aboutsummaryrefslogtreecommitdiff
path: root/block/qcow2-cluster.c
diff options
context:
space:
mode:
Diffstat (limited to 'block/qcow2-cluster.c')
-rw-r--r--block/qcow2-cluster.c144
1 files changed, 71 insertions, 73 deletions
diff --git a/block/qcow2-cluster.c b/block/qcow2-cluster.c
index 4d888c7..df0b2c9 100644
--- a/block/qcow2-cluster.c
+++ b/block/qcow2-cluster.c
@@ -1414,7 +1414,7 @@ int qcow2_decompress_cluster(BlockDriverState *bs, uint64_t cluster_offset)
* clusters.
*/
static int discard_single_l2(BlockDriverState *bs, uint64_t offset,
- unsigned int nb_clusters, enum qcow2_discard_type type)
+ unsigned int nb_clusters, enum qcow2_discard_type type, bool full_discard)
{
BDRVQcowState *s = bs->opaque;
uint64_t *l2_table;
@@ -1436,23 +1436,30 @@ static int discard_single_l2(BlockDriverState *bs, uint64_t offset,
old_l2_entry = be64_to_cpu(l2_table[l2_index + i]);
/*
- * Make sure that a discarded area reads back as zeroes for v3 images
- * (we cannot do it for v2 without actually writing a zero-filled
- * buffer). We can skip the operation if the cluster is already marked
- * as zero, or if it's unallocated and we don't have a backing file.
+ * If full_discard is false, make sure that a discarded area reads back
+ * as zeroes for v3 images (we cannot do it for v2 without actually
+ * writing a zero-filled buffer). We can skip the operation if the
+ * cluster is already marked as zero, or if it's unallocated and we
+ * don't have a backing file.
*
* TODO We might want to use bdrv_get_block_status(bs) here, but we're
* holding s->lock, so that doesn't work today.
+ *
+ * If full_discard is true, the sector should not read back as zeroes,
+ * but rather fall through to the backing file.
*/
switch (qcow2_get_cluster_type(old_l2_entry)) {
case QCOW2_CLUSTER_UNALLOCATED:
- if (!bs->backing_hd) {
+ if (full_discard || !bs->backing_hd) {
continue;
}
break;
case QCOW2_CLUSTER_ZERO:
- continue;
+ if (!full_discard) {
+ continue;
+ }
+ break;
case QCOW2_CLUSTER_NORMAL:
case QCOW2_CLUSTER_COMPRESSED:
@@ -1464,7 +1471,7 @@ static int discard_single_l2(BlockDriverState *bs, uint64_t offset,
/* First remove L2 entries */
qcow2_cache_entry_mark_dirty(s->l2_table_cache, l2_table);
- if (s->qcow_version >= 3) {
+ if (!full_discard && s->qcow_version >= 3) {
l2_table[l2_index + i] = cpu_to_be64(QCOW_OFLAG_ZERO);
} else {
l2_table[l2_index + i] = cpu_to_be64(0);
@@ -1483,7 +1490,7 @@ static int discard_single_l2(BlockDriverState *bs, uint64_t offset,
}
int qcow2_discard_clusters(BlockDriverState *bs, uint64_t offset,
- int nb_sectors, enum qcow2_discard_type type)
+ int nb_sectors, enum qcow2_discard_type type, bool full_discard)
{
BDRVQcowState *s = bs->opaque;
uint64_t end_offset;
@@ -1506,7 +1513,7 @@ int qcow2_discard_clusters(BlockDriverState *bs, uint64_t offset,
/* Each L2 table is handled by its own loop iteration */
while (nb_clusters > 0) {
- ret = discard_single_l2(bs, offset, nb_clusters, type);
+ ret = discard_single_l2(bs, offset, nb_clusters, type, full_discard);
if (ret < 0) {
goto fail;
}
@@ -1606,15 +1613,14 @@ fail:
* Expands all zero clusters in a specific L1 table (or deallocates them, for
* non-backed non-pre-allocated zero clusters).
*
- * expanded_clusters is a bitmap where every bit corresponds to one cluster in
- * the image file; a bit gets set if the corresponding cluster has been used for
- * zero expansion (i.e., has been filled with zeroes and is referenced from an
- * L2 table). nb_clusters contains the total cluster count of the image file,
- * i.e., the number of bits in expanded_clusters.
+ * l1_entries and *visited_l1_entries are used to keep track of progress for
+ * status_cb(). l1_entries contains the total number of L1 entries and
+ * *visited_l1_entries counts all visited L1 entries.
*/
static int expand_zero_clusters_in_l1(BlockDriverState *bs, uint64_t *l1_table,
- int l1_size, uint8_t **expanded_clusters,
- uint64_t *nb_clusters)
+ int l1_size, int64_t *visited_l1_entries,
+ int64_t l1_entries,
+ BlockDriverAmendStatusCB *status_cb)
{
BDRVQcowState *s = bs->opaque;
bool is_active_l1 = (l1_table == s->l1_table);
@@ -1634,9 +1640,14 @@ static int expand_zero_clusters_in_l1(BlockDriverState *bs, uint64_t *l1_table,
for (i = 0; i < l1_size; i++) {
uint64_t l2_offset = l1_table[i] & L1E_OFFSET_MASK;
bool l2_dirty = false;
+ int l2_refcount;
if (!l2_offset) {
/* unallocated */
+ (*visited_l1_entries)++;
+ if (status_cb) {
+ status_cb(bs, *visited_l1_entries, l1_entries);
+ }
continue;
}
@@ -1653,33 +1664,19 @@ static int expand_zero_clusters_in_l1(BlockDriverState *bs, uint64_t *l1_table,
goto fail;
}
+ l2_refcount = qcow2_get_refcount(bs, l2_offset >> s->cluster_bits);
+ if (l2_refcount < 0) {
+ ret = l2_refcount;
+ goto fail;
+ }
+
for (j = 0; j < s->l2_size; j++) {
uint64_t l2_entry = be64_to_cpu(l2_table[j]);
- int64_t offset = l2_entry & L2E_OFFSET_MASK, cluster_index;
+ int64_t offset = l2_entry & L2E_OFFSET_MASK;
int cluster_type = qcow2_get_cluster_type(l2_entry);
bool preallocated = offset != 0;
- if (cluster_type == QCOW2_CLUSTER_NORMAL) {
- cluster_index = offset >> s->cluster_bits;
- assert((cluster_index >= 0) && (cluster_index < *nb_clusters));
- if ((*expanded_clusters)[cluster_index / 8] &
- (1 << (cluster_index % 8))) {
- /* Probably a shared L2 table; this cluster was a zero
- * cluster which has been expanded, its refcount
- * therefore most likely requires an update. */
- ret = qcow2_update_cluster_refcount(bs, cluster_index, 1,
- QCOW2_DISCARD_NEVER);
- if (ret < 0) {
- goto fail;
- }
- /* Since we just increased the refcount, the COPIED flag may
- * no longer be set. */
- l2_table[j] = cpu_to_be64(l2_entry & ~QCOW_OFLAG_COPIED);
- l2_dirty = true;
- }
- continue;
- }
- else if (qcow2_get_cluster_type(l2_entry) != QCOW2_CLUSTER_ZERO) {
+ if (cluster_type != QCOW2_CLUSTER_ZERO) {
continue;
}
@@ -1697,6 +1694,19 @@ static int expand_zero_clusters_in_l1(BlockDriverState *bs, uint64_t *l1_table,
ret = offset;
goto fail;
}
+
+ if (l2_refcount > 1) {
+ /* For shared L2 tables, set the refcount accordingly (it is
+ * already 1 and needs to be l2_refcount) */
+ ret = qcow2_update_cluster_refcount(bs,
+ offset >> s->cluster_bits, l2_refcount - 1,
+ QCOW2_DISCARD_OTHER);
+ if (ret < 0) {
+ qcow2_free_clusters(bs, offset, s->cluster_size,
+ QCOW2_DISCARD_OTHER);
+ goto fail;
+ }
+ }
}
ret = qcow2_pre_write_overlap_check(bs, 0, offset, s->cluster_size);
@@ -1718,29 +1728,12 @@ static int expand_zero_clusters_in_l1(BlockDriverState *bs, uint64_t *l1_table,
goto fail;
}
- l2_table[j] = cpu_to_be64(offset | QCOW_OFLAG_COPIED);
- l2_dirty = true;
-
- cluster_index = offset >> s->cluster_bits;
-
- if (cluster_index >= *nb_clusters) {
- uint64_t old_bitmap_size = (*nb_clusters + 7) / 8;
- uint64_t new_bitmap_size;
- /* The offset may lie beyond the old end of the underlying image
- * file for growable files only */
- assert(bs->file->growable);
- *nb_clusters = size_to_clusters(s, bs->file->total_sectors *
- BDRV_SECTOR_SIZE);
- new_bitmap_size = (*nb_clusters + 7) / 8;
- *expanded_clusters = g_realloc(*expanded_clusters,
- new_bitmap_size);
- /* clear the newly allocated space */
- memset(&(*expanded_clusters)[old_bitmap_size], 0,
- new_bitmap_size - old_bitmap_size);
+ if (l2_refcount == 1) {
+ l2_table[j] = cpu_to_be64(offset | QCOW_OFLAG_COPIED);
+ } else {
+ l2_table[j] = cpu_to_be64(offset);
}
-
- assert((cluster_index >= 0) && (cluster_index < *nb_clusters));
- (*expanded_clusters)[cluster_index / 8] |= 1 << (cluster_index % 8);
+ l2_dirty = true;
}
if (is_active_l1) {
@@ -1769,6 +1762,11 @@ static int expand_zero_clusters_in_l1(BlockDriverState *bs, uint64_t *l1_table,
}
}
}
+
+ (*visited_l1_entries)++;
+ if (status_cb) {
+ status_cb(bs, *visited_l1_entries, l1_entries);
+ }
}
ret = 0;
@@ -1795,25 +1793,25 @@ fail:
* allocation for pre-allocated ones). This is important for downgrading to a
* qcow2 version which doesn't yet support metadata zero clusters.
*/
-int qcow2_expand_zero_clusters(BlockDriverState *bs)
+int qcow2_expand_zero_clusters(BlockDriverState *bs,
+ BlockDriverAmendStatusCB *status_cb)
{
BDRVQcowState *s = bs->opaque;
uint64_t *l1_table = NULL;
- uint64_t nb_clusters;
- uint8_t *expanded_clusters;
+ int64_t l1_entries = 0, visited_l1_entries = 0;
int ret;
int i, j;
- nb_clusters = size_to_clusters(s, bs->file->total_sectors *
- BDRV_SECTOR_SIZE);
- expanded_clusters = g_try_malloc0((nb_clusters + 7) / 8);
- if (expanded_clusters == NULL) {
- ret = -ENOMEM;
- goto fail;
+ if (status_cb) {
+ l1_entries = s->l1_size;
+ for (i = 0; i < s->nb_snapshots; i++) {
+ l1_entries += s->snapshots[i].l1_size;
+ }
}
ret = expand_zero_clusters_in_l1(bs, s->l1_table, s->l1_size,
- &expanded_clusters, &nb_clusters);
+ &visited_l1_entries, l1_entries,
+ status_cb);
if (ret < 0) {
goto fail;
}
@@ -1847,7 +1845,8 @@ int qcow2_expand_zero_clusters(BlockDriverState *bs)
}
ret = expand_zero_clusters_in_l1(bs, l1_table, s->snapshots[i].l1_size,
- &expanded_clusters, &nb_clusters);
+ &visited_l1_entries, l1_entries,
+ status_cb);
if (ret < 0) {
goto fail;
}
@@ -1856,7 +1855,6 @@ int qcow2_expand_zero_clusters(BlockDriverState *bs)
ret = 0;
fail:
- g_free(expanded_clusters);
g_free(l1_table);
return ret;
}