nir: Support deref instructions in loop_analyze

Reviewed-by: Caio Marcelo de Oliveira Filho <[email protected]>
Acked-by: Rob Clark <[email protected]>
Acked-by: Bas Nieuwenhuizen <[email protected]>
Acked-by: Dave Airlie <[email protected]>
Reviewed-by: Kenneth Graunke <[email protected]>

1 files changed, 47 insertions, 4 deletions

diff --git a/src/compiler/nir/nir_loop_analyze.c b/src/compiler/nir/nir_loop_analyze.c
index 870b4ffa5f0..bc0c44ab1bb 100644
--- a/src/compiler/nir/nir_loop_analyze.c
+++ b/src/compiler/nir/nir_loop_analyze.c
@@ -623,8 +623,8 @@ find_trip_count(loop_info_state *state)
  * due to array access heuristics.
  */
 static bool
-force_unroll_array_access(loop_info_state *state, nir_shader *ns,
-                          nir_deref_var *variable)
+force_unroll_array_access_var(loop_info_state *state, nir_shader *ns,
+                              nir_deref_var *variable)
 {
    nir_deref *tail = &variable->deref;
 
@@ -665,6 +665,37 @@ force_unroll_array_access(loop_info_state *state, nir_shader *ns,
 }
 
 static bool
+force_unroll_array_access(loop_info_state *state, nir_shader *ns,
+                          nir_deref_instr *deref)
+{
+   for (nir_deref_instr *d = deref; d; d = nir_deref_instr_parent(d)) {
+      if (d->deref_type != nir_deref_type_array)
+         continue;
+
+      assert(d->arr.index.is_ssa);
+      nir_loop_variable *array_index = get_loop_var(d->arr.index.ssa, state);
+
+      if (array_index->type != basic_induction)
+         continue;
+
+      nir_deref_instr *parent = nir_deref_instr_parent(d);
+      assert(glsl_type_is_array(parent->type) ||
+             glsl_type_is_matrix(parent->type));
+      if (glsl_get_length(parent->type) == state->loop->info->trip_count) {
+         state->loop->info->force_unroll = true;
+         return true;
+      }
+
+      if (deref->mode & state->indirect_mask) {
+         state->loop->info->force_unroll = true;
+         return true;
+      }
+   }
+
+   return false;
+}
+
+static bool
 force_unroll_heuristics(loop_info_state *state, nir_shader *ns,
                         nir_block *block)
 {
@@ -683,10 +714,23 @@ force_unroll_heuristics(loop_info_state *state, nir_shader *ns,
          unsigned num_vars =
             nir_intrinsic_infos[intrin->intrinsic].num_variables;
          for (unsigned i = 0; i < num_vars; i++) {
-            if (force_unroll_array_access(state, ns, intrin->variables[i]))
+            if (force_unroll_array_access_var(state, ns, intrin->variables[i]))
                return true;
          }
       }
+
+      if (intrin->intrinsic == nir_intrinsic_load_deref ||
+          intrin->intrinsic == nir_intrinsic_store_deref ||
+          intrin->intrinsic == nir_intrinsic_copy_deref) {
+         if (force_unroll_array_access(state, ns,
+                                       nir_src_as_deref(intrin->src[0])))
+            return true;
+
+         if (intrin->intrinsic == nir_intrinsic_copy_deref &&
+             force_unroll_array_access(state, ns,
+                                       nir_src_as_deref(intrin->src[1])))
+            return true;
+      }
    }
 
    return false;
@@ -827,7 +871,6 @@ void
 nir_loop_analyze_impl(nir_function_impl *impl,
                       nir_variable_mode indirect_mask)
 {
-   nir_assert_lowered_derefs(impl->function->shader, nir_lower_load_store_derefs);
    nir_index_ssa_defs(impl);
    foreach_list_typed(nir_cf_node, node, node, &impl->body)
       process_loops(node, indirect_mask);