#include <fcntl.h>
#include <unistd.h>
#include <getopt.h>
+#include <pthread.h>
+#include <stdbool.h>
#include "ctree.h"
#include "disk-io.h"
#include "convert/source-fs.h"
#include "fsfeatures.h"
-const struct btrfs_convert_operations ext2_convert_ops;
+extern const struct btrfs_convert_operations ext2_convert_ops;
+extern const struct btrfs_convert_operations reiserfs_convert_ops;
static const struct btrfs_convert_operations *convert_operations[] = {
#if BTRFSCONVERT_EXT2
&ext2_convert_ops,
#endif
+#if BTRFSCONVERT_REISERFS
+ &reiserfs_convert_ops,
+#endif
};
static void *print_copied_inodes(void *p)
task_period_start(priv->info, 1000 /* 1s */);
while (1) {
count++;
+ pthread_mutex_lock(&priv->mutex);
printf("copy inodes [%c] [%10llu/%10llu]\r",
work_indicator[count % 4],
(unsigned long long)priv->cur_copy_inodes,
(unsigned long long)priv->max_copy_inodes);
+ pthread_mutex_unlock(&priv->mutex);
fflush(stdout);
task_period_wait(priv->info);
}
struct btrfs_root *root,
u64 disk_bytenr, u64 num_bytes)
{
- u32 blocksize = root->sectorsize;
+ u32 blocksize = root->fs_info->sectorsize;
u64 offset;
char *buffer;
int ret = 0;
int ret;
u32 datacsum = convert_flags & CONVERT_FLAG_DATACSUM;
- if (bytenr != round_down(bytenr, root->sectorsize)) {
+ if (bytenr != round_down(bytenr, root->fs_info->sectorsize)) {
error("bytenr not sectorsize aligned: %llu",
(unsigned long long)bytenr);
return -EINVAL;
}
- if (len != round_down(len, root->sectorsize)) {
+ if (len != round_down(len, root->fs_info->sectorsize)) {
error("length not sectorsize aligned: %llu",
(unsigned long long)len);
return -EINVAL;
len = min_t(u64, len, BTRFS_MAX_EXTENT_SIZE);
/*
- * Skip sb ranges first
- * [0, 1M), [sb_offset(1), +64K), [sb_offset(2), +64K].
+ * Skip reserved ranges first
*
* Or we will insert a hole into current image file, and later
* migrate block will fail as there is already a file extent.
*/
- if (bytenr < 1024 * 1024) {
- *ret_len = 1024 * 1024 - bytenr;
- return 0;
- }
- for (i = 1; i < BTRFS_SUPER_MIRROR_MAX; i++) {
- u64 cur = btrfs_sb_offset(i);
+ for (i = 0; i < ARRAY_SIZE(btrfs_reserved_ranges); i++) {
+ const struct simple_range *reserved = &btrfs_reserved_ranges[i];
- if (bytenr >= cur && bytenr < cur + BTRFS_STRIPE_LEN) {
- *ret_len = cur + BTRFS_STRIPE_LEN - bytenr;
+ /*
+ * |-- reserved --|
+ * |--range---|
+ * or
+ * |---- reserved ----|
+ * |-- range --|
+ * Skip to reserved range end
+ */
+ if (bytenr >= reserved->start && bytenr < range_end(reserved)) {
+ *ret_len = range_end(reserved) - bytenr;
return 0;
}
- }
- for (i = 1; i < BTRFS_SUPER_MIRROR_MAX; i++) {
- u64 cur = btrfs_sb_offset(i);
/*
- * |--reserved--|
+ * |---reserved---|
* |----range-------|
- * May still need to go through file extent inserts
+ * Leading part may still create a file extent
*/
- if (bytenr < cur && bytenr + len >= cur) {
- len = min_t(u64, len, cur - bytenr);
+ if (bytenr < reserved->start &&
+ bytenr + len >= range_end(reserved)) {
+ len = min_t(u64, len, reserved->start - bytenr);
break;
}
- /*
- * |--reserved--|
- * |---range---|
- * Drop out, no need to insert anything
- */
- if (bytenr >= cur && bytenr < cur + BTRFS_STRIPE_LEN) {
- *ret_len = cur + BTRFS_STRIPE_LEN - bytenr;
- return 0;
- }
}
+ /* Check if we are going to insert regular file extent, or hole */
cache = search_cache_extent(used, bytenr);
if (cache) {
if (cache->start <= bytenr) {
* |///////Used///////|
* |<--insert--->|
* bytenr
+ * Insert one real file extent
*/
len = min_t(u64, len, cache->start + cache->size -
bytenr);
* |//Used//|
* |<-insert-->|
* bytenr
+ * Insert one hole
*/
len = min(len, cache->start - bytenr);
disk_bytenr = 0;
* |//Used//| |EOF
* |<-insert-->|
* bytenr
+ * Insert one hole
*/
disk_bytenr = 0;
datacsum = 0;
bg_cache->key.offset - bytenr);
}
- if (len != round_down(len, root->sectorsize)) {
+ if (len != round_down(len, root->fs_info->sectorsize)) {
error("remaining length not sectorsize aligned: %llu",
(unsigned long long)len);
return -EINVAL;
struct btrfs_root *root,
struct cache_tree *used,
struct btrfs_inode_item *inode, int fd,
- u64 ino, u64 start, u64 len,
+ u64 ino, const struct simple_range *range,
u32 convert_flags)
{
- u64 cur_off = start;
- u64 cur_len = len;
- u64 hole_start = start;
+ u64 cur_off = range->start;
+ u64 cur_len = range->len;
+ u64 hole_start = range->start;
u64 hole_len;
struct cache_extent *cache;
struct btrfs_key key;
struct extent_buffer *eb;
int ret = 0;
- while (cur_off < start + len) {
- cache = lookup_cache_extent(used, cur_off, cur_len);
+ /*
+ * It's possible that there are holes in reserved range:
+ * |<---------------- Reserved range ---------------------->|
+ * |<- Old fs data ->| |<- Old fs data ->|
+ * So here we need to iterate through old fs used space and only
+ * migrate ranges that covered by old fs data.
+ */
+ while (cur_off < range_end(range)) {
+ cache = search_cache_extent(used, cur_off);
if (!cache)
break;
cur_off = max(cache->start, cur_off);
- cur_len = min(cache->start + cache->size, start + len) -
+ if (cur_off >= range_end(range))
+ break;
+ cur_len = min(cache->start + cache->size, range_end(range)) -
cur_off;
- BUG_ON(cur_len < root->sectorsize);
+ BUG_ON(cur_len < root->fs_info->sectorsize);
/* reserve extent for the data */
ret = btrfs_reserve_extent(trans, root, cur_len, 0, 0, (u64)-1,
eb->len = key.offset;
/* Write the data */
- ret = write_and_map_eb(root, eb);
+ ret = write_and_map_eb(root->fs_info, eb);
free(eb);
if (ret < 0)
break;
cur_off += key.offset;
hole_start = cur_off;
- cur_len = start + len - cur_off;
+ cur_len = range_end(range) - cur_off;
}
- /* Last hole */
- if (start + len - hole_start > 0)
+ /*
+ * Last hole
+ * |<---- reserved -------->|
+ * |<- Old fs data ->| |
+ * | Hole |
+ */
+ if (range_end(range) - hole_start > 0)
ret = btrfs_record_file_extent(trans, root, ino, inode,
- hole_start, 0, start + len - hole_start);
+ hole_start, 0, range_end(range) - hole_start);
return ret;
}
/*
- * Relocate the used ext2 data in reserved ranges
- * [0,1M)
- * [btrfs_sb_offset(1), +BTRFS_STRIPE_LEN)
- * [btrfs_sb_offset(2), +BTRFS_STRIPE_LEN)
+ * Relocate the used source fs data in reserved ranges
*/
static int migrate_reserved_ranges(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
struct btrfs_inode_item *inode, int fd,
u64 ino, u64 total_bytes, u32 convert_flags)
{
- u64 cur_off;
- u64 cur_len;
+ int i;
int ret = 0;
- /* 0 ~ 1M */
- cur_off = 0;
- cur_len = 1024 * 1024;
- ret = migrate_one_reserved_range(trans, root, used, inode, fd, ino,
- cur_off, cur_len, convert_flags);
- if (ret < 0)
- return ret;
+ for (i = 0; i < ARRAY_SIZE(btrfs_reserved_ranges); i++) {
+ const struct simple_range *range = &btrfs_reserved_ranges[i];
- /* second sb(fisrt sb is included in 0~1M) */
- cur_off = btrfs_sb_offset(1);
- cur_len = min(total_bytes, cur_off + BTRFS_STRIPE_LEN) - cur_off;
- if (cur_off > total_bytes)
- return ret;
- ret = migrate_one_reserved_range(trans, root, used, inode, fd, ino,
- cur_off, cur_len, convert_flags);
- if (ret < 0)
- return ret;
+ if (range->start > total_bytes)
+ return ret;
+ ret = migrate_one_reserved_range(trans, root, used, inode, fd,
+ ino, range, convert_flags);
+ if (ret < 0)
+ return ret;
+ }
- /* Last sb */
- cur_off = btrfs_sb_offset(2);
- cur_len = min(total_bytes, cur_off + BTRFS_STRIPE_LEN) - cur_off;
- if (cur_off > total_bytes)
- return ret;
- ret = migrate_one_reserved_range(trans, root, used, inode, fd, ino,
- cur_off, cur_len, convert_flags);
return ret;
}
static int wipe_reserved_ranges(struct cache_tree *tree, u64 min_stripe_size,
int ensure_size)
{
+ int i;
int ret;
- ret = wipe_one_reserved_range(tree, 0, 1024 * 1024, min_stripe_size,
- ensure_size);
- if (ret < 0)
- return ret;
- ret = wipe_one_reserved_range(tree, btrfs_sb_offset(1),
- BTRFS_STRIPE_LEN, min_stripe_size, ensure_size);
- if (ret < 0)
- return ret;
- ret = wipe_one_reserved_range(tree, btrfs_sb_offset(2),
- BTRFS_STRIPE_LEN, min_stripe_size, ensure_size);
+ for (i = 0; i < ARRAY_SIZE(btrfs_reserved_ranges); i++) {
+ const struct simple_range *range = &btrfs_reserved_ranges[i];
+
+ ret = wipe_one_reserved_range(tree, range->start, range->len,
+ min_stripe_size, ensure_size);
+ if (ret < 0)
+ return ret;
+ }
return ret;
}
* Twice the minimal chunk size, to allow later wipe_reserved_ranges()
* works without need to consider overlap
*/
- u64 min_stripe_size = 2 * 16 * 1024 * 1024;
+ u64 min_stripe_size = SZ_32M;
int ret;
/* Calculate data_chunks */
flags |= BTRFS_INODE_NODATASUM;
trans = btrfs_start_transaction(root, 1);
- if (!trans)
- return -ENOMEM;
+ if (IS_ERR(trans))
+ return PTR_ERR(trans);
cache_tree_init(&used_tmp);
btrfs_init_path(&path);
if (ret < 0)
goto out;
ret = btrfs_add_link(trans, root, ino, BTRFS_FIRST_FREE_OBJECTID, name,
- strlen(name), BTRFS_FT_REG_FILE, NULL, 1);
+ strlen(name), BTRFS_FT_REG_FILE, NULL, 1, 0);
if (ret < 0)
goto out;
* Start from 1M, as 0~1M is reserved, and create_image_file_range()
* can't handle bytenr 0(will consider it as a hole)
*/
- cur = 1024 * 1024;
+ cur = SZ_1M;
while (cur < size) {
u64 len = size - cur;
return ret;
}
-static struct btrfs_root* link_subvol(struct btrfs_root *root,
- const char *base, u64 root_objectid)
-{
- struct btrfs_trans_handle *trans;
- struct btrfs_fs_info *fs_info = root->fs_info;
- struct btrfs_root *tree_root = fs_info->tree_root;
- struct btrfs_root *new_root = NULL;
- struct btrfs_path path;
- struct btrfs_inode_item *inode_item;
- struct extent_buffer *leaf;
- struct btrfs_key key;
- u64 dirid = btrfs_root_dirid(&root->root_item);
- u64 index = 2;
- char buf[BTRFS_NAME_LEN + 1]; /* for snprintf null */
- int len;
- int i;
- int ret;
-
- len = strlen(base);
- if (len == 0 || len > BTRFS_NAME_LEN)
- return NULL;
-
- btrfs_init_path(&path);
- key.objectid = dirid;
- key.type = BTRFS_DIR_INDEX_KEY;
- key.offset = (u64)-1;
-
- ret = btrfs_search_slot(NULL, root, &key, &path, 0, 0);
- if (ret <= 0) {
- error("search for DIR_INDEX dirid %llu failed: %d",
- (unsigned long long)dirid, ret);
- goto fail;
- }
-
- if (path.slots[0] > 0) {
- path.slots[0]--;
- btrfs_item_key_to_cpu(path.nodes[0], &key, path.slots[0]);
- if (key.objectid == dirid && key.type == BTRFS_DIR_INDEX_KEY)
- index = key.offset + 1;
- }
- btrfs_release_path(&path);
-
- trans = btrfs_start_transaction(root, 1);
- if (!trans) {
- error("unable to start transaction");
- goto fail;
- }
-
- key.objectid = dirid;
- key.offset = 0;
- key.type = BTRFS_INODE_ITEM_KEY;
-
- ret = btrfs_lookup_inode(trans, root, &path, &key, 1);
- if (ret) {
- error("search for INODE_ITEM %llu failed: %d",
- (unsigned long long)dirid, ret);
- goto fail;
- }
- leaf = path.nodes[0];
- inode_item = btrfs_item_ptr(leaf, path.slots[0],
- struct btrfs_inode_item);
-
- key.objectid = root_objectid;
- key.offset = (u64)-1;
- key.type = BTRFS_ROOT_ITEM_KEY;
-
- memcpy(buf, base, len);
- for (i = 0; i < 1024; i++) {
- ret = btrfs_insert_dir_item(trans, root, buf, len,
- dirid, &key, BTRFS_FT_DIR, index);
- if (ret != -EEXIST)
- break;
- len = snprintf(buf, ARRAY_SIZE(buf), "%s%d", base, i);
- if (len < 1 || len > BTRFS_NAME_LEN) {
- ret = -EINVAL;
- break;
- }
- }
- if (ret)
- goto fail;
-
- btrfs_set_inode_size(leaf, inode_item, len * 2 +
- btrfs_inode_size(leaf, inode_item));
- btrfs_mark_buffer_dirty(leaf);
- btrfs_release_path(&path);
-
- /* add the backref first */
- ret = btrfs_add_root_ref(trans, tree_root, root_objectid,
- BTRFS_ROOT_BACKREF_KEY,
- root->root_key.objectid,
- dirid, index, buf, len);
- if (ret) {
- error("unable to add root backref for %llu: %d",
- root->root_key.objectid, ret);
- goto fail;
- }
-
- /* now add the forward ref */
- ret = btrfs_add_root_ref(trans, tree_root, root->root_key.objectid,
- BTRFS_ROOT_REF_KEY, root_objectid,
- dirid, index, buf, len);
- if (ret) {
- error("unable to add root ref for %llu: %d",
- root->root_key.objectid, ret);
- goto fail;
- }
-
- ret = btrfs_commit_transaction(trans, root);
- if (ret) {
- error("transaction commit failed: %d", ret);
- goto fail;
- }
-
- new_root = btrfs_read_fs_root(fs_info, &key);
- if (IS_ERR(new_root)) {
- error("unable to fs read root: %lu", PTR_ERR(new_root));
- new_root = NULL;
- }
-fail:
- btrfs_init_path(&path);
- return new_root;
-}
-
static int create_subvol(struct btrfs_trans_handle *trans,
struct btrfs_root *root, u64 root_objectid)
{
* And for single chunk, don't create chunk larger than 1G.
*/
max_chunk_size = cfg->num_bytes / 10;
- max_chunk_size = min((u64)(1024 * 1024 * 1024), max_chunk_size);
- max_chunk_size = round_down(max_chunk_size, extent_root->sectorsize);
+ max_chunk_size = min((u64)(SZ_1G), max_chunk_size);
+ max_chunk_size = round_down(max_chunk_size,
+ extent_root->fs_info->sectorsize);
for (cache = first_cache_extent(data_chunks); cache;
cache = next_cache_extent(cache)) {
len = min(max_chunk_size,
cache->start + cache->size - cur);
- ret = btrfs_alloc_data_chunk(trans, extent_root,
+ ret = btrfs_alloc_data_chunk(trans, fs_info,
&cur_backup, len,
BTRFS_BLOCK_GROUP_DATA, 1);
if (ret < 0)
break;
- ret = btrfs_make_block_group(trans, extent_root, 0,
- BTRFS_BLOCK_GROUP_DATA,
- BTRFS_FIRST_CHUNK_TREE_OBJECTID,
- cur, len);
+ ret = btrfs_make_block_group(trans, fs_info, 0,
+ BTRFS_BLOCK_GROUP_DATA, cur, len);
if (ret < 0)
break;
cur += len;
fs_info->avoid_sys_chunk_alloc = 1;
fs_info->avoid_meta_chunk_alloc = 1;
trans = btrfs_start_transaction(root, 1);
- if (!trans) {
+ if (IS_ERR(trans)) {
error("unable to start transaction");
- ret = -EINVAL;
+ ret = PTR_ERR(trans);
goto err;
}
ret = btrfs_fix_block_accounting(trans, root);
BUG_ON(btrfs_super_bytenr(super) != old_bytenr);
btrfs_set_super_bytenr(super, BTRFS_SUPER_INFO_OFFSET);
- csum_tree_block_size(buf, BTRFS_CRC32_SIZE, 0);
+ csum_tree_block_size(buf, btrfs_csum_sizes[BTRFS_CSUM_TYPE_CRC32], 0);
ret = pwrite(fd, buf->data, BTRFS_SUPER_INFO_SIZE,
BTRFS_SUPER_INFO_OFFSET);
if (ret != BTRFS_SUPER_INFO_SIZE)
return ret;
}
-static int prepare_system_chunk_sb(struct btrfs_super_block *super)
-{
- struct btrfs_chunk *chunk;
- struct btrfs_disk_key *key;
- u32 sectorsize = btrfs_super_sectorsize(super);
-
- key = (struct btrfs_disk_key *)(super->sys_chunk_array);
- chunk = (struct btrfs_chunk *)(super->sys_chunk_array +
- sizeof(struct btrfs_disk_key));
-
- btrfs_set_disk_key_objectid(key, BTRFS_FIRST_CHUNK_TREE_OBJECTID);
- btrfs_set_disk_key_type(key, BTRFS_CHUNK_ITEM_KEY);
- btrfs_set_disk_key_offset(key, 0);
-
- btrfs_set_stack_chunk_length(chunk, btrfs_super_total_bytes(super));
- btrfs_set_stack_chunk_owner(chunk, BTRFS_EXTENT_TREE_OBJECTID);
- btrfs_set_stack_chunk_stripe_len(chunk, BTRFS_STRIPE_LEN);
- btrfs_set_stack_chunk_type(chunk, BTRFS_BLOCK_GROUP_SYSTEM);
- btrfs_set_stack_chunk_io_align(chunk, sectorsize);
- btrfs_set_stack_chunk_io_width(chunk, sectorsize);
- btrfs_set_stack_chunk_sector_size(chunk, sectorsize);
- btrfs_set_stack_chunk_num_stripes(chunk, 1);
- btrfs_set_stack_chunk_sub_stripes(chunk, 0);
- chunk->stripe.devid = super->dev_item.devid;
- btrfs_set_stack_stripe_offset(&chunk->stripe, 0);
- memcpy(chunk->stripe.dev_uuid, super->dev_item.uuid, BTRFS_UUID_SIZE);
- btrfs_set_super_sys_array_size(super, sizeof(*key) + sizeof(*chunk));
- return 0;
-}
-
static int convert_open_fs(const char *devname,
struct btrfs_convert_context *cctx)
{
goto fail;
fd = open(devname, O_RDWR);
if (fd < 0) {
- error("unable to open %s: %s", devname, strerror(errno));
+ error("unable to open %s: %m", devname);
goto fail;
}
btrfs_parse_features_to_string(features_buf, features);
goto fail;
}
- printf("creating btrfs metadata");
+ printf("creating btrfs metadata\n");
+ ret = pthread_mutex_init(&ctx.mutex, NULL);
+ if (ret) {
+ error("failed to initialize mutex: %d", ret);
+ goto fail;
+ }
ctx.max_copy_inodes = (cctx.inodes_count - cctx.free_inodes_count);
ctx.cur_copy_inodes = 0;
task_deinit(ctx.info);
}
- image_root = link_subvol(root, subvol_name, CONV_IMAGE_SUBVOL_OBJECTID);
+ image_root = btrfs_mksubvol(root, subvol_name,
+ CONV_IMAGE_SUBVOL_OBJECTID, true);
if (!image_root) {
error("unable to link subvolume %s", subvol_name);
goto fail;
close_ctree(root);
close(fd);
- printf("conversion complete");
+ printf("conversion complete\n");
return 0;
fail:
clean_convert_context(&cctx);
}
/*
- * Check if a non 1:1 mapped chunk can be rolled back.
- * For new convert, it's OK while for old convert it's not.
+ * Read out data of convert image which is in btrfs reserved ranges so we can
+ * use them to overwrite the ranges during rollback.
*/
-static int may_rollback_chunk(struct btrfs_fs_info *fs_info, u64 bytenr)
+static int read_reserved_ranges(struct btrfs_root *root, u64 ino,
+ u64 total_bytes, char *reserved_ranges[])
{
- struct btrfs_block_group_cache *bg;
- struct btrfs_key key;
- struct btrfs_path path;
- struct btrfs_root *extent_root = fs_info->extent_root;
- u64 bg_start;
- u64 bg_end;
- int ret;
-
- bg = btrfs_lookup_first_block_group(fs_info, bytenr);
- if (!bg)
- return -ENOENT;
- bg_start = bg->key.objectid;
- bg_end = bg->key.objectid + bg->key.offset;
-
- key.objectid = bg_end;
- key.type = BTRFS_METADATA_ITEM_KEY;
- key.offset = 0;
- btrfs_init_path(&path);
-
- ret = btrfs_search_slot(NULL, extent_root, &key, &path, 0, 0);
- if (ret < 0)
- return ret;
+ int i;
+ int ret = 0;
- while (1) {
- struct btrfs_extent_item *ei;
+ for (i = 0; i < ARRAY_SIZE(btrfs_reserved_ranges); i++) {
+ const struct simple_range *range = &btrfs_reserved_ranges[i];
- ret = btrfs_previous_extent_item(extent_root, &path, bg_start);
- if (ret > 0) {
- ret = 0;
+ if (range->start + range->len >= total_bytes)
break;
- }
- if (ret < 0)
+ ret = btrfs_read_file(root, ino, range->start, range->len,
+ reserved_ranges[i]);
+ if (ret < range->len) {
+ error(
+ "failed to read data of convert image, offset=%llu len=%llu ret=%d",
+ range->start, range->len, ret);
+ if (ret >= 0)
+ ret = -EIO;
break;
+ }
+ ret = 0;
+ }
+ return ret;
+}
- btrfs_item_key_to_cpu(path.nodes[0], &key, path.slots[0]);
- if (key.type == BTRFS_METADATA_ITEM_KEY)
- continue;
- /* Now it's EXTENT_ITEM_KEY only */
- ei = btrfs_item_ptr(path.nodes[0], path.slots[0],
- struct btrfs_extent_item);
- /*
- * Found data extent, means this is old convert must follow 1:1
- * mapping.
- */
- if (btrfs_extent_flags(path.nodes[0], ei)
- & BTRFS_EXTENT_FLAG_DATA) {
- ret = -EINVAL;
+static bool is_subset_of_reserved_ranges(u64 start, u64 len)
+{
+ int i;
+ bool ret = false;
+
+ for (i = 0; i < ARRAY_SIZE(btrfs_reserved_ranges); i++) {
+ const struct simple_range *range = &btrfs_reserved_ranges[i];
+
+ if (start >= range->start && start + len <= range_end(range)) {
+ ret = true;
break;
}
}
- btrfs_release_path(&path);
return ret;
}
-static int may_rollback(struct btrfs_root *root)
+static bool is_chunk_direct_mapped(struct btrfs_fs_info *fs_info, u64 start)
{
- struct btrfs_fs_info *info = root->fs_info;
- struct btrfs_multi_bio *multi = NULL;
- u64 bytenr;
- u64 length;
- u64 physical;
- u64 total_bytes;
- int num_stripes;
+ struct cache_extent *ce;
+ struct map_lookup *map;
+ bool ret = false;
+
+ ce = search_cache_extent(&fs_info->mapping_tree.cache_tree, start);
+ if (!ce)
+ goto out;
+ if (ce->start > start || ce->start + ce->size < start)
+ goto out;
+
+ map = container_of(ce, struct map_lookup, ce);
+
+ /* Not SINGLE chunk */
+ if (map->num_stripes != 1)
+ goto out;
+
+ /* Chunk's logical doesn't match with phisical, not 1:1 mapped */
+ if (map->ce.start != map->stripes[0].physical)
+ goto out;
+ ret = true;
+out:
+ return ret;
+}
+
+/*
+ * Iterate all file extents of the convert image.
+ *
+ * All file extents except ones in btrfs_reserved_ranges must be mapped 1:1
+ * on disk. (Means thier file_offset must match their on disk bytenr)
+ *
+ * File extents in reserved ranges can be relocated to other place, and in
+ * that case we will read them out for later use.
+ */
+static int check_convert_image(struct btrfs_root *image_root, u64 ino,
+ u64 total_size, char *reserved_ranges[])
+{
+ struct btrfs_key key;
+ struct btrfs_path path;
+ struct btrfs_fs_info *fs_info = image_root->fs_info;
+ u64 checked_bytes = 0;
int ret;
- if (btrfs_super_num_devices(info->super_copy) != 1)
- goto fail;
+ key.objectid = ino;
+ key.offset = 0;
+ key.type = BTRFS_EXTENT_DATA_KEY;
- bytenr = BTRFS_SUPER_INFO_OFFSET;
- total_bytes = btrfs_super_total_bytes(root->fs_info->super_copy);
+ btrfs_init_path(&path);
+ ret = btrfs_search_slot(NULL, image_root, &key, &path, 0, 0);
+ /*
+ * It's possible that some fs doesn't store any (including sb)
+ * data into 0~1M range, and NO_HOLES is enabled.
+ *
+ * So we only need to check if ret < 0
+ */
+ if (ret < 0) {
+ error("failed to iterate file extents at offset 0: %s",
+ strerror(-ret));
+ btrfs_release_path(&path);
+ return ret;
+ }
+ /* Loop from the first file extents */
while (1) {
- ret = btrfs_map_block(&info->mapping_tree, WRITE, bytenr,
- &length, &multi, 0, NULL);
- if (ret) {
- if (ret == -ENOENT) {
- /* removed block group at the tail */
- if (length == (u64)-1)
- break;
+ struct btrfs_file_extent_item *fi;
+ struct extent_buffer *leaf = path.nodes[0];
+ u64 disk_bytenr;
+ u64 file_offset;
+ u64 ram_bytes;
+ int slot = path.slots[0];
- /* removed block group in the middle */
- goto next;
- }
- goto fail;
+ if (slot >= btrfs_header_nritems(leaf))
+ goto next;
+ btrfs_item_key_to_cpu(leaf, &key, slot);
+
+ /*
+ * Iteration is done, exit normally, we have extra check out of
+ * the loop
+ */
+ if (key.objectid != ino || key.type != BTRFS_EXTENT_DATA_KEY) {
+ ret = 0;
+ break;
+ }
+ file_offset = key.offset;
+ fi = btrfs_item_ptr(leaf, slot, struct btrfs_file_extent_item);
+ if (btrfs_file_extent_type(leaf, fi) != BTRFS_FILE_EXTENT_REG) {
+ ret = -EINVAL;
+ error(
+ "ino %llu offset %llu doesn't have a regular file extent",
+ ino, file_offset);
+ break;
+ }
+ if (btrfs_file_extent_compression(leaf, fi) ||
+ btrfs_file_extent_encryption(leaf, fi) ||
+ btrfs_file_extent_other_encoding(leaf, fi)) {
+ ret = -EINVAL;
+ error(
+ "ino %llu offset %llu doesn't have a plain file extent",
+ ino, file_offset);
+ break;
}
- num_stripes = multi->num_stripes;
- physical = multi->stripes[0].physical;
- free(multi);
+ disk_bytenr = btrfs_file_extent_disk_bytenr(leaf, fi);
+ ram_bytes = btrfs_file_extent_ram_bytes(leaf, fi);
- if (num_stripes != 1) {
- error("num stripes for bytenr %llu is not 1", bytenr);
- goto fail;
- }
+ checked_bytes += ram_bytes;
+ /* Skip hole */
+ if (disk_bytenr == 0)
+ goto next;
/*
- * Extra check for new convert, as metadata chunk from new
- * convert is much more free than old convert, it doesn't need
- * to do 1:1 mapping.
+ * Most file extents must be 1:1 mapped, which means 2 things:
+ * 1) File extent file offset == disk_bytenr
+ * 2) That data chunk's logical == chunk's physical
+ *
+ * So file extent's file offset == physical position on disk.
+ *
+ * And after rolling back btrfs reserved range, other part
+ * remains what old fs used to be.
*/
- if (physical != bytenr) {
+ if (file_offset != disk_bytenr ||
+ !is_chunk_direct_mapped(fs_info, disk_bytenr)) {
/*
- * Check if it's a metadata chunk and has only metadata
- * extent.
+ * Only file extent in btrfs reserved ranges are
+ * allowed to be non-1:1 mapped
*/
- ret = may_rollback_chunk(info, bytenr);
- if (ret < 0)
- goto fail;
+ if (!is_subset_of_reserved_ranges(file_offset,
+ ram_bytes)) {
+ ret = -EINVAL;
+ error(
+ "ino %llu offset %llu file extent should not be relocated",
+ ino, file_offset);
+ break;
+ }
}
next:
- bytenr += length;
- if (bytenr >= total_bytes)
+ ret = btrfs_next_item(image_root, &path);
+ if (ret) {
+ if (ret > 0)
+ ret = 0;
break;
+ }
}
- return 0;
-fail:
- return -1;
+ btrfs_release_path(&path);
+ if (ret)
+ return ret;
+ /*
+ * For HOLES mode (without NO_HOLES), we must ensure file extents
+ * cover the whole range of the image
+ */
+ if (!ret && !btrfs_fs_incompat(fs_info, NO_HOLES)) {
+ if (checked_bytes != total_size) {
+ ret = -EINVAL;
+ error("inode %llu has some file extents not checked",
+ ino);
+ return ret;
+ }
+ }
+
+ /* So far so good, read old data located in btrfs reserved ranges */
+ ret = read_reserved_ranges(image_root, ino, total_size,
+ reserved_ranges);
+ return ret;
}
+/*
+ * btrfs rollback is just reverted convert:
+ * |<---------------Btrfs fs------------------------------>|
+ * |<- Old data chunk ->|< new chunk (D/M/S)>|<- ODC ->|
+ * |<-Old-FE->| |<-Old-FE->|<- Btrfs extents ->|<-Old-FE->|
+ * ||
+ * \/
+ * |<------------------Old fs----------------------------->|
+ * |<- used ->| |<- used ->| |<- used ->|
+ *
+ * However things are much easier than convert, we don't really need to
+ * do the complex space calculation, but only to handle btrfs reserved space
+ *
+ * |<---------------------------Btrfs fs----------------------------->|
+ * | RSV 1 | | Old | | RSV 2 | | Old | | RSV 3 |
+ * | 0~1M | | Fs | | SB2 + 64K | | Fs | | SB3 + 64K |
+ *
+ * On the other hand, the converted fs image in btrfs is a completely
+ * valid old fs.
+ *
+ * |<-----------------Converted fs image in btrfs-------------------->|
+ * | RSV 1 | | Old | | RSV 2 | | Old | | RSV 3 |
+ * | Relocated | | Fs | | Relocated | | Fs | | Relocated |
+ *
+ * Used space in fs image should be at the same physical position on disk.
+ * We only need to recover the data in reserved ranges, so the whole
+ * old fs is back.
+ *
+ * The idea to rollback is also straightforward, we just "read" out the data
+ * of reserved ranges, and write them back to there they should be.
+ * Then the old fs is back.
+ */
static int do_rollback(const char *devname)
{
- int fd = -1;
- int ret;
- int i;
struct btrfs_root *root;
struct btrfs_root *image_root;
- struct btrfs_root *chunk_root;
- struct btrfs_dir_item *dir;
- struct btrfs_inode_item *inode;
- struct btrfs_file_extent_item *fi;
- struct btrfs_trans_handle *trans;
- struct extent_buffer *leaf;
- struct btrfs_block_group_cache *cache1;
- struct btrfs_block_group_cache *cache2;
+ struct btrfs_fs_info *fs_info;
struct btrfs_key key;
struct btrfs_path path;
- struct extent_io_tree io_tree;
- char *buf = NULL;
- char *name;
- u64 bytenr;
- u64 num_bytes;
- u64 root_dir;
- u64 objectid;
- u64 offset;
- u64 start;
- u64 end;
- u64 sb_bytenr;
- u64 first_free;
+ struct btrfs_dir_item *dir;
+ struct btrfs_inode_item *inode_item;
+ char *image_name = "image";
+ char *reserved_ranges[ARRAY_SIZE(btrfs_reserved_ranges)] = { NULL };
u64 total_bytes;
- u32 sectorsize;
+ u64 fsize;
+ u64 root_dir;
+ u64 ino;
+ int fd = -1;
+ int ret;
+ int i;
- extent_io_tree_init(&io_tree);
+ for (i = 0; i < ARRAY_SIZE(btrfs_reserved_ranges); i++) {
+ const struct simple_range *range = &btrfs_reserved_ranges[i];
+ reserved_ranges[i] = calloc(1, range->len);
+ if (!reserved_ranges[i]) {
+ ret = -ENOMEM;
+ goto free_mem;
+ }
+ }
fd = open(devname, O_RDWR);
if (fd < 0) {
- error("unable to open %s: %s", devname, strerror(errno));
- goto fail;
+ error("unable to open %s: %m", devname);
+ ret = -EIO;
+ goto free_mem;
}
- root = open_ctree_fd(fd, devname, 0, OPEN_CTREE_WRITES);
+ fsize = lseek(fd, 0, SEEK_END);
+
+ /*
+ * For rollback, we don't really need to write anything so open it
+ * read-only. The write part will happen after we close the
+ * filesystem.
+ */
+ root = open_ctree_fd(fd, devname, 0, 0);
if (!root) {
error("unable to open ctree");
- goto fail;
- }
- ret = may_rollback(root);
- if (ret < 0) {
- error("unable to do rollback: %d", ret);
- goto fail;
+ ret = -EIO;
+ goto free_mem;
}
+ fs_info = root->fs_info;
- sectorsize = root->sectorsize;
- buf = malloc(sectorsize);
- if (!buf) {
- error("unable to allocate memory");
- goto fail;
- }
-
- btrfs_init_path(&path);
-
+ /*
+ * Search root backref first, or after subvolume deletion (orphan),
+ * we can still rollback the image.
+ */
key.objectid = CONV_IMAGE_SUBVOL_OBJECTID;
key.type = BTRFS_ROOT_BACKREF_KEY;
key.offset = BTRFS_FS_TREE_OBJECTID;
- ret = btrfs_search_slot(NULL, root->fs_info->tree_root, &key, &path, 0,
- 0);
+ btrfs_init_path(&path);
+ ret = btrfs_search_slot(NULL, fs_info->tree_root, &key, &path, 0, 0);
btrfs_release_path(&path);
if (ret > 0) {
- error("unable to convert ext2 image subvolume, is it deleted?");
- goto fail;
+ error("unable to find source fs image subvolume, is it deleted?");
+ ret = -ENOENT;
+ goto close_fs;
} else if (ret < 0) {
- error("unable to open ext2_saved, id %llu: %s",
- (unsigned long long)key.objectid, strerror(-ret));
- goto fail;
+ error("failed to find source fs image subvolume: %s",
+ strerror(-ret));
+ goto close_fs;
}
+ /* Search convert subvolume */
key.objectid = CONV_IMAGE_SUBVOL_OBJECTID;
key.type = BTRFS_ROOT_ITEM_KEY;
key.offset = (u64)-1;
- image_root = btrfs_read_fs_root(root->fs_info, &key);
- if (!image_root || IS_ERR(image_root)) {
- error("unable to open subvolume %llu: %ld",
- (unsigned long long)key.objectid, PTR_ERR(image_root));
- goto fail;
+ image_root = btrfs_read_fs_root(fs_info, &key);
+ if (IS_ERR(image_root)) {
+ ret = PTR_ERR(image_root);
+ error("failed to open convert image subvolume: %s",
+ strerror(-ret));
+ goto close_fs;
}
- name = "image";
- root_dir = btrfs_root_dirid(&root->root_item);
- dir = btrfs_lookup_dir_item(NULL, image_root, &path,
- root_dir, name, strlen(name), 0);
+ /* Search the image file */
+ root_dir = btrfs_root_dirid(&image_root->root_item);
+ dir = btrfs_lookup_dir_item(NULL, image_root, &path, root_dir,
+ image_name, strlen(image_name), 0);
+
if (!dir || IS_ERR(dir)) {
- error("unable to find file %s: %ld", name, PTR_ERR(dir));
- goto fail;
- }
- leaf = path.nodes[0];
- btrfs_dir_item_key_to_cpu(leaf, dir, &key);
+ btrfs_release_path(&path);
+ if (dir)
+ ret = PTR_ERR(dir);
+ else
+ ret = -ENOENT;
+ error("failed to locate file %s: %s", image_name,
+ strerror(-ret));
+ goto close_fs;
+ }
+ btrfs_dir_item_key_to_cpu(path.nodes[0], dir, &key);
btrfs_release_path(&path);
- objectid = key.objectid;
+ /* Get total size of the original image */
+ ino = key.objectid;
ret = btrfs_lookup_inode(NULL, image_root, &path, &key, 0);
- if (ret) {
- error("unable to find inode item: %d", ret);
- goto fail;
- }
- leaf = path.nodes[0];
- inode = btrfs_item_ptr(leaf, path.slots[0], struct btrfs_inode_item);
- total_bytes = btrfs_inode_size(leaf, inode);
- btrfs_release_path(&path);
- key.objectid = objectid;
- key.offset = 0;
- key.type = BTRFS_EXTENT_DATA_KEY;
- ret = btrfs_search_slot(NULL, image_root, &key, &path, 0, 0);
- if (ret != 0) {
- error("unable to find first file extent: %d", ret);
+ if (ret < 0) {
btrfs_release_path(&path);
- goto fail;
- }
-
- /* build mapping tree for the relocated blocks */
- for (offset = 0; offset < total_bytes; ) {
- leaf = path.nodes[0];
- if (path.slots[0] >= btrfs_header_nritems(leaf)) {
- ret = btrfs_next_leaf(root, &path);
- if (ret != 0)
- break;
- continue;
- }
-
- btrfs_item_key_to_cpu(leaf, &key, path.slots[0]);
- if (key.objectid != objectid || key.offset != offset ||
- key.type != BTRFS_EXTENT_DATA_KEY)
- break;
-
- fi = btrfs_item_ptr(leaf, path.slots[0],
- struct btrfs_file_extent_item);
- if (btrfs_file_extent_type(leaf, fi) != BTRFS_FILE_EXTENT_REG)
- break;
- if (btrfs_file_extent_compression(leaf, fi) ||
- btrfs_file_extent_encryption(leaf, fi) ||
- btrfs_file_extent_other_encoding(leaf, fi))
- break;
-
- bytenr = btrfs_file_extent_disk_bytenr(leaf, fi);
- /* skip holes and direct mapped extents */
- if (bytenr == 0 || bytenr == offset)
- goto next_extent;
-
- bytenr += btrfs_file_extent_offset(leaf, fi);
- num_bytes = btrfs_file_extent_num_bytes(leaf, fi);
-
- cache1 = btrfs_lookup_block_group(root->fs_info, offset);
- cache2 = btrfs_lookup_block_group(root->fs_info,
- offset + num_bytes - 1);
- /*
- * Here we must take consideration of old and new convert
- * behavior.
- * For old convert case, sign, there is no consist chunk type
- * that will cover the extent. META/DATA/SYS are all possible.
- * Just ensure relocate one is in SYS chunk.
- * For new convert case, they are all covered by DATA chunk.
- *
- * So, there is not valid chunk type check for it now.
- */
- if (cache1 != cache2)
- break;
-
- set_extent_bits(&io_tree, offset, offset + num_bytes - 1,
- EXTENT_LOCKED);
- set_state_private(&io_tree, offset, bytenr);
-next_extent:
- offset += btrfs_file_extent_num_bytes(leaf, fi);
- path.slots[0]++;
+ error("unable to find inode %llu: %s", ino, strerror(-ret));
+ goto close_fs;
}
+ inode_item = btrfs_item_ptr(path.nodes[0], path.slots[0],
+ struct btrfs_inode_item);
+ total_bytes = btrfs_inode_size(path.nodes[0], inode_item);
btrfs_release_path(&path);
- if (offset < total_bytes) {
- error("unable to build extent mapping (offset %llu, total_bytes %llu)",
- (unsigned long long)offset,
- (unsigned long long)total_bytes);
- error("converted filesystem after balance is unable to rollback");
- goto fail;
+ /* Check if we can rollback the image */
+ ret = check_convert_image(image_root, ino, total_bytes, reserved_ranges);
+ if (ret < 0) {
+ error("old fs image can't be rolled back");
+ goto close_fs;
}
+close_fs:
+ btrfs_release_path(&path);
+ close_ctree_fs_info(fs_info);
+ if (ret)
+ goto free_mem;
- first_free = BTRFS_SUPER_INFO_OFFSET + 2 * sectorsize - 1;
- first_free &= ~((u64)sectorsize - 1);
- /* backup for extent #0 should exist */
- if(!test_range_bit(&io_tree, 0, first_free - 1, EXTENT_LOCKED, 1)) {
- error("no backup for the first extent");
- goto fail;
- }
- /* force no allocation from system block group */
- root->fs_info->system_allocs = -1;
- trans = btrfs_start_transaction(root, 1);
- if (!trans) {
- error("unable to start transaction");
- goto fail;
- }
/*
- * recow the whole chunk tree, this will remove all chunk tree blocks
- * from system block group
+ * Everything is OK, just write back old fs data into btrfs reserved
+ * ranges
+ *
+ * Here, we starts from the backup blocks first, so if something goes
+ * wrong, the fs is still mountable
*/
- chunk_root = root->fs_info->chunk_root;
- memset(&key, 0, sizeof(key));
- while (1) {
- ret = btrfs_search_slot(trans, chunk_root, &key, &path, 0, 1);
- if (ret < 0)
- break;
-
- ret = btrfs_next_leaf(chunk_root, &path);
- if (ret)
- break;
-
- btrfs_item_key_to_cpu(path.nodes[0], &key, path.slots[0]);
- btrfs_release_path(&path);
- }
- btrfs_release_path(&path);
-
- offset = 0;
- num_bytes = 0;
- while(1) {
- cache1 = btrfs_lookup_block_group(root->fs_info, offset);
- if (!cache1)
- break;
-
- if (cache1->flags & BTRFS_BLOCK_GROUP_SYSTEM)
- num_bytes += btrfs_block_group_used(&cache1->item);
-
- offset = cache1->key.objectid + cache1->key.offset;
- }
- /* only extent #0 left in system block group? */
- if (num_bytes > first_free) {
- error(
- "unable to empty system block group (num_bytes %llu, first_free %llu",
- (unsigned long long)num_bytes,
- (unsigned long long)first_free);
- goto fail;
- }
- /* create a system chunk that maps the whole device */
- ret = prepare_system_chunk_sb(root->fs_info->super_copy);
- if (ret) {
- error("unable to update system chunk: %d", ret);
- goto fail;
- }
-
- ret = btrfs_commit_transaction(trans, root);
- if (ret) {
- error("transaction commit failed: %d", ret);
- goto fail;
- }
- ret = close_ctree(root);
- if (ret) {
- error("close_ctree failed: %d", ret);
- goto fail;
- }
-
- /* zero btrfs super block mirrors */
- memset(buf, 0, sectorsize);
- for (i = 1 ; i < BTRFS_SUPER_MIRROR_MAX; i++) {
- bytenr = btrfs_sb_offset(i);
- if (bytenr >= total_bytes)
- break;
- ret = pwrite(fd, buf, sectorsize, bytenr);
- if (ret != sectorsize) {
- error("zeroing superblock mirror %d failed: %d",
- i, ret);
- goto fail;
- }
- }
-
- sb_bytenr = (u64)-1;
- /* copy all relocated blocks back */
- while(1) {
- ret = find_first_extent_bit(&io_tree, 0, &start, &end,
- EXTENT_LOCKED);
- if (ret)
- break;
-
- ret = get_state_private(&io_tree, start, &bytenr);
- BUG_ON(ret);
+ for (i = ARRAY_SIZE(btrfs_reserved_ranges) - 1; i >= 0; i--) {
+ u64 real_size;
+ const struct simple_range *range = &btrfs_reserved_ranges[i];
- clear_extent_bits(&io_tree, start, end, EXTENT_LOCKED);
+ if (range_end(range) >= fsize)
+ continue;
- while (start <= end) {
- if (start == BTRFS_SUPER_INFO_OFFSET) {
- sb_bytenr = bytenr;
- goto next_sector;
- }
- ret = pread(fd, buf, sectorsize, bytenr);
- if (ret < 0) {
- error("reading superblock at %llu failed: %d",
- (unsigned long long)bytenr, ret);
- goto fail;
- }
- BUG_ON(ret != sectorsize);
- ret = pwrite(fd, buf, sectorsize, start);
- if (ret < 0) {
- error("writing superblock at %llu failed: %d",
- (unsigned long long)start, ret);
- goto fail;
- }
- BUG_ON(ret != sectorsize);
-next_sector:
- start += sectorsize;
- bytenr += sectorsize;
+ real_size = min(range_end(range), fsize) - range->start;
+ ret = pwrite(fd, reserved_ranges[i], real_size, range->start);
+ if (ret < real_size) {
+ if (ret < 0)
+ ret = -errno;
+ else
+ ret = -EIO;
+ error("failed to recover range [%llu, %llu): %s",
+ range->start, real_size, strerror(-ret));
+ goto free_mem;
}
+ ret = 0;
}
- ret = fsync(fd);
- if (ret < 0) {
- error("fsync failed: %s", strerror(errno));
- goto fail;
- }
- /*
- * finally, overwrite btrfs super block.
- */
- ret = pread(fd, buf, sectorsize, sb_bytenr);
- if (ret < 0) {
- error("reading primary superblock failed: %s",
- strerror(errno));
- goto fail;
- }
- BUG_ON(ret != sectorsize);
- ret = pwrite(fd, buf, sectorsize, BTRFS_SUPER_INFO_OFFSET);
- if (ret < 0) {
- error("writing primary superblock failed: %s",
- strerror(errno));
- goto fail;
- }
- BUG_ON(ret != sectorsize);
- ret = fsync(fd);
- if (ret < 0) {
- error("fsync failed: %s", strerror(errno));
- goto fail;
- }
-
- close(fd);
- free(buf);
- extent_io_tree_cleanup(&io_tree);
- printf("rollback complete\n");
- return 0;
-
-fail:
- if (fd != -1)
- close(fd);
- free(buf);
- error("rollback aborted");
- return -1;
+free_mem:
+ for (i = 0; i < ARRAY_SIZE(btrfs_reserved_ranges); i++)
+ free(reserved_ranges[i]);
+ if (ret)
+ error("rollback failed");
+ else
+ printf("rollback succeeded\n");
+ return ret;
}
static void print_usage(void)
printf("\n");
printf("Supported filesystems:\n");
printf("\text2/3/4: %s\n", BTRFSCONVERT_EXT2 ? "yes" : "no");
+ printf("\treiserfs: %s\n", BTRFSCONVERT_REISERFS ? "yes" : "no");
}
int main(int argc, char *argv[])