On 12/7/19 7:06 AM, Christoph Anton Mitterer wrote:
I'm also seeing these since quite a while on Debian sid: Jul 11 13:33:56 heisenberg kernel: BTRFS info (device dm-0): device fsid 60[...]3c devid 1 moved old:/dev/mapper/system new:/dev/dm-0 Jul 11 13:33:56 heisenberg kernel: BTRFS info (device dm-0): device fsid 60[...]3c devid 1 moved old:/dev/dm-0 new:/dev/mapper/system Jul 11 23:43:35 heisenberg kernel: BTRFS info (device dm-0): device fsid 60[...]3c devid 1 moved old:/dev/mapper/system new:/dev/dm-0 Jul 11 23:43:35 heisenberg kernel: BTRFS info (device dm-0): device fsid 60[...]3c devid 1 moved old:/dev/dm-0 new:/dev/mapper/system In my case it's a simply dm-crypt layer below the fs. Some years ago, there was a longer thread on this list about the fragility of btrfs with respect to accidentally or intentionally colliding UUIDs.
IIRC there were quite some concerns that this could have even a big security impact when an attacker e.g. plugs in a device with a certain UUID and the kernel or userland automatically adds or somehow else uses such device (just by UUID).
Back then it was said this would be looked into... has anything happened there?
Thanks for refreshing on that report. Looking into it. -Anand
Cheers, Chris.
