This patch series adds a new extension - zoned format - to the
qcow2 driver, allowing full zoned storage emulation on a qcow2
image file. A user can attach such an image to a guest and have
it appear as a host-managed zoned block device.
The zoned format is opt-in through a new qcow2 header extension
that pins the zone geometry. Behind the extension is a dedicated
zoned metadata region that stores one 8-byte write pointer (WP)
per zone. The extension is gated by an incompatible bit, so an
older qcow2 implementation cannot accidentally open the image.
Each write pointer is routed through the write pointer cache,
a Qcow2Cache object. The write pointer cache is written to disk
after the qcow2 metadata is written, thus guaranteeing that
the write pointer is updated after the corresponding data is
written.
Zone states are in memory. Read-only and offline states are
device-internal events, which are not modelled in qcow2
emulation for simplicity. The other zone states
(closed, empty, full) can be inferred from write pointer
values, persistent across QEMU reboots. The open states are
kept in memory using open zone lists.
To create a qcow2 file with the zoned format:
qemu-img create -f qcow2 zbc.qcow2 \
-o size=768M \
-o zone.size=64M \
-o zone.capacity=64M \
-o zone.conventional_zones=0 \
-o zone.max_open_zones=6 \
-o zone.max_active_zones=8 \
-o zone.mode=host-managed
Then attach it to a guest via the QEMU command line:
-blockdev
node-name=drive1,driver=qcow2,file.driver=file,file.filename=zbc.qcow2 \
-device virtio-blk-pci,drive=drive1 \
Tested with blktests, fio, zbd_state (an internal test suite) and
zonefs, comparing a zoned qcow2 image attached via virtio-blk with
zloop, null_blk and scsi_debug (passed/failed/skipped):
+--------------------+----------+---------+-----------+---------+
| Testrun | blktests | fio | zbd_state | zonefs |
+--------------------+----------+---------+-----------+---------+
| zloop | 9/0/10 | 57/0/18 | 29/0/8 | 113/0/0 |
| null_blk | 9/0/10 | 57/0/18 | 29/0/8 | 113/0/0 |
| scsi_debug | 10/0/12 | 61/0/14 | 29/0/8 | 113/0/0 |
| virtio-blk + qcow2 | 9/0/10 | 57/0/18 | 29/0/8 | 113/0/0 |
+--------------------+----------+---------+-----------+---------+
This series depends on the (still not merged) series "block: fix the
zone write granularity and the zone append limit":
https://lore.kernel.org/qemu-devel/[email protected]/
Git branch which includes the dependency series can be found at:
https://github.com/floatious/qemu/tree/qcow2-v15
Based-on: <[email protected]>
v14->v15:
- rebase on the series above, which now contains the v14 patches
"block: widen BlockLimits.zone_size to uint64_t" and "virtio-blk: do
not merge writes across a zone boundary", and which moves the zone
append checks to the block layer
- drop zone.max_append_bytes; zonedmeta_offset moves to bytes 24-31 of
the zoned header extension, so images created with v14 are not
compatible
- require a version 3 image, a zone size that is a multiple of the
cluster size, and a zone capacity that is a multiple of the zone
write granularity
- report zone errors with dedicated errnos (new patch) instead of EIO,
EINVAL and EBUSY
- manage the zone resources as in NVMe ZNS (2.1.1.4.1) and ZBC-3, and
fix OPEN ZONE of a full zone, writes that fill a closed zone, writes
from a conventional into a sequential zone, conventional zone and
zone capacity reporting, and a shorter last zone
- validate the zoned header extension, and the write pointer table (new
patch), when opening an image
- account for the zoned metadata in qemu-img check
- report zone errors with their zone status codes in virtio-blk (new
patch)
- extend the iotest with zone errors, zone resources and image checks
Niklas Cassel (3):
block: define errnos for zone protocol errors
qcow2: validate the write pointer table when opening an image
virtio-blk: report zone errors with their zone status codes
Sam Li (4):
docs/qcow2: add the zoned format feature
qcow2: add configurations for zoned format extension
qcow2: add zoned emulation capability
iotests: test the zoned format feature for qcow2 file
MAINTAINERS | 1 +
block/io.c | 13 +-
block/qcow2-cache.c | 11 +-
block/qcow2-refcount.c | 18 +
block/qcow2.c | 1674 ++++++++++++++++-
block/qcow2.h | 86 +-
block/trace-events | 2 +
docs/interop/qcow2.rst | 115 +-
docs/system/qemu-block-drivers.rst.inc | 44 +
hw/block/block.c | 7 +
hw/block/virtio-blk.c | 48 +-
include/block/block-common.h | 16 +
include/block/block-io.h | 6 +
include/block/block_int-common.h | 12 +
include/qemu/osdep.h | 3 +
qapi/block-core.json | 86 +-
tests/qemu-iotests/031.out | 8 +-
tests/qemu-iotests/036.out | 4 +-
tests/qemu-iotests/061.out | 14 +-
tests/qemu-iotests/082.out | 108 ++
tests/qemu-iotests/303.out | 4 +-
tests/qemu-iotests/tests/qcow2-encryption.out | 4 +-
tests/qemu-iotests/tests/zoned-qcow2 | 362 ++++
tests/qemu-iotests/tests/zoned-qcow2.out | 265 +++
24 files changed, 2874 insertions(+), 37 deletions(-)
create mode 100755 tests/qemu-iotests/tests/zoned-qcow2
create mode 100644 tests/qemu-iotests/tests/zoned-qcow2.out
--
2.56.0
