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

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/btrfs-unstable: (58 commits)
  Btrfs: use the device_list_mutex during write_dev_supers
  Btrfs: setup free ino caching in a more asynchronous way
  btrfs scrub: don't coalesce pages that are logically discontiguous
  Btrfs: return -ENOMEM in clear_extent_bit
  Btrfs: add mount -o auto_defrag
  Btrfs: using rcu lock in the reader side of devices list
  Btrfs: drop unnecessary device lock
  Btrfs: fix the race between remove dev and alloc chunk
  Btrfs: fix the race between reading and updating devices
  Btrfs: fix bh leak on __btrfs_open_devices path
  Btrfs: fix unsafe usage of merge_state
  Btrfs: allocate extent state and check the result properly
  fs/btrfs: Add missing btrfs_free_path
  Btrfs: check return value of btrfs_inc_extent_ref()
  Btrfs: return error to caller if read_one_inode() fails
  Btrfs: BUG_ON is deleted from the caller of btrfs_truncate_item & btrfs_extend_item
  Btrfs: return error code to caller when btrfs_del_item fails
  Btrfs: return error code to caller when btrfs_previous_item fails
  btrfs: fix typo 'testeing' -> 'testing'
  btrfs: typo: 'btrfS' -> 'btrfs'
  ...

fs/btrfs/Makefile

@@ -7,4 +7,4 @@ btrfs-y += super.o ctree.o extent-tree.o print-tree.o root-tree.o dir-item.o \
 	   extent_map.o sysfs.o struct-funcs.o xattr.o ordered-data.o \
 	   extent_io.o volumes.o async-thread.o ioctl.o locking.o orphan.o \
 	   export.o tree-log.o acl.o free-space-cache.o zlib.o lzo.o \
-	   compression.o delayed-ref.o relocation.o
+	   compression.o delayed-ref.o relocation.o delayed-inode.o scrub.o

fs/btrfs/acl.c

@@ -288,7 +288,7 @@ int btrfs_acl_chmod(struct inode *inode)
 		return 0;
 
 	acl = btrfs_get_acl(inode, ACL_TYPE_ACCESS);
-	if (IS_ERR(acl) || !acl)
+	if (IS_ERR_OR_NULL(acl))
 		return PTR_ERR(acl);
 
 	clone = posix_acl_clone(acl, GFP_KERNEL);

fs/btrfs/btrfs_inode.h

@@ -22,6 +22,7 @@
 #include "extent_map.h"
 #include "extent_io.h"
 #include "ordered-data.h"
+#include "delayed-inode.h"
 
 /* in memory btrfs inode */
 struct btrfs_inode {
@@ -152,20 +153,34 @@ struct btrfs_inode {
 	unsigned ordered_data_close:1;
 	unsigned orphan_meta_reserved:1;
 	unsigned dummy_inode:1;
+	unsigned in_defrag:1;
 
 	/*
 	 * always compress this one file
 	 */
 	unsigned force_compress:4;
 
+	struct btrfs_delayed_node *delayed_node;
+
 	struct inode vfs_inode;
 };
 
+extern unsigned char btrfs_filetype_table[];
+
 static inline struct btrfs_inode *BTRFS_I(struct inode *inode)
 {
 	return container_of(inode, struct btrfs_inode, vfs_inode);
 }
 
+static inline u64 btrfs_ino(struct inode *inode)
+{
+	u64 ino = BTRFS_I(inode)->location.objectid;
+
+	if (ino <= BTRFS_FIRST_FREE_OBJECTID)
+		ino = inode->i_ino;
+	return ino;
+}
+
 static inline void btrfs_i_size_write(struct inode *inode, u64 size)
 {
 	i_size_write(inode, size);

fs/btrfs/compression.c

@@ -125,9 +125,10 @@ static int check_compressed_csum(struct inode *inode,
 		kunmap_atomic(kaddr, KM_USER0);
 
 		if (csum != *cb_sum) {
-			printk(KERN_INFO "btrfs csum failed ino %lu "
+			printk(KERN_INFO "btrfs csum failed ino %llu "
 			       "extent %llu csum %u "
-			       "wanted %u mirror %d\n", inode->i_ino,
+			       "wanted %u mirror %d\n",
+			       (unsigned long long)btrfs_ino(inode),
 			       (unsigned long long)disk_start,
 			       csum, *cb_sum, cb->mirror_num);
 			ret = -EIO;
@@ -332,7 +333,7 @@ int btrfs_submit_compressed_write(struct inode *inode, u64 start,
 	struct compressed_bio *cb;
 	unsigned long bytes_left;
 	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
-	int page_index = 0;
+	int pg_index = 0;
 	struct page *page;
 	u64 first_byte = disk_start;
 	struct block_device *bdev;
@@ -366,8 +367,8 @@ int btrfs_submit_compressed_write(struct inode *inode, u64 start,
 
 	/* create and submit bios for the compressed pages */
 	bytes_left = compressed_len;
-	for (page_index = 0; page_index < cb->nr_pages; page_index++) {
-		page = compressed_pages[page_index];
+	for (pg_index = 0; pg_index < cb->nr_pages; pg_index++) {
+		page = compressed_pages[pg_index];
 		page->mapping = inode->i_mapping;
 		if (bio->bi_size)
 			ret = io_tree->ops->merge_bio_hook(page, 0,
@@ -432,7 +433,7 @@ static noinline int add_ra_bio_pages(struct inode *inode,
 				     struct compressed_bio *cb)
 {
 	unsigned long end_index;
-	unsigned long page_index;
+	unsigned long pg_index;
 	u64 last_offset;
 	u64 isize = i_size_read(inode);
 	int ret;
@@ -456,13 +457,13 @@ static noinline int add_ra_bio_pages(struct inode *inode,
 	end_index = (i_size_read(inode) - 1) >> PAGE_CACHE_SHIFT;
 
 	while (last_offset < compressed_end) {
-		page_index = last_offset >> PAGE_CACHE_SHIFT;
+		pg_index = last_offset >> PAGE_CACHE_SHIFT;
 
-		if (page_index > end_index)
+		if (pg_index > end_index)
 			break;
 
 		rcu_read_lock();
-		page = radix_tree_lookup(&mapping->page_tree, page_index);
+		page = radix_tree_lookup(&mapping->page_tree, pg_index);
 		rcu_read_unlock();
 		if (page) {
 			misses++;
@@ -476,7 +477,7 @@ static noinline int add_ra_bio_pages(struct inode *inode,
 		if (!page)
 			break;
 
-		if (add_to_page_cache_lru(page, mapping, page_index,
+		if (add_to_page_cache_lru(page, mapping, pg_index,
 								GFP_NOFS)) {
 			page_cache_release(page);
 			goto next;
@@ -560,7 +561,7 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio,
 	unsigned long uncompressed_len = bio->bi_vcnt * PAGE_CACHE_SIZE;
 	unsigned long compressed_len;
 	unsigned long nr_pages;
-	unsigned long page_index;
+	unsigned long pg_index;
 	struct page *page;
 	struct block_device *bdev;
 	struct bio *comp_bio;
@@ -613,10 +614,10 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio,
 
 	bdev = BTRFS_I(inode)->root->fs_info->fs_devices->latest_bdev;
 
-	for (page_index = 0; page_index < nr_pages; page_index++) {
-		cb->compressed_pages[page_index] = alloc_page(GFP_NOFS |
+	for (pg_index = 0; pg_index < nr_pages; pg_index++) {
+		cb->compressed_pages[pg_index] = alloc_page(GFP_NOFS |
 							      __GFP_HIGHMEM);
-		if (!cb->compressed_pages[page_index])
+		if (!cb->compressed_pages[pg_index])
 			goto fail2;
 	}
 	cb->nr_pages = nr_pages;
@@ -634,8 +635,8 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio,
 	comp_bio->bi_end_io = end_compressed_bio_read;
 	atomic_inc(&cb->pending_bios);
 
-	for (page_index = 0; page_index < nr_pages; page_index++) {
-		page = cb->compressed_pages[page_index];
+	for (pg_index = 0; pg_index < nr_pages; pg_index++) {
+		page = cb->compressed_pages[pg_index];
 		page->mapping = inode->i_mapping;
 		page->index = em_start >> PAGE_CACHE_SHIFT;
 
@@ -702,8 +703,8 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio,
 	return 0;
 
 fail2:
-	for (page_index = 0; page_index < nr_pages; page_index++)
-		free_page((unsigned long)cb->compressed_pages[page_index]);
+	for (pg_index = 0; pg_index < nr_pages; pg_index++)
+		free_page((unsigned long)cb->compressed_pages[pg_index]);
 
 	kfree(cb->compressed_pages);
 fail1:
@@ -945,7 +946,7 @@ void btrfs_exit_compress(void)
 int btrfs_decompress_buf2page(char *buf, unsigned long buf_start,
 			      unsigned long total_out, u64 disk_start,
 			      struct bio_vec *bvec, int vcnt,
-			      unsigned long *page_index,
+			      unsigned long *pg_index,
 			      unsigned long *pg_offset)
 {
 	unsigned long buf_offset;
@@ -954,7 +955,7 @@ int btrfs_decompress_buf2page(char *buf, unsigned long buf_start,
 	unsigned long working_bytes = total_out - buf_start;
 	unsigned long bytes;
 	char *kaddr;
-	struct page *page_out = bvec[*page_index].bv_page;
+	struct page *page_out = bvec[*pg_index].bv_page;
 
 	/*
 	 * start byte is the first byte of the page we're currently
@@ -995,11 +996,11 @@ int btrfs_decompress_buf2page(char *buf, unsigned long buf_start,
 
 		/* check if we need to pick another page */
 		if (*pg_offset == PAGE_CACHE_SIZE) {
-			(*page_index)++;
-			if (*page_index >= vcnt)
+			(*pg_index)++;
+			if (*pg_index >= vcnt)
 				return 0;
 
-			page_out = bvec[*page_index].bv_page;
+			page_out = bvec[*pg_index].bv_page;
 			*pg_offset = 0;
 			start_byte = page_offset(page_out) - disk_start;
 

fs/btrfs/compression.h

@@ -37,7 +37,7 @@ int btrfs_decompress(int type, unsigned char *data_in, struct page *dest_page,
 int btrfs_decompress_buf2page(char *buf, unsigned long buf_start,
 			      unsigned long total_out, u64 disk_start,
 			      struct bio_vec *bvec, int vcnt,
-			      unsigned long *page_index,
+			      unsigned long *pg_index,
 			      unsigned long *pg_offset);
 
 int btrfs_submit_compressed_write(struct inode *inode, u64 start,

fs/btrfs/ctree.c

@@ -38,11 +38,6 @@ static int balance_node_right(struct btrfs_trans_handle *trans,
 			      struct extent_buffer *src_buf);
 static int del_ptr(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 		   struct btrfs_path *path, int level, int slot);
-static int setup_items_for_insert(struct btrfs_trans_handle *trans,
-			struct btrfs_root *root, struct btrfs_path *path,
-			struct btrfs_key *cpu_key, u32 *data_size,
-			u32 total_data, u32 total_size, int nr);
-
 
 struct btrfs_path *btrfs_alloc_path(void)
 {
@@ -107,7 +102,7 @@ void btrfs_free_path(struct btrfs_path *p)
 {
 	if (!p)
 		return;
-	btrfs_release_path(NULL, p);
+	btrfs_release_path(p);
 	kmem_cache_free(btrfs_path_cachep, p);
 }
 
@@ -117,7 +112,7 @@ void btrfs_free_path(struct btrfs_path *p)
  *
  * It is safe to call this on paths that no locks or extent buffers held.
  */
-noinline void btrfs_release_path(struct btrfs_root *root, struct btrfs_path *p)
+noinline void btrfs_release_path(struct btrfs_path *p)
 {
 	int i;
 
@@ -1328,7 +1323,7 @@ static noinline int reada_for_balance(struct btrfs_root *root,
 		ret = -EAGAIN;
 
 		/* release the whole path */
-		btrfs_release_path(root, path);
+		btrfs_release_path(path);
 
 		/* read the blocks */
 		if (block1)
@@ -1475,7 +1470,7 @@ read_block_for_search(struct btrfs_trans_handle *trans,
 				return 0;
 			}
 			free_extent_buffer(tmp);
-			btrfs_release_path(NULL, p);
+			btrfs_release_path(p);
 			return -EIO;
 		}
 	}
@@ -1494,7 +1489,7 @@ read_block_for_search(struct btrfs_trans_handle *trans,
 	if (p->reada)
 		reada_for_search(root, p, level, slot, key->objectid);
 
-	btrfs_release_path(NULL, p);
+	btrfs_release_path(p);
 
 	ret = -EAGAIN;
 	tmp = read_tree_block(root, blocknr, blocksize, 0);
@@ -1563,7 +1558,7 @@ setup_nodes_for_search(struct btrfs_trans_handle *trans,
 		}
 		b = p->nodes[level];
 		if (!b) {
-			btrfs_release_path(NULL, p);
+			btrfs_release_path(p);
 			goto again;
 		}
 		BUG_ON(btrfs_header_nritems(b) == 1);
@@ -1753,7 +1748,7 @@ done:
 	if (!p->leave_spinning)
 		btrfs_set_path_blocking(p);
 	if (ret < 0)
-		btrfs_release_path(root, p);
+		btrfs_release_path(p);
 	return ret;
 }
 
@@ -3026,7 +3021,7 @@ static noinline int setup_leaf_for_split(struct btrfs_trans_handle *trans,
 				    struct btrfs_file_extent_item);
 		extent_len = btrfs_file_extent_num_bytes(leaf, fi);
 	}
-	btrfs_release_path(root, path);
+	btrfs_release_path(path);
 
 	path->keep_locks = 1;
 	path->search_for_split = 1;
@@ -3216,7 +3211,6 @@ int btrfs_truncate_item(struct btrfs_trans_handle *trans,
 			struct btrfs_path *path,
 			u32 new_size, int from_end)
 {
-	int ret = 0;
 	int slot;
 	struct extent_buffer *leaf;
 	struct btrfs_item *item;
@@ -3314,12 +3308,11 @@ int btrfs_truncate_item(struct btrfs_trans_handle *trans,
 	btrfs_set_item_size(leaf, item, new_size);
 	btrfs_mark_buffer_dirty(leaf);
 
-	ret = 0;
 	if (btrfs_leaf_free_space(root, leaf) < 0) {
 		btrfs_print_leaf(root, leaf);
 		BUG();
 	}
-	return ret;
+	return 0;
 }
 
 /*
@@ -3329,7 +3322,6 @@ int btrfs_extend_item(struct btrfs_trans_handle *trans,
 		      struct btrfs_root *root, struct btrfs_path *path,
 		      u32 data_size)
 {
-	int ret = 0;
 	int slot;
 	struct extent_buffer *leaf;
 	struct btrfs_item *item;
@@ -3394,12 +3386,11 @@ int btrfs_extend_item(struct btrfs_trans_handle *trans,
 	btrfs_set_item_size(leaf, item, old_size + data_size);
 	btrfs_mark_buffer_dirty(leaf);
 
-	ret = 0;
 	if (btrfs_leaf_free_space(root, leaf) < 0) {
 		btrfs_print_leaf(root, leaf);
 		BUG();
 	}
-	return ret;
+	return 0;
 }
 
 /*
@@ -3559,11 +3550,10 @@ out:
  * to save stack depth by doing the bulk of the work in a function
  * that doesn't call btrfs_search_slot
  */
-static noinline_for_stack int
-setup_items_for_insert(struct btrfs_trans_handle *trans,
-		      struct btrfs_root *root, struct btrfs_path *path,
-		      struct btrfs_key *cpu_key, u32 *data_size,
-		      u32 total_data, u32 total_size, int nr)
+int setup_items_for_insert(struct btrfs_trans_handle *trans,
+			   struct btrfs_root *root, struct btrfs_path *path,
+			   struct btrfs_key *cpu_key, u32 *data_size,
+			   u32 total_data, u32 total_size, int nr)
 {
 	struct btrfs_item *item;
 	int i;
@@ -3647,7 +3637,6 @@ setup_items_for_insert(struct btrfs_trans_handle *trans,
 
 	ret = 0;
 	if (slot == 0) {
-		struct btrfs_disk_key disk_key;
 		btrfs_cpu_key_to_disk(&disk_key, cpu_key);
 		ret = fixup_low_keys(trans, root, path, &disk_key, 1);
 	}
@@ -3949,7 +3938,7 @@ int btrfs_prev_leaf(struct btrfs_root *root, struct btrfs_path *path)
 	else
 		return 1;
 
-	btrfs_release_path(root, path);
+	btrfs_release_path(path);
 	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
 	if (ret < 0)
 		return ret;
@@ -4073,7 +4062,7 @@ find_next_key:
 			sret = btrfs_find_next_key(root, path, min_key, level,
 						  cache_only, min_trans);
 			if (sret == 0) {
-				btrfs_release_path(root, path);
+				btrfs_release_path(path);
 				goto again;
 			} else {
 				goto out;
@@ -4152,7 +4141,7 @@ next:
 				btrfs_node_key_to_cpu(c, &cur_key, slot);
 
 			orig_lowest = path->lowest_level;
-			btrfs_release_path(root, path);
+			btrfs_release_path(path);
 			path->lowest_level = level;
 			ret = btrfs_search_slot(NULL, root, &cur_key, path,
 						0, 0);
@@ -4229,7 +4218,7 @@ int btrfs_next_leaf(struct btrfs_root *root, struct btrfs_path *path)
 again:
 	level = 1;
 	next = NULL;
-	btrfs_release_path(root, path);
+	btrfs_release_path(path);
 
 	path->keep_locks = 1;
 
@@ -4285,7 +4274,7 @@ again:
 			goto again;
 
 		if (ret < 0) {
-			btrfs_release_path(root, path);
+			btrfs_release_path(path);
 			goto done;
 		}
 
@@ -4324,7 +4313,7 @@ again:
 			goto again;
 
 		if (ret < 0) {
-			btrfs_release_path(root, path);
+			btrfs_release_path(path);
 			goto done;
 		}
 

fs/btrfs/ctree.h

@@ -23,6 +23,7 @@
 #include <linux/mm.h>
 #include <linux/highmem.h>
 #include <linux/fs.h>
+#include <linux/rwsem.h>
 #include <linux/completion.h>
 #include <linux/backing-dev.h>
 #include <linux/wait.h>
@@ -33,6 +34,7 @@
 #include "extent_io.h"
 #include "extent_map.h"
 #include "async-thread.h"
+#include "ioctl.h"
 
 struct btrfs_trans_handle;
 struct btrfs_transaction;
@@ -105,6 +107,12 @@ struct btrfs_ordered_sum;
 /* For storing free space cache */
 #define BTRFS_FREE_SPACE_OBJECTID -11ULL
 
+/*
+ * The inode number assigned to the special inode for sotring
+ * free ino cache
+ */
+#define BTRFS_FREE_INO_OBJECTID -12ULL
+
 /* dummy objectid represents multiple objectids */
 #define BTRFS_MULTIPLE_OBJECTIDS -255ULL
 
@@ -187,7 +195,6 @@ struct btrfs_mapping_tree {
 	struct extent_map_tree map_tree;
 };
 
-#define BTRFS_UUID_SIZE 16
 struct btrfs_dev_item {
 	/* the internal btrfs device id */
 	__le64 devid;
@@ -294,7 +301,6 @@ static inline unsigned long btrfs_chunk_item_size(int num_stripes)
 		sizeof(struct btrfs_stripe) * (num_stripes - 1);
 }
 
-#define BTRFS_FSID_SIZE 16
 #define BTRFS_HEADER_FLAG_WRITTEN	(1ULL << 0)
 #define BTRFS_HEADER_FLAG_RELOC		(1ULL << 1)
 
@@ -510,6 +516,12 @@ struct btrfs_extent_item_v0 {
 /* use full backrefs for extent pointers in the block */
 #define BTRFS_BLOCK_FLAG_FULL_BACKREF	(1ULL << 8)
 
+/*
+ * this flag is only used internally by scrub and may be changed at any time
+ * it is only declared here to avoid collisions
+ */
+#define BTRFS_EXTENT_FLAG_SUPER		(1ULL << 48)
+
 struct btrfs_tree_block_info {
 	struct btrfs_disk_key key;
 	u8 level;
@@ -740,12 +752,12 @@ struct btrfs_space_info {
 	 */
 	unsigned long reservation_progress;
 
-	int full:1;		/* indicates that we cannot allocate any more
+	unsigned int full:1;	/* indicates that we cannot allocate any more
 				   chunks for this space */
-	int chunk_alloc:1;	/* set if we are allocating a chunk */
+	unsigned int chunk_alloc:1;	/* set if we are allocating a chunk */
 
-	int force_alloc;	/* set if we need to force a chunk alloc for
-				   this space */
+	unsigned int force_alloc;	/* set if we need to force a chunk
+					   alloc for this space */
 
 	struct list_head list;
 
@@ -830,9 +842,6 @@ struct btrfs_block_group_cache {
 	u64 bytes_super;
 	u64 flags;
 	u64 sectorsize;
-	int extents_thresh;
-	int free_extents;
-	int total_bitmaps;
 	unsigned int ro:1;
 	unsigned int dirty:1;
 	unsigned int iref:1;
@@ -847,9 +856,7 @@ struct btrfs_block_group_cache {
 	struct btrfs_space_info *space_info;
 
 	/* free space cache stuff */
-	spinlock_t tree_lock;
-	struct rb_root free_space_offset;
-	u64 free_space;
+	struct btrfs_free_space_ctl *free_space_ctl;
 
 	/* block group cache stuff */
 	struct rb_node cache_node;
@@ -869,6 +876,7 @@ struct btrfs_block_group_cache {
 struct reloc_control;
 struct btrfs_device;
 struct btrfs_fs_devices;
+struct btrfs_delayed_root;
 struct btrfs_fs_info {
 	u8 fsid[BTRFS_FSID_SIZE];
 	u8 chunk_tree_uuid[BTRFS_UUID_SIZE];
@@ -895,7 +903,10 @@ struct btrfs_fs_info {
 	/* logical->physical extent mapping */
 	struct btrfs_mapping_tree mapping_tree;
 
-	/* block reservation for extent, checksum and root tree */
+	/*
+	 * block reservation for extent, checksum, root tree and
+	 * delayed dir index item
+	 */
 	struct btrfs_block_rsv global_block_rsv;
 	/* block reservation for delay allocation */
 	struct btrfs_block_rsv delalloc_block_rsv;
@@ -1022,6 +1033,7 @@ struct btrfs_fs_info {
 	 * for the sys_munmap function call path
 	 */
 	struct btrfs_workers fixup_workers;
+	struct btrfs_workers delayed_workers;
 	struct task_struct *transaction_kthread;
 	struct task_struct *cleaner_kthread;
 	int thread_pool_size;
@@ -1062,6 +1074,11 @@ struct btrfs_fs_info {
 	/* all metadata allocations go through this cluster */
 	struct btrfs_free_cluster meta_alloc_cluster;
 
+	/* auto defrag inodes go here */
+	spinlock_t defrag_inodes_lock;
+	struct rb_root defrag_inodes;
+	atomic_t defrag_running;
+
 	spinlock_t ref_cache_lock;
 	u64 total_ref_cache_size;
 
@@ -1077,8 +1094,21 @@ struct btrfs_fs_info {
 
 	void *bdev_holder;
 
+	/* private scrub information */
+	struct mutex scrub_lock;
+	atomic_t scrubs_running;
+	atomic_t scrub_pause_req;
+	atomic_t scrubs_paused;
+	atomic_t scrub_cancel_req;
+	wait_queue_head_t scrub_pause_wait;
+	struct rw_semaphore scrub_super_lock;
+	int scrub_workers_refcnt;
+	struct btrfs_workers scrub_workers;
+
 	/* filesystem state */
 	u64 fs_state;
+
+	struct btrfs_delayed_root *delayed_root;
 };
 
 /*
@@ -1088,9 +1118,6 @@ struct btrfs_fs_info {
 struct btrfs_root {
 	struct extent_buffer *node;
 
-	/* the node lock is held while changing the node pointer */
-	spinlock_t node_lock;
-
 	struct extent_buffer *commit_root;
 	struct btrfs_root *log_root;
 	struct btrfs_root *reloc_root;
@@ -1107,6 +1134,16 @@ struct btrfs_root {
 	spinlock_t accounting_lock;
 	struct btrfs_block_rsv *block_rsv;
 
+	/* free ino cache stuff */
+	struct mutex fs_commit_mutex;
+	struct btrfs_free_space_ctl *free_ino_ctl;
+	enum btrfs_caching_type cached;
+	spinlock_t cache_lock;
+	wait_queue_head_t cache_wait;
+	struct btrfs_free_space_ctl *free_ino_pinned;
+	u64 cache_progress;
+	struct inode *cache_inode;
+
 	struct mutex log_mutex;
 	wait_queue_head_t log_writer_wait;
 	wait_queue_head_t log_commit_wait[2];
@@ -1161,6 +1198,11 @@ struct btrfs_root {
 	/* red-black tree that keeps track of in-memory inodes */
 	struct rb_root inode_tree;
 
+	/*
+	 * radix tree that keeps track of delayed nodes of every inode,
+	 * protected by inode_lock
+	 */
+	struct radix_tree_root delayed_nodes_tree;
 	/*
 	 * right now this just gets used so that a root has its own devid
 	 * for stat.  It may be used for more later
@@ -1168,6 +1210,38 @@ struct btrfs_root {
 	struct super_block anon_super;
 };
 
+struct btrfs_ioctl_defrag_range_args {
+	/* start of the defrag operation */
+	__u64 start;
+
+	/* number of bytes to defrag, use (u64)-1 to say all */
+	__u64 len;
+
+	/*
+	 * flags for the operation, which can include turning
+	 * on compression for this one defrag
+	 */
+	__u64 flags;
+
+	/*
+	 * any extent bigger than this will be considered
+	 * already defragged.  Use 0 to take the kernel default
+	 * Use 1 to say every single extent must be rewritten
+	 */
+	__u32 extent_thresh;
+
+	/*
+	 * which compression method to use if turning on compression
+	 * for this defrag operation.  If unspecified, zlib will
+	 * be used
+	 */
+	__u32 compress_type;
+
+	/* spare for later */
+	__u32 unused[4];
+};
+
+
 /*
  * inode items have the data typically returned from stat and store other
  * info about object characteristics.  There is one for every file and dir in
@@ -1265,6 +1339,7 @@ struct btrfs_root {
 #define BTRFS_MOUNT_CLEAR_CACHE		(1 << 13)
 #define BTRFS_MOUNT_USER_SUBVOL_RM_ALLOWED (1 << 14)
 #define BTRFS_MOUNT_ENOSPC_DEBUG	 (1 << 15)
+#define BTRFS_MOUNT_AUTO_DEFRAG		(1 << 16)
 
 #define btrfs_clear_opt(o, opt)		((o) &= ~BTRFS_MOUNT_##opt)
 #define btrfs_set_opt(o, opt)		((o) |= BTRFS_MOUNT_##opt)
@@ -1440,26 +1515,12 @@ static inline u64 btrfs_stripe_offset_nr(struct extent_buffer *eb,
 	return btrfs_stripe_offset(eb, btrfs_stripe_nr(c, nr));
 }
 
-static inline void btrfs_set_stripe_offset_nr(struct extent_buffer *eb,
-					     struct btrfs_chunk *c, int nr,
-					     u64 val)
-{
-	btrfs_set_stripe_offset(eb, btrfs_stripe_nr(c, nr), val);
-}
-
 static inline u64 btrfs_stripe_devid_nr(struct extent_buffer *eb,
 					 struct btrfs_chunk *c, int nr)
 {
 	return btrfs_stripe_devid(eb, btrfs_stripe_nr(c, nr));
 }
 
-static inline void btrfs_set_stripe_devid_nr(struct extent_buffer *eb,
-					     struct btrfs_chunk *c, int nr,
-					     u64 val)
-{
-	btrfs_set_stripe_devid(eb, btrfs_stripe_nr(c, nr), val);
-}
-
 /* struct btrfs_block_group_item */
 BTRFS_SETGET_STACK_FUNCS(block_group_used, struct btrfs_block_group_item,
 			 used, 64);
@@ -1517,14 +1578,6 @@ btrfs_inode_ctime(struct btrfs_inode_item *inode_item)
 	return (struct btrfs_timespec *)ptr;
 }
 
-static inline struct btrfs_timespec *
-btrfs_inode_otime(struct btrfs_inode_item *inode_item)
-{
-	unsigned long ptr = (unsigned long)inode_item;
-	ptr += offsetof(struct btrfs_inode_item, otime);
-	return (struct btrfs_timespec *)ptr;
-}
-
 BTRFS_SETGET_FUNCS(timespec_sec, struct btrfs_timespec, sec, 64);
 BTRFS_SETGET_FUNCS(timespec_nsec, struct btrfs_timespec, nsec, 32);
 
@@ -1875,33 +1928,6 @@ static inline u8 *btrfs_header_chunk_tree_uuid(struct extent_buffer *eb)
 	return (u8 *)ptr;
 }
 
-static inline u8 *btrfs_super_fsid(struct extent_buffer *eb)
-{
-	unsigned long ptr = offsetof(struct btrfs_super_block, fsid);
-	return (u8 *)ptr;
-}
-
-static inline u8 *btrfs_header_csum(struct extent_buffer *eb)
-{
-	unsigned long ptr = offsetof(struct btrfs_header, csum);
-	return (u8 *)ptr;
-}
-
-static inline struct btrfs_node *btrfs_buffer_node(struct extent_buffer *eb)
-{
-	return NULL;
-}
-
-static inline struct btrfs_leaf *btrfs_buffer_leaf(struct extent_buffer *eb)
-{
-	return NULL;
-}
-
-static inline struct btrfs_header *btrfs_buffer_header(struct extent_buffer *eb)
-{
-	return NULL;
-}
-
 static inline int btrfs_is_leaf(struct extent_buffer *eb)
 {
 	return btrfs_header_level(eb) == 0;
@@ -2055,22 +2081,6 @@ static inline struct btrfs_root *btrfs_sb(struct super_block *sb)
 	return sb->s_fs_info;
 }
 
-static inline int btrfs_set_root_name(struct btrfs_root *root,
-				      const char *name, int len)
-{
-	/* if we already have a name just free it */
-	kfree(root->name);
-
-	root->name = kmalloc(len+1, GFP_KERNEL);
-	if (!root->name)
-		return -ENOMEM;
-
-	memcpy(root->name, name, len);
-	root->name[len] = '\0';
-
-	return 0;
-}
-
 static inline u32 btrfs_level_size(struct btrfs_root *root, int level)
 {
 	if (level == 0)
@@ -2099,6 +2109,13 @@ static inline bool btrfs_mixed_space_info(struct btrfs_space_info *space_info)
 }
 
 /* extent-tree.c */
+static inline u64 btrfs_calc_trans_metadata_size(struct btrfs_root *root,
+						 int num_items)
+{
+	return (root->leafsize + root->nodesize * (BTRFS_MAX_LEVEL - 1)) *
+		3 * num_items;
+}
+
 void btrfs_put_block_group(struct btrfs_block_group_cache *cache);
 int btrfs_run_delayed_refs(struct btrfs_trans_handle *trans,
 			   struct btrfs_root *root, unsigned long count);
@@ -2108,12 +2125,9 @@ int btrfs_lookup_extent_info(struct btrfs_trans_handle *trans,
 			     u64 num_bytes, u64 *refs, u64 *flags);
 int btrfs_pin_extent(struct btrfs_root *root,
 		     u64 bytenr, u64 num, int reserved);
-int btrfs_drop_leaf_ref(struct btrfs_trans_handle *trans,
-			struct btrfs_root *root, struct extent_buffer *leaf);
 int btrfs_cross_ref_exist(struct btrfs_trans_handle *trans,
 			  struct btrfs_root *root,
 			  u64 objectid, u64 offset, u64 bytenr);
-int btrfs_copy_pinned(struct btrfs_root *root, struct extent_io_tree *copy);
 struct btrfs_block_group_cache *btrfs_lookup_block_group(
 						 struct btrfs_fs_info *info,
 						 u64 bytenr);
@@ -2290,10 +2304,12 @@ int btrfs_realloc_node(struct btrfs_trans_handle *trans,
 		       struct btrfs_root *root, struct extent_buffer *parent,
 		       int start_slot, int cache_only, u64 *last_ret,
 		       struct btrfs_key *progress);
-void btrfs_release_path(struct btrfs_root *root, struct btrfs_path *p);
+void btrfs_release_path(struct btrfs_path *p);
 struct btrfs_path *btrfs_alloc_path(void);
 void btrfs_free_path(struct btrfs_path *p);
 void btrfs_set_path_blocking(struct btrfs_path *p);
+void btrfs_clear_path_blocking(struct btrfs_path *p,
+			       struct extent_buffer *held);
 void btrfs_unlock_up_safe(struct btrfs_path *p, int level);
 
 int btrfs_del_items(struct btrfs_trans_handle *trans, struct btrfs_root *root,
@@ -2305,13 +2321,12 @@ static inline int btrfs_del_item(struct btrfs_trans_handle *trans,
 	return btrfs_del_items(trans, root, path, path->slots[0], 1);
 }
 
+int setup_items_for_insert(struct btrfs_trans_handle *trans,
+			   struct btrfs_root *root, struct btrfs_path *path,
+			   struct btrfs_key *cpu_key, u32 *data_size,
+			   u32 total_data, u32 total_size, int nr);
 int btrfs_insert_item(struct btrfs_trans_handle *trans, struct btrfs_root
 		      *root, struct btrfs_key *key, void *data, u32 data_size);
-int btrfs_insert_some_items(struct btrfs_trans_handle *trans,
-			    struct btrfs_root *root,
-			    struct btrfs_path *path,
-			    struct btrfs_key *cpu_key, u32 *data_size,
-			    int nr);
 int btrfs_insert_empty_items(struct btrfs_trans_handle *trans,
 			     struct btrfs_root *root,
 			     struct btrfs_path *path,
@@ -2357,8 +2372,6 @@ int btrfs_update_root(struct btrfs_trans_handle *trans, struct btrfs_root
 		      *item);
 int btrfs_find_last_root(struct btrfs_root *root, u64 objectid, struct
 			 btrfs_root_item *item, struct btrfs_key *key);
-int btrfs_search_root(struct btrfs_root *root, u64 search_start,
-		      u64 *found_objectid);
 int btrfs_find_dead_roots(struct btrfs_root *root, u64 objectid);
 int btrfs_find_orphan_roots(struct btrfs_root *tree_root);
 int btrfs_set_root_node(struct btrfs_root_item *item,
@@ -2368,7 +2381,7 @@ void btrfs_check_and_init_root_item(struct btrfs_root_item *item);
 /* dir-item.c */
 int btrfs_insert_dir_item(struct btrfs_trans_handle *trans,
 			  struct btrfs_root *root, const char *name,
-			  int name_len, u64 dir,
+			  int name_len, struct inode *dir,
 			  struct btrfs_key *location, u8 type, u64 index);
 struct btrfs_dir_item *btrfs_lookup_dir_item(struct btrfs_trans_handle *trans,
 					     struct btrfs_root *root,
@@ -2413,12 +2426,6 @@ int btrfs_del_orphan_item(struct btrfs_trans_handle *trans,
 			  struct btrfs_root *root, u64 offset);
 int btrfs_find_orphan_item(struct btrfs_root *root, u64 offset);
 
-/* inode-map.c */
-int btrfs_find_free_objectid(struct btrfs_trans_handle *trans,
-			     struct btrfs_root *fs_root,
-			     u64 dirid, u64 *objectid);
-int btrfs_find_highest_inode(struct btrfs_root *fs_root, u64 *objectid);
-
 /* inode-item.c */
 int btrfs_insert_inode_ref(struct btrfs_trans_handle *trans,
 			   struct btrfs_root *root,
@@ -2463,8 +2470,6 @@ int btrfs_csum_file_blocks(struct btrfs_trans_handle *trans,
 			   struct btrfs_ordered_sum *sums);
 int btrfs_csum_one_bio(struct btrfs_root *root, struct inode *inode,
 		       struct bio *bio, u64 file_start, int contig);
-int btrfs_csum_file_bytes(struct btrfs_root *root, struct inode *inode,
-			  u64 start, unsigned long len);
 struct btrfs_csum_item *btrfs_lookup_csum(struct btrfs_trans_handle *trans,
 					  struct btrfs_root *root,
 					  struct btrfs_path *path,
@@ -2472,8 +2477,8 @@ struct btrfs_csum_item *btrfs_lookup_csum(struct btrfs_trans_handle *trans,
 int btrfs_csum_truncate(struct btrfs_trans_handle *trans,
 			struct btrfs_root *root, struct btrfs_path *path,
 			u64 isize);
-int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start,
-			     u64 end, struct list_head *list);
+int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, u64 end,
+			     struct list_head *list, int search_commit);
 /* inode.c */
 
 /* RHEL and EL kernels have a patch that renames PG_checked to FsMisc */
@@ -2502,8 +2507,6 @@ int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
 			       u32 min_type);
 
 int btrfs_start_delalloc_inodes(struct btrfs_root *root, int delay_iput);
-int btrfs_start_one_delalloc_inode(struct btrfs_root *root, int delay_iput,
-				   int sync);
 int btrfs_set_extent_delalloc(struct inode *inode, u64 start, u64 end,
 			      struct extent_state **cached_state);
 int btrfs_writepages(struct address_space *mapping,
@@ -2520,7 +2523,6 @@ unsigned long btrfs_force_ra(struct address_space *mapping,
 int btrfs_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf);
 int btrfs_readpage(struct file *file, struct page *page);
 void btrfs_evict_inode(struct inode *inode);
-void btrfs_put_inode(struct inode *inode);
 int btrfs_write_inode(struct inode *inode, struct writeback_control *wbc);
 void btrfs_dirty_inode(struct inode *inode);
 struct inode *btrfs_alloc_inode(struct super_block *sb);
@@ -2531,10 +2533,8 @@ void btrfs_destroy_cachep(void);
 long btrfs_ioctl_trans_end(struct file *file);
 struct inode *btrfs_iget(struct super_block *s, struct btrfs_key *location,
 			 struct btrfs_root *root, int *was_new);
-int btrfs_commit_write(struct file *file, struct page *page,
-		       unsigned from, unsigned to);
 struct extent_map *btrfs_get_extent(struct inode *inode, struct page *page,
-				    size_t page_offset, u64 start, u64 end,
+				    size_t pg_offset, u64 start, u64 end,
 				    int create);
 int btrfs_update_inode(struct btrfs_trans_handle *trans,
 			      struct btrfs_root *root,
@@ -2566,12 +2566,16 @@ extern const struct dentry_operations btrfs_dentry_operations;
 long btrfs_ioctl(struct file *file, unsigned int cmd, unsigned long arg);
 void btrfs_update_iflags(struct inode *inode);
 void btrfs_inherit_iflags(struct inode *inode, struct inode *dir);
-
+int btrfs_defrag_file(struct inode *inode, struct file *file,
+		      struct btrfs_ioctl_defrag_range_args *range,
+		      u64 newer_than, unsigned long max_pages);
 /* file.c */
+int btrfs_add_inode_defrag(struct btrfs_trans_handle *trans,
+			   struct inode *inode);
+int btrfs_run_defrag_inodes(struct btrfs_fs_info *fs_info);
 int btrfs_sync_file(struct file *file, int datasync);
 int btrfs_drop_extent_cache(struct inode *inode, u64 start, u64 end,
 			    int skip_pinned);
-int btrfs_check_file(struct btrfs_root *root, struct inode *inode);
 extern const struct file_operations btrfs_file_operations;
 int btrfs_drop_extents(struct btrfs_trans_handle *trans, struct inode *inode,
 		       u64 start, u64 end, u64 *hint_byte, int drop_cache);
@@ -2591,10 +2595,6 @@ 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_super(struct btrfs_fs_info *fs);
-int btrfs_sysfs_add_root(struct btrfs_root *root);
-void btrfs_sysfs_del_root(struct btrfs_root *root);
-void btrfs_sysfs_del_super(struct btrfs_fs_info *root);
 
 /* xattr.c */
 ssize_t btrfs_listxattr(struct dentry *dentry, char *buffer, size_t size);
@@ -2637,4 +2637,18 @@ void btrfs_reloc_pre_snapshot(struct btrfs_trans_handle *trans,
 			      u64 *bytes_to_reserve);
 void btrfs_reloc_post_snapshot(struct btrfs_trans_handle *trans,
 			      struct btrfs_pending_snapshot *pending);
+
+/* scrub.c */
+int btrfs_scrub_dev(struct btrfs_root *root, u64 devid, u64 start, u64 end,
+		    struct btrfs_scrub_progress *progress, int readonly);
+int btrfs_scrub_pause(struct btrfs_root *root);
+int btrfs_scrub_pause_super(struct btrfs_root *root);
+int btrfs_scrub_continue(struct btrfs_root *root);
+int btrfs_scrub_continue_super(struct btrfs_root *root);
+int btrfs_scrub_cancel(struct btrfs_root *root);
+int btrfs_scrub_cancel_dev(struct btrfs_root *root, struct btrfs_device *dev);
+int btrfs_scrub_cancel_devid(struct btrfs_root *root, u64 devid);
+int btrfs_scrub_progress(struct btrfs_root *root, u64 devid,
+			 struct btrfs_scrub_progress *progress);
+
 #endif

fs/btrfs/delayed-inode.c

@@ -0,0 +1,1695 @@
+/*
+ * Copyright (C) 2011 Fujitsu.  All rights reserved.
+ * Written by Miao Xie <miaox@cn.fujitsu.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include <linux/slab.h>
+#include "delayed-inode.h"
+#include "disk-io.h"
+#include "transaction.h"
+
+#define BTRFS_DELAYED_WRITEBACK		400
+#define BTRFS_DELAYED_BACKGROUND	100
+
+static struct kmem_cache *delayed_node_cache;
+
+int __init btrfs_delayed_inode_init(void)
+{
+	delayed_node_cache = kmem_cache_create("delayed_node",
+					sizeof(struct btrfs_delayed_node),
+					0,
+					SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD,
+					NULL);
+	if (!delayed_node_cache)
+		return -ENOMEM;
+	return 0;
+}
+
+void btrfs_delayed_inode_exit(void)
+{
+	if (delayed_node_cache)
+		kmem_cache_destroy(delayed_node_cache);
+}
+
+static inline void btrfs_init_delayed_node(
+				struct btrfs_delayed_node *delayed_node,
+				struct btrfs_root *root, u64 inode_id)
+{
+	delayed_node->root = root;
+	delayed_node->inode_id = inode_id;
+	atomic_set(&delayed_node->refs, 0);
+	delayed_node->count = 0;
+	delayed_node->in_list = 0;
+	delayed_node->inode_dirty = 0;
+	delayed_node->ins_root = RB_ROOT;
+	delayed_node->del_root = RB_ROOT;
+	mutex_init(&delayed_node->mutex);
+	delayed_node->index_cnt = 0;
+	INIT_LIST_HEAD(&delayed_node->n_list);
+	INIT_LIST_HEAD(&delayed_node->p_list);
+	delayed_node->bytes_reserved = 0;
+}
+
+static inline int btrfs_is_continuous_delayed_item(
+					struct btrfs_delayed_item *item1,
+					struct btrfs_delayed_item *item2)
+{
+	if (item1->key.type == BTRFS_DIR_INDEX_KEY &&
+	    item1->key.objectid == item2->key.objectid &&
+	    item1->key.type == item2->key.type &&
+	    item1->key.offset + 1 == item2->key.offset)
+		return 1;
+	return 0;
+}
+
+static inline struct btrfs_delayed_root *btrfs_get_delayed_root(
+							struct btrfs_root *root)
+{
+	return root->fs_info->delayed_root;
+}
+
+static struct btrfs_delayed_node *btrfs_get_or_create_delayed_node(
+							struct inode *inode)
+{
+	struct btrfs_delayed_node *node;
+	struct btrfs_inode *btrfs_inode = BTRFS_I(inode);
+	struct btrfs_root *root = btrfs_inode->root;
+	u64 ino = btrfs_ino(inode);
+	int ret;
+
+again:
+	node = ACCESS_ONCE(btrfs_inode->delayed_node);
+	if (node) {
+		atomic_inc(&node->refs);	/* can be accessed */
+		return node;
+	}
+
+	spin_lock(&root->inode_lock);
+	node = radix_tree_lookup(&root->delayed_nodes_tree, ino);
+	if (node) {
+		if (btrfs_inode->delayed_node) {
+			spin_unlock(&root->inode_lock);
+			goto again;
+		}
+		btrfs_inode->delayed_node = node;
+		atomic_inc(&node->refs);	/* can be accessed */
+		atomic_inc(&node->refs);	/* cached in the inode */
+		spin_unlock(&root->inode_lock);
+		return node;
+	}
+	spin_unlock(&root->inode_lock);
+
+	node = kmem_cache_alloc(delayed_node_cache, GFP_NOFS);
+	if (!node)
+		return ERR_PTR(-ENOMEM);
+	btrfs_init_delayed_node(node, root, ino);
+
+	atomic_inc(&node->refs);	/* cached in the btrfs inode */
+	atomic_inc(&node->refs);	/* can be accessed */
+
+	ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM);
+	if (ret) {
+		kmem_cache_free(delayed_node_cache, node);
+		return ERR_PTR(ret);
+	}
+
+	spin_lock(&root->inode_lock);
+	ret = radix_tree_insert(&root->delayed_nodes_tree, ino, node);
+	if (ret == -EEXIST) {
+		kmem_cache_free(delayed_node_cache, node);
+		spin_unlock(&root->inode_lock);
+		radix_tree_preload_end();
+		goto again;
+	}
+	btrfs_inode->delayed_node = node;
+	spin_unlock(&root->inode_lock);
+	radix_tree_preload_end();
+
+	return node;
+}
+
+/*
+ * Call it when holding delayed_node->mutex
+ *
+ * If mod = 1, add this node into the prepared list.
+ */
+static void btrfs_queue_delayed_node(struct btrfs_delayed_root *root,
+				     struct btrfs_delayed_node *node,
+				     int mod)
+{
+	spin_lock(&root->lock);
+	if (node->in_list) {
+		if (!list_empty(&node->p_list))
+			list_move_tail(&node->p_list, &root->prepare_list);
+		else if (mod)
+			list_add_tail(&node->p_list, &root->prepare_list);
+	} else {
+		list_add_tail(&node->n_list, &root->node_list);
+		list_add_tail(&node->p_list, &root->prepare_list);
+		atomic_inc(&node->refs);	/* inserted into list */
+		root->nodes++;
+		node->in_list = 1;
+	}
+	spin_unlock(&root->lock);
+}
+
+/* Call it when holding delayed_node->mutex */
+static void btrfs_dequeue_delayed_node(struct btrfs_delayed_root *root,
+				       struct btrfs_delayed_node *node)
+{
+	spin_lock(&root->lock);
+	if (node->in_list) {
+		root->nodes--;
+		atomic_dec(&node->refs);	/* not in the list */
+		list_del_init(&node->n_list);
+		if (!list_empty(&node->p_list))
+			list_del_init(&node->p_list);
+		node->in_list = 0;
+	}
+	spin_unlock(&root->lock);
+}
+
+struct btrfs_delayed_node *btrfs_first_delayed_node(
+			struct btrfs_delayed_root *delayed_root)
+{
+	struct list_head *p;
+	struct btrfs_delayed_node *node = NULL;
+
+	spin_lock(&delayed_root->lock);
+	if (list_empty(&delayed_root->node_list))
+		goto out;
+
+	p = delayed_root->node_list.next;
+	node = list_entry(p, struct btrfs_delayed_node, n_list);
+	atomic_inc(&node->refs);
+out:
+	spin_unlock(&delayed_root->lock);
+
+	return node;
+}
+
+struct btrfs_delayed_node *btrfs_next_delayed_node(
+						struct btrfs_delayed_node *node)
+{
+	struct btrfs_delayed_root *delayed_root;
+	struct list_head *p;
+	struct btrfs_delayed_node *next = NULL;
+
+	delayed_root = node->root->fs_info->delayed_root;
+	spin_lock(&delayed_root->lock);
+	if (!node->in_list) {	/* not in the list */
+		if (list_empty(&delayed_root->node_list))
+			goto out;
+		p = delayed_root->node_list.next;
+	} else if (list_is_last(&node->n_list, &delayed_root->node_list))
+		goto out;
+	else
+		p = node->n_list.next;
+
+	next = list_entry(p, struct btrfs_delayed_node, n_list);
+	atomic_inc(&next->refs);
+out:
+	spin_unlock(&delayed_root->lock);
+
+	return next;
+}
+
+static void __btrfs_release_delayed_node(
+				struct btrfs_delayed_node *delayed_node,
+				int mod)
+{
+	struct btrfs_delayed_root *delayed_root;
+
+	if (!delayed_node)
+		return;
+
+	delayed_root = delayed_node->root->fs_info->delayed_root;
+
+	mutex_lock(&delayed_node->mutex);
+	if (delayed_node->count)
+		btrfs_queue_delayed_node(delayed_root, delayed_node, mod);
+	else
+		btrfs_dequeue_delayed_node(delayed_root, delayed_node);
+	mutex_unlock(&delayed_node->mutex);
+
+	if (atomic_dec_and_test(&delayed_node->refs)) {
+		struct btrfs_root *root = delayed_node->root;
+		spin_lock(&root->inode_lock);
+		if (atomic_read(&delayed_node->refs) == 0) {
+			radix_tree_delete(&root->delayed_nodes_tree,
+					  delayed_node->inode_id);
+			kmem_cache_free(delayed_node_cache, delayed_node);
+		}
+		spin_unlock(&root->inode_lock);
+	}
+}
+
+static inline void btrfs_release_delayed_node(struct btrfs_delayed_node *node)
+{
+	__btrfs_release_delayed_node(node, 0);
+}
+
+struct btrfs_delayed_node *btrfs_first_prepared_delayed_node(
+					struct btrfs_delayed_root *delayed_root)
+{
+	struct list_head *p;
+	struct btrfs_delayed_node *node = NULL;
+
+	spin_lock(&delayed_root->lock);
+	if (list_empty(&delayed_root->prepare_list))
+		goto out;
+
+	p = delayed_root->prepare_list.next;
+	list_del_init(p);
+	node = list_entry(p, struct btrfs_delayed_node, p_list);
+	atomic_inc(&node->refs);
+out:
+	spin_unlock(&delayed_root->lock);
+
+	return node;
+}
+
+static inline void btrfs_release_prepared_delayed_node(
+					struct btrfs_delayed_node *node)
+{
+	__btrfs_release_delayed_node(node, 1);
+}
+
+struct btrfs_delayed_item *btrfs_alloc_delayed_item(u32 data_len)
+{
+	struct btrfs_delayed_item *item;
+	item = kmalloc(sizeof(*item) + data_len, GFP_NOFS);
+	if (item) {
+		item->data_len = data_len;
+		item->ins_or_del = 0;
+		item->bytes_reserved = 0;
+		item->block_rsv = NULL;
+		item->delayed_node = NULL;
+		atomic_set(&item->refs, 1);
+	}
+	return item;
+}
+
+/*
+ * __btrfs_lookup_delayed_item - look up the delayed item by key
+ * @delayed_node: pointer to the delayed node
+ * @key:	  the key to look up
+ * @prev:	  used to store the prev item if the right item isn't found
+ * @next:	  used to store the next item if the right item isn't found
+ *
+ * Note: if we don't find the right item, we will return the prev item and
+ * the next item.
+ */
+static struct btrfs_delayed_item *__btrfs_lookup_delayed_item(
+				struct rb_root *root,
+				struct btrfs_key *key,
+				struct btrfs_delayed_item **prev,
+				struct btrfs_delayed_item **next)
+{
+	struct rb_node *node, *prev_node = NULL;
+	struct btrfs_delayed_item *delayed_item = NULL;
+	int ret = 0;
+
+	node = root->rb_node;
+
+	while (node) {
+		delayed_item = rb_entry(node, struct btrfs_delayed_item,
+					rb_node);
+		prev_node = node;
+		ret = btrfs_comp_cpu_keys(&delayed_item->key, key);
+		if (ret < 0)
+			node = node->rb_right;
+		else if (ret > 0)
+			node = node->rb_left;
+		else
+			return delayed_item;
+	}
+
+	if (prev) {
+		if (!prev_node)
+			*prev = NULL;
+		else if (ret < 0)
+			*prev = delayed_item;
+		else if ((node = rb_prev(prev_node)) != NULL) {
+			*prev = rb_entry(node, struct btrfs_delayed_item,
+					 rb_node);
+		} else
+			*prev = NULL;
+	}
+
+	if (next) {
+		if (!prev_node)
+			*next = NULL;
+		else if (ret > 0)
+			*next = delayed_item;
+		else if ((node = rb_next(prev_node)) != NULL) {
+			*next = rb_entry(node, struct btrfs_delayed_item,
+					 rb_node);
+		} else
+			*next = NULL;
+	}
+	return NULL;
+}
+
+struct btrfs_delayed_item *__btrfs_lookup_delayed_insertion_item(
+					struct btrfs_delayed_node *delayed_node,
+					struct btrfs_key *key)
+{
+	struct btrfs_delayed_item *item;
+
+	item = __btrfs_lookup_delayed_item(&delayed_node->ins_root, key,
+					   NULL, NULL);
+	return item;
+}
+
+struct btrfs_delayed_item *__btrfs_lookup_delayed_deletion_item(
+					struct btrfs_delayed_node *delayed_node,
+					struct btrfs_key *key)
+{
+	struct btrfs_delayed_item *item;
+
+	item = __btrfs_lookup_delayed_item(&delayed_node->del_root, key,
+					   NULL, NULL);
+	return item;
+}
+
+struct btrfs_delayed_item *__btrfs_search_delayed_insertion_item(
+					struct btrfs_delayed_node *delayed_node,
+					struct btrfs_key *key)
+{
+	struct btrfs_delayed_item *item, *next;
+
+	item = __btrfs_lookup_delayed_item(&delayed_node->ins_root, key,
+					   NULL, &next);
+	if (!item)
+		item = next;
+
+	return item;
+}
+
+struct btrfs_delayed_item *__btrfs_search_delayed_deletion_item(
+					struct btrfs_delayed_node *delayed_node,
+					struct btrfs_key *key)
+{
+	struct btrfs_delayed_item *item, *next;
+
+	item = __btrfs_lookup_delayed_item(&delayed_node->del_root, key,
+					   NULL, &next);
+	if (!item)
+		item = next;
+
+	return item;
+}
+
+static int __btrfs_add_delayed_item(struct btrfs_delayed_node *delayed_node,
+				    struct btrfs_delayed_item *ins,
+				    int action)
+{
+	struct rb_node **p, *node;
+	struct rb_node *parent_node = NULL;
+	struct rb_root *root;
+	struct btrfs_delayed_item *item;
+	int cmp;
+
+	if (action == BTRFS_DELAYED_INSERTION_ITEM)
+		root = &delayed_node->ins_root;
+	else if (action == BTRFS_DELAYED_DELETION_ITEM)
+		root = &delayed_node->del_root;
+	else
+		BUG();
+	p = &root->rb_node;
+	node = &ins->rb_node;
+
+	while (*p) {
+		parent_node = *p;
+		item = rb_entry(parent_node, struct btrfs_delayed_item,
+				 rb_node);
+
+		cmp = btrfs_comp_cpu_keys(&item->key, &ins->key);
+		if (cmp < 0)
+			p = &(*p)->rb_right;
+		else if (cmp > 0)
+			p = &(*p)->rb_left;
+		else
+			return -EEXIST;
+	}
+
+	rb_link_node(node, parent_node, p);
+	rb_insert_color(node, root);
+	ins->delayed_node = delayed_node;
+	ins->ins_or_del = action;
+
+	if (ins->key.type == BTRFS_DIR_INDEX_KEY &&
+	    action == BTRFS_DELAYED_INSERTION_ITEM &&
+	    ins->key.offset >= delayed_node->index_cnt)
+			delayed_node->index_cnt = ins->key.offset + 1;
+
+	delayed_node->count++;
+	atomic_inc(&delayed_node->root->fs_info->delayed_root->items);
+	return 0;
+}
+
+static int __btrfs_add_delayed_insertion_item(struct btrfs_delayed_node *node,
+					      struct btrfs_delayed_item *item)
+{
+	return __btrfs_add_delayed_item(node, item,
+					BTRFS_DELAYED_INSERTION_ITEM);
+}
+
+static int __btrfs_add_delayed_deletion_item(struct btrfs_delayed_node *node,
+					     struct btrfs_delayed_item *item)
+{
+	return __btrfs_add_delayed_item(node, item,
+					BTRFS_DELAYED_DELETION_ITEM);
+}
+
+static void __btrfs_remove_delayed_item(struct btrfs_delayed_item *delayed_item)
+{
+	struct rb_root *root;
+	struct btrfs_delayed_root *delayed_root;
+
+	delayed_root = delayed_item->delayed_node->root->fs_info->delayed_root;
+
+	BUG_ON(!delayed_root);
+	BUG_ON(delayed_item->ins_or_del != BTRFS_DELAYED_DELETION_ITEM &&
+	       delayed_item->ins_or_del != BTRFS_DELAYED_INSERTION_ITEM);
+
+	if (delayed_item->ins_or_del == BTRFS_DELAYED_INSERTION_ITEM)
+		root = &delayed_item->delayed_node->ins_root;
+	else
+		root = &delayed_item->delayed_node->del_root;
+
+	rb_erase(&delayed_item->rb_node, root);
+	delayed_item->delayed_node->count--;
+	atomic_dec(&delayed_root->items);
+	if (atomic_read(&delayed_root->items) < BTRFS_DELAYED_BACKGROUND &&
+	    waitqueue_active(&delayed_root->wait))
+		wake_up(&delayed_root->wait);
+}
+
+static void btrfs_release_delayed_item(struct btrfs_delayed_item *item)
+{
+	if (item) {
+		__btrfs_remove_delayed_item(item);
+		if (atomic_dec_and_test(&item->refs))
+			kfree(item);
+	}
+}
+
+struct btrfs_delayed_item *__btrfs_first_delayed_insertion_item(
+					struct btrfs_delayed_node *delayed_node)
+{
+	struct rb_node *p;
+	struct btrfs_delayed_item *item = NULL;
+
+	p = rb_first(&delayed_node->ins_root);
+	if (p)
+		item = rb_entry(p, struct btrfs_delayed_item, rb_node);
+
+	return item;
+}
+
+struct btrfs_delayed_item *__btrfs_first_delayed_deletion_item(
+					struct btrfs_delayed_node *delayed_node)
+{
+	struct rb_node *p;
+	struct btrfs_delayed_item *item = NULL;
+
+	p = rb_first(&delayed_node->del_root);
+	if (p)
+		item = rb_entry(p, struct btrfs_delayed_item, rb_node);
+
+	return item;
+}
+
+struct btrfs_delayed_item *__btrfs_next_delayed_item(
+						struct btrfs_delayed_item *item)
+{
+	struct rb_node *p;
+	struct btrfs_delayed_item *next = NULL;
+
+	p = rb_next(&item->rb_node);
+	if (p)
+		next = rb_entry(p, struct btrfs_delayed_item, rb_node);
+
+	return next;
+}
+
+static inline struct btrfs_delayed_node *btrfs_get_delayed_node(
+							struct inode *inode)
+{
+	struct btrfs_inode *btrfs_inode = BTRFS_I(inode);
+	struct btrfs_delayed_node *delayed_node;
+
+	delayed_node = btrfs_inode->delayed_node;
+	if (delayed_node)
+		atomic_inc(&delayed_node->refs);
+
+	return delayed_node;
+}
+
+static inline struct btrfs_root *btrfs_get_fs_root(struct btrfs_root *root,
+						   u64 root_id)
+{
+	struct btrfs_key root_key;
+
+	if (root->objectid == root_id)
+		return root;
+
+	root_key.objectid = root_id;
+	root_key.type = BTRFS_ROOT_ITEM_KEY;
+	root_key.offset = (u64)-1;
+	return btrfs_read_fs_root_no_name(root->fs_info, &root_key);
+}
+
+static int btrfs_delayed_item_reserve_metadata(struct btrfs_trans_handle *trans,
+					       struct btrfs_root *root,
+					       struct btrfs_delayed_item *item)
+{
+	struct btrfs_block_rsv *src_rsv;
+	struct btrfs_block_rsv *dst_rsv;
+	u64 num_bytes;
+	int ret;
+
+	if (!trans->bytes_reserved)
+		return 0;
+
+	src_rsv = trans->block_rsv;
+	dst_rsv = &root->fs_info->global_block_rsv;
+
+	num_bytes = btrfs_calc_trans_metadata_size(root, 1);
+	ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes);
+	if (!ret) {
+		item->bytes_reserved = num_bytes;
+		item->block_rsv = dst_rsv;
+	}
+
+	return ret;
+}
+
+static void btrfs_delayed_item_release_metadata(struct btrfs_root *root,
+						struct btrfs_delayed_item *item)
+{
+	if (!item->bytes_reserved)
+		return;
+
+	btrfs_block_rsv_release(root, item->block_rsv,
+				item->bytes_reserved);
+}
+
+static int btrfs_delayed_inode_reserve_metadata(
+					struct btrfs_trans_handle *trans,
+					struct btrfs_root *root,
+					struct btrfs_delayed_node *node)
+{
+	struct btrfs_block_rsv *src_rsv;
+	struct btrfs_block_rsv *dst_rsv;
+	u64 num_bytes;
+	int ret;
+
+	if (!trans->bytes_reserved)
+		return 0;
+
+	src_rsv = trans->block_rsv;
+	dst_rsv = &root->fs_info->global_block_rsv;
+
+	num_bytes = btrfs_calc_trans_metadata_size(root, 1);
+	ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes);
+	if (!ret)
+		node->bytes_reserved = num_bytes;
+
+	return ret;
+}
+
+static void btrfs_delayed_inode_release_metadata(struct btrfs_root *root,
+						struct btrfs_delayed_node *node)
+{
+	struct btrfs_block_rsv *rsv;
+
+	if (!node->bytes_reserved)
+		return;
+
+	rsv = &root->fs_info->global_block_rsv;
+	btrfs_block_rsv_release(root, rsv,
+				node->bytes_reserved);
+	node->bytes_reserved = 0;
+}
+
+/*
+ * This helper will insert some continuous items into the same leaf according
+ * to the free space of the leaf.
+ */
+static int btrfs_batch_insert_items(struct btrfs_trans_handle *trans,
+				struct btrfs_root *root,
+				struct btrfs_path *path,
+				struct btrfs_delayed_item *item)
+{
+	struct btrfs_delayed_item *curr, *next;
+	int free_space;
+	int total_data_size = 0, total_size = 0;
+	struct extent_buffer *leaf;
+	char *data_ptr;
+	struct btrfs_key *keys;
+	u32 *data_size;
+	struct list_head head;
+	int slot;
+	int nitems;
+	int i;
+	int ret = 0;
+
+	BUG_ON(!path->nodes[0]);
+
+	leaf = path->nodes[0];
+	free_space = btrfs_leaf_free_space(root, leaf);
+	INIT_LIST_HEAD(&head);
+
+	next = item;
+
+	/*
+	 * count the number of the continuous items that we can insert in batch
+	 */
+	while (total_size + next->data_len + sizeof(struct btrfs_item) <=
+	       free_space) {
+		total_data_size += next->data_len;
+		total_size += next->data_len + sizeof(struct btrfs_item);
+		list_add_tail(&next->tree_list, &head);
+		nitems++;
+
+		curr = next;
+		next = __btrfs_next_delayed_item(curr);
+		if (!next)
+			break;
+
+		if (!btrfs_is_continuous_delayed_item(curr, next))
+			break;
+	}
+
+	if (!nitems) {
+		ret = 0;
+		goto out;
+	}
+
+	/*
+	 * we need allocate some memory space, but it might cause the task
+	 * to sleep, so we set all locked nodes in the path to blocking locks
+	 * first.
+	 */
+	btrfs_set_path_blocking(path);
+
+	keys = kmalloc(sizeof(struct btrfs_key) * nitems, GFP_NOFS);
+	if (!keys) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	data_size = kmalloc(sizeof(u32) * nitems, GFP_NOFS);
+	if (!data_size) {
+		ret = -ENOMEM;
+		goto error;
+	}
+
+	/* get keys of all the delayed items */
+	i = 0;
+	list_for_each_entry(next, &head, tree_list) {
+		keys[i] = next->key;
+		data_size[i] = next->data_len;
+		i++;
+	}
+
+	/* reset all the locked nodes in the patch to spinning locks. */
+	btrfs_clear_path_blocking(path, NULL);
+
+	/* insert the keys of the items */
+	ret = setup_items_for_insert(trans, root, path, keys, data_size,
+				     total_data_size, total_size, nitems);
+	if (ret)
+		goto error;
+
+	/* insert the dir index items */
+	slot = path->slots[0];
+	list_for_each_entry_safe(curr, next, &head, tree_list) {
+		data_ptr = btrfs_item_ptr(leaf, slot, char);
+		write_extent_buffer(leaf, &curr->data,
+				    (unsigned long)data_ptr,
+				    curr->data_len);
+		slot++;
+
+		btrfs_delayed_item_release_metadata(root, curr);
+
+		list_del(&curr->tree_list);
+		btrfs_release_delayed_item(curr);
+	}
+
+error:
+	kfree(data_size);
+	kfree(keys);
+out:
+	return ret;
+}
+
+/*
+ * This helper can just do simple insertion that needn't extend item for new
+ * data, such as directory name index insertion, inode insertion.
+ */
+static int btrfs_insert_delayed_item(struct btrfs_trans_handle *trans,
+				     struct btrfs_root *root,
+				     struct btrfs_path *path,
+				     struct btrfs_delayed_item *delayed_item)
+{
+	struct extent_buffer *leaf;
+	struct btrfs_item *item;
+	char *ptr;
+	int ret;
+
+	ret = btrfs_insert_empty_item(trans, root, path, &delayed_item->key,
+				      delayed_item->data_len);
+	if (ret < 0 && ret != -EEXIST)
+		return ret;
+
+	leaf = path->nodes[0];
+
+	item = btrfs_item_nr(leaf, path->slots[0]);
+	ptr = btrfs_item_ptr(leaf, path->slots[0], char);
+
+	write_extent_buffer(leaf, delayed_item->data, (unsigned long)ptr,
+			    delayed_item->data_len);
+	btrfs_mark_buffer_dirty(leaf);
+
+	btrfs_delayed_item_release_metadata(root, delayed_item);
+	return 0;
+}
+
+/*
+ * we insert an item first, then if there are some continuous items, we try
+ * to insert those items into the same leaf.
+ */
+static int btrfs_insert_delayed_items(struct btrfs_trans_handle *trans,
+				      struct btrfs_path *path,
+				      struct btrfs_root *root,
+				      struct btrfs_delayed_node *node)
+{
+	struct btrfs_delayed_item *curr, *prev;
+	int ret = 0;
+
+do_again:
+	mutex_lock(&node->mutex);
+	curr = __btrfs_first_delayed_insertion_item(node);
+	if (!curr)
+		goto insert_end;
+
+	ret = btrfs_insert_delayed_item(trans, root, path, curr);
+	if (ret < 0) {
+		btrfs_release_path(path);
+		goto insert_end;
+	}
+
+	prev = curr;
+	curr = __btrfs_next_delayed_item(prev);
+	if (curr && btrfs_is_continuous_delayed_item(prev, curr)) {
+		/* insert the continuous items into the same leaf */
+		path->slots[0]++;
+		btrfs_batch_insert_items(trans, root, path, curr);
+	}
+	btrfs_release_delayed_item(prev);
+	btrfs_mark_buffer_dirty(path->nodes[0]);
+
+	btrfs_release_path(path);
+	mutex_unlock(&node->mutex);
+	goto do_again;
+
+insert_end:
+	mutex_unlock(&node->mutex);
+	return ret;
+}
+
+static int btrfs_batch_delete_items(struct btrfs_trans_handle *trans,
+				    struct btrfs_root *root,
+				    struct btrfs_path *path,
+				    struct btrfs_delayed_item *item)
+{
+	struct btrfs_delayed_item *curr, *next;
+	struct extent_buffer *leaf;
+	struct btrfs_key key;
+	struct list_head head;
+	int nitems, i, last_item;
+	int ret = 0;
+
+	BUG_ON(!path->nodes[0]);
+
+	leaf = path->nodes[0];
+
+	i = path->slots[0];
+	last_item = btrfs_header_nritems(leaf) - 1;
+	if (i > last_item)
+		return -ENOENT;	/* FIXME: Is errno suitable? */
+
+	next = item;
+	INIT_LIST_HEAD(&head);
+	btrfs_item_key_to_cpu(leaf, &key, i);
+	nitems = 0;
+	/*
+	 * count the number of the dir index items that we can delete in batch
+	 */
+	while (btrfs_comp_cpu_keys(&next->key, &key) == 0) {
+		list_add_tail(&next->tree_list, &head);
+		nitems++;
+
+		curr = next;
+		next = __btrfs_next_delayed_item(curr);
+		if (!next)
+			break;
+
+		if (!btrfs_is_continuous_delayed_item(curr, next))
+			break;
+
+		i++;
+		if (i > last_item)
+			break;
+		btrfs_item_key_to_cpu(leaf, &key, i);
+	}
+
+	if (!nitems)
+		return 0;
+
+	ret = btrfs_del_items(trans, root, path, path->slots[0], nitems);
+	if (ret)
+		goto out;
+
+	list_for_each_entry_safe(curr, next, &head, tree_list) {
+		btrfs_delayed_item_release_metadata(root, curr);
+		list_del(&curr->tree_list);
+		btrfs_release_delayed_item(curr);
+	}
+
+out:
+	return ret;
+}
+
+static int btrfs_delete_delayed_items(struct btrfs_trans_handle *trans,
+				      struct btrfs_path *path,
+				      struct btrfs_root *root,
+				      struct btrfs_delayed_node *node)
+{
+	struct btrfs_delayed_item *curr, *prev;
+	int ret = 0;
+
+do_again:
+	mutex_lock(&node->mutex);
+	curr = __btrfs_first_delayed_deletion_item(node);
+	if (!curr)
+		goto delete_fail;
+
+	ret = btrfs_search_slot(trans, root, &curr->key, path, -1, 1);
+	if (ret < 0)
+		goto delete_fail;
+	else if (ret > 0) {
+		/*
+		 * can't find the item which the node points to, so this node
+		 * is invalid, just drop it.
+		 */
+		prev = curr;
+		curr = __btrfs_next_delayed_item(prev);
+		btrfs_release_delayed_item(prev);
+		ret = 0;
+		btrfs_release_path(path);
+		if (curr)
+			goto do_again;
+		else
+			goto delete_fail;
+	}
+
+	btrfs_batch_delete_items(trans, root, path, curr);
+	btrfs_release_path(path);
+	mutex_unlock(&node->mutex);
+	goto do_again;
+
+delete_fail:
+	btrfs_release_path(path);
+	mutex_unlock(&node->mutex);
+	return ret;
+}
+
+static void btrfs_release_delayed_inode(struct btrfs_delayed_node *delayed_node)
+{
+	struct btrfs_delayed_root *delayed_root;
+
+	if (delayed_node && delayed_node->inode_dirty) {
+		BUG_ON(!delayed_node->root);
+		delayed_node->inode_dirty = 0;
+		delayed_node->count--;
+
+		delayed_root = delayed_node->root->fs_info->delayed_root;
+		atomic_dec(&delayed_root->items);
+		if (atomic_read(&delayed_root->items) <
+		    BTRFS_DELAYED_BACKGROUND &&
+		    waitqueue_active(&delayed_root->wait))
+			wake_up(&delayed_root->wait);
+	}
+}
+
+static int btrfs_update_delayed_inode(struct btrfs_trans_handle *trans,
+				      struct btrfs_root *root,
+				      struct btrfs_path *path,
+				      struct btrfs_delayed_node *node)
+{
+	struct btrfs_key key;
+	struct btrfs_inode_item *inode_item;
+	struct extent_buffer *leaf;
+	int ret;
+
+	mutex_lock(&node->mutex);
+	if (!node->inode_dirty) {
+		mutex_unlock(&node->mutex);
+		return 0;
+	}
+
+	key.objectid = node->inode_id;
+	btrfs_set_key_type(&key, BTRFS_INODE_ITEM_KEY);
+	key.offset = 0;
+	ret = btrfs_lookup_inode(trans, root, path, &key, 1);
+	if (ret > 0) {
+		btrfs_release_path(path);
+		mutex_unlock(&node->mutex);
+		return -ENOENT;
+	} else if (ret < 0) {
+		mutex_unlock(&node->mutex);
+		return ret;
+	}
+
+	btrfs_unlock_up_safe(path, 1);
+	leaf = path->nodes[0];
+	inode_item = btrfs_item_ptr(leaf, path->slots[0],
+				    struct btrfs_inode_item);
+	write_extent_buffer(leaf, &node->inode_item, (unsigned long)inode_item,
+			    sizeof(struct btrfs_inode_item));
+	btrfs_mark_buffer_dirty(leaf);
+	btrfs_release_path(path);
+
+	btrfs_delayed_inode_release_metadata(root, node);
+	btrfs_release_delayed_inode(node);
+	mutex_unlock(&node->mutex);
+
+	return 0;
+}
+
+/* Called when committing the transaction. */
+int btrfs_run_delayed_items(struct btrfs_trans_handle *trans,
+			    struct btrfs_root *root)
+{
+	struct btrfs_delayed_root *delayed_root;
+	struct btrfs_delayed_node *curr_node, *prev_node;
+	struct btrfs_path *path;
+	int ret = 0;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+	path->leave_spinning = 1;
+
+	delayed_root = btrfs_get_delayed_root(root);
+
+	curr_node = btrfs_first_delayed_node(delayed_root);
+	while (curr_node) {
+		root = curr_node->root;
+		ret = btrfs_insert_delayed_items(trans, path, root,
+						 curr_node);
+		if (!ret)
+			ret = btrfs_delete_delayed_items(trans, path, root,
+							 curr_node);
+		if (!ret)
+			ret = btrfs_update_delayed_inode(trans, root, path,
+							 curr_node);
+		if (ret) {
+			btrfs_release_delayed_node(curr_node);
+			break;
+		}
+
+		prev_node = curr_node;
+		curr_node = btrfs_next_delayed_node(curr_node);
+		btrfs_release_delayed_node(prev_node);
+	}
+
+	btrfs_free_path(path);
+	return ret;
+}
+
+static int __btrfs_commit_inode_delayed_items(struct btrfs_trans_handle *trans,
+					      struct btrfs_delayed_node *node)
+{
+	struct btrfs_path *path;
+	int ret;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+	path->leave_spinning = 1;
+
+	ret = btrfs_insert_delayed_items(trans, path, node->root, node);
+	if (!ret)
+		ret = btrfs_delete_delayed_items(trans, path, node->root, node);
+	if (!ret)
+		ret = btrfs_update_delayed_inode(trans, node->root, path, node);
+	btrfs_free_path(path);
+
+	return ret;
+}
+
+int btrfs_commit_inode_delayed_items(struct btrfs_trans_handle *trans,
+				     struct inode *inode)
+{
+	struct btrfs_delayed_node *delayed_node = btrfs_get_delayed_node(inode);
+	int ret;
+
+	if (!delayed_node)
+		return 0;
+
+	mutex_lock(&delayed_node->mutex);
+	if (!delayed_node->count) {
+		mutex_unlock(&delayed_node->mutex);
+		btrfs_release_delayed_node(delayed_node);
+		return 0;
+	}
+	mutex_unlock(&delayed_node->mutex);
+
+	ret = __btrfs_commit_inode_delayed_items(trans, delayed_node);
+	btrfs_release_delayed_node(delayed_node);
+	return ret;
+}
+
+void btrfs_remove_delayed_node(struct inode *inode)
+{
+	struct btrfs_delayed_node *delayed_node;
+
+	delayed_node = ACCESS_ONCE(BTRFS_I(inode)->delayed_node);
+	if (!delayed_node)
+		return;
+
+	BTRFS_I(inode)->delayed_node = NULL;
+	btrfs_release_delayed_node(delayed_node);
+}
+
+struct btrfs_async_delayed_node {
+	struct btrfs_root *root;
+	struct btrfs_delayed_node *delayed_node;
+	struct btrfs_work work;
+};
+
+static void btrfs_async_run_delayed_node_done(struct btrfs_work *work)
+{
+	struct btrfs_async_delayed_node *async_node;
+	struct btrfs_trans_handle *trans;
+	struct btrfs_path *path;
+	struct btrfs_delayed_node *delayed_node = NULL;
+	struct btrfs_root *root;
+	unsigned long nr = 0;
+	int need_requeue = 0;
+	int ret;
+
+	async_node = container_of(work, struct btrfs_async_delayed_node, work);
+
+	path = btrfs_alloc_path();
+	if (!path)
+		goto out;
+	path->leave_spinning = 1;
+
+	delayed_node = async_node->delayed_node;
+	root = delayed_node->root;
+
+	trans = btrfs_join_transaction(root, 0);
+	if (IS_ERR(trans))
+		goto free_path;
+
+	ret = btrfs_insert_delayed_items(trans, path, root, delayed_node);
+	if (!ret)
+		ret = btrfs_delete_delayed_items(trans, path, root,
+						 delayed_node);
+
+	if (!ret)
+		btrfs_update_delayed_inode(trans, root, path, delayed_node);
+
+	/*
+	 * Maybe new delayed items have been inserted, so we need requeue
+	 * the work. Besides that, we must dequeue the empty delayed nodes
+	 * to avoid the race between delayed items balance and the worker.
+	 * The race like this:
+	 * 	Task1				Worker thread
+	 * 					count == 0, needn't requeue
+	 * 					  also needn't insert the
+	 * 					  delayed node into prepare
+	 * 					  list again.
+	 * 	add lots of delayed items
+	 * 	queue the delayed node
+	 * 	  already in the list,
+	 * 	  and not in the prepare
+	 * 	  list, it means the delayed
+	 * 	  node is being dealt with
+	 * 	  by the worker.
+	 * 	do delayed items balance
+	 * 	  the delayed node is being
+	 * 	  dealt with by the worker
+	 * 	  now, just wait.
+	 * 	  				the worker goto idle.
+	 * Task1 will sleep until the transaction is commited.
+	 */
+	mutex_lock(&delayed_node->mutex);
+	if (delayed_node->count)
+		need_requeue = 1;
+	else
+		btrfs_dequeue_delayed_node(root->fs_info->delayed_root,
+					   delayed_node);
+	mutex_unlock(&delayed_node->mutex);
+
+	nr = trans->blocks_used;
+
+	btrfs_end_transaction_dmeta(trans, root);
+	__btrfs_btree_balance_dirty(root, nr);
+free_path:
+	btrfs_free_path(path);
+out:
+	if (need_requeue)
+		btrfs_requeue_work(&async_node->work);
+	else {
+		btrfs_release_prepared_delayed_node(delayed_node);
+		kfree(async_node);
+	}
+}
+
+static int btrfs_wq_run_delayed_node(struct btrfs_delayed_root *delayed_root,
+				     struct btrfs_root *root, int all)
+{
+	struct btrfs_async_delayed_node *async_node;
+	struct btrfs_delayed_node *curr;
+	int count = 0;
+
+again:
+	curr = btrfs_first_prepared_delayed_node(delayed_root);
+	if (!curr)
+		return 0;
+
+	async_node = kmalloc(sizeof(*async_node), GFP_NOFS);
+	if (!async_node) {
+		btrfs_release_prepared_delayed_node(curr);
+		return -ENOMEM;
+	}
+
+	async_node->root = root;
+	async_node->delayed_node = curr;
+
+	async_node->work.func = btrfs_async_run_delayed_node_done;
+	async_node->work.flags = 0;
+
+	btrfs_queue_worker(&root->fs_info->delayed_workers, &async_node->work);
+	count++;
+
+	if (all || count < 4)
+		goto again;
+
+	return 0;
+}
+
+void btrfs_balance_delayed_items(struct btrfs_root *root)
+{
+	struct btrfs_delayed_root *delayed_root;
+
+	delayed_root = btrfs_get_delayed_root(root);
+
+	if (atomic_read(&delayed_root->items) < BTRFS_DELAYED_BACKGROUND)
+		return;
+
+	if (atomic_read(&delayed_root->items) >= BTRFS_DELAYED_WRITEBACK) {
+		int ret;
+		ret = btrfs_wq_run_delayed_node(delayed_root, root, 1);
+		if (ret)
+			return;
+
+		wait_event_interruptible_timeout(
+				delayed_root->wait,
+				(atomic_read(&delayed_root->items) <
+				 BTRFS_DELAYED_BACKGROUND),
+				HZ);
+		return;
+	}
+
+	btrfs_wq_run_delayed_node(delayed_root, root, 0);
+}
+
+int btrfs_insert_delayed_dir_index(struct btrfs_trans_handle *trans,
+				   struct btrfs_root *root, const char *name,
+				   int name_len, struct inode *dir,
+				   struct btrfs_disk_key *disk_key, u8 type,
+				   u64 index)
+{
+	struct btrfs_delayed_node *delayed_node;
+	struct btrfs_delayed_item *delayed_item;
+	struct btrfs_dir_item *dir_item;
+	int ret;
+
+	delayed_node = btrfs_get_or_create_delayed_node(dir);
+	if (IS_ERR(delayed_node))
+		return PTR_ERR(delayed_node);
+
+	delayed_item = btrfs_alloc_delayed_item(sizeof(*dir_item) + name_len);
+	if (!delayed_item) {
+		ret = -ENOMEM;
+		goto release_node;
+	}
+
+	ret = btrfs_delayed_item_reserve_metadata(trans, root, delayed_item);
+	/*
+	 * we have reserved enough space when we start a new transaction,
+	 * so reserving metadata failure is impossible
+	 */
+	BUG_ON(ret);
+
+	delayed_item->key.objectid = btrfs_ino(dir);
+	btrfs_set_key_type(&delayed_item->key, BTRFS_DIR_INDEX_KEY);
+	delayed_item->key.offset = index;
+
+	dir_item = (struct btrfs_dir_item *)delayed_item->data;
+	dir_item->location = *disk_key;
+	dir_item->transid = cpu_to_le64(trans->transid);
+	dir_item->data_len = 0;
+	dir_item->name_len = cpu_to_le16(name_len);
+	dir_item->type = type;
+	memcpy((char *)(dir_item + 1), name, name_len);
+
+	mutex_lock(&delayed_node->mutex);
+	ret = __btrfs_add_delayed_insertion_item(delayed_node, delayed_item);
+	if (unlikely(ret)) {
+		printk(KERN_ERR "err add delayed dir index item(name: %s) into "
+				"the insertion tree of the delayed node"
+				"(root id: %llu, inode id: %llu, errno: %d)\n",
+				name,
+				(unsigned long long)delayed_node->root->objectid,
+				(unsigned long long)delayed_node->inode_id,
+				ret);
+		BUG();
+	}
+	mutex_unlock(&delayed_node->mutex);
+
+release_node:
+	btrfs_release_delayed_node(delayed_node);
+	return ret;
+}
+
+static int btrfs_delete_delayed_insertion_item(struct btrfs_root *root,
+					       struct btrfs_delayed_node *node,
+					       struct btrfs_key *key)
+{
+	struct btrfs_delayed_item *item;
+
+	mutex_lock(&node->mutex);
+	item = __btrfs_lookup_delayed_insertion_item(node, key);
+	if (!item) {
+		mutex_unlock(&node->mutex);
+		return 1;
+	}
+
+	btrfs_delayed_item_release_metadata(root, item);
+	btrfs_release_delayed_item(item);
+	mutex_unlock(&node->mutex);
+	return 0;
+}
+
+int btrfs_delete_delayed_dir_index(struct btrfs_trans_handle *trans,
+				   struct btrfs_root *root, struct inode *dir,
+				   u64 index)
+{
+	struct btrfs_delayed_node *node;
+	struct btrfs_delayed_item *item;
+	struct btrfs_key item_key;
+	int ret;
+
+	node = btrfs_get_or_create_delayed_node(dir);
+	if (IS_ERR(node))
+		return PTR_ERR(node);
+
+	item_key.objectid = btrfs_ino(dir);
+	btrfs_set_key_type(&item_key, BTRFS_DIR_INDEX_KEY);
+	item_key.offset = index;
+
+	ret = btrfs_delete_delayed_insertion_item(root, node, &item_key);
+	if (!ret)
+		goto end;
+
+	item = btrfs_alloc_delayed_item(0);
+	if (!item) {
+		ret = -ENOMEM;
+		goto end;
+	}
+
+	item->key = item_key;
+
+	ret = btrfs_delayed_item_reserve_metadata(trans, root, item);
+	/*
+	 * we have reserved enough space when we start a new transaction,
+	 * so reserving metadata failure is impossible.
+	 */
+	BUG_ON(ret);
+
+	mutex_lock(&node->mutex);
+	ret = __btrfs_add_delayed_deletion_item(node, item);
+	if (unlikely(ret)) {
+		printk(KERN_ERR "err add delayed dir index item(index: %llu) "
+				"into the deletion tree of the delayed node"
+				"(root id: %llu, inode id: %llu, errno: %d)\n",
+				(unsigned long long)index,
+				(unsigned long long)node->root->objectid,
+				(unsigned long long)node->inode_id,
+				ret);
+		BUG();
+	}
+	mutex_unlock(&node->mutex);
+end:
+	btrfs_release_delayed_node(node);
+	return ret;
+}
+
+int btrfs_inode_delayed_dir_index_count(struct inode *inode)
+{
+	struct btrfs_delayed_node *delayed_node = BTRFS_I(inode)->delayed_node;
+	int ret = 0;
+
+	if (!delayed_node)
+		return -ENOENT;
+
+	/*
+	 * Since we have held i_mutex of this directory, it is impossible that
+	 * a new directory index is added into the delayed node and index_cnt
+	 * is updated now. So we needn't lock the delayed node.
+	 */
+	if (!delayed_node->index_cnt)
+		return -EINVAL;
+
+	BTRFS_I(inode)->index_cnt = delayed_node->index_cnt;
+	return ret;
+}
+
+void btrfs_get_delayed_items(struct inode *inode, struct list_head *ins_list,
+			     struct list_head *del_list)
+{
+	struct btrfs_delayed_node *delayed_node;
+	struct btrfs_delayed_item *item;
+
+	delayed_node = btrfs_get_delayed_node(inode);
+	if (!delayed_node)
+		return;
+
+	mutex_lock(&delayed_node->mutex);
+	item = __btrfs_first_delayed_insertion_item(delayed_node);
+	while (item) {
+		atomic_inc(&item->refs);
+		list_add_tail(&item->readdir_list, ins_list);
+		item = __btrfs_next_delayed_item(item);
+	}
+
+	item = __btrfs_first_delayed_deletion_item(delayed_node);
+	while (item) {
+		atomic_inc(&item->refs);
+		list_add_tail(&item->readdir_list, del_list);
+		item = __btrfs_next_delayed_item(item);
+	}
+	mutex_unlock(&delayed_node->mutex);
+	/*
+	 * This delayed node is still cached in the btrfs inode, so refs
+	 * must be > 1 now, and we needn't check it is going to be freed
+	 * or not.
+	 *
+	 * Besides that, this function is used to read dir, we do not
+	 * insert/delete delayed items in this period. So we also needn't
+	 * requeue or dequeue this delayed node.
+	 */
+	atomic_dec(&delayed_node->refs);
+}
+
+void btrfs_put_delayed_items(struct list_head *ins_list,
+			     struct list_head *del_list)
+{
+	struct btrfs_delayed_item *curr, *next;
+
+	list_for_each_entry_safe(curr, next, ins_list, readdir_list) {
+		list_del(&curr->readdir_list);
+		if (atomic_dec_and_test(&curr->refs))
+			kfree(curr);
+	}
+
+	list_for_each_entry_safe(curr, next, del_list, readdir_list) {
+		list_del(&curr->readdir_list);
+		if (atomic_dec_and_test(&curr->refs))
+			kfree(curr);
+	}
+}
+
+int btrfs_should_delete_dir_index(struct list_head *del_list,
+				  u64 index)
+{
+	struct btrfs_delayed_item *curr, *next;
+	int ret;
+
+	if (list_empty(del_list))
+		return 0;
+
+	list_for_each_entry_safe(curr, next, del_list, readdir_list) {
+		if (curr->key.offset > index)
+			break;
+
+		list_del(&curr->readdir_list);
+		ret = (curr->key.offset == index);
+
+		if (atomic_dec_and_test(&curr->refs))
+			kfree(curr);
+
+		if (ret)
+			return 1;
+		else
+			continue;
+	}
+	return 0;
+}
+
+/*
+ * btrfs_readdir_delayed_dir_index - read dir info stored in the delayed tree
+ *
+ */
+int btrfs_readdir_delayed_dir_index(struct file *filp, void *dirent,
+				    filldir_t filldir,
+				    struct list_head *ins_list)
+{
+	struct btrfs_dir_item *di;
+	struct btrfs_delayed_item *curr, *next;
+	struct btrfs_key location;
+	char *name;
+	int name_len;
+	int over = 0;
+	unsigned char d_type;
+
+	if (list_empty(ins_list))
+		return 0;
+
+	/*
+	 * Changing the data of the delayed item is impossible. So
+	 * we needn't lock them. And we have held i_mutex of the
+	 * directory, nobody can delete any directory indexes now.
+	 */
+	list_for_each_entry_safe(curr, next, ins_list, readdir_list) {
+		list_del(&curr->readdir_list);
+
+		if (curr->key.offset < filp->f_pos) {
+			if (atomic_dec_and_test(&curr->refs))
+				kfree(curr);
+			continue;
+		}
+
+		filp->f_pos = curr->key.offset;
+
+		di = (struct btrfs_dir_item *)curr->data;
+		name = (char *)(di + 1);
+		name_len = le16_to_cpu(di->name_len);
+
+		d_type = btrfs_filetype_table[di->type];
+		btrfs_disk_key_to_cpu(&location, &di->location);
+
+		over = filldir(dirent, name, name_len, curr->key.offset,
+			       location.objectid, d_type);
+
+		if (atomic_dec_and_test(&curr->refs))
+			kfree(curr);
+
+		if (over)
+			return 1;
+	}
+	return 0;
+}
+
+BTRFS_SETGET_STACK_FUNCS(stack_inode_generation, struct btrfs_inode_item,
+			 generation, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_inode_sequence, struct btrfs_inode_item,
+			 sequence, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_inode_transid, struct btrfs_inode_item,
+			 transid, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_inode_size, struct btrfs_inode_item, size, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_inode_nbytes, struct btrfs_inode_item,
+			 nbytes, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_inode_block_group, struct btrfs_inode_item,
+			 block_group, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_inode_nlink, struct btrfs_inode_item, nlink, 32);
+BTRFS_SETGET_STACK_FUNCS(stack_inode_uid, struct btrfs_inode_item, uid, 32);
+BTRFS_SETGET_STACK_FUNCS(stack_inode_gid, struct btrfs_inode_item, gid, 32);
+BTRFS_SETGET_STACK_FUNCS(stack_inode_mode, struct btrfs_inode_item, mode, 32);
+BTRFS_SETGET_STACK_FUNCS(stack_inode_rdev, struct btrfs_inode_item, rdev, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_inode_flags, struct btrfs_inode_item, flags, 64);
+
+BTRFS_SETGET_STACK_FUNCS(stack_timespec_sec, struct btrfs_timespec, sec, 64);
+BTRFS_SETGET_STACK_FUNCS(stack_timespec_nsec, struct btrfs_timespec, nsec, 32);
+
+static void fill_stack_inode_item(struct btrfs_trans_handle *trans,
+				  struct btrfs_inode_item *inode_item,
+				  struct inode *inode)
+{
+	btrfs_set_stack_inode_uid(inode_item, inode->i_uid);
+	btrfs_set_stack_inode_gid(inode_item, inode->i_gid);
+	btrfs_set_stack_inode_size(inode_item, BTRFS_I(inode)->disk_i_size);
+	btrfs_set_stack_inode_mode(inode_item, inode->i_mode);
+	btrfs_set_stack_inode_nlink(inode_item, inode->i_nlink);
+	btrfs_set_stack_inode_nbytes(inode_item, inode_get_bytes(inode));
+	btrfs_set_stack_inode_generation(inode_item,
+					 BTRFS_I(inode)->generation);
+	btrfs_set_stack_inode_sequence(inode_item, BTRFS_I(inode)->sequence);
+	btrfs_set_stack_inode_transid(inode_item, trans->transid);
+	btrfs_set_stack_inode_rdev(inode_item, inode->i_rdev);
+	btrfs_set_stack_inode_flags(inode_item, BTRFS_I(inode)->flags);
+	btrfs_set_stack_inode_block_group(inode_item,
+					  BTRFS_I(inode)->block_group);
+
+	btrfs_set_stack_timespec_sec(btrfs_inode_atime(inode_item),
+				     inode->i_atime.tv_sec);
+	btrfs_set_stack_timespec_nsec(btrfs_inode_atime(inode_item),
+				      inode->i_atime.tv_nsec);
+
+	btrfs_set_stack_timespec_sec(btrfs_inode_mtime(inode_item),
+				     inode->i_mtime.tv_sec);
+	btrfs_set_stack_timespec_nsec(btrfs_inode_mtime(inode_item),
+				      inode->i_mtime.tv_nsec);
+
+	btrfs_set_stack_timespec_sec(btrfs_inode_ctime(inode_item),
+				     inode->i_ctime.tv_sec);
+	btrfs_set_stack_timespec_nsec(btrfs_inode_ctime(inode_item),
+				      inode->i_ctime.tv_nsec);
+}
+
+int btrfs_delayed_update_inode(struct btrfs_trans_handle *trans,
+			       struct btrfs_root *root, struct inode *inode)
+{
+	struct btrfs_delayed_node *delayed_node;
+	int ret;
+
+	delayed_node = btrfs_get_or_create_delayed_node(inode);
+	if (IS_ERR(delayed_node))
+		return PTR_ERR(delayed_node);
+
+	mutex_lock(&delayed_node->mutex);
+	if (delayed_node->inode_dirty) {
+		fill_stack_inode_item(trans, &delayed_node->inode_item, inode);
+		goto release_node;
+	}
+
+	ret = btrfs_delayed_inode_reserve_metadata(trans, root, delayed_node);
+	/*
+	 * we must reserve enough space when we start a new transaction,
+	 * so reserving metadata failure is impossible
+	 */
+	BUG_ON(ret);
+
+	fill_stack_inode_item(trans, &delayed_node->inode_item, inode);
+	delayed_node->inode_dirty = 1;
+	delayed_node->count++;
+	atomic_inc(&root->fs_info->delayed_root->items);
+release_node:
+	mutex_unlock(&delayed_node->mutex);
+	btrfs_release_delayed_node(delayed_node);
+	return ret;
+}
+
+static void __btrfs_kill_delayed_node(struct btrfs_delayed_node *delayed_node)
+{
+	struct btrfs_root *root = delayed_node->root;
+	struct btrfs_delayed_item *curr_item, *prev_item;
+
+	mutex_lock(&delayed_node->mutex);
+	curr_item = __btrfs_first_delayed_insertion_item(delayed_node);
+	while (curr_item) {
+		btrfs_delayed_item_release_metadata(root, curr_item);
+		prev_item = curr_item;
+		curr_item = __btrfs_next_delayed_item(prev_item);
+		btrfs_release_delayed_item(prev_item);
+	}
+
+	curr_item = __btrfs_first_delayed_deletion_item(delayed_node);
+	while (curr_item) {
+		btrfs_delayed_item_release_metadata(root, curr_item);
+		prev_item = curr_item;
+		curr_item = __btrfs_next_delayed_item(prev_item);
+		btrfs_release_delayed_item(prev_item);
+	}
+
+	if (delayed_node->inode_dirty) {
+		btrfs_delayed_inode_release_metadata(root, delayed_node);
+		btrfs_release_delayed_inode(delayed_node);
+	}
+	mutex_unlock(&delayed_node->mutex);
+}
+
+void btrfs_kill_delayed_inode_items(struct inode *inode)
+{
+	struct btrfs_delayed_node *delayed_node;
+
+	delayed_node = btrfs_get_delayed_node(inode);
+	if (!delayed_node)
+		return;
+
+	__btrfs_kill_delayed_node(delayed_node);
+	btrfs_release_delayed_node(delayed_node);
+}
+
+void btrfs_kill_all_delayed_nodes(struct btrfs_root *root)
+{
+	u64 inode_id = 0;
+	struct btrfs_delayed_node *delayed_nodes[8];
+	int i, n;
+
+	while (1) {
+		spin_lock(&root->inode_lock);
+		n = radix_tree_gang_lookup(&root->delayed_nodes_tree,
+					   (void **)delayed_nodes, inode_id,
+					   ARRAY_SIZE(delayed_nodes));
+		if (!n) {
+			spin_unlock(&root->inode_lock);
+			break;
+		}
+
+		inode_id = delayed_nodes[n - 1]->inode_id + 1;
+
+		for (i = 0; i < n; i++)
+			atomic_inc(&delayed_nodes[i]->refs);
+		spin_unlock(&root->inode_lock);
+
+		for (i = 0; i < n; i++) {
+			__btrfs_kill_delayed_node(delayed_nodes[i]);
+			btrfs_release_delayed_node(delayed_nodes[i]);
+		}
+	}
+}

fs/btrfs/delayed-inode.h

@@ -0,0 +1,141 @@
+/*
+ * Copyright (C) 2011 Fujitsu.  All rights reserved.
+ * Written by Miao Xie <miaox@cn.fujitsu.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#ifndef __DELAYED_TREE_OPERATION_H
+#define __DELAYED_TREE_OPERATION_H
+
+#include <linux/rbtree.h>
+#include <linux/spinlock.h>
+#include <linux/mutex.h>
+#include <linux/list.h>
+#include <linux/wait.h>
+#include <asm/atomic.h>
+
+#include "ctree.h"
+
+/* types of the delayed item */
+#define BTRFS_DELAYED_INSERTION_ITEM	1
+#define BTRFS_DELAYED_DELETION_ITEM	2
+
+struct btrfs_delayed_root {
+	spinlock_t lock;
+	struct list_head node_list;
+	/*
+	 * Used for delayed nodes which is waiting to be dealt with by the
+	 * worker. If the delayed node is inserted into the work queue, we
+	 * drop it from this list.
+	 */
+	struct list_head prepare_list;
+	atomic_t items;		/* for delayed items */
+	int nodes;		/* for delayed nodes */
+	wait_queue_head_t wait;
+};
+
+struct btrfs_delayed_node {
+	u64 inode_id;
+	u64 bytes_reserved;
+	struct btrfs_root *root;
+	/* Used to add the node into the delayed root's node list. */
+	struct list_head n_list;
+	/*
+	 * Used to add the node into the prepare list, the nodes in this list
+	 * is waiting to be dealt with by the async worker.
+	 */
+	struct list_head p_list;
+	struct rb_root ins_root;
+	struct rb_root del_root;
+	struct mutex mutex;
+	struct btrfs_inode_item inode_item;
+	atomic_t refs;
+	u64 index_cnt;
+	bool in_list;
+	bool inode_dirty;
+	int count;
+};
+
+struct btrfs_delayed_item {
+	struct rb_node rb_node;
+	struct btrfs_key key;
+	struct list_head tree_list;	/* used for batch insert/delete items */
+	struct list_head readdir_list;	/* used for readdir items */
+	u64 bytes_reserved;
+	struct btrfs_block_rsv *block_rsv;
+	struct btrfs_delayed_node *delayed_node;
+	atomic_t refs;
+	int ins_or_del;
+	u32 data_len;
+	char data[0];
+};
+
+static inline void btrfs_init_delayed_root(
+				struct btrfs_delayed_root *delayed_root)
+{
+	atomic_set(&delayed_root->items, 0);
+	delayed_root->nodes = 0;
+	spin_lock_init(&delayed_root->lock);
+	init_waitqueue_head(&delayed_root->wait);
+	INIT_LIST_HEAD(&delayed_root->node_list);
+	INIT_LIST_HEAD(&delayed_root->prepare_list);
+}
+
+int btrfs_insert_delayed_dir_index(struct btrfs_trans_handle *trans,
+				   struct btrfs_root *root, const char *name,
+				   int name_len, struct inode *dir,
+				   struct btrfs_disk_key *disk_key, u8 type,
+				   u64 index);
+
+int btrfs_delete_delayed_dir_index(struct btrfs_trans_handle *trans,
+				   struct btrfs_root *root, struct inode *dir,
+				   u64 index);
+
+int btrfs_inode_delayed_dir_index_count(struct inode *inode);
+
+int btrfs_run_delayed_items(struct btrfs_trans_handle *trans,
+			    struct btrfs_root *root);
+
+void btrfs_balance_delayed_items(struct btrfs_root *root);
+
+int btrfs_commit_inode_delayed_items(struct btrfs_trans_handle *trans,
+				     struct inode *inode);
+/* Used for evicting the inode. */
+void btrfs_remove_delayed_node(struct inode *inode);
+void btrfs_kill_delayed_inode_items(struct inode *inode);
+
+
+int btrfs_delayed_update_inode(struct btrfs_trans_handle *trans,
+			       struct btrfs_root *root, struct inode *inode);
+
+/* Used for drop dead root */
+void btrfs_kill_all_delayed_nodes(struct btrfs_root *root);
+
+/* Used for readdir() */
+void btrfs_get_delayed_items(struct inode *inode, struct list_head *ins_list,
+			     struct list_head *del_list);
+void btrfs_put_delayed_items(struct list_head *ins_list,
+			     struct list_head *del_list);
+int btrfs_should_delete_dir_index(struct list_head *del_list,
+				  u64 index);
+int btrfs_readdir_delayed_dir_index(struct file *filp, void *dirent,
+				    filldir_t filldir,
+				    struct list_head *ins_list);
+
+/* for init */
+int __init btrfs_delayed_inode_init(void);
+void btrfs_delayed_inode_exit(void);
+#endif

fs/btrfs/delayed-ref.c

@@ -280,44 +280,6 @@ again:
 	return 1;
 }
 
-/*
- * This checks to see if there are any delayed refs in the
- * btree for a given bytenr.  It returns one if it finds any
- * and zero otherwise.
- *
- * If it only finds a head node, it returns 0.
- *
- * The idea is to use this when deciding if you can safely delete an
- * extent from the extent allocation tree.  There may be a pending
- * ref in the rbtree that adds or removes references, so as long as this
- * returns one you need to leave the BTRFS_EXTENT_ITEM in the extent
- * allocation tree.
- */
-int btrfs_delayed_ref_pending(struct btrfs_trans_handle *trans, u64 bytenr)
-{
-	struct btrfs_delayed_ref_node *ref;
-	struct btrfs_delayed_ref_root *delayed_refs;
-	struct rb_node *prev_node;
-	int ret = 0;
-
-	delayed_refs = &trans->transaction->delayed_refs;
-	spin_lock(&delayed_refs->lock);
-
-	ref = find_ref_head(&delayed_refs->root, bytenr, NULL);
-	if (ref) {
-		prev_node = rb_prev(&ref->rb_node);
-		if (!prev_node)
-			goto out;
-		ref = rb_entry(prev_node, struct btrfs_delayed_ref_node,
-			       rb_node);
-		if (ref->bytenr == bytenr)
-			ret = 1;
-	}
-out:
-	spin_unlock(&delayed_refs->lock);
-	return ret;
-}
-
 /*
  * helper function to update an extent delayed ref in the
  * rbtree.  existing and update must both have the same
@@ -747,79 +709,3 @@ btrfs_find_delayed_ref_head(struct btrfs_trans_handle *trans, u64 bytenr)
 		return btrfs_delayed_node_to_head(ref);
 	return NULL;
 }
-
-/*
- * add a delayed ref to the tree.  This does all of the accounting required
- * to make sure the delayed ref is eventually processed before this
- * transaction commits.
- *
- * The main point of this call is to add and remove a backreference in a single
- * shot, taking the lock only once, and only searching for the head node once.
- *
- * It is the same as doing a ref add and delete in two separate calls.
- */
-#if 0
-int btrfs_update_delayed_ref(struct btrfs_trans_handle *trans,
-			  u64 bytenr, u64 num_bytes, u64 orig_parent,
-			  u64 parent, u64 orig_ref_root, u64 ref_root,
-			  u64 orig_ref_generation, u64 ref_generation,
-			  u64 owner_objectid, int pin)
-{
-	struct btrfs_delayed_ref *ref;
-	struct btrfs_delayed_ref *old_ref;
-	struct btrfs_delayed_ref_head *head_ref;
-	struct btrfs_delayed_ref_root *delayed_refs;
-	int ret;
-
-	ref = kmalloc(sizeof(*ref), GFP_NOFS);
-	if (!ref)
-		return -ENOMEM;
-
-	old_ref = kmalloc(sizeof(*old_ref), GFP_NOFS);
-	if (!old_ref) {
-		kfree(ref);
-		return -ENOMEM;
-	}
-
-	/*
-	 * the parent = 0 case comes from cases where we don't actually
-	 * know the parent yet.  It will get updated later via a add/drop
-	 * pair.
-	 */
-	if (parent == 0)
-		parent = bytenr;
-	if (orig_parent == 0)
-		orig_parent = bytenr;
-
-	head_ref = kmalloc(sizeof(*head_ref), GFP_NOFS);
-	if (!head_ref) {
-		kfree(ref);
-		kfree(old_ref);
-		return -ENOMEM;
-	}
-	delayed_refs = &trans->transaction->delayed_refs;
-	spin_lock(&delayed_refs->lock);
-
-	/*
-	 * insert both the head node and the new ref without dropping
-	 * the spin lock
-	 */
-	ret = __btrfs_add_delayed_ref(trans, &head_ref->node, bytenr, num_bytes,
-				      (u64)-1, 0, 0, 0,
-				      BTRFS_UPDATE_DELAYED_HEAD, 0);
-	BUG_ON(ret);
-
-	ret = __btrfs_add_delayed_ref(trans, &ref->node, bytenr, num_bytes,
-				      parent, ref_root, ref_generation,
-				      owner_objectid, BTRFS_ADD_DELAYED_REF, 0);
-	BUG_ON(ret);
-
-	ret = __btrfs_add_delayed_ref(trans, &old_ref->node, bytenr, num_bytes,
-				      orig_parent, orig_ref_root,
-				      orig_ref_generation, owner_objectid,
-				      BTRFS_DROP_DELAYED_REF, pin);
-	BUG_ON(ret);
-	spin_unlock(&delayed_refs->lock);
-	return 0;
-}
-#endif

fs/btrfs/delayed-ref.h

@@ -166,12 +166,6 @@ int btrfs_add_delayed_extent_op(struct btrfs_trans_handle *trans,
 
 struct btrfs_delayed_ref_head *
 btrfs_find_delayed_ref_head(struct btrfs_trans_handle *trans, u64 bytenr);
-int btrfs_delayed_ref_pending(struct btrfs_trans_handle *trans, u64 bytenr);
-int btrfs_update_delayed_ref(struct btrfs_trans_handle *trans,
-			  u64 bytenr, u64 num_bytes, u64 orig_parent,
-			  u64 parent, u64 orig_ref_root, u64 ref_root,
-			  u64 orig_ref_generation, u64 ref_generation,
-			  u64 owner_objectid, int pin);
 int btrfs_delayed_ref_lock(struct btrfs_trans_handle *trans,
 			   struct btrfs_delayed_ref_head *head);
 int btrfs_find_ref_cluster(struct btrfs_trans_handle *trans,

fs/btrfs/dir-item.c

@@ -50,7 +50,6 @@ static struct btrfs_dir_item *insert_with_overflow(struct btrfs_trans_handle
 		if (di)
 			return ERR_PTR(-EEXIST);
 		ret = btrfs_extend_item(trans, root, path, data_size);
-		WARN_ON(ret > 0);
 	}
 	if (ret < 0)
 		return ERR_PTR(ret);
@@ -124,8 +123,9 @@ int btrfs_insert_xattr_item(struct btrfs_trans_handle *trans,
  * to use for the second index (if one is created).
  */
 int btrfs_insert_dir_item(struct btrfs_trans_handle *trans, struct btrfs_root
-			  *root, const char *name, int name_len, u64 dir,
-			  struct btrfs_key *location, u8 type, u64 index)
+			  *root, const char *name, int name_len,
+			  struct inode *dir, struct btrfs_key *location,
+			  u8 type, u64 index)
 {
 	int ret = 0;
 	int ret2 = 0;
@@ -137,13 +137,17 @@ int btrfs_insert_dir_item(struct btrfs_trans_handle *trans, struct btrfs_root
 	struct btrfs_disk_key disk_key;
 	u32 data_size;
 
-	key.objectid = dir;
+	key.objectid = btrfs_ino(dir);
 	btrfs_set_key_type(&key, BTRFS_DIR_ITEM_KEY);
 	key.offset = btrfs_name_hash(name, name_len);
 
 	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
 	path->leave_spinning = 1;
 
+	btrfs_cpu_key_to_disk(&disk_key, location);
+
 	data_size = sizeof(*dir_item) + name_len;
 	dir_item = insert_with_overflow(trans, root, path, &key, data_size,
 					name, name_len);
@@ -155,7 +159,6 @@ int btrfs_insert_dir_item(struct btrfs_trans_handle *trans, struct btrfs_root
 	}
 
 	leaf = path->nodes[0];
-	btrfs_cpu_key_to_disk(&disk_key, location);
 	btrfs_set_dir_item_key(leaf, dir_item, &disk_key);
 	btrfs_set_dir_type(leaf, dir_item, type);
 	btrfs_set_dir_data_len(leaf, dir_item, 0);
@@ -172,29 +175,11 @@ second_insert:
 		ret = 0;
 		goto out_free;
 	}
-	btrfs_release_path(root, path);
-
-	btrfs_set_key_type(&key, BTRFS_DIR_INDEX_KEY);
-	key.offset = index;
-	dir_item = insert_with_overflow(trans, root, path, &key, data_size,
-					name, name_len);
-	if (IS_ERR(dir_item)) {
-		ret2 = PTR_ERR(dir_item);
-		goto out_free;
-	}
-	leaf = path->nodes[0];
-	btrfs_cpu_key_to_disk(&disk_key, location);
-	btrfs_set_dir_item_key(leaf, dir_item, &disk_key);
-	btrfs_set_dir_type(leaf, dir_item, type);
-	btrfs_set_dir_data_len(leaf, dir_item, 0);
-	btrfs_set_dir_name_len(leaf, dir_item, name_len);
-	btrfs_set_dir_transid(leaf, dir_item, trans->transid);
-	name_ptr = (unsigned long)(dir_item + 1);
-	write_extent_buffer(leaf, name, name_ptr, name_len);
-	btrfs_mark_buffer_dirty(leaf);
+	btrfs_release_path(path);
 
+	ret2 = btrfs_insert_delayed_dir_index(trans, root, name, name_len, dir,
+					      &disk_key, type, index);
 out_free:
-
 	btrfs_free_path(path);
 	if (ret)
 		return ret;
@@ -452,7 +437,7 @@ int verify_dir_item(struct btrfs_root *root,
 		namelen = XATTR_NAME_MAX;
 
 	if (btrfs_dir_name_len(leaf, dir_item) > namelen) {
-		printk(KERN_CRIT "btrfS: invalid dir item name len: %u\n",
+		printk(KERN_CRIT "btrfs: invalid dir item name len: %u\n",
 		       (unsigned)btrfs_dir_data_len(leaf, dir_item));
 		return 1;
 	}

fs/btrfs/disk-io.c

@@ -29,6 +29,7 @@
 #include <linux/crc32c.h>
 #include <linux/slab.h>
 #include <linux/migrate.h>
+#include <linux/ratelimit.h>
 #include <asm/unaligned.h>
 #include "compat.h"
 #include "ctree.h"
@@ -41,6 +42,7 @@
 #include "locking.h"
 #include "tree-log.h"
 #include "free-space-cache.h"
+#include "inode-map.h"
 
 static struct extent_io_ops btree_extent_io_ops;
 static void end_workqueue_fn(struct btrfs_work *work);
@@ -137,7 +139,7 @@ static const char *btrfs_eb_name[BTRFS_MAX_LEVEL + 1] = {
  * that covers the entire device
  */
 static struct extent_map *btree_get_extent(struct inode *inode,
-		struct page *page, size_t page_offset, u64 start, u64 len,
+		struct page *page, size_t pg_offset, u64 start, u64 len,
 		int create)
 {
 	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
@@ -154,7 +156,7 @@ static struct extent_map *btree_get_extent(struct inode *inode,
 	}
 	read_unlock(&em_tree->lock);
 
-	em = alloc_extent_map(GFP_NOFS);
+	em = alloc_extent_map();
 	if (!em) {
 		em = ERR_PTR(-ENOMEM);
 		goto out;
@@ -254,14 +256,12 @@ static int csum_tree_block(struct btrfs_root *root, struct extent_buffer *buf,
 			memcpy(&found, result, csum_size);
 
 			read_extent_buffer(buf, &val, 0, csum_size);
-			if (printk_ratelimit()) {
-				printk(KERN_INFO "btrfs: %s checksum verify "
+			printk_ratelimited(KERN_INFO "btrfs: %s checksum verify "
 				       "failed on %llu wanted %X found %X "
 				       "level %d\n",
 				       root->fs_info->sb->s_id,
 				       (unsigned long long)buf->start, val, found,
 				       btrfs_header_level(buf));
-			}
 			if (result != (char *)&inline_result)
 				kfree(result);
 			return 1;
@@ -296,13 +296,11 @@ static int verify_parent_transid(struct extent_io_tree *io_tree,
 		ret = 0;
 		goto out;
 	}
-	if (printk_ratelimit()) {
-		printk("parent transid verify failed on %llu wanted %llu "
+	printk_ratelimited("parent transid verify failed on %llu wanted %llu "
 		       "found %llu\n",
 		       (unsigned long long)eb->start,
 		       (unsigned long long)parent_transid,
 		       (unsigned long long)btrfs_header_generation(eb));
-	}
 	ret = 1;
 	clear_extent_buffer_uptodate(io_tree, eb, &cached_state);
 out:
@@ -380,7 +378,7 @@ static int csum_dirty_buffer(struct btrfs_root *root, struct page *page)
 	len = page->private >> 2;
 	WARN_ON(len == 0);
 
-	eb = alloc_extent_buffer(tree, start, len, page, GFP_NOFS);
+	eb = alloc_extent_buffer(tree, start, len, page);
 	if (eb == NULL) {
 		WARN_ON(1);
 		goto out;
@@ -525,7 +523,7 @@ static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end,
 	len = page->private >> 2;
 	WARN_ON(len == 0);
 
-	eb = alloc_extent_buffer(tree, start, len, page, GFP_NOFS);
+	eb = alloc_extent_buffer(tree, start, len, page);
 	if (eb == NULL) {
 		ret = -EIO;
 		goto out;
@@ -533,12 +531,10 @@ static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end,
 
 	found_start = btrfs_header_bytenr(eb);
 	if (found_start != start) {
-		if (printk_ratelimit()) {
-			printk(KERN_INFO "btrfs bad tree block start "
+		printk_ratelimited(KERN_INFO "btrfs bad tree block start "
 			       "%llu %llu\n",
 			       (unsigned long long)found_start,
 			       (unsigned long long)eb->start);
-		}
 		ret = -EIO;
 		goto err;
 	}
@@ -550,10 +546,8 @@ static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end,
 		goto err;
 	}
 	if (check_tree_block_fsid(root, eb)) {
-		if (printk_ratelimit()) {
-			printk(KERN_INFO "btrfs bad fsid on block %llu\n",
+		printk_ratelimited(KERN_INFO "btrfs bad fsid on block %llu\n",
 			       (unsigned long long)eb->start);
-		}
 		ret = -EIO;
 		goto err;
 	}
@@ -650,12 +644,6 @@ unsigned long btrfs_async_submit_limit(struct btrfs_fs_info *info)
 	return 256 * limit;
 }
 
-int btrfs_congested_async(struct btrfs_fs_info *info, int iodone)
-{
-	return atomic_read(&info->nr_async_bios) >
-		btrfs_async_submit_limit(info);
-}
-
 static void run_one_async_start(struct btrfs_work *work)
 {
 	struct async_submit_bio *async;
@@ -963,7 +951,7 @@ struct extent_buffer *btrfs_find_tree_block(struct btrfs_root *root,
 	struct inode *btree_inode = root->fs_info->btree_inode;
 	struct extent_buffer *eb;
 	eb = find_extent_buffer(&BTRFS_I(btree_inode)->io_tree,
-				bytenr, blocksize, GFP_NOFS);
+				bytenr, blocksize);
 	return eb;
 }
 
@@ -974,7 +962,7 @@ struct extent_buffer *btrfs_find_create_tree_block(struct btrfs_root *root,
 	struct extent_buffer *eb;
 
 	eb = alloc_extent_buffer(&BTRFS_I(btree_inode)->io_tree,
-				 bytenr, blocksize, NULL, GFP_NOFS);
+				 bytenr, blocksize, NULL);
 	return eb;
 }
 
@@ -1058,13 +1046,13 @@ static int __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize,
 	root->name = NULL;
 	root->in_sysfs = 0;
 	root->inode_tree = RB_ROOT;
+	INIT_RADIX_TREE(&root->delayed_nodes_tree, GFP_ATOMIC);
 	root->block_rsv = NULL;
 	root->orphan_block_rsv = NULL;
 
 	INIT_LIST_HEAD(&root->dirty_list);
 	INIT_LIST_HEAD(&root->orphan_list);
 	INIT_LIST_HEAD(&root->root_list);
-	spin_lock_init(&root->node_lock);
 	spin_lock_init(&root->orphan_lock);
 	spin_lock_init(&root->inode_lock);
 	spin_lock_init(&root->accounting_lock);
@@ -1080,7 +1068,7 @@ static int __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize,
 	root->log_transid = 0;
 	root->last_log_commit = 0;
 	extent_io_tree_init(&root->dirty_log_pages,
-			     fs_info->btree_inode->i_mapping, GFP_NOFS);
+			     fs_info->btree_inode->i_mapping);
 
 	memset(&root->root_key, 0, sizeof(root->root_key));
 	memset(&root->root_item, 0, sizeof(root->root_item));
@@ -1283,21 +1271,6 @@ out:
 	return root;
 }
 
-struct btrfs_root *btrfs_lookup_fs_root(struct btrfs_fs_info *fs_info,
-					u64 root_objectid)
-{
-	struct btrfs_root *root;
-
-	if (root_objectid == BTRFS_ROOT_TREE_OBJECTID)
-		return fs_info->tree_root;
-	if (root_objectid == BTRFS_EXTENT_TREE_OBJECTID)
-		return fs_info->extent_root;
-
-	root = radix_tree_lookup(&fs_info->fs_roots_radix,
-				 (unsigned long)root_objectid);
-	return root;
-}
-
 struct btrfs_root *btrfs_read_fs_root_no_name(struct btrfs_fs_info *fs_info,
 					      struct btrfs_key *location)
 {
@@ -1326,6 +1299,19 @@ again:
 	if (IS_ERR(root))
 		return root;
 
+	root->free_ino_ctl = kzalloc(sizeof(*root->free_ino_ctl), GFP_NOFS);
+	if (!root->free_ino_ctl)
+		goto fail;
+	root->free_ino_pinned = kzalloc(sizeof(*root->free_ino_pinned),
+					GFP_NOFS);
+	if (!root->free_ino_pinned)
+		goto fail;
+
+	btrfs_init_free_ino_ctl(root);
+	mutex_init(&root->fs_commit_mutex);
+	spin_lock_init(&root->cache_lock);
+	init_waitqueue_head(&root->cache_wait);
+
 	set_anon_super(&root->anon_super, NULL);
 
 	if (btrfs_root_refs(&root->root_item) == 0) {
@@ -1369,41 +1355,6 @@ fail:
 	return ERR_PTR(ret);
 }
 
-struct btrfs_root *btrfs_read_fs_root(struct btrfs_fs_info *fs_info,
-				      struct btrfs_key *location,
-				      const char *name, int namelen)
-{
-	return btrfs_read_fs_root_no_name(fs_info, location);
-#if 0
-	struct btrfs_root *root;
-	int ret;
-
-	root = btrfs_read_fs_root_no_name(fs_info, location);
-	if (!root)
-		return NULL;
-
-	if (root->in_sysfs)
-		return root;
-
-	ret = btrfs_set_root_name(root, name, namelen);
-	if (ret) {
-		free_extent_buffer(root->node);
-		kfree(root);
-		return ERR_PTR(ret);
-	}
-
-	ret = btrfs_sysfs_add_root(root);
-	if (ret) {
-		free_extent_buffer(root->node);
-		kfree(root->name);
-		kfree(root);
-		return ERR_PTR(ret);
-	}
-	root->in_sysfs = 1;
-	return root;
-#endif
-}
-
 static int btrfs_congested_fn(void *congested_data, int bdi_bits)
 {
 	struct btrfs_fs_info *info = (struct btrfs_fs_info *)congested_data;
@@ -1411,7 +1362,8 @@ static int btrfs_congested_fn(void *congested_data, int bdi_bits)
 	struct btrfs_device *device;
 	struct backing_dev_info *bdi;
 
-	list_for_each_entry(device, &info->fs_devices->devices, dev_list) {
+	rcu_read_lock();
+	list_for_each_entry_rcu(device, &info->fs_devices->devices, dev_list) {
 		if (!device->bdev)
 			continue;
 		bdi = blk_get_backing_dev_info(device->bdev);
@@ -1420,6 +1372,7 @@ static int btrfs_congested_fn(void *congested_data, int bdi_bits)
 			break;
 		}
 	}
+	rcu_read_unlock();
 	return ret;
 }
 
@@ -1522,6 +1475,7 @@ static int cleaner_kthread(void *arg)
 			btrfs_run_delayed_iputs(root);
 			btrfs_clean_old_snapshots(root);
 			mutex_unlock(&root->fs_info->cleaner_mutex);
+			btrfs_run_defrag_inodes(root->fs_info);
 		}
 
 		if (freezing(current)) {
@@ -1611,7 +1565,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	struct btrfs_root *csum_root = kzalloc(sizeof(struct btrfs_root),
 						 GFP_NOFS);
 	struct btrfs_root *tree_root = btrfs_sb(sb);
-	struct btrfs_fs_info *fs_info = tree_root->fs_info;
+	struct btrfs_fs_info *fs_info = NULL;
 	struct btrfs_root *chunk_root = kzalloc(sizeof(struct btrfs_root),
 						GFP_NOFS);
 	struct btrfs_root *dev_root = kzalloc(sizeof(struct btrfs_root),
@@ -1623,11 +1577,12 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 
 	struct btrfs_super_block *disk_super;
 
-	if (!extent_root || !tree_root || !fs_info ||
+	if (!extent_root || !tree_root || !tree_root->fs_info ||
 	    !chunk_root || !dev_root || !csum_root) {
 		err = -ENOMEM;
 		goto fail;
 	}
+	fs_info = tree_root->fs_info;
 
 	ret = init_srcu_struct(&fs_info->subvol_srcu);
 	if (ret) {
@@ -1662,6 +1617,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	spin_lock_init(&fs_info->ref_cache_lock);
 	spin_lock_init(&fs_info->fs_roots_radix_lock);
 	spin_lock_init(&fs_info->delayed_iput_lock);
+	spin_lock_init(&fs_info->defrag_inodes_lock);
 
 	init_completion(&fs_info->kobj_unregister);
 	fs_info->tree_root = tree_root;
@@ -1684,15 +1640,35 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	atomic_set(&fs_info->async_delalloc_pages, 0);
 	atomic_set(&fs_info->async_submit_draining, 0);
 	atomic_set(&fs_info->nr_async_bios, 0);
+	atomic_set(&fs_info->defrag_running, 0);
 	fs_info->sb = sb;
 	fs_info->max_inline = 8192 * 1024;
 	fs_info->metadata_ratio = 0;
+	fs_info->defrag_inodes = RB_ROOT;
 
 	fs_info->thread_pool_size = min_t(unsigned long,
 					  num_online_cpus() + 2, 8);
 
 	INIT_LIST_HEAD(&fs_info->ordered_extents);
 	spin_lock_init(&fs_info->ordered_extent_lock);
+	fs_info->delayed_root = kmalloc(sizeof(struct btrfs_delayed_root),
+					GFP_NOFS);
+	if (!fs_info->delayed_root) {
+		err = -ENOMEM;
+		goto fail_iput;
+	}
+	btrfs_init_delayed_root(fs_info->delayed_root);
+
+	mutex_init(&fs_info->scrub_lock);
+	atomic_set(&fs_info->scrubs_running, 0);
+	atomic_set(&fs_info->scrub_pause_req, 0);
+	atomic_set(&fs_info->scrubs_paused, 0);
+	atomic_set(&fs_info->scrub_cancel_req, 0);
+	init_waitqueue_head(&fs_info->scrub_pause_wait);
+	init_rwsem(&fs_info->scrub_super_lock);
+	fs_info->scrub_workers_refcnt = 0;
+	btrfs_init_workers(&fs_info->scrub_workers, "scrub",
+			   fs_info->thread_pool_size, &fs_info->generic_worker);
 
 	sb->s_blocksize = 4096;
 	sb->s_blocksize_bits = blksize_bits(4096);
@@ -1711,10 +1687,8 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 
 	RB_CLEAR_NODE(&BTRFS_I(fs_info->btree_inode)->rb_node);
 	extent_io_tree_init(&BTRFS_I(fs_info->btree_inode)->io_tree,
-			     fs_info->btree_inode->i_mapping,
-			     GFP_NOFS);
-	extent_map_tree_init(&BTRFS_I(fs_info->btree_inode)->extent_tree,
-			     GFP_NOFS);
+			     fs_info->btree_inode->i_mapping);
+	extent_map_tree_init(&BTRFS_I(fs_info->btree_inode)->extent_tree);
 
 	BTRFS_I(fs_info->btree_inode)->io_tree.ops = &btree_extent_io_ops;
 
@@ -1728,9 +1702,9 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	fs_info->block_group_cache_tree = RB_ROOT;
 
 	extent_io_tree_init(&fs_info->freed_extents[0],
-			     fs_info->btree_inode->i_mapping, GFP_NOFS);
+			     fs_info->btree_inode->i_mapping);
 	extent_io_tree_init(&fs_info->freed_extents[1],
-			     fs_info->btree_inode->i_mapping, GFP_NOFS);
+			     fs_info->btree_inode->i_mapping);
 	fs_info->pinned_extents = &fs_info->freed_extents[0];
 	fs_info->do_barriers = 1;
 
@@ -1760,7 +1734,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	bh = btrfs_read_dev_super(fs_devices->latest_bdev);
 	if (!bh) {
 		err = -EINVAL;
-		goto fail_iput;
+		goto fail_alloc;
 	}
 
 	memcpy(&fs_info->super_copy, bh->b_data, sizeof(fs_info->super_copy));
@@ -1772,7 +1746,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 
 	disk_super = &fs_info->super_copy;
 	if (!btrfs_super_root(disk_super))
-		goto fail_iput;
+		goto fail_alloc;
 
 	/* check FS state, whether FS is broken. */
 	fs_info->fs_state |= btrfs_super_flags(disk_super);
@@ -1788,7 +1762,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	ret = btrfs_parse_options(tree_root, options);
 	if (ret) {
 		err = ret;
-		goto fail_iput;
+		goto fail_alloc;
 	}
 
 	features = btrfs_super_incompat_flags(disk_super) &
@@ -1798,7 +1772,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 		       "unsupported optional features (%Lx).\n",
 		       (unsigned long long)features);
 		err = -EINVAL;
-		goto fail_iput;
+		goto fail_alloc;
 	}
 
 	features = btrfs_super_incompat_flags(disk_super);
@@ -1814,7 +1788,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 		       "unsupported option features (%Lx).\n",
 		       (unsigned long long)features);
 		err = -EINVAL;
-		goto fail_iput;
+		goto fail_alloc;
 	}
 
 	btrfs_init_workers(&fs_info->generic_worker,
@@ -1861,6 +1835,9 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 			   &fs_info->generic_worker);
 	btrfs_init_workers(&fs_info->endio_freespace_worker, "freespace-write",
 			   1, &fs_info->generic_worker);
+	btrfs_init_workers(&fs_info->delayed_workers, "delayed-meta",
+			   fs_info->thread_pool_size,
+			   &fs_info->generic_worker);
 
 	/*
 	 * endios are largely parallel and should have a very
@@ -1882,6 +1859,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	btrfs_start_workers(&fs_info->endio_meta_write_workers, 1);
 	btrfs_start_workers(&fs_info->endio_write_workers, 1);
 	btrfs_start_workers(&fs_info->endio_freespace_worker, 1);
+	btrfs_start_workers(&fs_info->delayed_workers, 1);
 
 	fs_info->bdi.ra_pages *= btrfs_super_num_devices(disk_super);
 	fs_info->bdi.ra_pages = max(fs_info->bdi.ra_pages,
@@ -2138,6 +2116,9 @@ fail_sb_buffer:
 	btrfs_stop_workers(&fs_info->endio_write_workers);
 	btrfs_stop_workers(&fs_info->endio_freespace_worker);
 	btrfs_stop_workers(&fs_info->submit_workers);
+	btrfs_stop_workers(&fs_info->delayed_workers);
+fail_alloc:
+	kfree(fs_info->delayed_root);
 fail_iput:
 	invalidate_inode_pages2(fs_info->btree_inode->i_mapping);
 	iput(fs_info->btree_inode);
@@ -2165,11 +2146,9 @@ static void btrfs_end_buffer_write_sync(struct buffer_head *bh, int uptodate)
 	if (uptodate) {
 		set_buffer_uptodate(bh);
 	} else {
-		if (printk_ratelimit()) {
-			printk(KERN_WARNING "lost page write due to "
+		printk_ratelimited(KERN_WARNING "lost page write due to "
 					"I/O error on %s\n",
 				       bdevname(bh->b_bdev, b));
-		}
 		/* note, we dont' set_buffer_write_io_error because we have
 		 * our own ways of dealing with the IO errors
 		 */
@@ -2333,7 +2312,7 @@ int write_all_supers(struct btrfs_root *root, int max_mirrors)
 
 	mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
 	head = &root->fs_info->fs_devices->devices;
-	list_for_each_entry(dev, head, dev_list) {
+	list_for_each_entry_rcu(dev, head, dev_list) {
 		if (!dev->bdev) {
 			total_errors++;
 			continue;
@@ -2366,7 +2345,7 @@ int write_all_supers(struct btrfs_root *root, int max_mirrors)
 	}
 
 	total_errors = 0;
-	list_for_each_entry(dev, head, dev_list) {
+	list_for_each_entry_rcu(dev, head, dev_list) {
 		if (!dev->bdev)
 			continue;
 		if (!dev->in_fs_metadata || !dev->writeable)
@@ -2404,12 +2383,15 @@ int btrfs_free_fs_root(struct btrfs_fs_info *fs_info, struct btrfs_root *root)
 	if (btrfs_root_refs(&root->root_item) == 0)
 		synchronize_srcu(&fs_info->subvol_srcu);
 
+	__btrfs_remove_free_space_cache(root->free_ino_pinned);
+	__btrfs_remove_free_space_cache(root->free_ino_ctl);
 	free_fs_root(root);
 	return 0;
 }
 
 static void free_fs_root(struct btrfs_root *root)
 {
+	iput(root->cache_inode);
 	WARN_ON(!RB_EMPTY_ROOT(&root->inode_tree));
 	if (root->anon_super.s_dev) {
 		down_write(&root->anon_super.s_umount);
@@ -2417,6 +2399,8 @@ static void free_fs_root(struct btrfs_root *root)
 	}
 	free_extent_buffer(root->node);
 	free_extent_buffer(root->commit_root);
+	kfree(root->free_ino_ctl);
+	kfree(root->free_ino_pinned);
 	kfree(root->name);
 	kfree(root);
 }
@@ -2520,6 +2504,15 @@ int close_ctree(struct btrfs_root *root)
 	fs_info->closing = 1;
 	smp_mb();
 
+	btrfs_scrub_cancel(root);
+
+	/* wait for any defraggers to finish */
+	wait_event(fs_info->transaction_wait,
+		   (atomic_read(&fs_info->defrag_running) == 0));
+
+	/* clear out the rbtree of defraggable inodes */
+	btrfs_run_defrag_inodes(root->fs_info);
+
 	btrfs_put_block_group_cache(fs_info);
 
 	/*
@@ -2578,6 +2571,7 @@ int close_ctree(struct btrfs_root *root)
 	del_fs_roots(fs_info);
 
 	iput(fs_info->btree_inode);
+	kfree(fs_info->delayed_root);
 
 	btrfs_stop_workers(&fs_info->generic_worker);
 	btrfs_stop_workers(&fs_info->fixup_workers);
@@ -2589,6 +2583,7 @@ int close_ctree(struct btrfs_root *root)
 	btrfs_stop_workers(&fs_info->endio_write_workers);
 	btrfs_stop_workers(&fs_info->endio_freespace_worker);
 	btrfs_stop_workers(&fs_info->submit_workers);
+	btrfs_stop_workers(&fs_info->delayed_workers);
 
 	btrfs_close_devices(fs_info->fs_devices);
 	btrfs_mapping_tree_free(&fs_info->mapping_tree);
@@ -2662,6 +2657,29 @@ void btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr)
 	u64 num_dirty;
 	unsigned long thresh = 32 * 1024 * 1024;
 
+	if (current->flags & PF_MEMALLOC)
+		return;
+
+	btrfs_balance_delayed_items(root);
+
+	num_dirty = root->fs_info->dirty_metadata_bytes;
+
+	if (num_dirty > thresh) {
+		balance_dirty_pages_ratelimited_nr(
+				   root->fs_info->btree_inode->i_mapping, 1);
+	}
+	return;
+}
+
+void __btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr)
+{
+	/*
+	 * looks as though older kernels can get into trouble with
+	 * this code, they end up stuck in balance_dirty_pages forever
+	 */
+	u64 num_dirty;
+	unsigned long thresh = 32 * 1024 * 1024;
+
 	if (current->flags & PF_MEMALLOC)
 		return;
 
@@ -2697,7 +2715,7 @@ int btree_lock_page_hook(struct page *page)
 		goto out;
 
 	len = page->private >> 2;
-	eb = find_extent_buffer(io_tree, bytenr, len, GFP_NOFS);
+	eb = find_extent_buffer(io_tree, bytenr, len);
 	if (!eb)
 		goto out;
 

fs/btrfs/disk-io.h

@@ -55,35 +55,20 @@ int btrfs_commit_super(struct btrfs_root *root);
 int btrfs_error_commit_super(struct btrfs_root *root);
 struct extent_buffer *btrfs_find_tree_block(struct btrfs_root *root,
 					    u64 bytenr, u32 blocksize);
-struct btrfs_root *btrfs_lookup_fs_root(struct btrfs_fs_info *fs_info,
-					u64 root_objectid);
-struct btrfs_root *btrfs_read_fs_root(struct btrfs_fs_info *fs_info,
-				      struct btrfs_key *location,
-				      const char *name, int namelen);
 struct btrfs_root *btrfs_read_fs_root_no_radix(struct btrfs_root *tree_root,
 					       struct btrfs_key *location);
 struct btrfs_root *btrfs_read_fs_root_no_name(struct btrfs_fs_info *fs_info,
 					      struct btrfs_key *location);
 int btrfs_cleanup_fs_roots(struct btrfs_fs_info *fs_info);
-int btrfs_insert_dev_radix(struct btrfs_root *root,
-			   struct block_device *bdev,
-			   u64 device_id,
-			   u64 block_start,
-			   u64 num_blocks);
 void btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr);
+void __btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr);
 int btrfs_free_fs_root(struct btrfs_fs_info *fs_info, struct btrfs_root *root);
 void btrfs_mark_buffer_dirty(struct extent_buffer *buf);
-void btrfs_mark_buffer_dirty_nonblocking(struct extent_buffer *buf);
 int btrfs_buffer_uptodate(struct extent_buffer *buf, u64 parent_transid);
 int btrfs_set_buffer_uptodate(struct extent_buffer *buf);
-int wait_on_tree_block_writeback(struct btrfs_root *root,
-				 struct extent_buffer *buf);
 int btrfs_read_buffer(struct extent_buffer *buf, u64 parent_transid);
 u32 btrfs_csum_data(struct btrfs_root *root, char *data, u32 seed, size_t len);
 void btrfs_csum_final(u32 crc, char *result);
-int btrfs_open_device(struct btrfs_device *dev);
-int btrfs_verify_block_csum(struct btrfs_root *root,
-			    struct extent_buffer *buf);
 int btrfs_bio_wq_end_io(struct btrfs_fs_info *info, struct bio *bio,
 			int metadata);
 int btrfs_wq_submit_bio(struct btrfs_fs_info *fs_info, struct inode *inode,
@@ -91,8 +76,6 @@ int btrfs_wq_submit_bio(struct btrfs_fs_info *fs_info, struct inode *inode,
 			unsigned long bio_flags, u64 bio_offset,
 			extent_submit_bio_hook_t *submit_bio_start,
 			extent_submit_bio_hook_t *submit_bio_done);
-
-int btrfs_congested_async(struct btrfs_fs_info *info, int iodone);
 unsigned long btrfs_async_submit_limit(struct btrfs_fs_info *info);
 int btrfs_write_tree_block(struct extent_buffer *buf);
 int btrfs_wait_tree_block_writeback(struct extent_buffer *buf);

fs/btrfs/export.c

@@ -32,7 +32,7 @@ static int btrfs_encode_fh(struct dentry *dentry, u32 *fh, int *max_len,
 	len  = BTRFS_FID_SIZE_NON_CONNECTABLE;
 	type = FILEID_BTRFS_WITHOUT_PARENT;
 
-	fid->objectid = inode->i_ino;
+	fid->objectid = btrfs_ino(inode);
 	fid->root_objectid = BTRFS_I(inode)->root->objectid;
 	fid->gen = inode->i_generation;
 
@@ -178,13 +178,13 @@ static struct dentry *btrfs_get_parent(struct dentry *child)
 	if (!path)
 		return ERR_PTR(-ENOMEM);
 
-	if (dir->i_ino == BTRFS_FIRST_FREE_OBJECTID) {
+	if (btrfs_ino(dir) == BTRFS_FIRST_FREE_OBJECTID) {
 		key.objectid = root->root_key.objectid;
 		key.type = BTRFS_ROOT_BACKREF_KEY;
 		key.offset = (u64)-1;
 		root = root->fs_info->tree_root;
 	} else {
-		key.objectid = dir->i_ino;
+		key.objectid = btrfs_ino(dir);
 		key.type = BTRFS_INODE_REF_KEY;
 		key.offset = (u64)-1;
 	}
@@ -244,6 +244,7 @@ static int btrfs_get_name(struct dentry *parent, char *name,
 	struct btrfs_key key;
 	int name_len;
 	int ret;
+	u64 ino;
 
 	if (!dir || !inode)
 		return -EINVAL;
@@ -251,19 +252,21 @@ static int btrfs_get_name(struct dentry *parent, char *name,
 	if (!S_ISDIR(dir->i_mode))
 		return -EINVAL;
 
+	ino = btrfs_ino(inode);
+
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
 	path->leave_spinning = 1;
 
-	if (inode->i_ino == BTRFS_FIRST_FREE_OBJECTID) {
+	if (ino == BTRFS_FIRST_FREE_OBJECTID) {
 		key.objectid = BTRFS_I(inode)->root->root_key.objectid;
 		key.type = BTRFS_ROOT_BACKREF_KEY;
 		key.offset = (u64)-1;
 		root = root->fs_info->tree_root;
 	} else {
-		key.objectid = inode->i_ino;
-		key.offset = dir->i_ino;
+		key.objectid = ino;
+		key.offset = btrfs_ino(dir);
 		key.type = BTRFS_INODE_REF_KEY;
 	}
 
@@ -272,7 +275,7 @@ static int btrfs_get_name(struct dentry *parent, char *name,
 		btrfs_free_path(path);
 		return ret;
 	} else if (ret > 0) {
-		if (inode->i_ino == BTRFS_FIRST_FREE_OBJECTID) {
+		if (ino == BTRFS_FIRST_FREE_OBJECTID) {
 			path->slots[0]--;
 		} else {
 			btrfs_free_path(path);
@@ -281,11 +284,11 @@ static int btrfs_get_name(struct dentry *parent, char *name,
 	}
 	leaf = path->nodes[0];
 
-	if (inode->i_ino == BTRFS_FIRST_FREE_OBJECTID) {
-	       rref = btrfs_item_ptr(leaf, path->slots[0],
+	if (ino == BTRFS_FIRST_FREE_OBJECTID) {
+		rref = btrfs_item_ptr(leaf, path->slots[0],
 				     struct btrfs_root_ref);
-	       name_ptr = (unsigned long)(rref + 1);
-	       name_len = btrfs_root_ref_name_len(leaf, rref);
+		name_ptr = (unsigned long)(rref + 1);
+		name_len = btrfs_root_ref_name_len(leaf, rref);
 	} else {
 		iref = btrfs_item_ptr(leaf, path->slots[0],
 				      struct btrfs_inode_ref);

fs/btrfs/extent_io.c

@@ -103,7 +103,7 @@ void extent_io_exit(void)
 }
 
 void extent_io_tree_init(struct extent_io_tree *tree,
-			  struct address_space *mapping, gfp_t mask)
+			 struct address_space *mapping)
 {
 	tree->state = RB_ROOT;
 	INIT_RADIX_TREE(&tree->buffer, GFP_ATOMIC);
@@ -441,6 +441,15 @@ static int clear_state_bit(struct extent_io_tree *tree,
 	return ret;
 }
 
+static struct extent_state *
+alloc_extent_state_atomic(struct extent_state *prealloc)
+{
+	if (!prealloc)
+		prealloc = alloc_extent_state(GFP_ATOMIC);
+
+	return prealloc;
+}
+
 /*
  * clear some bits on a range in the tree.  This may require splitting
  * or inserting elements in the tree, so the gfp mask is used to
@@ -531,8 +540,8 @@ hit_next:
 	 */
 
 	if (state->start < start) {
-		if (!prealloc)
-			prealloc = alloc_extent_state(GFP_ATOMIC);
+		prealloc = alloc_extent_state_atomic(prealloc);
+		BUG_ON(!prealloc);
 		err = split_state(tree, state, prealloc, start);
 		BUG_ON(err == -EEXIST);
 		prealloc = NULL;
@@ -553,8 +562,8 @@ hit_next:
 	 * on the first half
 	 */
 	if (state->start <= end && state->end > end) {
-		if (!prealloc)
-			prealloc = alloc_extent_state(GFP_ATOMIC);
+		prealloc = alloc_extent_state_atomic(prealloc);
+		BUG_ON(!prealloc);
 		err = split_state(tree, state, prealloc, end + 1);
 		BUG_ON(err == -EEXIST);
 		if (wake)
@@ -727,8 +736,7 @@ int set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
 again:
 	if (!prealloc && (mask & __GFP_WAIT)) {
 		prealloc = alloc_extent_state(mask);
-		if (!prealloc)
-			return -ENOMEM;
+		BUG_ON(!prealloc);
 	}
 
 	spin_lock(&tree->lock);
@@ -745,6 +753,8 @@ again:
 	 */
 	node = tree_search(tree, start);
 	if (!node) {
+		prealloc = alloc_extent_state_atomic(prealloc);
+		BUG_ON(!prealloc);
 		err = insert_state(tree, prealloc, start, end, &bits);
 		prealloc = NULL;
 		BUG_ON(err == -EEXIST);
@@ -773,20 +783,18 @@ hit_next:
 		if (err)
 			goto out;
 
+		next_node = rb_next(node);
 		cache_state(state, cached_state);
 		merge_state(tree, state);
 		if (last_end == (u64)-1)
 			goto out;
 
 		start = last_end + 1;
-		if (start < end && prealloc && !need_resched()) {
-			next_node = rb_next(node);
-			if (next_node) {
-				state = rb_entry(next_node, struct extent_state,
-						 rb_node);
-				if (state->start == start)
-					goto hit_next;
-			}
+		if (next_node && start < end && prealloc && !need_resched()) {
+			state = rb_entry(next_node, struct extent_state,
+					 rb_node);
+			if (state->start == start)
+				goto hit_next;
 		}
 		goto search_again;
 	}
@@ -813,6 +821,9 @@ hit_next:
 			err = -EEXIST;
 			goto out;
 		}
+
+		prealloc = alloc_extent_state_atomic(prealloc);
+		BUG_ON(!prealloc);
 		err = split_state(tree, state, prealloc, start);
 		BUG_ON(err == -EEXIST);
 		prealloc = NULL;
@@ -843,14 +854,25 @@ hit_next:
 			this_end = end;
 		else
 			this_end = last_start - 1;
+
+		prealloc = alloc_extent_state_atomic(prealloc);
+		BUG_ON(!prealloc);
+
+		/*
+		 * Avoid to free 'prealloc' if it can be merged with
+		 * the later extent.
+		 */
+		atomic_inc(&prealloc->refs);
 		err = insert_state(tree, prealloc, start, this_end,
 				   &bits);
 		BUG_ON(err == -EEXIST);
 		if (err) {
+			free_extent_state(prealloc);
 			prealloc = NULL;
 			goto out;
 		}
 		cache_state(prealloc, cached_state);
+		free_extent_state(prealloc);
 		prealloc = NULL;
 		start = this_end + 1;
 		goto search_again;
@@ -867,6 +889,9 @@ hit_next:
 			err = -EEXIST;
 			goto out;
 		}
+
+		prealloc = alloc_extent_state_atomic(prealloc);
+		BUG_ON(!prealloc);
 		err = split_state(tree, state, prealloc, end + 1);
 		BUG_ON(err == -EEXIST);
 
@@ -943,13 +968,6 @@ int set_extent_new(struct extent_io_tree *tree, u64 start, u64 end,
 			      NULL, mask);
 }
 
-static int clear_extent_new(struct extent_io_tree *tree, u64 start, u64 end,
-		       gfp_t mask)
-{
-	return clear_extent_bit(tree, start, end, EXTENT_NEW, 0, 0,
-				NULL, mask);
-}
-
 int set_extent_uptodate(struct extent_io_tree *tree, u64 start, u64 end,
 			struct extent_state **cached_state, gfp_t mask)
 {
@@ -965,11 +983,6 @@ static int clear_extent_uptodate(struct extent_io_tree *tree, u64 start,
 				cached_state, mask);
 }
 
-int wait_on_extent_writeback(struct extent_io_tree *tree, u64 start, u64 end)
-{
-	return wait_extent_bit(tree, start, end, EXTENT_WRITEBACK);
-}
-
 /*
  * either insert or lock state struct between start and end use mask to tell
  * us if waiting is desired.
@@ -1029,25 +1042,6 @@ int unlock_extent(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask)
 				mask);
 }
 
-/*
- * helper function to set pages and extents in the tree dirty
- */
-int set_range_dirty(struct extent_io_tree *tree, u64 start, u64 end)
-{
-	unsigned long index = start >> PAGE_CACHE_SHIFT;
-	unsigned long end_index = end >> PAGE_CACHE_SHIFT;
-	struct page *page;
-
-	while (index <= end_index) {
-		page = find_get_page(tree->mapping, index);
-		BUG_ON(!page);
-		__set_page_dirty_nobuffers(page);
-		page_cache_release(page);
-		index++;
-	}
-	return 0;
-}
-
 /*
  * helper function to set both pages and extents in the tree writeback
  */
@@ -1821,46 +1815,6 @@ static void end_bio_extent_readpage(struct bio *bio, int err)
 	bio_put(bio);
 }
 
-/*
- * IO done from prepare_write is pretty simple, we just unlock
- * the structs in the extent tree when done, and set the uptodate bits
- * as appropriate.
- */
-static void end_bio_extent_preparewrite(struct bio *bio, int err)
-{
-	const int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags);
-	struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
-	struct extent_io_tree *tree;
-	u64 start;
-	u64 end;
-
-	do {
-		struct page *page = bvec->bv_page;
-		struct extent_state *cached = NULL;
-		tree = &BTRFS_I(page->mapping->host)->io_tree;
-
-		start = ((u64)page->index << PAGE_CACHE_SHIFT) +
-			bvec->bv_offset;
-		end = start + bvec->bv_len - 1;
-
-		if (--bvec >= bio->bi_io_vec)
-			prefetchw(&bvec->bv_page->flags);
-
-		if (uptodate) {
-			set_extent_uptodate(tree, start, end, &cached,
-					    GFP_ATOMIC);
-		} else {
-			ClearPageUptodate(page);
-			SetPageError(page);
-		}
-
-		unlock_extent_cached(tree, start, end, &cached, GFP_ATOMIC);
-
-	} while (bvec >= bio->bi_io_vec);
-
-	bio_put(bio);
-}
-
 struct bio *
 btrfs_bio_alloc(struct block_device *bdev, u64 first_sector, int nr_vecs,
 		gfp_t gfp_flags)
@@ -2009,7 +1963,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 	struct btrfs_ordered_extent *ordered;
 	int ret;
 	int nr = 0;
-	size_t page_offset = 0;
+	size_t pg_offset = 0;
 	size_t iosize;
 	size_t disk_io_size;
 	size_t blocksize = inode->i_sb->s_blocksize;
@@ -2052,9 +2006,9 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 			char *userpage;
 			struct extent_state *cached = NULL;
 
-			iosize = PAGE_CACHE_SIZE - page_offset;
+			iosize = PAGE_CACHE_SIZE - pg_offset;
 			userpage = kmap_atomic(page, KM_USER0);
-			memset(userpage + page_offset, 0, iosize);
+			memset(userpage + pg_offset, 0, iosize);
 			flush_dcache_page(page);
 			kunmap_atomic(userpage, KM_USER0);
 			set_extent_uptodate(tree, cur, cur + iosize - 1,
@@ -2063,9 +2017,9 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 					     &cached, GFP_NOFS);
 			break;
 		}
-		em = get_extent(inode, page, page_offset, cur,
+		em = get_extent(inode, page, pg_offset, cur,
 				end - cur + 1, 0);
-		if (IS_ERR(em) || !em) {
+		if (IS_ERR_OR_NULL(em)) {
 			SetPageError(page);
 			unlock_extent(tree, cur, end, GFP_NOFS);
 			break;
@@ -2103,7 +2057,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 			struct extent_state *cached = NULL;
 
 			userpage = kmap_atomic(page, KM_USER0);
-			memset(userpage + page_offset, 0, iosize);
+			memset(userpage + pg_offset, 0, iosize);
 			flush_dcache_page(page);
 			kunmap_atomic(userpage, KM_USER0);
 
@@ -2112,7 +2066,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 			unlock_extent_cached(tree, cur, cur + iosize - 1,
 			                     &cached, GFP_NOFS);
 			cur = cur + iosize;
-			page_offset += iosize;
+			pg_offset += iosize;
 			continue;
 		}
 		/* the get_extent function already copied into the page */
@@ -2121,7 +2075,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 			check_page_uptodate(tree, page);
 			unlock_extent(tree, cur, cur + iosize - 1, GFP_NOFS);
 			cur = cur + iosize;
-			page_offset += iosize;
+			pg_offset += iosize;
 			continue;
 		}
 		/* we have an inline extent but it didn't get marked up
@@ -2131,7 +2085,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 			SetPageError(page);
 			unlock_extent(tree, cur, cur + iosize - 1, GFP_NOFS);
 			cur = cur + iosize;
-			page_offset += iosize;
+			pg_offset += iosize;
 			continue;
 		}
 
@@ -2144,7 +2098,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 			unsigned long pnr = (last_byte >> PAGE_CACHE_SHIFT) + 1;
 			pnr -= page->index;
 			ret = submit_extent_page(READ, tree, page,
-					 sector, disk_io_size, page_offset,
+					 sector, disk_io_size, pg_offset,
 					 bdev, bio, pnr,
 					 end_bio_extent_readpage, mirror_num,
 					 *bio_flags,
@@ -2155,7 +2109,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 		if (ret)
 			SetPageError(page);
 		cur = cur + iosize;
-		page_offset += iosize;
+		pg_offset += iosize;
 	}
 out:
 	if (!nr) {
@@ -2351,7 +2305,7 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc,
 		}
 		em = epd->get_extent(inode, page, pg_offset, cur,
 				     end - cur + 1, 1);
-		if (IS_ERR(em) || !em) {
+		if (IS_ERR_OR_NULL(em)) {
 			SetPageError(page);
 			break;
 		}
@@ -2729,128 +2683,6 @@ int extent_invalidatepage(struct extent_io_tree *tree,
 	return 0;
 }
 
-/*
- * simple commit_write call, set_range_dirty is used to mark both
- * the pages and the extent records as dirty
- */
-int extent_commit_write(struct extent_io_tree *tree,
-			struct inode *inode, struct page *page,
-			unsigned from, unsigned to)
-{
-	loff_t pos = ((loff_t)page->index << PAGE_CACHE_SHIFT) + to;
-
-	set_page_extent_mapped(page);
-	set_page_dirty(page);
-
-	if (pos > inode->i_size) {
-		i_size_write(inode, pos);
-		mark_inode_dirty(inode);
-	}
-	return 0;
-}
-
-int extent_prepare_write(struct extent_io_tree *tree,
-			 struct inode *inode, struct page *page,
-			 unsigned from, unsigned to, get_extent_t *get_extent)
-{
-	u64 page_start = (u64)page->index << PAGE_CACHE_SHIFT;
-	u64 page_end = page_start + PAGE_CACHE_SIZE - 1;
-	u64 block_start;
-	u64 orig_block_start;
-	u64 block_end;
-	u64 cur_end;
-	struct extent_map *em;
-	unsigned blocksize = 1 << inode->i_blkbits;
-	size_t page_offset = 0;
-	size_t block_off_start;
-	size_t block_off_end;
-	int err = 0;
-	int iocount = 0;
-	int ret = 0;
-	int isnew;
-
-	set_page_extent_mapped(page);
-
-	block_start = (page_start + from) & ~((u64)blocksize - 1);
-	block_end = (page_start + to - 1) | (blocksize - 1);
-	orig_block_start = block_start;
-
-	lock_extent(tree, page_start, page_end, GFP_NOFS);
-	while (block_start <= block_end) {
-		em = get_extent(inode, page, page_offset, block_start,
-				block_end - block_start + 1, 1);
-		if (IS_ERR(em) || !em)
-			goto err;
-
-		cur_end = min(block_end, extent_map_end(em) - 1);
-		block_off_start = block_start & (PAGE_CACHE_SIZE - 1);
-		block_off_end = block_off_start + blocksize;
-		isnew = clear_extent_new(tree, block_start, cur_end, GFP_NOFS);
-
-		if (!PageUptodate(page) && isnew &&
-		    (block_off_end > to || block_off_start < from)) {
-			void *kaddr;
-
-			kaddr = kmap_atomic(page, KM_USER0);
-			if (block_off_end > to)
-				memset(kaddr + to, 0, block_off_end - to);
-			if (block_off_start < from)
-				memset(kaddr + block_off_start, 0,
-				       from - block_off_start);
-			flush_dcache_page(page);
-			kunmap_atomic(kaddr, KM_USER0);
-		}
-		if ((em->block_start != EXTENT_MAP_HOLE &&
-		     em->block_start != EXTENT_MAP_INLINE) &&
-		    !isnew && !PageUptodate(page) &&
-		    (block_off_end > to || block_off_start < from) &&
-		    !test_range_bit(tree, block_start, cur_end,
-				    EXTENT_UPTODATE, 1, NULL)) {
-			u64 sector;
-			u64 extent_offset = block_start - em->start;
-			size_t iosize;
-			sector = (em->block_start + extent_offset) >> 9;
-			iosize = (cur_end - block_start + blocksize) &
-				~((u64)blocksize - 1);
-			/*
-			 * we've already got the extent locked, but we
-			 * need to split the state such that our end_bio
-			 * handler can clear the lock.
-			 */
-			set_extent_bit(tree, block_start,
-				       block_start + iosize - 1,
-				       EXTENT_LOCKED, 0, NULL, NULL, GFP_NOFS);
-			ret = submit_extent_page(READ, tree, page,
-					 sector, iosize, page_offset, em->bdev,
-					 NULL, 1,
-					 end_bio_extent_preparewrite, 0,
-					 0, 0);
-			if (ret && !err)
-				err = ret;
-			iocount++;
-			block_start = block_start + iosize;
-		} else {
-			struct extent_state *cached = NULL;
-
-			set_extent_uptodate(tree, block_start, cur_end, &cached,
-					    GFP_NOFS);
-			unlock_extent_cached(tree, block_start, cur_end,
-					     &cached, GFP_NOFS);
-			block_start = cur_end + 1;
-		}
-		page_offset = block_start & (PAGE_CACHE_SIZE - 1);
-		free_extent_map(em);
-	}
-	if (iocount) {
-		wait_extent_bit(tree, orig_block_start,
-				block_end, EXTENT_LOCKED);
-	}
-	check_page_uptodate(tree, page);
-err:
-	/* FIXME, zero out newly allocated blocks on error */
-	return err;
-}
-
 /*
  * a helper for releasepage, this tests for areas of the page that
  * are locked or under IO and drops the related state bits if it is safe
@@ -2909,7 +2741,7 @@ int try_release_extent_mapping(struct extent_map_tree *map,
 			len = end - start + 1;
 			write_lock(&map->lock);
 			em = lookup_extent_mapping(map, start, len);
-			if (!em || IS_ERR(em)) {
+			if (IS_ERR_OR_NULL(em)) {
 				write_unlock(&map->lock);
 				break;
 			}
@@ -2937,33 +2769,6 @@ int try_release_extent_mapping(struct extent_map_tree *map,
 	return try_release_extent_state(map, tree, page, mask);
 }
 
-sector_t extent_bmap(struct address_space *mapping, sector_t iblock,
-		get_extent_t *get_extent)
-{
-	struct inode *inode = mapping->host;
-	struct extent_state *cached_state = NULL;
-	u64 start = iblock << inode->i_blkbits;
-	sector_t sector = 0;
-	size_t blksize = (1 << inode->i_blkbits);
-	struct extent_map *em;
-
-	lock_extent_bits(&BTRFS_I(inode)->io_tree, start, start + blksize - 1,
-			 0, &cached_state, GFP_NOFS);
-	em = get_extent(inode, NULL, 0, start, blksize, 0);
-	unlock_extent_cached(&BTRFS_I(inode)->io_tree, start,
-			     start + blksize - 1, &cached_state, GFP_NOFS);
-	if (!em || IS_ERR(em))
-		return 0;
-
-	if (em->block_start > EXTENT_MAP_LAST_BYTE)
-		goto out;
-
-	sector = (em->block_start + start - em->start) >> inode->i_blkbits;
-out:
-	free_extent_map(em);
-	return sector;
-}
-
 /*
  * helper function for fiemap, which doesn't want to see any holes.
  * This maps until we find something past 'last'
@@ -2986,7 +2791,7 @@ static struct extent_map *get_extent_skip_holes(struct inode *inode,
 			break;
 		len = (len + sectorsize - 1) & ~(sectorsize - 1);
 		em = get_extent(inode, NULL, 0, offset, len, 0);
-		if (!em || IS_ERR(em))
+		if (IS_ERR_OR_NULL(em))
 			return em;
 
 		/* if this isn't a hole return it */
@@ -3040,7 +2845,7 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
 	 * because there might be preallocation past i_size
 	 */
 	ret = btrfs_lookup_file_extent(NULL, BTRFS_I(inode)->root,
-				       path, inode->i_ino, -1, 0);
+				       path, btrfs_ino(inode), -1, 0);
 	if (ret < 0) {
 		btrfs_free_path(path);
 		return ret;
@@ -3053,7 +2858,7 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
 	found_type = btrfs_key_type(&found_key);
 
 	/* No extents, but there might be delalloc bits */
-	if (found_key.objectid != inode->i_ino ||
+	if (found_key.objectid != btrfs_ino(inode) ||
 	    found_type != BTRFS_EXTENT_DATA_KEY) {
 		/* have to trust i_size as the end */
 		last = (u64)-1;
@@ -3276,8 +3081,7 @@ static inline void btrfs_release_extent_buffer(struct extent_buffer *eb)
 
 struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree,
 					  u64 start, unsigned long len,
-					  struct page *page0,
-					  gfp_t mask)
+					  struct page *page0)
 {
 	unsigned long num_pages = num_extent_pages(start, len);
 	unsigned long i;
@@ -3298,7 +3102,7 @@ struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree,
 	}
 	rcu_read_unlock();
 
-	eb = __alloc_extent_buffer(tree, start, len, mask);
+	eb = __alloc_extent_buffer(tree, start, len, GFP_NOFS);
 	if (!eb)
 		return NULL;
 
@@ -3315,7 +3119,7 @@ struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree,
 		i = 0;
 	}
 	for (; i < num_pages; i++, index++) {
-		p = find_or_create_page(mapping, index, mask | __GFP_HIGHMEM);
+		p = find_or_create_page(mapping, index, GFP_NOFS | __GFP_HIGHMEM);
 		if (!p) {
 			WARN_ON(1);
 			goto free_eb;
@@ -3387,8 +3191,7 @@ free_eb:
 }
 
 struct extent_buffer *find_extent_buffer(struct extent_io_tree *tree,
-					 u64 start, unsigned long len,
-					  gfp_t mask)
+					 u64 start, unsigned long len)
 {
 	struct extent_buffer *eb;
 
@@ -3449,13 +3252,6 @@ int clear_extent_buffer_dirty(struct extent_io_tree *tree,
 	return 0;
 }
 
-int wait_on_extent_buffer_writeback(struct extent_io_tree *tree,
-				    struct extent_buffer *eb)
-{
-	return wait_on_extent_writeback(tree, eb->start,
-					eb->start + eb->len - 1);
-}
-
 int set_extent_buffer_dirty(struct extent_io_tree *tree,
 			     struct extent_buffer *eb)
 {

fs/btrfs/extent_io.h

@@ -153,23 +153,14 @@ static inline int extent_compress_type(unsigned long bio_flags)
 
 struct extent_map_tree;
 
-static inline struct extent_state *extent_state_next(struct extent_state *state)
-{
-	struct rb_node *node;
-	node = rb_next(&state->rb_node);
-	if (!node)
-		return NULL;
-	return rb_entry(node, struct extent_state, rb_node);
-}
-
 typedef struct extent_map *(get_extent_t)(struct inode *inode,
 					  struct page *page,
-					  size_t page_offset,
+					  size_t pg_offset,
 					  u64 start, u64 len,
 					  int create);
 
 void extent_io_tree_init(struct extent_io_tree *tree,
-			  struct address_space *mapping, gfp_t mask);
+			 struct address_space *mapping);
 int try_release_extent_mapping(struct extent_map_tree *map,
 			       struct extent_io_tree *tree, struct page *page,
 			       gfp_t mask);
@@ -215,14 +206,8 @@ int set_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end,
 		     gfp_t mask);
 int clear_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end,
 		       gfp_t mask);
-int clear_extent_ordered(struct extent_io_tree *tree, u64 start, u64 end,
-		       gfp_t mask);
-int clear_extent_ordered_metadata(struct extent_io_tree *tree, u64 start,
-				  u64 end, gfp_t mask);
 int set_extent_delalloc(struct extent_io_tree *tree, u64 start, u64 end,
 			struct extent_state **cached_state, gfp_t mask);
-int set_extent_ordered(struct extent_io_tree *tree, u64 start, u64 end,
-		     gfp_t mask);
 int find_first_extent_bit(struct extent_io_tree *tree, u64 start,
 			  u64 *start_ret, u64 *end_ret, int bits);
 struct extent_state *find_first_extent_bit_state(struct extent_io_tree *tree,
@@ -243,28 +228,17 @@ int extent_readpages(struct extent_io_tree *tree,
 		     struct address_space *mapping,
 		     struct list_head *pages, unsigned nr_pages,
 		     get_extent_t get_extent);
-int extent_prepare_write(struct extent_io_tree *tree,
-			 struct inode *inode, struct page *page,
-			 unsigned from, unsigned to, get_extent_t *get_extent);
-int extent_commit_write(struct extent_io_tree *tree,
-			struct inode *inode, struct page *page,
-			unsigned from, unsigned to);
-sector_t extent_bmap(struct address_space *mapping, sector_t iblock,
-		get_extent_t *get_extent);
 int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
 		__u64 start, __u64 len, get_extent_t *get_extent);
-int set_range_dirty(struct extent_io_tree *tree, u64 start, u64 end);
 int set_state_private(struct extent_io_tree *tree, u64 start, u64 private);
 int get_state_private(struct extent_io_tree *tree, u64 start, u64 *private);
 void set_page_extent_mapped(struct page *page);
 
 struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree,
 					  u64 start, unsigned long len,
-					  struct page *page0,
-					  gfp_t mask);
+					  struct page *page0);
 struct extent_buffer *find_extent_buffer(struct extent_io_tree *tree,
-					 u64 start, unsigned long len,
-					  gfp_t mask);
+					 u64 start, unsigned long len);
 void free_extent_buffer(struct extent_buffer *eb);
 int read_extent_buffer_pages(struct extent_io_tree *tree,
 			     struct extent_buffer *eb, u64 start, int wait,
@@ -292,16 +266,11 @@ void memmove_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset,
 			   unsigned long src_offset, unsigned long len);
 void memset_extent_buffer(struct extent_buffer *eb, char c,
 			  unsigned long start, unsigned long len);
-int wait_on_extent_buffer_writeback(struct extent_io_tree *tree,
-				    struct extent_buffer *eb);
-int wait_on_extent_writeback(struct extent_io_tree *tree, u64 start, u64 end);
 int wait_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, int bits);
 int clear_extent_buffer_dirty(struct extent_io_tree *tree,
 			      struct extent_buffer *eb);
 int set_extent_buffer_dirty(struct extent_io_tree *tree,
 			     struct extent_buffer *eb);
-int test_extent_buffer_dirty(struct extent_io_tree *tree,
-			     struct extent_buffer *eb);
 int set_extent_buffer_uptodate(struct extent_io_tree *tree,
 			       struct extent_buffer *eb);
 int clear_extent_buffer_uptodate(struct extent_io_tree *tree,
@@ -319,7 +288,6 @@ int map_private_extent_buffer(struct extent_buffer *eb, unsigned long offset,
 		      unsigned long *map_start,
 		      unsigned long *map_len, int km);
 void unmap_extent_buffer(struct extent_buffer *eb, char *token, int km);
-int release_extent_buffer_tail_pages(struct extent_buffer *eb);
 int extent_range_uptodate(struct extent_io_tree *tree,
 			  u64 start, u64 end);
 int extent_clear_unlock_delalloc(struct inode *inode,

fs/btrfs/extent_map.c

@@ -28,12 +28,11 @@ void extent_map_exit(void)
 /**
  * extent_map_tree_init - initialize extent map tree
  * @tree:		tree to initialize
- * @mask:		flags for memory allocations during tree operations
  *
  * Initialize the extent tree @tree.  Should be called for each new inode
  * or other user of the extent_map interface.
  */
-void extent_map_tree_init(struct extent_map_tree *tree, gfp_t mask)
+void extent_map_tree_init(struct extent_map_tree *tree)
 {
 	tree->map = RB_ROOT;
 	rwlock_init(&tree->lock);
@@ -41,16 +40,15 @@ void extent_map_tree_init(struct extent_map_tree *tree, gfp_t mask)
 
 /**
  * alloc_extent_map - allocate new extent map structure
- * @mask:	memory allocation flags
  *
  * Allocate a new extent_map structure.  The new structure is
  * returned with a reference count of one and needs to be
  * freed using free_extent_map()
  */
-struct extent_map *alloc_extent_map(gfp_t mask)
+struct extent_map *alloc_extent_map(void)
 {
 	struct extent_map *em;
-	em = kmem_cache_alloc(extent_map_cache, mask);
+	em = kmem_cache_alloc(extent_map_cache, GFP_NOFS);
 	if (!em)
 		return NULL;
 	em->in_tree = 0;

fs/btrfs/extent_map.h

@@ -49,14 +49,14 @@ static inline u64 extent_map_block_end(struct extent_map *em)
 	return em->block_start + em->block_len;
 }
 
-void extent_map_tree_init(struct extent_map_tree *tree, gfp_t mask);
+void extent_map_tree_init(struct extent_map_tree *tree);
 struct extent_map *lookup_extent_mapping(struct extent_map_tree *tree,
 					 u64 start, u64 len);
 int add_extent_mapping(struct extent_map_tree *tree,
 		       struct extent_map *em);
 int remove_extent_mapping(struct extent_map_tree *tree, struct extent_map *em);
 
-struct extent_map *alloc_extent_map(gfp_t mask);
+struct extent_map *alloc_extent_map(void);
 void free_extent_map(struct extent_map *em);
 int __init extent_map_init(void);
 void extent_map_exit(void);

fs/btrfs/file-item.c

@@ -193,7 +193,7 @@ static int __btrfs_lookup_bio_sums(struct btrfs_root *root,
 			u32 item_size;
 
 			if (item)
-				btrfs_release_path(root, path);
+				btrfs_release_path(path);
 			item = btrfs_lookup_csum(NULL, root->fs_info->csum_root,
 						 path, disk_bytenr, 0);
 			if (IS_ERR(item)) {
@@ -208,12 +208,13 @@ static int __btrfs_lookup_bio_sums(struct btrfs_root *root,
 						EXTENT_NODATASUM, GFP_NOFS);
 				} else {
 					printk(KERN_INFO "btrfs no csum found "
-					       "for inode %lu start %llu\n",
-					       inode->i_ino,
+					       "for inode %llu start %llu\n",
+					       (unsigned long long)
+					       btrfs_ino(inode),
 					       (unsigned long long)offset);
 				}
 				item = NULL;
-				btrfs_release_path(root, path);
+				btrfs_release_path(path);
 				goto found;
 			}
 			btrfs_item_key_to_cpu(path->nodes[0], &found_key,
@@ -266,7 +267,7 @@ int btrfs_lookup_bio_sums_dio(struct btrfs_root *root, struct inode *inode,
 }
 
 int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, u64 end,
-			     struct list_head *list)
+			     struct list_head *list, int search_commit)
 {
 	struct btrfs_key key;
 	struct btrfs_path *path;
@@ -283,6 +284,12 @@ int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, u64 end,
 	path = btrfs_alloc_path();
 	BUG_ON(!path);
 
+	if (search_commit) {
+		path->skip_locking = 1;
+		path->reada = 2;
+		path->search_commit_root = 1;
+	}
+
 	key.objectid = BTRFS_EXTENT_CSUM_OBJECTID;
 	key.offset = start;
 	key.type = BTRFS_EXTENT_CSUM_KEY;
@@ -495,7 +502,6 @@ static noinline int truncate_one_csum(struct btrfs_trans_handle *trans,
 		u32 new_size = (bytenr - key->offset) >> blocksize_bits;
 		new_size *= csum_size;
 		ret = btrfs_truncate_item(trans, root, path, new_size, 1);
-		BUG_ON(ret);
 	} else if (key->offset >= bytenr && csum_end > end_byte &&
 		   end_byte > key->offset) {
 		/*
@@ -508,7 +514,6 @@ static noinline int truncate_one_csum(struct btrfs_trans_handle *trans,
 		new_size *= csum_size;
 
 		ret = btrfs_truncate_item(trans, root, path, new_size, 0);
-		BUG_ON(ret);
 
 		key->offset = end_byte;
 		ret = btrfs_set_item_key_safe(trans, root, path, key);
@@ -551,10 +556,10 @@ int btrfs_del_csums(struct btrfs_trans_handle *trans,
 		ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
 		if (ret > 0) {
 			if (path->slots[0] == 0)
-				goto out;
+				break;
 			path->slots[0]--;
 		} else if (ret < 0) {
-			goto out;
+			break;
 		}
 
 		leaf = path->nodes[0];
@@ -579,7 +584,8 @@ int btrfs_del_csums(struct btrfs_trans_handle *trans,
 		/* delete the entire item, it is inside our range */
 		if (key.offset >= bytenr && csum_end <= end_byte) {
 			ret = btrfs_del_item(trans, root, path);
-			BUG_ON(ret);
+			if (ret)
+				goto out;
 			if (key.offset == bytenr)
 				break;
 		} else if (key.offset < bytenr && csum_end > end_byte) {
@@ -631,11 +637,12 @@ int btrfs_del_csums(struct btrfs_trans_handle *trans,
 			if (key.offset < bytenr)
 				break;
 		}
-		btrfs_release_path(root, path);
+		btrfs_release_path(path);
 	}
+	ret = 0;
 out:
 	btrfs_free_path(path);
-	return 0;
+	return ret;
 }
 
 int btrfs_csum_file_blocks(struct btrfs_trans_handle *trans,
@@ -722,7 +729,7 @@ again:
 	 * at this point, we know the tree has an item, but it isn't big
 	 * enough yet to put our csum in.  Grow it
 	 */
-	btrfs_release_path(root, path);
+	btrfs_release_path(path);
 	ret = btrfs_search_slot(trans, root, &file_key, path,
 				csum_size, 1);
 	if (ret < 0)
@@ -761,12 +768,11 @@ again:
 			goto insert;
 
 		ret = btrfs_extend_item(trans, root, path, diff);
-		BUG_ON(ret);
 		goto csum;
 	}
 
 insert:
-	btrfs_release_path(root, path);
+	btrfs_release_path(path);
 	csum_offset = 0;
 	if (found_next) {
 		u64 tmp = total_bytes + root->sectorsize;
@@ -850,7 +856,7 @@ next_sector:
 	}
 	btrfs_mark_buffer_dirty(path->nodes[0]);
 	if (total_bytes < sums->len) {
-		btrfs_release_path(root, path);
+		btrfs_release_path(path);
 		cond_resched();
 		goto again;
 	}

fs/btrfs/file.c

@@ -40,6 +40,263 @@
 #include "locking.h"
 #include "compat.h"
 
+/*
+ * when auto defrag is enabled we
+ * queue up these defrag structs to remember which
+ * inodes need defragging passes
+ */
+struct inode_defrag {
+	struct rb_node rb_node;
+	/* objectid */
+	u64 ino;
+	/*
+	 * transid where the defrag was added, we search for
+	 * extents newer than this
+	 */
+	u64 transid;
+
+	/* root objectid */
+	u64 root;
+
+	/* last offset we were able to defrag */
+	u64 last_offset;
+
+	/* if we've wrapped around back to zero once already */
+	int cycled;
+};
+
+/* pop a record for an inode into the defrag tree.  The lock
+ * must be held already
+ *
+ * If you're inserting a record for an older transid than an
+ * existing record, the transid already in the tree is lowered
+ *
+ * If an existing record is found the defrag item you
+ * pass in is freed
+ */
+static int __btrfs_add_inode_defrag(struct inode *inode,
+				    struct inode_defrag *defrag)
+{
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct inode_defrag *entry;
+	struct rb_node **p;
+	struct rb_node *parent = NULL;
+
+	p = &root->fs_info->defrag_inodes.rb_node;
+	while (*p) {
+		parent = *p;
+		entry = rb_entry(parent, struct inode_defrag, rb_node);
+
+		if (defrag->ino < entry->ino)
+			p = &parent->rb_left;
+		else if (defrag->ino > entry->ino)
+			p = &parent->rb_right;
+		else {
+			/* if we're reinserting an entry for
+			 * an old defrag run, make sure to
+			 * lower the transid of our existing record
+			 */
+			if (defrag->transid < entry->transid)
+				entry->transid = defrag->transid;
+			if (defrag->last_offset > entry->last_offset)
+				entry->last_offset = defrag->last_offset;
+			goto exists;
+		}
+	}
+	BTRFS_I(inode)->in_defrag = 1;
+	rb_link_node(&defrag->rb_node, parent, p);
+	rb_insert_color(&defrag->rb_node, &root->fs_info->defrag_inodes);
+	return 0;
+
+exists:
+	kfree(defrag);
+	return 0;
+
+}
+
+/*
+ * insert a defrag record for this inode if auto defrag is
+ * enabled
+ */
+int btrfs_add_inode_defrag(struct btrfs_trans_handle *trans,
+			   struct inode *inode)
+{
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct inode_defrag *defrag;
+	int ret = 0;
+	u64 transid;
+
+	if (!btrfs_test_opt(root, AUTO_DEFRAG))
+		return 0;
+
+	if (root->fs_info->closing)
+		return 0;
+
+	if (BTRFS_I(inode)->in_defrag)
+		return 0;
+
+	if (trans)
+		transid = trans->transid;
+	else
+		transid = BTRFS_I(inode)->root->last_trans;
+
+	defrag = kzalloc(sizeof(*defrag), GFP_NOFS);
+	if (!defrag)
+		return -ENOMEM;
+
+	defrag->ino = inode->i_ino;
+	defrag->transid = transid;
+	defrag->root = root->root_key.objectid;
+
+	spin_lock(&root->fs_info->defrag_inodes_lock);
+	if (!BTRFS_I(inode)->in_defrag)
+		ret = __btrfs_add_inode_defrag(inode, defrag);
+	spin_unlock(&root->fs_info->defrag_inodes_lock);
+	return ret;
+}
+
+/*
+ * must be called with the defrag_inodes lock held
+ */
+struct inode_defrag *btrfs_find_defrag_inode(struct btrfs_fs_info *info, u64 ino,
+					     struct rb_node **next)
+{
+	struct inode_defrag *entry = NULL;
+	struct rb_node *p;
+	struct rb_node *parent = NULL;
+
+	p = info->defrag_inodes.rb_node;
+	while (p) {
+		parent = p;
+		entry = rb_entry(parent, struct inode_defrag, rb_node);
+
+		if (ino < entry->ino)
+			p = parent->rb_left;
+		else if (ino > entry->ino)
+			p = parent->rb_right;
+		else
+			return entry;
+	}
+
+	if (next) {
+		while (parent && ino > entry->ino) {
+			parent = rb_next(parent);
+			entry = rb_entry(parent, struct inode_defrag, rb_node);
+		}
+		*next = parent;
+	}
+	return NULL;
+}
+
+/*
+ * run through the list of inodes in the FS that need
+ * defragging
+ */
+int btrfs_run_defrag_inodes(struct btrfs_fs_info *fs_info)
+{
+	struct inode_defrag *defrag;
+	struct btrfs_root *inode_root;
+	struct inode *inode;
+	struct rb_node *n;
+	struct btrfs_key key;
+	struct btrfs_ioctl_defrag_range_args range;
+	u64 first_ino = 0;
+	int num_defrag;
+	int defrag_batch = 1024;
+
+	memset(&range, 0, sizeof(range));
+	range.len = (u64)-1;
+
+	atomic_inc(&fs_info->defrag_running);
+	spin_lock(&fs_info->defrag_inodes_lock);
+	while(1) {
+		n = NULL;
+
+		/* find an inode to defrag */
+		defrag = btrfs_find_defrag_inode(fs_info, first_ino, &n);
+		if (!defrag) {
+			if (n)
+				defrag = rb_entry(n, struct inode_defrag, rb_node);
+			else if (first_ino) {
+				first_ino = 0;
+				continue;
+			} else {
+				break;
+			}
+		}
+
+		/* remove it from the rbtree */
+		first_ino = defrag->ino + 1;
+		rb_erase(&defrag->rb_node, &fs_info->defrag_inodes);
+
+		if (fs_info->closing)
+			goto next_free;
+
+		spin_unlock(&fs_info->defrag_inodes_lock);
+
+		/* get the inode */
+		key.objectid = defrag->root;
+		btrfs_set_key_type(&key, BTRFS_ROOT_ITEM_KEY);
+		key.offset = (u64)-1;
+		inode_root = btrfs_read_fs_root_no_name(fs_info, &key);
+		if (IS_ERR(inode_root))
+			goto next;
+
+		key.objectid = defrag->ino;
+		btrfs_set_key_type(&key, BTRFS_INODE_ITEM_KEY);
+		key.offset = 0;
+
+		inode = btrfs_iget(fs_info->sb, &key, inode_root, NULL);
+		if (IS_ERR(inode))
+			goto next;
+
+		/* do a chunk of defrag */
+		BTRFS_I(inode)->in_defrag = 0;
+		range.start = defrag->last_offset;
+		num_defrag = btrfs_defrag_file(inode, NULL, &range, defrag->transid,
+					       defrag_batch);
+		/*
+		 * if we filled the whole defrag batch, there
+		 * must be more work to do.  Queue this defrag
+		 * again
+		 */
+		if (num_defrag == defrag_batch) {
+			defrag->last_offset = range.start;
+			__btrfs_add_inode_defrag(inode, defrag);
+			/*
+			 * we don't want to kfree defrag, we added it back to
+			 * the rbtree
+			 */
+			defrag = NULL;
+		} else if (defrag->last_offset && !defrag->cycled) {
+			/*
+			 * we didn't fill our defrag batch, but
+			 * we didn't start at zero.  Make sure we loop
+			 * around to the start of the file.
+			 */
+			defrag->last_offset = 0;
+			defrag->cycled = 1;
+			__btrfs_add_inode_defrag(inode, defrag);
+			defrag = NULL;
+		}
+
+		iput(inode);
+next:
+		spin_lock(&fs_info->defrag_inodes_lock);
+next_free:
+		kfree(defrag);
+	}
+	spin_unlock(&fs_info->defrag_inodes_lock);
+
+	atomic_dec(&fs_info->defrag_running);
+
+	/*
+	 * during unmount, we use the transaction_wait queue to
+	 * wait for the defragger to stop
+	 */
+	wake_up(&fs_info->transaction_wait);
+	return 0;
+}
 
 /* simple helper to fault in pages and copy.  This should go away
  * and be replaced with calls into generic code.
@@ -191,9 +448,9 @@ int btrfs_drop_extent_cache(struct inode *inode, u64 start, u64 end,
 	}
 	while (1) {
 		if (!split)
-			split = alloc_extent_map(GFP_NOFS);
+			split = alloc_extent_map();
 		if (!split2)
-			split2 = alloc_extent_map(GFP_NOFS);
+			split2 = alloc_extent_map();
 		BUG_ON(!split || !split2);
 
 		write_lock(&em_tree->lock);
@@ -298,6 +555,7 @@ int btrfs_drop_extents(struct btrfs_trans_handle *trans, struct inode *inode,
 	struct btrfs_path *path;
 	struct btrfs_key key;
 	struct btrfs_key new_key;
+	u64 ino = btrfs_ino(inode);
 	u64 search_start = start;
 	u64 disk_bytenr = 0;
 	u64 num_bytes = 0;
@@ -318,14 +576,14 @@ int btrfs_drop_extents(struct btrfs_trans_handle *trans, struct inode *inode,
 
 	while (1) {
 		recow = 0;
-		ret = btrfs_lookup_file_extent(trans, root, path, inode->i_ino,
+		ret = btrfs_lookup_file_extent(trans, root, path, ino,
 					       search_start, -1);
 		if (ret < 0)
 			break;
 		if (ret > 0 && path->slots[0] > 0 && search_start == start) {
 			leaf = path->nodes[0];
 			btrfs_item_key_to_cpu(leaf, &key, path->slots[0] - 1);
-			if (key.objectid == inode->i_ino &&
+			if (key.objectid == ino &&
 			    key.type == BTRFS_EXTENT_DATA_KEY)
 				path->slots[0]--;
 		}
@@ -346,7 +604,7 @@ next_slot:
 		}
 
 		btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
-		if (key.objectid > inode->i_ino ||
+		if (key.objectid > ino ||
 		    key.type > BTRFS_EXTENT_DATA_KEY || key.offset >= end)
 			break;
 
@@ -376,7 +634,7 @@ next_slot:
 
 		search_start = max(key.offset, start);
 		if (recow) {
-			btrfs_release_path(root, path);
+			btrfs_release_path(path);
 			continue;
 		}
 
@@ -393,7 +651,7 @@ next_slot:
 			ret = btrfs_duplicate_item(trans, root, path,
 						   &new_key);
 			if (ret == -EAGAIN) {
-				btrfs_release_path(root, path);
+				btrfs_release_path(path);
 				continue;
 			}
 			if (ret < 0)
@@ -516,7 +774,7 @@ next_slot:
 			del_nr = 0;
 			del_slot = 0;
 
-			btrfs_release_path(root, path);
+			btrfs_release_path(path);
 			continue;
 		}
 
@@ -592,6 +850,7 @@ int btrfs_mark_extent_written(struct btrfs_trans_handle *trans,
 	int del_slot = 0;
 	int recow;
 	int ret;
+	u64 ino = btrfs_ino(inode);
 
 	btrfs_drop_extent_cache(inode, start, end - 1, 0);
 
@@ -600,7 +859,7 @@ int btrfs_mark_extent_written(struct btrfs_trans_handle *trans,
 again:
 	recow = 0;
 	split = start;
-	key.objectid = inode->i_ino;
+	key.objectid = ino;
 	key.type = BTRFS_EXTENT_DATA_KEY;
 	key.offset = split;
 
@@ -612,8 +871,7 @@ again:
 
 	leaf = path->nodes[0];
 	btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
-	BUG_ON(key.objectid != inode->i_ino ||
-	       key.type != BTRFS_EXTENT_DATA_KEY);
+	BUG_ON(key.objectid != ino || key.type != BTRFS_EXTENT_DATA_KEY);
 	fi = btrfs_item_ptr(leaf, path->slots[0],
 			    struct btrfs_file_extent_item);
 	BUG_ON(btrfs_file_extent_type(leaf, fi) !=
@@ -630,7 +888,7 @@ again:
 		other_start = 0;
 		other_end = start;
 		if (extent_mergeable(leaf, path->slots[0] - 1,
-				     inode->i_ino, bytenr, orig_offset,
+				     ino, bytenr, orig_offset,
 				     &other_start, &other_end)) {
 			new_key.offset = end;
 			btrfs_set_item_key_safe(trans, root, path, &new_key);
@@ -653,7 +911,7 @@ again:
 		other_start = end;
 		other_end = 0;
 		if (extent_mergeable(leaf, path->slots[0] + 1,
-				     inode->i_ino, bytenr, orig_offset,
+				     ino, bytenr, orig_offset,
 				     &other_start, &other_end)) {
 			fi = btrfs_item_ptr(leaf, path->slots[0],
 					    struct btrfs_file_extent_item);
@@ -681,7 +939,7 @@ again:
 		new_key.offset = split;
 		ret = btrfs_duplicate_item(trans, root, path, &new_key);
 		if (ret == -EAGAIN) {
-			btrfs_release_path(root, path);
+			btrfs_release_path(path);
 			goto again;
 		}
 		BUG_ON(ret < 0);
@@ -702,7 +960,7 @@ again:
 
 		ret = btrfs_inc_extent_ref(trans, root, bytenr, num_bytes, 0,
 					   root->root_key.objectid,
-					   inode->i_ino, orig_offset);
+					   ino, orig_offset);
 		BUG_ON(ret);
 
 		if (split == start) {
@@ -718,10 +976,10 @@ again:
 	other_start = end;
 	other_end = 0;
 	if (extent_mergeable(leaf, path->slots[0] + 1,
-			     inode->i_ino, bytenr, orig_offset,
+			     ino, bytenr, orig_offset,
 			     &other_start, &other_end)) {
 		if (recow) {
-			btrfs_release_path(root, path);
+			btrfs_release_path(path);
 			goto again;
 		}
 		extent_end = other_end;
@@ -729,16 +987,16 @@ again:
 		del_nr++;
 		ret = btrfs_free_extent(trans, root, bytenr, num_bytes,
 					0, root->root_key.objectid,
-					inode->i_ino, orig_offset);
+					ino, orig_offset);
 		BUG_ON(ret);
 	}
 	other_start = 0;
 	other_end = start;
 	if (extent_mergeable(leaf, path->slots[0] - 1,
-			     inode->i_ino, bytenr, orig_offset,
+			     ino, bytenr, orig_offset,
 			     &other_start, &other_end)) {
 		if (recow) {
-			btrfs_release_path(root, path);
+			btrfs_release_path(path);
 			goto again;
 		}
 		key.offset = other_start;
@@ -746,7 +1004,7 @@ again:
 		del_nr++;
 		ret = btrfs_free_extent(trans, root, bytenr, num_bytes,
 					0, root->root_key.objectid,
-					inode->i_ino, orig_offset);
+					ino, orig_offset);
 		BUG_ON(ret);
 	}
 	if (del_nr == 0) {
@@ -1375,7 +1633,7 @@ static long btrfs_fallocate(struct file *file, int mode,
 	while (1) {
 		em = btrfs_get_extent(inode, NULL, 0, cur_offset,
 				      alloc_end - cur_offset, 0);
-		BUG_ON(IS_ERR(em) || !em);
+		BUG_ON(IS_ERR_OR_NULL(em));
 		last_byte = min(extent_map_end(em), alloc_end);
 		last_byte = (last_byte + mask) & ~mask;
 		if (em->block_start == EXTENT_MAP_HOLE ||

fs/btrfs/free-space-cache.h

@@ -27,6 +27,25 @@ struct btrfs_free_space {
 	struct list_head list;
 };
 
+struct btrfs_free_space_ctl {
+	spinlock_t tree_lock;
+	struct rb_root free_space_offset;
+	u64 free_space;
+	int extents_thresh;
+	int free_extents;
+	int total_bitmaps;
+	int unit;
+	u64 start;
+	struct btrfs_free_space_op *op;
+	void *private;
+};
+
+struct btrfs_free_space_op {
+	void (*recalc_thresholds)(struct btrfs_free_space_ctl *ctl);
+	bool (*use_bitmap)(struct btrfs_free_space_ctl *ctl,
+			   struct btrfs_free_space *info);
+};
+
 struct inode *lookup_free_space_inode(struct btrfs_root *root,
 				      struct btrfs_block_group_cache
 				      *block_group, struct btrfs_path *path);
@@ -45,17 +64,38 @@ int btrfs_write_out_cache(struct btrfs_root *root,
 			  struct btrfs_trans_handle *trans,
 			  struct btrfs_block_group_cache *block_group,
 			  struct btrfs_path *path);
-int btrfs_add_free_space(struct btrfs_block_group_cache *block_group,
-			 u64 bytenr, u64 size);
+
+struct inode *lookup_free_ino_inode(struct btrfs_root *root,
+				    struct btrfs_path *path);
+int create_free_ino_inode(struct btrfs_root *root,
+			  struct btrfs_trans_handle *trans,
+			  struct btrfs_path *path);
+int load_free_ino_cache(struct btrfs_fs_info *fs_info,
+			struct btrfs_root *root);
+int btrfs_write_out_ino_cache(struct btrfs_root *root,
+			      struct btrfs_trans_handle *trans,
+			      struct btrfs_path *path);
+
+void btrfs_init_free_space_ctl(struct btrfs_block_group_cache *block_group);
+int __btrfs_add_free_space(struct btrfs_free_space_ctl *ctl,
+			   u64 bytenr, u64 size);
+static inline int
+btrfs_add_free_space(struct btrfs_block_group_cache *block_group,
+		     u64 bytenr, u64 size)
+{
+	return __btrfs_add_free_space(block_group->free_space_ctl,
+				      bytenr, size);
+}
 int btrfs_remove_free_space(struct btrfs_block_group_cache *block_group,
 			    u64 bytenr, u64 size);
+void __btrfs_remove_free_space_cache(struct btrfs_free_space_ctl *ctl);
 void btrfs_remove_free_space_cache(struct btrfs_block_group_cache
-				   *block_group);
+				     *block_group);
 u64 btrfs_find_space_for_alloc(struct btrfs_block_group_cache *block_group,
 			       u64 offset, u64 bytes, u64 empty_size);
+u64 btrfs_find_ino_for_alloc(struct btrfs_root *fs_root);
 void btrfs_dump_free_space(struct btrfs_block_group_cache *block_group,
 			   u64 bytes);
-u64 btrfs_block_group_free_space(struct btrfs_block_group_cache *block_group);
 int btrfs_find_space_cluster(struct btrfs_trans_handle *trans,
 			     struct btrfs_root *root,
 			     struct btrfs_block_group_cache *block_group,

fs/btrfs/inode-item.c

@@ -130,7 +130,6 @@ int btrfs_del_inode_ref(struct btrfs_trans_handle *trans,
 			      item_size - (ptr + sub_item_len - item_start));
 	ret = btrfs_truncate_item(trans, root, path,
 				  item_size - sub_item_len, 1);
-	BUG_ON(ret);
 out:
 	btrfs_free_path(path);
 	return ret;
@@ -167,7 +166,6 @@ int btrfs_insert_inode_ref(struct btrfs_trans_handle *trans,
 
 		old_size = btrfs_item_size_nr(path->nodes[0], path->slots[0]);
 		ret = btrfs_extend_item(trans, root, path, ins_len);
-		BUG_ON(ret);
 		ref = btrfs_item_ptr(path->nodes[0], path->slots[0],
 				     struct btrfs_inode_ref);
 		ref = (struct btrfs_inode_ref *)((unsigned long)ref + old_size);

fs/btrfs/inode-map.c

@@ -16,11 +16,446 @@
  * Boston, MA 021110-1307, USA.
  */
 
+#include <linux/delay.h>
+#include <linux/kthread.h>
+#include <linux/pagemap.h>
+
 #include "ctree.h"
 #include "disk-io.h"
+#include "free-space-cache.h"
+#include "inode-map.h"
 #include "transaction.h"
 
-int btrfs_find_highest_inode(struct btrfs_root *root, u64 *objectid)
+static int caching_kthread(void *data)
+{
+	struct btrfs_root *root = data;
+	struct btrfs_fs_info *fs_info = root->fs_info;
+	struct btrfs_free_space_ctl *ctl = root->free_ino_ctl;
+	struct btrfs_key key;
+	struct btrfs_path *path;
+	struct extent_buffer *leaf;
+	u64 last = (u64)-1;
+	int slot;
+	int ret;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	/* Since the commit root is read-only, we can safely skip locking. */
+	path->skip_locking = 1;
+	path->search_commit_root = 1;
+	path->reada = 2;
+
+	key.objectid = BTRFS_FIRST_FREE_OBJECTID;
+	key.offset = 0;
+	key.type = BTRFS_INODE_ITEM_KEY;
+again:
+	/* need to make sure the commit_root doesn't disappear */
+	mutex_lock(&root->fs_commit_mutex);
+
+	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+	if (ret < 0)
+		goto out;
+
+	while (1) {
+		smp_mb();
+		if (fs_info->closing)
+			goto out;
+
+		leaf = path->nodes[0];
+		slot = path->slots[0];
+		if (slot >= btrfs_header_nritems(leaf)) {
+			ret = btrfs_next_leaf(root, path);
+			if (ret < 0)
+				goto out;
+			else if (ret > 0)
+				break;
+
+			if (need_resched() ||
+			    btrfs_transaction_in_commit(fs_info)) {
+				leaf = path->nodes[0];
+
+				if (btrfs_header_nritems(leaf) == 0) {
+					WARN_ON(1);
+					break;
+				}
+
+				/*
+				 * Save the key so we can advances forward
+				 * in the next search.
+				 */
+				btrfs_item_key_to_cpu(leaf, &key, 0);
+				btrfs_release_path(path);
+				root->cache_progress = last;
+				mutex_unlock(&root->fs_commit_mutex);
+				schedule_timeout(1);
+				goto again;
+			} else
+				continue;
+		}
+
+		btrfs_item_key_to_cpu(leaf, &key, slot);
+
+		if (key.type != BTRFS_INODE_ITEM_KEY)
+			goto next;
+
+		if (key.objectid >= root->highest_objectid)
+			break;
+
+		if (last != (u64)-1 && last + 1 != key.objectid) {
+			__btrfs_add_free_space(ctl, last + 1,
+					       key.objectid - last - 1);
+			wake_up(&root->cache_wait);
+		}
+
+		last = key.objectid;
+next:
+		path->slots[0]++;
+	}
+
+	if (last < root->highest_objectid - 1) {
+		__btrfs_add_free_space(ctl, last + 1,
+				       root->highest_objectid - last - 1);
+	}
+
+	spin_lock(&root->cache_lock);
+	root->cached = BTRFS_CACHE_FINISHED;
+	spin_unlock(&root->cache_lock);
+
+	root->cache_progress = (u64)-1;
+	btrfs_unpin_free_ino(root);
+out:
+	wake_up(&root->cache_wait);
+	mutex_unlock(&root->fs_commit_mutex);
+
+	btrfs_free_path(path);
+
+	return ret;
+}
+
+static void start_caching(struct btrfs_root *root)
+{
+	struct btrfs_free_space_ctl *ctl = root->free_ino_ctl;
+	struct task_struct *tsk;
+	int ret;
+	u64 objectid;
+
+	spin_lock(&root->cache_lock);
+	if (root->cached != BTRFS_CACHE_NO) {
+		spin_unlock(&root->cache_lock);
+		return;
+	}
+
+	root->cached = BTRFS_CACHE_STARTED;
+	spin_unlock(&root->cache_lock);
+
+	ret = load_free_ino_cache(root->fs_info, root);
+	if (ret == 1) {
+		spin_lock(&root->cache_lock);
+		root->cached = BTRFS_CACHE_FINISHED;
+		spin_unlock(&root->cache_lock);
+		return;
+	}
+
+	/*
+	 * It can be quite time-consuming to fill the cache by searching
+	 * through the extent tree, and this can keep ino allocation path
+	 * waiting. Therefore at start we quickly find out the highest
+	 * inode number and we know we can use inode numbers which fall in
+	 * [highest_ino + 1, BTRFS_LAST_FREE_OBJECTID].
+	 */
+	ret = btrfs_find_free_objectid(root, &objectid);
+	if (!ret && objectid <= BTRFS_LAST_FREE_OBJECTID) {
+		__btrfs_add_free_space(ctl, objectid,
+				       BTRFS_LAST_FREE_OBJECTID - objectid + 1);
+	}
+
+	tsk = kthread_run(caching_kthread, root, "btrfs-ino-cache-%llu\n",
+			  root->root_key.objectid);
+	BUG_ON(IS_ERR(tsk));
+}
+
+int btrfs_find_free_ino(struct btrfs_root *root, u64 *objectid)
+{
+again:
+	*objectid = btrfs_find_ino_for_alloc(root);
+
+	if (*objectid != 0)
+		return 0;
+
+	start_caching(root);
+
+	wait_event(root->cache_wait,
+		   root->cached == BTRFS_CACHE_FINISHED ||
+		   root->free_ino_ctl->free_space > 0);
+
+	if (root->cached == BTRFS_CACHE_FINISHED &&
+	    root->free_ino_ctl->free_space == 0)
+		return -ENOSPC;
+	else
+		goto again;
+}
+
+void btrfs_return_ino(struct btrfs_root *root, u64 objectid)
+{
+	struct btrfs_free_space_ctl *ctl = root->free_ino_ctl;
+	struct btrfs_free_space_ctl *pinned = root->free_ino_pinned;
+again:
+	if (root->cached == BTRFS_CACHE_FINISHED) {
+		__btrfs_add_free_space(ctl, objectid, 1);
+	} else {
+		/*
+		 * If we are in the process of caching free ino chunks,
+		 * to avoid adding the same inode number to the free_ino
+		 * tree twice due to cross transaction, we'll leave it
+		 * in the pinned tree until a transaction is committed
+		 * or the caching work is done.
+		 */
+
+		mutex_lock(&root->fs_commit_mutex);
+		spin_lock(&root->cache_lock);
+		if (root->cached == BTRFS_CACHE_FINISHED) {
+			spin_unlock(&root->cache_lock);
+			mutex_unlock(&root->fs_commit_mutex);
+			goto again;
+		}
+		spin_unlock(&root->cache_lock);
+
+		start_caching(root);
+
+		if (objectid <= root->cache_progress ||
+		    objectid > root->highest_objectid)
+			__btrfs_add_free_space(ctl, objectid, 1);
+		else
+			__btrfs_add_free_space(pinned, objectid, 1);
+
+		mutex_unlock(&root->fs_commit_mutex);
+	}
+}
+
+/*
+ * When a transaction is committed, we'll move those inode numbers which
+ * are smaller than root->cache_progress from pinned tree to free_ino tree,
+ * and others will just be dropped, because the commit root we were
+ * searching has changed.
+ *
+ * Must be called with root->fs_commit_mutex held
+ */
+void btrfs_unpin_free_ino(struct btrfs_root *root)
+{
+	struct btrfs_free_space_ctl *ctl = root->free_ino_ctl;
+	struct rb_root *rbroot = &root->free_ino_pinned->free_space_offset;
+	struct btrfs_free_space *info;
+	struct rb_node *n;
+	u64 count;
+
+	while (1) {
+		n = rb_first(rbroot);
+		if (!n)
+			break;
+
+		info = rb_entry(n, struct btrfs_free_space, offset_index);
+		BUG_ON(info->bitmap);
+
+		if (info->offset > root->cache_progress)
+			goto free;
+		else if (info->offset + info->bytes > root->cache_progress)
+			count = root->cache_progress - info->offset + 1;
+		else
+			count = info->bytes;
+
+		__btrfs_add_free_space(ctl, info->offset, count);
+free:
+		rb_erase(&info->offset_index, rbroot);
+		kfree(info);
+	}
+}
+
+#define INIT_THRESHOLD	(((1024 * 32) / 2) / sizeof(struct btrfs_free_space))
+#define INODES_PER_BITMAP (PAGE_CACHE_SIZE * 8)
+
+/*
+ * The goal is to keep the memory used by the free_ino tree won't
+ * exceed the memory if we use bitmaps only.
+ */
+static void recalculate_thresholds(struct btrfs_free_space_ctl *ctl)
+{
+	struct btrfs_free_space *info;
+	struct rb_node *n;
+	int max_ino;
+	int max_bitmaps;
+
+	n = rb_last(&ctl->free_space_offset);
+	if (!n) {
+		ctl->extents_thresh = INIT_THRESHOLD;
+		return;
+	}
+	info = rb_entry(n, struct btrfs_free_space, offset_index);
+
+	/*
+	 * Find the maximum inode number in the filesystem. Note we
+	 * ignore the fact that this can be a bitmap, because we are
+	 * not doing precise calculation.
+	 */
+	max_ino = info->bytes - 1;
+
+	max_bitmaps = ALIGN(max_ino, INODES_PER_BITMAP) / INODES_PER_BITMAP;
+	if (max_bitmaps <= ctl->total_bitmaps) {
+		ctl->extents_thresh = 0;
+		return;
+	}
+
+	ctl->extents_thresh = (max_bitmaps - ctl->total_bitmaps) *
+				PAGE_CACHE_SIZE / sizeof(*info);
+}
+
+/*
+ * We don't fall back to bitmap, if we are below the extents threshold
+ * or this chunk of inode numbers is a big one.
+ */
+static bool use_bitmap(struct btrfs_free_space_ctl *ctl,
+		       struct btrfs_free_space *info)
+{
+	if (ctl->free_extents < ctl->extents_thresh ||
+	    info->bytes > INODES_PER_BITMAP / 10)
+		return false;
+
+	return true;
+}
+
+static struct btrfs_free_space_op free_ino_op = {
+	.recalc_thresholds	= recalculate_thresholds,
+	.use_bitmap		= use_bitmap,
+};
+
+static void pinned_recalc_thresholds(struct btrfs_free_space_ctl *ctl)
+{
+}
+
+static bool pinned_use_bitmap(struct btrfs_free_space_ctl *ctl,
+			      struct btrfs_free_space *info)
+{
+	/*
+	 * We always use extents for two reasons:
+	 *
+	 * - The pinned tree is only used during the process of caching
+	 *   work.
+	 * - Make code simpler. See btrfs_unpin_free_ino().
+	 */
+	return false;
+}
+
+static struct btrfs_free_space_op pinned_free_ino_op = {
+	.recalc_thresholds	= pinned_recalc_thresholds,
+	.use_bitmap		= pinned_use_bitmap,
+};
+
+void btrfs_init_free_ino_ctl(struct btrfs_root *root)
+{
+	struct btrfs_free_space_ctl *ctl = root->free_ino_ctl;
+	struct btrfs_free_space_ctl *pinned = root->free_ino_pinned;
+
+	spin_lock_init(&ctl->tree_lock);
+	ctl->unit = 1;
+	ctl->start = 0;
+	ctl->private = NULL;
+	ctl->op = &free_ino_op;
+
+	/*
+	 * Initially we allow to use 16K of ram to cache chunks of
+	 * inode numbers before we resort to bitmaps. This is somewhat
+	 * arbitrary, but it will be adjusted in runtime.
+	 */
+	ctl->extents_thresh = INIT_THRESHOLD;
+
+	spin_lock_init(&pinned->tree_lock);
+	pinned->unit = 1;
+	pinned->start = 0;
+	pinned->private = NULL;
+	pinned->extents_thresh = 0;
+	pinned->op = &pinned_free_ino_op;
+}
+
+int btrfs_save_ino_cache(struct btrfs_root *root,
+			 struct btrfs_trans_handle *trans)
+{
+	struct btrfs_free_space_ctl *ctl = root->free_ino_ctl;
+	struct btrfs_path *path;
+	struct inode *inode;
+	u64 alloc_hint = 0;
+	int ret;
+	int prealloc;
+	bool retry = false;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+again:
+	inode = lookup_free_ino_inode(root, path);
+	if (IS_ERR(inode) && PTR_ERR(inode) != -ENOENT) {
+		ret = PTR_ERR(inode);
+		goto out;
+	}
+
+	if (IS_ERR(inode)) {
+		BUG_ON(retry);
+		retry = true;
+
+		ret = create_free_ino_inode(root, trans, path);
+		if (ret)
+			goto out;
+		goto again;
+	}
+
+	BTRFS_I(inode)->generation = 0;
+	ret = btrfs_update_inode(trans, root, inode);
+	WARN_ON(ret);
+
+	if (i_size_read(inode) > 0) {
+		ret = btrfs_truncate_free_space_cache(root, trans, path, inode);
+		if (ret)
+			goto out_put;
+	}
+
+	spin_lock(&root->cache_lock);
+	if (root->cached != BTRFS_CACHE_FINISHED) {
+		ret = -1;
+		spin_unlock(&root->cache_lock);
+		goto out_put;
+	}
+	spin_unlock(&root->cache_lock);
+
+	spin_lock(&ctl->tree_lock);
+	prealloc = sizeof(struct btrfs_free_space) * ctl->free_extents;
+	prealloc = ALIGN(prealloc, PAGE_CACHE_SIZE);
+	prealloc += ctl->total_bitmaps * PAGE_CACHE_SIZE;
+	spin_unlock(&ctl->tree_lock);
+
+	/* Just to make sure we have enough space */
+	prealloc += 8 * PAGE_CACHE_SIZE;
+
+	ret = btrfs_check_data_free_space(inode, prealloc);
+	if (ret)
+		goto out_put;
+
+	ret = btrfs_prealloc_file_range_trans(inode, trans, 0, 0, prealloc,
+					      prealloc, prealloc, &alloc_hint);
+	if (ret)
+		goto out_put;
+	btrfs_free_reserved_data_space(inode, prealloc);
+
+out_put:
+	iput(inode);
+out:
+	if (ret == 0)
+		ret = btrfs_write_out_ino_cache(root, trans, path);
+
+	btrfs_free_path(path);
+	return ret;
+}
+
+static int btrfs_find_highest_objectid(struct btrfs_root *root, u64 *objectid)
 {
 	struct btrfs_path *path;
 	int ret;
@@ -55,15 +490,14 @@ error:
 	return ret;
 }
 
-int btrfs_find_free_objectid(struct btrfs_trans_handle *trans,
-			     struct btrfs_root *root,
-			     u64 dirid, u64 *objectid)
+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_inode(root, &root->highest_objectid);
+		ret = btrfs_find_highest_objectid(root,
+						  &root->highest_objectid);
 		if (ret)
 			goto out;
 	}

fs/btrfs/inode-map.h

@@ -0,0 +1,13 @@
+#ifndef __BTRFS_INODE_MAP
+#define __BTRFS_INODE_MAP
+
+void btrfs_init_free_ino_ctl(struct btrfs_root *root);
+void btrfs_unpin_free_ino(struct btrfs_root *root);
+void btrfs_return_ino(struct btrfs_root *root, u64 objectid);
+int btrfs_find_free_ino(struct btrfs_root *root, u64 *objectid);
+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);
+
+#endif

fs/btrfs/ioctl.c

@@ -50,6 +50,7 @@
 #include "print-tree.h"
 #include "volumes.h"
 #include "locking.h"
+#include "inode-map.h"
 
 /* Mask out flags that are inappropriate for the given type of inode. */
 static inline __u32 btrfs_mask_flags(umode_t mode, __u32 flags)
@@ -281,8 +282,9 @@ static noinline int btrfs_ioctl_fitrim(struct file *file, void __user *arg)
 	if (!capable(CAP_SYS_ADMIN))
 		return -EPERM;
 
-	mutex_lock(&fs_info->fs_devices->device_list_mutex);
-	list_for_each_entry(device, &fs_info->fs_devices->devices, dev_list) {
+	rcu_read_lock();
+	list_for_each_entry_rcu(device, &fs_info->fs_devices->devices,
+				dev_list) {
 		if (!device->bdev)
 			continue;
 		q = bdev_get_queue(device->bdev);
@@ -292,7 +294,7 @@ static noinline int btrfs_ioctl_fitrim(struct file *file, void __user *arg)
 				     minlen);
 		}
 	}
-	mutex_unlock(&fs_info->fs_devices->device_list_mutex);
+	rcu_read_unlock();
 	if (!num_devices)
 		return -EOPNOTSUPP;
 
@@ -329,8 +331,7 @@ static noinline int create_subvol(struct btrfs_root *root,
 	u64 new_dirid = BTRFS_FIRST_FREE_OBJECTID;
 	u64 index = 0;
 
-	ret = btrfs_find_free_objectid(NULL, root->fs_info->tree_root,
-				       0, &objectid);
+	ret = btrfs_find_free_objectid(root->fs_info->tree_root, &objectid);
 	if (ret) {
 		dput(parent);
 		return ret;
@@ -422,7 +423,7 @@ static noinline int create_subvol(struct btrfs_root *root,
 	BUG_ON(ret);
 
 	ret = btrfs_insert_dir_item(trans, root,
-				    name, namelen, dir->i_ino, &key,
+				    name, namelen, dir, &key,
 				    BTRFS_FT_DIR, index);
 	if (ret)
 		goto fail;
@@ -433,7 +434,7 @@ static noinline int create_subvol(struct btrfs_root *root,
 
 	ret = btrfs_add_root_ref(trans, root->fs_info->tree_root,
 				 objectid, root->root_key.objectid,
-				 dir->i_ino, index, name, namelen);
+				 btrfs_ino(dir), index, name, namelen);
 
 	BUG_ON(ret);
 
@@ -655,6 +656,106 @@ out_unlock:
 	return error;
 }
 
+/*
+ * When we're defragging a range, we don't want to kick it off again
+ * if it is really just waiting for delalloc to send it down.
+ * If we find a nice big extent or delalloc range for the bytes in the
+ * file you want to defrag, we return 0 to let you know to skip this
+ * part of the file
+ */
+static int check_defrag_in_cache(struct inode *inode, u64 offset, int thresh)
+{
+	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
+	struct extent_map *em = NULL;
+	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
+	u64 end;
+
+	read_lock(&em_tree->lock);
+	em = lookup_extent_mapping(em_tree, offset, PAGE_CACHE_SIZE);
+	read_unlock(&em_tree->lock);
+
+	if (em) {
+		end = extent_map_end(em);
+		free_extent_map(em);
+		if (end - offset > thresh)
+			return 0;
+	}
+	/* if we already have a nice delalloc here, just stop */
+	thresh /= 2;
+	end = count_range_bits(io_tree, &offset, offset + thresh,
+			       thresh, EXTENT_DELALLOC, 1);
+	if (end >= thresh)
+		return 0;
+	return 1;
+}
+
+/*
+ * helper function to walk through a file and find extents
+ * newer than a specific transid, and smaller than thresh.
+ *
+ * This is used by the defragging code to find new and small
+ * extents
+ */
+static int find_new_extents(struct btrfs_root *root,
+			    struct inode *inode, u64 newer_than,
+			    u64 *off, int thresh)
+{
+	struct btrfs_path *path;
+	struct btrfs_key min_key;
+	struct btrfs_key max_key;
+	struct extent_buffer *leaf;
+	struct btrfs_file_extent_item *extent;
+	int type;
+	int ret;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	min_key.objectid = inode->i_ino;
+	min_key.type = BTRFS_EXTENT_DATA_KEY;
+	min_key.offset = *off;
+
+	max_key.objectid = inode->i_ino;
+	max_key.type = (u8)-1;
+	max_key.offset = (u64)-1;
+
+	path->keep_locks = 1;
+
+	while(1) {
+		ret = btrfs_search_forward(root, &min_key, &max_key,
+					   path, 0, newer_than);
+		if (ret != 0)
+			goto none;
+		if (min_key.objectid != inode->i_ino)
+			goto none;
+		if (min_key.type != BTRFS_EXTENT_DATA_KEY)
+			goto none;
+
+		leaf = path->nodes[0];
+		extent = btrfs_item_ptr(leaf, path->slots[0],
+					struct btrfs_file_extent_item);
+
+		type = btrfs_file_extent_type(leaf, extent);
+		if (type == BTRFS_FILE_EXTENT_REG &&
+		    btrfs_file_extent_num_bytes(leaf, extent) < thresh &&
+		    check_defrag_in_cache(inode, min_key.offset, thresh)) {
+			*off = min_key.offset;
+			btrfs_free_path(path);
+			return 0;
+		}
+
+		if (min_key.offset == (u64)-1)
+			goto none;
+
+		min_key.offset++;
+		btrfs_release_path(path);
+	}
+none:
+	btrfs_free_path(path);
+	return -ENOENT;
+}
+
 static int should_defrag_range(struct inode *inode, u64 start, u64 len,
 			       int thresh, u64 *last_len, u64 *skip,
 			       u64 *defrag_end)
@@ -664,10 +765,6 @@ static int should_defrag_range(struct inode *inode, u64 start, u64 len,
 	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
 	int ret = 1;
 
-
-	if (thresh == 0)
-		thresh = 256 * 1024;
-
 	/*
 	 * make sure that once we start defragging and extent, we keep on
 	 * defragging it
@@ -726,27 +823,176 @@ static int should_defrag_range(struct inode *inode, u64 start, u64 len,
 	return ret;
 }
 
-static int btrfs_defrag_file(struct file *file,
-			     struct btrfs_ioctl_defrag_range_args *range)
+/*
+ * it doesn't do much good to defrag one or two pages
+ * at a time.  This pulls in a nice chunk of pages
+ * to COW and defrag.
+ *
+ * It also makes sure the delalloc code has enough
+ * dirty data to avoid making new small extents as part
+ * of the defrag
+ *
+ * It's a good idea to start RA on this range
+ * before calling this.
+ */
+static int cluster_pages_for_defrag(struct inode *inode,
+				    struct page **pages,
+				    unsigned long start_index,
+				    int num_pages)
 {
-	struct inode *inode = fdentry(file)->d_inode;
-	struct btrfs_root *root = BTRFS_I(inode)->root;
-	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
+	unsigned long file_end;
+	u64 isize = i_size_read(inode);
+	u64 page_start;
+	u64 page_end;
+	int ret;
+	int i;
+	int i_done;
 	struct btrfs_ordered_extent *ordered;
-	struct page *page;
+	struct extent_state *cached_state = NULL;
+
+	if (isize == 0)
+		return 0;
+	file_end = (isize - 1) >> PAGE_CACHE_SHIFT;
+
+	ret = btrfs_delalloc_reserve_space(inode,
+					   num_pages << PAGE_CACHE_SHIFT);
+	if (ret)
+		return ret;
+again:
+	ret = 0;
+	i_done = 0;
+
+	/* step one, lock all the pages */
+	for (i = 0; i < num_pages; i++) {
+		struct page *page;
+		page = grab_cache_page(inode->i_mapping,
+					    start_index + i);
+		if (!page)
+			break;
+
+		if (!PageUptodate(page)) {
+			btrfs_readpage(NULL, page);
+			lock_page(page);
+			if (!PageUptodate(page)) {
+				unlock_page(page);
+				page_cache_release(page);
+				ret = -EIO;
+				break;
+			}
+		}
+		isize = i_size_read(inode);
+		file_end = (isize - 1) >> PAGE_CACHE_SHIFT;
+		if (!isize || page->index > file_end ||
+		    page->mapping != inode->i_mapping) {
+			/* whoops, we blew past eof, skip this page */
+			unlock_page(page);
+			page_cache_release(page);
+			break;
+		}
+		pages[i] = page;
+		i_done++;
+	}
+	if (!i_done || ret)
+		goto out;
+
+	if (!(inode->i_sb->s_flags & MS_ACTIVE))
+		goto out;
+
+	/*
+	 * so now we have a nice long stream of locked
+	 * and up to date pages, lets wait on them
+	 */
+	for (i = 0; i < i_done; i++)
+		wait_on_page_writeback(pages[i]);
+
+	page_start = page_offset(pages[0]);
+	page_end = page_offset(pages[i_done - 1]) + PAGE_CACHE_SIZE;
+
+	lock_extent_bits(&BTRFS_I(inode)->io_tree,
+			 page_start, page_end - 1, 0, &cached_state,
+			 GFP_NOFS);
+	ordered = btrfs_lookup_first_ordered_extent(inode, page_end - 1);
+	if (ordered &&
+	    ordered->file_offset + ordered->len > page_start &&
+	    ordered->file_offset < page_end) {
+		btrfs_put_ordered_extent(ordered);
+		unlock_extent_cached(&BTRFS_I(inode)->io_tree,
+				     page_start, page_end - 1,
+				     &cached_state, GFP_NOFS);
+		for (i = 0; i < i_done; i++) {
+			unlock_page(pages[i]);
+			page_cache_release(pages[i]);
+		}
+		btrfs_wait_ordered_range(inode, page_start,
+					 page_end - page_start);
+		goto again;
+	}
+	if (ordered)
+		btrfs_put_ordered_extent(ordered);
+
+	clear_extent_bit(&BTRFS_I(inode)->io_tree, page_start,
+			  page_end - 1, EXTENT_DIRTY | EXTENT_DELALLOC |
+			  EXTENT_DO_ACCOUNTING, 0, 0, &cached_state,
+			  GFP_NOFS);
+
+	if (i_done != num_pages) {
+		atomic_inc(&BTRFS_I(inode)->outstanding_extents);
+		btrfs_delalloc_release_space(inode,
+				     (num_pages - i_done) << PAGE_CACHE_SHIFT);
+	}
+
+
+	btrfs_set_extent_delalloc(inode, page_start, page_end - 1,
+				  &cached_state);
+
+	unlock_extent_cached(&BTRFS_I(inode)->io_tree,
+			     page_start, page_end - 1, &cached_state,
+			     GFP_NOFS);
+
+	for (i = 0; i < i_done; i++) {
+		clear_page_dirty_for_io(pages[i]);
+		ClearPageChecked(pages[i]);
+		set_page_extent_mapped(pages[i]);
+		set_page_dirty(pages[i]);
+		unlock_page(pages[i]);
+		page_cache_release(pages[i]);
+	}
+	return i_done;
+out:
+	for (i = 0; i < i_done; i++) {
+		unlock_page(pages[i]);
+		page_cache_release(pages[i]);
+	}
+	btrfs_delalloc_release_space(inode, num_pages << PAGE_CACHE_SHIFT);
+	return ret;
+
+}
+
+int btrfs_defrag_file(struct inode *inode, struct file *file,
+		      struct btrfs_ioctl_defrag_range_args *range,
+		      u64 newer_than, unsigned long max_to_defrag)
+{
+	struct btrfs_root *root = BTRFS_I(inode)->root;
 	struct btrfs_super_block *disk_super;
+	struct file_ra_state *ra = NULL;
 	unsigned long last_index;
-	unsigned long ra_pages = root->fs_info->bdi.ra_pages;
-	unsigned long total_read = 0;
 	u64 features;
-	u64 page_start;
-	u64 page_end;
 	u64 last_len = 0;
 	u64 skip = 0;
 	u64 defrag_end = 0;
+	u64 newer_off = range->start;
+	int newer_left = 0;
 	unsigned long i;
 	int ret;
+	int defrag_count = 0;
 	int compress_type = BTRFS_COMPRESS_ZLIB;
+	int extent_thresh = range->extent_thresh;
+	int newer_cluster = (256 * 1024) >> PAGE_CACHE_SHIFT;
+	u64 new_align = ~((u64)128 * 1024 - 1);
+	struct page **pages = NULL;
+
+	if (extent_thresh == 0)
+		extent_thresh = 256 * 1024;
 
 	if (range->flags & BTRFS_DEFRAG_RANGE_COMPRESS) {
 		if (range->compress_type > BTRFS_COMPRESS_TYPES)
@@ -758,6 +1004,27 @@ static int btrfs_defrag_file(struct file *file,
 	if (inode->i_size == 0)
 		return 0;
 
+	/*
+	 * if we were not given a file, allocate a readahead
+	 * context
+	 */
+	if (!file) {
+		ra = kzalloc(sizeof(*ra), GFP_NOFS);
+		if (!ra)
+			return -ENOMEM;
+		file_ra_state_init(ra, inode->i_mapping);
+	} else {
+		ra = &file->f_ra;
+	}
+
+	pages = kmalloc(sizeof(struct page *) * newer_cluster,
+			GFP_NOFS);
+	if (!pages) {
+		ret = -ENOMEM;
+		goto out_ra;
+	}
+
+	/* find the last page to defrag */
 	if (range->start + range->len > range->start) {
 		last_index = min_t(u64, inode->i_size - 1,
 			 range->start + range->len - 1) >> PAGE_CACHE_SHIFT;
@@ -765,11 +1032,37 @@ static int btrfs_defrag_file(struct file *file,
 		last_index = (inode->i_size - 1) >> PAGE_CACHE_SHIFT;
 	}
 
-	i = range->start >> PAGE_CACHE_SHIFT;
-	while (i <= last_index) {
-		if (!should_defrag_range(inode, (u64)i << PAGE_CACHE_SHIFT,
+	if (newer_than) {
+		ret = find_new_extents(root, inode, newer_than,
+				       &newer_off, 64 * 1024);
+		if (!ret) {
+			range->start = newer_off;
+			/*
+			 * we always align our defrag to help keep
+			 * the extents in the file evenly spaced
+			 */
+			i = (newer_off & new_align) >> PAGE_CACHE_SHIFT;
+			newer_left = newer_cluster;
+		} else
+			goto out_ra;
+	} else {
+		i = range->start >> PAGE_CACHE_SHIFT;
+	}
+	if (!max_to_defrag)
+		max_to_defrag = last_index - 1;
+
+	while (i <= last_index && defrag_count < max_to_defrag) {
+		/*
+		 * make sure we stop running if someone unmounts
+		 * the FS
+		 */
+		if (!(inode->i_sb->s_flags & MS_ACTIVE))
+			break;
+
+		if (!newer_than &&
+		    !should_defrag_range(inode, (u64)i << PAGE_CACHE_SHIFT,
 					PAGE_CACHE_SIZE,
-					range->extent_thresh,
+					extent_thresh,
 					&last_len, &skip,
 					&defrag_end)) {
 			unsigned long next;
@@ -781,92 +1074,39 @@ static int btrfs_defrag_file(struct file *file,
 			i = max(i + 1, next);
 			continue;
 		}
-
-		if (total_read % ra_pages == 0) {
-			btrfs_force_ra(inode->i_mapping, &file->f_ra, file, i,
-				       min(last_index, i + ra_pages - 1));
-		}
-		total_read++;
-		mutex_lock(&inode->i_mutex);
 		if (range->flags & BTRFS_DEFRAG_RANGE_COMPRESS)
 			BTRFS_I(inode)->force_compress = compress_type;
 
-		ret  = btrfs_delalloc_reserve_space(inode, PAGE_CACHE_SIZE);
-		if (ret)
-			goto err_unlock;
-again:
-		if (inode->i_size == 0 ||
-		    i > ((inode->i_size - 1) >> PAGE_CACHE_SHIFT)) {
-			ret = 0;
-			goto err_reservations;
-		}
+		btrfs_force_ra(inode->i_mapping, ra, file, i, newer_cluster);
 
-		page = grab_cache_page(inode->i_mapping, i);
-		if (!page) {
-			ret = -ENOMEM;
-			goto err_reservations;
-		}
-
-		if (!PageUptodate(page)) {
-			btrfs_readpage(NULL, page);
-			lock_page(page);
-			if (!PageUptodate(page)) {
-				unlock_page(page);
-				page_cache_release(page);
-				ret = -EIO;
-				goto err_reservations;
-			}
-		}
-
-		if (page->mapping != inode->i_mapping) {
-			unlock_page(page);
-			page_cache_release(page);
-			goto again;
-		}
+		ret = cluster_pages_for_defrag(inode, pages, i, newer_cluster);
+		if (ret < 0)
+			goto out_ra;
 
-		wait_on_page_writeback(page);
+		defrag_count += ret;
+		balance_dirty_pages_ratelimited_nr(inode->i_mapping, ret);
+		i += ret;
 
-		if (PageDirty(page)) {
-			btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE);
-			goto loop_unlock;
-		}
-
-		page_start = (u64)page->index << PAGE_CACHE_SHIFT;
-		page_end = page_start + PAGE_CACHE_SIZE - 1;
-		lock_extent(io_tree, page_start, page_end, GFP_NOFS);
+		if (newer_than) {
+			if (newer_off == (u64)-1)
+				break;
 
-		ordered = btrfs_lookup_ordered_extent(inode, page_start);
-		if (ordered) {
-			unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
-			unlock_page(page);
-			page_cache_release(page);
-			btrfs_start_ordered_extent(inode, ordered, 1);
-			btrfs_put_ordered_extent(ordered);
-			goto again;
+			newer_off = max(newer_off + 1,
+					(u64)i << PAGE_CACHE_SHIFT);
+
+			ret = find_new_extents(root, inode,
+					       newer_than, &newer_off,
+					       64 * 1024);
+			if (!ret) {
+				range->start = newer_off;
+				i = (newer_off & new_align) >> PAGE_CACHE_SHIFT;
+				newer_left = newer_cluster;
+			} else {
+				break;
+			}
+		} else {
+			i++;
 		}
-		set_page_extent_mapped(page);
-
-		/*
-		 * this makes sure page_mkwrite is called on the
-		 * page if it is dirtied again later
-		 */
-		clear_page_dirty_for_io(page);
-		clear_extent_bits(&BTRFS_I(inode)->io_tree, page_start,
-				  page_end, EXTENT_DIRTY | EXTENT_DELALLOC |
-				  EXTENT_DO_ACCOUNTING, GFP_NOFS);
-
-		btrfs_set_extent_delalloc(inode, page_start, page_end, NULL);
-		ClearPageChecked(page);
-		set_page_dirty(page);
-		unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
-
-loop_unlock:
-		unlock_page(page);
-		page_cache_release(page);
-		mutex_unlock(&inode->i_mutex);
-
-		balance_dirty_pages_ratelimited_nr(inode->i_mapping, 1);
-		i++;
 	}
 
 	if ((range->flags & BTRFS_DEFRAG_RANGE_START_IO))
@@ -898,12 +1138,14 @@ loop_unlock:
 		btrfs_set_super_incompat_flags(disk_super, features);
 	}
 
-	return 0;
+	if (!file)
+		kfree(ra);
+	return defrag_count;
 
-err_reservations:
-	btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE);
-err_unlock:
-	mutex_unlock(&inode->i_mutex);
+out_ra:
+	if (!file)
+		kfree(ra);
+	kfree(pages);
 	return ret;
 }
 
@@ -1129,7 +1371,7 @@ static noinline int btrfs_ioctl_subvol_getflags(struct file *file,
 	int ret = 0;
 	u64 flags = 0;
 
-	if (inode->i_ino != BTRFS_FIRST_FREE_OBJECTID)
+	if (btrfs_ino(inode) != BTRFS_FIRST_FREE_OBJECTID)
 		return -EINVAL;
 
 	down_read(&root->fs_info->subvol_sem);
@@ -1156,7 +1398,7 @@ static noinline int btrfs_ioctl_subvol_setflags(struct file *file,
 	if (root->fs_info->sb->s_flags & MS_RDONLY)
 		return -EROFS;
 
-	if (inode->i_ino != BTRFS_FIRST_FREE_OBJECTID)
+	if (btrfs_ino(inode) != BTRFS_FIRST_FREE_OBJECTID)
 		return -EINVAL;
 
 	if (copy_from_user(&flags, arg, sizeof(flags)))
@@ -1279,7 +1521,6 @@ static noinline int copy_to_sk(struct btrfs_root *root,
 	int nritems;
 	int i;
 	int slot;
-	int found = 0;
 	int ret = 0;
 
 	leaf = path->nodes[0];
@@ -1326,7 +1567,7 @@ static noinline int copy_to_sk(struct btrfs_root *root,
 					   item_off, item_len);
 			*sk_offset += item_len;
 		}
-		found++;
+		(*num_found)++;
 
 		if (*num_found >= sk->nr_items)
 			break;
@@ -1345,7 +1586,6 @@ advance_key:
 	} else
 		ret = 1;
 overflow:
-	*num_found += found;
 	return ret;
 }
 
@@ -1402,7 +1642,7 @@ static noinline int search_ioctl(struct inode *inode,
 		}
 		ret = copy_to_sk(root, path, &key, sk, args->buf,
 				 &sk_offset, &num_found);
-		btrfs_release_path(root, path);
+		btrfs_release_path(path);
 		if (ret || num_found >= sk->nr_items)
 			break;
 
@@ -1509,7 +1749,7 @@ static noinline int btrfs_search_path_in_tree(struct btrfs_fs_info *info,
 		if (key.offset == BTRFS_FIRST_FREE_OBJECTID)
 			break;
 
-		btrfs_release_path(root, path);
+		btrfs_release_path(path);
 		key.objectid = key.offset;
 		key.offset = (u64)-1;
 		dirid = key.objectid;
@@ -1639,7 +1879,7 @@ static noinline int btrfs_ioctl_snap_destroy(struct file *file,
 			goto out_dput;
 	}
 
-	if (inode->i_ino != BTRFS_FIRST_FREE_OBJECTID) {
+	if (btrfs_ino(inode) != BTRFS_FIRST_FREE_OBJECTID) {
 		err = -EINVAL;
 		goto out_dput;
 	}
@@ -1757,7 +1997,10 @@ static int btrfs_ioctl_defrag(struct file *file, void __user *argp)
 			/* the rest are all set to zero by kzalloc */
 			range->len = (u64)-1;
 		}
-		ret = btrfs_defrag_file(file, range);
+		ret = btrfs_defrag_file(fdentry(file)->d_inode, file,
+					range, 0, 0);
+		if (ret > 0)
+			ret = 0;
 		kfree(range);
 		break;
 	default:
@@ -1809,6 +2052,75 @@ static long btrfs_ioctl_rm_dev(struct btrfs_root *root, void __user *arg)
 	return ret;
 }
 
+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;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+
+	fi_args.num_devices = fs_devices->num_devices;
+	fi_args.max_id = 0;
+	memcpy(&fi_args.fsid, root->fs_info->fsid, sizeof(fi_args.fsid));
+
+	mutex_lock(&fs_devices->device_list_mutex);
+	list_for_each_entry_safe(device, next, &fs_devices->devices, dev_list) {
+		if (device->devid > fi_args.max_id)
+			fi_args.max_id = device->devid;
+	}
+	mutex_unlock(&fs_devices->device_list_mutex);
+
+	if (copy_to_user(arg, &fi_args, sizeof(fi_args)))
+		return -EFAULT;
+
+	return 0;
+}
+
+static long btrfs_ioctl_dev_info(struct btrfs_root *root, void __user *arg)
+{
+	struct btrfs_ioctl_dev_info_args *di_args;
+	struct btrfs_device *dev;
+	struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices;
+	int ret = 0;
+	char *s_uuid = NULL;
+	char empty_uuid[BTRFS_UUID_SIZE] = {0};
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+
+	di_args = memdup_user(arg, sizeof(*di_args));
+	if (IS_ERR(di_args))
+		return PTR_ERR(di_args);
+
+	if (memcmp(empty_uuid, di_args->uuid, BTRFS_UUID_SIZE) != 0)
+		s_uuid = di_args->uuid;
+
+	mutex_lock(&fs_devices->device_list_mutex);
+	dev = btrfs_find_device(root, di_args->devid, s_uuid, NULL);
+	mutex_unlock(&fs_devices->device_list_mutex);
+
+	if (!dev) {
+		ret = -ENODEV;
+		goto out;
+	}
+
+	di_args->devid = dev->devid;
+	di_args->bytes_used = dev->bytes_used;
+	di_args->total_bytes = dev->total_bytes;
+	memcpy(di_args->uuid, dev->uuid, sizeof(di_args->uuid));
+	strncpy(di_args->path, dev->name, sizeof(di_args->path));
+
+out:
+	if (ret == 0 && copy_to_user(arg, di_args, sizeof(*di_args)))
+		ret = -EFAULT;
+
+	kfree(di_args);
+	return ret;
+}
+
 static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
 				       u64 off, u64 olen, u64 destoff)
 {
@@ -1925,7 +2237,7 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
 	}
 
 	/* clone data */
-	key.objectid = src->i_ino;
+	key.objectid = btrfs_ino(src);
 	key.type = BTRFS_EXTENT_DATA_KEY;
 	key.offset = 0;
 
@@ -1952,7 +2264,7 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
 
 		btrfs_item_key_to_cpu(leaf, &key, slot);
 		if (btrfs_key_type(&key) > BTRFS_EXTENT_DATA_KEY ||
-		    key.objectid != src->i_ino)
+		    key.objectid != btrfs_ino(src))
 			break;
 
 		if (btrfs_key_type(&key) == BTRFS_EXTENT_DATA_KEY) {
@@ -1988,14 +2300,14 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
 				datal = btrfs_file_extent_ram_bytes(leaf,
 								    extent);
 			}
-			btrfs_release_path(root, path);
+			btrfs_release_path(path);
 
 			if (key.offset + datal <= off ||
 			    key.offset >= off+len)
 				goto next;
 
 			memcpy(&new_key, &key, sizeof(new_key));
-			new_key.objectid = inode->i_ino;
+			new_key.objectid = btrfs_ino(inode);
 			if (off <= key.offset)
 				new_key.offset = key.offset + destoff - off;
 			else
@@ -2049,7 +2361,7 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
 					ret = btrfs_inc_extent_ref(trans, root,
 							disko, diskl, 0,
 							root->root_key.objectid,
-							inode->i_ino,
+							btrfs_ino(inode),
 							new_key.offset - datao);
 					BUG_ON(ret);
 				}
@@ -2098,7 +2410,7 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
 			}
 
 			btrfs_mark_buffer_dirty(leaf);
-			btrfs_release_path(root, path);
+			btrfs_release_path(path);
 
 			inode->i_mtime = inode->i_ctime = CURRENT_TIME;
 
@@ -2119,12 +2431,12 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
 			btrfs_end_transaction(trans, root);
 		}
 next:
-		btrfs_release_path(root, path);
+		btrfs_release_path(path);
 		key.offset++;
 	}
 	ret = 0;
 out:
-	btrfs_release_path(root, path);
+	btrfs_release_path(path);
 	unlock_extent(&BTRFS_I(src)->io_tree, off, off+len, GFP_NOFS);
 out_unlock:
 	mutex_unlock(&src->i_mutex);
@@ -2471,6 +2783,58 @@ static noinline long btrfs_ioctl_wait_sync(struct file *file, void __user *argp)
 	return btrfs_wait_for_commit(root, transid);
 }
 
+static long btrfs_ioctl_scrub(struct btrfs_root *root, void __user *arg)
+{
+	int ret;
+	struct btrfs_ioctl_scrub_args *sa;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+
+	sa = memdup_user(arg, sizeof(*sa));
+	if (IS_ERR(sa))
+		return PTR_ERR(sa);
+
+	ret = btrfs_scrub_dev(root, sa->devid, sa->start, sa->end,
+			      &sa->progress, sa->flags & BTRFS_SCRUB_READONLY);
+
+	if (copy_to_user(arg, sa, sizeof(*sa)))
+		ret = -EFAULT;
+
+	kfree(sa);
+	return ret;
+}
+
+static long btrfs_ioctl_scrub_cancel(struct btrfs_root *root, void __user *arg)
+{
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+
+	return btrfs_scrub_cancel(root);
+}
+
+static long btrfs_ioctl_scrub_progress(struct btrfs_root *root,
+				       void __user *arg)
+{
+	struct btrfs_ioctl_scrub_args *sa;
+	int ret;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+
+	sa = memdup_user(arg, sizeof(*sa));
+	if (IS_ERR(sa))
+		return PTR_ERR(sa);
+
+	ret = btrfs_scrub_progress(root, sa->devid, &sa->progress);
+
+	if (copy_to_user(arg, sa, sizeof(*sa)))
+		ret = -EFAULT;
+
+	kfree(sa);
+	return ret;
+}
+
 long btrfs_ioctl(struct file *file, unsigned int
 		cmd, unsigned long arg)
 {
@@ -2510,6 +2874,10 @@ long btrfs_ioctl(struct file *file, unsigned int
 		return btrfs_ioctl_add_dev(root, argp);
 	case BTRFS_IOC_RM_DEV:
 		return btrfs_ioctl_rm_dev(root, argp);
+	case BTRFS_IOC_FS_INFO:
+		return btrfs_ioctl_fs_info(root, argp);
+	case BTRFS_IOC_DEV_INFO:
+		return btrfs_ioctl_dev_info(root, argp);
 	case BTRFS_IOC_BALANCE:
 		return btrfs_balance(root->fs_info->dev_root);
 	case BTRFS_IOC_CLONE:
@@ -2533,6 +2901,12 @@ long btrfs_ioctl(struct file *file, unsigned int
 		return btrfs_ioctl_start_sync(file, argp);
 	case BTRFS_IOC_WAIT_SYNC:
 		return btrfs_ioctl_wait_sync(file, argp);
+	case BTRFS_IOC_SCRUB:
+		return btrfs_ioctl_scrub(root, argp);
+	case BTRFS_IOC_SCRUB_CANCEL:
+		return btrfs_ioctl_scrub_cancel(root, argp);
+	case BTRFS_IOC_SCRUB_PROGRESS:
+		return btrfs_ioctl_scrub_progress(root, argp);
 	}
 
 	return -ENOTTY;

fs/btrfs/ioctl.h

@@ -32,6 +32,8 @@ struct btrfs_ioctl_vol_args {
 
 #define BTRFS_SUBVOL_CREATE_ASYNC	(1ULL << 0)
 #define BTRFS_SUBVOL_RDONLY		(1ULL << 1)
+#define BTRFS_FSID_SIZE 16
+#define BTRFS_UUID_SIZE 16
 
 #define BTRFS_SUBVOL_NAME_MAX 4039
 struct btrfs_ioctl_vol_args_v2 {
@@ -42,6 +44,71 @@ struct btrfs_ioctl_vol_args_v2 {
 	char name[BTRFS_SUBVOL_NAME_MAX + 1];
 };
 
+/*
+ * structure to report errors and progress to userspace, either as a
+ * result of a finished scrub, a canceled scrub or a progress inquiry
+ */
+struct btrfs_scrub_progress {
+	__u64 data_extents_scrubbed;	/* # of data extents scrubbed */
+	__u64 tree_extents_scrubbed;	/* # of tree extents scrubbed */
+	__u64 data_bytes_scrubbed;	/* # of data bytes scrubbed */
+	__u64 tree_bytes_scrubbed;	/* # of tree bytes scrubbed */
+	__u64 read_errors;		/* # of read errors encountered (EIO) */
+	__u64 csum_errors;		/* # of failed csum checks */
+	__u64 verify_errors;		/* # of occurences, where the metadata
+					 * of a tree block did not match the
+					 * expected values, like generation or
+					 * logical */
+	__u64 no_csum;			/* # of 4k data block for which no csum
+					 * is present, probably the result of
+					 * data written with nodatasum */
+	__u64 csum_discards;		/* # of csum for which no data was found
+					 * in the extent tree. */
+	__u64 super_errors;		/* # of bad super blocks encountered */
+	__u64 malloc_errors;		/* # of internal kmalloc errors. These
+					 * will likely cause an incomplete
+					 * scrub */
+	__u64 uncorrectable_errors;	/* # of errors where either no intact
+					 * copy was found or the writeback
+					 * failed */
+	__u64 corrected_errors;		/* # of errors corrected */
+	__u64 last_physical;		/* last physical address scrubbed. In
+					 * case a scrub was aborted, this can
+					 * be used to restart the scrub */
+	__u64 unverified_errors;	/* # of occurences where a read for a
+					 * full (64k) bio failed, but the re-
+					 * check succeeded for each 4k piece.
+					 * Intermittent error. */
+};
+
+#define BTRFS_SCRUB_READONLY	1
+struct btrfs_ioctl_scrub_args {
+	__u64 devid;				/* in */
+	__u64 start;				/* in */
+	__u64 end;				/* in */
+	__u64 flags;				/* in */
+	struct btrfs_scrub_progress progress;	/* out */
+	/* pad to 1k */
+	__u64 unused[(1024-32-sizeof(struct btrfs_scrub_progress))/8];
+};
+
+#define BTRFS_DEVICE_PATH_NAME_MAX 1024
+struct btrfs_ioctl_dev_info_args {
+	__u64 devid;				/* in/out */
+	__u8 uuid[BTRFS_UUID_SIZE];		/* in/out */
+	__u64 bytes_used;			/* out */
+	__u64 total_bytes;			/* out */
+	__u64 unused[379];			/* pad to 4k */
+	__u8 path[BTRFS_DEVICE_PATH_NAME_MAX];	/* out */
+};
+
+struct btrfs_ioctl_fs_info_args {
+	__u64 max_id;				/* out */
+	__u64 num_devices;			/* out */
+	__u8 fsid[BTRFS_FSID_SIZE];		/* out */
+	__u64 reserved[124];			/* pad to 1k */
+};
+
 #define BTRFS_INO_LOOKUP_PATH_MAX 4080
 struct btrfs_ioctl_ino_lookup_args {
 	__u64 treeid;
@@ -114,37 +181,6 @@ struct btrfs_ioctl_clone_range_args {
 #define BTRFS_DEFRAG_RANGE_COMPRESS 1
 #define BTRFS_DEFRAG_RANGE_START_IO 2
 
-struct btrfs_ioctl_defrag_range_args {
-	/* start of the defrag operation */
-	__u64 start;
-
-	/* number of bytes to defrag, use (u64)-1 to say all */
-	__u64 len;
-
-	/*
-	 * flags for the operation, which can include turning
-	 * on compression for this one defrag
-	 */
-	__u64 flags;
-
-	/*
-	 * any extent bigger than this will be considered
-	 * already defragged.  Use 0 to take the kernel default
-	 * Use 1 to say every single extent must be rewritten
-	 */
-	__u32 extent_thresh;
-
-	/*
-	 * which compression method to use if turning on compression
-	 * for this defrag operation.  If unspecified, zlib will
-	 * be used
-	 */
-	__u32 compress_type;
-
-	/* spare for later */
-	__u32 unused[4];
-};
-
 struct btrfs_ioctl_space_info {
 	__u64 flags;
 	__u64 total_bytes;
@@ -203,4 +239,13 @@ struct btrfs_ioctl_space_args {
 				   struct btrfs_ioctl_vol_args_v2)
 #define BTRFS_IOC_SUBVOL_GETFLAGS _IOW(BTRFS_IOCTL_MAGIC, 25, __u64)
 #define BTRFS_IOC_SUBVOL_SETFLAGS _IOW(BTRFS_IOCTL_MAGIC, 26, __u64)
+#define BTRFS_IOC_SCRUB _IOWR(BTRFS_IOCTL_MAGIC, 27, \
+			      struct btrfs_ioctl_scrub_args)
+#define BTRFS_IOC_SCRUB_CANCEL _IO(BTRFS_IOCTL_MAGIC, 28)
+#define BTRFS_IOC_SCRUB_PROGRESS _IOWR(BTRFS_IOCTL_MAGIC, 29, \
+				       struct btrfs_ioctl_scrub_args)
+#define BTRFS_IOC_DEV_INFO _IOWR(BTRFS_IOCTL_MAGIC, 30, \
+				 struct btrfs_ioctl_dev_info_args)
+#define BTRFS_IOC_FS_INFO _IOR(BTRFS_IOCTL_MAGIC, 31, \
+			       struct btrfs_ioctl_fs_info_args)
 #endif

fs/btrfs/locking.c

@@ -185,31 +185,6 @@ sleep:
 	return 0;
 }
 
-/*
- * Very quick trylock, this does not spin or schedule.  It returns
- * 1 with the spinlock held if it was able to take the lock, or it
- * returns zero if it was unable to take the lock.
- *
- * After this call, scheduling is not safe without first calling
- * btrfs_set_lock_blocking()
- */
-int btrfs_try_tree_lock(struct extent_buffer *eb)
-{
-	if (spin_trylock(&eb->lock)) {
-		if (test_bit(EXTENT_BUFFER_BLOCKING, &eb->bflags)) {
-			/*
-			 * we've got the spinlock, but the real owner is
-			 * blocking.  Drop the spinlock and return failure
-			 */
-			spin_unlock(&eb->lock);
-			return 0;
-		}
-		return 1;
-	}
-	/* someone else has the spinlock giveup */
-	return 0;
-}
-
 int btrfs_tree_unlock(struct extent_buffer *eb)
 {
 	/*

fs/btrfs/locking.h

@@ -21,8 +21,6 @@
 
 int btrfs_tree_lock(struct extent_buffer *eb);
 int btrfs_tree_unlock(struct extent_buffer *eb);
-
-int btrfs_try_tree_lock(struct extent_buffer *eb);
 int btrfs_try_spin_lock(struct extent_buffer *eb);
 
 void btrfs_set_lock_blocking(struct extent_buffer *eb);

fs/btrfs/ref-cache.c

@@ -23,56 +23,6 @@
 #include "ref-cache.h"
 #include "transaction.h"
 
-/*
- * leaf refs are used to cache the information about which extents
- * a given leaf has references on.  This allows us to process that leaf
- * in btrfs_drop_snapshot without needing to read it back from disk.
- */
-
-/*
- * kmalloc a leaf reference struct and update the counters for the
- * total ref cache size
- */
-struct btrfs_leaf_ref *btrfs_alloc_leaf_ref(struct btrfs_root *root,
-					    int nr_extents)
-{
-	struct btrfs_leaf_ref *ref;
-	size_t size = btrfs_leaf_ref_size(nr_extents);
-
-	ref = kmalloc(size, GFP_NOFS);
-	if (ref) {
-		spin_lock(&root->fs_info->ref_cache_lock);
-		root->fs_info->total_ref_cache_size += size;
-		spin_unlock(&root->fs_info->ref_cache_lock);
-
-		memset(ref, 0, sizeof(*ref));
-		atomic_set(&ref->usage, 1);
-		INIT_LIST_HEAD(&ref->list);
-	}
-	return ref;
-}
-
-/*
- * free a leaf reference struct and update the counters for the
- * total ref cache size
- */
-void btrfs_free_leaf_ref(struct btrfs_root *root, struct btrfs_leaf_ref *ref)
-{
-	if (!ref)
-		return;
-	WARN_ON(atomic_read(&ref->usage) == 0);
-	if (atomic_dec_and_test(&ref->usage)) {
-		size_t size = btrfs_leaf_ref_size(ref->nritems);
-
-		BUG_ON(ref->in_tree);
-		kfree(ref);
-
-		spin_lock(&root->fs_info->ref_cache_lock);
-		root->fs_info->total_ref_cache_size -= size;
-		spin_unlock(&root->fs_info->ref_cache_lock);
-	}
-}
-
 static struct rb_node *tree_insert(struct rb_root *root, u64 bytenr,
 				   struct rb_node *node)
 {
@@ -116,117 +66,3 @@ static struct rb_node *tree_search(struct rb_root *root, u64 bytenr)
 	}
 	return NULL;
 }
-
-int btrfs_remove_leaf_refs(struct btrfs_root *root, u64 max_root_gen,
-			   int shared)
-{
-	struct btrfs_leaf_ref *ref = NULL;
-	struct btrfs_leaf_ref_tree *tree = root->ref_tree;
-
-	if (shared)
-		tree = &root->fs_info->shared_ref_tree;
-	if (!tree)
-		return 0;
-
-	spin_lock(&tree->lock);
-	while (!list_empty(&tree->list)) {
-		ref = list_entry(tree->list.next, struct btrfs_leaf_ref, list);
-		BUG_ON(ref->tree != tree);
-		if (ref->root_gen > max_root_gen)
-			break;
-		if (!xchg(&ref->in_tree, 0)) {
-			cond_resched_lock(&tree->lock);
-			continue;
-		}
-
-		rb_erase(&ref->rb_node, &tree->root);
-		list_del_init(&ref->list);
-
-		spin_unlock(&tree->lock);
-		btrfs_free_leaf_ref(root, ref);
-		cond_resched();
-		spin_lock(&tree->lock);
-	}
-	spin_unlock(&tree->lock);
-	return 0;
-}
-
-/*
- * find the leaf ref for a given extent.  This returns the ref struct with
- * a usage reference incremented
- */
-struct btrfs_leaf_ref *btrfs_lookup_leaf_ref(struct btrfs_root *root,
-					     u64 bytenr)
-{
-	struct rb_node *rb;
-	struct btrfs_leaf_ref *ref = NULL;
-	struct btrfs_leaf_ref_tree *tree = root->ref_tree;
-again:
-	if (tree) {
-		spin_lock(&tree->lock);
-		rb = tree_search(&tree->root, bytenr);
-		if (rb)
-			ref = rb_entry(rb, struct btrfs_leaf_ref, rb_node);
-		if (ref)
-			atomic_inc(&ref->usage);
-		spin_unlock(&tree->lock);
-		if (ref)
-			return ref;
-	}
-	if (tree != &root->fs_info->shared_ref_tree) {
-		tree = &root->fs_info->shared_ref_tree;
-		goto again;
-	}
-	return NULL;
-}
-
-/*
- * add a fully filled in leaf ref struct
- * remove all the refs older than a given root generation
- */
-int btrfs_add_leaf_ref(struct btrfs_root *root, struct btrfs_leaf_ref *ref,
-		       int shared)
-{
-	int ret = 0;
-	struct rb_node *rb;
-	struct btrfs_leaf_ref_tree *tree = root->ref_tree;
-
-	if (shared)
-		tree = &root->fs_info->shared_ref_tree;
-
-	spin_lock(&tree->lock);
-	rb = tree_insert(&tree->root, ref->bytenr, &ref->rb_node);
-	if (rb) {
-		ret = -EEXIST;
-	} else {
-		atomic_inc(&ref->usage);
-		ref->tree = tree;
-		ref->in_tree = 1;
-		list_add_tail(&ref->list, &tree->list);
-	}
-	spin_unlock(&tree->lock);
-	return ret;
-}
-
-/*
- * remove a single leaf ref from the tree.  This drops the ref held by the tree
- * only
- */
-int btrfs_remove_leaf_ref(struct btrfs_root *root, struct btrfs_leaf_ref *ref)
-{
-	struct btrfs_leaf_ref_tree *tree;
-
-	if (!xchg(&ref->in_tree, 0))
-		return 0;
-
-	tree = ref->tree;
-	spin_lock(&tree->lock);
-
-	rb_erase(&ref->rb_node, &tree->root);
-	list_del_init(&ref->list);
-
-	spin_unlock(&tree->lock);
-
-	btrfs_free_leaf_ref(root, ref);
-	return 0;
-}

fs/btrfs/ref-cache.h

@@ -49,28 +49,4 @@ static inline size_t btrfs_leaf_ref_size(int nr_extents)
 	return sizeof(struct btrfs_leaf_ref) +
 	       sizeof(struct btrfs_extent_info) * nr_extents;
 }
-
-static inline void btrfs_leaf_ref_tree_init(struct btrfs_leaf_ref_tree *tree)
-{
-	tree->root = RB_ROOT;
-	INIT_LIST_HEAD(&tree->list);
-	spin_lock_init(&tree->lock);
-}
-
-static inline int btrfs_leaf_ref_tree_empty(struct btrfs_leaf_ref_tree *tree)
-{
-	return RB_EMPTY_ROOT(&tree->root);
-}
-
-void btrfs_leaf_ref_tree_init(struct btrfs_leaf_ref_tree *tree);
-struct btrfs_leaf_ref *btrfs_alloc_leaf_ref(struct btrfs_root *root,
-					    int nr_extents);
-void btrfs_free_leaf_ref(struct btrfs_root *root, struct btrfs_leaf_ref *ref);
-struct btrfs_leaf_ref *btrfs_lookup_leaf_ref(struct btrfs_root *root,
-					     u64 bytenr);
-int btrfs_add_leaf_ref(struct btrfs_root *root, struct btrfs_leaf_ref *ref,
-		       int shared);
-int btrfs_remove_leaf_refs(struct btrfs_root *root, u64 max_root_gen,
-			   int shared);
-int btrfs_remove_leaf_ref(struct btrfs_root *root, struct btrfs_leaf_ref *ref);
 #endif

fs/btrfs/relocation.c

@@ -30,6 +30,7 @@
 #include "btrfs_inode.h"
 #include "async-thread.h"
 #include "free-space-cache.h"
+#include "inode-map.h"
 
 /*
  * backref_node, mapping_node and tree_block start with this
@@ -507,6 +508,7 @@ static int update_backref_cache(struct btrfs_trans_handle *trans,
 	return 1;
 }
 
+
 static int should_ignore_root(struct btrfs_root *root)
 {
 	struct btrfs_root *reloc_root;
@@ -529,7 +531,6 @@ static int should_ignore_root(struct btrfs_root *root)
 	 */
 	return 1;
 }
-
 /*
  * find reloc tree by address of tree root
  */
@@ -961,7 +962,7 @@ again:
 			lower = upper;
 			upper = NULL;
 		}
-		btrfs_release_path(root, path2);
+		btrfs_release_path(path2);
 next:
 		if (ptr < end) {
 			ptr += btrfs_extent_inline_ref_size(key.type);
@@ -974,7 +975,7 @@ next:
 		if (ptr >= end)
 			path1->slots[0]++;
 	}
-	btrfs_release_path(rc->extent_root, path1);
+	btrfs_release_path(path1);
 
 	cur->checked = 1;
 	WARN_ON(exist);
@@ -1409,9 +1410,9 @@ again:
 		prev = node;
 		entry = rb_entry(node, struct btrfs_inode, rb_node);
 
-		if (objectid < entry->vfs_inode.i_ino)
+		if (objectid < btrfs_ino(&entry->vfs_inode))
 			node = node->rb_left;
-		else if (objectid > entry->vfs_inode.i_ino)
+		else if (objectid > btrfs_ino(&entry->vfs_inode))
 			node = node->rb_right;
 		else
 			break;
@@ -1419,7 +1420,7 @@ again:
 	if (!node) {
 		while (prev) {
 			entry = rb_entry(prev, struct btrfs_inode, rb_node);
-			if (objectid <= entry->vfs_inode.i_ino) {
+			if (objectid <= btrfs_ino(&entry->vfs_inode)) {
 				node = prev;
 				break;
 			}
@@ -1434,7 +1435,7 @@ again:
 			return inode;
 		}
 
-		objectid = entry->vfs_inode.i_ino + 1;
+		objectid = btrfs_ino(&entry->vfs_inode) + 1;
 		if (cond_resched_lock(&root->inode_lock))
 			goto again;
 
@@ -1470,7 +1471,7 @@ static int get_new_location(struct inode *reloc_inode, u64 *new_bytenr,
 		return -ENOMEM;
 
 	bytenr -= BTRFS_I(reloc_inode)->index_cnt;
-	ret = btrfs_lookup_file_extent(NULL, root, path, reloc_inode->i_ino,
+	ret = btrfs_lookup_file_extent(NULL, root, path, btrfs_ino(reloc_inode),
 				       bytenr, 0);
 	if (ret < 0)
 		goto out;
@@ -1558,11 +1559,11 @@ int replace_file_extents(struct btrfs_trans_handle *trans,
 			if (first) {
 				inode = find_next_inode(root, key.objectid);
 				first = 0;
-			} else if (inode && inode->i_ino < key.objectid) {
+			} else if (inode && btrfs_ino(inode) < key.objectid) {
 				btrfs_add_delayed_iput(inode);
 				inode = find_next_inode(root, key.objectid);
 			}
-			if (inode && inode->i_ino == key.objectid) {
+			if (inode && btrfs_ino(inode) == key.objectid) {
 				end = key.offset +
 				      btrfs_file_extent_num_bytes(leaf, fi);
 				WARN_ON(!IS_ALIGNED(key.offset,
@@ -1749,7 +1750,7 @@ again:
 
 		btrfs_node_key_to_cpu(path->nodes[level], &key,
 				      path->slots[level]);
-		btrfs_release_path(src, path);
+		btrfs_release_path(path);
 
 		path->lowest_level = level;
 		ret = btrfs_search_slot(trans, src, &key, path, 0, 1);
@@ -1893,6 +1894,7 @@ static int invalidate_extent_cache(struct btrfs_root *root,
 	struct inode *inode = NULL;
 	u64 objectid;
 	u64 start, end;
+	u64 ino;
 
 	objectid = min_key->objectid;
 	while (1) {
@@ -1905,17 +1907,18 @@ static int invalidate_extent_cache(struct btrfs_root *root,
 		inode = find_next_inode(root, objectid);
 		if (!inode)
 			break;
+		ino = btrfs_ino(inode);
 
-		if (inode->i_ino > max_key->objectid) {
+		if (ino > max_key->objectid) {
 			iput(inode);
 			break;
 		}
 
-		objectid = inode->i_ino + 1;
+		objectid = ino + 1;
 		if (!S_ISREG(inode->i_mode))
 			continue;
 
-		if (unlikely(min_key->objectid == inode->i_ino)) {
+		if (unlikely(min_key->objectid == ino)) {
 			if (min_key->type > BTRFS_EXTENT_DATA_KEY)
 				continue;
 			if (min_key->type < BTRFS_EXTENT_DATA_KEY)
@@ -1928,7 +1931,7 @@ static int invalidate_extent_cache(struct btrfs_root *root,
 			start = 0;
 		}
 
-		if (unlikely(max_key->objectid == inode->i_ino)) {
+		if (unlikely(max_key->objectid == ino)) {
 			if (max_key->type < BTRFS_EXTENT_DATA_KEY)
 				continue;
 			if (max_key->type > BTRFS_EXTENT_DATA_KEY) {
@@ -2496,7 +2499,7 @@ static int do_relocation(struct btrfs_trans_handle *trans,
 			path->locks[upper->level] = 0;
 
 			slot = path->slots[upper->level];
-			btrfs_release_path(NULL, path);
+			btrfs_release_path(path);
 		} else {
 			ret = btrfs_bin_search(upper->eb, key, upper->level,
 					       &slot);
@@ -2737,7 +2740,7 @@ static int relocate_tree_block(struct btrfs_trans_handle *trans,
 		} else {
 			path->lowest_level = node->level;
 			ret = btrfs_search_slot(trans, root, key, path, 0, 1);
-			btrfs_release_path(root, path);
+			btrfs_release_path(path);
 			if (ret > 0)
 				ret = 0;
 		}
@@ -2870,7 +2873,7 @@ int setup_extent_mapping(struct inode *inode, u64 start, u64 end,
 	struct extent_map *em;
 	int ret = 0;
 
-	em = alloc_extent_map(GFP_NOFS);
+	em = alloc_extent_map();
 	if (!em)
 		return -ENOMEM;
 
@@ -3119,7 +3122,7 @@ static int add_tree_block(struct reloc_control *rc,
 #endif
 	}
 
-	btrfs_release_path(rc->extent_root, path);
+	btrfs_release_path(path);
 
 	BUG_ON(level == -1);
 
@@ -3220,7 +3223,7 @@ static int delete_block_group_cache(struct btrfs_fs_info *fs_info,
 	key.offset = 0;
 
 	inode = btrfs_iget(fs_info->sb, &key, root, NULL);
-	if (!inode || IS_ERR(inode) || is_bad_inode(inode)) {
+	if (IS_ERR_OR_NULL(inode) || is_bad_inode(inode)) {
 		if (inode && !IS_ERR(inode))
 			iput(inode);
 		return -ENOENT;
@@ -3505,7 +3508,7 @@ int add_data_references(struct reloc_control *rc,
 		}
 		path->slots[0]++;
 	}
-	btrfs_release_path(rc->extent_root, path);
+	btrfs_release_path(path);
 	if (err)
 		free_block_list(blocks);
 	return err;
@@ -3568,7 +3571,7 @@ next:
 					    EXTENT_DIRTY);
 
 		if (ret == 0 && start <= key.objectid) {
-			btrfs_release_path(rc->extent_root, path);
+			btrfs_release_path(path);
 			rc->search_start = end + 1;
 		} else {
 			rc->search_start = key.objectid + key.offset;
@@ -3576,7 +3579,7 @@ next:
 			return 0;
 		}
 	}
-	btrfs_release_path(rc->extent_root, path);
+	btrfs_release_path(path);
 	return ret;
 }
 
@@ -3713,7 +3716,7 @@ restart:
 				flags = BTRFS_EXTENT_FLAG_DATA;
 
 			if (path_change) {
-				btrfs_release_path(rc->extent_root, path);
+				btrfs_release_path(path);
 
 				path->search_commit_root = 1;
 				path->skip_locking = 1;
@@ -3736,7 +3739,7 @@ restart:
 			   (flags & BTRFS_EXTENT_FLAG_DATA)) {
 			ret = add_data_references(rc, &key, path, &blocks);
 		} else {
-			btrfs_release_path(rc->extent_root, path);
+			btrfs_release_path(path);
 			ret = 0;
 		}
 		if (ret < 0) {
@@ -3799,7 +3802,7 @@ restart:
 		}
 	}
 
-	btrfs_release_path(rc->extent_root, path);
+	btrfs_release_path(path);
 	clear_extent_bits(&rc->processed_blocks, 0, (u64)-1, EXTENT_DIRTY,
 			  GFP_NOFS);
 
@@ -3867,7 +3870,7 @@ static int __insert_orphan_inode(struct btrfs_trans_handle *trans,
 	btrfs_set_inode_flags(leaf, item, BTRFS_INODE_NOCOMPRESS |
 					  BTRFS_INODE_PREALLOC);
 	btrfs_mark_buffer_dirty(leaf);
-	btrfs_release_path(root, path);
+	btrfs_release_path(path);
 out:
 	btrfs_free_path(path);
 	return ret;
@@ -3897,7 +3900,7 @@ struct inode *create_reloc_inode(struct btrfs_fs_info *fs_info,
 	if (IS_ERR(trans))
 		return ERR_CAST(trans);
 
-	err = btrfs_find_free_objectid(trans, root, objectid, &objectid);
+	err = btrfs_find_free_objectid(root, &objectid);
 	if (err)
 		goto out;
 
@@ -3935,7 +3938,7 @@ static struct reloc_control *alloc_reloc_control(void)
 	INIT_LIST_HEAD(&rc->reloc_roots);
 	backref_cache_init(&rc->backref_cache);
 	mapping_tree_init(&rc->reloc_root_tree);
-	extent_io_tree_init(&rc->processed_blocks, NULL, GFP_NOFS);
+	extent_io_tree_init(&rc->processed_blocks, NULL);
 	return rc;
 }
 
@@ -4109,7 +4112,7 @@ int btrfs_recover_relocation(struct btrfs_root *root)
 		}
 		leaf = path->nodes[0];
 		btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
-		btrfs_release_path(root->fs_info->tree_root, path);
+		btrfs_release_path(path);
 
 		if (key.objectid != BTRFS_TREE_RELOC_OBJECTID ||
 		    key.type != BTRFS_ROOT_ITEM_KEY)
@@ -4141,7 +4144,7 @@ int btrfs_recover_relocation(struct btrfs_root *root)
 
 		key.offset--;
 	}
-	btrfs_release_path(root->fs_info->tree_root, path);
+	btrfs_release_path(path);
 
 	if (list_empty(&reloc_roots))
 		goto out;
@@ -4242,7 +4245,7 @@ int btrfs_reloc_clone_csums(struct inode *inode, u64 file_pos, u64 len)
 
 	disk_bytenr = file_pos + BTRFS_I(inode)->index_cnt;
 	ret = btrfs_lookup_csums_range(root->fs_info->csum_root, disk_bytenr,
-				       disk_bytenr + len - 1, &list);
+				       disk_bytenr + len - 1, &list, 0);
 
 	while (!list_empty(&list)) {
 		sums = list_entry(list.next, struct btrfs_ordered_sum, list);

fs/btrfs/root-tree.c

@@ -21,53 +21,6 @@
 #include "disk-io.h"
 #include "print-tree.h"
 
-/*
- *  search forward for a root, starting with objectid 'search_start'
- *  if a root key is found, the objectid we find is filled into 'found_objectid'
- *  and 0 is returned.  < 0 is returned on error, 1 if there is nothing
- *  left in the tree.
- */
-int btrfs_search_root(struct btrfs_root *root, u64 search_start,
-		      u64 *found_objectid)
-{
-	struct btrfs_path *path;
-	struct btrfs_key search_key;
-	int ret;
-
-	root = root->fs_info->tree_root;
-	search_key.objectid = search_start;
-	search_key.type = (u8)-1;
-	search_key.offset = (u64)-1;
-
-	path = btrfs_alloc_path();
-	BUG_ON(!path);
-again:
-	ret = btrfs_search_slot(NULL, root, &search_key, path, 0, 0);
-	if (ret < 0)
-		goto out;
-	if (ret == 0) {
-		ret = 1;
-		goto out;
-	}
-	if (path->slots[0] >= btrfs_header_nritems(path->nodes[0])) {
-		ret = btrfs_next_leaf(root, path);
-		if (ret)
-			goto out;
-	}
-	btrfs_item_key_to_cpu(path->nodes[0], &search_key, path->slots[0]);
-	if (search_key.type != BTRFS_ROOT_ITEM_KEY) {
-		search_key.offset++;
-		btrfs_release_path(root, path);
-		goto again;
-	}
-	ret = 0;
-	*found_objectid = search_key.objectid;
-
-out:
-	btrfs_free_path(path);
-	return ret;
-}
-
 /*
  * lookup the root with the highest offset for a given objectid.  The key we do
  * find is copied into 'key'.  If we find something return 0, otherwise 1, < 0
@@ -230,7 +183,7 @@ again:
 
 		memcpy(&found_key, &key, sizeof(key));
 		key.offset++;
-		btrfs_release_path(root, path);
+		btrfs_release_path(path);
 		dead_root =
 			btrfs_read_fs_root_no_radix(root->fs_info->tree_root,
 						    &found_key);
@@ -292,7 +245,7 @@ int btrfs_find_orphan_roots(struct btrfs_root *tree_root)
 		}
 
 		btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
-		btrfs_release_path(tree_root, path);
+		btrfs_release_path(path);
 
 		if (key.objectid != BTRFS_ORPHAN_OBJECTID ||
 		    key.type != BTRFS_ORPHAN_ITEM_KEY)
@@ -385,18 +338,22 @@ again:
 		*sequence = btrfs_root_ref_sequence(leaf, ref);
 
 		ret = btrfs_del_item(trans, tree_root, path);
-		BUG_ON(ret);
+		if (ret) {
+			err = ret;
+			goto out;
+		}
 	} else
 		err = -ENOENT;
 
 	if (key.type == BTRFS_ROOT_BACKREF_KEY) {
-		btrfs_release_path(tree_root, path);
+		btrfs_release_path(path);
 		key.objectid = ref_id;
 		key.type = BTRFS_ROOT_REF_KEY;
 		key.offset = root_id;
 		goto again;
 	}
 
+out:
 	btrfs_free_path(path);
 	return err;
 }
@@ -463,7 +420,7 @@ again:
 	btrfs_mark_buffer_dirty(leaf);
 
 	if (key.type == BTRFS_ROOT_BACKREF_KEY) {
-		btrfs_release_path(tree_root, path);
+		btrfs_release_path(path);
 		key.objectid = ref_id;
 		key.type = BTRFS_ROOT_REF_KEY;
 		key.offset = root_id;

fs/btrfs/scrub.c

@@ -0,0 +1,1369 @@
+/*
+ * Copyright (C) 2011 STRATO.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include <linux/sched.h>
+#include <linux/pagemap.h>
+#include <linux/writeback.h>
+#include <linux/blkdev.h>
+#include <linux/rbtree.h>
+#include <linux/slab.h>
+#include <linux/workqueue.h>
+#include "ctree.h"
+#include "volumes.h"
+#include "disk-io.h"
+#include "ordered-data.h"
+
+/*
+ * This is only the first step towards a full-features scrub. It reads all
+ * extent and super block and verifies the checksums. In case a bad checksum
+ * is found or the extent cannot be read, good data will be written back if
+ * any can be found.
+ *
+ * Future enhancements:
+ *  - To enhance the performance, better read-ahead strategies for the
+ *    extent-tree can be employed.
+ *  - In case an unrepairable extent is encountered, track which files are
+ *    affected and report them
+ *  - In case of a read error on files with nodatasum, map the file and read
+ *    the extent to trigger a writeback of the good copy
+ *  - track and record media errors, throw out bad devices
+ *  - add a mode to also read unallocated space
+ *  - make the prefetch cancellable
+ */
+
+struct scrub_bio;
+struct scrub_page;
+struct scrub_dev;
+static void scrub_bio_end_io(struct bio *bio, int err);
+static void scrub_checksum(struct btrfs_work *work);
+static int scrub_checksum_data(struct scrub_dev *sdev,
+			       struct scrub_page *spag, void *buffer);
+static int scrub_checksum_tree_block(struct scrub_dev *sdev,
+				     struct scrub_page *spag, u64 logical,
+				     void *buffer);
+static int scrub_checksum_super(struct scrub_bio *sbio, void *buffer);
+static int scrub_fixup_check(struct scrub_bio *sbio, int ix);
+static void scrub_fixup_end_io(struct bio *bio, int err);
+static int scrub_fixup_io(int rw, struct block_device *bdev, sector_t sector,
+			  struct page *page);
+static void scrub_fixup(struct scrub_bio *sbio, int ix);
+
+#define SCRUB_PAGES_PER_BIO	16	/* 64k per bio */
+#define SCRUB_BIOS_PER_DEV	16	/* 1 MB per device in flight */
+
+struct scrub_page {
+	u64			flags;  /* extent flags */
+	u64			generation;
+	u64			mirror_num;
+	int			have_csum;
+	u8			csum[BTRFS_CSUM_SIZE];
+};
+
+struct scrub_bio {
+	int			index;
+	struct scrub_dev	*sdev;
+	struct bio		*bio;
+	int			err;
+	u64			logical;
+	u64			physical;
+	struct scrub_page	spag[SCRUB_PAGES_PER_BIO];
+	u64			count;
+	int			next_free;
+	struct btrfs_work	work;
+};
+
+struct scrub_dev {
+	struct scrub_bio	*bios[SCRUB_BIOS_PER_DEV];
+	struct btrfs_device	*dev;
+	int			first_free;
+	int			curr;
+	atomic_t		in_flight;
+	spinlock_t		list_lock;
+	wait_queue_head_t	list_wait;
+	u16			csum_size;
+	struct list_head	csum_list;
+	atomic_t		cancel_req;
+	int			readonly;
+	/*
+	 * statistics
+	 */
+	struct btrfs_scrub_progress stat;
+	spinlock_t		stat_lock;
+};
+
+static void scrub_free_csums(struct scrub_dev *sdev)
+{
+	while (!list_empty(&sdev->csum_list)) {
+		struct btrfs_ordered_sum *sum;
+		sum = list_first_entry(&sdev->csum_list,
+				       struct btrfs_ordered_sum, list);
+		list_del(&sum->list);
+		kfree(sum);
+	}
+}
+
+static noinline_for_stack void scrub_free_dev(struct scrub_dev *sdev)
+{
+	int i;
+	int j;
+	struct page *last_page;
+
+	if (!sdev)
+		return;
+
+	for (i = 0; i < SCRUB_BIOS_PER_DEV; ++i) {
+		struct scrub_bio *sbio = sdev->bios[i];
+		struct bio *bio;
+
+		if (!sbio)
+			break;
+
+		bio = sbio->bio;
+		if (bio) {
+			last_page = NULL;
+			for (j = 0; j < bio->bi_vcnt; ++j) {
+				if (bio->bi_io_vec[j].bv_page == last_page)
+					continue;
+				last_page = bio->bi_io_vec[j].bv_page;
+				__free_page(last_page);
+			}
+			bio_put(bio);
+		}
+		kfree(sbio);
+	}
+
+	scrub_free_csums(sdev);
+	kfree(sdev);
+}
+
+static noinline_for_stack
+struct scrub_dev *scrub_setup_dev(struct btrfs_device *dev)
+{
+	struct scrub_dev *sdev;
+	int		i;
+	int		j;
+	int		ret;
+	struct btrfs_fs_info *fs_info = dev->dev_root->fs_info;
+
+	sdev = kzalloc(sizeof(*sdev), GFP_NOFS);
+	if (!sdev)
+		goto nomem;
+	sdev->dev = dev;
+	for (i = 0; i < SCRUB_BIOS_PER_DEV; ++i) {
+		struct bio *bio;
+		struct scrub_bio *sbio;
+
+		sbio = kzalloc(sizeof(*sbio), GFP_NOFS);
+		if (!sbio)
+			goto nomem;
+		sdev->bios[i] = sbio;
+
+		bio = bio_kmalloc(GFP_NOFS, SCRUB_PAGES_PER_BIO);
+		if (!bio)
+			goto nomem;
+
+		sbio->index = i;
+		sbio->sdev = sdev;
+		sbio->bio = bio;
+		sbio->count = 0;
+		sbio->work.func = scrub_checksum;
+		bio->bi_private = sdev->bios[i];
+		bio->bi_end_io = scrub_bio_end_io;
+		bio->bi_sector = 0;
+		bio->bi_bdev = dev->bdev;
+		bio->bi_size = 0;
+
+		for (j = 0; j < SCRUB_PAGES_PER_BIO; ++j) {
+			struct page *page;
+			page = alloc_page(GFP_NOFS);
+			if (!page)
+				goto nomem;
+
+			ret = bio_add_page(bio, page, PAGE_SIZE, 0);
+			if (!ret)
+				goto nomem;
+		}
+		WARN_ON(bio->bi_vcnt != SCRUB_PAGES_PER_BIO);
+
+		if (i != SCRUB_BIOS_PER_DEV-1)
+			sdev->bios[i]->next_free = i + 1;
+		 else
+			sdev->bios[i]->next_free = -1;
+	}
+	sdev->first_free = 0;
+	sdev->curr = -1;
+	atomic_set(&sdev->in_flight, 0);
+	atomic_set(&sdev->cancel_req, 0);
+	sdev->csum_size = btrfs_super_csum_size(&fs_info->super_copy);
+	INIT_LIST_HEAD(&sdev->csum_list);
+
+	spin_lock_init(&sdev->list_lock);
+	spin_lock_init(&sdev->stat_lock);
+	init_waitqueue_head(&sdev->list_wait);
+	return sdev;
+
+nomem:
+	scrub_free_dev(sdev);
+	return ERR_PTR(-ENOMEM);
+}
+
+/*
+ * scrub_recheck_error gets called when either verification of the page
+ * failed or the bio failed to read, e.g. with EIO. In the latter case,
+ * recheck_error gets called for every page in the bio, even though only
+ * one may be bad
+ */
+static void scrub_recheck_error(struct scrub_bio *sbio, int ix)
+{
+	if (sbio->err) {
+		if (scrub_fixup_io(READ, sbio->sdev->dev->bdev,
+				   (sbio->physical + ix * PAGE_SIZE) >> 9,
+				   sbio->bio->bi_io_vec[ix].bv_page) == 0) {
+			if (scrub_fixup_check(sbio, ix) == 0)
+				return;
+		}
+	}
+
+	scrub_fixup(sbio, ix);
+}
+
+static int scrub_fixup_check(struct scrub_bio *sbio, int ix)
+{
+	int ret = 1;
+	struct page *page;
+	void *buffer;
+	u64 flags = sbio->spag[ix].flags;
+
+	page = sbio->bio->bi_io_vec[ix].bv_page;
+	buffer = kmap_atomic(page, KM_USER0);
+	if (flags & BTRFS_EXTENT_FLAG_DATA) {
+		ret = scrub_checksum_data(sbio->sdev,
+					  sbio->spag + ix, buffer);
+	} else if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) {
+		ret = scrub_checksum_tree_block(sbio->sdev,
+						sbio->spag + ix,
+						sbio->logical + ix * PAGE_SIZE,
+						buffer);
+	} else {
+		WARN_ON(1);
+	}
+	kunmap_atomic(buffer, KM_USER0);
+
+	return ret;
+}
+
+static void scrub_fixup_end_io(struct bio *bio, int err)
+{
+	complete((struct completion *)bio->bi_private);
+}
+
+static void scrub_fixup(struct scrub_bio *sbio, int ix)
+{
+	struct scrub_dev *sdev = sbio->sdev;
+	struct btrfs_fs_info *fs_info = sdev->dev->dev_root->fs_info;
+	struct btrfs_mapping_tree *map_tree = &fs_info->mapping_tree;
+	struct btrfs_multi_bio *multi = NULL;
+	u64 logical = sbio->logical + ix * PAGE_SIZE;
+	u64 length;
+	int i;
+	int ret;
+	DECLARE_COMPLETION_ONSTACK(complete);
+
+	if ((sbio->spag[ix].flags & BTRFS_EXTENT_FLAG_DATA) &&
+	    (sbio->spag[ix].have_csum == 0)) {
+		/*
+		 * nodatasum, don't try to fix anything
+		 * FIXME: we can do better, open the inode and trigger a
+		 * writeback
+		 */
+		goto uncorrectable;
+	}
+
+	length = PAGE_SIZE;
+	ret = btrfs_map_block(map_tree, REQ_WRITE, logical, &length,
+			      &multi, 0);
+	if (ret || !multi || length < PAGE_SIZE) {
+		printk(KERN_ERR
+		       "scrub_fixup: btrfs_map_block failed us for %llu\n",
+		       (unsigned long long)logical);
+		WARN_ON(1);
+		return;
+	}
+
+	if (multi->num_stripes == 1)
+		/* there aren't any replicas */
+		goto uncorrectable;
+
+	/*
+	 * first find a good copy
+	 */
+	for (i = 0; i < multi->num_stripes; ++i) {
+		if (i == sbio->spag[ix].mirror_num)
+			continue;
+
+		if (scrub_fixup_io(READ, multi->stripes[i].dev->bdev,
+				   multi->stripes[i].physical >> 9,
+				   sbio->bio->bi_io_vec[ix].bv_page)) {
+			/* I/O-error, this is not a good copy */
+			continue;
+		}
+
+		if (scrub_fixup_check(sbio, ix) == 0)
+			break;
+	}
+	if (i == multi->num_stripes)
+		goto uncorrectable;
+
+	if (!sdev->readonly) {
+		/*
+		 * bi_io_vec[ix].bv_page now contains good data, write it back
+		 */
+		if (scrub_fixup_io(WRITE, sdev->dev->bdev,
+				   (sbio->physical + ix * PAGE_SIZE) >> 9,
+				   sbio->bio->bi_io_vec[ix].bv_page)) {
+			/* I/O-error, writeback failed, give up */
+			goto uncorrectable;
+		}
+	}
+
+	kfree(multi);
+	spin_lock(&sdev->stat_lock);
+	++sdev->stat.corrected_errors;
+	spin_unlock(&sdev->stat_lock);
+
+	if (printk_ratelimit())
+		printk(KERN_ERR "btrfs: fixed up at %llu\n",
+		       (unsigned long long)logical);
+	return;
+
+uncorrectable:
+	kfree(multi);
+	spin_lock(&sdev->stat_lock);
+	++sdev->stat.uncorrectable_errors;
+	spin_unlock(&sdev->stat_lock);
+
+	if (printk_ratelimit())
+		printk(KERN_ERR "btrfs: unable to fixup at %llu\n",
+			 (unsigned long long)logical);
+}
+
+static int scrub_fixup_io(int rw, struct block_device *bdev, sector_t sector,
+			 struct page *page)
+{
+	struct bio *bio = NULL;
+	int ret;
+	DECLARE_COMPLETION_ONSTACK(complete);
+
+	/* we are going to wait on this IO */
+	rw |= REQ_SYNC;
+
+	bio = bio_alloc(GFP_NOFS, 1);
+	bio->bi_bdev = bdev;
+	bio->bi_sector = sector;
+	bio_add_page(bio, page, PAGE_SIZE, 0);
+	bio->bi_end_io = scrub_fixup_end_io;
+	bio->bi_private = &complete;
+	submit_bio(rw, bio);
+
+	wait_for_completion(&complete);
+
+	ret = !test_bit(BIO_UPTODATE, &bio->bi_flags);
+	bio_put(bio);
+	return ret;
+}
+
+static void scrub_bio_end_io(struct bio *bio, int err)
+{
+	struct scrub_bio *sbio = bio->bi_private;
+	struct scrub_dev *sdev = sbio->sdev;
+	struct btrfs_fs_info *fs_info = sdev->dev->dev_root->fs_info;
+
+	sbio->err = err;
+
+	btrfs_queue_worker(&fs_info->scrub_workers, &sbio->work);
+}
+
+static void scrub_checksum(struct btrfs_work *work)
+{
+	struct scrub_bio *sbio = container_of(work, struct scrub_bio, work);
+	struct scrub_dev *sdev = sbio->sdev;
+	struct page *page;
+	void *buffer;
+	int i;
+	u64 flags;
+	u64 logical;
+	int ret;
+
+	if (sbio->err) {
+		for (i = 0; i < sbio->count; ++i)
+			scrub_recheck_error(sbio, i);
+
+		sbio->bio->bi_flags &= ~(BIO_POOL_MASK - 1);
+		sbio->bio->bi_flags |= 1 << BIO_UPTODATE;
+		sbio->bio->bi_phys_segments = 0;
+		sbio->bio->bi_idx = 0;
+
+		for (i = 0; i < sbio->count; i++) {
+			struct bio_vec *bi;
+			bi = &sbio->bio->bi_io_vec[i];
+			bi->bv_offset = 0;
+			bi->bv_len = PAGE_SIZE;
+		}
+
+		spin_lock(&sdev->stat_lock);
+		++sdev->stat.read_errors;
+		spin_unlock(&sdev->stat_lock);
+		goto out;
+	}
+	for (i = 0; i < sbio->count; ++i) {
+		page = sbio->bio->bi_io_vec[i].bv_page;
+		buffer = kmap_atomic(page, KM_USER0);
+		flags = sbio->spag[i].flags;
+		logical = sbio->logical + i * PAGE_SIZE;
+		ret = 0;
+		if (flags & BTRFS_EXTENT_FLAG_DATA) {
+			ret = scrub_checksum_data(sdev, sbio->spag + i, buffer);
+		} else if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) {
+			ret = scrub_checksum_tree_block(sdev, sbio->spag + i,
+							logical, buffer);
+		} else if (flags & BTRFS_EXTENT_FLAG_SUPER) {
+			BUG_ON(i);
+			(void)scrub_checksum_super(sbio, buffer);
+		} else {
+			WARN_ON(1);
+		}
+		kunmap_atomic(buffer, KM_USER0);
+		if (ret)
+			scrub_recheck_error(sbio, i);
+	}
+
+out:
+	spin_lock(&sdev->list_lock);
+	sbio->next_free = sdev->first_free;
+	sdev->first_free = sbio->index;
+	spin_unlock(&sdev->list_lock);
+	atomic_dec(&sdev->in_flight);
+	wake_up(&sdev->list_wait);
+}
+
+static int scrub_checksum_data(struct scrub_dev *sdev,
+			       struct scrub_page *spag, void *buffer)
+{
+	u8 csum[BTRFS_CSUM_SIZE];
+	u32 crc = ~(u32)0;
+	int fail = 0;
+	struct btrfs_root *root = sdev->dev->dev_root;
+
+	if (!spag->have_csum)
+		return 0;
+
+	crc = btrfs_csum_data(root, buffer, crc, PAGE_SIZE);
+	btrfs_csum_final(crc, csum);
+	if (memcmp(csum, spag->csum, sdev->csum_size))
+		fail = 1;
+
+	spin_lock(&sdev->stat_lock);
+	++sdev->stat.data_extents_scrubbed;
+	sdev->stat.data_bytes_scrubbed += PAGE_SIZE;
+	if (fail)
+		++sdev->stat.csum_errors;
+	spin_unlock(&sdev->stat_lock);
+
+	return fail;
+}
+
+static int scrub_checksum_tree_block(struct scrub_dev *sdev,
+				     struct scrub_page *spag, u64 logical,
+				     void *buffer)
+{
+	struct btrfs_header *h;
+	struct btrfs_root *root = sdev->dev->dev_root;
+	struct btrfs_fs_info *fs_info = root->fs_info;
+	u8 csum[BTRFS_CSUM_SIZE];
+	u32 crc = ~(u32)0;
+	int fail = 0;
+	int crc_fail = 0;
+
+	/*
+	 * we don't use the getter functions here, as we
+	 * a) don't have an extent buffer and
+	 * b) the page is already kmapped
+	 */
+	h = (struct btrfs_header *)buffer;
+
+	if (logical != le64_to_cpu(h->bytenr))
+		++fail;
+
+	if (spag->generation != le64_to_cpu(h->generation))
+		++fail;
+
+	if (memcmp(h->fsid, fs_info->fsid, BTRFS_UUID_SIZE))
+		++fail;
+
+	if (memcmp(h->chunk_tree_uuid, fs_info->chunk_tree_uuid,
+		   BTRFS_UUID_SIZE))
+		++fail;
+
+	crc = btrfs_csum_data(root, buffer + BTRFS_CSUM_SIZE, crc,
+			      PAGE_SIZE - BTRFS_CSUM_SIZE);
+	btrfs_csum_final(crc, csum);
+	if (memcmp(csum, h->csum, sdev->csum_size))
+		++crc_fail;
+
+	spin_lock(&sdev->stat_lock);
+	++sdev->stat.tree_extents_scrubbed;
+	sdev->stat.tree_bytes_scrubbed += PAGE_SIZE;
+	if (crc_fail)
+		++sdev->stat.csum_errors;
+	if (fail)
+		++sdev->stat.verify_errors;
+	spin_unlock(&sdev->stat_lock);
+
+	return fail || crc_fail;
+}
+
+static int scrub_checksum_super(struct scrub_bio *sbio, void *buffer)
+{
+	struct btrfs_super_block *s;
+	u64 logical;
+	struct scrub_dev *sdev = sbio->sdev;
+	struct btrfs_root *root = sdev->dev->dev_root;
+	struct btrfs_fs_info *fs_info = root->fs_info;
+	u8 csum[BTRFS_CSUM_SIZE];
+	u32 crc = ~(u32)0;
+	int fail = 0;
+
+	s = (struct btrfs_super_block *)buffer;
+	logical = sbio->logical;
+
+	if (logical != le64_to_cpu(s->bytenr))
+		++fail;
+
+	if (sbio->spag[0].generation != le64_to_cpu(s->generation))
+		++fail;
+
+	if (memcmp(s->fsid, fs_info->fsid, BTRFS_UUID_SIZE))
+		++fail;
+
+	crc = btrfs_csum_data(root, buffer + BTRFS_CSUM_SIZE, crc,
+			      PAGE_SIZE - BTRFS_CSUM_SIZE);
+	btrfs_csum_final(crc, csum);
+	if (memcmp(csum, s->csum, sbio->sdev->csum_size))
+		++fail;
+
+	if (fail) {
+		/*
+		 * if we find an error in a super block, we just report it.
+		 * They will get written with the next transaction commit
+		 * anyway
+		 */
+		spin_lock(&sdev->stat_lock);
+		++sdev->stat.super_errors;
+		spin_unlock(&sdev->stat_lock);
+	}
+
+	return fail;
+}
+
+static int scrub_submit(struct scrub_dev *sdev)
+{
+	struct scrub_bio *sbio;
+
+	if (sdev->curr == -1)
+		return 0;
+
+	sbio = sdev->bios[sdev->curr];
+
+	sbio->bio->bi_sector = sbio->physical >> 9;
+	sbio->bio->bi_size = sbio->count * PAGE_SIZE;
+	sbio->bio->bi_next = NULL;
+	sbio->bio->bi_flags |= 1 << BIO_UPTODATE;
+	sbio->bio->bi_comp_cpu = -1;
+	sbio->bio->bi_bdev = sdev->dev->bdev;
+	sbio->err = 0;
+	sdev->curr = -1;
+	atomic_inc(&sdev->in_flight);
+
+	submit_bio(0, sbio->bio);
+
+	return 0;
+}
+
+static int scrub_page(struct scrub_dev *sdev, u64 logical, u64 len,
+		      u64 physical, u64 flags, u64 gen, u64 mirror_num,
+		      u8 *csum, int force)
+{
+	struct scrub_bio *sbio;
+
+again:
+	/*
+	 * grab a fresh bio or wait for one to become available
+	 */
+	while (sdev->curr == -1) {
+		spin_lock(&sdev->list_lock);
+		sdev->curr = sdev->first_free;
+		if (sdev->curr != -1) {
+			sdev->first_free = sdev->bios[sdev->curr]->next_free;
+			sdev->bios[sdev->curr]->next_free = -1;
+			sdev->bios[sdev->curr]->count = 0;
+			spin_unlock(&sdev->list_lock);
+		} else {
+			spin_unlock(&sdev->list_lock);
+			wait_event(sdev->list_wait, sdev->first_free != -1);
+		}
+	}
+	sbio = sdev->bios[sdev->curr];
+	if (sbio->count == 0) {
+		sbio->physical = physical;
+		sbio->logical = logical;
+	} else if (sbio->physical + sbio->count * PAGE_SIZE != physical ||
+		   sbio->logical + sbio->count * PAGE_SIZE != logical) {
+		scrub_submit(sdev);
+		goto again;
+	}
+	sbio->spag[sbio->count].flags = flags;
+	sbio->spag[sbio->count].generation = gen;
+	sbio->spag[sbio->count].have_csum = 0;
+	sbio->spag[sbio->count].mirror_num = mirror_num;
+	if (csum) {
+		sbio->spag[sbio->count].have_csum = 1;
+		memcpy(sbio->spag[sbio->count].csum, csum, sdev->csum_size);
+	}
+	++sbio->count;
+	if (sbio->count == SCRUB_PAGES_PER_BIO || force)
+		scrub_submit(sdev);
+
+	return 0;
+}
+
+static int scrub_find_csum(struct scrub_dev *sdev, u64 logical, u64 len,
+			   u8 *csum)
+{
+	struct btrfs_ordered_sum *sum = NULL;
+	int ret = 0;
+	unsigned long i;
+	unsigned long num_sectors;
+	u32 sectorsize = sdev->dev->dev_root->sectorsize;
+
+	while (!list_empty(&sdev->csum_list)) {
+		sum = list_first_entry(&sdev->csum_list,
+				       struct btrfs_ordered_sum, list);
+		if (sum->bytenr > logical)
+			return 0;
+		if (sum->bytenr + sum->len > logical)
+			break;
+
+		++sdev->stat.csum_discards;
+		list_del(&sum->list);
+		kfree(sum);
+		sum = NULL;
+	}
+	if (!sum)
+		return 0;
+
+	num_sectors = sum->len / sectorsize;
+	for (i = 0; i < num_sectors; ++i) {
+		if (sum->sums[i].bytenr == logical) {
+			memcpy(csum, &sum->sums[i].sum, sdev->csum_size);
+			ret = 1;
+			break;
+		}
+	}
+	if (ret && i == num_sectors - 1) {
+		list_del(&sum->list);
+		kfree(sum);
+	}
+	return ret;
+}
+
+/* scrub extent tries to collect up to 64 kB for each bio */
+static int scrub_extent(struct scrub_dev *sdev, u64 logical, u64 len,
+			u64 physical, u64 flags, u64 gen, u64 mirror_num)
+{
+	int ret;
+	u8 csum[BTRFS_CSUM_SIZE];
+
+	while (len) {
+		u64 l = min_t(u64, len, PAGE_SIZE);
+		int have_csum = 0;
+
+		if (flags & BTRFS_EXTENT_FLAG_DATA) {
+			/* push csums to sbio */
+			have_csum = scrub_find_csum(sdev, logical, l, csum);
+			if (have_csum == 0)
+				++sdev->stat.no_csum;
+		}
+		ret = scrub_page(sdev, logical, l, physical, flags, gen,
+				 mirror_num, have_csum ? csum : NULL, 0);
+		if (ret)
+			return ret;
+		len -= l;
+		logical += l;
+		physical += l;
+	}
+	return 0;
+}
+
+static noinline_for_stack int scrub_stripe(struct scrub_dev *sdev,
+	struct map_lookup *map, int num, u64 base, u64 length)
+{
+	struct btrfs_path *path;
+	struct btrfs_fs_info *fs_info = sdev->dev->dev_root->fs_info;
+	struct btrfs_root *root = fs_info->extent_root;
+	struct btrfs_root *csum_root = fs_info->csum_root;
+	struct btrfs_extent_item *extent;
+	u64 flags;
+	int ret;
+	int slot;
+	int i;
+	u64 nstripes;
+	int start_stripe;
+	struct extent_buffer *l;
+	struct btrfs_key key;
+	u64 physical;
+	u64 logical;
+	u64 generation;
+	u64 mirror_num;
+
+	u64 increment = map->stripe_len;
+	u64 offset;
+
+	nstripes = length;
+	offset = 0;
+	do_div(nstripes, map->stripe_len);
+	if (map->type & BTRFS_BLOCK_GROUP_RAID0) {
+		offset = map->stripe_len * num;
+		increment = map->stripe_len * map->num_stripes;
+		mirror_num = 0;
+	} else if (map->type & BTRFS_BLOCK_GROUP_RAID10) {
+		int factor = map->num_stripes / map->sub_stripes;
+		offset = map->stripe_len * (num / map->sub_stripes);
+		increment = map->stripe_len * factor;
+		mirror_num = num % map->sub_stripes;
+	} else if (map->type & BTRFS_BLOCK_GROUP_RAID1) {
+		increment = map->stripe_len;
+		mirror_num = num % map->num_stripes;
+	} else if (map->type & BTRFS_BLOCK_GROUP_DUP) {
+		increment = map->stripe_len;
+		mirror_num = num % map->num_stripes;
+	} else {
+		increment = map->stripe_len;
+		mirror_num = 0;
+	}
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	path->reada = 2;
+	path->search_commit_root = 1;
+	path->skip_locking = 1;
+
+	/*
+	 * find all extents for each stripe and just read them to get
+	 * them into the page cache
+	 * FIXME: we can do better. build a more intelligent prefetching
+	 */
+	logical = base + offset;
+	physical = map->stripes[num].physical;
+	ret = 0;
+	for (i = 0; i < nstripes; ++i) {
+		key.objectid = logical;
+		key.type = BTRFS_EXTENT_ITEM_KEY;
+		key.offset = (u64)0;
+
+		ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+		if (ret < 0)
+			goto out;
+
+		l = path->nodes[0];
+		slot = path->slots[0];
+		btrfs_item_key_to_cpu(l, &key, slot);
+		if (key.objectid != logical) {
+			ret = btrfs_previous_item(root, path, 0,
+						  BTRFS_EXTENT_ITEM_KEY);
+			if (ret < 0)
+				goto out;
+		}
+
+		while (1) {
+			l = path->nodes[0];
+			slot = path->slots[0];
+			if (slot >= btrfs_header_nritems(l)) {
+				ret = btrfs_next_leaf(root, path);
+				if (ret == 0)
+					continue;
+				if (ret < 0)
+					goto out;
+
+				break;
+			}
+			btrfs_item_key_to_cpu(l, &key, slot);
+
+			if (key.objectid >= logical + map->stripe_len)
+				break;
+
+			path->slots[0]++;
+		}
+		btrfs_release_path(path);
+		logical += increment;
+		physical += map->stripe_len;
+		cond_resched();
+	}
+
+	/*
+	 * collect all data csums for the stripe to avoid seeking during
+	 * the scrub. This might currently (crc32) end up to be about 1MB
+	 */
+	start_stripe = 0;
+again:
+	logical = base + offset + start_stripe * increment;
+	for (i = start_stripe; i < nstripes; ++i) {
+		ret = btrfs_lookup_csums_range(csum_root, logical,
+					       logical + map->stripe_len - 1,
+					       &sdev->csum_list, 1);
+		if (ret)
+			goto out;
+
+		logical += increment;
+		cond_resched();
+	}
+	/*
+	 * now find all extents for each stripe and scrub them
+	 */
+	logical = base + offset + start_stripe * increment;
+	physical = map->stripes[num].physical + start_stripe * map->stripe_len;
+	ret = 0;
+	for (i = start_stripe; i < nstripes; ++i) {
+		/*
+		 * canceled?
+		 */
+		if (atomic_read(&fs_info->scrub_cancel_req) ||
+		    atomic_read(&sdev->cancel_req)) {
+			ret = -ECANCELED;
+			goto out;
+		}
+		/*
+		 * check to see if we have to pause
+		 */
+		if (atomic_read(&fs_info->scrub_pause_req)) {
+			/* push queued extents */
+			scrub_submit(sdev);
+			wait_event(sdev->list_wait,
+				   atomic_read(&sdev->in_flight) == 0);
+			atomic_inc(&fs_info->scrubs_paused);
+			wake_up(&fs_info->scrub_pause_wait);
+			mutex_lock(&fs_info->scrub_lock);
+			while (atomic_read(&fs_info->scrub_pause_req)) {
+				mutex_unlock(&fs_info->scrub_lock);
+				wait_event(fs_info->scrub_pause_wait,
+				   atomic_read(&fs_info->scrub_pause_req) == 0);
+				mutex_lock(&fs_info->scrub_lock);
+			}
+			atomic_dec(&fs_info->scrubs_paused);
+			mutex_unlock(&fs_info->scrub_lock);
+			wake_up(&fs_info->scrub_pause_wait);
+			scrub_free_csums(sdev);
+			start_stripe = i;
+			goto again;
+		}
+
+		key.objectid = logical;
+		key.type = BTRFS_EXTENT_ITEM_KEY;
+		key.offset = (u64)0;
+
+		ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+		if (ret < 0)
+			goto out;
+
+		l = path->nodes[0];
+		slot = path->slots[0];
+		btrfs_item_key_to_cpu(l, &key, slot);
+		if (key.objectid != logical) {
+			ret = btrfs_previous_item(root, path, 0,
+						  BTRFS_EXTENT_ITEM_KEY);
+			if (ret < 0)
+				goto out;
+		}
+
+		while (1) {
+			l = path->nodes[0];
+			slot = path->slots[0];
+			if (slot >= btrfs_header_nritems(l)) {
+				ret = btrfs_next_leaf(root, path);
+				if (ret == 0)
+					continue;
+				if (ret < 0)
+					goto out;
+
+				break;
+			}
+			btrfs_item_key_to_cpu(l, &key, slot);
+
+			if (key.objectid + key.offset <= logical)
+				goto next;
+
+			if (key.objectid >= logical + map->stripe_len)
+				break;
+
+			if (btrfs_key_type(&key) != BTRFS_EXTENT_ITEM_KEY)
+				goto next;
+
+			extent = btrfs_item_ptr(l, slot,
+						struct btrfs_extent_item);
+			flags = btrfs_extent_flags(l, extent);
+			generation = btrfs_extent_generation(l, extent);
+
+			if (key.objectid < logical &&
+			    (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK)) {
+				printk(KERN_ERR
+				       "btrfs scrub: tree block %llu spanning "
+				       "stripes, ignored. logical=%llu\n",
+				       (unsigned long long)key.objectid,
+				       (unsigned long long)logical);
+				goto next;
+			}
+
+			/*
+			 * trim extent to this stripe
+			 */
+			if (key.objectid < logical) {
+				key.offset -= logical - key.objectid;
+				key.objectid = logical;
+			}
+			if (key.objectid + key.offset >
+			    logical + map->stripe_len) {
+				key.offset = logical + map->stripe_len -
+					     key.objectid;
+			}
+
+			ret = scrub_extent(sdev, key.objectid, key.offset,
+					   key.objectid - logical + physical,
+					   flags, generation, mirror_num);
+			if (ret)
+				goto out;
+
+next:
+			path->slots[0]++;
+		}
+		btrfs_release_path(path);
+		logical += increment;
+		physical += map->stripe_len;
+		spin_lock(&sdev->stat_lock);
+		sdev->stat.last_physical = physical;
+		spin_unlock(&sdev->stat_lock);
+	}
+	/* push queued extents */
+	scrub_submit(sdev);
+
+out:
+	btrfs_free_path(path);
+	return ret < 0 ? ret : 0;
+}
+
+static noinline_for_stack int scrub_chunk(struct scrub_dev *sdev,
+	u64 chunk_tree, u64 chunk_objectid, u64 chunk_offset, u64 length)
+{
+	struct btrfs_mapping_tree *map_tree =
+		&sdev->dev->dev_root->fs_info->mapping_tree;
+	struct map_lookup *map;
+	struct extent_map *em;
+	int i;
+	int ret = -EINVAL;
+
+	read_lock(&map_tree->map_tree.lock);
+	em = lookup_extent_mapping(&map_tree->map_tree, chunk_offset, 1);
+	read_unlock(&map_tree->map_tree.lock);
+
+	if (!em)
+		return -EINVAL;
+
+	map = (struct map_lookup *)em->bdev;
+	if (em->start != chunk_offset)
+		goto out;
+
+	if (em->len < length)
+		goto out;
+
+	for (i = 0; i < map->num_stripes; ++i) {
+		if (map->stripes[i].dev == sdev->dev) {
+			ret = scrub_stripe(sdev, map, i, chunk_offset, length);
+			if (ret)
+				goto out;
+		}
+	}
+out:
+	free_extent_map(em);
+
+	return ret;
+}
+
+static noinline_for_stack
+int scrub_enumerate_chunks(struct scrub_dev *sdev, u64 start, u64 end)
+{
+	struct btrfs_dev_extent *dev_extent = NULL;
+	struct btrfs_path *path;
+	struct btrfs_root *root = sdev->dev->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 slot;
+	struct extent_buffer *l;
+	struct btrfs_key key;
+	struct btrfs_key found_key;
+	struct btrfs_block_group_cache *cache;
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
+	path->reada = 2;
+	path->search_commit_root = 1;
+	path->skip_locking = 1;
+
+	key.objectid = sdev->dev->devid;
+	key.offset = 0ull;
+	key.type = BTRFS_DEV_EXTENT_KEY;
+
+
+	while (1) {
+		ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+		if (ret < 0)
+			goto out;
+		ret = 0;
+
+		l = path->nodes[0];
+		slot = path->slots[0];
+
+		btrfs_item_key_to_cpu(l, &found_key, slot);
+
+		if (found_key.objectid != sdev->dev->devid)
+			break;
+
+		if (btrfs_key_type(&key) != BTRFS_DEV_EXTENT_KEY)
+			break;
+
+		if (found_key.offset >= end)
+			break;
+
+		if (found_key.offset < key.offset)
+			break;
+
+		dev_extent = btrfs_item_ptr(l, slot, struct btrfs_dev_extent);
+		length = btrfs_dev_extent_length(l, dev_extent);
+
+		if (found_key.offset + length <= start) {
+			key.offset = found_key.offset + length;
+			btrfs_release_path(path);
+			continue;
+		}
+
+		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);
+
+		/*
+		 * get a reference on the corresponding block group to prevent
+		 * the chunk from going away while we scrub it
+		 */
+		cache = btrfs_lookup_block_group(fs_info, chunk_offset);
+		if (!cache) {
+			ret = -ENOENT;
+			goto out;
+		}
+		ret = scrub_chunk(sdev, chunk_tree, chunk_objectid,
+				  chunk_offset, length);
+		btrfs_put_block_group(cache);
+		if (ret)
+			break;
+
+		key.offset = found_key.offset + length;
+		btrfs_release_path(path);
+	}
+
+out:
+	btrfs_free_path(path);
+	return ret;
+}
+
+static noinline_for_stack int scrub_supers(struct scrub_dev *sdev)
+{
+	int	i;
+	u64	bytenr;
+	u64	gen;
+	int	ret;
+	struct btrfs_device *device = sdev->dev;
+	struct btrfs_root *root = device->dev_root;
+
+	gen = root->fs_info->last_trans_committed;
+
+	for (i = 0; i < BTRFS_SUPER_MIRROR_MAX; i++) {
+		bytenr = btrfs_sb_offset(i);
+		if (bytenr + BTRFS_SUPER_INFO_SIZE >= device->total_bytes)
+			break;
+
+		ret = scrub_page(sdev, bytenr, PAGE_SIZE, bytenr,
+				 BTRFS_EXTENT_FLAG_SUPER, gen, i, NULL, 1);
+		if (ret)
+			return ret;
+	}
+	wait_event(sdev->list_wait, atomic_read(&sdev->in_flight) == 0);
+
+	return 0;
+}
+
+/*
+ * get a reference count on fs_info->scrub_workers. start worker if necessary
+ */
+static noinline_for_stack int scrub_workers_get(struct btrfs_root *root)
+{
+	struct btrfs_fs_info *fs_info = root->fs_info;
+
+	mutex_lock(&fs_info->scrub_lock);
+	if (fs_info->scrub_workers_refcnt == 0)
+		btrfs_start_workers(&fs_info->scrub_workers, 1);
+	++fs_info->scrub_workers_refcnt;
+	mutex_unlock(&fs_info->scrub_lock);
+
+	return 0;
+}
+
+static noinline_for_stack void scrub_workers_put(struct btrfs_root *root)
+{
+	struct btrfs_fs_info *fs_info = root->fs_info;
+
+	mutex_lock(&fs_info->scrub_lock);
+	if (--fs_info->scrub_workers_refcnt == 0)
+		btrfs_stop_workers(&fs_info->scrub_workers);
+	WARN_ON(fs_info->scrub_workers_refcnt < 0);
+	mutex_unlock(&fs_info->scrub_lock);
+}
+
+
+int btrfs_scrub_dev(struct btrfs_root *root, u64 devid, u64 start, u64 end,
+		    struct btrfs_scrub_progress *progress, int readonly)
+{
+	struct scrub_dev *sdev;
+	struct btrfs_fs_info *fs_info = root->fs_info;
+	int ret;
+	struct btrfs_device *dev;
+
+	if (root->fs_info->closing)
+		return -EINVAL;
+
+	/*
+	 * check some assumptions
+	 */
+	if (root->sectorsize != PAGE_SIZE ||
+	    root->sectorsize != root->leafsize ||
+	    root->sectorsize != root->nodesize) {
+		printk(KERN_ERR "btrfs_scrub: size assumptions fail\n");
+		return -EINVAL;
+	}
+
+	ret = scrub_workers_get(root);
+	if (ret)
+		return ret;
+
+	mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
+	dev = btrfs_find_device(root, devid, NULL, NULL);
+	if (!dev || dev->missing) {
+		mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
+		scrub_workers_put(root);
+		return -ENODEV;
+	}
+	mutex_lock(&fs_info->scrub_lock);
+
+	if (!dev->in_fs_metadata) {
+		mutex_unlock(&fs_info->scrub_lock);
+		mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
+		scrub_workers_put(root);
+		return -ENODEV;
+	}
+
+	if (dev->scrub_device) {
+		mutex_unlock(&fs_info->scrub_lock);
+		mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
+		scrub_workers_put(root);
+		return -EINPROGRESS;
+	}
+	sdev = scrub_setup_dev(dev);
+	if (IS_ERR(sdev)) {
+		mutex_unlock(&fs_info->scrub_lock);
+		mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
+		scrub_workers_put(root);
+		return PTR_ERR(sdev);
+	}
+	sdev->readonly = readonly;
+	dev->scrub_device = sdev;
+
+	atomic_inc(&fs_info->scrubs_running);
+	mutex_unlock(&fs_info->scrub_lock);
+	mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
+
+	down_read(&fs_info->scrub_super_lock);
+	ret = scrub_supers(sdev);
+	up_read(&fs_info->scrub_super_lock);
+
+	if (!ret)
+		ret = scrub_enumerate_chunks(sdev, start, end);
+
+	wait_event(sdev->list_wait, atomic_read(&sdev->in_flight) == 0);
+
+	atomic_dec(&fs_info->scrubs_running);
+	wake_up(&fs_info->scrub_pause_wait);
+
+	if (progress)
+		memcpy(progress, &sdev->stat, sizeof(*progress));
+
+	mutex_lock(&fs_info->scrub_lock);
+	dev->scrub_device = NULL;
+	mutex_unlock(&fs_info->scrub_lock);
+
+	scrub_free_dev(sdev);
+	scrub_workers_put(root);
+
+	return ret;
+}
+
+int btrfs_scrub_pause(struct btrfs_root *root)
+{
+	struct btrfs_fs_info *fs_info = root->fs_info;
+
+	mutex_lock(&fs_info->scrub_lock);
+	atomic_inc(&fs_info->scrub_pause_req);
+	while (atomic_read(&fs_info->scrubs_paused) !=
+	       atomic_read(&fs_info->scrubs_running)) {
+		mutex_unlock(&fs_info->scrub_lock);
+		wait_event(fs_info->scrub_pause_wait,
+			   atomic_read(&fs_info->scrubs_paused) ==
+			   atomic_read(&fs_info->scrubs_running));
+		mutex_lock(&fs_info->scrub_lock);
+	}
+	mutex_unlock(&fs_info->scrub_lock);
+
+	return 0;
+}
+
+int btrfs_scrub_continue(struct btrfs_root *root)
+{
+	struct btrfs_fs_info *fs_info = root->fs_info;
+
+	atomic_dec(&fs_info->scrub_pause_req);
+	wake_up(&fs_info->scrub_pause_wait);
+	return 0;
+}
+
+int btrfs_scrub_pause_super(struct btrfs_root *root)
+{
+	down_write(&root->fs_info->scrub_super_lock);
+	return 0;
+}
+
+int btrfs_scrub_continue_super(struct btrfs_root *root)
+{
+	up_write(&root->fs_info->scrub_super_lock);
+	return 0;
+}
+
+int btrfs_scrub_cancel(struct btrfs_root *root)
+{
+	struct btrfs_fs_info *fs_info = root->fs_info;
+
+	mutex_lock(&fs_info->scrub_lock);
+	if (!atomic_read(&fs_info->scrubs_running)) {
+		mutex_unlock(&fs_info->scrub_lock);
+		return -ENOTCONN;
+	}
+
+	atomic_inc(&fs_info->scrub_cancel_req);
+	while (atomic_read(&fs_info->scrubs_running)) {
+		mutex_unlock(&fs_info->scrub_lock);
+		wait_event(fs_info->scrub_pause_wait,
+			   atomic_read(&fs_info->scrubs_running) == 0);
+		mutex_lock(&fs_info->scrub_lock);
+	}
+	atomic_dec(&fs_info->scrub_cancel_req);
+	mutex_unlock(&fs_info->scrub_lock);
+
+	return 0;
+}
+
+int btrfs_scrub_cancel_dev(struct btrfs_root *root, struct btrfs_device *dev)
+{
+	struct btrfs_fs_info *fs_info = root->fs_info;
+	struct scrub_dev *sdev;
+
+	mutex_lock(&fs_info->scrub_lock);
+	sdev = dev->scrub_device;
+	if (!sdev) {
+		mutex_unlock(&fs_info->scrub_lock);
+		return -ENOTCONN;
+	}
+	atomic_inc(&sdev->cancel_req);
+	while (dev->scrub_device) {
+		mutex_unlock(&fs_info->scrub_lock);
+		wait_event(fs_info->scrub_pause_wait,
+			   dev->scrub_device == NULL);
+		mutex_lock(&fs_info->scrub_lock);
+	}
+	mutex_unlock(&fs_info->scrub_lock);
+
+	return 0;
+}
+int btrfs_scrub_cancel_devid(struct btrfs_root *root, u64 devid)
+{
+	struct btrfs_fs_info *fs_info = root->fs_info;
+	struct btrfs_device *dev;
+	int ret;
+
+	/*
+	 * we have to hold the device_list_mutex here so the device
+	 * does not go away in cancel_dev. FIXME: find a better solution
+	 */
+	mutex_lock(&fs_info->fs_devices->device_list_mutex);
+	dev = btrfs_find_device(root, devid, NULL, NULL);
+	if (!dev) {
+		mutex_unlock(&fs_info->fs_devices->device_list_mutex);
+		return -ENODEV;
+	}
+	ret = btrfs_scrub_cancel_dev(root, dev);
+	mutex_unlock(&fs_info->fs_devices->device_list_mutex);
+
+	return ret;
+}
+
+int btrfs_scrub_progress(struct btrfs_root *root, u64 devid,
+			 struct btrfs_scrub_progress *progress)
+{
+	struct btrfs_device *dev;
+	struct scrub_dev *sdev = NULL;
+
+	mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
+	dev = btrfs_find_device(root, devid, NULL, NULL);
+	if (dev)
+		sdev = dev->scrub_device;
+	if (sdev)
+		memcpy(progress, &sdev->stat, sizeof(*progress));
+	mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
+
+	return dev ? (sdev ? 0 : -ENOTCONN) : -ENODEV;
+}

fs/btrfs/super.c

@@ -41,6 +41,7 @@
 #include <linux/slab.h>
 #include <linux/cleancache.h>
 #include "compat.h"
+#include "delayed-inode.h"
 #include "ctree.h"
 #include "disk-io.h"
 #include "transaction.h"
@@ -160,7 +161,7 @@ enum {
 	Opt_compress_type, Opt_compress_force, Opt_compress_force_type,
 	Opt_notreelog, Opt_ratio, Opt_flushoncommit, Opt_discard,
 	Opt_space_cache, Opt_clear_cache, Opt_user_subvol_rm_allowed,
-	Opt_enospc_debug, Opt_subvolrootid, Opt_err,
+	Opt_enospc_debug, Opt_subvolrootid, Opt_defrag, Opt_err,
 };
 
 static match_table_t tokens = {
@@ -191,6 +192,7 @@ static match_table_t tokens = {
 	{Opt_user_subvol_rm_allowed, "user_subvol_rm_allowed"},
 	{Opt_enospc_debug, "enospc_debug"},
 	{Opt_subvolrootid, "subvolrootid=%d"},
+	{Opt_defrag, "autodefrag"},
 	{Opt_err, NULL},
 };
 
@@ -369,6 +371,10 @@ int btrfs_parse_options(struct btrfs_root *root, char *options)
 		case Opt_enospc_debug:
 			btrfs_set_opt(info->mount_opt, ENOSPC_DEBUG);
 			break;
+		case Opt_defrag:
+			printk(KERN_INFO "btrfs: enabling auto defrag");
+			btrfs_set_opt(info->mount_opt, AUTO_DEFRAG);
+			break;
 		case Opt_err:
 			printk(KERN_INFO "btrfs: unrecognized mount option "
 			       "'%s'\n", p);
@@ -507,8 +513,10 @@ static struct dentry *get_default_root(struct super_block *sb,
 	 */
 	dir_id = btrfs_super_root_dir(&root->fs_info->super_copy);
 	di = btrfs_lookup_dir_item(NULL, root, path, dir_id, "default", 7, 0);
-	if (IS_ERR(di))
+	if (IS_ERR(di)) {
+		btrfs_free_path(path);
 		return ERR_CAST(di);
+	}
 	if (!di) {
 		/*
 		 * Ok the default dir item isn't there.  This is weird since
@@ -741,7 +749,7 @@ static int btrfs_set_super(struct super_block *s, void *data)
  *	  for multiple device setup.  Make sure to keep it in sync.
  */
 static struct dentry *btrfs_mount(struct file_system_type *fs_type, int flags,
-		const char *dev_name, void *data)
+		const char *device_name, void *data)
 {
 	struct block_device *bdev = NULL;
 	struct super_block *s;
@@ -764,7 +772,7 @@ static struct dentry *btrfs_mount(struct file_system_type *fs_type, int flags,
 	if (error)
 		return ERR_PTR(error);
 
-	error = btrfs_scan_one_device(dev_name, mode, fs_type, &fs_devices);
+	error = btrfs_scan_one_device(device_name, mode, fs_type, &fs_devices);
 	if (error)
 		goto error_free_subvol_name;
 
@@ -915,6 +923,32 @@ static int btrfs_remount(struct super_block *sb, int *flags, char *data)
 	return 0;
 }
 
+/* Used to sort the devices by max_avail(descending sort) */
+static int btrfs_cmp_device_free_bytes(const void *dev_info1,
+				       const void *dev_info2)
+{
+	if (((struct btrfs_device_info *)dev_info1)->max_avail >
+	    ((struct btrfs_device_info *)dev_info2)->max_avail)
+		return -1;
+	else if (((struct btrfs_device_info *)dev_info1)->max_avail <
+		 ((struct btrfs_device_info *)dev_info2)->max_avail)
+		return 1;
+	else
+	return 0;
+}
+
+/*
+ * sort the devices by max_avail, in which max free extent size of each device
+ * is stored.(Descending Sort)
+ */
+static inline void btrfs_descending_sort_devices(
+					struct btrfs_device_info *devices,
+					size_t nr_devices)
+{
+	sort(devices, nr_devices, sizeof(struct btrfs_device_info),
+	     btrfs_cmp_device_free_bytes, NULL);
+}
+
 /*
  * The helper to calc the free space on the devices that can be used to store
  * file data.
@@ -1208,10 +1242,14 @@ static int __init init_btrfs_fs(void)
 	if (err)
 		goto free_extent_io;
 
-	err = btrfs_interface_init();
+	err = btrfs_delayed_inode_init();
 	if (err)
 		goto free_extent_map;
 
+	err = btrfs_interface_init();
+	if (err)
+		goto free_delayed_inode;
+
 	err = register_filesystem(&btrfs_fs_type);
 	if (err)
 		goto unregister_ioctl;
@@ -1221,6 +1259,8 @@ static int __init init_btrfs_fs(void)
 
 unregister_ioctl:
 	btrfs_interface_exit();
+free_delayed_inode:
+	btrfs_delayed_inode_exit();
 free_extent_map:
 	extent_map_exit();
 free_extent_io:
@@ -1237,6 +1277,7 @@ free_sysfs:
 static void __exit exit_btrfs_fs(void)
 {
 	btrfs_destroy_cachep();
+	btrfs_delayed_inode_exit();
 	extent_map_exit();
 	extent_io_exit();
 	btrfs_interface_exit();

fs/btrfs/sysfs.c

@@ -174,86 +174,9 @@ static const struct sysfs_ops btrfs_root_attr_ops = {
 	.store	= btrfs_root_attr_store,
 };
 
-static struct kobj_type btrfs_root_ktype = {
-	.default_attrs	= btrfs_root_attrs,
-	.sysfs_ops	= &btrfs_root_attr_ops,
-	.release	= btrfs_root_release,
-};
-
-static struct kobj_type btrfs_super_ktype = {
-	.default_attrs	= btrfs_super_attrs,
-	.sysfs_ops	= &btrfs_super_attr_ops,
-	.release	= btrfs_super_release,
-};
-
 /* /sys/fs/btrfs/ entry */
 static struct kset *btrfs_kset;
 
-int btrfs_sysfs_add_super(struct btrfs_fs_info *fs)
-{
-	int error;
-	char *name;
-	char c;
-	int len = strlen(fs->sb->s_id) + 1;
-	int i;
-
-	name = kmalloc(len, GFP_NOFS);
-	if (!name) {
-		error = -ENOMEM;
-		goto fail;
-	}
-
-	for (i = 0; i < len; i++) {
-		c = fs->sb->s_id[i];
-		if (c == '/' || c == '\\')
-			c = '!';
-		name[i] = c;
-	}
-	name[len] = '\0';
-
-	fs->super_kobj.kset = btrfs_kset;
-	error = kobject_init_and_add(&fs->super_kobj, &btrfs_super_ktype,
-				     NULL, "%s", name);
-	kfree(name);
-	if (error)
-		goto fail;
-
-	return 0;
-
-fail:
-	printk(KERN_ERR "btrfs: sysfs creation for super failed\n");
-	return error;
-}
-
-int btrfs_sysfs_add_root(struct btrfs_root *root)
-{
-	int error;
-
-	error = kobject_init_and_add(&root->root_kobj, &btrfs_root_ktype,
-				     &root->fs_info->super_kobj,
-				     "%s", root->name);
-	if (error)
-		goto fail;
-
-	return 0;
-
-fail:
-	printk(KERN_ERR "btrfs: sysfs creation for root failed\n");
-	return error;
-}
-
-void btrfs_sysfs_del_root(struct btrfs_root *root)
-{
-	kobject_put(&root->root_kobj);
-	wait_for_completion(&root->kobj_unregister);
-}
-
-void btrfs_sysfs_del_super(struct btrfs_fs_info *fs)
-{
-	kobject_put(&fs->super_kobj);
-	wait_for_completion(&fs->kobj_unregister);
-}
-
 int btrfs_init_sysfs(void)
 {
 	btrfs_kset = kset_create_and_add("btrfs", NULL, fs_kobj);

fs/btrfs/transaction.c

@@ -27,6 +27,7 @@
 #include "transaction.h"
 #include "locking.h"
 #include "tree-log.h"
+#include "inode-map.h"
 
 #define BTRFS_ROOT_TRANS_TAG 0
 
@@ -80,8 +81,7 @@ static noinline int join_transaction(struct btrfs_root *root)
 		INIT_LIST_HEAD(&cur_trans->pending_snapshots);
 		list_add_tail(&cur_trans->list, &root->fs_info->trans_list);
 		extent_io_tree_init(&cur_trans->dirty_pages,
-				     root->fs_info->btree_inode->i_mapping,
-				     GFP_NOFS);
+				     root->fs_info->btree_inode->i_mapping);
 		spin_lock(&root->fs_info->new_trans_lock);
 		root->fs_info->running_transaction = cur_trans;
 		spin_unlock(&root->fs_info->new_trans_lock);
@@ -347,49 +347,6 @@ out_unlock:
 	return ret;
 }
 
-#if 0
-/*
- * rate limit against the drop_snapshot code.  This helps to slow down new
- * operations if the drop_snapshot code isn't able to keep up.
- */
-static void throttle_on_drops(struct btrfs_root *root)
-{
-	struct btrfs_fs_info *info = root->fs_info;
-	int harder_count = 0;
-
-harder:
-	if (atomic_read(&info->throttles)) {
-		DEFINE_WAIT(wait);
-		int thr;
-		thr = atomic_read(&info->throttle_gen);
-
-		do {
-			prepare_to_wait(&info->transaction_throttle,
-					&wait, TASK_UNINTERRUPTIBLE);
-			if (!atomic_read(&info->throttles)) {
-				finish_wait(&info->transaction_throttle, &wait);
-				break;
-			}
-			schedule();
-			finish_wait(&info->transaction_throttle, &wait);
-		} while (thr == atomic_read(&info->throttle_gen));
-		harder_count++;
-
-		if (root->fs_info->total_ref_cache_size > 1 * 1024 * 1024 &&
-		    harder_count < 2)
-			goto harder;
-
-		if (root->fs_info->total_ref_cache_size > 5 * 1024 * 1024 &&
-		    harder_count < 10)
-			goto harder;
-
-		if (root->fs_info->total_ref_cache_size > 10 * 1024 * 1024 &&
-		    harder_count < 20)
-			goto harder;
-	}
-}
-#endif
-
 void btrfs_throttle(struct btrfs_root *root)
 {
 	mutex_lock(&root->fs_info->trans_mutex);
@@ -487,19 +444,40 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans,
 int btrfs_end_transaction(struct btrfs_trans_handle *trans,
 			  struct btrfs_root *root)
 {
-	return __btrfs_end_transaction(trans, root, 0, 1);
+	int ret;
+
+	ret = __btrfs_end_transaction(trans, root, 0, 1);
+	if (ret)
+		return ret;
+	return 0;
 }
 
 int btrfs_end_transaction_throttle(struct btrfs_trans_handle *trans,
 				   struct btrfs_root *root)
 {
-	return __btrfs_end_transaction(trans, root, 1, 1);
+	int ret;
+
+	ret = __btrfs_end_transaction(trans, root, 1, 1);
+	if (ret)
+		return ret;
+	return 0;
 }
 
 int btrfs_end_transaction_nolock(struct btrfs_trans_handle *trans,
 				 struct btrfs_root *root)
 {
-	return __btrfs_end_transaction(trans, root, 0, 0);
+	int ret;
+
+	ret = __btrfs_end_transaction(trans, root, 0, 0);
+	if (ret)
+		return ret;
+	return 0;
+}
+
+int btrfs_end_transaction_dmeta(struct btrfs_trans_handle *trans,
+				struct btrfs_root *root)
+{
+	return __btrfs_end_transaction(trans, root, 1, 1);
 }
 
 /*
@@ -760,8 +738,14 @@ static noinline int commit_fs_roots(struct btrfs_trans_handle *trans,
 			btrfs_update_reloc_root(trans, root);
 			btrfs_orphan_commit_root(trans, root);
 
+			btrfs_save_ino_cache(root, trans);
+
 			if (root->commit_root != root->node) {
+				mutex_lock(&root->fs_commit_mutex);
 				switch_commit_root(root);
+				btrfs_unpin_free_ino(root);
+				mutex_unlock(&root->fs_commit_mutex);
+
 				btrfs_set_root_node(&root->root_item,
 						    root->node);
 			}
@@ -809,97 +793,6 @@ int btrfs_defrag_root(struct btrfs_root *root, int cacheonly)
 	return ret;
 }
 
-#if 0
-/*
- * when dropping snapshots, we generate a ton of delayed refs, and it makes
- * sense not to join the transaction while it is trying to flush the current
- * queue of delayed refs out.
- *
- * This is used by the drop snapshot code only
- */
-static noinline int wait_transaction_pre_flush(struct btrfs_fs_info *info)
-{
-	DEFINE_WAIT(wait);
-
-	mutex_lock(&info->trans_mutex);
-	while (info->running_transaction &&
-	       info->running_transaction->delayed_refs.flushing) {
-		prepare_to_wait(&info->transaction_wait, &wait,
-				TASK_UNINTERRUPTIBLE);
-		mutex_unlock(&info->trans_mutex);
-
-		schedule();
-
-		mutex_lock(&info->trans_mutex);
-		finish_wait(&info->transaction_wait, &wait);
-	}
-	mutex_unlock(&info->trans_mutex);
-	return 0;
-}
-
-/*
- * Given a list of roots that need to be deleted, call btrfs_drop_snapshot on
- * all of them
- */
-int btrfs_drop_dead_root(struct btrfs_root *root)
-{
-	struct btrfs_trans_handle *trans;
-	struct btrfs_root *tree_root = root->fs_info->tree_root;
-	unsigned long nr;
-	int ret;
-
-	while (1) {
-		/*
-		 * we don't want to jump in and create a bunch of
-		 * delayed refs if the transaction is starting to close
-		 */
-		wait_transaction_pre_flush(tree_root->fs_info);
-		trans = btrfs_start_transaction(tree_root, 1);
-
-		/*
-		 * we've joined a transaction, make sure it isn't
-		 * closing right now
-		 */
-		if (trans->transaction->delayed_refs.flushing) {
-			btrfs_end_transaction(trans, tree_root);
-			continue;
-		}
-
-		ret = btrfs_drop_snapshot(trans, root);
-		if (ret != -EAGAIN)
-			break;
-
-		ret = btrfs_update_root(trans, tree_root,
-					&root->root_key,
-					&root->root_item);
-		if (ret)
-			break;
-
-		nr = trans->blocks_used;
-		ret = btrfs_end_transaction(trans, tree_root);
-		BUG_ON(ret);
-
-		btrfs_btree_balance_dirty(tree_root, nr);
-		cond_resched();
-	}
-	BUG_ON(ret);
-
-	ret = btrfs_del_root(trans, tree_root, &root->root_key);
-	BUG_ON(ret);
-
-	nr = trans->blocks_used;
-	ret = btrfs_end_transaction(trans, tree_root);
-	BUG_ON(ret);
-
-	free_extent_buffer(root->node);
-	free_extent_buffer(root->commit_root);
-	kfree(root);
-
-	btrfs_btree_balance_dirty(tree_root, nr);
-	return ret;
-}
-#endif
-
 /*
  * new snapshots need to be created at a very specific time in the
  * transaction commit.  This does the actual creation
@@ -930,7 +823,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
 		goto fail;
 	}
 
-	ret = btrfs_find_free_objectid(trans, tree_root, 0, &objectid);
+	ret = btrfs_find_free_objectid(tree_root, &objectid);
 	if (ret) {
 		pending->error = ret;
 		goto fail;
@@ -967,7 +860,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
 	BUG_ON(ret);
 	ret = btrfs_insert_dir_item(trans, parent_root,
 				dentry->d_name.name, dentry->d_name.len,
-				parent_inode->i_ino, &key,
+				parent_inode, &key,
 				BTRFS_FT_DIR, index);
 	BUG_ON(ret);
 
@@ -1009,7 +902,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
 	 */
 	ret = btrfs_add_root_ref(trans, tree_root, objectid,
 				 parent_root->root_key.objectid,
-				 parent_inode->i_ino, index,
+				 btrfs_ino(parent_inode), index,
 				 dentry->d_name.name, dentry->d_name.len);
 	BUG_ON(ret);
 	dput(parent);
@@ -1037,6 +930,14 @@ static noinline int create_pending_snapshots(struct btrfs_trans_handle *trans,
 	int ret;
 
 	list_for_each_entry(pending, head, list) {
+		/*
+		 * We must deal with the delayed items before creating
+		 * snapshots, or we will create a snapthot with inconsistent
+		 * information.
+		*/
+		ret = btrfs_run_delayed_items(trans, fs_info->fs_root);
+		BUG_ON(ret);
+
 		ret = create_pending_snapshot(trans, fs_info, pending);
 		BUG_ON(ret);
 	}
@@ -1290,6 +1191,9 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 			BUG_ON(ret);
 		}
 
+		ret = btrfs_run_delayed_items(trans, root);
+		BUG_ON(ret);
+
 		/*
 		 * rename don't use btrfs_join_transaction, so, once we
 		 * set the transaction to blocked above, we aren't going
@@ -1316,11 +1220,15 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 	ret = create_pending_snapshots(trans, root->fs_info);
 	BUG_ON(ret);
 
+	ret = btrfs_run_delayed_items(trans, root);
+	BUG_ON(ret);
+
 	ret = btrfs_run_delayed_refs(trans, root, (unsigned long)-1);
 	BUG_ON(ret);
 
 	WARN_ON(cur_trans != trans->transaction);
 
+	btrfs_scrub_pause(root);
 	/* btrfs_commit_tree_roots is responsible for getting the
 	 * various roots consistent with each other.  Every pointer
 	 * in the tree of tree roots has to point to the most up to date
@@ -1405,6 +1313,8 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
 
 	mutex_unlock(&root->fs_info->trans_mutex);
 
+	btrfs_scrub_continue(root);
+
 	if (current->journal_info == trans)
 		current->journal_info = NULL;
 
@@ -1432,6 +1342,8 @@ int btrfs_clean_old_snapshots(struct btrfs_root *root)
 		root = list_entry(list.next, struct btrfs_root, root_list);
 		list_del(&root->root_list);
 
+		btrfs_kill_all_delayed_nodes(root);
+
 		if (btrfs_header_backref_rev(root->node) <
 		    BTRFS_MIXED_BACKREF_REV)
 			btrfs_drop_snapshot(root, NULL, 0);

fs/btrfs/transaction.h

@@ -101,11 +101,8 @@ struct btrfs_trans_handle *btrfs_start_ioctl_transaction(struct btrfs_root *r,
 int btrfs_wait_for_commit(struct btrfs_root *root, u64 transid);
 int btrfs_write_and_wait_transaction(struct btrfs_trans_handle *trans,
 				     struct btrfs_root *root);
-int btrfs_commit_tree_roots(struct btrfs_trans_handle *trans,
-			    struct btrfs_root *root);
 
 int btrfs_add_dead_root(struct btrfs_root *root);
-int btrfs_drop_dead_root(struct btrfs_root *root);
 int btrfs_defrag_root(struct btrfs_root *root, int cacheonly);
 int btrfs_clean_old_snapshots(struct btrfs_root *root);
 int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
@@ -115,6 +112,8 @@ int btrfs_commit_transaction_async(struct btrfs_trans_handle *trans,
 				   int wait_for_unblock);
 int btrfs_end_transaction_throttle(struct btrfs_trans_handle *trans,
 				   struct btrfs_root *root);
+int btrfs_end_transaction_dmeta(struct btrfs_trans_handle *trans,
+				struct btrfs_root *root);
 int btrfs_should_end_transaction(struct btrfs_trans_handle *trans,
 				 struct btrfs_root *root);
 void btrfs_throttle(struct btrfs_root *root);

fs/btrfs/tree-defrag.c

@@ -97,7 +97,7 @@ int btrfs_defrag_leaves(struct btrfs_trans_handle *trans,
 		ret = 0;
 		goto out;
 	}
-	btrfs_release_path(root, path);
+	btrfs_release_path(path);
 	wret = btrfs_search_slot(trans, root, &key, path, 0, 1);
 
 	if (wret < 0) {

fs/btrfs/tree-log.c

@@ -333,13 +333,13 @@ static noinline int overwrite_item(struct btrfs_trans_handle *trans,
 			goto insert;
 
 		if (item_size == 0) {
-			btrfs_release_path(root, path);
+			btrfs_release_path(path);
 			return 0;
 		}
 		dst_copy = kmalloc(item_size, GFP_NOFS);
 		src_copy = kmalloc(item_size, GFP_NOFS);
 		if (!dst_copy || !src_copy) {
-			btrfs_release_path(root, path);
+			btrfs_release_path(path);
 			kfree(dst_copy);
 			kfree(src_copy);
 			return -ENOMEM;
@@ -361,13 +361,13 @@ static noinline int overwrite_item(struct btrfs_trans_handle *trans,
 		 * sync
 		 */
 		if (ret == 0) {
-			btrfs_release_path(root, path);
+			btrfs_release_path(path);
 			return 0;
 		}
 
 	}
 insert:
-	btrfs_release_path(root, path);
+	btrfs_release_path(path);
 	/* try to insert the key into the destination tree */
 	ret = btrfs_insert_empty_item(trans, root, path,
 				      key, item_size);
@@ -382,7 +382,6 @@ insert:
 		} else if (found_size < item_size) {
 			ret = btrfs_extend_item(trans, root, path,
 						item_size - found_size);
-			BUG_ON(ret);
 		}
 	} else if (ret) {
 		return ret;
@@ -438,7 +437,7 @@ insert:
 	}
 no_copy:
 	btrfs_mark_buffer_dirty(path->nodes[0]);
-	btrfs_release_path(root, path);
+	btrfs_release_path(path);
 	return 0;
 }
 
@@ -519,7 +518,7 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans,
 	 * file.  This must be done before the btrfs_drop_extents run
 	 * so we don't try to drop this extent.
 	 */
-	ret = btrfs_lookup_file_extent(trans, root, path, inode->i_ino,
+	ret = btrfs_lookup_file_extent(trans, root, path, btrfs_ino(inode),
 				       start, 0);
 
 	if (ret == 0 &&
@@ -544,11 +543,11 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans,
 		 * we don't have to do anything
 		 */
 		if (memcmp(&cmp1, &cmp2, sizeof(cmp1)) == 0) {
-			btrfs_release_path(root, path);
+			btrfs_release_path(path);
 			goto out;
 		}
 	}
-	btrfs_release_path(root, path);
+	btrfs_release_path(path);
 
 	saved_nbytes = inode_get_bytes(inode);
 	/* drop any overlapping extents */
@@ -590,6 +589,7 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans,
 						ins.objectid, ins.offset,
 						0, root->root_key.objectid,
 						key->objectid, offset);
+				BUG_ON(ret);
 			} else {
 				/*
 				 * insert the extent pointer in the extent
@@ -600,7 +600,7 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans,
 						key->objectid, offset, &ins);
 				BUG_ON(ret);
 			}
-			btrfs_release_path(root, path);
+			btrfs_release_path(path);
 
 			if (btrfs_file_extent_compression(eb, item)) {
 				csum_start = ins.objectid;
@@ -614,7 +614,7 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans,
 
 			ret = btrfs_lookup_csums_range(root->log_root,
 						csum_start, csum_end - 1,
-						&ordered_sums);
+						&ordered_sums, 0);
 			BUG_ON(ret);
 			while (!list_empty(&ordered_sums)) {
 				struct btrfs_ordered_sum *sums;
@@ -629,7 +629,7 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans,
 				kfree(sums);
 			}
 		} else {
-			btrfs_release_path(root, path);
+			btrfs_release_path(path);
 		}
 	} else if (found_type == BTRFS_FILE_EXTENT_INLINE) {
 		/* inline extents are easy, we just overwrite them */
@@ -675,10 +675,13 @@ static noinline int drop_one_dir_item(struct btrfs_trans_handle *trans,
 		return -ENOMEM;
 
 	read_extent_buffer(leaf, name, (unsigned long)(di + 1), name_len);
-	btrfs_release_path(root, path);
+	btrfs_release_path(path);
 
 	inode = read_one_inode(root, location.objectid);
-	BUG_ON(!inode);
+	if (!inode) {
+		kfree(name);
+		return -EIO;
+	}
 
 	ret = link_to_fixup_dir(trans, root, path, location.objectid);
 	BUG_ON(ret);
@@ -713,7 +716,7 @@ static noinline int inode_in_dir(struct btrfs_root *root,
 			goto out;
 	} else
 		goto out;
-	btrfs_release_path(root, path);
+	btrfs_release_path(path);
 
 	di = btrfs_lookup_dir_item(NULL, root, path, dirid, name, name_len, 0);
 	if (di && !IS_ERR(di)) {
@@ -724,7 +727,7 @@ static noinline int inode_in_dir(struct btrfs_root *root,
 		goto out;
 	match = 1;
 out:
-	btrfs_release_path(root, path);
+	btrfs_release_path(path);
 	return match;
 }
 
@@ -817,7 +820,10 @@ static noinline int add_inode_ref(struct btrfs_trans_handle *trans,
 		return -ENOENT;
 
 	inode = read_one_inode(root, key->objectid);
-	BUG_ON(!inode);
+	if (!inode) {
+		iput(dir);
+		return -EIO;
+	}
 
 	ref_ptr = btrfs_item_ptr_offset(eb, slot);
 	ref_end = ref_ptr + btrfs_item_size_nr(eb, slot);
@@ -832,7 +838,7 @@ again:
 	read_extent_buffer(eb, name, (unsigned long)(ref + 1), namelen);
 
 	/* if we already have a perfect match, we're done */
-	if (inode_in_dir(root, path, dir->i_ino, inode->i_ino,
+	if (inode_in_dir(root, path, btrfs_ino(dir), btrfs_ino(inode),
 			 btrfs_inode_ref_index(eb, ref),
 			 name, namelen)) {
 		goto out;
@@ -884,7 +890,7 @@ again:
 			if (!backref_in_log(log, key, victim_name,
 					    victim_name_len)) {
 				btrfs_inc_nlink(inode);
-				btrfs_release_path(root, path);
+				btrfs_release_path(path);
 
 				ret = btrfs_unlink_inode(trans, root, dir,
 							 inode, victim_name,
@@ -901,7 +907,7 @@ again:
 		 */
 		search_done = 1;
 	}
-	btrfs_release_path(root, path);
+	btrfs_release_path(path);
 
 insert:
 	/* insert our name */
@@ -922,7 +928,7 @@ out:
 	BUG_ON(ret);
 
 out_nowrite:
-	btrfs_release_path(root, path);
+	btrfs_release_path(path);
 	iput(dir);
 	iput(inode);
 	return 0;
@@ -960,8 +966,9 @@ static noinline int fixup_inode_link_count(struct btrfs_trans_handle *trans,
 	unsigned long ptr;
 	unsigned long ptr_end;
 	int name_len;
+	u64 ino = btrfs_ino(inode);
 
-	key.objectid = inode->i_ino;
+	key.objectid = ino;
 	key.type = BTRFS_INODE_REF_KEY;
 	key.offset = (u64)-1;
 
@@ -980,7 +987,7 @@ static noinline int fixup_inode_link_count(struct btrfs_trans_handle *trans,
 		}
 		btrfs_item_key_to_cpu(path->nodes[0], &key,
 				      path->slots[0]);
-		if (key.objectid != inode->i_ino ||
+		if (key.objectid != ino ||
 		    key.type != BTRFS_INODE_REF_KEY)
 			break;
 		ptr = btrfs_item_ptr_offset(path->nodes[0], path->slots[0]);
@@ -999,9 +1006,9 @@ static noinline int fixup_inode_link_count(struct btrfs_trans_handle *trans,
 		if (key.offset == 0)
 			break;
 		key.offset--;
-		btrfs_release_path(root, path);
+		btrfs_release_path(path);
 	}
-	btrfs_release_path(root, path);
+	btrfs_release_path(path);
 	if (nlink != inode->i_nlink) {
 		inode->i_nlink = nlink;
 		btrfs_update_inode(trans, root, inode);
@@ -1011,10 +1018,10 @@ static noinline int fixup_inode_link_count(struct btrfs_trans_handle *trans,
 	if (inode->i_nlink == 0) {
 		if (S_ISDIR(inode->i_mode)) {
 			ret = replay_dir_deletes(trans, root, NULL, path,
-						 inode->i_ino, 1);
+						 ino, 1);
 			BUG_ON(ret);
 		}
-		ret = insert_orphan_item(trans, root, inode->i_ino);
+		ret = insert_orphan_item(trans, root, ino);
 		BUG_ON(ret);
 	}
 	btrfs_free_path(path);
@@ -1050,11 +1057,13 @@ static noinline int fixup_inode_link_counts(struct btrfs_trans_handle *trans,
 			break;
 
 		ret = btrfs_del_item(trans, root, path);
-		BUG_ON(ret);
+		if (ret)
+			goto out;
 
-		btrfs_release_path(root, path);
+		btrfs_release_path(path);
 		inode = read_one_inode(root, key.offset);
-		BUG_ON(!inode);
+		if (!inode)
+			return -EIO;
 
 		ret = fixup_inode_link_count(trans, root, inode);
 		BUG_ON(ret);
@@ -1068,8 +1077,10 @@ static noinline int fixup_inode_link_counts(struct btrfs_trans_handle *trans,
 		 */
 		key.offset = (u64)-1;
 	}
-	btrfs_release_path(root, path);
-	return 0;
+	ret = 0;
+out:
+	btrfs_release_path(path);
+	return ret;
 }
 
 
@@ -1088,7 +1099,8 @@ static noinline int link_to_fixup_dir(struct btrfs_trans_handle *trans,
 	struct inode *inode;
 
 	inode = read_one_inode(root, objectid);
-	BUG_ON(!inode);
+	if (!inode)
+		return -EIO;
 
 	key.objectid = BTRFS_TREE_LOG_FIXUP_OBJECTID;
 	btrfs_set_key_type(&key, BTRFS_ORPHAN_ITEM_KEY);
@@ -1096,7 +1108,7 @@ static noinline int link_to_fixup_dir(struct btrfs_trans_handle *trans,
 
 	ret = btrfs_insert_empty_item(trans, root, path, &key, 0);
 
-	btrfs_release_path(root, path);
+	btrfs_release_path(path);
 	if (ret == 0) {
 		btrfs_inc_nlink(inode);
 		btrfs_update_inode(trans, root, inode);
@@ -1175,7 +1187,8 @@ static noinline int replay_one_name(struct btrfs_trans_handle *trans,
 	int ret;
 
 	dir = read_one_inode(root, key->objectid);
-	BUG_ON(!dir);
+	if (!dir)
+		return -EIO;
 
 	name_len = btrfs_dir_name_len(eb, di);
 	name = kmalloc(name_len, GFP_NOFS);
@@ -1192,7 +1205,7 @@ static noinline int replay_one_name(struct btrfs_trans_handle *trans,
 		exists = 1;
 	else
 		exists = 0;
-	btrfs_release_path(root, path);
+	btrfs_release_path(path);
 
 	if (key->type == BTRFS_DIR_ITEM_KEY) {
 		dst_di = btrfs_lookup_dir_item(trans, root, path, key->objectid,
@@ -1205,7 +1218,7 @@ static noinline int replay_one_name(struct btrfs_trans_handle *trans,
 	} else {
 		BUG();
 	}
-	if (!dst_di || IS_ERR(dst_di)) {
+	if (IS_ERR_OR_NULL(dst_di)) {
 		/* we need a sequence number to insert, so we only
 		 * do inserts for the BTRFS_DIR_INDEX_KEY types
 		 */
@@ -1236,13 +1249,13 @@ static noinline int replay_one_name(struct btrfs_trans_handle *trans,
 	if (key->type == BTRFS_DIR_INDEX_KEY)
 		goto insert;
 out:
-	btrfs_release_path(root, path);
+	btrfs_release_path(path);
 	kfree(name);
 	iput(dir);
 	return 0;
 
 insert:
-	btrfs_release_path(root, path);
+	btrfs_release_path(path);
 	ret = insert_one_name(trans, root, path, key->objectid, key->offset,
 			      name, name_len, log_type, &log_key);
 
@@ -1363,7 +1376,7 @@ next:
 	*end_ret = found_end;
 	ret = 0;
 out:
-	btrfs_release_path(root, path);
+	btrfs_release_path(path);
 	return ret;
 }
 
@@ -1426,12 +1439,15 @@ again:
 						     dir_key->offset,
 						     name, name_len, 0);
 		}
-		if (!log_di || IS_ERR(log_di)) {
+		if (IS_ERR_OR_NULL(log_di)) {
 			btrfs_dir_item_key_to_cpu(eb, di, &location);
-			btrfs_release_path(root, path);
-			btrfs_release_path(log, log_path);
+			btrfs_release_path(path);
+			btrfs_release_path(log_path);
 			inode = read_one_inode(root, location.objectid);
-			BUG_ON(!inode);
+			if (!inode) {
+				kfree(name);
+				return -EIO;
+			}
 
 			ret = link_to_fixup_dir(trans, root,
 						path, location.objectid);
@@ -1453,7 +1469,7 @@ again:
 			ret = 0;
 			goto out;
 		}
-		btrfs_release_path(log, log_path);
+		btrfs_release_path(log_path);
 		kfree(name);
 
 		ptr = (unsigned long)(di + 1);
@@ -1461,8 +1477,8 @@ again:
 	}
 	ret = 0;
 out:
-	btrfs_release_path(root, path);
-	btrfs_release_path(log, log_path);
+	btrfs_release_path(path);
+	btrfs_release_path(log_path);
 	return ret;
 }
 
@@ -1550,7 +1566,7 @@ again:
 				break;
 			dir_key.offset = found_key.offset + 1;
 		}
-		btrfs_release_path(root, path);
+		btrfs_release_path(path);
 		if (range_end == (u64)-1)
 			break;
 		range_start = range_end + 1;
@@ -1561,11 +1577,11 @@ next_type:
 	if (key_type == BTRFS_DIR_LOG_ITEM_KEY) {
 		key_type = BTRFS_DIR_LOG_INDEX_KEY;
 		dir_key.type = BTRFS_DIR_INDEX_KEY;
-		btrfs_release_path(root, path);
+		btrfs_release_path(path);
 		goto again;
 	}
 out:
-	btrfs_release_path(root, path);
+	btrfs_release_path(path);
 	btrfs_free_path(log_path);
 	iput(dir);
 	return ret;
@@ -2093,7 +2109,9 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans,
 	 * the running transaction open, so a full commit can't hop
 	 * in and cause problems either.
 	 */
+	btrfs_scrub_pause_super(root);
 	write_ctree_super(trans, root->fs_info->tree_root, 1);
+	btrfs_scrub_continue_super(root);
 	ret = 0;
 
 	mutex_lock(&root->log_mutex);
@@ -2197,6 +2215,7 @@ int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans,
 	int ret;
 	int err = 0;
 	int bytes_del = 0;
+	u64 dir_ino = btrfs_ino(dir);
 
 	if (BTRFS_I(dir)->logged_trans < trans->transid)
 		return 0;
@@ -2214,7 +2233,7 @@ int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans,
 		goto out_unlock;
 	}
 
-	di = btrfs_lookup_dir_item(trans, log, path, dir->i_ino,
+	di = btrfs_lookup_dir_item(trans, log, path, dir_ino,
 				   name, name_len, -1);
 	if (IS_ERR(di)) {
 		err = PTR_ERR(di);
@@ -2225,8 +2244,8 @@ int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans,
 		bytes_del += name_len;
 		BUG_ON(ret);
 	}
-	btrfs_release_path(log, path);
-	di = btrfs_lookup_dir_index_item(trans, log, path, dir->i_ino,
+	btrfs_release_path(path);
+	di = btrfs_lookup_dir_index_item(trans, log, path, dir_ino,
 					 index, name, name_len, -1);
 	if (IS_ERR(di)) {
 		err = PTR_ERR(di);
@@ -2244,10 +2263,10 @@ int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans,
 	if (bytes_del) {
 		struct btrfs_key key;
 
-		key.objectid = dir->i_ino;
+		key.objectid = dir_ino;
 		key.offset = 0;
 		key.type = BTRFS_INODE_ITEM_KEY;
-		btrfs_release_path(log, path);
+		btrfs_release_path(path);
 
 		ret = btrfs_search_slot(trans, log, &key, path, 0, 1);
 		if (ret < 0) {
@@ -2269,7 +2288,7 @@ int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans,
 			btrfs_mark_buffer_dirty(path->nodes[0]);
 		} else
 			ret = 0;
-		btrfs_release_path(log, path);
+		btrfs_release_path(path);
 	}
 fail:
 	btrfs_free_path(path);
@@ -2303,7 +2322,7 @@ int btrfs_del_inode_ref_in_log(struct btrfs_trans_handle *trans,
 	log = root->log_root;
 	mutex_lock(&BTRFS_I(inode)->log_mutex);
 
-	ret = btrfs_del_inode_ref(trans, log, name, name_len, inode->i_ino,
+	ret = btrfs_del_inode_ref(trans, log, name, name_len, btrfs_ino(inode),
 				  dirid, &index);
 	mutex_unlock(&BTRFS_I(inode)->log_mutex);
 	if (ret == -ENOSPC) {
@@ -2344,7 +2363,7 @@ static noinline int insert_dir_log_key(struct btrfs_trans_handle *trans,
 			      struct btrfs_dir_log_item);
 	btrfs_set_dir_log_end(path->nodes[0], item, last_offset);
 	btrfs_mark_buffer_dirty(path->nodes[0]);
-	btrfs_release_path(log, path);
+	btrfs_release_path(path);
 	return 0;
 }
 
@@ -2369,13 +2388,14 @@ static noinline int log_dir_items(struct btrfs_trans_handle *trans,
 	int nritems;
 	u64 first_offset = min_offset;
 	u64 last_offset = (u64)-1;
+	u64 ino = btrfs_ino(inode);
 
 	log = root->log_root;
-	max_key.objectid = inode->i_ino;
+	max_key.objectid = ino;
 	max_key.offset = (u64)-1;
 	max_key.type = key_type;
 
-	min_key.objectid = inode->i_ino;
+	min_key.objectid = ino;
 	min_key.type = key_type;
 	min_key.offset = min_offset;
 
@@ -2388,18 +2408,17 @@ static noinline int log_dir_items(struct btrfs_trans_handle *trans,
 	 * we didn't find anything from this transaction, see if there
 	 * is anything at all
 	 */
-	if (ret != 0 || min_key.objectid != inode->i_ino ||
-	    min_key.type != key_type) {
-		min_key.objectid = inode->i_ino;
+	if (ret != 0 || min_key.objectid != ino || min_key.type != key_type) {
+		min_key.objectid = ino;
 		min_key.type = key_type;
 		min_key.offset = (u64)-1;
-		btrfs_release_path(root, path);
+		btrfs_release_path(path);
 		ret = btrfs_search_slot(NULL, root, &min_key, path, 0, 0);
 		if (ret < 0) {
-			btrfs_release_path(root, path);
+			btrfs_release_path(path);
 			return ret;
 		}
-		ret = btrfs_previous_item(root, path, inode->i_ino, key_type);
+		ret = btrfs_previous_item(root, path, ino, key_type);
 
 		/* if ret == 0 there are items for this type,
 		 * create a range to tell us the last key of this type.
@@ -2417,7 +2436,7 @@ static noinline int log_dir_items(struct btrfs_trans_handle *trans,
 	}
 
 	/* go backward to find any previous key */
-	ret = btrfs_previous_item(root, path, inode->i_ino, key_type);
+	ret = btrfs_previous_item(root, path, ino, key_type);
 	if (ret == 0) {
 		struct btrfs_key tmp;
 		btrfs_item_key_to_cpu(path->nodes[0], &tmp, path->slots[0]);
@@ -2432,7 +2451,7 @@ static noinline int log_dir_items(struct btrfs_trans_handle *trans,
 			}
 		}
 	}
-	btrfs_release_path(root, path);
+	btrfs_release_path(path);
 
 	/* find the first key from this transaction again */
 	ret = btrfs_search_slot(NULL, root, &min_key, path, 0, 0);
@@ -2452,8 +2471,7 @@ static noinline int log_dir_items(struct btrfs_trans_handle *trans,
 		for (i = path->slots[0]; i < nritems; i++) {
 			btrfs_item_key_to_cpu(src, &min_key, i);
 
-			if (min_key.objectid != inode->i_ino ||
-			    min_key.type != key_type)
+			if (min_key.objectid != ino || min_key.type != key_type)
 				goto done;
 			ret = overwrite_item(trans, log, dst_path, src, i,
 					     &min_key);
@@ -2474,7 +2492,7 @@ static noinline int log_dir_items(struct btrfs_trans_handle *trans,
 			goto done;
 		}
 		btrfs_item_key_to_cpu(path->nodes[0], &tmp, path->slots[0]);
-		if (tmp.objectid != inode->i_ino || tmp.type != key_type) {
+		if (tmp.objectid != ino || tmp.type != key_type) {
 			last_offset = (u64)-1;
 			goto done;
 		}
@@ -2490,8 +2508,8 @@ static noinline int log_dir_items(struct btrfs_trans_handle *trans,
 		}
 	}
 done:
-	btrfs_release_path(root, path);
-	btrfs_release_path(log, dst_path);
+	btrfs_release_path(path);
+	btrfs_release_path(dst_path);
 
 	if (err == 0) {
 		*last_offset_ret = last_offset;
@@ -2500,8 +2518,7 @@ done:
 		 * is valid
 		 */
 		ret = insert_dir_log_key(trans, log, path, key_type,
-					 inode->i_ino, first_offset,
-					 last_offset);
+					 ino, first_offset, last_offset);
 		if (ret)
 			err = ret;
 	}
@@ -2587,10 +2604,11 @@ static int drop_objectid_items(struct btrfs_trans_handle *trans,
 			break;
 
 		ret = btrfs_del_item(trans, log, path);
-		BUG_ON(ret);
-		btrfs_release_path(log, path);
+		if (ret)
+			break;
+		btrfs_release_path(path);
 	}
-	btrfs_release_path(log, path);
+	btrfs_release_path(path);
 	return ret;
 }
 
@@ -2665,6 +2683,9 @@ static noinline int copy_items(struct btrfs_trans_handle *trans,
 			extent = btrfs_item_ptr(src, start_slot + i,
 						struct btrfs_file_extent_item);
 
+			if (btrfs_file_extent_generation(src, extent) < trans->transid)
+				continue;
+
 			found_type = btrfs_file_extent_type(src, extent);
 			if (found_type == BTRFS_FILE_EXTENT_REG ||
 			    found_type == BTRFS_FILE_EXTENT_PREALLOC) {
@@ -2689,14 +2710,14 @@ static noinline int copy_items(struct btrfs_trans_handle *trans,
 				ret = btrfs_lookup_csums_range(
 						log->fs_info->csum_root,
 						ds + cs, ds + cs + cl - 1,
-						&ordered_sums);
+						&ordered_sums, 0);
 				BUG_ON(ret);
 			}
 		}
 	}
 
 	btrfs_mark_buffer_dirty(dst_path->nodes[0]);
-	btrfs_release_path(log, dst_path);
+	btrfs_release_path(dst_path);
 	kfree(ins_data);
 
 	/*
@@ -2745,6 +2766,7 @@ static int btrfs_log_inode(struct btrfs_trans_handle *trans,
 	int nritems;
 	int ins_start_slot = 0;
 	int ins_nr;
+	u64 ino = btrfs_ino(inode);
 
 	log = root->log_root;
 
@@ -2757,11 +2779,11 @@ static int btrfs_log_inode(struct btrfs_trans_handle *trans,
 		return -ENOMEM;
 	}
 
-	min_key.objectid = inode->i_ino;
+	min_key.objectid = ino;
 	min_key.type = BTRFS_INODE_ITEM_KEY;
 	min_key.offset = 0;
 
-	max_key.objectid = inode->i_ino;
+	max_key.objectid = ino;
 
 	/* today the code can only do partial logging of directories */
 	if (!S_ISDIR(inode->i_mode))
@@ -2773,6 +2795,13 @@ static int btrfs_log_inode(struct btrfs_trans_handle *trans,
 		max_key.type = (u8)-1;
 	max_key.offset = (u64)-1;
 
+	ret = btrfs_commit_inode_delayed_items(trans, inode);
+	if (ret) {
+		btrfs_free_path(path);
+		btrfs_free_path(dst_path);
+		return ret;
+	}
+
 	mutex_lock(&BTRFS_I(inode)->log_mutex);
 
 	/*
@@ -2784,8 +2813,7 @@ static int btrfs_log_inode(struct btrfs_trans_handle *trans,
 
 		if (inode_only == LOG_INODE_EXISTS)
 			max_key_type = BTRFS_XATTR_ITEM_KEY;
-		ret = drop_objectid_items(trans, log, path,
-					  inode->i_ino, max_key_type);
+		ret = drop_objectid_items(trans, log, path, ino, max_key_type);
 	} else {
 		ret = btrfs_truncate_inode_items(trans, log, inode, 0, 0);
 	}
@@ -2803,7 +2831,7 @@ static int btrfs_log_inode(struct btrfs_trans_handle *trans,
 			break;
 again:
 		/* note, ins_nr might be > 0 here, cleanup outside the loop */
-		if (min_key.objectid != inode->i_ino)
+		if (min_key.objectid != ino)
 			break;
 		if (min_key.type > max_key.type)
 			break;
@@ -2845,7 +2873,7 @@ next_slot:
 			}
 			ins_nr = 0;
 		}
-		btrfs_release_path(root, path);
+		btrfs_release_path(path);
 
 		if (min_key.offset < (u64)-1)
 			min_key.offset++;
@@ -2868,8 +2896,8 @@ next_slot:
 	}
 	WARN_ON(ins_nr);
 	if (inode_only == LOG_INODE_ALL && S_ISDIR(inode->i_mode)) {
-		btrfs_release_path(root, path);
-		btrfs_release_path(log, dst_path);
+		btrfs_release_path(path);
+		btrfs_release_path(dst_path);
 		ret = log_directory_changes(trans, root, inode, path, dst_path);
 		if (ret) {
 			err = ret;
@@ -3136,7 +3164,7 @@ again:
 		}
 		btrfs_item_key_to_cpu(path->nodes[0], &found_key,
 				      path->slots[0]);
-		btrfs_release_path(log_root_tree, path);
+		btrfs_release_path(path);
 		if (found_key.objectid != BTRFS_TREE_LOG_OBJECTID)
 			break;
 
@@ -3171,7 +3199,7 @@ again:
 		if (found_key.offset == 0)
 			break;
 	}
-	btrfs_release_path(log_root_tree, path);
+	btrfs_release_path(path);
 
 	/* step one is to pin it all, step two is to replay just inodes */
 	if (wc.pin) {

fs/btrfs/tree-log.h

@@ -38,7 +38,6 @@ int btrfs_del_inode_ref_in_log(struct btrfs_trans_handle *trans,
 			       struct btrfs_root *root,
 			       const char *name, int name_len,
 			       struct inode *inode, u64 dirid);
-int btrfs_join_running_log_trans(struct btrfs_root *root);
 int btrfs_end_log_trans(struct btrfs_root *root);
 int btrfs_pin_log_trans(struct btrfs_root *root);
 int btrfs_log_inode_parent(struct btrfs_trans_handle *trans,

fs/btrfs/version.sh

@@ -1,43 +0,0 @@
-#!/bin/bash
-#
-# determine-version -- report a useful version for releases
-#
-# Copyright 2008, Aron Griffis <agriffis@n01se.net>
-# Copyright 2008, Oracle
-# Released under the GNU GPLv2
- 
-v="v0.16"
-
-which git &> /dev/null
-if [ $? == 0 ]; then
-    git branch >& /dev/null
-    if [ $? == 0 ]; then
-	    if head=`git rev-parse --verify HEAD 2>/dev/null`; then
-		if tag=`git describe --tags 2>/dev/null`; then
-		    v="$tag"
-		fi
-
-		# Are there uncommitted changes?
-		git update-index --refresh --unmerged > /dev/null
-		if git diff-index --name-only HEAD | \
-		    grep -v "^scripts/package" \
-		    | read dummy; then
-		    v="$v"-dirty
-		fi
-	    fi
-    fi
-fi
- 
-echo "#ifndef __BUILD_VERSION" > .build-version.h
-echo "#define __BUILD_VERSION" >> .build-version.h
-echo "#define BTRFS_BUILD_VERSION \"Btrfs $v\"" >> .build-version.h
-echo "#endif" >> .build-version.h
-
-diff -q version.h .build-version.h >& /dev/null
-
-if [ $? == 0 ]; then
-    rm .build-version.h
-    exit 0
-fi
-
-mv .build-version.h version.h

fs/btrfs/volumes.h

@@ -85,7 +85,12 @@ struct btrfs_device {
 	/* physical drive uuid (or lvm uuid) */
 	u8 uuid[BTRFS_UUID_SIZE];
 
+	/* per-device scrub information */
+	struct scrub_dev *scrub_device;
+
 	struct btrfs_work work;
+	struct rcu_head rcu;
+	struct work_struct rcu_work;
 };
 
 struct btrfs_fs_devices {
@@ -144,6 +149,7 @@ struct btrfs_device_info {
 	struct btrfs_device *dev;
 	u64 dev_offset;
 	u64 max_avail;
+	u64 total_avail;
 };
 
 struct map_lookup {
@@ -157,20 +163,8 @@ struct map_lookup {
 	struct btrfs_bio_stripe stripes[];
 };
 
-/* Used to sort the devices by max_avail(descending sort) */
-int btrfs_cmp_device_free_bytes(const void *dev_info1, const void *dev_info2);
-
-/*
- * sort the devices by max_avail, in which max free extent size of each device
- * is stored.(Descending Sort)
- */
-static inline void btrfs_descending_sort_devices(
-					struct btrfs_device_info *devices,
-					size_t nr_devices)
-{
-	sort(devices, nr_devices, sizeof(struct btrfs_device_info),
-	     btrfs_cmp_device_free_bytes, NULL);
-}
+#define map_lookup_size(n) (sizeof(struct map_lookup) + \
+			    (sizeof(struct btrfs_bio_stripe) * (n)))
 
 int btrfs_account_dev_extents_size(struct btrfs_device *device, u64 start,
 				   u64 end, u64 *length);
@@ -196,7 +190,6 @@ void btrfs_mapping_init(struct btrfs_mapping_tree *tree);
 void btrfs_mapping_tree_free(struct btrfs_mapping_tree *tree);
 int btrfs_map_bio(struct btrfs_root *root, int rw, struct bio *bio,
 		  int mirror_num, int async_submit);
-int btrfs_read_super_device(struct btrfs_root *root, struct extent_buffer *buf);
 int btrfs_open_devices(struct btrfs_fs_devices *fs_devices,
 		       fmode_t flags, void *holder);
 int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder,
@@ -209,8 +202,6 @@ int btrfs_add_device(struct btrfs_trans_handle *trans,
 int btrfs_rm_device(struct btrfs_root *root, char *device_path);
 int btrfs_cleanup_fs_uuids(void);
 int btrfs_num_copies(struct btrfs_mapping_tree *map_tree, u64 logical, u64 len);
-int btrfs_unplug_page(struct btrfs_mapping_tree *map_tree,
-		      u64 logical, struct page *page);
 int btrfs_grow_device(struct btrfs_trans_handle *trans,
 		      struct btrfs_device *device, u64 new_size);
 struct btrfs_device *btrfs_find_device(struct btrfs_root *root, u64 devid,
@@ -218,8 +209,6 @@ struct btrfs_device *btrfs_find_device(struct btrfs_root *root, u64 devid,
 int btrfs_shrink_device(struct btrfs_device *device, u64 new_size);
 int btrfs_init_new_device(struct btrfs_root *root, char *path);
 int btrfs_balance(struct btrfs_root *dev_root);
-void btrfs_unlock_volumes(void);
-void btrfs_lock_volumes(void);
 int btrfs_chunk_readonly(struct btrfs_root *root, u64 chunk_offset);
 int find_free_dev_extent(struct btrfs_trans_handle *trans,
 			 struct btrfs_device *device, u64 num_bytes,

fs/btrfs/xattr.c

@@ -44,7 +44,7 @@ ssize_t __btrfs_getxattr(struct inode *inode, const char *name,
 		return -ENOMEM;
 
 	/* lookup the xattr by name */
-	di = btrfs_lookup_xattr(NULL, root, path, inode->i_ino, name,
+	di = btrfs_lookup_xattr(NULL, root, path, btrfs_ino(inode), name,
 				strlen(name), 0);
 	if (!di) {
 		ret = -ENODATA;
@@ -103,7 +103,7 @@ static int do_setxattr(struct btrfs_trans_handle *trans,
 		return -ENOMEM;
 
 	/* first lets see if we already have this xattr */
-	di = btrfs_lookup_xattr(trans, root, path, inode->i_ino, name,
+	di = btrfs_lookup_xattr(trans, root, path, btrfs_ino(inode), name,
 				strlen(name), -1);
 	if (IS_ERR(di)) {
 		ret = PTR_ERR(di);
@@ -120,13 +120,13 @@ static int do_setxattr(struct btrfs_trans_handle *trans,
 
 		ret = btrfs_delete_one_dir_name(trans, root, path, di);
 		BUG_ON(ret);
-		btrfs_release_path(root, path);
+		btrfs_release_path(path);
 
 		/* if we don't have a value then we are removing the xattr */
 		if (!value)
 			goto out;
 	} else {
-		btrfs_release_path(root, path);
+		btrfs_release_path(path);
 
 		if (flags & XATTR_REPLACE) {
 			/* we couldn't find the attr to replace */
@@ -136,7 +136,7 @@ static int do_setxattr(struct btrfs_trans_handle *trans,
 	}
 
 	/* ok we have to create a completely new xattr */
-	ret = btrfs_insert_xattr_item(trans, root, path, inode->i_ino,
+	ret = btrfs_insert_xattr_item(trans, root, path, btrfs_ino(inode),
 				      name, name_len, value, size);
 	BUG_ON(ret);
 out:
@@ -190,7 +190,7 @@ ssize_t btrfs_listxattr(struct dentry *dentry, char *buffer, size_t size)
 	 * NOTE: we set key.offset = 0; because we want to start with the
 	 * first xattr that we find and walk forward
 	 */
-	key.objectid = inode->i_ino;
+	key.objectid = btrfs_ino(inode);
 	btrfs_set_key_type(&key, BTRFS_XATTR_ITEM_KEY);
 	key.offset = 0;