[libre-riscv-dev] porting AMDVLK to the Libre RISC-V 3D GPU: NLNet EUR 50, 000 Grant application

Michael Pham pham.michael.98 at gmail.com
Tue Oct 1 22:00:38 BST 2019


Oh I remember what my third point is now.

Thirdly, even if we don't fund the RADV developers (which I don't see
in the original proposal anyway), that doesn't stop us from creating a
Libre RISC-V derivative of RADV. You guys (Jacob and Luke) can use the
funds for your own development (plus anyone who wants to help out with
the proposal).

Michael

On Tue, Oct 1, 2019 at 3:37 PM Michael Pham <pham.michael.98 at gmail.com> wrote:
>
> On Tue, Oct 1, 2019 at 12:35 AM Jacob Lifshay <programmerjake at gmail.com> wrote:
> >
> > I'd be inclined to use RADV anyway because it has much more community
> > support, is in mesa, and is not just code thrown over the wall (which
> > is what AMDVLK is, since AMDVLK is basically AMD's proprietary shared
> > windows/linux graphics driver rewired to use LLVM instead of their
> > proprietary compiler.
> >
> > Will discuss more on-list after Michael's response.
> >
> > Jacob
>
> I've responded to the whole amd graphics driver dispute off list. So
> let's return to the subject of RADV with technical arguments. First of
> all, even though RADV was basically started by Dave, he's not that
> involved with the development anymore and many other Mesa developers
> are working on it instead. We shouldn't throw away the possibility of
> using RADV just because of the person who started it -- that would be,
> in my opinion, disrespectful of everyone else working on RADV by
> ignoring their contributions.
>
> Secondly, if we do forget about RADV, then we should forget about
> using AMDVLK as well for the obvious reason that AMDVLK is by AMD that
> Luc seems to be so antagonistic towards. And we can continue to forget
> about Intel and ARM drivers as well because, well, as you can see on
> Luc's blog is antagonistic towards them as well. My point is that we
> shouldn't let Luc's accusations (true or not) stop us from reusing any
> of these drivers because at that point we might as well just ignore
> every corporation's open-source contributions. For example, Microsoft
> has done some very unethical things in the past and also has
> contributed code to the Linux kernel. What are you going to do? Fork
> the Linux kernel and rip out Microsoft's code? The RISC-V Foundation
> has acted unethical as well by not being as open as they claim. What
> are we going to do? Throw away all our work because it's based on
> RISC-V? I think you get what I'm trying to say.
>
> Thirdly.... ahhh I forgot what else I was trying to say :( If I
> remember I'll send out another email.
>
> Note to everyone: Luc is not Luke. Different people, not a misspelling.
>
> Michael



More information about the libre-riscv-dev mailing list