qcow2: Implement zstd cluster compression method
v9 cover qcow2: Implement zstd cluster compression method
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 | 205 +++++++++++++++++++++++++++++-- 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, 643 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 v9 2/4 qcow2: rework the cluster compression routine
The 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 v9 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
@@ -293,3 +293,4 @@ 283 auto quick 284 rw 286 rw quick +287 auto quick
Signature · 1 line
2.17.0
PATCH v9 1/4 qcow2: introduce compression type feature
In reply to Denis PlotnikovThe 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, 255--- 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' } } ##
@@ -4392,6 +4395,18 @@ '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: #
@@ -4415,6 +4430,8 @@ # 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 ##
@@ -4430,7 +4447,8 @@ '*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 (sizeof (QCowHeader) %8 !=0 ); + 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 */
@@ -369,6 +380,13 @@ 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 192 +length 240 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 192 +length 240 data <binary> Header extension:
@@ -164,7 +164,7 @@ No errors were found on the image. magic 0x514649fb version 3 -backing_file_offset 0x178 +backing_file_offset 0x1b0 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 192 +length 240 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 192 +length 240 data <binary>
@@ -38,7 +38,7 @@ compatible_features [] autoclear_features [] Header extension: magic 0x6803f857 -length 192 +length 240 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 192 +length 240 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 192 +length 240 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 192 +length 240 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 192 +length 240 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 192 +length 240 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 192 +length 240 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 192 +length 240 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
PATCH v9 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 | 134 +++++++++++++++++++++++++++++++++++++++++ block/qcow2.c | 7 +++ 5 files changed, 145 insertions(+), 2 deletions(-)
docs/interop/qcow2.txt +1 -0
@@ -208,6 +208,7 @@ version 2. Available compression type values: 0: zlib <https://www.zlib.net/> + 1: zstd <http://github.com/facebook/zstd> === Header padding ===
configure +1 -1
@@ -1835,7 +1835,7 @@ 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
@@ -4401,11 +4401,12 @@ # 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 +134 -0
158 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 :
@@ -3454,6 +3457,10 @@ 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
Re: [PATCH v9 1/4] qcow2: introduce compression type feature
In reply to Denis PlotnikovQuoted 50 lines · Denis Plotnikov
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, 255Signed-off-by Denis Plotnikov <[email protected]>--- 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(-)
[..]
Quoted 5 lines
block/qcow2.h +0
-0@@ -146,8 +146,16 @@ typedef struct QCowHeader {uint32_t refcount_order;
block/qcow2.h typedef struct QCowHeader { · line 157
Quoted 10 lines
block/qcow2.h +7
-0uint32_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 (sizeof (QCowHeader) %8 !=0 );
Better write
QEMU_BUILD_BUG_ON(QEMU_IS_ALIGNED(sizeof(QCowHeader), 8);
with or without this:
Quoted 36 lines
block/qcow2.h +12
-1+ 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 */ @@ -369,6 +380,13 @@ 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 {
[..]
Signature · 2 lines
Best regards, Vladimir
Re: [PATCH v9 3/4] qcow2: add zstd cluster compression
In reply to Denis PlotnikovShould we note somehow in qcow2 spec that we use streamed version of zstd with specific end byte?
Quoted 36 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 2.0Both 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:
---
[..]
Quoted 93 lines
block/qcow2-threads.c +93
-0+ +/* +* qcow2_zstd_compress() +* +* Compress @src_size bytes of data using zstd compression method +* +* @dest - destination buffer, @dest_size bytes +* @src - source buffer, @src_size bytes +* +* Returns: compressed size on success +* -ENOMEM destination buffer is not enough to store compressed data +* -EIO on any other error +*/ +static ssize_t qcow2_zstd_compress (void *dest,size_t dest_size, +const void *src,size_t src_size) +{ +size_t ret; +ZSTD_outBuffer output = { dest, dest_size,0 }; +ZSTD_inBuffer input = { src, src_size,0 }; +ZSTD_CCtx *cctx =ZSTD_createCCtx (); + +if (!cctx) { +return -EIO ; + } +/* +* ZSTD spec: "You must continue calling ZSTD_compressStream2() +* with ZSTD_e_end until it returns 0, at which point you are +* free to start a new frame" +*/ + { +/* +* zstd simple interface requires the exact compressed size. +* zstd stream interface reads the comressed size from +* the compressed stream frame. +* Instruct zstd to compress the whole buffer and write +* the frame which includes the compressed size. +* This allows as to use zstd streaming semantics and +* don't store the compressed size for the zstd decompression. +*/ + ret =ZSTD_compressStream2 (cctx, &output, &input, ZSTD_e_end); +if (ZSTD_isError (ret)) { + ret = -EIO ; + goto out; + } +/* Dest buffer isn't big enough to store compressed content */ +if (output.pos + ret > output.size) { + ret = -ENOMEM ; + goto out; + } + }while (ret); + +/* if no error, the input data must be fully consumed */ +assert (input.pos == input.size); + /* make sure we can safely return compressed buffer size with ssize_t *//z + assert(output.pos <= SSIZE_MAX); + ret = output.pos; + +out: +ZSTD_freeCCtx (cctx); +return ret; +} + +/* +* qcow2_zstd_decompress() +* +* Decompress some data (not more than @src_size bytes) to produce exactly +* @dest_size bytes using zstd compression method +* +* @dest - destination buffer, @dest_size bytes +* @src - source buffer, @src_size bytes +* +* Returns: 0 on success +* -EIO on any error +*/ +static ssize_t qcow2_zstd_decompress (void *dest,size_t dest_size, +const void *src,size_t src_size) +{ +size_t ret =0 ; +ZSTD_outBuffer output = { dest, dest_size,0 }; +ZSTD_inBuffer input = { src, src_size,0 }; +ZSTD_DCtx *dctx =ZSTD_createDCtx (); + +if (!dctx) { +return -EIO ; + } + + { + ret =ZSTD_decompressStream (dctx, &output, &input); +if (ZSTD_isError (ret)) { + ret = -EIO ; + goto out; + } +/*
block/qcow2-threads.c static ssize_t qcow2_zlib_decompress(void *dest, size_t dest_size, · line 277
Quoted 10 lines
block/qcow2-threads.c +10
-0+* Dest buffer size is the image cluster size. +* It should be big enough to store uncompressed content. +* There shouldn't be any cases when the decompressed content +* size is greater then the cluster size. +*/ +if (output.pos + ret > output.size) { + ret = -EIO ; + goto out; + } + }while (ret);
Hmm. Unfortunately, zstd spec is not enough verbose to understand how to use these functions :).
But I found this in comment in https://github.com/facebook/zstd/blob/dev/examples/streaming_decompression.c :
/* The return code is zero if the frame is complete, but there may
* be multiple frames concatenated together.
so, I think it would be safer to move to "while (input.pos < input.size)" loop like in example.
and drop next assertion. And possibly do same for compression to be consistent (and safer?).
Quoted 19 lines
block/qcow2-threads.c +15
-0+ +/* +* If decompression went fine we must have the compressed +* cluster fully consumed and flushed +*/ +if (output.pos != output.size) { + ret = -EIO ; + } +out: +ZSTD_freeDCtx (dctx); + return ret; + +} +#endif +static int qcow2_compress_pool_func (void *opaque) {Qcow2CompressData *data = opaque;@@ -217,6 +341,11 @@ qcow2_co_compress(BlockDriverState *bs, void *dest, size_t dest_size,
[..]
Signature · 2 lines
Best regards, Vladimir
Re: [PATCH v9 3/4] qcow2: add zstd cluster compression
Quoted 1 line · Vladimir Sementsov-Ogievskiy
Should we note somehow in qcow2 spec that we use streamed version of zstd with specific end byte?
We didn't do it for zlib. zstd does it the same way as zlib, saves the compression output to some buffer.
Quoted 152 lines
Quoted 36 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 2.0Both 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:
---
[..]
Quoted 96 lines
block/qcow2-threads.c +11
-0+ +/* +* qcow2_zstd_compress() +* +* Compress @src_size bytes of data using zstd compression method +* +* @dest - destination buffer, @dest_size bytes +* @src - source buffer, @src_size bytes +* +* Returns: compressed size on success +* -ENOMEM destination buffer is not enough to store compressed data
+ * -EIO on any other error
+ */block/qcow2-threads.c +41
-0+static ssize_t qcow2_zstd_compress (void *dest,size_t dest_size, + const void *src, size_t src_size) +{ + size_t ret; + ZSTD_outBuffer output = { dest, dest_size, 0 }; + ZSTD_inBuffer input = { src, src_size, 0 }; + ZSTD_CCtx *cctx = ZSTD_createCCtx(); + + if (!cctx) { + return -EIO; + } + /* + * ZSTD spec: "You must continue calling ZSTD_compressStream2() + * with ZSTD_e_end until it returns 0, at which point you are + * free to start a new frame" + */ + { + /* + * zstd simple interface requires the exact compressed size. + * zstd stream interface reads the comressed size from + * the compressed stream frame. + * Instruct zstd to compress the whole buffer and write + * the frame which includes the compressed size. + * This allows as to use zstd streaming semantics and + * don't store the compressed size for the zstd decompression. + */ + ret = ZSTD_compressStream2(cctx, &output, &input, ZSTD_e_end); + if (ZSTD_isError(ret)) { + ret = -EIO; + goto out; + } + /* Dest buffer isn't big enough to store compressed content */ + if (output.pos + ret > output.size) { + ret = -ENOMEM; + goto out; + } + } while (ret); + + /* if no error, the input data must be fully consumed */ + assert(input.pos == input.size); + /* make sure we can safely return compressed buffer size with
ssize_t*//z
+assert(output.pos <= SSIZE_MAX);
+ ret = output.pos;block/qcow2-threads.c +10
-0+ +out: + ZSTD_freeCCtx(cctx); + return ret; +} + +/* +* qcow2_zstd_decompress() +* +* Decompress some data (not more than @src_size bytes) to produce exactly
+ * @dest_size bytes using zstd compression method
+ *
+ * @dest - destination buffer, @dest_size bytes
+ * @src - source buffer, @src_size bytes
+ *
+ * Returns: 0 on success
+ *-EIOon any error
+ */block/qcow2-threads.c +19
-0+static ssize_t qcow2_zstd_decompress (void *dest,size_t dest_size, + const void *src, size_t src_size) +{ + size_t ret = 0; + ZSTD_outBuffer output = { dest, dest_size, 0 }; + ZSTD_inBuffer input = { src, src_size, 0 }; + ZSTD_DCtx *dctx = ZSTD_createDCtx(); + + if (!dctx) { + return -EIO; + } + + { + ret = ZSTD_decompressStream(dctx, &output, &input); + if (ZSTD_isError(ret)) { + ret = -EIO; + goto out; + } + /*block/qcow2-threads.c static ssize_t qcow2_zlib_decompress(void *dest, size_t dest_size, · line 277
Quoted 10 lines
block/qcow2-threads.c +10
-0+* Dest buffer size is the image cluster size. +* It should be big enough to store uncompressed content. +* There shouldn't be any cases when the decompressed content +* size is greater then the cluster size. +*/ + if (output.pos + ret > output.size) { + ret = -EIO; + goto out; + } + } while (ret);Hmm. Unfortunately, zstd spec is not enough verbose to understand how to use these functions :).
Quoted 4 lines
But I found this in comment in https://github.com/facebook/zstd/blob/dev/examples/streaming_decompression.c :
/* The return code is zero if the frame is complete, but there may
* be multiple frames concatenated together.
In our case, there should be the only frame as we stored it with ZSTD_compressStream2(..., ZSTD_e_end)
If explicitly state in the spec, that "zstd compressed cluster must contain the only zstd compressed frame"
On the other hand, we actually don't need such a restriction and we should be able to decompress whatever is in the buffer (cluster) if it's zstd stream.
Quoted 2 lines
so, I think it would be safer to move to "while (input.pos < input.size)" loop like in example.
yes, I think It worth doing
Quoted 29 lines
and drop next assertion. And possibly do same for compression to be consistent (and safer?).
Quoted 20 lines
block/qcow2-threads.c +15
-0+ + /* + * If decompression went fine we must have the compressed + * cluster fully consumed and flushed + */ + if (output.pos != output.size) { + ret = -EIO; + } +out: + ZSTD_freeDCtx(dctx); + return ret; + +} +#endif +static int
qcow2_compress_pool_func(void *opaque){
Qcow2CompressData *data = opaque;(quoted diff) +0
-0@@ -217,6 +341,11 @@ qcow2_co_compress(BlockDriverState *bs, void *dest, size_t dest_size,[..]
Re: [PATCH v9 3/4] qcow2: add zstd cluster compression
Quoted 5 lines · Denis Plotnikov
Quoted 1 line · Vladimir Sementsov-Ogievskiy
Should we note somehow in qcow2 spec that we use streamed version of zstd with specific end byte?
We didn't do it for zlib. zstd does it the same way as zlib, saves the compression output to some buffer.
Looked through zstd format spec, I don't thing that there are may be any principal difference between data compressed by simple api or by streamed, so now I don't think that we need any additional comment.
Quoted 191 lines
Quoted 150 lines
Quoted 36 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 2.0Both 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:
---
[..]
Quoted 93 lines
block/qcow2-threads.c +93
-0+ +/* +* qcow2_zstd_compress() +* +* Compress @src_size bytes of data using zstd compression method +* +* @dest - destination buffer, @dest_size bytes +* @src - source buffer, @src_size bytes +* +* Returns: compressed size on success +* -ENOMEM destination buffer is not enough to store compressed data +* -EIO on any other error +*/ +static ssize_t qcow2_zstd_compress (void *dest,size_t dest_size, + const void *src, size_t src_size) +{ + size_t ret; + ZSTD_outBuffer output = { dest, dest_size, 0 }; + ZSTD_inBuffer input = { src, src_size, 0 }; + ZSTD_CCtx *cctx = ZSTD_createCCtx(); + + if (!cctx) { + return -EIO; + } + /* + * ZSTD spec: "You must continue calling ZSTD_compressStream2() + * with ZSTD_e_end until it returns 0, at which point you are + * free to start a new frame" + */ + { + /* + * zstd simple interface requires the exact compressed size. + * zstd stream interface reads the comressed size from + * the compressed stream frame. + * Instruct zstd to compress the whole buffer and write + * the frame which includes the compressed size. + * This allows as to use zstd streaming semantics and + * don't store the compressed size for the zstd decompression. + */ + ret = ZSTD_compressStream2(cctx, &output, &input, ZSTD_e_end); + if (ZSTD_isError(ret)) { + ret = -EIO; + goto out; + } + /* Dest buffer isn't big enough to store compressed content */ + if (output.pos + ret > output.size) { + ret = -ENOMEM; + goto out; + } + } while (ret); + + /* if no error, the input data must be fully consumed */ + assert(input.pos == input.size); + /* make sure we can safely return compressed buffer size with ssize_t *//z + assert(output.pos <= SSIZE_MAX); + ret = output.pos; + +out: + ZSTD_freeCCtx(cctx); + return ret; +} + +/* +* qcow2_zstd_decompress() +* +* Decompress some data (not more than @src_size bytes) to produce exactly +* @dest_size bytes using zstd compression method +* +* @dest - destination buffer, @dest_size bytes +* @src - source buffer, @src_size bytes +* +* Returns: 0 on success +* -EIO on any error +*/ +static ssize_t qcow2_zstd_decompress (void *dest,size_t dest_size, + const void *src, size_t src_size) +{ + size_t ret = 0; + ZSTD_outBuffer output = { dest, dest_size, 0 }; + ZSTD_inBuffer input = { src, src_size, 0 }; + ZSTD_DCtx *dctx = ZSTD_createDCtx(); + + if (!dctx) { + return -EIO; + } + + { + ret = ZSTD_decompressStream(dctx, &output, &input); + if (ZSTD_isError(ret)) { + ret = -EIO; + goto out; + } + /*block/qcow2-threads.c static ssize_t qcow2_zlib_decompress(void *dest, size_t dest_size, · line 277
Quoted 10 lines
block/qcow2-threads.c +10
-0+* Dest buffer size is the image cluster size. +* It should be big enough to store uncompressed content. +* There shouldn't be any cases when the decompressed content +* size is greater then the cluster size. +*/ + if (output.pos + ret > output.size) { + ret = -EIO; + goto out; + } + } while (ret);Hmm. Unfortunately, zstd spec is not enough verbose to understand how to use these functions :).
Quoted 4 lines
But I found this in comment in https://github.com/facebook/zstd/blob/dev/examples/streaming_decompression.c :
/* The return code is zero if the frame is complete, but there may
* be multiple frames concatenated together.In our case, there should be the only frame as we stored it with
ZSTD_compressStream2(..., ZSTD_e_end)
If explicitly state in the spec, that "zstd compressed cluster must contain the only zstd compressed frame"
On the other hand, we actually don't need such a restriction and we should be able to decompress whatever is in the buffer (cluster) if it's zstd stream.Quoted 2 lines
so, I think it would be safer to move to "while (input.pos < input.size)" loop like in example.
yes, I think It worth doing
Quoted 28 lines
and drop next assertion. And possibly do same for compression to be consistent (and safer?).
Quoted 19 lines
block/qcow2-threads.c +15
-0+ + /* + * If decompression went fine we must have the compressed + * cluster fully consumed and flushed + */ + if (output.pos != output.size) { + ret = -EIO; + } +out: + ZSTD_freeDCtx(dctx); + return ret; + +} +#endif +static int
qcow2_compress_pool_func(void *opaque){
Qcow2CompressData *data = opaque;(quoted diff) +0
-0@@ -217,6 +341,11 @@ qcow2_co_compress(BlockDriverState *bs, void *dest, size_t dest_size,[..]
Signature · 2 lines
Best regards, Vladimir
Re: [PATCH v9 1/4] qcow2: introduce compression type feature
In reply to Denis PlotnikovQuoted 30 lines · Denis Plotnikov
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, 255
Quoted 1 line
Signed-off-by Denis Plotnikov <[email protected]>
Hmm, haven't I've seen this before? ... Yes, I provided my QAPI part:
Acked-by for v5, and the QAPI part hasn't changed since. You can save me a bit of labor by carrying my Acked-by forward.
Once again, QAPI part:
Re: [PATCH v9 1/4] qcow2: introduce compression type feature
Quoted 32 lines · Markus Armbruster
Quoted 30 lines · Denis Plotnikov
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, 255
Quoted 4 lines
Quoted 1 line
Signed-off-by Denis Plotnikov <[email protected]>Hmm, haven't I've seen this before? ... Yes, I provided my QAPI part:
Acked-by for v5, and the QAPI part hasn't changed since. You can save me a bit of labor by carrying my Acked-by forward.
For sure, next time. Thanks!
Quoted 4 lines
Once again, QAPI part: