nir/lower_idiv: add new llvm-based path

[mesa.git] / src / freedreno / ir3 / ir3_nir.c

diff --git a/src/freedreno/ir3/ir3_nir.c b/src/freedreno/ir3/ir3_nir.c
index 05426bc7a0ec9fa551d42bcd9433dc7ba0015951..99659a7ddef40253e3d671f069e84e189f3c8065 100644 (file)
--- a/src/freedreno/ir3/ir3_nir.c
+++ b/src/freedreno/ir3/ir3_nir.c
@@ -56,6 +56,8 @@ static const nir_shader_compiler_options options = {
                .lower_bitfield_extract_to_shifts = true,
                .use_interpolated_input_intrinsics = true,
                .lower_rotate = true,
+               .lower_to_scalar = true,
+               .has_imul24 = true,
 };
 
 /* we don't want to lower vertex_id to _zero_based on newer gpus: */
@@ -81,6 +83,9 @@ static const nir_shader_compiler_options options_a6xx = {
                .lower_bitfield_extract_to_shifts = true,
                .use_interpolated_input_intrinsics = true,
                .lower_rotate = true,
+               .vectorize_io = true,
+               .lower_to_scalar = true,
+               .has_imul24 = true,
 };
 
 const nir_shader_compiler_options *
@@ -98,7 +103,8 @@ ir3_key_lowers_nir(const struct ir3_shader_key *key)
        return key->fsaturate_s | key->fsaturate_t | key->fsaturate_r |
                        key->vsaturate_s | key->vsaturate_t | key->vsaturate_r |
                        key->ucp_enables | key->color_two_side |
-                       key->fclamp_color | key->vclamp_color;
+                       key->fclamp_color | key->vclamp_color |
+                       key->has_gs;
 }
 
 #define OPT(nir, pass, ...) ({                             \
@@ -124,7 +130,7 @@ ir3_optimize_loop(nir_shader *s)
                OPT_V(s, nir_lower_vars_to_ssa);
                progress |= OPT(s, nir_opt_copy_prop_vars);
                progress |= OPT(s, nir_opt_dead_write_vars);
-               progress |= OPT(s, nir_lower_alu_to_scalar, NULL);
+               progress |= OPT(s, nir_lower_alu_to_scalar, NULL, NULL);
                progress |= OPT(s, nir_lower_phis_to_scalar);
 
                progress |= OPT(s, nir_copy_prop);
@@ -183,6 +189,19 @@ ir3_optimize_nir(struct ir3_shader *shader, nir_shader *s,
                        .lower_tg4_offsets = true,
        };
 
+       if (key && key->has_gs) {
+               switch (shader->type) {
+               case MESA_SHADER_VERTEX:
+                       NIR_PASS_V(s, ir3_nir_lower_vs_to_explicit_io, shader);
+                       break;
+               case MESA_SHADER_GEOMETRY:
+                       NIR_PASS_V(s, ir3_nir_lower_gs, shader);
+                       break;
+               default:
+                       break;
+               }
+       }
+
        if (key) {
                switch (shader->type) {
                case MESA_SHADER_FRAGMENT:
@@ -220,11 +239,11 @@ ir3_optimize_nir(struct ir3_shader *shader, nir_shader *s,
 
        if (key) {
                if (s->info.stage == MESA_SHADER_VERTEX) {
-                       OPT_V(s, nir_lower_clip_vs, key->ucp_enables, false);
+                       OPT_V(s, nir_lower_clip_vs, key->ucp_enables, false, false, NULL);
                        if (key->vclamp_color)
                                OPT_V(s, nir_lower_clamp_color_outputs);
                } else if (s->info.stage == MESA_SHADER_FRAGMENT) {
-                       OPT_V(s, nir_lower_clip_fs, key->ucp_enables);
+                       OPT_V(s, nir_lower_clip_fs, key->ucp_enables, false);
                        if (key->fclamp_color)
                                OPT_V(s, nir_lower_clamp_color_outputs);
                }
@@ -258,10 +277,24 @@ ir3_optimize_nir(struct ir3_shader *shader, nir_shader *s,
         * NOTE that UBO analysis pass should only be done once, before variants
         */
        const bool ubo_progress = !key && OPT(s, ir3_nir_analyze_ubo_ranges, shader);
-       const bool idiv_progress = OPT(s, nir_lower_idiv);
+       const bool idiv_progress = OPT(s, nir_lower_idiv, nir_lower_idiv_fast);
        if (ubo_progress || idiv_progress)
                ir3_optimize_loop(s);
 
+       /* Do late algebraic optimization to turn add(a, neg(b)) back into
+       * subs, then the mandatory cleanup after algebraic.  Note that it may
+       * produce fnegs, and if so then we need to keep running to squash
+       * fneg(fneg(a)).
+       */
+       bool more_late_algebraic = true;
+       while (more_late_algebraic) {
+               more_late_algebraic = OPT(s, nir_opt_algebraic_late);
+               OPT_V(s, nir_opt_constant_folding);
+               OPT_V(s, nir_copy_prop);
+               OPT_V(s, nir_opt_dce);
+               OPT_V(s, nir_opt_cse);
+       }
+
        OPT_V(s, nir_remove_dead_variables, nir_var_function_temp);
 
        OPT_V(s, nir_opt_sink, nir_move_const_undef);
@@ -414,5 +447,19 @@ ir3_setup_const_state(struct ir3_shader *shader, nir_shader *nir)
                constoff += align(IR3_MAX_SO_BUFFERS * ptrsz, 4) / 4;
        }
 
+       switch (shader->type) {
+       case MESA_SHADER_VERTEX:
+               const_state->offsets.primitive_param = constoff;
+               constoff += 1;
+               break;
+       case MESA_SHADER_GEOMETRY:
+               const_state->offsets.primitive_param = constoff;
+               const_state->offsets.primitive_map = constoff + 1;
+               constoff += 1 + DIV_ROUND_UP(nir->num_inputs, 4);
+               break;
+       default:
+               break;
+       }
+
        const_state->offsets.immediate = constoff;
 }