Skip to contentSkip to conversation

btrfs: don't allow adding block device with 0 bytes

PATCH In review Mark Harmstone 3 participants 4 messages Aug 27, 2025 – Aug 28, 2025 Btrfs file system

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 btrfs: don't allow adding block device with 0 bytes

Commit 15ae0410 in btrfs-progs inadvertently changed it so that if the BLKGETSIZE64 ioctl on a block device returned a size of 0, this was no longer seen as an error condition.

Unfortunately this is how disconnected NBD devices behave, meaning that with btrfs-progs 6.16 it's now possible to add a device you can't remove:

~ # btrfs device add /dev/nbd0 /root/temp
~ # btrfs device remove /dev/nbd0 /root/temp
ERROR: error removing device '/dev/nbd0': Invalid argument

This check should always have been done kernel-side anyway, so add a check in btrfs_init_new_device() that the new device doesn't have a size of 0.

---
 fs/btrfs/volumes.c | 5 +++++
 1 file changed, 5 insertions(+)
fs/btrfs/volumes.c +5 -0
index 63b65f70a2b3..0757a546ff5c 100644
@@ -2726,6 +2726,11 @@ int btrfs_init_new_device(struct btrfs_fs_info *fs_info, const char *device_path
 		goto error;
 	}
 
+	if (bdev_nr_bytes(file_bdev(bdev_file)) == 0) {
+		ret = -EINVAL;
+		goto error;
+	}
+
 	if (fs_devices->seeding) {
 		seeding_dev = true;
 		down_write(&sb->s_umount);
Signature · 1 line
2.49.1
Reviewed-by on v1 1/1
#1 lore raw
Quoted 3 lines · Mark Harmstone

Commit 15ae0410 in btrfs-progs inadvertently changed it so that if the BLKGETSIZE64 ioctl on a block device returned a size of 0, this was no longer seen as an error condition.

My bad, that commit changed the handling of zero sized device, previously the caller who checks the returned size now only checks if the ioctl/stat works.

Feel free to send out a progs fix.

Quoted 9 lines

Unfortunately this is how disconnected NBD devices behave, meaning that with btrfs-progs 6.16 it's now possible to add a device you can't remove:

~ # btrfs device add /dev/nbd0 /root/temp
~ # btrfs device remove /dev/nbd0 /root/temp
ERROR: error removing device '/dev/nbd0': Invalid argument

Quoted 5 lines

This check should always have been done kernel-side anyway, so add a check in btrfs_init_new_device() that the new device doesn't have a size of 0.

The extra kernel checks looks good to me.

Thanks,
Qu

Quoted 20 lines
---
  fs/btrfs/volumes.c | 5 +++++
  1 file changed, 5 insertions(+)
fs/btrfs/volumes.c +5 -0
index 63b65f70a2b3..0757a546ff5c 100644
@@ -2726,6 +2726,11 @@ int btrfs_init_new_device(struct btrfs_fs_info *fs_info, const char *device_path
  		goto error;
  	}
 
+	if (bdev_nr_bytes(file_bdev(bdev_file)) == 0) {
+		ret = -EINVAL;
+		goto error;
+	}
+
  	if (fs_devices->seeding) {
  		seeding_dev = true;
  		down_write(&sb->s_umount);
#2 lore raw
Quoted 23 lines · Qu Wenruo
Quoted 3 lines · Mark Harmstone

Commit 15ae0410 in btrfs-progs inadvertently changed it so that if the BLKGETSIZE64 ioctl on a block device returned a size of 0, this was no longer seen as an error condition.

My bad, that commit changed the handling of zero sized device, previously the caller who checks the returned size now only checks if the ioctl/stat works.

Feel free to send out a progs fix.

Quoted 13 lines

Unfortunately this is how disconnected NBD devices behave, meaning that with btrfs-progs 6.16 it's now possible to add a device you can't remove:

~ # btrfs device add /dev/nbd0 /root/temp
~ # btrfs device remove /dev/nbd0 /root/temp
ERROR: error removing device '/dev/nbd0': Invalid argument

This check should always have been done kernel-side anyway, so add a check in btrfs_init_new_device() that the new device doesn't have a size of 0.

Quoted 5 lines
Quoted 1 line

The extra kernel checks looks good to me.

In fact this reminds me to try other extreme situations.

Like adding a 64K sized device.

The result is, the fs will never be able to be mounted again, as we saved the cursed device which can not have the primary super block written.

Since we're here, can we enlarge the bdev size to <= BTRFS_DEVICE_RANGE_RESERVED?

Thanks,
Qu

Quoted 25 lines

Thanks,
Qu

Quoted 20 lines

---
  fs/btrfs/volumes.c | 5 +++++

  1 file changed, 5 insertions(+)
fs/btrfs/volumes.c +5 -0
index 63b65f70a2b3..0757a546ff5c 100644
@@ -2726,6 +2726,11 @@ int btrfs_init_new_device(struct btrfs_fs_info 
 *fs_info, const char *device_path
           goto error;
       }
+    if (bdev_nr_bytes(file_bdev(bdev_file)) == 0) {
+        ret = -EINVAL;
+        goto error;
+    }
+
       if (fs_devices->seeding) {
           seeding_dev = true;
           down_write(&sb->s_umount);
#3 lore raw
Quoted 32 lines · Qu Wenruo
Quoted 23 lines · Qu Wenruo
Quoted 3 lines · Mark Harmstone

Commit 15ae0410 in btrfs-progs inadvertently changed it so that if the BLKGETSIZE64 ioctl on a block device returned a size of 0, this was no longer seen as an error condition.

My bad, that commit changed the handling of zero sized device, previously the caller who checks the returned size now only checks if the ioctl/stat works.

Feel free to send out a progs fix.

Quoted 13 lines

Unfortunately this is how disconnected NBD devices behave, meaning that with btrfs-progs 6.16 it's now possible to add a device you can't remove:

~ # btrfs device add /dev/nbd0 /root/temp
~ # btrfs device remove /dev/nbd0 /root/temp
ERROR: error removing device '/dev/nbd0': Invalid argument

This check should always have been done kernel-side anyway, so add a check in btrfs_init_new_device() that the new device doesn't have a size of 0.

Quoted 5 lines
Quoted 1 line

The extra kernel checks looks good to me.

Quoted 7 lines

In fact this reminds me to try other extreme situations.

Like adding a 64K sized device.

The result is, the fs will never be able to be mounted again, as we saved the cursed device which can not have the primary super block written.

Since we're here, can we enlarge the bdev size to <= BTRFS_DEVICE_RANGE_RESERVED?

Makes sense to me. I'll send another patch.

Quoted 28 lines

Thanks,
Qu

Quoted 24 lines

Thanks,
Qu

Quoted 19 lines

---
  fs/btrfs/volumes.c | 5 +++++

  1 file changed, 5 insertions(+)
fs/btrfs/volumes.c +5 -0
index 63b65f70a2b3..0757a546ff5c 100644
@@ -2726,6 +2726,11 @@ int btrfs_init_new_device(struct btrfs_fs_info *fs_info, const char *device_path
           goto error;
       }
+    if (bdev_nr_bytes(file_bdev(bdev_file)) == 0) {
+        ret = -EINVAL;
+        goto error;
+    }
+
       if (fs_devices->seeding) {
           seeding_dev = true;
           down_write(&sb->s_umount);