summaryrefslogtreecommitdiff
path: root/fs/btrfs/tree-log.c
diff options
context:
space:
mode:
authorChris Mason <chris.mason@oracle.com>2008-09-08 11:18:08 -0400
committerChris Mason <chris.mason@oracle.com>2008-09-25 11:04:07 -0400
commit4bef084857ab8fe71cf49eae349c25e440a49150 (patch)
tree7a9a850515538421c2976f8ee4c1bea5ceced59c /fs/btrfs/tree-log.c
parent325cd4bafeb6cfb44addd6e807a9b74170d1be31 (diff)
Btrfs: Tree logging fixes
* Pin down data blocks to prevent them from being reallocated like so: trans 1: allocate file extent trans 2: free file extent trans 3: free file extent during old snapshot deletion trans 3: allocate file extent to new file trans 3: fsync new file Before the tree logging code, this was legal because the fsync would commit the transation that did the final data extent free and the transaction that allocated the extent to the new file at the same time. With the tree logging code, the tree log subtransaction can commit before the transaction that freed the extent. If we crash, we're left with two different files using the extent. * Don't wait in start_transaction if log replay is going on. This avoids deadlocks from iput while we're cleaning up link counts in the replay code. * Don't deadlock in replay_one_name by trying to read an inode off the disk while holding paths for the directory * Hold the buffer lock while we mark a buffer as written. This closes a race where someone is changing a buffer while we write it. They are supposed to mark it dirty again after they change it, but this violates the cow rules. Signed-off-by: Chris Mason <chris.mason@oracle.com>
Diffstat (limited to 'fs/btrfs/tree-log.c')
-rw-r--r--fs/btrfs/tree-log.c13
1 files changed, 9 insertions, 4 deletions
diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c
index d1ce8314b948..13d7ee8e0c52 100644
--- a/fs/btrfs/tree-log.c
+++ b/fs/btrfs/tree-log.c
@@ -1176,8 +1176,8 @@ static noinline int replay_one_name(struct btrfs_trans_handle *trans,
struct btrfs_key found_key;
struct btrfs_key log_key;
struct inode *dir;
- struct inode *inode;
u8 log_type;
+ int exists;
int ret;
dir = read_one_inode(root, key->objectid);
@@ -1190,6 +1190,13 @@ static noinline int replay_one_name(struct btrfs_trans_handle *trans,
name_len);
btrfs_dir_item_key_to_cpu(eb, di, &log_key);
+ exists = btrfs_lookup_inode(trans, root, path, &log_key, 0);
+ if (exists == 0)
+ exists = 1;
+ else
+ exists = 0;
+ btrfs_release_path(root, path);
+
if (key->type == BTRFS_DIR_ITEM_KEY) {
dst_di = btrfs_lookup_dir_item(trans, root, path, key->objectid,
name, name_len, 1);
@@ -1224,11 +1231,9 @@ static noinline int replay_one_name(struct btrfs_trans_handle *trans,
* don't drop the conflicting directory entry if the inode
* for the new entry doesn't exist
*/
- inode = read_one_inode(root, log_key.objectid);
- if (!inode)
+ if (!exists)
goto out;
- iput(inode);
ret = drop_one_dir_item(trans, root, path, dir, dst_di);
BUG_ON(ret);