1 files changed, 454 insertions, 170 deletions

diff --git a/fs/btrfs/dev-replace.c b/fs/btrfs/dev-replace.c
index 8750c835f535..b6c7da8e1bc8 100644
--- a/fs/btrfs/dev-replace.c
+++ b/fs/btrfs/dev-replace.c
@@ -9,44 +9,84 @@
 #include <linux/blkdev.h>
 #include <linux/kthread.h>
 #include <linux/math64.h>
+#include "misc.h"
 #include "ctree.h"
-#include "extent_map.h"
 #include "disk-io.h"
 #include "transaction.h"
-#include "print-tree.h"
 #include "volumes.h"
 #include "async-thread.h"
-#include "check-integrity.h"
-#include "rcu-string.h"
 #include "dev-replace.h"
 #include "sysfs.h"
+#include "zoned.h"
+#include "block-group.h"
+#include "fs.h"
+#include "accessors.h"
+#include "scrub.h"
+
+/*
+ * Device replace overview
+ *
+ * [Objective]
+ * To copy all extents (both new and on-disk) from source device to target
+ * device, while still keeping the filesystem read-write.
+ *
+ * [Method]
+ * There are two main methods involved:
+ *
+ * - Write duplication
+ *
+ *   All new writes will be written to both target and source devices, so even
+ *   if replace gets canceled, sources device still contains up-to-date data.
+ *
+ *   Location:		handle_ops_on_dev_replace() from btrfs_map_block()
+ *   Start:		btrfs_dev_replace_start()
+ *   End:		btrfs_dev_replace_finishing()
+ *   Content:		Latest data/metadata
+ *
+ * - Copy existing extents
+ *
+ *   This happens by reusing scrub facility, as scrub also iterates through
+ *   existing extents from commit root.
+ *
+ *   Location:		scrub_write_block_to_dev_replace() from
+ *   			scrub_block_complete()
+ *   Content:		Data/meta from commit root.
+ *
+ * Due to the content difference, we need to avoid nocow write when dev-replace
+ * is happening.  This is done by marking the block group read-only and waiting
+ * for NOCOW writes.
+ *
+ * After replace is done, the finishing part is done by swapping the target and
+ * source devices.
+ *
+ *   Location:		btrfs_dev_replace_update_device_in_mapping_tree() from
+ *   			btrfs_dev_replace_finishing()
+ */
 
 static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
 				       int scrub_ret);
-static void btrfs_dev_replace_update_device_in_mapping_tree(
-						struct btrfs_fs_info *fs_info,
-						struct btrfs_device *srcdev,
-						struct btrfs_device *tgtdev);
 static int btrfs_dev_replace_kthread(void *data);
 
 int btrfs_init_dev_replace(struct btrfs_fs_info *fs_info)
 {
+	struct btrfs_dev_lookup_args args = { .devid = BTRFS_DEV_REPLACE_DEVID };
 	struct btrfs_key key;
 	struct btrfs_root *dev_root = fs_info->dev_root;
 	struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
 	struct extent_buffer *eb;
 	int slot;
 	int ret = 0;
-	struct btrfs_path *path = NULL;
+	BTRFS_PATH_AUTO_FREE(path);
 	int item_size;
 	struct btrfs_dev_replace_item *ptr;
 	u64 src_devid;
 
+	if (!dev_root)
+		return 0;
+
 	path = btrfs_alloc_path();
-	if (!path) {
-		ret = -ENOMEM;
-		goto out;
-	}
+	if (!path)
+		return -ENOMEM;
 
 	key.objectid = 0;
 	key.type = BTRFS_DEV_REPLACE_KEY;
@@ -54,9 +94,17 @@ int btrfs_init_dev_replace(struct btrfs_fs_info *fs_info)
 	ret = btrfs_search_slot(NULL, dev_root, &key, path, 0, 0);
 	if (ret) {
 no_valid_dev_replace_entry_found:
-		ret = 0;
+		/*
+		 * We don't have a replace item or it's corrupted.  If there is
+		 * a replace target, fail the mount.
+		 */
+		if (unlikely(btrfs_find_device(fs_info->fs_devices, &args))) {
+			btrfs_err(fs_info,
+			"found replace target device without a valid replace item");
+			return -EUCLEAN;
+		}
 		dev_replace->replace_state =
-			BTRFS_DEV_REPLACE_ITEM_STATE_NEVER_STARTED;
+			BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED;
 		dev_replace->cont_reading_from_srcdev_mode =
 		    BTRFS_DEV_REPLACE_ITEM_CONT_READING_FROM_SRCDEV_MODE_ALWAYS;
 		dev_replace->time_started = 0;
@@ -71,11 +119,11 @@ no_valid_dev_replace_entry_found:
 		dev_replace->tgtdev = NULL;
 		dev_replace->is_valid = 0;
 		dev_replace->item_needs_writeback = 0;
-		goto out;
+		return 0;
 	}
 	slot = path->slots[0];
 	eb = path->nodes[0];
-	item_size = btrfs_item_size_nr(eb, slot);
+	item_size = btrfs_item_size(eb, slot);
 	ptr = btrfs_item_ptr(eb, slot, struct btrfs_dev_replace_item);
 
 	if (item_size != sizeof(struct btrfs_dev_replace_item)) {
@@ -106,22 +154,30 @@ no_valid_dev_replace_entry_found:
 	case BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED:
 	case BTRFS_IOCTL_DEV_REPLACE_STATE_FINISHED:
 	case BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED:
-		dev_replace->srcdev = NULL;
-		dev_replace->tgtdev = NULL;
+		/*
+		 * We don't have an active replace item but if there is a
+		 * replace target, fail the mount.
+		 */
+		if (unlikely(btrfs_find_device(fs_info->fs_devices, &args))) {
+			btrfs_err(fs_info,
+"replace without active item, run 'device scan --forget' on the target device");
+			ret = -EUCLEAN;
+		} else {
+			dev_replace->srcdev = NULL;
+			dev_replace->tgtdev = NULL;
+		}
 		break;
 	case BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED:
 	case BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED:
-		dev_replace->srcdev = btrfs_find_device(fs_info, src_devid,
-							NULL, NULL);
-		dev_replace->tgtdev = btrfs_find_device(fs_info,
-							BTRFS_DEV_REPLACE_DEVID,
-							NULL, NULL);
+		dev_replace->tgtdev = btrfs_find_device(fs_info->fs_devices, &args);
+		args.devid = src_devid;
+		dev_replace->srcdev = btrfs_find_device(fs_info->fs_devices, &args);
+
 		/*
 		 * allow 'btrfs dev replace_cancel' if src/tgt device is
 		 * missing
 		 */
-		if (!dev_replace->srcdev &&
-		    !btrfs_test_opt(fs_info, DEGRADED)) {
+		if (unlikely(!dev_replace->srcdev && !btrfs_test_opt(fs_info, DEGRADED))) {
 			ret = -EIO;
 			btrfs_warn(fs_info,
 			   "cannot mount because device replace operation is ongoing and");
@@ -129,8 +185,7 @@ no_valid_dev_replace_entry_found:
 			   "srcdev (devid %llu) is missing, need to run 'btrfs dev scan'?",
 			   src_devid);
 		}
-		if (!dev_replace->tgtdev &&
-		    !btrfs_test_opt(fs_info, DEGRADED)) {
+		if (unlikely(!dev_replace->tgtdev && !btrfs_test_opt(fs_info, DEGRADED))) {
 			ret = -EIO;
 			btrfs_warn(fs_info,
 			   "cannot mount because device replace operation is ongoing and");
@@ -165,8 +220,6 @@ no_valid_dev_replace_entry_found:
 		break;
 	}
 
-out:
-	btrfs_free_path(path);
 	return ret;
 }
 
@@ -181,30 +234,37 @@ static int btrfs_init_dev_replace_tgtdev(struct btrfs_fs_info *fs_info,
 				  struct btrfs_device *srcdev,
 				  struct btrfs_device **device_out)
 {
+	struct btrfs_fs_devices *fs_devices = fs_info->fs_devices;
 	struct btrfs_device *device;
+	struct file *bdev_file;
 	struct block_device *bdev;
-	struct list_head *devices;
-	struct rcu_string *name;
 	u64 devid = BTRFS_DEV_REPLACE_DEVID;
 	int ret = 0;
 
 	*device_out = NULL;
-	if (fs_info->fs_devices->seeding) {
+	if (srcdev->fs_devices->seeding) {
 		btrfs_err(fs_info, "the filesystem is a seed filesystem!");
 		return -EINVAL;
 	}
 
-	bdev = blkdev_get_by_path(device_path, FMODE_WRITE | FMODE_EXCL,
-				  fs_info->bdev_holder);
-	if (IS_ERR(bdev)) {
+	bdev_file = bdev_file_open_by_path(device_path, BLK_OPEN_WRITE,
+					   fs_info->sb, &fs_holder_ops);
+	if (IS_ERR(bdev_file)) {
 		btrfs_err(fs_info, "target device %s is invalid!", device_path);
-		return PTR_ERR(bdev);
+		return PTR_ERR(bdev_file);
 	}
+	bdev = file_bdev(bdev_file);
 
-	filemap_write_and_wait(bdev->bd_inode->i_mapping);
+	if (!btrfs_check_device_zone_type(fs_info, bdev)) {
+		btrfs_err(fs_info,
+		"dev-replace: zoned type of target device mismatch with filesystem");
+		ret = -EINVAL;
+		goto error;
+	}
 
-	devices = &fs_info->fs_devices->devices;
-	list_for_each_entry(device, devices, dev_list) {
+	sync_blockdev(bdev);
+
+	list_for_each_entry(device, &fs_devices->devices, dev_list) {
 		if (device->bdev == bdev) {
 			btrfs_err(fs_info,
 				  "target device is in the filesystem!");
@@ -214,8 +274,7 @@ static int btrfs_init_dev_replace_tgtdev(struct btrfs_fs_info *fs_info,
 	}
 
 
-	if (i_size_read(bdev->bd_inode) <
-	    btrfs_device_get_total_bytes(srcdev)) {
+	if (bdev_nr_bytes(bdev) < btrfs_device_get_total_bytes(srcdev)) {
 		btrfs_err(fs_info,
 			  "target device is smaller than source device!");
 		ret = -EINVAL;
@@ -223,21 +282,16 @@ static int btrfs_init_dev_replace_tgtdev(struct btrfs_fs_info *fs_info,
 	}
 
 
-	device = btrfs_alloc_device(NULL, &devid, NULL);
+	device = btrfs_alloc_device(NULL, &devid, NULL, device_path);
 	if (IS_ERR(device)) {
 		ret = PTR_ERR(device);
 		goto error;
 	}
 
-	name = rcu_string_strdup(device_path, GFP_KERNEL);
-	if (!name) {
-		btrfs_free_device(device);
-		ret = -ENOMEM;
+	ret = lookup_bdev(device_path, &device->devt);
+	if (ret)
 		goto error;
-	}
-	rcu_assign_pointer(device->name, name);
 
-	mutex_lock(&fs_info->fs_devices->device_list_mutex);
 	set_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state);
 	device->generation = 0;
 	device->io_width = fs_info->sectorsize;
@@ -250,22 +304,28 @@ static int btrfs_init_dev_replace_tgtdev(struct btrfs_fs_info *fs_info,
 	device->commit_bytes_used = device->bytes_used;
 	device->fs_info = fs_info;
 	device->bdev = bdev;
+	device->bdev_file = bdev_file;
 	set_bit(BTRFS_DEV_STATE_IN_FS_METADATA, &device->dev_state);
 	set_bit(BTRFS_DEV_STATE_REPLACE_TGT, &device->dev_state);
-	device->mode = FMODE_EXCL;
 	device->dev_stats_valid = 1;
-	set_blocksize(device->bdev, BTRFS_BDEV_BLOCKSIZE);
-	device->fs_devices = fs_info->fs_devices;
-	list_add(&device->dev_list, &fs_info->fs_devices->devices);
-	fs_info->fs_devices->num_devices++;
-	fs_info->fs_devices->open_devices++;
-	mutex_unlock(&fs_info->fs_devices->device_list_mutex);
+	set_blocksize(bdev_file, BTRFS_BDEV_BLOCKSIZE);
+	device->fs_devices = fs_devices;
+
+	ret = btrfs_get_dev_zone_info(device, false);
+	if (ret)
+		goto error;
+
+	mutex_lock(&fs_devices->device_list_mutex);
+	list_add(&device->dev_list, &fs_devices->devices);
+	fs_devices->num_devices++;
+	fs_devices->open_devices++;
+	mutex_unlock(&fs_devices->device_list_mutex);
 
 	*device_out = device;
 	return 0;
 
 error:
-	blkdev_put(bdev, FMODE_EXCL);
+	bdev_fput(bdev_file);
 	return ret;
 }
 
@@ -273,12 +333,12 @@ error:
  * called from commit_transaction. Writes changed device replace state to
  * disk.
  */
-int btrfs_run_dev_replace(struct btrfs_trans_handle *trans,
-			  struct btrfs_fs_info *fs_info)
+int btrfs_run_dev_replace(struct btrfs_trans_handle *trans)
 {
+	struct btrfs_fs_info *fs_info = trans->fs_info;
 	int ret;
 	struct btrfs_root *dev_root = fs_info->dev_root;
-	struct btrfs_path *path;
+	BTRFS_PATH_AUTO_FREE(path);
 	struct btrfs_key key;
 	struct extent_buffer *eb;
 	struct btrfs_dev_replace_item *ptr;
@@ -297,20 +357,19 @@ int btrfs_run_dev_replace(struct btrfs_trans_handle *trans,
 	key.offset = 0;
 
 	path = btrfs_alloc_path();
-	if (!path) {
-		ret = -ENOMEM;
-		goto out;
-	}
+	if (!path)
+		return -ENOMEM;
+
 	ret = btrfs_search_slot(trans, dev_root, &key, path, -1, 1);
 	if (ret < 0) {
 		btrfs_warn(fs_info,
 			   "error %d while searching for dev_replace item!",
 			   ret);
-		goto out;
+		return ret;
 	}
 
 	if (ret == 0 &&
-	    btrfs_item_size_nr(path->nodes[0], path->slots[0]) < sizeof(*ptr)) {
+	    btrfs_item_size(path->nodes[0], path->slots[0]) < sizeof(*ptr)) {
 		/*
 		 * need to delete old one and insert a new one.
 		 * Since no attempt is made to recover any old state, if the
@@ -327,7 +386,7 @@ int btrfs_run_dev_replace(struct btrfs_trans_handle *trans,
 			btrfs_warn(fs_info,
 				   "delete too small dev_replace item failed %d!",
 				   ret);
-			goto out;
+			return ret;
 		}
 		ret = 1;
 	}
@@ -340,7 +399,7 @@ int btrfs_run_dev_replace(struct btrfs_trans_handle *trans,
 		if (ret < 0) {
 			btrfs_warn(fs_info,
 				   "insert dev_replace item failed %d!", ret);
-			goto out;
+			return ret;
 		}
 	}
 
@@ -373,20 +432,153 @@ int btrfs_run_dev_replace(struct btrfs_trans_handle *trans,
 	dev_replace->item_needs_writeback = 0;
 	up_write(&dev_replace->rwsem);
 
-	btrfs_mark_buffer_dirty(eb);
+	return ret;
+}
+
+static int mark_block_group_to_copy(struct btrfs_fs_info *fs_info,
+				    struct btrfs_device *src_dev)
+{
+	struct btrfs_path *path;
+	struct btrfs_key key;
+	struct btrfs_key found_key;
+	struct btrfs_root *root = fs_info->dev_root;
+	struct btrfs_dev_extent *dev_extent = NULL;
+	struct btrfs_block_group *cache;
+	struct btrfs_trans_handle *trans;
+	int iter_ret = 0;
+	int ret = 0;
+	u64 chunk_offset;
+
+	/* Do not use "to_copy" on non zoned filesystem for now */
+	if (!btrfs_is_zoned(fs_info))
+		return 0;
+
+	mutex_lock(&fs_info->chunk_mutex);
+
+	/* Ensure we don't have pending new block group */
+	spin_lock(&fs_info->trans_lock);
+	while (fs_info->running_transaction &&
+	       !list_empty(&fs_info->running_transaction->dev_update_list)) {
+		spin_unlock(&fs_info->trans_lock);
+		mutex_unlock(&fs_info->chunk_mutex);
+		trans = btrfs_attach_transaction(root);
+		if (IS_ERR(trans)) {
+			ret = PTR_ERR(trans);
+			mutex_lock(&fs_info->chunk_mutex);
+			if (ret == -ENOENT) {
+				spin_lock(&fs_info->trans_lock);
+				continue;
+			} else {
+				goto unlock;
+			}
+		}
+
+		ret = btrfs_commit_transaction(trans);
+		mutex_lock(&fs_info->chunk_mutex);
+		if (ret)
+			goto unlock;
+
+		spin_lock(&fs_info->trans_lock);
+	}
+	spin_unlock(&fs_info->trans_lock);
+
+	path = btrfs_alloc_path();
+	if (!path) {
+		ret = -ENOMEM;
+		goto unlock;
+	}
+
+	path->reada = READA_FORWARD;
+	path->search_commit_root = true;
+	path->skip_locking = true;
+
+	key.objectid = src_dev->devid;
+	key.type = BTRFS_DEV_EXTENT_KEY;
+	key.offset = 0;
+
+	btrfs_for_each_slot(root, &key, &found_key, path, iter_ret) {
+		struct extent_buffer *leaf = path->nodes[0];
+
+		if (found_key.objectid != src_dev->devid)
+			break;
+
+		if (found_key.type != BTRFS_DEV_EXTENT_KEY)
+			break;
+
+		if (found_key.offset < key.offset)
+			break;
+
+		dev_extent = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_dev_extent);
+
+		chunk_offset = btrfs_dev_extent_chunk_offset(leaf, dev_extent);
+
+		cache = btrfs_lookup_block_group(fs_info, chunk_offset);
+		if (!cache)
+			continue;
+
+		set_bit(BLOCK_GROUP_FLAG_TO_COPY, &cache->runtime_flags);
+		btrfs_put_block_group(cache);
+	}
+	if (iter_ret < 0)
+		ret = iter_ret;
 
-out:
 	btrfs_free_path(path);
+unlock:
+	mutex_unlock(&fs_info->chunk_mutex);
 
 	return ret;
 }
 
-static char* btrfs_dev_name(struct btrfs_device *device)
+bool btrfs_finish_block_group_to_copy(struct btrfs_device *srcdev,
+				      struct btrfs_block_group *cache,
+				      u64 physical)
 {
-	if (!device || test_bit(BTRFS_DEV_STATE_MISSING, &device->dev_state))
-		return "<missing disk>";
-	else
-		return rcu_str_deref(device->name);
+	struct btrfs_fs_info *fs_info = cache->fs_info;
+	struct btrfs_chunk_map *map;
+	u64 chunk_offset = cache->start;
+	int num_extents, cur_extent;
+	int i;
+
+	/* Do not use "to_copy" on non zoned filesystem for now */
+	if (!btrfs_is_zoned(fs_info))
+		return true;
+
+	spin_lock(&cache->lock);
+	if (test_bit(BLOCK_GROUP_FLAG_REMOVED, &cache->runtime_flags)) {
+		spin_unlock(&cache->lock);
+		return true;
+	}
+	spin_unlock(&cache->lock);
+
+	map = btrfs_get_chunk_map(fs_info, chunk_offset, 1);
+	ASSERT(!IS_ERR(map));
+
+	num_extents = 0;
+	cur_extent = 0;
+	for (i = 0; i < map->num_stripes; i++) {
+		/* We have more device extent to copy */
+		if (srcdev != map->stripes[i].dev)
+			continue;
+
+		num_extents++;
+		if (physical == map->stripes[i].physical)
+			cur_extent = i;
+	}
+
+	btrfs_free_chunk_map(map);
+
+	if (num_extents > 1 && cur_extent < num_extents - 1) {
+		/*
+		 * Has more stripes on this device. Keep this block group
+		 * readonly until we finish all the stripes.
+		 */
+		return false;
+	}
+
+	/* Last stripe on this device */
+	clear_bit(BLOCK_GROUP_FLAG_TO_COPY, &cache->runtime_flags);
+
+	return true;
 }
 
 static int btrfs_dev_replace_start(struct btrfs_fs_info *fs_info,
@@ -399,7 +591,6 @@ static int btrfs_dev_replace_start(struct btrfs_fs_info *fs_info,
 	int ret;
 	struct btrfs_device *tgt_device = NULL;
 	struct btrfs_device *src_device = NULL;
-	bool need_unlock;
 
 	src_device = btrfs_find_device_by_devspec(fs_info, srcdevid,
 						  srcdev_name);
@@ -407,17 +598,12 @@ static int btrfs_dev_replace_start(struct btrfs_fs_info *fs_info,
 		return PTR_ERR(src_device);
 
 	if (btrfs_pinned_by_swapfile(fs_info, src_device)) {
-		btrfs_warn_in_rcu(fs_info,
+		btrfs_warn(fs_info,
 	  "cannot replace device %s (devid %llu) due to active swapfile",
 			btrfs_dev_name(src_device), src_device->devid);
 		return -ETXTBSY;
 	}
 
-	ret = btrfs_init_dev_replace_tgtdev(fs_info, tgtdev_name,
-					    src_device, &tgt_device);
-	if (ret)
-		return ret;
-
 	/*
 	 * Here we commit the transaction to make sure commit_total_bytes
 	 * of all the devices are updated.
@@ -431,8 +617,17 @@ static int btrfs_dev_replace_start(struct btrfs_fs_info *fs_info,
 		return PTR_ERR(trans);
 	}
 
-	need_unlock = true;
+	ret = btrfs_init_dev_replace_tgtdev(fs_info, tgtdev_name,
+					    src_device, &tgt_device);
+	if (ret)
+		return ret;
+
+	ret = mark_block_group_to_copy(fs_info, src_device);
+	if (ret)
+		return ret;
+
 	down_write(&dev_replace->rwsem);
+	dev_replace->replace_task = current;
 	switch (dev_replace->replace_state) {
 	case BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED:
 	case BTRFS_IOCTL_DEV_REPLACE_STATE_FINISHED:
@@ -440,21 +635,21 @@ static int btrfs_dev_replace_start(struct btrfs_fs_info *fs_info,
 		break;
 	case BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED:
 	case BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED:
-		ASSERT(0);
+		DEBUG_WARN("unexpected STARTED or SUSPENDED dev-replace state");
 		ret = BTRFS_IOCTL_DEV_REPLACE_RESULT_ALREADY_STARTED;
+		up_write(&dev_replace->rwsem);
 		goto leave;
 	}
 
 	dev_replace->cont_reading_from_srcdev_mode = read_src;
-	WARN_ON(!src_device);
 	dev_replace->srcdev = src_device;
 	dev_replace->tgtdev = tgt_device;
 
-	btrfs_info_in_rcu(fs_info,
+	btrfs_info(fs_info,
 		      "dev_replace from %s (devid %llu) to %s started",
 		      btrfs_dev_name(src_device),
 		      src_device->devid,
-		      rcu_str_deref(tgt_device->name));
+		      btrfs_dev_name(tgt_device));
 
 	/*
 	 * from now on, the writes to the srcdev are all duplicated to
@@ -471,24 +666,28 @@ static int btrfs_dev_replace_start(struct btrfs_fs_info *fs_info,
 	atomic64_set(&dev_replace->num_write_errors, 0);
 	atomic64_set(&dev_replace->num_uncorrectable_read_errors, 0);
 	up_write(&dev_replace->rwsem);
-	need_unlock = false;
 
-	ret = btrfs_sysfs_add_device_link(tgt_device->fs_devices, tgt_device);
+	ret = btrfs_sysfs_add_device(tgt_device);
 	if (ret)
 		btrfs_err(fs_info, "kobj add dev failed %d", ret);
 
-	btrfs_wait_ordered_roots(fs_info, U64_MAX, 0, (u64)-1);
+	btrfs_wait_ordered_roots(fs_info, U64_MAX, NULL);
 
-	/* force writing the updated state information to disk */
-	trans = btrfs_start_transaction(root, 0);
+	/*
+	 * Commit dev_replace state and reserve 1 item for it.
+	 * This is crucial to ensure we won't miss copying extents for new block
+	 * groups that are allocated after we started the device replace, and
+	 * must be done after setting up the device replace state.
+	 */
+	trans = btrfs_start_transaction(root, 1);
 	if (IS_ERR(trans)) {
 		ret = PTR_ERR(trans);
-		need_unlock = true;
 		down_write(&dev_replace->rwsem);
 		dev_replace->replace_state =
 			BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED;
 		dev_replace->srcdev = NULL;
 		dev_replace->tgtdev = NULL;
+		up_write(&dev_replace->rwsem);
 		goto leave;
 	}
 
@@ -501,21 +700,33 @@ static int btrfs_dev_replace_start(struct btrfs_fs_info *fs_info,
 			      &dev_replace->scrub_progress, 0, 1);
 
 	ret = btrfs_dev_replace_finishing(fs_info, ret);
-	if (ret == -EINPROGRESS) {
+	if (ret == -EINPROGRESS)
 		ret = BTRFS_IOCTL_DEV_REPLACE_RESULT_SCRUB_INPROGRESS;
-	} else if (ret != -ECANCELED) {
-		WARN_ON(ret);
-	}
 
 	return ret;
 
 leave:
-	if (need_unlock)
-		up_write(&dev_replace->rwsem);
 	btrfs_destroy_dev_replace_tgtdev(tgt_device);
 	return ret;
 }
 
+static int btrfs_check_replace_dev_names(struct btrfs_ioctl_dev_replace_args *args)
+{
+	if (args->start.srcdevid == 0) {
+		if (memchr(args->start.srcdev_name, 0,
+			   sizeof(args->start.srcdev_name)) == NULL)
+			return -ENAMETOOLONG;
+	} else {
+		args->start.srcdev_name[0] = 0;
+	}
+
+	if (memchr(args->start.tgtdev_name, 0,
+		   sizeof(args->start.tgtdev_name)) == NULL)
+	    return -ENAMETOOLONG;
+
+	return 0;
+}
+
 int btrfs_dev_replace_by_ioctl(struct btrfs_fs_info *fs_info,
 			    struct btrfs_ioctl_dev_replace_args *args)
 {
@@ -528,10 +739,9 @@ int btrfs_dev_replace_by_ioctl(struct btrfs_fs_info *fs_info,
 	default:
 		return -EINVAL;
 	}
-
-	if ((args->start.srcdevid == 0 && args->start.srcdev_name[0] == '\0') ||
-	    args->start.tgtdev_name[0] == '\0')
-		return -EINVAL;
+	ret = btrfs_check_replace_dev_names(args);
+	if (ret < 0)
+		return ret;
 
 	ret = btrfs_dev_replace_start(fs_info, args->start.tgtdev_name,
 					args->start.srcdevid,
@@ -565,10 +775,89 @@ static void btrfs_rm_dev_replace_unblocked(struct btrfs_fs_info *fs_info)
 	wake_up(&fs_info->dev_replace.replace_wait);
 }
 
+/*
+ * When finishing the device replace, before swapping the source device with the
+ * target device we must update the chunk allocation state in the target device,
+ * as it is empty because replace works by directly copying the chunks and not
+ * through the normal chunk allocation path.
+ */
+static int btrfs_set_target_alloc_state(struct btrfs_device *srcdev,
+					struct btrfs_device *tgtdev)
+{
+	struct extent_state *cached_state = NULL;
+	u64 start = 0;
+	u64 found_start;
+	u64 found_end;
+	int ret = 0;
+
+	lockdep_assert_held(&srcdev->fs_info->chunk_mutex);
+
+	while (btrfs_find_first_extent_bit(&srcdev->alloc_state, start,
+					   &found_start, &found_end,
+					   CHUNK_ALLOCATED, &cached_state)) {
+		ret = btrfs_set_extent_bit(&tgtdev->alloc_state, found_start,
+					   found_end, CHUNK_ALLOCATED, NULL);
+		if (ret)
+			break;
+		start = found_end + 1;
+	}
+
+	btrfs_free_extent_state(cached_state);
+	return ret;
+}
+
+static void btrfs_dev_replace_update_device_in_mapping_tree(
+						struct btrfs_fs_info *fs_info,
+						struct btrfs_device *srcdev,
+						struct btrfs_device *tgtdev)
+{
+	struct rb_node *node;
+
+	/*
+	 * The chunk mutex must be held so that no new chunks can be created
+	 * while we are updating existing chunks. This guarantees we don't miss
+	 * any new chunk that gets created for a range that falls before the
+	 * range of the last chunk we processed.
+	 */
+	lockdep_assert_held(&fs_info->chunk_mutex);
+
+	write_lock(&fs_info->mapping_tree_lock);
+	node = rb_first_cached(&fs_info->mapping_tree);
+	while (node) {
+		struct rb_node *next = rb_next(node);
+		struct btrfs_chunk_map *map;
+		u64 next_start;
+
+		map = rb_entry(node, struct btrfs_chunk_map, rb_node);
+		next_start = map->start + map->chunk_len;
+
+		for (int i = 0; i < map->num_stripes; i++)
+			if (srcdev == map->stripes[i].dev)
+				map->stripes[i].dev = tgtdev;
+
+		if (cond_resched_rwlock_write(&fs_info->mapping_tree_lock)) {
+			map = btrfs_find_chunk_map_nolock(fs_info, next_start, U64_MAX);
+			if (!map)
+				break;
+			node = &map->rb_node;
+			/*
+			 * Drop the lookup reference since we are holding the
+			 * lock in write mode and no one can remove the chunk
+			 * map from the tree and drop its tree reference.
+			 */
+			btrfs_free_chunk_map(map);
+		} else {
+			node = next;
+		}
+	}
+	write_unlock(&fs_info->mapping_tree_lock);
+}
+
 static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
 				       int scrub_ret)
 {
 	struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
+	struct btrfs_fs_devices *fs_devices = fs_info->fs_devices;
 	struct btrfs_device *tgt_device;
 	struct btrfs_device *src_device;
 	struct btrfs_root *root = fs_info->tree_root;
@@ -596,24 +885,40 @@ 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
 	 */
-	ret = btrfs_start_delalloc_roots(fs_info, -1);
+	ret = btrfs_start_delalloc_roots(fs_info, LONG_MAX, false);
 	if (ret) {
 		mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
 		return ret;
 	}
-	btrfs_wait_ordered_roots(fs_info, U64_MAX, 0, (u64)-1);
+	btrfs_wait_ordered_roots(fs_info, U64_MAX, NULL);
 
-	trans = btrfs_start_transaction(root, 0);
-	if (IS_ERR(trans)) {
-		mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
-		return PTR_ERR(trans);
+	/*
+	 * We have to use this loop approach because at this point src_device
+	 * has to be available for transaction commit to complete, yet new
+	 * chunks shouldn't be allocated on the device.
+	 */
+	while (1) {
+		trans = btrfs_start_transaction(root, 0);
+		if (IS_ERR(trans)) {
+			mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
+			return PTR_ERR(trans);
+		}
+		ret = btrfs_commit_transaction(trans);
+		WARN_ON(ret);
+
+		/* Prevent write_all_supers() during the finishing procedure */
+		mutex_lock(&fs_devices->device_list_mutex);
+		/* Prevent new chunks being allocated on the source device */
+		mutex_lock(&fs_info->chunk_mutex);
+
+		if (!list_empty(&src_device->post_commit_list)) {
+			mutex_unlock(&fs_devices->device_list_mutex);
+			mutex_unlock(&fs_info->chunk_mutex);
+		} else {
+			break;
+		}
 	}
-	ret = btrfs_commit_transaction(trans);
-	WARN_ON(ret);
 
-	/* keep away write_all_supers() during the finishing procedure */
-	mutex_lock(&fs_info->fs_devices->device_list_mutex);
-	mutex_lock(&fs_info->chunk_mutex);
 	down_write(&dev_replace->rwsem);
 	dev_replace->replace_state =
 		scrub_ret ? BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED
@@ -623,21 +928,28 @@ static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
 	dev_replace->time_stopped = ktime_get_real_seconds();
 	dev_replace->item_needs_writeback = 1;
 
-	/* replace old device with new one in mapping tree */
+	/*
+	 * Update allocation state in the new device and replace the old device
+	 * with the new one in the mapping tree.
+	 */
 	if (!scrub_ret) {
+		scrub_ret = btrfs_set_target_alloc_state(src_device, tgt_device);
+		if (scrub_ret)
+			goto error;
 		btrfs_dev_replace_update_device_in_mapping_tree(fs_info,
 								src_device,
 								tgt_device);
 	} else {
 		if (scrub_ret != -ECANCELED)
-			btrfs_err_in_rcu(fs_info,
+			btrfs_err(fs_info,
 				 "btrfs_scrub_dev(%s, %llu, %s) failed %d",
 				 btrfs_dev_name(src_device),
 				 src_device->devid,
-				 rcu_str_deref(tgt_device->name), scrub_ret);
+				 btrfs_dev_name(tgt_device), scrub_ret);
+error:
 		up_write(&dev_replace->rwsem);
 		mutex_unlock(&fs_info->chunk_mutex);
-		mutex_unlock(&fs_info->fs_devices->device_list_mutex);
+		mutex_unlock(&fs_devices->device_list_mutex);
 		btrfs_rm_dev_replace_blocked(fs_info);
 		if (tgt_device)
 			btrfs_destroy_dev_replace_tgtdev(tgt_device);
@@ -647,11 +959,11 @@ static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
 		return scrub_ret;
 	}
 
-	btrfs_info_in_rcu(fs_info,
+	btrfs_info(fs_info,
 			  "dev_replace from %s (devid %llu) to %s finished",
 			  btrfs_dev_name(src_device),
 			  src_device->devid,
-			  rcu_str_deref(tgt_device->name));
+			  btrfs_dev_name(tgt_device));
 	clear_bit(BTRFS_DEV_STATE_REPLACE_TGT, &tgt_device->dev_state);
 	tgt_device->devid = src_device->devid;
 	src_device->devid = BTRFS_DEV_REPLACE_DEVID;
@@ -662,15 +974,14 @@ static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
 	btrfs_device_set_disk_total_bytes(tgt_device,
 					  src_device->disk_total_bytes);
 	btrfs_device_set_bytes_used(tgt_device, src_device->bytes_used);
-	ASSERT(list_empty(&src_device->resized_list));
-	tgt_device->commit_total_bytes = src_device->commit_total_bytes;
 	tgt_device->commit_bytes_used = src_device->bytes_used;
 
 	btrfs_assign_next_active_device(src_device, tgt_device);
 
-	list_add(&tgt_device->dev_alloc_list, &fs_info->fs_devices->alloc_list);
-	fs_info->fs_devices->rw_devices++;
+	list_add(&tgt_device->dev_alloc_list, &fs_devices->alloc_list);
+	fs_devices->rw_devices++;
 
+	dev_replace->replace_task = NULL;
 	up_write(&dev_replace->rwsem);
 	btrfs_rm_dev_replace_blocked(fs_info);
 
@@ -692,11 +1003,13 @@ static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
 	 * belong to this filesystem.
 	 */
 	mutex_unlock(&fs_info->chunk_mutex);
-	mutex_unlock(&fs_info->fs_devices->device_list_mutex);
+	mutex_unlock(&fs_devices->device_list_mutex);
 
 	/* replace the sysfs entry */
-	btrfs_sysfs_rm_device_link(fs_info->fs_devices, src_device);
-	btrfs_rm_dev_replace_free_srcdev(fs_info, src_device);
+	btrfs_sysfs_remove_device(src_device);
+	btrfs_sysfs_update_devid(tgt_device);
+	if (test_bit(BTRFS_DEV_STATE_WRITEABLE, &src_device->dev_state))
+		btrfs_scratch_superblocks(fs_info, src_device);
 
 	/* write back the superblocks */
 	trans = btrfs_start_transaction(root, 0);
@@ -705,33 +1018,9 @@ static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
 
 	mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
 
-	return 0;
-}
+	btrfs_rm_dev_replace_free_srcdev(src_device);
 
-static void btrfs_dev_replace_update_device_in_mapping_tree(
-						struct btrfs_fs_info *fs_info,
-						struct btrfs_device *srcdev,
-						struct btrfs_device *tgtdev)
-{
-	struct extent_map_tree *em_tree = &fs_info->mapping_tree.map_tree;
-	struct extent_map *em;
-	struct map_lookup *map;
-	u64 start = 0;
-	int i;
-
-	write_lock(&em_tree->lock);
-	do {
-		em = lookup_extent_mapping(em_tree, start, (u64)-1);
-		if (!em)
-			break;
-		map = em->map_lookup;
-		for (i = 0; i < map->num_stripes; i++)
-			if (srcdev == map->stripes[i].dev)
-				map->stripes[i].dev = tgtdev;
-		start = em->start + em->len;
-		free_extent_map(em);
-	} while (start);
-	write_unlock(&em_tree->lock);
+	return 0;
 }
 
 /*
@@ -818,7 +1107,7 @@ int btrfs_dev_replace_cancel(struct btrfs_fs_info *fs_info)
 			 * btrfs_dev_replace_finishing() will handle the
 			 * cleanup part
 			 */
-			btrfs_info_in_rcu(fs_info,
+			btrfs_info(fs_info,
 				"dev_replace from %s (devid %llu) to %s canceled",
 				btrfs_dev_name(src_device), src_device->devid,
 				btrfs_dev_name(tgt_device));
@@ -842,8 +1131,7 @@ int btrfs_dev_replace_cancel(struct btrfs_fs_info *fs_info)
 		up_write(&dev_replace->rwsem);
 
 		/* Scrub for replace must not be running in suspended state */
-		ret = btrfs_scrub_cancel(fs_info);
-		ASSERT(ret != -ENOTCONN);
+		btrfs_scrub_cancel(fs_info);
 
 		trans = btrfs_start_transaction(root, 0);
 		if (IS_ERR(trans)) {
@@ -853,7 +1141,7 @@ int btrfs_dev_replace_cancel(struct btrfs_fs_info *fs_info)
 		ret = btrfs_commit_transaction(trans);
 		WARN_ON(ret);
 
-		btrfs_info_in_rcu(fs_info,
+		btrfs_info(fs_info,
 		"suspended dev_replace from %s (devid %llu) to %s canceled",
 			btrfs_dev_name(src_device), src_device->devid,
 			btrfs_dev_name(tgt_device));
@@ -862,6 +1150,7 @@ int btrfs_dev_replace_cancel(struct btrfs_fs_info *fs_info)
 			btrfs_destroy_dev_replace_tgtdev(tgt_device);
 		break;
 	default:
+		up_write(&dev_replace->rwsem);
 		result = -EINVAL;
 	}
 
@@ -933,7 +1222,7 @@ int btrfs_resume_dev_replace_async(struct btrfs_fs_info *fs_info)
 	 * should never allow both to start and pause. We don't want to allow
 	 * dev-replace to start anyway.
 	 */
-	if (test_and_set_bit(BTRFS_FS_EXCL_OP, &fs_info->flags)) {
+	if (!btrfs_exclop_start(fs_info, BTRFS_EXCLOP_DEV_REPLACE)) {
 		down_write(&dev_replace->rwsem);
 		dev_replace->replace_state =
 					BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED;
@@ -956,7 +1245,7 @@ static int btrfs_dev_replace_kthread(void *data)
 
 	progress = btrfs_dev_replace_progress(fs_info);
 	progress = div_u64(progress, 10);
-	btrfs_info_in_rcu(fs_info,
+	btrfs_info(fs_info,
 		"continuing dev_replace from %s (devid %llu) to target %s @%u%%",
 		btrfs_dev_name(dev_replace->srcdev),
 		dev_replace->srcdev->devid,
@@ -970,20 +1259,20 @@ static int btrfs_dev_replace_kthread(void *data)
 	ret = btrfs_dev_replace_finishing(fs_info, ret);
 	WARN_ON(ret && ret != -ECANCELED);
 
-	clear_bit(BTRFS_FS_EXCL_OP, &fs_info->flags);
+	btrfs_exclop_finish(fs_info);
 	return 0;
 }
 
-int btrfs_dev_replace_is_ongoing(struct btrfs_dev_replace *dev_replace)
+bool __pure btrfs_dev_replace_is_ongoing(struct btrfs_dev_replace *dev_replace)
 {
 	if (!dev_replace->is_valid)
-		return 0;
+		return false;
 
 	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:
-		return 0;
+		return false;
 	case BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED:
 	case BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED:
 		/*
@@ -998,12 +1287,7 @@ int btrfs_dev_replace_is_ongoing(struct btrfs_dev_replace *dev_replace)
 		 */
 		break;
 	}
-	return 1;
-}
-
-void btrfs_bio_counter_inc_noblocked(struct btrfs_fs_info *fs_info)
-{
-	percpu_counter_inc(&fs_info->dev_replace.bio_counter);
+	return true;
 }
 
 void btrfs_bio_counter_sub(struct btrfs_fs_info *fs_info, s64 amount)