Merge tag 'for-5.9-rc7-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux

Pull btrfs fixes from David Sterba:
 "Two more fixes.

  One is for a lockdep warning/lockup (also caught by syzbot), that one
  has been seen in practice. Regarding the other syzbot reports
  mentioned last time, they don't seem to be urgent and reliably
  reproducible so they'll be fixed later.

  The second fix is for a potential corruption when device replace
  finishes and the in-memory state of trim is not copied to the new
  device"

* tag 'for-5.9-rc7-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:
  btrfs: fix filesystem corruption after a device replace
  btrfs: move btrfs_rm_dev_replace_free_srcdev outside of all locks
  btrfs: move btrfs_scratch_superblocks into btrfs_dev_replace_finishing

3 files changed, 52 insertions, 10 deletions

diff --git a/fs/btrfs/dev-replace.c b/fs/btrfs/dev-replace.c
index db93909b25e08..e4a1c6afe35dc 100644
--- a/fs/btrfs/dev-replace.c
+++ b/fs/btrfs/dev-replace.c
@@ -599,6 +599,37 @@ 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 (!find_first_extent_bit(&srcdev->alloc_state, start,
+				      &found_start, &found_end,
+				      CHUNK_ALLOCATED, &cached_state)) {
+		ret = set_extent_bits(&tgtdev->alloc_state, found_start,
+				      found_end, CHUNK_ALLOCATED);
+		if (ret)
+			break;
+		start = found_end + 1;
+	}
+
+	free_extent_state(cached_state);
+	return ret;
+}
+
 static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
 				       int scrub_ret)
 {
@@ -673,8 +704,14 @@ 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);
@@ -685,6 +722,7 @@ static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
 				 btrfs_dev_name(src_device),
 				 src_device->devid,
 				 rcu_str_deref(tgt_device->name), scrub_ret);
+error:
 		up_write(&dev_replace->rwsem);
 		mutex_unlock(&fs_info->chunk_mutex);
 		mutex_unlock(&fs_info->fs_devices->device_list_mutex);
@@ -745,7 +783,9 @@ static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
 	/* replace the sysfs entry */
 	btrfs_sysfs_remove_devices_dir(fs_info->fs_devices, src_device);
 	btrfs_sysfs_update_devid(tgt_device);
-	btrfs_rm_dev_replace_free_srcdev(src_device);
+	if (test_bit(BTRFS_DEV_STATE_WRITEABLE, &src_device->dev_state))
+		btrfs_scratch_superblocks(fs_info, src_device->bdev,
+					  src_device->name->str);
 
 	/* write back the superblocks */
 	trans = btrfs_start_transaction(root, 0);
@@ -754,6 +794,8 @@ static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
 
 	mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
 
+	btrfs_rm_dev_replace_free_srcdev(src_device);
+
 	return 0;
 }
 
diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index 117b43367629e..1997a7d67f22f 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -1999,9 +1999,9 @@ static u64 btrfs_num_devices(struct btrfs_fs_info *fs_info)
 	return num_devices;
 }
 
-static void btrfs_scratch_superblocks(struct btrfs_fs_info *fs_info,
-				      struct block_device *bdev,
-				      const char *device_path)
+void btrfs_scratch_superblocks(struct btrfs_fs_info *fs_info,
+			       struct block_device *bdev,
+			       const char *device_path)
 {
 	struct btrfs_super_block *disk_super;
 	int copy_num;
@@ -2224,11 +2224,7 @@ void btrfs_rm_dev_replace_free_srcdev(struct btrfs_device *srcdev)
 	struct btrfs_fs_info *fs_info = srcdev->fs_info;
 	struct btrfs_fs_devices *fs_devices = srcdev->fs_devices;
 
-	if (test_bit(BTRFS_DEV_STATE_WRITEABLE, &srcdev->dev_state)) {
-		/* zero out the old super if it is writable */
-		btrfs_scratch_superblocks(fs_info, srcdev->bdev,
-					  srcdev->name->str);
-	}
+	mutex_lock(&uuid_mutex);
 
 	btrfs_close_bdev(srcdev);
 	synchronize_rcu();
@@ -2258,6 +2254,7 @@ void btrfs_rm_dev_replace_free_srcdev(struct btrfs_device *srcdev)
 		close_fs_devices(fs_devices);
 		free_fs_devices(fs_devices);
 	}
+	mutex_unlock(&uuid_mutex);
 }
 
 void btrfs_destroy_dev_replace_tgtdev(struct btrfs_device *tgtdev)
diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h
index 5eea93916fbf1..302c9234f7d08 100644
--- a/fs/btrfs/volumes.h
+++ b/fs/btrfs/volumes.h
@@ -573,6 +573,9 @@ void btrfs_set_fs_info_ptr(struct btrfs_fs_info *fs_info);
 void btrfs_reset_fs_info_ptr(struct btrfs_fs_info *fs_info);
 bool btrfs_check_rw_degradable(struct btrfs_fs_info *fs_info,
 					struct btrfs_device *failing_dev);
+void btrfs_scratch_superblocks(struct btrfs_fs_info *fs_info,
+			       struct block_device *bdev,
+			       const char *device_path);
 
 int btrfs_bg_type_to_factor(u64 flags);
 const char *btrfs_bg_type_to_raid_name(u64 flags);