Btrfs: disallow some operations on the device replace target device

This patch adds some code to disallow operations on the device that
is used as the target for the device replace operation.

Signed-off-by: Stefan Behrens <sbehrens@giantdisaster.de>
Signed-off-by: Chris Mason <chris.mason@fusionio.com>

7 files changed, 54 insertions, 18 deletions

diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index e9dc78014f09..746cb6aa1f62 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -3649,7 +3649,7 @@ int btrfs_reloc_post_snapshot(struct btrfs_trans_handle *trans,
 /* scrub.c */
 int btrfs_scrub_dev(struct btrfs_fs_info *fs_info, u64 devid, u64 start,
 		    u64 end, struct btrfs_scrub_progress *progress,
-		    int readonly);
+		    int readonly, int is_dev_replace);
 void btrfs_scrub_pause(struct btrfs_root *root);
 void btrfs_scrub_pause_super(struct btrfs_root *root);
 void btrfs_scrub_continue(struct btrfs_root *root);
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index b4d438f6c2b3..98af8379895a 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -7468,7 +7468,8 @@ int btrfs_can_relocate(struct btrfs_root *root, u64 bytenr)
 		 * check to make sure we can actually find a chunk with enough
 		 * space to fit our block group in.
 		 */
-		if (device->total_bytes > device->bytes_used + min_free) {
+		if (device->total_bytes > device->bytes_used + min_free &&
+		    !device->is_tgtdev_for_dev_replace) {
 			ret = find_free_dev_extent(device, min_free,
 						   &dev_offset, NULL);
 			if (!ret)
diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index 26f46dad3b0e..e54b5e50c927 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -1375,6 +1375,11 @@ static noinline int btrfs_ioctl_resize(struct btrfs_root *root,
 		}
 	}
 
+	if (device->is_tgtdev_for_dev_replace) {
+		ret = -EINVAL;
+		goto out_free;
+	}
+
 	old_size = device->total_bytes;
 
 	if (mod < 0) {
@@ -3102,7 +3107,8 @@ static long btrfs_ioctl_scrub(struct btrfs_root *root, void __user *arg)
 		return PTR_ERR(sa);
 
 	ret = btrfs_scrub_dev(root->fs_info, sa->devid, sa->start, sa->end,
-			      &sa->progress, sa->flags & BTRFS_SCRUB_READONLY);
+			      &sa->progress, sa->flags & BTRFS_SCRUB_READONLY,
+			      0);
 
 	if (copy_to_user(arg, sa, sizeof(*sa)))
 		ret = -EFAULT;
diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c
index 6cf23f4f7bb7..460e30bb1884 100644
--- a/fs/btrfs/scrub.c
+++ b/fs/btrfs/scrub.c
@@ -116,6 +116,9 @@ struct scrub_ctx {
 	u32			sectorsize;
 	u32			nodesize;
 	u32			leafsize;
+
+	int			is_dev_replace;
+
 	/*
 	 * statistics
 	 */
@@ -284,7 +287,7 @@ static noinline_for_stack void scrub_free_ctx(struct scrub_ctx *sctx)
 }
 
 static noinline_for_stack
-struct scrub_ctx *scrub_setup_ctx(struct btrfs_device *dev)
+struct scrub_ctx *scrub_setup_ctx(struct btrfs_device *dev, int is_dev_replace)
 {
 	struct scrub_ctx *sctx;
 	int		i;
@@ -296,6 +299,7 @@ struct scrub_ctx *scrub_setup_ctx(struct btrfs_device *dev)
 	sctx = kzalloc(sizeof(*sctx), GFP_NOFS);
 	if (!sctx)
 		goto nomem;
+	sctx->is_dev_replace = is_dev_replace;
 	sctx->pages_per_bio = pages_per_bio;
 	sctx->curr = -1;
 	sctx->dev_root = dev->dev_root;
@@ -2293,7 +2297,7 @@ static noinline_for_stack void scrub_workers_put(struct btrfs_fs_info *fs_info)
 
 int btrfs_scrub_dev(struct btrfs_fs_info *fs_info, u64 devid, u64 start,
 		    u64 end, struct btrfs_scrub_progress *progress,
-		    int readonly)
+		    int readonly, int is_dev_replace)
 {
 	struct scrub_ctx *sctx;
 	int ret;
@@ -2356,14 +2360,14 @@ int btrfs_scrub_dev(struct btrfs_fs_info *fs_info, u64 devid, u64 start,
 
 	mutex_lock(&fs_info->fs_devices->device_list_mutex);
 	dev = btrfs_find_device(fs_info, devid, NULL, NULL);
-	if (!dev || dev->missing) {
+	if (!dev || (dev->missing && !is_dev_replace)) {
 		mutex_unlock(&fs_info->fs_devices->device_list_mutex);
 		scrub_workers_put(fs_info);
 		return -ENODEV;
 	}
 	mutex_lock(&fs_info->scrub_lock);
 
-	if (!dev->in_fs_metadata) {
+	if (!dev->in_fs_metadata || dev->is_tgtdev_for_dev_replace) {
 		mutex_unlock(&fs_info->scrub_lock);
 		mutex_unlock(&fs_info->fs_devices->device_list_mutex);
 		scrub_workers_put(fs_info);
@@ -2376,7 +2380,7 @@ int btrfs_scrub_dev(struct btrfs_fs_info *fs_info, u64 devid, u64 start,
 		scrub_workers_put(fs_info);
 		return -EINPROGRESS;
 	}
-	sctx = scrub_setup_ctx(dev);
+	sctx = scrub_setup_ctx(dev, is_dev_replace);
 	if (IS_ERR(sctx)) {
 		mutex_unlock(&fs_info->scrub_lock);
 		mutex_unlock(&fs_info->fs_devices->device_list_mutex);
diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c
index ef2415896b06..837ad2d27853 100644
--- a/fs/btrfs/super.c
+++ b/fs/btrfs/super.c
@@ -1354,7 +1354,8 @@ static int btrfs_calc_avail_data_space(struct btrfs_root *root, u64 *free_bytes)
 		min_stripe_size = BTRFS_STRIPE_LEN;
 
 	list_for_each_entry(device, &fs_devices->devices, dev_list) {
-		if (!device->in_fs_metadata || !device->bdev)
+		if (!device->in_fs_metadata || !device->bdev ||
+		    device->is_tgtdev_for_dev_replace)
 			continue;
 
 		avail_space = device->total_bytes - device->bytes_used;
diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index 33ca36b37a6a..31f7af878d96 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -518,8 +518,9 @@ again:
 	/* This is the initialized path, it is safe to release the devices. */
 	list_for_each_entry_safe(device, next, &fs_devices->devices, dev_list) {
 		if (device->in_fs_metadata) {
-			if (!latest_transid ||
-			    device->generation > latest_transid) {
+			if (!device->is_tgtdev_for_dev_replace &&
+			    (!latest_transid ||
+			     device->generation > latest_transid)) {
 				latest_devid = device->devid;
 				latest_transid = device->generation;
 				latest_bdev = device->bdev;
@@ -814,7 +815,7 @@ int btrfs_account_dev_extents_size(struct btrfs_device *device, u64 start,
 
 	*length = 0;
 
-	if (start >= device->total_bytes)
+	if (start >= device->total_bytes || device->is_tgtdev_for_dev_replace)
 		return 0;
 
 	path = btrfs_alloc_path();
@@ -931,7 +932,7 @@ int find_free_dev_extent(struct btrfs_device *device, u64 num_bytes,
 	max_hole_size = 0;
 	hole_size = 0;
 
-	if (search_start >= search_end) {
+	if (search_start >= search_end || device->is_tgtdev_for_dev_replace) {
 		ret = -ENOSPC;
 		goto error;
 	}
@@ -1114,6 +1115,7 @@ int btrfs_alloc_dev_extent(struct btrfs_trans_handle *trans,
 	struct btrfs_key key;
 
 	WARN_ON(!device->in_fs_metadata);
+	WARN_ON(device->is_tgtdev_for_dev_replace);
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
@@ -1375,7 +1377,9 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path)
 		 * is held.
 		 */
 		list_for_each_entry(tmp, devices, dev_list) {
-			if (tmp->in_fs_metadata && !tmp->bdev) {
+			if (tmp->in_fs_metadata &&
+			    !tmp->is_tgtdev_for_dev_replace &&
+			    !tmp->bdev) {
 				device = tmp;
 				break;
 			}
@@ -1406,6 +1410,12 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path)
 		}
 	}
 
+	if (device->is_tgtdev_for_dev_replace) {
+		pr_err("btrfs: unable to remove the dev_replace target dev\n");
+		ret = -EINVAL;
+		goto error_brelse;
+	}
+
 	if (device->writeable && root->fs_info->fs_devices->rw_devices == 1) {
 		printk(KERN_ERR "btrfs: unable to remove the only writeable "
 		       "device\n");
@@ -1425,6 +1435,11 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path)
 	if (ret)
 		goto error_undo;
 
+	/*
+	 * TODO: the superblock still includes this device in its num_devices
+	 * counter although write_all_supers() is not locked out. This
+	 * could give a filesystem state which requires a degraded mount.
+	 */
 	ret = btrfs_rm_dev_item(root->fs_info->chunk_root, device);
 	if (ret)
 		goto error_undo;
@@ -1808,6 +1823,7 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path)
 	device->dev_root = root->fs_info->dev_root;
 	device->bdev = bdev;
 	device->in_fs_metadata = 1;
+	device->is_tgtdev_for_dev_replace = 0;
 	device->mode = FMODE_EXCL;
 	set_blocksize(device->bdev, 4096);
 
@@ -1971,7 +1987,8 @@ static int __btrfs_grow_device(struct btrfs_trans_handle *trans,
 
 	if (!device->writeable)
 		return -EACCES;
-	if (new_size <= device->total_bytes)
+	if (new_size <= device->total_bytes ||
+	    device->is_tgtdev_for_dev_replace)
 		return -EINVAL;
 
 	btrfs_set_super_total_bytes(super_copy, old_total + diff);
@@ -2600,7 +2617,8 @@ static int __btrfs_balance(struct btrfs_fs_info *fs_info)
 		size_to_free = div_factor(old_size, 1);
 		size_to_free = min(size_to_free, (u64)1 * 1024 * 1024);
 		if (!device->writeable ||
-		    device->total_bytes - device->bytes_used > size_to_free)
+		    device->total_bytes - device->bytes_used > size_to_free ||
+		    device->is_tgtdev_for_dev_replace)
 			continue;
 
 		ret = btrfs_shrink_device(device, old_size - size_to_free);
@@ -3132,6 +3150,9 @@ int btrfs_shrink_device(struct btrfs_device *device, u64 new_size)
 	u64 old_size = device->total_bytes;
 	u64 diff = device->total_bytes - new_size;
 
+	if (device->is_tgtdev_for_dev_replace)
+		return -EINVAL;
+
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
@@ -3401,7 +3422,8 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans,
 			continue;
 		}
 
-		if (!device->in_fs_metadata)
+		if (!device->in_fs_metadata ||
+		    device->is_tgtdev_for_dev_replace)
 			continue;
 
 		if (device->total_bytes > device->bytes_used)
@@ -4612,6 +4634,7 @@ static void fill_device_from_item(struct extent_buffer *leaf,
 	device->io_align = btrfs_device_io_align(leaf, dev_item);
 	device->io_width = btrfs_device_io_width(leaf, dev_item);
 	device->sector_size = btrfs_device_sector_size(leaf, dev_item);
+	device->is_tgtdev_for_dev_replace = 0;
 
 	ptr = (unsigned long)btrfs_device_uuid(dev_item);
 	read_extent_buffer(leaf, device->uuid, ptr, BTRFS_UUID_SIZE);
@@ -4722,7 +4745,7 @@ static int read_one_dev(struct btrfs_root *root,
 	fill_device_from_item(leaf, dev_item, device);
 	device->dev_root = root->fs_info->dev_root;
 	device->in_fs_metadata = 1;
-	if (device->writeable) {
+	if (device->writeable && !device->is_tgtdev_for_dev_replace) {
 		device->fs_devices->total_rw_bytes += device->total_bytes;
 		spin_lock(&root->fs_info->free_chunk_lock);
 		root->fs_info->free_chunk_space += device->total_bytes -
diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h
index 802e2ba02f09..8fd5a4d8acc8 100644
--- a/fs/btrfs/volumes.h
+++ b/fs/btrfs/volumes.h
@@ -50,6 +50,7 @@ struct btrfs_device {
 	int in_fs_metadata;
 	int missing;
 	int can_discard;
+	int is_tgtdev_for_dev_replace;
 
 	spinlock_t io_lock;