[lowrisc-dev] Re: [GSoC_16] Porting Xinu/xv6 to the lowRISC platform

Alex Bradbury asb at asbradbury.org
Tue Mar 8 14:47:25 GMT 2016


On 8 March 2016 at 13:41, Shibjash Dutt <sjd96bvn at gmail.com> wrote:
> Hi,
>
> I'm a student interested in porting either Xinu or xv6 to the
> lowRISC platform, as part of this year's Google Summer of Code.
>
> There was a thread posted earlier by another interested student, and a
> subsequent
> followup by asb. However, I have some questions of my own and would be glad if
> the dev team could shed some light on them.
>
> 1) Alex Bradbury says targeting Spike or QEMU should be OK. As far as Spike is
>    concerned, how complete is it or expected to get? AFAIK Spike is somewhat
>    behind on some stuff, like tagged memory and full peripheral support. If my
>    project works on Spike, will it run (mostly) unmodified on the hardware?

Now QEMU is in a good shape, it's probably the better target but you
should be fine with either. Differences will come down to peripheral
support as you point out in your next question.

> 2) I've built the GNU toolchain, Spike and the kernel. So far, it's pretty bare
>    bones and all the functionality is exposed through the HTIF. Are there any
>    short term plans to standardize an interface?

It's very likely we'll have more standardisation here by the project
starts. In the mean time, planning to use standard virtio devices (as
supported by qemu) is probably a good start.

> 3) Apart from Wei Songs's excellent tutorial accompanying the v0.2 release,
>    and the phabricator wiki, are there any other sources of information or
>    documentation? Or is it read the code and figure it out :P

The various documents we've put out and Wei's excellent tutorials are
the best sources of information.

> 4) How strong would be the interest in such a project?

Well, I'm very interested in mentoring if we receive a sufficiently
compelling application, and I think it has the potential of having a
strong positive impact. RISC-V is excellent as a teaching tool and it
makes perfect sense to use it to teach operating systems principles as
well as computer architecture design. What I'd like is to find a
student who has a passion for this project, and perhaps has some of
their own ideas about what to do with the project beyond just getting
xv6 to boot.

Alex



More information about the lowrisc-dev mailing list