Btrfs: Data ordered fixes
* In btrfs_delete_inode, wait for ordered extents after calling
truncate_inode_pages. This is much faster, and more correct
* Properly clear our the PageChecked bit everywhere we redirty the page.
* Change the writepage fixup handler to lock the page range and check to
see if an ordered extent had been inserted since the improperly dirtied
page was discovered
* Wait for ordered extents outside the transaction. This isn't required
for locking rules but does improve transaction latencies
* Reduce contention on the alloc_mutex by dropping it while incrementing
refs on a node/leaf and while dropping refs on a leaf.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index be2aef1..ccd4932 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -934,7 +934,6 @@
if (!root->ref_cows)
return 0;
- mutex_lock(&root->fs_info->alloc_mutex);
level = btrfs_header_level(buf);
nritems = btrfs_header_nritems(buf);
for (i = 0; i < nritems; i++) {
@@ -951,29 +950,36 @@
disk_bytenr = btrfs_file_extent_disk_bytenr(buf, fi);
if (disk_bytenr == 0)
continue;
+
+ mutex_lock(&root->fs_info->alloc_mutex);
ret = __btrfs_inc_extent_ref(trans, root, disk_bytenr,
btrfs_file_extent_disk_num_bytes(buf, fi),
root->root_key.objectid, trans->transid,
key.objectid, key.offset);
+ mutex_unlock(&root->fs_info->alloc_mutex);
if (ret) {
faili = i;
+ WARN_ON(1);
goto fail;
}
} else {
bytenr = btrfs_node_blockptr(buf, i);
btrfs_node_key_to_cpu(buf, &key, i);
+
+ mutex_lock(&root->fs_info->alloc_mutex);
ret = __btrfs_inc_extent_ref(trans, root, bytenr,
btrfs_level_size(root, level - 1),
root->root_key.objectid,
trans->transid,
level - 1, key.objectid);
+ mutex_unlock(&root->fs_info->alloc_mutex);
if (ret) {
faili = i;
+ WARN_ON(1);
goto fail;
}
}
}
- mutex_unlock(&root->fs_info->alloc_mutex);
return 0;
fail:
WARN_ON(1);
@@ -1004,7 +1010,6 @@
}
}
#endif
- mutex_unlock(&root->fs_info->alloc_mutex);
return ret;
}
@@ -2180,6 +2185,8 @@
leaf_owner = btrfs_header_owner(leaf);
leaf_generation = btrfs_header_generation(leaf);
+ mutex_unlock(&root->fs_info->alloc_mutex);
+
for (i = 0; i < nritems; i++) {
u64 disk_bytenr;
@@ -2197,12 +2204,17 @@
disk_bytenr = btrfs_file_extent_disk_bytenr(leaf, fi);
if (disk_bytenr == 0)
continue;
+
+ mutex_lock(&root->fs_info->alloc_mutex);
ret = __btrfs_free_extent(trans, root, disk_bytenr,
btrfs_file_extent_disk_num_bytes(leaf, fi),
leaf_owner, leaf_generation,
key.objectid, key.offset, 0);
+ mutex_unlock(&root->fs_info->alloc_mutex);
BUG_ON(ret);
}
+
+ mutex_lock(&root->fs_info->alloc_mutex);
return 0;
}