Merge branch 'for-linus-4.6' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs

Pull btrfs updates from Chris Mason:
 "We have a good sized cleanup of our internal read ahead code, and the
  first series of commits from Chandan to enable PAGE_SIZE > sectorsize

  Otherwise, it's a normal series of cleanups and fixes, with many
  thanks to Dave Sterba for doing most of the patch wrangling this time"

* 'for-linus-4.6' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs: (82 commits)
  btrfs: make sure we stay inside the bvec during __btrfs_lookup_bio_sums
  btrfs: Fix misspellings in comments.
  btrfs: Print Warning only if ENOSPC_DEBUG is enabled
  btrfs: scrub: silence an uninitialized variable warning
  btrfs: move btrfs_compression_type to compression.h
  btrfs: rename btrfs_print_info to btrfs_print_mod_info
  Btrfs: Show a warning message if one of objectid reaches its highest value
  Documentation: btrfs: remove usage specific information
  btrfs: use kbasename in btrfsic_mount
  Btrfs: do not collect ordered extents when logging that inode exists
  Btrfs: fix race when checking if we can skip fsync'ing an inode
  Btrfs: fix listxattrs not listing all xattrs packed in the same item
  Btrfs: fix deadlock between direct IO reads and buffered writes
  Btrfs: fix extent_same allowing destination offset beyond i_size
  Btrfs: fix file loss on log replay after renaming a file and fsync
  Btrfs: fix unreplayable log after snapshot delete + parent dir fsync
  Btrfs: fix lockdep deadlock warning due to dev_replace
  btrfs: drop unused argument in btrfs_ioctl_get_supported_features
  btrfs: add GET_SUPPORTED_FEATURES to the control device ioctls
  btrfs: change max_inline default to 2048
  ...

35 files changed, 1091 insertions, 681 deletions

diff --git a/fs/btrfs/backref.c b/fs/btrfs/backref.c
index f6dac40f87ff..80e8472d618b 100644
--- a/fs/btrfs/backref.c
+++ b/fs/btrfs/backref.c
@@ -148,8 +148,7 @@ int __init btrfs_prelim_ref_init(void)
 
 void btrfs_prelim_ref_exit(void)
 {
-	if (btrfs_prelim_ref_cache)
-		kmem_cache_destroy(btrfs_prelim_ref_cache);
+	kmem_cache_destroy(btrfs_prelim_ref_cache);
 }
 
 /*
@@ -566,17 +565,14 @@ static void __merge_refs(struct list_head *head, int mode)
 		struct __prelim_ref *pos2 = pos1, *tmp;
 
 		list_for_each_entry_safe_continue(pos2, tmp, head, list) {
-			struct __prelim_ref *xchg, *ref1 = pos1, *ref2 = pos2;
+			struct __prelim_ref *ref1 = pos1, *ref2 = pos2;
 			struct extent_inode_elem *eie;
 
 			if (!ref_for_same_block(ref1, ref2))
 				continue;
 			if (mode == 1) {
-				if (!ref1->parent && ref2->parent) {
-					xchg = ref1;
-					ref1 = ref2;
-					ref2 = xchg;
-				}
+				if (!ref1->parent && ref2->parent)
+					swap(ref1, ref2);
 			} else {
 				if (ref1->parent != ref2->parent)
 					continue;
diff --git a/fs/btrfs/check-integrity.c b/fs/btrfs/check-integrity.c
index 861d472564c1..e34a71b3e225 100644
--- a/fs/btrfs/check-integrity.c
+++ b/fs/btrfs/check-integrity.c
@@ -95,6 +95,7 @@
 #include <linux/genhd.h>
 #include <linux/blkdev.h>
 #include <linux/vmalloc.h>
+#include <linux/string.h>
 #include "ctree.h"
 #include "disk-io.h"
 #include "hash.h"
@@ -105,6 +106,7 @@
 #include "locking.h"
 #include "check-integrity.h"
 #include "rcu-string.h"
+#include "compression.h"
 
 #define BTRFSIC_BLOCK_HASHTABLE_SIZE 0x10000
 #define BTRFSIC_BLOCK_LINK_HASHTABLE_SIZE 0x10000
@@ -176,7 +178,7 @@ struct btrfsic_block {
  * Elements of this type are allocated dynamically and required because
  * each block object can refer to and can be ref from multiple blocks.
  * The key to lookup them in the hashtable is the dev_bytenr of
- * the block ref to plus the one from the block refered from.
+ * the block ref to plus the one from the block referred from.
  * The fact that they are searchable via a hashtable and that a
  * ref_cnt is maintained is not required for the btrfs integrity
  * check algorithm itself, it is only used to make the output more
@@ -3076,7 +3078,7 @@ int btrfsic_mount(struct btrfs_root *root,
 
 	list_for_each_entry(device, dev_head, dev_list) {
 		struct btrfsic_dev_state *ds;
-		char *p;
+		const char *p;
 
 		if (!device->bdev || !device->name)
 			continue;
@@ -3092,11 +3094,7 @@ int btrfsic_mount(struct btrfs_root *root,
 		ds->state = state;
 		bdevname(ds->bdev, ds->name);
 		ds->name[BDEVNAME_SIZE - 1] = '\0';
-		for (p = ds->name; *p != '\0'; p++);
-		while (p > ds->name && *p != '/')
-			p--;
-		if (*p == '/')
-			p++;
+		p = kbasename(ds->name);
 		strlcpy(ds->name, p, sizeof(ds->name));
 		btrfsic_dev_state_hashtable_add(ds,
 						&btrfsic_dev_state_hashtable);
diff --git a/fs/btrfs/compression.h b/fs/btrfs/compression.h
index 13a4dc0436c9..f49d8b8c0f00 100644
--- a/fs/btrfs/compression.h
+++ b/fs/btrfs/compression.h
@@ -48,6 +48,15 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio,
 void btrfs_clear_biovec_end(struct bio_vec *bvec, int vcnt,
 				   unsigned long pg_index,
 				   unsigned long pg_offset);
+
+enum btrfs_compression_type {
+	BTRFS_COMPRESS_NONE  = 0,
+	BTRFS_COMPRESS_ZLIB  = 1,
+	BTRFS_COMPRESS_LZO   = 2,
+	BTRFS_COMPRESS_TYPES = 2,
+	BTRFS_COMPRESS_LAST  = 3,
+};
+
 struct btrfs_compress_op {
 	struct list_head *(*alloc_workspace)(void);
 
diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c
index 769e0ff1b4ce..77592931ab4f 100644
--- a/fs/btrfs/ctree.c
+++ b/fs/btrfs/ctree.c
@@ -311,7 +311,7 @@ struct tree_mod_root {
 
 struct tree_mod_elem {
 	struct rb_node node;
-	u64 index;		/* shifted logical */
+	u64 logical;
 	u64 seq;
 	enum mod_log_op op;
 
@@ -435,11 +435,11 @@ void btrfs_put_tree_mod_seq(struct btrfs_fs_info *fs_info,
 
 /*
  * key order of the log:
- *       index -> sequence
+ *       node/leaf start address -> sequence
  *
- * the index is the shifted logical of the *new* root node for root replace
- * operations, or the shifted logical of the affected block for all other
- * operations.
+ * The 'start address' is the logical address of the *new* root node
+ * for root replace operations, or the logical address of the affected
+ * block for all other operations.
  *
  * Note: must be called with write lock (tree_mod_log_write_lock).
  */
@@ -460,9 +460,9 @@ __tree_mod_log_insert(struct btrfs_fs_info *fs_info, struct tree_mod_elem *tm)
 	while (*new) {
 		cur = container_of(*new, struct tree_mod_elem, node);
 		parent = *new;
-		if (cur->index < tm->index)
+		if (cur->logical < tm->logical)
 			new = &((*new)->rb_left);
-		else if (cur->index > tm->index)
+		else if (cur->logical > tm->logical)
 			new = &((*new)->rb_right);
 		else if (cur->seq < tm->seq)
 			new = &((*new)->rb_left);
@@ -523,7 +523,7 @@ alloc_tree_mod_elem(struct extent_buffer *eb, int slot,
 	if (!tm)
 		return NULL;
 
-	tm->index = eb->start >> PAGE_CACHE_SHIFT;
+	tm->logical = eb->start;
 	if (op != MOD_LOG_KEY_ADD) {
 		btrfs_node_key(eb, &tm->key, slot);
 		tm->blockptr = btrfs_node_blockptr(eb, slot);
@@ -588,7 +588,7 @@ tree_mod_log_insert_move(struct btrfs_fs_info *fs_info,
 		goto free_tms;
 	}
 
-	tm->index = eb->start >> PAGE_CACHE_SHIFT;
+	tm->logical = eb->start;
 	tm->slot = src_slot;
 	tm->move.dst_slot = dst_slot;
 	tm->move.nr_items = nr_items;
@@ -699,7 +699,7 @@ tree_mod_log_insert_root(struct btrfs_fs_info *fs_info,
 		goto free_tms;
 	}
 
-	tm->index = new_root->start >> PAGE_CACHE_SHIFT;
+	tm->logical = new_root->start;
 	tm->old_root.logical = old_root->start;
 	tm->old_root.level = btrfs_header_level(old_root);
 	tm->generation = btrfs_header_generation(old_root);
@@ -739,16 +739,15 @@ __tree_mod_log_search(struct btrfs_fs_info *fs_info, u64 start, u64 min_seq,
 	struct rb_node *node;
 	struct tree_mod_elem *cur = NULL;
 	struct tree_mod_elem *found = NULL;
-	u64 index = start >> PAGE_CACHE_SHIFT;
 
 	tree_mod_log_read_lock(fs_info);
 	tm_root = &fs_info->tree_mod_log;
 	node = tm_root->rb_node;
 	while (node) {
 		cur = container_of(node, struct tree_mod_elem, node);
-		if (cur->index < index) {
+		if (cur->logical < start) {
 			node = node->rb_left;
-		} else if (cur->index > index) {
+		} else if (cur->logical > start) {
 			node = node->rb_right;
 		} else if (cur->seq < min_seq) {
 			node = node->rb_left;
@@ -1230,9 +1229,10 @@ __tree_mod_log_oldest_root(struct btrfs_fs_info *fs_info,
 		return NULL;
 
 	/*
-	 * the very last operation that's logged for a root is the replacement
-	 * operation (if it is replaced at all). this has the index of the *new*
-	 * root, making it the very first operation that's logged for this root.
+	 * the very last operation that's logged for a root is the
+	 * replacement operation (if it is replaced at all). this has
+	 * the logical address of the *new* root, making it the very
+	 * first operation that's logged for this root.
 	 */
 	while (1) {
 		tm = tree_mod_log_search_oldest(fs_info, root_logical,
@@ -1336,7 +1336,7 @@ __tree_mod_log_rewind(struct btrfs_fs_info *fs_info, struct extent_buffer *eb,
 		if (!next)
 			break;
 		tm = container_of(next, struct tree_mod_elem, node);
-		if (tm->index != first_tm->index)
+		if (tm->logical != first_tm->logical)
 			break;
 	}
 	tree_mod_log_read_unlock(fs_info);
@@ -5361,7 +5361,7 @@ int btrfs_compare_trees(struct btrfs_root *left_root,
 		goto out;
 	}
 
-	tmp_buf = kmalloc(left_root->nodesize, GFP_NOFS);
+	tmp_buf = kmalloc(left_root->nodesize, GFP_KERNEL);
 	if (!tmp_buf) {
 		ret = -ENOMEM;
 		goto out;
diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index bfe4a337fb4d..84a6a5b3384a 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -100,6 +100,9 @@ struct btrfs_ordered_sum;
 /* tracks free space in block groups. */
 #define BTRFS_FREE_SPACE_TREE_OBJECTID 10ULL
 
+/* device stats in the device tree */
+#define BTRFS_DEV_STATS_OBJECTID 0ULL
+
 /* for storing balance parameters in the root tree */
 #define BTRFS_BALANCE_OBJECTID -4ULL
 
@@ -715,14 +718,6 @@ struct btrfs_timespec {
 	__le32 nsec;
 } __attribute__ ((__packed__));
 
-enum btrfs_compression_type {
-	BTRFS_COMPRESS_NONE  = 0,
-	BTRFS_COMPRESS_ZLIB  = 1,
-	BTRFS_COMPRESS_LZO   = 2,
-	BTRFS_COMPRESS_TYPES = 2,
-	BTRFS_COMPRESS_LAST  = 3,
-};
-
 struct btrfs_inode_item {
 	/* nfs style generation number */
 	__le64 generation;
@@ -793,7 +788,7 @@ struct btrfs_root_item {
 
 	/*
 	 * This generation number is used to test if the new fields are valid
-	 * and up to date while reading the root item. Everytime the root item
+	 * and up to date while reading the root item. Every time the root item
 	 * is written out, the "generation" field is copied into this field. If
 	 * anyone ever mounted the fs with an older kernel, we will have
 	 * mismatching generation values here and thus must invalidate the
@@ -1002,8 +997,10 @@ struct btrfs_dev_replace {
 	pid_t lock_owner;
 	atomic_t nesting_level;
 	struct mutex lock_finishing_cancel_unmount;
-	struct mutex lock_management_lock;
-	struct mutex lock;
+	rwlock_t lock;
+	atomic_t read_locks;
+	atomic_t blocking_readers;
+	wait_queue_head_t read_lock_wq;
 
 	struct btrfs_scrub_progress scrub_progress;
 };
@@ -1222,10 +1219,10 @@ struct btrfs_space_info {
 	 * we've called update_block_group and dropped the bytes_used counter
 	 * and increased the bytes_pinned counter.  However this means that
 	 * bytes_pinned does not reflect the bytes that will be pinned once the
-	 * delayed refs are flushed, so this counter is inc'ed everytime we call
-	 * btrfs_free_extent so it is a realtime count of what will be freed
-	 * once the transaction is committed.  It will be zero'ed everytime the
-	 * transaction commits.
+	 * delayed refs are flushed, so this counter is inc'ed every time we
+	 * call btrfs_free_extent so it is a realtime count of what will be
+	 * freed once the transaction is committed.  It will be zero'ed every
+	 * time the transaction commits.
 	 */
 	struct percpu_counter total_bytes_pinned;
 
@@ -1822,6 +1819,9 @@ struct btrfs_fs_info {
 	spinlock_t reada_lock;
 	struct radix_tree_root reada_tree;
 
+	/* readahead works cnt */
+	atomic_t reada_works_cnt;
+
 	/* Extent buffer radix tree */
 	spinlock_t buffer_lock;
 	struct radix_tree_root buffer_radix;
@@ -2185,13 +2185,43 @@ struct btrfs_ioctl_defrag_range_args {
  */
 #define BTRFS_QGROUP_RELATION_KEY       246
 
+/*
+ * Obsolete name, see BTRFS_TEMPORARY_ITEM_KEY.
+ */
 #define BTRFS_BALANCE_ITEM_KEY	248
 
 /*
- * Persistantly stores the io stats in the device tree.
- * One key for all stats, (0, BTRFS_DEV_STATS_KEY, devid).
+ * The key type for tree items that are stored persistently, but do not need to
+ * exist for extended period of time. The items can exist in any tree.
+ *
+ * [subtype, BTRFS_TEMPORARY_ITEM_KEY, data]
+ *
+ * Existing items:
+ *
+ * - balance status item
+ *   (BTRFS_BALANCE_OBJECTID, BTRFS_TEMPORARY_ITEM_KEY, 0)
  */
-#define BTRFS_DEV_STATS_KEY	249
+#define BTRFS_TEMPORARY_ITEM_KEY	248
+
+/*
+ * Obsolete name, see BTRFS_PERSISTENT_ITEM_KEY
+ */
+#define BTRFS_DEV_STATS_KEY		249
+
+/*
+ * The key type for tree items that are stored persistently and usually exist
+ * for a long period, eg. filesystem lifetime. The item kinds can be status
+ * information, stats or preference values. The item can exist in any tree.
+ *
+ * [subtype, BTRFS_PERSISTENT_ITEM_KEY, data]
+ *
+ * Existing items:
+ *
+ * - device statistics, store IO stats in the device tree, one key for all
+ *   stats
+ *   (BTRFS_DEV_STATS_OBJECTID, BTRFS_DEV_STATS_KEY, 0)
+ */
+#define BTRFS_PERSISTENT_ITEM_KEY	249
 
 /*
  * Persistantly stores the device replace state in the device tree.
@@ -2241,7 +2271,7 @@ struct btrfs_ioctl_defrag_range_args {
 #define BTRFS_MOUNT_ENOSPC_DEBUG	 (1 << 15)
 #define BTRFS_MOUNT_AUTO_DEFRAG		(1 << 16)
 #define BTRFS_MOUNT_INODE_MAP_CACHE	(1 << 17)
-#define BTRFS_MOUNT_RECOVERY		(1 << 18)
+#define BTRFS_MOUNT_USEBACKUPROOT	(1 << 18)
 #define BTRFS_MOUNT_SKIP_BALANCE	(1 << 19)
 #define BTRFS_MOUNT_CHECK_INTEGRITY	(1 << 20)
 #define BTRFS_MOUNT_CHECK_INTEGRITY_INCLUDING_EXTENT_DATA (1 << 21)
@@ -2250,9 +2280,10 @@ struct btrfs_ioctl_defrag_range_args {
 #define BTRFS_MOUNT_FRAGMENT_DATA	(1 << 24)
 #define BTRFS_MOUNT_FRAGMENT_METADATA	(1 << 25)
 #define BTRFS_MOUNT_FREE_SPACE_TREE	(1 << 26)
+#define BTRFS_MOUNT_NOLOGREPLAY		(1 << 27)
 
 #define BTRFS_DEFAULT_COMMIT_INTERVAL	(30)
-#define BTRFS_DEFAULT_MAX_INLINE	(8192)
+#define BTRFS_DEFAULT_MAX_INLINE	(2048)
 
 #define btrfs_clear_opt(o, opt)		((o) &= ~BTRFS_MOUNT_##opt)
 #define btrfs_set_opt(o, opt)		((o) |= BTRFS_MOUNT_##opt)
@@ -2353,6 +2384,9 @@ struct btrfs_map_token {
 	unsigned long offset;
 };
 
+#define BTRFS_BYTES_TO_BLKS(fs_info, bytes) \
+				((bytes) >> (fs_info)->sb->s_blocksize_bits)
+
 static inline void btrfs_init_map_token (struct btrfs_map_token *token)
 {
 	token->kaddr = NULL;
@@ -3448,8 +3482,7 @@ u64 btrfs_csum_bytes_to_leaves(struct btrfs_root *root, u64 csum_bytes);
 static inline u64 btrfs_calc_trans_metadata_size(struct btrfs_root *root,
 						 unsigned num_items)
 {
-	return (root->nodesize + root->nodesize * (BTRFS_MAX_LEVEL - 1)) *
-		2 * num_items;
+	return root->nodesize * BTRFS_MAX_LEVEL * 2 * num_items;
 }
 
 /*
@@ -4027,7 +4060,7 @@ int btrfs_unlink_subvol(struct btrfs_trans_handle *trans,
 			struct btrfs_root *root,
 			struct inode *dir, u64 objectid,
 			const char *name, int name_len);
-int btrfs_truncate_page(struct inode *inode, loff_t from, loff_t len,
+int btrfs_truncate_block(struct inode *inode, loff_t from, loff_t len,
 			int front);
 int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans,
 			       struct btrfs_root *root,
@@ -4089,6 +4122,7 @@ void btrfs_test_inode_set_ops(struct inode *inode);
 
 /* ioctl.c */
 long btrfs_ioctl(struct file *file, unsigned int cmd, unsigned long arg);
+int btrfs_ioctl_get_supported_features(void __user *arg);
 void btrfs_update_iflags(struct inode *inode);
 void btrfs_inherit_iflags(struct inode *inode, struct inode *dir);
 int btrfs_is_empty_uuid(u8 *uuid);
@@ -4151,7 +4185,8 @@ void btrfs_sysfs_remove_mounted(struct btrfs_fs_info *fs_info);
 ssize_t btrfs_listxattr(struct dentry *dentry, char *buffer, size_t size);
 
 /* super.c */
-int btrfs_parse_options(struct btrfs_root *root, char *options);
+int btrfs_parse_options(struct btrfs_root *root, char *options,
+			unsigned long new_flags);
 int btrfs_sync_fs(struct super_block *sb, int wait);
 
 #ifdef CONFIG_PRINTK
@@ -4525,8 +4560,8 @@ struct reada_control *btrfs_reada_add(struct btrfs_root *root,
 			      struct btrfs_key *start, struct btrfs_key *end);
 int btrfs_reada_wait(void *handle);
 void btrfs_reada_detach(void *handle);
-int btree_readahead_hook(struct btrfs_root *root, struct extent_buffer *eb,
-			 u64 start, int err);
+int btree_readahead_hook(struct btrfs_fs_info *fs_info,
+			 struct extent_buffer *eb, u64 start, int err);
 
 static inline int is_fstree(u64 rootid)
 {
diff --git a/fs/btrfs/delayed-inode.c b/fs/btrfs/delayed-inode.c
index b57daa895cea..6cef0062f929 100644
--- a/fs/btrfs/delayed-inode.c
+++ b/fs/btrfs/delayed-inode.c
@@ -43,8 +43,7 @@ int __init btrfs_delayed_inode_init(void)
 
 void btrfs_delayed_inode_exit(void)
 {
-	if (delayed_node_cache)
-		kmem_cache_destroy(delayed_node_cache);
+	kmem_cache_destroy(delayed_node_cache);
 }
 
 static inline void btrfs_init_delayed_node(
@@ -651,9 +650,14 @@ static int btrfs_delayed_inode_reserve_metadata(
 			goto out;
 
 		ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes);
-		if (!WARN_ON(ret))
+		if (!ret)
 			goto out;
 
+		if (btrfs_test_opt(root, ENOSPC_DEBUG)) {
+			btrfs_debug(root->fs_info,
+				    "block rsv migrate returned %d", ret);
+			WARN_ON(1);
+		}
 		/*
 		 * Ok this is a problem, let's just steal from the global rsv
 		 * since this really shouldn't happen that often.
diff --git a/fs/btrfs/delayed-ref.c b/fs/btrfs/delayed-ref.c
index 914ac13bd92f..430b3689b112 100644
--- a/fs/btrfs/delayed-ref.c
+++ b/fs/btrfs/delayed-ref.c
@@ -929,14 +929,10 @@ btrfs_find_delayed_ref_head(struct btrfs_trans_handle *trans, u64 bytenr)
 
 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);
+	kmem_cache_destroy(btrfs_delayed_ref_head_cachep);
+	kmem_cache_destroy(btrfs_delayed_tree_ref_cachep);
+	kmem_cache_destroy(btrfs_delayed_data_ref_cachep);
+	kmem_cache_destroy(btrfs_delayed_extent_op_cachep);
 }
 
 int btrfs_delayed_ref_init(void)
diff --git a/fs/btrfs/dev-replace.c b/fs/btrfs/dev-replace.c
index cbb7dbfb3fff..a1d6652e0c47 100644
--- a/fs/btrfs/dev-replace.c
+++ b/fs/btrfs/dev-replace.c
@@ -202,13 +202,13 @@ int btrfs_run_dev_replace(struct btrfs_trans_handle *trans,
 	struct btrfs_dev_replace_item *ptr;
 	struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
 
-	btrfs_dev_replace_lock(dev_replace);
+	btrfs_dev_replace_lock(dev_replace, 0);
 	if (!dev_replace->is_valid ||
 	    !dev_replace->item_needs_writeback) {
-		btrfs_dev_replace_unlock(dev_replace);
+		btrfs_dev_replace_unlock(dev_replace, 0);
 		return 0;
 	}
-	btrfs_dev_replace_unlock(dev_replace);
+	btrfs_dev_replace_unlock(dev_replace, 0);
 
 	key.objectid = 0;
 	key.type = BTRFS_DEV_REPLACE_KEY;
@@ -264,7 +264,7 @@ int btrfs_run_dev_replace(struct btrfs_trans_handle *trans,
 	ptr = btrfs_item_ptr(eb, path->slots[0],
 			     struct btrfs_dev_replace_item);
 
-	btrfs_dev_replace_lock(dev_replace);
+	btrfs_dev_replace_lock(dev_replace, 1);
 	if (dev_replace->srcdev)
 		btrfs_set_dev_replace_src_devid(eb, ptr,
 			dev_replace->srcdev->devid);
@@ -287,7 +287,7 @@ int btrfs_run_dev_replace(struct btrfs_trans_handle *trans,
 	btrfs_set_dev_replace_cursor_right(eb, ptr,
 		dev_replace->cursor_right);
 	dev_replace->item_needs_writeback = 0;
-	btrfs_dev_replace_unlock(dev_replace);
+	btrfs_dev_replace_unlock(dev_replace, 1);
 
 	btrfs_mark_buffer_dirty(eb);
 
@@ -356,7 +356,7 @@ int btrfs_dev_replace_start(struct btrfs_root *root,
 		return PTR_ERR(trans);
 	}
 
-	btrfs_dev_replace_lock(dev_replace);
+	btrfs_dev_replace_lock(dev_replace, 1);
 	switch (dev_replace->replace_state) {
 	case BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED:
 	case BTRFS_IOCTL_DEV_REPLACE_STATE_FINISHED:
@@ -395,7 +395,7 @@ int btrfs_dev_replace_start(struct btrfs_root *root,
 	dev_replace->is_valid = 1;
 	dev_replace->item_needs_writeback = 1;
 	args->result = BTRFS_IOCTL_DEV_REPLACE_RESULT_NO_ERROR;
-	btrfs_dev_replace_unlock(dev_replace);
+	btrfs_dev_replace_unlock(dev_replace, 1);
 
 	ret = btrfs_sysfs_add_device_link(tgt_device->fs_devices, tgt_device);
 	if (ret)
@@ -407,7 +407,7 @@ int btrfs_dev_replace_start(struct btrfs_root *root,
 	trans = btrfs_start_transaction(root, 0);
 	if (IS_ERR(trans)) {
 		ret = PTR_ERR(trans);
-		btrfs_dev_replace_lock(dev_replace);
+		btrfs_dev_replace_lock(dev_replace, 1);
 		goto leave;
 	}
 
@@ -433,7 +433,7 @@ int btrfs_dev_replace_start(struct btrfs_root *root,
 leave:
 	dev_replace->srcdev = NULL;
 	dev_replace->tgtdev = NULL;
-	btrfs_dev_replace_unlock(dev_replace);
+	btrfs_dev_replace_unlock(dev_replace, 1);
 	btrfs_destroy_dev_replace_tgtdev(fs_info, tgt_device);
 	return ret;
 }
@@ -471,18 +471,18 @@ static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
 	/* don't allow cancel or unmount to disturb the finishing procedure */
 	mutex_lock(&dev_replace->lock_finishing_cancel_unmount);
 
-	btrfs_dev_replace_lock(dev_replace);
+	btrfs_dev_replace_lock(dev_replace, 0);
 	/* was the operation canceled, or is it finished? */
 	if (dev_replace->replace_state !=
 	    BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED) {
-		btrfs_dev_replace_unlock(dev_replace);
+		btrfs_dev_replace_unlock(dev_replace, 0);
 		mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
 		return 0;
 	}
 
 	tgt_device = dev_replace->tgtdev;
 	src_device = dev_replace->srcdev;
-	btrfs_dev_replace_unlock(dev_replace);
+	btrfs_dev_replace_unlock(dev_replace, 0);
 
 	/*
 	 * flush all outstanding I/O and inode extent mappings before the
@@ -507,7 +507,7 @@ static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
 	/* keep away write_all_supers() during the finishing procedure */
 	mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
 	mutex_lock(&root->fs_info->chunk_mutex);
-	btrfs_dev_replace_lock(dev_replace);
+	btrfs_dev_replace_lock(dev_replace, 1);
 	dev_replace->replace_state =
 		scrub_ret ? BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED
 			  : BTRFS_IOCTL_DEV_REPLACE_STATE_FINISHED;
@@ -528,7 +528,7 @@ static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
 			        rcu_str_deref(src_device->name),
 			      src_device->devid,
 			      rcu_str_deref(tgt_device->name), scrub_ret);
-		btrfs_dev_replace_unlock(dev_replace);
+		btrfs_dev_replace_unlock(dev_replace, 1);
 		mutex_unlock(&root->fs_info->chunk_mutex);
 		mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
 		mutex_unlock(&uuid_mutex);
@@ -565,7 +565,7 @@ static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
 	list_add(&tgt_device->dev_alloc_list, &fs_info->fs_devices->alloc_list);
 	fs_info->fs_devices->rw_devices++;
 
-	btrfs_dev_replace_unlock(dev_replace);
+	btrfs_dev_replace_unlock(dev_replace, 1);
 
 	btrfs_rm_dev_replace_blocked(fs_info);
 
@@ -649,7 +649,7 @@ void btrfs_dev_replace_status(struct btrfs_fs_info *fs_info,
 	struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
 	struct btrfs_device *srcdev;
 
-	btrfs_dev_replace_lock(dev_replace);
+	btrfs_dev_replace_lock(dev_replace, 0);
 	/* even if !dev_replace_is_valid, the values are good enough for
 	 * the replace_status ioctl */
 	args->result = BTRFS_IOCTL_DEV_REPLACE_RESULT_NO_ERROR;
@@ -675,7 +675,7 @@ void btrfs_dev_replace_status(struct btrfs_fs_info *fs_info,
 			div_u64(btrfs_device_get_total_bytes(srcdev), 1000));
 		break;
 	}
-	btrfs_dev_replace_unlock(dev_replace);
+	btrfs_dev_replace_unlock(dev_replace, 0);
 }
 
 int btrfs_dev_replace_cancel(struct btrfs_fs_info *fs_info,
@@ -698,13 +698,13 @@ static u64 __btrfs_dev_replace_cancel(struct btrfs_fs_info *fs_info)
 		return -EROFS;
 
 	mutex_lock(&dev_replace->lock_finishing_cancel_unmount);
-	btrfs_dev_replace_lock(dev_replace);
+	btrfs_dev_replace_lock(dev_replace, 1);
 	switch (dev_replace->replace_state) {
 	case BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED:
 	case BTRFS_IOCTL_DEV_REPLACE_STATE_FINISHED:
 	case BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED:
 		result = BTRFS_IOCTL_DEV_REPLACE_RESULT_NOT_STARTED;
-		btrfs_dev_replace_unlock(dev_replace);
+		btrfs_dev_replace_unlock(dev_replace, 1);
 		goto leave;
 	case BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED:
 	case BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED:
@@ -717,7 +717,7 @@ static u64 __btrfs_dev_replace_cancel(struct btrfs_fs_info *fs_info)
 	dev_replace->replace_state = BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED;
 	dev_replace->time_stopped = get_seconds();
 	dev_replace->item_needs_writeback = 1;
-	btrfs_dev_replace_unlock(dev_replace);
+	btrfs_dev_replace_unlock(dev_replace, 1);
 	btrfs_scrub_cancel(fs_info);
 
 	trans = btrfs_start_transaction(root, 0);
@@ -740,7 +740,7 @@ void btrfs_dev_replace_suspend_for_unmount(struct btrfs_fs_info *fs_info)
 	struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
 
 	mutex_lock(&dev_replace->lock_finishing_cancel_unmount);
-	btrfs_dev_replace_lock(dev_replace);
+	btrfs_dev_replace_lock(dev_replace, 1);
 	switch (dev_replace->replace_state) {
 	case BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED:
 	case BTRFS_IOCTL_DEV_REPLACE_STATE_FINISHED:
@@ -756,7 +756,7 @@ void btrfs_dev_replace_suspend_for_unmount(struct btrfs_fs_info *fs_info)
 		break;
 	}
 
-	btrfs_dev_replace_unlock(dev_replace);
+	btrfs_dev_replace_unlock(dev_replace, 1);
 	mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
 }
 
@@ -766,12 +766,12 @@ int btrfs_resume_dev_replace_async(struct btrfs_fs_info *fs_info)
 	struct task_struct *task;
 	struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
 
-	btrfs_dev_replace_lock(dev_replace);
+	btrfs_dev_replace_lock(dev_replace, 1);
 	switch (dev_replace->replace_state) {
 	case BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED:
 	case BTRFS_IOCTL_DEV_REPLACE_STATE_FINISHED:
 	case BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED:
-		btrfs_dev_replace_unlock(dev_replace);
+		btrfs_dev_replace_unlock(dev_replace, 1);
 		return 0;
 	case BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED:
 		break;
@@ -784,10 +784,10 @@ int btrfs_resume_dev_replace_async(struct btrfs_fs_info *fs_info)
 		btrfs_info(fs_info, "cannot continue dev_replace, tgtdev is missing");
 		btrfs_info(fs_info,
 			"you may cancel the operation after 'mount -o degraded'");
-		btrfs_dev_replace_unlock(dev_replace);
+		btrfs_dev_replace_unlock(dev_replace, 1);
 		return 0;
 	}
-	btrfs_dev_replace_unlock(dev_replace);
+	btrfs_dev_replace_unlock(dev_replace, 1);
 
 	WARN_ON(atomic_xchg(
 		&fs_info->mutually_exclusive_operation_running, 1));
@@ -802,7 +802,7 @@ static int btrfs_dev_replace_kthread(void *data)
 	struct btrfs_ioctl_dev_replace_args *status_args;
 	u64 progress;
 
-	status_args = kzalloc(sizeof(*status_args), GFP_NOFS);
+	status_args = kzalloc(sizeof(*status_args), GFP_KERNEL);
 	if (status_args) {
 		btrfs_dev_replace_status(fs_info, status_args);
 		progress = status_args->status.progress_1000;
@@ -858,55 +858,65 @@ int btrfs_dev_replace_is_ongoing(struct btrfs_dev_replace *dev_replace)
 		 * not called and the the filesystem is remounted
 		 * in degraded state. This does not stop the
 		 * dev_replace procedure. It needs to be canceled
-		 * manually if the cancelation is wanted.
+		 * manually if the cancellation is wanted.
 		 */
 		break;
 	}
 	return 1;
 }
 
-void btrfs_dev_replace_lock(struct btrfs_dev_replace *dev_replace)
+void btrfs_dev_replace_lock(struct btrfs_dev_replace *dev_replace, int rw)