qcow2: use one single memory block for the L2/refcount cache tables

The qcow2 L2/refcount cache contains one separate table for each cache
entry. Doing one allocation per table adds unnecessary overhead and it
also requires us to store the address of each table separately.

Since the size of the cache is constant during its lifetime, it's
better to have an array that contains all the tables using one single
allocation.

In my tests measuring freshly created caches with sizes 128MB (L2) and
32MB (refcount) this uses around 10MB of RAM less.

Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

1 files changed, 6 insertions, 6 deletions

diff --git a/block/qcow2-cluster.c b/block/qcow2-cluster.c
index ed2b44d..5cd418a 100644
--- a/block/qcow2-cluster.c
+++ b/block/qcow2-cluster.c
@@ -263,7 +263,7 @@ static int l2_allocate(BlockDriverState *bs, int l1_index, uint64_t **table)
     BLKDBG_EVENT(bs->file, BLKDBG_L2_ALLOC_WRITE);
 
     trace_qcow2_l2_allocate_write_l2(bs, l1_index);
-    qcow2_cache_entry_mark_dirty(s->l2_table_cache, l2_table);
+    qcow2_cache_entry_mark_dirty(bs, s->l2_table_cache, l2_table);
     ret = qcow2_cache_flush(bs, s->l2_table_cache);
     if (ret < 0) {
         goto fail;
@@ -692,7 +692,7 @@ uint64_t qcow2_alloc_compressed_cluster_offset(BlockDriverState *bs,
     /* compressed clusters never have the copied flag */
 
     BLKDBG_EVENT(bs->file, BLKDBG_L2_UPDATE_COMPRESSED);
-    qcow2_cache_entry_mark_dirty(s->l2_table_cache, l2_table);
+    qcow2_cache_entry_mark_dirty(bs, s->l2_table_cache, l2_table);
     l2_table[l2_index] = cpu_to_be64(cluster_offset);
     ret = qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
     if (ret < 0) {
@@ -771,7 +771,7 @@ int qcow2_alloc_cluster_link_l2(BlockDriverState *bs, QCowL2Meta *m)
     if (ret < 0) {
         goto err;
     }
-    qcow2_cache_entry_mark_dirty(s->l2_table_cache, l2_table);
+    qcow2_cache_entry_mark_dirty(bs, s->l2_table_cache, l2_table);
 
     assert(l2_index + m->nb_clusters <= s->l2_size);
     for (i = 0; i < m->nb_clusters; i++) {
@@ -1470,7 +1470,7 @@ 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);
+        qcow2_cache_entry_mark_dirty(bs, s->l2_table_cache, l2_table);
         if (!full_discard && s->qcow_version >= 3) {
             l2_table[l2_index + i] = cpu_to_be64(QCOW_OFLAG_ZERO);
         } else {
@@ -1558,7 +1558,7 @@ static int zero_single_l2(BlockDriverState *bs, uint64_t offset,
         old_offset = be64_to_cpu(l2_table[l2_index + i]);
 
         /* Update L2 entries */
-        qcow2_cache_entry_mark_dirty(s->l2_table_cache, l2_table);
+        qcow2_cache_entry_mark_dirty(bs, s->l2_table_cache, l2_table);
         if (old_offset & QCOW_OFLAG_COMPRESSED) {
             l2_table[l2_index + i] = cpu_to_be64(QCOW_OFLAG_ZERO);
             qcow2_free_any_clusters(bs, old_offset, 1, QCOW2_DISCARD_REQUEST);
@@ -1760,7 +1760,7 @@ static int expand_zero_clusters_in_l1(BlockDriverState *bs, uint64_t *l1_table,
 
         if (is_active_l1) {
             if (l2_dirty) {
-                qcow2_cache_entry_mark_dirty(s->l2_table_cache, l2_table);
+                qcow2_cache_entry_mark_dirty(bs, s->l2_table_cache, l2_table);
                 qcow2_cache_depends_on_flush(s->l2_table_cache);
             }
             ret = qcow2_cache_put(bs, s->l2_table_cache, (void **)&l2_table);