On Tue, 4 Mar 2025, Michael van Elst wrote:
netbsd%precedence.co.uk@localhost (Stephen Borrill) writes:Looking good! Also mmm, a kernel that's 800k in size :-)Are these changes commitable?I didn't change any code, just modified the superblock a little bit. In particular, it now says it is 8k in size and some bits that were undefined (all 1's) but are interpreted by our kernel, are now set.
Ah, I see. You adjusted the filesystem to fit the kernel rather than the other way round.
Shouldn't be difficult to make the kernel accept the original superblock.
This is a potentially live filesystem type, not a historical archive. There appears to have been a resurgence in running RISCiX both under emulation and on real hardware.
N.B. fsck and other userland tools are still confused by the "alternate superblocks". I guess these also need to be more relaxed about undefined bits.
Let me know if there's anything I can do to help. Stephen