Merge remote-tracking branch 'remotes/vivier2/tags/linux-user-for-4.2-pull-request...
[qemu/ar7.git] / block / qcow2-refcount.c
blobef965d78952637ed92d777db4e9a007122d9e481
1 /*
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
22 * THE SOFTWARE.
25 #include "qemu/osdep.h"
26 #include "qapi/error.h"
27 #include "qcow2.h"
28 #include "qemu/range.h"
29 #include "qemu/bswap.h"
30 #include "qemu/cutils.h"
31 #include "trace.h"
33 static int64_t alloc_clusters_noref(BlockDriverState *bs, uint64_t size,
34 uint64_t max);
35 static int QEMU_WARN_UNUSED_RESULT update_refcount(BlockDriverState *bs,
36 int64_t offset, int64_t length, uint64_t addend,
37 bool decrease, enum qcow2_discard_type type);
39 static uint64_t get_refcount_ro0(const void *refcount_array, uint64_t index);
40 static uint64_t get_refcount_ro1(const void *refcount_array, uint64_t index);
41 static uint64_t get_refcount_ro2(const void *refcount_array, uint64_t index);
42 static uint64_t get_refcount_ro3(const void *refcount_array, uint64_t index);
43 static uint64_t get_refcount_ro4(const void *refcount_array, uint64_t index);
44 static uint64_t get_refcount_ro5(const void *refcount_array, uint64_t index);
45 static uint64_t get_refcount_ro6(const void *refcount_array, uint64_t index);
47 static void set_refcount_ro0(void *refcount_array, uint64_t index,
48 uint64_t value);
49 static void set_refcount_ro1(void *refcount_array, uint64_t index,
50 uint64_t value);
51 static void set_refcount_ro2(void *refcount_array, uint64_t index,
52 uint64_t value);
53 static void set_refcount_ro3(void *refcount_array, uint64_t index,
54 uint64_t value);
55 static void set_refcount_ro4(void *refcount_array, uint64_t index,
56 uint64_t value);
57 static void set_refcount_ro5(void *refcount_array, uint64_t index,
58 uint64_t value);
59 static void set_refcount_ro6(void *refcount_array, uint64_t index,
60 uint64_t value);
63 static Qcow2GetRefcountFunc *const get_refcount_funcs[] = {
64 &get_refcount_ro0,
65 &get_refcount_ro1,
66 &get_refcount_ro2,
67 &get_refcount_ro3,
68 &get_refcount_ro4,
69 &get_refcount_ro5,
70 &get_refcount_ro6
73 static Qcow2SetRefcountFunc *const set_refcount_funcs[] = {
74 &set_refcount_ro0,
75 &set_refcount_ro1,
76 &set_refcount_ro2,
77 &set_refcount_ro3,
78 &set_refcount_ro4,
79 &set_refcount_ro5,
80 &set_refcount_ro6
84 /*********************************************************/
85 /* refcount handling */
87 static void update_max_refcount_table_index(BDRVQcow2State *s)
89 unsigned i = s->refcount_table_size - 1;
90 while (i > 0 && (s->refcount_table[i] & REFT_OFFSET_MASK) == 0) {
91 i--;
93 /* Set s->max_refcount_table_index to the index of the last used entry */
94 s->max_refcount_table_index = i;
97 int qcow2_refcount_init(BlockDriverState *bs)
99 BDRVQcow2State *s = bs->opaque;
100 unsigned int refcount_table_size2, i;
101 int ret;
103 assert(s->refcount_order >= 0 && s->refcount_order <= 6);
105 s->get_refcount = get_refcount_funcs[s->refcount_order];
106 s->set_refcount = set_refcount_funcs[s->refcount_order];
108 assert(s->refcount_table_size <= INT_MAX / sizeof(uint64_t));
109 refcount_table_size2 = s->refcount_table_size * sizeof(uint64_t);
110 s->refcount_table = g_try_malloc(refcount_table_size2);
112 if (s->refcount_table_size > 0) {
113 if (s->refcount_table == NULL) {
114 ret = -ENOMEM;
115 goto fail;
117 BLKDBG_EVENT(bs->file, BLKDBG_REFTABLE_LOAD);
118 ret = bdrv_pread(bs->file, s->refcount_table_offset,
119 s->refcount_table, refcount_table_size2);
120 if (ret < 0) {
121 goto fail;
123 for(i = 0; i < s->refcount_table_size; i++)
124 be64_to_cpus(&s->refcount_table[i]);
125 update_max_refcount_table_index(s);
127 return 0;
128 fail:
129 return ret;
132 void qcow2_refcount_close(BlockDriverState *bs)
134 BDRVQcow2State *s = bs->opaque;
135 g_free(s->refcount_table);
139 static uint64_t get_refcount_ro0(const void *refcount_array, uint64_t index)
141 return (((const uint8_t *)refcount_array)[index / 8] >> (index % 8)) & 0x1;
144 static void set_refcount_ro0(void *refcount_array, uint64_t index,
145 uint64_t value)
147 assert(!(value >> 1));
148 ((uint8_t *)refcount_array)[index / 8] &= ~(0x1 << (index % 8));
149 ((uint8_t *)refcount_array)[index / 8] |= value << (index % 8);
152 static uint64_t get_refcount_ro1(const void *refcount_array, uint64_t index)
154 return (((const uint8_t *)refcount_array)[index / 4] >> (2 * (index % 4)))
155 & 0x3;
158 static void set_refcount_ro1(void *refcount_array, uint64_t index,
159 uint64_t value)
161 assert(!(value >> 2));
162 ((uint8_t *)refcount_array)[index / 4] &= ~(0x3 << (2 * (index % 4)));
163 ((uint8_t *)refcount_array)[index / 4] |= value << (2 * (index % 4));
166 static uint64_t get_refcount_ro2(const void *refcount_array, uint64_t index)
168 return (((const uint8_t *)refcount_array)[index / 2] >> (4 * (index % 2)))
169 & 0xf;
172 static void set_refcount_ro2(void *refcount_array, uint64_t index,
173 uint64_t value)
175 assert(!(value >> 4));
176 ((uint8_t *)refcount_array)[index / 2] &= ~(0xf << (4 * (index % 2)));
177 ((uint8_t *)refcount_array)[index / 2] |= value << (4 * (index % 2));
180 static uint64_t get_refcount_ro3(const void *refcount_array, uint64_t index)
182 return ((const uint8_t *)refcount_array)[index];
185 static void set_refcount_ro3(void *refcount_array, uint64_t index,
186 uint64_t value)
188 assert(!(value >> 8));
189 ((uint8_t *)refcount_array)[index] = value;
192 static uint64_t get_refcount_ro4(const void *refcount_array, uint64_t index)
194 return be16_to_cpu(((const uint16_t *)refcount_array)[index]);
197 static void set_refcount_ro4(void *refcount_array, uint64_t index,
198 uint64_t value)
200 assert(!(value >> 16));
201 ((uint16_t *)refcount_array)[index] = cpu_to_be16(value);
204 static uint64_t get_refcount_ro5(const void *refcount_array, uint64_t index)
206 return be32_to_cpu(((const uint32_t *)refcount_array)[index]);
209 static void set_refcount_ro5(void *refcount_array, uint64_t index,
210 uint64_t value)
212 assert(!(value >> 32));
213 ((uint32_t *)refcount_array)[index] = cpu_to_be32(value);
216 static uint64_t get_refcount_ro6(const void *refcount_array, uint64_t index)
218 return be64_to_cpu(((const uint64_t *)refcount_array)[index]);
221 static void set_refcount_ro6(void *refcount_array, uint64_t index,
222 uint64_t value)
224 ((uint64_t *)refcount_array)[index] = cpu_to_be64(value);
228 static int load_refcount_block(BlockDriverState *bs,
229 int64_t refcount_block_offset,
230 void **refcount_block)
232 BDRVQcow2State *s = bs->opaque;
234 BLKDBG_EVENT(bs->file, BLKDBG_REFBLOCK_LOAD);
235 return qcow2_cache_get(bs, s->refcount_block_cache, refcount_block_offset,
236 refcount_block);
240 * Retrieves the refcount of the cluster given by its index and stores it in
241 * *refcount. Returns 0 on success and -errno on failure.
243 int qcow2_get_refcount(BlockDriverState *bs, int64_t cluster_index,
244 uint64_t *refcount)
246 BDRVQcow2State *s = bs->opaque;
247 uint64_t refcount_table_index, block_index;
248 int64_t refcount_block_offset;
249 int ret;
250 void *refcount_block;
252 refcount_table_index = cluster_index >> s->refcount_block_bits;
253 if (refcount_table_index >= s->refcount_table_size) {
254 *refcount = 0;
255 return 0;
257 refcount_block_offset =
258 s->refcount_table[refcount_table_index] & REFT_OFFSET_MASK;
259 if (!refcount_block_offset) {
260 *refcount = 0;
261 return 0;
264 if (offset_into_cluster(s, refcount_block_offset)) {
265 qcow2_signal_corruption(bs, true, -1, -1, "Refblock offset %#" PRIx64
266 " unaligned (reftable index: %#" PRIx64 ")",
267 refcount_block_offset, refcount_table_index);
268 return -EIO;
271 ret = qcow2_cache_get(bs, s->refcount_block_cache, refcount_block_offset,
272 &refcount_block);
273 if (ret < 0) {
274 return ret;
277 block_index = cluster_index & (s->refcount_block_size - 1);
278 *refcount = s->get_refcount(refcount_block, block_index);
280 qcow2_cache_put(s->refcount_block_cache, &refcount_block);
282 return 0;
285 /* Checks if two offsets are described by the same refcount block */
286 static int in_same_refcount_block(BDRVQcow2State *s, uint64_t offset_a,
287 uint64_t offset_b)
289 uint64_t block_a = offset_a >> (s->cluster_bits + s->refcount_block_bits);
290 uint64_t block_b = offset_b >> (s->cluster_bits + s->refcount_block_bits);
292 return (block_a == block_b);
296 * Loads a refcount block. If it doesn't exist yet, it is allocated first
297 * (including growing the refcount table if needed).
299 * Returns 0 on success or -errno in error case
301 static int alloc_refcount_block(BlockDriverState *bs,
302 int64_t cluster_index, void **refcount_block)
304 BDRVQcow2State *s = bs->opaque;
305 unsigned int refcount_table_index;
306 int64_t ret;
308 BLKDBG_EVENT(bs->file, BLKDBG_REFBLOCK_ALLOC);
310 /* Find the refcount block for the given cluster */
311 refcount_table_index = cluster_index >> s->refcount_block_bits;
313 if (refcount_table_index < s->refcount_table_size) {
315 uint64_t refcount_block_offset =
316 s->refcount_table[refcount_table_index] & REFT_OFFSET_MASK;
318 /* If it's already there, we're done */
319 if (refcount_block_offset) {
320 if (offset_into_cluster(s, refcount_block_offset)) {
321 qcow2_signal_corruption(bs, true, -1, -1, "Refblock offset %#"
322 PRIx64 " unaligned (reftable index: "
323 "%#x)", refcount_block_offset,
324 refcount_table_index);
325 return -EIO;
328 return load_refcount_block(bs, refcount_block_offset,
329 refcount_block);
334 * If we came here, we need to allocate something. Something is at least
335 * a cluster for the new refcount block. It may also include a new refcount
336 * table if the old refcount table is too small.
338 * Note that allocating clusters here needs some special care:
340 * - We can't use the normal qcow2_alloc_clusters(), it would try to
341 * increase the refcount and very likely we would end up with an endless
342 * recursion. Instead we must place the refcount blocks in a way that
343 * they can describe them themselves.
345 * - We need to consider that at this point we are inside update_refcounts
346 * and potentially doing an initial refcount increase. This means that
347 * some clusters have already been allocated by the caller, but their
348 * refcount isn't accurate yet. If we allocate clusters for metadata, we
349 * need to return -EAGAIN to signal the caller that it needs to restart
350 * the search for free clusters.
352 * - alloc_clusters_noref and qcow2_free_clusters may load a different
353 * refcount block into the cache
356 *refcount_block = NULL;
358 /* We write to the refcount table, so we might depend on L2 tables */
359 ret = qcow2_cache_flush(bs, s->l2_table_cache);
360 if (ret < 0) {
361 return ret;
364 /* Allocate the refcount block itself and mark it as used */
365 int64_t new_block = alloc_clusters_noref(bs, s->cluster_size, INT64_MAX);
366 if (new_block < 0) {
367 return new_block;
370 /* The offset must fit in the offset field of the refcount table entry */
371 assert((new_block & REFT_OFFSET_MASK) == new_block);
373 /* If we're allocating the block at offset 0 then something is wrong */
374 if (new_block == 0) {
375 qcow2_signal_corruption(bs, true, -1, -1, "Preventing invalid "
376 "allocation of refcount block at offset 0");
377 return -EIO;
380 #ifdef DEBUG_ALLOC2
381 fprintf(stderr, "qcow2: Allocate refcount block %d for %" PRIx64
382 " at %" PRIx64 "\n",
383 refcount_table_index, cluster_index << s->cluster_bits, new_block);
384 #endif
386 if (in_same_refcount_block(s, new_block, cluster_index << s->cluster_bits)) {
387 /* Zero the new refcount block before updating it */
388 ret = qcow2_cache_get_empty(bs, s->refcount_block_cache, new_block,
389 refcount_block);
390 if (ret < 0) {
391 goto fail;
394 memset(*refcount_block, 0, s->cluster_size);
396 /* The block describes itself, need to update the cache */
397 int block_index = (new_block >> s->cluster_bits) &
398 (s->refcount_block_size - 1);
399 s->set_refcount(*refcount_block, block_index, 1);
400 } else {
401 /* Described somewhere else. This can recurse at most twice before we
402 * arrive at a block that describes itself. */
403 ret = update_refcount(bs, new_block, s->cluster_size, 1, false,
404 QCOW2_DISCARD_NEVER);
405 if (ret < 0) {
406 goto fail;
409 ret = qcow2_cache_flush(bs, s->refcount_block_cache);
410 if (ret < 0) {
411 goto fail;
414 /* Initialize the new refcount block only after updating its refcount,
415 * update_refcount uses the refcount cache itself */
416 ret = qcow2_cache_get_empty(bs, s->refcount_block_cache, new_block,
417 refcount_block);
418 if (ret < 0) {
419 goto fail;
422 memset(*refcount_block, 0, s->cluster_size);
425 /* Now the new refcount block needs to be written to disk */
426 BLKDBG_EVENT(bs->file, BLKDBG_REFBLOCK_ALLOC_WRITE);
427 qcow2_cache_entry_mark_dirty(s->refcount_block_cache, *refcount_block);
428 ret = qcow2_cache_flush(bs, s->refcount_block_cache);
429 if (ret < 0) {
430 goto fail;
433 /* If the refcount table is big enough, just hook the block up there */
434 if (refcount_table_index < s->refcount_table_size) {
435 uint64_t data64 = cpu_to_be64(new_block);
436 BLKDBG_EVENT(bs->file, BLKDBG_REFBLOCK_ALLOC_HOOKUP);
437 ret = bdrv_pwrite_sync(bs->file,
438 s->refcount_table_offset + refcount_table_index * sizeof(uint64_t),
439 &data64, sizeof(data64));
440 if (ret < 0) {
441 goto fail;
444 s->refcount_table[refcount_table_index] = new_block;
445 /* If there's a hole in s->refcount_table then it can happen
446 * that refcount_table_index < s->max_refcount_table_index */
447 s->max_refcount_table_index =
448 MAX(s->max_refcount_table_index, refcount_table_index);
450 /* The new refcount block may be where the caller intended to put its
451 * data, so let it restart the search. */
452 return -EAGAIN;
455 qcow2_cache_put(s->refcount_block_cache, refcount_block);
458 * If we come here, we need to grow the refcount table. Again, a new
459 * refcount table needs some space and we can't simply allocate to avoid
460 * endless recursion.
462 * Therefore let's grab new refcount blocks at the end of the image, which
463 * will describe themselves and the new refcount table. This way we can
464 * reference them only in the new table and do the switch to the new
465 * refcount table at once without producing an inconsistent state in
466 * between.
468 BLKDBG_EVENT(bs->file, BLKDBG_REFTABLE_GROW);
470 /* Calculate the number of refcount blocks needed so far; this will be the
471 * basis for calculating the index of the first cluster used for the
472 * self-describing refcount structures which we are about to create.
474 * Because we reached this point, there cannot be any refcount entries for
475 * cluster_index or higher indices yet. However, because new_block has been
476 * allocated to describe that cluster (and it will assume this role later
477 * on), we cannot use that index; also, new_block may actually have a higher
478 * cluster index than cluster_index, so it needs to be taken into account
479 * here (and 1 needs to be added to its value because that cluster is used).
481 uint64_t blocks_used = DIV_ROUND_UP(MAX(cluster_index + 1,
482 (new_block >> s->cluster_bits) + 1),
483 s->refcount_block_size);
485 /* Create the new refcount table and blocks */
486 uint64_t meta_offset = (blocks_used * s->refcount_block_size) *
487 s->cluster_size;
489 ret = qcow2_refcount_area(bs, meta_offset, 0, false,
490 refcount_table_index, new_block);
491 if (ret < 0) {
492 return ret;
495 ret = load_refcount_block(bs, new_block, refcount_block);
496 if (ret < 0) {
497 return ret;
500 /* If we were trying to do the initial refcount update for some cluster
501 * allocation, we might have used the same clusters to store newly
502 * allocated metadata. Make the caller search some new space. */
503 return -EAGAIN;
505 fail:
506 if (*refcount_block != NULL) {
507 qcow2_cache_put(s->refcount_block_cache, refcount_block);
509 return ret;
513 * Starting at @start_offset, this function creates new self-covering refcount
514 * structures: A new refcount table and refcount blocks which cover all of
515 * themselves, and a number of @additional_clusters beyond their end.
516 * @start_offset must be at the end of the image file, that is, there must be
517 * only empty space beyond it.
518 * If @exact_size is false, the refcount table will have 50 % more entries than
519 * necessary so it will not need to grow again soon.
520 * If @new_refblock_offset is not zero, it contains the offset of a refcount
521 * block that should be entered into the new refcount table at index
522 * @new_refblock_index.
524 * Returns: The offset after the new refcount structures (i.e. where the
525 * @additional_clusters may be placed) on success, -errno on error.
527 int64_t qcow2_refcount_area(BlockDriverState *bs, uint64_t start_offset,
528 uint64_t additional_clusters, bool exact_size,
529 int new_refblock_index,
530 uint64_t new_refblock_offset)
532 BDRVQcow2State *s = bs->opaque;
533 uint64_t total_refblock_count_u64, additional_refblock_count;
534 int total_refblock_count, table_size, area_reftable_index, table_clusters;
535 int i;
536 uint64_t table_offset, block_offset, end_offset;
537 int ret;
538 uint64_t *new_table;
540 assert(!(start_offset % s->cluster_size));
542 qcow2_refcount_metadata_size(start_offset / s->cluster_size +
543 additional_clusters,
544 s->cluster_size, s->refcount_order,
545 !exact_size, &total_refblock_count_u64);
546 if (total_refblock_count_u64 > QCOW_MAX_REFTABLE_SIZE) {
547 return -EFBIG;
549 total_refblock_count = total_refblock_count_u64;
551 /* Index in the refcount table of the first refcount block to cover the area
552 * of refcount structures we are about to create; we know that
553 * @total_refblock_count can cover @start_offset, so this will definitely
554 * fit into an int. */
555 area_reftable_index = (start_offset / s->cluster_size) /
556 s->refcount_block_size;
558 if (exact_size) {
559 table_size = total_refblock_count;
560 } else {
561 table_size = total_refblock_count +
562 DIV_ROUND_UP(total_refblock_count, 2);
564 /* The qcow2 file can only store the reftable size in number of clusters */
565 table_size = ROUND_UP(table_size, s->cluster_size / sizeof(uint64_t));
566 table_clusters = (table_size * sizeof(uint64_t)) / s->cluster_size;
568 if (table_size > QCOW_MAX_REFTABLE_SIZE) {
569 return -EFBIG;
572 new_table = g_try_new0(uint64_t, table_size);
574 assert(table_size > 0);
575 if (new_table == NULL) {
576 ret = -ENOMEM;
577 goto fail;
580 /* Fill the new refcount table */
581 if (table_size > s->max_refcount_table_index) {
582 /* We're actually growing the reftable */
583 memcpy(new_table, s->refcount_table,
584 (s->max_refcount_table_index + 1) * sizeof(uint64_t));
585 } else {
586 /* Improbable case: We're shrinking the reftable. However, the caller
587 * has assured us that there is only empty space beyond @start_offset,
588 * so we can simply drop all of the refblocks that won't fit into the
589 * new reftable. */
590 memcpy(new_table, s->refcount_table, table_size * sizeof(uint64_t));
593 if (new_refblock_offset) {
594 assert(new_refblock_index < total_refblock_count);
595 new_table[new_refblock_index] = new_refblock_offset;
598 /* Count how many new refblocks we have to create */
599 additional_refblock_count = 0;
600 for (i = area_reftable_index; i < total_refblock_count; i++) {
601 if (!new_table[i]) {
602 additional_refblock_count++;
606 table_offset = start_offset + additional_refblock_count * s->cluster_size;
607 end_offset = table_offset + table_clusters * s->cluster_size;
609 /* Fill the refcount blocks, and create new ones, if necessary */
610 block_offset = start_offset;
611 for (i = area_reftable_index; i < total_refblock_count; i++) {
612 void *refblock_data;
613 uint64_t first_offset_covered;
615 /* Reuse an existing refblock if possible, create a new one otherwise */
616 if (new_table[i]) {
617 ret = qcow2_cache_get(bs, s->refcount_block_cache, new_table[i],
618 &refblock_data);
619 if (ret < 0) {
620 goto fail;
622 } else {
623 ret = qcow2_cache_get_empty(bs, s->refcount_block_cache,
624 block_offset, &refblock_data);
625 if (ret < 0) {
626 goto fail;
628 memset(refblock_data, 0, s->cluster_size);
629 qcow2_cache_entry_mark_dirty(s->refcount_block_cache,
630 refblock_data);
632 new_table[i] = block_offset;
633 block_offset += s->cluster_size;
636 /* First host offset covered by this refblock */
637 first_offset_covered = (uint64_t)i * s->refcount_block_size *
638 s->cluster_size;
639 if (first_offset_covered < end_offset) {
640 int j, end_index;
642 /* Set the refcount of all of the new refcount structures to 1 */
644 if (first_offset_covered < start_offset) {
645 assert(i == area_reftable_index);
646 j = (start_offset - first_offset_covered) / s->cluster_size;
647 assert(j < s->refcount_block_size);
648 } else {
649 j = 0;
652 end_index = MIN((end_offset - first_offset_covered) /
653 s->cluster_size,
654 s->refcount_block_size);
656 for (; j < end_index; j++) {
657 /* The caller guaranteed us this space would be empty */
658 assert(s->get_refcount(refblock_data, j) == 0);
659 s->set_refcount(refblock_data, j, 1);
662 qcow2_cache_entry_mark_dirty(s->refcount_block_cache,
663 refblock_data);
666 qcow2_cache_put(s->refcount_block_cache, &refblock_data);
669 assert(block_offset == table_offset);
671 /* Write refcount blocks to disk */
672 BLKDBG_EVENT(bs->file, BLKDBG_REFBLOCK_ALLOC_WRITE_BLOCKS);
673 ret = qcow2_cache_flush(bs, s->refcount_block_cache);
674 if (ret < 0) {
675 goto fail;
678 /* Write refcount table to disk */
679 for (i = 0; i < total_refblock_count; i++) {
680 cpu_to_be64s(&new_table[i]);
683 BLKDBG_EVENT(bs->file, BLKDBG_REFBLOCK_ALLOC_WRITE_TABLE);
684 ret = bdrv_pwrite_sync(bs->file, table_offset, new_table,
685 table_size * sizeof(uint64_t));
686 if (ret < 0) {
687 goto fail;
690 for (i = 0; i < total_refblock_count; i++) {
691 be64_to_cpus(&new_table[i]);
694 /* Hook up the new refcount table in the qcow2 header */
695 struct QEMU_PACKED {
696 uint64_t d64;
697 uint32_t d32;
698 } data;
699 data.d64 = cpu_to_be64(table_offset);
700 data.d32 = cpu_to_be32(table_clusters);
701 BLKDBG_EVENT(bs->file, BLKDBG_REFBLOCK_ALLOC_SWITCH_TABLE);
702 ret = bdrv_pwrite_sync(bs->file,
703 offsetof(QCowHeader, refcount_table_offset),
704 &data, sizeof(data));
705 if (ret < 0) {
706 goto fail;
709 /* And switch it in memory */
710 uint64_t old_table_offset = s->refcount_table_offset;
711 uint64_t old_table_size = s->refcount_table_size;
713 g_free(s->refcount_table);
714 s->refcount_table = new_table;
715 s->refcount_table_size = table_size;
716 s->refcount_table_offset = table_offset;
717 update_max_refcount_table_index(s);
719 /* Free old table. */
720 qcow2_free_clusters(bs, old_table_offset, old_table_size * sizeof(uint64_t),
721 QCOW2_DISCARD_OTHER);
723 return end_offset;
725 fail:
726 g_free(new_table);
727 return ret;
730 void qcow2_process_discards(BlockDriverState *bs, int ret)
732 BDRVQcow2State *s = bs->opaque;
733 Qcow2DiscardRegion *d, *next;
735 QTAILQ_FOREACH_SAFE(d, &s->discards, next, next) {
736 QTAILQ_REMOVE(&s->discards, d, next);
738 /* Discard is optional, ignore the return value */
739 if (ret >= 0) {
740 int r2 = bdrv_pdiscard(bs->file, d->offset, d->bytes);
741 if (r2 < 0) {
742 trace_qcow2_process_discards_failed_region(d->offset, d->bytes,
743 r2);
747 g_free(d);
751 static void update_refcount_discard(BlockDriverState *bs,
752 uint64_t offset, uint64_t length)
754 BDRVQcow2State *s = bs->opaque;
755 Qcow2DiscardRegion *d, *p, *next;
757 QTAILQ_FOREACH(d, &s->discards, next) {
758 uint64_t new_start = MIN(offset, d->offset);
759 uint64_t new_end = MAX(offset + length, d->offset + d->bytes);
761 if (new_end - new_start <= length + d->bytes) {
762 /* There can't be any overlap, areas ending up here have no
763 * references any more and therefore shouldn't get freed another
764 * time. */
765 assert(d->bytes + length == new_end - new_start);
766 d->offset = new_start;
767 d->bytes = new_end - new_start;
768 goto found;
772 d = g_malloc(sizeof(*d));
773 *d = (Qcow2DiscardRegion) {
774 .bs = bs,
775 .offset = offset,
776 .bytes = length,
778 QTAILQ_INSERT_TAIL(&s->discards, d, next);
780 found:
781 /* Merge discard requests if they are adjacent now */
782 QTAILQ_FOREACH_SAFE(p, &s->discards, next, next) {
783 if (p == d
784 || p->offset > d->offset + d->bytes
785 || d->offset > p->offset + p->bytes)
787 continue;
790 /* Still no overlap possible */
791 assert(p->offset == d->offset + d->bytes
792 || d->offset == p->offset + p->bytes);
794 QTAILQ_REMOVE(&s->discards, p, next);
795 d->offset = MIN(d->offset, p->offset);
796 d->bytes += p->bytes;
797 g_free(p);
801 /* XXX: cache several refcount block clusters ? */
802 /* @addend is the absolute value of the addend; if @decrease is set, @addend
803 * will be subtracted from the current refcount, otherwise it will be added */
804 static int QEMU_WARN_UNUSED_RESULT update_refcount(BlockDriverState *bs,
805 int64_t offset,
806 int64_t length,
807 uint64_t addend,
808 bool decrease,
809 enum qcow2_discard_type type)
811 BDRVQcow2State *s = bs->opaque;
812 int64_t start, last, cluster_offset;
813 void *refcount_block = NULL;
814 int64_t old_table_index = -1;
815 int ret;
817 #ifdef DEBUG_ALLOC2
818 fprintf(stderr, "update_refcount: offset=%" PRId64 " size=%" PRId64
819 " addend=%s%" PRIu64 "\n", offset, length, decrease ? "-" : "",
820 addend);
821 #endif
822 if (length < 0) {
823 return -EINVAL;
824 } else if (length == 0) {
825 return 0;
828 if (decrease) {
829 qcow2_cache_set_dependency(bs, s->refcount_block_cache,
830 s->l2_table_cache);
833 start = start_of_cluster(s, offset);
834 last = start_of_cluster(s, offset + length - 1);
835 for(cluster_offset = start; cluster_offset <= last;
836 cluster_offset += s->cluster_size)
838 int block_index;
839 uint64_t refcount;
840 int64_t cluster_index = cluster_offset >> s->cluster_bits;
841 int64_t table_index = cluster_index >> s->refcount_block_bits;
843 /* Load the refcount block and allocate it if needed */
844 if (table_index != old_table_index) {
845 if (refcount_block) {
846 qcow2_cache_put(s->refcount_block_cache, &refcount_block);
848 ret = alloc_refcount_block(bs, cluster_index, &refcount_block);
849 /* If the caller needs to restart the search for free clusters,
850 * try the same ones first to see if they're still free. */
851 if (ret == -EAGAIN) {
852 if (s->free_cluster_index > (start >> s->cluster_bits)) {
853 s->free_cluster_index = (start >> s->cluster_bits);
856 if (ret < 0) {
857 goto fail;
860 old_table_index = table_index;
862 qcow2_cache_entry_mark_dirty(s->refcount_block_cache, refcount_block);
864 /* we can update the count and save it */
865 block_index = cluster_index & (s->refcount_block_size - 1);
867 refcount = s->get_refcount(refcount_block, block_index);
868 if (decrease ? (refcount - addend > refcount)
869 : (refcount + addend < refcount ||
870 refcount + addend > s->refcount_max))
872 ret = -EINVAL;
873 goto fail;
875 if (decrease) {
876 refcount -= addend;
877 } else {
878 refcount += addend;
880 if (refcount == 0 && cluster_index < s->free_cluster_index) {
881 s->free_cluster_index = cluster_index;
883 s->set_refcount(refcount_block, block_index, refcount);
885 if (refcount == 0) {
886 void *table;
888 table = qcow2_cache_is_table_offset(s->refcount_block_cache,
889 offset);
890 if (table != NULL) {
891 qcow2_cache_put(s->refcount_block_cache, &refcount_block);
892 qcow2_cache_discard(s->refcount_block_cache, table);
895 table = qcow2_cache_is_table_offset(s->l2_table_cache, offset);
896 if (table != NULL) {
897 qcow2_cache_discard(s->l2_table_cache, table);
900 if (s->discard_passthrough[type]) {
901 update_refcount_discard(bs, cluster_offset, s->cluster_size);
906 ret = 0;
907 fail:
908 if (!s->cache_discards) {
909 qcow2_process_discards(bs, ret);
912 /* Write last changed block to disk */
913 if (refcount_block) {
914 qcow2_cache_put(s->refcount_block_cache, &refcount_block);
918 * Try do undo any updates if an error is returned (This may succeed in
919 * some cases like ENOSPC for allocating a new refcount block)
921 if (ret < 0) {
922 int dummy;
923 dummy = update_refcount(bs, offset, cluster_offset - offset, addend,
924 !decrease, QCOW2_DISCARD_NEVER);
925 (void)dummy;
928 return ret;
932 * Increases or decreases the refcount of a given cluster.
934 * @addend is the absolute value of the addend; if @decrease is set, @addend
935 * will be subtracted from the current refcount, otherwise it will be added.
937 * On success 0 is returned; on failure -errno is returned.
939 int qcow2_update_cluster_refcount(BlockDriverState *bs,
940 int64_t cluster_index,
941 uint64_t addend, bool decrease,
942 enum qcow2_discard_type type)
944 BDRVQcow2State *s = bs->opaque;
945 int ret;
947 ret = update_refcount(bs, cluster_index << s->cluster_bits, 1, addend,
948 decrease, type);
949 if (ret < 0) {
950 return ret;
953 return 0;
958 /*********************************************************/
959 /* cluster allocation functions */
963 /* return < 0 if error */
964 static int64_t alloc_clusters_noref(BlockDriverState *bs, uint64_t size,
965 uint64_t max)
967 BDRVQcow2State *s = bs->opaque;
968 uint64_t i, nb_clusters, refcount;
969 int ret;
971 /* We can't allocate clusters if they may still be queued for discard. */
972 if (s->cache_discards) {
973 qcow2_process_discards(bs, 0);
976 nb_clusters = size_to_clusters(s, size);
977 retry:
978 for(i = 0; i < nb_clusters; i++) {
979 uint64_t next_cluster_index = s->free_cluster_index++;
980 ret = qcow2_get_refcount(bs, next_cluster_index, &refcount);
982 if (ret < 0) {
983 return ret;
984 } else if (refcount != 0) {
985 goto retry;
989 /* Make sure that all offsets in the "allocated" range are representable
990 * in the requested max */
991 if (s->free_cluster_index > 0 &&
992 s->free_cluster_index - 1 > (max >> s->cluster_bits))
994 return -EFBIG;
997 #ifdef DEBUG_ALLOC2
998 fprintf(stderr, "alloc_clusters: size=%" PRId64 " -> %" PRId64 "\n",
999 size,
1000 (s->free_cluster_index - nb_clusters) << s->cluster_bits);
1001 #endif
1002 return (s->free_cluster_index - nb_clusters) << s->cluster_bits;
1005 int64_t qcow2_alloc_clusters(BlockDriverState *bs, uint64_t size)
1007 int64_t offset;
1008 int ret;
1010 BLKDBG_EVENT(bs->file, BLKDBG_CLUSTER_ALLOC);
1011 do {
1012 offset = alloc_clusters_noref(bs, size, QCOW_MAX_CLUSTER_OFFSET);
1013 if (offset < 0) {
1014 return offset;
1017 ret = update_refcount(bs, offset, size, 1, false, QCOW2_DISCARD_NEVER);
1018 } while (ret == -EAGAIN);
1020 if (ret < 0) {
1021 return ret;
1024 return offset;
1027 int64_t qcow2_alloc_clusters_at(BlockDriverState *bs, uint64_t offset,
1028 int64_t nb_clusters)
1030 BDRVQcow2State *s = bs->opaque;
1031 uint64_t cluster_index, refcount;
1032 uint64_t i;
1033 int ret;
1035 assert(nb_clusters >= 0);
1036 if (nb_clusters == 0) {
1037 return 0;
1040 do {
1041 /* Check how many clusters there are free */
1042 cluster_index = offset >> s->cluster_bits;
1043 for(i = 0; i < nb_clusters; i++) {
1044 ret = qcow2_get_refcount(bs, cluster_index++, &refcount);
1045 if (ret < 0) {
1046 return ret;
1047 } else if (refcount != 0) {
1048 break;
1052 /* And then allocate them */
1053 ret = update_refcount(bs, offset, i << s->cluster_bits, 1, false,
1054 QCOW2_DISCARD_NEVER);
1055 } while (ret == -EAGAIN);
1057 if (ret < 0) {
1058 return ret;
1061 return i;
1064 /* only used to allocate compressed sectors. We try to allocate
1065 contiguous sectors. size must be <= cluster_size */
1066 int64_t qcow2_alloc_bytes(BlockDriverState *bs, int size)
1068 BDRVQcow2State *s = bs->opaque;
1069 int64_t offset;
1070 size_t free_in_cluster;
1071 int ret;
1073 BLKDBG_EVENT(bs->file, BLKDBG_CLUSTER_ALLOC_BYTES);
1074 assert(size > 0 && size <= s->cluster_size);
1075 assert(!s->free_byte_offset || offset_into_cluster(s, s->free_byte_offset));
1077 offset = s->free_byte_offset;
1079 if (offset) {
1080 uint64_t refcount;
1081 ret = qcow2_get_refcount(bs, offset >> s->cluster_bits, &refcount);
1082 if (ret < 0) {
1083 return ret;
1086 if (refcount == s->refcount_max) {
1087 offset = 0;
1091 free_in_cluster = s->cluster_size - offset_into_cluster(s, offset);
1092 do {
1093 if (!offset || free_in_cluster < size) {
1094 int64_t new_cluster;
1096 new_cluster = alloc_clusters_noref(bs, s->cluster_size,
1097 MIN(s->cluster_offset_mask,
1098 QCOW_MAX_CLUSTER_OFFSET));
1099 if (new_cluster < 0) {
1100 return new_cluster;
1103 if (new_cluster == 0) {
1104 qcow2_signal_corruption(bs, true, -1, -1, "Preventing invalid "
1105 "allocation of compressed cluster "
1106 "at offset 0");
1107 return -EIO;
1110 if (!offset || ROUND_UP(offset, s->cluster_size) != new_cluster) {
1111 offset = new_cluster;
1112 free_in_cluster = s->cluster_size;
1113 } else {
1114 free_in_cluster += s->cluster_size;
1118 assert(offset);
1119 ret = update_refcount(bs, offset, size, 1, false, QCOW2_DISCARD_NEVER);
1120 if (ret < 0) {
1121 offset = 0;
1123 } while (ret == -EAGAIN);
1124 if (ret < 0) {
1125 return ret;
1128 /* The cluster refcount was incremented; refcount blocks must be flushed
1129 * before the caller's L2 table updates. */
1130 qcow2_cache_set_dependency(bs, s->l2_table_cache, s->refcount_block_cache);
1132 s->free_byte_offset = offset + size;
1133 if (!offset_into_cluster(s, s->free_byte_offset)) {
1134 s->free_byte_offset = 0;
1137 return offset;
1140 void qcow2_free_clusters(BlockDriverState *bs,
1141 int64_t offset, int64_t size,
1142 enum qcow2_discard_type type)
1144 int ret;
1146 BLKDBG_EVENT(bs->file, BLKDBG_CLUSTER_FREE);
1147 ret = update_refcount(bs, offset, size, 1, true, type);
1148 if (ret < 0) {
1149 fprintf(stderr, "qcow2_free_clusters failed: %s\n", strerror(-ret));
1150 /* TODO Remember the clusters to free them later and avoid leaking */
1155 * Free a cluster using its L2 entry (handles clusters of all types, e.g.
1156 * normal cluster, compressed cluster, etc.)
1158 void qcow2_free_any_clusters(BlockDriverState *bs, uint64_t l2_entry,
1159 int nb_clusters, enum qcow2_discard_type type)
1161 BDRVQcow2State *s = bs->opaque;
1162 QCow2ClusterType ctype = qcow2_get_cluster_type(bs, l2_entry);
1164 if (has_data_file(bs)) {
1165 if (s->discard_passthrough[type] &&
1166 (ctype == QCOW2_CLUSTER_NORMAL ||
1167 ctype == QCOW2_CLUSTER_ZERO_ALLOC))
1169 bdrv_pdiscard(s->data_file, l2_entry & L2E_OFFSET_MASK,
1170 nb_clusters << s->cluster_bits);
1172 return;
1175 switch (ctype) {
1176 case QCOW2_CLUSTER_COMPRESSED:
1178 int64_t offset = (l2_entry & s->cluster_offset_mask)
1179 & QCOW2_COMPRESSED_SECTOR_MASK;
1180 int size = QCOW2_COMPRESSED_SECTOR_SIZE *
1181 (((l2_entry >> s->csize_shift) & s->csize_mask) + 1);
1182 qcow2_free_clusters(bs, offset, size, type);
1184 break;
1185 case QCOW2_CLUSTER_NORMAL:
1186 case QCOW2_CLUSTER_ZERO_ALLOC:
1187 if (offset_into_cluster(s, l2_entry & L2E_OFFSET_MASK)) {
1188 qcow2_signal_corruption(bs, false, -1, -1,
1189 "Cannot free unaligned cluster %#llx",
1190 l2_entry & L2E_OFFSET_MASK);
1191 } else {
1192 qcow2_free_clusters(bs, l2_entry & L2E_OFFSET_MASK,
1193 nb_clusters << s->cluster_bits, type);
1195 break;
1196 case QCOW2_CLUSTER_ZERO_PLAIN:
1197 case QCOW2_CLUSTER_UNALLOCATED:
1198 break;
1199 default:
1200 abort();
1204 int coroutine_fn qcow2_write_caches(BlockDriverState *bs)
1206 BDRVQcow2State *s = bs->opaque;
1207 int ret;
1209 ret = qcow2_cache_write(bs, s->l2_table_cache);
1210 if (ret < 0) {
1211 return ret;
1214 if (qcow2_need_accurate_refcounts(s)) {
1215 ret = qcow2_cache_write(bs, s->refcount_block_cache);
1216 if (ret < 0) {
1217 return ret;
1221 return 0;
1224 int coroutine_fn qcow2_flush_caches(BlockDriverState *bs)
1226 int ret = qcow2_write_caches(bs);
1227 if (ret < 0) {
1228 return ret;
1231 return bdrv_flush(bs->file->bs);
1234 /*********************************************************/
1235 /* snapshots and image creation */
1239 /* update the refcounts of snapshots and the copied flag */
1240 int qcow2_update_snapshot_refcount(BlockDriverState *bs,
1241 int64_t l1_table_offset, int l1_size, int addend)
1243 BDRVQcow2State *s = bs->opaque;
1244 uint64_t *l1_table, *l2_slice, l2_offset, entry, l1_size2, refcount;
1245 bool l1_allocated = false;
1246 int64_t old_entry, old_l2_offset;
1247 unsigned slice, slice_size2, n_slices;
1248 int i, j, l1_modified = 0, nb_csectors;
1249 int ret;
1251 assert(addend >= -1 && addend <= 1);
1253 l2_slice = NULL;
1254 l1_table = NULL;
1255 l1_size2 = l1_size * sizeof(uint64_t);
1256 slice_size2 = s->l2_slice_size * sizeof(uint64_t);
1257 n_slices = s->cluster_size / slice_size2;
1259 s->cache_discards = true;
1261 /* WARNING: qcow2_snapshot_goto relies on this function not using the
1262 * l1_table_offset when it is the current s->l1_table_offset! Be careful
1263 * when changing this! */
1264 if (l1_table_offset != s->l1_table_offset) {
1265 l1_table = g_try_malloc0(ROUND_UP(l1_size2, 512));
1266 if (l1_size2 && l1_table == NULL) {
1267 ret = -ENOMEM;
1268 goto fail;
1270 l1_allocated = true;
1272 ret = bdrv_pread(bs->file, l1_table_offset, l1_table, l1_size2);
1273 if (ret < 0) {
1274 goto fail;
1277 for (i = 0; i < l1_size; i++) {
1278 be64_to_cpus(&l1_table[i]);
1280 } else {
1281 assert(l1_size == s->l1_size);
1282 l1_table = s->l1_table;
1283 l1_allocated = false;
1286 for (i = 0; i < l1_size; i++) {
1287 l2_offset = l1_table[i];
1288 if (l2_offset) {
1289 old_l2_offset = l2_offset;
1290 l2_offset &= L1E_OFFSET_MASK;
1292 if (offset_into_cluster(s, l2_offset)) {
1293 qcow2_signal_corruption(bs, true, -1, -1, "L2 table offset %#"
1294 PRIx64 " unaligned (L1 index: %#x)",
1295 l2_offset, i);
1296 ret = -EIO;
1297 goto fail;
1300 for (slice = 0; slice < n_slices; slice++) {
1301 ret = qcow2_cache_get(bs, s->l2_table_cache,
1302 l2_offset + slice * slice_size2,
1303 (void **) &l2_slice);
1304 if (ret < 0) {
1305 goto fail;
1308 for (j = 0; j < s->l2_slice_size; j++) {
1309 uint64_t cluster_index;
1310 uint64_t offset;
1312 entry = be64_to_cpu(l2_slice[j]);
1313 old_entry = entry;
1314 entry &= ~QCOW_OFLAG_COPIED;
1315 offset = entry & L2E_OFFSET_MASK;
1317 switch (qcow2_get_cluster_type(bs, entry)) {
1318 case QCOW2_CLUSTER_COMPRESSED:
1319 nb_csectors = ((entry >> s->csize_shift) &
1320 s->csize_mask) + 1;
1321 if (addend != 0) {
1322 uint64_t coffset = (entry & s->cluster_offset_mask)
1323 & QCOW2_COMPRESSED_SECTOR_MASK;
1324 ret = update_refcount(
1325 bs, coffset,
1326 nb_csectors * QCOW2_COMPRESSED_SECTOR_SIZE,
1327 abs(addend), addend < 0,
1328 QCOW2_DISCARD_SNAPSHOT);
1329 if (ret < 0) {
1330 goto fail;
1333 /* compressed clusters are never modified */
1334 refcount = 2;
1335 break;
1337 case QCOW2_CLUSTER_NORMAL:
1338 case QCOW2_CLUSTER_ZERO_ALLOC:
1339 if (offset_into_cluster(s, offset)) {
1340 /* Here l2_index means table (not slice) index */
1341 int l2_index = slice * s->l2_slice_size + j;
1342 qcow2_signal_corruption(
1343 bs, true, -1, -1, "Cluster "
1344 "allocation offset %#" PRIx64
1345 " unaligned (L2 offset: %#"
1346 PRIx64 ", L2 index: %#x)",
1347 offset, l2_offset, l2_index);
1348 ret = -EIO;
1349 goto fail;
1352 cluster_index = offset >> s->cluster_bits;
1353 assert(cluster_index);
1354 if (addend != 0) {
1355 ret = qcow2_update_cluster_refcount(
1356 bs, cluster_index, abs(addend), addend < 0,
1357 QCOW2_DISCARD_SNAPSHOT);
1358 if (ret < 0) {
1359 goto fail;
1363 ret = qcow2_get_refcount(bs, cluster_index, &refcount);
1364 if (ret < 0) {
1365 goto fail;
1367 break;
1369 case QCOW2_CLUSTER_ZERO_PLAIN:
1370 case QCOW2_CLUSTER_UNALLOCATED:
1371 refcount = 0;
1372 break;
1374 default:
1375 abort();
1378 if (refcount == 1) {
1379 entry |= QCOW_OFLAG_COPIED;
1381 if (entry != old_entry) {
1382 if (addend > 0) {
1383 qcow2_cache_set_dependency(bs, s->l2_table_cache,
1384 s->refcount_block_cache);
1386 l2_slice[j] = cpu_to_be64(entry);
1387 qcow2_cache_entry_mark_dirty(s->l2_table_cache,
1388 l2_slice);
1392 qcow2_cache_put(s->l2_table_cache, (void **) &l2_slice);
1395 if (addend != 0) {
1396 ret = qcow2_update_cluster_refcount(bs, l2_offset >>
1397 s->cluster_bits,
1398 abs(addend), addend < 0,
1399 QCOW2_DISCARD_SNAPSHOT);
1400 if (ret < 0) {
1401 goto fail;
1404 ret = qcow2_get_refcount(bs, l2_offset >> s->cluster_bits,
1405 &refcount);
1406 if (ret < 0) {
1407 goto fail;
1408 } else if (refcount == 1) {
1409 l2_offset |= QCOW_OFLAG_COPIED;
1411 if (l2_offset != old_l2_offset) {
1412 l1_table[i] = l2_offset;
1413 l1_modified = 1;
1418 ret = bdrv_flush(bs);
1419 fail:
1420 if (l2_slice) {
1421 qcow2_cache_put(s->l2_table_cache, (void **) &l2_slice);
1424 s->cache_discards = false;
1425 qcow2_process_discards(bs, ret);
1427 /* Update L1 only if it isn't deleted anyway (addend = -1) */
1428 if (ret == 0 && addend >= 0 && l1_modified) {
1429 for (i = 0; i < l1_size; i++) {
1430 cpu_to_be64s(&l1_table[i]);
1433 ret = bdrv_pwrite_sync(bs->file, l1_table_offset,
1434 l1_table, l1_size2);
1436 for (i = 0; i < l1_size; i++) {
1437 be64_to_cpus(&l1_table[i]);
1440 if (l1_allocated)
1441 g_free(l1_table);
1442 return ret;
1448 /*********************************************************/
1449 /* refcount checking functions */
1452 static uint64_t refcount_array_byte_size(BDRVQcow2State *s, uint64_t entries)
1454 /* This assertion holds because there is no way we can address more than
1455 * 2^(64 - 9) clusters at once (with cluster size 512 = 2^9, and because
1456 * offsets have to be representable in bytes); due to every cluster
1457 * corresponding to one refcount entry, we are well below that limit */
1458 assert(entries < (UINT64_C(1) << (64 - 9)));
1460 /* Thanks to the assertion this will not overflow, because
1461 * s->refcount_order < 7.
1462 * (note: x << s->refcount_order == x * s->refcount_bits) */
1463 return DIV_ROUND_UP(entries << s->refcount_order, 8);
1467 * Reallocates *array so that it can hold new_size entries. *size must contain
1468 * the current number of entries in *array. If the reallocation fails, *array
1469 * and *size will not be modified and -errno will be returned. If the
1470 * reallocation is successful, *array will be set to the new buffer, *size
1471 * will be set to new_size and 0 will be returned. The size of the reallocated
1472 * refcount array buffer will be aligned to a cluster boundary, and the newly
1473 * allocated area will be zeroed.
1475 static int realloc_refcount_array(BDRVQcow2State *s, void **array,
1476 int64_t *size, int64_t new_size)
1478 int64_t old_byte_size, new_byte_size;
1479 void *new_ptr;
1481 /* Round to clusters so the array can be directly written to disk */
1482 old_byte_size = size_to_clusters(s, refcount_array_byte_size(s, *size))
1483 * s->cluster_size;
1484 new_byte_size = size_to_clusters(s, refcount_array_byte_size(s, new_size))
1485 * s->cluster_size;
1487 if (new_byte_size == old_byte_size) {
1488 *size = new_size;
1489 return 0;
1492 assert(new_byte_size > 0);
1494 if (new_byte_size > SIZE_MAX) {
1495 return -ENOMEM;
1498 new_ptr = g_try_realloc(*array, new_byte_size);
1499 if (!new_ptr) {
1500 return -ENOMEM;
1503 if (new_byte_size > old_byte_size) {
1504 memset((char *)new_ptr + old_byte_size, 0,
1505 new_byte_size - old_byte_size);
1508 *array = new_ptr;
1509 *size = new_size;
1511 return 0;
1515 * Increases the refcount for a range of clusters in a given refcount table.
1516 * This is used to construct a temporary refcount table out of L1 and L2 tables
1517 * which can be compared to the refcount table saved in the image.
1519 * Modifies the number of errors in res.
1521 int qcow2_inc_refcounts_imrt(BlockDriverState *bs, BdrvCheckResult *res,
1522 void **refcount_table,
1523 int64_t *refcount_table_size,
1524 int64_t offset, int64_t size)
1526 BDRVQcow2State *s = bs->opaque;
1527 uint64_t start, last, cluster_offset, k, refcount;
1528 int64_t file_len;
1529 int ret;
1531 if (size <= 0) {
1532 return 0;
1535 file_len = bdrv_getlength(bs->file->bs);
1536 if (file_len < 0) {
1537 return file_len;
1541 * Last cluster of qcow2 image may be semi-allocated, so it may be OK to
1542 * reference some space after file end but it should be less than one
1543 * cluster.
1545 if (offset + size - file_len >= s->cluster_size) {
1546 fprintf(stderr, "ERROR: counting reference for region exceeding the "
1547 "end of the file by one cluster or more: offset 0x%" PRIx64
1548 " size 0x%" PRIx64 "\n", offset, size);
1549 res->corruptions++;
1550 return 0;
1553 start = start_of_cluster(s, offset);
1554 last = start_of_cluster(s, offset + size - 1);
1555 for(cluster_offset = start; cluster_offset <= last;
1556 cluster_offset += s->cluster_size) {
1557 k = cluster_offset >> s->cluster_bits;
1558 if (k >= *refcount_table_size) {
1559 ret = realloc_refcount_array(s, refcount_table,
1560 refcount_table_size, k + 1);
1561 if (ret < 0) {
1562 res->check_errors++;
1563 return ret;
1567 refcount = s->get_refcount(*refcount_table, k);
1568 if (refcount == s->refcount_max) {
1569 fprintf(stderr, "ERROR: overflow cluster offset=0x%" PRIx64
1570 "\n", cluster_offset);
1571 fprintf(stderr, "Use qemu-img amend to increase the refcount entry "
1572 "width or qemu-img convert to create a clean copy if the "
1573 "image cannot be opened for writing\n");
1574 res->corruptions++;
1575 continue;
1577 s->set_refcount(*refcount_table, k, refcount + 1);
1580 return 0;
1583 /* Flags for check_refcounts_l1() and check_refcounts_l2() */
1584 enum {
1585 CHECK_FRAG_INFO = 0x2, /* update BlockFragInfo counters */
1589 * Increases the refcount in the given refcount table for the all clusters
1590 * referenced in the L2 table. While doing so, performs some checks on L2
1591 * entries.
1593 * Returns the number of errors found by the checks or -errno if an internal
1594 * error occurred.
1596 static int check_refcounts_l2(BlockDriverState *bs, BdrvCheckResult *res,
1597 void **refcount_table,
1598 int64_t *refcount_table_size, int64_t l2_offset,
1599 int flags, BdrvCheckMode fix, bool active)
1601 BDRVQcow2State *s = bs->opaque;
1602 uint64_t *l2_table, l2_entry;
1603 uint64_t next_contiguous_offset = 0;
1604 int i, l2_size, nb_csectors, ret;
1606 /* Read L2 table from disk */
1607 l2_size = s->l2_size * sizeof(uint64_t);
1608 l2_table = g_malloc(l2_size);
1610 ret = bdrv_pread(bs->file, l2_offset, l2_table, l2_size);
1611 if (ret < 0) {
1612 fprintf(stderr, "ERROR: I/O error in check_refcounts_l2\n");
1613 res->check_errors++;
1614 goto fail;
1617 /* Do the actual checks */
1618 for(i = 0; i < s->l2_size; i++) {
1619 l2_entry = be64_to_cpu(l2_table[i]);
1621 switch (qcow2_get_cluster_type(bs, l2_entry)) {
1622 case QCOW2_CLUSTER_COMPRESSED:
1623 /* Compressed clusters don't have QCOW_OFLAG_COPIED */
1624 if (l2_entry & QCOW_OFLAG_COPIED) {
1625 fprintf(stderr, "ERROR: coffset=0x%" PRIx64 ": "
1626 "copied flag must never be set for compressed "
1627 "clusters\n", l2_entry & s->cluster_offset_mask);
1628 l2_entry &= ~QCOW_OFLAG_COPIED;
1629 res->corruptions++;
1632 if (has_data_file(bs)) {
1633 fprintf(stderr, "ERROR compressed cluster %d with data file, "
1634 "entry=0x%" PRIx64 "\n", i, l2_entry);
1635 res->corruptions++;
1636 break;
1639 /* Mark cluster as used */
1640 nb_csectors = ((l2_entry >> s->csize_shift) &
1641 s->csize_mask) + 1;
1642 l2_entry &= s->cluster_offset_mask;
1643 ret = qcow2_inc_refcounts_imrt(
1644 bs, res, refcount_table, refcount_table_size,
1645 l2_entry & QCOW2_COMPRESSED_SECTOR_MASK,
1646 nb_csectors * QCOW2_COMPRESSED_SECTOR_SIZE);
1647 if (ret < 0) {
1648 goto fail;
1651 if (flags & CHECK_FRAG_INFO) {
1652 res->bfi.allocated_clusters++;
1653 res->bfi.compressed_clusters++;
1655 /* Compressed clusters are fragmented by nature. Since they
1656 * take up sub-sector space but we only have sector granularity
1657 * I/O we need to re-read the same sectors even for adjacent
1658 * compressed clusters.
1660 res->bfi.fragmented_clusters++;
1662 break;
1664 case QCOW2_CLUSTER_ZERO_ALLOC:
1665 case QCOW2_CLUSTER_NORMAL:
1667 uint64_t offset = l2_entry & L2E_OFFSET_MASK;
1669 /* Correct offsets are cluster aligned */
1670 if (offset_into_cluster(s, offset)) {
1671 res->corruptions++;
1673 if (qcow2_get_cluster_type(bs, l2_entry) ==
1674 QCOW2_CLUSTER_ZERO_ALLOC)
1676 fprintf(stderr, "%s offset=%" PRIx64 ": Preallocated zero "
1677 "cluster is not properly aligned; L2 entry "
1678 "corrupted.\n",
1679 fix & BDRV_FIX_ERRORS ? "Repairing" : "ERROR",
1680 offset);
1681 if (fix & BDRV_FIX_ERRORS) {
1682 uint64_t l2e_offset =
1683 l2_offset + (uint64_t)i * sizeof(uint64_t);
1684 int ign = active ? QCOW2_OL_ACTIVE_L2 :
1685 QCOW2_OL_INACTIVE_L2;
1687 l2_entry = QCOW_OFLAG_ZERO;
1688 l2_table[i] = cpu_to_be64(l2_entry);
1689 ret = qcow2_pre_write_overlap_check(bs, ign,
1690 l2e_offset, sizeof(uint64_t), false);
1691 if (ret < 0) {
1692 fprintf(stderr, "ERROR: Overlap check failed\n");
1693 res->check_errors++;
1694 /* Something is seriously wrong, so abort checking
1695 * this L2 table */
1696 goto fail;
1699 ret = bdrv_pwrite_sync(bs->file, l2e_offset,
1700 &l2_table[i], sizeof(uint64_t));
1701 if (ret < 0) {
1702 fprintf(stderr, "ERROR: Failed to overwrite L2 "
1703 "table entry: %s\n", strerror(-ret));
1704 res->check_errors++;
1705 /* Do not abort, continue checking the rest of this
1706 * L2 table's entries */
1707 } else {
1708 res->corruptions--;
1709 res->corruptions_fixed++;
1710 /* Skip marking the cluster as used
1711 * (it is unused now) */
1712 continue;
1715 } else {
1716 fprintf(stderr, "ERROR offset=%" PRIx64 ": Data cluster is "
1717 "not properly aligned; L2 entry corrupted.\n", offset);
1721 if (flags & CHECK_FRAG_INFO) {
1722 res->bfi.allocated_clusters++;
1723 if (next_contiguous_offset &&
1724 offset != next_contiguous_offset) {
1725 res->bfi.fragmented_clusters++;
1727 next_contiguous_offset = offset + s->cluster_size;
1730 /* Mark cluster as used */
1731 if (!has_data_file(bs)) {
1732 ret = qcow2_inc_refcounts_imrt(bs, res, refcount_table,
1733 refcount_table_size,
1734 offset, s->cluster_size);
1735 if (ret < 0) {
1736 goto fail;
1739 break;
1742 case QCOW2_CLUSTER_ZERO_PLAIN:
1743 case QCOW2_CLUSTER_UNALLOCATED:
1744 break;
1746 default:
1747 abort();
1751 g_free(l2_table);
1752 return 0;
1754 fail:
1755 g_free(l2_table);
1756 return ret;
1760 * Increases the refcount for the L1 table, its L2 tables and all referenced
1761 * clusters in the given refcount table. While doing so, performs some checks
1762 * on L1 and L2 entries.
1764 * Returns the number of errors found by the checks or -errno if an internal
1765 * error occurred.
1767 static int check_refcounts_l1(BlockDriverState *bs,
1768 BdrvCheckResult *res,
1769 void **refcount_table,
1770 int64_t *refcount_table_size,
1771 int64_t l1_table_offset, int l1_size,
1772 int flags, BdrvCheckMode fix, bool active)
1774 BDRVQcow2State *s = bs->opaque;
1775 uint64_t *l1_table = NULL, l2_offset, l1_size2;
1776 int i, ret;
1778 l1_size2 = l1_size * sizeof(uint64_t);
1780 /* Mark L1 table as used */
1781 ret = qcow2_inc_refcounts_imrt(bs, res, refcount_table, refcount_table_size,
1782 l1_table_offset, l1_size2);
1783 if (ret < 0) {
1784 goto fail;
1787 /* Read L1 table entries from disk */
1788 if (l1_size2 > 0) {
1789 l1_table = g_try_malloc(l1_size2);
1790 if (l1_table == NULL) {
1791 ret = -ENOMEM;
1792 res->check_errors++;
1793 goto fail;
1795 ret = bdrv_pread(bs->file, l1_table_offset, l1_table, l1_size2);
1796 if (ret < 0) {
1797 fprintf(stderr, "ERROR: I/O error in check_refcounts_l1\n");
1798 res->check_errors++;
1799 goto fail;
1801 for(i = 0;i < l1_size; i++)
1802 be64_to_cpus(&l1_table[i]);
1805 /* Do the actual checks */
1806 for(i = 0; i < l1_size; i++) {
1807 l2_offset = l1_table[i];
1808 if (l2_offset) {
1809 /* Mark L2 table as used */
1810 l2_offset &= L1E_OFFSET_MASK;
1811 ret = qcow2_inc_refcounts_imrt(bs, res,
1812 refcount_table, refcount_table_size,
1813 l2_offset, s->cluster_size);
1814 if (ret < 0) {
1815 goto fail;
1818 /* L2 tables are cluster aligned */
1819 if (offset_into_cluster(s, l2_offset)) {
1820 fprintf(stderr, "ERROR l2_offset=%" PRIx64 ": Table is not "
1821 "cluster aligned; L1 entry corrupted\n", l2_offset);
1822 res->corruptions++;
1825 /* Process and check L2 entries */
1826 ret = check_refcounts_l2(bs, res, refcount_table,
1827 refcount_table_size, l2_offset, flags,
1828 fix, active);
1829 if (ret < 0) {
1830 goto fail;
1834 g_free(l1_table);
1835 return 0;
1837 fail:
1838 g_free(l1_table);
1839 return ret;
1843 * Checks the OFLAG_COPIED flag for all L1 and L2 entries.
1845 * This function does not print an error message nor does it increment
1846 * check_errors if qcow2_get_refcount fails (this is because such an error will
1847 * have been already detected and sufficiently signaled by the calling function
1848 * (qcow2_check_refcounts) by the time this function is called).
1850 static int check_oflag_copied(BlockDriverState *bs, BdrvCheckResult *res,
1851 BdrvCheckMode fix)
1853 BDRVQcow2State *s = bs->opaque;
1854 uint64_t *l2_table = qemu_blockalign(bs, s->cluster_size);
1855 int ret;
1856 uint64_t refcount;
1857 int i, j;
1858 bool repair;
1860 if (fix & BDRV_FIX_ERRORS) {
1861 /* Always repair */
1862 repair = true;
1863 } else if (fix & BDRV_FIX_LEAKS) {
1864 /* Repair only if that seems safe: This function is always
1865 * called after the refcounts have been fixed, so the refcount
1866 * is accurate if that repair was successful */
1867 repair = !res->check_errors && !res->corruptions && !res->leaks;
1868 } else {
1869 repair = false;
1872 for (i = 0; i < s->l1_size; i++) {
1873 uint64_t l1_entry = s->l1_table[i];
1874 uint64_t l2_offset = l1_entry & L1E_OFFSET_MASK;
1875 int l2_dirty = 0;
1877 if (!l2_offset) {
1878 continue;
1881 ret = qcow2_get_refcount(bs, l2_offset >> s->cluster_bits,
1882 &refcount);
1883 if (ret < 0) {
1884 /* don't print message nor increment check_errors */
1885 continue;
1887 if ((refcount == 1) != ((l1_entry & QCOW_OFLAG_COPIED) != 0)) {
1888 res->corruptions++;
1889 fprintf(stderr, "%s OFLAG_COPIED L2 cluster: l1_index=%d "
1890 "l1_entry=%" PRIx64 " refcount=%" PRIu64 "\n",
1891 repair ? "Repairing" : "ERROR", i, l1_entry, refcount);
1892 if (repair) {
1893 s->l1_table[i] = refcount == 1
1894 ? l1_entry | QCOW_OFLAG_COPIED
1895 : l1_entry & ~QCOW_OFLAG_COPIED;
1896 ret = qcow2_write_l1_entry(bs, i);
1897 if (ret < 0) {
1898 res->check_errors++;
1899 goto fail;
1901 res->corruptions--;
1902 res->corruptions_fixed++;
1906 ret = bdrv_pread(bs->file, l2_offset, l2_table,
1907 s->l2_size * sizeof(uint64_t));
1908 if (ret < 0) {
1909 fprintf(stderr, "ERROR: Could not read L2 table: %s\n",
1910 strerror(-ret));
1911 res->check_errors++;
1912 goto fail;
1915 for (j = 0; j < s->l2_size; j++) {
1916 uint64_t l2_entry = be64_to_cpu(l2_table[j]);
1917 uint64_t data_offset = l2_entry & L2E_OFFSET_MASK;
1918 QCow2ClusterType cluster_type = qcow2_get_cluster_type(bs, l2_entry);
1920 if (cluster_type == QCOW2_CLUSTER_NORMAL ||
1921 cluster_type == QCOW2_CLUSTER_ZERO_ALLOC) {
1922 if (has_data_file(bs)) {
1923 refcount = 1;
1924 } else {
1925 ret = qcow2_get_refcount(bs,
1926 data_offset >> s->cluster_bits,
1927 &refcount);
1928 if (ret < 0) {
1929 /* don't print message nor increment check_errors */
1930 continue;
1933 if ((refcount == 1) != ((l2_entry & QCOW_OFLAG_COPIED) != 0)) {
1934 res->corruptions++;
1935 fprintf(stderr, "%s OFLAG_COPIED data cluster: "
1936 "l2_entry=%" PRIx64 " refcount=%" PRIu64 "\n",
1937 repair ? "Repairing" : "ERROR", l2_entry, refcount);
1938 if (repair) {
1939 l2_table[j] = cpu_to_be64(refcount == 1
1940 ? l2_entry | QCOW_OFLAG_COPIED
1941 : l2_entry & ~QCOW_OFLAG_COPIED);
1942 l2_dirty++;
1948 if (l2_dirty > 0) {
1949 ret = qcow2_pre_write_overlap_check(bs, QCOW2_OL_ACTIVE_L2,
1950 l2_offset, s->cluster_size,
1951 false);
1952 if (ret < 0) {
1953 fprintf(stderr, "ERROR: Could not write L2 table; metadata "
1954 "overlap check failed: %s\n", strerror(-ret));
1955 res->check_errors++;
1956 goto fail;
1959 ret = bdrv_pwrite(bs->file, l2_offset, l2_table,
1960 s->cluster_size);
1961 if (ret < 0) {
1962 fprintf(stderr, "ERROR: Could not write L2 table: %s\n",
1963 strerror(-ret));
1964 res->check_errors++;
1965 goto fail;
1967 res->corruptions -= l2_dirty;
1968 res->corruptions_fixed += l2_dirty;
1972 ret = 0;
1974 fail:
1975 qemu_vfree(l2_table);
1976 return ret;
1980 * Checks consistency of refblocks and accounts for each refblock in
1981 * *refcount_table.
1983 static int check_refblocks(BlockDriverState *bs, BdrvCheckResult *res,
1984 BdrvCheckMode fix, bool *rebuild,
1985 void **refcount_table, int64_t *nb_clusters)
1987 BDRVQcow2State *s = bs->opaque;
1988 int64_t i, size;
1989 int ret;
1991 for(i = 0; i < s->refcount_table_size; i++) {
1992 uint64_t offset, cluster;
1993 offset = s->refcount_table[i];
1994 cluster = offset >> s->cluster_bits;
1996 /* Refcount blocks are cluster aligned */
1997 if (offset_into_cluster(s, offset)) {
1998 fprintf(stderr, "ERROR refcount block %" PRId64 " is not "
1999 "cluster aligned; refcount table entry corrupted\n", i);
2000 res->corruptions++;
2001 *rebuild = true;
2002 continue;
2005 if (cluster >= *nb_clusters) {
2006 res->corruptions++;
2007 fprintf(stderr, "%s refcount block %" PRId64 " is outside image\n",
2008 fix & BDRV_FIX_ERRORS ? "Repairing" : "ERROR", i);
2010 if (fix & BDRV_FIX_ERRORS) {
2011 int64_t new_nb_clusters;
2012 Error *local_err = NULL;
2014 if (offset > INT64_MAX - s->cluster_size) {
2015 ret = -EINVAL;
2016 goto resize_fail;
2019 ret = bdrv_truncate(bs->file, offset + s->cluster_size,
2020 PREALLOC_MODE_OFF, &local_err);
2021 if (ret < 0) {
2022 error_report_err(local_err);
2023 goto resize_fail;
2025 size = bdrv_getlength(bs->file->bs);
2026 if (size < 0) {
2027 ret = size;
2028 goto resize_fail;
2031 new_nb_clusters = size_to_clusters(s, size);
2032 assert(new_nb_clusters >= *nb_clusters);
2034 ret = realloc_refcount_array(s, refcount_table,
2035 nb_clusters, new_nb_clusters);
2036 if (ret < 0) {
2037 res->check_errors++;
2038 return ret;
2041 if (cluster >= *nb_clusters) {
2042 ret = -EINVAL;
2043 goto resize_fail;
2046 res->corruptions--;
2047 res->corruptions_fixed++;
2048 ret = qcow2_inc_refcounts_imrt(bs, res,
2049 refcount_table, nb_clusters,
2050 offset, s->cluster_size);
2051 if (ret < 0) {
2052 return ret;
2054 /* No need to check whether the refcount is now greater than 1:
2055 * This area was just allocated and zeroed, so it can only be
2056 * exactly 1 after qcow2_inc_refcounts_imrt() */
2057 continue;
2059 resize_fail:
2060 *rebuild = true;
2061 fprintf(stderr, "ERROR could not resize image: %s\n",
2062 strerror(-ret));
2064 continue;
2067 if (offset != 0) {
2068 ret = qcow2_inc_refcounts_imrt(bs, res, refcount_table, nb_clusters,
2069 offset, s->cluster_size);
2070 if (ret < 0) {
2071 return ret;
2073 if (s->get_refcount(*refcount_table, cluster) != 1) {
2074 fprintf(stderr, "ERROR refcount block %" PRId64
2075 " refcount=%" PRIu64 "\n", i,
2076 s->get_refcount(*refcount_table, cluster));
2077 res->corruptions++;
2078 *rebuild = true;
2083 return 0;
2087 * Calculates an in-memory refcount table.
2089 static int calculate_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
2090 BdrvCheckMode fix, bool *rebuild,
2091 void **refcount_table, int64_t *nb_clusters)
2093 BDRVQcow2State *s = bs->opaque;
2094 int64_t i;
2095 QCowSnapshot *sn;
2096 int ret;
2098 if (!*refcount_table) {
2099 int64_t old_size = 0;
2100 ret = realloc_refcount_array(s, refcount_table,
2101 &old_size, *nb_clusters);
2102 if (ret < 0) {
2103 res->check_errors++;
2104 return ret;
2108 /* header */
2109 ret = qcow2_inc_refcounts_imrt(bs, res, refcount_table, nb_clusters,
2110 0, s->cluster_size);
2111 if (ret < 0) {
2112 return ret;
2115 /* current L1 table */
2116 ret = check_refcounts_l1(bs, res, refcount_table, nb_clusters,
2117 s->l1_table_offset, s->l1_size, CHECK_FRAG_INFO,
2118 fix, true);
2119 if (ret < 0) {
2120 return ret;
2123 /* snapshots */
2124 if (has_data_file(bs) && s->nb_snapshots) {
2125 fprintf(stderr, "ERROR %d snapshots in image with data file\n",
2126 s->nb_snapshots);
2127 res->corruptions++;
2130 for (i = 0; i < s->nb_snapshots; i++) {
2131 sn = s->snapshots + i;
2132 if (offset_into_cluster(s, sn->l1_table_offset)) {
2133 fprintf(stderr, "ERROR snapshot %s (%s) l1_offset=%#" PRIx64 ": "
2134 "L1 table is not cluster aligned; snapshot table entry "
2135 "corrupted\n", sn->id_str, sn->name, sn->l1_table_offset);
2136 res->corruptions++;
2137 continue;
2139 if (sn->l1_size > QCOW_MAX_L1_SIZE / sizeof(uint64_t)) {
2140 fprintf(stderr, "ERROR snapshot %s (%s) l1_size=%#" PRIx32 ": "
2141 "L1 table is too large; snapshot table entry corrupted\n",
2142 sn->id_str, sn->name, sn->l1_size);
2143 res->corruptions++;
2144 continue;
2146 ret = check_refcounts_l1(bs, res, refcount_table, nb_clusters,
2147 sn->l1_table_offset, sn->l1_size, 0, fix,
2148 false);
2149 if (ret < 0) {
2150 return ret;
2153 ret = qcow2_inc_refcounts_imrt(bs, res, refcount_table, nb_clusters,
2154 s->snapshots_offset, s->snapshots_size);
2155 if (ret < 0) {
2156 return ret;
2159 /* refcount data */
2160 ret = qcow2_inc_refcounts_imrt(bs, res, refcount_table, nb_clusters,
2161 s->refcount_table_offset,
2162 s->refcount_table_size * sizeof(uint64_t));
2163 if (ret < 0) {
2164 return ret;
2167 /* encryption */
2168 if (s->crypto_header.length) {
2169 ret = qcow2_inc_refcounts_imrt(bs, res, refcount_table, nb_clusters,
2170 s->crypto_header.offset,
2171 s->crypto_header.length);
2172 if (ret < 0) {
2173 return ret;
2177 /* bitmaps */
2178 ret = qcow2_check_bitmaps_refcounts(bs, res, refcount_table, nb_clusters);
2179 if (ret < 0) {
2180 return ret;
2183 return check_refblocks(bs, res, fix, rebuild, refcount_table, nb_clusters);
2187 * Compares the actual reference count for each cluster in the image against the
2188 * refcount as reported by the refcount structures on-disk.
2190 static void compare_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
2191 BdrvCheckMode fix, bool *rebuild,
2192 int64_t *highest_cluster,
2193 void *refcount_table, int64_t nb_clusters)
2195 BDRVQcow2State *s = bs->opaque;
2196 int64_t i;
2197 uint64_t refcount1, refcount2;
2198 int ret;
2200 for (i = 0, *highest_cluster = 0; i < nb_clusters; i++) {
2201 ret = qcow2_get_refcount(bs, i, &refcount1);
2202 if (ret < 0) {
2203 fprintf(stderr, "Can't get refcount for cluster %" PRId64 ": %s\n",
2204 i, strerror(-ret));
2205 res->check_errors++;
2206 continue;
2209 refcount2 = s->get_refcount(refcount_table, i);
2211 if (refcount1 > 0 || refcount2 > 0) {
2212 *highest_cluster = i;
2215 if (refcount1 != refcount2) {
2216 /* Check if we're allowed to fix the mismatch */
2217 int *num_fixed = NULL;
2218 if (refcount1 == 0) {
2219 *rebuild = true;
2220 } else if (refcount1 > refcount2 && (fix & BDRV_FIX_LEAKS)) {
2221 num_fixed = &res->leaks_fixed;
2222 } else if (refcount1 < refcount2 && (fix & BDRV_FIX_ERRORS)) {
2223 num_fixed = &res->corruptions_fixed;
2226 fprintf(stderr, "%s cluster %" PRId64 " refcount=%" PRIu64
2227 " reference=%" PRIu64 "\n",
2228 num_fixed != NULL ? "Repairing" :
2229 refcount1 < refcount2 ? "ERROR" :
2230 "Leaked",
2231 i, refcount1, refcount2);
2233 if (num_fixed) {
2234 ret = update_refcount(bs, i << s->cluster_bits, 1,
2235 refcount_diff(refcount1, refcount2),
2236 refcount1 > refcount2,
2237 QCOW2_DISCARD_ALWAYS);
2238 if (ret >= 0) {
2239 (*num_fixed)++;
2240 continue;
2244 /* And if we couldn't, print an error */
2245 if (refcount1 < refcount2) {
2246 res->corruptions++;
2247 } else {
2248 res->leaks++;
2255 * Allocates clusters using an in-memory refcount table (IMRT) in contrast to
2256 * the on-disk refcount structures.
2258 * On input, *first_free_cluster tells where to start looking, and need not
2259 * actually be a free cluster; the returned offset will not be before that
2260 * cluster. On output, *first_free_cluster points to the first gap found, even
2261 * if that gap was too small to be used as the returned offset.
2263 * Note that *first_free_cluster is a cluster index whereas the return value is
2264 * an offset.
2266 static int64_t alloc_clusters_imrt(BlockDriverState *bs,
2267 int cluster_count,
2268 void **refcount_table,
2269 int64_t *imrt_nb_clusters,
2270 int64_t *first_free_cluster)
2272 BDRVQcow2State *s = bs->opaque;
2273 int64_t cluster = *first_free_cluster, i;
2274 bool first_gap = true;
2275 int contiguous_free_clusters;
2276 int ret;
2278 /* Starting at *first_free_cluster, find a range of at least cluster_count
2279 * continuously free clusters */
2280 for (contiguous_free_clusters = 0;
2281 cluster < *imrt_nb_clusters &&
2282 contiguous_free_clusters < cluster_count;
2283 cluster++)
2285 if (!s->get_refcount(*refcount_table, cluster)) {
2286 contiguous_free_clusters++;
2287 if (first_gap) {
2288 /* If this is the first free cluster found, update
2289 * *first_free_cluster accordingly */
2290 *first_free_cluster = cluster;
2291 first_gap = false;
2293 } else if (contiguous_free_clusters) {
2294 contiguous_free_clusters = 0;
2298 /* If contiguous_free_clusters is greater than zero, it contains the number
2299 * of continuously free clusters until the current cluster; the first free
2300 * cluster in the current "gap" is therefore
2301 * cluster - contiguous_free_clusters */
2303 /* If no such range could be found, grow the in-memory refcount table
2304 * accordingly to append free clusters at the end of the image */
2305 if (contiguous_free_clusters < cluster_count) {
2306 /* contiguous_free_clusters clusters are already empty at the image end;
2307 * we need cluster_count clusters; therefore, we have to allocate
2308 * cluster_count - contiguous_free_clusters new clusters at the end of
2309 * the image (which is the current value of cluster; note that cluster
2310 * may exceed old_imrt_nb_clusters if *first_free_cluster pointed beyond
2311 * the image end) */
2312 ret = realloc_refcount_array(s, refcount_table, imrt_nb_clusters,
2313 cluster + cluster_count
2314 - contiguous_free_clusters);
2315 if (ret < 0) {
2316 return ret;
2320 /* Go back to the first free cluster */
2321 cluster -= contiguous_free_clusters;
2322 for (i = 0; i < cluster_count; i++) {
2323 s->set_refcount(*refcount_table, cluster + i, 1);
2326 return cluster << s->cluster_bits;
2330 * Creates a new refcount structure based solely on the in-memory information
2331 * given through *refcount_table. All necessary allocations will be reflected
2332 * in that array.
2334 * On success, the old refcount structure is leaked (it will be covered by the
2335 * new refcount structure).
2337 static int rebuild_refcount_structure(BlockDriverState *bs,
2338 BdrvCheckResult *res,
2339 void **refcount_table,
2340 int64_t *nb_clusters)
2342 BDRVQcow2State *s = bs->opaque;
2343 int64_t first_free_cluster = 0, reftable_offset = -1, cluster = 0;
2344 int64_t refblock_offset, refblock_start, refblock_index;
2345 uint32_t reftable_size = 0;
2346 uint64_t *on_disk_reftable = NULL;
2347 void *on_disk_refblock;
2348 int ret = 0;
2349 struct {
2350 uint64_t reftable_offset;
2351 uint32_t reftable_clusters;
2352 } QEMU_PACKED reftable_offset_and_clusters;
2354 qcow2_cache_empty(bs, s->refcount_block_cache);
2356 write_refblocks:
2357 for (; cluster < *nb_clusters; cluster++) {
2358 if (!s->get_refcount(*refcount_table, cluster)) {
2359 continue;
2362 refblock_index = cluster >> s->refcount_block_bits;
2363 refblock_start = refblock_index << s->refcount_block_bits;
2365 /* Don't allocate a cluster in a refblock already written to disk */
2366 if (first_free_cluster < refblock_start) {
2367 first_free_cluster = refblock_start;
2369 refblock_offset = alloc_clusters_imrt(bs, 1, refcount_table,
2370 nb_clusters, &first_free_cluster);
2371 if (refblock_offset < 0) {
2372 fprintf(stderr, "ERROR allocating refblock: %s\n",
2373 strerror(-refblock_offset));
2374 res->check_errors++;
2375 ret = refblock_offset;
2376 goto fail;
2379 if (reftable_size <= refblock_index) {
2380 uint32_t old_reftable_size = reftable_size;
2381 uint64_t *new_on_disk_reftable;
2383 reftable_size = ROUND_UP((refblock_index + 1) * sizeof(uint64_t),
2384 s->cluster_size) / sizeof(uint64_t);
2385 new_on_disk_reftable = g_try_realloc(on_disk_reftable,
2386 reftable_size *
2387 sizeof(uint64_t));
2388 if (!new_on_disk_reftable) {
2389 res->check_errors++;
2390 ret = -ENOMEM;
2391 goto fail;
2393 on_disk_reftable = new_on_disk_reftable;
2395 memset(on_disk_reftable + old_reftable_size, 0,
2396 (reftable_size - old_reftable_size) * sizeof(uint64_t));
2398 /* The offset we have for the reftable is now no longer valid;
2399 * this will leak that range, but we can easily fix that by running
2400 * a leak-fixing check after this rebuild operation */
2401 reftable_offset = -1;
2402 } else {
2403 assert(on_disk_reftable);
2405 on_disk_reftable[refblock_index] = refblock_offset;
2407 /* If this is apparently the last refblock (for now), try to squeeze the
2408 * reftable in */
2409 if (refblock_index == (*nb_clusters - 1) >> s->refcount_block_bits &&
2410 reftable_offset < 0)
2412 uint64_t reftable_clusters = size_to_clusters(s, reftable_size *
2413 sizeof(uint64_t));
2414 reftable_offset = alloc_clusters_imrt(bs, reftable_clusters,
2415 refcount_table, nb_clusters,
2416 &first_free_cluster);
2417 if (reftable_offset < 0) {
2418 fprintf(stderr, "ERROR allocating reftable: %s\n",
2419 strerror(-reftable_offset));
2420 res->check_errors++;
2421 ret = reftable_offset;
2422 goto fail;
2426 ret = qcow2_pre_write_overlap_check(bs, 0, refblock_offset,
2427 s->cluster_size, false);
2428 if (ret < 0) {
2429 fprintf(stderr, "ERROR writing refblock: %s\n", strerror(-ret));
2430 goto fail;
2433 /* The size of *refcount_table is always cluster-aligned, therefore the
2434 * write operation will not overflow */
2435 on_disk_refblock = (void *)((char *) *refcount_table +
2436 refblock_index * s->cluster_size);
2438 ret = bdrv_pwrite(bs->file, refblock_offset, on_disk_refblock,
2439 s->cluster_size);
2440 if (ret < 0) {
2441 fprintf(stderr, "ERROR writing refblock: %s\n", strerror(-ret));
2442 goto fail;
2445 /* Go to the end of this refblock */
2446 cluster = refblock_start + s->refcount_block_size - 1;
2449 if (reftable_offset < 0) {
2450 uint64_t post_refblock_start, reftable_clusters;
2452 post_refblock_start = ROUND_UP(*nb_clusters, s->refcount_block_size);
2453 reftable_clusters = size_to_clusters(s,
2454 reftable_size * sizeof(uint64_t));
2455 /* Not pretty but simple */
2456 if (first_free_cluster < post_refblock_start) {
2457 first_free_cluster = post_refblock_start;
2459 reftable_offset = alloc_clusters_imrt(bs, reftable_clusters,
2460 refcount_table, nb_clusters,
2461 &first_free_cluster);
2462 if (reftable_offset < 0) {
2463 fprintf(stderr, "ERROR allocating reftable: %s\n",
2464 strerror(-reftable_offset));
2465 res->check_errors++;
2466 ret = reftable_offset;
2467 goto fail;
2470 goto write_refblocks;
2473 for (refblock_index = 0; refblock_index < reftable_size; refblock_index++) {
2474 cpu_to_be64s(&on_disk_reftable[refblock_index]);
2477 ret = qcow2_pre_write_overlap_check(bs, 0, reftable_offset,
2478 reftable_size * sizeof(uint64_t),
2479 false);
2480 if (ret < 0) {
2481 fprintf(stderr, "ERROR writing reftable: %s\n", strerror(-ret));
2482 goto fail;
2485 assert(reftable_size < INT_MAX / sizeof(uint64_t));
2486 ret = bdrv_pwrite(bs->file, reftable_offset, on_disk_reftable,
2487 reftable_size * sizeof(uint64_t));
2488 if (ret < 0) {
2489 fprintf(stderr, "ERROR writing reftable: %s\n", strerror(-ret));
2490 goto fail;
2493 /* Enter new reftable into the image header */
2494 reftable_offset_and_clusters.reftable_offset = cpu_to_be64(reftable_offset);
2495 reftable_offset_and_clusters.reftable_clusters =
2496 cpu_to_be32(size_to_clusters(s, reftable_size * sizeof(uint64_t)));
2497 ret = bdrv_pwrite_sync(bs->file,
2498 offsetof(QCowHeader, refcount_table_offset),
2499 &reftable_offset_and_clusters,
2500 sizeof(reftable_offset_and_clusters));
2501 if (ret < 0) {
2502 fprintf(stderr, "ERROR setting reftable: %s\n", strerror(-ret));
2503 goto fail;
2506 for (refblock_index = 0; refblock_index < reftable_size; refblock_index++) {
2507 be64_to_cpus(&on_disk_reftable[refblock_index]);
2509 s->refcount_table = on_disk_reftable;
2510 s->refcount_table_offset = reftable_offset;
2511 s->refcount_table_size = reftable_size;
2512 update_max_refcount_table_index(s);
2514 return 0;
2516 fail:
2517 g_free(on_disk_reftable);
2518 return ret;
2522 * Checks an image for refcount consistency.
2524 * Returns 0 if no errors are found, the number of errors in case the image is
2525 * detected as corrupted, and -errno when an internal error occurred.
2527 int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
2528 BdrvCheckMode fix)
2530 BDRVQcow2State *s = bs->opaque;
2531 BdrvCheckResult pre_compare_res;
2532 int64_t size, highest_cluster, nb_clusters;
2533 void *refcount_table = NULL;
2534 bool rebuild = false;
2535 int ret;
2537 size = bdrv_getlength(bs->file->bs);
2538 if (size < 0) {
2539 res->check_errors++;
2540 return size;
2543 nb_clusters = size_to_clusters(s, size);
2544 if (nb_clusters > INT_MAX) {
2545 res->check_errors++;
2546 return -EFBIG;
2549 res->bfi.total_clusters =
2550 size_to_clusters(s, bs->total_sectors * BDRV_SECTOR_SIZE);
2552 ret = calculate_refcounts(bs, res, fix, &rebuild, &refcount_table,
2553 &nb_clusters);
2554 if (ret < 0) {
2555 goto fail;
2558 /* In case we don't need to rebuild the refcount structure (but want to fix
2559 * something), this function is immediately called again, in which case the
2560 * result should be ignored */
2561 pre_compare_res = *res;
2562 compare_refcounts(bs, res, 0, &rebuild, &highest_cluster, refcount_table,
2563 nb_clusters);
2565 if (rebuild && (fix & BDRV_FIX_ERRORS)) {
2566 BdrvCheckResult old_res = *res;
2567 int fresh_leaks = 0;
2569 fprintf(stderr, "Rebuilding refcount structure\n");
2570 ret = rebuild_refcount_structure(bs, res, &refcount_table,
2571 &nb_clusters);
2572 if (ret < 0) {
2573 goto fail;
2576 res->corruptions = 0;
2577 res->leaks = 0;
2579 /* Because the old reftable has been exchanged for a new one the
2580 * references have to be recalculated */
2581 rebuild = false;
2582 memset(refcount_table, 0, refcount_array_byte_size(s, nb_clusters));
2583 ret = calculate_refcounts(bs, res, 0, &rebuild, &refcount_table,
2584 &nb_clusters);
2585 if (ret < 0) {
2586 goto fail;
2589 if (fix & BDRV_FIX_LEAKS) {
2590 /* The old refcount structures are now leaked, fix it; the result
2591 * can be ignored, aside from leaks which were introduced by
2592 * rebuild_refcount_structure() that could not be fixed */
2593 BdrvCheckResult saved_res = *res;
2594 *res = (BdrvCheckResult){ 0 };
2596 compare_refcounts(bs, res, BDRV_FIX_LEAKS, &rebuild,
2597 &highest_cluster, refcount_table, nb_clusters);
2598 if (rebuild) {
2599 fprintf(stderr, "ERROR rebuilt refcount structure is still "
2600 "broken\n");
2603 /* Any leaks accounted for here were introduced by
2604 * rebuild_refcount_structure() because that function has created a
2605 * new refcount structure from scratch */
2606 fresh_leaks = res->leaks;
2607 *res = saved_res;
2610 if (res->corruptions < old_res.corruptions) {
2611 res->corruptions_fixed += old_res.corruptions - res->corruptions;
2613 if (res->leaks < old_res.leaks) {
2614 res->leaks_fixed += old_res.leaks - res->leaks;
2616 res->leaks += fresh_leaks;
2617 } else if (fix) {
2618 if (rebuild) {
2619 fprintf(stderr, "ERROR need to rebuild refcount structures\n");
2620 res->check_errors++;
2621 ret = -EIO;
2622 goto fail;
2625 if (res->leaks || res->corruptions) {
2626 *res = pre_compare_res;
2627 compare_refcounts(bs, res, fix, &rebuild, &highest_cluster,
2628 refcount_table, nb_clusters);
2632 /* check OFLAG_COPIED */
2633 ret = check_oflag_copied(bs, res, fix);
2634 if (ret < 0) {
2635 goto fail;
2638 res->image_end_offset = (highest_cluster + 1) * s->cluster_size;
2639 ret = 0;
2641 fail:
2642 g_free(refcount_table);
2644 return ret;
2647 #define overlaps_with(ofs, sz) \
2648 ranges_overlap(offset, size, ofs, sz)
2651 * Checks if the given offset into the image file is actually free to use by
2652 * looking for overlaps with important metadata sections (L1/L2 tables etc.),
2653 * i.e. a sanity check without relying on the refcount tables.
2655 * The ign parameter specifies what checks not to perform (being a bitmask of
2656 * QCow2MetadataOverlap values), i.e., what sections to ignore.
2658 * Returns:
2659 * - 0 if writing to this offset will not affect the mentioned metadata
2660 * - a positive QCow2MetadataOverlap value indicating one overlapping section
2661 * - a negative value (-errno) indicating an error while performing a check,
2662 * e.g. when bdrv_read failed on QCOW2_OL_INACTIVE_L2
2664 int qcow2_check_metadata_overlap(BlockDriverState *bs, int ign, int64_t offset,
2665 int64_t size)
2667 BDRVQcow2State *s = bs->opaque;
2668 int chk = s->overlap_check & ~ign;
2669 int i, j;
2671 if (!size) {
2672 return 0;
2675 if (chk & QCOW2_OL_MAIN_HEADER) {
2676 if (offset < s->cluster_size) {
2677 return QCOW2_OL_MAIN_HEADER;
2681 /* align range to test to cluster boundaries */
2682 size = ROUND_UP(offset_into_cluster(s, offset) + size, s->cluster_size);
2683 offset = start_of_cluster(s, offset);
2685 if ((chk & QCOW2_OL_ACTIVE_L1) && s->l1_size) {
2686 if (overlaps_with(s->l1_table_offset, s->l1_size * sizeof(uint64_t))) {
2687 return QCOW2_OL_ACTIVE_L1;
2691 if ((chk & QCOW2_OL_REFCOUNT_TABLE) && s->refcount_table_size) {
2692 if (overlaps_with(s->refcount_table_offset,
2693 s->refcount_table_size * sizeof(uint64_t))) {
2694 return QCOW2_OL_REFCOUNT_TABLE;
2698 if ((chk & QCOW2_OL_SNAPSHOT_TABLE) && s->snapshots_size) {
2699 if (overlaps_with(s->snapshots_offset, s->snapshots_size)) {
2700 return QCOW2_OL_SNAPSHOT_TABLE;
2704 if ((chk & QCOW2_OL_INACTIVE_L1) && s->snapshots) {
2705 for (i = 0; i < s->nb_snapshots; i++) {
2706 if (s->snapshots[i].l1_size &&
2707 overlaps_with(s->snapshots[i].l1_table_offset,
2708 s->snapshots[i].l1_size * sizeof(uint64_t))) {
2709 return QCOW2_OL_INACTIVE_L1;
2714 if ((chk & QCOW2_OL_ACTIVE_L2) && s->l1_table) {
2715 for (i = 0; i < s->l1_size; i++) {
2716 if ((s->l1_table[i] & L1E_OFFSET_MASK) &&
2717 overlaps_with(s->l1_table[i] & L1E_OFFSET_MASK,
2718 s->cluster_size)) {
2719 return QCOW2_OL_ACTIVE_L2;
2724 if ((chk & QCOW2_OL_REFCOUNT_BLOCK) && s->refcount_table) {
2725 unsigned last_entry = s->max_refcount_table_index;
2726 assert(last_entry < s->refcount_table_size);
2727 assert(last_entry + 1 == s->refcount_table_size ||
2728 (s->refcount_table[last_entry + 1] & REFT_OFFSET_MASK) == 0);
2729 for (i = 0; i <= last_entry; i++) {
2730 if ((s->refcount_table[i] & REFT_OFFSET_MASK) &&
2731 overlaps_with(s->refcount_table[i] & REFT_OFFSET_MASK,
2732 s->cluster_size)) {
2733 return QCOW2_OL_REFCOUNT_BLOCK;
2738 if ((chk & QCOW2_OL_INACTIVE_L2) && s->snapshots) {
2739 for (i = 0; i < s->nb_snapshots; i++) {
2740 uint64_t l1_ofs = s->snapshots[i].l1_table_offset;
2741 uint32_t l1_sz = s->snapshots[i].l1_size;
2742 uint64_t l1_sz2 = l1_sz * sizeof(uint64_t);
2743 uint64_t *l1;
2744 int ret;
2746 ret = qcow2_validate_table(bs, l1_ofs, l1_sz, sizeof(uint64_t),
2747 QCOW_MAX_L1_SIZE, "", NULL);
2748 if (ret < 0) {
2749 return ret;
2752 l1 = g_try_malloc(l1_sz2);
2754 if (l1_sz2 && l1 == NULL) {
2755 return -ENOMEM;
2758 ret = bdrv_pread(bs->file, l1_ofs, l1, l1_sz2);
2759 if (ret < 0) {
2760 g_free(l1);
2761 return ret;
2764 for (j = 0; j < l1_sz; j++) {
2765 uint64_t l2_ofs = be64_to_cpu(l1[j]) & L1E_OFFSET_MASK;
2766 if (l2_ofs && overlaps_with(l2_ofs, s->cluster_size)) {
2767 g_free(l1);
2768 return QCOW2_OL_INACTIVE_L2;
2772 g_free(l1);
2776 if ((chk & QCOW2_OL_BITMAP_DIRECTORY) &&
2777 (s->autoclear_features & QCOW2_AUTOCLEAR_BITMAPS))
2779 if (overlaps_with(s->bitmap_directory_offset,
2780 s->bitmap_directory_size))
2782 return QCOW2_OL_BITMAP_DIRECTORY;
2786 return 0;
2789 static const char *metadata_ol_names[] = {
2790 [QCOW2_OL_MAIN_HEADER_BITNR] = "qcow2_header",
2791 [QCOW2_OL_ACTIVE_L1_BITNR] = "active L1 table",
2792 [QCOW2_OL_ACTIVE_L2_BITNR] = "active L2 table",
2793 [QCOW2_OL_REFCOUNT_TABLE_BITNR] = "refcount table",
2794 [QCOW2_OL_REFCOUNT_BLOCK_BITNR] = "refcount block",
2795 [QCOW2_OL_SNAPSHOT_TABLE_BITNR] = "snapshot table",
2796 [QCOW2_OL_INACTIVE_L1_BITNR] = "inactive L1 table",
2797 [QCOW2_OL_INACTIVE_L2_BITNR] = "inactive L2 table",
2798 [QCOW2_OL_BITMAP_DIRECTORY_BITNR] = "bitmap directory",
2800 QEMU_BUILD_BUG_ON(QCOW2_OL_MAX_BITNR != ARRAY_SIZE(metadata_ol_names));
2803 * First performs a check for metadata overlaps (through
2804 * qcow2_check_metadata_overlap); if that fails with a negative value (error
2805 * while performing a check), that value is returned. If an impending overlap
2806 * is detected, the BDS will be made unusable, the qcow2 file marked corrupt
2807 * and -EIO returned.
2809 * Returns 0 if there were neither overlaps nor errors while checking for
2810 * overlaps; or a negative value (-errno) on error.
2812 int qcow2_pre_write_overlap_check(BlockDriverState *bs, int ign, int64_t offset,
2813 int64_t size, bool data_file)
2815 int ret;
2817 if (data_file && has_data_file(bs)) {
2818 return 0;
2821 ret = qcow2_check_metadata_overlap(bs, ign, offset, size);
2822 if (ret < 0) {
2823 return ret;
2824 } else if (ret > 0) {
2825 int metadata_ol_bitnr = ctz32(ret);
2826 assert(metadata_ol_bitnr < QCOW2_OL_MAX_BITNR);
2828 qcow2_signal_corruption(bs, true, offset, size, "Preventing invalid "
2829 "write on metadata (overlaps with %s)",
2830 metadata_ol_names[metadata_ol_bitnr]);
2831 return -EIO;
2834 return 0;
2837 /* A pointer to a function of this type is given to walk_over_reftable(). That
2838 * function will create refblocks and pass them to a RefblockFinishOp once they
2839 * are completed (@refblock). @refblock_empty is set if the refblock is
2840 * completely empty.
2842 * Along with the refblock, a corresponding reftable entry is passed, in the
2843 * reftable @reftable (which may be reallocated) at @reftable_index.
2845 * @allocated should be set to true if a new cluster has been allocated.
2847 typedef int (RefblockFinishOp)(BlockDriverState *bs, uint64_t **reftable,
2848 uint64_t reftable_index, uint64_t *reftable_size,
2849 void *refblock, bool refblock_empty,
2850 bool *allocated, Error **errp);
2853 * This "operation" for walk_over_reftable() allocates the refblock on disk (if
2854 * it is not empty) and inserts its offset into the new reftable. The size of
2855 * this new reftable is increased as required.
2857 static int alloc_refblock(BlockDriverState *bs, uint64_t **reftable,
2858 uint64_t reftable_index, uint64_t *reftable_size,
2859 void *refblock, bool refblock_empty, bool *allocated,
2860 Error **errp)
2862 BDRVQcow2State *s = bs->opaque;
2863 int64_t offset;
2865 if (!refblock_empty && reftable_index >= *reftable_size) {
2866 uint64_t *new_reftable;
2867 uint64_t new_reftable_size;
2869 new_reftable_size = ROUND_UP(reftable_index + 1,
2870 s->cluster_size / sizeof(uint64_t));
2871 if (new_reftable_size > QCOW_MAX_REFTABLE_SIZE / sizeof(uint64_t)) {
2872 error_setg(errp,
2873 "This operation would make the refcount table grow "
2874 "beyond the maximum size supported by QEMU, aborting");
2875 return -ENOTSUP;
2878 new_reftable = g_try_realloc(*reftable, new_reftable_size *
2879 sizeof(uint64_t));
2880 if (!new_reftable) {
2881 error_setg(errp, "Failed to increase reftable buffer size");
2882 return -ENOMEM;
2885 memset(new_reftable + *reftable_size, 0,
2886 (new_reftable_size - *reftable_size) * sizeof(uint64_t));
2888 *reftable = new_reftable;
2889 *reftable_size = new_reftable_size;
2892 if (!refblock_empty && !(*reftable)[reftable_index]) {
2893 offset = qcow2_alloc_clusters(bs, s->cluster_size);
2894 if (offset < 0) {
2895 error_setg_errno(errp, -offset, "Failed to allocate refblock");
2896 return offset;
2898 (*reftable)[reftable_index] = offset;
2899 *allocated = true;
2902 return 0;
2906 * This "operation" for walk_over_reftable() writes the refblock to disk at the
2907 * offset specified by the new reftable's entry. It does not modify the new
2908 * reftable or change any refcounts.
2910 static int flush_refblock(BlockDriverState *bs, uint64_t **reftable,
2911 uint64_t reftable_index, uint64_t *reftable_size,
2912 void *refblock, bool refblock_empty, bool *allocated,
2913 Error **errp)
2915 BDRVQcow2State *s = bs->opaque;
2916 int64_t offset;
2917 int ret;
2919 if (reftable_index < *reftable_size && (*reftable)[reftable_index]) {
2920 offset = (*reftable)[reftable_index];
2922 ret = qcow2_pre_write_overlap_check(bs, 0, offset, s->cluster_size,
2923 false);
2924 if (ret < 0) {
2925 error_setg_errno(errp, -ret, "Overlap check failed");
2926 return ret;
2929 ret = bdrv_pwrite(bs->file, offset, refblock, s->cluster_size);
2930 if (ret < 0) {
2931 error_setg_errno(errp, -ret, "Failed to write refblock");
2932 return ret;
2934 } else {
2935 assert(refblock_empty);
2938 return 0;
2942 * This function walks over the existing reftable and every referenced refblock;
2943 * if @new_set_refcount is non-NULL, it is called for every refcount entry to
2944 * create an equal new entry in the passed @new_refblock. Once that
2945 * @new_refblock is completely filled, @operation will be called.
2947 * @status_cb and @cb_opaque are used for the amend operation's status callback.
2948 * @index is the index of the walk_over_reftable() calls and @total is the total
2949 * number of walk_over_reftable() calls per amend operation. Both are used for
2950 * calculating the parameters for the status callback.
2952 * @allocated is set to true if a new cluster has been allocated.
2954 static int walk_over_reftable(BlockDriverState *bs, uint64_t **new_reftable,
2955 uint64_t *new_reftable_index,
2956 uint64_t *new_reftable_size,
2957 void *new_refblock, int new_refblock_size,
2958 int new_refcount_bits,
2959 RefblockFinishOp *operation, bool *allocated,
2960 Qcow2SetRefcountFunc *new_set_refcount,
2961 BlockDriverAmendStatusCB *status_cb,
2962 void *cb_opaque, int index, int total,
2963 Error **errp)
2965 BDRVQcow2State *s = bs->opaque;
2966 uint64_t reftable_index;
2967 bool new_refblock_empty = true;
2968 int refblock_index;
2969 int new_refblock_index = 0;
2970 int ret;
2972 for (reftable_index = 0; reftable_index < s->refcount_table_size;
2973 reftable_index++)
2975 uint64_t refblock_offset = s->refcount_table[reftable_index]
2976 & REFT_OFFSET_MASK;
2978 status_cb(bs, (uint64_t)index * s->refcount_table_size + reftable_index,
2979 (uint64_t)total * s->refcount_table_size, cb_opaque);
2981 if (refblock_offset) {
2982 void *refblock;
2984 if (offset_into_cluster(s, refblock_offset)) {
2985 qcow2_signal_corruption(bs, true, -1, -1, "Refblock offset %#"
2986 PRIx64 " unaligned (reftable index: %#"
2987 PRIx64 ")", refblock_offset,
2988 reftable_index);
2989 error_setg(errp,
2990 "Image is corrupt (unaligned refblock offset)");
2991 return -EIO;
2994 ret = qcow2_cache_get(bs, s->refcount_block_cache, refblock_offset,
2995 &refblock);
2996 if (ret < 0) {
2997 error_setg_errno(errp, -ret, "Failed to retrieve refblock");
2998 return ret;
3001 for (refblock_index = 0; refblock_index < s->refcount_block_size;
3002 refblock_index++)
3004 uint64_t refcount;
3006 if (new_refblock_index >= new_refblock_size) {
3007 /* new_refblock is now complete */
3008 ret = operation(bs, new_reftable, *new_reftable_index,
3009 new_reftable_size, new_refblock,
3010 new_refblock_empty, allocated, errp);
3011 if (ret < 0) {
3012 qcow2_cache_put(s->refcount_block_cache, &refblock);
3013 return ret;
3016 (*new_reftable_index)++;
3017 new_refblock_index = 0;
3018 new_refblock_empty = true;
3021 refcount = s->get_refcount(refblock, refblock_index);
3022 if (new_refcount_bits < 64 && refcount >> new_refcount_bits) {
3023 uint64_t offset;
3025 qcow2_cache_put(s->refcount_block_cache, &refblock);
3027 offset = ((reftable_index << s->refcount_block_bits)
3028 + refblock_index) << s->cluster_bits;
3030 error_setg(errp, "Cannot decrease refcount entry width to "
3031 "%i bits: Cluster at offset %#" PRIx64 " has a "
3032 "refcount of %" PRIu64, new_refcount_bits,
3033 offset, refcount);
3034 return -EINVAL;
3037 if (new_set_refcount) {
3038 new_set_refcount(new_refblock, new_refblock_index++,
3039 refcount);
3040 } else {
3041 new_refblock_index++;
3043 new_refblock_empty = new_refblock_empty && refcount == 0;
3046 qcow2_cache_put(s->refcount_block_cache, &refblock);
3047 } else {
3048 /* No refblock means every refcount is 0 */
3049 for (refblock_index = 0; refblock_index < s->refcount_block_size;
3050 refblock_index++)
3052 if (new_refblock_index >= new_refblock_size) {
3053 /* new_refblock is now complete */
3054 ret = operation(bs, new_reftable, *new_reftable_index,
3055 new_reftable_size, new_refblock,
3056 new_refblock_empty, allocated, errp);
3057 if (ret < 0) {
3058 return ret;
3061 (*new_reftable_index)++;
3062 new_refblock_index = 0;
3063 new_refblock_empty = true;
3066 if (new_set_refcount) {
3067 new_set_refcount(new_refblock, new_refblock_index++, 0);
3068 } else {
3069 new_refblock_index++;
3075 if (new_refblock_index > 0) {
3076 /* Complete the potentially existing partially filled final refblock */
3077 if (new_set_refcount) {
3078 for (; new_refblock_index < new_refblock_size;
3079 new_refblock_index++)
3081 new_set_refcount(new_refblock, new_refblock_index, 0);
3085 ret = operation(bs, new_reftable, *new_reftable_index,
3086 new_reftable_size, new_refblock, new_refblock_empty,
3087 allocated, errp);
3088 if (ret < 0) {
3089 return ret;
3092 (*new_reftable_index)++;
3095 status_cb(bs, (uint64_t)(index + 1) * s->refcount_table_size,
3096 (uint64_t)total * s->refcount_table_size, cb_opaque);
3098 return 0;
3101 int qcow2_change_refcount_order(BlockDriverState *bs, int refcount_order,
3102 BlockDriverAmendStatusCB *status_cb,
3103 void *cb_opaque, Error **errp)
3105 BDRVQcow2State *s = bs->opaque;
3106 Qcow2GetRefcountFunc *new_get_refcount;
3107 Qcow2SetRefcountFunc *new_set_refcount;
3108 void *new_refblock = qemu_blockalign(bs->file->bs, s->cluster_size);
3109 uint64_t *new_reftable = NULL, new_reftable_size = 0;
3110 uint64_t *old_reftable, old_reftable_size, old_reftable_offset;
3111 uint64_t new_reftable_index = 0;
3112 uint64_t i;
3113 int64_t new_reftable_offset = 0, allocated_reftable_size = 0;
3114 int new_refblock_size, new_refcount_bits = 1 << refcount_order;
3115 int old_refcount_order;
3116 int walk_index = 0;
3117 int ret;
3118 bool new_allocation;
3120 assert(s->qcow_version >= 3);
3121 assert(refcount_order >= 0 && refcount_order <= 6);
3123 /* see qcow2_open() */
3124 new_refblock_size = 1 << (s->cluster_bits - (refcount_order - 3));
3126 new_get_refcount = get_refcount_funcs[refcount_order];
3127 new_set_refcount = set_refcount_funcs[refcount_order];
3130 do {
3131 int total_walks;
3133 new_allocation = false;
3135 /* At least we have to do this walk and the one which writes the
3136 * refblocks; also, at least we have to do this loop here at least
3137 * twice (normally), first to do the allocations, and second to
3138 * determine that everything is correctly allocated, this then makes
3139 * three walks in total */
3140 total_walks = MAX(walk_index + 2, 3);
3142 /* First, allocate the structures so they are present in the refcount
3143 * structures */
3144 ret = walk_over_reftable(bs, &new_reftable, &new_reftable_index,
3145 &new_reftable_size, NULL, new_refblock_size,
3146 new_refcount_bits, &alloc_refblock,
3147 &new_allocation, NULL, status_cb, cb_opaque,
3148 walk_index++, total_walks, errp);
3149 if (ret < 0) {
3150 goto done;
3153 new_reftable_index = 0;
3155 if (new_allocation) {
3156 if (new_reftable_offset) {
3157 qcow2_free_clusters(bs, new_reftable_offset,
3158 allocated_reftable_size * sizeof(uint64_t),
3159 QCOW2_DISCARD_NEVER);
3162 new_reftable_offset = qcow2_alloc_clusters(bs, new_reftable_size *
3163 sizeof(uint64_t));
3164 if (new_reftable_offset < 0) {
3165 error_setg_errno(errp, -new_reftable_offset,
3166 "Failed to allocate the new reftable");
3167 ret = new_reftable_offset;
3168 goto done;
3170 allocated_reftable_size = new_reftable_size;
3172 } while (new_allocation);
3174 /* Second, write the new refblocks */
3175 ret = walk_over_reftable(bs, &new_reftable, &new_reftable_index,
3176 &new_reftable_size, new_refblock,
3177 new_refblock_size, new_refcount_bits,
3178 &flush_refblock, &new_allocation, new_set_refcount,
3179 status_cb, cb_opaque, walk_index, walk_index + 1,
3180 errp);
3181 if (ret < 0) {
3182 goto done;
3184 assert(!new_allocation);
3187 /* Write the new reftable */
3188 ret = qcow2_pre_write_overlap_check(bs, 0, new_reftable_offset,
3189 new_reftable_size * sizeof(uint64_t),
3190 false);
3191 if (ret < 0) {
3192 error_setg_errno(errp, -ret, "Overlap check failed");
3193 goto done;
3196 for (i = 0; i < new_reftable_size; i++) {
3197 cpu_to_be64s(&new_reftable[i]);
3200 ret = bdrv_pwrite(bs->file, new_reftable_offset, new_reftable,
3201 new_reftable_size * sizeof(uint64_t));
3203 for (i = 0; i < new_reftable_size; i++) {
3204 be64_to_cpus(&new_reftable[i]);
3207 if (ret < 0) {
3208 error_setg_errno(errp, -ret, "Failed to write the new reftable");
3209 goto done;
3213 /* Empty the refcount cache */
3214 ret = qcow2_cache_flush(bs, s->refcount_block_cache);
3215 if (ret < 0) {
3216 error_setg_errno(errp, -ret, "Failed to flush the refblock cache");
3217 goto done;
3220 /* Update the image header to point to the new reftable; this only updates
3221 * the fields which are relevant to qcow2_update_header(); other fields
3222 * such as s->refcount_table or s->refcount_bits stay stale for now
3223 * (because we have to restore everything if qcow2_update_header() fails) */
3224 old_refcount_order = s->refcount_order;
3225 old_reftable_size = s->refcount_table_size;
3226 old_reftable_offset = s->refcount_table_offset;
3228 s->refcount_order = refcount_order;
3229 s->refcount_table_size = new_reftable_size;
3230 s->refcount_table_offset = new_reftable_offset;
3232 ret = qcow2_update_header(bs);
3233 if (ret < 0) {
3234 s->refcount_order = old_refcount_order;
3235 s->refcount_table_size = old_reftable_size;
3236 s->refcount_table_offset = old_reftable_offset;
3237 error_setg_errno(errp, -ret, "Failed to update the qcow2 header");
3238 goto done;
3241 /* Now update the rest of the in-memory information */
3242 old_reftable = s->refcount_table;
3243 s->refcount_table = new_reftable;
3244 update_max_refcount_table_index(s);
3246 s->refcount_bits = 1 << refcount_order;
3247 s->refcount_max = UINT64_C(1) << (s->refcount_bits - 1);
3248 s->refcount_max += s->refcount_max - 1;
3250 s->refcount_block_bits = s->cluster_bits - (refcount_order - 3);
3251 s->refcount_block_size = 1 << s->refcount_block_bits;
3253 s->get_refcount = new_get_refcount;
3254 s->set_refcount = new_set_refcount;
3256 /* For cleaning up all old refblocks and the old reftable below the "done"
3257 * label */
3258 new_reftable = old_reftable;
3259 new_reftable_size = old_reftable_size;
3260 new_reftable_offset = old_reftable_offset;
3262 done:
3263 if (new_reftable) {
3264 /* On success, new_reftable actually points to the old reftable (and
3265 * new_reftable_size is the old reftable's size); but that is just
3266 * fine */
3267 for (i = 0; i < new_reftable_size; i++) {
3268 uint64_t offset = new_reftable[i] & REFT_OFFSET_MASK;
3269 if (offset) {
3270 qcow2_free_clusters(bs, offset, s->cluster_size,
3271 QCOW2_DISCARD_OTHER);
3274 g_free(new_reftable);
3276 if (new_reftable_offset > 0) {
3277 qcow2_free_clusters(bs, new_reftable_offset,
3278 new_reftable_size * sizeof(uint64_t),
3279 QCOW2_DISCARD_OTHER);
3283 qemu_vfree(new_refblock);
3284 return ret;
3287 static int64_t get_refblock_offset(BlockDriverState *bs, uint64_t offset)
3289 BDRVQcow2State *s = bs->opaque;
3290 uint32_t index = offset_to_reftable_index(s, offset);
3291 int64_t covering_refblock_offset = 0;
3293 if (index < s->refcount_table_size) {
3294 covering_refblock_offset = s->refcount_table[index] & REFT_OFFSET_MASK;
3296 if (!covering_refblock_offset) {
3297 qcow2_signal_corruption(bs, true, -1, -1, "Refblock at %#" PRIx64 " is "
3298 "not covered by the refcount structures",
3299 offset);
3300 return -EIO;
3303 return covering_refblock_offset;
3306 static int qcow2_discard_refcount_block(BlockDriverState *bs,
3307 uint64_t discard_block_offs)
3309 BDRVQcow2State *s = bs->opaque;
3310 int64_t refblock_offs;
3311 uint64_t cluster_index = discard_block_offs >> s->cluster_bits;
3312 uint32_t block_index = cluster_index & (s->refcount_block_size - 1);
3313 void *refblock;
3314 int ret;
3316 refblock_offs = get_refblock_offset(bs, discard_block_offs);
3317 if (refblock_offs < 0) {
3318 return refblock_offs;
3321 assert(discard_block_offs != 0);
3323 ret = qcow2_cache_get(bs, s->refcount_block_cache, refblock_offs,
3324 &refblock);
3325 if (ret < 0) {
3326 return ret;
3329 if (s->get_refcount(refblock, block_index) != 1) {
3330 qcow2_signal_corruption(bs, true, -1, -1, "Invalid refcount:"
3331 " refblock offset %#" PRIx64
3332 ", reftable index %u"
3333 ", block offset %#" PRIx64
3334 ", refcount %#" PRIx64,
3335 refblock_offs,
3336 offset_to_reftable_index(s, discard_block_offs),
3337 discard_block_offs,
3338 s->get_refcount(refblock, block_index));
3339 qcow2_cache_put(s->refcount_block_cache, &refblock);
3340 return -EINVAL;
3342 s->set_refcount(refblock, block_index, 0);
3344 qcow2_cache_entry_mark_dirty(s->refcount_block_cache, refblock);
3346 qcow2_cache_put(s->refcount_block_cache, &refblock);
3348 if (cluster_index < s->free_cluster_index) {
3349 s->free_cluster_index = cluster_index;
3352 refblock = qcow2_cache_is_table_offset(s->refcount_block_cache,
3353 discard_block_offs);
3354 if (refblock) {
3355 /* discard refblock from the cache if refblock is cached */
3356 qcow2_cache_discard(s->refcount_block_cache, refblock);
3358 update_refcount_discard(bs, discard_block_offs, s->cluster_size);
3360 return 0;
3363 int qcow2_shrink_reftable(BlockDriverState *bs)
3365 BDRVQcow2State *s = bs->opaque;
3366 uint64_t *reftable_tmp =
3367 g_malloc(s->refcount_table_size * sizeof(uint64_t));
3368 int i, ret;
3370 for (i = 0; i < s->refcount_table_size; i++) {
3371 int64_t refblock_offs = s->refcount_table[i] & REFT_OFFSET_MASK;
3372 void *refblock;
3373 bool unused_block;
3375 if (refblock_offs == 0) {
3376 reftable_tmp[i] = 0;
3377 continue;
3379 ret = qcow2_cache_get(bs, s->refcount_block_cache, refblock_offs,
3380 &refblock);
3381 if (ret < 0) {
3382 goto out;
3385 /* the refblock has own reference */
3386 if (i == offset_to_reftable_index(s, refblock_offs)) {
3387 uint64_t block_index = (refblock_offs >> s->cluster_bits) &
3388 (s->refcount_block_size - 1);
3389 uint64_t refcount = s->get_refcount(refblock, block_index);
3391 s->set_refcount(refblock, block_index, 0);
3393 unused_block = buffer_is_zero(refblock, s->cluster_size);
3395 s->set_refcount(refblock, block_index, refcount);
3396 } else {
3397 unused_block = buffer_is_zero(refblock, s->cluster_size);
3399 qcow2_cache_put(s->refcount_block_cache, &refblock);
3401 reftable_tmp[i] = unused_block ? 0 : cpu_to_be64(s->refcount_table[i]);
3404 ret = bdrv_pwrite_sync(bs->file, s->refcount_table_offset, reftable_tmp,
3405 s->refcount_table_size * sizeof(uint64_t));
3407 * If the write in the reftable failed the image may contain a partially
3408 * overwritten reftable. In this case it would be better to clear the
3409 * reftable in memory to avoid possible image corruption.
3411 for (i = 0; i < s->refcount_table_size; i++) {
3412 if (s->refcount_table[i] && !reftable_tmp[i]) {
3413 if (ret == 0) {
3414 ret = qcow2_discard_refcount_block(bs, s->refcount_table[i] &
3415 REFT_OFFSET_MASK);
3417 s->refcount_table[i] = 0;
3421 if (!s->cache_discards) {
3422 qcow2_process_discards(bs, ret);
3425 out:
3426 g_free(reftable_tmp);
3427 return ret;
3430 int64_t qcow2_get_last_cluster(BlockDriverState *bs, int64_t size)
3432 BDRVQcow2State *s = bs->opaque;
3433 int64_t i;
3435 for (i = size_to_clusters(s, size) - 1; i >= 0; i--) {
3436 uint64_t refcount;
3437 int ret = qcow2_get_refcount(bs, i, &refcount);
3438 if (ret < 0) {
3439 fprintf(stderr, "Can't get refcount for cluster %" PRId64 ": %s\n",
3440 i, strerror(-ret));
3441 return ret;
3443 if (refcount > 0) {
3444 return i;
3447 qcow2_signal_corruption(bs, true, -1, -1,
3448 "There are no references in the refcount table.");
3449 return -EIO;
3452 int qcow2_detect_metadata_preallocation(BlockDriverState *bs)
3454 BDRVQcow2State *s = bs->opaque;
3455 int64_t i, end_cluster, cluster_count = 0, threshold;
3456 int64_t file_length, real_allocation, real_clusters;
3458 file_length = bdrv_getlength(bs->file->bs);
3459 if (file_length < 0) {
3460 return file_length;
3463 real_allocation = bdrv_get_allocated_file_size(bs->file->bs);
3464 if (real_allocation < 0) {
3465 return real_allocation;
3468 real_clusters = real_allocation / s->cluster_size;
3469 threshold = MAX(real_clusters * 10 / 9, real_clusters + 2);
3471 end_cluster = size_to_clusters(s, file_length);
3472 for (i = 0; i < end_cluster && cluster_count < threshold; i++) {
3473 uint64_t refcount;
3474 int ret = qcow2_get_refcount(bs, i, &refcount);
3475 if (ret < 0) {
3476 return ret;
3478 cluster_count += !!refcount;
3481 return cluster_count >= threshold;