All riscv64 development halts effective immediately

[kth5]

I've been considering this for quite some time but it's finally pissing me off so much that I can't be bothered to keep maintaining riscv64 anymore.

Every vendor is doing their ever so slighly different implementations of everything: no-vector, RVV 0.7 ! compatible with RVV 1.0 and other things such as firmware always being that tiny bit off between devices to prevent an at least dual-unified way to boot things. The future not looking very bright either...

The latest example being that a Milk-V provided kernel for the Sopgho2024 (employed on the Pioneer) just causing illegal instruction traps on the same code that runs on an older kernel. Why is irrellevant but in order to maintain an up to date vendor maintained kernel, I'd have to rebuild all of the repo which then of course has the great potential of breaking everything else, i.e. the JH7110 etc. I just can't justify spending any time at all on that and won't leave things as they are right now, thus it'll trickle into archive only mode.

With the riscv64 ecosystem moving into a similar direction that ARM has (very fragmented usual SBC style landscape), I just don't have the time or motivation to continue development here.

If you happen to use Arch POWER on riscv64 consider moving over to Archlinux RiscV which has a very different approach and focusses entirely only on riscv64 (which apparently is more work than 4 PowerPC arches).

https://archriscv.felixc.at/

My apologies if this comes at an inconvenient time but such is FOSS (also). 🙇