xref: /openbmc/qemu/docs/interop/qcow2.txt (revision e6546342)
1== General ==
2
3A qcow2 image file is organized in units of constant size, which are called
4(host) clusters. A cluster is the unit in which all allocations are done,
5both for actual guest data and for image metadata.
6
7Likewise, the virtual disk as seen by the guest is divided into (guest)
8clusters of the same size.
9
10All numbers in qcow2 are stored in Big Endian byte order.
11
12
13== Header ==
14
15The first cluster of a qcow2 image contains the file header:
16
17    Byte  0 -  3:   magic
18                    QCOW magic string ("QFI\xfb")
19
20          4 -  7:   version
21                    Version number (valid values are 2 and 3)
22
23          8 - 15:   backing_file_offset
24                    Offset into the image file at which the backing file name
25                    is stored (NB: The string is not null terminated). 0 if the
26                    image doesn't have a backing file.
27
28                    Note: backing files are incompatible with raw external data
29                    files (auto-clear feature bit 1).
30
31         16 - 19:   backing_file_size
32                    Length of the backing file name in bytes. Must not be
33                    longer than 1023 bytes. Undefined if the image doesn't have
34                    a backing file.
35
36         20 - 23:   cluster_bits
37                    Number of bits that are used for addressing an offset
38                    within a cluster (1 << cluster_bits is the cluster size).
39                    Must not be less than 9 (i.e. 512 byte clusters).
40
41                    Note: qemu as of today has an implementation limit of 2 MB
42                    as the maximum cluster size and won't be able to open images
43                    with larger cluster sizes.
44
45                    Note: if the image has Extended L2 Entries then cluster_bits
46                    must be at least 14 (i.e. 16384 byte clusters).
47
48         24 - 31:   size
49                    Virtual disk size in bytes.
50
51                    Note: qemu has an implementation limit of 32 MB as
52                    the maximum L1 table size.  With a 2 MB cluster
53                    size, it is unable to populate a virtual cluster
54                    beyond 2 EB (61 bits); with a 512 byte cluster
55                    size, it is unable to populate a virtual size
56                    larger than 128 GB (37 bits).  Meanwhile, L1/L2
57                    table layouts limit an image to no more than 64 PB
58                    (56 bits) of populated clusters, and an image may
59                    hit other limits first (such as a file system's
60                    maximum size).
61
62         32 - 35:   crypt_method
63                    0 for no encryption
64                    1 for AES encryption
65                    2 for LUKS encryption
66
67         36 - 39:   l1_size
68                    Number of entries in the active L1 table
69
70         40 - 47:   l1_table_offset
71                    Offset into the image file at which the active L1 table
72                    starts. Must be aligned to a cluster boundary.
73
74         48 - 55:   refcount_table_offset
75                    Offset into the image file at which the refcount table
76                    starts. Must be aligned to a cluster boundary.
77
78         56 - 59:   refcount_table_clusters
79                    Number of clusters that the refcount table occupies
80
81         60 - 63:   nb_snapshots
82                    Number of snapshots contained in the image
83
84         64 - 71:   snapshots_offset
85                    Offset into the image file at which the snapshot table
86                    starts. Must be aligned to a cluster boundary.
87
88For version 2, the header is exactly 72 bytes in length, and finishes here.
89For version 3 or higher, the header length is at least 104 bytes, including
90the next fields through header_length.
91
92         72 -  79:  incompatible_features
93                    Bitmask of incompatible features. An implementation must
94                    fail to open an image if an unknown bit is set.
95
96                    Bit 0:      Dirty bit.  If this bit is set then refcounts
97                                may be inconsistent, make sure to scan L1/L2
98                                tables to repair refcounts before accessing the
99                                image.
100
101                    Bit 1:      Corrupt bit.  If this bit is set then any data
102                                structure may be corrupt and the image must not
103                                be written to (unless for regaining
104                                consistency).
105
106                    Bit 2:      External data file bit.  If this bit is set, an
107                                external data file is used. Guest clusters are
108                                then stored in the external data file. For such
109                                images, clusters in the external data file are
110                                not refcounted. The offset field in the
111                                Standard Cluster Descriptor must match the
112                                guest offset and neither compressed clusters
113                                nor internal snapshots are supported.
114
115                                An External Data File Name header extension may
116                                be present if this bit is set.
117
118                    Bit 3:      Compression type bit.  If this bit is set,
119                                a non-default compression is used for compressed
120                                clusters. The compression_type field must be
121                                present and not zero.
122
123                    Bit 4:      Extended L2 Entries.  If this bit is set then
124                                L2 table entries use an extended format that
125                                allows subcluster-based allocation. See the
126                                Extended L2 Entries section for more details.
127
128                    Bits 5-63:  Reserved (set to 0)
129
130         80 -  87:  compatible_features
131                    Bitmask of compatible features. An implementation can
132                    safely ignore any unknown bits that are set.
133
134                    Bit 0:      Lazy refcounts bit.  If this bit is set then
135                                lazy refcount updates can be used.  This means
136                                marking the image file dirty and postponing
137                                refcount metadata updates.
138
139                    Bits 1-63:  Reserved (set to 0)
140
141         88 -  95:  autoclear_features
142                    Bitmask of auto-clear features. An implementation may only
143                    write to an image with unknown auto-clear features if it
144                    clears the respective bits from this field first.
145
146                    Bit 0:      Bitmaps extension bit
147                                This bit indicates consistency for the bitmaps
148                                extension data.
149
150                                It is an error if this bit is set without the
151                                bitmaps extension present.
152
153                                If the bitmaps extension is present but this
154                                bit is unset, the bitmaps extension data must be
155                                considered inconsistent.
156
157                    Bit 1:      Raw external data bit
158                                If this bit is set, the external data file can
159                                be read as a consistent standalone raw image
160                                without looking at the qcow2 metadata.
161
162                                Setting this bit has a performance impact for
163                                some operations on the image (e.g. writing
164                                zeros requires writing to the data file instead
165                                of only setting the zero flag in the L2 table
166                                entry) and conflicts with backing files.
167
168                                This bit may only be set if the External Data
169                                File bit (incompatible feature bit 1) is also
170                                set.
171
172                    Bits 2-63:  Reserved (set to 0)
173
174         96 -  99:  refcount_order
175                    Describes the width of a reference count block entry (width
176                    in bits: refcount_bits = 1 << refcount_order). For version 2
177                    images, the order is always assumed to be 4
178                    (i.e. refcount_bits = 16).
179                    This value may not exceed 6 (i.e. refcount_bits = 64).
180
181        100 - 103:  header_length
182                    Length of the header structure in bytes. For version 2
183                    images, the length is always assumed to be 72 bytes.
184                    For version 3 it's at least 104 bytes and must be a multiple
185                    of 8.
186
187
188=== Additional fields (version 3 and higher) ===
189
190In general, these fields are optional and may be safely ignored by the software,
191as well as filled by zeros (which is equal to field absence), if software needs
192to set field B, but does not care about field A which precedes B. More
193formally, additional fields have the following compatibility rules:
194
1951. If the value of the additional field must not be ignored for correct
196handling of the file, it will be accompanied by a corresponding incompatible
197feature bit.
198
1992. If there are no unrecognized incompatible feature bits set, an unknown
200additional field may be safely ignored other than preserving its value when
201rewriting the image header.
202
2033. An explicit value of 0 will have the same behavior as when the field is not
204present*, if not altered by a specific incompatible bit.
205
206*. A field is considered not present when header_length is less than or equal
207to the field's offset. Also, all additional fields are not present for
208version 2.
209
210              104:  compression_type
211
212                    Defines the compression method used for compressed clusters.
213                    All compressed clusters in an image use the same compression
214                    type.
215
216                    If the incompatible bit "Compression type" is set: the field
217                    must be present and non-zero (which means non-zlib
218                    compression type). Otherwise, this field must not be present
219                    or must be zero (which means zlib).
220
221                    Available compression type values:
222                        0: zlib <https://www.zlib.net/>
223                        1: zstd <http://github.com/facebook/zstd>
224
225
226=== Header padding ===
227
228@header_length must be a multiple of 8, which means that if the end of the last
229additional field is not aligned, some padding is needed. This padding must be
230zeroed, so that if some existing (or future) additional field will fall into
231the padding, it will be interpreted accordingly to point [3.] of the previous
232paragraph, i.e.  in the same manner as when this field is not present.
233
234
235=== Header extensions ===
236
237Directly after the image header, optional sections called header extensions can
238be stored. Each extension has a structure like the following:
239
240    Byte  0 -  3:   Header extension type:
241                        0x00000000 - End of the header extension area
242                        0xe2792aca - Backing file format name string
243                        0x6803f857 - Feature name table
244                        0x23852875 - Bitmaps extension
245                        0x0537be77 - Full disk encryption header pointer
246                        0x44415441 - External data file name string
247                        other      - Unknown header extension, can be safely
248                                     ignored
249
250          4 -  7:   Length of the header extension data
251
252          8 -  n:   Header extension data
253
254          n -  m:   Padding to round up the header extension size to the next
255                    multiple of 8.
256
257Unless stated otherwise, each header extension type shall appear at most once
258in the same image.
259
260If the image has a backing file then the backing file name should be stored in
261the remaining space between the end of the header extension area and the end of
262the first cluster. It is not allowed to store other data here, so that an
263implementation can safely modify the header and add extensions without harming
264data of compatible features that it doesn't support. Compatible features that
265need space for additional data can use a header extension.
266
267
268== String header extensions ==
269
270Some header extensions (such as the backing file format name and the external
271data file name) are just a single string. In this case, the header extension
272length is the string length and the string is not '\0' terminated. (The header
273extension padding can make it look like a string is '\0' terminated, but
274neither is padding always necessary nor is there a guarantee that zero bytes
275are used for padding.)
276
277
278== Feature name table ==
279
280The feature name table is an optional header extension that contains the name
281for features used by the image. It can be used by applications that don't know
282the respective feature (e.g. because the feature was introduced only later) to
283display a useful error message.
284
285The number of entries in the feature name table is determined by the length of
286the header extension data. Each entry look like this:
287
288    Byte       0:   Type of feature (select feature bitmap)
289                        0: Incompatible feature
290                        1: Compatible feature
291                        2: Autoclear feature
292
293               1:   Bit number within the selected feature bitmap (valid
294                    values: 0-63)
295
296          2 - 47:   Feature name (padded with zeros, but not necessarily null
297                    terminated if it has full length)
298
299
300== Bitmaps extension ==
301
302The bitmaps extension is an optional header extension. It provides the ability
303to store bitmaps related to a virtual disk. For now, there is only one bitmap
304type: the dirty tracking bitmap, which tracks virtual disk changes from some
305point in time.
306
307The data of the extension should be considered consistent only if the
308corresponding auto-clear feature bit is set, see autoclear_features above.
309
310The fields of the bitmaps extension are:
311
312    Byte  0 -  3:  nb_bitmaps
313                   The number of bitmaps contained in the image. Must be
314                   greater than or equal to 1.
315
316                   Note: Qemu currently only supports up to 65535 bitmaps per
317                   image.
318
319          4 -  7:  Reserved, must be zero.
320
321          8 - 15:  bitmap_directory_size
322                   Size of the bitmap directory in bytes. It is the cumulative
323                   size of all (nb_bitmaps) bitmap directory entries.
324
325         16 - 23:  bitmap_directory_offset
326                   Offset into the image file at which the bitmap directory
327                   starts. Must be aligned to a cluster boundary.
328
329== Full disk encryption header pointer ==
330
331The full disk encryption header must be present if, and only if, the
332'crypt_method' header requires metadata. Currently this is only true
333of the 'LUKS' crypt method. The header extension must be absent for
334other methods.
335
336This header provides the offset at which the crypt method can store
337its additional data, as well as the length of such data.
338
339    Byte  0 -  7:   Offset into the image file at which the encryption
340                    header starts in bytes. Must be aligned to a cluster
341                    boundary.
342    Byte  8 - 15:   Length of the written encryption header in bytes.
343                    Note actual space allocated in the qcow2 file may
344                    be larger than this value, since it will be rounded
345                    to the nearest multiple of the cluster size. Any
346                    unused bytes in the allocated space will be initialized
347                    to 0.
348
349For the LUKS crypt method, the encryption header works as follows.
350
351The first 592 bytes of the header clusters will contain the LUKS
352partition header. This is then followed by the key material data areas.
353The size of the key material data areas is determined by the number of
354stripes in the key slot and key size. Refer to the LUKS format
355specification ('docs/on-disk-format.pdf' in the cryptsetup source
356package) for details of the LUKS partition header format.
357
358In the LUKS partition header, the "payload-offset" field will be
359calculated as normal for the LUKS spec. ie the size of the LUKS
360header, plus key material regions, plus padding, relative to the
361start of the LUKS header. This offset value is not required to be
362qcow2 cluster aligned. Its value is currently never used in the
363context of qcow2, since the qcow2 file format itself defines where
364the real payload offset is, but none the less a valid payload offset
365should always be present.
366
367In the LUKS key slots header, the "key-material-offset" is relative
368to the start of the LUKS header clusters in the qcow2 container,
369not the start of the qcow2 file.
370
371Logically the layout looks like
372
373  +-----------------------------+
374  | QCow2 header                |
375  | QCow2 header extension X    |
376  | QCow2 header extension FDE  |
377  | QCow2 header extension ...  |
378  | QCow2 header extension Z    |
379  +-----------------------------+
380  | ....other QCow2 tables....  |
381  .                             .
382  .                             .
383  +-----------------------------+
384  | +-------------------------+ |
385  | | LUKS partition header   | |
386  | +-------------------------+ |
387  | | LUKS key material 1     | |
388  | +-------------------------+ |
389  | | LUKS key material 2     | |
390  | +-------------------------+ |
391  | | LUKS key material ...   | |
392  | +-------------------------+ |
393  | | LUKS key material 8     | |
394  | +-------------------------+ |
395  +-----------------------------+
396  | QCow2 cluster payload       |
397  .                             .
398  .                             .
399  .                             .
400  |                             |
401  +-----------------------------+
402
403== Data encryption ==
404
405When an encryption method is requested in the header, the image payload
406data must be encrypted/decrypted on every write/read. The image headers
407and metadata are never encrypted.
408
409The algorithms used for encryption vary depending on the method
410
411 - AES:
412
413   The AES cipher, in CBC mode, with 256 bit keys.
414
415   Initialization vectors generated using plain64 method, with
416   the virtual disk sector as the input tweak.
417
418   This format is no longer supported in QEMU system emulators, due
419   to a number of design flaws affecting its security. It is only
420   supported in the command line tools for the sake of back compatibility
421   and data liberation.
422
423 - LUKS:
424
425   The algorithms are specified in the LUKS header.
426
427   Initialization vectors generated using the method specified
428   in the LUKS header, with the physical disk sector as the
429   input tweak.
430
431== Host cluster management ==
432
433qcow2 manages the allocation of host clusters by maintaining a reference count
434for each host cluster. A refcount of 0 means that the cluster is free, 1 means
435that it is used, and >= 2 means that it is used and any write access must
436perform a COW (copy on write) operation.
437
438The refcounts are managed in a two-level table. The first level is called
439refcount table and has a variable size (which is stored in the header). The
440refcount table can cover multiple clusters, however it needs to be contiguous
441in the image file.
442
443It contains pointers to the second level structures which are called refcount
444blocks and are exactly one cluster in size.
445
446Although a large enough refcount table can reserve clusters past 64 PB
447(56 bits) (assuming the underlying protocol can even be sized that
448large), note that some qcow2 metadata such as L1/L2 tables must point
449to clusters prior to that point.
450
451Note: qemu has an implementation limit of 8 MB as the maximum refcount
452table size.  With a 2 MB cluster size and a default refcount_order of
4534, it is unable to reference host resources beyond 2 EB (61 bits); in
454the worst case, with a 512 cluster size and refcount_order of 6, it is
455unable to access beyond 32 GB (35 bits).
456
457Given an offset into the image file, the refcount of its cluster can be
458obtained as follows:
459
460    refcount_block_entries = (cluster_size * 8 / refcount_bits)
461
462    refcount_block_index = (offset / cluster_size) % refcount_block_entries
463    refcount_table_index = (offset / cluster_size) / refcount_block_entries
464
465    refcount_block = load_cluster(refcount_table[refcount_table_index]);
466    return refcount_block[refcount_block_index];
467
468Refcount table entry:
469
470    Bit  0 -  8:    Reserved (set to 0)
471
472         9 - 63:    Bits 9-63 of the offset into the image file at which the
473                    refcount block starts. Must be aligned to a cluster
474                    boundary.
475
476                    If this is 0, the corresponding refcount block has not yet
477                    been allocated. All refcounts managed by this refcount block
478                    are 0.
479
480Refcount block entry (x = refcount_bits - 1):
481
482    Bit  0 -  x:    Reference count of the cluster. If refcount_bits implies a
483                    sub-byte width, note that bit 0 means the least significant
484                    bit in this context.
485
486
487== Cluster mapping ==
488
489Just as for refcounts, qcow2 uses a two-level structure for the mapping of
490guest clusters to host clusters. They are called L1 and L2 table.
491
492The L1 table has a variable size (stored in the header) and may use multiple
493clusters, however it must be contiguous in the image file. L2 tables are
494exactly one cluster in size.
495
496The L1 and L2 tables have implications on the maximum virtual file
497size; for a given L1 table size, a larger cluster size is required for
498the guest to have access to more space.  Furthermore, a virtual
499cluster must currently map to a host offset below 64 PB (56 bits)
500(although this limit could be relaxed by putting reserved bits into
501use).  Additionally, as cluster size increases, the maximum host
502offset for a compressed cluster is reduced (a 2M cluster size requires
503compressed clusters to reside below 512 TB (49 bits), and this limit
504cannot be relaxed without an incompatible layout change).
505
506Given an offset into the virtual disk, the offset into the image file can be
507obtained as follows:
508
509    l2_entries = (cluster_size / sizeof(uint64_t))        [*]
510
511    l2_index = (offset / cluster_size) % l2_entries
512    l1_index = (offset / cluster_size) / l2_entries
513
514    l2_table = load_cluster(l1_table[l1_index]);
515    cluster_offset = l2_table[l2_index];
516
517    return cluster_offset + (offset % cluster_size)
518
519    [*] this changes if Extended L2 Entries are enabled, see next section
520
521L1 table entry:
522
523    Bit  0 -  8:    Reserved (set to 0)
524
525         9 - 55:    Bits 9-55 of the offset into the image file at which the L2
526                    table starts. Must be aligned to a cluster boundary. If the
527                    offset is 0, the L2 table and all clusters described by this
528                    L2 table are unallocated.
529
530        56 - 62:    Reserved (set to 0)
531
532             63:    0 for an L2 table that is unused or requires COW, 1 if its
533                    refcount is exactly one. This information is only accurate
534                    in the active L1 table.
535
536L2 table entry:
537
538    Bit  0 -  61:   Cluster descriptor
539
540              62:   0 for standard clusters
541                    1 for compressed clusters
542
543              63:   0 for clusters that are unused, compressed or require COW.
544                    1 for standard clusters whose refcount is exactly one.
545                    This information is only accurate in L2 tables
546                    that are reachable from the active L1 table.
547
548                    With external data files, all guest clusters have an
549                    implicit refcount of 1 (because of the fixed host = guest
550                    mapping for guest cluster offsets), so this bit should be 1
551                    for all allocated clusters.
552
553Standard Cluster Descriptor:
554
555    Bit       0:    If set to 1, the cluster reads as all zeros. The host
556                    cluster offset can be used to describe a preallocation,
557                    but it won't be used for reading data from this cluster,
558                    nor is data read from the backing file if the cluster is
559                    unallocated.
560
561                    With version 2 or with extended L2 entries (see the next
562                    section), this is always 0.
563
564         1 -  8:    Reserved (set to 0)
565
566         9 - 55:    Bits 9-55 of host cluster offset. Must be aligned to a
567                    cluster boundary. If the offset is 0 and bit 63 is clear,
568                    the cluster is unallocated. The offset may only be 0 with
569                    bit 63 set (indicating a host cluster offset of 0) when an
570                    external data file is used.
571
572        56 - 61:    Reserved (set to 0)
573
574
575Compressed Clusters Descriptor (x = 62 - (cluster_bits - 8)):
576
577    Bit  0 - x-1:   Host cluster offset. This is usually _not_ aligned to a
578                    cluster or sector boundary!  If cluster_bits is
579                    small enough that this field includes bits beyond
580                    55, those upper bits must be set to 0.
581
582         x - 61:    Number of additional 512-byte sectors used for the
583                    compressed data, beyond the sector containing the offset
584                    in the previous field. Some of these sectors may reside
585                    in the next contiguous host cluster.
586
587                    Note that the compressed data does not necessarily occupy
588                    all of the bytes in the final sector; rather, decompression
589                    stops when it has produced a cluster of data.
590
591                    Another compressed cluster may map to the tail of the final
592                    sector used by this compressed cluster.
593
594If a cluster is unallocated, read requests shall read the data from the backing
595file (except if bit 0 in the Standard Cluster Descriptor is set). If there is
596no backing file or the backing file is smaller than the image, they shall read
597zeros for all parts that are not covered by the backing file.
598
599== Extended L2 Entries ==
600
601An image uses Extended L2 Entries if bit 4 is set on the incompatible_features
602field of the header.
603
604In these images standard data clusters are divided into 32 subclusters of the
605same size. They are contiguous and start from the beginning of the cluster.
606Subclusters can be allocated independently and the L2 entry contains information
607indicating the status of each one of them. Compressed data clusters don't have
608subclusters so they are treated the same as in images without this feature.
609
610The size of an extended L2 entry is 128 bits so the number of entries per table
611is calculated using this formula:
612
613    l2_entries = (cluster_size / (2 * sizeof(uint64_t)))
614
615The first 64 bits have the same format as the standard L2 table entry described
616in the previous section, with the exception of bit 0 of the standard cluster
617descriptor.
618
619The last 64 bits contain a subcluster allocation bitmap with this format:
620
621Subcluster Allocation Bitmap (for standard clusters):
622
623    Bit  0 - 31:    Allocation status (one bit per subcluster)
624
625                    1: the subcluster is allocated. In this case the
626                       host cluster offset field must contain a valid
627                       offset.
628                    0: the subcluster is not allocated. In this case
629                       read requests shall go to the backing file or
630                       return zeros if there is no backing file data.
631
632                    Bits are assigned starting from the least significant
633                    one (i.e. bit x is used for subcluster x).
634
635        32 - 63     Subcluster reads as zeros (one bit per subcluster)
636
637                    1: the subcluster reads as zeros. In this case the
638                       allocation status bit must be unset. The host
639                       cluster offset field may or may not be set.
640                    0: no effect.
641
642                    Bits are assigned starting from the least significant
643                    one (i.e. bit x is used for subcluster x - 32).
644
645Subcluster Allocation Bitmap (for compressed clusters):
646
647    Bit  0 - 63:    Reserved (set to 0)
648                    Compressed clusters don't have subclusters,
649                    so this field is not used.
650
651== Snapshots ==
652
653qcow2 supports internal snapshots. Their basic principle of operation is to
654switch the active L1 table, so that a different set of host clusters are
655exposed to the guest.
656
657When creating a snapshot, the L1 table should be copied and the refcount of all
658L2 tables and clusters reachable from this L1 table must be increased, so that
659a write causes a COW and isn't visible in other snapshots.
660
661When loading a snapshot, bit 63 of all entries in the new active L1 table and
662all L2 tables referenced by it must be reconstructed from the refcount table
663as it doesn't need to be accurate in inactive L1 tables.
664
665A directory of all snapshots is stored in the snapshot table, a contiguous area
666in the image file, whose starting offset and length are given by the header
667fields snapshots_offset and nb_snapshots. The entries of the snapshot table
668have variable length, depending on the length of ID, name and extra data.
669
670Snapshot table entry:
671
672    Byte 0 -  7:    Offset into the image file at which the L1 table for the
673                    snapshot starts. Must be aligned to a cluster boundary.
674
675         8 - 11:    Number of entries in the L1 table of the snapshots
676
677        12 - 13:    Length of the unique ID string describing the snapshot
678
679        14 - 15:    Length of the name of the snapshot
680
681        16 - 19:    Time at which the snapshot was taken in seconds since the
682                    Epoch
683
684        20 - 23:    Subsecond part of the time at which the snapshot was taken
685                    in nanoseconds
686
687        24 - 31:    Time that the guest was running until the snapshot was
688                    taken in nanoseconds
689
690        32 - 35:    Size of the VM state in bytes. 0 if no VM state is saved.
691                    If there is VM state, it starts at the first cluster
692                    described by first L1 table entry that doesn't describe a
693                    regular guest cluster (i.e. VM state is stored like guest
694                    disk content, except that it is stored at offsets that are
695                    larger than the virtual disk presented to the guest)
696
697        36 - 39:    Size of extra data in the table entry (used for future
698                    extensions of the format)
699
700        variable:   Extra data for future extensions. Unknown fields must be
701                    ignored. Currently defined are (offset relative to snapshot
702                    table entry):
703
704                    Byte 40 - 47:   Size of the VM state in bytes. 0 if no VM
705                                    state is saved. If this field is present,
706                                    the 32-bit value in bytes 32-35 is ignored.
707
708                    Byte 48 - 55:   Virtual disk size of the snapshot in bytes
709
710                    Version 3 images must include extra data at least up to
711                    byte 55.
712
713        variable:   Unique ID string for the snapshot (not null terminated)
714
715        variable:   Name of the snapshot (not null terminated)
716
717        variable:   Padding to round up the snapshot table entry size to the
718                    next multiple of 8.
719
720
721== Bitmaps ==
722
723As mentioned above, the bitmaps extension provides the ability to store bitmaps
724related to a virtual disk. This section describes how these bitmaps are stored.
725
726All stored bitmaps are related to the virtual disk stored in the same image, so
727each bitmap size is equal to the virtual disk size.
728
729Each bit of the bitmap is responsible for strictly defined range of the virtual
730disk. For bit number bit_nr the corresponding range (in bytes) will be:
731
732    [bit_nr * bitmap_granularity .. (bit_nr + 1) * bitmap_granularity - 1]
733
734Granularity is a property of the concrete bitmap, see below.
735
736
737=== Bitmap directory ===
738
739Each bitmap saved in the image is described in a bitmap directory entry. The
740bitmap directory is a contiguous area in the image file, whose starting offset
741and length are given by the header extension fields bitmap_directory_offset and
742bitmap_directory_size. The entries of the bitmap directory have variable
743length, depending on the lengths of the bitmap name and extra data.
744
745Structure of a bitmap directory entry:
746
747    Byte 0 -  7:    bitmap_table_offset
748                    Offset into the image file at which the bitmap table
749                    (described below) for the bitmap starts. Must be aligned to
750                    a cluster boundary.
751
752         8 - 11:    bitmap_table_size
753                    Number of entries in the bitmap table of the bitmap.
754
755        12 - 15:    flags
756                    Bit
757                      0: in_use
758                         The bitmap was not saved correctly and may be
759                         inconsistent. Although the bitmap metadata is still
760                         well-formed from a qcow2 perspective, the metadata
761                         (such as the auto flag or bitmap size) or data
762                         contents may be outdated.
763
764                      1: auto
765                         The bitmap must reflect all changes of the virtual
766                         disk by any application that would write to this qcow2
767                         file (including writes, snapshot switching, etc.). The
768                         type of this bitmap must be 'dirty tracking bitmap'.
769
770                      2: extra_data_compatible
771                         This flags is meaningful when the extra data is
772                         unknown to the software (currently any extra data is
773                         unknown to Qemu).
774                         If it is set, the bitmap may be used as expected, extra
775                         data must be left as is.
776                         If it is not set, the bitmap must not be used, but
777                         both it and its extra data be left as is.
778
779                    Bits 3 - 31 are reserved and must be 0.
780
781             16:    type
782                    This field describes the sort of the bitmap.
783                    Values:
784                      1: Dirty tracking bitmap
785
786                    Values 0, 2 - 255 are reserved.
787
788             17:    granularity_bits
789                    Granularity bits. Valid values: 0 - 63.
790
791                    Note: Qemu currently supports only values 9 - 31.
792
793                    Granularity is calculated as
794                        granularity = 1 << granularity_bits
795
796                    A bitmap's granularity is how many bytes of the image
797                    accounts for one bit of the bitmap.
798
799        18 - 19:    name_size
800                    Size of the bitmap name. Must be non-zero.
801
802                    Note: Qemu currently doesn't support values greater than
803                    1023.
804
805        20 - 23:    extra_data_size
806                    Size of type-specific extra data.
807
808                    For now, as no extra data is defined, extra_data_size is
809                    reserved and should be zero. If it is non-zero the
810                    behavior is defined by extra_data_compatible flag.
811
812        variable:   extra_data
813                    Extra data for the bitmap, occupying extra_data_size bytes.
814                    Extra data must never contain references to clusters or in
815                    some other way allocate additional clusters.
816
817        variable:   name
818                    The name of the bitmap (not null terminated), occupying
819                    name_size bytes. Must be unique among all bitmap names
820                    within the bitmaps extension.
821
822        variable:   Padding to round up the bitmap directory entry size to the
823                    next multiple of 8. All bytes of the padding must be zero.
824
825
826=== Bitmap table ===
827
828Each bitmap is stored using a one-level structure (as opposed to two-level
829structures like for refcounts and guest clusters mapping) for the mapping of
830bitmap data to host clusters. This structure is called the bitmap table.
831
832Each bitmap table has a variable size (stored in the bitmap directory entry)
833and may use multiple clusters, however, it must be contiguous in the image
834file.
835
836Structure of a bitmap table entry:
837
838    Bit       0:    Reserved and must be zero if bits 9 - 55 are non-zero.
839                    If bits 9 - 55 are zero:
840                      0: Cluster should be read as all zeros.
841                      1: Cluster should be read as all ones.
842
843         1 -  8:    Reserved and must be zero.
844
845         9 - 55:    Bits 9 - 55 of the host cluster offset. Must be aligned to
846                    a cluster boundary. If the offset is 0, the cluster is
847                    unallocated; in that case, bit 0 determines how this
848                    cluster should be treated during reads.
849
850        56 - 63:    Reserved and must be zero.
851
852
853=== Bitmap data ===
854
855As noted above, bitmap data is stored in separate clusters, described by the
856bitmap table. Given an offset (in bytes) into the bitmap data, the offset into
857the image file can be obtained as follows:
858
859    image_offset(bitmap_data_offset) =
860        bitmap_table[bitmap_data_offset / cluster_size] +
861            (bitmap_data_offset % cluster_size)
862
863This offset is not defined if bits 9 - 55 of bitmap table entry are zero (see
864above).
865
866Given an offset byte_nr into the virtual disk and the bitmap's granularity, the
867bit offset into the image file to the corresponding bit of the bitmap can be
868calculated like this:
869
870    bit_offset(byte_nr) =
871        image_offset(byte_nr / granularity / 8) * 8 +
872            (byte_nr / granularity) % 8
873
874If the size of the bitmap data is not a multiple of the cluster size then the
875last cluster of the bitmap data contains some unused tail bits. These bits must
876be zero.
877
878
879=== Dirty tracking bitmaps ===
880
881Bitmaps with 'type' field equal to one are dirty tracking bitmaps.
882
883When the virtual disk is in use dirty tracking bitmap may be 'enabled' or
884'disabled'. While the bitmap is 'enabled', all writes to the virtual disk
885should be reflected in the bitmap. A set bit in the bitmap means that the
886corresponding range of the virtual disk (see above) was written to while the
887bitmap was 'enabled'. An unset bit means that this range was not written to.
888
889The software doesn't have to sync the bitmap in the image file with its
890representation in RAM after each write or metadata change. Flag 'in_use'
891should be set while the bitmap is not synced.
892
893In the image file the 'enabled' state is reflected by the 'auto' flag. If this
894flag is set, the software must consider the bitmap as 'enabled' and start
895tracking virtual disk changes to this bitmap from the first write to the
896virtual disk. If this flag is not set then the bitmap is disabled.
897