[Mesa-dev] [PATCH 1/2] panfrost: Initial stub for Panfrost driver

[Jason Ekstrand]

On Thu, Jan 31, 2019 at 10:36 AM Eric Anholt <eric at anholt.net> wrote:

> Alyssa Rosenzweig <alyssa at rosenzweig.io> writes:
>
> >> There's u_pipe_screen_get_param_defaults() that you really want to be
> >> using to avoid regressions when people add new pipe caps.  You can get
> >> rid of a lot of your switch statement that way, too.
> >
> > Oh, good to know. Is that newer? I'm pretty sure that whole block was
> > copied from vc4 at some point ;)
> >
> >> We should probably be sticking the kmsro entrypoints in a shared group,
> >> since there's nothing specific for the KMS .  Looks like we haven't been
> >> doing that, though.
> >
> > Yeah, I had the same thought, but I didn't want to disrupt other drivers
> > yet (since that delays merge due to pinging more people).
> >
> >> These are both things we can change later.  I'm mostly excited to see
> >> you finally in-tree!
>

Hear, Hear!  Whatever merge/patch you do is

Acked-by: Jason Ekstrand <jason at jlekstrand.net>

It's about time!


> > ^_^
> >
> > Should either of the above issues be resolved in a v2 now, or just a
> > follow-up patch after merging?
>
> Follow-up, imo.  New drivers have a long awkward period where they
> wouldn't pass review the way we would for incremental development, but
> continuing to work out-of-tree means rebase pain and losing git history.
>
> (If you have git history out of tree that you'd like to retain when
> merging in, I'm happy to try helping with that.  I've done fun git
> surgery before)
>

Doing merge commits for new drivers is not unheard of and preserves history
(if the history is something you want to preserve).

--Jason
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/mesa-dev/attachments/20190131/0a23c8cd/attachment.html>