ARMv8 or 64bit RISCOS?

Will there be a 64 bit version of RISC OS Open, assuming the compilers exist?

no

yes

aarch64 is essentially a totally different architecture, closer to MIPS than any previous ARM.
My vote is “no”.

There’s a lot of assembly language to port.

I asked and got a prompt reply, thanks :)

Maybe a ARMv8 RISCOS is possible:
“ARMv8-A allows (….) a 32-bit OS to be under the control of a 64-bit hypervisor.” “AArch64 provides user-space compatibility with ARMv7-A ISA, the 32-bit architecture (…..) and the old 32-bit instruction set (…..).”

That would neither be a 64-bit RISC OS nor an ARMv8 RISC OS. Just the same ARMv < 8 RISC OS running until ARMv8+n deprecates 32bit mode.

Another thing – given a 64 bit OS, what would we run on it?

Nothing, it would make the 26-32bit app porting issues look like a tiny bump.

Basically it’s as complicated as porting RISC OS to say x64.

> Basically it’s as complicated as porting RISC OS to say x64.

Would porting it even be permitted under the licence, if it is that different?

Would porting it even be permitted under the licence, if it is that different?

ARM yes, x86 no.

I’m sure ARM are going to be doing for many years 32bit and increasingly faster CPUs. The question is will they or at least some of them still use ARMv7 and be made available in suitable SoCs?
I would have thought the answer will be yes. I believe you can still buy Z80s, 6502s etc!

Edit: Just realised that for existing CPUs it is the licensees that matter not ARM. The licensees will still keep making them if they see a market.

In terms of refactoring the source code, it may be worth tackling a component at a time and translating/transcoding to C. A degree of incremental refactoring to make the code base more managable, would seem sensible and would get more control of the code i.e. some better modelling/documentation could be done in parallel. This might make a 64 bit port more plausible.

Then running legacy applications could involve a degree of sandboxing/emulation to enable legacy behaviours to be maintained. This can only be tackled incrementally in order to maintain stability, in my opinion, on the basis of my own experience in refactoring legacy commercial products.

I need to start trying to build the Risc OS tree really. More do and less talk (trouble is writing games is such fun). I think I might support ROOL and order the new Nut Pi….

I have gone and bought a Nuttier Pi. Looking forward to compiling RISC OS.