pan/bi: Implement load_const

In the laziest possible way...  We can just emit worst case moves which
DCE will eat for breakfast anyway, and inline constants on instructions
where that is supported directly. This approach eliminates a lot of
nasty corner cases that Midgard's crazy cache scheme has hit, at the
expense of slightly more work for DCE (but it's only a single iteration
of an O(N) pass that has to run anyway..)

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4097>

diff --git a/src/panfrost/bifrost/bifrost_compile.c b/src/panfrost/bifrost/bifrost_compile.c
index 3aedcd063fc87f64d3326dc6222a5673d5c3392d..a0b9d03b1eef488d9547d657276bbc3014b8884f 100644 (file)
--- a/src/panfrost/bifrost/bifrost_compile.c
+++ b/src/panfrost/bifrost/bifrost_compile.c
@@ -221,15 +221,34 @@ emit_intrinsic(bi_context *ctx, nir_intrinsic_instr *instr)
         }
 }
 
+static void
+emit_load_const(bi_context *ctx, nir_load_const_instr *instr)
+{
+        /* Make sure we've been lowered */
+        assert(instr->def.num_components == 1);
+
+        bi_instruction move = {
+                .type = BI_MOV,
+                .dest = bir_ssa_index(&instr->def),
+                .dest_type = instr->def.bit_size | nir_type_uint,
+                .src = {
+                        BIR_INDEX_CONSTANT
+                },
+                .constant = {
+                        .u64 = nir_const_value_as_uint(instr->value[0], instr->def.bit_size)
+                }
+        };
+
+        bi_emit(ctx, move);
+}
+
 static void
 emit_instr(bi_context *ctx, struct nir_instr *instr)
 {
         switch (instr->type) {
-#if 0
         case nir_instr_type_load_const:
                 emit_load_const(ctx, nir_instr_as_load_const(instr));
                 break;
-#endif
 
         case nir_instr_type_intrinsic:
                 emit_intrinsic(ctx, nir_instr_as_intrinsic(instr));
@@ -494,6 +513,8 @@ bi_optimize_nir(nir_shader *nir)
         };
 
         NIR_PASS(progress, nir, nir_lower_tex, &lower_tex_options);
+        NIR_PASS(progress, nir, nir_lower_alu_to_scalar, NULL, NULL);
+        NIR_PASS(progress, nir, nir_lower_load_const_to_scalar);
 
         do {
                 progress = false;
@@ -538,6 +559,8 @@ bi_optimize_nir(nir_shader *nir)
         } while (progress);
 
         NIR_PASS(progress, nir, nir_opt_algebraic_late);
+        NIR_PASS(progress, nir, nir_lower_alu_to_scalar, NULL, NULL);
+        NIR_PASS(progress, nir, nir_lower_load_const_to_scalar);
 
         /* Take us out of SSA */
         NIR_PASS(progress, nir, nir_lower_locals_to_regs);
@@ -571,10 +594,6 @@ bifrost_compile_shader_nir(nir_shader *nir, bifrost_program *program, unsigned p
         NIR_PASS_V(nir, nir_lower_io, nir_var_all, glsl_type_size, 0);
         NIR_PASS_V(nir, nir_lower_ssbo);
 
-        /* We have to lower ALU to scalar ourselves since viewport
-         * transformations produce vector ops */
-        NIR_PASS_V(nir, nir_lower_alu_to_scalar, NULL, NULL);
-
         bi_optimize_nir(nir);
         nir_print_shader(nir, stdout);