diff mbox series

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

Message ID ccab86c78fc2a1fd6dd320efa289863df7158ca9.1715167144.git.fdmanana@suse.com (mailing list archive)
State New, archived
Headers show
Series btrfs: zoned: fix use-after-free due to race with dev replace | expand

Commit Message

Filipe Manana May 8, 2024, 11:20 a.m. UTC 
From: Filipe Manana <fdmanana@suse.com>

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 operaton 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.

Signed-off-by: Filipe Manana <fdmanana@suse.com>
---
 fs/btrfs/zoned.c | 13 ++++++++++---
 1 file changed, 10 insertions(+), 3 deletions(-)

Comments

Johannes Thumshirn May 8, 2024, 3:29 p.m. UTC | #1 
Looks good,
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
David Sterba May 9, 2024, 5:13 p.m. UTC | #2 
On Wed, May 08, 2024 at 12:20:40PM +0100, fdmanana@kernel.org wrote:
> From: Filipe Manana <fdmanana@suse.com>
> 
> 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 operaton 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.
> 
> Signed-off-by: Filipe Manana <fdmanana@suse.com>

Reviewed-by: David Sterba <dsterba@suse.com>

I'll add CC for 6.1+ before sending a pull request, there are some
dependencies but otherwise it applies cleanly.
diff mbox series

Patch

diff --git a/fs/btrfs/zoned.c b/fs/btrfs/zoned.c
index aeab33708702..eb70fd80dbc5 100644
--- a/fs/btrfs/zoned.c
+++ b/fs/btrfs/zoned.c
@@ -1290,7 +1290,7 @@  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;
@@ -1298,7 +1298,11 @@  static int btrfs_load_zone_info(struct btrfs_fs_info *fs_info, int zone_idx,
 
 	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;
 	}
@@ -1308,6 +1312,7 @@  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;
 	}
@@ -1315,11 +1320,9 @@  static int btrfs_load_zone_info(struct btrfs_fs_info *fs_info, int zone_idx,
 	/* 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;
 }