btrfs: zoned: fix use-after-free due to race with dev replace

While loading a zone's info during creation of a block group, we can race
with a device replace operation and then trigger a use-after-free on the
device that was just replaced (source device of the replace operation).

This happens because at btrfs_load_zone_info() we extract a device from
the chunk map into a local variable and then use the device while not
under the protection of the device replace rwsem. So if there's a device
replace operation happening when we extract the device and that device
is the source of the replace operation, we will trigger a use-after-free
if before we finish using the device the replace operation finishes and
frees the device.

Fix this by enlarging the critical section under the protection of the
device replace rwsem so that all uses of the device are done inside the
critical section.

CC: stable@vger.kernel.org # 6.1.x: 15c12fc: btrfs: zoned: introduce a zone_info struct in btrfs_load_block_group_zone_info
CC: stable@vger.kernel.org # 6.1.x: 09a4672: btrfs: zoned: factor out per-zone logic from btrfs_load_block_group_zone_info
CC: stable@vger.kernel.org # 6.1.x: 9e0e3e7: btrfs: zoned: factor out single bg handling from btrfs_load_block_group_zone_info
CC: stable@vger.kernel.org # 6.1.x: 87463f7: btrfs: zoned: factor out DUP bg handling from btrfs_load_block_group_zone_info
CC: stable@vger.kernel.org # 6.1.x
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>

Author: fdmanana

fs/btrfs/zoned.c

@@ -1290,15 +1290,19 @@ static int btrfs_load_zone_info(struct btrfs_fs_info *fs_info, int zone_idx,
 				struct btrfs_chunk_map *map)
 {
 	struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
-	struct btrfs_device *device = map->stripes[zone_idx].dev;
+	struct btrfs_device *device;
 	int dev_replace_is_ongoing = 0;
 	unsigned int nofs_flag;
 	struct blk_zone zone;
 	int ret;
 
 	info->physical = map->stripes[zone_idx].physical;
 
+	down_read(&dev_replace->rwsem);
+	device = map->stripes[zone_idx].dev;
+
 	if (!device->bdev) {
+		up_read(&dev_replace->rwsem);
 		info->alloc_offset = WP_MISSING_DEV;
 		return 0;
 	}
@@ -1308,18 +1312,17 @@ static int btrfs_load_zone_info(struct btrfs_fs_info *fs_info, int zone_idx,
 		__set_bit(zone_idx, active);
 
 	if (!btrfs_dev_is_sequential(device, info->physical)) {
+		up_read(&dev_replace->rwsem);
 		info->alloc_offset = WP_CONVENTIONAL;
 		return 0;
 	}
 
 	/* This zone will be used for allocation, so mark this zone non-empty. */
 	btrfs_dev_clear_zone_empty(device, info->physical);
 
-	down_read(&dev_replace->rwsem);
 	dev_replace_is_ongoing = btrfs_dev_replace_is_ongoing(dev_replace);
 	if (dev_replace_is_ongoing && dev_replace->tgtdev != NULL)
 		btrfs_dev_clear_zone_empty(dev_replace->tgtdev, info->physical);
-	up_read(&dev_replace->rwsem);
 
 	/*
 	 * The group is mapped to a sequential zone. Get the zone write pointer
@@ -1330,6 +1333,7 @@ static int btrfs_load_zone_info(struct btrfs_fs_info *fs_info, int zone_idx,
 	ret = btrfs_get_dev_zone(device, info->physical, &zone);
 	memalloc_nofs_restore(nofs_flag);
 	if (ret) {
+		up_read(&dev_replace->rwsem);
 		if (ret != -EIO && ret != -EOPNOTSUPP)
 			return ret;
 		info->alloc_offset = WP_MISSING_DEV;
@@ -1341,6 +1345,7 @@ static int btrfs_load_zone_info(struct btrfs_fs_info *fs_info, int zone_idx,
 		"zoned: unexpected conventional zone %llu on device %s (devid %llu)",
 			zone.start << SECTOR_SHIFT, rcu_str_deref(device->name),
 			device->devid);
+		up_read(&dev_replace->rwsem);
 		return -EIO;
 	}
 
@@ -1368,6 +1373,8 @@ static int btrfs_load_zone_info(struct btrfs_fs_info *fs_info, int zone_idx,
 		break;
 	}
 
+	up_read(&dev_replace->rwsem);
+
 	return 0;
 }