Skip to contentSkip to conversation

qcow2: Implement zstd cluster compression method

PATCH Superseded Denis Plotnikov 3 participants 13 messages Mar 31, 2020 – Apr 1, 2020 Block layer

v14 of 24 Superseded Latest is v24 →

lore
Share & tools
lore
#0 lore raw

v14 cover qcow2: Implement zstd cluster compression method

v14:
* fix bug on compression - looping until compress == 0 [Me]
* apply reworked Vladimir's suggestions:
1. not mixing ssize_t with size_t
2. safe check for ENOMEM in compression part - avoid overflow
3. tolerate sanity check allow zstd to make progress only on one of the buffers
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            | 233 +++++++++++++++++++++++++++++--
 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, 671 insertions(+), 108 deletions(-)
 create mode 100755 tests/qemu-iotests/287
 create mode 100644 tests/qemu-iotests/287.out
Signature · 1 line
2.17.0
#1 lore raw

PATCH v14 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, 255

QAPI 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
index 943df1926a..d669ec0965 100644
@@ -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
index f4de0a27d5..6a8b82e6cc 100644
@@ -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
index 4c3587ea19..bc91b69f46 100644
@@ -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
index 46f97c5a4e..5a4beda6a2 100644
@@ -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
index 23b699ce06..e409acf60e 100644
@@ -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
index d27692a33c..3e47f537e8 100644
@@ -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
index 413cc4e0f4..45c029228c 100644
@@ -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
index 6426474271..18dc488c7a 100755
@@ -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
index a3d13c414e..7588c63b6c 100755
@@ -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
index c7aa2e4820..885a8874a5 100644
@@ -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
index a8eea166c3..ae43654d32 100644
@@ -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
index 7ac8404d11..091b9126ce 100644
@@ -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
index 348909fdef..a3c99fd62e 100644
@@ -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
index 3f8ee3e5f7..279e0bbb0d 100644
@@ -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
Reviewed-by on v14 1/4
#2 lore raw

Re: [PATCH v14 1/4] qcow2: introduce compression type feature

Quoted 38 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

QAPI part:

---

+++ b/qapi/block-core.json

qapi/block-core.json line 81

Quoted 4 lines
qapi/block-core.json +1 -0
@@ -78,6 +78,8 @@
  #
  # @bitmaps: A list of qcow2 bitmap details (since 4.0)
  #
+# @compression-type: the image cluster compression method (since 5.0)

At this point (-rc1 is due today), I'm thinking that this is going to land in 5.1 rather than 5.0. That will have some ripple effects on the patch.

Otherwise,

Signature · 3 lines
Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3226
Virtualization:  qemu.org | libvirt.org
#3 lore raw

PATCH v14 2/4 qcow2: rework the cluster compression routine

In reply to Denis Plotnikov

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
index a68126f291..7dbaf53489 100644
@@ -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_fn
 qcow2_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_fn
 qcow2_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
#4 lore raw

PATCH v14 3/4 qcow2: add zstd cluster compression

In reply to 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: 2.2G

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  | 162 +++++++++++++++++++++++++++++++++++++++++
 block/qcow2.c          |   7 ++
 5 files changed, 173 insertions(+), 2 deletions(-)
docs/interop/qcow2.txt +1 -0
index 640e0eca40..18a77f737e 100644
@@ -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
index e225a1e3ff..fdc991b010 100755
@@ -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
index d669ec0965..44690ed266 100644
@@ -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 +162 -0

186 lines not on this page · raw message

block/qcow2.c +7 -0
index 6cb000be19..3ae9624ba3 100644
@@ -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
#5 lore raw

Re: [PATCH v14 3/4] qcow2: add zstd cluster compression

Quoted 30 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.

QAPI part:

---

block/qcow2-threads.c +17 -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)
+{
+    ssize_t ret;
+    size_t zstd_ret = 0;

block/qcow2-threads.c static ssize_t qcow2_zlib_decompress(void *dest, size_t dest_size, · line 202

Quoted 10 lines
block/qcow2-threads.c +10 -0
+    ZSTD_outBuffer output = { dest, dest_size, 0 };
+    ZSTD_inBuffer input = { src, src_size, 0 };
+    ZSTD_CCtx *cctx = ZSTD_createCCtx();
+
+    if (!cctx) {
+        return -EIO;
+    }
+    /*
+     * We want to use zstd streamed interface on decompression,
+     * as we won't know the exact size of the compressed data.

This comment feels like it belongs to decompression rather than compression. We _do_ know the exact source size of the data to be compressed, but we don't know the resulting compressed size. Maybe:

'Use the zstd streamed interface for symmetry with decompression, where streaming is essential since we don't record the exact compressed size.'

Quoted 25 lines
block/qcow2-threads.c +25 -0
+     *
+     * In the loop, we try to compress all the data into one zstd frame.
+     * ZSTD_compressStream2 potentially can finish a frame earlier
+     * than the full input data is consumed. That's why we are looping
+     * until all the input data is consumed.
+     */
+    while (input.pos < input.size) {
+        /*
+         * 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". We assume that "start a new frame"
+         * means call ZSTD_compressStream2 in the very beginning or when
+         * ZSTD_compressStream2 has returned with 0.
+         */
+        do {
+            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 (zstd_ret > output.size - output.pos) {
+                ret = -ENOMEM;
+                goto out;

block/qcow2-threads.c static ssize_t qcow2_zlib_decompress(void *dest, size_t dest_size, · line 237

Quoted 10 lines
block/qcow2-threads.c +10 -0
+            }
+        } while (zstd_ret);
+    }
+    /* make sure we can safely return compressed buffer size with ssize_t */
+    assert(output.pos <= SSIZE_MAX);
+    ret = output.pos;
+out:
+    ZSTD_freeCCtx(cctx);
+    return ret;
+}

I didn't closely check the zstd documentation, but based on naming, this loop looks reasonable. And of course, if you can read an image created with this code, then the round trip proves it was used correctly for that case.

Quoted 18 lines
block/qcow2-threads.c +18 -0
+
+/*
+ * 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 };

block/qcow2-threads.c static ssize_t qcow2_zlib_decompress(void *dest, size_t dest_size, · line 265

Quoted 10 lines
block/qcow2-threads.c +10 -0
+    ZSTD_inBuffer input = { src, src_size, 0 };
+    ZSTD_DCtx *dctx = ZSTD_createDCtx();
+
+    if (!dctx) {
+        return -EIO;
+    }
+
+    /*
+     * The compressed stream from input buffer may consist from more
+     * than one zstd frames. So we iterate until we get a fully

The compressed stream from the input buffer may consist of more than one zstd frame. So we iterate...

block/qcow2-threads.c static ssize_t qcow2_zlib_decompress(void *dest, size_t dest_size, · line 270

Quoted 5 lines
block/qcow2-threads.c +5 -0
+     * uncompressed cluster.
+     * from zstd docs related to ZSTD_decompressStream:
+     * "return : 0 when a frame is completely decoded and fully flushed"
+     * We suppose that this means: each time ZSTD_decompressStream reads
+     * only ONE full frame and return 0 if and only if that frame

returns

block/qcow2-threads.c static ssize_t qcow2_zlib_decompress(void *dest, size_t dest_size, · line 279

Quoted 9 lines
block/qcow2-threads.c +9 -0
+     * is completely decoded and flushed. Only after returning 0,
+     * ZSTD_decompressStream reads another ONE full frame.
+     */
+    while (output.pos < output.size) {
+        size_t last_in_pos = input.pos;
+        size_t last_out_pos = output.pos;
+        ret = ZSTD_decompressStream(dctx, &output, &input);
+        /*
+         * zstd manual doesn't explicitly states what happens,

state

block/qcow2-threads.c static ssize_t qcow2_zlib_decompress(void *dest, size_t dest_size, · line 291

Quoted 12 lines
block/qcow2-threads.c +12 -0
+         * if ZSTD_decompressStream reads the frame partially.
+         * But, based on our tests, if we don't fully populate
+         * the output and have read all the frames from the input,
+         * we end up with error here.
+         */
+        if (ZSTD_isError(ret)) {
+            ret = -EIO;
+            break;
+        }
+
+        /*
+         * As ZSTD manual is vague about what to do if it

s/As/The/

block/qcow2-threads.c static ssize_t qcow2_zlib_decompress(void *dest, size_t dest_size, · line 294

Quoted 3 lines
block/qcow2-threads.c +3 -0
+         * reads the buffer partially, we afraid of case
+         * when we stuck in the infinite loop, because
+         * output isn't full and input has read partially, so

partially, and we don't want to get stuck in an infinite loop where ZSTD_decompressStream

Quoted 14 lines
block/qcow2-threads.c +14 -0
+         * ZSTD_decompressStream returns > 0 waiting for
+         * another input chunk. So, we add a paranoid check
+         * that on each step the loop makes some progress.
+         */
+        if (last_in_pos >= input.pos &&
+            last_out_pos >= output.pos) {
+            ret = -EIO;
+            break;
+        }
+    }
+    /*
+     * Make sure that we have the frame fully flushed here
+     * if not, we somehow managed to get uncompressed cluster
+     * greater then the cluster size, possibly because of its

block/qcow2-threads.c static ssize_t qcow2_zlib_decompress(void *dest, size_t dest_size, · line 318

Quoted 10 lines
block/qcow2-threads.c +10 -0
+     * damage.
+     */
+    if (ret > 0) {
+        ret = -EIO;
+    }
+
+    ZSTD_freeDCtx(dctx);
+    assert(ret == 0 || ret == -EIO);
+    return ret;
+}

Again, I didn't closely check the ZSTD documentation or implementation, but based on function names, the loop looks reasonable.

Signature · 3 lines
Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3226
Virtualization:  qemu.org | libvirt.org
#6 lore raw

Re: [PATCH v14 3/4] qcow2: add zstd cluster compression

In reply to Denis Plotnikov
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: 2.2G

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:

---

[..]

Quoted 12 lines
block/qcow2-threads.c +12 -0
+static ssize_t qcow2_zstd_compress(void *dest, size_t dest_size,
+                                   const void *src, size_t src_size)
+{
+    ssize_t ret;
+    size_t zstd_ret = 0;
+    ZSTD_outBuffer output = { dest, dest_size, 0 };
+    ZSTD_inBuffer input = { src, src_size, 0 };
+    ZSTD_CCtx *cctx = ZSTD_createCCtx();
+
+    if (!cctx) {
+        return -EIO;
+    }

block/qcow2-threads.c static ssize_t qcow2_zlib_decompress(void *dest, size_t dest_size, · line 209

Quoted 10 lines
block/qcow2-threads.c +10 -0
+    /*
+     * We want to use zstd streamed interface on decompression,
+     * as we won't know the exact size of the compressed data.
+     *
+     * In the loop, we try to compress all the data into one zstd frame.
+     * ZSTD_compressStream2 potentially can finish a frame earlier
+     * than the full input data is consumed. That's why we are looping
+     * until all the input data is consumed.
+     */
+    while (input.pos < input.size) {

zstd_ret may be defined here.

block/qcow2-threads.c static ssize_t qcow2_zlib_decompress(void *dest, size_t dest_size, · line 217

Quoted 8 lines
block/qcow2-threads.c +8 -0
+        /*
+         * 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". We assume that "start a new frame"
+         * means call ZSTD_compressStream2 in the very beginning or when
+         * ZSTD_compressStream2 has returned with 0.
+         */
+        do {

Hmm. Why did you decide to use nested loop? Ok, it works too.

Quoted 27 lines
block/qcow2-threads.c +27 -0
+            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 (zstd_ret > output.size - output.pos) {
+                ret = -ENOMEM;
+                goto out;
+            }
+        } while (zstd_ret);
+    }
+    /* make sure we can safely return compressed buffer size with ssize_t */
+    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
+ *

block/qcow2-threads.c static ssize_t qcow2_zlib_decompress(void *dest, size_t dest_size, · line 254

Quoted 10 lines
block/qcow2-threads.c +10 -0
+ * @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;

You forget to fix ret here. ret of the function should be ssize_t and for ZSTD - size_t.

Quoted 64 lines
block/qcow2-threads.c +64 -0
+    ZSTD_outBuffer output = { dest, dest_size, 0 };
+    ZSTD_inBuffer input = { src, src_size, 0 };
+    ZSTD_DCtx *dctx = ZSTD_createDCtx();
+
+    if (!dctx) {
+        return -EIO;
+    }
+
+    /*
+     * The compressed stream from input buffer may consist from more
+     * than one zstd frames. So we iterate until we get a fully
+     * uncompressed cluster.
+     * from zstd docs related to ZSTD_decompressStream:
+     * "return : 0 when a frame is completely decoded and fully flushed"
+     * We suppose that this means: each time ZSTD_decompressStream reads
+     * only ONE full frame and return 0 if and only if that frame
+     * is completely decoded and flushed. Only after returning 0,
+     * ZSTD_decompressStream reads another ONE full frame.
+     */
+    while (output.pos < output.size) {
+        size_t last_in_pos = input.pos;
+        size_t last_out_pos = output.pos;
+        ret = ZSTD_decompressStream(dctx, &output, &input);
+        /*
+         * zstd manual doesn't explicitly states what happens,
+         * if ZSTD_decompressStream reads the frame partially.
+         * But, based on our tests, if we don't fully populate
+         * the output and have read all the frames from the input,
+         * we end up with error here.
+         */
+        if (ZSTD_isError(ret)) {
+            ret = -EIO;
+            break;
+        }
+
+        /*
+         * As ZSTD manual is vague about what to do if it
+         * reads the buffer partially, we afraid of case
+         * when we stuck in the infinite loop, because
+         * output isn't full and input has read partially, so
+         * ZSTD_decompressStream returns > 0 waiting for
+         * another input chunk. So, we add a paranoid check
+         * that on each step the loop makes some progress.
+         */
+        if (last_in_pos >= input.pos &&
+            last_out_pos >= output.pos) {
+            ret = -EIO;
+            break;
+        }
+    }
+    /*
+     * Make sure that we have the frame fully flushed here
+     * if not, we somehow managed to get uncompressed cluster
+     * greater then the cluster size, possibly because of its
+     * damage.
+     */
+    if (ret > 0) {
+        ret = -EIO;
+    }
+
+    ZSTD_freeDCtx(dctx);
+    assert(ret == 0 || ret == -EIO);
+    return ret;
+}

[..]

Signature · 2 lines
Best regards,
Vladimir
#7 lore raw

Re: [PATCH v14 3/4] qcow2: add zstd cluster compression

Quoted 58 lines · Vladimir Sementsov-Ogievskiy
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: 2.2G

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:

---

[..]

Quoted 17 lines
block/qcow2-threads.c +17 -0
+static ssize_t qcow2_zstd_compress(void *dest, size_t dest_size,
+                                   const void *src, size_t src_size)
+{
+    ssize_t ret;
+    size_t zstd_ret = 0;
+    ZSTD_outBuffer output = { dest, dest_size, 0 };
+    ZSTD_inBuffer input = { src, src_size, 0 };
+    ZSTD_CCtx *cctx = ZSTD_createCCtx();
+
+    if (!cctx) {
+        return -EIO;
+    }
+    /*
+     * We want to use zstd streamed interface on decompression,
+     * as we won't know the exact size of the compressed data.
+     *
+     * In the loop, we try to compress all the data into one zstd 
Quoted 8 lines
Quoted 6 lines

frame.
+     * ZSTD_compressStream2 potentially can finish a frame earlier +     * than the full input data is consumed. That's why we are looping +     * until all the input data is consumed.
+     */
+    while (input.pos < input.size) {

zstd_ret may be defined here.

yep!

Quoted 12 lines
Quoted 11 lines

+        /*
+         * 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". We assume that "start a new frame"
+         * means call ZSTD_compressStream2 in the very beginning or when
+         * ZSTD_compressStream2 has returned with 0.
+         */
+        do {

Quoted 1 line

Hmm. Why did you decide to use nested loop?  Ok, it works too.

The previous condition was error prone. It led to loop ending once we finished the "first frame" but there could be a number of them.

Quoted 37 lines
Quoted 37 lines

+            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 (zstd_ret > output.size - output.pos) { +                ret = -ENOMEM;
+                goto out;
+            }
+        } while (zstd_ret);
+    }
+    /* make sure we can safely return compressed buffer size with ssize_t */
+    assert(output.pos <= SSIZE_MAX);
+    ret = output.pos;

block/qcow2-threads.c +9 -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
+ *          -EIO on any error

Quoted 8 lines

block/qcow2-threads.c static ssize_t qcow2_zlib_decompress(void *dest, size_t dest_size, · line 254

Quoted 5 lines

+ */

block/qcow2-threads.c +4 -0
+static ssize_t qcow2_zstd_decompress(void *dest, size_t dest_size,
+                                     const void *src, size_t src_size)
+{
+    size_t ret = 0;

You forget to fix ret here. ret of the function should be ssize_t and for ZSTD - size_t.

I decided not to use two vars since we can return only 0 or -EIO. I added an assert at the very end of the function to check it.

Quoted 72 lines
Quoted 67 lines

+    ZSTD_outBuffer output = { dest, dest_size, 0 }; +    ZSTD_inBuffer input = { src, src_size, 0 }; +    ZSTD_DCtx *dctx = ZSTD_createDCtx();

block/qcow2-threads.c +11 -0
+
+    if (!dctx) {
+        return -EIO;
+    }
+
+    /*
+     * The compressed stream from input buffer may consist from more
+     * than one zstd frames. So we iterate until we get a fully
+     * uncompressed cluster.
+     * from zstd docs related to ZSTD_decompressStream:
+     * "return : 0 when a frame is completely decoded and fully 

flushed"
+     * We suppose that this means: each time ZSTD_decompressStream reads
+     * only ONE full frame and return 0 if and only if that frame +     * is completely decoded and flushed. Only after returning 0, +     * ZSTD_decompressStream reads another ONE full frame. +     */
+    while (output.pos < output.size) {
+        size_t last_in_pos = input.pos;
+        size_t last_out_pos = output.pos;
+        ret = ZSTD_decompressStream(dctx, &output, &input); +        /*
+         * zstd manual doesn't explicitly states what happens, +         * if ZSTD_decompressStream reads the frame partially. +         * But, based on our tests, if we don't fully populate +         * the output and have read all the frames from the input, +         * we end up with error here.
+         */
+        if (ZSTD_isError(ret)) {
+            ret = -EIO;
+            break;
+        }

block/qcow2-threads.c +30 -0
+
+        /*
+         * As ZSTD manual is vague about what to do if it
+         * reads the buffer partially, we afraid of case
+         * when we stuck in the infinite loop, because
+         * output isn't full and input has read partially, so
+         * ZSTD_decompressStream returns > 0 waiting for
+         * another input chunk. So, we add a paranoid check
+         * that on each step the loop makes some progress.
+         */
+        if (last_in_pos >= input.pos &&
+            last_out_pos >= output.pos) {
+            ret = -EIO;
+            break;
+        }
+    }
+    /*
+     * Make sure that we have the frame fully flushed here
+     * if not, we somehow managed to get uncompressed cluster
+     * greater then the cluster size, possibly because of its
+     * damage.
+     */
+    if (ret > 0) {
+        ret = -EIO;
+    }
+
+    ZSTD_freeDCtx(dctx);
+    assert(ret == 0 || ret == -EIO);
+    return ret;
+}

[..]

#8 lore raw

Re: [PATCH v14 3/4] qcow2: add zstd cluster compression

Quoted 123 lines · Denis Plotnikov
Quoted 57 lines · Vladimir Sementsov-Ogievskiy
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: 2.2G

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:

---

[..]

Quoted 16 lines
block/qcow2-threads.c +16 -0
+static ssize_t qcow2_zstd_compress(void *dest, size_t dest_size,
+                                   const void *src, size_t src_size)
+{
+    ssize_t ret;
+    size_t zstd_ret = 0;
+    ZSTD_outBuffer output = { dest, dest_size, 0 };
+    ZSTD_inBuffer input = { src, src_size, 0 };
+    ZSTD_CCtx *cctx = ZSTD_createCCtx();
+
+    if (!cctx) {
+        return -EIO;
+    }
+    /*
+     * We want to use zstd streamed interface on decompression,
+     * as we won't know the exact size of the compressed data.
+     *
Quoted 8 lines
Quoted 6 lines

+     * In the loop, we try to compress all the data into one zstd frame. +     * ZSTD_compressStream2 potentially can finish a frame earlier +     * than the full input data is consumed. That's why we are looping +     * until all the input data is consumed.
+     */
+    while (input.pos < input.size) {

zstd_ret may be defined here.

yep!

Quoted 11 lines
Quoted 9 lines

+        /*
+         * 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". We assume that "start a new frame" +         * means call ZSTD_compressStream2 in the very beginning or when +         * ZSTD_compressStream2 has returned with 0.
+         */
+        do {

Hmm. Why did you decide to use nested loop?  Ok, it works too.

The previous condition was error prone. It led to loop ending once we finished the "first frame" but there could be a number of them.

Quoted 33 lines
Quoted 33 lines

+            zstd_ret = ZSTD_compressStream2(cctx, &output, &input, ZSTD_e_end);

block/qcow2-threads.c +31 -0
+
+            if (ZSTD_isError(zstd_ret)) {
+                ret = -EIO;
+                goto out;
+            }
+            /* Dest buffer isn't big enough to store compressed content */
+            if (zstd_ret > output.size - output.pos) {
+                ret = -ENOMEM;
+                goto out;
+            }
+        } while (zstd_ret);
+    }
+    /* make sure we can safely return compressed buffer size with ssize_t */
+    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
Quoted 8 lines

block/qcow2-threads.c static ssize_t qcow2_zlib_decompress(void *dest, size_t dest_size, · line 254

Quoted 5 lines

+ */

block/qcow2-threads.c +4 -0
+static ssize_t qcow2_zstd_decompress(void *dest, size_t dest_size,
+                                     const void *src, size_t src_size)
+{
+    size_t ret = 0;

You forget to fix ret here. ret of the function should be ssize_t and for ZSTD - size_t.

Quoted 1 line

I decided not to use two vars since we can return only 0 or -EIO. I added an assert at the very end of the function to check it.

But size_t is unsigned type. Bad idea to assign -EIO to it. And that's why we've moved to zstd_ret in _compress

Quoted 72 lines
Quoted 71 lines
Quoted 66 lines

+    ZSTD_outBuffer output = { dest, dest_size, 0 }; +    ZSTD_inBuffer input = { src, src_size, 0 }; +    ZSTD_DCtx *dctx = ZSTD_createDCtx();

block/qcow2-threads.c +61 -0
+
+    if (!dctx) {
+        return -EIO;
+    }
+
+    /*
+     * The compressed stream from input buffer may consist from more
+     * than one zstd frames. So we iterate until we get a fully
+     * uncompressed cluster.
+     * from zstd docs related to ZSTD_decompressStream:
+     * "return : 0 when a frame is completely decoded and fully flushed"
+     * We suppose that this means: each time ZSTD_decompressStream reads
+     * only ONE full frame and return 0 if and only if that frame
+     * is completely decoded and flushed. Only after returning 0,
+     * ZSTD_decompressStream reads another ONE full frame.
+     */
+    while (output.pos < output.size) {
+        size_t last_in_pos = input.pos;
+        size_t last_out_pos = output.pos;
+        ret = ZSTD_decompressStream(dctx, &output, &input);
+        /*
+         * zstd manual doesn't explicitly states what happens,
+         * if ZSTD_decompressStream reads the frame partially.
+         * But, based on our tests, if we don't fully populate
+         * the output and have read all the frames from the input,
+         * we end up with error here.
+         */
+        if (ZSTD_isError(ret)) {
+            ret = -EIO;
+            break;
+        }
+
+        /*
+         * As ZSTD manual is vague about what to do if it
+         * reads the buffer partially, we afraid of case
+         * when we stuck in the infinite loop, because
+         * output isn't full and input has read partially, so
+         * ZSTD_decompressStream returns > 0 waiting for
+         * another input chunk. So, we add a paranoid check
+         * that on each step the loop makes some progress.
+         */
+        if (last_in_pos >= input.pos &&
+            last_out_pos >= output.pos) {
+            ret = -EIO;
+            break;
+        }
+    }
+    /*
+     * Make sure that we have the frame fully flushed here
+     * if not, we somehow managed to get uncompressed cluster
+     * greater then the cluster size, possibly because of its
+     * damage.
+     */
+    if (ret > 0) {
+        ret = -EIO;
+    }
+
+    ZSTD_freeDCtx(dctx);
+    assert(ret == 0 || ret == -EIO);
+    return ret;
+}

[..]

Signature · 2 lines
Best regards,
Vladimir
#9 lore raw

PATCH v14 4/4 iotests: 287: add qcow2 compression type test

In reply to Denis Plotnikov

The 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
new file mode 100755
index 0000000000..49d15b3d43
@@ -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
index 79c6dfc85d..dacbcfc12d 100644
@@ -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
#10 lore raw

Re: [PATCH v14 4/4] iotests: 287: add qcow2 compression type test

Quoted 14 lines · Denis Plotnikov

The 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 line 53

Quoted 8 lines
tests/qemu-iotests/287 new file +8 -0
+
+# 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

This is a bash script, so (()) works; but the bash manual documents that (()) is not idiomatic. Also, what you have is rather verbose...

Review

Quoted 2 lines
(quoted diff) +2 -0
+    _notrun "ZSTD is disabled"
+fi

...I might have written:

if IMGOPTS='compression_type=zstd' _make_test_img 64M |
grep "Invalid parameter 'zstd'"; then _notrun "ZSTD is disabled"
fi

Quoted 1 line

+# Test: an image can't be openned if compression type is zlib and

opened

tests/qemu-iotests/287 line 85

Quoted 12 lines
tests/qemu-iotests/287 new file +12 -0
+#       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"

twice more

tests/qemu-iotests/287 line 88

Quoted 3 lines
tests/qemu-iotests/287 new file +3 -0
+fi
+
+# Test: an image can't be openned if compression type is NOT zlib and

and again. Multiple spots in the file, I'll quit pointing them out.

tests/qemu-iotests/287 line 91

Quoted 3 lines
tests/qemu-iotests/287 new file +3 -0
+#       incompatible feature compression type is UNSET
+echo
+echo "=== Testing zstd with incompatible bit unset  ==="

Why the double space?

tests/qemu-iotests/287 line 120

Quoted 7 lines
tests/qemu-iotests/287 new file +7 -0
+# 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

That's only one cluster. Wouldn't it be better to write more than one, to prove that we handle back-to-back compressed clusters resulting from back-to-back inputs?

Quoted 1 line

+$QEMU_IO -c "read -P 0xAC 65536 65536 " "$TEST_IMG" | _filter_qemu_io

And if you do compress more than one cluster, you may also want to use different patterns over those various clusters.

tests/qemu-iotests/287 line 126

Quoted 5 lines

+$QEMU_IO -c "read -v 131070 8 " "$TEST_IMG" | _filter_qemu_io +$QEMU_IO -c "read -v 65534 8" "$TEST_IMG" | _filter_qemu_io

tests/qemu-iotests/287 new file +3 -0
+
+# success, all done
+echo "*** done"

Is it also worth trying to generate an image with (pseudo-)random contents, and trying qemu-img convert to prove that uncompressable clusters are handled sanely? Ideally, there would be a way to use a fixed PRNG and seed that produces a deterministic sequence that cannot be compressed, but even if we can't meet the ideal, having a test that non-deterministically is likely to generate an uncompressable cluster in most runs is better than nothing (such as 1M of data copied from /dev/urandom, then qemu-img convert on that data).

Signature · 3 lines
Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3226
Virtualization:  qemu.org | libvirt.org
#11 lore raw

Re: [PATCH v14 4/4] iotests: 287: add qcow2 compression type test

Quoted 1 line · Eric Blake

Is it also worth trying to generate an image with (pseudo-)random contents, and trying qemu-img convert to prove that uncompressable clusters are handled sanely?  Ideally, there would be a way to use a fixed PRNG and seed that produces a deterministic sequence that cannot be compressed, but even if we can't meet the ideal, having a test that non-deterministically is likely to generate an uncompressable cluster in most runs is better than nothing (such as 1M of data copied from /dev/urandom, then qemu-img convert on that data).

For an example,

nbdkit -U - random 1M seed=1234 --run 'qemu-img convert $nbd file'

produces 'file' containing deterministic pseudo-random data. And 'zstd file' proves that the data was non-compressible in bulk:

$ zstd file
file                  :100.00%   (1048576 => 1048613 bytes, file.zst)

In fact, inspecting od output, I can see that the first few bytes of the original file are replayed verbatim after a 12-byte header in the .zst file.

But we don't have any other uses of nbdkit in our iotests, so I'm not sure if making this the first test to depend on that is ideal.

Signature · 3 lines
Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3226
Virtualization:  qemu.org | libvirt.org
#12 lore raw

Re: [PATCH v14 4/4] iotests: 287: add qcow2 compression type test

In reply to Eric Blake
Quoted 30 lines · Eric Blake
Quoted 19 lines · Denis Plotnikov

The 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 +4 -0
+
+# Check if we can run this test.
+
+IMGOPTS='compression_type=zstd' _make_test_img 64M | grep "Invalid 

parameter 'zstd'" 2>&1 1>/dev/null

tests/qemu-iotests/287 line 53

Quoted 4 lines
tests/qemu-iotests/287 new file +4 -0
+
+ZSTD_SUPPORTED=$?
+
+if (($ZSTD_SUPPORTED==0)); then

This is a bash script, so (()) works; but the bash manual documents that (()) is not idiomatic.  Also, what you have is rather verbose...

Quoted 2 lines

+    _notrun "ZSTD is disabled"
+fi

Quoted 6 lines

...I might have written:

if IMGOPTS='compression_type=zstd' _make_test_img 64M | grep "Invalid parameter 'zstd'"; then _notrun "ZSTD is disabled"
fi

yes, this on is shorter

Quoted 34 lines
Quoted 3 lines

+# Test: an image can't be openned if compression type is zlib and

opened

Quoted 12 lines
tests/qemu-iotests/287 new file +12 -0
+#       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 
Quoted 1 line

be openned"

twice more

tests/qemu-iotests/287 line 88

Quoted 3 lines
tests/qemu-iotests/287 new file +3 -0
+fi
+
+# Test: an image can't be openned if compression type is NOT zlib and

and again.  Multiple spots in the file, I'll quit pointing them out.

tests/qemu-iotests/287 line 91

Quoted 3 lines
tests/qemu-iotests/287 new file +3 -0
+#       incompatible feature compression type is UNSET
+echo
+echo "=== Testing zstd with incompatible bit unset  ==="

Why the double space?

Quoted 9 lines

tests/qemu-iotests/287 line 120

Quoted 7 lines
tests/qemu-iotests/287 new file +7 -0
+# 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

That's only one cluster.  Wouldn't it be better to write more than one, to prove that we handle back-to-back compressed clusters resulting from back-to-back inputs?

I'll add a case for back-to-back compressed clusters

Quoted 6 lines
Quoted 3 lines

+$QEMU_IO -c "read -P 0xAC 65536 65536 " "$TEST_IMG" | _filter_qemu_io

And if you do compress more than one cluster, you may also want to use different patterns over those various clusters.

Quoted 7 lines

tests/qemu-iotests/287 line 126

Quoted 5 lines

+$QEMU_IO -c "read -v 131070 8 " "$TEST_IMG" | _filter_qemu_io +$QEMU_IO -c "read -v 65534 8" "$TEST_IMG" | _filter_qemu_io

tests/qemu-iotests/287 new file +3 -0
+
+# success, all done
+echo "*** done"

Is it also worth trying to generate an image with (pseudo-)random contents, and trying qemu-img convert to prove that uncompressable clusters are handled sanely?  Ideally, there would be a way to use a fixed PRNG and seed that produces a deterministic sequence that cannot be compressed, but even if we can't meet the ideal, having a test that non-deterministically is likely to generate an uncompressable cluster in most runs is better than nothing (such as 1M of data copied from /dev/urandom, then qemu-img convert on that data).

And, I'll try to add a case for incompressible data.

Thanks