#include "qemu/range.h"
#include "qapi/qmp/types.h"
-static int64_t alloc_clusters_noref(BlockDriverState *bs, int64_t size);
+static int64_t alloc_clusters_noref(BlockDriverState *bs, uint64_t size);
static int QEMU_WARN_UNUSED_RESULT update_refcount(BlockDriverState *bs,
int64_t offset, int64_t length,
int addend, enum qcow2_discard_type type);
int qcow2_refcount_init(BlockDriverState *bs)
{
BDRVQcowState *s = bs->opaque;
- int ret, refcount_table_size2, i;
+ unsigned int refcount_table_size2, i;
+ int ret;
+ assert(s->refcount_table_size <= INT_MAX / sizeof(uint64_t));
refcount_table_size2 = s->refcount_table_size * sizeof(uint64_t);
s->refcount_table = g_malloc(refcount_table_size2);
if (s->refcount_table_size > 0) {
static int get_refcount(BlockDriverState *bs, int64_t cluster_index)
{
BDRVQcowState *s = bs->opaque;
- int refcount_table_index, block_index;
+ uint64_t refcount_table_index, block_index;
int64_t refcount_block_offset;
int ret;
uint16_t *refcount_block;
refcount_table_index = cluster_index >> (s->cluster_bits - REFCOUNT_SHIFT);
if (refcount_table_index >= s->refcount_table_size)
return 0;
- refcount_block_offset = s->refcount_table[refcount_table_index];
+ refcount_block_offset =
+ s->refcount_table[refcount_table_index] & REFT_OFFSET_MASK;
if (!refcount_block_offset)
return 0;
* they can describe them themselves.
*
* - We need to consider that at this point we are inside update_refcounts
- * and doing the initial refcount increase. This means that some clusters
- * have already been allocated by the caller, but their refcount isn't
- * accurate yet. free_cluster_index tells us where this allocation ends
- * as long as we don't overwrite it by freeing clusters.
+ * and potentially doing an initial refcount increase. This means that
+ * some clusters have already been allocated by the caller, but their
+ * refcount isn't accurate yet. If we allocate clusters for metadata, we
+ * need to return -EAGAIN to signal the caller that it needs to restart
+ * the search for free clusters.
*
* - alloc_clusters_noref and qcow2_free_clusters may load a different
* refcount block into the cache
}
s->refcount_table[refcount_table_index] = new_block;
- return 0;
+
+ /* The new refcount block may be where the caller intended to put its
+ * data, so let it restart the search. */
+ return -EAGAIN;
}
ret = qcow2_cache_put(bs, s->refcount_block_cache, (void**) refcount_block);
/* Calculate the number of refcount blocks needed so far */
uint64_t refcount_block_clusters = 1 << (s->cluster_bits - REFCOUNT_SHIFT);
- uint64_t blocks_used = (s->free_cluster_index +
- refcount_block_clusters - 1) / refcount_block_clusters;
+ uint64_t blocks_used = DIV_ROUND_UP(cluster_index, refcount_block_clusters);
+
+ if (blocks_used > QCOW_MAX_REFTABLE_SIZE / sizeof(uint64_t)) {
+ return -EFBIG;
+ }
/* And now we need at least one block more for the new metadata */
uint64_t table_size = next_refcount_table_size(s, blocks_used + 1);
uint16_t *new_blocks = g_malloc0(blocks_clusters * s->cluster_size);
uint64_t *new_table = g_malloc0(table_size * sizeof(uint64_t));
- assert(meta_offset >= (s->free_cluster_index * s->cluster_size));
-
/* Fill the new refcount table */
memcpy(new_table, s->refcount_table,
s->refcount_table_size * sizeof(uint64_t));
s->refcount_table_size = table_size;
s->refcount_table_offset = table_offset;
- /* Free old table. Remember, we must not change free_cluster_index */
- uint64_t old_free_cluster_index = s->free_cluster_index;
+ /* Free old table. */
qcow2_free_clusters(bs, old_table_offset, old_table_size * sizeof(uint64_t),
QCOW2_DISCARD_OTHER);
- s->free_cluster_index = old_free_cluster_index;
ret = load_refcount_block(bs, new_block, (void**) refcount_block);
if (ret < 0) {
return ret;
}
- return 0;
+ /* If we were trying to do the initial refcount update for some cluster
+ * allocation, we might have used the same clusters to store newly
+ * allocated metadata. Make the caller search some new space. */
+ return -EAGAIN;
fail_table:
g_free(new_table);
s->l2_table_cache);
}
- start = offset & ~(s->cluster_size - 1);
- last = (offset + length - 1) & ~(s->cluster_size - 1);
+ start = start_of_cluster(s, offset);
+ last = start_of_cluster(s, offset + length - 1);
for(cluster_offset = start; cluster_offset <= last;
cluster_offset += s->cluster_size)
{
/* return < 0 if error */
-static int64_t alloc_clusters_noref(BlockDriverState *bs, int64_t size)
+static int64_t alloc_clusters_noref(BlockDriverState *bs, uint64_t size)
{
BDRVQcowState *s = bs->opaque;
- int i, nb_clusters, refcount;
+ uint64_t i, nb_clusters;
+ int refcount;
nb_clusters = size_to_clusters(s, size);
retry:
for(i = 0; i < nb_clusters; i++) {
- int64_t next_cluster_index = s->free_cluster_index++;
+ uint64_t next_cluster_index = s->free_cluster_index++;
refcount = get_refcount(bs, next_cluster_index);
if (refcount < 0) {
goto retry;
}
}
+
+ /* Make sure that all offsets in the "allocated" range are representable
+ * in an int64_t */
+ if (s->free_cluster_index > 0 &&
+ s->free_cluster_index - 1 > (INT64_MAX >> s->cluster_bits))
+ {
+ return -EFBIG;
+ }
+
#ifdef DEBUG_ALLOC2
fprintf(stderr, "alloc_clusters: size=%" PRId64 " -> %" PRId64 "\n",
size,
return (s->free_cluster_index - nb_clusters) << s->cluster_bits;
}
-int64_t qcow2_alloc_clusters(BlockDriverState *bs, int64_t size)
+int64_t qcow2_alloc_clusters(BlockDriverState *bs, uint64_t size)
{
int64_t offset;
int ret;
BLKDBG_EVENT(bs->file, BLKDBG_CLUSTER_ALLOC);
- offset = alloc_clusters_noref(bs, size);
- if (offset < 0) {
- return offset;
- }
+ do {
+ offset = alloc_clusters_noref(bs, size);
+ if (offset < 0) {
+ return offset;
+ }
+
+ ret = update_refcount(bs, offset, size, 1, QCOW2_DISCARD_NEVER);
+ } while (ret == -EAGAIN);
- ret = update_refcount(bs, offset, size, 1, QCOW2_DISCARD_NEVER);
if (ret < 0) {
return ret;
}
{
BDRVQcowState *s = bs->opaque;
uint64_t cluster_index;
- uint64_t old_free_cluster_index;
- int i, refcount, ret;
+ uint64_t i;
+ int refcount, ret;
- /* Check how many clusters there are free */
- cluster_index = offset >> s->cluster_bits;
- for(i = 0; i < nb_clusters; i++) {
- refcount = get_refcount(bs, cluster_index++);
+ assert(nb_clusters >= 0);
+ if (nb_clusters == 0) {
+ return 0;
+ }
- if (refcount < 0) {
- return refcount;
- } else if (refcount != 0) {
- break;
+ do {
+ /* Check how many clusters there are free */
+ cluster_index = offset >> s->cluster_bits;
+ for(i = 0; i < nb_clusters; i++) {
+ refcount = get_refcount(bs, cluster_index++);
+
+ if (refcount < 0) {
+ return refcount;
+ } else if (refcount != 0) {
+ break;
+ }
}
- }
- /* And then allocate them */
- old_free_cluster_index = s->free_cluster_index;
- s->free_cluster_index = cluster_index + i;
+ /* And then allocate them */
+ ret = update_refcount(bs, offset, i << s->cluster_bits, 1,
+ QCOW2_DISCARD_NEVER);
+ } while (ret == -EAGAIN);
- ret = update_refcount(bs, offset, i << s->cluster_bits, 1,
- QCOW2_DISCARD_NEVER);
if (ret < 0) {
return ret;
}
- s->free_cluster_index = old_free_cluster_index;
-
return i;
}
}
redo:
free_in_cluster = s->cluster_size -
- (s->free_byte_offset & (s->cluster_size - 1));
+ offset_into_cluster(s, s->free_byte_offset);
if (size <= free_in_cluster) {
/* enough space in current cluster */
offset = s->free_byte_offset;
free_in_cluster -= size;
if (free_in_cluster == 0)
s->free_byte_offset = 0;
- if ((offset & (s->cluster_size - 1)) != 0)
+ if (offset_into_cluster(s, offset) != 0)
qcow2_update_cluster_refcount(bs, offset >> s->cluster_bits, 1,
QCOW2_DISCARD_NEVER);
} else {
if (offset < 0) {
return offset;
}
- cluster_offset = s->free_byte_offset & ~(s->cluster_size - 1);
+ cluster_offset = start_of_cluster(s, s->free_byte_offset);
if ((cluster_offset + s->cluster_size) == offset) {
/* we are lucky: contiguous data */
offset = s->free_byte_offset;
}
break;
case QCOW2_CLUSTER_NORMAL:
- qcow2_free_clusters(bs, l2_entry & L2E_OFFSET_MASK,
- nb_clusters << s->cluster_bits, type);
+ case QCOW2_CLUSTER_ZERO:
+ if (l2_entry & L2E_OFFSET_MASK) {
+ qcow2_free_clusters(bs, l2_entry & L2E_OFFSET_MASK,
+ nb_clusters << s->cluster_bits, type);
+ }
break;
case QCOW2_CLUSTER_UNALLOCATED:
- case QCOW2_CLUSTER_ZERO:
break;
default:
abort();
nb_csectors = ((offset >> s->csize_shift) &
s->csize_mask) + 1;
if (addend != 0) {
- int ret;
ret = update_refcount(bs,
(offset & s->cluster_offset_mask) & ~511,
nb_csectors * 512, addend,
int64_t offset, int64_t size)
{
BDRVQcowState *s = bs->opaque;
- int64_t start, last, cluster_offset;
- int k;
+ uint64_t start, last, cluster_offset, k;
if (size <= 0)
return;
- start = offset & ~(s->cluster_size - 1);
- last = (offset + size - 1) & ~(s->cluster_size - 1);
+ start = start_of_cluster(s, offset);
+ last = start_of_cluster(s, offset + size - 1);
for(cluster_offset = start; cluster_offset <= last;
cluster_offset += s->cluster_size) {
k = cluster_offset >> s->cluster_bits;
- if (k < 0) {
- fprintf(stderr, "ERROR: invalid cluster offset=0x%" PRIx64 "\n",
- cluster_offset);
- res->corruptions++;
- } else if (k >= refcount_table_size) {
+ if (k >= refcount_table_size) {
fprintf(stderr, "Warning: cluster offset=0x%" PRIx64 " is after "
"the end of the image file, can't properly check refcounts.\n",
cluster_offset);
/* Flags for check_refcounts_l1() and check_refcounts_l2() */
enum {
- CHECK_OFLAG_COPIED = 0x1, /* check QCOW_OFLAG_COPIED matches refcount */
CHECK_FRAG_INFO = 0x2, /* update BlockFragInfo counters */
};
offset, s->cluster_size);
/* Correct offsets are cluster aligned */
- if (offset & (s->cluster_size - 1)) {
+ if (offset_into_cluster(s, offset)) {
fprintf(stderr, "ERROR offset=%" PRIx64 ": Cluster is not "
"properly aligned; L2 entry corrupted.\n", offset);
res->corruptions++;
l2_offset, s->cluster_size);
/* L2 tables are cluster aligned */
- if (l2_offset & (s->cluster_size - 1)) {
+ if (offset_into_cluster(s, l2_offset)) {
fprintf(stderr, "ERROR l2_offset=%" PRIx64 ": Table is not "
"cluster aligned; L1 entry corrupted\n", l2_offset);
res->corruptions++;
}
if (l2_dirty) {
- ret = qcow2_pre_write_overlap_check(bs,
- QCOW2_OL_DEFAULT & ~QCOW2_OL_ACTIVE_L2, l2_offset,
- s->cluster_size);
+ ret = qcow2_pre_write_overlap_check(bs, QCOW2_OL_ACTIVE_L2,
+ l2_offset, s->cluster_size);
if (ret < 0) {
fprintf(stderr, "ERROR: Could not write L2 table; metadata "
"overlap check failed: %s\n", strerror(-ret));
buf[i] = cpu_to_be64(s->refcount_table[rt_start_index + i]);
}
- ret = qcow2_pre_write_overlap_check(bs,
- QCOW2_OL_DEFAULT & ~QCOW2_OL_REFCOUNT_TABLE,
+ ret = qcow2_pre_write_overlap_check(bs, QCOW2_OL_REFCOUNT_TABLE,
s->refcount_table_offset + rt_start_index * sizeof(uint64_t),
sizeof(buf));
if (ret < 0) {
* does _not_ decrement the reference count for the currently occupied cluster.
*
* This function prints an informative message to stderr on error (and returns
- * -errno); on success, 0 is returned.
+ * -errno); on success, the offset of the newly allocated cluster is returned.
*/
static int64_t realloc_refcount_block(BlockDriverState *bs, int reftable_index,
uint64_t offset)
fprintf(stderr, "Could not allocate new cluster: %s\n",
strerror(-new_offset));
ret = new_offset;
- goto fail;
+ goto done;
}
/* fetch current refcount block content */
ret = qcow2_cache_get(bs, s->refcount_block_cache, offset, &refcount_block);
if (ret < 0) {
fprintf(stderr, "Could not fetch refcount block: %s\n", strerror(-ret));
- goto fail;
+ goto fail_free_cluster;
}
/* new block has not yet been entered into refcount table, therefore it is
* no refcount block yet (regarding this check) */
- ret = qcow2_pre_write_overlap_check(bs, QCOW2_OL_DEFAULT, new_offset,
- s->cluster_size);
+ ret = qcow2_pre_write_overlap_check(bs, 0, new_offset, s->cluster_size);
if (ret < 0) {
fprintf(stderr, "Could not write refcount block; metadata overlap "
"check failed: %s\n", strerror(-ret));
/* the image will be marked corrupt, so don't even attempt on freeing
* the cluster */
- new_offset = 0;
- goto fail;
+ goto done;
}
/* write to new block */
s->cluster_sectors);
if (ret < 0) {
fprintf(stderr, "Could not write refcount block: %s\n", strerror(-ret));
- goto fail;
+ goto fail_free_cluster;
}
/* update refcount table */
- assert(!(new_offset & (s->cluster_size - 1)));
+ assert(!offset_into_cluster(s, new_offset));
s->refcount_table[reftable_index] = new_offset;
ret = write_reftable_entry(bs, reftable_index);
if (ret < 0) {
fprintf(stderr, "Could not update refcount table: %s\n",
strerror(-ret));
- goto fail;
+ goto fail_free_cluster;
}
-fail:
- if (new_offset && (ret < 0)) {
- qcow2_free_clusters(bs, new_offset, s->cluster_size,
- QCOW2_DISCARD_ALWAYS);
- }
+ goto done;
+
+fail_free_cluster:
+ qcow2_free_clusters(bs, new_offset, s->cluster_size, QCOW2_DISCARD_OTHER);
+
+done:
if (refcount_block) {
- if (ret < 0) {
- qcow2_cache_put(bs, s->refcount_block_cache, &refcount_block);
- } else {
- ret = qcow2_cache_put(bs, s->refcount_block_cache, &refcount_block);
- }
+ /* This should never fail, as it would only do so if the given refcount
+ * block cannot be found in the cache. As this is impossible as long as
+ * there are no bugs, assert the success. */
+ int tmp = qcow2_cache_put(bs, s->refcount_block_cache, &refcount_block);
+ assert(tmp == 0);
}
+
if (ret < 0) {
return ret;
}
+
return new_offset;
}
BdrvCheckMode fix)
{
BDRVQcowState *s = bs->opaque;
- int64_t size, i, highest_cluster;
- int nb_clusters, refcount1, refcount2;
+ int64_t size, i, highest_cluster, nb_clusters;
+ int refcount1, refcount2;
QCowSnapshot *sn;
uint16_t *refcount_table;
int ret;
size = bdrv_getlength(bs->file);
+ if (size < 0) {
+ res->check_errors++;
+ return size;
+ }
+
nb_clusters = size_to_clusters(s, size);
+ if (nb_clusters > INT_MAX) {
+ res->check_errors++;
+ return -EFBIG;
+ }
+
refcount_table = g_malloc0(nb_clusters * sizeof(uint16_t));
res->bfi.total_clusters =
/* current L1 table */
ret = check_refcounts_l1(bs, res, refcount_table, nb_clusters,
- s->l1_table_offset, s->l1_size,
- CHECK_OFLAG_COPIED | CHECK_FRAG_INFO);
+ s->l1_table_offset, s->l1_size, CHECK_FRAG_INFO);
if (ret < 0) {
goto fail;
}
cluster = offset >> s->cluster_bits;
/* Refcount blocks are cluster aligned */
- if (offset & (s->cluster_size - 1)) {
+ if (offset_into_cluster(s, offset)) {
fprintf(stderr, "ERROR refcount block %" PRId64 " is not "
"cluster aligned; refcount table entry corrupted\n", i);
res->corruptions++;
* looking for overlaps with important metadata sections (L1/L2 tables etc.),
* i.e. a sanity check without relying on the refcount tables.
*
- * The chk parameter specifies exactly what checks to perform (being a bitmask
- * of QCow2MetadataOverlap values).
+ * The ign parameter specifies what checks not to perform (being a bitmask of
+ * QCow2MetadataOverlap values), i.e., what sections to ignore.
*
* Returns:
* - 0 if writing to this offset will not affect the mentioned metadata
* - a negative value (-errno) indicating an error while performing a check,
* e.g. when bdrv_read failed on QCOW2_OL_INACTIVE_L2
*/
-int qcow2_check_metadata_overlap(BlockDriverState *bs, int chk, int64_t offset,
+int qcow2_check_metadata_overlap(BlockDriverState *bs, int ign, int64_t offset,
int64_t size)
{
BDRVQcowState *s = bs->opaque;
+ int chk = s->overlap_check & ~ign;
int i, j;
if (!size) {
for (i = 0; i < s->nb_snapshots; i++) {
uint64_t l1_ofs = s->snapshots[i].l1_table_offset;
uint32_t l1_sz = s->snapshots[i].l1_size;
- uint64_t *l1 = g_malloc(l1_sz * sizeof(uint64_t));
+ uint64_t l1_sz2 = l1_sz * sizeof(uint64_t);
+ uint64_t *l1 = g_malloc(l1_sz2);
int ret;
- ret = bdrv_read(bs->file, l1_ofs / BDRV_SECTOR_SIZE, (uint8_t *)l1,
- l1_sz * sizeof(uint64_t) / BDRV_SECTOR_SIZE);
-
+ ret = bdrv_pread(bs->file, l1_ofs, l1, l1_sz2);
if (ret < 0) {
g_free(l1);
return ret;
}
for (j = 0; j < l1_sz; j++) {
- if ((l1[j] & L1E_OFFSET_MASK) &&
- overlaps_with(l1[j] & L1E_OFFSET_MASK, s->cluster_size)) {
+ uint64_t l2_ofs = be64_to_cpu(l1[j]) & L1E_OFFSET_MASK;
+ if (l2_ofs && overlaps_with(l2_ofs, s->cluster_size)) {
g_free(l1);
return QCOW2_OL_INACTIVE_L2;
}
* Returns 0 if there were neither overlaps nor errors while checking for
* overlaps; or a negative value (-errno) on error.
*/
-int qcow2_pre_write_overlap_check(BlockDriverState *bs, int chk, int64_t offset,
+int qcow2_pre_write_overlap_check(BlockDriverState *bs, int ign, int64_t offset,
int64_t size)
{
- int ret = qcow2_check_metadata_overlap(bs, chk, offset, size);
+ int ret = qcow2_check_metadata_overlap(bs, ign, offset, size);
if (ret < 0) {
return ret;
projects / qemu.git / blobdiff