* [Qemu-devel] [PATCH V13 4/6] rename qcow2-cache.c to block-cache.c
2012-10-18 9:51 [Qemu-devel] [PATCH V13 0/6] add-cow file format Dong Xu Wang
` (2 preceding siblings ...)
2012-10-18 9:51 ` [Qemu-devel] [PATCH V13 3/6] qed_read_string to bdrv_read_string Dong Xu Wang
@ 2012-10-18 9:51 ` Dong Xu Wang
2012-10-22 8:22 ` Stefan Hajnoczi
2012-10-18 9:51 ` [Qemu-devel] [PATCH V13 5/6] add-cow file format core code Dong Xu Wang
2012-10-18 9:51 ` [Qemu-devel] [PATCH V13 6/6] qemu-iotests: add add-cow iotests support Dong Xu Wang
5 siblings, 1 reply; 12+ messages in thread
From: Dong Xu Wang @ 2012-10-18 9:51 UTC (permalink / raw)
To: qemu-devel; +Cc: kwolf, Dong Xu Wang
We will re-use qcow2-cache as block layer common cache code,
so change its name and made some changes, define a struct named
BlockTableType, pass BlockTableType and table size parameters to
block cache initialization function.
Signed-off-by: Dong Xu Wang <wdongxu@linux.vnet.ibm.com>
---
block/Makefile.objs | 3 +-
block/block-cache.c | 317 +++++++++++++++++++++++++++++++++++++++++++++++
block/block-cache.h | 76 +++++++++++
block/qcow2-cache.c | 323 ------------------------------------------------
block/qcow2-cluster.c | 54 +++++----
block/qcow2-refcount.c | 67 ++++++-----
block/qcow2.c | 21 ++--
block/qcow2.h | 24 +---
trace-events | 13 +-
9 files changed, 483 insertions(+), 415 deletions(-)
create mode 100644 block/block-cache.c
create mode 100644 block/block-cache.h
delete mode 100644 block/qcow2-cache.c
diff --git a/block/Makefile.objs b/block/Makefile.objs
index 554f429..f128b78 100644
--- a/block/Makefile.objs
+++ b/block/Makefile.objs
@@ -1,5 +1,6 @@
block-obj-y += raw.o cow.o qcow.o vdi.o vmdk.o cloop.o dmg.o bochs.o vpc.o vvfat.o
-block-obj-y += qcow2.o qcow2-refcount.o qcow2-cluster.o qcow2-snapshot.o qcow2-cache.o
+block-obj-y += qcow2.o qcow2-refcount.o qcow2-cluster.o qcow2-snapshot.o
+block-obj-y += block-cache.o
block-obj-y += qed.o qed-gencb.o qed-l2-cache.o qed-table.o qed-cluster.o
block-obj-y += qed-check.o
block-obj-y += parallels.o nbd.o blkdebug.o sheepdog.o blkverify.o
diff --git a/block/block-cache.c b/block/block-cache.c
new file mode 100644
index 0000000..bf5c57c
--- /dev/null
+++ b/block/block-cache.c
@@ -0,0 +1,317 @@
+/*
+ * QEMU Block Layer Cache
+ *
+ * Copyright IBM, Corp. 2012
+ *
+ * Authors:
+ * Dong Xu Wang <wdongxu@linux.vnet.ibm.com>
+ *
+ * This file is based on qcow2-cache.c, see its copyrights below:
+ *
+ * L2/refcount table cache for the QCOW2 format
+ *
+ * Copyright (c) 2010 Kevin Wolf <kwolf@redhat.com>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+
+#include "block_int.h"
+#include "qemu-common.h"
+#include "trace.h"
+#include "block-cache.h"
+
+BlockCache *block_cache_create(BlockDriverState *bs, int num_tables,
+ size_t cluster_size, BlockTableType type)
+{
+ BlockCache *c;
+ int i;
+
+ c = g_malloc0(sizeof(*c));
+ c->size = num_tables;
+ c->entries = g_malloc0(sizeof(*c->entries) * num_tables);
+ c->table_type = type;
+ c->cluster_size = cluster_size;
+
+ for (i = 0; i < c->size; i++) {
+ c->entries[i].table = qemu_blockalign(bs, cluster_size);
+ }
+
+ return c;
+}
+
+int block_cache_destroy(BlockDriverState *bs, BlockCache *c)
+{
+ int i;
+
+ for (i = 0; i < c->size; i++) {
+ assert(c->entries[i].ref == 0);
+ qemu_vfree(c->entries[i].table);
+ }
+
+ g_free(c->entries);
+ g_free(c);
+
+ return 0;
+}
+
+static int block_cache_flush_dependency(BlockDriverState *bs, BlockCache *c)
+{
+ int ret;
+
+ ret = block_cache_flush(bs, c->depends);
+ if (ret < 0) {
+ return ret;
+ }
+
+ c->depends = NULL;
+ c->depends_on_flush = false;
+
+ return 0;
+}
+
+static int block_cache_entry_flush(BlockDriverState *bs, BlockCache *c, int i)
+{
+ int ret = 0;
+
+ if (!c->entries[i].dirty || !c->entries[i].offset) {
+ return 0;
+ }
+
+ trace_block_cache_entry_flush(qemu_coroutine_self(), c->table_type, i);
+
+ if (c->depends) {
+ ret = block_cache_flush_dependency(bs, c);
+ } else if (c->depends_on_flush) {
+ ret = bdrv_flush(bs->file);
+ if (ret >= 0) {
+ c->depends_on_flush = false;
+ }
+ }
+
+ if (ret < 0) {
+ return ret;
+ }
+
+ if (c->table_type == BLOCK_TABLE_REF) {
+ BLKDBG_EVENT(bs->file, BLKDBG_REFBLOCK_UPDATE_PART);
+ } else if (c->table_type == BLOCK_TABLE_L2) {
+ BLKDBG_EVENT(bs->file, BLKDBG_L2_UPDATE);
+ }
+
+ ret = bdrv_pwrite(bs->file, c->entries[i].offset,
+ c->entries[i].table, c->cluster_size);
+ if (ret < 0) {
+ return ret;
+ }
+
+ c->entries[i].dirty = false;
+
+ return 0;
+}
+
+int block_cache_flush(BlockDriverState *bs, BlockCache *c)
+{
+ int result = 0;
+ int ret;
+ int i;
+
+ trace_block_cache_flush(qemu_coroutine_self(), c->table_type);
+
+ for (i = 0; i < c->size; i++) {
+ ret = block_cache_entry_flush(bs, c, i);
+ if (ret < 0 && result != -ENOSPC) {
+ result = ret;
+ }
+ }
+
+ if (result == 0) {
+ ret = bdrv_flush(bs->file);
+ if (ret < 0) {
+ result = ret;
+ }
+ }
+
+ return result;
+}
+
+int block_cache_set_dependency(BlockDriverState *bs,
+ BlockCache *c,
+ BlockCache *dependency)
+{
+ int ret;
+
+ if (dependency->depends) {
+ ret = block_cache_flush_dependency(bs, dependency);
+ if (ret < 0) {
+ return ret;
+ }
+ }
+
+ if (c->depends && (c->depends != dependency)) {
+ ret = block_cache_flush_dependency(bs, c);
+ if (ret < 0) {
+ return ret;
+ }
+ }
+
+ c->depends = dependency;
+ return 0;
+}
+
+void block_cache_depends_on_flush(BlockCache *c)
+{
+ c->depends_on_flush = true;
+}
+
+static int block_cache_find_entry_to_replace(BlockCache *c)
+{
+ int i;
+ int min_count = INT_MAX;
+ int min_index = -1;
+
+
+ for (i = 0; i < c->size; i++) {
+ if (c->entries[i].ref) {
+ continue;
+ }
+
+ if (c->entries[i].cache_hits < min_count) {
+ min_index = i;
+ min_count = c->entries[i].cache_hits;
+ }
+
+ /* Give newer hits priority */
+ /* TODO Check how to optimize the replacement strategy */
+ c->entries[i].cache_hits /= 2;
+ }
+
+ if (min_index == -1) {
+ /* This can't happen in current synchronous code, but leave the check
+ * here as a reminder for whoever starts using AIO with the cache */
+ abort();
+ }
+ return min_index;
+}
+
+static int block_cache_do_get(BlockDriverState *bs, BlockCache *c,
+ uint64_t offset, void **table,
+ bool read_from_disk)
+{
+ int i;
+ int ret;
+
+ trace_block_cache_get(qemu_coroutine_self(), c->table_type,
+ offset, read_from_disk);
+
+ /* Check if the table is already cached */
+ for (i = 0; i < c->size; i++) {
+ if (c->entries[i].offset == offset) {
+ goto found;
+ }
+ }
+
+ /* If not, write a table back and replace it */
+ i = block_cache_find_entry_to_replace(c);
+ trace_block_cache_get_replace_entry(qemu_coroutine_self(),
+ c->table_type, i);
+ if (i < 0) {
+ return i;
+ }
+
+ ret = block_cache_entry_flush(bs, c, i);
+ if (ret < 0) {
+ return ret;
+ }
+
+ trace_block_cache_get_read(qemu_coroutine_self(),
+ c->table_type, i);
+ c->entries[i].offset = 0;
+ if (read_from_disk) {
+ if (c->table_type == BLOCK_TABLE_L2) {
+ BLKDBG_EVENT(bs->file, BLKDBG_L2_LOAD);
+ }
+
+ ret = bdrv_pread(bs->file, offset, c->entries[i].table,
+ c->cluster_size);
+ if (ret < 0) {
+ return ret;
+ }
+ }
+
+ /* Give the table some hits for the start so that it won't be replaced
+ * immediately. The number 32 is completely arbitrary. */
+ c->entries[i].cache_hits = 32;
+ c->entries[i].offset = offset;
+
+ /* And return the right table */
+found:
+ c->entries[i].cache_hits++;
+ c->entries[i].ref++;
+ *table = c->entries[i].table;
+
+ trace_block_cache_get_done(qemu_coroutine_self(),
+ c->table_type, i);
+
+ return 0;
+}
+
+int block_cache_get(BlockDriverState *bs, BlockCache *c, uint64_t offset,
+ void **table)
+{
+ return block_cache_do_get(bs, c, offset, table, true);
+}
+
+int block_cache_get_empty(BlockDriverState *bs, BlockCache *c,
+ uint64_t offset, void **table)
+{
+ return block_cache_do_get(bs, c, offset, table, false);
+}
+
+int block_cache_put(BlockDriverState *bs, BlockCache *c, void **table)
+{
+ int i;
+
+ for (i = 0; i < c->size; i++) {
+ if (c->entries[i].table == *table) {
+ goto found;
+ }
+ }
+ return -ENOENT;
+
+found:
+ c->entries[i].ref--;
+ assert(c->entries[i].ref >= 0);
+ *table = NULL;
+ return 0;
+}
+
+void block_cache_entry_mark_dirty(BlockCache *c, void *table)
+{
+ int i;
+
+ for (i = 0; i < c->size; i++) {
+ if (c->entries[i].table == table) {
+ goto found;
+ }
+ }
+ abort();
+
+found:
+ c->entries[i].dirty = true;
+}
diff --git a/block/block-cache.h b/block/block-cache.h
new file mode 100644
index 0000000..4efa06e
--- /dev/null
+++ b/block/block-cache.h
@@ -0,0 +1,76 @@
+/*
+ * QEMU Block Layer Cache
+ *
+ * Copyright IBM, Corp. 2012
+ *
+ * Authors:
+ * Dong Xu Wang <wdongxu@linux.vnet.ibm.com>
+ *
+ * This file is based on qcow2-cache.c, see its copyrights below:
+ *
+ * L2/refcount table cache for the QCOW2 format
+ *
+ * Copyright (c) 2010 Kevin Wolf <kwolf@redhat.com>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+
+#ifndef BLOCK_CACHE_H
+#define BLOCK_CACHE_H
+
+typedef enum {
+ BLOCK_TABLE_REF,
+ BLOCK_TABLE_L2,
+} BlockTableType;
+
+typedef struct BlockCachedTable {
+ void *table;
+ int64_t offset;
+ bool dirty;
+ int cache_hits;
+ int ref;
+} BlockCachedTable;
+
+struct BlockCache {
+ BlockCachedTable *entries;
+ struct BlockCache *depends;
+ int size;
+ size_t cluster_size;
+ BlockTableType table_type;
+ bool depends_on_flush;
+};
+
+struct BlockCache;
+typedef struct BlockCache BlockCache;
+
+BlockCache *block_cache_create(BlockDriverState *bs, int num_tables,
+ size_t cluster_size, BlockTableType type);
+int block_cache_destroy(BlockDriverState *bs, BlockCache *c);
+int block_cache_flush(BlockDriverState *bs, BlockCache *c);
+int block_cache_set_dependency(BlockDriverState *bs,
+ BlockCache *c,
+ BlockCache *dependency);
+void block_cache_depends_on_flush(BlockCache *c);
+int block_cache_get(BlockDriverState *bs, BlockCache *c, uint64_t offset,
+ void **table);
+int block_cache_get_empty(BlockDriverState *bs, BlockCache *c,
+ uint64_t offset, void **table);
+int block_cache_put(BlockDriverState *bs, BlockCache *c, void **table);
+void block_cache_entry_mark_dirty(BlockCache *c, void *table);
+#endif /* BLOCK_CACHE_H */
diff --git a/block/qcow2-cache.c b/block/qcow2-cache.c
deleted file mode 100644
index 2d4322a..0000000
--- a/block/qcow2-cache.c
+++ /dev/null
@@ -1,323 +0,0 @@
-/*
- * L2/refcount table cache for the QCOW2 format
- *
- * Copyright (c) 2010 Kevin Wolf <kwolf@redhat.com>
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to deal
- * in the Software without restriction, including without limitation the rights
- * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
- * copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
- * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
- * THE SOFTWARE.
- */
-
-#include "block_int.h"
-#include "qemu-common.h"
-#include "qcow2.h"
-#include "trace.h"
-
-typedef struct Qcow2CachedTable {
- void* table;
- int64_t offset;
- bool dirty;
- int cache_hits;
- int ref;
-} Qcow2CachedTable;
-
-struct Qcow2Cache {
- Qcow2CachedTable* entries;
- struct Qcow2Cache* depends;
- int size;
- bool depends_on_flush;
-};
-
-Qcow2Cache *qcow2_cache_create(BlockDriverState *bs, int num_tables)
-{
- BDRVQcowState *s = bs->opaque;
- Qcow2Cache *c;
- int i;
-
- c = g_malloc0(sizeof(*c));
- c->size = num_tables;
- c->entries = g_malloc0(sizeof(*c->entries) * num_tables);
-
- for (i = 0; i < c->size; i++) {
- c->entries[i].table = qemu_blockalign(bs, s->cluster_size);
- }
-
- return c;
-}
-
-int qcow2_cache_destroy(BlockDriverState* bs, Qcow2Cache *c)
-{
- int i;
-
- for (i = 0; i < c->size; i++) {
- assert(c->entries[i].ref == 0);
- qemu_vfree(c->entries[i].table);
- }
-
- g_free(c->entries);
- g_free(c);
-
- return 0;
-}
-
-static int qcow2_cache_flush_dependency(BlockDriverState *bs, Qcow2Cache *c)
-{
- int ret;
-
- ret = qcow2_cache_flush(bs, c->depends);
- if (ret < 0) {
- return ret;
- }
-
- c->depends = NULL;
- c->depends_on_flush = false;
-
- return 0;
-}
-
-static int qcow2_cache_entry_flush(BlockDriverState *bs, Qcow2Cache *c, int i)
-{
- BDRVQcowState *s = bs->opaque;
- int ret = 0;
-
- if (!c->entries[i].dirty || !c->entries[i].offset) {
- return 0;
- }
-
- trace_qcow2_cache_entry_flush(qemu_coroutine_self(),
- c == s->l2_table_cache, i);
-
- if (c->depends) {
- ret = qcow2_cache_flush_dependency(bs, c);
- } else if (c->depends_on_flush) {
- ret = bdrv_flush(bs->file);
- if (ret >= 0) {
- c->depends_on_flush = false;
- }
- }
-
- if (ret < 0) {
- return ret;
- }
-
- if (c == s->refcount_block_cache) {
- BLKDBG_EVENT(bs->file, BLKDBG_REFBLOCK_UPDATE_PART);
- } else if (c == s->l2_table_cache) {
- BLKDBG_EVENT(bs->file, BLKDBG_L2_UPDATE);
- }
-
- ret = bdrv_pwrite(bs->file, c->entries[i].offset, c->entries[i].table,
- s->cluster_size);
- if (ret < 0) {
- return ret;
- }
-
- c->entries[i].dirty = false;
-
- return 0;
-}
-
-int qcow2_cache_flush(BlockDriverState *bs, Qcow2Cache *c)
-{
- BDRVQcowState *s = bs->opaque;
- int result = 0;
- int ret;
- int i;
-
- trace_qcow2_cache_flush(qemu_coroutine_self(), c == s->l2_table_cache);
-
- for (i = 0; i < c->size; i++) {
- ret = qcow2_cache_entry_flush(bs, c, i);
- if (ret < 0 && result != -ENOSPC) {
- result = ret;
- }
- }
-
- if (result == 0) {
- ret = bdrv_flush(bs->file);
- if (ret < 0) {
- result = ret;
- }
- }
-
- return result;
-}
-
-int qcow2_cache_set_dependency(BlockDriverState *bs, Qcow2Cache *c,
- Qcow2Cache *dependency)
-{
- int ret;
-
- if (dependency->depends) {
- ret = qcow2_cache_flush_dependency(bs, dependency);
- if (ret < 0) {
- return ret;
- }
- }
-
- if (c->depends && (c->depends != dependency)) {
- ret = qcow2_cache_flush_dependency(bs, c);
- if (ret < 0) {
- return ret;
- }
- }
-
- c->depends = dependency;
- return 0;
-}
-
-void qcow2_cache_depends_on_flush(Qcow2Cache *c)
-{
- c->depends_on_flush = true;
-}
-
-static int qcow2_cache_find_entry_to_replace(Qcow2Cache *c)
-{
- int i;
- int min_count = INT_MAX;
- int min_index = -1;
-
-
- for (i = 0; i < c->size; i++) {
- if (c->entries[i].ref) {
- continue;
- }
-
- if (c->entries[i].cache_hits < min_count) {
- min_index = i;
- min_count = c->entries[i].cache_hits;
- }
-
- /* Give newer hits priority */
- /* TODO Check how to optimize the replacement strategy */
- c->entries[i].cache_hits /= 2;
- }
-
- if (min_index == -1) {
- /* This can't happen in current synchronous code, but leave the check
- * here as a reminder for whoever starts using AIO with the cache */
- abort();
- }
- return min_index;
-}
-
-static int qcow2_cache_do_get(BlockDriverState *bs, Qcow2Cache *c,
- uint64_t offset, void **table, bool read_from_disk)
-{
- BDRVQcowState *s = bs->opaque;
- int i;
- int ret;
-
- trace_qcow2_cache_get(qemu_coroutine_self(), c == s->l2_table_cache,
- offset, read_from_disk);
-
- /* Check if the table is already cached */
- for (i = 0; i < c->size; i++) {
- if (c->entries[i].offset == offset) {
- goto found;
- }
- }
-
- /* If not, write a table back and replace it */
- i = qcow2_cache_find_entry_to_replace(c);
- trace_qcow2_cache_get_replace_entry(qemu_coroutine_self(),
- c == s->l2_table_cache, i);
- if (i < 0) {
- return i;
- }
-
- ret = qcow2_cache_entry_flush(bs, c, i);
- if (ret < 0) {
- return ret;
- }
-
- trace_qcow2_cache_get_read(qemu_coroutine_self(),
- c == s->l2_table_cache, i);
- c->entries[i].offset = 0;
- if (read_from_disk) {
- if (c == s->l2_table_cache) {
- BLKDBG_EVENT(bs->file, BLKDBG_L2_LOAD);
- }
-
- ret = bdrv_pread(bs->file, offset, c->entries[i].table, s->cluster_size);
- if (ret < 0) {
- return ret;
- }
- }
-
- /* Give the table some hits for the start so that it won't be replaced
- * immediately. The number 32 is completely arbitrary. */
- c->entries[i].cache_hits = 32;
- c->entries[i].offset = offset;
-
- /* And return the right table */
-found:
- c->entries[i].cache_hits++;
- c->entries[i].ref++;
- *table = c->entries[i].table;
-
- trace_qcow2_cache_get_done(qemu_coroutine_self(),
- c == s->l2_table_cache, i);
-
- return 0;
-}
-
-int qcow2_cache_get(BlockDriverState *bs, Qcow2Cache *c, uint64_t offset,
- void **table)
-{
- return qcow2_cache_do_get(bs, c, offset, table, true);
-}
-
-int qcow2_cache_get_empty(BlockDriverState *bs, Qcow2Cache *c, uint64_t offset,
- void **table)
-{
- return qcow2_cache_do_get(bs, c, offset, table, false);
-}
-
-int qcow2_cache_put(BlockDriverState *bs, Qcow2Cache *c, void **table)
-{
- int i;
-
- for (i = 0; i < c->size; i++) {
- if (c->entries[i].table == *table) {
- goto found;
- }
- }
- return -ENOENT;
-
-found:
- c->entries[i].ref--;
- *table = NULL;
-
- assert(c->entries[i].ref >= 0);
- return 0;
-}
-
-void qcow2_cache_entry_mark_dirty(Qcow2Cache *c, void *table)
-{
- int i;
-
- for (i = 0; i < c->size; i++) {
- if (c->entries[i].table == table) {
- goto found;
- }
- }
- abort();
-
-found:
- c->entries[i].dirty = true;
-}
diff --git a/block/qcow2-cluster.c b/block/qcow2-cluster.c
index e179211..171131f 100644
--- a/block/qcow2-cluster.c
+++ b/block/qcow2-cluster.c
@@ -28,6 +28,7 @@
#include "block_int.h"
#include "block/qcow2.h"
#include "trace.h"
+#include "block-cache.h"
int qcow2_grow_l1_table(BlockDriverState *bs, int min_size, bool exact_size)
{
@@ -69,7 +70,7 @@ int qcow2_grow_l1_table(BlockDriverState *bs, int min_size, bool exact_size)
return new_l1_table_offset;
}
- ret = qcow2_cache_flush(bs, s->refcount_block_cache);
+ ret = block_cache_flush(bs, s->refcount_block_cache);
if (ret < 0) {
goto fail;
}
@@ -119,7 +120,8 @@ static int l2_load(BlockDriverState *bs, uint64_t l2_offset,
BDRVQcowState *s = bs->opaque;
int ret;
- ret = qcow2_cache_get(bs, s->l2_table_cache, l2_offset, (void**) l2_table);
+ ret = block_cache_get(bs, s->l2_table_cache, l2_offset,
+ (void **) l2_table);
return ret;
}
@@ -180,7 +182,7 @@ static int l2_allocate(BlockDriverState *bs, int l1_index, uint64_t **table)
return l2_offset;
}
- ret = qcow2_cache_flush(bs, s->refcount_block_cache);
+ ret = block_cache_flush(bs, s->refcount_block_cache);
if (ret < 0) {
goto fail;
}
@@ -188,7 +190,8 @@ static int l2_allocate(BlockDriverState *bs, int l1_index, uint64_t **table)
/* allocate a new entry in the l2 cache */
trace_qcow2_l2_allocate_get_empty(bs, l1_index);
- ret = qcow2_cache_get_empty(bs, s->l2_table_cache, l2_offset, (void**) table);
+ ret = block_cache_get_empty(bs, s->l2_table_cache, l2_offset,
+ (void **) table);
if (ret < 0) {
return ret;
}
@@ -203,16 +206,16 @@ static int l2_allocate(BlockDriverState *bs, int l1_index, uint64_t **table)
/* if there was an old l2 table, read it from the disk */
BLKDBG_EVENT(bs->file, BLKDBG_L2_ALLOC_COW_READ);
- ret = qcow2_cache_get(bs, s->l2_table_cache,
- old_l2_offset & L1E_OFFSET_MASK,
- (void**) &old_table);
+ ret = block_cache_get(bs, s->l2_table_cache,
+ old_l2_offset & L1E_OFFSET_MASK,
+ (void **) &old_table);
if (ret < 0) {
goto fail;
}
memcpy(l2_table, old_table, s->cluster_size);
- ret = qcow2_cache_put(bs, s->l2_table_cache, (void**) &old_table);
+ ret = block_cache_put(bs, s->l2_table_cache, (void **) &old_table);
if (ret < 0) {
goto fail;
}
@@ -222,8 +225,8 @@ 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);
- ret = qcow2_cache_flush(bs, s->l2_table_cache);
+ block_cache_entry_mark_dirty(s->l2_table_cache, l2_table);
+ ret = block_cache_flush(bs, s->l2_table_cache);
if (ret < 0) {
goto fail;
}
@@ -242,7 +245,7 @@ static int l2_allocate(BlockDriverState *bs, int l1_index, uint64_t **table)
fail:
trace_qcow2_l2_allocate_done(bs, l1_index, ret);
- qcow2_cache_put(bs, s->l2_table_cache, (void**) table);
+ block_cache_put(bs, s->l2_table_cache, (void **) table);
s->l1_table[l1_index] = old_l2_offset;
return ret;
}
@@ -475,7 +478,7 @@ int qcow2_get_cluster_offset(BlockDriverState *bs, uint64_t offset,
abort();
}
- qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
+ block_cache_put(bs, s->l2_table_cache, (void **) &l2_table);
nb_available = (c * s->cluster_sectors);
@@ -584,13 +587,13 @@ uint64_t qcow2_alloc_compressed_cluster_offset(BlockDriverState *bs,
* allocated. */
cluster_offset = be64_to_cpu(l2_table[l2_index]);
if (cluster_offset & L2E_OFFSET_MASK) {
- qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
+ block_cache_put(bs, s->l2_table_cache, (void **) &l2_table);
return 0;
}
cluster_offset = qcow2_alloc_bytes(bs, compressed_size);
if (cluster_offset < 0) {
- qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
+ block_cache_put(bs, s->l2_table_cache, (void **) &l2_table);
return 0;
}
@@ -605,9 +608,9 @@ 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);
+ block_cache_entry_mark_dirty(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);
+ ret = block_cache_put(bs, s->l2_table_cache, (void **) &l2_table);
if (ret < 0) {
return 0;
}
@@ -659,18 +662,19 @@ int qcow2_alloc_cluster_link_l2(BlockDriverState *bs, QCowL2Meta *m)
* handled.
*/
if (cow) {
- qcow2_cache_depends_on_flush(s->l2_table_cache);
+ block_cache_depends_on_flush(s->l2_table_cache);
}
+
if (qcow2_need_accurate_refcounts(s)) {
- qcow2_cache_set_dependency(bs, s->l2_table_cache,
+ block_cache_set_dependency(bs, s->l2_table_cache,
s->refcount_block_cache);
}
ret = get_cluster_table(bs, m->offset, &l2_table, &l2_index);
if (ret < 0) {
goto err;
}
- qcow2_cache_entry_mark_dirty(s->l2_table_cache, l2_table);
+ block_cache_entry_mark_dirty(s->l2_table_cache, l2_table);
for (i = 0; i < m->nb_clusters; i++) {
/* if two concurrent writes happen to the same unallocated cluster
@@ -687,7 +691,7 @@ int qcow2_alloc_cluster_link_l2(BlockDriverState *bs, QCowL2Meta *m)
}
- ret = qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
+ ret = block_cache_put(bs, s->l2_table_cache, (void **) &l2_table);
if (ret < 0) {
goto err;
}
@@ -913,7 +917,7 @@ again:
* request to complete. If we still had the reference, we could use up the
* whole cache with sleeping requests.
*/
- ret = qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
+ ret = block_cache_put(bs, s->l2_table_cache, (void **) &l2_table);
if (ret < 0) {
return ret;
}
@@ -1077,14 +1081,14 @@ 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);
+ block_cache_entry_mark_dirty(s->l2_table_cache, l2_table);
l2_table[l2_index + i] = cpu_to_be64(0);
/* Then decrease the refcount */
qcow2_free_any_clusters(bs, old_offset, 1);
}
- ret = qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
+ ret = block_cache_put(bs, s->l2_table_cache, (void **) &l2_table);
if (ret < 0) {
return ret;
}
@@ -1154,7 +1158,7 @@ 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);
+ block_cache_entry_mark_dirty(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);
@@ -1163,7 +1167,7 @@ static int zero_single_l2(BlockDriverState *bs, uint64_t offset,
}
}
- ret = qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
+ ret = block_cache_put(bs, s->l2_table_cache, (void **) &l2_table);
if (ret < 0) {
return ret;
}
diff --git a/block/qcow2-refcount.c b/block/qcow2-refcount.c
index 5e3f915..a57c74b 100644
--- a/block/qcow2-refcount.c
+++ b/block/qcow2-refcount.c
@@ -25,6 +25,7 @@
#include "qemu-common.h"
#include "block_int.h"
#include "block/qcow2.h"
+#include "block-cache.h"
static int64_t alloc_clusters_noref(BlockDriverState *bs, int64_t size);
static int QEMU_WARN_UNUSED_RESULT update_refcount(BlockDriverState *bs,
@@ -71,8 +72,8 @@ static int load_refcount_block(BlockDriverState *bs,
int ret;
BLKDBG_EVENT(bs->file, BLKDBG_REFBLOCK_LOAD);
- ret = qcow2_cache_get(bs, s->refcount_block_cache, refcount_block_offset,
- refcount_block);
+ ret = block_cache_get(bs, s->refcount_block_cache, refcount_block_offset,
+ refcount_block);
return ret;
}
@@ -98,8 +99,8 @@ static int get_refcount(BlockDriverState *bs, int64_t cluster_index)
if (!refcount_block_offset)
return 0;
- ret = qcow2_cache_get(bs, s->refcount_block_cache, refcount_block_offset,
- (void**) &refcount_block);
+ ret = block_cache_get(bs, s->refcount_block_cache, refcount_block_offset,
+ (void **) &refcount_block);
if (ret < 0) {
return ret;
}
@@ -108,8 +109,8 @@ static int get_refcount(BlockDriverState *bs, int64_t cluster_index)
((1 << (s->cluster_bits - REFCOUNT_SHIFT)) - 1);
refcount = be16_to_cpu(refcount_block[block_index]);
- ret = qcow2_cache_put(bs, s->refcount_block_cache,
- (void**) &refcount_block);
+ ret = block_cache_put(bs, s->refcount_block_cache,
+ (void **) &refcount_block);
if (ret < 0) {
return ret;
}
@@ -201,7 +202,7 @@ static int alloc_refcount_block(BlockDriverState *bs,
*refcount_block = NULL;
/* We write to the refcount table, so we might depend on L2 tables */
- qcow2_cache_flush(bs, s->l2_table_cache);
+ block_cache_flush(bs, s->l2_table_cache);
/* Allocate the refcount block itself and mark it as used */
int64_t new_block = alloc_clusters_noref(bs, s->cluster_size);
@@ -217,8 +218,8 @@ static int alloc_refcount_block(BlockDriverState *bs,
if (in_same_refcount_block(s, new_block, cluster_index << s->cluster_bits)) {
/* Zero the new refcount block before updating it */
- ret = qcow2_cache_get_empty(bs, s->refcount_block_cache, new_block,
- (void**) refcount_block);
+ ret = block_cache_get_empty(bs, s->refcount_block_cache, new_block,
+ (void **) refcount_block);
if (ret < 0) {
goto fail_block;
}
@@ -241,8 +242,8 @@ static int alloc_refcount_block(BlockDriverState *bs,
/* Initialize the new refcount block only after updating its refcount,
* update_refcount uses the refcount cache itself */
- ret = qcow2_cache_get_empty(bs, s->refcount_block_cache, new_block,
- (void**) refcount_block);
+ ret = block_cache_get_empty(bs, s->refcount_block_cache, new_block,
+ (void **) refcount_block);
if (ret < 0) {
goto fail_block;
}
@@ -252,8 +253,8 @@ static int alloc_refcount_block(BlockDriverState *bs,
/* Now the new refcount block needs to be written to disk */
BLKDBG_EVENT(bs->file, BLKDBG_REFBLOCK_ALLOC_WRITE);
- qcow2_cache_entry_mark_dirty(s->refcount_block_cache, *refcount_block);
- ret = qcow2_cache_flush(bs, s->refcount_block_cache);
+ block_cache_entry_mark_dirty(s->refcount_block_cache, *refcount_block);
+ ret = block_cache_flush(bs, s->refcount_block_cache);
if (ret < 0) {
goto fail_block;
}
@@ -273,7 +274,8 @@ static int alloc_refcount_block(BlockDriverState *bs,
return 0;
}
- ret = qcow2_cache_put(bs, s->refcount_block_cache, (void**) refcount_block);
+ ret = block_cache_put(bs, s->refcount_block_cache,
+ (void **) refcount_block);
if (ret < 0) {
goto fail_block;
}
@@ -406,7 +408,8 @@ fail_table:
g_free(new_table);
fail_block:
if (*refcount_block != NULL) {
- qcow2_cache_put(bs, s->refcount_block_cache, (void**) refcount_block);
+ block_cache_put(bs, s->refcount_block_cache,
+ (void **) refcount_block);
}
return ret;
}
@@ -432,8 +435,8 @@ static int QEMU_WARN_UNUSED_RESULT update_refcount(BlockDriverState *bs,
}
if (addend < 0) {
- qcow2_cache_set_dependency(bs, s->refcount_block_cache,
- s->l2_table_cache);
+ block_cache_set_dependency(bs, s->refcount_block_cache,
+ s->l2_table_cache);
}
start = offset & ~(s->cluster_size - 1);
@@ -449,8 +452,8 @@ static int QEMU_WARN_UNUSED_RESULT update_refcount(BlockDriverState *bs,
/* Load the refcount block and allocate it if needed */
if (table_index != old_table_index) {
if (refcount_block) {
- ret = qcow2_cache_put(bs, s->refcount_block_cache,
- (void**) &refcount_block);
+ ret = block_cache_put(bs, s->refcount_block_cache,
+ (void **) &refcount_block);
if (ret < 0) {
goto fail;
}
@@ -463,7 +466,7 @@ static int QEMU_WARN_UNUSED_RESULT update_refcount(BlockDriverState *bs,
}
old_table_index = table_index;
- qcow2_cache_entry_mark_dirty(s->refcount_block_cache, refcount_block);
+ block_cache_entry_mark_dirty(s->refcount_block_cache, refcount_block);
/* we can update the count and save it */
block_index = cluster_index &
@@ -486,8 +489,8 @@ fail:
/* Write last changed block to disk */
if (refcount_block) {
int wret;
- wret = qcow2_cache_put(bs, s->refcount_block_cache,
- (void**) &refcount_block);
+ wret = block_cache_put(bs, s->refcount_block_cache,
+ (void **) &refcount_block);
if (wret < 0) {
return ret < 0 ? ret : wret;
}
@@ -763,8 +766,8 @@ int qcow2_update_snapshot_refcount(BlockDriverState *bs,
old_l2_offset = l2_offset;
l2_offset &= L1E_OFFSET_MASK;
- ret = qcow2_cache_get(bs, s->l2_table_cache, l2_offset,
- (void**) &l2_table);
+ ret = block_cache_get(bs, s->l2_table_cache, l2_offset,
+ (void **) &l2_table);
if (ret < 0) {
goto fail;
}
@@ -811,16 +814,18 @@ int qcow2_update_snapshot_refcount(BlockDriverState *bs,
}
if (offset != old_offset) {
if (addend > 0) {
- qcow2_cache_set_dependency(bs, s->l2_table_cache,
+ block_cache_set_dependency(bs, s->l2_table_cache,
s->refcount_block_cache);
}
l2_table[j] = cpu_to_be64(offset);
- qcow2_cache_entry_mark_dirty(s->l2_table_cache, l2_table);
+ block_cache_entry_mark_dirty(s->l2_table_cache,
+ l2_table);
}
}
}
- ret = qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
+ ret = block_cache_put(bs, s->l2_table_cache,
+ (void **) &l2_table);
if (ret < 0) {
goto fail;
}
@@ -847,7 +852,8 @@ int qcow2_update_snapshot_refcount(BlockDriverState *bs,
ret = 0;
fail:
if (l2_table) {
- qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
+ block_cache_put(bs, s->l2_table_cache,
+ (void **) &l2_table);
}
/* Update L1 only if it isn't deleted anyway (addend = -1) */
@@ -1130,8 +1136,9 @@ int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
0, s->cluster_size);
/* current L1 table */
- ret = check_refcounts_l1(bs, res, refcount_table, nb_clusters,
- s->l1_table_offset, s->l1_size, 1);
+ ret = check_refcounts_l1(bs, res, refcount_table,
+ nb_clusters, s->l1_table_offset,
+ s->l1_size, 1);
if (ret < 0) {
goto fail;
}
diff --git a/block/qcow2.c b/block/qcow2.c
index 025dce2..35950fb 100644
--- a/block/qcow2.c
+++ b/block/qcow2.c
@@ -430,8 +430,11 @@ static int qcow2_open(BlockDriverState *bs, int flags)
}
/* alloc L2 table/refcount block cache */
- s->l2_table_cache = qcow2_cache_create(bs, L2_CACHE_SIZE);
- s->refcount_block_cache = qcow2_cache_create(bs, REFCOUNT_CACHE_SIZE);
+ s->l2_table_cache = block_cache_create(bs, L2_CACHE_SIZE,
+ s->cluster_size, BLOCK_TABLE_L2);
+ s->refcount_block_cache = block_cache_create(bs, REFCOUNT_CACHE_SIZE,
+ s->cluster_size,
+ BLOCK_TABLE_REF);
s->cluster_cache = g_malloc(s->cluster_size);
/* one more sector for decompressed data alignment */
@@ -510,7 +513,7 @@ static int qcow2_open(BlockDriverState *bs, int flags)
qcow2_refcount_close(bs);
g_free(s->l1_table);
if (s->l2_table_cache) {
- qcow2_cache_destroy(bs, s->l2_table_cache);
+ block_cache_destroy(bs, s->l2_table_cache);
}
g_free(s->cluster_cache);
qemu_vfree(s->cluster_data);
@@ -878,13 +881,13 @@ static void qcow2_close(BlockDriverState *bs)
BDRVQcowState *s = bs->opaque;
g_free(s->l1_table);
- qcow2_cache_flush(bs, s->l2_table_cache);
- qcow2_cache_flush(bs, s->refcount_block_cache);
+ block_cache_flush(bs, s->l2_table_cache);
+ block_cache_flush(bs, s->refcount_block_cache);
qcow2_mark_clean(bs);
- qcow2_cache_destroy(bs, s->l2_table_cache);
- qcow2_cache_destroy(bs, s->refcount_block_cache);
+ block_cache_destroy(bs, s->l2_table_cache);
+ block_cache_destroy(bs, s->refcount_block_cache);
g_free(s->unknown_header_fields);
cleanup_unknown_header_ext(bs);
@@ -1553,14 +1556,14 @@ static coroutine_fn int qcow2_co_flush_to_os(BlockDriverState *bs)
int ret;
qemu_co_mutex_lock(&s->lock);
- ret = qcow2_cache_flush(bs, s->l2_table_cache);
+ ret = block_cache_flush(bs, s->l2_table_cache);
if (ret < 0) {
qemu_co_mutex_unlock(&s->lock);
return ret;
}
if (qcow2_need_accurate_refcounts(s)) {
- ret = qcow2_cache_flush(bs, s->refcount_block_cache);
+ ret = block_cache_flush(bs, s->refcount_block_cache);
if (ret < 0) {
qemu_co_mutex_unlock(&s->lock);
return ret;
diff --git a/block/qcow2.h b/block/qcow2.h
index b4eb654..cb6fd7a 100644
--- a/block/qcow2.h
+++ b/block/qcow2.h
@@ -27,6 +27,7 @@
#include "aes.h"
#include "qemu-coroutine.h"
+#include "block-cache.h"
//#define DEBUG_ALLOC
//#define DEBUG_ALLOC2
@@ -94,8 +95,6 @@ typedef struct QCowSnapshot {
uint64_t vm_clock_nsec;
} QCowSnapshot;
-struct Qcow2Cache;
-typedef struct Qcow2Cache Qcow2Cache;
typedef struct Qcow2UnknownHeaderExtension {
uint32_t magic;
@@ -146,8 +145,8 @@ typedef struct BDRVQcowState {
uint64_t l1_table_offset;
uint64_t *l1_table;
- Qcow2Cache* l2_table_cache;
- Qcow2Cache* refcount_block_cache;
+ BlockCache *l2_table_cache;
+ BlockCache *refcount_block_cache;
uint8_t *cluster_cache;
uint8_t *cluster_data;
@@ -316,21 +315,4 @@ int qcow2_snapshot_load_tmp(BlockDriverState *bs, const char *snapshot_name);
void qcow2_free_snapshots(BlockDriverState *bs);
int qcow2_read_snapshots(BlockDriverState *bs);
-
-/* qcow2-cache.c functions */
-Qcow2Cache *qcow2_cache_create(BlockDriverState *bs, int num_tables);
-int qcow2_cache_destroy(BlockDriverState* bs, Qcow2Cache *c);
-
-void qcow2_cache_entry_mark_dirty(Qcow2Cache *c, void *table);
-int qcow2_cache_flush(BlockDriverState *bs, Qcow2Cache *c);
-int qcow2_cache_set_dependency(BlockDriverState *bs, Qcow2Cache *c,
- Qcow2Cache *dependency);
-void qcow2_cache_depends_on_flush(Qcow2Cache *c);
-
-int qcow2_cache_get(BlockDriverState *bs, Qcow2Cache *c, uint64_t offset,
- void **table);
-int qcow2_cache_get_empty(BlockDriverState *bs, Qcow2Cache *c, uint64_t offset,
- void **table);
-int qcow2_cache_put(BlockDriverState *bs, Qcow2Cache *c, void **table);
-
#endif
diff --git a/trace-events b/trace-events
index 42b66f1..df1f12f 100644
--- a/trace-events
+++ b/trace-events
@@ -454,12 +454,13 @@ qcow2_l2_allocate_write_l2(void *bs, int l1_index) "bs %p l1_index %d"
qcow2_l2_allocate_write_l1(void *bs, int l1_index) "bs %p l1_index %d"
qcow2_l2_allocate_done(void *bs, int l1_index, int ret) "bs %p l1_index %d ret %d"
-qcow2_cache_get(void *co, int c, uint64_t offset, bool read_from_disk) "co %p is_l2_cache %d offset %" PRIx64 " read_from_disk %d"
-qcow2_cache_get_replace_entry(void *co, int c, int i) "co %p is_l2_cache %d index %d"
-qcow2_cache_get_read(void *co, int c, int i) "co %p is_l2_cache %d index %d"
-qcow2_cache_get_done(void *co, int c, int i) "co %p is_l2_cache %d index %d"
-qcow2_cache_flush(void *co, int c) "co %p is_l2_cache %d"
-qcow2_cache_entry_flush(void *co, int c, int i) "co %p is_l2_cache %d index %d"
+# block/block-cache.c
+block_cache_get(void *co, int c, uint64_t offset, bool read_from_disk) "co %p is_l2_cache %d offset %" PRIx64 " read_from_disk %d"
+block_cache_get_replace_entry(void *co, int c, int i) "co %p is_l2_cache %d index %d"
+block_cache_get_read(void *co, int c, int i) "co %p is_l2_cache %d index %d"
+block_cache_get_done(void *co, int c, int i) "co %p is_l2_cache %d index %d"
+block_cache_flush(void *co, int c) "co %p is_l2_cache %d"
+block_cache_entry_flush(void *co, int c, int i) "co %p is_l2_cache %d index %d"
# block/qed-l2-cache.c
qed_alloc_l2_cache_entry(void *l2_cache, void *entry) "l2_cache %p entry %p"
--
1.7.1
^ permalink raw reply related [flat|nested] 12+ messages in thread* [Qemu-devel] [PATCH V13 5/6] add-cow file format core code.
2012-10-18 9:51 [Qemu-devel] [PATCH V13 0/6] add-cow file format Dong Xu Wang
` (3 preceding siblings ...)
2012-10-18 9:51 ` [Qemu-devel] [PATCH V13 4/6] rename qcow2-cache.c to block-cache.c Dong Xu Wang
@ 2012-10-18 9:51 ` Dong Xu Wang
2012-10-22 9:29 ` Stefan Hajnoczi
2012-10-18 9:51 ` [Qemu-devel] [PATCH V13 6/6] qemu-iotests: add add-cow iotests support Dong Xu Wang
5 siblings, 1 reply; 12+ messages in thread
From: Dong Xu Wang @ 2012-10-18 9:51 UTC (permalink / raw)
To: qemu-devel; +Cc: kwolf, Dong Xu Wang
add-cow file format core code. It use block-cache.c as cache code.
It lacks of snapshot_blkdev support.
Signed-off-by: Dong Xu Wang <wdongxu@linux.vnet.ibm.com>
---
block/Makefile.objs | 1 +
block/add-cow.c | 693 +++++++++++++++++++++++++++++++++++++++++++++++++++
block/add-cow.h | 85 +++++++
block/block-cache.c | 4 +
block/block-cache.h | 1 +
block_int.h | 2 +
6 files changed, 786 insertions(+), 0 deletions(-)
create mode 100644 block/add-cow.c
create mode 100644 block/add-cow.h
diff --git a/block/Makefile.objs b/block/Makefile.objs
index f128b78..ed9222d 100644
--- a/block/Makefile.objs
+++ b/block/Makefile.objs
@@ -1,5 +1,6 @@
block-obj-y += raw.o cow.o qcow.o vdi.o vmdk.o cloop.o dmg.o bochs.o vpc.o vvfat.o
block-obj-y += qcow2.o qcow2-refcount.o qcow2-cluster.o qcow2-snapshot.o
+block-obj-y += add-cow.o
block-obj-y += block-cache.o
block-obj-y += qed.o qed-gencb.o qed-l2-cache.o qed-table.o qed-cluster.o
block-obj-y += qed-check.o
diff --git a/block/add-cow.c b/block/add-cow.c
new file mode 100644
index 0000000..15c86ab
--- /dev/null
+++ b/block/add-cow.c
@@ -0,0 +1,693 @@
+/*
+ * QEMU ADD-COW Disk Format
+ *
+ * Copyright IBM, Corp. 2012
+ *
+ * Authors:
+ * Dong Xu Wang <wdongxu@linux.vnet.ibm.com>
+ *
+ * This work is licensed under the terms of the GNU LGPL, version 2 or later.
+ * See the COPYING.LIB file in the top-level directory.
+ *
+ */
+
+#include "qemu-common.h"
+#include "block_int.h"
+#include "module.h"
+#include "add-cow.h"
+
+static void add_cow_header_le_to_cpu(const AddCowHeader *le, AddCowHeader *cpu)
+{
+ cpu->magic = le64_to_cpu(le->magic);
+ cpu->version = le32_to_cpu(le->version);
+
+ cpu->backing_filename_offset = le32_to_cpu(le->backing_filename_offset);
+ cpu->backing_filename_size = le32_to_cpu(le->backing_filename_size);
+
+ cpu->image_filename_offset = le32_to_cpu(le->image_filename_offset);
+ cpu->image_filename_size = le32_to_cpu(le->image_filename_size);
+
+ cpu->cluster_bits = le32_to_cpu(le->cluster_bits);
+ cpu->features = le64_to_cpu(le->features);
+ cpu->optional_features = le64_to_cpu(le->optional_features);
+ cpu->header_pages_size = le32_to_cpu(le->header_pages_size);
+
+ memcpy(cpu->backing_fmt, le->backing_fmt, sizeof(cpu->backing_fmt));
+ memcpy(cpu->image_fmt, le->image_fmt, sizeof(cpu->image_fmt));
+}
+
+static void add_cow_header_cpu_to_le(const AddCowHeader *cpu, AddCowHeader *le)
+{
+ le->magic = cpu_to_le64(cpu->magic);
+ le->version = cpu_to_le32(cpu->version);
+
+ le->backing_filename_offset = cpu_to_le32(cpu->backing_filename_offset);
+ le->backing_filename_size = cpu_to_le32(cpu->backing_filename_size);
+
+ le->image_filename_offset = cpu_to_le32(cpu->image_filename_offset);
+ le->image_filename_size = cpu_to_le32(cpu->image_filename_size);
+
+ le->cluster_bits = cpu_to_le32(cpu->cluster_bits);
+ le->features = cpu_to_le64(cpu->features);
+ le->optional_features = cpu_to_le64(cpu->optional_features);
+ le->header_pages_size = cpu_to_le32(cpu->header_pages_size);
+ memcpy(le->backing_fmt, cpu->backing_fmt, sizeof(cpu->backing_fmt));
+ memcpy(le->image_fmt, cpu->image_fmt, sizeof(cpu->image_fmt));
+}
+
+static int add_cow_probe(const uint8_t *buf, int buf_size, const char *filename)
+{
+ const AddCowHeader *header = (const AddCowHeader *)buf;
+
+ if (le64_to_cpu(header->magic) == ADD_COW_MAGIC &&
+ le32_to_cpu(header->version) == ADD_COW_VERSION) {
+ return 100;
+ } else {
+ return 0;
+ }
+}
+
+static int add_cow_create(const char *filename, QemuOpts *opts)
+{
+ AddCowHeader header = {
+ .magic = ADD_COW_MAGIC,
+ .version = ADD_COW_VERSION,
+ .features = 0,
+ .optional_features = 0,
+ .header_pages_size = ADD_COW_DEFAULT_PAGE_SIZE,
+ };
+ AddCowHeader le_header;
+ int64_t image_len = 0;
+ const char *backing_filename = NULL;
+ const char *backing_fmt = NULL;
+ const char *image_filename = NULL;
+ const char *image_format = NULL;
+ BlockDriverState *bs, *image_bs = NULL, *backing_bs = NULL;
+ BlockDriver *drv = bdrv_find_format("add-cow");
+ BDRVAddCowState s;
+ size_t cluster_size;
+ int ret;
+
+ image_len = qemu_opt_get_number(opts, BLOCK_OPT_SIZE, 0);
+ backing_filename = qemu_opt_get(opts, BLOCK_OPT_BACKING_FILE);
+ backing_fmt = qemu_opt_get(opts, BLOCK_OPT_BACKING_FMT);
+ image_filename = qemu_opt_get(opts, BLOCK_OPT_IMAGE_FILE);
+ image_format = qemu_opt_get(opts, BLOCK_OPT_IMAGE_FMT);
+ cluster_size = qemu_opt_get_size(opts, BLOCK_OPT_CLUSTER_SIZE,
+ ADD_COW_CLUSTER_SIZE);
+
+ header.cluster_bits = ffs(cluster_size) - 1;
+ if (header.cluster_bits < MIN_CLUSTER_BITS ||
+ header.cluster_bits > MAX_CLUSTER_BITS ||
+ (1 << header.cluster_bits) != cluster_size) {
+ error_report(
+ "Cluster size must be a power of two between %d and %dk",
+ 1 << MIN_CLUSTER_BITS, 1 << (MAX_CLUSTER_BITS - 10));
+ return -EINVAL;
+ }
+
+ if (backing_filename) {
+ header.backing_filename_offset = sizeof(header)
+ + sizeof(s.backing_file_format) + sizeof(s.image_file_format);
+ header.backing_filename_size = strlen(backing_filename);
+
+ if (!backing_fmt) {
+ backing_bs = bdrv_new("image");
+ ret = bdrv_open(backing_bs, backing_filename,
+ BDRV_O_RDWR | BDRV_O_CACHE_WB, NULL);
+ if (ret < 0) {
+ return ret;
+ }
+ backing_fmt = bdrv_get_format_name(backing_bs);
+ bdrv_delete(backing_bs);
+ }
+ } else {
+ header.features |= ADD_COW_F_ALL_ALLOCATED;
+ }
+
+ if (image_filename) {
+ header.image_filename_offset =
+ sizeof(header) + sizeof(s.backing_file_format)
+ + sizeof(s.image_file_format) + header.backing_filename_size;
+ header.image_filename_size = strlen(image_filename);
+ } else {
+ error_report("Error: image_file should be given.");
+ return -EINVAL;
+ }
+
+ if (backing_filename && !strcmp(backing_filename, image_filename)) {
+ error_report("Error: Trying to create an image with the "
+ "same backing file name as the image file name");
+ return -EINVAL;
+ }
+
+ if (!strcmp(filename, image_filename)) {
+ error_report("Error: Trying to create an image with the "
+ "same filename as the image file name");
+ return -EINVAL;
+ }
+
+ if (header.image_filename_offset + header.image_filename_size
+ > ADD_COW_PAGE_SIZE * ADD_COW_DEFAULT_PAGE_SIZE) {
+ error_report("image_file name or backing_file name too long.");
+ return -ENOSPC;
+ }
+
+ ret = bdrv_file_open(&image_bs, image_filename, BDRV_O_RDWR);
+ if (ret < 0) {
+ return ret;
+ }
+ bdrv_delete(image_bs);
+
+ ret = bdrv_create_file(filename, NULL);
+ if (ret < 0) {
+ return ret;
+ }
+
+ ret = bdrv_file_open(&bs, filename, BDRV_O_RDWR);
+ if (ret < 0) {
+ return ret;
+ }
+ snprintf(header.backing_fmt, sizeof(header.backing_fmt),
+ "%s", backing_fmt ? backing_fmt : "");
+ snprintf(header.image_fmt, sizeof(header.image_fmt),
+ "%s", image_format ? image_format : "raw");
+ add_cow_header_cpu_to_le(&header, &le_header);
+ ret = bdrv_pwrite(bs, 0, &le_header, sizeof(le_header));
+ if (ret < 0) {
+ bdrv_delete(bs);
+ return ret;
+ }
+
+ if (ret < 0) {
+ bdrv_delete(bs);
+ return ret;
+ }
+
+ if (backing_filename) {
+ ret = bdrv_pwrite(bs, header.backing_filename_offset,
+ backing_filename, header.backing_filename_size);
+ if (ret < 0) {
+ bdrv_delete(bs);
+ return ret;
+ }
+ }
+
+ ret = bdrv_pwrite(bs, header.image_filename_offset,
+ image_filename, header.image_filename_size);
+ if (ret < 0) {
+ bdrv_delete(bs);
+ return ret;
+ }
+
+ ret = bdrv_open(bs, filename, BDRV_O_RDWR | BDRV_O_NO_FLUSH, drv);
+ if (ret < 0) {
+ bdrv_delete(bs);
+ return ret;
+ }
+
+ ret = bdrv_truncate(bs, image_len);
+ bdrv_delete(bs);
+ return ret;
+}
+
+static int add_cow_open(BlockDriverState *bs, int flags)
+{
+ char image_filename[ADD_COW_FILE_LEN];
+ char tmp_name[ADD_COW_FILE_LEN];
+ int ret;
+ int sector_per_byte;
+ BDRVAddCowState *s = bs->opaque;
+ AddCowHeader le_header;
+
+ ret = bdrv_pread(bs->file, 0, &le_header, sizeof(le_header));
+ if (ret < 0) {
+ goto fail;
+ }
+
+ add_cow_header_le_to_cpu(&le_header, &s->header);
+
+ if (s->header.magic != ADD_COW_MAGIC) {
+ ret = -EINVAL;
+ goto fail;
+ }
+
+ if (s->header.version != ADD_COW_VERSION) {
+ char version[64];
+ snprintf(version, sizeof(version), "ADD-COW version %d",
+ s->header.version);
+ qerror_report(QERR_UNKNOWN_BLOCK_FORMAT_FEATURE,
+ bs->device_name, "add-cow", version);
+ ret = -ENOTSUP;
+ goto fail;
+ }
+
+ if (s->header.features & ~ADD_COW_FEATURE_MASK) {
+ char buf[64];
+ snprintf(buf, sizeof(buf), "Feature Flags: %" PRIx64,
+ s->header.features & ~ADD_COW_FEATURE_MASK);
+ qerror_report(QERR_UNKNOWN_BLOCK_FORMAT_FEATURE,
+ bs->device_name, "add-cow", buf);
+ return -ENOTSUP;
+ }
+
+ if ((s->header.features & ADD_COW_F_ALL_ALLOCATED) == 0) {
+ ret = bdrv_read_string(bs->file, sizeof(s->header),
+ sizeof(bs->backing_format) - 1,
+ bs->backing_format,
+ sizeof(bs->backing_format));
+ if (ret < 0) {
+ goto fail;
+ }
+ }
+
+ if (s->header.cluster_bits < MIN_CLUSTER_BITS ||
+ s->header.cluster_bits > MAX_CLUSTER_BITS) {
+ ret = -EINVAL;
+ goto fail;
+ }
+
+ if ((s->header.features & ADD_COW_F_ALL_ALLOCATED) == 0) {
+ ret = bdrv_read_string(bs->file, s->header.backing_filename_offset,
+ s->header.backing_filename_size,
+ bs->backing_file,
+ sizeof(bs->backing_file));
+ if (ret < 0) {
+ goto fail;
+ }
+ }
+
+ ret = bdrv_read_string(bs->file, s->header.image_filename_offset,
+ s->header.image_filename_size, tmp_name,
+ sizeof(tmp_name));
+ if (ret < 0) {
+ goto fail;
+ }
+
+ s->image_hd = bdrv_new("");
+ if (path_has_protocol(image_filename)) {
+ pstrcpy(image_filename, sizeof(image_filename), tmp_name);
+ } else {
+ path_combine(image_filename, sizeof(image_filename),
+ bs->filename, tmp_name);
+ }
+
+ ret = bdrv_open(s->image_hd, image_filename, flags, NULL);
+ if (ret < 0) {
+ bdrv_delete(s->image_hd);
+ goto fail;
+ }
+
+ bs->total_sectors = bdrv_getlength(s->image_hd) >> 9;
+ s->cluster_size = 1 << s->header.cluster_bits;
+ sector_per_byte = SECTORS_PER_CLUSTER * 8;
+ s->bitmap_size =
+ (bs->total_sectors + sector_per_byte - 1) / sector_per_byte;
+ s->bitmap_cache =
+ block_cache_create(bs, ADD_COW_CACHE_SIZE, ADD_COW_CACHE_ENTRY_SIZE,
+ BLOCK_TABLE_BITMAP);
+
+ qemu_co_mutex_init(&s->lock);
+ return 0;
+fail:
+ if (s->bitmap_cache) {
+ block_cache_destroy(bs, s->bitmap_cache);
+ }
+ return ret;
+}
+
+static void add_cow_close(BlockDriverState *bs)
+{
+ BDRVAddCowState *s = bs->opaque;
+ block_cache_destroy(bs, s->bitmap_cache);
+ bdrv_delete(s->image_hd);
+}
+
+static bool is_allocated(BlockDriverState *bs, int64_t sector_num)
+{
+ BDRVAddCowState *s = bs->opaque;
+ BlockCache *c = s->bitmap_cache;
+ int64_t cluster_num = sector_num / SECTORS_PER_CLUSTER;
+ uint8_t *table = NULL;
+ bool val = false;
+ int ret;
+
+ uint64_t offset = ADD_COW_PAGE_SIZE * s->header.header_pages_size
+ + (offset_in_bitmap(sector_num) & (~(c->cluster_size - 1)));
+ ret = block_cache_get(bs, s->bitmap_cache, offset, (void **)&table);
+ if (ret < 0) {
+ return ret;
+ }
+
+ val = table[cluster_num / 8 % ADD_COW_CACHE_ENTRY_SIZE]
+ & (1 << (cluster_num % 8));
+ ret = block_cache_put(bs, s->bitmap_cache, (void **)&table);
+ if (ret < 0) {
+ return ret;
+ }
+ return val;
+}
+
+static coroutine_fn int add_cow_is_allocated(BlockDriverState *bs,
+ int64_t sector_num, int nb_sectors, int *num_same)
+{
+ BDRVAddCowState *s = bs->opaque;
+ int changed;
+
+ if (nb_sectors == 0) {
+ *num_same = 0;
+ return 0;
+ }
+
+ if (s->header.features & ADD_COW_F_ALL_ALLOCATED) {
+ *num_same = nb_sectors;
+ return 1;
+ }
+ changed = is_allocated(bs, sector_num);
+
+ for (*num_same = 1; *num_same < nb_sectors; (*num_same)++) {
+ if (is_allocated(bs, sector_num + *num_same) != changed) {
+ break;
+ }
+ }
+ return changed;
+}
+
+static int add_cow_backing_read(BlockDriverState *bs, QEMUIOVector *qiov,
+ int64_t sector_num, int nb_sectors)
+{
+ int n1;
+ if ((sector_num + nb_sectors) <= bs->total_sectors) {
+ return nb_sectors;
+ }
+ if (sector_num >= bs->total_sectors) {
+ n1 = 0;
+ } else {
+ n1 = bs->total_sectors - sector_num;
+ }
+
+ qemu_iovec_memset(qiov, BDRV_SECTOR_SIZE * n1,
+ 0, BDRV_SECTOR_SIZE * (nb_sectors - n1));
+
+ return n1;
+}
+
+static coroutine_fn int add_cow_co_readv(BlockDriverState *bs,
+ int64_t sector_num,
+ int remaining_sectors,
+ QEMUIOVector *qiov)
+{
+ BDRVAddCowState *s = bs->opaque;
+ int cur_nr_sectors;
+ uint64_t bytes_done = 0;
+ QEMUIOVector hd_qiov;
+ int n1, ret = 0;
+
+ qemu_iovec_init(&hd_qiov, qiov->niov);
+ qemu_co_mutex_lock(&s->lock);
+ while (remaining_sectors != 0) {
+ cur_nr_sectors = remaining_sectors;
+ if (add_cow_is_allocated(bs, sector_num, cur_nr_sectors,
+ &cur_nr_sectors)) {
+ qemu_iovec_reset(&hd_qiov);
+ qemu_iovec_concat(&hd_qiov, qiov, bytes_done,
+ cur_nr_sectors * BDRV_SECTOR_SIZE);
+ qemu_co_mutex_unlock(&s->lock);
+ ret = bdrv_co_readv(s->image_hd, sector_num,
+ cur_nr_sectors, &hd_qiov);
+ qemu_co_mutex_lock(&s->lock);
+ if (ret < 0) {
+ goto fail;
+ }
+ } else {
+ if (bs->backing_hd) {
+ qemu_iovec_reset(&hd_qiov);
+ qemu_iovec_concat(&hd_qiov, qiov, bytes_done,
+ cur_nr_sectors * BDRV_SECTOR_SIZE);
+ n1 = add_cow_backing_read(bs->backing_hd, &hd_qiov,
+ sector_num, cur_nr_sectors);
+ if (n1 > 0) {
+ qemu_co_mutex_unlock(&s->lock);
+ ret = bdrv_co_readv(bs->backing_hd, sector_num,
+ cur_nr_sectors, &hd_qiov);
+ qemu_co_mutex_lock(&s->lock);
+ if (ret < 0) {
+ goto fail;
+ }
+ }
+ } else {
+ qemu_iovec_memset(&hd_qiov, 0, 0,
+ BDRV_SECTOR_SIZE * cur_nr_sectors);
+ }
+ }
+ remaining_sectors -= cur_nr_sectors;
+ sector_num += cur_nr_sectors;
+ bytes_done += cur_nr_sectors * BDRV_SECTOR_SIZE;
+ }
+fail:
+ qemu_co_mutex_unlock(&s->lock);
+ qemu_iovec_destroy(&hd_qiov);
+ return ret;
+}
+
+static int coroutine_fn copy_sectors(BlockDriverState *bs,
+ int n_start, int n_end)
+{
+ BDRVAddCowState *s = bs->opaque;
+ QEMUIOVector qiov;
+ struct iovec iov;
+ int n, ret;
+
+ n = n_end - n_start;
+ if (n <= 0) {
+ return 0;
+ }
+
+ iov.iov_len = n * BDRV_SECTOR_SIZE;
+ iov.iov_base = qemu_blockalign(bs, iov.iov_len);
+
+ qemu_iovec_init_external(&qiov, &iov, 1);
+
+ ret = bdrv_co_readv(bs->backing_hd, n_start, n, &qiov);
+ if (ret < 0) {
+ goto out;
+ }
+ ret = bdrv_co_writev(s->image_hd, n_start, n, &qiov);
+ if (ret < 0) {
+ goto out;
+ }
+
+ ret = 0;
+out:
+ qemu_vfree(iov.iov_base);
+ return ret;
+}
+
+static coroutine_fn int add_cow_co_writev(BlockDriverState *bs,
+ int64_t sector_num,
+ int remaining_sectors,
+ QEMUIOVector *qiov)
+{
+ BDRVAddCowState *s = bs->opaque;
+ BlockCache *c = s->bitmap_cache;
+ int ret = 0, i;
+ QEMUIOVector hd_qiov;
+ uint8_t *table;
+ uint64_t offset;
+ int mask = SECTORS_PER_CLUSTER - 1;
+ int table_mask = c->cluster_size - 1;
+
+ qemu_co_mutex_lock(&s->lock);
+ qemu_iovec_init(&hd_qiov, qiov->niov);
+ ret = bdrv_co_writev(s->image_hd, sector_num,
+ remaining_sectors, qiov);
+
+ if (ret < 0) {
+ goto fail;
+ }
+ if ((s->header.features & ADD_COW_F_ALL_ALLOCATED) == 0) {
+ /* Copy content of unmodified sectors */
+ if (!is_cluster_head(sector_num) && !is_allocated(bs, sector_num)) {
+ ret = copy_sectors(bs, sector_num & ~mask, sector_num);
+ if (ret < 0) {
+ goto fail;
+ }
+ }
+
+ if (!is_cluster_tail(sector_num + remaining_sectors - 1)
+ && !is_allocated(bs, sector_num + remaining_sectors - 1)) {
+ ret = copy_sectors(bs, sector_num + remaining_sectors,
+ ((sector_num + remaining_sectors) | mask) + 1);
+ if (ret < 0) {
+ goto fail;
+ }
+ }
+
+ for (i = sector_num / SECTORS_PER_CLUSTER;
+ i <= (sector_num + remaining_sectors - 1) / SECTORS_PER_CLUSTER;
+ i++) {
+ offset = ADD_COW_PAGE_SIZE * s->header.header_pages_size
+ + (offset_in_bitmap(i * SECTORS_PER_CLUSTER) & (~table_mask));
+ ret = block_cache_get(bs, s->bitmap_cache, offset, (void **)&table);
+ if (ret < 0) {
+ goto fail;
+ }
+ if ((table[i / 8] & (1 << (i % 8))) == 0) {
+ table[i / 8] |= (1 << (i % 8));
+ block_cache_entry_mark_dirty(s->bitmap_cache, table);
+ }
+
+ ret = block_cache_put(bs, s->bitmap_cache, (void **) &table);
+ if (ret < 0) {
+ goto fail;
+ }
+ }
+ }
+ ret = 0;
+fail:
+ qemu_co_mutex_unlock(&s->lock);
+ qemu_iovec_destroy(&hd_qiov);
+ return ret;
+}
+
+static int bdrv_add_cow_truncate(BlockDriverState *bs, int64_t size)
+{
+ BDRVAddCowState *s = bs->opaque;
+ int sector_per_byte = SECTORS_PER_CLUSTER * 8;
+ int ret;
+ uint32_t bitmap_pos = s->header.header_pages_size * ADD_COW_PAGE_SIZE;
+ int64_t bitmap_size =
+ (size / BDRV_SECTOR_SIZE + sector_per_byte - 1) / sector_per_byte;
+ bitmap_size = (bitmap_size + ADD_COW_CACHE_ENTRY_SIZE - 1)
+ & (~(ADD_COW_CACHE_ENTRY_SIZE - 1));
+
+ ret = bdrv_truncate(bs->file, bitmap_pos + bitmap_size);
+ if (ret < 0) {
+ return ret;
+ }
+
+ ret = bdrv_truncate(s->image_hd, size);
+ if (ret < 0) {
+ return ret;
+ }
+ return 0;
+}
+
+static int add_cow_reopen_prepare(BDRVReopenState *state,
+ BlockReopenQueue *queue, Error **errp)
+{
+ BDRVAddCowState *s;
+ int ret = -1;
+
+ assert(state != NULL);
+ assert(state->bs != NULL);
+
+ if (queue == NULL) {
+ error_set(errp, ERROR_CLASS_GENERIC_ERROR,
+ "No reopen queue for add-cow");
+ goto exit;
+ }
+
+ s = state->bs->opaque;
+
+ assert(s != NULL);
+
+
+ bdrv_reopen_queue(queue, s->image_hd, state->flags);
+ ret = 0;
+
+exit:
+ return ret;
+}
+
+
+static coroutine_fn int add_cow_co_flush(BlockDriverState *bs)
+{
+ BDRVAddCowState *s = bs->opaque;
+ int ret;
+
+ qemu_co_mutex_lock(&s->lock);
+ ret = block_cache_flush(bs, s->bitmap_cache);
+ if (ret < 0) {
+ return ret;
+ }
+ ret = bdrv_flush(s->image_hd);
+ qemu_co_mutex_unlock(&s->lock);
+ return ret;
+}
+
+static int add_cow_get_info(BlockDriverState *bs, BlockDriverInfo *bdi)
+{
+ BDRVAddCowState *s = bs->opaque;
+ bdi->cluster_size = s->cluster_size;
+ return 0;
+}
+
+static QemuOptsList add_cow_create_opts = {
+ .name = "add-cow-create-opts",
+ .head = QTAILQ_HEAD_INITIALIZER(add_cow_create_opts.head),
+ .desc = {
+ {
+ .name = BLOCK_OPT_SIZE,
+ .type = QEMU_OPT_NUMBER,
+ .help = "Virtual disk size"
+ },
+ {
+ .name = BLOCK_OPT_BACKING_FILE,
+ .type = QEMU_OPT_STRING,
+ .help = "File name of a base image"
+ },
+ {
+ .name = BLOCK_OPT_BACKING_FMT,
+ .type = QEMU_OPT_STRING,
+ .help = "Image format of the base image"
+ },
+ {
+ .name = BLOCK_OPT_IMAGE_FILE,
+ .type = QEMU_OPT_STRING,
+ .help = "File name of a image file"
+ },
+ {
+ .name = BLOCK_OPT_IMAGE_FMT,
+ .type = QEMU_OPT_STRING,
+ .help = "Image format of the image file"
+ },
+ {
+ .name = BLOCK_OPT_CLUSTER_SIZE,
+ .type = QEMU_OPT_SIZE,
+ .help = "add-cow cluster size",
+ .def_value = ADD_COW_CLUSTER_SIZE
+ },
+ { /* end of list */ }
+ }
+};
+
+static QemuOptsList *add_cow_create_options(void)
+{
+ return &add_cow_create_opts;
+}
+
+static BlockDriver bdrv_add_cow = {
+ .format_name = "add-cow",
+ .instance_size = sizeof(BDRVAddCowState),
+ .bdrv_probe = add_cow_probe,
+ .bdrv_open = add_cow_open,
+ .bdrv_close = add_cow_close,
+ .bdrv_create = add_cow_create,
+ .bdrv_co_readv = add_cow_co_readv,
+ .bdrv_co_writev = add_cow_co_writev,
+ .bdrv_truncate = bdrv_add_cow_truncate,
+ .bdrv_co_is_allocated = add_cow_is_allocated,
+ .bdrv_reopen_prepare = add_cow_reopen_prepare,
+ .bdrv_get_info = add_cow_get_info,
+
+ .bdrv_create_options = add_cow_create_options,
+ .bdrv_co_flush_to_os = add_cow_co_flush,
+};
+
+static void bdrv_add_cow_init(void)
+{
+ bdrv_register(&bdrv_add_cow);
+}
+
+block_init(bdrv_add_cow_init);
diff --git a/block/add-cow.h b/block/add-cow.h
new file mode 100644
index 0000000..ba9a61e
--- /dev/null
+++ b/block/add-cow.h
@@ -0,0 +1,85 @@
+/*
+ * QEMU ADD-COW Disk Format
+ *
+ * Copyright IBM, Corp. 2012
+ *
+ * Authors:
+ * Dong Xu Wang <wdongxu@linux.vnet.ibm.com>
+ *
+ * This work is licensed under the terms of the GNU LGPL, version 2 or later.
+ * See the COPYING.LIB file in the top-level directory.
+ *
+ */
+
+#ifndef BLOCK_ADD_COW_H
+#define BLOCK_ADD_COW_H
+#include "block-cache.h"
+
+enum {
+ ADD_COW_F_ALL_ALLOCATED = 0X01,
+ ADD_COW_FEATURE_MASK = ADD_COW_F_ALL_ALLOCATED,
+
+ ADD_COW_MAGIC = (((uint64_t)'A' << 56) | ((uint64_t)'D' << 48) | \
+ ((uint64_t)'D' << 40) | ((uint64_t)'_' << 32) | \
+ ((uint64_t)'C' << 24) | ((uint64_t)'O' << 16) | \
+ ((uint64_t)'W' << 8) | 0xFF),
+ ADD_COW_VERSION = 1,
+ ADD_COW_FILE_LEN = 1024,
+ ADD_COW_CACHE_SIZE = 16,
+ ADD_COW_CACHE_ENTRY_SIZE = 65536,
+ ADD_COW_CLUSTER_SIZE = 65536,
+ SECTORS_PER_CLUSTER = (ADD_COW_CLUSTER_SIZE / BDRV_SECTOR_SIZE),
+ ADD_COW_PAGE_SIZE = 4096,
+ ADD_COW_DEFAULT_PAGE_SIZE = 1,
+ MIN_CLUSTER_BITS = 9,
+ MAX_CLUSTER_BITS = 21,
+};
+
+typedef struct AddCowHeader {
+ uint64_t magic;
+ uint32_t version;
+
+ uint32_t backing_filename_offset;
+ uint32_t backing_filename_size;
+
+ uint32_t image_filename_offset;
+ uint32_t image_filename_size;
+
+ uint32_t cluster_bits;
+
+ uint64_t features;
+ uint64_t optional_features;
+ uint32_t header_pages_size;
+
+ char backing_fmt[16];
+ char image_fmt[16];
+} QEMU_PACKED AddCowHeader;
+
+typedef struct BDRVAddCowState {
+ BlockDriverState *image_hd;
+ CoMutex lock;
+ int cluster_size;
+ BlockCache *bitmap_cache;
+ uint64_t bitmap_size;
+ AddCowHeader header;
+ char backing_file_format[16];
+ char image_file_format[16];
+} BDRVAddCowState;
+
+/* Convert sector_num to offset in bitmap */
+static inline int64_t offset_in_bitmap(int64_t sector_num)
+{
+ int64_t cluster_num = sector_num / SECTORS_PER_CLUSTER;
+ return cluster_num / 8;
+}
+
+static inline bool is_cluster_head(int64_t sector_num)
+{
+ return sector_num % SECTORS_PER_CLUSTER == 0;
+}
+
+static inline bool is_cluster_tail(int64_t sector_num)
+{
+ return (sector_num + 1) % SECTORS_PER_CLUSTER == 0;
+}
+#endif
diff --git a/block/block-cache.c b/block/block-cache.c
index bf5c57c..1a30462 100644
--- a/block/block-cache.c
+++ b/block/block-cache.c
@@ -112,6 +112,8 @@ static int block_cache_entry_flush(BlockDriverState *bs, BlockCache *c, int i)
BLKDBG_EVENT(bs->file, BLKDBG_REFBLOCK_UPDATE_PART);
} else if (c->table_type == BLOCK_TABLE_L2) {
BLKDBG_EVENT(bs->file, BLKDBG_L2_UPDATE);
+ } else if (c->table_type == BLOCK_TABLE_BITMAP) {
+ BLKDBG_EVENT(bs->file, BLKDBG_COW_WRITE);
}
ret = bdrv_pwrite(bs->file, c->entries[i].offset,
@@ -245,6 +247,8 @@ static int block_cache_do_get(BlockDriverState *bs, BlockCache *c,
if (read_from_disk) {
if (c->table_type == BLOCK_TABLE_L2) {
BLKDBG_EVENT(bs->file, BLKDBG_L2_LOAD);
+ } else if (c->table_type == BLOCK_TABLE_BITMAP) {
+ BLKDBG_EVENT(bs->file, BLKDBG_COW_READ);
}
ret = bdrv_pread(bs->file, offset, c->entries[i].table,
diff --git a/block/block-cache.h b/block/block-cache.h
index 4efa06e..a3c4a1c 100644
--- a/block/block-cache.h
+++ b/block/block-cache.h
@@ -37,6 +37,7 @@
typedef enum {
BLOCK_TABLE_REF,
BLOCK_TABLE_L2,
+ BLOCK_TABLE_BITMAP,
} BlockTableType;
typedef struct BlockCachedTable {
diff --git a/block_int.h b/block_int.h
index a104e70..8a79045 100644
--- a/block_int.h
+++ b/block_int.h
@@ -55,6 +55,8 @@
#define BLOCK_OPT_SUBFMT "subformat"
#define BLOCK_OPT_COMPAT_LEVEL "compat"
#define BLOCK_OPT_LAZY_REFCOUNTS "lazy_refcounts"
+#define BLOCK_OPT_IMAGE_FILE "image_file"
+#define BLOCK_OPT_IMAGE_FMT "image_format"
typedef struct BdrvTrackedRequest BdrvTrackedRequest;
--
1.7.1
^ permalink raw reply related [flat|nested] 12+ messages in thread