src/amd/common/ac_binary.c

   1 /*
   2  * Copyright 2014 Advanced Micro Devices, Inc.
   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 FROM,
  20  * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
  21  * SOFTWARE.
  22  */
  23
  24 #include "ac_binary.h"
  25
  26 #include "util/u_math.h"
  27 #include "util/u_memory.h"
  28
  29 #include <gelf.h>
  30 #include <libelf.h>
  31 #include <stdio.h>
  32
  33 #include <sid.h>
  34
  35 #define SPILLED_SGPRS                                     0x4
  36 #define SPILLED_VGPRS                                     0x8
  37
  38 /* Parse configuration data in .AMDGPU.config section format. */
  39 void ac_parse_shader_binary_config(const char *data, size_t nbytes,
  40                                    unsigned wave_size,
  41                                    bool really_needs_scratch,
  42                                    struct ac_shader_config *conf)
  43 {
  44         uint32_t scratch_size = 0;
  45
  46         for (size_t i = 0; i < nbytes; i += 8) {
  47                 unsigned reg = util_le32_to_cpu(*(uint32_t*)(data + i));
  48                 unsigned value = util_le32_to_cpu(*(uint32_t*)(data + i + 4));
  49                 switch (reg) {
  50                 case R_00B028_SPI_SHADER_PGM_RSRC1_PS:
  51                 case R_00B128_SPI_SHADER_PGM_RSRC1_VS:
  52                 case R_00B228_SPI_SHADER_PGM_RSRC1_GS:
  53                 case R_00B848_COMPUTE_PGM_RSRC1:
  54                 case R_00B428_SPI_SHADER_PGM_RSRC1_HS:
  55                         if (wave_size == 32)
  56                                 conf->num_vgprs = MAX2(conf->num_vgprs, (G_00B028_VGPRS(value) + 1) * 8);
  57                         else
  58                                 conf->num_vgprs = MAX2(conf->num_vgprs, (G_00B028_VGPRS(value) + 1) * 4);
  59
  60                         conf->num_sgprs = MAX2(conf->num_sgprs, (G_00B028_SGPRS(value) + 1) * 8);
  61                         conf->float_mode =  G_00B028_FLOAT_MODE(value);
  62                         conf->rsrc1 = value;
  63                         break;
  64                 case R_00B02C_SPI_SHADER_PGM_RSRC2_PS:
  65                         conf->lds_size = MAX2(conf->lds_size, G_00B02C_EXTRA_LDS_SIZE(value));
  66                         break;
  67                 case R_00B84C_COMPUTE_PGM_RSRC2:
  68                         conf->lds_size = MAX2(conf->lds_size, G_00B84C_LDS_SIZE(value));
  69                         conf->rsrc2 = value;
  70                         break;
  71                 case R_0286CC_SPI_PS_INPUT_ENA:
  72                         conf->spi_ps_input_ena = value;
  73                         break;
  74                 case R_0286D0_SPI_PS_INPUT_ADDR:
  75                         conf->spi_ps_input_addr = value;
  76                         break;
  77                 case R_0286E8_SPI_TMPRING_SIZE:
  78                 case R_00B860_COMPUTE_TMPRING_SIZE:
  79                         /* WAVESIZE is in units of 256 dwords. */
  80                         scratch_size = value;
  81                         break;
  82                 case SPILLED_SGPRS:
  83                         conf->spilled_sgprs = value;
  84                         break;
  85                 case SPILLED_VGPRS:
  86                         conf->spilled_vgprs = value;
  87                         break;
  88                 default:
  89                         {
  90                                 static bool printed;
  91
  92                                 if (!printed) {
  93                                         fprintf(stderr, "Warning: LLVM emitted unknown "
  94                                                 "config register: 0x%x\n", reg);
  95                                         printed = true;
  96                                 }
  97                         }
  98                         break;
  99                 }
 100         }
 101
 102         if (!conf->spi_ps_input_addr)
 103                 conf->spi_ps_input_addr = conf->spi_ps_input_ena;
 104
 105         if (really_needs_scratch) {
 106                 /* sgprs spills aren't spilling */
 107                 conf->scratch_bytes_per_wave = G_00B860_WAVESIZE(scratch_size) * 256 * 4;
 108         }
 109 }