[Bug 89580] Implement a NIR -> vec4 pass

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=89580

--- Comment #28 from Iago Toral <itoral at igalia.com> ---
(In reply to Kenneth Graunke from comment #27)
> (In reply to Jason Ekstrand from comment #26)
> > It sounded, from what you said above, like you had already modified
> > nir_lower_io to handle vec4.  If that's  the case just go ahead and do the
> > rename and add the vec4 version.  No reason to wait for me or Connor to do
> > it.

Partially, I did that only when indirect indexing was involved, otherwise we
could work with scalar units, which is not great. In any case, it makes sense
that I add the change as part of the series so I'll do that.

> Yeah, I think your approach is great.
> 
> Instead of passing a gl_shader_stage and checking for !=
> MESA_SHADER_FRAGMENT, we should just pass a boolean indicating whether we
> should make things scalar or pad to vec4s.  Or have nir_lower_io_scalar and
> nir_lower_io_vec4 functions (effectively the same thing).
> 
> That way, i965's scalar-or-not choices aren't baked into the generic code,
> and everybody can pick what they want.

Yeah, that sounds good to me.

Thanks for the feedback!

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/intel-3d-bugs/attachments/20150622/39fbd75e/attachment.html>