tcg/i386: rdpmc: fix the the condtions
[qemu/ar7.git] / block / qcow2-refcount.c
blob8e649b008e8cd6b5f36f35e156e8d4a63e0b09b6
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 / REFTABLE_ENTRY_SIZE);
109 refcount_table_size2 = s->refcount_table_size * REFTABLE_ENTRY_SIZE;
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, s->refcount_table_offset +
438 refcount_table_index * REFTABLE_ENTRY_SIZE,
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 / REFTABLE_ENTRY_SIZE);
566 table_clusters = (table_size * REFTABLE_ENTRY_SIZE) / 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) * REFTABLE_ENTRY_SIZE);
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 * REFTABLE_ENTRY_SIZE);
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 * REFTABLE_ENTRY_SIZE);
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,
721 old_table_size * REFTABLE_ENTRY_SIZE,
722 QCOW2_DISCARD_OTHER);
724 return end_offset;
726 fail:
727 g_free(new_table);
728 return ret;
731 void qcow2_process_discards(BlockDriverState *bs, int ret)
733 BDRVQcow2State *s = bs->opaque;
734 Qcow2DiscardRegion *d, *next;
736 QTAILQ_FOREACH_SAFE(d, &s->discards, next, next) {
737 QTAILQ_REMOVE(&s->discards, d, next);
739 /* Discard is optional, ignore the return value */
740 if (ret >= 0) {
741 int r2 = bdrv_pdiscard(bs->file, d->offset, d->bytes);
742 if (r2 < 0) {
743 trace_qcow2_process_discards_failed_region(d->offset, d->bytes,
744 r2);
748 g_free(d);
752 static void update_refcount_discard(BlockDriverState *bs,
753 uint64_t offset, uint64_t length)
755 BDRVQcow2State *s = bs->opaque;
756 Qcow2DiscardRegion *d, *p, *next;
758 QTAILQ_FOREACH(d, &s->discards, next) {
759 uint64_t new_start = MIN(offset, d->offset);
760 uint64_t new_end = MAX(offset + length, d->offset + d->bytes);
762 if (new_end - new_start <= length + d->bytes) {
763 /* There can't be any overlap, areas ending up here have no
764 * references any more and therefore shouldn't get freed another
765 * time. */
766 assert(d->bytes + length == new_end - new_start);
767 d->offset = new_start;
768 d->bytes = new_end - new_start;
769 goto found;
773 d = g_malloc(sizeof(*d));
774 *d = (Qcow2DiscardRegion) {
775 .bs = bs,
776 .offset = offset,
777 .bytes = length,
779 QTAILQ_INSERT_TAIL(&s->discards, d, next);
781 found:
782 /* Merge discard requests if they are adjacent now */
783 QTAILQ_FOREACH_SAFE(p, &s->discards, next, next) {
784 if (p == d
785 || p->offset > d->offset + d->bytes
786 || d->offset > p->offset + p->bytes)
788 continue;
791 /* Still no overlap possible */
792 assert(p->offset == d->offset + d->bytes
793 || d->offset == p->offset + p->bytes);
795 QTAILQ_REMOVE(&s->discards, p, next);
796 d->offset = MIN(d->offset, p->offset);
797 d->bytes += p->bytes;
798 g_free(p);
802 /* XXX: cache several refcount block clusters ? */
803 /* @addend is the absolute value of the addend; if @decrease is set, @addend
804 * will be subtracted from the current refcount, otherwise it will be added */
805 static int QEMU_WARN_UNUSED_RESULT update_refcount(BlockDriverState *bs,
806 int64_t offset,
807 int64_t length,
808 uint64_t addend,
809 bool decrease,
810 enum qcow2_discard_type type)
812 BDRVQcow2State *s = bs->opaque;
813 int64_t start, last, cluster_offset;
814 void *refcount_block = NULL;
815 int64_t old_table_index = -1;
816 int ret;
818 #ifdef DEBUG_ALLOC2
819 fprintf(stderr, "update_refcount: offset=%" PRId64 " size=%" PRId64
820 " addend=%s%" PRIu64 "\n", offset, length, decrease ? "-" : "",
821 addend);
822 #endif
823 if (length < 0) {
824 return -EINVAL;
825 } else if (length == 0) {
826 return 0;
829 if (decrease) {
830 qcow2_cache_set_dependency(bs, s->refcount_block_cache,
831 s->l2_table_cache);
834 start = start_of_cluster(s, offset);
835 last = start_of_cluster(s, offset + length - 1);
836 for(cluster_offset = start; cluster_offset <= last;
837 cluster_offset += s->cluster_size)
839 int block_index;
840 uint64_t refcount;
841 int64_t cluster_index = cluster_offset >> s->cluster_bits;
842 int64_t table_index = cluster_index >> s->refcount_block_bits;
844 /* Load the refcount block and allocate it if needed */
845 if (table_index != old_table_index) {
846 if (refcount_block) {
847 qcow2_cache_put(s->refcount_block_cache, &refcount_block);
849 ret = alloc_refcount_block(bs, cluster_index, &refcount_block);
850 /* If the caller needs to restart the search for free clusters,
851 * try the same ones first to see if they're still free. */
852 if (ret == -EAGAIN) {
853 if (s->free_cluster_index > (start >> s->cluster_bits)) {
854 s->free_cluster_index = (start >> s->cluster_bits);
857 if (ret < 0) {
858 goto fail;
861 old_table_index = table_index;
863 qcow2_cache_entry_mark_dirty(s->refcount_block_cache, refcount_block);
865 /* we can update the count and save it */
866 block_index = cluster_index & (s->refcount_block_size - 1);
868 refcount = s->get_refcount(refcount_block, block_index);
869 if (decrease ? (refcount - addend > refcount)
870 : (refcount + addend < refcount ||
871 refcount + addend > s->refcount_max))
873 ret = -EINVAL;
874 goto fail;
876 if (decrease) {
877 refcount -= addend;
878 } else {
879 refcount += addend;
881 if (refcount == 0 && cluster_index < s->free_cluster_index) {
882 s->free_cluster_index = cluster_index;
884 s->set_refcount(refcount_block, block_index, refcount);
886 if (refcount == 0) {
887 void *table;
889 table = qcow2_cache_is_table_offset(s->refcount_block_cache,
890 offset);
891 if (table != NULL) {
892 qcow2_cache_put(s->refcount_block_cache, &refcount_block);
893 old_table_index = -1;
894 qcow2_cache_discard(s->refcount_block_cache, table);
897 table = qcow2_cache_is_table_offset(s->l2_table_cache, offset);
898 if (table != NULL) {
899 qcow2_cache_discard(s->l2_table_cache, table);
902 if (s->discard_passthrough[type]) {
903 update_refcount_discard(bs, cluster_offset, s->cluster_size);
908 ret = 0;
909 fail:
910 if (!s->cache_discards) {
911 qcow2_process_discards(bs, ret);
914 /* Write last changed block to disk */
915 if (refcount_block) {
916 qcow2_cache_put(s->refcount_block_cache, &refcount_block);
920 * Try do undo any updates if an error is returned (This may succeed in
921 * some cases like ENOSPC for allocating a new refcount block)
923 if (ret < 0) {
924 int dummy;
925 dummy = update_refcount(bs, offset, cluster_offset - offset, addend,
926 !decrease, QCOW2_DISCARD_NEVER);
927 (void)dummy;
930 return ret;
934 * Increases or decreases the refcount of a given cluster.
936 * @addend is the absolute value of the addend; if @decrease is set, @addend
937 * will be subtracted from the current refcount, otherwise it will be added.
939 * On success 0 is returned; on failure -errno is returned.
941 int qcow2_update_cluster_refcount(BlockDriverState *bs,
942 int64_t cluster_index,
943 uint64_t addend, bool decrease,
944 enum qcow2_discard_type type)
946 BDRVQcow2State *s = bs->opaque;
947 int ret;
949 ret = update_refcount(bs, cluster_index << s->cluster_bits, 1, addend,
950 decrease, type);
951 if (ret < 0) {
952 return ret;
955 return 0;
960 /*********************************************************/
961 /* cluster allocation functions */
965 /* return < 0 if error */
966 static int64_t alloc_clusters_noref(BlockDriverState *bs, uint64_t size,
967 uint64_t max)
969 BDRVQcow2State *s = bs->opaque;
970 uint64_t i, nb_clusters, refcount;
971 int ret;
973 /* We can't allocate clusters if they may still be queued for discard. */
974 if (s->cache_discards) {
975 qcow2_process_discards(bs, 0);
978 nb_clusters = size_to_clusters(s, size);
979 retry:
980 for(i = 0; i < nb_clusters; i++) {
981 uint64_t next_cluster_index = s->free_cluster_index++;
982 ret = qcow2_get_refcount(bs, next_cluster_index, &refcount);
984 if (ret < 0) {
985 return ret;
986 } else if (refcount != 0) {
987 goto retry;
991 /* Make sure that all offsets in the "allocated" range are representable
992 * in the requested max */
993 if (s->free_cluster_index > 0 &&
994 s->free_cluster_index - 1 > (max >> s->cluster_bits))
996 return -EFBIG;
999 #ifdef DEBUG_ALLOC2
1000 fprintf(stderr, "alloc_clusters: size=%" PRId64 " -> %" PRId64 "\n",
1001 size,
1002 (s->free_cluster_index - nb_clusters) << s->cluster_bits);
1003 #endif
1004 return (s->free_cluster_index - nb_clusters) << s->cluster_bits;
1007 int64_t qcow2_alloc_clusters(BlockDriverState *bs, uint64_t size)
1009 int64_t offset;
1010 int ret;
1012 BLKDBG_EVENT(bs->file, BLKDBG_CLUSTER_ALLOC);
1013 do {
1014 offset = alloc_clusters_noref(bs, size, QCOW_MAX_CLUSTER_OFFSET);
1015 if (offset < 0) {
1016 return offset;
1019 ret = update_refcount(bs, offset, size, 1, false, QCOW2_DISCARD_NEVER);
1020 } while (ret == -EAGAIN);
1022 if (ret < 0) {
1023 return ret;
1026 return offset;
1029 int64_t qcow2_alloc_clusters_at(BlockDriverState *bs, uint64_t offset,
1030 int64_t nb_clusters)
1032 BDRVQcow2State *s = bs->opaque;
1033 uint64_t cluster_index, refcount;
1034 uint64_t i;
1035 int ret;
1037 assert(nb_clusters >= 0);
1038 if (nb_clusters == 0) {
1039 return 0;
1042 do {
1043 /* Check how many clusters there are free */
1044 cluster_index = offset >> s->cluster_bits;
1045 for(i = 0; i < nb_clusters; i++) {
1046 ret = qcow2_get_refcount(bs, cluster_index++, &refcount);
1047 if (ret < 0) {
1048 return ret;
1049 } else if (refcount != 0) {
1050 break;
1054 /* And then allocate them */
1055 ret = update_refcount(bs, offset, i << s->cluster_bits, 1, false,
1056 QCOW2_DISCARD_NEVER);
1057 } while (ret == -EAGAIN);
1059 if (ret < 0) {
1060 return ret;
1063 return i;
1066 /* only used to allocate compressed sectors. We try to allocate
1067 contiguous sectors. size must be <= cluster_size */
1068 int64_t qcow2_alloc_bytes(BlockDriverState *bs, int size)
1070 BDRVQcow2State *s = bs->opaque;
1071 int64_t offset;
1072 size_t free_in_cluster;
1073 int ret;
1075 BLKDBG_EVENT(bs->file, BLKDBG_CLUSTER_ALLOC_BYTES);
1076 assert(size > 0 && size <= s->cluster_size);
1077 assert(!s->free_byte_offset || offset_into_cluster(s, s->free_byte_offset));
1079 offset = s->free_byte_offset;
1081 if (offset) {
1082 uint64_t refcount;
1083 ret = qcow2_get_refcount(bs, offset >> s->cluster_bits, &refcount);
1084 if (ret < 0) {
1085 return ret;
1088 if (refcount == s->refcount_max) {
1089 offset = 0;
1093 free_in_cluster = s->cluster_size - offset_into_cluster(s, offset);
1094 do {
1095 if (!offset || free_in_cluster < size) {
1096 int64_t new_cluster;
1098 new_cluster = alloc_clusters_noref(bs, s->cluster_size,
1099 MIN(s->cluster_offset_mask,
1100 QCOW_MAX_CLUSTER_OFFSET));
1101 if (new_cluster < 0) {
1102 return new_cluster;
1105 if (new_cluster == 0) {
1106 qcow2_signal_corruption(bs, true, -1, -1, "Preventing invalid "
1107 "allocation of compressed cluster "
1108 "at offset 0");
1109 return -EIO;
1112 if (!offset || ROUND_UP(offset, s->cluster_size) != new_cluster) {
1113 offset = new_cluster;
1114 free_in_cluster = s->cluster_size;
1115 } else {
1116 free_in_cluster += s->cluster_size;
1120 assert(offset);
1121 ret = update_refcount(bs, offset, size, 1, false, QCOW2_DISCARD_NEVER);
1122 if (ret < 0) {
1123 offset = 0;
1125 } while (ret == -EAGAIN);
1126 if (ret < 0) {
1127 return ret;
1130 /* The cluster refcount was incremented; refcount blocks must be flushed
1131 * before the caller's L2 table updates. */
1132 qcow2_cache_set_dependency(bs, s->l2_table_cache, s->refcount_block_cache);
1134 s->free_byte_offset = offset + size;
1135 if (!offset_into_cluster(s, s->free_byte_offset)) {
1136 s->free_byte_offset = 0;
1139 return offset;
1142 void qcow2_free_clusters(BlockDriverState *bs,
1143 int64_t offset, int64_t size,
1144 enum qcow2_discard_type type)
1146 int ret;
1148 BLKDBG_EVENT(bs->file, BLKDBG_CLUSTER_FREE);
1149 ret = update_refcount(bs, offset, size, 1, true, type);
1150 if (ret < 0) {
1151 fprintf(stderr, "qcow2_free_clusters failed: %s\n", strerror(-ret));
1152 /* TODO Remember the clusters to free them later and avoid leaking */
1157 * Free a cluster using its L2 entry (handles clusters of all types, e.g.
1158 * normal cluster, compressed cluster, etc.)
1160 void qcow2_free_any_cluster(BlockDriverState *bs, uint64_t l2_entry,
1161 enum qcow2_discard_type type)
1163 BDRVQcow2State *s = bs->opaque;
1164 QCow2ClusterType ctype = qcow2_get_cluster_type(bs, l2_entry);
1166 if (has_data_file(bs)) {
1167 if (s->discard_passthrough[type] &&
1168 (ctype == QCOW2_CLUSTER_NORMAL ||
1169 ctype == QCOW2_CLUSTER_ZERO_ALLOC))
1171 bdrv_pdiscard(s->data_file, l2_entry & L2E_OFFSET_MASK,
1172 s->cluster_size);
1174 return;
1177 switch (ctype) {
1178 case QCOW2_CLUSTER_COMPRESSED:
1180 int64_t offset = (l2_entry & s->cluster_offset_mask)
1181 & QCOW2_COMPRESSED_SECTOR_MASK;
1182 int size = QCOW2_COMPRESSED_SECTOR_SIZE *
1183 (((l2_entry >> s->csize_shift) & s->csize_mask) + 1);
1184 qcow2_free_clusters(bs, offset, size, type);
1186 break;
1187 case QCOW2_CLUSTER_NORMAL:
1188 case QCOW2_CLUSTER_ZERO_ALLOC:
1189 if (offset_into_cluster(s, l2_entry & L2E_OFFSET_MASK)) {
1190 qcow2_signal_corruption(bs, false, -1, -1,
1191 "Cannot free unaligned cluster %#llx",
1192 l2_entry & L2E_OFFSET_MASK);
1193 } else {
1194 qcow2_free_clusters(bs, l2_entry & L2E_OFFSET_MASK,
1195 s->cluster_size, type);
1197 break;
1198 case QCOW2_CLUSTER_ZERO_PLAIN:
1199 case QCOW2_CLUSTER_UNALLOCATED:
1200 break;
1201 default:
1202 abort();
1206 int coroutine_fn qcow2_write_caches(BlockDriverState *bs)
1208 BDRVQcow2State *s = bs->opaque;
1209 int ret;
1211 ret = qcow2_cache_write(bs, s->l2_table_cache);
1212 if (ret < 0) {
1213 return ret;
1216 if (qcow2_need_accurate_refcounts(s)) {
1217 ret = qcow2_cache_write(bs, s->refcount_block_cache);
1218 if (ret < 0) {
1219 return ret;
1223 return 0;
1226 int coroutine_fn qcow2_flush_caches(BlockDriverState *bs)
1228 int ret = qcow2_write_caches(bs);
1229 if (ret < 0) {
1230 return ret;
1233 return bdrv_flush(bs->file->bs);
1236 /*********************************************************/
1237 /* snapshots and image creation */
1241 /* update the refcounts of snapshots and the copied flag */
1242 int qcow2_update_snapshot_refcount(BlockDriverState *bs,
1243 int64_t l1_table_offset, int l1_size, int addend)
1245 BDRVQcow2State *s = bs->opaque;
1246 uint64_t *l1_table, *l2_slice, l2_offset, entry, l1_size2, refcount;
1247 bool l1_allocated = false;
1248 int64_t old_entry, old_l2_offset;
1249 unsigned slice, slice_size2, n_slices;
1250 int i, j, l1_modified = 0, nb_csectors;
1251 int ret;
1253 assert(addend >= -1 && addend <= 1);
1255 l2_slice = NULL;
1256 l1_table = NULL;
1257 l1_size2 = l1_size * L1E_SIZE;
1258 slice_size2 = s->l2_slice_size * l2_entry_size(s);
1259 n_slices = s->cluster_size / slice_size2;
1261 s->cache_discards = true;
1263 /* WARNING: qcow2_snapshot_goto relies on this function not using the
1264 * l1_table_offset when it is the current s->l1_table_offset! Be careful
1265 * when changing this! */
1266 if (l1_table_offset != s->l1_table_offset) {
1267 l1_table = g_try_malloc0(l1_size2);
1268 if (l1_size2 && l1_table == NULL) {
1269 ret = -ENOMEM;
1270 goto fail;
1272 l1_allocated = true;
1274 ret = bdrv_pread(bs->file, l1_table_offset, l1_table, l1_size2);
1275 if (ret < 0) {
1276 goto fail;
1279 for (i = 0; i < l1_size; i++) {
1280 be64_to_cpus(&l1_table[i]);
1282 } else {
1283 assert(l1_size == s->l1_size);
1284 l1_table = s->l1_table;
1285 l1_allocated = false;
1288 for (i = 0; i < l1_size; i++) {
1289 l2_offset = l1_table[i];
1290 if (l2_offset) {
1291 old_l2_offset = l2_offset;
1292 l2_offset &= L1E_OFFSET_MASK;
1294 if (offset_into_cluster(s, l2_offset)) {
1295 qcow2_signal_corruption(bs, true, -1, -1, "L2 table offset %#"
1296 PRIx64 " unaligned (L1 index: %#x)",
1297 l2_offset, i);
1298 ret = -EIO;
1299 goto fail;
1302 for (slice = 0; slice < n_slices; slice++) {
1303 ret = qcow2_cache_get(bs, s->l2_table_cache,
1304 l2_offset + slice * slice_size2,
1305 (void **) &l2_slice);
1306 if (ret < 0) {
1307 goto fail;
1310 for (j = 0; j < s->l2_slice_size; j++) {
1311 uint64_t cluster_index;
1312 uint64_t offset;
1314 entry = get_l2_entry(s, l2_slice, j);
1315 old_entry = entry;
1316 entry &= ~QCOW_OFLAG_COPIED;
1317 offset = entry & L2E_OFFSET_MASK;
1319 switch (qcow2_get_cluster_type(bs, entry)) {
1320 case QCOW2_CLUSTER_COMPRESSED:
1321 nb_csectors = ((entry >> s->csize_shift) &
1322 s->csize_mask) + 1;
1323 if (addend != 0) {
1324 uint64_t coffset = (entry & s->cluster_offset_mask)
1325 & QCOW2_COMPRESSED_SECTOR_MASK;
1326 ret = update_refcount(
1327 bs, coffset,
1328 nb_csectors * QCOW2_COMPRESSED_SECTOR_SIZE,
1329 abs(addend), addend < 0,
1330 QCOW2_DISCARD_SNAPSHOT);
1331 if (ret < 0) {
1332 goto fail;
1335 /* compressed clusters are never modified */
1336 refcount = 2;
1337 break;
1339 case QCOW2_CLUSTER_NORMAL:
1340 case QCOW2_CLUSTER_ZERO_ALLOC:
1341 if (offset_into_cluster(s, offset)) {
1342 /* Here l2_index means table (not slice) index */
1343 int l2_index = slice * s->l2_slice_size + j;
1344 qcow2_signal_corruption(
1345 bs, true, -1, -1, "Cluster "
1346 "allocation offset %#" PRIx64
1347 " unaligned (L2 offset: %#"
1348 PRIx64 ", L2 index: %#x)",
1349 offset, l2_offset, l2_index);
1350 ret = -EIO;
1351 goto fail;
1354 cluster_index = offset >> s->cluster_bits;
1355 assert(cluster_index);
1356 if (addend != 0) {
1357 ret = qcow2_update_cluster_refcount(
1358 bs, cluster_index, abs(addend), addend < 0,
1359 QCOW2_DISCARD_SNAPSHOT);
1360 if (ret < 0) {
1361 goto fail;
1365 ret = qcow2_get_refcount(bs, cluster_index, &refcount);
1366 if (ret < 0) {
1367 goto fail;
1369 break;
1371 case QCOW2_CLUSTER_ZERO_PLAIN:
1372 case QCOW2_CLUSTER_UNALLOCATED:
1373 refcount = 0;
1374 break;
1376 default:
1377 abort();
1380 if (refcount == 1) {
1381 entry |= QCOW_OFLAG_COPIED;
1383 if (entry != old_entry) {
1384 if (addend > 0) {
1385 qcow2_cache_set_dependency(bs, s->l2_table_cache,
1386 s->refcount_block_cache);
1388 set_l2_entry(s, l2_slice, j, entry);
1389 qcow2_cache_entry_mark_dirty(s->l2_table_cache,
1390 l2_slice);
1394 qcow2_cache_put(s->l2_table_cache, (void **) &l2_slice);
1397 if (addend != 0) {
1398 ret = qcow2_update_cluster_refcount(bs, l2_offset >>
1399 s->cluster_bits,
1400 abs(addend), addend < 0,
1401 QCOW2_DISCARD_SNAPSHOT);
1402 if (ret < 0) {
1403 goto fail;
1406 ret = qcow2_get_refcount(bs, l2_offset >> s->cluster_bits,
1407 &refcount);
1408 if (ret < 0) {
1409 goto fail;
1410 } else if (refcount == 1) {
1411 l2_offset |= QCOW_OFLAG_COPIED;
1413 if (l2_offset != old_l2_offset) {
1414 l1_table[i] = l2_offset;
1415 l1_modified = 1;
1420 ret = bdrv_flush(bs);
1421 fail:
1422 if (l2_slice) {
1423 qcow2_cache_put(s->l2_table_cache, (void **) &l2_slice);
1426 s->cache_discards = false;
1427 qcow2_process_discards(bs, ret);
1429 /* Update L1 only if it isn't deleted anyway (addend = -1) */
1430 if (ret == 0 && addend >= 0 && l1_modified) {
1431 for (i = 0; i < l1_size; i++) {
1432 cpu_to_be64s(&l1_table[i]);
1435 ret = bdrv_pwrite_sync(bs->file, l1_table_offset,
1436 l1_table, l1_size2);
1438 for (i = 0; i < l1_size; i++) {
1439 be64_to_cpus(&l1_table[i]);
1442 if (l1_allocated)
1443 g_free(l1_table);
1444 return ret;
1450 /*********************************************************/
1451 /* refcount checking functions */
1454 static uint64_t refcount_array_byte_size(BDRVQcow2State *s, uint64_t entries)
1456 /* This assertion holds because there is no way we can address more than
1457 * 2^(64 - 9) clusters at once (with cluster size 512 = 2^9, and because
1458 * offsets have to be representable in bytes); due to every cluster
1459 * corresponding to one refcount entry, we are well below that limit */
1460 assert(entries < (UINT64_C(1) << (64 - 9)));
1462 /* Thanks to the assertion this will not overflow, because
1463 * s->refcount_order < 7.
1464 * (note: x << s->refcount_order == x * s->refcount_bits) */
1465 return DIV_ROUND_UP(entries << s->refcount_order, 8);
1469 * Reallocates *array so that it can hold new_size entries. *size must contain
1470 * the current number of entries in *array. If the reallocation fails, *array
1471 * and *size will not be modified and -errno will be returned. If the
1472 * reallocation is successful, *array will be set to the new buffer, *size
1473 * will be set to new_size and 0 will be returned. The size of the reallocated
1474 * refcount array buffer will be aligned to a cluster boundary, and the newly
1475 * allocated area will be zeroed.
1477 static int realloc_refcount_array(BDRVQcow2State *s, void **array,
1478 int64_t *size, int64_t new_size)
1480 int64_t old_byte_size, new_byte_size;
1481 void *new_ptr;
1483 /* Round to clusters so the array can be directly written to disk */
1484 old_byte_size = size_to_clusters(s, refcount_array_byte_size(s, *size))
1485 * s->cluster_size;
1486 new_byte_size = size_to_clusters(s, refcount_array_byte_size(s, new_size))
1487 * s->cluster_size;
1489 if (new_byte_size == old_byte_size) {
1490 *size = new_size;
1491 return 0;
1494 assert(new_byte_size > 0);
1496 if (new_byte_size > SIZE_MAX) {
1497 return -ENOMEM;
1500 new_ptr = g_try_realloc(*array, new_byte_size);
1501 if (!new_ptr) {
1502 return -ENOMEM;
1505 if (new_byte_size > old_byte_size) {
1506 memset((char *)new_ptr + old_byte_size, 0,
1507 new_byte_size - old_byte_size);
1510 *array = new_ptr;
1511 *size = new_size;
1513 return 0;
1517 * Increases the refcount for a range of clusters in a given refcount table.
1518 * This is used to construct a temporary refcount table out of L1 and L2 tables
1519 * which can be compared to the refcount table saved in the image.
1521 * Modifies the number of errors in res.
1523 int qcow2_inc_refcounts_imrt(BlockDriverState *bs, BdrvCheckResult *res,
1524 void **refcount_table,
1525 int64_t *refcount_table_size,
1526 int64_t offset, int64_t size)
1528 BDRVQcow2State *s = bs->opaque;
1529 uint64_t start, last, cluster_offset, k, refcount;
1530 int64_t file_len;
1531 int ret;
1533 if (size <= 0) {
1534 return 0;
1537 file_len = bdrv_getlength(bs->file->bs);
1538 if (file_len < 0) {
1539 return file_len;
1543 * Last cluster of qcow2 image may be semi-allocated, so it may be OK to
1544 * reference some space after file end but it should be less than one
1545 * cluster.
1547 if (offset + size - file_len >= s->cluster_size) {
1548 fprintf(stderr, "ERROR: counting reference for region exceeding the "
1549 "end of the file by one cluster or more: offset 0x%" PRIx64
1550 " size 0x%" PRIx64 "\n", offset, size);
1551 res->corruptions++;
1552 return 0;
1555 start = start_of_cluster(s, offset);
1556 last = start_of_cluster(s, offset + size - 1);
1557 for(cluster_offset = start; cluster_offset <= last;
1558 cluster_offset += s->cluster_size) {
1559 k = cluster_offset >> s->cluster_bits;
1560 if (k >= *refcount_table_size) {
1561 ret = realloc_refcount_array(s, refcount_table,
1562 refcount_table_size, k + 1);
1563 if (ret < 0) {
1564 res->check_errors++;
1565 return ret;
1569 refcount = s->get_refcount(*refcount_table, k);
1570 if (refcount == s->refcount_max) {
1571 fprintf(stderr, "ERROR: overflow cluster offset=0x%" PRIx64
1572 "\n", cluster_offset);
1573 fprintf(stderr, "Use qemu-img amend to increase the refcount entry "
1574 "width or qemu-img convert to create a clean copy if the "
1575 "image cannot be opened for writing\n");
1576 res->corruptions++;
1577 continue;
1579 s->set_refcount(*refcount_table, k, refcount + 1);
1582 return 0;
1585 /* Flags for check_refcounts_l1() and check_refcounts_l2() */
1586 enum {
1587 CHECK_FRAG_INFO = 0x2, /* update BlockFragInfo counters */
1591 * Increases the refcount in the given refcount table for the all clusters
1592 * referenced in the L2 table. While doing so, performs some checks on L2
1593 * entries.
1595 * Returns the number of errors found by the checks or -errno if an internal
1596 * error occurred.
1598 static int check_refcounts_l2(BlockDriverState *bs, BdrvCheckResult *res,
1599 void **refcount_table,
1600 int64_t *refcount_table_size, int64_t l2_offset,
1601 int flags, BdrvCheckMode fix, bool active)
1603 BDRVQcow2State *s = bs->opaque;
1604 uint64_t *l2_table, l2_entry;
1605 uint64_t next_contiguous_offset = 0;
1606 int i, l2_size, nb_csectors, ret;
1608 /* Read L2 table from disk */
1609 l2_size = s->l2_size * l2_entry_size(s);
1610 l2_table = g_malloc(l2_size);
1612 ret = bdrv_pread(bs->file, l2_offset, l2_table, l2_size);
1613 if (ret < 0) {
1614 fprintf(stderr, "ERROR: I/O error in check_refcounts_l2\n");
1615 res->check_errors++;
1616 goto fail;
1619 /* Do the actual checks */
1620 for(i = 0; i < s->l2_size; i++) {
1621 l2_entry = get_l2_entry(s, l2_table, i);
1623 switch (qcow2_get_cluster_type(bs, l2_entry)) {
1624 case QCOW2_CLUSTER_COMPRESSED:
1625 /* Compressed clusters don't have QCOW_OFLAG_COPIED */
1626 if (l2_entry & QCOW_OFLAG_COPIED) {
1627 fprintf(stderr, "ERROR: coffset=0x%" PRIx64 ": "
1628 "copied flag must never be set for compressed "
1629 "clusters\n", l2_entry & s->cluster_offset_mask);
1630 l2_entry &= ~QCOW_OFLAG_COPIED;
1631 res->corruptions++;
1634 if (has_data_file(bs)) {
1635 fprintf(stderr, "ERROR compressed cluster %d with data file, "
1636 "entry=0x%" PRIx64 "\n", i, l2_entry);
1637 res->corruptions++;
1638 break;
1641 /* Mark cluster as used */
1642 nb_csectors = ((l2_entry >> s->csize_shift) &
1643 s->csize_mask) + 1;
1644 l2_entry &= s->cluster_offset_mask;
1645 ret = qcow2_inc_refcounts_imrt(
1646 bs, res, refcount_table, refcount_table_size,
1647 l2_entry & QCOW2_COMPRESSED_SECTOR_MASK,
1648 nb_csectors * QCOW2_COMPRESSED_SECTOR_SIZE);
1649 if (ret < 0) {
1650 goto fail;
1653 if (flags & CHECK_FRAG_INFO) {
1654 res->bfi.allocated_clusters++;
1655 res->bfi.compressed_clusters++;
1657 /* Compressed clusters are fragmented by nature. Since they
1658 * take up sub-sector space but we only have sector granularity
1659 * I/O we need to re-read the same sectors even for adjacent
1660 * compressed clusters.
1662 res->bfi.fragmented_clusters++;
1664 break;
1666 case QCOW2_CLUSTER_ZERO_ALLOC:
1667 case QCOW2_CLUSTER_NORMAL:
1669 uint64_t offset = l2_entry & L2E_OFFSET_MASK;
1671 /* Correct offsets are cluster aligned */
1672 if (offset_into_cluster(s, offset)) {
1673 bool contains_data;
1674 res->corruptions++;
1676 if (has_subclusters(s)) {
1677 uint64_t l2_bitmap = get_l2_bitmap(s, l2_table, i);
1678 contains_data = (l2_bitmap & QCOW_L2_BITMAP_ALL_ALLOC);
1679 } else {
1680 contains_data = !(l2_entry & QCOW_OFLAG_ZERO);
1683 if (!contains_data) {
1684 fprintf(stderr, "%s offset=%" PRIx64 ": Preallocated "
1685 "cluster is not properly aligned; L2 entry "
1686 "corrupted.\n",
1687 fix & BDRV_FIX_ERRORS ? "Repairing" : "ERROR",
1688 offset);
1689 if (fix & BDRV_FIX_ERRORS) {
1690 int idx = i * (l2_entry_size(s) / sizeof(uint64_t));
1691 uint64_t l2e_offset =
1692 l2_offset + (uint64_t)i * l2_entry_size(s);
1693 int ign = active ? QCOW2_OL_ACTIVE_L2 :
1694 QCOW2_OL_INACTIVE_L2;
1696 l2_entry = has_subclusters(s) ? 0 : QCOW_OFLAG_ZERO;
1697 set_l2_entry(s, l2_table, i, l2_entry);
1698 ret = qcow2_pre_write_overlap_check(bs, ign,
1699 l2e_offset, l2_entry_size(s), false);
1700 if (ret < 0) {
1701 fprintf(stderr, "ERROR: Overlap check failed\n");
1702 res->check_errors++;
1703 /* Something is seriously wrong, so abort checking
1704 * this L2 table */
1705 goto fail;
1708 ret = bdrv_pwrite_sync(bs->file, l2e_offset,
1709 &l2_table[idx],
1710 l2_entry_size(s));
1711 if (ret < 0) {
1712 fprintf(stderr, "ERROR: Failed to overwrite L2 "
1713 "table entry: %s\n", strerror(-ret));
1714 res->check_errors++;
1715 /* Do not abort, continue checking the rest of this
1716 * L2 table's entries */
1717 } else {
1718 res->corruptions--;
1719 res->corruptions_fixed++;
1720 /* Skip marking the cluster as used
1721 * (it is unused now) */
1722 continue;
1725 } else {
1726 fprintf(stderr, "ERROR offset=%" PRIx64 ": Data cluster is "
1727 "not properly aligned; L2 entry corrupted.\n", offset);
1731 if (flags & CHECK_FRAG_INFO) {
1732 res->bfi.allocated_clusters++;
1733 if (next_contiguous_offset &&
1734 offset != next_contiguous_offset) {
1735 res->bfi.fragmented_clusters++;
1737 next_contiguous_offset = offset + s->cluster_size;
1740 /* Mark cluster as used */
1741 if (!has_data_file(bs)) {
1742 ret = qcow2_inc_refcounts_imrt(bs, res, refcount_table,
1743 refcount_table_size,
1744 offset, s->cluster_size);
1745 if (ret < 0) {
1746 goto fail;
1749 break;
1752 case QCOW2_CLUSTER_ZERO_PLAIN:
1753 case QCOW2_CLUSTER_UNALLOCATED:
1754 break;
1756 default:
1757 abort();
1761 g_free(l2_table);
1762 return 0;
1764 fail:
1765 g_free(l2_table);
1766 return ret;
1770 * Increases the refcount for the L1 table, its L2 tables and all referenced
1771 * clusters in the given refcount table. While doing so, performs some checks
1772 * on L1 and L2 entries.
1774 * Returns the number of errors found by the checks or -errno if an internal
1775 * error occurred.
1777 static int check_refcounts_l1(BlockDriverState *bs,
1778 BdrvCheckResult *res,
1779 void **refcount_table,
1780 int64_t *refcount_table_size,
1781 int64_t l1_table_offset, int l1_size,
1782 int flags, BdrvCheckMode fix, bool active)
1784 BDRVQcow2State *s = bs->opaque;
1785 uint64_t *l1_table = NULL, l2_offset, l1_size2;
1786 int i, ret;
1788 l1_size2 = l1_size * L1E_SIZE;
1790 /* Mark L1 table as used */
1791 ret = qcow2_inc_refcounts_imrt(bs, res, refcount_table, refcount_table_size,
1792 l1_table_offset, l1_size2);
1793 if (ret < 0) {
1794 goto fail;
1797 /* Read L1 table entries from disk */
1798 if (l1_size2 > 0) {
1799 l1_table = g_try_malloc(l1_size2);
1800 if (l1_table == NULL) {
1801 ret = -ENOMEM;
1802 res->check_errors++;
1803 goto fail;
1805 ret = bdrv_pread(bs->file, l1_table_offset, l1_table, l1_size2);
1806 if (ret < 0) {
1807 fprintf(stderr, "ERROR: I/O error in check_refcounts_l1\n");
1808 res->check_errors++;
1809 goto fail;
1811 for(i = 0;i < l1_size; i++)
1812 be64_to_cpus(&l1_table[i]);
1815 /* Do the actual checks */
1816 for(i = 0; i < l1_size; i++) {
1817 l2_offset = l1_table[i];
1818 if (l2_offset) {
1819 /* Mark L2 table as used */
1820 l2_offset &= L1E_OFFSET_MASK;
1821 ret = qcow2_inc_refcounts_imrt(bs, res,
1822 refcount_table, refcount_table_size,
1823 l2_offset, s->cluster_size);
1824 if (ret < 0) {
1825 goto fail;
1828 /* L2 tables are cluster aligned */
1829 if (offset_into_cluster(s, l2_offset)) {
1830 fprintf(stderr, "ERROR l2_offset=%" PRIx64 ": Table is not "
1831 "cluster aligned; L1 entry corrupted\n", l2_offset);
1832 res->corruptions++;
1835 /* Process and check L2 entries */
1836 ret = check_refcounts_l2(bs, res, refcount_table,
1837 refcount_table_size, l2_offset, flags,
1838 fix, active);
1839 if (ret < 0) {
1840 goto fail;
1844 g_free(l1_table);
1845 return 0;
1847 fail:
1848 g_free(l1_table);
1849 return ret;
1853 * Checks the OFLAG_COPIED flag for all L1 and L2 entries.
1855 * This function does not print an error message nor does it increment
1856 * check_errors if qcow2_get_refcount fails (this is because such an error will
1857 * have been already detected and sufficiently signaled by the calling function
1858 * (qcow2_check_refcounts) by the time this function is called).
1860 static int check_oflag_copied(BlockDriverState *bs, BdrvCheckResult *res,
1861 BdrvCheckMode fix)
1863 BDRVQcow2State *s = bs->opaque;
1864 uint64_t *l2_table = qemu_blockalign(bs, s->cluster_size);
1865 int ret;
1866 uint64_t refcount;
1867 int i, j;
1868 bool repair;
1870 if (fix & BDRV_FIX_ERRORS) {
1871 /* Always repair */
1872 repair = true;
1873 } else if (fix & BDRV_FIX_LEAKS) {
1874 /* Repair only if that seems safe: This function is always
1875 * called after the refcounts have been fixed, so the refcount
1876 * is accurate if that repair was successful */
1877 repair = !res->check_errors && !res->corruptions && !res->leaks;
1878 } else {
1879 repair = false;
1882 for (i = 0; i < s->l1_size; i++) {
1883 uint64_t l1_entry = s->l1_table[i];
1884 uint64_t l2_offset = l1_entry & L1E_OFFSET_MASK;
1885 int l2_dirty = 0;
1887 if (!l2_offset) {
1888 continue;
1891 ret = qcow2_get_refcount(bs, l2_offset >> s->cluster_bits,
1892 &refcount);
1893 if (ret < 0) {
1894 /* don't print message nor increment check_errors */
1895 continue;
1897 if ((refcount == 1) != ((l1_entry & QCOW_OFLAG_COPIED) != 0)) {
1898 res->corruptions++;
1899 fprintf(stderr, "%s OFLAG_COPIED L2 cluster: l1_index=%d "
1900 "l1_entry=%" PRIx64 " refcount=%" PRIu64 "\n",
1901 repair ? "Repairing" : "ERROR", i, l1_entry, refcount);
1902 if (repair) {
1903 s->l1_table[i] = refcount == 1
1904 ? l1_entry | QCOW_OFLAG_COPIED
1905 : l1_entry & ~QCOW_OFLAG_COPIED;
1906 ret = qcow2_write_l1_entry(bs, i);
1907 if (ret < 0) {
1908 res->check_errors++;
1909 goto fail;
1911 res->corruptions--;
1912 res->corruptions_fixed++;
1916 ret = bdrv_pread(bs->file, l2_offset, l2_table,
1917 s->l2_size * l2_entry_size(s));
1918 if (ret < 0) {
1919 fprintf(stderr, "ERROR: Could not read L2 table: %s\n",
1920 strerror(-ret));
1921 res->check_errors++;
1922 goto fail;
1925 for (j = 0; j < s->l2_size; j++) {
1926 uint64_t l2_entry = get_l2_entry(s, l2_table, j);
1927 uint64_t data_offset = l2_entry & L2E_OFFSET_MASK;
1928 QCow2ClusterType cluster_type = qcow2_get_cluster_type(bs, l2_entry);
1930 if (cluster_type == QCOW2_CLUSTER_NORMAL ||
1931 cluster_type == QCOW2_CLUSTER_ZERO_ALLOC) {
1932 if (has_data_file(bs)) {
1933 refcount = 1;
1934 } else {
1935 ret = qcow2_get_refcount(bs,
1936 data_offset >> s->cluster_bits,
1937 &refcount);
1938 if (ret < 0) {
1939 /* don't print message nor increment check_errors */
1940 continue;
1943 if ((refcount == 1) != ((l2_entry & QCOW_OFLAG_COPIED) != 0)) {
1944 res->corruptions++;
1945 fprintf(stderr, "%s OFLAG_COPIED data cluster: "
1946 "l2_entry=%" PRIx64 " refcount=%" PRIu64 "\n",
1947 repair ? "Repairing" : "ERROR", l2_entry, refcount);
1948 if (repair) {
1949 set_l2_entry(s, l2_table, j,
1950 refcount == 1 ?
1951 l2_entry | QCOW_OFLAG_COPIED :
1952 l2_entry & ~QCOW_OFLAG_COPIED);
1953 l2_dirty++;
1959 if (l2_dirty > 0) {
1960 ret = qcow2_pre_write_overlap_check(bs, QCOW2_OL_ACTIVE_L2,
1961 l2_offset, s->cluster_size,
1962 false);
1963 if (ret < 0) {
1964 fprintf(stderr, "ERROR: Could not write L2 table; metadata "
1965 "overlap check failed: %s\n", strerror(-ret));
1966 res->check_errors++;
1967 goto fail;
1970 ret = bdrv_pwrite(bs->file, l2_offset, l2_table,
1971 s->cluster_size);
1972 if (ret < 0) {
1973 fprintf(stderr, "ERROR: Could not write L2 table: %s\n",
1974 strerror(-ret));
1975 res->check_errors++;
1976 goto fail;
1978 res->corruptions -= l2_dirty;
1979 res->corruptions_fixed += l2_dirty;
1983 ret = 0;
1985 fail:
1986 qemu_vfree(l2_table);
1987 return ret;
1991 * Checks consistency of refblocks and accounts for each refblock in
1992 * *refcount_table.
1994 static int check_refblocks(BlockDriverState *bs, BdrvCheckResult *res,
1995 BdrvCheckMode fix, bool *rebuild,
1996 void **refcount_table, int64_t *nb_clusters)
1998 BDRVQcow2State *s = bs->opaque;
1999 int64_t i, size;
2000 int ret;
2002 for(i = 0; i < s->refcount_table_size; i++) {
2003 uint64_t offset, cluster;
2004 offset = s->refcount_table[i];
2005 cluster = offset >> s->cluster_bits;
2007 /* Refcount blocks are cluster aligned */
2008 if (offset_into_cluster(s, offset)) {
2009 fprintf(stderr, "ERROR refcount block %" PRId64 " is not "
2010 "cluster aligned; refcount table entry corrupted\n", i);
2011 res->corruptions++;
2012 *rebuild = true;
2013 continue;
2016 if (cluster >= *nb_clusters) {
2017 res->corruptions++;
2018 fprintf(stderr, "%s refcount block %" PRId64 " is outside image\n",
2019 fix & BDRV_FIX_ERRORS ? "Repairing" : "ERROR", i);
2021 if (fix & BDRV_FIX_ERRORS) {
2022 int64_t new_nb_clusters;
2023 Error *local_err = NULL;
2025 if (offset > INT64_MAX - s->cluster_size) {
2026 ret = -EINVAL;
2027 goto resize_fail;
2030 ret = bdrv_truncate(bs->file, offset + s->cluster_size, false,
2031 PREALLOC_MODE_OFF, 0, &local_err);
2032 if (ret < 0) {
2033 error_report_err(local_err);
2034 goto resize_fail;
2036 size = bdrv_getlength(bs->file->bs);
2037 if (size < 0) {
2038 ret = size;
2039 goto resize_fail;
2042 new_nb_clusters = size_to_clusters(s, size);
2043 assert(new_nb_clusters >= *nb_clusters);
2045 ret = realloc_refcount_array(s, refcount_table,
2046 nb_clusters, new_nb_clusters);
2047 if (ret < 0) {
2048 res->check_errors++;
2049 return ret;
2052 if (cluster >= *nb_clusters) {
2053 ret = -EINVAL;
2054 goto resize_fail;
2057 res->corruptions--;
2058 res->corruptions_fixed++;
2059 ret = qcow2_inc_refcounts_imrt(bs, res,
2060 refcount_table, nb_clusters,
2061 offset, s->cluster_size);
2062 if (ret < 0) {
2063 return ret;
2065 /* No need to check whether the refcount is now greater than 1:
2066 * This area was just allocated and zeroed, so it can only be
2067 * exactly 1 after qcow2_inc_refcounts_imrt() */
2068 continue;
2070 resize_fail:
2071 *rebuild = true;
2072 fprintf(stderr, "ERROR could not resize image: %s\n",
2073 strerror(-ret));
2075 continue;
2078 if (offset != 0) {
2079 ret = qcow2_inc_refcounts_imrt(bs, res, refcount_table, nb_clusters,
2080 offset, s->cluster_size);
2081 if (ret < 0) {
2082 return ret;
2084 if (s->get_refcount(*refcount_table, cluster) != 1) {
2085 fprintf(stderr, "ERROR refcount block %" PRId64
2086 " refcount=%" PRIu64 "\n", i,
2087 s->get_refcount(*refcount_table, cluster));
2088 res->corruptions++;
2089 *rebuild = true;
2094 return 0;
2098 * Calculates an in-memory refcount table.
2100 static int calculate_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
2101 BdrvCheckMode fix, bool *rebuild,
2102 void **refcount_table, int64_t *nb_clusters)
2104 BDRVQcow2State *s = bs->opaque;
2105 int64_t i;
2106 QCowSnapshot *sn;
2107 int ret;
2109 if (!*refcount_table) {
2110 int64_t old_size = 0;
2111 ret = realloc_refcount_array(s, refcount_table,
2112 &old_size, *nb_clusters);
2113 if (ret < 0) {
2114 res->check_errors++;
2115 return ret;
2119 /* header */
2120 ret = qcow2_inc_refcounts_imrt(bs, res, refcount_table, nb_clusters,
2121 0, s->cluster_size);
2122 if (ret < 0) {
2123 return ret;
2126 /* current L1 table */
2127 ret = check_refcounts_l1(bs, res, refcount_table, nb_clusters,
2128 s->l1_table_offset, s->l1_size, CHECK_FRAG_INFO,
2129 fix, true);
2130 if (ret < 0) {
2131 return ret;
2134 /* snapshots */
2135 if (has_data_file(bs) && s->nb_snapshots) {
2136 fprintf(stderr, "ERROR %d snapshots in image with data file\n",
2137 s->nb_snapshots);
2138 res->corruptions++;
2141 for (i = 0; i < s->nb_snapshots; i++) {
2142 sn = s->snapshots + i;
2143 if (offset_into_cluster(s, sn->l1_table_offset)) {
2144 fprintf(stderr, "ERROR snapshot %s (%s) l1_offset=%#" PRIx64 ": "
2145 "L1 table is not cluster aligned; snapshot table entry "
2146 "corrupted\n", sn->id_str, sn->name, sn->l1_table_offset);
2147 res->corruptions++;
2148 continue;
2150 if (sn->l1_size > QCOW_MAX_L1_SIZE / L1E_SIZE) {
2151 fprintf(stderr, "ERROR snapshot %s (%s) l1_size=%#" PRIx32 ": "
2152 "L1 table is too large; snapshot table entry corrupted\n",
2153 sn->id_str, sn->name, sn->l1_size);
2154 res->corruptions++;
2155 continue;
2157 ret = check_refcounts_l1(bs, res, refcount_table, nb_clusters,
2158 sn->l1_table_offset, sn->l1_size, 0, fix,
2159 false);
2160 if (ret < 0) {
2161 return ret;
2164 ret = qcow2_inc_refcounts_imrt(bs, res, refcount_table, nb_clusters,
2165 s->snapshots_offset, s->snapshots_size);
2166 if (ret < 0) {
2167 return ret;
2170 /* refcount data */
2171 ret = qcow2_inc_refcounts_imrt(bs, res, refcount_table, nb_clusters,
2172 s->refcount_table_offset,
2173 s->refcount_table_size *
2174 REFTABLE_ENTRY_SIZE);
2175 if (ret < 0) {
2176 return ret;
2179 /* encryption */
2180 if (s->crypto_header.length) {
2181 ret = qcow2_inc_refcounts_imrt(bs, res, refcount_table, nb_clusters,
2182 s->crypto_header.offset,
2183 s->crypto_header.length);
2184 if (ret < 0) {
2185 return ret;
2189 /* bitmaps */
2190 ret = qcow2_check_bitmaps_refcounts(bs, res, refcount_table, nb_clusters);
2191 if (ret < 0) {
2192 return ret;
2195 return check_refblocks(bs, res, fix, rebuild, refcount_table, nb_clusters);
2199 * Compares the actual reference count for each cluster in the image against the
2200 * refcount as reported by the refcount structures on-disk.
2202 static void compare_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
2203 BdrvCheckMode fix, bool *rebuild,
2204 int64_t *highest_cluster,
2205 void *refcount_table, int64_t nb_clusters)
2207 BDRVQcow2State *s = bs->opaque;
2208 int64_t i;
2209 uint64_t refcount1, refcount2;
2210 int ret;
2212 for (i = 0, *highest_cluster = 0; i < nb_clusters; i++) {
2213 ret = qcow2_get_refcount(bs, i, &refcount1);
2214 if (ret < 0) {
2215 fprintf(stderr, "Can't get refcount for cluster %" PRId64 ": %s\n",
2216 i, strerror(-ret));
2217 res->check_errors++;
2218 continue;
2221 refcount2 = s->get_refcount(refcount_table, i);
2223 if (refcount1 > 0 || refcount2 > 0) {
2224 *highest_cluster = i;
2227 if (refcount1 != refcount2) {
2228 /* Check if we're allowed to fix the mismatch */
2229 int *num_fixed = NULL;
2230 if (refcount1 == 0) {
2231 *rebuild = true;
2232 } else if (refcount1 > refcount2 && (fix & BDRV_FIX_LEAKS)) {
2233 num_fixed = &res->leaks_fixed;
2234 } else if (refcount1 < refcount2 && (fix & BDRV_FIX_ERRORS)) {
2235 num_fixed = &res->corruptions_fixed;
2238 fprintf(stderr, "%s cluster %" PRId64 " refcount=%" PRIu64
2239 " reference=%" PRIu64 "\n",
2240 num_fixed != NULL ? "Repairing" :
2241 refcount1 < refcount2 ? "ERROR" :
2242 "Leaked",
2243 i, refcount1, refcount2);
2245 if (num_fixed) {
2246 ret = update_refcount(bs, i << s->cluster_bits, 1,
2247 refcount_diff(refcount1, refcount2),
2248 refcount1 > refcount2,
2249 QCOW2_DISCARD_ALWAYS);
2250 if (ret >= 0) {
2251 (*num_fixed)++;
2252 continue;
2256 /* And if we couldn't, print an error */
2257 if (refcount1 < refcount2) {
2258 res->corruptions++;
2259 } else {
2260 res->leaks++;
2267 * Allocates clusters using an in-memory refcount table (IMRT) in contrast to
2268 * the on-disk refcount structures.
2270 * On input, *first_free_cluster tells where to start looking, and need not
2271 * actually be a free cluster; the returned offset will not be before that
2272 * cluster. On output, *first_free_cluster points to the first gap found, even
2273 * if that gap was too small to be used as the returned offset.
2275 * Note that *first_free_cluster is a cluster index whereas the return value is
2276 * an offset.
2278 static int64_t alloc_clusters_imrt(BlockDriverState *bs,
2279 int cluster_count,
2280 void **refcount_table,
2281 int64_t *imrt_nb_clusters,
2282 int64_t *first_free_cluster)
2284 BDRVQcow2State *s = bs->opaque;
2285 int64_t cluster = *first_free_cluster, i;
2286 bool first_gap = true;
2287 int contiguous_free_clusters;
2288 int ret;
2290 /* Starting at *first_free_cluster, find a range of at least cluster_count
2291 * continuously free clusters */
2292 for (contiguous_free_clusters = 0;
2293 cluster < *imrt_nb_clusters &&
2294 contiguous_free_clusters < cluster_count;
2295 cluster++)
2297 if (!s->get_refcount(*refcount_table, cluster)) {
2298 contiguous_free_clusters++;
2299 if (first_gap) {
2300 /* If this is the first free cluster found, update
2301 * *first_free_cluster accordingly */
2302 *first_free_cluster = cluster;
2303 first_gap = false;
2305 } else if (contiguous_free_clusters) {
2306 contiguous_free_clusters = 0;
2310 /* If contiguous_free_clusters is greater than zero, it contains the number
2311 * of continuously free clusters until the current cluster; the first free
2312 * cluster in the current "gap" is therefore
2313 * cluster - contiguous_free_clusters */
2315 /* If no such range could be found, grow the in-memory refcount table
2316 * accordingly to append free clusters at the end of the image */
2317 if (contiguous_free_clusters < cluster_count) {
2318 /* contiguous_free_clusters clusters are already empty at the image end;
2319 * we need cluster_count clusters; therefore, we have to allocate
2320 * cluster_count - contiguous_free_clusters new clusters at the end of
2321 * the image (which is the current value of cluster; note that cluster
2322 * may exceed old_imrt_nb_clusters if *first_free_cluster pointed beyond
2323 * the image end) */
2324 ret = realloc_refcount_array(s, refcount_table, imrt_nb_clusters,
2325 cluster + cluster_count
2326 - contiguous_free_clusters);
2327 if (ret < 0) {
2328 return ret;
2332 /* Go back to the first free cluster */
2333 cluster -= contiguous_free_clusters;
2334 for (i = 0; i < cluster_count; i++) {
2335 s->set_refcount(*refcount_table, cluster + i, 1);
2338 return cluster << s->cluster_bits;
2342 * Creates a new refcount structure based solely on the in-memory information
2343 * given through *refcount_table. All necessary allocations will be reflected
2344 * in that array.
2346 * On success, the old refcount structure is leaked (it will be covered by the
2347 * new refcount structure).
2349 static int rebuild_refcount_structure(BlockDriverState *bs,
2350 BdrvCheckResult *res,
2351 void **refcount_table,
2352 int64_t *nb_clusters)
2354 BDRVQcow2State *s = bs->opaque;
2355 int64_t first_free_cluster = 0, reftable_offset = -1, cluster = 0;
2356 int64_t refblock_offset, refblock_start, refblock_index;
2357 uint32_t reftable_size = 0;
2358 uint64_t *on_disk_reftable = NULL;
2359 void *on_disk_refblock;
2360 int ret = 0;
2361 struct {
2362 uint64_t reftable_offset;
2363 uint32_t reftable_clusters;
2364 } QEMU_PACKED reftable_offset_and_clusters;
2366 qcow2_cache_empty(bs, s->refcount_block_cache);
2368 write_refblocks:
2369 for (; cluster < *nb_clusters; cluster++) {
2370 if (!s->get_refcount(*refcount_table, cluster)) {
2371 continue;
2374 refblock_index = cluster >> s->refcount_block_bits;
2375 refblock_start = refblock_index << s->refcount_block_bits;
2377 /* Don't allocate a cluster in a refblock already written to disk */
2378 if (first_free_cluster < refblock_start) {
2379 first_free_cluster = refblock_start;
2381 refblock_offset = alloc_clusters_imrt(bs, 1, refcount_table,
2382 nb_clusters, &first_free_cluster);
2383 if (refblock_offset < 0) {
2384 fprintf(stderr, "ERROR allocating refblock: %s\n",
2385 strerror(-refblock_offset));
2386 res->check_errors++;
2387 ret = refblock_offset;
2388 goto fail;
2391 if (reftable_size <= refblock_index) {
2392 uint32_t old_reftable_size = reftable_size;
2393 uint64_t *new_on_disk_reftable;
2395 reftable_size = ROUND_UP((refblock_index + 1) * REFTABLE_ENTRY_SIZE,
2396 s->cluster_size) / REFTABLE_ENTRY_SIZE;
2397 new_on_disk_reftable = g_try_realloc(on_disk_reftable,
2398 reftable_size *
2399 REFTABLE_ENTRY_SIZE);
2400 if (!new_on_disk_reftable) {
2401 res->check_errors++;
2402 ret = -ENOMEM;
2403 goto fail;
2405 on_disk_reftable = new_on_disk_reftable;
2407 memset(on_disk_reftable + old_reftable_size, 0,
2408 (reftable_size - old_reftable_size) * REFTABLE_ENTRY_SIZE);
2410 /* The offset we have for the reftable is now no longer valid;
2411 * this will leak that range, but we can easily fix that by running
2412 * a leak-fixing check after this rebuild operation */
2413 reftable_offset = -1;
2414 } else {
2415 assert(on_disk_reftable);
2417 on_disk_reftable[refblock_index] = refblock_offset;
2419 /* If this is apparently the last refblock (for now), try to squeeze the
2420 * reftable in */
2421 if (refblock_index == (*nb_clusters - 1) >> s->refcount_block_bits &&
2422 reftable_offset < 0)
2424 uint64_t reftable_clusters = size_to_clusters(s, reftable_size *
2425 REFTABLE_ENTRY_SIZE);
2426 reftable_offset = alloc_clusters_imrt(bs, reftable_clusters,
2427 refcount_table, nb_clusters,
2428 &first_free_cluster);
2429 if (reftable_offset < 0) {
2430 fprintf(stderr, "ERROR allocating reftable: %s\n",
2431 strerror(-reftable_offset));
2432 res->check_errors++;
2433 ret = reftable_offset;
2434 goto fail;
2438 ret = qcow2_pre_write_overlap_check(bs, 0, refblock_offset,
2439 s->cluster_size, false);
2440 if (ret < 0) {
2441 fprintf(stderr, "ERROR writing refblock: %s\n", strerror(-ret));
2442 goto fail;
2445 /* The size of *refcount_table is always cluster-aligned, therefore the
2446 * write operation will not overflow */
2447 on_disk_refblock = (void *)((char *) *refcount_table +
2448 refblock_index * s->cluster_size);
2450 ret = bdrv_pwrite(bs->file, refblock_offset, on_disk_refblock,
2451 s->cluster_size);
2452 if (ret < 0) {
2453 fprintf(stderr, "ERROR writing refblock: %s\n", strerror(-ret));
2454 goto fail;
2457 /* Go to the end of this refblock */
2458 cluster = refblock_start + s->refcount_block_size - 1;
2461 if (reftable_offset < 0) {
2462 uint64_t post_refblock_start, reftable_clusters;
2464 post_refblock_start = ROUND_UP(*nb_clusters, s->refcount_block_size);
2465 reftable_clusters =
2466 size_to_clusters(s, reftable_size * REFTABLE_ENTRY_SIZE);
2467 /* Not pretty but simple */
2468 if (first_free_cluster < post_refblock_start) {
2469 first_free_cluster = post_refblock_start;
2471 reftable_offset = alloc_clusters_imrt(bs, reftable_clusters,
2472 refcount_table, nb_clusters,
2473 &first_free_cluster);
2474 if (reftable_offset < 0) {
2475 fprintf(stderr, "ERROR allocating reftable: %s\n",
2476 strerror(-reftable_offset));
2477 res->check_errors++;
2478 ret = reftable_offset;
2479 goto fail;
2482 goto write_refblocks;
2485 for (refblock_index = 0; refblock_index < reftable_size; refblock_index++) {
2486 cpu_to_be64s(&on_disk_reftable[refblock_index]);
2489 ret = qcow2_pre_write_overlap_check(bs, 0, reftable_offset,
2490 reftable_size * REFTABLE_ENTRY_SIZE,
2491 false);
2492 if (ret < 0) {
2493 fprintf(stderr, "ERROR writing reftable: %s\n", strerror(-ret));
2494 goto fail;
2497 assert(reftable_size < INT_MAX / REFTABLE_ENTRY_SIZE);
2498 ret = bdrv_pwrite(bs->file, reftable_offset, on_disk_reftable,
2499 reftable_size * REFTABLE_ENTRY_SIZE);
2500 if (ret < 0) {
2501 fprintf(stderr, "ERROR writing reftable: %s\n", strerror(-ret));
2502 goto fail;
2505 /* Enter new reftable into the image header */
2506 reftable_offset_and_clusters.reftable_offset = cpu_to_be64(reftable_offset);
2507 reftable_offset_and_clusters.reftable_clusters =
2508 cpu_to_be32(size_to_clusters(s, reftable_size * REFTABLE_ENTRY_SIZE));
2509 ret = bdrv_pwrite_sync(bs->file,
2510 offsetof(QCowHeader, refcount_table_offset),
2511 &reftable_offset_and_clusters,
2512 sizeof(reftable_offset_and_clusters));
2513 if (ret < 0) {
2514 fprintf(stderr, "ERROR setting reftable: %s\n", strerror(-ret));
2515 goto fail;
2518 for (refblock_index = 0; refblock_index < reftable_size; refblock_index++) {
2519 be64_to_cpus(&on_disk_reftable[refblock_index]);
2521 s->refcount_table = on_disk_reftable;
2522 s->refcount_table_offset = reftable_offset;
2523 s->refcount_table_size = reftable_size;
2524 update_max_refcount_table_index(s);
2526 return 0;
2528 fail:
2529 g_free(on_disk_reftable);
2530 return ret;
2534 * Checks an image for refcount consistency.
2536 * Returns 0 if no errors are found, the number of errors in case the image is
2537 * detected as corrupted, and -errno when an internal error occurred.
2539 int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
2540 BdrvCheckMode fix)
2542 BDRVQcow2State *s = bs->opaque;
2543 BdrvCheckResult pre_compare_res;
2544 int64_t size, highest_cluster, nb_clusters;
2545 void *refcount_table = NULL;
2546 bool rebuild = false;
2547 int ret;
2549 size = bdrv_getlength(bs->file->bs);
2550 if (size < 0) {
2551 res->check_errors++;
2552 return size;
2555 nb_clusters = size_to_clusters(s, size);
2556 if (nb_clusters > INT_MAX) {
2557 res->check_errors++;
2558 return -EFBIG;
2561 res->bfi.total_clusters =
2562 size_to_clusters(s, bs->total_sectors * BDRV_SECTOR_SIZE);
2564 ret = calculate_refcounts(bs, res, fix, &rebuild, &refcount_table,
2565 &nb_clusters);
2566 if (ret < 0) {
2567 goto fail;
2570 /* In case we don't need to rebuild the refcount structure (but want to fix
2571 * something), this function is immediately called again, in which case the
2572 * result should be ignored */
2573 pre_compare_res = *res;
2574 compare_refcounts(bs, res, 0, &rebuild, &highest_cluster, refcount_table,
2575 nb_clusters);
2577 if (rebuild && (fix & BDRV_FIX_ERRORS)) {
2578 BdrvCheckResult old_res = *res;
2579 int fresh_leaks = 0;
2581 fprintf(stderr, "Rebuilding refcount structure\n");
2582 ret = rebuild_refcount_structure(bs, res, &refcount_table,
2583 &nb_clusters);
2584 if (ret < 0) {
2585 goto fail;
2588 res->corruptions = 0;
2589 res->leaks = 0;
2591 /* Because the old reftable has been exchanged for a new one the
2592 * references have to be recalculated */
2593 rebuild = false;
2594 memset(refcount_table, 0, refcount_array_byte_size(s, nb_clusters));
2595 ret = calculate_refcounts(bs, res, 0, &rebuild, &refcount_table,
2596 &nb_clusters);
2597 if (ret < 0) {
2598 goto fail;
2601 if (fix & BDRV_FIX_LEAKS) {
2602 /* The old refcount structures are now leaked, fix it; the result
2603 * can be ignored, aside from leaks which were introduced by
2604 * rebuild_refcount_structure() that could not be fixed */
2605 BdrvCheckResult saved_res = *res;
2606 *res = (BdrvCheckResult){ 0 };
2608 compare_refcounts(bs, res, BDRV_FIX_LEAKS, &rebuild,
2609 &highest_cluster, refcount_table, nb_clusters);
2610 if (rebuild) {
2611 fprintf(stderr, "ERROR rebuilt refcount structure is still "
2612 "broken\n");
2615 /* Any leaks accounted for here were introduced by
2616 * rebuild_refcount_structure() because that function has created a
2617 * new refcount structure from scratch */
2618 fresh_leaks = res->leaks;
2619 *res = saved_res;
2622 if (res->corruptions < old_res.corruptions) {
2623 res->corruptions_fixed += old_res.corruptions - res->corruptions;
2625 if (res->leaks < old_res.leaks) {
2626 res->leaks_fixed += old_res.leaks - res->leaks;
2628 res->leaks += fresh_leaks;
2629 } else if (fix) {
2630 if (rebuild) {
2631 fprintf(stderr, "ERROR need to rebuild refcount structures\n");
2632 res->check_errors++;
2633 ret = -EIO;
2634 goto fail;
2637 if (res->leaks || res->corruptions) {
2638 *res = pre_compare_res;
2639 compare_refcounts(bs, res, fix, &rebuild, &highest_cluster,
2640 refcount_table, nb_clusters);
2644 /* check OFLAG_COPIED */
2645 ret = check_oflag_copied(bs, res, fix);
2646 if (ret < 0) {
2647 goto fail;
2650 res->image_end_offset = (highest_cluster + 1) * s->cluster_size;
2651 ret = 0;
2653 fail:
2654 g_free(refcount_table);
2656 return ret;
2659 #define overlaps_with(ofs, sz) \
2660 ranges_overlap(offset, size, ofs, sz)
2663 * Checks if the given offset into the image file is actually free to use by
2664 * looking for overlaps with important metadata sections (L1/L2 tables etc.),
2665 * i.e. a sanity check without relying on the refcount tables.
2667 * The ign parameter specifies what checks not to perform (being a bitmask of
2668 * QCow2MetadataOverlap values), i.e., what sections to ignore.
2670 * Returns:
2671 * - 0 if writing to this offset will not affect the mentioned metadata
2672 * - a positive QCow2MetadataOverlap value indicating one overlapping section
2673 * - a negative value (-errno) indicating an error while performing a check,
2674 * e.g. when bdrv_pread failed on QCOW2_OL_INACTIVE_L2
2676 int qcow2_check_metadata_overlap(BlockDriverState *bs, int ign, int64_t offset,
2677 int64_t size)
2679 BDRVQcow2State *s = bs->opaque;
2680 int chk = s->overlap_check & ~ign;
2681 int i, j;
2683 if (!size) {
2684 return 0;
2687 if (chk & QCOW2_OL_MAIN_HEADER) {
2688 if (offset < s->cluster_size) {
2689 return QCOW2_OL_MAIN_HEADER;
2693 /* align range to test to cluster boundaries */
2694 size = ROUND_UP(offset_into_cluster(s, offset) + size, s->cluster_size);
2695 offset = start_of_cluster(s, offset);
2697 if ((chk & QCOW2_OL_ACTIVE_L1) && s->l1_size) {
2698 if (overlaps_with(s->l1_table_offset, s->l1_size * L1E_SIZE)) {
2699 return QCOW2_OL_ACTIVE_L1;
2703 if ((chk & QCOW2_OL_REFCOUNT_TABLE) && s->refcount_table_size) {
2704 if (overlaps_with(s->refcount_table_offset,
2705 s->refcount_table_size * REFTABLE_ENTRY_SIZE)) {
2706 return QCOW2_OL_REFCOUNT_TABLE;
2710 if ((chk & QCOW2_OL_SNAPSHOT_TABLE) && s->snapshots_size) {
2711 if (overlaps_with(s->snapshots_offset, s->snapshots_size)) {
2712 return QCOW2_OL_SNAPSHOT_TABLE;
2716 if ((chk & QCOW2_OL_INACTIVE_L1) && s->snapshots) {
2717 for (i = 0; i < s->nb_snapshots; i++) {
2718 if (s->snapshots[i].l1_size &&
2719 overlaps_with(s->snapshots[i].l1_table_offset,
2720 s->snapshots[i].l1_size * L1E_SIZE)) {
2721 return QCOW2_OL_INACTIVE_L1;
2726 if ((chk & QCOW2_OL_ACTIVE_L2) && s->l1_table) {
2727 for (i = 0; i < s->l1_size; i++) {
2728 if ((s->l1_table[i] & L1E_OFFSET_MASK) &&
2729 overlaps_with(s->l1_table[i] & L1E_OFFSET_MASK,
2730 s->cluster_size)) {
2731 return QCOW2_OL_ACTIVE_L2;
2736 if ((chk & QCOW2_OL_REFCOUNT_BLOCK) && s->refcount_table) {
2737 unsigned last_entry = s->max_refcount_table_index;
2738 assert(last_entry < s->refcount_table_size);
2739 assert(last_entry + 1 == s->refcount_table_size ||
2740 (s->refcount_table[last_entry + 1] & REFT_OFFSET_MASK) == 0);
2741 for (i = 0; i <= last_entry; i++) {
2742 if ((s->refcount_table[i] & REFT_OFFSET_MASK) &&
2743 overlaps_with(s->refcount_table[i] & REFT_OFFSET_MASK,
2744 s->cluster_size)) {
2745 return QCOW2_OL_REFCOUNT_BLOCK;
2750 if ((chk & QCOW2_OL_INACTIVE_L2) && s->snapshots) {
2751 for (i = 0; i < s->nb_snapshots; i++) {
2752 uint64_t l1_ofs = s->snapshots[i].l1_table_offset;
2753 uint32_t l1_sz = s->snapshots[i].l1_size;
2754 uint64_t l1_sz2 = l1_sz * L1E_SIZE;
2755 uint64_t *l1;
2756 int ret;
2758 ret = qcow2_validate_table(bs, l1_ofs, l1_sz, L1E_SIZE,
2759 QCOW_MAX_L1_SIZE, "", NULL);
2760 if (ret < 0) {
2761 return ret;
2764 l1 = g_try_malloc(l1_sz2);
2766 if (l1_sz2 && l1 == NULL) {
2767 return -ENOMEM;
2770 ret = bdrv_pread(bs->file, l1_ofs, l1, l1_sz2);
2771 if (ret < 0) {
2772 g_free(l1);
2773 return ret;
2776 for (j = 0; j < l1_sz; j++) {
2777 uint64_t l2_ofs = be64_to_cpu(l1[j]) & L1E_OFFSET_MASK;
2778 if (l2_ofs && overlaps_with(l2_ofs, s->cluster_size)) {
2779 g_free(l1);
2780 return QCOW2_OL_INACTIVE_L2;
2784 g_free(l1);
2788 if ((chk & QCOW2_OL_BITMAP_DIRECTORY) &&
2789 (s->autoclear_features & QCOW2_AUTOCLEAR_BITMAPS))
2791 if (overlaps_with(s->bitmap_directory_offset,
2792 s->bitmap_directory_size))
2794 return QCOW2_OL_BITMAP_DIRECTORY;
2798 return 0;
2801 static const char *metadata_ol_names[] = {
2802 [QCOW2_OL_MAIN_HEADER_BITNR] = "qcow2_header",
2803 [QCOW2_OL_ACTIVE_L1_BITNR] = "active L1 table",
2804 [QCOW2_OL_ACTIVE_L2_BITNR] = "active L2 table",
2805 [QCOW2_OL_REFCOUNT_TABLE_BITNR] = "refcount table",
2806 [QCOW2_OL_REFCOUNT_BLOCK_BITNR] = "refcount block",
2807 [QCOW2_OL_SNAPSHOT_TABLE_BITNR] = "snapshot table",
2808 [QCOW2_OL_INACTIVE_L1_BITNR] = "inactive L1 table",
2809 [QCOW2_OL_INACTIVE_L2_BITNR] = "inactive L2 table",
2810 [QCOW2_OL_BITMAP_DIRECTORY_BITNR] = "bitmap directory",
2812 QEMU_BUILD_BUG_ON(QCOW2_OL_MAX_BITNR != ARRAY_SIZE(metadata_ol_names));
2815 * First performs a check for metadata overlaps (through
2816 * qcow2_check_metadata_overlap); if that fails with a negative value (error
2817 * while performing a check), that value is returned. If an impending overlap
2818 * is detected, the BDS will be made unusable, the qcow2 file marked corrupt
2819 * and -EIO returned.
2821 * Returns 0 if there were neither overlaps nor errors while checking for
2822 * overlaps; or a negative value (-errno) on error.
2824 int qcow2_pre_write_overlap_check(BlockDriverState *bs, int ign, int64_t offset,
2825 int64_t size, bool data_file)
2827 int ret;
2829 if (data_file && has_data_file(bs)) {
2830 return 0;
2833 ret = qcow2_check_metadata_overlap(bs, ign, offset, size);
2834 if (ret < 0) {
2835 return ret;
2836 } else if (ret > 0) {
2837 int metadata_ol_bitnr = ctz32(ret);
2838 assert(metadata_ol_bitnr < QCOW2_OL_MAX_BITNR);
2840 qcow2_signal_corruption(bs, true, offset, size, "Preventing invalid "
2841 "write on metadata (overlaps with %s)",
2842 metadata_ol_names[metadata_ol_bitnr]);
2843 return -EIO;
2846 return 0;
2849 /* A pointer to a function of this type is given to walk_over_reftable(). That
2850 * function will create refblocks and pass them to a RefblockFinishOp once they
2851 * are completed (@refblock). @refblock_empty is set if the refblock is
2852 * completely empty.
2854 * Along with the refblock, a corresponding reftable entry is passed, in the
2855 * reftable @reftable (which may be reallocated) at @reftable_index.
2857 * @allocated should be set to true if a new cluster has been allocated.
2859 typedef int (RefblockFinishOp)(BlockDriverState *bs, uint64_t **reftable,
2860 uint64_t reftable_index, uint64_t *reftable_size,
2861 void *refblock, bool refblock_empty,
2862 bool *allocated, Error **errp);
2865 * This "operation" for walk_over_reftable() allocates the refblock on disk (if
2866 * it is not empty) and inserts its offset into the new reftable. The size of
2867 * this new reftable is increased as required.
2869 static int alloc_refblock(BlockDriverState *bs, uint64_t **reftable,
2870 uint64_t reftable_index, uint64_t *reftable_size,
2871 void *refblock, bool refblock_empty, bool *allocated,
2872 Error **errp)
2874 BDRVQcow2State *s = bs->opaque;
2875 int64_t offset;
2877 if (!refblock_empty && reftable_index >= *reftable_size) {
2878 uint64_t *new_reftable;
2879 uint64_t new_reftable_size;
2881 new_reftable_size = ROUND_UP(reftable_index + 1,
2882 s->cluster_size / REFTABLE_ENTRY_SIZE);
2883 if (new_reftable_size > QCOW_MAX_REFTABLE_SIZE / REFTABLE_ENTRY_SIZE) {
2884 error_setg(errp,
2885 "This operation would make the refcount table grow "
2886 "beyond the maximum size supported by QEMU, aborting");
2887 return -ENOTSUP;
2890 new_reftable = g_try_realloc(*reftable, new_reftable_size *
2891 REFTABLE_ENTRY_SIZE);
2892 if (!new_reftable) {
2893 error_setg(errp, "Failed to increase reftable buffer size");
2894 return -ENOMEM;
2897 memset(new_reftable + *reftable_size, 0,
2898 (new_reftable_size - *reftable_size) * REFTABLE_ENTRY_SIZE);
2900 *reftable = new_reftable;
2901 *reftable_size = new_reftable_size;
2904 if (!refblock_empty && !(*reftable)[reftable_index]) {
2905 offset = qcow2_alloc_clusters(bs, s->cluster_size);
2906 if (offset < 0) {
2907 error_setg_errno(errp, -offset, "Failed to allocate refblock");
2908 return offset;
2910 (*reftable)[reftable_index] = offset;
2911 *allocated = true;
2914 return 0;
2918 * This "operation" for walk_over_reftable() writes the refblock to disk at the
2919 * offset specified by the new reftable's entry. It does not modify the new
2920 * reftable or change any refcounts.
2922 static int flush_refblock(BlockDriverState *bs, uint64_t **reftable,
2923 uint64_t reftable_index, uint64_t *reftable_size,
2924 void *refblock, bool refblock_empty, bool *allocated,
2925 Error **errp)
2927 BDRVQcow2State *s = bs->opaque;
2928 int64_t offset;
2929 int ret;
2931 if (reftable_index < *reftable_size && (*reftable)[reftable_index]) {
2932 offset = (*reftable)[reftable_index];
2934 ret = qcow2_pre_write_overlap_check(bs, 0, offset, s->cluster_size,
2935 false);
2936 if (ret < 0) {
2937 error_setg_errno(errp, -ret, "Overlap check failed");
2938 return ret;
2941 ret = bdrv_pwrite(bs->file, offset, refblock, s->cluster_size);
2942 if (ret < 0) {
2943 error_setg_errno(errp, -ret, "Failed to write refblock");
2944 return ret;
2946 } else {
2947 assert(refblock_empty);
2950 return 0;
2954 * This function walks over the existing reftable and every referenced refblock;
2955 * if @new_set_refcount is non-NULL, it is called for every refcount entry to
2956 * create an equal new entry in the passed @new_refblock. Once that
2957 * @new_refblock is completely filled, @operation will be called.
2959 * @status_cb and @cb_opaque are used for the amend operation's status callback.
2960 * @index is the index of the walk_over_reftable() calls and @total is the total
2961 * number of walk_over_reftable() calls per amend operation. Both are used for
2962 * calculating the parameters for the status callback.
2964 * @allocated is set to true if a new cluster has been allocated.
2966 static int walk_over_reftable(BlockDriverState *bs, uint64_t **new_reftable,
2967 uint64_t *new_reftable_index,
2968 uint64_t *new_reftable_size,
2969 void *new_refblock, int new_refblock_size,
2970 int new_refcount_bits,
2971 RefblockFinishOp *operation, bool *allocated,
2972 Qcow2SetRefcountFunc *new_set_refcount,
2973 BlockDriverAmendStatusCB *status_cb,
2974 void *cb_opaque, int index, int total,
2975 Error **errp)
2977 BDRVQcow2State *s = bs->opaque;
2978 uint64_t reftable_index;
2979 bool new_refblock_empty = true;
2980 int refblock_index;
2981 int new_refblock_index = 0;
2982 int ret;
2984 for (reftable_index = 0; reftable_index < s->refcount_table_size;
2985 reftable_index++)
2987 uint64_t refblock_offset = s->refcount_table[reftable_index]
2988 & REFT_OFFSET_MASK;
2990 status_cb(bs, (uint64_t)index * s->refcount_table_size + reftable_index,
2991 (uint64_t)total * s->refcount_table_size, cb_opaque);
2993 if (refblock_offset) {
2994 void *refblock;
2996 if (offset_into_cluster(s, refblock_offset)) {
2997 qcow2_signal_corruption(bs, true, -1, -1, "Refblock offset %#"
2998 PRIx64 " unaligned (reftable index: %#"
2999 PRIx64 ")", refblock_offset,
3000 reftable_index);
3001 error_setg(errp,
3002 "Image is corrupt (unaligned refblock offset)");
3003 return -EIO;
3006 ret = qcow2_cache_get(bs, s->refcount_block_cache, refblock_offset,
3007 &refblock);
3008 if (ret < 0) {
3009 error_setg_errno(errp, -ret, "Failed to retrieve refblock");
3010 return ret;
3013 for (refblock_index = 0; refblock_index < s->refcount_block_size;
3014 refblock_index++)
3016 uint64_t refcount;
3018 if (new_refblock_index >= new_refblock_size) {
3019 /* new_refblock is now complete */
3020 ret = operation(bs, new_reftable, *new_reftable_index,
3021 new_reftable_size, new_refblock,
3022 new_refblock_empty, allocated, errp);
3023 if (ret < 0) {
3024 qcow2_cache_put(s->refcount_block_cache, &refblock);
3025 return ret;
3028 (*new_reftable_index)++;
3029 new_refblock_index = 0;
3030 new_refblock_empty = true;
3033 refcount = s->get_refcount(refblock, refblock_index);
3034 if (new_refcount_bits < 64 && refcount >> new_refcount_bits) {
3035 uint64_t offset;
3037 qcow2_cache_put(s->refcount_block_cache, &refblock);
3039 offset = ((reftable_index << s->refcount_block_bits)
3040 + refblock_index) << s->cluster_bits;
3042 error_setg(errp, "Cannot decrease refcount entry width to "
3043 "%i bits: Cluster at offset %#" PRIx64 " has a "
3044 "refcount of %" PRIu64, new_refcount_bits,
3045 offset, refcount);
3046 return -EINVAL;
3049 if (new_set_refcount) {
3050 new_set_refcount(new_refblock, new_refblock_index++,
3051 refcount);
3052 } else {
3053 new_refblock_index++;
3055 new_refblock_empty = new_refblock_empty && refcount == 0;
3058 qcow2_cache_put(s->refcount_block_cache, &refblock);
3059 } else {
3060 /* No refblock means every refcount is 0 */
3061 for (refblock_index = 0; refblock_index < s->refcount_block_size;
3062 refblock_index++)
3064 if (new_refblock_index >= new_refblock_size) {
3065 /* new_refblock is now complete */
3066 ret = operation(bs, new_reftable, *new_reftable_index,
3067 new_reftable_size, new_refblock,
3068 new_refblock_empty, allocated, errp);
3069 if (ret < 0) {
3070 return ret;
3073 (*new_reftable_index)++;
3074 new_refblock_index = 0;
3075 new_refblock_empty = true;
3078 if (new_set_refcount) {
3079 new_set_refcount(new_refblock, new_refblock_index++, 0);
3080 } else {
3081 new_refblock_index++;
3087 if (new_refblock_index > 0) {
3088 /* Complete the potentially existing partially filled final refblock */
3089 if (new_set_refcount) {
3090 for (; new_refblock_index < new_refblock_size;
3091 new_refblock_index++)
3093 new_set_refcount(new_refblock, new_refblock_index, 0);
3097 ret = operation(bs, new_reftable, *new_reftable_index,
3098 new_reftable_size, new_refblock, new_refblock_empty,
3099 allocated, errp);
3100 if (ret < 0) {
3101 return ret;
3104 (*new_reftable_index)++;
3107 status_cb(bs, (uint64_t)(index + 1) * s->refcount_table_size,
3108 (uint64_t)total * s->refcount_table_size, cb_opaque);
3110 return 0;
3113 int qcow2_change_refcount_order(BlockDriverState *bs, int refcount_order,
3114 BlockDriverAmendStatusCB *status_cb,
3115 void *cb_opaque, Error **errp)
3117 BDRVQcow2State *s = bs->opaque;
3118 Qcow2GetRefcountFunc *new_get_refcount;
3119 Qcow2SetRefcountFunc *new_set_refcount;
3120 void *new_refblock = qemu_blockalign(bs->file->bs, s->cluster_size);
3121 uint64_t *new_reftable = NULL, new_reftable_size = 0;
3122 uint64_t *old_reftable, old_reftable_size, old_reftable_offset;
3123 uint64_t new_reftable_index = 0;
3124 uint64_t i;
3125 int64_t new_reftable_offset = 0, allocated_reftable_size = 0;
3126 int new_refblock_size, new_refcount_bits = 1 << refcount_order;
3127 int old_refcount_order;
3128 int walk_index = 0;
3129 int ret;
3130 bool new_allocation;
3132 assert(s->qcow_version >= 3);
3133 assert(refcount_order >= 0 && refcount_order <= 6);
3135 /* see qcow2_open() */
3136 new_refblock_size = 1 << (s->cluster_bits - (refcount_order - 3));
3138 new_get_refcount = get_refcount_funcs[refcount_order];
3139 new_set_refcount = set_refcount_funcs[refcount_order];
3142 do {
3143 int total_walks;
3145 new_allocation = false;
3147 /* At least we have to do this walk and the one which writes the
3148 * refblocks; also, at least we have to do this loop here at least
3149 * twice (normally), first to do the allocations, and second to
3150 * determine that everything is correctly allocated, this then makes
3151 * three walks in total */
3152 total_walks = MAX(walk_index + 2, 3);
3154 /* First, allocate the structures so they are present in the refcount
3155 * structures */
3156 ret = walk_over_reftable(bs, &new_reftable, &new_reftable_index,
3157 &new_reftable_size, NULL, new_refblock_size,
3158 new_refcount_bits, &alloc_refblock,
3159 &new_allocation, NULL, status_cb, cb_opaque,
3160 walk_index++, total_walks, errp);
3161 if (ret < 0) {
3162 goto done;
3165 new_reftable_index = 0;
3167 if (new_allocation) {
3168 if (new_reftable_offset) {
3169 qcow2_free_clusters(
3170 bs, new_reftable_offset,
3171 allocated_reftable_size * REFTABLE_ENTRY_SIZE,
3172 QCOW2_DISCARD_NEVER);
3175 new_reftable_offset = qcow2_alloc_clusters(bs, new_reftable_size *
3176 REFTABLE_ENTRY_SIZE);
3177 if (new_reftable_offset < 0) {
3178 error_setg_errno(errp, -new_reftable_offset,
3179 "Failed to allocate the new reftable");
3180 ret = new_reftable_offset;
3181 goto done;
3183 allocated_reftable_size = new_reftable_size;
3185 } while (new_allocation);
3187 /* Second, write the new refblocks */
3188 ret = walk_over_reftable(bs, &new_reftable, &new_reftable_index,
3189 &new_reftable_size, new_refblock,
3190 new_refblock_size, new_refcount_bits,
3191 &flush_refblock, &new_allocation, new_set_refcount,
3192 status_cb, cb_opaque, walk_index, walk_index + 1,
3193 errp);
3194 if (ret < 0) {
3195 goto done;
3197 assert(!new_allocation);
3200 /* Write the new reftable */
3201 ret = qcow2_pre_write_overlap_check(bs, 0, new_reftable_offset,
3202 new_reftable_size * REFTABLE_ENTRY_SIZE,
3203 false);
3204 if (ret < 0) {
3205 error_setg_errno(errp, -ret, "Overlap check failed");
3206 goto done;
3209 for (i = 0; i < new_reftable_size; i++) {
3210 cpu_to_be64s(&new_reftable[i]);
3213 ret = bdrv_pwrite(bs->file, new_reftable_offset, new_reftable,
3214 new_reftable_size * REFTABLE_ENTRY_SIZE);
3216 for (i = 0; i < new_reftable_size; i++) {
3217 be64_to_cpus(&new_reftable[i]);
3220 if (ret < 0) {
3221 error_setg_errno(errp, -ret, "Failed to write the new reftable");
3222 goto done;
3226 /* Empty the refcount cache */
3227 ret = qcow2_cache_flush(bs, s->refcount_block_cache);
3228 if (ret < 0) {
3229 error_setg_errno(errp, -ret, "Failed to flush the refblock cache");
3230 goto done;
3233 /* Update the image header to point to the new reftable; this only updates
3234 * the fields which are relevant to qcow2_update_header(); other fields
3235 * such as s->refcount_table or s->refcount_bits stay stale for now
3236 * (because we have to restore everything if qcow2_update_header() fails) */
3237 old_refcount_order = s->refcount_order;
3238 old_reftable_size = s->refcount_table_size;
3239 old_reftable_offset = s->refcount_table_offset;
3241 s->refcount_order = refcount_order;
3242 s->refcount_table_size = new_reftable_size;
3243 s->refcount_table_offset = new_reftable_offset;
3245 ret = qcow2_update_header(bs);
3246 if (ret < 0) {
3247 s->refcount_order = old_refcount_order;
3248 s->refcount_table_size = old_reftable_size;
3249 s->refcount_table_offset = old_reftable_offset;
3250 error_setg_errno(errp, -ret, "Failed to update the qcow2 header");
3251 goto done;
3254 /* Now update the rest of the in-memory information */
3255 old_reftable = s->refcount_table;
3256 s->refcount_table = new_reftable;
3257 update_max_refcount_table_index(s);
3259 s->refcount_bits = 1 << refcount_order;
3260 s->refcount_max = UINT64_C(1) << (s->refcount_bits - 1);
3261 s->refcount_max += s->refcount_max - 1;
3263 s->refcount_block_bits = s->cluster_bits - (refcount_order - 3);
3264 s->refcount_block_size = 1 << s->refcount_block_bits;
3266 s->get_refcount = new_get_refcount;
3267 s->set_refcount = new_set_refcount;
3269 /* For cleaning up all old refblocks and the old reftable below the "done"
3270 * label */
3271 new_reftable = old_reftable;
3272 new_reftable_size = old_reftable_size;
3273 new_reftable_offset = old_reftable_offset;
3275 done:
3276 if (new_reftable) {
3277 /* On success, new_reftable actually points to the old reftable (and
3278 * new_reftable_size is the old reftable's size); but that is just
3279 * fine */
3280 for (i = 0; i < new_reftable_size; i++) {
3281 uint64_t offset = new_reftable[i] & REFT_OFFSET_MASK;
3282 if (offset) {
3283 qcow2_free_clusters(bs, offset, s->cluster_size,
3284 QCOW2_DISCARD_OTHER);
3287 g_free(new_reftable);
3289 if (new_reftable_offset > 0) {
3290 qcow2_free_clusters(bs, new_reftable_offset,
3291 new_reftable_size * REFTABLE_ENTRY_SIZE,
3292 QCOW2_DISCARD_OTHER);
3296 qemu_vfree(new_refblock);
3297 return ret;
3300 static int64_t get_refblock_offset(BlockDriverState *bs, uint64_t offset)
3302 BDRVQcow2State *s = bs->opaque;
3303 uint32_t index = offset_to_reftable_index(s, offset);
3304 int64_t covering_refblock_offset = 0;
3306 if (index < s->refcount_table_size) {
3307 covering_refblock_offset = s->refcount_table[index] & REFT_OFFSET_MASK;
3309 if (!covering_refblock_offset) {
3310 qcow2_signal_corruption(bs, true, -1, -1, "Refblock at %#" PRIx64 " is "
3311 "not covered by the refcount structures",
3312 offset);
3313 return -EIO;
3316 return covering_refblock_offset;
3319 static int qcow2_discard_refcount_block(BlockDriverState *bs,
3320 uint64_t discard_block_offs)
3322 BDRVQcow2State *s = bs->opaque;
3323 int64_t refblock_offs;
3324 uint64_t cluster_index = discard_block_offs >> s->cluster_bits;
3325 uint32_t block_index = cluster_index & (s->refcount_block_size - 1);
3326 void *refblock;
3327 int ret;
3329 refblock_offs = get_refblock_offset(bs, discard_block_offs);
3330 if (refblock_offs < 0) {
3331 return refblock_offs;
3334 assert(discard_block_offs != 0);
3336 ret = qcow2_cache_get(bs, s->refcount_block_cache, refblock_offs,
3337 &refblock);
3338 if (ret < 0) {
3339 return ret;
3342 if (s->get_refcount(refblock, block_index) != 1) {
3343 qcow2_signal_corruption(bs, true, -1, -1, "Invalid refcount:"
3344 " refblock offset %#" PRIx64
3345 ", reftable index %u"
3346 ", block offset %#" PRIx64
3347 ", refcount %#" PRIx64,
3348 refblock_offs,
3349 offset_to_reftable_index(s, discard_block_offs),
3350 discard_block_offs,
3351 s->get_refcount(refblock, block_index));
3352 qcow2_cache_put(s->refcount_block_cache, &refblock);
3353 return -EINVAL;
3355 s->set_refcount(refblock, block_index, 0);
3357 qcow2_cache_entry_mark_dirty(s->refcount_block_cache, refblock);
3359 qcow2_cache_put(s->refcount_block_cache, &refblock);
3361 if (cluster_index < s->free_cluster_index) {
3362 s->free_cluster_index = cluster_index;
3365 refblock = qcow2_cache_is_table_offset(s->refcount_block_cache,
3366 discard_block_offs);
3367 if (refblock) {
3368 /* discard refblock from the cache if refblock is cached */
3369 qcow2_cache_discard(s->refcount_block_cache, refblock);
3371 update_refcount_discard(bs, discard_block_offs, s->cluster_size);
3373 return 0;
3376 int qcow2_shrink_reftable(BlockDriverState *bs)
3378 BDRVQcow2State *s = bs->opaque;
3379 uint64_t *reftable_tmp =
3380 g_malloc(s->refcount_table_size * REFTABLE_ENTRY_SIZE);
3381 int i, ret;
3383 for (i = 0; i < s->refcount_table_size; i++) {
3384 int64_t refblock_offs = s->refcount_table[i] & REFT_OFFSET_MASK;
3385 void *refblock;
3386 bool unused_block;
3388 if (refblock_offs == 0) {
3389 reftable_tmp[i] = 0;
3390 continue;
3392 ret = qcow2_cache_get(bs, s->refcount_block_cache, refblock_offs,
3393 &refblock);
3394 if (ret < 0) {
3395 goto out;
3398 /* the refblock has own reference */
3399 if (i == offset_to_reftable_index(s, refblock_offs)) {
3400 uint64_t block_index = (refblock_offs >> s->cluster_bits) &
3401 (s->refcount_block_size - 1);
3402 uint64_t refcount = s->get_refcount(refblock, block_index);
3404 s->set_refcount(refblock, block_index, 0);
3406 unused_block = buffer_is_zero(refblock, s->cluster_size);
3408 s->set_refcount(refblock, block_index, refcount);
3409 } else {
3410 unused_block = buffer_is_zero(refblock, s->cluster_size);
3412 qcow2_cache_put(s->refcount_block_cache, &refblock);
3414 reftable_tmp[i] = unused_block ? 0 : cpu_to_be64(s->refcount_table[i]);
3417 ret = bdrv_pwrite_sync(bs->file, s->refcount_table_offset, reftable_tmp,
3418 s->refcount_table_size * REFTABLE_ENTRY_SIZE);
3420 * If the write in the reftable failed the image may contain a partially
3421 * overwritten reftable. In this case it would be better to clear the
3422 * reftable in memory to avoid possible image corruption.
3424 for (i = 0; i < s->refcount_table_size; i++) {
3425 if (s->refcount_table[i] && !reftable_tmp[i]) {
3426 if (ret == 0) {
3427 ret = qcow2_discard_refcount_block(bs, s->refcount_table[i] &
3428 REFT_OFFSET_MASK);
3430 s->refcount_table[i] = 0;
3434 if (!s->cache_discards) {
3435 qcow2_process_discards(bs, ret);
3438 out:
3439 g_free(reftable_tmp);
3440 return ret;
3443 int64_t qcow2_get_last_cluster(BlockDriverState *bs, int64_t size)
3445 BDRVQcow2State *s = bs->opaque;
3446 int64_t i;
3448 for (i = size_to_clusters(s, size) - 1; i >= 0; i--) {
3449 uint64_t refcount;
3450 int ret = qcow2_get_refcount(bs, i, &refcount);
3451 if (ret < 0) {
3452 fprintf(stderr, "Can't get refcount for cluster %" PRId64 ": %s\n",
3453 i, strerror(-ret));
3454 return ret;
3456 if (refcount > 0) {
3457 return i;
3460 qcow2_signal_corruption(bs, true, -1, -1,
3461 "There are no references in the refcount table.");
3462 return -EIO;
3465 int qcow2_detect_metadata_preallocation(BlockDriverState *bs)
3467 BDRVQcow2State *s = bs->opaque;
3468 int64_t i, end_cluster, cluster_count = 0, threshold;
3469 int64_t file_length, real_allocation, real_clusters;
3471 qemu_co_mutex_assert_locked(&s->lock);
3473 file_length = bdrv_getlength(bs->file->bs);
3474 if (file_length < 0) {
3475 return file_length;
3478 real_allocation = bdrv_get_allocated_file_size(bs->file->bs);
3479 if (real_allocation < 0) {
3480 return real_allocation;
3483 real_clusters = real_allocation / s->cluster_size;
3484 threshold = MAX(real_clusters * 10 / 9, real_clusters + 2);
3486 end_cluster = size_to_clusters(s, file_length);
3487 for (i = 0; i < end_cluster && cluster_count < threshold; i++) {
3488 uint64_t refcount;
3489 int ret = qcow2_get_refcount(bs, i, &refcount);
3490 if (ret < 0) {
3491 return ret;
3493 cluster_count += !!refcount;
3496 return cluster_count >= threshold;