Re: QEMU PPC VLE support

[Peter Maydell]

On Thu, 15 Dec 2022 at 13:38, Stefan Hajnoczi <stefanha@gmail.com> wrote:
>
> Hi,
> I came across this post where Ralf-Philipp is looking for a freelancer
> to implement PPC VLE support in QEMU:
> https://chaos.social/@rpw/109516326028642262
>
> It mentions upstreaming the code and I've included QEMU PPC
> maintainers in this email so they can discuss the project with
> Ralf-Philipp. That way the chances of a mergable result will be
> maximized.
>
> The Rust aspect is interesting, but QEMU does not have any existing
> targets implemented in Rust. It might be a major effort to create the
> necessary C<->Rust interfacing, so I'm not sure whether that's
> realistic given the timeframe for the project.

Yes, "in Rust" is at this point a good way to ensure it will *not*
get upstream. Not because we have an antipathy to Rust
(indeed I think it's a good long-term goal to have)
but because as Stefan suggests we would need to start by
figuring out what a transition plan would look like, where
would be a good place to start looking at Rust-like APIs
for things, whether the community as a whole has enough
Rust enthusiasts to push the project forward, etc. What we
don't want is "QEMU is 99% C except the PPC target has this
one weird corner that's in Rust and doesn't fit well with
everything else". I look at how much effort the
Rust-in-Linux-kernel folks have had to put in to get to the
point where they are now, and while I hope we won't need to
do that much work, their general approach to it seems worth
paying attention to.

thanks
-- PMM