriverState *bs = blk_bs(blk); + GLOBAL_STATE_CODE(); if (bs) { bdrv_op_block_all(bs, reason); @@ -XXX,XX +XXX,XX @@ void blk_op_block_all(BlockBackend *blk, Error *reason) void blk_op_unblock_all(BlockBackend *blk, Error *reason) { BlockDriverState *bs = blk_bs(blk); + GLOBAL_STATE_CODE(); if (bs) { bdrv_op_unblock_all(bs, reason); @@ -XXX,XX +XXX,XX @@ static int blk_do_set_aio_context(BlockBackend *blk, AioContext *new_context, int blk_set_aio_context(BlockBackend *blk, AioContext *new_context, Error **errp) { + GLOBAL_STATE_CODE(); return blk_do_set_aio_context(blk, new_context, true, errp); } @@ -XXX,XX +XXX,XX @@ void blk_add_aio_context_notifier(BlockBackend *blk, { BlockBackendAioNotifier *notifier; BlockDriverState *bs = blk_bs(blk); + GLOBAL_STATE_CODE(); notifier = g_new(BlockBackendAioNotifier, 1); notifier->attached_aio_context = attached_aio_context; @@ -XXX,XX +XXX,XX @@ void blk_remove_aio_context_notifier(BlockBackend *blk, BlockBackendAioNotifier *notifier; BlockDriverState *bs = blk_bs(blk); + GLOBAL_STATE_CODE(); + if (bs) { bdrv_remove_aio_context_notifier(bs, attached_aio_context, detach_aio_context, opaque); @@ -XXX,XX +XXX,XX @@ void blk_remove_aio_context_notifier(BlockBackend *blk, void blk_add_remove_bs_notifier(BlockBackend *blk, Notifier *notify) { + GLOBAL_STATE_CODE(); notifier_list_add(&blk->remove_bs_notifiers, notify); } void blk_add_insert_bs_notifier(BlockBackend *blk, Notifier *notify) { + GLOBAL_STATE_CODE(); notifier_list_add(&blk->insert_bs_notifiers, notify); } @@ -XXX,XX +XXX,XX @@ int blk_save_vmstate(BlockBackend *blk, const uint8_t *buf, int64_t pos, int size) { int ret; + GLOBAL_STATE_CODE(); if (!blk_is_available(blk)) { return -ENOMEDIUM; @@ -XXX,XX +XXX,XX @@ int blk_save_vmstate(BlockBackend *blk, const uint8_t *buf, int blk_load_vmstate(BlockBackend *blk, uint8_t *buf, int64_t pos, int size) { + GLOBAL_STATE_CODE(); if (!blk_is_available(blk)) { return -ENOMEDIUM; } @@ -XXX,XX +XXX,XX @@ int blk_load_vmstate(BlockBackend *blk, uint8_t *buf, int64_t pos, int size) int blk_probe_blocksizes(BlockBackend *blk, BlockSizes *bsz) { + GLOBAL_STATE_CODE(); if (!blk_is_available(blk)) { return -ENOMEDIUM; } @@ -XXX,XX +XXX,XX @@ int blk_probe_blocksizes(BlockBackend *blk, BlockSizes *bsz) int blk_probe_geometry(BlockBackend *blk, HDGeometry *geo) { + GLOBAL_STATE_CODE(); if (!blk_is_available(blk)) { return -ENOMEDIUM; } @@ -XXX,XX +XXX,XX @@ int blk_probe_geometry(BlockBackend *blk, HDGeometry *geo) */ void blk_update_root_state(BlockBackend *blk) { + GLOBAL_STATE_CODE(); assert(blk->root); blk->root_state.open_flags = blk->root->bs->open_flags; @@ -XXX,XX +XXX,XX @@ void blk_update_root_state(BlockBackend *blk) */ bool blk_get_detect_zeroes_from_root_state(BlockBackend *blk) { + GLOBAL_STATE_CODE(); return blk->root_state.detect_zeroes; } @@ -XXX,XX +XXX,XX @@ bool blk_get_detect_zeroes_from_root_state(BlockBackend *blk) */ int blk_get_open_flags_from_root_state(BlockBackend *blk) { + GLOBAL_STATE_CODE(); return blk->root_state.open_flags; } BlockBackendRootState *blk_get_root_state(BlockBackend *blk) { + GLOBAL_STATE_CODE(); return &blk->root_state; } int blk_commit_all(void) { BlockBackend *blk = NULL; + GLOBAL_STATE_CODE(); while ((blk = blk_all_next(blk)) != NULL) { AioContext *aio_context = blk_get_aio_context(blk); @@ -XXX,XX +XXX,XX @@ int blk_commit_all(void) /* throttling disk I/O limits */ void blk_set_io_limits(BlockBackend *blk, ThrottleConfig *cfg) { + GLOBAL_STATE_CODE(); throttle_group_config(&blk->public.throttle_group_member, cfg); } @@ -XXX,XX +XXX,XX @@ void blk_io_limits_disable(BlockBackend *blk) BlockDriverState *bs = blk_bs(blk); ThrottleGroupMember *tgm = &blk->public.throttle_group_member; assert(tgm->throttle_state); + GLOBAL_STATE_CODE(); if (bs) { bdrv_ref(bs); bdrv_drained_begin(bs); @@ -XXX,XX +XXX,XX @@ void blk_io_limits_disable(BlockBackend *blk) void blk_io_limits_enable(BlockBackend *blk, const char *group) { assert(!blk->public.throttle_group_member.throttle_state); + GLOBAL_STATE_CODE(); throttle_group_register_tgm(&blk->public.throttle_group_member, group, blk_get_aio_context(blk)); } void blk_io_limits_update_group(BlockBackend *blk, const char *group) { + GLOBAL_STATE_CODE(); /* this BB is not part of any group */ if (!blk->public.throttle_group_member.throttle_state) { return; @@ -XXX,XX +XXX,XX @@ static void blk_root_drained_end(BdrvChild *child, int *drained_end_counter) void blk_register_buf(BlockBackend *blk, void *host, size_t size) { + GLOBAL_STATE_CODE(); bdrv_register_buf(blk_bs(blk), host, size); } void blk_unregister_buf(BlockBackend *blk, void *host) { + GLOBAL_STATE_CODE(); bdrv_unregister_buf(blk_bs(blk), host); } @@ -XXX,XX +XXX,XX @@ int coroutine_fn blk_co_copy_range(BlockBackend *blk_in, int64_t off_in, const BdrvChild *blk_root(BlockBackend *blk) { + GLOBAL_STATE_CODE(); return blk->root; } int blk_make_empty(BlockBackend *blk, Error **errp) { + GLOBAL_STATE_CODE(); if (!blk_is_available(blk)) { error_setg(errp, "No medium inserted"); return -ENOMEDIUM; diff --git a/softmmu/qdev-monitor.c b/softmmu/qdev-monitor.c index XXXXXXX..XXXXXXX 100644 --- a/softmmu/qdev-monitor.c +++ b/softmmu/qdev-monitor.c @@ -XXX,XX +XXX,XX @@ BlockBackend *blk_by_qdev_id(const char *id, Error **errp) DeviceState *dev; BlockBackend *blk; + GLOBAL_STATE_CODE(); + dev = find_device_state(id, errp); if (dev == NULL) { return NULL; -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> Mark all I/O functions with IO_CODE, and all "I/O OR GS" with IO_OR_GS_CODE. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Message-Id: <20220303151616.325444-10-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/sysemu/block-backend-io.h | 2 ++ block/block-backend.c | 58 +++++++++++++++++++++++++++++++ 2 files changed, 60 insertions(+) diff --git a/include/sysemu/block-backend-io.h b/include/sysemu/block-backend-io.h index XXXXXXX..XXXXXXX 100644 --- a/include/sysemu/block-backend-io.h +++ b/include/sysemu/block-backend-io.h @@ -XXX,XX +XXX,XX @@ static inline int coroutine_fn blk_co_pread(BlockBackend *blk, int64_t offset, BdrvRequestFlags flags) { QEMUIOVector qiov = QEMU_IOVEC_INIT_BUF(qiov, buf, bytes); + IO_OR_GS_CODE(); assert(bytes <= SIZE_MAX); @@ -XXX,XX +XXX,XX @@ static inline int coroutine_fn blk_co_pwrite(BlockBackend *blk, int64_t offset, BdrvRequestFlags flags) { QEMUIOVector qiov = QEMU_IOVEC_INIT_BUF(qiov, buf, bytes); + IO_OR_GS_CODE(); assert(bytes <= SIZE_MAX); diff --git a/block/block-backend.c b/block/block-backend.c index XXXXXXX..XXXXXXX 100644 --- a/block/block-backend.c +++ b/block/block-backend.c @@ -XXX,XX +XXX,XX @@ void monitor_remove_blk(BlockBackend *blk) */ const char *blk_name(const BlockBackend *blk) { + IO_CODE(); return blk->name ?: ""; } @@ -XXX,XX +XXX,XX @@ BlockBackend *blk_by_name(const char *name) */ BlockDriverState *blk_bs(BlockBackend *blk) { + IO_CODE(); return blk->root ? blk->root->bs : NULL; } @@ -XXX,XX +XXX,XX @@ DeviceState *blk_get_attached_dev(BlockBackend *blk) char *blk_get_attached_dev_id(BlockBackend *blk) { DeviceState *dev = blk->dev; + IO_CODE(); if (!dev) { return g_strdup(""); @@ -XXX,XX +XXX,XX @@ void blk_iostatus_enable(BlockBackend *blk) * enables it _and_ the VM is configured to stop on errors */ bool blk_iostatus_is_enabled(const BlockBackend *blk) { + IO_CODE(); return (blk->iostatus_enabled && (blk->on_write_error == BLOCKDEV_ON_ERROR_ENOSPC || blk->on_write_error == BLOCKDEV_ON_ERROR_STOP || @@ -XXX,XX +XXX,XX @@ void blk_iostatus_reset(BlockBackend *blk) void blk_iostatus_set_err(BlockBackend *blk, int error) { + IO_CODE(); assert(blk_iostatus_is_enabled(blk)); if (blk->iostatus == BLOCK_DEVICE_IO_STATUS_OK) { blk->iostatus = error == ENOSPC ? BLOCK_DEVICE_IO_STATUS_NOSPACE : @@ -XXX,XX +XXX,XX @@ void blk_iostatus_set_err(BlockBackend *blk, int error) void blk_set_allow_write_beyond_eof(BlockBackend *blk, bool allow) { + IO_CODE(); blk->allow_write_beyond_eof = allow; } void blk_set_allow_aio_context_change(BlockBackend *blk, bool allow) { + IO_CODE(); blk->allow_aio_context_change = allow; } void blk_set_disable_request_queuing(BlockBackend *blk, bool disable) { + IO_CODE(); blk->disable_request_queuing = disable; } @@ -XXX,XX +XXX,XX @@ int coroutine_fn blk_co_preadv(BlockBackend *blk, int64_t offset, BdrvRequestFlags flags) { int ret; + IO_OR_GS_CODE(); blk_inc_in_flight(blk); ret = blk_co_do_preadv(blk, offset, bytes, qiov, flags); @@ -XXX,XX +XXX,XX @@ int coroutine_fn blk_co_pwritev_part(BlockBackend *blk, int64_t offset, BdrvRequestFlags flags) { int ret; + IO_OR_GS_CODE(); blk_inc_in_flight(blk); ret = blk_co_do_pwritev_part(blk, offset, bytes, qiov, qiov_offset, flags); @@ -XXX,XX +XXX,XX @@ int coroutine_fn blk_co_pwritev(BlockBackend *blk, int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags) { + IO_OR_GS_CODE(); return blk_co_pwritev_part(blk, offset, bytes, qiov, 0, flags); } @@ -XXX,XX +XXX,XX @@ typedef struct BlkRwCo { int blk_pwrite_zeroes(BlockBackend *blk, int64_t offset, int64_t bytes, BdrvRequestFlags flags) { + IO_OR_GS_CODE(); return blk_pwritev_part(blk, offset, bytes, NULL, 0, flags | BDRV_REQ_ZERO_WRITE); } @@ -XXX,XX +XXX,XX @@ int blk_make_zero(BlockBackend *blk, BdrvRequestFlags flags) void blk_inc_in_flight(BlockBackend *blk) { + IO_CODE(); qatomic_inc(&blk->in_flight); } void blk_dec_in_flight(BlockBackend *blk) { + IO_CODE(); qatomic_dec(&blk->in_flight); aio_wait_kick(); } @@ -XXX,XX +XXX,XX @@ BlockAIOCB *blk_abort_aio_request(BlockBackend *blk, void *opaque, int ret) { struct BlockBackendAIOCB *acb; + IO_CODE(); blk_inc_in_flight(blk); acb = blk_aio_get(&block_backend_aiocb_info, blk, cb, opaque); @@ -XXX,XX +XXX,XX @@ BlockAIOCB *blk_aio_pwrite_zeroes(BlockBackend *blk, int64_t offset, int64_t bytes, BdrvRequestFlags flags, BlockCompletionFunc *cb, void *opaque) { + IO_CODE(); return blk_aio_prwv(blk, offset, bytes, NULL, blk_aio_write_entry, flags | BDRV_REQ_ZERO_WRITE, cb, opaque); } @@ -XXX,XX +XXX,XX @@ int blk_pread(BlockBackend *blk, int64_t offset, void *buf, int bytes) { int ret; QEMUIOVector qiov = QEMU_IOVEC_INIT_BUF(qiov, buf, bytes); + IO_OR_GS_CODE(); blk_inc_in_flight(blk); ret = blk_do_preadv(blk, offset, bytes, &qiov, 0); @@ -XXX,XX +XXX,XX @@ int blk_pwrite(BlockBackend *blk, int64_t offset, const void *buf, int bytes, { int ret; QEMUIOVector qiov = QEMU_IOVEC_INIT_BUF(qiov, buf, bytes); + IO_OR_GS_CODE(); ret = blk_pwritev_part(blk, offset, bytes, &qiov, 0, flags); @@ -XXX,XX +XXX,XX @@ int blk_pwrite(BlockBackend *blk, int64_t offset, const void *buf, int bytes, int64_t blk_getlength(BlockBackend *blk) { + IO_CODE(); if (!blk_is_available(blk)) { return -ENOMEDIUM; } @@ -XXX,XX +XXX,XX @@ int64_t blk_getlength(BlockBackend *blk) void blk_get_geometry(BlockBackend *blk, uint64_t *nb_sectors_ptr) { + IO_CODE(); if (!blk_bs(blk)) { *nb_sectors_ptr = 0; } else { @@ -XXX,XX +XXX,XX @@ void blk_get_geometry(BlockBackend *blk, uint64_t *nb_sectors_ptr) int64_t blk_nb_sectors(BlockBackend *blk) { + IO_CODE(); if (!blk_is_available(blk)) { return -ENOMEDIUM; } @@ -XXX,XX +XXX,XX @@ BlockAIOCB *blk_aio_preadv(BlockBackend *blk, int64_t offset, QEMUIOVector *qiov, BdrvRequestFlags flags, BlockCompletionFunc *cb, void *opaque) { + IO_CODE(); assert((uint64_t)qiov->size <= INT64_MAX); return blk_aio_prwv(blk, offset, qiov->size, qiov, blk_aio_read_entry, flags, cb, opaque); @@ -XXX,XX +XXX,XX @@ BlockAIOCB *blk_aio_pwritev(BlockBackend *blk, int64_t offset, QEMUIOVector *qiov, BdrvRequestFlags flags, BlockCompletionFunc *cb, void *opaque) { + IO_CODE(); assert((uint64_t)qiov->size <= INT64_MAX); return blk_aio_prwv(blk, offset, qiov->size, qiov, blk_aio_write_entry, flags, cb, opaque); @@ -XXX,XX +XXX,XX @@ void blk_aio_cancel(BlockAIOCB *acb) void blk_aio_cancel_async(BlockAIOCB *acb) { + IO_CODE(); bdrv_aio_cancel_async(acb); } @@ -XXX,XX +XXX,XX @@ blk_co_do_ioctl(BlockBackend *blk, unsigned long int req, void *buf) int blk_ioctl(BlockBackend *blk, unsigned long int req, void *buf) { int ret; + IO_OR_GS_CODE(); blk_inc_in_flight(blk); ret = blk_do_ioctl(blk, req, buf); @@ -XXX,XX +XXX,XX @@ static void blk_aio_ioctl_entry(void *opaque) BlockAIOCB *blk_aio_ioctl(BlockBackend *blk, unsigned long int req, void *buf, BlockCompletionFunc *cb, void *opaque) { + IO_CODE(); return blk_aio_prwv(blk, req, 0, buf, blk_aio_ioctl_entry, 0, cb, opaque); } @@ -XXX,XX +XXX,XX @@ BlockAIOCB *blk_aio_pdiscard(BlockBackend *blk, int64_t offset, int64_t bytes, BlockCompletionFunc *cb, void *opaque) { + IO_CODE(); return blk_aio_prwv(blk, offset, bytes, NULL, blk_aio_pdiscard_entry, 0, cb, opaque); } @@ -XXX,XX +XXX,XX @@ int coroutine_fn blk_co_pdiscard(BlockBackend *blk, int64_t offset, int64_t bytes) { int ret; + IO_OR_GS_CODE(); blk_inc_in_flight(blk); ret = blk_co_do_pdiscard(blk, offset, bytes); @@ -XXX,XX +XXX,XX @@ int coroutine_fn blk_co_pdiscard(BlockBackend *blk, int64_t offset, int blk_pdiscard(BlockBackend *blk, int64_t offset, int64_t bytes) { int ret; + IO_OR_GS_CODE(); blk_inc_in_flight(blk); ret = blk_do_pdiscard(blk, offset, bytes); @@ -XXX,XX +XXX,XX @@ static void blk_aio_flush_entry(void *opaque) BlockAIOCB *blk_aio_flush(BlockBackend *blk, BlockCompletionFunc *cb, void *opaque) { + IO_CODE(); return blk_aio_prwv(blk, 0, 0, NULL, blk_aio_flush_entry, 0, cb, opaque); } int coroutine_fn blk_co_flush(BlockBackend *blk) { int ret; + IO_OR_GS_CODE(); blk_inc_in_flight(blk); ret = blk_co_do_flush(blk); @@ -XXX,XX +XXX,XX @@ void blk_set_on_error(BlockBackend *blk, BlockdevOnError on_read_error, BlockdevOnError blk_get_on_error(BlockBackend *blk, bool is_read) { + IO_CODE(); return is_read ? blk->on_read_error : blk->on_write_error; } @@ -XXX,XX +XXX,XX @@ BlockErrorAction blk_get_error_action(BlockBackend *blk, bool is_read, int error) { BlockdevOnError on_err = blk_get_on_error(blk, is_read); + IO_CODE(); switch (on_err) { case BLOCKDEV_ON_ERROR_ENOSPC: @@ -XXX,XX +XXX,XX @@ void blk_error_action(BlockBackend *blk, BlockErrorAction action, bool is_read, int error) { assert(error >= 0); + IO_CODE(); if (action == BLOCK_ERROR_ACTION_STOP) { /* First set the iostatus, so that "info block" returns an iostatus @@ -XXX,XX +XXX,XX @@ bool blk_supports_write_perm(BlockBackend *blk) */ bool blk_is_writable(BlockBackend *blk) { + IO_CODE(); return blk->perm & BLK_PERM_WRITE; } @@ -XXX,XX +XXX,XX @@ bool blk_is_sg(BlockBackend *blk) bool blk_enable_write_cache(BlockBackend *blk) { + IO_CODE(); return blk->enable_write_cache; } @@ -XXX,XX +XXX,XX @@ void blk_activate(BlockBackend *blk, Error **errp) bool blk_is_inserted(BlockBackend *blk) { BlockDriverState *bs = blk_bs(blk); + IO_CODE(); return bs && bdrv_is_inserted(bs); } bool blk_is_available(BlockBackend *blk) { + IO_CODE(); return blk_is_inserted(blk) && !blk_dev_is_tray_open(blk); } void blk_lock_medium(BlockBackend *blk, bool locked) { BlockDriverState *bs = blk_bs(blk); + IO_CODE(); if (bs) { bdrv_lock_medium(bs, locked); @@ -XXX,XX +XXX,XX @@ void blk_eject(BlockBackend *blk, bool eject_flag) { BlockDriverState *bs = blk_bs(blk); char *id; + IO_CODE(); if (bs) { bdrv_eject(bs, eject_flag); @@ -XXX,XX +XXX,XX @@ int blk_get_flags(BlockBackend *blk) uint32_t blk_get_request_alignment(BlockBackend *blk) { BlockDriverState *bs = blk_bs(blk); + IO_CODE(); return bs ? bs->bl.request_alignment : BDRV_SECTOR_SIZE; } @@ -XXX,XX +XXX,XX @@ uint64_t blk_get_max_hw_transfer(BlockBackend *blk) { BlockDriverState *bs = blk_bs(blk); uint64_t max = INT_MAX; + IO_CODE(); if (bs) { max = MIN_NON_ZERO(max, bs->bl.max_hw_transfer); @@ -XXX,XX +XXX,XX @@ uint32_t blk_get_max_transfer(BlockBackend *blk) { BlockDriverState *bs = blk_bs(blk); uint32_t max = INT_MAX; + IO_CODE(); if (bs) { max = MIN_NON_ZERO(max, bs->bl.max_transfer); @@ -XXX,XX +XXX,XX @@ uint32_t blk_get_max_transfer(BlockBackend *blk) int blk_get_max_hw_iov(BlockBackend *blk) { + IO_CODE(); return MIN_NON_ZERO(blk->root->bs->bl.max_hw_iov, blk->root->bs->bl.max_iov); } int blk_get_max_iov(BlockBackend *blk) { + IO_CODE(); return blk->root->bs->bl.max_iov; } void blk_set_guest_block_size(BlockBackend *blk, int align) { + IO_CODE(); blk->guest_block_size = align; } void *blk_try_blockalign(BlockBackend *blk, size_t size) { + IO_CODE(); return qemu_try_blockalign(blk ? blk_bs(blk) : NULL, size); } void *blk_blockalign(BlockBackend *blk, size_t size) { + IO_CODE(); return qemu_blockalign(blk ? blk_bs(blk) : NULL, size); } @@ -XXX,XX +XXX,XX @@ void blk_op_unblock_all(BlockBackend *blk, Error *reason) AioContext *blk_get_aio_context(BlockBackend *blk) { BlockDriverState *bs = blk_bs(blk); + IO_CODE(); if (bs) { AioContext *ctx = bdrv_get_aio_context(blk_bs(blk)); @@ -XXX,XX +XXX,XX @@ void blk_add_insert_bs_notifier(BlockBackend *blk, Notifier *notify) void blk_io_plug(BlockBackend *blk) { BlockDriverState *bs = blk_bs(blk); + IO_CODE(); if (bs) { bdrv_io_plug(bs); @@ -XXX,XX +XXX,XX @@ void blk_io_plug(BlockBackend *blk) void blk_io_unplug(BlockBackend *blk) { BlockDriverState *bs = blk_bs(blk); + IO_CODE(); if (bs) { bdrv_io_unplug(bs); @@ -XXX,XX +XXX,XX @@ void blk_io_unplug(BlockBackend *blk) BlockAcctStats *blk_get_stats(BlockBackend *blk) { + IO_CODE(); return &blk->stats; } void *blk_aio_get(const AIOCBInfo *aiocb_info, BlockBackend *blk, BlockCompletionFunc *cb, void *opaque) { + IO_CODE(); return qemu_aio_get(aiocb_info, blk_bs(blk), cb, opaque); } int coroutine_fn blk_co_pwrite_zeroes(BlockBackend *blk, int64_t offset, int64_t bytes, BdrvRequestFlags flags) { + IO_OR_GS_CODE(); return blk_co_pwritev(blk, offset, bytes, NULL, flags | BDRV_REQ_ZERO_WRITE); } @@ -XXX,XX +XXX,XX @@ int blk_pwrite_compressed(BlockBackend *blk, int64_t offset, const void *buf, int64_t bytes) { QEMUIOVector qiov = QEMU_IOVEC_INIT_BUF(qiov, buf, bytes); + IO_OR_GS_CODE(); return blk_pwritev_part(blk, offset, bytes, &qiov, 0, BDRV_REQ_WRITE_COMPRESSED); } @@ -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, BdrvRequestFlags flags, Error **errp) { + IO_OR_GS_CODE(); if (!blk_is_available(blk)) { error_setg(errp, "No medium inserted"); return -ENOMEDIUM; @@ -XXX,XX +XXX,XX @@ int coroutine_fn blk_co_copy_range(BlockBackend *blk_in, int64_t off_in, BdrvRequestFlags write_flags) { int r; + IO_CODE(); + r = blk_check_byte_request(blk_in, off_in, bytes); if (r) { return r; -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> Now that we "covered" the three main cases where the permission API was being used under BQL (fuse, amend and invalidate_cache), we can safely assert for the permission functions implemented in block.c Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Message-Id: <20220303151616.325444-11-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block.c | 12 ++++++++++++ 1 file changed, 12 insertions(+) diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ static bool bdrv_a_allow_b(BdrvChild *a, BdrvChild *b, Error **errp) assert(a->bs); assert(a->bs == b->bs); + GLOBAL_STATE_CODE(); if ((b->perm & a->shared_perm) == b->perm) { return true; @@ -XXX,XX +XXX,XX @@ static bool bdrv_a_allow_b(BdrvChild *a, BdrvChild *b, Error **errp) static bool bdrv_parent_perms_conflict(BlockDriverState *bs, Error **errp) { BdrvChild *a, *b; + GLOBAL_STATE_CODE(); /* * During the loop we'll look at each pair twice. That's correct because @@ -XXX,XX +XXX,XX @@ static void bdrv_child_set_perm_abort(void *opaque) { BdrvChildSetPermState *s = opaque; + GLOBAL_STATE_CODE(); + s->child->perm = s->old_perm; s->child->shared_perm = s->old_shared_perm; } @@ -XXX,XX +XXX,XX @@ static void bdrv_child_set_perm(BdrvChild *c, uint64_t perm, uint64_t shared, Transaction *tran) { BdrvChildSetPermState *s = g_new(BdrvChildSetPermState, 1); + GLOBAL_STATE_CODE(); *s = (BdrvChildSetPermState) { .child = c, @@ -XXX,XX +XXX,XX @@ static int bdrv_node_refresh_perm(BlockDriverState *bs, BlockReopenQueue *q, BdrvChild *c; int ret; uint64_t cumulative_perms, cumulative_shared_perms; + GLOBAL_STATE_CODE(); bdrv_get_cumulative_perm(bs, &cumulative_perms, &cumulative_shared_perms); @@ -XXX,XX +XXX,XX @@ static int bdrv_list_refresh_perms(GSList *list, BlockReopenQueue *q, { int ret; BlockDriverState *bs; + GLOBAL_STATE_CODE(); for ( ; list; list = list->next) { bs = list->data; @@ -XXX,XX +XXX,XX @@ static int bdrv_refresh_perms(BlockDriverState *bs, Error **errp) int ret; Transaction *tran = tran_new(); g_autoptr(GSList) list = bdrv_topological_dfs(NULL, NULL, bs); + GLOBAL_STATE_CODE(); ret = bdrv_list_refresh_perms(list, NULL, tran, errp); tran_finalize(tran, ret); @@ -XXX,XX +XXX,XX @@ static void bdrv_filter_default_perms(BlockDriverState *bs, BdrvChild *c, uint64_t perm, uint64_t shared, uint64_t *nperm, uint64_t *nshared) { + GLOBAL_STATE_CODE(); *nperm = perm & DEFAULT_PERM_PASSTHROUGH; *nshared = (shared & DEFAULT_PERM_PASSTHROUGH) | DEFAULT_PERM_UNCHANGED; } @@ -XXX,XX +XXX,XX @@ static void bdrv_default_perms_for_cow(BlockDriverState *bs, BdrvChild *c, uint64_t *nperm, uint64_t *nshared) { assert(role & BDRV_CHILD_COW); + GLOBAL_STATE_CODE(); /* * We want consistent read from backing files if the parent needs it. @@ -XXX,XX +XXX,XX @@ static void bdrv_default_perms_for_storage(BlockDriverState *bs, BdrvChild *c, { int flags; + GLOBAL_STATE_CODE(); assert(role & (BDRV_CHILD_METADATA | BDRV_CHILD_DATA)); flags = bdrv_reopen_get_flags(reopen_queue, bs); @@ -XXX,XX +XXX,XX @@ static void xdbg_graph_add_edge(XDbgBlockGraphConstructor *gr, void *parent, { BlockPermission qapi_perm; XDbgBlockGraphEdge *edge; + GLOBAL_STATE_CODE(); edge = g_new0(XDbgBlockGraphEdge, 1); -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> Similarly to the previous patch, split block_int.h in block_int-io.h and block_int-global-state.h block_int-common.h contains the structures shared between the two headers, and the functions that can't be categorized as I/O or global state. Assertions are added in the next patch. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Message-Id: <20220303151616.325444-12-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block_int-common.h | 1180 +++++++++++++++++++ include/block/block_int-global-state.h | 312 +++++ include/block/block_int-io.h | 179 +++ include/block/block_int.h | 1489 +----------------------- blockdev.c | 5 + 5 files changed, 1679 insertions(+), 1486 deletions(-) create mode 100644 include/block/block_int-common.h create mode 100644 include/block/block_int-global-state.h create mode 100644 include/block/block_int-io.h diff --git a/include/block/block_int-common.h b/include/block/block_int-common.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/block/block_int-common.h @@ -XXX,XX +XXX,XX @@ +/* + * QEMU System Emulator block driver + * + * Copyright (c) 2003 Fabrice Bellard + * + * 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. + */ +#ifndef BLOCK_INT_COMMON_H +#define BLOCK_INT_COMMON_H + +#include "block/accounting.h" +#include "block/block.h" +#include "block/aio-wait.h" +#include "qemu/queue.h" +#include "qemu/coroutine.h" +#include "qemu/stats64.h" +#include "qemu/timer.h" +#include "qemu/hbitmap.h" +#include "block/snapshot.h" +#include "qemu/throttle.h" +#include "qemu/rcu.h" + +#define BLOCK_FLAG_LAZY_REFCOUNTS 8 + +#define BLOCK_OPT_SIZE "size" +#define BLOCK_OPT_ENCRYPT "encryption" +#define BLOCK_OPT_ENCRYPT_FORMAT "encrypt.format" +#define BLOCK_OPT_COMPAT6 "compat6" +#define BLOCK_OPT_HWVERSION "hwversion" +#define BLOCK_OPT_BACKING_FILE "backing_file" +#define BLOCK_OPT_BACKING_FMT "backing_fmt" +#define BLOCK_OPT_CLUSTER_SIZE "cluster_size" +#define BLOCK_OPT_TABLE_SIZE "table_size" +#define BLOCK_OPT_PREALLOC "preallocation" +#define BLOCK_OPT_SUBFMT "subformat" +#define BLOCK_OPT_COMPAT_LEVEL "compat" +#define BLOCK_OPT_LAZY_REFCOUNTS "lazy_refcounts" +#define BLOCK_OPT_ADAPTER_TYPE "adapter_type" +#define BLOCK_OPT_REDUNDANCY "redundancy" +#define BLOCK_OPT_NOCOW "nocow" +#define BLOCK_OPT_EXTENT_SIZE_HINT "extent_size_hint" +#define BLOCK_OPT_OBJECT_SIZE "object_size" +#define BLOCK_OPT_REFCOUNT_BITS "refcount_bits" +#define BLOCK_OPT_DATA_FILE "data_file" +#define BLOCK_OPT_DATA_FILE_RAW "data_file_raw" +#define BLOCK_OPT_COMPRESSION_TYPE "compression_type" +#define BLOCK_OPT_EXTL2 "extended_l2" + +#define BLOCK_PROBE_BUF_SIZE 512 + +enum BdrvTrackedRequestType { + BDRV_TRACKED_READ, + BDRV_TRACKED_WRITE, + BDRV_TRACKED_DISCARD, + BDRV_TRACKED_TRUNCATE, +}; + +/* + * That is not quite good that BdrvTrackedRequest structure is public, + * as block/io.c is very careful about incoming offset/bytes being + * correct. Be sure to assert bdrv_check_request() succeeded after any + * modification of BdrvTrackedRequest object out of block/io.c + */ +typedef struct BdrvTrackedRequest { + BlockDriverState *bs; + int64_t offset; + int64_t bytes; + enum BdrvTrackedRequestType type; + + bool serialising; + int64_t overlap_offset; + int64_t overlap_bytes; + + QLIST_ENTRY(BdrvTrackedRequest) list; + Coroutine *co; /* owner, used for deadlock detection */ + CoQueue wait_queue; /* coroutines blocked on this request */ + + struct BdrvTrackedRequest *waiting_for; +} BdrvTrackedRequest; + + +struct BlockDriver { + const char *format_name; + int instance_size; + + /* + * Set to true if the BlockDriver is a block filter. Block filters pass + * certain callbacks that refer to data (see block.c) to their bs->file + * or bs->backing (whichever one exists) if the driver doesn't implement + * them. Drivers that do not wish to forward must implement them and return + * -ENOTSUP. + * Note that filters are not allowed to modify data. + * + * Filters generally cannot have more than a single filtered child, + * because the data they present must at all times be the same as + * that on their filtered child. That would be impossible to + * achieve for multiple filtered children. + * (And this filtered child must then be bs->file or bs->backing.) + */ + bool is_filter; + /* + * Set to true if the BlockDriver is a format driver. Format nodes + * generally do not expect their children to be other format nodes + * (except for backing files), and so format probing is disabled + * on those children. + */ + bool is_format; + + /* + * This function is invoked under BQL before .bdrv_co_amend() + * (which in contrast does not necessarily run under the BQL) + * to allow driver-specific initialization code that requires + * the BQL, like setting up specific permission flags. + */ + int (*bdrv_amend_pre_run)(BlockDriverState *bs, Error **errp); + /* + * This function is invoked under BQL after .bdrv_co_amend() + * to allow cleaning up what was done in .bdrv_amend_pre_run(). + */ + void (*bdrv_amend_clean)(BlockDriverState *bs); + + /* + * Return true if @to_replace can be replaced by a BDS with the + * same data as @bs without it affecting @bs's behavior (that is, + * without it being visible to @bs's parents). + */ + bool (*bdrv_recurse_can_replace)(BlockDriverState *bs, + BlockDriverState *to_replace); + + int (*bdrv_probe)(const uint8_t *buf, int buf_size, const char *filename); + int (*bdrv_probe_device)(const char *filename); + + /* + * Any driver implementing this callback is expected to be able to handle + * NULL file names in its .bdrv_open() implementation. + */ + void (*bdrv_parse_filename)(const char *filename, QDict *options, + Error **errp); + /* + * Drivers not implementing bdrv_parse_filename nor bdrv_open should have + * this field set to true, except ones that are defined only by their + * child's bs. + * An example of the last type will be the quorum block driver. + */ + bool bdrv_needs_filename; + + /* + * Set if a driver can support backing files. This also implies the + * following semantics: + * + * - Return status 0 of .bdrv_co_block_status means that corresponding + * blocks are not allocated in this layer of backing-chain + * - For such (unallocated) blocks, read will: + * - fill buffer with zeros if there is no backing file + * - read from the backing file otherwise, where the block layer + * takes care of reading zeros beyond EOF if backing file is short + */ + bool supports_backing; + + /* For handling image reopen for split or non-split files */ + int (*bdrv_reopen_prepare)(BDRVReopenState *reopen_state, + BlockReopenQueue *queue, Error **errp); + void (*bdrv_reopen_commit)(BDRVReopenState *reopen_state); + void (*bdrv_reopen_commit_post)(BDRVReopenState *reopen_state); + void (*bdrv_reopen_abort)(BDRVReopenState *reopen_state); + void (*bdrv_join_options)(QDict *options, QDict *old_options); + + int (*bdrv_open)(BlockDriverState *bs, QDict *options, int flags, + Error **errp); + + /* Protocol drivers should implement this instead of bdrv_open */ + int (*bdrv_file_open)(BlockDriverState *bs, QDict *options, int flags, + Error **errp); + void (*bdrv_close)(BlockDriverState *bs); + + + int coroutine_fn (*bdrv_co_create)(BlockdevCreateOptions *opts, + Error **errp); + int coroutine_fn (*bdrv_co_create_opts)(BlockDriver *drv, + const char *filename, + QemuOpts *opts, + Error **errp); + + int coroutine_fn (*bdrv_co_amend)(BlockDriverState *bs, + BlockdevAmendOptions *opts, + bool force, + Error **errp); + + int (*bdrv_amend_options)(BlockDriverState *bs, + QemuOpts *opts, + BlockDriverAmendStatusCB *status_cb, + void *cb_opaque, + bool force, + Error **errp); + + int (*bdrv_make_empty)(BlockDriverState *bs); + + /* + * Refreshes the bs->exact_filename field. If that is impossible, + * bs->exact_filename has to be left empty. + */ + void (*bdrv_refresh_filename)(BlockDriverState *bs); + + /* + * Gathers the open options for all children into @target. + * A simple format driver (without backing file support) might + * implement this function like this: + * + * QINCREF(bs->file->bs->full_open_options); + * qdict_put(target, "file", bs->file->bs->full_open_options); + * + * If not specified, the generic implementation will simply put + * all children's options under their respective name. + * + * @backing_overridden is true when bs->backing seems not to be + * the child that would result from opening bs->backing_file. + * Therefore, if it is true, the backing child's options should be + * gathered; otherwise, there is no need since the backing child + * is the one implied by the image header. + * + * Note that ideally this function would not be needed. Every + * block driver which implements it is probably doing something + * shady regarding its runtime option structure. + */ + void (*bdrv_gather_child_options)(BlockDriverState *bs, QDict *target, + bool backing_overridden); + + /* + * Returns an allocated string which is the directory name of this BDS: It + * will be used to make relative filenames absolute by prepending this + * function's return value to them. + */ + char *(*bdrv_dirname)(BlockDriverState *bs, Error **errp); + + /* aio */ + BlockAIOCB *(*bdrv_aio_preadv)(BlockDriverState *bs, + int64_t offset, int64_t bytes, QEMUIOVector *qiov, + BdrvRequestFlags flags, BlockCompletionFunc *cb, void *opaque); + BlockAIOCB *(*bdrv_aio_pwritev)(BlockDriverState *bs, + int64_t offset, int64_t bytes, QEMUIOVector *qiov, + BdrvRequestFlags flags, BlockCompletionFunc *cb, void *opaque); + BlockAIOCB *(*bdrv_aio_flush)(BlockDriverState *bs, + BlockCompletionFunc *cb, void *opaque); + BlockAIOCB *(*bdrv_aio_pdiscard)(BlockDriverState *bs, + int64_t offset, int bytes, + BlockCompletionFunc *cb, void *opaque); + + int coroutine_fn (*bdrv_co_readv)(BlockDriverState *bs, + int64_t sector_num, int nb_sectors, QEMUIOVector *qiov); + + /** + * @offset: position in bytes to read at + * @bytes: number of bytes to read + * @qiov: the buffers to fill with read data + * @flags: currently unused, always 0 + * + * @offset and @bytes will be a multiple of 'request_alignment', + * but the length of individual @qiov elements does not have to + * be a multiple. + * + * @bytes will always equal the total size of @qiov, and will be + * no larger than 'max_transfer'. + * + * The buffer in @qiov may point directly to guest memory. + */ + int coroutine_fn (*bdrv_co_preadv)(BlockDriverState *bs, + int64_t offset, int64_t bytes, QEMUIOVector *qiov, + BdrvRequestFlags flags); + + int coroutine_fn (*bdrv_co_preadv_part)(BlockDriverState *bs, + int64_t offset, int64_t bytes, + QEMUIOVector *qiov, size_t qiov_offset, + BdrvRequestFlags flags); + + int coroutine_fn (*bdrv_co_writev)(BlockDriverState *bs, + int64_t sector_num, int nb_sectors, QEMUIOVector *qiov, + int flags); + /** + * @offset: position in bytes to write at + * @bytes: number of bytes to write + * @qiov: the buffers containing data to write + * @flags: zero or more bits allowed by 'supported_write_flags' + * + * @offset and @bytes will be a multiple of 'request_alignment', + * but the length of individual @qiov elements does not have to + * be a multiple. + * + * @bytes will always equal the total size of @qiov, and will be + * no larger than 'max_transfer'. + * + * The buffer in @qiov may point directly to guest memory. + */ + int coroutine_fn (*bdrv_co_pwritev)(BlockDriverState *bs, + int64_t offset, int64_t bytes, QEMUIOVector *qiov, + BdrvRequestFlags flags); + int coroutine_fn (*bdrv_co_pwritev_part)(BlockDriverState *bs, + int64_t offset, int64_t bytes, QEMUIOVector *qiov, size_t qiov_offset, + BdrvRequestFlags flags); + + /* + * Efficiently zero a region of the disk image. Typically an image format + * would use a compact metadata representation to implement this. This + * function pointer may be NULL or return -ENOSUP and .bdrv_co_writev() + * will be called instead. + */ + int coroutine_fn (*bdrv_co_pwrite_zeroes)(BlockDriverState *bs, + int64_t offset, int64_t bytes, BdrvRequestFlags flags); + int coroutine_fn (*bdrv_co_pdiscard)(BlockDriverState *bs, + int64_t offset, int64_t bytes); + + /* + * Map [offset, offset + nbytes) range onto a child of @bs to copy from, + * and invoke bdrv_co_copy_range_from(child, ...), or invoke + * bdrv_co_copy_range_to() if @bs is the leaf child to copy data from. + * + * See the comment of bdrv_co_copy_range for the parameter and return value + * semantics. + */ + int coroutine_fn (*bdrv_co_copy_range_from)(BlockDriverState *bs, + BdrvChild *src, + int64_t offset, + BdrvChild *dst, + int64_t dst_offset, + int64_t bytes, + BdrvRequestFlags read_flags, + BdrvRequestFlags write_flags); + + /* + * Map [offset, offset + nbytes) range onto a child of bs to copy data to, + * and invoke bdrv_co_copy_range_to(child, src, ...), or perform the copy + * operation if @bs is the leaf and @src has the same BlockDriver. Return + * -ENOTSUP if @bs is the leaf but @src has a different BlockDriver. + * + * See the comment of bdrv_co_copy_range for the parameter and return value + * semantics. + */ + int coroutine_fn (*bdrv_co_copy_range_to)(BlockDriverState *bs, + BdrvChild *src, + int64_t src_offset, + BdrvChild *dst, + int64_t dst_offset, + int64_t bytes, + BdrvRequestFlags read_flags, + BdrvRequestFlags write_flags); + + /* + * Building block for bdrv_block_status[_above] and + * bdrv_is_allocated[_above]. The driver should answer only + * according to the current layer, and should only need to set + * BDRV_BLOCK_DATA, BDRV_BLOCK_ZERO, BDRV_BLOCK_OFFSET_VALID, + * and/or BDRV_BLOCK_RAW; if the current layer defers to a backing + * layer, the result should be 0 (and not BDRV_BLOCK_ZERO). See + * block.h for the overall meaning of the bits. As a hint, the + * flag want_zero is true if the caller cares more about precise + * mappings (favor accurate _OFFSET_VALID/_ZERO) or false for + * overall allocation (favor larger *pnum, perhaps by reporting + * _DATA instead of _ZERO). The block layer guarantees input + * clamped to bdrv_getlength() and aligned to request_alignment, + * as well as non-NULL pnum, map, and file; in turn, the driver + * must return an error or set pnum to an aligned non-zero value. + * + * Note that @bytes is just a hint on how big of a region the + * caller wants to inspect. It is not a limit on *pnum. + * Implementations are free to return larger values of *pnum if + * doing so does not incur a performance penalty. + * + * block/io.c's bdrv_co_block_status() will utilize an unclamped + * *pnum value for the block-status cache on protocol nodes, prior + * to clamping *pnum for return to its caller. + */ + int coroutine_fn (*bdrv_co_block_status)(BlockDriverState *bs, + bool want_zero, int64_t offset, int64_t bytes, int64_t *pnum, + int64_t *map, BlockDriverState **file); + + /* + * This informs the driver that we are no longer interested in the result + * of in-flight requests, so don't waste the time if possible. + * + * One example usage is to avoid waiting for an nbd target node reconnect + * timeout during job-cancel with force=true. + */ + void (*bdrv_cancel_in_flight)(BlockDriverState *bs); + + /* + * Invalidate any cached meta-data. + */ + void coroutine_fn (*bdrv_co_invalidate_cache)(BlockDriverState *bs, + Error **errp); + int (*bdrv_inactivate)(BlockDriverState *bs); + + /* + * Flushes all data for all layers by calling bdrv_co_flush for underlying + * layers, if needed. This function is needed for deterministic + * synchronization of the flush finishing callback. + */ + int coroutine_fn (*bdrv_co_flush)(BlockDriverState *bs); + + /* Delete a created file. */ + int coroutine_fn (*bdrv_co_delete_file)(BlockDriverState *bs, + Error **errp); + + /* + * Flushes all data that was already written to the OS all the way down to + * the disk (for example file-posix.c calls fsync()). + */ + int coroutine_fn (*bdrv_co_flush_to_disk)(BlockDriverState *bs); + + /* + * Flushes all internal caches to the OS. The data may still sit in a + * writeback cache of the host OS, but it will survive a crash of the qemu + * process. + */ + int coroutine_fn (*bdrv_co_flush_to_os)(BlockDriverState *bs); + + /* + * Drivers setting this field must be able to work with just a plain + * filename with '<protocol_name>:' as a prefix, and no other options. + * Options may be extracted from the filename by implementing + * bdrv_parse_filename. + */ + const char *protocol_name; + + /* + * Truncate @bs to @offset bytes using the given @prealloc mode + * when growing. Modes other than PREALLOC_MODE_OFF should be + * rejected when shrinking @bs. + * + * If @exact is true, @bs must be resized to exactly @offset. + * Otherwise, it is sufficient for @bs (if it is a host block + * device and thus there is no way to resize it) to be at least + * @offset bytes in length. + * + * If @exact is true and this function fails but would succeed + * with @exact = false, it should return -ENOTSUP. + */ + int coroutine_fn (*bdrv_co_truncate)(BlockDriverState *bs, int64_t offset, + bool exact, PreallocMode prealloc, + BdrvRequestFlags flags, Error **errp); + int64_t (*bdrv_getlength)(BlockDriverState *bs); + bool has_variable_length; + int64_t (*bdrv_get_allocated_file_size)(BlockDriverState *bs); + BlockMeasureInfo *(*bdrv_measure)(QemuOpts *opts, BlockDriverState *in_bs, + Error **errp); + + int coroutine_fn (*bdrv_co_pwritev_compressed)(BlockDriverState *bs, + int64_t offset, int64_t bytes, QEMUIOVector *qiov); + int coroutine_fn (*bdrv_co_pwritev_compressed_part)(BlockDriverState *bs, + int64_t offset, int64_t bytes, QEMUIOVector *qiov, + size_t qiov_offset); + + int (*bdrv_snapshot_create)(BlockDriverState *bs, + QEMUSnapshotInfo *sn_info); + int (*bdrv_snapshot_goto)(BlockDriverState *bs, + const char *snapshot_id); + int (*bdrv_snapshot_delete)(BlockDriverState *bs, + const char *snapshot_id, + const char *name, + Error **errp); + int (*bdrv_snapshot_list)(BlockDriverState *bs, + QEMUSnapshotInfo **psn_info); + int (*bdrv_snapshot_load_tmp)(BlockDriverState *bs, + const char *snapshot_id, + const char *name, + Error **errp); + int (*bdrv_get_info)(BlockDriverState *bs, BlockDriverInfo *bdi); + + ImageInfoSpecific *(*bdrv_get_specific_info)(BlockDriverState *bs, + Error **errp); + BlockStatsSpecific *(*bdrv_get_specific_stats)(BlockDriverState *bs); + + int coroutine_fn (*bdrv_save_vmstate)(BlockDriverState *bs, + QEMUIOVector *qiov, + int64_t pos); + int coroutine_fn (*bdrv_load_vmstate)(BlockDriverState *bs, + QEMUIOVector *qiov, + int64_t pos); + + int (*bdrv_change_backing_file)(BlockDriverState *bs, + const char *backing_file, const char *backing_fmt); + + /* removable device specific */ + bool (*bdrv_is_inserted)(BlockDriverState *bs); + void (*bdrv_eject)(BlockDriverState *bs, bool eject_flag); + void (*bdrv_lock_medium)(BlockDriverState *bs, bool locked); + + /* to control generic scsi devices */ + BlockAIOCB *(*bdrv_aio_ioctl)(BlockDriverState *bs, + unsigned long int req, void *buf, + BlockCompletionFunc *cb, void *opaque); + int coroutine_fn (*bdrv_co_ioctl)(BlockDriverState *bs, + unsigned long int req, void *buf); + + /* List of options for creating images, terminated by name == NULL */ + QemuOptsList *create_opts; + + /* List of options for image amend */ + QemuOptsList *amend_opts; + + /* + * If this driver supports reopening images this contains a + * NULL-terminated list of the runtime options that can be + * modified. If an option in this list is unspecified during + * reopen then it _must_ be reset to its default value or return + * an error. + */ + const char *const *mutable_opts; + + /* + * Returns 0 for completed check, -errno for internal errors. + * The check results are stored in result. + */ + int coroutine_fn (*bdrv_co_check)(BlockDriverState *bs, + BdrvCheckResult *result, + BdrvCheckMode fix); + + void (*bdrv_debug_event)(BlockDriverState *bs, BlkdebugEvent event); + + /* TODO Better pass a option string/QDict/QemuOpts to add any rule? */ + int (*bdrv_debug_breakpoint)(BlockDriverState *bs, const char *event, + const char *tag); + int (*bdrv_debug_remove_breakpoint)(BlockDriverState *bs, + const char *tag); + int (*bdrv_debug_resume)(BlockDriverState *bs, const char *tag); + bool (*bdrv_debug_is_suspended)(BlockDriverState *bs, const char *tag); + + void (*bdrv_refresh_limits)(BlockDriverState *bs, Error **errp); + + /* + * Returns 1 if newly created images are guaranteed to contain only + * zeros, 0 otherwise. + */ + int (*bdrv_has_zero_init)(BlockDriverState *bs); + + /* + * Remove fd handlers, timers, and other event loop callbacks so the event + * loop is no longer in use. Called with no in-flight requests and in + * depth-first traversal order with parents before child nodes. + */ + void (*bdrv_detach_aio_context)(BlockDriverState *bs); + + /* + * Add fd handlers, timers, and other event loop callbacks so I/O requests + * can be processed again. Called with no in-flight requests and in + * depth-first traversal order with child nodes before parent nodes. + */ + void (*bdrv_attach_aio_context)(BlockDriverState *bs, + AioContext *new_context); + + /* io queue for linux-aio */ + void (*bdrv_io_plug)(BlockDriverState *bs); + void (*bdrv_io_unplug)(BlockDriverState *bs); + + /** + * Try to get @bs's logical and physical block size. + * On success, store them in @bsz and return zero. + * On failure, return negative errno. + */ + int (*bdrv_probe_blocksizes)(BlockDriverState *bs, BlockSizes *bsz); + /** + * Try to get @bs's geometry (cyls, heads, sectors) + * On success, store them in @geo and return 0. + * On failure return -errno. + * Only drivers that want to override guest geometry implement this + * callback; see hd_geometry_guess(). + */ + int (*bdrv_probe_geometry)(BlockDriverState *bs, HDGeometry *geo); + + /** + * bdrv_co_drain_begin is called if implemented in the beginning of a + * drain operation to drain and stop any internal sources of requests in + * the driver. + * bdrv_co_drain_end is called if implemented at the end of the drain. + * + * They should be used by the driver to e.g. manage scheduled I/O + * requests, or toggle an internal state. After the end of the drain new + * requests will continue normally. + */ + void coroutine_fn (*bdrv_co_drain_begin)(BlockDriverState *bs); + void coroutine_fn (*bdrv_co_drain_end)(BlockDriverState *bs); + + void (*bdrv_add_child)(BlockDriverState *parent, BlockDriverState *child, + Error **errp); + void (*bdrv_del_child)(BlockDriverState *parent, BdrvChild *child, + Error **errp); + + /** + * Informs the block driver that a permission change is intended. The + * driver checks whether the change is permissible and may take other + * preparations for the change (e.g. get file system locks). This operation + * is always followed either by a call to either .bdrv_set_perm or + * .bdrv_abort_perm_update. + * + * Checks whether the requested set of cumulative permissions in @perm + * can be granted for accessing @bs and whether no other users are using + * permissions other than those given in @shared (both arguments take + * BLK_PERM_* bitmasks). + * + * If both conditions are met, 0 is returned. Otherwise, -errno is returned + * and errp is set to an error describing the conflict. + */ + int (*bdrv_check_perm)(BlockDriverState *bs, uint64_t perm, + uint64_t shared, Error **errp); + + /** + * Called to inform the driver that the set of cumulative set of used + * permissions for @bs has changed to @perm, and the set of sharable + * permission to @shared. The driver can use this to propagate changes to + * its children (i.e. request permissions only if a parent actually needs + * them). + * + * This function is only invoked after bdrv_check_perm(), so block drivers + * may rely on preparations made in their .bdrv_check_perm implementation. + */ + void (*bdrv_set_perm)(BlockDriverState *bs, uint64_t perm, uint64_t shared); + + /* + * Called to inform the driver that after a previous bdrv_check_perm() + * call, the permission update is not performed and any preparations made + * for it (e.g. taken file locks) need to be undone. + * + * This function can be called even for nodes that never saw a + * bdrv_check_perm() call. It is a no-op then. + */ + void (*bdrv_abort_perm_update)(BlockDriverState *bs); + + /** + * Returns in @nperm and @nshared the permissions that the driver for @bs + * needs on its child @c, based on the cumulative permissions requested by + * the parents in @parent_perm and @parent_shared. + * + * If @c is NULL, return the permissions for attaching a new child for the + * given @child_class and @role. + * + * If @reopen_queue is non-NULL, don't return the currently needed + * permissions, but those that will be needed after applying the + * @reopen_queue. + */ + void (*bdrv_child_perm)(BlockDriverState *bs, BdrvChild *c, + BdrvChildRole role, + BlockReopenQueue *reopen_queue, + uint64_t parent_perm, uint64_t parent_shared, + uint64_t *nperm, uint64_t *nshared); + + bool (*bdrv_supports_persistent_dirty_bitmap)(BlockDriverState *bs); + bool (*bdrv_co_can_store_new_dirty_bitmap)(BlockDriverState *bs, + const char *name, + uint32_t granularity, + Error **errp); + int (*bdrv_co_remove_persistent_dirty_bitmap)(BlockDriverState *bs, + const char *name, + Error **errp); + + /** + * Register/unregister a buffer for I/O. For example, when the driver is + * interested to know the memory areas that will later be used in iovs, so + * that it can do IOMMU mapping with VFIO etc., in order to get better + * performance. In the case of VFIO drivers, this callback is used to do + * DMA mapping for hot buffers. + */ + void (*bdrv_register_buf)(BlockDriverState *bs, void *host, size_t size); + void (*bdrv_unregister_buf)(BlockDriverState *bs, void *host); + QLIST_ENTRY(BlockDriver) list; + + /* + * Pointer to a NULL-terminated array of names of strong options + * that can be specified for bdrv_open(). A strong option is one + * that changes the data of a BDS. + * If this pointer is NULL, the array is considered empty. + * "filename" and "driver" are always considered strong. + */ + const char *const *strong_runtime_opts; +}; + +static inline bool block_driver_can_compress(BlockDriver *drv) +{ + return drv->bdrv_co_pwritev_compressed || + drv->bdrv_co_pwritev_compressed_part; +} + +typedef struct BlockLimits { + /* + * Alignment requirement, in bytes, for offset/length of I/O + * requests. Must be a power of 2 less than INT_MAX; defaults to + * 1 for drivers with modern byte interfaces, and to 512 + * otherwise. + */ + uint32_t request_alignment; + + /* + * Maximum number of bytes that can be discarded at once. Must be multiple + * of pdiscard_alignment, but need not be power of 2. May be 0 if no + * inherent 64-bit limit. + */ + int64_t max_pdiscard; + + /* + * Optimal alignment for discard requests in bytes. A power of 2 + * is best but not mandatory. Must be a multiple of + * bl.request_alignment, and must be less than max_pdiscard if + * that is set. May be 0 if bl.request_alignment is good enough + */ + uint32_t pdiscard_alignment; + + /* + * Maximum number of bytes that can zeroized at once. Must be multiple of + * pwrite_zeroes_alignment. 0 means no limit. + */ + int64_t max_pwrite_zeroes; + + /* + * Optimal alignment for write zeroes requests in bytes. A power + * of 2 is best but not mandatory. Must be a multiple of + * bl.request_alignment, and must be less than max_pwrite_zeroes + * if that is set. May be 0 if bl.request_alignment is good + * enough + */ + uint32_t pwrite_zeroes_alignment; + + /* + * Optimal transfer length in bytes. A power of 2 is best but not + * mandatory. Must be a multiple of bl.request_alignment, or 0 if + * no preferred size + */ + uint32_t opt_transfer; + + /* + * Maximal transfer length in bytes. Need not be power of 2, but + * must be multiple of opt_transfer and bl.request_alignment, or 0 + * for no 32-bit limit. For now, anything larger than INT_MAX is + * clamped down. + */ + uint32_t max_transfer; + + /* + * Maximal hardware transfer length in bytes. Applies whenever + * transfers to the device bypass the kernel I/O scheduler, for + * example with SG_IO. If larger than max_transfer or if zero, + * blk_get_max_hw_transfer will fall back to max_transfer. + */ + uint64_t max_hw_transfer; + + /* + * Maximal number of scatter/gather elements allowed by the hardware. + * Applies whenever transfers to the device bypass the kernel I/O + * scheduler, for example with SG_IO. If larger than max_iov + * or if zero, blk_get_max_hw_iov will fall back to max_iov. + */ + int max_hw_iov; + + + /* memory alignment, in bytes so that no bounce buffer is needed */ + size_t min_mem_alignment; + + /* memory alignment, in bytes, for bounce buffer */ + size_t opt_mem_alignment; + + /* maximum number of iovec elements */ + int max_iov; +} BlockLimits; + +typedef struct BdrvOpBlocker BdrvOpBlocker; + +typedef struct BdrvAioNotifier { + void (*attached_aio_context)(AioContext *new_context, void *opaque); + void (*detach_aio_context)(void *opaque); + + void *opaque; + bool deleted; + + QLIST_ENTRY(BdrvAioNotifier) list; +} BdrvAioNotifier; + +struct BdrvChildClass { + /* + * If true, bdrv_replace_node() doesn't change the node this BdrvChild + * points to. + */ + bool stay_at_node; + + /* + * If true, the parent is a BlockDriverState and bdrv_next_all_states() + * will return it. This information is used for drain_all, where every node + * will be drained separately, so the drain only needs to be propagated to + * non-BDS parents. + */ + bool parent_is_bds; + + void (*inherit_options)(BdrvChildRole role, bool parent_is_format, + int *child_flags, QDict *child_options, + int parent_flags, QDict *parent_options); + + void (*change_media)(BdrvChild *child, bool load); + void (*resize)(BdrvChild *child); + + /* + * Returns a name that is supposedly more useful for human users than the + * node name for identifying the node in question (in particular, a BB + * name), or NULL if the parent can't provide a better name. + */ + const char *(*get_name)(BdrvChild *child); + + /* + * Returns a malloced string that describes the parent of the child for a + * human reader. This could be a node-name, BlockBackend name, qdev ID or + * QOM path of the device owning the BlockBackend, job type and ID etc. The + * caller is responsible for freeing the memory. + */ + char *(*get_parent_desc)(BdrvChild *child); + + /* + * If this pair of functions is implemented, the parent doesn't issue new + * requests after returning from .drained_begin() until .drained_end() is + * called. + * + * These functions must not change the graph (and therefore also must not + * call aio_poll(), which could change the graph indirectly). + * + * If drained_end() schedules background operations, it must atomically + * increment *drained_end_counter for each such operation and atomically + * decrement it once the operation has settled. + * + * Note that this can be nested. If drained_begin() was called twice, new + * I/O is allowed only after drained_end() was called twice, too. + */ + void (*drained_begin)(BdrvChild *child); + void (*drained_end)(BdrvChild *child, int *drained_end_counter); + + /* + * Returns whether the parent has pending requests for the child. This + * callback is polled after .drained_begin() has been called until all + * activity on the child has stopped. + */ + bool (*drained_poll)(BdrvChild *child); + + /* + * Notifies the parent that the child has been activated/inactivated (e.g. + * when migration is completing) and it can start/stop requesting + * permissions and doing I/O on it. + */ + void (*activate)(BdrvChild *child, Error **errp); + int (*inactivate)(BdrvChild *child); + + void (*attach)(BdrvChild *child); + void (*detach)(BdrvChild *child); + + /* + * Notifies the parent that the filename of its child has changed (e.g. + * because the direct child was removed from the backing chain), so that it + * can update its reference. + */ + int (*update_filename)(BdrvChild *child, BlockDriverState *new_base, + const char *filename, Error **errp); + + bool (*can_set_aio_ctx)(BdrvChild *child, AioContext *ctx, + GSList **ignore, Error **errp); + void (*set_aio_ctx)(BdrvChild *child, AioContext *ctx, GSList **ignore); + + AioContext *(*get_parent_aio_context)(BdrvChild *child); +}; + +extern const BdrvChildClass child_of_bds; + +struct BdrvChild { + BlockDriverState *bs; + char *name; + const BdrvChildClass *klass; + BdrvChildRole role; + void *opaque; + + /** + * Granted permissions for operating on this BdrvChild (BLK_PERM_* bitmask) + */ + uint64_t perm; + + /** + * Permissions that can still be granted to other users of @bs while this + * BdrvChild is still attached to it. (BLK_PERM_* bitmask) + */ + uint64_t shared_perm; + + /* + * This link is frozen: the child can neither be replaced nor + * detached from the parent. + */ + bool frozen; + + /* + * How many times the parent of this child has been drained + * (through klass->drained_*). + * Usually, this is equal to bs->quiesce_counter (potentially + * reduced by bdrv_drain_all_count). It may differ while the + * child is entering or leaving a drained section. + */ + int parent_quiesce_counter; + + QLIST_ENTRY(BdrvChild) next; + QLIST_ENTRY(BdrvChild) next_parent; +}; + +/* + * Allows bdrv_co_block_status() to cache one data region for a + * protocol node. + * + * @valid: Whether the cache is valid (should be accessed with atomic + * functions so this can be reset by RCU readers) + * @data_start: Offset where we know (or strongly assume) is data + * @data_end: Offset where the data region ends (which is not necessarily + * the start of a zeroed region) + */ +typedef struct BdrvBlockStatusCache { + struct rcu_head rcu; + + bool valid; + int64_t data_start; + int64_t data_end; +} BdrvBlockStatusCache; + +struct BlockDriverState { + /* + * Protected by big QEMU lock or read-only after opening. No special + * locking needed during I/O... + */ + int open_flags; /* flags used to open the file, re-used for re-open */ + bool encrypted; /* if true, the media is encrypted */ + bool sg; /* if true, the device is a /dev/sg* */ + bool probed; /* if true, format was probed rather than specified */ + bool force_share; /* if true, always allow all shared permissions */ + bool implicit; /* if true, this filter node was automatically inserted */ + + BlockDriver *drv; /* NULL means no media */ + void *opaque; + + AioContext *aio_context; /* event loop used for fd handlers, timers, etc */ + /* + * long-running tasks intended to always use the same AioContext as this + * BDS may register themselves in this list to be notified of changes + * regarding this BDS's context + */ + QLIST_HEAD(, BdrvAioNotifier) aio_notifiers; + bool walking_aio_notifiers; /* to make removal during iteration safe */ + + char filename[PATH_MAX]; + /* + * If not empty, this image is a diff in relation to backing_file. + * Note that this is the name given in the image header and + * therefore may or may not be equal to .backing->bs->filename. + * If this field contains a relative path, it is to be resolved + * relatively to the overlay's location. + */ + char backing_file[PATH_MAX]; + /* + * The backing filename indicated by the image header. Contrary + * to backing_file, if we ever open this file, auto_backing_file + * is replaced by the resulting BDS's filename (i.e. after a + * bdrv_refresh_filename() run). + */ + char auto_backing_file[PATH_MAX]; + char backing_format[16]; /* if non-zero and backing_file exists */ + + QDict *full_open_options; + char exact_filename[PATH_MAX]; + + BdrvChild *backing; + BdrvChild *file; + + /* I/O Limits */ + BlockLimits bl; + + /* + * Flags honored during pread + */ + unsigned int supported_read_flags; + /* + * Flags honored during pwrite (so far: BDRV_REQ_FUA, + * BDRV_REQ_WRITE_UNCHANGED). + * If a driver does not support BDRV_REQ_WRITE_UNCHANGED, those + * writes will be issued as normal writes without the flag set. + * This is important to note for drivers that do not explicitly + * request a WRITE permission for their children and instead take + * the same permissions as their parent did (this is commonly what + * block filters do). Such drivers have to be aware that the + * parent may have taken a WRITE_UNCHANGED permission only and is + * issuing such requests. Drivers either must make sure that + * these requests do not result in plain WRITE accesses (usually + * by supporting BDRV_REQ_WRITE_UNCHANGED, and then forwarding + * every incoming write request as-is, including potentially that + * flag), or they have to explicitly take the WRITE permission for + * their children. + */ + unsigned int supported_write_flags; + /* + * 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]; + /* element of the list of named nodes building the graph */ + QTAILQ_ENTRY(BlockDriverState) node_list; + /* element of the list of all BlockDriverStates (all_bdrv_states) */ + QTAILQ_ENTRY(BlockDriverState) bs_list; + /* element of the list of monitor-owned BDS */ + QTAILQ_ENTRY(BlockDriverState) monitor_list; + int refcnt; + + /* operation blockers. Protected by BQL. */ + QLIST_HEAD(, BdrvOpBlocker) op_blockers[BLOCK_OP_TYPE_MAX]; + + /* + * The node that this node inherited default options from (and a reopen on + * which can affect this node by changing these defaults). This is always a + * parent node of this node. + */ + BlockDriverState *inherits_from; + QLIST_HEAD(, BdrvChild) children; + QLIST_HEAD(, BdrvChild) parents; + + QDict *options; + QDict *explicit_options; + BlockdevDetectZeroesOptions detect_zeroes; + + /* The error object in use for blocking operations on backing_hd */ + Error *backing_blocker; + + /* Protected by AioContext lock */ + + /* + * If we are reading a disk image, give its size in sectors. + * Generally read-only; it is written to by load_snapshot and + * save_snaphost, but the block layer is quiescent during those. + */ + int64_t total_sectors; + + /* threshold limit for writes, in bytes. "High water mark". */ + uint64_t write_threshold_offset; + + /* + * Writing to the list requires the BQL _and_ the dirty_bitmap_mutex. + * Reading from the list can be done with either the BQL or the + * dirty_bitmap_mutex. Modifying a bitmap only requires + * dirty_bitmap_mutex. + */ + QemuMutex dirty_bitmap_mutex; + QLIST_HEAD(, BdrvDirtyBitmap) dirty_bitmaps; + + /* Offset after the highest byte written to */ + Stat64 wr_highest_offset; + + /* + * If true, copy read backing sectors into image. Can be >1 if more + * than one client has requested copy-on-read. Accessed with atomic + * ops. + */ + int copy_on_read; + + /* + * number of in-flight requests; overall and serialising. + * Accessed with atomic ops. + */ + unsigned int in_flight; + unsigned int serialising_in_flight; + + /* + * counter for nested bdrv_io_plug. + * Accessed with atomic ops. + */ + unsigned io_plugged; + + /* do we need to tell the quest if we have a volatile write cache? */ + int enable_write_cache; + + /* Accessed with atomic ops. */ + int quiesce_counter; + int recursive_quiesce_counter; + + unsigned int write_gen; /* Current data generation */ + + /* Protected by reqs_lock. */ + CoMutex reqs_lock; + QLIST_HEAD(, BdrvTrackedRequest) tracked_requests; + CoQueue flush_queue; /* Serializing flush queue */ + bool active_flush_req; /* Flush request in flight? */ + + /* Only read/written by whoever has set active_flush_req to true. */ + unsigned int flushed_gen; /* Flushed write generation */ + + /* BdrvChild links to this node may never be frozen */ + bool never_freeze; + + /* Lock for block-status cache RCU writers */ + CoMutex bsc_modify_lock; + /* Always non-NULL, but must only be dereferenced under an RCU read guard */ + BdrvBlockStatusCache *block_status_cache; +}; + +struct BlockBackendRootState { + int open_flags; + BlockdevDetectZeroesOptions detect_zeroes; +}; + +typedef enum BlockMirrorBackingMode { + /* + * Reuse the existing backing chain from the source for the target. + * - sync=full: Set backing BDS to NULL. + * - sync=top: Use source's backing BDS. + * - sync=none: Use source as the backing BDS. + */ + MIRROR_SOURCE_BACKING_CHAIN, + + /* Open the target's backing chain completely anew */ + MIRROR_OPEN_BACKING_CHAIN, + + /* Do not change the target's backing BDS after job completion */ + MIRROR_LEAVE_BACKING_CHAIN, +} BlockMirrorBackingMode; + + +/* + * Essential block drivers which must always be statically linked into qemu, and + * which therefore can be accessed without using bdrv_find_format() + */ +extern BlockDriver bdrv_file; +extern BlockDriver bdrv_raw; +extern BlockDriver bdrv_qcow2; + +extern unsigned int bdrv_drain_all_count; +extern QemuOptsList bdrv_create_opts_simple; + +/* + * Common functions that are neither I/O nor Global State. + * + * See include/block/block-commmon.h for more information about + * the Common API. + */ + +static inline BlockDriverState *child_bs(BdrvChild *child) +{ + return child ? child->bs : NULL; +} + +int bdrv_check_request(int64_t offset, int64_t bytes, Error **errp); +int get_tmp_filename(char *filename, int size); +void bdrv_parse_filename_strip_prefix(const char *filename, const char *prefix, + QDict *options); + + +int bdrv_check_qiov_request(int64_t offset, int64_t bytes, + QEMUIOVector *qiov, size_t qiov_offset, + Error **errp); + +#ifdef _WIN32 +int is_windows_drive(const char *filename); +#endif + +#endif /* BLOCK_INT_COMMON_H */ diff --git a/include/block/block_int-global-state.h b/include/block/block_int-global-state.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/block/block_int-global-state.h @@ -XXX,XX +XXX,XX @@ +/* + * QEMU System Emulator block driver + * + * Copyright (c) 2003 Fabrice Bellard + * + * 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. + */ +#ifndef BLOCK_INT_GLOBAL_STATE_H +#define BLOCK_INT_GLOBAL_STATE_H + +#include "block_int-common.h" + +/* + * Global state (GS) API. These functions run under the BQL. + * + * See include/block/block-global-state.h for more information about + * the GS API. + */ + +/** + * stream_start: + * @job_id: The id of the newly-created job, or %NULL to use the + * device name of @bs. + * @bs: Block device to operate on. + * @base: Block device that will become the new base, or %NULL to + * flatten the whole backing file chain onto @bs. + * @backing_file_str: The file name that will be written to @bs as the + * the new backing file if the job completes. Ignored if @base is %NULL. + * @creation_flags: Flags that control the behavior of the Job lifetime. + * See @BlockJobCreateFlags + * @speed: The maximum speed, in bytes per second, or 0 for unlimited. + * @on_error: The action to take upon error. + * @filter_node_name: The node name that should be assigned to the filter + * driver that the stream job inserts into the graph above + * @bs. NULL means that a node name should be autogenerated. + * @errp: Error object. + * + * Start a streaming operation on @bs. Clusters that are unallocated + * in @bs, but allocated in any image between @base and @bs (both + * exclusive) will be written to @bs. At the end of a successful + * streaming job, the backing file of @bs will be changed to + * @backing_file_str in the written image and to @base in the live + * BlockDriverState. + */ +void stream_start(const char *job_id, BlockDriverState *bs, + BlockDriverState *base, const char *backing_file_str, + BlockDriverState *bottom, + int creation_flags, int64_t speed, + BlockdevOnError on_error, + const char *filter_node_name, + Error **errp); + +/** + * commit_start: + * @job_id: The id of the newly-created job, or %NULL to use the + * device name of @bs. + * @bs: Active block device. + * @top: Top block device to be committed. + * @base: Block device that will be written into, and become the new top. + * @creation_flags: Flags that control the behavior of the Job lifetime. + * See @BlockJobCreateFlags + * @speed: The maximum speed, in bytes per second, or 0 for unlimited. + * @on_error: The action to take upon error. + * @backing_file_str: String to use as the backing file in @top's overlay + * @filter_node_name: The node name that should be assigned to the filter + * driver that the commit job inserts into the graph above @top. NULL means + * that a node name should be autogenerated. + * @errp: Error object. + * + */ +void commit_start(const char *job_id, BlockDriverState *bs, + BlockDriverState *base, BlockDriverState *top, + int creation_flags, int64_t speed, + BlockdevOnError on_error, const char *backing_file_str, + const char *filter_node_name, Error **errp); +/** + * commit_active_start: + * @job_id: The id of the newly-created job, or %NULL to use the + * device name of @bs. + * @bs: Active block device to be committed. + * @base: Block device that will be written into, and become the new top. + * @creation_flags: Flags that control the behavior of the Job lifetime. + * See @BlockJobCreateFlags + * @speed: The maximum speed, in bytes per second, or 0 for unlimited. + * @on_error: The action to take upon error. + * @filter_node_name: The node name that should be assigned to the filter + * driver that the commit job inserts into the graph above @bs. NULL means that + * a node name should be autogenerated. + * @cb: Completion function for the job. + * @opaque: Opaque pointer value passed to @cb. + * @auto_complete: Auto complete the job. + * @errp: Error object. + * + */ +BlockJob *commit_active_start(const char *job_id, BlockDriverState *bs, + BlockDriverState *base, int creation_flags, + int64_t speed, BlockdevOnError on_error, + const char *filter_node_name, + BlockCompletionFunc *cb, void *opaque, + bool auto_complete, Error **errp); +/* + * mirror_start: + * @job_id: The id of the newly-created job, or %NULL to use the + * device name of @bs. + * @bs: Block device to operate on. + * @target: Block device to write to. + * @replaces: Block graph node name to replace once the mirror is done. Can + * only be used when full mirroring is selected. + * @creation_flags: Flags that control the behavior of the Job lifetime. + * See @BlockJobCreateFlags + * @speed: The maximum speed, in bytes per second, or 0 for unlimited. + * @granularity: The chosen granularity for the dirty bitmap. + * @buf_size: The amount of data that can be in flight at one time. + * @mode: Whether to collapse all images in the chain to the target. + * @backing_mode: How to establish the target's backing chain after completion. + * @zero_target: Whether the target should be explicitly zero-initialized + * @on_source_error: The action to take upon error reading from the source. + * @on_target_error: The action to take upon error writing to the target. + * @unmap: Whether to unmap target where source sectors only contain zeroes. + * @filter_node_name: The node name that should be assigned to the filter + * driver that the mirror job inserts into the graph above @bs. NULL means that + * a node name should be autogenerated. + * @copy_mode: When to trigger writes to the target. + * @errp: Error object. + * + * Start a mirroring operation on @bs. Clusters that are allocated + * in @bs will be written to @target until the job is cancelled or + * manually completed. At the end of a successful mirroring job, + * @bs will be switched to read from @target. + */ +void mirror_start(const char *job_id, BlockDriverState *bs, + BlockDriverState *target, const char *replaces, + int creation_flags, int64_t speed, + uint32_t granularity, int64_t buf_size, + MirrorSyncMode mode, BlockMirrorBackingMode backing_mode, + bool zero_target, + BlockdevOnError on_source_error, + BlockdevOnError on_target_error, + bool unmap, const char *filter_node_name, + MirrorCopyMode copy_mode, Error **errp); + +/* + * backup_job_create: + * @job_id: The id of the newly-created job, or %NULL to use the + * device name of @bs. + * @bs: Block device to operate on. + * @target: Block device to write to. + * @speed: The maximum speed, in bytes per second, or 0 for unlimited. + * @sync_mode: What parts of the disk image should be copied to the destination. + * @sync_bitmap: The dirty bitmap if sync_mode is 'bitmap' or 'incremental' + * @bitmap_mode: The bitmap synchronization policy to use. + * @perf: Performance options. All actual fields assumed to be present, + * all ".has_*" fields are ignored. + * @on_source_error: The action to take upon error reading from the source. + * @on_target_error: The action to take upon error writing to the target. + * @creation_flags: Flags that control the behavior of the Job lifetime. + * See @BlockJobCreateFlags + * @cb: Completion function for the job. + * @opaque: Opaque pointer value passed to @cb. + * @txn: Transaction that this job is part of (may be NULL). + * + * Create a backup operation on @bs. Clusters in @bs are written to @target + * until the job is cancelled or manually completed. + */ +BlockJob *backup_job_create(const char *job_id, BlockDriverState *bs, + BlockDriverState *target, int64_t speed, + MirrorSyncMode sync_mode, + BdrvDirtyBitmap *sync_bitmap, + BitmapSyncMode bitmap_mode, + bool compress, + const char *filter_node_name, + BackupPerf *perf, + BlockdevOnError on_source_error, + BlockdevOnError on_target_error, + int creation_flags, + BlockCompletionFunc *cb, void *opaque, + JobTxn *txn, Error **errp); + +BdrvChild *bdrv_root_attach_child(BlockDriverState *child_bs, + const char *child_name, + const BdrvChildClass *child_class, + BdrvChildRole child_role, + uint64_t perm, uint64_t shared_perm, + void *opaque, Error **errp); +void bdrv_root_unref_child(BdrvChild *child); + +void bdrv_get_cumulative_perm(BlockDriverState *bs, uint64_t *perm, + uint64_t *shared_perm); + +/** + * Sets a BdrvChild's permissions. Avoid if the parent is a BDS; use + * bdrv_child_refresh_perms() instead and make the parent's + * .bdrv_child_perm() implementation return the correct values. + */ +int bdrv_child_try_set_perm(BdrvChild *c, uint64_t perm, uint64_t shared, + Error **errp); + +/** + * Calls bs->drv->bdrv_child_perm() and updates the child's permission + * masks with the result. + * Drivers should invoke this function whenever an event occurs that + * makes their .bdrv_child_perm() implementation return different + * values than before, but which will not result in the block layer + * automatically refreshing the permissions. + */ +int bdrv_child_refresh_perms(BlockDriverState *bs, BdrvChild *c, Error **errp); + +bool bdrv_recurse_can_replace(BlockDriverState *bs, + BlockDriverState *to_replace); + +/* + * Default implementation for BlockDriver.bdrv_child_perm() that can + * be used by block filters and image formats, as long as they use the + * child_of_bds child class and set an appropriate BdrvChildRole. + */ +void bdrv_default_perms(BlockDriverState *bs, BdrvChild *c, + BdrvChildRole role, BlockReopenQueue *reopen_queue, + uint64_t perm, uint64_t shared, + uint64_t *nperm, uint64_t *nshared); + +void blk_dev_change_media_cb(BlockBackend *blk, bool load, Error **errp); +bool blk_dev_has_removable_media(BlockBackend *blk); +void blk_dev_eject_request(BlockBackend *blk, bool force); +bool blk_dev_is_medium_locked(BlockBackend *blk); + +void bdrv_restore_dirty_bitmap(BdrvDirtyBitmap *bitmap, HBitmap *backup); + +void bdrv_set_monitor_owned(BlockDriverState *bs); + +void blockdev_close_all_bdrv_states(void); + +BlockDriverState *bds_tree_init(QDict *bs_opts, Error **errp); + +/** + * Simple implementation of bdrv_co_create_opts for protocol drivers + * which only support creation via opening a file + * (usually existing raw storage device) + */ +int coroutine_fn bdrv_co_create_opts_simple(BlockDriver *drv, + const char *filename, + QemuOpts *opts, + Error **errp); + +BdrvDirtyBitmap *block_dirty_bitmap_lookup(const char *node, + const char *name, + BlockDriverState **pbs, + Error **errp); +BdrvDirtyBitmap *block_dirty_bitmap_merge(const char *node, const char *target, + BlockDirtyBitmapMergeSourceList *bms, + HBitmap **backup, Error **errp); +BdrvDirtyBitmap *block_dirty_bitmap_remove(const char *node, const char *name, + bool release, + BlockDriverState **bitmap_bs, + Error **errp); + + +BlockDriverState *bdrv_skip_implicit_filters(BlockDriverState *bs); + +/** + * bdrv_add_aio_context_notifier: + * + * If a long-running job intends to be always run in the same AioContext as a + * certain BDS, it may use this function to be notified of changes regarding the + * association of the BDS to an AioContext. + * + * attached_aio_context() is called after the target BDS has been attached to a + * new AioContext; detach_aio_context() is called before the target BDS is being + * detached from its old AioContext. + */ +void bdrv_add_aio_context_notifier(BlockDriverState *bs, + void (*attached_aio_context)(AioContext *new_context, void *opaque), + void (*detach_aio_context)(void *opaque), void *opaque); + +/** + * bdrv_remove_aio_context_notifier: + * + * Unsubscribe of change notifications regarding the BDS's AioContext. The + * parameters given here have to be the same as those given to + * bdrv_add_aio_context_notifier(). + */ +void bdrv_remove_aio_context_notifier(BlockDriverState *bs, + void (*aio_context_attached)(AioContext *, + void *), + void (*aio_context_detached)(void *), + void *opaque); + +/** + * End all quiescent sections started by bdrv_drain_all_begin(). This is + * needed when deleting a BDS before bdrv_drain_all_end() is called. + * + * NOTE: this is an internal helper for bdrv_close() *only*. No one else + * should call it. + */ +void bdrv_drain_all_end_quiesce(BlockDriverState *bs); + +#endif /* BLOCK_INT_GLOBAL_STATE */ diff --git a/include/block/block_int-io.h b/include/block/block_int-io.h new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/include/block/block_int-io.h @@ -XXX,XX +XXX,XX @@ +/* + * QEMU System Emulator block driver + * + * Copyright (c) 2003 Fabrice Bellard + * + * 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. + */ +#ifndef BLOCK_INT_IO_H +#define BLOCK_INT_IO_H + +#include "block_int-common.h" + +/* + * I/O API functions. These functions are thread-safe. + * + * See include/block/block-io.h for more information about + * the I/O API. + */ + +int coroutine_fn bdrv_co_preadv(BdrvChild *child, + int64_t offset, int64_t bytes, QEMUIOVector *qiov, + BdrvRequestFlags flags); +int coroutine_fn bdrv_co_preadv_part(BdrvChild *child, + int64_t offset, int64_t bytes, + QEMUIOVector *qiov, size_t qiov_offset, BdrvRequestFlags flags); +int coroutine_fn bdrv_co_pwritev(BdrvChild *child, + int64_t offset, int64_t bytes, QEMUIOVector *qiov, + BdrvRequestFlags flags); +int coroutine_fn bdrv_co_pwritev_part(BdrvChild *child, + int64_t offset, int64_t bytes, + QEMUIOVector *qiov, size_t qiov_offset, BdrvRequestFlags flags); + +static inline int coroutine_fn bdrv_co_pread(BdrvChild *child, + int64_t offset, unsigned int bytes, void *buf, BdrvRequestFlags flags) +{ + QEMUIOVector qiov = QEMU_IOVEC_INIT_BUF(qiov, buf, bytes); + + return bdrv_co_preadv(child, offset, bytes, &qiov, flags); +} + +static inline int coroutine_fn bdrv_co_pwrite(BdrvChild *child, + int64_t offset, unsigned int bytes, void *buf, BdrvRequestFlags flags) +{ + QEMUIOVector qiov = QEMU_IOVEC_INIT_BUF(qiov, buf, bytes); + + return bdrv_co_pwritev(child, offset, bytes, &qiov, flags); +} + +bool coroutine_fn bdrv_make_request_serialising(BdrvTrackedRequest *req, + uint64_t align); +BdrvTrackedRequest *coroutine_fn bdrv_co_get_self_request(BlockDriverState *bs); + +BlockDriver *bdrv_probe_all(const uint8_t *buf, int buf_size, + const char *filename); + +/** + * bdrv_wakeup: + * @bs: The BlockDriverState for which an I/O operation has been completed. + * + * Wake up the main thread if it is waiting on BDRV_POLL_WHILE. During + * synchronous I/O on a BlockDriverState that is attached to another + * I/O thread, the main thread lets the I/O thread's event loop run, + * waiting for the I/O operation to complete. A bdrv_wakeup will wake + * up the main thread if necessary. + * + * Manual calls to bdrv_wakeup are rarely necessary, because + * bdrv_dec_in_flight already calls it. + */ +void bdrv_wakeup(BlockDriverState *bs); + +const char *bdrv_get_parent_name(const BlockDriverState *bs); +bool blk_dev_has_tray(BlockBackend *blk); +bool blk_dev_is_tray_open(BlockBackend *blk); + +void bdrv_set_dirty(BlockDriverState *bs, int64_t offset, int64_t bytes); + +void bdrv_clear_dirty_bitmap(BdrvDirtyBitmap *bitmap, HBitmap **out); +bool bdrv_dirty_bitmap_merge_internal(BdrvDirtyBitmap *dest, + const BdrvDirtyBitmap *src, + HBitmap **backup, bool lock); + +void bdrv_inc_in_flight(BlockDriverState *bs); +void bdrv_dec_in_flight(BlockDriverState *bs); + +int coroutine_fn bdrv_co_copy_range_from(BdrvChild *src, int64_t src_offset, + BdrvChild *dst, int64_t dst_offset, + int64_t bytes, + BdrvRequestFlags read_flags, + BdrvRequestFlags write_flags); +int coroutine_fn bdrv_co_copy_range_to(BdrvChild *src, int64_t src_offset, + BdrvChild *dst, int64_t dst_offset, + int64_t bytes, + BdrvRequestFlags read_flags, + BdrvRequestFlags write_flags); + +int refresh_total_sectors(BlockDriverState *bs, int64_t hint); + +BdrvChild *bdrv_cow_child(BlockDriverState *bs); +BdrvChild *bdrv_filter_child(BlockDriverState *bs); +BdrvChild *bdrv_filter_or_cow_child(BlockDriverState *bs); +BdrvChild *bdrv_primary_child(BlockDriverState *bs); +BlockDriverState *bdrv_skip_filters(BlockDriverState *bs); +BlockDriverState *bdrv_backing_chain_next(BlockDriverState *bs); + +static inline BlockDriverState *bdrv_cow_bs(BlockDriverState *bs) +{ + return child_bs(bdrv_cow_child(bs)); +} + +static inline BlockDriverState *bdrv_filter_bs(BlockDriverState *bs) +{ + return child_bs(bdrv_filter_child(bs)); +} + +static inline BlockDriverState *bdrv_filter_or_cow_bs(BlockDriverState *bs) +{ + return child_bs(bdrv_filter_or_cow_child(bs)); +} + +static inline BlockDriverState *bdrv_primary_bs(BlockDriverState *bs) +{ + return child_bs(bdrv_primary_child(bs)); +} + +/** + * Check whether the given offset is in the cached block-status data + * region. + * + * If it is, and @pnum is not NULL, *pnum is set to + * `bsc.data_end - offset`, i.e. how many bytes, starting from + * @offset, are data (according to the cache). + * Otherwise, *pnum is not touched. + */ +bool bdrv_bsc_is_data(BlockDriverState *bs, int64_t offset, int64_t *pnum); + +/** + * If [offset, offset + bytes) overlaps with the currently cached + * block-status region, invalidate the cache. + * + * (To be used by I/O paths that cause data regions to be zero or + * holes.) + */ +void bdrv_bsc_invalidate_range(BlockDriverState *bs, + int64_t offset, int64_t bytes); + +/** + * Mark the range [offset, offset + bytes) as a data region. + */ +void bdrv_bsc_fill(BlockDriverState *bs, int64_t offset, int64_t bytes); + + +/* + * "I/O or GS" API functions. These functions can run without + * the BQL, but only in one specific iothread/main loop. + * + * See include/block/block-io.h for more information about + * the "I/O or GS" API. + */ + +void bdrv_apply_subtree_drain(BdrvChild *child, BlockDriverState *new_parent); +void bdrv_unapply_subtree_drain(BdrvChild *child, BlockDriverState *old_parent); + +#endif /* BLOCK_INT_IO_H */ 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 @@ #ifndef BLOCK_INT_H #define BLOCK_INT_H -#include "block/accounting.h" -#include "block/block.h" -#include "block/aio-wait.h" -#include "qemu/queue.h" -#include "qemu/coroutine.h" -#include "qemu/stats64.h" -#include "qemu/timer.h" -#include "qemu/hbitmap.h" -#include "block/snapshot.h" -#include "qemu/throttle.h" -#include "qemu/rcu.h" +#include "block_int-global-state.h" +#include "block_int-io.h" -#define BLOCK_FLAG_LAZY_REFCOUNTS 8 - -#define BLOCK_OPT_SIZE "size" -#define BLOCK_OPT_ENCRYPT "encryption" -#define BLOCK_OPT_ENCRYPT_FORMAT "encrypt.format" -#define BLOCK_OPT_COMPAT6 "compat6" -#define BLOCK_OPT_HWVERSION "hwversion" -#define BLOCK_OPT_BACKING_FILE "backing_file" -#define BLOCK_OPT_BACKING_FMT "backing_fmt" -#define BLOCK_OPT_CLUSTER_SIZE "cluster_size" -#define BLOCK_OPT_TABLE_SIZE "table_size" -#define BLOCK_OPT_PREALLOC "preallocation" -#define BLOCK_OPT_SUBFMT "subformat" -#define BLOCK_OPT_COMPAT_LEVEL "compat" -#define BLOCK_OPT_LAZY_REFCOUNTS "lazy_refcounts" -#define BLOCK_OPT_ADAPTER_TYPE "adapter_type" -#define BLOCK_OPT_REDUNDANCY "redundancy" -#define BLOCK_OPT_NOCOW "nocow" -#define BLOCK_OPT_EXTENT_SIZE_HINT "extent_size_hint" -#define BLOCK_OPT_OBJECT_SIZE "object_size" -#define BLOCK_OPT_REFCOUNT_BITS "refcount_bits" -#define BLOCK_OPT_DATA_FILE "data_file" -#define BLOCK_OPT_DATA_FILE_RAW "data_file_raw" -#define BLOCK_OPT_COMPRESSION_TYPE "compression_type" -#define BLOCK_OPT_EXTL2 "extended_l2" - -#define BLOCK_PROBE_BUF_SIZE 512 - -enum BdrvTrackedRequestType { - BDRV_TRACKED_READ, - BDRV_TRACKED_WRITE, - BDRV_TRACKED_DISCARD, - BDRV_TRACKED_TRUNCATE, -}; - -/* - * That is not quite good that BdrvTrackedRequest structure is public, - * as block/io.c is very careful about incoming offset/bytes being - * correct. Be sure to assert bdrv_check_request() succeeded after any - * modification of BdrvTrackedRequest object out of block/io.c - */ -typedef struct BdrvTrackedRequest { - BlockDriverState *bs; - int64_t offset; - int64_t bytes; - enum BdrvTrackedRequestType type; - - bool serialising; - int64_t overlap_offset; - int64_t overlap_bytes; - - QLIST_ENTRY(BdrvTrackedRequest) list; - Coroutine *co; /* owner, used for deadlock detection */ - CoQueue wait_queue; /* coroutines blocked on this request */ - - struct BdrvTrackedRequest *waiting_for; -} BdrvTrackedRequest; - -int bdrv_check_qiov_request(int64_t offset, int64_t bytes, - QEMUIOVector *qiov, size_t qiov_offset, - Error **errp); -int bdrv_check_request(int64_t offset, int64_t bytes, Error **errp); - -struct BlockDriver { - const char *format_name; - int instance_size; - - /* set to true if the BlockDriver is a block filter. Block filters pass - * certain callbacks that refer to data (see block.c) to their bs->file - * or bs->backing (whichever one exists) if the driver doesn't implement - * them. Drivers that do not wish to forward must implement them and return - * -ENOTSUP. - * Note that filters are not allowed to modify data. - * - * Filters generally cannot have more than a single filtered child, - * because the data they present must at all times be the same as - * that on their filtered child. That would be impossible to - * achieve for multiple filtered children. - * (And this filtered child must then be bs->file or bs->backing.) - */ - bool is_filter; - /* - * Set to true if the BlockDriver is a format driver. Format nodes - * generally do not expect their children to be other format nodes - * (except for backing files), and so format probing is disabled - * on those children. - */ - bool is_format; - - /* - * This function is invoked under BQL before .bdrv_co_amend() - * (which in contrast does not necessarily run under the BQL) - * to allow driver-specific initialization code that requires - * the BQL, like setting up specific permission flags. - */ - int (*bdrv_amend_pre_run)(BlockDriverState *bs, Error **errp); - /* - * This function is invoked under BQL after .bdrv_co_amend() - * to allow cleaning up what was done in .bdrv_amend_pre_run(). - */ - void (*bdrv_amend_clean)(BlockDriverState *bs); - - /* - * Return true if @to_replace can be replaced by a BDS with the - * same data as @bs without it affecting @bs's behavior (that is, - * without it being visible to @bs's parents). - */ - bool (*bdrv_recurse_can_replace)(BlockDriverState *bs, - BlockDriverState *to_replace); - - int (*bdrv_probe)(const uint8_t *buf, int buf_size, const char *filename); - int (*bdrv_probe_device)(const char *filename); - - /* Any driver implementing this callback is expected to be able to handle - * NULL file names in its .bdrv_open() implementation */ - void (*bdrv_parse_filename)(const char *filename, QDict *options, Error **errp); - /* Drivers not implementing bdrv_parse_filename nor bdrv_open should have - * this field set to true, except ones that are defined only by their - * child's bs. - * An example of the last type will be the quorum block driver. - */ - bool bdrv_needs_filename; - - /* - * Set if a driver can support backing files. This also implies the - * following semantics: - * - * - Return status 0 of .bdrv_co_block_status means that corresponding - * blocks are not allocated in this layer of backing-chain - * - For such (unallocated) blocks, read will: - * - fill buffer with zeros if there is no backing file - * - read from the backing file otherwise, where the block layer - * takes care of reading zeros beyond EOF if backing file is short - */ - bool supports_backing; - - /* For handling image reopen for split or non-split files */ - int (*bdrv_reopen_prepare)(BDRVReopenState *reopen_state, - BlockReopenQueue *queue, Error **errp); - void (*bdrv_reopen_commit)(BDRVReopenState *reopen_state); - void (*bdrv_reopen_commit_post)(BDRVReopenState *reopen_state); - void (*bdrv_reopen_abort)(BDRVReopenState *reopen_state); - void (*bdrv_join_options)(QDict *options, QDict *old_options); - - int (*bdrv_open)(BlockDriverState *bs, QDict *options, int flags, - Error **errp); - - /* Protocol drivers should implement this instead of bdrv_open */ - int (*bdrv_file_open)(BlockDriverState *bs, QDict *options, int flags, - Error **errp); - void (*bdrv_close)(BlockDriverState *bs); - - - int coroutine_fn (*bdrv_co_create)(BlockdevCreateOptions *opts, - Error **errp); - int coroutine_fn (*bdrv_co_create_opts)(BlockDriver *drv, - const char *filename, - QemuOpts *opts, - Error **errp); - - int coroutine_fn (*bdrv_co_amend)(BlockDriverState *bs, - BlockdevAmendOptions *opts, - bool force, - Error **errp); - - int (*bdrv_amend_options)(BlockDriverState *bs, - QemuOpts *opts, - BlockDriverAmendStatusCB *status_cb, - void *cb_opaque, - bool force, - Error **errp); - - int (*bdrv_make_empty)(BlockDriverState *bs); - - /* - * Refreshes the bs->exact_filename field. If that is impossible, - * bs->exact_filename has to be left empty. - */ - void (*bdrv_refresh_filename)(BlockDriverState *bs); - - /* - * Gathers the open options for all children into @target. - * A simple format driver (without backing file support) might - * implement this function like this: - * - * QINCREF(bs->file->bs->full_open_options); - * qdict_put(target, "file", bs->file->bs->full_open_options); - * - * If not specified, the generic implementation will simply put - * all children's options under their respective name. - * - * @backing_overridden is true when bs->backing seems not to be - * the child that would result from opening bs->backing_file. - * Therefore, if it is true, the backing child's options should be - * gathered; otherwise, there is no need since the backing child - * is the one implied by the image header. - * - * Note that ideally this function would not be needed. Every - * block driver which implements it is probably doing something - * shady regarding its runtime option structure. - */ - void (*bdrv_gather_child_options)(BlockDriverState *bs, QDict *target, - bool backing_overridden); - - /* - * Returns an allocated string which is the directory name of this BDS: It - * will be used to make relative filenames absolute by prepending this - * function's return value to them. - */ - char *(*bdrv_dirname)(BlockDriverState *bs, Error **errp); - - /* aio */ - BlockAIOCB *(*bdrv_aio_preadv)(BlockDriverState *bs, - int64_t offset, int64_t bytes, QEMUIOVector *qiov, - BdrvRequestFlags flags, BlockCompletionFunc *cb, void *opaque); - BlockAIOCB *(*bdrv_aio_pwritev)(BlockDriverState *bs, - int64_t offset, int64_t bytes, QEMUIOVector *qiov, - BdrvRequestFlags flags, BlockCompletionFunc *cb, void *opaque); - BlockAIOCB *(*bdrv_aio_flush)(BlockDriverState *bs, - BlockCompletionFunc *cb, void *opaque); - BlockAIOCB *(*bdrv_aio_pdiscard)(BlockDriverState *bs, - int64_t offset, int bytes, - BlockCompletionFunc *cb, void *opaque); - - int coroutine_fn (*bdrv_co_readv)(BlockDriverState *bs, - int64_t sector_num, int nb_sectors, QEMUIOVector *qiov); - - /** - * @offset: position in bytes to read at - * @bytes: number of bytes to read - * @qiov: the buffers to fill with read data - * @flags: currently unused, always 0 - * - * @offset and @bytes will be a multiple of 'request_alignment', - * but the length of individual @qiov elements does not have to - * be a multiple. - * - * @bytes will always equal the total size of @qiov, and will be - * no larger than 'max_transfer'. - * - * The buffer in @qiov may point directly to guest memory. - */ - int coroutine_fn (*bdrv_co_preadv)(BlockDriverState *bs, - int64_t offset, int64_t bytes, QEMUIOVector *qiov, - BdrvRequestFlags flags); - int coroutine_fn (*bdrv_co_preadv_part)(BlockDriverState *bs, - int64_t offset, int64_t bytes, - QEMUIOVector *qiov, size_t qiov_offset, BdrvRequestFlags flags); - int coroutine_fn (*bdrv_co_writev)(BlockDriverState *bs, - int64_t sector_num, int nb_sectors, QEMUIOVector *qiov, int flags); - /** - * @offset: position in bytes to write at - * @bytes: number of bytes to write - * @qiov: the buffers containing data to write - * @flags: zero or more bits allowed by 'supported_write_flags' - * - * @offset and @bytes will be a multiple of 'request_alignment', - * but the length of individual @qiov elements does not have to - * be a multiple. - * - * @bytes will always equal the total size of @qiov, and will be - * no larger than 'max_transfer'. - * - * The buffer in @qiov may point directly to guest memory. - */ - int coroutine_fn (*bdrv_co_pwritev)(BlockDriverState *bs, - int64_t offset, int64_t bytes, QEMUIOVector *qiov, - BdrvRequestFlags flags); - int coroutine_fn (*bdrv_co_pwritev_part)(BlockDriverState *bs, - int64_t offset, int64_t bytes, QEMUIOVector *qiov, size_t qiov_offset, - BdrvRequestFlags flags); - - /* - * Efficiently zero a region of the disk image. Typically an image format - * would use a compact metadata representation to implement this. This - * function pointer may be NULL or return -ENOSUP and .bdrv_co_writev() - * will be called instead. - */ - int coroutine_fn (*bdrv_co_pwrite_zeroes)(BlockDriverState *bs, - int64_t offset, int64_t bytes, BdrvRequestFlags flags); - int coroutine_fn (*bdrv_co_pdiscard)(BlockDriverState *bs, - int64_t offset, int64_t bytes); - - /* Map [offset, offset + nbytes) range onto a child of @bs to copy from, - * and invoke bdrv_co_copy_range_from(child, ...), or invoke - * bdrv_co_copy_range_to() if @bs is the leaf child to copy data from. - * - * See the comment of bdrv_co_copy_range for the parameter and return value - * semantics. - */ - int coroutine_fn (*bdrv_co_copy_range_from)(BlockDriverState *bs, - BdrvChild *src, - int64_t offset, - BdrvChild *dst, - int64_t dst_offset, - int64_t bytes, - BdrvRequestFlags read_flags, - BdrvRequestFlags write_flags); - - /* Map [offset, offset + nbytes) range onto a child of bs to copy data to, - * and invoke bdrv_co_copy_range_to(child, src, ...), or perform the copy - * operation if @bs is the leaf and @src has the same BlockDriver. Return - * -ENOTSUP if @bs is the leaf but @src has a different BlockDriver. - * - * See the comment of bdrv_co_copy_range for the parameter and return value - * semantics. - */ - int coroutine_fn (*bdrv_co_copy_range_to)(BlockDriverState *bs, - BdrvChild *src, - int64_t src_offset, - BdrvChild *dst, - int64_t dst_offset, - int64_t bytes, - BdrvRequestFlags read_flags, - BdrvRequestFlags write_flags); - - /* - * Building block for bdrv_block_status[_above] and - * bdrv_is_allocated[_above]. The driver should answer only - * according to the current layer, and should only need to set - * BDRV_BLOCK_DATA, BDRV_BLOCK_ZERO, BDRV_BLOCK_OFFSET_VALID, - * and/or BDRV_BLOCK_RAW; if the current layer defers to a backing - * layer, the result should be 0 (and not BDRV_BLOCK_ZERO). See - * block.h for the overall meaning of the bits. As a hint, the - * flag want_zero is true if the caller cares more about precise - * mappings (favor accurate _OFFSET_VALID/_ZERO) or false for - * overall allocation (favor larger *pnum, perhaps by reporting - * _DATA instead of _ZERO). The block layer guarantees input - * clamped to bdrv_getlength() and aligned to request_alignment, - * as well as non-NULL pnum, map, and file; in turn, the driver - * must return an error or set pnum to an aligned non-zero value. - * - * Note that @bytes is just a hint on how big of a region the - * caller wants to inspect. It is not a limit on *pnum. - * Implementations are free to return larger values of *pnum if - * doing so does not incur a performance penalty. - * - * block/io.c's bdrv_co_block_status() will utilize an unclamped - * *pnum value for the block-status cache on protocol nodes, prior - * to clamping *pnum for return to its caller. - */ - int coroutine_fn (*bdrv_co_block_status)(BlockDriverState *bs, - bool want_zero, int64_t offset, int64_t bytes, int64_t *pnum, - int64_t *map, BlockDriverState **file); - - /* - * This informs the driver that we are no longer interested in the result - * of in-flight requests, so don't waste the time if possible. - * - * One example usage is to avoid waiting for an nbd target node reconnect - * timeout during job-cancel with force=true. - */ - void (*bdrv_cancel_in_flight)(BlockDriverState *bs); - - /* - * Invalidate any cached meta-data. - */ - void coroutine_fn (*bdrv_co_invalidate_cache)(BlockDriverState *bs, - Error **errp); - int (*bdrv_inactivate)(BlockDriverState *bs); - - /* - * Flushes all data for all layers by calling bdrv_co_flush for underlying - * layers, if needed. This function is needed for deterministic - * synchronization of the flush finishing callback. - */ - int coroutine_fn (*bdrv_co_flush)(BlockDriverState *bs); - - /* Delete a created file. */ - int coroutine_fn (*bdrv_co_delete_file)(BlockDriverState *bs, - Error **errp); - - /* - * Flushes all data that was already written to the OS all the way down to - * the disk (for example file-posix.c calls fsync()). - */ - int coroutine_fn (*bdrv_co_flush_to_disk)(BlockDriverState *bs); - - /* - * Flushes all internal caches to the OS. The data may still sit in a - * writeback cache of the host OS, but it will survive a crash of the qemu - * process. - */ - int coroutine_fn (*bdrv_co_flush_to_os)(BlockDriverState *bs); - - /* - * Drivers setting this field must be able to work with just a plain - * filename with '<protocol_name>:' as a prefix, and no other options. - * Options may be extracted from the filename by implementing - * bdrv_parse_filename. - */ - const char *protocol_name; - - /* - * Truncate @bs to @offset bytes using the given @prealloc mode - * when growing. Modes other than PREALLOC_MODE_OFF should be - * rejected when shrinking @bs. - * - * If @exact is true, @bs must be resized to exactly @offset. - * Otherwise, it is sufficient for @bs (if it is a host block - * device and thus there is no way to resize it) to be at least - * @offset bytes in length. - * - * If @exact is true and this function fails but would succeed - * with @exact = false, it should return -ENOTSUP. - */ - int coroutine_fn (*bdrv_co_truncate)(BlockDriverState *bs, int64_t offset, - bool exact, PreallocMode prealloc, - BdrvRequestFlags flags, Error **errp); - - int64_t (*bdrv_getlength)(BlockDriverState *bs); - bool has_variable_length; - int64_t (*bdrv_get_allocated_file_size)(BlockDriverState *bs); - BlockMeasureInfo *(*bdrv_measure)(QemuOpts *opts, BlockDriverState *in_bs, - Error **errp); - - int coroutine_fn (*bdrv_co_pwritev_compressed)(BlockDriverState *bs, - int64_t offset, int64_t bytes, QEMUIOVector *qiov); - int coroutine_fn (*bdrv_co_pwritev_compressed_part)(BlockDriverState *bs, - int64_t offset, int64_t bytes, QEMUIOVector *qiov, size_t qiov_offset); - - int (*bdrv_snapshot_create)(BlockDriverState *bs, - QEMUSnapshotInfo *sn_info); - int (*bdrv_snapshot_goto)(BlockDriverState *bs, - const char *snapshot_id); - int (*bdrv_snapshot_delete)(BlockDriverState *bs, - const char *snapshot_id, - const char *name, - Error **errp); - int (*bdrv_snapshot_list)(BlockDriverState *bs, - QEMUSnapshotInfo **psn_info); - int (*bdrv_snapshot_load_tmp)(BlockDriverState *bs, - const char *snapshot_id, - const char *name, - Error **errp); - int (*bdrv_get_info)(BlockDriverState *bs, BlockDriverInfo *bdi); - ImageInfoSpecific *(*bdrv_get_specific_info)(BlockDriverState *bs, - Error **errp); - BlockStatsSpecific *(*bdrv_get_specific_stats)(BlockDriverState *bs); - - int coroutine_fn (*bdrv_save_vmstate)(BlockDriverState *bs, - QEMUIOVector *qiov, - int64_t pos); - int coroutine_fn (*bdrv_load_vmstate)(BlockDriverState *bs, - QEMUIOVector *qiov, - int64_t pos); - - int (*bdrv_change_backing_file)(BlockDriverState *bs, - const char *backing_file, const char *backing_fmt); - - /* removable device specific */ - bool (*bdrv_is_inserted)(BlockDriverState *bs); - void (*bdrv_eject)(BlockDriverState *bs, bool eject_flag); - void (*bdrv_lock_medium)(BlockDriverState *bs, bool locked); - - /* to control generic scsi devices */ - BlockAIOCB *(*bdrv_aio_ioctl)(BlockDriverState *bs, - unsigned long int req, void *buf, - BlockCompletionFunc *cb, void *opaque); - int coroutine_fn (*bdrv_co_ioctl)(BlockDriverState *bs, - unsigned long int req, void *buf); - - /* List of options for creating images, terminated by name == NULL */ - QemuOptsList *create_opts; - - /* List of options for image amend */ - QemuOptsList *amend_opts; - - /* - * If this driver supports reopening images this contains a - * NULL-terminated list of the runtime options that can be - * modified. If an option in this list is unspecified during - * reopen then it _must_ be reset to its default value or return - * an error. - */ - const char *const *mutable_opts; - - /* - * Returns 0 for completed check, -errno for internal errors. - * The check results are stored in result. - */ - int coroutine_fn (*bdrv_co_check)(BlockDriverState *bs, - BdrvCheckResult *result, - BdrvCheckMode fix); - - void (*bdrv_debug_event)(BlockDriverState *bs, BlkdebugEvent event); - - /* TODO Better pass a option string/QDict/QemuOpts to add any rule? */ - int (*bdrv_debug_breakpoint)(BlockDriverState *bs, const char *event, - const char *tag); - int (*bdrv_debug_remove_breakpoint)(BlockDriverState *bs, - const char *tag); - int (*bdrv_debug_resume)(BlockDriverState *bs, const char *tag); - bool (*bdrv_debug_is_suspended)(BlockDriverState *bs, const char *tag); - - void (*bdrv_refresh_limits)(BlockDriverState *bs, Error **errp); - - /* - * Returns 1 if newly created images are guaranteed to contain only - * zeros, 0 otherwise. - */ - int (*bdrv_has_zero_init)(BlockDriverState *bs); - - /* Remove fd handlers, timers, and other event loop callbacks so the event - * loop is no longer in use. Called with no in-flight requests and in - * depth-first traversal order with parents before child nodes. - */ - void (*bdrv_detach_aio_context)(BlockDriverState *bs); - - /* Add fd handlers, timers, and other event loop callbacks so I/O requests - * can be processed again. Called with no in-flight requests and in - * depth-first traversal order with child nodes before parent nodes. - */ - void (*bdrv_attach_aio_context)(BlockDriverState *bs, - AioContext *new_context); - - /* io queue for linux-aio */ - void (*bdrv_io_plug)(BlockDriverState *bs); - void (*bdrv_io_unplug)(BlockDriverState *bs); - - /** - * Try to get @bs's logical and physical block size. - * On success, store them in @bsz and return zero. - * On failure, return negative errno. - */ - int (*bdrv_probe_blocksizes)(BlockDriverState *bs, BlockSizes *bsz); - /** - * Try to get @bs's geometry (cyls, heads, sectors) - * On success, store them in @geo and return 0. - * On failure return -errno. - * Only drivers that want to override guest geometry implement this - * callback; see hd_geometry_guess(). - */ - int (*bdrv_probe_geometry)(BlockDriverState *bs, HDGeometry *geo); - - /** - * bdrv_co_drain_begin is called if implemented in the beginning of a - * drain operation to drain and stop any internal sources of requests in - * the driver. - * bdrv_co_drain_end is called if implemented at the end of the drain. - * - * They should be used by the driver to e.g. manage scheduled I/O - * requests, or toggle an internal state. After the end of the drain new - * requests will continue normally. - */ - void coroutine_fn (*bdrv_co_drain_begin)(BlockDriverState *bs); - void coroutine_fn (*bdrv_co_drain_end)(BlockDriverState *bs); - - void (*bdrv_add_child)(BlockDriverState *parent, BlockDriverState *child, - Error **errp); - void (*bdrv_del_child)(BlockDriverState *parent, BdrvChild *child, - Error **errp); - - /** - * Informs the block driver that a permission change is intended. The - * driver checks whether the change is permissible and may take other - * preparations for the change (e.g. get file system locks). This operation - * is always followed either by a call to either .bdrv_set_perm or - * .bdrv_abort_perm_update. - * - * Checks whether the requested set of cumulative permissions in @perm - * can be granted for accessing @bs and whether no other users are using - * permissions other than those given in @shared (both arguments take - * BLK_PERM_* bitmasks). - * - * If both conditions are met, 0 is returned. Otherwise, -errno is returned - * and errp is set to an error describing the conflict. - */ - int (*bdrv_check_perm)(BlockDriverState *bs, uint64_t perm, - uint64_t shared, Error **errp); - - /** - * Called to inform the driver that the set of cumulative set of used - * permissions for @bs has changed to @perm, and the set of sharable - * permission to @shared. The driver can use this to propagate changes to - * its children (i.e. request permissions only if a parent actually needs - * them). - * - * This function is only invoked after bdrv_check_perm(), so block drivers - * may rely on preparations made in their .bdrv_check_perm implementation. - */ - void (*bdrv_set_perm)(BlockDriverState *bs, uint64_t perm, uint64_t shared); - - /* - * Called to inform the driver that after a previous bdrv_check_perm() - * call, the permission update is not performed and any preparations made - * for it (e.g. taken file locks) need to be undone. - * - * This function can be called even for nodes that never saw a - * bdrv_check_perm() call. It is a no-op then. - */ - void (*bdrv_abort_perm_update)(BlockDriverState *bs); - - /** - * Returns in @nperm and @nshared the permissions that the driver for @bs - * needs on its child @c, based on the cumulative permissions requested by - * the parents in @parent_perm and @parent_shared. - * - * If @c is NULL, return the permissions for attaching a new child for the - * given @child_class and @role. - * - * If @reopen_queue is non-NULL, don't return the currently needed - * permissions, but those that will be needed after applying the - * @reopen_queue. - */ - void (*bdrv_child_perm)(BlockDriverState *bs, BdrvChild *c, - BdrvChildRole role, - BlockReopenQueue *reopen_queue, - uint64_t parent_perm, uint64_t parent_shared, - uint64_t *nperm, uint64_t *nshared); - - bool (*bdrv_supports_persistent_dirty_bitmap)(BlockDriverState *bs); - bool (*bdrv_co_can_store_new_dirty_bitmap)(BlockDriverState *bs, - const char *name, - uint32_t granularity, - Error **errp); - int (*bdrv_co_remove_persistent_dirty_bitmap)(BlockDriverState *bs, - const char *name, - Error **errp); - - /** - * Register/unregister a buffer for I/O. For example, when the driver is - * interested to know the memory areas that will later be used in iovs, so - * that it can do IOMMU mapping with VFIO etc., in order to get better - * performance. In the case of VFIO drivers, this callback is used to do - * DMA mapping for hot buffers. - */ - void (*bdrv_register_buf)(BlockDriverState *bs, void *host, size_t size); - void (*bdrv_unregister_buf)(BlockDriverState *bs, void *host); - QLIST_ENTRY(BlockDriver) list; - - /* Pointer to a NULL-terminated array of names of strong options - * that can be specified for bdrv_open(). A strong option is one - * that changes the data of a BDS. - * If this pointer is NULL, the array is considered empty. - * "filename" and "driver" are always considered strong. */ - const char *const *strong_runtime_opts; -}; - -static inline bool block_driver_can_compress(BlockDriver *drv) -{ - return drv->bdrv_co_pwritev_compressed || - drv->bdrv_co_pwritev_compressed_part; -} - -typedef struct BlockLimits { - /* Alignment requirement, in bytes, for offset/length of I/O - * requests. Must be a power of 2 less than INT_MAX; defaults to - * 1 for drivers with modern byte interfaces, and to 512 - * otherwise. */ - uint32_t request_alignment; - - /* - * Maximum number of bytes that can be discarded at once. Must be multiple - * of pdiscard_alignment, but need not be power of 2. May be 0 if no - * inherent 64-bit limit. - */ - int64_t max_pdiscard; - - /* Optimal alignment for discard requests in bytes. A power of 2 - * is best but not mandatory. Must be a multiple of - * bl.request_alignment, and must be less than max_pdiscard if - * that is set. May be 0 if bl.request_alignment is good enough */ - uint32_t pdiscard_alignment; - - /* - * Maximum number of bytes that can zeroized at once. Must be multiple of - * pwrite_zeroes_alignment. 0 means no limit. - */ - int64_t max_pwrite_zeroes; - - /* Optimal alignment for write zeroes requests in bytes. A power - * of 2 is best but not mandatory. Must be a multiple of - * bl.request_alignment, and must be less than max_pwrite_zeroes - * if that is set. May be 0 if bl.request_alignment is good - * enough */ - uint32_t pwrite_zeroes_alignment; - - /* Optimal transfer length in bytes. A power of 2 is best but not - * mandatory. Must be a multiple of bl.request_alignment, or 0 if - * no preferred size */ - uint32_t opt_transfer; - - /* Maximal transfer length in bytes. Need not be power of 2, but - * must be multiple of opt_transfer and bl.request_alignment, or 0 - * for no 32-bit limit. For now, anything larger than INT_MAX is - * clamped down. */ - uint32_t max_transfer; - - /* Maximal hardware transfer length in bytes. Applies whenever - * transfers to the device bypass the kernel I/O scheduler, for - * example with SG_IO. If larger than max_transfer or if zero, - * blk_get_max_hw_transfer will fall back to max_transfer. - */ - uint64_t max_hw_transfer; - - /* Maximal number of scatter/gather elements allowed by the hardware. - * Applies whenever transfers to the device bypass the kernel I/O - * scheduler, for example with SG_IO. If larger than max_iov - * or if zero, blk_get_max_hw_iov will fall back to max_iov. - */ - int max_hw_iov; - - /* memory alignment, in bytes so that no bounce buffer is needed */ - size_t min_mem_alignment; - - /* memory alignment, in bytes, for bounce buffer */ - size_t opt_mem_alignment; - - /* maximum number of iovec elements */ - int max_iov; -} BlockLimits; - -typedef struct BdrvOpBlocker BdrvOpBlocker; - -typedef struct BdrvAioNotifier { - void (*attached_aio_context)(AioContext *new_context, void *opaque); - void (*detach_aio_context)(void *opaque); - - void *opaque; - bool deleted; - - QLIST_ENTRY(BdrvAioNotifier) list; -} BdrvAioNotifier; - -struct BdrvChildClass { - /* If true, bdrv_replace_node() doesn't change the node this BdrvChild - * points to. */ - bool stay_at_node; - - /* If true, the parent is a BlockDriverState and bdrv_next_all_states() - * will return it. This information is used for drain_all, where every node - * will be drained separately, so the drain only needs to be propagated to - * non-BDS parents. */ - bool parent_is_bds; - - void (*inherit_options)(BdrvChildRole role, bool parent_is_format, - int *child_flags, QDict *child_options, - int parent_flags, QDict *parent_options); - - void (*change_media)(BdrvChild *child, bool load); - void (*resize)(BdrvChild *child); - - /* Returns a name that is supposedly more useful for human users than the - * node name for identifying the node in question (in particular, a BB - * name), or NULL if the parent can't provide a better name. */ - const char *(*get_name)(BdrvChild *child); - - /* Returns a malloced string that describes the parent of the child for a - * human reader. This could be a node-name, BlockBackend name, qdev ID or - * QOM path of the device owning the BlockBackend, job type and ID etc. The - * caller is responsible for freeing the memory. */ - char *(*get_parent_desc)(BdrvChild *child); - - /* - * If this pair of functions is implemented, the parent doesn't issue new - * requests after returning from .drained_begin() until .drained_end() is - * called. - * - * These functions must not change the graph (and therefore also must not - * call aio_poll(), which could change the graph indirectly). - * - * If drained_end() schedules background operations, it must atomically - * increment *drained_end_counter for each such operation and atomically - * decrement it once the operation has settled. - * - * Note that this can be nested. If drained_begin() was called twice, new - * I/O is allowed only after drained_end() was called twice, too. - */ - void (*drained_begin)(BdrvChild *child); - void (*drained_end)(BdrvChild *child, int *drained_end_counter); - - /* - * Returns whether the parent has pending requests for the child. This - * callback is polled after .drained_begin() has been called until all - * activity on the child has stopped. - */ - bool (*drained_poll)(BdrvChild *child); - - /* Notifies the parent that the child has been activated/inactivated (e.g. - * when migration is completing) and it can start/stop requesting - * permissions and doing I/O on it. */ - void (*activate)(BdrvChild *child, Error **errp); - int (*inactivate)(BdrvChild *child); - - void (*attach)(BdrvChild *child); - void (*detach)(BdrvChild *child); - - /* Notifies the parent that the filename of its child has changed (e.g. - * because the direct child was removed from the backing chain), so that it - * can update its reference. */ - int (*update_filename)(BdrvChild *child, BlockDriverState *new_base, - const char *filename, Error **errp); - - bool (*can_set_aio_ctx)(BdrvChild *child, AioContext *ctx, - GSList **ignore, Error **errp); - void (*set_aio_ctx)(BdrvChild *child, AioContext *ctx, GSList **ignore); - - AioContext *(*get_parent_aio_context)(BdrvChild *child); -}; - -extern const BdrvChildClass child_of_bds; - -struct BdrvChild { - BlockDriverState *bs; - char *name; - const BdrvChildClass *klass; - BdrvChildRole role; - void *opaque; - - /** - * Granted permissions for operating on this BdrvChild (BLK_PERM_* bitmask) - */ - uint64_t perm; - - /** - * Permissions that can still be granted to other users of @bs while this - * BdrvChild is still attached to it. (BLK_PERM_* bitmask) - */ - uint64_t shared_perm; - - /* - * This link is frozen: the child can neither be replaced nor - * detached from the parent. - */ - bool frozen; - - /* - * How many times the parent of this child has been drained - * (through klass->drained_*). - * Usually, this is equal to bs->quiesce_counter (potentially - * reduced by bdrv_drain_all_count). It may differ while the - * child is entering or leaving a drained section. - */ - int parent_quiesce_counter; - - QLIST_ENTRY(BdrvChild) next; - QLIST_ENTRY(BdrvChild) next_parent; -}; - -/* - * Allows bdrv_co_block_status() to cache one data region for a - * protocol node. - * - * @valid: Whether the cache is valid (should be accessed with atomic - * functions so this can be reset by RCU readers) - * @data_start: Offset where we know (or strongly assume) is data - * @data_end: Offset where the data region ends (which is not necessarily - * the start of a zeroed region) - */ -typedef struct BdrvBlockStatusCache { - struct rcu_head rcu; - - bool valid; - int64_t data_start; - int64_t data_end; -} BdrvBlockStatusCache; - -struct BlockDriverState { - /* Protected by big QEMU lock or read-only after opening. No special - * locking needed during I/O... - */ - int open_flags; /* flags used to open the file, re-used for re-open */ - bool encrypted; /* if true, the media is encrypted */ - bool sg; /* if true, the device is a /dev/sg* */ - bool probed; /* if true, format was probed rather than specified */ - bool force_share; /* if true, always allow all shared permissions */ - bool implicit; /* if true, this filter node was automatically inserted */ - - BlockDriver *drv; /* NULL means no media */ - void *opaque; - - AioContext *aio_context; /* event loop used for fd handlers, timers, etc */ - /* long-running tasks intended to always use the same AioContext as this - * BDS may register themselves in this list to be notified of changes - * regarding this BDS's context */ - QLIST_HEAD(, BdrvAioNotifier) aio_notifiers; - bool walking_aio_notifiers; /* to make removal during iteration safe */ - - char filename[PATH_MAX]; - /* - * If not empty, this image is a diff in relation to backing_file. - * Note that this is the name given in the image header and - * therefore may or may not be equal to .backing->bs->filename. - * If this field contains a relative path, it is to be resolved - * relatively to the overlay's location. - */ - char backing_file[PATH_MAX]; - /* - * The backing filename indicated by the image header. Contrary - * to backing_file, if we ever open this file, auto_backing_file - * is replaced by the resulting BDS's filename (i.e. after a - * bdrv_refresh_filename() run). - */ - char auto_backing_file[PATH_MAX]; - char backing_format[16]; /* if non-zero and backing_file exists */ - - QDict *full_open_options; - char exact_filename[PATH_MAX]; - - BdrvChild *backing; - BdrvChild *file; - - /* I/O Limits */ - BlockLimits bl; - - /* - * Flags honored during pread - */ - unsigned int supported_read_flags; - /* Flags honored during pwrite (so far: BDRV_REQ_FUA, - * BDRV_REQ_WRITE_UNCHANGED). - * If a driver does not support BDRV_REQ_WRITE_UNCHANGED, those - * writes will be issued as normal writes without the flag set. - * This is important to note for drivers that do not explicitly - * request a WRITE permission for their children and instead take - * the same permissions as their parent did (this is commonly what - * block filters do). Such drivers have to be aware that the - * parent may have taken a WRITE_UNCHANGED permission only and is - * issuing such requests. Drivers either must make sure that - * these requests do not result in plain WRITE accesses (usually - * by supporting BDRV_REQ_WRITE_UNCHANGED, and then forwarding - * every incoming write request as-is, including potentially that - * flag), or they have to explicitly take the WRITE permission for - * their children. */ - unsigned int supported_write_flags; - /* 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]; - /* element of the list of named nodes building the graph */ - QTAILQ_ENTRY(BlockDriverState) node_list; - /* element of the list of all BlockDriverStates (all_bdrv_states) */ - QTAILQ_ENTRY(BlockDriverState) bs_list; - /* element of the list of monitor-owned BDS */ - QTAILQ_ENTRY(BlockDriverState) monitor_list; - int refcnt; - - /* operation blockers */ - QLIST_HEAD(, BdrvOpBlocker) op_blockers[BLOCK_OP_TYPE_MAX]; - - /* The node that this node inherited default options from (and a reopen on - * which can affect this node by changing these defaults). This is always a - * parent node of this node. */ - BlockDriverState *inherits_from; - QLIST_HEAD(, BdrvChild) children; - QLIST_HEAD(, BdrvChild) parents; - - QDict *options; - QDict *explicit_options; - BlockdevDetectZeroesOptions detect_zeroes; - - /* The error object in use for blocking operations on backing_hd */ - Error *backing_blocker; - - /* Protected by AioContext lock */ - - /* If we are reading a disk image, give its size in sectors. - * Generally read-only; it is written to by load_snapshot and - * save_snaphost, but the block layer is quiescent during those. - */ - int64_t total_sectors; - - /* threshold limit for writes, in bytes. "High water mark". */ - uint64_t write_threshold_offset; - - /* Writing to the list requires the BQL _and_ the dirty_bitmap_mutex. - * Reading from the list can be done with either the BQL or the - * dirty_bitmap_mutex. Modifying a bitmap only requires - * dirty_bitmap_mutex. */ - QemuMutex dirty_bitmap_mutex; - QLIST_HEAD(, BdrvDirtyBitmap) dirty_bitmaps; - - /* Offset after the highest byte written to */ - Stat64 wr_highest_offset; - - /* If true, copy read backing sectors into image. Can be >1 if more - * than one client has requested copy-on-read. Accessed with atomic - * ops. - */ - int copy_on_read; - - /* number of in-flight requests; overall and serialising. - * Accessed with atomic ops. - */ - unsigned int in_flight; - unsigned int serialising_in_flight; - - /* counter for nested bdrv_io_plug. - * Accessed with atomic ops. - */ - unsigned io_plugged; - - /* do we need to tell the quest if we have a volatile write cache? */ - int enable_write_cache; - - /* Accessed with atomic ops. */ - int quiesce_counter; - int recursive_quiesce_counter; - - unsigned int write_gen; /* Current data generation */ - - /* Protected by reqs_lock. */ - CoMutex reqs_lock; - QLIST_HEAD(, BdrvTrackedRequest) tracked_requests; - CoQueue flush_queue; /* Serializing flush queue */ - bool active_flush_req; /* Flush request in flight? */ - - /* Only read/written by whoever has set active_flush_req to true. */ - unsigned int flushed_gen; /* Flushed write generation */ - - /* BdrvChild links to this node may never be frozen */ - bool never_freeze; - - /* Lock for block-status cache RCU writers */ - CoMutex bsc_modify_lock; - /* Always non-NULL, but must only be dereferenced under an RCU read guard */ - BdrvBlockStatusCache *block_status_cache; -}; - -struct BlockBackendRootState { - int open_flags; - BlockdevDetectZeroesOptions detect_zeroes; -}; - -typedef enum BlockMirrorBackingMode { - /* Reuse the existing backing chain from the source for the target. - * - sync=full: Set backing BDS to NULL. - * - sync=top: Use source's backing BDS. - * - sync=none: Use source as the backing BDS. */ - MIRROR_SOURCE_BACKING_CHAIN, - - /* Open the target's backing chain completely anew */ - MIRROR_OPEN_BACKING_CHAIN, - - /* Do not change the target's backing BDS after job completion */ - MIRROR_LEAVE_BACKING_CHAIN, -} BlockMirrorBackingMode; - - -/* Essential block drivers which must always be statically linked into qemu, and - * which therefore can be accessed without using bdrv_find_format() */ -extern BlockDriver bdrv_file; -extern BlockDriver bdrv_raw; -extern BlockDriver bdrv_qcow2; - -int coroutine_fn bdrv_co_preadv(BdrvChild *child, - int64_t offset, int64_t bytes, QEMUIOVector *qiov, - BdrvRequestFlags flags); -int coroutine_fn bdrv_co_preadv_part(BdrvChild *child, - int64_t offset, int64_t bytes, - QEMUIOVector *qiov, size_t qiov_offset, BdrvRequestFlags flags); -int coroutine_fn bdrv_co_pwritev(BdrvChild *child, - int64_t offset, int64_t bytes, QEMUIOVector *qiov, - BdrvRequestFlags flags); -int coroutine_fn bdrv_co_pwritev_part(BdrvChild *child, - int64_t offset, int64_t bytes, - QEMUIOVector *qiov, size_t qiov_offset, BdrvRequestFlags flags); - -static inline int coroutine_fn bdrv_co_pread(BdrvChild *child, - int64_t offset, unsigned int bytes, void *buf, BdrvRequestFlags flags) -{ - QEMUIOVector qiov = QEMU_IOVEC_INIT_BUF(qiov, buf, bytes); - - return bdrv_co_preadv(child, offset, bytes, &qiov, flags); -} - -static inline int coroutine_fn bdrv_co_pwrite(BdrvChild *child, - int64_t offset, unsigned int bytes, void *buf, BdrvRequestFlags flags) -{ - QEMUIOVector qiov = QEMU_IOVEC_INIT_BUF(qiov, buf, bytes); - - return bdrv_co_pwritev(child, offset, bytes, &qiov, flags); -} - -extern unsigned int bdrv_drain_all_count; -void bdrv_apply_subtree_drain(BdrvChild *child, BlockDriverState *new_parent); -void bdrv_unapply_subtree_drain(BdrvChild *child, BlockDriverState *old_parent); - -bool coroutine_fn bdrv_make_request_serialising(BdrvTrackedRequest *req, - uint64_t align); -BdrvTrackedRequest *coroutine_fn bdrv_co_get_self_request(BlockDriverState *bs); - -int get_tmp_filename(char *filename, int size); -BlockDriver *bdrv_probe_all(const uint8_t *buf, int buf_size, - const char *filename); - -void bdrv_parse_filename_strip_prefix(const char *filename, const char *prefix, - QDict *options); - -/** - * bdrv_add_aio_context_notifier: - * - * If a long-running job intends to be always run in the same AioContext as a - * certain BDS, it may use this function to be notified of changes regarding the - * association of the BDS to an AioContext. - * - * attached_aio_context() is called after the target BDS has been attached to a - * new AioContext; detach_aio_context() is called before the target BDS is being - * detached from its old AioContext. - */ -void bdrv_add_aio_context_notifier(BlockDriverState *bs, - void (*attached_aio_context)(AioContext *new_context, void *opaque), - void (*detach_aio_context)(void *opaque), void *opaque); - -/** - * bdrv_remove_aio_context_notifier: - * - * Unsubscribe of change notifications regarding the BDS's AioContext. The - * parameters given here have to be the same as those given to - * bdrv_add_aio_context_notifier(). - */ -void bdrv_remove_aio_context_notifier(BlockDriverState *bs, - void (*aio_context_attached)(AioContext *, - void *), - void (*aio_context_detached)(void *), - void *opaque); - -/** - * bdrv_wakeup: - * @bs: The BlockDriverState for which an I/O operation has been completed. - * - * Wake up the main thread if it is waiting on BDRV_POLL_WHILE. During - * synchronous I/O on a BlockDriverState that is attached to another - * I/O thread, the main thread lets the I/O thread's event loop run, - * waiting for the I/O operation to complete. A bdrv_wakeup will wake - * up the main thread if necessary. - * - * Manual calls to bdrv_wakeup are rarely necessary, because - * bdrv_dec_in_flight already calls it. - */ -void bdrv_wakeup(BlockDriverState *bs); - -#ifdef _WIN32 -int is_windows_drive(const char *filename); -#endif - -/** - * stream_start: - * @job_id: The id of the newly-created job, or %NULL to use the - * device name of @bs. - * @bs: Block device to operate on. - * @base: Block device that will become the new base, or %NULL to - * flatten the whole backing file chain onto @bs. - * @backing_file_str: The file name that will be written to @bs as the - * the new backing file if the job completes. Ignored if @base is %NULL. - * @creation_flags: Flags that control the behavior of the Job lifetime. - * See @BlockJobCreateFlags - * @speed: The maximum speed, in bytes per second, or 0 for unlimited. - * @on_error: The action to take upon error. - * @filter_node_name: The node name that should be assigned to the filter - * driver that the stream job inserts into the graph above - * @bs. NULL means that a node name should be autogenerated. - * @errp: Error object. - * - * Start a streaming operation on @bs. Clusters that are unallocated - * in @bs, but allocated in any image between @base and @bs (both - * exclusive) will be written to @bs. At the end of a successful - * streaming job, the backing file of @bs will be changed to - * @backing_file_str in the written image and to @base in the live - * BlockDriverState. - */ -void stream_start(const char *job_id, BlockDriverState *bs, - BlockDriverState *base, const char *backing_file_str, - BlockDriverState *bottom, - int creation_flags, int64_t speed, - BlockdevOnError on_error, - const char *filter_node_name, - Error **errp); - -/** - * commit_start: - * @job_id: The id of the newly-created job, or %NULL to use the - * device name of @bs. - * @bs: Active block device. - * @top: Top block device to be committed. - * @base: Block device that will be written into, and become the new top. - * @creation_flags: Flags that control the behavior of the Job lifetime. - * See @BlockJobCreateFlags - * @speed: The maximum speed, in bytes per second, or 0 for unlimited. - * @on_error: The action to take upon error. - * @backing_file_str: String to use as the backing file in @top's overlay - * @filter_node_name: The node name that should be assigned to the filter - * driver that the commit job inserts into the graph above @top. NULL means - * that a node name should be autogenerated. - * @errp: Error object. - * - */ -void commit_start(const char *job_id, BlockDriverState *bs, - BlockDriverState *base, BlockDriverState *top, - int creation_flags, int64_t speed, - BlockdevOnError on_error, const char *backing_file_str, - const char *filter_node_name, Error **errp); -/** - * commit_active_start: - * @job_id: The id of the newly-created job, or %NULL to use the - * device name of @bs. - * @bs: Active block device to be committed. - * @base: Block device that will be written into, and become the new top. - * @creation_flags: Flags that control the behavior of the Job lifetime. - * See @BlockJobCreateFlags - * @speed: The maximum speed, in bytes per second, or 0 for unlimited. - * @on_error: The action to take upon error. - * @filter_node_name: The node name that should be assigned to the filter - * driver that the commit job inserts into the graph above @bs. NULL means that - * a node name should be autogenerated. - * @cb: Completion function for the job. - * @opaque: Opaque pointer value passed to @cb. - * @auto_complete: Auto complete the job. - * @errp: Error object. - * - */ -BlockJob *commit_active_start(const char *job_id, BlockDriverState *bs, - BlockDriverState *base, int creation_flags, - int64_t speed, BlockdevOnError on_error, - const char *filter_node_name, - BlockCompletionFunc *cb, void *opaque, - bool auto_complete, Error **errp); -/* - * mirror_start: - * @job_id: The id of the newly-created job, or %NULL to use the - * device name of @bs. - * @bs: Block device to operate on. - * @target: Block device to write to. - * @replaces: Block graph node name to replace once the mirror is done. Can - * only be used when full mirroring is selected. - * @creation_flags: Flags that control the behavior of the Job lifetime. - * See @BlockJobCreateFlags - * @speed: The maximum speed, in bytes per second, or 0 for unlimited. - * @granularity: The chosen granularity for the dirty bitmap. - * @buf_size: The amount of data that can be in flight at one time. - * @mode: Whether to collapse all images in the chain to the target. - * @backing_mode: How to establish the target's backing chain after completion. - * @zero_target: Whether the target should be explicitly zero-initialized - * @on_source_error: The action to take upon error reading from the source. - * @on_target_error: The action to take upon error writing to the target. - * @unmap: Whether to unmap target where source sectors only contain zeroes. - * @filter_node_name: The node name that should be assigned to the filter - * driver that the mirror job inserts into the graph above @bs. NULL means that - * a node name should be autogenerated. - * @copy_mode: When to trigger writes to the target. - * @errp: Error object. - * - * Start a mirroring operation on @bs. Clusters that are allocated - * in @bs will be written to @target until the job is cancelled or - * manually completed. At the end of a successful mirroring job, - * @bs will be switched to read from @target. - */ -void mirror_start(const char *job_id, BlockDriverState *bs, - BlockDriverState *target, const char *replaces, - int creation_flags, int64_t speed, - uint32_t granularity, int64_t buf_size, - MirrorSyncMode mode, BlockMirrorBackingMode backing_mode, - bool zero_target, - BlockdevOnError on_source_error, - BlockdevOnError on_target_error, - bool unmap, const char *filter_node_name, - MirrorCopyMode copy_mode, Error **errp); - -/* - * backup_job_create: - * @job_id: The id of the newly-created job, or %NULL to use the - * device name of @bs. - * @bs: Block device to operate on. - * @target: Block device to write to. - * @speed: The maximum speed, in bytes per second, or 0 for unlimited. - * @sync_mode: What parts of the disk image should be copied to the destination. - * @sync_bitmap: The dirty bitmap if sync_mode is 'bitmap' or 'incremental' - * @bitmap_mode: The bitmap synchronization policy to use. - * @perf: Performance options. All actual fields assumed to be present, - * all ".has_*" fields are ignored. - * @on_source_error: The action to take upon error reading from the source. - * @on_target_error: The action to take upon error writing to the target. - * @creation_flags: Flags that control the behavior of the Job lifetime. - * See @BlockJobCreateFlags - * @cb: Completion function for the job. - * @opaque: Opaque pointer value passed to @cb. - * @txn: Transaction that this job is part of (may be NULL). - * - * Create a backup operation on @bs. Clusters in @bs are written to @target - * until the job is cancelled or manually completed. - */ -BlockJob *backup_job_create(const char *job_id, BlockDriverState *bs, - BlockDriverState *target, int64_t speed, - MirrorSyncMode sync_mode, - BdrvDirtyBitmap *sync_bitmap, - BitmapSyncMode bitmap_mode, - bool compress, - const char *filter_node_name, - BackupPerf *perf, - BlockdevOnError on_source_error, - BlockdevOnError on_target_error, - int creation_flags, - BlockCompletionFunc *cb, void *opaque, - JobTxn *txn, Error **errp); - -BdrvChild *bdrv_root_attach_child(BlockDriverState *child_bs, - const char *child_name, - const BdrvChildClass *child_class, - BdrvChildRole child_role, - uint64_t perm, uint64_t shared_perm, - void *opaque, Error **errp); -void bdrv_root_unref_child(BdrvChild *child); - -void bdrv_get_cumulative_perm(BlockDriverState *bs, uint64_t *perm, - uint64_t *shared_perm); - -/** - * Sets a BdrvChild's permissions. Avoid if the parent is a BDS; use - * bdrv_child_refresh_perms() instead and make the parent's - * .bdrv_child_perm() implementation return the correct values. - */ -int bdrv_child_try_set_perm(BdrvChild *c, uint64_t perm, uint64_t shared, - Error **errp); - -/** - * Calls bs->drv->bdrv_child_perm() and updates the child's permission - * masks with the result. - * Drivers should invoke this function whenever an event occurs that - * makes their .bdrv_child_perm() implementation return different - * values than before, but which will not result in the block layer - * automatically refreshing the permissions. - */ -int bdrv_child_refresh_perms(BlockDriverState *bs, BdrvChild *c, Error **errp); - -bool bdrv_recurse_can_replace(BlockDriverState *bs, - BlockDriverState *to_replace); - -/* - * Default implementation for BlockDriver.bdrv_child_perm() that can - * be used by block filters and image formats, as long as they use the - * child_of_bds child class and set an appropriate BdrvChildRole. - */ -void bdrv_default_perms(BlockDriverState *bs, BdrvChild *c, - BdrvChildRole role, BlockReopenQueue *reopen_queue, - uint64_t perm, uint64_t shared, - uint64_t *nperm, uint64_t *nshared); - -const char *bdrv_get_parent_name(const BlockDriverState *bs); -void blk_dev_change_media_cb(BlockBackend *blk, bool load, Error **errp); -bool blk_dev_has_removable_media(BlockBackend *blk); -bool blk_dev_has_tray(BlockBackend *blk); -void blk_dev_eject_request(BlockBackend *blk, bool force); -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); - -void bdrv_clear_dirty_bitmap(BdrvDirtyBitmap *bitmap, HBitmap **out); -void bdrv_restore_dirty_bitmap(BdrvDirtyBitmap *bitmap, HBitmap *backup); -bool bdrv_dirty_bitmap_merge_internal(BdrvDirtyBitmap *dest, - const BdrvDirtyBitmap *src, - HBitmap **backup, bool lock); - -void bdrv_inc_in_flight(BlockDriverState *bs); -void bdrv_dec_in_flight(BlockDriverState *bs); - -void blockdev_close_all_bdrv_states(void); - -int coroutine_fn bdrv_co_copy_range_from(BdrvChild *src, int64_t src_offset, - BdrvChild *dst, int64_t dst_offset, - int64_t bytes, - BdrvRequestFlags read_flags, - BdrvRequestFlags write_flags); -int coroutine_fn bdrv_co_copy_range_to(BdrvChild *src, int64_t src_offset, - BdrvChild *dst, int64_t dst_offset, - int64_t bytes, - BdrvRequestFlags read_flags, - BdrvRequestFlags write_flags); - -int refresh_total_sectors(BlockDriverState *bs, int64_t hint); - -void bdrv_set_monitor_owned(BlockDriverState *bs); -BlockDriverState *bds_tree_init(QDict *bs_opts, Error **errp); - -/** - * Simple implementation of bdrv_co_create_opts for protocol drivers - * which only support creation via opening a file - * (usually existing raw storage device) - */ -int coroutine_fn bdrv_co_create_opts_simple(BlockDriver *drv, - const char *filename, - QemuOpts *opts, - Error **errp); -extern QemuOptsList bdrv_create_opts_simple; - -BdrvDirtyBitmap *block_dirty_bitmap_lookup(const char *node, - const char *name, - BlockDriverState **pbs, - Error **errp); -BdrvDirtyBitmap *block_dirty_bitmap_merge(const char *node, const char *target, - BlockDirtyBitmapMergeSourceList *bms, - HBitmap **backup, Error **errp); -BdrvDirtyBitmap *block_dirty_bitmap_remove(const char *node, const char *name, - bool release, - BlockDriverState **bitmap_bs, - Error **errp); - -BdrvChild *bdrv_cow_child(BlockDriverState *bs); -BdrvChild *bdrv_filter_child(BlockDriverState *bs); -BdrvChild *bdrv_filter_or_cow_child(BlockDriverState *bs); -BdrvChild *bdrv_primary_child(BlockDriverState *bs); -BlockDriverState *bdrv_skip_implicit_filters(BlockDriverState *bs); -BlockDriverState *bdrv_skip_filters(BlockDriverState *bs); -BlockDriverState *bdrv_backing_chain_next(BlockDriverState *bs); - -static inline BlockDriverState *child_bs(BdrvChild *child) -{ - return child ? child->bs : NULL; -} - -static inline BlockDriverState *bdrv_cow_bs(BlockDriverState *bs) -{ - return child_bs(bdrv_cow_child(bs)); -} - -static inline BlockDriverState *bdrv_filter_bs(BlockDriverState *bs) -{ - return child_bs(bdrv_filter_child(bs)); -} - -static inline BlockDriverState *bdrv_filter_or_cow_bs(BlockDriverState *bs) -{ - return child_bs(bdrv_filter_or_cow_child(bs)); -} - -static inline BlockDriverState *bdrv_primary_bs(BlockDriverState *bs) -{ - return child_bs(bdrv_primary_child(bs)); -} - -/** - * End all quiescent sections started by bdrv_drain_all_begin(). This is - * needed when deleting a BDS before bdrv_drain_all_end() is called. - * - * NOTE: this is an internal helper for bdrv_close() *only*. No one else - * should call it. - */ -void bdrv_drain_all_end_quiesce(BlockDriverState *bs); - -/** - * Check whether the given offset is in the cached block-status data - * region. - * - * If it is, and @pnum is not NULL, *pnum is set to - * `bsc.data_end - offset`, i.e. how many bytes, starting from - * @offset, are data (according to the cache). - * Otherwise, *pnum is not touched. - */ -bool bdrv_bsc_is_data(BlockDriverState *bs, int64_t offset, int64_t *pnum); - -/** - * If [offset, offset + bytes) overlaps with the currently cached - * block-status region, invalidate the cache. - * - * (To be used by I/O paths that cause data regions to be zero or - * holes.) - */ -void bdrv_bsc_invalidate_range(BlockDriverState *bs, - int64_t offset, int64_t bytes); - -/** - * Mark the range [offset, offset + bytes) as a data region. - */ -void bdrv_bsc_fill(BlockDriverState *bs, int64_t offset, int64_t bytes); +/* DO NOT ADD ANYTHING IN HERE. USE ONE OF THE HEADERS INCLUDED ABOVE */ #endif /* BLOCK_INT_H */ diff --git a/blockdev.c b/blockdev.c index XXXXXXX..XXXXXXX 100644 --- a/blockdev.c +++ b/blockdev.c @@ -XXX,XX +XXX,XX @@ #include "qemu/main-loop.h" #include "qemu/throttle-options.h" +/* Protected by BQL */ QTAILQ_HEAD(, BlockDriverState) monitor_bdrv_states = QTAILQ_HEAD_INITIALIZER(monitor_bdrv_states); @@ -XXX,XX +XXX,XX @@ typedef struct BlkActionState BlkActionState; * * Only prepare() may fail. In a single transaction, only one of commit() or * abort() will be called. clean() will always be called if it is present. + * + * Always run under BQL. */ typedef struct BlkActionOps { size_t instance_size; @@ -XXX,XX +XXX,XX @@ static TransactionProperties *get_transaction_properties( /* * 'Atomic' group operations. The operations are performed as a set, and if * any fail then we roll back all operations in the group. + * + * Always run under BQL. */ void qmp_transaction(TransactionActionList *dev_list, bool has_props, -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Message-Id: <20220303151616.325444-13-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block.c | 15 +++++++++++++++ block/backup.c | 1 + block/block-backend.c | 3 +++ block/commit.c | 2 ++ block/dirty-bitmap.c | 1 + block/io.c | 1 + block/mirror.c | 4 ++++ block/monitor/bitmap-qmp-cmds.c | 6 ++++++ block/stream.c | 2 ++ blockdev.c | 7 +++++++ 10 files changed, 42 insertions(+) diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_create_opts_simple(BlockDriver *drv, Error *local_err = NULL; int ret; + GLOBAL_STATE_CODE(); + size = qemu_opt_get_size_del(opts, BLOCK_OPT_SIZE, 0); buf = qemu_opt_get_del(opts, BLOCK_OPT_PREALLOC); prealloc = qapi_enum_parse(&PreallocMode_lookup, buf, @@ -XXX,XX +XXX,XX @@ void bdrv_get_cumulative_perm(BlockDriverState *bs, uint64_t *perm, uint64_t cumulative_perms = 0; uint64_t cumulative_shared_perms = BLK_PERM_ALL; + GLOBAL_STATE_CODE(); + QLIST_FOREACH(c, &bs->parents, next_parent) { cumulative_perms |= c->perm; cumulative_shared_perms &= c->shared_perm; @@ -XXX,XX +XXX,XX @@ int bdrv_child_try_set_perm(BdrvChild *c, uint64_t perm, uint64_t shared, Transaction *tran = tran_new(); int ret; + GLOBAL_STATE_CODE(); + bdrv_child_set_perm(c, perm, shared, tran); ret = bdrv_refresh_perms(c->bs, &local_err); @@ -XXX,XX +XXX,XX @@ int bdrv_child_refresh_perms(BlockDriverState *bs, BdrvChild *c, Error **errp) uint64_t parent_perms, parent_shared; uint64_t perms, shared; + GLOBAL_STATE_CODE(); + bdrv_get_cumulative_perm(bs, &parent_perms, &parent_shared); bdrv_child_perm(bs, c->bs, c, c->role, NULL, parent_perms, parent_shared, &perms, &shared); @@ -XXX,XX +XXX,XX @@ void bdrv_default_perms(BlockDriverState *bs, BdrvChild *c, uint64_t perm, uint64_t shared, uint64_t *nperm, uint64_t *nshared) { + GLOBAL_STATE_CODE(); if (role & BDRV_CHILD_FILTERED) { assert(!(role & (BDRV_CHILD_DATA | BDRV_CHILD_METADATA | BDRV_CHILD_COW))); @@ -XXX,XX +XXX,XX @@ BdrvChild *bdrv_root_attach_child(BlockDriverState *child_bs, BdrvChild *child = NULL; Transaction *tran = tran_new(); + GLOBAL_STATE_CODE(); + ret = bdrv_attach_child_common(child_bs, child_name, child_class, child_role, perm, shared_perm, opaque, &child, tran, errp); @@ -XXX,XX +XXX,XX @@ bool bdrv_recurse_can_replace(BlockDriverState *bs, { BlockDriverState *filtered; + GLOBAL_STATE_CODE(); + if (!bs || !bs->drv) { return false; } @@ -XXX,XX +XXX,XX @@ static bool append_strong_runtime_options(QDict *d, BlockDriverState *bs) * would result in exactly bs->backing. */ static bool bdrv_backing_overridden(BlockDriverState *bs) { + GLOBAL_STATE_CODE(); if (bs->backing) { return strcmp(bs->auto_backing_file, bs->backing->bs->filename); @@ -XXX,XX +XXX,XX @@ static BlockDriverState *bdrv_do_skip_filters(BlockDriverState *bs, */ BlockDriverState *bdrv_skip_implicit_filters(BlockDriverState *bs) { + GLOBAL_STATE_CODE(); return bdrv_do_skip_filters(bs, true); } diff --git a/block/backup.c b/block/backup.c index XXXXXXX..XXXXXXX 100644 --- a/block/backup.c +++ b/block/backup.c @@ -XXX,XX +XXX,XX @@ BlockJob *backup_job_create(const char *job_id, BlockDriverState *bs, assert(bs); assert(target); + GLOBAL_STATE_CODE(); /* QMP interface protects us from these cases */ assert(sync_mode != MIRROR_SYNC_MODE_INCREMENTAL); diff --git a/block/block-backend.c b/block/block-backend.c index XXXXXXX..XXXXXXX 100644 --- a/block/block-backend.c +++ b/block/block-backend.c @@ -XXX,XX +XXX,XX @@ static void blk_root_change_media(BdrvChild *child, bool load) */ bool blk_dev_has_removable_media(BlockBackend *blk) { + GLOBAL_STATE_CODE(); return !blk->dev || (blk->dev_ops && blk->dev_ops->change_media_cb); } @@ -XXX,XX +XXX,XX @@ bool blk_dev_has_tray(BlockBackend *blk) */ void blk_dev_eject_request(BlockBackend *blk, bool force) { + GLOBAL_STATE_CODE(); if (blk->dev_ops && blk->dev_ops->eject_request_cb) { blk->dev_ops->eject_request_cb(blk->dev_opaque, force); } @@ -XXX,XX +XXX,XX @@ bool blk_dev_is_tray_open(BlockBackend *blk) */ bool blk_dev_is_medium_locked(BlockBackend *blk) { + GLOBAL_STATE_CODE(); if (blk->dev_ops && blk->dev_ops->is_medium_locked) { return blk->dev_ops->is_medium_locked(blk->dev_opaque); } 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, uint64_t base_perms, iter_shared_perms; int ret; + GLOBAL_STATE_CODE(); + assert(top != bs); if (bdrv_skip_filters(top) == bdrv_skip_filters(base)) { error_setg(errp, "Invalid files for merge: top and base are the same"); diff --git a/block/dirty-bitmap.c b/block/dirty-bitmap.c index XXXXXXX..XXXXXXX 100644 --- a/block/dirty-bitmap.c +++ b/block/dirty-bitmap.c @@ -XXX,XX +XXX,XX @@ void bdrv_restore_dirty_bitmap(BdrvDirtyBitmap *bitmap, HBitmap *backup) { HBitmap *tmp = bitmap->bitmap; assert(!bdrv_dirty_bitmap_readonly(bitmap)); + GLOBAL_STATE_CODE(); bitmap->bitmap = backup; hbitmap_free(tmp); } 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) void bdrv_drain_all_end_quiesce(BlockDriverState *bs) { int drained_end_counter = 0; + GLOBAL_STATE_CODE(); g_assert(bs->quiesce_counter > 0); g_assert(!bs->refcnt); 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 @@ void mirror_start(const char *job_id, BlockDriverState *bs, bool is_none_mode; BlockDriverState *base; + GLOBAL_STATE_CODE(); + if ((mode == MIRROR_SYNC_MODE_INCREMENTAL) || (mode == MIRROR_SYNC_MODE_BITMAP)) { error_setg(errp, "Sync mode '%s' not supported", @@ -XXX,XX +XXX,XX @@ BlockJob *commit_active_start(const char *job_id, BlockDriverState *bs, bool base_read_only; BlockJob *job; + GLOBAL_STATE_CODE(); + base_read_only = bdrv_is_read_only(base); if (base_read_only) { diff --git a/block/monitor/bitmap-qmp-cmds.c b/block/monitor/bitmap-qmp-cmds.c index XXXXXXX..XXXXXXX 100644 --- a/block/monitor/bitmap-qmp-cmds.c +++ b/block/monitor/bitmap-qmp-cmds.c @@ -XXX,XX +XXX,XX @@ BdrvDirtyBitmap *block_dirty_bitmap_lookup(const char *node, BlockDriverState *bs; BdrvDirtyBitmap *bitmap; + GLOBAL_STATE_CODE(); + if (!node) { error_setg(errp, "Node cannot be NULL"); return NULL; @@ -XXX,XX +XXX,XX @@ BdrvDirtyBitmap *block_dirty_bitmap_remove(const char *node, const char *name, BdrvDirtyBitmap *bitmap; AioContext *aio_context; + GLOBAL_STATE_CODE(); + bitmap = block_dirty_bitmap_lookup(node, name, &bs, errp); if (!bitmap || !bs) { return NULL; @@ -XXX,XX +XXX,XX @@ BdrvDirtyBitmap *block_dirty_bitmap_merge(const char *node, const char *target, BlockDirtyBitmapMergeSourceList *lst; Error *local_err = NULL; + GLOBAL_STATE_CODE(); + dst = block_dirty_bitmap_lookup(node, target, &bs, errp); if (!dst) { return NULL; diff --git a/block/stream.c b/block/stream.c index XXXXXXX..XXXXXXX 100644 --- a/block/stream.c +++ b/block/stream.c @@ -XXX,XX +XXX,XX @@ void stream_start(const char *job_id, BlockDriverState *bs, QDict *opts; int ret; + GLOBAL_STATE_CODE(); + assert(!(base && bottom)); assert(!(backing_file_str && bottom)); diff --git a/blockdev.c b/blockdev.c index XXXXXXX..XXXXXXX 100644 --- a/blockdev.c +++ b/blockdev.c @@ -XXX,XX +XXX,XX @@ QTAILQ_HEAD(, BlockDriverState) monitor_bdrv_states = void bdrv_set_monitor_owned(BlockDriverState *bs) { + GLOBAL_STATE_CODE(); QTAILQ_INSERT_TAIL(&monitor_bdrv_states, bs, monitor_list); } @@ -XXX,XX +XXX,XX @@ BlockDriverState *bds_tree_init(QDict *bs_opts, Error **errp) { int bdrv_flags = 0; + GLOBAL_STATE_CODE(); /* bdrv_open() defaults to the values in bdrv_flags (for compatibility * with other callers) rather than what we want as the real defaults. * Apply the defaults here instead. */ @@ -XXX,XX +XXX,XX @@ void blockdev_close_all_bdrv_states(void) { BlockDriverState *bs, *next_bs; + GLOBAL_STATE_CODE(); QTAILQ_FOREACH_SAFE(bs, &monitor_bdrv_states, monitor_list, next_bs) { AioContext *ctx = bdrv_get_aio_context(bs); @@ -XXX,XX +XXX,XX @@ void qmp_transaction(TransactionActionList *dev_list, BlkActionState *state, *next; Error *local_err = NULL; + GLOBAL_STATE_CODE(); + QTAILQ_HEAD(, BlkActionState) snap_bdrv_states; QTAILQ_INIT(&snap_bdrv_states); @@ -XXX,XX +XXX,XX @@ void qmp_blockdev_del(const char *node_name, Error **errp) AioContext *aio_context; BlockDriverState *bs; + GLOBAL_STATE_CODE(); + bs = bdrv_find_node(node_name); if (!bs) { error_setg(errp, "Failed to find node with node-name='%s'", node_name); -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> Mark all I/O functions with IO_CODE, and all "I/O OR GS" with IO_OR_GS_CODE. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Message-Id: <20220303151616.325444-14-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block_int-io.h | 6 ++++++ block.c | 14 +++++++++++++- block/block-backend.c | 2 ++ block/dirty-bitmap.c | 3 +++ block/io.c | 13 +++++++++++++ 5 files changed, 37 insertions(+), 1 deletion(-) diff --git a/include/block/block_int-io.h b/include/block/block_int-io.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block_int-io.h +++ b/include/block/block_int-io.h @@ -XXX,XX +XXX,XX @@ static inline int coroutine_fn bdrv_co_pread(BdrvChild *child, int64_t offset, unsigned int bytes, void *buf, BdrvRequestFlags flags) { QEMUIOVector qiov = QEMU_IOVEC_INIT_BUF(qiov, buf, bytes); + IO_CODE(); return bdrv_co_preadv(child, offset, bytes, &qiov, flags); } @@ -XXX,XX +XXX,XX @@ static inline int coroutine_fn bdrv_co_pwrite(BdrvChild *child, int64_t offset, unsigned int bytes, void *buf, BdrvRequestFlags flags) { QEMUIOVector qiov = QEMU_IOVEC_INIT_BUF(qiov, buf, bytes); + IO_CODE(); return bdrv_co_pwritev(child, offset, bytes, &qiov, flags); } @@ -XXX,XX +XXX,XX @@ BlockDriverState *bdrv_backing_chain_next(BlockDriverState *bs); static inline BlockDriverState *bdrv_cow_bs(BlockDriverState *bs) { + IO_CODE(); return child_bs(bdrv_cow_child(bs)); } static inline BlockDriverState *bdrv_filter_bs(BlockDriverState *bs) { + IO_CODE(); return child_bs(bdrv_filter_child(bs)); } static inline BlockDriverState *bdrv_filter_or_cow_bs(BlockDriverState *bs) { + IO_CODE(); return child_bs(bdrv_filter_or_cow_child(bs)); } static inline BlockDriverState *bdrv_primary_bs(BlockDriverState *bs) { + IO_CODE(); return child_bs(bdrv_primary_child(bs)); } diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ BlockDriver *bdrv_probe_all(const uint8_t *buf, int buf_size, { int score_max = 0, score; BlockDriver *drv = NULL, *d; + IO_CODE(); QLIST_FOREACH(d, &bdrv_drivers, list) { if (d->bdrv_probe) { @@ -XXX,XX +XXX,XX @@ static int find_image_format(BlockBackend *file, const char *filename, int refresh_total_sectors(BlockDriverState *bs, int64_t hint) { BlockDriver *drv = bs->drv; + IO_CODE(); if (!drv) { return -ENOMEDIUM; @@ -XXX,XX +XXX,XX @@ const char *bdrv_get_parent_name(const BlockDriverState *bs) { BdrvChild *c; const char *name; + IO_CODE(); /* If multiple parents have a name, just pick the first one. */ QLIST_FOREACH(c, &bs->parents, next_parent) { @@ -XXX,XX +XXX,XX @@ int bdrv_make_empty(BdrvChild *c, Error **errp) */ BdrvChild *bdrv_cow_child(BlockDriverState *bs) { + IO_CODE(); + if (!bs || !bs->drv) { return NULL; } @@ -XXX,XX +XXX,XX @@ BdrvChild *bdrv_cow_child(BlockDriverState *bs) BdrvChild *bdrv_filter_child(BlockDriverState *bs) { BdrvChild *c; + IO_CODE(); if (!bs || !bs->drv) { return NULL; @@ -XXX,XX +XXX,XX @@ BdrvChild *bdrv_filter_or_cow_child(BlockDriverState *bs) { BdrvChild *cow_child = bdrv_cow_child(bs); BdrvChild *filter_child = bdrv_filter_child(bs); + IO_CODE(); /* Filter nodes cannot have COW backing files */ assert(!(cow_child && filter_child)); @@ -XXX,XX +XXX,XX @@ BdrvChild *bdrv_filter_or_cow_child(BlockDriverState *bs) BdrvChild *bdrv_primary_child(BlockDriverState *bs) { BdrvChild *c, *found = NULL; + IO_CODE(); QLIST_FOREACH(c, &bs->children, next) { if (c->role & BDRV_CHILD_PRIMARY) { @@ -XXX,XX +XXX,XX @@ BlockDriverState *bdrv_skip_implicit_filters(BlockDriverState *bs) */ BlockDriverState *bdrv_skip_filters(BlockDriverState *bs) { + IO_CODE(); return bdrv_do_skip_filters(bs, false); } @@ -XXX,XX +XXX,XX @@ BlockDriverState *bdrv_skip_filters(BlockDriverState *bs) */ BlockDriverState *bdrv_backing_chain_next(BlockDriverState *bs) { + IO_CODE(); return bdrv_skip_filters(bdrv_cow_bs(bdrv_skip_filters(bs))); } @@ -XXX,XX +XXX,XX @@ static bool bdrv_bsc_range_overlaps_locked(BlockDriverState *bs, */ bool bdrv_bsc_is_data(BlockDriverState *bs, int64_t offset, int64_t *pnum) { + IO_CODE(); RCU_READ_LOCK_GUARD(); - return bdrv_bsc_range_overlaps_locked(bs, offset, 1, pnum); } @@ -XXX,XX +XXX,XX @@ bool bdrv_bsc_is_data(BlockDriverState *bs, int64_t offset, int64_t *pnum) void bdrv_bsc_invalidate_range(BlockDriverState *bs, int64_t offset, int64_t bytes) { + IO_CODE(); RCU_READ_LOCK_GUARD(); if (bdrv_bsc_range_overlaps_locked(bs, offset, bytes, NULL)) { @@ -XXX,XX +XXX,XX @@ void bdrv_bsc_fill(BlockDriverState *bs, int64_t offset, int64_t bytes) { BdrvBlockStatusCache *new_bsc = g_new(BdrvBlockStatusCache, 1); BdrvBlockStatusCache *old_bsc; + IO_CODE(); *new_bsc = (BdrvBlockStatusCache) { .valid = true, diff --git a/block/block-backend.c b/block/block-backend.c index XXXXXXX..XXXXXXX 100644 --- a/block/block-backend.c +++ b/block/block-backend.c @@ -XXX,XX +XXX,XX @@ bool blk_dev_has_removable_media(BlockBackend *blk) */ bool blk_dev_has_tray(BlockBackend *blk) { + IO_CODE(); return blk->dev_ops && blk->dev_ops->is_tray_open; } @@ -XXX,XX +XXX,XX @@ void blk_dev_eject_request(BlockBackend *blk, bool force) */ bool blk_dev_is_tray_open(BlockBackend *blk) { + IO_CODE(); if (blk_dev_has_tray(blk)) { return blk->dev_ops->is_tray_open(blk->dev_opaque); } diff --git a/block/dirty-bitmap.c b/block/dirty-bitmap.c index XXXXXXX..XXXXXXX 100644 --- a/block/dirty-bitmap.c +++ b/block/dirty-bitmap.c @@ -XXX,XX +XXX,XX @@ void bdrv_reset_dirty_bitmap(BdrvDirtyBitmap *bitmap, void bdrv_clear_dirty_bitmap(BdrvDirtyBitmap *bitmap, HBitmap **out) { + IO_CODE(); assert(!bdrv_dirty_bitmap_readonly(bitmap)); bdrv_dirty_bitmaps_lock(bitmap->bs); if (!out) { @@ -XXX,XX +XXX,XX @@ void bdrv_dirty_bitmap_deserialize_finish(BdrvDirtyBitmap *bitmap) void bdrv_set_dirty(BlockDriverState *bs, int64_t offset, int64_t bytes) { BdrvDirtyBitmap *bitmap; + IO_CODE(); if (QLIST_EMPTY(&bs->dirty_bitmaps)) { return; @@ -XXX,XX +XXX,XX @@ bool bdrv_dirty_bitmap_merge_internal(BdrvDirtyBitmap *dest, bool lock) { bool ret; + IO_CODE(); assert(!bdrv_dirty_bitmap_readonly(dest)); assert(!bdrv_dirty_bitmap_inconsistent(dest)); 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_subtree_drained_end(BlockDriverState *bs) void bdrv_apply_subtree_drain(BdrvChild *child, BlockDriverState *new_parent) { int i; + IO_OR_GS_CODE(); for (i = 0; i < new_parent->recursive_quiesce_counter; i++) { bdrv_do_drained_begin(child->bs, true, child, false, true); @@ -XXX,XX +XXX,XX @@ void bdrv_unapply_subtree_drain(BdrvChild *child, BlockDriverState *old_parent) { int drained_end_counter = 0; int i; + IO_OR_GS_CODE(); for (i = 0; i < old_parent->recursive_quiesce_counter; i++) { bdrv_do_drained_end(child->bs, true, child, false, @@ -XXX,XX +XXX,XX @@ BdrvTrackedRequest *coroutine_fn bdrv_co_get_self_request(BlockDriverState *bs) { BdrvTrackedRequest *req; Coroutine *self = qemu_coroutine_self(); + IO_CODE(); QLIST_FOREACH(req, &bs->tracked_requests, list) { if (req->co == self) { @@ -XXX,XX +XXX,XX @@ static int bdrv_get_cluster_size(BlockDriverState *bs) void bdrv_inc_in_flight(BlockDriverState *bs) { + IO_CODE(); qatomic_inc(&bs->in_flight); } void bdrv_wakeup(BlockDriverState *bs) { + IO_CODE(); aio_wait_kick(); } void bdrv_dec_in_flight(BlockDriverState *bs) { + IO_CODE(); qatomic_dec(&bs->in_flight); bdrv_wakeup(bs); } @@ -XXX,XX +XXX,XX @@ bool coroutine_fn bdrv_make_request_serialising(BdrvTrackedRequest *req, uint64_t align) { bool waited; + IO_CODE(); qemu_co_mutex_lock(&req->bs->reqs_lock); @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_preadv(BdrvChild *child, int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags) { + IO_CODE(); return bdrv_co_preadv_part(child, offset, bytes, qiov, 0, flags); } @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_preadv_part(BdrvChild *child, BdrvTrackedRequest req; BdrvRequestPadding pad; int ret; + IO_CODE(); trace_bdrv_co_preadv_part(bs, offset, bytes, flags); @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_pwritev(BdrvChild *child, int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags) { + IO_CODE(); return bdrv_co_pwritev_part(child, offset, bytes, qiov, 0, flags); } @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_pwritev_part(BdrvChild *child, BdrvRequestPadding pad; int ret; bool padded = false; + IO_CODE(); trace_bdrv_co_pwritev_part(child->bs, offset, bytes, flags); @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_copy_range_from(BdrvChild *src, int64_t src_offset, BdrvRequestFlags read_flags, BdrvRequestFlags write_flags) { + IO_CODE(); trace_bdrv_co_copy_range_from(src, src_offset, dst, dst_offset, bytes, read_flags, write_flags); return bdrv_co_copy_range_internal(src, src_offset, dst, dst_offset, @@ -XXX,XX +XXX,XX @@ int coroutine_fn bdrv_co_copy_range_to(BdrvChild *src, int64_t src_offset, BdrvRequestFlags read_flags, BdrvRequestFlags write_flags) { + IO_CODE(); trace_bdrv_co_copy_range_to(src, src_offset, dst, dst_offset, bytes, read_flags, write_flags); return bdrv_co_copy_range_internal(src, src_offset, dst, dst_offset, -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> We want to be sure that the functions that write the child and parent list of a bs are under BQL and drain. BQL prevents from concurrent writings from the GS API, while drains protect from I/O. TODO: drains are missing in some functions using this assert. Therefore a proper assertion will fail. Because adding drains requires additional discussions, they will be added in future series. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Message-Id: <20220303151616.325444-15-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block_int-global-state.h | 17 +++++++++++++++++ block.c | 4 ++++ 2 files changed, 21 insertions(+) diff --git a/include/block/block_int-global-state.h b/include/block/block_int-global-state.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block_int-global-state.h +++ b/include/block/block_int-global-state.h @@ -XXX,XX +XXX,XX @@ void bdrv_remove_aio_context_notifier(BlockDriverState *bs, */ void bdrv_drain_all_end_quiesce(BlockDriverState *bs); +/** + * Make sure that the function is running under both drain and BQL. + * The latter protects from concurrent writings + * from the GS API, while the former prevents concurrent reads + * from I/O. + */ +static inline void assert_bdrv_graph_writable(BlockDriverState *bs) +{ + /* + * TODO: this function is incomplete. Because the users of this + * assert lack the necessary drains, check only for BQL. + * Once the necessary drains are added, + * assert also for qatomic_read(&bs->quiesce_counter) > 0 + */ + assert(qemu_in_main_thread()); +} + #endif /* BLOCK_INT_GLOBAL_STATE */ 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_attach(BdrvChild *child) { BlockDriverState *bs = child->opaque; + assert_bdrv_graph_writable(bs); QLIST_INSERT_HEAD(&bs->children, child, next); if (child->role & BDRV_CHILD_COW) { @@ -XXX,XX +XXX,XX @@ static void bdrv_child_cb_detach(BdrvChild *child) bdrv_unapply_subtree_drain(child, bs); + assert_bdrv_graph_writable(bs); QLIST_REMOVE(child, next); } @@ -XXX,XX +XXX,XX @@ static void bdrv_replace_child_noperm(BdrvChild **childp, if (child->klass->detach) { child->klass->detach(child); } + assert_bdrv_graph_writable(old_bs); QLIST_REMOVE(child, next_parent); } @@ -XXX,XX +XXX,XX @@ static void bdrv_replace_child_noperm(BdrvChild **childp, } if (new_bs) { + assert_bdrv_graph_writable(new_bs); QLIST_INSERT_HEAD(&new_bs->parents, child, next_parent); /* -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> Since the I/O functions are not many, keep a single file. Also split the function pointers in BlockJobDriver. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Message-Id: <20220303151616.325444-16-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/blockjob_int.h | 28 ++++++++++++++++++++++++++++ 1 file changed, 28 insertions(+) diff --git a/include/block/blockjob_int.h b/include/block/blockjob_int.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/blockjob_int.h +++ b/include/block/blockjob_int.h @@ -XXX,XX +XXX,XX @@ struct BlockJobDriver { /** Generic JobDriver callbacks and settings */ JobDriver job_driver; + /* + * I/O API functions. These functions are thread-safe. + * + * See include/block/block-io.h for more information about + * the I/O API. + */ + /* * Returns whether the job has pending requests for the child or will * submit new requests before the next pause point. This callback is polled @@ -XXX,XX +XXX,XX @@ struct BlockJobDriver { */ bool (*drained_poll)(BlockJob *job); + /* + * Global state (GS) API. These functions run under the BQL. + * + * See include/block/block-global-state.h for more information about + * the GS API. + */ + /* * If the callback is not NULL, it will be invoked before the job is * resumed in a new AioContext. This is the place to move any resources @@ -XXX,XX +XXX,XX @@ struct BlockJobDriver { void (*set_speed)(BlockJob *job, int64_t speed); }; +/* + * Global state (GS) API. These functions run under the BQL. + * + * See include/block/block-global-state.h for more information about + * the GS API. + */ + /** * block_job_create: * @job_id: The id of the newly-created job, or %NULL to have one @@ -XXX,XX +XXX,XX @@ void block_job_free(Job *job); */ void block_job_user_resume(Job *job); +/* + * I/O API functions. These functions are thread-safe. + * + * See include/block/block-io.h for more information about + * the I/O API. + */ + /** * block_job_ratelimit_get_delay: * -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Message-Id: <20220303151616.325444-17-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- blockjob.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/blockjob.c b/blockjob.c index XXXXXXX..XXXXXXX 100644 --- a/blockjob.c +++ b/blockjob.c @@ -XXX,XX +XXX,XX @@ BlockJob *block_job_get(const char *id) void block_job_free(Job *job) { BlockJob *bjob = container_of(job, BlockJob, job); + GLOBAL_STATE_CODE(); block_job_remove_all_bdrv(bjob); ratelimit_destroy(&bjob->limit); @@ -XXX,XX +XXX,XX @@ bool block_job_set_speed(BlockJob *job, int64_t speed, Error **errp) int64_t block_job_ratelimit_get_delay(BlockJob *job, uint64_t n) { + IO_CODE(); return ratelimit_calculate_delay(&job->limit, n); } @@ -XXX,XX +XXX,XX @@ void *block_job_create(const char *job_id, const BlockJobDriver *driver, { BlockJob *job; int ret; + GLOBAL_STATE_CODE(); if (job_id == NULL && !(flags & JOB_INTERNAL)) { job_id = bdrv_get_device_name(bs); @@ -XXX,XX +XXX,XX @@ void block_job_iostatus_reset(BlockJob *job) void block_job_user_resume(Job *job) { BlockJob *bjob = container_of(job, BlockJob, job); + GLOBAL_STATE_CODE(); block_job_iostatus_reset(bjob); } @@ -XXX,XX +XXX,XX @@ BlockErrorAction block_job_error_action(BlockJob *job, BlockdevOnError on_err, int is_read, int error) { BlockErrorAction action; + IO_CODE(); switch (on_err) { case BLOCKDEV_ON_ERROR_ENOSPC: -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> Following the assertion derived from the API split, propagate the assertion also in the static functions. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Message-Id: <20220303151616.325444-18-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block.c | 46 ++++++++++++++++++++++++++++++++++++++++++- block/block-backend.c | 3 +++ 2 files changed, 48 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 @@ BlockDriverState *bdrv_new(void) static BlockDriver *bdrv_do_find_format(const char *format_name) { BlockDriver *drv1; + GLOBAL_STATE_CODE(); QLIST_FOREACH(drv1, &bdrv_drivers, list) { if (!strcmp(drv1->format_name, format_name)) { @@ -XXX,XX +XXX,XX @@ static int64_t create_file_fallback_truncate(BlockBackend *blk, int64_t size; int ret; + GLOBAL_STATE_CODE(); + ret = blk_truncate(blk, minimum_size, false, PREALLOC_MODE_OFF, 0, &local_err); if (ret < 0 && ret != -ENOTSUP) { @@ -XXX,XX +XXX,XX @@ static int create_file_fallback_zero_first_sector(BlockBackend *blk, int64_t bytes_to_clear; int ret; + GLOBAL_STATE_CODE(); + bytes_to_clear = MIN(current_size, BDRV_SECTOR_SIZE); if (bytes_to_clear) { ret = blk_pwrite_zeroes(blk, 0, bytes_to_clear, BDRV_REQ_MAY_UNMAP); @@ -XXX,XX +XXX,XX @@ static BlockDriver *find_hdev_driver(const char *filename) { int score_max = 0, score; BlockDriver *drv = NULL, *d; + GLOBAL_STATE_CODE(); QLIST_FOREACH(d, &bdrv_drivers, list) { if (d->bdrv_probe_device) { @@ -XXX,XX +XXX,XX @@ static BlockDriver *find_hdev_driver(const char *filename) static BlockDriver *bdrv_do_find_protocol(const char *protocol) { BlockDriver *drv1; + GLOBAL_STATE_CODE(); QLIST_FOREACH(drv1, &bdrv_drivers, list) { if (drv1->protocol_name && !strcmp(drv1->protocol_name, protocol)) { @@ -XXX,XX +XXX,XX @@ static int find_image_format(BlockBackend *file, const char *filename, uint8_t buf[BLOCK_PROBE_BUF_SIZE]; int ret = 0; + GLOBAL_STATE_CODE(); + /* Return the raw BlockDriver * to scsi-generic devices or empty drives */ if (blk_is_sg(file) || !blk_is_inserted(file) || blk_getlength(file) == 0) { *pdrv = &bdrv_raw; @@ -XXX,XX +XXX,XX @@ static BlockdevDetectZeroesOptions bdrv_parse_detect_zeroes(QemuOpts *opts, BlockdevDetectZeroesOptions detect_zeroes = qapi_enum_parse(&BlockdevDetectZeroesOptions_lookup, value, BLOCKDEV_DETECT_ZEROES_OPTIONS_OFF, &local_err); + GLOBAL_STATE_CODE(); g_free(value); if (local_err) { error_propagate(errp, local_err); @@ -XXX,XX +XXX,XX @@ static void bdrv_child_cb_drained_end(BdrvChild *child, static int bdrv_child_cb_inactivate(BdrvChild *child) { BlockDriverState *bs = child->opaque; + GLOBAL_STATE_CODE(); assert(bs->open_flags & BDRV_O_INACTIVE); return 0; } @@ -XXX,XX +XXX,XX @@ static void bdrv_child_cb_set_aio_ctx(BdrvChild *child, AioContext *ctx, static void bdrv_temp_snapshot_options(int *child_flags, QDict *child_options, int parent_flags, QDict *parent_options) { + GLOBAL_STATE_CODE(); *child_flags = (parent_flags & ~BDRV_O_SNAPSHOT) | BDRV_O_TEMPORARY; /* For temporary files, unconditional cache=unsafe is fine */ @@ -XXX,XX +XXX,XX @@ static void bdrv_backing_attach(BdrvChild *c) BlockDriverState *parent = c->opaque; BlockDriverState *backing_hd = c->bs; + GLOBAL_STATE_CODE(); assert(!parent->backing_blocker); error_setg(&parent->backing_blocker, "node is used as backing hd of '%s'", @@ -XXX,XX +XXX,XX @@ static void bdrv_backing_detach(BdrvChild *c) { BlockDriverState *parent = c->opaque; + GLOBAL_STATE_CODE(); assert(parent->backing_blocker); bdrv_op_unblock_all(c->bs, parent->backing_blocker); error_free(parent->backing_blocker); @@ -XXX,XX +XXX,XX @@ static int bdrv_backing_update_filename(BdrvChild *c, BlockDriverState *base, BlockDriverState *parent = c->opaque; bool read_only = bdrv_is_read_only(parent); int ret; + GLOBAL_STATE_CODE(); if (read_only) { ret = bdrv_reopen_set_read_only(parent, false, errp); @@ -XXX,XX +XXX,XX @@ static void bdrv_inherited_options(BdrvChildRole role, bool parent_is_format, int parent_flags, QDict *parent_options) { int flags = parent_flags; + GLOBAL_STATE_CODE(); /* * First, decide whether to set, clear, or leave BDRV_O_PROTOCOL. @@ -XXX,XX +XXX,XX @@ AioContext *bdrv_child_get_parent_aio_context(BdrvChild *c) static int bdrv_open_flags(BlockDriverState *bs, int flags) { int open_flags = flags; + GLOBAL_STATE_CODE(); /* * Clear flags that are internal to the block layer before opening the @@ -XXX,XX +XXX,XX @@ static int bdrv_open_flags(BlockDriverState *bs, int flags) static void update_flags_from_options(int *flags, QemuOpts *opts) { + GLOBAL_STATE_CODE(); + *flags &= ~(BDRV_O_CACHE_MASK | BDRV_O_RDWR | BDRV_O_AUTO_RDONLY); if (qemu_opt_get_bool_del(opts, BDRV_OPT_CACHE_NO_FLUSH, false)) { @@ -XXX,XX +XXX,XX @@ static void update_flags_from_options(int *flags, QemuOpts *opts) static void update_options_from_flags(QDict *options, int flags) { + GLOBAL_STATE_CODE(); if (!qdict_haskey(options, BDRV_OPT_CACHE_DIRECT)) { qdict_put_bool(options, BDRV_OPT_CACHE_DIRECT, flags & BDRV_O_NOCACHE); } @@ -XXX,XX +XXX,XX @@ static void bdrv_assign_node_name(BlockDriverState *bs, Error **errp) { char *gen_node_name = NULL; + GLOBAL_STATE_CODE(); if (!node_name) { node_name = gen_node_name = id_generate(ID_BLOCK); @@ -XXX,XX +XXX,XX @@ static int bdrv_open_common(BlockDriverState *bs, BlockBackend *file, assert(bs->file == NULL); assert(options != NULL && bs->options != options); + GLOBAL_STATE_CODE(); opts = qemu_opts_create(&bdrv_runtime_opts, NULL, 0, &error_abort); if (!qemu_opts_absorb_qdict(opts, options, errp)) { @@ -XXX,XX +XXX,XX @@ static QDict *parse_json_filename(const char *filename, Error **errp) QObject *options_obj; QDict *options; int ret; + GLOBAL_STATE_CODE(); ret = strstart(filename, "json:", &filename); assert(ret); @@ -XXX,XX +XXX,XX @@ static void parse_json_protocol(QDict *options, const char **pfilename, { QDict *json_options; Error *local_err = NULL; + GLOBAL_STATE_CODE(); /* Parse json: pseudo-protocol */ if (!*pfilename || !g_str_has_prefix(*pfilename, "json:")) { @@ -XXX,XX +XXX,XX @@ static GSList *bdrv_topological_dfs(GSList *list, GHashTable *found, BdrvChild *child; g_autoptr(GHashTable) local_found = NULL; + GLOBAL_STATE_CODE(); + if (!found) { assert(!list); found = local_found = g_hash_table_new(NULL, NULL); @@ -XXX,XX +XXX,XX @@ typedef struct BdrvReplaceChildState { static void bdrv_replace_child_commit(void *opaque) { BdrvReplaceChildState *s = opaque; + GLOBAL_STATE_CODE(); if (s->free_empty_child && !s->child->bs) { bdrv_child_free(s->child); @@ -XXX,XX +XXX,XX @@ static void bdrv_replace_child_abort(void *opaque) BdrvReplaceChildState *s = opaque; BlockDriverState *new_bs = s->child->bs; + GLOBAL_STATE_CODE(); /* * old_bs reference is transparently moved from @s to s->child. * @@ -XXX,XX +XXX,XX @@ static void bdrv_replace_child_noperm(BdrvChild **childp, static void bdrv_child_free(BdrvChild *child) { assert(!child->bs); + GLOBAL_STATE_CODE(); assert(!child->next.le_prev); /* not in children list */ g_free(child->name); @@ -XXX,XX +XXX,XX @@ static int bdrv_attach_child_common(BlockDriverState *child_bs, assert(child); assert(*child == NULL); assert(child_class->get_parent_desc); + GLOBAL_STATE_CODE(); new_child = g_new(BdrvChild, 1); *new_child = (BdrvChild) { @@ -XXX,XX +XXX,XX @@ static int bdrv_attach_child_noperm(BlockDriverState *parent_bs, uint64_t perm, shared_perm; assert(parent_bs->drv); + GLOBAL_STATE_CODE(); if (bdrv_recurse_has_child(child_bs, parent_bs)) { error_setg(errp, "Making '%s' a %s child of '%s' would create a cycle", @@ -XXX,XX +XXX,XX @@ static void bdrv_detach_child(BdrvChild **childp) { BlockDriverState *old_bs = (*childp)->bs; + GLOBAL_STATE_CODE(); bdrv_replace_child_noperm(childp, NULL, true); if (old_bs) { @@ -XXX,XX +XXX,XX @@ static int bdrv_set_file_or_backing_noperm(BlockDriverState *parent_bs, BdrvChild *child = is_backing ? parent_bs->backing : parent_bs->file; BdrvChildRole role; + GLOBAL_STATE_CODE(); + if (!parent_bs->drv) { /* * Node without drv is an object without a class :/. TODO: finally fix @@ -XXX,XX +XXX,XX @@ static int bdrv_set_backing_noperm(BlockDriverState *bs, BlockDriverState *backing_hd, Transaction *tran, Error **errp) { + GLOBAL_STATE_CODE(); return bdrv_set_file_or_backing_noperm(bs, backing_hd, true, tran, errp); } @@ -XXX,XX +XXX,XX @@ static BlockDriverState *bdrv_append_temp_snapshot(BlockDriverState *bs, BlockDriverState *bs_snapshot = NULL; int ret; + GLOBAL_STATE_CODE(); + /* if snapshot, we create a temporary backing file and open it instead of opening 'filename' directly */ @@ -XXX,XX +XXX,XX @@ static int bdrv_reopen_parse_file_or_backing(BDRVReopenState *reopen_state, QObject *value; const char *str; + GLOBAL_STATE_CODE(); + value = qdict_get(reopen_state->options, child_name); if (value == NULL) { return 0; @@ -XXX,XX +XXX,XX @@ static void bdrv_remove_filter_or_cow_child_abort(void *opaque) static void bdrv_remove_filter_or_cow_child_commit(void *opaque) { BdrvRemoveFilterOrCowChild *s = opaque; - + GLOBAL_STATE_CODE(); bdrv_child_free(s->child); } @@ -XXX,XX +XXX,XX @@ static int bdrv_replace_node_noperm(BlockDriverState *from, BdrvChild *c, *next; assert(to != NULL); + GLOBAL_STATE_CODE(); QLIST_FOREACH_SAFE(c, &from->parents, next_parent, next) { assert(c->bs == from); @@ -XXX,XX +XXX,XX @@ static int bdrv_replace_node_common(BlockDriverState *from, BlockDriverState *to_cow_parent = NULL; int ret; + GLOBAL_STATE_CODE(); assert(to != NULL); if (detach_subchain) { @@ -XXX,XX +XXX,XX @@ void bdrv_debug_event(BlockDriverState *bs, BlkdebugEvent event) static BlockDriverState *bdrv_find_debug_node(BlockDriverState *bs) { + GLOBAL_STATE_CODE(); while (bs && bs->drv && !bs->drv->bdrv_debug_breakpoint) { bs = bdrv_primary_bs(bs); } @@ -XXX,XX +XXX,XX @@ void bdrv_activate_all(Error **errp) static bool bdrv_has_bds_parent(BlockDriverState *bs, bool only_active) { BdrvChild *parent; + GLOBAL_STATE_CODE(); QLIST_FOREACH(parent, &bs->parents, next_parent) { if (parent->klass->parent_is_bds) { @@ -XXX,XX +XXX,XX @@ void bdrv_coroutine_enter(BlockDriverState *bs, Coroutine *co) static void bdrv_do_remove_aio_context_notifier(BdrvAioNotifier *ban) { + GLOBAL_STATE_CODE(); QLIST_REMOVE(ban, list); g_free(ban); } 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 @@ BlockDriverState *blk_bs(BlockBackend *blk) static BlockBackend *bdrv_first_blk(BlockDriverState *bs) { BdrvChild *child; + + GLOBAL_STATE_CODE(); + QLIST_FOREACH(child, &bs->parents, next_parent) { if (child->klass == &child_root) { return child->opaque; -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> blockjob functions run always under the BQL lock. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Message-Id: <20220303151616.325444-19-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/blockjob.h | 29 ++++++++++++++++++++++------- 1 file changed, 22 insertions(+), 7 deletions(-) diff --git a/include/block/blockjob.h b/include/block/blockjob.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/blockjob.h +++ b/include/block/blockjob.h @@ -XXX,XX +XXX,XX @@ typedef struct BlockJob { GSList *nodes; } BlockJob; +/* + * Global state (GS) API. These functions run under the BQL. + * + * See include/block/block-global-state.h for more information about + * the GS API. + */ + /** * block_job_next: * @job: A block job, or %NULL. @@ -XXX,XX +XXX,XX @@ BlockJobInfo *block_job_query(BlockJob *job, Error **errp); */ void block_job_iostatus_reset(BlockJob *job); +/* + * block_job_get_aio_context: + * + * Returns aio context associated with a block job. + */ +AioContext *block_job_get_aio_context(BlockJob *job); + + +/* + * Common functions that are neither I/O nor Global State. + * + * See include/block/block-common.h for more information about + * the Common API. + */ + /** * block_job_is_internal: * @job: The job to determine if it is user-visible or not. @@ -XXX,XX +XXX,XX @@ bool block_job_is_internal(BlockJob *job); */ const BlockJobDriver *block_job_driver(BlockJob *job); -/* - * block_job_get_aio_context: - * - * Returns aio context associated with a block job. - */ -AioContext *block_job_get_aio_context(BlockJob *job); - #endif -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Message-Id: <20220303151616.325444-20-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- blockjob.c | 11 +++++++++++ 1 file changed, 11 insertions(+) diff --git a/blockjob.c b/blockjob.c index XXXXXXX..XXXXXXX 100644 --- a/blockjob.c +++ b/blockjob.c @@ -XXX,XX +XXX,XX @@ static bool is_block_job(Job *job) BlockJob *block_job_next(BlockJob *bjob) { Job *job = bjob ? &bjob->job : NULL; + GLOBAL_STATE_CODE(); do { job = job_next(job); @@ -XXX,XX +XXX,XX @@ BlockJob *block_job_next(BlockJob *bjob) BlockJob *block_job_get(const char *id) { Job *job = job_get(id); + GLOBAL_STATE_CODE(); if (job && is_block_job(job)) { return container_of(job, BlockJob, job); @@ -XXX,XX +XXX,XX @@ static const BdrvChildClass child_job = { void block_job_remove_all_bdrv(BlockJob *job) { + GLOBAL_STATE_CODE(); /* * bdrv_root_unref_child() may reach child_job_[can_]set_aio_ctx(), * which will also traverse job->nodes, so consume the list one by @@ -XXX,XX +XXX,XX @@ void block_job_remove_all_bdrv(BlockJob *job) bool block_job_has_bdrv(BlockJob *job, BlockDriverState *bs) { GSList *el; + GLOBAL_STATE_CODE(); for (el = job->nodes; el; el = el->next) { BdrvChild *c = el->data; @@ -XXX,XX +XXX,XX @@ int block_job_add_bdrv(BlockJob *job, const char *name, BlockDriverState *bs, { BdrvChild *c; bool need_context_ops; + GLOBAL_STATE_CODE(); bdrv_ref(bs); @@ -XXX,XX +XXX,XX @@ bool block_job_set_speed(BlockJob *job, int64_t speed, Error **errp) const BlockJobDriver *drv = block_job_driver(job); int64_t old_speed = job->speed; + GLOBAL_STATE_CODE(); + if (job_apply_verb(&job->job, JOB_VERB_SET_SPEED, errp) < 0) { return false; } @@ -XXX,XX +XXX,XX @@ BlockJobInfo *block_job_query(BlockJob *job, Error **errp) BlockJobInfo *info; uint64_t progress_current, progress_total; + GLOBAL_STATE_CODE(); + if (block_job_is_internal(job)) { error_setg(errp, "Cannot query QEMU internal jobs"); return NULL; @@ -XXX,XX +XXX,XX @@ fail: void block_job_iostatus_reset(BlockJob *job) { + GLOBAL_STATE_CODE(); if (job->iostatus == BLOCK_DEVICE_IO_STATUS_OK) { return; } @@ -XXX,XX +XXX,XX @@ BlockErrorAction block_job_error_action(BlockJob *job, BlockdevOnError on_err, AioContext *block_job_get_aio_context(BlockJob *job) { + GLOBAL_STATE_CODE(); return job->job.aio_context; } -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> blockdev functions run always under the BQL lock. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Message-Id: <20220303151616.325444-21-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/sysemu/blockdev.h | 13 ++++++++++--- 1 file changed, 10 insertions(+), 3 deletions(-) diff --git a/include/sysemu/blockdev.h b/include/sysemu/blockdev.h index XXXXXXX..XXXXXXX 100644 --- a/include/sysemu/blockdev.h +++ b/include/sysemu/blockdev.h @@ -XXX,XX +XXX,XX @@ #include "block/block.h" #include "qemu/queue.h" -void blockdev_mark_auto_del(BlockBackend *blk); -void blockdev_auto_del(BlockBackend *blk); - typedef enum { IF_DEFAULT = -1, /* for use with drive_add() only */ /* @@ -XXX,XX +XXX,XX @@ struct DriveInfo { QTAILQ_ENTRY(DriveInfo) next; }; +/* + * Global state (GS) API. These functions run under the BQL. + * + * See include/block/block-global-state.h for more information about + * the GS API. + */ + +void blockdev_mark_auto_del(BlockBackend *blk); +void blockdev_auto_del(BlockBackend *blk); + DriveInfo *blk_legacy_dinfo(BlockBackend *blk); DriveInfo *blk_set_legacy_dinfo(BlockBackend *blk, DriveInfo *dinfo); BlockBackend *blk_by_legacy_dinfo(DriveInfo *dinfo); -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Message-Id: <20220303151616.325444-22-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/block-backend.c | 3 +++ blockdev.c | 16 ++++++++++++++++ 2 files changed, 19 insertions(+) diff --git a/block/block-backend.c b/block/block-backend.c index XXXXXXX..XXXXXXX 100644 --- a/block/block-backend.c +++ b/block/block-backend.c @@ -XXX,XX +XXX,XX @@ bool bdrv_is_root_node(BlockDriverState *bs) */ DriveInfo *blk_legacy_dinfo(BlockBackend *blk) { + GLOBAL_STATE_CODE(); return blk->legacy_dinfo; } @@ -XXX,XX +XXX,XX @@ DriveInfo *blk_legacy_dinfo(BlockBackend *blk) DriveInfo *blk_set_legacy_dinfo(BlockBackend *blk, DriveInfo *dinfo) { assert(!blk->legacy_dinfo); + GLOBAL_STATE_CODE(); return blk->legacy_dinfo = dinfo; } @@ -XXX,XX +XXX,XX @@ DriveInfo *blk_set_legacy_dinfo(BlockBackend *blk, DriveInfo *dinfo) BlockBackend *blk_by_legacy_dinfo(DriveInfo *dinfo) { BlockBackend *blk = NULL; + GLOBAL_STATE_CODE(); while ((blk = blk_next(blk)) != NULL) { if (blk->legacy_dinfo == dinfo) { diff --git a/blockdev.c b/blockdev.c index XXXXXXX..XXXXXXX 100644 --- a/blockdev.c +++ b/blockdev.c @@ -XXX,XX +XXX,XX @@ void override_max_devs(BlockInterfaceType type, int max_devs) BlockBackend *blk; DriveInfo *dinfo; + GLOBAL_STATE_CODE(); + if (max_devs <= 0) { return; } @@ -XXX,XX +XXX,XX @@ void blockdev_mark_auto_del(BlockBackend *blk) DriveInfo *dinfo = blk_legacy_dinfo(blk); BlockJob *job; + GLOBAL_STATE_CODE(); + if (!dinfo) { return; } @@ -XXX,XX +XXX,XX @@ void blockdev_mark_auto_del(BlockBackend *blk) void blockdev_auto_del(BlockBackend *blk) { DriveInfo *dinfo = blk_legacy_dinfo(blk); + GLOBAL_STATE_CODE(); if (dinfo && dinfo->auto_del) { monitor_remove_blk(blk); @@ -XXX,XX +XXX,XX @@ QemuOpts *drive_add(BlockInterfaceType type, int index, const char *file, { QemuOpts *opts; + GLOBAL_STATE_CODE(); + opts = qemu_opts_parse_noisily(qemu_find_opts("drive"), optstr, false); if (!opts) { return NULL; @@ -XXX,XX +XXX,XX @@ DriveInfo *drive_get(BlockInterfaceType type, int bus, int unit) BlockBackend *blk; DriveInfo *dinfo; + GLOBAL_STATE_CODE(); + for (blk = blk_next(NULL); blk; blk = blk_next(blk)) { dinfo = blk_legacy_dinfo(blk); if (dinfo && dinfo->type == type @@ -XXX,XX +XXX,XX @@ void drive_check_orphaned(void) Location loc; bool orphans = false; + GLOBAL_STATE_CODE(); + for (blk = blk_next(NULL); blk; blk = blk_next(blk)) { dinfo = blk_legacy_dinfo(blk); /* @@ -XXX,XX +XXX,XX @@ void drive_check_orphaned(void) DriveInfo *drive_get_by_index(BlockInterfaceType type, int index) { + GLOBAL_STATE_CODE(); return drive_get(type, drive_index_to_bus_id(type, index), drive_index_to_unit_id(type, index)); @@ -XXX,XX +XXX,XX @@ int drive_get_max_bus(BlockInterfaceType type) BlockBackend *blk; DriveInfo *dinfo; + GLOBAL_STATE_CODE(); + max_bus = -1; for (blk = blk_next(NULL); blk; blk = blk_next(blk)) { dinfo = blk_legacy_dinfo(blk); @@ -XXX,XX +XXX,XX @@ DriveInfo *drive_new(QemuOpts *all_opts, BlockInterfaceType block_default_type, const char *filename; int i; + GLOBAL_STATE_CODE(); + /* Change legacy command line options into QMP ones */ static const struct { const char *from; -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> Snapshots run also under the BQL, so they all are in the global state API. The aiocontext lock that they hold is currently an overkill and in future could be removed. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Message-Id: <20220303151616.325444-23-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/snapshot.h | 13 +++++++++++-- block/snapshot.c | 28 ++++++++++++++++++++++++++++ migration/savevm.c | 2 ++ 3 files changed, 41 insertions(+), 2 deletions(-) diff --git a/include/block/snapshot.h b/include/block/snapshot.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/snapshot.h +++ b/include/block/snapshot.h @@ -XXX,XX +XXX,XX @@ typedef struct QEMUSnapshotInfo { uint64_t icount; /* record/replay step */ } QEMUSnapshotInfo; +/* + * Global state (GS) API. These functions run under the BQL. + * + * See include/block/block-global-state.h for more information about + * the GS API. + */ + int bdrv_snapshot_find(BlockDriverState *bs, QEMUSnapshotInfo *sn_info, const char *name); bool bdrv_snapshot_find_by_id_and_name(BlockDriverState *bs, @@ -XXX,XX +XXX,XX @@ int bdrv_snapshot_load_tmp_by_id_or_name(BlockDriverState *bs, Error **errp); -/* Group operations. All block drivers are involved. +/* + * Group operations. All block drivers are involved. * These functions will properly handle dataplane (take aio_context_acquire - * when appropriate for appropriate block drivers */ + * when appropriate for appropriate block drivers + */ bool bdrv_all_can_snapshot(bool has_devices, strList *devices, Error **errp); diff --git a/block/snapshot.c b/block/snapshot.c index XXXXXXX..XXXXXXX 100644 --- a/block/snapshot.c +++ b/block/snapshot.c @@ -XXX,XX +XXX,XX @@ int bdrv_snapshot_find(BlockDriverState *bs, QEMUSnapshotInfo *sn_info, QEMUSnapshotInfo *sn_tab, *sn; int nb_sns, i, ret; + GLOBAL_STATE_CODE(); + ret = -ENOENT; nb_sns = bdrv_snapshot_list(bs, &sn_tab); if (nb_sns < 0) { @@ -XXX,XX +XXX,XX @@ bool bdrv_snapshot_find_by_id_and_name(BlockDriverState *bs, bool ret = false; assert(id || name); + GLOBAL_STATE_CODE(); nb_sns = bdrv_snapshot_list(bs, &sn_tab); if (nb_sns < 0) { @@ -XXX,XX +XXX,XX @@ static BlockDriverState *bdrv_snapshot_fallback(BlockDriverState *bs) int bdrv_can_snapshot(BlockDriverState *bs) { BlockDriver *drv = bs->drv; + GLOBAL_STATE_CODE(); if (!drv || !bdrv_is_inserted(bs) || bdrv_is_read_only(bs)) { return 0; } @@ -XXX,XX +XXX,XX @@ int bdrv_snapshot_create(BlockDriverState *bs, { BlockDriver *drv = bs->drv; BlockDriverState *fallback_bs = bdrv_snapshot_fallback(bs); + + GLOBAL_STATE_CODE(); + if (!drv) { return -ENOMEDIUM; } @@ -XXX,XX +XXX,XX @@ int bdrv_snapshot_goto(BlockDriverState *bs, BdrvChild **fallback_ptr; int ret, open_ret; + GLOBAL_STATE_CODE(); + if (!drv) { error_setg(errp, "Block driver is closed"); return -ENOMEDIUM; @@ -XXX,XX +XXX,XX @@ int bdrv_snapshot_delete(BlockDriverState *bs, BlockDriverState *fallback_bs = bdrv_snapshot_fallback(bs); int ret; + GLOBAL_STATE_CODE(); + if (!drv) { error_setg(errp, QERR_DEVICE_HAS_NO_MEDIUM, bdrv_get_device_name(bs)); return -ENOMEDIUM; @@ -XXX,XX +XXX,XX @@ int bdrv_snapshot_list(BlockDriverState *bs, { BlockDriver *drv = bs->drv; BlockDriverState *fallback_bs = bdrv_snapshot_fallback(bs); + + GLOBAL_STATE_CODE(); if (!drv) { return -ENOMEDIUM; } @@ -XXX,XX +XXX,XX @@ int bdrv_snapshot_load_tmp(BlockDriverState *bs, { BlockDriver *drv = bs->drv; + GLOBAL_STATE_CODE(); + if (!drv) { error_setg(errp, QERR_DEVICE_HAS_NO_MEDIUM, bdrv_get_device_name(bs)); return -ENOMEDIUM; @@ -XXX,XX +XXX,XX @@ int bdrv_snapshot_load_tmp_by_id_or_name(BlockDriverState *bs, int ret; Error *local_err = NULL; + GLOBAL_STATE_CODE(); + ret = bdrv_snapshot_load_tmp(bs, id_or_name, NULL, &local_err); if (ret == -ENOENT || ret == -EINVAL) { error_free(local_err); @@ -XXX,XX +XXX,XX @@ bool bdrv_all_can_snapshot(bool has_devices, strList *devices, g_autoptr(GList) bdrvs = NULL; GList *iterbdrvs; + GLOBAL_STATE_CODE(); + if (bdrv_all_get_snapshot_devices(has_devices, devices, &bdrvs, errp) < 0) { return false; } @@ -XXX,XX +XXX,XX @@ int bdrv_all_delete_snapshot(const char *name, g_autoptr(GList) bdrvs = NULL; GList *iterbdrvs; + GLOBAL_STATE_CODE(); + if (bdrv_all_get_snapshot_devices(has_devices, devices, &bdrvs, errp) < 0) { return -1; } @@ -XXX,XX +XXX,XX @@ int bdrv_all_goto_snapshot(const char *name, g_autoptr(GList) bdrvs = NULL; GList *iterbdrvs; + GLOBAL_STATE_CODE(); + if (bdrv_all_get_snapshot_devices(has_devices, devices, &bdrvs, errp) < 0) { return -1; } @@ -XXX,XX +XXX,XX @@ int bdrv_all_has_snapshot(const char *name, g_autoptr(GList) bdrvs = NULL; GList *iterbdrvs; + GLOBAL_STATE_CODE(); + if (bdrv_all_get_snapshot_devices(has_devices, devices, &bdrvs, errp) < 0) { return -1; } @@ -XXX,XX +XXX,XX @@ int bdrv_all_create_snapshot(QEMUSnapshotInfo *sn, { g_autoptr(GList) bdrvs = NULL; GList *iterbdrvs; + GLOBAL_STATE_CODE(); if (bdrv_all_get_snapshot_devices(has_devices, devices, &bdrvs, errp) < 0) { return -1; @@ -XXX,XX +XXX,XX @@ BlockDriverState *bdrv_all_find_vmstate_bs(const char *vmstate_bs, g_autoptr(GList) bdrvs = NULL; GList *iterbdrvs; + GLOBAL_STATE_CODE(); + if (bdrv_all_get_snapshot_devices(has_devices, devices, &bdrvs, errp) < 0) { return NULL; } diff --git a/migration/savevm.c b/migration/savevm.c index XXXXXXX..XXXXXXX 100644 --- a/migration/savevm.c +++ b/migration/savevm.c @@ -XXX,XX +XXX,XX @@ bool save_snapshot(const char *name, bool overwrite, const char *vmstate, g_autoptr(GDateTime) now = g_date_time_new_now_local(); AioContext *aio_context; + GLOBAL_STATE_CODE(); + if (migration_is_blocked(errp)) { return false; } -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> copy-before-write functions always run under BQL. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Message-Id: <20220303151616.325444-24-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/copy-before-write.h | 7 +++++++ block/copy-before-write.c | 2 ++ 2 files changed, 9 insertions(+) diff --git a/block/copy-before-write.h b/block/copy-before-write.h index XXXXXXX..XXXXXXX 100644 --- a/block/copy-before-write.h +++ b/block/copy-before-write.h @@ -XXX,XX +XXX,XX @@ #include "block/block_int.h" #include "block/block-copy.h" +/* + * Global state (GS) API. These functions run under the BQL. + * + * See include/block/block-global-state.h for more information about + * the GS API. + */ + BlockDriverState *bdrv_cbw_append(BlockDriverState *source, BlockDriverState *target, const char *filter_node_name, diff --git a/block/copy-before-write.c b/block/copy-before-write.c index XXXXXXX..XXXXXXX 100644 --- a/block/copy-before-write.c +++ b/block/copy-before-write.c @@ -XXX,XX +XXX,XX @@ BlockDriverState *bdrv_cbw_append(BlockDriverState *source, QDict *opts; assert(source->total_sectors == target->total_sectors); + GLOBAL_STATE_CODE(); opts = qdict_new(); qdict_put_str(opts, "driver", "copy-before-write"); @@ -XXX,XX +XXX,XX @@ BlockDriverState *bdrv_cbw_append(BlockDriverState *source, void bdrv_cbw_drop(BlockDriverState *bs) { + GLOBAL_STATE_CODE(); bdrv_drop_filter(bs, &error_abort); bdrv_unref(bs); } -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> block coroutines functions run in different aiocontext, and are not protected by the BQL. Therefore are I/O. On the other side, generated_co_wrapper functions use BDRV_POLL_WHILE, meaning the caller can either be the main loop or a specific iothread. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Message-Id: <20220303151616.325444-25-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/coroutines.h | 81 +++++++++++++++++++++++++++---------------- block.c | 2 ++ block/block-backend.c | 6 ++++ block/io.c | 3 ++ block/nbd.c | 1 + 5 files changed, 64 insertions(+), 29 deletions(-) diff --git a/block/coroutines.h b/block/coroutines.h index XXXXXXX..XXXXXXX 100644 --- a/block/coroutines.h +++ b/block/coroutines.h @@ -XXX,XX +XXX,XX @@ /* For blk_bs() in generated block/block-gen.c */ #include "sysemu/block-backend.h" +/* + * I/O API functions. These functions are thread-safe. + * + * See include/block/block-io.h for more information about + * the I/O API. + */ + int coroutine_fn bdrv_co_check(BlockDriverState *bs, BdrvCheckResult *res, BdrvCheckMode fix); int coroutine_fn bdrv_co_invalidate_cache(BlockDriverState *bs, Error **errp); -int generated_co_wrapper -bdrv_preadv(BdrvChild *child, int64_t offset, unsigned int bytes, - QEMUIOVector *qiov, BdrvRequestFlags flags); -int generated_co_wrapper -bdrv_pwritev(BdrvChild *child, int64_t offset, unsigned int bytes, - QEMUIOVector *qiov, BdrvRequestFlags flags); - int coroutine_fn bdrv_co_common_block_status_above(BlockDriverState *bs, BlockDriverState *base, @@ -XXX,XX +XXX,XX @@ bdrv_co_common_block_status_above(BlockDriverState *bs, int64_t *map, BlockDriverState **file, int *depth); + +int coroutine_fn bdrv_co_readv_vmstate(BlockDriverState *bs, + QEMUIOVector *qiov, int64_t pos); +int coroutine_fn bdrv_co_writev_vmstate(BlockDriverState *bs, + QEMUIOVector *qiov, int64_t pos); + +int coroutine_fn +nbd_co_do_establish_connection(BlockDriverState *bs, Error **errp); + + +int coroutine_fn +blk_co_do_preadv(BlockBackend *blk, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, BdrvRequestFlags flags); + + +int coroutine_fn +blk_co_do_pwritev_part(BlockBackend *blk, int64_t offset, int64_t bytes, + QEMUIOVector *qiov, size_t qiov_offset, + BdrvRequestFlags flags); + +int coroutine_fn +blk_co_do_ioctl(BlockBackend *blk, unsigned long int req, void *buf); + +int coroutine_fn +blk_co_do_pdiscard(BlockBackend *blk, int64_t offset, int64_t bytes); + +int coroutine_fn blk_co_do_flush(BlockBackend *blk); + + +/* + * "I/O or GS" API functions. These functions can run without + * the BQL, but only in one specific iothread/main loop. + * + * See include/block/block-io.h for more information about + * the "I/O or GS" API. + */ + +int generated_co_wrapper +bdrv_preadv(BdrvChild *child, int64_t offset, unsigned int bytes, + QEMUIOVector *qiov, BdrvRequestFlags flags); + +int generated_co_wrapper +bdrv_pwritev(BdrvChild *child, int64_t offset, unsigned int bytes, + QEMUIOVector *qiov, BdrvRequestFlags flags); + int generated_co_wrapper bdrv_common_block_status_above(BlockDriverState *bs, BlockDriverState *base, @@ -XXX,XX +XXX,XX @@ bdrv_common_block_status_above(BlockDriverState *bs, int64_t *map, BlockDriverState **file, int *depth); - -int coroutine_fn bdrv_co_readv_vmstate(BlockDriverState *bs, - QEMUIOVector *qiov, int64_t pos); -int coroutine_fn bdrv_co_writev_vmstate(BlockDriverState *bs, - QEMUIOVector *qiov, int64_t pos); - int generated_co_wrapper nbd_do_establish_connection(BlockDriverState *bs, Error **errp); -int coroutine_fn -nbd_co_do_establish_connection(BlockDriverState *bs, Error **errp); - int generated_co_wrapper blk_do_preadv(BlockBackend *blk, int64_t offset, int64_t bytes, QEMUIOVector *qiov, BdrvRequestFlags flags); -int coroutine_fn -blk_co_do_preadv(BlockBackend *blk, int64_t offset, int64_t bytes, - QEMUIOVector *qiov, BdrvRequestFlags flags); - int generated_co_wrapper blk_do_pwritev_part(BlockBackend *blk, int64_t offset, int64_t bytes, QEMUIOVector *qiov, size_t qiov_offset, BdrvRequestFlags flags); -int coroutine_fn -blk_co_do_pwritev_part(BlockBackend *blk, int64_t offset, int64_t bytes, - QEMUIOVector *qiov, size_t qiov_offset, - BdrvRequestFlags flags); int generated_co_wrapper blk_do_ioctl(BlockBackend *blk, unsigned long int req, void *buf); -int coroutine_fn -blk_co_do_ioctl(BlockBackend *blk, unsigned long int req, void *buf); int generated_co_wrapper blk_do_pdiscard(BlockBackend *blk, int64_t offset, int64_t bytes); -int coroutine_fn -blk_co_do_pdiscard(BlockBackend *blk, int64_t offset, int64_t bytes); int generated_co_wrapper blk_do_flush(BlockBackend *blk); -int coroutine_fn blk_co_do_flush(BlockBackend *blk); #endif /* BLOCK_COROUTINES_INT_H */ diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ fail: int coroutine_fn bdrv_co_check(BlockDriverState *bs, BdrvCheckResult *res, BdrvCheckMode fix) { + IO_CODE(); if (bs->drv == NULL) { return -ENOMEDIUM; } @@ -XXX,XX +XXX,XX @@ int bdrv_activate(BlockDriverState *bs, Error **errp) int coroutine_fn bdrv_co_invalidate_cache(BlockDriverState *bs, Error **errp) { Error *local_err = NULL; + IO_CODE(); assert(!(bs->open_flags & BDRV_O_INACTIVE)); diff --git a/block/block-backend.c b/block/block-backend.c index XXXXXXX..XXXXXXX 100644 --- a/block/block-backend.c +++ b/block/block-backend.c @@ -XXX,XX +XXX,XX @@ blk_co_do_preadv(BlockBackend *blk, int64_t offset, int64_t bytes, { int ret; BlockDriverState *bs; + IO_CODE(); blk_wait_while_drained(blk); @@ -XXX,XX +XXX,XX @@ blk_co_do_pwritev_part(BlockBackend *blk, int64_t offset, int64_t bytes, { int ret; BlockDriverState *bs; + IO_CODE(); blk_wait_while_drained(blk); @@ -XXX,XX +XXX,XX @@ void blk_aio_cancel_async(BlockAIOCB *acb) int coroutine_fn blk_co_do_ioctl(BlockBackend *blk, unsigned long int req, void *buf) { + IO_CODE(); + blk_wait_while_drained(blk); if (!blk_is_available(blk)) { @@ -XXX,XX +XXX,XX @@ int coroutine_fn blk_co_do_pdiscard(BlockBackend *blk, int64_t offset, int64_t bytes) { int ret; + IO_CODE(); blk_wait_while_drained(blk); @@ -XXX,XX +XXX,XX @@ int blk_pdiscard(BlockBackend *blk, int64_t offset, int64_t bytes) int coroutine_fn blk_co_do_flush(BlockBackend *blk) { blk_wait_while_drained(blk); + IO_CODE(); if (!blk_is_available(blk)) { return -ENOMEDIUM; diff --git a/block/io.c b/block/io.c index XXXXXXX..XXXXXXX 100644 --- a/block/io.c +++ b/block/io.c @@ -XXX,XX +XXX,XX @@ bdrv_co_common_block_status_above(BlockDriverState *bs, BlockDriverState *p; int64_t eof = 0; int dummy; + IO_CODE(); assert(!include_base || base); /* Can't include NULL base */ @@ -XXX,XX +XXX,XX @@ bdrv_co_readv_vmstate(BlockDriverState *bs, QEMUIOVector *qiov, int64_t pos) BlockDriver *drv = bs->drv; BlockDriverState *child_bs = bdrv_primary_bs(bs); int ret; + IO_CODE(); ret = bdrv_check_qiov_request(pos, qiov->size, qiov, 0, NULL); if (ret < 0) { @@ -XXX,XX +XXX,XX @@ bdrv_co_writev_vmstate(BlockDriverState *bs, QEMUIOVector *qiov, int64_t pos) BlockDriver *drv = bs->drv; BlockDriverState *child_bs = bdrv_primary_bs(bs); int ret; + IO_CODE(); ret = bdrv_check_qiov_request(pos, qiov->size, qiov, 0, NULL); if (ret < 0) { diff --git a/block/nbd.c b/block/nbd.c index XXXXXXX..XXXXXXX 100644 --- a/block/nbd.c +++ b/block/nbd.c @@ -XXX,XX +XXX,XX @@ int coroutine_fn nbd_co_do_establish_connection(BlockDriverState *bs, BDRVNBDState *s = (BDRVNBDState *)bs->opaque; int ret; bool blocking = nbd_client_connecting_wait(s); + IO_CODE(); assert(!s->ioc); -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> Similar to the header split, also the function pointers in BlockDriver can be split in I/O and global state. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Message-Id: <20220303151616.325444-26-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block_int-common.h | 445 ++++++++++++++++--------------- 1 file changed, 237 insertions(+), 208 deletions(-) diff --git a/include/block/block_int-common.h b/include/block/block_int-common.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block_int-common.h +++ b/include/block/block_int-common.h @@ -XXX,XX +XXX,XX @@ typedef struct BdrvTrackedRequest { struct BlockDriver { + /* + * These fields are initialized when this object is created, + * and are never changed afterwards. + */ + const char *format_name; int instance_size; @@ -XXX,XX +XXX,XX @@ struct BlockDriver { */ bool is_format; + /* + * Drivers not implementing bdrv_parse_filename nor bdrv_open should have + * this field set to true, except ones that are defined only by their + * child's bs. + * An example of the last type will be the quorum block driver. + */ + bool bdrv_needs_filename; + + /* + * Set if a driver can support backing files. This also implies the + * following semantics: + * + * - Return status 0 of .bdrv_co_block_status means that corresponding + * blocks are not allocated in this layer of backing-chain + * - For such (unallocated) blocks, read will: + * - fill buffer with zeros if there is no backing file + * - read from the backing file otherwise, where the block layer + * takes care of reading zeros beyond EOF if backing file is short + */ + bool supports_backing; + + bool has_variable_length; + + /* + * Drivers setting this field must be able to work with just a plain + * filename with '<protocol_name>:' as a prefix, and no other options. + * Options may be extracted from the filename by implementing + * bdrv_parse_filename. + */ + const char *protocol_name; + + /* List of options for creating images, terminated by name == NULL */ + QemuOptsList *create_opts; + + /* List of options for image amend */ + QemuOptsList *amend_opts; + + /* + * If this driver supports reopening images this contains a + * NULL-terminated list of the runtime options that can be + * modified. If an option in this list is unspecified during + * reopen then it _must_ be reset to its default value or return + * an error. + */ + const char *const *mutable_opts; + + /* + * Pointer to a NULL-terminated array of names of strong options + * that can be specified for bdrv_open(). A strong option is one + * that changes the data of a BDS. + * If this pointer is NULL, the array is considered empty. + * "filename" and "driver" are always considered strong. + */ + const char *const *strong_runtime_opts; + + + /* + * Global state (GS) API. These functions run under the BQL. + * + * See include/block/block-global-state.h for more information about + * the GS API. + */ + /* * This function is invoked under BQL before .bdrv_co_amend() * (which in contrast does not necessarily run under the BQL) @@ -XXX,XX +XXX,XX @@ struct BlockDriver { bool (*bdrv_recurse_can_replace)(BlockDriverState *bs, BlockDriverState *to_replace); - int (*bdrv_probe)(const uint8_t *buf, int buf_size, const char *filename); int (*bdrv_probe_device)(const char *filename); /* @@ -XXX,XX +XXX,XX @@ struct BlockDriver { */ void (*bdrv_parse_filename)(const char *filename, QDict *options, Error **errp); - /* - * Drivers not implementing bdrv_parse_filename nor bdrv_open should have - * this field set to true, except ones that are defined only by their - * child's bs. - * An example of the last type will be the quorum block driver. - */ - bool bdrv_needs_filename; - - /* - * Set if a driver can support backing files. This also implies the - * following semantics: - * - * - Return status 0 of .bdrv_co_block_status means that corresponding - * blocks are not allocated in this layer of backing-chain - * - For such (unallocated) blocks, read will: - * - fill buffer with zeros if there is no backing file - * - read from the backing file otherwise, where the block layer - * takes care of reading zeros beyond EOF if backing file is short - */ - bool supports_backing; - /* For handling image reopen for split or non-split files */ + /* For handling image reopen for split or non-split files. */ int (*bdrv_reopen_prepare)(BDRVReopenState *reopen_state, BlockReopenQueue *queue, Error **errp); void (*bdrv_reopen_commit)(BDRVReopenState *reopen_state); @@ -XXX,XX +XXX,XX @@ struct BlockDriver { Error **errp); void (*bdrv_close)(BlockDriverState *bs); - int coroutine_fn (*bdrv_co_create)(BlockdevCreateOptions *opts, Error **errp); int coroutine_fn (*bdrv_co_create_opts)(BlockDriver *drv, @@ -XXX,XX +XXX,XX @@ struct BlockDriver { QemuOpts *opts, Error **errp); - int coroutine_fn (*bdrv_co_amend)(BlockDriverState *bs, - BlockdevAmendOptions *opts, - bool force, - Error **errp); - int (*bdrv_amend_options)(BlockDriverState *bs, QemuOpts *opts, BlockDriverAmendStatusCB *status_cb, @@ -XXX,XX +XXX,XX @@ struct BlockDriver { */ char *(*bdrv_dirname)(BlockDriverState *bs, Error **errp); + /* + * This informs the driver that we are no longer interested in the result + * of in-flight requests, so don't waste the time if possible. + * + * One example usage is to avoid waiting for an nbd target node reconnect + * timeout during job-cancel with force=true. + */ + void (*bdrv_cancel_in_flight)(BlockDriverState *bs); + + int (*bdrv_inactivate)(BlockDriverState *bs); + + int (*bdrv_snapshot_create)(BlockDriverState *bs, + QEMUSnapshotInfo *sn_info); + int (*bdrv_snapshot_goto)(BlockDriverState *bs, + const char *snapshot_id); + int (*bdrv_snapshot_delete)(BlockDriverState *bs, + const char *snapshot_id, + const char *name, + Error **errp); + int (*bdrv_snapshot_list)(BlockDriverState *bs, + QEMUSnapshotInfo **psn_info); + int (*bdrv_snapshot_load_tmp)(BlockDriverState *bs, + const char *snapshot_id, + const char *name, + Error **errp); + + int (*bdrv_change_backing_file)(BlockDriverState *bs, + const char *backing_file, const char *backing_fmt); + + /* TODO Better pass a option string/QDict/QemuOpts to add any rule? */ + int (*bdrv_debug_breakpoint)(BlockDriverState *bs, const char *event, + const char *tag); + int (*bdrv_debug_remove_breakpoint)(BlockDriverState *bs, + const char *tag); + int (*bdrv_debug_resume)(BlockDriverState *bs, const char *tag); + bool (*bdrv_debug_is_suspended)(BlockDriverState *bs, const char *tag); + + void (*bdrv_refresh_limits)(BlockDriverState *bs, Error **errp); + + /* + * Returns 1 if newly created images are guaranteed to contain only + * zeros, 0 otherwise. + */ + int (*bdrv_has_zero_init)(BlockDriverState *bs); + + /* + * Remove fd handlers, timers, and other event loop callbacks so the event + * loop is no longer in use. Called with no in-flight requests and in + * depth-first traversal order with parents before child nodes. + */ + void (*bdrv_detach_aio_context)(BlockDriverState *bs); + + /* + * Add fd handlers, timers, and other event loop callbacks so I/O requests + * can be processed again. Called with no in-flight requests and in + * depth-first traversal order with child nodes before parent nodes. + */ + void (*bdrv_attach_aio_context)(BlockDriverState *bs, + AioContext *new_context); + + /** + * Try to get @bs's logical and physical block size. + * On success, store them in @bsz and return zero. + * On failure, return negative errno. + */ + int (*bdrv_probe_blocksizes)(BlockDriverState *bs, BlockSizes *bsz); + /** + * Try to get @bs's geometry (cyls, heads, sectors) + * On success, store them in @geo and return 0. + * On failure return -errno. + * Only drivers that want to override guest geometry implement this + * callback; see hd_geometry_guess(). + */ + int (*bdrv_probe_geometry)(BlockDriverState *bs, HDGeometry *geo); + + void (*bdrv_add_child)(BlockDriverState *parent, BlockDriverState *child, + Error **errp); + void (*bdrv_del_child)(BlockDriverState *parent, BdrvChild *child, + Error **errp); + + /** + * Informs the block driver that a permission change is intended. The + * driver checks whether the change is permissible and may take other + * preparations for the change (e.g. get file system locks). This operation + * is always followed either by a call to either .bdrv_set_perm or + * .bdrv_abort_perm_update. + * + * Checks whether the requested set of cumulative permissions in @perm + * can be granted for accessing @bs and whether no other users are using + * permissions other than those given in @shared (both arguments take + * BLK_PERM_* bitmasks). + * + * If both conditions are met, 0 is returned. Otherwise, -errno is returned + * and errp is set to an error describing the conflict. + */ + int (*bdrv_check_perm)(BlockDriverState *bs, uint64_t perm, + uint64_t shared, Error **errp); + + /** + * Called to inform the driver that the set of cumulative set of used + * permissions for @bs has changed to @perm, and the set of sharable + * permission to @shared. The driver can use this to propagate changes to + * its children (i.e. request permissions only if a parent actually needs + * them). + * + * This function is only invoked after bdrv_check_perm(), so block drivers + * may rely on preparations made in their .bdrv_check_perm implementation. + */ + void (*bdrv_set_perm)(BlockDriverState *bs, uint64_t perm, uint64_t shared); + + /* + * Called to inform the driver that after a previous bdrv_check_perm() + * call, the permission update is not performed and any preparations made + * for it (e.g. taken file locks) need to be undone. + * + * This function can be called even for nodes that never saw a + * bdrv_check_perm() call. It is a no-op then. + */ + void (*bdrv_abort_perm_update)(BlockDriverState *bs); + + /** + * Returns in @nperm and @nshared the permissions that the driver for @bs + * needs on its child @c, based on the cumulative permissions requested by + * the parents in @parent_perm and @parent_shared. + * + * If @c is NULL, return the permissions for attaching a new child for the + * given @child_class and @role. + * + * If @reopen_queue is non-NULL, don't return the currently needed + * permissions, but those that will be needed after applying the + * @reopen_queue. + */ + void (*bdrv_child_perm)(BlockDriverState *bs, BdrvChild *c, + BdrvChildRole role, + BlockReopenQueue *reopen_queue, + uint64_t parent_perm, uint64_t parent_shared, + uint64_t *nperm, uint64_t *nshared); + + /** + * Register/unregister a buffer for I/O. For example, when the driver is + * interested to know the memory areas that will later be used in iovs, so + * that it can do IOMMU mapping with VFIO etc., in order to get better + * performance. In the case of VFIO drivers, this callback is used to do + * DMA mapping for hot buffers. + */ + void (*bdrv_register_buf)(BlockDriverState *bs, void *host, size_t size); + void (*bdrv_unregister_buf)(BlockDriverState *bs, void *host); + + /* + * This field is modified only under the BQL, and is part of + * the global state. + */ + QLIST_ENTRY(BlockDriver) list; + + /* + * I/O API functions. These functions are thread-safe. + * + * See include/block/block-io.h for more information about + * the I/O API. + */ + + int (*bdrv_probe)(const uint8_t *buf, int buf_size, const char *filename); + + int coroutine_fn (*bdrv_co_amend)(BlockDriverState *bs, + BlockdevAmendOptions *opts, + bool force, + Error **errp); + /* aio */ BlockAIOCB *(*bdrv_aio_preadv)(BlockDriverState *bs, int64_t offset, int64_t bytes, QEMUIOVector *qiov, @@ -XXX,XX +XXX,XX @@ struct BlockDriver { bool want_zero, int64_t offset, int64_t bytes, int64_t *pnum, int64_t *map, BlockDriverState **file); - /* - * This informs the driver that we are no longer interested in the result - * of in-flight requests, so don't waste the time if possible. - * - * One example usage is to avoid waiting for an nbd target node reconnect - * timeout during job-cancel with force=true. - */ - void (*bdrv_cancel_in_flight)(BlockDriverState *bs); - /* * Invalidate any cached meta-data. */ void coroutine_fn (*bdrv_co_invalidate_cache)(BlockDriverState *bs, Error **errp); - int (*bdrv_inactivate)(BlockDriverState *bs); /* * Flushes all data for all layers by calling bdrv_co_flush for underlying @@ -XXX,XX +XXX,XX @@ struct BlockDriver { */ int coroutine_fn (*bdrv_co_flush_to_os)(BlockDriverState *bs); - /* - * Drivers setting this field must be able to work with just a plain - * filename with '<protocol_name>:' as a prefix, and no other options. - * Options may be extracted from the filename by implementing - * bdrv_parse_filename. - */ - const char *protocol_name; - /* * Truncate @bs to @offset bytes using the given @prealloc mode * when growing. Modes other than PREALLOC_MODE_OFF should be @@ -XXX,XX +XXX,XX @@ struct BlockDriver { bool exact, PreallocMode prealloc, BdrvRequestFlags flags, Error **errp); int64_t (*bdrv_getlength)(BlockDriverState *bs); - bool has_variable_length; int64_t (*bdrv_get_allocated_file_size)(BlockDriverState *bs); BlockMeasureInfo *(*bdrv_measure)(QemuOpts *opts, BlockDriverState *in_bs, Error **errp); @@ -XXX,XX +XXX,XX @@ struct BlockDriver { int64_t offset, int64_t bytes, QEMUIOVector *qiov, size_t qiov_offset); - int (*bdrv_snapshot_create)(BlockDriverState *bs, - QEMUSnapshotInfo *sn_info); - int (*bdrv_snapshot_goto)(BlockDriverState *bs, - const char *snapshot_id); - int (*bdrv_snapshot_delete)(BlockDriverState *bs, - const char *snapshot_id, - const char *name, - Error **errp); - int (*bdrv_snapshot_list)(BlockDriverState *bs, - QEMUSnapshotInfo **psn_info); - int (*bdrv_snapshot_load_tmp)(BlockDriverState *bs, - const char *snapshot_id, - const char *name, - Error **errp); int (*bdrv_get_info)(BlockDriverState *bs, BlockDriverInfo *bdi); ImageInfoSpecific *(*bdrv_get_specific_info)(BlockDriverState *bs, @@ -XXX,XX +XXX,XX @@ struct BlockDriver { QEMUIOVector *qiov, int64_t pos); - int (*bdrv_change_backing_file)(BlockDriverState *bs, - const char *backing_file, const char *backing_fmt); - /* removable device specific */ bool (*bdrv_is_inserted)(BlockDriverState *bs); void (*bdrv_eject)(BlockDriverState *bs, bool eject_flag); @@ -XXX,XX +XXX,XX @@ struct BlockDriver { int coroutine_fn (*bdrv_co_ioctl)(BlockDriverState *bs, unsigned long int req, void *buf); - /* List of options for creating images, terminated by name == NULL */ - QemuOptsList *create_opts; - - /* List of options for image amend */ - QemuOptsList *amend_opts; - - /* - * If this driver supports reopening images this contains a - * NULL-terminated list of the runtime options that can be - * modified. If an option in this list is unspecified during - * reopen then it _must_ be reset to its default value or return - * an error. - */ - const char *const *mutable_opts; - /* * Returns 0 for completed check, -errno for internal errors. * The check results are stored in result. @@ -XXX,XX +XXX,XX @@ struct BlockDriver { void (*bdrv_debug_event)(BlockDriverState *bs, BlkdebugEvent event); - /* TODO Better pass a option string/QDict/QemuOpts to add any rule? */ - int (*bdrv_debug_breakpoint)(BlockDriverState *bs, const char *event, - const char *tag); - int (*bdrv_debug_remove_breakpoint)(BlockDriverState *bs, - const char *tag); - int (*bdrv_debug_resume)(BlockDriverState *bs, const char *tag); - bool (*bdrv_debug_is_suspended)(BlockDriverState *bs, const char *tag); - - void (*bdrv_refresh_limits)(BlockDriverState *bs, Error **errp); - - /* - * Returns 1 if newly created images are guaranteed to contain only - * zeros, 0 otherwise. - */ - int (*bdrv_has_zero_init)(BlockDriverState *bs); - - /* - * Remove fd handlers, timers, and other event loop callbacks so the event - * loop is no longer in use. Called with no in-flight requests and in - * depth-first traversal order with parents before child nodes. - */ - void (*bdrv_detach_aio_context)(BlockDriverState *bs); - - /* - * Add fd handlers, timers, and other event loop callbacks so I/O requests - * can be processed again. Called with no in-flight requests and in - * depth-first traversal order with child nodes before parent nodes. - */ - void (*bdrv_attach_aio_context)(BlockDriverState *bs, - AioContext *new_context); - /* io queue for linux-aio */ void (*bdrv_io_plug)(BlockDriverState *bs); void (*bdrv_io_unplug)(BlockDriverState *bs); - /** - * Try to get @bs's logical and physical block size. - * On success, store them in @bsz and return zero. - * On failure, return negative errno. - */ - int (*bdrv_probe_blocksizes)(BlockDriverState *bs, BlockSizes *bsz); - /** - * Try to get @bs's geometry (cyls, heads, sectors) - * On success, store them in @geo and return 0. - * On failure return -errno. - * Only drivers that want to override guest geometry implement this - * callback; see hd_geometry_guess(). - */ - int (*bdrv_probe_geometry)(BlockDriverState *bs, HDGeometry *geo); - /** * bdrv_co_drain_begin is called if implemented in the beginning of a * drain operation to drain and stop any internal sources of requests in @@ -XXX,XX +XXX,XX @@ struct BlockDriver { void coroutine_fn (*bdrv_co_drain_begin)(BlockDriverState *bs); void coroutine_fn (*bdrv_co_drain_end)(BlockDriverState *bs); - void (*bdrv_add_child)(BlockDriverState *parent, BlockDriverState *child, - Error **errp); - void (*bdrv_del_child)(BlockDriverState *parent, BdrvChild *child, - Error **errp); - - /** - * Informs the block driver that a permission change is intended. The - * driver checks whether the change is permissible and may take other - * preparations for the change (e.g. get file system locks). This operation - * is always followed either by a call to either .bdrv_set_perm or - * .bdrv_abort_perm_update. - * - * Checks whether the requested set of cumulative permissions in @perm - * can be granted for accessing @bs and whether no other users are using - * permissions other than those given in @shared (both arguments take - * BLK_PERM_* bitmasks). - * - * If both conditions are met, 0 is returned. Otherwise, -errno is returned - * and errp is set to an error describing the conflict. - */ - int (*bdrv_check_perm)(BlockDriverState *bs, uint64_t perm, - uint64_t shared, Error **errp); - - /** - * Called to inform the driver that the set of cumulative set of used - * permissions for @bs has changed to @perm, and the set of sharable - * permission to @shared. The driver can use this to propagate changes to - * its children (i.e. request permissions only if a parent actually needs - * them). - * - * This function is only invoked after bdrv_check_perm(), so block drivers - * may rely on preparations made in their .bdrv_check_perm implementation. - */ - void (*bdrv_set_perm)(BlockDriverState *bs, uint64_t perm, uint64_t shared); - - /* - * Called to inform the driver that after a previous bdrv_check_perm() - * call, the permission update is not performed and any preparations made - * for it (e.g. taken file locks) need to be undone. - * - * This function can be called even for nodes that never saw a - * bdrv_check_perm() call. It is a no-op then. - */ - void (*bdrv_abort_perm_update)(BlockDriverState *bs); - - /** - * Returns in @nperm and @nshared the permissions that the driver for @bs - * needs on its child @c, based on the cumulative permissions requested by - * the parents in @parent_perm and @parent_shared. - * - * If @c is NULL, return the permissions for attaching a new child for the - * given @child_class and @role. - * - * If @reopen_queue is non-NULL, don't return the currently needed - * permissions, but those that will be needed after applying the - * @reopen_queue. - */ - void (*bdrv_child_perm)(BlockDriverState *bs, BdrvChild *c, - BdrvChildRole role, - BlockReopenQueue *reopen_queue, - uint64_t parent_perm, uint64_t parent_shared, - uint64_t *nperm, uint64_t *nshared); - bool (*bdrv_supports_persistent_dirty_bitmap)(BlockDriverState *bs); bool (*bdrv_co_can_store_new_dirty_bitmap)(BlockDriverState *bs, const char *name, @@ -XXX,XX +XXX,XX @@ struct BlockDriver { int (*bdrv_co_remove_persistent_dirty_bitmap)(BlockDriverState *bs, const char *name, Error **errp); - - /** - * Register/unregister a buffer for I/O. For example, when the driver is - * interested to know the memory areas that will later be used in iovs, so - * that it can do IOMMU mapping with VFIO etc., in order to get better - * performance. In the case of VFIO drivers, this callback is used to do - * DMA mapping for hot buffers. - */ - void (*bdrv_register_buf)(BlockDriverState *bs, void *host, size_t size); - void (*bdrv_unregister_buf)(BlockDriverState *bs, void *host); - QLIST_ENTRY(BlockDriver) list; - - /* - * Pointer to a NULL-terminated array of names of strong options - * that can be specified for bdrv_open(). A strong option is one - * that changes the data of a BDS. - * If this pointer is NULL, the array is considered empty. - * "filename" and "driver" are always considered strong. - */ - const char *const *strong_runtime_opts; }; static inline bool block_driver_can_compress(BlockDriver *drv) -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Message-Id: <20220303151616.325444-27-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block.c | 17 +++++++++++++++++ block/create.c | 2 ++ 2 files changed, 19 insertions(+) diff --git a/block.c b/block.c index XXXXXXX..XXXXXXX 100644 --- a/block.c +++ b/block.c @@ -XXX,XX +XXX,XX @@ static void coroutine_fn bdrv_create_co_entry(void *opaque) CreateCo *cco = opaque; assert(cco->drv); + GLOBAL_STATE_CODE(); ret = cco->drv->bdrv_co_create_opts(cco->drv, cco->filename, cco->opts, &local_err); @@ -XXX,XX +XXX,XX @@ int refresh_total_sectors(BlockDriverState *bs, int64_t hint) static void bdrv_join_options(BlockDriverState *bs, QDict *options, QDict *old_options) { + GLOBAL_STATE_CODE(); if (bs->drv && bs->drv->bdrv_join_options) { bs->drv->bdrv_join_options(options, old_options); } else { @@ -XXX,XX +XXX,XX @@ static int bdrv_open_driver(BlockDriverState *bs, BlockDriver *drv, { Error *local_err = NULL; int i, ret; + GLOBAL_STATE_CODE(); bdrv_assign_node_name(bs, node_name, &local_err); if (local_err) { @@ -XXX,XX +XXX,XX @@ static int bdrv_fill_options(QDict **options, const char *filename, BlockDriver *drv = NULL; Error *local_err = NULL; + GLOBAL_STATE_CODE(); + /* * Caution: while qdict_get_try_str() is fine, getting non-string * types would require more care. When @options come from @@ -XXX,XX +XXX,XX @@ static void bdrv_child_perm(BlockDriverState *bs, BlockDriverState *child_bs, uint64_t *nperm, uint64_t *nshared) { assert(bs->drv && bs->drv->bdrv_child_perm); + GLOBAL_STATE_CODE(); bs->drv->bdrv_child_perm(bs, c, role, reopen_queue, parent_perm, parent_shared, nperm, nshared); @@ -XXX,XX +XXX,XX @@ static void bdrv_drv_set_perm_commit(void *opaque) { BlockDriverState *bs = opaque; uint64_t cumulative_perms, cumulative_shared_perms; + GLOBAL_STATE_CODE(); if (bs->drv->bdrv_set_perm) { bdrv_get_cumulative_perm(bs, &cumulative_perms, @@ -XXX,XX +XXX,XX @@ static void bdrv_drv_set_perm_commit(void *opaque) static void bdrv_drv_set_perm_abort(void *opaque) { BlockDriverState *bs = opaque; + GLOBAL_STATE_CODE(); if (bs->drv->bdrv_abort_perm_update) { bs->drv->bdrv_abort_perm_update(bs); @@ -XXX,XX +XXX,XX @@ static int bdrv_drv_set_perm(BlockDriverState *bs, uint64_t perm, uint64_t shared_perm, Transaction *tran, Error **errp) { + GLOBAL_STATE_CODE(); if (!bs->drv) { return 0; } @@ -XXX,XX +XXX,XX @@ int bdrv_reopen_multiple(BlockReopenQueue *bs_queue, Error **errp) assert(qemu_get_current_aio_context() == qemu_get_aio_context()); assert(bs_queue != NULL); + GLOBAL_STATE_CODE(); QTAILQ_FOREACH(bs_entry, bs_queue, entry) { ctx = bdrv_get_aio_context(bs_entry->state.bs); @@ -XXX,XX +XXX,XX @@ static int bdrv_reopen_prepare(BDRVReopenState *reopen_state, assert(reopen_state != NULL); assert(reopen_state->bs->drv != NULL); + GLOBAL_STATE_CODE(); drv = reopen_state->bs->drv; /* This function and each driver's bdrv_reopen_prepare() remove @@ -XXX,XX +XXX,XX @@ static void bdrv_reopen_commit(BDRVReopenState *reopen_state) bs = reopen_state->bs; drv = bs->drv; assert(drv != NULL); + GLOBAL_STATE_CODE(); /* If there are any driver level actions to take */ if (drv->bdrv_reopen_commit) { @@ -XXX,XX +XXX,XX @@ static void bdrv_reopen_abort(BDRVReopenState *reopen_state) assert(reopen_state != NULL); drv = reopen_state->bs->drv; assert(drv != NULL); + GLOBAL_STATE_CODE(); if (drv->bdrv_reopen_abort) { drv->bdrv_reopen_abort(reopen_state); @@ -XXX,XX +XXX,XX @@ static int bdrv_inactivate_recurse(BlockDriverState *bs) int ret; uint64_t cumulative_perms, cumulative_shared_perms; + GLOBAL_STATE_CODE(); + if (!bs->drv) { return -ENOMEDIUM; } @@ -XXX,XX +XXX,XX @@ static void bdrv_detach_aio_context(BlockDriverState *bs) BdrvAioNotifier *baf, *baf_tmp; assert(!bs->walking_aio_notifiers); + GLOBAL_STATE_CODE(); bs->walking_aio_notifiers = true; QLIST_FOREACH_SAFE(baf, &bs->aio_notifiers, list, baf_tmp) { if (baf->deleted) { @@ -XXX,XX +XXX,XX @@ static void bdrv_attach_aio_context(BlockDriverState *bs, AioContext *new_context) { BdrvAioNotifier *ban, *ban_tmp; + GLOBAL_STATE_CODE(); if (bs->quiesce_counter) { aio_disable_external(new_context); diff --git a/block/create.c b/block/create.c index XXXXXXX..XXXXXXX 100644 --- a/block/create.c +++ b/block/create.c @@ -XXX,XX +XXX,XX @@ static int coroutine_fn blockdev_create_run(Job *job, Error **errp) BlockdevCreateJob *s = container_of(job, BlockdevCreateJob, common); int ret; + GLOBAL_STATE_CODE(); + job_progress_set_remaining(&s->common, 1); ret = s->drv->bdrv_co_create(s->opts, errp); job_progress_update(&s->common, 1); -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Message-Id: <20220303151616.325444-28-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/block/block_int-common.h | 81 ++++++++++++++++++-------------- 1 file changed, 47 insertions(+), 34 deletions(-) diff --git a/include/block/block_int-common.h b/include/block/block_int-common.h index XXXXXXX..XXXXXXX 100644 --- a/include/block/block_int-common.h +++ b/include/block/block_int-common.h @@ -XXX,XX +XXX,XX @@ struct BdrvChildClass { */ bool parent_is_bds; + /* + * Global state (GS) API. These functions run under the BQL. + * + * See include/block/block-global-state.h for more information about + * the GS API. + */ void (*inherit_options)(BdrvChildRole role, bool parent_is_format, int *child_flags, QDict *child_options, int parent_flags, QDict *parent_options); - void (*change_media)(BdrvChild *child, bool load); - void (*resize)(BdrvChild *child); - - /* - * Returns a name that is supposedly more useful for human users than the - * node name for identifying the node in question (in particular, a BB - * name), or NULL if the parent can't provide a better name. - */ - const char *(*get_name)(BdrvChild *child); /* * Returns a malloced string that describes the parent of the child for a @@ -XXX,XX +XXX,XX @@ struct BdrvChildClass { */ char *(*get_parent_desc)(BdrvChild *child); + /* + * Notifies the parent that the child has been activated/inactivated (e.g. + * when migration is completing) and it can start/stop requesting + * permissions and doing I/O on it. + */ + void (*activate)(BdrvChild *child, Error **errp); + int (*inactivate)(BdrvChild *child); + + void (*attach)(BdrvChild *child); + void (*detach)(BdrvChild *child); + + /* + * Notifies the parent that the filename of its child has changed (e.g. + * because the direct child was removed from the backing chain), so that it + * can update its reference. + */ + int (*update_filename)(BdrvChild *child, BlockDriverState *new_base, + const char *filename, Error **errp); + + bool (*can_set_aio_ctx)(BdrvChild *child, AioContext *ctx, + GSList **ignore, Error **errp); + void (*set_aio_ctx)(BdrvChild *child, AioContext *ctx, GSList **ignore); + + AioContext *(*get_parent_aio_context)(BdrvChild *child); + + /* + * I/O API functions. These functions are thread-safe. + * + * See include/block/block-io.h for more information about + * the I/O API. + */ + + void (*resize)(BdrvChild *child); + + /* + * Returns a name that is supposedly more useful for human users than the + * node name for identifying the node in question (in particular, a BB + * name), or NULL if the parent can't provide a better name. + */ + const char *(*get_name)(BdrvChild *child); + /* * If this pair of functions is implemented, the parent doesn't issue new * requests after returning from .drained_begin() until .drained_end() is @@ -XXX,XX +XXX,XX @@ struct BdrvChildClass { * activity on the child has stopped. */ bool (*drained_poll)(BdrvChild *child); - - /* - * Notifies the parent that the child has been activated/inactivated (e.g. - * when migration is completing) and it can start/stop requesting - * permissions and doing I/O on it. - */ - void (*activate)(BdrvChild *child, Error **errp); - int (*inactivate)(BdrvChild *child); - - void (*attach)(BdrvChild *child); - void (*detach)(BdrvChild *child); - - /* - * Notifies the parent that the filename of its child has changed (e.g. - * because the direct child was removed from the backing chain), so that it - * can update its reference. - */ - int (*update_filename)(BdrvChild *child, BlockDriverState *new_base, - const char *filename, Error **errp); - - bool (*can_set_aio_ctx)(BdrvChild *child, AioContext *ctx, - GSList **ignore, Error **errp); - void (*set_aio_ctx)(BdrvChild *child, AioContext *ctx, GSList **ignore); - - AioContext *(*get_parent_aio_context)(BdrvChild *child); }; extern const BdrvChildClass child_of_bds; -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Message-Id: <20220303151616.325444-29-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block.c | 10 +++++++++- 1 file changed, 9 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 @@ const BdrvChildClass child_of_bds = { AioContext *bdrv_child_get_parent_aio_context(BdrvChild *c) { - IO_CODE(); + GLOBAL_STATE_CODE(); return c->klass->get_parent_aio_context(c); } @@ -XXX,XX +XXX,XX @@ bool bdrv_is_writable(BlockDriverState *bs) static char *bdrv_child_user_desc(BdrvChild *c) { + GLOBAL_STATE_CODE(); return c->klass->get_parent_desc(c); } @@ -XXX,XX +XXX,XX @@ static void bdrv_replace_child_noperm(BdrvChild **childp, assert(!child->frozen); assert(old_bs != new_bs); + GLOBAL_STATE_CODE(); if (old_bs && new_bs) { assert(bdrv_get_aio_context(old_bs) == bdrv_get_aio_context(new_bs)); @@ -XXX,XX +XXX,XX @@ static void bdrv_attach_child_common_abort(void *opaque) BdrvChild *child = *s->child; BlockDriverState *bs = child->bs; + GLOBAL_STATE_CODE(); /* * Pass free_empty_child=false, because we still need the child * for the AioContext operations on the parent below; those @@ -XXX,XX +XXX,XX @@ void bdrv_unref_child(BlockDriverState *parent, BdrvChild *child) static void bdrv_parent_cb_change_media(BlockDriverState *bs, bool load) { BdrvChild *c; + GLOBAL_STATE_CODE(); QLIST_FOREACH(c, &bs->parents, next_parent) { if (c->klass->change_media) { c->klass->change_media(c, load); @@ -XXX,XX +XXX,XX @@ static BlockDriverState *bdrv_open_inherit(const char *filename, assert(!child_class || !flags); assert(!child_class == !parent); + GLOBAL_STATE_CODE(); if (reference) { bool options_non_empty = options ? qdict_size(options) : false; @@ -XXX,XX +XXX,XX @@ static BlockReopenQueue *bdrv_reopen_queue_child(BlockReopenQueue *bs_queue, * important to avoid graph changes between the recursive queuing here and * bdrv_reopen_multiple(). */ assert(bs->quiesce_counter > 0); + GLOBAL_STATE_CODE(); if (bs_queue == NULL) { bs_queue = g_new0(BlockReopenQueue, 1); @@ -XXX,XX +XXX,XX @@ void bdrv_set_aio_context_ignore(BlockDriverState *bs, BdrvChild *child, *parent; g_assert(qemu_get_current_aio_context() == qemu_get_aio_context()); + GLOBAL_STATE_CODE(); if (old_context == new_context) { return; @@ -XXX,XX +XXX,XX @@ void bdrv_set_aio_context_ignore(BlockDriverState *bs, static bool bdrv_parent_can_set_aio_context(BdrvChild *c, AioContext *ctx, GSList **ignore, Error **errp) { + GLOBAL_STATE_CODE(); if (g_slist_find(*ignore, c)) { return true; } -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> Assertions in the callers of the function pointrs are already added by previous patches. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com> Message-Id: <20220303151616.325444-30-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/sysemu/block-backend-common.h | 28 ++++++++++++++++++++++----- 1 file changed, 23 insertions(+), 5 deletions(-) diff --git a/include/sysemu/block-backend-common.h b/include/sysemu/block-backend-common.h index XXXXXXX..XXXXXXX 100644 --- a/include/sysemu/block-backend-common.h +++ b/include/sysemu/block-backend-common.h @@ -XXX,XX +XXX,XX @@ /* Callbacks for block device models */ typedef struct BlockDevOps { + + /* + * Global state (GS) API. These functions run under the BQL. + * + * See include/block/block-global-state.h for more information about + * the GS API. + */ + /* * Runs when virtual media changed (monitor commands eject, change) * Argument load is true on load and false on eject. @@ -XXX,XX +XXX,XX @@ typedef struct BlockDevOps { * true, even if they do not support eject requests. */ void (*eject_request_cb)(void *opaque, bool force); - /* - * Is the virtual tray open? - * Device models implement this only when the device has a tray. - */ - bool (*is_tray_open)(void *opaque); + /* * Is the virtual medium locked into the device? * Device models implement this only when device has such a lock. */ bool (*is_medium_locked)(void *opaque); + + /* + * I/O API functions. These functions are thread-safe. + * + * See include/block/block-io.h for more information about + * the I/O API. + */ + + /* + * Is the virtual tray open? + * Device models implement this only when the device has a tray. + */ + bool (*is_tray_open)(void *opaque); + /* * Runs when the size changed (e.g. monitor command block_resize) */ -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> The job API will be handled separately in another serie. Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Message-Id: <20220303151616.325444-31-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- include/qemu/job.h | 22 ++++++++++++++++++++++ 1 file changed, 22 insertions(+) diff --git a/include/qemu/job.h b/include/qemu/job.h index XXXXXXX..XXXXXXX 100644 --- a/include/qemu/job.h +++ b/include/qemu/job.h @@ -XXX,XX +XXX,XX @@ typedef struct Job { * Callbacks and other information about a Job driver. */ struct JobDriver { + + /* + * These fields are initialized when this object is created, + * and are never changed afterwards + */ + /** Derived Job struct size */ size_t instance_size; @@ -XXX,XX +XXX,XX @@ struct JobDriver { * aborted. If it returns zero, the job moves into the WAITING state. If it * is the last job to complete in its transaction, all jobs in the * transaction move from WAITING to PENDING. + * + * This callback must be run in the job's context. */ int coroutine_fn (*run)(Job *job, Error **errp); + /* + * Functions run without regard to the BQL that may run in any + * arbitrary thread. These functions do not need to be thread-safe + * because the caller ensures that they are invoked from one + * thread at time. + */ + /** * If the callback is not NULL, it will be invoked when the job transitions * into the paused state. Paused jobs must not perform any asynchronous @@ -XXX,XX +XXX,XX @@ struct JobDriver { */ void coroutine_fn (*resume)(Job *job); + /* + * Global state (GS) API. These functions run under the BQL. + * + * See include/block/block-global-state.h for more information about + * the GS API. + */ + /** * Called when the job is resumed by the user (i.e. user_paused becomes * false). .user_resume is called before .resume. -- 2.35.1 From: Emanuele Giuseppe Esposito <eesposit@redhat.com> Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com> Message-Id: <20220303151616.325444-32-eesposit@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- job.c | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/job.c b/job.c index XXXXXXX..XXXXXXX 100644 --- a/job.c +++ b/job.c @@ -XXX,XX +XXX,XX @@ void job_ref(Job *job) void job_unref(Job *job) { + GLOBAL_STATE_CODE(); + if (--job->refcnt == 0) { assert(job->status == JOB_STATUS_NULL); assert(!timer_pending(&job->sleep_timer)); @@ -XXX,XX +XXX,XX @@ bool job_user_paused(Job *job) void job_user_resume(Job *job, Error **errp) { assert(job); + GLOBAL_STATE_CODE(); if (!job->user_paused || job->pause_count <= 0) { error_setg(errp, "Can't resume a job that was not paused"); return; @@ -XXX,XX +XXX,XX @@ static void job_update_rc(Job *job) static void job_commit(Job *job) { assert(!job->ret); + GLOBAL_STATE_CODE(); if (job->driver->commit) { job->driver->commit(job); } @@ -XXX,XX +XXX,XX @@ static void job_commit(Job *job) static void job_abort(Job *job) { assert(job->ret); + GLOBAL_STATE_CODE(); if (job->driver->abort) { job->driver->abort(job); } @@ -XXX,XX +XXX,XX @@ static void job_abort(Job *job) static void job_clean(Job *job) { + GLOBAL_STATE_CODE(); if (job->driver->clean) { job->driver->clean(job); } @@ -XXX,XX +XXX,XX @@ static int job_finalize_single(Job *job) static void job_cancel_async(Job *job, bool force) { + GLOBAL_STATE_CODE(); if (job->driver->cancel) { force = job->driver->cancel(job, force); } else { @@ -XXX,XX +XXX,XX @@ static void job_completed_txn_abort(Job *job) static int job_prepare(Job *job) { + GLOBAL_STATE_CODE(); if (job->ret == 0 && job->driver->prepare) { job->ret = job->driver->prepare(job); job_update_rc(job); @@ -XXX,XX +XXX,XX @@ static void coroutine_fn job_co_entry(void *opaque) Job *job = opaque; assert(job && job->driver && job->driver->run); + assert(job->aio_context == qemu_get_current_aio_context()); job_pause_point(job); job->ret = job->driver->run(job, &job->err); job->deferred_to_main_loop = true; @@ -XXX,XX +XXX,XX @@ void job_complete(Job *job, Error **errp) { /* Should not be reachable via external interface for internal jobs */ assert(job->id); + GLOBAL_STATE_CODE(); if (job_apply_verb(job, JOB_VERB_COMPLETE, errp)) { return; } -- 2.35.1 From: Hanna Reitz <hreitz@redhat.com> bdrv_refresh_limits() recurses down to the node's children. That does not seem necessary: When we refresh limits on some node, and then recurse down and were to change one of its children's BlockLimits, then that would mean we noticed the changed limits by pure chance. The fact that we refresh the parent's limits has nothing to do with it, so the reason for the change probably happened before this point in time, and we should have refreshed the limits then. Consequently, we should actually propagate block limits changes upwards, not downwards.  That is a separate and pre-existing issue, though, and so will not be addressed in this patch. The problem with recursing is that bdrv_refresh_limits() is not atomic. It begins with zeroing BDS.bl, and only then sets proper, valid limits. If we do not drain all nodes whose limits are refreshed, then concurrent I/O requests can encounter invalid request_alignment values and crash qemu. Therefore, a recursing bdrv_refresh_limits() requires the whole subtree to be drained, which is currently not ensured by most callers. A non-recursive bdrv_refresh_limits() only requires the node in question to not receive I/O requests, and this is done by most callers in some way or another: - bdrv_open_driver() deals with a new node with no parents yet - bdrv_set_file_or_backing_noperm() acts on a drained node - bdrv_reopen_commit() acts only on drained nodes - bdrv_append() should in theory require the node to be drained; in practice most callers just lock the AioContext, which should at least be enough to prevent concurrent I/O requests from accessing invalid limits So we can resolve the bug by making bdrv_refresh_limits() non-recursive. Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=1879437 Signed-off-by: Hanna Reitz <hreitz@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <20220216105355.30729-2-hreitz@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/io.c | 4 ---- 1 file changed, 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_refresh_limits(BlockDriverState *bs, Transaction *tran, Error **errp) QLIST_FOREACH(c, &bs->children, next) { if (c->role & (BDRV_CHILD_DATA | BDRV_CHILD_FILTERED | BDRV_CHILD_COW)) { - bdrv_refresh_limits(c->bs, tran, errp); - if (*errp) { - return; - } bdrv_merge_limits(&bs->bl, &c->bs->bl); have_limits = true; } -- 2.35.1 From: Hanna Reitz <hreitz@redhat.com> Add a parameter to optionally open a QMP connection when creating a QemuStorageDaemon instance. Signed-off-by: Hanna Reitz <hreitz@redhat.com> Message-Id: <20220216105355.30729-3-hreitz@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- tests/qemu-iotests/iotests.py | 32 +++++++++++++++++++++++++++++++- 1 file changed, 31 insertions(+), 1 deletion(-) 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 @@ from qemu.machine import qtest from qemu.qmp import QMPMessage +from qemu.aqmp.legacy import QEMUMonitorProtocol # Use this logger for logging messages directly from the iotests module logger = logging.getLogger('qemu.iotests') @@ -XXX,XX +XXX,XX @@ def cmd(self, cmd): class QemuStorageDaemon: - def __init__(self, *args: str, instance_id: str = 'a'): + _qmp: Optional[QEMUMonitorProtocol] = None + _qmpsock: Optional[str] = None + # Python < 3.8 would complain if this type were not a string literal + # (importing `annotations` from `__future__` would work; but not on <= 3.6) + _p: 'Optional[subprocess.Popen[bytes]]' = None + + def __init__(self, *args: str, instance_id: str = 'a', qmp: bool = False): assert '--pidfile' not in args self.pidfile = os.path.join(test_dir, f'qsd-{instance_id}-pid') all_args = [qsd_prog] + list(args) + ['--pidfile', self.pidfile] + if qmp: + self._qmpsock = os.path.join(sock_dir, f'qsd-{instance_id}.sock') + all_args += ['--chardev', + f'socket,id=qmp-sock,path={self._qmpsock}', + '--monitor', 'qmp-sock'] + + self._qmp = QEMUMonitorProtocol(self._qmpsock, server=True) + # Cannot use with here, we want the subprocess to stay around # pylint: disable=consider-using-with self._p = subprocess.Popen(all_args) + if self._qmp is not None: + self._qmp.accept() while not os.path.exists(self.pidfile): if self._p.poll() is not None: cmd = ' '.join(all_args) @@ -XXX,XX +XXX,XX @@ def __init__(self, *args: str, instance_id: str = 'a'): assert self._pid == self._p.pid + def qmp(self, cmd: str, args: Optional[Dict[str, object]] = None) \ + -> QMPMessage: + assert self._qmp is not None + return self._qmp.cmd(cmd, args) + def stop(self, kill_signal=15): self._p.send_signal(kill_signal) self._p.wait() self._p = None + if self._qmp: + self._qmp.close() + + if self._qmpsock is not None: + try: + os.remove(self._qmpsock) + except OSError: + pass try: os.remove(self.pidfile) except OSError: -- 2.35.1 From: Hanna Reitz <hreitz@redhat.com> Test the following scenario: 1. Some block node (null-co) attached to a user (here: NBD server) that performs I/O and keeps the node in an I/O thread 2. Repeatedly run blockdev-add/blockdev-del to add/remove an overlay to/from that node Each blockdev-add triggers bdrv_refresh_limits(), and because blockdev-add runs in the main thread, it does not stop the I/O requests. I/O can thus happen while the limits are refreshed, and when such a request sees a temporarily invalid block limit (e.g. alignment is 0), this may easily crash qemu (or the storage daemon in this case). The block layer needs to ensure that I/O requests to a node are paused while that node's BlockLimits are refreshed. Signed-off-by: Hanna Reitz <hreitz@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <20220216105355.30729-4-hreitz@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- .../qemu-iotests/tests/graph-changes-while-io | 91 +++++++++++++++++++ .../tests/graph-changes-while-io.out | 5 + 2 files changed, 96 insertions(+) create mode 100755 tests/qemu-iotests/tests/graph-changes-while-io create mode 100644 tests/qemu-iotests/tests/graph-changes-while-io.out diff --git a/tests/qemu-iotests/tests/graph-changes-while-io b/tests/qemu-iotests/tests/graph-changes-while-io new file mode 100755 index XXXXXXX..XXXXXXX --- /dev/null +++ b/tests/qemu-iotests/tests/graph-changes-while-io @@ -XXX,XX +XXX,XX @@ +#!/usr/bin/env python3 +# group: rw +# +# Test graph changes while I/O is happening +# +# Copyright (C) 2022 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/>. +# + +import os +from threading import Thread +import iotests +from iotests import imgfmt, qemu_img, qemu_img_create, QMPTestCase, \ + QemuStorageDaemon + + +top = os.path.join(iotests.test_dir, 'top.img') +nbd_sock = os.path.join(iotests.sock_dir, 'nbd.sock') + + +def do_qemu_img_bench() -> None: + """ + Do some I/O requests on `nbd_sock`. + """ + assert qemu_img('bench', '-f', 'raw', '-c', '2000000', + f'nbd+unix:///node0?socket={nbd_sock}') == 0 + + +class TestGraphChangesWhileIO(QMPTestCase): + def setUp(self) -> None: + # Create an overlay that can be added at runtime on top of the + # null-co block node that will receive I/O + assert qemu_img_create('-f', imgfmt, '-F', 'raw', '-b', 'null-co://', + top) == 0 + + # QSD instance with a null-co block node in an I/O thread, + # exported over NBD (on `nbd_sock`, export name "node0") + self.qsd = QemuStorageDaemon( + '--object', 'iothread,id=iothread0', + '--blockdev', 'null-co,node-name=node0,read-zeroes=true', + '--nbd-server', f'addr.type=unix,addr.path={nbd_sock}', + '--export', 'nbd,id=exp0,node-name=node0,iothread=iothread0,' + + 'fixed-iothread=true,writable=true', + qmp=True + ) + + def tearDown(self) -> None: + self.qsd.stop() + + def test_blockdev_add_while_io(self) -> None: + # Run qemu-img bench in the background + bench_thr = Thread(target=do_qemu_img_bench) + bench_thr.start() + + # While qemu-img bench is running, repeatedly add and remove an + # overlay to/from node0 + while bench_thr.is_alive(): + result = self.qsd.qmp('blockdev-add', { + 'driver': imgfmt, + 'node-name': 'overlay', + 'backing': 'node0', + 'file': { + 'driver': 'file', + 'filename': top + } + }) + self.assert_qmp(result, 'return', {}) + + result = self.qsd.qmp('blockdev-del', { + 'node-name': 'overlay' + }) + self.assert_qmp(result, 'return', {}) + + bench_thr.join() + +if __name__ == '__main__': + # Format must support raw backing files + iotests.main(supported_fmts=['qcow', 'qcow2', 'qed'], + supported_protocols=['file']) diff --git a/tests/qemu-iotests/tests/graph-changes-while-io.out b/tests/qemu-iotests/tests/graph-changes-while-io.out new file mode 100644 index XXXXXXX..XXXXXXX --- /dev/null +++ b/tests/qemu-iotests/tests/graph-changes-while-io.out @@ -XXX,XX +XXX,XX @@ +. +---------------------------------------------------------------------- +Ran 1 tests + +OK -- 2.35.1 From: Thomas Huth <thuth@redhat.com> Instead of failing the iotests if GNU sed is not available (or skipping them completely in the check-block.sh script), it would be better to simply skip the bash-based tests that rely on GNU sed, so that the other tests could still be run. Thus we now explicitely use "gsed" (either as direct program or as a wrapper around "sed" if it's the GNU version) in the spots that rely on the GNU sed behavior. Statements that use the "-r" parameter of sed have been switched to use "-E" instead, since this switch is supported by all sed versions on our supported build hosts (most also support "-r", but macOS' sed only supports "-E"). With all these changes in place, we then can also remove the sed checks from the check-block.sh script, so that "make check-block" can now be run on systems without GNU sed, too. Signed-off-by: Thomas Huth <thuth@redhat.com> Message-Id: <20220216125454.465041-1-thuth@redhat.com> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> Reviewed-by: Eric Blake <eblake@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- tests/check-block.sh | 12 ------ tests/qemu-iotests/271 | 2 +- tests/qemu-iotests/common.filter | 65 ++++++++++++++++---------------- tests/qemu-iotests/common.rc | 45 +++++++++++----------- 4 files changed, 57 insertions(+), 67 deletions(-) diff --git a/tests/check-block.sh b/tests/check-block.sh index XXXXXXX..XXXXXXX 100755 --- a/tests/check-block.sh +++ b/tests/check-block.sh @@ -XXX,XX +XXX,XX @@ if LANG=C bash --version | grep -q 'GNU bash, version [123]' ; then skip "bash version too old ==> Not running the qemu-iotests." fi -if ! (sed --version | grep 'GNU sed') > /dev/null 2>&1 ; then - if ! command -v gsed >/dev/null 2>&1; then - skip "GNU sed not available ==> Not running the qemu-iotests." - fi -else - # Double-check that we're not using BusyBox' sed which says - # that "This is not GNU sed version 4.0" ... - if sed --version | grep -q 'not GNU sed' ; then - skip "BusyBox sed not supported ==> Not running the qemu-iotests." - fi -fi - cd tests/qemu-iotests # QEMU_CHECK_BLOCK_AUTO is used to disable some unstable sub-tests diff --git a/tests/qemu-iotests/271 b/tests/qemu-iotests/271 index XXXXXXX..XXXXXXX 100755 --- a/tests/qemu-iotests/271 +++ b/tests/qemu-iotests/271 @@ -XXX,XX +XXX,XX @@ _make_test_img -o extended_l2=on 1M # Second and third writes in _concurrent_io() are independent and may finish in # different order. So, filter offset out to match both possible variants. _concurrent_io | $QEMU_IO | _filter_qemu_io | \ - $SED -e 's/\(20480\|40960\)/OFFSET/' + sed -e 's/\(20480\|40960\)/OFFSET/' _concurrent_verify | $QEMU_IO | _filter_qemu_io # success, all done 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_date() { - $SED -re 's/[0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}:[0-9]{2}/yyyy-mm-dd hh:mm:ss/' + sed -Ee 's/[0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}:[0-9]{2}/yyyy-mm-dd hh:mm:ss/' } _filter_vmstate_size() { - $SED -r -e 's/[0-9. ]{5} [KMGT]iB/ SIZE/' \ - -e 's/[0-9. ]{5} B/ SIZE/' + sed -E -e 's/[0-9. ]{5} [KMGT]iB/ SIZE/' \ + -e 's/[0-9. ]{5} B/ SIZE/' } _filter_generated_node_ids() { - $SED -re 's/\#block[0-9]{3,}/NODE_NAME/' + sed -Ee 's/\#block[0-9]{3,}/NODE_NAME/' } _filter_qom_path() { - $SED -e '/Attached to:/s/\device[[0-9]\+\]/device[N]/g' + gsed -e '/Attached to:/s/\device[[0-9]\+\]/device[N]/g' } # replace occurrences of the actual TEST_DIR value with TEST_DIR _filter_testdir() { - $SED -e "s#$TEST_DIR/#TEST_DIR/#g" \ - -e "s#$SOCK_DIR/#SOCK_DIR/#g" \ - -e "s#SOCK_DIR/fuse-#TEST_DIR/#g" + sed -e "s#$TEST_DIR/#TEST_DIR/#g" \ + -e "s#$SOCK_DIR/#SOCK_DIR/#g" \ + -e "s#SOCK_DIR/fuse-#TEST_DIR/#g" } # replace occurrences of the actual IMGFMT value with IMGFMT _filter_imgfmt() { - $SED -e "s#$IMGFMT#IMGFMT#g" + sed -e "s#$IMGFMT#IMGFMT#g" } # Replace error message when the format is not supported and delete # the output lines after the first one _filter_qemu_img_check() { - $SED -e '/allocated.*fragmented.*compressed clusters/d' \ + gsed -e '/allocated.*fragmented.*compressed clusters/d' \ -e 's/qemu-img: This image format does not support checks/No errors were found on the image./' \ -e '/Image end offset: [0-9]\+/d' } @@ -XXX,XX +XXX,XX @@ _filter_qemu_img_check() # Removes \r from messages _filter_win32() { - $SED -e 's/\r//g' + gsed -e 's/\r//g' } # sanitize qemu-io output _filter_qemu_io() { - _filter_win32 | $SED -e "s/[0-9]* ops\; [0-9/:. sec]* ([0-9/.inf]* [EPTGMKiBbytes]*\/sec and [0-9/.inf]* ops\/sec)/X ops\; XX:XX:XX.X (XXX YYY\/sec and XXX ops\/sec)/" \ + _filter_win32 | \ + gsed -e "s/[0-9]* ops\; [0-9/:. sec]* ([0-9/.inf]* [EPTGMKiBbytes]*\/sec and [0-9/.inf]* ops\/sec)/X ops\; XX:XX:XX.X (XXX YYY\/sec and XXX ops\/sec)/" \ -e "s/: line [0-9][0-9]*: *[0-9][0-9]*\( Aborted\| Killed\)/:\1/" \ -e "s/qemu-io> //g" } @@ -XXX,XX +XXX,XX @@ _filter_qemu_io() # replace occurrences of QEMU_PROG with "qemu" _filter_qemu() { - $SED -e "s#\\(^\\|(qemu) \\)$(basename $QEMU_PROG):#\1QEMU_PROG:#" \ + gsed -e "s#\\(^\\|(qemu) \\)$(basename $QEMU_PROG):#\1QEMU_PROG:#" \ -e 's#^QEMU [0-9]\+\.[0-9]\+\.[0-9]\+ monitor#QEMU X.Y.Z monitor#' \ -e $'s#\r##' # QEMU monitor uses \r\n line endings } @@ -XXX,XX +XXX,XX @@ _filter_qemu() _filter_qmp() { _filter_win32 | \ - $SED -e 's#\("\(micro\)\?seconds": \)[0-9]\+#\1 TIMESTAMP#g' \ + gsed -e 's#\("\(micro\)\?seconds": \)[0-9]\+#\1 TIMESTAMP#g' \ -e 's#^{"QMP":.*}$#QMP_VERSION#' \ -e '/^ "QMP": {\s*$/, /^ }\s*$/ c\' \ -e ' QMP_VERSION' @@ -XXX,XX +XXX,XX @@ _filter_qmp() # readline makes HMP command strings so long that git complains _filter_hmp() { - $SED -e $'s/^\\((qemu) \\)\\?.*\e\\[D/\\1/g' \ + gsed -e $'s/^\\((qemu) \\)\\?.*\e\\[D/\\1/g' \ -e $'s/\e\\[K//g' } # replace block job offset _filter_block_job_offset() { - $SED -e 's/, "offset": [0-9]\+,/, "offset": OFFSET,/' + sed -e 's/, "offset": [0-9]\+,/, "offset": OFFSET,/' } # replace block job len _filter_block_job_len() { - $SED -e 's/, "len": [0-9]\+,/, "len": LEN,/g' + sed -e 's/, "len": [0-9]\+,/, "len": LEN,/g' } # replace actual image size (depends on the host filesystem) _filter_actual_image_size() { - $SED -s 's/\("actual-size":\s*\)[0-9]\+/\1SIZE/g' + gsed -s 's/\("actual-size":\s*\)[0-9]\+/\1SIZE/g' } # Filename filters for qemu-img create _filter_img_create_filenames() { - $SED \ + sed \ -e "s#$REMOTE_TEST_DIR#TEST_DIR#g" \ -e "s#$IMGPROTO:$TEST_DIR#TEST_DIR#g" \ -e "s#$TEST_DIR#TEST_DIR#g" \ @@ -XXX,XX +XXX,XX @@ _do_filter_img_create() # precedes ", fmt=") and the options part ($options, which starts # with "fmt=") # (And just echo everything before the first "^Formatting") - readarray formatting_line < <($SED -e 's/, fmt=/\n/') + readarray formatting_line < <(gsed -e 's/, fmt=/\n/') filename_part=${formatting_line[0]} unset formatting_line[0] @@ -XXX,XX +XXX,XX @@ _do_filter_img_create() options=$( echo "$options" \ | tr '\n' '\0' \ - | $SED -e 's/ \([a-z0-9_.-]*\)=/\n\1=/g' \ + | gsed -e 's/ \([a-z0-9_.-]*\)=/\n\1=/g' \ | grep -a -e '^fmt' -e '^size' -e '^backing' -e '^preallocation' \ -e '^encryption' "${grep_data_file[@]}" \ | _filter_img_create_filenames \ - | $SED \ + | sed \ -e 's/^\(fmt\)/0-\1/' \ -e 's/^\(size\)/1-\1/' \ -e 's/^\(backing\)/2-\1/' \ @@ -XXX,XX +XXX,XX @@ _do_filter_img_create() -e 's/^\(encryption\)/4-\1/' \ -e 's/^\(preallocation\)/8-\1/' \ | LC_ALL=C sort \ - | $SED -e 's/^[0-9]-//' \ + | sed -e 's/^[0-9]-//' \ | tr '\n\0' ' \n' \ - | $SED -e 's/^ *$//' -e 's/ *$//' + | sed -e 's/^ *$//' -e 's/ *$//' ) if [ -n "$options" ]; then @@ -XXX,XX +XXX,XX @@ _filter_img_create() _filter_img_create_size() { - $SED -e "s# size=[0-9]\\+# size=SIZE#g" + gsed -e "s# size=[0-9]\\+# size=SIZE#g" } _filter_img_info() @@ -XXX,XX +XXX,XX @@ _filter_img_info() discard=0 regex_json_spec_start='^ *"format-specific": \{' - $SED -e "s#$REMOTE_TEST_DIR#TEST_DIR#g" \ + gsed -e "s#$REMOTE_TEST_DIR#TEST_DIR#g" \ -e "s#$IMGPROTO:$TEST_DIR#TEST_DIR#g" \ -e "s#$TEST_DIR#TEST_DIR#g" \ -e "s#$SOCK_DIR#SOCK_DIR#g" \ @@ -XXX,XX +XXX,XX @@ _filter_qemu_img_map() data_file_filter=(-e "s#$data_file_pattern#\\1#") fi - $SED -e 's/\([0-9a-fx]* *[0-9a-fx]* *\)[0-9a-fx]* */\1/g' \ + sed -e 's/\([0-9a-fx]* *[0-9a-fx]* *\)[0-9a-fx]* */\1/g' \ -e 's/"offset": [0-9]\+/"offset": OFFSET/g' \ -e 's/Mapped to *//' \ "${data_file_filter[@]}" \ @@ -XXX,XX +XXX,XX @@ _filter_nbd() # receive callbacks sometimes, making them unreliable. # # Filter out the TCP port number since this changes between runs. - $SED -e '/nbd\/.*\.c:/d' \ + sed -e '/nbd\/.*\.c:/d' \ -e 's#127\.0\.0\.1:[0-9]*#127.0.0.1:PORT#g' \ -e "s#?socket=$SOCK_DIR#?socket=SOCK_DIR#g" \ -e 's#\(foo\|PORT/\?\|.sock\): Failed to .*$#\1#' @@ -XXX,XX +XXX,XX @@ sys.stdout.write(result)' _filter_authz_check_tls() { - $SED -e 's/TLS x509 authz check for .* is denied/TLS x509 authz check for DISTINGUISHED-NAME is denied/' + sed -e 's/TLS x509 authz check for .* is denied/TLS x509 authz check for DISTINGUISHED-NAME is denied/' } _filter_qcow2_compression_type_bit() { - $SED -e 's/\(incompatible_features\s\+\)\[3\(, \)\?/\1[/' \ - -e 's/\(incompatible_features.*\), 3\]/\1]/' \ - -e 's/\(incompatible_features.*\), 3\(,.*\)/\1\2/' + gsed -e 's/\(incompatible_features\s\+\)\[3\(, \)\?/\1[/' \ + -e 's/\(incompatible_features.*\), 3\]/\1]/' \ + -e 's/\(incompatible_features.*\), 3\(,.*\)/\1\2/' } # make sure this script returns success diff --git a/tests/qemu-iotests/common.rc b/tests/qemu-iotests/common.rc index XXXXXXX..XXXXXXX 100644 --- a/tests/qemu-iotests/common.rc +++ b/tests/qemu-iotests/common.rc @@ -XXX,XX +XXX,XX @@ # along with this program. If not, see <http://www.gnu.org/licenses/>. # -SED= -for sed in sed gsed; do - ($sed --version | grep 'GNU sed') > /dev/null 2>&1 - if [ "$?" -eq 0 ]; then - SED=$sed - break - fi -done -if [ -z "$SED" ]; then - echo "$0: GNU sed not found" - exit 1 +# bail out, setting up .notrun file +_notrun() +{ + echo "$*" >"$OUTPUT_DIR/$seq.notrun" + echo "$seq not run: $*" + status=0 + exit +} + +if ! command -v gsed >/dev/null 2>&1; then + if sed --version 2>&1 | grep -v 'not GNU sed' | grep 'GNU sed' > /dev/null; + then + gsed() + { + sed "$@" + } + else + gsed() + { + _notrun "GNU sed not available" + } + fi fi dd() @@ -XXX,XX +XXX,XX @@ _img_info() done } -# bail out, setting up .notrun file -# -_notrun() -{ - echo "$*" >"$OUTPUT_DIR/$seq.notrun" - echo "$seq not run: $*" - status=0 - exit -} - # bail out, setting up .casenotrun file # The function _casenotrun() is used as a notifier. It is the # caller's responsibility to make skipped a particular test. @@ -XXX,XX +XXX,XX @@ _require_working_luks() IMGFMT='luks' _rm_test_img "$file" if [ $status != 0 ]; then - reason=$(echo "$output" | grep "$file:" | $SED -e "s#.*$file: *##") + reason=$(echo "$output" | grep "$file:" | sed -e "s#.*$file: *##") if [ -z "$reason" ]; then reason="Failed to create a LUKS image" fi -- 2.35.1 From: Hanna Reitz <hreitz@redhat.com> .bdrv_amend_clean() says block drivers can use it to clean up what was done in .bdrv_amend_pre_run(). Therefore, it should always be called after .bdrv_amend_pre_run(), which means we need it to call it in the JobDriver.free() callback, not in JobDriver.clean(). Signed-off-by: Hanna Reitz <hreitz@redhat.com> Message-Id: <20220304153729.711387-3-hreitz@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/amend.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/block/amend.c b/block/amend.c index XXXXXXX..XXXXXXX 100644 --- a/block/amend.c +++ b/block/amend.c @@ -XXX,XX +XXX,XX @@ static int blockdev_amend_pre_run(BlockdevAmendJob *s, Error **errp) return 0; } -static void blockdev_amend_clean(Job *job) +static void blockdev_amend_free(Job *job) { BlockdevAmendJob *s = container_of(job, BlockdevAmendJob, common); @@ -XXX,XX +XXX,XX @@ static const JobDriver blockdev_amend_job_driver = { .instance_size = sizeof(BlockdevAmendJob), .job_type = JOB_TYPE_AMEND, .run = blockdev_amend_run, - .clean = blockdev_amend_clean, + .free = blockdev_amend_free, }; void qmp_x_blockdev_amend(const char *job_id, -- 2.35.1 From: Hanna Reitz <hreitz@redhat.com> Otherwise, the BDS might be freed while the job is running, which would cause a use-after-free. Signed-off-by: Hanna Reitz <hreitz@redhat.com> Message-Id: <20220304153729.711387-5-hreitz@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com> --- block/amend.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/block/amend.c b/block/amend.c index XXXXXXX..XXXXXXX 100644 --- a/block/amend.c +++ b/block/amend.c @@ -XXX,XX +XXX,XX @@ static void blockdev_amend_free(Job *job) if (s->bs->drv->bdrv_amend_clean) { s->bs->drv->bdrv_amend_clean(s->bs); } + + bdrv_unref(s->bs); } static const JobDriver blockdev_amend_job_driver = { @@ -XXX,XX +XXX,XX @@ void qmp_x_blockdev_amend(const char *job_id, return; } + bdrv_ref(bs); s->bs = bs, s->opts = QAPI_CLONE(BlockdevAmendOptions, options), s->force = has_force ? force : false; -- 2.35.1