xref: /openbmc/qemu/docs/interop/qcow2.txt (revision 35c7f525)
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         24 - 31:   size
46                    Virtual disk size in bytes.
47
48                    Note: qemu has an implementation limit of 32 MB as
49                    the maximum L1 table size.  With a 2 MB cluster
50                    size, it is unable to populate a virtual cluster
51                    beyond 2 EB (61 bits); with a 512 byte cluster
52                    size, it is unable to populate a virtual size
53                    larger than 128 GB (37 bits).  Meanwhile, L1/L2
54                    table layouts limit an image to no more than 64 PB
55                    (56 bits) of populated clusters, and an image may
56                    hit other limits first (such as a file system's
57                    maximum size).
58
59         32 - 35:   crypt_method
60                    0 for no encryption
61                    1 for AES encryption
62                    2 for LUKS encryption
63
64         36 - 39:   l1_size
65                    Number of entries in the active L1 table
66
67         40 - 47:   l1_table_offset
68                    Offset into the image file at which the active L1 table
69                    starts. Must be aligned to a cluster boundary.
70
71         48 - 55:   refcount_table_offset
72                    Offset into the image file at which the refcount table
73                    starts. Must be aligned to a cluster boundary.
74
75         56 - 59:   refcount_table_clusters
76                    Number of clusters that the refcount table occupies
77
78         60 - 63:   nb_snapshots
79                    Number of snapshots contained in the image
80
81         64 - 71:   snapshots_offset
82                    Offset into the image file at which the snapshot table
83                    starts. Must be aligned to a cluster boundary.
84
85For version 2, the header is exactly 72 bytes in length, and finishes here.
86For version 3 or higher, the header length is at least 104 bytes, including
87the next fields through header_length.
88
89         72 -  79:  incompatible_features
90                    Bitmask of incompatible features. An implementation must
91                    fail to open an image if an unknown bit is set.
92
93                    Bit 0:      Dirty bit.  If this bit is set then refcounts
94                                may be inconsistent, make sure to scan L1/L2
95                                tables to repair refcounts before accessing the
96                                image.
97
98                    Bit 1:      Corrupt bit.  If this bit is set then any data
99                                structure may be corrupt and the image must not
100                                be written to (unless for regaining
101                                consistency).
102
103                    Bit 2:      External data file bit.  If this bit is set, an
104                                external data file is used. Guest clusters are
105                                then stored in the external data file. For such
106                                images, clusters in the external data file are
107                                not refcounted. The offset field in the
108                                Standard Cluster Descriptor must match the
109                                guest offset and neither compressed clusters
110                                nor internal snapshots are supported.
111
112                                An External Data File Name header extension may
113                                be present if this bit is set.
114
115                    Bit 3:      Compression type bit.  If this bit is set,
116                                a non-default compression is used for compressed
117                                clusters. The compression_type field must be
118                                present and not zero.
119
120                    Bits 4-63:  Reserved (set to 0)
121
122         80 -  87:  compatible_features
123                    Bitmask of compatible features. An implementation can
124                    safely ignore any unknown bits that are set.
125
126                    Bit 0:      Lazy refcounts bit.  If this bit is set then
127                                lazy refcount updates can be used.  This means
128                                marking the image file dirty and postponing
129                                refcount metadata updates.
130
131                    Bits 1-63:  Reserved (set to 0)
132
133         88 -  95:  autoclear_features
134                    Bitmask of auto-clear features. An implementation may only
135                    write to an image with unknown auto-clear features if it
136                    clears the respective bits from this field first.
137
138                    Bit 0:      Bitmaps extension bit
139                                This bit indicates consistency for the bitmaps
140                                extension data.
141
142                                It is an error if this bit is set without the
143                                bitmaps extension present.
144
145                                If the bitmaps extension is present but this
146                                bit is unset, the bitmaps extension data must be
147                                considered inconsistent.
148
149                    Bit 1:      Raw external data bit
150                                If this bit is set, the external data file can
151                                be read as a consistent standalone raw image
152                                without looking at the qcow2 metadata.
153
154                                Setting this bit has a performance impact for
155                                some operations on the image (e.g. writing
156                                zeros requires writing to the data file instead
157                                of only setting the zero flag in the L2 table
158                                entry) and conflicts with backing files.
159
160                                This bit may only be set if the External Data
161                                File bit (incompatible feature bit 1) is also
162                                set.
163
164                    Bits 2-63:  Reserved (set to 0)
165
166         96 -  99:  refcount_order
167                    Describes the width of a reference count block entry (width
168                    in bits: refcount_bits = 1 << refcount_order). For version 2
169                    images, the order is always assumed to be 4
170                    (i.e. refcount_bits = 16).
171                    This value may not exceed 6 (i.e. refcount_bits = 64).
172
173        100 - 103:  header_length
174                    Length of the header structure in bytes. For version 2
175                    images, the length is always assumed to be 72 bytes.
176                    For version 3 it's at least 104 bytes and must be a multiple
177                    of 8.
178
179
180=== Additional fields (version 3 and higher) ===
181
182In general, these fields are optional and may be safely ignored by the software,
183as well as filled by zeros (which is equal to field absence), if software needs
184to set field B, but does not care about field A which precedes B. More
185formally, additional fields have the following compatibility rules:
186
1871. If the value of the additional field must not be ignored for correct
188handling of the file, it will be accompanied by a corresponding incompatible
189feature bit.
190
1912. If there are no unrecognized incompatible feature bits set, an unknown
192additional field may be safely ignored other than preserving its value when
193rewriting the image header.
194
1953. An explicit value of 0 will have the same behavior as when the field is not
196present*, if not altered by a specific incompatible bit.
197
198*. A field is considered not present when header_length is less than or equal
199to the field's offset. Also, all additional fields are not present for
200version 2.
201
202              104:  compression_type
203
204                    Defines the compression method used for compressed clusters.
205                    All compressed clusters in an image use the same compression
206                    type.
207
208                    If the incompatible bit "Compression type" is set: the field
209                    must be present and non-zero (which means non-zlib
210                    compression type). Otherwise, this field must not be present
211                    or must be zero (which means zlib).
212
213                    Available compression type values:
214                        0: zlib <https://www.zlib.net/>
215                        1: zstd <http://github.com/facebook/zstd>
216
217
218=== Header padding ===
219
220@header_length must be a multiple of 8, which means that if the end of the last
221additional field is not aligned, some padding is needed. This padding must be
222zeroed, so that if some existing (or future) additional field will fall into
223the padding, it will be interpreted accordingly to point [3.] of the previous
224paragraph, i.e.  in the same manner as when this field is not present.
225
226
227=== Header extensions ===
228
229Directly after the image header, optional sections called header extensions can
230be stored. Each extension has a structure like the following:
231
232    Byte  0 -  3:   Header extension type:
233                        0x00000000 - End of the header extension area
234                        0xe2792aca - Backing file format name string
235                        0x6803f857 - Feature name table
236                        0x23852875 - Bitmaps extension
237                        0x0537be77 - Full disk encryption header pointer
238                        0x44415441 - External data file name string
239                        other      - Unknown header extension, can be safely
240                                     ignored
241
242          4 -  7:   Length of the header extension data
243
244          8 -  n:   Header extension data
245
246          n -  m:   Padding to round up the header extension size to the next
247                    multiple of 8.
248
249Unless stated otherwise, each header extension type shall appear at most once
250in the same image.
251
252If the image has a backing file then the backing file name should be stored in
253the remaining space between the end of the header extension area and the end of
254the first cluster. It is not allowed to store other data here, so that an
255implementation can safely modify the header and add extensions without harming
256data of compatible features that it doesn't support. Compatible features that
257need space for additional data can use a header extension.
258
259
260== String header extensions ==
261
262Some header extensions (such as the backing file format name and the external
263data file name) are just a single string. In this case, the header extension
264length is the string length and the string is not '\0' terminated. (The header
265extension padding can make it look like a string is '\0' terminated, but
266neither is padding always necessary nor is there a guarantee that zero bytes
267are used for padding.)
268
269
270== Feature name table ==
271
272The feature name table is an optional header extension that contains the name
273for features used by the image. It can be used by applications that don't know
274the respective feature (e.g. because the feature was introduced only later) to
275display a useful error message.
276
277The number of entries in the feature name table is determined by the length of
278the header extension data. Each entry look like this:
279
280    Byte       0:   Type of feature (select feature bitmap)
281                        0: Incompatible feature
282                        1: Compatible feature
283                        2: Autoclear feature
284
285               1:   Bit number within the selected feature bitmap (valid
286                    values: 0-63)
287
288          2 - 47:   Feature name (padded with zeros, but not necessarily null
289                    terminated if it has full length)
290
291
292== Bitmaps extension ==
293
294The bitmaps extension is an optional header extension. It provides the ability
295to store bitmaps related to a virtual disk. For now, there is only one bitmap
296type: the dirty tracking bitmap, which tracks virtual disk changes from some
297point in time.
298
299The data of the extension should be considered consistent only if the
300corresponding auto-clear feature bit is set, see autoclear_features above.
301
302The fields of the bitmaps extension are:
303
304    Byte  0 -  3:  nb_bitmaps
305                   The number of bitmaps contained in the image. Must be
306                   greater than or equal to 1.
307
308                   Note: Qemu currently only supports up to 65535 bitmaps per
309                   image.
310
311          4 -  7:  Reserved, must be zero.
312
313          8 - 15:  bitmap_directory_size
314                   Size of the bitmap directory in bytes. It is the cumulative
315                   size of all (nb_bitmaps) bitmap directory entries.
316
317         16 - 23:  bitmap_directory_offset
318                   Offset into the image file at which the bitmap directory
319                   starts. Must be aligned to a cluster boundary.
320
321== Full disk encryption header pointer ==
322
323The full disk encryption header must be present if, and only if, the
324'crypt_method' header requires metadata. Currently this is only true
325of the 'LUKS' crypt method. The header extension must be absent for
326other methods.
327
328This header provides the offset at which the crypt method can store
329its additional data, as well as the length of such data.
330
331    Byte  0 -  7:   Offset into the image file at which the encryption
332                    header starts in bytes. Must be aligned to a cluster
333                    boundary.
334    Byte  8 - 15:   Length of the written encryption header in bytes.
335                    Note actual space allocated in the qcow2 file may
336                    be larger than this value, since it will be rounded
337                    to the nearest multiple of the cluster size. Any
338                    unused bytes in the allocated space will be initialized
339                    to 0.
340
341For the LUKS crypt method, the encryption header works as follows.
342
343The first 592 bytes of the header clusters will contain the LUKS
344partition header. This is then followed by the key material data areas.
345The size of the key material data areas is determined by the number of
346stripes in the key slot and key size. Refer to the LUKS format
347specification ('docs/on-disk-format.pdf' in the cryptsetup source
348package) for details of the LUKS partition header format.
349
350In the LUKS partition header, the "payload-offset" field will be
351calculated as normal for the LUKS spec. ie the size of the LUKS
352header, plus key material regions, plus padding, relative to the
353start of the LUKS header. This offset value is not required to be
354qcow2 cluster aligned. Its value is currently never used in the
355context of qcow2, since the qcow2 file format itself defines where
356the real payload offset is, but none the less a valid payload offset
357should always be present.
358
359In the LUKS key slots header, the "key-material-offset" is relative
360to the start of the LUKS header clusters in the qcow2 container,
361not the start of the qcow2 file.
362
363Logically the layout looks like
364
365  +-----------------------------+
366  | QCow2 header                |
367  | QCow2 header extension X    |
368  | QCow2 header extension FDE  |
369  | QCow2 header extension ...  |
370  | QCow2 header extension Z    |
371  +-----------------------------+
372  | ....other QCow2 tables....  |
373  .                             .
374  .                             .
375  +-----------------------------+
376  | +-------------------------+ |
377  | | LUKS partition header   | |
378  | +-------------------------+ |
379  | | LUKS key material 1     | |
380  | +-------------------------+ |
381  | | LUKS key material 2     | |
382  | +-------------------------+ |
383  | | LUKS key material ...   | |
384  | +-------------------------+ |
385  | | LUKS key material 8     | |
386  | +-------------------------+ |
387  +-----------------------------+
388  | QCow2 cluster payload       |
389  .                             .
390  .                             .
391  .                             .
392  |                             |
393  +-----------------------------+
394
395== Data encryption ==
396
397When an encryption method is requested in the header, the image payload
398data must be encrypted/decrypted on every write/read. The image headers
399and metadata are never encrypted.
400
401The algorithms used for encryption vary depending on the method
402
403 - AES:
404
405   The AES cipher, in CBC mode, with 256 bit keys.
406
407   Initialization vectors generated using plain64 method, with
408   the virtual disk sector as the input tweak.
409
410   This format is no longer supported in QEMU system emulators, due
411   to a number of design flaws affecting its security. It is only
412   supported in the command line tools for the sake of back compatibility
413   and data liberation.
414
415 - LUKS:
416
417   The algorithms are specified in the LUKS header.
418
419   Initialization vectors generated using the method specified
420   in the LUKS header, with the physical disk sector as the
421   input tweak.
422
423== Host cluster management ==
424
425qcow2 manages the allocation of host clusters by maintaining a reference count
426for each host cluster. A refcount of 0 means that the cluster is free, 1 means
427that it is used, and >= 2 means that it is used and any write access must
428perform a COW (copy on write) operation.
429
430The refcounts are managed in a two-level table. The first level is called
431refcount table and has a variable size (which is stored in the header). The
432refcount table can cover multiple clusters, however it needs to be contiguous
433in the image file.
434
435It contains pointers to the second level structures which are called refcount
436blocks and are exactly one cluster in size.
437
438Although a large enough refcount table can reserve clusters past 64 PB
439(56 bits) (assuming the underlying protocol can even be sized that
440large), note that some qcow2 metadata such as L1/L2 tables must point
441to clusters prior to that point.
442
443Note: qemu has an implementation limit of 8 MB as the maximum refcount
444table size.  With a 2 MB cluster size and a default refcount_order of
4454, it is unable to reference host resources beyond 2 EB (61 bits); in
446the worst case, with a 512 cluster size and refcount_order of 6, it is
447unable to access beyond 32 GB (35 bits).
448
449Given an offset into the image file, the refcount of its cluster can be
450obtained as follows:
451
452    refcount_block_entries = (cluster_size * 8 / refcount_bits)
453
454    refcount_block_index = (offset / cluster_size) % refcount_block_entries
455    refcount_table_index = (offset / cluster_size) / refcount_block_entries
456
457    refcount_block = load_cluster(refcount_table[refcount_table_index]);
458    return refcount_block[refcount_block_index];
459
460Refcount table entry:
461
462    Bit  0 -  8:    Reserved (set to 0)
463
464         9 - 63:    Bits 9-63 of the offset into the image file at which the
465                    refcount block starts. Must be aligned to a cluster
466                    boundary.
467
468                    If this is 0, the corresponding refcount block has not yet
469                    been allocated. All refcounts managed by this refcount block
470                    are 0.
471
472Refcount block entry (x = refcount_bits - 1):
473
474    Bit  0 -  x:    Reference count of the cluster. If refcount_bits implies a
475                    sub-byte width, note that bit 0 means the least significant
476                    bit in this context.
477
478
479== Cluster mapping ==
480
481Just as for refcounts, qcow2 uses a two-level structure for the mapping of
482guest clusters to host clusters. They are called L1 and L2 table.
483
484The L1 table has a variable size (stored in the header) and may use multiple
485clusters, however it must be contiguous in the image file. L2 tables are
486exactly one cluster in size.
487
488The L1 and L2 tables have implications on the maximum virtual file
489size; for a given L1 table size, a larger cluster size is required for
490the guest to have access to more space.  Furthermore, a virtual
491cluster must currently map to a host offset below 64 PB (56 bits)
492(although this limit could be relaxed by putting reserved bits into
493use).  Additionally, as cluster size increases, the maximum host
494offset for a compressed cluster is reduced (a 2M cluster size requires
495compressed clusters to reside below 512 TB (49 bits), and this limit
496cannot be relaxed without an incompatible layout change).
497
498Given an offset into the virtual disk, the offset into the image file can be
499obtained as follows:
500
501    l2_entries = (cluster_size / sizeof(uint64_t))
502
503    l2_index = (offset / cluster_size) % l2_entries
504    l1_index = (offset / cluster_size) / l2_entries
505
506    l2_table = load_cluster(l1_table[l1_index]);
507    cluster_offset = l2_table[l2_index];
508
509    return cluster_offset + (offset % cluster_size)
510
511L1 table entry:
512
513    Bit  0 -  8:    Reserved (set to 0)
514
515         9 - 55:    Bits 9-55 of the offset into the image file at which the L2
516                    table starts. Must be aligned to a cluster boundary. If the
517                    offset is 0, the L2 table and all clusters described by this
518                    L2 table are unallocated.
519
520        56 - 62:    Reserved (set to 0)
521
522             63:    0 for an L2 table that is unused or requires COW, 1 if its
523                    refcount is exactly one. This information is only accurate
524                    in the active L1 table.
525
526L2 table entry:
527
528    Bit  0 -  61:   Cluster descriptor
529
530              62:   0 for standard clusters
531                    1 for compressed clusters
532
533              63:   0 for clusters that are unused, compressed or require COW.
534                    1 for standard clusters whose refcount is exactly one.
535                    This information is only accurate in L2 tables
536                    that are reachable from the active L1 table.
537
538                    With external data files, all guest clusters have an
539                    implicit refcount of 1 (because of the fixed host = guest
540                    mapping for guest cluster offsets), so this bit should be 1
541                    for all allocated clusters.
542
543Standard Cluster Descriptor:
544
545    Bit       0:    If set to 1, the cluster reads as all zeros. The host
546                    cluster offset can be used to describe a preallocation,
547                    but it won't be used for reading data from this cluster,
548                    nor is data read from the backing file if the cluster is
549                    unallocated.
550
551                    With version 2, this is always 0.
552
553         1 -  8:    Reserved (set to 0)
554
555         9 - 55:    Bits 9-55 of host cluster offset. Must be aligned to a
556                    cluster boundary. If the offset is 0 and bit 63 is clear,
557                    the cluster is unallocated. The offset may only be 0 with
558                    bit 63 set (indicating a host cluster offset of 0) when an
559                    external data file is used.
560
561        56 - 61:    Reserved (set to 0)
562
563
564Compressed Clusters Descriptor (x = 62 - (cluster_bits - 8)):
565
566    Bit  0 - x-1:   Host cluster offset. This is usually _not_ aligned to a
567                    cluster or sector boundary!  If cluster_bits is
568                    small enough that this field includes bits beyond
569                    55, those upper bits must be set to 0.
570
571         x - 61:    Number of additional 512-byte sectors used for the
572                    compressed data, beyond the sector containing the offset
573                    in the previous field. Some of these sectors may reside
574                    in the next contiguous host cluster.
575
576                    Note that the compressed data does not necessarily occupy
577                    all of the bytes in the final sector; rather, decompression
578                    stops when it has produced a cluster of data.
579
580                    Another compressed cluster may map to the tail of the final
581                    sector used by this compressed cluster.
582
583If a cluster is unallocated, read requests shall read the data from the backing
584file (except if bit 0 in the Standard Cluster Descriptor is set). If there is
585no backing file or the backing file is smaller than the image, they shall read
586zeros for all parts that are not covered by the backing file.
587
588
589== Snapshots ==
590
591qcow2 supports internal snapshots. Their basic principle of operation is to
592switch the active L1 table, so that a different set of host clusters are
593exposed to the guest.
594
595When creating a snapshot, the L1 table should be copied and the refcount of all
596L2 tables and clusters reachable from this L1 table must be increased, so that
597a write causes a COW and isn't visible in other snapshots.
598
599When loading a snapshot, bit 63 of all entries in the new active L1 table and
600all L2 tables referenced by it must be reconstructed from the refcount table
601as it doesn't need to be accurate in inactive L1 tables.
602
603A directory of all snapshots is stored in the snapshot table, a contiguous area
604in the image file, whose starting offset and length are given by the header
605fields snapshots_offset and nb_snapshots. The entries of the snapshot table
606have variable length, depending on the length of ID, name and extra data.
607
608Snapshot table entry:
609
610    Byte 0 -  7:    Offset into the image file at which the L1 table for the
611                    snapshot starts. Must be aligned to a cluster boundary.
612
613         8 - 11:    Number of entries in the L1 table of the snapshots
614
615        12 - 13:    Length of the unique ID string describing the snapshot
616
617        14 - 15:    Length of the name of the snapshot
618
619        16 - 19:    Time at which the snapshot was taken in seconds since the
620                    Epoch
621
622        20 - 23:    Subsecond part of the time at which the snapshot was taken
623                    in nanoseconds
624
625        24 - 31:    Time that the guest was running until the snapshot was
626                    taken in nanoseconds
627
628        32 - 35:    Size of the VM state in bytes. 0 if no VM state is saved.
629                    If there is VM state, it starts at the first cluster
630                    described by first L1 table entry that doesn't describe a
631                    regular guest cluster (i.e. VM state is stored like guest
632                    disk content, except that it is stored at offsets that are
633                    larger than the virtual disk presented to the guest)
634
635        36 - 39:    Size of extra data in the table entry (used for future
636                    extensions of the format)
637
638        variable:   Extra data for future extensions. Unknown fields must be
639                    ignored. Currently defined are (offset relative to snapshot
640                    table entry):
641
642                    Byte 40 - 47:   Size of the VM state in bytes. 0 if no VM
643                                    state is saved. If this field is present,
644                                    the 32-bit value in bytes 32-35 is ignored.
645
646                    Byte 48 - 55:   Virtual disk size of the snapshot in bytes
647
648                    Version 3 images must include extra data at least up to
649                    byte 55.
650
651        variable:   Unique ID string for the snapshot (not null terminated)
652
653        variable:   Name of the snapshot (not null terminated)
654
655        variable:   Padding to round up the snapshot table entry size to the
656                    next multiple of 8.
657
658
659== Bitmaps ==
660
661As mentioned above, the bitmaps extension provides the ability to store bitmaps
662related to a virtual disk. This section describes how these bitmaps are stored.
663
664All stored bitmaps are related to the virtual disk stored in the same image, so
665each bitmap size is equal to the virtual disk size.
666
667Each bit of the bitmap is responsible for strictly defined range of the virtual
668disk. For bit number bit_nr the corresponding range (in bytes) will be:
669
670    [bit_nr * bitmap_granularity .. (bit_nr + 1) * bitmap_granularity - 1]
671
672Granularity is a property of the concrete bitmap, see below.
673
674
675=== Bitmap directory ===
676
677Each bitmap saved in the image is described in a bitmap directory entry. The
678bitmap directory is a contiguous area in the image file, whose starting offset
679and length are given by the header extension fields bitmap_directory_offset and
680bitmap_directory_size. The entries of the bitmap directory have variable
681length, depending on the lengths of the bitmap name and extra data.
682
683Structure of a bitmap directory entry:
684
685    Byte 0 -  7:    bitmap_table_offset
686                    Offset into the image file at which the bitmap table
687                    (described below) for the bitmap starts. Must be aligned to
688                    a cluster boundary.
689
690         8 - 11:    bitmap_table_size
691                    Number of entries in the bitmap table of the bitmap.
692
693        12 - 15:    flags
694                    Bit
695                      0: in_use
696                         The bitmap was not saved correctly and may be
697                         inconsistent. Although the bitmap metadata is still
698                         well-formed from a qcow2 perspective, the metadata
699                         (such as the auto flag or bitmap size) or data
700                         contents may be outdated.
701
702                      1: auto
703                         The bitmap must reflect all changes of the virtual
704                         disk by any application that would write to this qcow2
705                         file (including writes, snapshot switching, etc.). The
706                         type of this bitmap must be 'dirty tracking bitmap'.
707
708                      2: extra_data_compatible
709                         This flags is meaningful when the extra data is
710                         unknown to the software (currently any extra data is
711                         unknown to Qemu).
712                         If it is set, the bitmap may be used as expected, extra
713                         data must be left as is.
714                         If it is not set, the bitmap must not be used, but
715                         both it and its extra data be left as is.
716
717                    Bits 3 - 31 are reserved and must be 0.
718
719             16:    type
720                    This field describes the sort of the bitmap.
721                    Values:
722                      1: Dirty tracking bitmap
723
724                    Values 0, 2 - 255 are reserved.
725
726             17:    granularity_bits
727                    Granularity bits. Valid values: 0 - 63.
728
729                    Note: Qemu currently supports only values 9 - 31.
730
731                    Granularity is calculated as
732                        granularity = 1 << granularity_bits
733
734                    A bitmap's granularity is how many bytes of the image
735                    accounts for one bit of the bitmap.
736
737        18 - 19:    name_size
738                    Size of the bitmap name. Must be non-zero.
739
740                    Note: Qemu currently doesn't support values greater than
741                    1023.
742
743        20 - 23:    extra_data_size
744                    Size of type-specific extra data.
745
746                    For now, as no extra data is defined, extra_data_size is
747                    reserved and should be zero. If it is non-zero the
748                    behavior is defined by extra_data_compatible flag.
749
750        variable:   extra_data
751                    Extra data for the bitmap, occupying extra_data_size bytes.
752                    Extra data must never contain references to clusters or in
753                    some other way allocate additional clusters.
754
755        variable:   name
756                    The name of the bitmap (not null terminated), occupying
757                    name_size bytes. Must be unique among all bitmap names
758                    within the bitmaps extension.
759
760        variable:   Padding to round up the bitmap directory entry size to the
761                    next multiple of 8. All bytes of the padding must be zero.
762
763
764=== Bitmap table ===
765
766Each bitmap is stored using a one-level structure (as opposed to two-level
767structures like for refcounts and guest clusters mapping) for the mapping of
768bitmap data to host clusters. This structure is called the bitmap table.
769
770Each bitmap table has a variable size (stored in the bitmap directory entry)
771and may use multiple clusters, however, it must be contiguous in the image
772file.
773
774Structure of a bitmap table entry:
775
776    Bit       0:    Reserved and must be zero if bits 9 - 55 are non-zero.
777                    If bits 9 - 55 are zero:
778                      0: Cluster should be read as all zeros.
779                      1: Cluster should be read as all ones.
780
781         1 -  8:    Reserved and must be zero.
782
783         9 - 55:    Bits 9 - 55 of the host cluster offset. Must be aligned to
784                    a cluster boundary. If the offset is 0, the cluster is
785                    unallocated; in that case, bit 0 determines how this
786                    cluster should be treated during reads.
787
788        56 - 63:    Reserved and must be zero.
789
790
791=== Bitmap data ===
792
793As noted above, bitmap data is stored in separate clusters, described by the
794bitmap table. Given an offset (in bytes) into the bitmap data, the offset into
795the image file can be obtained as follows:
796
797    image_offset(bitmap_data_offset) =
798        bitmap_table[bitmap_data_offset / cluster_size] +
799            (bitmap_data_offset % cluster_size)
800
801This offset is not defined if bits 9 - 55 of bitmap table entry are zero (see
802above).
803
804Given an offset byte_nr into the virtual disk and the bitmap's granularity, the
805bit offset into the image file to the corresponding bit of the bitmap can be
806calculated like this:
807
808    bit_offset(byte_nr) =
809        image_offset(byte_nr / granularity / 8) * 8 +
810            (byte_nr / granularity) % 8
811
812If the size of the bitmap data is not a multiple of the cluster size then the
813last cluster of the bitmap data contains some unused tail bits. These bits must
814be zero.
815
816
817=== Dirty tracking bitmaps ===
818
819Bitmaps with 'type' field equal to one are dirty tracking bitmaps.
820
821When the virtual disk is in use dirty tracking bitmap may be 'enabled' or
822'disabled'. While the bitmap is 'enabled', all writes to the virtual disk
823should be reflected in the bitmap. A set bit in the bitmap means that the
824corresponding range of the virtual disk (see above) was written to while the
825bitmap was 'enabled'. An unset bit means that this range was not written to.
826
827The software doesn't have to sync the bitmap in the image file with its
828representation in RAM after each write or metadata change. Flag 'in_use'
829should be set while the bitmap is not synced.
830
831In the image file the 'enabled' state is reflected by the 'auto' flag. If this
832flag is set, the software must consider the bitmap as 'enabled' and start
833tracking virtual disk changes to this bitmap from the first write to the
834virtual disk. If this flag is not set then the bitmap is disabled.
835