These changes are the raw update to linux-4.4.6-rt14. Kernel sources

are taken from kernel.org, and rt patch from the rt wiki download page.

During the rebasing, the following patch collided:

Force tick interrupt and get rid of softirq magic(I70131fb85).

Collisions have been removed because its logic was found on the
source already.

Change-Id: I7f57a4081d9deaa0d9ccfc41a6c8daccdee3b769
Signed-off-by: José Pekkarinen <jose.pekkarinen@nokia.com>

54 files changed, 6921 insertions, 3638 deletions

diff --git a/kernel/fs/btrfs/async-thread.c b/kernel/fs/btrfs/async-thread.c
index df9932b00..9aba42b78 100644
--- a/kernel/fs/btrfs/async-thread.c
+++ b/kernel/fs/btrfs/async-thread.c
@@ -42,8 +42,14 @@ struct __btrfs_workqueue {
 
 	/* Thresholding related variants */
 	atomic_t pending;
-	int max_active;
-	int current_max;
+
+	/* Up limit of concurrency workers */
+	int limit_active;
+
+	/* Current number of concurrency workers */
+	int current_active;
+
+	/* Threshold to change current_active */
 	int thresh;
 	unsigned int count;
 	spinlock_t thres_lock;
@@ -85,9 +91,10 @@ BTRFS_WORK_HELPER(extent_refs_helper);
 BTRFS_WORK_HELPER(scrub_helper);
 BTRFS_WORK_HELPER(scrubwrc_helper);
 BTRFS_WORK_HELPER(scrubnc_helper);
+BTRFS_WORK_HELPER(scrubparity_helper);
 
 static struct __btrfs_workqueue *
-__btrfs_alloc_workqueue(const char *name, unsigned int flags, int max_active,
+__btrfs_alloc_workqueue(const char *name, unsigned int flags, int limit_active,
 			 int thresh)
 {
 	struct __btrfs_workqueue *ret = kzalloc(sizeof(*ret), GFP_NOFS);
@@ -95,26 +102,31 @@ __btrfs_alloc_workqueue(const char *name, unsigned int flags, int max_active,
 	if (!ret)
 		return NULL;
 
-	ret->max_active = max_active;
+	ret->limit_active = limit_active;
 	atomic_set(&ret->pending, 0);
 	if (thresh == 0)
 		thresh = DFT_THRESHOLD;
 	/* For low threshold, disabling threshold is a better choice */
 	if (thresh < DFT_THRESHOLD) {
-		ret->current_max = max_active;
+		ret->current_active = limit_active;
 		ret->thresh = NO_THRESHOLD;
 	} else {
-		ret->current_max = 1;
+		/*
+		 * For threshold-able wq, let its concurrency grow on demand.
+		 * Use minimal max_active at alloc time to reduce resource
+		 * usage.
+		 */
+		ret->current_active = 1;
 		ret->thresh = thresh;
 	}
 
 	if (flags & WQ_HIGHPRI)
 		ret->normal_wq = alloc_workqueue("%s-%s-high", flags,
-						 ret->max_active,
-						 "btrfs", name);
+						 ret->current_active, "btrfs",
+						 name);
 	else
 		ret->normal_wq = alloc_workqueue("%s-%s", flags,
-						 ret->max_active, "btrfs",
+						 ret->current_active, "btrfs",
 						 name);
 	if (!ret->normal_wq) {
 		kfree(ret);
@@ -133,7 +145,7 @@ __btrfs_destroy_workqueue(struct __btrfs_workqueue *wq);
 
 struct btrfs_workqueue *btrfs_alloc_workqueue(const char *name,
 					      unsigned int flags,
-					      int max_active,
+					      int limit_active,
 					      int thresh)
 {
 	struct btrfs_workqueue *ret = kzalloc(sizeof(*ret), GFP_NOFS);
@@ -142,14 +154,14 @@ struct btrfs_workqueue *btrfs_alloc_workqueue(const char *name,
 		return NULL;
 
 	ret->normal = __btrfs_alloc_workqueue(name, flags & ~WQ_HIGHPRI,
-					      max_active, thresh);
+					      limit_active, thresh);
 	if (!ret->normal) {
 		kfree(ret);
 		return NULL;
 	}
 
 	if (flags & WQ_HIGHPRI) {
-		ret->high = __btrfs_alloc_workqueue(name, flags, max_active,
+		ret->high = __btrfs_alloc_workqueue(name, flags, limit_active,
 						    thresh);
 		if (!ret->high) {
 			__btrfs_destroy_workqueue(ret->normal);
@@ -179,7 +191,7 @@ static inline void thresh_queue_hook(struct __btrfs_workqueue *wq)
  */
 static inline void thresh_exec_hook(struct __btrfs_workqueue *wq)
 {
-	int new_max_active;
+	int new_current_active;
 	long pending;
 	int need_change = 0;
 
@@ -196,7 +208,7 @@ static inline void thresh_exec_hook(struct __btrfs_workqueue *wq)
 	wq->count %= (wq->thresh / 4);
 	if (!wq->count)
 		goto  out;
-	new_max_active = wq->current_max;
+	new_current_active = wq->current_active;
 
 	/*
 	 * pending may be changed later, but it's OK since we really
@@ -204,19 +216,19 @@ static inline void thresh_exec_hook(struct __btrfs_workqueue *wq)
 	 */
 	pending = atomic_read(&wq->pending);
 	if (pending > wq->thresh)
-		new_max_active++;
+		new_current_active++;
 	if (pending < wq->thresh / 2)
-		new_max_active--;
-	new_max_active = clamp_val(new_max_active, 1, wq->max_active);
-	if (new_max_active != wq->current_max)  {
+		new_current_active--;
+	new_current_active = clamp_val(new_current_active, 1, wq->limit_active);
+	if (new_current_active != wq->current_active)  {
 		need_change = 1;
-		wq->current_max = new_max_active;
+		wq->current_active = new_current_active;
 	}
 out:
 	spin_unlock(&wq->thres_lock);
 
 	if (need_change) {
-		workqueue_set_max_active(wq->normal_wq, wq->current_max);
+		workqueue_set_max_active(wq->normal_wq, wq->current_active);
 	}
 }
 
@@ -316,8 +328,8 @@ static inline void __btrfs_queue_work(struct __btrfs_workqueue *wq,
 		list_add_tail(&work->ordered_list, &wq->ordered_list);
 		spin_unlock_irqrestore(&wq->list_lock, flags);
 	}
-	queue_work(wq->normal_wq, &work->normal_work);
 	trace_btrfs_work_queued(work);
+	queue_work(wq->normal_wq, &work->normal_work);
 }
 
 void btrfs_queue_work(struct btrfs_workqueue *wq,
@@ -350,13 +362,13 @@ void btrfs_destroy_workqueue(struct btrfs_workqueue *wq)
 	kfree(wq);
 }
 
-void btrfs_workqueue_set_max(struct btrfs_workqueue *wq, int max)
+void btrfs_workqueue_set_max(struct btrfs_workqueue *wq, int limit_active)
 {
 	if (!wq)
 		return;
-	wq->normal->max_active = max;
+	wq->normal->limit_active = limit_active;
 	if (wq->high)
-		wq->high->max_active = max;
+		wq->high->limit_active = limit_active;
 }
 
 void btrfs_set_work_high_priority(struct btrfs_work *work)
diff --git a/kernel/fs/btrfs/async-thread.h b/kernel/fs/btrfs/async-thread.h
index ec2ee477f..ad4d0647d 100644
--- a/kernel/fs/btrfs/async-thread.h
+++ b/kernel/fs/btrfs/async-thread.h
@@ -64,10 +64,12 @@ BTRFS_WORK_HELPER_PROTO(extent_refs_helper);
 BTRFS_WORK_HELPER_PROTO(scrub_helper);
 BTRFS_WORK_HELPER_PROTO(scrubwrc_helper);
 BTRFS_WORK_HELPER_PROTO(scrubnc_helper);
+BTRFS_WORK_HELPER_PROTO(scrubparity_helper);
+
 
 struct btrfs_workqueue *btrfs_alloc_workqueue(const char *name,
 					      unsigned int flags,
-					      int max_active,
+					      int limit_active,
 					      int thresh);
 void btrfs_init_work(struct btrfs_work *work, btrfs_work_func_t helper,
 		     btrfs_func_t func,
diff --git a/kernel/fs/btrfs/backref.c b/kernel/fs/btrfs/backref.c
index 614aaa196..e2f659dc5 100644
--- a/kernel/fs/btrfs/backref.c
+++ b/kernel/fs/btrfs/backref.c
@@ -206,10 +206,33 @@ static int __add_prelim_ref(struct list_head *head, u64 root_id,
 		return -ENOMEM;
 
 	ref->root_id = root_id;
-	if (key)
+	if (key) {
 		ref->key_for_search = *key;
-	else
+		/*
+		 * We can often find data backrefs with an offset that is too
+		 * large (>= LLONG_MAX, maximum allowed file offset) due to
+		 * underflows when subtracting a file's offset with the data
+		 * offset of its corresponding extent data item. This can
+		 * happen for example in the clone ioctl.
+		 * So if we detect such case we set the search key's offset to
+		 * zero to make sure we will find the matching file extent item
+		 * at add_all_parents(), otherwise we will miss it because the
+		 * offset taken form the backref is much larger then the offset
+		 * of the file extent item. This can make us scan a very large
+		 * number of file extent items, but at least it will not make
+		 * us miss any.
+		 * This is an ugly workaround for a behaviour that should have
+		 * never existed, but it does and a fix for the clone ioctl
+		 * would touch a lot of places, cause backwards incompatibility
+		 * and would not fix the problem for extents cloned with older
+		 * kernels.
+		 */
+		if (ref->key_for_search.type == BTRFS_EXTENT_DATA_KEY &&
+		    ref->key_for_search.offset >= LLONG_MAX)
+			ref->key_for_search.offset = 0;
+	} else {
 		memset(&ref->key_for_search, 0, sizeof(ref->key_for_search));
+	}
 
 	ref->inode_list = NULL;
 	ref->level = level;
@@ -250,8 +273,12 @@ static int add_all_parents(struct btrfs_root *root, struct btrfs_path *path,
 	 * the first item to check. But sometimes, we may enter it with
 	 * slot==nritems. In that case, go to the next leaf before we continue.
 	 */
-	if (path->slots[0] >= btrfs_header_nritems(path->nodes[0]))
-		ret = btrfs_next_old_leaf(root, path, time_seq);
+	if (path->slots[0] >= btrfs_header_nritems(path->nodes[0])) {
+		if (time_seq == (u64)-1)
+			ret = btrfs_next_leaf(root, path);
+		else
+			ret = btrfs_next_old_leaf(root, path, time_seq);
+	}
 
 	while (!ret && count < total_refs) {
 		eb = path->nodes[0];
@@ -291,7 +318,10 @@ static int add_all_parents(struct btrfs_root *root, struct btrfs_path *path,
 			eie = NULL;
 		}
 next:
-		ret = btrfs_next_old_item(root, path, time_seq);
+		if (time_seq == (u64)-1)
+			ret = btrfs_next_item(root, path);
+		else
+			ret = btrfs_next_old_item(root, path, time_seq);
 	}
 
 	if (ret > 0)
@@ -325,15 +355,23 @@ static int __resolve_indirect_ref(struct btrfs_fs_info *fs_info,
 
 	index = srcu_read_lock(&fs_info->subvol_srcu);
 
-	root = btrfs_read_fs_root_no_name(fs_info, &root_key);
+	root = btrfs_get_fs_root(fs_info, &root_key, false);
 	if (IS_ERR(root)) {
 		srcu_read_unlock(&fs_info->subvol_srcu, index);
 		ret = PTR_ERR(root);
 		goto out;
 	}
 
+	if (btrfs_test_is_dummy_root(root)) {
+		srcu_read_unlock(&fs_info->subvol_srcu, index);
+		ret = -ENOENT;
+		goto out;
+	}
+
 	if (path->search_commit_root)
 		root_level = btrfs_header_level(root->commit_root);
+	else if (time_seq == (u64)-1)
+		root_level = btrfs_header_level(root->node);
 	else
 		root_level = btrfs_old_root_level(root, time_seq);
 
@@ -343,7 +381,12 @@ static int __resolve_indirect_ref(struct btrfs_fs_info *fs_info,
 	}
 
 	path->lowest_level = level;
-	ret = btrfs_search_old_slot(root, &ref->key_for_search, path, time_seq);
+	if (time_seq == (u64)-1)
+		ret = btrfs_search_slot(NULL, root, &ref->key_for_search, path,
+					0, 0);
+	else
+		ret = btrfs_search_old_slot(root, &ref->key_for_search, path,
+					    time_seq);
 
 	/* root node has been locked, we can release @subvol_srcu safely here */
 	srcu_read_unlock(&fs_info->subvol_srcu, index);
@@ -491,7 +534,9 @@ static int __add_missing_keys(struct btrfs_fs_info *fs_info,
 		BUG_ON(!ref->wanted_disk_byte);
 		eb = read_tree_block(fs_info->tree_root, ref->wanted_disk_byte,
 				     0);
-		if (!eb || !extent_buffer_uptodate(eb)) {
+		if (IS_ERR(eb)) {
+			return PTR_ERR(eb);
+		} else if (!extent_buffer_uptodate(eb)) {
 			free_extent_buffer(eb);
 			return -EIO;
 		}
@@ -507,7 +552,7 @@ static int __add_missing_keys(struct btrfs_fs_info *fs_info,
 }
 
 /*
- * merge two lists of backrefs and adjust counts accordingly
+ * merge backrefs and adjust counts accordingly
  *
  * mode = 1: merge identical keys, if key is set
  *    FIXME: if we add more keys in __add_prelim_ref, we can merge more here.
@@ -535,9 +580,9 @@ static void __merge_refs(struct list_head *head, int mode)
 
 			ref2 = list_entry(pos2, struct __prelim_ref, list);
 
+			if (!ref_for_same_block(ref1, ref2))
+				continue;
 			if (mode == 1) {
-				if (!ref_for_same_block(ref1, ref2))
-					continue;
 				if (!ref1->parent && ref2->parent) {
 					xchg = ref1;
 					ref1 = ref2;
@@ -572,8 +617,8 @@ static int __add_delayed_refs(struct btrfs_delayed_ref_head *head, u64 seq,
 			      struct list_head *prefs, u64 *total_refs,
 			      u64 inum)
 {
+	struct btrfs_delayed_ref_node *node;
 	struct btrfs_delayed_extent_op *extent_op = head->extent_op;
-	struct rb_node *n = &head->node.rb_node;
 	struct btrfs_key key;
 	struct btrfs_key op_key = {0};
 	int sgn;
@@ -583,12 +628,7 @@ static int __add_delayed_refs(struct btrfs_delayed_ref_head *head, u64 seq,
 		btrfs_disk_key_to_cpu(&op_key, &extent_op->key);
 
 	spin_lock(&head->lock);
-	n = rb_first(&head->ref_root);
-	while (n) {
-		struct btrfs_delayed_ref_node *node;
-		node = rb_entry(n, struct btrfs_delayed_ref_node,
-				rb_node);
-		n = rb_next(n);
+	list_for_each_entry(node, &head->ref_list, list) {
 		if (node->seq > seq)
 			continue;
 
@@ -621,7 +661,7 @@ static int __add_delayed_refs(struct btrfs_delayed_ref_head *head, u64 seq,
 			struct btrfs_delayed_tree_ref *ref;
 
 			ref = btrfs_delayed_node_to_tree_ref(node);
-			ret = __add_prelim_ref(prefs, ref->root, NULL,
+			ret = __add_prelim_ref(prefs, 0, NULL,
 					       ref->level + 1, ref->parent,
 					       node->bytenr,
 					       node->ref_mod * sgn, GFP_ATOMIC);
@@ -653,11 +693,7 @@ static int __add_delayed_refs(struct btrfs_delayed_ref_head *head, u64 seq,
 			struct btrfs_delayed_data_ref *ref;
 
 			ref = btrfs_delayed_node_to_data_ref(node);
-
-			key.objectid = ref->objectid;
-			key.type = BTRFS_EXTENT_DATA_KEY;
-			key.offset = ref->offset;
-			ret = __add_prelim_ref(prefs, ref->root, &key, 0,
+			ret = __add_prelim_ref(prefs, 0, NULL, 0,
 					       ref->parent, node->bytenr,
 					       node->ref_mod * sgn, GFP_ATOMIC);
 			break;
@@ -882,6 +918,11 @@ static int __add_keyed_refs(struct btrfs_fs_info *fs_info,
  *
  * NOTE: This can return values > 0
  *
+ * If time_seq is set to (u64)-1, it will not search delayed_refs, and behave
+ * much like trans == NULL case, the difference only lies in it will not
+ * commit root.
+ * The special case is for qgroup to search roots in commit_transaction().
+ *
  * FIXME some caching might speed things up
  */
 static int find_parent_nodes(struct btrfs_trans_handle *trans,
@@ -920,6 +961,9 @@ static int find_parent_nodes(struct btrfs_trans_handle *trans,
 		path->skip_locking = 1;
 	}
 
+	if (time_seq == (u64)-1)
+		path->skip_locking = 1;
+
 	/*
 	 * grab both a lock on the path and a lock on the delayed ref head.
 	 * We need both to get a consistent picture of how the refs look
@@ -934,9 +978,10 @@ again:
 	BUG_ON(ret == 0);
 
 #ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS
-	if (trans && likely(trans->type != __TRANS_DUMMY)) {
+	if (trans && likely(trans->type != __TRANS_DUMMY) &&
+	    time_seq != (u64)-1) {
 #else
-	if (trans) {
+	if (trans && time_seq != (u64)-1) {
 #endif
 		/*
 		 * look if there are updates for this ref queued and lock the
@@ -1034,7 +1079,10 @@ again:
 
 				eb = read_tree_block(fs_info->extent_root,
 							   ref->parent, 0);
-				if (!eb || !extent_buffer_uptodate(eb)) {
+				if (IS_ERR(eb)) {
+					ret = PTR_ERR(eb);
+					goto out;
+				} else if (!extent_buffer_uptodate(eb)) {
 					free_extent_buffer(eb);
 					ret = -EIO;
 					goto out;
@@ -1369,7 +1417,8 @@ char *btrfs_ref_to_path(struct btrfs_root *fs_root, struct btrfs_path *path,
 			read_extent_buffer(eb, dest + bytes_left,
 					   name_off, name_len);
 		if (eb != eb_in) {
-			btrfs_tree_read_unlock_blocking(eb);
+			if (!path->skip_locking)
+				btrfs_tree_read_unlock_blocking(eb);
 			free_extent_buffer(eb);
 		}
 		ret = btrfs_find_item(fs_root, path, parent, 0,
@@ -1389,9 +1438,10 @@ char *btrfs_ref_to_path(struct btrfs_root *fs_root, struct btrfs_path *path,
 		eb = path->nodes[0];
 		/* make sure we can use eb after releasing the path */
 		if (eb != eb_in) {
-			atomic_inc(&eb->refs);
-			btrfs_tree_read_lock(eb);
-			btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK);
+			if (!path->skip_locking)
+				btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK);
+			path->nodes[0] = NULL;
+			path->locks[0] = 0;
 		}
 		btrfs_release_path(path);
 		iref = btrfs_item_ptr(eb, slot, struct btrfs_inode_ref);
@@ -1786,7 +1836,6 @@ static int iterate_inode_extrefs(u64 inum, struct btrfs_root *fs_root,
 	int found = 0;
 	struct extent_buffer *eb;
 	struct btrfs_inode_extref *extref;
-	struct extent_buffer *leaf;
 	u32 item_size;
 	u32 cur_offset;
 	unsigned long ptr;
@@ -1814,9 +1863,8 @@ static int iterate_inode_extrefs(u64 inum, struct btrfs_root *fs_root,
 		btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK);
 		btrfs_release_path(path);
 
-		leaf = path->nodes[0];
-		item_size = btrfs_item_size_nr(leaf, slot);
-		ptr = btrfs_item_ptr_offset(leaf, slot);
+		item_size = btrfs_item_size_nr(eb, slot);
+		ptr = btrfs_item_ptr_offset(eb, slot);
 		cur_offset = 0;
 
 		while (cur_offset < item_size) {
@@ -1830,7 +1878,7 @@ static int iterate_inode_extrefs(u64 inum, struct btrfs_root *fs_root,
 			if (ret)
 				break;
 
-			cur_offset += btrfs_inode_extref_name_len(leaf, extref);
+			cur_offset += btrfs_inode_extref_name_len(eb, extref);
 			cur_offset += sizeof(*extref);
 		}
 		btrfs_tree_read_unlock_blocking(eb);
diff --git a/kernel/fs/btrfs/check-integrity.c b/kernel/fs/btrfs/check-integrity.c
index ce7dec88f..0340c57bf 100644
--- a/kernel/fs/btrfs/check-integrity.c
+++ b/kernel/fs/btrfs/check-integrity.c
@@ -343,7 +343,7 @@ static int btrfsic_process_written_superblock(
 		struct btrfsic_state *state,
 		struct btrfsic_block *const block,
 		struct btrfs_super_block *const super_hdr);
-static void btrfsic_bio_end_io(struct bio *bp, int bio_error_status);
+static void btrfsic_bio_end_io(struct bio *bp);
 static void btrfsic_bh_end_io(struct buffer_head *bh, int uptodate);
 static int btrfsic_is_block_ref_by_superblock(const struct btrfsic_state *state,
 					      const struct btrfsic_block *block,
@@ -667,7 +667,7 @@ static int btrfsic_process_superblock(struct btrfsic_state *state,
 	selected_super = kzalloc(sizeof(*selected_super), GFP_NOFS);
 	if (NULL == selected_super) {
 		printk(KERN_INFO "btrfsic: error, kmalloc failed!\n");
-		return -1;
+		return -ENOMEM;
 	}
 
 	list_for_each_entry(device, dev_head, dev_list) {
@@ -845,8 +845,8 @@ static int btrfsic_process_superblock_dev_mirror(
 		superblock_tmp->never_written = 0;
 		superblock_tmp->mirror_num = 1 + superblock_mirror_num;
 		if (state->print_mask & BTRFSIC_PRINT_MASK_SUPERBLOCK_WRITE)
-			printk_in_rcu(KERN_INFO "New initial S-block (bdev %p, %s)"
-				     " @%llu (%s/%llu/%d)\n",
+			btrfs_info_in_rcu(device->dev_root->fs_info,
+				"new initial S-block (bdev %p, %s) @%llu (%s/%llu/%d)",
 				     superblock_bdev,
 				     rcu_str_deref(device->name), dev_bytenr,
 				     dev_state->name, dev_bytenr,
@@ -1660,7 +1660,7 @@ static int btrfsic_read_block(struct btrfsic_state *state,
 					  sizeof(*block_ctx->pagev)) *
 					 num_pages, GFP_NOFS);
 	if (!block_ctx->mem_to_free)
-		return -1;
+		return -ENOMEM;
 	block_ctx->datav = block_ctx->mem_to_free;
 	block_ctx->pagev = (struct page **)(block_ctx->datav + num_pages);
 	for (i = 0; i < num_pages; i++) {
@@ -2207,7 +2207,7 @@ continue_loop:
 	goto again;
 }
 
-static void btrfsic_bio_end_io(struct bio *bp, int bio_error_status)
+static void btrfsic_bio_end_io(struct bio *bp)
 {
 	struct btrfsic_block *block = (struct btrfsic_block *)bp->bi_private;
 	int iodone_w_error;
@@ -2215,7 +2215,7 @@ static void btrfsic_bio_end_io(struct bio *bp, int bio_error_status)
 	/* mutex is not held! This is not save if IO is not yet completed
 	 * on umount */
 	iodone_w_error = 0;
-	if (bio_error_status)
+	if (bp->bi_error)
 		iodone_w_error = 1;
 
 	BUG_ON(NULL == block);
@@ -2230,7 +2230,7 @@ static void btrfsic_bio_end_io(struct bio *bp, int bio_error_status)
 		     BTRFSIC_PRINT_MASK_END_IO_BIO_BH))
 			printk(KERN_INFO
 			       "bio_end_io(err=%d) for %c @%llu (%s/%llu/%d)\n",
-			       bio_error_status,
+			       bp->bi_error,
 			       btrfsic_get_block_type(dev_state->state, block),
 			       block->logical_bytenr, dev_state->name,
 			       block->dev_bytenr, block->mirror_num);
@@ -2252,7 +2252,7 @@ static void btrfsic_bio_end_io(struct bio *bp, int bio_error_status)
 		block = next_block;
 	} while (NULL != block);
 
-	bp->bi_end_io(bp, bio_error_status);
+	bp->bi_end_io(bp);
 }
 
 static void btrfsic_bh_end_io(struct buffer_head *bh, int uptodate)
diff --git a/kernel/fs/btrfs/compression.c b/kernel/fs/btrfs/compression.c
index ce62324c7..c473c42d7 100644
--- a/kernel/fs/btrfs/compression.c
+++ b/kernel/fs/btrfs/compression.c
@@ -97,10 +97,7 @@ static inline int compressed_bio_size(struct btrfs_root *root,
 static struct bio *compressed_bio_alloc(struct block_device *bdev,
 					u64 first_byte, gfp_t gfp_flags)
 {
-	int nr_vecs;
-
-	nr_vecs = bio_get_nr_vecs(bdev);
-	return btrfs_bio_alloc(bdev, first_byte >> 9, nr_vecs, gfp_flags);
+	return btrfs_bio_alloc(bdev, first_byte >> 9, BIO_MAX_PAGES, gfp_flags);
 }
 
 static int check_compressed_csum(struct inode *inode,
@@ -152,7 +149,7 @@ fail:
  * The compressed pages are freed here, and it must be run
  * in process context
  */
-static void end_compressed_bio_read(struct bio *bio, int err)
+static void end_compressed_bio_read(struct bio *bio)
 {
 	struct compressed_bio *cb = bio->bi_private;
 	struct inode *inode;
@@ -160,7 +157,7 @@ static void end_compressed_bio_read(struct bio *bio, int err)
 	unsigned long index;
 	int ret;
 
-	if (err)
+	if (bio->bi_error)
 		cb->errors = 1;
 
 	/* if there are more bios still pending for this compressed
@@ -210,7 +207,7 @@ csum_failed:
 		bio_for_each_segment_all(bvec, cb->orig_bio, i)
 			SetPageChecked(bvec->bv_page);
 
-		bio_endio(cb->orig_bio, 0);
+		bio_endio(cb->orig_bio);
 	}
 
 	/* finally free the cb struct */
@@ -266,7 +263,7 @@ static noinline void end_compressed_writeback(struct inode *inode,
  * This also calls the writeback end hooks for the file pages so that
  * metadata and checksums can be updated in the file.
  */
-static void end_compressed_bio_write(struct bio *bio, int err)
+static void end_compressed_bio_write(struct bio *bio)
 {
 	struct extent_io_tree *tree;
 	struct compressed_bio *cb = bio->bi_private;
@@ -274,7 +271,7 @@ static void end_compressed_bio_write(struct bio *bio, int err)
 	struct page *page;
 	unsigned long index;
 
-	if (err)
+	if (bio->bi_error)
 		cb->errors = 1;
 
 	/* if there are more bios still pending for this compressed
@@ -293,7 +290,7 @@ static void end_compressed_bio_write(struct bio *bio, int err)
 					 cb->start,
 					 cb->start + cb->len - 1,
 					 NULL,
-					 err ? 0 : 1);
+					 bio->bi_error ? 0 : 1);
 	cb->compressed_pages[0]->mapping = NULL;
 
 	end_compressed_writeback(inode, cb);
@@ -485,13 +482,12 @@ static noinline int add_ra_bio_pages(struct inode *inode,
 			goto next;
 		}
 
-		page = __page_cache_alloc(mapping_gfp_mask(mapping) &
-								~__GFP_FS);
+		page = __page_cache_alloc(mapping_gfp_constraint(mapping,
+								 ~__GFP_FS));
 		if (!page)
 			break;
 
-		if (add_to_page_cache_lru(page, mapping, pg_index,
-								GFP_NOFS)) {
+		if (add_to_page_cache_lru(page, mapping, pg_index, GFP_NOFS)) {
 			page_cache_release(page);
 			goto next;
 		}
@@ -697,8 +693,10 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio,
 
 			ret = btrfs_map_bio(root, READ, comp_bio,
 					    mirror_num, 0);
-			if (ret)
-				bio_endio(comp_bio, ret);
+			if (ret) {
+				bio->bi_error = ret;
+				bio_endio(comp_bio);
+			}
 
 			bio_put(comp_bio);
 
@@ -724,8 +722,10 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio,
 	}
 
 	ret = btrfs_map_bio(root, READ, comp_bio, mirror_num, 0);
-	if (ret)
-		bio_endio(comp_bio, ret);
+	if (ret) {
+		bio->bi_error = ret;
+		bio_endio(comp_bio);
+	}
 
 	bio_put(comp_bio);
 	return 0;
@@ -744,11 +744,13 @@ out:
 	return ret;
 }
 
-static struct list_head comp_idle_workspace[BTRFS_COMPRESS_TYPES];
-static spinlock_t comp_workspace_lock[BTRFS_COMPRESS_TYPES];
-static int comp_num_workspace[BTRFS_COMPRESS_TYPES];
-static atomic_t comp_alloc_workspace[BTRFS_COMPRESS_TYPES];
-static wait_queue_head_t comp_workspace_wait[BTRFS_COMPRESS_TYPES];
+static struct {
+	struct list_head idle_ws;
+	spinlock_t ws_lock;
+	int num_ws;
+	atomic_t alloc_ws;
+	wait_queue_head_t ws_wait;
+} btrfs_comp_ws[BTRFS_COMPRESS_TYPES];
 
 static const struct btrfs_compress_op * const btrfs_compress_op[] = {
 	&btrfs_zlib_compress,
@@ -760,10 +762,10 @@ void __init btrfs_init_compress(void)
 	int i;
 
 	for (i = 0; i < BTRFS_COMPRESS_TYPES; i++) {
-		INIT_LIST_HEAD(&comp_idle_workspace[i]);
-		spin_lock_init(&comp_workspace_lock[i]);
-		atomic_set(&comp_alloc_workspace[i], 0);
-		init_waitqueue_head(&comp_workspace_wait[i]);
+		INIT_LIST_HEAD(&btrfs_comp_ws[i].idle_ws);
+		spin_lock_init(&btrfs_comp_ws[i].ws_lock);
+		atomic_set(&btrfs_comp_ws[i].alloc_ws, 0);
+		init_waitqueue_head(&btrfs_comp_ws[i].ws_wait);
 	}
 }
 
@@ -777,38 +779,38 @@ static struct list_head *find_workspace(int type)
 	int cpus = num_online_cpus();
 	int idx = type - 1;
 
-	struct list_head *idle_workspace	= &comp_idle_workspace[idx];
-	spinlock_t *workspace_lock		= &comp_workspace_lock[idx];
-	atomic_t *alloc_workspace		= &comp_alloc_workspace[idx];
-	wait_queue_head_t *workspace_wait	= &comp_workspace_wait[idx];
-	int *num_workspace			= &comp_num_workspace[idx];
+	struct list_head *idle_ws	= &btrfs_comp_ws[idx].idle_ws;
+	spinlock_t *ws_lock		= &btrfs_comp_ws[idx].ws_lock;
+	atomic_t *alloc_ws		= &btrfs_comp_ws[idx].alloc_ws;
+	wait_queue_head_t *ws_wait	= &btrfs_comp_ws[idx].ws_wait;
+	int *num_ws			= &btrfs_comp_ws[idx].num_ws;
 again:
-	spin_lock(workspace_lock);
-	if (!list_empty(idle_workspace)) {
-		workspace = idle_workspace->next;
+	spin_lock(ws_lock);
+	if (!list_empty(idle_ws)) {
+		workspace = idle_ws->next;
 		list_del(workspace);
-		(*num_workspace)--;
-		spin_unlock(workspace_lock);
+		(*num_ws)--;
+		spin_unlock(ws_lock);
 		return workspace;
 
 	}
-	if (atomic_read(alloc_workspace) > cpus) {
+	if (atomic_read(alloc_ws) > cpus) {
 		DEFINE_WAIT(wait);
 
-		spin_unlock(workspace_lock);
-		prepare_to_wait(workspace_wait, &wait, TASK_UNINTERRUPTIBLE);
-		if (atomic_read(alloc_workspace) > cpus && !*num_workspace)
+		spin_unlock(ws_lock);
+		prepare_to_wait(ws_wait, &wait, TASK_UNINTERRUPTIBLE);
+		if (atomic_read(alloc_ws) > cpus && !*num_ws)
 			schedule();
-		finish_wait(workspace_wait, &wait);
+		finish_wait(ws_wait, &wait);
 		goto again;
 	}
-	atomic_inc(alloc_workspace);
-	spin_unlock(workspace_lock);
+	atomic_inc(alloc_ws);
+	spin_unlock(ws_lock);
 
 	workspace = btrfs_compress_op[idx]->alloc_workspace();
 	if (IS_ERR(workspace)) {
-		atomic_dec(alloc_workspace);
-		wake_up(workspace_wait);
+		atomic_dec(alloc_ws);
+		wake_up(ws_wait);
 	}
 	return workspace;
 }
@@ -820,27 +822,30 @@ again:
 static void free_workspace(int type, struct list_head *workspace)
 {
 	int idx = type - 1;
-	struct list_head *idle_workspace	= &comp_idle_workspace[idx];
-	spinlock_t *workspace_lock		= &comp_workspace_lock[idx];
-	atomic_t *alloc_workspace		= &comp_alloc_workspace[idx];
-	wait_queue_head_t *workspace_wait	= &comp_workspace_wait[idx];
-	int *num_workspace			= &comp_num_workspace[idx];
-
-	spin_lock(workspace_lock);
-	if (*num_workspace < num_online_cpus()) {
-		list_add(workspace, idle_workspace);
-		(*num_workspace)++;
-		spin_unlock(workspace_lock);
+	struct list_head *idle_ws	= &btrfs_comp_ws[idx].idle_ws;
+	spinlock_t *ws_lock		= &btrfs_comp_ws[idx].ws_lock;
+	atomic_t *alloc_ws		= &btrfs_comp_ws[idx].alloc_ws;
+	wait_queue_head_t *ws_wait	= &btrfs_comp_ws[idx].ws_wait;
+	int *num_ws			= &btrfs_comp_ws[idx].num_ws;
+
+	spin_lock(ws_lock);
+	if (*num_ws < num_online_cpus()) {
+		list_add(workspace, idle_ws);
+		(*num_ws)++;
+		spin_unlock(ws_lock);
 		goto wake;
 	}
-	spin_unlock(workspace_lock);
+	spin_unlock(ws_lock);
 
 	btrfs_compress_op[idx]->free_workspace(workspace);
-	atomic_dec(alloc_workspace);
+	atomic_dec(alloc_ws);
 wake:
+	/*
+	 * Make sure counter is updated before we wake up waiters.
+	 */
 	smp_mb();
-	if (waitqueue_active(workspace_wait))
-		wake_up(workspace_wait);
+	if (waitqueue_active(ws_wait))
+		wake_up(ws_wait);
 }
 
 /*
@@ -852,11 +857,11 @@ static void free_workspaces(void)
 	int i;
 
 	for (i = 0; i < BTRFS_COMPRESS_TYPES; i++) {
-		while (!list_empty(&comp_idle_workspace[i])) {
-			workspace = comp_idle_workspace[i].next;
+		while (!list_empty(&btrfs_comp_ws[i].idle_ws)) {
+			workspace = btrfs_comp_ws[i].idle_ws.next;
 			list_del(workspace);
 			btrfs_compress_op[i]->free_workspace(workspace);
-			atomic_dec(&comp_alloc_workspace[i]);
+			atomic_dec(&btrfs_comp_ws[i].alloc_ws);
 		}
 	}
 }
diff --git a/kernel/fs/btrfs/ctree.c b/kernel/fs/btrfs/ctree.c
index 0f11ebc92..5b8e235c4 100644
--- a/kernel/fs/btrfs/ctree.c
+++ b/kernel/fs/btrfs/ctree.c
@@ -1011,7 +1011,7 @@ static noinline int update_ref_for_cow(struct btrfs_trans_handle *trans,
 			return ret;
 		if (refs == 0) {
 			ret = -EROFS;
-			btrfs_std_error(root->fs_info, ret);
+			btrfs_std_error(root->fs_info, ret, NULL);
 			return ret;
 		}
 	} else {
@@ -1159,8 +1159,10 @@ static noinline int __btrfs_cow_block(struct btrfs_trans_handle *trans,
 
 	if (test_bit(BTRFS_ROOT_REF_COWS, &root->state)) {
 		ret = btrfs_reloc_cow_block(trans, root, buf, cow);
-		if (ret)
+		if (ret) {
+			btrfs_abort_transaction(trans, root, ret);
 			return ret;
+		}
 	}
 
 	if (buf == root->node) {
@@ -1439,8 +1441,9 @@ get_old_root(struct btrfs_root *root, u64 time_seq)
 		btrfs_tree_read_unlock(eb_root);
 		free_extent_buffer(eb_root);
 		old = read_tree_block(root, logical, 0);
-		if (WARN_ON(!old || !extent_buffer_uptodate(old))) {
-			free_extent_buffer(old);
+		if (WARN_ON(IS_ERR(old) || !extent_buffer_uptodate(old))) {
+			if (!IS_ERR(old))
+				free_extent_buffer(old);
 			btrfs_warn(root->fs_info,
 				"failed to read tree block %llu from get_old_root", logical);
 		} else {
@@ -1685,7 +1688,9 @@ int btrfs_realloc_node(struct btrfs_trans_handle *trans,
 		if (!cur || !uptodate) {
 			if (!cur) {
 				cur = read_tree_block(root, blocknr, gen);
-				if (!cur || !extent_buffer_uptodate(cur)) {
+				if (IS_ERR(cur)) {
+					return PTR_ERR(cur);
+				} else if (!extent_buffer_uptodate(cur)) {
 					free_extent_buffer(cur);
 					return -EIO;
 				}
@@ -1864,8 +1869,9 @@ static noinline struct extent_buffer *read_node_slot(struct btrfs_root *root,
 
 	eb = read_tree_block(root, btrfs_node_blockptr(parent, slot),
 			     btrfs_node_ptr_generation(parent, slot));
-	if (eb && !extent_buffer_uptodate(eb)) {
-		free_extent_buffer(eb);
+	if (IS_ERR(eb) || !extent_buffer_uptodate(eb)) {
+		if (!IS_ERR(eb))
+			free_extent_buffer(eb);
 		eb = NULL;
 	}
 
@@ -1921,7 +1927,7 @@ static noinline int balance_level(struct btrfs_trans_handle *trans,
 		child = read_node_slot(root, mid, 0);
 		if (!child) {
 			ret = -EROFS;
-			btrfs_std_error(root->fs_info, ret);
+			btrfs_std_error(root->fs_info, ret, NULL);
 			goto enospc;
 		}
 
@@ -2024,7 +2030,7 @@ static noinline int balance_level(struct btrfs_trans_handle *trans,
 		 */
 		if (!left) {
 			ret = -EROFS;
-			btrfs_std_error(root->fs_info, ret);
+			btrfs_std_error(root->fs_info, ret, NULL);
 			goto enospc;
 		}
 		wret = balance_node_right(trans, root, mid, left);
@@ -2494,7 +2500,7 @@ read_block_for_search(struct btrfs_trans_handle *trans,
 
 	ret = -EAGAIN;
 	tmp = read_tree_block(root, blocknr, 0);
-	if (tmp) {
+	if (!IS_ERR(tmp)) {
 		/*
 		 * If the read above didn't mark this buffer up to date,
 		 * it will never end up being up to date.  Set ret to EIO now
@@ -4934,8 +4940,8 @@ int btrfs_del_items(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 {
 	struct extent_buffer *leaf;
 	struct btrfs_item *item;
-	int last_off;
-	int dsize = 0;
+	u32 last_off;
+	u32 dsize = 0;
 	int ret = 0;
 	int wret;
 	int i;
diff --git a/kernel/fs/btrfs/ctree.h b/kernel/fs/btrfs/ctree.h
index 6f364e1d8..385b449fd 100644
--- a/kernel/fs/btrfs/ctree.h
+++ b/kernel/fs/btrfs/ctree.h
@@ -174,7 +174,7 @@ struct btrfs_ordered_sum;
 /* csum types */
 #define BTRFS_CSUM_TYPE_CRC32	0
 
-static int btrfs_csum_sizes[] = { 4, 0 };
+static int btrfs_csum_sizes[] = { 4 };
 
 /* four bytes for CRC32 */
 #define BTRFS_EMPTY_DIR_SIZE 0
@@ -823,8 +823,18 @@ struct btrfs_disk_balance_args {
 	 */
 	__le64 profiles;
 
-	/* usage filter */
-	__le64 usage;
+	/*
+	 * usage filter
+	 * BTRFS_BALANCE_ARGS_USAGE with a single value means '0..N'
+	 * BTRFS_BALANCE_ARGS_USAGE_RANGE - range syntax, min..max
+	 */
+	union {
+		__le64 usage;
+		struct {
+			__le32 usage_min;
+			__le32 usage_max;
+		};
+	};
 
 	/* devid filter */
 	__le64 devid;
@@ -846,10 +856,27 @@ struct btrfs_disk_balance_args {
 	/* BTRFS_BALANCE_ARGS_* */
 	__le64 flags;
 
-	/* BTRFS_BALANCE_ARGS_LIMIT value */
-	__le64 limit;
+	/*
+	 * BTRFS_BALANCE_ARGS_LIMIT with value 'limit'
+	 * BTRFS_BALANCE_ARGS_LIMIT_RANGE - the extend version can use minimum
+	 * and maximum
+	 */
+	union {
+		__le64 limit;
+		struct {
+			__le32 limit_min;
+			__le32 limit_max;
+		};
+	};
+
+	/*
+	 * Process chunks that cross stripes_min..stripes_max devices,
+	 * BTRFS_BALANCE_ARGS_STRIPES_RANGE
+	 */
+	__le32 stripes_min;
+	__le32 stripes_max;
 
-	__le64 unused[7];
+	__le64 unused[6];
 } __attribute__ ((__packed__));
 
 /*
@@ -1154,6 +1181,10 @@ struct btrfs_space_info {
 				   delalloc/allocations */
 	u64 bytes_readonly;	/* total bytes that are read only */
 
+	u64 max_extent_size;	/* This will hold the maximum extent size of
+				   the space info if we had an ENOSPC in the
+				   allocator. */
+
 	unsigned int full:1;	/* indicates that we cannot allocate any more
 				   chunks for this space */
 	unsigned int chunk_alloc:1;	/* set if we are allocating a chunk */
@@ -1228,6 +1259,9 @@ struct btrfs_free_cluster {
 	/* first extent starting offset */
 	u64 window_start;
 
+	/* We did a full search and couldn't create a cluster */
+	bool fragmented;
+
 	struct btrfs_block_group_cache *block_group;
 	/*
 	 * when a cluster is allocated from a block group, we put the
@@ -1300,7 +1334,7 @@ struct btrfs_block_group_cache {
 	/* for raid56, this is a full stripe, without parity */
 	unsigned long full_stripe_len;
 
-	unsigned int ro:1;
+	unsigned int ro;
 	unsigned int iref:1;
 	unsigned int has_caching_ctl:1;
 	unsigned int removed:1;
@@ -1518,12 +1552,6 @@ struct btrfs_fs_info {
 	 */
 	struct mutex ordered_operations_mutex;
 
-	/*
-	 * Same as ordered_operations_mutex except this is for ordered extents
-	 * and not the operations.
-	 */
-	struct mutex ordered_extent_flush_mutex;
-
 	struct rw_semaphore commit_root_sem;
 
 	struct rw_semaphore cleanup_work_sem;
@@ -1544,7 +1572,7 @@ struct btrfs_fs_info {
 
 	spinlock_t delayed_iput_lock;
 	struct list_head delayed_iputs;
-	struct rw_semaphore delayed_iput_sem;
+	struct mutex cleaner_delayed_iput_mutex;
 
 	/* this protects tree_mod_seq_list */
 	spinlock_t tree_mod_seq_lock;
@@ -1619,10 +1647,7 @@ struct btrfs_fs_info {
 	struct task_struct *cleaner_kthread;
 	int thread_pool_size;
 
-	struct kobject super_kobj;
 	struct kobject *space_info_kobj;
-	struct kobject *device_dir_kobj;
-	struct completion kobj_unregister;
 	int do_barriers;
 	int closing;
 	int log_root_recovering;
@@ -1698,6 +1723,7 @@ struct btrfs_fs_info {
 	struct btrfs_workqueue *scrub_workers;
 	struct btrfs_workqueue *scrub_wr_completion_workers;
 	struct btrfs_workqueue *scrub_nocow_workers;
+	struct btrfs_workqueue *scrub_parity_workers;
 
 #ifdef CONFIG_BTRFS_FS_CHECK_INTEGRITY
 	u32 check_integrity_print_mask;
@@ -1735,7 +1761,7 @@ struct btrfs_fs_info {
 	/* list of dirty qgroups to be written at next commit */
 	struct list_head dirty_qgroups;
 
-	/* used by btrfs_qgroup_record_ref for an efficient tree traversal */
+	/* used by qgroup for an efficient tree traversal */
 	u64 qgroup_seq;
 
 	/* qgroup rescan items */
@@ -1780,6 +1806,7 @@ struct btrfs_fs_info {
 	spinlock_t unused_bgs_lock;
 	struct list_head unused_bgs;
 	struct mutex unused_bg_unpin_mutex;
+	struct mutex delete_unused_bgs_mutex;
 
 	/* For btrfs to record security options */
 	struct security_mnt_opts security_opts;
@@ -1950,6 +1977,9 @@ struct btrfs_root {
 	int send_in_progress;
 	struct btrfs_subvolume_writers *subv_writers;
 	atomic_t will_be_snapshoted;
+
+	/* For qgroup metadata space reserve */
+	atomic_t qgroup_meta_rsv;
 };
 
 struct btrfs_ioctl_defrag_range_args {
@@ -2152,6 +2182,8 @@ struct btrfs_ioctl_defrag_range_args {
 #define BTRFS_MOUNT_CHECK_INTEGRITY_INCLUDING_EXTENT_DATA (1 << 21)
 #define BTRFS_MOUNT_PANIC_ON_FATAL_ERROR	(1 << 22)
 #define BTRFS_MOUNT_RESCAN_UUID_TREE	(1 << 23)
+#define BTRFS_MOUNT_FRAGMENT_DATA	(1 << 24)
+#define BTRFS_MOUNT_FRAGMENT_METADATA	(1 << 25)
 
 #define BTRFS_DEFAULT_COMMIT_INTERVAL	(30)
 #define BTRFS_DEFAULT_MAX_INLINE	(8192)
@@ -2176,6 +2208,18 @@ struct btrfs_ioctl_defrag_range_args {
 	btrfs_clear_opt(root->fs_info->mount_opt, opt);			\
 }
 
+#ifdef CONFIG_BTRFS_DEBUG
+static inline int
+btrfs_should_fragment_free_space(struct btrfs_root *root,
+				 struct btrfs_block_group_cache *block_group)
+{
+	return (btrfs_test_opt(root, FRAGMENT_METADATA) &&
+		block_group->flags & BTRFS_BLOCK_GROUP_METADATA) ||
+	       (btrfs_test_opt(root, FRAGMENT_DATA) &&
+		block_group->flags &  BTRFS_BLOCK_GROUP_DATA);
+}
+#endif
+
 /*
  * Requests for changes that need to be done during transaction commit.
  *
@@ -3323,7 +3367,7 @@ static inline bool btrfs_mixed_space_info(struct btrfs_space_info *space_info)
 
 static inline gfp_t btrfs_alloc_write_mask(struct address_space *mapping)
 {
-	return mapping_gfp_mask(mapping) & ~__GFP_FS;
+	return mapping_gfp_constraint(mapping, ~__GFP_FS);
 }
 
 /* extent-tree.c */
@@ -3372,6 +3416,7 @@ int btrfs_cross_ref_exist(struct btrfs_trans_handle *trans,
 struct btrfs_block_group_cache *btrfs_lookup_block_group(
 						 struct btrfs_fs_info *info,
 						 u64 bytenr);
+void btrfs_get_block_group(struct btrfs_block_group_cache *cache);
 void btrfs_put_block_group(struct btrfs_block_group_cache *cache);
 int get_block_group_index(struct btrfs_block_group_cache *cache);
 struct extent_buffer *btrfs_alloc_tree_block(struct btrfs_trans_handle *trans,
@@ -3386,7 +3431,8 @@ void btrfs_free_tree_block(struct btrfs_trans_handle *trans,
 int btrfs_alloc_reserved_file_extent(struct btrfs_trans_handle *trans,
 				     struct btrfs_root *root,
 				     u64 root_objectid, u64 owner,
-				     u64 offset, struct btrfs_key *ins);
+				     u64 offset, u64 ram_bytes,
+				     struct btrfs_key *ins);
 int btrfs_alloc_logged_file_extent(struct btrfs_trans_handle *trans,
 				   struct btrfs_root *root,
 				   u64 root_objectid, u64 owner, u64 offset,
@@ -3405,7 +3451,7 @@ int btrfs_set_disk_extent_flags(struct btrfs_trans_handle *trans,
 int btrfs_free_extent(struct btrfs_trans_handle *trans,
 		      struct btrfs_root *root,
 		      u64 bytenr, u64 num_bytes, u64 parent, u64 root_objectid,
-		      u64 owner, u64 offset, int no_quota);
+		      u64 owner, u64 offset);
 
 int btrfs_free_reserved_extent(struct btrfs_root *root, u64 start, u64 len,
 			       int delalloc);
@@ -3418,7 +3464,7 @@ int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans,
 int btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
 			 struct btrfs_root *root,
 			 u64 bytenr, u64 num_bytes, u64 parent,
-			 u64 root_objectid, u64 owner, u64 offset, int no_quota);
+			 u64 root_objectid, u64 owner, u64 offset);
 
 int btrfs_start_dirty_block_groups(struct btrfs_trans_handle *trans,
 				   struct btrfs_root *root);
@@ -3434,10 +3480,15 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans,
 			   struct btrfs_root *root, u64 bytes_used,
 			   u64 type, u64 chunk_objectid, u64 chunk_offset,
 			   u64 size);
+struct btrfs_trans_handle *btrfs_start_trans_remove_block_group(
+				struct btrfs_fs_info *fs_info,
+				const u64 chunk_offset);
 int btrfs_remove_block_group(struct btrfs_trans_handle *trans,
 			     struct btrfs_root *root, u64 group_start,
 			     struct extent_map *em);
 void btrfs_delete_unused_bgs(struct btrfs_fs_info *fs_info);
+void btrfs_get_block_group_trimming(struct btrfs_block_group_cache *cache);
+void btrfs_put_block_group_trimming(struct btrfs_block_group_cache *cache);
 void btrfs_create_pending_block_groups(struct btrfs_trans_handle *trans,
 				       struct btrfs_root *root);
 u64 btrfs_get_alloc_profile(struct btrfs_root *root, int data);
@@ -3454,10 +3505,14 @@ enum btrfs_reserve_flush_enum {
 	BTRFS_RESERVE_FLUSH_ALL,
 };
 
-int btrfs_check_data_free_space(struct inode *inode, u64 bytes, u64 write_bytes);
-void btrfs_free_reserved_data_space(struct inode *inode, u64 bytes);
+int btrfs_check_data_free_space(struct inode *inode, u64 start, u64 len);
+int btrfs_alloc_data_chunk_ondemand(struct inode *inode, u64 bytes);
+void btrfs_free_reserved_data_space(struct inode *inode, u64 start, u64 len);
+void btrfs_free_reserved_data_space_noquota(struct inode *inode, u64 start,
+					    u64 len);
 void btrfs_trans_release_metadata(struct btrfs_trans_handle *trans,
 				struct btrfs_root *root);
+void btrfs_trans_release_chunk_metadata(struct btrfs_trans_handle *trans);
 int btrfs_orphan_reserve_metadata(struct btrfs_trans_handle *trans,
 				  struct inode *inode);
 void btrfs_orphan_release_metadata(struct inode *inode);
@@ -3470,8 +3525,8 @@ void btrfs_subvolume_release_metadata(struct btrfs_root *root,
 				      u64 qgroup_reserved);
 int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes);
 void btrfs_delalloc_release_metadata(struct inode *inode, u64 num_bytes);
-int btrfs_delalloc_reserve_space(struct inode *inode, u64 num_bytes);
-void btrfs_delalloc_release_space(struct inode *inode, u64 num_bytes);
+int btrfs_delalloc_reserve_space(struct inode *inode, u64 start, u64 len);
+void btrfs_delalloc_release_space(struct inode *inode, u64 start, u64 len);
 void btrfs_init_block_rsv(struct btrfs_block_rsv *rsv, unsigned short type);
 struct btrfs_block_rsv *btrfs_alloc_block_rsv(struct btrfs_root *root,
 					      unsigned short type);
@@ -3495,9 +3550,9 @@ int btrfs_cond_migrate_bytes(struct btrfs_fs_info *fs_info,
 void btrfs_block_rsv_release(struct btrfs_root *root,
 			     struct btrfs_block_rsv *block_rsv,
 			     u64 num_bytes);
-int btrfs_set_block_group_ro(struct btrfs_root *root,
+int btrfs_inc_block_group_ro(struct btrfs_root *root,
 			     struct btrfs_block_group_cache *cache);
-void btrfs_set_block_group_rw(struct btrfs_root *root,
+void btrfs_dec_block_group_ro(struct btrfs_root *root,
 			      struct btrfs_block_group_cache *cache);
 void btrfs_put_block_group_cache(struct btrfs_fs_info *info);
 u64 btrfs_account_ro_block_groups_free_space(struct btrfs_space_info *sinfo);
@@ -3515,6 +3570,9 @@ int btrfs_delayed_refs_qgroup_accounting(struct btrfs_trans_handle *trans,
 int __get_raid_index(u64 flags);
 int btrfs_start_write_no_snapshoting(struct btrfs_root *root);
 void btrfs_end_write_no_snapshoting(struct btrfs_root *root);
+void check_system_chunk(struct btrfs_trans_handle *trans,
+			struct btrfs_root *root,
+			const u64 type);
 /* ctree.c */
 int btrfs_bin_search(struct extent_buffer *eb, struct btrfs_key *key,
 		     int level, int *slot);
@@ -4005,8 +4063,8 @@ int btrfs_defrag_leaves(struct btrfs_trans_handle *trans,
 /* sysfs.c */
 int btrfs_init_sysfs(void);
 void btrfs_exit_sysfs(void);
-int btrfs_sysfs_add_one(struct btrfs_fs_info *fs_info);
-void btrfs_sysfs_remove_one(struct btrfs_fs_info *fs_info);
+int btrfs_sysfs_add_mounted(struct btrfs_fs_info *fs_info);
+void btrfs_sysfs_remove_mounted(struct btrfs_fs_info *fs_info);
 
 /* xattr.c */
 ssize_t btrfs_listxattr(struct dentry *dentry, char *buffer, size_t size);
@@ -4040,16 +4098,105 @@ void btrfs_printk(const struct btrfs_fs_info *fs_info, const char *fmt, ...)
 #define btrfs_info(fs_info, fmt, args...) \
 	btrfs_printk(fs_info, KERN_INFO fmt, ##args)
 
+/*
+ * Wrappers that use printk_in_rcu
+ */
+#define btrfs_emerg_in_rcu(fs_info, fmt, args...) \
+	btrfs_printk_in_rcu(fs_info, KERN_EMERG fmt, ##args)
+#define btrfs_alert_in_rcu(fs_info, fmt, args...) \
+	btrfs_printk_in_rcu(fs_info, KERN_ALERT fmt, ##args)
+#define btrfs_crit_in_rcu(fs_info, fmt, args...) \
+	btrfs_printk_in_rcu(fs_info, KERN_CRIT fmt, ##args)
+#define btrfs_err_in_rcu(fs_info, fmt, args...) \
+	btrfs_printk_in_rcu(fs_info, KERN_ERR fmt, ##args)
+#define btrfs_warn_in_rcu(fs_info, fmt, args...) \
+	btrfs_printk_in_rcu(fs_info, KERN_WARNING fmt, ##args)
+#define btrfs_notice_in_rcu(fs_info, fmt, args...) \
+	btrfs_printk_in_rcu(fs_info, KERN_NOTICE fmt, ##args)
+#define btrfs_info_in_rcu(fs_info, fmt, args...) \
+	btrfs_printk_in_rcu(fs_info, KERN_INFO fmt, ##args)
+
+/*
+ * Wrappers that use a ratelimited printk_in_rcu
+ */
+#define btrfs_emerg_rl_in_rcu(fs_info, fmt, args...) \
+	btrfs_printk_rl_in_rcu(fs_info, KERN_EMERG fmt, ##args)
+#define btrfs_alert_rl_in_rcu(fs_info, fmt, args...) \
+	btrfs_printk_rl_in_rcu(fs_info, KERN_ALERT fmt, ##args)
+#define btrfs_crit_rl_in_rcu(fs_info, fmt, args...) \
+	btrfs_printk_rl_in_rcu(fs_info, KERN_CRIT fmt, ##args)
+#define btrfs_err_rl_in_rcu(fs_info, fmt, args...) \
+	btrfs_printk_rl_in_rcu(fs_info, KERN_ERR fmt, ##args)
+#define btrfs_warn_rl_in_rcu(fs_info, fmt, args...) \
+	btrfs_printk_rl_in_rcu(fs_info, KERN_WARNING fmt, ##args)
+#define btrfs_notice_rl_in_rcu(fs_info, fmt, args...) \
+	btrfs_printk_rl_in_rcu(fs_info, KERN_NOTICE fmt, ##args)
+#define btrfs_info_rl_in_rcu(fs_info, fmt, args...) \
+	btrfs_printk_rl_in_rcu(fs_info, KERN_INFO fmt, ##args)
+
+/*
+ * Wrappers that use a ratelimited printk
+ */
+#define btrfs_emerg_rl(fs_info, fmt, args...) \
+	btrfs_printk_ratelimited(fs_info, KERN_EMERG fmt, ##args)
+#define btrfs_alert_rl(fs_info, fmt, args...) \
+	btrfs_printk_ratelimited(fs_info, KERN_ALERT fmt, ##args)
+#define btrfs_crit_rl(fs_info, fmt, args...) \
+	btrfs_printk_ratelimited(fs_info, KERN_CRIT fmt, ##args)
+#define btrfs_err_rl(fs_info, fmt, args...) \
+	btrfs_printk_ratelimited(fs_info, KERN_ERR fmt, ##args)
+#define btrfs_warn_rl(fs_info, fmt, args...) \
+	btrfs_printk_ratelimited(fs_info, KERN_WARNING fmt, ##args)
+#define btrfs_notice_rl(fs_info, fmt, args...) \
+	btrfs_printk_ratelimited(fs_info, KERN_NOTICE fmt, ##args)
+#define btrfs_info_rl(fs_info, fmt, args...) \
+	btrfs_printk_ratelimited(fs_info, KERN_INFO fmt, ##args)
 #ifdef DEBUG
 #define btrfs_debug(fs_info, fmt, args...) \
 	btrfs_printk(fs_info, KERN_DEBUG fmt, ##args)
+#define btrfs_debug_in_rcu(fs_info, fmt, args...) \
+	btrfs_printk_in_rcu(fs_info, KERN_DEBUG fmt, ##args)
+#define btrfs_debug_rl_in_rcu(fs_info, fmt, args...) \
+	btrfs_printk_rl_in_rcu(fs_info, KERN_DEBUG fmt, ##args)
+#define btrfs_debug_rl(fs_info, fmt, args...) \
+	btrfs_printk_ratelimited(fs_info, KERN_DEBUG fmt, ##args)
 #else
 #define btrfs_debug(fs_info, fmt, args...) \
     no_printk(KERN_DEBUG fmt, ##args)
+#define btrfs_debug_in_rcu(fs_info, fmt, args...) \
+	no_printk(KERN_DEBUG fmt, ##args)
+#define btrfs_debug_rl_in_rcu(fs_info, fmt, args...) \
+	no_printk(KERN_DEBUG fmt, ##args)
+#define btrfs_debug_rl(fs_info, fmt, args...) \
+	no_printk(KERN_DEBUG fmt, ##args)
 #endif
 
+#define btrfs_printk_in_rcu(fs_info, fmt, args...)	\
+do {							\
+	rcu_read_lock();				\
+	btrfs_printk(fs_info, fmt, ##args);		\
+	rcu_read_unlock();				\
+} while (0)
+
+#define btrfs_printk_ratelimited(fs_info, fmt, args...)		\
+do {								\
+	static DEFINE_RATELIMIT_STATE(_rs,			\
+		DEFAULT_RATELIMIT_INTERVAL,			\
+		DEFAULT_RATELIMIT_BURST);       		\
+	if (__ratelimit(&_rs))					\
+		btrfs_printk(fs_info, fmt, ##args);		\
+} while (0)
+
+#define btrfs_printk_rl_in_rcu(fs_info, fmt, args...)		\
+do {								\
+	rcu_read_lock();					\
+	btrfs_printk_ratelimited(fs_info, fmt, ##args);		\
+	rcu_read_unlock();					\
+} while (0)
+
 #ifdef CONFIG_BTRFS_ASSERT
 
+__cold
 static inline void assfail(char *expr, char *file, int line)
 {
 	pr_err("BTRFS: assertion failed: %s, file: %s, line: %d",
@@ -4065,10 +4212,13 @@ static inline void assfail(char *expr, char *file, int line)
 
 #define btrfs_assert()
 __printf(5, 6)
+__cold
 void __btrfs_std_error(struct btrfs_fs_info *fs_info, const char *function,
 		     unsigned int line, int errno, const char *fmt, ...);
 
+const char *btrfs_decode_error(int errno);
 
+__cold
 void __btrfs_abort_transaction(struct btrfs_trans_handle *trans,
 			       struct btrfs_root *root, const char *function,
 			       unsigned int line, int errno);
@@ -4111,27 +4261,27 @@ static inline int __btrfs_fs_incompat(struct btrfs_fs_info *fs_info, u64 flag)
  * Call btrfs_abort_transaction as early as possible when an error condition is
  * detected, that way the exact line number is reported.
  */
-
 #define btrfs_abort_transaction(trans, root, errno)		\
 do {								\
-	__btrfs_abort_transaction(trans, root, __func__,	\
-				  __LINE__, errno);		\
-} while (0)
-
-#define btrfs_std_error(fs_info, errno)				\
-do {								\
-	if ((errno))						\
-		__btrfs_std_error((fs_info), __func__,		\
-				   __LINE__, (errno), NULL);	\
+	/* Report first abort since mount */			\
+	if (!test_and_set_bit(BTRFS_FS_STATE_TRANS_ABORTED,	\
+			&((root)->fs_info->fs_state))) {	\
+		WARN(1, KERN_DEBUG				\
+		"BTRFS: Transaction aborted (error %d)\n",	\
+		(errno));					\
+	}							\
+	__btrfs_abort_transaction((trans), (root), __func__,	\
+				  __LINE__, (errno));		\
 } while (0)
 
-#define btrfs_error(fs_info, errno, fmt, args...)		\
+#define btrfs_std_error(fs_info, errno, fmt, args...)		\
 do {								\
 	__btrfs_std_error((fs_info), __func__, __LINE__,	\
 			  (errno), fmt, ##args);		\
 } while (0)
 
 __printf(5, 6)
+__cold
 void __btrfs_panic(struct btrfs_fs_info *fs_info, const char *function,
 		   unsigned int line, int errno, const char *fmt, ...);
 
@@ -4172,8 +4322,7 @@ int btrfs_reloc_clone_csums(struct inode *inode, u64 file_pos, u64 len);
 int btrfs_reloc_cow_block(struct btrfs_trans_handle *trans,
 			  struct btrfs_root *root, struct extent_buffer *buf,
 			  struct extent_buffer *cow);
-void btrfs_reloc_pre_snapshot(struct btrfs_trans_handle *trans,
-			      struct btrfs_pending_snapshot *pending,
+void btrfs_reloc_pre_snapshot(struct btrfs_pending_snapshot *pending,
 			      u64 *bytes_to_reserve);
 int btrfs_reloc_post_snapshot(struct btrfs_trans_handle *trans,
 			      struct btrfs_pending_snapshot *pending);
diff --git a/kernel/fs/btrfs/delayed-inode.c b/kernel/fs/btrfs/delayed-inode.c
index a2ae42720..02b934d0e 100644
--- a/kernel/fs/btrfs/delayed-inode.c
+++ b/kernel/fs/btrfs/delayed-inode.c
@@ -463,6 +463,10 @@ static int __btrfs_add_delayed_deletion_item(struct btrfs_delayed_node *node,
 static void finish_one_item(struct btrfs_delayed_root *delayed_root)
 {
 	int seq = atomic_inc_return(&delayed_root->items_seq);
+
+	/*
+	 * atomic_dec_return implies a barrier for waitqueue_active
+	 */
 	if ((atomic_dec_return(&delayed_root->items) <
 	    BTRFS_DELAYED_BACKGROUND || seq % BTRFS_DELAYED_BATCH == 0) &&
 	    waitqueue_active(&delayed_root->wait))
@@ -1690,7 +1694,7 @@ int btrfs_should_delete_dir_index(struct list_head *del_list,
  *
  */
 int btrfs_readdir_delayed_dir_index(struct dir_context *ctx,
-				    struct list_head *ins_list)
+				    struct list_head *ins_list, bool *emitted)
 {
 	struct btrfs_dir_item *di;
 	struct btrfs_delayed_item *curr, *next;
@@ -1734,6 +1738,7 @@ int btrfs_readdir_delayed_dir_index(struct dir_context *ctx,
 
 		if (over)
 			return 1;
+		*emitted = true;
 	}
 	return 0;
 }
diff --git a/kernel/fs/btrfs/delayed-inode.h b/kernel/fs/btrfs/delayed-inode.h
index f70119f25..0167853c8 100644
--- a/kernel/fs/btrfs/delayed-inode.h
+++ b/kernel/fs/btrfs/delayed-inode.h
@@ -144,7 +144,7 @@ void btrfs_put_delayed_items(struct list_head *ins_list,
 int btrfs_should_delete_dir_index(struct list_head *del_list,
 				  u64 index);
 int btrfs_readdir_delayed_dir_index(struct dir_context *ctx,
-				    struct list_head *ins_list);
+				    struct list_head *ins_list, bool *emitted);
 
 /* for init */
 int __init btrfs_delayed_inode_init(void);
diff --git a/kernel/fs/btrfs/delayed-ref.c b/kernel/fs/btrfs/delayed-ref.c
index 8f8ed7d20..e06dd75ad 100644
--- a/kernel/fs/btrfs/delayed-ref.c
+++ b/kernel/fs/btrfs/delayed-ref.c
@@ -22,6 +22,7 @@
 #include "ctree.h"
 #include "delayed-ref.h"
 #include "transaction.h"
+#include "qgroup.h"
 
 struct kmem_cache *btrfs_delayed_ref_head_cachep;
 struct kmem_cache *btrfs_delayed_tree_ref_cachep;
@@ -84,87 +85,6 @@ static int comp_data_refs(struct btrfs_delayed_data_ref *ref2,
 	return 0;
 }
 
-/*
- * entries in the rb tree are ordered by the byte number of the extent,
- * type of the delayed backrefs and content of delayed backrefs.
- */
-static int comp_entry(struct btrfs_delayed_ref_node *ref2,
-		      struct btrfs_delayed_ref_node *ref1,
-		      bool compare_seq)
-{
-	if (ref1->bytenr < ref2->bytenr)
-		return -1;
-	if (ref1->bytenr > ref2->bytenr)
-		return 1;
-	if (ref1->is_head && ref2->is_head)
-		return 0;
-	if (ref2->is_head)
-		return -1;
-	if (ref1->is_head)
-		return 1;
-	if (ref1->type < ref2->type)
-		return -1;
-	if (ref1->type > ref2->type)
-		return 1;
-	if (ref1->no_quota > ref2->no_quota)
-		return 1;
-	if (ref1->no_quota < ref2->no_quota)
-		return -1;
-	/* merging of sequenced refs is not allowed */
-	if (compare_seq) {
-		if (ref1->seq < ref2->seq)
-			return -1;
-		if (ref1->seq > ref2->seq)
-			return 1;
-	}
-	if (ref1->type == BTRFS_TREE_BLOCK_REF_KEY ||
-	    ref1->type == BTRFS_SHARED_BLOCK_REF_KEY) {
-		return comp_tree_refs(btrfs_delayed_node_to_tree_ref(ref2),
-				      btrfs_delayed_node_to_tree_ref(ref1),
-				      ref1->type);
-	} else if (ref1->type == BTRFS_EXTENT_DATA_REF_KEY ||
-		   ref1->type == BTRFS_SHARED_DATA_REF_KEY) {
-		return comp_data_refs(btrfs_delayed_node_to_data_ref(ref2),
-				      btrfs_delayed_node_to_data_ref(ref1));
-	}
-	BUG();
-	return 0;
-}
-
-/*
- * insert a new ref into the rbtree.  This returns any existing refs
- * for the same (bytenr,parent) tuple, or NULL if the new node was properly
- * inserted.
- */
-static struct btrfs_delayed_ref_node *tree_insert(struct rb_root *root,
-						  struct rb_node *node)
-{
-	struct rb_node **p = &root->rb_node;
-	struct rb_node *parent_node = NULL;
-	struct btrfs_delayed_ref_node *entry;
-	struct btrfs_delayed_ref_node *ins;
-	int cmp;
-
-	ins = rb_entry(node, struct btrfs_delayed_ref_node, rb_node);
-	while (*p) {
-		parent_node = *p;
-		entry = rb_entry(parent_node, struct btrfs_delayed_ref_node,
-				 rb_node);
-
-		cmp = comp_entry(entry, ins, 1);
-		if (cmp < 0)
-			p = &(*p)->rb_left;
-		else if (cmp > 0)
-			p = &(*p)->rb_right;
-		else
-			return entry;
-	}
-
-	rb_link_node(node, parent_node, p);
-	rb_insert_color(node, root);
-	return NULL;
-}
-
 /* insert a new ref to head ref rbtree */
 static struct btrfs_delayed_ref_head *htree_insert(struct rb_root *root,
 						   struct rb_node *node)
@@ -268,7 +188,7 @@ static inline void drop_delayed_ref(struct btrfs_trans_handle *trans,
 		rb_erase(&head->href_node, &delayed_refs->href_root);
 	} else {
 		assert_spin_locked(&head->lock);
-		rb_erase(&ref->rb_node, &head->ref_root);
+		list_del(&ref->list);
 	}
 	ref->in_tree = 0;
 	btrfs_put_delayed_ref(ref);
@@ -277,36 +197,50 @@ static inline void drop_delayed_ref(struct btrfs_trans_handle *trans,
 		trans->delayed_ref_updates--;
 }
 
-static int merge_ref(struct btrfs_trans_handle *trans,
-		     struct btrfs_delayed_ref_root *delayed_refs,
-		     struct btrfs_delayed_ref_head *head,
-		     struct btrfs_delayed_ref_node *ref, u64 seq)
+static bool merge_ref(struct btrfs_trans_handle *trans,
+		      struct btrfs_delayed_ref_root *delayed_refs,
+		      struct btrfs_delayed_ref_head *head,
+		      struct btrfs_delayed_ref_node *ref,
+		      u64 seq)
 {
-	struct rb_node *node;
-	int mod = 0;
-	int done = 0;
+	struct btrfs_delayed_ref_node *next;
+	bool done = false;
+
+	next = list_first_entry(&head->ref_list, struct btrfs_delayed_ref_node,
+				list);
+	while (!done && &next->list != &head->ref_list) {
+		int mod;
+		struct btrfs_delayed_ref_node *next2;
 
-	node = rb_next(&ref->rb_node);
-	while (!done && node) {
-		struct btrfs_delayed_ref_node *next;
+		next2 = list_next_entry(next, list);
+
+		if (next == ref)
+			goto next;
 
-		next = rb_entry(node, struct btrfs_delayed_ref_node, rb_node);
-		node = rb_next(node);
 		if (seq && next->seq >= seq)
-			break;
-		if (comp_entry(ref, next, 0))
-			continue;
+			goto next;
+
+		if (next->type != ref->type)
+			goto next;
+
+		if ((ref->type == BTRFS_TREE_BLOCK_REF_KEY ||
+		     ref->type == BTRFS_SHARED_BLOCK_REF_KEY) &&
+		    comp_tree_refs(btrfs_delayed_node_to_tree_ref(ref),
+				   btrfs_delayed_node_to_tree_ref(next),
+				   ref->type))
+			goto next;
+		if ((ref->type == BTRFS_EXTENT_DATA_REF_KEY ||
+		     ref->type == BTRFS_SHARED_DATA_REF_KEY) &&
+		    comp_data_refs(btrfs_delayed_node_to_data_ref(ref),
+				   btrfs_delayed_node_to_data_ref(next)))
+			goto next;
 
 		if (ref->action == next->action) {
 			mod = next->ref_mod;
 		} else {
 			if (ref->ref_mod < next->ref_mod) {
-				struct btrfs_delayed_ref_node *tmp;
-
-				tmp = ref;
-				ref = next;
-				next = tmp;
-				done = 1;
+				swap(ref, next);
+				done = true;
 			}
 			mod = -next->ref_mod;
 		}
@@ -315,16 +249,18 @@ static int merge_ref(struct btrfs_trans_handle *trans,
 		ref->ref_mod += mod;
 		if (ref->ref_mod == 0) {
 			drop_delayed_ref(trans, delayed_refs, head, ref);
-			done = 1;
+			done = true;
 		} else {
 			/*
-			 * You can't have multiples of the same ref on a tree
-			 * block.
+			 * Can't have multiples of the same ref on a tree block.
 			 */
 			WARN_ON(ref->type == BTRFS_TREE_BLOCK_REF_KEY ||
 				ref->type == BTRFS_SHARED_BLOCK_REF_KEY);
 		}
+next:
+		next = next2;
 	}
+
 	return done;
 }
 
@@ -333,14 +269,15 @@ void btrfs_merge_delayed_refs(struct btrfs_trans_handle *trans,
 			      struct btrfs_delayed_ref_root *delayed_refs,
 			      struct btrfs_delayed_ref_head *head)
 {
-	struct rb_node *node;
+	struct btrfs_delayed_ref_node *ref;
 	u64 seq = 0;
 
 	assert_spin_locked(&head->lock);
-	/*
-	 * We don't have too much refs to merge in the case of delayed data
-	 * refs.
-	 */
+
+	if (list_empty(&head->ref_list))
+		return;
+
+	/* We don't have too many refs to merge for data. */
 	if (head->is_data)
 		return;
 
@@ -354,19 +291,22 @@ void btrfs_merge_delayed_refs(struct btrfs_trans_handle *trans,
 	}
 	spin_unlock(&fs_info->tree_mod_seq_lock);
 
-	node = rb_first(&head->ref_root);
-	while (node) {
-		struct btrfs_delayed_ref_node *ref;
-
-		ref = rb_entry(node, struct btrfs_delayed_ref_node,
-			       rb_node);
-		/* We can't merge refs that are outside of our seq count */
+	ref = list_first_entry(&head->ref_list, struct btrfs_delayed_ref_node,
+			       list);
+	while (&ref->list != &head->ref_list) {
 		if (seq && ref->seq >= seq)
-			break;
-		if (merge_ref(trans, delayed_refs, head, ref, seq))
-			node = rb_first(&head->ref_root);
-		else
-			node = rb_next(&ref->rb_node);
+			goto next;
+
+		if (merge_ref(trans, delayed_refs, head, ref, seq)) {
+			if (list_empty(&head->ref_list))
+				break;
+			ref = list_first_entry(&head->ref_list,
+					       struct btrfs_delayed_ref_node,
+					       list);
+			continue;
+		}
+next:
+		ref = list_next_entry(ref, list);
 	}
 }
 
@@ -443,45 +383,70 @@ again:
 }
 
 /*
- * helper function to update an extent delayed ref in the
- * rbtree.  existing and update must both have the same
- * bytenr and parent
+ * Helper to insert the ref_node to the tail or merge with tail.
  *
- * This may free existing if the update cancels out whatever
- * operation it was doing.
+ * Return 0 for insert.
+ * Return >0 for merge.
  */
-static noinline void
-update_existing_ref(struct btrfs_trans_handle *trans,
-		    struct btrfs_delayed_ref_root *delayed_refs,
-		    struct btrfs_delayed_ref_head *head,
-		    struct btrfs_delayed_ref_node *existing,
-		    struct btrfs_delayed_ref_node *update)
+static int
+add_delayed_ref_tail_merge(struct btrfs_trans_handle *trans,
+			   struct btrfs_delayed_ref_root *root,
+			   struct btrfs_delayed_ref_head *href,
+			   struct btrfs_delayed_ref_node *ref)
 {
-	if (update->action != existing->action) {
-		/*
-		 * this is effectively undoing either an add or a
-		 * drop.  We decrement the ref_mod, and if it goes
-		 * down to zero we just delete the entry without
-		 * every changing the extent allocation tree.
-		 */
-		existing->ref_mod--;
-		if (existing->ref_mod == 0)
-			drop_delayed_ref(trans, delayed_refs, head, existing);
-		else
-			WARN_ON(existing->type == BTRFS_TREE_BLOCK_REF_KEY ||
-				existing->type == BTRFS_SHARED_BLOCK_REF_KEY);
+	struct btrfs_delayed_ref_node *exist;
+	int mod;
+	int ret = 0;
+
+	spin_lock(&href->lock);
+	/* Check whether we can merge the tail node with ref */
+	if (list_empty(&href->ref_list))
+		goto add_tail;
+	exist = list_entry(href->ref_list.prev, struct btrfs_delayed_ref_node,
+			   list);
+	/* No need to compare bytenr nor is_head */
+	if (exist->type != ref->type || exist->seq != ref->seq)
+		goto add_tail;
+
+	if ((exist->type == BTRFS_TREE_BLOCK_REF_KEY ||
+	     exist->type == BTRFS_SHARED_BLOCK_REF_KEY) &&
+	    comp_tree_refs(btrfs_delayed_node_to_tree_ref(exist),
+			   btrfs_delayed_node_to_tree_ref(ref),
+			   ref->type))
+		goto add_tail;
+	if ((exist->type == BTRFS_EXTENT_DATA_REF_KEY ||
+	     exist->type == BTRFS_SHARED_DATA_REF_KEY) &&
+	    comp_data_refs(btrfs_delayed_node_to_data_ref(exist),
+			   btrfs_delayed_node_to_data_ref(ref)))
+		goto add_tail;
+
+	/* Now we are sure we can merge */
+	ret = 1;
+	if (exist->action == ref->action) {
+		mod = ref->ref_mod;
 	} else {
-		WARN_ON(existing->type == BTRFS_TREE_BLOCK_REF_KEY ||
-			existing->type == BTRFS_SHARED_BLOCK_REF_KEY);
-		/*
-		 * the action on the existing ref matches
-		 * the action on the ref we're trying to add.
-		 * Bump the ref_mod by one so the backref that
-		 * is eventually added/removed has the correct
-		 * reference count
-		 */
-		existing->ref_mod += update->ref_mod;
+		/* Need to change action */
+		if (exist->ref_mod < ref->ref_mod) {
+			exist->action = ref->action;
+			mod = -exist->ref_mod;
+			exist->ref_mod = ref->ref_mod;
+		} else
+			mod = -ref->ref_mod;
 	}
+	exist->ref_mod += mod;
+
+	/* remove existing tail if its ref_mod is zero */
+	if (exist->ref_mod == 0)
+		drop_delayed_ref(trans, root, href, exist);
+	spin_unlock(&href->lock);
+	return ret;
+
+add_tail:
+	list_add_tail(&ref->list, &href->ref_list);
+	atomic_inc(&root->num_entries);
+	trans->delayed_ref_updates++;
+	spin_unlock(&href->lock);
+	return ret;
 }
 
 /*
@@ -568,15 +533,21 @@ update_existing_head_ref(struct btrfs_delayed_ref_root *delayed_refs,
 static noinline struct btrfs_delayed_ref_head *
 add_delayed_ref_head(struct btrfs_fs_info *fs_info,
 		     struct btrfs_trans_handle *trans,
-		     struct btrfs_delayed_ref_node *ref, u64 bytenr,
-		     u64 num_bytes, int action, int is_data)
+		     struct btrfs_delayed_ref_node *ref,
+		     struct btrfs_qgroup_extent_record *qrecord,
+		     u64 bytenr, u64 num_bytes, u64 ref_root, u64 reserved,
+		     int action, int is_data)
 {
 	struct btrfs_delayed_ref_head *existing;
 	struct btrfs_delayed_ref_head *head_ref = NULL;
 	struct btrfs_delayed_ref_root *delayed_refs;
+	struct btrfs_qgroup_extent_record *qexisting;
 	int count_mod = 1;
 	int must_insert_reserved = 0;
 
+	/* If reserved is provided, it must be a data extent. */
+	BUG_ON(!is_data && reserved);
+
 	/*
 	 * the head node stores the sum of all the mods, so dropping a ref
 	 * should drop the sum in the head node by one.
@@ -618,9 +589,28 @@ add_delayed_ref_head(struct btrfs_fs_info *fs_info,
 	head_ref = btrfs_delayed_node_to_head(ref);
 	head_ref->must_insert_reserved = must_insert_reserved;
 	head_ref->is_data = is_data;
-	head_ref->ref_root = RB_ROOT;
+	INIT_LIST_HEAD(&head_ref->ref_list);
 	head_ref->processing = 0;
 	head_ref->total_ref_mod = count_mod;
+	head_ref->qgroup_reserved = 0;
+	head_ref->qgroup_ref_root = 0;
+
+	/* Record qgroup extent info if provided */
+	if (qrecord) {
+		if (ref_root && reserved) {
+			head_ref->qgroup_ref_root = ref_root;
+			head_ref->qgroup_reserved = reserved;
+		}
+
+		qrecord->bytenr = bytenr;
+		qrecord->num_bytes = num_bytes;
+		qrecord->old_roots = NULL;
+
+		qexisting = btrfs_qgroup_insert_dirty_extent(delayed_refs,
+							     qrecord);
+		if (qexisting)
+			kfree(qrecord);
+	}
 
 	spin_lock_init(&head_ref->lock);
 	mutex_init(&head_ref->mutex);
@@ -630,6 +620,8 @@ add_delayed_ref_head(struct btrfs_fs_info *fs_info,
 	existing = htree_insert(&delayed_refs->href_root,
 				&head_ref->href_node);
 	if (existing) {
+		WARN_ON(ref_root && reserved && existing->qgroup_ref_root
+			&& existing->qgroup_reserved);
 		update_existing_head_ref(delayed_refs, &existing->node, ref);
 		/*
 		 * we've updated the existing ref, free the newly
@@ -657,12 +649,12 @@ add_delayed_tree_ref(struct btrfs_fs_info *fs_info,
 		     struct btrfs_delayed_ref_head *head_ref,
 		     struct btrfs_delayed_ref_node *ref, u64 bytenr,
 		     u64 num_bytes, u64 parent, u64 ref_root, int level,
-		     int action, int no_quota)
+		     int action)
 {
-	struct btrfs_delayed_ref_node *existing;
 	struct btrfs_delayed_tree_ref *full_ref;
 	struct btrfs_delayed_ref_root *delayed_refs;
 	u64 seq = 0;
+	int ret;
 
 	if (action == BTRFS_ADD_DELAYED_EXTENT)
 		action = BTRFS_ADD_DELAYED_REF;
@@ -679,7 +671,6 @@ add_delayed_tree_ref(struct btrfs_fs_info *fs_info,
 	ref->action = action;
 	ref->is_head = 0;
 	ref->in_tree = 1;
-	ref->no_quota = no_quota;
 	ref->seq = seq;
 
 	full_ref = btrfs_delayed_node_to_tree_ref(ref);
@@ -693,21 +684,14 @@ add_delayed_tree_ref(struct btrfs_fs_info *fs_info,
 
 	trace_add_delayed_tree_ref(ref, full_ref, action);
 
-	spin_lock(&head_ref->lock);
-	existing = tree_insert(&head_ref->ref_root, &ref->rb_node);
-	if (existing) {
-		update_existing_ref(trans, delayed_refs, head_ref, existing,
-				    ref);
-		/*
-		 * we've updated the existing ref, free the newly
-		 * allocated ref
-		 */
+	ret = add_delayed_ref_tail_merge(trans, delayed_refs, head_ref, ref);
+
+	/*
+	 * XXX: memory should be freed at the same level allocated.
+	 * But bad practice is anywhere... Follow it now. Need cleanup.
+	 */
+	if (ret > 0)
 		kmem_cache_free(btrfs_delayed_tree_ref_cachep, full_ref);
-	} else {
-		atomic_inc(&delayed_refs->num_entries);
-		trans->delayed_ref_updates++;
-	}
-	spin_unlock(&head_ref->lock);
 }
 
 /*
@@ -719,12 +703,12 @@ add_delayed_data_ref(struct btrfs_fs_info *fs_info,
 		     struct btrfs_delayed_ref_head *head_ref,
 		     struct btrfs_delayed_ref_node *ref, u64 bytenr,
 		     u64 num_bytes, u64 parent, u64 ref_root, u64 owner,
-		     u64 offset, int action, int no_quota)
+		     u64 offset, int action)
 {
-	struct btrfs_delayed_ref_node *existing;
 	struct btrfs_delayed_data_ref *full_ref;
 	struct btrfs_delayed_ref_root *delayed_refs;
 	u64 seq = 0;
+	int ret;
 
 	if (action == BTRFS_ADD_DELAYED_EXTENT)
 		action = BTRFS_ADD_DELAYED_REF;
@@ -742,7 +726,6 @@ add_delayed_data_ref(struct btrfs_fs_info *fs_info,
 	ref->action = action;
 	ref->is_head = 0;
 	ref->in_tree = 1;
-	ref->no_quota = no_quota;
 	ref->seq = seq;
 
 	full_ref = btrfs_delayed_node_to_data_ref(ref);
@@ -758,21 +741,10 @@ add_delayed_data_ref(struct btrfs_fs_info *fs_info,
 
 	trace_add_delayed_data_ref(ref, full_ref, action);
 
-	spin_lock(&head_ref->lock);
-	existing = tree_insert(&head_ref->ref_root, &ref->rb_node);
-	if (existing) {
-		update_existing_ref(trans, delayed_refs, head_ref, existing,
-				    ref);
-		/*
-		 * we've updated the existing ref, free the newly
-		 * allocated ref
-		 */
+	ret = add_delayed_ref_tail_merge(trans, delayed_refs, head_ref, ref);
+
+	if (ret > 0)
 		kmem_cache_free(btrfs_delayed_data_ref_cachep, full_ref);
-	} else {
-		atomic_inc(&delayed_refs->num_entries);
-		trans->delayed_ref_updates++;
-	}
-	spin_unlock(&head_ref->lock);
 }
 
 /*
@@ -784,15 +756,12 @@ int btrfs_add_delayed_tree_ref(struct btrfs_fs_info *fs_info,
 			       struct btrfs_trans_handle *trans,
 			       u64 bytenr, u64 num_bytes, u64 parent,
 			       u64 ref_root,  int level, int action,
-			       struct btrfs_delayed_extent_op *extent_op,
-			       int no_quota)
+			       struct btrfs_delayed_extent_op *extent_op)
 {
 	struct btrfs_delayed_tree_ref *ref;
 	struct btrfs_delayed_ref_head *head_ref;
 	struct btrfs_delayed_ref_root *delayed_refs;
-
-	if (!is_fstree(ref_root) || !fs_info->quota_enabled)
-		no_quota = 0;
+	struct btrfs_qgroup_extent_record *record = NULL;
 
 	BUG_ON(extent_op && extent_op->is_data);
 	ref = kmem_cache_alloc(btrfs_delayed_tree_ref_cachep, GFP_NOFS);
@@ -800,9 +769,13 @@ int btrfs_add_delayed_tree_ref(struct btrfs_fs_info *fs_info,
 		return -ENOMEM;
 
 	head_ref = kmem_cache_alloc(btrfs_delayed_ref_head_cachep, GFP_NOFS);
-	if (!head_ref) {
-		kmem_cache_free(btrfs_delayed_tree_ref_cachep, ref);
-		return -ENOMEM;
+	if (!head_ref)
+		goto free_ref;
+
+	if (fs_info->quota_enabled && is_fstree(ref_root)) {
+		record = kmalloc(sizeof(*record), GFP_NOFS);
+		if (!record)
+			goto free_head_ref;
 	}
 
 	head_ref->extent_op = extent_op;
@@ -814,15 +787,21 @@ int btrfs_add_delayed_tree_ref(struct btrfs_fs_info *fs_info,
 	 * insert both the head node and the new ref without dropping
 	 * the spin lock
 	 */
-	head_ref = add_delayed_ref_head(fs_info, trans, &head_ref->node,
-					bytenr, num_bytes, action, 0);
+	head_ref = add_delayed_ref_head(fs_info, trans, &head_ref->node, record,
+					bytenr, num_bytes, 0, 0, action, 0);
 
 	add_delayed_tree_ref(fs_info, trans, head_ref, &ref->node, bytenr,
-				   num_bytes, parent, ref_root, level, action,
-				   no_quota);
+			     num_bytes, parent, ref_root, level, action);
 	spin_unlock(&delayed_refs->lock);
 
 	return 0;
+
+free_head_ref:
+	kmem_cache_free(btrfs_delayed_ref_head_cachep, head_ref);
+free_ref:
+	kmem_cache_free(btrfs_delayed_tree_ref_cachep, ref);
+
+	return -ENOMEM;
 }
 
 /*
@@ -832,16 +811,13 @@ int btrfs_add_delayed_data_ref(struct btrfs_fs_info *fs_info,
 			       struct btrfs_trans_handle *trans,
 			       u64 bytenr, u64 num_bytes,
 			       u64 parent, u64 ref_root,
-			       u64 owner, u64 offset, int action,
-			       struct btrfs_delayed_extent_op *extent_op,
-			       int no_quota)
+			       u64 owner, u64 offset, u64 reserved, int action,
+			       struct btrfs_delayed_extent_op *extent_op)
 {
 	struct btrfs_delayed_data_ref *ref;
 	struct btrfs_delayed_ref_head *head_ref;
 	struct btrfs_delayed_ref_root *delayed_refs;
-
-	if (!is_fstree(ref_root) || !fs_info->quota_enabled)
-		no_quota = 0;
+	struct btrfs_qgroup_extent_record *record = NULL;
 
 	BUG_ON(extent_op && !extent_op->is_data);
 	ref = kmem_cache_alloc(btrfs_delayed_data_ref_cachep, GFP_NOFS);
@@ -854,6 +830,16 @@ int btrfs_add_delayed_data_ref(struct btrfs_fs_info *fs_info,
 		return -ENOMEM;
 	}
 
+	if (fs_info->quota_enabled && is_fstree(ref_root)) {
+		record = kmalloc(sizeof(*record), GFP_NOFS);
+		if (!record) {
+			kmem_cache_free(btrfs_delayed_data_ref_cachep, ref);
+			kmem_cache_free(btrfs_delayed_ref_head_cachep,
+					head_ref);
+			return -ENOMEM;
+		}
+	}
+
 	head_ref->extent_op = extent_op;
 
 	delayed_refs = &trans->transaction->delayed_refs;
@@ -863,17 +849,45 @@ int btrfs_add_delayed_data_ref(struct btrfs_fs_info *fs_info,
 	 * insert both the head node and the new ref without dropping
 	 * the spin lock
 	 */
-	head_ref = add_delayed_ref_head(fs_info, trans, &head_ref->node,
-					bytenr, num_bytes, action, 1);
+	head_ref = add_delayed_ref_head(fs_info, trans, &head_ref->node, record,
+					bytenr, num_bytes, ref_root, reserved,
+					action, 1);
 
 	add_delayed_data_ref(fs_info, trans, head_ref, &ref->node, bytenr,
 				   num_bytes, parent, ref_root, owner, offset,
-				   action, no_quota);
+				   action);
 	spin_unlock(&delayed_refs->lock);
 
 	return 0;
 }
 
+int btrfs_add_delayed_qgroup_reserve(struct btrfs_fs_info *fs_info,
+				     struct btrfs_trans_handle *trans,
+				     u64 ref_root, u64 bytenr, u64 num_bytes)
+{
+	struct btrfs_delayed_ref_root *delayed_refs;
+	struct btrfs_delayed_ref_head *ref_head;
+	int ret = 0;
+
+	if (!fs_info->quota_enabled || !is_fstree(ref_root))
+		return 0;
+
+	delayed_refs = &trans->transaction->delayed_refs;
+
+	spin_lock(&delayed_refs->lock);
+	ref_head = find_ref_head(&delayed_refs->href_root, bytenr, 0);
+	if (!ref_head) {
+		ret = -ENOENT;
+		goto out;
+	}
+	WARN_ON(ref_head->qgroup_reserved || ref_head->qgroup_ref_root);
+	ref_head->qgroup_ref_root = ref_root;
+	ref_head->qgroup_reserved = num_bytes;
+out:
+	spin_unlock(&delayed_refs->lock);
+	return ret;
+}
+
 int btrfs_add_delayed_extent_op(struct btrfs_fs_info *fs_info,
 				struct btrfs_trans_handle *trans,
 				u64 bytenr, u64 num_bytes,
@@ -891,9 +905,9 @@ int btrfs_add_delayed_extent_op(struct btrfs_fs_info *fs_info,
 	delayed_refs = &trans->transaction->delayed_refs;
 	spin_lock(&delayed_refs->lock);
 
-	add_delayed_ref_head(fs_info, trans, &head_ref->node, bytenr,
-				   num_bytes, BTRFS_UPDATE_DELAYED_HEAD,
-				   extent_op->is_data);
+	add_delayed_ref_head(fs_info, trans, &head_ref->node, NULL, bytenr,
+			     num_bytes, 0, 0, BTRFS_UPDATE_DELAYED_HEAD,
+			     extent_op->is_data);
 
 	spin_unlock(&delayed_refs->lock);
 	return 0;
diff --git a/kernel/fs/btrfs/delayed-ref.h b/kernel/fs/btrfs/delayed-ref.h
index 5eb089239..00ed02cbf 100644
--- a/kernel/fs/btrfs/delayed-ref.h
+++ b/kernel/fs/btrfs/delayed-ref.h
@@ -24,9 +24,25 @@
 #define BTRFS_ADD_DELAYED_EXTENT 3 /* record a full extent allocation */
 #define BTRFS_UPDATE_DELAYED_HEAD 4 /* not changing ref count on head ref */
 
+/*
+ * XXX: Qu: I really hate the design that ref_head and tree/data ref shares the
+ * same ref_node structure.
+ * Ref_head is in a higher logic level than tree/data ref, and duplicated
+ * bytenr/num_bytes in ref_node is really a waste or memory, they should be
+ * referred from ref_head.
+ * This gets more disgusting after we use list to store tree/data ref in
+ * ref_head. Must clean this mess up later.
+ */
 struct btrfs_delayed_ref_node {
+	/*
+	 * ref_head use rb tree, stored in ref_root->href.
+	 * indexed by bytenr
+	 */
 	struct rb_node rb_node;
 
+	/*data/tree ref use list, stored in ref_head->ref_list. */
+	struct list_head list;
+
 	/* the starting bytenr of the extent */
 	u64 bytenr;
 
@@ -52,7 +68,6 @@ struct btrfs_delayed_ref_node {
 
 	unsigned int action:8;
 	unsigned int type:8;
-	unsigned int no_quota:1;
 	/* is this node still in the rbtree? */
 	unsigned int is_head:1;
 	unsigned int in_tree:1;
@@ -83,7 +98,7 @@ struct btrfs_delayed_ref_head {
 	struct mutex mutex;
 
 	spinlock_t lock;
-	struct rb_root ref_root;
+	struct list_head ref_list;
 
 	struct rb_node href_node;
 
@@ -97,6 +112,17 @@ struct btrfs_delayed_ref_head {
 	int total_ref_mod;
 
 	/*
+	 * For qgroup reserved space freeing.
+	 *
+	 * ref_root and reserved will be recorded after
+	 * BTRFS_ADD_DELAYED_EXTENT is called.
+	 * And will be used to free reserved qgroup space at
+	 * run_delayed_refs() time.
+	 */
+	u64 qgroup_ref_root;
+	u64 qgroup_reserved;
+
+	/*
 	 * when a new extent is allocated, it is just reserved in memory
 	 * The actual extent isn't inserted into the extent allocation tree
 	 * until the delayed ref is processed.  must_insert_reserved is
@@ -132,6 +158,9 @@ struct btrfs_delayed_ref_root {
 	/* head ref rbtree */
 	struct rb_root href_root;
 
+	/* dirty extent records */
+	struct rb_root dirty_extent_root;
+
 	/* this spin lock protects the rbtree and the entries inside */
 	spinlock_t lock;
 
@@ -156,6 +185,14 @@ struct btrfs_delayed_ref_root {
 	int flushing;
 
 	u64 run_delayed_start;
+
+	/*
+	 * To make qgroup to skip given root.
+	 * This is for snapshot, as btrfs_qgroup_inherit() will manully
+	 * modify counters for snapshot and its source, so we should skip
+	 * the snapshot in new_root/old_roots or it will get calculated twice
+	 */
+	u64 qgroup_to_skip;
 };
 
 extern struct kmem_cache *btrfs_delayed_ref_head_cachep;
@@ -206,15 +243,16 @@ int btrfs_add_delayed_tree_ref(struct btrfs_fs_info *fs_info,
 			       struct btrfs_trans_handle *trans,
 			       u64 bytenr, u64 num_bytes, u64 parent,
 			       u64 ref_root, int level, int action,
-			       struct btrfs_delayed_extent_op *extent_op,
-			       int no_quota);
+			       struct btrfs_delayed_extent_op *extent_op);
 int btrfs_add_delayed_data_ref(struct btrfs_fs_info *fs_info,
 			       struct btrfs_trans_handle *trans,
 			       u64 bytenr, u64 num_bytes,
 			       u64 parent, u64 ref_root,
-			       u64 owner, u64 offset, int action,
-			       struct btrfs_delayed_extent_op *extent_op,
-			       int no_quota);
+			       u64 owner, u64 offset, u64 reserved, int action,
+			       struct btrfs_delayed_extent_op *extent_op);
+int btrfs_add_delayed_qgroup_reserve(struct btrfs_fs_info *fs_info,
+				     struct btrfs_trans_handle *trans,
+				     u64 ref_root, u64 bytenr, u64 num_bytes);
 int btrfs_add_delayed_extent_op(struct btrfs_fs_info *fs_info,
 				struct btrfs_trans_handle *trans,
 				u64 bytenr, u64 num_bytes,
diff --git a/kernel/fs/btrfs/dev-replace.c b/kernel/fs/btrfs/dev-replace.c
index 0573848c7..1e668fb7d 100644
--- a/kernel/fs/btrfs/dev-replace.c
+++ b/kernel/fs/btrfs/dev-replace.c
@@ -183,8 +183,7 @@ no_valid_dev_replace_entry_found:
 	}
 
 out:
-	if (path)
-		btrfs_free_path(path);
+	btrfs_free_path(path);
 	return ret;
 }
 
@@ -328,19 +327,6 @@ int btrfs_dev_replace_start(struct btrfs_root *root,
 	    args->start.tgtdev_name[0] == '\0')
 		return -EINVAL;
 
-	/*
-	 * Here we commit the transaction to make sure commit_total_bytes
-	 * of all the devices are updated.
-	 */
-	trans = btrfs_attach_transaction(root);
-	if (!IS_ERR(trans)) {
-		ret = btrfs_commit_transaction(trans, root);
-		if (ret)
-			return ret;
-	} else if (PTR_ERR(trans) != -ENOENT) {
-		return PTR_ERR(trans);
-	}
-
 	/* the disk copy procedure reuses the scrub code */
 	mutex_lock(&fs_info->volume_mutex);
 	ret = btrfs_dev_replace_find_srcdev(root, args->start.srcdevid,
@@ -357,6 +343,19 @@ int btrfs_dev_replace_start(struct btrfs_root *root,
 	if (ret)
 		return ret;
 
+	/*
+	 * Here we commit the transaction to make sure commit_total_bytes
+	 * of all the devices are updated.
+	 */
+	trans = btrfs_attach_transaction(root);
+	if (!IS_ERR(trans)) {
+		ret = btrfs_commit_transaction(trans, root);
+		if (ret)
+			return ret;
+	} else if (PTR_ERR(trans) != -ENOENT) {
+		return PTR_ERR(trans);
+	}
+
 	btrfs_dev_replace_lock(dev_replace);
 	switch (dev_replace->replace_state) {
 	case BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED:
@@ -376,8 +375,8 @@ int btrfs_dev_replace_start(struct btrfs_root *root,
 	WARN_ON(!tgt_device);
 	dev_replace->tgtdev = tgt_device;
 
-	printk_in_rcu(KERN_INFO
-		      "BTRFS: dev_replace from %s (devid %llu) to %s started\n",
+	btrfs_info_in_rcu(root->fs_info,
+		      "dev_replace from %s (devid %llu) to %s started",
 		      src_device->missing ? "<missing disk>" :
 		        rcu_str_deref(src_device->name),
 		      src_device->devid,
@@ -398,6 +397,10 @@ int btrfs_dev_replace_start(struct btrfs_root *root,
 	args->result = BTRFS_IOCTL_DEV_REPLACE_RESULT_NO_ERROR;
 	btrfs_dev_replace_unlock(dev_replace);
 
+	ret = btrfs_sysfs_add_device_link(tgt_device->fs_devices, tgt_device);
+	if (ret)
+		btrfs_err(root->fs_info, "kobj add dev failed %d\n", ret);
+
 	btrfs_wait_ordered_roots(root->fs_info, -1);
 
 	/* force writing the updated state information to disk */
@@ -451,8 +454,7 @@ static void btrfs_rm_dev_replace_blocked(struct btrfs_fs_info *fs_info)
 static void btrfs_rm_dev_replace_unblocked(struct btrfs_fs_info *fs_info)
 {
 	clear_bit(BTRFS_FS_STATE_DEV_REPLACING, &fs_info->fs_state);
-	if (waitqueue_active(&fs_info->replace_wait))
-		wake_up(&fs_info->replace_wait);
+	wake_up(&fs_info->replace_wait);
 }
 
 static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
@@ -520,8 +522,8 @@ static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
 								src_device,
 								tgt_device);
 	} else {
-		printk_in_rcu(KERN_ERR
-			      "BTRFS: btrfs_scrub_dev(%s, %llu, %s) failed %d\n",
+		btrfs_err_in_rcu(root->fs_info,
+			      "btrfs_scrub_dev(%s, %llu, %s) failed %d",
 			      src_device->missing ? "<missing disk>" :
 			        rcu_str_deref(src_device->name),
 			      src_device->devid,
@@ -537,8 +539,8 @@ static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
 		return scrub_ret;
 	}
 
-	printk_in_rcu(KERN_INFO
-		      "BTRFS: dev_replace from %s (devid %llu) to %s finished\n",
+	btrfs_info_in_rcu(root->fs_info,
+		      "dev_replace from %s (devid %llu) to %s finished",
 		      src_device->missing ? "<missing disk>" :
 		        rcu_str_deref(src_device->name),
 		      src_device->devid,
@@ -583,8 +585,7 @@ static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
 	mutex_unlock(&uuid_mutex);
 
 	/* replace the sysfs entry */
-	btrfs_kobj_rm_device(fs_info, src_device);
-	btrfs_kobj_add_device(fs_info, tgt_device);
+	btrfs_sysfs_rm_device_link(fs_info->fs_devices, src_device);
 	btrfs_rm_dev_replace_free_srcdev(fs_info, src_device);
 
 	/* write back the superblocks */
@@ -807,8 +808,8 @@ static int btrfs_dev_replace_kthread(void *data)
 		progress = status_args->status.progress_1000;
 		kfree(status_args);
 		progress = div_u64(progress, 10);
-		printk_in_rcu(KERN_INFO
-			"BTRFS: continuing dev_replace from %s (devid %llu) to %s @%u%%\n",
+		btrfs_info_in_rcu(fs_info,
+			"continuing dev_replace from %s (devid %llu) to %s @%u%%",
 			dev_replace->srcdev->missing ? "<missing disk>" :
 			rcu_str_deref(dev_replace->srcdev->name),
 			dev_replace->srcdev->devid,
diff --git a/kernel/fs/btrfs/disk-io.c b/kernel/fs/btrfs/disk-io.c
index 2ef9a4b72..41fb43183 100644
--- a/kernel/fs/btrfs/disk-io.c
+++ b/kernel/fs/btrfs/disk-io.c
@@ -319,9 +319,9 @@ static int csum_tree_block(struct btrfs_fs_info *fs_info,
 			memcpy(&found, result, csum_size);
 
 			read_extent_buffer(buf, &val, 0, csum_size);
-			printk_ratelimited(KERN_WARNING
-				"BTRFS: %s checksum verify failed on %llu wanted %X found %X "
-				"level %d\n",
+			btrfs_warn_rl(fs_info,
+				"%s checksum verify failed on %llu wanted %X found %X "
+				"level %d",
 				fs_info->sb->s_id, buf->start,
 				val, found, btrfs_header_level(buf));
 			if (result != (char *)&inline_result)
@@ -368,9 +368,9 @@ static int verify_parent_transid(struct extent_io_tree *io_tree,
 		ret = 0;
 		goto out;
 	}
-	printk_ratelimited(KERN_ERR
-	    "BTRFS (device %s): parent transid verify failed on %llu wanted %llu found %llu\n",
-			eb->fs_info->sb->s_id, eb->start,
+	btrfs_err_rl(eb->fs_info,
+		"parent transid verify failed on %llu wanted %llu found %llu",
+			eb->start,
 			parent_transid, btrfs_header_generation(eb));
 	ret = 1;
 
@@ -629,15 +629,14 @@ static int btree_readpage_end_io_hook(struct btrfs_io_bio *io_bio,
 
 	found_start = btrfs_header_bytenr(eb);
 	if (found_start != eb->start) {
-		printk_ratelimited(KERN_ERR "BTRFS (device %s): bad tree block start "
-			       "%llu %llu\n",
-			       eb->fs_info->sb->s_id, found_start, eb->start);
+		btrfs_err_rl(eb->fs_info, "bad tree block start %llu %llu",
+			       found_start, eb->start);
 		ret = -EIO;
 		goto err;
 	}
 	if (check_tree_block_fsid(root->fs_info, eb)) {
-		printk_ratelimited(KERN_ERR "BTRFS (device %s): bad fsid on block %llu\n",
-			       eb->fs_info->sb->s_id, eb->start);
+		btrfs_err_rl(eb->fs_info, "bad fsid on block %llu",
+			       eb->start);
 		ret = -EIO;
 		goto err;
 	}
@@ -703,7 +702,7 @@ static int btree_io_failed_hook(struct page *page, int failed_mirror)
 	return -EIO;	/* we fixed nothing */
 }
 
-static void end_workqueue_bio(struct bio *bio, int err)
+static void end_workqueue_bio(struct bio *bio)
 {
 	struct btrfs_end_io_wq *end_io_wq = bio->bi_private;
 	struct btrfs_fs_info *fs_info;
@@ -711,7 +710,7 @@ static void end_workqueue_bio(struct bio *bio, int err)
 	btrfs_work_func_t func;
 
 	fs_info = end_io_wq->info;
-	end_io_wq->error = err;
+	end_io_wq->error = bio->bi_error;
 
 	if (bio->bi_rw & REQ_WRITE) {
 		if (end_io_wq->metadata == BTRFS_WQ_ENDIO_METADATA) {
@@ -802,13 +801,17 @@ static void run_one_async_done(struct btrfs_work *work)
 	limit = btrfs_async_submit_limit(fs_info);
 	limit = limit * 2 / 3;
 
+	/*
+	 * atomic_dec_return implies a barrier for waitqueue_active
+	 */
 	if (atomic_dec_return(&fs_info->nr_async_submits) < limit &&
 	    waitqueue_active(&fs_info->async_submit_wait))
 		wake_up(&fs_info->async_submit_wait);
 
 	/* If an error occured we just want to clean up the bio and move on */
 	if (async->error) {
-		bio_endio(async->bio, async->error);
+		async->bio->bi_error = async->error;
+		bio_endio(async->bio);
 		return;
 	}
 
@@ -908,8 +911,10 @@ static int __btree_submit_bio_done(struct inode *inode, int rw, struct bio *bio,
 	 * submission context.  Just jump into btrfs_map_bio
 	 */
 	ret = btrfs_map_bio(BTRFS_I(inode)->root, rw, bio, mirror_num, 1);
-	if (ret)
-		bio_endio(bio, ret);
+	if (ret) {
+		bio->bi_error = ret;
+		bio_endio(bio);
+	}
 	return ret;
 }
 
@@ -960,10 +965,13 @@ static int btree_submit_bio_hook(struct inode *inode, int rw, struct bio *bio,
 					  __btree_submit_bio_done);
 	}
 
-	if (ret) {
+	if (ret)
+		goto out_w_error;
+	return 0;
+
 out_w_error:
-		bio_endio(bio, ret);
-	}
+	bio->bi_error = ret;
+	bio_endio(bio);
 	return ret;
 }
 
@@ -1149,12 +1157,12 @@ struct extent_buffer *read_tree_block(struct btrfs_root *root, u64 bytenr,
 
 	buf = btrfs_find_create_tree_block(root, bytenr);
 	if (!buf)
-		return NULL;
+		return ERR_PTR(-ENOMEM);
 
 	ret = btree_read_extent_buffer_pages(root, buf, 0, parent_transid);
 	if (ret) {
 		free_extent_buffer(buf);
-		return NULL;
+		return ERR_PTR(ret);
 	}
 	return buf;
 
@@ -1259,6 +1267,7 @@ static void __setup_root(u32 nodesize, u32 sectorsize, u32 stripesize,
 	atomic_set(&root->orphan_inodes, 0);
 	atomic_set(&root->refs, 1);
 	atomic_set(&root->will_be_snapshoted, 0);
+	atomic_set(&root->qgroup_meta_rsv, 0);
 	root->log_transid = 0;
 	root->log_transid_committed = -1;
 	root->last_log_commit = 0;
@@ -1509,20 +1518,19 @@ static struct btrfs_root *btrfs_read_tree_root(struct btrfs_root *tree_root,
 	generation = btrfs_root_generation(&root->root_item);
 	root->node = read_tree_block(root, btrfs_root_bytenr(&root->root_item),
 				     generation);
-	if (!root->node) {
-		ret = -ENOMEM;
+	if (IS_ERR(root->node)) {
+		ret = PTR_ERR(root->node);
 		goto find_fail;
 	} else if (!btrfs_buffer_uptodate(root->node, generation, 0)) {
 		ret = -EIO;
-		goto read_fail;
+		free_extent_buffer(root->node);
+		goto find_fail;
 	}
 	root->commit_root = btrfs_root_node(root);
 out:
 	btrfs_free_path(path);
 	return root;
 
-read_fail:
-	free_extent_buffer(root->node);
 find_fail:
 	kfree(root);
 alloc_fail:
@@ -1574,8 +1582,23 @@ int btrfs_init_fs_root(struct btrfs_root *root)
 	ret = get_anon_bdev(&root->anon_dev);
 	if (ret)
 		goto free_writers;
+
+	mutex_lock(&root->objectid_mutex);
+	ret = btrfs_find_highest_objectid(root,
+					&root->highest_objectid);
+	if (ret) {
+		mutex_unlock(&root->objectid_mutex);
+		goto free_root_dev;
+	}
+
+	ASSERT(root->highest_objectid <= BTRFS_LAST_FREE_OBJECTID);
+
+	mutex_unlock(&root->objectid_mutex);
+
 	return 0;
 
+free_root_dev:
+	free_anon_bdev(root->anon_dev);
 free_writers:
 	btrfs_free_subvolume_writers(root->subv_writers);
 fail:
@@ -1725,6 +1748,7 @@ static int setup_bdi(struct btrfs_fs_info *info, struct backing_dev_info *bdi)
 	bdi->ra_pages = VM_MAX_READAHEAD * 1024 / PAGE_CACHE_SIZE;
 	bdi->congested_fn	= btrfs_congested_fn;
 	bdi->congested_data	= info;
+	bdi->capabilities |= BDI_CAP_CGROUP_WRITEBACK;
 	return 0;
 }
 
@@ -1736,22 +1760,22 @@ static void end_workqueue_fn(struct btrfs_work *work)
 {
 	struct bio *bio;
 	struct btrfs_end_io_wq *end_io_wq;
-	int error;
 
 	end_io_wq = container_of(work, struct btrfs_end_io_wq, work);
 	bio = end_io_wq->bio;
 
-	error = end_io_wq->error;
+	bio->bi_error = end_io_wq->error;
 	bio->bi_private = end_io_wq->private;
 	bio->bi_end_io = end_io_wq->end_io;
 	kmem_cache_free(btrfs_end_io_wq_cache, end_io_wq);
-	bio_endio_nodec(bio, error);
+	bio_endio(bio);
 }
 
 static int cleaner_kthread(void *arg)
 {
 	struct btrfs_root *root = arg;
 	int again;
+	struct btrfs_trans_handle *trans;
 
 	do {
 		again = 0;
@@ -1772,8 +1796,10 @@ static int cleaner_kthread(void *arg)
 			goto sleep;
 		}
 
+		mutex_lock(&root->fs_info->cleaner_delayed_iput_mutex);
 		btrfs_run_delayed_iputs(root);
-		btrfs_delete_unused_bgs(root->fs_info);
+		mutex_unlock(&root->fs_info->cleaner_delayed_iput_mutex);
+
 		again = btrfs_clean_one_deleted_snapshot(root);
 		mutex_unlock(&root->fs_info->cleaner_mutex);
 
@@ -1782,6 +1808,16 @@ static int cleaner_kthread(void *arg)
 		 * needn't do anything special here.
 		 */
 		btrfs_run_defrag_inodes(root->fs_info);
+
+		/*
+		 * Acquires fs_info->delete_unused_bgs_mutex to avoid racing
+		 * with relocation (btrfs_relocate_chunk) and relocation
+		 * acquires fs_info->cleaner_mutex (btrfs_relocate_block_group)
+		 * after acquiring fs_info->delete_unused_bgs_mutex. So we
+		 * can't hold, nor need to, fs_info->cleaner_mutex when deleting
+		 * unused block groups.
+		 */
+		btrfs_delete_unused_bgs(root->fs_info);
 sleep:
 		if (!try_to_freeze() && !again) {
 			set_current_state(TASK_INTERRUPTIBLE);
@@ -1790,6 +1826,34 @@ sleep:
 			__set_current_state(TASK_RUNNING);
 		}
 	} while (!kthread_should_stop());
+
+	/*
+	 * Transaction kthread is stopped before us and wakes us up.
+	 * However we might have started a new transaction and COWed some
+	 * tree blocks when deleting unused block groups for example. So
+	 * make sure we commit the transaction we started to have a clean
+	 * shutdown when evicting the btree inode - if it has dirty pages
+	 * when we do the final iput() on it, eviction will trigger a
+	 * writeback for it which will fail with null pointer dereferences
+	 * since work queues and other resources were already released and
+	 * destroyed by the time the iput/eviction/writeback is made.
+	 */
+	trans = btrfs_attach_transaction(root);
+	if (IS_ERR(trans)) {
+		if (PTR_ERR(trans) != -ENOENT)
+			btrfs_err(root->fs_info,
+				  "cleaner transaction attach returned %ld",
+				  PTR_ERR(trans));
+	} else {
+		int ret;
+
+		ret = btrfs_commit_transaction(trans, root);
+		if (ret)
+			btrfs_err(root->fs_info,
+				  "cleaner open transaction commit returned %d",
+				  ret);
+	}
+
 	return 0;
 }
 
@@ -2305,8 +2369,7 @@ static int btrfs_replay_log(struct btrfs_fs_info *fs_info,
 	u64 bytenr = btrfs_super_log_root(disk_super);
 
 	if (fs_devices->rw_devices == 0) {
-		printk(KERN_WARNING "BTRFS: log replay required "
-		       "on RO media\n");
+		btrfs_warn(fs_info, "log replay required on RO media");
 		return -EIO;
 	}
 
@@ -2320,9 +2383,13 @@ static int btrfs_replay_log(struct btrfs_fs_info *fs_info,
 
 	log_tree_root->node = read_tree_block(tree_root, bytenr,
 			fs_info->generation + 1);
-	if (!log_tree_root->node ||
-	    !extent_buffer_uptodate(log_tree_root->node)) {
-		printk(KERN_ERR "BTRFS: failed to read log tree\n");
+	if (IS_ERR(log_tree_root->node)) {
+		btrfs_warn(fs_info, "failed to read log tree");
+		ret = PTR_ERR(log_tree_root->node);
+		kfree(log_tree_root);
+		return ret;
+	} else if (!extent_buffer_uptodate(log_tree_root->node)) {
+		btrfs_err(fs_info, "failed to read log tree");
 		free_extent_buffer(log_tree_root->node);
 		kfree(log_tree_root);
 		return -EIO;
@@ -2330,7 +2397,7 @@ static int btrfs_replay_log(struct btrfs_fs_info *fs_info,
 	/* returns with log_tree_root freed on success */
 	ret = btrfs_recover_log_trees(log_tree_root);
 	if (ret) {
-		btrfs_error(tree_root->fs_info, ret,
+		btrfs_std_error(tree_root->fs_info, ret,
 			    "Failed to recover log tree");
 		free_extent_buffer(log_tree_root->node);
 		kfree(log_tree_root);
@@ -2489,12 +2556,12 @@ int open_ctree(struct super_block *sb,
 	spin_lock_init(&fs_info->unused_bgs_lock);
 	rwlock_init(&fs_info->tree_mod_log_lock);
 	mutex_init(&fs_info->unused_bg_unpin_mutex);
+	mutex_init(&fs_info->delete_unused_bgs_mutex);
 	mutex_init(&fs_info->reloc_mutex);
 	mutex_init(&fs_info->delalloc_root_mutex);
+	mutex_init(&fs_info->cleaner_delayed_iput_mutex);
 	seqlock_init(&fs_info->profiles_lock);
-	init_rwsem(&fs_info->delayed_iput_sem);
 
-	init_completion(&fs_info->kobj_unregister);
 	INIT_LIST_HEAD(&fs_info->dirty_cowonly_roots);
 	INIT_LIST_HEAD(&fs_info->space_info);
 	INIT_LIST_HEAD(&fs_info->tree_mod_seq_list);
@@ -2525,7 +2592,7 @@ int open_ctree(struct super_block *sb,
 	fs_info->commit_interval = BTRFS_DEFAULT_COMMIT_INTERVAL;
 	fs_info->avg_delayed_ref_runtime = NSEC_PER_SEC >> 6; /* div by 64 */
 	/* readahead state */
-	INIT_RADIX_TREE(&fs_info->reada_tree, GFP_NOFS & ~__GFP_WAIT);
+	INIT_RADIX_TREE(&fs_info->reada_tree, GFP_NOFS & ~__GFP_DIRECT_RECLAIM);
 	spin_lock_init(&fs_info->reada_lock);
 
 	fs_info->thread_pool_size = min_t(unsigned long,
@@ -2567,7 +2634,6 @@ int open_ctree(struct super_block *sb,
 
 
 	mutex_init(&fs_info->ordered_operations_mutex);
-	mutex_init(&fs_info->ordered_extent_flush_mutex);
 	mutex_init(&fs_info->tree_log_mutex);
 	mutex_init(&fs_info->chunk_mutex);
 	mutex_init(&fs_info->transaction_kthread_mutex);
@@ -2607,8 +2673,8 @@ int open_ctree(struct super_block *sb,
 	 * Read super block and check the signature bytes only
 	 */
 	bh = btrfs_read_dev_super(fs_devices->latest_bdev);
-	if (!bh) {
-		err = -EINVAL;
+	if (IS_ERR(bh)) {
+		err = PTR_ERR(bh);
 		goto fail_alloc;
 	}
 
@@ -2619,6 +2685,7 @@ int open_ctree(struct super_block *sb,
 	if (btrfs_check_super_csum(bh->b_data)) {
 		printk(KERN_ERR "BTRFS: superblock checksum mismatch\n");
 		err = -EINVAL;
+		brelse(bh);
 		goto fail_alloc;
 	}
 
@@ -2797,10 +2864,13 @@ int open_ctree(struct super_block *sb,
 	chunk_root->node = read_tree_block(chunk_root,
 					   btrfs_super_chunk_root(disk_super),
 					   generation);
-	if (!chunk_root->node ||
-	    !test_bit(EXTENT_BUFFER_UPTODATE, &chunk_root->node->bflags)) {
+	if (IS_ERR(chunk_root->node) ||
+	    !extent_buffer_uptodate(chunk_root->node)) {
 		printk(KERN_ERR "BTRFS: failed to read chunk root on %s\n",
 		       sb->s_id);
+		if (!IS_ERR(chunk_root->node))
+			free_extent_buffer(chunk_root->node);
+		chunk_root->node = NULL;
 		goto fail_tree_roots;
 	}
 	btrfs_set_root_node(&chunk_root->root_item, chunk_root->node);
@@ -2834,11 +2904,13 @@ retry_root_backup:
 	tree_root->node = read_tree_block(tree_root,
 					  btrfs_super_root(disk_super),
 					  generation);
-	if (!tree_root->node ||
-	    !test_bit(EXTENT_BUFFER_UPTODATE, &tree_root->node->bflags)) {
+	if (IS_ERR(tree_root->node) ||
+	    !extent_buffer_uptodate(tree_root->node)) {
 		printk(KERN_WARNING "BTRFS: failed to read tree root on %s\n",
 		       sb->s_id);
-
+		if (!IS_ERR(tree_root->node))
+			free_extent_buffer(tree_root->node);
+		tree_root->node = NULL;
 		goto recovery_tree_root;
 	}
 
@@ -2846,6 +2918,18 @@ retry_root_backup:
 	tree_root->commit_root = btrfs_root_node(tree_root);
 	btrfs_set_root_refs(&tree_root->root_item, 1);
 
+	mutex_lock(&tree_root->objectid_mutex);
+	ret = btrfs_find_highest_objectid(tree_root,
+					&tree_root->highest_objectid);
+	if (ret) {
+		mutex_unlock(&tree_root->objectid_mutex);
+		goto recovery_tree_root;
+	}
+
+	ASSERT(tree_root->highest_objectid <= BTRFS_LAST_FREE_OBJECTID);
+
+	mutex_unlock(&tree_root->objectid_mutex);
+
 	ret = btrfs_read_roots(fs_info, tree_root);
 	if (ret)
 		goto recovery_tree_root;
@@ -2874,12 +2958,24 @@ retry_root_backup:
 
 	btrfs_close_extra_devices(fs_devices, 1);
 
-	ret = btrfs_sysfs_add_one(fs_info);
+	ret = btrfs_sysfs_add_fsid(fs_devices, NULL);
 	if (ret) {
-		pr_err("BTRFS: failed to init sysfs interface: %d\n", ret);
+		pr_err("BTRFS: failed to init sysfs fsid interface: %d\n", ret);
 		goto fail_block_groups;
 	}
 
+	ret = btrfs_sysfs_add_device(fs_devices);
+	if (ret) {
+		pr_err("BTRFS: failed to init sysfs device interface: %d\n", ret);
+		goto fail_fsdev_sysfs;
+	}
+
+	ret = btrfs_sysfs_add_mounted(fs_info);
+	if (ret) {
+		pr_err("BTRFS: failed to init sysfs interface: %d\n", ret);
+		goto fail_fsdev_sysfs;
+	}
+
 	ret = btrfs_init_space_info(fs_info);
 	if (ret) {
 		printk(KERN_ERR "BTRFS: Failed to initial space info: %d\n", ret);
@@ -2896,8 +2992,9 @@ retry_root_backup:
 	if (fs_info->fs_devices->missing_devices >
 	     fs_info->num_tolerated_disk_barrier_failures &&
 	    !(sb->s_flags & MS_RDONLY)) {
-		printk(KERN_WARNING "BTRFS: "
-			"too many missing devices, writeable mount is not allowed\n");
+		pr_warn("BTRFS: missing devices(%llu) exceeds the limit(%d), writeable mount is not allowed\n",
+			fs_info->fs_devices->missing_devices,
+			fs_info->num_tolerated_disk_barrier_failures);
 		goto fail_sysfs;
 	}
 
@@ -3053,7 +3150,10 @@ fail_cleaner:
 	filemap_write_and_wait(fs_info->btree_inode->i_mapping);
 
 fail_sysfs:
-	btrfs_sysfs_remove_one(fs_info);
+	btrfs_sysfs_remove_mounted(fs_info);
+
+fail_fsdev_sysfs:
+	btrfs_sysfs_remove_fsid(fs_info->fs_devices);
 
 fail_block_groups:
 	btrfs_put_block_group_cache(fs_info);
@@ -3112,8 +3212,8 @@ static void btrfs_end_buffer_write_sync(struct buffer_head *bh, int uptodate)
 		struct btrfs_device *device = (struct btrfs_device *)
 			bh->b_private;
 
-		printk_ratelimited_in_rcu(KERN_WARNING "BTRFS: lost page write due to "
-					  "I/O error on %s\n",
+		btrfs_warn_rl_in_rcu(device->dev_root->fs_info,
+				"lost page write due to IO error on %s",
 					  rcu_str_deref(device->name));
 		/* note, we dont' set_buffer_write_io_error because we have
 		 * our own ways of dealing with the IO errors
@@ -3125,6 +3225,37 @@ static void btrfs_end_buffer_write_sync(struct buffer_head *bh, int uptodate)
 	put_bh(bh);
 }
 
+int btrfs_read_dev_one_super(struct block_device *bdev, int copy_num,
+			struct buffer_head **bh_ret)
+{
+	struct buffer_head *bh;
+	struct btrfs_super_block *super;
+	u64 bytenr;
+
+	bytenr = btrfs_sb_offset(copy_num);
+	if (bytenr + BTRFS_SUPER_INFO_SIZE >= i_size_read(bdev->bd_inode))
+		return -EINVAL;
+
+	bh = __bread(bdev, bytenr / 4096, BTRFS_SUPER_INFO_SIZE);
+	/*
+	 * If we fail to read from the underlying devices, as of now
+	 * the best option we have is to mark it EIO.
+	 */
+	if (!bh)
+		return -EIO;
+
+	super = (struct btrfs_super_block *)bh->b_data;
+	if (btrfs_super_bytenr(super) != bytenr ||
+		    btrfs_super_magic(super) != BTRFS_MAGIC) {
+		brelse(bh);
+		return -EINVAL;
+	}
+
+	*bh_ret = bh;
+	return 0;
+}
+
+
 struct buffer_head *btrfs_read_dev_super(struct block_device *bdev)
 {
 	struct buffer_head *bh;
@@ -3132,7 +3263,7 @@ struct buffer_head *btrfs_read_dev_super(struct block_device *bdev)
 	struct btrfs_super_block *super;
 	int i;
 	u64 transid = 0;
-	u64 bytenr;
+	int ret = -EINVAL;
 
 	/* we would like to check all the supers, but that would make
 	 * a btrfs mount succeed after a mkfs from a different FS.
@@ -3140,21 +3271,11 @@ struct buffer_head *btrfs_read_dev_super(struct block_device *bdev)
 	 * later supers, using BTRFS_SUPER_MIRROR_MAX instead
 	 */
 	for (i = 0; i < 1; i++) {
-		bytenr = btrfs_sb_offset(i);
-		if (bytenr + BTRFS_SUPER_INFO_SIZE >=
-					i_size_read(bdev->bd_inode))
-			break;
-		bh = __bread(bdev, bytenr / 4096,
-					BTRFS_SUPER_INFO_SIZE);
-		if (!bh)
+		ret = btrfs_read_dev_one_super(bdev, i, &bh);
+		if (ret)
 			continue;
 
 		super = (struct btrfs_super_block *)bh->b_data;
-		if (btrfs_super_bytenr(super) != bytenr ||
-		    btrfs_super_magic(super) != BTRFS_MAGIC) {
-			brelse(bh);
-			continue;
-		}
 
 		if (!latest || btrfs_super_generation(super) > transid) {
 			brelse(latest);
@@ -3164,6 +3285,10 @@ struct buffer_head *btrfs_read_dev_super(struct block_device *bdev)
 			brelse(bh);
 		}
 	}
+
+	if (!latest)
+		return ERR_PTR(ret);
+
 	return latest;
 }
 
@@ -3232,8 +3357,9 @@ static int write_dev_supers(struct btrfs_device *device,
 			bh = __getblk(device->bdev, bytenr / 4096,
 				      BTRFS_SUPER_INFO_SIZE);
 			if (!bh) {
-				printk(KERN_ERR "BTRFS: couldn't get super "
-				       "buffer head for bytenr %Lu\n", bytenr);
+				btrfs_err(device->dev_root->fs_info,
+				    "couldn't get super buffer head for bytenr %llu",
+				    bytenr);
 				errors++;
 				continue;
 			}
@@ -3267,13 +3393,8 @@ static int write_dev_supers(struct btrfs_device *device,
  * endio for the write_dev_flush, this will wake anyone waiting
  * for the barrier when it is done
  */
-static void btrfs_end_empty_barrier(struct bio *bio, int err)
+static void btrfs_end_empty_barrier(struct bio *bio)
 {
-	if (err) {
-		if (err == -EOPNOTSUPP)
-			set_bit(BIO_EOPNOTSUPP, &bio->bi_flags);
-		clear_bit(BIO_UPTODATE, &bio->bi_flags);
-	}
 	if (bio->bi_private)
 		complete(bio->bi_private);
 	bio_put(bio);
@@ -3301,12 +3422,8 @@ static int write_dev_flush(struct btrfs_device *device, int wait)
 
 		wait_for_completion(&device->flush_wait);
 
-		if (bio_flagged(bio, BIO_EOPNOTSUPP)) {
-			printk_in_rcu("BTRFS: disabling barriers on dev %s\n",
-				      rcu_str_deref(device->name));
-			device->nobarriers = 1;
-		} else if (!bio_flagged(bio, BIO_UPTODATE)) {
-			ret = -EIO;
+		if (bio->bi_error) {
+			ret = bio->bi_error;
 			btrfs_dev_stat_inc_and_print(device,
 				BTRFS_DEV_STAT_FLUSH_ERRS);
 		}
@@ -3389,6 +3506,35 @@ static int barrier_all_devices(struct btrfs_fs_info *info)
 	return 0;
 }
 
+int btrfs_get_num_tolerated_disk_barrier_failures(u64 flags)
+{
+	int raid_type;
+	int min_tolerated = INT_MAX;
+
+	if ((flags & BTRFS_BLOCK_GROUP_PROFILE_MASK) == 0 ||
+	    (flags & BTRFS_AVAIL_ALLOC_BIT_SINGLE))
+		min_tolerated = min(min_tolerated,
+				    btrfs_raid_array[BTRFS_RAID_SINGLE].
+				    tolerated_failures);
+
+	for (raid_type = 0; raid_type < BTRFS_NR_RAID_TYPES; raid_type++) {
+		if (raid_type == BTRFS_RAID_SINGLE)
+			continue;
+		if (!(flags & btrfs_raid_group[raid_type]))
+			continue;
+		min_tolerated = min(min_tolerated,
+				    btrfs_raid_array[raid_type].
+				    tolerated_failures);
+	}
+
+	if (min_tolerated == INT_MAX) {
+		pr_warn("BTRFS: unknown raid flag: %llu\n", flags);
+		min_tolerated = 0;
+	}
+
+	return min_tolerated;
+}
+
 int btrfs_calc_num_tolerated_disk_barrier_failures(
 	struct btrfs_fs_info *fs_info)
 {
@@ -3398,13 +3544,12 @@ int btrfs_calc_num_tolerated_disk_barrier_failures(
 		       BTRFS_BLOCK_GROUP_SYSTEM,
 		       BTRFS_BLOCK_GROUP_METADATA,
 		       BTRFS_BLOCK_GROUP_DATA | BTRFS_BLOCK_GROUP_METADATA};
-	int num_types = 4;
 	int i;
 	int c;
 	int num_tolerated_disk_barrier_failures =
 		(int)fs_info->fs_devices->num_devices;
 
-	for (i = 0; i < num_types; i++) {
+	for (i = 0; i < ARRAY_SIZE(types); i++) {
 		struct btrfs_space_info *tmp;
 
 		sinfo = NULL;
@@ -3422,44 +3567,21 @@ int btrfs_calc_num_tolerated_disk_barrier_failures(
 
 		down_read(&sinfo->groups_sem);
 		for (c = 0; c < BTRFS_NR_RAID_TYPES; c++) {
-			if (!list_empty(&sinfo->block_groups[c])) {
-				u64 flags;
-
-				btrfs_get_block_group_info(
-					&sinfo->block_groups[c], &space);
-				if (space.total_bytes == 0 ||
-				    space.used_bytes == 0)
-					continue;
-				flags = space.flags;
-				/*
-				 * return
-				 * 0: if dup, single or RAID0 is configured for
-				 *    any of metadata, system or data, else
-				 * 1: if RAID5 is configured, or if RAID1 or
-				 *    RAID10 is configured and only two mirrors
-				 *    are used, else
-				 * 2: if RAID6 is configured, else
-				 * num_mirrors - 1: if RAID1 or RAID10 is
-				 *                  configured and more than
-				 *                  2 mirrors are used.
-				 */
-				if (num_tolerated_disk_barrier_failures > 0 &&
-				    ((flags & (BTRFS_BLOCK_GROUP_DUP |
-					       BTRFS_BLOCK_GROUP_RAID0)) ||
-				     ((flags & BTRFS_BLOCK_GROUP_PROFILE_MASK)
-				      == 0)))
-					num_tolerated_disk_barrier_failures = 0;
-				else if (num_tolerated_disk_barrier_failures > 1) {
-					if (flags & (BTRFS_BLOCK_GROUP_RAID1 |
-					    BTRFS_BLOCK_GROUP_RAID5 |
-					    BTRFS_BLOCK_GROUP_RAID10)) {
-						num_tolerated_disk_barrier_failures = 1;
-					} else if (flags &
-						   BTRFS_BLOCK_GROUP_RAID6) {
-						num_tolerated_disk_barrier_failures = 2;
-					}
-				}
-			}
+			u64 flags;
+
+			if (list_empty(&sinfo->block_groups[c]))
+				continue;
+
+			btrfs_get_block_group_info(&sinfo->block_groups[c],
+						   &space);
+			if (space.total_bytes == 0 || space.used_bytes == 0)
+				continue;
+			flags = space.flags;
+
+			num_tolerated_disk_barrier_failures = min(
+				num_tolerated_disk_barrier_failures,
+				btrfs_get_num_tolerated_disk_barrier_failures(
+					flags));
 		}
 		up_read(&sinfo->groups_sem);
 	}
@@ -3494,7 +3616,7 @@ static int write_all_supers(struct btrfs_root *root, int max_mirrors)
 		if (ret) {
 			mutex_unlock(
 				&root->fs_info->fs_devices->device_list_mutex);
-			btrfs_error(root->fs_info, ret,
+			btrfs_std_error(root->fs_info, ret,
 				    "errors while submitting device barriers.");
 			return ret;
 		}
@@ -3534,7 +3656,7 @@ static int write_all_supers(struct btrfs_root *root, int max_mirrors)
 		mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
 
 		/* FUA is masked off if unsupported and can't be the reason */
-		btrfs_error(root->fs_info, -EIO,
+		btrfs_std_error(root->fs_info, -EIO,
 			    "%d errors while writing supers", total_errors);
 		return -EIO;
 	}
@@ -3552,7 +3674,7 @@ static int write_all_supers(struct btrfs_root *root, int max_mirrors)
 	}
 	mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
 	if (total_errors > max_errors) {
-		btrfs_error(root->fs_info, -EIO,
+		btrfs_std_error(root->fs_info, -EIO,
 			    "%d errors while writing supers", total_errors);
 		return -EIO;
 	}
@@ -3688,6 +3810,9 @@ void close_ctree(struct btrfs_root *root)
 	fs_info->closing = 1;
 	smp_mb();
 
+	/* wait for the qgroup rescan worker to stop */
+	btrfs_qgroup_wait_for_completion(fs_info);
+
 	/* wait for the uuid_scan task to finish */
 	down(&fs_info->uuid_tree_rescan_sem);
 	/* avoid complains from lockdep et al., set sem back to initial state */
@@ -3710,6 +3835,13 @@ void close_ctree(struct btrfs_root *root)
 	cancel_work_sync(&fs_info->async_reclaim_work);
 
 	if (!(fs_info->sb->s_flags & MS_RDONLY)) {
+		/*
+		 * If the cleaner thread is stopped and there are
+		 * block groups queued for removal, the deletion will be
+		 * skipped when we quit the cleaner thread.
+		 */
+		btrfs_delete_unused_bgs(root->fs_info);
+
 		ret = btrfs_commit_super(root);
 		if (ret)
 			btrfs_err(fs_info, "commit super ret %d", ret);
@@ -3731,7 +3863,8 @@ void close_ctree(struct btrfs_root *root)
 		       percpu_counter_sum(&fs_info->delalloc_bytes));
 	}
 
-	btrfs_sysfs_remove_one(fs_info);
+	btrfs_sysfs_remove_mounted(fs_info);
+	btrfs_sysfs_remove_fsid(fs_info->fs_devices);
 
 	btrfs_free_fs_roots(fs_info);
 
@@ -4060,6 +4193,7 @@ static int btrfs_destroy_delayed_refs(struct btrfs_transaction *trans,
 
 	while ((node = rb_first(&delayed_refs->href_root)) != NULL) {
 		struct btrfs_delayed_ref_head *head;
+		struct btrfs_delayed_ref_node *tmp;
 		bool pin_bytes = false;
 
 		head = rb_entry(node, struct btrfs_delayed_ref_head,
@@ -4075,11 +4209,10 @@ static int btrfs_destroy_delayed_refs(struct btrfs_transaction *trans,
 			continue;
 		}
 		spin_lock(&head->lock);
-		while ((node = rb_first(&head->ref_root)) != NULL) {
-			ref = rb_entry(node, struct btrfs_delayed_ref_node,
-				       rb_node);
+		list_for_each_entry_safe_reverse(ref, tmp, &head->ref_list,
+						 list) {
 			ref->in_tree = 0;
-			rb_erase(&ref->rb_node, &head->ref_root);
+			list_del(&ref->list);
 			atomic_dec(&delayed_refs->num_entries);
 			btrfs_put_delayed_ref(ref);
 		}
@@ -4228,25 +4361,6 @@ again:
 	return 0;
 }
 
-static void btrfs_free_pending_ordered(struct btrfs_transaction *cur_trans,
-				       struct btrfs_fs_info *fs_info)
-{
-	struct btrfs_ordered_extent *ordered;
-
-	spin_lock(&fs_info->trans_lock);
-	while (!list_empty(&cur_trans->pending_ordered)) {
-		ordered = list_first_entry(&cur_trans->pending_ordered,
-					   struct btrfs_ordered_extent,
-					   trans_list);
-		list_del_init(&ordered->trans_list);
-		spin_unlock(&fs_info->trans_lock);
-
-		btrfs_put_ordered_extent(ordered);
-		spin_lock(&fs_info->trans_lock);
-	}
-	spin_unlock(&fs_info->trans_lock);
-}
-
 void btrfs_cleanup_one_transaction(struct btrfs_transaction *cur_trans,
 				   struct btrfs_root *root)
 {
@@ -4258,7 +4372,6 @@ void btrfs_cleanup_one_transaction(struct btrfs_transaction *cur_trans,
 	cur_trans->state = TRANS_STATE_UNBLOCKED;
 	wake_up(&root->fs_info->transaction_wait);
 
-	btrfs_free_pending_ordered(cur_trans, root->fs_info);
 	btrfs_destroy_delayed_inodes(root);
 	btrfs_assert_delayed_root_empty(root);
 
diff --git a/kernel/fs/btrfs/disk-io.h b/kernel/fs/btrfs/disk-io.h
index d4cbfeeee..adeb31830 100644
--- a/kernel/fs/btrfs/disk-io.h
+++ b/kernel/fs/btrfs/disk-io.h
@@ -60,6 +60,8 @@ void close_ctree(struct btrfs_root *root);
 int write_ctree_super(struct btrfs_trans_handle *trans,
 		      struct btrfs_root *root, int max_mirrors);
 struct buffer_head *btrfs_read_dev_super(struct block_device *bdev);
+int btrfs_read_dev_one_super(struct block_device *bdev, int copy_num,
+			struct buffer_head **bh_ret);
 int btrfs_commit_super(struct btrfs_root *root);
 struct extent_buffer *btrfs_find_tree_block(struct btrfs_fs_info *fs_info,
 					    u64 bytenr);
@@ -139,6 +141,7 @@ struct btrfs_root *btrfs_create_tree(struct btrfs_trans_handle *trans,
 				     u64 objectid);
 int btree_lock_page_hook(struct page *page, void *data,
 				void (*flush_fn)(void *));
+int btrfs_get_num_tolerated_disk_barrier_failures(u64 flags);
 int btrfs_calc_num_tolerated_disk_barrier_failures(
 	struct btrfs_fs_info *fs_info);
 int __init btrfs_end_io_wq_init(void);
diff --git a/kernel/fs/btrfs/export.c b/kernel/fs/btrfs/export.c
index 8d052209f..2513a7f53 100644
--- a/kernel/fs/btrfs/export.c
+++ b/kernel/fs/btrfs/export.c
@@ -112,11 +112,11 @@ static struct dentry *btrfs_fh_to_parent(struct super_block *sb, struct fid *fh,
 	u32 generation;
 
 	if (fh_type == FILEID_BTRFS_WITH_PARENT) {
-		if (fh_len !=  BTRFS_FID_SIZE_CONNECTABLE)
+		if (fh_len <  BTRFS_FID_SIZE_CONNECTABLE)
 			return NULL;
 		root_objectid = fid->root_objectid;
 	} else if (fh_type == FILEID_BTRFS_WITH_PARENT_ROOT) {
-		if (fh_len != BTRFS_FID_SIZE_CONNECTABLE_ROOT)
+		if (fh_len < BTRFS_FID_SIZE_CONNECTABLE_ROOT)
 			return NULL;
 		root_objectid = fid->parent_root_objectid;
 	} else
@@ -136,11 +136,11 @@ static struct dentry *btrfs_fh_to_dentry(struct super_block *sb, struct fid *fh,
 	u32 generation;
 
 	if ((fh_type != FILEID_BTRFS_WITH_PARENT ||
-	     fh_len != BTRFS_FID_SIZE_CONNECTABLE) &&
+	     fh_len < BTRFS_FID_SIZE_CONNECTABLE) &&
 	    (fh_type != FILEID_BTRFS_WITH_PARENT_ROOT ||
-	     fh_len != BTRFS_FID_SIZE_CONNECTABLE_ROOT) &&
+	     fh_len < BTRFS_FID_SIZE_CONNECTABLE_ROOT) &&
 	    (fh_type != FILEID_BTRFS_WITHOUT_PARENT ||
-	     fh_len != BTRFS_FID_SIZE_NON_CONNECTABLE))
+	     fh_len < BTRFS_FID_SIZE_NON_CONNECTABLE))
 		return NULL;
 
 	objectid = fid->objectid;
diff --git a/kernel/fs/btrfs/extent-tree.c b/kernel/fs/btrfs/extent-tree.c
index 0ec3acd14..2368cac11 100644
--- a/kernel/fs/btrfs/extent-tree.c
+++ b/kernel/fs/btrfs/extent-tree.c
@@ -79,11 +79,10 @@ static int update_block_group(struct btrfs_trans_handle *trans,
 			      u64 num_bytes, int alloc);
 static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
 				struct btrfs_root *root,
-				u64 bytenr, u64 num_bytes, u64 parent,
+				struct btrfs_delayed_ref_node *node, u64 parent,
 				u64 root_objectid, u64 owner_objectid,
 				u64 owner_offset, int refs_to_drop,
-				struct btrfs_delayed_extent_op *extra_op,
-				int no_quota);
+				struct btrfs_delayed_extent_op *extra_op);
 static void __run_delayed_extent_op(struct btrfs_delayed_extent_op *extent_op,
 				    struct extent_buffer *leaf,
 				    struct btrfs_extent_item *ei);
@@ -96,8 +95,7 @@ static int alloc_reserved_tree_block(struct btrfs_trans_handle *trans,
 				     struct btrfs_root *root,
 				     u64 parent, u64 root_objectid,
 				     u64 flags, struct btrfs_disk_key *key,
-				     int level, struct btrfs_key *ins,
-				     int no_quota);
+				     int level, struct btrfs_key *ins);
 static int do_chunk_alloc(struct btrfs_trans_handle *trans,
 			  struct btrfs_root *extent_root, u64 flags,
 			  int force);
@@ -126,7 +124,7 @@ static int block_group_bits(struct btrfs_block_group_cache *cache, u64 bits)
 	return (cache->flags & bits) == bits;
 }
 
-static void btrfs_get_block_group(struct btrfs_block_group_cache *cache)
+void btrfs_get_block_group(struct btrfs_block_group_cache *cache)
 {
 	atomic_inc(&cache->count);
 }
@@ -333,6 +331,27 @@ static void put_caching_control(struct btrfs_caching_control *ctl)
 		kfree(ctl);
 }
 
+#ifdef CONFIG_BTRFS_DEBUG
+static void fragment_free_space(struct btrfs_root *root,
+				struct btrfs_block_group_cache *block_group)
+{
+	u64 start = block_group->key.objectid;
+	u64 len = block_group->key.offset;
+	u64 chunk = block_group->flags & BTRFS_BLOCK_GROUP_METADATA ?
+		root->nodesize : root->sectorsize;
+	u64 step = chunk << 1;
+
+	while (len > chunk) {
+		btrfs_remove_free_space(block_group, start, chunk);
+		start += step;
+		if (len < step)
+			len = 0;
+		else
+			len -= step;
+	}
+}
+#endif
+
 /*
  * this is only called by cache_block_group, since we could have freed extents
  * we need to check the pinned_extents for any extents that can't be used yet
@@ -389,6 +408,7 @@ static noinline void caching_thread(struct btrfs_work *work)
 	u64 last = 0;
 	u32 nritems;
 	int ret = -ENOMEM;
+	bool wakeup = true;
 
 	caching_ctl = container_of(work, struct btrfs_caching_control, work);
 	block_group = caching_ctl->block_group;
@@ -401,6 +421,15 @@ static noinline void caching_thread(struct btrfs_work *work)
 
 	last = max_t(u64, block_group->key.objectid, BTRFS_SUPER_INFO_OFFSET);
 
+#ifdef CONFIG_BTRFS_DEBUG
+	/*
+	 * If we're fragmenting we don't want to make anybody think we can
+	 * allocate from this block group until we've had a chance to fragment
+	 * the free space.
+	 */
+	if (btrfs_should_fragment_free_space(extent_root, block_group))
+		wakeup = false;
+#endif
 	/*
 	 * We don't want to deadlock with somebody trying to allocate a new
 	 * extent for the extent root while also trying to search the extent
@@ -442,7 +471,8 @@ next:
 
 			if (need_resched() ||
 			    rwsem_is_contended(&fs_info->commit_root_sem)) {
-				caching_ctl->progress = last;
+				if (wakeup)
+					caching_ctl->progress = last;
 				btrfs_release_path(path);
 				up_read(&fs_info->commit_root_sem);
 				mutex_unlock(&caching_ctl->mutex);
@@ -465,7 +495,8 @@ next:
 			key.offset = 0;
 			key.type = BTRFS_EXTENT_ITEM_KEY;
 
-			caching_ctl->progress = last;
+			if (wakeup)
+				caching_ctl->progress = last;
 			btrfs_release_path(path);
 			goto next;
 		}
@@ -492,7 +523,8 @@ next:
 
 			if (total_found > (1024 * 1024 * 2)) {
 				total_found = 0;
-				wake_up(&caching_ctl->wait);
+				if (wakeup)
+					wake_up(&caching_ctl->wait);
 			}
 		}
 		path->slots[0]++;
@@ -502,13 +534,27 @@ next:
 	total_found += add_new_free_space(block_group, fs_info, last,
 					  block_group->key.objectid +
 					  block_group->key.offset);
-	caching_ctl->progress = (u64)-1;
-
 	spin_lock(&block_group->lock);
 	block_group->caching_ctl = NULL;
 	block_group->cached = BTRFS_CACHE_FINISHED;
 	spin_unlock(&block_group->lock);
 
+#ifdef CONFIG_BTRFS_DEBUG
+	if (btrfs_should_fragment_free_space(extent_root, block_group)) {
+		u64 bytes_used;
+
+		spin_lock(&block_group->space_info->lock);
+		spin_lock(&block_group->lock);
+		bytes_used = block_group->key.offset -
+			btrfs_block_group_used(&block_group->item);
+		block_group->space_info->bytes_used += bytes_used >> 1;
+		spin_unlock(&block_group->lock);
+		spin_unlock(&block_group->space_info->lock);
+		fragment_free_space(extent_root, block_group);
+	}
+#endif
+
+	caching_ctl->progress = (u64)-1;
 err:
 	btrfs_free_path(path);
 	up_read(&fs_info->commit_root_sem);
@@ -608,6 +654,22 @@ static int cache_block_group(struct btrfs_block_group_cache *cache,
 			}
 		}
 		spin_unlock(&cache->lock);
+#ifdef CONFIG_BTRFS_DEBUG
+		if (ret == 1 &&
+		    btrfs_should_fragment_free_space(fs_info->extent_root,
+						     cache)) {
+			u64 bytes_used;
+
+			spin_lock(&cache->space_info->lock);
+			spin_lock(&cache->lock);
+			bytes_used = cache->key.offset -
+				btrfs_block_group_used(&cache->item);
+			cache->space_info->bytes_used += bytes_used >> 1;
+			spin_unlock(&cache->lock);
+			spin_unlock(&cache->space_info->lock);
+			fragment_free_space(fs_info->extent_root, cache);
+		}
+#endif
 		mutex_unlock(&caching_ctl->mutex);
 
 		wake_up(&caching_ctl->wait);
@@ -1317,8 +1379,7 @@ static noinline int remove_extent_data_ref(struct btrfs_trans_handle *trans,
 	return ret;
 }
 
-static noinline u32 extent_data_ref_count(struct btrfs_root *root,
-					  struct btrfs_path *path,
+static noinline u32 extent_data_ref_count(struct btrfs_path *path,
 					  struct btrfs_extent_inline_ref *iref)
 {
 	struct btrfs_key key;
@@ -1884,10 +1945,77 @@ static int remove_extent_backref(struct btrfs_trans_handle *trans,
 	return ret;
 }
 
-static int btrfs_issue_discard(struct block_device *bdev,
-				u64 start, u64 len)
+#define in_range(b, first, len)        ((b) >= (first) && (b) < (first) + (len))
+static int btrfs_issue_discard(struct block_device *bdev, u64 start, u64 len,
+			       u64 *discarded_bytes)
 {
-	return blkdev_issue_discard(bdev, start >> 9, len >> 9, GFP_NOFS, 0);
+	int j, ret = 0;
+	u64 bytes_left, end;
+	u64 aligned_start = ALIGN(start, 1 << 9);
+
+	if (WARN_ON(start != aligned_start)) {
+		len -= aligned_start - start;
+		len = round_down(len, 1 << 9);
+		start = aligned_start;
+	}
+
+	*discarded_bytes = 0;
+
+	if (!len)
+		return 0;
+
+	end = start + len;
+	bytes_left = len;
+
+	/* Skip any superblocks on this device. */
+	for (j = 0; j < BTRFS_SUPER_MIRROR_MAX; j++) {
+		u64 sb_start = btrfs_sb_offset(j);
+		u64 sb_end = sb_start + BTRFS_SUPER_INFO_SIZE;
+		u64 size = sb_start - start;
+
+		if (!in_range(sb_start, start, bytes_left) &&
+		    !in_range(sb_end, start, bytes_left) &&
+		    !in_range(start, sb_start, BTRFS_SUPER_INFO_SIZE))
+			continue;
+
+		/*
+		 * Superblock spans beginning of range.  Adjust start and
+		 * try again.
+		 */
+		if (sb_start <= start) {
+			start += sb_end - start;
+			if (start > end) {
+				bytes_left = 0;
+				break;
+			}
+			bytes_left = end - start;
+			continue;
+		}
+
+		if (size) {
+			ret = blkdev_issue_discard(bdev, start >> 9, size >> 9,
+						   GFP_NOFS, 0);
+			if (!ret)
+				*discarded_bytes += size;
+			else if (ret != -EOPNOTSUPP)
+				return ret;
+		}
+
+		start = sb_end;
+		if (start > end) {
+			bytes_left = 0;
+			break;
+		}
+		bytes_left = end - start;
+	}
+
+	if (bytes_left) {
+		ret = blkdev_issue_discard(bdev, start >> 9, bytes_left >> 9,
+					   GFP_NOFS, 0);
+		if (!ret)
+			*discarded_bytes += bytes_left;
+	}
+	return ret;
 }
 
 int btrfs_discard_extent(struct btrfs_root *root, u64 bytenr,
@@ -1908,14 +2036,16 @@ int btrfs_discard_extent(struct btrfs_root *root, u64 bytenr,
 
 
 		for (i = 0; i < bbio->num_stripes; i++, stripe++) {
+			u64 bytes;
 			if (!stripe->dev->can_discard)
 				continue;
 
 			ret = btrfs_issue_discard(stripe->dev->bdev,
 						  stripe->physical,
-						  stripe->length);
+						  stripe->length,
+						  &bytes);
 			if (!ret)
-				discarded_bytes += stripe->length;
+				discarded_bytes += bytes;
 			else if (ret != -EOPNOTSUPP)
 				break; /* Logic errors or -ENOMEM, or -EIO but I don't know how that could happen JDM */
 
@@ -1942,8 +2072,7 @@ int btrfs_discard_extent(struct btrfs_root *root, u64 bytenr,
 int btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
 			 struct btrfs_root *root,
 			 u64 bytenr, u64 num_bytes, u64 parent,
-			 u64 root_objectid, u64 owner, u64 offset,
-			 int no_quota)
+			 u64 root_objectid, u64 owner, u64 offset)
 {
 	int ret;
 	struct btrfs_fs_info *fs_info = root->fs_info;
@@ -1955,22 +2084,21 @@ int btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
 		ret = btrfs_add_delayed_tree_ref(fs_info, trans, bytenr,
 					num_bytes,
 					parent, root_objectid, (int)owner,
-					BTRFS_ADD_DELAYED_REF, NULL, no_quota);
+					BTRFS_ADD_DELAYED_REF, NULL);
 	} else {
 		ret = btrfs_add_delayed_data_ref(fs_info, trans, bytenr,
-					num_bytes,
-					parent, root_objectid, owner, offset,
-					BTRFS_ADD_DELAYED_REF, NULL, no_quota);
+					num_bytes, parent, root_objectid,
+					owner, offset, 0,
+					BTRFS_ADD_DELAYED_REF, NULL);
 	}
 	return ret;
 }
 
 static int __btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
 				  struct btrfs_root *root,
-				  u64 bytenr, u64 num_bytes,
+				  struct btrfs_delayed_ref_node *node,
 				  u64 parent, u64 root_objectid,
 				  u64 owner, u64 offset, int refs_to_add,
-				  int no_quota,
 				  struct btrfs_delayed_extent_op *extent_op)
 {
 	struct btrfs_fs_info *fs_info = root->fs_info;
@@ -1978,17 +2106,15 @@ static int __btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
 	struct extent_buffer *leaf;
 	struct btrfs_extent_item *item;
 	struct btrfs_key key;
+	u64 bytenr = node->bytenr;
+	u64 num_bytes = node->num_bytes;
 	u64 refs;
 	int ret;
-	enum btrfs_qgroup_operation_type type = BTRFS_QGROUP_OPER_ADD_EXCL;
 
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
 
-	if (!is_fstree(root_objectid) || !root->fs_info->quota_enabled)
-		no_quota = 1;
-
 	path->reada = 1;
 	path->leave_spinning = 1;
 	/* this will setup the path even if it fails to insert the back ref */
@@ -1996,26 +2122,8 @@ static int __btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
 					   bytenr, num_bytes, parent,
 					   root_objectid, owner, offset,
 					   refs_to_add, extent_op);
-	if ((ret < 0 && ret != -EAGAIN) || (!ret && no_quota))
+	if ((ret < 0 && ret != -EAGAIN) || !ret)
 		goto out;
-	/*
-	 * Ok we were able to insert an inline extent and it appears to be a new
-	 * reference, deal with the qgroup accounting.
-	 */
-	if (!ret && !no_quota) {
-		ASSERT(root->fs_info->quota_enabled);
-		leaf = path->nodes[0];
-		btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
-		item = btrfs_item_ptr(leaf, path->slots[0],
-				      struct btrfs_extent_item);
-		if (btrfs_extent_refs(leaf, item) > (u64)refs_to_add)
-			type = BTRFS_QGROUP_OPER_ADD_SHARED;
-		btrfs_release_path(path);
-
-		ret = btrfs_qgroup_record_ref(trans, fs_info, root_objectid,
-					      bytenr, num_bytes, type, 0);
-		goto out;
-	}
 
 	/*
 	 * Ok we had -EAGAIN which means we didn't have space to insert and
@@ -2026,8 +2134,6 @@ static int __btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
 	btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
 	item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
 	refs = btrfs_extent_refs(leaf, item);
-	if (refs)
-		type = BTRFS_QGROUP_OPER_ADD_SHARED;
 	btrfs_set_extent_refs(leaf, item, refs + refs_to_add);
 	if (extent_op)
 		__run_delayed_extent_op(extent_op, leaf, item);
@@ -2035,13 +2141,6 @@ static int __btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
 	btrfs_mark_buffer_dirty(leaf);
 	btrfs_release_path(path);
 
-	if (!no_quota) {
-		ret = btrfs_qgroup_record_ref(trans, fs_info, root_objectid,
-					      bytenr, num_bytes, type, 0);
-		if (ret)
-			goto out;
-	}
-
 	path->reada = 1;
 	path->leave_spinning = 1;
 	/* now insert the actual backref */
@@ -2087,17 +2186,15 @@ static int run_delayed_data_ref(struct btrfs_trans_handle *trans,
 						 ref->objectid, ref->offset,
 						 &ins, node->ref_mod);
 	} else if (node->action == BTRFS_ADD_DELAYED_REF) {
-		ret = __btrfs_inc_extent_ref(trans, root, node->bytenr,
-					     node->num_bytes, parent,
+		ret = __btrfs_inc_extent_ref(trans, root, node, parent,
 					     ref_root, ref->objectid,
 					     ref->offset, node->ref_mod,
-					     node->no_quota, extent_op);
+					     extent_op);
 	} else if (node->action == BTRFS_DROP_DELAYED_REF) {
-		ret = __btrfs_free_extent(trans, root, node->bytenr,
-					  node->num_bytes, parent,
+		ret = __btrfs_free_extent(trans, root, node, parent,
 					  ref_root, ref->objectid,
 					  ref->offset, node->ref_mod,
-					  extent_op, node->no_quota);
+					  extent_op);
 	} else {
 		BUG();
 	}
@@ -2252,18 +2349,16 @@ static int run_delayed_tree_ref(struct btrfs_trans_handle *trans,
 						parent, ref_root,
 						extent_op->flags_to_set,
 						&extent_op->key,
-						ref->level, &ins,
-						node->no_quota);
+						ref->level, &ins);
 	} else if (node->action == BTRFS_ADD_DELAYED_REF) {
-		ret = __btrfs_inc_extent_ref(trans, root, node->bytenr,
-					     node->num_bytes, parent, ref_root,
-					     ref->level, 0, 1, node->no_quota,
+		ret = __btrfs_inc_extent_ref(trans, root, node,
+					     parent, ref_root,
+					     ref->level, 0, 1,
 					     extent_op);
 	} else if (node->action == BTRFS_DROP_DELAYED_REF) {
-		ret = __btrfs_free_extent(trans, root, node->bytenr,
-					  node->num_bytes, parent, ref_root,
-					  ref->level, 0, 1, extent_op,
-					  node->no_quota);
+		ret = __btrfs_free_extent(trans, root, node,
+					  parent, ref_root,
+					  ref->level, 0, 1, extent_op);
 	} else {
 		BUG();
 	}
@@ -2307,6 +2402,11 @@ static int run_one_delayed_ref(struct btrfs_trans_handle *trans,
 						      node->num_bytes);
 			}
 		}
+
+		/* Also free its reserved qgroup space */
+		btrfs_qgroup_free_delayed_ref(root->fs_info,
+					      head->qgroup_ref_root,
+					      head->qgroup_reserved);
 		return ret;
 	}
 
@@ -2323,28 +2423,27 @@ static int run_one_delayed_ref(struct btrfs_trans_handle *trans,
 	return ret;
 }
 
-static noinline struct btrfs_delayed_ref_node *
+static inline struct btrfs_delayed_ref_node *
 select_delayed_ref(struct btrfs_delayed_ref_head *head)
 {
-	struct rb_node *node;
-	struct btrfs_delayed_ref_node *ref, *last = NULL;;
+	struct btrfs_delayed_ref_node *ref;
+
+	if (list_empty(&head->ref_list))
+		return NULL;
 
 	/*
-	 * select delayed ref of type BTRFS_ADD_DELAYED_REF first.
-	 * this prevents ref count from going down to zero when
-	 * there still are pending delayed ref.
+	 * Select a delayed ref of type BTRFS_ADD_DELAYED_REF first.
+	 * This is to prevent a ref count from going down to zero, which deletes
+	 * the extent item from the extent tree, when there still are references
+	 * to add, which would fail because they would not find the extent item.
 	 */
-	node = rb_first(&head->ref_root);
-	while (node) {
-		ref = rb_entry(node, struct btrfs_delayed_ref_node,
-				rb_node);
+	list_for_each_entry(ref, &head->ref_list, list) {
 		if (ref->action == BTRFS_ADD_DELAYED_REF)
 			return ref;
-		else if (last == NULL)
-			last = ref;
-		node = rb_next(node);
 	}
-	return last;
+
+	return list_entry(head->ref_list.next, struct btrfs_delayed_ref_node,
+			  list);
 }
 
 /*
@@ -2402,6 +2501,11 @@ static noinline int __btrfs_run_delayed_refs(struct btrfs_trans_handle *trans,
 		 * and then it being added back again before the drop can
 		 * finish.  If we merged anything we need to re-loop so we can
 		 * get a good ref.
+		 * Or we can get node references of the same type that weren't
+		 * merged when created due to bumps in the tree mod seq, and
+		 * we need to merge them to prevent adding an inline extent
+		 * backref before dropping it (triggering a BUG_ON at
+		 * insert_inline_extent_backref()).
 		 */
 		spin_lock(&locked_ref->lock);
 		btrfs_merge_delayed_refs(trans, fs_info, delayed_refs,
@@ -2482,7 +2586,7 @@ static noinline int __btrfs_run_delayed_refs(struct btrfs_trans_handle *trans,
 			spin_unlock(&locked_ref->lock);
 			spin_lock(&delayed_refs->lock);
 			spin_lock(&locked_ref->lock);
-			if (rb_first(&locked_ref->ref_root) ||
+			if (!list_empty(&locked_ref->ref_list) ||
 			    locked_ref->extent_op) {
 				spin_unlock(&locked_ref->lock);
 				spin_unlock(&delayed_refs->lock);
@@ -2496,7 +2600,7 @@ static noinline int __btrfs_run_delayed_refs(struct btrfs_trans_handle *trans,
 		} else {
 			actual_count++;
 			ref->in_tree = 0;
-			rb_erase(&ref->rb_node, &locked_ref->ref_root);
+			list_del(&ref->list);
 		}
 		atomic_dec(&delayed_refs->num_entries);
 
@@ -2800,6 +2904,7 @@ int btrfs_run_delayed_refs(struct btrfs_trans_handle *trans,
 	struct btrfs_delayed_ref_head *head;
 	int ret;
 	int run_all = count == (unsigned long)-1;
+	bool can_flush_pending_bgs = trans->can_flush_pending_bgs;
 
 	/* We'll clean this up in btrfs_cleanup_transaction */
 	if (trans->aborted)
@@ -2816,6 +2921,7 @@ again:
 #ifdef SCRAMBLE_DELAYED_REFS
 	delayed_refs->run_delayed_start = find_middle(&delayed_refs->root);
 #endif
+	trans->can_flush_pending_bgs = false;
 	ret = __btrfs_run_delayed_refs(trans, root, count);
 	if (ret < 0) {
 		btrfs_abort_transaction(trans, root, ret);
@@ -2864,10 +2970,8 @@ again:
 		goto again;
 	}
 out:
-	ret = btrfs_delayed_qgroup_accounting(trans, root->fs_info);
-	if (ret)
-		return ret;
 	assert_qgroups_uptodate(trans);
+	trans->can_flush_pending_bgs = can_flush_pending_bgs;
 	return 0;
 }
 
@@ -2905,7 +3009,6 @@ static noinline int check_delayed_ref(struct btrfs_trans_handle *trans,
 	struct btrfs_delayed_ref_node *ref;
 	struct btrfs_delayed_data_ref *data_ref;
 	struct btrfs_delayed_ref_root *delayed_refs;
-	struct rb_node *node;
 	int ret = 0;
 
 	delayed_refs = &trans->transaction->delayed_refs;
@@ -2934,11 +3037,7 @@ static noinline int check_delayed_ref(struct btrfs_trans_handle *trans,
 	spin_unlock(&delayed_refs->lock);
 
 	spin_lock(&head->lock);
-	node = rb_first(&head->ref_root);
-	while (node) {
-		ref = rb_entry(node, struct btrfs_delayed_ref_node, rb_node);
-		node = rb_next(node);
-
+	list_for_each_entry(ref, &head->ref_list, list) {
 		/* If it's a shared ref we know a cross reference exists */
 		if (ref->type != BTRFS_EXTENT_DATA_REF_KEY) {
 			ret = 1;
@@ -3086,7 +3185,7 @@ static int __btrfs_mod_ref(struct btrfs_trans_handle *trans,
 	int level;
 	int ret = 0;
 	int (*process_func)(struct btrfs_trans_handle *, struct btrfs_root *,
-			    u64, u64, u64, u64, u64, u64, int);
+			    u64, u64, u64, u64, u64, u64);
 
 
 	if (btrfs_test_is_dummy_root(root))
@@ -3127,15 +3226,14 @@ static int __btrfs_mod_ref(struct btrfs_trans_handle *trans,
 			key.offset -= btrfs_file_extent_offset(buf, fi);
 			ret = process_func(trans, root, bytenr, num_bytes,
 					   parent, ref_root, key.objectid,
-					   key.offset, 1);
+					   key.offset);
 			if (ret)
 				goto fail;
 		} else {
 			bytenr = btrfs_node_blockptr(buf, i);
 			num_bytes = root->nodesize;
 			ret = process_func(trans, root, bytenr, num_bytes,
-					   parent, ref_root, level - 1, 0,
-					   1);
+					   parent, ref_root, level - 1, 0);
 			if (ret)
 				goto fail;
 		}
@@ -3316,6 +3414,15 @@ again:
 	spin_unlock(&block_group->lock);
 
 	/*
+	 * We hit an ENOSPC when setting up the cache in this transaction, just
+	 * skip doing the setup, we've already cleared the cache so we're safe.
+	 */
+	if (test_bit(BTRFS_TRANS_CACHE_ENOSPC, &trans->transaction->flags)) {
+		ret = -ENOSPC;
+		goto out_put;
+	}
+
+	/*
 	 * Try to preallocate enough space based on how big the block group is.
 	 * Keep in mind this has to include any pinned space which could end up
 	 * taking up quite a bit since it's not folded into the other space
@@ -3328,16 +3435,26 @@ again:
 	num_pages *= 16;
 	num_pages *= PAGE_CACHE_SIZE;
 
-	ret = btrfs_check_data_free_space(inode, num_pages, num_pages);
+	ret = btrfs_check_data_free_space(inode, 0, 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);
+	/*
+	 * Our cache requires contiguous chunks so that we don't modify a bunch
+	 * of metadata or split extents when writing the cache out, which means
+	 * we can enospc if we are heavily fragmented in addition to just normal
+	 * out of space conditions.  So if we hit this just skip setting up any
+	 * other block groups for this transaction, maybe we'll unpin enough
+	 * space the next time around.
+	 */
 	if (!ret)
 		dcs = BTRFS_DC_SETUP;
-	btrfs_free_reserved_data_space(inode, num_pages);
+	else if (ret == -ENOSPC)
+		set_bit(BTRFS_TRANS_CACHE_ENOSPC, &trans->transaction->flags);
+	btrfs_free_reserved_data_space(inode, 0, num_pages);
 
 out_put:
 	iput(inode);
@@ -3693,7 +3810,8 @@ static int update_space_info(struct btrfs_fs_info *info, u64 flags,
 		found->disk_total += total_bytes * factor;
 		found->bytes_used += bytes_used;
 		found->disk_used += bytes_used * factor;
-		found->full = 0;
+		if (total_bytes > 0)
+			found->full = 0;
 		spin_unlock(&found->lock);
 		*space_info = found;
 		return 0;
@@ -3722,6 +3840,7 @@ static int update_space_info(struct btrfs_fs_info *info, u64 flags,
 	found->bytes_readonly = 0;
 	found->bytes_may_use = 0;
 	found->full = 0;
+	found->max_extent_size = 0;
 	found->force_alloc = CHUNK_ALLOC_NO_FORCE;
 	found->chunk_alloc = 0;
 	found->flush = 0;
@@ -3798,7 +3917,8 @@ static u64 btrfs_reduce_alloc_profile(struct btrfs_root *root, u64 flags)
 {
 	u64 num_devices = root->fs_info->fs_devices->rw_devices;
 	u64 target;
-	u64 tmp;
+	u64 raid_type;
+	u64 allowed = 0;
 
 	/*
 	 * see if restripe for this chunk_type is in progress, if so
@@ -3816,31 +3936,26 @@ static u64 btrfs_reduce_alloc_profile(struct btrfs_root *root, u64 flags)
 	spin_unlock(&root->fs_info->balance_lock);
 
 	/* First, mask out the RAID levels which aren't possible */
-	if (num_devices == 1)
-		flags &= ~(BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID0 |
-			   BTRFS_BLOCK_GROUP_RAID5);
-	if (num_devices < 3)
-		flags &= ~BTRFS_BLOCK_GROUP_RAID6;
-	if (num_devices < 4)
-		flags &= ~BTRFS_BLOCK_GROUP_RAID10;
-
-	tmp = flags & (BTRFS_BLOCK_GROUP_DUP | BTRFS_BLOCK_GROUP_RAID0 |
-		       BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID5 |
-		       BTRFS_BLOCK_GROUP_RAID6 | BTRFS_BLOCK_GROUP_RAID10);
-	flags &= ~tmp;
-
-	if (tmp & BTRFS_BLOCK_GROUP_RAID6)
-		tmp = BTRFS_BLOCK_GROUP_RAID6;
-	else if (tmp & BTRFS_BLOCK_GROUP_RAID5)
-		tmp = BTRFS_BLOCK_GROUP_RAID5;
-	else if (tmp & BTRFS_BLOCK_GROUP_RAID10)
-		tmp = BTRFS_BLOCK_GROUP_RAID10;
-	else if (tmp & BTRFS_BLOCK_GROUP_RAID1)
-		tmp = BTRFS_BLOCK_GROUP_RAID1;
-	else if (tmp & BTRFS_BLOCK_GROUP_RAID0)
-		tmp = BTRFS_BLOCK_GROUP_RAID0;
-
-	return extended_to_chunk(flags | tmp);
+	for (raid_type = 0; raid_type < BTRFS_NR_RAID_TYPES; raid_type++) {
+		if (num_devices >= btrfs_raid_array[raid_type].devs_min)
+			allowed |= btrfs_raid_group[raid_type];
+	}
+	allowed &= flags;
+
+	if (allowed & BTRFS_BLOCK_GROUP_RAID6)
+		allowed = BTRFS_BLOCK_GROUP_RAID6;
+	else if (allowed & BTRFS_BLOCK_GROUP_RAID5)
+		allowed = BTRFS_BLOCK_GROUP_RAID5;
+	else if (allowed & BTRFS_BLOCK_GROUP_RAID10)
+		allowed = BTRFS_BLOCK_GROUP_RAID10;
+	else if (allowed & BTRFS_BLOCK_GROUP_RAID1)
+		allowed = BTRFS_BLOCK_GROUP_RAID1;
+	else if (allowed & BTRFS_BLOCK_GROUP_RAID0)
+		allowed = BTRFS_BLOCK_GROUP_RAID0;
+
+	flags &= ~BTRFS_BLOCK_GROUP_PROFILE_MASK;
+
+	return extended_to_chunk(flags | allowed);
 }
 
 static u64 get_alloc_profile(struct btrfs_root *root, u64 orig_flags)
@@ -3879,11 +3994,7 @@ u64 btrfs_get_alloc_profile(struct btrfs_root *root, int data)
 	return ret;
 }
 
-/*
- * This will check the space that the inode allocates from to make sure we have
- * enough space for bytes.
- */
-int btrfs_check_data_free_space(struct inode *inode, u64 bytes, u64 write_bytes)
+int btrfs_alloc_data_chunk_ondemand(struct inode *inode, u64 bytes)
 {
 	struct btrfs_space_info *data_sinfo;
 	struct btrfs_root *root = BTRFS_I(inode)->root;
@@ -3975,21 +4086,28 @@ commit_trans:
 		    !atomic_read(&root->fs_info->open_ioctl_trans)) {
 			need_commit--;
 
+			if (need_commit > 0) {
+				btrfs_start_delalloc_roots(fs_info, 0, -1);
+				btrfs_wait_ordered_roots(fs_info, -1);
+			}
+
 			trans = btrfs_join_transaction(root);
 			if (IS_ERR(trans))
 				return PTR_ERR(trans);
 			if (have_pinned_space >= 0 ||
-			    trans->transaction->have_free_bgs ||
+			    test_bit(BTRFS_TRANS_HAVE_FREE_BGS,
+				     &trans->transaction->flags) ||
 			    need_commit > 0) {
 				ret = btrfs_commit_transaction(trans, root);
 				if (ret)
 					return ret;
 				/*
-				 * make sure that all running delayed iput are
-				 * done
+				 * The cleaner kthread might still be doing iput
+				 * operations. Wait for it to finish so that
+				 * more space is released.
 				 */
-				down_write(&root->fs_info->delayed_iput_sem);
-				up_write(&root->fs_info->delayed_iput_sem);
+				mutex_lock(&root->fs_info->cleaner_delayed_iput_mutex);
+				mutex_unlock(&root->fs_info->cleaner_delayed_iput_mutex);
 				goto again;
 			} else {
 				btrfs_end_transaction(trans, root);
@@ -4001,38 +4119,86 @@ commit_trans:
 					      data_sinfo->flags, bytes, 1);
 		return -ENOSPC;
 	}
-	ret = btrfs_qgroup_reserve(root, write_bytes);
-	if (ret)
-		goto out;
 	data_sinfo->bytes_may_use += bytes;
 	trace_btrfs_space_reservation(root->fs_info, "space_info",
 				      data_sinfo->flags, bytes, 1);
-out:
 	spin_unlock(&data_sinfo->lock);
 
 	return ret;
 }
 
 /*
- * Called if we need to clear a data reservation for this inode.
+ * New check_data_free_space() with ability for precious data reservation
+ * Will replace old btrfs_check_data_free_space(), but for patch split,
+ * add a new function first and then replace it.
  */
-void btrfs_free_reserved_data_space(struct inode *inode, u64 bytes)
+int btrfs_check_data_free_space(struct inode *inode, u64 start, u64 len)
+{
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	int ret;
+
+	/* align the range */
+	len = round_up(start + len, root->sectorsize) -
+	      round_down(start, root->sectorsize);
+	start = round_down(start, root->sectorsize);
+
+	ret = btrfs_alloc_data_chunk_ondemand(inode, len);
+	if (ret < 0)
+		return ret;
+
+	/*
+	 * Use new btrfs_qgroup_reserve_data to reserve precious data space
+	 *
+	 * TODO: Find a good method to avoid reserve data space for NOCOW
+	 * range, but don't impact performance on quota disable case.
+	 */
+	ret = btrfs_qgroup_reserve_data(inode, start, len);
+	return ret;
+}
+
+/*
+ * Called if we need to clear a data reservation for this inode
+ * Normally in a error case.
+ *
+ * This one will *NOT* use accurate qgroup reserved space API, just for case
+ * which we can't sleep and is sure it won't affect qgroup reserved space.
+ * Like clear_bit_hook().
+ */
+void btrfs_free_reserved_data_space_noquota(struct inode *inode, u64 start,
+					    u64 len)
 {
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	struct btrfs_space_info *data_sinfo;
 
-	/* make sure bytes are sectorsize aligned */
-	bytes = ALIGN(bytes, root->sectorsize);
+	/* Make sure the range is aligned to sectorsize */
+	len = round_up(start + len, root->sectorsize) -
+	      round_down(start, root->sectorsize);
+	start = round_down(start, root->sectorsize);
 
 	data_sinfo = root->fs_info->data_sinfo;
 	spin_lock(&data_sinfo->lock);
-	WARN_ON(data_sinfo->bytes_may_use < bytes);
-	data_sinfo->bytes_may_use -= bytes;
+	if (WARN_ON(data_sinfo->bytes_may_use < len))
+		data_sinfo->bytes_may_use = 0;
+	else
+		data_sinfo->bytes_may_use -= len;
 	trace_btrfs_space_reservation(root->fs_info, "space_info",
-				      data_sinfo->flags, bytes, 0);
+				      data_sinfo->flags, len, 0);
 	spin_unlock(&data_sinfo->lock);
 }
 
+/*
+ * Called if we need to clear a data reservation for this inode
+ * Normally in a error case.
+ *
+ * This one will handle the per-indoe data rsv map for accurate reserved
+ * space framework.
+ */
+void btrfs_free_reserved_data_space(struct inode *inode, u64 start, u64 len)
+{
+	btrfs_free_reserved_data_space_noquota(inode, start, len);
+	btrfs_qgroup_free_data(inode, start, len);
+}
+
 static void force_metadata_allocation(struct btrfs_fs_info *info)
 {
 	struct list_head *head = &info->space_info;
@@ -4088,7 +4254,7 @@ static int should_alloc_chunk(struct btrfs_root *root,
 	return 1;
 }
 
-static u64 get_system_chunk_thresh(struct btrfs_root *root, u64 type)
+static u64 get_profile_num_devs(struct btrfs_root *root, u64 type)
 {
 	u64 num_dev;
 
@@ -4102,24 +4268,43 @@ static u64 get_system_chunk_thresh(struct btrfs_root *root, u64 type)
 	else
 		num_dev = 1;	/* DUP or single */
 
-	/* metadata for updaing devices and chunk tree */
-	return btrfs_calc_trans_metadata_size(root, num_dev + 1);
+	return num_dev;
 }
 
-static void check_system_chunk(struct btrfs_trans_handle *trans,
-			       struct btrfs_root *root, u64 type)
+/*
+ * If @is_allocation is true, reserve space in the system space info necessary
+ * for allocating a chunk, otherwise if it's false, reserve space necessary for
+ * removing a chunk.
+ */
+void check_system_chunk(struct btrfs_trans_handle *trans,
+			struct btrfs_root *root,
+			u64 type)
 {
 	struct btrfs_space_info *info;
 	u64 left;
 	u64 thresh;
+	int ret = 0;
+	u64 num_devs;
+
+	/*
+	 * Needed because we can end up allocating a system chunk and for an
+	 * atomic and race free space reservation in the chunk block reserve.
+	 */
+	ASSERT(mutex_is_locked(&root->fs_info->chunk_mutex));
 
 	info = __find_space_info(root->fs_info, BTRFS_BLOCK_GROUP_SYSTEM);
 	spin_lock(&info->lock);
 	left = info->total_bytes - info->bytes_used - info->bytes_pinned -
-		info->bytes_reserved - info->bytes_readonly;
+		info->bytes_reserved - info->bytes_readonly -
+		info->bytes_may_use;
 	spin_unlock(&info->lock);
 
-	thresh = get_system_chunk_thresh(root, type);
+	num_devs = get_profile_num_devs(root, type);
+
+	/* num_devs device items to update and 1 chunk item to add or remove */
+	thresh = btrfs_calc_trunc_metadata_size(root, num_devs) +
+		btrfs_calc_trans_metadata_size(root, 1);
+
 	if (left < thresh && btrfs_test_opt(root, ENOSPC_DEBUG)) {
 		btrfs_info(root->fs_info, "left=%llu, need=%llu, flags=%llu",
 			left, thresh, type);
@@ -4130,7 +4315,21 @@ static void check_system_chunk(struct btrfs_trans_handle *trans,
 		u64 flags;
 
 		flags = btrfs_get_alloc_profile(root->fs_info->chunk_root, 0);
-		btrfs_alloc_chunk(trans, root, flags);
+		/*
+		 * Ignore failure to create system chunk. We might end up not
+		 * needing it, as we might not need to COW all nodes/leafs from
+		 * the paths we visit in the chunk tree (they were already COWed
+		 * or created in the current transaction for example).
+		 */
+		ret = btrfs_alloc_chunk(trans, root, flags);
+	}
+
+	if (!ret) {
+		ret = btrfs_block_rsv_add(root->fs_info->chunk_root,
+					  &root->fs_info->chunk_block_rsv,
+					  thresh, BTRFS_RESERVE_NO_FLUSH);
+		if (!ret)
+			trans->chunk_bytes_reserved += thresh;
 	}
 }
 
@@ -4235,6 +4434,25 @@ out:
 	space_info->chunk_alloc = 0;
 	spin_unlock(&space_info->lock);
 	mutex_unlock(&fs_info->chunk_mutex);
+	/*
+	 * When we allocate a new chunk we reserve space in the chunk block
+	 * reserve to make sure we can COW nodes/leafs in the chunk tree or
+	 * add new nodes/leafs to it if we end up needing to do it when
+	 * inserting the chunk item and updating device items as part of the
+	 * second phase of chunk allocation, performed by
+	 * btrfs_finish_chunk_alloc(). So make sure we don't accumulate a
+	 * large number of new block groups to create in our transaction
+	 * handle's new_bgs list to avoid exhausting the chunk block reserve
+	 * in extreme cases - like having a single transaction create many new
+	 * block groups when starting to write out the free space caches of all
+	 * the block groups that were made dirty during the lifetime of the
+	 * transaction.
+	 */
+	if (trans->can_flush_pending_bgs &&
+	    trans->chunk_bytes_reserved >= (2 * 1024 * 1024ull)) {
+		btrfs_create_pending_block_groups(trans, trans->root);
+		btrfs_trans_release_chunk_metadata(trans);
+	}
 	return ret;
 }
 
@@ -4812,13 +5030,9 @@ static struct btrfs_block_rsv *get_block_rsv(
 {
 	struct btrfs_block_rsv *block_rsv = NULL;
 
-	if (test_bit(BTRFS_ROOT_REF_COWS, &root->state))
-		block_rsv = trans->block_rsv;
-
-	if (root == root->fs_info->csum_root && trans->adding_csums)
-		block_rsv = trans->block_rsv;
-
-	if (root == root->fs_info->uuid_root)
+	if (test_bit(BTRFS_ROOT_REF_COWS, &root->state) ||
+	    (root == root->fs_info->csum_root && trans->adding_csums) ||
+	     (root == root->fs_info->uuid_root))
 		block_rsv = trans->block_rsv;
 
 	if (!block_rsv)
@@ -5188,6 +5402,24 @@ void btrfs_trans_release_metadata(struct btrfs_trans_handle *trans,
 	trans->bytes_reserved = 0;
 }
 
+/*
+ * To be called after all the new block groups attached to the transaction
+ * handle have been created (btrfs_create_pending_block_groups()).
+ */
+void btrfs_trans_release_chunk_metadata(struct btrfs_trans_handle *trans)
+{
+	struct btrfs_fs_info *fs_info = trans->root->fs_info;
+
+	if (!trans->chunk_bytes_reserved)
+		return;
+
+	WARN_ON_ONCE(!list_empty(&trans->new_bgs));
+
+	block_rsv_release_bytes(fs_info, &fs_info->chunk_block_rsv, NULL,
+				trans->chunk_bytes_reserved);
+	trans->chunk_bytes_reserved = 0;
+}
+
 /* Can only return 0 or -ENOSPC */
 int btrfs_orphan_reserve_metadata(struct btrfs_trans_handle *trans,
 				  struct inode *inode)
@@ -5243,7 +5475,7 @@ int btrfs_subvolume_reserve_metadata(struct btrfs_root *root,
 	if (root->fs_info->quota_enabled) {
 		/* One for parent inode, two for dir entries */
 		num_bytes = 3 * root->nodesize;
-		ret = btrfs_qgroup_reserve(root, num_bytes);
+		ret = btrfs_qgroup_reserve_meta(root, num_bytes);
 		if (ret)
 			return ret;
 	} else {
@@ -5261,10 +5493,8 @@ int btrfs_subvolume_reserve_metadata(struct btrfs_root *root,
 	if (ret == -ENOSPC && use_global_rsv)
 		ret = btrfs_block_rsv_migrate(global_rsv, rsv, num_bytes);
 
-	if (ret) {
-		if (*qgroup_reserved)
-			btrfs_qgroup_free(root, *qgroup_reserved);
-	}
+	if (ret && *qgroup_reserved)
+		btrfs_qgroup_free_meta(root, *qgroup_reserved);
 
 	return ret;
 }
@@ -5425,15 +5655,15 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes)
 	spin_unlock(&BTRFS_I(inode)->lock);
 
 	if (root->fs_info->quota_enabled) {
-		ret = btrfs_qgroup_reserve(root, nr_extents * root->nodesize);
+		ret = btrfs_qgroup_reserve_meta(root,
+				nr_extents * root->nodesize);
 		if (ret)
 			goto out_fail;
 	}
 
 	ret = reserve_metadata_bytes(root, block_rsv, to_reserve, flush);
 	if (unlikely(ret)) {
-		if (root->fs_info->quota_enabled)
-			btrfs_qgroup_free(root, nr_extents * root->nodesize);
+		btrfs_qgroup_free_meta(root, nr_extents * root->nodesize);
 		goto out_fail;
 	}
 
@@ -5556,41 +5786,48 @@ void btrfs_delalloc_release_metadata(struct inode *inode, u64 num_bytes)
 }
 
 /**
- * btrfs_delalloc_reserve_space - reserve data and metadata space for delalloc
+ * btrfs_delalloc_reserve_space - reserve data and metadata space for
+ * delalloc
  * @inode: inode we're writing to
- * @num_bytes: the number of bytes we want to allocate
+ * @start: start range we are writing to
+ * @len: how long the range we are writing to
+ *
+ * TODO: This function will finally replace old btrfs_delalloc_reserve_space()
  *
  * This will do the following things
  *
- * o reserve space in the data space info for num_bytes
- * o reserve space in the metadata space info based on number of outstanding
+ * o reserve space in data space info for num bytes
+ *   and reserve precious corresponding qgroup space
+ *   (Done in check_data_free_space)
+ *
+ * o reserve space for metadata space, based on the number of outstanding
  *   extents and how much csums will be needed
- * o add to the inodes ->delalloc_bytes
+ *   also reserve metadata space in a per root over-reserve method.
+ * o add to the inodes->delalloc_bytes
  * o add it to the fs_info's delalloc inodes list.
+ *   (Above 3 all done in delalloc_reserve_metadata)
  *
- * This will return 0 for success and -ENOSPC if there is no space left.
+ * Return 0 for success
+ * Return <0 for error(-ENOSPC or -EQUOT)
  */
-int btrfs_delalloc_reserve_space(struct inode *inode, u64 num_bytes)
+int btrfs_delalloc_reserve_space(struct inode *inode, u64 start, u64 len)
 {
 	int ret;
 
-	ret = btrfs_check_data_free_space(inode, num_bytes, num_bytes);
-	if (ret)
-		return ret;
-
-	ret = btrfs_delalloc_reserve_metadata(inode, num_bytes);
-	if (ret) {
-		btrfs_free_reserved_data_space(inode, num_bytes);
+	ret = btrfs_check_data_free_space(inode, start, len);
+	if (ret < 0)
 		return ret;
-	}
-
-	return 0;
+	ret = btrfs_delalloc_reserve_metadata(inode, len);
+	if (ret < 0)
+		btrfs_free_reserved_data_space(inode, start, len);
+	return ret;
 }
 
 /**
  * btrfs_delalloc_release_space - release data and metadata space for delalloc
  * @inode: inode we're releasing space for
- * @num_bytes: the number of bytes we want to free up
+ * @start: start position of the space already reserved
+ * @len: the len of the space already reserved
  *
  * This must be matched with a call to btrfs_delalloc_reserve_space.  This is
  * called in the case that we don't need the metadata AND data reservations
@@ -5599,11 +5836,12 @@ int btrfs_delalloc_reserve_space(struct inode *inode, u64 num_bytes)
  * This function will release the metadata space that was not used and will
  * decrement ->delalloc_bytes and remove it from the fs_info delalloc_inodes
  * list if there are no delalloc bytes left.
+ * Also it will handle the qgroup reserved space.
  */
-void btrfs_delalloc_release_space(struct inode *inode, u64 num_bytes)
+void btrfs_delalloc_release_space(struct inode *inode, u64 start, u64 len)
 {
-	btrfs_delalloc_release_metadata(inode, num_bytes);
-	btrfs_free_reserved_data_space(inode, num_bytes);
+	btrfs_delalloc_release_metadata(inode, len);
+	btrfs_free_reserved_data_space(inode, start, len);
 }
 
 static int update_block_group(struct btrfs_trans_handle *trans,
@@ -5680,19 +5918,6 @@ static int update_block_group(struct btrfs_trans_handle *trans,
 			set_extent_dirty(info->pinned_extents,
 					 bytenr, bytenr + num_bytes - 1,
 					 GFP_NOFS | __GFP_NOFAIL);
-			/*
-			 * No longer have used bytes in this block group, queue
-			 * it for deletion.
-			 */
-			if (old_val == 0) {
-				spin_lock(&info->unused_bgs_lock);
-				if (list_empty(&cache->bg_list)) {
-					btrfs_get_block_group(cache);
-					list_add_tail(&cache->bg_list,
-						      &info->unused_bgs);
-				}
-				spin_unlock(&info->unused_bgs_lock);
-			}
 		}
 
 		spin_lock(&trans->transaction->dirty_bgs_lock);
@@ -5704,6 +5929,22 @@ static int update_block_group(struct btrfs_trans_handle *trans,
 		}
 		spin_unlock(&trans->transaction->dirty_bgs_lock);
 
+		/*
+		 * No longer have used bytes in this block group, queue it for
+		 * deletion. We do this after adding the block group to the
+		 * dirty list to avoid races between cleaner kthread and space
+		 * cache writeout.
+		 */
+		if (!alloc && old_val == 0) {
+			spin_lock(&info->unused_bgs_lock);
+			if (list_empty(&cache->bg_list)) {
+				btrfs_get_block_group(cache);
+				list_add_tail(&cache->bg_list,
+					      &info->unused_bgs);
+			}
+			spin_unlock(&info->unused_bgs_lock);
+		}
+
 		btrfs_put_block_group(cache);
 		total -= num_bytes;
 		bytenr += num_bytes;
@@ -5968,6 +6209,34 @@ void btrfs_prepare_extent_commit(struct btrfs_trans_handle *trans,
 	update_global_block_rsv(fs_info);
 }
 
+/*
+ * Returns the free cluster for the given space info and sets empty_cluster to
+ * what it should be based on the mount options.
+ */
+static struct btrfs_free_cluster *
+fetch_cluster_info(struct btrfs_root *root, struct btrfs_space_info *space_info,
+		   u64 *empty_cluster)
+{
+	struct btrfs_free_cluster *ret = NULL;
+	bool ssd = btrfs_test_opt(root, SSD);
+
+	*empty_cluster = 0;
+	if (btrfs_mixed_space_info(space_info))
+		return ret;
+
+	if (ssd)
+		*empty_cluster = 2 * 1024 * 1024;
+	if (space_info->flags & BTRFS_BLOCK_GROUP_METADATA) {
+		ret = &root->fs_info->meta_alloc_cluster;
+		if (!ssd)
+			*empty_cluster = 64 * 1024;
+	} else if ((space_info->flags & BTRFS_BLOCK_GROUP_DATA) && ssd) {
+		ret = &root->fs_info->data_alloc_cluster;
+	}
+
+	return ret;
+}
+
 static int unpin_extent_range(struct btrfs_root *root, u64 start, u64 end,
 			      const bool return_free_space)
 {
@@ -5975,7 +6244,10 @@ static int unpin_extent_range(struct btrfs_root *root, u64 start, u64 end,
 	struct btrfs_block_group_cache *cache = NULL;
 	struct btrfs_space_info *space_info;
 	struct btrfs_block_rsv *global_rsv = &fs_info->global_block_rsv;
+	struct btrfs_free_cluster *cluster = NULL;
 	u64 len;
+	u64 total_unpinned = 0;
+	u64 empty_cluster = 0;
 	bool readonly;
 
 	while (start <= end) {
@@ -5984,8 +6256,14 @@ static int unpin_extent_range(struct btrfs_root *root, u64 start, u64 end,
 		    start >= cache->key.objectid + cache->key.offset) {
 			if (cache)
 				btrfs_put_block_group(cache);
+			total_unpinned = 0;
 			cache = btrfs_lookup_block_group(fs_info, start);
 			BUG_ON(!cache); /* Logic error */
+
+			cluster = fetch_cluster_info(root,
+						     cache->space_info,
+						     &empty_cluster);
+			empty_cluster <<= 1;
 		}
 
 		len = cache->key.objectid + cache->key.offset - start;
@@ -5998,12 +6276,27 @@ static int unpin_extent_range(struct btrfs_root *root, u64 start, u64 end,
 		}
 
 		start += len;
+		total_unpinned += len;
 		space_info = cache->space_info;
 
+		/*
+		 * If this space cluster has been marked as fragmented and we've
+		 * unpinned enough in this block group to potentially allow a
+		 * cluster to be created inside of it go ahead and clear the
+		 * fragmented check.
+		 */
+		if (cluster && cluster->fragmented &&
+		    total_unpinned > empty_cluster) {
+			spin_lock(&cluster->lock);
+			cluster->fragmented = 0;
+			spin_unlock(&cluster->lock);
+		}
+
 		spin_lock(&space_info->lock);
 		spin_lock(&cache->lock);
 		cache->pinned -= len;
 		space_info->bytes_pinned -= len;
+		space_info->max_extent_size = 0;
 		percpu_counter_add(&space_info->total_bytes_pinned, -len);
 		if (cache->ro) {
 			space_info->bytes_readonly += len;
@@ -6034,20 +6327,19 @@ int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans,
 			       struct btrfs_root *root)
 {
 	struct btrfs_fs_info *fs_info = root->fs_info;
+	struct btrfs_block_group_cache *block_group, *tmp;
+	struct list_head *deleted_bgs;
 	struct extent_io_tree *unpin;
 	u64 start;
 	u64 end;
 	int ret;
 
-	if (trans->aborted)
-		return 0;
-
 	if (fs_info->pinned_extents == &fs_info->freed_extents[0])
 		unpin = &fs_info->freed_extents[1];
 	else
 		unpin = &fs_info->freed_extents[0];
 
-	while (1) {
+	while (!trans->aborted) {
 		mutex_lock(&fs_info->unused_bg_unpin_mutex);
 		ret = find_first_extent_bit(unpin, 0, &start, &end,
 					    EXTENT_DIRTY, NULL);
@@ -6066,6 +6358,34 @@ int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans,
 		cond_resched();
 	}
 
+	/*
+	 * Transaction is finished.  We don't need the lock anymore.  We
+	 * do need to clean up the block groups in case of a transaction
+	 * abort.
+	 */
+	deleted_bgs = &trans->transaction->deleted_bgs;
+	list_for_each_entry_safe(block_group, tmp, deleted_bgs, bg_list) {
+		u64 trimmed = 0;
+
+		ret = -EROFS;
+		if (!trans->aborted)
+			ret = btrfs_discard_extent(root,
+						   block_group->key.objectid,
+						   block_group->key.offset,
+						   &trimmed);
+
+		list_del_init(&block_group->bg_list);
+		btrfs_put_block_group_trimming(block_group);
+		btrfs_put_block_group(block_group);
+
+		if (ret) {
+			const char *errstr = btrfs_decode_error(ret);
+			btrfs_warn(fs_info,
+				   "Discard failed while removing blockgroup: errno=%d %s\n",
+				   ret, errstr);
+		}
+	}
+
 	return 0;
 }
 
@@ -6092,11 +6412,10 @@ static void add_pinned_bytes(struct btrfs_fs_info *fs_info, u64 num_bytes,
 
 static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
 				struct btrfs_root *root,
-				u64 bytenr, u64 num_bytes, u64 parent,
+				struct btrfs_delayed_ref_node *node, u64 parent,
 				u64 root_objectid, u64 owner_objectid,
 				u64 owner_offset, int refs_to_drop,
-				struct btrfs_delayed_extent_op *extent_op,
-				int no_quota)
+				struct btrfs_delayed_extent_op *extent_op)
 {
 	struct btrfs_key key;
 	struct btrfs_path *path;
@@ -6112,14 +6431,12 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
 	int num_to_del = 1;
 	u32 item_size;
 	u64 refs;
+	u64 bytenr = node->bytenr;
+	u64 num_bytes = node->num_bytes;
 	int last_ref = 0;
-	enum btrfs_qgroup_operation_type type = BTRFS_QGROUP_OPER_SUB_EXCL;
 	bool skinny_metadata = btrfs_fs_incompat(root->fs_info,
 						 SKINNY_METADATA);
 
-	if (!info->quota_enabled || !is_fstree(root_objectid))
-		no_quota = 1;
-
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
@@ -6294,7 +6611,6 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
 	refs -= refs_to_drop;
 
 	if (refs > 0) {
-		type = BTRFS_QGROUP_OPER_SUB_SHARED;
 		if (extent_op)
 			__run_delayed_extent_op(extent_op, leaf, ei);
 		/*
@@ -6321,7 +6637,7 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
 	} else {
 		if (found_extent) {
 			BUG_ON(is_data && refs_to_drop !=
-			       extent_data_ref_count(root, path, iref));
+			       extent_data_ref_count(path, iref));
 			if (iref) {
 				BUG_ON(path->slots[0] != extent_slot);
 			} else {
@@ -6356,18 +6672,6 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
 	}
 	btrfs_release_path(path);
 
-	/* Deal with the quota accounting */
-	if (!ret && last_ref && !no_quota) {
-		int mod_seq = 0;
-
-		if (owner_objectid >= BTRFS_FIRST_FREE_OBJECTID &&
-		    type == BTRFS_QGROUP_OPER_SUB_SHARED)
-			mod_seq = 1;
-
-		ret = btrfs_qgroup_record_ref(trans, info, root_objectid,
-					      bytenr, num_bytes, type,
-					      mod_seq);
-	}
 out:
 	btrfs_free_path(path);
 	return ret;
@@ -6393,7 +6697,7 @@ static noinline int check_ref_cleanup(struct btrfs_trans_handle *trans,
 		goto out_delayed_unlock;
 
 	spin_lock(&head->lock);
-	if (rb_first(&head->ref_root))
+	if (!list_empty(&head->ref_list))
 		goto out;
 
 	if (head->extent_op) {
@@ -6458,7 +6762,7 @@ void btrfs_free_tree_block(struct btrfs_trans_handle *trans,
 					buf->start, buf->len,
 					parent, root->root_key.objectid,
 					btrfs_header_level(buf),
-					BTRFS_DROP_DELAYED_REF, NULL, 0);
+					BTRFS_DROP_DELAYED_REF, NULL);
 		BUG_ON(ret); /* -ENOMEM */
 	}
 
@@ -6506,7 +6810,7 @@ out:
 /* Can return -ENOMEM */
 int btrfs_free_extent(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 		      u64 bytenr, u64 num_bytes, u64 parent, u64 root_objectid,
-		      u64 owner, u64 offset, int no_quota)
+		      u64 owner, u64 offset)
 {
 	int ret;
 	struct btrfs_fs_info *fs_info = root->fs_info;
@@ -6529,13 +6833,13 @@ int btrfs_free_extent(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 		ret = btrfs_add_delayed_tree_ref(fs_info, trans, bytenr,
 					num_bytes,
 					parent, root_objectid, (int)owner,
-					BTRFS_DROP_DELAYED_REF, NULL, no_quota);
+					BTRFS_DROP_DELAYED_REF, NULL);
 	} else {
 		ret = btrfs_add_delayed_data_ref(fs_info, trans, bytenr,
 						num_bytes,
 						parent, root_objectid, owner,
-						offset, BTRFS_DROP_DELAYED_REF,
-						NULL, no_quota);
+						offset, 0,
+						BTRFS_DROP_DELAYED_REF, NULL);
 	}
 	return ret;
 }
@@ -6721,7 +7025,7 @@ static noinline int find_free_extent(struct btrfs_root *orig_root,
 	struct btrfs_block_group_cache *block_group = NULL;
 	u64 search_start = 0;
 	u64 max_extent_size = 0;
-	int empty_cluster = 2 * 1024 * 1024;
+	u64 empty_cluster = 0;
 	struct btrfs_space_info *space_info;
 	int loop = 0;
 	int index = __get_raid_index(flags);
@@ -6731,6 +7035,8 @@ static noinline int find_free_extent(struct btrfs_root *orig_root,
 	bool failed_alloc = false;
 	bool use_cluster = true;
 	bool have_caching_bg = false;
+	bool orig_have_caching_bg = false;
+	bool full_search = false;
 
 	WARN_ON(num_bytes < root->sectorsize);
 	ins->type = BTRFS_EXTENT_ITEM_KEY;
@@ -6746,36 +7052,47 @@ static noinline int find_free_extent(struct btrfs_root *orig_root,
 	}
 
 	/*
-	 * If the space info is for both data and metadata it means we have a
-	 * small filesystem and we can't use the clustering stuff.
+	 * If our free space is heavily fragmented we may not be able to make
+	 * big contiguous allocations, so instead of doing the expensive search
+	 * for free space, simply return ENOSPC with our max_extent_size so we
+	 * can go ahead and search for a more manageable chunk.
+	 *
+	 * If our max_extent_size is large enough for our allocation simply
+	 * disable clustering since we will likely not be able to find enough
+	 * space to create a cluster and induce latency trying.
 	 */
-	if (btrfs_mixed_space_info(space_info))
-		use_cluster = false;
-
-	if (flags & BTRFS_BLOCK_GROUP_METADATA && use_cluster) {
-		last_ptr = &root->fs_info->meta_alloc_cluster;
-		if (!btrfs_test_opt(root, SSD))
-			empty_cluster = 64 * 1024;
-	}
-
-	if ((flags & BTRFS_BLOCK_GROUP_DATA) && use_cluster &&
-	    btrfs_test_opt(root, SSD)) {
-		last_ptr = &root->fs_info->data_alloc_cluster;
+	if (unlikely(space_info->max_extent_size)) {
+		spin_lock(&space_info->lock);
+		if (space_info->max_extent_size &&
+		    num_bytes > space_info->max_extent_size) {
+			ins->offset = space_info->max_extent_size;
+			spin_unlock(&space_info->lock);
+			return -ENOSPC;
+		} else if (space_info->max_extent_size) {
+			use_cluster = false;
+		}
+		spin_unlock(&space_info->lock);
 	}
 
+	last_ptr = fetch_cluster_info(orig_root, space_info, &empty_cluster);
 	if (last_ptr) {
 		spin_lock(&last_ptr->lock);
 		if (last_ptr->block_group)
 			hint_byte = last_ptr->window_start;
+		if (last_ptr->fragmented) {
+			/*
+			 * We still set window_start so we can keep track of the
+			 * last place we found an allocation to try and save
+			 * some time.
+			 */
+			hint_byte = last_ptr->window_start;
+			use_cluster = false;
+		}
 		spin_unlock(&last_ptr->lock);
 	}
 
 	search_start = max(search_start, first_logical_byte(root, 0));
 	search_start = max(search_start, hint_byte);
-
-	if (!last_ptr)
-		empty_cluster = 0;
-
 	if (search_start == hint_byte) {
 		block_group = btrfs_lookup_block_group(root->fs_info,
 						       search_start);
@@ -6810,6 +7127,8 @@ static noinline int find_free_extent(struct btrfs_root *orig_root,
 	}
 search:
 	have_caching_bg = false;
+	if (index == 0 || index == __get_raid_index(flags))
+		full_search = true;
 	down_read(&space_info->groups_sem);
 	list_for_each_entry(block_group, &space_info->block_groups[index],
 			    list) {
@@ -6843,6 +7162,7 @@ search:
 have_block_group:
 		cached = block_group_cache_done(block_group);
 		if (unlikely(!cached)) {
+			have_caching_bg = true;
 			ret = cache_block_group(block_group, 0);
 			BUG_ON(ret < 0);
 			ret = 0;
@@ -6857,7 +7177,7 @@ have_block_group:
 		 * Ok we want to try and use the cluster allocator, so
 		 * lets look there
 		 */
-		if (last_ptr) {
+		if (last_ptr && use_cluster) {
 			struct btrfs_block_group_cache *used_block_group;
 			unsigned long aligned_cluster;
 			/*
@@ -6983,6 +7303,16 @@ refill_cluster:
 		}
 
 unclustered_alloc:
+		/*
+		 * We are doing an unclustered alloc, set the fragmented flag so
+		 * we don't bother trying to setup a cluster again until we get
+		 * more space.
+		 */
+		if (unlikely(last_ptr)) {
+			spin_lock(&last_ptr->lock);
+			last_ptr->fragmented = 1;
+			spin_unlock(&last_ptr->lock);
+		}
 		spin_lock(&block_group->free_space_ctl->tree_lock);
 		if (cached &&
 		    block_group->free_space_ctl->free_space <
@@ -7015,8 +7345,6 @@ unclustered_alloc:
 			failed_alloc = true;
 			goto have_block_group;
 		} else if (!offset) {
-			if (!cached)
-				have_caching_bg = true;
 			goto loop;
 		}
 checks:
@@ -7057,6 +7385,10 @@ loop:
 	}
 	up_read(&space_info->groups_sem);
 
+	if ((loop == LOOP_CACHING_NOWAIT) && have_caching_bg
+		&& !orig_have_caching_bg)
+		orig_have_caching_bg = true;
+
 	if (!ins->objectid && loop >= LOOP_CACHING_WAIT && have_caching_bg)
 		goto search;
 
@@ -7073,7 +7405,20 @@ loop:
 	 */
 	if (!ins->objectid && loop < LOOP_NO_EMPTY_SIZE) {
 		index = 0;
-		loop++;
+		if (loop == LOOP_CACHING_NOWAIT) {
+			/*
+			 * We want to skip the LOOP_CACHING_WAIT step if we
+			 * don't have any unached bgs and we've alrelady done a
+			 * full search through.
+			 */
+			if (orig_have_caching_bg || !full_search)
+				loop = LOOP_CACHING_WAIT;
+			else
+				loop = LOOP_ALLOC_CHUNK;
+		} else {
+			loop++;
+		}
+
 		if (loop == LOOP_ALLOC_CHUNK) {
 			struct btrfs_trans_handle *trans;
 			int exist = 0;
@@ -7091,6 +7436,15 @@ loop:
 
 			ret = do_chunk_alloc(trans, root, flags,
 					     CHUNK_ALLOC_FORCE);
+
+			/*
+			 * If we can't allocate a new chunk we've already looped
+			 * through at least once, move on to the NO_EMPTY_SIZE
+			 * case.
+			 */
+			if (ret == -ENOSPC)
+				loop = LOOP_NO_EMPTY_SIZE;
+
 			/*
 			 * Do not bail out on ENOSPC since we
 			 * can do more things.
@@ -7107,6 +7461,15 @@ loop:
 		}
 
 		if (loop == LOOP_NO_EMPTY_SIZE) {
+			/*
+			 * Don't loop again if we already have no empty_size and
+			 * no empty_cluster.
+			 */
+			if (empty_size == 0 &&
+			    empty_cluster == 0) {
+				ret = -ENOSPC;
+				goto out;
+			}
 			empty_size = 0;
 			empty_cluster = 0;
 		}
@@ -7115,11 +7478,20 @@ loop:
 	} else if (!ins->objectid) {
 		ret = -ENOSPC;
 	} else if (ins->objectid) {
+		if (!use_cluster && last_ptr) {
+			spin_lock(&last_ptr->lock);
+			last_ptr->window_start = ins->objectid;
+			spin_unlock(&last_ptr->lock);
+		}
 		ret = 0;
 	}
 out:
-	if (ret == -ENOSPC)
+	if (ret == -ENOSPC) {
+		spin_lock(&space_info->lock);
+		space_info->max_extent_size = max_extent_size;
+		spin_unlock(&space_info->lock);
 		ins->offset = max_extent_size;
+	}
 	return ret;
 }
 
@@ -7168,7 +7540,7 @@ int btrfs_reserve_extent(struct btrfs_root *root,
 			 u64 empty_size, u64 hint_byte,
 			 struct btrfs_key *ins, int is_data, int delalloc)
 {
-	bool final_tried = false;
+	bool final_tried = num_bytes == min_alloc_size;
 	u64 flags;
 	int ret;
 
@@ -7303,13 +7675,6 @@ static int alloc_reserved_file_extent(struct btrfs_trans_handle *trans,
 	btrfs_mark_buffer_dirty(path->nodes[0]);
 	btrfs_free_path(path);
 
-	/* Always set parent to 0 here since its exclusive anyway. */
-	ret = btrfs_qgroup_record_ref(trans, fs_info, root_objectid,
-				      ins->objectid, ins->offset,
-				      BTRFS_QGROUP_OPER_ADD_EXCL, 0);
-	if (ret)
-		return ret;
-
 	ret = update_block_group(trans, root, ins->objectid, ins->offset, 1);
 	if (ret) { /* -ENOENT, logic error */
 		btrfs_err(fs_info, "update block group failed for %llu %llu",
@@ -7324,8 +7689,7 @@ static int alloc_reserved_tree_block(struct btrfs_trans_handle *trans,
 				     struct btrfs_root *root,
 				     u64 parent, u64 root_objectid,
 				     u64 flags, struct btrfs_disk_key *key,
-				     int level, struct btrfs_key *ins,
-				     int no_quota)
+				     int level, struct btrfs_key *ins)
 {
 	int ret;
 	struct btrfs_fs_info *fs_info = root->fs_info;
@@ -7391,14 +7755,6 @@ static int alloc_reserved_tree_block(struct btrfs_trans_handle *trans,
 	btrfs_mark_buffer_dirty(leaf);
 	btrfs_free_path(path);
 
-	if (!no_quota) {
-		ret = btrfs_qgroup_record_ref(trans, fs_info, root_objectid,
-					      ins->objectid, num_bytes,
-					      BTRFS_QGROUP_OPER_ADD_EXCL, 0);
-		if (ret)
-			return ret;
-	}
-
 	ret = update_block_group(trans, root, ins->objectid, root->nodesize,
 				 1);
 	if (ret) { /* -ENOENT, logic error */
@@ -7414,7 +7770,8 @@ static int alloc_reserved_tree_block(struct btrfs_trans_handle *trans,
 int btrfs_alloc_reserved_file_extent(struct btrfs_trans_handle *trans,
 				     struct btrfs_root *root,
 				     u64 root_objectid, u64 owner,
-				     u64 offset, struct btrfs_key *ins)
+				     u64 offset, u64 ram_bytes,
+				     struct btrfs_key *ins)
 {
 	int ret;
 
@@ -7423,7 +7780,8 @@ int btrfs_alloc_reserved_file_extent(struct btrfs_trans_handle *trans,
 	ret = btrfs_add_delayed_data_ref(root->fs_info, trans, ins->objectid,
 					 ins->offset, 0,
 					 root_objectid, owner, offset,
-					 BTRFS_ADD_DELAYED_EXTENT, NULL, 0);
+					 ram_bytes, BTRFS_ADD_DELAYED_EXTENT,
+					 NULL);
 	return ret;
 }
 
@@ -7566,9 +7924,6 @@ static void unuse_block_rsv(struct btrfs_fs_info *fs_info,
 
 /*
  * finds a free extent and does all the dirty work required for allocation
- * returns the key for the extent through ins, and a tree buffer for
- * the first block of the extent through buf.
- *
  * returns the tree buffer or an ERR_PTR on error.
  */
 struct extent_buffer *btrfs_alloc_tree_block(struct btrfs_trans_handle *trans,
@@ -7640,7 +7995,7 @@ struct extent_buffer *btrfs_alloc_tree_block(struct btrfs_trans_handle *trans,
 						 ins.objectid, ins.offset,
 						 parent, root_objectid, level,
 						 BTRFS_ADD_DELAYED_EXTENT,
-						 extent_op, 0);
+						 extent_op);
 		if (ret)
 			goto out_free_delayed;
 	}
@@ -7755,6 +8110,34 @@ reada:
 	wc->reada_slot = slot;
 }
 
+/*
+ * These may not be seen by the usual inc/dec ref code so we have to
+ * add them here.
+ */
+static int record_one_subtree_extent(struct btrfs_trans_handle *trans,
+				     struct btrfs_root *root, u64 bytenr,
+				     u64 num_bytes)
+{
+	struct btrfs_qgroup_extent_record *qrecord;
+	struct btrfs_delayed_ref_root *delayed_refs;
+
+	qrecord = kmalloc(sizeof(*qrecord), GFP_NOFS);
+	if (!qrecord)
+		return -ENOMEM;
+
+	qrecord->bytenr = bytenr;
+	qrecord->num_bytes = num_bytes;
+	qrecord->old_roots = NULL;
+
+	delayed_refs = &trans->transaction->delayed_refs;
+	spin_lock(&delayed_refs->lock);
+	if (btrfs_qgroup_insert_dirty_extent(delayed_refs, qrecord))
+		kfree(qrecord);
+	spin_unlock(&delayed_refs->lock);
+
+	return 0;
+}
+
 static int account_leaf_items(struct btrfs_trans_handle *trans,
 			      struct btrfs_root *root,
 			      struct extent_buffer *eb)
@@ -7765,6 +8148,10 @@ static int account_leaf_items(struct btrfs_trans_handle *trans,
 	struct btrfs_file_extent_item *fi;
 	u64 bytenr, num_bytes;
 
+	/* We can be called directly from walk_up_proc() */
+	if (!root->fs_info->quota_enabled)
+		return 0;
+
 	for (i = 0; i < nr; i++) {
 		btrfs_item_key_to_cpu(eb, &key, i);
 
@@ -7784,10 +8171,7 @@ static int account_leaf_items(struct btrfs_trans_handle *trans,
 
 		num_bytes = btrfs_file_extent_disk_num_bytes(eb, fi);
 
-		ret = btrfs_qgroup_record_ref(trans, root->fs_info,
-					      root->objectid,
-					      bytenr, num_bytes,
-					      BTRFS_QGROUP_OPER_SUB_SUBTREE, 0);
+		ret = record_one_subtree_extent(trans, root, bytenr, num_bytes);
 		if (ret)
 			return ret;
 	}
@@ -7920,7 +8304,11 @@ walk_down:
 			child_gen = btrfs_node_ptr_generation(eb, parent_slot);
 
 			eb = read_tree_block(root, child_bytenr, child_gen);
-			if (!eb || !extent_buffer_uptodate(eb)) {
+			if (IS_ERR(eb)) {
+				ret = PTR_ERR(eb);
+				goto out;
+			} else if (!extent_buffer_uptodate(eb)) {
+				free_extent_buffer(eb);
 				ret = -EIO;
 				goto out;
 			}
@@ -7932,15 +8320,10 @@ walk_down:
 			btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK);
 			path->locks[level] = BTRFS_READ_LOCK_BLOCKING;
 
-			ret = btrfs_qgroup_record_ref(trans, root->fs_info,
-						root->objectid,
-						child_bytenr,
-						root->nodesize,
-						BTRFS_QGROUP_OPER_SUB_SUBTREE,
-						0);
+			ret = record_one_subtree_extent(trans, root, child_bytenr,
+							root->nodesize);
 			if (ret)
 				goto out;
-
 		}
 
 		if (level == 0) {
@@ -8151,7 +8534,9 @@ static noinline int do_walk_down(struct btrfs_trans_handle *trans,
 		if (reada && level == 1)
 			reada_walk_down(trans, root, wc, path);
 		next = read_tree_block(root, bytenr, generation);
-		if (!next || !extent_buffer_uptodate(next)) {
+		if (IS_ERR(next)) {
+			return PTR_ERR(next);
+		} else if (!extent_buffer_uptodate(next)) {
 			free_extent_buffer(next);
 			return -EIO;
 		}
@@ -8184,14 +8569,15 @@ skip:
 			ret = account_shared_subtree(trans, root, next,
 						     generation, level - 1);
 			if (ret) {
-				printk_ratelimited(KERN_ERR "BTRFS: %s Error "
+				btrfs_err_rl(root->fs_info,
+					"Error "
 					"%d accounting shared subtree. Quota "
-					"is out of sync, rescan required.\n",
-					root->fs_info->sb->s_id, ret);
+					"is out of sync, rescan required.",
+					ret);
 			}
 		}
 		ret = btrfs_free_extent(trans, root, bytenr, blocksize, parent,
-				root->root_key.objectid, level - 1, 0, 0);
+				root->root_key.objectid, level - 1, 0);
 		BUG_ON(ret); /* -ENOMEM */
 	}
 	btrfs_tree_unlock(next);
@@ -8276,10 +8662,11 @@ static noinline int walk_up_proc(struct btrfs_trans_handle *trans,
 			BUG_ON(ret); /* -ENOMEM */
 			ret = account_leaf_items(trans, root, eb);
 			if (ret) {
-				printk_ratelimited(KERN_ERR "BTRFS: %s Error "
+				btrfs_err_rl(root->fs_info,
+					"error "
 					"%d accounting leaf items. Quota "
-					"is out of sync, rescan required.\n",
-					root->fs_info->sb->s_id, ret);
+					"is out of sync, rescan required.",
+					ret);
 			}
 		}
 		/* make block locked assertion in clean_tree_block happy */
@@ -8533,24 +8920,6 @@ int btrfs_drop_snapshot(struct btrfs_root *root,
 				goto out_end_trans;
 			}
 
-			/*
-			 * Qgroup update accounting is run from
-			 * delayed ref handling. This usually works
-			 * out because delayed refs are normally the
-			 * only way qgroup updates are added. However,
-			 * we may have added updates during our tree
-			 * walk so run qgroups here to make sure we
-			 * don't lose any updates.
-			 */
-			ret = btrfs_delayed_qgroup_accounting(trans,
-							      root->fs_info);
-			if (ret)
-				printk_ratelimited(KERN_ERR "BTRFS: Failure %d "
-						   "running qgroup updates "
-						   "during snapshot delete. "
-						   "Quota is out of sync, "
-						   "rescan required.\n", ret);
-
 			btrfs_end_transaction_throttle(trans, tree_root);
 			if (!for_reloc && btrfs_need_cleaner_sleep(root)) {
 				pr_debug("BTRFS: drop snapshot early exit\n");
@@ -8596,7 +8965,7 @@ int btrfs_drop_snapshot(struct btrfs_root *root,
 	}
 
 	if (test_bit(BTRFS_ROOT_IN_RADIX, &root->state)) {
-		btrfs_drop_and_free_fs_root(tree_root->fs_info, root);
+		btrfs_add_dropped_root(trans, root);
 	} else {
 		free_extent_buffer(root->node);
 		free_extent_buffer(root->commit_root);
@@ -8604,14 +8973,6 @@ int btrfs_drop_snapshot(struct btrfs_root *root,
 	}
 	root_dropped = true;
 out_end_trans:
-	ret = btrfs_delayed_qgroup_accounting(trans, tree_root->fs_info);
-	if (ret)
-		printk_ratelimited(KERN_ERR "BTRFS: Failure %d "
-				   "running qgroup updates "
-				   "during snapshot delete. "
-				   "Quota is out of sync, "
-				   "rescan required.\n", ret);
-
 	btrfs_end_transaction_throttle(trans, tree_root);
 out_free:
 	kfree(wc);
@@ -8627,7 +8988,7 @@ out:
 	if (!for_reloc && root_dropped == false)
 		btrfs_add_dead_root(root);
 	if (err && err != -EAGAIN)
-		btrfs_std_error(root->fs_info, err);
+		btrfs_std_error(root->fs_info, err, NULL);
 	return err;
 }
 
@@ -8751,14 +9112,13 @@ static u64 update_block_group_flags(struct btrfs_root *root, u64 flags)
 	return flags;
 }
 
-static int set_block_group_ro(struct btrfs_block_group_cache *cache, int force)
+static int inc_block_group_ro(struct btrfs_block_group_cache *cache, int force)
 {
 	struct btrfs_space_info *sinfo = cache->space_info;
 	u64 num_bytes;
 	u64 min_allocable_bytes;
 	int ret = -ENOSPC;
 
-
 	/*
 	 * We need some metadata space and system metadata space for
 	 * allocating chunks in some corner cases until we force to set
@@ -8775,6 +9135,7 @@ static int set_block_group_ro(struct btrfs_block_group_cache *cache, int force)
 	spin_lock(&cache->lock);
 
 	if (cache->ro) {
+		cache->ro++;
 		ret = 0;
 		goto out;
 	}
@@ -8786,7 +9147,7 @@ static int set_block_group_ro(struct btrfs_block_group_cache *cache, int force)
 	    sinfo->bytes_may_use + sinfo->bytes_readonly + num_bytes +
 	    min_allocable_bytes <= sinfo->total_bytes) {
 		sinfo->bytes_readonly += num_bytes;
-		cache->ro = 1;
+		cache->ro++;
 		list_add_tail(&cache->ro_list, &sinfo->ro_bgs);
 		ret = 0;
 	}
@@ -8796,7 +9157,7 @@ out:
 	return ret;
 }
 
-int btrfs_set_block_group_ro(struct btrfs_root *root,
+int btrfs_inc_block_group_ro(struct btrfs_root *root,
 			     struct btrfs_block_group_cache *cache)
 
 {
@@ -8804,8 +9165,6 @@ int btrfs_set_block_group_ro(struct btrfs_root *root,
 	u64 alloc_flags;
 	int ret;
 
-	BUG_ON(cache->ro);
-
 again:
 	trans = btrfs_join_transaction(root);
 	if (IS_ERR(trans))
@@ -8817,7 +9176,7 @@ again:
 	 * back off and let this transaction commit
 	 */
 	mutex_lock(&root->fs_info->ro_block_group_mutex);
-	if (trans->transaction->dirty_bg_run) {
+	if (test_bit(BTRFS_TRANS_DIRTY_BG_RUN, &trans->transaction->flags)) {
 		u64 transid = trans->transid;
 
 		mutex_unlock(&root->fs_info->ro_block_group_mutex);
@@ -8848,7 +9207,7 @@ again:
 			goto out;
 	}
 
-	ret = set_block_group_ro(cache, 0);
+	ret = inc_block_group_ro(cache, 0);
 	if (!ret)
 		goto out;
 	alloc_flags = get_alloc_profile(root, cache->space_info->flags);
@@ -8856,7 +9215,7 @@ again:
 			     CHUNK_ALLOC_FORCE);
 	if (ret < 0)
 		goto out;
-	ret = set_block_group_ro(cache, 0);
+	ret = inc_block_group_ro(cache, 0);
 out:
 	if (cache->flags & BTRFS_BLOCK_GROUP_SYSTEM) {
 		alloc_flags = update_block_group_flags(root, cache->flags);
@@ -8919,7 +9278,7 @@ u64 btrfs_account_ro_block_groups_free_space(struct btrfs_space_info *sinfo)
 	return free_bytes;
 }
 
-void btrfs_set_block_group_rw(struct btrfs_root *root,
+void btrfs_dec_block_group_ro(struct btrfs_root *root,
 			      struct btrfs_block_group_cache *cache)
 {
 	struct btrfs_space_info *sinfo = cache->space_info;
@@ -8929,11 +9288,13 @@ void btrfs_set_block_group_rw(struct btrfs_root *root,
 
 	spin_lock(&sinfo->lock);
 	spin_lock(&cache->lock);
-	num_bytes = cache->key.offset - cache->reserved - cache->pinned -
-		    cache->bytes_super - btrfs_block_group_used(&cache->item);
-	sinfo->bytes_readonly -= num_bytes;
-	cache->ro = 0;
-	list_del_init(&cache->ro_list);
+	if (!--cache->ro) {
+		num_bytes = cache->key.offset - cache->reserved -
+			    cache->pinned - cache->bytes_super -
+			    btrfs_block_group_used(&cache->item);
+		sinfo->bytes_readonly -= num_bytes;
+		list_del_init(&cache->ro_list);
+	}
 	spin_unlock(&cache->lock);
 	spin_unlock(&sinfo->lock);
 }
@@ -9449,7 +9810,7 @@ int btrfs_read_block_groups(struct btrfs_root *root)
 
 		set_avail_alloc_bits(root->fs_info, cache->flags);
 		if (btrfs_chunk_readonly(root, cache->key.objectid)) {
-			set_block_group_ro(cache, 1);
+			inc_block_group_ro(cache, 1);
 		} else if (btrfs_block_group_used(&cache->item) == 0) {
 			spin_lock(&info->unused_bgs_lock);
 			/* Should always be true but just in case. */
@@ -9477,11 +9838,11 @@ int btrfs_read_block_groups(struct btrfs_root *root)
 		list_for_each_entry(cache,
 				&space_info->block_groups[BTRFS_RAID_RAID0],
 				list)
-			set_block_group_ro(cache, 1);
+			inc_block_group_ro(cache, 1);
 		list_for_each_entry(cache,
 				&space_info->block_groups[BTRFS_RAID_SINGLE],
 				list)
-			set_block_group_ro(cache, 1);
+			inc_block_group_ro(cache, 1);
 	}
 
 	init_global_block_rsv(info);
@@ -9499,7 +9860,9 @@ void btrfs_create_pending_block_groups(struct btrfs_trans_handle *trans,
 	struct btrfs_block_group_item item;
 	struct btrfs_key key;
 	int ret = 0;
+	bool can_flush_pending_bgs = trans->can_flush_pending_bgs;
 
+	trans->can_flush_pending_bgs = false;
 	list_for_each_entry_safe(block_group, tmp, &trans->new_bgs, bg_list) {
 		if (ret)
 			goto next;
@@ -9520,6 +9883,7 @@ void btrfs_create_pending_block_groups(struct btrfs_trans_handle *trans,
 next:
 		list_del_init(&block_group->bg_list);
 	}
+	trans->can_flush_pending_bgs = can_flush_pending_bgs;
 }
 
 int btrfs_make_block_group(struct btrfs_trans_handle *trans,
@@ -9562,6 +9926,27 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans,
 
 	free_excluded_extents(root, cache);
 
+#ifdef CONFIG_BTRFS_DEBUG
+	if (btrfs_should_fragment_free_space(root, cache)) {
+		u64 new_bytes_used = size - bytes_used;
+
+		bytes_used += new_bytes_used >> 1;
+		fragment_free_space(root, cache);
+	}
+#endif
+	/*
+	 * Call to ensure the corresponding space_info object is created and
+	 * assigned to our block group, but don't update its counters just yet.
+	 * We want our bg to be added to the rbtree with its ->space_info set.
+	 */
+	ret = update_space_info(root->fs_info, cache->flags, 0, 0,
+				&cache->space_info);
+	if (ret) {
+		btrfs_remove_free_space_cache(cache);
+		btrfs_put_block_group(cache);
+		return ret;
+	}
+
 	ret = btrfs_add_block_group_cache(root->fs_info, cache);
 	if (ret) {
 		btrfs_remove_free_space_cache(cache);
@@ -9569,6 +9954,10 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans,
 		return ret;
 	}
 
+	/*
+	 * Now that our block group has its ->space_info set and is inserted in
+	 * the rbtree, update the space info's counters.
+	 */
 	ret = update_space_info(root->fs_info, cache->flags, size, bytes_used,
 				&cache->space_info);
 	if (ret) {
@@ -9845,6 +10234,11 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans,
 	 * currently running transaction might finish and a new one start,
 	 * allowing for new block groups to be created that can reuse the same
 	 * physical device locations unless we take this special care.
+	 *
+	 * There may also be an implicit trim operation if the file system
+	 * is mounted with -odiscard. The same protections must remain
+	 * in place until the extents have been discarded completely when
+	 * the transaction commit has completed.
 	 */
 	remove_em = (atomic_read(&block_group->trimming) == 0);
 	/*
@@ -9901,6 +10295,47 @@ out:
 	return ret;
 }
 
+struct btrfs_trans_handle *
+btrfs_start_trans_remove_block_group(struct btrfs_fs_info *fs_info,
+				     const u64 chunk_offset)
+{
+	struct extent_map_tree *em_tree = &fs_info->mapping_tree.map_tree;
+	struct extent_map *em;
+	struct map_lookup *map;
+	unsigned int num_items;
+
+	read_lock(&em_tree->lock);
+	em = lookup_extent_mapping(em_tree, chunk_offset, 1);
+	read_unlock(&em_tree->lock);
+	ASSERT(em && em->start == chunk_offset);
+
+	/*
+	 * We need to reserve 3 + N units from the metadata space info in order
+	 * to remove a block group (done at btrfs_remove_chunk() and at
+	 * btrfs_remove_block_group()), which are used for:
+	 *
+	 * 1 unit for adding the free space inode's orphan (located in the tree
+	 * of tree roots).
+	 * 1 unit for deleting the block group item (located in the extent
+	 * tree).
+	 * 1 unit for deleting the free space item (located in tree of tree
+	 * roots).
+	 * N units for deleting N device extent items corresponding to each
+	 * stripe (located in the device tree).
+	 *
+	 * In order to remove a block group we also need to reserve units in the
+	 * system space info in order to update the chunk tree (update one or
+	 * more device items and remove one chunk item), but this is done at
+	 * btrfs_remove_chunk() through a call to check_system_chunk().
+	 */
+	map = (struct map_lookup *)em->bdev;
+	num_items = 3 + map->num_stripes;
+	free_extent_map(em);
+
+	return btrfs_start_transaction_fallback_global_rsv(fs_info->extent_root,
+							   num_items, 1);
+}
+
 /*
  * Process the unused_bgs list and remove any that don't have any allocated
  * space inside of them.
@@ -9919,24 +10354,30 @@ void btrfs_delete_unused_bgs(struct btrfs_fs_info *fs_info)
 	spin_lock(&fs_info->unused_bgs_lock);
 	while (!list_empty(&fs_info->unused_bgs)) {
 		u64 start, end;
+		int trimming;
 
 		block_group = list_first_entry(&fs_info->unused_bgs,
 					       struct btrfs_block_group_cache,
 					       bg_list);
-		space_info = block_group->space_info;
 		list_del_init(&block_group->bg_list);
+
+		space_info = block_group->space_info;
+
 		if (ret || btrfs_mixed_space_info(space_info)) {
 			btrfs_put_block_group(block_group);
 			continue;
 		}
 		spin_unlock(&fs_info->unused_bgs_lock);
 
+		mutex_lock(&fs_info->delete_unused_bgs_mutex);
+
 		/* Don't want to race with allocators so take the groups_sem */
 		down_write(&space_info->groups_sem);
 		spin_lock(&block_group->lock);
 		if (block_group->reserved ||
 		    btrfs_block_group_used(&block_group->item) ||
-		    block_group->ro) {
+		    block_group->ro ||
+		    list_is_singular(&block_group->list)) {
 			/*
 			 * We want to bail if we made new allocations or have
 			 * outstanding allocations in this block group.  We do
@@ -9950,7 +10391,7 @@ void btrfs_delete_unused_bgs(struct btrfs_fs_info *fs_info)
 		spin_unlock(&block_group->lock);
 
 		/* We don't want to force the issue, only flip if it's ok. */
-		ret = set_block_group_ro(block_group, 0);
+		ret = inc_block_group_ro(block_group, 0);
 		up_write(&space_info->groups_sem);
 		if (ret < 0) {
 			ret = 0;
@@ -9961,10 +10402,10 @@ void btrfs_delete_unused_bgs(struct btrfs_fs_info *fs_info)
 		 * Want to do this before we do anything else so we can recover
 		 * properly if we fail to join the transaction.
 		 */
-		/* 1 for btrfs_orphan_reserve_metadata() */
-		trans = btrfs_start_transaction(root, 1);
+		trans = btrfs_start_trans_remove_block_group(fs_info,
+						     block_group->key.objectid);
 		if (IS_ERR(trans)) {
-			btrfs_set_block_group_rw(root, block_group);
+			btrfs_dec_block_group_ro(root, block_group);
 			ret = PTR_ERR(trans);
 			goto next;
 		}
@@ -9991,14 +10432,14 @@ void btrfs_delete_unused_bgs(struct btrfs_fs_info *fs_info)
 				  EXTENT_DIRTY, GFP_NOFS);
 		if (ret) {
 			mutex_unlock(&fs_info->unused_bg_unpin_mutex);
-			btrfs_set_block_group_rw(root, block_group);
+			btrfs_dec_block_group_ro(root, block_group);
 			goto end_trans;
 		}
 		ret = clear_extent_bits(&fs_info->freed_extents[1], start, end,
 				  EXTENT_DIRTY, GFP_NOFS);
 		if (ret) {
 			mutex_unlock(&fs_info->unused_bg_unpin_mutex);
-			btrfs_set_block_group_rw(root, block_group);
+			btrfs_dec_block_group_ro(root, block_group);
 			goto end_trans;
 		}
 		mutex_unlock(&fs_info->unused_bg_unpin_mutex);
@@ -10016,15 +10457,47 @@ void btrfs_delete_unused_bgs(struct btrfs_fs_info *fs_info)
 		spin_unlock(&block_group->lock);
 		spin_unlock(&space_info->lock);
 
+		/* DISCARD can flip during remount */
+		trimming = btrfs_test_opt(root, DISCARD);
+
+		/* Implicit trim during transaction commit. */
+		if (trimming)
+			btrfs_get_block_group_trimming(block_group);
+
 		/*
 		 * Btrfs_remove_chunk will abort the transaction if things go
 		 * horribly wrong.
 		 */
 		ret = btrfs_remove_chunk(trans, root,
 					 block_group->key.objectid);
+
+		if (ret) {
+			if (trimming)
+				btrfs_put_block_group_trimming(block_group);
+			goto end_trans;
+		}
+
+		/*
+		 * If we're not mounted with -odiscard, we can just forget
+		 * about this block group. Otherwise we'll need to wait
+		 * until transaction commit to do the actual discard.
+		 */
+		if (trimming) {
+			spin_lock(&fs_info->unused_bgs_lock);
+			/*
+			 * A concurrent scrub might have added us to the list
+			 * fs_info->unused_bgs, so use a list_move operation
+			 * to add the block group to the deleted_bgs list.
+			 */
+			list_move(&block_group->bg_list,
+				  &trans->transaction->deleted_bgs);
+			spin_unlock(&fs_info->unused_bgs_lock);
+			btrfs_get_block_group(block_group);
+		}
 end_trans:
 		btrfs_end_transaction(trans, root);
 next:
+		mutex_unlock(&fs_info->delete_unused_bgs_mutex);
 		btrfs_put_block_group(block_group);
 		spin_lock(&fs_info->unused_bgs_lock);
 	}
@@ -10074,10 +10547,99 @@ int btrfs_error_unpin_extent_range(struct btrfs_root *root, u64 start, u64 end)
 	return unpin_extent_range(root, start, end, false);
 }
 
+/*
+ * It used to be that old block groups would be left around forever.
+ * Iterating over them would be enough to trim unused space.  Since we
+ * now automatically remove them, we also need to iterate over unallocated
+ * space.
+ *
+ * We don't want a transaction for this since the discard may take a
+ * substantial amount of time.  We don't require that a transaction be
+ * running, but we do need to take a running transaction into account
+ * to ensure that we're not discarding chunks that were released in
+ * the current transaction.
+ *
+ * Holding the chunks lock will prevent other threads from allocating
+ * or releasing chunks, but it won't prevent a running transaction
+ * from committing and releasing the memory that the pending chunks
+ * list head uses.  For that, we need to take a reference to the
+ * transaction.
+ */
+static int btrfs_trim_free_extents(struct btrfs_device *device,
+				   u64 minlen, u64 *trimmed)
+{
+	u64 start = 0, len = 0;
+	int ret;
+
+	*trimmed = 0;
+
+	/* Not writeable = nothing to do. */
+	if (!device->writeable)
+		return 0;
+
+	/* No free space = nothing to do. */
+	if (device->total_bytes <= device->bytes_used)
+		return 0;
+
+	ret = 0;
+
+	while (1) {
+		struct btrfs_fs_info *fs_info = device->dev_root->fs_info;
+		struct btrfs_transaction *trans;
+		u64 bytes;
+
+		ret = mutex_lock_interruptible(&fs_info->chunk_mutex);
+		if (ret)
+			return ret;
+
+		down_read(&fs_info->commit_root_sem);
+
+		spin_lock(&fs_info->trans_lock);
+		trans = fs_info->running_transaction;
+		if (trans)
+			atomic_inc(&trans->use_count);
+		spin_unlock(&fs_info->trans_lock);
+
+		ret = find_free_dev_extent_start(trans, device, minlen, start,
+						 &start, &len);
+		if (trans)
+			btrfs_put_transaction(trans);
+
+		if (ret) {
+			up_read(&fs_info->commit_root_sem);
+			mutex_unlock(&fs_info->chunk_mutex);
+			if (ret == -ENOSPC)
+				ret = 0;
+			break;
+		}
+
+		ret = btrfs_issue_discard(device->bdev, start, len, &bytes);
+		up_read(&fs_info->commit_root_sem);
+		mutex_unlock(&fs_info->chunk_mutex);
+
+		if (ret)
+			break;
+
+		start += len;
+		*trimmed += bytes;
+
+		if (fatal_signal_pending(current)) {
+			ret = -ERESTARTSYS;
+			break;
+		}
+
+		cond_resched();
+	}
+
+	return ret;
+}
+
 int btrfs_trim_fs(struct btrfs_root *root, struct fstrim_range *range)
 {
 	struct btrfs_fs_info *fs_info = root->fs_info;
 	struct btrfs_block_group_cache *cache = NULL;
+	struct btrfs_device *device;
+	struct list_head *devices;
 	u64 group_trimmed;
 	u64 start;
 	u64 end;
@@ -10132,6 +10694,18 @@ int btrfs_trim_fs(struct btrfs_root *root, struct fstrim_range *range)
 		cache = next_block_group(fs_info->tree_root, cache);
 	}
 
+	mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
+	devices = &root->fs_info->fs_devices->alloc_list;
+	list_for_each_entry(device, devices, dev_alloc_list) {
+		ret = btrfs_trim_free_extents(device, range->minlen,
+					      &group_trimmed);
+		if (ret)
+			break;
+
+		trimmed += group_trimmed;
+	}
+	mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
+
 	range->len = trimmed;
 	return ret;
 }
@@ -10148,8 +10722,7 @@ void btrfs_end_write_no_snapshoting(struct btrfs_root *root)
 {
 	percpu_counter_dec(&root->subv_writers->counter);
 	/*
-	 * Make sure counter is updated before we wake up
-	 * waiters.
+	 * Make sure counter is updated before we wake up waiters.
 	 */
 	smp_mb();
 	if (waitqueue_active(&root->subv_writers->wait))
diff --git a/kernel/fs/btrfs/extent-tree.h b/kernel/fs/btrfs/extent-tree.h
new file mode 100644
index 000000000..e69de29bb
--- /dev/null
+++ b/kernel/fs/btrfs/extent-tree.h
diff --git a/kernel/fs/btrfs/extent_io.c b/kernel/fs/btrfs/extent_io.c
index c32d226bf..9abe18763 100644
--- a/kernel/fs/btrfs/extent_io.c
+++ b/kernel/fs/btrfs/extent_io.c
@@ -96,8 +96,8 @@ static inline void __btrfs_debug_check_extent_io_range(const char *caller,
 	inode = tree->mapping->host;
 	isize = i_size_read(inode);
 	if (end >= PAGE_SIZE && (end % 2) == 0 && end != isize - 1) {
-		printk_ratelimited(KERN_DEBUG
-		    "BTRFS: %s: ino %llu isize %llu odd range [%llu,%llu]\n",
+		btrfs_debug_rl(BTRFS_I(inode)->root->fs_info,
+		    "%s: ino %llu isize %llu odd range [%llu,%llu]",
 				caller, btrfs_ino(inode), isize, start, end);
 	}
 }
@@ -131,6 +131,25 @@ struct extent_page_data {
 	unsigned int sync_io:1;
 };
 
+static void add_extent_changeset(struct extent_state *state, unsigned bits,
+				 struct extent_changeset *changeset,
+				 int set)
+{
+	int ret;
+
+	if (!changeset)
+		return;
+	if (set && (state->state & bits) == bits)
+		return;
+	if (!set && (state->state & bits) == 0)
+		return;
+	changeset->bytes_changed += state->end - state->start + 1;
+	ret = ulist_add(changeset->range_changed, state->start, state->end,
+			GFP_ATOMIC);
+	/* ENOMEM */
+	BUG_ON(ret < 0);
+}
+
 static noinline void flush_write_bio(void *data);
 static inline struct btrfs_fs_info *
 tree_fs_info(struct extent_io_tree *tree)
@@ -410,7 +429,8 @@ static void clear_state_cb(struct extent_io_tree *tree,
 }
 
 static void set_state_bits(struct extent_io_tree *tree,
-			   struct extent_state *state, unsigned *bits);
+			   struct extent_state *state, unsigned *bits,
+			   struct extent_changeset *changeset);
 
 /*
  * insert an extent_state struct into the tree.  'bits' are set on the
@@ -426,7 +446,7 @@ static int insert_state(struct extent_io_tree *tree,
 			struct extent_state *state, u64 start, u64 end,
 			struct rb_node ***p,
 			struct rb_node **parent,
-			unsigned *bits)
+			unsigned *bits, struct extent_changeset *changeset)
 {
 	struct rb_node *node;
 
@@ -436,7 +456,7 @@ static int insert_state(struct extent_io_tree *tree,
 	state->start = start;
 	state->end = end;
 
-	set_state_bits(tree, state, bits);
+	set_state_bits(tree, state, bits, changeset);
 
 	node = tree_insert(&tree->state, NULL, end, &state->rb_node, p, parent);
 	if (node) {
@@ -511,7 +531,8 @@ static struct extent_state *next_state(struct extent_state *state)
  */
 static struct extent_state *clear_state_bit(struct extent_io_tree *tree,
 					    struct extent_state *state,
-					    unsigned *bits, int wake)
+					    unsigned *bits, int wake,
+					    struct extent_changeset *changeset)
 {
 	struct extent_state *next;
 	unsigned bits_to_clear = *bits & ~EXTENT_CTLBITS;
@@ -522,6 +543,7 @@ static struct extent_state *clear_state_bit(struct extent_io_tree *tree,
 		tree->dirty_bytes -= range;
 	}
 	clear_state_cb(tree, state, bits);
+	add_extent_changeset(state, bits_to_clear, changeset, 0);
 	state->state &= ~bits_to_clear;
 	if (wake)
 		wake_up(&state->wq);
@@ -569,10 +591,10 @@ static void extent_io_tree_panic(struct extent_io_tree *tree, int err)
  *
  * This takes the tree lock, and returns 0 on success and < 0 on error.
  */
-int clear_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
-		     unsigned bits, int wake, int delete,
-		     struct extent_state **cached_state,
-		     gfp_t mask)
+static int __clear_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
+			      unsigned bits, int wake, int delete,
+			      struct extent_state **cached_state,
+			      gfp_t mask, struct extent_changeset *changeset)
 {
 	struct extent_state *state;
 	struct extent_state *cached;
@@ -594,7 +616,7 @@ int clear_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
 	if (bits & (EXTENT_IOBITS | EXTENT_BOUNDARY))
 		clear = 1;
 again:
-	if (!prealloc && (mask & __GFP_WAIT)) {
+	if (!prealloc && gfpflags_allow_blocking(mask)) {
 		/*
 		 * Don't care for allocation failure here because we might end
 		 * up not needing the pre-allocated extent state at all, which
@@ -671,7 +693,8 @@ hit_next:
 		if (err)
 			goto out;
 		if (state->end <= end) {
-			state = clear_state_bit(tree, state, &bits, wake);
+			state = clear_state_bit(tree, state, &bits, wake,
+						changeset);
 			goto next;
 		}
 		goto search_again;
@@ -692,13 +715,13 @@ hit_next:
 		if (wake)
 			wake_up(&state->wq);
 
-		clear_state_bit(tree, prealloc, &bits, wake);
+		clear_state_bit(tree, prealloc, &bits, wake, changeset);
 
 		prealloc = NULL;
 		goto out;
 	}
 
-	state = clear_state_bit(tree, state, &bits, wake);
+	state = clear_state_bit(tree, state, &bits, wake, changeset);
 next:
 	if (last_end == (u64)-1)
 		goto out;
@@ -718,7 +741,7 @@ search_again:
 	if (start > end)
 		goto out;
 	spin_unlock(&tree->lock);
-	if (mask & __GFP_WAIT)
+	if (gfpflags_allow_blocking(mask))
 		cond_resched();
 	goto again;
 }
@@ -789,7 +812,7 @@ out:
 
 static void set_state_bits(struct extent_io_tree *tree,
 			   struct extent_state *state,
-			   unsigned *bits)
+			   unsigned *bits, struct extent_changeset *changeset)
 {
 	unsigned bits_to_set = *bits & ~EXTENT_CTLBITS;
 
@@ -798,6 +821,7 @@ static void set_state_bits(struct extent_io_tree *tree,
 		u64 range = state->end - state->start + 1;
 		tree->dirty_bytes += range;
 	}
+	add_extent_changeset(state, bits_to_set, changeset, 1);
 	state->state |= bits_to_set;
 }
 
@@ -835,7 +859,7 @@ static int __must_check
 __set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
 		 unsigned bits, unsigned exclusive_bits,
 		 u64 *failed_start, struct extent_state **cached_state,
-		 gfp_t mask)
+		 gfp_t mask, struct extent_changeset *changeset)
 {
 	struct extent_state *state;
 	struct extent_state *prealloc = NULL;
@@ -850,7 +874,7 @@ __set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
 
 	bits |= EXTENT_FIRST_DELALLOC;
 again:
-	if (!prealloc && (mask & __GFP_WAIT)) {
+	if (!prealloc && gfpflags_allow_blocking(mask)) {
 		prealloc = alloc_extent_state(mask);
 		BUG_ON(!prealloc);
 	}
@@ -873,7 +897,7 @@ again:
 		prealloc = alloc_extent_state_atomic(prealloc);
 		BUG_ON(!prealloc);
 		err = insert_state(tree, prealloc, start, end,
-				   &p, &parent, &bits);
+				   &p, &parent, &bits, changeset);
 		if (err)
 			extent_io_tree_panic(tree, err);
 
@@ -899,7 +923,7 @@ hit_next:
 			goto out;
 		}
 
-		set_state_bits(tree, state, &bits);
+		set_state_bits(tree, state, &bits, changeset);
 		cache_state(state, cached_state);
 		merge_state(tree, state);
 		if (last_end == (u64)-1)
@@ -945,7 +969,7 @@ hit_next:
 		if (err)
 			goto out;
 		if (state->end <= end) {
-			set_state_bits(tree, state, &bits);
+			set_state_bits(tree, state, &bits, changeset);
 			cache_state(state, cached_state);
 			merge_state(tree, state);
 			if (last_end == (u64)-1)
@@ -980,7 +1004,7 @@ hit_next:
 		 * the later extent.
 		 */
 		err = insert_state(tree, prealloc, start, this_end,
-				   NULL, NULL, &bits);
+				   NULL, NULL, &bits, changeset);
 		if (err)
 			extent_io_tree_panic(tree, err);
 
@@ -1008,7 +1032,7 @@ hit_next:
 		if (err)
 			extent_io_tree_panic(tree, err);
 
-		set_state_bits(tree, prealloc, &bits);
+		set_state_bits(tree, prealloc, &bits, changeset);
 		cache_state(prealloc, cached_state);
 		merge_state(tree, prealloc);
 		prealloc = NULL;
@@ -1028,7 +1052,7 @@ search_again:
 	if (start > end)
 		goto out;
 	spin_unlock(&tree->lock);
-	if (mask & __GFP_WAIT)
+	if (gfpflags_allow_blocking(mask))
 		cond_resched();
 	goto again;
 }
@@ -1038,7 +1062,7 @@ int set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
 		   struct extent_state **cached_state, gfp_t mask)
 {
 	return __set_extent_bit(tree, start, end, bits, 0, failed_start,
-				cached_state, mask);
+				cached_state, mask, NULL);
 }
 
 
@@ -1076,7 +1100,7 @@ int convert_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
 	btrfs_debug_check_extent_io_range(tree, start, end);
 
 again:
-	if (!prealloc && (mask & __GFP_WAIT)) {
+	if (!prealloc && gfpflags_allow_blocking(mask)) {
 		/*
 		 * Best effort, don't worry if extent state allocation fails
 		 * here for the first iteration. We might have a cached state
@@ -1111,7 +1135,7 @@ again:
 			goto out;
 		}
 		err = insert_state(tree, prealloc, start, end,
-				   &p, &parent, &bits);
+				   &p, &parent, &bits, NULL);
 		if (err)
 			extent_io_tree_panic(tree, err);
 		cache_state(prealloc, cached_state);
@@ -1130,9 +1154,9 @@ hit_next:
 	 * Just lock what we found and keep going
 	 */
 	if (state->start == start && state->end <= end) {
-		set_state_bits(tree, state, &bits);
+		set_state_bits(tree, state, &bits, NULL);
 		cache_state(state, cached_state);
-		state = clear_state_bit(tree, state, &clear_bits, 0);
+		state = clear_state_bit(tree, state, &clear_bits, 0, NULL);
 		if (last_end == (u64)-1)
 			goto out;
 		start = last_end + 1;
@@ -1171,9 +1195,10 @@ hit_next:
 		if (err)
 			goto out;
 		if (state->end <= end) {
-			set_state_bits(tree, state, &bits);
+			set_state_bits(tree, state, &bits, NULL);
 			cache_state(state, cached_state);
-			state = clear_state_bit(tree, state, &clear_bits, 0);
+			state = clear_state_bit(tree, state, &clear_bits, 0,
+						NULL);
 			if (last_end == (u64)-1)
 				goto out;
 			start = last_end + 1;
@@ -1208,7 +1233,7 @@ hit_next:
 		 * the later extent.
 		 */
 		err = insert_state(tree, prealloc, start, this_end,
-				   NULL, NULL, &bits);
+				   NULL, NULL, &bits, NULL);
 		if (err)
 			extent_io_tree_panic(tree, err);
 		cache_state(prealloc, cached_state);
@@ -1233,9 +1258,9 @@ hit_next:
 		if (err)
 			extent_io_tree_panic(tree, err);
 
-		set_state_bits(tree, prealloc, &bits);
+		set_state_bits(tree, prealloc, &bits, NULL);
 		cache_state(prealloc, cached_state);
-		clear_state_bit(tree, prealloc, &clear_bits, 0);
+		clear_state_bit(tree, prealloc, &clear_bits, 0, NULL);
 		prealloc = NULL;
 		goto out;
 	}
@@ -1253,7 +1278,7 @@ search_again:
 	if (start > end)
 		goto out;
 	spin_unlock(&tree->lock);
-	if (mask & __GFP_WAIT)
+	if (gfpflags_allow_blocking(mask))
 		cond_resched();
 	first_iteration = false;
 	goto again;
@@ -1274,10 +1299,53 @@ int set_extent_bits(struct extent_io_tree *tree, u64 start, u64 end,
 			      NULL, mask);
 }
 
+int set_record_extent_bits(struct extent_io_tree *tree, u64 start, u64 end,
+			   unsigned bits, gfp_t mask,
+			   struct extent_changeset *changeset)
+{
+	/*
+	 * We don't support EXTENT_LOCKED yet, as current changeset will
+	 * record any bits changed, so for EXTENT_LOCKED case, it will
+	 * either fail with -EEXIST or changeset will record the whole
+	 * range.
+	 */
+	BUG_ON(bits & EXTENT_LOCKED);
+
+	return __set_extent_bit(tree, start, end, bits, 0, NULL, NULL, mask,
+				changeset);
+}
+
+int clear_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
+		     unsigned bits, int wake, int delete,
+		     struct extent_state **cached, gfp_t mask)
+{
+	return __clear_extent_bit(tree, start, end, bits, wake, delete,
+				  cached, mask, NULL);
+}
+
 int clear_extent_bits(struct extent_io_tree *tree, u64 start, u64 end,
 		      unsigned bits, gfp_t mask)
 {
-	return clear_extent_bit(tree, start, end, bits, 0, 0, NULL, mask);
+	int wake = 0;
+
+	if (bits & EXTENT_LOCKED)
+		wake = 1;
+
+	return clear_extent_bit(tree, start, end, bits, wake, 0, NULL, mask);
+}
+
+int clear_record_extent_bits(struct extent_io_tree *tree, u64 start, u64 end,
+			     unsigned bits, gfp_t mask,
+			     struct extent_changeset *changeset)
+{
+	/*
+	 * Don't support EXTENT_LOCKED case, same reason as
+	 * set_record_extent_bits().
+	 */
+	BUG_ON(bits & EXTENT_LOCKED);
+
+	return __clear_extent_bit(tree, start, end, bits, 0, 0, NULL, mask,
+				  changeset);
 }
 
 int set_extent_delalloc(struct extent_io_tree *tree, u64 start, u64 end,
@@ -1338,7 +1406,7 @@ int lock_extent_bits(struct extent_io_tree *tree, u64 start, u64 end,
 	while (1) {
 		err = __set_extent_bit(tree, start, end, EXTENT_LOCKED | bits,
 				       EXTENT_LOCKED, &failed_start,
-				       cached_state, GFP_NOFS);
+				       cached_state, GFP_NOFS, NULL);
 		if (err == -EEXIST) {
 			wait_extent_bit(tree, failed_start, end, EXTENT_LOCKED);
 			start = failed_start;
@@ -1360,7 +1428,7 @@ int try_lock_extent(struct extent_io_tree *tree, u64 start, u64 end)
 	u64 failed_start;
 
 	err = __set_extent_bit(tree, start, end, EXTENT_LOCKED, EXTENT_LOCKED,
-			       &failed_start, NULL, GFP_NOFS);
+			       &failed_start, NULL, GFP_NOFS, NULL);
 	if (err == -EEXIST) {
 		if (failed_start > start)
 			clear_extent_bit(tree, start, failed_start - 1,
@@ -2073,8 +2141,8 @@ int repair_io_failure(struct inode *inode, u64 start, u64 length, u64 logical,
 		return -EIO;
 	}
 
-	printk_ratelimited_in_rcu(KERN_INFO
-				  "BTRFS: read error corrected: ino %llu off %llu (dev %s sector %llu)\n",
+	btrfs_info_rl_in_rcu(fs_info,
+		"read error corrected: ino %llu off %llu (dev %s sector %llu)",
 				  btrfs_ino(inode), start,
 				  rcu_str_deref(dev->name), sector);
 	bio_put(bio);
@@ -2481,7 +2549,7 @@ int end_extent_writepage(struct page *page, int err, u64 start, u64 end)
  * Scheduling is not allowed, so the extent state tree is expected
  * to have one and only one object corresponding to this IO.
  */
-static void end_bio_extent_writepage(struct bio *bio, int err)
+static void end_bio_extent_writepage(struct bio *bio)
 {
 	struct bio_vec *bvec;
 	u64 start;
@@ -2511,7 +2579,7 @@ static void end_bio_extent_writepage(struct bio *bio, int err)
 		start = page_offset(page);
 		end = start + bvec->bv_offset + bvec->bv_len - 1;
 
-		if (end_extent_writepage(page, err, start, end))
+		if (end_extent_writepage(page, bio->bi_error, start, end))
 			continue;
 
 		end_page_writeback(page);
@@ -2543,10 +2611,10 @@ endio_readpage_release_extent(struct extent_io_tree *tree, u64 start, u64 len,
  * Scheduling is not allowed, so the extent state tree is expected
  * to have one and only one object corresponding to this IO.
  */
-static void end_bio_extent_readpage(struct bio *bio, int err)
+static void end_bio_extent_readpage(struct bio *bio)
 {
 	struct bio_vec *bvec;
-	int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags);
+	int uptodate = !bio->bi_error;
 	struct btrfs_io_bio *io_bio = btrfs_io_bio(bio);
 	struct extent_io_tree *tree;
 	u64 offset = 0;
@@ -2559,16 +2627,13 @@ static void end_bio_extent_readpage(struct bio *bio, int err)
 	int ret;
 	int i;
 
-	if (err)
-		uptodate = 0;
-
 	bio_for_each_segment_all(bvec, bio, i) {
 		struct page *page = bvec->bv_page;
 		struct inode *inode = page->mapping->host;
 
 		pr_debug("end_bio_extent_readpage: bi_sector=%llu, err=%d, "
-			 "mirror=%u\n", (u64)bio->bi_iter.bi_sector, err,
-			 io_bio->mirror_num);
+			 "mirror=%u\n", (u64)bio->bi_iter.bi_sector,
+			 bio->bi_error, io_bio->mirror_num);
 		tree = &BTRFS_I(inode)->io_tree;
 
 		/* We always issue full-page reads, but if some block
@@ -2609,8 +2674,7 @@ static void end_bio_extent_readpage(struct bio *bio, int err)
 
 		if (tree->ops && tree->ops->readpage_io_failed_hook) {
 			ret = tree->ops->readpage_io_failed_hook(page, mirror);
-			if (!ret && !err &&
-			    test_bit(BIO_UPTODATE, &bio->bi_flags))
+			if (!ret && !bio->bi_error)
 				uptodate = 1;
 		} else {
 			/*
@@ -2626,10 +2690,7 @@ static void end_bio_extent_readpage(struct bio *bio, int err)
 			ret = bio_readpage_error(bio, offset, page, start, end,
 						 mirror);
 			if (ret == 0) {
-				uptodate =
-					test_bit(BIO_UPTODATE, &bio->bi_flags);
-				if (err)
-					uptodate = 0;
+				uptodate = !bio->bi_error;
 				offset += len;
 				continue;
 			}
@@ -2679,7 +2740,7 @@ readpage_ok:
 		endio_readpage_release_extent(tree, extent_start, extent_len,
 					      uptodate);
 	if (io_bio->end_io)
-		io_bio->end_io(io_bio, err);
+		io_bio->end_io(io_bio, bio->bi_error);
 	bio_put(bio);
 }
 
@@ -2725,6 +2786,12 @@ struct bio *btrfs_bio_clone(struct bio *bio, gfp_t gfp_mask)
 		btrfs_bio->csum = NULL;
 		btrfs_bio->csum_allocated = NULL;
 		btrfs_bio->end_io = NULL;
+
+#ifdef CONFIG_BLK_CGROUP
+		/* FIXME, put this into bio_clone_bioset */
+		if (bio->bi_css)
+			bio_associate_blkcg(new, bio->bi_css);
+#endif
 	}
 	return new;
 }
@@ -2767,8 +2834,6 @@ static int __must_check submit_one_bio(int rw, struct bio *bio,
 	else
 		btrfsic_submit_bio(rw, bio);
 
-	if (bio_flagged(bio, BIO_EOPNOTSUPP))
-		ret = -EOPNOTSUPP;
 	bio_put(bio);
 	return ret;
 }
@@ -2787,6 +2852,7 @@ static int merge_bio(int rw, struct extent_io_tree *tree, struct page *page,
 }
 
 static int submit_extent_page(int rw, struct extent_io_tree *tree,
+			      struct writeback_control *wbc,
 			      struct page *page, sector_t sector,
 			      size_t size, unsigned long offset,
 			      struct block_device *bdev,
@@ -2795,13 +2861,12 @@ static int submit_extent_page(int rw, struct extent_io_tree *tree,
 			      bio_end_io_t end_io_func,
 			      int mirror_num,
 			      unsigned long prev_bio_flags,
-			      unsigned long bio_flags)
+			      unsigned long bio_flags,
+			      bool force_bio_submit)
 {
 	int ret = 0;
 	struct bio *bio;
-	int nr;
 	int contig = 0;
-	int this_compressed = bio_flags & EXTENT_BIO_COMPRESSED;
 	int old_compressed = prev_bio_flags & EXTENT_BIO_COMPRESSED;
 	size_t page_size = min_t(size_t, size, PAGE_CACHE_SIZE);
 
@@ -2813,6 +2878,7 @@ static int submit_extent_page(int rw, struct extent_io_tree *tree,
 			contig = bio_end_sector(bio) == sector;
 
 		if (prev_bio_flags != bio_flags || !contig ||
+		    force_bio_submit ||
 		    merge_bio(rw, tree, page, offset, page_size, bio, bio_flags) ||
 		    bio_add_page(bio, page, page_size, offset) < page_size) {
 			ret = submit_one_bio(rw, bio, mirror_num,
@@ -2823,21 +2889,24 @@ static int submit_extent_page(int rw, struct extent_io_tree *tree,
 			}
 			bio = NULL;
 		} else {
+			if (wbc)
+				wbc_account_io(wbc, page, page_size);
 			return 0;
 		}
 	}
-	if (this_compressed)
-		nr = BIO_MAX_PAGES;
-	else
-		nr = bio_get_nr_vecs(bdev);
 
-	bio = btrfs_bio_alloc(bdev, sector, nr, GFP_NOFS | __GFP_HIGH);
+	bio = btrfs_bio_alloc(bdev, sector, BIO_MAX_PAGES,
+			GFP_NOFS | __GFP_HIGH);
 	if (!bio)
 		return -ENOMEM;
 
 	bio_add_page(bio, page, page_size, offset);
 	bio->bi_end_io = end_io_func;
 	bio->bi_private = tree;
+	if (wbc) {
+		wbc_init_bio(wbc, bio);
+		wbc_account_io(wbc, page, page_size);
+	}
 
 	if (bio_ret)
 		*bio_ret = bio;
@@ -2906,7 +2975,8 @@ static int __do_readpage(struct extent_io_tree *tree,
 			 get_extent_t *get_extent,
 			 struct extent_map **em_cached,
 			 struct bio **bio, int mirror_num,
-			 unsigned long *bio_flags, int rw)
+			 unsigned long *bio_flags, int rw,
+			 u64 *prev_em_start)
 {
 	struct inode *inode = page->mapping->host;
 	u64 start = page_offset(page);
@@ -2954,6 +3024,7 @@ static int __do_readpage(struct extent_io_tree *tree,
 	}
 	while (cur <= end) {
 		unsigned long pnr = (last_byte >> PAGE_CACHE_SHIFT) + 1;
+		bool force_bio_submit = false;
 
 		if (cur >= last_byte) {
 			char *userpage;
@@ -3004,6 +3075,49 @@ static int __do_readpage(struct extent_io_tree *tree,
 		block_start = em->block_start;
 		if (test_bit(EXTENT_FLAG_PREALLOC, &em->flags))
 			block_start = EXTENT_MAP_HOLE;
+
+		/*
+		 * If we have a file range that points to a compressed extent
+		 * and it's followed by a consecutive file range that points to
+		 * to the same compressed extent (possibly with a different
+		 * offset and/or length, so it either points to the whole extent
+		 * or only part of it), we must make sure we do not submit a
+		 * single bio to populate the pages for the 2 ranges because
+		 * this makes the compressed extent read zero out the pages
+		 * belonging to the 2nd range. Imagine the following scenario:
+		 *
+		 *  File layout
+		 *  [0 - 8K]                     [8K - 24K]
+		 *    |                               |
+		 *    |                               |
+		 * points to extent X,         points to extent X,
+		 * offset 4K, length of 8K     offset 0, length 16K
+		 *
+		 * [extent X, compressed length = 4K uncompressed length = 16K]
+		 *
+		 * If the bio to read the compressed extent covers both ranges,
+		 * it will decompress extent X into the pages belonging to the
+		 * first range and then it will stop, zeroing out the remaining
+		 * pages that belong to the other range that points to extent X.
+		 * So here we make sure we submit 2 bios, one for the first
+		 * range and another one for the third range. Both will target
+		 * the same physical extent from disk, but we can't currently
+		 * make the compressed bio endio callback populate the pages
+		 * for both ranges because each compressed bio is tightly
+		 * coupled with a single extent map, and each range can have
+		 * an extent map with a different offset value relative to the
+		 * uncompressed data of our extent and different lengths. This
+		 * is a corner case so we prioritize correctness over
+		 * non-optimal behavior (submitting 2 bios for the same extent).
+		 */
+		if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags) &&
+		    prev_em_start && *prev_em_start != (u64)-1 &&
+		    *prev_em_start != em->orig_start)
+			force_bio_submit = true;
+
+		if (prev_em_start)
+			*prev_em_start = em->orig_start;
+
 		free_extent_map(em);
 		em = NULL;
 
@@ -3019,8 +3133,12 @@ static int __do_readpage(struct extent_io_tree *tree,
 
 			set_extent_uptodate(tree, cur, cur + iosize - 1,
 					    &cached, GFP_NOFS);
-			unlock_extent_cached(tree, cur, cur + iosize - 1,
-			                     &cached, GFP_NOFS);
+			if (parent_locked)
+				free_extent_state(cached);
+			else
+				unlock_extent_cached(tree, cur,
+						     cur + iosize - 1,
+						     &cached, GFP_NOFS);
 			cur = cur + iosize;
 			pg_offset += iosize;
 			continue;
@@ -3048,12 +3166,13 @@ static int __do_readpage(struct extent_io_tree *tree,
 		}
 
 		pnr -= page->index;
-		ret = submit_extent_page(rw, tree, page,
+		ret = submit_extent_page(rw, tree, NULL, page,
 					 sector, disk_io_size, pg_offset,
 					 bdev, bio, pnr,
 					 end_bio_extent_readpage, mirror_num,
 					 *bio_flags,
-					 this_bio_flag);
+					 this_bio_flag,
+					 force_bio_submit);
 		if (!ret) {
 			nr++;
 			*bio_flags = this_bio_flag;
@@ -3080,7 +3199,8 @@ static inline void __do_contiguous_readpages(struct extent_io_tree *tree,
 					     get_extent_t *get_extent,
 					     struct extent_map **em_cached,
 					     struct bio **bio, int mirror_num,
-					     unsigned long *bio_flags, int rw)
+					     unsigned long *bio_flags, int rw,
+					     u64 *prev_em_start)
 {
 	struct inode *inode;
 	struct btrfs_ordered_extent *ordered;
@@ -3100,7 +3220,7 @@ static inline void __do_contiguous_readpages(struct extent_io_tree *tree,
 
 	for (index = 0; index < nr_pages; index++) {
 		__do_readpage(tree, pages[index], get_extent, em_cached, bio,
-			      mirror_num, bio_flags, rw);
+			      mirror_num, bio_flags, rw, prev_em_start);
 		page_cache_release(pages[index]);
 	}
 }
@@ -3110,7 +3230,8 @@ static void __extent_readpages(struct extent_io_tree *tree,
 			       int nr_pages, get_extent_t *get_extent,
 			       struct extent_map **em_cached,
 			       struct bio **bio, int mirror_num,
-			       unsigned long *bio_flags, int rw)
+			       unsigned long *bio_flags, int rw,
+			       u64 *prev_em_start)
 {
 	u64 start = 0;
 	u64 end = 0;
@@ -3131,7 +3252,7 @@ static void __extent_readpages(struct extent_io_tree *tree,
 						  index - first_index, start,
 						  end, get_extent, em_cached,
 						  bio, mirror_num, bio_flags,
-						  rw);
+						  rw, prev_em_start);
 			start = page_start;
 			end = start + PAGE_CACHE_SIZE - 1;
 			first_index = index;
@@ -3142,7 +3263,8 @@ static void __extent_readpages(struct extent_io_tree *tree,
 		__do_contiguous_readpages(tree, &pages[first_index],
 					  index - first_index, start,
 					  end, get_extent, em_cached, bio,
-					  mirror_num, bio_flags, rw);
+					  mirror_num, bio_flags, rw,
+					  prev_em_start);
 }
 
 static int __extent_read_full_page(struct extent_io_tree *tree,
@@ -3168,7 +3290,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 	}
 
 	ret = __do_readpage(tree, page, get_extent, NULL, bio, mirror_num,
-			    bio_flags, rw);
+			    bio_flags, rw, NULL);
 	return ret;
 }
 
@@ -3194,7 +3316,7 @@ int extent_read_full_page_nolock(struct extent_io_tree *tree, struct page *page,
 	int ret;
 
 	ret = __do_readpage(tree, page, get_extent, NULL, &bio, mirror_num,
-				      &bio_flags, READ);
+			    &bio_flags, READ, NULL);
 	if (bio)
 		ret = submit_one_bio(READ, bio, mirror_num, bio_flags);
 	return ret;
@@ -3443,11 +3565,11 @@ static noinline_for_stack int __extent_writepage_io(struct inode *inode,
 				       page->index, cur, end);
 			}
 
-			ret = submit_extent_page(write_flags, tree, page,
+			ret = submit_extent_page(write_flags, tree, wbc, page,
 						 sector, iosize, pg_offset,
 						 bdev, &epd->bio, max_nr,
 						 end_bio_extent_writepage,
-						 0, 0, 0);
+						 0, 0, 0, false);
 			if (ret)
 				SetPageError(page);
 		}
@@ -3693,7 +3815,7 @@ static void set_btree_ioerr(struct page *page)
 	}
 }
 
-static void end_bio_extent_buffer_writepage(struct bio *bio, int err)
+static void end_bio_extent_buffer_writepage(struct bio *bio)
 {
 	struct bio_vec *bvec;
 	struct extent_buffer *eb;
@@ -3706,7 +3828,8 @@ static void end_bio_extent_buffer_writepage(struct bio *bio, int err)
 		BUG_ON(!eb);
 		done = atomic_dec_and_test(&eb->io_pages);
 
-		if (err || test_bit(EXTENT_BUFFER_WRITE_ERR, &eb->bflags)) {
+		if (bio->bi_error ||
+		    test_bit(EXTENT_BUFFER_WRITE_ERR, &eb->bflags)) {
 			ClearPageUptodate(page);
 			set_btree_ioerr(page);
 		}
@@ -3746,10 +3869,10 @@ static noinline_for_stack int write_one_eb(struct extent_buffer *eb,
 
 		clear_page_dirty_for_io(p);
 		set_page_writeback(p);
-		ret = submit_extent_page(rw, tree, p, offset >> 9,
+		ret = submit_extent_page(rw, tree, wbc, p, offset >> 9,
 					 PAGE_CACHE_SIZE, 0, bdev, &epd->bio,
 					 -1, end_bio_extent_buffer_writepage,
-					 0, epd->bio_flags, bio_flags);
+					 0, epd->bio_flags, bio_flags, false);
 		epd->bio_flags = bio_flags;
 		if (ret) {
 			set_btree_ioerr(p);
@@ -4153,6 +4276,7 @@ int extent_readpages(struct extent_io_tree *tree,
 	struct page *page;
 	struct extent_map *em_cached = NULL;
 	int nr = 0;
+	u64 prev_em_start = (u64)-1;
 
 	for (page_idx = 0; page_idx < nr_pages; page_idx++) {
 		page = list_entry(pages->prev, struct page, lru);
@@ -4169,12 +4293,12 @@ int extent_readpages(struct extent_io_tree *tree,
 		if (nr < ARRAY_SIZE(pagepool))
 			continue;
 		__extent_readpages(tree, pagepool, nr, get_extent, &em_cached,
-				   &bio, 0, &bio_flags, READ);
+				   &bio, 0, &bio_flags, READ, &prev_em_start);
 		nr = 0;
 	}
 	if (nr)
 		__extent_readpages(tree, pagepool, nr, get_extent, &em_cached,
-				   &bio, 0, &bio_flags, READ);
+				   &bio, 0, &bio_flags, READ, &prev_em_start);
 
 	if (em_cached)
 		free_extent_map(em_cached);
@@ -4262,7 +4386,7 @@ int try_release_extent_mapping(struct extent_map_tree *map,
 	u64 start = page_offset(page);
 	u64 end = start + PAGE_CACHE_SIZE - 1;
 
-	if ((mask & __GFP_WAIT) &&
+	if (gfpflags_allow_blocking(mask) &&
 	    page->mapping->host->i_size > 16 * 1024 * 1024) {
 		u64 len;
 		while (start <= end) {
@@ -4492,6 +4616,8 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
 		}
 		if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags))
 			flags |= FIEMAP_EXTENT_ENCODED;
+		if (test_bit(EXTENT_FLAG_PREALLOC, &em->flags))
+			flags |= FIEMAP_EXTENT_UNWRITTEN;
 
 		free_extent_map(em);
 		em = NULL;
@@ -4609,9 +4735,7 @@ __alloc_extent_buffer(struct btrfs_fs_info *fs_info, u64 start,
 {
 	struct extent_buffer *eb = NULL;
 
-	eb = kmem_cache_zalloc(extent_buffer_cache, GFP_NOFS);
-	if (eb == NULL)
-		return NULL;
+	eb = kmem_cache_zalloc(extent_buffer_cache, GFP_NOFS|__GFP_NOFAIL);
 	eb->start = start;
 	eb->len = len;
 	eb->fs_info = fs_info;
@@ -4869,7 +4993,7 @@ struct extent_buffer *alloc_extent_buffer(struct btrfs_fs_info *fs_info,
 		return NULL;
 
 	for (i = 0; i < num_pages; i++, index++) {
-		p = find_or_create_page(mapping, index, GFP_NOFS);
+		p = find_or_create_page(mapping, index, GFP_NOFS|__GFP_NOFAIL);
 		if (!p)
 			goto free_eb;
 
@@ -5509,13 +5633,15 @@ void memcpy_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset,
 	unsigned long src_i;
 
 	if (src_offset + len > dst->len) {
-		printk(KERN_ERR "BTRFS: memmove bogus src_offset %lu move "
-		       "len %lu dst len %lu\n", src_offset, len, dst->len);
+		btrfs_err(dst->fs_info,
+			"memmove bogus src_offset %lu move "
+		       "len %lu dst len %lu", src_offset, len, dst->len);
 		BUG_ON(1);
 	}
 	if (dst_offset + len > dst->len) {
-		printk(KERN_ERR "BTRFS: memmove bogus dst_offset %lu move "
-		       "len %lu dst len %lu\n", dst_offset, len, dst->len);
+		btrfs_err(dst->fs_info,
+			"memmove bogus dst_offset %lu move "
+		       "len %lu dst len %lu", dst_offset, len, dst->len);
 		BUG_ON(1);
 	}
 
@@ -5555,13 +5681,13 @@ void memmove_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset,
 	unsigned long src_i;
 
 	if (src_offset + len > dst->len) {
-		printk(KERN_ERR "BTRFS: memmove bogus src_offset %lu move "
-		       "len %lu len %lu\n", src_offset, len, dst->len);
+		btrfs_err(dst->fs_info, "memmove bogus src_offset %lu move "
+		       "len %lu len %lu", src_offset, len, dst->len);
 		BUG_ON(1);
 	}
 	if (dst_offset + len > dst->len) {
-		printk(KERN_ERR "BTRFS: memmove bogus dst_offset %lu move "
-		       "len %lu len %lu\n", dst_offset, len, dst->len);
+		btrfs_err(dst->fs_info, "memmove bogus dst_offset %lu move "
+		       "len %lu len %lu", dst_offset, len, dst->len);
 		BUG_ON(1);
 	}
 	if (dst_offset < src_offset) {
diff --git a/kernel/fs/btrfs/extent_io.h b/kernel/fs/btrfs/extent_io.h
index c668f3689..f4c1ae118 100644
--- a/kernel/fs/btrfs/extent_io.h
+++ b/kernel/fs/btrfs/extent_io.h
@@ -2,6 +2,7 @@
 #define __EXTENTIO__
 
 #include <linux/rbtree.h>
+#include "ulist.h"
 
 /* bits for the extent state */
 #define EXTENT_DIRTY		(1U << 0)
@@ -18,6 +19,7 @@
 #define EXTENT_NEED_WAIT	(1U << 13)
 #define EXTENT_DAMAGED		(1U << 14)
 #define EXTENT_NORESERVE	(1U << 15)
+#define EXTENT_QGROUP_RESERVED	(1U << 16)
 #define EXTENT_IOBITS		(EXTENT_LOCKED | EXTENT_WRITEBACK)
 #define EXTENT_CTLBITS		(EXTENT_DO_ACCOUNTING | EXTENT_FIRST_DELALLOC)
 
@@ -161,6 +163,17 @@ struct extent_buffer {
 #endif
 };
 
+/*
+ * Structure to record how many bytes and which ranges are set/cleared
+ */
+struct extent_changeset {
+	/* How many bytes are set/cleared in this operation */
+	u64 bytes_changed;
+
+	/* Changed ranges */
+	struct ulist *range_changed;
+};
+
 static inline void extent_set_compress_type(unsigned long *bio_flags,
 					    int compress_type)
 {
@@ -210,11 +223,17 @@ int test_range_bit(struct extent_io_tree *tree, u64 start, u64 end,
 		   struct extent_state *cached_state);
 int clear_extent_bits(struct extent_io_tree *tree, u64 start, u64 end,
 		      unsigned bits, gfp_t mask);
+int clear_record_extent_bits(struct extent_io_tree *tree, u64 start, u64 end,
+			     unsigned bits, gfp_t mask,
+			     struct extent_changeset *changeset);
 int clear_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
 		     unsigned bits, int wake, int delete,
 		     struct extent_state **cached, gfp_t mask);
 int set_extent_bits(struct extent_io_tree *tree, u64 start, u64 end,
 		    unsigned bits, gfp_t mask);
+int set_record_extent_bits(struct extent_io_tree *tree, u64 start, u64 end,
+			   unsigned bits, gfp_t mask,
+			   struct extent_changeset *changeset);
 int set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
 		   unsigned bits, u64 *failed_start,
 		   struct extent_state **cached_state, gfp_t mask);
diff --git a/kernel/fs/btrfs/file.c b/kernel/fs/btrfs/file.c
index b072e1747..0f09526aa 100644
--- a/kernel/fs/btrfs/file.c
+++ b/kernel/fs/btrfs/file.c
@@ -756,8 +756,16 @@ next_slot:
 		}
 
 		btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
-		if (key.objectid > ino ||
-		    key.type > BTRFS_EXTENT_DATA_KEY || key.offset >= end)
+
+		if (key.objectid > ino)
+			break;
+		if (WARN_ON_ONCE(key.objectid < ino) ||
+		    key.type < BTRFS_EXTENT_DATA_KEY) {
+			ASSERT(del_nr == 0);
+			path->slots[0]++;
+			goto next_slot;
+		}
+		if (key.type > BTRFS_EXTENT_DATA_KEY || key.offset >= end)
 			break;
 
 		fi = btrfs_item_ptr(leaf, path->slots[0],
@@ -776,8 +784,8 @@ next_slot:
 				btrfs_file_extent_inline_len(leaf,
 						     path->slots[0], fi);
 		} else {
-			WARN_ON(1);
-			extent_end = search_start;
+			/* can't happen */
+			BUG();
 		}
 
 		/*
@@ -847,7 +855,7 @@ next_slot:
 						disk_bytenr, num_bytes, 0,
 						root->root_key.objectid,
 						new_key.objectid,
-						start - extent_offset, 1);
+						start - extent_offset);
 				BUG_ON(ret); /* -ENOMEM */
 			}
 			key.offset = start;
@@ -925,7 +933,7 @@ delete_extent_item:
 						disk_bytenr, num_bytes, 0,
 						root->root_key.objectid,
 						key.objectid, key.offset -
-						extent_offset, 0);
+						extent_offset);
 				BUG_ON(ret); /* -ENOMEM */
 				inode_sub_bytes(inode,
 						extent_end - key.offset);
@@ -1204,7 +1212,7 @@ again:
 
 		ret = btrfs_inc_extent_ref(trans, root, bytenr, num_bytes, 0,
 					   root->root_key.objectid,
-					   ino, orig_offset, 1);
+					   ino, orig_offset);
 		BUG_ON(ret); /* -ENOMEM */
 
 		if (split == start) {
@@ -1231,7 +1239,7 @@ again:
 		del_nr++;
 		ret = btrfs_free_extent(trans, root, bytenr, num_bytes,
 					0, root->root_key.objectid,
-					ino, orig_offset, 0);
+					ino, orig_offset);
 		BUG_ON(ret); /* -ENOMEM */
 	}
 	other_start = 0;
@@ -1248,7 +1256,7 @@ again:
 		del_nr++;
 		ret = btrfs_free_extent(trans, root, bytenr, num_bytes,
 					0, root->root_key.objectid,
-					ino, orig_offset, 0);
+					ino, orig_offset);
 		BUG_ON(ret); /* -ENOMEM */
 	}
 	if (del_nr == 0) {
@@ -1283,7 +1291,8 @@ out:
  * on error we return an unlocked page and the error value
  * on success we return a locked page and 0
  */
-static int prepare_uptodate_page(struct page *page, u64 pos,
+static int prepare_uptodate_page(struct inode *inode,
+				 struct page *page, u64 pos,
 				 bool force_uptodate)
 {
 	int ret = 0;
@@ -1298,6 +1307,10 @@ static int prepare_uptodate_page(struct page *page, u64 pos,
 			unlock_page(page);
 			return -EIO;
 		}
+		if (page->mapping != inode->i_mapping) {
+			unlock_page(page);
+			return -EAGAIN;
+		}
 	}
 	return 0;
 }
@@ -1316,6 +1329,7 @@ static noinline int prepare_pages(struct inode *inode, struct page **pages,
 	int faili;
 
 	for (i = 0; i < num_pages; i++) {
+again:
 		pages[i] = find_or_create_page(inode->i_mapping, index + i,
 					       mask | __GFP_WRITE);
 		if (!pages[i]) {
@@ -1325,13 +1339,17 @@ static noinline int prepare_pages(struct inode *inode, struct page **pages,
 		}
 
 		if (i == 0)
-			err = prepare_uptodate_page(pages[i], pos,
+			err = prepare_uptodate_page(inode, pages[i], pos,
 						    force_uptodate);
-		if (i == num_pages - 1)
-			err = prepare_uptodate_page(pages[i],
+		if (!err && i == num_pages - 1)
+			err = prepare_uptodate_page(inode, pages[i],
 						    pos + write_bytes, false);
 		if (err) {
 			page_cache_release(pages[i]);
+			if (err == -EAGAIN) {
+				err = 0;
+				goto again;
+			}
 			faili = i - 1;
 			goto fail;
 		}
@@ -1469,7 +1487,6 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file,
 	u64 release_bytes = 0;
 	u64 lockstart;
 	u64 lockend;
-	unsigned long first_index;
 	size_t num_written = 0;
 	int nrptrs;
 	int ret = 0;
@@ -1485,8 +1502,6 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file,
 	if (!pages)
 		return -ENOMEM;
 
-	first_index = pos >> PAGE_CACHE_SHIFT;
-
 	while (iov_iter_count(i) > 0) {
 		size_t offset = pos & (PAGE_CACHE_SIZE - 1);
 		size_t write_bytes = min(iov_iter_count(i),
@@ -1510,12 +1525,17 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file,
 		}
 
 		reserve_bytes = num_pages << PAGE_CACHE_SHIFT;
-		ret = btrfs_check_data_free_space(inode, reserve_bytes, write_bytes);
-		if (ret == -ENOSPC &&
-		    (BTRFS_I(inode)->flags & (BTRFS_INODE_NODATACOW |
-					      BTRFS_INODE_PREALLOC))) {
+
+		if (BTRFS_I(inode)->flags & (BTRFS_INODE_NODATACOW |
+					     BTRFS_INODE_PREALLOC)) {
 			ret = check_can_nocow(inode, pos, &write_bytes);
+			if (ret < 0)
+				break;
 			if (ret > 0) {
+				/*
+				 * For nodata cow case, no need to reserve
+				 * data space.
+				 */
 				only_release_metadata = true;
 				/*
 				 * our prealloc extent may be smaller than
@@ -1524,20 +1544,19 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file,
 				num_pages = DIV_ROUND_UP(write_bytes + offset,
 							 PAGE_CACHE_SIZE);
 				reserve_bytes = num_pages << PAGE_CACHE_SHIFT;
-				ret = 0;
-			} else {
-				ret = -ENOSPC;
+				goto reserve_metadata;
 			}
 		}
-
-		if (ret)
+		ret = btrfs_check_data_free_space(inode, pos, write_bytes);
+		if (ret < 0)
 			break;
 
+reserve_metadata:
 		ret = btrfs_delalloc_reserve_metadata(inode, reserve_bytes);
 		if (ret) {
 			if (!only_release_metadata)
-				btrfs_free_reserved_data_space(inode,
-							       reserve_bytes);
+				btrfs_free_reserved_data_space(inode, pos,
+							       write_bytes);
 			else
 				btrfs_end_write_no_snapshoting(root);
 			break;
@@ -1603,12 +1622,17 @@ again:
 				BTRFS_I(inode)->outstanding_extents++;
 				spin_unlock(&BTRFS_I(inode)->lock);
 			}
-			if (only_release_metadata)
+			if (only_release_metadata) {
 				btrfs_delalloc_release_metadata(inode,
 								release_bytes);
-			else
-				btrfs_delalloc_release_space(inode,
+			} else {
+				u64 __pos;
+
+				__pos = round_down(pos, root->sectorsize) +
+					(dirty_pages << PAGE_CACHE_SHIFT);
+				btrfs_delalloc_release_space(inode, __pos,
 							     release_bytes);
+			}
 		}
 
 		release_bytes = dirty_pages << PAGE_CACHE_SHIFT;
@@ -1660,7 +1684,7 @@ again:
 			btrfs_end_write_no_snapshoting(root);
 			btrfs_delalloc_release_metadata(inode, release_bytes);
 		} else {
-			btrfs_delalloc_release_space(inode, release_bytes);
+			btrfs_delalloc_release_space(inode, pos, release_bytes);
 		}
 	}
 
@@ -1748,7 +1772,7 @@ static ssize_t btrfs_file_write_iter(struct kiocb *iocb,
 	}
 
 	current->backing_dev_info = inode_to_bdi(inode);
-	err = file_remove_suid(file);
+	err = file_remove_privs(file);
 	if (err) {
 		mutex_unlock(&inode->i_mutex);
 		goto out;
@@ -1868,7 +1892,13 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 	struct btrfs_log_ctx ctx;
 	int ret = 0;
 	bool full_sync = 0;
+	u64 len;
 
+	/*
+	 * The range length can be represented by u64, we have to do the typecasts
+	 * to avoid signed overflow if it's [0, LLONG_MAX] eg. from fsync()
+	 */
+	len = (u64)end - (u64)start + 1;
 	trace_btrfs_sync_file(file, datasync);
 
 	/*
@@ -1896,7 +1926,7 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 		 * all extents are persisted and the respective file extent
 		 * items are in the fs/subvol btree.
 		 */
-		ret = btrfs_wait_ordered_range(inode, start, end - start + 1);
+		ret = btrfs_wait_ordered_range(inode, start, len);
 	} else {
 		/*
 		 * Start any new ordered operations before starting to log the
@@ -1968,8 +1998,10 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 	 */
 	smp_mb();
 	if (btrfs_inode_in_log(inode, root->fs_info->generation) ||
-	    (full_sync && BTRFS_I(inode)->last_trans <=
-	     root->fs_info->last_trans_committed)) {
+	    (BTRFS_I(inode)->last_trans <=
+	     root->fs_info->last_trans_committed &&
+	     (full_sync ||
+	      !btrfs_have_ordered_extents_in_range(inode, start, len)))) {
 		/*
 		 * We'v had everything committed since the last time we were
 		 * modified so clear this flag in case it was set for whatever
@@ -2054,8 +2086,7 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 			}
 		}
 		if (!full_sync) {
-			ret = btrfs_wait_ordered_range(inode, start,
-						       end - start + 1);
+			ret = btrfs_wait_ordered_range(inode, start, len);
 			if (ret) {
 				btrfs_end_transaction(trans, root);
 				goto out;
@@ -2263,7 +2294,7 @@ static int btrfs_punch_hole(struct inode *inode, loff_t offset, loff_t len)
 	u64 drop_end;
 	int ret = 0;
 	int err = 0;
-	int rsv_count;
+	unsigned int rsv_count;
 	bool same_page;
 	bool no_holes = btrfs_fs_incompat(root->fs_info, NO_HOLES);
 	u64 ino_size;
@@ -2485,6 +2516,19 @@ static int btrfs_punch_hole(struct inode *inode, loff_t offset, loff_t len)
 
 	trans->block_rsv = &root->fs_info->trans_block_rsv;
 	/*
+	 * If we are using the NO_HOLES feature we might have had already an
+	 * hole that overlaps a part of the region [lockstart, lockend] and
+	 * ends at (or beyond) lockend. Since we have no file extent items to
+	 * represent holes, drop_end can be less than lockend and so we must
+	 * make sure we have an extent map representing the existing hole (the
+	 * call to __btrfs_drop_extents() might have dropped the existing extent
+	 * map representing the existing hole), otherwise the fast fsync path
+	 * will not record the existence of the hole region
+	 * [existing_hole_start, lockend].
+	 */
+	if (drop_end <= lockend)
+		drop_end = lockend + 1;
+	/*
 	 * Don't insert file hole extent item if it's for a range beyond eof
 	 * (because it's useless) or if it represents a 0 bytes range (when
 	 * cur_offset == drop_end).
@@ -2538,17 +2582,61 @@ out_only_mutex:
 	return err;
 }
 
+/* Helper structure to record which range is already reserved */
+struct falloc_range {
+	struct list_head list;
+	u64 start;
+	u64 len;
+};
+
+/*
+ * Helper function to add falloc range
+ *
+ * Caller should have locked the larger range of extent containing
+ * [start, len)
+ */
+static int add_falloc_range(struct list_head *head, u64 start, u64 len)
+{
+	struct falloc_range *prev = NULL;
+	struct falloc_range *range = NULL;
+
+	if (list_empty(head))
+		goto insert;
+
+	/*
+	 * As fallocate iterate by bytenr order, we only need to check
+	 * the last range.
+	 */
+	prev = list_entry(head->prev, struct falloc_range, list);
+	if (prev->start + prev->len == start) {
+		prev->len += len;
+		return 0;
+	}
+insert:
+	range = kmalloc(sizeof(*range), GFP_NOFS);
+	if (!range)
+		return -ENOMEM;
+	range->start = start;
+	range->len = len;
+	list_add_tail(&range->list, head);
+	return 0;
+}
+
 static long btrfs_fallocate(struct file *file, int mode,
 			    loff_t offset, loff_t len)
 {
 	struct inode *inode = file_inode(file);
 	struct extent_state *cached_state = NULL;
+	struct falloc_range *range;
+	struct falloc_range *tmp;
+	struct list_head reserve_list;
 	u64 cur_offset;
 	u64 last_byte;
 	u64 alloc_start;
 	u64 alloc_end;
 	u64 alloc_hint = 0;
 	u64 locked_end;
+	u64 actual_end = 0;
 	struct extent_map *em;
 	int blocksize = BTRFS_I(inode)->root->sectorsize;
 	int ret;
@@ -2564,11 +2652,12 @@ static long btrfs_fallocate(struct file *file, int mode,
 		return btrfs_punch_hole(inode, offset, len);
 
 	/*
-	 * Make sure we have enough space before we do the
-	 * allocation.
+	 * Only trigger disk allocation, don't trigger qgroup reserve
+	 *
+	 * For qgroup space, it will be checked later.
 	 */
-	ret = btrfs_check_data_free_space(inode, alloc_end - alloc_start, alloc_end - alloc_start);
-	if (ret)
+	ret = btrfs_alloc_data_chunk_ondemand(inode, alloc_end - alloc_start);
+	if (ret < 0)
 		return ret;
 
 	mutex_lock(&inode->i_mutex);
@@ -2576,12 +2665,19 @@ static long btrfs_fallocate(struct file *file, int mode,
 	if (ret)
 		goto out;
 
+	/*
+	 * TODO: Move these two operations after we have checked
+	 * accurate reserved space, or fallocate can still fail but
+	 * with page truncated or size expanded.
+	 *
+	 * But that's a minor problem and won't do much harm BTW.
+	 */
 	if (alloc_start > inode->i_size) {
 		ret = btrfs_cont_expand(inode, i_size_read(inode),
 					alloc_start);
 		if (ret)
 			goto out;
-	} else {
+	} else if (offset + len > inode->i_size) {
 		/*
 		 * If we are fallocating from the end of the file onward we
 		 * need to zero out the end of the page if i_size lands in the
@@ -2634,10 +2730,10 @@ static long btrfs_fallocate(struct file *file, int mode,
 		}
 	}
 
+	/* First, check if we exceed the qgroup limit */
+	INIT_LIST_HEAD(&reserve_list);
 	cur_offset = alloc_start;
 	while (1) {
-		u64 actual_end;
-
 		em = btrfs_get_extent(inode, NULL, 0, cur_offset,
 				      alloc_end - cur_offset, 0);
 		if (IS_ERR_OR_NULL(em)) {
@@ -2650,57 +2746,82 @@ static long btrfs_fallocate(struct file *file, int mode,
 		last_byte = min(extent_map_end(em), alloc_end);
 		actual_end = min_t(u64, extent_map_end(em), offset + len);
 		last_byte = ALIGN(last_byte, blocksize);
-
 		if (em->block_start == EXTENT_MAP_HOLE ||
 		    (cur_offset >= inode->i_size &&
 		     !test_bit(EXTENT_FLAG_PREALLOC, &em->flags))) {
-			ret = btrfs_prealloc_file_range(inode, mode, cur_offset,
-							last_byte - cur_offset,
-							1 << inode->i_blkbits,
-							offset + len,
-							&alloc_hint);
-		} else if (actual_end > inode->i_size &&
-			   !(mode & FALLOC_FL_KEEP_SIZE)) {
-			struct btrfs_trans_handle *trans;
-			struct btrfs_root *root = BTRFS_I(inode)->root;
-
-			/*
-			 * We didn't need to allocate any more space, but we
-			 * still extended the size of the file so we need to
-			 * update i_size and the inode item.
-			 */
-			trans = btrfs_start_transaction(root, 1);
-			if (IS_ERR(trans)) {
-				ret = PTR_ERR(trans);
-			} else {
-				inode->i_ctime = CURRENT_TIME;
-				i_size_write(inode, actual_end);
-				btrfs_ordered_update_i_size(inode, actual_end,
-							    NULL);
-				ret = btrfs_update_inode(trans, root, inode);
-				if (ret)
-					btrfs_end_transaction(trans, root);
-				else
-					ret = btrfs_end_transaction(trans,
-								    root);
+			ret = add_falloc_range(&reserve_list, cur_offset,
+					       last_byte - cur_offset);
+			if (ret < 0) {
+				free_extent_map(em);
+				break;
 			}
+			ret = btrfs_qgroup_reserve_data(inode, cur_offset,
+					last_byte - cur_offset);
+			if (ret < 0)
+				break;
 		}
 		free_extent_map(em);
-		if (ret < 0)
-			break;
-
 		cur_offset = last_byte;
-		if (cur_offset >= alloc_end) {
-			ret = 0;
+		if (cur_offset >= alloc_end)
 			break;
+	}
+
+	/*
+	 * If ret is still 0, means we're OK to fallocate.
+	 * Or just cleanup the list and exit.
+	 */
+	list_for_each_entry_safe(range, tmp, &reserve_list, list) {
+		if (!ret)
+			ret = btrfs_prealloc_file_range(inode, mode,
+					range->start,
+					range->len, 1 << inode->i_blkbits,
+					offset + len, &alloc_hint);
+		list_del(&range->list);
+		kfree(range);
+	}
+	if (ret < 0)
+		goto out_unlock;
+
+	if (actual_end > inode->i_size &&
+	    !(mode & FALLOC_FL_KEEP_SIZE)) {
+		struct btrfs_trans_handle *trans;
+		struct btrfs_root *root = BTRFS_I(inode)->root;
+
+		/*
+		 * We didn't need to allocate any more space, but we
+		 * still extended the size of the file so we need to
+		 * update i_size and the inode item.
+		 */
+		trans = btrfs_start_transaction(root, 1);
+		if (IS_ERR(trans)) {
+			ret = PTR_ERR(trans);
+		} else {
+			inode->i_ctime = CURRENT_TIME;
+			i_size_write(inode, actual_end);
+			btrfs_ordered_update_i_size(inode, actual_end, NULL);
+			ret = btrfs_update_inode(trans, root, inode);
+			if (ret)
+				btrfs_end_transaction(trans, root);
+			else
+				ret = btrfs_end_transaction(trans, root);
 		}
 	}
+out_unlock:
 	unlock_extent_cached(&BTRFS_I(inode)->io_tree, alloc_start, locked_end,
 			     &cached_state, GFP_NOFS);
 out:
+	/*
+	 * As we waited the extent range, the data_rsv_map must be empty
+	 * in the range, as written data range will be released from it.
+	 * And for prealloacted extent, it will also be released when
+	 * its metadata is written.
+	 * So this is completely used as cleanup.
+	 */
+	btrfs_qgroup_free_data(inode, alloc_start, alloc_end - alloc_start);
 	mutex_unlock(&inode->i_mutex);
 	/* Let go of our reservation. */
-	btrfs_free_reserved_data_space(inode, alloc_end - alloc_start);
+	btrfs_free_reserved_data_space(inode, alloc_start,
+				       alloc_end - alloc_start);
 	return ret;
 }
 
diff --git a/kernel/fs/btrfs/free-space-cache.c b/kernel/fs/btrfs/free-space-cache.c
index 9dbe5b548..cfe99bec4 100644
--- a/kernel/fs/btrfs/free-space-cache.c
+++ b/kernel/fs/btrfs/free-space-cache.c
@@ -85,8 +85,8 @@ static struct inode *__lookup_free_space_inode(struct btrfs_root *root,
 	}
 
 	mapping_set_gfp_mask(inode->i_mapping,
-			mapping_gfp_mask(inode->i_mapping) &
-			~(__GFP_FS | __GFP_HIGHMEM));
+			mapping_gfp_constraint(inode->i_mapping,
+			~(__GFP_FS | __GFP_HIGHMEM)));
 
 	return inode;
 }
@@ -231,6 +231,7 @@ int btrfs_truncate_free_space_cache(struct btrfs_root *root,
 {
 	int ret = 0;
 	struct btrfs_path *path = btrfs_alloc_path();
+	bool locked = false;
 
 	if (!path) {
 		ret = -ENOMEM;
@@ -238,6 +239,7 @@ int btrfs_truncate_free_space_cache(struct btrfs_root *root,
 	}
 
 	if (block_group) {
+		locked = true;
 		mutex_lock(&trans->transaction->cache_write_mutex);
 		if (!list_empty(&block_group->io_list)) {
 			list_del_init(&block_group->io_list);
@@ -269,18 +271,14 @@ int btrfs_truncate_free_space_cache(struct btrfs_root *root,
 	 */
 	ret = btrfs_truncate_inode_items(trans, root, inode,
 					 0, BTRFS_EXTENT_DATA_KEY);
-	if (ret) {
-		mutex_unlock(&trans->transaction->cache_write_mutex);
-		btrfs_abort_transaction(trans, root, ret);
-		return ret;
-	}
+	if (ret)
+		goto fail;
 
 	ret = btrfs_update_inode(trans, root, inode);
 
-	if (block_group)
-		mutex_unlock(&trans->transaction->cache_write_mutex);
-
 fail:
+	if (locked)
+		mutex_unlock(&trans->transaction->cache_write_mutex);
 	if (ret)
 		btrfs_abort_transaction(trans, root, ret);
 
@@ -452,9 +450,9 @@ static int io_ctl_check_generation(struct btrfs_io_ctl *io_ctl, u64 generation)
 
 	gen = io_ctl->cur;
 	if (le64_to_cpu(*gen) != generation) {
-		printk_ratelimited(KERN_ERR "BTRFS: space cache generation "
-				   "(%Lu) does not match inode (%Lu)\n", *gen,
-				   generation);
+		btrfs_err_rl(io_ctl->root->fs_info,
+			"space cache generation (%llu) does not match inode (%llu)",
+				*gen, generation);
 		io_ctl_unmap_page(io_ctl);
 		return -EIO;
 	}
@@ -508,8 +506,8 @@ static int io_ctl_check_crc(struct btrfs_io_ctl *io_ctl, int index)
 			      PAGE_CACHE_SIZE - offset);
 	btrfs_csum_final(crc, (char *)&crc);
 	if (val != crc) {
-		printk_ratelimited(KERN_ERR "BTRFS: csum mismatch on free "
-				   "space cache\n");
+		btrfs_err_rl(io_ctl->root->fs_info,
+			"csum mismatch on free space cache");
 		io_ctl_unmap_page(io_ctl);
 		return -EIO;
 	}
@@ -893,7 +891,7 @@ out:
 		spin_unlock(&block_group->lock);
 		ret = 0;
 
-		btrfs_warn(fs_info, "failed to load free space cache for block group %llu, rebuild it now",
+		btrfs_warn(fs_info, "failed to load free space cache for block group %llu, rebuilding it now",
 			block_group->key.objectid);
 	}
 
@@ -1217,7 +1215,7 @@ out:
  * @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,
+ * on mount.  This will return 0 if it was successful in writing the cache out,
  * or an errno if it was not.
  */
 static int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
@@ -1732,7 +1730,7 @@ static void bitmap_set_bits(struct btrfs_free_space_ctl *ctl,
  */
 static int search_bitmap(struct btrfs_free_space_ctl *ctl,
 			 struct btrfs_free_space *bitmap_info, u64 *offset,
-			 u64 *bytes)
+			 u64 *bytes, bool for_alloc)
 {
 	unsigned long found_bits = 0;
 	unsigned long max_bits = 0;
@@ -1740,11 +1738,26 @@ static int search_bitmap(struct btrfs_free_space_ctl *ctl,
 	unsigned long next_zero;
 	unsigned long extent_bits;
 
+	/*
+	 * Skip searching the bitmap if we don't have a contiguous section that
+	 * is large enough for this allocation.
+	 */
+	if (for_alloc &&
+	    bitmap_info->max_extent_size &&
+	    bitmap_info->max_extent_size < *bytes) {
+		*bytes = bitmap_info->max_extent_size;
+		return -1;
+	}
+
 	i = offset_to_bit(bitmap_info->offset, ctl->unit,
 			  max_t(u64, *offset, bitmap_info->offset));
 	bits = bytes_to_bits(*bytes, ctl->unit);
 
 	for_each_set_bit_from(i, bitmap_info->bitmap, BITS_PER_BITMAP) {
+		if (for_alloc && bits == 1) {
+			found_bits = 1;
+			break;
+		}
 		next_zero = find_next_zero_bit(bitmap_info->bitmap,
 					       BITS_PER_BITMAP, i);
 		extent_bits = next_zero - i;
@@ -1764,6 +1777,7 @@ static int search_bitmap(struct btrfs_free_space_ctl *ctl,
 	}
 
 	*bytes = (u64)(max_bits) * ctl->unit;
+	bitmap_info->max_extent_size = *bytes;
 	return -1;
 }
 
@@ -1815,7 +1829,7 @@ find_free_space(struct btrfs_free_space_ctl *ctl, u64 *offset, u64 *bytes,
 		if (entry->bitmap) {
 			u64 size = *bytes;
 
-			ret = search_bitmap(ctl, entry, &tmp, &size);
+			ret = search_bitmap(ctl, entry, &tmp, &size, true);
 			if (!ret) {
 				*offset = tmp;
 				*bytes = size;
@@ -1876,7 +1890,8 @@ again:
 	search_start = *offset;
 	search_bytes = ctl->unit;
 	search_bytes = min(search_bytes, end - search_start + 1);
-	ret = search_bitmap(ctl, bitmap_info, &search_start, &search_bytes);
+	ret = search_bitmap(ctl, bitmap_info, &search_start, &search_bytes,
+			    false);
 	if (ret < 0 || search_start != *offset)
 		return -EINVAL;
 
@@ -1921,7 +1936,7 @@ again:
 		search_start = *offset;
 		search_bytes = ctl->unit;
 		ret = search_bitmap(ctl, bitmap_info, &search_start,
-				    &search_bytes);
+				    &search_bytes, false);
 		if (ret < 0 || search_start != *offset)
 			return -EAGAIN;
 
@@ -1945,6 +1960,12 @@ static u64 add_bytes_to_bitmap(struct btrfs_free_space_ctl *ctl,
 
 	bitmap_set_bits(ctl, info, offset, bytes_to_set);
 
+	/*
+	 * We set some bytes, we have no idea what the max extent size is
+	 * anymore.
+	 */
+	info->max_extent_size = 0;
+
 	return bytes_to_set;
 
 }
@@ -1953,12 +1974,19 @@ static bool use_bitmap(struct btrfs_free_space_ctl *ctl,
 		      struct btrfs_free_space *info)
 {
 	struct btrfs_block_group_cache *block_group = ctl->private;
+	bool forced = false;
+
+#ifdef CONFIG_BTRFS_DEBUG
+	if (btrfs_should_fragment_free_space(block_group->fs_info->extent_root,
+					     block_group))
+		forced = true;
+#endif
 
 	/*
 	 * If we are below the extents threshold then we can add this as an
 	 * extent, and don't have to deal with the bitmap
 	 */
-	if (ctl->free_extents < ctl->extents_thresh) {
+	if (!forced && ctl->free_extents < ctl->extents_thresh) {
 		/*
 		 * If this block group has some small extents we don't want to
 		 * use up all of our free slots in the cache with them, we want
@@ -2663,7 +2691,7 @@ static u64 btrfs_alloc_from_bitmap(struct btrfs_block_group_cache *block_group,
 	search_start = min_start;
 	search_bytes = bytes;
 
-	err = search_bitmap(ctl, entry, &search_start, &search_bytes);
+	err = search_bitmap(ctl, entry, &search_start, &search_bytes, true);
 	if (err) {
 		if (search_bytes > *max_extent_size)
 			*max_extent_size = search_bytes;
@@ -2777,6 +2805,7 @@ static int btrfs_bitmap_cluster(struct btrfs_block_group_cache *block_group,
 	unsigned long want_bits;
 	unsigned long min_bits;
 	unsigned long found_bits;
+	unsigned long max_bits = 0;
 	unsigned long start = 0;
 	unsigned long total_found = 0;
 	int ret;
@@ -2786,6 +2815,13 @@ static int btrfs_bitmap_cluster(struct btrfs_block_group_cache *block_group,
 	want_bits = bytes_to_bits(bytes, ctl->unit);
 	min_bits = bytes_to_bits(min_bytes, ctl->unit);
 
+	/*
+	 * Don't bother looking for a cluster in this bitmap if it's heavily
+	 * fragmented.
+	 */
+	if (entry->max_extent_size &&
+	    entry->max_extent_size < cont1_bytes)
+		return -ENOSPC;
 again:
 	found_bits = 0;
 	for_each_set_bit_from(i, entry->bitmap, BITS_PER_BITMAP) {
@@ -2793,13 +2829,19 @@ again:
 					       BITS_PER_BITMAP, i);
 		if (next_zero - i >= min_bits) {
 			found_bits = next_zero - i;
+			if (found_bits > max_bits)
+				max_bits = found_bits;
 			break;
 		}
+		if (next_zero - i > max_bits)
+			max_bits = next_zero - i;
 		i = next_zero;
 	}
 
-	if (!found_bits)
+	if (!found_bits) {
+		entry->max_extent_size = (u64)max_bits * ctl->unit;
 		return -ENOSPC;
+	}
 
 	if (!total_found) {
 		start = i;
@@ -2930,7 +2972,7 @@ setup_cluster_bitmap(struct btrfs_block_group_cache *block_group,
 		     u64 cont1_bytes, u64 min_bytes)
 {
 	struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
-	struct btrfs_free_space *entry;
+	struct btrfs_free_space *entry = NULL;
 	int ret = -ENOSPC;
 	u64 bitmap_offset = offset_to_bitmap(ctl, offset);
 
@@ -2941,8 +2983,10 @@ setup_cluster_bitmap(struct btrfs_block_group_cache *block_group,
 	 * The bitmap that covers offset won't be in the list unless offset
 	 * is just its start offset.
 	 */
-	entry = list_first_entry(bitmaps, struct btrfs_free_space, list);
-	if (entry->offset != bitmap_offset) {
+	if (!list_empty(bitmaps))
+		entry = list_first_entry(bitmaps, struct btrfs_free_space, list);
+
+	if (!entry || entry->offset != bitmap_offset) {
 		entry = tree_search_offset(ctl, bitmap_offset, 1, 0);
 		if (entry && list_empty(&entry->list))
 			list_add(&entry->list, bitmaps);
@@ -3058,6 +3102,7 @@ void btrfs_init_free_cluster(struct btrfs_free_cluster *cluster)
 	spin_lock_init(&cluster->refill_lock);
 	cluster->root = RB_ROOT;
 	cluster->max_size = 0;
+	cluster->fragmented = false;
 	INIT_LIST_HEAD(&cluster->block_group_list);
 	cluster->block_group = NULL;
 }
@@ -3225,7 +3270,7 @@ static int trim_bitmaps(struct btrfs_block_group_cache *block_group,
 		}
 
 		bytes = minlen;
-		ret2 = search_bitmap(ctl, entry, &start, &bytes);
+		ret2 = search_bitmap(ctl, entry, &start, &bytes, false);
 		if (ret2 || start >= end) {
 			spin_unlock(&ctl->tree_lock);
 			mutex_unlock(&ctl->cache_writeout_mutex);
@@ -3274,35 +3319,23 @@ next:
 	return ret;
 }
 
-int btrfs_trim_block_group(struct btrfs_block_group_cache *block_group,
-			   u64 *trimmed, u64 start, u64 end, u64 minlen)
+void btrfs_get_block_group_trimming(struct btrfs_block_group_cache *cache)
 {
-	int ret;
+	atomic_inc(&cache->trimming);
+}
 
-	*trimmed = 0;
+void btrfs_put_block_group_trimming(struct btrfs_block_group_cache *block_group)
+{
+	struct extent_map_tree *em_tree;
+	struct extent_map *em;
+	bool cleanup;
 
 	spin_lock(&block_group->lock);
-	if (block_group->removed) {
-		spin_unlock(&block_group->lock);
-		return 0;
-	}
-	atomic_inc(&block_group->trimming);
+	cleanup = (atomic_dec_and_test(&block_group->trimming) &&
+		   block_group->removed);
 	spin_unlock(&block_group->lock);
 
-	ret = trim_no_bitmap(block_group, trimmed, start, end, minlen);
-	if (ret)
-		goto out;
-
-	ret = trim_bitmaps(block_group, trimmed, start, end, minlen);
-out:
-	spin_lock(&block_group->lock);
-	if (atomic_dec_and_test(&block_group->trimming) &&
-	    block_group->removed) {
-		struct extent_map_tree *em_tree;
-		struct extent_map *em;
-
-		spin_unlock(&block_group->lock);
-
+	if (cleanup) {
 		lock_chunks(block_group->fs_info->chunk_root);
 		em_tree = &block_group->fs_info->mapping_tree.map_tree;
 		write_lock(&em_tree->lock);
@@ -3326,10 +3359,31 @@ out:
 		 * this block group have left 1 entry each one. Free them.
 		 */
 		__btrfs_remove_free_space_cache(block_group->free_space_ctl);
-	} else {
+	}
+}
+
+int btrfs_trim_block_group(struct btrfs_block_group_cache *block_group,
+			   u64 *trimmed, u64 start, u64 end, u64 minlen)
+{
+	int ret;
+
+	*trimmed = 0;
+
+	spin_lock(&block_group->lock);
+	if (block_group->removed) {
 		spin_unlock(&block_group->lock);
+		return 0;
 	}
+	btrfs_get_block_group_trimming(block_group);
+	spin_unlock(&block_group->lock);
 
+	ret = trim_no_bitmap(block_group, trimmed, start, end, minlen);
+	if (ret)
+		goto out;
+
+	ret = trim_bitmaps(block_group, trimmed, start, end, minlen);
+out:
+	btrfs_put_block_group_trimming(block_group);
 	return ret;
 }
 
@@ -3369,7 +3423,7 @@ u64 btrfs_find_ino_for_alloc(struct btrfs_root *fs_root)
 		u64 count = 1;
 		int ret;
 
-		ret = search_bitmap(ctl, entry, &offset, &count);
+		ret = search_bitmap(ctl, entry, &offset, &count, true);
 		/* Logic error; Should be empty if it can't find anything */
 		ASSERT(!ret);
 
@@ -3525,6 +3579,7 @@ again:
 		spin_lock(&ctl->tree_lock);
 		info->offset = offset;
 		info->bytes = bytes;
+		info->max_extent_size = 0;
 		ret = link_free_space(ctl, info);
 		spin_unlock(&ctl->tree_lock);
 		if (ret)
@@ -3552,6 +3607,7 @@ again:
 	}
 
 	bytes_added = add_bytes_to_bitmap(ctl, bitmap_info, offset, bytes);
+
 	bytes -= bytes_added;
 	offset += bytes_added;
 	spin_unlock(&ctl->tree_lock);
@@ -3595,7 +3651,7 @@ have_info:
 
 		bit_off = offset;
 		bit_bytes = ctl->unit;
-		ret = search_bitmap(ctl, info, &bit_off, &bit_bytes);
+		ret = search_bitmap(ctl, info, &bit_off, &bit_bytes, false);
 		if (!ret) {
 			if (bit_off == offset) {
 				ret = 1;
diff --git a/kernel/fs/btrfs/free-space-cache.h b/kernel/fs/btrfs/free-space-cache.h
index a16a029ad..f251865eb 100644
--- a/kernel/fs/btrfs/free-space-cache.h
+++ b/kernel/fs/btrfs/free-space-cache.h
@@ -23,6 +23,7 @@ struct btrfs_free_space {
 	struct rb_node offset_index;
 	u64 offset;
 	u64 bytes;
+	u64 max_extent_size;
 	unsigned long *bitmap;
 	struct list_head list;
 };
diff --git a/kernel/fs/btrfs/inode-item.c b/kernel/fs/btrfs/inode-item.c
index 265e03c73..be4d22a50 100644
--- a/kernel/fs/btrfs/inode-item.c
+++ b/kernel/fs/btrfs/inode-item.c
@@ -157,7 +157,7 @@ static int btrfs_del_inode_extref(struct btrfs_trans_handle *trans,
 	 */
 	if (!btrfs_find_name_in_ext_backref(path, ref_objectid,
 					    name, name_len, &extref)) {
-		btrfs_std_error(root->fs_info, -ENOENT);
+		btrfs_std_error(root->fs_info, -ENOENT, NULL);
 		ret = -EROFS;
 		goto out;
 	}
diff --git a/kernel/fs/btrfs/inode-map.c b/kernel/fs/btrfs/inode-map.c
index d4a582ac3..07573dc16 100644
--- a/kernel/fs/btrfs/inode-map.c
+++ b/kernel/fs/btrfs/inode-map.c
@@ -488,17 +488,17 @@ again:
 	/* Just to make sure we have enough space */
 	prealloc += 8 * PAGE_CACHE_SIZE;
 
-	ret = btrfs_delalloc_reserve_space(inode, prealloc);
+	ret = btrfs_delalloc_reserve_space(inode, 0, prealloc);
 	if (ret)
 		goto out_put;
 
 	ret = btrfs_prealloc_file_range_trans(inode, trans, 0, 0, prealloc,
 					      prealloc, prealloc, &alloc_hint);
 	if (ret) {
-		btrfs_delalloc_release_space(inode, prealloc);
+		btrfs_delalloc_release_space(inode, 0, prealloc);
 		goto out_put;
 	}
-	btrfs_free_reserved_data_space(inode, prealloc);
+	btrfs_free_reserved_data_space(inode, 0, prealloc);
 
 	ret = btrfs_write_out_ino_cache(root, trans, path, inode);
 out_put:
@@ -515,7 +515,7 @@ out:
 	return ret;
 }
 
-static int btrfs_find_highest_objectid(struct btrfs_root *root, u64 *objectid)
+int btrfs_find_highest_objectid(struct btrfs_root *root, u64 *objectid)
 {
 	struct btrfs_path *path;
 	int ret;
@@ -555,13 +555,6 @@ int btrfs_find_free_objectid(struct btrfs_root *root, u64 *objectid)
 	int ret;
 	mutex_lock(&root->objectid_mutex);
 
-	if (unlikely(root->highest_objectid < BTRFS_FIRST_FREE_OBJECTID)) {
-		ret = btrfs_find_highest_objectid(root,
-						  &root->highest_objectid);
-		if (ret)
-			goto out;
-	}
-
 	if (unlikely(root->highest_objectid >= BTRFS_LAST_FREE_OBJECTID)) {
 		ret = -ENOSPC;
 		goto out;
diff --git a/kernel/fs/btrfs/inode-map.h b/kernel/fs/btrfs/inode-map.h
index ddb347bfe..c8e864b2d 100644
--- a/kernel/fs/btrfs/inode-map.h
+++ b/kernel/fs/btrfs/inode-map.h
@@ -9,5 +9,6 @@ int btrfs_save_ino_cache(struct btrfs_root *root,
 			 struct btrfs_trans_handle *trans);
 
 int btrfs_find_free_objectid(struct btrfs_root *root, u64 *objectid);
+int btrfs_find_highest_objectid(struct btrfs_root *root, u64 *objectid);
 
 #endif
diff --git a/kernel/fs/btrfs/inode.c b/kernel/fs/btrfs/inode.c
index 8bb013672..4bc9dbf29 100644
--- a/kernel/fs/btrfs/inode.c
+++ b/kernel/fs/btrfs/inode.c
@@ -310,6 +310,13 @@ static noinline int cow_file_range_inline(struct btrfs_root *root,
 	btrfs_delalloc_release_metadata(inode, end + 1 - start);
 	btrfs_drop_extent_cache(inode, start, aligned_end - 1, 0);
 out:
+	/*
+	 * Don't forget to free the reserved space, as for inlined extent
+	 * it won't count as data extent, free them directly here.
+	 * And at reserve time, it's always aligned to page size, so
+	 * just free one page here.
+	 */
+	btrfs_qgroup_free_data(inode, 0, PAGE_CACHE_SIZE);
 	btrfs_free_path(path);
 	btrfs_end_transaction(trans, root);
 	return ret;
@@ -1096,6 +1103,9 @@ static noinline void async_cow_submit(struct btrfs_work *work)
 	nr_pages = (async_cow->end - async_cow->start + PAGE_CACHE_SIZE) >>
 		PAGE_CACHE_SHIFT;
 
+	/*
+	 * atomic_sub_return implies a barrier for waitqueue_active
+	 */
 	if (atomic_sub_return(nr_pages, &root->fs_info->async_delalloc_pages) <
 	    5 * 1024 * 1024 &&
 	    waitqueue_active(&root->fs_info->async_submit_wait))
@@ -1294,8 +1304,14 @@ next_slot:
 		num_bytes = 0;
 		btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
 
-		if (found_key.objectid > ino ||
-		    found_key.type > BTRFS_EXTENT_DATA_KEY ||
+		if (found_key.objectid > ino)
+			break;
+		if (WARN_ON_ONCE(found_key.objectid < ino) ||
+		    found_key.type < BTRFS_EXTENT_DATA_KEY) {
+			path->slots[0]++;
+			goto next_slot;
+		}
+		if (found_key.type > BTRFS_EXTENT_DATA_KEY ||
 		    found_key.offset > end)
 			break;
 
@@ -1766,7 +1782,8 @@ static void btrfs_clear_bit_hook(struct inode *inode,
 
 		if (root->root_key.objectid != BTRFS_DATA_RELOC_TREE_OBJECTID
 		    && do_list && !(state->state & EXTENT_NORESERVE))
-			btrfs_free_reserved_data_space(inode, len);
+			btrfs_free_reserved_data_space_noquota(inode,
+					state->start, len);
 
 		__percpu_counter_add(&root->fs_info->delalloc_bytes, -len,
 				     root->fs_info->delalloc_batch);
@@ -1845,8 +1862,10 @@ static int __btrfs_submit_bio_done(struct inode *inode, int rw, struct bio *bio,
 	int ret;
 
 	ret = btrfs_map_bio(root, rw, bio, mirror_num, 1);
-	if (ret)
-		bio_endio(bio, ret);
+	if (ret) {
+		bio->bi_error = ret;
+		bio_endio(bio);
+	}
 	return ret;
 }
 
@@ -1859,15 +1878,15 @@ static int btrfs_submit_bio_hook(struct inode *inode, int rw, struct bio *bio,
 			  u64 bio_offset)
 {
 	struct btrfs_root *root = BTRFS_I(inode)->root;
+	enum btrfs_wq_endio_type metadata = BTRFS_WQ_ENDIO_DATA;
 	int ret = 0;
 	int skip_sum;
-	int metadata = 0;
 	int async = !atomic_read(&BTRFS_I(inode)->sync_writers);
 
 	skip_sum = BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM;
 
 	if (btrfs_is_free_space_inode(inode))
-		metadata = 2;
+		metadata = BTRFS_WQ_ENDIO_FREE_SPACE;
 
 	if (!(rw & REQ_WRITE)) {
 		ret = btrfs_bio_wq_end_io(root->fs_info, bio, metadata);
@@ -1906,8 +1925,10 @@ mapit:
 	ret = btrfs_map_bio(root, rw, bio, mirror_num, 0);
 
 out:
-	if (ret < 0)
-		bio_endio(bio, ret);
+	if (ret < 0) {
+		bio->bi_error = ret;
+		bio_endio(bio);
+	}
 	return ret;
 }
 
@@ -1985,7 +2006,8 @@ again:
 		goto again;
 	}
 
-	ret = btrfs_delalloc_reserve_space(inode, PAGE_CACHE_SIZE);
+	ret = btrfs_delalloc_reserve_space(inode, page_start,
+					   PAGE_CACHE_SIZE);
 	if (ret) {
 		mapping_set_error(page->mapping, ret);
 		end_extent_writepage(page, ret, page_start, page_end);
@@ -2111,7 +2133,13 @@ static int insert_reserved_file_extent(struct btrfs_trans_handle *trans,
 	ins.type = BTRFS_EXTENT_ITEM_KEY;
 	ret = btrfs_alloc_reserved_file_extent(trans, root,
 					root->root_key.objectid,
-					btrfs_ino(inode), file_pos, &ins);
+					btrfs_ino(inode), file_pos,
+					ram_bytes, &ins);
+	/*
+	 * Release the reserved range from inode dirty range map, as it is
+	 * already moved into delayed_ref_head
+	 */
+	btrfs_qgroup_release_data(inode, file_pos, ram_bytes);
 out:
 	btrfs_free_path(path);
 
@@ -2569,7 +2597,7 @@ again:
 	ret = btrfs_inc_extent_ref(trans, root, new->bytenr,
 			new->disk_len, 0,
 			backref->root_id, backref->inum,
-			new->file_pos, 0);	/* start - extent_offset */
+			new->file_pos);	/* start - extent_offset */
 	if (ret) {
 		btrfs_abort_transaction(trans, root, ret);
 		goto out_free_path;
@@ -2595,7 +2623,6 @@ static void free_sa_defrag_extent(struct new_sa_defrag_extent *new)
 		return;
 
 	list_for_each_entry_safe(old, tmp, &new->head, list) {
-		list_del(&old->list);
 		kfree(old);
 	}
 	kfree(new);
@@ -2820,6 +2847,14 @@ static int btrfs_finish_ordered_io(struct btrfs_ordered_extent *ordered_extent)
 
 	if (test_bit(BTRFS_ORDERED_NOCOW, &ordered_extent->flags)) {
 		BUG_ON(!list_empty(&ordered_extent->list)); /* Logic error */
+
+		/*
+		 * For mwrite(mmap + memset to write) case, we still reserve
+		 * space for NOCOW range.
+		 * As NOCOW won't cause a new delayed ref, just free the space
+		 */
+		btrfs_qgroup_free_data(inode, ordered_extent->file_offset,
+				       ordered_extent->len);
 		btrfs_ordered_update_i_size(inode, 0, ordered_extent);
 		if (nolock)
 			trans = btrfs_join_transaction_nolock(root);
@@ -3014,8 +3049,6 @@ static int __readpage_endio_check(struct inode *inode,
 	char *kaddr;
 	u32 csum_expected;
 	u32 csum = ~(u32)0;
-	static DEFINE_RATELIMIT_STATE(_rs, DEFAULT_RATELIMIT_INTERVAL,
-				      DEFAULT_RATELIMIT_BURST);
 
 	csum_expected = *(((u32 *)io_bio->csum) + icsum);
 
@@ -3028,9 +3061,8 @@ static int __readpage_endio_check(struct inode *inode,
 	kunmap_atomic(kaddr);
 	return 0;
 zeroit:
-	if (__ratelimit(&_rs))
-		btrfs_warn(BTRFS_I(inode)->root->fs_info,
-			   "csum failed ino %llu off %llu csum %u expected csum %u",
+	btrfs_warn_rl(BTRFS_I(inode)->root->fs_info,
+		"csum failed ino %llu off %llu csum %u expected csum %u",
 			   btrfs_ino(inode), start, csum, csum_expected);
 	memset(kaddr + pgoff, 1, len);
 	flush_dcache_page(page);
@@ -3110,8 +3142,6 @@ void btrfs_run_delayed_iputs(struct btrfs_root *root)
 	if (empty)
 		return;
 
-	down_read(&fs_info->delayed_iput_sem);
-
 	spin_lock(&fs_info->delayed_iput_lock);
 	list_splice_init(&fs_info->delayed_iputs, &list);
 	spin_unlock(&fs_info->delayed_iput_lock);
@@ -3122,8 +3152,6 @@ void btrfs_run_delayed_iputs(struct btrfs_root *root)
 		iput(delayed->inode);
 		kfree(delayed);
 	}
-
-	up_read(&root->fs_info->delayed_iput_sem);
 }
 
 /*
@@ -3654,6 +3682,35 @@ cache_index:
 		set_bit(BTRFS_INODE_NEEDS_FULL_SYNC,
 			&BTRFS_I(inode)->runtime_flags);
 
+	/*
+	 * We don't persist the id of the transaction where an unlink operation
+	 * against the inode was last made. So here we assume the inode might
+	 * have been evicted, and therefore the exact value of last_unlink_trans
+	 * lost, and set it to last_trans to avoid metadata inconsistencies
+	 * between the inode and its parent if the inode is fsync'ed and the log
+	 * replayed. For example, in the scenario:
+	 *
+	 * touch mydir/foo
+	 * ln mydir/foo mydir/bar
+	 * sync
+	 * unlink mydir/bar
+	 * echo 2 > /proc/sys/vm/drop_caches   # evicts inode
+	 * xfs_io -c fsync mydir/foo
+	 * <power failure>
+	 * mount fs, triggers fsync log replay
+	 *
+	 * We must make sure that when we fsync our inode foo we also log its
+	 * parent inode, otherwise after log replay the parent still has the
+	 * dentry with the "bar" name but our inode foo has a link count of 1
+	 * and doesn't have an inode ref with the name "bar" anymore.
+	 *
+	 * Setting last_unlink_trans to last_trans is a pessimistic approach,
+	 * but it guarantees correctness at the expense of ocassional full
+	 * transaction commits on fsync if our inode is a directory, or if our
+	 * inode is not a directory, logging its parent unnecessarily.
+	 */
+	BTRFS_I(inode)->last_unlink_trans = BTRFS_I(inode)->last_trans;
+
 	path->slots[0]++;
 	if (inode->i_nlink != 1 ||
 	    path->slots[0] >= btrfs_header_nritems(leaf))
@@ -3985,9 +4042,7 @@ int btrfs_unlink_inode(struct btrfs_trans_handle *trans,
  */
 static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir)
 {
-	struct btrfs_trans_handle *trans;
 	struct btrfs_root *root = BTRFS_I(dir)->root;
-	int ret;
 
 	/*
 	 * 1 for the possible orphan item
@@ -3996,27 +4051,7 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir)
 	 * 1 for the inode ref
 	 * 1 for the inode
 	 */
-	trans = btrfs_start_transaction(root, 5);
-	if (!IS_ERR(trans) || PTR_ERR(trans) != -ENOSPC)
-		return trans;
-
-	if (PTR_ERR(trans) == -ENOSPC) {
-		u64 num_bytes = btrfs_calc_trans_metadata_size(root, 5);
-
-		trans = btrfs_start_transaction(root, 0);
-		if (IS_ERR(trans))
-			return trans;
-		ret = btrfs_cond_migrate_bytes(root->fs_info,
-					       &root->fs_info->trans_block_rsv,
-					       num_bytes, 5);
-		if (ret) {
-			btrfs_end_transaction(trans, root);
-			return ERR_PTR(ret);
-		}
-		trans->block_rsv = &root->fs_info->trans_block_rsv;
-		trans->bytes_reserved = num_bytes;
-	}
-	return trans;
+	return btrfs_start_transaction_fallback_global_rsv(root, 5, 5);
 }
 
 static int btrfs_unlink(struct inode *dir, struct dentry *dentry)
@@ -4184,6 +4219,47 @@ static int truncate_space_check(struct btrfs_trans_handle *trans,
 
 }
 
+static int truncate_inline_extent(struct inode *inode,
+				  struct btrfs_path *path,
+				  struct btrfs_key *found_key,
+				  const u64 item_end,
+				  const u64 new_size)
+{
+	struct extent_buffer *leaf = path->nodes[0];
+	int slot = path->slots[0];
+	struct btrfs_file_extent_item *fi;
+	u32 size = (u32)(new_size - found_key->offset);
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+
+	fi = btrfs_item_ptr(leaf, slot, struct btrfs_file_extent_item);
+
+	if (btrfs_file_extent_compression(leaf, fi) != BTRFS_COMPRESS_NONE) {
+		loff_t offset = new_size;
+		loff_t page_end = ALIGN(offset, PAGE_CACHE_SIZE);
+
+		/*
+		 * Zero out the remaining of the last page of our inline extent,
+		 * instead of directly truncating our inline extent here - that
+		 * would be much more complex (decompressing all the data, then
+		 * compressing the truncated data, which might be bigger than
+		 * the size of the inline extent, resize the extent, etc).
+		 * We release the path because to get the page we might need to
+		 * read the extent item from disk (data not in the page cache).
+		 */
+		btrfs_release_path(path);
+		return btrfs_truncate_page(inode, offset, page_end - offset, 0);
+	}
+
+	btrfs_set_file_extent_ram_bytes(leaf, fi, size);
+	size = btrfs_file_extent_calc_inline_size(size);
+	btrfs_truncate_item(root, path, size, 1);
+
+	if (test_bit(BTRFS_ROOT_REF_COWS, &root->state))
+		inode_sub_bytes(inode, item_end + 1 - new_size);
+
+	return 0;
+}
+
 /*
  * this can truncate away extent items, csum items and directory items.
  * It starts at a high offset and removes keys until it can't find
@@ -4209,7 +4285,7 @@ int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
 	u64 extent_num_bytes = 0;
 	u64 extent_offset = 0;
 	u64 item_end = 0;
-	u64 last_size = (u64)-1;
+	u64 last_size = new_size;
 	u32 found_type = (u8)-1;
 	int found_extent;
 	int del_item;
@@ -4378,27 +4454,40 @@ search_again:
 			 * special encodings
 			 */
 			if (!del_item &&
-			    btrfs_file_extent_compression(leaf, fi) == 0 &&
 			    btrfs_file_extent_encryption(leaf, fi) == 0 &&
 			    btrfs_file_extent_other_encoding(leaf, fi) == 0) {
-				u32 size = new_size - found_key.offset;
-
-				if (test_bit(BTRFS_ROOT_REF_COWS, &root->state))
-					inode_sub_bytes(inode, item_end + 1 -
-							new_size);
 
 				/*
-				 * update the ram bytes to properly reflect
-				 * the new size of our item
+				 * Need to release path in order to truncate a
+				 * compressed extent. So delete any accumulated
+				 * extent items so far.
 				 */
-				btrfs_set_file_extent_ram_bytes(leaf, fi, size);
-				size =
-				    btrfs_file_extent_calc_inline_size(size);
-				btrfs_truncate_item(root, path, size, 1);
+				if (btrfs_file_extent_compression(leaf, fi) !=
+				    BTRFS_COMPRESS_NONE && pending_del_nr) {
+					err = btrfs_del_items(trans, root, path,
+							      pending_del_slot,
+							      pending_del_nr);
+					if (err) {
+						btrfs_abort_transaction(trans,
+									root,
+									err);
+						goto error;
+					}
+					pending_del_nr = 0;
+				}
+
+				err = truncate_inline_extent(inode, path,
+							     &found_key,
+							     item_end,
+							     new_size);
+				if (err) {
+					btrfs_abort_transaction(trans,
+								root, err);
+					goto error;
+				}
 			} else if (test_bit(BTRFS_ROOT_REF_COWS,
 					    &root->state)) {
-				inode_sub_bytes(inode, item_end + 1 -
-						found_key.offset);
+				inode_sub_bytes(inode, item_end + 1 - new_size);
 			}
 		}
 delete:
@@ -4428,7 +4517,7 @@ delete:
 			ret = btrfs_free_extent(trans, root, extent_start,
 						extent_num_bytes, 0,
 						btrfs_header_owner(leaf),
-						ino, extent_offset, 0);
+						ino, extent_offset);
 			BUG_ON(ret);
 			if (btrfs_should_throttle_delayed_refs(trans, root))
 				btrfs_async_run_delayed_refs(root,
@@ -4493,8 +4582,7 @@ out:
 			btrfs_abort_transaction(trans, root, ret);
 	}
 error:
-	if (last_size != (u64)-1 &&
-	    root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID)
+	if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID)
 		btrfs_ordered_update_i_size(inode, last_size, NULL);
 
 	btrfs_free_path(path);
@@ -4543,14 +4631,17 @@ int btrfs_truncate_page(struct inode *inode, loff_t from, loff_t len,
 	if ((offset & (blocksize - 1)) == 0 &&
 	    (!len || ((len & (blocksize - 1)) == 0)))
 		goto out;
-	ret = btrfs_delalloc_reserve_space(inode, PAGE_CACHE_SIZE);
+	ret = btrfs_delalloc_reserve_space(inode,
+			round_down(from, PAGE_CACHE_SIZE), PAGE_CACHE_SIZE);
 	if (ret)
 		goto out;
 
 again:
 	page = find_or_create_page(mapping, index, mask);
 	if (!page) {
-		btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE);
+		btrfs_delalloc_release_space(inode,
+				round_down(from, PAGE_CACHE_SIZE),
+				PAGE_CACHE_SIZE);
 		ret = -ENOMEM;
 		goto out;
 	}
@@ -4618,7 +4709,8 @@ again:
 
 out_unlock:
 	if (ret)
-		btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE);
+		btrfs_delalloc_release_space(inode, page_start,
+					     PAGE_CACHE_SIZE);
 	unlock_page(page);
 	page_cache_release(page);
 out:
@@ -4986,24 +5078,53 @@ static void evict_inode_truncate_pages(struct inode *inode)
 	}
 	write_unlock(&map_tree->lock);
 
+	/*
+	 * Keep looping until we have no more ranges in the io tree.
+	 * We can have ongoing bios started by readpages (called from readahead)
+	 * that have their endio callback (extent_io.c:end_bio_extent_readpage)
+	 * still in progress (unlocked the pages in the bio but did not yet
+	 * unlocked the ranges in the io tree). Therefore this means some
+	 * ranges can still be locked and eviction started because before
+	 * submitting those bios, which are executed by a separate task (work
+	 * queue kthread), inode references (inode->i_count) were not taken
+	 * (which would be dropped in the end io callback of each bio).
+	 * Therefore here we effectively end up waiting for those bios and
+	 * anyone else holding locked ranges without having bumped the inode's
+	 * reference count - if we don't do it, when they access the inode's
+	 * io_tree to unlock a range it may be too late, leading to an
+	 * use-after-free issue.
+	 */
 	spin_lock(&io_tree->lock);
 	while (!RB_EMPTY_ROOT(&io_tree->state)) {
 		struct extent_state *state;
 		struct extent_state *cached_state = NULL;
+		u64 start;
+		u64 end;
 
 		node = rb_first(&io_tree->state);
 		state = rb_entry(node, struct extent_state, rb_node);
-		atomic_inc(&state->refs);
+		start = state->start;
+		end = state->end;
 		spin_unlock(&io_tree->lock);
 
-		lock_extent_bits(io_tree, state->start, state->end,
-				 0, &cached_state);
-		clear_extent_bit(io_tree, state->start, state->end,
+		lock_extent_bits(io_tree, start, end, 0, &cached_state);
+
+		/*
+		 * If still has DELALLOC flag, the extent didn't reach disk,
+		 * and its reserved space won't be freed by delayed_ref.
+		 * So we need to free its reserved space here.
+		 * (Refer to comment in btrfs_invalidatepage, case 2)
+		 *
+		 * Note, end is the bytenr of last byte, so we need + 1 here.
+		 */
+		if (state->state & EXTENT_DELALLOC)
+			btrfs_qgroup_free_data(inode, start, end - start + 1);
+
+		clear_extent_bit(io_tree, start, end,
 				 EXTENT_LOCKED | EXTENT_DIRTY |
 				 EXTENT_DELALLOC | EXTENT_DO_ACCOUNTING |
 				 EXTENT_DEFRAG, 1, 1,
 				 &cached_state, GFP_NOFS);
-		free_extent_state(state);
 
 		cond_resched();
 		spin_lock(&io_tree->lock);
@@ -5035,7 +5156,8 @@ void btrfs_evict_inode(struct inode *inode)
 		goto no_delete;
 	}
 	/* do we really want it for ->i_nlink > 0 and zero btrfs_root_refs? */
-	btrfs_wait_ordered_range(inode, 0, (u64)-1);
+	if (!special_file(inode->i_mode))
+		btrfs_wait_ordered_range(inode, 0, (u64)-1);
 
 	btrfs_free_io_failure_record(inode, 0, (u64)-1);
 
@@ -5615,6 +5737,7 @@ static int btrfs_real_readdir(struct file *file, struct dir_context *ctx)
 	char *name_ptr;
 	int name_len;
 	int is_curr = 0;	/* ctx->pos points to the current index? */
+	bool emitted;
 
 	/* FIXME, use a real flag for deciding about the key type */
 	if (root->fs_info->tree_root == root)
@@ -5643,6 +5766,7 @@ static int btrfs_real_readdir(struct file *file, struct dir_context *ctx)
 	if (ret < 0)
 		goto err;
 
+	emitted = false;
 	while (1) {
 		leaf = path->nodes[0];
 		slot = path->slots[0];
@@ -5722,6 +5846,7 @@ skip:
 
 			if (over)
 				goto nopos;
+			emitted = true;
 			di_len = btrfs_dir_name_len(leaf, di) +
 				 btrfs_dir_data_len(leaf, di) + sizeof(*di);
 			di_cur += di_len;
@@ -5734,11 +5859,20 @@ next:
 	if (key_type == BTRFS_DIR_INDEX_KEY) {
 		if (is_curr)
 			ctx->pos++;
-		ret = btrfs_readdir_delayed_dir_index(ctx, &ins_list);
+		ret = btrfs_readdir_delayed_dir_index(ctx, &ins_list, &emitted);
 		if (ret)
 			goto nopos;
 	}
 
+	/*
+	 * If we haven't emitted any dir entry, we must not touch ctx->pos as
+	 * it was was set to the termination value in previous call. We assume
+	 * that "." and ".." were emitted if we reach this point and set the
+	 * termination value as well for an empty directory.
+	 */
+	if (ctx->pos > 2 && !emitted)
+		goto nopos;
+
 	/* Reached end of directory/root. Bump pos past the last item. */
 	ctx->pos++;
 
@@ -6218,9 +6352,6 @@ static int btrfs_mknod(struct inode *dir, struct dentry *dentry,
 	u64 objectid;
 	u64 index = 0;
 
-	if (!new_valid_dev(rdev))
-		return -EINVAL;
-
 	/*
 	 * 2 for inode item and ref
 	 * 2 for dir items
@@ -6358,7 +6489,7 @@ out_unlock_inode:
 static int btrfs_link(struct dentry *old_dentry, struct inode *dir,
 		      struct dentry *dentry)
 {
-	struct btrfs_trans_handle *trans;
+	struct btrfs_trans_handle *trans = NULL;
 	struct btrfs_root *root = BTRFS_I(dir)->root;
 	struct inode *inode = d_inode(old_dentry);
 	u64 index;
@@ -6384,6 +6515,7 @@ static int btrfs_link(struct dentry *old_dentry, struct inode *dir,
 	trans = btrfs_start_transaction(root, 5);
 	if (IS_ERR(trans)) {
 		err = PTR_ERR(trans);
+		trans = NULL;
 		goto fail;
 	}
 
@@ -6417,9 +6549,10 @@ static int btrfs_link(struct dentry *old_dentry, struct inode *dir,
 		btrfs_log_new_name(trans, inode, NULL, parent);
 	}
 
-	btrfs_end_transaction(trans, root);
 	btrfs_balance_delayed_items(root);
 fail:
+	if (trans)
+		btrfs_end_transaction(trans, root);
 	if (drop_inode) {
 		inode_dec_link_count(inode);
 		iput(inode);
@@ -6860,8 +6993,7 @@ out:
 
 	trace_btrfs_get_extent(root, em);
 
-	if (path)
-		btrfs_free_path(path);
+	btrfs_free_path(path);
 	if (trans) {
 		ret = btrfs_end_transaction(trans, root);
 		if (!err)
@@ -7360,6 +7492,32 @@ static struct extent_map *create_pinned_em(struct inode *inode, u64 start,
 	return em;
 }
 
+struct btrfs_dio_data {
+	u64 outstanding_extents;
+	u64 reserve;
+};
+
+static void adjust_dio_outstanding_extents(struct inode *inode,
+					   struct btrfs_dio_data *dio_data,
+					   const u64 len)
+{
+	unsigned num_extents;
+
+	num_extents = (unsigned) div64_u64(len + BTRFS_MAX_EXTENT_SIZE - 1,
+					   BTRFS_MAX_EXTENT_SIZE);
+	/*
+	 * If we have an outstanding_extents count still set then we're
+	 * within our reservation, otherwise we need to adjust our inode
+	 * counter appropriately.
+	 */
+	if (dio_data->outstanding_extents) {
+		dio_data->outstanding_extents -= num_extents;
+	} else {
+		spin_lock(&BTRFS_I(inode)->lock);
+		BTRFS_I(inode)->outstanding_extents += num_extents;
+		spin_unlock(&BTRFS_I(inode)->lock);
+	}
+}
 
 static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock,
 				   struct buffer_head *bh_result, int create)
@@ -7367,10 +7525,10 @@ static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock,
 	struct extent_map *em;
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	struct extent_state *cached_state = NULL;
+	struct btrfs_dio_data *dio_data = NULL;
 	u64 start = iblock << inode->i_blkbits;
 	u64 lockstart, lockend;
 	u64 len = bh_result->b_size;
-	u64 *outstanding_extents = NULL;
 	int unlock_bits = EXTENT_LOCKED;
 	int ret = 0;
 
@@ -7388,7 +7546,7 @@ static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock,
 		 * that anything that needs to check if there's a transction doesn't get
 		 * confused.
 		 */
-		outstanding_extents = current->journal_info;
+		dio_data = current->journal_info;
 		current->journal_info = NULL;
 	}
 
@@ -7396,8 +7554,11 @@ static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock,
 	 * If this errors out it's because we couldn't invalidate pagecache for
 	 * this range and we need to fallback to buffered.
 	 */
-	if (lock_extent_direct(inode, lockstart, lockend, &cached_state, create))
-		return -ENOTBLK;
+	if (lock_extent_direct(inode, lockstart, lockend, &cached_state,
+			       create)) {
+		ret = -ENOTBLK;
+		goto err;
+	}
 
 	em = btrfs_get_extent(inode, NULL, 0, start, len, 0);
 	if (IS_ERR(em)) {
@@ -7515,21 +7676,11 @@ unlock:
 		if (start + len > i_size_read(inode))
 			i_size_write(inode, start + len);
 
-		/*
-		 * If we have an outstanding_extents count still set then we're
-		 * within our reservation, otherwise we need to adjust our inode
-		 * counter appropriately.
-		 */
-		if (*outstanding_extents) {
-			(*outstanding_extents)--;
-		} else {
-			spin_lock(&BTRFS_I(inode)->lock);
-			BTRFS_I(inode)->outstanding_extents++;
-			spin_unlock(&BTRFS_I(inode)->lock);
-		}
-
-		current->journal_info = outstanding_extents;
-		btrfs_free_reserved_data_space(inode, len);
+		adjust_dio_outstanding_extents(inode, dio_data, len);
+		btrfs_free_reserved_data_space(inode, start, len);
+		WARN_ON(dio_data->reserve < len);
+		dio_data->reserve -= len;
+		current->journal_info = dio_data;
 	}
 
 	/*
@@ -7552,8 +7703,17 @@ unlock:
 unlock_err:
 	clear_extent_bit(&BTRFS_I(inode)->io_tree, lockstart, lockend,
 			 unlock_bits, 1, 0, &cached_state, GFP_NOFS);
-	if (outstanding_extents)
-		current->journal_info = outstanding_extents;
+err:
+	if (dio_data)
+		current->journal_info = dio_data;
+	/*
+	 * Compensate the delalloc release we do in btrfs_direct_IO() when we
+	 * write less data then expected, so that we don't underflow our inode's
+	 * outstanding extents counter.
+	 */
+	if (create && dio_data)
+		adjust_dio_outstanding_extents(inode, dio_data, len);
+
 	return ret;
 }
 
@@ -7671,13 +7831,13 @@ struct btrfs_retry_complete {
 	int uptodate;
 };
 
-static void btrfs_retry_endio_nocsum(struct bio *bio, int err)
+static void btrfs_retry_endio_nocsum(struct bio *bio)
 {
 	struct btrfs_retry_complete *done = bio->bi_private;
 	struct bio_vec *bvec;
 	int i;
 
-	if (err)
+	if (bio->bi_error)
 		goto end;
 
 	done->uptodate = 1;
@@ -7726,7 +7886,7 @@ try_again:
 	return 0;
 }
 
-static void btrfs_retry_endio(struct bio *bio, int err)
+static void btrfs_retry_endio(struct bio *bio)
 {
 	struct btrfs_retry_complete *done = bio->bi_private;
 	struct btrfs_io_bio *io_bio = btrfs_io_bio(bio);
@@ -7735,7 +7895,7 @@ static void btrfs_retry_endio(struct bio *bio, int err)
 	int ret;
 	int i;
 
-	if (err)
+	if (bio->bi_error)
 		goto end;
 
 	uptodate = 1;
@@ -7818,12 +7978,13 @@ static int btrfs_subio_endio_read(struct inode *inode,
 	}
 }
 
-static void btrfs_endio_direct_read(struct bio *bio, int err)
+static void btrfs_endio_direct_read(struct bio *bio)
 {
 	struct btrfs_dio_private *dip = bio->bi_private;
 	struct inode *inode = dip->inode;
 	struct bio *dio_bio;
 	struct btrfs_io_bio *io_bio = btrfs_io_bio(bio);
+	int err = bio->bi_error;
 
 	if (dip->flags & BTRFS_DIO_ORIG_BIO_SUBMITTED)
 		err = btrfs_subio_endio_read(inode, io_bio, err);
@@ -7834,17 +7995,15 @@ static void btrfs_endio_direct_read(struct bio *bio, int err)
 
 	kfree(dip);
 
-	/* If we had a csum failure make sure to clear the uptodate flag */
-	if (err)
-		clear_bit(BIO_UPTODATE, &dio_bio->bi_flags);
-	dio_end_io(dio_bio, err);
+	dio_bio->bi_error = bio->bi_error;
+	dio_end_io(dio_bio, bio->bi_error);
 
 	if (io_bio->end_io)
 		io_bio->end_io(io_bio, err);
 	bio_put(bio);
 }
 
-static void btrfs_endio_direct_write(struct bio *bio, int err)
+static void btrfs_endio_direct_write(struct bio *bio)
 {
 	struct btrfs_dio_private *dip = bio->bi_private;
 	struct inode *inode = dip->inode;
@@ -7855,12 +8014,11 @@ static void btrfs_endio_direct_write(struct bio *bio, int err)
 	struct bio *dio_bio;
 	int ret;
 
-	if (err)
-		goto out_done;
 again:
 	ret = btrfs_dec_test_first_ordered_pending(inode, &ordered,
 						   &ordered_offset,
-						   ordered_bytes, !err);
+						   ordered_bytes,
+						   !bio->bi_error);
 	if (!ret)
 		goto out_test;
 
@@ -7879,15 +8037,12 @@ out_test:
 		ordered = NULL;
 		goto again;
 	}
-out_done:
 	dio_bio = dip->dio_bio;
 
 	kfree(dip);
 
-	/* If we had an error make sure to clear the uptodate flag */
-	if (err)
-		clear_bit(BIO_UPTODATE, &dio_bio->bi_flags);
-	dio_end_io(dio_bio, err);
+	dio_bio->bi_error = bio->bi_error;
+	dio_end_io(dio_bio, bio->bi_error);
 	bio_put(bio);
 }
 
@@ -7902,9 +8057,10 @@ static int __btrfs_submit_bio_start_direct_io(struct inode *inode, int rw,
 	return 0;
 }
 
-static void btrfs_end_dio_bio(struct bio *bio, int err)
+static void btrfs_end_dio_bio(struct bio *bio)
 {
 	struct btrfs_dio_private *dip = bio->bi_private;
+	int err = bio->bi_error;
 
 	if (err)
 		btrfs_warn(BTRFS_I(dip->inode)->root->fs_info,
@@ -7933,8 +8089,8 @@ static void btrfs_end_dio_bio(struct bio *bio, int err)
 	if (dip->errors) {
 		bio_io_error(dip->orig_bio);
 	} else {
-		set_bit(BIO_UPTODATE, &dip->dio_bio->bi_flags);
-		bio_endio(dip->orig_bio, 0);
+		dip->dio_bio->bi_error = 0;
+		bio_endio(dip->orig_bio);
 	}
 out:
 	bio_put(bio);
@@ -7943,8 +8099,11 @@ out:
 static struct bio *btrfs_dio_bio_alloc(struct block_device *bdev,
 				       u64 first_sector, gfp_t gfp_flags)
 {
-	int nr_vecs = bio_get_nr_vecs(bdev);
-	return btrfs_bio_alloc(bdev, first_sector, nr_vecs, gfp_flags);
+	struct bio *bio;
+	bio = btrfs_bio_alloc(bdev, first_sector, BIO_MAX_PAGES, gfp_flags);
+	if (bio)
+		bio_associate_current(bio);
+	return bio;
 }
 
 static inline int btrfs_lookup_and_bind_dio_csum(struct btrfs_root *root,
@@ -8147,9 +8306,8 @@ out_err:
 static void btrfs_submit_direct(int rw, struct bio *dio_bio,
 				struct inode *inode, loff_t file_offset)
 {
-	struct btrfs_root *root = BTRFS_I(inode)->root;
-	struct btrfs_dio_private *dip;
-	struct bio *io_bio;
+	struct btrfs_dio_private *dip = NULL;
+	struct bio *io_bio = NULL;
 	struct btrfs_io_bio *btrfs_bio;
 	int skip_sum;
 	int write = rw & REQ_WRITE;
@@ -8166,7 +8324,7 @@ static void btrfs_submit_direct(int rw, struct bio *dio_bio,
 	dip = kzalloc(sizeof(*dip), GFP_NOFS);
 	if (!dip) {
 		ret = -ENOMEM;
-		goto free_io_bio;
+		goto free_ordered;
 	}
 
 	dip->private = dio_bio->bi_private;
@@ -8194,25 +8352,56 @@ static void btrfs_submit_direct(int rw, struct bio *dio_bio,
 
 	if (btrfs_bio->end_io)
 		btrfs_bio->end_io(btrfs_bio, ret);
-free_io_bio:
-	bio_put(io_bio);
 
 free_ordered:
 	/*
-	 * If this is a write, we need to clean up the reserved space and kill
-	 * the ordered extent.
+	 * If we arrived here it means either we failed to submit the dip
+	 * or we either failed to clone the dio_bio or failed to allocate the
+	 * dip. If we cloned the dio_bio and allocated the dip, we can just
+	 * call bio_endio against our io_bio so that we get proper resource
+	 * cleanup if we fail to submit the dip, otherwise, we must do the
+	 * same as btrfs_endio_direct_[write|read] because we can't call these
+	 * callbacks - they require an allocated dip and a clone of dio_bio.
 	 */
-	if (write) {
-		struct btrfs_ordered_extent *ordered;
-		ordered = btrfs_lookup_ordered_extent(inode, file_offset);
-		if (!test_bit(BTRFS_ORDERED_PREALLOC, &ordered->flags) &&
-		    !test_bit(BTRFS_ORDERED_NOCOW, &ordered->flags))
-			btrfs_free_reserved_extent(root, ordered->start,
-						   ordered->disk_len, 1);
-		btrfs_put_ordered_extent(ordered);
-		btrfs_put_ordered_extent(ordered);
+	if (io_bio && dip) {
+		io_bio->bi_error = -EIO;
+		bio_endio(io_bio);
+		/*
+		 * The end io callbacks free our dip, do the final put on io_bio
+		 * and all the cleanup and final put for dio_bio (through
+		 * dio_end_io()).
+		 */
+		dip = NULL;
+		io_bio = NULL;
+	} else {
+		if (write) {
+			struct btrfs_ordered_extent *ordered;
+
+			ordered = btrfs_lookup_ordered_extent(inode,
+							      file_offset);
+			set_bit(BTRFS_ORDERED_IOERR, &ordered->flags);
+			/*
+			 * Decrements our ref on the ordered extent and removes
+			 * the ordered extent from the inode's ordered tree,
+			 * doing all the proper resource cleanup such as for the
+			 * reserved space and waking up any waiters for this
+			 * ordered extent (through btrfs_remove_ordered_extent).
+			 */
+			btrfs_finish_ordered_io(ordered);
+		} else {
+			unlock_extent(&BTRFS_I(inode)->io_tree, file_offset,
+			      file_offset + dio_bio->bi_iter.bi_size - 1);
+		}
+		dio_bio->bi_error = -EIO;
+		/*
+		 * Releases and cleans up our dio_bio, no need to bio_put()
+		 * nor bio_endio()/bio_io_error() against dio_bio.
+		 */
+		dio_end_io(dio_bio, ret);
 	}
-	bio_endio(dio_bio, ret);
+	if (io_bio)
+		bio_put(io_bio);
+	kfree(dip);
 }
 
 static ssize_t check_direct_IO(struct btrfs_root *root, struct kiocb *iocb,
@@ -8253,7 +8442,8 @@ static ssize_t btrfs_direct_IO(struct kiocb *iocb, struct iov_iter *iter,
 {
 	struct file *file = iocb->ki_filp;
 	struct inode *inode = file->f_mapping->host;
-	u64 outstanding_extents = 0;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct btrfs_dio_data dio_data = { 0 };
 	size_t count = 0;
 	int flags = 0;
 	bool wakeup = true;
@@ -8288,10 +8478,10 @@ static ssize_t btrfs_direct_IO(struct kiocb *iocb, struct iov_iter *iter,
 			mutex_unlock(&inode->i_mutex);
 			relock = true;
 		}
-		ret = btrfs_delalloc_reserve_space(inode, count);
+		ret = btrfs_delalloc_reserve_space(inode, offset, count);
 		if (ret)
 			goto out;
-		outstanding_extents = div64_u64(count +
+		dio_data.outstanding_extents = div64_u64(count +
 						BTRFS_MAX_EXTENT_SIZE - 1,
 						BTRFS_MAX_EXTENT_SIZE);
 
@@ -8300,7 +8490,8 @@ static ssize_t btrfs_direct_IO(struct kiocb *iocb, struct iov_iter *iter,
 		 * do the accounting properly if we go over the number we
 		 * originally calculated.  Abuse current->journal_info for this.
 		 */
-		current->journal_info = &outstanding_extents;
+		dio_data.reserve = round_up(count, root->sectorsize);
+		current->journal_info = &dio_data;
 	} else if (test_bit(BTRFS_INODE_READDIO_NEED_LOCK,
 				     &BTRFS_I(inode)->runtime_flags)) {
 		inode_dio_end(inode);
@@ -8314,10 +8505,12 @@ static ssize_t btrfs_direct_IO(struct kiocb *iocb, struct iov_iter *iter,
 				   btrfs_submit_direct, flags);
 	if (iov_iter_rw(iter) == WRITE) {
 		current->journal_info = NULL;
-		if (ret < 0 && ret != -EIOCBQUEUED)
-			btrfs_delalloc_release_space(inode, count);
-		else if (ret >= 0 && (size_t)ret < count)
-			btrfs_delalloc_release_space(inode,
+		if (ret < 0 && ret != -EIOCBQUEUED) {
+			if (dio_data.reserve)
+				btrfs_delalloc_release_space(inode, offset,
+							     dio_data.reserve);
+		} else if (ret >= 0 && (size_t)ret < count)
+			btrfs_delalloc_release_space(inode, offset,
 						     count - (size_t)ret);
 	}
 out:
@@ -8353,15 +8546,28 @@ int btrfs_readpage(struct file *file, struct page *page)
 static int btrfs_writepage(struct page *page, struct writeback_control *wbc)
 {
 	struct extent_io_tree *tree;
-
+	struct inode *inode = page->mapping->host;
+	int ret;
 
 	if (current->flags & PF_MEMALLOC) {
 		redirty_page_for_writepage(wbc, page);
 		unlock_page(page);
 		return 0;
 	}
+
+	/*
+	 * If we are under memory pressure we will call this directly from the
+	 * VM, we need to make sure we have the inode referenced for the ordered
+	 * extent.  If not just return like we didn't do anything.
+	 */
+	if (!igrab(inode)) {
+		redirty_page_for_writepage(wbc, page);
+		return AOP_WRITEPAGE_ACTIVATE;
+	}
 	tree = &BTRFS_I(page->mapping->host)->io_tree;
-	return extent_write_full_page(tree, page, btrfs_get_extent, wbc);
+	ret = extent_write_full_page(tree, page, btrfs_get_extent, wbc);
+	btrfs_add_delayed_iput(inode);
+	return ret;
 }
 
 static int btrfs_writepages(struct address_space *mapping,
@@ -8476,6 +8682,18 @@ static void btrfs_invalidatepage(struct page *page, unsigned int offset,
 		}
 	}
 
+	/*
+	 * Qgroup reserved space handler
+	 * Page here will be either
+	 * 1) Already written to disk
+	 *    In this case, its reserved space is released from data rsv map
+	 *    and will be freed by delayed_ref handler finally.
+	 *    So even we call qgroup_free_data(), it won't decrease reserved
+	 *    space.
+	 * 2) Not written to disk
+	 *    This means the reserved space should be freed here.
+	 */
+	btrfs_qgroup_free_data(inode, page_start, PAGE_CACHE_SIZE);
 	if (!inode_evicting) {
 		clear_extent_bit(tree, page_start, page_end,
 				 EXTENT_LOCKED | EXTENT_DIRTY |
@@ -8526,7 +8744,11 @@ int btrfs_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf)
 	u64 page_end;
 
 	sb_start_pagefault(inode->i_sb);
-	ret  = btrfs_delalloc_reserve_space(inode, PAGE_CACHE_SIZE);
+	page_start = page_offset(page);
+	page_end = page_start + PAGE_CACHE_SIZE - 1;
+
+	ret = btrfs_delalloc_reserve_space(inode, page_start,
+					   PAGE_CACHE_SIZE);
 	if (!ret) {
 		ret = file_update_time(vma->vm_file);
 		reserved = 1;
@@ -8545,8 +8767,6 @@ int btrfs_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf)
 again:
 	lock_page(page);
 	size = i_size_read(inode);
-	page_start = page_offset(page);
-	page_end = page_start + PAGE_CACHE_SIZE - 1;
 
 	if ((page->mapping != inode->i_mapping) ||
 	    (page_start >= size)) {
@@ -8623,7 +8843,7 @@ out_unlock:
 	}
 	unlock_page(page);
 out:
-	btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE);
+	btrfs_delalloc_release_space(inode, page_start, PAGE_CACHE_SIZE);
 out_noreserve:
 	sb_end_pagefault(inode->i_sb);
 	return ret;
@@ -8912,6 +9132,7 @@ void btrfs_destroy_inode(struct inode *inode)
 			btrfs_put_ordered_extent(ordered);
 		}
 	}
+	btrfs_qgroup_check_reserved_leak(inode);
 	inode_tree_del(inode);
 	btrfs_drop_extent_cache(inode, 0, (u64)-1, 0);
 free:
@@ -9440,9 +9661,11 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry,
 	/*
 	 * 2 items for inode item and ref
 	 * 2 items for dir items
+	 * 1 item for updating parent inode item
+	 * 1 item for the inline extent item
 	 * 1 item for xattr if selinux is on
 	 */
-	trans = btrfs_start_transaction(root, 5);
+	trans = btrfs_start_transaction(root, 7);
 	if (IS_ERR(trans))
 		return PTR_ERR(trans);
 
@@ -9548,6 +9771,7 @@ static int __btrfs_prealloc_file_range(struct inode *inode, int mode,
 	u64 cur_offset = start;
 	u64 i_size;
 	u64 cur_bytes;
+	u64 last_alloc = (u64)-1;
 	int ret = 0;
 	bool own_trans = true;
 
@@ -9564,6 +9788,13 @@ static int __btrfs_prealloc_file_range(struct inode *inode, int mode,
 
 		cur_bytes = min(num_bytes, 256ULL * 1024 * 1024);
 		cur_bytes = max(cur_bytes, min_size);
+		/*
+		 * If we are severely fragmented we could end up with really
+		 * small allocations, so if the allocator is returning small
+		 * chunks lets make its job easier by only searching for those
+		 * sized chunks.
+		 */
+		cur_bytes = min(cur_bytes, last_alloc);
 		ret = btrfs_reserve_extent(root, cur_bytes, min_size, 0,
 					   *alloc_hint, &ins, 1, 0);
 		if (ret) {
@@ -9572,6 +9803,7 @@ static int __btrfs_prealloc_file_range(struct inode *inode, int mode,
 			break;
 		}
 
+		last_alloc = ins.offset;
 		ret = insert_reserved_file_extent(trans, inode,
 						  cur_offset, ins.objectid,
 						  ins.offset, ins.offset,
diff --git a/kernel/fs/btrfs/ioctl.c b/kernel/fs/btrfs/ioctl.c
index 37d456a9a..f07d01bc4 100644
--- a/kernel/fs/btrfs/ioctl.c
+++ b/kernel/fs/btrfs/ioctl.c
@@ -87,7 +87,8 @@ struct btrfs_ioctl_received_subvol_args_32 {
 
 
 static int btrfs_clone(struct inode *src, struct inode *inode,
-		       u64 off, u64 olen, u64 olen_aligned, u64 destoff);
+		       u64 off, u64 olen, u64 olen_aligned, u64 destoff,
+		       int no_time_update);
 
 /* Mask out flags that are inappropriate for the given type of inode. */
 static inline __u32 btrfs_mask_flags(umode_t mode, __u32 flags)
@@ -553,8 +554,8 @@ static noinline int create_subvol(struct inode *dir,
 	key.offset = (u64)-1;
 	new_root = btrfs_read_fs_root_no_name(root->fs_info, &key);
 	if (IS_ERR(new_root)) {
-		btrfs_abort_transaction(trans, root, PTR_ERR(new_root));
 		ret = PTR_ERR(new_root);
+		btrfs_abort_transaction(trans, root, ret);
 		goto fail;
 	}
 
@@ -567,6 +568,10 @@ static noinline int create_subvol(struct inode *dir,
 		goto fail;
 	}
 
+	mutex_lock(&new_root->objectid_mutex);
+	new_root->highest_objectid = new_dirid;
+	mutex_unlock(&new_root->objectid_mutex);
+
 	/*
 	 * insert the directory item
 	 */
@@ -1029,6 +1034,7 @@ static int should_defrag_range(struct inode *inode, u64 start, u32 thresh,
 	struct extent_map *em;
 	int ret = 1;
 	bool next_mergeable = true;
+	bool prev_mergeable = true;
 
 	/*
 	 * make sure that once we start defragging an extent, we keep on
@@ -1049,13 +1055,16 @@ static int should_defrag_range(struct inode *inode, u64 start, u32 thresh,
 		goto out;
 	}
 
+	if (!*defrag_end)
+		prev_mergeable = false;
+
 	next_mergeable = defrag_check_next_extent(inode, em);
 	/*
 	 * we hit a real extent, if it is big or the next extent is not a
 	 * real extent, don't bother defragging it
 	 */
 	if (!compress && (*last_len == 0 || *last_len >= thresh) &&
-	    (em->len >= thresh || !next_mergeable))
+	    (em->len >= thresh || (!next_mergeable && !prev_mergeable)))
 		ret = 0;
 out:
 	/*
@@ -1115,7 +1124,8 @@ static int cluster_pages_for_defrag(struct inode *inode,
 	page_cnt = min_t(u64, (u64)num_pages, (u64)file_end - start_index + 1);
 
 	ret = btrfs_delalloc_reserve_space(inode,
-					   page_cnt << PAGE_CACHE_SHIFT);
+			start_index << PAGE_CACHE_SHIFT,
+			page_cnt << PAGE_CACHE_SHIFT);
 	if (ret)
 		return ret;
 	i_done = 0;
@@ -1205,7 +1215,8 @@ again:
 		BTRFS_I(inode)->outstanding_extents++;
 		spin_unlock(&BTRFS_I(inode)->lock);
 		btrfs_delalloc_release_space(inode,
-				     (page_cnt - i_done) << PAGE_CACHE_SHIFT);
+				start_index << PAGE_CACHE_SHIFT,
+				(page_cnt - i_done) << PAGE_CACHE_SHIFT);
 	}
 
 
@@ -1230,7 +1241,9 @@ out:
 		unlock_page(pages[i]);
 		page_cache_release(pages[i]);
 	}
-	btrfs_delalloc_release_space(inode, page_cnt << PAGE_CACHE_SHIFT);
+	btrfs_delalloc_release_space(inode,
+			start_index << PAGE_CACHE_SHIFT,
+			page_cnt << PAGE_CACHE_SHIFT);
 	return ret;
 
 }
@@ -1318,7 +1331,7 @@ int btrfs_defrag_file(struct inode *inode, struct file *file,
 		i = range->start >> PAGE_CACHE_SHIFT;
 	}
 	if (!max_to_defrag)
-		max_to_defrag = last_index + 1;
+		max_to_defrag = last_index - i + 1;
 
 	/*
 	 * make writeback starts from i, so the defrag range can be
@@ -1337,7 +1350,7 @@ int btrfs_defrag_file(struct inode *inode, struct file *file,
 			break;
 
 		if (btrfs_defrag_cancelled(root->fs_info)) {
-			printk(KERN_DEBUG "BTRFS: defrag_file cancelled\n");
+			btrfs_debug(root->fs_info, "defrag_file cancelled");
 			ret = -EAGAIN;
 			break;
 		}
@@ -1368,7 +1381,7 @@ int btrfs_defrag_file(struct inode *inode, struct file *file,
 			ra_index = max(i, ra_index);
 			btrfs_force_ra(inode->i_mapping, ra, file, ra_index,
 				       cluster);
-			ra_index += max_cluster;
+			ra_index += cluster;
 		}
 
 		mutex_lock(&inode->i_mutex);
@@ -1574,7 +1587,7 @@ static noinline int btrfs_ioctl_resize(struct file *file,
 	new_size = div_u64(new_size, root->sectorsize);
 	new_size *= root->sectorsize;
 
-	printk_in_rcu(KERN_INFO "BTRFS: new size for %s is %llu\n",
+	btrfs_info_in_rcu(root->fs_info, "new size for %s is %llu",
 		      rcu_str_deref(device->name), new_size);
 
 	if (new_size > old_size) {
@@ -1932,6 +1945,7 @@ static noinline int copy_to_sk(struct btrfs_root *root,
 	u64 found_transid;
 	struct extent_buffer *leaf;
 	struct btrfs_ioctl_search_header sh;
+	struct btrfs_key test;
 	unsigned long item_off;
 	unsigned long item_len;
 	int nritems;
@@ -2015,12 +2029,17 @@ static noinline int copy_to_sk(struct btrfs_root *root,
 	}
 advance_key:
 	ret = 0;
-	if (key->offset < (u64)-1 && key->offset < sk->max_offset)
+	test.objectid = sk->max_objectid;
+	test.type = sk->max_type;
+	test.offset = sk->max_offset;
+	if (btrfs_comp_cpu_keys(key, &test) >= 0)
+		ret = 1;
+	else if (key->offset < (u64)-1)
 		key->offset++;
-	else if (key->type < (u8)-1 && key->type < sk->max_type) {
+	else if (key->type < (u8)-1) {
 		key->offset = 0;
 		key->type++;
-	} else if (key->objectid < (u64)-1 && key->objectid < sk->max_objectid) {
+	} else if (key->objectid < (u64)-1) {
 		key->offset = 0;
 		key->type = 0;
 		key->objectid++;
@@ -2070,7 +2089,7 @@ static noinline int search_ioctl(struct inode *inode,
 		key.offset = (u64)-1;
 		root = btrfs_read_fs_root_no_name(info, &key);
 		if (IS_ERR(root)) {
-			printk(KERN_ERR "BTRFS: could not find root %llu\n",
+			btrfs_err(info, "could not find root %llu",
 			       sk->tree_id);
 			btrfs_free_path(path);
 			return -ENOENT;
@@ -2210,7 +2229,7 @@ static noinline int btrfs_search_path_in_tree(struct btrfs_fs_info *info,
 	key.offset = (u64)-1;
 	root = btrfs_read_fs_root_no_name(info, &key);
 	if (IS_ERR(root)) {
-		printk(KERN_ERR "BTRFS: could not find root %llu\n", tree_id);
+		btrfs_err(info, "could not find root %llu", tree_id);
 		ret = -ENOENT;
 		goto out;
 	}
@@ -2271,10 +2290,7 @@ static noinline int btrfs_ioctl_ino_lookup(struct file *file,
 {
 	 struct btrfs_ioctl_ino_lookup_args *args;
 	 struct inode *inode;
-	 int ret;
-
-	if (!capable(CAP_SYS_ADMIN))
-		return -EPERM;
+	int ret = 0;
 
 	args = memdup_user(argp, sizeof(*args));
 	if (IS_ERR(args))
@@ -2282,13 +2298,28 @@ static noinline int btrfs_ioctl_ino_lookup(struct file *file,
 
 	inode = file_inode(file);
 
+	/*
+	 * Unprivileged query to obtain the containing subvolume root id. The
+	 * path is reset so it's consistent with btrfs_search_path_in_tree.
+	 */
 	if (args->treeid == 0)
 		args->treeid = BTRFS_I(inode)->root->root_key.objectid;
 
+	if (args->objectid == BTRFS_FIRST_FREE_OBJECTID) {
+		args->name[0] = 0;
+		goto out;
+	}
+
+	if (!capable(CAP_SYS_ADMIN)) {
+		ret = -EPERM;
+		goto out;
+	}
+
 	ret = btrfs_search_path_in_tree(BTRFS_I(inode)->root->fs_info,
 					args->treeid, args->objectid,
 					args->name);
 
+out:
 	if (ret == 0 && copy_to_user(argp, args, sizeof(*args)))
 		ret = -EFAULT;
 
@@ -2676,7 +2707,6 @@ static long btrfs_ioctl_fs_info(struct btrfs_root *root, void __user *arg)
 {
 	struct btrfs_ioctl_fs_info_args *fi_args;
 	struct btrfs_device *device;
-	struct btrfs_device *next;
 	struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices;
 	int ret = 0;
 
@@ -2688,7 +2718,7 @@ static long btrfs_ioctl_fs_info(struct btrfs_root *root, void __user *arg)
 	fi_args->num_devices = fs_devices->num_devices;
 	memcpy(&fi_args->fsid, root->fs_info->fsid, sizeof(fi_args->fsid));
 
-	list_for_each_entry_safe(device, next, &fs_devices->devices, dev_list) {
+	list_for_each_entry(device, &fs_devices->devices, dev_list) {
 		if (device->devid > fi_args->max_id)
 			fi_args->max_id = device->devid;
 	}
@@ -2753,38 +2783,68 @@ out:
 	return ret;
 }
 
-static struct page *extent_same_get_page(struct inode *inode, u64 off)
+static struct page *extent_same_get_page(struct inode *inode, pgoff_t index)
 {
 	struct page *page;
-	pgoff_t index;
-	struct extent_io_tree *tree = &BTRFS_I(inode)->io_tree;
-
-	index = off >> PAGE_CACHE_SHIFT;
 
 	page = grab_cache_page(inode->i_mapping, index);
 	if (!page)
-		return NULL;
+		return ERR_PTR(-ENOMEM);
 
 	if (!PageUptodate(page)) {
-		if (extent_read_full_page_nolock(tree, page, btrfs_get_extent,
-						 0))
-			return NULL;
+		int ret;
+
+		ret = btrfs_readpage(NULL, page);
+		if (ret)
+			return ERR_PTR(ret);
 		lock_page(page);
 		if (!PageUptodate(page)) {
 			unlock_page(page);
 			page_cache_release(page);
-			return NULL;
+			return ERR_PTR(-EIO);
+		}
+		if (page->mapping != inode->i_mapping) {
+			unlock_page(page);
+			page_cache_release(page);
+			return ERR_PTR(-EAGAIN);
 		}
 	}
-	unlock_page(page);
 
 	return page;
 }
 
-static inline void lock_extent_range(struct inode *inode, u64 off, u64 len)
+static int gather_extent_pages(struct inode *inode, struct page **pages,
+			       int num_pages, u64 off)
+{
+	int i;
+	pgoff_t index = off >> PAGE_CACHE_SHIFT;
+
+	for (i = 0; i < num_pages; i++) {
+again:
+		pages[i] = extent_same_get_page(inode, index + i);
+		if (IS_ERR(pages[i])) {
+			int err = PTR_ERR(pages[i]);
+
+			if (err == -EAGAIN)
+				goto again;
+			pages[i] = NULL;
+			return err;
+		}
+	}
+	return 0;
+}
+
+static int lock_extent_range(struct inode *inode, u64 off, u64 len,
+			     bool retry_range_locking)
 {
-	/* do any pending delalloc/csum calc on src, one way or
-	   another, and lock file content */
+	/*
+	 * Do any pending delalloc/csum calculations on inode, one way or
+	 * another, and lock file content.
+	 * The locking order is:
+	 *
+	 *   1) pages
+	 *   2) range in the inode's io tree
+	 */
 	while (1) {
 		struct btrfs_ordered_extent *ordered;
 		lock_extent(&BTRFS_I(inode)->io_tree, off, off + len - 1);
@@ -2802,56 +2862,140 @@ static inline void lock_extent_range(struct inode *inode, u64 off, u64 len)
 		unlock_extent(&BTRFS_I(inode)->io_tree, off, off + len - 1);
 		if (ordered)
 			btrfs_put_ordered_extent(ordered);
+		if (!retry_range_locking)
+			return -EAGAIN;
 		btrfs_wait_ordered_range(inode, off, len);
 	}
+	return 0;
 }
 
-static void btrfs_double_unlock(struct inode *inode1, u64 loff1,
-				struct inode *inode2, u64 loff2, u64 len)
+static void btrfs_double_inode_unlock(struct inode *inode1, struct inode *inode2)
 {
-	unlock_extent(&BTRFS_I(inode1)->io_tree, loff1, loff1 + len - 1);
-	unlock_extent(&BTRFS_I(inode2)->io_tree, loff2, loff2 + len - 1);
-
 	mutex_unlock(&inode1->i_mutex);
 	mutex_unlock(&inode2->i_mutex);
 }
 
-static void btrfs_double_lock(struct inode *inode1, u64 loff1,
-			      struct inode *inode2, u64 loff2, u64 len)
+static void btrfs_double_inode_lock(struct inode *inode1, struct inode *inode2)
 {
+	if (inode1 < inode2)
+		swap(inode1, inode2);
+
+	mutex_lock_nested(&inode1->i_mutex, I_MUTEX_PARENT);
+	mutex_lock_nested(&inode2->i_mutex, I_MUTEX_CHILD);
+}
+
+static void btrfs_double_extent_unlock(struct inode *inode1, u64 loff1,
+				      struct inode *inode2, u64 loff2, u64 len)
+{
+	unlock_extent(&BTRFS_I(inode1)->io_tree, loff1, loff1 + len - 1);
+	unlock_extent(&BTRFS_I(inode2)->io_tree, loff2, loff2 + len - 1);
+}
+
+static int btrfs_double_extent_lock(struct inode *inode1, u64 loff1,
+				    struct inode *inode2, u64 loff2, u64 len,
+				    bool retry_range_locking)
+{
+	int ret;
+
 	if (inode1 < inode2) {
 		swap(inode1, inode2);
 		swap(loff1, loff2);
 	}
+	ret = lock_extent_range(inode1, loff1, len, retry_range_locking);
+	if (ret)
+		return ret;
+	ret = lock_extent_range(inode2, loff2, len, retry_range_locking);
+	if (ret)
+		unlock_extent(&BTRFS_I(inode1)->io_tree, loff1,
+			      loff1 + len - 1);
+	return ret;
+}
 
-	mutex_lock_nested(&inode1->i_mutex, I_MUTEX_PARENT);
-	lock_extent_range(inode1, loff1, len);
-	if (inode1 != inode2) {
-		mutex_lock_nested(&inode2->i_mutex, I_MUTEX_CHILD);
-		lock_extent_range(inode2, loff2, len);
+struct cmp_pages {
+	int		num_pages;
+	struct page	**src_pages;
+	struct page	**dst_pages;
+};
+
+static void btrfs_cmp_data_free(struct cmp_pages *cmp)
+{
+	int i;
+	struct page *pg;
+
+	for (i = 0; i < cmp->num_pages; i++) {
+		pg = cmp->src_pages[i];
+		if (pg) {
+			unlock_page(pg);
+			page_cache_release(pg);
+		}
+		pg = cmp->dst_pages[i];
+		if (pg) {
+			unlock_page(pg);
+			page_cache_release(pg);
+		}
 	}
+	kfree(cmp->src_pages);
+	kfree(cmp->dst_pages);
+}
+
+static int btrfs_cmp_data_prepare(struct inode *src, u64 loff,
+				  struct inode *dst, u64 dst_loff,
+				  u64 len, struct cmp_pages *cmp)
+{
+	int ret;
+	int num_pages = PAGE_CACHE_ALIGN(len) >> PAGE_CACHE_SHIFT;
+	struct page **src_pgarr, **dst_pgarr;
+
+	/*
+	 * We must gather up all the pages before we initiate our
+	 * extent locking. We use an array for the page pointers. Size
+	 * of the array is bounded by len, which is in turn bounded by
+	 * BTRFS_MAX_DEDUPE_LEN.
+	 */
+	src_pgarr = kzalloc(num_pages * sizeof(struct page *), GFP_NOFS);
+	dst_pgarr = kzalloc(num_pages * sizeof(struct page *), GFP_NOFS);
+	if (!src_pgarr || !dst_pgarr) {
+		kfree(src_pgarr);
+		kfree(dst_pgarr);
+		return -ENOMEM;
+	}
+	cmp->num_pages = num_pages;
+	cmp->src_pages = src_pgarr;
+	cmp->dst_pages = dst_pgarr;
+
+	ret = gather_extent_pages(src, cmp->src_pages, cmp->num_pages, loff);
+	if (ret)
+		goto out;
+
+	ret = gather_extent_pages(dst, cmp->dst_pages, cmp->num_pages, dst_loff);
+
+out:
+	if (ret)
+		btrfs_cmp_data_free(cmp);
+	return 0;
 }
 
 static int btrfs_cmp_data(struct inode *src, u64 loff, struct inode *dst,
-			  u64 dst_loff, u64 len)
+			  u64 dst_loff, u64 len, struct cmp_pages *cmp)
 {
 	int ret = 0;
+	int i;
 	struct page *src_page, *dst_page;
 	unsigned int cmp_len = PAGE_CACHE_SIZE;
 	void *addr, *dst_addr;
 
+	i = 0;
 	while (len) {
 		if (len < PAGE_CACHE_SIZE)
 			cmp_len = len;
 
-		src_page = extent_same_get_page(src, loff);
-		if (!src_page)
-			return -EINVAL;
-		dst_page = extent_same_get_page(dst, dst_loff);
-		if (!dst_page) {
-			page_cache_release(src_page);
-			return -EINVAL;
-		}
+		BUG_ON(i >= cmp->num_pages);
+
+		src_page = cmp->src_pages[i];
+		dst_page = cmp->dst_pages[i];
+		ASSERT(PageLocked(src_page));
+		ASSERT(PageLocked(dst_page));
+
 		addr = kmap_atomic(src_page);
 		dst_addr = kmap_atomic(dst_page);
 
@@ -2863,26 +3007,30 @@ static int btrfs_cmp_data(struct inode *src, u64 loff, struct inode *dst,
 
 		kunmap_atomic(addr);
 		kunmap_atomic(dst_addr);
-		page_cache_release(src_page);
-		page_cache_release(dst_page);
 
 		if (ret)
 			break;
 
-		loff += cmp_len;
-		dst_loff += cmp_len;
 		len -= cmp_len;
+		i++;
 	}
 
 	return ret;
 }
 
-static int extent_same_check_offsets(struct inode *inode, u64 off, u64 len)
+static int extent_same_check_offsets(struct inode *inode, u64 off, u64 *plen,
+				     u64 olen)
 {
+	u64 len = *plen;
 	u64 bs = BTRFS_I(inode)->root->fs_info->sb->s_blocksize;
 
-	if (off + len > inode->i_size || off + len < off)
+	if (off + olen > inode->i_size || off + olen < off)
 		return -EINVAL;
+
+	/* if we extend to eof, continue to block boundary */
+	if (off + len == inode->i_size)
+		*plen = len = ALIGN(inode->i_size, bs) - off;
+
 	/* Check that we are block aligned - btrfs_clone() requires this */
 	if (!IS_ALIGNED(off, bs) || !IS_ALIGNED(off + len, bs))
 		return -EINVAL;
@@ -2890,31 +3038,67 @@ static int extent_same_check_offsets(struct inode *inode, u64 off, u64 len)
 	return 0;
 }
 
-static int btrfs_extent_same(struct inode *src, u64 loff, u64 len,
+static int btrfs_extent_same(struct inode *src, u64 loff, u64 olen,
 			     struct inode *dst, u64 dst_loff)
 {
 	int ret;
+	u64 len = olen;
+	struct cmp_pages cmp;
+	int same_inode = 0;
+	u64 same_lock_start = 0;
+	u64 same_lock_len = 0;
 
-	/*
-	 * btrfs_clone() can't handle extents in the same file
-	 * yet. Once that works, we can drop this check and replace it
-	 * with a check for the same inode, but overlapping extents.
-	 */
 	if (src == dst)
-		return -EINVAL;
+		same_inode = 1;
 
 	if (len == 0)
 		return 0;
 
-	btrfs_double_lock(src, loff, dst, dst_loff, len);
+	if (same_inode) {
+		mutex_lock(&src->i_mutex);
 
-	ret = extent_same_check_offsets(src, loff, len);
-	if (ret)
-		goto out_unlock;
+		ret = extent_same_check_offsets(src, loff, &len, olen);
+		if (ret)
+			goto out_unlock;
 
-	ret = extent_same_check_offsets(dst, dst_loff, len);
-	if (ret)
-		goto out_unlock;
+		/*
+		 * Single inode case wants the same checks, except we
+		 * don't want our length pushed out past i_size as
+		 * comparing that data range makes no sense.
+		 *
+		 * extent_same_check_offsets() will do this for an
+		 * unaligned length at i_size, so catch it here and
+		 * reject the request.
+		 *
+		 * This effectively means we require aligned extents
+		 * for the single-inode case, whereas the other cases
+		 * allow an unaligned length so long as it ends at
+		 * i_size.
+		 */
+		if (len != olen) {
+			ret = -EINVAL;
+			goto out_unlock;
+		}
+
+		/* Check for overlapping ranges */
+		if (dst_loff + len > loff && dst_loff < loff + len) {
+			ret = -EINVAL;
+			goto out_unlock;
+		}
+
+		same_lock_start = min_t(u64, loff, dst_loff);
+		same_lock_len = max_t(u64, loff, dst_loff) + len - same_lock_start;
+	} else {
+		btrfs_double_inode_lock(src, dst);
+
+		ret = extent_same_check_offsets(src, loff, &len, olen);
+		if (ret)
+			goto out_unlock;
+
+		ret = extent_same_check_offsets(dst, dst_loff, &len, olen);
+		if (ret)
+			goto out_unlock;
+	}
 
 	/* don't make the dst file partly checksummed */
 	if ((BTRFS_I(src)->flags & BTRFS_INODE_NODATASUM) !=
@@ -2923,12 +3107,64 @@ static int btrfs_extent_same(struct inode *src, u64 loff, u64 len,
 		goto out_unlock;
 	}
 
-	ret = btrfs_cmp_data(src, loff, dst, dst_loff, len);
+again:
+	ret = btrfs_cmp_data_prepare(src, loff, dst, dst_loff, olen, &cmp);
+	if (ret)
+		goto out_unlock;
+
+	if (same_inode)
+		ret = lock_extent_range(src, same_lock_start, same_lock_len,
+					false);
+	else
+		ret = btrfs_double_extent_lock(src, loff, dst, dst_loff, len,
+					       false);
+	/*
+	 * If one of the inodes has dirty pages in the respective range or
+	 * ordered extents, we need to flush dellaloc and wait for all ordered
+	 * extents in the range. We must unlock the pages and the ranges in the
+	 * io trees to avoid deadlocks when flushing delalloc (requires locking
+	 * pages) and when waiting for ordered extents to complete (they require
+	 * range locking).
+	 */
+	if (ret == -EAGAIN) {
+		/*
+		 * Ranges in the io trees already unlocked. Now unlock all
+		 * pages before waiting for all IO to complete.
+		 */
+		btrfs_cmp_data_free(&cmp);
+		if (same_inode) {
+			btrfs_wait_ordered_range(src, same_lock_start,
+						 same_lock_len);
+		} else {
+			btrfs_wait_ordered_range(src, loff, len);
+			btrfs_wait_ordered_range(dst, dst_loff, len);
+		}
+		goto again;
+	}
+	ASSERT(ret == 0);
+	if (WARN_ON(ret)) {
+		/* ranges in the io trees already unlocked */
+		btrfs_cmp_data_free(&cmp);
+		return ret;
+	}
+
+	/* pass original length for comparison so we stay within i_size */
+	ret = btrfs_cmp_data(src, loff, dst, dst_loff, olen, &cmp);
 	if (ret == 0)
-		ret = btrfs_clone(src, dst, loff, len, len, dst_loff);
+		ret = btrfs_clone(src, dst, loff, olen, len, dst_loff, 1);
 
+	if (same_inode)
+		unlock_extent(&BTRFS_I(src)->io_tree, same_lock_start,
+			      same_lock_start + same_lock_len - 1);
+	else
+		btrfs_double_extent_unlock(src, loff, dst, dst_loff, len);
+
+	btrfs_cmp_data_free(&cmp);
 out_unlock:
-	btrfs_double_unlock(src, loff, dst, dst_loff, len);
+	if (same_inode)
+		mutex_unlock(&src->i_mutex);
+	else
+		btrfs_double_inode_unlock(src, dst);
 
 	return ret;
 }
@@ -3043,52 +3279,19 @@ out:
 	return ret;
 }
 
-/* Helper to check and see if this root currently has a ref on the given disk
- * bytenr.  If it does then we need to update the quota for this root.  This
- * doesn't do anything if quotas aren't enabled.
- */
-static int check_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
-		     u64 disko)
-{
-	struct seq_list tree_mod_seq_elem = SEQ_LIST_INIT(tree_mod_seq_elem);
-	struct ulist *roots;
-	struct ulist_iterator uiter;
-	struct ulist_node *root_node = NULL;
-	int ret;
-
-	if (!root->fs_info->quota_enabled)
-		return 1;
-
-	btrfs_get_tree_mod_seq(root->fs_info, &tree_mod_seq_elem);
-	ret = btrfs_find_all_roots(trans, root->fs_info, disko,
-				   tree_mod_seq_elem.seq, &roots);
-	if (ret < 0)
-		goto out;
-	ret = 0;
-	ULIST_ITER_INIT(&uiter);
-	while ((root_node = ulist_next(roots, &uiter))) {
-		if (root_node->val == root->objectid) {
-			ret = 1;
-			break;
-		}
-	}
-	ulist_free(roots);
-out:
-	btrfs_put_tree_mod_seq(root->fs_info, &tree_mod_seq_elem);
-	return ret;
-}
-
 static int clone_finish_inode_update(struct btrfs_trans_handle *trans,
 				     struct inode *inode,
 				     u64 endoff,
 				     const u64 destoff,
-				     const u64 olen)
+				     const u64 olen,
+				     int no_time_update)
 {
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	int ret;
 
 	inode_inc_iversion(inode);
-	inode->i_mtime = inode->i_ctime = CURRENT_TIME;
+	if (!no_time_update)
+		inode->i_mtime = inode->i_ctime = CURRENT_TIME;
 	/*
 	 * We round up to the block size at eof when determining which
 	 * extents to clone above, but shouldn't round up the file size.
@@ -3166,6 +3369,150 @@ static void clone_update_extent_map(struct inode *inode,
 			&BTRFS_I(inode)->runtime_flags);
 }
 
+/*
+ * Make sure we do not end up inserting an inline extent into a file that has
+ * already other (non-inline) extents. If a file has an inline extent it can
+ * not have any other extents and the (single) inline extent must start at the
+ * file offset 0. Failing to respect these rules will lead to file corruption,
+ * resulting in EIO errors on read/write operations, hitting BUG_ON's in mm, etc
+ *
+ * We can have extents that have been already written to disk or we can have
+ * dirty ranges still in delalloc, in which case the extent maps and items are
+ * created only when we run delalloc, and the delalloc ranges might fall outside
+ * the range we are currently locking in the inode's io tree. So we check the
+ * inode's i_size because of that (i_size updates are done while holding the
+ * i_mutex, which we are holding here).
+ * We also check to see if the inode has a size not greater than "datal" but has
+ * extents beyond it, due to an fallocate with FALLOC_FL_KEEP_SIZE (and we are
+ * protected against such concurrent fallocate calls by the i_mutex).
+ *
+ * If the file has no extents but a size greater than datal, do not allow the
+ * copy because we would need turn the inline extent into a non-inline one (even
+ * with NO_HOLES enabled). If we find our destination inode only has one inline
+ * extent, just overwrite it with the source inline extent if its size is less
+ * than the source extent's size, or we could copy the source inline extent's
+ * data into the destination inode's inline extent if the later is greater then
+ * the former.
+ */
+static int clone_copy_inline_extent(struct inode *src,
+				    struct inode *dst,
+				    struct btrfs_trans_handle *trans,
+				    struct btrfs_path *path,
+				    struct btrfs_key *new_key,
+				    const u64 drop_start,
+				    const u64 datal,
+				    const u64 skip,
+				    const u64 size,
+				    char *inline_data)
+{
+	struct btrfs_root *root = BTRFS_I(dst)->root;
+	const u64 aligned_end = ALIGN(new_key->offset + datal,
+				      root->sectorsize);
+	int ret;
+	struct btrfs_key key;
+
+	if (new_key->offset > 0)
+		return -EOPNOTSUPP;
+
+	key.objectid = btrfs_ino(dst);
+	key.type = BTRFS_EXTENT_DATA_KEY;
+	key.offset = 0;
+	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+	if (ret < 0) {
+		return ret;
+	} else if (ret > 0) {
+		if (path->slots[0] >= btrfs_header_nritems(path->nodes[0])) {
+			ret = btrfs_next_leaf(root, path);
+			if (ret < 0)
+				return ret;
+			else if (ret > 0)
+				goto copy_inline_extent;
+		}
+		btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
+		if (key.objectid == btrfs_ino(dst) &&
+		    key.type == BTRFS_EXTENT_DATA_KEY) {
+			ASSERT(key.offset > 0);
+			return -EOPNOTSUPP;
+		}
+	} else if (i_size_read(dst) <= datal) {
+		struct btrfs_file_extent_item *ei;
+		u64 ext_len;
+
+		/*
+		 * If the file size is <= datal, make sure there are no other
+		 * extents following (can happen do to an fallocate call with
+		 * the flag FALLOC_FL_KEEP_SIZE).
+		 */
+		ei = btrfs_item_ptr(path->nodes[0], path->slots[0],
+				    struct btrfs_file_extent_item);
+		/*
+		 * If it's an inline extent, it can not have other extents
+		 * following it.
+		 */
+		if (btrfs_file_extent_type(path->nodes[0], ei) ==
+		    BTRFS_FILE_EXTENT_INLINE)
+			goto copy_inline_extent;
+
+		ext_len = btrfs_file_extent_num_bytes(path->nodes[0], ei);
+		if (ext_len > aligned_end)
+			return -EOPNOTSUPP;
+
+		ret = btrfs_next_item(root, path);
+		if (ret < 0) {
+			return ret;
+		} else if (ret == 0) {
+			btrfs_item_key_to_cpu(path->nodes[0], &key,
+					      path->slots[0]);
+			if (key.objectid == btrfs_ino(dst) &&
+			    key.type == BTRFS_EXTENT_DATA_KEY)
+				return -EOPNOTSUPP;
+		}
+	}
+
+copy_inline_extent:
+	/*
+	 * We have no extent items, or we have an extent at offset 0 which may
+	 * or may not be inlined. All these cases are dealt the same way.
+	 */
+	if (i_size_read(dst) > datal) {
+		/*
+		 * If the destination inode has an inline extent...
+		 * This would require copying the data from the source inline
+		 * extent into the beginning of the destination's inline extent.
+		 * But this is really complex, both extents can be compressed
+		 * or just one of them, which would require decompressing and
+		 * re-compressing data (which could increase the new compressed
+		 * size, not allowing the compressed data to fit anymore in an
+		 * inline extent).
+		 * So just don't support this case for now (it should be rare,
+		 * we are not really saving space when cloning inline extents).
+		 */
+		return -EOPNOTSUPP;
+	}
+
+	btrfs_release_path(path);
+	ret = btrfs_drop_extents(trans, root, dst, drop_start, aligned_end, 1);
+	if (ret)
+		return ret;
+	ret = btrfs_insert_empty_item(trans, root, path, new_key, size);
+	if (ret)
+		return ret;
+
+	if (skip) {
+		const u32 start = btrfs_file_extent_calc_inline_size(0);
+
+		memmove(inline_data + start, inline_data + start + skip, datal);
+	}
+
+	write_extent_buffer(path->nodes[0], inline_data,
+			    btrfs_item_ptr_offset(path->nodes[0],
+						  path->slots[0]),
+			    size);
+	inode_add_bytes(dst, datal);
+
+	return 0;
+}
+
 /**
  * btrfs_clone() - clone a range from inode file to another
  *
@@ -3173,13 +3520,13 @@ static void clone_update_extent_map(struct inode *inode,
  * @inode: Inode to clone to
  * @off: Offset within source to start clone from
  * @olen: Original length, passed by user, of range to clone
- * @olen_aligned: Block-aligned value of olen, extent_same uses
- *               identical values here
+ * @olen_aligned: Block-aligned value of olen
  * @destoff: Offset within @inode to start clone
+ * @no_time_update: Whether to update mtime/ctime on the target inode
  */
 static int btrfs_clone(struct inode *src, struct inode *inode,
 		       const u64 off, const u64 olen, const u64 olen_aligned,
-		       const u64 destoff)
+		       const u64 destoff, int no_time_update)
 {
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	struct btrfs_path *path = NULL;
@@ -3190,9 +3537,7 @@ static int btrfs_clone(struct inode *src, struct inode *inode,
 	u32 nritems;
 	int slot;
 	int ret;
-	int no_quota;
 	const u64 len = olen_aligned;
-	u64 last_disko = 0;
 	u64 last_dest_end = destoff;
 
 	ret = -ENOMEM;
@@ -3238,7 +3583,6 @@ static int btrfs_clone(struct inode *src, struct inode *inode,
 
 		nritems = btrfs_header_nritems(path->nodes[0]);
 process_slot:
-		no_quota = 1;
 		if (path->slots[0] >= nritems) {
 			ret = btrfs_next_leaf(BTRFS_I(src)->root, path);
 			if (ret < 0)
@@ -3390,35 +3734,13 @@ process_slot:
 				btrfs_set_file_extent_num_bytes(leaf, extent,
 								datal);
 
-				/*
-				 * We need to look up the roots that point at
-				 * this bytenr and see if the new root does.  If
-				 * it does not we need to make sure we update
-				 * quotas appropriately.
-				 */
-				if (disko && root != BTRFS_I(src)->root &&
-				    disko != last_disko) {
-					no_quota = check_ref(trans, root,
-							     disko);
-					if (no_quota < 0) {
-						btrfs_abort_transaction(trans,
-									root,
-									ret);
-						btrfs_end_transaction(trans,
-								      root);
-						ret = no_quota;
-						goto out;
-					}
-				}
-
 				if (disko) {
 					inode_add_bytes(inode, datal);
 					ret = btrfs_inc_extent_ref(trans, root,
 							disko, diskl, 0,
 							root->root_key.objectid,
 							btrfs_ino(inode),
-							new_key.offset - datao,
-							no_quota);
+							new_key.offset - datao);
 					if (ret) {
 						btrfs_abort_transaction(trans,
 									root,
@@ -3432,21 +3754,6 @@ process_slot:
 			} else if (type == BTRFS_FILE_EXTENT_INLINE) {
 				u64 skip = 0;
 				u64 trim = 0;
-				u64 aligned_end = 0;
-
-				/*
-				 * Don't copy an inline extent into an offset
-				 * greater than zero. Having an inline extent
-				 * at such an offset results in chaos as btrfs
-				 * isn't prepared for such cases. Just skip
-				 * this case for the same reasons as commented
-				 * at btrfs_ioctl_clone().
-				 */
-				if (last_dest_end > 0) {
-					ret = -EOPNOTSUPP;
-					btrfs_end_transaction(trans, root);
-					goto out;
-				}
 
 				if (off > key.offset) {
 					skip = off - key.offset;
@@ -3464,42 +3771,22 @@ process_slot:
 				size -= skip + trim;
 				datal -= skip + trim;
 
-				aligned_end = ALIGN(new_key.offset + datal,
-						    root->sectorsize);
-				ret = btrfs_drop_extents(trans, root, inode,
-							 drop_start,
-							 aligned_end,
-							 1);
+				ret = clone_copy_inline_extent(src, inode,
+							       trans, path,
+							       &new_key,
+							       drop_start,
+							       datal,
+							       skip, size, buf);
 				if (ret) {
 					if (ret != -EOPNOTSUPP)
 						btrfs_abort_transaction(trans,
-							root, ret);
-					btrfs_end_transaction(trans, root);
-					goto out;
-				}
-
-				ret = btrfs_insert_empty_item(trans, root, path,
-							      &new_key, size);
-				if (ret) {
-					btrfs_abort_transaction(trans, root,
-								ret);
+									root,
+									ret);
 					btrfs_end_transaction(trans, root);
 					goto out;
 				}
-
-				if (skip) {
-					u32 start =
-					  btrfs_file_extent_calc_inline_size(0);
-					memmove(buf+start, buf+start+skip,
-						datal);
-				}
-
 				leaf = path->nodes[0];
 				slot = path->slots[0];
-				write_extent_buffer(leaf, buf,
-					    btrfs_item_ptr_offset(leaf, slot),
-					    size);
-				inode_add_bytes(inode, datal);
 			}
 
 			/* If we have an implicit hole (NO_HOLES feature). */
@@ -3517,7 +3804,8 @@ process_slot:
 					      root->sectorsize);
 			ret = clone_finish_inode_update(trans, inode,
 							last_dest_end,
-							destoff, olen);
+							destoff, olen,
+							no_time_update);
 			if (ret)
 				goto out;
 			if (new_key.offset + datal >= destoff + len)
@@ -3555,7 +3843,7 @@ process_slot:
 		clone_update_extent_map(inode, trans, NULL, last_dest_end,
 					destoff + len - last_dest_end);
 		ret = clone_finish_inode_update(trans, inode, destoff + len,
-						destoff, olen);
+						destoff, olen, no_time_update);
 	}
 
 out:
@@ -3632,13 +3920,7 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
 		goto out_fput;
 
 	if (!same_inode) {
-		if (inode < src) {
-			mutex_lock_nested(&inode->i_mutex, I_MUTEX_PARENT);
-			mutex_lock_nested(&src->i_mutex, I_MUTEX_CHILD);
-		} else {
-			mutex_lock_nested(&src->i_mutex, I_MUTEX_PARENT);
-			mutex_lock_nested(&inode->i_mutex, I_MUTEX_CHILD);
-		}
+		btrfs_double_inode_lock(src, inode);
 	} else {
 		mutex_lock(&src->i_mutex);
 	}
@@ -3686,13 +3968,18 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
 		u64 lock_start = min_t(u64, off, destoff);
 		u64 lock_len = max_t(u64, off, destoff) + len - lock_start;
 
-		lock_extent_range(src, lock_start, lock_len);
+		ret = lock_extent_range(src, lock_start, lock_len, true);
 	} else {
-		lock_extent_range(src, off, len);
-		lock_extent_range(inode, destoff, len);
+		ret = btrfs_double_extent_lock(src, off, inode, destoff, len,
+					       true);
+	}
+	ASSERT(ret == 0);
+	if (WARN_ON(ret)) {
+		/* ranges in the io trees already unlocked */
+		goto out_unlock;
 	}
 
-	ret = btrfs_clone(src, inode, off, olen, len, destoff);
+	ret = btrfs_clone(src, inode, off, olen, len, destoff, 0);
 
 	if (same_inode) {
 		u64 lock_start = min_t(u64, off, destoff);
@@ -3700,9 +3987,7 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
 
 		unlock_extent(&BTRFS_I(src)->io_tree, lock_start, lock_end);
 	} else {
-		unlock_extent(&BTRFS_I(src)->io_tree, off, off + len - 1);
-		unlock_extent(&BTRFS_I(inode)->io_tree, destoff,
-			      destoff + len - 1);
+		btrfs_double_extent_unlock(src, off, inode, destoff, len);
 	}
 	/*
 	 * Truncate page cache pages so that future reads will see the cloned
@@ -3711,17 +3996,10 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
 	truncate_inode_pages_range(&inode->i_data, destoff,
 				   PAGE_CACHE_ALIGN(destoff + len) - 1);
 out_unlock:
-	if (!same_inode) {
-		if (inode < src) {
-			mutex_unlock(&src->i_mutex);
-			mutex_unlock(&inode->i_mutex);
-		} else {
-			mutex_unlock(&inode->i_mutex);
-			mutex_unlock(&src->i_mutex);
-		}
-	} else {
+	if (!same_inode)
+		btrfs_double_inode_unlock(src, inode);
+	else
 		mutex_unlock(&src->i_mutex);
-	}
 out_fput:
 	fdput(src_file);
 out_drop_write:
@@ -4492,6 +4770,11 @@ locked:
 		bctl->flags |= BTRFS_BALANCE_TYPE_MASK;
 	}
 
+	if (bctl->flags & ~(BTRFS_BALANCE_ARGS_MASK | BTRFS_BALANCE_TYPE_MASK)) {
+		ret = -EINVAL;
+		goto out_bctl;
+	}
+
 do_balance:
 	/*
 	 * Ownership of bctl and mutually_exclusive_operation_running
@@ -4503,12 +4786,15 @@ do_balance:
 	need_unlock = false;
 
 	ret = btrfs_balance(bctl, bargs);
+	bctl = NULL;
 
 	if (arg) {
 		if (copy_to_user(arg, bargs, sizeof(*bargs)))
 			ret = -EFAULT;
 	}
 
+out_bctl:
+	kfree(bctl);
 out_bargs:
 	kfree(bargs);
 out_unlock:
@@ -4659,7 +4945,7 @@ static long btrfs_ioctl_qgroup_assign(struct file *file, void __user *arg)
 	/* update qgroup status and info */
 	err = btrfs_run_qgroups(trans, root->fs_info);
 	if (err < 0)
-		btrfs_error(root->fs_info, ret,
+		btrfs_std_error(root->fs_info, ret,
 			    "failed to update qgroup status and info\n");
 	err = btrfs_end_transaction(trans, root);
 	if (err && !ret)
diff --git a/kernel/fs/btrfs/locking.c b/kernel/fs/btrfs/locking.c
index f8229ef1b..8077461fc 100644
--- a/kernel/fs/btrfs/locking.c
+++ b/kernel/fs/btrfs/locking.c
@@ -79,6 +79,9 @@ void btrfs_clear_lock_blocking_rw(struct extent_buffer *eb, int rw)
 		write_lock(&eb->lock);
 		WARN_ON(atomic_read(&eb->spinning_writers));
 		atomic_inc(&eb->spinning_writers);
+		/*
+		 * atomic_dec_and_test implies a barrier for waitqueue_active
+		 */
 		if (atomic_dec_and_test(&eb->blocking_writers) &&
 		    waitqueue_active(&eb->write_lock_wq))
 			wake_up(&eb->write_lock_wq);
@@ -86,6 +89,9 @@ void btrfs_clear_lock_blocking_rw(struct extent_buffer *eb, int rw)
 		BUG_ON(atomic_read(&eb->blocking_readers) == 0);
 		read_lock(&eb->lock);
 		atomic_inc(&eb->spinning_readers);
+		/*
+		 * atomic_dec_and_test implies a barrier for waitqueue_active
+		 */
 		if (atomic_dec_and_test(&eb->blocking_readers) &&
 		    waitqueue_active(&eb->read_lock_wq))
 			wake_up(&eb->read_lock_wq);
@@ -229,6 +235,9 @@ void btrfs_tree_read_unlock_blocking(struct extent_buffer *eb)
 	}
 	btrfs_assert_tree_read_locked(eb);
 	WARN_ON(atomic_read(&eb->blocking_readers) == 0);
+	/*
+	 * atomic_dec_and_test implies a barrier for waitqueue_active
+	 */
 	if (atomic_dec_and_test(&eb->blocking_readers) &&
 	    waitqueue_active(&eb->read_lock_wq))
 		wake_up(&eb->read_lock_wq);
@@ -241,6 +250,7 @@ void btrfs_tree_read_unlock_blocking(struct extent_buffer *eb)
  */
 void btrfs_tree_lock(struct extent_buffer *eb)
 {
+	WARN_ON(eb->lock_owner == current->pid);
 again:
 	wait_event(eb->read_lock_wq, atomic_read(&eb->blocking_readers) == 0);
 	wait_event(eb->write_lock_wq, atomic_read(&eb->blocking_writers) == 0);
@@ -279,6 +289,9 @@ void btrfs_tree_unlock(struct extent_buffer *eb)
 	if (blockers) {
 		WARN_ON(atomic_read(&eb->spinning_writers));
 		atomic_dec(&eb->blocking_writers);
+		/*
+		 * Make sure counter is updated before we wake up waiters.
+		 */
 		smp_mb();
 		if (waitqueue_active(&eb->write_lock_wq))
 			wake_up(&eb->write_lock_wq);
diff --git a/kernel/fs/btrfs/ordered-data.c b/kernel/fs/btrfs/ordered-data.c
index 760c4a5e0..8c27292ea 100644
--- a/kernel/fs/btrfs/ordered-data.c
+++ b/kernel/fs/btrfs/ordered-data.c
@@ -198,9 +198,6 @@ static int __btrfs_add_ordered_extent(struct inode *inode, u64 file_offset,
 	entry->file_offset = file_offset;
 	entry->start = start;
 	entry->len = len;
-	if (!(BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM) &&
-	    !(type == BTRFS_ORDERED_NOCOW))
-		entry->csum_bytes_left = disk_len;
 	entry->disk_len = disk_len;
 	entry->bytes_left = len;
 	entry->inode = igrab(inode);
@@ -286,10 +283,6 @@ void btrfs_add_ordered_sum(struct inode *inode,
 	tree = &BTRFS_I(inode)->ordered_tree;
 	spin_lock_irq(&tree->lock);
 	list_add_tail(&sum->list, &entry->list);
-	WARN_ON(entry->csum_bytes_left < sum->len);
-	entry->csum_bytes_left -= sum->len;
-	if (entry->csum_bytes_left == 0)
-		wake_up(&entry->wait);
 	spin_unlock_irq(&tree->lock);
 }
 
@@ -352,6 +345,9 @@ int btrfs_dec_test_first_ordered_pending(struct inode *inode,
 
 	if (entry->bytes_left == 0) {
 		ret = test_and_set_bit(BTRFS_ORDERED_IO_DONE, &entry->flags);
+		/*
+		 * Implicit memory barrier after test_and_set_bit
+		 */
 		if (waitqueue_active(&entry->wait))
 			wake_up(&entry->wait);
 	} else {
@@ -416,6 +412,9 @@ have_entry:
 
 	if (entry->bytes_left == 0) {
 		ret = test_and_set_bit(BTRFS_ORDERED_IO_DONE, &entry->flags);
+		/*
+		 * Implicit memory barrier after test_and_set_bit
+		 */
 		if (waitqueue_active(&entry->wait))
 			wake_up(&entry->wait);
 	} else {
@@ -491,15 +490,16 @@ void btrfs_wait_logged_extents(struct btrfs_trans_handle *trans,
 
 	spin_lock_irq(&log->log_extents_lock[index]);
 	while (!list_empty(&log->logged_list[index])) {
+		struct inode *inode;
 		ordered = list_first_entry(&log->logged_list[index],
 					   struct btrfs_ordered_extent,
 					   log_list);
 		list_del_init(&ordered->log_list);
+		inode = ordered->inode;
 		spin_unlock_irq(&log->log_extents_lock[index]);
 
 		if (!test_bit(BTRFS_ORDERED_IO_DONE, &ordered->flags) &&
 		    !test_bit(BTRFS_ORDERED_DIRECT, &ordered->flags)) {
-			struct inode *inode = ordered->inode;
 			u64 start = ordered->file_offset;
 			u64 end = ordered->file_offset + ordered->len - 1;
 
@@ -509,7 +509,26 @@ void btrfs_wait_logged_extents(struct btrfs_trans_handle *trans,
 		wait_event(ordered->wait, test_bit(BTRFS_ORDERED_IO_DONE,
 						   &ordered->flags));
 
-		list_add_tail(&ordered->trans_list, &trans->ordered);
+		/*
+		 * In order to keep us from losing our ordered extent
+		 * information when committing the transaction we have to make
+		 * sure that any logged extents are completed when we go to
+		 * commit the transaction.  To do this we simply increase the
+		 * current transactions pending_ordered counter and decrement it
+		 * when the ordered extent completes.
+		 */
+		if (!test_bit(BTRFS_ORDERED_COMPLETE, &ordered->flags)) {
+			struct btrfs_ordered_inode_tree *tree;
+
+			tree = &BTRFS_I(inode)->ordered_tree;
+			spin_lock_irq(&tree->lock);
+			if (!test_bit(BTRFS_ORDERED_COMPLETE, &ordered->flags)) {
+				set_bit(BTRFS_ORDERED_PENDING, &ordered->flags);
+				atomic_inc(&trans->transaction->pending_ordered);
+			}
+			spin_unlock_irq(&tree->lock);
+		}
+		btrfs_put_ordered_extent(ordered);
 		spin_lock_irq(&log->log_extents_lock[index]);
 	}
 	spin_unlock_irq(&log->log_extents_lock[index]);
@@ -545,6 +564,10 @@ void btrfs_put_ordered_extent(struct btrfs_ordered_extent *entry)
 	trace_btrfs_ordered_extent_put(entry->inode, entry);
 
 	if (atomic_dec_and_test(&entry->refs)) {
+		ASSERT(list_empty(&entry->log_list));
+		ASSERT(list_empty(&entry->trans_list));
+		ASSERT(list_empty(&entry->root_extent_list));
+		ASSERT(RB_EMPTY_NODE(&entry->rb_node));
 		if (entry->inode)
 			btrfs_add_delayed_iput(entry->inode);
 		while (!list_empty(&entry->list)) {
@@ -567,16 +590,47 @@ void btrfs_remove_ordered_extent(struct inode *inode,
 	struct btrfs_ordered_inode_tree *tree;
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	struct rb_node *node;
+	bool dec_pending_ordered = false;
 
 	tree = &BTRFS_I(inode)->ordered_tree;
 	spin_lock_irq(&tree->lock);
 	node = &entry->rb_node;
 	rb_erase(node, &tree->tree);
+	RB_CLEAR_NODE(node);
 	if (tree->last == node)
 		tree->last = NULL;
 	set_bit(BTRFS_ORDERED_COMPLETE, &entry->flags);
+	if (test_and_clear_bit(BTRFS_ORDERED_PENDING, &entry->flags))
+		dec_pending_ordered = true;
 	spin_unlock_irq(&tree->lock);
 
+	/*
+	 * The current running transaction is waiting on us, we need to let it
+	 * know that we're complete and wake it up.
+	 */
+	if (dec_pending_ordered) {
+		struct btrfs_transaction *trans;
+
+		/*
+		 * The checks for trans are just a formality, it should be set,
+		 * but if it isn't we don't want to deref/assert under the spin
+		 * lock, so be nice and check if trans is set, but ASSERT() so
+		 * if it isn't set a developer will notice.
+		 */
+		spin_lock(&root->fs_info->trans_lock);
+		trans = root->fs_info->running_transaction;
+		if (trans)
+			atomic_inc(&trans->use_count);
+		spin_unlock(&root->fs_info->trans_lock);
+
+		ASSERT(trans);
+		if (trans) {
+			if (atomic_dec_and_test(&trans->pending_ordered))
+				wake_up(&trans->pending_wait);
+			btrfs_put_transaction(trans);
+		}
+	}
+
 	spin_lock(&root->ordered_extent_lock);
 	list_del_init(&entry->root_extent_list);
 	root->nr_ordered_extents--;
@@ -844,6 +898,20 @@ out:
 	return entry;
 }
 
+bool btrfs_have_ordered_extents_in_range(struct inode *inode,
+					 u64 file_offset,
+					 u64 len)
+{
+	struct btrfs_ordered_extent *oe;
+
+	oe = btrfs_lookup_ordered_range(inode, file_offset, len);
+	if (oe) {
+		btrfs_put_ordered_extent(oe);
+		return true;
+	}
+	return false;
+}
+
 /*
  * lookup and return any extent before 'file_offset'.  NULL is returned
  * if none is found
diff --git a/kernel/fs/btrfs/ordered-data.h b/kernel/fs/btrfs/ordered-data.h
index e96cd4ccd..23c96059c 100644
--- a/kernel/fs/btrfs/ordered-data.h
+++ b/kernel/fs/btrfs/ordered-data.h
@@ -73,6 +73,8 @@ struct btrfs_ordered_sum {
 
 #define BTRFS_ORDERED_LOGGED 10 /* Set when we've waited on this ordered extent
 				 * in the logging code. */
+#define BTRFS_ORDERED_PENDING 11 /* We are waiting for this ordered extent to
+				  * complete in the current transaction. */
 struct btrfs_ordered_extent {
 	/* logical offset in the file */
 	u64 file_offset;
@@ -89,9 +91,6 @@ struct btrfs_ordered_extent {
 	/* number of bytes that still need writing */
 	u64 bytes_left;
 
-	/* number of bytes that still need csumming */
-	u64 csum_bytes_left;
-
 	/*
 	 * the end of the ordered extent which is behind it but
 	 * didn't update disk_i_size. Please see the comment of
@@ -191,6 +190,9 @@ btrfs_lookup_first_ordered_extent(struct inode * inode, u64 file_offset);
 struct btrfs_ordered_extent *btrfs_lookup_ordered_range(struct inode *inode,
 							u64 file_offset,
 							u64 len);
+bool btrfs_have_ordered_extents_in_range(struct inode *inode,
+					 u64 file_offset,
+					 u64 len);
 int btrfs_ordered_update_i_size(struct inode *inode, u64 offset,
 				struct btrfs_ordered_extent *ordered);
 int btrfs_find_ordered_sum(struct inode *inode, u64 offset, u64 disk_bytenr,
diff --git a/kernel/fs/btrfs/props.c b/kernel/fs/btrfs/props.c
index dca137b04..f9e60231f 100644
--- a/kernel/fs/btrfs/props.c
+++ b/kernel/fs/btrfs/props.c
@@ -49,18 +49,16 @@ static struct prop_handler prop_handlers[] = {
 		.extract = prop_compression_extract,
 		.inheritable = 1
 	},
-	{
-		.xattr_name = NULL
-	}
 };
 
 void __init btrfs_props_init(void)
 {
-	struct prop_handler *p;
+	int i;
 
 	hash_init(prop_handlers_ht);
 
-	for (p = &prop_handlers[0]; p->xattr_name; p++) {
+	for (i = 0; i < ARRAY_SIZE(prop_handlers); i++) {
+		struct prop_handler *p = &prop_handlers[i];
 		u64 h = btrfs_name_hash(p->xattr_name, strlen(p->xattr_name));
 
 		hash_add(prop_handlers_ht, &p->node, h);
@@ -301,15 +299,16 @@ static int inherit_props(struct btrfs_trans_handle *trans,
 			 struct inode *inode,
 			 struct inode *parent)
 {
-	const struct prop_handler *h;
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	int ret;
+	int i;
 
 	if (!test_bit(BTRFS_INODE_HAS_PROPS,
 		      &BTRFS_I(parent)->runtime_flags))
 		return 0;
 
-	for (h = &prop_handlers[0]; h->xattr_name; h++) {
+	for (i = 0; i < ARRAY_SIZE(prop_handlers); i++) {
+		const struct prop_handler *h = &prop_handlers[i];
 		const char *value;
 		u64 num_bytes;
 
diff --git a/kernel/fs/btrfs/qgroup.c b/kernel/fs/btrfs/qgroup.c
index 3d6546581..5279fdae7 100644
--- a/kernel/fs/btrfs/qgroup.c
+++ b/kernel/fs/btrfs/qgroup.c
@@ -34,6 +34,7 @@
 #include "extent_io.h"
 #include "qgroup.h"
 
+
 /* TODO XXX FIXME
  *  - subvol delete -> delete when ref goes to 0? delete limits also?
  *  - reorganize keys
@@ -84,11 +85,42 @@ struct btrfs_qgroup {
 
 	/*
 	 * temp variables for accounting operations
+	 * Refer to qgroup_shared_accouting() for details.
 	 */
 	u64 old_refcnt;
 	u64 new_refcnt;
 };
 
+static void btrfs_qgroup_update_old_refcnt(struct btrfs_qgroup *qg, u64 seq,
+					   int mod)
+{
+	if (qg->old_refcnt < seq)
+		qg->old_refcnt = seq;
+	qg->old_refcnt += mod;
+}
+
+static void btrfs_qgroup_update_new_refcnt(struct btrfs_qgroup *qg, u64 seq,
+					   int mod)
+{
+	if (qg->new_refcnt < seq)
+		qg->new_refcnt = seq;
+	qg->new_refcnt += mod;
+}
+
+static inline u64 btrfs_qgroup_get_old_refcnt(struct btrfs_qgroup *qg, u64 seq)
+{
+	if (qg->old_refcnt < seq)
+		return 0;
+	return qg->old_refcnt - seq;
+}
+
+static inline u64 btrfs_qgroup_get_new_refcnt(struct btrfs_qgroup *qg, u64 seq)
+{
+	if (qg->new_refcnt < seq)
+		return 0;
+	return qg->new_refcnt - seq;
+}
+
 /*
  * glue structure to represent the relations between qgroups.
  */
@@ -344,7 +376,7 @@ int btrfs_read_qgroup_config(struct btrfs_fs_info *fs_info)
 		qgroup = find_qgroup_rb(fs_info, found_key.offset);
 		if ((qgroup && found_key.type == BTRFS_QGROUP_INFO_KEY) ||
 		    (!qgroup && found_key.type == BTRFS_QGROUP_LIMIT_KEY)) {
-			btrfs_err(fs_info, "inconsitent qgroup config");
+			btrfs_err(fs_info, "inconsistent qgroup config");
 			flags |= BTRFS_QGROUP_STATUS_FLAG_INCONSISTENT;
 		}
 		if (!qgroup) {
@@ -961,9 +993,10 @@ int btrfs_quota_disable(struct btrfs_trans_handle *trans,
 	mutex_lock(&fs_info->qgroup_ioctl_lock);
 	if (!fs_info->quota_root)
 		goto out;
-	spin_lock(&fs_info->qgroup_lock);
 	fs_info->quota_enabled = 0;
 	fs_info->pending_quota_state = 0;
+	btrfs_qgroup_wait_for_completion(fs_info);
+	spin_lock(&fs_info->qgroup_lock);
 	quota_root = fs_info->quota_root;
 	fs_info->quota_root = NULL;
 	fs_info->qgroup_flags &= ~BTRFS_QGROUP_STATUS_FLAG_ON;
@@ -1115,14 +1148,14 @@ int btrfs_add_qgroup_relation(struct btrfs_trans_handle *trans,
 	struct ulist *tmp;
 	int ret = 0;
 
-	tmp = ulist_alloc(GFP_NOFS);
-	if (!tmp)
-		return -ENOMEM;
-
 	/* Check the level of src and dst first */
 	if (btrfs_qgroup_level(src) >= btrfs_qgroup_level(dst))
 		return -EINVAL;
 
+	tmp = ulist_alloc(GFP_NOFS);
+	if (!tmp)
+		return -ENOMEM;
+
 	mutex_lock(&fs_info->qgroup_ioctl_lock);
 	quota_root = fs_info->quota_root;
 	if (!quota_root) {
@@ -1317,6 +1350,11 @@ int btrfs_limit_qgroup(struct btrfs_trans_handle *trans,
 	struct btrfs_root *quota_root;
 	struct btrfs_qgroup *qgroup;
 	int ret = 0;
+	/* Sometimes we would want to clear the limit on this qgroup.
+	 * To meet this requirement, we treat the -1 as a special value
+	 * which tell kernel to clear the limit on this qgroup.
+	 */
+	const u64 CLEAR_VALUE = -1;
 
 	mutex_lock(&fs_info->qgroup_ioctl_lock);
 	quota_root = fs_info->quota_root;
@@ -1332,14 +1370,42 @@ int btrfs_limit_qgroup(struct btrfs_trans_handle *trans,
 	}
 
 	spin_lock(&fs_info->qgroup_lock);
-	if (limit->flags & BTRFS_QGROUP_LIMIT_MAX_RFER)
-		qgroup->max_rfer = limit->max_rfer;
-	if (limit->flags & BTRFS_QGROUP_LIMIT_MAX_EXCL)
-		qgroup->max_excl = limit->max_excl;
-	if (limit->flags & BTRFS_QGROUP_LIMIT_RSV_RFER)
-		qgroup->rsv_rfer = limit->rsv_rfer;
-	if (limit->flags & BTRFS_QGROUP_LIMIT_RSV_EXCL)
-		qgroup->rsv_excl = limit->rsv_excl;
+	if (limit->flags & BTRFS_QGROUP_LIMIT_MAX_RFER) {
+		if (limit->max_rfer == CLEAR_VALUE) {
+			qgroup->lim_flags &= ~BTRFS_QGROUP_LIMIT_MAX_RFER;
+			limit->flags &= ~BTRFS_QGROUP_LIMIT_MAX_RFER;
+			qgroup->max_rfer = 0;
+		} else {
+			qgroup->max_rfer = limit->max_rfer;
+		}
+	}
+	if (limit->flags & BTRFS_QGROUP_LIMIT_MAX_EXCL) {
+		if (limit->max_excl == CLEAR_VALUE) {
+			qgroup->lim_flags &= ~BTRFS_QGROUP_LIMIT_MAX_EXCL;
+			limit->flags &= ~BTRFS_QGROUP_LIMIT_MAX_EXCL;
+			qgroup->max_excl = 0;
+		} else {
+			qgroup->max_excl = limit->max_excl;
+		}
+	}
+	if (limit->flags & BTRFS_QGROUP_LIMIT_RSV_RFER) {
+		if (limit->rsv_rfer == CLEAR_VALUE) {
+			qgroup->lim_flags &= ~BTRFS_QGROUP_LIMIT_RSV_RFER;
+			limit->flags &= ~BTRFS_QGROUP_LIMIT_RSV_RFER;
+			qgroup->rsv_rfer = 0;
+		} else {
+			qgroup->rsv_rfer = limit->rsv_rfer;
+		}
+	}
+	if (limit->flags & BTRFS_QGROUP_LIMIT_RSV_EXCL) {
+		if (limit->rsv_excl == CLEAR_VALUE) {
+			qgroup->lim_flags &= ~BTRFS_QGROUP_LIMIT_RSV_EXCL;
+			limit->flags &= ~BTRFS_QGROUP_LIMIT_RSV_EXCL;
+			qgroup->rsv_excl = 0;
+		} else {
+			qgroup->rsv_excl = limit->rsv_excl;
+		}
+	}
 	qgroup->lim_flags |= limit->flags;
 
 	spin_unlock(&fs_info->qgroup_lock);
@@ -1356,239 +1422,88 @@ out:
 	return ret;
 }
 
-static int comp_oper_exist(struct btrfs_qgroup_operation *oper1,
-			   struct btrfs_qgroup_operation *oper2)
+int btrfs_qgroup_prepare_account_extents(struct btrfs_trans_handle *trans,
+					 struct btrfs_fs_info *fs_info)
 {
-	/*
-	 * Ignore seq and type here, we're looking for any operation
-	 * at all related to this extent on that root.
-	 */
-	if (oper1->bytenr < oper2->bytenr)
-		return -1;
-	if (oper1->bytenr > oper2->bytenr)
-		return 1;
-	if (oper1->ref_root < oper2->ref_root)
-		return -1;
-	if (oper1->ref_root > oper2->ref_root)
-		return 1;
-	return 0;
-}
+	struct btrfs_qgroup_extent_record *record;
+	struct btrfs_delayed_ref_root *delayed_refs;
+	struct rb_node *node;
+	u64 qgroup_to_skip;
+	int ret = 0;
 
-static int qgroup_oper_exists(struct btrfs_fs_info *fs_info,
-			      struct btrfs_qgroup_operation *oper)
-{
-	struct rb_node *n;
-	struct btrfs_qgroup_operation *cur;
-	int cmp;
+	delayed_refs = &trans->transaction->delayed_refs;
+	qgroup_to_skip = delayed_refs->qgroup_to_skip;
 
-	spin_lock(&fs_info->qgroup_op_lock);
-	n = fs_info->qgroup_op_tree.rb_node;
-	while (n) {
-		cur = rb_entry(n, struct btrfs_qgroup_operation, n);
-		cmp = comp_oper_exist(cur, oper);
-		if (cmp < 0) {
-			n = n->rb_right;
-		} else if (cmp) {
-			n = n->rb_left;
-		} else {
-			spin_unlock(&fs_info->qgroup_op_lock);
-			return -EEXIST;
-		}
+	/*
+	 * No need to do lock, since this function will only be called in
+	 * btrfs_commmit_transaction().
+	 */
+	node = rb_first(&delayed_refs->dirty_extent_root);
+	while (node) {
+		record = rb_entry(node, struct btrfs_qgroup_extent_record,
+				  node);
+		ret = btrfs_find_all_roots(NULL, fs_info, record->bytenr, 0,
+					   &record->old_roots);
+		if (ret < 0)
+			break;
+		if (qgroup_to_skip)
+			ulist_del(record->old_roots, qgroup_to_skip, 0);
+		node = rb_next(node);
 	}
-	spin_unlock(&fs_info->qgroup_op_lock);
-	return 0;
+	return ret;
 }
 
-static int comp_oper(struct btrfs_qgroup_operation *oper1,
-		     struct btrfs_qgroup_operation *oper2)
+struct btrfs_qgroup_extent_record
+*btrfs_qgroup_insert_dirty_extent(struct btrfs_delayed_ref_root *delayed_refs,
+				  struct btrfs_qgroup_extent_record *record)
 {
-	if (oper1->bytenr < oper2->bytenr)
-		return -1;
-	if (oper1->bytenr > oper2->bytenr)
-		return 1;
-	if (oper1->ref_root < oper2->ref_root)
-		return -1;
-	if (oper1->ref_root > oper2->ref_root)
-		return 1;
-	if (oper1->seq < oper2->seq)
-		return -1;
-	if (oper1->seq > oper2->seq)
-		return 1;
-	if (oper1->type < oper2->type)
-		return -1;
-	if (oper1->type > oper2->type)
-		return 1;
-	return 0;
-}
+	struct rb_node **p = &delayed_refs->dirty_extent_root.rb_node;
+	struct rb_node *parent_node = NULL;
+	struct btrfs_qgroup_extent_record *entry;
+	u64 bytenr = record->bytenr;
 
-static int insert_qgroup_oper(struct btrfs_fs_info *fs_info,
-			      struct btrfs_qgroup_operation *oper)
-{
-	struct rb_node **p;
-	struct rb_node *parent = NULL;
-	struct btrfs_qgroup_operation *cur;
-	int cmp;
+	assert_spin_locked(&delayed_refs->lock);
 
-	spin_lock(&fs_info->qgroup_op_lock);
-	p = &fs_info->qgroup_op_tree.rb_node;
 	while (*p) {
-		parent = *p;
-		cur = rb_entry(parent, struct btrfs_qgroup_operation, n);
-		cmp = comp_oper(cur, oper);
-		if (cmp < 0) {
-			p = &(*p)->rb_right;
-		} else if (cmp) {
+		parent_node = *p;
+		entry = rb_entry(parent_node, struct btrfs_qgroup_extent_record,
+				 node);
+		if (bytenr < entry->bytenr)
 			p = &(*p)->rb_left;
-		} else {
-			spin_unlock(&fs_info->qgroup_op_lock);
-			return -EEXIST;
-		}
-	}
-	rb_link_node(&oper->n, parent, p);
-	rb_insert_color(&oper->n, &fs_info->qgroup_op_tree);
-	spin_unlock(&fs_info->qgroup_op_lock);
-	return 0;
-}
-
-/*
- * Record a quota operation for processing later on.
- * @trans: the transaction we are adding the delayed op to.
- * @fs_info: the fs_info for this fs.
- * @ref_root: the root of the reference we are acting on,
- * @bytenr: the bytenr we are acting on.
- * @num_bytes: the number of bytes in the reference.
- * @type: the type of operation this is.
- * @mod_seq: do we need to get a sequence number for looking up roots.
- *
- * We just add it to our trans qgroup_ref_list and carry on and process these
- * operations in order at some later point.  If the reference root isn't a fs
- * root then we don't bother with doing anything.
- *
- * MUST BE HOLDING THE REF LOCK.
- */
-int btrfs_qgroup_record_ref(struct btrfs_trans_handle *trans,
-			    struct btrfs_fs_info *fs_info, u64 ref_root,
-			    u64 bytenr, u64 num_bytes,
-			    enum btrfs_qgroup_operation_type type, int mod_seq)
-{
-	struct btrfs_qgroup_operation *oper;
-	int ret;
-
-	if (!is_fstree(ref_root) || !fs_info->quota_enabled)
-		return 0;
-
-	oper = kmalloc(sizeof(*oper), GFP_NOFS);
-	if (!oper)
-		return -ENOMEM;
-
-	oper->ref_root = ref_root;
-	oper->bytenr = bytenr;
-	oper->num_bytes = num_bytes;
-	oper->type = type;
-	oper->seq = atomic_inc_return(&fs_info->qgroup_op_seq);
-	INIT_LIST_HEAD(&oper->elem.list);
-	oper->elem.seq = 0;
-
-	trace_btrfs_qgroup_record_ref(oper);
-
-	if (type == BTRFS_QGROUP_OPER_SUB_SUBTREE) {
-		/*
-		 * If any operation for this bytenr/ref_root combo
-		 * exists, then we know it's not exclusively owned and
-		 * shouldn't be queued up.
-		 *
-		 * This also catches the case where we have a cloned
-		 * extent that gets queued up multiple times during
-		 * drop snapshot.
-		 */
-		if (qgroup_oper_exists(fs_info, oper)) {
-			kfree(oper);
-			return 0;
-		}
-	}
-
-	ret = insert_qgroup_oper(fs_info, oper);
-	if (ret) {
-		/* Shouldn't happen so have an assert for developers */
-		ASSERT(0);
-		kfree(oper);
-		return ret;
+		else if (bytenr > entry->bytenr)
+			p = &(*p)->rb_right;
+		else
+			return entry;
 	}
-	list_add_tail(&oper->list, &trans->qgroup_ref_list);
-
-	if (mod_seq)
-		btrfs_get_tree_mod_seq(fs_info, &oper->elem);
 
-	return 0;
-}
-
-static int qgroup_excl_accounting(struct btrfs_fs_info *fs_info,
-				  struct btrfs_qgroup_operation *oper)
-{
-	struct ulist *tmp;
-	int sign = 0;
-	int ret = 0;
-
-	tmp = ulist_alloc(GFP_NOFS);
-	if (!tmp)
-		return -ENOMEM;
-
-	spin_lock(&fs_info->qgroup_lock);
-	if (!fs_info->quota_root)
-		goto out;
-
-	switch (oper->type) {
-	case BTRFS_QGROUP_OPER_ADD_EXCL:
-		sign = 1;
-		break;
-	case BTRFS_QGROUP_OPER_SUB_EXCL:
-		sign = -1;
-		break;
-	default:
-		ASSERT(0);
-	}
-	ret = __qgroup_excl_accounting(fs_info, tmp, oper->ref_root,
-				       oper->num_bytes, sign);
-out:
-	spin_unlock(&fs_info->qgroup_lock);
-	ulist_free(tmp);
-	return ret;
+	rb_link_node(&record->node, parent_node, p);
+	rb_insert_color(&record->node, &delayed_refs->dirty_extent_root);
+	return NULL;
 }
 
+#define UPDATE_NEW	0
+#define UPDATE_OLD	1
 /*
- * Walk all of the roots that pointed to our bytenr and adjust their refcnts as
- * properly.
+ * Walk all of the roots that points to the bytenr and adjust their refcnts.
  */
-static int qgroup_calc_old_refcnt(struct btrfs_fs_info *fs_info,
-				  u64 root_to_skip, struct ulist *tmp,
-				  struct ulist *roots, struct ulist *qgroups,
-				  u64 seq, int *old_roots, int rescan)
+static int qgroup_update_refcnt(struct btrfs_fs_info *fs_info,
+				struct ulist *roots, struct ulist *tmp,
+				struct ulist *qgroups, u64 seq, int update_old)
 {
 	struct ulist_node *unode;
 	struct ulist_iterator uiter;
 	struct ulist_node *tmp_unode;
 	struct ulist_iterator tmp_uiter;
 	struct btrfs_qgroup *qg;
-	int ret;
+	int ret = 0;
 
+	if (!roots)
+		return 0;
 	ULIST_ITER_INIT(&uiter);
 	while ((unode = ulist_next(roots, &uiter))) {
-		/* We don't count our current root here */
-		if (unode->val == root_to_skip)
-			continue;
 		qg = find_qgroup_rb(fs_info, unode->val);
 		if (!qg)
 			continue;
-		/*
-		 * We could have a pending removal of this same ref so we may
-		 * not have actually found our ref root when doing
-		 * btrfs_find_all_roots, so we need to keep track of how many
-		 * old roots we find in case we removed ours and added a
-		 * different one at the same time.  I don't think this could
-		 * happen in practice but that sort of thinking leads to pain
-		 * and suffering and to the dark side.
-		 */
-		(*old_roots)++;
 
 		ulist_reinit(tmp);
 		ret = ulist_add(qgroups, qg->qgroupid, ptr_to_u64(qg),
@@ -1603,29 +1518,10 @@ static int qgroup_calc_old_refcnt(struct btrfs_fs_info *fs_info,
 			struct btrfs_qgroup_list *glist;
 
 			qg = u64_to_ptr(tmp_unode->aux);
-			/*
-			 * We use this sequence number to keep from having to
-			 * run the whole list and 0 out the refcnt every time.
-			 * We basically use sequnce as the known 0 count and
-			 * then add 1 everytime we see a qgroup.  This is how we
-			 * get how many of the roots actually point up to the
-			 * upper level qgroups in order to determine exclusive
-			 * counts.
-			 *
-			 * For rescan we want to set old_refcnt to seq so our
-			 * exclusive calculations end up correct.
-			 */
-			if (rescan)
-				qg->old_refcnt = seq;
-			else if (qg->old_refcnt < seq)
-				qg->old_refcnt = seq + 1;
+			if (update_old)
+				btrfs_qgroup_update_old_refcnt(qg, seq, 1);
 			else
-				qg->old_refcnt++;
-
-			if (qg->new_refcnt < seq)
-				qg->new_refcnt = seq + 1;
-			else
-				qg->new_refcnt++;
+				btrfs_qgroup_update_new_refcnt(qg, seq, 1);
 			list_for_each_entry(glist, &qg->groups, next_group) {
 				ret = ulist_add(qgroups, glist->group->qgroupid,
 						ptr_to_u64(glist->group),
@@ -1644,161 +1540,46 @@ static int qgroup_calc_old_refcnt(struct btrfs_fs_info *fs_info,
 }
 
 /*
- * We need to walk forward in our operation tree and account for any roots that
- * were deleted after we made this operation.
- */
-static int qgroup_account_deleted_refs(struct btrfs_fs_info *fs_info,
-				       struct btrfs_qgroup_operation *oper,
-				       struct ulist *tmp,
-				       struct ulist *qgroups, u64 seq,
-				       int *old_roots)
-{
-	struct ulist_node *unode;
-	struct ulist_iterator uiter;
-	struct btrfs_qgroup *qg;
-	struct btrfs_qgroup_operation *tmp_oper;
-	struct rb_node *n;
-	int ret;
-
-	ulist_reinit(tmp);
-
-	/*
-	 * We only walk forward in the tree since we're only interested in
-	 * removals that happened _after_  our operation.
-	 */
-	spin_lock(&fs_info->qgroup_op_lock);
-	n = rb_next(&oper->n);
-	spin_unlock(&fs_info->qgroup_op_lock);
-	if (!n)
-		return 0;
-	tmp_oper = rb_entry(n, struct btrfs_qgroup_operation, n);
-	while (tmp_oper->bytenr == oper->bytenr) {
-		/*
-		 * If it's not a removal we don't care, additions work out
-		 * properly with our refcnt tracking.
-		 */
-		if (tmp_oper->type != BTRFS_QGROUP_OPER_SUB_SHARED &&
-		    tmp_oper->type != BTRFS_QGROUP_OPER_SUB_EXCL)
-			goto next;
-		qg = find_qgroup_rb(fs_info, tmp_oper->ref_root);
-		if (!qg)
-			goto next;
-		ret = ulist_add(qgroups, qg->qgroupid, ptr_to_u64(qg),
-				GFP_ATOMIC);
-		if (ret) {
-			if (ret < 0)
-				return ret;
-			/*
-			 * We only want to increase old_roots if this qgroup is
-			 * not already in the list of qgroups.  If it is already
-			 * there then that means it must have been re-added or
-			 * the delete will be discarded because we had an
-			 * existing ref that we haven't looked up yet.  In this
-			 * case we don't want to increase old_roots.  So if ret
-			 * == 1 then we know that this is the first time we've
-			 * seen this qgroup and we can bump the old_roots.
-			 */
-			(*old_roots)++;
-			ret = ulist_add(tmp, qg->qgroupid, ptr_to_u64(qg),
-					GFP_ATOMIC);
-			if (ret < 0)
-				return ret;
-		}
-next:
-		spin_lock(&fs_info->qgroup_op_lock);
-		n = rb_next(&tmp_oper->n);
-		spin_unlock(&fs_info->qgroup_op_lock);
-		if (!n)
-			break;
-		tmp_oper = rb_entry(n, struct btrfs_qgroup_operation, n);
-	}
-
-	/* Ok now process the qgroups we found */
-	ULIST_ITER_INIT(&uiter);
-	while ((unode = ulist_next(tmp, &uiter))) {
-		struct btrfs_qgroup_list *glist;
-
-		qg = u64_to_ptr(unode->aux);
-		if (qg->old_refcnt < seq)
-			qg->old_refcnt = seq + 1;
-		else
-			qg->old_refcnt++;
-		if (qg->new_refcnt < seq)
-			qg->new_refcnt = seq + 1;
-		else
-			qg->new_refcnt++;
-		list_for_each_entry(glist, &qg->groups, next_group) {
-			ret = ulist_add(qgroups, glist->group->qgroupid,
-					ptr_to_u64(glist->group), GFP_ATOMIC);
-			if (ret < 0)
-				return ret;
-			ret = ulist_add(tmp, glist->group->qgroupid,
-					ptr_to_u64(glist->group), GFP_ATOMIC);
-			if (ret < 0)
-				return ret;
-		}
-	}
-	return 0;
-}
-
-/* Add refcnt for the newly added reference. */
-static int qgroup_calc_new_refcnt(struct btrfs_fs_info *fs_info,
-				  struct btrfs_qgroup_operation *oper,
-				  struct btrfs_qgroup *qgroup,
-				  struct ulist *tmp, struct ulist *qgroups,
-				  u64 seq)
-{
-	struct ulist_node *unode;
-	struct ulist_iterator uiter;
-	struct btrfs_qgroup *qg;
-	int ret;
-
-	ulist_reinit(tmp);
-	ret = ulist_add(qgroups, qgroup->qgroupid, ptr_to_u64(qgroup),
-			GFP_ATOMIC);
-	if (ret < 0)
-		return ret;
-	ret = ulist_add(tmp, qgroup->qgroupid, ptr_to_u64(qgroup),
-			GFP_ATOMIC);
-	if (ret < 0)
-		return ret;
-	ULIST_ITER_INIT(&uiter);
-	while ((unode = ulist_next(tmp, &uiter))) {
-		struct btrfs_qgroup_list *glist;
-
-		qg = u64_to_ptr(unode->aux);
-		if (oper->type == BTRFS_QGROUP_OPER_ADD_SHARED) {
-			if (qg->new_refcnt < seq)
-				qg->new_refcnt = seq + 1;
-			else
-				qg->new_refcnt++;
-		} else {
-			if (qg->old_refcnt < seq)
-				qg->old_refcnt = seq + 1;
-			else
-				qg->old_refcnt++;
-		}
-		list_for_each_entry(glist, &qg->groups, next_group) {
-			ret = ulist_add(tmp, glist->group->qgroupid,
-					ptr_to_u64(glist->group), GFP_ATOMIC);
-			if (ret < 0)
-				return ret;
-			ret = ulist_add(qgroups, glist->group->qgroupid,
-					ptr_to_u64(glist->group), GFP_ATOMIC);
-			if (ret < 0)
-				return ret;
-		}
-	}
-	return 0;
-}
-
-/*
- * This adjusts the counters for all referenced qgroups if need be.
+ * Update qgroup rfer/excl counters.
+ * Rfer update is easy, codes can explain themselves.
+ *
+ * Excl update is tricky, the update is split into 2 part.
+ * Part 1: Possible exclusive <-> sharing detect:
+ *	|	A	|	!A	|
+ *  -------------------------------------
+ *  B	|	*	|	-	|
+ *  -------------------------------------
+ *  !B	|	+	|	**	|
+ *  -------------------------------------
+ *
+ * Conditions:
+ * A:	cur_old_roots < nr_old_roots	(not exclusive before)
+ * !A:	cur_old_roots == nr_old_roots	(possible exclusive before)
+ * B:	cur_new_roots < nr_new_roots	(not exclusive now)
+ * !B:	cur_new_roots == nr_new_roots	(possible exclsuive now)
+ *
+ * Results:
+ * +: Possible sharing -> exclusive	-: Possible exclusive -> sharing
+ * *: Definitely not changed.		**: Possible unchanged.
+ *
+ * For !A and !B condition, the exception is cur_old/new_roots == 0 case.
+ *
+ * To make the logic clear, we first use condition A and B to split
+ * combination into 4 results.
+ *
+ * Then, for result "+" and "-", check old/new_roots == 0 case, as in them
+ * only on variant maybe 0.
+ *
+ * Lastly, check result **, since there are 2 variants maybe 0, split them
+ * again(2x2).
+ * But this time we don't need to consider other things, the codes and logic
+ * is easy to understand now.
  */
-static int qgroup_adjust_counters(struct btrfs_fs_info *fs_info,
-				  u64 root_to_skip, u64 num_bytes,
-				  struct ulist *qgroups, u64 seq,
-				  int old_roots, int new_roots, int rescan)
+static int qgroup_update_counters(struct btrfs_fs_info *fs_info,
+				  struct ulist *qgroups,
+				  u64 nr_old_roots,
+				  u64 nr_new_roots,
+				  u64 num_bytes, u64 seq)
 {
 	struct ulist_node *unode;
 	struct ulist_iterator uiter;
@@ -1810,57 +1591,68 @@ static int qgroup_adjust_counters(struct btrfs_fs_info *fs_info,
 		bool dirty = false;
 
 		qg = u64_to_ptr(unode->aux);
-		/*
-		 * Wasn't referenced before but is now, add to the reference
-		 * counters.
-		 */
-		if (qg->old_refcnt <= seq && qg->new_refcnt > seq) {
+		cur_old_count = btrfs_qgroup_get_old_refcnt(qg, seq);
+		cur_new_count = btrfs_qgroup_get_new_refcnt(qg, seq);
+
+		/* Rfer update part */
+		if (cur_old_count == 0 && cur_new_count > 0) {
 			qg->rfer += num_bytes;
 			qg->rfer_cmpr += num_bytes;
 			dirty = true;
 		}
-
-		/*
-		 * Was referenced before but isn't now, subtract from the
-		 * reference counters.
-		 */
-		if (qg->old_refcnt > seq && qg->new_refcnt <= seq) {
+		if (cur_old_count > 0 && cur_new_count == 0) {
 			qg->rfer -= num_bytes;
 			qg->rfer_cmpr -= num_bytes;
 			dirty = true;
 		}
 
-		if (qg->old_refcnt < seq)
-			cur_old_count = 0;
-		else
-			cur_old_count = qg->old_refcnt - seq;
-		if (qg->new_refcnt < seq)
-			cur_new_count = 0;
-		else
-			cur_new_count = qg->new_refcnt - seq;
+		/* Excl update part */
+		/* Exclusive/none -> shared case */
+		if (cur_old_count == nr_old_roots &&
+		    cur_new_count < nr_new_roots) {
+			/* Exclusive -> shared */
+			if (cur_old_count != 0) {
+				qg->excl -= num_bytes;
+				qg->excl_cmpr -= num_bytes;
+				dirty = true;
+			}
+		}
 
-		/*
-		 * If our refcount was the same as the roots previously but our
-		 * new count isn't the same as the number of roots now then we
-		 * went from having a exclusive reference on this range to not.
-		 */
-		if (old_roots && cur_old_count == old_roots &&
-		    (cur_new_count != new_roots || new_roots == 0)) {
-			WARN_ON(cur_new_count != new_roots && new_roots == 0);
-			qg->excl -= num_bytes;
-			qg->excl_cmpr -= num_bytes;
-			dirty = true;
+		/* Shared -> exclusive/none case */
+		if (cur_old_count < nr_old_roots &&
+		    cur_new_count == nr_new_roots) {
+			/* Shared->exclusive */
+			if (cur_new_count != 0) {
+				qg->excl += num_bytes;
+				qg->excl_cmpr += num_bytes;
+				dirty = true;
+			}
 		}
 
-		/*
-		 * If we didn't reference all the roots before but now we do we
-		 * have an exclusive reference to this range.
-		 */
-		if ((!old_roots || (old_roots && cur_old_count != old_roots))
-		    && cur_new_count == new_roots) {
-			qg->excl += num_bytes;
-			qg->excl_cmpr += num_bytes;
-			dirty = true;
+		/* Exclusive/none -> exclusive/none case */
+		if (cur_old_count == nr_old_roots &&
+		    cur_new_count == nr_new_roots) {
+			if (cur_old_count == 0) {
+				/* None -> exclusive/none */
+
+				if (cur_new_count != 0) {
+					/* None -> exclusive */
+					qg->excl += num_bytes;
+					qg->excl_cmpr += num_bytes;
+					dirty = true;
+				}
+				/* None -> none, nothing changed */
+			} else {
+				/* Exclusive -> exclusive/none */
+
+				if (cur_new_count == 0) {
+					/* Exclusive -> none */
+					qg->excl -= num_bytes;
+					qg->excl_cmpr -= num_bytes;
+					dirty = true;
+				}
+				/* Exclusive -> exclusive, nothing changed */
+			}
 		}
 
 		if (dirty)
@@ -1869,364 +1661,122 @@ static int qgroup_adjust_counters(struct btrfs_fs_info *fs_info,
 	return 0;
 }
 
-/*
- * If we removed a data extent and there were other references for that bytenr
- * then we need to lookup all referenced roots to make sure we still don't
- * reference this bytenr.  If we do then we can just discard this operation.
- */
-static int check_existing_refs(struct btrfs_trans_handle *trans,
-			       struct btrfs_fs_info *fs_info,
-			       struct btrfs_qgroup_operation *oper)
-{
-	struct ulist *roots = NULL;
-	struct ulist_node *unode;
-	struct ulist_iterator uiter;
-	int ret = 0;
-
-	ret = btrfs_find_all_roots(trans, fs_info, oper->bytenr,
-				   oper->elem.seq, &roots);
-	if (ret < 0)
-		return ret;
-	ret = 0;
-
-	ULIST_ITER_INIT(&uiter);
-	while ((unode = ulist_next(roots, &uiter))) {
-		if (unode->val == oper->ref_root) {
-			ret = 1;
-			break;
-		}
-	}
-	ulist_free(roots);
-	btrfs_put_tree_mod_seq(fs_info, &oper->elem);
-
-	return ret;
-}
-
-/*
- * If we share a reference across multiple roots then we may need to adjust
- * various qgroups referenced and exclusive counters.  The basic premise is this
- *
- * 1) We have seq to represent a 0 count.  Instead of looping through all of the
- * qgroups and resetting their refcount to 0 we just constantly bump this
- * sequence number to act as the base reference count.  This means that if
- * anybody is equal to or below this sequence they were never referenced.  We
- * jack this sequence up by the number of roots we found each time in order to
- * make sure we don't have any overlap.
- *
- * 2) We first search all the roots that reference the area _except_ the root
- * we're acting on currently.  This makes up the old_refcnt of all the qgroups
- * before.
- *
- * 3) We walk all of the qgroups referenced by the root we are currently acting
- * on, and will either adjust old_refcnt in the case of a removal or the
- * new_refcnt in the case of an addition.
- *
- * 4) Finally we walk all the qgroups that are referenced by this range
- * including the root we are acting on currently.  We will adjust the counters
- * based on the number of roots we had and will have after this operation.
- *
- * Take this example as an illustration
- *
- *			[qgroup 1/0]
- *		     /         |          \
- *		[qg 0/0]   [qg 0/1]	[qg 0/2]
- *		   \          |            /
- *		  [	   extent	    ]
- *
- * Say we are adding a reference that is covered by qg 0/0.  The first step
- * would give a refcnt of 1 to qg 0/1 and 0/2 and a refcnt of 2 to qg 1/0 with
- * old_roots being 2.  Because it is adding new_roots will be 1.  We then go
- * through qg 0/0 which will get the new_refcnt set to 1 and add 1 to qg 1/0's
- * new_refcnt, bringing it to 3.  We then walk through all of the qgroups, we
- * notice that the old refcnt for qg 0/0 < the new refcnt, so we added a
- * reference and thus must add the size to the referenced bytes.  Everything
- * else is the same so nothing else changes.
- */
-static int qgroup_shared_accounting(struct btrfs_trans_handle *trans,
-				    struct btrfs_fs_info *fs_info,
-				    struct btrfs_qgroup_operation *oper)
+int
+btrfs_qgroup_account_extent(struct btrfs_trans_handle *trans,
+			    struct btrfs_fs_info *fs_info,
+			    u64 bytenr, u64 num_bytes,
+			    struct ulist *old_roots, struct ulist *new_roots)
 {
-	struct ulist *roots = NULL;
-	struct ulist *qgroups, *tmp;
-	struct btrfs_qgroup *qgroup;
-	struct seq_list elem = SEQ_LIST_INIT(elem);
+	struct ulist *qgroups = NULL;
+	struct ulist *tmp = NULL;
 	u64 seq;
-	int old_roots = 0;
-	int new_roots = 0;
+	u64 nr_new_roots = 0;
+	u64 nr_old_roots = 0;
 	int ret = 0;
 
-	if (oper->elem.seq) {
-		ret = check_existing_refs(trans, fs_info, oper);
-		if (ret < 0)
-			return ret;
-		if (ret)
-			return 0;
-	}
+	if (new_roots)
+		nr_new_roots = new_roots->nnodes;
+	if (old_roots)
+		nr_old_roots = old_roots->nnodes;
 
-	qgroups = ulist_alloc(GFP_NOFS);
-	if (!qgroups)
-		return -ENOMEM;
+	if (!fs_info->quota_enabled)
+		goto out_free;
+	BUG_ON(!fs_info->quota_root);
 
+	qgroups = ulist_alloc(GFP_NOFS);
+	if (!qgroups) {
+		ret = -ENOMEM;
+		goto out_free;
+	}
 	tmp = ulist_alloc(GFP_NOFS);
 	if (!tmp) {
-		ulist_free(qgroups);
-		return -ENOMEM;
+		ret = -ENOMEM;
+		goto out_free;
 	}
 
-	btrfs_get_tree_mod_seq(fs_info, &elem);
-	ret = btrfs_find_all_roots(trans, fs_info, oper->bytenr, elem.seq,
-				   &roots);
-	btrfs_put_tree_mod_seq(fs_info, &elem);
-	if (ret < 0) {
-		ulist_free(qgroups);
-		ulist_free(tmp);
-		return ret;
+	mutex_lock(&fs_info->qgroup_rescan_lock);
+	if (fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_RESCAN) {
+		if (fs_info->qgroup_rescan_progress.objectid <= bytenr) {
+			mutex_unlock(&fs_info->qgroup_rescan_lock);
+			ret = 0;
+			goto out_free;
+		}
 	}
+	mutex_unlock(&fs_info->qgroup_rescan_lock);
+
 	spin_lock(&fs_info->qgroup_lock);
-	qgroup = find_qgroup_rb(fs_info, oper->ref_root);
-	if (!qgroup)
-		goto out;
 	seq = fs_info->qgroup_seq;
 
-	/*
-	 * So roots is the list of all the roots currently pointing at the
-	 * bytenr, including the ref we are adding if we are adding, or not if
-	 * we are removing a ref.  So we pass in the ref_root to skip that root
-	 * in our calculations.  We set old_refnct and new_refcnt cause who the
-	 * hell knows what everything looked like before, and it doesn't matter
-	 * except...
-	 */
-	ret = qgroup_calc_old_refcnt(fs_info, oper->ref_root, tmp, roots, qgroups,
-				     seq, &old_roots, 0);
-	if (ret < 0)
-		goto out;
-
-	/*
-	 * Now adjust the refcounts of the qgroups that care about this
-	 * reference, either the old_count in the case of removal or new_count
-	 * in the case of an addition.
-	 */
-	ret = qgroup_calc_new_refcnt(fs_info, oper, qgroup, tmp, qgroups,
-				     seq);
+	/* Update old refcnts using old_roots */
+	ret = qgroup_update_refcnt(fs_info, old_roots, tmp, qgroups, seq,
+				   UPDATE_OLD);
 	if (ret < 0)
 		goto out;
 
-	/*
-	 * ...in the case of removals.  If we had a removal before we got around
-	 * to processing this operation then we need to find that guy and count
-	 * his references as if they really existed so we don't end up screwing
-	 * up the exclusive counts.  Then whenever we go to process the delete
-	 * everything will be grand and we can account for whatever exclusive
-	 * changes need to be made there.  We also have to pass in old_roots so
-	 * we have an accurate count of the roots as it pertains to this
-	 * operations view of the world.
-	 */
-	ret = qgroup_account_deleted_refs(fs_info, oper, tmp, qgroups, seq,
-					  &old_roots);
+	/* Update new refcnts using new_roots */
+	ret = qgroup_update_refcnt(fs_info, new_roots, tmp, qgroups, seq,
+				   UPDATE_NEW);
 	if (ret < 0)
 		goto out;
 
-	/*
-	 * We are adding our root, need to adjust up the number of roots,
-	 * otherwise old_roots is the number of roots we want.
-	 */
-	if (oper->type == BTRFS_QGROUP_OPER_ADD_SHARED) {
-		new_roots = old_roots + 1;
-	} else {
-		new_roots = old_roots;
-		old_roots++;
-	}
-	fs_info->qgroup_seq += old_roots + 1;
-
+	qgroup_update_counters(fs_info, qgroups, nr_old_roots, nr_new_roots,
+			       num_bytes, seq);
 
 	/*
-	 * And now the magic happens, bless Arne for having a pretty elegant
-	 * solution for this.
+	 * Bump qgroup_seq to avoid seq overlap
 	 */
-	qgroup_adjust_counters(fs_info, oper->ref_root, oper->num_bytes,
-			       qgroups, seq, old_roots, new_roots, 0);
+	fs_info->qgroup_seq += max(nr_old_roots, nr_new_roots) + 1;
 out:
 	spin_unlock(&fs_info->qgroup_lock);
-	ulist_free(qgroups);
-	ulist_free(roots);
+out_free:
 	ulist_free(tmp);
+	ulist_free(qgroups);
+	ulist_free(old_roots);
+	ulist_free(new_roots);
 	return ret;
 }
 
-/*
- * Process a reference to a shared subtree. This type of operation is
- * queued during snapshot removal when we encounter extents which are
- * shared between more than one root.
- */
-static int qgroup_subtree_accounting(struct btrfs_trans_handle *trans,
-				     struct btrfs_fs_info *fs_info,
-				     struct btrfs_qgroup_operation *oper)
-{
-	struct ulist *roots = NULL;
-	struct ulist_node *unode;
-	struct ulist_iterator uiter;
-	struct btrfs_qgroup_list *glist;
-	struct ulist *parents;
-	int ret = 0;
-	int err;
-	struct btrfs_qgroup *qg;
-	u64 root_obj = 0;
-	struct seq_list elem = SEQ_LIST_INIT(elem);
-
-	parents = ulist_alloc(GFP_NOFS);
-	if (!parents)
-		return -ENOMEM;
-
-	btrfs_get_tree_mod_seq(fs_info, &elem);
-	ret = btrfs_find_all_roots(trans, fs_info, oper->bytenr,
-				   elem.seq, &roots);
-	btrfs_put_tree_mod_seq(fs_info, &elem);
-	if (ret < 0)
-		goto out;
-
-	if (roots->nnodes != 1)
-		goto out;
-
-	ULIST_ITER_INIT(&uiter);
-	unode = ulist_next(roots, &uiter); /* Only want 1 so no need to loop */
-	/*
-	 * If we find our ref root then that means all refs
-	 * this extent has to the root have not yet been
-	 * deleted. In that case, we do nothing and let the
-	 * last ref for this bytenr drive our update.
-	 *
-	 * This can happen for example if an extent is
-	 * referenced multiple times in a snapshot (clone,
-	 * etc). If we are in the middle of snapshot removal,
-	 * queued updates for such an extent will find the
-	 * root if we have not yet finished removing the
-	 * snapshot.
-	 */
-	if (unode->val == oper->ref_root)
-		goto out;
-
-	root_obj = unode->val;
-	BUG_ON(!root_obj);
-
-	spin_lock(&fs_info->qgroup_lock);
-	qg = find_qgroup_rb(fs_info, root_obj);
-	if (!qg)
-		goto out_unlock;
-
-	qg->excl += oper->num_bytes;
-	qg->excl_cmpr += oper->num_bytes;
-	qgroup_dirty(fs_info, qg);
-
-	/*
-	 * Adjust counts for parent groups. First we find all
-	 * parents, then in the 2nd loop we do the adjustment
-	 * while adding parents of the parents to our ulist.
-	 */
-	list_for_each_entry(glist, &qg->groups, next_group) {
-		err = ulist_add(parents, glist->group->qgroupid,
-				ptr_to_u64(glist->group), GFP_ATOMIC);
-		if (err < 0) {
-			ret = err;
-			goto out_unlock;
-		}
-	}
-
-	ULIST_ITER_INIT(&uiter);
-	while ((unode = ulist_next(parents, &uiter))) {
-		qg = u64_to_ptr(unode->aux);
-		qg->excl += oper->num_bytes;
-		qg->excl_cmpr += oper->num_bytes;
-		qgroup_dirty(fs_info, qg);
-
-		/* Add any parents of the parents */
-		list_for_each_entry(glist, &qg->groups, next_group) {
-			err = ulist_add(parents, glist->group->qgroupid,
-					ptr_to_u64(glist->group), GFP_ATOMIC);
-			if (err < 0) {
-				ret = err;
-				goto out_unlock;
-			}
-		}
-	}
-
-out_unlock:
-	spin_unlock(&fs_info->qgroup_lock);
-
-out:
-	ulist_free(roots);
-	ulist_free(parents);
-	return ret;
-}
-
-/*
- * btrfs_qgroup_account_ref is called for every ref that is added to or deleted
- * from the fs. First, all roots referencing the extent are searched, and
- * then the space is accounted accordingly to the different roots. The
- * accounting algorithm works in 3 steps documented inline.
- */
-static int btrfs_qgroup_account(struct btrfs_trans_handle *trans,
-				struct btrfs_fs_info *fs_info,
-				struct btrfs_qgroup_operation *oper)
+int btrfs_qgroup_account_extents(struct btrfs_trans_handle *trans,
+				 struct btrfs_fs_info *fs_info)
 {
+	struct btrfs_qgroup_extent_record *record;
+	struct btrfs_delayed_ref_root *delayed_refs;
+	struct ulist *new_roots = NULL;
+	struct rb_node *node;
+	u64 qgroup_to_skip;
 	int ret = 0;
 
-	if (!fs_info->quota_enabled)
-		return 0;
-
-	BUG_ON(!fs_info->quota_root);
+	delayed_refs = &trans->transaction->delayed_refs;
+	qgroup_to_skip = delayed_refs->qgroup_to_skip;
+	while ((node = rb_first(&delayed_refs->dirty_extent_root))) {
+		record = rb_entry(node, struct btrfs_qgroup_extent_record,
+				  node);
 
-	mutex_lock(&fs_info->qgroup_rescan_lock);
-	if (fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_RESCAN) {
-		if (fs_info->qgroup_rescan_progress.objectid <= oper->bytenr) {
-			mutex_unlock(&fs_info->qgroup_rescan_lock);
-			return 0;
+		if (!ret) {
+			/*
+			 * Use (u64)-1 as time_seq to do special search, which
+			 * doesn't lock tree or delayed_refs and search current
+			 * root. It's safe inside commit_transaction().
+			 */
+			ret = btrfs_find_all_roots(trans, fs_info,
+					record->bytenr, (u64)-1, &new_roots);
+			if (ret < 0)
+				goto cleanup;
+			if (qgroup_to_skip)
+				ulist_del(new_roots, qgroup_to_skip, 0);
+			ret = btrfs_qgroup_account_extent(trans, fs_info,
+					record->bytenr, record->num_bytes,
+					record->old_roots, new_roots);
+			record->old_roots = NULL;
+			new_roots = NULL;
 		}
-	}
-	mutex_unlock(&fs_info->qgroup_rescan_lock);
+cleanup:
+		ulist_free(record->old_roots);
+		ulist_free(new_roots);
+		new_roots = NULL;
+		rb_erase(node, &delayed_refs->dirty_extent_root);
+		kfree(record);
 
-	ASSERT(is_fstree(oper->ref_root));
-
-	trace_btrfs_qgroup_account(oper);
-
-	switch (oper->type) {
-	case BTRFS_QGROUP_OPER_ADD_EXCL:
-	case BTRFS_QGROUP_OPER_SUB_EXCL:
-		ret = qgroup_excl_accounting(fs_info, oper);
-		break;
-	case BTRFS_QGROUP_OPER_ADD_SHARED:
-	case BTRFS_QGROUP_OPER_SUB_SHARED:
-		ret = qgroup_shared_accounting(trans, fs_info, oper);
-		break;
-	case BTRFS_QGROUP_OPER_SUB_SUBTREE:
-		ret = qgroup_subtree_accounting(trans, fs_info, oper);
-		break;
-	default:
-		ASSERT(0);
-	}
-	return ret;
-}
-
-/*
- * Needs to be called everytime we run delayed refs, even if there is an error
- * in order to cleanup outstanding operations.
- */
-int btrfs_delayed_qgroup_accounting(struct btrfs_trans_handle *trans,
-				    struct btrfs_fs_info *fs_info)
-{
-	struct btrfs_qgroup_operation *oper;
-	int ret = 0;
-
-	while (!list_empty(&trans->qgroup_ref_list)) {
-		oper = list_first_entry(&trans->qgroup_ref_list,
-					struct btrfs_qgroup_operation, list);
-		list_del_init(&oper->list);
-		if (!ret || !trans->aborted)
-			ret = btrfs_qgroup_account(trans, fs_info, oper);
-		spin_lock(&fs_info->qgroup_op_lock);
-		rb_erase(&oper->n, &fs_info->qgroup_op_tree);
-		spin_unlock(&fs_info->qgroup_op_lock);
-		btrfs_put_tree_mod_seq(fs_info, &oper->elem);
-		kfree(oper);
 	}
 	return ret;
 }
@@ -2484,7 +2034,7 @@ out:
 	return ret;
 }
 
-int btrfs_qgroup_reserve(struct btrfs_root *root, u64 num_bytes)
+static int qgroup_reserve(struct btrfs_root *root, u64 num_bytes)
 {
 	struct btrfs_root *quota_root;
 	struct btrfs_qgroup *qgroup;
@@ -2565,14 +2115,13 @@ out:
 	return ret;
 }
 
-void btrfs_qgroup_free(struct btrfs_root *root, u64 num_bytes)
+void btrfs_qgroup_free_refroot(struct btrfs_fs_info *fs_info,
+			       u64 ref_root, u64 num_bytes)
 {
 	struct btrfs_root *quota_root;
 	struct btrfs_qgroup *qgroup;
-	struct btrfs_fs_info *fs_info = root->fs_info;
 	struct ulist_node *unode;
 	struct ulist_iterator uiter;
-	u64 ref_root = root->root_key.objectid;
 	int ret = 0;
 
 	if (!is_fstree(ref_root))
@@ -2618,6 +2167,11 @@ out:
 	spin_unlock(&fs_info->qgroup_lock);
 }
 
+static inline void qgroup_free(struct btrfs_root *root, u64 num_bytes)
+{
+	return btrfs_qgroup_free_refroot(root->fs_info, root->objectid,
+					 num_bytes);
+}
 void assert_qgroups_uptodate(struct btrfs_trans_handle *trans)
 {
 	if (list_empty(&trans->qgroup_ref_list) && !trans->delayed_ref_elem.seq)
@@ -2637,19 +2191,16 @@ void assert_qgroups_uptodate(struct btrfs_trans_handle *trans)
  */
 static int
 qgroup_rescan_leaf(struct btrfs_fs_info *fs_info, struct btrfs_path *path,
-		   struct btrfs_trans_handle *trans, struct ulist *qgroups,
-		   struct ulist *tmp, struct extent_buffer *scratch_leaf)
+		   struct btrfs_trans_handle *trans)
 {
 	struct btrfs_key found;
+	struct extent_buffer *scratch_leaf = NULL;
 	struct ulist *roots = NULL;
 	struct seq_list tree_mod_seq_elem = SEQ_LIST_INIT(tree_mod_seq_elem);
 	u64 num_bytes;
-	u64 seq;
-	int new_roots;
 	int slot;
 	int ret;
 
-	path->leave_spinning = 1;
 	mutex_lock(&fs_info->qgroup_rescan_lock);
 	ret = btrfs_search_slot_for_read(fs_info->extent_root,
 					 &fs_info->qgroup_rescan_progress,
@@ -2680,7 +2231,15 @@ qgroup_rescan_leaf(struct btrfs_fs_info *fs_info, struct btrfs_path *path,
 	fs_info->qgroup_rescan_progress.objectid = found.objectid + 1;
 
 	btrfs_get_tree_mod_seq(fs_info, &tree_mod_seq_elem);
-	memcpy(scratch_leaf, path->nodes[0], sizeof(*scratch_leaf));
+	scratch_leaf = btrfs_clone_extent_buffer(path->nodes[0]);
+	if (!scratch_leaf) {
+		ret = -ENOMEM;
+		mutex_unlock(&fs_info->qgroup_rescan_lock);
+		goto out;
+	}
+	extent_buffer_get(scratch_leaf);
+	btrfs_tree_read_lock(scratch_leaf);
+	btrfs_set_lock_blocking_rw(scratch_leaf, BTRFS_READ_LOCK);
 	slot = path->slots[0];
 	btrfs_release_path(path);
 	mutex_unlock(&fs_info->qgroup_rescan_lock);
@@ -2695,35 +2254,21 @@ qgroup_rescan_leaf(struct btrfs_fs_info *fs_info, struct btrfs_path *path,
 		else
 			num_bytes = found.offset;
 
-		ulist_reinit(qgroups);
 		ret = btrfs_find_all_roots(NULL, fs_info, found.objectid, 0,
 					   &roots);
 		if (ret < 0)
 			goto out;
-		spin_lock(&fs_info->qgroup_lock);
-		seq = fs_info->qgroup_seq;
-		fs_info->qgroup_seq += roots->nnodes + 1; /* max refcnt */
-
-		new_roots = 0;
-		ret = qgroup_calc_old_refcnt(fs_info, 0, tmp, roots, qgroups,
-					     seq, &new_roots, 1);
-		if (ret < 0) {
-			spin_unlock(&fs_info->qgroup_lock);
-			ulist_free(roots);
-			goto out;
-		}
-
-		ret = qgroup_adjust_counters(fs_info, 0, num_bytes, qgroups,
-					     seq, 0, new_roots, 1);
-		if (ret < 0) {
-			spin_unlock(&fs_info->qgroup_lock);
-			ulist_free(roots);
+		/* For rescan, just pass old_roots as NULL */
+		ret = btrfs_qgroup_account_extent(trans, fs_info,
+				found.objectid, num_bytes, NULL, roots);
+		if (ret < 0)
 			goto out;
-		}
-		spin_unlock(&fs_info->qgroup_lock);
-		ulist_free(roots);
 	}
 out:
+	if (scratch_leaf) {
+		btrfs_tree_read_unlock_blocking(scratch_leaf);
+		free_extent_buffer(scratch_leaf);
+	}
 	btrfs_put_tree_mod_seq(fs_info, &tree_mod_seq_elem);
 
 	return ret;
@@ -2735,26 +2280,15 @@ static void btrfs_qgroup_rescan_worker(struct btrfs_work *work)
 						     qgroup_rescan_work);
 	struct btrfs_path *path;
 	struct btrfs_trans_handle *trans = NULL;
-	struct ulist *tmp = NULL, *qgroups = NULL;
-	struct extent_buffer *scratch_leaf = NULL;
 	int err = -ENOMEM;
 	int ret = 0;
 
 	path = btrfs_alloc_path();
 	if (!path)
 		goto out;
-	qgroups = ulist_alloc(GFP_NOFS);
-	if (!qgroups)
-		goto out;
-	tmp = ulist_alloc(GFP_NOFS);
-	if (!tmp)
-		goto out;
-	scratch_leaf = kmalloc(sizeof(*scratch_leaf), GFP_NOFS);
-	if (!scratch_leaf)
-		goto out;
 
 	err = 0;
-	while (!err) {
+	while (!err && !btrfs_fs_closing(fs_info)) {
 		trans = btrfs_start_transaction(fs_info->fs_root, 0);
 		if (IS_ERR(trans)) {
 			err = PTR_ERR(trans);
@@ -2763,8 +2297,7 @@ static void btrfs_qgroup_rescan_worker(struct btrfs_work *work)
 		if (!fs_info->quota_enabled) {
 			err = -EINTR;
 		} else {
-			err = qgroup_rescan_leaf(fs_info, path, trans,
-						 qgroups, tmp, scratch_leaf);
+			err = qgroup_rescan_leaf(fs_info, path, trans);
 		}
 		if (err > 0)
 			btrfs_commit_transaction(trans, fs_info->fs_root);
@@ -2773,13 +2306,11 @@ static void btrfs_qgroup_rescan_worker(struct btrfs_work *work)
 	}
 
 out:
-	kfree(scratch_leaf);
-	ulist_free(qgroups);
-	ulist_free(tmp);
 	btrfs_free_path(path);
 
 	mutex_lock(&fs_info->qgroup_rescan_lock);
-	fs_info->qgroup_flags &= ~BTRFS_QGROUP_STATUS_FLAG_RESCAN;
+	if (!btrfs_fs_closing(fs_info))
+		fs_info->qgroup_flags &= ~BTRFS_QGROUP_STATUS_FLAG_RESCAN;
 
 	if (err > 0 &&
 	    fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_INCONSISTENT) {
@@ -2808,7 +2339,9 @@ out:
 	}
 	btrfs_end_transaction(trans, fs_info->quota_root);
 
-	if (err >= 0) {
+	if (btrfs_fs_closing(fs_info)) {
+		btrfs_info(fs_info, "qgroup scan paused");
+	} else if (err >= 0) {
 		btrfs_info(fs_info, "qgroup scan completed%s",
 			err > 0 ? " (inconsistency flag cleared)" : "");
 	} else {
@@ -2856,12 +2389,11 @@ qgroup_rescan_init(struct btrfs_fs_info *fs_info, u64 progress_objectid,
 	memset(&fs_info->qgroup_rescan_progress, 0,
 		sizeof(fs_info->qgroup_rescan_progress));
 	fs_info->qgroup_rescan_progress.objectid = progress_objectid;
+	init_completion(&fs_info->qgroup_rescan_completion);
 
 	spin_unlock(&fs_info->qgroup_lock);
 	mutex_unlock(&fs_info->qgroup_rescan_lock);
 
-	init_completion(&fs_info->qgroup_rescan_completion);
-
 	memset(&fs_info->qgroup_rescan_work, 0,
 	       sizeof(fs_info->qgroup_rescan_work));
 	btrfs_init_work(&fs_info->qgroup_rescan_work,
@@ -2964,3 +2496,190 @@ btrfs_qgroup_rescan_resume(struct btrfs_fs_info *fs_info)
 		btrfs_queue_work(fs_info->qgroup_rescan_workers,
 				 &fs_info->qgroup_rescan_work);
 }
+
+/*
+ * Reserve qgroup space for range [start, start + len).
+ *
+ * This function will either reserve space from related qgroups or doing
+ * nothing if the range is already reserved.
+ *
+ * Return 0 for successful reserve
+ * Return <0 for error (including -EQUOT)
+ *
+ * NOTE: this function may sleep for memory allocation.
+ */
+int btrfs_qgroup_reserve_data(struct inode *inode, u64 start, u64 len)
+{
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct extent_changeset changeset;
+	struct ulist_node *unode;
+	struct ulist_iterator uiter;
+	int ret;
+
+	if (!root->fs_info->quota_enabled || !is_fstree(root->objectid) ||
+	    len == 0)
+		return 0;
+
+	changeset.bytes_changed = 0;
+	changeset.range_changed = ulist_alloc(GFP_NOFS);
+	ret = set_record_extent_bits(&BTRFS_I(inode)->io_tree, start,
+			start + len -1, EXTENT_QGROUP_RESERVED, GFP_NOFS,
+			&changeset);
+	trace_btrfs_qgroup_reserve_data(inode, start, len,
+					changeset.bytes_changed,
+					QGROUP_RESERVE);
+	if (ret < 0)
+		goto cleanup;
+	ret = qgroup_reserve(root, changeset.bytes_changed);
+	if (ret < 0)
+		goto cleanup;
+
+	ulist_free(changeset.range_changed);
+	return ret;
+
+cleanup:
+	/* cleanup already reserved ranges */
+	ULIST_ITER_INIT(&uiter);
+	while ((unode = ulist_next(changeset.range_changed, &uiter)))
+		clear_extent_bit(&BTRFS_I(inode)->io_tree, unode->val,
+				 unode->aux, EXTENT_QGROUP_RESERVED, 0, 0, NULL,
+				 GFP_NOFS);
+	ulist_free(changeset.range_changed);
+	return ret;
+}
+
+static int __btrfs_qgroup_release_data(struct inode *inode, u64 start, u64 len,
+				       int free)
+{
+	struct extent_changeset changeset;
+	int trace_op = QGROUP_RELEASE;
+	int ret;
+
+	changeset.bytes_changed = 0;
+	changeset.range_changed = ulist_alloc(GFP_NOFS);
+	if (!changeset.range_changed)
+		return -ENOMEM;
+
+	ret = clear_record_extent_bits(&BTRFS_I(inode)->io_tree, start, 
+			start + len -1, EXTENT_QGROUP_RESERVED, GFP_NOFS,
+			&changeset);
+	if (ret < 0)
+		goto out;
+
+	if (free) {
+		qgroup_free(BTRFS_I(inode)->root, changeset.bytes_changed);
+		trace_op = QGROUP_FREE;
+	}
+	trace_btrfs_qgroup_release_data(inode, start, len,
+					changeset.bytes_changed, trace_op);
+out:
+	ulist_free(changeset.range_changed);
+	return ret;
+}
+
+/*
+ * Free a reserved space range from io_tree and related qgroups
+ *
+ * Should be called when a range of pages get invalidated before reaching disk.
+ * Or for error cleanup case.
+ *
+ * For data written to disk, use btrfs_qgroup_release_data().
+ *
+ * NOTE: This function may sleep for memory allocation.
+ */
+int btrfs_qgroup_free_data(struct inode *inode, u64 start, u64 len)
+{
+	return __btrfs_qgroup_release_data(inode, start, len, 1);
+}
+
+/*
+ * Release a reserved space range from io_tree only.
+ *
+ * Should be called when a range of pages get written to disk and corresponding
+ * FILE_EXTENT is inserted into corresponding root.
+ *
+ * Since new qgroup accounting framework will only update qgroup numbers at
+ * commit_transaction() time, its reserved space shouldn't be freed from
+ * related qgroups.
+ *
+ * But we should release the range from io_tree, to allow further write to be
+ * COWed.
+ *
+ * NOTE: This function may sleep for memory allocation.
+ */
+int btrfs_qgroup_release_data(struct inode *inode, u64 start, u64 len)
+{
+	return __btrfs_qgroup_release_data(inode, start, len, 0);
+}
+
+int btrfs_qgroup_reserve_meta(struct btrfs_root *root, int num_bytes)
+{
+	int ret;
+
+	if (!root->fs_info->quota_enabled || !is_fstree(root->objectid) ||
+	    num_bytes == 0)
+		return 0;
+
+	BUG_ON(num_bytes != round_down(num_bytes, root->nodesize));
+	ret = qgroup_reserve(root, num_bytes);
+	if (ret < 0)
+		return ret;
+	atomic_add(num_bytes, &root->qgroup_meta_rsv);
+	return ret;
+}
+
+void btrfs_qgroup_free_meta_all(struct btrfs_root *root)
+{
+	int reserved;
+
+	if (!root->fs_info->quota_enabled || !is_fstree(root->objectid))
+		return;
+
+	reserved = atomic_xchg(&root->qgroup_meta_rsv, 0);
+	if (reserved == 0)
+		return;
+	qgroup_free(root, reserved);
+}
+
+void btrfs_qgroup_free_meta(struct btrfs_root *root, int num_bytes)
+{
+	if (!root->fs_info->quota_enabled || !is_fstree(root->objectid))
+		return;
+
+	BUG_ON(num_bytes != round_down(num_bytes, root->nodesize));
+	WARN_ON(atomic_read(&root->qgroup_meta_rsv) < num_bytes);
+	atomic_sub(num_bytes, &root->qgroup_meta_rsv);
+	qgroup_free(root, num_bytes);
+}
+
+/*
+ * Check qgroup reserved space leaking, normally at destory inode
+ * time
+ */
+void btrfs_qgroup_check_reserved_leak(struct inode *inode)
+{
+	struct extent_changeset changeset;
+	struct ulist_node *unode;
+	struct ulist_iterator iter;
+	int ret;
+
+	changeset.bytes_changed = 0;
+	changeset.range_changed = ulist_alloc(GFP_NOFS);
+	if (WARN_ON(!changeset.range_changed))
+		return;
+
+	ret = clear_record_extent_bits(&BTRFS_I(inode)->io_tree, 0, (u64)-1,
+			EXTENT_QGROUP_RESERVED, GFP_NOFS, &changeset);
+
+	WARN_ON(ret < 0);
+	if (WARN_ON(changeset.bytes_changed)) {
+		ULIST_ITER_INIT(&iter);
+		while ((unode = ulist_next(changeset.range_changed, &iter))) {
+			btrfs_warn(BTRFS_I(inode)->root->fs_info,
+				"leaking qgroup reserved space, ino: %lu, start: %llu, end: %llu",
+				inode->i_ino, unode->val, unode->aux);
+		}
+		qgroup_free(BTRFS_I(inode)->root, changeset.bytes_changed);
+	}
+	ulist_free(changeset.range_changed);
+}
diff --git a/kernel/fs/btrfs/qgroup.h b/kernel/fs/btrfs/qgroup.h
index c5242aa9a..ecb2c143e 100644
--- a/kernel/fs/btrfs/qgroup.h
+++ b/kernel/fs/btrfs/qgroup.h
@@ -19,45 +19,27 @@
 #ifndef __BTRFS_QGROUP__
 #define __BTRFS_QGROUP__
 
+#include "ulist.h"
+#include "delayed-ref.h"
+
 /*
- * A description of the operations, all of these operations only happen when we
- * are adding the 1st reference for that subvolume in the case of adding space
- * or on the last reference delete in the case of subtraction.  The only
- * exception is the last one, which is added for confusion.
- *
- * BTRFS_QGROUP_OPER_ADD_EXCL: adding bytes where this subvolume is the only
- * one pointing at the bytes we are adding.  This is called on the first
- * allocation.
- *
- * BTRFS_QGROUP_OPER_ADD_SHARED: adding bytes where this bytenr is going to be
- * shared between subvols.  This is called on the creation of a ref that already
- * has refs from a different subvolume, so basically reflink.
- *
- * BTRFS_QGROUP_OPER_SUB_EXCL: removing bytes where this subvolume is the only
- * one referencing the range.
- *
- * BTRFS_QGROUP_OPER_SUB_SHARED: removing bytes where this subvolume shares with
- * refs with other subvolumes.
+ * Record a dirty extent, and info qgroup to update quota on it
+ * TODO: Use kmem cache to alloc it.
  */
-enum btrfs_qgroup_operation_type {
-	BTRFS_QGROUP_OPER_ADD_EXCL,
-	BTRFS_QGROUP_OPER_ADD_SHARED,
-	BTRFS_QGROUP_OPER_SUB_EXCL,
-	BTRFS_QGROUP_OPER_SUB_SHARED,
-	BTRFS_QGROUP_OPER_SUB_SUBTREE,
-};
-
-struct btrfs_qgroup_operation {
-	u64 ref_root;
+struct btrfs_qgroup_extent_record {
+	struct rb_node node;
 	u64 bytenr;
 	u64 num_bytes;
-	u64 seq;
-	enum btrfs_qgroup_operation_type type;
-	struct seq_list elem;
-	struct rb_node n;
-	struct list_head list;
+	struct ulist *old_roots;
 };
 
+/*
+ * For qgroup event trace points only
+ */
+#define QGROUP_RESERVE		(1<<0)
+#define QGROUP_RELEASE		(1<<1)
+#define QGROUP_FREE		(1<<2)
+
 int btrfs_quota_enable(struct btrfs_trans_handle *trans,
 		       struct btrfs_fs_info *fs_info);
 int btrfs_quota_disable(struct btrfs_trans_handle *trans,
@@ -79,24 +61,35 @@ int btrfs_limit_qgroup(struct btrfs_trans_handle *trans,
 int btrfs_read_qgroup_config(struct btrfs_fs_info *fs_info);
 void btrfs_free_qgroup_config(struct btrfs_fs_info *fs_info);
 struct btrfs_delayed_extent_op;
-int btrfs_qgroup_record_ref(struct btrfs_trans_handle *trans,
-			    struct btrfs_fs_info *fs_info, u64 ref_root,
+int btrfs_qgroup_prepare_account_extents(struct btrfs_trans_handle *trans,
+					 struct btrfs_fs_info *fs_info);
+struct btrfs_qgroup_extent_record
+*btrfs_qgroup_insert_dirty_extent(struct btrfs_delayed_ref_root *delayed_refs,
+				  struct btrfs_qgroup_extent_record *record);
+int
+btrfs_qgroup_account_extent(struct btrfs_trans_handle *trans,
+			    struct btrfs_fs_info *fs_info,
 			    u64 bytenr, u64 num_bytes,
-			    enum btrfs_qgroup_operation_type type,
-			    int mod_seq);
-int btrfs_delayed_qgroup_accounting(struct btrfs_trans_handle *trans,
-				    struct btrfs_fs_info *fs_info);
-void btrfs_remove_qgroup_operation(struct btrfs_trans_handle *trans,
-				   struct btrfs_fs_info *fs_info,
-				   struct btrfs_qgroup_operation *oper);
+			    struct ulist *old_roots, struct ulist *new_roots);
+int btrfs_qgroup_account_extents(struct btrfs_trans_handle *trans,
+				 struct btrfs_fs_info *fs_info);
 int btrfs_run_qgroups(struct btrfs_trans_handle *trans,
 		      struct btrfs_fs_info *fs_info);
 int btrfs_qgroup_inherit(struct btrfs_trans_handle *trans,
 			 struct btrfs_fs_info *fs_info, u64 srcid, u64 objectid,
 			 struct btrfs_qgroup_inherit *inherit);
-int btrfs_qgroup_reserve(struct btrfs_root *root, u64 num_bytes);
-void btrfs_qgroup_free(struct btrfs_root *root, u64 num_bytes);
-
+void btrfs_qgroup_free_refroot(struct btrfs_fs_info *fs_info,
+			       u64 ref_root, u64 num_bytes);
+/*
+ * TODO: Add proper trace point for it, as btrfs_qgroup_free() is
+ * called by everywhere, can't provide good trace for delayed ref case.
+ */
+static inline void btrfs_qgroup_free_delayed_ref(struct btrfs_fs_info *fs_info,
+						 u64 ref_root, u64 num_bytes)
+{
+	btrfs_qgroup_free_refroot(fs_info, ref_root, num_bytes);
+	trace_btrfs_qgroup_free_delayed_ref(ref_root, num_bytes);
+}
 void assert_qgroups_uptodate(struct btrfs_trans_handle *trans);
 
 #ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS
@@ -104,4 +97,13 @@ int btrfs_verify_qgroup_counts(struct btrfs_fs_info *fs_info, u64 qgroupid,
 			       u64 rfer, u64 excl);
 #endif
 
+/* New io_tree based accurate qgroup reserve API */
+int btrfs_qgroup_reserve_data(struct inode *inode, u64 start, u64 len);
+int btrfs_qgroup_release_data(struct inode *inode, u64 start, u64 len);
+int btrfs_qgroup_free_data(struct inode *inode, u64 start, u64 len);
+
+int btrfs_qgroup_reserve_meta(struct btrfs_root *root, int num_bytes);
+void btrfs_qgroup_free_meta_all(struct btrfs_root *root);
+void btrfs_qgroup_free_meta(struct btrfs_root *root, int num_bytes);
+void btrfs_qgroup_check_reserved_leak(struct inode *inode);
 #endif /* __BTRFS_QGROUP__ */
diff --git a/kernel/fs/btrfs/raid56.c b/kernel/fs/btrfs/raid56.c
index fa72068bd..1a33d3eb3 100644
--- a/kernel/fs/btrfs/raid56.c
+++ b/kernel/fs/btrfs/raid56.c
@@ -61,9 +61,10 @@
 #define RBIO_CACHE_SIZE 1024
 
 enum btrfs_rbio_ops {
-	BTRFS_RBIO_WRITE	= 0,
-	BTRFS_RBIO_READ_REBUILD	= 1,
-	BTRFS_RBIO_PARITY_SCRUB	= 2,
+	BTRFS_RBIO_WRITE,
+	BTRFS_RBIO_READ_REBUILD,
+	BTRFS_RBIO_PARITY_SCRUB,
+	BTRFS_RBIO_REBUILD_MISSING,
 };
 
 struct btrfs_raid_bio {
@@ -602,6 +603,10 @@ static int rbio_can_merge(struct btrfs_raid_bio *last,
 	    cur->operation == BTRFS_RBIO_PARITY_SCRUB)
 		return 0;
 
+	if (last->operation == BTRFS_RBIO_REBUILD_MISSING ||
+	    cur->operation == BTRFS_RBIO_REBUILD_MISSING)
+		return 0;
+
 	return 1;
 }
 
@@ -793,7 +798,10 @@ static noinline void unlock_stripe(struct btrfs_raid_bio *rbio)
 
 			if (next->operation == BTRFS_RBIO_READ_REBUILD)
 				async_read_rebuild(next);
-			else if (next->operation == BTRFS_RBIO_WRITE) {
+			else if (next->operation == BTRFS_RBIO_REBUILD_MISSING) {
+				steal_rbio(rbio, next);
+				async_read_rebuild(next);
+			} else if (next->operation == BTRFS_RBIO_WRITE) {
 				steal_rbio(rbio, next);
 				async_rmw_stripe(next);
 			} else if (next->operation == BTRFS_RBIO_PARITY_SCRUB) {
@@ -802,7 +810,11 @@ static noinline void unlock_stripe(struct btrfs_raid_bio *rbio)
 			}
 
 			goto done_nolock;
-		} else  if (waitqueue_active(&h->wait)) {
+			/*
+			 * The barrier for this waitqueue_active is not needed,
+			 * we're protected by h->lock and can't miss a wakeup.
+			 */
+		} else if (waitqueue_active(&h->wait)) {
 			spin_unlock(&rbio->bio_list_lock);
 			spin_unlock_irqrestore(&h->lock, flags);
 			wake_up(&h->wait);
@@ -851,7 +863,7 @@ static void free_raid_bio(struct btrfs_raid_bio *rbio)
  * this frees the rbio and runs through all the bios in the
  * bio_list and calls end_io on them
  */
-static void rbio_orig_end_io(struct btrfs_raid_bio *rbio, int err, int uptodate)
+static void rbio_orig_end_io(struct btrfs_raid_bio *rbio, int err)
 {
 	struct bio *cur = bio_list_get(&rbio->bio_list);
 	struct bio *next;
@@ -864,9 +876,8 @@ static void rbio_orig_end_io(struct btrfs_raid_bio *rbio, int err, int uptodate)
 	while (cur) {
 		next = cur->bi_next;
 		cur->bi_next = NULL;
-		if (uptodate)
-			set_bit(BIO_UPTODATE, &cur->bi_flags);
-		bio_endio(cur, err);
+		cur->bi_error = err;
+		bio_endio(cur);
 		cur = next;
 	}
 }
@@ -875,9 +886,10 @@ static void rbio_orig_end_io(struct btrfs_raid_bio *rbio, int err, int uptodate)
  * end io function used by finish_rmw.  When we finally
  * get here, we've written a full stripe
  */
-static void raid_write_end_io(struct bio *bio, int err)
+static void raid_write_end_io(struct bio *bio)
 {
 	struct btrfs_raid_bio *rbio = bio->bi_private;
+	int err = bio->bi_error;
 
 	if (err)
 		fail_bio_stripe(rbio, bio);
@@ -893,7 +905,7 @@ static void raid_write_end_io(struct bio *bio, int err)
 	if (atomic_read(&rbio->error) > rbio->bbio->max_errors)
 		err = -EIO;
 
-	rbio_orig_end_io(rbio, err, 0);
+	rbio_orig_end_io(rbio, err);
 	return;
 }
 
@@ -1071,7 +1083,7 @@ static int rbio_add_io_page(struct btrfs_raid_bio *rbio,
 		 * devices or if they are not contiguous
 		 */
 		if (last_end == disk_start && stripe->dev->bdev &&
-		    test_bit(BIO_UPTODATE, &last->bi_flags) &&
+		    !last->bi_error &&
 		    last->bi_bdev == stripe->dev->bdev) {
 			ret = bio_add_page(last, page, PAGE_CACHE_SIZE, 0);
 			if (ret == PAGE_CACHE_SIZE)
@@ -1087,7 +1099,6 @@ static int rbio_add_io_page(struct btrfs_raid_bio *rbio,
 	bio->bi_iter.bi_size = 0;
 	bio->bi_bdev = stripe->dev->bdev;
 	bio->bi_iter.bi_sector = disk_start >> 9;
-	set_bit(BIO_UPTODATE, &bio->bi_flags);
 
 	bio_add_page(bio, page, PAGE_CACHE_SIZE, 0);
 	bio_list_add(bio_list, bio);
@@ -1312,13 +1323,12 @@ write_data:
 
 		bio->bi_private = rbio;
 		bio->bi_end_io = raid_write_end_io;
-		BUG_ON(!test_bit(BIO_UPTODATE, &bio->bi_flags));
 		submit_bio(WRITE, bio);
 	}
 	return;
 
 cleanup:
-	rbio_orig_end_io(rbio, -EIO, 0);
+	rbio_orig_end_io(rbio, -EIO);
 }
 
 /*
@@ -1441,11 +1451,11 @@ static void set_bio_pages_uptodate(struct bio *bio)
  * This will usually kick off finish_rmw once all the bios are read in, but it
  * may trigger parity reconstruction if we had any errors along the way
  */
-static void raid_rmw_end_io(struct bio *bio, int err)
+static void raid_rmw_end_io(struct bio *bio)
 {
 	struct btrfs_raid_bio *rbio = bio->bi_private;
 
-	if (err)
+	if (bio->bi_error)
 		fail_bio_stripe(rbio, bio);
 	else
 		set_bio_pages_uptodate(bio);
@@ -1455,7 +1465,6 @@ static void raid_rmw_end_io(struct bio *bio, int err)
 	if (!atomic_dec_and_test(&rbio->stripes_pending))
 		return;
 
-	err = 0;
 	if (atomic_read(&rbio->error) > rbio->bbio->max_errors)
 		goto cleanup;
 
@@ -1469,7 +1478,7 @@ static void raid_rmw_end_io(struct bio *bio, int err)
 
 cleanup:
 
-	rbio_orig_end_io(rbio, -EIO, 0);
+	rbio_orig_end_io(rbio, -EIO);
 }
 
 static void async_rmw_stripe(struct btrfs_raid_bio *rbio)
@@ -1572,14 +1581,13 @@ static int raid56_rmw_stripe(struct btrfs_raid_bio *rbio)
 		btrfs_bio_wq_end_io(rbio->fs_info, bio,
 				    BTRFS_WQ_ENDIO_RAID56);
 
-		BUG_ON(!test_bit(BIO_UPTODATE, &bio->bi_flags));
 		submit_bio(READ, bio);
 	}
 	/* the actual write will happen once the reads are done */
 	return 0;
 
 cleanup:
-	rbio_orig_end_io(rbio, -EIO, 0);
+	rbio_orig_end_io(rbio, -EIO);
 	return -EIO;
 
 finish:
@@ -1809,7 +1817,8 @@ static void __raid_recover_end_io(struct btrfs_raid_bio *rbio)
 	faila = rbio->faila;
 	failb = rbio->failb;
 
-	if (rbio->operation == BTRFS_RBIO_READ_REBUILD) {
+	if (rbio->operation == BTRFS_RBIO_READ_REBUILD ||
+	    rbio->operation == BTRFS_RBIO_REBUILD_MISSING) {
 		spin_lock_irq(&rbio->bio_list_lock);
 		set_bit(RBIO_RMW_LOCKED_BIT, &rbio->flags);
 		spin_unlock_irq(&rbio->bio_list_lock);
@@ -1834,7 +1843,8 @@ static void __raid_recover_end_io(struct btrfs_raid_bio *rbio)
 			 * if we're rebuilding a read, we have to use
 			 * pages from the bio list
 			 */
-			if (rbio->operation == BTRFS_RBIO_READ_REBUILD &&
+			if ((rbio->operation == BTRFS_RBIO_READ_REBUILD ||
+			     rbio->operation == BTRFS_RBIO_REBUILD_MISSING) &&
 			    (stripe == faila || stripe == failb)) {
 				page = page_in_rbio(rbio, stripe, pagenr, 0);
 			} else {
@@ -1943,7 +1953,8 @@ pstripe:
 			 * if we're rebuilding a read, we have to use
 			 * pages from the bio list
 			 */
-			if (rbio->operation == BTRFS_RBIO_READ_REBUILD &&
+			if ((rbio->operation == BTRFS_RBIO_READ_REBUILD ||
+			     rbio->operation == BTRFS_RBIO_REBUILD_MISSING) &&
 			    (stripe == faila || stripe == failb)) {
 				page = page_in_rbio(rbio, stripe, pagenr, 0);
 			} else {
@@ -1964,7 +1975,9 @@ cleanup_io:
 		else
 			clear_bit(RBIO_CACHE_READY_BIT, &rbio->flags);
 
-		rbio_orig_end_io(rbio, err, err == 0);
+		rbio_orig_end_io(rbio, err);
+	} else if (rbio->operation == BTRFS_RBIO_REBUILD_MISSING) {
+		rbio_orig_end_io(rbio, err);
 	} else if (err == 0) {
 		rbio->faila = -1;
 		rbio->failb = -1;
@@ -1976,7 +1989,7 @@ cleanup_io:
 		else
 			BUG();
 	} else {
-		rbio_orig_end_io(rbio, err, 0);
+		rbio_orig_end_io(rbio, err);
 	}
 }
 
@@ -1984,7 +1997,7 @@ cleanup_io:
  * This is called only for stripes we've read from disk to
  * reconstruct the parity.
  */
-static void raid_recover_end_io(struct bio *bio, int err)
+static void raid_recover_end_io(struct bio *bio)
 {
 	struct btrfs_raid_bio *rbio = bio->bi_private;
 
@@ -1992,7 +2005,7 @@ static void raid_recover_end_io(struct bio *bio, int err)
 	 * we only read stripe pages off the disk, set them
 	 * up to date if there were no errors
 	 */
-	if (err)
+	if (bio->bi_error)
 		fail_bio_stripe(rbio, bio);
 	else
 		set_bio_pages_uptodate(bio);
@@ -2002,7 +2015,7 @@ static void raid_recover_end_io(struct bio *bio, int err)
 		return;
 
 	if (atomic_read(&rbio->error) > rbio->bbio->max_errors)
-		rbio_orig_end_io(rbio, -EIO, 0);
+		rbio_orig_end_io(rbio, -EIO);
 	else
 		__raid_recover_end_io(rbio);
 }
@@ -2094,15 +2107,15 @@ static int __raid56_parity_recover(struct btrfs_raid_bio *rbio)
 		btrfs_bio_wq_end_io(rbio->fs_info, bio,
 				    BTRFS_WQ_ENDIO_RAID56);
 
-		BUG_ON(!test_bit(BIO_UPTODATE, &bio->bi_flags));
 		submit_bio(READ, bio);
 	}
 out:
 	return 0;
 
 cleanup:
-	if (rbio->operation == BTRFS_RBIO_READ_REBUILD)
-		rbio_orig_end_io(rbio, -EIO, 0);
+	if (rbio->operation == BTRFS_RBIO_READ_REBUILD ||
+	    rbio->operation == BTRFS_RBIO_REBUILD_MISSING)
+		rbio_orig_end_io(rbio, -EIO);
 	return -EIO;
 }
 
@@ -2232,8 +2245,9 @@ raid56_parity_alloc_scrub_rbio(struct btrfs_root *root, struct bio *bio,
 	return rbio;
 }
 
-void raid56_parity_add_scrub_pages(struct btrfs_raid_bio *rbio,
-				   struct page *page, u64 logical)
+/* Used for both parity scrub and missing. */
+void raid56_add_scrub_pages(struct btrfs_raid_bio *rbio, struct page *page,
+			    u64 logical)
 {
 	int stripe_offset;
 	int index;
@@ -2277,11 +2291,12 @@ static int alloc_rbio_essential_pages(struct btrfs_raid_bio *rbio)
  * end io function used by finish_rmw.  When we finally
  * get here, we've written a full stripe
  */
-static void raid_write_parity_end_io(struct bio *bio, int err)
+static void raid_write_parity_end_io(struct bio *bio)
 {
 	struct btrfs_raid_bio *rbio = bio->bi_private;
+	int err = bio->bi_error;
 
-	if (err)
+	if (bio->bi_error)
 		fail_bio_stripe(rbio, bio);
 
 	bio_put(bio);
@@ -2294,7 +2309,7 @@ static void raid_write_parity_end_io(struct bio *bio, int err)
 	if (atomic_read(&rbio->error))
 		err = -EIO;
 
-	rbio_orig_end_io(rbio, err, 0);
+	rbio_orig_end_io(rbio, err);
 }
 
 static noinline void finish_parity_scrub(struct btrfs_raid_bio *rbio,
@@ -2437,7 +2452,7 @@ submit_write:
 	nr_data = bio_list_size(&bio_list);
 	if (!nr_data) {
 		/* Every parity is right */
-		rbio_orig_end_io(rbio, 0, 0);
+		rbio_orig_end_io(rbio, 0);
 		return;
 	}
 
@@ -2450,13 +2465,12 @@ submit_write:
 
 		bio->bi_private = rbio;
 		bio->bi_end_io = raid_write_parity_end_io;
-		BUG_ON(!test_bit(BIO_UPTODATE, &bio->bi_flags));
 		submit_bio(WRITE, bio);
 	}
 	return;
 
 cleanup:
-	rbio_orig_end_io(rbio, -EIO, 0);
+	rbio_orig_end_io(rbio, -EIO);
 }
 
 static inline int is_data_stripe(struct btrfs_raid_bio *rbio, int stripe)
@@ -2524,7 +2538,7 @@ static void validate_rbio_for_parity_scrub(struct btrfs_raid_bio *rbio)
 	return;
 
 cleanup:
-	rbio_orig_end_io(rbio, -EIO, 0);
+	rbio_orig_end_io(rbio, -EIO);
 }
 
 /*
@@ -2535,11 +2549,11 @@ cleanup:
  * This will usually kick off finish_rmw once all the bios are read in, but it
  * may trigger parity reconstruction if we had any errors along the way
  */
-static void raid56_parity_scrub_end_io(struct bio *bio, int err)
+static void raid56_parity_scrub_end_io(struct bio *bio)
 {
 	struct btrfs_raid_bio *rbio = bio->bi_private;
 
-	if (err)
+	if (bio->bi_error)
 		fail_bio_stripe(rbio, bio);
 	else
 		set_bio_pages_uptodate(bio);
@@ -2632,14 +2646,13 @@ static void raid56_parity_scrub_stripe(struct btrfs_raid_bio *rbio)
 		btrfs_bio_wq_end_io(rbio->fs_info, bio,
 				    BTRFS_WQ_ENDIO_RAID56);
 
-		BUG_ON(!test_bit(BIO_UPTODATE, &bio->bi_flags));
 		submit_bio(READ, bio);
 	}
 	/* the actual write will happen once the reads are done */
 	return;
 
 cleanup:
-	rbio_orig_end_io(rbio, -EIO, 0);
+	rbio_orig_end_io(rbio, -EIO);
 	return;
 
 finish:
@@ -2668,3 +2681,55 @@ void raid56_parity_submit_scrub_rbio(struct btrfs_raid_bio *rbio)
 	if (!lock_stripe_add(rbio))
 		async_scrub_parity(rbio);
 }
+
+/* The following code is used for dev replace of a missing RAID 5/6 device. */
+
+struct btrfs_raid_bio *
+raid56_alloc_missing_rbio(struct btrfs_root *root, struct bio *bio,
+			  struct btrfs_bio *bbio, u64 length)
+{
+	struct btrfs_raid_bio *rbio;
+
+	rbio = alloc_rbio(root, bbio, length);
+	if (IS_ERR(rbio))
+		return NULL;
+
+	rbio->operation = BTRFS_RBIO_REBUILD_MISSING;
+	bio_list_add(&rbio->bio_list, bio);
+	/*
+	 * This is a special bio which is used to hold the completion handler
+	 * and make the scrub rbio is similar to the other types
+	 */
+	ASSERT(!bio->bi_iter.bi_size);
+
+	rbio->faila = find_logical_bio_stripe(rbio, bio);
+	if (rbio->faila == -1) {
+		BUG();
+		kfree(rbio);
+		return NULL;
+	}
+
+	return rbio;
+}
+
+static void missing_raid56_work(struct btrfs_work *work)
+{
+	struct btrfs_raid_bio *rbio;
+
+	rbio = container_of(work, struct btrfs_raid_bio, work);
+	__raid56_parity_recover(rbio);
+}
+
+static void async_missing_raid56(struct btrfs_raid_bio *rbio)
+{
+	btrfs_init_work(&rbio->work, btrfs_rmw_helper,
+			missing_raid56_work, NULL, NULL);
+
+	btrfs_queue_work(rbio->fs_info->rmw_workers, &rbio->work);
+}
+
+void raid56_submit_missing_rbio(struct btrfs_raid_bio *rbio)
+{
+	if (!lock_stripe_add(rbio))
+		async_missing_raid56(rbio);
+}
diff --git a/kernel/fs/btrfs/raid56.h b/kernel/fs/btrfs/raid56.h
index 2b5d7977d..8b694699d 100644
--- a/kernel/fs/btrfs/raid56.h
+++ b/kernel/fs/btrfs/raid56.h
@@ -48,15 +48,21 @@ int raid56_parity_recover(struct btrfs_root *root, struct bio *bio,
 int raid56_parity_write(struct btrfs_root *root, struct bio *bio,
 			       struct btrfs_bio *bbio, u64 stripe_len);
 
+void raid56_add_scrub_pages(struct btrfs_raid_bio *rbio, struct page *page,
+			    u64 logical);
+
 struct btrfs_raid_bio *
 raid56_parity_alloc_scrub_rbio(struct btrfs_root *root, struct bio *bio,
 			       struct btrfs_bio *bbio, u64 stripe_len,
 			       struct btrfs_device *scrub_dev,
 			       unsigned long *dbitmap, int stripe_nsectors);
-void raid56_parity_add_scrub_pages(struct btrfs_raid_bio *rbio,
-				   struct page *page, u64 logical);
 void raid56_parity_submit_scrub_rbio(struct btrfs_raid_bio *rbio);
 
+struct btrfs_raid_bio *
+raid56_alloc_missing_rbio(struct btrfs_root *root, struct bio *bio,
+			  struct btrfs_bio *bbio, u64 length);
+void raid56_submit_missing_rbio(struct btrfs_raid_bio *rbio);
+
 int btrfs_alloc_stripe_hash_table(struct btrfs_fs_info *info);
 void btrfs_free_stripe_hash_table(struct btrfs_fs_info *info);
 #endif
diff --git a/kernel/fs/btrfs/reada.c b/kernel/fs/btrfs/reada.c
index 0e7beea92..619f92963 100644
--- a/kernel/fs/btrfs/reada.c
+++ b/kernel/fs/btrfs/reada.c
@@ -328,6 +328,7 @@ static struct reada_extent *reada_find_extent(struct btrfs_root *root,
 	struct btrfs_device *prev_dev;
 	u32 blocksize;
 	u64 length;
+	int real_stripes;
 	int nzones = 0;
 	int i;
 	unsigned long index = logical >> PAGE_CACHE_SHIFT;
@@ -369,7 +370,8 @@ static struct reada_extent *reada_find_extent(struct btrfs_root *root,
 		goto error;
 	}
 
-	for (nzones = 0; nzones < bbio->num_stripes; ++nzones) {
+	real_stripes = bbio->num_stripes - bbio->num_tgtdevs;
+	for (nzones = 0; nzones < real_stripes; ++nzones) {
 		struct reada_zone *zone;
 
 		dev = bbio->stripes[nzones].dev;
@@ -567,7 +569,7 @@ static int reada_add_block(struct reada_control *rc, u64 logical,
 	rec = kzalloc(sizeof(*rec), GFP_NOFS);
 	if (!rec) {
 		reada_extent_put(root->fs_info, re);
-		return -1;
+		return -ENOMEM;
 	}
 
 	rec->rc = rc;
@@ -916,6 +918,7 @@ struct reada_control *btrfs_reada_add(struct btrfs_root *root,
 	u64 start;
 	u64 generation;
 	int level;
+	int ret;
 	struct extent_buffer *node;
 	static struct btrfs_key max_key = {
 		.objectid = (u64)-1,
@@ -941,9 +944,10 @@ struct reada_control *btrfs_reada_add(struct btrfs_root *root,
 	generation = btrfs_header_generation(node);
 	free_extent_buffer(node);
 
-	if (reada_add_block(rc, start, &max_key, level, generation)) {
+	ret = reada_add_block(rc, start, &max_key, level, generation);
+	if (ret) {
 		kfree(rc);
-		return ERR_PTR(-ENOMEM);
+		return ERR_PTR(ret);
 	}
 
 	reada_start_machine(root->fs_info);
diff --git a/kernel/fs/btrfs/relocation.c b/kernel/fs/btrfs/relocation.c
index 74b24b01d..b4ca5454e 100644
--- a/kernel/fs/btrfs/relocation.c
+++ b/kernel/fs/btrfs/relocation.c
@@ -1716,7 +1716,7 @@ int replace_file_extents(struct btrfs_trans_handle *trans,
 		ret = btrfs_inc_extent_ref(trans, root, new_bytenr,
 					   num_bytes, parent,
 					   btrfs_header_owner(leaf),
-					   key.objectid, key.offset, 1);
+					   key.objectid, key.offset);
 		if (ret) {
 			btrfs_abort_transaction(trans, root, ret);
 			break;
@@ -1724,7 +1724,7 @@ int replace_file_extents(struct btrfs_trans_handle *trans,
 
 		ret = btrfs_free_extent(trans, root, bytenr, num_bytes,
 					parent, btrfs_header_owner(leaf),
-					key.objectid, key.offset, 1);
+					key.objectid, key.offset);
 		if (ret) {
 			btrfs_abort_transaction(trans, root, ret);
 			break;
@@ -1847,8 +1847,10 @@ again:
 			}
 
 			eb = read_tree_block(dest, old_bytenr, old_ptr_gen);
-			if (!eb || !extent_buffer_uptodate(eb)) {
-				ret = (!eb) ? -ENOMEM : -EIO;
+			if (IS_ERR(eb)) {
+				ret = PTR_ERR(eb);
+			} else if (!extent_buffer_uptodate(eb)) {
+				ret = -EIO;
 				free_extent_buffer(eb);
 				break;
 			}
@@ -1898,23 +1900,21 @@ again:
 
 		ret = btrfs_inc_extent_ref(trans, src, old_bytenr, blocksize,
 					path->nodes[level]->start,
-					src->root_key.objectid, level - 1, 0,
-					1);
+					src->root_key.objectid, level - 1, 0);
 		BUG_ON(ret);
 		ret = btrfs_inc_extent_ref(trans, dest, new_bytenr, blocksize,
 					0, dest->root_key.objectid, level - 1,
-					0, 1);
+					0);
 		BUG_ON(ret);
 
 		ret = btrfs_free_extent(trans, src, new_bytenr, blocksize,
 					path->nodes[level]->start,
-					src->root_key.objectid, level - 1, 0,
-					1);
+					src->root_key.objectid, level - 1, 0);
 		BUG_ON(ret);
 
 		ret = btrfs_free_extent(trans, dest, old_bytenr, blocksize,
 					0, dest->root_key.objectid, level - 1,
-					0, 1);
+					0);
 		BUG_ON(ret);
 
 		btrfs_unlock_up_safe(path, 0);
@@ -2002,7 +2002,9 @@ int walk_down_reloc_tree(struct btrfs_root *root, struct btrfs_path *path,
 
 		bytenr = btrfs_node_blockptr(eb, path->slots[i]);
 		eb = read_tree_block(root, bytenr, ptr_gen);
-		if (!eb || !extent_buffer_uptodate(eb)) {
+		if (IS_ERR(eb)) {
+			return PTR_ERR(eb);
+		} else if (!extent_buffer_uptodate(eb)) {
 			free_extent_buffer(eb);
 			return -EIO;
 		}
@@ -2414,7 +2416,7 @@ again:
 	}
 out:
 	if (ret) {
-		btrfs_std_error(root->fs_info, ret);
+		btrfs_std_error(root->fs_info, ret, NULL);
 		if (!list_empty(&reloc_roots))
 			free_reloc_roots(&reloc_roots);
 
@@ -2519,8 +2521,7 @@ struct btrfs_root *select_reloc_root(struct btrfs_trans_handle *trans,
  * counted. return -ENOENT if the block is root of reloc tree.
  */
 static noinline_for_stack
-struct btrfs_root *select_one_root(struct btrfs_trans_handle *trans,
-				   struct backref_node *node)
+struct btrfs_root *select_one_root(struct backref_node *node)
 {
 	struct backref_node *next;
 	struct btrfs_root *root;
@@ -2710,7 +2711,10 @@ static int do_relocation(struct btrfs_trans_handle *trans,
 		blocksize = root->nodesize;
 		generation = btrfs_node_ptr_generation(upper->eb, slot);
 		eb = read_tree_block(root, bytenr, generation);
-		if (!eb || !extent_buffer_uptodate(eb)) {
+		if (IS_ERR(eb)) {
+			err = PTR_ERR(eb);
+			goto next;
+		} else if (!extent_buffer_uptodate(eb)) {
 			free_extent_buffer(eb);
 			err = -EIO;
 			goto next;
@@ -2739,7 +2743,7 @@ static int do_relocation(struct btrfs_trans_handle *trans,
 						node->eb->start, blocksize,
 						upper->eb->start,
 						btrfs_header_owner(upper->eb),
-						node->level, 0, 1);
+						node->level, 0);
 			BUG_ON(ret);
 
 			ret = btrfs_drop_subtree(trans, root, eb, upper->eb);
@@ -2873,7 +2877,9 @@ static int get_tree_block_key(struct reloc_control *rc,
 	BUG_ON(block->key_ready);
 	eb = read_tree_block(rc->extent_root, block->bytenr,
 			     block->key.offset);
-	if (!eb || !extent_buffer_uptodate(eb)) {
+	if (IS_ERR(eb)) {
+		return PTR_ERR(eb);
+	} else if (!extent_buffer_uptodate(eb)) {
 		free_extent_buffer(eb);
 		return -EIO;
 	}
@@ -2903,7 +2909,7 @@ static int relocate_tree_block(struct btrfs_trans_handle *trans,
 		return 0;
 
 	BUG_ON(node->processed);
-	root = select_one_root(trans, node);
+	root = select_one_root(node);
 	if (root == ERR_PTR(-ENOENT)) {
 		update_processed_blocks(rc, node);
 		goto out;
@@ -3026,8 +3032,8 @@ int prealloc_file_extent_cluster(struct inode *inode,
 	BUG_ON(cluster->start != cluster->boundary[0]);
 	mutex_lock(&inode->i_mutex);
 
-	ret = btrfs_check_data_free_space(inode, cluster->end +
-					  1 - cluster->start, 0);
+	ret = btrfs_check_data_free_space(inode, cluster->start,
+					  cluster->end + 1 - cluster->start);
 	if (ret)
 		goto out;
 
@@ -3048,8 +3054,8 @@ int prealloc_file_extent_cluster(struct inode *inode,
 			break;
 		nr++;
 	}
-	btrfs_free_reserved_data_space(inode, cluster->end +
-				       1 - cluster->start);
+	btrfs_free_reserved_data_space(inode, cluster->start,
+				       cluster->end + 1 - cluster->start);
 out:
 	mutex_unlock(&inode->i_mutex);
 	return ret;
@@ -3746,8 +3752,7 @@ out:
  * helper to find next unprocessed extent
  */
 static noinline_for_stack
-int find_next_extent(struct btrfs_trans_handle *trans,
-		     struct reloc_control *rc, struct btrfs_path *path,
+int find_next_extent(struct reloc_control *rc, struct btrfs_path *path,
 		     struct btrfs_key *extent_key)
 {
 	struct btrfs_key key;
@@ -3942,7 +3947,7 @@ restart:
 			continue;
 		}
 
-		ret = find_next_extent(trans, rc, path, &key);
+		ret = find_next_extent(rc, path, &key);
 		if (ret < 0)
 			err = ret;
 		if (ret != 0)
@@ -3967,6 +3972,10 @@ restart:
 			       sizeof(struct btrfs_extent_item_v0));
 			ret = get_ref_objectid_v0(rc, path, &key, &ref_owner,
 						  &path_change);
+			if (ret < 0) {
+				err = ret;
+				break;
+			}
 			if (ref_owner < BTRFS_FIRST_FREE_OBJECTID)
 				flags = BTRFS_EXTENT_FLAG_TREE_BLOCK;
 			else
@@ -4040,7 +4049,7 @@ restart:
 	if (trans && progress && err == -ENOSPC) {
 		ret = btrfs_force_chunk_alloc(trans, rc->extent_root,
 					      rc->block_group->flags);
-		if (ret == 0) {
+		if (ret == 1) {
 			err = 0;
 			progress = 0;
 			goto restart;
@@ -4131,7 +4140,7 @@ struct inode *create_reloc_inode(struct btrfs_fs_info *fs_info,
 	struct btrfs_trans_handle *trans;
 	struct btrfs_root *root;
 	struct btrfs_key key;
-	u64 objectid = BTRFS_FIRST_FREE_OBJECTID;
+	u64 objectid;
 	int err = 0;
 
 	root = read_fs_root(fs_info, BTRFS_DATA_RELOC_TREE_OBJECTID);
@@ -4206,14 +4215,12 @@ int btrfs_relocate_block_group(struct btrfs_root *extent_root, u64 group_start)
 	rc->block_group = btrfs_lookup_block_group(fs_info, group_start);
 	BUG_ON(!rc->block_group);
 
-	if (!rc->block_group->ro) {
-		ret = btrfs_set_block_group_ro(extent_root, rc->block_group);
-		if (ret) {
-			err = ret;
-			goto out;
-		}
-		rw = 1;
+	ret = btrfs_inc_block_group_ro(extent_root, rc->block_group);
+	if (ret) {
+		err = ret;
+		goto out;
 	}
+	rw = 1;
 
 	path = btrfs_alloc_path();
 	if (!path) {
@@ -4285,7 +4292,7 @@ int btrfs_relocate_block_group(struct btrfs_root *extent_root, u64 group_start)
 	WARN_ON(btrfs_block_group_used(&rc->block_group->item) > 0);
 out:
 	if (err && rw)
-		btrfs_set_block_group_rw(extent_root, rc->block_group);
+		btrfs_dec_block_group_ro(extent_root, rc->block_group);
 	iput(rc->data_inode);
 	btrfs_put_block_group(rc->block_group);
 	kfree(rc);
@@ -4585,8 +4592,7 @@ int btrfs_reloc_cow_block(struct btrfs_trans_handle *trans,
  * called before creating snapshot. it calculates metadata reservation
  * requried for relocating tree blocks in the snapshot
  */
-void btrfs_reloc_pre_snapshot(struct btrfs_trans_handle *trans,
-			      struct btrfs_pending_snapshot *pending,
+void btrfs_reloc_pre_snapshot(struct btrfs_pending_snapshot *pending,
 			      u64 *bytes_to_reserve)
 {
 	struct btrfs_root *root;
diff --git a/kernel/fs/btrfs/root-tree.c b/kernel/fs/btrfs/root-tree.c
index 360a728a6..2c849b08a 100644
--- a/kernel/fs/btrfs/root-tree.c
+++ b/kernel/fs/btrfs/root-tree.c
@@ -45,12 +45,13 @@ static void btrfs_read_root_item(struct extent_buffer *eb, int slot,
 	if (!need_reset && btrfs_root_generation(item)
 		!= btrfs_root_generation_v2(item)) {
 		if (btrfs_root_generation_v2(item) != 0) {
-			printk(KERN_WARNING "BTRFS: mismatching "
+			btrfs_warn(eb->fs_info,
+					"mismatching "
 					"generation and generation_v2 "
 					"found in root item. This root "
 					"was probably mounted with an "
 					"older kernel. Resetting all "
-					"new fields.\n");
+					"new fields.");
 		}
 		need_reset = 1;
 	}
@@ -141,7 +142,7 @@ int btrfs_update_root(struct btrfs_trans_handle *trans, struct btrfs_root
 	int ret;
 	int slot;
 	unsigned long ptr;
-	int old_len;
+	u32 old_len;
 
 	path = btrfs_alloc_path();
 	if (!path)
@@ -283,7 +284,7 @@ int btrfs_find_orphan_roots(struct btrfs_root *tree_root)
 			trans = btrfs_join_transaction(tree_root);
 			if (IS_ERR(trans)) {
 				err = PTR_ERR(trans);
-				btrfs_error(tree_root->fs_info, err,
+				btrfs_std_error(tree_root->fs_info, err,
 					    "Failed to start trans to delete "
 					    "orphan item");
 				break;
@@ -292,7 +293,7 @@ int btrfs_find_orphan_roots(struct btrfs_root *tree_root)
 						    root_key.objectid);
 			btrfs_end_transaction(trans, tree_root);
 			if (err) {
-				btrfs_error(tree_root->fs_info, err,
+				btrfs_std_error(tree_root->fs_info, err,
 					    "Failed to delete root orphan "
 					    "item");
 				break;
@@ -309,8 +310,16 @@ int btrfs_find_orphan_roots(struct btrfs_root *tree_root)
 		set_bit(BTRFS_ROOT_ORPHAN_ITEM_INSERTED, &root->state);
 
 		err = btrfs_insert_fs_root(root->fs_info, root);
+		/*
+		 * The root might have been inserted already, as before we look
+		 * for orphan roots, log replay might have happened, which
+		 * triggers a transaction commit and qgroup accounting, which
+		 * in turn reads and inserts fs roots while doing backref
+		 * walking.
+		 */
+		if (err == -EEXIST)
+			err = 0;
 		if (err) {
-			BUG_ON(err == -EEXIST);
 			btrfs_free_fs_root(root);
 			break;
 		}
diff --git a/kernel/fs/btrfs/scrub.c b/kernel/fs/btrfs/scrub.c
index ab5811545..b091d94ce 100644
--- a/kernel/fs/btrfs/scrub.c
+++ b/kernel/fs/btrfs/scrub.c
@@ -125,6 +125,7 @@ struct scrub_block {
 		/* It is for the data with checksum */
 		unsigned int	data_corrected:1;
 	};
+	struct btrfs_work	work;
 };
 
 /* Used for the chunks with parity stripe such RAID5/6 */
@@ -247,14 +248,9 @@ static int scrub_handle_errored_block(struct scrub_block *sblock_to_check);
 static int scrub_setup_recheck_block(struct scrub_block *original_sblock,
 				     struct scrub_block *sblocks_for_recheck);
 static void scrub_recheck_block(struct btrfs_fs_info *fs_info,
-				struct scrub_block *sblock, int is_metadata,
-				int have_csum, u8 *csum, u64 generation,
-				u16 csum_size, int retry_failed_mirror);
-static void scrub_recheck_block_checksum(struct btrfs_fs_info *fs_info,
-					 struct scrub_block *sblock,
-					 int is_metadata, int have_csum,
-					 const u8 *csum, u64 generation,
-					 u16 csum_size);
+				struct scrub_block *sblock,
+				int retry_failed_mirror);
+static void scrub_recheck_block_checksum(struct scrub_block *sblock);
 static int scrub_repair_block_from_good_copy(struct scrub_block *sblock_bad,
 					     struct scrub_block *sblock_good);
 static int scrub_repair_page_from_good_copy(struct scrub_block *sblock_bad,
@@ -278,7 +274,7 @@ static int scrub_pages(struct scrub_ctx *sctx, u64 logical, u64 len,
 		       u64 physical, struct btrfs_device *dev, u64 flags,
 		       u64 gen, int mirror_num, u8 *csum, int force,
 		       u64 physical_for_dev_replace);
-static void scrub_bio_end_io(struct bio *bio, int err);
+static void scrub_bio_end_io(struct bio *bio);
 static void scrub_bio_end_io_worker(struct btrfs_work *work);
 static void scrub_block_complete(struct scrub_block *sblock);
 static void scrub_remap_extent(struct btrfs_fs_info *fs_info,
@@ -295,7 +291,7 @@ static void scrub_free_wr_ctx(struct scrub_wr_ctx *wr_ctx);
 static int scrub_add_page_to_wr_bio(struct scrub_ctx *sctx,
 				    struct scrub_page *spage);
 static void scrub_wr_submit(struct scrub_ctx *sctx);
-static void scrub_wr_bio_end_io(struct bio *bio, int err);
+static void scrub_wr_bio_end_io(struct bio *bio);
 static void scrub_wr_bio_end_io_worker(struct btrfs_work *work);
 static int write_page_nocow(struct scrub_ctx *sctx,
 			    u64 physical_for_dev_replace, struct page *page);
@@ -332,11 +328,14 @@ static void __scrub_blocked_if_needed(struct btrfs_fs_info *fs_info)
 	}
 }
 
-static void scrub_blocked_if_needed(struct btrfs_fs_info *fs_info)
+static void scrub_pause_on(struct btrfs_fs_info *fs_info)
 {
 	atomic_inc(&fs_info->scrubs_paused);
 	wake_up(&fs_info->scrub_pause_wait);
+}
 
+static void scrub_pause_off(struct btrfs_fs_info *fs_info)
+{
 	mutex_lock(&fs_info->scrub_lock);
 	__scrub_blocked_if_needed(fs_info);
 	atomic_dec(&fs_info->scrubs_paused);
@@ -345,6 +344,12 @@ static void scrub_blocked_if_needed(struct btrfs_fs_info *fs_info)
 	wake_up(&fs_info->scrub_pause_wait);
 }
 
+static void scrub_blocked_if_needed(struct btrfs_fs_info *fs_info)
+{
+	scrub_pause_on(fs_info);
+	scrub_pause_off(fs_info);
+}
+
 /*
  * used for workers that require transaction commits (i.e., for the
  * NOCOW case)
@@ -454,27 +459,14 @@ struct scrub_ctx *scrub_setup_ctx(struct btrfs_device *dev, int is_dev_replace)
 	struct scrub_ctx *sctx;
 	int		i;
 	struct btrfs_fs_info *fs_info = dev->dev_root->fs_info;
-	int pages_per_rd_bio;
 	int ret;
 
-	/*
-	 * the setting of pages_per_rd_bio is correct for scrub but might
-	 * be wrong for the dev_replace code where we might read from
-	 * different devices in the initial huge bios. However, that
-	 * code is able to correctly handle the case when adding a page
-	 * to a bio fails.
-	 */
-	if (dev->bdev)
-		pages_per_rd_bio = min_t(int, SCRUB_PAGES_PER_RD_BIO,
-					 bio_get_nr_vecs(dev->bdev));
-	else
-		pages_per_rd_bio = SCRUB_PAGES_PER_RD_BIO;
 	sctx = kzalloc(sizeof(*sctx), GFP_NOFS);
 	if (!sctx)
 		goto nomem;
 	atomic_set(&sctx->refs, 1);
 	sctx->is_dev_replace = is_dev_replace;
-	sctx->pages_per_rd_bio = pages_per_rd_bio;
+	sctx->pages_per_rd_bio = SCRUB_PAGES_PER_RD_BIO;
 	sctx->curr = -1;
 	sctx->dev_root = dev->dev_root;
 	for (i = 0; i < SCRUB_BIOS_PER_SCTX; ++i) {
@@ -583,9 +575,9 @@ static int scrub_print_warning_inode(u64 inum, u64 offset, u64 root,
 	 * hold all of the paths here
 	 */
 	for (i = 0; i < ipath->fspath->elem_cnt; ++i)
-		printk_in_rcu(KERN_WARNING "BTRFS: %s at logical %llu on dev "
+		btrfs_warn_in_rcu(fs_info, "%s at logical %llu on dev "
 			"%s, sector %llu, root %llu, inode %llu, offset %llu, "
-			"length %llu, links %u (path: %s)\n", swarn->errstr,
+			"length %llu, links %u (path: %s)", swarn->errstr,
 			swarn->logical, rcu_str_deref(swarn->dev->name),
 			(unsigned long long)swarn->sector, root, inum, offset,
 			min(isize - offset, (u64)PAGE_SIZE), nlink,
@@ -595,9 +587,9 @@ static int scrub_print_warning_inode(u64 inum, u64 offset, u64 root,
 	return 0;
 
 err:
-	printk_in_rcu(KERN_WARNING "BTRFS: %s at logical %llu on dev "
+	btrfs_warn_in_rcu(fs_info, "%s at logical %llu on dev "
 		"%s, sector %llu, root %llu, inode %llu, offset %llu: path "
-		"resolving failed with ret=%d\n", swarn->errstr,
+		"resolving failed with ret=%d", swarn->errstr,
 		swarn->logical, rcu_str_deref(swarn->dev->name),
 		(unsigned long long)swarn->sector, root, inum, offset, ret);
 
@@ -652,10 +644,10 @@ static void scrub_print_warning(const char *errstr, struct scrub_block *sblock)
 			ret = tree_backref_for_extent(&ptr, eb, &found_key, ei,
 						      item_size, &ref_root,
 						      &ref_level);
-			printk_in_rcu(KERN_WARNING
-				"BTRFS: %s at logical %llu on dev %s, "
+			btrfs_warn_in_rcu(fs_info,
+				"%s at logical %llu on dev %s, "
 				"sector %llu: metadata %s (level %d) in tree "
-				"%llu\n", errstr, swarn.logical,
+				"%llu", errstr, swarn.logical,
 				rcu_str_deref(dev->name),
 				(unsigned long long)swarn.sector,
 				ref_level ? "node" : "leaf",
@@ -853,8 +845,8 @@ out:
 		btrfs_dev_replace_stats_inc(
 			&sctx->dev_root->fs_info->dev_replace.
 			num_uncorrectable_read_errors);
-		printk_ratelimited_in_rcu(KERN_ERR "BTRFS: "
-		    "unable to fixup (nodatasum) error at logical %llu on dev %s\n",
+		btrfs_err_rl_in_rcu(sctx->dev_root->fs_info,
+		    "unable to fixup (nodatasum) error at logical %llu on dev %s",
 			fixup->logical, rcu_str_deref(fixup->dev->name));
 	}
 
@@ -892,11 +884,9 @@ static int scrub_handle_errored_block(struct scrub_block *sblock_to_check)
 	struct btrfs_fs_info *fs_info;
 	u64 length;
 	u64 logical;
-	u64 generation;
 	unsigned int failed_mirror_index;
 	unsigned int is_metadata;
 	unsigned int have_csum;
-	u8 *csum;
 	struct scrub_block *sblocks_for_recheck; /* holds one for each mirror */
 	struct scrub_block *sblock_bad;
 	int ret;
@@ -921,13 +911,11 @@ static int scrub_handle_errored_block(struct scrub_block *sblock_to_check)
 	}
 	length = sblock_to_check->page_count * PAGE_SIZE;
 	logical = sblock_to_check->pagev[0]->logical;
-	generation = sblock_to_check->pagev[0]->generation;
 	BUG_ON(sblock_to_check->pagev[0]->mirror_num < 1);
 	failed_mirror_index = sblock_to_check->pagev[0]->mirror_num - 1;
 	is_metadata = !(sblock_to_check->pagev[0]->flags &
 			BTRFS_EXTENT_FLAG_DATA);
 	have_csum = sblock_to_check->pagev[0]->have_csum;
-	csum = sblock_to_check->pagev[0]->csum;
 	dev = sblock_to_check->pagev[0]->dev;
 
 	if (sctx->is_dev_replace && !is_metadata && !have_csum) {
@@ -990,8 +978,7 @@ static int scrub_handle_errored_block(struct scrub_block *sblock_to_check)
 	sblock_bad = sblocks_for_recheck + failed_mirror_index;
 
 	/* build and submit the bios for the failed mirror, check checksums */
-	scrub_recheck_block(fs_info, sblock_bad, is_metadata, have_csum,
-			    csum, generation, sctx->csum_size, 1);
+	scrub_recheck_block(fs_info, sblock_bad, 1);
 
 	if (!sblock_bad->header_error && !sblock_bad->checksum_error &&
 	    sblock_bad->no_io_error_seen) {
@@ -1104,9 +1091,7 @@ nodatasum_case:
 		sblock_other = sblocks_for_recheck + mirror_index;
 
 		/* build and submit the bios, check checksums */
-		scrub_recheck_block(fs_info, sblock_other, is_metadata,
-				    have_csum, csum, generation,
-				    sctx->csum_size, 0);
+		scrub_recheck_block(fs_info, sblock_other, 0);
 
 		if (!sblock_other->header_error &&
 		    !sblock_other->checksum_error &&
@@ -1218,9 +1203,7 @@ nodatasum_case:
 			 * is verified, but most likely the data comes out
 			 * of the page cache.
 			 */
-			scrub_recheck_block(fs_info, sblock_bad,
-					    is_metadata, have_csum, csum,
-					    generation, sctx->csum_size, 1);
+			scrub_recheck_block(fs_info, sblock_bad, 1);
 			if (!sblock_bad->header_error &&
 			    !sblock_bad->checksum_error &&
 			    sblock_bad->no_io_error_seen)
@@ -1233,8 +1216,8 @@ corrected_error:
 			sctx->stat.corrected_errors++;
 			sblock_to_check->data_corrected = 1;
 			spin_unlock(&sctx->stat_lock);
-			printk_ratelimited_in_rcu(KERN_ERR
-				"BTRFS: fixed up error at logical %llu on dev %s\n",
+			btrfs_err_rl_in_rcu(fs_info,
+				"fixed up error at logical %llu on dev %s",
 				logical, rcu_str_deref(dev->name));
 		}
 	} else {
@@ -1242,8 +1225,8 @@ did_not_correct_error:
 		spin_lock(&sctx->stat_lock);
 		sctx->stat.uncorrectable_errors++;
 		spin_unlock(&sctx->stat_lock);
-		printk_ratelimited_in_rcu(KERN_ERR
-			"BTRFS: unable to fixup (regular) error at logical %llu on dev %s\n",
+		btrfs_err_rl_in_rcu(fs_info,
+			"unable to fixup (regular) error at logical %llu on dev %s",
 			logical, rcu_str_deref(dev->name));
 	}
 
@@ -1321,6 +1304,9 @@ static int scrub_setup_recheck_block(struct scrub_block *original_sblock,
 	struct btrfs_fs_info *fs_info = sctx->dev_root->fs_info;
 	u64 length = original_sblock->page_count * PAGE_SIZE;
 	u64 logical = original_sblock->pagev[0]->logical;
+	u64 generation = original_sblock->pagev[0]->generation;
+	u64 flags = original_sblock->pagev[0]->flags;
+	u64 have_csum = original_sblock->pagev[0]->have_csum;
 	struct scrub_recover *recover;
 	struct btrfs_bio *bbio;
 	u64 sublen;
@@ -1375,6 +1361,7 @@ static int scrub_setup_recheck_block(struct scrub_block *original_sblock,
 
 			sblock = sblocks_for_recheck + mirror_index;
 			sblock->sctx = sctx;
+
 			page = kzalloc(sizeof(*page), GFP_NOFS);
 			if (!page) {
 leave_nomem:
@@ -1386,7 +1373,15 @@ leave_nomem:
 			}
 			scrub_page_get(page);
 			sblock->pagev[page_index] = page;
+			page->sblock = sblock;
+			page->flags = flags;
+			page->generation = generation;
 			page->logical = logical;
+			page->have_csum = have_csum;
+			if (have_csum)
+				memcpy(page->csum,
+				       original_sblock->pagev[0]->csum,
+				       sctx->csum_size);
 
 			scrub_stripe_index_and_offset(logical,
 						      bbio->map_type,
@@ -1429,11 +1424,11 @@ struct scrub_bio_ret {
 	int error;
 };
 
-static void scrub_bio_wait_endio(struct bio *bio, int error)
+static void scrub_bio_wait_endio(struct bio *bio)
 {
 	struct scrub_bio_ret *ret = bio->bi_private;
 
-	ret->error = error;
+	ret->error = bio->bi_error;
 	complete(&ret->event);
 }
 
@@ -1477,15 +1472,12 @@ static int scrub_submit_raid56_bio_wait(struct btrfs_fs_info *fs_info,
  * the pages that are errored in the just handled mirror can be repaired.
  */
 static void scrub_recheck_block(struct btrfs_fs_info *fs_info,
-				struct scrub_block *sblock, int is_metadata,
-				int have_csum, u8 *csum, u64 generation,
-				u16 csum_size, int retry_failed_mirror)
+				struct scrub_block *sblock,
+				int retry_failed_mirror)
 {
 	int page_num;
 
 	sblock->no_io_error_seen = 1;
-	sblock->header_error = 0;
-	sblock->checksum_error = 0;
 
 	for (page_num = 0; page_num < sblock->page_count; page_num++) {
 		struct bio *bio;
@@ -1521,9 +1513,7 @@ static void scrub_recheck_block(struct btrfs_fs_info *fs_info,
 	}
 
 	if (sblock->no_io_error_seen)
-		scrub_recheck_block_checksum(fs_info, sblock, is_metadata,
-					     have_csum, csum, generation,
-					     csum_size);
+		scrub_recheck_block_checksum(sblock);
 
 	return;
 }
@@ -1538,61 +1528,16 @@ static inline int scrub_check_fsid(u8 fsid[],
 	return !ret;
 }
 
-static void scrub_recheck_block_checksum(struct btrfs_fs_info *fs_info,
-					 struct scrub_block *sblock,
-					 int is_metadata, int have_csum,
-					 const u8 *csum, u64 generation,
-					 u16 csum_size)
+static void scrub_recheck_block_checksum(struct scrub_block *sblock)
 {
-	int page_num;
-	u8 calculated_csum[BTRFS_CSUM_SIZE];
-	u32 crc = ~(u32)0;
-	void *mapped_buffer;
-
-	WARN_ON(!sblock->pagev[0]->page);
-	if (is_metadata) {
-		struct btrfs_header *h;
-
-		mapped_buffer = kmap_atomic(sblock->pagev[0]->page);
-		h = (struct btrfs_header *)mapped_buffer;
-
-		if (sblock->pagev[0]->logical != btrfs_stack_header_bytenr(h) ||
-		    !scrub_check_fsid(h->fsid, sblock->pagev[0]) ||
-		    memcmp(h->chunk_tree_uuid, fs_info->chunk_tree_uuid,
-			   BTRFS_UUID_SIZE)) {
-			sblock->header_error = 1;
-		} else if (generation != btrfs_stack_header_generation(h)) {
-			sblock->header_error = 1;
-			sblock->generation_error = 1;
-		}
-		csum = h->csum;
-	} else {
-		if (!have_csum)
-			return;
-
-		mapped_buffer = kmap_atomic(sblock->pagev[0]->page);
-	}
-
-	for (page_num = 0;;) {
-		if (page_num == 0 && is_metadata)
-			crc = btrfs_csum_data(
-				((u8 *)mapped_buffer) + BTRFS_CSUM_SIZE,
-				crc, PAGE_SIZE - BTRFS_CSUM_SIZE);
-		else
-			crc = btrfs_csum_data(mapped_buffer, crc, PAGE_SIZE);
-
-		kunmap_atomic(mapped_buffer);
-		page_num++;
-		if (page_num >= sblock->page_count)
-			break;
-		WARN_ON(!sblock->pagev[page_num]->page);
-
-		mapped_buffer = kmap_atomic(sblock->pagev[page_num]->page);
-	}
+	sblock->header_error = 0;
+	sblock->checksum_error = 0;
+	sblock->generation_error = 0;
 
-	btrfs_csum_final(crc, calculated_csum);
-	if (memcmp(calculated_csum, csum, csum_size))
-		sblock->checksum_error = 1;
+	if (sblock->pagev[0]->flags & BTRFS_EXTENT_FLAG_DATA)
+		scrub_checksum_data(sblock);
+	else
+		scrub_checksum_tree_block(sblock);
 }
 
 static int scrub_repair_block_from_good_copy(struct scrub_block *sblock_bad,
@@ -1629,9 +1574,9 @@ static int scrub_repair_page_from_good_copy(struct scrub_block *sblock_bad,
 		int ret;
 
 		if (!page_bad->dev->bdev) {
-			printk_ratelimited(KERN_WARNING "BTRFS: "
+			btrfs_warn_rl(sblock_bad->sctx->dev_root->fs_info,
 				"scrub_repair_page_from_good_copy(bdev == NULL) "
-				"is unexpected!\n");
+				"is unexpected");
 			return -EIO;
 		}
 
@@ -1790,12 +1735,12 @@ static void scrub_wr_submit(struct scrub_ctx *sctx)
 	btrfsic_submit_bio(WRITE, sbio->bio);
 }
 
-static void scrub_wr_bio_end_io(struct bio *bio, int err)
+static void scrub_wr_bio_end_io(struct bio *bio)
 {
 	struct scrub_bio *sbio = bio->bi_private;
 	struct btrfs_fs_info *fs_info = sbio->dev->dev_root->fs_info;
 
-	sbio->err = err;
+	sbio->err = bio->bi_error;
 	sbio->bio = bio;
 
 	btrfs_init_work(&sbio->work, btrfs_scrubwrc_helper,
@@ -1836,6 +1781,18 @@ static int scrub_checksum(struct scrub_block *sblock)
 	u64 flags;
 	int ret;
 
+	/*
+	 * No need to initialize these stats currently,
+	 * because this function only use return value
+	 * instead of these stats value.
+	 *
+	 * Todo:
+	 * always use stats
+	 */
+	sblock->header_error = 0;
+	sblock->generation_error = 0;
+	sblock->checksum_error = 0;
+
 	WARN_ON(sblock->page_count < 1);
 	flags = sblock->pagev[0]->flags;
 	ret = 0;
@@ -1861,7 +1818,6 @@ static int scrub_checksum_data(struct scrub_block *sblock)
 	struct page *page;
 	void *buffer;
 	u32 crc = ~(u32)0;
-	int fail = 0;
 	u64 len;
 	int index;
 
@@ -1892,9 +1848,9 @@ static int scrub_checksum_data(struct scrub_block *sblock)
 
 	btrfs_csum_final(crc, csum);
 	if (memcmp(csum, on_disk_csum, sctx->csum_size))
-		fail = 1;
+		sblock->checksum_error = 1;
 
-	return fail;
+	return sblock->checksum_error;
 }
 
 static int scrub_checksum_tree_block(struct scrub_block *sblock)
@@ -1910,8 +1866,6 @@ static int scrub_checksum_tree_block(struct scrub_block *sblock)
 	u64 mapped_size;
 	void *p;
 	u32 crc = ~(u32)0;
-	int fail = 0;
-	int crc_fail = 0;
 	u64 len;
 	int index;
 
@@ -1926,19 +1880,20 @@ static int scrub_checksum_tree_block(struct scrub_block *sblock)
 	 * a) don't have an extent buffer and
 	 * b) the page is already kmapped
 	 */
-
 	if (sblock->pagev[0]->logical != btrfs_stack_header_bytenr(h))
-		++fail;
+		sblock->header_error = 1;
 
-	if (sblock->pagev[0]->generation != btrfs_stack_header_generation(h))
-		++fail;
+	if (sblock->pagev[0]->generation != btrfs_stack_header_generation(h)) {
+		sblock->header_error = 1;
+		sblock->generation_error = 1;
+	}
 
 	if (!scrub_check_fsid(h->fsid, sblock->pagev[0]))
-		++fail;
+		sblock->header_error = 1;
 
 	if (memcmp(h->chunk_tree_uuid, fs_info->chunk_tree_uuid,
 		   BTRFS_UUID_SIZE))
-		++fail;
+		sblock->header_error = 1;
 
 	len = sctx->nodesize - BTRFS_CSUM_SIZE;
 	mapped_size = PAGE_SIZE - BTRFS_CSUM_SIZE;
@@ -1963,9 +1918,9 @@ static int scrub_checksum_tree_block(struct scrub_block *sblock)
 
 	btrfs_csum_final(crc, calculated_csum);
 	if (memcmp(calculated_csum, on_disk_csum, sctx->csum_size))
-		++crc_fail;
+		sblock->checksum_error = 1;
 
-	return fail || crc_fail;
+	return sblock->header_error || sblock->checksum_error;
 }
 
 static int scrub_checksum_super(struct scrub_block *sblock)
@@ -2087,21 +2042,7 @@ static void scrub_submit(struct scrub_ctx *sctx)
 	sbio = sctx->bios[sctx->curr];
 	sctx->curr = -1;
 	scrub_pending_bio_inc(sctx);
-
-	if (!sbio->bio->bi_bdev) {
-		/*
-		 * this case should not happen. If btrfs_map_block() is
-		 * wrong, it could happen for dev-replace operations on
-		 * missing devices when no mirrors are available, but in
-		 * this case it should already fail the mount.
-		 * This case is handled correctly (but _very_ slowly).
-		 */
-		printk_ratelimited(KERN_WARNING
-			"BTRFS: scrub_submit(bio bdev == NULL) is unexpected!\n");
-		bio_endio(sbio->bio, -EIO);
-	} else {
-		btrfsic_submit_bio(READ, sbio->bio);
-	}
+	btrfsic_submit_bio(READ, sbio->bio);
 }
 
 static int scrub_add_page_to_rd_bio(struct scrub_ctx *sctx,
@@ -2178,6 +2119,122 @@ again:
 	return 0;
 }
 
+static void scrub_missing_raid56_end_io(struct bio *bio)
+{
+	struct scrub_block *sblock = bio->bi_private;
+	struct btrfs_fs_info *fs_info = sblock->sctx->dev_root->fs_info;
+
+	if (bio->bi_error)
+		sblock->no_io_error_seen = 0;
+
+	btrfs_queue_work(fs_info->scrub_workers, &sblock->work);
+}
+
+static void scrub_missing_raid56_worker(struct btrfs_work *work)
+{
+	struct scrub_block *sblock = container_of(work, struct scrub_block, work);
+	struct scrub_ctx *sctx = sblock->sctx;
+	u64 logical;
+	struct btrfs_device *dev;
+
+	logical = sblock->pagev[0]->logical;
+	dev = sblock->pagev[0]->dev;
+
+	if (sblock->no_io_error_seen)
+		scrub_recheck_block_checksum(sblock);
+
+	if (!sblock->no_io_error_seen) {
+		spin_lock(&sctx->stat_lock);
+		sctx->stat.read_errors++;
+		spin_unlock(&sctx->stat_lock);
+		btrfs_err_rl_in_rcu(sctx->dev_root->fs_info,
+			"IO error rebuilding logical %llu for dev %s",
+			logical, rcu_str_deref(dev->name));
+	} else if (sblock->header_error || sblock->checksum_error) {
+		spin_lock(&sctx->stat_lock);
+		sctx->stat.uncorrectable_errors++;
+		spin_unlock(&sctx->stat_lock);
+		btrfs_err_rl_in_rcu(sctx->dev_root->fs_info,
+			"failed to rebuild valid logical %llu for dev %s",
+			logical, rcu_str_deref(dev->name));
+	} else {
+		scrub_write_block_to_dev_replace(sblock);
+	}
+
+	scrub_block_put(sblock);
+
+	if (sctx->is_dev_replace &&
+	    atomic_read(&sctx->wr_ctx.flush_all_writes)) {
+		mutex_lock(&sctx->wr_ctx.wr_lock);
+		scrub_wr_submit(sctx);
+		mutex_unlock(&sctx->wr_ctx.wr_lock);
+	}
+
+	scrub_pending_bio_dec(sctx);
+}
+
+static void scrub_missing_raid56_pages(struct scrub_block *sblock)
+{
+	struct scrub_ctx *sctx = sblock->sctx;
+	struct btrfs_fs_info *fs_info = sctx->dev_root->fs_info;
+	u64 length = sblock->page_count * PAGE_SIZE;
+	u64 logical = sblock->pagev[0]->logical;
+	struct btrfs_bio *bbio;
+	struct bio *bio;
+	struct btrfs_raid_bio *rbio;
+	int ret;
+	int i;
+
+	ret = btrfs_map_sblock(fs_info, REQ_GET_READ_MIRRORS, logical, &length,
+			       &bbio, 0, 1);
+	if (ret || !bbio || !bbio->raid_map)
+		goto bbio_out;
+
+	if (WARN_ON(!sctx->is_dev_replace ||
+		    !(bbio->map_type & BTRFS_BLOCK_GROUP_RAID56_MASK))) {
+		/*
+		 * We shouldn't be scrubbing a missing device. Even for dev
+		 * replace, we should only get here for RAID 5/6. We either
+		 * managed to mount something with no mirrors remaining or
+		 * there's a bug in scrub_remap_extent()/btrfs_map_block().
+		 */
+		goto bbio_out;
+	}
+
+	bio = btrfs_io_bio_alloc(GFP_NOFS, 0);
+	if (!bio)
+		goto bbio_out;
+
+	bio->bi_iter.bi_sector = logical >> 9;
+	bio->bi_private = sblock;
+	bio->bi_end_io = scrub_missing_raid56_end_io;
+
+	rbio = raid56_alloc_missing_rbio(sctx->dev_root, bio, bbio, length);
+	if (!rbio)
+		goto rbio_out;
+
+	for (i = 0; i < sblock->page_count; i++) {
+		struct scrub_page *spage = sblock->pagev[i];
+
+		raid56_add_scrub_pages(rbio, spage->page, spage->logical);
+	}
+
+	btrfs_init_work(&sblock->work, btrfs_scrub_helper,
+			scrub_missing_raid56_worker, NULL, NULL);
+	scrub_block_get(sblock);
+	scrub_pending_bio_inc(sctx);
+	raid56_submit_missing_rbio(rbio);
+	return;
+
+rbio_out:
+	bio_put(bio);
+bbio_out:
+	btrfs_put_bbio(bbio);
+	spin_lock(&sctx->stat_lock);
+	sctx->stat.malloc_errors++;
+	spin_unlock(&sctx->stat_lock);
+}
+
 static int scrub_pages(struct scrub_ctx *sctx, u64 logical, u64 len,
 		       u64 physical, struct btrfs_device *dev, u64 flags,
 		       u64 gen, int mirror_num, u8 *csum, int force,
@@ -2241,31 +2298,39 @@ leave_nomem:
 	}
 
 	WARN_ON(sblock->page_count == 0);
-	for (index = 0; index < sblock->page_count; index++) {
-		struct scrub_page *spage = sblock->pagev[index];
-		int ret;
+	if (dev->missing) {
+		/*
+		 * This case should only be hit for RAID 5/6 device replace. See
+		 * the comment in scrub_missing_raid56_pages() for details.
+		 */
+		scrub_missing_raid56_pages(sblock);
+	} else {
+		for (index = 0; index < sblock->page_count; index++) {
+			struct scrub_page *spage = sblock->pagev[index];
+			int ret;
 
-		ret = scrub_add_page_to_rd_bio(sctx, spage);
-		if (ret) {
-			scrub_block_put(sblock);
-			return ret;
+			ret = scrub_add_page_to_rd_bio(sctx, spage);
+			if (ret) {
+				scrub_block_put(sblock);
+				return ret;
+			}
 		}
-	}
 
-	if (force)
-		scrub_submit(sctx);
+		if (force)
+			scrub_submit(sctx);
+	}
 
 	/* last one frees, either here or in bio completion for last page */
 	scrub_block_put(sblock);
 	return 0;
 }
 
-static void scrub_bio_end_io(struct bio *bio, int err)
+static void scrub_bio_end_io(struct bio *bio)
 {
 	struct scrub_bio *sbio = bio->bi_private;
 	struct btrfs_fs_info *fs_info = sbio->dev->dev_root->fs_info;
 
-	sbio->err = err;
+	sbio->err = bio->bi_error;
 	sbio->bio = bio;
 
 	btrfs_queue_work(fs_info->scrub_workers, &sbio->work);
@@ -2381,8 +2446,7 @@ static void scrub_block_complete(struct scrub_block *sblock)
 	}
 }
 
-static int scrub_find_csum(struct scrub_ctx *sctx, u64 logical, u64 len,
-			   u8 *csum)
+static int scrub_find_csum(struct scrub_ctx *sctx, u64 logical, u8 *csum)
 {
 	struct btrfs_ordered_sum *sum = NULL;
 	unsigned long index;
@@ -2446,7 +2510,7 @@ static int scrub_extent(struct scrub_ctx *sctx, u64 logical, u64 len,
 
 		if (flags & BTRFS_EXTENT_FLAG_DATA) {
 			/* push csums to sbio */
-			have_csum = scrub_find_csum(sctx, logical, l, csum);
+			have_csum = scrub_find_csum(sctx, logical, csum);
 			if (have_csum == 0)
 				++sctx->stat.no_csum;
 			if (sctx->is_dev_replace && !have_csum) {
@@ -2564,6 +2628,11 @@ static int scrub_extent_for_parity(struct scrub_parity *sparity,
 	u8 csum[BTRFS_CSUM_SIZE];
 	u32 blocksize;
 
+	if (dev->missing) {
+		scrub_parity_mark_sectors_error(sparity, logical, len);
+		return 0;
+	}
+
 	if (flags & BTRFS_EXTENT_FLAG_DATA) {
 		blocksize = sctx->sectorsize;
 	} else if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) {
@@ -2579,7 +2648,7 @@ static int scrub_extent_for_parity(struct scrub_parity *sparity,
 
 		if (flags & BTRFS_EXTENT_FLAG_DATA) {
 			/* push csums to sbio */
-			have_csum = scrub_find_csum(sctx, logical, l, csum);
+			have_csum = scrub_find_csum(sctx, logical, csum);
 			if (have_csum == 0)
 				goto skip;
 		}
@@ -2662,18 +2731,30 @@ static void scrub_free_parity(struct scrub_parity *sparity)
 	kfree(sparity);
 }
 
-static void scrub_parity_bio_endio(struct bio *bio, int error)
+static void scrub_parity_bio_endio_worker(struct btrfs_work *work)
 {
-	struct scrub_parity *sparity = (struct scrub_parity *)bio->bi_private;
+	struct scrub_parity *sparity = container_of(work, struct scrub_parity,
+						    work);
 	struct scrub_ctx *sctx = sparity->sctx;
 
-	if (error)
+	scrub_free_parity(sparity);
+	scrub_pending_bio_dec(sctx);
+}
+
+static void scrub_parity_bio_endio(struct bio *bio)
+{
+	struct scrub_parity *sparity = (struct scrub_parity *)bio->bi_private;
+
+	if (bio->bi_error)
 		bitmap_or(sparity->ebitmap, sparity->ebitmap, sparity->dbitmap,
 			  sparity->nsectors);
 
-	scrub_free_parity(sparity);
-	scrub_pending_bio_dec(sctx);
 	bio_put(bio);
+
+	btrfs_init_work(&sparity->work, btrfs_scrubparity_helper,
+			scrub_parity_bio_endio_worker, NULL, NULL);
+	btrfs_queue_work(sparity->sctx->dev_root->fs_info->scrub_parity_workers,
+			 &sparity->work);
 }
 
 static void scrub_parity_check_and_repair(struct scrub_parity *sparity)
@@ -2690,7 +2771,7 @@ static void scrub_parity_check_and_repair(struct scrub_parity *sparity)
 			   sparity->nsectors))
 		goto out;
 
-	length = sparity->logic_end - sparity->logic_start + 1;
+	length = sparity->logic_end - sparity->logic_start;
 	ret = btrfs_map_sblock(sctx->dev_root->fs_info, WRITE,
 			       sparity->logic_start,
 			       &length, &bbio, 0, 1);
@@ -2713,8 +2794,7 @@ static void scrub_parity_check_and_repair(struct scrub_parity *sparity)
 		goto rbio_out;
 
 	list_for_each_entry(spage, &sparity->spages, list)
-		raid56_parity_add_scrub_pages(rbio, spage->page,
-					      spage->logical);
+		raid56_add_scrub_pages(rbio, spage->page, spage->logical);
 
 	scrub_pending_bio_inc(sctx);
 	raid56_parity_submit_scrub_rbio(rbio);
@@ -2762,6 +2842,7 @@ static noinline_for_stack int scrub_raid56_parity(struct scrub_ctx *sctx,
 	struct btrfs_root *root = fs_info->extent_root;
 	struct btrfs_root *csum_root = fs_info->csum_root;
 	struct btrfs_extent_item *extent;
+	struct btrfs_bio *bbio = NULL;
 	u64 flags;
 	int ret;
 	int slot;
@@ -2771,6 +2852,7 @@ static noinline_for_stack int scrub_raid56_parity(struct scrub_ctx *sctx,
 	u64 extent_logical;
 	u64 extent_physical;
 	u64 extent_len;
+	u64 mapped_length;
 	struct btrfs_device *extent_dev;
 	struct scrub_parity *sparity;
 	int nsectors;
@@ -2844,6 +2926,10 @@ static noinline_for_stack int scrub_raid56_parity(struct scrub_ctx *sctx,
 			}
 			btrfs_item_key_to_cpu(l, &key, slot);
 
+			if (key.type != BTRFS_EXTENT_ITEM_KEY &&
+			    key.type != BTRFS_METADATA_ITEM_KEY)
+				goto next;
+
 			if (key.type == BTRFS_METADATA_ITEM_KEY)
 				bytes = root->nodesize;
 			else
@@ -2852,11 +2938,7 @@ static noinline_for_stack int scrub_raid56_parity(struct scrub_ctx *sctx,
 			if (key.objectid + bytes <= logic_start)
 				goto next;
 
-			if (key.type != BTRFS_EXTENT_ITEM_KEY &&
-			    key.type != BTRFS_METADATA_ITEM_KEY)
-				goto next;
-
-			if (key.objectid > logic_end) {
+			if (key.objectid >= logic_end) {
 				stop_loop = 1;
 				break;
 			}
@@ -2869,11 +2951,15 @@ static noinline_for_stack int scrub_raid56_parity(struct scrub_ctx *sctx,
 			flags = btrfs_extent_flags(l, extent);
 			generation = btrfs_extent_generation(l, extent);
 
-			if (key.objectid < logic_start &&
-			    (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK)) {
-				btrfs_err(fs_info,
-					  "scrub: tree block %llu spanning stripes, ignored. logical=%llu",
-					   key.objectid, logic_start);
+			if ((flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) &&
+			    (key.objectid < logic_start ||
+			     key.objectid + bytes >
+			     logic_start + map->stripe_len)) {
+				btrfs_err(fs_info, "scrub: tree block %llu spanning stripes, ignored. logical=%llu",
+					  key.objectid, logic_start);
+				spin_lock(&sctx->stat_lock);
+				sctx->stat.uncorrectable_errors++;
+				spin_unlock(&sctx->stat_lock);
 				goto next;
 			}
 again:
@@ -2893,10 +2979,21 @@ again:
 			scrub_parity_mark_sectors_data(sparity, extent_logical,
 						       extent_len);
 
-			scrub_remap_extent(fs_info, extent_logical,
-					   extent_len, &extent_physical,
-					   &extent_dev,
-					   &extent_mirror_num);
+			mapped_length = extent_len;
+			ret = btrfs_map_block(fs_info, READ, extent_logical,
+					      &mapped_length, &bbio, 0);
+			if (!ret) {
+				if (!bbio || mapped_length < extent_len)
+					ret = -EIO;
+			}
+			if (ret) {
+				btrfs_put_bbio(bbio);
+				goto out;
+			}
+			extent_physical = bbio->stripes[0].physical;
+			extent_mirror_num = bbio->mirror_num;
+			extent_dev = bbio->stripes[0].dev;
+			btrfs_put_bbio(bbio);
 
 			ret = btrfs_lookup_csums_range(csum_root,
 						extent_logical,
@@ -2911,10 +3008,12 @@ again:
 						      extent_dev, flags,
 						      generation,
 						      extent_mirror_num);
+
+			scrub_free_csums(sctx);
+
 			if (ret)
 				goto out;
 
-			scrub_free_csums(sctx);
 			if (extent_logical + extent_len <
 			    key.objectid + bytes) {
 				logic_start += map->stripe_len;
@@ -2943,7 +3042,7 @@ next:
 out:
 	if (ret < 0)
 		scrub_parity_mark_sectors_error(sparity, logic_start,
-						logic_end - logic_start + 1);
+						logic_end - logic_start);
 	scrub_parity_put(sparity);
 	scrub_submit(sctx);
 	mutex_lock(&sctx->wr_ctx.wr_lock);
@@ -3092,22 +3191,6 @@ static noinline_for_stack int scrub_stripe(struct scrub_ctx *sctx,
 	 */
 	ret = 0;
 	while (physical < physical_end) {
-		/* for raid56, we skip parity stripe */
-		if (map->type & BTRFS_BLOCK_GROUP_RAID56_MASK) {
-			ret = get_raid56_logic_offset(physical, num,
-					map, &logical, &stripe_logical);
-			logical += base;
-			if (ret) {
-				stripe_logical += base;
-				stripe_end = stripe_logical + increment - 1;
-				ret = scrub_raid56_parity(sctx, map, scrub_dev,
-						ppath, stripe_logical,
-						stripe_end);
-				if (ret)
-					goto out;
-				goto skip;
-			}
-		}
 		/*
 		 * canceled?
 		 */
@@ -3132,6 +3215,24 @@ static noinline_for_stack int scrub_stripe(struct scrub_ctx *sctx,
 			scrub_blocked_if_needed(fs_info);
 		}
 
+		if (map->type & BTRFS_BLOCK_GROUP_RAID56_MASK) {
+			ret = get_raid56_logic_offset(physical, num, map,
+						      &logical,
+						      &stripe_logical);
+			logical += base;
+			if (ret) {
+				/* it is parity strip */
+				stripe_logical += base;
+				stripe_end = stripe_logical + increment;
+				ret = scrub_raid56_parity(sctx, map, scrub_dev,
+							  ppath, stripe_logical,
+							  stripe_end);
+				if (ret)
+					goto out;
+				goto skip;
+			}
+		}
+
 		if (btrfs_fs_incompat(fs_info, SKINNY_METADATA))
 			key.type = BTRFS_METADATA_ITEM_KEY;
 		else
@@ -3176,6 +3277,10 @@ static noinline_for_stack int scrub_stripe(struct scrub_ctx *sctx,
 			}
 			btrfs_item_key_to_cpu(l, &key, slot);
 
+			if (key.type != BTRFS_EXTENT_ITEM_KEY &&
+			    key.type != BTRFS_METADATA_ITEM_KEY)
+				goto next;
+
 			if (key.type == BTRFS_METADATA_ITEM_KEY)
 				bytes = root->nodesize;
 			else
@@ -3184,10 +3289,6 @@ static noinline_for_stack int scrub_stripe(struct scrub_ctx *sctx,
 			if (key.objectid + bytes <= logical)
 				goto next;
 
-			if (key.type != BTRFS_EXTENT_ITEM_KEY &&
-			    key.type != BTRFS_METADATA_ITEM_KEY)
-				goto next;
-
 			if (key.objectid >= logical + map->stripe_len) {
 				/* out of this device extent */
 				if (key.objectid >= logic_end)
@@ -3200,12 +3301,17 @@ static noinline_for_stack int scrub_stripe(struct scrub_ctx *sctx,
 			flags = btrfs_extent_flags(l, extent);
 			generation = btrfs_extent_generation(l, extent);
 
-			if (key.objectid < logical &&
-			    (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK)) {
+			if ((flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) &&
+			    (key.objectid < logical ||
+			     key.objectid + bytes >
+			     logical + map->stripe_len)) {
 				btrfs_err(fs_info,
 					   "scrub: tree block %llu spanning "
 					   "stripes, ignored. logical=%llu",
 				       key.objectid, logical);
+				spin_lock(&sctx->stat_lock);
+				sctx->stat.uncorrectable_errors++;
+				spin_unlock(&sctx->stat_lock);
 				goto next;
 			}
 
@@ -3235,9 +3341,11 @@ again:
 						   &extent_dev,
 						   &extent_mirror_num);
 
-			ret = btrfs_lookup_csums_range(csum_root, logical,
-						logical + map->stripe_len - 1,
-						&sctx->csum_list, 1);
+			ret = btrfs_lookup_csums_range(csum_root,
+						       extent_logical,
+						       extent_logical +
+						       extent_len - 1,
+						       &sctx->csum_list, 1);
 			if (ret)
 				goto out;
 
@@ -3245,10 +3353,12 @@ again:
 					   extent_physical, extent_dev, flags,
 					   generation, extent_mirror_num,
 					   extent_logical - logical + physical);
+
+			scrub_free_csums(sctx);
+
 			if (ret)
 				goto out;
 
-			scrub_free_csums(sctx);
 			if (extent_logical + extent_len <
 			    key.objectid + bytes) {
 				if (map->type & BTRFS_BLOCK_GROUP_RAID56_MASK) {
@@ -3266,7 +3376,7 @@ loop:
 					if (ret && physical < physical_end) {
 						stripe_logical += base;
 						stripe_end = stripe_logical +
-								increment - 1;
+								increment;
 						ret = scrub_raid56_parity(sctx,
 							map, scrub_dev, ppath,
 							stripe_logical,
@@ -3321,9 +3431,10 @@ out:
 
 static noinline_for_stack int scrub_chunk(struct scrub_ctx *sctx,
 					  struct btrfs_device *scrub_dev,
-					  u64 chunk_tree, u64 chunk_objectid,
 					  u64 chunk_offset, u64 length,
-					  u64 dev_offset, int is_dev_replace)
+					  u64 dev_offset,
+					  struct btrfs_block_group_cache *cache,
+					  int is_dev_replace)
 {
 	struct btrfs_mapping_tree *map_tree =
 		&sctx->dev_root->fs_info->mapping_tree;
@@ -3336,8 +3447,18 @@ static noinline_for_stack int scrub_chunk(struct scrub_ctx *sctx,
 	em = lookup_extent_mapping(&map_tree->map_tree, chunk_offset, 1);
 	read_unlock(&map_tree->map_tree.lock);
 
-	if (!em)
-		return -EINVAL;
+	if (!em) {
+		/*
+		 * Might have been an unused block group deleted by the cleaner
+		 * kthread or relocation.
+		 */
+		spin_lock(&cache->lock);
+		if (!cache->removed)
+			ret = -EINVAL;
+		spin_unlock(&cache->lock);
+
+		return ret;
+	}
 
 	map = (struct map_lookup *)em->bdev;
 	if (em->start != chunk_offset)
@@ -3372,10 +3493,9 @@ int scrub_enumerate_chunks(struct scrub_ctx *sctx,
 	struct btrfs_root *root = sctx->dev_root;
 	struct btrfs_fs_info *fs_info = root->fs_info;
 	u64 length;
-	u64 chunk_tree;
-	u64 chunk_objectid;
 	u64 chunk_offset;
-	int ret;
+	int ret = 0;
+	int ro_set;
 	int slot;
 	struct extent_buffer *l;
 	struct btrfs_key key;
@@ -3403,8 +3523,14 @@ int scrub_enumerate_chunks(struct scrub_ctx *sctx,
 			if (path->slots[0] >=
 			    btrfs_header_nritems(path->nodes[0])) {
 				ret = btrfs_next_leaf(root, path);
-				if (ret)
+				if (ret < 0)
 					break;
+				if (ret > 0) {
+					ret = 0;
+					break;
+				}
+			} else {
+				ret = 0;
 			}
 		}
 
@@ -3431,8 +3557,6 @@ int scrub_enumerate_chunks(struct scrub_ctx *sctx,
 		if (found_key.offset + length <= start)
 			goto skip;
 
-		chunk_tree = btrfs_dev_extent_chunk_tree(l, dev_extent);
-		chunk_objectid = btrfs_dev_extent_chunk_objectid(l, dev_extent);
 		chunk_offset = btrfs_dev_extent_chunk_offset(l, dev_extent);
 
 		/*
@@ -3446,12 +3570,41 @@ int scrub_enumerate_chunks(struct scrub_ctx *sctx,
 		if (!cache)
 			goto skip;
 
+		/*
+		 * we need call btrfs_inc_block_group_ro() with scrubs_paused,
+		 * to avoid deadlock caused by:
+		 * btrfs_inc_block_group_ro()
+		 * -> btrfs_wait_for_commit()
+		 * -> btrfs_commit_transaction()
+		 * -> btrfs_scrub_pause()
+		 */
+		scrub_pause_on(fs_info);
+		ret = btrfs_inc_block_group_ro(root, cache);
+		scrub_pause_off(fs_info);
+
+		if (ret == 0) {
+			ro_set = 1;
+		} else if (ret == -ENOSPC) {
+			/*
+			 * btrfs_inc_block_group_ro return -ENOSPC when it
+			 * failed in creating new chunk for metadata.
+			 * It is not a problem for scrub/replace, because
+			 * metadata are always cowed, and our scrub paused
+			 * commit_transactions.
+			 */
+			ro_set = 0;
+		} else {
+			btrfs_warn(fs_info, "failed setting block group ro, ret=%d\n",
+				   ret);
+			btrfs_put_block_group(cache);
+			break;
+		}
+
 		dev_replace->cursor_right = found_key.offset + length;
 		dev_replace->cursor_left = found_key.offset;
 		dev_replace->item_needs_writeback = 1;
-		ret = scrub_chunk(sctx, scrub_dev, chunk_tree, chunk_objectid,
-				  chunk_offset, length, found_key.offset,
-				  is_dev_replace);
+		ret = scrub_chunk(sctx, scrub_dev, chunk_offset, length,
+				  found_key.offset, cache, is_dev_replace);
 
 		/*
 		 * flush, submit all pending read and write bios, afterwards
@@ -3471,8 +3624,8 @@ int scrub_enumerate_chunks(struct scrub_ctx *sctx,
 
 		wait_event(sctx->list_wait,
 			   atomic_read(&sctx->bios_in_flight) == 0);
-		atomic_inc(&fs_info->scrubs_paused);
-		wake_up(&fs_info->scrub_pause_wait);
+
+		scrub_pause_on(fs_info);
 
 		/*
 		 * must be called before we decrease @scrub_paused.
@@ -3483,11 +3636,32 @@ int scrub_enumerate_chunks(struct scrub_ctx *sctx,
 			   atomic_read(&sctx->workers_pending) == 0);
 		atomic_set(&sctx->wr_ctx.flush_all_writes, 0);
 
-		mutex_lock(&fs_info->scrub_lock);
-		__scrub_blocked_if_needed(fs_info);
-		atomic_dec(&fs_info->scrubs_paused);
-		mutex_unlock(&fs_info->scrub_lock);
-		wake_up(&fs_info->scrub_pause_wait);
+		scrub_pause_off(fs_info);
+
+		if (ro_set)
+			btrfs_dec_block_group_ro(root, cache);
+
+		/*
+		 * We might have prevented the cleaner kthread from deleting
+		 * this block group if it was already unused because we raced
+		 * and set it to RO mode first. So add it back to the unused
+		 * list, otherwise it might not ever be deleted unless a manual
+		 * balance is triggered or it becomes used and unused again.
+		 */
+		spin_lock(&cache->lock);
+		if (!cache->removed && !cache->ro && cache->reserved == 0 &&
+		    btrfs_block_group_used(&cache->item) == 0) {
+			spin_unlock(&cache->lock);
+			spin_lock(&fs_info->unused_bgs_lock);
+			if (list_empty(&cache->bg_list)) {
+				btrfs_get_block_group(cache);
+				list_add_tail(&cache->bg_list,
+					      &fs_info->unused_bgs);
+			}
+			spin_unlock(&fs_info->unused_bgs_lock);
+		} else {
+			spin_unlock(&cache->lock);
+		}
 
 		btrfs_put_block_group(cache);
 		if (ret)
@@ -3511,11 +3685,7 @@ skip:
 
 	btrfs_free_path(path);
 
-	/*
-	 * ret can still be 1 from search_slot or next_leaf,
-	 * that's not an error
-	 */
-	return ret < 0 ? ret : 0;
+	return ret;
 }
 
 static noinline_for_stack int scrub_supers(struct scrub_ctx *sctx,
@@ -3559,7 +3729,6 @@ static noinline_for_stack int scrub_supers(struct scrub_ctx *sctx,
 static noinline_for_stack int scrub_workers_get(struct btrfs_fs_info *fs_info,
 						int is_dev_replace)
 {
-	int ret = 0;
 	unsigned int flags = WQ_FREEZABLE | WQ_UNBOUND;
 	int max_active = fs_info->thread_pool_size;
 
@@ -3572,27 +3741,36 @@ static noinline_for_stack int scrub_workers_get(struct btrfs_fs_info *fs_info,
 			fs_info->scrub_workers =
 				btrfs_alloc_workqueue("btrfs-scrub", flags,
 						      max_active, 4);
-		if (!fs_info->scrub_workers) {
-			ret = -ENOMEM;
-			goto out;
-		}
+		if (!fs_info->scrub_workers)
+			goto fail_scrub_workers;
+
 		fs_info->scrub_wr_completion_workers =
 			btrfs_alloc_workqueue("btrfs-scrubwrc", flags,
 					      max_active, 2);
-		if (!fs_info->scrub_wr_completion_workers) {
-			ret = -ENOMEM;
-			goto out;
-		}
+		if (!fs_info->scrub_wr_completion_workers)
+			goto fail_scrub_wr_completion_workers;
+
 		fs_info->scrub_nocow_workers =
 			btrfs_alloc_workqueue("btrfs-scrubnc", flags, 1, 0);
-		if (!fs_info->scrub_nocow_workers) {
-			ret = -ENOMEM;
-			goto out;
-		}
+		if (!fs_info->scrub_nocow_workers)
+			goto fail_scrub_nocow_workers;
+		fs_info->scrub_parity_workers =
+			btrfs_alloc_workqueue("btrfs-scrubparity", flags,
+					      max_active, 2);
+		if (!fs_info->scrub_parity_workers)
+			goto fail_scrub_parity_workers;
 	}
 	++fs_info->scrub_workers_refcnt;
-out:
-	return ret;
+	return 0;
+
+fail_scrub_parity_workers:
+	btrfs_destroy_workqueue(fs_info->scrub_nocow_workers);
+fail_scrub_nocow_workers:
+	btrfs_destroy_workqueue(fs_info->scrub_wr_completion_workers);
+fail_scrub_wr_completion_workers:
+	btrfs_destroy_workqueue(fs_info->scrub_workers);
+fail_scrub_workers:
+	return -ENOMEM;
 }
 
 static noinline_for_stack void scrub_workers_put(struct btrfs_fs_info *fs_info)
@@ -3601,6 +3779,7 @@ static noinline_for_stack void scrub_workers_put(struct btrfs_fs_info *fs_info)
 		btrfs_destroy_workqueue(fs_info->scrub_workers);
 		btrfs_destroy_workqueue(fs_info->scrub_wr_completion_workers);
 		btrfs_destroy_workqueue(fs_info->scrub_nocow_workers);
+		btrfs_destroy_workqueue(fs_info->scrub_parity_workers);
 	}
 	WARN_ON(fs_info->scrub_workers_refcnt < 0);
 }
@@ -3875,8 +4054,7 @@ static int scrub_setup_wr_ctx(struct scrub_ctx *sctx,
 		return 0;
 
 	WARN_ON(!dev->bdev);
-	wr_ctx->pages_per_wr_bio = min_t(int, SCRUB_PAGES_PER_WR_BIO,
-					 bio_get_nr_vecs(dev->bdev));
+	wr_ctx->pages_per_wr_bio = SCRUB_PAGES_PER_WR_BIO;
 	wr_ctx->tgtdev = dev;
 	atomic_set(&wr_ctx->flush_all_writes, 0);
 	return 0;
@@ -4198,8 +4376,8 @@ static int write_page_nocow(struct scrub_ctx *sctx,
 	if (!dev)
 		return -EIO;
 	if (!dev->bdev) {
-		printk_ratelimited(KERN_WARNING
-			"BTRFS: scrub write_page_nocow(bdev == NULL) is unexpected!\n");
+		btrfs_warn_rl(dev->dev_root->fs_info,
+			"scrub write_page_nocow(bdev == NULL) is unexpected");
 		return -EIO;
 	}
 	bio = btrfs_io_bio_alloc(GFP_NOFS, 1);
diff --git a/kernel/fs/btrfs/send.c b/kernel/fs/btrfs/send.c
index a1216f9b4..63a6152be 100644
--- a/kernel/fs/btrfs/send.c
+++ b/kernel/fs/btrfs/send.c
@@ -243,6 +243,7 @@ struct waiting_dir_move {
 	 * after this directory is moved, we can try to rmdir the ino rmdir_ino.
 	 */
 	u64 rmdir_ino;
+	bool orphanized;
 };
 
 struct orphan_dir_info {
@@ -1158,6 +1159,9 @@ struct backref_ctx {
 	/* may be truncated in case it's the last extent in a file */
 	u64 extent_len;
 
+	/* data offset in the file extent item */
+	u64 data_offset;
+
 	/* Just to check for bugs in backref resolving */
 	int found_itself;
 };
@@ -1221,7 +1225,7 @@ static int __iterate_backrefs(u64 ino, u64 offset, u64 root, void *ctx_)
 	if (ret < 0)
 		return ret;
 
-	if (offset + bctx->extent_len > i_size)
+	if (offset + bctx->data_offset + bctx->extent_len > i_size)
 		return 0;
 
 	/*
@@ -1363,6 +1367,19 @@ static int find_extent_clone(struct send_ctx *sctx,
 	backref_ctx->cur_offset = data_offset;
 	backref_ctx->found_itself = 0;
 	backref_ctx->extent_len = num_bytes;
+	/*
+	 * For non-compressed extents iterate_extent_inodes() gives us extent
+	 * offsets that already take into account the data offset, but not for
+	 * compressed extents, since the offset is logical and not relative to
+	 * the physical extent locations. We must take this into account to
+	 * avoid sending clone offsets that go beyond the source file's size,
+	 * which would result in the clone ioctl failing with -EINVAL on the
+	 * receiving end.
+	 */
+	if (compressed == BTRFS_COMPRESS_NONE)
+		backref_ctx->data_offset = 0;
+	else
+		backref_ctx->data_offset = btrfs_file_extent_offset(eb, fi);
 
 	/*
 	 * The last extent of a file may be too large due to page alignment.
@@ -1417,16 +1434,6 @@ verbose_printk(KERN_DEBUG "btrfs: find_extent_clone: data_offset=%llu, "
 	}
 
 	if (cur_clone_root) {
-		if (compressed != BTRFS_COMPRESS_NONE) {
-			/*
-			 * Offsets given by iterate_extent_inodes() are relative
-			 * to the start of the extent, we need to add logical
-			 * offset from the file extent item.
-			 * (See why at backref.c:check_extent_in_eb())
-			 */
-			cur_clone_root->offset += btrfs_file_extent_offset(eb,
-									   fi);
-		}
 		*found = cur_clone_root;
 		ret = 0;
 	} else {
@@ -1462,7 +1469,21 @@ static int read_symlink(struct btrfs_root *root,
 	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
 	if (ret < 0)
 		goto out;
-	BUG_ON(ret);
+	if (ret) {
+		/*
+		 * An empty symlink inode. Can happen in rare error paths when
+		 * creating a symlink (transaction committed before the inode
+		 * eviction handler removed the symlink inode items and a crash
+		 * happened in between or the subvol was snapshoted in between).
+		 * Print an informative message to dmesg/syslog so that the user
+		 * can delete the symlink.
+		 */
+		btrfs_err(root->fs_info,
+			  "Found empty symlink inode %llu at root %llu",
+			  ino, root->root_key.objectid);
+		ret = -EIO;
+		goto out;
+	}
 
 	ei = btrfs_item_ptr(path->nodes[0], path->slots[0],
 			struct btrfs_file_extent_item);
@@ -1900,8 +1921,15 @@ static int did_overwrite_ref(struct send_ctx *sctx,
 		goto out;
 	}
 
-	/* we know that it is or will be overwritten. check this now */
-	if (ow_inode < sctx->send_progress)
+	/*
+	 * We know that it is or will be overwritten. Check this now.
+	 * The current inode being processed might have been the one that caused
+	 * inode 'ino' to be orphanized, therefore check if ow_inode matches
+	 * the current inode being processed.
+	 */
+	if ((ow_inode < sctx->send_progress) ||
+	    (ino != sctx->cur_ino && ow_inode == sctx->cur_ino &&
+	     gen == sctx->cur_inode_gen))
 		ret = 1;
 	else
 		ret = 0;
@@ -2223,6 +2251,8 @@ static int get_cur_path(struct send_ctx *sctx, u64 ino, u64 gen,
 	fs_path_reset(dest);
 
 	while (!stop && ino != BTRFS_FIRST_FREE_OBJECTID) {
+		struct waiting_dir_move *wdm;
+
 		fs_path_reset(name);
 
 		if (is_waiting_for_rm(sctx, ino)) {
@@ -2233,7 +2263,11 @@ static int get_cur_path(struct send_ctx *sctx, u64 ino, u64 gen,
 			break;
 		}
 
-		if (is_waiting_for_move(sctx, ino)) {
+		wdm = get_waiting_dir_move(sctx, ino);
+		if (wdm && wdm->orphanized) {
+			ret = gen_unique_name(sctx, ino, gen, name);
+			stop = 1;
+		} else if (wdm) {
 			ret = get_first_ref(sctx->parent_root, ino,
 					    &parent_inode, &parent_gen, name);
 		} else {
@@ -2323,13 +2357,23 @@ static int send_subvol_begin(struct send_ctx *sctx)
 	}
 
 	TLV_PUT_STRING(sctx, BTRFS_SEND_A_PATH, name, namelen);
-	TLV_PUT_UUID(sctx, BTRFS_SEND_A_UUID,
-			sctx->send_root->root_item.uuid);
+
+	if (!btrfs_is_empty_uuid(sctx->send_root->root_item.received_uuid))
+		TLV_PUT_UUID(sctx, BTRFS_SEND_A_UUID,
+			    sctx->send_root->root_item.received_uuid);
+	else
+		TLV_PUT_UUID(sctx, BTRFS_SEND_A_UUID,
+			    sctx->send_root->root_item.uuid);
+
 	TLV_PUT_U64(sctx, BTRFS_SEND_A_CTRANSID,
 		    le64_to_cpu(sctx->send_root->root_item.ctransid));
 	if (parent_root) {
-		TLV_PUT_UUID(sctx, BTRFS_SEND_A_CLONE_UUID,
-				sctx->parent_root->root_item.uuid);
+		if (!btrfs_is_empty_uuid(parent_root->root_item.received_uuid))
+			TLV_PUT_UUID(sctx, BTRFS_SEND_A_CLONE_UUID,
+				     parent_root->root_item.received_uuid);
+		else
+			TLV_PUT_UUID(sctx, BTRFS_SEND_A_CLONE_UUID,
+				     parent_root->root_item.uuid);
 		TLV_PUT_U64(sctx, BTRFS_SEND_A_CLONE_CTRANSID,
 			    le64_to_cpu(sctx->parent_root->root_item.ctransid));
 	}
@@ -2530,7 +2574,7 @@ verbose_printk("btrfs: send_create_inode %llu\n", ino);
 	} else if (S_ISSOCK(mode)) {
 		cmd = BTRFS_SEND_C_MKSOCK;
 	} else {
-		printk(KERN_WARNING "btrfs: unexpected inode type %o",
+		btrfs_warn(sctx->send_root->fs_info, "unexpected inode type %o",
 				(int)(mode & S_IFMT));
 		ret = -ENOTSUPP;
 		goto out;
@@ -2923,7 +2967,7 @@ static int is_waiting_for_move(struct send_ctx *sctx, u64 ino)
 	return entry != NULL;
 }
 
-static int add_waiting_dir_move(struct send_ctx *sctx, u64 ino)
+static int add_waiting_dir_move(struct send_ctx *sctx, u64 ino, bool orphanized)
 {
 	struct rb_node **p = &sctx->waiting_dir_moves.rb_node;
 	struct rb_node *parent = NULL;
@@ -2934,6 +2978,7 @@ static int add_waiting_dir_move(struct send_ctx *sctx, u64 ino)
 		return -ENOMEM;
 	dm->ino = ino;
 	dm->rmdir_ino = 0;
+	dm->orphanized = orphanized;
 
 	while (*p) {
 		parent = *p;
@@ -3030,7 +3075,7 @@ static int add_pending_dir_move(struct send_ctx *sctx,
 			goto out;
 	}
 
-	ret = add_waiting_dir_move(sctx, pm->ino);
+	ret = add_waiting_dir_move(sctx, pm->ino, is_orphan);
 	if (ret)
 		goto out;
 
@@ -3353,8 +3398,40 @@ out:
 	return ret;
 }
 
+/*
+ * Check if ino ino1 is an ancestor of inode ino2 in the given root.
+ * Return 1 if true, 0 if false and < 0 on error.
+ */
+static int is_ancestor(struct btrfs_root *root,
+		       const u64 ino1,
+		       const u64 ino1_gen,
+		       const u64 ino2,
+		       struct fs_path *fs_path)
+{
+	u64 ino = ino2;
+
+	while (ino > BTRFS_FIRST_FREE_OBJECTID) {
+		int ret;
+		u64 parent;
+		u64 parent_gen;
+
+		fs_path_reset(fs_path);
+		ret = get_first_ref(root, ino, &parent, &parent_gen, fs_path);
+		if (ret < 0) {
+			if (ret == -ENOENT && ino == ino2)
+				ret = 0;
+			return ret;
+		}
+		if (parent == ino1)
+			return parent_gen == ino1_gen ? 1 : 0;
+		ino = parent;
+	}
+	return 0;
+}
+
 static int wait_for_parent_move(struct send_ctx *sctx,
-				struct recorded_ref *parent_ref)
+				struct recorded_ref *parent_ref,
+				const bool is_orphan)
 {
 	int ret = 0;
 	u64 ino = parent_ref->dir;
@@ -3374,11 +3451,24 @@ static int wait_for_parent_move(struct send_ctx *sctx,
 	 * Our current directory inode may not yet be renamed/moved because some
 	 * ancestor (immediate or not) has to be renamed/moved first. So find if
 	 * such ancestor exists and make sure our own rename/move happens after
-	 * that ancestor is processed.
+	 * that ancestor is processed to avoid path build infinite loops (done
+	 * at get_cur_path()).
 	 */
 	while (ino > BTRFS_FIRST_FREE_OBJECTID) {
 		if (is_waiting_for_move(sctx, ino)) {
-			ret = 1;
+			/*
+			 * If the current inode is an ancestor of ino in the
+			 * parent root, we need to delay the rename of the
+			 * current inode, otherwise don't delayed the rename
+			 * because we can end up with a circular dependency
+			 * of renames, resulting in some directories never
+			 * getting the respective rename operations issued in
+			 * the send stream or getting into infinite path build
+			 * loops.
+			 */
+			ret = is_ancestor(sctx->parent_root,
+					  sctx->cur_ino, sctx->cur_inode_gen,
+					  ino, path_before);
 			break;
 		}
 
@@ -3420,7 +3510,7 @@ out:
 					   ino,
 					   &sctx->new_refs,
 					   &sctx->deleted_refs,
-					   false);
+					   is_orphan);
 		if (!ret)
 			ret = 1;
 	}
@@ -3589,6 +3679,17 @@ verbose_printk("btrfs: process_recorded_refs %llu\n", sctx->cur_ino);
 			}
 		}
 
+		if (S_ISDIR(sctx->cur_inode_mode) && sctx->parent_root &&
+		    can_rename) {
+			ret = wait_for_parent_move(sctx, cur, is_orphan);
+			if (ret < 0)
+				goto out;
+			if (ret == 1) {
+				can_rename = false;
+				*pending_move = 1;
+			}
+		}
+
 		/*
 		 * link/move the ref to the new place. If we have an orphan
 		 * inode, move it and update valid_path. If not, link or move
@@ -3609,18 +3710,11 @@ verbose_printk("btrfs: process_recorded_refs %llu\n", sctx->cur_ino);
 				 * dirs, we always have one new and one deleted
 				 * ref. The deleted ref is ignored later.
 				 */
-				ret = wait_for_parent_move(sctx, cur);
-				if (ret < 0)
-					goto out;
-				if (ret) {
-					*pending_move = 1;
-				} else {
-					ret = send_rename(sctx, valid_path,
-							  cur->full_path);
-					if (!ret)
-						ret = fs_path_copy(valid_path,
-							       cur->full_path);
-				}
+				ret = send_rename(sctx, valid_path,
+						  cur->full_path);
+				if (!ret)
+					ret = fs_path_copy(valid_path,
+							   cur->full_path);
 				if (ret < 0)
 					goto out;
 			} else {
@@ -4508,8 +4602,21 @@ verbose_printk("btrfs: send_clone offset=%llu, len=%d, clone_root=%llu, "
 	if (ret < 0)
 		goto out;
 
-	TLV_PUT_UUID(sctx, BTRFS_SEND_A_CLONE_UUID,
-			clone_root->root->root_item.uuid);
+	/*
+	 * If the parent we're using has a received_uuid set then use that as
+	 * our clone source as that is what we will look for when doing a
+	 * receive.
+	 *
+	 * This covers the case that we create a snapshot off of a received
+	 * subvolume and then use that as the parent and try to receive on a
+	 * different host.
+	 */
+	if (!btrfs_is_empty_uuid(clone_root->root->root_item.received_uuid))
+		TLV_PUT_UUID(sctx, BTRFS_SEND_A_CLONE_UUID,
+			     clone_root->root->root_item.received_uuid);
+	else
+		TLV_PUT_UUID(sctx, BTRFS_SEND_A_CLONE_UUID,
+			     clone_root->root->root_item.uuid);
 	TLV_PUT_U64(sctx, BTRFS_SEND_A_CLONE_CTRANSID,
 		    le64_to_cpu(clone_root->root->root_item.ctransid));
 	TLV_PUT_PATH(sctx, BTRFS_SEND_A_CLONE_PATH, p);
@@ -4590,6 +4697,171 @@ tlv_put_failure:
 	return ret;
 }
 
+static int send_extent_data(struct send_ctx *sctx,
+			    const u64 offset,
+			    const u64 len)
+{
+	u64 sent = 0;
+
+	if (sctx->flags & BTRFS_SEND_FLAG_NO_FILE_DATA)
+		return send_update_extent(sctx, offset, len);
+
+	while (sent < len) {
+		u64 size = len - sent;
+		int ret;
+
+		if (size > BTRFS_SEND_READ_SIZE)
+			size = BTRFS_SEND_READ_SIZE;
+		ret = send_write(sctx, offset + sent, size);
+		if (ret < 0)
+			return ret;
+		if (!ret)
+			break;
+		sent += ret;
+	}
+	return 0;
+}
+
+static int clone_range(struct send_ctx *sctx,
+		       struct clone_root *clone_root,
+		       const u64 disk_byte,
+		       u64 data_offset,
+		       u64 offset,
+		       u64 len)
+{
+	struct btrfs_path *path;
+	struct btrfs_key key;
+	int ret;
+
+	path = alloc_path_for_send();
+	if (!path)
+		return -ENOMEM;
+
+	/*
+	 * We can't send a clone operation for the entire range if we find
+	 * extent items in the respective range in the source file that
+	 * refer to different extents or if we find holes.
+	 * So check for that and do a mix of clone and regular write/copy
+	 * operations if needed.
+	 *
+	 * Example:
+	 *
+	 * mkfs.btrfs -f /dev/sda
+	 * mount /dev/sda /mnt
+	 * xfs_io -f -c "pwrite -S 0xaa 0K 100K" /mnt/foo
+	 * cp --reflink=always /mnt/foo /mnt/bar
+	 * xfs_io -c "pwrite -S 0xbb 50K 50K" /mnt/foo
+	 * btrfs subvolume snapshot -r /mnt /mnt/snap
+	 *
+	 * If when we send the snapshot and we are processing file bar (which
+	 * has a higher inode number than foo) we blindly send a clone operation
+	 * for the [0, 100K[ range from foo to bar, the receiver ends up getting
+	 * a file bar that matches the content of file foo - iow, doesn't match
+	 * the content from bar in the original filesystem.
+	 */
+	key.objectid = clone_root->ino;
+	key.type = BTRFS_EXTENT_DATA_KEY;
+	key.offset = clone_root->offset;
+	ret = btrfs_search_slot(NULL, clone_root->root, &key, path, 0, 0);
+	if (ret < 0)
+		goto out;
+	if (ret > 0 && path->slots[0] > 0) {
+		btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0] - 1);
+		if (key.objectid == clone_root->ino &&
+		    key.type == BTRFS_EXTENT_DATA_KEY)
+			path->slots[0]--;
+	}
+
+	while (true) {
+		struct extent_buffer *leaf = path->nodes[0];
+		int slot = path->slots[0];
+		struct btrfs_file_extent_item *ei;
+		u8 type;
+		u64 ext_len;
+		u64 clone_len;
+
+		if (slot >= btrfs_header_nritems(leaf)) {
+			ret = btrfs_next_leaf(clone_root->root, path);
+			if (ret < 0)
+				goto out;
+			else if (ret > 0)
+				break;
+			continue;
+		}
+
+		btrfs_item_key_to_cpu(leaf, &key, slot);
+
+		/*
+		 * We might have an implicit trailing hole (NO_HOLES feature
+		 * enabled). We deal with it after leaving this loop.
+		 */
+		if (key.objectid != clone_root->ino ||
+		    key.type != BTRFS_EXTENT_DATA_KEY)
+			break;
+
+		ei = btrfs_item_ptr(leaf, slot, struct btrfs_file_extent_item);
+		type = btrfs_file_extent_type(leaf, ei);
+		if (type == BTRFS_FILE_EXTENT_INLINE) {
+			ext_len = btrfs_file_extent_inline_len(leaf, slot, ei);
+			ext_len = PAGE_CACHE_ALIGN(ext_len);
+		} else {
+			ext_len = btrfs_file_extent_num_bytes(leaf, ei);
+		}
+
+		if (key.offset + ext_len <= clone_root->offset)
+			goto next;
+
+		if (key.offset > clone_root->offset) {
+			/* Implicit hole, NO_HOLES feature enabled. */
+			u64 hole_len = key.offset - clone_root->offset;
+
+			if (hole_len > len)
+				hole_len = len;
+			ret = send_extent_data(sctx, offset, hole_len);
+			if (ret < 0)
+				goto out;
+
+			len -= hole_len;
+			if (len == 0)
+				break;
+			offset += hole_len;
+			clone_root->offset += hole_len;
+			data_offset += hole_len;
+		}
+
+		if (key.offset >= clone_root->offset + len)
+			break;
+
+		clone_len = min_t(u64, ext_len, len);
+
+		if (btrfs_file_extent_disk_bytenr(leaf, ei) == disk_byte &&
+		    btrfs_file_extent_offset(leaf, ei) == data_offset)
+			ret = send_clone(sctx, offset, clone_len, clone_root);
+		else
+			ret = send_extent_data(sctx, offset, clone_len);
+
+		if (ret < 0)
+			goto out;
+
+		len -= clone_len;
+		if (len == 0)
+			break;
+		offset += clone_len;
+		clone_root->offset += clone_len;
+		data_offset += clone_len;
+next:
+		path->slots[0]++;
+	}
+
+	if (len > 0)
+		ret = send_extent_data(sctx, offset, len);
+	else
+		ret = 0;
+out:
+	btrfs_free_path(path);
+	return ret;
+}
+
 static int send_write_or_clone(struct send_ctx *sctx,
 			       struct btrfs_path *path,
 			       struct btrfs_key *key,
@@ -4598,9 +4870,7 @@ static int send_write_or_clone(struct send_ctx *sctx,
 	int ret = 0;
 	struct btrfs_file_extent_item *ei;
 	u64 offset = key->offset;
-	u64 pos = 0;
 	u64 len;
-	u32 l;
 	u8 type;
 	u64 bs = sctx->send_root->fs_info->sb->s_blocksize;
 
@@ -4628,22 +4898,15 @@ static int send_write_or_clone(struct send_ctx *sctx,
 	}
 
 	if (clone_root && IS_ALIGNED(offset + len, bs)) {
-		ret = send_clone(sctx, offset, len, clone_root);
-	} else if (sctx->flags & BTRFS_SEND_FLAG_NO_FILE_DATA) {
-		ret = send_update_extent(sctx, offset, len);
+		u64 disk_byte;
+		u64 data_offset;
+
+		disk_byte = btrfs_file_extent_disk_bytenr(path->nodes[0], ei);
+		data_offset = btrfs_file_extent_offset(path->nodes[0], ei);
+		ret = clone_range(sctx, clone_root, disk_byte, data_offset,
+				  offset, len);
 	} else {
-		while (pos < len) {
-			l = len - pos;
-			if (l > BTRFS_SEND_READ_SIZE)
-				l = BTRFS_SEND_READ_SIZE;
-			ret = send_write(sctx, pos + offset, l);
-			if (ret < 0)
-				goto out;
-			if (!ret)
-				break;
-			pos += ret;
-		}
-		ret = 0;
+		ret = send_extent_data(sctx, offset, len);
 	}
 out:
 	return ret;
diff --git a/kernel/fs/btrfs/super.c b/kernel/fs/btrfs/super.c
index 9e66f5e72..fe609b81d 100644
--- a/kernel/fs/btrfs/super.c
+++ b/kernel/fs/btrfs/super.c
@@ -69,7 +69,7 @@ static struct file_system_type btrfs_fs_type;
 
 static int btrfs_remount(struct super_block *sb, int *flags, char *data);
 
-static const char *btrfs_decode_error(int errno)
+const char *btrfs_decode_error(int errno)
 {
 	char *errstr = "unknown";
 
@@ -130,16 +130,18 @@ static void btrfs_handle_error(struct btrfs_fs_info *fs_info)
 	}
 }
 
-#ifdef CONFIG_PRINTK
 /*
  * __btrfs_std_error decodes expected errors from the caller and
  * invokes the approciate error response.
  */
+__cold
 void __btrfs_std_error(struct btrfs_fs_info *fs_info, const char *function,
 		       unsigned int line, int errno, const char *fmt, ...)
 {
 	struct super_block *sb = fs_info->sb;
+#ifdef CONFIG_PRINTK
 	const char *errstr;
+#endif
 
 	/*
 	 * Special case: if the error is EROFS, and we're already
@@ -148,6 +150,7 @@ void __btrfs_std_error(struct btrfs_fs_info *fs_info, const char *function,
 	if (errno == -EROFS && (sb->s_flags & MS_RDONLY))
   		return;
 
+#ifdef CONFIG_PRINTK
 	errstr = btrfs_decode_error(errno);
 	if (fmt) {
 		struct va_format vaf;
@@ -165,6 +168,7 @@ void __btrfs_std_error(struct btrfs_fs_info *fs_info, const char *function,
 		printk(KERN_CRIT "BTRFS: error (device %s) in %s:%d: errno=%d %s\n",
 			sb->s_id, function, line, errno, errstr);
 	}
+#endif
 
 	/* Don't go through full error handling during mount */
 	save_error_info(fs_info);
@@ -172,6 +176,7 @@ void __btrfs_std_error(struct btrfs_fs_info *fs_info, const char *function,
 		btrfs_handle_error(fs_info);
 }
 
+#ifdef CONFIG_PRINTK
 static const char * const logtypes[] = {
 	"emergency",
 	"alert",
@@ -211,27 +216,6 @@ void btrfs_printk(const struct btrfs_fs_info *fs_info, const char *fmt, ...)
 
 	va_end(args);
 }
-
-#else
-
-void __btrfs_std_error(struct btrfs_fs_info *fs_info, const char *function,
-		       unsigned int line, int errno, const char *fmt, ...)
-{
-	struct super_block *sb = fs_info->sb;
-
-	/*
-	 * Special case: if the error is EROFS, and we're already
-	 * under MS_RDONLY, then it is safe here.
-	 */
-	if (errno == -EROFS && (sb->s_flags & MS_RDONLY))
-		return;
-
-	/* Don't go through full error handling during mount */
-	if (sb->s_flags & MS_BORN) {
-		save_error_info(fs_info);
-		btrfs_handle_error(fs_info);
-	}
-}
 #endif
 
 /*
@@ -247,18 +231,11 @@ void __btrfs_std_error(struct btrfs_fs_info *fs_info, const char *function,
  * We'll complete the cleanup in btrfs_end_transaction and
  * btrfs_commit_transaction.
  */
+__cold
 void __btrfs_abort_transaction(struct btrfs_trans_handle *trans,
 			       struct btrfs_root *root, const char *function,
 			       unsigned int line, int errno)
 {
-	/*
-	 * Report first abort since mount
-	 */
-	if (!test_and_set_bit(BTRFS_FS_STATE_TRANS_ABORTED,
-				&root->fs_info->fs_state)) {
-		WARN(1, KERN_DEBUG "BTRFS: Transaction aborted (error %d)\n",
-				errno);
-	}
 	trans->aborted = errno;
 	/* Nothing used. The other threads that have joined this
 	 * transaction may be able to continue. */
@@ -281,6 +258,7 @@ void __btrfs_abort_transaction(struct btrfs_trans_handle *trans,
  * __btrfs_panic decodes unexpected, fatal errors from the caller,
  * issues an alert, and either panics or BUGs, depending on mount options.
  */
+__cold
 void __btrfs_panic(struct btrfs_fs_info *fs_info, const char *function,
 		   unsigned int line, int errno, const char *fmt, ...)
 {
@@ -325,6 +303,9 @@ enum {
 	Opt_commit_interval, Opt_barrier, Opt_nodefrag, Opt_nodiscard,
 	Opt_noenospc_debug, Opt_noflushoncommit, Opt_acl, Opt_datacow,
 	Opt_datasum, Opt_treelog, Opt_noinode_cache,
+#ifdef CONFIG_BTRFS_DEBUG
+	Opt_fragment_data, Opt_fragment_metadata, Opt_fragment_all,
+#endif
 	Opt_err,
 };
 
@@ -377,6 +358,11 @@ static match_table_t tokens = {
 	{Opt_rescan_uuid_tree, "rescan_uuid_tree"},
 	{Opt_fatal_errors, "fatal_errors=%s"},
 	{Opt_commit_interval, "commit=%d"},
+#ifdef CONFIG_BTRFS_DEBUG
+	{Opt_fragment_data, "fragment=data"},
+	{Opt_fragment_metadata, "fragment=metadata"},
+	{Opt_fragment_all, "fragment=all"},
+#endif
 	{Opt_err, NULL},
 };
 
@@ -743,6 +729,22 @@ int btrfs_parse_options(struct btrfs_root *root, char *options)
 				info->commit_interval = BTRFS_DEFAULT_COMMIT_INTERVAL;
 			}
 			break;
+#ifdef CONFIG_BTRFS_DEBUG
+		case Opt_fragment_all:
+			btrfs_info(root->fs_info, "fragmenting all space");
+			btrfs_set_opt(info->mount_opt, FRAGMENT_DATA);
+			btrfs_set_opt(info->mount_opt, FRAGMENT_METADATA);
+			break;
+		case Opt_fragment_metadata:
+			btrfs_info(root->fs_info, "fragmenting metadata");
+			btrfs_set_opt(info->mount_opt,
+				      FRAGMENT_METADATA);
+			break;
+		case Opt_fragment_data:
+			btrfs_info(root->fs_info, "fragmenting data");
+			btrfs_set_opt(info->mount_opt, FRAGMENT_DATA);
+			break;
+#endif
 		case Opt_err:
 			btrfs_info(root->fs_info, "unrecognized mount option '%s'", p);
 			ret = -EINVAL;
@@ -841,33 +843,153 @@ out:
 	return error;
 }
 
-static struct dentry *get_default_root(struct super_block *sb,
-				       u64 subvol_objectid)
+static char *get_subvol_name_from_objectid(struct btrfs_fs_info *fs_info,
+					   u64 subvol_objectid)
 {
-	struct btrfs_fs_info *fs_info = btrfs_sb(sb);
 	struct btrfs_root *root = fs_info->tree_root;
-	struct btrfs_root *new_root;
-	struct btrfs_dir_item *di;
-	struct btrfs_path *path;
-	struct btrfs_key location;
-	struct inode *inode;
-	u64 dir_id;
-	int new = 0;
+	struct btrfs_root *fs_root;
+	struct btrfs_root_ref *root_ref;
+	struct btrfs_inode_ref *inode_ref;
+	struct btrfs_key key;
+	struct btrfs_path *path = NULL;
+	char *name = NULL, *ptr;
+	u64 dirid;
+	int len;
+	int ret;
+
+	path = btrfs_alloc_path();
+	if (!path) {
+		ret = -ENOMEM;
+		goto err;
+	}
+	path->leave_spinning = 1;
+
+	name = kmalloc(PATH_MAX, GFP_NOFS);
+	if (!name) {
+		ret = -ENOMEM;
+		goto err;
+	}
+	ptr = name + PATH_MAX - 1;
+	ptr[0] = '\0';
 
 	/*
-	 * We have a specific subvol we want to mount, just setup location and
-	 * go look up the root.
+	 * Walk up the subvolume trees in the tree of tree roots by root
+	 * backrefs until we hit the top-level subvolume.
 	 */
-	if (subvol_objectid) {
-		location.objectid = subvol_objectid;
-		location.type = BTRFS_ROOT_ITEM_KEY;
-		location.offset = (u64)-1;
-		goto find_root;
+	while (subvol_objectid != BTRFS_FS_TREE_OBJECTID) {
+		key.objectid = subvol_objectid;
+		key.type = BTRFS_ROOT_BACKREF_KEY;
+		key.offset = (u64)-1;
+
+		ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+		if (ret < 0) {
+			goto err;
+		} else if (ret > 0) {
+			ret = btrfs_previous_item(root, path, subvol_objectid,
+						  BTRFS_ROOT_BACKREF_KEY);
+			if (ret < 0) {
+				goto err;
+			} else if (ret > 0) {
+				ret = -ENOENT;
+				goto err;
+			}
+		}
+
+		btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
+		subvol_objectid = key.offset;
+
+		root_ref = btrfs_item_ptr(path->nodes[0], path->slots[0],
+					  struct btrfs_root_ref);
+		len = btrfs_root_ref_name_len(path->nodes[0], root_ref);
+		ptr -= len + 1;
+		if (ptr < name) {
+			ret = -ENAMETOOLONG;
+			goto err;
+		}
+		read_extent_buffer(path->nodes[0], ptr + 1,
+				   (unsigned long)(root_ref + 1), len);
+		ptr[0] = '/';
+		dirid = btrfs_root_ref_dirid(path->nodes[0], root_ref);
+		btrfs_release_path(path);
+
+		key.objectid = subvol_objectid;
+		key.type = BTRFS_ROOT_ITEM_KEY;
+		key.offset = (u64)-1;
+		fs_root = btrfs_read_fs_root_no_name(fs_info, &key);
+		if (IS_ERR(fs_root)) {
+			ret = PTR_ERR(fs_root);
+			goto err;
+		}
+
+		/*
+		 * Walk up the filesystem tree by inode refs until we hit the
+		 * root directory.
+		 */
+		while (dirid != BTRFS_FIRST_FREE_OBJECTID) {
+			key.objectid = dirid;
+			key.type = BTRFS_INODE_REF_KEY;
+			key.offset = (u64)-1;
+
+			ret = btrfs_search_slot(NULL, fs_root, &key, path, 0, 0);
+			if (ret < 0) {
+				goto err;
+			} else if (ret > 0) {
+				ret = btrfs_previous_item(fs_root, path, dirid,
+							  BTRFS_INODE_REF_KEY);
+				if (ret < 0) {
+					goto err;
+				} else if (ret > 0) {
+					ret = -ENOENT;
+					goto err;
+				}
+			}
+
+			btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
+			dirid = key.offset;
+
+			inode_ref = btrfs_item_ptr(path->nodes[0],
+						   path->slots[0],
+						   struct btrfs_inode_ref);
+			len = btrfs_inode_ref_name_len(path->nodes[0],
+						       inode_ref);
+			ptr -= len + 1;
+			if (ptr < name) {
+				ret = -ENAMETOOLONG;
+				goto err;
+			}
+			read_extent_buffer(path->nodes[0], ptr + 1,
+					   (unsigned long)(inode_ref + 1), len);
+			ptr[0] = '/';
+			btrfs_release_path(path);
+		}
 	}
 
+	btrfs_free_path(path);
+	if (ptr == name + PATH_MAX - 1) {
+		name[0] = '/';
+		name[1] = '\0';
+	} else {
+		memmove(name, ptr, name + PATH_MAX - ptr);
+	}
+	return name;
+
+err:
+	btrfs_free_path(path);
+	kfree(name);
+	return ERR_PTR(ret);
+}
+
+static int get_default_subvol_objectid(struct btrfs_fs_info *fs_info, u64 *objectid)
+{
+	struct btrfs_root *root = fs_info->tree_root;
+	struct btrfs_dir_item *di;
+	struct btrfs_path *path;
+	struct btrfs_key location;
+	u64 dir_id;
+
 	path = btrfs_alloc_path();
 	if (!path)
-		return ERR_PTR(-ENOMEM);
+		return -ENOMEM;
 	path->leave_spinning = 1;
 
 	/*
@@ -879,58 +1001,23 @@ static struct dentry *get_default_root(struct super_block *sb,
 	di = btrfs_lookup_dir_item(NULL, root, path, dir_id, "default", 7, 0);
 	if (IS_ERR(di)) {
 		btrfs_free_path(path);
-		return ERR_CAST(di);
+		return PTR_ERR(di);
 	}
 	if (!di) {
 		/*
 		 * Ok the default dir item isn't there.  This is weird since
 		 * it's always been there, but don't freak out, just try and
-		 * mount to root most subvolume.
+		 * mount the top-level subvolume.
 		 */
 		btrfs_free_path(path);
-		dir_id = BTRFS_FIRST_FREE_OBJECTID;
-		new_root = fs_info->fs_root;
-		goto setup_root;
+		*objectid = BTRFS_FS_TREE_OBJECTID;
+		return 0;
 	}
 
 	btrfs_dir_item_key_to_cpu(path->nodes[0], di, &location);
 	btrfs_free_path(path);
-
-find_root:
-	new_root = btrfs_read_fs_root_no_name(fs_info, &location);
-	if (IS_ERR(new_root))
-		return ERR_CAST(new_root);
-
-	if (!(sb->s_flags & MS_RDONLY)) {
-		int ret;
-		down_read(&fs_info->cleanup_work_sem);
-		ret = btrfs_orphan_cleanup(new_root);
-		up_read(&fs_info->cleanup_work_sem);
-		if (ret)
-			return ERR_PTR(ret);
-	}
-
-	dir_id = btrfs_root_dirid(&new_root->root_item);
-setup_root:
-	location.objectid = dir_id;
-	location.type = BTRFS_INODE_ITEM_KEY;
-	location.offset = 0;
-
-	inode = btrfs_iget(sb, &location, new_root, &new);
-	if (IS_ERR(inode))
-		return ERR_CAST(inode);
-
-	/*
-	 * If we're just mounting the root most subvol put the inode and return
-	 * a reference to the dentry.  We will have already gotten a reference
-	 * to the inode in btrfs_fill_super so we're good to go.
-	 */
-	if (!new && d_inode(sb->s_root) == inode) {
-		iput(inode);
-		return dget(sb->s_root);
-	}
-
-	return d_obtain_root(inode);
+	*objectid = location.objectid;
+	return 0;
 }
 
 static int btrfs_fill_super(struct super_block *sb,
@@ -953,6 +1040,7 @@ static int btrfs_fill_super(struct super_block *sb,
 	sb->s_flags |= MS_POSIXACL;
 #endif
 	sb->s_flags |= MS_I_VERSION;
+	sb->s_iflags |= SB_I_CGROUPWB;
 	err = open_ctree(sb, fs_devices, (char *)data);
 	if (err) {
 		printk(KERN_ERR "BTRFS: open_ctree failed\n");
@@ -1108,6 +1196,16 @@ static int btrfs_show_options(struct seq_file *seq, struct dentry *dentry)
 		seq_puts(seq, ",fatal_errors=panic");
 	if (info->commit_interval != BTRFS_DEFAULT_COMMIT_INTERVAL)
 		seq_printf(seq, ",commit=%d", info->commit_interval);
+#ifdef CONFIG_BTRFS_DEBUG
+	if (btrfs_test_opt(root, FRAGMENT_DATA))
+		seq_puts(seq, ",fragment=data");
+	if (btrfs_test_opt(root, FRAGMENT_METADATA))
+		seq_puts(seq, ",fragment=metadata");
+#endif
+	seq_printf(seq, ",subvolid=%llu",
+		  BTRFS_I(d_inode(dentry))->root->root_key.objectid);
+	seq_puts(seq, ",subvol=");
+	seq_dentry(seq, dentry, " \t\n\\");
 	return 0;
 }
 
@@ -1138,107 +1236,139 @@ static inline int is_subvolume_inode(struct inode *inode)
 }
 
 /*
- * This will strip out the subvol=%s argument for an argument string and add
- * subvolid=0 to make sure we get the actual tree root for path walking to the
- * subvol we want.
+ * This will add subvolid=0 to the argument string while removing any subvol=
+ * and subvolid= arguments to make sure we get the top-level root for path
+ * walking to the subvol we want.
  */
 static char *setup_root_args(char *args)
 {
-	unsigned len = strlen(args) + 2 + 1;
-	char *src, *dst, *buf;
-
-	/*
-	 * We need the same args as before, but with this substitution:
-	 * s!subvol=[^,]+!subvolid=0!
-	 *
-	 * Since the replacement string is up to 2 bytes longer than the
-	 * original, allocate strlen(args) + 2 + 1 bytes.
-	 */
+	char *buf, *dst, *sep;
 
-	src = strstr(args, "subvol=");
-	/* This shouldn't happen, but just in case.. */
-	if (!src)
-		return NULL;
+	if (!args)
+		return kstrdup("subvolid=0", GFP_NOFS);
 
-	buf = dst = kmalloc(len, GFP_NOFS);
+	/* The worst case is that we add ",subvolid=0" to the end. */
+	buf = dst = kmalloc(strlen(args) + strlen(",subvolid=0") + 1, GFP_NOFS);
 	if (!buf)
 		return NULL;
 
-	/*
-	 * If the subvol= arg is not at the start of the string,
-	 * copy whatever precedes it into buf.
-	 */
-	if (src != args) {
-		*src++ = '\0';
-		strcpy(buf, args);
-		dst += strlen(args);
+	while (1) {
+		sep = strchrnul(args, ',');
+		if (!strstarts(args, "subvol=") &&
+		    !strstarts(args, "subvolid=")) {
+			memcpy(dst, args, sep - args);
+			dst += sep - args;
+			*dst++ = ',';
+		}
+		if (*sep)
+			args = sep + 1;
+		else
+			break;
 	}
-
 	strcpy(dst, "subvolid=0");
-	dst += strlen("subvolid=0");
-
-	/*
-	 * If there is a "," after the original subvol=... string,
-	 * copy that suffix into our buffer.  Otherwise, we're done.
-	 */
-	src = strchr(src, ',');
-	if (src)
-		strcpy(dst, src);
 
 	return buf;
 }
 
-static struct dentry *mount_subvol(const char *subvol_name, int flags,
-				   const char *device_name, char *data)
+static struct dentry *mount_subvol(const char *subvol_name, u64 subvol_objectid,
+				   int flags, const char *device_name,
+				   char *data)
 {
 	struct dentry *root;
-	struct vfsmount *mnt;
+	struct vfsmount *mnt = NULL;
 	char *newargs;
+	int ret;
 
 	newargs = setup_root_args(data);
-	if (!newargs)
-		return ERR_PTR(-ENOMEM);
-	mnt = vfs_kern_mount(&btrfs_fs_type, flags, device_name,
-			     newargs);
+	if (!newargs) {
+		root = ERR_PTR(-ENOMEM);
+		goto out;
+	}
 
-	if (PTR_RET(mnt) == -EBUSY) {
+	mnt = vfs_kern_mount(&btrfs_fs_type, flags, device_name, newargs);
+	if (PTR_ERR_OR_ZERO(mnt) == -EBUSY) {
 		if (flags & MS_RDONLY) {
-			mnt = vfs_kern_mount(&btrfs_fs_type, flags & ~MS_RDONLY, device_name,
-					     newargs);
+			mnt = vfs_kern_mount(&btrfs_fs_type, flags & ~MS_RDONLY,
+					     device_name, newargs);
 		} else {
-			int r;
-			mnt = vfs_kern_mount(&btrfs_fs_type, flags | MS_RDONLY, device_name,
-					     newargs);
+			mnt = vfs_kern_mount(&btrfs_fs_type, flags | MS_RDONLY,
+					     device_name, newargs);
 			if (IS_ERR(mnt)) {
-				kfree(newargs);
-				return ERR_CAST(mnt);
+				root = ERR_CAST(mnt);
+				mnt = NULL;
+				goto out;
 			}
 
-			r = btrfs_remount(mnt->mnt_sb, &flags, NULL);
-			if (r < 0) {
-				/* FIXME: release vfsmount mnt ??*/
-				kfree(newargs);
-				return ERR_PTR(r);
+			down_write(&mnt->mnt_sb->s_umount);
+			ret = btrfs_remount(mnt->mnt_sb, &flags, NULL);
+			up_write(&mnt->mnt_sb->s_umount);
+			if (ret < 0) {
+				root = ERR_PTR(ret);
+				goto out;
 			}
 		}
 	}
+	if (IS_ERR(mnt)) {
+		root = ERR_CAST(mnt);
+		mnt = NULL;
+		goto out;
+	}
 
-	kfree(newargs);
+	if (!subvol_name) {
+		if (!subvol_objectid) {
+			ret = get_default_subvol_objectid(btrfs_sb(mnt->mnt_sb),
+							  &subvol_objectid);
+			if (ret) {
+				root = ERR_PTR(ret);
+				goto out;
+			}
+		}
+		subvol_name = get_subvol_name_from_objectid(btrfs_sb(mnt->mnt_sb),
+							    subvol_objectid);
+		if (IS_ERR(subvol_name)) {
+			root = ERR_CAST(subvol_name);
+			subvol_name = NULL;
+			goto out;
+		}
 
-	if (IS_ERR(mnt))
-		return ERR_CAST(mnt);
+	}
 
 	root = mount_subtree(mnt, subvol_name);
+	/* mount_subtree() drops our reference on the vfsmount. */
+	mnt = NULL;
 
-	if (!IS_ERR(root) && !is_subvolume_inode(d_inode(root))) {
+	if (!IS_ERR(root)) {
 		struct super_block *s = root->d_sb;
-		dput(root);
-		root = ERR_PTR(-EINVAL);
-		deactivate_locked_super(s);
-		printk(KERN_ERR "BTRFS: '%s' is not a valid subvolume\n",
-				subvol_name);
+		struct inode *root_inode = d_inode(root);
+		u64 root_objectid = BTRFS_I(root_inode)->root->root_key.objectid;
+
+		ret = 0;
+		if (!is_subvolume_inode(root_inode)) {
+			pr_err("BTRFS: '%s' is not a valid subvolume\n",
+			       subvol_name);
+			ret = -EINVAL;
+		}
+		if (subvol_objectid && root_objectid != subvol_objectid) {
+			/*
+			 * This will also catch a race condition where a
+			 * subvolume which was passed by ID is renamed and
+			 * another subvolume is renamed over the old location.
+			 */
+			pr_err("BTRFS: subvol '%s' does not match subvolid %llu\n",
+			       subvol_name, subvol_objectid);
+			ret = -EINVAL;
+		}
+		if (ret) {
+			dput(root);
+			root = ERR_PTR(ret);
+			deactivate_locked_super(s);
+		}
 	}
 
+out:
+	mntput(mnt);
+	kfree(newargs);
+	kfree(subvol_name);
 	return root;
 }
 
@@ -1303,7 +1433,6 @@ static struct dentry *btrfs_mount(struct file_system_type *fs_type, int flags,
 {
 	struct block_device *bdev = NULL;
 	struct super_block *s;
-	struct dentry *root;
 	struct btrfs_fs_devices *fs_devices = NULL;
 	struct btrfs_fs_info *fs_info = NULL;
 	struct security_mnt_opts new_sec_opts;
@@ -1323,10 +1452,10 @@ static struct dentry *btrfs_mount(struct file_system_type *fs_type, int flags,
 		return ERR_PTR(error);
 	}
 
-	if (subvol_name) {
-		root = mount_subvol(subvol_name, flags, device_name, data);
-		kfree(subvol_name);
-		return root;
+	if (subvol_name || subvol_objectid != BTRFS_FS_TREE_OBJECTID) {
+		/* mount_subvol() will free subvol_name. */
+		return mount_subvol(subvol_name, subvol_objectid, flags,
+				    device_name, data);
 	}
 
 	security_init_mnt_opts(&new_sec_opts);
@@ -1392,23 +1521,19 @@ static struct dentry *btrfs_mount(struct file_system_type *fs_type, int flags,
 		error = btrfs_fill_super(s, fs_devices, data,
 					 flags & MS_SILENT ? 1 : 0);
 	}
-
-	root = !error ? get_default_root(s, subvol_objectid) : ERR_PTR(error);
-	if (IS_ERR(root)) {
+	if (error) {
 		deactivate_locked_super(s);
-		error = PTR_ERR(root);
 		goto error_sec_opts;
 	}
 
 	fs_info = btrfs_sb(s);
 	error = setup_security_options(fs_info, s, &new_sec_opts);
 	if (error) {
-		dput(root);
 		deactivate_locked_super(s);
 		goto error_sec_opts;
 	}
 
-	return root;
+	return dget(s->s_root);
 
 error_close_devices:
 	btrfs_close_devices(fs_devices);
@@ -1539,6 +1664,15 @@ static int btrfs_remount(struct super_block *sb, int *flags, char *data)
 
 		sb->s_flags |= MS_RDONLY;
 
+		/*
+		 * Setting MS_RDONLY will put the cleaner thread to
+		 * sleep at the next loop if it's already active.
+		 * If it's already asleep, we'll leave unused block
+		 * groups on disk until we're mounted read-write again
+		 * unless we clean them up here.
+		 */
+		btrfs_delete_unused_bgs(fs_info);
+
 		btrfs_dev_replace_suspend_for_unmount(fs_info);
 		btrfs_scrub_cancel(fs_info);
 		btrfs_pause_balance(fs_info);
@@ -1822,6 +1956,8 @@ static int btrfs_calc_avail_data_space(struct btrfs_root *root, u64 *free_bytes)
  * there are other factors that may change the result (like a new metadata
  * chunk).
  *
+ * If metadata is exhausted, f_bavail will be 0.
+ *
  * FIXME: not accurate for mixed block groups, total and free/used are ok,
  * available appears slightly larger.
  */
@@ -1833,11 +1969,13 @@ static int btrfs_statfs(struct dentry *dentry, struct kstatfs *buf)
 	struct btrfs_space_info *found;
 	u64 total_used = 0;
 	u64 total_free_data = 0;
+	u64 total_free_meta = 0;
 	int bits = dentry->d_sb->s_blocksize_bits;
 	__be32 *fsid = (__be32 *)fs_info->fsid;
 	unsigned factor = 1;
 	struct btrfs_block_rsv *block_rsv = &fs_info->global_block_rsv;
 	int ret;
+	u64 thresh = 0;
 
 	/*
 	 * holding chunk_muext to avoid allocating new chunks, holding
@@ -1863,6 +2001,8 @@ static int btrfs_statfs(struct dentry *dentry, struct kstatfs *buf)
 				}
 			}
 		}
+		if (found->flags & BTRFS_BLOCK_GROUP_METADATA)
+			total_free_meta += found->disk_total - found->disk_used;
 
 		total_used += found->disk_used;
 	}
@@ -1885,6 +2025,24 @@ static int btrfs_statfs(struct dentry *dentry, struct kstatfs *buf)
 	buf->f_bavail += div_u64(total_free_data, factor);
 	buf->f_bavail = buf->f_bavail >> bits;
 
+	/*
+	 * We calculate the remaining metadata space minus global reserve. If
+	 * this is (supposedly) smaller than zero, there's no space. But this
+	 * does not hold in practice, the exhausted state happens where's still
+	 * some positive delta. So we apply some guesswork and compare the
+	 * delta to a 4M threshold.  (Practically observed delta was ~2M.)
+	 *
+	 * We probably cannot calculate the exact threshold value because this
+	 * depends on the internal reservations requested by various
+	 * operations, so some operations that consume a few metadata will
+	 * succeed even if the Avail is zero. But this is better than the other
+	 * way around.
+	 */
+	thresh = 4 * 1024 * 1024;
+
+	if (total_free_meta - thresh < block_rsv->size)
+		buf->f_bavail = 0;
+
 	buf->f_type = BTRFS_SUPER_MAGIC;
 	buf->f_bsize = dentry->d_sb->s_blocksize;
 	buf->f_namelen = BTRFS_NAME_LEN;
@@ -2052,8 +2210,7 @@ static int btrfs_interface_init(void)
 
 static void btrfs_interface_exit(void)
 {
-	if (misc_deregister(&btrfs_misc) < 0)
-		printk(KERN_INFO "BTRFS: misc_deregister failed for control device\n");
+	misc_deregister(&btrfs_misc);
 }
 
 static void btrfs_print_info(void)
diff --git a/kernel/fs/btrfs/sysfs.c b/kernel/fs/btrfs/sysfs.c
index e8a4c86d2..e0ac85949 100644
--- a/kernel/fs/btrfs/sysfs.c
+++ b/kernel/fs/btrfs/sysfs.c
@@ -33,6 +33,7 @@
 #include "volumes.h"
 
 static inline struct btrfs_fs_info *to_fs_info(struct kobject *kobj);
+static inline struct btrfs_fs_devices *to_fs_devs(struct kobject *kobj);
 
 static u64 get_features(struct btrfs_fs_info *fs_info,
 			enum btrfs_feature_set set)
@@ -428,7 +429,7 @@ static ssize_t btrfs_clone_alignment_show(struct kobject *kobj,
 
 BTRFS_ATTR(clone_alignment, btrfs_clone_alignment_show);
 
-static struct attribute *btrfs_attrs[] = {
+static const struct attribute *btrfs_attrs[] = {
 	BTRFS_ATTR_PTR(label),
 	BTRFS_ATTR_PTR(nodesize),
 	BTRFS_ATTR_PTR(sectorsize),
@@ -436,23 +437,31 @@ static struct attribute *btrfs_attrs[] = {
 	NULL,
 };
 
-static void btrfs_release_super_kobj(struct kobject *kobj)
+static void btrfs_release_fsid_kobj(struct kobject *kobj)
 {
-	struct btrfs_fs_info *fs_info = to_fs_info(kobj);
-	complete(&fs_info->kobj_unregister);
+	struct btrfs_fs_devices *fs_devs = to_fs_devs(kobj);
+
+	memset(&fs_devs->fsid_kobj, 0, sizeof(struct kobject));
+	complete(&fs_devs->kobj_unregister);
 }
 
 static struct kobj_type btrfs_ktype = {
 	.sysfs_ops	= &kobj_sysfs_ops,
-	.release	= btrfs_release_super_kobj,
-	.default_attrs	= btrfs_attrs,
+	.release	= btrfs_release_fsid_kobj,
 };
 
+static inline struct btrfs_fs_devices *to_fs_devs(struct kobject *kobj)
+{
+	if (kobj->ktype != &btrfs_ktype)
+		return NULL;
+	return container_of(kobj, struct btrfs_fs_devices, fsid_kobj);
+}
+
 static inline struct btrfs_fs_info *to_fs_info(struct kobject *kobj)
 {
 	if (kobj->ktype != &btrfs_ktype)
 		return NULL;
-	return container_of(kobj, struct btrfs_fs_info, super_kobj);
+	return to_fs_devs(kobj)->fs_info;
 }
 
 #define NUM_FEATURE_BITS 64
@@ -493,12 +502,12 @@ static int addrm_unknown_feature_attrs(struct btrfs_fs_info *fs_info, bool add)
 			attrs[0] = &fa->kobj_attr.attr;
 			if (add) {
 				int ret;
-				ret = sysfs_merge_group(&fs_info->super_kobj,
+				ret = sysfs_merge_group(&fs_info->fs_devices->fsid_kobj,
 							&agroup);
 				if (ret)
 					return ret;
 			} else
-				sysfs_unmerge_group(&fs_info->super_kobj,
+				sysfs_unmerge_group(&fs_info->fs_devices->fsid_kobj,
 						    &agroup);
 		}
 
@@ -506,25 +515,49 @@ static int addrm_unknown_feature_attrs(struct btrfs_fs_info *fs_info, bool add)
 	return 0;
 }
 
-static void __btrfs_sysfs_remove_one(struct btrfs_fs_info *fs_info)
+static void __btrfs_sysfs_remove_fsid(struct btrfs_fs_devices *fs_devs)
+{
+	if (fs_devs->device_dir_kobj) {
+		kobject_del(fs_devs->device_dir_kobj);
+		kobject_put(fs_devs->device_dir_kobj);
+		fs_devs->device_dir_kobj = NULL;
+	}
+
+	if (fs_devs->fsid_kobj.state_initialized) {
+		kobject_del(&fs_devs->fsid_kobj);
+		kobject_put(&fs_devs->fsid_kobj);
+		wait_for_completion(&fs_devs->kobj_unregister);
+	}
+}
+
+/* when fs_devs is NULL it will remove all fsid kobject */
+void btrfs_sysfs_remove_fsid(struct btrfs_fs_devices *fs_devs)
 {
-	kobject_del(&fs_info->super_kobj);
-	kobject_put(&fs_info->super_kobj);
-	wait_for_completion(&fs_info->kobj_unregister);
+	struct list_head *fs_uuids = btrfs_get_fs_uuids();
+
+	if (fs_devs) {
+		__btrfs_sysfs_remove_fsid(fs_devs);
+		return;
+	}
+
+	list_for_each_entry(fs_devs, fs_uuids, list) {
+		__btrfs_sysfs_remove_fsid(fs_devs);
+	}
 }
 
-void btrfs_sysfs_remove_one(struct btrfs_fs_info *fs_info)
+void btrfs_sysfs_remove_mounted(struct btrfs_fs_info *fs_info)
 {
+	btrfs_reset_fs_info_ptr(fs_info);
+
 	if (fs_info->space_info_kobj) {
 		sysfs_remove_files(fs_info->space_info_kobj, allocation_attrs);
 		kobject_del(fs_info->space_info_kobj);
 		kobject_put(fs_info->space_info_kobj);
 	}
-	kobject_del(fs_info->device_dir_kobj);
-	kobject_put(fs_info->device_dir_kobj);
 	addrm_unknown_feature_attrs(fs_info, false);
-	sysfs_remove_group(&fs_info->super_kobj, &btrfs_feature_attr_group);
-	__btrfs_sysfs_remove_one(fs_info);
+	sysfs_remove_group(&fs_info->fs_devices->fsid_kobj, &btrfs_feature_attr_group);
+	sysfs_remove_files(&fs_info->fs_devices->fsid_kobj, btrfs_attrs);
+	btrfs_sysfs_rm_device_link(fs_info->fs_devices, NULL);
 }
 
 const char * const btrfs_feature_set_names[3] = {
@@ -602,40 +635,60 @@ static void init_feature_attrs(void)
 	}
 }
 
-int btrfs_kobj_rm_device(struct btrfs_fs_info *fs_info,
+/* when one_device is NULL, it removes all device links */
+
+int btrfs_sysfs_rm_device_link(struct btrfs_fs_devices *fs_devices,
 		struct btrfs_device *one_device)
 {
 	struct hd_struct *disk;
 	struct kobject *disk_kobj;
 
-	if (!fs_info->device_dir_kobj)
+	if (!fs_devices->device_dir_kobj)
 		return -EINVAL;
 
 	if (one_device && one_device->bdev) {
 		disk = one_device->bdev->bd_part;
 		disk_kobj = &part_to_dev(disk)->kobj;
 
-		sysfs_remove_link(fs_info->device_dir_kobj,
+		sysfs_remove_link(fs_devices->device_dir_kobj,
+						disk_kobj->name);
+	}
+
+	if (one_device)
+		return 0;
+
+	list_for_each_entry(one_device,
+			&fs_devices->devices, dev_list) {
+		if (!one_device->bdev)
+			continue;
+		disk = one_device->bdev->bd_part;
+		disk_kobj = &part_to_dev(disk)->kobj;
+
+		sysfs_remove_link(fs_devices->device_dir_kobj,
 						disk_kobj->name);
 	}
 
 	return 0;
 }
 
-int btrfs_kobj_add_device(struct btrfs_fs_info *fs_info,
-		struct btrfs_device *one_device)
+int btrfs_sysfs_add_device(struct btrfs_fs_devices *fs_devs)
 {
-	int error = 0;
-	struct btrfs_fs_devices *fs_devices = fs_info->fs_devices;
-	struct btrfs_device *dev;
-
-	if (!fs_info->device_dir_kobj)
-		fs_info->device_dir_kobj = kobject_create_and_add("devices",
-						&fs_info->super_kobj);
+	if (!fs_devs->device_dir_kobj)
+		fs_devs->device_dir_kobj = kobject_create_and_add("devices",
+						&fs_devs->fsid_kobj);
 
-	if (!fs_info->device_dir_kobj)
+	if (!fs_devs->device_dir_kobj)
 		return -ENOMEM;
 
+	return 0;
+}
+
+int btrfs_sysfs_add_device_link(struct btrfs_fs_devices *fs_devices,
+				struct btrfs_device *one_device)
+{
+	int error = 0;
+	struct btrfs_device *dev;
+
 	list_for_each_entry(dev, &fs_devices->devices, dev_list) {
 		struct hd_struct *disk;
 		struct kobject *disk_kobj;
@@ -649,7 +702,7 @@ int btrfs_kobj_add_device(struct btrfs_fs_info *fs_info,
 		disk = dev->bdev->bd_part;
 		disk_kobj = &part_to_dev(disk)->kobj;
 
-		error = sysfs_create_link(fs_info->device_dir_kobj,
+		error = sysfs_create_link(fs_devices->device_dir_kobj,
 					  disk_kobj, disk_kobj->name);
 		if (error)
 			break;
@@ -667,34 +720,51 @@ static struct dentry *btrfs_debugfs_root_dentry;
 /* Debugging tunables and exported data */
 u64 btrfs_debugfs_test;
 
-int btrfs_sysfs_add_one(struct btrfs_fs_info *fs_info)
+/*
+ * Can be called by the device discovery thread.
+ * And parent can be specified for seed device
+ */
+int btrfs_sysfs_add_fsid(struct btrfs_fs_devices *fs_devs,
+				struct kobject *parent)
 {
 	int error;
 
-	init_completion(&fs_info->kobj_unregister);
-	fs_info->super_kobj.kset = btrfs_kset;
-	error = kobject_init_and_add(&fs_info->super_kobj, &btrfs_ktype, NULL,
-				     "%pU", fs_info->fsid);
+	init_completion(&fs_devs->kobj_unregister);
+	fs_devs->fsid_kobj.kset = btrfs_kset;
+	error = kobject_init_and_add(&fs_devs->fsid_kobj,
+				&btrfs_ktype, parent, "%pU", fs_devs->fsid);
+	return error;
+}
+
+int btrfs_sysfs_add_mounted(struct btrfs_fs_info *fs_info)
+{
+	int error;
+	struct btrfs_fs_devices *fs_devs = fs_info->fs_devices;
+	struct kobject *fsid_kobj = &fs_devs->fsid_kobj;
+
+	btrfs_set_fs_info_ptr(fs_info);
+
+	error = btrfs_sysfs_add_device_link(fs_devs, NULL);
 	if (error)
 		return error;
 
-	error = sysfs_create_group(&fs_info->super_kobj,
-				   &btrfs_feature_attr_group);
+	error = sysfs_create_files(fsid_kobj, btrfs_attrs);
 	if (error) {
-		__btrfs_sysfs_remove_one(fs_info);
+		btrfs_sysfs_rm_device_link(fs_devs, NULL);
 		return error;
 	}
 
-	error = addrm_unknown_feature_attrs(fs_info, true);
+	error = sysfs_create_group(fsid_kobj,
+				   &btrfs_feature_attr_group);
 	if (error)
 		goto failure;
 
-	error = btrfs_kobj_add_device(fs_info, NULL);
+	error = addrm_unknown_feature_attrs(fs_info, true);
 	if (error)
 		goto failure;
 
 	fs_info->space_info_kobj = kobject_create_and_add("allocation",
-						  &fs_info->super_kobj);
+						  fsid_kobj);
 	if (!fs_info->space_info_kobj) {
 		error = -ENOMEM;
 		goto failure;
@@ -706,7 +776,7 @@ int btrfs_sysfs_add_one(struct btrfs_fs_info *fs_info)
 
 	return 0;
 failure:
-	btrfs_sysfs_remove_one(fs_info);
+	btrfs_sysfs_remove_mounted(fs_info);
 	return error;
 }
 
diff --git a/kernel/fs/btrfs/sysfs.h b/kernel/fs/btrfs/sysfs.h
index 3a4bbed72..9c0952212 100644
--- a/kernel/fs/btrfs/sysfs.h
+++ b/kernel/fs/btrfs/sysfs.h
@@ -82,8 +82,12 @@ char *btrfs_printable_features(enum btrfs_feature_set set, u64 flags);
 extern const char * const btrfs_feature_set_names[3];
 extern struct kobj_type space_info_ktype;
 extern struct kobj_type btrfs_raid_ktype;
-int btrfs_kobj_add_device(struct btrfs_fs_info *fs_info,
+int btrfs_sysfs_add_device_link(struct btrfs_fs_devices *fs_devices,
 		struct btrfs_device *one_device);
-int btrfs_kobj_rm_device(struct btrfs_fs_info *fs_info,
+int btrfs_sysfs_rm_device_link(struct btrfs_fs_devices *fs_devices,
                 struct btrfs_device *one_device);
+int btrfs_sysfs_add_fsid(struct btrfs_fs_devices *fs_devs,
+				struct kobject *parent);
+int btrfs_sysfs_add_device(struct btrfs_fs_devices *fs_devs);
+void btrfs_sysfs_remove_fsid(struct btrfs_fs_devices *fs_devs);
 #endif /* _BTRFS_SYSFS_H_ */
diff --git a/kernel/fs/btrfs/tests/free-space-tests.c b/kernel/fs/btrfs/tests/free-space-tests.c
index 2299bfde3..8b72b005b 100644
--- a/kernel/fs/btrfs/tests/free-space-tests.c
+++ b/kernel/fs/btrfs/tests/free-space-tests.c
@@ -19,6 +19,7 @@
 #include <linux/slab.h>
 #include "btrfs-tests.h"
 #include "../ctree.h"
+#include "../disk-io.h"
 #include "../free-space-cache.h"
 
 #define BITS_PER_BITMAP		(PAGE_CACHE_SIZE * 8)
@@ -35,6 +36,12 @@ static struct btrfs_block_group_cache *init_test_block_group(void)
 		kfree(cache);
 		return NULL;
 	}
+	cache->fs_info = btrfs_alloc_dummy_fs_info();
+	if (!cache->fs_info) {
+		kfree(cache->free_space_ctl);
+		kfree(cache);
+		return NULL;
+	}
 
 	cache->key.objectid = 0;
 	cache->key.offset = 1024 * 1024 * 1024;
@@ -879,7 +886,8 @@ test_steal_space_from_bitmap_to_extent(struct btrfs_block_group_cache *cache)
 int btrfs_test_free_space_cache(void)
 {
 	struct btrfs_block_group_cache *cache;
-	int ret;
+	struct btrfs_root *root = NULL;
+	int ret = -ENOMEM;
 
 	test_msg("Running btrfs free space cache tests\n");
 
@@ -889,6 +897,19 @@ int btrfs_test_free_space_cache(void)
 		return 0;
 	}
 
+	root = btrfs_alloc_dummy_root();
+	if (IS_ERR(root)) {
+		ret = PTR_ERR(root);
+		goto out;
+	}
+
+	root->fs_info = btrfs_alloc_dummy_fs_info();
+	if (!root->fs_info)
+		goto out;
+
+	root->fs_info->extent_root = root;
+	cache->fs_info = root->fs_info;
+
 	ret = test_extents(cache);
 	if (ret)
 		goto out;
@@ -904,6 +925,7 @@ out:
 	__btrfs_remove_free_space_cache(cache->free_space_ctl);
 	kfree(cache->free_space_ctl);
 	kfree(cache);
+	btrfs_free_dummy_root(root);
 	test_msg("Free space cache tests finished\n");
 	return ret;
 }
diff --git a/kernel/fs/btrfs/tests/qgroup-tests.c b/kernel/fs/btrfs/tests/qgroup-tests.c
index c32a7ba76..846d277b1 100644
--- a/kernel/fs/btrfs/tests/qgroup-tests.c
+++ b/kernel/fs/btrfs/tests/qgroup-tests.c
@@ -21,6 +21,7 @@
 #include "../transaction.h"
 #include "../disk-io.h"
 #include "../qgroup.h"
+#include "../backref.h"
 
 static void init_dummy_trans(struct btrfs_trans_handle *trans)
 {
@@ -227,6 +228,8 @@ static int test_no_shared_qgroup(struct btrfs_root *root)
 {
 	struct btrfs_trans_handle trans;
 	struct btrfs_fs_info *fs_info = root->fs_info;
+	struct ulist *old_roots = NULL;
+	struct ulist *new_roots = NULL;
 	int ret;
 
 	init_dummy_trans(&trans);
@@ -238,10 +241,15 @@ static int test_no_shared_qgroup(struct btrfs_root *root)
 		return ret;
 	}
 
-	ret = btrfs_qgroup_record_ref(&trans, fs_info, 5, 4096, 4096,
-				      BTRFS_QGROUP_OPER_ADD_EXCL, 0);
+	/*
+	 * Since the test trans doesn't havee the complicated delayed refs,
+	 * we can only call btrfs_qgroup_account_extent() directly to test
+	 * quota.
+	 */
+	ret = btrfs_find_all_roots(&trans, fs_info, 4096, 0, &old_roots);
 	if (ret) {
-		test_msg("Couldn't add space to a qgroup %d\n", ret);
+		ulist_free(old_roots);
+		test_msg("Couldn't find old roots: %d\n", ret);
 		return ret;
 	}
 
@@ -249,9 +257,18 @@ static int test_no_shared_qgroup(struct btrfs_root *root)
 	if (ret)
 		return ret;
 
-	ret = btrfs_delayed_qgroup_accounting(&trans, fs_info);
+	ret = btrfs_find_all_roots(&trans, fs_info, 4096, 0, &new_roots);
+	if (ret) {
+		ulist_free(old_roots);
+		ulist_free(new_roots);
+		test_msg("Couldn't find old roots: %d\n", ret);
+		return ret;
+	}
+
+	ret = btrfs_qgroup_account_extent(&trans, fs_info, 4096, 4096,
+					  old_roots, new_roots);
 	if (ret) {
-		test_msg("Delayed qgroup accounting failed %d\n", ret);
+		test_msg("Couldn't account space for a qgroup %d\n", ret);
 		return ret;
 	}
 
@@ -259,21 +276,32 @@ static int test_no_shared_qgroup(struct btrfs_root *root)
 		test_msg("Qgroup counts didn't match expected values\n");
 		return -EINVAL;
 	}
+	old_roots = NULL;
+	new_roots = NULL;
+
+	ret = btrfs_find_all_roots(&trans, fs_info, 4096, 0, &old_roots);
+	if (ret) {
+		ulist_free(old_roots);
+		test_msg("Couldn't find old roots: %d\n", ret);
+		return ret;
+	}
 
 	ret = remove_extent_item(root, 4096, 4096);
 	if (ret)
 		return -EINVAL;
 
-	ret = btrfs_qgroup_record_ref(&trans, fs_info, 5, 4096, 4096,
-				      BTRFS_QGROUP_OPER_SUB_EXCL, 0);
+	ret = btrfs_find_all_roots(&trans, fs_info, 4096, 0, &new_roots);
 	if (ret) {
-		test_msg("Couldn't remove space from the qgroup %d\n", ret);
-		return -EINVAL;
+		ulist_free(old_roots);
+		ulist_free(new_roots);
+		test_msg("Couldn't find old roots: %d\n", ret);
+		return ret;
 	}
 
-	ret = btrfs_delayed_qgroup_accounting(&trans, fs_info);
+	ret = btrfs_qgroup_account_extent(&trans, fs_info, 4096, 4096,
+					  old_roots, new_roots);
 	if (ret) {
-		test_msg("Qgroup accounting failed %d\n", ret);
+		test_msg("Couldn't account space for a qgroup %d\n", ret);
 		return -EINVAL;
 	}
 
@@ -294,6 +322,8 @@ static int test_multiple_refs(struct btrfs_root *root)
 {
 	struct btrfs_trans_handle trans;
 	struct btrfs_fs_info *fs_info = root->fs_info;
+	struct ulist *old_roots = NULL;
+	struct ulist *new_roots = NULL;
 	int ret;
 
 	init_dummy_trans(&trans);
@@ -307,20 +337,29 @@ static int test_multiple_refs(struct btrfs_root *root)
 		return ret;
 	}
 
+	ret = btrfs_find_all_roots(&trans, fs_info, 4096, 0, &old_roots);
+	if (ret) {
+		ulist_free(old_roots);
+		test_msg("Couldn't find old roots: %d\n", ret);
+		return ret;
+	}
+
 	ret = insert_normal_tree_ref(root, 4096, 4096, 0, 5);
 	if (ret)
 		return ret;
 
-	ret = btrfs_qgroup_record_ref(&trans, fs_info, 5, 4096, 4096,
-				      BTRFS_QGROUP_OPER_ADD_EXCL, 0);
+	ret = btrfs_find_all_roots(&trans, fs_info, 4096, 0, &new_roots);
 	if (ret) {
-		test_msg("Couldn't add space to a qgroup %d\n", ret);
+		ulist_free(old_roots);
+		ulist_free(new_roots);
+		test_msg("Couldn't find old roots: %d\n", ret);
 		return ret;
 	}
 
-	ret = btrfs_delayed_qgroup_accounting(&trans, fs_info);
+	ret = btrfs_qgroup_account_extent(&trans, fs_info, 4096, 4096,
+					  old_roots, new_roots);
 	if (ret) {
-		test_msg("Delayed qgroup accounting failed %d\n", ret);
+		test_msg("Couldn't account space for a qgroup %d\n", ret);
 		return ret;
 	}
 
@@ -329,20 +368,29 @@ static int test_multiple_refs(struct btrfs_root *root)
 		return -EINVAL;
 	}
 
+	ret = btrfs_find_all_roots(&trans, fs_info, 4096, 0, &old_roots);
+	if (ret) {
+		ulist_free(old_roots);
+		test_msg("Couldn't find old roots: %d\n", ret);
+		return ret;
+	}
+
 	ret = add_tree_ref(root, 4096, 4096, 0, 256);
 	if (ret)
 		return ret;
 
-	ret = btrfs_qgroup_record_ref(&trans, fs_info, 256, 4096, 4096,
-				      BTRFS_QGROUP_OPER_ADD_SHARED, 0);
+	ret = btrfs_find_all_roots(&trans, fs_info, 4096, 0, &new_roots);
 	if (ret) {
-		test_msg("Qgroup record ref failed %d\n", ret);
+		ulist_free(old_roots);
+		ulist_free(new_roots);
+		test_msg("Couldn't find old roots: %d\n", ret);
 		return ret;
 	}
 
-	ret = btrfs_delayed_qgroup_accounting(&trans, fs_info);
+	ret = btrfs_qgroup_account_extent(&trans, fs_info, 4096, 4096,
+					  old_roots, new_roots);
 	if (ret) {
-		test_msg("Qgroup accounting failed %d\n", ret);
+		test_msg("Couldn't account space for a qgroup %d\n", ret);
 		return ret;
 	}
 
@@ -356,20 +404,29 @@ static int test_multiple_refs(struct btrfs_root *root)
 		return -EINVAL;
 	}
 
+	ret = btrfs_find_all_roots(&trans, fs_info, 4096, 0, &old_roots);
+	if (ret) {
+		ulist_free(old_roots);
+		test_msg("Couldn't find old roots: %d\n", ret);
+		return ret;
+	}
+
 	ret = remove_extent_ref(root, 4096, 4096, 0, 256);
 	if (ret)
 		return ret;
 
-	ret = btrfs_qgroup_record_ref(&trans, fs_info, 256, 4096, 4096,
-				      BTRFS_QGROUP_OPER_SUB_SHARED, 0);
+	ret = btrfs_find_all_roots(&trans, fs_info, 4096, 0, &new_roots);
 	if (ret) {
-		test_msg("Qgroup record ref failed %d\n", ret);
+		ulist_free(old_roots);
+		ulist_free(new_roots);
+		test_msg("Couldn't find old roots: %d\n", ret);
 		return ret;
 	}
 
-	ret = btrfs_delayed_qgroup_accounting(&trans, fs_info);
+	ret = btrfs_qgroup_account_extent(&trans, fs_info, 4096, 4096,
+					  old_roots, new_roots);
 	if (ret) {
-		test_msg("Qgroup accounting failed %d\n", ret);
+		test_msg("Couldn't account space for a qgroup %d\n", ret);
 		return ret;
 	}
 
diff --git a/kernel/fs/btrfs/transaction.c b/kernel/fs/btrfs/transaction.c
index 00d18c2bd..be8eae80f 100644
--- a/kernel/fs/btrfs/transaction.c
+++ b/kernel/fs/btrfs/transaction.c
@@ -82,6 +82,12 @@ void btrfs_put_transaction(struct btrfs_transaction *transaction)
 static void clear_btree_io_tree(struct extent_io_tree *tree)
 {
 	spin_lock(&tree->lock);
+	/*
+	 * Do a single barrier for the waitqueue_active check here, the state
+	 * of the waitqueue should not change once clear_btree_io_tree is
+	 * called.
+	 */
+	smp_mb();
 	while (!RB_EMPTY_ROOT(&tree->state)) {
 		struct rb_node *node;
 		struct extent_state *state;
@@ -117,6 +123,18 @@ static noinline void switch_commit_roots(struct btrfs_transaction *trans,
 			btrfs_unpin_free_ino(root);
 		clear_btree_io_tree(&root->dirty_log_pages);
 	}
+
+	/* We can free old roots now. */
+	spin_lock(&trans->dropped_roots_lock);
+	while (!list_empty(&trans->dropped_roots)) {
+		root = list_first_entry(&trans->dropped_roots,
+					struct btrfs_root, root_list);
+		list_del_init(&root->root_list);
+		spin_unlock(&trans->dropped_roots_lock);
+		btrfs_drop_and_free_fs_root(fs_info, root);
+		spin_lock(&trans->dropped_roots_lock);
+	}
+	spin_unlock(&trans->dropped_roots_lock);
 	up_write(&fs_info->commit_root_sem);
 }
 
@@ -214,23 +232,22 @@ loop:
 	extwriter_counter_init(cur_trans, type);
 	init_waitqueue_head(&cur_trans->writer_wait);
 	init_waitqueue_head(&cur_trans->commit_wait);
+	init_waitqueue_head(&cur_trans->pending_wait);
 	cur_trans->state = TRANS_STATE_RUNNING;
 	/*
 	 * One for this trans handle, one so it will live on until we
 	 * commit the transaction.
 	 */
 	atomic_set(&cur_trans->use_count, 2);
-	cur_trans->have_free_bgs = 0;
+	atomic_set(&cur_trans->pending_ordered, 0);
+	cur_trans->flags = 0;
 	cur_trans->start_time = get_seconds();
-	cur_trans->dirty_bg_run = 0;
+
+	memset(&cur_trans->delayed_refs, 0, sizeof(cur_trans->delayed_refs));
 
 	cur_trans->delayed_refs.href_root = RB_ROOT;
+	cur_trans->delayed_refs.dirty_extent_root = RB_ROOT;
 	atomic_set(&cur_trans->delayed_refs.num_entries, 0);
-	cur_trans->delayed_refs.num_heads_ready = 0;
-	cur_trans->delayed_refs.pending_csums = 0;
-	cur_trans->delayed_refs.num_heads = 0;
-	cur_trans->delayed_refs.flushing = 0;
-	cur_trans->delayed_refs.run_delayed_start = 0;
 
 	/*
 	 * although the tree mod log is per file system and not per transaction,
@@ -250,12 +267,14 @@ loop:
 	INIT_LIST_HEAD(&cur_trans->pending_snapshots);
 	INIT_LIST_HEAD(&cur_trans->pending_chunks);
 	INIT_LIST_HEAD(&cur_trans->switch_commits);
-	INIT_LIST_HEAD(&cur_trans->pending_ordered);
 	INIT_LIST_HEAD(&cur_trans->dirty_bgs);
 	INIT_LIST_HEAD(&cur_trans->io_bgs);
+	INIT_LIST_HEAD(&cur_trans->dropped_roots);
 	mutex_init(&cur_trans->cache_write_mutex);
 	cur_trans->num_dirty_bgs = 0;
 	spin_lock_init(&cur_trans->dirty_bgs_lock);
+	INIT_LIST_HEAD(&cur_trans->deleted_bgs);
+	spin_lock_init(&cur_trans->dropped_roots_lock);
 	list_add_tail(&cur_trans->list, &fs_info->trans_list);
 	extent_io_tree_init(&cur_trans->dirty_pages,
 			     fs_info->btree_inode->i_mapping);
@@ -332,6 +351,24 @@ static int record_root_in_trans(struct btrfs_trans_handle *trans,
 }
 
 
+void btrfs_add_dropped_root(struct btrfs_trans_handle *trans,
+			    struct btrfs_root *root)
+{
+	struct btrfs_transaction *cur_trans = trans->transaction;
+
+	/* Add ourselves to the transaction dropped list */
+	spin_lock(&cur_trans->dropped_roots_lock);
+	list_add_tail(&root->root_list, &cur_trans->dropped_roots);
+	spin_unlock(&cur_trans->dropped_roots_lock);
+
+	/* Make sure we don't try to update the root at commit time */
+	spin_lock(&root->fs_info->fs_roots_radix_lock);
+	radix_tree_tag_clear(&root->fs_info->fs_roots_radix,
+			     (unsigned long)root->root_key.objectid,
+			     BTRFS_ROOT_TRANS_TAG);
+	spin_unlock(&root->fs_info->fs_roots_radix_lock);
+}
+
 int btrfs_record_root_in_trans(struct btrfs_trans_handle *trans,
 			       struct btrfs_root *root)
 {
@@ -411,8 +448,8 @@ static inline bool need_reserve_reloc_root(struct btrfs_root *root)
 }
 
 static struct btrfs_trans_handle *
-start_transaction(struct btrfs_root *root, u64 num_items, unsigned int type,
-		  enum btrfs_reserve_flush_enum flush)
+start_transaction(struct btrfs_root *root, unsigned int num_items,
+		  unsigned int type, enum btrfs_reserve_flush_enum flush)
 {
 	struct btrfs_trans_handle *h;
 	struct btrfs_transaction *cur_trans;
@@ -442,13 +479,10 @@ start_transaction(struct btrfs_root *root, u64 num_items, unsigned int type,
 	 * the appropriate flushing if need be.
 	 */
 	if (num_items > 0 && root != root->fs_info->chunk_root) {
-		if (root->fs_info->quota_enabled &&
-		    is_fstree(root->root_key.objectid)) {
-			qgroup_reserved = num_items * root->nodesize;
-			ret = btrfs_qgroup_reserve(root, qgroup_reserved);
-			if (ret)
-				return ERR_PTR(ret);
-		}
+		qgroup_reserved = num_items * root->nodesize;
+		ret = btrfs_qgroup_reserve_meta(root, qgroup_reserved);
+		if (ret)
+			return ERR_PTR(ret);
 
 		num_bytes = btrfs_calc_trans_metadata_size(root, num_items);
 		/*
@@ -466,7 +500,7 @@ start_transaction(struct btrfs_root *root, u64 num_items, unsigned int type,
 			goto reserve_fail;
 	}
 again:
-	h = kmem_cache_alloc(btrfs_trans_handle_cachep, GFP_NOFS);
+	h = kmem_cache_zalloc(btrfs_trans_handle_cachep, GFP_NOFS);
 	if (!h) {
 		ret = -ENOMEM;
 		goto alloc_fail;
@@ -507,24 +541,13 @@ again:
 
 	h->transid = cur_trans->transid;
 	h->transaction = cur_trans;
-	h->blocks_used = 0;
-	h->bytes_reserved = 0;
 	h->root = root;
-	h->delayed_ref_updates = 0;
 	h->use_count = 1;
-	h->adding_csums = 0;
-	h->block_rsv = NULL;
-	h->orig_rsv = NULL;
-	h->aborted = 0;
-	h->qgroup_reserved = 0;
-	h->delayed_ref_elem.seq = 0;
+
 	h->type = type;
-	h->allocating_chunk = false;
-	h->reloc_reserved = false;
-	h->sync = false;
+	h->can_flush_pending_bgs = true;
 	INIT_LIST_HEAD(&h->qgroup_ref_list);
 	INIT_LIST_HEAD(&h->new_bgs);
-	INIT_LIST_HEAD(&h->ordered);
 
 	smp_mb();
 	if (cur_trans->state >= TRANS_STATE_BLOCKED &&
@@ -541,7 +564,6 @@ again:
 		h->bytes_reserved = num_bytes;
 		h->reloc_reserved = reloc_reserved;
 	}
-	h->qgroup_reserved = qgroup_reserved;
 
 got_it:
 	btrfs_record_root_in_trans(h, root);
@@ -559,20 +581,52 @@ alloc_fail:
 		btrfs_block_rsv_release(root, &root->fs_info->trans_block_rsv,
 					num_bytes);
 reserve_fail:
-	if (qgroup_reserved)
-		btrfs_qgroup_free(root, qgroup_reserved);
+	btrfs_qgroup_free_meta(root, qgroup_reserved);
 	return ERR_PTR(ret);
 }
 
 struct btrfs_trans_handle *btrfs_start_transaction(struct btrfs_root *root,
-						   int num_items)
+						   unsigned int num_items)
 {
 	return start_transaction(root, num_items, TRANS_START,
 				 BTRFS_RESERVE_FLUSH_ALL);
 }
+struct btrfs_trans_handle *btrfs_start_transaction_fallback_global_rsv(
+					struct btrfs_root *root,
+					unsigned int num_items,
+					int min_factor)
+{
+	struct btrfs_trans_handle *trans;
+	u64 num_bytes;
+	int ret;
+
+	trans = btrfs_start_transaction(root, num_items);
+	if (!IS_ERR(trans) || PTR_ERR(trans) != -ENOSPC)
+		return trans;
+
+	trans = btrfs_start_transaction(root, 0);
+	if (IS_ERR(trans))
+		return trans;
+
+	num_bytes = btrfs_calc_trans_metadata_size(root, num_items);
+	ret = btrfs_cond_migrate_bytes(root->fs_info,
+				       &root->fs_info->trans_block_rsv,
+				       num_bytes,
+				       min_factor);
+	if (ret) {
+		btrfs_end_transaction(trans, root);
+		return ERR_PTR(ret);
+	}
+
+	trans->block_rsv = &root->fs_info->trans_block_rsv;
+	trans->bytes_reserved = num_bytes;
+
+	return trans;
+}
 
 struct btrfs_trans_handle *btrfs_start_transaction_lflush(
-					struct btrfs_root *root, int num_items)
+					struct btrfs_root *root,
+					unsigned int num_items)
 {
 	return start_transaction(root, num_items, TRANS_START,
 				 BTRFS_RESERVE_FLUSH_LIMIT);
@@ -756,12 +810,6 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans,
 	if (!list_empty(&trans->new_bgs))
 		btrfs_create_pending_block_groups(trans, root);
 
-	if (!list_empty(&trans->ordered)) {
-		spin_lock(&info->trans_lock);
-		list_splice_init(&trans->ordered, &cur_trans->pending_ordered);
-		spin_unlock(&info->trans_lock);
-	}
-
 	trans->delayed_ref_updates = 0;
 	if (!trans->sync) {
 		must_run_delayed_refs =
@@ -777,21 +825,14 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans,
 			must_run_delayed_refs = 2;
 	}
 
-	if (trans->qgroup_reserved) {
-		/*
-		 * the same root has to be passed here between start_transaction
-		 * and end_transaction. Subvolume quota depends on this.
-		 */
-		btrfs_qgroup_free(trans->root, trans->qgroup_reserved);
-		trans->qgroup_reserved = 0;
-	}
-
 	btrfs_trans_release_metadata(trans, root);
 	trans->block_rsv = NULL;
 
 	if (!list_empty(&trans->new_bgs))
 		btrfs_create_pending_block_groups(trans, root);
 
+	btrfs_trans_release_chunk_metadata(trans);
+
 	if (lock && !atomic_read(&root->fs_info->open_ioctl_trans) &&
 	    should_end_transaction(trans, root) &&
 	    ACCESS_ONCE(cur_trans->state) == TRANS_STATE_RUNNING) {
@@ -816,6 +857,9 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans,
 	atomic_dec(&cur_trans->num_writers);
 	extwriter_counter_dec(cur_trans, trans->type);
 
+	/*
+	 * Make sure counter is updated before we wake up waiters.
+	 */
 	smp_mb();
 	if (waitqueue_active(&cur_trans->writer_wait))
 		wake_up(&cur_trans->writer_wait);
@@ -1198,6 +1242,7 @@ static noinline int commit_fs_roots(struct btrfs_trans_handle *trans,
 			spin_lock(&fs_info->fs_roots_radix_lock);
 			if (err)
 				break;
+			btrfs_qgroup_free_meta_all(root);
 		}
 	}
 	spin_unlock(&fs_info->fs_roots_radix_lock);
@@ -1290,7 +1335,13 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
 	if (pending->error)
 		goto no_free_objectid;
 
-	btrfs_reloc_pre_snapshot(trans, pending, &to_reserve);
+	/*
+	 * Make qgroup to skip current new snapshot's qgroupid, as it is
+	 * accounted by later btrfs_qgroup_inherit().
+	 */
+	btrfs_set_skip_qgroup(trans, objectid);
+
+	btrfs_reloc_pre_snapshot(pending, &to_reserve);
 
 	if (to_reserve > 0) {
 		pending->error = btrfs_block_rsv_add(root,
@@ -1298,7 +1349,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
 						     to_reserve,
 						     BTRFS_RESERVE_NO_FLUSH);
 		if (pending->error)
-			goto no_free_objectid;
+			goto clear_skip_qgroup;
 	}
 
 	key.objectid = objectid;
@@ -1396,25 +1447,6 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
 		btrfs_abort_transaction(trans, root, ret);
 		goto fail;
 	}
-
-	/*
-	 * We need to flush delayed refs in order to make sure all of our quota
-	 * operations have been done before we call btrfs_qgroup_inherit.
-	 */
-	ret = btrfs_run_delayed_refs(trans, root, (unsigned long)-1);
-	if (ret) {
-		btrfs_abort_transaction(trans, root, ret);
-		goto fail;
-	}
-
-	ret = btrfs_qgroup_inherit(trans, fs_info,
-				   root->root_key.objectid,
-				   objectid, pending->inherit);
-	if (ret) {
-		btrfs_abort_transaction(trans, root, ret);
-		goto fail;
-	}
-
 	/* see comments in should_cow_block() */
 	set_bit(BTRFS_ROOT_FORCE_COW, &root->state);
 	smp_wmb();
@@ -1497,11 +1529,37 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
 			goto fail;
 		}
 	}
+
+	ret = btrfs_run_delayed_refs(trans, root, (unsigned long)-1);
+	if (ret) {
+		btrfs_abort_transaction(trans, root, ret);
+		goto fail;
+	}
+
+	/*
+	 * account qgroup counters before qgroup_inherit()
+	 */
+	ret = btrfs_qgroup_prepare_account_extents(trans, fs_info);
+	if (ret)
+		goto fail;
+	ret = btrfs_qgroup_account_extents(trans, fs_info);
+	if (ret)
+		goto fail;
+	ret = btrfs_qgroup_inherit(trans, fs_info,
+				   root->root_key.objectid,
+				   objectid, pending->inherit);
+	if (ret) {
+		btrfs_abort_transaction(trans, root, ret);
+		goto fail;
+	}
+
 fail:
 	pending->error = ret;
 dir_item_existed:
 	trans->block_rsv = rsv;
 	trans->bytes_reserved = 0;
+clear_skip_qgroup:
+	btrfs_clear_skip_qgroup(trans);
 no_free_objectid:
 	kfree(new_root_item);
 root_item_alloc_fail:
@@ -1620,9 +1678,7 @@ static void do_async_commit(struct work_struct *work)
 	 * Tell lockdep about it.
 	 */
 	if (ac->newtrans->type & __TRANS_FREEZABLE)
-		rwsem_acquire_read(
-		     &ac->root->fs_info->sb->s_writers.lock_map[SB_FREEZE_FS-1],
-		     0, 1, _THIS_IP_);
+		__sb_writers_acquired(ac->root->fs_info->sb, SB_FREEZE_FS);
 
 	current->journal_info = ac->newtrans;
 
@@ -1661,9 +1717,7 @@ int btrfs_commit_transaction_async(struct btrfs_trans_handle *trans,
 	 * async commit thread will be the one to unlock it.
 	 */
 	if (ac->newtrans->type & __TRANS_FREEZABLE)
-		rwsem_release(
-			&root->fs_info->sb->s_writers.lock_map[SB_FREEZE_FS-1],
-			1, _THIS_IP_);
+		__sb_writers_release(root->fs_info->sb, SB_FREEZE_FS);
 
 	schedule_work(&ac->work);
 
@@ -1746,25 +1800,10 @@ static inline void btrfs_wait_delalloc_flush(struct btrfs_fs_info *fs_info)
 }
 
 static inline void
-btrfs_wait_pending_ordered(struct btrfs_transaction *cur_trans,
-			   struct btrfs_fs_info *fs_info)
+btrfs_wait_pending_ordered(struct btrfs_transaction *cur_trans)
 {
-	struct btrfs_ordered_extent *ordered;
-
-	spin_lock(&fs_info->trans_lock);
-	while (!list_empty(&cur_trans->pending_ordered)) {
-		ordered = list_first_entry(&cur_trans->pending_ordered,
-					   struct btrfs_ordered_extent,
-					   trans_list);
-		list_del_init(&ordered->trans_list);
-		spin_unlock(&fs_info->trans_lock);
-
-		wait_event(ordered->wait, test_bit(BTRFS_ORDERED_COMPLETE,
-						   &ordered->flags));
-		btrfs_put_ordered_extent(ordered);
-		spin_lock(&fs_info->trans_lock);
-	}
-	spin_unlock(&fs_info->trans_lock);
+	wait_event(cur_trans->pending_wait,
+		   atomic_read(&cur_trans->pending_ordered) == 0);
 }
 
 int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
@@ -1793,10 +1832,6 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 
 	btrfs_trans_release_metadata(trans, root);
 	trans->block_rsv = NULL;
-	if (trans->qgroup_reserved) {
-		btrfs_qgroup_free(root, trans->qgroup_reserved);
-		trans->qgroup_reserved = 0;
-	}
 
 	cur_trans = trans->transaction;
 
@@ -1816,7 +1851,7 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 		return ret;
 	}
 
-	if (!cur_trans->dirty_bg_run) {
+	if (!test_bit(BTRFS_TRANS_DIRTY_BG_RUN, &cur_trans->flags)) {
 		int run_it = 0;
 
 		/* this mutex is also taken before trying to set
@@ -1825,18 +1860,17 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 		 * after a extents from that block group have been
 		 * allocated for cache files.  btrfs_set_block_group_ro
 		 * will wait for the transaction to commit if it
-		 * finds dirty_bg_run = 1
+		 * finds BTRFS_TRANS_DIRTY_BG_RUN set.
 		 *
-		 * The dirty_bg_run flag is also used to make sure only
-		 * one process starts all the block group IO.  It wouldn't
+		 * The BTRFS_TRANS_DIRTY_BG_RUN flag is also used to make sure
+		 * only one process starts all the block group IO.  It wouldn't
 		 * hurt to have more than one go through, but there's no
 		 * real advantage to it either.
 		 */
 		mutex_lock(&root->fs_info->ro_block_group_mutex);
-		if (!cur_trans->dirty_bg_run) {
+		if (!test_and_set_bit(BTRFS_TRANS_DIRTY_BG_RUN,
+				      &cur_trans->flags))
 			run_it = 1;
-			cur_trans->dirty_bg_run = 1;
-		}
 		mutex_unlock(&root->fs_info->ro_block_group_mutex);
 
 		if (run_it)
@@ -1848,7 +1882,6 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 	}
 
 	spin_lock(&root->fs_info->trans_lock);
-	list_splice_init(&trans->ordered, &cur_trans->pending_ordered);
 	if (cur_trans->state >= TRANS_STATE_COMMIT_START) {
 		spin_unlock(&root->fs_info->trans_lock);
 		atomic_inc(&cur_trans->use_count);
@@ -1907,7 +1940,7 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 
 	btrfs_wait_delalloc_flush(root->fs_info);
 
-	btrfs_wait_pending_ordered(cur_trans, root->fs_info);
+	btrfs_wait_pending_ordered(cur_trans);
 
 	btrfs_scrub_pause(root);
 	/*
@@ -1966,6 +1999,13 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 		goto scrub_continue;
 	}
 
+	/* Reocrd old roots for later qgroup accounting */
+	ret = btrfs_qgroup_prepare_account_extents(trans, root->fs_info);
+	if (ret) {
+		mutex_unlock(&root->fs_info->reloc_mutex);
+		goto scrub_continue;
+	}
+
 	/*
 	 * make sure none of the code above managed to slip in a
 	 * delayed item
@@ -2007,6 +2047,17 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 	 */
 	btrfs_free_log_root_tree(trans, root->fs_info);
 
+	/*
+	 * Since fs roots are all committed, we can get a quite accurate
+	 * new_roots. So let's do quota accounting.
+	 */
+	ret = btrfs_qgroup_account_extents(trans, root->fs_info);
+	if (ret < 0) {
+		mutex_unlock(&root->fs_info->tree_log_mutex);
+		mutex_unlock(&root->fs_info->reloc_mutex);
+		goto scrub_continue;
+	}
+
 	ret = commit_cowonly_roots(trans, root);
 	if (ret) {
 		mutex_unlock(&root->fs_info->tree_log_mutex);
@@ -2057,6 +2108,8 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 	clear_bit(BTRFS_INODE_BTREE_LOG1_ERR, &btree_ino->runtime_flags);
 	clear_bit(BTRFS_INODE_BTREE_LOG2_ERR, &btree_ino->runtime_flags);
 
+	btrfs_trans_release_chunk_metadata(trans);
+
 	spin_lock(&root->fs_info->trans_lock);
 	cur_trans->state = TRANS_STATE_UNBLOCKED;
 	root->fs_info->running_transaction = NULL;
@@ -2067,7 +2120,7 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 
 	ret = btrfs_write_and_wait_transaction(trans, root);
 	if (ret) {
-		btrfs_error(root->fs_info, ret,
+		btrfs_std_error(root->fs_info, ret,
 			    "Error while writing out transaction");
 		mutex_unlock(&root->fs_info->tree_log_mutex);
 		goto scrub_continue;
@@ -2087,7 +2140,7 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 
 	btrfs_finish_extent_commit(trans, root);
 
-	if (cur_trans->have_free_bgs)
+	if (test_bit(BTRFS_TRANS_HAVE_FREE_BGS, &cur_trans->flags))
 		btrfs_clear_space_info_full(root->fs_info);
 
 	root->fs_info->last_trans_committed = cur_trans->transid;
@@ -2117,7 +2170,8 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 
 	kmem_cache_free(btrfs_trans_handle_cachep, trans);
 
-	if (current != root->fs_info->transaction_kthread)
+	if (current != root->fs_info->transaction_kthread &&
+	    current != root->fs_info->cleaner_kthread)
 		btrfs_run_delayed_iputs(root);
 
 	return ret;
@@ -2126,11 +2180,8 @@ scrub_continue:
 	btrfs_scrub_continue(root);
 cleanup_transaction:
 	btrfs_trans_release_metadata(trans, root);
+	btrfs_trans_release_chunk_metadata(trans);
 	trans->block_rsv = NULL;
-	if (trans->qgroup_reserved) {
-		btrfs_qgroup_free(root, trans->qgroup_reserved);
-		trans->qgroup_reserved = 0;
-	}
 	btrfs_warn(root->fs_info, "Skipping commit of aborted transaction.");
 	if (current->journal_info == trans)
 		current->journal_info = NULL;
diff --git a/kernel/fs/btrfs/transaction.h b/kernel/fs/btrfs/transaction.h
index 0b2475559..64c8221b6 100644
--- a/kernel/fs/btrfs/transaction.h
+++ b/kernel/fs/btrfs/transaction.h
@@ -32,6 +32,10 @@ enum btrfs_trans_state {
 	TRANS_STATE_MAX			= 6,
 };
 
+#define BTRFS_TRANS_HAVE_FREE_BGS	0
+#define BTRFS_TRANS_DIRTY_BG_RUN	1
+#define BTRFS_TRANS_CACHE_ENOSPC	2
+
 struct btrfs_transaction {
 	u64 transid;
 	/*
@@ -46,11 +50,9 @@ struct btrfs_transaction {
 	 */
 	atomic_t num_writers;
 	atomic_t use_count;
+	atomic_t pending_ordered;
 
-	/*
-	 * true if there is free bgs operations in this transaction
-	 */
-	int have_free_bgs;
+	unsigned long flags;
 
 	/* Be protected by fs_info->trans_lock when we want to change it. */
 	enum btrfs_trans_state state;
@@ -59,12 +61,13 @@ struct btrfs_transaction {
 	unsigned long start_time;
 	wait_queue_head_t writer_wait;
 	wait_queue_head_t commit_wait;
+	wait_queue_head_t pending_wait;
 	struct list_head pending_snapshots;
 	struct list_head pending_chunks;
-	struct list_head pending_ordered;
 	struct list_head switch_commits;
 	struct list_head dirty_bgs;
 	struct list_head io_bgs;
+	struct list_head dropped_roots;
 	u64 num_dirty_bgs;
 
 	/*
@@ -74,9 +77,11 @@ struct btrfs_transaction {
 	 */
 	struct mutex cache_write_mutex;
 	spinlock_t dirty_bgs_lock;
+	/* Protected by spin lock fs_info->unused_bgs_lock. */
+	struct list_head deleted_bgs;
+	spinlock_t dropped_roots_lock;
 	struct btrfs_delayed_ref_root delayed_refs;
 	int aborted;
-	int dirty_bg_run;
 };
 
 #define __TRANS_FREEZABLE	(1U << 0)
@@ -102,7 +107,7 @@ struct btrfs_transaction {
 struct btrfs_trans_handle {
 	u64 transid;
 	u64 bytes_reserved;
-	u64 qgroup_reserved;
+	u64 chunk_bytes_reserved;
 	unsigned long use_count;
 	unsigned long blocks_reserved;
 	unsigned long blocks_used;
@@ -113,6 +118,7 @@ struct btrfs_trans_handle {
 	short aborted;
 	short adding_csums;
 	bool allocating_chunk;
+	bool can_flush_pending_bgs;
 	bool reloc_reserved;
 	bool sync;
 	unsigned int type;
@@ -123,7 +129,6 @@ struct btrfs_trans_handle {
 	 */
 	struct btrfs_root *root;
 	struct seq_list delayed_ref_elem;
-	struct list_head ordered;
 	struct list_head qgroup_ref_list;
 	struct list_head new_bgs;
 };
@@ -153,12 +158,40 @@ static inline void btrfs_set_inode_last_trans(struct btrfs_trans_handle *trans,
 	spin_unlock(&BTRFS_I(inode)->lock);
 }
 
+/*
+ * Make qgroup codes to skip given qgroupid, means the old/new_roots for
+ * qgroup won't contain the qgroupid in it.
+ */
+static inline void btrfs_set_skip_qgroup(struct btrfs_trans_handle *trans,
+					 u64 qgroupid)
+{
+	struct btrfs_delayed_ref_root *delayed_refs;
+
+	delayed_refs = &trans->transaction->delayed_refs;
+	WARN_ON(delayed_refs->qgroup_to_skip);
+	delayed_refs->qgroup_to_skip = qgroupid;
+}
+
+static inline void btrfs_clear_skip_qgroup(struct btrfs_trans_handle *trans)
+{
+	struct btrfs_delayed_ref_root *delayed_refs;
+
+	delayed_refs = &trans->transaction->delayed_refs;
+	WARN_ON(!delayed_refs->qgroup_to_skip);
+	delayed_refs->qgroup_to_skip = 0;
+}
+
 int btrfs_end_transaction(struct btrfs_trans_handle *trans,
 			  struct btrfs_root *root);
 struct btrfs_trans_handle *btrfs_start_transaction(struct btrfs_root *root,
-						   int num_items);
+						   unsigned int num_items);
+struct btrfs_trans_handle *btrfs_start_transaction_fallback_global_rsv(
+					struct btrfs_root *root,
+					unsigned int num_items,
+					int min_factor);
 struct btrfs_trans_handle *btrfs_start_transaction_lflush(
-					struct btrfs_root *root, int num_items);
+					struct btrfs_root *root,
+					unsigned int num_items);
 struct btrfs_trans_handle *btrfs_join_transaction(struct btrfs_root *root);
 struct btrfs_trans_handle *btrfs_join_transaction_nolock(struct btrfs_root *root);
 struct btrfs_trans_handle *btrfs_attach_transaction(struct btrfs_root *root);
@@ -190,5 +223,6 @@ int btrfs_transaction_blocked(struct btrfs_fs_info *info);
 int btrfs_transaction_in_commit(struct btrfs_fs_info *info);
 void btrfs_put_transaction(struct btrfs_transaction *transaction);
 void btrfs_apply_pending_changes(struct btrfs_fs_info *fs_info);
-
+void btrfs_add_dropped_root(struct btrfs_trans_handle *trans,
+			    struct btrfs_root *root);
 #endif
diff --git a/kernel/fs/btrfs/tree-defrag.c b/kernel/fs/btrfs/tree-defrag.c
index a63719cc9..f31db4325 100644
--- a/kernel/fs/btrfs/tree-defrag.c
+++ b/kernel/fs/btrfs/tree-defrag.c
@@ -52,9 +52,6 @@ int btrfs_defrag_leaves(struct btrfs_trans_handle *trans,
 	if (!test_bit(BTRFS_ROOT_REF_COWS, &root->state))
 		goto out;
 
-	if (btrfs_test_opt(root, SSD))
-		goto out;
-
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
@@ -118,8 +115,7 @@ int btrfs_defrag_leaves(struct btrfs_trans_handle *trans,
 		ret = -EAGAIN;
 	}
 out:
-	if (path)
-		btrfs_free_path(path);
+	btrfs_free_path(path);
 	if (ret == -EAGAIN) {
 		if (root->defrag_max.objectid > root->defrag_progress.objectid)
 			goto done;
diff --git a/kernel/fs/btrfs/tree-log.c b/kernel/fs/btrfs/tree-log.c
index 4920fceff..323e12cc9 100644
--- a/kernel/fs/btrfs/tree-log.c
+++ b/kernel/fs/btrfs/tree-log.c
@@ -140,55 +140,46 @@ static int start_log_trans(struct btrfs_trans_handle *trans,
 			   struct btrfs_root *root,
 			   struct btrfs_log_ctx *ctx)
 {
-	int index;
-	int ret;
+	int ret = 0;
 
 	mutex_lock(&root->log_mutex);
+
 	if (root->log_root) {
 		if (btrfs_need_log_full_commit(root->fs_info, trans)) {
 			ret = -EAGAIN;
 			goto out;
 		}
+
 		if (!root->log_start_pid) {
-			root->log_start_pid = current->pid;
 			clear_bit(BTRFS_ROOT_MULTI_LOG_TASKS, &root->state);
+			root->log_start_pid = current->pid;
 		} else if (root->log_start_pid != current->pid) {
 			set_bit(BTRFS_ROOT_MULTI_LOG_TASKS, &root->state);
 		}
+	} else {
+		mutex_lock(&root->fs_info->tree_log_mutex);
+		if (!root->fs_info->log_root_tree)
+			ret = btrfs_init_log_root_tree(trans, root->fs_info);
+		mutex_unlock(&root->fs_info->tree_log_mutex);
+		if (ret)
+			goto out;
 
-		atomic_inc(&root->log_batch);
-		atomic_inc(&root->log_writers);
-		if (ctx) {
-			index = root->log_transid % 2;
-			list_add_tail(&ctx->list, &root->log_ctxs[index]);
-			ctx->log_transid = root->log_transid;
-		}
-		mutex_unlock(&root->log_mutex);
-		return 0;
-	}
-
-	ret = 0;
-	mutex_lock(&root->fs_info->tree_log_mutex);
-	if (!root->fs_info->log_root_tree)
-		ret = btrfs_init_log_root_tree(trans, root->fs_info);
-	mutex_unlock(&root->fs_info->tree_log_mutex);
-	if (ret)
-		goto out;
-
-	if (!root->log_root) {
 		ret = btrfs_add_log_tree(trans, root);
 		if (ret)
 			goto out;
+
+		clear_bit(BTRFS_ROOT_MULTI_LOG_TASKS, &root->state);
+		root->log_start_pid = current->pid;
 	}
-	clear_bit(BTRFS_ROOT_MULTI_LOG_TASKS, &root->state);
-	root->log_start_pid = current->pid;
+
 	atomic_inc(&root->log_batch);
 	atomic_inc(&root->log_writers);
 	if (ctx) {
-		index = root->log_transid % 2;
+		int index = root->log_transid % 2;
 		list_add_tail(&ctx->list, &root->log_ctxs[index]);
 		ctx->log_transid = root->log_transid;
 	}
+
 out:
 	mutex_unlock(&root->log_mutex);
 	return ret;
@@ -238,7 +229,9 @@ int btrfs_pin_log_trans(struct btrfs_root *root)
 void btrfs_end_log_trans(struct btrfs_root *root)
 {
 	if (atomic_dec_and_test(&root->log_writers)) {
-		smp_mb();
+		/*
+		 * Implicit memory barrier after atomic_dec_and_test
+		 */
 		if (waitqueue_active(&root->log_writer_wait))
 			wake_up(&root->log_writer_wait);
 	}
@@ -700,7 +693,7 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans,
 				ret = btrfs_inc_extent_ref(trans, root,
 						ins.objectid, ins.offset,
 						0, root->root_key.objectid,
-						key->objectid, offset, 0);
+						key->objectid, offset);
 				if (ret)
 					goto out;
 			} else {
@@ -731,12 +724,66 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans,
 						&ordered_sums, 0);
 			if (ret)
 				goto out;
+			/*
+			 * Now delete all existing cums in the csum root that
+			 * cover our range. We do this because we can have an
+			 * extent that is completely referenced by one file
+			 * extent item and partially referenced by another
+			 * file extent item (like after using the clone or
+			 * extent_same ioctls). In this case if we end up doing
+			 * the replay of the one that partially references the
+			 * extent first, and we do not do the csum deletion
+			 * below, we can get 2 csum items in the csum tree that
+			 * overlap each other. For example, imagine our log has
+			 * the two following file extent items:
+			 *
+			 * key (257 EXTENT_DATA 409600)
+			 *     extent data disk byte 12845056 nr 102400
+			 *     extent data offset 20480 nr 20480 ram 102400
+			 *
+			 * key (257 EXTENT_DATA 819200)
+			 *     extent data disk byte 12845056 nr 102400
+			 *     extent data offset 0 nr 102400 ram 102400
+			 *
+			 * Where the second one fully references the 100K extent
+			 * that starts at disk byte 12845056, and the log tree
+			 * has a single csum item that covers the entire range
+			 * of the extent:
+			 *
+			 * key (EXTENT_CSUM EXTENT_CSUM 12845056) itemsize 100
+			 *
+			 * After the first file extent item is replayed, the
+			 * csum tree gets the following csum item:
+			 *
+			 * key (EXTENT_CSUM EXTENT_CSUM 12865536) itemsize 20
+			 *
+			 * Which covers the 20K sub-range starting at offset 20K
+			 * of our extent. Now when we replay the second file
+			 * extent item, if we do not delete existing csum items
+			 * that cover any of its blocks, we end up getting two
+			 * csum items in our csum tree that overlap each other:
+			 *
+			 * key (EXTENT_CSUM EXTENT_CSUM 12845056) itemsize 100
+			 * key (EXTENT_CSUM EXTENT_CSUM 12865536) itemsize 20
+			 *
+			 * Which is a problem, because after this anyone trying
+			 * to lookup up for the checksum of any block of our
+			 * extent starting at an offset of 40K or higher, will
+			 * end up looking at the second csum item only, which
+			 * does not contain the checksum for any block starting
+			 * at offset 40K or higher of our extent.
+			 */
 			while (!list_empty(&ordered_sums)) {
 				struct btrfs_ordered_sum *sums;
 				sums = list_entry(ordered_sums.next,
 						struct btrfs_ordered_sum,
 						list);
 				if (!ret)
+					ret = btrfs_del_csums(trans,
+						      root->fs_info->csum_root,
+						      sums->bytenr,
+						      sums->len);
+				if (!ret)
 					ret = btrfs_csum_file_blocks(trans,
 						root->fs_info->csum_root,
 						sums);
@@ -1549,9 +1596,8 @@ static noinline int link_to_fixup_dir(struct btrfs_trans_handle *trans,
  */
 static noinline int insert_one_name(struct btrfs_trans_handle *trans,
 				    struct btrfs_root *root,
-				    struct btrfs_path *path,
 				    u64 dirid, u64 index,
-				    char *name, int name_len, u8 type,
+				    char *name, int name_len,
 				    struct btrfs_key *location)
 {
 	struct inode *inode;
@@ -1613,6 +1659,9 @@ static bool name_in_log_ref(struct btrfs_root *log_root,
  * not exist in the FS, it is skipped.  fsyncs on directories
  * do not force down inodes inside that directory, just changes to the
  * names or unlinks in a directory.
+ *
+ * Returns < 0 on error, 0 if the name wasn't replayed (dentry points to a
+ * non-existing inode) and 1 if the name was replayed.
  */
 static noinline int replay_one_name(struct btrfs_trans_handle *trans,
 				    struct btrfs_root *root,
@@ -1631,6 +1680,7 @@ static noinline int replay_one_name(struct btrfs_trans_handle *trans,
 	int exists;
 	int ret = 0;
 	bool update_size = (key->type == BTRFS_DIR_INDEX_KEY);
+	bool name_added = false;
 
 	dir = read_one_inode(root, key->objectid);
 	if (!dir)
@@ -1708,6 +1758,8 @@ out:
 	}
 	kfree(name);
 	iput(dir);
+	if (!ret && name_added)
+		ret = 1;
 	return ret;
 
 insert:
@@ -1719,10 +1771,12 @@ insert:
 		goto out;
 	}
 	btrfs_release_path(path);
-	ret = insert_one_name(trans, root, path, key->objectid, key->offset,
-			      name, name_len, log_type, &log_key);
+	ret = insert_one_name(trans, root, key->objectid, key->offset,
+			      name, name_len, &log_key);
 	if (ret && ret != -ENOENT && ret != -EEXIST)
 		goto out;
+	if (!ret)
+		name_added = true;
 	update_size = false;
 	ret = 0;
 	goto out;
@@ -1740,12 +1794,13 @@ static noinline int replay_one_dir_item(struct btrfs_trans_handle *trans,
 					struct extent_buffer *eb, int slot,
 					struct btrfs_key *key)
 {
-	int ret;
+	int ret = 0;
 	u32 item_size = btrfs_item_size_nr(eb, slot);
 	struct btrfs_dir_item *di;
 	int name_len;
 	unsigned long ptr;
 	unsigned long ptr_end;
+	struct btrfs_path *fixup_path = NULL;
 
 	ptr = btrfs_item_ptr_offset(eb, slot);
 	ptr_end = ptr + item_size;
@@ -1755,12 +1810,59 @@ static noinline int replay_one_dir_item(struct btrfs_trans_handle *trans,
 			return -EIO;
 		name_len = btrfs_dir_name_len(eb, di);
 		ret = replay_one_name(trans, root, path, eb, di, key);
-		if (ret)
-			return ret;
+		if (ret < 0)
+			break;
 		ptr = (unsigned long)(di + 1);
 		ptr += name_len;
+
+		/*
+		 * If this entry refers to a non-directory (directories can not
+		 * have a link count > 1) and it was added in the transaction
+		 * that was not committed, make sure we fixup the link count of
+		 * the inode it the entry points to. Otherwise something like
+		 * the following would result in a directory pointing to an
+		 * inode with a wrong link that does not account for this dir
+		 * entry:
+		 *
+		 * mkdir testdir
+		 * touch testdir/foo
+		 * touch testdir/bar
+		 * sync
+		 *
+		 * ln testdir/bar testdir/bar_link
+		 * ln testdir/foo testdir/foo_link
+		 * xfs_io -c "fsync" testdir/bar
+		 *
+		 * <power failure>
+		 *
+		 * mount fs, log replay happens
+		 *
+		 * File foo would remain with a link count of 1 when it has two
+		 * entries pointing to it in the directory testdir. This would
+		 * make it impossible to ever delete the parent directory has
+		 * it would result in stale dentries that can never be deleted.
+		 */
+		if (ret == 1 && btrfs_dir_type(eb, di) != BTRFS_FT_DIR) {
+			struct btrfs_key di_key;
+
+			if (!fixup_path) {
+				fixup_path = btrfs_alloc_path();
+				if (!fixup_path) {
+					ret = -ENOMEM;
+					break;
+				}
+			}
+
+			btrfs_dir_item_key_to_cpu(eb, di, &di_key);
+			ret = link_to_fixup_dir(trans, root, fixup_path,
+						di_key.objectid);
+			if (ret)
+				break;
+		}
+		ret = 0;
 	}
-	return 0;
+	btrfs_free_path(fixup_path);
+	return ret;
 }
 
 /*
@@ -2535,8 +2637,7 @@ static int update_log_root(struct btrfs_trans_handle *trans,
 	return ret;
 }
 
-static void wait_log_commit(struct btrfs_trans_handle *trans,
-			    struct btrfs_root *root, int transid)
+static void wait_log_commit(struct btrfs_root *root, int transid)
 {
 	DEFINE_WAIT(wait);
 	int index = transid % 2;
@@ -2561,8 +2662,7 @@ static void wait_log_commit(struct btrfs_trans_handle *trans,
 		 atomic_read(&root->log_commit[index]));
 }
 
-static void wait_for_writer(struct btrfs_trans_handle *trans,
-			    struct btrfs_root *root)
+static void wait_for_writer(struct btrfs_root *root)
 {
 	DEFINE_WAIT(wait);
 
@@ -2642,7 +2742,7 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans,
 
 	index1 = log_transid % 2;
 	if (atomic_read(&root->log_commit[index1])) {
-		wait_log_commit(trans, root, log_transid);
+		wait_log_commit(root, log_transid);
 		mutex_unlock(&root->log_mutex);
 		return ctx->log_ret;
 	}
@@ -2651,7 +2751,7 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans,
 
 	/* wait for previous tree log sync to complete */
 	if (atomic_read(&root->log_commit[(index1 + 1) % 2]))
-		wait_log_commit(trans, root, log_transid - 1);
+		wait_log_commit(root, log_transid - 1);
 
 	while (1) {
 		int batch = atomic_read(&root->log_batch);
@@ -2662,7 +2762,7 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans,
 			schedule_timeout_uninterruptible(1);
 			mutex_lock(&root->log_mutex);
 		}
-		wait_for_writer(trans, root);
+		wait_for_writer(root);
 		if (batch == atomic_read(&root->log_batch))
 			break;
 	}
@@ -2722,7 +2822,9 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans,
 
 	mutex_lock(&log_root_tree->log_mutex);
 	if (atomic_dec_and_test(&log_root_tree->log_writers)) {
-		smp_mb();
+		/*
+		 * Implicit memory barrier after atomic_dec_and_test
+		 */
 		if (waitqueue_active(&log_root_tree->log_writer_wait))
 			wake_up(&log_root_tree->log_writer_wait);
 	}
@@ -2759,7 +2861,7 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans,
 		ret = btrfs_wait_marked_extents(log, &log->dirty_log_pages,
 						mark);
 		btrfs_wait_logged_extents(trans, log, log_transid);
-		wait_log_commit(trans, log_root_tree,
+		wait_log_commit(log_root_tree,
 				root_log_ctx.log_transid);
 		mutex_unlock(&log_root_tree->log_mutex);
 		if (!ret)
@@ -2770,11 +2872,11 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans,
 	atomic_set(&log_root_tree->log_commit[index2], 1);
 
 	if (atomic_read(&log_root_tree->log_commit[(index2 + 1) % 2])) {
-		wait_log_commit(trans, log_root_tree,
+		wait_log_commit(log_root_tree,
 				root_log_ctx.log_transid - 1);
 	}
 
-	wait_for_writer(trans, log_root_tree);
+	wait_for_writer(log_root_tree);
 
 	/*
 	 * now that we've moved on to the tree of log tree roots,
@@ -2852,6 +2954,9 @@ out_wake_log_root:
 	atomic_set(&log_root_tree->log_commit[index2], 0);
 	mutex_unlock(&log_root_tree->log_mutex);
 
+	/*
+	 * The barrier before waitqueue_active is implied by mutex_unlock
+	 */
 	if (waitqueue_active(&log_root_tree->log_commit_wait[index2]))
 		wake_up(&log_root_tree->log_commit_wait[index2]);
 out:
@@ -2863,6 +2968,9 @@ out:
 	atomic_set(&root->log_commit[index1], 0);
 	mutex_unlock(&root->log_mutex);
 
+	/*
+	 * The barrier before waitqueue_active is implied by mutex_unlock
+	 */
 	if (waitqueue_active(&root->log_commit_wait[index1]))
 		wake_up(&root->log_commit_wait[index1]);
 	return ret;
@@ -3881,12 +3989,6 @@ static int wait_ordered_extents(struct btrfs_trans_handle *trans,
 				     &ordered->flags))
 			continue;
 
-		if (ordered->csum_bytes_left) {
-			btrfs_start_ordered_extent(inode, ordered, 0);
-			wait_event(ordered->wait,
-				   ordered->csum_bytes_left == 0);
-		}
-
 		list_for_each_entry(sum, &ordered->list, list) {
 			ret = btrfs_csum_file_blocks(trans, log, sum);
 			if (ret)
@@ -4123,6 +4225,187 @@ static int logged_inode_size(struct btrfs_root *log, struct inode *inode,
 	return 0;
 }
 
+/*
+ * At the moment we always log all xattrs. This is to figure out at log replay
+ * time which xattrs must have their deletion replayed. If a xattr is missing
+ * in the log tree and exists in the fs/subvol tree, we delete it. This is
+ * because if a xattr is deleted, the inode is fsynced and a power failure
+ * happens, causing the log to be replayed the next time the fs is mounted,
+ * we want the xattr to not exist anymore (same behaviour as other filesystems
+ * with a journal, ext3/4, xfs, f2fs, etc).
+ */
+static int btrfs_log_all_xattrs(struct btrfs_trans_handle *trans,
+				struct btrfs_root *root,
+				struct inode *inode,
+				struct btrfs_path *path,
+				struct btrfs_path *dst_path)
+{
+	int ret;
+	struct btrfs_key key;
+	const u64 ino = btrfs_ino(inode);
+	int ins_nr = 0;
+	int start_slot = 0;
+
+	key.objectid = ino;
+	key.type = BTRFS_XATTR_ITEM_KEY;
+	key.offset = 0;
+
+	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+	if (ret < 0)
+		return ret;
+
+	while (true) {
+		int slot = path->slots[0];
+		struct extent_buffer *leaf = path->nodes[0];
+		int nritems = btrfs_header_nritems(leaf);
+
+		if (slot >= nritems) {
+			if (ins_nr > 0) {
+				u64 last_extent = 0;
+
+				ret = copy_items(trans, inode, dst_path, path,
+						 &last_extent, start_slot,
+						 ins_nr, 1, 0);
+				/* can't be 1, extent items aren't processed */
+				ASSERT(ret <= 0);
+				if (ret < 0)
+					return ret;
+				ins_nr = 0;
+			}
+			ret = btrfs_next_leaf(root, path);
+			if (ret < 0)
+				return ret;
+			else if (ret > 0)
+				break;
+			continue;
+		}
+
+		btrfs_item_key_to_cpu(leaf, &key, slot);
+		if (key.objectid != ino || key.type != BTRFS_XATTR_ITEM_KEY)
+			break;
+
+		if (ins_nr == 0)
+			start_slot = slot;
+		ins_nr++;
+		path->slots[0]++;
+		cond_resched();
+	}
+	if (ins_nr > 0) {
+		u64 last_extent = 0;
+
+		ret = copy_items(trans, inode, dst_path, path,
+				 &last_extent, start_slot,
+				 ins_nr, 1, 0);
+		/* can't be 1, extent items aren't processed */
+		ASSERT(ret <= 0);
+		if (ret < 0)
+			return ret;
+	}
+
+	return 0;
+}
+
+/*
+ * If the no holes feature is enabled we need to make sure any hole between the
+ * last extent and the i_size of our inode is explicitly marked in the log. This
+ * is to make sure that doing something like:
+ *
+ *      1) create file with 128Kb of data
+ *      2) truncate file to 64Kb
+ *      3) truncate file to 256Kb
+ *      4) fsync file
+ *      5) <crash/power failure>
+ *      6) mount fs and trigger log replay
+ *
+ * Will give us a file with a size of 256Kb, the first 64Kb of data match what
+ * the file had in its first 64Kb of data at step 1 and the last 192Kb of the
+ * file correspond to a hole. The presence of explicit holes in a log tree is
+ * what guarantees that log replay will remove/adjust file extent items in the
+ * fs/subvol tree.
+ *
+ * Here we do not need to care about holes between extents, that is already done
+ * by copy_items(). We also only need to do this in the full sync path, where we
+ * lookup for extents from the fs/subvol tree only. In the fast path case, we
+ * lookup the list of modified extent maps and if any represents a hole, we
+ * insert a corresponding extent representing a hole in the log tree.
+ */
+static int btrfs_log_trailing_hole(struct btrfs_trans_handle *trans,
+				   struct btrfs_root *root,
+				   struct inode *inode,
+				   struct btrfs_path *path)
+{
+	int ret;
+	struct btrfs_key key;
+	u64 hole_start;
+	u64 hole_size;
+	struct extent_buffer *leaf;
+	struct btrfs_root *log = root->log_root;
+	const u64 ino = btrfs_ino(inode);
+	const u64 i_size = i_size_read(inode);
+
+	if (!btrfs_fs_incompat(root->fs_info, NO_HOLES))
+		return 0;
+
+	key.objectid = ino;
+	key.type = BTRFS_EXTENT_DATA_KEY;
+	key.offset = (u64)-1;
+
+	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+	ASSERT(ret != 0);
+	if (ret < 0)
+		return ret;
+
+	ASSERT(path->slots[0] > 0);
+	path->slots[0]--;
+	leaf = path->nodes[0];
+	btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
+
+	if (key.objectid != ino || key.type != BTRFS_EXTENT_DATA_KEY) {
+		/* inode does not have any extents */
+		hole_start = 0;
+		hole_size = i_size;
+	} else {
+		struct btrfs_file_extent_item *extent;
+		u64 len;
+
+		/*
+		 * If there's an extent beyond i_size, an explicit hole was
+		 * already inserted by copy_items().
+		 */
+		if (key.offset >= i_size)
+			return 0;
+
+		extent = btrfs_item_ptr(leaf, path->slots[0],
+					struct btrfs_file_extent_item);
+
+		if (btrfs_file_extent_type(leaf, extent) ==
+		    BTRFS_FILE_EXTENT_INLINE) {
+			len = btrfs_file_extent_inline_len(leaf,
+							   path->slots[0],
+							   extent);
+			ASSERT(len == i_size);
+			return 0;
+		}
+
+		len = btrfs_file_extent_num_bytes(leaf, extent);
+		/* Last extent goes beyond i_size, no need to log a hole. */
+		if (key.offset + len > i_size)
+			return 0;
+		hole_start = key.offset + len;
+		hole_size = i_size - hole_start;
+	}
+	btrfs_release_path(path);
+
+	/* Last extent ends at i_size. */
+	if (hole_size == 0)
+		return 0;
+
+	hole_size = ALIGN(hole_size, root->sectorsize);
+	ret = btrfs_insert_file_extent(trans, log, ino, hole_start, 0, 0,
+				       hole_size, 0, hole_size, 0, 0, 0);
+	return ret;
+}
+
 /* log a single inode in the tree log.
  * At least one parent directory for this inode must exist in the tree
  * or be logged already.
@@ -4295,6 +4578,25 @@ again:
 		if (min_key.type == BTRFS_INODE_ITEM_KEY)
 			need_log_inode_item = false;
 
+		/* Skip xattrs, we log them later with btrfs_log_all_xattrs() */
+		if (min_key.type == BTRFS_XATTR_ITEM_KEY) {
+			if (ins_nr == 0)
+				goto next_slot;
+			ret = copy_items(trans, inode, dst_path, path,
+					 &last_extent, ins_start_slot,
+					 ins_nr, inode_only, logged_isize);
+			if (ret < 0) {
+				err = ret;
+				goto out_unlock;
+			}
+			ins_nr = 0;
+			if (ret) {
+				btrfs_release_path(path);
+				continue;
+			}
+			goto next_slot;
+		}
+
 		src = path->nodes[0];
 		if (ins_nr && ins_start_slot + ins_nr == path->slots[0]) {
 			ins_nr++;
@@ -4362,6 +4664,18 @@ next_slot:
 		ins_nr = 0;
 	}
 
+	btrfs_release_path(path);
+	btrfs_release_path(dst_path);
+	err = btrfs_log_all_xattrs(trans, root, inode, path, dst_path);
+	if (err)
+		goto out_unlock;
+	if (max_key.type >= BTRFS_EXTENT_DATA_KEY && !fast_search) {
+		btrfs_release_path(path);
+		btrfs_release_path(dst_path);
+		err = btrfs_log_trailing_hole(trans, root, inode, path);
+		if (err)
+			goto out_unlock;
+	}
 log_extents:
 	btrfs_release_path(path);
 	btrfs_release_path(dst_path);
@@ -4698,6 +5012,94 @@ next_dir_inode:
 	return ret;
 }
 
+static int btrfs_log_all_parents(struct btrfs_trans_handle *trans,
+				 struct inode *inode,
+				 struct btrfs_log_ctx *ctx)
+{
+	int ret;
+	struct btrfs_path *path;
+	struct btrfs_key key;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	const u64 ino = btrfs_ino(inode);
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+	path->skip_locking = 1;
+	path->search_commit_root = 1;
+
+	key.objectid = ino;
+	key.type = BTRFS_INODE_REF_KEY;
+	key.offset = 0;
+	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+	if (ret < 0)
+		goto out;
+
+	while (true) {
+		struct extent_buffer *leaf = path->nodes[0];
+		int slot = path->slots[0];
+		u32 cur_offset = 0;
+		u32 item_size;
+		unsigned long ptr;
+
+		if (slot >= btrfs_header_nritems(leaf)) {
+			ret = btrfs_next_leaf(root, path);
+			if (ret < 0)
+				goto out;
+			else if (ret > 0)
+				break;
+			continue;
+		}
+
+		btrfs_item_key_to_cpu(leaf, &key, slot);
+		/* BTRFS_INODE_EXTREF_KEY is BTRFS_INODE_REF_KEY + 1 */
+		if (key.objectid != ino || key.type > BTRFS_INODE_EXTREF_KEY)
+			break;
+
+		item_size = btrfs_item_size_nr(leaf, slot);
+		ptr = btrfs_item_ptr_offset(leaf, slot);
+		while (cur_offset < item_size) {
+			struct btrfs_key inode_key;
+			struct inode *dir_inode;
+
+			inode_key.type = BTRFS_INODE_ITEM_KEY;
+			inode_key.offset = 0;
+
+			if (key.type == BTRFS_INODE_EXTREF_KEY) {
+				struct btrfs_inode_extref *extref;
+
+				extref = (struct btrfs_inode_extref *)
+					(ptr + cur_offset);
+				inode_key.objectid = btrfs_inode_extref_parent(
+					leaf, extref);
+				cur_offset += sizeof(*extref);
+				cur_offset += btrfs_inode_extref_name_len(leaf,
+					extref);
+			} else {
+				inode_key.objectid = key.offset;
+				cur_offset = item_size;
+			}
+
+			dir_inode = btrfs_iget(root->fs_info->sb, &inode_key,
+					       root, NULL);
+			/* If parent inode was deleted, skip it. */
+			if (IS_ERR(dir_inode))
+				continue;
+
+			ret = btrfs_log_inode(trans, root, dir_inode,
+					      LOG_INODE_ALL, 0, LLONG_MAX, ctx);
+			iput(dir_inode);
+			if (ret)
+				goto out;
+		}
+		path->slots[0]++;
+	}
+	ret = 0;
+out:
+	btrfs_free_path(path);
+	return ret;
+}
+
 /*
  * helper function around btrfs_log_inode to make sure newly created
  * parent directories also end up in the log.  A minimal inode and backref
@@ -4717,9 +5119,6 @@ static int btrfs_log_inode_parent(struct btrfs_trans_handle *trans,
 	struct dentry *old_parent = NULL;
 	int ret = 0;
 	u64 last_committed = root->fs_info->last_trans_committed;
-	const struct dentry * const first_parent = parent;
-	const bool did_unlink = (BTRFS_I(inode)->last_unlink_trans >
-				 last_committed);
 	bool log_dentries = false;
 	struct inode *orig_inode = inode;
 
@@ -4780,6 +5179,53 @@ static int btrfs_log_inode_parent(struct btrfs_trans_handle *trans,
 	if (S_ISDIR(inode->i_mode) && ctx && ctx->log_new_dentries)
 		log_dentries = true;
 
+	/*
+	 * On unlink we must make sure all our current and old parent directores
+	 * inodes are fully logged. This is to prevent leaving dangling
+	 * directory index entries in directories that were our parents but are
+	 * not anymore. Not doing this results in old parent directory being
+	 * impossible to delete after log replay (rmdir will always fail with
+	 * error -ENOTEMPTY).
+	 *
+	 * Example 1:
+	 *
+	 * mkdir testdir
+	 * touch testdir/foo
+	 * ln testdir/foo testdir/bar
+	 * sync
+	 * unlink testdir/bar
+	 * xfs_io -c fsync testdir/foo
+	 * <power failure>
+	 * mount fs, triggers log replay
+	 *
+	 * If we don't log the parent directory (testdir), after log replay the
+	 * directory still has an entry pointing to the file inode using the bar
+	 * name, but a matching BTRFS_INODE_[REF|EXTREF]_KEY does not exist and
+	 * the file inode has a link count of 1.
+	 *
+	 * Example 2:
+	 *
+	 * mkdir testdir
+	 * touch foo
+	 * ln foo testdir/foo2
+	 * ln foo testdir/foo3
+	 * sync
+	 * unlink testdir/foo3
+	 * xfs_io -c fsync foo
+	 * <power failure>
+	 * mount fs, triggers log replay
+	 *
+	 * Similar as the first example, after log replay the parent directory
+	 * testdir still has an entry pointing to the inode file with name foo3
+	 * but the file inode does not have a matching BTRFS_INODE_REF_KEY item
+	 * and has a link count of 2.
+	 */
+	if (BTRFS_I(inode)->last_unlink_trans > last_committed) {
+		ret = btrfs_log_all_parents(trans, orig_inode, ctx);
+		if (ret)
+			goto end_trans;
+	}
+
 	while (1) {
 		if (!parent || d_really_is_negative(parent) || sb != d_inode(parent)->i_sb)
 			break;
@@ -4788,23 +5234,9 @@ static int btrfs_log_inode_parent(struct btrfs_trans_handle *trans,
 		if (root != BTRFS_I(inode)->root)
 			break;
 
-		/*
-		 * On unlink we must make sure our immediate parent directory
-		 * inode is fully logged. This is to prevent leaving dangling
-		 * directory index entries and a wrong directory inode's i_size.
-		 * Not doing so can result in a directory being impossible to
-		 * delete after log replay (rmdir will always fail with error
-		 * -ENOTEMPTY).
-		 */
-		if (did_unlink && parent == first_parent)
-			inode_only = LOG_INODE_ALL;
-		else
-			inode_only = LOG_INODE_EXISTS;
-
-		if (BTRFS_I(inode)->generation >
-		    root->fs_info->last_trans_committed ||
-		    inode_only == LOG_INODE_ALL) {
-			ret = btrfs_log_inode(trans, root, inode, inode_only,
+		if (BTRFS_I(inode)->generation > last_committed) {
+			ret = btrfs_log_inode(trans, root, inode,
+					      LOG_INODE_EXISTS,
 					      0, LLONG_MAX, ctx);
 			if (ret)
 				goto end_trans;
@@ -4892,7 +5324,7 @@ int btrfs_recover_log_trees(struct btrfs_root *log_root_tree)
 
 	ret = walk_log_tree(trans, log_root_tree, &wc);
 	if (ret) {
-		btrfs_error(fs_info, ret, "Failed to pin buffers while "
+		btrfs_std_error(fs_info, ret, "Failed to pin buffers while "
 			    "recovering log root tree.");
 		goto error;
 	}
@@ -4906,7 +5338,7 @@ again:
 		ret = btrfs_search_slot(NULL, log_root_tree, &key, path, 0, 0);
 
 		if (ret < 0) {
-			btrfs_error(fs_info, ret,
+			btrfs_std_error(fs_info, ret,
 				    "Couldn't find tree log root.");
 			goto error;
 		}
@@ -4924,7 +5356,7 @@ again:
 		log = btrfs_read_fs_root(log_root_tree, &found_key);
 		if (IS_ERR(log)) {
 			ret = PTR_ERR(log);
-			btrfs_error(fs_info, ret,
+			btrfs_std_error(fs_info, ret,
 				    "Couldn't read tree log root.");
 			goto error;
 		}
@@ -4939,7 +5371,7 @@ again:
 			free_extent_buffer(log->node);
 			free_extent_buffer(log->commit_root);
 			kfree(log);
-			btrfs_error(fs_info, ret, "Couldn't read target root "
+			btrfs_std_error(fs_info, ret, "Couldn't read target root "
 				    "for tree log recovery.");
 			goto error;
 		}
diff --git a/kernel/fs/btrfs/ulist.c b/kernel/fs/btrfs/ulist.c
index 840a38b27..91feb2bde 100644
--- a/kernel/fs/btrfs/ulist.c
+++ b/kernel/fs/btrfs/ulist.c
@@ -132,6 +132,15 @@ static struct ulist_node *ulist_rbtree_search(struct ulist *ulist, u64 val)
 	return NULL;
 }
 
+static void ulist_rbtree_erase(struct ulist *ulist, struct ulist_node *node)
+{
+	rb_erase(&node->rb_node, &ulist->root);
+	list_del(&node->list);
+	kfree(node);
+	BUG_ON(ulist->nnodes == 0);
+	ulist->nnodes--;
+}
+
 static int ulist_rbtree_insert(struct ulist *ulist, struct ulist_node *ins)
 {
 	struct rb_node **p = &ulist->root.rb_node;
@@ -197,9 +206,6 @@ int ulist_add_merge(struct ulist *ulist, u64 val, u64 aux,
 
 	node->val = val;
 	node->aux = aux;
-#ifdef CONFIG_BTRFS_DEBUG
-	node->seqnum = ulist->nnodes;
-#endif
 
 	ret = ulist_rbtree_insert(ulist, node);
 	ASSERT(!ret);
@@ -209,6 +215,33 @@ int ulist_add_merge(struct ulist *ulist, u64 val, u64 aux,
 	return 1;
 }
 
+/*
+ * ulist_del - delete one node from ulist
+ * @ulist:	ulist to remove node from
+ * @val:	value to delete
+ * @aux:	aux to delete
+ *
+ * The deletion will only be done when *BOTH* val and aux matches.
+ * Return 0 for successful delete.
+ * Return > 0 for not found.
+ */
+int ulist_del(struct ulist *ulist, u64 val, u64 aux)
+{
+	struct ulist_node *node;
+
+	node = ulist_rbtree_search(ulist, val);
+	/* Not found */
+	if (!node)
+		return 1;
+
+	if (node->aux != aux)
+		return 1;
+
+	/* Found and delete */
+	ulist_rbtree_erase(ulist, node);
+	return 0;
+}
+
 /**
  * ulist_next - iterate ulist
  * @ulist:	ulist to iterate
@@ -237,15 +270,7 @@ struct ulist_node *ulist_next(struct ulist *ulist, struct ulist_iterator *uiter)
 		uiter->cur_list = uiter->cur_list->next;
 	} else {
 		uiter->cur_list = ulist->nodes.next;
-#ifdef CONFIG_BTRFS_DEBUG
-		uiter->i = 0;
-#endif
 	}
 	node = list_entry(uiter->cur_list, struct ulist_node, list);
-#ifdef CONFIG_BTRFS_DEBUG
-	ASSERT(node->seqnum == uiter->i);
-	ASSERT(uiter->i >= 0 && uiter->i < ulist->nnodes);
-	uiter->i++;
-#endif
 	return node;
 }
diff --git a/kernel/fs/btrfs/ulist.h b/kernel/fs/btrfs/ulist.h
index 4c29db604..a01a2c458 100644
--- a/kernel/fs/btrfs/ulist.h
+++ b/kernel/fs/btrfs/ulist.h
@@ -57,6 +57,7 @@ void ulist_free(struct ulist *ulist);
 int ulist_add(struct ulist *ulist, u64 val, u64 aux, gfp_t gfp_mask);
 int ulist_add_merge(struct ulist *ulist, u64 val, u64 aux,
 		    u64 *old_aux, gfp_t gfp_mask);
+int ulist_del(struct ulist *ulist, u64 val, u64 aux);
 
 /* just like ulist_add_merge() but take a pointer for the aux data */
 static inline int ulist_add_merge_ptr(struct ulist *ulist, u64 val, void *aux,
diff --git a/kernel/fs/btrfs/volumes.c b/kernel/fs/btrfs/volumes.c
index 174f5e1e0..9c62a6f97 100644
--- a/kernel/fs/btrfs/volumes.c
+++ b/kernel/fs/btrfs/volumes.c
@@ -42,6 +42,82 @@
 #include "dev-replace.h"
 #include "sysfs.h"
 
+const struct btrfs_raid_attr btrfs_raid_array[BTRFS_NR_RAID_TYPES] = {
+	[BTRFS_RAID_RAID10] = {
+		.sub_stripes	= 2,
+		.dev_stripes	= 1,
+		.devs_max	= 0,	/* 0 == as many as possible */
+		.devs_min	= 4,
+		.tolerated_failures = 1,
+		.devs_increment	= 2,
+		.ncopies	= 2,
+	},
+	[BTRFS_RAID_RAID1] = {
+		.sub_stripes	= 1,
+		.dev_stripes	= 1,
+		.devs_max	= 2,
+		.devs_min	= 2,
+		.tolerated_failures = 1,
+		.devs_increment	= 2,
+		.ncopies	= 2,
+	},
+	[BTRFS_RAID_DUP] = {
+		.sub_stripes	= 1,
+		.dev_stripes	= 2,
+		.devs_max	= 1,
+		.devs_min	= 1,
+		.tolerated_failures = 0,
+		.devs_increment	= 1,
+		.ncopies	= 2,
+	},
+	[BTRFS_RAID_RAID0] = {
+		.sub_stripes	= 1,
+		.dev_stripes	= 1,
+		.devs_max	= 0,
+		.devs_min	= 2,
+		.tolerated_failures = 0,
+		.devs_increment	= 1,
+		.ncopies	= 1,
+	},
+	[BTRFS_RAID_SINGLE] = {
+		.sub_stripes	= 1,
+		.dev_stripes	= 1,
+		.devs_max	= 1,
+		.devs_min	= 1,
+		.tolerated_failures = 0,
+		.devs_increment	= 1,
+		.ncopies	= 1,
+	},
+	[BTRFS_RAID_RAID5] = {
+		.sub_stripes	= 1,
+		.dev_stripes	= 1,
+		.devs_max	= 0,
+		.devs_min	= 2,
+		.tolerated_failures = 1,
+		.devs_increment	= 1,
+		.ncopies	= 2,
+	},
+	[BTRFS_RAID_RAID6] = {
+		.sub_stripes	= 1,
+		.dev_stripes	= 1,
+		.devs_max	= 0,
+		.devs_min	= 3,
+		.tolerated_failures = 2,
+		.devs_increment	= 1,
+		.ncopies	= 3,
+	},
+};
+
+const u64 const btrfs_raid_group[BTRFS_NR_RAID_TYPES] = {
+	[BTRFS_RAID_RAID10] = BTRFS_BLOCK_GROUP_RAID10,
+	[BTRFS_RAID_RAID1]  = BTRFS_BLOCK_GROUP_RAID1,
+	[BTRFS_RAID_DUP]    = BTRFS_BLOCK_GROUP_DUP,
+	[BTRFS_RAID_RAID0]  = BTRFS_BLOCK_GROUP_RAID0,
+	[BTRFS_RAID_SINGLE] = 0,
+	[BTRFS_RAID_RAID5]  = BTRFS_BLOCK_GROUP_RAID5,
+	[BTRFS_RAID_RAID6]  = BTRFS_BLOCK_GROUP_RAID6,
+};
+
 static int init_first_rw_device(struct btrfs_trans_handle *trans,
 				struct btrfs_root *root,
 				struct btrfs_device *device);
@@ -52,6 +128,10 @@ static void btrfs_dev_stat_print_on_load(struct btrfs_device *device);
 
 DEFINE_MUTEX(uuid_mutex);
 static LIST_HEAD(fs_uuids);
+struct list_head *btrfs_get_fs_uuids(void)
+{
+	return &fs_uuids;
+}
 
 static struct btrfs_fs_devices *__alloc_fs_devices(void)
 {
@@ -152,8 +232,9 @@ static struct btrfs_device *__alloc_device(void)
 	spin_lock_init(&dev->reada_lock);
 	atomic_set(&dev->reada_in_flight, 0);
 	atomic_set(&dev->dev_stats_ccnt, 0);
-	INIT_RADIX_TREE(&dev->reada_zones, GFP_NOFS & ~__GFP_WAIT);
-	INIT_RADIX_TREE(&dev->reada_extents, GFP_NOFS & ~__GFP_WAIT);
+	btrfs_device_data_ordered_init(dev);
+	INIT_RADIX_TREE(&dev->reada_zones, GFP_NOFS & ~__GFP_DIRECT_RECLAIM);
+	INIT_RADIX_TREE(&dev->reada_extents, GFP_NOFS & ~__GFP_DIRECT_RECLAIM);
 
 	return dev;
 }
@@ -194,7 +275,6 @@ btrfs_get_bdev_and_sb(const char *device_path, fmode_t flags, void *holder,
 
 	if (IS_ERR(*bdev)) {
 		ret = PTR_ERR(*bdev);
-		printk(KERN_INFO "BTRFS: open %s failed\n", device_path);
 		goto error;
 	}
 
@@ -207,8 +287,8 @@ btrfs_get_bdev_and_sb(const char *device_path, fmode_t flags, void *holder,
 	}
 	invalidate_bdev(*bdev);
 	*bh = btrfs_read_dev_super(*bdev);
-	if (!*bh) {
-		ret = -EINVAL;
+	if (IS_ERR(*bh)) {
+		ret = PTR_ERR(*bh);
 		blkdev_put(*bdev, flags);
 		goto error;
 	}
@@ -341,11 +421,14 @@ loop_lock:
 		pending = pending->bi_next;
 		cur->bi_next = NULL;
 
+		/*
+		 * atomic_dec_return implies a barrier for waitqueue_active
+		 */
 		if (atomic_dec_return(&fs_info->nr_async_bios) < limit &&
 		    waitqueue_active(&fs_info->async_submit_wait))
 			wake_up(&fs_info->async_submit_wait);
 
-		BUG_ON(atomic_read(&cur->bi_cnt) == 0);
+		BUG_ON(atomic_read(&cur->__bi_cnt) == 0);
 
 		/*
 		 * if we're doing the sync list, record that our
@@ -441,6 +524,61 @@ static void pending_bios_fn(struct btrfs_work *work)
 	run_scheduled_bios(device);
 }
 
+
+void btrfs_free_stale_device(struct btrfs_device *cur_dev)
+{
+	struct btrfs_fs_devices *fs_devs;
+	struct btrfs_device *dev;
+
+	if (!cur_dev->name)
+		return;
+
+	list_for_each_entry(fs_devs, &fs_uuids, list) {
+		int del = 1;
+
+		if (fs_devs->opened)
+			continue;
+		if (fs_devs->seeding)
+			continue;
+
+		list_for_each_entry(dev, &fs_devs->devices, dev_list) {
+
+			if (dev == cur_dev)
+				continue;
+			if (!dev->name)
+				continue;
+
+			/*
+			 * Todo: This won't be enough. What if the same device
+			 * comes back (with new uuid and) with its mapper path?
+			 * But for now, this does help as mostly an admin will
+			 * either use mapper or non mapper path throughout.
+			 */
+			rcu_read_lock();
+			del = strcmp(rcu_str_deref(dev->name),
+						rcu_str_deref(cur_dev->name));
+			rcu_read_unlock();
+			if (!del)
+				break;
+		}
+
+		if (!del) {
+			/* delete the stale device */
+			if (fs_devs->num_devices == 1) {
+				btrfs_sysfs_remove_fsid(fs_devs);
+				list_del(&fs_devs->list);
+				free_fs_devices(fs_devs);
+			} else {
+				fs_devs->num_devices--;
+				list_del(&dev->dev_list);
+				rcu_string_free(dev->name);
+				kfree(dev);
+			}
+			break;
+		}
+	}
+}
+
 /*
  * Add new device to list of registered devices
  *
@@ -556,6 +694,12 @@ static noinline int device_list_add(const char *path,
 	if (!fs_devices->opened)
 		device->generation = found_transid;
 
+	/*
+	 * if there is new btrfs on an already registered device,
+	 * then remove the stale device entry.
+	 */
+	btrfs_free_stale_device(device);
+
 	*fs_devices_ret = fs_devices;
 
 	return ret;
@@ -693,43 +837,14 @@ static void free_device(struct rcu_head *head)
 
 static int __btrfs_close_devices(struct btrfs_fs_devices *fs_devices)
 {
-	struct btrfs_device *device;
+	struct btrfs_device *device, *tmp;
 
 	if (--fs_devices->opened > 0)
 		return 0;
 
 	mutex_lock(&fs_devices->device_list_mutex);
-	list_for_each_entry(device, &fs_devices->devices, dev_list) {
-		struct btrfs_device *new_device;
-		struct rcu_string *name;
-
-		if (device->bdev)
-			fs_devices->open_devices--;
-
-		if (device->writeable &&
-		    device->devid != BTRFS_DEV_REPLACE_DEVID) {
-			list_del_init(&device->dev_alloc_list);
-			fs_devices->rw_devices--;
-		}
-
-		if (device->missing)
-			fs_devices->missing_devices--;
-
-		new_device = btrfs_alloc_device(NULL, &device->devid,
-						device->uuid);
-		BUG_ON(IS_ERR(new_device)); /* -ENOMEM */
-
-		/* Safe because we are under uuid_mutex */
-		if (device->name) {
-			name = rcu_string_strdup(device->name->str, GFP_NOFS);
-			BUG_ON(!name); /* -ENOMEM */
-			rcu_assign_pointer(new_device->name, name);
-		}
-
-		list_replace_rcu(&device->dev_list, &new_device->dev_list);
-		new_device->fs_devices = device->fs_devices;
-
-		call_rcu(&device->rcu, free_device);
+	list_for_each_entry_safe(device, tmp, &fs_devices->devices, dev_list) {
+		btrfs_close_one_device(device);
 	}
 	mutex_unlock(&fs_devices->device_list_mutex);
 
@@ -1051,15 +1166,18 @@ out:
 	return ret;
 }
 
-static int contains_pending_extent(struct btrfs_trans_handle *trans,
+static int contains_pending_extent(struct btrfs_transaction *transaction,
 				   struct btrfs_device *device,
 				   u64 *start, u64 len)
 {
+	struct btrfs_fs_info *fs_info = device->dev_root->fs_info;
 	struct extent_map *em;
-	struct list_head *search_list = &trans->transaction->pending_chunks;
+	struct list_head *search_list = &fs_info->pinned_chunks;
 	int ret = 0;
 	u64 physical_start = *start;
 
+	if (transaction)
+		search_list = &transaction->pending_chunks;
 again:
 	list_for_each_entry(em, search_list, list) {
 		struct map_lookup *map;
@@ -1067,19 +1185,35 @@ again:
 
 		map = (struct map_lookup *)em->bdev;
 		for (i = 0; i < map->num_stripes; i++) {
+			u64 end;
+
 			if (map->stripes[i].dev != device)
 				continue;
 			if (map->stripes[i].physical >= physical_start + len ||
 			    map->stripes[i].physical + em->orig_block_len <=
 			    physical_start)
 				continue;
-			*start = map->stripes[i].physical +
-				em->orig_block_len;
-			ret = 1;
+			/*
+			 * Make sure that while processing the pinned list we do
+			 * not override our *start with a lower value, because
+			 * we can have pinned chunks that fall within this
+			 * device hole and that have lower physical addresses
+			 * than the pending chunks we processed before. If we
+			 * do not take this special care we can end up getting
+			 * 2 pending chunks that start at the same physical
+			 * device offsets because the end offset of a pinned
+			 * chunk can be equal to the start offset of some
+			 * pending chunk.
+			 */
+			end = map->stripes[i].physical + em->orig_block_len;
+			if (end > *start) {
+				*start = end;
+				ret = 1;
+			}
 		}
 	}
-	if (search_list == &trans->transaction->pending_chunks) {
-		search_list = &trans->root->fs_info->pinned_chunks;
+	if (search_list != &fs_info->pinned_chunks) {
+		search_list = &fs_info->pinned_chunks;
 		goto again;
 	}
 
@@ -1088,12 +1222,13 @@ again:
 
 
 /*
- * find_free_dev_extent - find free space in the specified device
- * @device:	the device which we search the free space in
- * @num_bytes:	the size of the free space that we need
- * @start:	store the start of the free space.
- * @len:	the size of the free space. that we find, or the size of the max
- * 		free space if we don't find suitable free space
+ * find_free_dev_extent_start - find free space in the specified device
+ * @device:	  the device which we search the free space in
+ * @num_bytes:	  the size of the free space that we need
+ * @search_start: the position from which to begin the search
+ * @start:	  store the start of the free space.
+ * @len:	  the size of the free space. that we find, or the size
+ *		  of the max free space if we don't find suitable free space
  *
  * this uses a pretty simple search, the expectation is that it is
  * called very infrequently and that a given device has a small number
@@ -1107,9 +1242,9 @@ again:
  * But if we don't find suitable free space, it is used to store the size of
  * the max free space.
  */
-int find_free_dev_extent(struct btrfs_trans_handle *trans,
-			 struct btrfs_device *device, u64 num_bytes,
-			 u64 *start, u64 *len)
+int find_free_dev_extent_start(struct btrfs_transaction *transaction,
+			       struct btrfs_device *device, u64 num_bytes,
+			       u64 search_start, u64 *start, u64 *len)
 {
 	struct btrfs_key key;
 	struct btrfs_root *root = device->dev_root;
@@ -1119,18 +1254,19 @@ int find_free_dev_extent(struct btrfs_trans_handle *trans,
 	u64 max_hole_start;
 	u64 max_hole_size;
 	u64 extent_end;
-	u64 search_start;
 	u64 search_end = device->total_bytes;
 	int ret;
 	int slot;
 	struct extent_buffer *l;
+	u64 min_search_start;
 
-	/* FIXME use last free of some kind */
-
-	/* we don't want to overwrite the superblock on the drive,
-	 * so we make sure to start at an offset of at least 1MB
+	/*
+	 * We don't want to overwrite the superblock on the drive nor any area
+	 * used by the boot loader (grub for example), so we make sure to start
+	 * at an offset of at least 1MB.
 	 */
-	search_start = max(root->fs_info->alloc_start, 1024ull * 1024);
+	min_search_start = max(root->fs_info->alloc_start, 1024ull * 1024);
+	search_start = max(search_start, min_search_start);
 
 	path = btrfs_alloc_path();
 	if (!path)
@@ -1192,7 +1328,7 @@ again:
 			 * Have to check before we set max_hole_start, otherwise
 			 * we could end up sending back this offset anyway.
 			 */
-			if (contains_pending_extent(trans, device,
+			if (contains_pending_extent(transaction, device,
 						    &search_start,
 						    hole_size)) {
 				if (key.offset >= search_start) {
@@ -1241,7 +1377,7 @@ next:
 	if (search_end > search_start) {
 		hole_size = search_end - search_start;
 
-		if (contains_pending_extent(trans, device, &search_start,
+		if (contains_pending_extent(transaction, device, &search_start,
 					    hole_size)) {
 			btrfs_release_path(path);
 			goto again;
@@ -1267,6 +1403,15 @@ out:
 	return ret;
 }
 
+int find_free_dev_extent(struct btrfs_trans_handle *trans,
+			 struct btrfs_device *device, u64 num_bytes,
+			 u64 *start, u64 *len)
+{
+	/* FIXME use last free of some kind */
+	return find_free_dev_extent_start(trans->transaction, device,
+					  num_bytes, 0, start, len);
+}
+
 static int btrfs_free_dev_extent(struct btrfs_trans_handle *trans,
 			  struct btrfs_device *device,
 			  u64 start, u64 *dev_extent_len)
@@ -1307,7 +1452,7 @@ again:
 		extent = btrfs_item_ptr(leaf, path->slots[0],
 					struct btrfs_dev_extent);
 	} else {
-		btrfs_error(root->fs_info, ret, "Slot search failed");
+		btrfs_std_error(root->fs_info, ret, "Slot search failed");
 		goto out;
 	}
 
@@ -1315,10 +1460,10 @@ again:
 
 	ret = btrfs_del_item(trans, root, path);
 	if (ret) {
-		btrfs_error(root->fs_info, ret,
+		btrfs_std_error(root->fs_info, ret,
 			    "Failed to remove dev extent item");
 	} else {
-		trans->transaction->have_free_bgs = 1;
+		set_bit(BTRFS_TRANS_HAVE_FREE_BGS, &trans->transaction->flags);
 	}
 out:
 	btrfs_free_path(path);
@@ -1706,7 +1851,7 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path)
 	if (device->bdev) {
 		device->fs_devices->open_devices--;
 		/* remove sysfs entry */
-		btrfs_kobj_rm_device(root->fs_info, device);
+		btrfs_sysfs_rm_device_link(root->fs_info->fs_devices, device);
 	}
 
 	call_rcu(&device->rcu, free_device);
@@ -1829,7 +1974,7 @@ void btrfs_rm_dev_replace_remove_srcdev(struct btrfs_fs_info *fs_info,
 	if (srcdev->writeable) {
 		fs_devices->rw_devices--;
 		/* zero out the old super if it is writable */
-		btrfs_scratch_superblock(srcdev);
+		btrfs_scratch_superblocks(srcdev->bdev, srcdev->name->str);
 	}
 
 	if (srcdev->bdev)
@@ -1875,8 +2020,11 @@ void btrfs_destroy_dev_replace_tgtdev(struct btrfs_fs_info *fs_info,
 	mutex_lock(&uuid_mutex);
 	WARN_ON(!tgtdev);
 	mutex_lock(&fs_info->fs_devices->device_list_mutex);
+
+	btrfs_sysfs_rm_device_link(fs_info->fs_devices, tgtdev);
+
 	if (tgtdev->bdev) {
-		btrfs_scratch_superblock(tgtdev);
+		btrfs_scratch_superblocks(tgtdev->bdev, tgtdev->name->str);
 		fs_info->fs_devices->open_devices--;
 	}
 	fs_info->fs_devices->num_devices--;
@@ -1943,10 +2091,8 @@ int btrfs_find_device_missing_or_by_path(struct btrfs_root *root,
 			}
 		}
 
-		if (!*device) {
-			btrfs_err(root->fs_info, "no missing device found");
-			return -ENOENT;
-		}
+		if (!*device)
+			return BTRFS_ERROR_DEV_MISSING_NOT_FOUND;
 
 		return 0;
 	} else {
@@ -2211,7 +2357,7 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path)
 				    tmp + 1);
 
 	/* add sysfs device entry */
-	btrfs_kobj_add_device(root->fs_info, device);
+	btrfs_sysfs_add_device_link(root->fs_info->fs_devices, device);
 
 	/*
 	 * we've got more storage, clear any full flags on the space
@@ -2252,8 +2398,10 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path)
 		 */
 		snprintf(fsid_buf, BTRFS_UUID_UNPARSED_SIZE, "%pU",
 						root->fs_info->fsid);
-		if (kobject_rename(&root->fs_info->super_kobj, fsid_buf))
-			goto error_trans;
+		if (kobject_rename(&root->fs_info->fs_devices->fsid_kobj,
+								fsid_buf))
+			btrfs_warn(root->fs_info,
+				"sysfs: failed to create fsid for sprout");
 	}
 
 	root->fs_info->num_tolerated_disk_barrier_failures =
@@ -2269,7 +2417,7 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path)
 
 		ret = btrfs_relocate_sys_chunks(root);
 		if (ret < 0)
-			btrfs_error(root->fs_info, ret,
+			btrfs_std_error(root->fs_info, ret,
 				    "Failed to relocate sys chunks after "
 				    "device initialization. This can be fixed "
 				    "using the \"btrfs balance\" command.");
@@ -2289,7 +2437,7 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path)
 error_trans:
 	btrfs_end_transaction(trans, root);
 	rcu_string_free(device->name);
-	btrfs_kobj_rm_device(root->fs_info, device);
+	btrfs_sysfs_rm_device_link(root->fs_info->fs_devices, device);
 	kfree(device);
 error:
 	blkdev_put(bdev, FMODE_EXCL);
@@ -2514,7 +2662,7 @@ static int btrfs_free_chunk(struct btrfs_trans_handle *trans,
 	if (ret < 0)
 		goto out;
 	else if (ret > 0) { /* Logic error or corruption */
-		btrfs_error(root->fs_info, -ENOENT,
+		btrfs_std_error(root->fs_info, -ENOENT,
 			    "Failed lookup while freeing chunk.");
 		ret = -ENOENT;
 		goto out;
@@ -2522,7 +2670,7 @@ static int btrfs_free_chunk(struct btrfs_trans_handle *trans,
 
 	ret = btrfs_del_item(trans, root, path);
 	if (ret < 0)
-		btrfs_error(root->fs_info, ret,
+		btrfs_std_error(root->fs_info, ret,
 			    "Failed to delete chunk item.");
 out:
 	btrfs_free_path(path);
@@ -2609,6 +2757,9 @@ int btrfs_remove_chunk(struct btrfs_trans_handle *trans,
 		return -EINVAL;
 	}
 	map = (struct map_lookup *)em->bdev;
+	lock_chunks(root->fs_info->chunk_root);
+	check_system_chunk(trans, extent_root, map->type);
+	unlock_chunks(root->fs_info->chunk_root);
 
 	for (i = 0; i < map->num_stripes; i++) {
 		struct btrfs_device *device = map->stripes[i].dev;
@@ -2667,9 +2818,7 @@ out:
 	return ret;
 }
 
-static int btrfs_relocate_chunk(struct btrfs_root *root,
-				u64 chunk_objectid,
-				u64 chunk_offset)
+static int btrfs_relocate_chunk(struct btrfs_root *root, u64 chunk_offset)
 {
 	struct btrfs_root *extent_root;
 	struct btrfs_trans_handle *trans;
@@ -2678,19 +2827,36 @@ static int btrfs_relocate_chunk(struct btrfs_root *root,
 	root = root->fs_info->chunk_root;
 	extent_root = root->fs_info->extent_root;
 
+	/*
+	 * Prevent races with automatic removal of unused block groups.
+	 * After we relocate and before we remove the chunk with offset
+	 * chunk_offset, automatic removal of the block group can kick in,
+	 * resulting in a failure when calling btrfs_remove_chunk() below.
+	 *
+	 * Make sure to acquire this mutex before doing a tree search (dev
+	 * or chunk trees) to find chunks. Otherwise the cleaner kthread might
+	 * call btrfs_remove_chunk() (through btrfs_delete_unused_bgs()) after
+	 * we release the path used to search the chunk/dev tree and before
+	 * the current task acquires this mutex and calls us.
+	 */
+	ASSERT(mutex_is_locked(&root->fs_info->delete_unused_bgs_mutex));
+
 	ret = btrfs_can_relocate(extent_root, chunk_offset);
 	if (ret)
 		return -ENOSPC;
 
 	/* step one, relocate all the extents inside this chunk */
+	btrfs_scrub_pause(root);
 	ret = btrfs_relocate_block_group(extent_root, chunk_offset);
+	btrfs_scrub_continue(root);
 	if (ret)
 		return ret;
 
-	trans = btrfs_start_transaction(root, 0);
+	trans = btrfs_start_trans_remove_block_group(root->fs_info,
+						     chunk_offset);
 	if (IS_ERR(trans)) {
 		ret = PTR_ERR(trans);
-		btrfs_std_error(root->fs_info, ret);
+		btrfs_std_error(root->fs_info, ret, NULL);
 		return ret;
 	}
 
@@ -2726,13 +2892,18 @@ again:
 	key.type = BTRFS_CHUNK_ITEM_KEY;
 
 	while (1) {
+		mutex_lock(&root->fs_info->delete_unused_bgs_mutex);
 		ret = btrfs_search_slot(NULL, chunk_root, &key, path, 0, 0);
-		if (ret < 0)
+		if (ret < 0) {
+			mutex_unlock(&root->fs_info->delete_unused_bgs_mutex);
 			goto error;
+		}
 		BUG_ON(ret == 0); /* Corruption */
 
 		ret = btrfs_previous_item(chunk_root, path, key.objectid,
 					  key.type);
+		if (ret)
+			mutex_unlock(&root->fs_info->delete_unused_bgs_mutex);
 		if (ret < 0)
 			goto error;
 		if (ret > 0)
@@ -2748,13 +2919,13 @@ again:
 
 		if (chunk_type & BTRFS_BLOCK_GROUP_SYSTEM) {
 			ret = btrfs_relocate_chunk(chunk_root,
-						   found_key.objectid,
 						   found_key.offset);
 			if (ret == -ENOSPC)
 				failed++;
 			else
 				BUG_ON(ret);
 		}
+		mutex_unlock(&root->fs_info->delete_unused_bgs_mutex);
 
 		if (found_key.offset == 0)
 			break;
@@ -2888,16 +3059,19 @@ static void update_balance_args(struct btrfs_balance_control *bctl)
 	 * (albeit full) chunks.
 	 */
 	if (!(bctl->data.flags & BTRFS_BALANCE_ARGS_USAGE) &&
+	    !(bctl->data.flags & BTRFS_BALANCE_ARGS_USAGE_RANGE) &&
 	    !(bctl->data.flags & BTRFS_BALANCE_ARGS_CONVERT)) {
 		bctl->data.flags |= BTRFS_BALANCE_ARGS_USAGE;
 		bctl->data.usage = 90;
 	}
 	if (!(bctl->sys.flags & BTRFS_BALANCE_ARGS_USAGE) &&
+	    !(bctl->sys.flags & BTRFS_BALANCE_ARGS_USAGE_RANGE) &&
 	    !(bctl->sys.flags & BTRFS_BALANCE_ARGS_CONVERT)) {
 		bctl->sys.flags |= BTRFS_BALANCE_ARGS_USAGE;
 		bctl->sys.usage = 90;
 	}
 	if (!(bctl->meta.flags & BTRFS_BALANCE_ARGS_USAGE) &&
+	    !(bctl->meta.flags & BTRFS_BALANCE_ARGS_USAGE_RANGE) &&
 	    !(bctl->meta.flags & BTRFS_BALANCE_ARGS_CONVERT)) {
 		bctl->meta.flags |= BTRFS_BALANCE_ARGS_USAGE;
 		bctl->meta.usage = 90;
@@ -2949,17 +3123,50 @@ static int chunk_profiles_filter(u64 chunk_type,
 	return 1;
 }
 
-static int chunk_usage_filter(struct btrfs_fs_info *fs_info, u64 chunk_offset,
+static int chunk_usage_range_filter(struct btrfs_fs_info *fs_info, u64 chunk_offset,
 			      struct btrfs_balance_args *bargs)
 {
 	struct btrfs_block_group_cache *cache;
+	u64 chunk_used;
+	u64 user_thresh_min;
+	u64 user_thresh_max;
+	int ret = 1;
+
+	cache = btrfs_lookup_block_group(fs_info, chunk_offset);
+	chunk_used = btrfs_block_group_used(&cache->item);
+
+	if (bargs->usage_min == 0)
+		user_thresh_min = 0;
+	else
+		user_thresh_min = div_factor_fine(cache->key.offset,
+					bargs->usage_min);
+
+	if (bargs->usage_max == 0)
+		user_thresh_max = 1;
+	else if (bargs->usage_max > 100)
+		user_thresh_max = cache->key.offset;
+	else
+		user_thresh_max = div_factor_fine(cache->key.offset,
+					bargs->usage_max);
+
+	if (user_thresh_min <= chunk_used && chunk_used < user_thresh_max)
+		ret = 0;
+
+	btrfs_put_block_group(cache);
+	return ret;
+}
+
+static int chunk_usage_filter(struct btrfs_fs_info *fs_info,
+		u64 chunk_offset, struct btrfs_balance_args *bargs)
+{
+	struct btrfs_block_group_cache *cache;
 	u64 chunk_used, user_thresh;
 	int ret = 1;
 
 	cache = btrfs_lookup_block_group(fs_info, chunk_offset);
 	chunk_used = btrfs_block_group_used(&cache->item);
 
-	if (bargs->usage == 0)
+	if (bargs->usage_min == 0)
 		user_thresh = 1;
 	else if (bargs->usage > 100)
 		user_thresh = cache->key.offset;
@@ -3049,6 +3256,19 @@ static int chunk_vrange_filter(struct extent_buffer *leaf,
 	return 1;
 }
 
+static int chunk_stripes_range_filter(struct extent_buffer *leaf,
+			       struct btrfs_chunk *chunk,
+			       struct btrfs_balance_args *bargs)
+{
+	int num_stripes = btrfs_chunk_num_stripes(leaf, chunk);
+
+	if (bargs->stripes_min <= num_stripes
+			&& num_stripes <= bargs->stripes_max)
+		return 0;
+
+	return 1;
+}
+
 static int chunk_soft_convert_filter(u64 chunk_type,
 				     struct btrfs_balance_args *bargs)
 {
@@ -3095,6 +3315,9 @@ static int should_balance_chunk(struct btrfs_root *root,
 	if ((bargs->flags & BTRFS_BALANCE_ARGS_USAGE) &&
 	    chunk_usage_filter(bctl->fs_info, chunk_offset, bargs)) {
 		return 0;
+	} else if ((bargs->flags & BTRFS_BALANCE_ARGS_USAGE_RANGE) &&
+	    chunk_usage_range_filter(bctl->fs_info, chunk_offset, bargs)) {
+		return 0;
 	}
 
 	/* devid filter */
@@ -3115,6 +3338,12 @@ static int should_balance_chunk(struct btrfs_root *root,
 		return 0;
 	}
 
+	/* stripes filter */
+	if ((bargs->flags & BTRFS_BALANCE_ARGS_STRIPES_RANGE) &&
+	    chunk_stripes_range_filter(leaf, chunk, bargs)) {
+		return 0;
+	}
+
 	/* soft profile changing mode */
 	if ((bargs->flags & BTRFS_BALANCE_ARGS_SOFT) &&
 	    chunk_soft_convert_filter(chunk_type, bargs)) {
@@ -3129,6 +3358,16 @@ static int should_balance_chunk(struct btrfs_root *root,
 			return 0;
 		else
 			bargs->limit--;
+	} else if ((bargs->flags & BTRFS_BALANCE_ARGS_LIMIT_RANGE)) {
+		/*
+		 * Same logic as the 'limit' filter; the minimum cannot be
+		 * determined here because we do not have the global informatoin
+		 * about the count of all chunks that satisfy the filters.
+		 */
+		if (bargs->limit_max == 0)
+			return 0;
+		else
+			bargs->limit_max--;
 	}
 
 	return 1;
@@ -3143,6 +3382,7 @@ static int __btrfs_balance(struct btrfs_fs_info *fs_info)
 	struct btrfs_device *device;
 	u64 old_size;
 	u64 size_to_free;
+	u64 chunk_type;
 	struct btrfs_chunk *chunk;
 	struct btrfs_path *path;
 	struct btrfs_key key;
@@ -3153,9 +3393,14 @@ static int __btrfs_balance(struct btrfs_fs_info *fs_info)
 	int ret;
 	int enospc_errors = 0;
 	bool counting = true;
+	/* The single value limit and min/max limits use the same bytes in the */
 	u64 limit_data = bctl->data.limit;
 	u64 limit_meta = bctl->meta.limit;
 	u64 limit_sys = bctl->sys.limit;
+	u32 count_data = 0;
+	u32 count_meta = 0;
+	u32 count_sys = 0;
+	int chunk_reserved = 0;
 
 	/* step one make some room on all the devices */
 	devices = &fs_info->fs_devices->devices;
@@ -3196,6 +3441,10 @@ static int __btrfs_balance(struct btrfs_fs_info *fs_info)
 	spin_unlock(&fs_info->balance_lock);
 again:
 	if (!counting) {
+		/*
+		 * The single value limit and min/max limits use the same bytes
+		 * in the
+		 */
 		bctl->data.limit = limit_data;
 		bctl->meta.limit = limit_meta;
 		bctl->sys.limit = limit_sys;
@@ -3211,9 +3460,12 @@ again:
 			goto error;
 		}
 
+		mutex_lock(&fs_info->delete_unused_bgs_mutex);
 		ret = btrfs_search_slot(NULL, chunk_root, &key, path, 0, 0);
-		if (ret < 0)
+		if (ret < 0) {
+			mutex_unlock(&fs_info->delete_unused_bgs_mutex);
 			goto error;
+		}
 
 		/*
 		 * this shouldn't happen, it means the last relocate
@@ -3225,6 +3477,7 @@ again:
 		ret = btrfs_previous_item(chunk_root, path, 0,
 					  BTRFS_CHUNK_ITEM_KEY);
 		if (ret) {
+			mutex_unlock(&fs_info->delete_unused_bgs_mutex);
 			ret = 0;
 			break;
 		}
@@ -3233,10 +3486,13 @@ again:
 		slot = path->slots[0];
 		btrfs_item_key_to_cpu(leaf, &found_key, slot);
 
-		if (found_key.objectid != key.objectid)
+		if (found_key.objectid != key.objectid) {
+			mutex_unlock(&fs_info->delete_unused_bgs_mutex);
 			break;
+		}
 
 		chunk = btrfs_item_ptr(leaf, slot, struct btrfs_chunk);
+		chunk_type = btrfs_chunk_type(leaf, chunk);
 
 		if (!counting) {
 			spin_lock(&fs_info->balance_lock);
@@ -3246,20 +3502,64 @@ again:
 
 		ret = should_balance_chunk(chunk_root, leaf, chunk,
 					   found_key.offset);
+
 		btrfs_release_path(path);
-		if (!ret)
+		if (!ret) {
+			mutex_unlock(&fs_info->delete_unused_bgs_mutex);
 			goto loop;
+		}
 
 		if (counting) {
+			mutex_unlock(&fs_info->delete_unused_bgs_mutex);
 			spin_lock(&fs_info->balance_lock);
 			bctl->stat.expected++;
 			spin_unlock(&fs_info->balance_lock);
+
+			if (chunk_type & BTRFS_BLOCK_GROUP_DATA)
+				count_data++;
+			else if (chunk_type & BTRFS_BLOCK_GROUP_SYSTEM)
+				count_sys++;
+			else if (chunk_type & BTRFS_BLOCK_GROUP_METADATA)
+				count_meta++;
+
+			goto loop;
+		}
+
+		/*
+		 * Apply limit_min filter, no need to check if the LIMITS
+		 * filter is used, limit_min is 0 by default
+		 */
+		if (((chunk_type & BTRFS_BLOCK_GROUP_DATA) &&
+					count_data < bctl->data.limit_min)
+				|| ((chunk_type & BTRFS_BLOCK_GROUP_METADATA) &&
+					count_meta < bctl->meta.limit_min)
+				|| ((chunk_type & BTRFS_BLOCK_GROUP_SYSTEM) &&
+					count_sys < bctl->sys.limit_min)) {
+			mutex_unlock(&fs_info->delete_unused_bgs_mutex);
 			goto loop;
 		}
 
+		if ((chunk_type & BTRFS_BLOCK_GROUP_DATA) && !chunk_reserved) {
+			trans = btrfs_start_transaction(chunk_root, 0);
+			if (IS_ERR(trans)) {
+				mutex_unlock(&fs_info->delete_unused_bgs_mutex);
+				ret = PTR_ERR(trans);
+				goto error;
+			}
+
+			ret = btrfs_force_chunk_alloc(trans, chunk_root,
+						      BTRFS_BLOCK_GROUP_DATA);
+			btrfs_end_transaction(trans, chunk_root);
+			if (ret < 0) {
+				mutex_unlock(&fs_info->delete_unused_bgs_mutex);
+				goto error;
+			}
+			chunk_reserved = 1;
+		}
+
 		ret = btrfs_relocate_chunk(chunk_root,
-					   found_key.objectid,
 					   found_key.offset);
+		mutex_unlock(&fs_info->delete_unused_bgs_mutex);
 		if (ret && ret != -ENOSPC)
 			goto error;
 		if (ret == -ENOSPC) {
@@ -3331,11 +3631,20 @@ static void __cancel_balance(struct btrfs_fs_info *fs_info)
 	unset_balance_control(fs_info);
 	ret = del_balance_item(fs_info->tree_root);
 	if (ret)
-		btrfs_std_error(fs_info, ret);
+		btrfs_std_error(fs_info, ret, NULL);
 
 	atomic_set(&fs_info->mutually_exclusive_operation_running, 0);
 }
 
+/* Non-zero return value signifies invalidity */
+static inline int validate_convert_profile(struct btrfs_balance_args *bctl_arg,
+		u64 allowed)
+{
+	return ((bctl_arg->flags & BTRFS_BALANCE_ARGS_CONVERT) &&
+		(!alloc_profile_is_valid(bctl_arg->target, 1) ||
+		 (bctl_arg->target & ~allowed)));
+}
+
 /*
  * Should be called with both balance and volume mutexes held
  */
@@ -3393,27 +3702,21 @@ int btrfs_balance(struct btrfs_balance_control *bctl,
 	if (num_devices > 3)
 		allowed |= (BTRFS_BLOCK_GROUP_RAID10 |
 			    BTRFS_BLOCK_GROUP_RAID6);
-	if ((bctl->data.flags & BTRFS_BALANCE_ARGS_CONVERT) &&
-	    (!alloc_profile_is_valid(bctl->data.target, 1) ||
-	     (bctl->data.target & ~allowed))) {
+	if (validate_convert_profile(&bctl->data, allowed)) {
 		btrfs_err(fs_info, "unable to start balance with target "
 			   "data profile %llu",
 		       bctl->data.target);
 		ret = -EINVAL;
 		goto out;
 	}
-	if ((bctl->meta.flags & BTRFS_BALANCE_ARGS_CONVERT) &&
-	    (!alloc_profile_is_valid(bctl->meta.target, 1) ||
-	     (bctl->meta.target & ~allowed))) {
+	if (validate_convert_profile(&bctl->meta, allowed)) {
 		btrfs_err(fs_info,
 			   "unable to start balance with target metadata profile %llu",
 		       bctl->meta.target);
 		ret = -EINVAL;
 		goto out;
 	}
-	if ((bctl->sys.flags & BTRFS_BALANCE_ARGS_CONVERT) &&
-	    (!alloc_profile_is_valid(bctl->sys.target, 1) ||
-	     (bctl->sys.target & ~allowed))) {
+	if (validate_convert_profile(&bctl->sys, allowed)) {
 		btrfs_err(fs_info,
 			   "unable to start balance with target system profile %llu",
 		       bctl->sys.target);
@@ -3455,23 +3758,10 @@ int btrfs_balance(struct btrfs_balance_control *bctl,
 	} while (read_seqretry(&fs_info->profiles_lock, seq));
 
 	if (bctl->sys.flags & BTRFS_BALANCE_ARGS_CONVERT) {
-		int num_tolerated_disk_barrier_failures;
-		u64 target = bctl->sys.target;
-
-		num_tolerated_disk_barrier_failures =
-			btrfs_calc_num_tolerated_disk_barrier_failures(fs_info);
-		if (num_tolerated_disk_barrier_failures > 0 &&
-		    (target &
-		     (BTRFS_BLOCK_GROUP_DUP | BTRFS_BLOCK_GROUP_RAID0 |
-		      BTRFS_AVAIL_ALLOC_BIT_SINGLE)))
-			num_tolerated_disk_barrier_failures = 0;
-		else if (num_tolerated_disk_barrier_failures > 1 &&
-			 (target &
-			  (BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID10)))
-			num_tolerated_disk_barrier_failures = 1;
-
-		fs_info->num_tolerated_disk_barrier_failures =
-			num_tolerated_disk_barrier_failures;
+		fs_info->num_tolerated_disk_barrier_failures = min(
+			btrfs_calc_num_tolerated_disk_barrier_failures(fs_info),
+			btrfs_get_num_tolerated_disk_barrier_failures(
+				bctl->sys.target));
 	}
 
 	ret = insert_balance_item(fs_info->tree_root, bctl);
@@ -3908,9 +4198,9 @@ int btrfs_create_uuid_tree(struct btrfs_fs_info *fs_info)
 	uuid_root = btrfs_create_tree(trans, fs_info,
 				      BTRFS_UUID_TREE_OBJECTID);
 	if (IS_ERR(uuid_root)) {
-		btrfs_abort_transaction(trans, tree_root,
-					PTR_ERR(uuid_root));
-		return PTR_ERR(uuid_root);
+		ret = PTR_ERR(uuid_root);
+		btrfs_abort_transaction(trans, tree_root, ret);
+		return ret;
 	}
 
 	fs_info->uuid_root = uuid_root;
@@ -3959,12 +4249,12 @@ int btrfs_shrink_device(struct btrfs_device *device, u64 new_size)
 	struct btrfs_dev_extent *dev_extent = NULL;
 	struct btrfs_path *path;
 	u64 length;
-	u64 chunk_objectid;
 	u64 chunk_offset;
 	int ret;
 	int slot;
 	int failed = 0;
 	bool retried = false;
+	bool checked_pending_chunks = false;
 	struct extent_buffer *l;
 	struct btrfs_key key;
 	struct btrfs_super_block *super_copy = root->fs_info->super_copy;
@@ -3998,11 +4288,16 @@ again:
 	key.type = BTRFS_DEV_EXTENT_KEY;
 
 	do {
+		mutex_lock(&root->fs_info->delete_unused_bgs_mutex);
 		ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
-		if (ret < 0)
+		if (ret < 0) {
+			mutex_unlock(&root->fs_info->delete_unused_bgs_mutex);
 			goto done;
+		}
 
 		ret = btrfs_previous_item(root, path, 0, key.type);
+		if (ret)
+			mutex_unlock(&root->fs_info->delete_unused_bgs_mutex);
 		if (ret < 0)
 			goto done;
 		if (ret) {
@@ -4016,6 +4311,7 @@ again:
 		btrfs_item_key_to_cpu(l, &key, path->slots[0]);
 
 		if (key.objectid != device->devid) {
+			mutex_unlock(&root->fs_info->delete_unused_bgs_mutex);
 			btrfs_release_path(path);
 			break;
 		}
@@ -4024,15 +4320,16 @@ again:
 		length = btrfs_dev_extent_length(l, dev_extent);
 
 		if (key.offset + length <= new_size) {
+			mutex_unlock(&root->fs_info->delete_unused_bgs_mutex);
 			btrfs_release_path(path);
 			break;
 		}
 
-		chunk_objectid = btrfs_dev_extent_chunk_objectid(l, dev_extent);
 		chunk_offset = btrfs_dev_extent_chunk_offset(l, dev_extent);
 		btrfs_release_path(path);
 
-		ret = btrfs_relocate_chunk(root, chunk_objectid, chunk_offset);
+		ret = btrfs_relocate_chunk(root, chunk_offset);
+		mutex_unlock(&root->fs_info->delete_unused_bgs_mutex);
 		if (ret && ret != -ENOSPC)
 			goto done;
 		if (ret == -ENOSPC)
@@ -4045,15 +4342,6 @@ again:
 		goto again;
 	} else if (failed && retried) {
 		ret = -ENOSPC;
-		lock_chunks(root);
-
-		btrfs_device_set_total_bytes(device, old_size);
-		if (device->writeable)
-			device->fs_devices->total_rw_bytes += diff;
-		spin_lock(&root->fs_info->free_chunk_lock);
-		root->fs_info->free_chunk_space += diff;
-		spin_unlock(&root->fs_info->free_chunk_lock);
-		unlock_chunks(root);
 		goto done;
 	}
 
@@ -4065,6 +4353,36 @@ again:
 	}
 
 	lock_chunks(root);
+
+	/*
+	 * We checked in the above loop all device extents that were already in
+	 * the device tree. However before we have updated the device's
+	 * total_bytes to the new size, we might have had chunk allocations that
+	 * have not complete yet (new block groups attached to transaction
+	 * handles), and therefore their device extents were not yet in the
+	 * device tree and we missed them in the loop above. So if we have any
+	 * pending chunk using a device extent that overlaps the device range
+	 * that we can not use anymore, commit the current transaction and
+	 * repeat the search on the device tree - this way we guarantee we will
+	 * not have chunks using device extents that end beyond 'new_size'.
+	 */
+	if (!checked_pending_chunks) {
+		u64 start = new_size;
+		u64 len = old_size - new_size;
+
+		if (contains_pending_extent(trans->transaction, device,
+					    &start, len)) {
+			unlock_chunks(root);
+			checked_pending_chunks = true;
+			failed = 0;
+			retried = false;
+			ret = btrfs_commit_transaction(trans, root);
+			if (ret)
+				goto done;
+			goto again;
+		}
+	}
+
 	btrfs_device_set_disk_total_bytes(device, new_size);
 	if (list_empty(&device->resized_list))
 		list_add_tail(&device->resized_list,
@@ -4079,6 +4397,16 @@ again:
 	btrfs_end_transaction(trans, root);
 done:
 	btrfs_free_path(path);
+	if (ret) {
+		lock_chunks(root);
+		btrfs_device_set_total_bytes(device, old_size);
+		if (device->writeable)
+			device->fs_devices->total_rw_bytes += diff;
+		spin_lock(&root->fs_info->free_chunk_lock);
+		root->fs_info->free_chunk_space += diff;
+		spin_unlock(&root->fs_info->free_chunk_lock);
+		unlock_chunks(root);
+	}
 	return ret;
 }
 
@@ -4130,65 +4458,6 @@ static int btrfs_cmp_device_info(const void *a, const void *b)
 	return 0;
 }
 
-static const struct btrfs_raid_attr btrfs_raid_array[BTRFS_NR_RAID_TYPES] = {
-	[BTRFS_RAID_RAID10] = {
-		.sub_stripes	= 2,
-		.dev_stripes	= 1,
-		.devs_max	= 0,	/* 0 == as many as possible */
-		.devs_min	= 4,
-		.devs_increment	= 2,
-		.ncopies	= 2,
-	},
-	[BTRFS_RAID_RAID1] = {
-		.sub_stripes	= 1,
-		.dev_stripes	= 1,
-		.devs_max	= 2,
-		.devs_min	= 2,
-		.devs_increment	= 2,
-		.ncopies	= 2,
-	},
-	[BTRFS_RAID_DUP] = {
-		.sub_stripes	= 1,
-		.dev_stripes	= 2,
-		.devs_max	= 1,
-		.devs_min	= 1,
-		.devs_increment	= 1,
-		.ncopies	= 2,
-	},
-	[BTRFS_RAID_RAID0] = {
-		.sub_stripes	= 1,
-		.dev_stripes	= 1,
-		.devs_max	= 0,
-		.devs_min	= 2,
-		.devs_increment	= 1,
-		.ncopies	= 1,
-	},
-	[BTRFS_RAID_SINGLE] = {
-		.sub_stripes	= 1,
-		.dev_stripes	= 1,
-		.devs_max	= 1,
-		.devs_min	= 1,
-		.devs_increment	= 1,
-		.ncopies	= 1,
-	},
-	[BTRFS_RAID_RAID5] = {
-		.sub_stripes	= 1,
-		.dev_stripes	= 1,
-		.devs_max	= 0,
-		.devs_min	= 2,
-		.devs_increment	= 1,
-		.ncopies	= 2,
-	},
-	[BTRFS_RAID_RAID6] = {
-		.sub_stripes	= 1,
-		.dev_stripes	= 1,
-		.devs_max	= 0,
-		.devs_min	= 3,
-		.devs_increment	= 1,
-		.ncopies	= 3,
-	},
-};
-
 static u32 find_raid56_stripe_len(u32 data_devices, u32 dev_stripe_target)
 {
 	/* TODO allow them to set a preferred stripe size */
@@ -4914,9 +5183,7 @@ static struct btrfs_bio *alloc_btrfs_bio(int total_stripes, int real_stripes)
 		 * and the stripes
 		 */
 		sizeof(u64) * (total_stripes),
-		GFP_NOFS);
-	if (!bbio)
-		return NULL;
+		GFP_NOFS|__GFP_NOFAIL);
 
 	atomic_set(&bbio->error, 0);
 	atomic_set(&bbio->refs, 1);
@@ -5584,26 +5851,26 @@ int btrfs_rmap_block(struct btrfs_mapping_tree *map_tree,
 	return 0;
 }
 
-static inline void btrfs_end_bbio(struct btrfs_bio *bbio, struct bio *bio, int err)
+static inline void btrfs_end_bbio(struct btrfs_bio *bbio, struct bio *bio)
 {
-	if (likely(bbio->flags & BTRFS_BIO_ORIG_BIO_SUBMITTED))
-		bio_endio_nodec(bio, err);
-	else
-		bio_endio(bio, err);
+	bio->bi_private = bbio->private;
+	bio->bi_end_io = bbio->end_io;
+	bio_endio(bio);
+
 	btrfs_put_bbio(bbio);
 }
 
-static void btrfs_end_bio(struct bio *bio, int err)
+static void btrfs_end_bio(struct bio *bio)
 {
 	struct btrfs_bio *bbio = bio->bi_private;
-	struct btrfs_device *dev = bbio->stripes[0].dev;
 	int is_orig_bio = 0;
 
-	if (err) {
+	if (bio->bi_error) {
 		atomic_inc(&bbio->error);
-		if (err == -EIO || err == -EREMOTEIO) {
+		if (bio->bi_error == -EIO || bio->bi_error == -EREMOTEIO) {
 			unsigned int stripe_index =
 				btrfs_io_bio(bio)->stripe_index;
+			struct btrfs_device *dev;
 
 			BUG_ON(stripe_index >= bbio->num_stripes);
 			dev = bbio->stripes[stripe_index].dev;
@@ -5633,24 +5900,21 @@ static void btrfs_end_bio(struct bio *bio, int err)
 			bio = bbio->orig_bio;
 		}
 
-		bio->bi_private = bbio->private;
-		bio->bi_end_io = bbio->end_io;
 		btrfs_io_bio(bio)->mirror_num = bbio->mirror_num;
 		/* only send an error to the higher layers if it is
 		 * beyond the tolerance of the btrfs bio
 		 */
 		if (atomic_read(&bbio->error) > bbio->max_errors) {
-			err = -EIO;
+			bio->bi_error = -EIO;
 		} else {
 			/*
 			 * this bio is actually up to date, we didn't
 			 * go over the max number of errors
 			 */
-			set_bit(BIO_UPTODATE, &bio->bi_flags);
-			err = 0;
+			bio->bi_error = 0;
 		}
 
-		btrfs_end_bbio(bbio, bio, err);
+		btrfs_end_bbio(bbio, bio);
 	} else if (!is_orig_bio) {
 		bio_put(bio);
 	}
@@ -5671,7 +5935,7 @@ static noinline void btrfs_schedule_bio(struct btrfs_root *root,
 	struct btrfs_pending_bios *pending_bios;
 
 	if (device->missing || !device->bdev) {
-		bio_endio(bio, -EIO);
+		bio_io_error(bio);
 		return;
 	}
 
@@ -5716,34 +5980,6 @@ static noinline void btrfs_schedule_bio(struct btrfs_root *root,
 				 &device->work);
 }
 
-static int bio_size_ok(struct block_device *bdev, struct bio *bio,
-		       sector_t sector)
-{
-	struct bio_vec *prev;
-	struct request_queue *q = bdev_get_queue(bdev);
-	unsigned int max_sectors = queue_max_sectors(q);
-	struct bvec_merge_data bvm = {
-		.bi_bdev = bdev,
-		.bi_sector = sector,
-		.bi_rw = bio->bi_rw,
-	};
-
-	if (WARN_ON(bio->bi_vcnt == 0))
-		return 1;
-
-	prev = &bio->bi_io_vec[bio->bi_vcnt - 1];
-	if (bio_sectors(bio) > max_sectors)
-		return 0;
-
-	if (!q->merge_bvec_fn)
-		return 1;
-
-	bvm.bi_size = bio->bi_iter.bi_size - prev->bv_len;
-	if (q->merge_bvec_fn(q, &bvm, prev) < prev->bv_len)
-		return 0;
-	return 1;
-}
-
 static void submit_stripe_bio(struct btrfs_root *root, struct btrfs_bio *bbio,
 			      struct bio *bio, u64 physical, int dev_nr,
 			      int rw, int async)
@@ -5777,38 +6013,6 @@ static void submit_stripe_bio(struct btrfs_root *root, struct btrfs_bio *bbio,
 		btrfsic_submit_bio(rw, bio);
 }
 
-static int breakup_stripe_bio(struct btrfs_root *root, struct btrfs_bio *bbio,
-			      struct bio *first_bio, struct btrfs_device *dev,
-			      int dev_nr, int rw, int async)
-{
-	struct bio_vec *bvec = first_bio->bi_io_vec;
-	struct bio *bio;
-	int nr_vecs = bio_get_nr_vecs(dev->bdev);
-	u64 physical = bbio->stripes[dev_nr].physical;
-
-again:
-	bio = btrfs_bio_alloc(dev->bdev, physical >> 9, nr_vecs, GFP_NOFS);
-	if (!bio)
-		return -ENOMEM;
-
-	while (bvec <= (first_bio->bi_io_vec + first_bio->bi_vcnt - 1)) {
-		if (bio_add_page(bio, bvec->bv_page, bvec->bv_len,
-				 bvec->bv_offset) < bvec->bv_len) {
-			u64 len = bio->bi_iter.bi_size;
-
-			atomic_inc(&bbio->stripes_pending);
-			submit_stripe_bio(root, bbio, bio, physical, dev_nr,
-					  rw, async);
-			physical += len;
-			goto again;
-		}
-		bvec++;
-	}
-
-	submit_stripe_bio(root, bbio, bio, physical, dev_nr, rw, async);
-	return 0;
-}
-
 static void bbio_error(struct btrfs_bio *bbio, struct bio *bio, u64 logical)
 {
 	atomic_inc(&bbio->error);
@@ -5816,12 +6020,10 @@ static void bbio_error(struct btrfs_bio *bbio, struct bio *bio, u64 logical)
 		/* Shoud be the original bio. */
 		WARN_ON(bio != bbio->orig_bio);
 
-		bio->bi_private = bbio->private;
-		bio->bi_end_io = bbio->end_io;
 		btrfs_io_bio(bio)->mirror_num = bbio->mirror_num;
 		bio->bi_iter.bi_sector = logical >> 9;
-
-		btrfs_end_bbio(bbio, bio, -EIO);
+		bio->bi_error = -EIO;
+		btrfs_end_bbio(bbio, bio);
 	}
 }
 
@@ -5883,25 +6085,11 @@ int btrfs_map_bio(struct btrfs_root *root, int rw, struct bio *bio,
 			continue;
 		}
 
-		/*
-		 * Check and see if we're ok with this bio based on it's size
-		 * and offset with the given device.
-		 */
-		if (!bio_size_ok(dev->bdev, first_bio,
-				 bbio->stripes[dev_nr].physical >> 9)) {
-			ret = breakup_stripe_bio(root, bbio, first_bio, dev,
-						 dev_nr, rw, async_submit);
-			BUG_ON(ret);
-			continue;
-		}
-
 		if (dev_nr < total_devs - 1) {
 			bio = btrfs_bio_clone(first_bio, GFP_NOFS);
 			BUG_ON(!bio); /* -ENOMEM */
-		} else {
+		} else
 			bio = first_bio;
-			bbio->flags |= BTRFS_BIO_ORIG_BIO_SUBMITTED;
-		}
 
 		submit_stripe_bio(root, bbio, bio,
 				  bbio->stripes[dev_nr].physical, dev_nr, rw,
@@ -6078,6 +6266,8 @@ static int read_one_chunk(struct btrfs_root *root, struct btrfs_key *key,
 				free_extent_map(em);
 				return -EIO;
 			}
+			btrfs_warn(root->fs_info, "devid %llu uuid %pU is missing",
+						devid, uuid);
 		}
 		map->stripes[i].dev->in_fs_metadata = 1;
 	}
@@ -6197,10 +6387,11 @@ static int read_one_dev(struct btrfs_root *root,
 		if (!btrfs_test_opt(root, DEGRADED))
 			return -EIO;
 
-		btrfs_warn(root->fs_info, "devid %llu missing", devid);
 		device = add_missing_dev(root, fs_devices, devid, dev_uuid);
 		if (!device)
 			return -ENOMEM;
+		btrfs_warn(root->fs_info, "devid %llu uuid %pU missing",
+				devid, dev_uuid);
 	} else {
 		if (!device->bdev && !btrfs_test_opt(root, DEGRADED))
 			return -EIO;
@@ -6322,6 +6513,14 @@ int btrfs_read_sys_array(struct btrfs_root *root)
 				goto out_short_read;
 
 			num_stripes = btrfs_chunk_num_stripes(sb, chunk);
+			if (!num_stripes) {
+				printk(KERN_ERR
+	    "BTRFS: invalid number of stripes %u in sys_array at offset %u\n",
+					num_stripes, cur_offset);
+				ret = -EIO;
+				break;
+			}
+
 			len = btrfs_chunk_item_size(num_stripes);
 			if (cur_offset + len > array_size)
 				goto out_short_read;
@@ -6517,8 +6716,8 @@ static int update_dev_stat_item(struct btrfs_trans_handle *trans,
 	BUG_ON(!path);
 	ret = btrfs_search_slot(trans, dev_root, &key, path, -1, 1);
 	if (ret < 0) {
-		printk_in_rcu(KERN_WARNING "BTRFS: "
-			"error %d while searching for dev_stats item for device %s!\n",
+		btrfs_warn_in_rcu(dev_root->fs_info,
+			"error %d while searching for dev_stats item for device %s",
 			      ret, rcu_str_deref(device->name));
 		goto out;
 	}
@@ -6528,8 +6727,8 @@ static int update_dev_stat_item(struct btrfs_trans_handle *trans,
 		/* need to delete old one and insert a new one */
 		ret = btrfs_del_item(trans, dev_root, path);
 		if (ret != 0) {
-			printk_in_rcu(KERN_WARNING "BTRFS: "
-				"delete too small dev_stats item for device %s failed %d!\n",
+			btrfs_warn_in_rcu(dev_root->fs_info,
+				"delete too small dev_stats item for device %s failed %d",
 				      rcu_str_deref(device->name), ret);
 			goto out;
 		}
@@ -6542,9 +6741,9 @@ static int update_dev_stat_item(struct btrfs_trans_handle *trans,
 		ret = btrfs_insert_empty_item(trans, dev_root, path,
 					      &key, sizeof(*ptr));
 		if (ret < 0) {
-			printk_in_rcu(KERN_WARNING "BTRFS: "
-					  "insert dev_stats item for device %s failed %d!\n",
-				      rcu_str_deref(device->name), ret);
+			btrfs_warn_in_rcu(dev_root->fs_info,
+				"insert dev_stats item for device %s failed %d",
+				rcu_str_deref(device->name), ret);
 			goto out;
 		}
 	}
@@ -6598,8 +6797,8 @@ static void btrfs_dev_stat_print_on_error(struct btrfs_device *dev)
 {
 	if (!dev->dev_stats_valid)
 		return;
-	printk_ratelimited_in_rcu(KERN_ERR "BTRFS: "
-			   "bdev %s errs: wr %u, rd %u, flush %u, corrupt %u, gen %u\n",
+	btrfs_err_rl_in_rcu(dev->dev_root->fs_info,
+		"bdev %s errs: wr %u, rd %u, flush %u, corrupt %u, gen %u",
 			   rcu_str_deref(dev->name),
 			   btrfs_dev_stat_read(dev, BTRFS_DEV_STAT_WRITE_ERRS),
 			   btrfs_dev_stat_read(dev, BTRFS_DEV_STAT_READ_ERRS),
@@ -6618,8 +6817,8 @@ static void btrfs_dev_stat_print_on_load(struct btrfs_device *dev)
 	if (i == BTRFS_DEV_STAT_VALUES_MAX)
 		return; /* all values == 0, suppress message */
 
-	printk_in_rcu(KERN_INFO "BTRFS: "
-		   "bdev %s errs: wr %u, rd %u, flush %u, corrupt %u, gen %u\n",
+	btrfs_info_in_rcu(dev->dev_root->fs_info,
+		"bdev %s errs: wr %u, rd %u, flush %u, corrupt %u, gen %u",
 	       rcu_str_deref(dev->name),
 	       btrfs_dev_stat_read(dev, BTRFS_DEV_STAT_WRITE_ERRS),
 	       btrfs_dev_stat_read(dev, BTRFS_DEV_STAT_READ_ERRS),
@@ -6663,22 +6862,34 @@ int btrfs_get_dev_stats(struct btrfs_root *root,
 	return 0;
 }
 
-int btrfs_scratch_superblock(struct btrfs_device *device)
+void btrfs_scratch_superblocks(struct block_device *bdev, char *device_path)
 {
 	struct buffer_head *bh;
 	struct btrfs_super_block *disk_super;
+	int copy_num;
 
-	bh = btrfs_read_dev_super(device->bdev);
-	if (!bh)
-		return -EINVAL;
-	disk_super = (struct btrfs_super_block *)bh->b_data;
+	if (!bdev)
+		return;
 
-	memset(&disk_super->magic, 0, sizeof(disk_super->magic));
-	set_buffer_dirty(bh);
-	sync_dirty_buffer(bh);
-	brelse(bh);
+	for (copy_num = 0; copy_num < BTRFS_SUPER_MIRROR_MAX;
+		copy_num++) {
 
-	return 0;
+		if (btrfs_read_dev_one_super(bdev, copy_num, &bh))
+			continue;
+
+		disk_super = (struct btrfs_super_block *)bh->b_data;
+
+		memset(&disk_super->magic, 0, sizeof(disk_super->magic));
+		set_buffer_dirty(bh);
+		sync_dirty_buffer(bh);
+		brelse(bh);
+	}
+
+	/* Notify udev that device has changed */
+	btrfs_kobject_uevent(bdev, KOBJ_CHANGE);
+
+	/* Update ctime/mtime for device path for libblkid */
+	update_dev_time(device_path);
 }
 
 /*
@@ -6728,3 +6939,56 @@ void btrfs_update_commit_device_bytes_used(struct btrfs_root *root,
 	}
 	unlock_chunks(root);
 }
+
+void btrfs_set_fs_info_ptr(struct btrfs_fs_info *fs_info)
+{
+	struct btrfs_fs_devices *fs_devices = fs_info->fs_devices;
+	while (fs_devices) {
+		fs_devices->fs_info = fs_info;
+		fs_devices = fs_devices->seed;
+	}
+}
+
+void btrfs_reset_fs_info_ptr(struct btrfs_fs_info *fs_info)
+{
+	struct btrfs_fs_devices *fs_devices = fs_info->fs_devices;
+	while (fs_devices) {
+		fs_devices->fs_info = NULL;
+		fs_devices = fs_devices->seed;
+	}
+}
+
+void btrfs_close_one_device(struct btrfs_device *device)
+{
+	struct btrfs_fs_devices *fs_devices = device->fs_devices;
+	struct btrfs_device *new_device;
+	struct rcu_string *name;
+
+	if (device->bdev)
+		fs_devices->open_devices--;
+
+	if (device->writeable &&
+	    device->devid != BTRFS_DEV_REPLACE_DEVID) {
+		list_del_init(&device->dev_alloc_list);
+		fs_devices->rw_devices--;
+	}
+
+	if (device->missing)
+		fs_devices->missing_devices--;
+
+	new_device = btrfs_alloc_device(NULL, &device->devid,
+					device->uuid);
+	BUG_ON(IS_ERR(new_device)); /* -ENOMEM */
+
+	/* Safe because we are under uuid_mutex */
+	if (device->name) {
+		name = rcu_string_strdup(device->name->str, GFP_NOFS);
+		BUG_ON(!name); /* -ENOMEM */
+		rcu_assign_pointer(new_device->name, name);
+	}
+
+	list_replace_rcu(&device->dev_list, &new_device->dev_list);
+	new_device->fs_devices = device->fs_devices;
+
+	call_rcu(&device->rcu, free_device);
+}
diff --git a/kernel/fs/btrfs/volumes.h b/kernel/fs/btrfs/volumes.h
index ebc31331a..d5c84f6b1 100644
--- a/kernel/fs/btrfs/volumes.h
+++ b/kernel/fs/btrfs/volumes.h
@@ -253,6 +253,12 @@ struct btrfs_fs_devices {
 	 * nonrot flag set
 	 */
 	int rotating;
+
+	struct btrfs_fs_info *fs_info;
+	/* sysfs kobjects */
+	struct kobject fsid_kobj;
+	struct kobject *device_dir_kobj;
+	struct completion kobj_unregister;
 };
 
 #define BTRFS_BIO_INLINE_CSUM_SIZE	64
@@ -292,8 +298,6 @@ struct btrfs_bio_stripe {
 struct btrfs_bio;
 typedef void (btrfs_bio_end_io_t) (struct btrfs_bio *bio, int err);
 
-#define BTRFS_BIO_ORIG_BIO_SUBMITTED	(1 << 0)
-
 struct btrfs_bio {
 	atomic_t refs;
 	atomic_t stripes_pending;
@@ -330,10 +334,15 @@ struct btrfs_raid_attr {
 	int dev_stripes;	/* stripes per dev */
 	int devs_max;		/* max devs to use */
 	int devs_min;		/* min devs needed */
+	int tolerated_failures; /* max tolerated fail devs */
 	int devs_increment;	/* ndevs has to be a multiple of this */
 	int ncopies;		/* how many copies to data has */
 };
 
+extern const struct btrfs_raid_attr btrfs_raid_array[BTRFS_NR_RAID_TYPES];
+
+extern const u64 btrfs_raid_group[BTRFS_NR_RAID_TYPES];
+
 struct map_lookup {
 	u64 type;
 	int io_align;
@@ -371,6 +380,20 @@ struct map_lookup {
 #define BTRFS_BALANCE_ARGS_DRANGE	(1ULL << 3)
 #define BTRFS_BALANCE_ARGS_VRANGE	(1ULL << 4)
 #define BTRFS_BALANCE_ARGS_LIMIT	(1ULL << 5)
+#define BTRFS_BALANCE_ARGS_LIMIT_RANGE	(1ULL << 6)
+#define BTRFS_BALANCE_ARGS_STRIPES_RANGE (1ULL << 7)
+#define BTRFS_BALANCE_ARGS_USAGE_RANGE	(1ULL << 10)
+
+#define BTRFS_BALANCE_ARGS_MASK			\
+	(BTRFS_BALANCE_ARGS_PROFILES |		\
+	 BTRFS_BALANCE_ARGS_USAGE |		\
+	 BTRFS_BALANCE_ARGS_DEVID | 		\
+	 BTRFS_BALANCE_ARGS_DRANGE |		\
+	 BTRFS_BALANCE_ARGS_VRANGE |		\
+	 BTRFS_BALANCE_ARGS_LIMIT |		\
+	 BTRFS_BALANCE_ARGS_LIMIT_RANGE |	\
+	 BTRFS_BALANCE_ARGS_STRIPES_RANGE |	\
+	 BTRFS_BALANCE_ARGS_USAGE_RANGE)
 
 /*
  * Profile changing flags.  When SOFT is set we won't relocate chunk if
@@ -449,6 +472,9 @@ int btrfs_cancel_balance(struct btrfs_fs_info *fs_info);
 int btrfs_create_uuid_tree(struct btrfs_fs_info *fs_info);
 int btrfs_check_uuid_tree(struct btrfs_fs_info *fs_info);
 int btrfs_chunk_readonly(struct btrfs_root *root, u64 chunk_offset);
+int find_free_dev_extent_start(struct btrfs_transaction *transaction,
+			 struct btrfs_device *device, u64 num_bytes,
+			 u64 search_start, u64 *start, u64 *max_avail);
 int find_free_dev_extent(struct btrfs_trans_handle *trans,
 			 struct btrfs_device *device, u64 num_bytes,
 			 u64 *start, u64 *max_avail);
@@ -467,7 +493,7 @@ void btrfs_destroy_dev_replace_tgtdev(struct btrfs_fs_info *fs_info,
 				      struct btrfs_device *tgtdev);
 void btrfs_init_dev_replace_tgtdev_for_resume(struct btrfs_fs_info *fs_info,
 					      struct btrfs_device *tgtdev);
-int btrfs_scratch_superblock(struct btrfs_device *device);
+void btrfs_scratch_superblocks(struct block_device *bdev, char *device_path);
 int btrfs_is_parity_mirror(struct btrfs_mapping_tree *map_tree,
 			   u64 logical, u64 len, int mirror_num);
 unsigned long btrfs_full_stripe_len(struct btrfs_root *root,
@@ -537,5 +563,9 @@ static inline void unlock_chunks(struct btrfs_root *root)
 	mutex_unlock(&root->fs_info->chunk_mutex);
 }
 
+struct list_head *btrfs_get_fs_uuids(void);
+void btrfs_set_fs_info_ptr(struct btrfs_fs_info *fs_info);
+void btrfs_reset_fs_info_ptr(struct btrfs_fs_info *fs_info);
+void btrfs_close_one_device(struct btrfs_device *device);
 
 #endif
diff --git a/kernel/fs/btrfs/xattr.c b/kernel/fs/btrfs/xattr.c
index 6f518c90e..1fcd7b6e7 100644
--- a/kernel/fs/btrfs/xattr.c
+++ b/kernel/fs/btrfs/xattr.c
@@ -313,8 +313,10 @@ ssize_t btrfs_listxattr(struct dentry *dentry, char *buffer, size_t size)
 		/* check to make sure this item is what we want */
 		if (found_key.objectid != key.objectid)
 			break;
-		if (found_key.type != BTRFS_XATTR_ITEM_KEY)
+		if (found_key.type > BTRFS_XATTR_ITEM_KEY)
 			break;
+		if (found_key.type < BTRFS_XATTR_ITEM_KEY)
+			goto next;
 
 		di = btrfs_item_ptr(leaf, slot, struct btrfs_dir_item);
 		if (verify_dir_item(root, leaf, di))