[Mesa-dev] [PATCH v2 01/25] glsl: remember per-component vertex streams for packed varyings

[Nicolai Hähnle]

From: Nicolai Hähnle <nicolai.haehnle at amd.com>

---
 src/compiler/glsl/ir.h                      |  3 +++
 src/compiler/glsl/ir_print_visitor.cpp      | 14 ++++++++++++--
 src/compiler/glsl/lower_packed_varyings.cpp |  9 +++++++++
 3 files changed, 24 insertions(+), 2 deletions(-)

diff --git a/src/compiler/glsl/ir.h b/src/compiler/glsl/ir.h
index 24f510e..df3ccfd 100644
--- a/src/compiler/glsl/ir.h
+++ b/src/compiler/glsl/ir.h
@@ -910,20 +910,23 @@ public:
       /**
        * for glsl->tgsi/mesa IR we need to store the index into the
        * parameters for uniforms, initially the code overloaded location
        * but this causes problems with indirect samplers and AoA.
        * This is assigned in _mesa_generate_parameters_list_for_uniforms.
        */
       int param_index;
 
       /**
        * Vertex stream output identifier.
+       *
+       * For packed outputs, bit 31 is set and bits [2*i+1,2*i] indicate the
+       * stream of the i-th component.
        */
       unsigned stream;
 
       /**
        * Atomic, transform feedback or block member offset.
        */
       unsigned offset;
 
       /**
        * Highest element accessed with a constant expression array index
diff --git a/src/compiler/glsl/ir_print_visitor.cpp b/src/compiler/glsl/ir_print_visitor.cpp
index 703169e..2b77c14 100644
--- a/src/compiler/glsl/ir_print_visitor.cpp
+++ b/src/compiler/glsl/ir_print_visitor.cpp
@@ -170,37 +170,47 @@ void ir_print_visitor::visit(ir_variable *ir)
       snprintf(binding, sizeof(binding), "binding=%i ", ir->data.binding);
 
    char loc[32] = {0};
    if (ir->data.location != -1)
       snprintf(loc, sizeof(loc), "location=%i ", ir->data.location);
 
    char component[32] = {0};
    if (ir->data.explicit_component)
       snprintf(component, sizeof(component), "component=%i ", ir->data.location_frac);
 
+   char stream[32] = {0};
+   if (ir->data.stream & (1u << 31)) {
+      if (ir->data.stream & ~(1u << 31)) {
+         snprintf(stream, sizeof(stream), "stream(%u,%u,%u,%u)",
+                  ir->data.stream & 3, (ir->data.stream >> 2) & 3,
+                  (ir->data.stream >> 4) & 3, (ir->data.stream >> 6) & 3);
+      }
+   } else if (ir->data.stream) {
+      snprintf(stream, sizeof(stream), "stream%u", ir->data.stream);
+   }
+
    const char *const cent = (ir->data.centroid) ? "centroid " : "";
    const char *const samp = (ir->data.sample) ? "sample " : "";
    const char *const patc = (ir->data.patch) ? "patch " : "";
    const char *const inv = (ir->data.invariant) ? "invariant " : "";
    const char *const prec = (ir->data.precise) ? "precise " : "";
    const char *const mode[] = { "", "uniform ", "shader_storage ",
                                 "shader_shared ", "shader_in ", "shader_out ",
                                 "in ", "out ", "inout ",
 			        "const_in ", "sys ", "temporary " };
    STATIC_ASSERT(ARRAY_SIZE(mode) == ir_var_mode_count);
-   const char *const stream [] = {"", "stream1 ", "stream2 ", "stream3 "};
    const char *const interp[] = { "", "smooth", "flat", "noperspective" };
    STATIC_ASSERT(ARRAY_SIZE(interp) == INTERP_MODE_COUNT);
 
    fprintf(f, "(%s%s%s%s%s%s%s%s%s%s%s) ",
            binding, loc, component, cent, samp, patc, inv, prec, mode[ir->data.mode],
-           stream[ir->data.stream],
+           stream,
            interp[ir->data.interpolation]);
 
    print_type(f, ir->type);
    fprintf(f, " %s)", unique_name(ir));
 }
 
 
 void ir_print_visitor::visit(ir_function_signature *ir)
 {
    _mesa_symbol_table_push_scope(symbols);
diff --git a/src/compiler/glsl/lower_packed_varyings.cpp b/src/compiler/glsl/lower_packed_varyings.cpp
index b16f25f..7a2f187 100644
--- a/src/compiler/glsl/lower_packed_varyings.cpp
+++ b/src/compiler/glsl/lower_packed_varyings.cpp
@@ -524,20 +524,28 @@ lower_packed_varyings_visitor::lower_rvalue(ir_rvalue *rvalue,
        */
       unsigned swizzle_values[4] = { 0, 0, 0, 0 };
       unsigned components = rvalue->type->vector_elements * dmul;
       unsigned location = fine_location / 4;
       unsigned location_frac = fine_location % 4;
       for (unsigned i = 0; i < components; ++i)
          swizzle_values[i] = i + location_frac;
       ir_dereference *packed_deref =
          this->get_packed_varying_deref(location, unpacked_var, name,
                                         vertex_index);
+      if (unpacked_var->data.stream != 0) {
+         assert(unpacked_var->data.stream < 4);
+         ir_variable *packed_var = packed_deref->variable_referenced();
+         for (unsigned i = 0; i < components; ++i) {
+            packed_var->data.stream |=
+               unpacked_var->data.stream << (2 * (location_frac + i));
+         }
+      }
       ir_swizzle *swizzle = new(this->mem_ctx)
          ir_swizzle(packed_deref, swizzle_values, components);
       if (this->mode == ir_var_shader_out) {
          this->bitwise_assign_pack(swizzle, rvalue);
       } else {
          this->bitwise_assign_unpack(rvalue, swizzle);
       }
       return fine_location + components;
    }
 }
@@ -632,20 +640,21 @@ lower_packed_varyings_visitor::get_packed_varying_deref(
          packed_var->data.max_array_access = this->gs_input_vertices - 1;
       }
       packed_var->data.centroid = unpacked_var->data.centroid;
       packed_var->data.sample = unpacked_var->data.sample;
       packed_var->data.patch = unpacked_var->data.patch;
       packed_var->data.interpolation = packed_type == glsl_type::ivec4_type
          ? unsigned(INTERP_MODE_FLAT) : unpacked_var->data.interpolation;
       packed_var->data.location = location;
       packed_var->data.precision = unpacked_var->data.precision;
       packed_var->data.always_active_io = unpacked_var->data.always_active_io;
+      packed_var->data.stream = 1u << 31;
       unpacked_var->insert_before(packed_var);
       this->packed_varyings[slot] = packed_var;
    } else {
       /* For geometry shader inputs, only update the packed variable name the
        * first time we visit each component.
        */
       if (this->gs_input_vertices == 0 || vertex_index == 0) {
          ir_variable *var = this->packed_varyings[slot];
 
          if (var->is_name_ralloced())
-- 
2.7.4