qcow2: Implement zstd cluster compression method
PATCH v13 2/4 qcow2: rework the cluster compression routine
In reply to Denis PlotnikovThe patch enables processing the image compression type defined for the image and chooses an appropriate method for image clusters (de)compression.
--- block/qcow2-threads.c | 71 ++++++++++++++++++++++++++++++++++++------- 1 file changed, 60 insertions(+), 11 deletions(-)
block/qcow2-threads.c +60 -11
@@ -74,7 +74,9 @@ typedef struct Qcow2CompressData { } Qcow2CompressData;/* -* qcow2_compress() +* qcow2_zlib_compress() +* +* Compress @src_size bytes of data using zlib compression method * * @dest - destination buffer, @dest_size bytes * @src - source buffer, @src_size bytes
@@ -83,8 +85,8 @@ typedef struct Qcow2CompressData {* -ENOMEM destination buffer is not enough to store compressed data * -EIO on any other error */ -static ssize_t qcow2_compress (void *dest,size_t dest_size, -const void *src,size_t src_size) +static ssize_t qcow2_zlib_compress (void *dest,size_t dest_size, +const void *src,size_t src_size) {ssize_t ret;z_stream strm;
@@ -119,10 +121,10 @@ static ssize_t qcow2_compress(void *dest, size_t dest_size, }/* -* qcow2_decompress() +* qcow2_zlib_decompress() * * Decompress some data (not more than @src_size bytes) to produce exactly -* @dest_size bytes. +* @dest_size bytes using zlib compression method * * @dest - destination buffer, @dest_size bytes * @src - source buffer, @src_size bytes
@@ -130,8 +132,8 @@ static ssize_t qcow2_compress(void *dest, size_t dest_size,* Returns: 0 on success * -EIO on fail */ -static ssize_t qcow2_decompress (void *dest,size_t dest_size, -const void *src,size_t src_size) +static ssize_t qcow2_zlib_decompress (void *dest,size_t dest_size, +const void *src,size_t src_size) {int ret;z_stream strm;
@@ -191,20 +193,67 @@ qcow2_co_do_compress(BlockDriverState *bs, void *dest, size_t dest_size,return arg.ret; } +/* +* qcow2_co_compress() +* +* Compress @src_size bytes of data using the compression +* method defined by the image compression type +* +* @dest - destination buffer, @dest_size bytes +* @src - source buffer, @src_size bytes +* +* Returns: compressed size on success +* a negative error code on failure +*/ ssize_t coroutine_fnqcow2_co_compress (BlockDriverState *bs,void *dest,size_t dest_size,const void *src,size_t src_size) { -return qcow2_co_do_compress (bs, dest, dest_size, src, src_size, - qcow2_compress); +BDRVQcow2State *s = bs->opaque; +Qcow2CompressFunc fn; + +switch (s->compression_type) { +case QCOW2_COMPRESSION_TYPE_ZLIB : + fn = qcow2_zlib_compress; +break ; + +default : +abort (); + } + +return qcow2_co_do_compress (bs, dest, dest_size, src, src_size, fn); } +/* +* qcow2_co_decompress() +* +* Decompress some data (not more than @src_size bytes) to produce exactly +* @dest_size bytes using the compression method defined by the image +* compression type +* +* @dest - destination buffer, @dest_size bytes +* @src - source buffer, @src_size bytes +* +* Returns: 0 on success +* a negative error code on failure +*/ ssize_t coroutine_fnqcow2_co_decompress (BlockDriverState *bs,void *dest,size_t dest_size,const void *src,size_t src_size) { -return qcow2_co_do_compress (bs, dest, dest_size, src, src_size, - qcow2_decompress); +BDRVQcow2State *s = bs->opaque; +Qcow2CompressFunc fn; + +switch (s->compression_type) { +case QCOW2_COMPRESSION_TYPE_ZLIB : + fn = qcow2_zlib_decompress; +break ; + +default : +abort (); + } + +return qcow2_co_do_compress (bs, dest, dest_size, src, src_size, fn); }
Signature · 1 line
2.17.0
PATCH v13 3/4 qcow2: add zstd cluster compression
In reply to Denis Plotnikovzstd significantly reduces cluster compression time. It provides better compression performance maintaining the same level of the compression ratio in comparison with zlib, which, at the moment, is the only compression method available.
The performance test results:
Test compresses and decompresses qemu qcow2 image with just installed rhel-7.6 guest.
Image cluster size: 64K. Image on disk size: 2.2G
The test was conducted with brd disk to reduce the influence of disk subsystem to the test results.
The results is given in seconds.
compress cmd:
time ./qemu-img convert -O qcow2 -c -o compression_type=[zlib|zstd]
src.img [zlib|zstd]_compressed.img
decompress cmd
time ./qemu-img convert -O qcow2
[zlib|zstd]_compressed.img uncompressed.img
compression decompression
zlib zstd zlib zstd
------------------------------------------------------------
real 65.5 16.3 (-75 %) 1.9 1.6 (-16 %) user 65.0 15.8 5.3 2.5 sys 3.3 0.2 2.0 2.0
Both ZLIB and ZSTD gave the same compression ratio: 1.57 compressed image size in both cases: 1.4G
QAPI part:
--- docs/interop/qcow2.txt | 1 + configure | 2 +- qapi/block-core.json | 3 +- block/qcow2-threads.c | 150 +++++++++++++++++++++++++++++++++++++++++ block/qcow2.c | 7 ++ 5 files changed, 161 insertions(+), 2 deletions(-)
docs/interop/qcow2.txt +1 -0
@@ -209,6 +209,7 @@ version 2. Available compression type values: 0: zlib <https://www.zlib.net/> + 1: zstd <http://github.com/facebook/zstd> === Header padding ===
configure +1 -1
@@ -1861,7 +1861,7 @@ disabled with --disable-FEATURE, default is enabled if available: lzfse support of lzfse compression library (for reading lzfse-compressed dmg images) zstd support for zstd compression library - (for migration compression) + (for migration compression and qcow2 cluster compression) seccomp seccomp support coroutine-pool coroutine freelist (better performance) glusterfs GlusterFS backend
qapi/block-core.json +2 -1
@@ -4293,11 +4293,12 @@ # Compression type used in qcow2 image file # # @zlib: zlib compression, see <http://zlib.net/> +# @zstd: zstd compression, see <http://github.com/facebook/zstd> # # Since: 5.0 ## { 'enum': 'Qcow2CompressionType', - 'data': [ 'zlib' ] } + 'data': [ 'zlib', { 'name': 'zstd', 'if': 'defined(CONFIG_ZSTD)' } ] } ## # @BlockdevCreateOptionsQcow2:
block/qcow2-threads.c +150 -0
174 lines not on this page · raw message
block/qcow2.c +7 -0
@@ -1246,6 +1246,9 @@ static int validate_compression_type(BDRVQcow2State *s, Error **errp) {switch (s->compression_type) {case QCOW2_COMPRESSION_TYPE_ZLIB : +#ifdef CONFIG_ZSTD +case QCOW2_COMPRESSION_TYPE_ZSTD : +#endif break ;default :
@@ -3478,6 +3481,10 @@ qcow2_co_create(BlockdevCreateOptions *create_options, Error **errp) }switch (qcow2_opts->compression_type) { +#ifdef CONFIG_ZSTD +case QCOW2_COMPRESSION_TYPE_ZSTD : +break ; +#endif default :error_setg (errp,"Unknown compression type" ); goto out;
Signature · 1 line
2.17.0
PATCH v13 4/4 iotests: 287: add qcow2 compression type test
In reply to Denis PlotnikovThe test checks fulfilling qcow2 requiriements for the compression type feature and zstd compression type operability.
--- tests/qemu-iotests/287 | 128 +++++++++++++++++++++++++++++++++++++ tests/qemu-iotests/287.out | 43 +++++++++++++ tests/qemu-iotests/group | 1 + 3 files changed, 172 insertions(+) create mode 100755 tests/qemu-iotests/287 create mode 100644 tests/qemu-iotests/287.out
tests/qemu-iotests/287 new file +128 -0
@@ -0,0 +1,128 @@ +#!/usr/bin/env bash +# +# Test case for an image using zstd compression +# +# Copyright (c) 2020 Virtuozzo International GmbH +# +# 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 [email protected] + +seq="$(basename $0)" +echo "QA output created by $seq" + +status=1 # failure is the default! + +_cleanup() +{ + _cleanup_test_img +} +trap "_cleanup; exit \$status" 0 1 2 3 15 + +# standard environment +. ./common.rc +. ./common.filter + +# This tests qocw2-specific low-level functionality +_supported_fmt qcow2 +_supported_proto file +_supported_os Linux + +# for all the cases +CLUSTER_SIZE=65536 + +# Check if we can run this test. + +IMGOPTS='compression_type=zstd' _make_test_img 64M | grep "Invalid parameter 'zstd'" 2>&1 1>/dev/null + +ZSTD_SUPPORTED=$? + +if (($ZSTD_SUPPORTED==0)); then + _notrun "ZSTD is disabled" +fi + +# Test: when compression is zlib the incompatible bit is unset +echo +echo "=== Testing compression type incompatible bit setting for zlib ===" +echo + +IMGOPTS='compression_type=zlib' _make_test_img 64M +$PYTHON qcow2.py "$TEST_IMG" dump-header | grep incompatible_features + +# Test: when compression differs from zlib the incompatible bit is set +echo +echo "=== Testing compression type incompatible bit setting for zstd ===" +echo + +IMGOPTS='compression_type=zstd' _make_test_img 64M +$PYTHON qcow2.py "$TEST_IMG" dump-header | grep incompatible_features + +# Test: an image can't be openned if compression type is zlib and +# incompatible feature compression type is set +echo +echo "=== Testing zlib with incompatible bit set ===" +echo + +IMGOPTS='compression_type=zlib' _make_test_img 64M +$PYTHON qcow2.py "$TEST_IMG" set-feature-bit incompatible 3 +# to make sure the bit was actually set +$PYTHON qcow2.py "$TEST_IMG" dump-header | grep incompatible_features +$QEMU_IMG info "$TEST_IMG" 2>1 1>/dev/null +if (($?==0)); then + echo "Error: The image openned successfully. The image must not be openned" +fi + +# Test: an image can't be openned if compression type is NOT zlib and +# incompatible feature compression type is UNSET +echo +echo "=== Testing zstd with incompatible bit unset ===" +echo + +IMGOPTS='compression_type=zstd' _make_test_img 64M +$PYTHON qcow2.py "$TEST_IMG" set-header incompatible_features 0 +# to make sure the bit was actually unset +$PYTHON qcow2.py "$TEST_IMG" dump-header | grep incompatible_features +$QEMU_IMG info "$TEST_IMG" 2>1 1>/dev/null +if (($?==0)); then + echo "Error: The image openned successfully. The image must not be openned" +fi +# Test: check compression type values +echo +echo "=== Testing compression type values ===" +echo +# zlib=0 +IMGOPTS='compression_type=zlib' _make_test_img 64M +od -j104 -N1 -An -vtu1 "$TEST_IMG" + +# zstd=1 +IMGOPTS='compression_type=zstd' _make_test_img 64M +od -j104 -N1 -An -vtu1 "$TEST_IMG" + +# Test: using zstd compression, write to and read from an image +echo +echo "=== Testing reading and writing with zstd ===" +echo + +IMGOPTS='compression_type=zstd' _make_test_img 64M +$QEMU_IO -c "write -c -P 0xAC 65536 64k " "$TEST_IMG" | _filter_qemu_io +$QEMU_IO -c "read -P 0xAC 65536 65536 " "$TEST_IMG" | _filter_qemu_io +$QEMU_IO -c "read -v 131070 8 " "$TEST_IMG" | _filter_qemu_io +$QEMU_IO -c "read -v 65534 8" "$TEST_IMG" | _filter_qemu_io + +# success, all done +echo "*** done" +rm -f $seq.full +status=0
tests/qemu-iotests/287.out new file +43 -0
43 lines not on this page · raw message
tests/qemu-iotests/group +1 -0
@@ -294,5 +294,6 @@ 283 auto quick 284 rw 286 rw quick +287 auto quick 288 quick 289 rw quick
Signature · 1 line
2.17.0
v13 cover qcow2: Implement zstd cluster compression method
v13:
* 03: add progress sanity check to decompression loop [Vladimir] 03: add successful decompression check [Me]
v12:
* 03: again, rework compression and decompression loops
to make them more correct [Vladimir]
03: move assert in compression to more appropriate place
[Vladimir]
v11:
* 03: the loops don't need "do{}while" form anymore and
the they were buggy (missed "do" in the beginning)
replace them with usual "while(){}" loops [Vladimir]
v10:
* 03: fix zstd (de)compressed loops for multi-frame
cases [Vladimir]
v9:
* 01: fix error checking and reporting in qcow2_amend compression type part [Vladimir]
* 03: replace asserts with -EIO in qcow2_zstd_decompression [Vladimir, Alberto]
* 03: reword/amend/add comments, fix typos [Vladimir]
v8:
* 03: switch zstd API from simple to stream [Eric] No need to state a special cluster layout for zstd compressed clusters.
v7:
* use qapi_enum_parse instead of the open-coding [Eric]
* fix wording, typos and spelling [Eric]
v6:
* "block/qcow2-threads: fix qcow2_decompress" is removed from the series since it has been accepted by Max already
* add compile time checking for Qcow2Header to be a multiple of 8 [Max, Alberto]
* report error on qcow2 amending when the compression type is actually chnged [Max]
* remove the extra space and the extra new line [Max]
* re-arrange acks and signed-off-s [Vladimir]
v5:
* replace -ENOTSUP with abort in qcow2_co_decompress [Vladimir]
* set cluster size for all test cases in the beginning of the 287 test
v4:
* the series is rebased on top of 01 "block/qcow2-threads: fix qcow2_decompress"
* 01 is just a no-change resend to avoid extra dependencies. Still, it may be merged in separate
v3:
* remove redundant max compression type value check [Vladimir, Eric] (the switch below checks everything)
* prevent compression type changing on "qemu-img amend" [Vladimir]
* remove zstd config setting, since it has been added already by "migration" patches [Vladimir]
* change the compression type error message [Vladimir]
* fix alignment and 80-chars exceeding [Vladimir]
v2:
* rework compression type setting [Vladimir]
* squash iotest changes to the compression type introduction patch [Vladimir, Eric]
* fix zstd availability checking in zstd iotest [Vladimir]
* remove unnecessry casting [Eric]
* remove rudundant checks [Eric]
* fix compressed cluster layout in qcow2 spec [Vladimir]
* fix wording [Eric, Vladimir]
* fix compression type filtering in iotests [Eric]
v1:
the initial series
Denis Plotnikov (4):
qcow2: introduce compression type feature
qcow2: rework the cluster compression routine
qcow2: add zstd cluster compression
iotests: 287: add qcow2 compression type test
docs/interop/qcow2.txt | 1 + configure | 2 +- qapi/block-core.json | 23 +++- block/qcow2.h | 20 ++- include/block/block_int.h | 1 + block/qcow2-threads.c | 221 +++++++++++++++++++++++++++++-- block/qcow2.c | 120 +++++++++++++++++ tests/qemu-iotests/031.out | 14 +- tests/qemu-iotests/036.out | 4 +- tests/qemu-iotests/049.out | 102 +++++++------- tests/qemu-iotests/060.out | 1 + tests/qemu-iotests/061.out | 34 +++-- tests/qemu-iotests/065 | 28 ++-- tests/qemu-iotests/080 | 2 +- tests/qemu-iotests/144.out | 4 +- tests/qemu-iotests/182.out | 2 +- tests/qemu-iotests/242.out | 5 + tests/qemu-iotests/255.out | 8 +- tests/qemu-iotests/287 | 128 ++++++++++++++++++ tests/qemu-iotests/287.out | 43 ++++++ tests/qemu-iotests/common.filter | 3 +- tests/qemu-iotests/group | 1 + 22 files changed, 659 insertions(+), 108 deletions(-) create mode 100755 tests/qemu-iotests/287 create mode 100644 tests/qemu-iotests/287.out
Signature · 1 line
2.17.0
PATCH v13 1/4 qcow2: introduce compression type feature
The patch adds some preparation parts for incompatible compression type feature to qcow2 allowing the use different compression methods for image clusters (de)compressing.
It is implied that the compression type is set on the image creation and can be changed only later by image conversion, thus compression type defines the only compression algorithm used for the image, and thus, for all image clusters.
The goal of the feature is to add support of other compression methods to qcow2. For example, ZSTD which is more effective on compression than ZLIB.
The default compression is ZLIB. Images created with ZLIB compression type are backward compatible with older qemu versions.
Adding of the compression type breaks a number of tests because now the compression type is reported on image creation and there are some changes in the qcow2 header in size and offsets.
The tests are fixed in the following ways:
* filter out compression_type for many tests
* fix header size, feature table size and backing file offset
affected tests: 031, 036, 061, 080
header_size +=8: 1 byte compression type
7 bytes padding
feature_table += 48: incompatible feature compression type
backing_file_offset += 56 (8 + 48 -> header_change + feature_table_change)
* add "compression type" for test output matching when it isn't filtered
affected tests: 049, 060, 061, 065, 144, 182, 242, 255QAPI part:
--- qapi/block-core.json | 22 +++++- block/qcow2.h | 20 +++++- include/block/block_int.h | 1 + block/qcow2.c | 113 +++++++++++++++++++++++++++++++ tests/qemu-iotests/031.out | 14 ++-- tests/qemu-iotests/036.out | 4 +- tests/qemu-iotests/049.out | 102 ++++++++++++++-------------- tests/qemu-iotests/060.out | 1 + tests/qemu-iotests/061.out | 34 ++++++---- tests/qemu-iotests/065 | 28 +++++--- tests/qemu-iotests/080 | 2 +- tests/qemu-iotests/144.out | 4 +- tests/qemu-iotests/182.out | 2 +- tests/qemu-iotests/242.out | 5 ++ tests/qemu-iotests/255.out | 8 +-- tests/qemu-iotests/common.filter | 3 +- 16 files changed, 267 insertions(+), 96 deletions(-)
qapi/block-core.json +20 -2
@@ -78,6 +78,8 @@ # # @bitmaps: A list of qcow2 bitmap details (since 4.0) # +# @compression-type: the image cluster compression method (since 5.0) +# # Since: 1.7 ## { 'struct': 'ImageInfoSpecificQCow2',
@@ -89,7 +91,8 @@ '*corrupt': 'bool', 'refcount-bits': 'int', '*encrypt': 'ImageInfoSpecificQCow2Encryption', - '*bitmaps': ['Qcow2BitmapInfo'] + '*bitmaps': ['Qcow2BitmapInfo'], + 'compression-type': 'Qcow2CompressionType' } } ##
@@ -4284,6 +4287,18 @@ 'data': [ 'v2', 'v3' ] } +## +# @Qcow2CompressionType: +# +# Compression type used in qcow2 image file +# +# @zlib: zlib compression, see <http://zlib.net/> +# +# Since: 5.0 +## +{ 'enum': 'Qcow2CompressionType', + 'data': [ 'zlib' ] } + ## # @BlockdevCreateOptionsQcow2: #
@@ -4307,6 +4322,8 @@ # allowed values: off, falloc, full, metadata) # @lazy-refcounts: True if refcounts may be updated lazily (default: off) # @refcount-bits: Width of reference counts in bits (default: 16) +# @compression-type: The image cluster compression method +# (default: zlib, since 5.0) # # Since: 2.12 ##
@@ -4322,7 +4339,8 @@ '*cluster-size': 'size', '*preallocation': 'PreallocMode', '*lazy-refcounts': 'bool', - '*refcount-bits': 'int' } } + '*refcount-bits': 'int', + '*compression-type':'Qcow2CompressionType' } } ## # @BlockdevCreateOptionsQed:
block/qcow2.h +19 -1
@@ -146,8 +146,16 @@ typedef struct QCowHeader {uint32_t refcount_order;uint32_t header_length; + +/* Additional fields */ +uint8_t compression_type; + +/* header must be a multiple of 8 */ +uint8_t padding[7 ]; }QEMU_PACKED QCowHeader; +QEMU_BUILD_BUG_ON (!QEMU_IS_ALIGNED (sizeof (QCowHeader),8 )); + typedef struct QEMU_PACKED QCowSnapshotHeader {/* header is 8 byte aligned */ uint64_t l1_table_offset;
@@ -216,13 +224,16 @@ enum {QCOW2_INCOMPAT_DIRTY_BITNR =0 ,QCOW2_INCOMPAT_CORRUPT_BITNR =1 ,QCOW2_INCOMPAT_DATA_FILE_BITNR =2 , +QCOW2_INCOMPAT_COMPRESSION_BITNR =3 ,QCOW2_INCOMPAT_DIRTY =1 <<QCOW2_INCOMPAT_DIRTY_BITNR ,QCOW2_INCOMPAT_CORRUPT =1 <<QCOW2_INCOMPAT_CORRUPT_BITNR ,QCOW2_INCOMPAT_DATA_FILE =1 <<QCOW2_INCOMPAT_DATA_FILE_BITNR , +QCOW2_INCOMPAT_COMPRESSION =1 <<QCOW2_INCOMPAT_COMPRESSION_BITNR ,QCOW2_INCOMPAT_MASK =QCOW2_INCOMPAT_DIRTY |QCOW2_INCOMPAT_CORRUPT - | QCOW2_INCOMPAT_DATA_FILE, + |QCOW2_INCOMPAT_DATA_FILE + | QCOW2_INCOMPAT_COMPRESSION, };/* Compatible feature bits */
@@ -366,6 +377,13 @@ typedef struct BDRVQcow2State {bool metadata_preallocation_checked;bool metadata_preallocation; +/* +* Compression type used for the image. Default: 0 - ZLIB +* The image compression type is set on image creation. +* For now, the only way to change the compression type +* is to convert the image with the desired compression type set. +*/ +Qcow2CompressionType compression_type; } BDRVQcow2State; typedef struct Qcow2COWRegion {
include/block/block_int.h +1 -0
@@ -57,6 +57,7 @@#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_PROBE_BUF_SIZE 512
block/qcow2.c +113 -0
191 lines not on this page · raw message
tests/qemu-iotests/031.out +7 -7
@@ -113,11 +113,11 @@ incompatible_features [] compatible_features [] autoclear_features [] refcount_order 4 -header_length 104 +header_length 112 Header extension: magic 0x6803f857 -length 288 +length 336 data <binary> Header extension:
@@ -146,11 +146,11 @@ incompatible_features [] compatible_features [] autoclear_features [] refcount_order 4 -header_length 104 +header_length 112 Header extension: magic 0x6803f857 -length 288 +length 336 data <binary> Header extension:
@@ -164,7 +164,7 @@ No errors were found on the image. magic 0x514649fb version 3 -backing_file_offset 0x1d8 +backing_file_offset 0x210 backing_file_size 0x17 cluster_bits 16 size 67108864
@@ -179,7 +179,7 @@ incompatible_features [] compatible_features [] autoclear_features [] refcount_order 4 -header_length 104 +header_length 112 Header extension: magic 0xe2792aca
@@ -188,7 +188,7 @@ data 'host_device' Header extension: magic 0x6803f857 -length 288 +length 336 data <binary> Header extension:
tests/qemu-iotests/036.out +2 -2
@@ -26,7 +26,7 @@ compatible_features [] autoclear_features [63] Header extension: magic 0x6803f857 -length 288 +length 336 data <binary>
@@ -38,7 +38,7 @@ compatible_features [] autoclear_features [] Header extension: magic 0x6803f857 -length 288 +length 336 data <binary> *** done
tests/qemu-iotests/049.out +51 -51
225 lines not on this page · raw message
tests/qemu-iotests/060.out +1 -0
@@ -17,6 +17,7 @@ virtual size: 64 MiB (67108864 bytes) cluster_size: 65536 Format specific information: compat: 1.1 + compression type: zlib lazy refcounts: false refcount bits: 16 corrupt: true
tests/qemu-iotests/061.out +20 -14
@@ -22,11 +22,11 @@ incompatible_features [] compatible_features [0] autoclear_features [] refcount_order 4 -header_length 104 +header_length 112 Header extension: magic 0x6803f857 -length 288 +length 336 data <binary> magic 0x514649fb
@@ -80,11 +80,11 @@ incompatible_features [] compatible_features [0] autoclear_features [] refcount_order 4 -header_length 104 +header_length 112 Header extension: magic 0x6803f857 -length 288 +length 336 data <binary> magic 0x514649fb
@@ -136,11 +136,11 @@ incompatible_features [0] compatible_features [0] autoclear_features [] refcount_order 4 -header_length 104 +header_length 112 Header extension: magic 0x6803f857 -length 288 +length 336 data <binary> ERROR cluster 5 refcount=0 reference=1
@@ -191,11 +191,11 @@ incompatible_features [] compatible_features [42] autoclear_features [42] refcount_order 4 -header_length 104 +header_length 112 Header extension: magic 0x6803f857 -length 288 +length 336 data <binary> magic 0x514649fb
@@ -260,11 +260,11 @@ incompatible_features [] compatible_features [0] autoclear_features [] refcount_order 4 -header_length 104 +header_length 112 Header extension: magic 0x6803f857 -length 288 +length 336 data <binary> read 65536/65536 bytes at offset 44040192
@@ -294,11 +294,11 @@ incompatible_features [0] compatible_features [0] autoclear_features [] refcount_order 4 -header_length 104 +header_length 112 Header extension: magic 0x6803f857 -length 288 +length 336 data <binary> ERROR cluster 5 refcount=0 reference=1
@@ -323,11 +323,11 @@ incompatible_features [] compatible_features [] autoclear_features [] refcount_order 4 -header_length 104 +header_length 112 Header extension: magic 0x6803f857 -length 288 +length 336 data <binary> read 131072/131072 bytes at offset 0
@@ -491,6 +491,7 @@ virtual size: 64 MiB (67108864 bytes) cluster_size: 65536 Format specific information: compat: 1.1 + compression type: zlib lazy refcounts: false refcount bits: 16 data file: TEST_DIR/t.IMGFMT.data
@@ -511,6 +512,7 @@ virtual size: 64 MiB (67108864 bytes) cluster_size: 65536 Format specific information: compat: 1.1 + compression type: zlib lazy refcounts: false refcount bits: 16 data file: foo
@@ -524,6 +526,7 @@ virtual size: 64 MiB (67108864 bytes) cluster_size: 65536 Format specific information: compat: 1.1 + compression type: zlib lazy refcounts: false refcount bits: 16 data file raw: false
@@ -538,6 +541,7 @@ virtual size: 64 MiB (67108864 bytes) cluster_size: 65536 Format specific information: compat: 1.1 + compression type: zlib lazy refcounts: false refcount bits: 16 data file: TEST_DIR/t.IMGFMT.data
@@ -550,6 +554,7 @@ virtual size: 64 MiB (67108864 bytes) cluster_size: 65536 Format specific information: compat: 1.1 + compression type: zlib lazy refcounts: false refcount bits: 16 data file: TEST_DIR/t.IMGFMT.data
@@ -563,6 +568,7 @@ virtual size: 64 MiB (67108864 bytes) cluster_size: 65536 Format specific information: compat: 1.1 + compression type: zlib lazy refcounts: false refcount bits: 16 data file: TEST_DIR/t.IMGFMT.data
tests/qemu-iotests/065 +18 -10
@@ -88,24 +88,30 @@ class TestQMP(TestImageInfoSpecific): class TestQCow2(TestQemuImgInfo): '''Testing a qcow2 version 2 image''' img_options = 'compat=0.10' - json_compare = { 'compat': '0.10', 'refcount-bits': 16 } - human_compare = [ 'compat: 0.10', 'refcount bits: 16' ] + json_compare = { 'compat': '0.10', 'refcount-bits': 16, + 'compression-type': 'zlib' } + human_compare = [ 'compat: 0.10', 'compression type: zlib', + 'refcount bits: 16' ] class TestQCow3NotLazy(TestQemuImgInfo): '''Testing a qcow2 version 3 image with lazy refcounts disabled''' img_options = 'compat=1.1,lazy_refcounts=off' json_compare = { 'compat': '1.1', 'lazy-refcounts': False, - 'refcount-bits': 16, 'corrupt': False } - human_compare = [ 'compat: 1.1', 'lazy refcounts: false', - 'refcount bits: 16', 'corrupt: false' ] + 'refcount-bits': 16, 'corrupt': False, + 'compression-type': 'zlib' } + human_compare = [ 'compat: 1.1', 'compression type: zlib', + 'lazy refcounts: false', 'refcount bits: 16', + 'corrupt: false' ] class TestQCow3Lazy(TestQemuImgInfo): '''Testing a qcow2 version 3 image with lazy refcounts enabled''' img_options = 'compat=1.1,lazy_refcounts=on' json_compare = { 'compat': '1.1', 'lazy-refcounts': True, - 'refcount-bits': 16, 'corrupt': False } - human_compare = [ 'compat: 1.1', 'lazy refcounts: true', - 'refcount bits: 16', 'corrupt: false' ] + 'refcount-bits': 16, 'corrupt': False, + 'compression-type': 'zlib' } + human_compare = [ 'compat: 1.1', 'compression type: zlib', + 'lazy refcounts: true', 'refcount bits: 16', + 'corrupt: false' ] class TestQCow3NotLazyQMP(TestQMP): '''Testing a qcow2 version 3 image with lazy refcounts disabled, opening
@@ -113,7 +119,8 @@ class TestQCow3NotLazyQMP(TestQMP): img_options = 'compat=1.1,lazy_refcounts=off' qemu_options = 'lazy-refcounts=on' compare = { 'compat': '1.1', 'lazy-refcounts': False, - 'refcount-bits': 16, 'corrupt': False } + 'refcount-bits': 16, 'corrupt': False, + 'compression-type': 'zlib' } class TestQCow3LazyQMP(TestQMP):
@@ -122,7 +129,8 @@ class TestQCow3LazyQMP(TestQMP): img_options = 'compat=1.1,lazy_refcounts=on' qemu_options = 'lazy-refcounts=off' compare = { 'compat': '1.1', 'lazy-refcounts': True, - 'refcount-bits': 16, 'corrupt': False } + 'refcount-bits': 16, 'corrupt': False, + 'compression-type': 'zlib' } TestImageInfoSpecific = None TestQemuImgInfo = None
tests/qemu-iotests/080 +1 -1
@@ -45,7 +45,7 @@ _supported_os Linux # - This is generally a test for compat=1.1 images _unsupported_imgopts 'refcount_bits=1[^0-9]' data_file 'compat=0.10' -header_size=104 +header_size=112 offset_backing_file_offset=8 offset_backing_file_size=16
tests/qemu-iotests/144.out +2 -2
@@ -9,7 +9,7 @@ Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=536870912 { 'execute': 'qmp_capabilities' } {"return": {}} { 'execute': 'blockdev-snapshot-sync', 'arguments': { 'device': 'virtio0', 'snapshot-file':'TEST_DIR/tmp.IMGFMT', 'format': 'IMGFMT' } } -Formatting 'TEST_DIR/tmp.qcow2', fmt=qcow2 size=536870912 backing_file=TEST_DIR/t.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 +Formatting 'TEST_DIR/tmp.qcow2', fmt=qcow2 size=536870912 backing_file=TEST_DIR/t.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib {"return": {}} === Performing block-commit on active layer ===
@@ -31,6 +31,6 @@ Formatting 'TEST_DIR/tmp.qcow2', fmt=qcow2 size=536870912 backing_file=TEST_DIR/ === Performing Live Snapshot 2 === { 'execute': 'blockdev-snapshot-sync', 'arguments': { 'device': 'virtio0', 'snapshot-file':'TEST_DIR/tmp2.IMGFMT', 'format': 'IMGFMT' } } -Formatting 'TEST_DIR/tmp2.qcow2', fmt=qcow2 size=536870912 backing_file=TEST_DIR/t.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 +Formatting 'TEST_DIR/tmp2.qcow2', fmt=qcow2 size=536870912 backing_file=TEST_DIR/t.qcow2 backing_fmt=qcow2 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib {"return": {}} *** done
tests/qemu-iotests/182.out +1 -1
@@ -13,7 +13,7 @@ Is another process using the image [TEST_DIR/t.qcow2]? {'execute': 'blockdev-add', 'arguments': { 'node-name': 'node0', 'driver': 'file', 'filename': 'TEST_DIR/t.IMGFMT', 'locking': 'on' } } {"return": {}} {'execute': 'blockdev-snapshot-sync', 'arguments': { 'node-name': 'node0', 'snapshot-file': 'TEST_DIR/t.IMGFMT.overlay', 'snapshot-node-name': 'node1' } } -Formatting 'TEST_DIR/t.qcow2.overlay', fmt=qcow2 size=197120 backing_file=TEST_DIR/t.qcow2 backing_fmt=file cluster_size=65536 lazy_refcounts=off refcount_bits=16 +Formatting 'TEST_DIR/t.qcow2.overlay', fmt=qcow2 size=197120 backing_file=TEST_DIR/t.qcow2 backing_fmt=file cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib {"return": {}} {'execute': 'blockdev-add', 'arguments': { 'node-name': 'node1', 'driver': 'file', 'filename': 'TEST_DIR/t.IMGFMT', 'locking': 'on' } } {"return": {}}
tests/qemu-iotests/242.out +5 -0
@@ -12,6 +12,7 @@ virtual size: 1 MiB (1048576 bytes) cluster_size: 65536 Format specific information: compat: 1.1 + compression type: zlib lazy refcounts: false refcount bits: 16 corrupt: false
@@ -32,6 +33,7 @@ virtual size: 1 MiB (1048576 bytes) cluster_size: 65536 Format specific information: compat: 1.1 + compression type: zlib lazy refcounts: false bitmaps: [0]:
@@ -64,6 +66,7 @@ virtual size: 1 MiB (1048576 bytes) cluster_size: 65536 Format specific information: compat: 1.1 + compression type: zlib lazy refcounts: false bitmaps: [0]:
@@ -104,6 +107,7 @@ virtual size: 1 MiB (1048576 bytes) cluster_size: 65536 Format specific information: compat: 1.1 + compression type: zlib lazy refcounts: false bitmaps: [0]:
@@ -153,6 +157,7 @@ virtual size: 1 MiB (1048576 bytes) cluster_size: 65536 Format specific information: compat: 1.1 + compression type: zlib lazy refcounts: false bitmaps: [0]:
tests/qemu-iotests/255.out +4 -4
@@ -3,9 +3,9 @@ Finishing a commit job with background reads === Create backing chain and start VM === -Formatting 'TEST_DIR/PID-t.qcow2.mid', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16 +Formatting 'TEST_DIR/PID-t.qcow2.mid', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib -Formatting 'TEST_DIR/PID-t.qcow2', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16 +Formatting 'TEST_DIR/PID-t.qcow2', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib === Start background read requests ===
@@ -23,9 +23,9 @@ Closing the VM while a job is being cancelled === Create images and start VM === -Formatting 'TEST_DIR/PID-src.qcow2', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16 +Formatting 'TEST_DIR/PID-src.qcow2', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib -Formatting 'TEST_DIR/PID-dst.qcow2', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16 +Formatting 'TEST_DIR/PID-dst.qcow2', fmt=qcow2 size=134217728 cluster_size=65536 lazy_refcounts=off refcount_bits=16 compression_type=zlib wrote 1048576/1048576 bytes at offset 0 1 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
tests/qemu-iotests/common.filter +2 -1
@@ -152,7 +152,8 @@ _filter_img_create() -e "s# refcount_bits=[0-9]\\+##g" \ -e "s# key-secret=[a-zA-Z0-9]\\+##g" \ -e "s# iter-time=[0-9]\\+##g" \ - -e "s# force_size=\\(on\\|off\\)##g" + -e "s# force_size=\\(on\\|off\\)##g" \ + -e "s# compression_type=[a-zA-Z0-9]\\+##g" } _filter_img_info()
Signature · 1 line
2.17.0
Re: [PATCH v13 3/4] qcow2: add zstd cluster compression
In reply to Denis PlotnikovQuoted 29 lines · Denis Plotnikov
zstd significantly reduces cluster compression time. It provides better compression performance maintaining the same level of the compression ratio in comparison with zlib, which, at the moment, is the only compression method available.
The performance test results:
Test compresses and decompresses qemu qcow2 image with just installed rhel-7.6 guest.
Image cluster size: 64K. Image on disk size: 2.2GThe test was conducted with brd disk to reduce the influence of disk subsystem to the test results.
The results is given in seconds.compress cmd:
time ./qemu-img convert -O qcow2 -c -o compression_type=[zlib|zstd]
src.img [zlib|zstd]_compressed.img
decompress cmd
time ./qemu-img convert -O qcow2
[zlib|zstd]_compressed.img uncompressed.imgcompression decompression
zlib zstd zlib zstd
------------------------------------------------------------real 65.5 16.3 (-75 %) 1.9 1.6 (-16 %) user 65.0 15.8 5.3 2.5 sys 3.3 0.2 2.0 2.0
Quoted 7 lines
Both ZLIB and ZSTD gave the same compression ratio: 1.57 compressed image size in both cases: 1.4G
Signed-off-by Denis Plotnikov <[email protected]>QAPI part:
---
Hi!
I have three suggestions:
1. return type of qcow2 compression functions is ssize_t, and return type of zstd is size_t. I think better is not to mix them.
2. safe check for ENOMEM in compression part - avoid overflow ( maximum of paranoia :)
3. slightly more tolerate sanity check in decompression part: allow zstd to make progress only on one of the buffers
Here is it, and if you like it, take it together with my
block/qcow2-threads.c +13 -10
@@ -188,7 +188,7 @@ static ssize_t qcow2_zlib_decompress(void *dest, size_t dest_size,static ssize_t qcow2_zstd_compress (void *dest,size_t dest_size,const void *src,size_t src_size) { -size_t ret; +ssize_t ret;ZSTD_outBuffer output = { dest, dest_size,0 };ZSTD_inBuffer input = { src, src_size,0 };ZSTD_CCtx *cctx;
@@ -214,13 +214,14 @@ static ssize_t qcow2_zstd_compress(void *dest, size_t dest_size,* We want to use zstd streamed interface on decompression, * as we won't know the exact size of the compressed data. */ - ret =ZSTD_compressStream2 (cctx, &output, &input, ZSTD_e_end); -if (ZSTD_isError (ret)) { +size_t zstd_ret =ZSTD_compressStream2 (cctx, &output, &input, + ZSTD_e_end); +if (ZSTD_isError (zstd_ret)) { ret = -EIO ; goto out; }/* Dest buffer isn't big enough to store compressed content */ -if (output.pos + ret > output.size) { +if (zstd_ret > output.size - output.pos) { ret = -ENOMEM ; goto out; }
@@ -248,7 +249,8 @@ out:static ssize_t qcow2_zstd_decompress (void *dest,size_t dest_size,const void *src,size_t src_size) { -size_t ret =0 ; +ssize_t ret =0 ; +size_t zstd_ret =0 ;ZSTD_outBuffer output = { dest, dest_size,0 };ZSTD_inBuffer input = { src, src_size,0 };ZSTD_DCtx *dctx =ZSTD_createDCtx ();
@@ -269,8 +271,9 @@ static ssize_t qcow2_zstd_decompress(void *dest, size_t dest_size,* ZSTD_decompressStream reads another ONE full frame. */ while (output.pos < output.size) { -size_t last_input_pos = input.pos; - ret =ZSTD_decompressStream (dctx, &output, &input); +size_t last_progress = input.pos + output.pos; + + zstd_ret =ZSTD_decompressStream (dctx, &output, &input);/* * zstd manual doesn't explicitly states what happens, * if the read the frame partially. But, based on the
@@ -278,7 +281,7 @@ static ssize_t qcow2_zstd_decompress(void *dest, size_t dest_size,* and have read all the frames from the input, * we end up with error here. */ -if (ZSTD_isError (ret)) { +if (ZSTD_isError (zstd_ret)) { ret = -EIO ;break ; }
@@ -286,7 +289,7 @@ static ssize_t qcow2_zstd_decompress(void *dest, size_t dest_size,/* * Sanity check that each time we do some progress */ -if (last_input_pos >= input.pos) { +if (last_progress >= input.pos + output.pos) { ret = -EIO ;break ; }
@@ -297,7 +300,7 @@ static ssize_t qcow2_zstd_decompress(void *dest, size_t dest_size,* if not, we somehow managed to get uncompressed cluster * greater then the cluster size. */ -if (ret >0 ) { +if (zstd_ret >0 ) { ret = -EIO ; }
Signature · 2 lines
Best regards, Vladimir
Re: [PATCH v13 3/4] qcow2: add zstd cluster compression
Quoted 81 lines · Vladimir Sementsov-Ogievskiy
Quoted 29 lines · Denis Plotnikov
zstd significantly reduces cluster compression time. It provides better compression performance maintaining the same level of the compression ratio in comparison with zlib, which, at the moment, is the only compression method available.
The performance test results:
Test compresses and decompresses qemu qcow2 image with just installed rhel-7.6 guest.
Image cluster size: 64K. Image on disk size: 2.2GThe test was conducted with brd disk to reduce the influence of disk subsystem to the test results.
The results is given in seconds.compress cmd:
time ./qemu-img convert -O qcow2 -c -o compression_type=[zlib|zstd] src.img [zlib|zstd]_compressed.img decompress cmd
time ./qemu-img convert -O qcow2 [zlib|zstd]_compressed.img uncompressed.imgcompression decompression zlib zstd zlib zstd
------------------------------------------------------------
real 65.5 16.3 (-75 %) 1.9 1.6 (-16 %) user 65.0 15.8 5.3 2.5 sys 3.3 0.2 2.0 2.0Quoted 7 lines
Both ZLIB and ZSTD gave the same compression ratio: 1.57 compressed image size in both cases: 1.4G
Signed-off-by Denis Plotnikov <[email protected]>QAPI part:
---
Hi!
I have three suggestions:
1. return type of qcow2 compression functions is
ssize_t, and return type of zstd issize_t. I think better is not to mix them.
2. safe check for ENOMEM in compression part - avoid overflow ( maximum of paranoia :)
3. slightly more tolerate sanity check in decompression part: allow zstd to make progress only on one of the buffersHere is it, and if you like it, take it together with my
Reviewed-by Vladimir Sementsov-Ogievskiy <[email protected]>block/qcow2-threads.c +1
-1@@ -188,7 +188,7 @@ static ssize_t qcow2_zlib_decompress(void *dest,size_t dest_size, static ssize_t qcow2_zstd_compress(void *dest, size_t dest_size, const void *src, size_t src_size) { - size_t ret; + ssize_t ret; ZSTD_outBuffer output = { dest, dest_size, 0 }; ZSTD_inBuffer input = { src, src_size, 0 };
ZSTD_CCtx*cctx;block/qcow2-threads.c +4
-3@@ -214,13 +214,14 @@ static ssize_t qcow2_zstd_compress(void *dest,size_t dest_size, * We want to use zstd streamed interface on decompression, * as we won't know the exact size of the compressed data. */ - ret = ZSTD_compressStream2(cctx, &output, &input, ZSTD_e_end); - if (ZSTD_isError(ret)) { + size_t zstd_ret = ZSTD_compressStream2(cctx, &output, &input, + ZSTD_e_end); + if (ZSTD_isError(zstd_ret)) { ret = -EIO; goto out; } /* Dest buffer isn't big enough to store compressed content */ - if (output.pos + ret > output.size) { + if (zstd_ret > output.size - output.pos) { ret = -ENOMEM; goto out;
Quoted 1 line
}
I'll apply this part as you suggested
Quoted 11 lines
block/qcow2-threads.c +2
-1@@ -248,7 +249,8 @@ out: static ssize_t qcow2_zstd_decompress(void *dest, size_t dest_size, const void *src, size_t src_size) { - size_t ret = 0; + ssize_t ret = 0; + size_t zstd_ret = 0; ZSTD_outBuffer output = { dest, dest_size, 0 }; ZSTD_inBuffer input = { src, src_size, 0 }; ZSTD_DCtx *dctx = ZSTD_createDCtx();
block/qcow2-threads.c static ssize_t qcow2_zstd_decompress(void *dest, · line 275
Quoted 7 lines
block/qcow2-threads.c +1
-2@@ -269,8 +271,9 @@ static ssize_t qcow2_zstd_decompress(void *dest,size_t dest_size, * ZSTD_decompressStream reads another ONE full frame. */ while (output.pos < output.size) { - size_t last_input_pos = input.pos; - ret = ZSTD_decompressStream(dctx, &output, &input); + size_t last_progress = input.pos + output.pos;
These two may overflow last_progress. In that case we have to check them in separate
Quoted 26 lines
block/qcow2-threads.c +2
-0+ + zstd_ret = ZSTD_decompressStream(dctx, &output, &input); /** zstd manual doesn't explicitly states what happens, * if the read the frame partially. But, based on the
block/qcow2-threads.c +1
-1@@ -278,7 +281,7 @@ static ssize_t qcow2_zstd_decompress(void *dest,size_t dest_size, * and have read all the frames from the input, * we end up with error here. */ - if (ZSTD_isError(ret)) { + if (ZSTD_isError(zstd_ret)) { ret = -EIO; break;}
block/qcow2-threads.c +1
-1@@ -286,7 +289,7 @@ static ssize_t qcow2_zstd_decompress(void *dest,size_t dest_size, /* * Sanity check that each time we do some progress */ - if (last_input_pos >= input.pos) { + if (last_progress >= input.pos + output.pos) { ret = -EIO; break;}
block/qcow2-threads.c static ssize_t qcow2_zstd_decompress(void *dest, · line 306
Quoted 8 lines
block/qcow2-threads.c +1
-1@@ -297,7 +300,7 @@ static ssize_t qcow2_zstd_decompress(void *dest,size_t dest_size, * if not, we somehow managed to get uncompressed cluster * greater then the cluster size. */ - if (ret > 0) { + if (zstd_ret > 0) { ret = -EIO; }
I'll apply the changes and resend the series.
Thanks!