Module: Mesa
Branch: master
Commit: edb7890750f8230021c1efa8e4d0d7a383f076c4
URL:    
http://cgit.freedesktop.org/mesa/mesa/commit/?id=edb7890750f8230021c1efa8e4d0d7a383f076c4

Author: Eric Anholt <[email protected]>
Date:   Wed Jun 20 10:22:44 2018 -0700

v3d: Fix min vs mag determination when not doing mip filtering.

Fixes all 128 failing tests in
dEQP-GLES3.functional.texture.filtering.*.combinations

---

 src/gallium/drivers/v3d/v3dx_state.c | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)

diff --git a/src/gallium/drivers/v3d/v3dx_state.c 
b/src/gallium/drivers/v3d/v3dx_state.c
index 70c596855f..c0f43d800e 100644
--- a/src/gallium/drivers/v3d/v3dx_state.c
+++ b/src/gallium/drivers/v3d/v3dx_state.c
@@ -556,9 +556,17 @@ v3d_create_sampler_state(struct pipe_context *pctx,
                                                    15);
                 sampler.max_level_of_detail = MIN2(cso->max_lod, 15);
 
+                /* If we're not doing inter-miplevel filtering, we need to
+                 * clamp the LOD so that we only sample from baselevel.
+                 * However, we need to still allow the calculated LOD to be
+                 * fractionally over the baselevel, so that the HW can decide
+                 * between the min and mag filters.
+                 */
                 if (cso->min_mip_filter == PIPE_TEX_MIPFILTER_NONE) {
-                        sampler.min_level_of_detail = 0;
-                        sampler.max_level_of_detail = 0;
+                        sampler.min_level_of_detail =
+                                MIN2(sampler.min_level_of_detail, 1.0 / 256.0);
+                        sampler.max_level_of_detail =
+                                MIN2(sampler.max_level_of_detail, 1.0 / 256.0);
                 }
 
                 if (cso->max_anisotropy) {

_______________________________________________
mesa-commit mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/mesa-commit

Reply via email to