Re: [PATCH v1 0/2] RISC-V: enable rust

[Conor Dooley <conor@kernel.org>]

[-- Attachment #1: Type: text/plain, Size: 1809 bytes --]

On Wed, Jan 17, 2024 at 07:23:17PM +0100, Miguel Ojeda wrote:
> On Wed, Jan 17, 2024 at 12:31 PM Conor Dooley <conor@kernel.org> wrote:
> >
> > 6.6 came and went, and I have been busy dealing with the other
> > responsibilities I mentioned and have not had a chance to look here.
> > I rebased this today and things still work as they did when I submitted
> > this version, but things have gotten muddier on the LLVM side of things,
> > as more recent versions have added yet more extension support.
> 
> Sounds fun :)
> 
> > My inclination at this point is to engage in a bit of LARPing as an
> > ostrich, and sorta ignore these concerns initially. Specifically, I'd
> > like to drop the idea of having the gcc support, and restrict to LLVM=1.
> 
> Yeah, if `LLVM=1` works, then I would suggest going ahead with that.
> 
> (Now that `rustc_codegen_gcc` is here, we will move to that and forget
> about mixed compiler builds, but we still have to handle `bindgen`
> flags until we have an alternative for that)

The bit that worries me most is bindgen, and in particular detecting the
version of libclang used. I mentioned to Nathan or Nick about needing a
buildtime test for the version of LIBCLANG being used.
I'm less worried about this for LLVM=1 builds, since while I think it is
possible to provide a LIBCLANG path to the build system, I suspect that
for LLVM=1 builds it's almost always going to match the LLVM toolchain
in use.

> > When it comes to asymmetrical extension support between the C and Rust
> > toolchains, I'm think we deal with that as we do for the C toolchains,
> > sort issues out as-and-when they arrive rather than punt this again.
> 
> Sounds good, thanks a lot!

I'll do another rebase and resend after the merge window closes I
suppose :)

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]