David Chisnall

> uintptr_t needs to have full XLEN range for arithmetic, that's non-negotiable in the real world. That means the address needs to be the full XLEN bits. Not quite. `uintptr_t`...

> I don't think a world where CGetAddr(CClearTag(foo)) != CGetAddr(foo) is a good idea. We've been down roads like that before and it's been porting pain. I'd like to see...

I see. Do you have a reference to where you've seen that? I've heard people mention it but never actually seen it.

Thanks. I shouldn’t be surprised it’s xlib.

I think it is the right approach. I believe @nwf-msr has a concrete proposal along these lines underway.

> The downside would be that we do need a representability check and have to use the CHERI ALU rather than the integer one for these instructions. Should not matter...

There's a difference between a single manual and a single non-decomposable ISA. The unprivileged spec manual describes several extensions which can be implemented, not implemented, or replaced with others. This...

> where does it say this? Have you got a reference? That's not the way RISC-V works. Any extension is optional. Extensions that depend on other extensions are explicitly documented...

It looks as if you have failed to clone the submodules. I thought we had a CMake check for that.

Ah, yes, we stopped using submodules. We now either use the systemwide installation or use FetchContent to download it. Unfortunately, your report elides all of the CMake output, which is...