2 * Block driver for the QCOW version 2 format
4 * Copyright (c) 2004-2006 Fabrice Bellard
6 * Permission is hereby granted, free of charge, to any person obtaining a copy
7 * of this software and associated documentation files (the "Software"), to deal
8 * in the Software without restriction, including without limitation the rights
9 * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
10 * copies of the Software, and to permit persons to whom the Software is
11 * furnished to do so, subject to the following conditions:
13 * The above copyright notice and this permission notice shall be included in
14 * all copies or substantial portions of the Software.
16 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
17 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
18 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
19 * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
20 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
21 * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
24 #include "qemu-common.h"
25 #include "block_int.h"
31 Differences with QCOW:
33 - Support for multiple incremental snapshots.
34 - Memory management by reference counts.
35 - Clusters which have a reference count of one have the bit
36 QCOW_OFLAG_COPIED to optimize write performance.
37 - Size of compressed clusters is stored in sectors to reduce bit usage
38 in the cluster offsets.
39 - Support for storing additional data (such as the VM state) in the
41 - If a backing store is used, the cluster size is not constrained
42 (could be backported to QCOW).
43 - L2 tables have always a size of one cluster.
47 //#define DEBUG_ALLOC2
49 #define QCOW_MAGIC (('Q' << 24) | ('F' << 16) | ('I' << 8) | 0xfb)
50 #define QCOW_VERSION 2
52 #define QCOW_CRYPT_NONE 0
53 #define QCOW_CRYPT_AES 1
55 /* indicate that the refcount of the referenced cluster is exactly one. */
56 #define QCOW_OFLAG_COPIED (1LL << 63)
57 /* indicate that the cluster is compressed (they never have the copied flag) */
58 #define QCOW_OFLAG_COMPRESSED (1LL << 62)
60 #define REFCOUNT_SHIFT 1 /* refcount size is 2 bytes */
63 #define offsetof(type, field) ((size_t) &((type *)0)->field)
66 typedef struct QCowHeader {
69 uint64_t backing_file_offset;
70 uint32_t backing_file_size;
71 uint32_t cluster_bits;
72 uint64_t size; /* in bytes */
73 uint32_t crypt_method;
74 uint32_t l1_size; /* XXX: save number of clusters instead ? */
75 uint64_t l1_table_offset;
76 uint64_t refcount_table_offset;
77 uint32_t refcount_table_clusters;
78 uint32_t nb_snapshots;
79 uint64_t snapshots_offset;
82 typedef struct __attribute__((packed)) QCowSnapshotHeader {
83 /* header is 8 byte aligned */
84 uint64_t l1_table_offset;
93 uint64_t vm_clock_nsec;
95 uint32_t vm_state_size;
96 uint32_t extra_data_size; /* for extension */
97 /* extra data follows */
100 } QCowSnapshotHeader;
102 #define L2_CACHE_SIZE 16
104 typedef struct QCowSnapshot {
105 uint64_t l1_table_offset;
109 uint32_t vm_state_size;
112 uint64_t vm_clock_nsec;
115 typedef struct BDRVQcowState {
116 BlockDriverState *hd;
123 int l1_vm_state_index;
126 uint64_t cluster_offset_mask;
127 uint64_t l1_table_offset;
130 uint64_t l2_cache_offsets[L2_CACHE_SIZE];
131 uint32_t l2_cache_counts[L2_CACHE_SIZE];
132 uint8_t *cluster_cache;
133 uint8_t *cluster_data;
134 uint64_t cluster_cache_offset;
136 uint64_t *refcount_table;
137 uint64_t refcount_table_offset;
138 uint32_t refcount_table_size;
139 uint64_t refcount_block_cache_offset;
140 uint16_t *refcount_block_cache;
141 int64_t free_cluster_index;
142 int64_t free_byte_offset;
144 uint32_t crypt_method; /* current crypt method, 0 if no key yet */
145 uint32_t crypt_method_header;
146 AES_KEY aes_encrypt_key;
147 AES_KEY aes_decrypt_key;
148 uint64_t snapshots_offset;
151 QCowSnapshot *snapshots;
154 static int decompress_cluster(BDRVQcowState *s, uint64_t cluster_offset);
155 static int qcow_read(BlockDriverState *bs, int64_t sector_num,
156 uint8_t *buf, int nb_sectors);
157 static int qcow_read_snapshots(BlockDriverState *bs);
158 static void qcow_free_snapshots(BlockDriverState *bs);
159 static int refcount_init(BlockDriverState *bs);
160 static void refcount_close(BlockDriverState *bs);
161 static int get_refcount(BlockDriverState *bs, int64_t cluster_index);
162 static int update_cluster_refcount(BlockDriverState *bs,
163 int64_t cluster_index,
165 static void update_refcount(BlockDriverState *bs,
166 int64_t offset, int64_t length,
168 static int64_t alloc_clusters(BlockDriverState *bs, int64_t size);
169 static int64_t alloc_bytes(BlockDriverState *bs, int size);
170 static void free_clusters(BlockDriverState *bs,
171 int64_t offset, int64_t size);
173 static void check_refcounts(BlockDriverState *bs);
176 static int qcow_probe(const uint8_t *buf, int buf_size, const char *filename)
178 const QCowHeader *cow_header = (const void *)buf;
180 if (buf_size >= sizeof(QCowHeader) &&
181 be32_to_cpu(cow_header->magic) == QCOW_MAGIC &&
182 be32_to_cpu(cow_header->version) == QCOW_VERSION)
188 static int qcow_open(BlockDriverState *bs, const char *filename, int flags)
190 BDRVQcowState *s = bs->opaque;
191 int len, i, shift, ret;
194 ret = bdrv_file_open(&s->hd, filename, flags);
197 if (bdrv_pread(s->hd, 0, &header, sizeof(header)) != sizeof(header))
199 be32_to_cpus(&header.magic);
200 be32_to_cpus(&header.version);
201 be64_to_cpus(&header.backing_file_offset);
202 be32_to_cpus(&header.backing_file_size);
203 be64_to_cpus(&header.size);
204 be32_to_cpus(&header.cluster_bits);
205 be32_to_cpus(&header.crypt_method);
206 be64_to_cpus(&header.l1_table_offset);
207 be32_to_cpus(&header.l1_size);
208 be64_to_cpus(&header.refcount_table_offset);
209 be32_to_cpus(&header.refcount_table_clusters);
210 be64_to_cpus(&header.snapshots_offset);
211 be32_to_cpus(&header.nb_snapshots);
213 if (header.magic != QCOW_MAGIC || header.version != QCOW_VERSION)
215 if (header.size <= 1 ||
216 header.cluster_bits < 9 ||
217 header.cluster_bits > 16)
219 if (header.crypt_method > QCOW_CRYPT_AES)
221 s->crypt_method_header = header.crypt_method;
222 if (s->crypt_method_header)
224 s->cluster_bits = header.cluster_bits;
225 s->cluster_size = 1 << s->cluster_bits;
226 s->cluster_sectors = 1 << (s->cluster_bits - 9);
227 s->l2_bits = s->cluster_bits - 3; /* L2 is always one cluster */
228 s->l2_size = 1 << s->l2_bits;
229 bs->total_sectors = header.size / 512;
230 s->csize_shift = (62 - (s->cluster_bits - 8));
231 s->csize_mask = (1 << (s->cluster_bits - 8)) - 1;
232 s->cluster_offset_mask = (1LL << s->csize_shift) - 1;
233 s->refcount_table_offset = header.refcount_table_offset;
234 s->refcount_table_size =
235 header.refcount_table_clusters << (s->cluster_bits - 3);
237 s->snapshots_offset = header.snapshots_offset;
238 s->nb_snapshots = header.nb_snapshots;
240 /* read the level 1 table */
241 s->l1_size = header.l1_size;
242 shift = s->cluster_bits + s->l2_bits;
243 s->l1_vm_state_index = (header.size + (1LL << shift) - 1) >> shift;
244 /* the L1 table must contain at least enough entries to put
246 if (s->l1_size < s->l1_vm_state_index)
248 s->l1_table_offset = header.l1_table_offset;
249 s->l1_table = qemu_malloc(s->l1_size * sizeof(uint64_t));
252 if (bdrv_pread(s->hd, s->l1_table_offset, s->l1_table, s->l1_size * sizeof(uint64_t)) !=
253 s->l1_size * sizeof(uint64_t))
255 for(i = 0;i < s->l1_size; i++) {
256 be64_to_cpus(&s->l1_table[i]);
259 s->l2_cache = qemu_malloc(s->l2_size * L2_CACHE_SIZE * sizeof(uint64_t));
262 s->cluster_cache = qemu_malloc(s->cluster_size);
263 if (!s->cluster_cache)
265 /* one more sector for decompressed data alignment */
266 s->cluster_data = qemu_malloc(s->cluster_size + 512);
267 if (!s->cluster_data)
269 s->cluster_cache_offset = -1;
271 if (refcount_init(bs) < 0)
274 /* read the backing file name */
275 if (header.backing_file_offset != 0) {
276 len = header.backing_file_size;
279 if (bdrv_pread(s->hd, header.backing_file_offset, bs->backing_file, len) != len)
281 bs->backing_file[len] = '\0';
283 if (qcow_read_snapshots(bs) < 0)
292 qcow_free_snapshots(bs);
294 qemu_free(s->l1_table);
295 qemu_free(s->l2_cache);
296 qemu_free(s->cluster_cache);
297 qemu_free(s->cluster_data);
302 static int qcow_set_key(BlockDriverState *bs, const char *key)
304 BDRVQcowState *s = bs->opaque;
308 memset(keybuf, 0, 16);
312 /* XXX: we could compress the chars to 7 bits to increase
314 for(i = 0;i < len;i++) {
317 s->crypt_method = s->crypt_method_header;
319 if (AES_set_encrypt_key(keybuf, 128, &s->aes_encrypt_key) != 0)
321 if (AES_set_decrypt_key(keybuf, 128, &s->aes_decrypt_key) != 0)
331 AES_encrypt(in, tmp, &s->aes_encrypt_key);
332 AES_decrypt(tmp, out, &s->aes_decrypt_key);
333 for(i = 0; i < 16; i++)
334 printf(" %02x", tmp[i]);
336 for(i = 0; i < 16; i++)
337 printf(" %02x", out[i]);
344 /* The crypt function is compatible with the linux cryptoloop
345 algorithm for < 4 GB images. NOTE: out_buf == in_buf is
347 static void encrypt_sectors(BDRVQcowState *s, int64_t sector_num,
348 uint8_t *out_buf, const uint8_t *in_buf,
349 int nb_sectors, int enc,
358 for(i = 0; i < nb_sectors; i++) {
359 ivec.ll[0] = cpu_to_le64(sector_num);
361 AES_cbc_encrypt(in_buf, out_buf, 512, key,
369 static int copy_sectors(BlockDriverState *bs, uint64_t start_sect,
370 uint64_t cluster_offset, int n_start, int n_end)
372 BDRVQcowState *s = bs->opaque;
378 ret = qcow_read(bs, start_sect + n_start, s->cluster_data, n);
381 if (s->crypt_method) {
382 encrypt_sectors(s, start_sect + n_start,
384 s->cluster_data, n, 1,
385 &s->aes_encrypt_key);
387 ret = bdrv_write(s->hd, (cluster_offset >> 9) + n_start,
394 static void l2_cache_reset(BlockDriverState *bs)
396 BDRVQcowState *s = bs->opaque;
398 memset(s->l2_cache, 0, s->l2_size * L2_CACHE_SIZE * sizeof(uint64_t));
399 memset(s->l2_cache_offsets, 0, L2_CACHE_SIZE * sizeof(uint64_t));
400 memset(s->l2_cache_counts, 0, L2_CACHE_SIZE * sizeof(uint32_t));
403 static inline int l2_cache_new_entry(BlockDriverState *bs)
405 BDRVQcowState *s = bs->opaque;
409 /* find a new entry in the least used one */
411 min_count = 0xffffffff;
412 for(i = 0; i < L2_CACHE_SIZE; i++) {
413 if (s->l2_cache_counts[i] < min_count) {
414 min_count = s->l2_cache_counts[i];
421 static int64_t align_offset(int64_t offset, int n)
423 offset = (offset + n - 1) & ~(n - 1);
427 static int grow_l1_table(BlockDriverState *bs, int min_size)
429 BDRVQcowState *s = bs->opaque;
430 int new_l1_size, new_l1_size2, ret, i;
431 uint64_t *new_l1_table;
432 uint64_t new_l1_table_offset;
436 new_l1_size = s->l1_size;
437 if (min_size <= new_l1_size)
439 while (min_size > new_l1_size) {
440 new_l1_size = (new_l1_size * 3 + 1) / 2;
443 printf("grow l1_table from %d to %d\n", s->l1_size, new_l1_size);
446 new_l1_size2 = sizeof(uint64_t) * new_l1_size;
447 new_l1_table = qemu_mallocz(new_l1_size2);
450 memcpy(new_l1_table, s->l1_table, s->l1_size * sizeof(uint64_t));
452 /* write new table (align to cluster) */
453 new_l1_table_offset = alloc_clusters(bs, new_l1_size2);
455 for(i = 0; i < s->l1_size; i++)
456 new_l1_table[i] = cpu_to_be64(new_l1_table[i]);
457 ret = bdrv_pwrite(s->hd, new_l1_table_offset, new_l1_table, new_l1_size2);
458 if (ret != new_l1_size2)
460 for(i = 0; i < s->l1_size; i++)
461 new_l1_table[i] = be64_to_cpu(new_l1_table[i]);
464 data64 = cpu_to_be64(new_l1_table_offset);
465 if (bdrv_pwrite(s->hd, offsetof(QCowHeader, l1_table_offset),
466 &data64, sizeof(data64)) != sizeof(data64))
468 data32 = cpu_to_be32(new_l1_size);
469 if (bdrv_pwrite(s->hd, offsetof(QCowHeader, l1_size),
470 &data32, sizeof(data32)) != sizeof(data32))
472 qemu_free(s->l1_table);
473 free_clusters(bs, s->l1_table_offset, s->l1_size * sizeof(uint64_t));
474 s->l1_table_offset = new_l1_table_offset;
475 s->l1_table = new_l1_table;
476 s->l1_size = new_l1_size;
479 qemu_free(s->l1_table);
486 * seek l2_offset in the l2_cache table
487 * if not found, return NULL,
489 * increments the l2 cache hit count of the entry,
490 * if counter overflow, divide by two all counters
491 * return the pointer to the l2 cache entry
495 static uint64_t *seek_l2_table(BDRVQcowState *s, uint64_t l2_offset)
499 for(i = 0; i < L2_CACHE_SIZE; i++) {
500 if (l2_offset == s->l2_cache_offsets[i]) {
501 /* increment the hit count */
502 if (++s->l2_cache_counts[i] == 0xffffffff) {
503 for(j = 0; j < L2_CACHE_SIZE; j++) {
504 s->l2_cache_counts[j] >>= 1;
507 return s->l2_cache + (i << s->l2_bits);
516 * Loads a L2 table into memory. If the table is in the cache, the cache
517 * is used; otherwise the L2 table is loaded from the image file.
519 * Returns a pointer to the L2 table on success, or NULL if the read from
520 * the image file failed.
523 static uint64_t *l2_load(BlockDriverState *bs, uint64_t l2_offset)
525 BDRVQcowState *s = bs->opaque;
529 /* seek if the table for the given offset is in the cache */
531 l2_table = seek_l2_table(s, l2_offset);
532 if (l2_table != NULL)
535 /* not found: load a new entry in the least used one */
537 min_index = l2_cache_new_entry(bs);
538 l2_table = s->l2_cache + (min_index << s->l2_bits);
539 if (bdrv_pread(s->hd, l2_offset, l2_table, s->l2_size * sizeof(uint64_t)) !=
540 s->l2_size * sizeof(uint64_t))
542 s->l2_cache_offsets[min_index] = l2_offset;
543 s->l2_cache_counts[min_index] = 1;
551 * Allocate a new l2 entry in the file. If l1_index points to an already
552 * used entry in the L2 table (i.e. we are doing a copy on write for the L2
553 * table) copy the contents of the old L2 table into the newly allocated one.
554 * Otherwise the new table is initialized with zeros.
558 static uint64_t *l2_allocate(BlockDriverState *bs, int l1_index)
560 BDRVQcowState *s = bs->opaque;
562 uint64_t old_l2_offset, tmp;
563 uint64_t *l2_table, l2_offset;
565 old_l2_offset = s->l1_table[l1_index];
567 /* allocate a new l2 entry */
569 l2_offset = alloc_clusters(bs, s->l2_size * sizeof(uint64_t));
571 /* update the L1 entry */
573 s->l1_table[l1_index] = l2_offset | QCOW_OFLAG_COPIED;
575 tmp = cpu_to_be64(l2_offset | QCOW_OFLAG_COPIED);
576 if (bdrv_pwrite(s->hd, s->l1_table_offset + l1_index * sizeof(tmp),
577 &tmp, sizeof(tmp)) != sizeof(tmp))
580 /* allocate a new entry in the l2 cache */
582 min_index = l2_cache_new_entry(bs);
583 l2_table = s->l2_cache + (min_index << s->l2_bits);
585 if (old_l2_offset == 0) {
586 /* if there was no old l2 table, clear the new table */
587 memset(l2_table, 0, s->l2_size * sizeof(uint64_t));
589 /* if there was an old l2 table, read it from the disk */
590 if (bdrv_pread(s->hd, old_l2_offset,
591 l2_table, s->l2_size * sizeof(uint64_t)) !=
592 s->l2_size * sizeof(uint64_t))
595 /* write the l2 table to the file */
596 if (bdrv_pwrite(s->hd, l2_offset,
597 l2_table, s->l2_size * sizeof(uint64_t)) !=
598 s->l2_size * sizeof(uint64_t))
601 /* update the l2 cache entry */
603 s->l2_cache_offsets[min_index] = l2_offset;
604 s->l2_cache_counts[min_index] = 1;
609 static uint64_t get_cluster_offset(BlockDriverState *bs,
610 uint64_t offset, int allocate,
612 int n_start, int n_end)
614 BDRVQcowState *s = bs->opaque;
615 int l1_index, l2_index, ret;
616 uint64_t l2_offset, *l2_table, cluster_offset, tmp;
618 /* seek the the l2 offset in the l1 table */
620 l1_index = offset >> (s->l2_bits + s->cluster_bits);
621 if (l1_index >= s->l1_size) {
622 /* outside l1 table is allowed: we grow the table if needed */
625 ret = grow_l1_table(bs, l1_index + 1);
629 l2_offset = s->l1_table[l1_index];
631 /* seek the l2 table of the given l2 offset */
634 /* the l2 table doesn't exist */
637 /* allocate a new l2 table for this offset */
638 l2_table = l2_allocate(bs, l1_index);
639 if (l2_table == NULL)
641 l2_offset = s->l1_table[l1_index] & ~QCOW_OFLAG_COPIED;
643 /* the l2 table exists */
644 if (!(l2_offset & QCOW_OFLAG_COPIED) && allocate) {
645 /* duplicate the l2 table, and free the old table */
646 free_clusters(bs, l2_offset, s->l2_size * sizeof(uint64_t));
647 l2_table = l2_allocate(bs, l1_index);
648 if (l2_table == NULL)
650 l2_offset = s->l1_table[l1_index] & ~QCOW_OFLAG_COPIED;
652 /* load the l2 table in memory */
653 l2_offset &= ~QCOW_OFLAG_COPIED;
654 l2_table = l2_load(bs, l2_offset);
655 if (l2_table == NULL)
660 /* find the cluster offset for the given disk offset */
662 l2_index = (offset >> s->cluster_bits) & (s->l2_size - 1);
663 cluster_offset = be64_to_cpu(l2_table[l2_index]);
664 if (!cluster_offset) {
665 /* cluster doesn't exist */
668 } else if (!(cluster_offset & QCOW_OFLAG_COPIED)) {
670 return cluster_offset;
671 /* free the cluster */
672 if (cluster_offset & QCOW_OFLAG_COMPRESSED) {
674 nb_csectors = ((cluster_offset >> s->csize_shift) &
676 free_clusters(bs, (cluster_offset & s->cluster_offset_mask) & ~511,
679 free_clusters(bs, cluster_offset, s->cluster_size);
682 cluster_offset &= ~QCOW_OFLAG_COPIED;
683 return cluster_offset;
687 /* allocate a new cluster */
688 cluster_offset = alloc_clusters(bs, s->cluster_size);
690 /* we must initialize the cluster content which won't be
692 if ((n_end - n_start) < s->cluster_sectors) {
695 start_sect = (offset & ~(s->cluster_size - 1)) >> 9;
696 ret = copy_sectors(bs, start_sect,
697 cluster_offset, 0, n_start);
700 ret = copy_sectors(bs, start_sect,
701 cluster_offset, n_end, s->cluster_sectors);
705 tmp = cpu_to_be64(cluster_offset | QCOW_OFLAG_COPIED);
708 cluster_offset = alloc_bytes(bs, compressed_size);
709 nb_csectors = ((cluster_offset + compressed_size - 1) >> 9) -
710 (cluster_offset >> 9);
711 cluster_offset |= QCOW_OFLAG_COMPRESSED |
712 ((uint64_t)nb_csectors << s->csize_shift);
713 /* compressed clusters never have the copied flag */
714 tmp = cpu_to_be64(cluster_offset);
716 /* update L2 table */
717 l2_table[l2_index] = tmp;
718 if (bdrv_pwrite(s->hd,
719 l2_offset + l2_index * sizeof(tmp), &tmp, sizeof(tmp)) != sizeof(tmp))
721 return cluster_offset;
724 static int qcow_is_allocated(BlockDriverState *bs, int64_t sector_num,
725 int nb_sectors, int *pnum)
727 BDRVQcowState *s = bs->opaque;
728 int index_in_cluster, n;
729 uint64_t cluster_offset;
731 cluster_offset = get_cluster_offset(bs, sector_num << 9, 0, 0, 0, 0);
732 index_in_cluster = sector_num & (s->cluster_sectors - 1);
733 n = s->cluster_sectors - index_in_cluster;
737 return (cluster_offset != 0);
740 static int decompress_buffer(uint8_t *out_buf, int out_buf_size,
741 const uint8_t *buf, int buf_size)
743 z_stream strm1, *strm = &strm1;
746 memset(strm, 0, sizeof(*strm));
748 strm->next_in = (uint8_t *)buf;
749 strm->avail_in = buf_size;
750 strm->next_out = out_buf;
751 strm->avail_out = out_buf_size;
753 ret = inflateInit2(strm, -12);
756 ret = inflate(strm, Z_FINISH);
757 out_len = strm->next_out - out_buf;
758 if ((ret != Z_STREAM_END && ret != Z_BUF_ERROR) ||
759 out_len != out_buf_size) {
767 static int decompress_cluster(BDRVQcowState *s, uint64_t cluster_offset)
769 int ret, csize, nb_csectors, sector_offset;
772 coffset = cluster_offset & s->cluster_offset_mask;
773 if (s->cluster_cache_offset != coffset) {
774 nb_csectors = ((cluster_offset >> s->csize_shift) & s->csize_mask) + 1;
775 sector_offset = coffset & 511;
776 csize = nb_csectors * 512 - sector_offset;
777 ret = bdrv_read(s->hd, coffset >> 9, s->cluster_data, nb_csectors);
781 if (decompress_buffer(s->cluster_cache, s->cluster_size,
782 s->cluster_data + sector_offset, csize) < 0) {
785 s->cluster_cache_offset = coffset;
790 /* handle reading after the end of the backing file */
791 static int backing_read1(BlockDriverState *bs,
792 int64_t sector_num, uint8_t *buf, int nb_sectors)
795 if ((sector_num + nb_sectors) <= bs->total_sectors)
797 if (sector_num >= bs->total_sectors)
800 n1 = bs->total_sectors - sector_num;
801 memset(buf + n1 * 512, 0, 512 * (nb_sectors - n1));
805 static int qcow_read(BlockDriverState *bs, int64_t sector_num,
806 uint8_t *buf, int nb_sectors)
808 BDRVQcowState *s = bs->opaque;
809 int ret, index_in_cluster, n, n1;
810 uint64_t cluster_offset;
812 while (nb_sectors > 0) {
813 cluster_offset = get_cluster_offset(bs, sector_num << 9, 0, 0, 0, 0);
814 index_in_cluster = sector_num & (s->cluster_sectors - 1);
815 n = s->cluster_sectors - index_in_cluster;
818 if (!cluster_offset) {
819 if (bs->backing_hd) {
820 /* read from the base image */
821 n1 = backing_read1(bs->backing_hd, sector_num, buf, n);
823 ret = bdrv_read(bs->backing_hd, sector_num, buf, n1);
828 memset(buf, 0, 512 * n);
830 } else if (cluster_offset & QCOW_OFLAG_COMPRESSED) {
831 if (decompress_cluster(s, cluster_offset) < 0)
833 memcpy(buf, s->cluster_cache + index_in_cluster * 512, 512 * n);
835 ret = bdrv_pread(s->hd, cluster_offset + index_in_cluster * 512, buf, n * 512);
838 if (s->crypt_method) {
839 encrypt_sectors(s, sector_num, buf, buf, n, 0,
840 &s->aes_decrypt_key);
850 static int qcow_write(BlockDriverState *bs, int64_t sector_num,
851 const uint8_t *buf, int nb_sectors)
853 BDRVQcowState *s = bs->opaque;
854 int ret, index_in_cluster, n;
855 uint64_t cluster_offset;
857 while (nb_sectors > 0) {
858 index_in_cluster = sector_num & (s->cluster_sectors - 1);
859 n = s->cluster_sectors - index_in_cluster;
862 cluster_offset = get_cluster_offset(bs, sector_num << 9, 1, 0,
864 index_in_cluster + n);
867 if (s->crypt_method) {
868 encrypt_sectors(s, sector_num, s->cluster_data, buf, n, 1,
869 &s->aes_encrypt_key);
870 ret = bdrv_pwrite(s->hd, cluster_offset + index_in_cluster * 512,
871 s->cluster_data, n * 512);
873 ret = bdrv_pwrite(s->hd, cluster_offset + index_in_cluster * 512, buf, n * 512);
881 s->cluster_cache_offset = -1; /* disable compressed cache */
885 typedef struct QCowAIOCB {
886 BlockDriverAIOCB common;
891 uint64_t cluster_offset;
892 uint8_t *cluster_data;
893 BlockDriverAIOCB *hd_aiocb;
896 static void qcow_aio_read_cb(void *opaque, int ret)
898 QCowAIOCB *acb = opaque;
899 BlockDriverState *bs = acb->common.bs;
900 BDRVQcowState *s = bs->opaque;
901 int index_in_cluster, n1;
903 acb->hd_aiocb = NULL;
906 acb->common.cb(acb->common.opaque, ret);
907 qemu_aio_release(acb);
912 /* post process the read buffer */
913 if (!acb->cluster_offset) {
915 } else if (acb->cluster_offset & QCOW_OFLAG_COMPRESSED) {
918 if (s->crypt_method) {
919 encrypt_sectors(s, acb->sector_num, acb->buf, acb->buf,
921 &s->aes_decrypt_key);
925 acb->nb_sectors -= acb->n;
926 acb->sector_num += acb->n;
927 acb->buf += acb->n * 512;
929 if (acb->nb_sectors == 0) {
930 /* request completed */
931 acb->common.cb(acb->common.opaque, 0);
932 qemu_aio_release(acb);
936 /* prepare next AIO request */
937 acb->cluster_offset = get_cluster_offset(bs, acb->sector_num << 9,
939 index_in_cluster = acb->sector_num & (s->cluster_sectors - 1);
940 acb->n = s->cluster_sectors - index_in_cluster;
941 if (acb->n > acb->nb_sectors)
942 acb->n = acb->nb_sectors;
944 if (!acb->cluster_offset) {
945 if (bs->backing_hd) {
946 /* read from the base image */
947 n1 = backing_read1(bs->backing_hd, acb->sector_num,
950 acb->hd_aiocb = bdrv_aio_read(bs->backing_hd, acb->sector_num,
951 acb->buf, acb->n, qcow_aio_read_cb, acb);
952 if (acb->hd_aiocb == NULL)
958 /* Note: in this case, no need to wait */
959 memset(acb->buf, 0, 512 * acb->n);
962 } else if (acb->cluster_offset & QCOW_OFLAG_COMPRESSED) {
963 /* add AIO support for compressed blocks ? */
964 if (decompress_cluster(s, acb->cluster_offset) < 0)
967 s->cluster_cache + index_in_cluster * 512, 512 * acb->n);
970 if ((acb->cluster_offset & 511) != 0) {
974 acb->hd_aiocb = bdrv_aio_read(s->hd,
975 (acb->cluster_offset >> 9) + index_in_cluster,
976 acb->buf, acb->n, qcow_aio_read_cb, acb);
977 if (acb->hd_aiocb == NULL)
982 static QCowAIOCB *qcow_aio_setup(BlockDriverState *bs,
983 int64_t sector_num, uint8_t *buf, int nb_sectors,
984 BlockDriverCompletionFunc *cb, void *opaque)
988 acb = qemu_aio_get(bs, cb, opaque);
991 acb->hd_aiocb = NULL;
992 acb->sector_num = sector_num;
994 acb->nb_sectors = nb_sectors;
996 acb->cluster_offset = 0;
1000 static BlockDriverAIOCB *qcow_aio_read(BlockDriverState *bs,
1001 int64_t sector_num, uint8_t *buf, int nb_sectors,
1002 BlockDriverCompletionFunc *cb, void *opaque)
1006 acb = qcow_aio_setup(bs, sector_num, buf, nb_sectors, cb, opaque);
1010 qcow_aio_read_cb(acb, 0);
1011 return &acb->common;
1014 static void qcow_aio_write_cb(void *opaque, int ret)
1016 QCowAIOCB *acb = opaque;
1017 BlockDriverState *bs = acb->common.bs;
1018 BDRVQcowState *s = bs->opaque;
1019 int index_in_cluster;
1020 uint64_t cluster_offset;
1021 const uint8_t *src_buf;
1023 acb->hd_aiocb = NULL;
1027 acb->common.cb(acb->common.opaque, ret);
1028 qemu_aio_release(acb);
1032 acb->nb_sectors -= acb->n;
1033 acb->sector_num += acb->n;
1034 acb->buf += acb->n * 512;
1036 if (acb->nb_sectors == 0) {
1037 /* request completed */
1038 acb->common.cb(acb->common.opaque, 0);
1039 qemu_aio_release(acb);
1043 index_in_cluster = acb->sector_num & (s->cluster_sectors - 1);
1044 acb->n = s->cluster_sectors - index_in_cluster;
1045 if (acb->n > acb->nb_sectors)
1046 acb->n = acb->nb_sectors;
1047 cluster_offset = get_cluster_offset(bs, acb->sector_num << 9, 1, 0,
1049 index_in_cluster + acb->n);
1050 if (!cluster_offset || (cluster_offset & 511) != 0) {
1054 if (s->crypt_method) {
1055 if (!acb->cluster_data) {
1056 acb->cluster_data = qemu_mallocz(s->cluster_size);
1057 if (!acb->cluster_data) {
1062 encrypt_sectors(s, acb->sector_num, acb->cluster_data, acb->buf,
1063 acb->n, 1, &s->aes_encrypt_key);
1064 src_buf = acb->cluster_data;
1068 acb->hd_aiocb = bdrv_aio_write(s->hd,
1069 (cluster_offset >> 9) + index_in_cluster,
1071 qcow_aio_write_cb, acb);
1072 if (acb->hd_aiocb == NULL)
1076 static BlockDriverAIOCB *qcow_aio_write(BlockDriverState *bs,
1077 int64_t sector_num, const uint8_t *buf, int nb_sectors,
1078 BlockDriverCompletionFunc *cb, void *opaque)
1080 BDRVQcowState *s = bs->opaque;
1083 s->cluster_cache_offset = -1; /* disable compressed cache */
1085 acb = qcow_aio_setup(bs, sector_num, (uint8_t*)buf, nb_sectors, cb, opaque);
1089 qcow_aio_write_cb(acb, 0);
1090 return &acb->common;
1093 static void qcow_aio_cancel(BlockDriverAIOCB *blockacb)
1095 QCowAIOCB *acb = (QCowAIOCB *)blockacb;
1097 bdrv_aio_cancel(acb->hd_aiocb);
1098 qemu_aio_release(acb);
1101 static void qcow_close(BlockDriverState *bs)
1103 BDRVQcowState *s = bs->opaque;
1104 qemu_free(s->l1_table);
1105 qemu_free(s->l2_cache);
1106 qemu_free(s->cluster_cache);
1107 qemu_free(s->cluster_data);
1112 /* XXX: use std qcow open function ? */
1113 typedef struct QCowCreateState {
1116 uint16_t *refcount_block;
1117 uint64_t *refcount_table;
1118 int64_t l1_table_offset;
1119 int64_t refcount_table_offset;
1120 int64_t refcount_block_offset;
1123 static void create_refcount_update(QCowCreateState *s,
1124 int64_t offset, int64_t size)
1127 int64_t start, last, cluster_offset;
1130 start = offset & ~(s->cluster_size - 1);
1131 last = (offset + size - 1) & ~(s->cluster_size - 1);
1132 for(cluster_offset = start; cluster_offset <= last;
1133 cluster_offset += s->cluster_size) {
1134 p = &s->refcount_block[cluster_offset >> s->cluster_bits];
1135 refcount = be16_to_cpu(*p);
1137 *p = cpu_to_be16(refcount);
1141 static int qcow_create(const char *filename, int64_t total_size,
1142 const char *backing_file, int flags)
1144 int fd, header_size, backing_filename_len, l1_size, i, shift, l2_bits;
1146 uint64_t tmp, offset;
1147 QCowCreateState s1, *s = &s1;
1149 memset(s, 0, sizeof(*s));
1151 fd = open(filename, O_WRONLY | O_CREAT | O_TRUNC | O_BINARY, 0644);
1154 memset(&header, 0, sizeof(header));
1155 header.magic = cpu_to_be32(QCOW_MAGIC);
1156 header.version = cpu_to_be32(QCOW_VERSION);
1157 header.size = cpu_to_be64(total_size * 512);
1158 header_size = sizeof(header);
1159 backing_filename_len = 0;
1161 header.backing_file_offset = cpu_to_be64(header_size);
1162 backing_filename_len = strlen(backing_file);
1163 header.backing_file_size = cpu_to_be32(backing_filename_len);
1164 header_size += backing_filename_len;
1166 s->cluster_bits = 12; /* 4 KB clusters */
1167 s->cluster_size = 1 << s->cluster_bits;
1168 header.cluster_bits = cpu_to_be32(s->cluster_bits);
1169 header_size = (header_size + 7) & ~7;
1170 if (flags & BLOCK_FLAG_ENCRYPT) {
1171 header.crypt_method = cpu_to_be32(QCOW_CRYPT_AES);
1173 header.crypt_method = cpu_to_be32(QCOW_CRYPT_NONE);
1175 l2_bits = s->cluster_bits - 3;
1176 shift = s->cluster_bits + l2_bits;
1177 l1_size = (((total_size * 512) + (1LL << shift) - 1) >> shift);
1178 offset = align_offset(header_size, s->cluster_size);
1179 s->l1_table_offset = offset;
1180 header.l1_table_offset = cpu_to_be64(s->l1_table_offset);
1181 header.l1_size = cpu_to_be32(l1_size);
1182 offset += align_offset(l1_size * sizeof(uint64_t), s->cluster_size);
1184 s->refcount_table = qemu_mallocz(s->cluster_size);
1185 if (!s->refcount_table)
1187 s->refcount_block = qemu_mallocz(s->cluster_size);
1188 if (!s->refcount_block)
1191 s->refcount_table_offset = offset;
1192 header.refcount_table_offset = cpu_to_be64(offset);
1193 header.refcount_table_clusters = cpu_to_be32(1);
1194 offset += s->cluster_size;
1196 s->refcount_table[0] = cpu_to_be64(offset);
1197 s->refcount_block_offset = offset;
1198 offset += s->cluster_size;
1200 /* update refcounts */
1201 create_refcount_update(s, 0, header_size);
1202 create_refcount_update(s, s->l1_table_offset, l1_size * sizeof(uint64_t));
1203 create_refcount_update(s, s->refcount_table_offset, s->cluster_size);
1204 create_refcount_update(s, s->refcount_block_offset, s->cluster_size);
1206 /* write all the data */
1207 write(fd, &header, sizeof(header));
1209 write(fd, backing_file, backing_filename_len);
1211 lseek(fd, s->l1_table_offset, SEEK_SET);
1213 for(i = 0;i < l1_size; i++) {
1214 write(fd, &tmp, sizeof(tmp));
1216 lseek(fd, s->refcount_table_offset, SEEK_SET);
1217 write(fd, s->refcount_table, s->cluster_size);
1219 lseek(fd, s->refcount_block_offset, SEEK_SET);
1220 write(fd, s->refcount_block, s->cluster_size);
1222 qemu_free(s->refcount_table);
1223 qemu_free(s->refcount_block);
1227 qemu_free(s->refcount_table);
1228 qemu_free(s->refcount_block);
1233 static int qcow_make_empty(BlockDriverState *bs)
1236 /* XXX: not correct */
1237 BDRVQcowState *s = bs->opaque;
1238 uint32_t l1_length = s->l1_size * sizeof(uint64_t);
1241 memset(s->l1_table, 0, l1_length);
1242 if (bdrv_pwrite(s->hd, s->l1_table_offset, s->l1_table, l1_length) < 0)
1244 ret = bdrv_truncate(s->hd, s->l1_table_offset + l1_length);
1253 /* XXX: put compressed sectors first, then all the cluster aligned
1254 tables to avoid losing bytes in alignment */
1255 static int qcow_write_compressed(BlockDriverState *bs, int64_t sector_num,
1256 const uint8_t *buf, int nb_sectors)
1258 BDRVQcowState *s = bs->opaque;
1262 uint64_t cluster_offset;
1264 if (nb_sectors == 0) {
1265 /* align end of file to a sector boundary to ease reading with
1266 sector based I/Os */
1267 cluster_offset = bdrv_getlength(s->hd);
1268 cluster_offset = (cluster_offset + 511) & ~511;
1269 bdrv_truncate(s->hd, cluster_offset);
1273 if (nb_sectors != s->cluster_sectors)
1276 out_buf = qemu_malloc(s->cluster_size + (s->cluster_size / 1000) + 128);
1280 /* best compression, small window, no zlib header */
1281 memset(&strm, 0, sizeof(strm));
1282 ret = deflateInit2(&strm, Z_DEFAULT_COMPRESSION,
1284 9, Z_DEFAULT_STRATEGY);
1290 strm.avail_in = s->cluster_size;
1291 strm.next_in = (uint8_t *)buf;
1292 strm.avail_out = s->cluster_size;
1293 strm.next_out = out_buf;
1295 ret = deflate(&strm, Z_FINISH);
1296 if (ret != Z_STREAM_END && ret != Z_OK) {
1301 out_len = strm.next_out - out_buf;
1305 if (ret != Z_STREAM_END || out_len >= s->cluster_size) {
1306 /* could not compress: write normal cluster */
1307 qcow_write(bs, sector_num, buf, s->cluster_sectors);
1309 cluster_offset = get_cluster_offset(bs, sector_num << 9, 2,
1311 cluster_offset &= s->cluster_offset_mask;
1312 if (bdrv_pwrite(s->hd, cluster_offset, out_buf, out_len) != out_len) {
1322 static void qcow_flush(BlockDriverState *bs)
1324 BDRVQcowState *s = bs->opaque;
1328 static int qcow_get_info(BlockDriverState *bs, BlockDriverInfo *bdi)
1330 BDRVQcowState *s = bs->opaque;
1331 bdi->cluster_size = s->cluster_size;
1332 bdi->vm_state_offset = (int64_t)s->l1_vm_state_index <<
1333 (s->cluster_bits + s->l2_bits);
1337 /*********************************************************/
1338 /* snapshot support */
1340 /* update the refcounts of snapshots and the copied flag */
1341 static int update_snapshot_refcount(BlockDriverState *bs,
1342 int64_t l1_table_offset,
1346 BDRVQcowState *s = bs->opaque;
1347 uint64_t *l1_table, *l2_table, l2_offset, offset, l1_size2, l1_allocated;
1348 int64_t old_offset, old_l2_offset;
1349 int l2_size, i, j, l1_modified, l2_modified, nb_csectors, refcount;
1355 l1_size2 = l1_size * sizeof(uint64_t);
1357 if (l1_table_offset != s->l1_table_offset) {
1358 l1_table = qemu_malloc(l1_size2);
1362 if (bdrv_pread(s->hd, l1_table_offset,
1363 l1_table, l1_size2) != l1_size2)
1365 for(i = 0;i < l1_size; i++)
1366 be64_to_cpus(&l1_table[i]);
1368 assert(l1_size == s->l1_size);
1369 l1_table = s->l1_table;
1373 l2_size = s->l2_size * sizeof(uint64_t);
1374 l2_table = qemu_malloc(l2_size);
1378 for(i = 0; i < l1_size; i++) {
1379 l2_offset = l1_table[i];
1381 old_l2_offset = l2_offset;
1382 l2_offset &= ~QCOW_OFLAG_COPIED;
1384 if (bdrv_pread(s->hd, l2_offset, l2_table, l2_size) != l2_size)
1386 for(j = 0; j < s->l2_size; j++) {
1387 offset = be64_to_cpu(l2_table[j]);
1389 old_offset = offset;
1390 offset &= ~QCOW_OFLAG_COPIED;
1391 if (offset & QCOW_OFLAG_COMPRESSED) {
1392 nb_csectors = ((offset >> s->csize_shift) &
1395 update_refcount(bs, (offset & s->cluster_offset_mask) & ~511,
1396 nb_csectors * 512, addend);
1397 /* compressed clusters are never modified */
1401 refcount = update_cluster_refcount(bs, offset >> s->cluster_bits, addend);
1403 refcount = get_refcount(bs, offset >> s->cluster_bits);
1407 if (refcount == 1) {
1408 offset |= QCOW_OFLAG_COPIED;
1410 if (offset != old_offset) {
1411 l2_table[j] = cpu_to_be64(offset);
1417 if (bdrv_pwrite(s->hd,
1418 l2_offset, l2_table, l2_size) != l2_size)
1423 refcount = update_cluster_refcount(bs, l2_offset >> s->cluster_bits, addend);
1425 refcount = get_refcount(bs, l2_offset >> s->cluster_bits);
1427 if (refcount == 1) {
1428 l2_offset |= QCOW_OFLAG_COPIED;
1430 if (l2_offset != old_l2_offset) {
1431 l1_table[i] = l2_offset;
1437 for(i = 0; i < l1_size; i++)
1438 cpu_to_be64s(&l1_table[i]);
1439 if (bdrv_pwrite(s->hd, l1_table_offset, l1_table,
1440 l1_size2) != l1_size2)
1442 for(i = 0; i < l1_size; i++)
1443 be64_to_cpus(&l1_table[i]);
1446 qemu_free(l1_table);
1447 qemu_free(l2_table);
1451 qemu_free(l1_table);
1452 qemu_free(l2_table);
1456 static void qcow_free_snapshots(BlockDriverState *bs)
1458 BDRVQcowState *s = bs->opaque;
1461 for(i = 0; i < s->nb_snapshots; i++) {
1462 qemu_free(s->snapshots[i].name);
1463 qemu_free(s->snapshots[i].id_str);
1465 qemu_free(s->snapshots);
1466 s->snapshots = NULL;
1467 s->nb_snapshots = 0;
1470 static int qcow_read_snapshots(BlockDriverState *bs)
1472 BDRVQcowState *s = bs->opaque;
1473 QCowSnapshotHeader h;
1475 int i, id_str_size, name_size;
1477 uint32_t extra_data_size;
1479 offset = s->snapshots_offset;
1480 s->snapshots = qemu_mallocz(s->nb_snapshots * sizeof(QCowSnapshot));
1483 for(i = 0; i < s->nb_snapshots; i++) {
1484 offset = align_offset(offset, 8);
1485 if (bdrv_pread(s->hd, offset, &h, sizeof(h)) != sizeof(h))
1487 offset += sizeof(h);
1488 sn = s->snapshots + i;
1489 sn->l1_table_offset = be64_to_cpu(h.l1_table_offset);
1490 sn->l1_size = be32_to_cpu(h.l1_size);
1491 sn->vm_state_size = be32_to_cpu(h.vm_state_size);
1492 sn->date_sec = be32_to_cpu(h.date_sec);
1493 sn->date_nsec = be32_to_cpu(h.date_nsec);
1494 sn->vm_clock_nsec = be64_to_cpu(h.vm_clock_nsec);
1495 extra_data_size = be32_to_cpu(h.extra_data_size);
1497 id_str_size = be16_to_cpu(h.id_str_size);
1498 name_size = be16_to_cpu(h.name_size);
1500 offset += extra_data_size;
1502 sn->id_str = qemu_malloc(id_str_size + 1);
1505 if (bdrv_pread(s->hd, offset, sn->id_str, id_str_size) != id_str_size)
1507 offset += id_str_size;
1508 sn->id_str[id_str_size] = '\0';
1510 sn->name = qemu_malloc(name_size + 1);
1513 if (bdrv_pread(s->hd, offset, sn->name, name_size) != name_size)
1515 offset += name_size;
1516 sn->name[name_size] = '\0';
1518 s->snapshots_size = offset - s->snapshots_offset;
1521 qcow_free_snapshots(bs);
1525 /* add at the end of the file a new list of snapshots */
1526 static int qcow_write_snapshots(BlockDriverState *bs)
1528 BDRVQcowState *s = bs->opaque;
1530 QCowSnapshotHeader h;
1531 int i, name_size, id_str_size, snapshots_size;
1534 int64_t offset, snapshots_offset;
1536 /* compute the size of the snapshots */
1538 for(i = 0; i < s->nb_snapshots; i++) {
1539 sn = s->snapshots + i;
1540 offset = align_offset(offset, 8);
1541 offset += sizeof(h);
1542 offset += strlen(sn->id_str);
1543 offset += strlen(sn->name);
1545 snapshots_size = offset;
1547 snapshots_offset = alloc_clusters(bs, snapshots_size);
1548 offset = snapshots_offset;
1550 for(i = 0; i < s->nb_snapshots; i++) {
1551 sn = s->snapshots + i;
1552 memset(&h, 0, sizeof(h));
1553 h.l1_table_offset = cpu_to_be64(sn->l1_table_offset);
1554 h.l1_size = cpu_to_be32(sn->l1_size);
1555 h.vm_state_size = cpu_to_be32(sn->vm_state_size);
1556 h.date_sec = cpu_to_be32(sn->date_sec);
1557 h.date_nsec = cpu_to_be32(sn->date_nsec);
1558 h.vm_clock_nsec = cpu_to_be64(sn->vm_clock_nsec);
1560 id_str_size = strlen(sn->id_str);
1561 name_size = strlen(sn->name);
1562 h.id_str_size = cpu_to_be16(id_str_size);
1563 h.name_size = cpu_to_be16(name_size);
1564 offset = align_offset(offset, 8);
1565 if (bdrv_pwrite(s->hd, offset, &h, sizeof(h)) != sizeof(h))
1567 offset += sizeof(h);
1568 if (bdrv_pwrite(s->hd, offset, sn->id_str, id_str_size) != id_str_size)
1570 offset += id_str_size;
1571 if (bdrv_pwrite(s->hd, offset, sn->name, name_size) != name_size)
1573 offset += name_size;
1576 /* update the various header fields */
1577 data64 = cpu_to_be64(snapshots_offset);
1578 if (bdrv_pwrite(s->hd, offsetof(QCowHeader, snapshots_offset),
1579 &data64, sizeof(data64)) != sizeof(data64))
1581 data32 = cpu_to_be32(s->nb_snapshots);
1582 if (bdrv_pwrite(s->hd, offsetof(QCowHeader, nb_snapshots),
1583 &data32, sizeof(data32)) != sizeof(data32))
1586 /* free the old snapshot table */
1587 free_clusters(bs, s->snapshots_offset, s->snapshots_size);
1588 s->snapshots_offset = snapshots_offset;
1589 s->snapshots_size = snapshots_size;
1595 static void find_new_snapshot_id(BlockDriverState *bs,
1596 char *id_str, int id_str_size)
1598 BDRVQcowState *s = bs->opaque;
1600 int i, id, id_max = 0;
1602 for(i = 0; i < s->nb_snapshots; i++) {
1603 sn = s->snapshots + i;
1604 id = strtoul(sn->id_str, NULL, 10);
1608 snprintf(id_str, id_str_size, "%d", id_max + 1);
1611 static int find_snapshot_by_id(BlockDriverState *bs, const char *id_str)
1613 BDRVQcowState *s = bs->opaque;
1616 for(i = 0; i < s->nb_snapshots; i++) {
1617 if (!strcmp(s->snapshots[i].id_str, id_str))
1623 static int find_snapshot_by_id_or_name(BlockDriverState *bs, const char *name)
1625 BDRVQcowState *s = bs->opaque;
1628 ret = find_snapshot_by_id(bs, name);
1631 for(i = 0; i < s->nb_snapshots; i++) {
1632 if (!strcmp(s->snapshots[i].name, name))
1638 /* if no id is provided, a new one is constructed */
1639 static int qcow_snapshot_create(BlockDriverState *bs,
1640 QEMUSnapshotInfo *sn_info)
1642 BDRVQcowState *s = bs->opaque;
1643 QCowSnapshot *snapshots1, sn1, *sn = &sn1;
1645 uint64_t *l1_table = NULL;
1647 memset(sn, 0, sizeof(*sn));
1649 if (sn_info->id_str[0] == '\0') {
1650 /* compute a new id */
1651 find_new_snapshot_id(bs, sn_info->id_str, sizeof(sn_info->id_str));
1654 /* check that the ID is unique */
1655 if (find_snapshot_by_id(bs, sn_info->id_str) >= 0)
1658 sn->id_str = qemu_strdup(sn_info->id_str);
1661 sn->name = qemu_strdup(sn_info->name);
1664 sn->vm_state_size = sn_info->vm_state_size;
1665 sn->date_sec = sn_info->date_sec;
1666 sn->date_nsec = sn_info->date_nsec;
1667 sn->vm_clock_nsec = sn_info->vm_clock_nsec;
1669 ret = update_snapshot_refcount(bs, s->l1_table_offset, s->l1_size, 1);
1673 /* create the L1 table of the snapshot */
1674 sn->l1_table_offset = alloc_clusters(bs, s->l1_size * sizeof(uint64_t));
1675 sn->l1_size = s->l1_size;
1677 l1_table = qemu_malloc(s->l1_size * sizeof(uint64_t));
1680 for(i = 0; i < s->l1_size; i++) {
1681 l1_table[i] = cpu_to_be64(s->l1_table[i]);
1683 if (bdrv_pwrite(s->hd, sn->l1_table_offset,
1684 l1_table, s->l1_size * sizeof(uint64_t)) !=
1685 (s->l1_size * sizeof(uint64_t)))
1687 qemu_free(l1_table);
1690 snapshots1 = qemu_malloc((s->nb_snapshots + 1) * sizeof(QCowSnapshot));
1693 memcpy(snapshots1, s->snapshots, s->nb_snapshots * sizeof(QCowSnapshot));
1694 s->snapshots = snapshots1;
1695 s->snapshots[s->nb_snapshots++] = *sn;
1697 if (qcow_write_snapshots(bs) < 0)
1700 check_refcounts(bs);
1704 qemu_free(sn->name);
1705 qemu_free(l1_table);
1709 /* copy the snapshot 'snapshot_name' into the current disk image */
1710 static int qcow_snapshot_goto(BlockDriverState *bs,
1711 const char *snapshot_id)
1713 BDRVQcowState *s = bs->opaque;
1715 int i, snapshot_index, l1_size2;
1717 snapshot_index = find_snapshot_by_id_or_name(bs, snapshot_id);
1718 if (snapshot_index < 0)
1720 sn = &s->snapshots[snapshot_index];
1722 if (update_snapshot_refcount(bs, s->l1_table_offset, s->l1_size, -1) < 0)
1725 if (grow_l1_table(bs, sn->l1_size) < 0)
1728 s->l1_size = sn->l1_size;
1729 l1_size2 = s->l1_size * sizeof(uint64_t);
1730 /* copy the snapshot l1 table to the current l1 table */
1731 if (bdrv_pread(s->hd, sn->l1_table_offset,
1732 s->l1_table, l1_size2) != l1_size2)
1734 if (bdrv_pwrite(s->hd, s->l1_table_offset,
1735 s->l1_table, l1_size2) != l1_size2)
1737 for(i = 0;i < s->l1_size; i++) {
1738 be64_to_cpus(&s->l1_table[i]);
1741 if (update_snapshot_refcount(bs, s->l1_table_offset, s->l1_size, 1) < 0)
1745 check_refcounts(bs);
1752 static int qcow_snapshot_delete(BlockDriverState *bs, const char *snapshot_id)
1754 BDRVQcowState *s = bs->opaque;
1756 int snapshot_index, ret;
1758 snapshot_index = find_snapshot_by_id_or_name(bs, snapshot_id);
1759 if (snapshot_index < 0)
1761 sn = &s->snapshots[snapshot_index];
1763 ret = update_snapshot_refcount(bs, sn->l1_table_offset, sn->l1_size, -1);
1766 /* must update the copied flag on the current cluster offsets */
1767 ret = update_snapshot_refcount(bs, s->l1_table_offset, s->l1_size, 0);
1770 free_clusters(bs, sn->l1_table_offset, sn->l1_size * sizeof(uint64_t));
1772 qemu_free(sn->id_str);
1773 qemu_free(sn->name);
1774 memmove(sn, sn + 1, (s->nb_snapshots - snapshot_index - 1) * sizeof(*sn));
1776 ret = qcow_write_snapshots(bs);
1778 /* XXX: restore snapshot if error ? */
1782 check_refcounts(bs);
1787 static int qcow_snapshot_list(BlockDriverState *bs,
1788 QEMUSnapshotInfo **psn_tab)
1790 BDRVQcowState *s = bs->opaque;
1791 QEMUSnapshotInfo *sn_tab, *sn_info;
1795 sn_tab = qemu_mallocz(s->nb_snapshots * sizeof(QEMUSnapshotInfo));
1798 for(i = 0; i < s->nb_snapshots; i++) {
1799 sn_info = sn_tab + i;
1800 sn = s->snapshots + i;
1801 pstrcpy(sn_info->id_str, sizeof(sn_info->id_str),
1803 pstrcpy(sn_info->name, sizeof(sn_info->name),
1805 sn_info->vm_state_size = sn->vm_state_size;
1806 sn_info->date_sec = sn->date_sec;
1807 sn_info->date_nsec = sn->date_nsec;
1808 sn_info->vm_clock_nsec = sn->vm_clock_nsec;
1811 return s->nb_snapshots;
1818 /*********************************************************/
1819 /* refcount handling */
1821 static int refcount_init(BlockDriverState *bs)
1823 BDRVQcowState *s = bs->opaque;
1824 int ret, refcount_table_size2, i;
1826 s->refcount_block_cache = qemu_malloc(s->cluster_size);
1827 if (!s->refcount_block_cache)
1829 refcount_table_size2 = s->refcount_table_size * sizeof(uint64_t);
1830 s->refcount_table = qemu_malloc(refcount_table_size2);
1831 if (!s->refcount_table)
1833 if (s->refcount_table_size > 0) {
1834 ret = bdrv_pread(s->hd, s->refcount_table_offset,
1835 s->refcount_table, refcount_table_size2);
1836 if (ret != refcount_table_size2)
1838 for(i = 0; i < s->refcount_table_size; i++)
1839 be64_to_cpus(&s->refcount_table[i]);
1846 static void refcount_close(BlockDriverState *bs)
1848 BDRVQcowState *s = bs->opaque;
1849 qemu_free(s->refcount_block_cache);
1850 qemu_free(s->refcount_table);
1854 static int load_refcount_block(BlockDriverState *bs,
1855 int64_t refcount_block_offset)
1857 BDRVQcowState *s = bs->opaque;
1859 ret = bdrv_pread(s->hd, refcount_block_offset, s->refcount_block_cache,
1861 if (ret != s->cluster_size)
1863 s->refcount_block_cache_offset = refcount_block_offset;
1867 static int get_refcount(BlockDriverState *bs, int64_t cluster_index)
1869 BDRVQcowState *s = bs->opaque;
1870 int refcount_table_index, block_index;
1871 int64_t refcount_block_offset;
1873 refcount_table_index = cluster_index >> (s->cluster_bits - REFCOUNT_SHIFT);
1874 if (refcount_table_index >= s->refcount_table_size)
1876 refcount_block_offset = s->refcount_table[refcount_table_index];
1877 if (!refcount_block_offset)
1879 if (refcount_block_offset != s->refcount_block_cache_offset) {
1880 /* better than nothing: return allocated if read error */
1881 if (load_refcount_block(bs, refcount_block_offset) < 0)
1884 block_index = cluster_index &
1885 ((1 << (s->cluster_bits - REFCOUNT_SHIFT)) - 1);
1886 return be16_to_cpu(s->refcount_block_cache[block_index]);
1889 /* return < 0 if error */
1890 static int64_t alloc_clusters_noref(BlockDriverState *bs, int64_t size)
1892 BDRVQcowState *s = bs->opaque;
1895 nb_clusters = (size + s->cluster_size - 1) >> s->cluster_bits;
1897 if (get_refcount(bs, s->free_cluster_index) == 0) {
1898 s->free_cluster_index++;
1899 for(i = 1; i < nb_clusters; i++) {
1900 if (get_refcount(bs, s->free_cluster_index) != 0)
1902 s->free_cluster_index++;
1905 printf("alloc_clusters: size=%lld -> %lld\n",
1907 (s->free_cluster_index - nb_clusters) << s->cluster_bits);
1909 return (s->free_cluster_index - nb_clusters) << s->cluster_bits;
1912 s->free_cluster_index++;
1917 static int64_t alloc_clusters(BlockDriverState *bs, int64_t size)
1921 offset = alloc_clusters_noref(bs, size);
1922 update_refcount(bs, offset, size, 1);
1926 /* only used to allocate compressed sectors. We try to allocate
1927 contiguous sectors. size must be <= cluster_size */
1928 static int64_t alloc_bytes(BlockDriverState *bs, int size)
1930 BDRVQcowState *s = bs->opaque;
1931 int64_t offset, cluster_offset;
1932 int free_in_cluster;
1934 assert(size > 0 && size <= s->cluster_size);
1935 if (s->free_byte_offset == 0) {
1936 s->free_byte_offset = alloc_clusters(bs, s->cluster_size);
1939 free_in_cluster = s->cluster_size -
1940 (s->free_byte_offset & (s->cluster_size - 1));
1941 if (size <= free_in_cluster) {
1942 /* enough space in current cluster */
1943 offset = s->free_byte_offset;
1944 s->free_byte_offset += size;
1945 free_in_cluster -= size;
1946 if (free_in_cluster == 0)
1947 s->free_byte_offset = 0;
1948 if ((offset & (s->cluster_size - 1)) != 0)
1949 update_cluster_refcount(bs, offset >> s->cluster_bits, 1);
1951 offset = alloc_clusters(bs, s->cluster_size);
1952 cluster_offset = s->free_byte_offset & ~(s->cluster_size - 1);
1953 if ((cluster_offset + s->cluster_size) == offset) {
1954 /* we are lucky: contiguous data */
1955 offset = s->free_byte_offset;
1956 update_cluster_refcount(bs, offset >> s->cluster_bits, 1);
1957 s->free_byte_offset += size;
1959 s->free_byte_offset = offset;
1966 static void free_clusters(BlockDriverState *bs,
1967 int64_t offset, int64_t size)
1969 update_refcount(bs, offset, size, -1);
1972 static int grow_refcount_table(BlockDriverState *bs, int min_size)
1974 BDRVQcowState *s = bs->opaque;
1975 int new_table_size, new_table_size2, refcount_table_clusters, i, ret;
1976 uint64_t *new_table;
1977 int64_t table_offset;
1981 int64_t old_table_offset;
1983 if (min_size <= s->refcount_table_size)
1985 /* compute new table size */
1986 refcount_table_clusters = s->refcount_table_size >> (s->cluster_bits - 3);
1988 if (refcount_table_clusters == 0) {
1989 refcount_table_clusters = 1;
1991 refcount_table_clusters = (refcount_table_clusters * 3 + 1) / 2;
1993 new_table_size = refcount_table_clusters << (s->cluster_bits - 3);
1994 if (min_size <= new_table_size)
1998 printf("grow_refcount_table from %d to %d\n",
1999 s->refcount_table_size,
2002 new_table_size2 = new_table_size * sizeof(uint64_t);
2003 new_table = qemu_mallocz(new_table_size2);
2006 memcpy(new_table, s->refcount_table,
2007 s->refcount_table_size * sizeof(uint64_t));
2008 for(i = 0; i < s->refcount_table_size; i++)
2009 cpu_to_be64s(&new_table[i]);
2010 /* Note: we cannot update the refcount now to avoid recursion */
2011 table_offset = alloc_clusters_noref(bs, new_table_size2);
2012 ret = bdrv_pwrite(s->hd, table_offset, new_table, new_table_size2);
2013 if (ret != new_table_size2)
2015 for(i = 0; i < s->refcount_table_size; i++)
2016 be64_to_cpus(&new_table[i]);
2018 data64 = cpu_to_be64(table_offset);
2019 if (bdrv_pwrite(s->hd, offsetof(QCowHeader, refcount_table_offset),
2020 &data64, sizeof(data64)) != sizeof(data64))
2022 data32 = cpu_to_be32(refcount_table_clusters);
2023 if (bdrv_pwrite(s->hd, offsetof(QCowHeader, refcount_table_clusters),
2024 &data32, sizeof(data32)) != sizeof(data32))
2026 qemu_free(s->refcount_table);
2027 old_table_offset = s->refcount_table_offset;
2028 old_table_size = s->refcount_table_size;
2029 s->refcount_table = new_table;
2030 s->refcount_table_size = new_table_size;
2031 s->refcount_table_offset = table_offset;
2033 update_refcount(bs, table_offset, new_table_size2, 1);
2034 free_clusters(bs, old_table_offset, old_table_size * sizeof(uint64_t));
2037 free_clusters(bs, table_offset, new_table_size2);
2038 qemu_free(new_table);
2042 /* addend must be 1 or -1 */
2043 /* XXX: cache several refcount block clusters ? */
2044 static int update_cluster_refcount(BlockDriverState *bs,
2045 int64_t cluster_index,
2048 BDRVQcowState *s = bs->opaque;
2049 int64_t offset, refcount_block_offset;
2050 int ret, refcount_table_index, block_index, refcount;
2053 refcount_table_index = cluster_index >> (s->cluster_bits - REFCOUNT_SHIFT);
2054 if (refcount_table_index >= s->refcount_table_size) {
2057 ret = grow_refcount_table(bs, refcount_table_index + 1);
2061 refcount_block_offset = s->refcount_table[refcount_table_index];
2062 if (!refcount_block_offset) {
2065 /* create a new refcount block */
2066 /* Note: we cannot update the refcount now to avoid recursion */
2067 offset = alloc_clusters_noref(bs, s->cluster_size);
2068 memset(s->refcount_block_cache, 0, s->cluster_size);
2069 ret = bdrv_pwrite(s->hd, offset, s->refcount_block_cache, s->cluster_size);
2070 if (ret != s->cluster_size)
2072 s->refcount_table[refcount_table_index] = offset;
2073 data64 = cpu_to_be64(offset);
2074 ret = bdrv_pwrite(s->hd, s->refcount_table_offset +
2075 refcount_table_index * sizeof(uint64_t),
2076 &data64, sizeof(data64));
2077 if (ret != sizeof(data64))
2080 refcount_block_offset = offset;
2081 s->refcount_block_cache_offset = offset;
2082 update_refcount(bs, offset, s->cluster_size, 1);
2084 if (refcount_block_offset != s->refcount_block_cache_offset) {
2085 if (load_refcount_block(bs, refcount_block_offset) < 0)
2089 /* we can update the count and save it */
2090 block_index = cluster_index &
2091 ((1 << (s->cluster_bits - REFCOUNT_SHIFT)) - 1);
2092 refcount = be16_to_cpu(s->refcount_block_cache[block_index]);
2094 if (refcount < 0 || refcount > 0xffff)
2096 if (refcount == 0 && cluster_index < s->free_cluster_index) {
2097 s->free_cluster_index = cluster_index;
2099 s->refcount_block_cache[block_index] = cpu_to_be16(refcount);
2100 if (bdrv_pwrite(s->hd,
2101 refcount_block_offset + (block_index << REFCOUNT_SHIFT),
2102 &s->refcount_block_cache[block_index], 2) != 2)
2107 static void update_refcount(BlockDriverState *bs,
2108 int64_t offset, int64_t length,
2111 BDRVQcowState *s = bs->opaque;
2112 int64_t start, last, cluster_offset;
2115 printf("update_refcount: offset=%lld size=%lld addend=%d\n",
2116 offset, length, addend);
2120 start = offset & ~(s->cluster_size - 1);
2121 last = (offset + length - 1) & ~(s->cluster_size - 1);
2122 for(cluster_offset = start; cluster_offset <= last;
2123 cluster_offset += s->cluster_size) {
2124 update_cluster_refcount(bs, cluster_offset >> s->cluster_bits, addend);
2129 static void inc_refcounts(BlockDriverState *bs,
2130 uint16_t *refcount_table,
2131 int refcount_table_size,
2132 int64_t offset, int64_t size)
2134 BDRVQcowState *s = bs->opaque;
2135 int64_t start, last, cluster_offset;
2141 start = offset & ~(s->cluster_size - 1);
2142 last = (offset + size - 1) & ~(s->cluster_size - 1);
2143 for(cluster_offset = start; cluster_offset <= last;
2144 cluster_offset += s->cluster_size) {
2145 k = cluster_offset >> s->cluster_bits;
2146 if (k < 0 || k >= refcount_table_size) {
2147 printf("ERROR: invalid cluster offset=0x%llx\n", cluster_offset);
2149 if (++refcount_table[k] == 0) {
2150 printf("ERROR: overflow cluster offset=0x%llx\n", cluster_offset);
2156 static int check_refcounts_l1(BlockDriverState *bs,
2157 uint16_t *refcount_table,
2158 int refcount_table_size,
2159 int64_t l1_table_offset, int l1_size,
2162 BDRVQcowState *s = bs->opaque;
2163 uint64_t *l1_table, *l2_table, l2_offset, offset, l1_size2;
2164 int l2_size, i, j, nb_csectors, refcount;
2167 l1_size2 = l1_size * sizeof(uint64_t);
2169 inc_refcounts(bs, refcount_table, refcount_table_size,
2170 l1_table_offset, l1_size2);
2172 l1_table = qemu_malloc(l1_size2);
2175 if (bdrv_pread(s->hd, l1_table_offset,
2176 l1_table, l1_size2) != l1_size2)
2178 for(i = 0;i < l1_size; i++)
2179 be64_to_cpus(&l1_table[i]);
2181 l2_size = s->l2_size * sizeof(uint64_t);
2182 l2_table = qemu_malloc(l2_size);
2185 for(i = 0; i < l1_size; i++) {
2186 l2_offset = l1_table[i];
2189 refcount = get_refcount(bs, (l2_offset & ~QCOW_OFLAG_COPIED) >> s->cluster_bits);
2190 if ((refcount == 1) != ((l2_offset & QCOW_OFLAG_COPIED) != 0)) {
2191 printf("ERROR OFLAG_COPIED: l2_offset=%llx refcount=%d\n",
2192 l2_offset, refcount);
2195 l2_offset &= ~QCOW_OFLAG_COPIED;
2196 if (bdrv_pread(s->hd, l2_offset, l2_table, l2_size) != l2_size)
2198 for(j = 0; j < s->l2_size; j++) {
2199 offset = be64_to_cpu(l2_table[j]);
2201 if (offset & QCOW_OFLAG_COMPRESSED) {
2202 if (offset & QCOW_OFLAG_COPIED) {
2203 printf("ERROR: cluster %lld: copied flag must never be set for compressed clusters\n",
2204 offset >> s->cluster_bits);
2205 offset &= ~QCOW_OFLAG_COPIED;
2207 nb_csectors = ((offset >> s->csize_shift) &
2209 offset &= s->cluster_offset_mask;
2210 inc_refcounts(bs, refcount_table,
2211 refcount_table_size,
2212 offset & ~511, nb_csectors * 512);
2215 refcount = get_refcount(bs, (offset & ~QCOW_OFLAG_COPIED) >> s->cluster_bits);
2216 if ((refcount == 1) != ((offset & QCOW_OFLAG_COPIED) != 0)) {
2217 printf("ERROR OFLAG_COPIED: offset=%llx refcount=%d\n",
2221 offset &= ~QCOW_OFLAG_COPIED;
2222 inc_refcounts(bs, refcount_table,
2223 refcount_table_size,
2224 offset, s->cluster_size);
2228 inc_refcounts(bs, refcount_table,
2229 refcount_table_size,
2234 qemu_free(l1_table);
2235 qemu_free(l2_table);
2238 printf("ERROR: I/O error in check_refcounts_l1\n");
2239 qemu_free(l1_table);
2240 qemu_free(l2_table);
2244 static void check_refcounts(BlockDriverState *bs)
2246 BDRVQcowState *s = bs->opaque;
2248 int nb_clusters, refcount1, refcount2, i;
2250 uint16_t *refcount_table;
2252 size = bdrv_getlength(s->hd);
2253 nb_clusters = (size + s->cluster_size - 1) >> s->cluster_bits;
2254 refcount_table = qemu_mallocz(nb_clusters * sizeof(uint16_t));
2257 inc_refcounts(bs, refcount_table, nb_clusters,
2258 0, s->cluster_size);
2260 check_refcounts_l1(bs, refcount_table, nb_clusters,
2261 s->l1_table_offset, s->l1_size, 1);
2264 for(i = 0; i < s->nb_snapshots; i++) {
2265 sn = s->snapshots + i;
2266 check_refcounts_l1(bs, refcount_table, nb_clusters,
2267 sn->l1_table_offset, sn->l1_size, 0);
2269 inc_refcounts(bs, refcount_table, nb_clusters,
2270 s->snapshots_offset, s->snapshots_size);
2273 inc_refcounts(bs, refcount_table, nb_clusters,
2274 s->refcount_table_offset,
2275 s->refcount_table_size * sizeof(uint64_t));
2276 for(i = 0; i < s->refcount_table_size; i++) {
2278 offset = s->refcount_table[i];
2280 inc_refcounts(bs, refcount_table, nb_clusters,
2281 offset, s->cluster_size);
2285 /* compare ref counts */
2286 for(i = 0; i < nb_clusters; i++) {
2287 refcount1 = get_refcount(bs, i);
2288 refcount2 = refcount_table[i];
2289 if (refcount1 != refcount2)
2290 printf("ERROR cluster %d refcount=%d reference=%d\n",
2291 i, refcount1, refcount2);
2294 qemu_free(refcount_table);
2298 static void dump_refcounts(BlockDriverState *bs)
2300 BDRVQcowState *s = bs->opaque;
2301 int64_t nb_clusters, k, k1, size;
2304 size = bdrv_getlength(s->hd);
2305 nb_clusters = (size + s->cluster_size - 1) >> s->cluster_bits;
2306 for(k = 0; k < nb_clusters;) {
2308 refcount = get_refcount(bs, k);
2310 while (k < nb_clusters && get_refcount(bs, k) == refcount)
2312 printf("%lld: refcount=%d nb=%lld\n", k, refcount, k - k1);
2318 BlockDriver bdrv_qcow2 = {
2320 sizeof(BDRVQcowState),
2332 .bdrv_aio_read = qcow_aio_read,
2333 .bdrv_aio_write = qcow_aio_write,
2334 .bdrv_aio_cancel = qcow_aio_cancel,
2335 .aiocb_size = sizeof(QCowAIOCB),
2336 .bdrv_write_compressed = qcow_write_compressed,
2338 .bdrv_snapshot_create = qcow_snapshot_create,
2339 .bdrv_snapshot_goto = qcow_snapshot_goto,
2340 .bdrv_snapshot_delete = qcow_snapshot_delete,
2341 .bdrv_snapshot_list = qcow_snapshot_list,
2342 .bdrv_get_info = qcow_get_info,