On Tue, Nov 29, 2022 at 9:31 PM Dave Chinner <david@xxxxxxxxxxxxx> wrote: > Looks like something to do with the gfs2 inline data functionality - > syzbot probably corrupted the resource index inode given the > gfs2_fill_super() context. Hmm, interesting. We're not checking the size of inline (stuffed) inodes when reading them from disk in gfs2_dinode_in(). I'll fix that. Thanks, Andreas
