xref: /openbmc/qemu/docs/interop/qcow2.txt (revision 89aafcf2)
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-deflate
218                    compression type). Otherwise, this field must not be present
219                    or must be zero (which means deflate).
220
221                    Available compression type values:
222                        0: deflate <https://www.ietf.org/rfc/rfc1951.txt>
223                        1: zstd <http://github.com/facebook/zstd>
224
225                    The deflate compression type is called "zlib"
226                    <https://www.zlib.net/> in QEMU. However, clusters with the
227                    deflate compression type do not have zlib headers.
228
229
230=== Header padding ===
231
232@header_length must be a multiple of 8, which means that if the end of the last
233additional field is not aligned, some padding is needed. This padding must be
234zeroed, so that if some existing (or future) additional field will fall into
235the padding, it will be interpreted accordingly to point [3.] of the previous
236paragraph, i.e.  in the same manner as when this field is not present.
237
238
239=== Header extensions ===
240
241Directly after the image header, optional sections called header extensions can
242be stored. Each extension has a structure like the following:
243
244    Byte  0 -  3:   Header extension type:
245                        0x00000000 - End of the header extension area
246                        0xe2792aca - Backing file format name string
247                        0x6803f857 - Feature name table
248                        0x23852875 - Bitmaps extension
249                        0x0537be77 - Full disk encryption header pointer
250                        0x44415441 - External data file name string
251                        other      - Unknown header extension, can be safely
252                                     ignored
253
254          4 -  7:   Length of the header extension data
255
256          8 -  n:   Header extension data
257
258          n -  m:   Padding to round up the header extension size to the next
259                    multiple of 8.
260
261Unless stated otherwise, each header extension type shall appear at most once
262in the same image.
263
264If the image has a backing file then the backing file name should be stored in
265the remaining space between the end of the header extension area and the end of
266the first cluster. It is not allowed to store other data here, so that an
267implementation can safely modify the header and add extensions without harming
268data of compatible features that it doesn't support. Compatible features that
269need space for additional data can use a header extension.
270
271
272== String header extensions ==
273
274Some header extensions (such as the backing file format name and the external
275data file name) are just a single string. In this case, the header extension
276length is the string length and the string is not '\0' terminated. (The header
277extension padding can make it look like a string is '\0' terminated, but
278neither is padding always necessary nor is there a guarantee that zero bytes
279are used for padding.)
280
281
282== Feature name table ==
283
284The feature name table is an optional header extension that contains the name
285for features used by the image. It can be used by applications that don't know
286the respective feature (e.g. because the feature was introduced only later) to
287display a useful error message.
288
289The number of entries in the feature name table is determined by the length of
290the header extension data. Each entry look like this:
291
292    Byte       0:   Type of feature (select feature bitmap)
293                        0: Incompatible feature
294                        1: Compatible feature
295                        2: Autoclear feature
296
297               1:   Bit number within the selected feature bitmap (valid
298                    values: 0-63)
299
300          2 - 47:   Feature name (padded with zeros, but not necessarily null
301                    terminated if it has full length)
302
303
304== Bitmaps extension ==
305
306The bitmaps extension is an optional header extension. It provides the ability
307to store bitmaps related to a virtual disk. For now, there is only one bitmap
308type: the dirty tracking bitmap, which tracks virtual disk changes from some
309point in time.
310
311The data of the extension should be considered consistent only if the
312corresponding auto-clear feature bit is set, see autoclear_features above.
313
314The fields of the bitmaps extension are:
315
316    Byte  0 -  3:  nb_bitmaps
317                   The number of bitmaps contained in the image. Must be
318                   greater than or equal to 1.
319
320                   Note: QEMU currently only supports up to 65535 bitmaps per
321                   image.
322
323          4 -  7:  Reserved, must be zero.
324
325          8 - 15:  bitmap_directory_size
326                   Size of the bitmap directory in bytes. It is the cumulative
327                   size of all (nb_bitmaps) bitmap directory entries.
328
329         16 - 23:  bitmap_directory_offset
330                   Offset into the image file at which the bitmap directory
331                   starts. Must be aligned to a cluster boundary.
332
333== Full disk encryption header pointer ==
334
335The full disk encryption header must be present if, and only if, the
336'crypt_method' header requires metadata. Currently this is only true
337of the 'LUKS' crypt method. The header extension must be absent for
338other methods.
339
340This header provides the offset at which the crypt method can store
341its additional data, as well as the length of such data.
342
343    Byte  0 -  7:   Offset into the image file at which the encryption
344                    header starts in bytes. Must be aligned to a cluster
345                    boundary.
346    Byte  8 - 15:   Length of the written encryption header in bytes.
347                    Note actual space allocated in the qcow2 file may
348                    be larger than this value, since it will be rounded
349                    to the nearest multiple of the cluster size. Any
350                    unused bytes in the allocated space will be initialized
351                    to 0.
352
353For the LUKS crypt method, the encryption header works as follows.
354
355The first 592 bytes of the header clusters will contain the LUKS
356partition header. This is then followed by the key material data areas.
357The size of the key material data areas is determined by the number of
358stripes in the key slot and key size. Refer to the LUKS format
359specification ('docs/on-disk-format.pdf' in the cryptsetup source
360package) for details of the LUKS partition header format.
361
362In the LUKS partition header, the "payload-offset" field will be
363calculated as normal for the LUKS spec. ie the size of the LUKS
364header, plus key material regions, plus padding, relative to the
365start of the LUKS header. This offset value is not required to be
366qcow2 cluster aligned. Its value is currently never used in the
367context of qcow2, since the qcow2 file format itself defines where
368the real payload offset is, but none the less a valid payload offset
369should always be present.
370
371In the LUKS key slots header, the "key-material-offset" is relative
372to the start of the LUKS header clusters in the qcow2 container,
373not the start of the qcow2 file.
374
375Logically the layout looks like
376
377  +-----------------------------+
378  | QCow2 header                |
379  | QCow2 header extension X    |
380  | QCow2 header extension FDE  |
381  | QCow2 header extension ...  |
382  | QCow2 header extension Z    |
383  +-----------------------------+
384  | ....other QCow2 tables....  |
385  .                             .
386  .                             .
387  +-----------------------------+
388  | +-------------------------+ |
389  | | LUKS partition header   | |
390  | +-------------------------+ |
391  | | LUKS key material 1     | |
392  | +-------------------------+ |
393  | | LUKS key material 2     | |
394  | +-------------------------+ |
395  | | LUKS key material ...   | |
396  | +-------------------------+ |
397  | | LUKS key material 8     | |
398  | +-------------------------+ |
399  +-----------------------------+
400  | QCow2 cluster payload       |
401  .                             .
402  .                             .
403  .                             .
404  |                             |
405  +-----------------------------+
406
407== Data encryption ==
408
409When an encryption method is requested in the header, the image payload
410data must be encrypted/decrypted on every write/read. The image headers
411and metadata are never encrypted.
412
413The algorithms used for encryption vary depending on the method
414
415 - AES:
416
417   The AES cipher, in CBC mode, with 256 bit keys.
418
419   Initialization vectors generated using plain64 method, with
420   the virtual disk sector as the input tweak.
421
422   This format is no longer supported in QEMU system emulators, due
423   to a number of design flaws affecting its security. It is only
424   supported in the command line tools for the sake of back compatibility
425   and data liberation.
426
427 - LUKS:
428
429   The algorithms are specified in the LUKS header.
430
431   Initialization vectors generated using the method specified
432   in the LUKS header, with the physical disk sector as the
433   input tweak.
434
435== Host cluster management ==
436
437qcow2 manages the allocation of host clusters by maintaining a reference count
438for each host cluster. A refcount of 0 means that the cluster is free, 1 means
439that it is used, and >= 2 means that it is used and any write access must
440perform a COW (copy on write) operation.
441
442The refcounts are managed in a two-level table. The first level is called
443refcount table and has a variable size (which is stored in the header). The
444refcount table can cover multiple clusters, however it needs to be contiguous
445in the image file.
446
447It contains pointers to the second level structures which are called refcount
448blocks and are exactly one cluster in size.
449
450Although a large enough refcount table can reserve clusters past 64 PB
451(56 bits) (assuming the underlying protocol can even be sized that
452large), note that some qcow2 metadata such as L1/L2 tables must point
453to clusters prior to that point.
454
455Note: qemu has an implementation limit of 8 MB as the maximum refcount
456table size.  With a 2 MB cluster size and a default refcount_order of
4574, it is unable to reference host resources beyond 2 EB (61 bits); in
458the worst case, with a 512 cluster size and refcount_order of 6, it is
459unable to access beyond 32 GB (35 bits).
460
461Given an offset into the image file, the refcount of its cluster can be
462obtained as follows:
463
464    refcount_block_entries = (cluster_size * 8 / refcount_bits)
465
466    refcount_block_index = (offset / cluster_size) % refcount_block_entries
467    refcount_table_index = (offset / cluster_size) / refcount_block_entries
468
469    refcount_block = load_cluster(refcount_table[refcount_table_index]);
470    return refcount_block[refcount_block_index];
471
472Refcount table entry:
473
474    Bit  0 -  8:    Reserved (set to 0)
475
476         9 - 63:    Bits 9-63 of the offset into the image file at which the
477                    refcount block starts. Must be aligned to a cluster
478                    boundary.
479
480                    If this is 0, the corresponding refcount block has not yet
481                    been allocated. All refcounts managed by this refcount block
482                    are 0.
483
484Refcount block entry (x = refcount_bits - 1):
485
486    Bit  0 -  x:    Reference count of the cluster. If refcount_bits implies a
487                    sub-byte width, note that bit 0 means the least significant
488                    bit in this context.
489
490
491== Cluster mapping ==
492
493Just as for refcounts, qcow2 uses a two-level structure for the mapping of
494guest clusters to host clusters. They are called L1 and L2 table.
495
496The L1 table has a variable size (stored in the header) and may use multiple
497clusters, however it must be contiguous in the image file. L2 tables are
498exactly one cluster in size.
499
500The L1 and L2 tables have implications on the maximum virtual file
501size; for a given L1 table size, a larger cluster size is required for
502the guest to have access to more space.  Furthermore, a virtual
503cluster must currently map to a host offset below 64 PB (56 bits)
504(although this limit could be relaxed by putting reserved bits into
505use).  Additionally, as cluster size increases, the maximum host
506offset for a compressed cluster is reduced (a 2M cluster size requires
507compressed clusters to reside below 512 TB (49 bits), and this limit
508cannot be relaxed without an incompatible layout change).
509
510Given an offset into the virtual disk, the offset into the image file can be
511obtained as follows:
512
513    l2_entries = (cluster_size / sizeof(uint64_t))        [*]
514
515    l2_index = (offset / cluster_size) % l2_entries
516    l1_index = (offset / cluster_size) / l2_entries
517
518    l2_table = load_cluster(l1_table[l1_index]);
519    cluster_offset = l2_table[l2_index];
520
521    return cluster_offset + (offset % cluster_size)
522
523    [*] this changes if Extended L2 Entries are enabled, see next section
524
525L1 table entry:
526
527    Bit  0 -  8:    Reserved (set to 0)
528
529         9 - 55:    Bits 9-55 of the offset into the image file at which the L2
530                    table starts. Must be aligned to a cluster boundary. If the
531                    offset is 0, the L2 table and all clusters described by this
532                    L2 table are unallocated.
533
534        56 - 62:    Reserved (set to 0)
535
536             63:    0 for an L2 table that is unused or requires COW, 1 if its
537                    refcount is exactly one. This information is only accurate
538                    in the active L1 table.
539
540L2 table entry:
541
542    Bit  0 -  61:   Cluster descriptor
543
544              62:   0 for standard clusters
545                    1 for compressed clusters
546
547              63:   0 for clusters that are unused, compressed or require COW.
548                    1 for standard clusters whose refcount is exactly one.
549                    This information is only accurate in L2 tables
550                    that are reachable from the active L1 table.
551
552                    With external data files, all guest clusters have an
553                    implicit refcount of 1 (because of the fixed host = guest
554                    mapping for guest cluster offsets), so this bit should be 1
555                    for all allocated clusters.
556
557Standard Cluster Descriptor:
558
559    Bit       0:    If set to 1, the cluster reads as all zeros. The host
560                    cluster offset can be used to describe a preallocation,
561                    but it won't be used for reading data from this cluster,
562                    nor is data read from the backing file if the cluster is
563                    unallocated.
564
565                    With version 2 or with extended L2 entries (see the next
566                    section), this is always 0.
567
568         1 -  8:    Reserved (set to 0)
569
570         9 - 55:    Bits 9-55 of host cluster offset. Must be aligned to a
571                    cluster boundary. If the offset is 0 and bit 63 is clear,
572                    the cluster is unallocated. The offset may only be 0 with
573                    bit 63 set (indicating a host cluster offset of 0) when an
574                    external data file is used.
575
576        56 - 61:    Reserved (set to 0)
577
578
579Compressed Clusters Descriptor (x = 62 - (cluster_bits - 8)):
580
581    Bit  0 - x-1:   Host cluster offset. This is usually _not_ aligned to a
582                    cluster or sector boundary!  If cluster_bits is
583                    small enough that this field includes bits beyond
584                    55, those upper bits must be set to 0.
585
586         x - 61:    Number of additional 512-byte sectors used for the
587                    compressed data, beyond the sector containing the offset
588                    in the previous field. Some of these sectors may reside
589                    in the next contiguous host cluster.
590
591                    Note that the compressed data does not necessarily occupy
592                    all of the bytes in the final sector; rather, decompression
593                    stops when it has produced a cluster of data.
594
595                    Another compressed cluster may map to the tail of the final
596                    sector used by this compressed cluster.
597
598If a cluster is unallocated, read requests shall read the data from the backing
599file (except if bit 0 in the Standard Cluster Descriptor is set). If there is
600no backing file or the backing file is smaller than the image, they shall read
601zeros for all parts that are not covered by the backing file.
602
603== Extended L2 Entries ==
604
605An image uses Extended L2 Entries if bit 4 is set on the incompatible_features
606field of the header.
607
608In these images standard data clusters are divided into 32 subclusters of the
609same size. They are contiguous and start from the beginning of the cluster.
610Subclusters can be allocated independently and the L2 entry contains information
611indicating the status of each one of them. Compressed data clusters don't have
612subclusters so they are treated the same as in images without this feature.
613
614The size of an extended L2 entry is 128 bits so the number of entries per table
615is calculated using this formula:
616
617    l2_entries = (cluster_size / (2 * sizeof(uint64_t)))
618
619The first 64 bits have the same format as the standard L2 table entry described
620in the previous section, with the exception of bit 0 of the standard cluster
621descriptor.
622
623The last 64 bits contain a subcluster allocation bitmap with this format:
624
625Subcluster Allocation Bitmap (for standard clusters):
626
627    Bit  0 - 31:    Allocation status (one bit per subcluster)
628
629                    1: the subcluster is allocated. In this case the
630                       host cluster offset field must contain a valid
631                       offset.
632                    0: the subcluster is not allocated. In this case
633                       read requests shall go to the backing file or
634                       return zeros if there is no backing file data.
635
636                    Bits are assigned starting from the least significant
637                    one (i.e. bit x is used for subcluster x).
638
639        32 - 63     Subcluster reads as zeros (one bit per subcluster)
640
641                    1: the subcluster reads as zeros. In this case the
642                       allocation status bit must be unset. The host
643                       cluster offset field may or may not be set.
644                    0: no effect.
645
646                    Bits are assigned starting from the least significant
647                    one (i.e. bit x is used for subcluster x - 32).
648
649Subcluster Allocation Bitmap (for compressed clusters):
650
651    Bit  0 - 63:    Reserved (set to 0)
652                    Compressed clusters don't have subclusters,
653                    so this field is not used.
654
655== Snapshots ==
656
657qcow2 supports internal snapshots. Their basic principle of operation is to
658switch the active L1 table, so that a different set of host clusters are
659exposed to the guest.
660
661When creating a snapshot, the L1 table should be copied and the refcount of all
662L2 tables and clusters reachable from this L1 table must be increased, so that
663a write causes a COW and isn't visible in other snapshots.
664
665When loading a snapshot, bit 63 of all entries in the new active L1 table and
666all L2 tables referenced by it must be reconstructed from the refcount table
667as it doesn't need to be accurate in inactive L1 tables.
668
669A directory of all snapshots is stored in the snapshot table, a contiguous area
670in the image file, whose starting offset and length are given by the header
671fields snapshots_offset and nb_snapshots. The entries of the snapshot table
672have variable length, depending on the length of ID, name and extra data.
673
674Snapshot table entry:
675
676    Byte 0 -  7:    Offset into the image file at which the L1 table for the
677                    snapshot starts. Must be aligned to a cluster boundary.
678
679         8 - 11:    Number of entries in the L1 table of the snapshots
680
681        12 - 13:    Length of the unique ID string describing the snapshot
682
683        14 - 15:    Length of the name of the snapshot
684
685        16 - 19:    Time at which the snapshot was taken in seconds since the
686                    Epoch
687
688        20 - 23:    Subsecond part of the time at which the snapshot was taken
689                    in nanoseconds
690
691        24 - 31:    Time that the guest was running until the snapshot was
692                    taken in nanoseconds
693
694        32 - 35:    Size of the VM state in bytes. 0 if no VM state is saved.
695                    If there is VM state, it starts at the first cluster
696                    described by first L1 table entry that doesn't describe a
697                    regular guest cluster (i.e. VM state is stored like guest
698                    disk content, except that it is stored at offsets that are
699                    larger than the virtual disk presented to the guest)
700
701        36 - 39:    Size of extra data in the table entry (used for future
702                    extensions of the format)
703
704        variable:   Extra data for future extensions. Unknown fields must be
705                    ignored. Currently defined are (offset relative to snapshot
706                    table entry):
707
708                    Byte 40 - 47:   Size of the VM state in bytes. 0 if no VM
709                                    state is saved. If this field is present,
710                                    the 32-bit value in bytes 32-35 is ignored.
711
712                    Byte 48 - 55:   Virtual disk size of the snapshot in bytes
713
714                    Byte 56 - 63:   icount value which corresponds to
715                                    the record/replay instruction count
716                                    when the snapshot was taken. Set to -1
717                                    if icount was disabled
718
719                    Version 3 images must include extra data at least up to
720                    byte 55.
721
722        variable:   Unique ID string for the snapshot (not null terminated)
723
724        variable:   Name of the snapshot (not null terminated)
725
726        variable:   Padding to round up the snapshot table entry size to the
727                    next multiple of 8.
728
729
730== Bitmaps ==
731
732As mentioned above, the bitmaps extension provides the ability to store bitmaps
733related to a virtual disk. This section describes how these bitmaps are stored.
734
735All stored bitmaps are related to the virtual disk stored in the same image, so
736each bitmap size is equal to the virtual disk size.
737
738Each bit of the bitmap is responsible for strictly defined range of the virtual
739disk. For bit number bit_nr the corresponding range (in bytes) will be:
740
741    [bit_nr * bitmap_granularity .. (bit_nr + 1) * bitmap_granularity - 1]
742
743Granularity is a property of the concrete bitmap, see below.
744
745
746=== Bitmap directory ===
747
748Each bitmap saved in the image is described in a bitmap directory entry. The
749bitmap directory is a contiguous area in the image file, whose starting offset
750and length are given by the header extension fields bitmap_directory_offset and
751bitmap_directory_size. The entries of the bitmap directory have variable
752length, depending on the lengths of the bitmap name and extra data.
753
754Structure of a bitmap directory entry:
755
756    Byte 0 -  7:    bitmap_table_offset
757                    Offset into the image file at which the bitmap table
758                    (described below) for the bitmap starts. Must be aligned to
759                    a cluster boundary.
760
761         8 - 11:    bitmap_table_size
762                    Number of entries in the bitmap table of the bitmap.
763
764        12 - 15:    flags
765                    Bit
766                      0: in_use
767                         The bitmap was not saved correctly and may be
768                         inconsistent. Although the bitmap metadata is still
769                         well-formed from a qcow2 perspective, the metadata
770                         (such as the auto flag or bitmap size) or data
771                         contents may be outdated.
772
773                      1: auto
774                         The bitmap must reflect all changes of the virtual
775                         disk by any application that would write to this qcow2
776                         file (including writes, snapshot switching, etc.). The
777                         type of this bitmap must be 'dirty tracking bitmap'.
778
779                      2: extra_data_compatible
780                         This flags is meaningful when the extra data is
781                         unknown to the software (currently any extra data is
782                         unknown to QEMU).
783                         If it is set, the bitmap may be used as expected, extra
784                         data must be left as is.
785                         If it is not set, the bitmap must not be used, but
786                         both it and its extra data be left as is.
787
788                    Bits 3 - 31 are reserved and must be 0.
789
790             16:    type
791                    This field describes the sort of the bitmap.
792                    Values:
793                      1: Dirty tracking bitmap
794
795                    Values 0, 2 - 255 are reserved.
796
797             17:    granularity_bits
798                    Granularity bits. Valid values: 0 - 63.
799
800                    Note: QEMU currently supports only values 9 - 31.
801
802                    Granularity is calculated as
803                        granularity = 1 << granularity_bits
804
805                    A bitmap's granularity is how many bytes of the image
806                    accounts for one bit of the bitmap.
807
808        18 - 19:    name_size
809                    Size of the bitmap name. Must be non-zero.
810
811                    Note: QEMU currently doesn't support values greater than
812                    1023.
813
814        20 - 23:    extra_data_size
815                    Size of type-specific extra data.
816
817                    For now, as no extra data is defined, extra_data_size is
818                    reserved and should be zero. If it is non-zero the
819                    behavior is defined by extra_data_compatible flag.
820
821        variable:   extra_data
822                    Extra data for the bitmap, occupying extra_data_size bytes.
823                    Extra data must never contain references to clusters or in
824                    some other way allocate additional clusters.
825
826        variable:   name
827                    The name of the bitmap (not null terminated), occupying
828                    name_size bytes. Must be unique among all bitmap names
829                    within the bitmaps extension.
830
831        variable:   Padding to round up the bitmap directory entry size to the
832                    next multiple of 8. All bytes of the padding must be zero.
833
834
835=== Bitmap table ===
836
837Each bitmap is stored using a one-level structure (as opposed to two-level
838structures like for refcounts and guest clusters mapping) for the mapping of
839bitmap data to host clusters. This structure is called the bitmap table.
840
841Each bitmap table has a variable size (stored in the bitmap directory entry)
842and may use multiple clusters, however, it must be contiguous in the image
843file.
844
845Structure of a bitmap table entry:
846
847    Bit       0:    Reserved and must be zero if bits 9 - 55 are non-zero.
848                    If bits 9 - 55 are zero:
849                      0: Cluster should be read as all zeros.
850                      1: Cluster should be read as all ones.
851
852         1 -  8:    Reserved and must be zero.
853
854         9 - 55:    Bits 9 - 55 of the host cluster offset. Must be aligned to
855                    a cluster boundary. If the offset is 0, the cluster is
856                    unallocated; in that case, bit 0 determines how this
857                    cluster should be treated during reads.
858
859        56 - 63:    Reserved and must be zero.
860
861
862=== Bitmap data ===
863
864As noted above, bitmap data is stored in separate clusters, described by the
865bitmap table. Given an offset (in bytes) into the bitmap data, the offset into
866the image file can be obtained as follows:
867
868    image_offset(bitmap_data_offset) =
869        bitmap_table[bitmap_data_offset / cluster_size] +
870            (bitmap_data_offset % cluster_size)
871
872This offset is not defined if bits 9 - 55 of bitmap table entry are zero (see
873above).
874
875Given an offset byte_nr into the virtual disk and the bitmap's granularity, the
876bit offset into the image file to the corresponding bit of the bitmap can be
877calculated like this:
878
879    bit_offset(byte_nr) =
880        image_offset(byte_nr / granularity / 8) * 8 +
881            (byte_nr / granularity) % 8
882
883If the size of the bitmap data is not a multiple of the cluster size then the
884last cluster of the bitmap data contains some unused tail bits. These bits must
885be zero.
886
887
888=== Dirty tracking bitmaps ===
889
890Bitmaps with 'type' field equal to one are dirty tracking bitmaps.
891
892When the virtual disk is in use dirty tracking bitmap may be 'enabled' or
893'disabled'. While the bitmap is 'enabled', all writes to the virtual disk
894should be reflected in the bitmap. A set bit in the bitmap means that the
895corresponding range of the virtual disk (see above) was written to while the
896bitmap was 'enabled'. An unset bit means that this range was not written to.
897
898The software doesn't have to sync the bitmap in the image file with its
899representation in RAM after each write or metadata change. Flag 'in_use'
900should be set while the bitmap is not synced.
901
902In the image file the 'enabled' state is reflected by the 'auto' flag. If this
903flag is set, the software must consider the bitmap as 'enabled' and start
904tracking virtual disk changes to this bitmap from the first write to the
905virtual disk. If this flag is not set then the bitmap is disabled.
906