nir: Add a "compact array" flag and IO lowering code.

Certain built-in arrays, such as gl_ClipDistance[], gl_CullDistance[],
gl_TessLevelInner[], and gl_TessLevelOuter[] are specified as scalar
arrays.  Normal scalar arrays are sparse - each array element usually
occupies a whole vec4 slot.  However, most hardware assumes these
built-in arrays are tightly packed.

The new var->data.compact flag indicates that a scalar array should
be tightly packed, so a float[4] array would take up a single vec4
slot, and a float[8] array would take up two slots.

They are still arrays, not vec4s, however.  nir_lower_io will generate
intrinsics using ARB_enhanced_layouts style component qualifiers.

v2: Add nir_validate code to enforce type restrictions.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>

1 files changed, 6 insertions, 2 deletions

diff --git a/src/compiler/nir/nir_lower_indirect_derefs.c b/src/compiler/nir/nir_lower_indirect_derefs.c
index 356373e2788..5c97dc8e5fe 100644
--- a/src/compiler/nir/nir_lower_indirect_derefs.c
+++ b/src/compiler/nir/nir_lower_indirect_derefs.c
@@ -175,8 +175,12 @@ lower_indirect_block(nir_block *block, nir_builder *b,
       if (!deref_has_indirect(intrin->variables[0]))
          continue;
 
-      /* Only lower variables whose mode is in the mask */
-      if (!(modes & intrin->variables[0]->var->data.mode))
+      /* Only lower variables whose mode is in the mask, or compact
+       * array variables.  (We can't handle indirects on tightly packed
+       * scalar arrays, so we need to lower them regardless.)
+       */
+      if (!(modes & intrin->variables[0]->var->data.mode) &&
+          !intrin->variables[0]->var->data.compact)
          continue;
 
       b->cursor = nir_before_instr(&intrin->instr);