mesa/src/compiler/nir/nir_opt_constant_folding.c

/*
 * Copyright © 2014 Intel Corporation
 *
 * Permission is hereby granted, free of charge, to any person obtaining a
 * copy of this software and associated documentation files (the "Software"),
 * to deal in the Software without restriction, including without limitation
 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
 * and/or sell copies of the Software, and to permit persons to whom the
 * Software is furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice (including the next
 * paragraph) shall be included in all copies or substantial portions of the
 * Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
 * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
 * IN THE SOFTWARE.
 *
 * Authors:
 *    Jason Ekstrand (jason@jlekstrand.net)
 *
 */

#include "nir.h"
#include "nir_builder.h"
#include "nir_constant_expressions.h"
#include <math.h>

/*
 * Implements SSA-based constant folding.
 */

struct constant_fold_state {
   bool has_load_constant;
   bool has_indirect_load_const;
};

static bool
try_fold_alu(nir_builder *b, nir_alu_instr *instr)
{
   nir_const_value src[NIR_MAX_VEC_COMPONENTS][NIR_MAX_VEC_COMPONENTS];

   if (!instr->dest.dest.is_ssa)
      return false;

   /* In the case that any outputs/inputs have unsized types, then we need to
    * guess the bit-size. In this case, the validator ensures that all
    * bit-sizes match so we can just take the bit-size from first
    * output/input with an unsized type. If all the outputs/inputs are sized
    * then we don't need to guess the bit-size at all because the code we
    * generate for constant opcodes in this case already knows the sizes of
    * the types involved and does not need the provided bit-size for anything
    * (although it still requires to receive a valid bit-size).
    */
   unsigned bit_size = 0;
   if (!nir_alu_type_get_type_size(nir_op_infos[instr->op].output_type))
      bit_size = instr->dest.dest.ssa.bit_size;

   for (unsigned i = 0; i < nir_op_infos[instr->op].num_inputs; i++) {
      if (!instr->src[i].src.is_ssa)
         return false;

      if (bit_size == 0 &&
          !nir_alu_type_get_type_size(nir_op_infos[instr->op].input_types[i]))
         bit_size = instr->src[i].src.ssa->bit_size;

      nir_instr *src_instr = instr->src[i].src.ssa->parent_instr;

      if (src_instr->type != nir_instr_type_load_const)
         return false;
      nir_load_const_instr* load_const = nir_instr_as_load_const(src_instr);

      for (unsigned j = 0; j < nir_ssa_alu_instr_src_components(instr, i);
           j++) {
         src[i][j] = load_const->value[instr->src[i].swizzle[j]];
      }

      /* We shouldn't have any source modifiers in the optimization loop. */
      assert(!instr->src[i].abs && !instr->src[i].negate);
   }

   if (bit_size == 0)
      bit_size = 32;

   /* We shouldn't have any saturate modifiers in the optimization loop. */
   assert(!instr->dest.saturate);

   nir_const_value dest[NIR_MAX_VEC_COMPONENTS];
   nir_const_value *srcs[NIR_MAX_VEC_COMPONENTS];
   memset(dest, 0, sizeof(dest));
   for (unsigned i = 0; i < nir_op_infos[instr->op].num_inputs; ++i)
      srcs[i] = src[i];
   nir_eval_const_opcode(instr->op, dest, instr->dest.dest.ssa.num_components,
                         bit_size, srcs,
                         b->shader->info.float_controls_execution_mode);

   b->cursor = nir_before_instr(&instr->instr);
   nir_ssa_def *imm = nir_build_imm(b, instr->dest.dest.ssa.num_components,
                                       instr->dest.dest.ssa.bit_size,
                                       dest);
   nir_ssa_def_rewrite_uses(&instr->dest.dest.ssa, nir_src_for_ssa(imm));
   nir_instr_remove(&instr->instr);

   ralloc_free(instr);

   return true;
}

static bool
try_fold_intrinsic(nir_builder *b, nir_intrinsic_instr *instr,
                   struct constant_fold_state *state)
{
   bool progress = false;

   if ((instr->intrinsic == nir_intrinsic_demote_if ||
        instr->intrinsic == nir_intrinsic_discard_if) &&
       nir_src_is_const(instr->src[0])) {
      if (nir_src_as_bool(instr->src[0])) {
         b->cursor = nir_before_instr(&instr->instr);
         nir_intrinsic_op op = instr->intrinsic == nir_intrinsic_discard_if ?
                               nir_intrinsic_discard :
                               nir_intrinsic_demote;
         nir_intrinsic_instr *new_instr =
            nir_intrinsic_instr_create(b->shader, op);
         nir_builder_instr_insert(b, &new_instr->instr);
      }
      nir_instr_remove(&instr->instr);
      progress = true;
   } else if (instr->intrinsic == nir_intrinsic_load_constant) {
      state->has_load_constant = true;

      if (!nir_src_is_const(instr->src[0])) {
         state->has_indirect_load_const = true;
         return progress;
      }

      unsigned offset = nir_src_as_uint(instr->src[0]);
      unsigned base = nir_intrinsic_base(instr);
      unsigned range = nir_intrinsic_range(instr);
      assert(base + range <= b->shader->constant_data_size);

      b->cursor = nir_before_instr(&instr->instr);
      nir_ssa_def *val;
      if (offset >= range) {
         val = nir_ssa_undef(b, instr->dest.ssa.num_components,
                                instr->dest.ssa.bit_size);
      } else {
         nir_const_value imm[NIR_MAX_VEC_COMPONENTS];
         uint8_t *data = (uint8_t*)b->shader->constant_data + base;
         for (unsigned i = 0; i < instr->num_components; i++) {
            unsigned bytes = instr->dest.ssa.bit_size / 8;
            bytes = MIN2(bytes, range - offset);

            memcpy(&imm[i].u64, data + offset, bytes);
            offset += bytes;
         }
         val = nir_build_imm(b, instr->dest.ssa.num_components,
                                instr->dest.ssa.bit_size, imm);
      }
      nir_ssa_def_rewrite_uses(&instr->dest.ssa, nir_src_for_ssa(val));
      nir_instr_remove(&instr->instr);
      progress = true;
   }

   return progress;
}

static bool
try_fold_instr(nir_builder *b, nir_instr *instr, void *_state)
{
   switch (instr->type) {
   case nir_instr_type_alu:
      return try_fold_alu(b, nir_instr_as_alu(instr));
   case nir_instr_type_intrinsic:
      return try_fold_intrinsic(b, nir_instr_as_intrinsic(instr), _state);
   default:
      /* Don't know how to constant fold */
      return false;
   }
}

bool
nir_opt_constant_folding(nir_shader *shader)
{
   struct constant_fold_state state;
   state.has_load_constant = false;
   state.has_indirect_load_const = false;

   bool progress = nir_shader_instructions_pass(shader, try_fold_instr,
                                                nir_metadata_block_index |
                                                nir_metadata_dominance,
                                                &state);

   /* This doesn't free the constant data if there are no constant loads because
    * the data might still be used but the loads have been lowered to load_ubo
    */
   if (state.has_load_constant && !state.has_indirect_load_const &&
       shader->constant_data_size) {
      ralloc_free(shader->constant_data);
      shader->constant_data = NULL;
      shader->constant_data_size = 0;
   }

   return progress;
}