[Bug 89580] Implement a NIR -> vec4 pass

[bugzilla-daemon at freedesktop.org]

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

--- Comment #12 from Jason Ekstrand <jason at jlekstrand.net> ---
(In reply to Eduardo Lima Mitev from comment #11)
> This is odd. In some piglit tests (like "bin/glean -o -v -v -v -t +api2"),
> I'm getting a nir_shader with no uniforms declared, yet it loads and uses
> some:
> 
> NIR (final form) for vertex shader:
> decl_var shader_in  vec4 in_0 (0, 0)
> decl_var shader_out  vec4 out_0 (0, 0)
> decl_overload main returning void
> 
> impl main {
> 	decl_reg vec4 r2
> 	decl_reg vec4 r3
> 	decl_reg vec1 r4
> 	decl_reg vec4 r5
> 	decl_reg vec1 r6
> 	decl_reg vec4 r7
> 	decl_reg vec1 r8
> 	decl_reg vec4 r9
> 	decl_reg vec1 r10
> 	decl_reg vec4 r11
> 	block block_0:
> 	/* preds: */
> 	r2 = intrinsic load_input () () (0, 1)
> 	r3 = intrinsic load_uniform () () (0, 1)
> 	r4 = fdot4 r2, r3
> 	r5 = intrinsic load_uniform () () (4, 1)
> 	r6 = fdot4 r2, r5
> 	r7 = intrinsic load_uniform () () (8, 1)
> 	r8 = fdot4 r2, r7
> 	r9 = intrinsic load_uniform () () (12, 1)
> 	r10 = fdot4 r2, r9
> 	r11 = vec4 r4, r6, r8, r10
> 	intrinsic store_output (r11) () (0, 1)
> 	/* succs: block_1 */
> 	block block_1:
> }
> 
> This looks like a bug to me. Am I missing something?

Taking a wild stab in the dark, I'm going to guess that you're hitting an ARB
or fixed-function program.  Those come in via the prog_to_nir path and the
uniform setup is a bit different.

-- 
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/20150505/93afa4a6/attachment.html>