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

Pull btrfs update from Chris Mason:
 "The biggest feature in the pull is the new (and still experimental)
  raid56 code that David Woodhouse started long ago.  I'm still working
  on the parity logging setup that will avoid inconsistent parity after
  a crash, so this is only for testing right now.  But, I'd really like
  to get it out to a broader audience to hammer out any performance
  issues or other problems.

  scrub does not yet correct errors on raid5/6 either.

  Josef has another pass at fsync performance.  The big change here is
  to combine waiting for metadata with waiting for data, which is a big
  latency win.  It is also step one toward using atomics from the
  hardware during a commit.

  Mark Fasheh has a new way to use btrfs send/receive to send only the
  metadata changes.  SUSE is using this to make snapper more efficient
  at finding changes between snapshosts.

  Snapshot-aware defrag is also included.

  Otherwise we have a large number of fixes and cleanups.  Eric Sandeen
  wins the award for removing the most lines, and I'm hoping we steal
  this idea from XFS over and over again."

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs: (118 commits)
  btrfs: fixup/remove module.h usage as required
  Btrfs: delete inline extents when we find them during logging
  btrfs: try harder to allocate raid56 stripe cache
  Btrfs: cleanup to make the function btrfs_delalloc_reserve_metadata more logic
  Btrfs: don't call btrfs_qgroup_free if just btrfs_qgroup_reserve fails
  Btrfs: remove reduplicate check about root in the function btrfs_clean_quota_tree
  Btrfs: return ENOMEM rather than use BUG_ON when btrfs_alloc_path fails
  Btrfs: fix missing deleted items in btrfs_clean_quota_tree
  btrfs: use only inline_pages from extent buffer
  Btrfs: fix wrong reserved space when deleting a snapshot/subvolume
  Btrfs: fix wrong reserved space in qgroup during snap/subv creation
  Btrfs: remove unnecessary dget_parent/dput when creating the pending snapshot
  btrfs: remove a printk from scan_one_device
  Btrfs: fix NULL pointer after aborting a transaction
  Btrfs: fix memory leak of log roots
  Btrfs: copy everything if we've created an inline extent
  btrfs: cleanup for open-coded alignment
  Btrfs: do not change inode flags in rename
  Btrfs: use reserved space for creating a snapshot
  clear chunk_alloc flag on retryable failure
  ...

fs/btrfs/Kconfig

@@ -5,6 +5,9 @@ config BTRFS_FS
 	select ZLIB_DEFLATE
 	select LZO_COMPRESS
 	select LZO_DECOMPRESS
+	select RAID6_PQ
+	select XOR_BLOCKS
+
 	help
 	  Btrfs is a new filesystem with extents, writable snapshotting,
 	  support for multiple devices and many more features.

fs/btrfs/Makefile

@@ -8,7 +8,7 @@ btrfs-y += super.o ctree.o extent-tree.o print-tree.o root-tree.o dir-item.o \
 	   extent_io.o volumes.o async-thread.o ioctl.o locking.o orphan.o \
 	   export.o tree-log.o free-space-cache.o zlib.o lzo.o \
 	   compression.o delayed-ref.o relocation.o delayed-inode.o scrub.o \
-	   reada.o backref.o ulist.o qgroup.o send.o dev-replace.o
+	   reada.o backref.o ulist.o qgroup.o send.o dev-replace.o raid56.o
 
 btrfs-$(CONFIG_BTRFS_FS_POSIX_ACL) += acl.o
 btrfs-$(CONFIG_BTRFS_FS_CHECK_INTEGRITY) += check-integrity.o

fs/btrfs/backref.c

@@ -352,11 +352,8 @@ static int __resolve_indirect_refs(struct btrfs_fs_info *fs_info,
 		err = __resolve_indirect_ref(fs_info, search_commit_root,
 					     time_seq, ref, parents,
 					     extent_item_pos);
-		if (err) {
-			if (ret == 0)
-				ret = err;
+		if (err)
 			continue;
-		}
 
 		/* we put the first parent into the ref at hand */
 		ULIST_ITER_INIT(&uiter);

fs/btrfs/backref.h

@@ -19,7 +19,7 @@
 #ifndef __BTRFS_BACKREF__
 #define __BTRFS_BACKREF__
 
-#include "ioctl.h"
+#include <linux/btrfs.h>
 #include "ulist.h"
 #include "extent_io.h"
 

fs/btrfs/btrfs_inode.h

@@ -40,6 +40,8 @@
 #define BTRFS_INODE_HAS_ASYNC_EXTENT		6
 #define BTRFS_INODE_NEEDS_FULL_SYNC		7
 #define BTRFS_INODE_COPY_EVERYTHING		8
+#define BTRFS_INODE_IN_DELALLOC_LIST		9
+#define BTRFS_INODE_READDIO_NEED_LOCK		10
 
 /* in memory btrfs inode */
 struct btrfs_inode {
@@ -216,4 +218,22 @@ static inline int btrfs_inode_in_log(struct inode *inode, u64 generation)
 	return 0;
 }
 
+/*
+ * Disable DIO read nolock optimization, so new dio readers will be forced
+ * to grab i_mutex. It is used to avoid the endless truncate due to
+ * nonlocked dio read.
+ */
+static inline void btrfs_inode_block_unlocked_dio(struct inode *inode)
+{
+	set_bit(BTRFS_INODE_READDIO_NEED_LOCK, &BTRFS_I(inode)->runtime_flags);
+	smp_mb();
+}
+
+static inline void btrfs_inode_resume_unlocked_dio(struct inode *inode)
+{
+	smp_mb__before_clear_bit();
+	clear_bit(BTRFS_INODE_READDIO_NEED_LOCK,
+		  &BTRFS_I(inode)->runtime_flags);
+}
+
 #endif

fs/btrfs/check-integrity.c

@@ -813,8 +813,7 @@ static int btrfsic_process_superblock_dev_mirror(
 	    (bh->b_data + (dev_bytenr & 4095));
 
 	if (btrfs_super_bytenr(super_tmp) != dev_bytenr ||
-	    strncmp((char *)(&(super_tmp->magic)), BTRFS_MAGIC,
-		    sizeof(super_tmp->magic)) ||
+	    super_tmp->magic != cpu_to_le64(BTRFS_MAGIC) ||
 	    memcmp(device->uuid, super_tmp->dev_item.uuid, BTRFS_UUID_SIZE) ||
 	    btrfs_super_nodesize(super_tmp) != state->metablock_size ||
 	    btrfs_super_leafsize(super_tmp) != state->metablock_size ||

fs/btrfs/compression.c

@@ -372,7 +372,7 @@ int btrfs_submit_compressed_write(struct inode *inode, u64 start,
 		page = compressed_pages[pg_index];
 		page->mapping = inode->i_mapping;
 		if (bio->bi_size)
-			ret = io_tree->ops->merge_bio_hook(page, 0,
+			ret = io_tree->ops->merge_bio_hook(WRITE, page, 0,
 							   PAGE_CACHE_SIZE,
 							   bio, 0);
 		else
@@ -655,7 +655,7 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio,
 		page->index = em_start >> PAGE_CACHE_SHIFT;
 
 		if (comp_bio->bi_size)
-			ret = tree->ops->merge_bio_hook(page, 0,
+			ret = tree->ops->merge_bio_hook(READ, page, 0,
 							PAGE_CACHE_SIZE,
 							comp_bio, 0);
 		else

fs/btrfs/ctree.c

@@ -1138,6 +1138,7 @@ __tree_mod_log_rewind(struct extent_buffer *eb, u64 time_seq,
 		switch (tm->op) {
 		case MOD_LOG_KEY_REMOVE_WHILE_FREEING:
 			BUG_ON(tm->slot < n);
+			/* Fallthrough */
 		case MOD_LOG_KEY_REMOVE_WHILE_MOVING:
 		case MOD_LOG_KEY_REMOVE:
 			btrfs_set_node_key(eb, &tm->key, tm->slot);
@@ -1222,7 +1223,7 @@ tree_mod_log_rewind(struct btrfs_fs_info *fs_info, struct extent_buffer *eb,
 
 	__tree_mod_log_rewind(eb_rewin, time_seq, tm);
 	WARN_ON(btrfs_header_nritems(eb_rewin) >
-		BTRFS_NODEPTRS_PER_BLOCK(fs_info->fs_root));
+		BTRFS_NODEPTRS_PER_BLOCK(fs_info->tree_root));
 
 	return eb_rewin;
 }
@@ -1441,7 +1442,7 @@ int btrfs_comp_cpu_keys(struct btrfs_key *k1, struct btrfs_key *k2)
  */
 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,
+		       int start_slot, u64 *last_ret,
 		       struct btrfs_key *progress)
 {
 	struct extent_buffer *cur;
@@ -1461,8 +1462,6 @@ int btrfs_realloc_node(struct btrfs_trans_handle *trans,
 	struct btrfs_disk_key disk_key;
 
 	parent_level = btrfs_header_level(parent);
-	if (cache_only && parent_level != 1)
-		return 0;
 
 	WARN_ON(trans->transaction != root->fs_info->running_transaction);
 	WARN_ON(trans->transid != root->fs_info->generation);
@@ -1508,10 +1507,6 @@ int btrfs_realloc_node(struct btrfs_trans_handle *trans,
 		else
 			uptodate = 0;
 		if (!cur || !uptodate) {
-			if (cache_only) {
-				free_extent_buffer(cur);
-				continue;
-			}
 			if (!cur) {
 				cur = read_tree_block(root, blocknr,
 							 blocksize, gen);
@@ -4825,8 +4820,8 @@ int btrfs_prev_leaf(struct btrfs_root *root, struct btrfs_path *path)
 
 /*
  * A helper function to walk down the tree starting at min_key, and looking
- * for nodes or leaves that are either in cache or have a minimum
- * transaction id.  This is used by the btree defrag code, and tree logging
+ * for nodes or leaves that are have a minimum transaction id.
+ * This is used by the btree defrag code, and tree logging
  *
  * This does not cow, but it does stuff the starting key it finds back
  * into min_key, so you can call btrfs_search_slot with cow=1 on the
@@ -4847,7 +4842,7 @@ int btrfs_prev_leaf(struct btrfs_root *root, struct btrfs_path *path)
  */
 int btrfs_search_forward(struct btrfs_root *root, struct btrfs_key *min_key,
 			 struct btrfs_key *max_key,
-			 struct btrfs_path *path, int cache_only,
+			 struct btrfs_path *path,
 			 u64 min_trans)
 {
 	struct extent_buffer *cur;
@@ -4887,15 +4882,12 @@ int btrfs_search_forward(struct btrfs_root *root, struct btrfs_key *min_key,
 		if (sret && slot > 0)
 			slot--;
 		/*
-		 * check this node pointer against the cache_only and
-		 * min_trans parameters.  If it isn't in cache or is too
-		 * old, skip to the next one.
+		 * check this node pointer against the min_trans parameters.
+		 * If it is too old, old, skip to the next one.
 		 */
 		while (slot < nritems) {
 			u64 blockptr;
 			u64 gen;
-			struct extent_buffer *tmp;
-			struct btrfs_disk_key disk_key;
 
 			blockptr = btrfs_node_blockptr(cur, slot);
 			gen = btrfs_node_ptr_generation(cur, slot);
@@ -4903,27 +4895,7 @@ int btrfs_search_forward(struct btrfs_root *root, struct btrfs_key *min_key,
 				slot++;
 				continue;
 			}
-			if (!cache_only)
-				break;
-
-			if (max_key) {
-				btrfs_node_key(cur, &disk_key, slot);
-				if (comp_keys(&disk_key, max_key) >= 0) {
-					ret = 1;
-					goto out;
-				}
-			}
-
-			tmp = btrfs_find_tree_block(root, blockptr,
-					    btrfs_level_size(root, level - 1));
-
-			if (tmp && btrfs_buffer_uptodate(tmp, gen, 1) > 0) {
-				free_extent_buffer(tmp);
-				break;
-			}
-			if (tmp)
-				free_extent_buffer(tmp);
-			slot++;
+			break;
 		}
 find_next_key:
 		/*
@@ -4934,7 +4906,7 @@ int btrfs_search_forward(struct btrfs_root *root, struct btrfs_key *min_key,
 			path->slots[level] = slot;
 			btrfs_set_path_blocking(path);
 			sret = btrfs_find_next_key(root, path, min_key, level,
-						  cache_only, min_trans);
+						  min_trans);
 			if (sret == 0) {
 				btrfs_release_path(path);
 				goto again;
@@ -5399,8 +5371,7 @@ int btrfs_compare_trees(struct btrfs_root *left_root,
 /*
  * this is similar to btrfs_next_leaf, but does not try to preserve
  * and fixup the path.  It looks for and returns the next key in the
- * tree based on the current path and the cache_only and min_trans
- * parameters.
+ * tree based on the current path and the min_trans parameters.
  *
  * 0 is returned if another key is found, < 0 if there are any errors
  * and 1 is returned if there are no higher keys in the tree
@@ -5409,8 +5380,7 @@ int btrfs_compare_trees(struct btrfs_root *left_root,
  * calling this function.
  */
 int btrfs_find_next_key(struct btrfs_root *root, struct btrfs_path *path,
-			struct btrfs_key *key, int level,
-			int cache_only, u64 min_trans)
+			struct btrfs_key *key, int level, u64 min_trans)
 {
 	int slot;
 	struct extent_buffer *c;
@@ -5461,22 +5431,8 @@ int btrfs_find_next_key(struct btrfs_root *root, struct btrfs_path *path,
 		if (level == 0)
 			btrfs_item_key_to_cpu(c, key, slot);
 		else {
-			u64 blockptr = btrfs_node_blockptr(c, slot);
 			u64 gen = btrfs_node_ptr_generation(c, slot);
 
-			if (cache_only) {
-				struct extent_buffer *cur;
-				cur = btrfs_find_tree_block(root, blockptr,
-					    btrfs_level_size(root, level - 1));
-				if (!cur ||
-				    btrfs_buffer_uptodate(cur, gen, 1) <= 0) {
-					slot++;
-					if (cur)
-						free_extent_buffer(cur);
-					goto next;
-				}
-				free_extent_buffer(cur);
-			}
 			if (gen < min_trans) {
 				slot++;
 				goto next;

fs/btrfs/ctree.h

@@ -31,10 +31,10 @@
 #include <trace/events/btrfs.h>
 #include <asm/kmap_types.h>
 #include <linux/pagemap.h>
+#include <linux/btrfs.h>
 #include "extent_io.h"
 #include "extent_map.h"
 #include "async-thread.h"
-#include "ioctl.h"
 
 struct btrfs_trans_handle;
 struct btrfs_transaction;
@@ -46,7 +46,7 @@ extern struct kmem_cache *btrfs_path_cachep;
 extern struct kmem_cache *btrfs_free_space_cachep;
 struct btrfs_ordered_sum;
 
-#define BTRFS_MAGIC "_BHRfS_M"
+#define BTRFS_MAGIC 0x4D5F53665248425FULL /* ascii _BHRfS_M, no null */
 
 #define BTRFS_MAX_MIRRORS 3
 
@@ -191,6 +191,8 @@ static int btrfs_csum_sizes[] = { 4, 0 };
 /* ioprio of readahead is set to idle */
 #define BTRFS_IOPRIO_READA (IOPRIO_PRIO_VALUE(IOPRIO_CLASS_IDLE, 0))
 
+#define BTRFS_DIRTY_METADATA_THRESH	(32 * 1024 * 1024)
+
 /*
  * The key defines the order in the tree, and so it also defines (optimal)
  * block layout.
@@ -336,7 +338,10 @@ static inline unsigned long btrfs_chunk_item_size(int num_stripes)
 /*
  * File system states
  */
+#define BTRFS_FS_STATE_ERROR		0
+#define BTRFS_FS_STATE_REMOUNTING	1
 
+/* Super block flags */
 /* Errors detected */
 #define BTRFS_SUPER_FLAG_ERROR		(1ULL << 2)
 
@@ -502,6 +507,7 @@ struct btrfs_super_block {
 #define BTRFS_FEATURE_INCOMPAT_BIG_METADATA	(1ULL << 5)
 
 #define BTRFS_FEATURE_INCOMPAT_EXTENDED_IREF	(1ULL << 6)
+#define BTRFS_FEATURE_INCOMPAT_RAID56		(1ULL << 7)
 
 #define BTRFS_FEATURE_COMPAT_SUPP		0ULL
 #define BTRFS_FEATURE_COMPAT_RO_SUPP		0ULL
@@ -511,6 +517,7 @@ struct btrfs_super_block {
 	 BTRFS_FEATURE_INCOMPAT_MIXED_GROUPS |		\
 	 BTRFS_FEATURE_INCOMPAT_BIG_METADATA |		\
 	 BTRFS_FEATURE_INCOMPAT_COMPRESS_LZO |		\
+	 BTRFS_FEATURE_INCOMPAT_RAID56 |		\
 	 BTRFS_FEATURE_INCOMPAT_EXTENDED_IREF)
 
 /*
@@ -952,8 +959,20 @@ struct btrfs_dev_replace_item {
 #define BTRFS_BLOCK_GROUP_RAID1		(1ULL << 4)
 #define BTRFS_BLOCK_GROUP_DUP		(1ULL << 5)
 #define BTRFS_BLOCK_GROUP_RAID10	(1ULL << 6)
+#define BTRFS_BLOCK_GROUP_RAID5    (1 << 7)
+#define BTRFS_BLOCK_GROUP_RAID6    (1 << 8)
 #define BTRFS_BLOCK_GROUP_RESERVED	BTRFS_AVAIL_ALLOC_BIT_SINGLE
-#define BTRFS_NR_RAID_TYPES		5
+
+enum btrfs_raid_types {
+	BTRFS_RAID_RAID10,
+	BTRFS_RAID_RAID1,
+	BTRFS_RAID_DUP,
+	BTRFS_RAID_RAID0,
+	BTRFS_RAID_SINGLE,
+	BTRFS_RAID_RAID5,
+	BTRFS_RAID_RAID6,
+	BTRFS_NR_RAID_TYPES
+};
 
 #define BTRFS_BLOCK_GROUP_TYPE_MASK	(BTRFS_BLOCK_GROUP_DATA |    \
 					 BTRFS_BLOCK_GROUP_SYSTEM |  \
@@ -961,6 +980,8 @@ struct btrfs_dev_replace_item {
 
 #define BTRFS_BLOCK_GROUP_PROFILE_MASK	(BTRFS_BLOCK_GROUP_RAID0 |   \
 					 BTRFS_BLOCK_GROUP_RAID1 |   \
+					 BTRFS_BLOCK_GROUP_RAID5 |   \
+					 BTRFS_BLOCK_GROUP_RAID6 |   \
 					 BTRFS_BLOCK_GROUP_DUP |     \
 					 BTRFS_BLOCK_GROUP_RAID10)
 /*
@@ -1185,6 +1206,10 @@ struct btrfs_block_group_cache {
 	u64 flags;
 	u64 sectorsize;
 	u64 cache_generation;
+
+	/* for raid56, this is a full stripe, without parity */
+	unsigned long full_stripe_len;
+
 	unsigned int ro:1;
 	unsigned int dirty:1;
 	unsigned int iref:1;
@@ -1225,6 +1250,28 @@ struct seq_list {
 	u64 seq;
 };
 
+enum btrfs_orphan_cleanup_state {
+	ORPHAN_CLEANUP_STARTED	= 1,
+	ORPHAN_CLEANUP_DONE	= 2,
+};
+
+/* used by the raid56 code to lock stripes for read/modify/write */
+struct btrfs_stripe_hash {
+	struct list_head hash_list;
+	wait_queue_head_t wait;
+	spinlock_t lock;
+};
+
+/* used by the raid56 code to lock stripes for read/modify/write */
+struct btrfs_stripe_hash_table {
+	struct list_head stripe_cache;
+	spinlock_t cache_lock;
+	int cache_size;
+	struct btrfs_stripe_hash table[];
+};
+
+#define BTRFS_STRIPE_HASH_TABLE_BITS 11
+
 /* fs_info */
 struct reloc_control;
 struct btrfs_device;
@@ -1250,6 +1297,7 @@ struct btrfs_fs_info {
 
 	/* block group cache stuff */
 	spinlock_t block_group_cache_lock;
+	u64 first_logical_byte;
 	struct rb_root block_group_cache_tree;
 
 	/* keep track of unallocated space */
@@ -1288,7 +1336,23 @@ struct btrfs_fs_info {
 	u64 last_trans_log_full_commit;
 	unsigned long mount_opt;
 	unsigned long compress_type:4;
+	/*
+	 * It is a suggestive number, the read side is safe even it gets a
+	 * wrong number because we will write out the data into a regular
+	 * extent. The write side(mount/remount) is under ->s_umount lock,
+	 * so it is also safe.
+	 */
 	u64 max_inline;
+	/*
+	 * Protected by ->chunk_mutex and sb->s_umount.
+	 *
+	 * The reason that we use two lock to protect it is because only
+	 * remount and mount operations can change it and these two operations
+	 * are under sb->s_umount, but the read side (chunk allocation) can not
+	 * acquire sb->s_umount or the deadlock would happen. So we use two
+	 * locks to protect it. On the write side, we must acquire two locks,
+	 * and on the read side, we just need acquire one of them.
+	 */
 	u64 alloc_start;
 	struct btrfs_transaction *running_transaction;
 	wait_queue_head_t transaction_throttle;
@@ -1307,6 +1371,13 @@ struct btrfs_fs_info {
 	struct mutex cleaner_mutex;
 	struct mutex chunk_mutex;
 	struct mutex volume_mutex;
+
+	/* this is used during read/modify/write to make sure
+	 * no two ios are trying to mod the same stripe at the same
+	 * time
+	 */
+	struct btrfs_stripe_hash_table *stripe_hash_table;
+
 	/*
 	 * this protects the ordered operations list only while we are
 	 * processing all of the entries on it.  This way we make
@@ -1365,6 +1436,7 @@ struct btrfs_fs_info {
 	 */
 	struct list_head ordered_extents;
 
+	spinlock_t delalloc_lock;
 	/*
 	 * all of the inodes that have delalloc bytes.  It is possible for
 	 * this list to be empty even when there is still dirty data=ordered
@@ -1372,13 +1444,6 @@ struct btrfs_fs_info {
 	 */
 	struct list_head delalloc_inodes;
 
-	/*
-	 * special rename and truncate targets that must be on disk before
-	 * we're allowed to commit.  This is basically the ext3 style
-	 * data=ordered list.
-	 */
-	struct list_head ordered_operations;
-
 	/*
 	 * there is a pool of worker threads for checksumming during writes
 	 * and a pool for checksumming after reads.  This is because readers
@@ -1395,6 +1460,8 @@ struct btrfs_fs_info {
 	struct btrfs_workers flush_workers;
 	struct btrfs_workers endio_workers;
 	struct btrfs_workers endio_meta_workers;
+	struct btrfs_workers endio_raid56_workers;
+	struct btrfs_workers rmw_workers;
 	struct btrfs_workers endio_meta_write_workers;
 	struct btrfs_workers endio_write_workers;
 	struct btrfs_workers endio_freespace_worker;
@@ -1423,10 +1490,12 @@ struct btrfs_fs_info {
 
 	u64 total_pinned;
 
-	/* protected by the delalloc lock, used to keep from writing
-	 * metadata until there is a nice batch
-	 */
-	u64 dirty_metadata_bytes;
+	/* used to keep from writing metadata until there is a nice batch */
+	struct percpu_counter dirty_metadata_bytes;
+	struct percpu_counter delalloc_bytes;
+	s32 dirty_metadata_batch;
+	s32 delalloc_batch;
+
 	struct list_head dirty_cowonly_roots;
 
 	struct btrfs_fs_devices *fs_devices;
@@ -1442,9 +1511,6 @@ struct btrfs_fs_info {
 
 	struct reloc_control *reloc_ctl;
 
-	spinlock_t delalloc_lock;
-	u64 delalloc_bytes;
-
 	/* data_alloc_cluster is only used in ssd mode */
 	struct btrfs_free_cluster data_alloc_cluster;
 
@@ -1456,6 +1522,8 @@ struct btrfs_fs_info {
 	struct rb_root defrag_inodes;
 	atomic_t defrag_running;
 
+	/* Used to protect avail_{data, metadata, system}_alloc_bits */
+	seqlock_t profiles_lock;
 	/*
 	 * these three are in extended format (availability of single
 	 * chunks is denoted by BTRFS_AVAIL_ALLOC_BIT_SINGLE bit, other
@@ -1520,7 +1588,7 @@ struct btrfs_fs_info {
 	u64 qgroup_seq;
 
 	/* filesystem state */
-	u64 fs_state;
+	unsigned long fs_state;
 
 	struct btrfs_delayed_root *delayed_root;
 
@@ -1623,6 +1691,9 @@ struct btrfs_root {
 
 	struct list_head root_list;
 
+	spinlock_t log_extents_lock[2];
+	struct list_head logged_list[2];
+
 	spinlock_t orphan_lock;
 	atomic_t orphan_inodes;
 	struct btrfs_block_rsv *orphan_block_rsv;
@@ -1832,6 +1903,7 @@ struct btrfs_ioctl_defrag_range_args {
 
 #define btrfs_clear_opt(o, opt)		((o) &= ~BTRFS_MOUNT_##opt)
 #define btrfs_set_opt(o, opt)		((o) |= BTRFS_MOUNT_##opt)
+#define btrfs_raw_test_opt(o, opt)	((o) & BTRFS_MOUNT_##opt)
 #define btrfs_test_opt(root, opt)	((root)->fs_info->mount_opt & \
 					 BTRFS_MOUNT_##opt)
 /*
@@ -2936,8 +3008,7 @@ 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_pin_extent_for_log_replay(struct btrfs_trans_handle *trans,
-				    struct btrfs_root *root,
+int btrfs_pin_extent_for_log_replay(struct btrfs_root *root,
 				    u64 bytenr, u64 num_bytes);
 int btrfs_cross_ref_exist(struct btrfs_trans_handle *trans,
 			  struct btrfs_root *root,
@@ -3035,8 +3106,13 @@ void btrfs_trans_release_metadata(struct btrfs_trans_handle *trans,
 int btrfs_orphan_reserve_metadata(struct btrfs_trans_handle *trans,
 				  struct inode *inode);
 void btrfs_orphan_release_metadata(struct inode *inode);
-int btrfs_snap_reserve_metadata(struct btrfs_trans_handle *trans,
-				struct btrfs_pending_snapshot *pending);
+int btrfs_subvolume_reserve_metadata(struct btrfs_root *root,
+				     struct btrfs_block_rsv *rsv,
+				     int nitems,
+				     u64 *qgroup_reserved);
+void btrfs_subvolume_release_metadata(struct btrfs_root *root,
+				      struct btrfs_block_rsv *rsv,
+				      u64 qgroup_reserved);
 int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes);
 void btrfs_delalloc_release_metadata(struct inode *inode, u64 num_bytes);
 int btrfs_delalloc_reserve_space(struct inode *inode, u64 num_bytes);
@@ -3092,10 +3168,10 @@ struct extent_buffer *btrfs_root_node(struct btrfs_root *root);
 struct extent_buffer *btrfs_lock_root_node(struct btrfs_root *root);
 int btrfs_find_next_key(struct btrfs_root *root, struct btrfs_path *path,
 			struct btrfs_key *key, int lowest_level,
-			int cache_only, u64 min_trans);
+			u64 min_trans);
 int btrfs_search_forward(struct btrfs_root *root, struct btrfs_key *min_key,
 			 struct btrfs_key *max_key,
-			 struct btrfs_path *path, int cache_only,
+			 struct btrfs_path *path,
 			 u64 min_trans);
 enum btrfs_compare_tree_result {
 	BTRFS_COMPARE_TREE_NEW,
@@ -3148,7 +3224,7 @@ int btrfs_search_slot_for_read(struct btrfs_root *root,
 			       int find_higher, int return_any);
 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,
+		       int start_slot, u64 *last_ret,
 		       struct btrfs_key *progress);
 void btrfs_release_path(struct btrfs_path *p);
 struct btrfs_path *btrfs_alloc_path(void);
@@ -3459,9 +3535,9 @@ int btrfs_writepages(struct address_space *mapping,
 		     struct writeback_control *wbc);
 int btrfs_create_subvol_root(struct btrfs_trans_handle *trans,
 			     struct btrfs_root *new_root, u64 new_dirid);
-int btrfs_merge_bio_hook(struct page *page, unsigned long offset,
-			 size_t size, struct bio *bio, unsigned long bio_flags);
-
+int btrfs_merge_bio_hook(int rw, struct page *page, unsigned long offset,
+			 size_t size, struct bio *bio,
+			 unsigned long bio_flags);
 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);
@@ -3543,7 +3619,7 @@ int btrfs_dirty_pages(struct btrfs_root *root, struct inode *inode,
 
 /* tree-defrag.c */
 int btrfs_defrag_leaves(struct btrfs_trans_handle *trans,
-			struct btrfs_root *root, int cache_only);
+			struct btrfs_root *root);
 
 /* sysfs.c */
 int btrfs_init_sysfs(void);
@@ -3620,11 +3696,14 @@ __printf(5, 6)
 void __btrfs_panic(struct btrfs_fs_info *fs_info, const char *function,
 		   unsigned int line, int errno, const char *fmt, ...);
 
+/*
+ * If BTRFS_MOUNT_PANIC_ON_FATAL_ERROR is in mount_opt, __btrfs_panic
+ * will panic().  Otherwise we BUG() here.
+ */
 #define btrfs_panic(fs_info, errno, fmt, args...)			\
 do {									\
-	struct btrfs_fs_info *_i = (fs_info);				\
-	__btrfs_panic(_i, __func__, __LINE__, errno, fmt, ##args);	\
-	BUG_ON(!(_i->mount_opt & BTRFS_MOUNT_PANIC_ON_FATAL_ERROR));	\
+	__btrfs_panic(fs_info, __func__, __LINE__, errno, fmt, ##args);	\
+	BUG();								\
 } while (0)
 
 /* acl.c */
@@ -3745,4 +3824,11 @@ static inline int is_fstree(u64 rootid)
 		return 1;
 	return 0;
 }
+
+static inline int btrfs_defrag_cancelled(struct btrfs_fs_info *fs_info)
+{
+	return signal_pending(current);
+}
+
+
 #endif

fs/btrfs/delayed-inode.c

@@ -875,7 +875,6 @@ static int btrfs_insert_delayed_item(struct btrfs_trans_handle *trans,
 				     struct btrfs_delayed_item *delayed_item)
 {
 	struct extent_buffer *leaf;
-	struct btrfs_item *item;
 	char *ptr;
 	int ret;
 
@@ -886,7 +885,6 @@ static int btrfs_insert_delayed_item(struct btrfs_trans_handle *trans,
 
 	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,
@@ -1065,32 +1063,25 @@ static void btrfs_release_delayed_inode(struct btrfs_delayed_node *delayed_node)
 	}
 }
 
-static int btrfs_update_delayed_inode(struct btrfs_trans_handle *trans,
-				      struct btrfs_root *root,
-				      struct btrfs_path *path,
-				      struct btrfs_delayed_node *node)
+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;
 	}
 
@@ -1105,11 +1096,47 @@ static int btrfs_update_delayed_inode(struct btrfs_trans_handle *trans,
 
 	btrfs_delayed_inode_release_metadata(root, node);
 	btrfs_release_delayed_inode(node);
-	mutex_unlock(&node->mutex);
 
 	return 0;
 }
 
+static inline int btrfs_update_delayed_inode(struct btrfs_trans_handle *trans,
+					     struct btrfs_root *root,
+					     struct btrfs_path *path,
+					     struct btrfs_delayed_node *node)
+{
+	int ret;
+
+	mutex_lock(&node->mutex);
+	if (!node->inode_dirty) {
+		mutex_unlock(&node->mutex);
+		return 0;
+	}
+
+	ret = __btrfs_update_delayed_inode(trans, root, path, node);
+	mutex_unlock(&node->mutex);
+	return ret;
+}
+
+static inline int
+__btrfs_commit_inode_delayed_items(struct btrfs_trans_handle *trans,
+				   struct btrfs_path *path,
+				   struct btrfs_delayed_node *node)
+{
+	int ret;
+
+	ret = btrfs_insert_delayed_items(trans, path, node->root, node);
+	if (ret)
+		return ret;
+
+	ret = btrfs_delete_delayed_items(trans, path, node->root, node);
+	if (ret)
+		return ret;
+
+	ret = btrfs_update_delayed_inode(trans, node->root, path, node);
+	return ret;
+}
+
 /*
  * Called when committing the transaction.
  * Returns 0 on success.
@@ -1119,7 +1146,6 @@ static int btrfs_update_delayed_inode(struct btrfs_trans_handle *trans,
 static int __btrfs_run_delayed_items(struct btrfs_trans_handle *trans,
 				     struct btrfs_root *root, int nr)
 {
-	struct btrfs_root *curr_root = root;
 	struct btrfs_delayed_root *delayed_root;
 	struct btrfs_delayed_node *curr_node, *prev_node;
 	struct btrfs_path *path;
@@ -1142,15 +1168,8 @@ static int __btrfs_run_delayed_items(struct btrfs_trans_handle *trans,
 
 	curr_node = btrfs_first_delayed_node(delayed_root);
 	while (curr_node && (!count || (count && nr--))) {
-		curr_root = curr_node->root;
-		ret = btrfs_insert_delayed_items(trans, path, curr_root,
-						 curr_node);
-		if (!ret)
-			ret = btrfs_delete_delayed_items(trans, path,
-						curr_root, curr_node);
-		if (!ret)
-			ret = btrfs_update_delayed_inode(trans, curr_root,
-						path, curr_node);
+		ret = __btrfs_commit_inode_delayed_items(trans, path,
+							 curr_node);
 		if (ret) {
 			btrfs_release_delayed_node(curr_node);
 			curr_node = NULL;
@@ -1183,51 +1202,93 @@ int btrfs_run_delayed_items_nr(struct btrfs_trans_handle *trans,
 	return __btrfs_run_delayed_items(trans, root, nr);
 }
 
-static int __btrfs_commit_inode_delayed_items(struct btrfs_trans_handle *trans,
-					      struct btrfs_delayed_node *node)
+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);
 	struct btrfs_path *path;
 	struct btrfs_block_rsv *block_rsv;
 	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);
+
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
 	path->leave_spinning = 1;
 
 	block_rsv = trans->block_rsv;
-	trans->block_rsv = &node->root->fs_info->delayed_block_rsv;
+	trans->block_rsv = &delayed_node->root->fs_info->delayed_block_rsv;
 
-	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);
+	ret = __btrfs_commit_inode_delayed_items(trans, path, delayed_node);
 
+	btrfs_release_delayed_node(delayed_node);
+	btrfs_free_path(path);
 	trans->block_rsv = block_rsv;
+
 	return ret;
 }
 
-int btrfs_commit_inode_delayed_items(struct btrfs_trans_handle *trans,
-				     struct inode *inode)
+int btrfs_commit_inode_delayed_inode(struct inode *inode)
 {
+	struct btrfs_trans_handle *trans;
 	struct btrfs_delayed_node *delayed_node = btrfs_get_delayed_node(inode);
+	struct btrfs_path *path;
+	struct btrfs_block_rsv *block_rsv;
 	int ret;
 
 	if (!delayed_node)
 		return 0;
 
 	mutex_lock(&delayed_node->mutex);
-	if (!delayed_node->count) {
+	if (!delayed_node->inode_dirty) {
 		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);
+	trans = btrfs_join_transaction(delayed_node->root);
+	if (IS_ERR(trans)) {
+		ret = PTR_ERR(trans);
+		goto out;
+	}
+
+	path = btrfs_alloc_path();
+	if (!path) {
+		ret = -ENOMEM;
+		goto trans_out;
+	}
+	path->leave_spinning = 1;
+
+	block_rsv = trans->block_rsv;
+	trans->block_rsv = &delayed_node->root->fs_info->delayed_block_rsv;
+
+	mutex_lock(&delayed_node->mutex);
+	if (delayed_node->inode_dirty)
+		ret = __btrfs_update_delayed_inode(trans, delayed_node->root,
+						   path, delayed_node);
+	else
+		ret = 0;
+	mutex_unlock(&delayed_node->mutex);
+
+	btrfs_free_path(path);
+	trans->block_rsv = block_rsv;
+trans_out:
+	btrfs_end_transaction(trans, delayed_node->root);
+	btrfs_btree_balance_dirty(delayed_node->root);
+out:
 	btrfs_release_delayed_node(delayed_node);
+
 	return ret;
 }
 
@@ -1258,7 +1319,6 @@ static void btrfs_async_run_delayed_node_done(struct btrfs_work *work)
 	struct btrfs_root *root;
 	struct btrfs_block_rsv *block_rsv;
 	int need_requeue = 0;
-	int ret;
 
 	async_node = container_of(work, struct btrfs_async_delayed_node, work);
 
@@ -1277,14 +1337,7 @@ static void btrfs_async_run_delayed_node_done(struct btrfs_work *work)
 	block_rsv = trans->block_rsv;
 	trans->block_rsv = &root->fs_info->delayed_block_rsv;
 
-	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);
-
+	__btrfs_commit_inode_delayed_items(trans, 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

fs/btrfs/delayed-inode.h

@@ -117,6 +117,7 @@ int btrfs_commit_inode_delayed_items(struct btrfs_trans_handle *trans,
 /* Used for evicting the inode. */
 void btrfs_remove_delayed_node(struct inode *inode);
 void btrfs_kill_delayed_inode_items(struct inode *inode);
+int btrfs_commit_inode_delayed_inode(struct inode *inode);
 
 
 int btrfs_delayed_update_inode(struct btrfs_trans_handle *trans,

fs/btrfs/delayed-ref.c

@@ -23,6 +23,10 @@
 #include "delayed-ref.h"
 #include "transaction.h"
 
+struct kmem_cache *btrfs_delayed_ref_head_cachep;
+struct kmem_cache *btrfs_delayed_tree_ref_cachep;
+struct kmem_cache *btrfs_delayed_data_ref_cachep;
+struct kmem_cache *btrfs_delayed_extent_op_cachep;
 /*
  * delayed back reference update tracking.  For subvolume trees
  * we queue up extent allocations and backref maintenance for
@@ -422,6 +426,14 @@ int btrfs_find_ref_cluster(struct btrfs_trans_handle *trans,
 	return 1;
 }
 
+void btrfs_release_ref_cluster(struct list_head *cluster)
+{
+	struct list_head *pos, *q;
+
+	list_for_each_safe(pos, q, cluster)
+		list_del_init(pos);
+}
+
 /*
  * helper function to update an extent delayed ref in the
  * rbtree.  existing and update must both have the same
@@ -511,7 +523,7 @@ update_existing_head_ref(struct btrfs_delayed_ref_node *existing,
 					ref->extent_op->flags_to_set;
 				existing_ref->extent_op->update_flags = 1;
 			}
-			kfree(ref->extent_op);
+			btrfs_free_delayed_extent_op(ref->extent_op);
 		}
 	}
 	/*
@@ -592,7 +604,7 @@ static noinline void add_delayed_ref_head(struct btrfs_fs_info *fs_info,
 		 * we've updated the existing ref, free the newly
 		 * allocated ref
 		 */
-		kfree(head_ref);
+		kmem_cache_free(btrfs_delayed_ref_head_cachep, head_ref);
 	} else {
 		delayed_refs->num_heads++;
 		delayed_refs->num_heads_ready++;
@@ -653,7 +665,7 @@ static noinline void add_delayed_tree_ref(struct btrfs_fs_info *fs_info,
 		 * we've updated the existing ref, free the newly
 		 * allocated ref
 		 */
-		kfree(full_ref);
+		kmem_cache_free(btrfs_delayed_tree_ref_cachep, full_ref);
 	} else {
 		delayed_refs->num_entries++;
 		trans->delayed_ref_updates++;
@@ -714,7 +726,7 @@ static noinline void add_delayed_data_ref(struct btrfs_fs_info *fs_info,
 		 * we've updated the existing ref, free the newly
 		 * allocated ref
 		 */
-		kfree(full_ref);
+		kmem_cache_free(btrfs_delayed_data_ref_cachep, full_ref);
 	} else {
 		delayed_refs->num_entries++;
 		trans->delayed_ref_updates++;
@@ -738,13 +750,13 @@ int btrfs_add_delayed_tree_ref(struct btrfs_fs_info *fs_info,
 	struct btrfs_delayed_ref_root *delayed_refs;
 
 	BUG_ON(extent_op && extent_op->is_data);
-	ref = kmalloc(sizeof(*ref), GFP_NOFS);
+	ref = kmem_cache_alloc(btrfs_delayed_tree_ref_cachep, GFP_NOFS);
 	if (!ref)
 		return -ENOMEM;
 
-	head_ref = kmalloc(sizeof(*head_ref), GFP_NOFS);
+	head_ref = kmem_cache_alloc(btrfs_delayed_ref_head_cachep, GFP_NOFS);
 	if (!head_ref) {
-		kfree(ref);
+		kmem_cache_free(btrfs_delayed_tree_ref_cachep, ref);
 		return -ENOMEM;
 	}
 
@@ -786,13 +798,13 @@ int btrfs_add_delayed_data_ref(struct btrfs_fs_info *fs_info,
 	struct btrfs_delayed_ref_root *delayed_refs;
 
 	BUG_ON(extent_op && !extent_op->is_data);
-	ref = kmalloc(sizeof(*ref), GFP_NOFS);
+	ref = kmem_cache_alloc(btrfs_delayed_data_ref_cachep, GFP_NOFS);
 	if (!ref)
 		return -ENOMEM;
 
-	head_ref = kmalloc(sizeof(*head_ref), GFP_NOFS);
+	head_ref = kmem_cache_alloc(btrfs_delayed_ref_head_cachep, GFP_NOFS);
 	if (!head_ref) {
-		kfree(ref);
+		kmem_cache_free(btrfs_delayed_data_ref_cachep, ref);
 		return -ENOMEM;
 	}
 
@@ -826,7 +838,7 @@ int btrfs_add_delayed_extent_op(struct btrfs_fs_info *fs_info,
 	struct btrfs_delayed_ref_head *head_ref;
 	struct btrfs_delayed_ref_root *delayed_refs;
 
-	head_ref = kmalloc(sizeof(*head_ref), GFP_NOFS);
+	head_ref = kmem_cache_alloc(btrfs_delayed_ref_head_cachep, GFP_NOFS);
 	if (!head_ref)
 		return -ENOMEM;
 
@@ -860,3 +872,51 @@ btrfs_find_delayed_ref_head(struct btrfs_trans_handle *trans, u64 bytenr)
 		return btrfs_delayed_node_to_head(ref);
 	return NULL;
 }
+
+void btrfs_delayed_ref_exit(void)
+{
+	if (btrfs_delayed_ref_head_cachep)
+		kmem_cache_destroy(btrfs_delayed_ref_head_cachep);
+	if (btrfs_delayed_tree_ref_cachep)
+		kmem_cache_destroy(btrfs_delayed_tree_ref_cachep);
+	if (btrfs_delayed_data_ref_cachep)
+		kmem_cache_destroy(btrfs_delayed_data_ref_cachep);
+	if (btrfs_delayed_extent_op_cachep)
+		kmem_cache_destroy(btrfs_delayed_extent_op_cachep);
+}
+
+int btrfs_delayed_ref_init(void)
+{
+	btrfs_delayed_ref_head_cachep = kmem_cache_create(
+				"btrfs_delayed_ref_head",
+				sizeof(struct btrfs_delayed_ref_head), 0,
+				SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL);
+	if (!btrfs_delayed_ref_head_cachep)
+		goto fail;
+
+	btrfs_delayed_tree_ref_cachep = kmem_cache_create(
+				"btrfs_delayed_tree_ref",
+				sizeof(struct btrfs_delayed_tree_ref), 0,
+				SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL);
+	if (!btrfs_delayed_tree_ref_cachep)
+		goto fail;
+
+	btrfs_delayed_data_ref_cachep = kmem_cache_create(
+				"btrfs_delayed_data_ref",
+				sizeof(struct btrfs_delayed_data_ref), 0,
+				SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL);
+	if (!btrfs_delayed_data_ref_cachep)
+		goto fail;
+
+	btrfs_delayed_extent_op_cachep = kmem_cache_create(
+				"btrfs_delayed_extent_op",
+				sizeof(struct btrfs_delayed_extent_op), 0,
+				SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL);
+	if (!btrfs_delayed_extent_op_cachep)
+		goto fail;
+
+	return 0;
+fail:
+	btrfs_delayed_ref_exit();
+	return -ENOMEM;
+}

fs/btrfs/delayed-ref.h

@@ -131,6 +131,15 @@ struct btrfs_delayed_ref_root {
 	/* total number of head nodes ready for processing */
 	unsigned long num_heads_ready;
 
+	/*
+	 * bumped when someone is making progress on the delayed
+	 * refs, so that other procs know they are just adding to
+	 * contention intead of helping
+	 */
+	atomic_t procs_running_refs;
+	atomic_t ref_seq;
+	wait_queue_head_t wait;
+
 	/*
 	 * set when the tree is flushing before a transaction commit,
 	 * used by the throttling code to decide if new updates need
@@ -141,12 +150,47 @@ struct btrfs_delayed_ref_root {
 	u64 run_delayed_start;
 };
 
+extern struct kmem_cache *btrfs_delayed_ref_head_cachep;
+extern struct kmem_cache *btrfs_delayed_tree_ref_cachep;
+extern struct kmem_cache *btrfs_delayed_data_ref_cachep;
+extern struct kmem_cache *btrfs_delayed_extent_op_cachep;
+
+int btrfs_delayed_ref_init(void);
+void btrfs_delayed_ref_exit(void);
+
+static inline struct btrfs_delayed_extent_op *
+btrfs_alloc_delayed_extent_op(void)
+{
+	return kmem_cache_alloc(btrfs_delayed_extent_op_cachep, GFP_NOFS);
+}
+
+static inline void
+btrfs_free_delayed_extent_op(struct btrfs_delayed_extent_op *op)
+{
+	if (op)
+		kmem_cache_free(btrfs_delayed_extent_op_cachep, op);
+}
+
 static inline void btrfs_put_delayed_ref(struct btrfs_delayed_ref_node *ref)
 {
 	WARN_ON(atomic_read(&ref->refs) == 0);
 	if (atomic_dec_and_test(&ref->refs)) {
 		WARN_ON(ref->in_tree);
-		kfree(ref);
+		switch (ref->type) {
+		case BTRFS_TREE_BLOCK_REF_KEY:
+		case BTRFS_SHARED_BLOCK_REF_KEY:
+			kmem_cache_free(btrfs_delayed_tree_ref_cachep, ref);
+			break;
+		case BTRFS_EXTENT_DATA_REF_KEY:
+		case BTRFS_SHARED_DATA_REF_KEY:
+			kmem_cache_free(btrfs_delayed_data_ref_cachep, ref);
+			break;
+		case 0:
+			kmem_cache_free(btrfs_delayed_ref_head_cachep, ref);
+			break;
+		default:
+			BUG();
+		}
 	}
 }
 
@@ -176,8 +220,14 @@ struct btrfs_delayed_ref_head *
 btrfs_find_delayed_ref_head(struct btrfs_trans_handle *trans, u64 bytenr);
 int btrfs_delayed_ref_lock(struct btrfs_trans_handle *trans,
 			   struct btrfs_delayed_ref_head *head);
+static inline void btrfs_delayed_ref_unlock(struct btrfs_delayed_ref_head *head)
+{
+	mutex_unlock(&head->mutex);
+}
+
 int btrfs_find_ref_cluster(struct btrfs_trans_handle *trans,
 			   struct list_head *cluster, u64 search_start);
+void btrfs_release_ref_cluster(struct list_head *cluster);
 
 int btrfs_check_delayed_seq(struct btrfs_fs_info *fs_info,
 			    struct btrfs_delayed_ref_root *delayed_refs,

fs/btrfs/dev-replace.c

@@ -465,7 +465,11 @@ static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
 	 * flush all outstanding I/O and inode extent mappings before the
 	 * copy operation is declared as being finished
 	 */
-	btrfs_start_delalloc_inodes(root, 0);
+	ret = btrfs_start_delalloc_inodes(root, 0);
+	if (ret) {
+		mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
+		return ret;
+	}
 	btrfs_wait_ordered_extents(root, 0);
 
 	trans = btrfs_start_transaction(root, 0);

fs/btrfs/disk-io.c

@@ -46,6 +46,7 @@
 #include "check-integrity.h"
 #include "rcu-string.h"
 #include "dev-replace.h"
+#include "raid56.h"
 
 #ifdef CONFIG_X86
 #include <asm/cpufeature.h>
@@ -56,7 +57,8 @@ static void end_workqueue_fn(struct btrfs_work *work);
 static void free_fs_root(struct btrfs_root *root);
 static int btrfs_check_super_valid(struct btrfs_fs_info *fs_info,
 				    int read_only);
-static void btrfs_destroy_ordered_operations(struct btrfs_root *root);
+static void btrfs_destroy_ordered_operations(struct btrfs_transaction *t,
+					     struct btrfs_root *root);
 static void btrfs_destroy_ordered_extents(struct btrfs_root *root);
 static int btrfs_destroy_delayed_refs(struct btrfs_transaction *trans,
 				      struct btrfs_root *root);
@@ -420,7 +422,7 @@ static int btree_read_extent_buffer_pages(struct btrfs_root *root,
 static int csum_dirty_buffer(struct btrfs_root *root, struct page *page)
 {
 	struct extent_io_tree *tree;
-	u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
+	u64 start = page_offset(page);
 	u64 found_start;
 	struct extent_buffer *eb;
 
@@ -639,8 +641,15 @@ static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end,
 		btree_readahead_hook(root, eb, eb->start, ret);
 	}
 
-	if (ret)
+	if (ret) {
+		/*
+		 * our io error hook is going to dec the io pages
+		 * again, we have to make sure it has something
+		 * to decrement
+		 */
+		atomic_inc(&eb->io_pages);
 		clear_extent_buffer_uptodate(eb);
+	}
 	free_extent_buffer(eb);
 out:
 	return ret;
@@ -654,6 +663,7 @@ static int btree_io_failed_hook(struct page *page, int failed_mirror)
 	eb = (struct extent_buffer *)page->private;
 	set_bit(EXTENT_BUFFER_IOERR, &eb->bflags);
 	eb->read_mirror = failed_mirror;
+	atomic_dec(&eb->io_pages);
 	if (test_and_clear_bit(EXTENT_BUFFER_READAHEAD, &eb->bflags))
 		btree_readahead_hook(root, eb, eb->start, -EIO);
 	return -EIO;	/* we fixed nothing */
@@ -670,17 +680,23 @@ static void end_workqueue_bio(struct bio *bio, int err)
 	end_io_wq->work.flags = 0;
 
 	if (bio->bi_rw & REQ_WRITE) {
-		if (end_io_wq->metadata == 1)
+		if (end_io_wq->metadata == BTRFS_WQ_ENDIO_METADATA)
 			btrfs_queue_worker(&fs_info->endio_meta_write_workers,
 					   &end_io_wq->work);
-		else if (end_io_wq->metadata == 2)
+		else if (end_io_wq->metadata == BTRFS_WQ_ENDIO_FREE_SPACE)
 			btrfs_queue_worker(&fs_info->endio_freespace_worker,
 					   &end_io_wq->work);
+		else if (end_io_wq->metadata == BTRFS_WQ_ENDIO_RAID56)
+			btrfs_queue_worker(&fs_info->endio_raid56_workers,
+					   &end_io_wq->work);
 		else
 			btrfs_queue_worker(&fs_info->endio_write_workers,
 					   &end_io_wq->work);
 	} else {
-		if (end_io_wq->metadata)
+		if (end_io_wq->metadata == BTRFS_WQ_ENDIO_RAID56)
+			btrfs_queue_worker(&fs_info->endio_raid56_workers,
+					   &end_io_wq->work);
+		else if (end_io_wq->metadata)
 			btrfs_queue_worker(&fs_info->endio_meta_workers,
 					   &end_io_wq->work);
 		else
@@ -695,6 +711,7 @@ static void end_workqueue_bio(struct bio *bio, int err)
  * 0 - if data
  * 1 - if normal metadta
  * 2 - if writing to the free space cache area
+ * 3 - raid parity work
  */
 int btrfs_bio_wq_end_io(struct btrfs_fs_info *info, struct bio *bio,
 			int metadata)
@@ -946,18 +963,20 @@ static int btree_writepages(struct address_space *mapping,
 			    struct writeback_control *wbc)
 {
 	struct extent_io_tree *tree;
+	struct btrfs_fs_info *fs_info;
+	int ret;
+
 	tree = &BTRFS_I(mapping->host)->io_tree;
 	if (wbc->sync_mode == WB_SYNC_NONE) {
-		struct btrfs_root *root = BTRFS_I(mapping->host)->root;
-		u64 num_dirty;
-		unsigned long thresh = 32 * 1024 * 1024;
 
 		if (wbc->for_kupdate)
 			return 0;
 
+		fs_info = BTRFS_I(mapping->host)->root->fs_info;
 		/* this is a bit racy, but that's ok */
-		num_dirty = root->fs_info->dirty_metadata_bytes;
-		if (num_dirty < thresh)
+		ret = percpu_counter_compare(&fs_info->dirty_metadata_bytes,
+					     BTRFS_DIRTY_METADATA_THRESH);
+		if (ret < 0)
 			return 0;
 	}
 	return btree_write_cache_pages(mapping, wbc);
@@ -1125,24 +1144,16 @@ struct extent_buffer *read_tree_block(struct btrfs_root *root, u64 bytenr,
 void clean_tree_block(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 		      struct extent_buffer *buf)
 {
+	struct btrfs_fs_info *fs_info = root->fs_info;
+
 	if (btrfs_header_generation(buf) ==
-	    root->fs_info->running_transaction->transid) {
+	    fs_info->running_transaction->transid) {
 		btrfs_assert_tree_locked(buf);
 
 		if (test_and_clear_bit(EXTENT_BUFFER_DIRTY, &buf->bflags)) {
-			spin_lock(&root->fs_info->delalloc_lock);
-			if (root->fs_info->dirty_metadata_bytes >= buf->len)
-				root->fs_info->dirty_metadata_bytes -= buf->len;
-			else {
-				spin_unlock(&root->fs_info->delalloc_lock);
-				btrfs_panic(root->fs_info, -EOVERFLOW,
-					  "Can't clear %lu bytes from "
-					  " dirty_mdatadata_bytes (%llu)",
-					  buf->len,
-					  root->fs_info->dirty_metadata_bytes);
-			}
-			spin_unlock(&root->fs_info->delalloc_lock);
-
+			__percpu_counter_add(&fs_info->dirty_metadata_bytes,
+					     -buf->len,
+					     fs_info->dirty_metadata_batch);
 			/* ugh, clear_extent_buffer_dirty needs to lock the page */
 			btrfs_set_lock_blocking(buf);
 			clear_extent_buffer_dirty(buf);
@@ -1178,9 +1189,13 @@ static void __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize,
 
 	INIT_LIST_HEAD(&root->dirty_list);
 	INIT_LIST_HEAD(&root->root_list);
+	INIT_LIST_HEAD(&root->logged_list[0]);
+	INIT_LIST_HEAD(&root->logged_list[1]);
 	spin_lock_init(&root->orphan_lock);
 	spin_lock_init(&root->inode_lock);
 	spin_lock_init(&root->accounting_lock);
+	spin_lock_init(&root->log_extents_lock[0]);
+	spin_lock_init(&root->log_extents_lock[1]);
 	mutex_init(&root->objectid_mutex);
 	mutex_init(&root->log_mutex);
 	init_waitqueue_head(&root->log_writer_wait);
@@ -2004,10 +2019,24 @@ int open_ctree(struct super_block *sb,
 		goto fail_srcu;
 	}
 
+	ret = percpu_counter_init(&fs_info->dirty_metadata_bytes, 0);
+	if (ret) {
+		err = ret;
+		goto fail_bdi;
+	}
+	fs_info->dirty_metadata_batch = PAGE_CACHE_SIZE *
+					(1 + ilog2(nr_cpu_ids));
+
+	ret = percpu_counter_init(&fs_info->delalloc_bytes, 0);
+	if (ret) {
+		err = ret;
+		goto fail_dirty_metadata_bytes;
+	}
+
 	fs_info->btree_inode = new_inode(sb);
 	if (!fs_info->btree_inode) {
 		err = -ENOMEM;
-		goto fail_bdi;
+		goto fail_delalloc_bytes;
 	}
 
 	mapping_set_gfp_mask(fs_info->btree_inode->i_mapping, GFP_NOFS);
@@ -2017,7 +2046,6 @@ int open_ctree(struct super_block *sb,
 	INIT_LIST_HEAD(&fs_info->dead_roots);
 	INIT_LIST_HEAD(&fs_info->delayed_iputs);
 	INIT_LIST_HEAD(&fs_info->delalloc_inodes);
-	INIT_LIST_HEAD(&fs_info->ordered_operations);
 	INIT_LIST_HEAD(&fs_info->caching_block_groups);
 	spin_lock_init(&fs_info->delalloc_lock);
 	spin_lock_init(&fs_info->trans_lock);
@@ -2028,6 +2056,7 @@ int open_ctree(struct super_block *sb,
 	spin_lock_init(&fs_info->tree_mod_seq_lock);
 	rwlock_init(&fs_info->tree_mod_log_lock);
 	mutex_init(&fs_info->reloc_mutex);
+	seqlock_init(&fs_info->profiles_lock);
 
 	init_completion(&fs_info->kobj_unregister);
 	INIT_LIST_HEAD(&fs_info->dirty_cowonly_roots);
@@ -2126,6 +2155,7 @@ int open_ctree(struct super_block *sb,
 
 	spin_lock_init(&fs_info->block_group_cache_lock);
 	fs_info->block_group_cache_tree = RB_ROOT;
+	fs_info->first_logical_byte = (u64)-1;
 
 	extent_io_tree_init(&fs_info->freed_extents[0],
 			     fs_info->btree_inode->i_mapping);
@@ -2165,6 +2195,12 @@ int open_ctree(struct super_block *sb,
 	init_waitqueue_head(&fs_info->transaction_blocked_wait);
 	init_waitqueue_head(&fs_info->async_submit_wait);
 
+	ret = btrfs_alloc_stripe_hash_table(fs_info);
+	if (ret) {
+		err = ret;
+		goto fail_alloc;
+	}
+
 	__setup_root(4096, 4096, 4096, 4096, tree_root,
 		     fs_info, BTRFS_ROOT_TREE_OBJECTID);
 
@@ -2187,7 +2223,8 @@ int open_ctree(struct super_block *sb,
 		goto fail_alloc;
 
 	/* check FS state, whether FS is broken. */
-	fs_info->fs_state |= btrfs_super_flags(disk_super);
+	if (btrfs_super_flags(disk_super) & BTRFS_SUPER_FLAG_ERROR)
+		set_bit(BTRFS_FS_STATE_ERROR, &fs_info->fs_state);
 
 	ret = btrfs_check_super_valid(fs_info, sb->s_flags & MS_RDONLY);
 	if (ret) {
@@ -2261,6 +2298,8 @@ int open_ctree(struct super_block *sb,
 	leafsize = btrfs_super_leafsize(disk_super);
 	sectorsize = btrfs_super_sectorsize(disk_super);
 	stripesize = btrfs_super_stripesize(disk_super);
+	fs_info->dirty_metadata_batch = leafsize * (1 + ilog2(nr_cpu_ids));
+	fs_info->delalloc_batch = sectorsize * 512 * (1 + ilog2(nr_cpu_ids));
 
 	/*
 	 * mixed block groups end up with duplicate but slightly offset
@@ -2332,6 +2371,12 @@ int open_ctree(struct super_block *sb,
 	btrfs_init_workers(&fs_info->endio_meta_write_workers,
 			   "endio-meta-write", fs_info->thread_pool_size,
 			   &fs_info->generic_worker);
+	btrfs_init_workers(&fs_info->endio_raid56_workers,
+			   "endio-raid56", fs_info->thread_pool_size,
+			   &fs_info->generic_worker);
+	btrfs_init_workers(&fs_info->rmw_workers,
+			   "rmw", fs_info->thread_pool_size,
+			   &fs_info->generic_worker);
 	btrfs_init_workers(&fs_info->endio_write_workers, "endio-write",
 			   fs_info->thread_pool_size,
 			   &fs_info->generic_worker);
@@ -2350,6 +2395,8 @@ int open_ctree(struct super_block *sb,
 	 */
 	fs_info->endio_workers.idle_thresh = 4;
 	fs_info->endio_meta_workers.idle_thresh = 4;
+	fs_info->endio_raid56_workers.idle_thresh = 4;
+	fs_info->rmw_workers.idle_thresh = 2;
 
 	fs_info->endio_write_workers.idle_thresh = 2;
 	fs_info->endio_meta_write_workers.idle_thresh = 2;
@@ -2366,6 +2413,8 @@ int open_ctree(struct super_block *sb,
 	ret |= btrfs_start_workers(&fs_info->fixup_workers);
 	ret |= btrfs_start_workers(&fs_info->endio_workers);
 	ret |= btrfs_start_workers(&fs_info->endio_meta_workers);
+	ret |= btrfs_start_workers(&fs_info->rmw_workers);
+	ret |= btrfs_start_workers(&fs_info->endio_raid56_workers);
 	ret |= btrfs_start_workers(&fs_info->endio_meta_write_workers);
 	ret |= btrfs_start_workers(&fs_info->endio_write_workers);
 	ret |= btrfs_start_workers(&fs_info->endio_freespace_worker);
@@ -2390,8 +2439,7 @@ int open_ctree(struct super_block *sb,
 	sb->s_blocksize = sectorsize;
 	sb->s_blocksize_bits = blksize_bits(sectorsize);
 
-	if (strncmp((char *)(&disk_super->magic), BTRFS_MAGIC,
-		    sizeof(disk_super->magic))) {
+	if (disk_super->magic != cpu_to_le64(BTRFS_MAGIC)) {
 		printk(KERN_INFO "btrfs: valid FS not found on %s\n", sb->s_id);
 		goto fail_sb_buffer;
 	}
@@ -2694,13 +2742,13 @@ int open_ctree(struct super_block *sb,
 	 * kthreads
 	 */
 	filemap_write_and_wait(fs_info->btree_inode->i_mapping);
-	invalidate_inode_pages2(fs_info->btree_inode->i_mapping);
 
 fail_block_groups:
 	btrfs_free_block_groups(fs_info);
 
 fail_tree_roots:
 	free_root_pointers(fs_info, 1);
+	invalidate_inode_pages2(fs_info->btree_inode->i_mapping);
 
 fail_sb_buffer:
 	btrfs_stop_workers(&fs_info->generic_worker);
@@ -2710,6 +2758,8 @@ int open_ctree(struct super_block *sb,
 	btrfs_stop_workers(&fs_info->workers);
 	btrfs_stop_workers(&fs_info->endio_workers);
 	btrfs_stop_workers(&fs_info->endio_meta_workers);
+	btrfs_stop_workers(&fs_info->endio_raid56_workers);
+	btrfs_stop_workers(&fs_info->rmw_workers);
 	btrfs_stop_workers(&fs_info->endio_meta_write_workers);
 	btrfs_stop_workers(&fs_info->endio_write_workers);
 	btrfs_stop_workers(&fs_info->endio_freespace_worker);
@@ -2721,13 +2771,17 @@ int open_ctree(struct super_block *sb,
 fail_iput:
 	btrfs_mapping_tree_free(&fs_info->mapping_tree);
 
-	invalidate_inode_pages2(fs_info->btree_inode->i_mapping);
 	iput(fs_info->btree_inode);
+fail_delalloc_bytes:
+	percpu_counter_destroy(&fs_info->delalloc_bytes);
+fail_dirty_metadata_bytes:
+	percpu_counter_destroy(&fs_info->dirty_metadata_bytes);
 fail_bdi:
 	bdi_destroy(&fs_info->bdi);
 fail_srcu:
 	cleanup_srcu_struct(&fs_info->subvol_srcu);
 fail:
+	btrfs_free_stripe_hash_table(fs_info);
 	btrfs_close_devices(fs_info->fs_devices);
 	return err;
 
@@ -2795,8 +2849,7 @@ struct buffer_head *btrfs_read_dev_super(struct block_device *bdev)
 
 		super = (struct btrfs_super_block *)bh->b_data;
 		if (btrfs_super_bytenr(super) != bytenr ||
-		    strncmp((char *)(&super->magic), BTRFS_MAGIC,
-			    sizeof(super->magic))) {
+		    super->magic != cpu_to_le64(BTRFS_MAGIC)) {
 			brelse(bh);
 			continue;
 		}
@@ -3076,11 +3129,16 @@ int btrfs_calc_num_tolerated_disk_barrier_failures(
 				     ((flags & BTRFS_BLOCK_GROUP_PROFILE_MASK)
 				      == 0)))
 					num_tolerated_disk_barrier_failures = 0;
-				else if (num_tolerated_disk_barrier_failures > 1
-					 &&
-					 (flags & (BTRFS_BLOCK_GROUP_RAID1 |
-						   BTRFS_BLOCK_GROUP_RAID10)))
-					num_tolerated_disk_barrier_failures = 1;
+				else if (num_tolerated_disk_barrier_failures > 1) {
+					if (flags & (BTRFS_BLOCK_GROUP_RAID1 |
+					    BTRFS_BLOCK_GROUP_RAID5 |
+					    BTRFS_BLOCK_GROUP_RAID10)) {
+						num_tolerated_disk_barrier_failures = 1;
+					} else if (flags &
+						   BTRFS_BLOCK_GROUP_RAID5) {
+						num_tolerated_disk_barrier_failures = 2;
+					}
+				}
 			}
 		}
 		up_read(&sinfo->groups_sem);
@@ -3195,6 +3253,11 @@ void 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);
 
+	if (fs_info->fs_state & BTRFS_SUPER_FLAG_ERROR) {
+		btrfs_free_log(NULL, root);
+		btrfs_free_log_root_tree(NULL, fs_info);
+	}
+
 	__btrfs_remove_free_space_cache(root->free_ino_pinned);
 	__btrfs_remove_free_space_cache(root->free_ino_ctl);
 	free_fs_root(root);
@@ -3339,7 +3402,7 @@ int close_ctree(struct btrfs_root *root)
 			printk(KERN_ERR "btrfs: commit super ret %d\n", ret);
 	}
 
-	if (fs_info->fs_state & BTRFS_SUPER_FLAG_ERROR)
+	if (test_bit(BTRFS_FS_STATE_ERROR, &fs_info->fs_state))
 		btrfs_error_commit_super(root);
 
 	btrfs_put_block_group_cache(fs_info);
@@ -3352,9 +3415,9 @@ int close_ctree(struct btrfs_root *root)
 
 	btrfs_free_qgroup_config(root->fs_info);
 
-	if (fs_info->delalloc_bytes) {
-		printk(KERN_INFO "btrfs: at unmount delalloc count %llu\n",
-		       (unsigned long long)fs_info->delalloc_bytes);
+	if (percpu_counter_sum(&fs_info->delalloc_bytes)) {
+		printk(KERN_INFO "btrfs: at unmount delalloc count %lld\n",
+		       percpu_counter_sum(&fs_info->delalloc_bytes));
 	}
 
 	free_extent_buffer(fs_info->extent_root->node);
@@ -3384,6 +3447,8 @@ int close_ctree(struct btrfs_root *root)
 	btrfs_stop_workers(&fs_info->workers);
 	btrfs_stop_workers(&fs_info->endio_workers);
 	btrfs_stop_workers(&fs_info->endio_meta_workers);
+	btrfs_stop_workers(&fs_info->endio_raid56_workers);
+	btrfs_stop_workers(&fs_info->rmw_workers);
 	btrfs_stop_workers(&fs_info->endio_meta_write_workers);
 	btrfs_stop_workers(&fs_info->endio_write_workers);
 	btrfs_stop_workers(&fs_info->endio_freespace_worker);
@@ -3401,9 +3466,13 @@ int close_ctree(struct btrfs_root *root)
 	btrfs_close_devices(fs_info->fs_devices);
 	btrfs_mapping_tree_free(&fs_info->mapping_tree);
 
+	percpu_counter_destroy(&fs_info->dirty_metadata_bytes);
+	percpu_counter_destroy(&fs_info->delalloc_bytes);
 	bdi_destroy(&fs_info->bdi);
 	cleanup_srcu_struct(&fs_info->subvol_srcu);
 
+	btrfs_free_stripe_hash_table(fs_info);
+
 	return 0;
 }
 
@@ -3443,11 +3512,10 @@ void btrfs_mark_buffer_dirty(struct extent_buffer *buf)
 			(unsigned long long)transid,
 			(unsigned long long)root->fs_info->generation);
 	was_dirty = set_extent_buffer_dirty(buf);
-	if (!was_dirty) {
-		spin_lock(&root->fs_info->delalloc_lock);
-		root->fs_info->dirty_metadata_bytes += buf->len;
-		spin_unlock(&root->fs_info->delalloc_lock);
-	}
+	if (!was_dirty)
+		__percpu_counter_add(&root->fs_info->dirty_metadata_bytes,
+				     buf->len,
+				     root->fs_info->dirty_metadata_batch);
 }
 
 static void __btrfs_btree_balance_dirty(struct btrfs_root *root,
@@ -3457,8 +3525,7 @@ static void __btrfs_btree_balance_dirty(struct btrfs_root *root,
 	 * 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;
+	int ret;
 
 	if (current->flags & PF_MEMALLOC)
 		return;
@@ -3466,9 +3533,9 @@ static void __btrfs_btree_balance_dirty(struct btrfs_root *root,
 	if (flush_delayed)
 		btrfs_balance_delayed_items(root);
 
-	num_dirty = root->fs_info->dirty_metadata_bytes;
-
-	if (num_dirty > thresh) {
+	ret = percpu_counter_compare(&root->fs_info->dirty_metadata_bytes,
+				     BTRFS_DIRTY_METADATA_THRESH);
+	if (ret > 0) {
 		balance_dirty_pages_ratelimited(
 				   root->fs_info->btree_inode->i_mapping);
 	}
@@ -3518,7 +3585,8 @@ void btrfs_error_commit_super(struct btrfs_root *root)
 	btrfs_cleanup_transaction(root);
 }
 
-static void btrfs_destroy_ordered_operations(struct btrfs_root *root)
+static void btrfs_destroy_ordered_operations(struct btrfs_transaction *t,
+					     struct btrfs_root *root)
 {
 	struct btrfs_inode *btrfs_inode;
 	struct list_head splice;
@@ -3528,7 +3596,7 @@ static void btrfs_destroy_ordered_operations(struct btrfs_root *root)
 	mutex_lock(&root->fs_info->ordered_operations_mutex);
 	spin_lock(&root->fs_info->ordered_extent_lock);
 
-	list_splice_init(&root->fs_info->ordered_operations, &splice);
+	list_splice_init(&t->ordered_operations, &splice);
 	while (!list_empty(&splice)) {
 		btrfs_inode = list_entry(splice.next, struct btrfs_inode,
 					 ordered_operations);
@@ -3544,35 +3612,16 @@ static void btrfs_destroy_ordered_operations(struct btrfs_root *root)
 
 static void btrfs_destroy_ordered_extents(struct btrfs_root *root)
 {
-	struct list_head splice;
 	struct btrfs_ordered_extent *ordered;
-	struct inode *inode;
-
-	INIT_LIST_HEAD(&splice);
 
 	spin_lock(&root->fs_info->ordered_extent_lock);
-
-	list_splice_init(&root->fs_info->ordered_extents, &splice);
-	while (!list_empty(&splice)) {
-		ordered = list_entry(splice.next, struct btrfs_ordered_extent,
-				     root_extent_list);
-
-		list_del_init(&ordered->root_extent_list);
-		atomic_inc(&ordered->refs);
-
-		/* the inode may be getting freed (in sys_unlink path). */
-		inode = igrab(ordered->inode);
-
-		spin_unlock(&root->fs_info->ordered_extent_lock);
-		if (inode)
-			iput(inode);
-
-		atomic_set(&ordered->refs, 1);
-		btrfs_put_ordered_extent(ordered);
-
-		spin_lock(&root->fs_info->ordered_extent_lock);
-	}
-
+	/*
+	 * This will just short circuit the ordered completion stuff which will
+	 * make sure the ordered extent gets properly cleaned up.
+	 */
+	list_for_each_entry(ordered, &root->fs_info->ordered_extents,
+			    root_extent_list)
+		set_bit(BTRFS_ORDERED_IOERR, &ordered->flags);
 	spin_unlock(&root->fs_info->ordered_extent_lock);
 }
 
@@ -3594,11 +3643,11 @@ int btrfs_destroy_delayed_refs(struct btrfs_transaction *trans,
 	}
 
 	while ((node = rb_first(&delayed_refs->root)) != NULL) {
-		ref = rb_entry(node, struct btrfs_delayed_ref_node, rb_node);
+		struct btrfs_delayed_ref_head *head = NULL;
 
+		ref = rb_entry(node, struct btrfs_delayed_ref_node, rb_node);
 		atomic_set(&ref->refs, 1);
 		if (btrfs_delayed_ref_is_head(ref)) {
-			struct btrfs_delayed_ref_head *head;
 
 			head = btrfs_delayed_node_to_head(ref);
 			if (!mutex_trylock(&head->mutex)) {
@@ -3614,16 +3663,18 @@ int btrfs_destroy_delayed_refs(struct btrfs_transaction *trans,
 				continue;
 			}
 
-			kfree(head->extent_op);
+			btrfs_free_delayed_extent_op(head->extent_op);
 			delayed_refs->num_heads--;
 			if (list_empty(&head->cluster))
 				delayed_refs->num_heads_ready--;
 			list_del_init(&head->cluster);
 		}
+
 		ref->in_tree = 0;
 		rb_erase(&ref->rb_node, &delayed_refs->root);
 		delayed_refs->num_entries--;
-
+		if (head)
+			mutex_unlock(&head->mutex);
 		spin_unlock(&delayed_refs->lock);
 		btrfs_put_delayed_ref(ref);
 
@@ -3671,6 +3722,8 @@ static void btrfs_destroy_delalloc_inodes(struct btrfs_root *root)
 				    delalloc_inodes);
 
 		list_del_init(&btrfs_inode->delalloc_inodes);
+		clear_bit(BTRFS_INODE_IN_DELALLOC_LIST,
+			  &btrfs_inode->runtime_flags);
 
 		btrfs_invalidate_inodes(btrfs_inode->root);
 	}
@@ -3823,10 +3876,8 @@ int btrfs_cleanup_transaction(struct btrfs_root *root)
 
 	while (!list_empty(&list)) {
 		t = list_entry(list.next, struct btrfs_transaction, list);
-		if (!t)
-			break;
 
-		btrfs_destroy_ordered_operations(root);
+		btrfs_destroy_ordered_operations(t, root);
 
 		btrfs_destroy_ordered_extents(root);
 

fs/btrfs/disk-io.h

@@ -25,6 +25,13 @@
 #define BTRFS_SUPER_MIRROR_MAX	 3
 #define BTRFS_SUPER_MIRROR_SHIFT 12
 
+enum {
+	BTRFS_WQ_ENDIO_DATA = 0,
+	BTRFS_WQ_ENDIO_METADATA = 1,
+	BTRFS_WQ_ENDIO_FREE_SPACE = 2,
+	BTRFS_WQ_ENDIO_RAID56 = 3,
+};
+
 static inline u64 btrfs_sb_offset(int mirror)
 {
 	u64 start = 16 * 1024;

fs/btrfs/extent_io.c

@@ -4,7 +4,6 @@
 #include <linux/mm.h>
 #include <linux/pagemap.h>
 #include <linux/page-flags.h>
-#include <linux/module.h>
 #include <linux/spinlock.h>
 #include <linux/blkdev.h>
 #include <linux/swap.h>
@@ -1834,7 +1833,7 @@ int test_range_bit(struct extent_io_tree *tree, u64 start, u64 end,
  */
 static void check_page_uptodate(struct extent_io_tree *tree, struct page *page)
 {
-	u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
+	u64 start = page_offset(page);
 	u64 end = start + PAGE_CACHE_SIZE - 1;
 	if (test_range_bit(tree, start, end, EXTENT_UPTODATE, 1, NULL))
 		SetPageUptodate(page);
@@ -1846,7 +1845,7 @@ static void check_page_uptodate(struct extent_io_tree *tree, struct page *page)
  */
 static void check_page_locked(struct extent_io_tree *tree, struct page *page)
 {
-	u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
+	u64 start = page_offset(page);
 	u64 end = start + PAGE_CACHE_SIZE - 1;
 	if (!test_range_bit(tree, start, end, EXTENT_LOCKED, 0, NULL))
 		unlock_page(page);
@@ -1895,13 +1894,11 @@ static int free_io_failure(struct inode *inode, struct io_failure_record *rec,
 	if (ret)
 		err = ret;
 
-	if (did_repair) {
-		ret = clear_extent_bits(&BTRFS_I(inode)->io_tree, rec->start,
-					rec->start + rec->len - 1,
-					EXTENT_DAMAGED, GFP_NOFS);
-		if (ret && !err)
-			err = ret;
-	}
+	ret = clear_extent_bits(&BTRFS_I(inode)->io_tree, rec->start,
+				rec->start + rec->len - 1,
+				EXTENT_DAMAGED, GFP_NOFS);
+	if (ret && !err)
+		err = ret;
 
 	kfree(rec);
 	return err;
@@ -1932,10 +1929,15 @@ int repair_io_failure(struct btrfs_fs_info *fs_info, u64 start,
 	u64 map_length = 0;
 	u64 sector;
 	struct btrfs_bio *bbio = NULL;
+	struct btrfs_mapping_tree *map_tree = &fs_info->mapping_tree;
 	int ret;
 
 	BUG_ON(!mirror_num);
 
+	/* we can't repair anything in raid56 yet */
+	if (btrfs_is_parity_mirror(map_tree, logical, length, mirror_num))
+		return 0;
+
 	bio = bio_alloc(GFP_NOFS, 1);
 	if (!bio)
 		return -EIO;
@@ -1960,7 +1962,7 @@ int repair_io_failure(struct btrfs_fs_info *fs_info, u64 start,
 		return -EIO;
 	}
 	bio->bi_bdev = dev->bdev;
-	bio_add_page(bio, page, length, start-page_offset(page));
+	bio_add_page(bio, page, length, start - page_offset(page));
 	btrfsic_submit_bio(WRITE_SYNC, bio);
 	wait_for_completion(&compl);
 
@@ -2052,6 +2054,7 @@ static int clean_io_failure(u64 start, struct page *page)
 						failrec->failed_mirror);
 			did_repair = !ret;
 		}
+		ret = 0;
 	}
 
 out:
@@ -2293,8 +2296,7 @@ static void end_bio_extent_writepage(struct bio *bio, int err)
 		struct page *page = bvec->bv_page;
 		tree = &BTRFS_I(page->mapping->host)->io_tree;
 
-		start = ((u64)page->index << PAGE_CACHE_SHIFT) +
-			 bvec->bv_offset;
+		start = page_offset(page) + bvec->bv_offset;
 		end = start + bvec->bv_len - 1;
 
 		if (bvec->bv_offset == 0 && bvec->bv_len == PAGE_CACHE_SIZE)
@@ -2353,8 +2355,7 @@ static void end_bio_extent_readpage(struct bio *bio, int err)
 			 (long int)bio->bi_bdev);
 		tree = &BTRFS_I(page->mapping->host)->io_tree;
 
-		start = ((u64)page->index << PAGE_CACHE_SHIFT) +
-			bvec->bv_offset;
+		start = page_offset(page) + bvec->bv_offset;
 		end = start + bvec->bv_len - 1;
 
 		if (bvec->bv_offset == 0 && bvec->bv_len == PAGE_CACHE_SIZE)
@@ -2471,7 +2472,7 @@ static int __must_check submit_one_bio(int rw, struct bio *bio,
 	struct extent_io_tree *tree = bio->bi_private;
 	u64 start;
 
-	start = ((u64)page->index << PAGE_CACHE_SHIFT) + bvec->bv_offset;
+	start = page_offset(page) + bvec->bv_offset;
 
 	bio->bi_private = NULL;
 
@@ -2489,13 +2490,13 @@ static int __must_check submit_one_bio(int rw, struct bio *bio,
 	return ret;
 }
 
-static int merge_bio(struct extent_io_tree *tree, struct page *page,
+static int merge_bio(int rw, struct extent_io_tree *tree, struct page *page,
 		     unsigned long offset, size_t size, struct bio *bio,
 		     unsigned long bio_flags)
 {
 	int ret = 0;
 	if (tree->ops && tree->ops->merge_bio_hook)
-		ret = tree->ops->merge_bio_hook(page, offset, size, bio,
+		ret = tree->ops->merge_bio_hook(rw, page, offset, size, bio,
 						bio_flags);
 	BUG_ON(ret < 0);
 	return ret;
@@ -2530,7 +2531,7 @@ static int submit_extent_page(int rw, struct extent_io_tree *tree,
 				sector;
 
 		if (prev_bio_flags != bio_flags || !contig ||
-		    merge_bio(tree, page, offset, page_size, bio, bio_flags) ||
+		    merge_bio(rw, tree, page, offset, page_size, bio, bio_flags) ||
 		    bio_add_page(bio, page, page_size, offset) < page_size) {
 			ret = submit_one_bio(rw, bio, mirror_num,
 					     prev_bio_flags);
@@ -2595,7 +2596,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 				   unsigned long *bio_flags)
 {
 	struct inode *inode = page->mapping->host;
-	u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
+	u64 start = page_offset(page);
 	u64 page_end = start + PAGE_CACHE_SIZE - 1;
 	u64 end;
 	u64 cur = start;
@@ -2648,6 +2649,8 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 		}
 	}
 	while (cur <= end) {
+		unsigned long pnr = (last_byte >> PAGE_CACHE_SHIFT) + 1;
+
 		if (cur >= last_byte) {
 			char *userpage;
 			struct extent_state *cached = NULL;
@@ -2682,7 +2685,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 
 		iosize = min(extent_map_end(em) - cur, end - cur + 1);
 		cur_end = min(extent_map_end(em) - 1, end);
-		iosize = (iosize + blocksize - 1) & ~((u64)blocksize - 1);
+		iosize = ALIGN(iosize, blocksize);
 		if (this_bio_flag & EXTENT_BIO_COMPRESSED) {
 			disk_io_size = em->block_len;
 			sector = em->block_start >> 9;
@@ -2735,26 +2738,17 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 			continue;
 		}
 
-		ret = 0;
-		if (tree->ops && tree->ops->readpage_io_hook) {
-			ret = tree->ops->readpage_io_hook(page, cur,
-							  cur + iosize - 1);
-		}
-		if (!ret) {
-			unsigned long pnr = (last_byte >> PAGE_CACHE_SHIFT) + 1;
-			pnr -= page->index;
-			ret = submit_extent_page(READ, tree, page,
+		pnr -= page->index;
+		ret = submit_extent_page(READ, tree, page,
 					 sector, disk_io_size, pg_offset,
 					 bdev, bio, pnr,
 					 end_bio_extent_readpage, mirror_num,
 					 *bio_flags,
 					 this_bio_flag);
-			if (!ret) {
-				nr++;
-				*bio_flags = this_bio_flag;
-			}
-		}
-		if (ret) {
+		if (!ret) {
+			nr++;
+			*bio_flags = this_bio_flag;
+		} else {
 			SetPageError(page);
 			unlock_extent(tree, cur, cur + iosize - 1);
 		}
@@ -2806,7 +2800,7 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc,
 	struct inode *inode = page->mapping->host;
 	struct extent_page_data *epd = data;
 	struct extent_io_tree *tree = epd->tree;
-	u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
+	u64 start = page_offset(page);
 	u64 delalloc_start;
 	u64 page_end = start + PAGE_CACHE_SIZE - 1;
 	u64 end;
@@ -2982,7 +2976,7 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc,
 		BUG_ON(extent_map_end(em) <= cur);
 		BUG_ON(end < cur);
 		iosize = min(extent_map_end(em) - cur, end - cur + 1);
-		iosize = (iosize + blocksize - 1) & ~((u64)blocksize - 1);
+		iosize = ALIGN(iosize, blocksize);
 		sector = (em->block_start + extent_offset) >> 9;
 		bdev = em->bdev;
 		block_start = em->block_start;
@@ -3124,12 +3118,9 @@ static int lock_extent_buffer_for_io(struct extent_buffer *eb,
 		set_bit(EXTENT_BUFFER_WRITEBACK, &eb->bflags);
 		spin_unlock(&eb->refs_lock);
 		btrfs_set_header_flag(eb, BTRFS_HEADER_FLAG_WRITTEN);
-		spin_lock(&fs_info->delalloc_lock);
-		if (fs_info->dirty_metadata_bytes >= eb->len)
-			fs_info->dirty_metadata_bytes -= eb->len;
-		else
-			WARN_ON(1);
-		spin_unlock(&fs_info->delalloc_lock);
+		__percpu_counter_add(&fs_info->dirty_metadata_bytes,
+				     -eb->len,
+				     fs_info->dirty_metadata_batch);
 		ret = 1;
 	} else {
 		spin_unlock(&eb->refs_lock);
@@ -3446,15 +3437,9 @@ static int extent_write_cache_pages(struct extent_io_tree *tree,
 			 * swizzled back from swapper_space to tmpfs file
 			 * mapping
 			 */
-			if (tree->ops &&
-			    tree->ops->write_cache_pages_lock_hook) {
-				tree->ops->write_cache_pages_lock_hook(page,
-							       data, flush_fn);
-			} else {
-				if (!trylock_page(page)) {
-					flush_fn(data);
-					lock_page(page);
-				}
+			if (!trylock_page(page)) {
+				flush_fn(data);
+				lock_page(page);
 			}
 
 			if (unlikely(page->mapping != mapping)) {
@@ -3674,11 +3659,11 @@ int extent_invalidatepage(struct extent_io_tree *tree,
 			  struct page *page, unsigned long offset)
 {
 	struct extent_state *cached_state = NULL;
-	u64 start = ((u64)page->index << PAGE_CACHE_SHIFT);
+	u64 start = page_offset(page);
 	u64 end = start + PAGE_CACHE_SIZE - 1;
 	size_t blocksize = page->mapping->host->i_sb->s_blocksize;
 
-	start += (offset + blocksize - 1) & ~(blocksize - 1);
+	start += ALIGN(offset, blocksize);
 	if (start > end)
 		return 0;
 
@@ -3700,7 +3685,7 @@ int try_release_extent_state(struct extent_map_tree *map,
 			     struct extent_io_tree *tree, struct page *page,
 			     gfp_t mask)
 {
-	u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
+	u64 start = page_offset(page);
 	u64 end = start + PAGE_CACHE_SIZE - 1;
 	int ret = 1;
 
@@ -3739,7 +3724,7 @@ int try_release_extent_mapping(struct extent_map_tree *map,
 			       gfp_t mask)
 {
 	struct extent_map *em;
-	u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
+	u64 start = page_offset(page);
 	u64 end = start + PAGE_CACHE_SIZE - 1;
 
 	if ((mask & __GFP_WAIT) &&
@@ -3797,7 +3782,7 @@ static struct extent_map *get_extent_skip_holes(struct inode *inode,
 		len = last - offset;
 		if (len == 0)
 			break;
-		len = (len + sectorsize - 1) & ~(sectorsize - 1);
+		len = ALIGN(len, sectorsize);
 		em = get_extent(inode, NULL, 0, offset, len, 0);
 		if (IS_ERR_OR_NULL(em))
 			return em;
@@ -3995,8 +3980,6 @@ static void __free_extent_buffer(struct extent_buffer *eb)
 	list_del(&eb->leak_list);
 	spin_unlock_irqrestore(&leak_lock, flags);
 #endif
-	if (eb->pages && eb->pages != eb->inline_pages)
-		kfree(eb->pages);
 	kmem_cache_free(extent_buffer_cache, eb);
 }
 
@@ -4037,19 +4020,12 @@ static struct extent_buffer *__alloc_extent_buffer(struct extent_io_tree *tree,
 	atomic_set(&eb->refs, 1);
 	atomic_set(&eb->io_pages, 0);
 
-	if (len > MAX_INLINE_EXTENT_BUFFER_SIZE) {
-		struct page **pages;
-		int num_pages = (len + PAGE_CACHE_SIZE - 1) >>
-			PAGE_CACHE_SHIFT;
-		pages = kzalloc(num_pages, mask);
-		if (!pages) {
-			__free_extent_buffer(eb);
-			return NULL;
-		}
-		eb->pages = pages;
-	} else {
-		eb->pages = eb->inline_pages;
-	}
+	/*
+	 * Sanity checks, currently the maximum is 64k covered by 16x 4k pages
+	 */
+	BUILD_BUG_ON(BTRFS_MAX_METADATA_BLOCKSIZE
+		> MAX_INLINE_EXTENT_BUFFER_SIZE);
+	BUG_ON(len > MAX_INLINE_EXTENT_BUFFER_SIZE);
 
 	return eb;
 }
@@ -4180,6 +4156,7 @@ static inline void btrfs_release_extent_buffer(struct extent_buffer *eb)
 
 static void check_buffer_tree_ref(struct extent_buffer *eb)
 {
+	int refs;
 	/* the ref bit is tricky.  We have to make sure it is set
 	 * if we have the buffer dirty.   Otherwise the
 	 * code to free a buffer can end up dropping a dirty
@@ -4200,6 +4177,10 @@ static void check_buffer_tree_ref(struct extent_buffer *eb)
 	 * So bump the ref count first, then set the bit.  If someone
 	 * beat us to it, drop the ref we added.
 	 */
+	refs = atomic_read(&eb->refs);
+	if (refs >= 2 && test_bit(EXTENT_BUFFER_TREE_REF, &eb->bflags))
+		return;
+
 	spin_lock(&eb->refs_lock);
 	if (!test_and_set_bit(EXTENT_BUFFER_TREE_REF, &eb->bflags))
 		atomic_inc(&eb->refs);
@@ -4401,9 +4382,20 @@ static int release_extent_buffer(struct extent_buffer *eb, gfp_t mask)
 
 void free_extent_buffer(struct extent_buffer *eb)
 {
+	int refs;
+	int old;
 	if (!eb)
 		return;
 
+	while (1) {
+		refs = atomic_read(&eb->refs);
+		if (refs <= 3)
+			break;
+		old = atomic_cmpxchg(&eb->refs, refs, refs - 1);
+		if (old == refs)
+			return;
+	}
+
 	spin_lock(&eb->refs_lock);
 	if (atomic_read(&eb->refs) == 2 &&
 	    test_bit(EXTENT_BUFFER_DUMMY, &eb->bflags))

fs/btrfs/extent_io.h

@@ -72,10 +72,9 @@ struct extent_io_ops {
 	int (*writepage_start_hook)(struct page *page, u64 start, u64 end);
 	int (*writepage_io_hook)(struct page *page, u64 start, u64 end);
 	extent_submit_bio_hook_t *submit_bio_hook;
-	int (*merge_bio_hook)(struct page *page, unsigned long offset,
+	int (*merge_bio_hook)(int rw, struct page *page, unsigned long offset,
 			      size_t size, struct bio *bio,
 			      unsigned long bio_flags);
-	int (*readpage_io_hook)(struct page *page, u64 start, u64 end);
 	int (*readpage_io_failed_hook)(struct page *page, int failed_mirror);
 	int (*readpage_end_io_hook)(struct page *page, u64 start, u64 end,
 				    struct extent_state *state, int mirror);
@@ -90,8 +89,6 @@ struct extent_io_ops {
 				  struct extent_state *other);
 	void (*split_extent_hook)(struct inode *inode,
 				  struct extent_state *orig, u64 split);
-	int (*write_cache_pages_lock_hook)(struct page *page, void *data,
-					   void (*flush_fn)(void *));
 };
 
 struct extent_io_tree {
@@ -161,8 +158,7 @@ struct extent_buffer {
 	 */
 	wait_queue_head_t read_lock_wq;
 	wait_queue_head_t lock_wq;
-	struct page *inline_pages[INLINE_EXTENT_BUFFER_PAGES];
-	struct page **pages;
+	struct page *pages[INLINE_EXTENT_BUFFER_PAGES];
 };
 
 static inline void extent_set_compress_type(unsigned long *bio_flags,

fs/btrfs/extent_map.c

 #include <linux/err.h>
 #include <linux/slab.h>
-#include <linux/module.h>
 #include <linux/spinlock.h>
 #include <linux/hardirq.h>
 #include "ctree.h"

fs/btrfs/file-item.c

@@ -684,6 +684,24 @@ int btrfs_del_csums(struct btrfs_trans_handle *trans,
 	return ret;
 }
 
+static u64 btrfs_sector_sum_left(struct btrfs_ordered_sum *sums,
+				 struct btrfs_sector_sum *sector_sum,
+				 u64 total_bytes, u64 sectorsize)
+{
+	u64 tmp = sectorsize;
+	u64 next_sector = sector_sum->bytenr;
+	struct btrfs_sector_sum *next = sector_sum + 1;
+
+	while ((tmp + total_bytes) < sums->len) {
+		if (next_sector + sectorsize != next->bytenr)
+			break;
+		tmp += sectorsize;
+		next_sector = next->bytenr;
+		next++;
+	}
+	return tmp;
+}
+
 int btrfs_csum_file_blocks(struct btrfs_trans_handle *trans,
 			   struct btrfs_root *root,
 			   struct btrfs_ordered_sum *sums)
@@ -789,20 +807,32 @@ int btrfs_csum_file_blocks(struct btrfs_trans_handle *trans,
 		goto insert;
 	}
 
-	if (csum_offset >= btrfs_item_size_nr(leaf, path->slots[0]) /
+	if (csum_offset == btrfs_item_size_nr(leaf, path->slots[0]) /
 	    csum_size) {
-		u32 diff = (csum_offset + 1) * csum_size;
+		int extend_nr;
+		u64 tmp;
+		u32 diff;
+		u32 free_space;
 
-		/*
-		 * is the item big enough already?  we dropped our lock
-		 * before and need to recheck
-		 */
-		if (diff < btrfs_item_size_nr(leaf, path->slots[0]))
-			goto csum;
+		if (btrfs_leaf_free_space(root, leaf) <
+				 sizeof(struct btrfs_item) + csum_size * 2)
+			goto insert;
+
+		free_space = btrfs_leaf_free_space(root, leaf) -
+					 sizeof(struct btrfs_item) - csum_size;
+		tmp = btrfs_sector_sum_left(sums, sector_sum, total_bytes,
+					    root->sectorsize);
+		tmp >>= root->fs_info->sb->s_blocksize_bits;
+		WARN_ON(tmp < 1);
+
+		extend_nr = max_t(int, 1, (int)tmp);
+		diff = (csum_offset + extend_nr) * csum_size;
+		diff = min(diff, MAX_CSUM_ITEMS(root, csum_size) * csum_size);
 
 		diff = diff - btrfs_item_size_nr(leaf, path->slots[0]);
-		if (diff != csum_size)
-			goto insert;
+		diff = min(free_space, diff);
+		diff /= csum_size;
+		diff *= csum_size;
 
 		btrfs_extend_item(trans, root, path, diff);
 		goto csum;
@@ -812,19 +842,14 @@ int btrfs_csum_file_blocks(struct btrfs_trans_handle *trans,
 	btrfs_release_path(path);
 	csum_offset = 0;
 	if (found_next) {
-		u64 tmp = total_bytes + root->sectorsize;
-		u64 next_sector = sector_sum->bytenr;
-		struct btrfs_sector_sum *next = sector_sum + 1;
+		u64 tmp;
 
-		while (tmp < sums->len) {
-			if (next_sector + root->sectorsize != next->bytenr)
-				break;
-			tmp += root->sectorsize;
-			next_sector = next->bytenr;
-			next++;
-		}
-		tmp = min(tmp, next_offset - file_key.offset);
+		tmp = btrfs_sector_sum_left(sums, sector_sum, total_bytes,
+					    root->sectorsize);
 		tmp >>= root->fs_info->sb->s_blocksize_bits;
+		tmp = min(tmp, (next_offset - file_key.offset) >>
+					 root->fs_info->sb->s_blocksize_bits);
+
 		tmp = max((u64)1, tmp);
 		tmp = min(tmp, (u64)MAX_CSUM_ITEMS(root, csum_size));
 		ins_size = csum_size * tmp;

fs/btrfs/file.c

@@ -30,11 +30,11 @@
 #include <linux/statfs.h>
 #include <linux/compat.h>
 #include <linux/slab.h>
+#include <linux/btrfs.h>
 #include "ctree.h"
 #include "disk-io.h"
 #include "transaction.h"
 #include "btrfs_inode.h"
-#include "ioctl.h"
 #include "print-tree.h"
 #include "tree-log.h"
 #include "locking.h"
@@ -374,6 +374,11 @@ int btrfs_run_defrag_inodes(struct btrfs_fs_info *fs_info)
 
 	atomic_inc(&fs_info->defrag_running);
 	while(1) {
+		/* Pause the auto defragger. */
+		if (test_bit(BTRFS_FS_STATE_REMOUNTING,
+			     &fs_info->fs_state))
+			break;
+
 		if (!__need_auto_defrag(fs_info->tree_root))
 			break;
 
@@ -505,8 +510,7 @@ int btrfs_dirty_pages(struct btrfs_root *root, struct inode *inode,
 	loff_t isize = i_size_read(inode);
 
 	start_pos = pos & ~((u64)root->sectorsize - 1);
-	num_bytes = (write_bytes + pos - start_pos +
-		    root->sectorsize - 1) & ~((u64)root->sectorsize - 1);
+	num_bytes = ALIGN(write_bytes + pos - start_pos, root->sectorsize);
 
 	end_of_last_block = start_pos + num_bytes - 1;
 	err = btrfs_set_extent_delalloc(inode, start_pos, end_of_last_block,
@@ -1544,7 +1548,7 @@ static ssize_t btrfs_file_aio_write(struct kiocb *iocb,
 	 * although we have opened a file as writable, we have
 	 * to stop this write operation to ensure FS consistency.
 	 */
-	if (root->fs_info->fs_state & BTRFS_SUPER_FLAG_ERROR) {
+	if (test_bit(BTRFS_FS_STATE_ERROR, &root->fs_info->fs_state)) {
 		mutex_unlock(&inode->i_mutex);
 		err = -EROFS;
 		goto out;
@@ -1627,7 +1631,20 @@ int btrfs_release_file(struct inode *inode, struct file *filp)
 	 */
 	if (test_and_clear_bit(BTRFS_INODE_ORDERED_DATA_CLOSE,
 			       &BTRFS_I(inode)->runtime_flags)) {
-		btrfs_add_ordered_operation(NULL, BTRFS_I(inode)->root, inode);
+		struct btrfs_trans_handle *trans;
+		struct btrfs_root *root = BTRFS_I(inode)->root;
+
+		/*
+		 * We need to block on a committing transaction to keep us from
+		 * throwing a ordered operation on to the list and causing
+		 * something like sync to deadlock trying to flush out this
+		 * inode.
+		 */
+		trans = btrfs_start_transaction(root, 0);
+		if (IS_ERR(trans))
+			return PTR_ERR(trans);
+		btrfs_add_ordered_operation(trans, BTRFS_I(inode)->root, inode);
+		btrfs_end_transaction(trans, root);
 		if (inode->i_size > BTRFS_ORDERED_OPERATIONS_FLUSH_LIMIT)
 			filemap_flush(inode->i_mapping);
 	}
@@ -1654,16 +1671,21 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	int ret = 0;
 	struct btrfs_trans_handle *trans;
+	bool full_sync = 0;
 
 	trace_btrfs_sync_file(file, datasync);
 
 	/*
 	 * We write the dirty pages in the range and wait until they complete
 	 * out of the ->i_mutex. If so, we can flush the dirty pages by
-	 * multi-task, and make the performance up.
+	 * multi-task, and make the performance up.  See
+	 * btrfs_wait_ordered_range for an explanation of the ASYNC check.
 	 */
 	atomic_inc(&BTRFS_I(inode)->sync_writers);
-	ret = filemap_write_and_wait_range(inode->i_mapping, start, end);
+	ret = filemap_fdatawrite_range(inode->i_mapping, start, end);
+	if (!ret && test_bit(BTRFS_INODE_HAS_ASYNC_EXTENT,
+			     &BTRFS_I(inode)->runtime_flags))
+		ret = filemap_fdatawrite_range(inode->i_mapping, start, end);
 	atomic_dec(&BTRFS_I(inode)->sync_writers);
 	if (ret)
 		return ret;
@@ -1675,7 +1697,10 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 	 * range being left.
 	 */
 	atomic_inc(&root->log_batch);
-	btrfs_wait_ordered_range(inode, start, end - start + 1);
+	full_sync = test_bit(BTRFS_INODE_NEEDS_FULL_SYNC,
+			     &BTRFS_I(inode)->runtime_flags);
+	if (full_sync)
+		btrfs_wait_ordered_range(inode, start, end - start + 1);
 	atomic_inc(&root->log_batch);
 
 	/*
@@ -1742,13 +1767,25 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 
 	if (ret != BTRFS_NO_LOG_SYNC) {
 		if (ret > 0) {
+			/*
+			 * If we didn't already wait for ordered extents we need
+			 * to do that now.
+			 */
+			if (!full_sync)
+				btrfs_wait_ordered_range(inode, start,
+							 end - start + 1);
 			ret = btrfs_commit_transaction(trans, root);
 		} else {
 			ret = btrfs_sync_log(trans, root);
-			if (ret == 0)
+			if (ret == 0) {
 				ret = btrfs_end_transaction(trans, root);
-			else
+			} else {
+				if (!full_sync)
+					btrfs_wait_ordered_range(inode, start,
+								 end -
+								 start + 1);
 				ret = btrfs_commit_transaction(trans, root);
+			}
 		}
 	} else {
 		ret = btrfs_end_transaction(trans, root);

fs/btrfs/free-space-cache.c

@@ -1356,6 +1356,8 @@ static void recalculate_thresholds(struct btrfs_free_space_ctl *ctl)
 	u64 bytes_per_bg = BITS_PER_BITMAP * ctl->unit;
 	int max_bitmaps = div64_u64(size + bytes_per_bg - 1, bytes_per_bg);
 
+	max_bitmaps = max(max_bitmaps, 1);
+
 	BUG_ON(ctl->total_bitmaps > max_bitmaps);
 
 	/*
@@ -1463,10 +1465,14 @@ static int search_bitmap(struct btrfs_free_space_ctl *ctl,
 }
 
 static struct btrfs_free_space *
-find_free_space(struct btrfs_free_space_ctl *ctl, u64 *offset, u64 *bytes)
+find_free_space(struct btrfs_free_space_ctl *ctl, u64 *offset, u64 *bytes,
+		unsigned long align)
 {
 	struct btrfs_free_space *entry;
 	struct rb_node *node;
+	u64 ctl_off;
+	u64 tmp;
+	u64 align_off;
 	int ret;
 
 	if (!ctl->free_space_offset.rb_node)
@@ -1481,15 +1487,34 @@ find_free_space(struct btrfs_free_space_ctl *ctl, u64 *offset, u64 *bytes)
 		if (entry->bytes < *bytes)
 			continue;
 
+		/* make sure the space returned is big enough
+		 * to match our requested alignment
+		 */
+		if (*bytes >= align) {
+			ctl_off = entry->offset - ctl->start;
+			tmp = ctl_off + align - 1;;
+			do_div(tmp, align);
+			tmp = tmp * align + ctl->start;
+			align_off = tmp - entry->offset;
+		} else {
+			align_off = 0;
+			tmp = entry->offset;
+		}
+
+		if (entry->bytes < *bytes + align_off)
+			continue;
+
 		if (entry->bitmap) {
-			ret = search_bitmap(ctl, entry, offset, bytes);
-			if (!ret)
+			ret = search_bitmap(ctl, entry, &tmp, bytes);
+			if (!ret) {
+				*offset = tmp;
 				return entry;
+			}
 			continue;
 		}
 
-		*offset = entry->offset;
-		*bytes = entry->bytes;
+		*offset = tmp;
+		*bytes = entry->bytes - align_off;
 		return entry;
 	}
 
@@ -1636,10 +1661,14 @@ static bool use_bitmap(struct btrfs_free_space_ctl *ctl,
 	}
 
 	/*
-	 * some block groups are so tiny they can't be enveloped by a bitmap, so
-	 * don't even bother to create a bitmap for this
+	 * The original block groups from mkfs can be really small, like 8
+	 * megabytes, so don't bother with a bitmap for those entries.  However
+	 * some block groups can be smaller than what a bitmap would cover but
+	 * are still large enough that they could overflow the 32k memory limit,
+	 * so allow those block groups to still be allowed to have a bitmap
+	 * entry.
 	 */
-	if (BITS_PER_BITMAP * ctl->unit > block_group->key.offset)
+	if (((BITS_PER_BITMAP * ctl->unit) >> 1) > block_group->key.offset)
 		return false;
 
 	return true;
@@ -2095,9 +2124,12 @@ u64 btrfs_find_space_for_alloc(struct btrfs_block_group_cache *block_group,
 	struct btrfs_free_space *entry = NULL;
 	u64 bytes_search = bytes + empty_size;
 	u64 ret = 0;
+	u64 align_gap = 0;
+	u64 align_gap_len = 0;
 
 	spin_lock(&ctl->tree_lock);
-	entry = find_free_space(ctl, &offset, &bytes_search);
+	entry = find_free_space(ctl, &offset, &bytes_search,
+				block_group->full_stripe_len);
 	if (!entry)
 		goto out;
 
@@ -2107,9 +2139,15 @@ u64 btrfs_find_space_for_alloc(struct btrfs_block_group_cache *block_group,
 		if (!entry->bytes)
 			free_bitmap(ctl, entry);
 	} else {
+
 		unlink_free_space(ctl, entry);
-		entry->offset += bytes;
-		entry->bytes -= bytes;
+		align_gap_len = offset - entry->offset;
+		align_gap = entry->offset;
+
+		entry->offset = offset + bytes;
+		WARN_ON(entry->bytes < bytes + align_gap_len);
+
+		entry->bytes -= bytes + align_gap_len;
 		if (!entry->bytes)
 			kmem_cache_free(btrfs_free_space_cachep, entry);
 		else
@@ -2119,6 +2157,8 @@ u64 btrfs_find_space_for_alloc(struct btrfs_block_group_cache *block_group,
 out:
 	spin_unlock(&ctl->tree_lock);
 
+	if (align_gap_len)
+		__btrfs_add_free_space(ctl, align_gap, align_gap_len);
 	return ret;
 }
 

fs/btrfs/locking.c

@@ -113,11 +113,10 @@ void btrfs_tree_read_lock(struct extent_buffer *eb)
 		read_unlock(&eb->lock);
 		return;
 	}
-	read_unlock(&eb->lock);
-	wait_event(eb->write_lock_wq, atomic_read(&eb->blocking_writers) == 0);
-	read_lock(&eb->lock);
 	if (atomic_read(&eb->blocking_writers)) {
 		read_unlock(&eb->lock);
+		wait_event(eb->write_lock_wq,
+			   atomic_read(&eb->blocking_writers) == 0);
 		goto again;
 	}
 	atomic_inc(&eb->read_locks);

fs/btrfs/ordered-data.h

@@ -79,6 +79,8 @@ struct btrfs_ordered_sum {
 #define BTRFS_ORDERED_UPDATED_ISIZE 7 /* indicates whether this ordered extent
 				       * has done its due diligence in updating
 				       * the isize. */
+#define BTRFS_ORDERED_LOGGED_CSUM 8 /* We've logged the csums on this ordered
+				       ordered extent */
 
 struct btrfs_ordered_extent {
 	/* logical offset in the file */
@@ -96,6 +98,9 @@ struct btrfs_ordered_extent {
 	/* number of bytes that still need writing */
 	u64 bytes_left;
 
+	/* number of bytes that still need csumming */
+	u64 csum_bytes_left;
+
 	/*
 	 * the end of the ordered extent which is behind it but
 	 * didn't update disk_i_size. Please see the comment of
@@ -118,6 +123,9 @@ struct btrfs_ordered_extent {
 	/* list of checksums for insertion when the extent io is done */
 	struct list_head list;
 
+	/* If we need to wait on this to be done */
+	struct list_head log_list;
+
 	/* used to wait for the BTRFS_ORDERED_COMPLETE bit */
 	wait_queue_head_t wait;
 
@@ -189,11 +197,15 @@ struct btrfs_ordered_extent *btrfs_lookup_ordered_range(struct inode *inode,
 int btrfs_ordered_update_i_size(struct inode *inode, u64 offset,
 				struct btrfs_ordered_extent *ordered);
 int btrfs_find_ordered_sum(struct inode *inode, u64 offset, u64 disk_bytenr, u32 *sum);
-int btrfs_run_ordered_operations(struct btrfs_root *root, int wait);
+int btrfs_run_ordered_operations(struct btrfs_trans_handle *trans,
+				 struct btrfs_root *root, int wait);
 void btrfs_add_ordered_operation(struct btrfs_trans_handle *trans,
 				 struct btrfs_root *root,
 				 struct inode *inode);
 void btrfs_wait_ordered_extents(struct btrfs_root *root, int delay_iput);
+void btrfs_get_logged_extents(struct btrfs_root *log, struct inode *inode);
+void btrfs_wait_logged_extents(struct btrfs_root *log, u64 transid);
+void btrfs_free_logged_extents(struct btrfs_root *log, u64 transid);
 int __init ordered_data_init(void);
 void ordered_data_exit(void);
 #endif

fs/btrfs/print-tree.c

@@ -294,6 +294,7 @@ void btrfs_print_leaf(struct btrfs_root *root, struct extent_buffer *l)
 			       btrfs_dev_extent_chunk_offset(l, dev_extent),
 			       (unsigned long long)
 			       btrfs_dev_extent_length(l, dev_extent));
+			break;
 		case BTRFS_DEV_STATS_KEY:
 			printk(KERN_INFO "\t\tdevice stats\n");
 			break;