diff options
Diffstat (limited to 'drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c')
| -rw-r--r-- | drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c | 2101 |
1 files changed, 867 insertions, 1234 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c index 72793f93e2fc..bb8709066fd8 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c @@ -31,8 +31,6 @@ #include "amdgpu_ucode.h" #include "clearstate_ci.h" -#include "uvd/uvd_4_2_d.h" - #include "dce/dce_8_0_d.h" #include "dce/dce_8_0_sh_mask.h" @@ -1006,9 +1004,15 @@ out: */ static void gfx_v7_0_tiling_mode_table_init(struct amdgpu_device *adev) { - const u32 num_tile_mode_states = 32; - const u32 num_secondary_tile_mode_states = 16; - u32 reg_offset, gb_tile_moden, split_equal_to_row_size; + const u32 num_tile_mode_states = + ARRAY_SIZE(adev->gfx.config.tile_mode_array); + const u32 num_secondary_tile_mode_states = + ARRAY_SIZE(adev->gfx.config.macrotile_mode_array); + u32 reg_offset, split_equal_to_row_size; + uint32_t *tile, *macrotile; + + tile = adev->gfx.config.tile_mode_array; + macrotile = adev->gfx.config.macrotile_mode_array; switch (adev->gfx.config.mem_row_size_in_kb) { case 1: @@ -1023,832 +1027,531 @@ static void gfx_v7_0_tiling_mode_table_init(struct amdgpu_device *adev) break; } + for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) + tile[reg_offset] = 0; + for (reg_offset = 0; reg_offset < num_secondary_tile_mode_states; reg_offset++) + macrotile[reg_offset] = 0; + switch (adev->asic_type) { case CHIP_BONAIRE: - for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) { - switch (reg_offset) { - case 0: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); - break; - case 1: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); - break; - case 2: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); - break; - case 3: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); - break; - case 4: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | - TILE_SPLIT(split_equal_to_row_size)); - break; - case 5: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); - break; - case 6: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | - TILE_SPLIT(split_equal_to_row_size)); - break; - case 7: - gb_tile_moden = (TILE_SPLIT(split_equal_to_row_size)); - break; - - case 8: - gb_tile_moden = (ARRAY_MODE(ARRAY_LINEAR_ALIGNED) | - PIPE_CONFIG(ADDR_SURF_P4_16x16)); - break; - case 9: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING)); - break; - case 10: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); - break; - case 11: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); - break; - case 12: - gb_tile_moden = (TILE_SPLIT(split_equal_to_row_size)); - break; - case 13: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING)); - break; - case 14: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); - break; - case 15: - gb_tile_moden = (ARRAY_MODE(ARRAY_3D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); - break; - case 16: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); - break; - case 17: - gb_tile_moden = (TILE_SPLIT(split_equal_to_row_size)); - break; - case 18: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 19: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING)); - break; - case 20: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 21: - gb_tile_moden = (ARRAY_MODE(ARRAY_3D_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 22: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 23: - gb_tile_moden = (TILE_SPLIT(split_equal_to_row_size)); - break; - case 24: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 25: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_XTHICK) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 26: - gb_tile_moden = (ARRAY_MODE(ARRAY_3D_TILED_XTHICK) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 27: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING)); - break; - case 28: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); - break; - case 29: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); - break; - case 30: - gb_tile_moden = (TILE_SPLIT(split_equal_to_row_size)); - break; - default: - gb_tile_moden = 0; - break; - } - adev->gfx.config.tile_mode_array[reg_offset] = gb_tile_moden; - WREG32(mmGB_TILE_MODE0 + reg_offset, gb_tile_moden); - } - for (reg_offset = 0; reg_offset < num_secondary_tile_mode_states; reg_offset++) { - switch (reg_offset) { - case 0: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 1: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 2: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 3: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 4: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 5: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_8_BANK)); - break; - case 6: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | - NUM_BANKS(ADDR_SURF_4_BANK)); - break; - case 8: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_8) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 9: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 10: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 11: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 12: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 13: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_8_BANK)); - break; - case 14: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | - NUM_BANKS(ADDR_SURF_4_BANK)); - break; - default: - gb_tile_moden = 0; - break; - } - adev->gfx.config.macrotile_mode_array[reg_offset] = gb_tile_moden; - WREG32(mmGB_MACROTILE_MODE0 + reg_offset, gb_tile_moden); - } + tile[0] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); + tile[1] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); + tile[2] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); + tile[3] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); + tile[4] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | + TILE_SPLIT(split_equal_to_row_size)); + tile[5] = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); + tile[6] = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | + TILE_SPLIT(split_equal_to_row_size)); + tile[7] = (TILE_SPLIT(split_equal_to_row_size)); + tile[8] = (ARRAY_MODE(ARRAY_LINEAR_ALIGNED) | + PIPE_CONFIG(ADDR_SURF_P4_16x16)); + tile[9] = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING)); + tile[10] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + tile[11] = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); + tile[12] = (TILE_SPLIT(split_equal_to_row_size)); + tile[13] = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING)); + tile[14] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + tile[15] = (ARRAY_MODE(ARRAY_3D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + tile[16] = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); + tile[17] = (TILE_SPLIT(split_equal_to_row_size)); + tile[18] = (ARRAY_MODE(ARRAY_1D_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[19] = (ARRAY_MODE(ARRAY_1D_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING)); + tile[20] = (ARRAY_MODE(ARRAY_2D_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[21] = (ARRAY_MODE(ARRAY_3D_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[22] = (ARRAY_MODE(ARRAY_PRT_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[23] = (TILE_SPLIT(split_equal_to_row_size)); + tile[24] = (ARRAY_MODE(ARRAY_2D_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[25] = (ARRAY_MODE(ARRAY_2D_TILED_XTHICK) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[26] = (ARRAY_MODE(ARRAY_3D_TILED_XTHICK) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[27] = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING)); + tile[28] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + tile[29] = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); + tile[30] = (TILE_SPLIT(split_equal_to_row_size)); + + macrotile[0] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[1] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[2] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[3] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[4] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[5] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_8_BANK)); + macrotile[6] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | + NUM_BANKS(ADDR_SURF_4_BANK)); + macrotile[8] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_8) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[9] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[10] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[11] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[12] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[13] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_8_BANK)); + macrotile[14] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | + NUM_BANKS(ADDR_SURF_4_BANK)); + + for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) + WREG32(mmGB_TILE_MODE0 + reg_offset, tile[reg_offset]); + for (reg_offset = 0; reg_offset < num_secondary_tile_mode_states; reg_offset++) + if (reg_offset != 7) + WREG32(mmGB_MACROTILE_MODE0 + reg_offset, macrotile[reg_offset]); break; case CHIP_HAWAII: - for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) { - switch (reg_offset) { - case 0: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); - break; - case 1: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); - break; - case 2: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); - break; - case 3: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); - break; - case 4: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | - TILE_SPLIT(split_equal_to_row_size)); - break; - case 5: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | - TILE_SPLIT(split_equal_to_row_size)); - break; - case 6: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | - TILE_SPLIT(split_equal_to_row_size)); - break; - case 7: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | - TILE_SPLIT(split_equal_to_row_size)); - break; - - case 8: - gb_tile_moden = (ARRAY_MODE(ARRAY_LINEAR_ALIGNED) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16)); - break; - case 9: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING)); - break; - case 10: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); - break; - case 11: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); - break; - case 12: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); - break; - case 13: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING)); - break; - case 14: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); - break; - case 15: - gb_tile_moden = (ARRAY_MODE(ARRAY_3D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); - break; - case 16: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); - break; - case 17: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); - break; - case 18: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 19: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING)); - break; - case 20: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 21: - gb_tile_moden = (ARRAY_MODE(ARRAY_3D_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 22: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 23: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 24: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 25: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_XTHICK) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 26: - gb_tile_moden = (ARRAY_MODE(ARRAY_3D_TILED_XTHICK) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 27: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING)); - break; - case 28: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); - break; - case 29: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); - break; - case 30: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); - break; - default: - gb_tile_moden = 0; - break; - } - adev->gfx.config.tile_mode_array[reg_offset] = gb_tile_moden; - WREG32(mmGB_TILE_MODE0 + reg_offset, gb_tile_moden); - } - for (reg_offset = 0; reg_offset < num_secondary_tile_mode_states; reg_offset++) { - switch (reg_offset) { - case 0: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 1: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 2: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 3: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 4: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | - NUM_BANKS(ADDR_SURF_8_BANK)); - break; - case 5: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | - NUM_BANKS(ADDR_SURF_4_BANK)); - break; - case 6: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | - NUM_BANKS(ADDR_SURF_4_BANK)); - break; - case 8: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 9: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 10: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 11: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | - NUM_BANKS(ADDR_SURF_8_BANK)); - break; - case 12: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 13: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_8_BANK)); - break; - case 14: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | - NUM_BANKS(ADDR_SURF_4_BANK)); - break; - default: - gb_tile_moden = 0; - break; - } - adev->gfx.config.macrotile_mode_array[reg_offset] = gb_tile_moden; - WREG32(mmGB_MACROTILE_MODE0 + reg_offset, gb_tile_moden); - } + tile[0] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); + tile[1] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); + tile[2] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); + tile[3] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); + tile[4] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | + TILE_SPLIT(split_equal_to_row_size)); + tile[5] = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | + TILE_SPLIT(split_equal_to_row_size)); + tile[6] = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | + TILE_SPLIT(split_equal_to_row_size)); + tile[7] = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | + TILE_SPLIT(split_equal_to_row_size)); + tile[8] = (ARRAY_MODE(ARRAY_LINEAR_ALIGNED) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16)); + tile[9] = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING)); + tile[10] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + tile[11] = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); + tile[12] = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); + tile[13] = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING)); + tile[14] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + tile[15] = (ARRAY_MODE(ARRAY_3D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + tile[16] = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); + tile[17] = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); + tile[18] = (ARRAY_MODE(ARRAY_1D_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[19] = (ARRAY_MODE(ARRAY_1D_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING)); + tile[20] = (ARRAY_MODE(ARRAY_2D_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[21] = (ARRAY_MODE(ARRAY_3D_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[22] = (ARRAY_MODE(ARRAY_PRT_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[23] = (ARRAY_MODE(ARRAY_PRT_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[24] = (ARRAY_MODE(ARRAY_2D_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[25] = (ARRAY_MODE(ARRAY_2D_TILED_XTHICK) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[26] = (ARRAY_MODE(ARRAY_3D_TILED_XTHICK) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[27] = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING)); + tile[28] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + tile[29] = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); + tile[30] = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); + + macrotile[0] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[1] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[2] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[3] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[4] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | + NUM_BANKS(ADDR_SURF_8_BANK)); + macrotile[5] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | + NUM_BANKS(ADDR_SURF_4_BANK)); + macrotile[6] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | + NUM_BANKS(ADDR_SURF_4_BANK)); + macrotile[8] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[9] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[10] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[11] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | + NUM_BANKS(ADDR_SURF_8_BANK)); + macrotile[12] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[13] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_8_BANK)); + macrotile[14] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | + NUM_BANKS(ADDR_SURF_4_BANK)); + + for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) + WREG32(mmGB_TILE_MODE0 + reg_offset, tile[reg_offset]); + for (reg_offset = 0; reg_offset < num_secondary_tile_mode_states; reg_offset++) + if (reg_offset != 7) + WREG32(mmGB_MACROTILE_MODE0 + reg_offset, macrotile[reg_offset]); break; case CHIP_KABINI: case CHIP_KAVERI: case CHIP_MULLINS: default: - for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) { - switch (reg_offset) { - case 0: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); - break; - case 1: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); - break; - case 2: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); - break; - case 3: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); - break; - case 4: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | - TILE_SPLIT(split_equal_to_row_size)); - break; - case 5: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); - break; - case 6: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | - TILE_SPLIT(split_equal_to_row_size)); - break; - case 7: - gb_tile_moden = (TILE_SPLIT(split_equal_to_row_size)); - break; - - case 8: - gb_tile_moden = (ARRAY_MODE(ARRAY_LINEAR_ALIGNED) | - PIPE_CONFIG(ADDR_SURF_P2)); - break; - case 9: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING)); - break; - case 10: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); - break; - case 11: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); - break; - case 12: - gb_tile_moden = (TILE_SPLIT(split_equal_to_row_size)); - break; - case 13: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING)); - break; - case 14: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); - break; - case 15: - gb_tile_moden = (ARRAY_MODE(ARRAY_3D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); - break; - case 16: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); - break; - case 17: - gb_tile_moden = (TILE_SPLIT(split_equal_to_row_size)); - break; - case 18: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 19: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING)); - break; - case 20: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 21: - gb_tile_moden = (ARRAY_MODE(ARRAY_3D_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 22: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 23: - gb_tile_moden = (TILE_SPLIT(split_equal_to_row_size)); - break; - case 24: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 25: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_XTHICK) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 26: - gb_tile_moden = (ARRAY_MODE(ARRAY_3D_TILED_XTHICK) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 27: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING)); - break; - case 28: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); - break; - case 29: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); - break; - case 30: - gb_tile_moden = (TILE_SPLIT(split_equal_to_row_size)); - break; - default: - gb_tile_moden = 0; - break; - } - adev->gfx.config.tile_mode_array[reg_offset] = gb_tile_moden; - WREG32(mmGB_TILE_MODE0 + reg_offset, gb_tile_moden); - } - for (reg_offset = 0; reg_offset < num_secondary_tile_mode_states; reg_offset++) { - switch (reg_offset) { - case 0: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | - NUM_BANKS(ADDR_SURF_8_BANK)); - break; - case 1: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | - NUM_BANKS(ADDR_SURF_8_BANK)); - break; - case 2: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_8_BANK)); - break; - case 3: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_8_BANK)); - break; - case 4: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_8_BANK)); - break; - case 5: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_8_BANK)); - break; - case 6: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_8_BANK)); - break; - case 8: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_4) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_8) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 9: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_4) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 10: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 11: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 12: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 13: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 14: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_8_BANK)); - break; - default: - gb_tile_moden = 0; - break; - } - adev->gfx.config.macrotile_mode_array[reg_offset] = gb_tile_moden; - WREG32(mmGB_MACROTILE_MODE0 + reg_offset, gb_tile_moden); - } + tile[0] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); + tile[1] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); + tile[2] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); + tile[3] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); + tile[4] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | + TILE_SPLIT(split_equal_to_row_size)); + tile[5] = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); + tile[6] = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | + TILE_SPLIT(split_equal_to_row_size)); + tile[7] = (TILE_SPLIT(split_equal_to_row_size)); + tile[8] = (ARRAY_MODE(ARRAY_LINEAR_ALIGNED) | + PIPE_CONFIG(ADDR_SURF_P2)); + tile[9] = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING)); + tile[10] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + tile[11] = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); + tile[12] = (TILE_SPLIT(split_equal_to_row_size)); + tile[13] = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING)); + tile[14] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + tile[15] = (ARRAY_MODE(ARRAY_3D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + tile[16] = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); + tile[17] = (TILE_SPLIT(split_equal_to_row_size)); + tile[18] = (ARRAY_MODE(ARRAY_1D_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[19] = (ARRAY_MODE(ARRAY_1D_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING)); + tile[20] = (ARRAY_MODE(ARRAY_2D_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[21] = (ARRAY_MODE(ARRAY_3D_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[22] = (ARRAY_MODE(ARRAY_PRT_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[23] = (TILE_SPLIT(split_equal_to_row_size)); + tile[24] = (ARRAY_MODE(ARRAY_2D_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[25] = (ARRAY_MODE(ARRAY_2D_TILED_XTHICK) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[26] = (ARRAY_MODE(ARRAY_3D_TILED_XTHICK) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[27] = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING)); + tile[28] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + tile[29] = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); + tile[30] = (TILE_SPLIT(split_equal_to_row_size)); + + macrotile[0] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | + NUM_BANKS(ADDR_SURF_8_BANK)); + macrotile[1] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | + NUM_BANKS(ADDR_SURF_8_BANK)); + macrotile[2] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_8_BANK)); + macrotile[3] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_8_BANK)); + macrotile[4] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_8_BANK)); + macrotile[5] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_8_BANK)); + macrotile[6] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_8_BANK)); + macrotile[8] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_4) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_8) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[9] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_4) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[10] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[11] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[12] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[13] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[14] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_8_BANK)); + + for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) + WREG32(mmGB_TILE_MODE0 + reg_offset, tile[reg_offset]); + for (reg_offset = 0; reg_offset < num_secondary_tile_mode_states; reg_offset++) + if (reg_offset != 7) + WREG32(mmGB_MACROTILE_MODE0 + reg_offset, macrotile[reg_offset]); break; } } @@ -1893,45 +1596,31 @@ void gfx_v7_0_select_se_sh(struct amdgpu_device *adev, u32 se_num, u32 sh_num) */ static u32 gfx_v7_0_create_bitmask(u32 bit_width) { - u32 i, mask = 0; - - for (i = 0; i < bit_width; i++) { - mask <<= 1; - mask |= 1; - } - return mask; + return (u32)((1ULL << bit_width) - 1); } /** - * gfx_v7_0_get_rb_disabled - computes the mask of disabled RBs + * gfx_v7_0_get_rb_active_bitmap - computes the mask of enabled RBs * * @adev: amdgpu_device pointer - * @max_rb_num: max RBs (render backends) for the asic - * @se_num: number of SEs (shader engines) for the asic - * @sh_per_se: number of SH blocks per SE for the asic * - * Calculates the bitmask of disabled RBs (CIK). - * Returns the disabled RB bitmask. + * Calculates the bitmask of enabled RBs (CIK). + * Returns the enabled RB bitmask. */ -static u32 gfx_v7_0_get_rb_disabled(struct amdgpu_device *adev, - u32 max_rb_num_per_se, - u32 sh_per_se) +static u32 gfx_v7_0_get_rb_active_bitmap(struct amdgpu_device *adev) { u32 data, mask; data = RREG32(mmCC_RB_BACKEND_DISABLE); - if (data & 1) - data &= CC_RB_BACKEND_DISABLE__BACKEND_DISABLE_MASK; - else - data = 0; - data |= RREG32(mmGC_USER_RB_BACKEND_DISABLE); + data &= CC_RB_BACKEND_DISABLE__BACKEND_DISABLE_MASK; data >>= GC_USER_RB_BACKEND_DISABLE__BACKEND_DISABLE__SHIFT; - mask = gfx_v7_0_create_bitmask(max_rb_num_per_se / sh_per_se); + mask = gfx_v7_0_create_bitmask(adev->gfx.config.max_backends_per_se / + adev->gfx.config.max_sh_per_se); - return data & mask; + return (~data) & mask; } /** @@ -1940,73 +1629,31 @@ static u32 gfx_v7_0_get_rb_disabled(struct amdgpu_device *adev, * @adev: amdgpu_device pointer * @se_num: number of SEs (shader engines) for the asic * @sh_per_se: number of SH blocks per SE for the asic - * @max_rb_num: max RBs (render backends) for the asic * * Configures per-SE/SH RB registers (CIK). */ -static void gfx_v7_0_setup_rb(struct amdgpu_device *adev, - u32 se_num, u32 sh_per_se, - u32 max_rb_num_per_se) +static void gfx_v7_0_setup_rb(struct amdgpu_device *adev) { int i, j; - u32 data, mask; - u32 disabled_rbs = 0; - u32 enabled_rbs = 0; + u32 data; + u32 active_rbs = 0; + u32 rb_bitmap_width_per_sh = adev->gfx.config.max_backends_per_se / + adev->gfx.config.max_sh_per_se; mutex_lock(&adev->grbm_idx_mutex); - for (i = 0; i < se_num; i++) { - for (j = 0; j < sh_per_se; j++) { + for (i = 0; i < adev->gfx.config.max_shader_engines; i++) { + for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) { gfx_v7_0_select_se_sh(adev, i, j); - data = gfx_v7_0_get_rb_disabled(adev, max_rb_num_per_se, sh_per_se); - if (adev->asic_type == CHIP_HAWAII) - disabled_rbs |= data << ((i * sh_per_se + j) * HAWAII_RB_BITMAP_WIDTH_PER_SH); - else - disabled_rbs |= data << ((i * sh_per_se + j) * CIK_RB_BITMAP_WIDTH_PER_SH); + data = gfx_v7_0_get_rb_active_bitmap(adev); + active_rbs |= data << ((i * adev->gfx.config.max_sh_per_se + j) * + rb_bitmap_width_per_sh); } } gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff); mutex_unlock(&adev->grbm_idx_mutex); - mask = 1; - for (i = 0; i < max_rb_num_per_se * se_num; i++) { - if (!(disabled_rbs & mask)) - enabled_rbs |= mask; - mask <<= 1; - } - - adev->gfx.config.backend_enable_mask = enabled_rbs; - - mutex_lock(&adev->grbm_idx_mutex); - for (i = 0; i < se_num; i++) { - gfx_v7_0_select_se_sh(adev, i, 0xffffffff); - data = 0; - for (j = 0; j < sh_per_se; j++) { - switch (enabled_rbs & 3) { - case 0: - if (j == 0) - data |= (RASTER_CONFIG_RB_MAP_3 << - PA_SC_RASTER_CONFIG__PKR_MAP__SHIFT); - else - data |= (RASTER_CONFIG_RB_MAP_0 << - PA_SC_RASTER_CONFIG__PKR_MAP__SHIFT); - break; - case 1: - data |= (RASTER_CONFIG_RB_MAP_0 << (i * sh_per_se + j) * 2); - break; - case 2: - data |= (RASTER_CONFIG_RB_MAP_3 << (i * sh_per_se + j) * 2); - break; - case 3: - default: - data |= (RASTER_CONFIG_RB_MAP_2 << (i * sh_per_se + j) * 2); - break; - } - enabled_rbs >>= 2; - } - WREG32(mmPA_SC_RASTER_CONFIG, data); - } - gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff); - mutex_unlock(&adev->grbm_idx_mutex); + adev->gfx.config.backend_enable_mask = active_rbs; + adev->gfx.config.num_rbs = hweight32(active_rbs); } /** @@ -2059,192 +1706,23 @@ static void gmc_v7_0_init_compute_vmid(struct amdgpu_device *adev) */ static void gfx_v7_0_gpu_init(struct amdgpu_device *adev) { - u32 gb_addr_config; - u32 mc_shared_chmap, mc_arb_ramcfg; - u32 dimm00_addr_map, dimm01_addr_map, dimm10_addr_map, dimm11_addr_map; - u32 sh_mem_cfg; - u32 tmp; + u32 tmp, sh_mem_cfg; int i; - switch (adev->asic_type) { - case CHIP_BONAIRE: - adev->gfx.config.max_shader_engines = 2; - adev->gfx.config.max_tile_pipes = 4; - adev->gfx.config.max_cu_per_sh = 7; - adev->gfx.config.max_sh_per_se = 1; - adev->gfx.config.max_backends_per_se = 2; - adev->gfx.config.max_texture_channel_caches = 4; - adev->gfx.config.max_gprs = 256; - adev->gfx.config.max_gs_threads = 32; - adev->gfx.config.max_hw_contexts = 8; - - adev->gfx.config.sc_prim_fifo_size_frontend = 0x20; - adev->gfx.config.sc_prim_fifo_size_backend = 0x100; - adev->gfx.config.sc_hiz_tile_fifo_size = 0x30; - adev->gfx.config.sc_earlyz_tile_fifo_size = 0x130; - gb_addr_config = BONAIRE_GB_ADDR_CONFIG_GOLDEN; - break; - case CHIP_HAWAII: - adev->gfx.config.max_shader_engines = 4; - adev->gfx.config.max_tile_pipes = 16; - adev->gfx.config.max_cu_per_sh = 11; - adev->gfx.config.max_sh_per_se = 1; - adev->gfx.config.max_backends_per_se = 4; - adev->gfx.config.max_texture_channel_caches = 16; - adev->gfx.config.max_gprs = 256; - adev->gfx.config.max_gs_threads = 32; - adev->gfx.config.max_hw_contexts = 8; - - adev->gfx.config.sc_prim_fifo_size_frontend = 0x20; - adev->gfx.config.sc_prim_fifo_size_backend = 0x100; - adev->gfx.config.sc_hiz_tile_fifo_size = 0x30; - adev->gfx.config.sc_earlyz_tile_fifo_size = 0x130; - gb_addr_config = HAWAII_GB_ADDR_CONFIG_GOLDEN; - break; - case CHIP_KAVERI: - adev->gfx.config.max_shader_engines = 1; - adev->gfx.config.max_tile_pipes = 4; - if ((adev->pdev->device == 0x1304) || - (adev->pdev->device == 0x1305) || - (adev->pdev->device == 0x130C) || - (adev->pdev->device == 0x130F) || - (adev->pdev->device == 0x1310) || - (adev->pdev->device == 0x1311) || - (adev->pdev->device == 0x131C)) { - adev->gfx.config.max_cu_per_sh = 8; - adev->gfx.config.max_backends_per_se = 2; - } else if ((adev->pdev->device == 0x1309) || - (adev->pdev->device == 0x130A) || - (adev->pdev->device == 0x130D) || - (adev->pdev->device == 0x1313) || - (adev->pdev->device == 0x131D)) { - adev->gfx.config.max_cu_per_sh = 6; - adev->gfx.config.max_backends_per_se = 2; - } else if ((adev->pdev->device == 0x1306) || - (adev->pdev->device == 0x1307) || - (adev->pdev->device == 0x130B) || - (adev->pdev->device == 0x130E) || - (adev->pdev->device == 0x1315) || - (adev->pdev->device == 0x131B)) { - adev->gfx.config.max_cu_per_sh = 4; - adev->gfx.config.max_backends_per_se = 1; - } else { - adev->gfx.config.max_cu_per_sh = 3; - adev->gfx.config.max_backends_per_se = 1; - } - adev->gfx.config.max_sh_per_se = 1; - adev->gfx.config.max_texture_channel_caches = 4; - adev->gfx.config.max_gprs = 256; - adev->gfx.config.max_gs_threads = 16; - adev->gfx.config.max_hw_contexts = 8; - - adev->gfx.config.sc_prim_fifo_size_frontend = 0x20; - adev->gfx.config.sc_prim_fifo_size_backend = 0x100; - adev->gfx.config.sc_hiz_tile_fifo_size = 0x30; - adev->gfx.config.sc_earlyz_tile_fifo_size = 0x130; - gb_addr_config = BONAIRE_GB_ADDR_CONFIG_GOLDEN; - break; - case CHIP_KABINI: - case CHIP_MULLINS: - default: - adev->gfx.config.max_shader_engines = 1; - adev->gfx.config.max_tile_pipes = 2; - adev->gfx.config.max_cu_per_sh = 2; - adev->gfx.config.max_sh_per_se = 1; - adev->gfx.config.max_backends_per_se = 1; - adev->gfx.config.max_texture_channel_caches = 2; - adev->gfx.config.max_gprs = 256; - adev->gfx.config.max_gs_threads = 16; - adev->gfx.config.max_hw_contexts = 8; - - adev->gfx.config.sc_prim_fifo_size_frontend = 0x20; - adev->gfx.config.sc_prim_fifo_size_backend = 0x100; - adev->gfx.config.sc_hiz_tile_fifo_size = 0x30; - adev->gfx.config.sc_earlyz_tile_fifo_size = 0x130; - gb_addr_config = BONAIRE_GB_ADDR_CONFIG_GOLDEN; - break; - } - WREG32(mmGRBM_CNTL, (0xff << GRBM_CNTL__READ_TIMEOUT__SHIFT)); - mc_shared_chmap = RREG32(mmMC_SHARED_CHMAP); - adev->gfx.config.mc_arb_ramcfg = RREG32(mmMC_ARB_RAMCFG); - mc_arb_ramcfg = adev->gfx.config.mc_arb_ramcfg; - - adev->gfx.config.num_tile_pipes = adev->gfx.config.max_tile_pipes; - adev->gfx.config.mem_max_burst_length_bytes = 256; - if (adev->flags & AMD_IS_APU) { - /* Get memory bank mapping mode. */ - tmp = RREG32(mmMC_FUS_DRAM0_BANK_ADDR_MAPPING); - dimm00_addr_map = REG_GET_FIELD(tmp, MC_FUS_DRAM0_BANK_ADDR_MAPPING, DIMM0ADDRMAP); - dimm01_addr_map = REG_GET_FIELD(tmp, MC_FUS_DRAM0_BANK_ADDR_MAPPING, DIMM1ADDRMAP); - - tmp = RREG32(mmMC_FUS_DRAM1_BANK_ADDR_MAPPING); - dimm10_addr_map = REG_GET_FIELD(tmp, MC_FUS_DRAM1_BANK_ADDR_MAPPING, DIMM0ADDRMAP); - dimm11_addr_map = REG_GET_FIELD(tmp, MC_FUS_DRAM1_BANK_ADDR_MAPPING, DIMM1ADDRMAP); - - /* Validate settings in case only one DIMM installed. */ - if ((dimm00_addr_map == 0) || (dimm00_addr_map == 3) || (dimm00_addr_map == 4) || (dimm00_addr_map > 12)) - dimm00_addr_map = 0; - if ((dimm01_addr_map == 0) || (dimm01_addr_map == 3) || (dimm01_addr_map == 4) || (dimm01_addr_map > 12)) - dimm01_addr_map = 0; - if ((dimm10_addr_map == 0) || (dimm10_addr_map == 3) || (dimm10_addr_map == 4) || (dimm10_addr_map > 12)) - dimm10_addr_map = 0; - if ((dimm11_addr_map == 0) || (dimm11_addr_map == 3) || (dimm11_addr_map == 4) || (dimm11_addr_map > 12)) - dimm11_addr_map = 0; - - /* If DIMM Addr map is 8GB, ROW size should be 2KB. Otherwise 1KB. */ - /* If ROW size(DIMM1) != ROW size(DMIMM0), ROW size should be larger one. */ - if ((dimm00_addr_map == 11) || (dimm01_addr_map == 11) || (dimm10_addr_map == 11) || (dimm11_addr_map == 11)) - adev->gfx.config.mem_row_size_in_kb = 2; - else - adev->gfx.config.mem_row_size_in_kb = 1; - } else { - tmp = (mc_arb_ramcfg & MC_ARB_RAMCFG__NOOFCOLS_MASK) >> MC_ARB_RAMCFG__NOOFCOLS__SHIFT; - adev->gfx.config.mem_row_size_in_kb = (4 * (1 << (8 + tmp))) / 1024; - if (adev->gfx.config.mem_row_size_in_kb > 4) - adev->gfx.config.mem_row_size_in_kb = 4; - } - /* XXX use MC settings? */ - adev->gfx.config.shader_engine_tile_size = 32; - adev->gfx.config.num_gpus = 1; - adev->gfx.config.multi_gpu_tile_size = 64; - - /* fix up row size */ - gb_addr_config &= ~GB_ADDR_CONFIG__ROW_SIZE_MASK; - switch (adev->gfx.config.mem_row_size_in_kb) { - case 1: - default: - gb_addr_config |= (0 << GB_ADDR_CONFIG__ROW_SIZE__SHIFT); - break; - case 2: - gb_addr_config |= (1 << GB_ADDR_CONFIG__ROW_SIZE__SHIFT); - break; - case 4: - gb_addr_config |= (2 << GB_ADDR_CONFIG__ROW_SIZE__SHIFT); - break; - } - adev->gfx.config.gb_addr_config = gb_addr_config; - - WREG32(mmGB_ADDR_CONFIG, gb_addr_config); - WREG32(mmHDP_ADDR_CONFIG, gb_addr_config); - WREG32(mmDMIF_ADDR_CALC, gb_addr_config); - WREG32(mmSDMA0_TILING_CONFIG + SDMA0_REGISTER_OFFSET, gb_addr_config & 0x70); - WREG32(mmSDMA0_TILING_CONFIG + SDMA1_REGISTER_OFFSET, gb_addr_config & 0x70); - WREG32(mmUVD_UDEC_ADDR_CONFIG, gb_addr_config); - WREG32(mmUVD_UDEC_DB_ADDR_CONFIG, gb_addr_config); - WREG32(mmUVD_UDEC_DBW_ADDR_CONFIG, gb_addr_config); + WREG32(mmGB_ADDR_CONFIG, adev->gfx.config.gb_addr_config); + WREG32(mmHDP_ADDR_CONFIG, adev->gfx.config.gb_addr_config); + WREG32(mmDMIF_ADDR_CALC, adev->gfx.config.gb_addr_config); gfx_v7_0_tiling_mode_table_init(adev); - gfx_v7_0_setup_rb(adev, adev->gfx.config.max_shader_engines, - adev->gfx.config.max_sh_per_se, - adev->gfx.config.max_backends_per_se); + gfx_v7_0_setup_rb(adev); /* set HW defaults for 3D engine */ WREG32(mmCP_MEQ_THRESHOLDS, - (0x30 << CP_MEQ_THRESHOLDS__MEQ1_START__SHIFT) | - (0x60 << CP_MEQ_THRESHOLDS__MEQ2_START__SHIFT)); + (0x30 << CP_MEQ_THRESHOLDS__MEQ1_START__SHIFT) | + (0x60 << CP_MEQ_THRESHOLDS__MEQ2_START__SHIFT)); mutex_lock(&adev->grbm_idx_mutex); /* @@ -2255,7 +1733,7 @@ static void gfx_v7_0_gpu_init(struct amdgpu_device *adev) /* XXX SH_MEM regs */ /* where to put LDS, scratch, GPUVM in FSA64 space */ - sh_mem_cfg = REG_SET_FIELD(0, SH_MEM_CONFIG, ALIGNMENT_MODE, + sh_mem_cfg = REG_SET_FIELD(0, SH_MEM_CONFIG, ALIGNMENT_MODE, SH_MEM_ALIGNMENT_MODE_UNALIGNED); mutex_lock(&adev->srbm_mutex); @@ -2379,7 +1857,7 @@ static int gfx_v7_0_ring_test_ring(struct amdgpu_ring *ring) return r; } WREG32(scratch, 0xCAFEDEAD); - r = amdgpu_ring_lock(ring, 3); + r = amdgpu_ring_alloc(ring, 3); if (r) { DRM_ERROR("amdgpu: cp failed to lock ring %d (%d).\n", ring->idx, r); amdgpu_gfx_scratch_free(adev, scratch); @@ -2388,7 +1866,7 @@ static int gfx_v7_0_ring_test_ring(struct amdgpu_ring *ring) amdgpu_ring_write(ring, PACKET3(PACKET3_SET_UCONFIG_REG, 1)); amdgpu_ring_write(ring, (scratch - PACKET3_SET_UCONFIG_REG_START)); amdgpu_ring_write(ring, 0xDEADBEEF); - amdgpu_ring_unlock_commit(ring); + amdgpu_ring_commit(ring); for (i = 0; i < adev->usec_timeout; i++) { tmp = RREG32(scratch); @@ -2447,6 +1925,25 @@ static void gfx_v7_0_ring_emit_hdp_flush(struct amdgpu_ring *ring) } /** + * gfx_v7_0_ring_emit_hdp_invalidate - emit an hdp invalidate on the cp + * + * @adev: amdgpu_device pointer + * @ridx: amdgpu ring index + * + * Emits an hdp invalidate on the cp. + */ +static void gfx_v7_0_ring_emit_hdp_invalidate(struct amdgpu_ring *ring) +{ + amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3)); + amdgpu_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) | + WRITE_DATA_DST_SEL(0) | + WR_CONFIRM)); + amdgpu_ring_write(ring, mmHDP_DEBUG0); + amdgpu_ring_write(ring, 0); + amdgpu_ring_write(ring, 1); +} + +/** * gfx_v7_0_ring_emit_fence_gfx - emit a fence on the gfx ring * * @adev: amdgpu_device pointer @@ -2516,36 +2013,6 @@ static void gfx_v7_0_ring_emit_fence_compute(struct amdgpu_ring *ring, amdgpu_ring_write(ring, upper_32_bits(seq)); } -/** - * gfx_v7_0_ring_emit_semaphore - emit a semaphore on the CP ring - * - * @ring: amdgpu ring buffer object - * @semaphore: amdgpu semaphore object - * @emit_wait: Is this a sempahore wait? - * - * Emits a semaphore signal/wait packet to the CP ring and prevents the PFP - * from running ahead of semaphore waits. - */ -static bool gfx_v7_0_ring_emit_semaphore(struct amdgpu_ring *ring, - struct amdgpu_semaphore *semaphore, - bool emit_wait) -{ - uint64_t addr = semaphore->gpu_addr; - unsigned sel = emit_wait ? PACKET3_SEM_SEL_WAIT : PACKET3_SEM_SEL_SIGNAL; - - amdgpu_ring_write(ring, PACKET3(PACKET3_MEM_SEMAPHORE, 1)); - amdgpu_ring_write(ring, addr & 0xffffffff); - amdgpu_ring_write(ring, (upper_32_bits(addr) & 0xffff) | sel); - - if (emit_wait && (ring->type == AMDGPU_RING_TYPE_GFX)) { - /* Prevent the PFP from running ahead of the semaphore wait */ - amdgpu_ring_write(ring, PACKET3(PACKET3_PFP_SYNC_ME, 0)); - amdgpu_ring_write(ring, 0x0); - } - - return true; -} - /* * IB stuff */ @@ -2593,8 +2060,7 @@ static void gfx_v7_0_ring_emit_ib_gfx(struct amdgpu_ring *ring, else header = PACKET3(PACKET3_INDIRECT_BUFFER, 2); - control |= ib->length_dw | - (ib->vm ? (ib->vm->ids[ring->idx].id << 24) : 0); + control |= ib->length_dw | (ib->vm_id << 24); amdgpu_ring_write(ring, header); amdgpu_ring_write(ring, @@ -2622,8 +2088,7 @@ static void gfx_v7_0_ring_emit_ib_compute(struct amdgpu_ring *ring, header = PACKET3(PACKET3_INDIRECT_BUFFER, 2); - control |= ib->length_dw | - (ib->vm ? (ib->vm->ids[ring->idx].id << 24) : 0); + control |= ib->length_dw | (ib->vm_id << 24); amdgpu_ring_write(ring, header); amdgpu_ring_write(ring, @@ -2661,7 +2126,7 @@ static int gfx_v7_0_ring_test_ib(struct amdgpu_ring *ring) } WREG32(scratch, 0xCAFEDEAD); memset(&ib, 0, sizeof(ib)); - r = amdgpu_ib_get(ring, NULL, 256, &ib); + r = amdgpu_ib_get(adev, NULL, 256, &ib); if (r) { DRM_ERROR("amdgpu: failed to get ib (%d).\n", r); goto err1; @@ -2671,9 +2136,7 @@ static int gfx_v7_0_ring_test_ib(struct amdgpu_ring *ring) ib.ptr[2] = 0xDEADBEEF; ib.length_dw = 3; - r = amdgpu_sched_ib_submit_kernel_helper(adev, ring, &ib, 1, NULL, - AMDGPU_FENCE_OWNER_UNDEFINED, - &f); + r = amdgpu_ib_schedule(ring, 1, &ib, NULL, &f); if (r) goto err2; @@ -2700,7 +2163,8 @@ static int gfx_v7_0_ring_test_ib(struct amdgpu_ring *ring) err2: fence_put(f); - amdgpu_ib_free(adev, &ib); + amdgpu_ib_free(adev, &ib, NULL); + fence_put(f); err1: amdgpu_gfx_scratch_free(adev, scratch); return r; @@ -2842,7 +2306,7 @@ static int gfx_v7_0_cp_gfx_start(struct amdgpu_device *adev) gfx_v7_0_cp_gfx_enable(adev, true); - r = amdgpu_ring_lock(ring, gfx_v7_0_get_csb_size(adev) + 8); + r = amdgpu_ring_alloc(ring, gfx_v7_0_get_csb_size(adev) + 8); if (r) { DRM_ERROR("amdgpu: cp failed to lock ring (%d).\n", r); return r; @@ -2911,7 +2375,7 @@ static int gfx_v7_0_cp_gfx_start(struct amdgpu_device *adev) amdgpu_ring_write(ring, 0x0000000e); /* VGT_VERTEX_REUSE_BLOCK_CNTL */ amdgpu_ring_write(ring, 0x00000010); /* VGT_OUT_DEALLOC_CNTL */ - amdgpu_ring_unlock_commit(ring); + amdgpu_ring_commit(ring); return 0; } @@ -2989,21 +2453,14 @@ static int gfx_v7_0_cp_gfx_resume(struct amdgpu_device *adev) static u32 gfx_v7_0_ring_get_rptr_gfx(struct amdgpu_ring *ring) { - u32 rptr; - - rptr = ring->adev->wb.wb[ring->rptr_offs]; - - return rptr; + return ring->adev->wb.wb[ring->rptr_offs]; } static u32 gfx_v7_0_ring_get_wptr_gfx(struct amdgpu_ring *ring) { struct amdgpu_device *adev = ring->adev; - u32 wptr; - wptr = RREG32(mmCP_RB0_WPTR); - - return wptr; + return RREG32(mmCP_RB0_WPTR); } static void gfx_v7_0_ring_set_wptr_gfx(struct amdgpu_ring *ring) @@ -3016,21 +2473,13 @@ static void gfx_v7_0_ring_set_wptr_gfx(struct amdgpu_ring *ring) static u32 gfx_v7_0_ring_get_rptr_compute(struct amdgpu_ring *ring) { - u32 rptr; - - rptr = ring->adev->wb.wb[ring->rptr_offs]; - - return rptr; + return ring->adev->wb.wb[ring->rptr_offs]; } static u32 gfx_v7_0_ring_get_wptr_compute(struct amdgpu_ring *ring) { - u32 wptr; - /* XXX check if swapping is necessary on BE */ - wptr = ring->adev->wb.wb[ring->wptr_offs]; - - return wptr; + return ring->adev->wb.wb[ring->wptr_offs]; } static void gfx_v7_0_ring_set_wptr_compute(struct amdgpu_ring *ring) @@ -3126,21 +2575,6 @@ static int gfx_v7_0_cp_compute_load_microcode(struct amdgpu_device *adev) } /** - * gfx_v7_0_cp_compute_start - start the compute queues - * - * @adev: amdgpu_device pointer - * - * Enable the compute queues. - * Returns 0 for success, error for failure. - */ -static int gfx_v7_0_cp_compute_start(struct amdgpu_device *adev) -{ - gfx_v7_0_cp_compute_enable(adev, true); - - return 0; -} - -/** * gfx_v7_0_cp_compute_fini - stop the compute queues * * @adev: amdgpu_device pointer @@ -3330,9 +2764,7 @@ static int gfx_v7_0_cp_compute_resume(struct amdgpu_device *adev) u32 *buf; struct bonaire_mqd *mqd; - r = gfx_v7_0_cp_compute_start(adev); - if (r) - return r; + gfx_v7_0_cp_compute_enable(adev, true); /* fix up chicken bits */ tmp = RREG32(mmCP_CPF_DEBUG); @@ -3610,6 +3042,26 @@ static int gfx_v7_0_cp_resume(struct amdgpu_device *adev) return 0; } +/** + * gfx_v7_0_ring_emit_vm_flush - cik vm flush using the CP + * + * @ring: the ring to emmit the commands to + * + * Sync the command pipeline with the PFP. E.g. wait for everything + * to be completed. + */ +static void gfx_v7_0_ring_emit_pipeline_sync(struct amdgpu_ring *ring) +{ + int usepfp = (ring->type == AMDGPU_RING_TYPE_GFX); + if (usepfp) { + /* synce CE with ME to prevent CE fetch CEIB before context switch done */ + amdgpu_ring_write(ring, PACKET3(PACKET3_SWITCH_BUFFER, 0)); + amdgpu_ring_write(ring, 0); + amdgpu_ring_write(ring, PACKET3(PACKET3_SWITCH_BUFFER, 0)); + amdgpu_ring_write(ring, 0); + } +} + /* * vm * VMID 0 is the physical GPU addresses as used by the kernel. @@ -3628,13 +3080,18 @@ static void gfx_v7_0_ring_emit_vm_flush(struct amdgpu_ring *ring, unsigned vm_id, uint64_t pd_addr) { int usepfp = (ring->type == AMDGPU_RING_TYPE_GFX); - if (usepfp) { - /* synce CE with ME to prevent CE fetch CEIB before context switch done */ - amdgpu_ring_write(ring, PACKET3(PACKET3_SWITCH_BUFFER, 0)); - amdgpu_ring_write(ring, 0); - amdgpu_ring_write(ring, PACKET3(PACKET3_SWITCH_BUFFER, 0)); - amdgpu_ring_write(ring, 0); - } + uint32_t seq = ring->fence_drv.sync_seq; + uint64_t addr = ring->fence_drv.gpu_addr; + + amdgpu_ring_write(ring, PACKET3(PACKET3_WAIT_REG_MEM, 5)); + amdgpu_ring_write(ring, (WAIT_REG_MEM_MEM_SPACE(1) | /* memory */ + WAIT_REG_MEM_FUNCTION(3) | /* equal */ + WAIT_REG_MEM_ENGINE(usepfp))); /* pfp or me */ + amdgpu_ring_write(ring, addr & 0xfffffffc); + amdgpu_ring_write(ring, upper_32_bits(addr) & 0xffffffff); + amdgpu_ring_write(ring, seq); + amdgpu_ring_write(ring, 0xffffffff); + amdgpu_ring_write(ring, 4); /* poll interval */ amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3)); amdgpu_ring_write(ring, (WRITE_DATA_ENGINE_SEL(usepfp) | @@ -4109,7 +3566,7 @@ static void gfx_v7_0_enable_cgcg(struct amdgpu_device *adev, bool enable) orig = data = RREG32(mmRLC_CGCG_CGLS_CTRL); - if (enable && (adev->cg_flags & AMDGPU_CG_SUPPORT_GFX_CGCG)) { + if (enable && (adev->cg_flags & AMD_CG_SUPPORT_GFX_CGCG)) { gfx_v7_0_enable_gui_idle_interrupt(adev, true); tmp = gfx_v7_0_halt_rlc(adev); @@ -4147,9 +3604,9 @@ static void gfx_v7_0_enable_mgcg(struct amdgpu_device *adev, bool enable) { u32 data, orig, tmp = 0; - if (enable && (adev->cg_flags & AMDGPU_CG_SUPPORT_GFX_MGCG)) { - if (adev->cg_flags & AMDGPU_CG_SUPPORT_GFX_MGLS) { - if (adev->cg_flags & AMDGPU_CG_SUPPORT_GFX_CP_LS) { + if (enable && (adev->cg_flags & AMD_CG_SUPPORT_GFX_MGCG)) { + if (adev->cg_flags & AMD_CG_SUPPORT_GFX_MGLS) { + if (adev->cg_flags & AMD_CG_SUPPORT_GFX_CP_LS) { orig = data = RREG32(mmCP_MEM_SLP_CNTL); data |= CP_MEM_SLP_CNTL__CP_MEM_LS_EN_MASK; if (orig != data) @@ -4176,14 +3633,14 @@ static void gfx_v7_0_enable_mgcg(struct amdgpu_device *adev, bool enable) gfx_v7_0_update_rlc(adev, tmp); - if (adev->cg_flags & AMDGPU_CG_SUPPORT_GFX_CGTS) { + if (adev->cg_flags & AMD_CG_SUPPORT_GFX_CGTS) { orig = data = RREG32(mmCGTS_SM_CTRL_REG); data &= ~CGTS_SM_CTRL_REG__SM_MODE_MASK; data |= (0x2 << CGTS_SM_CTRL_REG__SM_MODE__SHIFT); data |= CGTS_SM_CTRL_REG__SM_MODE_ENABLE_MASK; data &= ~CGTS_SM_CTRL_REG__OVERRIDE_MASK; - if ((adev->cg_flags & AMDGPU_CG_SUPPORT_GFX_MGLS) && - (adev->cg_flags & AMDGPU_CG_SUPPORT_GFX_CGTS_LS)) + if ((adev->cg_flags & AMD_CG_SUPPORT_GFX_MGLS) && + (adev->cg_flags & AMD_CG_SUPPORT_GFX_CGTS_LS)) data &= ~CGTS_SM_CTRL_REG__LS_OVERRIDE_MASK; data &= ~CGTS_SM_CTRL_REG__ON_MONITOR_ADD_MASK; data |= CGTS_SM_CTRL_REG__ON_MONITOR_ADD_EN_MASK; @@ -4249,7 +3706,7 @@ static void gfx_v7_0_enable_sclk_slowdown_on_pu(struct amdgpu_device *adev, u32 data, orig; orig = data = RREG32(mmRLC_PG_CNTL); - if (enable && (adev->pg_flags & AMDGPU_PG_SUPPORT_RLC_SMU_HS)) + if (enable && (adev->pg_flags & AMD_PG_SUPPORT_RLC_SMU_HS)) data |= RLC_PG_CNTL__SMU_CLK_SLOWDOWN_ON_PU_ENABLE_MASK; else data &= ~RLC_PG_CNTL__SMU_CLK_SLOWDOWN_ON_PU_ENABLE_MASK; @@ -4263,7 +3720,7 @@ static void gfx_v7_0_enable_sclk_slowdown_on_pd(struct amdgpu_device *adev, u32 data, orig; orig = data = RREG32(mmRLC_PG_CNTL); - if (enable && (adev->pg_flags & AMDGPU_PG_SUPPORT_RLC_SMU_HS)) + if (enable && (adev->pg_flags & AMD_PG_SUPPORT_RLC_SMU_HS)) data |= RLC_PG_CNTL__SMU_CLK_SLOWDOWN_ON_PD_ENABLE_MASK; else data &= ~RLC_PG_CNTL__SMU_CLK_SLOWDOWN_ON_PD_ENABLE_MASK; @@ -4276,7 +3733,7 @@ static void gfx_v7_0_enable_cp_pg(struct amdgpu_device *adev, bool enable) u32 data, orig; orig = data = RREG32(mmRLC_PG_CNTL); - if (enable && (adev->pg_flags & AMDGPU_PG_SUPPORT_CP)) + if (enable && (adev->pg_flags & AMD_PG_SUPPORT_CP)) data &= ~0x8000; else data |= 0x8000; @@ -4289,7 +3746,7 @@ static void gfx_v7_0_enable_gds_pg(struct amdgpu_device *adev, bool enable) u32 data, orig; orig = data = RREG32(mmRLC_PG_CNTL); - if (enable && (adev->pg_flags & AMDGPU_PG_SUPPORT_GDS)) + if (enable && (adev->pg_flags & AMD_PG_SUPPORT_GDS)) data &= ~0x2000; else data |= 0x2000; @@ -4370,7 +3827,7 @@ static void gfx_v7_0_enable_gfx_cgpg(struct amdgpu_device *adev, { u32 data, orig; - if (enable && (adev->pg_flags & AMDGPU_PG_SUPPORT_GFX_PG)) { + if (enable && (adev->pg_flags & AMD_PG_SUPPORT_GFX_PG)) { orig = data = RREG32(mmRLC_PG_CNTL); data |= RLC_PG_CNTL__GFX_POWER_GATING_ENABLE_MASK; if (orig != data) @@ -4395,28 +3852,19 @@ static void gfx_v7_0_enable_gfx_cgpg(struct amdgpu_device *adev, } } -static u32 gfx_v7_0_get_cu_active_bitmap(struct amdgpu_device *adev, - u32 se, u32 sh) +static u32 gfx_v7_0_get_cu_active_bitmap(struct amdgpu_device *adev) { - u32 mask = 0, tmp, tmp1; - int i; - - gfx_v7_0_select_se_sh(adev, se, sh); - tmp = RREG32(mmCC_GC_SHADER_ARRAY_CONFIG); - tmp1 = RREG32(mmGC_USER_SHADER_ARRAY_CONFIG); - gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff); + u32 data, mask; - tmp &= 0xffff0000; + data = RREG32(mmCC_GC_SHADER_ARRAY_CONFIG); + data |= RREG32(mmGC_USER_SHADER_ARRAY_CONFIG); - tmp |= tmp1; - tmp >>= 16; + data &= CC_GC_SHADER_ARRAY_CONFIG__INACTIVE_CUS_MASK; + data >>= CC_GC_SHADER_ARRAY_CONFIG__INACTIVE_CUS__SHIFT; - for (i = 0; i < adev->gfx.config.max_cu_per_sh; i ++) { - mask <<= 1; - mask |= 1; - } + mask = gfx_v7_0_create_bitmask(adev->gfx.config.max_cu_per_sh); - return (~tmp) & mask; + return (~data) & mask; } static void gfx_v7_0_init_ao_cu_mask(struct amdgpu_device *adev) @@ -4442,7 +3890,7 @@ static void gfx_v7_0_enable_gfx_static_mgpg(struct amdgpu_device *adev, u32 data, orig; orig = data = RREG32(mmRLC_PG_CNTL); - if (enable && (adev->pg_flags & AMDGPU_PG_SUPPORT_GFX_SMG)) + if (enable && (adev->pg_flags & AMD_PG_SUPPORT_GFX_SMG)) data |= RLC_PG_CNTL__STATIC_PER_CU_PG_ENABLE_MASK; else data &= ~RLC_PG_CNTL__STATIC_PER_CU_PG_ENABLE_MASK; @@ -4456,7 +3904,7 @@ static void gfx_v7_0_enable_gfx_dynamic_mgpg(struct amdgpu_device *adev, u32 data, orig; orig = data = RREG32(mmRLC_PG_CNTL); - if (enable && (adev->pg_flags & AMDGPU_PG_SUPPORT_GFX_DMG)) + if (enable && (adev->pg_flags & AMD_PG_SUPPORT_GFX_DMG)) data |= RLC_PG_CNTL__DYN_PER_CU_PG_ENABLE_MASK; else data &= ~RLC_PG_CNTL__DYN_PER_CU_PG_ENABLE_MASK; @@ -4623,15 +4071,15 @@ static void gfx_v7_0_get_csb_buffer(struct amdgpu_device *adev, static void gfx_v7_0_init_pg(struct amdgpu_device *adev) { - if (adev->pg_flags & (AMDGPU_PG_SUPPORT_GFX_PG | - AMDGPU_PG_SUPPORT_GFX_SMG | - AMDGPU_PG_SUPPORT_GFX_DMG | - AMDGPU_PG_SUPPORT_CP | - AMDGPU_PG_SUPPORT_GDS | - AMDGPU_PG_SUPPORT_RLC_SMU_HS)) { + if (adev->pg_flags & (AMD_PG_SUPPORT_GFX_PG | + AMD_PG_SUPPORT_GFX_SMG | + AMD_PG_SUPPORT_GFX_DMG | + AMD_PG_SUPPORT_CP | + AMD_PG_SUPPORT_GDS | + AMD_PG_SUPPORT_RLC_SMU_HS)) { gfx_v7_0_enable_sclk_slowdown_on_pu(adev, true); gfx_v7_0_enable_sclk_slowdown_on_pd(adev, true); - if (adev->pg_flags & AMDGPU_PG_SUPPORT_GFX_PG) { + if (adev->pg_flags & AMD_PG_SUPPORT_GFX_PG) { gfx_v7_0_init_gfx_cgpg(adev); gfx_v7_0_enable_cp_pg(adev, true); gfx_v7_0_enable_gds_pg(adev, true); @@ -4643,14 +4091,14 @@ static void gfx_v7_0_init_pg(struct amdgpu_device *adev) static void gfx_v7_0_fini_pg(struct amdgpu_device *adev) { - if (adev->pg_flags & (AMDGPU_PG_SUPPORT_GFX_PG | - AMDGPU_PG_SUPPORT_GFX_SMG | - AMDGPU_PG_SUPPORT_GFX_DMG | - AMDGPU_PG_SUPPORT_CP | - AMDGPU_PG_SUPPORT_GDS | - AMDGPU_PG_SUPPORT_RLC_SMU_HS)) { + if (adev->pg_flags & (AMD_PG_SUPPORT_GFX_PG | + AMD_PG_SUPPORT_GFX_SMG | + AMD_PG_SUPPORT_GFX_DMG | + AMD_PG_SUPPORT_CP | + AMD_PG_SUPPORT_GDS | + AMD_PG_SUPPORT_RLC_SMU_HS)) { gfx_v7_0_update_gfx_pg(adev, false); - if (adev->pg_flags & AMDGPU_PG_SUPPORT_GFX_PG) { + if (adev->pg_flags & AMD_PG_SUPPORT_GFX_PG) { gfx_v7_0_enable_cp_pg(adev, false); gfx_v7_0_enable_gds_pg(adev, false); } @@ -4738,6 +4186,188 @@ static int gfx_v7_0_early_init(void *handle) return 0; } +static int gfx_v7_0_late_init(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + int r; + + r = amdgpu_irq_get(adev, &adev->gfx.priv_reg_irq, 0); + if (r) + return r; + + r = amdgpu_irq_get(adev, &adev->gfx.priv_inst_irq, 0); + if (r) + return r; + + return 0; +} + +static void gfx_v7_0_gpu_early_init(struct amdgpu_device *adev) +{ + u32 gb_addr_config; + u32 mc_shared_chmap, mc_arb_ramcfg; + u32 dimm00_addr_map, dimm01_addr_map, dimm10_addr_map, dimm11_addr_map; + u32 tmp; + + switch (adev->asic_type) { + case CHIP_BONAIRE: + adev->gfx.config.max_shader_engines = 2; + adev->gfx.config.max_tile_pipes = 4; + adev->gfx.config.max_cu_per_sh = 7; + adev->gfx.config.max_sh_per_se = 1; + adev->gfx.config.max_backends_per_se = 2; + adev->gfx.config.max_texture_channel_caches = 4; + adev->gfx.config.max_gprs = 256; + adev->gfx.config.max_gs_threads = 32; + adev->gfx.config.max_hw_contexts = 8; + + adev->gfx.config.sc_prim_fifo_size_frontend = 0x20; + adev->gfx.config.sc_prim_fifo_size_backend = 0x100; + adev->gfx.config.sc_hiz_tile_fifo_size = 0x30; + adev->gfx.config.sc_earlyz_tile_fifo_size = 0x130; + gb_addr_config = BONAIRE_GB_ADDR_CONFIG_GOLDEN; + break; + case CHIP_HAWAII: + adev->gfx.config.max_shader_engines = 4; + adev->gfx.config.max_tile_pipes = 16; + adev->gfx.config.max_cu_per_sh = 11; + adev->gfx.config.max_sh_per_se = 1; + adev->gfx.config.max_backends_per_se = 4; + adev->gfx.config.max_texture_channel_caches = 16; + adev->gfx.config.max_gprs = 256; + adev->gfx.config.max_gs_threads = 32; + adev->gfx.config.max_hw_contexts = 8; + + adev->gfx.config.sc_prim_fifo_size_frontend = 0x20; + adev->gfx.config.sc_prim_fifo_size_backend = 0x100; + adev->gfx.config.sc_hiz_tile_fifo_size = 0x30; + adev->gfx.config.sc_earlyz_tile_fifo_size = 0x130; + gb_addr_config = HAWAII_GB_ADDR_CONFIG_GOLDEN; + break; + case CHIP_KAVERI: + adev->gfx.config.max_shader_engines = 1; + adev->gfx.config.max_tile_pipes = 4; + if ((adev->pdev->device == 0x1304) || + (adev->pdev->device == 0x1305) || + (adev->pdev->device == 0x130C) || + (adev->pdev->device == 0x130F) || + (adev->pdev->device == 0x1310) || + (adev->pdev->device == 0x1311) || + (adev->pdev->device == 0x131C)) { + adev->gfx.config.max_cu_per_sh = 8; + adev->gfx.config.max_backends_per_se = 2; + } else if ((adev->pdev->device == 0x1309) || + (adev->pdev->device == 0x130A) || + (adev->pdev->device == 0x130D) || + (adev->pdev->device == 0x1313) || + (adev->pdev->device == 0x131D)) { + adev->gfx.config.max_cu_per_sh = 6; + adev->gfx.config.max_backends_per_se = 2; + } else if ((adev->pdev->device == 0x1306) || + (adev->pdev->device == 0x1307) || + (adev->pdev->device == 0x130B) || + (adev->pdev->device == 0x130E) || + (adev->pdev->device == 0x1315) || + (adev->pdev->device == 0x131B)) { + adev->gfx.config.max_cu_per_sh = 4; + adev->gfx.config.max_backends_per_se = 1; + } else { + adev->gfx.config.max_cu_per_sh = 3; + adev->gfx.config.max_backends_per_se = 1; + } + adev->gfx.config.max_sh_per_se = 1; + adev->gfx.config.max_texture_channel_caches = 4; + adev->gfx.config.max_gprs = 256; + adev->gfx.config.max_gs_threads = 16; + adev->gfx.config.max_hw_contexts = 8; + + adev->gfx.config.sc_prim_fifo_size_frontend = 0x20; + adev->gfx.config.sc_prim_fifo_size_backend = 0x100; + adev->gfx.config.sc_hiz_tile_fifo_size = 0x30; + adev->gfx.config.sc_earlyz_tile_fifo_size = 0x130; + gb_addr_config = BONAIRE_GB_ADDR_CONFIG_GOLDEN; + break; + case CHIP_KABINI: + case CHIP_MULLINS: + default: + adev->gfx.config.max_shader_engines = 1; + adev->gfx.config.max_tile_pipes = 2; + adev->gfx.config.max_cu_per_sh = 2; + adev->gfx.config.max_sh_per_se = 1; + adev->gfx.config.max_backends_per_se = 1; + adev->gfx.config.max_texture_channel_caches = 2; + adev->gfx.config.max_gprs = 256; + adev->gfx.config.max_gs_threads = 16; + adev->gfx.config.max_hw_contexts = 8; + + adev->gfx.config.sc_prim_fifo_size_frontend = 0x20; + adev->gfx.config.sc_prim_fifo_size_backend = 0x100; + adev->gfx.config.sc_hiz_tile_fifo_size = 0x30; + adev->gfx.config.sc_earlyz_tile_fifo_size = 0x130; + gb_addr_config = BONAIRE_GB_ADDR_CONFIG_GOLDEN; + break; + } + + mc_shared_chmap = RREG32(mmMC_SHARED_CHMAP); + adev->gfx.config.mc_arb_ramcfg = RREG32(mmMC_ARB_RAMCFG); + mc_arb_ramcfg = adev->gfx.config.mc_arb_ramcfg; + + adev->gfx.config.num_tile_pipes = adev->gfx.config.max_tile_pipes; + adev->gfx.config.mem_max_burst_length_bytes = 256; + if (adev->flags & AMD_IS_APU) { + /* Get memory bank mapping mode. */ + tmp = RREG32(mmMC_FUS_DRAM0_BANK_ADDR_MAPPING); + dimm00_addr_map = REG_GET_FIELD(tmp, MC_FUS_DRAM0_BANK_ADDR_MAPPING, DIMM0ADDRMAP); + dimm01_addr_map = REG_GET_FIELD(tmp, MC_FUS_DRAM0_BANK_ADDR_MAPPING, DIMM1ADDRMAP); + + tmp = RREG32(mmMC_FUS_DRAM1_BANK_ADDR_MAPPING); + dimm10_addr_map = REG_GET_FIELD(tmp, MC_FUS_DRAM1_BANK_ADDR_MAPPING, DIMM0ADDRMAP); + dimm11_addr_map = REG_GET_FIELD(tmp, MC_FUS_DRAM1_BANK_ADDR_MAPPING, DIMM1ADDRMAP); + + /* Validate settings in case only one DIMM installed. */ + if ((dimm00_addr_map == 0) || (dimm00_addr_map == 3) || (dimm00_addr_map == 4) || (dimm00_addr_map > 12)) + dimm00_addr_map = 0; + if ((dimm01_addr_map == 0) || (dimm01_addr_map == 3) || (dimm01_addr_map == 4) || (dimm01_addr_map > 12)) + dimm01_addr_map = 0; + if ((dimm10_addr_map == 0) || (dimm10_addr_map == 3) || (dimm10_addr_map == 4) || (dimm10_addr_map > 12)) + dimm10_addr_map = 0; + if ((dimm11_addr_map == 0) || (dimm11_addr_map == 3) || (dimm11_addr_map == 4) || (dimm11_addr_map > 12)) + dimm11_addr_map = 0; + + /* If DIMM Addr map is 8GB, ROW size should be 2KB. Otherwise 1KB. */ + /* If ROW size(DIMM1) != ROW size(DMIMM0), ROW size should be larger one. */ + if ((dimm00_addr_map == 11) || (dimm01_addr_map == 11) || (dimm10_addr_map == 11) || (dimm11_addr_map == 11)) + adev->gfx.config.mem_row_size_in_kb = 2; + else + adev->gfx.config.mem_row_size_in_kb = 1; + } else { + tmp = (mc_arb_ramcfg & MC_ARB_RAMCFG__NOOFCOLS_MASK) >> MC_ARB_RAMCFG__NOOFCOLS__SHIFT; + adev->gfx.config.mem_row_size_in_kb = (4 * (1 << (8 + tmp))) / 1024; + if (adev->gfx.config.mem_row_size_in_kb > 4) + adev->gfx.config.mem_row_size_in_kb = 4; + } + /* XXX use MC settings? */ + adev->gfx.config.shader_engine_tile_size = 32; + adev->gfx.config.num_gpus = 1; + adev->gfx.config.multi_gpu_tile_size = 64; + + /* fix up row size */ + gb_addr_config &= ~GB_ADDR_CONFIG__ROW_SIZE_MASK; + switch (adev->gfx.config.mem_row_size_in_kb) { + case 1: + default: + gb_addr_config |= (0 << GB_ADDR_CONFIG__ROW_SIZE__SHIFT); + break; + case 2: + gb_addr_config |= (1 << GB_ADDR_CONFIG__ROW_SIZE__SHIFT); + break; + case 4: + gb_addr_config |= (2 << GB_ADDR_CONFIG__ROW_SIZE__SHIFT); + break; + } + adev->gfx.config.gb_addr_config = gb_addr_config; +} + static int gfx_v7_0_sw_init(void *handle) { struct amdgpu_ring *ring; @@ -4841,6 +4471,10 @@ static int gfx_v7_0_sw_init(void *handle) if (r) return r; + adev->gfx.ce_ram_size = 0x8000; + + gfx_v7_0_gpu_early_init(adev); + return r; } @@ -4881,8 +4515,6 @@ static int gfx_v7_0_hw_init(void *handle) if (r) return r; - adev->gfx.ce_ram_size = 0x8000; - return r; } @@ -4890,6 +4522,8 @@ static int gfx_v7_0_hw_fini(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; + amdgpu_irq_put(adev, &adev->gfx.priv_reg_irq, 0); + amdgpu_irq_put(adev, &adev->gfx.priv_inst_irq, 0); gfx_v7_0_cp_enable(adev, false); gfx_v7_0_rlc_stop(adev); gfx_v7_0_fini_pg(adev); @@ -4997,16 +4631,6 @@ static void gfx_v7_0_print_status(void *handle) RREG32(mmHDP_ADDR_CONFIG)); dev_info(adev->dev, " DMIF_ADDR_CALC=0x%08X\n", RREG32(mmDMIF_ADDR_CALC)); - dev_info(adev->dev, " SDMA0_TILING_CONFIG=0x%08X\n", - RREG32(mmSDMA0_TILING_CONFIG + SDMA0_REGISTER_OFFSET)); - dev_info(adev->dev, " SDMA1_TILING_CONFIG=0x%08X\n", - RREG32(mmSDMA0_TILING_CONFIG + SDMA1_REGISTER_OFFSET)); - dev_info(adev->dev, " UVD_UDEC_ADDR_CONFIG=0x%08X\n", - RREG32(mmUVD_UDEC_ADDR_CONFIG)); - dev_info(adev->dev, " UVD_UDEC_DB_ADDR_CONFIG=0x%08X\n", - RREG32(mmUVD_UDEC_DB_ADDR_CONFIG)); - dev_info(adev->dev, " UVD_UDEC_DBW_ADDR_CONFIG=0x%08X\n", - RREG32(mmUVD_UDEC_DBW_ADDR_CONFIG)); dev_info(adev->dev, " CP_MEQ_THRESHOLDS=0x%08X\n", RREG32(mmCP_MEQ_THRESHOLDS)); @@ -5509,14 +5133,14 @@ static int gfx_v7_0_set_powergating_state(void *handle, if (state == AMD_PG_STATE_GATE) gate = true; - if (adev->pg_flags & (AMDGPU_PG_SUPPORT_GFX_PG | - AMDGPU_PG_SUPPORT_GFX_SMG | - AMDGPU_PG_SUPPORT_GFX_DMG | - AMDGPU_PG_SUPPORT_CP | - AMDGPU_PG_SUPPORT_GDS | - AMDGPU_PG_SUPPORT_RLC_SMU_HS)) { + if (adev->pg_flags & (AMD_PG_SUPPORT_GFX_PG | + AMD_PG_SUPPORT_GFX_SMG | + AMD_PG_SUPPORT_GFX_DMG | + AMD_PG_SUPPORT_CP | + AMD_PG_SUPPORT_GDS | + AMD_PG_SUPPORT_RLC_SMU_HS)) { gfx_v7_0_update_gfx_pg(adev, gate); - if (adev->pg_flags & AMDGPU_PG_SUPPORT_GFX_PG) { + if (adev->pg_flags & AMD_PG_SUPPORT_GFX_PG) { gfx_v7_0_enable_cp_pg(adev, gate); gfx_v7_0_enable_gds_pg(adev, gate); } @@ -5527,7 +5151,7 @@ static int gfx_v7_0_set_powergating_state(void *handle, const struct amd_ip_funcs gfx_v7_0_ip_funcs = { .early_init = gfx_v7_0_early_init, - .late_init = NULL, + .late_init = gfx_v7_0_late_init, .sw_init = gfx_v7_0_sw_init, .sw_fini = gfx_v7_0_sw_fini, .hw_init = gfx_v7_0_hw_init, @@ -5549,13 +5173,15 @@ static const struct amdgpu_ring_funcs gfx_v7_0_ring_funcs_gfx = { .parse_cs = NULL, .emit_ib = gfx_v7_0_ring_emit_ib_gfx, .emit_fence = gfx_v7_0_ring_emit_fence_gfx, - .emit_semaphore = gfx_v7_0_ring_emit_semaphore, + .emit_pipeline_sync = gfx_v7_0_ring_emit_pipeline_sync, .emit_vm_flush = gfx_v7_0_ring_emit_vm_flush, .emit_gds_switch = gfx_v7_0_ring_emit_gds_switch, .emit_hdp_flush = gfx_v7_0_ring_emit_hdp_flush, + .emit_hdp_invalidate = gfx_v7_0_ring_emit_hdp_invalidate, .test_ring = gfx_v7_0_ring_test_ring, .test_ib = gfx_v7_0_ring_test_ib, .insert_nop = amdgpu_ring_insert_nop, + .pad_ib = amdgpu_ring_generic_pad_ib, }; static const struct amdgpu_ring_funcs gfx_v7_0_ring_funcs_compute = { @@ -5565,13 +5191,15 @@ static const struct amdgpu_ring_funcs gfx_v7_0_ring_funcs_compute = { .parse_cs = NULL, .emit_ib = gfx_v7_0_ring_emit_ib_compute, .emit_fence = gfx_v7_0_ring_emit_fence_compute, - .emit_semaphore = gfx_v7_0_ring_emit_semaphore, + .emit_pipeline_sync = gfx_v7_0_ring_emit_pipeline_sync, .emit_vm_flush = gfx_v7_0_ring_emit_vm_flush, .emit_gds_switch = gfx_v7_0_ring_emit_gds_switch, .emit_hdp_flush = gfx_v7_0_ring_emit_hdp_flush, + .emit_hdp_invalidate = gfx_v7_0_ring_emit_hdp_invalidate, .test_ring = gfx_v7_0_ring_test_ring, .test_ib = gfx_v7_0_ring_test_ib, .insert_nop = amdgpu_ring_insert_nop, + .pad_ib = amdgpu_ring_generic_pad_ib, }; static void gfx_v7_0_set_ring_funcs(struct amdgpu_device *adev) @@ -5641,7 +5269,7 @@ static void gfx_v7_0_set_gds_init(struct amdgpu_device *adev) int gfx_v7_0_get_cu_info(struct amdgpu_device *adev, - struct amdgpu_cu_info *cu_info) + struct amdgpu_cu_info *cu_info) { int i, j, k, counter, active_cu_number = 0; u32 mask, bitmap, ao_bitmap, ao_cu_mask = 0; @@ -5649,16 +5277,19 @@ int gfx_v7_0_get_cu_info(struct amdgpu_device *adev, if (!adev || !cu_info) return -EINVAL; + memset(cu_info, 0, sizeof(*cu_info)); + mutex_lock(&adev->grbm_idx_mutex); for (i = 0; i < adev->gfx.config.max_shader_engines; i++) { for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) { mask = 1; ao_bitmap = 0; counter = 0; - bitmap = gfx_v7_0_get_cu_active_bitmap(adev, i, j); + gfx_v7_0_select_se_sh(adev, i, j); + bitmap = gfx_v7_0_get_cu_active_bitmap(adev); cu_info->bitmap[i][j] = bitmap; - for (k = 0; k < adev->gfx.config.max_cu_per_sh; k ++) { + for (k = 0; k < 16; k ++) { if (bitmap & mask) { if (counter < 2) ao_bitmap |= mask; @@ -5670,9 +5301,11 @@ int gfx_v7_0_get_cu_info(struct amdgpu_device *adev, ao_cu_mask |= (ao_bitmap << (i * 16 + j * 8)); } } + gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff); + mutex_unlock(&adev->grbm_idx_mutex); cu_info->number = active_cu_number; cu_info->ao_cu_mask = ao_cu_mask; - mutex_unlock(&adev->grbm_idx_mutex); + return 0; } |