On 12/11/2017 11:06 AM, Marek Olšák wrote:
On Thu, Dec 7, 2017 at 5:09 PM, Brian Paul <bri...@vmware.com> wrote:
If we're doing a Z -> Z MSAA blit (for example) we need to enable
msaa rasterization when drawing the quads so that we can properly
write the per-sample values.
This fixes a number of Piglit ext_framebuffer_multisample blit tests
such as ext_framebuffer_multisample/no-color 2 depth combined with
the VMware driver.
---
Marek, I'm not sure how this could have worked without MSAA enabled.
Perhaps you can comment.
Enabling MSAA rasterization should have no effect, because the blitter
draws an axis-aligned quad, so each pixel will generate only one
sample.
radeon drivers still obey the sample mask even if MSAA is disabled. I
guess svga doesn't.
If we'd set the sample mask in the fs (with a TGSI_SEMANTIC_SAMPLEMASK
output register) we'd get automatic MSAA, but we don't do that with the
pipe_context::set_sample_mask() state.
I guess enabling MSAA is OK, but now we also have to save and restore
the "min_samples" value if it's not 1, because we don't want to use
sample shading because u_blitter doesn't need it.
OK, I can add a predecessor patch to take care of that. I'll send a v2...
-Brian
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev
