nvc0: fix gl_SampleMaskIn computation

The SAMPLEMASK semantic should only return the bits set covered by the
current invocation. However we were always retrieving the covmask, which
returns the covered samples of the whole pixel.

When not doing per-sample invocation, this is precisely what we want.
However when doing per-sample invocation, we have to select the
sampleid'th bit and only return that. Furthermore, this means that we
have to have a 1:1 correlation for invocations and samples.

This fixes most

dEQP-GLES31.functional.shaders.sample_variables.sample_mask_in.*

tests. A few failures remain due to disagreements about nr_samples==1
logic as well as what happens with MSAA x2 RTs when the shading fraction
is 0.5.

Signed-off-by: Ilia Mirkin <[email protected]>

3 files changed, 13 insertions, 2 deletions

diff --git a/src/gallium/drivers/nouveau/nvc0/nvc0_program.c b/src/gallium/drivers/nouveau/nvc0/nvc0_program.c
index 944efa042bf..9db45c0759a 100644
--- a/src/gallium/drivers/nouveau/nvc0/nvc0_program.c
+++ b/src/gallium/drivers/nouveau/nvc0/nvc0_program.c
@@ -464,6 +464,7 @@ nvc0_fp_gen_header(struct nvc0_program *fp, struct nv50_ir_prog_info *info)
       fp->hdr[18] |= 0xf;
 
    fp->fp.early_z = info->prop.fp.earlyFragTests;
+   fp->fp.sample_mask_in = info->prop.fp.usesSampleMaskIn;
 
    return 0;
 }
diff --git a/src/gallium/drivers/nouveau/nvc0/nvc0_program.h b/src/gallium/drivers/nouveau/nvc0/nvc0_program.h
index bd852e27c36..08af3c823b8 100644
--- a/src/gallium/drivers/nouveau/nvc0/nvc0_program.h
+++ b/src/gallium/drivers/nouveau/nvc0/nvc0_program.h
@@ -48,6 +48,7 @@ struct nvc0_program {
       uint8_t early_z;
       uint8_t colors;
       uint8_t color_interp[2];
+      bool sample_mask_in;
       bool force_persample_interp;
       bool flatshade;
    } fp;
diff --git a/src/gallium/drivers/nouveau/nvc0/nvc0_state_validate.c b/src/gallium/drivers/nouveau/nvc0/nvc0_state_validate.c
index e8d41729392..4280db44bb6 100644
--- a/src/gallium/drivers/nouveau/nvc0/nvc0_state_validate.c
+++ b/src/gallium/drivers/nouveau/nvc0/nvc0_state_validate.c
@@ -1,5 +1,6 @@
 
 #include "util/u_format.h"
+#include "util/u_framebuffer.h"
 #include "util/u_math.h"
 
 #include "nvc0/nvc0_context.h"
@@ -551,8 +552,14 @@ nvc0_validate_min_samples(struct nvc0_context *nvc0)
    int samples;
 
    samples = util_next_power_of_two(nvc0->min_samples);
-   if (samples > 1)
+   if (samples > 1) {
+      // If we're using the incoming sample mask and doing sample shading, we
+      // have to do sample shading "to the max", otherwise there's no way to
+      // tell which sets of samples are covered by the current invocation.
+      if (nvc0->fragprog->fp.sample_mask_in)
+         samples = util_framebuffer_get_num_samples(&nvc0->framebuffer);
       samples |= NVC0_3D_SAMPLE_SHADING_ENABLE;
+   }
 
    IMMED_NVC0(push, NVC0_3D(SAMPLE_SHADING), samples);
 }
@@ -708,6 +715,9 @@ validate_list_3d[] = {
     { nvc0_tevlprog_validate,      NVC0_NEW_3D_TEVLPROG },
     { nvc0_validate_tess_state,    NVC0_NEW_3D_TESSFACTOR },
     { nvc0_gmtyprog_validate,      NVC0_NEW_3D_GMTYPROG },
+    { nvc0_validate_min_samples,   NVC0_NEW_3D_MIN_SAMPLES |
+                                   NVC0_NEW_3D_FRAGPROG |
+                                   NVC0_NEW_3D_FRAMEBUFFER },
     { nvc0_fragprog_validate,      NVC0_NEW_3D_FRAGPROG | NVC0_NEW_3D_RASTERIZER },
     { nvc0_validate_derived_1,     NVC0_NEW_3D_FRAGPROG | NVC0_NEW_3D_ZSA |
                                    NVC0_NEW_3D_RASTERIZER },
@@ -726,7 +736,6 @@ validate_list_3d[] = {
     { nvc0_validate_buffers,       NVC0_NEW_3D_BUFFERS },
     { nvc0_idxbuf_validate,        NVC0_NEW_3D_IDXBUF },
     { nvc0_tfb_validate,           NVC0_NEW_3D_TFB_TARGETS | NVC0_NEW_3D_GMTYPROG },
-    { nvc0_validate_min_samples,   NVC0_NEW_3D_MIN_SAMPLES },
     { nvc0_validate_driverconst,   NVC0_NEW_3D_DRIVERCONST },
 };