Chris Webb writes: > I'd like to support metadata types 0.90 and 1.0 too, relocating the > superblock. I think it'd be sufficient to write a new one at the end of > the device then update rdev->sb_offset. I'll try this next week. I've had a quick stab at this before I leave along with fixing a couple of other oversights. (I just return EBUSY if there's a bitmap present in 0.90 and 1.0.) It appears to work correctly for 0.90 but generates a corrupt superblock for 1.0, which I'll need to fix on Monday. Cheers, Chris.