src/gallium/drivers/vc4/vc4_opt_coalesce_ff_writes.c

   1 /*
   2  * Copyright © 2014 Broadcom
   3  *
   4  * Permission is hereby granted, free of charge, to any person obtaining a
   5  * copy of this software and associated documentation files (the "Software"),
   6  * to deal in the Software without restriction, including without limitation
   7  * the rights to use, copy, modify, merge, publish, distribute, sublicense,
   8  * and/or sell copies of the Software, and to permit persons to whom the
   9  * Software is furnished to do so, subject to the following conditions:
  10  *
  11  * The above copyright notice and this permission notice (including the next
  12  * paragraph) shall be included in all copies or substantial portions of the
  13  * Software.
  14  *
  15  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
  16  * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  17  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
  18  * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
  19  * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
  20  * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
  21  * IN THE SOFTWARE.
  22  */
  23
  24 /**
  25  * @file vc4_opt_coalesce_ff_writes.c
  26  *
  27  * This modifies instructions that generate the value consumed by a VPM write
  28  * to write directly into the VPM.
  29  */
  30
  31 #include "vc4_qir.h"
  32
  33 bool
  34 qir_opt_coalesce_ff_writes(struct vc4_compile *c)
  35 {
  36         if (c->stage == QSTAGE_FRAG)
  37                 return false;
  38
  39         /* For now, only do this pass when we don't have control flow. */
  40         struct qblock *block = qir_entry_block(c);
  41         if (block != qir_exit_block(c))
  42                 return false;
  43
  44         bool progress = false;
  45         struct qinst *vpm_writes[64] = { 0 };
  46         uint32_t use_count[c->num_temps];
  47         uint32_t vpm_write_count = 0;
  48         memset(&use_count, 0, sizeof(use_count));
  49
  50         qir_for_each_inst_inorder(inst, c) {
  51                 switch (inst->dst.file) {
  52                 case QFILE_VPM:
  53                         vpm_writes[vpm_write_count++] = inst;
  54                         break;
  55                 default:
  56                         break;
  57                 }
  58
  59                 for (int i = 0; i < qir_get_nsrc(inst); i++) {
  60                         if (inst->src[i].file == QFILE_TEMP) {
  61                                 uint32_t temp = inst->src[i].index;
  62                                 use_count[temp]++;
  63                         }
  64                 }
  65         }
  66
  67         for (int i = 0; i < vpm_write_count; i++) {
  68                 if (!qir_is_raw_mov(vpm_writes[i]) ||
  69                     vpm_writes[i]->src[0].file != QFILE_TEMP) {
  70                         continue;
  71                 }
  72
  73                 uint32_t temp = vpm_writes[i]->src[0].index;
  74                 if (use_count[temp] != 1)
  75                         continue;
  76
  77                 struct qinst *inst = c->defs[temp];
  78                 if (!inst)
  79                         continue;
  80
  81                 if (qir_depends_on_flags(inst) || inst->sf)
  82                         continue;
  83
  84                 if (qir_has_side_effects(c, inst) ||
  85                     qir_has_side_effect_reads(c, inst)) {
  86                         continue;
  87                 }
  88
  89                 /* Move the generating instruction to the end of the program
  90                  * to maintain the order of the VPM writes.
  91                  */
  92                 assert(!vpm_writes[i]->sf);
  93                 list_del(&inst->link);
  94                 list_addtail(&inst->link, &vpm_writes[i]->link);
  95                 qir_remove_instruction(c, vpm_writes[i]);
  96
  97                 c->defs[inst->dst.index] = NULL;
  98                 inst->dst.file = QFILE_VPM;
  99                 inst->dst.index = 0;
 100
 101                 progress = true;
 102         }
 103
 104         return progress;
 105 }