qemu-img: Add convert --bitmaps
v6 cover qemu-img: Add convert --bitmaps
v5 was here:
https://lists.gnu.org/archive/html/qemu-devel/2020-05/msg05958.html
original cover letter here:
https://lists.gnu.org/archive/html/qemu-devel/2020-04/msg03464.html
Hopefully this version is close enough to do a pull request Monday.
Since then:
- patch 2: drop 'measure --bitmaps' sugar on top of QMP [Nir]
- patch 3: add function comment
- patch 4: fallout from patch 2
001/5:[----] [--] 'iotests: Fix test 178'
002/5:[0073] [FC] 'qcow2: Expose bitmaps' size during measure' 003/5:[0001] [FC] 'qemu-img: Factor out code for merging bitmaps' 004/5:[0001] [FC] 'qemu-img: Add convert --bitmaps option' 005/5:[----] [--] 'iotests: Add test 291 to for qemu-img bitmap coverage'
Series can also be downloaded at:
https://repo.or.cz/qemu/ericb.git/shortlog/refs/tags/qemu-img-bitmaps-v6
Eric Blake (5):
iotests: Fix test 178
qcow2: Expose bitmaps' size during measure
qemu-img: Factor out code for merging bitmaps
qemu-img: Add convert --bitmaps option
iotests: Add test 291 to for qemu-img bitmap coverage
docs/tools/qemu-img.rst | 13 +++- qapi/block-core.json | 16 +++-- block/qcow2.h | 2 + block/crypto.c | 2 +- block/qcow2-bitmap.c | 36 ++++++++++ block/qcow2.c | 14 +++- block/raw-format.c | 2 +- qemu-img.c | 115 ++++++++++++++++++++++++++----- qemu-img-cmds.hx | 4 +- tests/qemu-iotests/178.out.qcow2 | 18 ++++- tests/qemu-iotests/178.out.raw | 2 +- tests/qemu-iotests/190 | 45 +++++++++++- tests/qemu-iotests/190.out | 27 +++++++- tests/qemu-iotests/291 | 112 ++++++++++++++++++++++++++++++ tests/qemu-iotests/291.out | 80 +++++++++++++++++++++ tests/qemu-iotests/group | 1 + 16 files changed, 455 insertions(+), 34 deletions(-) create mode 100755 tests/qemu-iotests/291 create mode 100644 tests/qemu-iotests/291.out
Signature · 1 line
2.26.2
PATCH v6 1/5 iotests: Fix test 178
A recent change to qemu-img changed expected error message output, but 178 takes long enough to execute that it does not get run by 'make check' or './check -g quick'.
--- tests/qemu-iotests/178.out.qcow2 | 2 +- tests/qemu-iotests/178.out.raw | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)
tests/qemu-iotests/178.out.qcow2 +1 -1
@@ -13,7 +13,7 @@ qemu-img: Invalid option list: , qemu-img: Invalid parameter 'snapshot.foo' qemu-img: Failed in parsing snapshot param 'snapshot.foo' qemu-img: --output must be used with human or json as argument. -qemu-img: Image size must be less than 8 EiB! +qemu-img: Invalid image size specified. Must be between 0 and 9223372036854775807. qemu-img: Unknown file format 'foo' == Size calculation for a new file (human) ==
tests/qemu-iotests/178.out.raw +1 -1
@@ -13,7 +13,7 @@ qemu-img: Invalid option list: , qemu-img: Invalid parameter 'snapshot.foo' qemu-img: Failed in parsing snapshot param 'snapshot.foo' qemu-img: --output must be used with human or json as argument. -qemu-img: Image size must be less than 8 EiB! +qemu-img: Invalid image size specified. Must be between 0 and 9223372036854775807. qemu-img: Unknown file format 'foo' == Size calculation for a new file (human) ==
Signature · 1 line
2.26.2
PATCH v6 2/5 qcow2: Expose bitmaps' size during measure
In reply to Eric BlakeIt's useful to know how much space can be occupied by qcow2 persistent bitmaps, even though such metadata is unrelated to the guest-visible data. Report this value as an additional QMP field, present when measuring an existing image and output format that both support bitmaps. Update iotest 178 and 190 to updated output, as well as new coverage in 190 demonstrating non-zero values made possible with the recently-added qemu-img bitmap command (see 3b51ab4b).
The new 'bitmaps size:' field is displayed automatically as part of 'qemu-img measure' any time it is present in QMP (that is, any time both the source image being measured and destination format support bitmaps, even if the measurement is 0 because there are no bitmaps present). If the field is absent, it means that no bitmaps can be copied (source, destination, or both lack bitmaps, including when measuring based on size rather than on a source image). This behavior is compatible with an upcoming patch adding 'qemu-img convert --bitmaps': that command will fail in the same situations where this patch omits the field.
The addition of a new field demonstrates why we should always zero-initialize qapi C structs; while the qcow2 driver still fully populates all fields, the raw and crypto drivers had to be tweaked to avoid uninitialized data.
See also: https://bugzilla.redhat.com/1779904
--- docs/tools/qemu-img.rst | 7 +++++ qapi/block-core.json | 16 ++++++++---- block/qcow2.h | 2 ++ block/crypto.c | 2 +- block/qcow2-bitmap.c | 36 +++++++++++++++++++++++++ block/qcow2.c | 14 +++++++--- block/raw-format.c | 2 +- qemu-img.c | 3 +++ tests/qemu-iotests/178.out.qcow2 | 16 ++++++++++++ tests/qemu-iotests/190 | 45 ++++++++++++++++++++++++++++++-- tests/qemu-iotests/190.out | 27 ++++++++++++++++++- 11 files changed, 157 insertions(+), 13 deletions(-)
docs/tools/qemu-img.rst +7 -0
@@ -616,6 +616,7 @@ Command description: required size: 524288 fully allocated size: 1074069504 + bitmaps size: 0 The ``required size`` is the file size of the new image. It may be smaller than the virtual disk size if the image format supports compact representation.
@@ -625,6 +626,12 @@ Command description: occupy with the exception of internal snapshots, dirty bitmaps, vmstate data, and other advanced image format features. + The ``bitmaps size`` is the additional size required in order to + copy bitmaps from a source image in addition to the guest-visible + data; the line is omitted if either source or destination lacks + bitmap support, or 0 if bitmaps are supported but there is nothing + to copy. + .. option:: snapshot [--object OBJECTDEF] [--image-opts] [-U] [-q] [-l | -a SNAPSHOT | -c SNAPSHOT | -d SNAPSHOT] FILENAME List, apply, create or delete snapshots in image *FILENAME*.
qapi/block-core.json +11 -5
@@ -636,18 +636,24 @@ # efficiently so file size may be smaller than virtual disk size. # # The values are upper bounds that are guaranteed to fit the new image file. -# Subsequent modification, such as internal snapshot or bitmap creation, may -# require additional space and is not covered here. +# Subsequent modification, such as internal snapshot or further bitmap +# creation, may require additional space and is not covered here. # -# @required: Size required for a new image file, in bytes. +# @required: Size required for a new image file, in bytes, when copying just +# allocated guest-visible contents. # # @fully-allocated: Image file size, in bytes, once data has been written -# to all sectors. +# to all sectors, when copying just guest-visible contents. +# +# @bitmaps: Additional size required if all the top-level bitmap metadata +# in the source image were to be copied to the destination, +# present only when source and destination both support +# persistent bitmaps. (since 5.1) # # Since: 2.10 ## { 'struct': 'BlockMeasureInfo', - 'data': {'required': 'int', 'fully-allocated': 'int'} } + 'data': {'required': 'int', 'fully-allocated': 'int', '*bitmaps': 'int'} } ## # @query-block:
block/qcow2.h +2 -0
@@ -783,6 +783,8 @@ int qcow2_co_remove_persistent_dirty_bitmap(BlockDriverState *bs,const char *name, Error **errp);bool qcow2_supports_persistent_dirty_bitmap (BlockDriverState *bs); +uint64_t qcow2_get_persistent_dirty_bitmap_size (BlockDriverState *bs, +uint32_t cluster_size); ssize_t coroutine_fn qcow2_co_compress(BlockDriverState *bs, void *dest, size_t dest_size,
block/crypto.c +1 -1
@@ -552,7 +552,7 @@ static BlockMeasureInfo *block_crypto_measure(QemuOpts *opts,* Unallocated blocks are still encrypted so allocation status makes no * difference to the file size. */ - info =g_new (BlockMeasureInfo,1 ); + info =g_new0 (BlockMeasureInfo,1 ); info->fully_allocated = luks_payload_size + size; info->required = luks_payload_size + size;return info;
block/qcow2-bitmap.c +36 -0
@@ -1755,3 +1755,39 @@ bool qcow2_supports_persistent_dirty_bitmap(BlockDriverState *bs)return s->qcow_version >=3 ; } + +/* +* Compute the space required for bitmaps in @bs. +* +* The computation is based as if copying to a new image with the +* given @cluster_size, which may differ from the cluster size in @bs. +*/ +uint64_t qcow2_get_persistent_dirty_bitmap_size (BlockDriverState *bs, +uint32_t cluster_size) +{ +uint64_t bitmaps_size =0 ; +BdrvDirtyBitmap *bm; +size_t bitmap_dir_size =0 ; + + FOR_EACH_DIRTY_BITMAP(bs, bm) { +if (bdrv_dirty_bitmap_get_persistence (bm)) { +const char *name =bdrv_dirty_bitmap_name (bm); +uint32_t granularity =bdrv_dirty_bitmap_granularity (bm); +uint64_t bmbytes = +get_bitmap_bytes_needed (bdrv_dirty_bitmap_size (bm), + granularity); +uint64_t bmclusters =DIV_ROUND_UP (bmbytes, cluster_size); + +/* Assume the entire bitmap is allocated */ + bitmaps_size += bmclusters * cluster_size; +/* Also reserve space for the bitmap table entries */ + bitmaps_size +=ROUND_UP (bmclusters *sizeof (uint64_t ), + cluster_size); +/* And space for contribution to bitmap directory size */ + bitmap_dir_size +=calc_dir_entry_size (strlen (name),0 ); + } + } + bitmaps_size +=ROUND_UP (bitmap_dir_size, cluster_size); + +return bitmaps_size; +}
block/qcow2.c +11 -3
@@ -4953,16 +4953,24 @@ static BlockMeasureInfo *qcow2_measure(QemuOpts *opts, BlockDriverState *in_bs, required = virtual_size; } - info =g_new (BlockMeasureInfo,1 ); + info =g_new0 (BlockMeasureInfo,1 ); info->fully_allocated =qcow2_calc_prealloc_size (virtual_size, cluster_size,ctz32 (refcount_bits)) + luks_payload_size; -/* Remove data clusters that are not required. This overestimates the +/* +* Remove data clusters that are not required. This overestimates the * required size because metadata needed for the fully allocated file is -* still counted. +* still counted. Show bitmaps only if both source and destination +* would support them. */ info->required = info->fully_allocated - virtual_size + required; + info->has_bitmaps = version >=3 && in_bs && +bdrv_supports_persistent_dirty_bitmap (in_bs); +if (info->has_bitmaps) { + info->bitmaps =qcow2_get_persistent_dirty_bitmap_size (in_bs, + cluster_size); + }return info; err:
block/raw-format.c +1 -1
@@ -359,7 +359,7 @@ static BlockMeasureInfo *raw_measure(QemuOpts *opts, BlockDriverState *in_bs, BDRV_SECTOR_SIZE); } - info =g_new (BlockMeasureInfo,1 ); + info =g_new0 (BlockMeasureInfo,1 ); info->required = required;/* Unallocated sectors count towards the file size in raw images */
qemu-img.c +3 -0
@@ -5302,6 +5302,9 @@ static int img_measure(int argc, char **argv)if (output_format ==OFORMAT_HUMAN ) {printf ("required size: %" PRIu64"\n" , info->required);printf ("fully allocated size: %" PRIu64"\n" , info->fully_allocated); +if (info->has_bitmaps) { +printf ("bitmaps size: %" PRIu64"\n" , info->bitmaps); + } }else {dump_json_block_measure_info (info); }
tests/qemu-iotests/178.out.qcow2 +16 -0
@@ -37,6 +37,7 @@ qemu-img: The image size is too large (try using a larger cluster size) Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=0 required size: 196608 fully allocated size: 196608 +bitmaps size: 0 converted image file size in bytes: 196608
@@ -45,6 +46,7 @@ converted image file size in bytes: 196608 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 required size: 393216 fully allocated size: 1074135040 +bitmaps size: 0 wrote 512/512 bytes at offset 512 512 bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) wrote 65536/65536 bytes at offset 65536
@@ -53,6 +55,7 @@ wrote 64512/64512 bytes at offset 134217728 63 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) required size: 589824 fully allocated size: 1074135040 +bitmaps size: 0 converted image file size in bytes: 524288
@@ -60,6 +63,7 @@ converted image file size in bytes: 524288 required size: 524288 fully allocated size: 1074135040 +bitmaps size: 0 converted image file size in bytes: 458752
@@ -67,16 +71,19 @@ converted image file size in bytes: 458752 required size: 1074135040 fully allocated size: 1074135040 +bitmaps size: 0 == qcow2 input image and LUKS encryption == required size: 2686976 fully allocated size: 1076232192 +bitmaps size: 0 == qcow2 input image and preallocation (human) == required size: 1074135040 fully allocated size: 1074135040 +bitmaps size: 0 converted image file size in bytes: 1074135040
@@ -87,6 +94,7 @@ wrote 8388608/8388608 bytes at offset 0 8 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) required size: 8716288 fully allocated size: 8716288 +bitmaps size: 0 converted image file size in bytes: 8716288
@@ -173,6 +181,7 @@ qemu-img: The image size is too large (try using a larger cluster size) Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=0 { + "bitmaps": 0, "required": 196608, "fully-allocated": 196608 }
@@ -183,6 +192,7 @@ converted image file size in bytes: 196608 Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=1073741824 { + "bitmaps": 0, "required": 393216, "fully-allocated": 1074135040 }
@@ -193,6 +203,7 @@ wrote 65536/65536 bytes at offset 65536 wrote 64512/64512 bytes at offset 134217728 63 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) { + "bitmaps": 0, "required": 589824, "fully-allocated": 1074135040 }
@@ -202,6 +213,7 @@ converted image file size in bytes: 524288 == qcow2 input image with internal snapshot (json) == { + "bitmaps": 0, "required": 524288, "fully-allocated": 1074135040 }
@@ -211,6 +223,7 @@ converted image file size in bytes: 458752 == qcow2 input image and a backing file (json) == { + "bitmaps": 0, "required": 1074135040, "fully-allocated": 1074135040 }
@@ -218,6 +231,7 @@ converted image file size in bytes: 458752 == qcow2 input image and LUKS encryption == { + "bitmaps": 0, "required": 2686976, "fully-allocated": 1076232192 }
@@ -225,6 +239,7 @@ converted image file size in bytes: 458752 == qcow2 input image and preallocation (json) == { + "bitmaps": 0, "required": 1074135040, "fully-allocated": 1074135040 }
@@ -237,6 +252,7 @@ Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=8388608 wrote 8388608/8388608 bytes at offset 0 8 MiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) { + "bitmaps": 0, "required": 8716288, "fully-allocated": 8716288 }
tests/qemu-iotests/190 +43 -2
63 lines not on this page · raw message
tests/qemu-iotests/190.out +26 -1
@@ -1,11 +1,36 @@ QA output created by 190 -== Huge file == +== Huge file without bitmaps == Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=2199023255552 required size: 2199023255552 fully allocated size: 2199023255552 required size: 335806464 fully allocated size: 2199359062016 +bitmaps size: 0 required size: 18874368 fully allocated size: 2199042129920 +bitmaps size: 0 + +== Huge file with bitmaps == + +required size: 327680 +fully allocated size: 10813440 +required size: 2199023255552 +fully allocated size: 2199023255552 +required size: 7012352 +fully allocated size: 17170432 +required size: 335806464 +fully allocated size: 2199359062016 + +expected bitmap 537198592 +required size: 335806464 +fully allocated size: 2199359062016 +bitmaps size: 537198592 + +expected bitmap 545259520 +{ + "bitmaps": 545259520, + "required": 18874368, + "fully-allocated": 2199042129920 +} *** done
Signature · 1 line
2.26.2
Re: [PATCH v6 2/5] qcow2: Expose bitmaps' size during measure
Quoted 24 lines · Eric Blake
It's useful to know how much space can be occupied by qcow2 persistent bitmaps, even though such metadata is unrelated to the guest-visible data. Report this value as an additional QMP field, present when measuring an existing image and output format that both support bitmaps. Update iotest 178 and 190 to updated output, as well as new coverage in 190 demonstrating non-zero values made possible with the recently-added qemu-img bitmap command (see
3b51ab4b).The new 'bitmaps size:' field is displayed automatically as part of 'qemu-img measure' any time it is present in QMP (that is, any time both the source image being measured and destination format support bitmaps, even if the measurement is 0 because there are no bitmaps present). If the field is absent, it means that no bitmaps can be copied (source, destination, or both lack bitmaps, including when measuring based on size rather than on a source image). This behavior is compatible with an upcoming patch adding 'qemu-img convert --bitmaps': that command will fail in the same situations where this patch omits the field.
The addition of a new field demonstrates why we should always zero-initialize qapi C structs; while the qcow2 driver still fully populates all fields, the raw and crypto drivers had to be tweaked to avoid uninitialized data.
Quoted 4 lines
See also:https://bugzilla.redhat.com/1779904
Reported-by Nir Soffer<[email protected]>Signed-off-by Eric Blake<[email protected]>
Signature · 2 lines
Best regards, Vladimir
PATCH v6 3/5 qemu-img: Factor out code for merging bitmaps
In reply to Eric BlakeThe next patch will add another client that wants to merge dirty bitmaps; it will be easier to refactor the code to construct the QAPI struct correctly into a helper function.
--- qemu-img.c | 34 +++++++++++++++++++++------------- 1 file changed, 21 insertions(+), 13 deletions(-)
qemu-img.c +21 -13
@@ -1638,6 +1638,24 @@ out4:return ret; } +/* Convenience wrapper around qmp_block_dirty_bitmap_merge */ +static void do_dirty_bitmap_merge (const char *dst_node,const char *dst_name, +const char *src_node,const char *src_name, +Error **errp) +{ +BlockDirtyBitmapMergeSource *merge_src; +BlockDirtyBitmapMergeSourceList *list; + + merge_src =g_new0 (BlockDirtyBitmapMergeSource,1 ); + merge_src->type =QTYPE_QDICT ; + merge_src->u.external.node =g_strdup (src_node); + merge_src->u.external.name =g_strdup (src_name); + list =g_new0 (BlockDirtyBitmapMergeSourceList,1 ); + list->value = merge_src; +qmp_block_dirty_bitmap_merge (dst_node, dst_name, list, errp); +qapi_free_BlockDirtyBitmapMergeSourceList (list); +} +enum ImgConvertBlockStatus {BLK_DATA ,BLK_ZERO ,
@@ -4714,21 +4732,11 @@ static int img_bitmap(int argc, char **argv)qmp_block_dirty_bitmap_disable (bs->node_name, bitmap, &err); op ="disable" ;break ; -case BITMAP_MERGE : { -BlockDirtyBitmapMergeSource *merge_src; -BlockDirtyBitmapMergeSourceList *list; - - merge_src =g_new0 (BlockDirtyBitmapMergeSource,1 ); - merge_src->type =QTYPE_QDICT ; - merge_src->u.external.node =g_strdup (src_bs->node_name); - merge_src->u.external.name =g_strdup (act->src); - list =g_new0 (BlockDirtyBitmapMergeSourceList,1 ); - list->value = merge_src; -qmp_block_dirty_bitmap_merge (bs->node_name, bitmap, list, &err); -qapi_free_BlockDirtyBitmapMergeSourceList (list); +case BITMAP_MERGE : +do_dirty_bitmap_merge (bs->node_name, bitmap, src_bs->node_name, + act->src, &err); op ="merge" ;break ; - }default :g_assert_not_reached (); }
Signature · 1 line
2.26.2
Re: [PATCH v6 3/5] qemu-img: Factor out code for merging bitmaps
Quoted 5 lines · Eric Blake
The next patch will add another client that wants to merge dirty bitmaps; it will be easier to refactor the code to construct the QAPI struct correctly into a helper function.
Signed-off-by Eric Blake<[email protected]>
Signature · 2 lines
Best regards, Vladimir
PATCH v6 4/5 qemu-img: Add convert --bitmaps option
In reply to Eric BlakeMake it easier to copy all the persistent bitmaps of (the top layer of) a source image along with its guest-visible contents, by adding a boolean flag for use with qemu-img convert. This is basically shorthand, as the same effect could be accomplished with a series of 'qemu-img bitmap --add' and 'qemu-img bitmap --merge -b source' commands, or by their corresponding QMP commands.
Note that this command will fail in the same scenarios where 'qemu-img measure' omits a 'bitmaps size:' line, namely, when either the source or the destination lacks persistent bitmap support altogether.
See also https://bugzilla.redhat.com/show_bug.cgi?id=1779893
While touching this, clean up a couple coding issues spotted in the same function: an extra blank line, and merging back-to-back 'if (!skip_create)' blocks.
--- docs/tools/qemu-img.rst | 6 +++- qemu-img.c | 78 +++++++++++++++++++++++++++++++++++++++-- qemu-img-cmds.hx | 4 +-- 3 files changed, 82 insertions(+), 6 deletions(-)
docs/tools/qemu-img.rst +5 -1
@@ -162,6 +162,10 @@ Parameters to convert subcommand: .. program:: qemu-img-convert +.. option:: --bitmaps + + Additionally copy all persistent bitmaps from the top layer of the source + .. option:: -n Skip the creation of the target volume
@@ -397,7 +401,7 @@ Command description: 4 Error on reading data -.. option:: convert [--object OBJECTDEF] [--image-opts] [--target-image-opts] [--target-is-zero] [-U] [-C] [-c] [-p] [-q] [-n] [-f FMT] [-t CACHE] [-T SRC_CACHE] [-O OUTPUT_FMT] [-B BACKING_FILE] [-o OPTIONS] [-l SNAPSHOT_PARAM] [-S SPARSE_SIZE] [-m NUM_COROUTINES] [-W] FILENAME [FILENAME2 [...]] OUTPUT_FILENAME +.. option:: convert [--object OBJECTDEF] [--image-opts] [--target-image-opts] [--target-is-zero] [--bitmaps] [-U] [-C] [-c] [-p] [-q] [-n] [-f FMT] [-t CACHE] [-T SRC_CACHE] [-O OUTPUT_FMT] [-B BACKING_FILE] [-o OPTIONS] [-l SNAPSHOT_PARAM] [-S SPARSE_SIZE] [-m NUM_COROUTINES] [-W] FILENAME [FILENAME2 [...]] OUTPUT_FILENAME Convert the disk image *FILENAME* or a snapshot *SNAPSHOT_PARAM* to disk image *OUTPUT_FILENAME* using format *OUTPUT_FMT*. It can
qemu-img.c +75 -3
145 lines not on this page · raw message
qemu-img-cmds.hx +2 -2
@@ -46,9 +46,9 @@ SRST ERST DEF("convert", img_convert, - "convert [--object objectdef] [--image-opts] [--target-image-opts] [--target-is-zero] [-U] [-C] [-c] [-p] [-q] [-n] [-f fmt] [-t cache] [-T src_cache] [-O output_fmt] [-B backing_file] [-o options] [-l snapshot_param] [-S sparse_size] [-m num_coroutines] [-W] [--salvage] filename [filename2 [...]] output_filename") + "convert [--object objectdef] [--image-opts] [--target-image-opts] [--target-is-zero] [--bitmaps] [-U] [-C] [-c] [-p] [-q] [-n] [-f fmt] [-t cache] [-T src_cache] [-O output_fmt] [-B backing_file] [-o options] [-l snapshot_param] [-S sparse_size] [-m num_coroutines] [-W] [--salvage] filename [filename2 [...]] output_filename") SRST -.. option:: convert [--object OBJECTDEF] [--image-opts] [--target-image-opts] [--target-is-zero] [-U] [-C] [-c] [-p] [-q] [-n] [-f FMT] [-t CACHE] [-T SRC_CACHE] [-O OUTPUT_FMT] [-B BACKING_FILE] [-o OPTIONS] [-l SNAPSHOT_PARAM] [-S SPARSE_SIZE] [-m NUM_COROUTINES] [-W] [--salvage] FILENAME [FILENAME2 [...]] OUTPUT_FILENAME +.. option:: convert [--object OBJECTDEF] [--image-opts] [--target-image-opts] [--target-is-zero] [--bitmaps] [-U] [-C] [-c] [-p] [-q] [-n] [-f FMT] [-t CACHE] [-T SRC_CACHE] [-O OUTPUT_FMT] [-B BACKING_FILE] [-o OPTIONS] [-l SNAPSHOT_PARAM] [-S SPARSE_SIZE] [-m NUM_COROUTINES] [-W] [--salvage] FILENAME [FILENAME2 [...]] OUTPUT_FILENAME ERST DEF("create", img_create,
Signature · 1 line
2.26.2
Re: [PATCH v6 4/5] qemu-img: Add convert --bitmaps option
Quoted 19 lines · Eric Blake
Make it easier to copy all the persistent bitmaps of (the top layer of) a source image along with its guest-visible contents, by adding a boolean flag for use with qemu-img convert. This is basically shorthand, as the same effect could be accomplished with a series of 'qemu-img bitmap --add' and 'qemu-img bitmap --merge -b source' commands, or by their corresponding QMP commands.
Note that this command will fail in the same scenarios where 'qemu-img measure' omits a 'bitmaps size:' line, namely, when either the source or the destination lacks persistent bitmap support altogether.
See also https://bugzilla.redhat.com/show_bug.cgi?id=1779893
While touching this, clean up a couple coding issues spotted in the same function: an extra blank line, and merging back-to-back 'if (!skip_create)' blocks.
Signed-off-by Eric Blake <[email protected]>---
[..]
Quoted 8 lines
qemu-img.c +0
-0@@ -78,6 +78,7 @@ enum {OPTION_ENABLE =272 ,OPTION_DISABLE =273 , OPTION_MERGE = 274,
[..]
Quoted 76 lines
static int
img_convert(int argc, char **argv)qemu-img.c +27
-3@@ -2160,6 +2195,8 @@ static int img_convert(int argc, char **argv)int64_t ret = -EINVAL ;bool force_share = false;bool explict_min_sparse = false; +bool bitmaps = false; +size_t nbitmaps =0 ;ImgConvertState s = (ImgConvertState ) {/* Need at least 4k of zeros for sparse detection */ @@ -2179,6 +2216,7 @@ static int img_convert(int argc, char **argv) {"target-image-opts", no_argument, 0, OPTION_TARGET_IMAGE_OPTS}, {"salvage", no_argument, 0, OPTION_SALVAGE}, {"target-is-zero", no_argument, 0, OPTION_TARGET_IS_ZERO}, + {"bitmaps", no_argument, 0, OPTION_BITMAPS}, {0, 0, 0, 0} }; c = getopt_long(argc, argv, ":hf:O:B:Cco:l:S:pt:T:qnm:WU",@@ -2304,6 +2342,9 @@ static int img_convert(int argc, char **argv)*/ s.has_zero_init = true;break ; +case OPTION_BITMAPS : + bitmaps = true; +break ; } }@@ -2365,7 +2406,6 @@ static int img_convert(int argc, char **argv) goto fail_getopt; } -/* ret is still -EINVAL until here */ ret =bdrv_parse_cache_mode (src_cache, &src_flags, &src_writethrough);if (ret <0 ) {@@ -2525,6 +2565,27 @@ static int img_convert(int argc, char **argv) } } +/* Determine how many bitmaps need copying */ +if (bitmaps) { +BdrvDirtyBitmap *bm; + +if (s.src_num >1 ) { +error_report ("Copying bitmaps only possible with single source" ); + ret =-1 ; + goto out; + } +if (!bdrv_supports_persistent_dirty_bitmap (blk_bs (s.src[0 ]))) { +error_report ("Source lacks bitmap support" ); + ret =-1 ; + goto out; + } + FOR_EACH_DIRTY_BITMAP(blk_bs(s.src[0]), bm) { +if (bdrv_dirty_bitmap_get_persistence (bm)) { + nbitmaps++; + } + } + } +/* * The later open call will need any decryption secrets, and * bdrv_create() will purge "opts", so extract them now before @@ -2533,9 +2594,7 @@ static int img_convert(int argc, char **argv)if (!skip_create) { open_opts =qdict_new ();qemu_opt_foreach (opts, img_add_key_secrets, open_opts, &error_abort); - } -if (!skip_create) {/* Create the new image */ ret =bdrv_create (drv, out_filename, opts, &local_err);if (ret <0 ) {
qemu-img.c static int img_convert(int argc, char **argv) · line 2635
Quoted 4 lines
qemu-img.c +1
-0@@ -2573,6 +2632,13 @@ static int img_convert(int argc, char **argv) } out_bs =blk_bs (s.target); +if (nbitmaps >0 && !bdrv_supports_persistent_dirty_bitmap (out_bs)) {
We will not fail, if target doesn't support bitmaps, source supports them but has no bitmaps? Doesn't seem to be a problem, but a bit less strict than you write in commit message.
So, maybe, s/nbitmaps > 0/bitmaps/
qemu-img.c static int img_convert(int argc, char **argv) · line 2637
Quoted 2 lines
qemu-img.c +2
-0+error_report ("Format driver '%s' does not support bitmaps" , + out_fmt);
Hmm seems, out_fmt may be NULL at this point, consider the path:
const char *out_fmt = NULL
...
[no -O option]
--target-image-opts, so out_fmt doesn't default to "raw" but remains NULL
...
So, with s/out_fmt/out_bs->drv->format_name/ (and with or without s/nbitmaps > 0/bitmaps/):
Signature · 2 lines
Best regards, Vladimir
Re: [PATCH v6 4/5] qemu-img: Add convert --bitmaps option
Quoted 26 lines · Vladimir Sementsov-Ogievskiy
Quoted 25 lines · Eric Blake
Make it easier to copy all the persistent bitmaps of (the top layer of) a source image along with its guest-visible contents, by adding a boolean flag for use with qemu-img convert. This is basically shorthand, as the same effect could be accomplished with a series of 'qemu-img bitmap --add' and 'qemu-img bitmap --merge -b source' commands, or by their corresponding QMP commands.
Note that this command will fail in the same scenarios where 'qemu-img measure' omits a 'bitmaps size:' line, namely, when either the source or the destination lacks persistent bitmap support altogether.
See also https://bugzilla.redhat.com/show_bug.cgi?id=1779893
While touching this, clean up a couple coding issues spotted in the same function: an extra blank line, and merging back-to-back 'if (!skip_create)' blocks.
Signed-off-by Eric Blake <[email protected]>---
(quoted diff) +0
-0@@ -2573,6 +2632,13 @@ static int img_convert(int argc, char **argv) } out_bs = blk_bs(s.target);
Quoted 6 lines
Review static int img_convert(int argc, char **argv) · line 2636
Quoted 2 lines
(quoted diff) +1
-0+ if (nbitmaps > 0 && !bdrv_supports_persistent_dirty_bitmap(out_bs)) {We will not fail, if target doesn't support bitmaps, source supports them but has no bitmaps? Doesn't seem to be a problem, but a bit less strict than you write in commit message.
So, maybe, s/nbitmaps > 0/bitmaps/
In fact, nbitmaps is not needed at all (it was useful in earlier iterations, but as the series has morphed, it is no longer buying me anything useful).
Quoted 11 lines
Review static int img_convert(int argc, char **argv) · line 2638
Quoted 3 lines
(quoted diff) +2
-0+ error_report("Format driver '%s' does not support bitmaps", + out_fmt);Hmm seems,
out_fmtmay be NULL at this point, consider the path:
const char*out_fmt= NULL
...
[no -O option]
--target-image-opts, soout_fmtdoesn't default to "raw" but remains NULL
...
Quoted 4 lines
So, with s/out_fmt/out_bs->drv->format_name/ (and with or without s/nbitmaps > 0/bitmaps/):
Reviewed-by Vladimir Sementsov-Ogievskiy <[email protected]>
Okay, I'm squashing this in, and adding your R-b. Pull request coming shortly.
qemu-img.c +1 -4
@@ -2196,7 +2196,6 @@ static int img_convert(int argc, char **argv)bool force_share = false;bool explict_min_sparse = false;bool bitmaps = false; -size_t nbitmaps =0 ;ImgConvertState s = (ImgConvertState ) {/* Need at least 4k of zeros for sparse detection */
@@ -2565,10 +2564,8 @@ static int img_convert(int argc, char **argv) } } -/* Determine how many bitmaps need copying */ +/* Determine if bitmaps need copying */ if (bitmaps) { -BdrvDirtyBitmap *bm; -if (s.src_num >1 ) { error_report("Copying bitmaps only possible with single
source");
ret = -1;
@@ -2579,11 +2576,6 @@ static int img_convert(int argc, char **argv)
ret = -1;
goto out;
}- FOR_EACH_DIRTY_BITMAP(blk_bs(s.src[0]), bm) {
- if (bdrv_dirty_bitmap_get_persistence(bm)) {
- nbitmaps++;
- }
- }
}
/*
@@ -2632,9 +2624,9 @@ static int img_convert(int argc, char **argv)
}
out_bs = blk_bs(s.target);- if (nbitmaps > 0 && !bdrv_supports_persistent_dirty_bitmap(out_bs)) {
+ if (bitmaps && !bdrv_supports_persistent_dirty_bitmap(out_bs)) {
error_report("Format driver '%s' does not support bitmaps",
- out_fmt);
+ out_bs->drv->format_name);
ret = -1;
goto out;
}
@@ -2700,7 +2692,7 @@ static int img_convert(int argc, char **argv)
ret = convert_do_copy(&s);
/* Now copy the bitmaps */
- if (nbitmaps > 0 && ret == 0) {
+ if (bitmaps && ret == 0) {
ret = convert_copy_bitmaps(blk_bs(s.src[0]), out_bs);
}Signature · 3 lines
Eric Blake, Principal Software Engineer Red Hat, Inc. +1-919-301-3226 Virtualization: qemu.org | libvirt.org
PATCH v6 5/5 iotests: Add test 291 to for qemu-img bitmap coverage
In reply to Eric BlakeAdd a new test covering the 'qemu-img bitmap' subcommand, as well as 'qemu-img convert --bitmaps', both added in recent patches.
--- tests/qemu-iotests/291 | 112 +++++++++++++++++++++++++++++++++++++ tests/qemu-iotests/291.out | 80 ++++++++++++++++++++++++++ tests/qemu-iotests/group | 1 + 3 files changed, 193 insertions(+) create mode 100755 tests/qemu-iotests/291 create mode 100644 tests/qemu-iotests/291.out
tests/qemu-iotests/291 new file +112 -0
@@ -0,0 +1,112 @@ +#!/usr/bin/env bash +# +# Test qemu-img bitmap handling +# +# Copyright (C) 2018-2020 Red Hat, Inc. +# +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 2 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see <http://www.gnu.org/licenses/>. +# + +seq="$(basename $0)" +echo "QA output created by $seq" + +status=1 # failure is the default! + +_cleanup() +{ + _cleanup_test_img + nbd_server_stop +} +trap "_cleanup; exit \$status" 0 1 2 3 15 + +# get standard environment, filters and checks +. ./common.rc +. ./common.filter +. ./common.nbd + +_supported_fmt qcow2 +_supported_proto file +_supported_os Linux +_require_command QEMU_NBD + +echo +echo "=== Initial image setup ===" +echo + +# Create backing image with one bitmap +TEST_IMG="$TEST_IMG.base" _make_test_img 10M +$QEMU_IMG bitmap --add -f $IMGFMT "$TEST_IMG.base" b0 +$QEMU_IO -c 'w 3M 1M' -f $IMGFMT "$TEST_IMG.base" | _filter_qemu_io + +# Create initial image and populate two bitmaps: one active, one inactive. +ORIG_IMG=$TEST_IMG +TEST_IMG=$TEST_IMG.orig +_make_test_img -b "$ORIG_IMG.base" -F $IMGFMT 10M +$QEMU_IO -c 'w 0 1M' -f $IMGFMT "$TEST_IMG" | _filter_qemu_io +$QEMU_IMG bitmap --add -g 512k -f $IMGFMT "$TEST_IMG" b1 +$QEMU_IMG bitmap --add --disable -f $IMGFMT "$TEST_IMG" b2 +$QEMU_IO -c 'w 3M 1M' -f $IMGFMT "$TEST_IMG" | _filter_qemu_io +$QEMU_IMG bitmap --clear -f $IMGFMT "$TEST_IMG" b1 +$QEMU_IO -c 'w 1M 1M' -f $IMGFMT "$TEST_IMG" | _filter_qemu_io +$QEMU_IMG bitmap --disable -f $IMGFMT "$TEST_IMG" b1 +$QEMU_IMG bitmap --enable -f $IMGFMT "$TEST_IMG" b2 +$QEMU_IO -c 'w 2M 1M' -f $IMGFMT "$TEST_IMG" | _filter_qemu_io + +echo +echo "=== Bitmap preservation not possible to non-qcow2 ===" +echo + +TEST_IMG=$ORIG_IMG +$QEMU_IMG convert --bitmaps -O raw "$TEST_IMG.orig" "$TEST_IMG" && + echo "unexpected success" + +echo +echo "=== Convert with bitmap preservation ===" +echo + +# Only bitmaps from the active layer are copied +$QEMU_IMG convert --bitmaps -O qcow2 "$TEST_IMG.orig" "$TEST_IMG" +$QEMU_IMG info "$TEST_IMG" | _filter_img_info --format-specific +# But we can also merge in bitmaps from other layers. This test is a bit +# contrived to cover more code paths, in reality, you could merge directly +# into b0 without going through tmp +$QEMU_IMG bitmap --add --disable -f $IMGFMT "$TEST_IMG" b0 +$QEMU_IMG bitmap --add --merge b0 -b "$TEST_IMG.base" -F $IMGFMT \ + -f $IMGFMT "$TEST_IMG" tmp +$QEMU_IMG bitmap --merge tmp -f $IMGFMT "$TEST_IMG" b0 +$QEMU_IMG bitmap --remove --image-opts \ + driver=$IMGFMT,file.driver=file,file.filename="$TEST_IMG" tmp +$QEMU_IMG info "$TEST_IMG" | _filter_img_info --format-specific + +echo +echo "=== Check bitmap contents ===" +echo + +# x-dirty-bitmap is a hack for reading bitmaps; it abuses block status to +# report "data":false for portions of the bitmap which are set +IMG="driver=nbd,server.type=unix,server.path=$nbd_unix_socket" +nbd_server_start_unix_socket -r -f qcow2 -B b0 "$TEST_IMG" +$QEMU_IMG map --output=json --image-opts \ + "$IMG,x-dirty-bitmap=qemu:dirty-bitmap:b0" | _filter_qemu_img_map +nbd_server_start_unix_socket -r -f qcow2 -B b1 "$TEST_IMG" +$QEMU_IMG map --output=json --image-opts \ + "$IMG,x-dirty-bitmap=qemu:dirty-bitmap:b1" | _filter_qemu_img_map +nbd_server_start_unix_socket -r -f qcow2 -B b2 "$TEST_IMG" +$QEMU_IMG map --output=json --image-opts \ + "$IMG,x-dirty-bitmap=qemu:dirty-bitmap:b2" | _filter_qemu_img_map + +# success, all done +echo '*** done' +rm -f $seq.full +status=0
tests/qemu-iotests/291.out new file +80 -0
80 lines not on this page · raw message
tests/qemu-iotests/group +1 -0
@@ -299,5 +299,6 @@ 288 quick 289 rw quick 290 rw auto quick +291 rw quick 292 rw auto quick 297 meta
Signature · 1 line
2.26.2
Re: [PATCH v6 5/5] iotests: Add test 291 to for qemu-img bitmap coverage
Am 21.05.2020 um 21:21 hat Eric Blake geschrieben:
Quoted 6 lines
Add a new test covering the 'qemu-img bitmap' subcommand, as well as 'qemu-img convert --bitmaps', both added in recent patches.
Signed-off-by Eric Blake <[email protected]>Reviewed-by Max Reitz <[email protected]>Reviewed-by Vladimir Sementsov-Ogievskiy <[email protected]>
This one fails for me. We need to filter out "disk size" because it depends on the filesystem.
Kevin
291 fail [21:10:19] [21:10:21] output mismatch (see 291.out.bad)
home/kwolf/source/qemu/tests/qemu-iotests/291.out.bad 2020-06-08 21:10:21.043763527 +0200 renamed from home/kwolf/source/qemu/tests/qemu-iotests/291.out 2020-06-08 21:04:38.967014949 +0200 +2 -2
@@ -24,7 +24,7 @@ image: TEST_DIR/t.IMGFMT file format: IMGFMT virtual size: 10 MiB (10485760 bytes) -disk size: 4.39 MiB +disk size: 4.51 MiB Format specific information: compat: 1.1 compression type: zlib
@@ -44,7 +44,7 @@ image: TEST_DIR/t.IMGFMT file format: IMGFMT virtual size: 10 MiB (10485760 bytes) -disk size: 4.48 MiB +disk size: 5.08 MiB Format specific information: compat: 1.1 compression type: zlib
Failures: 291
Failed 1 of 1 iotests
Re: [PATCH v6 5/5] iotests: Add test 291 to for qemu-img bitmap coverage
Quoted 10 lines · Kevin Wolf
Am 21.05.2020 um 21:21 hat Eric Blake geschrieben:
Quoted 6 lines
Add a new test covering the 'qemu-img bitmap' subcommand, as well as 'qemu-img convert --bitmaps', both added in recent patches.
Signed-off-by Eric Blake <[email protected]>Reviewed-by Max Reitz <[email protected]>Reviewed-by Vladimir Sementsov-Ogievskiy <[email protected]>This one fails for me. We need to filter out "disk size" because it depends on the filesystem.
Will fix. Thanks for the heads up.
Signature · 3 lines
Eric Blake, Principal Software Engineer Red Hat, Inc. +1-919-301-3226 Virtualization: qemu.org | libvirt.org