Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs

Pull btrfs update from Chris Mason:
 "A big set of fixes and features.

  In terms of line count, most of the code comes from Stefan, who added
  the ability to replace a single drive in place.  This is different
  from how btrfs normally replaces drives, and is much much much faster.

  Josef is plowing through our synchronous write performance.  This pull
  request does not include the DIO_OWN_WAITING patch that was discussed
  on the list, but it has a number of other improvements to cut down our
  latencies and CPU time during fsync/O_DIRECT writes.

  Miao Xie has a big series of fixes and is spreading out ordered
  operations over more CPUs.  This improves performance and reduces
  contention.

  I've put in fixes for error handling around hash collisions.  These
  are going back to individual stable kernels as I test against them.

  Otherwise we have a lot of fixes and cleanups, thanks everyone!
  raid5/6 is being rebased against the device replacement code.  I'll
  have it posted this Friday along with a nice series of benchmarks."

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs: (115 commits)
  Btrfs: fix a bug of per-file nocow
  Btrfs: fix hash overflow handling
  Btrfs: don't take inode delalloc mutex if we're a free space inode
  Btrfs: fix autodefrag and umount lockup
  Btrfs: fix permissions of empty files not affected by umask
  Btrfs: put raid properties into global table
  Btrfs: fix BUG() in scrub when first superblock reading gives EIO
  Btrfs: do not call file_update_time in aio_write
  Btrfs: only unlock and relock if we have to
  Btrfs: use tokens where we can in the tree log
  Btrfs: optimize leaf_space_used
  Btrfs: don't memset new tokens
  Btrfs: only clear dirty on the buffer if it is marked as dirty
  Btrfs: move checks in set_page_dirty under DEBUG
  Btrfs: log changed inodes based on the extent map tree
  Btrfs: add path->really_keep_locks
  Btrfs: do not mark ems as prealloc if we are writing to them
  Btrfs: keep track of the extents original block length
  Btrfs: inline csums if we're fsyncing
  Btrfs: don't bother copying if we're only logging the inode
  ...

42 files changed, 5255 insertions, 1745 deletions

diff --git a/fs/btrfs/Makefile b/fs/btrfs/Makefile
index d7fcdba141a2..7df3e0f0ee51 100644
--- a/fs/btrfs/Makefile
+++ b/fs/btrfs/Makefile
@@ -8,7 +8,7 @@ btrfs-y += super.o ctree.o extent-tree.o print-tree.o root-tree.o dir-item.o \
 	   extent_io.o volumes.o async-thread.o ioctl.o locking.o orphan.o \
 	   export.o tree-log.o free-space-cache.o zlib.o lzo.o \
 	   compression.o delayed-ref.o relocation.o delayed-inode.o scrub.o \
-	   reada.o backref.o ulist.o qgroup.o send.o
+	   reada.o backref.o ulist.o qgroup.o send.o dev-replace.o
 
 btrfs-$(CONFIG_BTRFS_FS_POSIX_ACL) += acl.o
 btrfs-$(CONFIG_BTRFS_FS_CHECK_INTEGRITY) += check-integrity.o
diff --git a/fs/btrfs/acl.c b/fs/btrfs/acl.c
index 0c16e3dbfd56..e15d2b0d8d3b 100644
--- a/fs/btrfs/acl.c
+++ b/fs/btrfs/acl.c
@@ -121,6 +121,8 @@ static int btrfs_set_acl(struct btrfs_trans_handle *trans,
 			ret = posix_acl_equiv_mode(acl, &inode->i_mode);
 			if (ret < 0)
 				return ret;
+			if (ret == 0)
+				acl = NULL;
 		}
 		ret = 0;
 		break;
diff --git a/fs/btrfs/backref.c b/fs/btrfs/backref.c
index 208d8aa5b07e..04edf69be875 100644
--- a/fs/btrfs/backref.c
+++ b/fs/btrfs/backref.c
@@ -461,6 +461,7 @@ static int __merge_refs(struct list_head *head, int mode)
 		     pos2 = n2, n2 = pos2->next) {
 			struct __prelim_ref *ref2;
 			struct __prelim_ref *xchg;
+			struct extent_inode_elem *eie;
 
 			ref2 = list_entry(pos2, struct __prelim_ref, list);
 
@@ -472,12 +473,20 @@ static int __merge_refs(struct list_head *head, int mode)
 					ref1 = ref2;
 					ref2 = xchg;
 				}
-				ref1->count += ref2->count;
 			} else {
 				if (ref1->parent != ref2->parent)
 					continue;
-				ref1->count += ref2->count;
 			}
+
+			eie = ref1->inode_list;
+			while (eie && eie->next)
+				eie = eie->next;
+			if (eie)
+				eie->next = ref2->inode_list;
+			else
+				ref1->inode_list = ref2->inode_list;
+			ref1->count += ref2->count;
+
 			list_del(&ref2->list);
 			kfree(ref2);
 		}
@@ -890,8 +899,7 @@ again:
 	while (!list_empty(&prefs)) {
 		ref = list_first_entry(&prefs, struct __prelim_ref, list);
 		list_del(&ref->list);
-		if (ref->count < 0)
-			WARN_ON(1);
+		WARN_ON(ref->count < 0);
 		if (ref->count && ref->root_id && ref->parent == 0) {
 			/* no parent == root of tree */
 			ret = ulist_add(roots, ref->root_id, 0, GFP_NOFS);
diff --git a/fs/btrfs/btrfs_inode.h b/fs/btrfs/btrfs_inode.h
index ed8ca7ca5eff..2a8c242bc4f5 100644
--- a/fs/btrfs/btrfs_inode.h
+++ b/fs/btrfs/btrfs_inode.h
@@ -39,6 +39,7 @@
 #define BTRFS_INODE_HAS_ORPHAN_ITEM		5
 #define BTRFS_INODE_HAS_ASYNC_EXTENT		6
 #define BTRFS_INODE_NEEDS_FULL_SYNC		7
+#define BTRFS_INODE_COPY_EVERYTHING		8
 
 /* in memory btrfs inode */
 struct btrfs_inode {
@@ -90,6 +91,9 @@ struct btrfs_inode {
 
 	unsigned long runtime_flags;
 
+	/* Keep track of who's O_SYNC/fsycing currently */
+	atomic_t sync_writers;
+
 	/* full 64 bit generation number, struct vfs_inode doesn't have a big
 	 * enough field for this.
 	 */
diff --git a/fs/btrfs/check-integrity.c b/fs/btrfs/check-integrity.c
index 5a3e45db642a..11d47bfb62b4 100644
--- a/fs/btrfs/check-integrity.c
+++ b/fs/btrfs/check-integrity.c
@@ -137,7 +137,7 @@ struct btrfsic_block {
 	unsigned int never_written:1;	/* block was added because it was
 					 * referenced, not because it was
 					 * written */
-	unsigned int mirror_num:2;	/* large enough to hold
+	unsigned int mirror_num;	/* large enough to hold
 					 * BTRFS_SUPER_MIRROR_MAX */
 	struct btrfsic_dev_state *dev_state;
 	u64 dev_bytenr;		/* key, physical byte num on disk */
@@ -723,7 +723,7 @@ static int btrfsic_process_superblock(struct btrfsic_state *state,
 		}
 
 		num_copies =
-		    btrfs_num_copies(&state->root->fs_info->mapping_tree,
+		    btrfs_num_copies(state->root->fs_info,
 				     next_bytenr, state->metablock_size);
 		if (state->print_mask & BTRFSIC_PRINT_MASK_NUM_COPIES)
 			printk(KERN_INFO "num_copies(log_bytenr=%llu) = %d\n",
@@ -903,7 +903,7 @@ static int btrfsic_process_superblock_dev_mirror(
 		}
 
 		num_copies =
-		    btrfs_num_copies(&state->root->fs_info->mapping_tree,
+		    btrfs_num_copies(state->root->fs_info,
 				     next_bytenr, state->metablock_size);
 		if (state->print_mask & BTRFSIC_PRINT_MASK_NUM_COPIES)
 			printk(KERN_INFO "num_copies(log_bytenr=%llu) = %d\n",
@@ -1287,7 +1287,7 @@ static int btrfsic_create_link_to_next_block(
 	*next_blockp = NULL;
 	if (0 == *num_copiesp) {
 		*num_copiesp =
-		    btrfs_num_copies(&state->root->fs_info->mapping_tree,
+		    btrfs_num_copies(state->root->fs_info,
 				     next_bytenr, state->metablock_size);
 		if (state->print_mask & BTRFSIC_PRINT_MASK_NUM_COPIES)
 			printk(KERN_INFO "num_copies(log_bytenr=%llu) = %d\n",
@@ -1489,7 +1489,7 @@ static int btrfsic_handle_extent_data(
 			chunk_len = num_bytes;
 
 		num_copies =
-		    btrfs_num_copies(&state->root->fs_info->mapping_tree,
+		    btrfs_num_copies(state->root->fs_info,
 				     next_bytenr, state->datablock_size);
 		if (state->print_mask & BTRFSIC_PRINT_MASK_NUM_COPIES)
 			printk(KERN_INFO "num_copies(log_bytenr=%llu) = %d\n",
@@ -1582,9 +1582,21 @@ static int btrfsic_map_block(struct btrfsic_state *state, u64 bytenr, u32 len,
 	struct btrfs_device *device;
 
 	length = len;
-	ret = btrfs_map_block(&state->root->fs_info->mapping_tree, READ,
+	ret = btrfs_map_block(state->root->fs_info, READ,
 			      bytenr, &length, &multi, mirror_num);
 
+	if (ret) {
+		block_ctx_out->start = 0;
+		block_ctx_out->dev_bytenr = 0;
+		block_ctx_out->len = 0;
+		block_ctx_out->dev = NULL;
+		block_ctx_out->datav = NULL;
+		block_ctx_out->pagev = NULL;
+		block_ctx_out->mem_to_free = NULL;
+
+		return ret;
+	}
+
 	device = multi->stripes[0].dev;
 	block_ctx_out->dev = btrfsic_dev_state_lookup(device->bdev);
 	block_ctx_out->dev_bytenr = multi->stripes[0].physical;
@@ -1594,8 +1606,7 @@ static int btrfsic_map_block(struct btrfsic_state *state, u64 bytenr, u32 len,
 	block_ctx_out->pagev = NULL;
 	block_ctx_out->mem_to_free = NULL;
 
-	if (0 == ret)
-		kfree(multi);
+	kfree(multi);
 	if (NULL == block_ctx_out->dev) {
 		ret = -ENXIO;
 		printk(KERN_INFO "btrfsic: error, cannot lookup dev (#1)!\n");
@@ -2463,7 +2474,7 @@ static int btrfsic_process_written_superblock(
 		}
 
 		num_copies =
-		    btrfs_num_copies(&state->root->fs_info->mapping_tree,
+		    btrfs_num_copies(state->root->fs_info,
 				     next_bytenr, BTRFS_SUPER_INFO_SIZE);
 		if (state->print_mask & BTRFSIC_PRINT_MASK_NUM_COPIES)
 			printk(KERN_INFO "num_copies(log_bytenr=%llu) = %d\n",
@@ -2960,7 +2971,7 @@ static void btrfsic_cmp_log_and_dev_bytenr(struct btrfsic_state *state,
 	struct btrfsic_block_data_ctx block_ctx;
 	int match = 0;
 
-	num_copies = btrfs_num_copies(&state->root->fs_info->mapping_tree,
+	num_copies = btrfs_num_copies(state->root->fs_info,
 				      bytenr, state->metablock_size);
 
 	for (mirror_num = 1; mirror_num <= num_copies; mirror_num++) {
diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c
index c6467aa88bee..94ab2f80e7e3 100644
--- a/fs/btrfs/compression.c
+++ b/fs/btrfs/compression.c
@@ -687,7 +687,8 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio,
 
 			ret = btrfs_map_bio(root, READ, comp_bio,
 					    mirror_num, 0);
-			BUG_ON(ret); /* -ENOMEM */
+			if (ret)
+				bio_endio(comp_bio, ret);
 
 			bio_put(comp_bio);
 
@@ -712,7 +713,8 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio,
 	}
 
 	ret = btrfs_map_bio(root, READ, comp_bio, mirror_num, 0);
-	BUG_ON(ret); /* -ENOMEM */
+	if (ret)
+		bio_endio(comp_bio, ret);
 
 	bio_put(comp_bio);
 	return 0;
diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c
index cdfb4c49a806..c7b67cf24bba 100644
--- a/fs/btrfs/ctree.c
+++ b/fs/btrfs/ctree.c
@@ -38,8 +38,7 @@ static int balance_node_right(struct btrfs_trans_handle *trans,
 			      struct extent_buffer *dst_buf,
 			      struct extent_buffer *src_buf);
 static void del_ptr(struct btrfs_trans_handle *trans, struct btrfs_root *root,
-		    struct btrfs_path *path, int level, int slot,
-		    int tree_mod_log);
+		    struct btrfs_path *path, int level, int slot);
 static void tree_mod_log_free_eb(struct btrfs_fs_info *fs_info,
 				 struct extent_buffer *eb);
 struct extent_buffer *read_old_tree_block(struct btrfs_root *root, u64 bytenr,
@@ -776,8 +775,7 @@ tree_mod_log_eb_move(struct btrfs_fs_info *fs_info, struct extent_buffer *dst,
 
 static noinline void
 tree_mod_log_set_node_key(struct btrfs_fs_info *fs_info,
-			  struct extent_buffer *eb,
-			  struct btrfs_disk_key *disk_key, int slot, int atomic)
+			  struct extent_buffer *eb, int slot, int atomic)
 {
 	int ret;
 
@@ -1140,13 +1138,13 @@ __tree_mod_log_rewind(struct extent_buffer *eb, u64 time_seq,
 		switch (tm->op) {
 		case MOD_LOG_KEY_REMOVE_WHILE_FREEING:
 			BUG_ON(tm->slot < n);
-		case MOD_LOG_KEY_REMOVE_WHILE_MOVING:
 		case MOD_LOG_KEY_REMOVE:
+			n++;
+		case MOD_LOG_KEY_REMOVE_WHILE_MOVING:
 			btrfs_set_node_key(eb, &tm->key, tm->slot);
 			btrfs_set_node_blockptr(eb, tm->slot, tm->blockptr);
 			btrfs_set_node_ptr_generation(eb, tm->slot,
 						      tm->generation);
-			n++;
 			break;
 		case MOD_LOG_KEY_REPLACE:
 			BUG_ON(tm->slot >= n);
@@ -1361,19 +1359,16 @@ noinline int btrfs_cow_block(struct btrfs_trans_handle *trans,
 	u64 search_start;
 	int ret;
 
-	if (trans->transaction != root->fs_info->running_transaction) {
-		printk(KERN_CRIT "trans %llu running %llu\n",
+	if (trans->transaction != root->fs_info->running_transaction)
+		WARN(1, KERN_CRIT "trans %llu running %llu\n",
 		       (unsigned long long)trans->transid,
 		       (unsigned long long)
 		       root->fs_info->running_transaction->transid);
-		WARN_ON(1);
-	}
-	if (trans->transid != root->fs_info->generation) {
-		printk(KERN_CRIT "trans %llu running %llu\n",
+
+	if (trans->transid != root->fs_info->generation)
+		WARN(1, KERN_CRIT "trans %llu running %llu\n",
 		       (unsigned long long)trans->transid,
 		       (unsigned long long)root->fs_info->generation);
-		WARN_ON(1);
-	}
 
 	if (!should_cow_block(trans, root, buf)) {
 		*cow_ret = buf;
@@ -1469,10 +1464,8 @@ int btrfs_realloc_node(struct btrfs_trans_handle *trans,
 	if (cache_only && parent_level != 1)
 		return 0;
 
-	if (trans->transaction != root->fs_info->running_transaction)
-		WARN_ON(1);
-	if (trans->transid != root->fs_info->generation)
-		WARN_ON(1);
+	WARN_ON(trans->transaction != root->fs_info->running_transaction);
+	WARN_ON(trans->transid != root->fs_info->generation);
 
 	parent_nritems = btrfs_header_nritems(parent);
 	blocksize = btrfs_level_size(root, parent_level - 1);
@@ -1827,7 +1820,7 @@ static noinline int balance_level(struct btrfs_trans_handle *trans,
 		if (btrfs_header_nritems(right) == 0) {
 			clean_tree_block(trans, root, right);
 			btrfs_tree_unlock(right);
-			del_ptr(trans, root, path, level + 1, pslot + 1, 1);
+			del_ptr(trans, root, path, level + 1, pslot + 1);
 			root_sub_used(root, right->len);
 			btrfs_free_tree_block(trans, root, right, 0, 1);
 			free_extent_buffer_stale(right);
@@ -1836,7 +1829,7 @@ static noinline int balance_level(struct btrfs_trans_handle *trans,
 			struct btrfs_disk_key right_key;
 			btrfs_node_key(right, &right_key, 0);
 			tree_mod_log_set_node_key(root->fs_info, parent,
-						  &right_key, pslot + 1, 0);
+						  pslot + 1, 0);
 			btrfs_set_node_key(parent, &right_key, pslot + 1);
 			btrfs_mark_buffer_dirty(parent);
 		}
@@ -1871,7 +1864,7 @@ static noinline int balance_level(struct btrfs_trans_handle *trans,
 	if (btrfs_header_nritems(mid) == 0) {
 		clean_tree_block(trans, root, mid);
 		btrfs_tree_unlock(mid);
-		del_ptr(trans, root, path, level + 1, pslot, 1);
+		del_ptr(trans, root, path, level + 1, pslot);
 		root_sub_used(root, mid->len);
 		btrfs_free_tree_block(trans, root, mid, 0, 1);
 		free_extent_buffer_stale(mid);
@@ -1880,7 +1873,7 @@ static noinline int balance_level(struct btrfs_trans_handle *trans,
 		/* update the parent key to reflect our changes */
 		struct btrfs_disk_key mid_key;
 		btrfs_node_key(mid, &mid_key, 0);
-		tree_mod_log_set_node_key(root->fs_info, parent, &mid_key,
+		tree_mod_log_set_node_key(root->fs_info, parent,
 					  pslot, 0);
 		btrfs_set_node_key(parent, &mid_key, pslot);
 		btrfs_mark_buffer_dirty(parent);
@@ -1980,7 +1973,7 @@ static noinline int push_nodes_for_insert(struct btrfs_trans_handle *trans,
 			orig_slot += left_nr;
 			btrfs_node_key(mid, &disk_key, 0);
 			tree_mod_log_set_node_key(root->fs_info, parent,
-						  &disk_key, pslot, 0);
+						  pslot, 0);
 			btrfs_set_node_key(parent, &disk_key, pslot);
 			btrfs_mark_buffer_dirty(parent);
 			if (btrfs_header_nritems(left) > orig_slot) {
@@ -2033,7 +2026,7 @@ static noinline int push_nodes_for_insert(struct btrfs_trans_handle *trans,
 
 			btrfs_node_key(right, &disk_key, 0);
 			tree_mod_log_set_node_key(root->fs_info, parent,
-						  &disk_key, pslot + 1, 0);
+						  pslot + 1, 0);
 			btrfs_set_node_key(parent, &disk_key, pslot + 1);
 			btrfs_mark_buffer_dirty(parent);
 
@@ -2219,6 +2212,9 @@ static noinline void unlock_up(struct btrfs_path *path, int level,
 	int no_skips = 0;
 	struct extent_buffer *t;
 
+	if (path->really_keep_locks)
+		return;
+
 	for (i = level; i < BTRFS_MAX_LEVEL; i++) {
 		if (!path->nodes[i])
 			break;
@@ -2266,7 +2262,7 @@ noinline void btrfs_unlock_up_safe(struct btrfs_path *path, int level)
 {
 	int i;
 
-	if (path->keep_locks)
+	if (path->keep_locks || path->really_keep_locks)
 		return;
 
 	for (i = level; i < BTRFS_MAX_LEVEL; i++) {
@@ -2499,7 +2495,7 @@ int btrfs_search_slot(struct btrfs_trans_handle *trans, struct btrfs_root
 	if (!cow)
 		write_lock_level = -1;
 
-	if (cow && (p->keep_locks || p->lowest_level))
+	if (cow && (p->really_keep_locks || p->keep_locks || p->lowest_level))
 		write_lock_level = BTRFS_MAX_LEVEL;
 
 	min_write_lock_level = write_lock_level;
@@ -2568,7 +2564,10 @@ again:
 			 * must have write locks on this node and the
 			 * parent
 			 */
-			if (level + 1 > write_lock_level) {
+			if (level > write_lock_level ||
+			    (level + 1 > write_lock_level &&
+			    level + 1 < BTRFS_MAX_LEVEL &&
+			    p->nodes[level + 1])) {
 				write_lock_level = level + 1;
 				btrfs_release_path(p);
 				goto again;
@@ -2917,7 +2916,7 @@ static void fixup_low_keys(struct btrfs_trans_handle *trans,
 		if (!path->nodes[i])
 			break;
 		t = path->nodes[i];
-		tree_mod_log_set_node_key(root->fs_info, t, key, tslot, 1);
+		tree_mod_log_set_node_key(root->fs_info, t, tslot, 1);
 		btrfs_set_node_key(t, key, tslot);
 		btrfs_mark_buffer_dirty(path->nodes[i]);
 		if (tslot != 0)
@@ -3302,14 +3301,21 @@ static noinline int split_node(struct btrfs_trans_handle *trans,
  */
 static int leaf_space_used(struct extent_buffer *l, int start, int nr)
 {
+	struct btrfs_item *start_item;
+	struct btrfs_item *end_item;
+	struct btrfs_map_token token;
 	int data_len;
 	int nritems = btrfs_header_nritems(l);
 	int end = min(nritems, start + nr) - 1;
 
 	if (!nr)
 		return 0;
-	data_len = btrfs_item_end_nr(l, start);
-	data_len = data_len - btrfs_item_offset_nr(l, end);
+	btrfs_init_map_token(&token);
+	start_item = btrfs_item_nr(l, start);
+	end_item = btrfs_item_nr(l, end);
+	data_len = btrfs_token_item_offset(l, start_item, &token) +
+		btrfs_token_item_size(l, start_item, &token);
+	data_len = data_len - btrfs_token_item_offset(l, end_item, &token);
 	data_len += sizeof(struct btrfs_item) * nr;
 	WARN_ON(data_len < 0);
 	return data_len;
@@ -3403,8 +3409,7 @@ static noinline int __push_leaf_right(struct btrfs_trans_handle *trans,
 	if (push_items == 0)
 		goto out_unlock;
 
-	if (!empty && push_items == left_nritems)
-		WARN_ON(1);
+	WARN_ON(!empty && push_items == left_nritems);
 
 	/* push left to right */
 	right_nritems = btrfs_header_nritems(right);
@@ -3642,11 +3647,9 @@ static noinline int __push_leaf_left(struct btrfs_trans_handle *trans,
 	btrfs_set_header_nritems(left, old_left_nritems + push_items);
 
 	/* fixup right node */
-	if (push_items > right_nritems) {
-		printk(KERN_CRIT "push items %d nr %u\n", push_items,
+	if (push_items > right_nritems)
+		WARN(1, KERN_CRIT "push items %d nr %u\n", push_items,
 		       right_nritems);
-		WARN_ON(1);
-	}
 
 	if (push_items < right_nritems) {
 		push_space = btrfs_item_offset_nr(right, push_items - 1) -
@@ -4602,16 +4605,21 @@ int btrfs_insert_item(struct btrfs_trans_handle *trans, struct btrfs_root
  * empty a node.
  */
 static void del_ptr(struct btrfs_trans_handle *trans, struct btrfs_root *root,
-		    struct btrfs_path *path, int level, int slot,
-		    int tree_mod_log)
+		    struct btrfs_path *path, int level, int slot)
 {
 	struct extent_buffer *parent = path->nodes[level];
 	u32 nritems;
 	int ret;
 
+	if (level) {
+		ret = tree_mod_log_insert_key(root->fs_info, parent, slot,
+					      MOD_LOG_KEY_REMOVE);
+		BUG_ON(ret < 0);
+	}
+
 	nritems = btrfs_header_nritems(parent);
 	if (slot != nritems - 1) {
-		if (tree_mod_log && level)
+		if (level)
 			tree_mod_log_eb_move(root->fs_info, parent, slot,
 					     slot + 1, nritems - slot - 1);
 		memmove_extent_buffer(parent,
@@ -4619,10 +4627,6 @@ static void del_ptr(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 			      btrfs_node_key_ptr_offset(slot + 1),
 			      sizeof(struct btrfs_key_ptr) *
 			      (nritems - slot - 1));
-	} else if (tree_mod_log && level) {
-		ret = tree_mod_log_insert_key(root->fs_info, parent, slot,
-					      MOD_LOG_KEY_REMOVE);
-		BUG_ON(ret < 0);
 	}
 
 	nritems--;
@@ -4656,7 +4660,7 @@ static noinline void btrfs_del_leaf(struct btrfs_trans_handle *trans,
 				    struct extent_buffer *leaf)
 {
 	WARN_ON(btrfs_header_generation(leaf) != trans->transid);
-	del_ptr(trans, root, path, 1, path->slots[1], 1);
+	del_ptr(trans, root, path, 1, path->slots[1]);
 
 	/*
 	 * btrfs_free_extent is expensive, we want to make sure we
@@ -5123,13 +5127,13 @@ int btrfs_compare_trees(struct btrfs_root *left_root,
 	right_path->search_commit_root = 1;
 	right_path->skip_locking = 1;
 
-	spin_lock(&left_root->root_times_lock);
+	spin_lock(&left_root->root_item_lock);
 	left_start_ctransid = btrfs_root_ctransid(&left_root->root_item);
-	spin_unlock(&left_root->root_times_lock);
+	spin_unlock(&left_root->root_item_lock);
 
-	spin_lock(&right_root->root_times_lock);
+	spin_lock(&right_root->root_item_lock);
 	right_start_ctransid = btrfs_root_ctransid(&right_root->root_item);
-	spin_unlock(&right_root->root_times_lock);
+	spin_unlock(&right_root->root_item_lock);
 
 	trans = btrfs_join_transaction(left_root);
 	if (IS_ERR(trans)) {
@@ -5224,15 +5228,15 @@ int btrfs_compare_trees(struct btrfs_root *left_root,
 				goto out;
 			}
 
-			spin_lock(&left_root->root_times_lock);
+			spin_lock(&left_root->root_item_lock);
 			ctransid = btrfs_root_ctransid(&left_root->root_item);
-			spin_unlock(&left_root->root_times_lock);
+			spin_unlock(&left_root->root_item_lock);
 			if (ctransid != left_start_ctransid)
 				left_start_ctransid = 0;
 
-			spin_lock(&right_root->root_times_lock);
+			spin_lock(&right_root->root_item_lock);
 			ctransid = btrfs_root_ctransid(&right_root->root_item);
-			spin_unlock(&right_root->root_times_lock);
+			spin_unlock(&right_root->root_item_lock);
 			if (ctransid != right_start_ctransid)
 				right_start_ctransid = 0;
 
@@ -5496,6 +5500,139 @@ int btrfs_next_leaf(struct btrfs_root *root, struct btrfs_path *path)
 	return btrfs_next_old_leaf(root, path, 0);
 }
 
+/* Release the path up to but not including the given level */
+static void btrfs_release_level(struct btrfs_path *path, int level)
+{
+	int i;
+
+	for (i = 0; i < level; i++) {
+		path->slots[i] = 0;
+		if (!path->nodes[i])
+			continue;
+		if (path->locks[i]) {
+			btrfs_tree_unlock_rw(path->nodes[i], path->locks[i]);
+			path->locks[i] = 0;
+		}
+		free_extent_buffer(path->nodes[i]);
+		path->nodes[i] = NULL;
+	}
+}
+
+/*
+ * This function assumes 2 things
+ *
+ * 1) You are using path->keep_locks
+ * 2) You are not inserting items.
+ *
+ * If either of these are not true do not use this function. If you need a next
+ * leaf with either of these not being true then this function can be easily
+ * adapted to do that, but at the moment these are the limitations.
+ */
+int btrfs_next_leaf_write(struct btrfs_trans_handle *trans,
+			  struct btrfs_root *root, struct btrfs_path *path,
+			  int del)
+{
+	struct extent_buffer *b;
+	struct btrfs_key key;
+	u32 nritems;
+	int level = 1;
+	int slot;
+	int ret = 1;
+	int write_lock_level = BTRFS_MAX_LEVEL;
+	int ins_len = del ? -1 : 0;
+
+	WARN_ON(!(path->keep_locks || path->really_keep_locks));
+
+	nritems = btrfs_header_nritems(path->nodes[0]);
+	btrfs_item_key_to_cpu(path->nodes[0], &key, nritems - 1);
+
+	while (path->nodes[level]) {
+		nritems = btrfs_header_nritems(path->nodes[level]);
+		if (!(path->locks[level] & BTRFS_WRITE_LOCK)) {
+search:
+			btrfs_release_path(path);
+			ret = btrfs_search_slot(trans, root, &key, path,
+						ins_len, 1);
+			if (ret < 0)
+				goto out;
+			level = 1;
+			continue;
+		}
+
+		if (path->slots[level] >= nritems - 1) {
+			level++;
+			continue;
+		}
+
+		btrfs_release_level(path, level);
+		break;
+	}
+
+	if (!path->nodes[level]) {
+		ret = 1;
+		goto out;
+	}
+
+	path->slots[level]++;
+	b = path->nodes[level];
+
+	while (b) {
+		level = btrfs_header_level(b);
+
+		if (!should_cow_block(trans, root, b))
+			goto cow_done;
+
+		btrfs_set_path_blocking(path);
+		ret = btrfs_cow_block(trans, root, b,
+				      path->nodes[level + 1],
+				      path->slots[level + 1], &b);
+		if (ret)
+			goto out;
+cow_done:
+		path->nodes[level] = b;
+		btrfs_clear_path_blocking(path, NULL, 0);
+		if (level != 0) {
+			ret = setup_nodes_for_search(trans, root, path, b,
+						     level, ins_len,
+						     &write_lock_level);
+			if (ret == -EAGAIN)
+				goto search;
+			if (ret)
+				goto out;
+
+			b = path->nodes[level];
+			slot = path->slots[level];
+
+			ret = read_block_for_search(trans, root, path,
+						    &b, level, slot, &key, 0);
+			if (ret == -EAGAIN)
+				goto search;
+			if (ret)
+				goto out;
+			level = btrfs_header_level(b);
+			if (!btrfs_try_tree_write_lock(b)) {
+				btrfs_set_path_blocking(path);
+				btrfs_tree_lock(b);
+				btrfs_clear_path_blocking(path, b,
+							  BTRFS_WRITE_LOCK);
+			}
+			path->locks[level] = BTRFS_WRITE_LOCK;
+			path->nodes[level] = b;
+			path->slots[level] = 0;
+		} else {
+			path->slots[level] = 0;
+			ret = 0;
+			break;
+		}
+	}
+
+out:
+	if (ret)
+		btrfs_release_path(path);<