@@ -2755,16 +2755,20 @@ again:
num_pages *= 16;
num_pages *= PAGE_CACHE_SIZE;
- ret = btrfs_check_data_free_space(inode, num_pages);
+ ret = btrfs_delalloc_reserve_space(inode, num_pages);
if (ret)
goto out_put;
ret = btrfs_prealloc_file_range_trans(inode, trans, 0, 0, num_pages,
num_pages, num_pages,
&alloc_hint);
- if (!ret)
+ if (!ret) {
dcs = BTRFS_DC_SETUP;
- btrfs_free_reserved_data_space(inode, num_pages);
+ btrfs_free_reserved_data_space(inode, num_pages);
+ } else {
+ btrfs_delalloc_release_space(inode, num_pages);
+ }
+
out_put:
iput(inode);
out_free:
@@ -4002,9 +4006,13 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes)
struct btrfs_block_rsv *block_rsv = &root->fs_info->delalloc_block_rsv;
u64 to_reserve = 0;
unsigned nr_extents = 0;
+ int flush = 1;
int ret;
- if (btrfs_transaction_in_commit(root->fs_info))
+ if (btrfs_is_free_space_inode(root, inode))
+ flush = 0;
+
+ if (flush && btrfs_transaction_in_commit(root->fs_info))
schedule_timeout(1);
num_bytes = ALIGN(num_bytes, root->sectorsize);
@@ -4023,7 +4031,7 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes)
to_reserve += calc_csum_metadata_size(inode, num_bytes, 1);
spin_unlock(&BTRFS_I(inode)->lock);
- ret = reserve_metadata_bytes(NULL, root, block_rsv, to_reserve, 1);
+ ret = reserve_metadata_bytes(NULL, root, block_rsv, to_reserve, flush);
if (ret) {
u64 to_free = 0;
unsigned dropped;
@@ -528,6 +528,19 @@ out:
return ret;
}
+/**
+ * __btrfs_write_out_cache - write out cached info to an inode
+ * @root - the root the inode belongs to
+ * @ctl - the free space cache we are going to write out
+ * @block_group - the block_group for this cache if it belongs to a block_group
+ * @trans - the trans handle
+ * @path - the path to use
+ * @offset - the offset for the key we'll insert
+ *
+ * This function writes out a free space cache struct to disk for quick recovery
+ * on mount. This will return 0 if it was successfull in writing the cache out,
+ * and -1 if it was not.
+ */
int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
struct btrfs_free_space_ctl *ctl,
struct btrfs_block_group_cache *block_group,
@@ -551,7 +564,8 @@ int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
int index = 0, num_pages = 0;
int entries = 0;
int bitmaps = 0;
- int ret = -1;
+ int ret;
+ int err = -1;
bool next_page = false;
bool out_of_space = false;
@@ -559,7 +573,7 @@ int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
node = rb_first(&ctl->free_space_offset);
if (!node)
- return 0;
+ return -1;
if (!i_size_read(inode))
return -1;
@@ -763,7 +777,6 @@ int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
unlock_extent_cached(&BTRFS_I(inode)->io_tree, 0,
i_size_read(inode) - 1, &cached_state,
GFP_NOFS);
- ret = 0;
goto out;
}
@@ -785,10 +798,8 @@ int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
unlock_extent_cached(&BTRFS_I(inode)->io_tree, 0,
i_size_read(inode) - 1, &cached_state, GFP_NOFS);
- if (ret) {
- ret = 0;
+ if (ret)
goto out;
- }
BTRFS_I(inode)->generation = trans->transid;
@@ -800,7 +811,6 @@ int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
if (ret < 0) {
- ret = -1;
clear_extent_bit(&BTRFS_I(inode)->io_tree, 0, bytes - 1,
EXTENT_DIRTY | EXTENT_DELALLOC |
EXTENT_DO_ACCOUNTING, 0, 0, NULL, GFP_NOFS);
@@ -814,7 +824,6 @@ int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
if (found_key.objectid != BTRFS_FREE_SPACE_OBJECTID ||
found_key.offset != offset) {
- ret = -1;
clear_extent_bit(&BTRFS_I(inode)->io_tree, 0, bytes - 1,
EXTENT_DIRTY | EXTENT_DELALLOC |
EXTENT_DO_ACCOUNTING, 0, 0, NULL,
@@ -831,16 +840,15 @@ int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
btrfs_mark_buffer_dirty(leaf);
btrfs_release_path(path);
- ret = 1;
-
+ err = 0;
out:
kfree(pages);
- if (ret != 1) {
+ if (err) {
invalidate_inode_pages2_range(inode->i_mapping, 0, index);
BTRFS_I(inode)->generation = 0;
}
btrfs_update_inode(trans, root, inode);
- return ret;
+ return err;
}
int btrfs_write_out_cache(struct btrfs_root *root,
@@ -867,14 +875,16 @@ int btrfs_write_out_cache(struct btrfs_root *root,
ret = __btrfs_write_out_cache(root, inode, ctl, block_group, trans,
path, block_group->key.objectid);
- if (ret < 0) {
+ if (ret) {
+ btrfs_delalloc_release_metadata(inode, inode->i_size);
spin_lock(&block_group->lock);
block_group->disk_cache_state = BTRFS_DC_ERROR;
spin_unlock(&block_group->lock);
ret = 0;
-
+#ifdef DEBUG
printk(KERN_ERR "btrfs: failed to write free space cace "
"for block group %llu\n", block_group->key.objectid);
+#endif
}
iput(inode);
@@ -2658,9 +2668,13 @@ int btrfs_write_out_ino_cache(struct btrfs_root *root,
return 0;
ret = __btrfs_write_out_cache(root, inode, ctl, NULL, trans, path, 0);
- if (ret < 0)
+ if (ret) {
+ btrfs_delalloc_release_metadata(inode, inode->i_size);
+#ifdef DEBUG
printk(KERN_ERR "btrfs: failed to write free ino cache "
"for root %llu\n", root->root_key.objectid);
+#endif
+ }
iput(inode);
return ret;
@@ -465,14 +465,16 @@ again:
/* Just to make sure we have enough space */
prealloc += 8 * PAGE_CACHE_SIZE;
- ret = btrfs_check_data_free_space(inode, prealloc);
+ ret = btrfs_delalloc_reserve_space(inode, prealloc);
if (ret)
goto out_put;
ret = btrfs_prealloc_file_range_trans(inode, trans, 0, 0, prealloc,
prealloc, prealloc, &alloc_hint);
- if (ret)
+ if (ret) {
+ btrfs_delalloc_release_space(inode, prealloc);
goto out_put;
+ }
btrfs_free_reserved_data_space(inode, prealloc);
out_put:
@@ -1792,11 +1792,11 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end)
}
ret = 0;
out:
+ btrfs_delalloc_release_metadata(inode, ordered_extent->len);
if (nolock) {
if (trans)
btrfs_end_transaction_nolock(trans, root);
} else {
- btrfs_delalloc_release_metadata(inode, ordered_extent->len);
if (trans)
btrfs_end_transaction(trans, root);
}
Since free space inodes now use normal checksumming we need to make sure to account for their metadata use. So reserve metadata space, and then if we fail to write out the metadata we can just release it, otherwise it will be freed up when the io completes. Thanks, Signed-off-by: Josef Bacik <josef@redhat.com> --- fs/btrfs/extent-tree.c | 18 ++++++++++++---- fs/btrfs/free-space-cache.c | 44 ++++++++++++++++++++++++++++-------------- fs/btrfs/inode-map.c | 6 +++- fs/btrfs/inode.c | 2 +- 4 files changed, 47 insertions(+), 23 deletions(-)