summaryrefslogtreecommitdiffstats
path: root/drivers/gpu/drm/amd/display/dc
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/gpu/drm/amd/display/dc')
-rw-r--r--drivers/gpu/drm/amd/display/dc/bios/bios_parser2.c9
-rw-r--r--drivers/gpu/drm/amd/display/dc/clk_mgr/Makefile2
-rw-r--r--drivers/gpu/drm/amd/display/dc/clk_mgr/clk_mgr.c5
-rw-r--r--drivers/gpu/drm/amd/display/dc/clk_mgr/dcn201/dcn201_clk_mgr.c4
-rw-r--r--drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn351_clk_mgr.c140
-rw-r--r--drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c243
-rw-r--r--drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.h4
-rw-r--r--drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_smu.h41
-rw-r--r--drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dalsmc.h4
-rw-r--r--drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr.c311
-rw-r--r--drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr.h1
-rw-r--r--drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr_smu_msg.c37
-rw-r--r--drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr_smu_msg.h4
-rw-r--r--drivers/gpu/drm/amd/display/dc/core/dc.c141
-rw-r--r--drivers/gpu/drm/amd/display/dc/core/dc_hw_sequencer.c38
-rw-r--r--drivers/gpu/drm/amd/display/dc/core/dc_link_exports.c8
-rw-r--r--drivers/gpu/drm/amd/display/dc/core/dc_resource.c58
-rw-r--r--drivers/gpu/drm/amd/display/dc/core/dc_state.c8
-rw-r--r--drivers/gpu/drm/amd/display/dc/core/dc_stream.c13
-rw-r--r--drivers/gpu/drm/amd/display/dc/core/dc_surface.c53
-rw-r--r--drivers/gpu/drm/amd/display/dc/dc.h52
-rw-r--r--drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c37
-rw-r--r--drivers/gpu/drm/amd/display/dc/dc_dp_types.h16
-rw-r--r--drivers/gpu/drm/amd/display/dc/dc_dsc.h5
-rw-r--r--drivers/gpu/drm/amd/display/dc/dc_hw_types.h179
-rw-r--r--drivers/gpu/drm/amd/display/dc/dc_plane.h3
-rw-r--r--drivers/gpu/drm/amd/display/dc/dc_spl_translate.c19
-rw-r--r--drivers/gpu/drm/amd/display/dc/dc_stream.h17
-rw-r--r--drivers/gpu/drm/amd/display/dc/dc_types.h35
-rw-r--r--drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.c3
-rw-r--r--drivers/gpu/drm/amd/display/dc/dce/dce_mem_input.c44
-rw-r--r--drivers/gpu/drm/amd/display/dc/dce/dmub_hw_lock_mgr.c3
-rw-r--r--drivers/gpu/drm/amd/display/dc/dce/dmub_psr.c3
-rw-r--r--drivers/gpu/drm/amd/display/dc/dce110/dce110_mem_input_v.c8
-rw-r--r--drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator.c201
-rw-r--r--drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator.h2
-rw-r--r--drivers/gpu/drm/amd/display/dc/dce120/dce120_timing_generator.c124
-rw-r--r--drivers/gpu/drm/amd/display/dc/dcn31/dcn31_panel_cntl.c34
-rw-r--r--drivers/gpu/drm/amd/display/dc/dio/dcn31/dcn31_dio_link_encoder.c5
-rw-r--r--drivers/gpu/drm/amd/display/dc/dio/dcn35/dcn35_dio_link_encoder.c80
-rw-r--r--drivers/gpu/drm/amd/display/dc/dio/dcn35/dcn35_dio_link_encoder.h18
-rw-r--r--drivers/gpu/drm/amd/display/dc/dm_helpers.h5
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml/calcs/dcn_calcs.c132
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml/dcn30/display_rq_dlg_calc_30.c1
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_util_32.c4
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c4
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml/dml_inline_defs.h8
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/Makefile19
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/display_mode_core.c42
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/display_mode_core_structs.h109
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/display_mode_util.c6
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.c162
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.h2
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_utils.c209
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_utils.h8
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_wrapper.c8
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/bounding_boxes/dcn3_soc_bb.h401
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/bounding_boxes/dcn4_soc_bb.h3
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top_display_cfg_types.h52
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top_soc_parameter_types.h2
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top_types.h77
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4.c42
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c1412
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.h2
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_shared_types.h135
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_utils.c225
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_utils.h6
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_dpmm/dml2_dpmm_dcn4.c59
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn3.c12
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn4_fams2.c428
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn4_fams2.h7
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_factory.c1
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_interfaces.c50
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_legacy.c10
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_legacy.h9
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_optimization.c307
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_optimization.h33
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_soc15.c1178
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_soc15.h (renamed from drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml_top_mcache.h)19
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml_top_mcache.c549
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/src/inc/dml2_debug.c5
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/src/inc/dml2_debug.h46
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/src/inc/dml2_internal_shared_types.h58
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml2_mall_phantom.c2
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c54
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.c49
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml_display_rq_dlg_calc.c12
-rw-r--r--drivers/gpu/drm/amd/display/dc/dsc/dc_dsc.c45
-rw-r--r--drivers/gpu/drm/amd/display/dc/dwb/dcn30/dcn30_dwb.c13
-rw-r--r--drivers/gpu/drm/amd/display/dc/dwb/dcn30/dcn30_dwb.h1
-rw-r--r--drivers/gpu/drm/amd/display/dc/hubp/dcn10/dcn10_hubp.c19
-rw-r--r--drivers/gpu/drm/amd/display/dc/hubp/dcn10/dcn10_hubp.h6
-rw-r--r--drivers/gpu/drm/amd/display/dc/hubp/dcn20/dcn20_hubp.c19
-rw-r--r--drivers/gpu/drm/amd/display/dc/hubp/dcn20/dcn20_hubp.h4
-rw-r--r--drivers/gpu/drm/amd/display/dc/hubp/dcn201/dcn201_hubp.c3
-rw-r--r--drivers/gpu/drm/amd/display/dc/hubp/dcn21/dcn21_hubp.c1
-rw-r--r--drivers/gpu/drm/amd/display/dc/hubp/dcn30/dcn30_hubp.c21
-rw-r--r--drivers/gpu/drm/amd/display/dc/hubp/dcn30/dcn30_hubp.h6
-rw-r--r--drivers/gpu/drm/amd/display/dc/hubp/dcn31/dcn31_hubp.c1
-rw-r--r--drivers/gpu/drm/amd/display/dc/hubp/dcn32/dcn32_hubp.c3
-rw-r--r--drivers/gpu/drm/amd/display/dc/hubp/dcn35/dcn35_hubp.c3
-rw-r--r--drivers/gpu/drm/amd/display/dc/hubp/dcn35/dcn35_hubp.h2
-rw-r--r--drivers/gpu/drm/amd/display/dc/hubp/dcn401/dcn401_hubp.c131
-rw-r--r--drivers/gpu/drm/amd/display/dc/hubp/dcn401/dcn401_hubp.h64
-rw-r--r--drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c257
-rw-r--r--drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.h17
-rw-r--r--drivers/gpu/drm/amd/display/dc/hwss/dcn30/dcn30_init.c1
-rw-r--r--drivers/gpu/drm/amd/display/dc/hwss/dcn301/dcn301_init.c1
-rw-r--r--drivers/gpu/drm/amd/display/dc/hwss/dcn31/dcn31_init.c3
-rw-r--r--drivers/gpu/drm/amd/display/dc/hwss/dcn314/dcn314_hwseq.c12
-rw-r--r--drivers/gpu/drm/amd/display/dc/hwss/dcn314/dcn314_init.c3
-rw-r--r--drivers/gpu/drm/amd/display/dc/hwss/dcn32/dcn32_hwseq.c16
-rw-r--r--drivers/gpu/drm/amd/display/dc/hwss/dcn32/dcn32_init.c1
-rw-r--r--drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.c53
-rw-r--r--drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.h2
-rw-r--r--drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_init.c6
-rw-r--r--drivers/gpu/drm/amd/display/dc/hwss/dcn351/dcn351_init.c1
-rw-r--r--drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c841
-rw-r--r--drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.h11
-rw-r--r--drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_init.c11
-rw-r--r--drivers/gpu/drm/amd/display/dc/hwss/hw_sequencer.h22
-rw-r--r--drivers/gpu/drm/amd/display/dc/inc/core_types.h11
-rw-r--r--drivers/gpu/drm/amd/display/dc/inc/dcn_calcs.h4
-rw-r--r--drivers/gpu/drm/amd/display/dc/inc/hw/clk_mgr.h3
-rw-r--r--drivers/gpu/drm/amd/display/dc/inc/hw/clk_mgr_internal.h59
-rw-r--r--drivers/gpu/drm/amd/display/dc/inc/hw/hubp.h16
-rw-r--r--drivers/gpu/drm/amd/display/dc/inc/hw/link_encoder.h8
-rw-r--r--drivers/gpu/drm/amd/display/dc/inc/hw/mem_input.h6
-rw-r--r--drivers/gpu/drm/amd/display/dc/inc/hw/optc.h2
-rw-r--r--drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h5
-rw-r--r--drivers/gpu/drm/amd/display/dc/inc/link.h4
-rw-r--r--drivers/gpu/drm/amd/display/dc/irq/dcn201/irq_service_dcn201.c4
-rw-r--r--drivers/gpu/drm/amd/display/dc/link/accessories/link_dp_cts.c4
-rw-r--r--drivers/gpu/drm/amd/display/dc/link/hwss/link_hwss_dio.c4
-rw-r--r--drivers/gpu/drm/amd/display/dc/link/hwss/link_hwss_dio_fixed_vs_pe_retimer.c2
-rw-r--r--drivers/gpu/drm/amd/display/dc/link/hwss/link_hwss_dpia.c61
-rw-r--r--drivers/gpu/drm/amd/display/dc/link/hwss/link_hwss_dpia.h3
-rw-r--r--drivers/gpu/drm/amd/display/dc/link/link_detection.c3
-rw-r--r--drivers/gpu/drm/amd/display/dc/link/link_dpms.c50
-rw-r--r--drivers/gpu/drm/amd/display/dc/link/link_factory.c3
-rw-r--r--drivers/gpu/drm/amd/display/dc/link/link_validation.c179
-rw-r--r--drivers/gpu/drm/amd/display/dc/link/link_validation.h5
-rw-r--r--drivers/gpu/drm/amd/display/dc/link/protocols/link_ddc.c2
-rw-r--r--drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c36
-rw-r--r--drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_irq_handler.c8
-rw-r--r--drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_phy.c2
-rw-r--r--drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training.c4
-rw-r--r--drivers/gpu/drm/amd/display/dc/mpc/dcn30/dcn30_mpc.c18
-rw-r--r--drivers/gpu/drm/amd/display/dc/mpc/dcn30/dcn30_mpc.h7
-rw-r--r--drivers/gpu/drm/amd/display/dc/optc/dcn10/dcn10_optc.c120
-rw-r--r--drivers/gpu/drm/amd/display/dc/optc/dcn10/dcn10_optc.h19
-rw-r--r--drivers/gpu/drm/amd/display/dc/optc/dcn35/dcn35_optc.c101
-rw-r--r--drivers/gpu/drm/amd/display/dc/optc/dcn401/dcn401_optc.c4
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn10/dcn10_resource.c8
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn10/dcn10_resource.h1
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn20/dcn20_resource.c43
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn201/dcn201_resource.c7
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn21/dcn21_resource.c1
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn30/dcn30_resource.c2
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn301/dcn301_resource.c11
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn302/dcn302_resource.c2
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn303/dcn303_resource.c2
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn31/dcn31_resource.c8
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn31/dcn31_resource.h3
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn314/dcn314_resource.c2
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn315/dcn315_resource.c2
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn316/dcn316_resource.c2
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn32/dcn32_resource.c3
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn321/dcn321_resource.c2
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn35/dcn35_resource.c12
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn35/dcn35_resource.h1
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn351/dcn351_resource.c5
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn401/dcn401_resource.c62
-rw-r--r--drivers/gpu/drm/amd/display/dc/spl/dc_spl.c177
-rw-r--r--drivers/gpu/drm/amd/display/dc/spl/dc_spl_types.h16
175 files changed, 6877 insertions, 3828 deletions
diff --git a/drivers/gpu/drm/amd/display/dc/bios/bios_parser2.c b/drivers/gpu/drm/amd/display/dc/bios/bios_parser2.c
index c9a6de110b74..a62f6c51301c 100644
--- a/drivers/gpu/drm/amd/display/dc/bios/bios_parser2.c
+++ b/drivers/gpu/drm/amd/display/dc/bios/bios_parser2.c
@@ -3088,11 +3088,12 @@ static enum bp_result construct_integrated_info(
info->ext_disp_conn_info.path[i].ext_encoder_obj_id.id,
info->ext_disp_conn_info.path[i].caps
);
- if (info->ext_disp_conn_info.path[i].caps & EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN)
- DC_LOG_BIOS("BIOS EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN on path %d\n", i);
+ if ((info->ext_disp_conn_info.path[i].caps & AMD_EXT_DISPLAY_PATH_CAPS__EXT_CHIP_MASK) == AMD_EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN)
+ DC_LOG_BIOS("BIOS AMD_EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN on path %d\n", i);
else if (bp->base.ctx->dc->config.force_bios_fixed_vs) {
- info->ext_disp_conn_info.path[i].caps |= EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN;
- DC_LOG_BIOS("driver forced EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN on path %d\n", i);
+ info->ext_disp_conn_info.path[i].caps &= ~AMD_EXT_DISPLAY_PATH_CAPS__EXT_CHIP_MASK;
+ info->ext_disp_conn_info.path[i].caps |= AMD_EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN;
+ DC_LOG_BIOS("driver forced AMD_EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN on path %d\n", i);
}
}
// Log the Checksum and Voltage Swing
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/Makefile b/drivers/gpu/drm/amd/display/dc/clk_mgr/Makefile
index ab1132bc896a..d9955c5d2e5e 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/Makefile
@@ -174,7 +174,7 @@ AMD_DISPLAY_FILES += $(AMD_DAL_CLK_MGR_DCN32)
###############################################################################
# DCN35
###############################################################################
-CLK_MGR_DCN35 = dcn35_smu.o dcn35_clk_mgr.o
+CLK_MGR_DCN35 = dcn35_smu.o dcn351_clk_mgr.o dcn35_clk_mgr.o
AMD_DAL_CLK_MGR_DCN35 = $(addprefix $(AMDDALPATH)/dc/clk_mgr/dcn35/,$(CLK_MGR_DCN35))
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/clk_mgr.c
index 0e243f4344d0..4c3e58c730b1 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/clk_mgr.c
@@ -355,8 +355,11 @@ struct clk_mgr *dc_clk_mgr_create(struct dc_context *ctx, struct pp_smu_funcs *p
BREAK_TO_DEBUGGER();
return NULL;
}
+ if (ctx->dce_version == DCN_VERSION_3_51)
+ dcn351_clk_mgr_construct(ctx, clk_mgr, pp_smu, dccg);
+ else
+ dcn35_clk_mgr_construct(ctx, clk_mgr, pp_smu, dccg);
- dcn35_clk_mgr_construct(ctx, clk_mgr, pp_smu, dccg);
return &clk_mgr->base.base;
}
break;
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn201/dcn201_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn201/dcn201_clk_mgr.c
index 7920f6f1aa62..76c612ecfe3c 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn201/dcn201_clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn201/dcn201_clk_mgr.c
@@ -34,8 +34,8 @@
#include "dm_services.h"
#include "cyan_skillfish_ip_offset.h"
-#include "dcn/dcn_2_0_3_offset.h"
-#include "dcn/dcn_2_0_3_sh_mask.h"
+#include "dcn/dcn_2_0_1_offset.h"
+#include "dcn/dcn_2_0_1_sh_mask.h"
#include "clk/clk_11_0_1_offset.h"
#include "clk/clk_11_0_1_sh_mask.h"
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn351_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn351_clk_mgr.c
new file mode 100644
index 000000000000..6a6ae618650b
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn351_clk_mgr.c
@@ -0,0 +1,140 @@
+/*
+ * Copyright 2024 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#include "core_types.h"
+#include "dcn35_clk_mgr.h"
+
+#define DCN_BASE__INST0_SEG1 0x000000C0
+#define mmCLK1_CLK_PLL_REQ 0x16E37
+
+#define mmCLK1_CLK0_DFS_CNTL 0x16E69
+#define mmCLK1_CLK1_DFS_CNTL 0x16E6C
+#define mmCLK1_CLK2_DFS_CNTL 0x16E6F
+#define mmCLK1_CLK3_DFS_CNTL 0x16E72
+#define mmCLK1_CLK4_DFS_CNTL 0x16E75
+#define mmCLK1_CLK5_DFS_CNTL 0x16E78
+
+#define mmCLK1_CLK0_CURRENT_CNT 0x16EFC
+#define mmCLK1_CLK1_CURRENT_CNT 0x16EFD
+#define mmCLK1_CLK2_CURRENT_CNT 0x16EFE
+#define mmCLK1_CLK3_CURRENT_CNT 0x16EFF
+#define mmCLK1_CLK4_CURRENT_CNT 0x16F00
+#define mmCLK1_CLK5_CURRENT_CNT 0x16F01
+
+#define mmCLK1_CLK0_BYPASS_CNTL 0x16E8A
+#define mmCLK1_CLK1_BYPASS_CNTL 0x16E93
+#define mmCLK1_CLK2_BYPASS_CNTL 0x16E9C
+#define mmCLK1_CLK3_BYPASS_CNTL 0x16EA5
+#define mmCLK1_CLK4_BYPASS_CNTL 0x16EAE
+#define mmCLK1_CLK5_BYPASS_CNTL 0x16EB7
+
+#define mmCLK1_CLK0_DS_CNTL 0x16E83
+#define mmCLK1_CLK1_DS_CNTL 0x16E8C
+#define mmCLK1_CLK2_DS_CNTL 0x16E95
+#define mmCLK1_CLK3_DS_CNTL 0x16E9E
+#define mmCLK1_CLK4_DS_CNTL 0x16EA7
+#define mmCLK1_CLK5_DS_CNTL 0x16EB0
+
+#define mmCLK1_CLK0_ALLOW_DS 0x16E84
+#define mmCLK1_CLK1_ALLOW_DS 0x16E8D
+#define mmCLK1_CLK2_ALLOW_DS 0x16E96
+#define mmCLK1_CLK3_ALLOW_DS 0x16E9F
+#define mmCLK1_CLK4_ALLOW_DS 0x16EA8
+#define mmCLK1_CLK5_ALLOW_DS 0x16EB1
+
+#define mmCLK5_spll_field_8 0x1B04B
+#define mmDENTIST_DISPCLK_CNTL 0x0124
+#define regDENTIST_DISPCLK_CNTL 0x0064
+#define regDENTIST_DISPCLK_CNTL_BASE_IDX 1
+
+#define CLK1_CLK_PLL_REQ__FbMult_int__SHIFT 0x0
+#define CLK1_CLK_PLL_REQ__PllSpineDiv__SHIFT 0xc
+#define CLK1_CLK_PLL_REQ__FbMult_frac__SHIFT 0x10
+#define CLK1_CLK_PLL_REQ__FbMult_int_MASK 0x000001FFL
+#define CLK1_CLK_PLL_REQ__PllSpineDiv_MASK 0x0000F000L
+#define CLK1_CLK_PLL_REQ__FbMult_frac_MASK 0xFFFF0000L
+
+#define CLK1_CLK2_BYPASS_CNTL__CLK2_BYPASS_SEL_MASK 0x00000007L
+
+// DENTIST_DISPCLK_CNTL
+#define DENTIST_DISPCLK_CNTL__DENTIST_DISPCLK_WDIVIDER__SHIFT 0x0
+#define DENTIST_DISPCLK_CNTL__DENTIST_DISPCLK_RDIVIDER__SHIFT 0x8
+#define DENTIST_DISPCLK_CNTL__DENTIST_DISPCLK_CHG_DONE__SHIFT 0x13
+#define DENTIST_DISPCLK_CNTL__DENTIST_DPPCLK_CHG_DONE__SHIFT 0x14
+#define DENTIST_DISPCLK_CNTL__DENTIST_DPPCLK_WDIVIDER__SHIFT 0x18
+#define DENTIST_DISPCLK_CNTL__DENTIST_DISPCLK_WDIVIDER_MASK 0x0000007FL
+#define DENTIST_DISPCLK_CNTL__DENTIST_DISPCLK_RDIVIDER_MASK 0x00007F00L
+#define DENTIST_DISPCLK_CNTL__DENTIST_DISPCLK_CHG_DONE_MASK 0x00080000L
+#define DENTIST_DISPCLK_CNTL__DENTIST_DPPCLK_CHG_DONE_MASK 0x00100000L
+#define DENTIST_DISPCLK_CNTL__DENTIST_DPPCLK_WDIVIDER_MASK 0x7F000000L
+
+#define CLK5_spll_field_8__spll_ssc_en_MASK 0x00002000L
+
+#define REG(reg) \
+ (clk_mgr->regs->reg)
+
+#define BASE_INNER(seg) DCN_BASE__INST0_SEG ## seg
+
+#define BASE(seg) BASE_INNER(seg)
+
+#define SR(reg_name)\
+ .reg_name = BASE(reg ## reg_name ## _BASE_IDX) + \
+ reg ## reg_name
+
+#define CLK_SR_DCN35(reg_name)\
+ .reg_name = mm ## reg_name
+
+static const struct clk_mgr_registers clk_mgr_regs_dcn351 = {
+ CLK_REG_LIST_DCN35()
+};
+
+static const struct clk_mgr_shift clk_mgr_shift_dcn351 = {
+ CLK_COMMON_MASK_SH_LIST_DCN32(__SHIFT)
+};
+
+static const struct clk_mgr_mask clk_mgr_mask_dcn351 = {
+ CLK_COMMON_MASK_SH_LIST_DCN32(_MASK)
+};
+
+#define TO_CLK_MGR_DCN35(clk_mgr)\
+ container_of(clk_mgr, struct clk_mgr_dcn35, base)
+
+
+void dcn351_clk_mgr_construct(
+ struct dc_context *ctx,
+ struct clk_mgr_dcn35 *clk_mgr,
+ struct pp_smu_funcs *pp_smu,
+ struct dccg *dccg)
+{
+ /*register offset changed*/
+ clk_mgr->base.regs = &clk_mgr_regs_dcn351;
+ clk_mgr->base.clk_mgr_shift = &clk_mgr_shift_dcn351;
+ clk_mgr->base.clk_mgr_mask = &clk_mgr_mask_dcn351;
+
+ dcn35_clk_mgr_construct(ctx, clk_mgr, pp_smu, dccg);
+
+}
+
+
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c
index b77333817f18..1f974ea3b0c6 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c
@@ -36,15 +36,11 @@
#include "dcn20/dcn20_clk_mgr.h"
-
-
#include "reg_helper.h"
#include "core_types.h"
#include "dcn35_smu.h"
#include "dm_helpers.h"
-/* TODO: remove this include once we ported over remaining clk mgr functions*/
-#include "dcn30/dcn30_clk_mgr.h"
#include "dcn31/dcn31_clk_mgr.h"
#include "dc_dmub_srv.h"
@@ -55,35 +51,102 @@
#define DC_LOGGER \
clk_mgr->base.base.ctx->logger
+#define DCN_BASE__INST0_SEG1 0x000000C0
+#define mmCLK1_CLK_PLL_REQ 0x16E37
+
+#define mmCLK1_CLK0_DFS_CNTL 0x16E69
+#define mmCLK1_CLK1_DFS_CNTL 0x16E6C
+#define mmCLK1_CLK2_DFS_CNTL 0x16E6F
+#define mmCLK1_CLK3_DFS_CNTL 0x16E72
+#define mmCLK1_CLK4_DFS_CNTL 0x16E75
+#define mmCLK1_CLK5_DFS_CNTL 0x16E78
+
+#define mmCLK1_CLK0_CURRENT_CNT 0x16EFB
+#define mmCLK1_CLK1_CURRENT_CNT 0x16EFC
+#define mmCLK1_CLK2_CURRENT_CNT 0x16EFD
+#define mmCLK1_CLK3_CURRENT_CNT 0x16EFE
+#define mmCLK1_CLK4_CURRENT_CNT 0x16EFF
+#define mmCLK1_CLK5_CURRENT_CNT 0x16F00
+
+#define mmCLK1_CLK0_BYPASS_CNTL 0x16E8A
+#define mmCLK1_CLK1_BYPASS_CNTL 0x16E93
+#define mmCLK1_CLK2_BYPASS_CNTL 0x16E9C
+#define mmCLK1_CLK3_BYPASS_CNTL 0x16EA5
+#define mmCLK1_CLK4_BYPASS_CNTL 0x16EAE
+#define mmCLK1_CLK5_BYPASS_CNTL 0x16EB7
+
+#define mmCLK1_CLK0_DS_CNTL 0x16E83
+#define mmCLK1_CLK1_DS_CNTL 0x16E8C
+#define mmCLK1_CLK2_DS_CNTL 0x16E95
+#define mmCLK1_CLK3_DS_CNTL 0x16E9E
+#define mmCLK1_CLK4_DS_CNTL 0x16EA7
+#define mmCLK1_CLK5_DS_CNTL 0x16EB0
+
+#define mmCLK1_CLK0_ALLOW_DS 0x16E84
+#define mmCLK1_CLK1_ALLOW_DS 0x16E8D
+#define mmCLK1_CLK2_ALLOW_DS 0x16E96
+#define mmCLK1_CLK3_ALLOW_DS 0x16E9F
+#define mmCLK1_CLK4_ALLOW_DS 0x16EA8
+#define mmCLK1_CLK5_ALLOW_DS 0x16EB1
+
+#define mmCLK5_spll_field_8 0x1B04B
+#define mmDENTIST_DISPCLK_CNTL 0x0124
+#define regDENTIST_DISPCLK_CNTL 0x0064
+#define regDENTIST_DISPCLK_CNTL_BASE_IDX 1
+
+#define CLK1_CLK_PLL_REQ__FbMult_int__SHIFT 0x0
+#define CLK1_CLK_PLL_REQ__PllSpineDiv__SHIFT 0xc
+#define CLK1_CLK_PLL_REQ__FbMult_frac__SHIFT 0x10
+#define CLK1_CLK_PLL_REQ__FbMult_int_MASK 0x000001FFL
+#define CLK1_CLK_PLL_REQ__PllSpineDiv_MASK 0x0000F000L
+#define CLK1_CLK_PLL_REQ__FbMult_frac_MASK 0xFFFF0000L
+
+#define CLK1_CLK2_BYPASS_CNTL__CLK2_BYPASS_SEL_MASK 0x00000007L
+#define CLK1_CLK2_BYPASS_CNTL__CLK2_BYPASS_DIV_MASK 0x000F0000L
+// DENTIST_DISPCLK_CNTL
+#define DENTIST_DISPCLK_CNTL__DENTIST_DISPCLK_WDIVIDER__SHIFT 0x0
+#define DENTIST_DISPCLK_CNTL__DENTIST_DISPCLK_RDIVIDER__SHIFT 0x8
+#define DENTIST_DISPCLK_CNTL__DENTIST_DISPCLK_CHG_DONE__SHIFT 0x13
+#define DENTIST_DISPCLK_CNTL__DENTIST_DPPCLK_CHG_DONE__SHIFT 0x14
+#define DENTIST_DISPCLK_CNTL__DENTIST_DPPCLK_WDIVIDER__SHIFT 0x18
+#define DENTIST_DISPCLK_CNTL__DENTIST_DISPCLK_WDIVIDER_MASK 0x0000007FL
+#define DENTIST_DISPCLK_CNTL__DENTIST_DISPCLK_RDIVIDER_MASK 0x00007F00L
+#define DENTIST_DISPCLK_CNTL__DENTIST_DISPCLK_CHG_DONE_MASK 0x00080000L
+#define DENTIST_DISPCLK_CNTL__DENTIST_DPPCLK_CHG_DONE_MASK 0x00100000L
+#define DENTIST_DISPCLK_CNTL__DENTIST_DPPCLK_WDIVIDER_MASK 0x7F000000L
+
+#define CLK5_spll_field_8__spll_ssc_en_MASK 0x00002000L
-#define regCLK1_CLK_PLL_REQ 0x0237
-#define regCLK1_CLK_PLL_REQ_BASE_IDX 0
+#define SMU_VER_THRESHOLD 0x5D4A00 //93.74.0
+#undef FN
+#define FN(reg_name, field_name) \
+ clk_mgr->clk_mgr_shift->field_name, clk_mgr->clk_mgr_mask->field_name
-#define CLK1_CLK_PLL_REQ__FbMult_int__SHIFT 0x0
-#define CLK1_CLK_PLL_REQ__PllSpineDiv__SHIFT 0xc
-#define CLK1_CLK_PLL_REQ__FbMult_frac__SHIFT 0x10
-#define CLK1_CLK_PLL_REQ__FbMult_int_MASK 0x000001FFL
-#define CLK1_CLK_PLL_REQ__PllSpineDiv_MASK 0x0000F000L
-#define CLK1_CLK_PLL_REQ__FbMult_frac_MASK 0xFFFF0000L
+#define REG(reg) \
+ (clk_mgr->regs->reg)
-#define regCLK1_CLK2_BYPASS_CNTL 0x029c
-#define regCLK1_CLK2_BYPASS_CNTL_BASE_IDX 0
+#define BASE_INNER(seg) DCN_BASE__INST0_SEG ## seg
-#define CLK1_CLK2_BYPASS_CNTL__CLK2_BYPASS_SEL__SHIFT 0x0
-#define CLK1_CLK2_BYPASS_CNTL__CLK2_BYPASS_DIV__SHIFT 0x10
-#define CLK1_CLK2_BYPASS_CNTL__CLK2_BYPASS_SEL_MASK 0x00000007L
-#define CLK1_CLK2_BYPASS_CNTL__CLK2_BYPASS_DIV_MASK 0x000F0000L
+#define BASE(seg) BASE_INNER(seg)
-#define regCLK5_0_CLK5_spll_field_8 0x464b
-#define regCLK5_0_CLK5_spll_field_8_BASE_IDX 0
+#define SR(reg_name)\
+ .reg_name = BASE(reg ## reg_name ## _BASE_IDX) + \
+ reg ## reg_name
-#define CLK5_0_CLK5_spll_field_8__spll_ssc_en__SHIFT 0xd
-#define CLK5_0_CLK5_spll_field_8__spll_ssc_en_MASK 0x00002000L
+#define CLK_SR_DCN35(reg_name)\
+ .reg_name = mm ## reg_name
-#define SMU_VER_THRESHOLD 0x5D4A00 //93.74.0
+static const struct clk_mgr_registers clk_mgr_regs_dcn35 = {
+ CLK_REG_LIST_DCN35()
+};
+
+static const struct clk_mgr_shift clk_mgr_shift_dcn35 = {
+ CLK_COMMON_MASK_SH_LIST_DCN32(__SHIFT)
+};
-#define REG(reg_name) \
- (ctx->clk_reg_offsets[reg ## reg_name ## _BASE_IDX] + reg ## reg_name)
+static const struct clk_mgr_mask clk_mgr_mask_dcn35 = {
+ CLK_COMMON_MASK_SH_LIST_DCN32(_MASK)
+};
#define TO_CLK_MGR_DCN35(clk_mgr)\
container_of(clk_mgr, struct clk_mgr_dcn35, base)
@@ -338,6 +401,7 @@ void dcn35_update_clocks(struct clk_mgr *clk_mgr_base,
if (clk_mgr_base->clks.dtbclk_en && !new_clocks->dtbclk_en) {
if (clk_mgr->base.ctx->dc->config.allow_0_dtb_clk)
dcn35_smu_set_dtbclk(clk_mgr, false);
+
clk_mgr_base->clks.dtbclk_en = new_clocks->dtbclk_en;
}
/* check that we're not already in lower */
@@ -355,11 +419,17 @@ void dcn35_update_clocks(struct clk_mgr *clk_mgr_base,
}
if (!clk_mgr_base->clks.dtbclk_en && new_clocks->dtbclk_en) {
- dcn35_smu_set_dtbclk(clk_mgr, true);
- clk_mgr_base->clks.dtbclk_en = new_clocks->dtbclk_en;
+ int actual_dtbclk = 0;
dcn35_update_clocks_update_dtb_dto(clk_mgr, context, new_clocks->ref_dtbclk_khz);
- clk_mgr_base->clks.ref_dtbclk_khz = new_clocks->ref_dtbclk_khz;
+ dcn35_smu_set_dtbclk(clk_mgr, true);
+
+ actual_dtbclk = REG_READ(CLK1_CLK4_CURRENT_CNT);
+
+ if (actual_dtbclk) {
+ clk_mgr_base->clks.ref_dtbclk_khz = new_clocks->ref_dtbclk_khz;
+ clk_mgr_base->clks.dtbclk_en = new_clocks->dtbclk_en;
+ }
}
/* check that we're not already in D0 */
@@ -452,7 +522,6 @@ static int get_vco_frequency_from_reg(struct clk_mgr_internal *clk_mgr)
struct fixed31_32 pll_req;
unsigned int fbmult_frac_val = 0;
unsigned int fbmult_int_val = 0;
- struct dc_context *ctx = clk_mgr->base.ctx;
/*
* Register value of fbmult is in 8.16 format, we are converting to 314.32
@@ -512,22 +581,20 @@ static void dcn35_dump_clk_registers(struct clk_state_registers_and_bypass *regs
static bool dcn35_is_spll_ssc_enabled(struct clk_mgr *clk_mgr_base)
{
struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
- struct dc_context *ctx = clk_mgr->base.ctx;
+
uint32_t ssc_enable;
- REG_GET(CLK5_0_CLK5_spll_field_8, spll_ssc_en, &ssc_enable);
+ ssc_enable = REG_READ(CLK5_spll_field_8) & CLK5_spll_field_8__spll_ssc_en_MASK;
- return ssc_enable == 1;
+ return ssc_enable != 0;
}
static void init_clk_states(struct clk_mgr *clk_mgr)
{
- struct clk_mgr_internal *clk_mgr_int = TO_CLK_MGR_INTERNAL(clk_mgr);
uint32_t ref_dtbclk = clk_mgr->clks.ref_dtbclk_khz;
+
memset(&(clk_mgr->clks), 0, sizeof(struct dc_clocks));
- if (clk_mgr_int->smu_ver >= SMU_VER_THRESHOLD)
- clk_mgr->clks.dtbclk_en = true; // request DTBCLK disable on first commit
clk_mgr->clks.ref_dtbclk_khz = ref_dtbclk; // restore ref_dtbclk
clk_mgr->clks.p_state_change_support = true;
clk_mgr->clks.prev_p_state_change_support = true;
@@ -538,6 +605,7 @@ static void init_clk_states(struct clk_mgr *clk_mgr)
void dcn35_init_clocks(struct clk_mgr *clk_mgr)
{
struct clk_mgr_internal *clk_mgr_int = TO_CLK_MGR_INTERNAL(clk_mgr);
+
init_clk_states(clk_mgr);
// to adjust dp_dto reference clock if ssc is enable otherwise to apply dprefclk
@@ -632,6 +700,7 @@ static struct wm_table lpddr5_wm_table = {
};
static DpmClocks_t_dcn35 dummy_clocks;
+static DpmClocks_t_dcn351 dummy_clocks_dcn351;
static struct dcn35_watermarks dummy_wms = { 0 };
@@ -642,10 +711,10 @@ static struct dcn35_ss_info_table ss_info_table = {
static void dcn35_read_ss_info_from_lut(struct clk_mgr_internal *clk_mgr)
{
- struct dc_context *ctx = clk_mgr->base.ctx;
- uint32_t clock_source;
+ uint32_t clock_source = 0;
+
+ clock_source = REG_READ(CLK1_CLK2_BYPASS_CNTL) & CLK1_CLK2_BYPASS_CNTL__CLK2_BYPASS_SEL_MASK;
- REG_GET(CLK1_CLK2_BYPASS_CNTL, CLK2_BYPASS_SEL, &clock_source);
// If it's DFS mode, clock_source is 0.
if (dcn35_is_spll_ssc_enabled(&clk_mgr->base) && (clock_source < ARRAY_SIZE(ss_info_table.ss_percentage))) {
clk_mgr->dprefclk_ss_percentage = ss_info_table.ss_percentage[clock_source];
@@ -755,6 +824,22 @@ static void dcn35_get_dpm_table_from_smu(struct clk_mgr_internal *clk_mgr,
dcn35_smu_transfer_dpm_table_smu_2_dram(clk_mgr);
}
+static void dcn351_get_dpm_table_from_smu(struct clk_mgr_internal *clk_mgr,
+ struct dcn351_smu_dpm_clks *smu_dpm_clks)
+{
+ DpmClocks_t_dcn351 *table = smu_dpm_clks->dpm_clks;
+
+ if (!clk_mgr->smu_ver)
+ return;
+ if (!table || smu_dpm_clks->mc_address.quad_part == 0)
+ return;
+ memset(table, 0, sizeof(*table));
+ dcn35_smu_set_dram_addr_high(clk_mgr,
+ smu_dpm_clks->mc_address.high_part);
+ dcn35_smu_set_dram_addr_low(clk_mgr,
+ smu_dpm_clks->mc_address.low_part);
+ dcn35_smu_transfer_dpm_table_smu_2_dram(clk_mgr);
+}
static uint32_t find_max_clk_value(const uint32_t clocks[], uint32_t num_clocks)
{
uint32_t max = 0;
@@ -1093,6 +1178,57 @@ struct clk_mgr_funcs dcn35_fpga_funcs = {
.get_dtb_ref_clk_frequency = dcn31_get_dtb_ref_freq_khz,
};
+static void translate_to_DpmClocks_t_dcn35(struct dcn351_smu_dpm_clks *smu_dpm_clks_a,
+ struct dcn35_smu_dpm_clks *smu_dpm_clks_b)
+{
+ /*translate two structures and only take need clock tables*/
+ uint8_t i;
+
+ if (smu_dpm_clks_a == NULL || smu_dpm_clks_b == NULL ||
+ smu_dpm_clks_a->dpm_clks == NULL || smu_dpm_clks_b->dpm_clks == NULL)
+ return;
+
+ for (i = 0; i < NUM_DCFCLK_DPM_LEVELS; i++)
+ smu_dpm_clks_b->dpm_clks->DcfClocks[i] = smu_dpm_clks_a->dpm_clks->DcfClocks[i];
+
+ for (i = 0; i < NUM_DISPCLK_DPM_LEVELS; i++)
+ smu_dpm_clks_b->dpm_clks->DispClocks[i] = smu_dpm_clks_a->dpm_clks->DispClocks[i];
+
+ for (i = 0; i < NUM_DPPCLK_DPM_LEVELS; i++)
+ smu_dpm_clks_b->dpm_clks->DppClocks[i] = smu_dpm_clks_a->dpm_clks->DppClocks[i];
+
+ for (i = 0; i < NUM_FCLK_DPM_LEVELS; i++) {
+ smu_dpm_clks_b->dpm_clks->FclkClocks_Freq[i] = smu_dpm_clks_a->dpm_clks->FclkClocks_Freq[i];
+ smu_dpm_clks_b->dpm_clks->FclkClocks_Voltage[i] = smu_dpm_clks_a->dpm_clks->FclkClocks_Voltage[i];
+ }
+ for (i = 0; i < NUM_MEM_PSTATE_LEVELS; i++) {
+ smu_dpm_clks_b->dpm_clks->MemPstateTable[i].MemClk =
+ smu_dpm_clks_a->dpm_clks->MemPstateTable[i].MemClk;
+ smu_dpm_clks_b->dpm_clks->MemPstateTable[i].UClk =
+ smu_dpm_clks_a->dpm_clks->MemPstateTable[i].UClk;
+ smu_dpm_clks_b->dpm_clks->MemPstateTable[i].Voltage =
+ smu_dpm_clks_a->dpm_clks->MemPstateTable[i].Voltage;
+ smu_dpm_clks_b->dpm_clks->MemPstateTable[i].WckRatio =
+ smu_dpm_clks_a->dpm_clks->MemPstateTable[i].WckRatio;
+ }
+ smu_dpm_clks_b->dpm_clks->MaxGfxClk = smu_dpm_clks_a->dpm_clks->MaxGfxClk;
+ smu_dpm_clks_b->dpm_clks->MinGfxClk = smu_dpm_clks_a->dpm_clks->MinGfxClk;
+ smu_dpm_clks_b->dpm_clks->NumDcfClkLevelsEnabled =
+ smu_dpm_clks_a->dpm_clks->NumDcfClkLevelsEnabled;
+ smu_dpm_clks_b->dpm_clks->NumDispClkLevelsEnabled =
+ smu_dpm_clks_a->dpm_clks->NumDispClkLevelsEnabled;
+ smu_dpm_clks_b->dpm_clks->NumFclkLevelsEnabled =
+ smu_dpm_clks_a->dpm_clks->NumFclkLevelsEnabled;
+ smu_dpm_clks_b->dpm_clks->NumMemPstatesEnabled =
+ smu_dpm_clks_a->dpm_clks->NumMemPstatesEnabled;
+ smu_dpm_clks_b->dpm_clks->NumSocClkLevelsEnabled =
+ smu_dpm_clks_a->dpm_clks->NumSocClkLevelsEnabled;
+
+ for (i = 0; i < NUM_SOC_VOLTAGE_LEVELS; i++) {
+ smu_dpm_clks_b->dpm_clks->SocClocks[i] = smu_dpm_clks_a->dpm_clks->SocClocks[i];
+ smu_dpm_clks_b->dpm_clks->SocVoltage[i] = smu_dpm_clks_a->dpm_clks->SocVoltage[i];
+ }
+}
void dcn35_clk_mgr_construct(
struct dc_context *ctx,
struct clk_mgr_dcn35 *clk_mgr,
@@ -1100,6 +1236,7 @@ void dcn35_clk_mgr_construct(
struct dccg *dccg)
{
struct dcn35_smu_dpm_clks smu_dpm_clks = { 0 };
+ struct dcn351_smu_dpm_clks smu_dpm_clks_dcn351 = { 0 };
clk_mgr->base.base.ctx = ctx;
clk_mgr->base.base.funcs = &dcn35_funcs;
@@ -1112,6 +1249,12 @@ void dcn35_clk_mgr_construct(
clk_mgr->base.dprefclk_ss_divider = 1000;
clk_mgr->base.ss_on_dprefclk = false;
clk_mgr->base.dfs_ref_freq_khz = 48000;
+ if (ctx->dce_version == DCN_VERSION_3_5) {
+ clk_mgr->base.regs = &clk_mgr_regs_dcn35;
+ clk_mgr->base.clk_mgr_shift = &clk_mgr_shift_dcn35;
+ clk_mgr->base.clk_mgr_mask = &clk_mgr_mask_dcn35;
+ }
+
clk_mgr->smu_wm_set.wm_set = (struct dcn35_watermarks *)dm_helpers_allocate_gpu_mem(
clk_mgr->base.base.ctx,
@@ -1130,14 +1273,24 @@ void dcn35_clk_mgr_construct(
DC_MEM_ALLOC_TYPE_GART,
sizeof(DpmClocks_t_dcn35),
&smu_dpm_clks.mc_address.quad_part);
-
if (smu_dpm_clks.dpm_clks == NULL) {
smu_dpm_clks.dpm_clks = &dummy_clocks;
smu_dpm_clks.mc_address.quad_part = 0;
}
-
ASSERT(smu_dpm_clks.dpm_clks);
+ if (ctx->dce_version == DCN_VERSION_3_51) {
+ smu_dpm_clks_dcn351.dpm_clks = (DpmClocks_t_dcn351 *)dm_helpers_allocate_gpu_mem(
+ clk_mgr->base.base.ctx,
+ DC_MEM_ALLOC_TYPE_GART,
+ sizeof(DpmClocks_t_dcn351),
+ &smu_dpm_clks_dcn351.mc_address.quad_part);
+ if (smu_dpm_clks_dcn351.dpm_clks == NULL) {
+ smu_dpm_clks_dcn351.dpm_clks = &dummy_clocks_dcn351;
+ smu_dpm_clks_dcn351.mc_address.quad_part = 0;
+ }
+ }
+
clk_mgr->base.smu_ver = dcn35_smu_get_smu_version(&clk_mgr->base);
if (clk_mgr->base.smu_ver)
@@ -1166,7 +1319,11 @@ void dcn35_clk_mgr_construct(
if (clk_mgr->base.base.ctx->dc->debug.pstate_enabled) {
int i;
- dcn35_get_dpm_table_from_smu(&clk_mgr->base, &smu_dpm_clks);
+ if (ctx->dce_version == DCN_VERSION_3_51) {
+ dcn351_get_dpm_table_from_smu(&clk_mgr->base, &smu_dpm_clks_dcn351);
+ translate_to_DpmClocks_t_dcn35(&smu_dpm_clks_dcn351, &smu_dpm_clks);
+ } else
+ dcn35_get_dpm_table_from_smu(&clk_mgr->base, &smu_dpm_clks);
DC_LOG_SMU("NumDcfClkLevelsEnabled: %d\n"
"NumDispClkLevelsEnabled: %d\n"
"NumSocClkLevelsEnabled: %d\n"
@@ -1227,6 +1384,10 @@ void dcn35_clk_mgr_construct(
dm_helpers_free_gpu_mem(clk_mgr->base.base.ctx, DC_MEM_ALLOC_TYPE_GART,
smu_dpm_clks.dpm_clks);
+ if (smu_dpm_clks_dcn351.dpm_clks && smu_dpm_clks_dcn351.mc_address.quad_part != 0)
+ dm_helpers_free_gpu_mem(clk_mgr->base.base.ctx, DC_MEM_ALLOC_TYPE_GART,
+ smu_dpm_clks_dcn351.dpm_clks);
+
if (ctx->dc->config.disable_ips != DMUB_IPS_DISABLE_ALL) {
bool ips_support = false;
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.h
index 1203dc605b12..a12a9bf90806 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.h
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.h
@@ -60,4 +60,8 @@ void dcn35_clk_mgr_construct(struct dc_context *ctx,
void dcn35_clk_mgr_destroy(struct clk_mgr_internal *clk_mgr_int);
+void dcn351_clk_mgr_construct(struct dc_context *ctx,
+ struct clk_mgr_dcn35 *clk_mgr,
+ struct pp_smu_funcs *pp_smu,
+ struct dccg *dccg);
#endif //__DCN35_CLK_MGR_H__
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_smu.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_smu.h
index 3fae13c73934..ab9d21ba0c43 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_smu.h
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_smu.h
@@ -126,18 +126,31 @@ typedef struct {
uint32_t MaxGfxClk;
} DpmClocks_t_dcn35;
-
-// Throttler Status Bitmask
-
-
-
-
-
-
-
-
-
-
+typedef struct {
+ uint32_t DcfClocks[NUM_DCFCLK_DPM_LEVELS];
+ uint32_t DispClocks[NUM_DISPCLK_DPM_LEVELS];
+ uint32_t DppClocks[NUM_DPPCLK_DPM_LEVELS];
+ uint32_t SocClocks[NUM_SOCCLK_DPM_LEVELS];
+ uint32_t VClocks0[NUM_VCN_DPM_LEVELS];
+ uint32_t VClocks1[NUM_VCN_DPM_LEVELS];
+ uint32_t DClocks0[NUM_VCN_DPM_LEVELS];
+ uint32_t DClocks1[NUM_VCN_DPM_LEVELS];
+ uint32_t VPEClocks[NUM_VPE_DPM_LEVELS];
+ uint32_t FclkClocks_Freq[NUM_FCLK_DPM_LEVELS];
+ uint32_t FclkClocks_Voltage[NUM_FCLK_DPM_LEVELS];
+ uint32_t SocVoltage[NUM_SOC_VOLTAGE_LEVELS];
+ MemPstateTable_t MemPstateTable[NUM_MEM_PSTATE_LEVELS];
+ uint8_t NumDcfClkLevelsEnabled;
+ uint8_t NumDispClkLevelsEnabled; // Applies to both Dispclk and Dppclk
+ uint8_t NumSocClkLevelsEnabled;
+ uint8_t Vcn0ClkLevelsEnabled; // Applies to both Vclk0 and Dclk0
+ uint8_t Vcn1ClkLevelsEnabled; // Applies to both Vclk1 and Dclk1
+ uint8_t VpeClkLevelsEnabled;
+ uint8_t NumMemPstatesEnabled;
+ uint8_t NumFclkLevelsEnabled;
+ uint32_t MinGfxClk;
+ uint32_t MaxGfxClk;
+} DpmClocks_t_dcn351;
#define TABLE_BIOS_IF 0 // Called by BIOS
#define TABLE_WATERMARKS 1 // Called by DAL through VBIOS
@@ -163,6 +176,10 @@ struct dcn35_smu_dpm_clks {
union large_integer mc_address;
};
+struct dcn351_smu_dpm_clks {
+ DpmClocks_t_dcn351 *dpm_clks;
+ union large_integer mc_address;
+};
/* TODO: taken from vgh, may not be correct */
struct display_idle_optimization {
unsigned int df_request_disabled : 1;
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dalsmc.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dalsmc.h
index dbfdd3487da5..2e0d34fd7512 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dalsmc.h
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dalsmc.h
@@ -43,7 +43,9 @@
#define DALSMC_MSG_ActiveUclkFclk 0x18
#define DALSMC_MSG_IdleUclkFclk 0x19
#define DALSMC_MSG_SetUclkPstateAllow 0x1A
-#define DALSMC_Message_Count 0x1B
+#define DALSMC_MSG_SubvpUclkFclk 0x1B
+#define DALSMC_MSG_GetNumUmcChannels 0x1C
+#define DALSMC_Message_Count 0x1D
typedef enum {
FCLK_SWITCH_DISALLOW,
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr.c
index 8cfc5f435937..8082bb877611 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr.c
@@ -141,6 +141,20 @@ static bool dcn401_is_ppclk_idle_dpm_enabled(struct clk_mgr_internal *clk_mgr, P
return ppclk_idle_dpm_enabled;
}
+static bool dcn401_is_df_throttle_opt_enabled(struct clk_mgr_internal *clk_mgr)
+{
+ bool is_df_throttle_opt_enabled = false;
+
+ if (ASICREV_IS_GC_12_0_1_A0(clk_mgr->base.ctx->asic_id.hw_internal_rev) &&
+ clk_mgr->smu_ver >= 0x663500) {
+ is_df_throttle_opt_enabled = !clk_mgr->base.ctx->dc->debug.force_subvp_df_throttle;
+ }
+
+ is_df_throttle_opt_enabled &= clk_mgr->smu_present;
+
+ return is_df_throttle_opt_enabled;
+}
+
/* Query SMU for all clock states for a particular clock */
static void dcn401_init_single_clock(struct clk_mgr_internal *clk_mgr, PPCLK_e clk, unsigned int *entry_0,
unsigned int *num_levels)
@@ -614,207 +628,6 @@ static void dcn401_update_clocks_update_dentist(
}
-static void dcn401_update_clocks_legacy(struct clk_mgr *clk_mgr_base,
- struct dc_state *context,
- bool safe_to_lower)
-{
- struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
- struct dc_clocks *new_clocks = &context->bw_ctx.bw.dcn.clk;
- struct dc *dc = clk_mgr_base->ctx->dc;
- int display_count;
- bool update_dppclk = false;
- bool update_dispclk = false;
- bool enter_display_off = false;
- bool dpp_clock_lowered = false;
- struct dmcu *dmcu = clk_mgr_base->ctx->dc->res_pool->dmcu;
- bool force_reset = false;
- bool update_uclk = false, update_fclk = false;
- bool p_state_change_support;
- bool fclk_p_state_change_support;
- int total_plane_count;
-
- if (dc->work_arounds.skip_clock_update)
- return;
-
- if (clk_mgr_base->clks.dispclk_khz == 0 ||
- (dc->debug.force_clock_mode & 0x1)) {
- /* This is from resume or boot up, if forced_clock cfg option used,
- * we bypass program dispclk and DPPCLK, but need set them for S3.
- */
- force_reset = true;
-
- dcn2_read_clocks_from_hw_dentist(clk_mgr_base);
-
- /* Force_clock_mode 0x1: force reset the clock even it is the same clock
- * as long as it is in Passive level.
- */
- }
- display_count = clk_mgr_helper_get_active_display_cnt(dc, context);
-
- if (display_count == 0)
- enter_display_off = true;
-
- if (clk_mgr->smu_present) {
- if (enter_display_off == safe_to_lower)
- dcn401_smu_set_num_of_displays(clk_mgr, display_count);
-
- clk_mgr_base->clks.fclk_prev_p_state_change_support = clk_mgr_base->clks.fclk_p_state_change_support;
-
- total_plane_count = clk_mgr_helper_get_active_plane_cnt(dc, context);
- fclk_p_state_change_support = new_clocks->fclk_p_state_change_support || (total_plane_count == 0);
-
- if (should_update_pstate_support(safe_to_lower, fclk_p_state_change_support, clk_mgr_base->clks.fclk_p_state_change_support)) {
- clk_mgr_base->clks.fclk_p_state_change_support = fclk_p_state_change_support;
-
- /* To enable FCLK P-state switching, send PSTATE_SUPPORTED message to PMFW */
- if (clk_mgr_base->clks.fclk_p_state_change_support) {
- /* Handle the code for sending a message to PMFW that FCLK P-state change is supported */
- dcn401_smu_send_fclk_pstate_message(clk_mgr, true);
- }
- }
-
- if (dc->debug.force_min_dcfclk_mhz > 0)
- new_clocks->dcfclk_khz = (new_clocks->dcfclk_khz > (dc->debug.force_min_dcfclk_mhz * 1000)) ?
- new_clocks->dcfclk_khz : (dc->debug.force_min_dcfclk_mhz * 1000);
-
- if (should_set_clock(safe_to_lower, new_clocks->dcfclk_khz, clk_mgr_base->clks.dcfclk_khz)) {
- clk_mgr_base->clks.dcfclk_khz = new_clocks->dcfclk_khz;
- if (dcn401_is_ppclk_dpm_enabled(clk_mgr, PPCLK_DCFCLK))
- dcn401_smu_set_hard_min_by_freq(clk_mgr, PPCLK_DCFCLK, khz_to_mhz_ceil(clk_mgr_base->clks.dcfclk_khz));
- }
-
- if (should_set_clock(safe_to_lower, new_clocks->dcfclk_deep_sleep_khz, clk_mgr_base->clks.dcfclk_deep_sleep_khz)) {
- clk_mgr_base->clks.dcfclk_deep_sleep_khz = new_clocks->dcfclk_deep_sleep_khz;
- if (dcn401_is_ppclk_dpm_enabled(clk_mgr, PPCLK_DCFCLK))
- dcn401_smu_set_min_deep_sleep_dcef_clk(clk_mgr, khz_to_mhz_ceil(clk_mgr_base->clks.dcfclk_deep_sleep_khz));
- }
-
- if (should_set_clock(safe_to_lower, new_clocks->socclk_khz, clk_mgr_base->clks.socclk_khz))
- /* We don't actually care about socclk, don't notify SMU of hard min */
- clk_mgr_base->clks.socclk_khz = new_clocks->socclk_khz;
-
- clk_mgr_base->clks.prev_p_state_change_support = clk_mgr_base->clks.p_state_change_support;
- clk_mgr_base->clks.prev_num_ways = clk_mgr_base->clks.num_ways;
-
- if (clk_mgr_base->clks.num_ways != new_clocks->num_ways &&
- clk_mgr_base->clks.num_ways < new_clocks->num_ways) {
- clk_mgr_base->clks.num_ways = new_clocks->num_ways;
- if (dcn401_is_ppclk_dpm_enabled(clk_mgr, PPCLK_UCLK))
- dcn401_smu_send_cab_for_uclk_message(clk_mgr, clk_mgr_base->clks.num_ways);
- }
-
-
- p_state_change_support = new_clocks->p_state_change_support || (total_plane_count == 0);
- if (should_update_pstate_support(safe_to_lower, p_state_change_support, clk_mgr_base->clks.prev_p_state_change_support)) {
- clk_mgr_base->clks.p_state_change_support = p_state_change_support;
- clk_mgr_base->clks.fw_based_mclk_switching = p_state_change_support && new_clocks->fw_based_mclk_switching;
-
- /* to disable P-State switching, set UCLK min = max */
- if (!clk_mgr_base->clks.p_state_change_support && dcn401_is_ppclk_dpm_enabled(clk_mgr, PPCLK_UCLK))
- dcn401_smu_set_hard_min_by_freq(clk_mgr, PPCLK_UCLK,
- clk_mgr_base->bw_params->clk_table.entries[clk_mgr_base->bw_params->clk_table.num_entries_per_clk.num_memclk_levels - 1].memclk_mhz);
- }
-
- /* Always update saved value, even if new value not set due to P-State switching unsupported. Also check safe_to_lower for FCLK */
- if (safe_to_lower && (clk_mgr_base->clks.fclk_p_state_change_support != clk_mgr_base->clks.fclk_prev_p_state_change_support)) {
- update_fclk = true;
- }
-
- if (!clk_mgr_base->clks.fclk_p_state_change_support &&
- update_fclk &&
- dcn401_is_ppclk_dpm_enabled(clk_mgr, PPCLK_FCLK)) {
- /* Handle code for sending a message to PMFW that FCLK P-state change is not supported */
- dcn401_smu_send_fclk_pstate_message(clk_mgr, false);
- }
-
- /* Always update saved value, even if new value not set due to P-State switching unsupported */
- if (should_set_clock(safe_to_lower, new_clocks->dramclk_khz, clk_mgr_base->clks.dramclk_khz)) {
- clk_mgr_base->clks.dramclk_khz = new_clocks->dramclk_khz;
- update_uclk = true;
- }
-
- /* set UCLK to requested value if P-State switching is supported, or to re-enable P-State switching */
- if (clk_mgr_base->clks.p_state_change_support &&
- (update_uclk || !clk_mgr_base->clks.prev_p_state_change_support) &&
- dcn401_is_ppclk_dpm_enabled(clk_mgr, PPCLK_UCLK))
- dcn401_smu_set_hard_min_by_freq(clk_mgr, PPCLK_UCLK, khz_to_mhz_ceil(clk_mgr_base->clks.dramclk_khz));
-
- if (clk_mgr_base->clks.num_ways != new_clocks->num_ways &&
- clk_mgr_base->clks.num_ways > new_clocks->num_ways) {
- clk_mgr_base->clks.num_ways = new_clocks->num_ways;
- if (dcn401_is_ppclk_dpm_enabled(clk_mgr, PPCLK_UCLK))
- dcn401_smu_send_cab_for_uclk_message(clk_mgr, clk_mgr_base->clks.num_ways);
- }
- }
-
- if (should_set_clock(safe_to_lower, new_clocks->dppclk_khz, clk_mgr_base->clks.dppclk_khz)) {
- if (clk_mgr_base->clks.dppclk_khz > new_clocks->dppclk_khz)
- dpp_clock_lowered = true;
-
- clk_mgr_base->clks.dppclk_khz = new_clocks->dppclk_khz;
- clk_mgr_base->clks.actual_dppclk_khz = new_clocks->dppclk_khz;
-
- if (clk_mgr->smu_present && !dpp_clock_lowered && dcn401_is_ppclk_dpm_enabled(clk_mgr, PPCLK_DPPCLK))
- clk_mgr_base->clks.actual_dppclk_khz = dcn401_set_hard_min_by_freq_optimized(clk_mgr, PPCLK_DPPCLK, clk_mgr_base->clks.dppclk_khz);
- update_dppclk = true;
- }
-
- if (should_set_clock(safe_to_lower, new_clocks->dispclk_khz, clk_mgr_base->clks.dispclk_khz)) {
- clk_mgr_base->clks.dispclk_khz = new_clocks->dispclk_khz;
-
- if (clk_mgr->smu_present && dcn401_is_ppclk_dpm_enabled(clk_mgr, PPCLK_DISPCLK))
- clk_mgr_base->clks.actual_dispclk_khz = dcn401_set_hard_min_by_freq_optimized(clk_mgr, PPCLK_DISPCLK, clk_mgr_base->clks.dispclk_khz);
-
- update_dispclk = true;
- }
-
- if (!new_clocks->dtbclk_en && dcn401_is_ppclk_dpm_enabled(clk_mgr, PPCLK_DTBCLK)) {
- new_clocks->ref_dtbclk_khz = clk_mgr_base->bw_params->clk_table.entries[0].dtbclk_mhz * 1000;
- }
-
- /* clock limits are received with MHz precision, divide by 1000 to prevent setting clocks at every call */
- if (!dc->debug.disable_dtb_ref_clk_switch &&
- should_set_clock(safe_to_lower, new_clocks->ref_dtbclk_khz / 1000, clk_mgr_base->clks.ref_dtbclk_khz / 1000) &&
- dcn401_is_ppclk_dpm_enabled(clk_mgr, PPCLK_DTBCLK)) {
- /* DCCG requires KHz precision for DTBCLK */
- clk_mgr_base->clks.ref_dtbclk_khz =
- dcn401_smu_set_hard_min_by_freq(clk_mgr, PPCLK_DTBCLK, khz_to_mhz_ceil(new_clocks->ref_dtbclk_khz));
-
- dcn401_update_clocks_update_dtb_dto(clk_mgr, context, clk_mgr_base->clks.ref_dtbclk_khz);
- }
-
- if (dc->config.forced_clocks == false || (force_reset && safe_to_lower)) {
- if (dpp_clock_lowered) {
- /* if clock is being lowered, increase DTO before lowering refclk */
- dcn401_update_clocks_update_dpp_dto(clk_mgr, context,
- safe_to_lower, clk_mgr_base->clks.dppclk_khz);
- dcn401_update_clocks_update_dentist(clk_mgr, context);
- if (clk_mgr->smu_present && dcn401_is_ppclk_dpm_enabled(clk_mgr, PPCLK_DPPCLK)) {
- clk_mgr_base->clks.actual_dppclk_khz = dcn401_set_hard_min_by_freq_optimized(clk_mgr, PPCLK_DPPCLK,
- clk_mgr_base->clks.dppclk_khz);
- dcn401_update_clocks_update_dpp_dto(clk_mgr, context, safe_to_lower,
- clk_mgr_base->clks.actual_dppclk_khz);
- }
-
- } else {
- /* if clock is being raised, increase refclk before lowering DTO */
- if (update_dppclk || update_dispclk)
- dcn401_update_clocks_update_dentist(clk_mgr, context);
- /* There is a check inside dcn20_update_clocks_update_dpp_dto which ensures
- * that we do not lower dto when it is not safe to lower. We do not need to
- * compare the current and new dppclk before calling this function.
- */
- dcn401_update_clocks_update_dpp_dto(clk_mgr, context,
- safe_to_lower, clk_mgr_base->clks.actual_dppclk_khz);
- }
- }
-
- if (update_dispclk && dmcu && dmcu->funcs->is_dmcu_initialized(dmcu))
- /*update dmcu for wait_loop count*/
- dmcu->funcs->set_psr_wait_loop(dmcu,
- clk_mgr_base->clks.dispclk_khz / 1000 / 7);
-}
-
static void dcn401_execute_block_sequence(struct clk_mgr *clk_mgr_base, unsigned int num_steps)
{
struct clk_mgr_internal *clk_mgr_internal = TO_CLK_MGR_INTERNAL(clk_mgr_base);
@@ -869,6 +682,12 @@ static void dcn401_execute_block_sequence(struct clk_mgr *clk_mgr_base, unsigned
params->update_idle_hardmin_params.uclk_mhz,
params->update_idle_hardmin_params.fclk_mhz);
break;
+ case CLK_MGR401_UPDATE_SUBVP_HARDMINS:
+ dcn401_smu_set_subvp_uclk_fclk_hardmin(
+ clk_mgr_internal,
+ params->update_idle_hardmin_params.uclk_mhz,
+ params->update_idle_hardmin_params.fclk_mhz);
+ break;
case CLK_MGR401_UPDATE_DEEP_SLEEP_DCFCLK:
dcn401_smu_set_min_deep_sleep_dcef_clk(
clk_mgr_internal,
@@ -945,15 +764,21 @@ static unsigned int dcn401_build_update_bandwidth_clocks_sequence(
bool update_active_uclk = false;
bool update_idle_fclk = false;
bool update_idle_uclk = false;
+ bool update_subvp_prefetch_dramclk = false;
+ bool update_subvp_prefetch_fclk = false;
bool is_idle_dpm_enabled = dcn401_is_ppclk_dpm_enabled(clk_mgr_internal, PPCLK_UCLK) &&
dcn401_is_ppclk_dpm_enabled(clk_mgr_internal, PPCLK_FCLK) &&
dcn401_is_ppclk_idle_dpm_enabled(clk_mgr_internal, PPCLK_UCLK) &&
dcn401_is_ppclk_idle_dpm_enabled(clk_mgr_internal, PPCLK_FCLK);
+ bool is_df_throttle_opt_enabled = is_idle_dpm_enabled &&
+ dcn401_is_df_throttle_opt_enabled(clk_mgr_internal);
int total_plane_count = clk_mgr_helper_get_active_plane_cnt(dc, context);
int active_uclk_mhz = khz_to_mhz_ceil(clk_mgr_base->clks.dramclk_khz);
int active_fclk_mhz = khz_to_mhz_ceil(clk_mgr_base->clks.fclk_khz);
int idle_uclk_mhz = khz_to_mhz_ceil(clk_mgr_base->clks.idle_dramclk_khz);
int idle_fclk_mhz = khz_to_mhz_ceil(clk_mgr_base->clks.idle_fclk_khz);
+ int subvp_prefetch_dramclk_mhz = khz_to_mhz_ceil(clk_mgr_base->clks.subvp_prefetch_dramclk_khz);
+ int subvp_prefetch_fclk_mhz = khz_to_mhz_ceil(clk_mgr_base->clks.subvp_prefetch_fclk_khz);
unsigned int num_steps = 0;
@@ -982,15 +807,15 @@ static unsigned int dcn401_build_update_bandwidth_clocks_sequence(
update_active_fclk = true;
update_idle_fclk = true;
- /* To enable FCLK P-state switching, send PSTATE_SUPPORTED message to PMFW */
- if (clk_mgr_base->clks.fclk_p_state_change_support) {
- /* Handle the code for sending a message to PMFW that FCLK P-state change is supported */
- if (dcn401_is_ppclk_dpm_enabled(clk_mgr_internal, PPCLK_FCLK)) {
- block_sequence[num_steps].params.update_pstate_support_params.support = true;
- block_sequence[num_steps].func = CLK_MGR401_UPDATE_FCLK_PSTATE_SUPPORT;
- num_steps++;
- }
- }
+ /* To enable FCLK P-state switching, send PSTATE_SUPPORTED message to PMFW (message not supported on DCN401)*/
+ // if (clk_mgr_base->clks.fclk_p_state_change_support) {
+ // /* Handle the code for sending a message to PMFW that FCLK P-state change is supported */
+ // if (dcn401_is_ppclk_dpm_enabled(clk_mgr_internal, PPCLK_FCLK)) {
+ // block_sequence[num_steps].params.update_pstate_support_params.support = true;
+ // block_sequence[num_steps].func = CLK_MGR401_UPDATE_FCLK_PSTATE_SUPPORT;
+ // num_steps++;
+ // }
+ // }
}
if (!clk_mgr_base->clks.fclk_p_state_change_support && dcn401_is_ppclk_dpm_enabled(clk_mgr_internal, PPCLK_FCLK)) {
@@ -1109,6 +934,12 @@ static unsigned int dcn401_build_update_bandwidth_clocks_sequence(
}
}
+ if (should_set_clock(safe_to_lower, new_clocks->subvp_prefetch_dramclk_khz, clk_mgr_base->clks.subvp_prefetch_dramclk_khz)) {
+ clk_mgr_base->clks.subvp_prefetch_dramclk_khz = new_clocks->subvp_prefetch_dramclk_khz;
+ update_subvp_prefetch_dramclk = true;
+ subvp_prefetch_dramclk_mhz = khz_to_mhz_ceil(clk_mgr_base->clks.subvp_prefetch_dramclk_khz);
+ }
+
/* FCLK */
/* Always update saved value, even if new value not set due to P-State switching unsupported */
if (should_set_clock(safe_to_lower, new_clocks->fclk_khz, clk_mgr_base->clks.fclk_khz)) {
@@ -1129,6 +960,12 @@ static unsigned int dcn401_build_update_bandwidth_clocks_sequence(
}
}
+ if (should_set_clock(safe_to_lower, new_clocks->subvp_prefetch_fclk_khz, clk_mgr_base->clks.subvp_prefetch_fclk_khz)) {
+ clk_mgr_base->clks.subvp_prefetch_fclk_khz = new_clocks->subvp_prefetch_fclk_khz;
+ update_subvp_prefetch_fclk = true;
+ subvp_prefetch_fclk_mhz = khz_to_mhz_ceil(clk_mgr_base->clks.subvp_prefetch_fclk_khz);
+ }
+
/* When idle DPM is enabled, need to send active and idle hardmins separately */
/* CLK_MGR401_UPDATE_ACTIVE_HARDMINS */
if ((update_active_uclk || update_active_fclk) && is_idle_dpm_enabled) {
@@ -1146,6 +983,14 @@ static unsigned int dcn401_build_update_bandwidth_clocks_sequence(
num_steps++;
}
+ /* CLK_MGR401_UPDATE_SUBVP_HARDMINS */
+ if ((update_subvp_prefetch_dramclk || update_subvp_prefetch_fclk) && is_df_throttle_opt_enabled) {
+ block_sequence[num_steps].params.update_idle_hardmin_params.uclk_mhz = subvp_prefetch_dramclk_mhz;
+ block_sequence[num_steps].params.update_idle_hardmin_params.fclk_mhz = subvp_prefetch_fclk_mhz;
+ block_sequence[num_steps].func = CLK_MGR401_UPDATE_SUBVP_HARDMINS;
+ num_steps++;
+ }
+
/* set UCLK to requested value if P-State switching is supported, or to re-enable P-State switching */
if (update_active_uclk || update_idle_uclk) {
if (!is_idle_dpm_enabled) {
@@ -1178,14 +1023,14 @@ static unsigned int dcn401_build_update_bandwidth_clocks_sequence(
// (*num_steps)++;
// }
- /* disable FCLK P-State support if needed */
- if (!fclk_p_state_change_support &&
- should_update_pstate_support(safe_to_lower, fclk_p_state_change_support, clk_mgr_base->clks.fclk_prev_p_state_change_support) &&
- dcn401_is_ppclk_dpm_enabled(clk_mgr_internal, PPCLK_FCLK)) {
- block_sequence[num_steps].params.update_pstate_support_params.support = false;
- block_sequence[num_steps].func = CLK_MGR401_UPDATE_FCLK_PSTATE_SUPPORT;
- num_steps++;
- }
+ /* disable FCLK P-State support if needed (message not supported on DCN401)*/
+ // if (!fclk_p_state_change_support &&
+ // should_update_pstate_support(safe_to_lower, fclk_p_state_change_support, clk_mgr_base->clks.fclk_prev_p_state_change_support) &&
+ // dcn401_is_ppclk_dpm_enabled(clk_mgr_internal, PPCLK_FCLK)) {
+ // block_sequence[num_steps].params.update_pstate_support_params.support = false;
+ // block_sequence[num_steps].func = CLK_MGR401_UPDATE_FCLK_PSTATE_SUPPORT;
+ // num_steps++;
+ // }
}
if (new_clocks->fw_based_mclk_switching != clk_mgr_base->clks.fw_based_mclk_switching &&
@@ -1366,11 +1211,6 @@ static void dcn401_update_clocks(struct clk_mgr *clk_mgr_base,
unsigned int num_steps = 0;
- if (dc->debug.enable_legacy_clock_update) {
- dcn401_update_clocks_legacy(clk_mgr_base, context, safe_to_lower);
- return;
- }
-
/* build bandwidth related clocks update sequence */
num_steps = dcn401_build_update_bandwidth_clocks_sequence(clk_mgr_base,
context,
@@ -1505,6 +1345,20 @@ static void dcn401_set_hard_min_memclk(struct clk_mgr *clk_mgr_base, bool curren
dcn401_execute_block_sequence(clk_mgr_base, num_steps);
}
+static int dcn401_get_hard_min_memclk(struct clk_mgr *clk_mgr_base)
+{
+ struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+
+ return clk_mgr->base.ctx->dc->current_state->bw_ctx.bw.dcn.clk.dramclk_khz;
+}
+
+static int dcn401_get_hard_min_fclk(struct clk_mgr *clk_mgr_base)
+{
+ struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+
+ return clk_mgr->base.ctx->dc->current_state->bw_ctx.bw.dcn.clk.fclk_khz;
+}
+
/* Get current memclk states, update bounding box */
static void dcn401_get_memclk_states_from_smu(struct clk_mgr *clk_mgr_base)
{
@@ -1549,6 +1403,15 @@ static void dcn401_get_memclk_states_from_smu(struct clk_mgr *clk_mgr_base)
if (clk_mgr->dpm_present && !num_levels)
clk_mgr->dpm_present = false;
+ clk_mgr_base->bw_params->num_channels = dcn401_smu_get_num_of_umc_channels(clk_mgr);
+ if (clk_mgr_base->ctx->dc_bios) {
+ /* use BIOS values if none provided by PMFW */
+ if (clk_mgr_base->bw_params->num_channels == 0) {
+ clk_mgr_base->bw_params->num_channels = clk_mgr_base->ctx->dc_bios->vram_info.num_chans;
+ }
+ clk_mgr_base->bw_params->dram_channel_width_bytes = clk_mgr_base->ctx->dc_bios->vram_info.dram_channel_width_bytes;
+ }
+
/* Refresh bounding box */
clk_mgr_base->ctx->dc->res_pool->funcs->update_bw_bounding_box(
clk_mgr->base.ctx->dc, clk_mgr_base->bw_params);
@@ -1638,6 +1501,8 @@ static struct clk_mgr_funcs dcn401_funcs = {
.enable_pme_wa = dcn401_enable_pme_wa,
.is_smu_present = dcn401_is_smu_present,
.get_dispclk_from_dentist = dcn401_get_dispclk_from_dentist,
+ .get_hard_min_memclk = dcn401_get_hard_min_memclk,
+ .get_hard_min_fclk = dcn401_get_hard_min_fclk,
};
struct clk_mgr_internal *dcn401_clk_mgr_construct(
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr.h
index 8b0461992b22..6c9ae5ca2c7e 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr.h
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr.h
@@ -90,6 +90,7 @@ enum dcn401_clk_mgr_block_sequence_func {
CLK_MGR401_UPDATE_DTBCLK_DTO,
CLK_MGR401_UPDATE_DENTIST,
CLK_MGR401_UPDATE_PSR_WAIT_LOOP,
+ CLK_MGR401_UPDATE_SUBVP_HARDMINS,
};
struct dcn401_clk_mgr_block_sequence {
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr_smu_msg.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr_smu_msg.c
index 7700477d019b..21c35528f61f 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr_smu_msg.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr_smu_msg.c
@@ -21,6 +21,14 @@
#define smu_print(str, ...) {DC_LOG_SMU(str, ##__VA_ARGS__); }
+/* temporary define */
+#ifndef DALSMC_MSG_SubvpUclkFclk
+#define DALSMC_MSG_SubvpUclkFclk 0x1B
+#endif
+#ifndef DALSMC_MSG_GetNumUmcChannels
+#define DALSMC_MSG_GetNumUmcChannels 0x1C
+#endif
+
/*
* Function to be used instead of REG_WAIT macro because the wait ends when
* the register is NOT EQUAL to zero, and because the translation in msg_if.h
@@ -296,6 +304,24 @@ bool dcn401_smu_set_active_uclk_fclk_hardmin(struct clk_mgr_internal *clk_mgr,
return success;
}
+bool dcn401_smu_set_subvp_uclk_fclk_hardmin(struct clk_mgr_internal *clk_mgr,
+ uint16_t uclk_freq_mhz,
+ uint16_t fclk_freq_mhz)
+{
+ uint32_t response = 0;
+ bool success;
+
+ /* 15:0 for uclk, 32:16 for fclk */
+ uint32_t param = (fclk_freq_mhz << 16) | uclk_freq_mhz;
+
+ smu_print("SMU Set active hardmin by freq: uclk_freq_mhz = %d MHz, fclk_freq_mhz = %d MHz\n", uclk_freq_mhz, fclk_freq_mhz);
+
+ success = dcn401_smu_send_msg_with_param(clk_mgr,
+ DALSMC_MSG_SubvpUclkFclk, param, &response);
+
+ return success;
+}
+
void dcn401_smu_set_min_deep_sleep_dcef_clk(struct clk_mgr_internal *clk_mgr, uint32_t freq_mhz)
{
smu_print("SMU Set min deep sleep dcef clk: freq_mhz = %d MHz\n", freq_mhz);
@@ -311,3 +337,14 @@ void dcn401_smu_set_num_of_displays(struct clk_mgr_internal *clk_mgr, uint32_t n
dcn401_smu_send_msg_with_param(clk_mgr,
DALSMC_MSG_NumOfDisplays, num_displays, NULL);
}
+
+unsigned int dcn401_smu_get_num_of_umc_channels(struct clk_mgr_internal *clk_mgr)
+{
+ unsigned int response = 0;
+
+ dcn401_smu_send_msg_with_param(clk_mgr, DALSMC_MSG_GetNumUmcChannels, 0, &response);
+
+ smu_print("SMU Get Num UMC Channels: num_umc_channels = %d\n", response);
+
+ return response;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr_smu_msg.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr_smu_msg.h
index 651fb8d62864..e02eb1294b37 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr_smu_msg.h
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr_smu_msg.h
@@ -23,7 +23,11 @@ bool dcn401_smu_set_idle_uclk_fclk_hardmin(struct clk_mgr_internal *clk_mgr,
bool dcn401_smu_set_active_uclk_fclk_hardmin(struct clk_mgr_internal *clk_mgr,
uint16_t uclk_freq_mhz,
uint16_t fclk_freq_mhz);
+bool dcn401_smu_set_subvp_uclk_fclk_hardmin(struct clk_mgr_internal *clk_mgr,
+ uint16_t uclk_freq_mhz,
+ uint16_t fclk_freq_mhz);
void dcn401_smu_set_min_deep_sleep_dcef_clk(struct clk_mgr_internal *clk_mgr, uint32_t freq_mhz);
void dcn401_smu_set_num_of_displays(struct clk_mgr_internal *clk_mgr, uint32_t num_displays);
+unsigned int dcn401_smu_get_num_of_umc_channels(struct clk_mgr_internal *clk_mgr);
#endif /* __DCN401_CLK_MGR_SMU_MSG_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c
index 1dd26d5df6b9..cecaadf741ad 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc.c
@@ -579,7 +579,7 @@ dc_stream_forward_dmcu_crc_window(struct dmcu *dmcu,
bool
dc_stream_forward_crc_window(struct dc_stream_state *stream,
- struct rect *rect, bool is_stop)
+ struct rect *rect, uint8_t phy_id, bool is_stop)
{
struct dmcu *dmcu;
struct dc_dmub_srv *dmub_srv;
@@ -598,7 +598,7 @@ dc_stream_forward_crc_window(struct dc_stream_state *stream,
if (i == MAX_PIPES)
return false;
- mux_mapping.phy_output_num = stream->link->link_enc_hw_inst;
+ mux_mapping.phy_output_num = phy_id;
mux_mapping.otg_output_num = pipe->stream_res.tg->inst;
dmcu = dc->res_pool->dmcu;
@@ -615,6 +615,68 @@ dc_stream_forward_crc_window(struct dc_stream_state *stream,
return true;
}
+
+static void
+dc_stream_forward_dmub_multiple_crc_window(struct dc_dmub_srv *dmub_srv,
+ struct crc_window *window, struct otg_phy_mux *mux_mapping, bool stop)
+{
+ int i;
+ union dmub_rb_cmd cmd = {0};
+
+ cmd.secure_display.mul_roi_ctl.phy_id = mux_mapping->phy_output_num;
+ cmd.secure_display.mul_roi_ctl.otg_id = mux_mapping->otg_output_num;
+
+ cmd.secure_display.header.type = DMUB_CMD__SECURE_DISPLAY;
+
+ if (stop) {
+ cmd.secure_display.header.sub_type = DMUB_CMD__SECURE_DISPLAY_MULTIPLE_CRC_STOP_UPDATE;
+ } else {
+ cmd.secure_display.header.sub_type = DMUB_CMD__SECURE_DISPLAY_MULTIPLE_CRC_WIN_NOTIFY;
+ for (i = 0; i < MAX_CRC_WINDOW_NUM; i++) {
+ cmd.secure_display.mul_roi_ctl.roi_ctl[i].x_start = window[i].rect.x;
+ cmd.secure_display.mul_roi_ctl.roi_ctl[i].y_start = window[i].rect.y;
+ cmd.secure_display.mul_roi_ctl.roi_ctl[i].x_end = window[i].rect.x + window[i].rect.width;
+ cmd.secure_display.mul_roi_ctl.roi_ctl[i].y_end = window[i].rect.y + window[i].rect.height;
+ cmd.secure_display.mul_roi_ctl.roi_ctl[i].enable = window[i].enable;
+ }
+ }
+
+ dc_wake_and_execute_dmub_cmd(dmub_srv->ctx, &cmd, DM_DMUB_WAIT_TYPE_NO_WAIT);
+}
+
+bool
+dc_stream_forward_multiple_crc_window(struct dc_stream_state *stream,
+ struct crc_window *window, uint8_t phy_id, bool stop)
+{
+ struct dc_dmub_srv *dmub_srv;
+ struct otg_phy_mux mux_mapping;
+ struct pipe_ctx *pipe;
+ int i;
+ struct dc *dc = stream->ctx->dc;
+
+ for (i = 0; i < MAX_PIPES; i++) {
+ pipe = &dc->current_state->res_ctx.pipe_ctx[i];
+ if (pipe->stream == stream && !pipe->top_pipe && !pipe->prev_odm_pipe)
+ break;
+ }
+
+ /* Stream not found */
+ if (i == MAX_PIPES)
+ return false;
+
+ mux_mapping.phy_output_num = phy_id;
+ mux_mapping.otg_output_num = pipe->stream_res.tg->inst;
+
+ dmub_srv = dc->ctx->dmub_srv;
+
+ /* forward to dmub only. no dmcu support*/
+ if (dmub_srv)
+ dc_stream_forward_dmub_multiple_crc_window(dmub_srv, window, &mux_mapping, stop);
+ else
+ return false;
+
+ return true;
+}
#endif /* CONFIG_DRM_AMD_SECURE_DISPLAY */
/**
@@ -625,15 +687,17 @@ dc_stream_forward_crc_window(struct dc_stream_state *stream,
* @enable: Enable CRC if true, disable otherwise.
* @continuous: Capture CRC on every frame if true. Otherwise, only capture
* once.
+ * @idx: Capture CRC on which CRC engine instance
+ * @reset: Reset CRC engine before the configuration
*
- * By default, only CRC0 is configured, and the entire frame is used to
- * calculate the CRC.
+ * By default, the entire frame is used to calculate the CRC.
*
* Return: %false if the stream is not found or CRC capture is not supported;
* %true if the stream has been configured.
*/
bool dc_stream_configure_crc(struct dc *dc, struct dc_stream_state *stream,
- struct crc_params *crc_window, bool enable, bool continuous)
+ struct crc_params *crc_window, bool enable, bool continuous,
+ uint8_t idx, bool reset)
{
struct pipe_ctx *pipe;
struct crc_params param;
@@ -677,6 +741,9 @@ bool dc_stream_configure_crc(struct dc *dc, struct dc_stream_state *stream,
param.continuous_mode = continuous;
param.enable = enable;
+ param.crc_eng_inst = idx;
+ param.reset = reset;
+
tg = pipe->stream_res.tg;
/* Only call if supported */
@@ -691,6 +758,7 @@ bool dc_stream_configure_crc(struct dc *dc, struct dc_stream_state *stream,
*
* @dc: DC object.
* @stream: The DC stream state of the stream to get CRCs from.
+ * @idx: index of crc engine to get CRC from
* @r_cr: CRC value for the red component.
* @g_y: CRC value for the green component.
* @b_cb: CRC value for the blue component.
@@ -700,7 +768,7 @@ bool dc_stream_configure_crc(struct dc *dc, struct dc_stream_state *stream,
* Return:
* %false if stream is not found, or if CRCs are not enabled.
*/
-bool dc_stream_get_crc(struct dc *dc, struct dc_stream_state *stream,
+bool dc_stream_get_crc(struct dc *dc, struct dc_stream_state *stream, uint8_t idx,
uint32_t *r_cr, uint32_t *g_y, uint32_t *b_cb)
{
int i;
@@ -721,7 +789,7 @@ bool dc_stream_get_crc(struct dc *dc, struct dc_stream_state *stream,
tg = pipe->stream_res.tg;
if (tg->funcs->get_crc)
- return tg->funcs->get_crc(tg, r_cr, g_y, b_cb);
+ return tg->funcs->get_crc(tg, idx, r_cr, g_y, b_cb);
DC_LOG_WARNING("CRC capture not supported.");
return false;
}
@@ -1173,6 +1241,8 @@ static void dc_update_visual_confirm_color(struct dc *dc, struct dc_state *conte
get_mclk_switch_visual_confirm_color(pipe_ctx, &(pipe_ctx->visual_confirm_color));
else if (dc->debug.visual_confirm == VISUAL_CONFIRM_FAMS2)
get_fams2_visual_confirm_color(dc, context, pipe_ctx, &(pipe_ctx->visual_confirm_color));
+ else if (dc->debug.visual_confirm == VISUAL_CONFIRM_VABC)
+ get_vabc_visual_confirm_color(pipe_ctx, &(pipe_ctx->visual_confirm_color));
}
}
}
@@ -2153,6 +2223,11 @@ enum dc_status dc_commit_streams(struct dc *dc, struct dc_commit_streams_params
struct dc_stream_state *stream = params->streams[i];
struct dc_stream_status *status = dc_stream_get_status(stream);
+ /* revalidate streams */
+ res = dc_validate_stream(dc, stream);
+ if (res != DC_OK)
+ return res;
+
dc_stream_log(dc, stream);
set[i].stream = stream;
@@ -2487,7 +2562,7 @@ static enum surface_update_type get_plane_info_update_type(const struct dc *dc,
if (memcmp(&u->plane_info->tiling_info, &u->surface->tiling_info,
- sizeof(union dc_tiling_info)) != 0) {
+ sizeof(struct dc_tiling_info)) != 0) {
update_flags->bits.swizzle_change = 1;
elevate_update_type(&update_type, UPDATE_TYPE_MED);
@@ -2982,6 +3057,10 @@ static void copy_surface_update_to_plane(
if (srf_update->cursor_csc_color_matrix)
surface->cursor_csc_color_matrix =
*srf_update->cursor_csc_color_matrix;
+
+ if (srf_update->bias_and_scale.bias_and_scale_valid)
+ surface->bias_and_scale =
+ srf_update->bias_and_scale;
}
static void copy_stream_update_to_stream(struct dc *dc,
@@ -4510,7 +4589,7 @@ static bool commit_minimal_transition_based_on_current_context(struct dc *dc,
struct pipe_split_policy_backup policy;
struct dc_state *intermediate_context;
struct dc_state *old_current_state = dc->current_state;
- struct dc_surface_update srf_updates[MAX_SURFACE_NUM] = {0};
+ struct dc_surface_update srf_updates[MAX_SURFACES] = {0};
int surface_count;
/*
@@ -5307,11 +5386,9 @@ void dc_set_power_state(struct dc *dc, enum dc_acpi_cm_power_state power_state)
dc->vm_pa_config.valid) {
dc->hwss.init_sys_ctx(dc->hwseq, dc, &dc->vm_pa_config);
}
-
break;
default:
ASSERT(dc->current_state->stream_count == 0);
-
dc_dmub_srv_notify_fw_dc_power_state(dc->ctx->dmub_srv, power_state);
dc_state_destruct(dc->current_state);
@@ -5435,6 +5512,11 @@ bool dc_set_ips_disable(struct dc *dc, unsigned int disable_ips)
void dc_allow_idle_optimizations_internal(struct dc *dc, bool allow, char const *caller_name)
{
+ int idle_fclk_khz = 0, idle_dramclk_khz = 0, i = 0;
+ enum mall_stream_type subvp_pipe_type[MAX_PIPES] = {0};
+ struct pipe_ctx *pipe = NULL;
+ struct dc_state *context = dc->current_state;
+
if (dc->debug.disable_idle_power_optimizations) {
DC_LOG_DEBUG("%s: disabled\n", __func__);
return;
@@ -5459,6 +5541,23 @@ void dc_allow_idle_optimizations_internal(struct dc *dc, bool allow, char const
dc->idle_optimizations_allowed = allow;
DC_LOG_DEBUG("%s: %s\n", __func__, allow ? "enabled" : "disabled");
}
+
+ // log idle clocks and sub vp pipe types at idle optimization time
+ if (dc->clk_mgr != NULL && dc->clk_mgr->funcs->get_hard_min_fclk)
+ idle_fclk_khz = dc->clk_mgr->funcs->get_hard_min_fclk(dc->clk_mgr);
+
+ if (dc->clk_mgr != NULL && dc->clk_mgr->funcs->get_hard_min_memclk)
+ idle_dramclk_khz = dc->clk_mgr->funcs->get_hard_min_memclk(dc->clk_mgr);
+
+ for (i = 0; i < dc->res_pool->pipe_count; i++) {
+ pipe = &context->res_ctx.pipe_ctx[i];
+ subvp_pipe_type[i] = dc_state_get_pipe_subvp_type(context, pipe);
+ }
+
+ DC_LOG_DC("%s: allow_idle=%d\n HardMinUClk_Khz=%d HardMinDramclk_Khz=%d\n Pipe_0=%d Pipe_1=%d Pipe_2=%d Pipe_3=%d Pipe_4=%d Pipe_5=%d (caller=%s)\n",
+ __func__, allow, idle_fclk_khz, idle_dramclk_khz, subvp_pipe_type[0], subvp_pipe_type[1], subvp_pipe_type[2],
+ subvp_pipe_type[3], subvp_pipe_type[4], subvp_pipe_type[5], caller_name);
+
}
void dc_exit_ips_for_hw_access_internal(struct dc *dc, const char *caller_name)
@@ -6056,7 +6155,7 @@ void dc_query_current_properties(struct dc *dc, struct dc_current_properties *pr
bool subvp_sw_cursor_req = false;
for (i = 0; i < dc->current_state->stream_count; i++) {
- if (check_subvp_sw_cursor_fallback_req(dc, dc->current_state->streams[i])) {
+ if (check_subvp_sw_cursor_fallback_req(dc, dc->current_state->streams[i]) && !dc->current_state->streams[i]->hw_cursor_req) {
subvp_sw_cursor_req = true;
break;
}
@@ -6109,3 +6208,21 @@ struct dc_power_profile dc_get_power_profile_for_dc_state(const struct dc_state
profile.power_level = dc->res_pool->funcs->get_power_profile(context);
return profile;
}
+
+/*
+ **********************************************************************************
+ * dc_get_det_buffer_size_from_state() - extracts detile buffer size from dc state
+ *
+ * Called when DM wants to log detile buffer size from dc_state
+ *
+ **********************************************************************************
+ */
+unsigned int dc_get_det_buffer_size_from_state(const struct dc_state *context)
+{
+ struct dc *dc = context->clk_mgr->ctx->dc;
+
+ if (dc->res_pool->funcs->get_det_buffer_size)
+ return dc->res_pool->funcs->get_det_buffer_size(context);
+ else
+ return 0;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/core/dc_hw_sequencer.c
index 252af83e34a5..6eb9bae3af91 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_hw_sequencer.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_hw_sequencer.c
@@ -425,6 +425,44 @@ void get_hdr_visual_confirm_color(
}
}
+/* Visual Confirm color definition for VABC */
+void get_vabc_visual_confirm_color(
+ struct pipe_ctx *pipe_ctx,
+ struct tg_color *color)
+{
+ uint32_t color_value = MAX_TG_COLOR_VALUE;
+ struct dc_link *edp_link = NULL;
+
+ if (pipe_ctx && pipe_ctx->stream && pipe_ctx->stream->link) {
+ if (pipe_ctx->stream->link->connector_signal == SIGNAL_TYPE_EDP)
+ edp_link = pipe_ctx->stream->link;
+ }
+
+ if (edp_link) {
+ switch (edp_link->backlight_control_type) {
+ case BACKLIGHT_CONTROL_PWM:
+ color->color_r_cr = color_value;
+ color->color_g_y = 0;
+ color->color_b_cb = 0;
+ break;
+ case BACKLIGHT_CONTROL_AMD_AUX:
+ color->color_r_cr = 0;
+ color->color_g_y = color_value;
+ color->color_b_cb = 0;
+ break;
+ case BACKLIGHT_CONTROL_VESA_AUX:
+ color->color_r_cr = 0;
+ color->color_g_y = 0;
+ color->color_b_cb = color_value;
+ break;
+ }
+ } else {
+ color->color_r_cr = 0;
+ color->color_g_y = 0;
+ color->color_b_cb = 0;
+ }
+}
+
void get_subvp_visual_confirm_color(
struct pipe_ctx *pipe_ctx,
struct tg_color *color)
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link_exports.c b/drivers/gpu/drm/amd/display/dc/core/dc_link_exports.c
index 457d60eeb486..c1b79b379447 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_link_exports.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_link_exports.c
@@ -125,6 +125,14 @@ uint32_t dc_link_bandwidth_kbps(
return link->dc->link_srv->dp_link_bandwidth_kbps(link, link_settings);
}
+uint32_t dc_link_required_hblank_size_bytes(
+ const struct dc_link *link,
+ struct dp_audio_bandwidth_params *audio_params)
+{
+ return link->dc->link_srv->dp_required_hblank_size_bytes(link,
+ audio_params);
+}
+
void dc_get_cur_link_res_map(const struct dc *dc, uint32_t *map)
{
dc->link_srv->get_cur_res_map(dc, map);
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_resource.c b/drivers/gpu/drm/amd/display/dc/core/dc_resource.c
index 619fad17de55..520a34a42827 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_resource.c
@@ -2094,7 +2094,8 @@ int resource_get_odm_slice_dst_width(struct pipe_ctx *otg_master,
count = resource_get_odm_slice_count(otg_master);
h_active = timing->h_addressable +
timing->h_border_left +
- timing->h_border_right;
+ timing->h_border_right +
+ otg_master->hblank_borrow;
width = h_active / count;
if (otg_master->stream_res.tg)
@@ -4027,6 +4028,41 @@ fail:
}
/**
+ * decide_hblank_borrow - Decides the horizontal blanking borrow value for a given pipe context.
+ * @pipe_ctx: Pointer to the pipe context structure.
+ *
+ * This function calculates the horizontal blanking borrow value for a given pipe context based on the
+ * display stream compression (DSC) configuration. If the horizontal active pixels (hactive) are less
+ * than the total width of the DSC slices, it sets the hblank_borrow value to the difference. If the
+ * total horizontal timing minus the hblank_borrow value is less than 32, it resets the hblank_borrow
+ * value to 0.
+ */
+static void decide_hblank_borrow(struct pipe_ctx *pipe_ctx)
+{
+ uint32_t hactive;
+ uint32_t ceil_slice_width;
+ struct dc_stream_state *stream = NULL;
+
+ if (!pipe_ctx)
+ return;
+
+ stream = pipe_ctx->stream;
+
+ if (stream->timing.flags.DSC) {
+ hactive = stream->timing.h_addressable + stream->timing.h_border_left + stream->timing.h_border_right;
+
+ /* Assume if determined slices does not divide Hactive evenly, Hborrow is needed for padding*/
+ if (hactive % stream->timing.dsc_cfg.num_slices_h != 0) {
+ ceil_slice_width = (hactive / stream->timing.dsc_cfg.num_slices_h) + 1;
+ pipe_ctx->hblank_borrow = ceil_slice_width * stream->timing.dsc_cfg.num_slices_h - hactive;
+
+ if (stream->timing.h_total - hactive - pipe_ctx->hblank_borrow < 32)
+ pipe_ctx->hblank_borrow = 0;
+ }
+ }
+}
+
+/**
* dc_validate_global_state() - Determine if hardware can support a given state
*
* @dc: dc struct for this driver
@@ -4064,6 +4100,10 @@ enum dc_status dc_validate_global_state(
if (pipe_ctx->stream != stream)
continue;
+ /* Decide whether hblank borrow is needed and save it in pipe_ctx */
+ if (dc->debug.enable_hblank_borrow)
+ decide_hblank_borrow(pipe_ctx);
+
if (dc->res_pool->funcs->patch_unknown_plane_state &&
pipe_ctx->plane_state &&
pipe_ctx->plane_state->tiling_info.gfx9.swizzle == DC_SW_UNKNOWN) {
@@ -4438,7 +4478,7 @@ static void set_hfvs_info_packet(
static void adaptive_sync_override_dp_info_packets_sdp_line_num(
const struct dc_crtc_timing *timing,
struct enc_sdp_line_num *sdp_line_num,
- struct _vcs_dpi_display_pipe_dest_params_st *pipe_dlg_param)
+ unsigned int vstartup_start)
{
uint32_t asic_blank_start = 0;
uint32_t asic_blank_end = 0;
@@ -4453,8 +4493,8 @@ static void adaptive_sync_override_dp_info_packets_sdp_line_num(
asic_blank_end = (asic_blank_start - tg->v_border_bottom -
tg->v_addressable - tg->v_border_top);
- if (pipe_dlg_param->vstartup_start > asic_blank_end) {
- v_update = (tg->v_total - (pipe_dlg_param->vstartup_start - asic_blank_end));
+ if (vstartup_start > asic_blank_end) {
+ v_update = (tg->v_total - (vstartup_start - asic_blank_end));
sdp_line_num->adaptive_sync_line_num_valid = true;
sdp_line_num->adaptive_sync_line_num = (tg->v_total - v_update - 1);
} else {
@@ -4467,7 +4507,7 @@ static void set_adaptive_sync_info_packet(
struct dc_info_packet *info_packet,
const struct dc_stream_state *stream,
struct encoder_info_frame *info_frame,
- struct _vcs_dpi_display_pipe_dest_params_st *pipe_dlg_param)
+ unsigned int vstartup_start)
{
if (!stream->adaptive_sync_infopacket.valid)
return;
@@ -4475,7 +4515,7 @@ static void set_adaptive_sync_info_packet(
adaptive_sync_override_dp_info_packets_sdp_line_num(
&stream->timing,
&info_frame->sdp_line_num,
- pipe_dlg_param);
+ vstartup_start);
*info_packet = stream->adaptive_sync_infopacket;
}
@@ -4508,6 +4548,7 @@ void resource_build_info_frame(struct pipe_ctx *pipe_ctx)
{
enum signal_type signal = SIGNAL_TYPE_NONE;
struct encoder_info_frame *info = &pipe_ctx->stream_res.encoder_info_frame;
+ unsigned int vstartup_start = 0;
/* default all packets to invalid */
info->avi.valid = false;
@@ -4521,6 +4562,9 @@ void resource_build_info_frame(struct pipe_ctx *pipe_ctx)
info->adaptive_sync.valid = false;
signal = pipe_ctx->stream->signal;
+ if (pipe_ctx->stream->ctx->dc->res_pool->funcs->get_vstartup_for_pipe)
+ vstartup_start = pipe_ctx->stream->ctx->dc->res_pool->funcs->get_vstartup_for_pipe(pipe_ctx);
+
/* HDMi and DP have different info packets*/
if (dc_is_hdmi_signal(signal)) {
set_avi_info_frame(&info->avi, pipe_ctx);
@@ -4542,7 +4586,7 @@ void resource_build_info_frame(struct pipe_ctx *pipe_ctx)
set_adaptive_sync_info_packet(&info->adaptive_sync,
pipe_ctx->stream,
info,
- &pipe_ctx->pipe_dlg_param);
+ vstartup_start);
}
patch_gamut_packet_checksum(&info->gamut);
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_state.c b/drivers/gpu/drm/amd/display/dc/core/dc_state.c
index e006f816ff2f..1b2cce127981 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_state.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_state.c
@@ -483,9 +483,9 @@ bool dc_state_add_plane(
if (stream_status == NULL) {
dm_error("Existing stream not found; failed to attach surface!\n");
goto out;
- } else if (stream_status->plane_count == MAX_SURFACE_NUM) {
+ } else if (stream_status->plane_count == MAX_SURFACES) {
dm_error("Surface: can not attach plane_state %p! Maximum is: %d\n",
- plane_state, MAX_SURFACE_NUM);
+ plane_state, MAX_SURFACES);
goto out;
} else if (!otg_master_pipe) {
goto out;
@@ -600,7 +600,7 @@ bool dc_state_rem_all_planes_for_stream(
{
int i, old_plane_count;
struct dc_stream_status *stream_status = NULL;
- struct dc_plane_state *del_planes[MAX_SURFACE_NUM] = { 0 };
+ struct dc_plane_state *del_planes[MAX_SURFACES] = { 0 };
for (i = 0; i < state->stream_count; i++)
if (state->streams[i] == stream) {
@@ -875,7 +875,7 @@ bool dc_state_rem_all_phantom_planes_for_stream(
{
int i, old_plane_count;
struct dc_stream_status *stream_status = NULL;
- struct dc_plane_state *del_planes[MAX_SURFACE_NUM] = { 0 };
+ struct dc_plane_state *del_planes[MAX_SURFACES] = { 0 };
for (i = 0; i < state->stream_count; i++)
if (state->streams[i] == phantom_stream) {
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_stream.c b/drivers/gpu/drm/amd/display/dc/core/dc_stream.c
index 55dc482d9b36..e8134c47fe0d 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_stream.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_stream.c
@@ -37,6 +37,8 @@
#define DC_LOGGER dc->ctx->logger
#ifndef MIN
#define MIN(X, Y) ((X) < (Y) ? (X) : (Y))
+#endif
+#ifndef MAX
#define MAX(x, y) ((x > y) ? x : y)
#endif
@@ -605,17 +607,6 @@ bool dc_stream_remove_writeback(struct dc *dc,
return true;
}
-bool dc_stream_warmup_writeback(struct dc *dc,
- int num_dwb,
- struct dc_writeback_info *wb_info)
-{
- dc_exit_ips_for_hw_access(dc);
-
- if (dc->hwss.mmhubbub_warmup)
- return dc->hwss.mmhubbub_warmup(dc, num_dwb, wb_info);
- else
- return false;
-}
uint32_t dc_stream_get_vblank_counter(const struct dc_stream_state *stream)
{
uint8_t i;
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_surface.c b/drivers/gpu/drm/amd/display/dc/core/dc_surface.c
index ccbb15f1638c..f3471d45b312 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_surface.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_surface.c
@@ -83,13 +83,6 @@ uint8_t dc_plane_get_pipe_mask(struct dc_state *dc_state, const struct dc_plane
/*******************************************************************************
* Public functions
******************************************************************************/
-void enable_surface_flip_reporting(struct dc_plane_state *plane_state,
- uint32_t controller_id)
-{
- plane_state->irq_source = controller_id + DC_IRQ_SOURCE_PFLIP1 - 1;
- /*register_flip_interrupt(surface);*/
-}
-
struct dc_plane_state *dc_create_plane_state(const struct dc *dc)
{
struct dc_plane_state *plane_state = kvzalloc(sizeof(*plane_state),
@@ -277,4 +270,50 @@ void dc_3dlut_func_retain(struct dc_3dlut *lut)
kref_get(&lut->refcount);
}
+void dc_plane_force_update_for_panic(struct dc_plane_state *plane_state,
+ bool clear_tiling)
+{
+ struct dc *dc;
+ int i;
+
+ if (!plane_state)
+ return;
+
+ dc = plane_state->ctx->dc;
+ if (!dc || !dc->current_state)
+ return;
+
+ for (i = 0; i < dc->res_pool->pipe_count; i++) {
+ struct pipe_ctx *pipe_ctx = &dc->current_state->res_ctx.pipe_ctx[i];
+
+ if (!pipe_ctx)
+ continue;
+
+ if (dc->ctx->dce_version >= DCE_VERSION_MAX) {
+ struct hubp *hubp = pipe_ctx->plane_res.hubp;
+ if (!hubp)
+ continue;
+ /* if framebuffer is tiled, disable tiling */
+ if (clear_tiling && hubp->funcs->hubp_clear_tiling)
+ hubp->funcs->hubp_clear_tiling(hubp);
+
+ /* force page flip to see the new content of the framebuffer */
+ hubp->funcs->hubp_program_surface_flip_and_addr(hubp,
+ &plane_state->address,
+ true);
+ } else {
+ struct mem_input *mi = pipe_ctx->plane_res.mi;
+ if (!mi)
+ continue;
+ /* if framebuffer is tiled, disable tiling */
+ if (clear_tiling && mi->funcs->mem_input_clear_tiling)
+ mi->funcs->mem_input_clear_tiling(mi);
+
+ /* force page flip to see the new content of the framebuffer */
+ mi->funcs->mem_input_program_surface_flip_and_addr(mi,
+ &plane_state->address,
+ true);
+ }
+ }
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dc.h b/drivers/gpu/drm/amd/display/dc/dc.h
index 104051935884..053481ab69ef 100644
--- a/drivers/gpu/drm/amd/display/dc/dc.h
+++ b/drivers/gpu/drm/amd/display/dc/dc.h
@@ -55,9 +55,9 @@ struct aux_payload;
struct set_config_cmd_payload;
struct dmub_notification;
-#define DC_VER "3.2.310"
+#define DC_VER "3.2.316"
-#define MAX_SURFACES 3
+#define MAX_SURFACES 4
#define MAX_PLANES 6
#define MAX_STREAMS 6
#define MIN_VIEWPORT_SIZE 12
@@ -290,6 +290,7 @@ struct dc_caps {
uint16_t subvp_vertical_int_margin_us;
bool seamless_odm;
uint32_t max_v_total;
+ bool vtotal_limited_by_fp2;
uint32_t max_disp_clock_khz_at_vmin;
uint8_t subvp_drr_vblank_start_margin_us;
bool cursor_not_scaled;
@@ -462,6 +463,7 @@ struct dc_config {
bool enable_auto_dpm_test_logs;
unsigned int disable_ips;
unsigned int disable_ips_in_vpb;
+ bool disable_ips_in_dpms_off;
bool usb4_bw_alloc_support;
bool allow_0_dtb_clk;
bool use_assr_psp_message;
@@ -470,6 +472,7 @@ struct dc_config {
bool disable_hbr_audio_dp2;
bool consolidated_dpia_dp_lt;
bool set_pipe_unlock_order;
+ bool enable_dpia_pre_training;
};
enum visual_confirm {
@@ -486,6 +489,7 @@ enum visual_confirm {
VISUAL_CONFIRM_MCLK_SWITCH = 16,
VISUAL_CONFIRM_FAMS2 = 19,
VISUAL_CONFIRM_HW_CURSOR = 20,
+ VISUAL_CONFIRM_VABC = 21,
};
enum dc_psr_power_opts {
@@ -627,6 +631,8 @@ struct dc_clocks {
int bw_dispclk_khz;
int idle_dramclk_khz;
int idle_fclk_khz;
+ int subvp_prefetch_dramclk_khz;
+ int subvp_prefetch_fclk_khz;
};
struct dc_bw_validation_profile {
@@ -771,7 +777,8 @@ union dpia_debug_options {
uint32_t enable_force_tbt3_work_around:1; /* bit 4 */
uint32_t disable_usb4_pm_support:1; /* bit 5 */
uint32_t enable_consolidated_dpia_dp_lt:1; /* bit 6 */
- uint32_t reserved:25;
+ uint32_t enable_dpia_pre_training:1; /* bit 7 */
+ uint32_t reserved:24;
} bits;
uint32_t raw;
};
@@ -1054,8 +1061,8 @@ struct dc_debug_options {
bool dml21_force_pstate_method;
uint32_t dml21_force_pstate_method_values[MAX_PIPES];
uint32_t dml21_disable_pstate_method_mask;
+ union fw_assisted_mclk_switch_version fams_version;
union dmub_fams2_global_feature_config fams2_config;
- bool enable_legacy_clock_update;
unsigned int force_cositing;
unsigned int disable_spl;
unsigned int force_easf;
@@ -1068,6 +1075,8 @@ struct dc_debug_options {
unsigned int scale_to_sharpness_policy;
bool skip_full_updated_if_possible;
unsigned int enable_oled_edp_power_up_opt;
+ bool enable_hblank_borrow;
+ bool force_subvp_df_throttle;
};
@@ -1298,7 +1307,7 @@ struct dc_plane_state {
struct rect clip_rect;
struct plane_size plane_size;
- union dc_tiling_info tiling_info;
+ struct dc_tiling_info tiling_info;
struct dc_plane_dcc_param dcc;
@@ -1369,7 +1378,7 @@ struct dc_plane_state {
struct dc_plane_info {
struct plane_size plane_size;
- union dc_tiling_info tiling_info;
+ struct dc_tiling_info tiling_info;
struct dc_plane_dcc_param dcc;
enum surface_pixel_format format;
enum dc_rotation_angle rotation;
@@ -1396,7 +1405,7 @@ struct dc_scratch_space {
* store current value in plane states so we can still recover
* a valid current state during dc update.
*/
- struct dc_plane_state plane_states[MAX_SURFACE_NUM];
+ struct dc_plane_state plane_states[MAX_SURFACES];
struct dc_stream_state stream_state;
};
@@ -1524,6 +1533,7 @@ struct dc_surface_update {
const struct dc_cm2_parameters *cm2_params;
const struct dc_csc_transform *cursor_csc_color_matrix;
unsigned int sdr_white_level_nits;
+ struct dc_bias_and_scale bias_and_scale;
};
/*
@@ -2017,6 +2027,24 @@ uint32_t dc_link_bandwidth_kbps(
const struct dc_link *link,
const struct dc_link_settings *link_setting);
+struct dp_audio_bandwidth_params {
+ const struct dc_crtc_timing *crtc_timing;
+ enum dp_link_encoding link_encoding;
+ uint32_t channel_count;
+ uint32_t sample_rate_hz;
+};
+
+/* The function calculates the minimum size of hblank (in bytes) needed to
+ * support the specified channel count and sample rate combination, given the
+ * link encoding and timing to be used. This calculation is not supported
+ * for 8b/10b SST.
+ *
+ * return - min hblank size in bytes, 0 if 8b/10b SST.
+ */
+uint32_t dc_link_required_hblank_size_bytes(
+ const struct dc_link *link,
+ struct dp_audio_bandwidth_params *audio_params);
+
/* The function takes a snapshot of current link resource allocation state
* @dc: pointer to dc of the dm calling this
* @map: a dc link resource snapshot defined internally to dc.
@@ -2376,6 +2404,13 @@ struct dc_sink_dsc_caps {
struct dsc_dec_dpcd_caps dsc_dec_caps;
};
+struct dc_sink_hblank_expansion_caps {
+ // 'true' if these are virtual DPCD's HBlank expansion caps (immediately upstream of sink in MST topology),
+ // 'false' if they are sink's HBlank expansion caps
+ bool is_virtual_dpcd_hblank_expansion;
+ struct hblank_expansion_dpcd_caps dpcd_caps;
+};
+
struct dc_sink_fec_caps {
bool is_rx_fec_supported;
bool is_topology_fec_supported;
@@ -2402,6 +2437,7 @@ struct dc_sink {
struct scdc_caps scdc_caps;
struct dc_sink_dsc_caps dsc_caps;
struct dc_sink_fec_caps fec_caps;
+ struct dc_sink_hblank_expansion_caps hblank_expansion_caps;
bool is_vsc_sdp_colorimetry_supported;
@@ -2550,6 +2586,8 @@ struct dc_power_profile {
struct dc_power_profile dc_get_power_profile_for_dc_state(const struct dc_state *context);
+unsigned int dc_get_det_buffer_size_from_state(const struct dc_state *context);
+
/* DSC Interfaces */
#include "dc_dsc.h"
diff --git a/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c b/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c
index f90fc154549a..44ff9abe2880 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c
+++ b/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c
@@ -1245,7 +1245,7 @@ static int count_active_streams(const struct dc *dc)
for (i = 0; i < dc->current_state->stream_count; ++i) {
struct dc_stream_state *stream = dc->current_state->streams[i];
- if (stream && !stream->dpms_off)
+ if (stream && (!stream->dpms_off || dc->config.disable_ips_in_dpms_off))
count += 1;
}
@@ -1694,10 +1694,10 @@ void dc_dmub_srv_fams2_update_config(struct dc *dc,
{
uint8_t num_cmds = 1;
uint32_t i;
- union dmub_rb_cmd cmd[MAX_STREAMS + 1];
+ union dmub_rb_cmd cmd[2 * MAX_STREAMS + 1];
struct dmub_rb_cmd_fams2 *global_cmd = &cmd[0].fams2_config;
- memset(cmd, 0, sizeof(union dmub_rb_cmd) * (MAX_STREAMS + 1));
+ memset(cmd, 0, sizeof(union dmub_rb_cmd) * (2 * MAX_STREAMS + 1));
/* fill in generic command header */
global_cmd->header.type = DMUB_CMD__FW_ASSISTED_MCLK_SWITCH;
global_cmd->header.sub_type = DMUB_CMD__FAMS2_CONFIG;
@@ -1714,17 +1714,26 @@ void dc_dmub_srv_fams2_update_config(struct dc *dc,
/* construct per-stream configs */
for (i = 0; i < context->bw_ctx.bw.dcn.fams2_global_config.num_streams; i++) {
- struct dmub_rb_cmd_fams2 *stream_cmd = &cmd[i+1].fams2_config;
+ struct dmub_rb_cmd_fams2 *stream_base_cmd = &cmd[i+1].fams2_config;
+ struct dmub_rb_cmd_fams2 *stream_sub_state_cmd = &cmd[i+1+context->bw_ctx.bw.dcn.fams2_global_config.num_streams].fams2_config;
/* configure command header */
- stream_cmd->header.type = DMUB_CMD__FW_ASSISTED_MCLK_SWITCH;
- stream_cmd->header.sub_type = DMUB_CMD__FAMS2_CONFIG;
- stream_cmd->header.payload_bytes = sizeof(struct dmub_rb_cmd_fams2) - sizeof(struct dmub_cmd_header);
- stream_cmd->header.multi_cmd_pending = 1;
- /* copy stream static state */
- memcpy(&stream_cmd->config.stream,
- &context->bw_ctx.bw.dcn.fams2_stream_params[i],
- sizeof(struct dmub_fams2_stream_static_state));
+ stream_base_cmd->header.type = DMUB_CMD__FW_ASSISTED_MCLK_SWITCH;
+ stream_base_cmd->header.sub_type = DMUB_CMD__FAMS2_CONFIG;
+ stream_base_cmd->header.payload_bytes = sizeof(struct dmub_rb_cmd_fams2) - sizeof(struct dmub_cmd_header);
+ stream_base_cmd->header.multi_cmd_pending = 1;
+ stream_sub_state_cmd->header.type = DMUB_CMD__FW_ASSISTED_MCLK_SWITCH;
+ stream_sub_state_cmd->header.sub_type = DMUB_CMD__FAMS2_CONFIG;
+ stream_sub_state_cmd->header.payload_bytes = sizeof(struct dmub_rb_cmd_fams2) - sizeof(struct dmub_cmd_header);
+ stream_sub_state_cmd->header.multi_cmd_pending = 1;
+ /* copy stream static base state */
+ memcpy(&stream_base_cmd->config,
+ &context->bw_ctx.bw.dcn.fams2_stream_base_params[i],
+ sizeof(union dmub_cmd_fams2_config));
+ /* copy stream static sub state */
+ memcpy(&stream_sub_state_cmd->config,
+ &context->bw_ctx.bw.dcn.fams2_stream_sub_params[i],
+ sizeof(union dmub_cmd_fams2_config));
}
}
@@ -1735,8 +1744,8 @@ void dc_dmub_srv_fams2_update_config(struct dc *dc,
if (enable && context->bw_ctx.bw.dcn.fams2_global_config.features.bits.enable) {
/* set multi pending for global, and unset for last stream cmd */
global_cmd->header.multi_cmd_pending = 1;
- cmd[context->bw_ctx.bw.dcn.fams2_global_config.num_streams].fams2_config.header.multi_cmd_pending = 0;
- num_cmds += context->bw_ctx.bw.dcn.fams2_global_config.num_streams;
+ cmd[2 * context->bw_ctx.bw.dcn.fams2_global_config.num_streams].fams2_config.header.multi_cmd_pending = 0;
+ num_cmds += 2 * context->bw_ctx.bw.dcn.fams2_global_config.num_streams;
}
dm_execute_dmub_cmd_list(dc->ctx, num_cmds, cmd, DM_DMUB_WAIT_TYPE_WAIT);
diff --git a/drivers/gpu/drm/amd/display/dc/dc_dp_types.h b/drivers/gpu/drm/amd/display/dc/dc_dp_types.h
index 8dd6eb044829..94ce8fe74481 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_dp_types.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_dp_types.h
@@ -969,6 +969,21 @@ union dp_sink_video_fallback_formats {
uint8_t raw;
};
+union dp_receive_port0_cap {
+ struct {
+ uint8_t RESERVED :1;
+ uint8_t LOCAL_EDID_PRESENT :1;
+ uint8_t ASSOCIATED_TO_PRECEDING_PORT:1;
+ uint8_t HBLANK_EXPANSION_CAPABLE :1;
+ uint8_t BUFFER_SIZE_UNIT :1;
+ uint8_t BUFFER_SIZE_PER_PORT :1;
+ uint8_t HBLANK_REDUCTION_CAPABLE :1;
+ uint8_t RESERVED2:1;
+ uint8_t BUFFER_SIZE:8;
+ } bits;
+ uint8_t raw[2];
+};
+
union dpcd_max_uncompressed_pixel_rate_cap {
struct {
uint16_t max_uncompressed_pixel_rate_cap :15;
@@ -1193,6 +1208,7 @@ struct dpcd_caps {
struct replay_info pr_info;
uint16_t edp_oled_emission_rate;
+ union dp_receive_port0_cap receive_port0_cap;
};
union dpcd_sink_ext_caps {
diff --git a/drivers/gpu/drm/amd/display/dc/dc_dsc.h b/drivers/gpu/drm/amd/display/dc/dc_dsc.h
index 9014c2409817..9d18f1c08079 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_dsc.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_dsc.h
@@ -94,6 +94,11 @@ uint32_t dc_dsc_stream_bandwidth_overhead_in_kbps(
const int num_slices_h,
const bool is_dp);
+void dc_dsc_dump_decoder_caps(const struct display_stream_compressor *dsc,
+ const struct dsc_dec_dpcd_caps *dsc_sink_caps);
+void dc_dsc_dump_encoder_caps(const struct display_stream_compressor *dsc,
+ const struct dc_crtc_timing *timing);
+
/* TODO - Hardware/specs limitation should be owned by dc dsc and returned to DM,
* and DM can choose to OVERRIDE the limitation on CASE BY CASE basis.
* Hardware/specs limitation should not be writable by DM.
diff --git a/drivers/gpu/drm/amd/display/dc/dc_hw_types.h b/drivers/gpu/drm/amd/display/dc/dc_hw_types.h
index c10567ec1c81..5ac55601a6da 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_hw_types.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_hw_types.h
@@ -341,89 +341,101 @@ enum swizzle_mode_addr3_values {
DC_ADDR3_SW_UNKNOWN = DC_ADDR3_SW_MAX
};
-union dc_tiling_info {
-
- struct {
- /* Specifies the number of memory banks for tiling
- * purposes.
- * Only applies to 2D and 3D tiling modes.
- * POSSIBLE VALUES: 2,4,8,16
- */
- unsigned int num_banks;
- /* Specifies the number of tiles in the x direction
- * to be incorporated into the same bank.
- * Only applies to 2D and 3D tiling modes.
- * POSSIBLE VALUES: 1,2,4,8
- */
- unsigned int bank_width;
- unsigned int bank_width_c;
- /* Specifies the number of tiles in the y direction to
- * be incorporated into the same bank.
- * Only applies to 2D and 3D tiling modes.
- * POSSIBLE VALUES: 1,2,4,8
- */
- unsigned int bank_height;
- unsigned int bank_height_c;
- /* Specifies the macro tile aspect ratio. Only applies
- * to 2D and 3D tiling modes.
- */
- unsigned int tile_aspect;
- unsigned int tile_aspect_c;
- /* Specifies the number of bytes that will be stored
- * contiguously for each tile.
- * If the tile data requires more storage than this
- * amount, it is split into multiple slices.
- * This field must not be larger than
- * GB_ADDR_CONFIG.DRAM_ROW_SIZE.
- * Only applies to 2D and 3D tiling modes.
- * For color render targets, TILE_SPLIT >= 256B.
- */
- enum tile_split_values tile_split;
- enum tile_split_values tile_split_c;
- /* Specifies the addressing within a tile.
- * 0x0 - DISPLAY_MICRO_TILING
- * 0x1 - THIN_MICRO_TILING
- * 0x2 - DEPTH_MICRO_TILING
- * 0x3 - ROTATED_MICRO_TILING
- */
- enum tile_mode_values tile_mode;
- enum tile_mode_values tile_mode_c;
- /* Specifies the number of pipes and how they are
- * interleaved in the surface.
- * Refer to memory addressing document for complete
- * details and constraints.
- */
- unsigned int pipe_config;
- /* Specifies the tiling mode of the surface.
- * THIN tiles use an 8x8x1 tile size.
- * THICK tiles use an 8x8x4 tile size.
- * 2D tiling modes rotate banks for successive Z slices
- * 3D tiling modes rotate pipes and banks for Z slices
- * Refer to memory addressing document for complete
- * details and constraints.
- */
- enum array_mode_values array_mode;
- } gfx8;
+enum dc_gfxversion {
+ DcGfxVersion7 = 0,
+ DcGfxVersion8,
+ DcGfxVersion9,
+ DcGfxVersion10,
+ DcGfxVersion11,
+ DcGfxAddr3,
+ DcGfxVersionUnknown
+};
+
+ struct dc_tiling_info {
+ unsigned int gfxversion; // Specifies which part of the union to use. Must use DalGfxVersion enum
+ union {
+ struct {
+ /* Specifies the number of memory banks for tiling
+ * purposes.
+ * Only applies to 2D and 3D tiling modes.
+ * POSSIBLE VALUES: 2,4,8,16
+ */
+ unsigned int num_banks;
+ /* Specifies the number of tiles in the x direction
+ * to be incorporated into the same bank.
+ * Only applies to 2D and 3D tiling modes.
+ * POSSIBLE VALUES: 1,2,4,8
+ */
+ unsigned int bank_width;
+ unsigned int bank_width_c;
+ /* Specifies the number of tiles in the y direction to
+ * be incorporated into the same bank.
+ * Only applies to 2D and 3D tiling modes.
+ * POSSIBLE VALUES: 1,2,4,8
+ */
+ unsigned int bank_height;
+ unsigned int bank_height_c;
+ /* Specifies the macro tile aspect ratio. Only applies
+ * to 2D and 3D tiling modes.
+ */
+ unsigned int tile_aspect;
+ unsigned int tile_aspect_c;
+ /* Specifies the number of bytes that will be stored
+ * contiguously for each tile.
+ * If the tile data requires more storage than this
+ * amount, it is split into multiple slices.
+ * This field must not be larger than
+ * GB_ADDR_CONFIG.DRAM_ROW_SIZE.
+ * Only applies to 2D and 3D tiling modes.
+ * For color render targets, TILE_SPLIT >= 256B.
+ */
+ enum tile_split_values tile_split;
+ enum tile_split_values tile_split_c;
+ /* Specifies the addressing within a tile.
+ * 0x0 - DISPLAY_MICRO_TILING
+ * 0x1 - THIN_MICRO_TILING
+ * 0x2 - DEPTH_MICRO_TILING
+ * 0x3 - ROTATED_MICRO_TILING
+ */
+ enum tile_mode_values tile_mode;
+ enum tile_mode_values tile_mode_c;
+ /* Specifies the number of pipes and how they are
+ * interleaved in the surface.
+ * Refer to memory addressing document for complete
+ * details and constraints.
+ */
+ unsigned int pipe_config;
+ /* Specifies the tiling mode of the surface.
+ * THIN tiles use an 8x8x1 tile size.
+ * THICK tiles use an 8x8x4 tile size.
+ * 2D tiling modes rotate banks for successive Z slices
+ * 3D tiling modes rotate pipes and banks for Z slices
+ * Refer to memory addressing document for complete
+ * details and constraints.
+ */
+ enum array_mode_values array_mode;
+ } gfx8;
- struct {
- enum swizzle_mode_values swizzle;
- unsigned int num_pipes;
- unsigned int max_compressed_frags;
- unsigned int pipe_interleave;
-
- unsigned int num_banks;
- unsigned int num_shader_engines;
- unsigned int num_rb_per_se;
- bool shaderEnable;
-
- bool meta_linear;
- bool rb_aligned;
- bool pipe_aligned;
- unsigned int num_pkrs;
- } gfx9;/*gfx9, gfx10 and above*/
- struct {
- enum swizzle_mode_addr3_values swizzle;
- } gfx_addr3;/*gfx with addr3 and above*/
+ struct {
+ enum swizzle_mode_values swizzle;
+ unsigned int num_pipes;
+ unsigned int max_compressed_frags;
+ unsigned int pipe_interleave;
+
+ unsigned int num_banks;
+ unsigned int num_shader_engines;
+ unsigned int num_rb_per_se;
+ bool shaderEnable;
+
+ bool meta_linear;
+ bool rb_aligned;
+ bool pipe_aligned;
+ unsigned int num_pkrs;
+ } gfx9;/*gfx9, gfx10 and above*/
+ struct {
+ enum swizzle_mode_addr3_values swizzle;
+ } gfx_addr3;/*gfx with addr3 and above*/
+ };
};
/* Rotation angle */
@@ -975,6 +987,9 @@ struct dc_crtc_timing {
struct dc_crtc_timing_flags flags;
uint32_t dsc_fixed_bits_per_pixel_x16; /* DSC target bitrate in 1/16 of bpp (e.g. 128 -> 8bpp) */
struct dc_dsc_config dsc_cfg;
+
+ /* The number of pixels that HBlank has been expanded by from the original EDID timing. */
+ uint32_t expanded_hblank;
};
enum trigger_delay {
diff --git a/drivers/gpu/drm/amd/display/dc/dc_plane.h b/drivers/gpu/drm/amd/display/dc/dc_plane.h
index bd37ec82b42d..fabcefeda288 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_plane.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_plane.h
@@ -34,4 +34,7 @@ const struct dc_plane_status *dc_plane_get_status(
void dc_plane_state_retain(struct dc_plane_state *plane_state);
void dc_plane_state_release(struct dc_plane_state *plane_state);
+void dc_plane_force_update_for_panic(struct dc_plane_state *plane_state,
+ bool clear_tiling);
+
#endif /* _DC_PLANE_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/dc_spl_translate.c b/drivers/gpu/drm/amd/display/dc/dc_spl_translate.c
index c8d8e335fa37..3518eb1b8cd1 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_spl_translate.c
+++ b/drivers/gpu/drm/amd/display/dc/dc_spl_translate.c
@@ -64,6 +64,13 @@ static void populate_inits_from_splinits(struct scl_inits *inits,
inits->h_c = dc_fixpt_from_int_dy(spl_inits->h_filter_init_int_c, spl_inits->h_filter_init_frac_c >> 5, 0, 19);
inits->v_c = dc_fixpt_from_int_dy(spl_inits->v_filter_init_int_c, spl_inits->v_filter_init_frac_c >> 5, 0, 19);
}
+static void populate_splformat_from_format(enum spl_pixel_format *spl_pixel_format, const enum pixel_format pixel_format)
+{
+ if (pixel_format < PIXEL_FORMAT_INVALID)
+ *spl_pixel_format = (enum spl_pixel_format)pixel_format;
+ else
+ *spl_pixel_format = SPL_PIXEL_FORMAT_INVALID;
+}
/// @brief Translate SPL input parameters from pipe context
/// @param pipe_ctx
/// @param spl_in
@@ -89,7 +96,7 @@ void translate_SPL_in_params_from_pipe_ctx(struct pipe_ctx *pipe_ctx, struct spl
spl_in->callbacks = dcn2_spl_callbacks;
}
// Make format field from spl_in point to plane_res scl_data format
- spl_in->basic_in.format = (enum spl_pixel_format)pipe_ctx->plane_res.scl_data.format;
+ populate_splformat_from_format(&spl_in->basic_in.format, pipe_ctx->plane_res.scl_data.format);
// Make view_format from basic_out point to view_format from stream
spl_in->basic_out.view_format = (enum spl_view_3d)stream->view_format;
// Populate spl input basic input clip rect from plane state clip rect
@@ -108,19 +115,21 @@ void translate_SPL_in_params_from_pipe_ctx(struct pipe_ctx *pipe_ctx, struct spl
spl_in->basic_in.horizontal_mirror = plane_state->horizontal_mirror;
// Calculate horizontal splits and split index
- spl_in->basic_in.mpc_combine_h = resource_get_mpc_slice_count(pipe_ctx);
+ spl_in->basic_in.num_h_slices_recout_width_align.use_recout_width_aligned = false;
+ spl_in->basic_in.num_h_slices_recout_width_align.num_slices_recout_width.mpc_num_h_slices =
+ resource_get_mpc_slice_count(pipe_ctx);
if (stream->view_format == VIEW_3D_FORMAT_SIDE_BY_SIDE)
- spl_in->basic_in.mpc_combine_v = 0;
+ spl_in->basic_in.mpc_h_slice_index = 0;
else
- spl_in->basic_in.mpc_combine_v = resource_get_mpc_slice_index(pipe_ctx);
+ spl_in->basic_in.mpc_h_slice_index = resource_get_mpc_slice_index(pipe_ctx);
populate_splrect_from_rect(&spl_in->basic_out.odm_slice_rect, &odm_slice_src);
spl_in->basic_out.odm_combine_factor = 0;
spl_in->odm_slice_index = resource_get_odm_slice_index(pipe_ctx);
// Make spl input basic out info output_size width point to stream h active
spl_in->basic_out.output_size.width =
- stream->timing.h_addressable + stream->timing.h_border_left + stream->timing.h_border_right;
+ stream->timing.h_addressable + stream->timing.h_border_left + stream->timing.h_border_right + pipe_ctx->hblank_borrow;
// Make spl input basic out info output_size height point to v active
spl_in->basic_out.output_size.height =
stream->timing.v_addressable + stream->timing.v_border_bottom + stream->timing.v_border_top;
diff --git a/drivers/gpu/drm/amd/display/dc/dc_stream.h b/drivers/gpu/drm/amd/display/dc/dc_stream.h
index 413970588a26..3e303c7808fb 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_stream.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_stream.h
@@ -56,7 +56,7 @@ struct dc_stream_status {
int plane_count;
int audio_inst;
struct timing_sync_info timing_sync_info;
- struct dc_plane_state *plane_states[MAX_SURFACE_NUM];
+ struct dc_plane_state *plane_states[MAX_SURFACES];
bool is_abm_supported;
struct mall_stream_config mall_stream_config;
bool fpo_in_use;
@@ -447,10 +447,6 @@ enum dc_status dc_stream_add_dsc_to_resource(struct dc *dc,
struct dc_state *state,
struct dc_stream_state *stream);
-bool dc_stream_warmup_writeback(struct dc *dc,
- int num_dwb,
- struct dc_writeback_info *wb_info);
-
bool dc_stream_dmdata_status_done(struct dc *dc, struct dc_stream_state *stream);
bool dc_stream_set_dynamic_metadata(struct dc *dc,
@@ -541,17 +537,26 @@ bool dc_stream_get_crtc_position(struct dc *dc,
#if defined(CONFIG_DRM_AMD_SECURE_DISPLAY)
bool dc_stream_forward_crc_window(struct dc_stream_state *stream,
struct rect *rect,
+ uint8_t phy_id,
bool is_stop);
+
+bool dc_stream_forward_multiple_crc_window(struct dc_stream_state *stream,
+ struct crc_window *window,
+ uint8_t phy_id,
+ bool stop);
#endif
bool dc_stream_configure_crc(struct dc *dc,
struct dc_stream_state *stream,
struct crc_params *crc_window,
bool enable,
- bool continuous);
+ bool continuous,
+ uint8_t idx,
+ bool reset);
bool dc_stream_get_crc(struct dc *dc,
struct dc_stream_state *stream,
+ uint8_t idx,
uint32_t *r_cr,
uint32_t *g_y,
uint32_t *b_cb);
diff --git a/drivers/gpu/drm/amd/display/dc/dc_types.h b/drivers/gpu/drm/amd/display/dc/dc_types.h
index edf4df1d03b5..0c2aa91f0a11 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_types.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_types.h
@@ -76,7 +76,6 @@ struct dc_perf_trace {
unsigned long last_entry_write;
};
-#define MAX_SURFACE_NUM 6
#define NUM_PIXEL_FORMATS 10
enum tiling_mode {
@@ -875,6 +874,14 @@ struct dsc_dec_dpcd_caps {
bool is_dp; /* Decoded format */
};
+struct hblank_expansion_dpcd_caps {
+ bool expansion_supported;
+ bool reduction_supported;
+ bool buffer_unit_bytes; /* True: buffer size in bytes. False: buffer size in pixels*/
+ bool buffer_per_port; /* True: buffer size per port. False: buffer size per lane*/
+ uint32_t buffer_size; /* Add 1 to value and multiply by 32 */
+};
+
struct dc_golden_table {
uint16_t dc_golden_table_ver;
uint32_t aux_dphy_rx_control0_val;
@@ -932,10 +939,17 @@ enum backlight_control_type {
};
#if defined(CONFIG_DRM_AMD_SECURE_DISPLAY)
+#define MAX_CRC_WINDOW_NUM 2
+
struct otg_phy_mux {
uint8_t phy_output_num;
uint8_t otg_output_num;
};
+
+struct crc_window {
+ struct rect rect;
+ bool enable;
+};
#endif
enum dc_detect_reason {
@@ -1052,10 +1066,13 @@ enum replay_FW_Message_type {
union replay_error_status {
struct {
- unsigned char STATE_TRANSITION_ERROR :1;
- unsigned char LINK_CRC_ERROR :1;
- unsigned char DESYNC_ERROR :1;
- unsigned char RESERVED :5;
+ unsigned int STATE_TRANSITION_ERROR :1;
+ unsigned int LINK_CRC_ERROR :1;
+ unsigned int DESYNC_ERROR :1;
+ unsigned int RESERVED_3 :1;
+ unsigned int LOW_RR_INCORRECT_VTOTAL :1;
+ unsigned int NO_DOUBLED_RR :1;
+ unsigned int RESERVED_6_7 :2;
} bits;
unsigned char raw;
};
@@ -1102,6 +1119,8 @@ struct replay_config {
union replay_error_status replay_error_status;
/* Replay Low Hz enable Options */
union replay_low_refresh_rate_enable_options low_rr_enable_options;
+ /* Replay coasting vtotal is within low refresh rate range. */
+ bool low_rr_activated;
};
/* Replay feature flags*/
@@ -1126,10 +1145,12 @@ struct replay_settings {
uint32_t defer_update_coasting_vtotal_table[PR_COASTING_TYPE_NUM];
/* Maximum link off frame count */
uint32_t link_off_frame_count;
- /* Replay pseudo vtotal for abm + ips on full screen video which can improve ips residency */
- uint16_t abm_with_ips_on_full_screen_video_pseudo_vtotal;
+ /* Replay pseudo vtotal for low refresh rate*/
+ uint16_t low_rr_full_screen_video_pseudo_vtotal;
/* Replay last pseudo vtotal set to DMUB */
uint16_t last_pseudo_vtotal;
+ /* Replay desync error */
+ uint32_t replay_desync_error_fail_count;
};
/* To split out "global" and "per-panel" config settings.
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.c b/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.c
index b700608e4240..077337698e0a 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.c
@@ -1105,6 +1105,9 @@ static bool dcn401_program_pix_clk(
&dto_params);
} else {
+ if (pll_settings->actual_pix_clk_100hz > 6000000UL)
+ return false;
+
/* disables DP DTO when provided with TMDS signal type */
clock_source->ctx->dc->res_pool->dccg->funcs->set_dp_dto(
clock_source->ctx->dc->res_pool->dccg,
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_mem_input.c b/drivers/gpu/drm/amd/display/dc/dce/dce_mem_input.c
index f5e1d9caee4c..1c2009e38aa1 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_mem_input.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_mem_input.c
@@ -98,7 +98,7 @@ static enum mi_bits_per_pixel get_mi_bpp(
}
static enum mi_tiling_format get_mi_tiling(
- union dc_tiling_info *tiling_info)
+ struct dc_tiling_info *tiling_info)
{
switch (tiling_info->gfx8.array_mode) {
case DC_ARRAY_1D_TILED_THIN1:
@@ -133,7 +133,7 @@ static bool is_vert_scan(enum dc_rotation_angle rotation)
static void dce_mi_program_pte_vm(
struct mem_input *mi,
enum surface_pixel_format format,
- union dc_tiling_info *tiling_info,
+ struct dc_tiling_info *tiling_info,
enum dc_rotation_angle rotation)
{
struct dce_mem_input *dce_mi = TO_DCE_MEM_INPUT(mi);
@@ -430,7 +430,7 @@ static void dce120_mi_program_display_marks(struct mem_input *mi,
}
static void program_tiling(
- struct dce_mem_input *dce_mi, const union dc_tiling_info *info)
+ struct dce_mem_input *dce_mi, const struct dc_tiling_info *info)
{
if (dce_mi->masks->GRPH_SW_MODE) { /* GFX9 */
REG_UPDATE_6(GRPH_CONTROL,
@@ -481,7 +481,6 @@ static void program_tiling(
}
}
-
static void program_size_and_rotation(
struct dce_mem_input *dce_mi,
enum dc_rotation_angle rotation,
@@ -627,10 +626,31 @@ static void program_grph_pixel_format(
GRPH_PRESCALE_B_SIGN, sign);
}
+static void dce_mi_clear_tiling(
+ struct mem_input *mi)
+{
+ struct dce_mem_input *dce_mi = TO_DCE_MEM_INPUT(mi);
+
+ if (dce_mi->masks->GRPH_SW_MODE) { /* GFX9 */
+ REG_UPDATE(GRPH_CONTROL,
+ GRPH_SW_MODE, DC_SW_LINEAR);
+ }
+
+ if (dce_mi->masks->GRPH_MICRO_TILE_MODE) { /* GFX8 */
+ REG_UPDATE(GRPH_CONTROL,
+ GRPH_ARRAY_MODE, DC_SW_LINEAR);
+ }
+
+ if (dce_mi->masks->GRPH_ARRAY_MODE) { /* GFX6 but reuses gfx8 struct */
+ REG_UPDATE(GRPH_CONTROL,
+ GRPH_ARRAY_MODE, DC_SW_LINEAR);
+ }
+}
+
static void dce_mi_program_surface_config(
struct mem_input *mi,
enum surface_pixel_format format,
- union dc_tiling_info *tiling_info,
+ struct dc_tiling_info *tiling_info,
struct plane_size *plane_size,
enum dc_rotation_angle rotation,
struct dc_plane_dcc_param *dcc,
@@ -650,7 +670,7 @@ static void dce_mi_program_surface_config(
static void dce60_mi_program_surface_config(
struct mem_input *mi,
enum surface_pixel_format format,
- union dc_tiling_info *tiling_info,
+ struct dc_tiling_info *tiling_info,
struct plane_size *plane_size,
enum dc_rotation_angle rotation, /* not used in DCE6 */
struct dc_plane_dcc_param *dcc,
@@ -884,7 +904,8 @@ static const struct mem_input_funcs dce_mi_funcs = {
.mem_input_program_pte_vm = dce_mi_program_pte_vm,
.mem_input_program_surface_config =
dce_mi_program_surface_config,
- .mem_input_is_flip_pending = dce_mi_is_flip_pending
+ .mem_input_is_flip_pending = dce_mi_is_flip_pending,
+ .mem_input_clear_tiling = dce_mi_clear_tiling,
};
#if defined(CONFIG_DRM_AMD_DC_SI)
@@ -897,7 +918,8 @@ static const struct mem_input_funcs dce60_mi_funcs = {
.mem_input_program_pte_vm = dce_mi_program_pte_vm,
.mem_input_program_surface_config =
dce60_mi_program_surface_config,
- .mem_input_is_flip_pending = dce_mi_is_flip_pending
+ .mem_input_is_flip_pending = dce_mi_is_flip_pending,
+ .mem_input_clear_tiling = dce_mi_clear_tiling,
};
#endif
@@ -910,7 +932,8 @@ static const struct mem_input_funcs dce112_mi_funcs = {
.mem_input_program_pte_vm = dce_mi_program_pte_vm,
.mem_input_program_surface_config =
dce_mi_program_surface_config,
- .mem_input_is_flip_pending = dce_mi_is_flip_pending
+ .mem_input_is_flip_pending = dce_mi_is_flip_pending,
+ .mem_input_clear_tiling = dce_mi_clear_tiling,
};
static const struct mem_input_funcs dce120_mi_funcs = {
@@ -922,7 +945,8 @@ static const struct mem_input_funcs dce120_mi_funcs = {
.mem_input_program_pte_vm = dce_mi_program_pte_vm,
.mem_input_program_surface_config =
dce_mi_program_surface_config,
- .mem_input_is_flip_pending = dce_mi_is_flip_pending
+ .mem_input_is_flip_pending = dce_mi_is_flip_pending,
+ .mem_input_clear_tiling = dce_mi_clear_tiling,
};
void dce_mem_input_construct(
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dmub_hw_lock_mgr.c b/drivers/gpu/drm/amd/display/dc/dce/dmub_hw_lock_mgr.c
index bf636b28e3e1..5bb8b78bf250 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dmub_hw_lock_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dmub_hw_lock_mgr.c
@@ -63,7 +63,8 @@ void dmub_hw_lock_mgr_inbox0_cmd(struct dc_dmub_srv *dmub_srv,
bool should_use_dmub_lock(struct dc_link *link)
{
- if (link->psr_settings.psr_version == DC_PSR_VERSION_SU_1)
+ if (link->psr_settings.psr_version == DC_PSR_VERSION_SU_1 ||
+ link->psr_settings.psr_version == DC_PSR_VERSION_1)
return true;
if (link->replay_settings.replay_feature_enabled)
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dmub_psr.c b/drivers/gpu/drm/amd/display/dc/dce/dmub_psr.c
index cae18f8c1c9a..88c75c243bf8 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dmub_psr.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dmub_psr.c
@@ -390,8 +390,7 @@ static bool dmub_psr_copy_settings(struct dmub_psr *dmub,
!memcmp(link->dpcd_caps.sink_dev_id_str, DP_SINK_DEVICE_STR_ID_1,
sizeof(DP_SINK_DEVICE_STR_ID_1)))
link->psr_settings.force_ffu_mode = 1;
- else
- link->psr_settings.force_ffu_mode = 0;
+
copy_settings_data->force_ffu_mode = link->psr_settings.force_ffu_mode;
if (((link->dpcd_caps.fec_cap.bits.FEC_CAPABLE &&
diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_mem_input_v.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_mem_input_v.c
index 8a3fbf95c48f..2c43c2422638 100644
--- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_mem_input_v.c
+++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_mem_input_v.c
@@ -162,7 +162,7 @@ static void enable(struct dce_mem_input *mem_input110)
static void program_tiling(
struct dce_mem_input *mem_input110,
- const union dc_tiling_info *info,
+ const struct dc_tiling_info *info,
const enum surface_pixel_format pixel_format)
{
uint32_t value = 0;
@@ -523,7 +523,7 @@ static const unsigned int dvmm_Hw_Setting_Linear[4][9] = {
/* Helper to get table entry from surface info */
static const unsigned int *get_dvmm_hw_setting(
- union dc_tiling_info *tiling_info,
+ struct dc_tiling_info *tiling_info,
enum surface_pixel_format format,
bool chroma)
{
@@ -563,7 +563,7 @@ static const unsigned int *get_dvmm_hw_setting(
static void dce_mem_input_v_program_pte_vm(
struct mem_input *mem_input,
enum surface_pixel_format format,
- union dc_tiling_info *tiling_info,
+ struct dc_tiling_info *tiling_info,
enum dc_rotation_angle rotation)
{
struct dce_mem_input *mem_input110 = TO_DCE_MEM_INPUT(mem_input);
@@ -636,7 +636,7 @@ static void dce_mem_input_v_program_pte_vm(
static void dce_mem_input_v_program_surface_config(
struct mem_input *mem_input,
enum surface_pixel_format format,
- union dc_tiling_info *tiling_info,
+ struct dc_tiling_info *tiling_info,
struct plane_size *plane_size,
enum dc_rotation_angle rotation,
struct dc_plane_dcc_param *dcc,
diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator.c
index fa422a8cbced..61b0807693fb 100644
--- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator.c
+++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator.c
@@ -2127,70 +2127,131 @@ bool dce110_configure_crc(struct timing_generator *tg,
cntl_addr = CRTC_REG(mmCRTC_CRC_CNTL);
- /* First, disable CRC before we configure it. */
- dm_write_reg(tg->ctx, cntl_addr, 0);
+ if (!params->enable || params->reset)
+ /* First, disable CRC before we configure it. */
+ dm_write_reg(tg->ctx, cntl_addr, 0);
if (!params->enable)
return true;
/* Program frame boundaries */
- /* Window A x axis start and end. */
- value = 0;
- addr = CRTC_REG(mmCRTC_CRC0_WINDOWA_X_CONTROL);
- set_reg_field_value(value, params->windowa_x_start,
- CRTC_CRC0_WINDOWA_X_CONTROL,
- CRTC_CRC0_WINDOWA_X_START);
- set_reg_field_value(value, params->windowa_x_end,
- CRTC_CRC0_WINDOWA_X_CONTROL,
- CRTC_CRC0_WINDOWA_X_END);
- dm_write_reg(tg->ctx, addr, value);
-
- /* Window A y axis start and end. */
- value = 0;
- addr = CRTC_REG(mmCRTC_CRC0_WINDOWA_Y_CONTROL);
- set_reg_field_value(value, params->windowa_y_start,
- CRTC_CRC0_WINDOWA_Y_CONTROL,
- CRTC_CRC0_WINDOWA_Y_START);
- set_reg_field_value(value, params->windowa_y_end,
- CRTC_CRC0_WINDOWA_Y_CONTROL,
- CRTC_CRC0_WINDOWA_Y_END);
- dm_write_reg(tg->ctx, addr, value);
-
- /* Window B x axis start and end. */
- value = 0;
- addr = CRTC_REG(mmCRTC_CRC0_WINDOWB_X_CONTROL);
- set_reg_field_value(value, params->windowb_x_start,
- CRTC_CRC0_WINDOWB_X_CONTROL,
- CRTC_CRC0_WINDOWB_X_START);
- set_reg_field_value(value, params->windowb_x_end,
- CRTC_CRC0_WINDOWB_X_CONTROL,
- CRTC_CRC0_WINDOWB_X_END);
- dm_write_reg(tg->ctx, addr, value);
-
- /* Window B y axis start and end. */
- value = 0;
- addr = CRTC_REG(mmCRTC_CRC0_WINDOWB_Y_CONTROL);
- set_reg_field_value(value, params->windowb_y_start,
- CRTC_CRC0_WINDOWB_Y_CONTROL,
- CRTC_CRC0_WINDOWB_Y_START);
- set_reg_field_value(value, params->windowb_y_end,
- CRTC_CRC0_WINDOWB_Y_CONTROL,
- CRTC_CRC0_WINDOWB_Y_END);
- dm_write_reg(tg->ctx, addr, value);
-
- /* Set crc mode and selection, and enable. Only using CRC0*/
- value = 0;
- set_reg_field_value(value, params->continuous_mode ? 1 : 0,
- CRTC_CRC_CNTL, CRTC_CRC_CONT_EN);
- set_reg_field_value(value, params->selection,
- CRTC_CRC_CNTL, CRTC_CRC0_SELECT);
- set_reg_field_value(value, 1, CRTC_CRC_CNTL, CRTC_CRC_EN);
- dm_write_reg(tg->ctx, cntl_addr, value);
+ switch (params->crc_eng_inst) {
+ case 0:
+ /* Window A x axis start and end. */
+ value = 0;
+ addr = CRTC_REG(mmCRTC_CRC0_WINDOWA_X_CONTROL);
+ set_reg_field_value(value, params->windowa_x_start,
+ CRTC_CRC0_WINDOWA_X_CONTROL,
+ CRTC_CRC0_WINDOWA_X_START);
+ set_reg_field_value(value, params->windowa_x_end,
+ CRTC_CRC0_WINDOWA_X_CONTROL,
+ CRTC_CRC0_WINDOWA_X_END);
+ dm_write_reg(tg->ctx, addr, value);
+
+ /* Window A y axis start and end. */
+ value = 0;
+ addr = CRTC_REG(mmCRTC_CRC0_WINDOWA_Y_CONTROL);
+ set_reg_field_value(value, params->windowa_y_start,
+ CRTC_CRC0_WINDOWA_Y_CONTROL,
+ CRTC_CRC0_WINDOWA_Y_START);
+ set_reg_field_value(value, params->windowa_y_end,
+ CRTC_CRC0_WINDOWA_Y_CONTROL,
+ CRTC_CRC0_WINDOWA_Y_END);
+ dm_write_reg(tg->ctx, addr, value);
+
+ /* Window B x axis start and end. */
+ value = 0;
+ addr = CRTC_REG(mmCRTC_CRC0_WINDOWB_X_CONTROL);
+ set_reg_field_value(value, params->windowb_x_start,
+ CRTC_CRC0_WINDOWB_X_CONTROL,
+ CRTC_CRC0_WINDOWB_X_START);
+ set_reg_field_value(value, params->windowb_x_end,
+ CRTC_CRC0_WINDOWB_X_CONTROL,
+ CRTC_CRC0_WINDOWB_X_END);
+ dm_write_reg(tg->ctx, addr, value);
+
+ /* Window B y axis start and end. */
+ value = 0;
+ addr = CRTC_REG(mmCRTC_CRC0_WINDOWB_Y_CONTROL);
+ set_reg_field_value(value, params->windowb_y_start,
+ CRTC_CRC0_WINDOWB_Y_CONTROL,
+ CRTC_CRC0_WINDOWB_Y_START);
+ set_reg_field_value(value, params->windowb_y_end,
+ CRTC_CRC0_WINDOWB_Y_CONTROL,
+ CRTC_CRC0_WINDOWB_Y_END);
+ dm_write_reg(tg->ctx, addr, value);
+
+ /* Set crc mode and selection, and enable.*/
+ value = 0;
+ set_reg_field_value(value, params->continuous_mode ? 1 : 0,
+ CRTC_CRC_CNTL, CRTC_CRC_CONT_EN);
+ set_reg_field_value(value, params->selection,
+ CRTC_CRC_CNTL, CRTC_CRC0_SELECT);
+ set_reg_field_value(value, 1, CRTC_CRC_CNTL, CRTC_CRC_EN);
+ dm_write_reg(tg->ctx, cntl_addr, value);
+ break;
+ case 1:
+ /* Window A x axis start and end. */
+ value = 0;
+ addr = CRTC_REG(mmCRTC_CRC1_WINDOWA_X_CONTROL);
+ set_reg_field_value(value, params->windowa_x_start,
+ CRTC_CRC1_WINDOWA_X_CONTROL,
+ CRTC_CRC1_WINDOWA_X_START);
+ set_reg_field_value(value, params->windowa_x_end,
+ CRTC_CRC1_WINDOWA_X_CONTROL,
+ CRTC_CRC1_WINDOWA_X_END);
+ dm_write_reg(tg->ctx, addr, value);
+
+ /* Window A y axis start and end. */
+ value = 0;
+ addr = CRTC_REG(mmCRTC_CRC1_WINDOWA_Y_CONTROL);
+ set_reg_field_value(value, params->windowa_y_start,
+ CRTC_CRC1_WINDOWA_Y_CONTROL,
+ CRTC_CRC1_WINDOWA_Y_START);
+ set_reg_field_value(value, params->windowa_y_end,
+ CRTC_CRC1_WINDOWA_Y_CONTROL,
+ CRTC_CRC1_WINDOWA_Y_END);
+ dm_write_reg(tg->ctx, addr, value);
+
+ /* Window B x axis start and end. */
+ value = 0;
+ addr = CRTC_REG(mmCRTC_CRC1_WINDOWB_X_CONTROL);
+ set_reg_field_value(value, params->windowb_x_start,
+ CRTC_CRC1_WINDOWB_X_CONTROL,
+ CRTC_CRC1_WINDOWB_X_START);
+ set_reg_field_value(value, params->windowb_x_end,
+ CRTC_CRC1_WINDOWB_X_CONTROL,
+ CRTC_CRC1_WINDOWB_X_END);
+ dm_write_reg(tg->ctx, addr, value);
+
+ /* Window B y axis start and end. */
+ value = 0;
+ addr = CRTC_REG(mmCRTC_CRC1_WINDOWB_Y_CONTROL);
+ set_reg_field_value(value, params->windowb_y_start,
+ CRTC_CRC1_WINDOWB_Y_CONTROL,
+ CRTC_CRC1_WINDOWB_Y_START);
+ set_reg_field_value(value, params->windowb_y_end,
+ CRTC_CRC1_WINDOWB_Y_CONTROL,
+ CRTC_CRC1_WINDOWB_Y_END);
+ dm_write_reg(tg->ctx, addr, value);
+
+ /* Set crc mode and selection, and enable.*/
+ value = 0;
+ set_reg_field_value(value, params->continuous_mode ? 1 : 0,
+ CRTC_CRC_CNTL, CRTC_CRC_CONT_EN);
+ set_reg_field_value(value, params->selection,
+ CRTC_CRC_CNTL, CRTC_CRC1_SELECT);
+ set_reg_field_value(value, 1, CRTC_CRC_CNTL, CRTC_CRC_EN);
+ dm_write_reg(tg->ctx, cntl_addr, value);
+ break;
+ default:
+ return false;
+ }
return true;
}
-bool dce110_get_crc(struct timing_generator *tg,
+bool dce110_get_crc(struct timing_generator *tg, uint8_t idx,
uint32_t *r_cr, uint32_t *g_y, uint32_t *b_cb)
{
uint32_t addr = 0;
@@ -2206,14 +2267,30 @@ bool dce110_get_crc(struct timing_generator *tg,
if (!field)
return false;
- addr = CRTC_REG(mmCRTC_CRC0_DATA_RG);
- value = dm_read_reg(tg->ctx, addr);
- *r_cr = get_reg_field_value(value, CRTC_CRC0_DATA_RG, CRC0_R_CR);
- *g_y = get_reg_field_value(value, CRTC_CRC0_DATA_RG, CRC0_G_Y);
+ switch (idx) {
+ case 0:
+ addr = CRTC_REG(mmCRTC_CRC0_DATA_RG);
+ value = dm_read_reg(tg->ctx, addr);
+ *r_cr = get_reg_field_value(value, CRTC_CRC0_DATA_RG, CRC0_R_CR);
+ *g_y = get_reg_field_value(value, CRTC_CRC0_DATA_RG, CRC0_G_Y);
- addr = CRTC_REG(mmCRTC_CRC0_DATA_B);
- value = dm_read_reg(tg->ctx, addr);
- *b_cb = get_reg_field_value(value, CRTC_CRC0_DATA_B, CRC0_B_CB);
+ addr = CRTC_REG(mmCRTC_CRC0_DATA_B);
+ value = dm_read_reg(tg->ctx, addr);
+ *b_cb = get_reg_field_value(value, CRTC_CRC0_DATA_B, CRC0_B_CB);
+ break;
+ case 1:
+ addr = CRTC_REG(mmCRTC_CRC1_DATA_RG);
+ value = dm_read_reg(tg->ctx, addr);
+ *r_cr = get_reg_field_value(value, CRTC_CRC1_DATA_RG, CRC1_R_CR);
+ *g_y = get_reg_field_value(value, CRTC_CRC1_DATA_RG, CRC1_G_Y);
+
+ addr = CRTC_REG(mmCRTC_CRC1_DATA_B);
+ value = dm_read_reg(tg->ctx, addr);
+ *b_cb = get_reg_field_value(value, CRTC_CRC1_DATA_B, CRC1_B_CB);
+ break;
+ default:
+ return false;
+ }
return true;
}
diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator.h b/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator.h
index ee4de740aceb..e4f5cad64f32 100644
--- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator.h
+++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator.h
@@ -286,7 +286,7 @@ bool dce110_arm_vert_intr(
bool dce110_configure_crc(struct timing_generator *tg,
const struct crc_params *params);
-bool dce110_get_crc(struct timing_generator *tg,
+bool dce110_get_crc(struct timing_generator *tg, uint8_t idx,
uint32_t *r_cr, uint32_t *g_y, uint32_t *b_cb);
bool dce110_is_two_pixels_per_container(const struct dc_crtc_timing *timing);
diff --git a/drivers/gpu/drm/amd/display/dc/dce120/dce120_timing_generator.c b/drivers/gpu/drm/amd/display/dc/dce120/dce120_timing_generator.c
index fcf59348eb62..31c4f44ceaac 100644
--- a/drivers/gpu/drm/amd/display/dc/dce120/dce120_timing_generator.c
+++ b/drivers/gpu/drm/amd/display/dc/dce120/dce120_timing_generator.c
@@ -1100,45 +1100,79 @@ static bool dce120_configure_crc(struct timing_generator *tg,
if (!dce120_is_tg_enabled(tg))
return false;
- /* First, disable CRC before we configure it. */
- dm_write_reg_soc15(tg->ctx, mmCRTC0_CRTC_CRC_CNTL,
- tg110->offsets.crtc, 0);
+ if (!params->enable || params->reset)
+ /* First, disable CRC before we configure it. */
+ dm_write_reg_soc15(tg->ctx, mmCRTC0_CRTC_CRC_CNTL,
+ tg110->offsets.crtc, 0);
if (!params->enable)
return true;
/* Program frame boundaries */
- /* Window A x axis start and end. */
- CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC0_WINDOWA_X_CONTROL,
- CRTC_CRC0_WINDOWA_X_START, params->windowa_x_start,
- CRTC_CRC0_WINDOWA_X_END, params->windowa_x_end);
-
- /* Window A y axis start and end. */
- CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC0_WINDOWA_Y_CONTROL,
- CRTC_CRC0_WINDOWA_Y_START, params->windowa_y_start,
- CRTC_CRC0_WINDOWA_Y_END, params->windowa_y_end);
-
- /* Window B x axis start and end. */
- CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC0_WINDOWB_X_CONTROL,
- CRTC_CRC0_WINDOWB_X_START, params->windowb_x_start,
- CRTC_CRC0_WINDOWB_X_END, params->windowb_x_end);
-
- /* Window B y axis start and end. */
- CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC0_WINDOWB_Y_CONTROL,
- CRTC_CRC0_WINDOWB_Y_START, params->windowb_y_start,
- CRTC_CRC0_WINDOWB_Y_END, params->windowb_y_end);
-
- /* Set crc mode and selection, and enable. Only using CRC0*/
- CRTC_REG_UPDATE_3(CRTC0_CRTC_CRC_CNTL,
- CRTC_CRC_EN, params->continuous_mode ? 1 : 0,
- CRTC_CRC0_SELECT, params->selection,
- CRTC_CRC_EN, 1);
+ switch (params->crc_eng_inst) {
+ case 0:
+ /* Window A x axis start and end. */
+ CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC0_WINDOWA_X_CONTROL,
+ CRTC_CRC0_WINDOWA_X_START, params->windowa_x_start,
+ CRTC_CRC0_WINDOWA_X_END, params->windowa_x_end);
+
+ /* Window A y axis start and end. */
+ CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC0_WINDOWA_Y_CONTROL,
+ CRTC_CRC0_WINDOWA_Y_START, params->windowa_y_start,
+ CRTC_CRC0_WINDOWA_Y_END, params->windowa_y_end);
+
+ /* Window B x axis start and end. */
+ CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC0_WINDOWB_X_CONTROL,
+ CRTC_CRC0_WINDOWB_X_START, params->windowb_x_start,
+ CRTC_CRC0_WINDOWB_X_END, params->windowb_x_end);
+
+ /* Window B y axis start and end. */
+ CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC0_WINDOWB_Y_CONTROL,
+ CRTC_CRC0_WINDOWB_Y_START, params->windowb_y_start,
+ CRTC_CRC0_WINDOWB_Y_END, params->windowb_y_end);
+
+ /* Set crc mode and selection, and enable.*/
+ CRTC_REG_UPDATE_3(CRTC0_CRTC_CRC_CNTL,
+ CRTC_CRC_CONT_EN, params->continuous_mode ? 1 : 0,
+ CRTC_CRC0_SELECT, params->selection,
+ CRTC_CRC_EN, 1);
+ break;
+ case 1:
+ /* Window A x axis start and end. */
+ CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC1_WINDOWA_X_CONTROL,
+ CRTC_CRC1_WINDOWA_X_START, params->windowa_x_start,
+ CRTC_CRC1_WINDOWA_X_END, params->windowa_x_end);
+
+ /* Window A y axis start and end. */
+ CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC1_WINDOWA_Y_CONTROL,
+ CRTC_CRC1_WINDOWA_Y_START, params->windowa_y_start,
+ CRTC_CRC1_WINDOWA_Y_END, params->windowa_y_end);
+
+ /* Window B x axis start and end. */
+ CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC1_WINDOWB_X_CONTROL,
+ CRTC_CRC1_WINDOWB_X_START, params->windowb_x_start,
+ CRTC_CRC1_WINDOWB_X_END, params->windowb_x_end);
+
+ /* Window B y axis start and end. */
+ CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC1_WINDOWB_Y_CONTROL,
+ CRTC_CRC1_WINDOWB_Y_START, params->windowb_y_start,
+ CRTC_CRC1_WINDOWB_Y_END, params->windowb_y_end);
+
+ /* Set crc mode and selection, and enable */
+ CRTC_REG_UPDATE_3(CRTC0_CRTC_CRC_CNTL,
+ CRTC_CRC_CONT_EN, params->continuous_mode ? 1 : 0,
+ CRTC_CRC1_SELECT, params->selection,
+ CRTC_CRC_EN, 1);
+ break;
+ default:
+ return false;
+ }
return true;
}
-static bool dce120_get_crc(struct timing_generator *tg, uint32_t *r_cr,
- uint32_t *g_y, uint32_t *b_cb)
+static bool dce120_get_crc(struct timing_generator *tg, uint8_t idx,
+ uint32_t *r_cr, uint32_t *g_y, uint32_t *b_cb)
{
struct dce110_timing_generator *tg110 = DCE110TG_FROM_TG(tg);
uint32_t value, field;
@@ -1151,14 +1185,30 @@ static bool dce120_get_crc(struct timing_generator *tg, uint32_t *r_cr,
if (!field)
return false;
- value = dm_read_reg_soc15(tg->ctx, mmCRTC0_CRTC_CRC0_DATA_RG,
- tg110->offsets.crtc);
- *r_cr = get_reg_field_value(value, CRTC0_CRTC_CRC0_DATA_RG, CRC0_R_CR);
- *g_y = get_reg_field_value(value, CRTC0_CRTC_CRC0_DATA_RG, CRC0_G_Y);
+ switch (idx) {
+ case 0:
+ value = dm_read_reg_soc15(tg->ctx, mmCRTC0_CRTC_CRC0_DATA_RG,
+ tg110->offsets.crtc);
+ *r_cr = get_reg_field_value(value, CRTC0_CRTC_CRC0_DATA_RG, CRC0_R_CR);
+ *g_y = get_reg_field_value(value, CRTC0_CRTC_CRC0_DATA_RG, CRC0_G_Y);
- value = dm_read_reg_soc15(tg->ctx, mmCRTC0_CRTC_CRC0_DATA_B,
- tg110->offsets.crtc);
- *b_cb = get_reg_field_value(value, CRTC0_CRTC_CRC0_DATA_B, CRC0_B_CB);
+ value = dm_read_reg_soc15(tg->ctx, mmCRTC0_CRTC_CRC0_DATA_B,
+ tg110->offsets.crtc);
+ *b_cb = get_reg_field_value(value, CRTC0_CRTC_CRC0_DATA_B, CRC0_B_CB);
+ break;
+ case 1:
+ value = dm_read_reg_soc15(tg->ctx, mmCRTC0_CRTC_CRC1_DATA_RG,
+ tg110->offsets.crtc);
+ *r_cr = get_reg_field_value(value, CRTC0_CRTC_CRC1_DATA_RG, CRC1_R_CR);
+ *g_y = get_reg_field_value(value, CRTC0_CRTC_CRC1_DATA_RG, CRC1_G_Y);
+
+ value = dm_read_reg_soc15(tg->ctx, mmCRTC0_CRTC_CRC1_DATA_B,
+ tg110->offsets.crtc);
+ *b_cb = get_reg_field_value(value, CRTC0_CRTC_CRC1_DATA_B, CRC1_B_CB);
+ break;
+ default:
+ return false;
+ }
return true;
}
diff --git a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_panel_cntl.c b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_panel_cntl.c
index 573898984726..f9961a6446f3 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_panel_cntl.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn31/dcn31_panel_cntl.c
@@ -168,31 +168,33 @@ void dcn31_panel_cntl_construct(
struct dcn31_panel_cntl *dcn31_panel_cntl,
const struct panel_cntl_init_data *init_data)
{
- uint8_t pwrseq_inst = 0xF;
dcn31_panel_cntl->base.funcs = &dcn31_link_panel_cntl_funcs;
dcn31_panel_cntl->base.ctx = init_data->ctx;
dcn31_panel_cntl->base.inst = init_data->inst;
- switch (init_data->eng_id) {
- case ENGINE_ID_DIGA:
- pwrseq_inst = 0;
- break;
- case ENGINE_ID_DIGB:
- pwrseq_inst = 1;
- break;
- default:
- DC_LOG_WARNING("Unsupported pwrseq engine id: %d!\n", init_data->eng_id);
- ASSERT(false);
- break;
- }
-
- if (dcn31_panel_cntl->base.ctx->dc->config.support_edp0_on_dp1)
+ if (dcn31_panel_cntl->base.ctx->dc->config.support_edp0_on_dp1) {
//If supported, power sequencer mapping shall follow the DIG instance
+ uint8_t pwrseq_inst = 0xF;
+
+ switch (init_data->eng_id) {
+ case ENGINE_ID_DIGA:
+ pwrseq_inst = 0;
+ break;
+ case ENGINE_ID_DIGB:
+ pwrseq_inst = 1;
+ break;
+ default:
+ DC_LOG_WARNING("Unsupported pwrseq engine id: %d!\n", init_data->eng_id);
+ ASSERT(false);
+ break;
+ }
+
dcn31_panel_cntl->base.pwrseq_inst = pwrseq_inst;
- else
+ } else {
/* If not supported, pwrseq will be assigned in order,
* so first pwrseq will be assigned to first panel instance (legacy behavior)
*/
dcn31_panel_cntl->base.pwrseq_inst = dcn31_panel_cntl->base.inst;
+ }
}
diff --git a/drivers/gpu/drm/amd/display/dc/dio/dcn31/dcn31_dio_link_encoder.c b/drivers/gpu/drm/amd/display/dc/dio/dcn31/dcn31_dio_link_encoder.c
index b2cea59ba5d4..9a92f73d5b7f 100644
--- a/drivers/gpu/drm/amd/display/dc/dio/dcn31/dcn31_dio_link_encoder.c
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn31/dcn31_dio_link_encoder.c
@@ -653,8 +653,9 @@ void dcn31_link_encoder_get_max_link_cap(struct link_encoder *enc, struct dc_lin
if (!query_dp_alt_from_dmub(enc, &cmd))
return;
- if (cmd.query_dp_alt.data.is_usb &&
- cmd.query_dp_alt.data.is_dp4 == 0)
+ if (cmd.query_dp_alt.data.is_dp_alt_disable == 0 &&
+ cmd.query_dp_alt.data.is_usb &&
+ cmd.query_dp_alt.data.is_dp4 == 0)
link_settings->lane_count = MIN(LANE_COUNT_TWO, link_settings->lane_count);
return;
diff --git a/drivers/gpu/drm/amd/display/dc/dio/dcn35/dcn35_dio_link_encoder.c b/drivers/gpu/drm/amd/display/dc/dio/dcn35/dcn35_dio_link_encoder.c
index d4a3e811aa39..ea0c9a9d0bd6 100644
--- a/drivers/gpu/drm/amd/display/dc/dio/dcn35/dcn35_dio_link_encoder.c
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn35/dcn35_dio_link_encoder.c
@@ -28,6 +28,7 @@
#include "link_encoder.h"
#include "dcn31/dcn31_dio_link_encoder.h"
#include "dcn35_dio_link_encoder.h"
+#include "dc_dmub_srv.h"
#define CTX \
enc10->base.ctx
#define DC_LOGGER \
@@ -159,6 +160,8 @@ static const struct link_encoder_funcs dcn35_link_enc_funcs = {
.is_in_alt_mode = dcn31_link_encoder_is_in_alt_mode,
.get_max_link_cap = dcn31_link_encoder_get_max_link_cap,
.set_dio_phy_mux = dcn31_link_encoder_set_dio_phy_mux,
+ .enable_dpia_output = dcn35_link_encoder_enable_dpia_output,
+ .disable_dpia_output = dcn35_link_encoder_disable_dpia_output,
};
void dcn35_link_encoder_construct(
@@ -265,3 +268,80 @@ void dcn35_link_encoder_construct(
enc10->base.features.flags.bits.HDMI_6GB_EN = 0;
}
+
+/* DPIA equivalent of link_transmitter_control. */
+static bool link_dpia_control(struct dc_context *dc_ctx,
+ struct dmub_cmd_dig_dpia_control_data *dpia_control)
+{
+ union dmub_rb_cmd cmd;
+
+ memset(&cmd, 0, sizeof(cmd));
+
+ cmd.dig1_dpia_control.header.type = DMUB_CMD__DPIA;
+ cmd.dig1_dpia_control.header.sub_type =
+ DMUB_CMD__DPIA_DIG1_DPIA_CONTROL;
+ cmd.dig1_dpia_control.header.payload_bytes =
+ sizeof(cmd.dig1_dpia_control) -
+ sizeof(cmd.dig1_dpia_control.header);
+
+ cmd.dig1_dpia_control.dpia_control = *dpia_control;
+
+ dc_wake_and_execute_dmub_cmd(dc_ctx, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
+
+ return true;
+}
+
+static void link_encoder_disable(struct dcn10_link_encoder *enc10)
+{
+ /* reset training complete */
+ REG_UPDATE(DP_LINK_CNTL, DP_LINK_TRAINING_COMPLETE, 0);
+}
+
+void dcn35_link_encoder_enable_dpia_output(
+ struct link_encoder *enc,
+ const struct dc_link_settings *link_settings,
+ uint8_t dpia_id,
+ uint8_t digmode,
+ uint8_t fec_rdy)
+{
+ struct dcn10_link_encoder *enc10 = TO_DCN10_LINK_ENC(enc);
+ struct dmub_cmd_dig_dpia_control_data dpia_control = { 0 };
+
+ enc1_configure_encoder(enc10, link_settings);
+
+ dpia_control.action = (uint8_t)TRANSMITTER_CONTROL_ENABLE;
+ dpia_control.enc_id = enc->preferred_engine;
+ dpia_control.mode_laneset.digmode = digmode;
+ dpia_control.lanenum = (uint8_t)link_settings->lane_count;
+ dpia_control.symclk_10khz = link_settings->link_rate *
+ LINK_RATE_REF_FREQ_IN_KHZ / 10;
+ /* DIG_BE_CNTL.DIG_HPD_SELECT set to 5 (hpdsel - 1) to indicate HPD pin unused by DPIA. */
+ dpia_control.hpdsel = 6;
+ dpia_control.dpia_id = dpia_id;
+ dpia_control.fec_rdy = fec_rdy;
+
+ DC_LOG_DEBUG("%s: DPIA(%d) - enc_id(%d)\n", __func__, dpia_control.dpia_id, dpia_control.enc_id);
+ link_dpia_control(enc->ctx, &dpia_control);
+}
+
+void dcn35_link_encoder_disable_dpia_output(
+ struct link_encoder *enc,
+ uint8_t dpia_id,
+ uint8_t digmode)
+{
+ struct dcn10_link_encoder *enc10 = TO_DCN10_LINK_ENC(enc);
+ struct dmub_cmd_dig_dpia_control_data dpia_control = { 0 };
+
+ if (enc->funcs->is_dig_enabled && !enc->funcs->is_dig_enabled(enc))
+ return;
+
+ dpia_control.action = (uint8_t)TRANSMITTER_CONTROL_DISABLE;
+ dpia_control.enc_id = enc->preferred_engine;
+ dpia_control.mode_laneset.digmode = digmode;
+ dpia_control.dpia_id = dpia_id;
+
+ DC_LOG_DEBUG("%s: DPIA(%d) - enc_id(%d)\n", __func__, dpia_control.dpia_id, dpia_control.enc_id);
+ link_dpia_control(enc->ctx, &dpia_control);
+
+ link_encoder_disable(enc10);
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dio/dcn35/dcn35_dio_link_encoder.h b/drivers/gpu/drm/amd/display/dc/dio/dcn35/dcn35_dio_link_encoder.h
index d546a3676304..f9d4221f4b43 100644
--- a/drivers/gpu/drm/amd/display/dc/dio/dcn35/dcn35_dio_link_encoder.h
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn35/dcn35_dio_link_encoder.h
@@ -144,4 +144,22 @@ bool dcn35_is_dig_enabled(struct link_encoder *enc);
enum signal_type dcn35_get_dig_mode(struct link_encoder *enc);
void dcn35_link_encoder_setup(struct link_encoder *enc, enum signal_type signal);
+/*
+ * Enable DP transmitter and its encoder for dpia port.
+ */
+void dcn35_link_encoder_enable_dpia_output(
+ struct link_encoder *enc,
+ const struct dc_link_settings *link_settings,
+ uint8_t dpia_id,
+ uint8_t digmode,
+ uint8_t fec_rdy);
+
+/*
+ * Disable transmitter and its encoder for dpia port.
+ */
+void dcn35_link_encoder_disable_dpia_output(
+ struct link_encoder *enc,
+ uint8_t dpia_id,
+ uint8_t digmode);
+
#endif /* __DC_LINK_ENCODER__DCN35_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dm_helpers.h b/drivers/gpu/drm/amd/display/dc/dm_helpers.h
index 2e4a46f1b499..5efddd48d5c5 100644
--- a/drivers/gpu/drm/amd/display/dc/dm_helpers.h
+++ b/drivers/gpu/drm/amd/display/dc/dm_helpers.h
@@ -158,6 +158,11 @@ bool dm_helpers_dp_write_dsc_enable(
const struct dc_stream_state *stream,
bool enable
);
+
+bool dm_helpers_dp_write_hblank_reduction(
+ struct dc_context *ctx,
+ const struct dc_stream_state *stream);
+
bool dm_helpers_is_dp_sink_present(
struct dc_link *link);
diff --git a/drivers/gpu/drm/amd/display/dc/dml/calcs/dcn_calcs.c b/drivers/gpu/drm/amd/display/dc/dml/calcs/dcn_calcs.c
index 39525721c976..f1235bf9a596 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/calcs/dcn_calcs.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/calcs/dcn_calcs.c
@@ -1312,138 +1312,6 @@ bool dcn_validate_bandwidth(
return false;
}
-static unsigned int dcn_find_normalized_clock_vdd_Level(
- const struct dc *dc,
- enum dm_pp_clock_type clocks_type,
- int clocks_in_khz)
-{
- int vdd_level = dcn_bw_v_min0p65;
-
- if (clocks_in_khz == 0)/*todo some clock not in the considerations*/
- return vdd_level;
-
- switch (clocks_type) {
- case DM_PP_CLOCK_TYPE_DISPLAY_CLK:
- if (clocks_in_khz > dc->dcn_soc->max_dispclk_vmax0p9*1000) {
- vdd_level = dcn_bw_v_max0p91;
- BREAK_TO_DEBUGGER();
- } else if (clocks_in_khz > dc->dcn_soc->max_dispclk_vnom0p8*1000) {
- vdd_level = dcn_bw_v_max0p9;
- } else if (clocks_in_khz > dc->dcn_soc->max_dispclk_vmid0p72*1000) {
- vdd_level = dcn_bw_v_nom0p8;
- } else if (clocks_in_khz > dc->dcn_soc->max_dispclk_vmin0p65*1000) {
- vdd_level = dcn_bw_v_mid0p72;
- } else
- vdd_level = dcn_bw_v_min0p65;
- break;
- case DM_PP_CLOCK_TYPE_DISPLAYPHYCLK:
- if (clocks_in_khz > dc->dcn_soc->phyclkv_max0p9*1000) {
- vdd_level = dcn_bw_v_max0p91;
- BREAK_TO_DEBUGGER();
- } else if (clocks_in_khz > dc->dcn_soc->phyclkv_nom0p8*1000) {
- vdd_level = dcn_bw_v_max0p9;
- } else if (clocks_in_khz > dc->dcn_soc->phyclkv_mid0p72*1000) {
- vdd_level = dcn_bw_v_nom0p8;
- } else if (clocks_in_khz > dc->dcn_soc->phyclkv_min0p65*1000) {
- vdd_level = dcn_bw_v_mid0p72;
- } else
- vdd_level = dcn_bw_v_min0p65;
- break;
-
- case DM_PP_CLOCK_TYPE_DPPCLK:
- if (clocks_in_khz > dc->dcn_soc->max_dppclk_vmax0p9*1000) {
- vdd_level = dcn_bw_v_max0p91;
- BREAK_TO_DEBUGGER();
- } else if (clocks_in_khz > dc->dcn_soc->max_dppclk_vnom0p8*1000) {
- vdd_level = dcn_bw_v_max0p9;
- } else if (clocks_in_khz > dc->dcn_soc->max_dppclk_vmid0p72*1000) {
- vdd_level = dcn_bw_v_nom0p8;
- } else if (clocks_in_khz > dc->dcn_soc->max_dppclk_vmin0p65*1000) {
- vdd_level = dcn_bw_v_mid0p72;
- } else
- vdd_level = dcn_bw_v_min0p65;
- break;
-
- case DM_PP_CLOCK_TYPE_MEMORY_CLK:
- {
- unsigned factor = (ddr4_dram_factor_single_Channel * dc->dcn_soc->number_of_channels);
-
- if (clocks_in_khz > dc->dcn_soc->fabric_and_dram_bandwidth_vmax0p9*1000000/factor) {
- vdd_level = dcn_bw_v_max0p91;
- BREAK_TO_DEBUGGER();
- } else if (clocks_in_khz > dc->dcn_soc->fabric_and_dram_bandwidth_vnom0p8*1000000/factor) {
- vdd_level = dcn_bw_v_max0p9;
- } else if (clocks_in_khz > dc->dcn_soc->fabric_and_dram_bandwidth_vmid0p72*1000000/factor) {
- vdd_level = dcn_bw_v_nom0p8;
- } else if (clocks_in_khz > dc->dcn_soc->fabric_and_dram_bandwidth_vmin0p65*1000000/factor) {
- vdd_level = dcn_bw_v_mid0p72;
- } else
- vdd_level = dcn_bw_v_min0p65;
- }
- break;
-
- case DM_PP_CLOCK_TYPE_DCFCLK:
- if (clocks_in_khz > dc->dcn_soc->dcfclkv_max0p9*1000) {
- vdd_level = dcn_bw_v_max0p91;
- BREAK_TO_DEBUGGER();
- } else if (clocks_in_khz > dc->dcn_soc->dcfclkv_nom0p8*1000) {
- vdd_level = dcn_bw_v_max0p9;
- } else if (clocks_in_khz > dc->dcn_soc->dcfclkv_mid0p72*1000) {
- vdd_level = dcn_bw_v_nom0p8;
- } else if (clocks_in_khz > dc->dcn_soc->dcfclkv_min0p65*1000) {
- vdd_level = dcn_bw_v_mid0p72;
- } else
- vdd_level = dcn_bw_v_min0p65;
- break;
-
- default:
- break;
- }
- return vdd_level;
-}
-
-unsigned int dcn_find_dcfclk_suits_all(
- const struct dc *dc,
- struct dc_clocks *clocks)
-{
- unsigned vdd_level, vdd_level_temp;
- unsigned dcf_clk;
-
- /*find a common supported voltage level*/
- vdd_level = dcn_find_normalized_clock_vdd_Level(
- dc, DM_PP_CLOCK_TYPE_DISPLAY_CLK, clocks->dispclk_khz);
- vdd_level_temp = dcn_find_normalized_clock_vdd_Level(
- dc, DM_PP_CLOCK_TYPE_DISPLAYPHYCLK, clocks->phyclk_khz);
-
- vdd_level = dcn_bw_max(vdd_level, vdd_level_temp);
- vdd_level_temp = dcn_find_normalized_clock_vdd_Level(
- dc, DM_PP_CLOCK_TYPE_DPPCLK, clocks->dppclk_khz);
- vdd_level = dcn_bw_max(vdd_level, vdd_level_temp);
-
- vdd_level_temp = dcn_find_normalized_clock_vdd_Level(
- dc, DM_PP_CLOCK_TYPE_MEMORY_CLK, clocks->fclk_khz);
- vdd_level = dcn_bw_max(vdd_level, vdd_level_temp);
- vdd_level_temp = dcn_find_normalized_clock_vdd_Level(
- dc, DM_PP_CLOCK_TYPE_DCFCLK, clocks->dcfclk_khz);
-
- /*find that level conresponding dcfclk*/
- vdd_level = dcn_bw_max(vdd_level, vdd_level_temp);
- if (vdd_level == dcn_bw_v_max0p91) {
- BREAK_TO_DEBUGGER();
- dcf_clk = dc->dcn_soc->dcfclkv_max0p9*1000;
- } else if (vdd_level == dcn_bw_v_max0p9)
- dcf_clk = dc->dcn_soc->dcfclkv_max0p9*1000;
- else if (vdd_level == dcn_bw_v_nom0p8)
- dcf_clk = dc->dcn_soc->dcfclkv_nom0p8*1000;
- else if (vdd_level == dcn_bw_v_mid0p72)
- dcf_clk = dc->dcn_soc->dcfclkv_mid0p72*1000;
- else
- dcf_clk = dc->dcn_soc->dcfclkv_min0p65*1000;
-
- DC_LOG_BANDWIDTH_CALCS("\tdcf_clk for voltage = %d\n", dcf_clk);
- return dcf_clk;
-}
-
void dcn_bw_update_from_pplib_fclks(
struct dc *dc,
struct dm_pp_clock_levels_with_voltage *fclks)
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_rq_dlg_calc_30.c b/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_rq_dlg_calc_30.c
index 76d3bb3c9155..8d4873f80df0 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_rq_dlg_calc_30.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn30/display_rq_dlg_calc_30.c
@@ -1562,6 +1562,7 @@ static void dml_rq_dlg_get_dlg_params(struct display_mode_lib *mode_lib,
dml_print("DML_DLG: %s: disp_dlg_regs->dst_y_per_row_vblank = 0x%x\n", __func__, disp_dlg_regs->dst_y_per_row_vblank);
dml_print("DML_DLG: %s: disp_dlg_regs->dst_y_per_vm_flip = 0x%x\n", __func__, disp_dlg_regs->dst_y_per_vm_flip);
dml_print("DML_DLG: %s: disp_dlg_regs->dst_y_per_row_flip = 0x%x\n", __func__, disp_dlg_regs->dst_y_per_row_flip);
+
disp_dlg_regs->refcyc_per_pte_group_vblank_l =
(unsigned int)(dst_y_per_row_vblank * (double)htotal
* ref_freq_to_pix_freq / (double)dpte_groups_per_row_ub_l);
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_util_32.c b/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_util_32.c
index 86ac7d59fd32..0748ef36a16a 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_util_32.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_util_32.c
@@ -1595,6 +1595,7 @@ double dml32_TruncToValidBPP(
unsigned int NonDSCBPP0;
unsigned int NonDSCBPP1;
unsigned int NonDSCBPP2;
+ unsigned int NonDSCBPP3 = BPP_INVALID;
if (Format == dm_420) {
NonDSCBPP0 = 12;
@@ -1603,6 +1604,7 @@ double dml32_TruncToValidBPP(
MinDSCBPP = 6;
MaxDSCBPP = 1.5 * DSCInputBitPerComponent - 1.0 / 16;
} else if (Format == dm_444) {
+ NonDSCBPP3 = 18;
NonDSCBPP0 = 24;
NonDSCBPP1 = 30;
NonDSCBPP2 = 36;
@@ -1667,6 +1669,8 @@ double dml32_TruncToValidBPP(
return NonDSCBPP1;
else if (MaxLinkBPP >= NonDSCBPP0)
return 16.0;
+ else if ((Output == dm_dp2p0 || Output == dm_dp) && NonDSCBPP3 != BPP_INVALID && MaxLinkBPP >= NonDSCBPP3)
+ return NonDSCBPP3; // Special case to allow 6bpc RGB for DP connections.
else
return BPP_INVALID;
}
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c
index beed7adbbd43..47d785204f29 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c
@@ -195,9 +195,9 @@ struct _vcs_dpi_soc_bounding_box_st dcn3_5_soc = {
.dcn_downspread_percent = 0.5,
.gpuvm_min_page_size_bytes = 4096,
.hostvm_min_page_size_bytes = 4096,
- .do_urgent_latency_adjustment = 1,
+ .do_urgent_latency_adjustment = 0,
.urgent_latency_adjustment_fabric_clock_component_us = 0,
- .urgent_latency_adjustment_fabric_clock_reference_mhz = 3000,
+ .urgent_latency_adjustment_fabric_clock_reference_mhz = 0,
};
void dcn35_build_wm_range_table_fpu(struct clk_mgr *clk_mgr)
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dml_inline_defs.h b/drivers/gpu/drm/amd/display/dc/dml/dml_inline_defs.h
index 072bd0539605..6b2ab4ec2b5f 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dml_inline_defs.h
+++ b/drivers/gpu/drm/amd/display/dc/dml/dml_inline_defs.h
@@ -66,11 +66,15 @@ static inline double dml_max5(double a, double b, double c, double d, double e)
static inline double dml_ceil(double a, double granularity)
{
+ if (granularity == 0)
+ return 0;
return (double) dcn_bw_ceil2(a, granularity);
}
static inline double dml_floor(double a, double granularity)
{
+ if (granularity == 0)
+ return 0;
return (double) dcn_bw_floor2(a, granularity);
}
@@ -114,11 +118,15 @@ static inline double dml_ceil_2(double f)
static inline double dml_ceil_ex(double x, double granularity)
{
+ if (granularity == 0)
+ return 0;
return (double) dcn_bw_ceil2(x, granularity);
}
static inline double dml_floor_ex(double x, double granularity)
{
+ if (granularity == 0)
+ return 0;
return (double) dcn_bw_floor2(x, granularity);
}
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/Makefile b/drivers/gpu/drm/amd/display/dc/dml2/Makefile
index c4378e620cbf..91c4f3b4bd5f 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/dml2/Makefile
@@ -29,7 +29,11 @@ dml2_rcflags := $(CC_FLAGS_NO_FPU)
ifneq ($(CONFIG_FRAME_WARN),0)
ifeq ($(filter y,$(CONFIG_KASAN)$(CONFIG_KCSAN)),y)
+ifeq ($(CONFIG_CC_IS_CLANG)$(CONFIG_COMPILE_TEST),yy)
+frame_warn_flag := -Wframe-larger-than=4096
+else
frame_warn_flag := -Wframe-larger-than=3072
+endif
else
frame_warn_flag := -Wframe-larger-than=2048
endif
@@ -73,9 +77,8 @@ AMD_DAL_DML2 = $(addprefix $(AMDDALPATH)/dc/dml2/,$(DML2))
AMD_DISPLAY_FILES += $(AMD_DAL_DML2)
-CFLAGS_$(AMDDALPATH)/dc/dml2/dml21/src/dml2_top/dml_top.o := $(dml2_ccflags)
-CFLAGS_$(AMDDALPATH)/dc/dml2/dml21/src/dml2_top/dml_top_mcache.o := $(dml2_ccflags)
-CFLAGS_$(AMDDALPATH)/dc/dml2/dml21/src/dml2_top/dml2_top_optimization := $(dml2_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml2/dml21/src/dml2_top/dml2_top_interfaces.o := $(dml2_ccflags)
+CFLAGS_$(AMDDALPATH)/dc/dml2/dml21/src/dml2_top/dml2_top_soc15.o := $(dml2_ccflags)
CFLAGS_$(AMDDALPATH)/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4.o := $(dml2_ccflags)
CFLAGS_$(AMDDALPATH)/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.o := $(dml2_ccflags) $(frame_warn_flag)
CFLAGS_$(AMDDALPATH)/dc/dml2/dml21/src/dml2_core/dml2_core_factory.o := $(dml2_ccflags)
@@ -94,9 +97,8 @@ CFLAGS_$(AMDDALPATH)/dc/dml2/dml21/dml21_translation_helper.o := $(dml2_ccflags)
CFLAGS_$(AMDDALPATH)/dc/dml2/dml21/dml21_utils.o := $(dml2_ccflags)
CFLAGS_$(AMDDALPATH)/dc/dml2/dml21/inc/dml2_debug.o := $(dml2_ccflags)
-CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2/dml21/src/dml2_top/dml_top.o := $(dml2_rcflags)
-CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2/dml21/src/dml2_top/dml_top_mcache.o := $(dml2_rcflags)
-CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2/dml21/src/dml2_top/dml2_top_optimization.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2/dml21/src/dml2_top/dml2_top_interfaces.o := $(dml2_rcflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2/dml21/src/dml2_top/dml2_top_soc15.o := $(dml2_rcflags)
CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4.o := $(dml2_rcflags)
CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.o := $(dml2_rcflags)
CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2/dml21/src/dml2_core/dml2_core_factory.o := $(dml2_rcflags)
@@ -113,9 +115,8 @@ CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2/dml21/dml21_translation_helper.o := $(dml2_r
CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2/dml21/dml21_utils.o := $(dml2_rcflags)
CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml2/dml21/inc/dml2_debug.o := $(dml2_rcflags)
-DML21 := src/dml2_top/dml_top.o
-DML21 += src/dml2_top/dml_top_mcache.o
-DML21 += src/dml2_top/dml2_top_optimization.o
+DML21 := src/dml2_top/dml2_top_interfaces.o
+DML21 += src/dml2_top/dml2_top_soc15.o
DML21 += src/inc/dml2_debug.o
DML21 += src/dml2_core/dml2_core_dcn4.o
DML21 += src/dml2_core/dml2_core_factory.o
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/display_mode_core.c b/drivers/gpu/drm/amd/display/dc/dml2/display_mode_core.c
index d851c081e376..35bc917631ae 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/display_mode_core.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/display_mode_core.c
@@ -1222,6 +1222,7 @@ static dml_bool_t CalculatePrefetchSchedule(struct display_mode_lib_scratch_st *
s->dst_y_prefetch_oto = s->Tvm_oto_lines + 2 * s->Tr0_oto_lines + s->Lsw_oto;
s->dst_y_prefetch_equ = p->VStartup - (*p->TSetup + dml_max(p->TWait + p->TCalc, *p->Tdmdl)) / s->LineTime - (*p->DSTYAfterScaler + (dml_float_t) *p->DSTXAfterScaler / (dml_float_t)p->myPipe->HTotal);
+ s->dst_y_prefetch_equ = dml_min(s->dst_y_prefetch_equ, 63.75); // limit to the reg limit of U6.2 for DST_Y_PREFETCH
#ifdef __DML_VBA_DEBUG__
dml_print("DML::%s: HTotal = %u\n", __func__, p->myPipe->HTotal);
@@ -6300,9 +6301,9 @@ static void dml_prefetch_check(struct display_mode_lib_st *mode_lib)
mode_lib->ms.meta_row_bandwidth_this_state,
mode_lib->ms.dpte_row_bandwidth_this_state,
mode_lib->ms.NoOfDPPThisState,
- mode_lib->ms.UrgentBurstFactorLuma,
- mode_lib->ms.UrgentBurstFactorChroma,
- mode_lib->ms.UrgentBurstFactorCursor);
+ mode_lib->ms.UrgentBurstFactorLuma[j],
+ mode_lib->ms.UrgentBurstFactorChroma[j],
+ mode_lib->ms.UrgentBurstFactorCursor[j]);
s->VMDataOnlyReturnBWPerState = dml_get_return_bw_mbps_vm_only(
&mode_lib->ms.soc,
@@ -6433,7 +6434,7 @@ static void dml_prefetch_check(struct display_mode_lib_st *mode_lib)
/* Output */
&mode_lib->ms.UrgentBurstFactorCursorPre[k],
&mode_lib->ms.UrgentBurstFactorLumaPre[k],
- &mode_lib->ms.UrgentBurstFactorChroma[k],
+ &mode_lib->ms.UrgentBurstFactorChromaPre[k],
&mode_lib->ms.NotUrgentLatencyHidingPre[k]);
mode_lib->ms.cursor_bw_pre[k] = mode_lib->ms.cache_display_cfg.plane.NumberOfCursors[k] * mode_lib->ms.cache_display_cfg.plane.CursorWidth[k] *
@@ -6457,9 +6458,9 @@ static void dml_prefetch_check(struct display_mode_lib_st *mode_lib)
mode_lib->ms.cursor_bw_pre,
mode_lib->ms.prefetch_vmrow_bw,
mode_lib->ms.NoOfDPPThisState,
- mode_lib->ms.UrgentBurstFactorLuma,
- mode_lib->ms.UrgentBurstFactorChroma,
- mode_lib->ms.UrgentBurstFactorCursor,
+ mode_lib->ms.UrgentBurstFactorLuma[j],
+ mode_lib->ms.UrgentBurstFactorChroma[j],
+ mode_lib->ms.UrgentBurstFactorCursor[j],
mode_lib->ms.UrgentBurstFactorLumaPre,
mode_lib->ms.UrgentBurstFactorChromaPre,
mode_lib->ms.UrgentBurstFactorCursorPre,
@@ -6516,9 +6517,9 @@ static void dml_prefetch_check(struct display_mode_lib_st *mode_lib)
mode_lib->ms.cursor_bw,
mode_lib->ms.cursor_bw_pre,
mode_lib->ms.NoOfDPPThisState,
- mode_lib->ms.UrgentBurstFactorLuma,
- mode_lib->ms.UrgentBurstFactorChroma,
- mode_lib->ms.UrgentBurstFactorCursor,
+ mode_lib->ms.UrgentBurstFactorLuma[j],
+ mode_lib->ms.UrgentBurstFactorChroma[j],
+ mode_lib->ms.UrgentBurstFactorCursor[j],
mode_lib->ms.UrgentBurstFactorLumaPre,
mode_lib->ms.UrgentBurstFactorChromaPre,
mode_lib->ms.UrgentBurstFactorCursorPre);
@@ -6585,9 +6586,9 @@ static void dml_prefetch_check(struct display_mode_lib_st *mode_lib)
mode_lib->ms.cursor_bw_pre,
mode_lib->ms.prefetch_vmrow_bw,
mode_lib->ms.NoOfDPP[j], // VBA_ERROR DPPPerSurface is not assigned at this point, should use NoOfDpp here
- mode_lib->ms.UrgentBurstFactorLuma,
- mode_lib->ms.UrgentBurstFactorChroma,
- mode_lib->ms.UrgentBurstFactorCursor,
+ mode_lib->ms.UrgentBurstFactorLuma[j],
+ mode_lib->ms.UrgentBurstFactorChroma[j],
+ mode_lib->ms.UrgentBurstFactorCursor[j],
mode_lib->ms.UrgentBurstFactorLumaPre,
mode_lib->ms.UrgentBurstFactorChromaPre,
mode_lib->ms.UrgentBurstFactorCursorPre,
@@ -7808,9 +7809,9 @@ dml_bool_t dml_core_mode_support(struct display_mode_lib_st *mode_lib)
mode_lib->ms.DETBufferSizeYThisState[k],
mode_lib->ms.DETBufferSizeCThisState[k],
/* Output */
- &mode_lib->ms.UrgentBurstFactorCursor[k],
- &mode_lib->ms.UrgentBurstFactorLuma[k],
- &mode_lib->ms.UrgentBurstFactorChroma[k],
+ &mode_lib->ms.UrgentBurstFactorCursor[j][k],
+ &mode_lib->ms.UrgentBurstFactorLuma[j][k],
+ &mode_lib->ms.UrgentBurstFactorChroma[j][k],
&mode_lib->ms.NotUrgentLatencyHiding[k]);
}
@@ -8317,7 +8318,7 @@ void dml_core_mode_programming(struct display_mode_lib_st *mode_lib, const struc
if (clk_cfg->dcfclk_option != dml_use_override_freq)
locals->Dcfclk = mode_lib->ms.DCFCLK;
else
- locals->Dcfclk = clk_cfg->dcfclk_freq_mhz;
+ locals->Dcfclk = clk_cfg->dcfclk_mhz;
#ifdef __DML_VBA_DEBUG__
dml_print_dml_policy(&mode_lib->ms.policy);
@@ -8370,7 +8371,7 @@ void dml_core_mode_programming(struct display_mode_lib_st *mode_lib, const struc
if (clk_cfg->dispclk_option == dml_use_required_freq)
locals->Dispclk = locals->Dispclk_calculated;
else if (clk_cfg->dispclk_option == dml_use_override_freq)
- locals->Dispclk = clk_cfg->dispclk_freq_mhz;
+ locals->Dispclk = clk_cfg->dispclk_mhz;
else
locals->Dispclk = mode_lib->ms.state.dispclk_mhz;
#ifdef __DML_VBA_DEBUG__
@@ -8411,7 +8412,7 @@ void dml_core_mode_programming(struct display_mode_lib_st *mode_lib, const struc
if (clk_cfg->dppclk_option[k] == dml_use_required_freq)
locals->Dppclk[k] = locals->Dppclk_calculated[k];
else if (clk_cfg->dppclk_option[k] == dml_use_override_freq)
- locals->Dppclk[k] = clk_cfg->dppclk_freq_mhz[k];
+ locals->Dppclk[k] = clk_cfg->dppclk_mhz[k];
else
locals->Dppclk[k] = mode_lib->ms.state.dppclk_mhz;
#ifdef __DML_VBA_DEBUG__
@@ -9189,6 +9190,8 @@ void dml_core_mode_programming(struct display_mode_lib_st *mode_lib, const struc
&locals->FractionOfUrgentBandwidth,
&s->dummy_boolean[0]); // dml_bool_t *PrefetchBandwidthSupport
+
+
if (s->VRatioPrefetchMoreThanMax != false || s->DestinationLineTimesForPrefetchLessThan2 != false) {
dml_print("DML::%s: VRatioPrefetchMoreThanMax = %u\n", __func__, s->VRatioPrefetchMoreThanMax);
dml_print("DML::%s: DestinationLineTimesForPrefetchLessThan2 = %u\n", __func__, s->DestinationLineTimesForPrefetchLessThan2);
@@ -9203,6 +9206,7 @@ void dml_core_mode_programming(struct display_mode_lib_st *mode_lib, const struc
}
}
+
if (locals->PrefetchModeSupported == true && mode_lib->ms.support.ImmediateFlipSupport == true) {
locals->BandwidthAvailableForImmediateFlip = CalculateBandwidthAvailableForImmediateFlip(
mode_lib->ms.num_active_planes,
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/display_mode_core_structs.h b/drivers/gpu/drm/amd/display/dc/dml2/display_mode_core_structs.h
index f951936bb579..dd3f43181a6e 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/display_mode_core_structs.h
+++ b/drivers/gpu/drm/amd/display/dc/dml2/display_mode_core_structs.h
@@ -28,6 +28,7 @@
#define __DISPLAY_MODE_CORE_STRUCT_H__
#include "display_mode_lib_defines.h"
+#include "dml_top_display_cfg_types.h"
enum dml_project_id {
dml_project_invalid = 0,
@@ -49,7 +50,9 @@ enum dml_use_mall_for_pstate_change_mode {
dml_use_mall_pstate_change_disable = 0,
dml_use_mall_pstate_change_full_frame = 1,
dml_use_mall_pstate_change_sub_viewport = 2,
- dml_use_mall_pstate_change_phantom_pipe = 3
+ dml_use_mall_pstate_change_phantom_pipe = 3,
+ dml_use_mall_pstate_change_phantom_pipe_no_data_return = 4,
+ dml_use_mall_pstate_change_imall = 5
};
enum dml_use_mall_for_static_screen_mode {
dml_use_mall_static_screen_disable = 0,
@@ -171,7 +174,11 @@ enum dml_swizzle_mode {
dml_sw_256kb_z_x = 28,
dml_sw_256kb_s_x = 29,
dml_sw_256kb_d_x = 30,
- dml_sw_256kb_r_x = 31
+ dml_sw_256kb_r_x = 31,
+ dml_sw_256b_2d = 32,
+ dml_sw_4kb_2d = 33,
+ dml_sw_64kb_2d = 34,
+ dml_sw_256kb_2d = 35
};
enum dml_lb_depth {
dml_lb_6 = 0,
@@ -223,24 +230,28 @@ enum dml_mpc_use_policy {
dml_mpc_disabled = 0,
dml_mpc_as_possible = 1,
dml_mpc_as_needed_for_voltage = 2,
- dml_mpc_as_needed_for_pstate_and_voltage = 3
+ dml_mpc_as_needed_for_pstate_and_voltage = 3,
+ dml_mpc_as_needed = 4,
+ dml_mpc_2to1 = 5
};
enum dml_odm_use_policy {
dml_odm_use_policy_bypass = 0,
dml_odm_use_policy_combine_as_needed = 1,
dml_odm_use_policy_combine_2to1 = 2,
- dml_odm_use_policy_combine_4to1 = 3,
- dml_odm_use_policy_split_1to2 = 4,
- dml_odm_use_policy_mso_1to2 = 5,
- dml_odm_use_policy_mso_1to4 = 6
+ dml_odm_use_policy_combine_3to1 = 3,
+ dml_odm_use_policy_combine_4to1 = 4,
+ dml_odm_use_policy_split_1to2 = 5,
+ dml_odm_use_policy_mso_1to2 = 6,
+ dml_odm_use_policy_mso_1to4 = 7
};
enum dml_odm_mode {
dml_odm_mode_bypass = 0,
dml_odm_mode_combine_2to1 = 1,
- dml_odm_mode_combine_4to1 = 2,
- dml_odm_mode_split_1to2 = 3,
- dml_odm_mode_mso_1to2 = 4,
- dml_odm_mode_mso_1to4 = 5
+ dml_odm_mode_combine_3to1 = 2,
+ dml_odm_mode_combine_4to1 = 3,
+ dml_odm_mode_split_1to2 = 4,
+ dml_odm_mode_mso_1to2 = 5,
+ dml_odm_mode_mso_1to4 = 6
};
enum dml_writeback_configuration {
dml_whole_buffer_for_single_stream_no_interleave = 0,
@@ -289,6 +300,17 @@ struct soc_state_bounding_box_st {
dml_float_t fclk_change_latency_us;
dml_float_t usr_retraining_latency_us;
dml_bool_t use_ideal_dram_bw_strobe;
+ dml_float_t g6_temp_read_blackout_us;
+
+ struct {
+ dml_uint_t urgent_ramp_uclk_cycles;
+ dml_uint_t trip_to_memory_uclk_cycles;
+ dml_uint_t meta_trip_to_memory_uclk_cycles;
+ dml_uint_t maximum_latency_when_urgent_uclk_cycles;
+ dml_uint_t average_latency_when_urgent_uclk_cycles;
+ dml_uint_t maximum_latency_when_non_urgent_uclk_cycles;
+ dml_uint_t average_latency_when_non_urgent_uclk_cycles;
+ } dml_dcn401_uclk_dpm_dependent_soc_qos_params;
};
struct soc_bounding_box_st {
@@ -297,7 +319,7 @@ struct soc_bounding_box_st {
dml_float_t pcierefclk_mhz;
dml_float_t refclk_mhz;
dml_float_t amclk_mhz;
- dml_float_t max_outstanding_reqs;
+ dml_uint_t max_outstanding_reqs;
dml_float_t pct_ideal_sdp_bw_after_urgent;
dml_float_t pct_ideal_fabric_bw_after_urgent;
dml_float_t pct_ideal_dram_bw_after_urgent_pixel_only;
@@ -308,6 +330,16 @@ struct soc_bounding_box_st {
dml_float_t max_avg_fabric_bw_use_normal_percent;
dml_float_t max_avg_dram_bw_use_normal_percent;
dml_float_t max_avg_dram_bw_use_normal_strobe_percent;
+
+ dml_float_t svp_prefetch_pct_ideal_sdp_bw_after_urgent;
+ dml_float_t svp_prefetch_pct_ideal_fabric_bw_after_urgent;
+ dml_float_t svp_prefetch_pct_ideal_dram_bw_after_urgent_pixel_only;
+ dml_float_t svp_prefetch_pct_ideal_dram_bw_after_urgent_pixel_and_vm;
+ dml_float_t svp_prefetch_pct_ideal_dram_bw_after_urgent_vm_only;
+ dml_float_t svp_prefetch_max_avg_sdp_bw_use_normal_percent;
+ dml_float_t svp_prefetch_max_avg_fabric_bw_use_normal_percent;
+ dml_float_t svp_prefetch_max_avg_dram_bw_use_normal_percent;
+
dml_uint_t round_trip_ping_latency_dcfclk_cycles;
dml_uint_t urgent_out_of_order_return_per_channel_pixel_only_bytes;
dml_uint_t urgent_out_of_order_return_per_channel_pixel_and_vm_bytes;
@@ -324,6 +356,26 @@ struct soc_bounding_box_st {
dml_uint_t mall_allocated_for_dcn_mbytes;
dml_float_t dispclk_dppclk_vco_speed_mhz;
dml_bool_t do_urgent_latency_adjustment;
+
+ dml_uint_t mem_word_bytes;
+ dml_uint_t num_dcc_mcaches;
+ dml_uint_t mcache_size_bytes;
+ dml_uint_t mcache_line_size_bytes;
+
+ struct {
+ dml_bool_t UseNewDCN401SOCParameters;
+ dml_uint_t df_qos_response_time_fclk_cycles;
+ dml_uint_t max_round_trip_to_furthest_cs_fclk_cycles;
+ dml_uint_t mall_overhead_fclk_cycles;
+ dml_uint_t meta_trip_adder_fclk_cycles;
+ dml_uint_t average_transport_distance_fclk_cycles;
+ dml_float_t umc_urgent_ramp_latency_margin;
+ dml_float_t umc_max_latency_margin;
+ dml_float_t umc_average_latency_margin;
+ dml_float_t fabric_max_transport_latency_margin;
+ dml_float_t fabric_average_transport_latency_margin;
+ } dml_dcn401_soc_qos_params;
+
};
struct ip_params_st {
@@ -515,6 +567,10 @@ struct dml_plane_cfg_st {
dml_uint_t CursorWidth[__DML_NUM_PLANES__];
dml_uint_t CursorBPP[__DML_NUM_PLANES__];
+ dml_bool_t setup_for_tdlut[__DML_NUM_PLANES__];
+ enum dml2_tdlut_addressing_mode tdlut_addressing_mode[__DML_NUM_PLANES__];
+ enum dml2_tdlut_width_mode tdlut_width_mode[__DML_NUM_PLANES__];
+
enum dml_use_mall_for_static_screen_mode UseMALLForStaticScreen[__DML_NUM_PLANES__];
enum dml_use_mall_for_pstate_change_mode UseMALLForPStateChange[__DML_NUM_PLANES__];
@@ -604,6 +660,17 @@ struct dml_hw_resource_st {
dml_float_t DLGRefClkFreqMHz; /// <brief DLG Global Reference timer
};
+/// @brief To control the clk usage for model programming
+struct dml_clk_cfg_st {
+ enum dml_clk_cfg_policy dcfclk_option; ///< brief Use for mode_program; user can select between use the min require clk req as calculated by DML or use the test-specific freq
+ enum dml_clk_cfg_policy dispclk_option; ///< brief Use for mode_program; user can select between use the min require clk req as calculated by DML or use the test-specific freq
+ enum dml_clk_cfg_policy dppclk_option[__DML_NUM_PLANES__];
+
+ dml_float_t dcfclk_mhz;
+ dml_float_t dispclk_mhz;
+ dml_float_t dppclk_mhz[__DML_NUM_PLANES__];
+}; // dml_clk_cfg_st
+
/// @brief DML display configuration.
/// Describe how to display a surface in multi-plane setup and output to different output and writeback using the specified timgin
struct dml_display_cfg_st {
@@ -616,19 +683,9 @@ struct dml_display_cfg_st {
unsigned int num_timings;
struct dml_hw_resource_st hw; //< brief for mode programming
+ struct dml_clk_cfg_st clk_overrides; //< brief for mode programming clk override
}; // dml_display_cfg_st
-/// @brief To control the clk usage for model programming
-struct dml_clk_cfg_st {
- enum dml_clk_cfg_policy dcfclk_option; ///< brief Use for mode_program; user can select between use the min require clk req as calculated by DML or use the test-specific freq
- enum dml_clk_cfg_policy dispclk_option; ///< brief Use for mode_program; user can select between use the min require clk req as calculated by DML or use the test-specific freq
- enum dml_clk_cfg_policy dppclk_option[__DML_NUM_PLANES__];
-
- dml_float_t dcfclk_freq_mhz;
- dml_float_t dispclk_freq_mhz;
- dml_float_t dppclk_freq_mhz[__DML_NUM_PLANES__];
-}; // dml_clk_cfg_st
-
/// @brief DML mode evaluation and programming policy
/// Those knobs that affect mode support and mode programming
struct dml_mode_eval_policy_st {
@@ -884,11 +941,11 @@ struct mode_support_st {
dml_uint_t meta_row_height[__DML_NUM_PLANES__];
dml_uint_t meta_row_height_chroma[__DML_NUM_PLANES__];
dml_float_t UrgLatency;
- dml_float_t UrgentBurstFactorCursor[__DML_NUM_PLANES__];
+ dml_float_t UrgentBurstFactorCursor[2][__DML_NUM_PLANES__];
dml_float_t UrgentBurstFactorCursorPre[__DML_NUM_PLANES__];
- dml_float_t UrgentBurstFactorLuma[__DML_NUM_PLANES__];
+ dml_float_t UrgentBurstFactorLuma[2][__DML_NUM_PLANES__];
dml_float_t UrgentBurstFactorLumaPre[__DML_NUM_PLANES__];
- dml_float_t UrgentBurstFactorChroma[__DML_NUM_PLANES__];
+ dml_float_t UrgentBurstFactorChroma[2][__DML_NUM_PLANES__];
dml_float_t UrgentBurstFactorChromaPre[__DML_NUM_PLANES__];
dml_float_t MaximumSwathWidthInLineBufferLuma;
dml_float_t MaximumSwathWidthInLineBufferChroma;
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/display_mode_util.c b/drivers/gpu/drm/amd/display/dc/dml2/display_mode_util.c
index c247aee89caf..89890c88fd66 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/display_mode_util.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/display_mode_util.c
@@ -690,12 +690,12 @@ __DML_DLL_EXPORT__ void dml_print_clk_cfg(const struct dml_clk_cfg_st *clk_cfg)
dml_print("DML: clk_cfg: dcfclk_option = %d\n", clk_cfg->dcfclk_option);
dml_print("DML: clk_cfg: dispclk_option = %d\n", clk_cfg->dispclk_option);
- dml_print("DML: clk_cfg: dcfclk_freq_mhz = %f\n", clk_cfg->dcfclk_freq_mhz);
- dml_print("DML: clk_cfg: dispclk_freq_mhz = %f\n", clk_cfg->dispclk_freq_mhz);
+ dml_print("DML: clk_cfg: dcfclk_mhz = %f\n", clk_cfg->dcfclk_mhz);
+ dml_print("DML: clk_cfg: dispclk_mhz = %f\n", clk_cfg->dispclk_mhz);
for (dml_uint_t i = 0; i < DCN_DML__NUM_PLANE; i++) {
dml_print("DML: clk_cfg: i=%d, dppclk_option = %d\n", i, clk_cfg->dppclk_option[i]);
- dml_print("DML: clk_cfg: i=%d, dppclk_freq_mhz = %f\n", i, clk_cfg->dppclk_freq_mhz[i]);
+ dml_print("DML: clk_cfg: i=%d, dppclk_mhz = %f\n", i, clk_cfg->dppclk_mhz[i]);
}
}
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.c
index 138b4b1e42ed..b9c6b45f6872 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.c
@@ -10,7 +10,6 @@
#include "dml21_utils.h"
#include "dml21_translation_helper.h"
#include "bounding_boxes/dcn4_soc_bb.h"
-#include "bounding_boxes/dcn3_soc_bb.h"
static void dml21_init_socbb_params(struct dml2_initialize_instance_in_out *dml_init,
const struct dml2_configuration_options *config,
@@ -20,10 +19,6 @@ static void dml21_init_socbb_params(struct dml2_initialize_instance_in_out *dml_
const struct dml2_soc_qos_parameters *qos_params;
switch (in_dc->ctx->dce_version) {
- case DCN_VERSION_3_2: // TODO : Temporary for N-1 validation. Remove this after N-1 validation phase is complete.
- soc_bb = &dml2_socbb_dcn31;
- qos_params = &dml_dcn31_soc_qos_params;
- break;
case DCN_VERSION_4_01:
default:
if (config->bb_from_dmub)
@@ -60,9 +55,6 @@ static void dml21_init_ip_params(struct dml2_initialize_instance_in_out *dml_ini
const struct dml2_ip_capabilities *ip_caps;
switch (in_dc->ctx->dce_version) {
- case DCN_VERSION_3_2: // TODO : Temporary for N-1 validation. Remove this after N-1 validation phase is complete.
- ip_caps = &dml2_dcn31_max_ip_caps;
- break;
case DCN_VERSION_4_01:
default:
ip_caps = &dml2_dcn401_max_ip_caps;
@@ -302,12 +294,17 @@ void dml21_apply_soc_bb_overrides(struct dml2_initialize_instance_in_out *dml_in
dml_soc_bb->power_management_parameters.stutter_exit_latency_us =
(in_dc->ctx->dc_bios->bb_info.dram_sr_exit_latency_100ns + 9) / 10;
- if (in_dc->ctx->dc_bios->vram_info.num_chans) {
+ if (dc_bw_params->num_channels) {
+ dml_clk_table->dram_config.channel_count = dc_bw_params->num_channels;
+ dml_soc_bb->mall_allocated_for_dcn_mbytes = in_dc->caps.mall_size_total / 1048576;
+ } else if (in_dc->ctx->dc_bios->vram_info.num_chans) {
dml_clk_table->dram_config.channel_count = in_dc->ctx->dc_bios->vram_info.num_chans;
dml_soc_bb->mall_allocated_for_dcn_mbytes = in_dc->caps.mall_size_total / 1048576;
}
- if (in_dc->ctx->dc_bios->vram_info.dram_channel_width_bytes) {
+ if (dc_bw_params->dram_channel_width_bytes) {
+ dml_clk_table->dram_config.channel_width_bytes = dc_bw_params->dram_channel_width_bytes;
+ } else if (in_dc->ctx->dc_bios->vram_info.dram_channel_width_bytes) {
dml_clk_table->dram_config.channel_width_bytes = in_dc->ctx->dc_bios->vram_info.dram_channel_width_bytes;
}
@@ -339,11 +336,22 @@ void dml21_apply_soc_bb_overrides(struct dml2_initialize_instance_in_out *dml_in
// }
}
+static unsigned int calc_max_hardware_v_total(const struct dc_stream_state *stream)
+{
+ unsigned int max_hw_v_total = stream->ctx->dc->caps.max_v_total;
+
+ if (stream->ctx->dc->caps.vtotal_limited_by_fp2) {
+ max_hw_v_total -= stream->timing.v_front_porch + 1;
+ }
+
+ return max_hw_v_total;
+}
+
static void populate_dml21_timing_config_from_stream_state(struct dml2_timing_cfg *timing,
struct dc_stream_state *stream,
struct dml2_context *dml_ctx)
{
- unsigned int hblank_start, vblank_start;
+ unsigned int hblank_start, vblank_start, min_hardware_refresh_in_uhz;
timing->h_active = stream->timing.h_addressable + stream->timing.h_border_left + stream->timing.h_border_right;
timing->v_active = stream->timing.v_addressable + stream->timing.v_border_bottom + stream->timing.v_border_top;
@@ -371,11 +379,23 @@ static void populate_dml21_timing_config_from_stream_state(struct dml2_timing_cf
- stream->timing.v_border_top - stream->timing.v_border_bottom;
timing->drr_config.enabled = stream->ignore_msa_timing_param;
- timing->drr_config.min_refresh_uhz = stream->timing.min_refresh_in_uhz;
timing->drr_config.drr_active_variable = stream->vrr_active_variable;
timing->drr_config.drr_active_fixed = stream->vrr_active_fixed;
timing->drr_config.disallowed = !stream->allow_freesync;
+ /* limit min refresh rate to DC cap */
+ min_hardware_refresh_in_uhz = stream->timing.min_refresh_in_uhz;
+ if (stream->ctx->dc->caps.max_v_total != 0) {
+ min_hardware_refresh_in_uhz = div64_u64((stream->timing.pix_clk_100hz * 100000000ULL),
+ (stream->timing.h_total * (long long)calc_max_hardware_v_total(stream)));
+ }
+
+ if (stream->timing.min_refresh_in_uhz > min_hardware_refresh_in_uhz) {
+ timing->drr_config.min_refresh_uhz = stream->timing.min_refresh_in_uhz;
+ } else {
+ timing->drr_config.min_refresh_uhz = min_hardware_refresh_in_uhz;
+ }
+
if (dml_ctx->config.callbacks.get_max_flickerless_instant_vtotal_increase &&
stream->ctx->dc->config.enable_fpo_flicker_detection == 1)
timing->drr_config.max_instant_vtotal_delta = dml_ctx->config.callbacks.get_max_flickerless_instant_vtotal_increase(stream, false);
@@ -422,6 +442,21 @@ static void populate_dml21_timing_config_from_stream_state(struct dml2_timing_cf
timing->vblank_nom = timing->v_total - timing->v_active;
}
+/**
+ * adjust_dml21_hblank_timing_config_from_pipe_ctx - Adjusts the horizontal blanking timing configuration
+ * based on the pipe context.
+ * @timing: Pointer to the dml2_timing_cfg structure to be adjusted.
+ * @pipe: Pointer to the pipe_ctx structure containing the horizontal blanking borrow value.
+ *
+ * This function modifies the horizontal active and blank end timings by adding and subtracting
+ * the horizontal blanking borrow value from the pipe context, respectively.
+ */
+static void adjust_dml21_hblank_timing_config_from_pipe_ctx(struct dml2_timing_cfg *timing, struct pipe_ctx *pipe)
+{
+ timing->h_active += pipe->hblank_borrow;
+ timing->h_blank_end -= pipe->hblank_borrow;
+}
+
static void populate_dml21_output_config_from_stream_state(struct dml2_link_output_cfg *output,
struct dc_stream_state *stream, const struct pipe_ctx *pipe)
{
@@ -683,11 +718,21 @@ static void populate_dml21_surface_config_from_plane_state(
surface->dcc.informative.fraction_of_zero_size_request_plane1 = plane_state->dcc.independent_64b_blks_c;
surface->dcc.plane0.pitch = plane_state->dcc.meta_pitch;
surface->dcc.plane1.pitch = plane_state->dcc.meta_pitch_c;
- if (in_dc->ctx->dce_version < DCN_VERSION_4_01) {
- /* needed for N-1 testing */
+
+ // Update swizzle / array mode based on the gfx_format
+ switch (plane_state->tiling_info.gfxversion) {
+ case DcGfxVersion7:
+ case DcGfxVersion8:
+ // Placeholder for programming the array_mode
+ break;
+ case DcGfxVersion9:
+ case DcGfxVersion10:
+ case DcGfxVersion11:
surface->tiling = gfx9_to_dml2_swizzle_mode(plane_state->tiling_info.gfx9.swizzle);
- } else {
+ break;
+ case DcGfxAddr3:
surface->tiling = gfx_addr3_to_dml2_swizzle_mode(plane_state->tiling_info.gfx_addr3.swizzle);
+ break;
}
}
@@ -709,6 +754,7 @@ static const struct scaler_data *get_scaler_data_for_plane(
temp_pipe->plane_state = pipe->plane_state;
temp_pipe->plane_res.scl_data.taps = pipe->plane_res.scl_data.taps;
temp_pipe->stream_res = pipe->stream_res;
+ temp_pipe->hblank_borrow = pipe->hblank_borrow;
dml_ctx->config.callbacks.build_scaling_params(temp_pipe);
break;
}
@@ -973,6 +1019,7 @@ bool dml21_map_dc_state_into_dml_display_cfg(const struct dc *in_dc, struct dc_s
ASSERT(disp_cfg_stream_location >= 0 && disp_cfg_stream_location <= __DML2_WRAPPER_MAX_STREAMS_PLANES__);
populate_dml21_timing_config_from_stream_state(&dml_dispcfg->stream_descriptors[disp_cfg_stream_location].timing, context->streams[stream_index], dml_ctx);
+ adjust_dml21_hblank_timing_config_from_pipe_ctx(&dml_dispcfg->stream_descriptors[disp_cfg_stream_location].timing, &context->res_ctx.pipe_ctx[stream_index]);
populate_dml21_output_config_from_stream_state(&dml_dispcfg->stream_descriptors[disp_cfg_stream_location].output, context->streams[stream_index], &context->res_ctx.pipe_ctx[stream_index]);
populate_dml21_stream_overrides_from_stream_state(&dml_dispcfg->stream_descriptors[disp_cfg_stream_location], context->streams[stream_index]);
@@ -1037,28 +1084,8 @@ void dml21_copy_clocks_to_dc_state(struct dml2_context *in_ctx, struct dc_state
context->bw_ctx.bw.dcn.clk.dtbclk_en = in_ctx->v21.mode_programming.programming->min_clocks.dcn4x.dtbrefclk_khz > 0;
context->bw_ctx.bw.dcn.clk.ref_dtbclk_khz = in_ctx->v21.mode_programming.programming->min_clocks.dcn4x.dtbrefclk_khz;
context->bw_ctx.bw.dcn.clk.socclk_khz = in_ctx->v21.mode_programming.programming->min_clocks.dcn4x.socclk_khz;
-}
-
-void dml21_extract_legacy_watermark_set(const struct dc *in_dc, struct dcn_watermarks *watermark, enum dml2_dchub_watermark_reg_set_index reg_set_idx, struct dml2_context *in_ctx)
-{
- struct dml2_core_internal_display_mode_lib *mode_lib = &in_ctx->v21.dml_init.dml2_instance->core_instance.clean_me_up.mode_lib;
- double refclk_freq_in_mhz = (in_ctx->v21.display_config.overrides.hw.dlg_ref_clk_mhz > 0) ? (double)in_ctx->v21.display_config.overrides.hw.dlg_ref_clk_mhz : mode_lib->soc.dchub_refclk_mhz;
-
- if (reg_set_idx >= DML2_DCHUB_WATERMARK_SET_NUM) {
- /* invalid register set index */
- return;
- }
-
- /* convert to legacy format (time in ns) */
- watermark->urgent_ns = ((double)in_ctx->v21.mode_programming.programming->global_regs.wm_regs[reg_set_idx].urgent / refclk_freq_in_mhz) * 1000.0;
- watermark->pte_meta_urgent_ns = ((double)in_ctx->v21.mode_programming.programming->global_regs.wm_regs[reg_set_idx].urgent / refclk_freq_in_mhz) * 1000.0;
- watermark->cstate_pstate.cstate_enter_plus_exit_ns = ((double)in_ctx->v21.mode_programming.programming->global_regs.wm_regs[reg_set_idx].sr_enter / refclk_freq_in_mhz) * 1000.0;
- watermark->cstate_pstate.cstate_exit_ns = ((double)in_ctx->v21.mode_programming.programming->global_regs.wm_regs[reg_set_idx].sr_exit / refclk_freq_in_mhz) * 1000.0;
- watermark->cstate_pstate.pstate_change_ns = ((double)in_ctx->v21.mode_programming.programming->global_regs.wm_regs[reg_set_idx].uclk_pstate / refclk_freq_in_mhz) * 1000.0;
- watermark->urgent_latency_ns = ((double)in_ctx->v21.mode_programming.programming->global_regs.wm_regs[reg_set_idx].urgent / refclk_freq_in_mhz) * 1000.0;
- watermark->cstate_pstate.fclk_pstate_change_ns = ((double)in_ctx->v21.mode_programming.programming->global_regs.wm_regs[reg_set_idx].fclk_pstate / refclk_freq_in_mhz) * 1000.0;
- watermark->frac_urg_bw_flip = in_ctx->v21.mode_programming.programming->global_regs.wm_regs[reg_set_idx].frac_urg_bw_flip;
- watermark->frac_urg_bw_nom = in_ctx->v21.mode_programming.programming->global_regs.wm_regs[reg_set_idx].frac_urg_bw_nom;
+ context->bw_ctx.bw.dcn.clk.subvp_prefetch_dramclk_khz = in_ctx->v21.mode_programming.programming->min_clocks.dcn4x.svp_prefetch_no_throttle.uclk_khz;
+ context->bw_ctx.bw.dcn.clk.subvp_prefetch_fclk_khz = in_ctx->v21.mode_programming.programming->min_clocks.dcn4x.svp_prefetch_no_throttle.fclk_khz;
}
static struct dml2_dchub_watermark_regs *wm_set_index_to_dc_wm_set(union dcn_watermark_set *watermarks, const enum dml2_dchub_watermark_reg_set_index wm_index)
@@ -1104,53 +1131,6 @@ void dml21_extract_watermark_sets(const struct dc *in_dc, union dcn_watermark_se
}
}
-
-void dml21_populate_pipe_ctx_dlg_params(struct dml2_context *dml_ctx, struct dc_state *context, struct pipe_ctx *pipe_ctx, struct dml2_per_stream_programming *stream_programming)
-{
- unsigned int hactive, vactive, hblank_start, vblank_start, hblank_end, vblank_end;
- struct dc_crtc_timing *timing = &pipe_ctx->stream->timing;
- union dml2_global_sync_programming *global_sync = &stream_programming->global_sync;
-
- hactive = timing->h_addressable + timing->h_border_left + timing->h_border_right;
- vactive = timing->v_addressable + timing->v_border_bottom + timing->v_border_top;
- hblank_start = pipe_ctx->stream->timing.h_total - pipe_ctx->stream->timing.h_front_porch;
- vblank_start = pipe_ctx->stream->timing.v_total - pipe_ctx->stream->timing.v_front_porch;
-
- hblank_end = hblank_start - timing->h_addressable - timing->h_border_left - timing->h_border_right;
- vblank_end = vblank_start - timing->v_addressable - timing->v_border_top - timing->v_border_bottom;
-
- if (dml_ctx->config.svp_pstate.callbacks.get_pipe_subvp_type(context, pipe_ctx) == SUBVP_PHANTOM) {
- /* phantom has its own global sync */
- global_sync = &stream_programming->phantom_stream.global_sync;
- }
-
- pipe_ctx->pipe_dlg_param.vstartup_start = global_sync->dcn4x.vstartup_lines;
- pipe_ctx->pipe_dlg_param.vupdate_offset = global_sync->dcn4x.vupdate_offset_pixels;
- pipe_ctx->pipe_dlg_param.vupdate_width = global_sync->dcn4x.vupdate_vupdate_width_pixels;
- pipe_ctx->pipe_dlg_param.vready_offset = global_sync->dcn4x.vready_offset_pixels;
- pipe_ctx->pipe_dlg_param.pstate_keepout = global_sync->dcn4x.pstate_keepout_start_lines;
-
- pipe_ctx->pipe_dlg_param.otg_inst = pipe_ctx->stream_res.tg->inst;
-
- pipe_ctx->pipe_dlg_param.hactive = hactive;
- pipe_ctx->pipe_dlg_param.vactive = vactive;
- pipe_ctx->pipe_dlg_param.htotal = pipe_ctx->stream->timing.h_total;
- pipe_ctx->pipe_dlg_param.vtotal = pipe_ctx->stream->timing.v_total;
- pipe_ctx->pipe_dlg_param.hblank_end = hblank_end;
- pipe_ctx->pipe_dlg_param.vblank_end = vblank_end;
- pipe_ctx->pipe_dlg_param.hblank_start = hblank_start;
- pipe_ctx->pipe_dlg_param.vblank_start = vblank_start;
- pipe_ctx->pipe_dlg_param.vfront_porch = pipe_ctx->stream->timing.v_front_porch;
- pipe_ctx->pipe_dlg_param.pixel_rate_mhz = pipe_ctx->stream->timing.pix_clk_100hz / 10000.00;
- pipe_ctx->pipe_dlg_param.refresh_rate = ((timing->pix_clk_100hz * 100) / timing->h_total) / timing->v_total;
- pipe_ctx->pipe_dlg_param.vtotal_max = pipe_ctx->stream->adjust.v_total_max;
- pipe_ctx->pipe_dlg_param.vtotal_min = pipe_ctx->stream->adjust.v_total_min;
- pipe_ctx->pipe_dlg_param.recout_height = pipe_ctx->plane_res.scl_data.recout.height;
- pipe_ctx->pipe_dlg_param.recout_width = pipe_ctx->plane_res.scl_data.recout.width;
- pipe_ctx->pipe_dlg_param.full_recout_height = pipe_ctx->plane_res.scl_data.recout.height;
- pipe_ctx->pipe_dlg_param.full_recout_width = pipe_ctx->plane_res.scl_data.recout.width;
-}
-
void dml21_map_hw_resources(struct dml2_context *dml_ctx)
{
unsigned int i = 0;
@@ -1186,22 +1166,22 @@ void dml21_set_dc_p_state_type(
bool sub_vp_enabled)
{
switch (stream_programming->uclk_pstate_method) {
- case dml2_uclk_pstate_support_method_vactive:
- case dml2_uclk_pstate_support_method_fw_vactive_drr:
+ case dml2_pstate_method_vactive:
+ case dml2_pstate_method_fw_vactive_drr:
pipe_ctx->p_state_type = P_STATE_V_ACTIVE;
break;
- case dml2_uclk_pstate_support_method_vblank:
- case dml2_uclk_pstate_support_method_fw_vblank_drr:
+ case dml2_pstate_method_vblank:
+ case dml2_pstate_method_fw_vblank_drr:
if (sub_vp_enabled)
pipe_ctx->p_state_type = P_STATE_V_BLANK_SUB_VP;
else
pipe_ctx->p_state_type = P_STATE_V_BLANK;
break;
- case dml2_uclk_pstate_support_method_fw_subvp_phantom:
- case dml2_uclk_pstate_support_method_fw_subvp_phantom_drr:
+ case dml2_pstate_method_fw_svp:
+ case dml2_pstate_method_fw_svp_drr:
pipe_ctx->p_state_type = P_STATE_SUB_VP;
break;
- case dml2_uclk_pstate_support_method_fw_drr:
+ case dml2_pstate_method_fw_drr:
if (sub_vp_enabled)
pipe_ctx->p_state_type = P_STATE_DRR_SUB_VP;
else
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.h b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.h
index 476a7f6e4875..069b939c672a 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.h
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.h
@@ -21,8 +21,6 @@ void dml21_initialize_soc_bb_params(struct dml2_initialize_instance_in_out *dml_
void dml21_initialize_ip_params(struct dml2_initialize_instance_in_out *dml_init, const struct dml2_configuration_options *config, const struct dc *in_dc);
bool dml21_map_dc_state_into_dml_display_cfg(const struct dc *in_dc, struct dc_state *context, struct dml2_context *dml_ctx);
void dml21_copy_clocks_to_dc_state(struct dml2_context *in_ctx, struct dc_state *context);
-void dml21_populate_pipe_ctx_dlg_params(struct dml2_context *dml_ctx, struct dc_state *context, struct pipe_ctx *pipe_ctx, struct dml2_per_stream_programming *stream_programming);
-void dml21_extract_legacy_watermark_set(const struct dc *in_dc, struct dcn_watermarks *watermark, enum dml2_dchub_watermark_reg_set_index reg_set_idx, struct dml2_context *in_ctx);
void dml21_extract_watermark_sets(const struct dc *in_dc, union dcn_watermark_set *watermarks, struct dml2_context *in_ctx);
void dml21_map_hw_resources(struct dml2_context *dml_ctx);
void dml21_get_pipe_mcache_config(struct dc_state *context, struct pipe_ctx *pipe_ctx, struct dml2_per_plane_programming *pln_prog, struct dml2_pipe_configuration_descriptor *mcache_pipe_config);
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_utils.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_utils.c
index 51d491bffa32..1e56d995cd0e 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_utils.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_utils.c
@@ -142,108 +142,21 @@ int dml21_find_dc_pipes_for_plane(const struct dc *in_dc,
return num_pipes;
}
-
-void dml21_update_pipe_ctx_dchub_regs(struct dml2_display_rq_regs *rq_regs,
- struct dml2_display_dlg_regs *disp_dlg_regs,
- struct dml2_display_ttu_regs *disp_ttu_regs,
- struct pipe_ctx *out)
+void dml21_pipe_populate_global_sync(struct dml2_context *dml_ctx,
+ struct dc_state *context,
+ struct pipe_ctx *pipe_ctx,
+ struct dml2_per_stream_programming *stream_programming)
{
- memset(&out->rq_regs, 0, sizeof(out->rq_regs));
- out->rq_regs.rq_regs_l.chunk_size = rq_regs->rq_regs_l.chunk_size;
- out->rq_regs.rq_regs_l.min_chunk_size = rq_regs->rq_regs_l.min_chunk_size;
- //out->rq_regs.rq_regs_l.meta_chunk_size = rq_regs->rq_regs_l.meta_chunk_size;
- //out->rq_regs.rq_regs_l.min_meta_chunk_size = rq_regs->rq_regs_l.min_meta_chunk_size;
- out->rq_regs.rq_regs_l.dpte_group_size = rq_regs->rq_regs_l.dpte_group_size;
- out->rq_regs.rq_regs_l.mpte_group_size = rq_regs->rq_regs_l.mpte_group_size;
- out->rq_regs.rq_regs_l.swath_height = rq_regs->rq_regs_l.swath_height;
- out->rq_regs.rq_regs_l.pte_row_height_linear = rq_regs->rq_regs_l.pte_row_height_linear;
-
- out->rq_regs.rq_regs_c.chunk_size = rq_regs->rq_regs_c.chunk_size;
- out->rq_regs.rq_regs_c.min_chunk_size = rq_regs->rq_regs_c.min_chunk_size;
- //out->rq_regs.rq_regs_c.meta_chunk_size = rq_regs->rq_regs_c.meta_chunk_size;
- //out->rq_regs.rq_regs_c.min_meta_chunk_size = rq_regs->rq_regs_c.min_meta_chunk_size;
- out->rq_regs.rq_regs_c.dpte_group_size = rq_regs->rq_regs_c.dpte_group_size;
- out->rq_regs.rq_regs_c.mpte_group_size = rq_regs->rq_regs_c.mpte_group_size;
- out->rq_regs.rq_regs_c.swath_height = rq_regs->rq_regs_c.swath_height;
- out->rq_regs.rq_regs_c.pte_row_height_linear = rq_regs->rq_regs_c.pte_row_height_linear;
-
- out->rq_regs.drq_expansion_mode = rq_regs->drq_expansion_mode;
- out->rq_regs.prq_expansion_mode = rq_regs->prq_expansion_mode;
- //out->rq_regs.mrq_expansion_mode = rq_regs->mrq_expansion_mode;
- out->rq_regs.crq_expansion_mode = rq_regs->crq_expansion_mode;
- out->rq_regs.plane1_base_address = rq_regs->plane1_base_address;
- out->unbounded_req = rq_regs->unbounded_request_enabled;
-
- memset(&out->dlg_regs, 0, sizeof(out->dlg_regs));
- out->dlg_regs.refcyc_h_blank_end = disp_dlg_regs->refcyc_h_blank_end;
- out->dlg_regs.dlg_vblank_end = disp_dlg_regs->dlg_vblank_end;
- out->dlg_regs.min_dst_y_next_start = disp_dlg_regs->min_dst_y_next_start;
- out->dlg_regs.refcyc_per_htotal = disp_dlg_regs->refcyc_per_htotal;
- out->dlg_regs.refcyc_x_after_scaler = disp_dlg_regs->refcyc_x_after_scaler;
- out->dlg_regs.dst_y_after_scaler = disp_dlg_regs->dst_y_after_scaler;
- out->dlg_regs.dst_y_prefetch = disp_dlg_regs->dst_y_prefetch;
- out->dlg_regs.dst_y_per_vm_vblank = disp_dlg_regs->dst_y_per_vm_vblank;
- out->dlg_regs.dst_y_per_row_vblank = disp_dlg_regs->dst_y_per_row_vblank;
- out->dlg_regs.dst_y_per_vm_flip = disp_dlg_regs->dst_y_per_vm_flip;
- out->dlg_regs.dst_y_per_row_flip = disp_dlg_regs->dst_y_per_row_flip;
- out->dlg_regs.ref_freq_to_pix_freq = disp_dlg_regs->ref_freq_to_pix_freq;
- out->dlg_regs.vratio_prefetch = disp_dlg_regs->vratio_prefetch;
- out->dlg_regs.vratio_prefetch_c = disp_dlg_regs->vratio_prefetch_c;
- out->dlg_regs.refcyc_per_tdlut_group = disp_dlg_regs->refcyc_per_tdlut_group;
- out->dlg_regs.refcyc_per_pte_group_vblank_l = disp_dlg_regs->refcyc_per_pte_group_vblank_l;
- out->dlg_regs.refcyc_per_pte_group_vblank_c = disp_dlg_regs->refcyc_per_pte_group_vblank_c;
- //out->dlg_regs.refcyc_per_meta_chunk_vblank_l = disp_dlg_regs->refcyc_per_meta_chunk_vblank_l;
- //out->dlg_regs.refcyc_per_meta_chunk_vblank_c = disp_dlg_regs->refcyc_per_meta_chunk_vblank_c;
- out->dlg_regs.refcyc_per_pte_group_flip_l = disp_dlg_regs->refcyc_per_pte_group_flip_l;
- out->dlg_regs.refcyc_per_pte_group_flip_c = disp_dlg_regs->refcyc_per_pte_group_flip_c;
- //out->dlg_regs.refcyc_per_meta_chunk_flip_l = disp_dlg_regs->refcyc_per_meta_chunk_flip_l;
- //out->dlg_regs.refcyc_per_meta_chunk_flip_c = disp_dlg_regs->refcyc_per_meta_chunk_flip_c;
- out->dlg_regs.dst_y_per_pte_row_nom_l = disp_dlg_regs->dst_y_per_pte_row_nom_l;
- out->dlg_regs.dst_y_per_pte_row_nom_c = disp_dlg_regs->dst_y_per_pte_row_nom_c;
- out->dlg_regs.refcyc_per_pte_group_nom_l = disp_dlg_regs->refcyc_per_pte_group_nom_l;
- out->dlg_regs.refcyc_per_pte_group_nom_c = disp_dlg_regs->refcyc_per_pte_group_nom_c;
- //out->dlg_regs.dst_y_per_meta_row_nom_l = disp_dlg_regs->dst_y_per_meta_row_nom_l;
- //out->dlg_regs.dst_y_per_meta_row_nom_c = disp_dlg_regs->dst_y_per_meta_row_nom_c;
- //out->dlg_regs.refcyc_per_meta_chunk_nom_l = disp_dlg_regs->refcyc_per_meta_chunk_nom_l;
- //out->dlg_regs.refcyc_per_meta_chunk_nom_c = disp_dlg_regs->refcyc_per_meta_chunk_nom_c;
- out->dlg_regs.refcyc_per_line_delivery_pre_l = disp_dlg_regs->refcyc_per_line_delivery_pre_l;
- out->dlg_regs.refcyc_per_line_delivery_pre_c = disp_dlg_regs->refcyc_per_line_delivery_pre_c;
- out->dlg_regs.refcyc_per_line_delivery_l = disp_dlg_regs->refcyc_per_line_delivery_l;
- out->dlg_regs.refcyc_per_line_delivery_c = disp_dlg_regs->refcyc_per_line_delivery_c;
- out->dlg_regs.refcyc_per_vm_group_vblank = disp_dlg_regs->refcyc_per_vm_group_vblank;
- out->dlg_regs.refcyc_per_vm_group_flip = disp_dlg_regs->refcyc_per_vm_group_flip;
- out->dlg_regs.refcyc_per_vm_req_vblank = disp_dlg_regs->refcyc_per_vm_req_vblank;
- out->dlg_regs.refcyc_per_vm_req_flip = disp_dlg_regs->refcyc_per_vm_req_flip;
- out->dlg_regs.dst_y_offset_cur0 = disp_dlg_regs->dst_y_offset_cur0;
- out->dlg_regs.chunk_hdl_adjust_cur0 = disp_dlg_regs->chunk_hdl_adjust_cur0;
- //out->dlg_regs.dst_y_offset_cur1 = disp_dlg_regs->dst_y_offset_cur1;
- //out->dlg_regs.chunk_hdl_adjust_cur1 = disp_dlg_regs->chunk_hdl_adjust_cur1;
- out->dlg_regs.vready_after_vcount0 = disp_dlg_regs->vready_after_vcount0;
- out->dlg_regs.dst_y_delta_drq_limit = disp_dlg_regs->dst_y_delta_drq_limit;
- out->dlg_regs.refcyc_per_vm_dmdata = disp_dlg_regs->refcyc_per_vm_dmdata;
- out->dlg_regs.dmdata_dl_delta = disp_dlg_regs->dmdata_dl_delta;
-
- memset(&out->ttu_regs, 0, sizeof(out->ttu_regs));
- out->ttu_regs.qos_level_low_wm = disp_ttu_regs->qos_level_low_wm;
- out->ttu_regs.qos_level_high_wm = disp_ttu_regs->qos_level_high_wm;
- out->ttu_regs.min_ttu_vblank = disp_ttu_regs->min_ttu_vblank;
- out->ttu_regs.qos_level_flip = disp_ttu_regs->qos_level_flip;
- out->ttu_regs.refcyc_per_req_delivery_l = disp_ttu_regs->refcyc_per_req_delivery_l;
- out->ttu_regs.refcyc_per_req_delivery_c = disp_ttu_regs->refcyc_per_req_delivery_c;
- out->ttu_regs.refcyc_per_req_delivery_cur0 = disp_ttu_regs->refcyc_per_req_delivery_cur0;
- //out->ttu_regs.refcyc_per_req_delivery_cur1 = disp_ttu_regs->refcyc_per_req_delivery_cur1;
- out->ttu_regs.refcyc_per_req_delivery_pre_l = disp_ttu_regs->refcyc_per_req_delivery_pre_l;
- out->ttu_regs.refcyc_per_req_delivery_pre_c = disp_ttu_regs->refcyc_per_req_delivery_pre_c;
- out->ttu_regs.refcyc_per_req_delivery_pre_cur0 = disp_ttu_regs->refcyc_per_req_delivery_pre_cur0;
- //out->ttu_regs.refcyc_per_req_delivery_pre_cur1 = disp_ttu_regs->refcyc_per_req_delivery_pre_cur1;
- out->ttu_regs.qos_level_fixed_l = disp_ttu_regs->qos_level_fixed_l;
- out->ttu_regs.qos_level_fixed_c = disp_ttu_regs->qos_level_fixed_c;
- out->ttu_regs.qos_level_fixed_cur0 = disp_ttu_regs->qos_level_fixed_cur0;
- //out->ttu_regs.qos_level_fixed_cur1 = disp_ttu_regs->qos_level_fixed_cur1;
- out->ttu_regs.qos_ramp_disable_l = disp_ttu_regs->qos_ramp_disable_l;
- out->ttu_regs.qos_ramp_disable_c = disp_ttu_regs->qos_ramp_disable_c;
- out->ttu_regs.qos_ramp_disable_cur0 = disp_ttu_regs->qos_ramp_disable_cur0;
- //out->ttu_regs.qos_ramp_disable_cur1 = disp_ttu_regs->qos_ramp_disable_cur1;
+ union dml2_global_sync_programming *global_sync = &stream_programming->global_sync;
+
+ if (dml_ctx->config.svp_pstate.callbacks.get_pipe_subvp_type(context, pipe_ctx) == SUBVP_PHANTOM) {
+ /* phantom has its own global sync */
+ global_sync = &stream_programming->phantom_stream.global_sync;
+ }
+
+ memcpy(&pipe_ctx->global_sync,
+ global_sync,
+ sizeof(union dml2_global_sync_programming));
}
void dml21_populate_mall_allocation_size(struct dc_state *context,
@@ -301,28 +214,16 @@ void dml21_program_dc_pipe(struct dml2_context *dml_ctx, struct dc_state *contex
{
unsigned int pipe_reg_index = 0;
- dml21_populate_pipe_ctx_dlg_params(dml_ctx, context, pipe_ctx, stream_prog);
+ dml21_pipe_populate_global_sync(dml_ctx, context, pipe_ctx, stream_prog);
find_pipe_regs_idx(dml_ctx, pipe_ctx, &pipe_reg_index);
if (dml_ctx->config.svp_pstate.callbacks.get_pipe_subvp_type(context, pipe_ctx) == SUBVP_PHANTOM) {
memcpy(&pipe_ctx->hubp_regs, pln_prog->phantom_plane.pipe_regs[pipe_reg_index], sizeof(struct dml2_dchub_per_pipe_register_set));
pipe_ctx->unbounded_req = false;
-
- /* legacy only, should be removed later */
- dml21_update_pipe_ctx_dchub_regs(&pln_prog->phantom_plane.pipe_regs[pipe_reg_index]->rq_regs,
- &pln_prog->phantom_plane.pipe_regs[pipe_reg_index]->dlg_regs,
- &pln_prog->phantom_plane.pipe_regs[pipe_reg_index]->ttu_regs, pipe_ctx);
-
pipe_ctx->det_buffer_size_kb = 0;
} else {
memcpy(&pipe_ctx->hubp_regs, pln_prog->pipe_regs[pipe_reg_index], sizeof(struct dml2_dchub_per_pipe_register_set));
pipe_ctx->unbounded_req = pln_prog->pipe_regs[pipe_reg_index]->rq_regs.unbounded_request_enabled;
-
- /* legacy only, should be removed later */
- dml21_update_pipe_ctx_dchub_regs(&pln_prog->pipe_regs[pipe_reg_index]->rq_regs,
- &pln_prog->pipe_regs[pipe_reg_index]->dlg_regs,
- &pln_prog->pipe_regs[pipe_reg_index]->ttu_regs, pipe_ctx);
-
pipe_ctx->det_buffer_size_kb = pln_prog->pipe_regs[pipe_reg_index]->det_size * 64;
}
@@ -482,7 +383,8 @@ void dml21_build_fams2_programming(const struct dc *dc,
unsigned int num_fams2_streams = 0;
/* reset fams2 data */
- memset(&context->bw_ctx.bw.dcn.fams2_stream_params, 0, sizeof(struct dmub_fams2_stream_static_state) * DML2_MAX_PLANES);
+ memset(&context->bw_ctx.bw.dcn.fams2_stream_base_params, 0, sizeof(union dmub_cmd_fams2_config) * DML2_MAX_PLANES);
+ memset(&context->bw_ctx.bw.dcn.fams2_stream_sub_params, 0, sizeof(union dmub_cmd_fams2_config) * DML2_MAX_PLANES);
memset(&context->bw_ctx.bw.dcn.fams2_global_config, 0, sizeof(struct dmub_cmd_fams2_global_config));
if (dml_ctx->v21.mode_programming.programming->fams2_required) {
@@ -490,8 +392,10 @@ void dml21_build_fams2_programming(const struct dc *dc,
int dml_stream_idx;
struct dc_stream_state *phantom_stream;
struct dc_stream_status *phantom_status;
+ enum fams2_stream_type type = 0;
- struct dmub_fams2_stream_static_state *static_state = &context->bw_ctx.bw.dcn.fams2_stream_params[num_fams2_streams];
+ union dmub_cmd_fams2_config *static_base_state = &context->bw_ctx.bw.dcn.fams2_stream_base_params[num_fams2_streams];
+ union dmub_cmd_fams2_config *static_sub_state = &context->bw_ctx.bw.dcn.fams2_stream_sub_params[num_fams2_streams];
struct dc_stream_state *stream = context->streams[i];
@@ -508,28 +412,38 @@ void dml21_build_fams2_programming(const struct dc *dc,
}
/* copy static state from PMO */
- memcpy(static_state,
- &dml_ctx->v21.mode_programming.programming->stream_programming[dml_stream_idx].fams2_params,
- sizeof(struct dmub_fams2_stream_static_state));
-
- /* get information from context */
- static_state->num_planes = context->stream_status[i].plane_count;
- static_state->otg_inst = context->stream_status[i].primary_otg_inst;
-
- /* populate pipe masks for planes */
- for (j = 0; j < context->stream_status[i].plane_count; j++) {
- for (k = 0; k < dc->res_pool->pipe_count; k++) {
- if (context->res_ctx.pipe_ctx[k].stream &&
- context->res_ctx.pipe_ctx[k].stream->stream_id == stream->stream_id &&
- context->res_ctx.pipe_ctx[k].plane_state == context->stream_status[i].plane_states[j]) {
- static_state->pipe_mask |= (1 << k);
- static_state->plane_pipe_masks[j] |= (1 << k);
+ memcpy(static_base_state,
+ &dml_ctx->v21.mode_programming.programming->stream_programming[dml_stream_idx].fams2_base_params,
+ sizeof(union dmub_cmd_fams2_config));
+ memcpy(static_sub_state,
+ &dml_ctx->v21.mode_programming.programming->stream_programming[dml_stream_idx].fams2_sub_params,
+ sizeof(union dmub_cmd_fams2_config));
+
+ switch (dc->debug.fams_version.minor) {
+ case 1:
+ default:
+ type = static_base_state->stream_v1.base.type;
+
+ /* get information from context */
+ static_base_state->stream_v1.base.num_planes = context->stream_status[i].plane_count;
+ static_base_state->stream_v1.base.otg_inst = context->stream_status[i].primary_otg_inst;
+
+ /* populate pipe masks for planes */
+ for (j = 0; j < context->stream_status[i].plane_count; j++) {
+ for (k = 0; k < dc->res_pool->pipe_count; k++) {
+ if (context->res_ctx.pipe_ctx[k].stream &&
+ context->res_ctx.pipe_ctx[k].stream->stream_id == stream->stream_id &&
+ context->res_ctx.pipe_ctx[k].plane_state == context->stream_status[i].plane_states[j]) {
+ static_base_state->stream_v1.base.pipe_mask |= (1 << k);
+ static_base_state->stream_v1.base.plane_pipe_masks[j] |= (1 << k);
+ }
}
}
}
+
/* get per method programming */
- switch (static_state->type) {
+ switch (type) {
case FAMS2_STREAM_TYPE_VBLANK:
case FAMS2_STREAM_TYPE_VACTIVE:
case FAMS2_STREAM_TYPE_DRR:
@@ -543,16 +457,27 @@ void dml21_build_fams2_programming(const struct dc *dc,
/* phantom status should always be present */
ASSERT(phantom_status);
- static_state->sub_state.subvp.phantom_otg_inst = phantom_status->primary_otg_inst;
+ if (!phantom_status)
+ break;
- /* populate pipe masks for phantom planes */
- for (j = 0; j < phantom_status->plane_count; j++) {
- for (k = 0; k < dc->res_pool->pipe_count; k++) {
- if (context->res_ctx.pipe_ctx[k].stream &&
- context->res_ctx.pipe_ctx[k].stream->stream_id == phantom_stream->stream_id &&
- context->res_ctx.pipe_ctx[k].plane_state == phantom_status->plane_states[j]) {
- static_state->sub_state.subvp.phantom_pipe_mask |= (1 << k);
- static_state->sub_state.subvp.phantom_plane_pipe_masks[j] |= (1 << k);
+ switch (dc->debug.fams_version.minor) {
+ case 1:
+ default:
+ static_sub_state->stream_v1.sub_state.subvp.phantom_otg_inst = phantom_status->primary_otg_inst;
+
+ /* populate pipe masks for phantom planes */
+ for (j = 0; j < phantom_status->plane_count; j++) {
+ for (k = 0; k < dc->res_pool->pipe_count; k++) {
+ if (context->res_ctx.pipe_ctx[k].stream &&
+ context->res_ctx.pipe_ctx[k].stream->stream_id == phantom_stream->stream_id &&
+ context->res_ctx.pipe_ctx[k].plane_state == phantom_status->plane_states[j]) {
+ switch (dc->debug.fams_version.minor) {
+ case 1:
+ default:
+ static_sub_state->stream_v1.sub_state.subvp.phantom_pipe_mask |= (1 << k);
+ static_sub_state->stream_v1.sub_state.subvp.phantom_plane_pipe_masks[j] |= (1 << k);
+ }
+ }
}
}
}
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_utils.h b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_utils.h
index d5153fbac921..4bff52eaaef8 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_utils.h
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_utils.h
@@ -18,10 +18,10 @@ struct dml2_display_ttu_regs;
int dml21_helper_find_dml_pipe_idx_by_stream_id(struct dml2_context *ctx, unsigned int stream_id);
int dml21_find_dml_pipe_idx_by_plane_id(struct dml2_context *ctx, unsigned int plane_id);
bool dml21_get_plane_id(const struct dc_state *state, const struct dc_plane_state *plane, unsigned int *plane_id);
-void dml21_update_pipe_ctx_dchub_regs(struct dml2_display_rq_regs *rq_regs,
- struct dml2_display_dlg_regs *disp_dlg_regs,
- struct dml2_display_ttu_regs *disp_ttu_regs,
- struct pipe_ctx *out);
+void dml21_pipe_populate_global_sync(struct dml2_context *dml_ctx,
+ struct dc_state *context,
+ struct pipe_ctx *pipe_ctx,
+ struct dml2_per_stream_programming *stream_programming);
void dml21_populate_mall_allocation_size(struct dc_state *context,
struct dml2_context *in_ctx,
struct dml2_per_plane_programming *pln_prog,
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_wrapper.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_wrapper.c
index bbc28b9a15a3..fb80ba9287b6 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_wrapper.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_wrapper.c
@@ -75,7 +75,6 @@ static void dml21_init(const struct dc *in_dc, struct dml2_context **dml_ctx, co
{
switch (in_dc->ctx->dce_version) {
case DCN_VERSION_4_01:
- case DCN_VERSION_3_2: // TODO : Temporary for N-1 validation. Remove this after N-1 validation phase is complete.
(*dml_ctx)->v21.dml_init.options.project_id = dml2_project_dcn4x_stage2_auto_drr_svp;
break;
default:
@@ -233,13 +232,6 @@ static bool dml21_mode_check_and_programming(const struct dc *in_dc, struct dc_s
dml21_calculate_rq_and_dlg_params(in_dc, context, &context->res_ctx, dml_ctx, in_dc->res_pool->pipe_count);
dml21_copy_clocks_to_dc_state(dml_ctx, context);
dml21_extract_watermark_sets(in_dc, &context->bw_ctx.bw.dcn.watermarks, dml_ctx);
- if (in_dc->ctx->dce_version == DCN_VERSION_3_2) {
- dml21_extract_legacy_watermark_set(in_dc, &context->bw_ctx.bw.dcn.watermarks.a, DML2_DCHUB_WATERMARK_SET_A, dml_ctx);
- dml21_extract_legacy_watermark_set(in_dc, &context->bw_ctx.bw.dcn.watermarks.b, DML2_DCHUB_WATERMARK_SET_A, dml_ctx);
- dml21_extract_legacy_watermark_set(in_dc, &context->bw_ctx.bw.dcn.watermarks.c, DML2_DCHUB_WATERMARK_SET_A, dml_ctx);
- dml21_extract_legacy_watermark_set(in_dc, &context->bw_ctx.bw.dcn.watermarks.d, DML2_DCHUB_WATERMARK_SET_A, dml_ctx);
- }
-
dml21_build_fams2_programming(in_dc, context, dml_ctx);
}
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/bounding_boxes/dcn3_soc_bb.h b/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/bounding_boxes/dcn3_soc_bb.h
deleted file mode 100644
index d82c681a5402..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/bounding_boxes/dcn3_soc_bb.h
+++ /dev/null
@@ -1,401 +0,0 @@
-/*
- * Copyright 2022 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: AMD
- *
- */
-
-#ifndef __DML_DML_DCN3_SOC_BB__
-#define __DML_DML_DCN3_SOC_BB__
-
-#include "dml_top_soc_parameter_types.h"
-
-static const struct dml2_soc_qos_parameters dml_dcn31_soc_qos_params = {
- .derate_table = {
- .system_active_urgent = {
- .dram_derate_percent_pixel = 22,
- .dram_derate_percent_vm = 0,
- .dram_derate_percent_pixel_and_vm = 0,
- .fclk_derate_percent = 76,
- .dcfclk_derate_percent = 100,
- },
- .system_active_average = {
- .dram_derate_percent_pixel = 17,
- .dram_derate_percent_vm = 0,
- .dram_derate_percent_pixel_and_vm = 0,
- .fclk_derate_percent = 57,
- .dcfclk_derate_percent = 75,
- },
- .dcn_mall_prefetch_urgent = {
- .dram_derate_percent_pixel = 22,
- .dram_derate_percent_vm = 0,
- .dram_derate_percent_pixel_and_vm = 0,
- .fclk_derate_percent = 76,
- .dcfclk_derate_percent = 100,
- },
- .dcn_mall_prefetch_average = {
- .dram_derate_percent_pixel = 17,
- .dram_derate_percent_vm = 0,
- .dram_derate_percent_pixel_and_vm = 0,
- .fclk_derate_percent = 57,
- .dcfclk_derate_percent = 75,
- },
- .system_idle_average = {
- .dram_derate_percent_pixel = 17,
- .dram_derate_percent_vm = 0,
- .dram_derate_percent_pixel_and_vm = 0,
- .fclk_derate_percent = 57,
- .dcfclk_derate_percent = 100,
- },
- },
- .writeback = {
- .base_latency_us = 12,
- .scaling_factor_us = 0,
- .scaling_factor_mhz = 0,
- },
- .qos_params = {
- .dcn4x = {
- .df_qos_response_time_fclk_cycles = 300,
- .max_round_trip_to_furthest_cs_fclk_cycles = 350,
- .mall_overhead_fclk_cycles = 50,
- .meta_trip_adder_fclk_cycles = 36,
- .average_transport_distance_fclk_cycles = 257,
- .umc_urgent_ramp_latency_margin = 50,
- .umc_max_latency_margin = 30,
- .umc_average_latency_margin = 20,
- .fabric_max_transport_latency_margin = 20,
- .fabric_average_transport_latency_margin = 10,
-
- .per_uclk_dpm_params = {
- {
- .minimum_uclk_khz = 97,
- .urgent_ramp_uclk_cycles = 472,
- .trip_to_memory_uclk_cycles = 827,
- .meta_trip_to_memory_uclk_cycles = 827,
- .maximum_latency_when_urgent_uclk_cycles = 72,
- .average_latency_when_urgent_uclk_cycles = 61,
- .maximum_latency_when_non_urgent_uclk_cycles = 827,
- .average_latency_when_non_urgent_uclk_cycles = 118,
- },
- {
- .minimum_uclk_khz = 435,
- .urgent_ramp_uclk_cycles = 546,
- .trip_to_memory_uclk_cycles = 848,
- .meta_trip_to_memory_uclk_cycles = 848,
- .maximum_latency_when_urgent_uclk_cycles = 146,
- .average_latency_when_urgent_uclk_cycles = 90,
- .maximum_latency_when_non_urgent_uclk_cycles = 848,
- .average_latency_when_non_urgent_uclk_cycles = 135,
- },
- {
- .minimum_uclk_khz = 731,
- .urgent_ramp_uclk_cycles = 632,
- .trip_to_memory_uclk_cycles = 874,
- .meta_trip_to_memory_uclk_cycles = 874,
- .maximum_latency_when_urgent_uclk_cycles = 232,
- .average_latency_when_urgent_uclk_cycles = 124,
- .maximum_latency_when_non_urgent_uclk_cycles = 874,
- .average_latency_when_non_urgent_uclk_cycles = 155,
- },
- {
- .minimum_uclk_khz = 1187,
- .urgent_ramp_uclk_cycles = 716,
- .trip_to_memory_uclk_cycles = 902,
- .meta_trip_to_memory_uclk_cycles = 902,
- .maximum_latency_when_urgent_uclk_cycles = 316,
- .average_latency_when_urgent_uclk_cycles = 160,
- .maximum_latency_when_non_urgent_uclk_cycles = 902,
- .average_latency_when_non_urgent_uclk_cycles = 177,
- },
- },
- },
- },
- .qos_type = dml2_qos_param_type_dcn4x,
-};
-
-static const struct dml2_soc_bb dml2_socbb_dcn31 = {
- .clk_table = {
- .uclk = {
- .clk_values_khz = {97000, 435000, 731000, 1187000},
- .num_clk_values = 4,
- },
- .fclk = {
- .clk_values_khz = {300000, 2500000},
- .num_clk_values = 2,
- },
- .dcfclk = {
- .clk_values_khz = {200000, 1800000},
- .num_clk_values = 2,
- },
- .dispclk = {
- .clk_values_khz = {100000, 2000000},
- .num_clk_values = 2,
- },
- .dppclk = {
- .clk_values_khz = {100000, 2000000},
- .num_clk_values = 2,
- },
- .dtbclk = {
- .clk_values_khz = {100000, 2000000},
- .num_clk_values = 2,
- },
- .phyclk = {
- .clk_values_khz = {810000, 810000},
- .num_clk_values = 2,
- },
- .socclk = {
- .clk_values_khz = {300000, 1600000},
- .num_clk_values = 2,
- },
- .dscclk = {
- .clk_values_khz = {666667, 666667},
- .num_clk_values = 2,
- },
- .phyclk_d18 = {
- .clk_values_khz = {625000, 625000},
- .num_clk_values = 2,
- },
- .phyclk_d32 = {
- .clk_values_khz = {2000000, 2000000},
- .num_clk_values = 2,
- },
- .dram_config = {
- .channel_width_bytes = 2,
- .channel_count = 16,
- .transactions_per_clock = 16,
- },
- },
-
- .qos_parameters = {
- .derate_table = {
- .system_active_urgent = {
- .dram_derate_percent_pixel = 22,
- .dram_derate_percent_vm = 0,
- .dram_derate_percent_pixel_and_vm = 0,
- .fclk_derate_percent = 76,
- .dcfclk_derate_percent = 100,
- },
- .system_active_average = {
- .dram_derate_percent_pixel = 17,
- .dram_derate_percent_vm = 0,
- .dram_derate_percent_pixel_and_vm = 0,
- .fclk_derate_percent = 57,
- .dcfclk_derate_percent = 75,
- },
- .dcn_mall_prefetch_urgent = {
- .dram_derate_percent_pixel = 22,
- .dram_derate_percent_vm = 0,
- .dram_derate_percent_pixel_and_vm = 0,
- .fclk_derate_percent = 76,
- .dcfclk_derate_percent = 100,
- },
- .dcn_mall_prefetch_average = {
- .dram_derate_percent_pixel = 17,
- .dram_derate_percent_vm = 0,
- .dram_derate_percent_pixel_and_vm = 0,
- .fclk_derate_percent = 57,
- .dcfclk_derate_percent = 75,
- },
- .system_idle_average = {
- .dram_derate_percent_pixel = 17,
- .dram_derate_percent_vm = 0,
- .dram_derate_percent_pixel_and_vm = 0,
- .fclk_derate_percent = 57,
- .dcfclk_derate_percent = 100,
- },
- },
- .writeback = {
- .base_latency_us = 0,
- .scaling_factor_us = 0,
- .scaling_factor_mhz = 0,
- },
- .qos_params = {
- .dcn4x = {
- .df_qos_response_time_fclk_cycles = 300,
- .max_round_trip_to_furthest_cs_fclk_cycles = 350,
- .mall_overhead_fclk_cycles = 50,
- .meta_trip_adder_fclk_cycles = 36,
- .average_transport_distance_fclk_cycles = 260,
- .umc_urgent_ramp_latency_margin = 50,
- .umc_max_latency_margin = 30,
- .umc_average_latency_margin = 20,
- .fabric_max_transport_latency_margin = 20,
- .fabric_average_transport_latency_margin = 10,
-
- .per_uclk_dpm_params = {
- {
- // State 1
- .minimum_uclk_khz = 0,
- .urgent_ramp_uclk_cycles = 472,
- .trip_to_memory_uclk_cycles = 827,
- .meta_trip_to_memory_uclk_cycles = 827,
- .maximum_latency_when_urgent_uclk_cycles = 72,
- .average_latency_when_urgent_uclk_cycles = 72,
- .maximum_latency_when_non_urgent_uclk_cycles = 827,
- .average_latency_when_non_urgent_uclk_cycles = 117,
- },
- {
- // State 2
- .minimum_uclk_khz = 0,
- .urgent_ramp_uclk_cycles = 546,
- .trip_to_memory_uclk_cycles = 848,
- .meta_trip_to_memory_uclk_cycles = 848,
- .maximum_latency_when_urgent_uclk_cycles = 146,
- .average_latency_when_urgent_uclk_cycles = 146,
- .maximum_latency_when_non_urgent_uclk_cycles = 848,
- .average_latency_when_non_urgent_uclk_cycles = 133,
- },
- {
- // State 3
- .minimum_uclk_khz = 0,
- .urgent_ramp_uclk_cycles = 564,
- .trip_to_memory_uclk_cycles = 853,
- .meta_trip_to_memory_uclk_cycles = 853,
- .maximum_latency_when_urgent_uclk_cycles = 164,
- .average_latency_when_urgent_uclk_cycles = 164,
- .maximum_latency_when_non_urgent_uclk_cycles = 853,
- .average_latency_when_non_urgent_uclk_cycles = 136,
- },
- {
- // State 4
- .minimum_uclk_khz = 0,
- .urgent_ramp_uclk_cycles = 613,
- .trip_to_memory_uclk_cycles = 869,
- .meta_trip_to_memory_uclk_cycles = 869,
- .maximum_latency_when_urgent_uclk_cycles = 213,
- .average_latency_when_urgent_uclk_cycles = 213,
- .maximum_latency_when_non_urgent_uclk_cycles = 869,
- .average_latency_when_non_urgent_uclk_cycles = 149,
- },
- {
- // State 5
- .minimum_uclk_khz = 0,
- .urgent_ramp_uclk_cycles = 632,
- .trip_to_memory_uclk_cycles = 874,
- .meta_trip_to_memory_uclk_cycles = 874,
- .maximum_latency_when_urgent_uclk_cycles = 232,
- .average_latency_when_urgent_uclk_cycles = 232,
- .maximum_latency_when_non_urgent_uclk_cycles = 874,
- .average_latency_when_non_urgent_uclk_cycles = 153,
- },
- {
- // State 6
- .minimum_uclk_khz = 0,
- .urgent_ramp_uclk_cycles = 665,
- .trip_to_memory_uclk_cycles = 885,
- .meta_trip_to_memory_uclk_cycles = 885,
- .maximum_latency_when_urgent_uclk_cycles = 265,
- .average_latency_when_urgent_uclk_cycles = 265,
- .maximum_latency_when_non_urgent_uclk_cycles = 885,
- .average_latency_when_non_urgent_uclk_cycles = 161,
- },
- {
- // State 7
- .minimum_uclk_khz = 0,
- .urgent_ramp_uclk_cycles = 689,
- .trip_to_memory_uclk_cycles = 895,
- .meta_trip_to_memory_uclk_cycles = 895,
- .maximum_latency_when_urgent_uclk_cycles = 289,
- .average_latency_when_urgent_uclk_cycles = 289,
- .maximum_latency_when_non_urgent_uclk_cycles = 895,
- .average_latency_when_non_urgent_uclk_cycles = 167,
- },
- {
- // State 8
- .minimum_uclk_khz = 0,
- .urgent_ramp_uclk_cycles = 716,
- .trip_to_memory_uclk_cycles = 902,
- .meta_trip_to_memory_uclk_cycles = 902,
- .maximum_latency_when_urgent_uclk_cycles = 316,
- .average_latency_when_urgent_uclk_cycles = 316,
- .maximum_latency_when_non_urgent_uclk_cycles = 902,
- .average_latency_when_non_urgent_uclk_cycles = 174,
- },
- },
- },
- },
- .qos_type = dml2_qos_param_type_dcn4x,
- },
-
- .power_management_parameters = {
- .dram_clk_change_blackout_us = 400,
- .fclk_change_blackout_us = 0,
- .g7_ppt_blackout_us = 0,
- .stutter_enter_plus_exit_latency_us = 50,
- .stutter_exit_latency_us = 43,
- .z8_stutter_enter_plus_exit_latency_us = 0,
- .z8_stutter_exit_latency_us = 0,
- },
-
- .vmin_limit = {
- .dispclk_khz = 600 * 1000,
- },
-
- .dprefclk_mhz = 700,
- .xtalclk_mhz = 100,
- .pcie_refclk_mhz = 100,
- .dchub_refclk_mhz = 50,
- .mall_allocated_for_dcn_mbytes = 64,
- .max_outstanding_reqs = 512,
- .fabric_datapath_to_dcn_data_return_bytes = 64,
- .return_bus_width_bytes = 64,
- .hostvm_min_page_size_kbytes = 0,
- .gpuvm_min_page_size_kbytes = 256,
- .phy_downspread_percent = 0,
- .dcn_downspread_percent = 0,
- .dispclk_dppclk_vco_speed_mhz = 4500,
- .do_urgent_latency_adjustment = 0,
- .mem_word_bytes = 32,
- .num_dcc_mcaches = 8,
- .mcache_size_bytes = 2048,
- .mcache_line_size_bytes = 32,
- .max_fclk_for_uclk_dpm_khz = 1250 * 1000,
-};
-
-static const struct dml2_ip_capabilities dml2_dcn31_max_ip_caps = {
- .pipe_count = 4,
- .otg_count = 4,
- .num_dsc = 4,
- .max_num_dp2p0_streams = 4,
- .max_num_hdmi_frl_outputs = 1,
- .max_num_dp2p0_outputs = 4,
- .rob_buffer_size_kbytes = 192,
- .config_return_buffer_size_in_kbytes = 1152,
- .meta_fifo_size_in_kentries = 22,
- .compressed_buffer_segment_size_in_kbytes = 64,
- .subvp_drr_scheduling_margin_us = 100,
- .subvp_prefetch_end_to_mall_start_us = 15,
- .subvp_fw_processing_delay = 15,
-
- .fams2 = {
- .max_allow_delay_us = 100 * 1000,
- .scheduling_delay_us = 50,
- .vertical_interrupt_ack_delay_us = 18,
- .allow_programming_delay_us = 18,
- .min_allow_width_us = 20,
- .subvp_df_throttle_delay_us = 100,
- .subvp_programming_delay_us = 18,
- .subvp_prefetch_to_mall_delay_us = 18,
- .drr_programming_delay_us = 18,
- },
-};
-
-#endif /* __DML_DML_DCN3_SOC_BB__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/bounding_boxes/dcn4_soc_bb.h b/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/bounding_boxes/dcn4_soc_bb.h
index 8ef7977841de..793e1c038efd 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/bounding_boxes/dcn4_soc_bb.h
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/bounding_boxes/dcn4_soc_bb.h
@@ -344,6 +344,7 @@ static const struct dml2_ip_capabilities dml2_dcn401_max_ip_caps = {
.config_return_buffer_segment_size_in_kbytes = 64,
.meta_fifo_size_in_kentries = 22,
.compressed_buffer_segment_size_in_kbytes = 64,
+ .cursor_buffer_size = 24,
.max_flip_time_us = 80,
.max_flip_time_lines = 32,
.hostvm_mode = 0,
@@ -354,7 +355,7 @@ static const struct dml2_ip_capabilities dml2_dcn401_max_ip_caps = {
.fams2 = {
.max_allow_delay_us = 100 * 1000,
- .scheduling_delay_us = 125,
+ .scheduling_delay_us = 550,
.vertical_interrupt_ack_delay_us = 40,
.allow_programming_delay_us = 18,
.min_allow_width_us = 20,
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top_display_cfg_types.h b/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top_display_cfg_types.h
index b132f676a68d..5e1ab6d97640 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top_display_cfg_types.h
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top_display_cfg_types.h
@@ -10,9 +10,10 @@
#define DML2_MAX_PLANES 8
#define DML2_MAX_DCN_PIPES 8
#define DML2_MAX_MCACHES 8 // assume plane is going to be supported by a max of 8 mcaches
+#define DML2_MAX_WRITEBACK 3
enum dml2_swizzle_mode {
- dml2_sw_linear,
+ dml2_sw_linear, // SW_LINEAR accepts 256 byte aligned pitch and also 128 byte aligned pitch if DCC is not enabled
dml2_sw_256b_2d,
dml2_sw_4kb_2d,
dml2_sw_64kb_2d,
@@ -24,7 +25,8 @@ enum dml2_swizzle_mode {
dml2_gfx11_sw_64kb_d_x,
dml2_gfx11_sw_64kb_r_x,
dml2_gfx11_sw_256kb_d_x,
- dml2_gfx11_sw_256kb_r_x
+ dml2_gfx11_sw_256kb_r_x,
+
};
enum dml2_source_format_class {
@@ -38,7 +40,13 @@ enum dml2_source_format_class {
dml2_rgbe_alpha = 9,
dml2_rgbe = 10,
dml2_mono_8 = 11,
- dml2_mono_16 = 12
+ dml2_mono_16 = 12,
+ dml2_422_planar_8 = 13,
+ dml2_422_planar_10 = 14,
+ dml2_422_planar_12 = 15,
+ dml2_422_packed_8 = 16,
+ dml2_422_packed_10 = 17,
+ dml2_422_packed_12 = 18
};
enum dml2_rotation_angle {
@@ -121,15 +129,6 @@ enum dml2_dsc_enable_option {
dml2_dsc_enable_if_necessary = 2
};
-enum dml2_pstate_support_method {
- dml2_pstate_method_uninitialized,
- dml2_pstate_method_not_supported,
- dml2_pstate_method_vactive,
- dml2_pstate_method_vblank,
- dml2_pstate_method_svp,
- dml2_pstate_method_drr
-};
-
enum dml2_tdlut_addressing_mode {
dml2_tdlut_sw_linear = 0,
dml2_tdlut_simple_linear = 1
@@ -287,22 +286,23 @@ struct dml2_link_output_cfg {
bool validate_output; // Do not validate the link configuration for this display stream.
};
-struct dml2_writeback_cfg {
- bool enable;
+struct dml2_writeback_info {
enum dml2_source_format_class pixel_format;
- unsigned int active_writebacks_per_surface;
+ unsigned long input_width;
+ unsigned long input_height;
+ unsigned long output_width;
+ unsigned long output_height;
+ unsigned long v_taps;
+ unsigned long h_taps;
+ unsigned long v_taps_chroma;
+ unsigned long h_taps_chroma;
+ double h_ratio;
+ double v_ratio;
+};
- struct {
- bool enabled;
- unsigned long input_width;
- unsigned long input_height;
- unsigned long output_width;
- unsigned long output_height;
- unsigned long v_taps;
- unsigned long h_taps;
- double h_ratio;
- double v_ratio;
- } scaling_info;
+struct dml2_writeback_cfg {
+ unsigned int active_writebacks_per_stream;
+ struct dml2_writeback_info writeback_stream[DML2_MAX_WRITEBACK];
};
struct dml2_plane_parameters {
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top_soc_parameter_types.h b/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top_soc_parameter_types.h
index ebd8abe894a9..5f0bc42d1d2f 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top_soc_parameter_types.h
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top_soc_parameter_types.h
@@ -167,11 +167,13 @@ struct dml2_ip_capabilities {
unsigned int max_num_dp2p0_streams;
unsigned int max_num_hdmi_frl_outputs;
unsigned int max_num_dp2p0_outputs;
+ unsigned int max_num_wb;
unsigned int rob_buffer_size_kbytes;
unsigned int config_return_buffer_size_in_kbytes;
unsigned int config_return_buffer_segment_size_in_kbytes;
unsigned int meta_fifo_size_in_kentries;
unsigned int compressed_buffer_segment_size_in_kbytes;
+ unsigned int cursor_buffer_size;
unsigned int max_flip_time_us;
unsigned int max_flip_time_lines;
unsigned int hostvm_mode;
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top_types.h b/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top_types.h
index eeb96c455658..d2d053f2354d 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top_types.h
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top_types.h
@@ -26,20 +26,14 @@ enum dml2_project_id {
dml2_project_dcn4x_stage2_auto_drr_svp = 3,
};
-enum dml2_dram_clock_change_support {
- dml2_dram_clock_change_vactive = 0,
- dml2_dram_clock_change_vblank = 1,
- dml2_dram_clock_change_vblank_and_vactive = 2,
- dml2_dram_clock_change_drr = 3,
- dml2_dram_clock_change_mall_svp = 4,
- dml2_dram_clock_change_mall_full_frame = 6,
- dml2_dram_clock_change_unsupported = 7
-};
-
-enum dml2_fclock_change_support {
- dml2_fclock_change_vactive = 0,
- dml2_fclock_change_vblank = 1,
- dml2_fclock_change_unsupported = 2
+enum dml2_pstate_change_support {
+ dml2_pstate_change_vactive = 0,
+ dml2_pstate_change_vblank = 1,
+ dml2_pstate_change_vblank_and_vactive = 2,
+ dml2_pstate_change_drr = 3,
+ dml2_pstate_change_mall_svp = 4,
+ dml2_pstate_change_mall_full_frame = 6,
+ dml2_pstate_change_unsupported = 7
};
enum dml2_output_type_and_rate__type {
@@ -202,24 +196,23 @@ struct dml2_mcache_surface_allocation {
} informative;
};
-enum dml2_uclk_pstate_support_method {
- dml2_uclk_pstate_support_method_not_supported = 0,
- /* hw */
- dml2_uclk_pstate_support_method_vactive = 1,
- dml2_uclk_pstate_support_method_vblank = 2,
- dml2_uclk_pstate_support_method_reserved_hw = 5,
- /* fw */
- dml2_uclk_pstate_support_method_fw_subvp_phantom = 6,
- dml2_uclk_pstate_support_method_reserved_fw = 10,
- /* fw w/drr */
- dml2_uclk_pstate_support_method_fw_vactive_drr = 11,
- dml2_uclk_pstate_support_method_fw_vblank_drr = 12,
- dml2_uclk_pstate_support_method_fw_subvp_phantom_drr = 13,
- dml2_uclk_pstate_support_method_reserved_fw_drr_fixed = 20,
- dml2_uclk_pstate_support_method_fw_drr = 21,
- dml2_uclk_pstate_support_method_reserved_fw_drr_var = 22,
-
- dml2_uclk_pstate_support_method_count
+enum dml2_pstate_method {
+ dml2_pstate_method_na = 0,
+ /* hw exclusive modes */
+ dml2_pstate_method_vactive = 1,
+ dml2_pstate_method_vblank = 2,
+ dml2_pstate_method_reserved_hw = 5,
+ /* fw assisted exclusive modes */
+ dml2_pstate_method_fw_svp = 6,
+ dml2_pstate_method_reserved_fw = 10,
+ /* fw assisted modes requiring drr modulation */
+ dml2_pstate_method_fw_vactive_drr = 11,
+ dml2_pstate_method_fw_vblank_drr = 12,
+ dml2_pstate_method_fw_svp_drr = 13,
+ dml2_pstate_method_reserved_fw_drr_clamped = 20,
+ dml2_pstate_method_fw_drr = 21,
+ dml2_pstate_method_reserved_fw_drr_var = 22,
+ dml2_pstate_method_count
};
struct dml2_per_plane_programming {
@@ -241,7 +234,7 @@ struct dml2_per_plane_programming {
// If a stream is using odm split, then this value is always 1
unsigned int num_dpps_required;
- enum dml2_uclk_pstate_support_method uclk_pstate_support_method;
+ enum dml2_pstate_method uclk_pstate_support_method;
// MALL size requirements for MALL SS and SubVP
unsigned int surface_size_mall_bytes;
@@ -281,7 +274,7 @@ struct dml2_per_stream_programming {
unsigned int num_odms_required;
- enum dml2_uclk_pstate_support_method uclk_pstate_method;
+ enum dml2_pstate_method uclk_pstate_method;
struct {
bool enabled;
@@ -289,7 +282,8 @@ struct dml2_per_stream_programming {
union dml2_global_sync_programming global_sync;
} phantom_stream;
- struct dmub_fams2_stream_static_state fams2_params;
+ union dmub_cmd_fams2_config fams2_base_params;
+ union dmub_cmd_fams2_config fams2_sub_params;
};
//-----------------
@@ -339,7 +333,7 @@ struct dml2_mode_support_info {
bool DCCMetaBufferSizeNotExceeded;
bool TotalVerticalActiveBandwidthSupport;
bool VActiveBandwidthSupport;
- enum dml2_fclock_change_support FCLKChangeSupport[DML2_MAX_PLANES];
+ enum dml2_pstate_change_support FCLKChangeSupport[DML2_MAX_PLANES];
bool USRRetrainingSupport;
bool PrefetchSupported;
bool DynamicMetadataSupported;
@@ -361,6 +355,7 @@ struct dml2_mode_support_info {
unsigned int AlignedYPitch[DML2_MAX_PLANES];
unsigned int AlignedCPitch[DML2_MAX_PLANES];
bool g6_temp_read_support;
+ bool temp_read_or_ppt_support;
}; // dml2_mode_support_info
struct dml2_display_cfg_programming {
@@ -392,6 +387,11 @@ struct dml2_display_cfg_programming {
unsigned long fclk_khz;
unsigned long dcfclk_khz;
} svp_prefetch;
+ struct {
+ unsigned long uclk_khz;
+ unsigned long fclk_khz;
+ unsigned long dcfclk_khz;
+ } svp_prefetch_no_throttle;
unsigned long deepsleep_dcfclk_khz;
unsigned long dispclk_khz;
@@ -444,7 +444,7 @@ struct dml2_display_cfg_programming {
double pstate_change_us;
double fclk_pstate_change_us;
double usr_retraining_us;
- double g6_temp_read_watermark_us;
+ double temp_read_or_ppt_watermark_us;
} watermarks;
struct {
@@ -653,6 +653,7 @@ struct dml2_display_cfg_programming {
double DisplayPipeLineDeliveryTimeLumaPrefetch[DML2_MAX_PLANES];
double DisplayPipeLineDeliveryTimeChromaPrefetch[DML2_MAX_PLANES];
+ double WritebackRequiredBandwidth;
double WritebackAllowDRAMClockChangeEndPosition[DML2_MAX_PLANES];
double WritebackAllowFCLKChangeEndPosition[DML2_MAX_PLANES];
double DSCCLK_calculated[DML2_MAX_PLANES];
@@ -662,6 +663,7 @@ struct dml2_display_cfg_programming {
double MaxActiveDRAMClockChangeLatencySupported[DML2_MAX_PLANES];
unsigned int PrefetchMode[DML2_MAX_PLANES]; // LEGACY_ONLY
bool ROBUrgencyAvoidance;
+ double LowestPrefetchMargin;
} misc;
struct dml2_mode_support_info mode_support_info;
@@ -675,6 +677,7 @@ struct dml2_display_cfg_programming {
bool failed_mcache_validation;
bool failed_dpmm;
bool failed_mode_programming;
+ bool failed_map_watermarks;
} informative;
};
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4.c
index 3d41ffde91c1..d68b4567e218 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4.c
@@ -9,7 +9,7 @@
#include "dml2_debug.h"
#include "lib_float_math.h"
-static const struct dml2_core_ip_params core_dcn4_ip_caps_base = {
+struct dml2_core_ip_params core_dcn4_ip_caps_base = {
// Hardcoded values for DCN3x
.vblank_nom_default_us = 668,
.remote_iommu_outstanding_translations = 256,
@@ -90,6 +90,7 @@ static void patch_ip_caps_with_explicit_ip_params(struct dml2_ip_capabilities *i
ip_caps->config_return_buffer_segment_size_in_kbytes = ip_params->config_return_buffer_segment_size_in_kbytes;
ip_caps->meta_fifo_size_in_kentries = ip_params->meta_fifo_size_in_kentries;
ip_caps->compressed_buffer_segment_size_in_kbytes = ip_params->compressed_buffer_segment_size_in_kbytes;
+ ip_caps->cursor_buffer_size = ip_params->cursor_buffer_size;
ip_caps->max_flip_time_us = ip_params->max_flip_time_us;
ip_caps->max_flip_time_lines = ip_params->max_flip_time_lines;
ip_caps->hostvm_mode = ip_params->hostvm_mode;
@@ -114,6 +115,7 @@ static void patch_ip_params_with_ip_caps(struct dml2_core_ip_params *ip_params,
ip_params->config_return_buffer_segment_size_in_kbytes = ip_caps->config_return_buffer_segment_size_in_kbytes;
ip_params->meta_fifo_size_in_kentries = ip_caps->meta_fifo_size_in_kentries;
ip_params->compressed_buffer_segment_size_in_kbytes = ip_caps->compressed_buffer_segment_size_in_kbytes;
+ ip_params->cursor_buffer_size = ip_caps->cursor_buffer_size;
ip_params->max_flip_time_us = ip_caps->max_flip_time_us;
ip_params->max_flip_time_lines = ip_caps->max_flip_time_lines;
ip_params->hostvm_mode = ip_caps->hostvm_mode;
@@ -316,28 +318,9 @@ static void pack_mode_programming_params_with_implicit_subvp(struct dml2_core_in
// Setup the appropriate p-state strategy
if (display_cfg->stage3.performed && display_cfg->stage3.success) {
- switch (display_cfg->stage3.pstate_switch_modes[plane_index]) {
- case dml2_uclk_pstate_support_method_vactive:
- case dml2_uclk_pstate_support_method_vblank:
- case dml2_uclk_pstate_support_method_fw_subvp_phantom:
- case dml2_uclk_pstate_support_method_fw_drr:
- case dml2_uclk_pstate_support_method_fw_vactive_drr:
- case dml2_uclk_pstate_support_method_fw_vblank_drr:
- case dml2_uclk_pstate_support_method_fw_subvp_phantom_drr:
- programming->plane_programming[plane_index].uclk_pstate_support_method = display_cfg->stage3.pstate_switch_modes[plane_index];
- break;
- case dml2_uclk_pstate_support_method_reserved_hw:
- case dml2_uclk_pstate_support_method_reserved_fw:
- case dml2_uclk_pstate_support_method_reserved_fw_drr_fixed:
- case dml2_uclk_pstate_support_method_reserved_fw_drr_var:
- case dml2_uclk_pstate_support_method_not_supported:
- case dml2_uclk_pstate_support_method_count:
- default:
- programming->plane_programming[plane_index].uclk_pstate_support_method = dml2_uclk_pstate_support_method_not_supported;
- break;
- }
+ programming->plane_programming[plane_index].uclk_pstate_support_method = display_cfg->stage3.pstate_switch_modes[plane_index];
} else {
- programming->plane_programming[plane_index].uclk_pstate_support_method = dml2_uclk_pstate_support_method_not_supported;
+ programming->plane_programming[plane_index].uclk_pstate_support_method = dml2_pstate_method_na;
}
dml2_core_calcs_get_mall_allocation(&core->clean_me_up.mode_lib, &programming->plane_programming[plane_index].surface_size_mall_bytes, dml_internal_pipe_index);
@@ -360,7 +343,8 @@ static void pack_mode_programming_params_with_implicit_subvp(struct dml2_core_in
/* unconditionally populate fams2 params */
dml2_core_calcs_get_stream_fams2_programming(&core->clean_me_up.mode_lib,
display_cfg,
- &programming->stream_programming[main_plane->stream_index].fams2_params,
+ &programming->stream_programming[main_plane->stream_index].fams2_base_params,
+ &programming->stream_programming[main_plane->stream_index].fams2_sub_params,
programming->stream_programming[main_plane->stream_index].uclk_pstate_method,
plane_index);
@@ -572,18 +556,18 @@ bool core_dcn4_mode_programming(struct dml2_core_mode_programming_in_out *in_out
in_out->programming->plane_programming[plane_index].num_dpps_required = core->clean_me_up.mode_lib.mp.NoOfDPP[plane_index];
if (in_out->programming->display_config.plane_descriptors[plane_index].overrides.legacy_svp_config == dml2_svp_mode_override_main_pipe)
- in_out->programming->plane_programming[plane_index].uclk_pstate_support_method = dml2_uclk_pstate_support_method_fw_subvp_phantom;
+ in_out->programming->plane_programming[plane_index].uclk_pstate_support_method = dml2_pstate_method_fw_svp;
else if (in_out->programming->display_config.plane_descriptors[plane_index].overrides.legacy_svp_config == dml2_svp_mode_override_phantom_pipe)
- in_out->programming->plane_programming[plane_index].uclk_pstate_support_method = dml2_uclk_pstate_support_method_fw_subvp_phantom;
+ in_out->programming->plane_programming[plane_index].uclk_pstate_support_method = dml2_pstate_method_fw_svp;
else if (in_out->programming->display_config.plane_descriptors[plane_index].overrides.legacy_svp_config == dml2_svp_mode_override_phantom_pipe_no_data_return)
- in_out->programming->plane_programming[plane_index].uclk_pstate_support_method = dml2_uclk_pstate_support_method_fw_subvp_phantom;
+ in_out->programming->plane_programming[plane_index].uclk_pstate_support_method = dml2_pstate_method_fw_svp;
else {
if (core->clean_me_up.mode_lib.mp.MaxActiveDRAMClockChangeLatencySupported[plane_index] >= core->clean_me_up.mode_lib.soc.power_management_parameters.dram_clk_change_blackout_us)
- in_out->programming->plane_programming[plane_index].uclk_pstate_support_method = dml2_uclk_pstate_support_method_vactive;
+ in_out->programming->plane_programming[plane_index].uclk_pstate_support_method = dml2_pstate_method_vactive;
else if (core->clean_me_up.mode_lib.mp.TWait[plane_index] >= core->clean_me_up.mode_lib.soc.power_management_parameters.dram_clk_change_blackout_us)
- in_out->programming->plane_programming[plane_index].uclk_pstate_support_method = dml2_uclk_pstate_support_method_vblank;
+ in_out->programming->plane_programming[plane_index].uclk_pstate_support_method = dml2_pstate_method_vblank;
else
- in_out->programming->plane_programming[plane_index].uclk_pstate_support_method = dml2_uclk_pstate_support_method_not_supported;
+ in_out->programming->plane_programming[plane_index].uclk_pstate_support_method = dml2_pstate_method_na;
}
dml2_core_calcs_get_mall_allocation(&core->clean_me_up.mode_lib, &in_out->programming->plane_programming[plane_index].surface_size_mall_bytes, dml_internal_pipe_index);
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c
index 601320b1be81..c4dbf27abaf8 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c
@@ -11,6 +11,9 @@
#define DML2_MAX_FMT_420_BUFFER_WIDTH 4096
#define DML_MAX_NUM_OF_SLICES_PER_DSC 4
+#define DML_MAX_COMPRESSION_RATIO 4
+//#define DML_MODE_SUPPORT_USE_DPM_DRAM_BW
+//#define DML_GLOBAL_PREFETCH_CHECK
#define ALLOW_SDPIF_RATE_LIMIT_PRE_CSTATE
const char *dml2_core_internal_bw_type_str(enum dml2_core_internal_bw_type bw_type)
@@ -132,9 +135,9 @@ static void dml2_print_mode_support_info(const struct dml2_core_internal_mode_su
dml2_printf("DML: support: DynamicMetadataSupported = %d\n", support->DynamicMetadataSupported);
if (!fail_only || support->VRatioInPrefetchSupported == 0)
dml2_printf("DML: support: VRatioInPrefetchSupported = %d\n", support->VRatioInPrefetchSupported);
- if (!fail_only || support->PTEBufferSizeNotExceeded == 1)
+ if (!fail_only || support->PTEBufferSizeNotExceeded == 0)
dml2_printf("DML: support: PTEBufferSizeNotExceeded = %d\n", support->PTEBufferSizeNotExceeded);
- if (!fail_only || support->DCCMetaBufferSizeNotExceeded == 1)
+ if (!fail_only || support->DCCMetaBufferSizeNotExceeded == 0)
dml2_printf("DML: support: DCCMetaBufferSizeNotExceeded = %d\n", support->DCCMetaBufferSizeNotExceeded);
if (!fail_only || support->ExceededMALLSize == 1)
dml2_printf("DML: support: ExceededMALLSize = %d\n", support->ExceededMALLSize);
@@ -315,12 +318,11 @@ dml_get_var_func(meta_trip_memory_us, double, mode_lib->mp.MetaTripToMemory);
dml_get_var_func(wm_fclk_change, double, mode_lib->mp.Watermark.FCLKChangeWatermark);
dml_get_var_func(wm_usr_retraining, double, mode_lib->mp.Watermark.USRRetrainingWatermark);
-dml_get_var_func(wm_g6_temp_read, double, mode_lib->mp.Watermark.g6_temp_read_watermark_us);
+dml_get_var_func(wm_temp_read_or_ppt, double, mode_lib->mp.Watermark.temp_read_or_ppt_watermark_us);
dml_get_var_func(wm_dram_clock_change, double, mode_lib->mp.Watermark.DRAMClockChangeWatermark);
dml_get_var_func(fraction_of_urgent_bandwidth, double, mode_lib->mp.FractionOfUrgentBandwidth);
dml_get_var_func(fraction_of_urgent_bandwidth_imm_flip, double, mode_lib->mp.FractionOfUrgentBandwidthImmediateFlip);
dml_get_var_func(fraction_of_urgent_bandwidth_mall, double, mode_lib->mp.FractionOfUrgentBandwidthMALL);
-dml_get_var_func(urgent_latency, double, mode_lib->mp.UrgentLatency);
dml_get_var_func(wm_writeback_dram_clock_change, double, mode_lib->mp.Watermark.WritebackDRAMClockChangeWatermark);
dml_get_var_func(wm_writeback_fclk_change, double, mode_lib->mp.Watermark.WritebackFCLKChangeWatermark);
dml_get_var_func(stutter_efficiency, double, mode_lib->mp.StutterEfficiency);
@@ -355,7 +357,9 @@ dml_get_var_func(svp_prefetch_urg_bw_available_sdp, double, mode_lib->mp.urg_ban
dml_get_var_func(svp_prefetch_urg_bw_available_dram, double, mode_lib->mp.urg_bandwidth_available[dml2_core_internal_soc_state_svp_prefetch][dml2_core_internal_bw_dram]);
dml_get_var_func(svp_prefetch_urg_bw_available_dram_vm_only, double, mode_lib->mp.urg_bandwidth_available_vm_only[dml2_core_internal_soc_state_svp_prefetch]);
+dml_get_var_func(urgent_latency, double, mode_lib->mp.UrgentLatency);
dml_get_var_func(max_urgent_latency_us, double, mode_lib->ms.support.max_urgent_latency_us);
+dml_get_var_func(max_non_urgent_latency_us, double, mode_lib->ms.support.max_non_urgent_latency_us);
dml_get_var_func(avg_non_urgent_latency_us, double, mode_lib->ms.support.avg_non_urgent_latency_us);
dml_get_var_func(avg_urgent_latency_us, double, mode_lib->ms.support.avg_urgent_latency_us);
@@ -466,6 +470,24 @@ static bool dml_is_420(enum dml2_source_format_class source_format)
case dml2_420_12:
val = 1;
break;
+ case dml2_422_planar_8:
+ val = 0;
+ break;
+ case dml2_422_planar_10:
+ val = 0;
+ break;
+ case dml2_422_planar_12:
+ val = 0;
+ break;
+ case dml2_422_packed_8:
+ val = 0;
+ break;
+ case dml2_422_packed_10:
+ val = 0;
+ break;
+ case dml2_422_packed_12:
+ val = 0;
+ break;
case dml2_rgbe_alpha:
val = 0;
break;
@@ -487,32 +509,31 @@ static bool dml_is_420(enum dml2_source_format_class source_format)
static unsigned int dml_get_tile_block_size_bytes(enum dml2_swizzle_mode sw_mode)
{
- switch (sw_mode) {
- case (dml2_sw_linear):
- return 256; break;
- case (dml2_sw_256b_2d):
- return 256; break;
- case (dml2_sw_4kb_2d):
- return 4096; break;
- case (dml2_sw_64kb_2d):
- return 65536; break;
- case (dml2_sw_256kb_2d):
- return 262144; break;
- case (dml2_gfx11_sw_linear):
- return 256; break;
- case (dml2_gfx11_sw_64kb_d):
- return 65536; break;
- case (dml2_gfx11_sw_64kb_d_t):
- return 65536; break;
- case (dml2_gfx11_sw_64kb_d_x):
- return 65536; break;
- case (dml2_gfx11_sw_64kb_r_x):
- return 65536; break;
- case (dml2_gfx11_sw_256kb_d_x):
- return 262144; break;
- case (dml2_gfx11_sw_256kb_r_x):
- return 262144; break;
- default:
+ if (sw_mode == dml2_sw_linear)
+ return 256;
+ else if (sw_mode == dml2_sw_256b_2d)
+ return 256;
+ else if (sw_mode == dml2_sw_4kb_2d)
+ return 4096;
+ else if (sw_mode == dml2_sw_64kb_2d)
+ return 65536;
+ else if (sw_mode == dml2_sw_256kb_2d)
+ return 262144;
+ else if (sw_mode == dml2_gfx11_sw_linear)
+ return 256;
+ else if (sw_mode == dml2_gfx11_sw_64kb_d)
+ return 65536;
+ else if (sw_mode == dml2_gfx11_sw_64kb_d_t)
+ return 65536;
+ else if (sw_mode == dml2_gfx11_sw_64kb_d_x)
+ return 65536;
+ else if (sw_mode == dml2_gfx11_sw_64kb_r_x)
+ return 65536;
+ else if (sw_mode == dml2_gfx11_sw_256kb_d_x)
+ return 262144;
+ else if (sw_mode == dml2_gfx11_sw_256kb_r_x)
+ return 262144;
+ else {
DML2_ASSERT(0);
return 256;
}
@@ -579,8 +600,8 @@ static void CalculateBytePerPixelAndBlockSizes(
{
*BytePerPixelDETY = 0;
*BytePerPixelDETC = 0;
- *BytePerPixelY = 0;
- *BytePerPixelC = 0;
+ *BytePerPixelY = 1;
+ *BytePerPixelC = 1;
if (SourcePixelFormat == dml2_444_64) {
*BytePerPixelDETY = 8;
@@ -820,7 +841,7 @@ static void CalculateSwathWidth(
// Output
unsigned int req_per_swath_ub_l[],
unsigned int req_per_swath_ub_c[],
- unsigned int SwathWidthSingleDPPY[],
+ unsigned int SwathWidthSingleDPPY[], // post-rotated plane width
unsigned int SwathWidthSingleDPPC[],
unsigned int SwathWidthY[], // per-pipe
unsigned int SwathWidthC[], // per-pipe
@@ -1403,7 +1424,6 @@ static unsigned int dscceComputeDelay(
// N422/N420 operate at 2 pixels per clock
unsigned int pixelsPerClock, padding_pixels, ssm_group_priming_delay, ssm_pipeline_delay, obsm_pipeline_delay, slice_padded_pixels, ixd_plus_padding, ixd_plus_padding_groups, cycles_per_group, group_delay, pipeline_delay, pixels, additional_group_delay, lines_to_reach_ixd, groups_to_reach_ixd, slice_width_groups, initial_xmit_delay, number_of_lines_to_reach_ixd, slice_width_modified;
-
if (pixelFormat == dml2_420)
pixelsPerClock = 2;
// #all other modes operate at 1 pixel per clock
@@ -1428,7 +1448,6 @@ static unsigned int dscceComputeDelay(
}
}
-
//sub-stream multiplexer balance fifo priming delay in groups as per dsc standard
if (bpc == 8)
ssm_group_priming_delay = 83;
@@ -1447,9 +1466,6 @@ static unsigned int dscceComputeDelay(
//determine number of padded pixels in the last group of a slice line, computed as
slice_padded_pixels = 3 * slice_width_groups - slice_width_modified;
-
-
-
//determine integer number of complete slice lines required to reach initial transmit delay without ssm delay considered
number_of_lines_to_reach_ixd = initial_xmit_delay / slice_width_modified;
@@ -1463,7 +1479,6 @@ static unsigned int dscceComputeDelay(
//number of groups required for a slice to reach initial transmit delay is the sum of the padded initial transmit delay plus the ssm group priming delay
groups_to_reach_ixd = ixd_plus_padding_groups + ssm_group_priming_delay;
-
//number of lines required to reach padded initial transmit delay in groups in slices to the left of the last horizontal slice
//needs to be rounded up as a complete slice lines are buffered prior to initial transmit delay being reached in the last horizontal slice
lines_to_reach_ixd = (groups_to_reach_ixd + slice_width_groups - 1) / slice_width_groups; //round up lines to reach ixd to next
@@ -1506,7 +1521,6 @@ static unsigned int dscceComputeDelay(
return pixels;
}
-
//updated in dcn4
static unsigned int dscComputeDelay(enum dml2_output_format_class pixelFormat, enum dml2_output_encoder_class Output)
{
@@ -2090,7 +2104,6 @@ static void CalculateDCCConfiguration(
yuv420 = 1;
else
yuv420 = 0;
-
horz_div_l = 1;
horz_div_c = 1;
vert_div_l = 1;
@@ -2561,8 +2574,7 @@ static void calculate_mcache_setting(
if (*p->num_mcaches_l) {
l->avg_mcache_element_size_l = l->meta_row_width_l / *p->num_mcaches_l;
}
-
- if (l->is_dual_plane && *p->num_mcaches_c) {
+ if (l->is_dual_plane) {
l->avg_mcache_element_size_c = l->meta_row_width_c / *p->num_mcaches_c;
if (!p->imall_enable || (*p->mall_comb_mcache_l == *p->mall_comb_mcache_c)) {
@@ -2682,12 +2694,12 @@ static double dml_get_return_bandwidth_available(
bool is_avg_bw,
bool is_hvm_en,
bool is_hvm_only,
- double dcflk_mhz,
+ double dcfclk_mhz,
double fclk_mhz,
double dram_bw_mbps)
{
double return_bw_mbps = 0.;
- double ideal_sdp_bandwidth = (double)soc->return_bus_width_bytes * dcflk_mhz;
+ double ideal_sdp_bandwidth = (double)soc->return_bus_width_bytes * dcfclk_mhz;
double ideal_fabric_bandwidth = fclk_mhz * (double)soc->fabric_datapath_to_dcn_data_return_bytes;
double ideal_dram_bandwidth = dram_bw_mbps; //dram_speed_mts * soc->clk_table.dram_config.channel_count * soc->clk_table.dram_config.channel_width_bytes;
@@ -2753,7 +2765,7 @@ static double dml_get_return_bandwidth_available(
dml2_printf("DML::%s: is_hvm_only = %u\n", __func__, is_hvm_only);
dml2_printf("DML::%s: state_type = %s\n", __func__, dml2_core_internal_soc_state_type_str(state_type));
dml2_printf("DML::%s: bw_type = %s\n", __func__, dml2_core_internal_bw_type_str(bw_type));
- dml2_printf("DML::%s: dcflk_mhz = %f\n", __func__, dcflk_mhz);
+ dml2_printf("DML::%s: dcfclk_mhz = %f\n", __func__, dcfclk_mhz);
dml2_printf("DML::%s: fclk_mhz = %f\n", __func__, fclk_mhz);
dml2_printf("DML::%s: ideal_sdp_bandwidth = %f\n", __func__, ideal_sdp_bandwidth);
dml2_printf("DML::%s: ideal_fabric_bandwidth = %f\n", __func__, ideal_fabric_bandwidth);
@@ -3516,10 +3528,9 @@ static void CalculateUrgentBurstFactor(
dml2_printf("DML::%s: UrgentBurstFactorChroma = %f\n", __func__, *UrgentBurstFactorChroma);
dml2_printf("DML::%s: NotEnoughUrgentLatencyHiding = %d\n", __func__, *NotEnoughUrgentLatencyHiding);
#endif
-
}
-static void CalculateDCFCLKDeepSleep(
+static void CalculateDCFCLKDeepSleepTdlut(
const struct dml2_display_cfg *display_cfg,
unsigned int NumberOfActiveSurfaces,
unsigned int BytePerPixelY[],
@@ -3534,6 +3545,10 @@ static void CalculateDCFCLKDeepSleep(
double ReadBandwidthChroma[],
unsigned int ReturnBusWidth,
+ double dispclk,
+ unsigned int tdlut_bytes_to_deliver[],
+ double prefetch_swath_time_us[],
+
// Output
double *DCFClkDeepSleep)
{
@@ -3568,6 +3583,22 @@ static void CalculateDCFCLKDeepSleep(
}
DCFClkDeepSleepPerSurface[k] = math_max2(DCFClkDeepSleepPerSurface[k], pixel_rate_mhz / 16);
+ // adjust for 3dlut delivery time
+ if (display_cfg->plane_descriptors[k].tdlut.setup_for_tdlut && tdlut_bytes_to_deliver[k] > 0) {
+ double tdlut_required_deepsleep_dcfclk = (double) tdlut_bytes_to_deliver[k] / 64.0 / prefetch_swath_time_us[k];
+
+ dml2_printf("DML::%s: k=%d, DCFClkDeepSleepPerSurface = %f\n", __func__, k, DCFClkDeepSleepPerSurface[k]);
+ dml2_printf("DML::%s: k=%d, tdlut_bytes_to_deliver = %d\n", __func__, k, tdlut_bytes_to_deliver[k]);
+ dml2_printf("DML::%s: k=%d, prefetch_swath_time_us = %f\n", __func__, k, prefetch_swath_time_us[k]);
+ dml2_printf("DML::%s: k=%d, tdlut_required_deepsleep_dcfclk = %f\n", __func__, k, tdlut_required_deepsleep_dcfclk);
+
+ // increase the deepsleep dcfclk to match the original dispclk throughput rate
+ if (tdlut_required_deepsleep_dcfclk > DCFClkDeepSleepPerSurface[k]) {
+ DCFClkDeepSleepPerSurface[k] = math_max2(DCFClkDeepSleepPerSurface[k], tdlut_required_deepsleep_dcfclk);
+ DCFClkDeepSleepPerSurface[k] = math_max2(DCFClkDeepSleepPerSurface[k], dispclk / 4.0);
+ }
+ }
+
#ifdef __DML_VBA_DEBUG__
dml2_printf("DML::%s: k=%u, PixelClock = %f\n", __func__, k, pixel_rate_mhz);
dml2_printf("DML::%s: k=%u, DCFClkDeepSleepPerSurface = %f\n", __func__, k, DCFClkDeepSleepPerSurface[k]);
@@ -3590,9 +3621,56 @@ static void CalculateDCFCLKDeepSleep(
for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) {
*DCFClkDeepSleep = math_max2(*DCFClkDeepSleep, DCFClkDeepSleepPerSurface[k]);
}
+
dml2_printf("DML::%s: DCFClkDeepSleep = %f (final)\n", __func__, *DCFClkDeepSleep);
}
+static void CalculateDCFCLKDeepSleep(
+ const struct dml2_display_cfg *display_cfg,
+ unsigned int NumberOfActiveSurfaces,
+ unsigned int BytePerPixelY[],
+ unsigned int BytePerPixelC[],
+ unsigned int SwathWidthY[],
+ unsigned int SwathWidthC[],
+ unsigned int DPPPerSurface[],
+ double PSCL_THROUGHPUT[],
+ double PSCL_THROUGHPUT_CHROMA[],
+ double Dppclk[],
+ double ReadBandwidthLuma[],
+ double ReadBandwidthChroma[],
+ unsigned int ReturnBusWidth,
+
+ // Output
+ double *DCFClkDeepSleep)
+{
+ double zero_double[DML2_MAX_PLANES];
+ unsigned int zero_integer[DML2_MAX_PLANES];
+
+ memset(zero_double, 0, DML2_MAX_PLANES * sizeof(double));
+ memset(zero_integer, 0, DML2_MAX_PLANES * sizeof(unsigned int));
+
+ CalculateDCFCLKDeepSleepTdlut(
+ display_cfg,
+ NumberOfActiveSurfaces,
+ BytePerPixelY,
+ BytePerPixelC,
+ SwathWidthY,
+ SwathWidthC,
+ DPPPerSurface,
+ PSCL_THROUGHPUT,
+ PSCL_THROUGHPUT_CHROMA,
+ Dppclk,
+ ReadBandwidthLuma,
+ ReadBandwidthChroma,
+ ReturnBusWidth,
+ 0,
+ zero_integer, //tdlut_bytes_to_deliver,
+ zero_double, //prefetch_swath_time_us,
+
+ // Output
+ DCFClkDeepSleep);
+}
+
static double CalculateWriteBackDelay(
enum dml2_source_format_class WritebackPixelFormat,
double WritebackHRatio,
@@ -3816,8 +3894,8 @@ static void CalculateSwathAndDETConfiguration(struct dml2_core_internal_scratch
p->SwathHeightC[k] = MaximumSwathHeightC[k] / 2;
RoundedUpSwathSizeBytesY[k] = p->full_swath_bytes_l[k] / 2;
RoundedUpSwathSizeBytesC[k] = p->full_swath_bytes_c[k] / 2;
- p->request_size_bytes_luma[k] = ((p->BytePerPixY[k] == 2) == dml_is_vertical_rotation(p->display_cfg->plane_descriptors[k].composition.rotation_angle)) ? 128 : 64;
- p->request_size_bytes_chroma[k] = ((p->BytePerPixC[k] == 2) == dml_is_vertical_rotation(p->display_cfg->plane_descriptors[k].composition.rotation_angle)) ? 128 : 64;
+ p->request_size_bytes_luma[k] = ((p->BytePerPixY[k] == 2) == dml_is_vertical_rotation(p->display_cfg->plane_descriptors[k].composition.rotation_angle)) ? 128 : 64;;
+ p->request_size_bytes_chroma[k] = ((p->BytePerPixC[k] == 2) == dml_is_vertical_rotation(p->display_cfg->plane_descriptors[k].composition.rotation_angle)) ? 128 : 64;;
}
if (p->SwathHeightC[k] == 0)
@@ -4592,6 +4670,7 @@ static void calculate_tdlut_setting(
*p->tdlut_groups_per_2row_ub = 0;
*p->tdlut_opt_time = 0;
*p->tdlut_drain_time = 0;
+ *p->tdlut_bytes_to_deliver = 0;
*p->tdlut_bytes_per_group = 0;
*p->tdlut_pte_bytes_per_frame = 0;
*p->tdlut_bytes_per_frame = 0;
@@ -4660,6 +4739,7 @@ static void calculate_tdlut_setting(
*p->tdlut_groups_per_2row_ub = (unsigned int)math_ceil2((double) *p->tdlut_bytes_per_frame / *p->tdlut_bytes_per_group, 1);
*p->tdlut_opt_time = (*p->tdlut_bytes_per_frame - p->cursor_buffer_size * 1024) / tdlut_drain_rate;
*p->tdlut_drain_time = p->cursor_buffer_size * 1024 / tdlut_drain_rate;
+ *p->tdlut_bytes_to_deliver = (unsigned int) (p->cursor_buffer_size * 1024.0);
}
#ifdef __DML_VBA_DEBUG__
@@ -4680,6 +4760,7 @@ static void calculate_tdlut_setting(
dml2_printf("DML::%s: tdlut_delivery_cycles = %u\n", __func__, tdlut_delivery_cycles);
dml2_printf("DML::%s: tdlut_opt_time = %f\n", __func__, *p->tdlut_opt_time);
dml2_printf("DML::%s: tdlut_drain_time = %f\n", __func__, *p->tdlut_drain_time);
+ dml2_printf("DML::%s: tdlut_bytes_to_deliver = %d\n", __func__, *p->tdlut_bytes_to_deliver);
dml2_printf("DML::%s: tdlut_groups_per_2row_ub = %d\n", __func__, *p->tdlut_groups_per_2row_ub);
#endif
}
@@ -5069,20 +5150,18 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch
s->trip_to_mem = 0.0;
*p->Tvm_trips = 0.0;
*p->Tr0_trips = 0.0;
- s->Tvm_no_trip_oto = 0.0;
- s->Tr0_no_trip_oto = 0.0;
s->Tvm_trips_rounded = 0.0;
s->Tr0_trips_rounded = 0.0;
s->max_Tsw = 0.0;
s->Lsw_oto = 0.0;
- s->Tpre_rounded = 0.0;
+ *p->Tpre_rounded = 0.0;
s->prefetch_bw_equ = 0.0;
s->Tvm_equ = 0.0;
s->Tr0_equ = 0.0;
s->Tdmbf = 0.0;
s->Tdmec = 0.0;
s->Tdmsks = 0.0;
- s->prefetch_sw_bytes = 0.0;
+ *p->prefetch_sw_bytes = 0.0;
s->prefetch_bw_pr = 0.0;
s->bytes_pp = 0.0;
s->dep_bytes = 0.0;
@@ -5207,6 +5286,7 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch
dml2_printf("DML::%s: setup_for_tdlut = %u\n", __func__, p->setup_for_tdlut);
dml2_printf("DML::%s: tdlut_opt_time = %f\n", __func__, p->tdlut_opt_time);
dml2_printf("DML::%s: tdlut_pte_bytes_per_frame = %u\n", __func__, p->tdlut_pte_bytes_per_frame);
+ dml2_printf("DML::%s: tdlut_drain_time = %f\n", __func__, p->tdlut_drain_time);
#endif
if (p->OutputFormat == dml2_420 || (p->myPipe->InterlaceEnable && p->myPipe->ProgressiveToInterlaceUnitInOPP))
@@ -5277,23 +5357,8 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch
s->bytes_pp = p->myPipe->BytePerPixelY + p->myPipe->BytePerPixelC;
}
- s->prefetch_bw_pr = s->bytes_pp * p->myPipe->PixelClock / (double)p->myPipe->DPPPerSurface;
- if (p->myPipe->VRatio < 1.0)
- s->prefetch_bw_pr = p->myPipe->VRatio * s->prefetch_bw_pr;
- s->max_Tsw = (math_max2(p->PrefetchSourceLinesY, p->PrefetchSourceLinesC) * s->LineTime);
-
- s->prefetch_sw_bytes = p->PrefetchSourceLinesY * p->swath_width_luma_ub * p->myPipe->BytePerPixelY + p->PrefetchSourceLinesC * p->swath_width_chroma_ub * p->myPipe->BytePerPixelC;
- s->prefetch_bw_pr = s->prefetch_bw_pr * p->mall_prefetch_sdp_overhead_factor;
- s->prefetch_sw_bytes = s->prefetch_sw_bytes * p->mall_prefetch_sdp_overhead_factor;
- s->prefetch_bw_oto = math_max2(s->prefetch_bw_pr, s->prefetch_sw_bytes / s->max_Tsw);
-
- s->min_Lsw_oto = math_max2(p->PrefetchSourceLinesY, p->PrefetchSourceLinesC) / __DML2_CALCS_MAX_VRATIO_PRE_OTO__;
- s->min_Lsw_oto = math_max2(s->min_Lsw_oto, 2.0);
- s->min_Lsw_oto = math_max2(s->min_Lsw_oto, p->tdlut_drain_time / s->LineTime);
-
- s->min_Lsw_equ = math_max2(p->PrefetchSourceLinesY, p->PrefetchSourceLinesC) / __DML2_CALCS_MAX_VRATIO_PRE_EQU__;
- s->min_Lsw_equ = math_max2(s->min_Lsw_equ, 2.0);
- s->min_Lsw_equ = math_max2(s->min_Lsw_equ, p->tdlut_drain_time / s->LineTime);
+ *p->prefetch_sw_bytes = p->PrefetchSourceLinesY * p->swath_width_luma_ub * p->myPipe->BytePerPixelY + p->PrefetchSourceLinesC * p->swath_width_chroma_ub * p->myPipe->BytePerPixelC;
+ *p->prefetch_sw_bytes = *p->prefetch_sw_bytes * p->mall_prefetch_sdp_overhead_factor;
vm_bytes = p->vm_bytes; // vm_bytes is dpde0_bytes_per_frame_ub_l + dpde0_bytes_per_frame_ub_c + 2*extra_dpde_bytes;
extra_tdpe_bytes = (unsigned int)math_max2(0, (p->display_cfg->gpuvm_max_page_table_levels - 1) * 128);
@@ -5302,57 +5367,103 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch
vm_bytes = vm_bytes + p->tdlut_pte_bytes_per_frame + (p->display_cfg->gpuvm_enable ? extra_tdpe_bytes : 0);
tdlut_row_bytes = (unsigned long) math_ceil2(p->tdlut_bytes_per_frame/2.0, 1.0);
+
+ s->min_Lsw_oto = math_max2(p->PrefetchSourceLinesY, p->PrefetchSourceLinesC) / __DML2_CALCS_MAX_VRATIO_PRE_OTO__;
+ s->min_Lsw_oto = math_max2(s->min_Lsw_oto, p->tdlut_drain_time / s->LineTime);
+ s->min_Lsw_oto = math_max2(s->min_Lsw_oto, 2.0);
+
+ // use vactive swath bw for prefetch oto and also cap prefetch_bw_oto to max_vratio_oto
+ // Note: in prefetch calculation, acounting is done mostly per-pipe.
+ // vactive swath bw represents the per-surface (aka per dml plane) bw to move vratio_l/c lines of bytes_l/c per line time
+ s->per_pipe_vactive_sw_bw = p->vactive_sw_bw_l / (double)p->myPipe->DPPPerSurface;
+
+ // one-to-one prefetch bw as one line of bytes per line time (as per vratio_pre_l/c = 1)
+ s->prefetch_bw_oto = (p->swath_width_luma_ub * p->myPipe->BytePerPixelY) / s->LineTime;
+
+ if (p->myPipe->BytePerPixelC > 0) {
+ s->per_pipe_vactive_sw_bw += p->vactive_sw_bw_c / (double)p->myPipe->DPPPerSurface;
+ s->prefetch_bw_oto += (p->swath_width_chroma_ub * p->myPipe->BytePerPixelC) / s->LineTime;
+ }
+
+ s->prefetch_bw_oto = math_max2(s->per_pipe_vactive_sw_bw, s->prefetch_bw_oto) * p->mall_prefetch_sdp_overhead_factor;
+
+ s->prefetch_bw_oto = math_min2(s->prefetch_bw_oto, *p->prefetch_sw_bytes/(s->min_Lsw_oto*s->LineTime));
+
+ s->Lsw_oto = math_ceil2(4.0 * *p->prefetch_sw_bytes / s->prefetch_bw_oto / s->LineTime, 1.0) / 4.0;
+
s->prefetch_bw_oto = math_max3(s->prefetch_bw_oto,
p->vm_bytes * p->HostVMInefficiencyFactor / (31 * s->LineTime) - *p->Tno_bw,
(p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes) / (15 * s->LineTime));
- s->Lsw_oto = math_ceil2(4.0 * math_max2(s->prefetch_sw_bytes / s->prefetch_bw_oto / s->LineTime, s->min_Lsw_oto), 1.0) / 4.0;
+
+#ifdef __DML_VBA_DEBUG__
+ dml2_printf("DML::%s: vactive_sw_bw_l = %f\n", __func__, p->vactive_sw_bw_l);
+ dml2_printf("DML::%s: vactive_sw_bw_c = %f\n", __func__, p->vactive_sw_bw_c);
+ dml2_printf("DML::%s: per_pipe_vactive_sw_bw = %f\n", __func__, s->per_pipe_vactive_sw_bw);
+#endif
if (p->display_cfg->gpuvm_enable == true) {
- s->Tvm_no_trip_oto = math_max2(
+ s->Tvm_oto = math_max3(
+ *p->Tvm_trips,
*p->Tno_bw + vm_bytes * p->HostVMInefficiencyFactor / s->prefetch_bw_oto,
s->LineTime / 4.0);
- s->Tvm_oto = math_max2(
- *p->Tvm_trips,
- s->Tvm_no_trip_oto);
+
#ifdef __DML_VBA_DEBUG__
dml2_printf("DML::%s: Tvm_oto max0 = %f\n", __func__, *p->Tvm_trips);
dml2_printf("DML::%s: Tvm_oto max1 = %f\n", __func__, *p->Tno_bw + vm_bytes * p->HostVMInefficiencyFactor / s->prefetch_bw_oto);
dml2_printf("DML::%s: Tvm_oto max2 = %f\n", __func__, s->LineTime / 4.0);
#endif
} else {
- s->Tvm_no_trip_oto = s->Tvm_trips_rounded;
s->Tvm_oto = s->Tvm_trips_rounded;
}
if ((p->display_cfg->gpuvm_enable == true || p->setup_for_tdlut || dcc_mrq_enable)) {
- s->Tr0_no_trip_oto = math_max2(
+ s->Tr0_oto = math_max3(
+ *p->Tr0_trips,
(p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes) / s->prefetch_bw_oto,
s->LineTime / 4.0);
- s->Tr0_oto = math_max2(
- *p->Tr0_trips,
- s->Tr0_no_trip_oto);
#ifdef __DML_VBA_DEBUG__
dml2_printf("DML::%s: Tr0_oto max0 = %f\n", __func__, *p->Tr0_trips);
dml2_printf("DML::%s: Tr0_oto max1 = %f\n", __func__, (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes) / s->prefetch_bw_oto);
dml2_printf("DML::%s: Tr0_oto max2 = %f\n", __func__, s->LineTime / 4);
#endif
- } else {
- s->Tr0_no_trip_oto = (s->LineTime - s->Tvm_oto) / 4.0;
- s->Tr0_oto = s->Tr0_no_trip_oto;
- }
+ } else
+ s->Tr0_oto = s->LineTime / 4.0;
s->Tvm_oto_lines = math_ceil2(4.0 * s->Tvm_oto / s->LineTime, 1) / 4.0;
s->Tr0_oto_lines = math_ceil2(4.0 * s->Tr0_oto / s->LineTime, 1) / 4.0;
s->dst_y_prefetch_oto = s->Tvm_oto_lines + 2 * s->Tr0_oto_lines + s->Lsw_oto;
+#ifdef DML_GLOBAL_PREFETCH_CHECK
+ dml2_printf("DML::%s: impacted_Tpre = %f\n", __func__, p->impacted_dst_y_pre);
+ if (p->impacted_dst_y_pre > 0) {
+ dml2_printf("DML::%s: dst_y_prefetch_oto = %f\n", __func__, s->dst_y_prefetch_oto);
+ s->dst_y_prefetch_oto = math_max2(s->dst_y_prefetch_oto, p->impacted_dst_y_pre);
+ dml2_printf("DML::%s: dst_y_prefetch_oto = %f (impacted)\n", __func__, s->dst_y_prefetch_oto);
+ }
+#endif
+ *p->Tpre_oto = s->dst_y_prefetch_oto * s->LineTime;
+
//To (time for delay after scaler) in line time
Lo = (unsigned int)(*p->DSTYAfterScaler + (double)*p->DSTXAfterScaler / (double)p->myPipe->HTotal);
+ s->min_Lsw_equ = math_max2(p->PrefetchSourceLinesY, p->PrefetchSourceLinesC) / __DML2_CALCS_MAX_VRATIO_PRE_EQU__;
+ s->min_Lsw_equ = math_max2(s->min_Lsw_equ, p->tdlut_drain_time / s->LineTime);
+ s->min_Lsw_equ = math_max2(s->min_Lsw_equ, 2.0);
//Tpre_equ in line time
if (p->DynamicMetadataVMEnabled && p->DynamicMetadataEnable)
s->dst_y_prefetch_equ = p->VStartup - (*p->TSetup + math_max2(p->TCalc, *p->Tvm_trips) + s->TWait_p) / s->LineTime - Lo;
else
s->dst_y_prefetch_equ = p->VStartup - (*p->TSetup + math_max2(p->TCalc, p->ExtraLatencyPrefetch) + s->TWait_p) / s->LineTime - Lo;
+
+#ifdef DML_GLOBAL_PREFETCH_CHECK
+ s->dst_y_prefetch_equ_impacted = math_max2(p->impacted_dst_y_pre, s->dst_y_prefetch_equ);
+
+ s->dst_y_prefetch_equ_impacted = math_min2(s->dst_y_prefetch_equ_impacted, 63.75); // limit to the reg limit of U6.2 for DST_Y_PREFETCH
+
+ if (s->dst_y_prefetch_equ_impacted > s->dst_y_prefetch_equ)
+ s->dst_y_prefetch_equ -= s->dst_y_prefetch_equ_impacted - s->dst_y_prefetch_equ;
+#endif
+
s->dst_y_prefetch_equ = math_min2(s->dst_y_prefetch_equ, 63.75); // limit to the reg limit of U6.2 for DST_Y_PREFETCH
#ifdef __DML_VBA_DEBUG__
@@ -5370,7 +5481,7 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch
dml2_printf("DML::%s: BytePerPixelC = %u\n", __func__, p->myPipe->BytePerPixelC);
dml2_printf("DML::%s: PrefetchSourceLinesC = %f\n", __func__, p->PrefetchSourceLinesC);
dml2_printf("DML::%s: swath_width_chroma_ub = %u\n", __func__, p->swath_width_chroma_ub);
- dml2_printf("DML::%s: prefetch_sw_bytes = %f\n", __func__, s->prefetch_sw_bytes);
+ dml2_printf("DML::%s: prefetch_sw_bytes = %f\n", __func__, *p->prefetch_sw_bytes);
dml2_printf("DML::%s: max_Tsw = %f\n", __func__, s->max_Tsw);
dml2_printf("DML::%s: bytes_pp = %f\n", __func__, s->bytes_pp);
dml2_printf("DML::%s: vm_bytes = %u\n", __func__, vm_bytes);
@@ -5394,7 +5505,7 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch
#endif
double Tpre = s->dst_y_prefetch_equ * s->LineTime;
s->dst_y_prefetch_equ = math_floor2(4.0 * (s->dst_y_prefetch_equ + 0.125), 1) / 4.0;
- s->Tpre_rounded = s->dst_y_prefetch_equ * s->LineTime;
+ *p->Tpre_rounded = s->dst_y_prefetch_equ * s->LineTime;
#ifdef __DML_VBA_DEBUG__
dml2_printf("DML::%s: dst_y_prefetch_equ: %f (after round)\n", __func__, s->dst_y_prefetch_equ);
@@ -5420,7 +5531,7 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch
dml2_printf("DML::%s: vm_bytes: %f (hvm inefficiency scaled)\n", __func__, vm_bytes*p->HostVMInefficiencyFactor);
dml2_printf("DML::%s: row_bytes: %f (hvm inefficiency scaled, 1 row)\n", __func__, p->PixelPTEBytesPerRow*p->HostVMInefficiencyFactor+p->meta_row_bytes+tdlut_row_bytes);
dml2_printf("DML::%s: Tno_bw: %f\n", __func__, *p->Tno_bw);
- dml2_printf("DML::%s: Tpre=%f Tpre_rounded: %f, delta=%f\n", __func__, Tpre, s->Tpre_rounded, (s->Tpre_rounded - Tpre));
+ dml2_printf("DML::%s: Tpre=%f Tpre_rounded: %f, delta=%f\n", __func__, Tpre, *p->Tpre_rounded, (*p->Tpre_rounded - Tpre));
dml2_printf("DML::%s: Tvm_trips=%f Tvm_trips_rounded: %f, delta=%f\n", __func__, *p->Tvm_trips, s->Tvm_trips_rounded, (s->Tvm_trips_rounded - *p->Tvm_trips));
#endif
@@ -5434,78 +5545,85 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch
// Tpre_rounded is Tpre rounding to 2-bit fraction
// Tvm_trips_rounded is Tvm_trips ceiling to 1/4 line time
// Tr0_trips_rounded is Tr0_trips ceiling to 1/4 line time
- // So that means prefetch bw calculated can be higher since the total time availabe for prefetch is less
- bool min_Lsw_equ_ok = s->Tpre_rounded >= s->Tvm_trips_rounded + 2.0*s->Tr0_trips_rounded + s->min_Lsw_equ*s->LineTime;
+ // So that means prefetch bw calculated can be higher since the total time available for prefetch is less
+ bool min_Lsw_equ_ok = *p->Tpre_rounded >= s->Tvm_trips_rounded + 2.0*s->Tr0_trips_rounded + s->min_Lsw_equ*s->LineTime;
+ bool tpre_gt_req_latency = true;
+#if 0
+ // Check that Tpre_rounded is big enough if all of the stages of the prefetch are time constrained.
+ // The terms Tvm_trips_rounded and Tr0_trips_rounded represent the min time constraints for the VM and row stages.
+ // Normally, these terms cover the overall time constraint for Tpre >= (Tex + max{Ttrip, Turg}), but if these terms are at their minimum, an explicit check is necessary.
+ tpre_gt_req_latency = *p->Tpre_rounded > (math_max2(p->Turg, s->trip_to_mem) + p->ExtraLatencyPrefetch);
+#endif
- if (s->dst_y_prefetch_equ > 1 && min_Lsw_equ_ok) {
+ if (s->dst_y_prefetch_equ > 1 && min_Lsw_equ_ok && tpre_gt_req_latency) {
s->prefetch_bw1 = 0.;
s->prefetch_bw2 = 0.;
s->prefetch_bw3 = 0.;
s->prefetch_bw4 = 0.;
// prefetch_bw1: VM + 2*R0 + SW
- if (s->Tpre_rounded - *p->Tno_bw > 0) {
+ if (*p->Tpre_rounded - *p->Tno_bw > 0) {
s->prefetch_bw1 = (vm_bytes * p->HostVMInefficiencyFactor
+ 2 * (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes)
- + s->prefetch_sw_bytes)
- / (s->Tpre_rounded - *p->Tno_bw);
- s->Tsw_est1 = s->prefetch_sw_bytes / s->prefetch_bw1;
+ + *p->prefetch_sw_bytes)
+ / (*p->Tpre_rounded - *p->Tno_bw);
+ s->Tsw_est1 = *p->prefetch_sw_bytes / s->prefetch_bw1;
} else
s->prefetch_bw1 = 0;
dml2_printf("DML::%s: prefetch_bw1: %f\n", __func__, s->prefetch_bw1);
- if ((s->Tsw_est1 < s->min_Lsw_equ * s->LineTime) && (s->Tpre_rounded - s->min_Lsw_equ * s->LineTime - 0.75 * s->LineTime - *p->Tno_bw > 0)) {
+ if ((s->Tsw_est1 < s->min_Lsw_equ * s->LineTime) && (*p->Tpre_rounded - s->min_Lsw_equ * s->LineTime - 0.75 * s->LineTime - *p->Tno_bw > 0)) {
s->prefetch_bw1 = (vm_bytes * p->HostVMInefficiencyFactor + 2 * (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes)) /
- (s->Tpre_rounded - s->min_Lsw_equ * s->LineTime - 0.75 * s->LineTime - *p->Tno_bw);
+ (*p->Tpre_rounded - s->min_Lsw_equ * s->LineTime - 0.75 * s->LineTime - *p->Tno_bw);
#ifdef __DML_VBA_DEBUG__
dml2_printf("DML::%s: vm and 2 rows bytes = %f\n", __func__, (vm_bytes * p->HostVMInefficiencyFactor + 2 * (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes)));
- dml2_printf("DML::%s: Tpre_rounded = %f\n", __func__, s->Tpre_rounded);
+ dml2_printf("DML::%s: Tpre_rounded = %f\n", __func__, *p->Tpre_rounded);
dml2_printf("DML::%s: minus term = %f\n", __func__, s->min_Lsw_equ * s->LineTime + 0.75 * s->LineTime + *p->Tno_bw);
dml2_printf("DML::%s: min_Lsw_equ = %f\n", __func__, s->min_Lsw_equ);
dml2_printf("DML::%s: LineTime = %f\n", __func__, s->LineTime);
dml2_printf("DML::%s: Tno_bw = %f\n", __func__, *p->Tno_bw);
- dml2_printf("DML::%s: Time to fetch vm and 2 rows = %f\n", __func__, (s->Tpre_rounded - s->min_Lsw_equ * s->LineTime - 0.75 * s->LineTime - *p->Tno_bw));
+ dml2_printf("DML::%s: Time to fetch vm and 2 rows = %f\n", __func__, (*p->Tpre_rounded - s->min_Lsw_equ * s->LineTime - 0.75 * s->LineTime - *p->Tno_bw));
dml2_printf("DML::%s: prefetch_bw1: %f (updated)\n", __func__, s->prefetch_bw1);
#endif
}
// prefetch_bw2: VM + SW
- if (s->Tpre_rounded - *p->Tno_bw - 2.0 * s->Tr0_trips_rounded > 0) {
- s->prefetch_bw2 = (vm_bytes * p->HostVMInefficiencyFactor + s->prefetch_sw_bytes) /
- (s->Tpre_rounded - *p->Tno_bw - 2.0 * s->Tr0_trips_rounded);
- s->Tsw_est2 = s->prefetch_sw_bytes / s->prefetch_bw2;
+ if (*p->Tpre_rounded - *p->Tno_bw - 2.0 * s->Tr0_trips_rounded > 0) {
+ s->prefetch_bw2 = (vm_bytes * p->HostVMInefficiencyFactor + *p->prefetch_sw_bytes) /
+ (*p->Tpre_rounded - *p->Tno_bw - 2.0 * s->Tr0_trips_rounded);
+ s->Tsw_est2 = *p->prefetch_sw_bytes / s->prefetch_bw2;
} else
s->prefetch_bw2 = 0;
dml2_printf("DML::%s: prefetch_bw2: %f\n", __func__, s->prefetch_bw2);
- if ((s->Tsw_est2 < s->min_Lsw_equ * s->LineTime) && ((s->Tpre_rounded - *p->Tno_bw - 2.0 * s->Tr0_trips_rounded - s->min_Lsw_equ * s->LineTime - 0.25 * s->LineTime) > 0)) {
- s->prefetch_bw2 = vm_bytes * p->HostVMInefficiencyFactor / (s->Tpre_rounded - *p->Tno_bw - 2.0 * s->Tr0_trips_rounded - s->min_Lsw_equ * s->LineTime - 0.25 * s->LineTime);
+ if ((s->Tsw_est2 < s->min_Lsw_equ * s->LineTime) && ((*p->Tpre_rounded - *p->Tno_bw - 2.0 * s->Tr0_trips_rounded - s->min_Lsw_equ * s->LineTime - 0.25 * s->LineTime) > 0)) {
+ s->prefetch_bw2 = vm_bytes * p->HostVMInefficiencyFactor / (*p->Tpre_rounded - *p->Tno_bw - 2.0 * s->Tr0_trips_rounded - s->min_Lsw_equ * s->LineTime - 0.25 * s->LineTime);
dml2_printf("DML::%s: prefetch_bw2: %f (updated)\n", __func__, s->prefetch_bw2);
}
// prefetch_bw3: 2*R0 + SW
- if (s->Tpre_rounded - s->Tvm_trips_rounded > 0) {
- s->prefetch_bw3 = (2 * (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes) + s->prefetch_sw_bytes) /
- (s->Tpre_rounded - s->Tvm_trips_rounded);
- s->Tsw_est3 = s->prefetch_sw_bytes / s->prefetch_bw3;
+ if (*p->Tpre_rounded - s->Tvm_trips_rounded > 0) {
+ s->prefetch_bw3 = (2 * (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes) + *p->prefetch_sw_bytes) /
+ (*p->Tpre_rounded - s->Tvm_trips_rounded);
+ s->Tsw_est3 = *p->prefetch_sw_bytes / s->prefetch_bw3;
} else
s->prefetch_bw3 = 0;
dml2_printf("DML::%s: prefetch_bw3: %f\n", __func__, s->prefetch_bw3);
- if ((s->Tsw_est3 < s->min_Lsw_equ * s->LineTime) && ((s->Tpre_rounded - s->min_Lsw_equ * s->LineTime - 0.5 * s->LineTime - s->Tvm_trips_rounded) > 0)) {
- s->prefetch_bw3 = (2 * (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes)) / (s->Tpre_rounded - s->min_Lsw_equ * s->LineTime - 0.5 * s->LineTime - s->Tvm_trips_rounded);
+ if ((s->Tsw_est3 < s->min_Lsw_equ * s->LineTime) && ((*p->Tpre_rounded - s->min_Lsw_equ * s->LineTime - 0.5 * s->LineTime - s->Tvm_trips_rounded) > 0)) {
+ s->prefetch_bw3 = (2 * (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes)) / (*p->Tpre_rounded - s->min_Lsw_equ * s->LineTime - 0.5 * s->LineTime - s->Tvm_trips_rounded);
dml2_printf("DML::%s: prefetch_bw3: %f (updated)\n", __func__, s->prefetch_bw3);
}
// prefetch_bw4: SW
- if (s->Tpre_rounded - s->Tvm_trips_rounded - 2 * s->Tr0_trips_rounded > 0)
- s->prefetch_bw4 = s->prefetch_sw_bytes / (s->Tpre_rounded - s->Tvm_trips_rounded - 2 * s->Tr0_trips_rounded);
+ if (*p->Tpre_rounded - s->Tvm_trips_rounded - 2 * s->Tr0_trips_rounded > 0)
+ s->prefetch_bw4 = *p->prefetch_sw_bytes / (*p->Tpre_rounded - s->Tvm_trips_rounded - 2 * s->Tr0_trips_rounded);
else
s->prefetch_bw4 = 0;
#ifdef __DML_VBA_DEBUG__
dml2_printf("DML::%s: Tno_bw: %f\n", __func__, *p->Tno_bw);
- dml2_printf("DML::%s: Tpre=%f Tpre_rounded: %f, delta=%f\n", __func__, Tpre, s->Tpre_rounded, (s->Tpre_rounded - Tpre));
+ dml2_printf("DML::%s: Tpre=%f Tpre_rounded: %f, delta=%f\n", __func__, Tpre, *p->Tpre_rounded, (*p->Tpre_rounded - Tpre));
dml2_printf("DML::%s: Tvm_trips=%f Tvm_trips_rounded: %f, delta=%f\n", __func__, *p->Tvm_trips, s->Tvm_trips_rounded, (s->Tvm_trips_rounded - *p->Tvm_trips));
dml2_printf("DML::%s: Tr0_trips=%f Tr0_trips_rounded: %f, delta=%f\n", __func__, *p->Tr0_trips, s->Tr0_trips_rounded, (s->Tr0_trips_rounded - *p->Tr0_trips));
dml2_printf("DML::%s: Tsw_est1: %f\n", __func__, s->Tsw_est1);
@@ -5617,9 +5735,6 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch
dml2_printf("DML::%s: Tvm_equ = %f\n", __func__, s->Tvm_equ);
dml2_printf("DML::%s: Tr0_equ = %f\n", __func__, s->Tr0_equ);
#endif
- // Lsw = dst_y_prefetch - (dst_y_per_vm_vblank + 2*dst_y_per_row_vblank)
- s->Lsw_equ = s->dst_y_prefetch_equ - math_ceil2(4.0 * (s->Tvm_equ + 2 * s->Tr0_equ) / s->LineTime, 1.0) / 4.0;
-
// Use the more stressful prefetch schedule
if (s->dst_y_prefetch_oto < s->dst_y_prefetch_equ) {
*p->dst_y_prefetch = s->dst_y_prefetch_oto;
@@ -5628,31 +5743,33 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch
*p->dst_y_per_vm_vblank = math_ceil2(4.0 * s->TimeForFetchingVM / s->LineTime, 1.0) / 4.0;
*p->dst_y_per_row_vblank = math_ceil2(4.0 * s->TimeForFetchingRowInVBlank / s->LineTime, 1.0) / 4.0;
- s->dst_y_per_vm_no_trip_vblank = math_ceil2(4.0 * s->Tvm_no_trip_oto / s->LineTime, 1.0) / 4.0;
- s->dst_y_per_row_no_trip_vblank = math_ceil2(4.0 * s->Tr0_no_trip_oto / s->LineTime, 1.0) / 4.0;
#ifdef __DML_VBA_DEBUG__
dml2_printf("DML::%s: Using oto scheduling for prefetch\n", __func__);
#endif
+
} else {
*p->dst_y_prefetch = s->dst_y_prefetch_equ;
+
+ if (s->dst_y_prefetch_equ < s->dst_y_prefetch_equ_impacted)
+ *p->dst_y_prefetch = s->dst_y_prefetch_equ_impacted;
+
s->TimeForFetchingVM = s->Tvm_equ;
s->TimeForFetchingRowInVBlank = s->Tr0_equ;
- *p->dst_y_per_vm_vblank = math_ceil2(4.0 * s->TimeForFetchingVM / s->LineTime, 1.0) / 4.0;
- *p->dst_y_per_row_vblank = math_ceil2(4.0 * s->TimeForFetchingRowInVBlank / s->LineTime, 1.0) / 4.0;
- s->dst_y_per_vm_no_trip_vblank = *p->dst_y_per_vm_vblank;
- s->dst_y_per_row_no_trip_vblank = *p->dst_y_per_row_vblank;
+ *p->dst_y_per_vm_vblank = math_ceil2(4.0 * s->TimeForFetchingVM / s->LineTime, 1.0) / 4.0;
+ *p->dst_y_per_row_vblank = math_ceil2(4.0 * s->TimeForFetchingRowInVBlank / s->LineTime, 1.0) / 4.0;
#ifdef __DML_VBA_DEBUG__
dml2_printf("DML::%s: Using equ bw scheduling for prefetch\n", __func__);
#endif
}
- /* take worst case Lsw to calculate bandwidth requirement regardless of schedule */
- s->LinesToRequestPrefetchPixelData = math_min2(s->Lsw_equ, s->Lsw_oto); // Lsw
+ // Lsw = dst_y_prefetch - (dst_y_per_vm_vblank + 2*dst_y_per_row_vblank)
+ s->LinesToRequestPrefetchPixelData = *p->dst_y_prefetch - *p->dst_y_per_vm_vblank - 2 * *p->dst_y_per_row_vblank; // Lsw
s->cursor_prefetch_bytes = (unsigned int)math_max2(p->cursor_bytes_per_chunk, 4 * p->cursor_bytes_per_line);
*p->prefetch_cursor_bw = p->num_cursors * s->cursor_prefetch_bytes / (s->LinesToRequestPrefetchPixelData * s->LineTime);
+ *p->prefetch_swath_time_us = (s->LinesToRequestPrefetchPixelData * s->LineTime);
#ifdef __DML_VBA_DEBUG__
dml2_printf("DML::%s: TimeForFetchingVM = %f\n", __func__, s->TimeForFetchingVM);
@@ -5663,6 +5780,7 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch
dml2_printf("DML::%s: dst_y_per_row_vblank = %f\n", __func__, *p->dst_y_per_row_vblank);
dml2_printf("DML::%s: LinesToRequestPrefetchPixelData = %f\n", __func__, s->LinesToRequestPrefetchPixelData);
dml2_printf("DML::%s: PrefetchSourceLinesY = %f\n", __func__, p->PrefetchSourceLinesY);
+ dml2_printf("DML::%s: prefetch_swath_time_us = %f\n", __func__, *p->prefetch_swath_time_us);
dml2_printf("DML::%s: cursor_bytes_per_chunk = %d\n", __func__, p->cursor_bytes_per_chunk);
dml2_printf("DML::%s: cursor_bytes_per_line = %d\n", __func__, p->cursor_bytes_per_line);
@@ -5749,8 +5867,10 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch
} else {
dml2_printf("DML::%s: No time to prefetch! dst_y_prefetch_equ = %f (should be > 1)\n", __func__, s->dst_y_prefetch_equ);
- dml2_printf("DML::%s: No time to prefetch! min_Lsw_equ_ok = %d, Tpre_rounded (%f) should be >= Tvm_trips_rounded (%f) + 2.0*Tr0_trips_rounded (%f) + min_Tsw_equ (%f)\n",
- __func__, min_Lsw_equ_ok, s->Tpre_rounded, s->Tvm_trips_rounded, 2.0*s->Tr0_trips_rounded, s->min_Lsw_equ*s->LineTime);
+ dml2_printf("DML::%s: No time to prefetch! min_Lsw_equ_ok = %d, Tpre_rounded (%f) should be >= Tvm_trips_rounded (%f) + 2.0*Tr0_trips_rounded (%f) + min_Tsw_equ (%f)\n",
+ __func__, min_Lsw_equ_ok, *p->Tpre_rounded, s->Tvm_trips_rounded, 2.0*s->Tr0_trips_rounded, s->min_Lsw_equ*s->LineTime);
+ dml2_printf("DML::%s: No time to prefetch! min_Lsw_equ_ok = %d, Tpre_rounded+Tvm_trips_rounded+2.0*Tr0_trips_rounded+min_Tsw_equ (%f) should be > \n",
+ __func__, tpre_gt_req_latency, (s->min_Lsw_equ*s->LineTime + s->Tvm_trips_rounded + 2.0*s->Tr0_trips_rounded), p->Turg, s->trip_to_mem, p->ExtraLatencyPrefetch);
s->NoTimeToPrefetch = true;
s->TimeForFetchingVM = 0;
s->TimeForFetchingRowInVBlank = 0;
@@ -5769,13 +5889,13 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch
if (vm_bytes == 0) {
prefetch_vm_bw = 0;
- } else if (s->dst_y_per_vm_no_trip_vblank > 0) {
+ } else if (*p->dst_y_per_vm_vblank > 0) {
#ifdef __DML_VBA_DEBUG__
dml2_printf("DML::%s: HostVMInefficiencyFactor = %f\n", __func__, p->HostVMInefficiencyFactor);
dml2_printf("DML::%s: dst_y_per_vm_vblank = %f\n", __func__, *p->dst_y_per_vm_vblank);
dml2_printf("DML::%s: LineTime = %f\n", __func__, s->LineTime);
#endif
- prefetch_vm_bw = vm_bytes * p->HostVMInefficiencyFactor / (s->dst_y_per_vm_no_trip_vblank * s->LineTime);
+ prefetch_vm_bw = vm_bytes * p->HostVMInefficiencyFactor / (*p->dst_y_per_vm_vblank * s->LineTime);
#ifdef __DML_VBA_DEBUG__
dml2_printf("DML::%s: prefetch_vm_bw = %f\n", __func__, prefetch_vm_bw);
#endif
@@ -5787,8 +5907,8 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch
if (p->PixelPTEBytesPerRow == 0 && tdlut_row_bytes == 0) {
prefetch_row_bw = 0;
- } else if (s->dst_y_per_row_no_trip_vblank > 0) {
- prefetch_row_bw = (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + tdlut_row_bytes) / (s->dst_y_per_row_no_trip_vblank * s->LineTime);
+ } else if (*p->dst_y_per_row_vblank > 0) {
+ prefetch_row_bw = (p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + tdlut_row_bytes) / (*p->dst_y_per_row_vblank * s->LineTime);
#ifdef __DML_VBA_DEBUG__
dml2_printf("DML::%s: PixelPTEBytesPerRow = %u\n", __func__, p->PixelPTEBytesPerRow);
@@ -5828,6 +5948,171 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch
return s->NoTimeToPrefetch;
}
+static unsigned int get_num_lb_source_lines(unsigned int max_line_buffer_lines,
+ unsigned int line_buffer_size_bits,
+ unsigned int num_pipes,
+ unsigned int vp_width,
+ unsigned int vp_height,
+ double h_ratio,
+ enum dml2_rotation_angle rotation_angle)
+{
+ unsigned int num_lb_source_lines = 0;
+ double lb_bit_per_pixel = 57.0;
+ unsigned recin_width = vp_width/num_pipes;
+
+ if (dml_is_vertical_rotation(rotation_angle))
+ recin_width = vp_height/num_pipes;
+
+ num_lb_source_lines = (unsigned int) math_min2((double) max_line_buffer_lines,
+ math_floor2(line_buffer_size_bits / lb_bit_per_pixel / (recin_width / math_max2(h_ratio, 1.0)), 1.0));
+
+ return num_lb_source_lines;
+}
+
+static unsigned int find_max_impact_plane(unsigned int this_plane_idx, unsigned int num_planes, unsigned int Trpd_dcfclk_cycles[])
+{
+ int max_value = -1;
+ int max_idx = -1;
+ for (unsigned int i = 0; i < num_planes; i++) {
+ if (i != this_plane_idx && (int) Trpd_dcfclk_cycles[i] > max_value) {
+ max_value = Trpd_dcfclk_cycles[i];
+ max_idx = i;
+ }
+ }
+ if (max_idx <= 0) {
+ dml2_assert(max_idx >= 0);
+ max_idx = this_plane_idx;
+ }
+
+ return max_idx;
+}
+
+static double calculate_impacted_Tsw(unsigned int exclude_plane_idx, unsigned int num_planes, double *prefetch_swath_bytes, double bw_mbps)
+{
+ double sum = 0.;
+ for (unsigned int i = 0; i < num_planes; i++) {
+ if (i != exclude_plane_idx) {
+ sum += prefetch_swath_bytes[i];
+ }
+ }
+ return sum / bw_mbps;
+}
+
+// a global check against the aggregate effect of the per plane prefetch schedule
+static bool CheckGlobalPrefetchAdmissibility(struct dml2_core_internal_scratch *scratch,
+ struct dml2_core_calcs_CheckGlobalPrefetchAdmissibility_params *p)
+{
+ struct dml2_core_calcs_CheckGlobalPrefetchAdmissibility_locals *s = &scratch->CheckGlobalPrefetchAdmissibility_locals;
+ unsigned int i, k;
+
+ memset(s, 0, sizeof(struct dml2_core_calcs_CheckGlobalPrefetchAdmissibility_locals));
+
+ *p->recalc_prefetch_schedule = 0;
+ s->prefetch_global_check_passed = 1;
+ // worst case if the rob and cdb is fully hogged
+ s->max_Trpd_dcfclk_cycles = (unsigned int) math_ceil2((p->rob_buffer_size_kbytes*1024 + p->compressed_buffer_size_kbytes*DML_MAX_COMPRESSION_RATIO*1024)/64.0, 1.0);
+#ifdef __DML_VBA_DEBUG__
+ dml2_printf("DML::%s: num_active_planes = %d\n", __func__, p->num_active_planes);
+ dml2_printf("DML::%s: rob_buffer_size_kbytes = %d\n", __func__, p->rob_buffer_size_kbytes);
+ dml2_printf("DML::%s: compressed_buffer_size_kbytes = %d\n", __func__, p->compressed_buffer_size_kbytes);
+ dml2_printf("DML::%s: estimated_urg_bandwidth_required_mbps = %f\n", __func__, p->estimated_urg_bandwidth_required_mbps);
+ dml2_printf("DML::%s: estimated_dcfclk_mhz = %f\n", __func__, p->estimated_dcfclk_mhz);
+ dml2_printf("DML::%s: max_Trpd_dcfclk_cycles = %u\n", __func__, s->max_Trpd_dcfclk_cycles);
+#endif
+
+ // calculate the return impact from each plane, request is 256B per dcfclk
+ for (i = 0; i < p->num_active_planes; i++) {
+ s->src_detile_buf_size_bytes_l[i] = p->detile_buffer_size_bytes_l[i];
+ s->src_detile_buf_size_bytes_c[i] = p->detile_buffer_size_bytes_c[i];
+ s->src_swath_bytes_l[i] = p->full_swath_bytes_l[i];
+ s->src_swath_bytes_c[i] = p->full_swath_bytes_c[i];
+
+ if (p->pixel_format[i] == dml2_420_10) {
+ s->src_detile_buf_size_bytes_l[i] = (unsigned int) (s->src_detile_buf_size_bytes_l[i] * 1.5);
+ s->src_detile_buf_size_bytes_c[i] = (unsigned int) (s->src_detile_buf_size_bytes_c[i] * 1.5);
+ s->src_swath_bytes_l[i] = (unsigned int) (s->src_swath_bytes_l[i] * 1.5);
+ s->src_swath_bytes_c[i] = (unsigned int) (s->src_swath_bytes_c[i] * 1.5);
+ }
+
+ s->burst_bytes_to_fill_det = (unsigned int) (math_floor2(s->src_detile_buf_size_bytes_l[i] / p->chunk_bytes_l, 1) * p->chunk_bytes_l);
+ s->burst_bytes_to_fill_det += (unsigned int) (math_floor2(p->lb_source_lines_l[i] / p->swath_height_l[i], 1) * s->src_swath_bytes_l[i]);
+
+#ifdef __DML_VBA_DEBUG__
+ dml2_printf("DML::%s: i=%u pixel_format = %d\n", __func__, i, p->pixel_format[i]);
+ dml2_printf("DML::%s: i=%u chunk_bytes_l = %d\n", __func__, i, p->chunk_bytes_l);
+ dml2_printf("DML::%s: i=%u lb_source_lines_l = %d\n", __func__, i, p->lb_source_lines_l[i]);
+ dml2_printf("DML::%s: i=%u src_detile_buf_size_bytes_l=%d\n", __func__, i, s->src_detile_buf_size_bytes_l[i]);
+ dml2_printf("DML::%s: i=%u src_swath_bytes_l=%d\n", __func__, i, s->src_swath_bytes_l[i]);
+ dml2_printf("DML::%s: i=%u burst_bytes_to_fill_det=%d (luma)\n", __func__, i, s->burst_bytes_to_fill_det);
+#endif
+
+ if (s->src_swath_bytes_c[i] > 0) { // dual_plane
+ s->burst_bytes_to_fill_det += (unsigned int) (math_floor2(s->src_detile_buf_size_bytes_c[i] / p->chunk_bytes_c, 1) * p->chunk_bytes_c);
+
+ if (p->pixel_format[i] == dml2_422_planar_8 || p->pixel_format[i] == dml2_422_planar_10 || p->pixel_format[i] == dml2_422_planar_12) {
+ s->burst_bytes_to_fill_det += (unsigned int) (math_floor2(p->lb_source_lines_c[i] / p->swath_height_c[i], 1) * s->src_swath_bytes_c[i]);
+ }
+
+#ifdef __DML_VBA_DEBUG__
+ dml2_printf("DML::%s: i=%u chunk_bytes_c = %d\n", __func__, i, p->chunk_bytes_c);
+ dml2_printf("DML::%s: i=%u lb_source_lines_c = %d\n", __func__, i, p->lb_source_lines_c[i]);
+ dml2_printf("DML::%s: i=%u src_detile_buf_size_bytes_c=%d\n", __func__, i, s->src_detile_buf_size_bytes_c[i]);
+ dml2_printf("DML::%s: i=%u src_swath_bytes_c=%d\n", __func__, i, s->src_swath_bytes_c[i]);
+#endif
+ }
+
+ s->time_to_fill_det_us = (double) s->burst_bytes_to_fill_det / (256 * p->estimated_dcfclk_mhz); // fill time assume full burst at request rate
+ s->accumulated_return_path_dcfclk_cycles[i] = (unsigned int) math_ceil2(((DML_MAX_COMPRESSION_RATIO-1) * 64 * p->estimated_dcfclk_mhz) * s->time_to_fill_det_us / 64.0, 1.0); //for 64B per DCFClk
+
+#ifdef __DML_VBA_DEBUG__
+ dml2_printf("DML::%s: i=%u burst_bytes_to_fill_det=%d\n", __func__, i, s->burst_bytes_to_fill_det);
+ dml2_printf("DML::%s: i=%u time_to_fill_det_us=%f\n", __func__, i, s->time_to_fill_det_us);
+ dml2_printf("DML::%s: i=%u accumulated_return_path_dcfclk_cycles=%u\n", __func__, i, s->accumulated_return_path_dcfclk_cycles[i]);
+#endif
+ // clamping to worst case delay which is one which occupy the full rob+cdb
+ if (s->accumulated_return_path_dcfclk_cycles[i] > s->max_Trpd_dcfclk_cycles)
+ s->accumulated_return_path_dcfclk_cycles[i] = s->max_Trpd_dcfclk_cycles;
+ }
+
+ // Figure out the impacted prefetch time for each plane
+ // if impacted_Tre is > equ bw Tpre, we need to fail the prefetch schedule as we need a higher state to support the bw
+ for (i = 0; i < p->num_active_planes; i++) {
+ k = find_max_impact_plane(i, p->num_active_planes, s->accumulated_return_path_dcfclk_cycles); // plane k causes most impact to plane i
+ // the rest of planes (except for k) complete for bw
+ p->impacted_dst_y_pre[i] = s->accumulated_return_path_dcfclk_cycles[k]/p->estimated_dcfclk_mhz;
+ p->impacted_dst_y_pre[i] += calculate_impacted_Tsw(k, p->num_active_planes, p->prefetch_sw_bytes, p->estimated_urg_bandwidth_required_mbps);
+ p->impacted_dst_y_pre[i] = math_ceil2(p->impacted_dst_y_pre[i] / p->line_time[i], 0.25);
+
+#ifdef __DML_VBA_DEBUG__
+ dml2_printf("DML::%s: i=%u impacted_Tpre=%f (k=%u)\n", __func__, i, p->impacted_dst_y_pre[i], k);
+#endif
+ }
+
+ if (p->Tpre_rounded != NULL && p->Tpre_oto != NULL) {
+ for (i = 0; i < p->num_active_planes; i++) {
+ if (p->impacted_dst_y_pre[i] > p->dst_y_prefetch[i]) {
+ s->prefetch_global_check_passed = 0;
+ *p->recalc_prefetch_schedule = 1;
+ }
+#ifdef __DML_VBA_DEBUG__
+ dml2_printf("DML::%s: i=%u Tpre_rounded=%f\n", __func__, i, p->Tpre_rounded[i]);
+ dml2_printf("DML::%s: i=%u Tpre_oto=%f\n", __func__, i, p->Tpre_oto[i]);
+#endif
+ }
+ } else {
+ // likely a mode programming calls, assume support, and no recalc - not used anyways
+ s->prefetch_global_check_passed = 1;
+ *p->recalc_prefetch_schedule = 0;
+ }
+
+#ifdef __DML_VBA_DEBUG__
+ dml2_printf("DML::%s: prefetch_global_check_passed=%u\n", __func__, s->prefetch_global_check_passed);
+ dml2_printf("DML::%s: recalc_prefetch_schedule=%u\n", __func__, *p->recalc_prefetch_schedule);
+#endif
+
+ return s->prefetch_global_check_passed;
+}
+
static void calculate_peak_bandwidth_required(
struct dml2_core_internal_scratch *s,
struct dml2_core_calcs_calculate_peak_bandwidth_required_params *p)
@@ -6046,7 +6331,7 @@ static void check_urgent_bandwidth_support(
double *frac_urg_bandwidth_nom,
double *frac_urg_bandwidth_mall,
bool *vactive_bandwidth_support_ok, // vactive ok
- bool *bandwidth_support_ok, // max of vm, prefetch, vactive all ok
+ bool *bandwidth_support_ok,// max of vm, prefetch, vactive all ok
unsigned int mall_allocated_for_dcn_mbytes,
double non_urg_bandwidth_required[dml2_core_internal_soc_state_max][dml2_core_internal_bw_max],
@@ -6116,7 +6401,6 @@ static void check_urgent_bandwidth_support(
}
}
#endif
-
}
static double get_bandwidth_available_for_immediate_flip(enum dml2_core_internal_soc_state_type eval_state,
@@ -6438,7 +6722,7 @@ static void CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport(
p->Watermark->Z8StutterExitWatermark += p->mmSOCParameters.max_urgent_latency_us + p->mmSOCParameters.df_response_time_us;
p->Watermark->Z8StutterEnterPlusExitWatermark += p->mmSOCParameters.max_urgent_latency_us + p->mmSOCParameters.df_response_time_us;
}
- p->Watermark->g6_temp_read_watermark_us = p->mmSOCParameters.g6_temp_read_blackout_us + p->Watermark->UrgentWatermark;
+ p->Watermark->temp_read_or_ppt_watermark_us = p->mmSOCParameters.g6_temp_read_blackout_us + p->Watermark->UrgentWatermark;
#ifdef __DML_VBA_DEBUG__
dml2_printf("DML::%s: UrgentLatency = %f\n", __func__, p->mmSOCParameters.UrgentLatency);
@@ -6454,12 +6738,12 @@ static void CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport(
dml2_printf("DML::%s: StutterEnterPlusExitWatermark = %f\n", __func__, p->Watermark->StutterEnterPlusExitWatermark);
dml2_printf("DML::%s: Z8StutterExitWatermark = %f\n", __func__, p->Watermark->Z8StutterExitWatermark);
dml2_printf("DML::%s: Z8StutterEnterPlusExitWatermark = %f\n", __func__, p->Watermark->Z8StutterEnterPlusExitWatermark);
- dml2_printf("DML::%s: g6_temp_read_watermark_us = %f\n", __func__, p->Watermark->g6_temp_read_watermark_us);
+ dml2_printf("DML::%s: temp_read_or_ppt_watermark_us = %f\n", __func__, p->Watermark->temp_read_or_ppt_watermark_us);
#endif
s->TotalActiveWriteback = 0;
for (unsigned int k = 0; k < p->NumberOfActiveSurfaces; ++k) {
- if (p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].writeback.enable == true) {
+ if (p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].writeback.active_writebacks_per_stream > 0) {
s->TotalActiveWriteback = s->TotalActiveWriteback + 1;
}
}
@@ -6522,7 +6806,7 @@ static void CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport(
s->LBLatencyHidingSourceLinesC[k] = (unsigned int)(math_min2((double)p->MaxLineBufferLines, math_floor2((double)p->LineBufferSize / LBBitPerPixel / ((double)p->SwathWidthC[k] / math_max2(h_ratio_c, 1.0)), 1)) - (v_taps_c - 1));
#ifdef __DML_VBA_DEBUG__
- dml2_printf("DML::%s: k=%u, MaxLineBufferLines= %u\n", __func__, k, p->MaxLineBufferLines);
+ dml2_printf("DML::%s: k=%u, MaxLineBufferLines = %u\n", __func__, k, p->MaxLineBufferLines);
dml2_printf("DML::%s: k=%u, LineBufferSize = %u\n", __func__, k, p->LineBufferSize);
dml2_printf("DML::%s: k=%u, LBBitPerPixel = %u\n", __func__, k, LBBitPerPixel);
dml2_printf("DML::%s: k=%u, HRatio = %f\n", __func__, k, h_ratio);
@@ -6563,7 +6847,7 @@ static void CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport(
s->ActiveDRAMClockChangeLatencyMargin[k] = s->ActiveClockChangeLatencyHiding - p->Watermark->DRAMClockChangeWatermark;
s->ActiveFCLKChangeLatencyMargin[k] = s->ActiveClockChangeLatencyHiding - p->Watermark->FCLKChangeWatermark;
s->USRRetrainingLatencyMargin[k] = s->ActiveClockChangeLatencyHiding - p->Watermark->USRRetrainingWatermark;
- s->g6_temp_read_latency_margin[k] = s->ActiveClockChangeLatencyHiding - p->Watermark->g6_temp_read_watermark_us;
+ s->g6_temp_read_latency_margin[k] = s->ActiveClockChangeLatencyHiding - p->Watermark->temp_read_or_ppt_watermark_us;
if (p->VActiveLatencyHidingMargin)
p->VActiveLatencyHidingMargin[k] = s->ActiveDRAMClockChangeLatencyMargin[k];
@@ -6571,9 +6855,12 @@ static void CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport(
if (p->VActiveLatencyHidingUs)
p->VActiveLatencyHidingUs[k] = s->ActiveClockChangeLatencyHiding;
- if (p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].writeback.enable) {
- s->WritebackLatencyHiding = (double)p->WritebackInterfaceBufferSize * 1024.0 / ((double)p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.output_height * (double)p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.output_width / ((double)p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.input_height * (double)h_total / pixel_clock_mhz) * 4.0);
- if (p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].writeback.pixel_format == dml2_444_64) {
+ if (p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].writeback.active_writebacks_per_stream > 0) {
+ s->WritebackLatencyHiding = (double)p->WritebackInterfaceBufferSize * 1024.0
+ / ((double)p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].output_height
+ * (double)p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].output_width
+ / ((double)p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].input_height * (double)h_total / pixel_clock_mhz) * 4.0);
+ if (p->display_cfg->stream_descriptors[p->display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].pixel_format == dml2_444_64) {
s->WritebackLatencyHiding = s->WritebackLatencyHiding / 2;
}
s->WritebackDRAMClockChangeLatencyMargin = s->WritebackLatencyHiding - p->Watermark->WritebackDRAMClockChangeWatermark;
@@ -6588,36 +6875,36 @@ static void CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport(
uclk_pstate_change_strategy = p->display_cfg->plane_descriptors[k].overrides.uclk_pstate_change_strategy;
reserved_vblank_time_us = (double)p->display_cfg->plane_descriptors[k].overrides.reserved_vblank_time_ns / 1000;
- p->FCLKChangeSupport[k] = dml2_fclock_change_unsupported;
+ p->FCLKChangeSupport[k] = dml2_pstate_change_unsupported;
if (s->ActiveFCLKChangeLatencyMargin[k] > 0)
- p->FCLKChangeSupport[k] = dml2_fclock_change_vactive;
+ p->FCLKChangeSupport[k] = dml2_pstate_change_vactive;
else if (reserved_vblank_time_us >= p->mmSOCParameters.FCLKChangeLatency)
- p->FCLKChangeSupport[k] = dml2_fclock_change_vblank;
+ p->FCLKChangeSupport[k] = dml2_pstate_change_vblank;
- if (p->FCLKChangeSupport[k] == dml2_fclock_change_unsupported)
+ if (p->FCLKChangeSupport[k] == dml2_pstate_change_unsupported)
*p->global_fclk_change_supported = false;
- p->DRAMClockChangeSupport[k] = dml2_dram_clock_change_unsupported;
+ p->DRAMClockChangeSupport[k] = dml2_pstate_change_unsupported;
if (uclk_pstate_change_strategy == dml2_uclk_pstate_change_strategy_auto) {
if (p->display_cfg->overrides.all_streams_blanked ||
(s->ActiveDRAMClockChangeLatencyMargin[k] > 0 && reserved_vblank_time_us >= p->mmSOCParameters.DRAMClockChangeLatency))
- p->DRAMClockChangeSupport[k] = dml2_dram_clock_change_vblank_and_vactive;
+ p->DRAMClockChangeSupport[k] = dml2_pstate_change_vblank_and_vactive;
else if (s->ActiveDRAMClockChangeLatencyMargin[k] > 0)
- p->DRAMClockChangeSupport[k] = dml2_dram_clock_change_vactive;
+ p->DRAMClockChangeSupport[k] = dml2_pstate_change_vactive;
else if (reserved_vblank_time_us >= p->mmSOCParameters.DRAMClockChangeLatency)
- p->DRAMClockChangeSupport[k] = dml2_dram_clock_change_vblank;
+ p->DRAMClockChangeSupport[k] = dml2_pstate_change_vblank;
} else if (uclk_pstate_change_strategy == dml2_uclk_pstate_change_strategy_force_vactive && s->ActiveDRAMClockChangeLatencyMargin[k] > 0)
- p->DRAMClockChangeSupport[k] = dml2_dram_clock_change_vactive;
+ p->DRAMClockChangeSupport[k] = dml2_pstate_change_vactive;
else if (uclk_pstate_change_strategy == dml2_uclk_pstate_change_strategy_force_vblank && reserved_vblank_time_us >= p->mmSOCParameters.DRAMClockChangeLatency)
- p->DRAMClockChangeSupport[k] = dml2_dram_clock_change_vblank;
+ p->DRAMClockChangeSupport[k] = dml2_pstate_change_vblank;
else if (uclk_pstate_change_strategy == dml2_uclk_pstate_change_strategy_force_drr)
- p->DRAMClockChangeSupport[k] = dml2_dram_clock_change_drr;
+ p->DRAMClockChangeSupport[k] = dml2_pstate_change_drr;
else if (uclk_pstate_change_strategy == dml2_uclk_pstate_change_strategy_force_mall_svp)
- p->DRAMClockChangeSupport[k] = dml2_dram_clock_change_mall_svp;
+ p->DRAMClockChangeSupport[k] = dml2_pstate_change_mall_svp;
else if (uclk_pstate_change_strategy == dml2_uclk_pstate_change_strategy_force_mall_full_frame)
- p->DRAMClockChangeSupport[k] = dml2_dram_clock_change_mall_full_frame;
+ p->DRAMClockChangeSupport[k] = dml2_pstate_change_mall_full_frame;
- if (p->DRAMClockChangeSupport[k] == dml2_dram_clock_change_unsupported)
+ if (p->DRAMClockChangeSupport[k] == dml2_pstate_change_unsupported)
*p->global_dram_clock_change_supported = false;
s->dst_y_pstate = (unsigned int)(math_ceil2((p->mmSOCParameters.DRAMClockChangeLatency + p->mmSOCParameters.UrgentLatency) / (h_total / pixel_clock_mhz), 1));
@@ -6915,8 +7202,7 @@ struct dml2_core_internal_g6_temp_read_blackouts_table {
} entries[DML_MAX_CLK_TABLE_SIZE];
};
-static const struct dml2_core_internal_g6_temp_read_blackouts_table
- core_dcn4_g6_temp_read_blackout_table = {
+struct dml2_core_internal_g6_temp_read_blackouts_table core_dcn4_g6_temp_read_blackout_table = {
.entries = {
{
.uclk_khz = 96000,
@@ -7036,6 +7322,9 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out
struct dml2_core_calcs_CalculateVMRowAndSwath_params *CalculateVMRowAndSwath_params = &mode_lib->scratch.CalculateVMRowAndSwath_params;
struct dml2_core_calcs_CalculateSwathAndDETConfiguration_params *CalculateSwathAndDETConfiguration_params = &mode_lib->scratch.CalculateSwathAndDETConfiguration_params;
struct dml2_core_calcs_CalculatePrefetchSchedule_params *CalculatePrefetchSchedule_params = &mode_lib->scratch.CalculatePrefetchSchedule_params;
+#ifdef DML_GLOBAL_PREFETCH_CHECK
+ struct dml2_core_calcs_CheckGlobalPrefetchAdmissibility_params *CheckGlobalPrefetchAdmissibility_params = &mode_lib->scratch.CheckGlobalPrefetchAdmissibility_params;
+#endif
struct dml2_core_calcs_calculate_tdlut_setting_params *calculate_tdlut_setting_params = &mode_lib->scratch.calculate_tdlut_setting_params;
struct dml2_core_calcs_calculate_mcache_setting_params *calculate_mcache_setting_params = &mode_lib->scratch.calculate_mcache_setting_params;
struct dml2_core_calcs_calculate_peak_bandwidth_required_params *calculate_peak_bandwidth_params = &mode_lib->scratch.calculate_peak_bandwidth_params;
@@ -7083,12 +7372,6 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out
for (k = 0; k < mode_lib->ms.num_active_planes; k++)
dml2_printf("DML::%s: plane_%d: reserved_vblank_time_ns = %u\n", __func__, k, display_cfg->plane_descriptors[k].overrides.reserved_vblank_time_ns);
-
- // dml2_printf_dml_policy(&mode_lib->ms.policy);
- // dml2_printf_dml_display_cfg_timing(&display_cfg->timing, mode_lib->ms.num_active_planes);
- // dml2_printf_dml_display_cfg_plane(&display_cfg->plane, mode_lib->ms.num_active_planes);
- // dml2_printf_dml_display_cfg_surface(&display_cfg->surface, mode_lib->ms.num_active_planes);
- // dml2_printf_dml_display_cfg_output(&display_cfg->output, mode_lib->ms.num_active_planes);
#endif
CalculateMaxDETAndMinCompressedBufferSize(
@@ -7183,8 +7466,8 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out
}
for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
- mode_lib->ms.SurfaceReadBandwidthLuma[k] = mode_lib->ms.SwathWidthYSingleDPP[k] * math_ceil2(mode_lib->ms.BytePerPixelY[k], 1.0) / (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)) * display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio;
- mode_lib->ms.SurfaceReadBandwidthChroma[k] = mode_lib->ms.SwathWidthCSingleDPP[k] * math_ceil2(mode_lib->ms.BytePerPixelC[k], 2.0) / (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)) * display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio;
+ mode_lib->ms.vactive_sw_bw_l[k] = mode_lib->ms.SwathWidthYSingleDPP[k] * math_ceil2(mode_lib->ms.BytePerPixelY[k], 1.0) / (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)) * display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio;
+ mode_lib->ms.vactive_sw_bw_c[k] = mode_lib->ms.SwathWidthCSingleDPP[k] * math_ceil2(mode_lib->ms.BytePerPixelC[k], 2.0) / (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)) * display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio;
mode_lib->ms.cursor_bw[k] = display_cfg->plane_descriptors[k].cursor.num_cursors * display_cfg->plane_descriptors[k].cursor.cursor_width *
display_cfg->plane_descriptors[k].cursor.cursor_bpp / 8.0 / (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000));
@@ -7194,35 +7477,35 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out
old_ReadBandwidthChroma = mode_lib->ms.SwathWidthYSingleDPP[k] / 2 * math_ceil2(mode_lib->ms.BytePerPixelInDETC[k], 2.0) / (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)) * display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio / 2.0;
dml2_printf("DML::%s: k=%u, old_ReadBandwidthLuma = %f\n", __func__, k, old_ReadBandwidthLuma);
dml2_printf("DML::%s: k=%u, old_ReadBandwidthChroma = %f\n", __func__, k, old_ReadBandwidthChroma);
- dml2_printf("DML::%s: k=%u, ReadBandwidthLuma = %f\n", __func__, k, mode_lib->ms.SurfaceReadBandwidthLuma[k]);
- dml2_printf("DML::%s: k=%u, ReadBandwidthChroma = %f\n", __func__, k, mode_lib->ms.SurfaceReadBandwidthChroma[k]);
+ dml2_printf("DML::%s: k=%u, vactive_sw_bw_l = %f\n", __func__, k, mode_lib->ms.vactive_sw_bw_l[k]);
+ dml2_printf("DML::%s: k=%u, vactive_sw_bw_c = %f\n", __func__, k, mode_lib->ms.vactive_sw_bw_c[k]);
#endif
}
// Writeback bandwidth
for (k = 0; k < mode_lib->ms.num_active_planes; k++) {
- if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.enable == true && display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.pixel_format == dml2_444_64) {
- mode_lib->ms.WriteBandwidth[k] = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.output_height
- * display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.output_width
- / (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.input_height
+ if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.active_writebacks_per_stream > 0 && display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].pixel_format == dml2_444_64) {
+ mode_lib->ms.WriteBandwidth[k][0] = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].output_height
+ * display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].output_width
+ / (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].input_height
* display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total
/ ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)) * 8.0;
- } else if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.enable == true) {
- mode_lib->ms.WriteBandwidth[k] = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.output_height
- * display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.output_width
- / (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.input_height
+ } else if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.active_writebacks_per_stream > 0) {
+ mode_lib->ms.WriteBandwidth[k][0] = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].output_height
+ * display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].output_width
+ / (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].input_height
* display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total
/ ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)) * 4.0;
} else {
- mode_lib->ms.WriteBandwidth[k] = 0.0;
+ mode_lib->ms.WriteBandwidth[k][0] = 0.0;
}
}
/*Writeback Latency support check*/
mode_lib->ms.support.WritebackLatencySupport = true;
for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
- if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.enable == true &&
- (mode_lib->ms.WriteBandwidth[k] > mode_lib->ip.writeback_interface_buffer_size_kbytes * 1024 / ((double)mode_lib->soc.qos_parameters.writeback.base_latency_us))) {
+ if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.active_writebacks_per_stream > 0 &&
+ (mode_lib->ms.WriteBandwidth[k][0] > mode_lib->ip.writeback_interface_buffer_size_kbytes * 1024 / ((double)mode_lib->soc.qos_parameters.writeback.base_latency_us))) {
mode_lib->ms.support.WritebackLatencySupport = false;
}
}
@@ -7231,19 +7514,19 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out
/* Writeback Scale Ratio and Taps Support Check */
mode_lib->ms.support.WritebackScaleRatioAndTapsSupport = true;
for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
- if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.enable == true) {
- if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.h_ratio > mode_lib->ip.writeback_max_hscl_ratio
- || display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.v_ratio > mode_lib->ip.writeback_max_vscl_ratio
- || display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.h_ratio < mode_lib->ip.writeback_min_hscl_ratio
- || display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.v_ratio < mode_lib->ip.writeback_min_vscl_ratio
- || display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.h_taps > (unsigned int) mode_lib->ip.writeback_max_hscl_taps
- || display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.v_taps > (unsigned int) mode_lib->ip.writeback_max_vscl_taps
- || display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.h_ratio > (unsigned int)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.h_taps
- || display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.v_ratio > (unsigned int)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.v_taps
- || (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.h_taps > 2.0 && ((display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.h_taps % 2) == 1))) {
+ if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.active_writebacks_per_stream > 0) {
+ if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].h_ratio > mode_lib->ip.writeback_max_hscl_ratio
+ || display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].v_ratio > mode_lib->ip.writeback_max_vscl_ratio
+ || display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].h_ratio < mode_lib->ip.writeback_min_hscl_ratio
+ || display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].v_ratio < mode_lib->ip.writeback_min_vscl_ratio
+ || display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].h_taps > (unsigned int) mode_lib->ip.writeback_max_hscl_taps
+ || display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].v_taps > (unsigned int) mode_lib->ip.writeback_max_vscl_taps
+ || display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].h_ratio > (unsigned int)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].h_taps
+ || display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].v_ratio > (unsigned int)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].v_taps
+ || (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].h_taps > 2.0 && ((display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].h_taps % 2) == 1))) {
mode_lib->ms.support.WritebackScaleRatioAndTapsSupport = false;
}
- if (2.0 * display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.output_height * (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.v_taps - 1) * 57 > mode_lib->ip.writeback_line_buffer_buffer_size) {
+ if (2.0 * display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].output_height * (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].v_taps - 1) * 57 > mode_lib->ip.writeback_line_buffer_buffer_size) {
mode_lib->ms.support.WritebackScaleRatioAndTapsSupport = false;
}
}
@@ -7423,8 +7706,8 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out
CalculateSwathAndDETConfiguration_params->nomDETInKByte = mode_lib->ms.NomDETInKByte;
CalculateSwathAndDETConfiguration_params->ConfigReturnBufferSegmentSizeInkByte = mode_lib->ip.config_return_buffer_segment_size_in_kbytes;
CalculateSwathAndDETConfiguration_params->CompressedBufferSegmentSizeInkByte = mode_lib->ip.compressed_buffer_segment_size_in_kbytes;
- CalculateSwathAndDETConfiguration_params->ReadBandwidthLuma = mode_lib->ms.SurfaceReadBandwidthLuma;
- CalculateSwathAndDETConfiguration_params->ReadBandwidthChroma = mode_lib->ms.SurfaceReadBandwidthChroma;
+ CalculateSwathAndDETConfiguration_params->ReadBandwidthLuma = mode_lib->ms.vactive_sw_bw_l;
+ CalculateSwathAndDETConfiguration_params->ReadBandwidthChroma = mode_lib->ms.vactive_sw_bw_c;
CalculateSwathAndDETConfiguration_params->MaximumSwathWidthLuma = mode_lib->ms.MaximumSwathWidthLuma;
CalculateSwathAndDETConfiguration_params->MaximumSwathWidthChroma = mode_lib->ms.MaximumSwathWidthChroma;
CalculateSwathAndDETConfiguration_params->Read256BytesBlockHeightY = mode_lib->ms.Read256BlockHeightY;
@@ -7671,16 +7954,16 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out
//DISPCLK/DPPCLK
mode_lib->ms.WritebackRequiredDISPCLK = 0;
for (k = 0; k < mode_lib->ms.num_active_planes; ++k) {
- if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.enable) {
+ if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.active_writebacks_per_stream > 0) {
mode_lib->ms.WritebackRequiredDISPCLK = math_max2(mode_lib->ms.WritebackRequiredDISPCLK,
- CalculateWriteBackDISPCLK(display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.pixel_format,
+ CalculateWriteBackDISPCLK(display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].pixel_format,
((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000),
- display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.h_ratio,
- display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.v_ratio,
- display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.h_taps,
- display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.v_taps,
- display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.input_width,
- display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.output_width,
+ display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].h_ratio,
+ display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].v_ratio,
+ display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].h_taps,
+ display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].v_taps,
+ display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].input_width,
+ display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].output_width,
display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total,
mode_lib->ip.writeback_line_buffer_buffer_size));
}
@@ -7712,7 +7995,7 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out
if (!s->stream_visited[display_cfg->plane_descriptors[k].stream_index]) {
s->stream_visited[display_cfg->plane_descriptors[k].stream_index] = 1;
- if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.enable == true)
+ if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.active_writebacks_per_stream > 0)
s->TotalNumberOfActiveWriteback = s->TotalNumberOfActiveWriteback + 1;
s->TotalNumberOfActiveOTG = s->TotalNumberOfActiveOTG + 1;
@@ -8256,23 +8539,23 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out
mode_lib->ms.PSCL_FACTOR,
mode_lib->ms.PSCL_FACTOR_CHROMA,
mode_lib->ms.RequiredDPPCLK,
- mode_lib->ms.SurfaceReadBandwidthLuma,
- mode_lib->ms.SurfaceReadBandwidthChroma,
+ mode_lib->ms.vactive_sw_bw_l,
+ mode_lib->ms.vactive_sw_bw_c,
mode_lib->soc.return_bus_width_bytes,
/* Output */
&mode_lib->ms.dcfclk_deepsleep);
for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
- if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.enable == true) {
+ if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.active_writebacks_per_stream > 0) {
mode_lib->ms.WritebackDelayTime[k] = mode_lib->soc.qos_parameters.writeback.base_latency_us + CalculateWriteBackDelay(
- display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.pixel_format,
- display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.h_ratio,
- display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.v_ratio,
- display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.v_taps,
- display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.output_width,
- display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.output_height,
- display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.input_height,
+ display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].pixel_format,
+ display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].h_ratio,
+ display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].v_ratio,
+ display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].v_taps,
+ display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].output_width,
+ display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].output_height,
+ display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].input_height,
display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total) / mode_lib->ms.RequiredDISPCLK;
} else {
mode_lib->ms.WritebackDelayTime[k] = 0.0;
@@ -8349,7 +8632,7 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out
dml2_printf("DML::%s: mode_lib->ms.DCFCLK = %f\n", __func__, mode_lib->ms.DCFCLK);
dml2_printf("DML::%s: mode_lib->ms.FabricClock = %f\n", __func__, mode_lib->ms.FabricClock);
dml2_printf("DML::%s: mode_lib->ms.uclk_freq_mhz = %f\n", __func__, mode_lib->ms.uclk_freq_mhz);
- dml2_printf("DML::%s: urgent latency tolerance = %f\n", __func__, ((mode_lib->ip.rob_buffer_size_kbytes - mode_lib->ip.pixel_chunk_size_kbytes) * 1024 / (mode_lib->ms.DCFCLK * mode_lib->soc.return_bus_width_bytes)));
+ dml2_printf("DML::%s: urgent latency tolarance = %f\n", __func__, ((mode_lib->ip.rob_buffer_size_kbytes - mode_lib->ip.pixel_chunk_size_kbytes) * 1024 / (mode_lib->ms.DCFCLK * mode_lib->soc.return_bus_width_bytes)));
#endif
mode_lib->ms.support.OutstandingRequestsSupport = true;
@@ -8367,6 +8650,13 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out
+ mode_lib->soc.qos_parameters.qos_params.dcn4x.average_transport_distance_fclk_cycles / mode_lib->ms.FabricClock)
* (1 + mode_lib->soc.qos_parameters.qos_params.dcn4x.fabric_average_transport_latency_margin / 100.0);
+ mode_lib->ms.support.max_non_urgent_latency_us
+ = mode_lib->soc.qos_parameters.qos_params.dcn4x.per_uclk_dpm_params[mode_lib->ms.qos_param_index].maximum_latency_when_non_urgent_uclk_cycles
+ / mode_lib->ms.uclk_freq_mhz * (1 + mode_lib->soc.qos_parameters.qos_params.dcn4x.umc_max_latency_margin / 100.0)
+ + mode_lib->soc.qos_parameters.qos_params.dcn4x.mall_overhead_fclk_cycles / mode_lib->ms.FabricClock
+ + mode_lib->soc.qos_parameters.qos_params.dcn4x.max_round_trip_to_furthest_cs_fclk_cycles / mode_lib->ms.FabricClock
+ * (1 + mode_lib->soc.qos_parameters.qos_params.dcn4x.fabric_max_transport_latency_margin / 100.0);
+
for (k = 0; k < mode_lib->ms.num_active_planes; k++) {
if (mode_lib->soc.qos_parameters.qos_type == dml2_qos_param_type_dcn4x) {
@@ -8408,7 +8698,7 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out
}
memset(calculate_mcache_setting_params, 0, sizeof(struct dml2_core_calcs_calculate_mcache_setting_params));
- if (mode_lib->soc.mall_allocated_for_dcn_mbytes == 0 || mode_lib->ip.dcn_mrq_present) {
+ if (mode_lib->soc.mcache_size_bytes == 0 || mode_lib->ip.dcn_mrq_present) {
for (k = 0; k < mode_lib->ms.num_active_planes; k++) {
mode_lib->ms.mall_prefetch_sdp_overhead_factor[k] = 1.0;
mode_lib->ms.mall_prefetch_dram_overhead_factor[k] = 1.0;
@@ -8515,8 +8805,11 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out
display_cfg->hostvm_enable,
mode_lib->ms.MaxDCFCLK,
mode_lib->ms.MaxFabricClock,
+#ifdef DML_MODE_SUPPORT_USE_DPM_DRAM_BW
+ mode_lib->ms.dram_bw_mbps);
+#else
mode_lib->ms.max_dram_bw_mbps);
-
+#endif
// Average BW support check
calculate_avg_bandwidth_required(
@@ -8524,8 +8817,8 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out
// input
display_cfg,
mode_lib->ms.num_active_planes,
- mode_lib->ms.SurfaceReadBandwidthLuma,
- mode_lib->ms.SurfaceReadBandwidthChroma,
+ mode_lib->ms.vactive_sw_bw_l,
+ mode_lib->ms.vactive_sw_bw_c,
mode_lib->ms.cursor_bw,
mode_lib->ms.dcc_dram_bw_nom_overhead_factor_p0,
mode_lib->ms.dcc_dram_bw_nom_overhead_factor_p1,
@@ -8595,6 +8888,7 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out
calculate_tdlut_setting_params->tdlut_groups_per_2row_ub = &s->tdlut_groups_per_2row_ub[k];
calculate_tdlut_setting_params->tdlut_opt_time = &s->tdlut_opt_time[k];
calculate_tdlut_setting_params->tdlut_drain_time = &s->tdlut_drain_time[k];
+ calculate_tdlut_setting_params->tdlut_bytes_to_deliver = &s->tdlut_bytes_to_deliver[k];
calculate_tdlut_setting_params->tdlut_bytes_per_group = &s->tdlut_bytes_per_group[k];
calculate_tdlut_setting(&mode_lib->scratch, calculate_tdlut_setting_params);
@@ -8638,9 +8932,32 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out
&mode_lib->ms.ExtraLatency_sr,
&mode_lib->ms.ExtraLatencyPrefetch);
- {
+ for (k = 0; k < mode_lib->ms.num_active_planes; k++)
+ s->impacted_dst_y_pre[k] = 0;
+
+ s->recalc_prefetch_schedule = 0;
+ s->recalc_prefetch_done = 0;
+ do {
mode_lib->ms.support.PrefetchSupported = true;
+
for (k = 0; k < mode_lib->ms.num_active_planes; k++) {
+ s->line_times[k] = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000);
+ s->pixel_format[k] = display_cfg->plane_descriptors[k].pixel_format;
+
+ s->lb_source_lines_l[k] = get_num_lb_source_lines(mode_lib->ip.max_line_buffer_lines, mode_lib->ip.line_buffer_size_bits,
+ mode_lib->ms.NoOfDPP[k],
+ display_cfg->plane_descriptors[k].composition.viewport.plane0.width,
+ display_cfg->plane_descriptors[k].composition.viewport.plane0.height,
+ display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_ratio,
+ display_cfg->plane_descriptors[k].composition.rotation_angle);
+
+ s->lb_source_lines_c[k] = get_num_lb_source_lines(mode_lib->ip.max_line_buffer_lines, mode_lib->ip.line_buffer_size_bits,
+ mode_lib->ms.NoOfDPP[k],
+ display_cfg->plane_descriptors[k].composition.viewport.plane1.width,
+ display_cfg->plane_descriptors[k].composition.viewport.plane1.height,
+ display_cfg->plane_descriptors[k].composition.scaler_info.plane1.h_ratio,
+ display_cfg->plane_descriptors[k].composition.rotation_angle);
+
struct dml2_core_internal_DmlPipe *myPipe = &s->myPipe;
mode_lib->ms.TWait[k] = CalculateTWait(
@@ -8730,6 +9047,9 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out
CalculatePrefetchSchedule_params->mrq_present = mode_lib->ip.dcn_mrq_present;
CalculatePrefetchSchedule_params->meta_row_bytes = mode_lib->ms.meta_row_bytes[k];
CalculatePrefetchSchedule_params->mall_prefetch_sdp_overhead_factor = mode_lib->ms.mall_prefetch_sdp_overhead_factor[k];
+ CalculatePrefetchSchedule_params->impacted_dst_y_pre = s->impacted_dst_y_pre[k];
+ CalculatePrefetchSchedule_params->vactive_sw_bw_l = mode_lib->ms.vactive_sw_bw_l[k];
+ CalculatePrefetchSchedule_params->vactive_sw_bw_c = mode_lib->ms.vactive_sw_bw_c[k];
// output
CalculatePrefetchSchedule_params->DSTXAfterScaler = &s->DSTXAfterScaler[k];
@@ -8758,6 +9078,10 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out
CalculatePrefetchSchedule_params->VUpdateWidthPix = &s->dummy_integer[1];
CalculatePrefetchSchedule_params->VReadyOffsetPix = &s->dummy_integer[2];
CalculatePrefetchSchedule_params->prefetch_cursor_bw = &mode_lib->ms.prefetch_cursor_bw[k];
+ CalculatePrefetchSchedule_params->prefetch_sw_bytes = &s->prefetch_sw_bytes[k];
+ CalculatePrefetchSchedule_params->Tpre_rounded = &s->Tpre_rounded[k];
+ CalculatePrefetchSchedule_params->Tpre_oto = &s->Tpre_oto[k];
+ CalculatePrefetchSchedule_params->prefetch_swath_time_us = &s->prefetch_swath_time_us[k];
mode_lib->ms.NoTimeForPrefetch[k] = CalculatePrefetchSchedule(&mode_lib->scratch, CalculatePrefetchSchedule_params);
@@ -8766,6 +9090,27 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out
dml2_printf("DML::%s: k=%d, dst_y_per_row_vblank = %f\n", __func__, k, *CalculatePrefetchSchedule_params->dst_y_per_row_vblank);
} // for k num_planes
+ CalculateDCFCLKDeepSleepTdlut(
+ display_cfg,
+ mode_lib->ms.num_active_planes,
+ mode_lib->ms.BytePerPixelY,
+ mode_lib->ms.BytePerPixelC,
+ mode_lib->ms.SwathWidthY,
+ mode_lib->ms.SwathWidthC,
+ mode_lib->ms.NoOfDPP,
+ mode_lib->ms.PSCL_FACTOR,
+ mode_lib->ms.PSCL_FACTOR_CHROMA,
+ mode_lib->ms.RequiredDPPCLK,
+ mode_lib->ms.vactive_sw_bw_l,
+ mode_lib->ms.vactive_sw_bw_c,
+ mode_lib->soc.return_bus_width_bytes,
+ mode_lib->ms.RequiredDISPCLK,
+ s->tdlut_bytes_to_deliver,
+ s->prefetch_swath_time_us,
+
+ /* Output */
+ &mode_lib->ms.dcfclk_deepsleep);
+
for (k = 0; k < mode_lib->ms.num_active_planes; k++) {
if (mode_lib->ms.dst_y_prefetch[k] < 2.0
|| mode_lib->ms.LinesForVM[k] >= 32.0
@@ -8789,7 +9134,7 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out
}
mode_lib->ms.support.VRatioInPrefetchSupported = true;
- for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+ for (k = 0; k < mode_lib->ms.num_active_planes; k++) {
if (mode_lib->ms.VRatioPreY[k] > __DML2_CALCS_MAX_VRATIO_PRE__ ||
mode_lib->ms.VRatioPreC[k] > __DML2_CALCS_MAX_VRATIO_PRE__) {
mode_lib->ms.support.VRatioInPrefetchSupported = false;
@@ -8799,10 +9144,14 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out
}
}
+ mode_lib->ms.support.PrefetchSupported &= mode_lib->ms.support.VRatioInPrefetchSupported;
+
+ // By default, do not recalc prefetch schedule
+ s->recalc_prefetch_schedule = 0;
+
// Only do urg vs prefetch bandwidth check, flip schedule check, power saving feature support check IF the Prefetch Schedule Check is ok
if (mode_lib->ms.support.PrefetchSupported) {
- for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
- double line_time_us = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000);
+ for (k = 0; k < mode_lib->ms.num_active_planes; k++) {
// Calculate Urgent burst factor for prefetch
#ifdef __DML_VBA_DEBUG__
dml2_printf("DML::%s: k=%d, Calling CalculateUrgentBurstFactor (for prefetch)\n", __func__, k);
@@ -8815,7 +9164,7 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out
mode_lib->ms.swath_width_chroma_ub[k],
mode_lib->ms.SwathHeightY[k],
mode_lib->ms.SwathHeightC[k],
- line_time_us,
+ s->line_times[k],
mode_lib->ms.UrgLatency,
mode_lib->ms.VRatioPreY[k],
mode_lib->ms.VRatioPreC[k],
@@ -8852,8 +9201,8 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out
calculate_peak_bandwidth_params->mall_prefetch_sdp_overhead_factor = mode_lib->ms.mall_prefetch_sdp_overhead_factor;
calculate_peak_bandwidth_params->mall_prefetch_dram_overhead_factor = mode_lib->ms.mall_prefetch_dram_overhead_factor;
- calculate_peak_bandwidth_params->surface_read_bandwidth_l = mode_lib->ms.SurfaceReadBandwidthLuma;
- calculate_peak_bandwidth_params->surface_read_bandwidth_c = mode_lib->ms.SurfaceReadBandwidthChroma;
+ calculate_peak_bandwidth_params->surface_read_bandwidth_l = mode_lib->ms.vactive_sw_bw_l;
+ calculate_peak_bandwidth_params->surface_read_bandwidth_c = mode_lib->ms.vactive_sw_bw_c;
calculate_peak_bandwidth_params->prefetch_bandwidth_l = mode_lib->ms.RequiredPrefetchPixelDataBWLuma;
calculate_peak_bandwidth_params->prefetch_bandwidth_c = mode_lib->ms.RequiredPrefetchPixelDataBWChroma;
calculate_peak_bandwidth_params->excess_vactive_fill_bw_l = mode_lib->ms.excess_vactive_fill_bw_l;
@@ -8899,127 +9248,164 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out
}
}
+#ifdef DML_GLOBAL_PREFETCH_CHECK
+ if (mode_lib->ms.support.PrefetchSupported && mode_lib->ms.num_active_planes > 1 && s->recalc_prefetch_done == 0) {
+ CheckGlobalPrefetchAdmissibility_params->num_active_planes = mode_lib->ms.num_active_planes;
+ CheckGlobalPrefetchAdmissibility_params->pixel_format = s->pixel_format;
+ CheckGlobalPrefetchAdmissibility_params->chunk_bytes_l = mode_lib->ip.pixel_chunk_size_kbytes * 1024;
+ CheckGlobalPrefetchAdmissibility_params->chunk_bytes_c = mode_lib->ip.pixel_chunk_size_kbytes * 1024;
+ CheckGlobalPrefetchAdmissibility_params->lb_source_lines_l = s->lb_source_lines_l;
+ CheckGlobalPrefetchAdmissibility_params->lb_source_lines_c = s->lb_source_lines_c;
+ CheckGlobalPrefetchAdmissibility_params->swath_height_l = mode_lib->ms.SwathHeightY;
+ CheckGlobalPrefetchAdmissibility_params->swath_height_c = mode_lib->ms.SwathHeightC;
+ CheckGlobalPrefetchAdmissibility_params->rob_buffer_size_kbytes = mode_lib->ip.rob_buffer_size_kbytes;
+ CheckGlobalPrefetchAdmissibility_params->compressed_buffer_size_kbytes = mode_lib->ms.CompressedBufferSizeInkByte;
+ CheckGlobalPrefetchAdmissibility_params->detile_buffer_size_bytes_l = mode_lib->ms.DETBufferSizeY;
+ CheckGlobalPrefetchAdmissibility_params->detile_buffer_size_bytes_c = mode_lib->ms.DETBufferSizeC;
+ CheckGlobalPrefetchAdmissibility_params->full_swath_bytes_l = s->full_swath_bytes_l;
+ CheckGlobalPrefetchAdmissibility_params->full_swath_bytes_c = s->full_swath_bytes_c;
+ CheckGlobalPrefetchAdmissibility_params->prefetch_sw_bytes = s->prefetch_sw_bytes;
+ CheckGlobalPrefetchAdmissibility_params->Tpre_rounded = s->Tpre_rounded;
+ CheckGlobalPrefetchAdmissibility_params->Tpre_oto = s->Tpre_oto;
+ CheckGlobalPrefetchAdmissibility_params->estimated_urg_bandwidth_required_mbps = mode_lib->ms.support.urg_bandwidth_required[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_sdp];
+ CheckGlobalPrefetchAdmissibility_params->line_time = s->line_times;
+ CheckGlobalPrefetchAdmissibility_params->dst_y_prefetch = mode_lib->ms.dst_y_prefetch;
+ if (CheckGlobalPrefetchAdmissibility_params->estimated_urg_bandwidth_required_mbps < 10 * 1024)
+ CheckGlobalPrefetchAdmissibility_params->estimated_urg_bandwidth_required_mbps = 10 * 1024;
+
+ CheckGlobalPrefetchAdmissibility_params->estimated_dcfclk_mhz = (CheckGlobalPrefetchAdmissibility_params->estimated_urg_bandwidth_required_mbps / (double) mode_lib->soc.return_bus_width_bytes) /
+ ((double)mode_lib->soc.qos_parameters.derate_table.system_active_urgent.dcfclk_derate_percent / 100.0);
+
+ // if recalc_prefetch_schedule is set, recalculate the prefetch schedule with the new impacted_Tpre, prefetch should be possible
+ CheckGlobalPrefetchAdmissibility_params->recalc_prefetch_schedule = &s->recalc_prefetch_schedule;
+ CheckGlobalPrefetchAdmissibility_params->impacted_dst_y_pre = s->impacted_dst_y_pre;
+ mode_lib->ms.support.PrefetchSupported = CheckGlobalPrefetchAdmissibility(&mode_lib->scratch, CheckGlobalPrefetchAdmissibility_params);
+ s->recalc_prefetch_done = 1;
+ s->recalc_prefetch_schedule = 1;
+ }
+#endif
+ } // prefetch schedule ok, do urg bw and flip schedule
+ } while (s->recalc_prefetch_schedule);
- // Both prefetch schedule and BW okay
- if (mode_lib->ms.support.PrefetchSupported == true && mode_lib->ms.support.VRatioInPrefetchSupported == true) {
- mode_lib->ms.BandwidthAvailableForImmediateFlip =
- get_bandwidth_available_for_immediate_flip(
- dml2_core_internal_soc_state_sys_active,
- mode_lib->ms.support.urg_bandwidth_required_qual, // no flip
- mode_lib->ms.support.urg_bandwidth_available);
-
- mode_lib->ms.TotImmediateFlipBytes = 0;
- for (k = 0; k < mode_lib->ms.num_active_planes; k++) {
- if (display_cfg->plane_descriptors[k].immediate_flip) {
- s->per_pipe_flip_bytes[k] = get_pipe_flip_bytes(
- s->HostVMInefficiencyFactor,
- mode_lib->ms.vm_bytes[k],
- mode_lib->ms.DPTEBytesPerRow[k],
- mode_lib->ms.meta_row_bytes[k]);
- } else {
- s->per_pipe_flip_bytes[k] = 0;
- }
- mode_lib->ms.TotImmediateFlipBytes += s->per_pipe_flip_bytes[k] * mode_lib->ms.NoOfDPP[k];
+ // Flip Schedule
+ // Both prefetch schedule and BW okay
+ if (mode_lib->ms.support.PrefetchSupported == true) {
+ mode_lib->ms.BandwidthAvailableForImmediateFlip =
+ get_bandwidth_available_for_immediate_flip(
+ dml2_core_internal_soc_state_sys_active,
+ mode_lib->ms.support.urg_bandwidth_required_qual, // no flip
+ mode_lib->ms.support.urg_bandwidth_available);
- }
+ mode_lib->ms.TotImmediateFlipBytes = 0;
+ for (k = 0; k < mode_lib->ms.num_active_planes; k++) {
+ if (display_cfg->plane_descriptors[k].immediate_flip) {
+ s->per_pipe_flip_bytes[k] = get_pipe_flip_bytes(
+ s->HostVMInefficiencyFactor,
+ mode_lib->ms.vm_bytes[k],
+ mode_lib->ms.DPTEBytesPerRow[k],
+ mode_lib->ms.meta_row_bytes[k]);
+ } else {
+ s->per_pipe_flip_bytes[k] = 0;
+ }
+ mode_lib->ms.TotImmediateFlipBytes += s->per_pipe_flip_bytes[k] * mode_lib->ms.NoOfDPP[k];
- for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
- CalculateFlipSchedule(
- &mode_lib->scratch,
- display_cfg->plane_descriptors[k].immediate_flip,
- 1, // use_lb_flip_bw
- s->HostVMInefficiencyFactor,
- s->Tvm_trips_flip[k],
- s->Tr0_trips_flip[k],
- s->Tvm_trips_flip_rounded[k],
- s->Tr0_trips_flip_rounded[k],
- display_cfg->gpuvm_enable,
- mode_lib->ms.vm_bytes[k],
- mode_lib->ms.DPTEBytesPerRow[k],
- mode_lib->ms.BandwidthAvailableForImmediateFlip,
- mode_lib->ms.TotImmediateFlipBytes,
- display_cfg->plane_descriptors[k].pixel_format,
- (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)),
- display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio,
- display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio,
- mode_lib->ms.Tno_bw_flip[k],
- mode_lib->ms.dpte_row_height[k],
- mode_lib->ms.dpte_row_height_chroma[k],
- mode_lib->ms.use_one_row_for_frame_flip[k],
- mode_lib->ip.max_flip_time_us,
- mode_lib->ip.max_flip_time_lines,
- s->per_pipe_flip_bytes[k],
- mode_lib->ms.meta_row_bytes[k],
- s->meta_row_height_luma[k],
- s->meta_row_height_chroma[k],
- mode_lib->ip.dcn_mrq_present && display_cfg->plane_descriptors[k].surface.dcc.enable,
-
- /* Output */
- &mode_lib->ms.dst_y_per_vm_flip[k],
- &mode_lib->ms.dst_y_per_row_flip[k],
- &mode_lib->ms.final_flip_bw[k],
- &mode_lib->ms.ImmediateFlipSupportedForPipe[k]);
- }
+ }
- calculate_peak_bandwidth_params->urg_vactive_bandwidth_required = s->dummy_bw;
- calculate_peak_bandwidth_params->urg_bandwidth_required = mode_lib->ms.support.urg_bandwidth_required_flip;
- calculate_peak_bandwidth_params->urg_bandwidth_required_qual = s->dummy_bw;
- calculate_peak_bandwidth_params->non_urg_bandwidth_required = mode_lib->ms.support.non_urg_bandwidth_required_flip;
- calculate_peak_bandwidth_params->surface_avg_vactive_required_bw = s->surface_dummy_bw;
- calculate_peak_bandwidth_params->surface_peak_required_bw = mode_lib->ms.surface_peak_required_bw;
-
- calculate_peak_bandwidth_params->display_cfg = display_cfg;
- calculate_peak_bandwidth_params->inc_flip_bw = 1;
- calculate_peak_bandwidth_params->num_active_planes = mode_lib->ms.num_active_planes;
- calculate_peak_bandwidth_params->num_of_dpp = mode_lib->ms.NoOfDPP;
- calculate_peak_bandwidth_params->dcc_dram_bw_nom_overhead_factor_p0 = mode_lib->ms.dcc_dram_bw_nom_overhead_factor_p0;
- calculate_peak_bandwidth_params->dcc_dram_bw_nom_overhead_factor_p1 = mode_lib->ms.dcc_dram_bw_nom_overhead_factor_p1;
- calculate_peak_bandwidth_params->dcc_dram_bw_pref_overhead_factor_p0 = mode_lib->ms.dcc_dram_bw_pref_overhead_factor_p0;
- calculate_peak_bandwidth_params->dcc_dram_bw_pref_overhead_factor_p1 = mode_lib->ms.dcc_dram_bw_pref_overhead_factor_p1;
- calculate_peak_bandwidth_params->mall_prefetch_sdp_overhead_factor = mode_lib->ms.mall_prefetch_sdp_overhead_factor;
- calculate_peak_bandwidth_params->mall_prefetch_dram_overhead_factor = mode_lib->ms.mall_prefetch_dram_overhead_factor;
-
- calculate_peak_bandwidth_params->surface_read_bandwidth_l = mode_lib->ms.SurfaceReadBandwidthLuma;
- calculate_peak_bandwidth_params->surface_read_bandwidth_c = mode_lib->ms.SurfaceReadBandwidthChroma;
- calculate_peak_bandwidth_params->prefetch_bandwidth_l = mode_lib->ms.RequiredPrefetchPixelDataBWLuma;
- calculate_peak_bandwidth_params->prefetch_bandwidth_c = mode_lib->ms.RequiredPrefetchPixelDataBWChroma;
- calculate_peak_bandwidth_params->excess_vactive_fill_bw_l = mode_lib->ms.excess_vactive_fill_bw_l;
- calculate_peak_bandwidth_params->excess_vactive_fill_bw_c = mode_lib->ms.excess_vactive_fill_bw_c;
- calculate_peak_bandwidth_params->cursor_bw = mode_lib->ms.cursor_bw;
- calculate_peak_bandwidth_params->dpte_row_bw = mode_lib->ms.dpte_row_bw;
- calculate_peak_bandwidth_params->meta_row_bw = mode_lib->ms.meta_row_bw;
- calculate_peak_bandwidth_params->prefetch_cursor_bw = mode_lib->ms.prefetch_cursor_bw;
- calculate_peak_bandwidth_params->prefetch_vmrow_bw = mode_lib->ms.prefetch_vmrow_bw;
- calculate_peak_bandwidth_params->flip_bw = mode_lib->ms.final_flip_bw;
- calculate_peak_bandwidth_params->urgent_burst_factor_l = mode_lib->ms.UrgentBurstFactorLuma;
- calculate_peak_bandwidth_params->urgent_burst_factor_c = mode_lib->ms.UrgentBurstFactorChroma;
- calculate_peak_bandwidth_params->urgent_burst_factor_cursor = mode_lib->ms.UrgentBurstFactorCursor;
- calculate_peak_bandwidth_params->urgent_burst_factor_prefetch_l = mode_lib->ms.UrgentBurstFactorLumaPre;
- calculate_peak_bandwidth_params->urgent_burst_factor_prefetch_c = mode_lib->ms.UrgentBurstFactorChromaPre;
- calculate_peak_bandwidth_params->urgent_burst_factor_prefetch_cursor = mode_lib->ms.UrgentBurstFactorCursorPre;
-
- calculate_peak_bandwidth_required(
- &mode_lib->scratch,
- calculate_peak_bandwidth_params);
-
- calculate_immediate_flip_bandwidth_support(
- &s->dummy_single[0], // double* frac_urg_bandwidth_flip
- &mode_lib->ms.support.ImmediateFlipSupport,
-
- dml2_core_internal_soc_state_sys_active,
- mode_lib->ms.support.urg_bandwidth_required_flip,
- mode_lib->ms.support.non_urg_bandwidth_required_flip,
- mode_lib->ms.support.urg_bandwidth_available);
-
- for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
- if (display_cfg->plane_descriptors[k].immediate_flip == true && mode_lib->ms.ImmediateFlipSupportedForPipe[k] == false)
- mode_lib->ms.support.ImmediateFlipSupport = false;
- }
+ for (k = 0; k < mode_lib->ms.num_active_planes; k++) {
+ CalculateFlipSchedule(
+ &mode_lib->scratch,
+ display_cfg->plane_descriptors[k].immediate_flip,
+ 1, // use_lb_flip_bw
+ s->HostVMInefficiencyFactor,
+ s->Tvm_trips_flip[k],
+ s->Tr0_trips_flip[k],
+ s->Tvm_trips_flip_rounded[k],
+ s->Tr0_trips_flip_rounded[k],
+ display_cfg->gpuvm_enable,
+ mode_lib->ms.vm_bytes[k],
+ mode_lib->ms.DPTEBytesPerRow[k],
+ mode_lib->ms.BandwidthAvailableForImmediateFlip,
+ mode_lib->ms.TotImmediateFlipBytes,
+ display_cfg->plane_descriptors[k].pixel_format,
+ (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)),
+ display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio,
+ display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio,
+ mode_lib->ms.Tno_bw_flip[k],
+ mode_lib->ms.dpte_row_height[k],
+ mode_lib->ms.dpte_row_height_chroma[k],
+ mode_lib->ms.use_one_row_for_frame_flip[k],
+ mode_lib->ip.max_flip_time_us,
+ mode_lib->ip.max_flip_time_lines,
+ s->per_pipe_flip_bytes[k],
+ mode_lib->ms.meta_row_bytes[k],
+ s->meta_row_height_luma[k],
+ s->meta_row_height_chroma[k],
+ mode_lib->ip.dcn_mrq_present && display_cfg->plane_descriptors[k].surface.dcc.enable,
+
+ /* Output */
+ &mode_lib->ms.dst_y_per_vm_flip[k],
+ &mode_lib->ms.dst_y_per_row_flip[k],
+ &mode_lib->ms.final_flip_bw[k],
+ &mode_lib->ms.ImmediateFlipSupportedForPipe[k]);
+ }
+
+ calculate_peak_bandwidth_params->urg_vactive_bandwidth_required = s->dummy_bw;
+ calculate_peak_bandwidth_params->urg_bandwidth_required = mode_lib->ms.support.urg_bandwidth_required_flip;
+ calculate_peak_bandwidth_params->urg_bandwidth_required_qual = s->dummy_bw;
+ calculate_peak_bandwidth_params->non_urg_bandwidth_required = mode_lib->ms.support.non_urg_bandwidth_required_flip;
+ calculate_peak_bandwidth_params->surface_avg_vactive_required_bw = s->surface_dummy_bw;
+ calculate_peak_bandwidth_params->surface_peak_required_bw = mode_lib->ms.surface_peak_required_bw;
+
+ calculate_peak_bandwidth_params->display_cfg = display_cfg;
+ calculate_peak_bandwidth_params->inc_flip_bw = 1;
+ calculate_peak_bandwidth_params->num_active_planes = mode_lib->ms.num_active_planes;
+ calculate_peak_bandwidth_params->num_of_dpp = mode_lib->ms.NoOfDPP;
+ calculate_peak_bandwidth_params->dcc_dram_bw_nom_overhead_factor_p0 = mode_lib->ms.dcc_dram_bw_nom_overhead_factor_p0;
+ calculate_peak_bandwidth_params->dcc_dram_bw_nom_overhead_factor_p1 = mode_lib->ms.dcc_dram_bw_nom_overhead_factor_p1;
+ calculate_peak_bandwidth_params->dcc_dram_bw_pref_overhead_factor_p0 = mode_lib->ms.dcc_dram_bw_pref_overhead_factor_p0;
+ calculate_peak_bandwidth_params->dcc_dram_bw_pref_overhead_factor_p1 = mode_lib->ms.dcc_dram_bw_pref_overhead_factor_p1;
+ calculate_peak_bandwidth_params->mall_prefetch_sdp_overhead_factor = mode_lib->ms.mall_prefetch_sdp_overhead_factor;
+ calculate_peak_bandwidth_params->mall_prefetch_dram_overhead_factor = mode_lib->ms.mall_prefetch_dram_overhead_factor;
+
+ calculate_peak_bandwidth_params->surface_read_bandwidth_l = mode_lib->ms.vactive_sw_bw_l;
+ calculate_peak_bandwidth_params->surface_read_bandwidth_c = mode_lib->ms.vactive_sw_bw_c;
+ calculate_peak_bandwidth_params->prefetch_bandwidth_l = mode_lib->ms.RequiredPrefetchPixelDataBWLuma;
+ calculate_peak_bandwidth_params->prefetch_bandwidth_c = mode_lib->ms.RequiredPrefetchPixelDataBWChroma;
+ calculate_peak_bandwidth_params->excess_vactive_fill_bw_l = mode_lib->ms.excess_vactive_fill_bw_l;
+ calculate_peak_bandwidth_params->excess_vactive_fill_bw_c = mode_lib->ms.excess_vactive_fill_bw_c;
+ calculate_peak_bandwidth_params->cursor_bw = mode_lib->ms.cursor_bw;
+ calculate_peak_bandwidth_params->dpte_row_bw = mode_lib->ms.dpte_row_bw;
+ calculate_peak_bandwidth_params->meta_row_bw = mode_lib->ms.meta_row_bw;
+ calculate_peak_bandwidth_params->prefetch_cursor_bw = mode_lib->ms.prefetch_cursor_bw;
+ calculate_peak_bandwidth_params->prefetch_vmrow_bw = mode_lib->ms.prefetch_vmrow_bw;
+ calculate_peak_bandwidth_params->flip_bw = mode_lib->ms.final_flip_bw;
+ calculate_peak_bandwidth_params->urgent_burst_factor_l = mode_lib->ms.UrgentBurstFactorLuma;
+ calculate_peak_bandwidth_params->urgent_burst_factor_c = mode_lib->ms.UrgentBurstFactorChroma;
+ calculate_peak_bandwidth_params->urgent_burst_factor_cursor = mode_lib->ms.UrgentBurstFactorCursor;
+ calculate_peak_bandwidth_params->urgent_burst_factor_prefetch_l = mode_lib->ms.UrgentBurstFactorLumaPre;
+ calculate_peak_bandwidth_params->urgent_burst_factor_prefetch_c = mode_lib->ms.UrgentBurstFactorChromaPre;
+ calculate_peak_bandwidth_params->urgent_burst_factor_prefetch_cursor = mode_lib->ms.UrgentBurstFactorCursorPre;
+
+ calculate_peak_bandwidth_required(
+ &mode_lib->scratch,
+ calculate_peak_bandwidth_params);
- } else { // if prefetch not support, assume iflip is not supported too
+ calculate_immediate_flip_bandwidth_support(
+ &s->dummy_single[0], // double* frac_urg_bandwidth_flip
+ &mode_lib->ms.support.ImmediateFlipSupport,
+
+ dml2_core_internal_soc_state_sys_active,
+ mode_lib->ms.support.urg_bandwidth_required_flip,
+ mode_lib->ms.support.non_urg_bandwidth_required_flip,
+ mode_lib->ms.support.urg_bandwidth_available);
+
+ for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+ if (display_cfg->plane_descriptors[k].immediate_flip == true && mode_lib->ms.ImmediateFlipSupportedForPipe[k] == false)
mode_lib->ms.support.ImmediateFlipSupport = false;
- }
- } // prefetch schedule
+ }
+
+ } else { // if prefetch not support, assume iflip is not supported too
+ mode_lib->ms.support.ImmediateFlipSupport = false;
}
s->mSOCParameters.UrgentLatency = mode_lib->ms.UrgLatency;
@@ -9116,8 +9502,8 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out
s->pstate_bytes_required_c,
mode_lib->ms.dcc_dram_bw_nom_overhead_factor_p0,
mode_lib->ms.dcc_dram_bw_nom_overhead_factor_p1,
- mode_lib->ms.SurfaceReadBandwidthLuma,
- mode_lib->ms.SurfaceReadBandwidthChroma,
+ mode_lib->ms.vactive_sw_bw_l,
+ mode_lib->ms.vactive_sw_bw_c,
mode_lib->ms.surface_avg_vactive_required_bw,
mode_lib->ms.surface_peak_required_bw,
/* outputs */
@@ -9187,12 +9573,12 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out
dml2_printf("DML::%s: ModeSupport = %u\n", __func__, mode_lib->ms.support.ModeSupport);
dml2_printf("DML::%s: ImmediateFlipSupport = %u\n", __func__, mode_lib->ms.support.ImmediateFlipSupport);
- for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+ for (k = 0; k < mode_lib->ms.num_active_planes; k++) {
mode_lib->ms.support.MPCCombineEnable[k] = mode_lib->ms.MPCCombine[k];
mode_lib->ms.support.DPPPerSurface[k] = mode_lib->ms.NoOfDPP[k];
}
- for (k = 0; k <= mode_lib->ms.num_active_planes - 1; k++) {
+ for (k = 0; k < mode_lib->ms.num_active_planes; k++) {
mode_lib->ms.support.ODMMode[k] = mode_lib->ms.ODMMode[k];
mode_lib->ms.support.DSCEnabled[k] = mode_lib->ms.RequiresDSC[k];
mode_lib->ms.support.FECEnabled[k] = mode_lib->ms.RequiresFEC[k];
@@ -9229,7 +9615,7 @@ unsigned int dml2_core_calcs_mode_support_ex(struct dml2_core_calcs_mode_support
dml2_printf("DML::%s: is_mode_support = %u (min_clk_index=%d)\n", __func__, result, in_out_params->min_clk_index);
for (unsigned int k = 0; k < in_out_params->in_display_cfg->num_planes; k++)
- dml2_printf("DML::%s: plane_%d: reserved_vblank_time_ns = %u\n", __func__, k, in_out_params->in_display_cfg->plane_descriptors[k].overrides.reserved_vblank_time_ns);
+ dml2_printf("DML::%s: plane_%d: reserved_vblank_time_ns = %u\n", __func__, k, in_out_params->in_display_cfg->plane_descriptors[k].overrides.reserved_vblank_time_ns);
dml2_printf("DML::%s: ------------- DONE ----------\n", __func__);
@@ -9882,7 +10268,7 @@ static void CalculateStutterEfficiency(struct dml2_core_internal_scratch *scratc
if (!dml_is_phantom_pipe(&p->display_cfg->plane_descriptors[k])) {
if (!l->stream_visited[p->display_cfg->plane_descriptors[k].stream_index]) {
- if (p->display_cfg->stream_descriptors[k].writeback.enable)
+ if (p->display_cfg->stream_descriptors[k].writeback.active_writebacks_per_stream > 0)
l->TotalActiveWriteback = l->TotalActiveWriteback + 1;
if (TotalNumberOfActiveOTG == 0) { // first otg
@@ -9984,6 +10370,7 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex
struct dml2_core_calcs_CalculateSwathAndDETConfiguration_params *CalculateSwathAndDETConfiguration_params = &mode_lib->scratch.CalculateSwathAndDETConfiguration_params;
struct dml2_core_calcs_CalculateStutterEfficiency_params *CalculateStutterEfficiency_params = &mode_lib->scratch.CalculateStutterEfficiency_params;
struct dml2_core_calcs_CalculatePrefetchSchedule_params *CalculatePrefetchSchedule_params = &mode_lib->scratch.CalculatePrefetchSchedule_params;
+ struct dml2_core_calcs_CheckGlobalPrefetchAdmissibility_params *CheckGlobalPrefetchAdmissibility_params = &mode_lib->scratch.CheckGlobalPrefetchAdmissibility_params;
struct dml2_core_calcs_calculate_mcache_setting_params *calculate_mcache_setting_params = &mode_lib->scratch.calculate_mcache_setting_params;
struct dml2_core_calcs_calculate_tdlut_setting_params *calculate_tdlut_setting_params = &mode_lib->scratch.calculate_tdlut_setting_params;
struct dml2_core_shared_CalculateMetaAndPTETimes_params *CalculateMetaAndPTETimes_params = &mode_lib->scratch.CalculateMetaAndPTETimes_params;
@@ -10075,12 +10462,6 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex
dml2_assert(s->SOCCLK > 0);
#ifdef __DML_VBA_DEBUG__
- // dml2_printf_dml_display_cfg_timing(&display_cfg->timing, s->num_active_planes);
- // dml2_printf_dml_display_cfg_plane(&display_cfg->plane, s->num_active_planes);
- // dml2_printf_dml_display_cfg_surface(&display_cfg->surface, s->num_active_planes);
- // dml2_printf_dml_display_cfg_output(&display_cfg->output, s->num_active_planes);
- // dml2_printf_dml_display_cfg_hw_resource(&display_cfg->hw, s->num_active_planes);
-
dml2_printf("DML::%s: num_active_planes = %u\n", __func__, s->num_active_planes);
dml2_printf("DML::%s: num_active_pipes = %u\n", __func__, mode_lib->mp.num_active_pipes);
dml2_printf("DML::%s: Dcfclk = %f\n", __func__, mode_lib->mp.Dcfclk);
@@ -10198,10 +10579,10 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex
for (k = 0; k < s->num_active_planes; ++k) {
mode_lib->mp.cursor_bw[k] = display_cfg->plane_descriptors[k].cursor.num_cursors * display_cfg->plane_descriptors[k].cursor.cursor_width * display_cfg->plane_descriptors[k].cursor.cursor_bpp / 8.0 /
((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000));
- mode_lib->mp.SurfaceReadBandwidthLuma[k] = mode_lib->mp.SwathWidthSingleDPPY[k] * mode_lib->mp.BytePerPixelY[k] / (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)) * display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio;
- mode_lib->mp.SurfaceReadBandwidthChroma[k] = mode_lib->mp.SwathWidthSingleDPPC[k] * mode_lib->mp.BytePerPixelC[k] / (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)) * display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio;
- dml2_printf("DML::%s: ReadBandwidthSurfaceLuma[%i] = %fBps\n", __func__, k, mode_lib->mp.SurfaceReadBandwidthLuma[k]);
- dml2_printf("DML::%s: ReadBandwidthSurfaceChroma[%i] = %fBps\n", __func__, k, mode_lib->mp.SurfaceReadBandwidthChroma[k]);
+ mode_lib->mp.vactive_sw_bw_l[k] = mode_lib->mp.SwathWidthSingleDPPY[k] * mode_lib->mp.BytePerPixelY[k] / (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)) * display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio;
+ mode_lib->mp.vactive_sw_bw_c[k] = mode_lib->mp.SwathWidthSingleDPPC[k] * mode_lib->mp.BytePerPixelC[k] / (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)) * display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio;
+ dml2_printf("DML::%s: vactive_sw_bw_l[%i] = %fBps\n", __func__, k, mode_lib->mp.vactive_sw_bw_l[k]);
+ dml2_printf("DML::%s: vactive_sw_bw_c[%i] = %fBps\n", __func__, k, mode_lib->mp.vactive_sw_bw_c[k]);
}
CalculateSwathAndDETConfiguration_params->display_cfg = display_cfg;
@@ -10217,8 +10598,8 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex
CalculateSwathAndDETConfiguration_params->nomDETInKByte = s->NomDETInKByte;
CalculateSwathAndDETConfiguration_params->ConfigReturnBufferSegmentSizeInkByte = mode_lib->ip.config_return_buffer_segment_size_in_kbytes;
CalculateSwathAndDETConfiguration_params->CompressedBufferSegmentSizeInkByte = mode_lib->ip.compressed_buffer_segment_size_in_kbytes;
- CalculateSwathAndDETConfiguration_params->ReadBandwidthLuma = mode_lib->mp.SurfaceReadBandwidthLuma;
- CalculateSwathAndDETConfiguration_params->ReadBandwidthChroma = mode_lib->mp.SurfaceReadBandwidthChroma;
+ CalculateSwathAndDETConfiguration_params->ReadBandwidthLuma = mode_lib->mp.vactive_sw_bw_l;
+ CalculateSwathAndDETConfiguration_params->ReadBandwidthChroma = mode_lib->mp.vactive_sw_bw_c;
CalculateSwathAndDETConfiguration_params->MaximumSwathWidthLuma = s->dummy_single_array[0];
CalculateSwathAndDETConfiguration_params->MaximumSwathWidthChroma = s->dummy_single_array[1];
CalculateSwathAndDETConfiguration_params->Read256BytesBlockHeightY = mode_lib->mp.Read256BlockHeightY;
@@ -10539,8 +10920,8 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex
calculate_tdlut_setting_params->tdlut_groups_per_2row_ub = &s->tdlut_groups_per_2row_ub[k];
calculate_tdlut_setting_params->tdlut_opt_time = &s->tdlut_opt_time[k];
calculate_tdlut_setting_params->tdlut_drain_time = &s->tdlut_drain_time[k];
+ calculate_tdlut_setting_params->tdlut_bytes_to_deliver = &s->tdlut_bytes_to_deliver[k];
calculate_tdlut_setting_params->tdlut_bytes_per_group = &s->tdlut_bytes_per_group[k];
-
calculate_tdlut_setting(&mode_lib->scratch, calculate_tdlut_setting_params);
}
@@ -10583,17 +10964,17 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex
mode_lib->mp.TCalc = 24.0 / mode_lib->mp.DCFCLKDeepSleep;
for (k = 0; k < s->num_active_planes; ++k) {
- if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.enable == true) {
+ if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.active_writebacks_per_stream > 0) {
mode_lib->mp.WritebackDelay[k] =
mode_lib->soc.qos_parameters.writeback.base_latency_us
+ CalculateWriteBackDelay(
- display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.pixel_format,
- display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.h_ratio,
- display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.v_ratio,
- display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.v_taps,
- display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.output_width,
- display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.output_height,
- display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.input_height,
+ display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].pixel_format,
+ display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].h_ratio,
+ display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].v_ratio,
+ display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].v_taps,
+ display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].output_width,
+ display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].output_height,
+ display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.writeback_stream[0].input_height,
display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total) / mode_lib->mp.Dispclk;
} else
mode_lib->mp.WritebackDelay[k] = 0;
@@ -10679,10 +11060,25 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex
for (k = 0; k < s->num_active_planes; ++k) {
bool cursor_not_enough_urgent_latency_hiding = 0;
- double line_time_us = 0.0;
-
- line_time_us = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total /
+ s->line_times[k] = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total /
((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000);
+
+ s->pixel_format[k] = display_cfg->plane_descriptors[k].pixel_format;
+
+ s->lb_source_lines_l[k] = get_num_lb_source_lines(mode_lib->ip.max_line_buffer_lines, mode_lib->ip.line_buffer_size_bits,
+ mode_lib->mp.NoOfDPP[k],
+ display_cfg->plane_descriptors[k].composition.viewport.plane0.width,
+ display_cfg->plane_descriptors[k].composition.viewport.plane0.height,
+ display_cfg->plane_descriptors[k].composition.scaler_info.plane0.h_ratio,
+ display_cfg->plane_descriptors[k].composition.rotation_angle);
+
+ s->lb_source_lines_c[k] = get_num_lb_source_lines(mode_lib->ip.max_line_buffer_lines, mode_lib->ip.line_buffer_size_bits,
+ mode_lib->mp.NoOfDPP[k],
+ display_cfg->plane_descriptors[k].composition.viewport.plane1.width,
+ display_cfg->plane_descriptors[k].composition.viewport.plane1.height,
+ display_cfg->plane_descriptors[k].composition.scaler_info.plane1.h_ratio,
+ display_cfg->plane_descriptors[k].composition.rotation_angle);
+
if (display_cfg->plane_descriptors[k].cursor.num_cursors > 0) {
calculate_cursor_req_attributes(
display_cfg->plane_descriptors[k].cursor.cursor_width,
@@ -10699,7 +11095,7 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex
display_cfg->plane_descriptors[k].cursor.cursor_width,
s->cursor_bytes_per_chunk[k],
s->cursor_lines_per_chunk[k],
- line_time_us,
+ s->line_times[k],
mode_lib->mp.UrgentLatency,
// output
@@ -10714,7 +11110,7 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex
mode_lib->mp.swath_width_chroma_ub[k],
mode_lib->mp.SwathHeightY[k],
mode_lib->mp.SwathHeightC[k],
- line_time_us,
+ s->line_times[k],
mode_lib->mp.UrgentLatency,
display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio,
display_cfg->plane_descriptors[k].composition.scaler_info.plane1.v_ratio,
@@ -10752,6 +11148,35 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex
dml2_printf("DML::%s: immediate_flip_required = %u\n", __func__, s->immediate_flip_required);
#endif
+ if (s->num_active_planes > 1) {
+ CheckGlobalPrefetchAdmissibility_params->num_active_planes = s->num_active_planes;
+ CheckGlobalPrefetchAdmissibility_params->pixel_format = s->pixel_format;
+ CheckGlobalPrefetchAdmissibility_params->chunk_bytes_l = mode_lib->ip.pixel_chunk_size_kbytes * 1024;
+ CheckGlobalPrefetchAdmissibility_params->chunk_bytes_c = mode_lib->ip.pixel_chunk_size_kbytes * 1024;
+ CheckGlobalPrefetchAdmissibility_params->lb_source_lines_l = s->lb_source_lines_l;
+ CheckGlobalPrefetchAdmissibility_params->lb_source_lines_c = s->lb_source_lines_c;
+ CheckGlobalPrefetchAdmissibility_params->swath_height_l = mode_lib->mp.SwathHeightY;
+ CheckGlobalPrefetchAdmissibility_params->swath_height_c = mode_lib->mp.SwathHeightC;
+ CheckGlobalPrefetchAdmissibility_params->rob_buffer_size_kbytes = mode_lib->ip.rob_buffer_size_kbytes;
+ CheckGlobalPrefetchAdmissibility_params->compressed_buffer_size_kbytes = mode_lib->mp.CompressedBufferSizeInkByte;
+ CheckGlobalPrefetchAdmissibility_params->detile_buffer_size_bytes_l = mode_lib->mp.DETBufferSizeY;
+ CheckGlobalPrefetchAdmissibility_params->detile_buffer_size_bytes_c = mode_lib->mp.DETBufferSizeC;
+ CheckGlobalPrefetchAdmissibility_params->full_swath_bytes_l = s->full_swath_bytes_l;
+ CheckGlobalPrefetchAdmissibility_params->full_swath_bytes_c = s->full_swath_bytes_c;
+ CheckGlobalPrefetchAdmissibility_params->prefetch_sw_bytes = s->prefetch_sw_bytes;
+ CheckGlobalPrefetchAdmissibility_params->Tpre_rounded = 0; // don't care
+ CheckGlobalPrefetchAdmissibility_params->Tpre_oto = 0; // don't care
+ CheckGlobalPrefetchAdmissibility_params->estimated_urg_bandwidth_required_mbps = mode_lib->mp.urg_bandwidth_available[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_sdp];
+ CheckGlobalPrefetchAdmissibility_params->estimated_dcfclk_mhz = mode_lib->mp.Dcfclk;
+ CheckGlobalPrefetchAdmissibility_params->line_time = s->line_times;
+ CheckGlobalPrefetchAdmissibility_params->dst_y_prefetch = mode_lib->mp.dst_y_prefetch;
+
+ // if recalc_prefetch_schedule is set, recalculate the prefetch schedule with the new impacted_Tpre, prefetch should be possible
+ CheckGlobalPrefetchAdmissibility_params->recalc_prefetch_schedule = &s->dummy_boolean[0];
+ CheckGlobalPrefetchAdmissibility_params->impacted_dst_y_pre = s->impacted_dst_y_pre;
+ CheckGlobalPrefetchAdmissibility(&mode_lib->scratch, CheckGlobalPrefetchAdmissibility_params); // dont care about the check output for mode programming
+ }
+
{
s->DestinationLineTimesForPrefetchLessThan2 = false;
s->VRatioPrefetchMoreThanMax = false;
@@ -10763,11 +11188,11 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex
dml2_printf("DML::%s: k=%d MaxVStartupLines = %u\n", __func__, k, s->MaxVStartupLines[k]);
mode_lib->mp.TWait[k] = CalculateTWait(
- display_cfg->plane_descriptors[k].overrides.reserved_vblank_time_ns,
- mode_lib->mp.UrgentLatency,
- mode_lib->mp.TripToMemory,
- !dml_is_phantom_pipe(&display_cfg->plane_descriptors[k]) && display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.drr_config.enabled ?
- get_g6_temp_read_blackout_us(&mode_lib->soc, (unsigned int)(mode_lib->mp.uclk_freq_mhz * 1000), in_out_params->min_clk_index) : 0.0);
+ display_cfg->plane_descriptors[k].overrides.reserved_vblank_time_ns,
+ mode_lib->mp.UrgentLatency,
+ mode_lib->mp.TripToMemory,
+ !dml_is_phantom_pipe(&display_cfg->plane_descriptors[k]) && display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.drr_config.enabled ?
+ get_g6_temp_read_blackout_us(&mode_lib->soc, (unsigned int)(mode_lib->mp.uclk_freq_mhz * 1000), in_out_params->min_clk_index) : 0.0);
myPipe->Dppclk = mode_lib->mp.Dppclk[k];
myPipe->Dispclk = mode_lib->mp.Dispclk;
@@ -10848,6 +11273,9 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex
CalculatePrefetchSchedule_params->mrq_present = mode_lib->ip.dcn_mrq_present;
CalculatePrefetchSchedule_params->meta_row_bytes = mode_lib->mp.meta_row_bytes[k];
CalculatePrefetchSchedule_params->mall_prefetch_sdp_overhead_factor = mode_lib->mp.mall_prefetch_sdp_overhead_factor[k];
+ CalculatePrefetchSchedule_params->impacted_dst_y_pre = s->impacted_dst_y_pre[k];
+ CalculatePrefetchSchedule_params->vactive_sw_bw_l = mode_lib->mp.vactive_sw_bw_l[k];
+ CalculatePrefetchSchedule_params->vactive_sw_bw_c = mode_lib->mp.vactive_sw_bw_c[k];
// output
CalculatePrefetchSchedule_params->DSTXAfterScaler = &mode_lib->mp.DSTXAfterScaler[k];
@@ -10876,9 +11304,18 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex
CalculatePrefetchSchedule_params->VUpdateWidthPix = &mode_lib->mp.VUpdateWidthPix[k];
CalculatePrefetchSchedule_params->VReadyOffsetPix = &mode_lib->mp.VReadyOffsetPix[k];
CalculatePrefetchSchedule_params->prefetch_cursor_bw = &mode_lib->mp.prefetch_cursor_bw[k];
+ CalculatePrefetchSchedule_params->prefetch_sw_bytes = &s->prefetch_sw_bytes[k];
+ CalculatePrefetchSchedule_params->Tpre_rounded = &s->Tpre_rounded[k];
+ CalculatePrefetchSchedule_params->Tpre_oto = &s->Tpre_oto[k];
+ CalculatePrefetchSchedule_params->prefetch_swath_time_us = &s->dummy_single[0];
mode_lib->mp.NoTimeToPrefetch[k] = CalculatePrefetchSchedule(&mode_lib->scratch, CalculatePrefetchSchedule_params);
+ if (s->impacted_dst_y_pre[k] > 0)
+ mode_lib->mp.impacted_prefetch_margin_us[k] = (mode_lib->mp.dst_y_prefetch[k] - s->impacted_dst_y_pre[k]) * s->line_times[k];
+ else
+ mode_lib->mp.impacted_prefetch_margin_us[k] = 0;
+
#ifdef __DML_VBA_DEBUG__
dml2_printf("DML::%s: k=%0u NoTimeToPrefetch=%0d\n", __func__, k, mode_lib->mp.NoTimeToPrefetch[k]);
#endif
@@ -10956,8 +11393,8 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex
dml2_printf("DML::%s: k=%0u VRatioY=%f\n", __func__, k, display_cfg->plane_descriptors[k].composition.scaler_info.plane0.v_ratio);
dml2_printf("DML::%s: k=%0u prefetch_vmrow_bw=%f\n", __func__, k, mode_lib->mp.prefetch_vmrow_bw[k]);
- dml2_printf("DML::%s: k=%0u ReadBandwidthSurfaceLuma=%f\n", __func__, k, mode_lib->mp.SurfaceReadBandwidthLuma[k]);
- dml2_printf("DML::%s: k=%0u ReadBandwidthSurfaceChroma=%f\n", __func__, k, mode_lib->mp.SurfaceReadBandwidthChroma[k]);
+ dml2_printf("DML::%s: k=%0u vactive_sw_bw_l=%f\n", __func__, k, mode_lib->mp.vactive_sw_bw_l[k]);
+ dml2_printf("DML::%s: k=%0u vactive_sw_bw_c=%f\n", __func__, k, mode_lib->mp.vactive_sw_bw_c[k]);
dml2_printf("DML::%s: k=%0u cursor_bw=%f\n", __func__, k, mode_lib->mp.cursor_bw[k]);
dml2_printf("DML::%s: k=%0u dpte_row_bw=%f\n", __func__, k, mode_lib->mp.dpte_row_bw[k]);
dml2_printf("DML::%s: k=%0u meta_row_bw=%f\n", __func__, k, mode_lib->mp.meta_row_bw[k]);
@@ -10988,8 +11425,8 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex
calculate_peak_bandwidth_params->mall_prefetch_sdp_overhead_factor = mode_lib->mp.mall_prefetch_sdp_overhead_factor;
calculate_peak_bandwidth_params->mall_prefetch_dram_overhead_factor = mode_lib->mp.mall_prefetch_dram_overhead_factor;
- calculate_peak_bandwidth_params->surface_read_bandwidth_l = mode_lib->mp.SurfaceReadBandwidthLuma;
- calculate_peak_bandwidth_params->surface_read_bandwidth_c = mode_lib->mp.SurfaceReadBandwidthChroma;
+ calculate_peak_bandwidth_params->surface_read_bandwidth_l = mode_lib->mp.vactive_sw_bw_l;
+ calculate_peak_bandwidth_params->surface_read_bandwidth_c = mode_lib->mp.vactive_sw_bw_c;
calculate_peak_bandwidth_params->prefetch_bandwidth_l = mode_lib->mp.RequiredPrefetchPixelDataBWLuma;
calculate_peak_bandwidth_params->prefetch_bandwidth_c = mode_lib->mp.RequiredPrefetchPixelDataBWChroma;
calculate_peak_bandwidth_params->excess_vactive_fill_bw_l = mode_lib->mp.excess_vactive_fill_bw_l;
@@ -11120,8 +11557,8 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex
calculate_peak_bandwidth_params->mall_prefetch_sdp_overhead_factor = mode_lib->mp.mall_prefetch_sdp_overhead_factor;
calculate_peak_bandwidth_params->mall_prefetch_dram_overhead_factor = mode_lib->mp.mall_prefetch_dram_overhead_factor;
- calculate_peak_bandwidth_params->surface_read_bandwidth_l = mode_lib->mp.SurfaceReadBandwidthLuma;
- calculate_peak_bandwidth_params->surface_read_bandwidth_c = mode_lib->mp.SurfaceReadBandwidthChroma;
+ calculate_peak_bandwidth_params->surface_read_bandwidth_l = mode_lib->mp.vactive_sw_bw_l;
+ calculate_peak_bandwidth_params->surface_read_bandwidth_c = mode_lib->mp.vactive_sw_bw_c;
calculate_peak_bandwidth_params->prefetch_bandwidth_l = mode_lib->mp.RequiredPrefetchPixelDataBWLuma;
calculate_peak_bandwidth_params->prefetch_bandwidth_c = mode_lib->mp.RequiredPrefetchPixelDataBWChroma;
calculate_peak_bandwidth_params->excess_vactive_fill_bw_l = mode_lib->mp.excess_vactive_fill_bw_l;
@@ -11238,8 +11675,8 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex
s->mmSOCParameters.USRRetrainingLatency = 0;
s->mmSOCParameters.SMNLatency = 0;
s->mmSOCParameters.g6_temp_read_blackout_us = get_g6_temp_read_blackout_us(&mode_lib->soc, (unsigned int)(mode_lib->mp.uclk_freq_mhz * 1000), in_out_params->min_clk_index);
- s->mmSOCParameters.max_urgent_latency_us = get_max_urgent_latency_us(&mode_lib->soc.qos_parameters.qos_params.dcn4x, mode_lib->ms.uclk_freq_mhz, mode_lib->ms.FabricClock, in_out_params->min_clk_index);
- s->mmSOCParameters.df_response_time_us = mode_lib->soc.qos_parameters.qos_params.dcn4x.df_qos_response_time_fclk_cycles / mode_lib->ms.FabricClock;
+ s->mmSOCParameters.max_urgent_latency_us = get_max_urgent_latency_us(&mode_lib->soc.qos_parameters.qos_params.dcn4x, mode_lib->mp.uclk_freq_mhz, mode_lib->mp.FabricClock, in_out_params->min_clk_index);
+ s->mmSOCParameters.df_response_time_us = mode_lib->soc.qos_parameters.qos_params.dcn4x.df_qos_response_time_fclk_cycles / mode_lib->mp.FabricClock;
s->mmSOCParameters.qos_type = mode_lib->soc.qos_parameters.qos_type;
CalculateWatermarks_params->display_cfg = display_cfg;
@@ -11289,7 +11726,7 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex
CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport(&mode_lib->scratch, CalculateWatermarks_params);
for (k = 0; k < s->num_active_planes; ++k) {
- if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.enable == true) {
+ if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.active_writebacks_per_stream > 0) {
mode_lib->mp.WritebackAllowDRAMClockChangeEndPosition[k] = math_max2(0, mode_lib->mp.VStartupMin[k] * display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total /
((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000) - mode_lib->mp.Watermark.WritebackDRAMClockChangeWatermark);
mode_lib->mp.WritebackAllowFCLKChangeEndPosition[k] = math_max2(0, mode_lib->mp.VStartupMin[k] * display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total /
@@ -11475,25 +11912,25 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex
//Maximum Bandwidth Used
s->TotalWRBandwidth = 0;
- s->WRBandwidth = 0;
- for (k = 0; k < s->num_active_planes; ++k) {
- if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.enable == true && display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.pixel_format == dml2_444_32) {
- s->WRBandwidth = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.output_height * display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.output_width /
- (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total * display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.input_height / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)) * 4;
- } else if (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.enable == true) {
- s->WRBandwidth = display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.output_height * display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.output_width /
- (display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.h_total * display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].writeback.scaling_info.input_height / ((double)display_cfg->stream_descriptors[display_cfg->plane_descriptors[k].stream_index].timing.pixel_clock_khz / 1000)) * 8;
+ for (k = 0; k < display_cfg->num_streams; ++k) {
+ s->WRBandwidth = 0;
+ if (display_cfg->stream_descriptors[k].writeback.active_writebacks_per_stream > 0) {
+ s->WRBandwidth = display_cfg->stream_descriptors[k].writeback.writeback_stream[0].output_height
+ * display_cfg->stream_descriptors[k].writeback.writeback_stream[0].output_width /
+ (display_cfg->stream_descriptors[k].timing.h_total * display_cfg->stream_descriptors[k].writeback.writeback_stream[0].input_height
+ / ((double)display_cfg->stream_descriptors[k].timing.pixel_clock_khz / 1000))
+ * (display_cfg->stream_descriptors[k].writeback.writeback_stream[0].pixel_format == dml2_444_32 ? 4.0 : 8.0);
+ s->TotalWRBandwidth = s->TotalWRBandwidth + s->WRBandwidth;
}
- s->TotalWRBandwidth = s->TotalWRBandwidth + s->WRBandwidth;
}
mode_lib->mp.TotalDataReadBandwidth = 0;
for (k = 0; k < s->num_active_planes; ++k) {
- mode_lib->mp.TotalDataReadBandwidth = mode_lib->mp.TotalDataReadBandwidth + mode_lib->mp.SurfaceReadBandwidthLuma[k] + mode_lib->mp.SurfaceReadBandwidthChroma[k];
+ mode_lib->mp.TotalDataReadBandwidth = mode_lib->mp.TotalDataReadBandwidth + mode_lib->mp.vactive_sw_bw_l[k] + mode_lib->mp.vactive_sw_bw_c[k];
#ifdef __DML_VBA_DEBUG__
dml2_printf("DML::%s: k=%u, TotalDataReadBandwidth = %f\n", __func__, k, mode_lib->mp.TotalDataReadBandwidth);
- dml2_printf("DML::%s: k=%u, ReadBandwidthSurfaceLuma = %f\n", __func__, k, mode_lib->mp.SurfaceReadBandwidthLuma[k]);
- dml2_printf("DML::%s: k=%u, ReadBandwidthSurfaceChroma = %f\n", __func__, k, mode_lib->mp.SurfaceReadBandwidthChroma[k]);
+ dml2_printf("DML::%s: k=%u, vactive_sw_bw_l = %f\n", __func__, k, mode_lib->mp.vactive_sw_bw_l[k]);
+ dml2_printf("DML::%s: k=%u, vactive_sw_bw_c = %f\n", __func__, k, mode_lib->mp.vactive_sw_bw_c[k]);
#endif
}
@@ -11530,8 +11967,8 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex
CalculateStutterEfficiency_params->BlockWidth256BytesC = mode_lib->mp.Read256BlockWidthC;
CalculateStutterEfficiency_params->DCCYMaxUncompressedBlock = mode_lib->mp.DCCYMaxUncompressedBlock;
CalculateStutterEfficiency_params->DCCCMaxUncompressedBlock = mode_lib->mp.DCCCMaxUncompressedBlock;
- CalculateStutterEfficiency_params->ReadBandwidthSurfaceLuma = mode_lib->mp.SurfaceReadBandwidthLuma;
- CalculateStutterEfficiency_params->ReadBandwidthSurfaceChroma = mode_lib->mp.SurfaceReadBandwidthChroma;
+ CalculateStutterEfficiency_params->ReadBandwidthSurfaceLuma = mode_lib->mp.vactive_sw_bw_l;
+ CalculateStutterEfficiency_params->ReadBandwidthSurfaceChroma = mode_lib->mp.vactive_sw_bw_c;
CalculateStutterEfficiency_params->dpte_row_bw = mode_lib->mp.dpte_row_bw;
CalculateStutterEfficiency_params->meta_row_bw = mode_lib->mp.meta_row_bw;
CalculateStutterEfficiency_params->rob_alloc_compressed = mode_lib->ip.dcn_mrq_present;
@@ -11742,7 +12179,7 @@ static void rq_dlg_get_wm_regs(const struct dml2_display_cfg *display_cfg, const
wm_regs->fclk_pstate = (int unsigned)(mode_lib->mp.Watermark.FCLKChangeWatermark * refclk_freq_in_mhz);
wm_regs->sr_enter = (int unsigned)(mode_lib->mp.Watermark.StutterEnterPlusExitWatermark * refclk_freq_in_mhz);
wm_regs->sr_exit = (int unsigned)(mode_lib->mp.Watermark.StutterExitWatermark * refclk_freq_in_mhz);
- wm_regs->temp_read_or_ppt = (int unsigned)(mode_lib->mp.Watermark.g6_temp_read_watermark_us * refclk_freq_in_mhz);
+ wm_regs->temp_read_or_ppt = (int unsigned)(mode_lib->mp.Watermark.temp_read_or_ppt_watermark_us * refclk_freq_in_mhz);
wm_regs->uclk_pstate = (int unsigned)(mode_lib->mp.Watermark.DRAMClockChangeWatermark * refclk_freq_in_mhz);
wm_regs->urgent = (int unsigned)(mode_lib->mp.Watermark.UrgentWatermark * refclk_freq_in_mhz);
wm_regs->usr = (int unsigned)(mode_lib->mp.Watermark.USRRetrainingWatermark * refclk_freq_in_mhz);
@@ -12321,14 +12758,18 @@ void dml2_core_calcs_get_global_fams2_programming(const struct dml2_core_interna
void dml2_core_calcs_get_stream_fams2_programming(const struct dml2_core_internal_display_mode_lib *mode_lib,
const struct display_configuation_with_meta *display_cfg,
- struct dmub_fams2_stream_static_state *fams2_programming,
- enum dml2_uclk_pstate_support_method pstate_method,
+ union dmub_cmd_fams2_config *fams2_base_programming,
+ union dmub_cmd_fams2_config *fams2_sub_programming,
+ enum dml2_pstate_method pstate_method,
int plane_index)
{
const struct dml2_plane_parameters *plane_descriptor = &display_cfg->display_config.plane_descriptors[plane_index];
const struct dml2_stream_parameters *stream_descriptor = &display_cfg->display_config.stream_descriptors[plane_descriptor->stream_index];
const struct dml2_fams2_meta *stream_fams2_meta = &display_cfg->stage3.stream_fams2_meta[plane_descriptor->stream_index];
+ struct dmub_fams2_cmd_stream_static_base_state *base_programming = &fams2_base_programming->stream_v1.base;
+ union dmub_fams2_cmd_stream_static_sub_state *sub_programming = &fams2_sub_programming->stream_v1.sub_state;
+
unsigned int i;
if (display_cfg->display_config.overrides.all_streams_blanked) {
@@ -12337,110 +12778,110 @@ void dml2_core_calcs_get_stream_fams2_programming(const struct dml2_core_interna
}
/* from display configuration */
- fams2_programming->htotal = (uint16_t)stream_descriptor->timing.h_total;
- fams2_programming->vtotal = (uint16_t)stream_descriptor->timing.v_total;
- fams2_programming->vblank_start = (uint16_t)(stream_fams2_meta->nom_vtotal -
+ base_programming->htotal = (uint16_t)stream_descriptor->timing.h_total;
+ base_programming->vtotal = (uint16_t)stream_descriptor->timing.v_total;
+ base_programming->vblank_start = (uint16_t)(stream_fams2_meta->nom_vtotal -
stream_descriptor->timing.v_front_porch);
- fams2_programming->vblank_end = (uint16_t)(stream_fams2_meta->nom_vtotal -
+ base_programming->vblank_end = (uint16_t)(stream_fams2_meta->nom_vtotal -
stream_descriptor->timing.v_front_porch -
stream_descriptor->timing.v_active);
- fams2_programming->config.bits.is_drr = stream_descriptor->timing.drr_config.enabled;
+ base_programming->config.bits.is_drr = stream_descriptor->timing.drr_config.enabled;
/* from meta */
- fams2_programming->otg_vline_time_ns =
+ base_programming->otg_vline_time_ns =
(unsigned int)(stream_fams2_meta->otg_vline_time_us * 1000.0);
- fams2_programming->scheduling_delay_otg_vlines = (uint8_t)stream_fams2_meta->scheduling_delay_otg_vlines;
- fams2_programming->contention_delay_otg_vlines = (uint8_t)stream_fams2_meta->contention_delay_otg_vlines;
- fams2_programming->vline_int_ack_delay_otg_vlines = (uint8_t)stream_fams2_meta->vertical_interrupt_ack_delay_otg_vlines;
- fams2_programming->drr_keepout_otg_vline = (uint16_t)(stream_fams2_meta->nom_vtotal -
+ base_programming->scheduling_delay_otg_vlines = (uint8_t)stream_fams2_meta->scheduling_delay_otg_vlines;
+ base_programming->contention_delay_otg_vlines = (uint8_t)stream_fams2_meta->contention_delay_otg_vlines;
+ base_programming->vline_int_ack_delay_otg_vlines = (uint8_t)stream_fams2_meta->vertical_interrupt_ack_delay_otg_vlines;
+ base_programming->drr_keepout_otg_vline = (uint16_t)(stream_fams2_meta->nom_vtotal -
stream_descriptor->timing.v_front_porch -
stream_fams2_meta->method_drr.programming_delay_otg_vlines);
- fams2_programming->allow_to_target_delay_otg_vlines = (uint8_t)stream_fams2_meta->allow_to_target_delay_otg_vlines;
- fams2_programming->max_vtotal = (uint16_t)stream_fams2_meta->max_vtotal;
+ base_programming->allow_to_target_delay_otg_vlines = (uint8_t)stream_fams2_meta->allow_to_target_delay_otg_vlines;
+ base_programming->max_vtotal = (uint16_t)stream_fams2_meta->max_vtotal;
/* from core */
- fams2_programming->config.bits.min_ttu_vblank_usable = true;
+ base_programming->config.bits.min_ttu_vblank_usable = true;
for (i = 0; i < display_cfg->display_config.num_planes; i++) {
/* check if all planes support p-state in blank */
if (display_cfg->display_config.plane_descriptors[i].stream_index == plane_descriptor->stream_index &&
mode_lib->mp.MinTTUVBlank[i] <= mode_lib->mp.Watermark.DRAMClockChangeWatermark) {
- fams2_programming->config.bits.min_ttu_vblank_usable = false;
+ base_programming->config.bits.min_ttu_vblank_usable = false;
break;
}
}
switch (pstate_method) {
- case dml2_uclk_pstate_support_method_vactive:
- case dml2_uclk_pstate_support_method_fw_vactive_drr:
+ case dml2_pstate_method_vactive:
+ case dml2_pstate_method_fw_vactive_drr:
/* legacy vactive */
- fams2_programming->type = FAMS2_STREAM_TYPE_VACTIVE;
- fams2_programming->sub_state.legacy.vactive_det_fill_delay_otg_vlines =
- (uint8_t)stream_fams2_meta->method_vactive.max_vactive_det_fill_delay_otg_vlines;
- fams2_programming->allow_start_otg_vline =
- (uint16_t)stream_fams2_meta->method_vactive.common.allow_start_otg_vline;
- fams2_programming->allow_end_otg_vline =
- (uint16_t)stream_fams2_meta->method_vactive.common.allow_end_otg_vline;
- fams2_programming->config.bits.clamp_vtotal_min = true;
+ base_programming->type = FAMS2_STREAM_TYPE_VACTIVE;
+ sub_programming->legacy.vactive_det_fill_delay_otg_vlines =
+ (uint8_t)stream_fams2_meta->method_vactive.max_vactive_det_fill_delay_otg_vlines;
+ base_programming->allow_start_otg_vline =
+ (uint16_t)stream_fams2_meta->method_vactive.common.allow_start_otg_vline;
+ base_programming->allow_end_otg_vline =
+ (uint16_t)stream_fams2_meta->method_vactive.common.allow_end_otg_vline;
+ base_programming->config.bits.clamp_vtotal_min = true;
break;
- case dml2_uclk_pstate_support_method_vblank:
- case dml2_uclk_pstate_support_method_fw_vblank_drr:
+ case dml2_pstate_method_vblank:
+ case dml2_pstate_method_fw_vblank_drr:
/* legacy vblank */
- fams2_programming->type = FAMS2_STREAM_TYPE_VBLANK;
- fams2_programming->allow_start_otg_vline =
- (uint16_t)stream_fams2_meta->method_vblank.common.allow_start_otg_vline;
- fams2_programming->allow_end_otg_vline =
- (uint16_t)stream_fams2_meta->method_vblank.common.allow_end_otg_vline;
- fams2_programming->config.bits.clamp_vtotal_min = true;
+ base_programming->type = FAMS2_STREAM_TYPE_VBLANK;
+ base_programming->allow_start_otg_vline =
+ (uint16_t)stream_fams2_meta->method_vblank.common.allow_start_otg_vline;
+ base_programming->allow_end_otg_vline =
+ (uint16_t)stream_fams2_meta->method_vblank.common.allow_end_otg_vline;
+ base_programming->config.bits.clamp_vtotal_min = true;
break;
- case dml2_uclk_pstate_support_method_fw_drr:
+ case dml2_pstate_method_fw_drr:
/* drr */
- fams2_programming->type = FAMS2_STREAM_TYPE_DRR;
- fams2_programming->sub_state.drr.programming_delay_otg_vlines =
- (uint8_t)stream_fams2_meta->method_drr.programming_delay_otg_vlines;
- fams2_programming->sub_state.drr.nom_stretched_vtotal =
- (uint16_t)stream_fams2_meta->method_drr.stretched_vtotal;
- fams2_programming->allow_start_otg_vline =
- (uint16_t)stream_fams2_meta->method_drr.common.allow_start_otg_vline;
- fams2_programming->allow_end_otg_vline =
- (uint16_t)stream_fams2_meta->method_drr.common.allow_end_otg_vline;
+ base_programming->type = FAMS2_STREAM_TYPE_DRR;
+ sub_programming->drr.programming_delay_otg_vlines =
+ (uint8_t)stream_fams2_meta->method_drr.programming_delay_otg_vlines;
+ sub_programming->drr.nom_stretched_vtotal =
+ (uint16_t)stream_fams2_meta->method_drr.stretched_vtotal;
+ base_programming->allow_start_otg_vline =
+ (uint16_t)stream_fams2_meta->method_drr.common.allow_start_otg_vline;
+ base_programming->allow_end_otg_vline =
+ (uint16_t)stream_fams2_meta->method_drr.common.allow_end_otg_vline;
/* drr only clamps to vtotal min for single display */
- fams2_programming->config.bits.clamp_vtotal_min = display_cfg->display_config.num_streams == 1;
- fams2_programming->sub_state.drr.only_stretch_if_required = true;
+ base_programming->config.bits.clamp_vtotal_min = display_cfg->display_config.num_streams == 1;
+ sub_programming->drr.only_stretch_if_required = true;
break;
- case dml2_uclk_pstate_support_method_fw_subvp_phantom:
- case dml2_uclk_pstate_support_method_fw_subvp_phantom_drr:
+ case dml2_pstate_method_fw_svp:
+ case dml2_pstate_method_fw_svp_drr:
/* subvp */
- fams2_programming->type = FAMS2_STREAM_TYPE_SUBVP;
- fams2_programming->sub_state.subvp.vratio_numerator =
- (uint16_t)(plane_descriptor->composition.scaler_info.plane0.v_ratio * 1000.0);
- fams2_programming->sub_state.subvp.vratio_denominator = 1000;
- fams2_programming->sub_state.subvp.programming_delay_otg_vlines =
- (uint8_t)stream_fams2_meta->method_subvp.programming_delay_otg_vlines;
- fams2_programming->sub_state.subvp.prefetch_to_mall_otg_vlines =
- (uint8_t)stream_fams2_meta->method_subvp.prefetch_to_mall_delay_otg_vlines;
- fams2_programming->sub_state.subvp.phantom_vtotal =
- (uint16_t)stream_fams2_meta->method_subvp.phantom_vtotal;
- fams2_programming->sub_state.subvp.phantom_vactive =
- (uint16_t)stream_fams2_meta->method_subvp.phantom_vactive;
- fams2_programming->sub_state.subvp.config.bits.is_multi_planar =
- plane_descriptor->surface.plane1.height > 0;
- fams2_programming->sub_state.subvp.config.bits.is_yuv420 =
- plane_descriptor->pixel_format == dml2_420_8 ||
- plane_descriptor->pixel_format == dml2_420_10 ||
- plane_descriptor->pixel_format == dml2_420_12;
-
- fams2_programming->allow_start_otg_vline =
- (uint16_t)stream_fams2_meta->method_subvp.common.allow_start_otg_vline;
- fams2_programming->allow_end_otg_vline =
- (uint16_t)stream_fams2_meta->method_subvp.common.allow_end_otg_vline;
- fams2_programming->config.bits.clamp_vtotal_min = true;
+ base_programming->type = FAMS2_STREAM_TYPE_SUBVP;
+ sub_programming->subvp.vratio_numerator =
+ (uint16_t)(plane_descriptor->composition.scaler_info.plane0.v_ratio * 1000.0);
+ sub_programming->subvp.vratio_denominator = 1000;
+ sub_programming->subvp.programming_delay_otg_vlines =
+ (uint8_t)stream_fams2_meta->method_subvp.programming_delay_otg_vlines;
+ sub_programming->subvp.prefetch_to_mall_otg_vlines =
+ (uint8_t)stream_fams2_meta->method_subvp.prefetch_to_mall_delay_otg_vlines;
+ sub_programming->subvp.phantom_vtotal =
+ (uint16_t)stream_fams2_meta->method_subvp.phantom_vtotal;
+ sub_programming->subvp.phantom_vactive =
+ (uint16_t)stream_fams2_meta->method_subvp.phantom_vactive;
+ sub_programming->subvp.config.bits.is_multi_planar =
+ plane_descriptor->surface.plane1.height > 0;
+ sub_programming->subvp.config.bits.is_yuv420 =
+ plane_descriptor->pixel_format == dml2_420_8 ||
+ plane_descriptor->pixel_format == dml2_420_10 ||
+ plane_descriptor->pixel_format == dml2_420_12;
+
+ base_programming->allow_start_otg_vline =
+ (uint16_t)stream_fams2_meta->method_subvp.common.allow_start_otg_vline;
+ base_programming->allow_end_otg_vline =
+ (uint16_t)stream_fams2_meta->method_subvp.common.allow_end_otg_vline;
+ base_programming->config.bits.clamp_vtotal_min = true;
break;
- case dml2_uclk_pstate_support_method_reserved_hw:
- case dml2_uclk_pstate_support_method_reserved_fw:
- case dml2_uclk_pstate_support_method_reserved_fw_drr_fixed:
- case dml2_uclk_pstate_support_method_reserved_fw_drr_var:
- case dml2_uclk_pstate_support_method_not_supported:
- case dml2_uclk_pstate_support_method_count:
+ case dml2_pstate_method_reserved_hw:
+ case dml2_pstate_method_reserved_fw:
+ case dml2_pstate_method_reserved_fw_drr_clamped:
+ case dml2_pstate_method_reserved_fw_drr_var:
+ case dml2_pstate_method_na:
+ case dml2_pstate_method_count:
default:
/* this should never happen */
break;
@@ -12569,6 +13010,8 @@ void dml2_core_calcs_get_informative(const struct dml2_core_internal_display_mod
out->informative.mode_support_info.InvalidCombinationOfMALLUseForPState = mode_lib->ms.support.InvalidCombinationOfMALLUseForPState;
out->informative.mode_support_info.ExceededMALLSize = mode_lib->ms.support.ExceededMALLSize;
out->informative.mode_support_info.EnoughWritebackUnits = mode_lib->ms.support.EnoughWritebackUnits;
+ out->informative.mode_support_info.temp_read_or_ppt_support = mode_lib->ms.support.temp_read_or_ppt_support;
+ out->informative.mode_support_info.g6_temp_read_support = mode_lib->ms.support.g6_temp_read_support;
out->informative.mode_support_info.ExceededMultistreamSlots = mode_lib->ms.support.ExceededMultistreamSlots;
out->informative.mode_support_info.NotEnoughDSCUnits = mode_lib->ms.support.NotEnoughDSCUnits;
@@ -12662,7 +13105,7 @@ void dml2_core_calcs_get_informative(const struct dml2_core_internal_display_mod
out->informative.watermarks.pstate_change_us = dml_get_wm_dram_clock_change(mode_lib);
out->informative.watermarks.fclk_pstate_change_us = dml_get_wm_fclk_change(mode_lib);
out->informative.watermarks.usr_retraining_us = dml_get_wm_usr_retraining(mode_lib);
- out->informative.watermarks.g6_temp_read_watermark_us = dml_get_wm_g6_temp_read(mode_lib);
+ out->informative.watermarks.temp_read_or_ppt_watermark_us = dml_get_wm_temp_read_or_ppt(mode_lib);
out->informative.mall.total_surface_size_in_mall_bytes = 0;
for (k = 0; k < out->display_config.num_planes; ++k)
@@ -12745,6 +13188,8 @@ void dml2_core_calcs_get_informative(const struct dml2_core_internal_display_mod
out->informative.qos.max_active_fclk_change_latency_supported = dml_get_fclk_change_latency(mode_lib);
+ out->informative.misc.LowestPrefetchMargin = 10 * 1000 * 1000;
+
for (k = 0; k < out->display_config.num_planes; k++) {
if ((out->display_config.plane_descriptors->overrides.reserved_vblank_time_ns >= 1000.0 * mode_lib->soc.power_management_parameters.dram_clk_change_blackout_us)
@@ -12824,6 +13269,7 @@ void dml2_core_calcs_get_informative(const struct dml2_core_internal_display_mod
out->informative.misc.DisplayPipeLineDeliveryTimeLumaPrefetch[k] = mode_lib->mp.DisplayPipeLineDeliveryTimeLumaPrefetch[k];
out->informative.misc.DisplayPipeLineDeliveryTimeChromaPrefetch[k] = mode_lib->mp.DisplayPipeLineDeliveryTimeChromaPrefetch[k];
+ out->informative.misc.WritebackRequiredBandwidth = mode_lib->scratch.dml_core_mode_programming_locals.TotalWRBandwidth / 1000.0;
out->informative.misc.WritebackAllowDRAMClockChangeEndPosition[k] = mode_lib->mp.WritebackAllowDRAMClockChangeEndPosition[k];
out->informative.misc.WritebackAllowFCLKChangeEndPosition[k] = mode_lib->mp.WritebackAllowFCLKChangeEndPosition[k];
out->informative.misc.DSCCLK_calculated[k] = mode_lib->mp.DSCCLK[k];
@@ -12831,6 +13277,9 @@ void dml2_core_calcs_get_informative(const struct dml2_core_internal_display_mod
out->informative.misc.PTE_BUFFER_MODE[k] = mode_lib->mp.PTE_BUFFER_MODE[k];
out->informative.misc.DSCDelay[k] = mode_lib->mp.DSCDelay[k];
out->informative.misc.MaxActiveDRAMClockChangeLatencySupported[k] = mode_lib->mp.MaxActiveDRAMClockChangeLatencySupported[k];
+
+ if (mode_lib->mp.impacted_prefetch_margin_us[k] < out->informative.misc.LowestPrefetchMargin)
+ out->informative.misc.LowestPrefetchMargin = mode_lib->mp.impacted_prefetch_margin_us[k];
}
// For this DV informative layer, all pipes in the same planes will just use the same id
@@ -12853,16 +13302,11 @@ void dml2_core_calcs_get_informative(const struct dml2_core_internal_display_mod
out->informative.non_optimized_mcache_allocation[k].global_mcache_ids_plane1[n] = k;
}
}
-
- out->informative.qos.max_non_urgent_latency_us = mode_lib->soc.qos_parameters.qos_params.dcn4x.per_uclk_dpm_params[mode_lib->mp.qos_param_index].maximum_latency_when_non_urgent_uclk_cycles
- / mode_lib->mp.uclk_freq_mhz * (1 + mode_lib->soc.qos_parameters.qos_params.dcn4x.umc_max_latency_margin / 100.0)
- + mode_lib->soc.qos_parameters.qos_params.dcn4x.mall_overhead_fclk_cycles / mode_lib->mp.FabricClock
- + mode_lib->soc.qos_parameters.qos_params.dcn4x.max_round_trip_to_furthest_cs_fclk_cycles / mode_lib->mp.FabricClock
- * (1 + mode_lib->soc.qos_parameters.qos_params.dcn4x.fabric_max_transport_latency_margin / 100.0);
+ out->informative.qos.max_non_urgent_latency_us = dml_get_max_non_urgent_latency_us(mode_lib);
if (mode_lib->soc.qos_parameters.qos_type == dml2_qos_param_type_dcn4x) {
if (((mode_lib->ip.rob_buffer_size_kbytes - mode_lib->ip.pixel_chunk_size_kbytes) * 1024
- / mode_lib->mp.non_urg_bandwidth_required[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_sdp]) >= out->informative.qos.max_non_urgent_latency_us) {
+ / mode_lib->ms.support.non_urg_bandwidth_required[dml2_core_internal_soc_state_sys_active][dml2_core_internal_bw_sdp]) >= out->informative.qos.max_non_urgent_latency_us) {
out->informative.misc.ROBUrgencyAvoidance = true;
} else {
out->informative.misc.ROBUrgencyAvoidance = false;
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.h b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.h
index df2d1550a14b..27ef0e096b25 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.h
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.h
@@ -28,7 +28,7 @@ void dml2_core_calcs_get_plane_support_info(const struct dml2_display_cfg *displ
void dml2_core_calcs_get_informative(const struct dml2_core_internal_display_mode_lib *mode_lib, struct dml2_display_cfg_programming *out);
void dml2_core_calcs_get_stream_support_info(const struct dml2_display_cfg *display_cfg, const struct dml2_core_internal_display_mode_lib *mode_lib, struct core_stream_support_info *out, int plane_index);
void dml2_core_calcs_get_mall_allocation(struct dml2_core_internal_display_mode_lib *mode_lib, unsigned int *out, int pipe_index);
-void dml2_core_calcs_get_stream_fams2_programming(const struct dml2_core_internal_display_mode_lib *mode_lib, const struct display_configuation_with_meta *display_cfg, struct dmub_fams2_stream_static_state *fams2_programming, enum dml2_uclk_pstate_support_method pstate_method, int plane_index);
+void dml2_core_calcs_get_stream_fams2_programming(const struct dml2_core_internal_display_mode_lib *mode_lib, const struct display_configuation_with_meta *display_cfg, union dmub_cmd_fams2_config *fams2_base_programming, union dmub_cmd_fams2_config *fams2_sub_programming, enum dml2_pstate_method pstate_method, int plane_index);
void dml2_core_calcs_get_global_fams2_programming(const struct dml2_core_internal_display_mode_lib *mode_lib, const struct display_configuation_with_meta *display_cfg, struct dmub_cmd_fams2_global_config *fams2_global_config);
void dml2_core_calcs_get_dpte_row_height(unsigned int *dpte_row_height, struct dml2_core_internal_display_mode_lib *mode_lib, bool is_plane1, enum dml2_source_format_class SourcePixelFormat, enum dml2_swizzle_mode SurfaceTiling, enum dml2_rotation_angle ScanDirection, unsigned int pitch, unsigned int GPUVMMinPageSizeKBytes);
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_shared_types.h b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_shared_types.h
index cbdfbd5a0bde..23c0fca5515f 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_shared_types.h
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_shared_types.h
@@ -201,7 +201,7 @@ struct dml2_core_internal_watermarks {
double Z8StutterExitWatermark;
double Z8StutterEnterPlusExitWatermark;
double USRRetrainingWatermark;
- double g6_temp_read_watermark_us;
+ double temp_read_or_ppt_watermark_us;
};
struct dml2_core_internal_mode_support_info {
@@ -252,8 +252,8 @@ struct dml2_core_internal_mode_support_info {
bool PTEBufferSizeNotExceeded;
bool DCCMetaBufferSizeNotExceeded;
- enum dml2_dram_clock_change_support DRAMClockChangeSupport[DML2_MAX_PLANES];
- enum dml2_fclock_change_support FCLKChangeSupport[DML2_MAX_PLANES];
+ enum dml2_pstate_change_support DRAMClockChangeSupport[DML2_MAX_PLANES];
+ enum dml2_pstate_change_support FCLKChangeSupport[DML2_MAX_PLANES];
bool global_dram_clock_change_supported;
bool global_fclk_change_supported;
bool USRRetrainingSupport;
@@ -318,12 +318,15 @@ struct dml2_core_internal_mode_support_info {
bool avg_bandwidth_support_ok[dml2_core_internal_soc_state_max][dml2_core_internal_bw_max];
double max_urgent_latency_us;
+ double max_non_urgent_latency_us;
double avg_non_urgent_latency_us;
double avg_urgent_latency_us;
+ double df_response_time_us;
bool incorrect_imall_usage;
bool g6_temp_read_support;
+ bool temp_read_or_ppt_support;
struct dml2_core_internal_watermarks watermarks;
};
@@ -378,8 +381,8 @@ struct dml2_core_internal_mode_support {
unsigned int DETBufferSizeC[DML2_MAX_PLANES];
unsigned int SwathHeightY[DML2_MAX_PLANES];
unsigned int SwathHeightC[DML2_MAX_PLANES];
- unsigned int SwathWidthY[DML2_MAX_PLANES];
- unsigned int SwathWidthC[DML2_MAX_PLANES];
+ unsigned int SwathWidthY[DML2_MAX_PLANES]; // per-pipe
+ unsigned int SwathWidthC[DML2_MAX_PLANES]; // per-pipe
// ----------------------------------
// Intermediates/Informational
@@ -476,9 +479,9 @@ struct dml2_core_internal_mode_support {
// Bandwidth Related Info
double BandwidthAvailableForImmediateFlip;
- double SurfaceReadBandwidthLuma[DML2_MAX_PLANES]; // no dcc overhead, for the plane
- double SurfaceReadBandwidthChroma[DML2_MAX_PLANES];
- double WriteBandwidth[DML2_MAX_PLANES];
+ double vactive_sw_bw_l[DML2_MAX_PLANES]; // no dcc overhead, for the plane
+ double vactive_sw_bw_c[DML2_MAX_PLANES];
+ double WriteBandwidth[DML2_MAX_PLANES][DML2_MAX_WRITEBACK];
double RequiredPrefetchPixelDataBWLuma[DML2_MAX_PLANES];
double RequiredPrefetchPixelDataBWChroma[DML2_MAX_PLANES];
double cursor_bw[DML2_MAX_PLANES];
@@ -539,7 +542,7 @@ struct dml2_core_internal_mode_program {
unsigned int qos_param_index; // to access the uclk dependent dpm table
unsigned int active_min_uclk_dpm_index; // to access the min_clk table
double FabricClock; /// <brief Basically just the clock freq at the min (or given) state
- double DCFCLK; /// <brief Basically just the clock freq at the min (or given) state and max combine setting
+ //double DCFCLK; /// <brief Basically just the clock freq at the min (or given) state and max combine setting
double dram_bw_mbps;
double uclk_freq_mhz;
unsigned int NoOfDPP[DML2_MAX_PLANES];
@@ -562,14 +565,14 @@ struct dml2_core_internal_mode_program {
double BytePerPixelInDETC[DML2_MAX_PLANES];
unsigned int BytePerPixelY[DML2_MAX_PLANES];
unsigned int BytePerPixelC[DML2_MAX_PLANES];
- unsigned int SwathWidthY[DML2_MAX_PLANES];
- unsigned int SwathWidthC[DML2_MAX_PLANES];
+ unsigned int SwathWidthY[DML2_MAX_PLANES]; // per-pipe
+ unsigned int SwathWidthC[DML2_MAX_PLANES]; // per-pipe
unsigned int req_per_swath_ub_l[DML2_MAX_PLANES];
unsigned int req_per_swath_ub_c[DML2_MAX_PLANES];
unsigned int SwathWidthSingleDPPY[DML2_MAX_PLANES];
unsigned int SwathWidthSingleDPPC[DML2_MAX_PLANES];
- double SurfaceReadBandwidthLuma[DML2_MAX_PLANES];
- double SurfaceReadBandwidthChroma[DML2_MAX_PLANES];
+ double vactive_sw_bw_l[DML2_MAX_PLANES];
+ double vactive_sw_bw_c[DML2_MAX_PLANES];
double excess_vactive_fill_bw_l[DML2_MAX_PLANES];
double excess_vactive_fill_bw_c[DML2_MAX_PLANES];
@@ -797,8 +800,9 @@ struct dml2_core_internal_mode_program {
double MaxActiveFCLKChangeLatencySupported;
bool USRRetrainingSupport;
bool g6_temp_read_support;
- enum dml2_fclock_change_support FCLKChangeSupport[DML2_MAX_PLANES];
- enum dml2_dram_clock_change_support DRAMClockChangeSupport[DML2_MAX_PLANES];
+ bool temp_read_or_ppt_support;
+ enum dml2_pstate_change_support FCLKChangeSupport[DML2_MAX_PLANES];
+ enum dml2_pstate_change_support DRAMClockChangeSupport[DML2_MAX_PLANES];
bool global_dram_clock_change_supported;
bool global_fclk_change_supported;
double MaxActiveDRAMClockChangeLatencySupported[DML2_MAX_PLANES];
@@ -846,6 +850,8 @@ struct dml2_core_internal_mode_program {
bool mall_comb_mcache_l[DML2_MAX_PLANES];
bool mall_comb_mcache_c[DML2_MAX_PLANES];
bool lc_comb_mcache[DML2_MAX_PLANES];
+
+ double impacted_prefetch_margin_us[DML2_MAX_PLANES];
};
struct dml2_core_internal_SOCParametersList {
@@ -862,6 +868,7 @@ struct dml2_core_internal_SOCParametersList {
double USRRetrainingLatency;
double SMNLatency;
double g6_temp_read_blackout_us;
+ double temp_read_or_ppt_blackout_us;
double max_urgent_latency_us;
double df_response_time_us;
enum dml2_qos_param_type qos_type;
@@ -951,6 +958,7 @@ struct dml2_core_calcs_mode_support_locals {
unsigned int tdlut_groups_per_2row_ub[DML2_MAX_PLANES];
double tdlut_opt_time[DML2_MAX_PLANES];
double tdlut_drain_time[DML2_MAX_PLANES];
+ unsigned int tdlut_bytes_to_deliver[DML2_MAX_PLANES];
unsigned int tdlut_bytes_per_group[DML2_MAX_PLANES];
unsigned int cursor_bytes_per_chunk[DML2_MAX_PLANES];
@@ -961,6 +969,18 @@ struct dml2_core_calcs_mode_support_locals {
unsigned int pstate_bytes_required_l[DML2_MAX_PLANES];
unsigned int pstate_bytes_required_c[DML2_MAX_PLANES];
+
+ double prefetch_sw_bytes[DML2_MAX_PLANES];
+ double Tpre_rounded[DML2_MAX_PLANES];
+ double Tpre_oto[DML2_MAX_PLANES];
+ bool recalc_prefetch_schedule;
+ bool recalc_prefetch_done;
+ double impacted_dst_y_pre[DML2_MAX_PLANES];
+ double line_times[DML2_MAX_PLANES];
+ enum dml2_source_format_class pixel_format[DML2_MAX_PLANES];
+ unsigned int lb_source_lines_l[DML2_MAX_PLANES];
+ unsigned int lb_source_lines_c[DML2_MAX_PLANES];
+ double prefetch_swath_time_us[DML2_MAX_PLANES];
};
struct dml2_core_calcs_mode_programming_locals {
@@ -1024,6 +1044,7 @@ struct dml2_core_calcs_mode_programming_locals {
unsigned int tdlut_groups_per_2row_ub[DML2_MAX_PLANES];
double tdlut_opt_time[DML2_MAX_PLANES];
double tdlut_drain_time[DML2_MAX_PLANES];
+ unsigned int tdlut_bytes_to_deliver[DML2_MAX_PLANES];
unsigned int tdlut_bytes_per_group[DML2_MAX_PLANES];
unsigned int cursor_bytes_per_chunk[DML2_MAX_PLANES];
@@ -1041,6 +1062,16 @@ struct dml2_core_calcs_mode_programming_locals {
unsigned int pstate_bytes_required_l[DML2_MAX_PLANES];
unsigned int pstate_bytes_required_c[DML2_MAX_PLANES];
+
+ double prefetch_sw_bytes[DML2_MAX_PLANES];
+ double Tpre_rounded[DML2_MAX_PLANES];
+ double Tpre_oto[DML2_MAX_PLANES];
+ bool recalc_prefetch_schedule;
+ double impacted_dst_y_pre[DML2_MAX_PLANES];
+ double line_times[DML2_MAX_PLANES];
+ enum dml2_source_format_class pixel_format[DML2_MAX_PLANES];
+ unsigned int lb_source_lines_l[DML2_MAX_PLANES];
+ unsigned int lb_source_lines_c[DML2_MAX_PLANES];
};
struct dml2_core_calcs_CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_locals {
@@ -1048,6 +1079,7 @@ struct dml2_core_calcs_CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_local
double ActiveFCLKChangeLatencyMargin[DML2_MAX_PLANES];
double USRRetrainingLatencyMargin[DML2_MAX_PLANES];
double g6_temp_read_latency_margin[DML2_MAX_PLANES];
+ double temp_read_or_ppt_latency_margin[DML2_MAX_PLANES];
double EffectiveLBLatencyHidingY;
double EffectiveLBLatencyHidingC;
@@ -1185,17 +1217,14 @@ struct dml2_core_calcs_CalculatePrefetchSchedule_locals {
double LineTime;
double dst_y_prefetch_equ;
double prefetch_bw_oto;
+ double per_pipe_vactive_sw_bw;
double Tvm_oto;
double Tr0_oto;
- double Tvm_no_trip_oto;
- double Tr0_no_trip_oto;
double Tvm_oto_lines;
double Tr0_oto_lines;
double dst_y_prefetch_oto;
double TimeForFetchingVM;
double TimeForFetchingRowInVBlank;
- double dst_y_per_vm_no_trip_vblank;
- double dst_y_per_row_no_trip_vblank;
double LinesToRequestPrefetchPixelData;
unsigned int HostVMDynamicLevelsTrips;
double trip_to_mem;
@@ -1203,15 +1232,12 @@ struct dml2_core_calcs_CalculatePrefetchSchedule_locals {
double Tr0_trips_rounded;
double max_Tsw;
double Lsw_oto;
- double Lsw_equ;
- double Tpre_rounded;
double prefetch_bw_equ;
double Tvm_equ;
double Tr0_equ;
double Tdmbf;
double Tdmec;
double Tdmsks;
- double prefetch_sw_bytes;
double total_row_bytes;
double prefetch_bw_pr;
double bytes_pp;
@@ -1225,6 +1251,7 @@ struct dml2_core_calcs_CalculatePrefetchSchedule_locals {
double prefetch_bw2;
double prefetch_bw3;
double prefetch_bw4;
+ double dst_y_prefetch_equ_impacted;
double TWait_p;
unsigned int cursor_prefetch_bytes;
@@ -1545,17 +1572,18 @@ struct dml2_core_calcs_CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_param
// Output
struct dml2_core_internal_watermarks *Watermark;
- enum dml2_dram_clock_change_support *DRAMClockChangeSupport;
+ enum dml2_pstate_change_support *DRAMClockChangeSupport;
bool *global_dram_clock_change_supported;
double *MaxActiveDRAMClockChangeLatencySupported;
unsigned int *SubViewportLinesNeededInMALL;
- enum dml2_fclock_change_support *FCLKChangeSupport;
+ enum dml2_pstate_change_support *FCLKChangeSupport;
bool *global_fclk_change_supported;
double *MaxActiveFCLKChangeLatencySupported;
bool *USRRetrainingSupport;
double *VActiveLatencyHidingMargin;
double *VActiveLatencyHidingUs;
bool *g6_temp_read_support;
+ bool *temp_read_or_ppt_support;
};
@@ -1727,8 +1755,8 @@ struct dml2_core_calcs_CalculatePrefetchSchedule_params {
double PrefetchSourceLinesC;
unsigned int VInitPreFillC;
unsigned int MaxNumSwathC;
- unsigned int swath_width_luma_ub;
- unsigned int swath_width_chroma_ub;
+ unsigned int swath_width_luma_ub; // per-pipe
+ unsigned int swath_width_chroma_ub; // per-pipe
unsigned int SwathHeightY;
unsigned int SwathHeightC;
double TWait;
@@ -1750,6 +1778,10 @@ struct dml2_core_calcs_CalculatePrefetchSchedule_params {
unsigned int meta_row_bytes;
double mall_prefetch_sdp_overhead_factor;
+ double impacted_dst_y_pre;
+ double vactive_sw_bw_l; // per surface bw
+ double vactive_sw_bw_c; // per surface bw
+
// output
unsigned int *DSTXAfterScaler;
unsigned int *DSTYAfterScaler;
@@ -1767,6 +1799,8 @@ struct dml2_core_calcs_CalculatePrefetchSchedule_params {
double *Tdmdl_vm;
double *Tdmdl;
double *TSetup;
+ double *Tpre_rounded;
+ double *Tpre_oto;
double *Tvm_trips;
double *Tr0_trips;
double *Tvm_trips_flip;
@@ -1777,6 +1811,48 @@ struct dml2_core_calcs_CalculatePrefetchSchedule_params {
unsigned int *VUpdateWidthPix;
unsigned int *VReadyOffsetPix;
double *prefetch_cursor_bw;
+ double *prefetch_sw_bytes;
+ double *prefetch_swath_time_us;
+};
+
+struct dml2_core_calcs_CheckGlobalPrefetchAdmissibility_params {
+ unsigned int num_active_planes;
+ enum dml2_source_format_class *pixel_format;
+ unsigned int rob_buffer_size_kbytes;
+ unsigned int compressed_buffer_size_kbytes;
+ unsigned int chunk_bytes_l; // same for all planes
+ unsigned int chunk_bytes_c;
+ unsigned int *detile_buffer_size_bytes_l;
+ unsigned int *detile_buffer_size_bytes_c;
+ unsigned int *full_swath_bytes_l;
+ unsigned int *full_swath_bytes_c;
+ unsigned int *lb_source_lines_l;
+ unsigned int *lb_source_lines_c;
+ unsigned int *swath_height_l;
+ unsigned int *swath_height_c;
+ double *prefetch_sw_bytes;
+ double *Tpre_rounded;
+ double *Tpre_oto;
+ double estimated_dcfclk_mhz;
+ double estimated_urg_bandwidth_required_mbps;
+ double *line_time;
+ double *dst_y_prefetch;
+
+ // output
+ bool *recalc_prefetch_schedule;
+ double *impacted_dst_y_pre;
+};
+
+struct dml2_core_calcs_CheckGlobalPrefetchAdmissibility_locals {
+ unsigned int max_Trpd_dcfclk_cycles;
+ unsigned int burst_bytes_to_fill_det;
+ double time_to_fill_det_us;
+ unsigned int accumulated_return_path_dcfclk_cycles[DML2_MAX_PLANES];
+ bool prefetch_global_check_passed;
+ unsigned int src_swath_bytes_l[DML2_MAX_PLANES];
+ unsigned int src_swath_bytes_c[DML2_MAX_PLANES];
+ unsigned int src_detile_buf_size_bytes_l[DML2_MAX_PLANES];
+ unsigned int src_detile_buf_size_bytes_c[DML2_MAX_PLANES];
};
struct dml2_core_calcs_calculate_mcache_row_bytes_params {
@@ -1921,6 +1997,7 @@ struct dml2_core_calcs_calculate_tdlut_setting_params {
unsigned int *tdlut_groups_per_2row_ub;
double *tdlut_opt_time;
double *tdlut_drain_time;
+ unsigned int *tdlut_bytes_to_deliver;
unsigned int *tdlut_bytes_per_group;
};
@@ -2004,6 +2081,7 @@ struct dml2_core_internal_scratch {
struct dml2_core_calcs_CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_locals CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport_locals;
struct dml2_core_calcs_CalculateVMRowAndSwath_locals CalculateVMRowAndSwath_locals;
struct dml2_core_calcs_CalculatePrefetchSchedule_locals CalculatePrefetchSchedule_locals;
+ struct dml2_core_calcs_CheckGlobalPrefetchAdmissibility_locals CheckGlobalPrefetchAdmissibility_locals;
struct dml2_core_shared_CalculateSwathAndDETConfiguration_locals CalculateSwathAndDETConfiguration_locals;
struct dml2_core_shared_TruncToValidBPP_locals TruncToValidBPP_locals;
struct dml2_core_shared_CalculateDETBufferSize_locals CalculateDETBufferSize_locals;
@@ -2019,6 +2097,7 @@ struct dml2_core_internal_scratch {
struct dml2_core_calcs_CalculateSwathAndDETConfiguration_params CalculateSwathAndDETConfiguration_params;
struct dml2_core_calcs_CalculateStutterEfficiency_params CalculateStutterEfficiency_params;
struct dml2_core_calcs_CalculatePrefetchSchedule_params CalculatePrefetchSchedule_params;
+ struct dml2_core_calcs_CheckGlobalPrefetchAdmissibility_params CheckGlobalPrefetchAdmissibility_params;
struct dml2_core_calcs_calculate_mcache_setting_params calculate_mcache_setting_params;
struct dml2_core_calcs_calculate_tdlut_setting_params calculate_tdlut_setting_params;
struct dml2_core_shared_calculate_vm_and_row_bytes_params calculate_vm_and_row_bytes_params;
@@ -2038,7 +2117,6 @@ struct dml2_core_internal_display_mode_lib {
// Used to hold input; intermediate and output of the calculations
struct dml2_core_internal_mode_support ms; // struct for mode support
struct dml2_core_internal_mode_program mp; // struct for mode programming
-
// Available overridable calculators for core_shared.
// if null, core_shared will use default calculators.
struct dml2_core_shared_calculation_funcs funcs;
@@ -2051,7 +2129,6 @@ struct dml2_core_calcs_mode_support_ex {
const struct dml2_display_cfg *in_display_cfg;
const struct dml2_mcg_min_clock_table *min_clk_table;
int min_clk_index;
-
//unsigned int in_state_index;
struct dml2_core_internal_mode_support_info *out_evaluation_info;
};
@@ -2064,9 +2141,7 @@ struct dml2_core_calcs_mode_programming_ex {
const struct dml2_mcg_min_clock_table *min_clk_table;
const struct core_display_cfg_support_info *cfg_support_info;
int min_clk_index;
-
struct dml2_display_cfg_programming *programming;
-
};
#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_utils.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_utils.c
index 714b5c39b7e6..456b3f8a6d38 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_utils.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_utils.c
@@ -63,6 +63,150 @@ bool dml2_core_utils_is_420(enum dml2_source_format_class source_format)
case dml2_mono_16:
val = 0;
break;
+ case dml2_422_planar_8:
+ val = 0;
+ break;
+ case dml2_422_planar_10:
+ val = 0;
+ break;
+ case dml2_422_planar_12:
+ val = 0;
+ break;
+ case dml2_422_packed_8:
+ val = 0;
+ break;
+ case dml2_422_packed_10:
+ val = 0;
+ break;
+ case dml2_422_packed_12:
+ val = 0;
+ break;
+ default:
+ DML2_ASSERT(0);
+ break;
+ }
+ return val;
+}
+
+bool dml2_core_utils_is_422_planar(enum dml2_source_format_class source_format)
+{
+ bool val = false;
+
+ switch (source_format) {
+ case dml2_444_8:
+ val = 0;
+ break;
+ case dml2_444_16:
+ val = 0;
+ break;
+ case dml2_444_32:
+ val = 0;
+ break;
+ case dml2_444_64:
+ val = 0;
+ break;
+ case dml2_420_8:
+ val = 0;
+ break;
+ case dml2_420_10:
+ val = 0;
+ break;
+ case dml2_420_12:
+ val = 0;
+ break;
+ case dml2_rgbe_alpha:
+ val = 0;
+ break;
+ case dml2_rgbe:
+ val = 0;
+ break;
+ case dml2_mono_8:
+ val = 0;
+ break;
+ case dml2_mono_16:
+ val = 0;
+ break;
+ case dml2_422_planar_8:
+ val = 1;
+ break;
+ case dml2_422_planar_10:
+ val = 1;
+ break;
+ case dml2_422_planar_12:
+ val = 1;
+ break;
+ case dml2_422_packed_8:
+ val = 0;
+ break;
+ case dml2_422_packed_10:
+ val = 0;
+ break;
+ case dml2_422_packed_12:
+ val = 0;
+ break;
+ default:
+ DML2_ASSERT(0);
+ break;
+ }
+ return val;
+}
+
+bool dml2_core_utils_is_422_packed(enum dml2_source_format_class source_format)
+{
+ bool val = false;
+
+ switch (source_format) {
+ case dml2_444_8:
+ val = 0;
+ break;
+ case dml2_444_16:
+ val = 0;
+ break;
+ case dml2_444_32:
+ val = 0;
+ break;
+ case dml2_444_64:
+ val = 0;
+ break;
+ case dml2_420_8:
+ val = 0;
+ break;
+ case dml2_420_10:
+ val = 0;
+ break;
+ case dml2_420_12:
+ val = 0;
+ break;
+ case dml2_rgbe_alpha:
+ val = 0;
+ break;
+ case dml2_rgbe:
+ val = 0;
+ break;
+ case dml2_mono_8:
+ val = 0;
+ break;
+ case dml2_mono_16:
+ val = 0;
+ break;
+ case dml2_422_planar_8:
+ val = 0;
+ break;
+ case dml2_422_planar_10:
+ val = 0;
+ break;
+ case dml2_422_planar_12:
+ val = 0;
+ break;
+ case dml2_422_packed_8:
+ val = 1;
+ break;
+ case dml2_422_packed_10:
+ val = 1;
+ break;
+ case dml2_422_packed_12:
+ val = 1;
+ break;
default:
DML2_ASSERT(0);
break;
@@ -154,9 +298,9 @@ void dml2_core_utils_print_mode_support_info(const struct dml2_core_internal_mod
dml2_printf("DML: support: DynamicMetadataSupported = %d\n", support->DynamicMetadataSupported);
if (!fail_only || support->VRatioInPrefetchSupported == 0)
dml2_printf("DML: support: VRatioInPrefetchSupported = %d\n", support->VRatioInPrefetchSupported);
- if (!fail_only || support->PTEBufferSizeNotExceeded == 1)
+ if (!fail_only || support->PTEBufferSizeNotExceeded == 0)
dml2_printf("DML: support: PTEBufferSizeNotExceeded = %d\n", support->PTEBufferSizeNotExceeded);
- if (!fail_only || support->DCCMetaBufferSizeNotExceeded == 1)
+ if (!fail_only || support->DCCMetaBufferSizeNotExceeded == 0)
dml2_printf("DML: support: DCCMetaBufferSizeNotExceeded = %d\n", support->DCCMetaBufferSizeNotExceeded);
if (!fail_only || support->ExceededMALLSize == 1)
dml2_printf("DML: support: ExceededMALLSize = %d\n", support->ExceededMALLSize);
@@ -280,39 +424,49 @@ bool dml2_core_utils_is_phantom_pipe(const struct dml2_plane_parameters *plane_c
return is_phantom;
}
-unsigned int dml2_core_utils_get_tile_block_size_bytes(enum dml2_swizzle_mode sw_mode)
-{
- switch (sw_mode) {
- case (dml2_sw_linear):
- return 256; break;
- case (dml2_sw_256b_2d):
- return 256; break;
- case (dml2_sw_4kb_2d):
- return 4096; break;
- case (dml2_sw_64kb_2d):
- return 65536; break;
- case (dml2_sw_256kb_2d):
- return 262144; break;
- case (dml2_gfx11_sw_linear):
- return 256; break;
- case (dml2_gfx11_sw_64kb_d):
- return 65536; break;
- case (dml2_gfx11_sw_64kb_d_t):
- return 65536; break;
- case (dml2_gfx11_sw_64kb_d_x):
- return 65536; break;
- case (dml2_gfx11_sw_64kb_r_x):
- return 65536; break;
- case (dml2_gfx11_sw_256kb_d_x):
- return 262144; break;
- case (dml2_gfx11_sw_256kb_r_x):
- return 262144; break;
- default:
+unsigned int dml2_core_utils_get_tile_block_size_bytes(enum dml2_swizzle_mode sw_mode, unsigned int byte_per_pixel)
+{
+
+ if (sw_mode == dml2_sw_linear)
+ return 256;
+ else if (sw_mode == dml2_sw_256b_2d)
+ return 256;
+ else if (sw_mode == dml2_sw_4kb_2d)
+ return 4096;
+ else if (sw_mode == dml2_sw_64kb_2d)
+ return 65536;
+ else if (sw_mode == dml2_sw_256kb_2d)
+ return 262144;
+ else if (sw_mode == dml2_gfx11_sw_linear)
+ return 256;
+ else if (sw_mode == dml2_gfx11_sw_64kb_d)
+ return 65536;
+ else if (sw_mode == dml2_gfx11_sw_64kb_d_t)
+ return 65536;
+ else if (sw_mode == dml2_gfx11_sw_64kb_d_x)
+ return 65536;
+ else if (sw_mode == dml2_gfx11_sw_64kb_r_x)
+ return 65536;
+ else if (sw_mode == dml2_gfx11_sw_256kb_d_x)
+ return 262144;
+ else if (sw_mode == dml2_gfx11_sw_256kb_r_x)
+ return 262144;
+ else {
DML2_ASSERT(0);
return 256;
};
}
+bool dml2_core_utils_get_segment_horizontal_contiguous(enum dml2_swizzle_mode sw_mode, unsigned int byte_per_pixel)
+{
+ return (byte_per_pixel != 2);
+}
+
+bool dml2_core_utils_is_linear(enum dml2_swizzle_mode sw_mode)
+{
+ return (sw_mode == dml2_sw_linear || sw_mode == dml2_sw_linear_256b || sw_mode == dml2_linear_64elements);
+};
+
bool dml2_core_utils_is_vertical_rotation(enum dml2_rotation_angle Scan)
{
@@ -325,7 +479,6 @@ bool dml2_core_utils_is_vertical_rotation(enum dml2_rotation_angle Scan)
return is_vert;
}
-
int unsigned dml2_core_utils_get_gfx_version(enum dml2_swizzle_mode sw_mode)
{
int unsigned version = 0;
@@ -334,17 +487,17 @@ int unsigned dml2_core_utils_get_gfx_version(enum dml2_swizzle_mode sw_mode)
sw_mode == dml2_sw_256b_2d ||
sw_mode == dml2_sw_4kb_2d ||
sw_mode == dml2_sw_64kb_2d ||
- sw_mode == dml2_sw_256kb_2d) {
+ sw_mode == dml2_sw_256kb_2d)
version = 12;
- } else if (sw_mode == dml2_gfx11_sw_linear ||
+ else if (sw_mode == dml2_gfx11_sw_linear ||
sw_mode == dml2_gfx11_sw_64kb_d ||
sw_mode == dml2_gfx11_sw_64kb_d_t ||
sw_mode == dml2_gfx11_sw_64kb_d_x ||
sw_mode == dml2_gfx11_sw_64kb_r_x ||
sw_mode == dml2_gfx11_sw_256kb_d_x ||
- sw_mode == dml2_gfx11_sw_256kb_r_x) {
+ sw_mode == dml2_gfx11_sw_256kb_r_x)
version = 11;
- } else {
+ else {
dml2_printf("ERROR: Invalid sw_mode setting! val=%u\n", sw_mode);
DML2_ASSERT(0);
}
@@ -403,7 +556,7 @@ bool dml2_core_utils_is_dual_plane(enum dml2_source_format_class source_format)
{
bool ret_val = 0;
- if ((source_format == dml2_420_12) || (source_format == dml2_420_8) || (source_format == dml2_420_10) || (source_format == dml2_rgbe_alpha))
+ if (dml2_core_utils_is_420(source_format) || dml2_core_utils_is_422_planar(source_format) || (source_format == dml2_rgbe_alpha))
ret_val = 1;
return ret_val;
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_utils.h b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_utils.h
index a5cc6a07167a..95f0d017add4 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_utils.h
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_utils.h
@@ -11,6 +11,8 @@
double dml2_core_utils_div_rem(double dividend, unsigned int divisor, unsigned int *remainder);
const char *dml2_core_utils_internal_bw_type_str(enum dml2_core_internal_bw_type bw_type);
bool dml2_core_utils_is_420(enum dml2_source_format_class source_format);
+bool dml2_core_utils_is_422_planar(enum dml2_source_format_class source_format);
+bool dml2_core_utils_is_422_packed(enum dml2_source_format_class source_format);
void dml2_core_utils_print_mode_support_info(const struct dml2_core_internal_mode_support_info *support, bool fail_only);
const char *dml2_core_utils_internal_soc_state_type_str(enum dml2_core_internal_soc_state_type dml2_core_internal_soc_state_type);
void dml2_core_utils_get_stream_output_bpp(double *out_bpp, const struct dml2_display_cfg *display_cfg);
@@ -18,8 +20,10 @@ unsigned int dml2_core_utils_round_to_multiple(unsigned int num, unsigned int mu
unsigned int dml2_core_util_get_num_active_pipes(int unsigned num_planes, const struct core_display_cfg_support_info *cfg_support_info);
void dml2_core_utils_pipe_plane_mapping(const struct core_display_cfg_support_info *cfg_support_info, unsigned int *pipe_plane);
bool dml2_core_utils_is_phantom_pipe(const struct dml2_plane_parameters *plane_cfg);
-unsigned int dml2_core_utils_get_tile_block_size_bytes(enum dml2_swizzle_mode sw_mode);
+unsigned int dml2_core_utils_get_tile_block_size_bytes(enum dml2_swizzle_mode sw_mode, unsigned int byte_per_pixel);
+bool dml2_core_utils_get_segment_horizontal_contiguous(enum dml2_swizzle_mode sw_mode, unsigned int byte_per_pixel);
bool dml2_core_utils_is_vertical_rotation(enum dml2_rotation_angle Scan);
+bool dml2_core_utils_is_linear(enum dml2_swizzle_mode sw_mode);
int unsigned dml2_core_utils_get_gfx_version(enum dml2_swizzle_mode sw_mode);
unsigned int dml2_core_utils_get_qos_param_index(unsigned long uclk_freq_khz, const struct dml2_dcn4_uclk_dpm_dependent_qos_params *per_uclk_dpm_params);
unsigned int dml2_core_utils_get_active_min_uclk_dpm_index(unsigned long uclk_freq_khz, const struct dml2_soc_state_table *clk_table);
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_dpmm/dml2_dpmm_dcn4.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_dpmm/dml2_dpmm_dcn4.c
index 8869ea089312..fc77fb34a19a 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_dpmm/dml2_dpmm_dcn4.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_dpmm/dml2_dpmm_dcn4.c
@@ -96,6 +96,7 @@ static void calculate_svp_prefetch_minimums(struct dml2_dpmm_map_mode_to_soc_dpm
double min_uclk_latency;
const struct dml2_core_mode_support_result *mode_support_result = &in_out->display_cfg->mode_support_result;
+ /* assumes DF throttling is enabled */
min_uclk_avg = dram_bw_kbps_to_uclk_khz(mode_support_result->global.svp_prefetch.average_bw_dram_kbps, &in_out->soc_bb->clk_table.dram_config);
min_uclk_avg = (double)min_uclk_avg / ((double)in_out->soc_bb->qos_parameters.derate_table.dcn_mall_prefetch_average.dram_derate_percent_pixel / 100);
@@ -125,6 +126,37 @@ static void calculate_svp_prefetch_minimums(struct dml2_dpmm_map_mode_to_soc_dpm
in_out->programming->min_clocks.dcn4x.svp_prefetch.uclk_khz = dml_round_up(min_uclk_bw > min_uclk_latency ? min_uclk_bw : min_uclk_latency);
in_out->programming->min_clocks.dcn4x.svp_prefetch.fclk_khz = dml_round_up(min_fclk_bw > min_fclk_latency ? min_fclk_bw : min_fclk_latency);
in_out->programming->min_clocks.dcn4x.svp_prefetch.dcfclk_khz = dml_round_up(min_dcfclk_bw > min_dcfclk_latency ? min_dcfclk_bw : min_dcfclk_latency);
+
+ /* assumes DF throttling is disabled */
+ min_uclk_avg = dram_bw_kbps_to_uclk_khz(mode_support_result->global.svp_prefetch.average_bw_dram_kbps, &in_out->soc_bb->clk_table.dram_config);
+ min_uclk_avg = (double)min_uclk_avg / ((double)in_out->soc_bb->qos_parameters.derate_table.system_active_average.dram_derate_percent_pixel / 100);
+
+ min_uclk_urgent = dram_bw_kbps_to_uclk_khz(mode_support_result->global.svp_prefetch.urgent_bw_dram_kbps, &in_out->soc_bb->clk_table.dram_config);
+ min_uclk_urgent = (double)min_uclk_urgent / ((double)in_out->soc_bb->qos_parameters.derate_table.system_active_urgent.dram_derate_percent_pixel / 100);
+
+ min_uclk_bw = min_uclk_urgent > min_uclk_avg ? min_uclk_urgent : min_uclk_avg;
+
+ min_fclk_avg = (double)mode_support_result->global.svp_prefetch.average_bw_sdp_kbps / in_out->soc_bb->fabric_datapath_to_dcn_data_return_bytes;
+ min_fclk_avg = (double)min_fclk_avg / ((double)in_out->soc_bb->qos_parameters.derate_table.system_active_average.fclk_derate_percent / 100);
+
+ min_fclk_urgent = (double)mode_support_result->global.svp_prefetch.urgent_bw_sdp_kbps / in_out->soc_bb->fabric_datapath_to_dcn_data_return_bytes;
+ min_fclk_urgent = (double)min_fclk_urgent / ((double)in_out->soc_bb->qos_parameters.derate_table.system_active_urgent.fclk_derate_percent / 100);
+
+ min_fclk_bw = min_fclk_urgent > min_fclk_avg ? min_fclk_urgent : min_fclk_avg;
+
+ min_dcfclk_avg = (double)mode_support_result->global.svp_prefetch.average_bw_sdp_kbps / in_out->soc_bb->return_bus_width_bytes;
+ min_dcfclk_avg = (double)min_dcfclk_avg / ((double)in_out->soc_bb->qos_parameters.derate_table.system_active_average.dcfclk_derate_percent / 100);
+
+ min_dcfclk_urgent = (double)mode_support_result->global.svp_prefetch.urgent_bw_sdp_kbps / in_out->soc_bb->return_bus_width_bytes;
+ min_dcfclk_urgent = (double)min_dcfclk_urgent / ((double)in_out->soc_bb->qos_parameters.derate_table.system_active_urgent.dcfclk_derate_percent / 100);
+
+ min_dcfclk_bw = min_dcfclk_urgent > min_dcfclk_avg ? min_dcfclk_urgent : min_dcfclk_avg;
+
+ get_minimum_clocks_for_latency(in_out, &min_uclk_latency, &min_fclk_latency, &min_dcfclk_latency);
+
+ in_out->programming->min_clocks.dcn4x.svp_prefetch_no_throttle.uclk_khz = dml_round_up(min_uclk_bw > min_uclk_latency ? min_uclk_bw : min_uclk_latency);
+ in_out->programming->min_clocks.dcn4x.svp_prefetch_no_throttle.fclk_khz = dml_round_up(min_fclk_bw > min_fclk_latency ? min_fclk_bw : min_fclk_latency);
+ in_out->programming->min_clocks.dcn4x.svp_prefetch_no_throttle.dcfclk_khz = dml_round_up(min_dcfclk_bw > min_dcfclk_latency ? min_dcfclk_bw : min_dcfclk_latency);
}
static void calculate_idle_minimums(struct dml2_dpmm_map_mode_to_soc_dpm_params_in_out *in_out)
@@ -272,6 +304,17 @@ static bool map_soc_min_clocks_to_dpm_fine_grained(struct dml2_display_cfg_progr
if (result)
result = round_up_to_next_dpm(&display_cfg->min_clocks.dcn4x.idle.uclk_khz, &state_table->uclk);
+ /* these clocks are optional, so they can fail to map, in which case map all to 0 */
+ if (result) {
+ if (!round_up_to_next_dpm(&display_cfg->min_clocks.dcn4x.svp_prefetch_no_throttle.dcfclk_khz, &state_table->dcfclk) ||
+ !round_up_to_next_dpm(&display_cfg->min_clocks.dcn4x.svp_prefetch_no_throttle.fclk_khz, &state_table->fclk) ||
+ !round_up_to_next_dpm(&display_cfg->min_clocks.dcn4x.svp_prefetch_no_throttle.uclk_khz, &state_table->uclk)) {
+ display_cfg->min_clocks.dcn4x.svp_prefetch_no_throttle.dcfclk_khz = 0;
+ display_cfg->min_clocks.dcn4x.svp_prefetch_no_throttle.fclk_khz = 0;
+ display_cfg->min_clocks.dcn4x.svp_prefetch_no_throttle.uclk_khz = 0;
+ }
+ }
+
return result;
}
@@ -374,11 +417,11 @@ static bool map_min_clocks_to_dpm(const struct dml2_core_mode_support_result *mo
static bool are_timings_trivially_synchronizable(struct dml2_display_cfg *display_config, int mask)
{
- unsigned char i;
+ unsigned int i;
bool identical = true;
bool contains_drr = false;
- unsigned char remap_array[DML2_MAX_PLANES];
- unsigned char remap_array_size = 0;
+ unsigned int remap_array[DML2_MAX_PLANES];
+ unsigned int remap_array_size = 0;
// Create a remap array to enable simple iteration through only masked stream indicies
for (i = 0; i < display_config->num_streams; i++) {
@@ -413,10 +456,10 @@ static bool are_timings_trivially_synchronizable(struct dml2_display_cfg *displa
static int find_smallest_idle_time_in_vblank_us(struct dml2_dpmm_map_mode_to_soc_dpm_params_in_out *in_out, int mask)
{
- unsigned char i;
+ unsigned int i;
int min_idle_us = 0;
- unsigned char remap_array[DML2_MAX_PLANES];
- unsigned char remap_array_size = 0;
+ unsigned int remap_array[DML2_MAX_PLANES];
+ unsigned int remap_array_size = 0;
const struct dml2_core_mode_support_result *mode_support_result = &in_out->display_cfg->mode_support_result;
// Create a remap array to enable simple iteration through only masked stream indicies
@@ -711,7 +754,7 @@ bool dpmm_dcn4_map_watermarks(struct dml2_dpmm_map_watermarks_params_in_out *in_
dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_A].fclk_pstate = (int unsigned)(mode_lib->mp.Watermark.FCLKChangeWatermark * refclk_freq_in_mhz);
dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_A].sr_enter = (int unsigned)(mode_lib->mp.Watermark.StutterEnterPlusExitWatermark * refclk_freq_in_mhz);
dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_A].sr_exit = (int unsigned)(mode_lib->mp.Watermark.StutterExitWatermark * refclk_freq_in_mhz);
- dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_A].temp_read_or_ppt = (int unsigned)(mode_lib->mp.Watermark.g6_temp_read_watermark_us * refclk_freq_in_mhz);
+ dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_A].temp_read_or_ppt = (int unsigned)(mode_lib->mp.Watermark.temp_read_or_ppt_watermark_us * refclk_freq_in_mhz);
dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_A].uclk_pstate = (int unsigned)(mode_lib->mp.Watermark.DRAMClockChangeWatermark * refclk_freq_in_mhz);
dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_A].urgent = (int unsigned)(mode_lib->mp.Watermark.UrgentWatermark * refclk_freq_in_mhz);
dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_A].usr = (int unsigned)(mode_lib->mp.Watermark.USRRetrainingWatermark * refclk_freq_in_mhz);
@@ -725,7 +768,7 @@ bool dpmm_dcn4_map_watermarks(struct dml2_dpmm_map_watermarks_params_in_out *in_
dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_B].fclk_pstate = (int unsigned)(mode_lib->mp.Watermark.FCLKChangeWatermark * refclk_freq_in_mhz);
dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_B].sr_enter = (int unsigned)(mode_lib->mp.Watermark.StutterEnterPlusExitWatermark * refclk_freq_in_mhz);
dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_B].sr_exit = (int unsigned)(mode_lib->mp.Watermark.StutterExitWatermark * refclk_freq_in_mhz);
- dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_B].temp_read_or_ppt = (int unsigned)(mode_lib->mp.Watermark.g6_temp_read_watermark_us * refclk_freq_in_mhz);
+ dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_B].temp_read_or_ppt = (int unsigned)(mode_lib->mp.Watermark.temp_read_or_ppt_watermark_us * refclk_freq_in_mhz);
dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_B].uclk_pstate = (int unsigned)(mode_lib->mp.Watermark.DRAMClockChangeWatermark * refclk_freq_in_mhz);
dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_B].urgent = (int unsigned)(mode_lib->mp.Watermark.UrgentWatermark * refclk_freq_in_mhz);
dchubbub_regs->wm_regs[DML2_DCHUB_WATERMARK_SET_B].usr = (int unsigned)(mode_lib->mp.Watermark.USRRetrainingWatermark * refclk_freq_in_mhz);
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn3.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn3.c
index a31db5742675..e763c8e45da8 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn3.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn3.c
@@ -195,11 +195,11 @@ static int count_planes_with_stream_index(const struct dml2_display_cfg *display
static bool are_timings_trivially_synchronizable(struct display_configuation_with_meta *display_config, int mask)
{
- unsigned char i;
+ unsigned int i;
bool identical = true;
bool contains_drr = false;
- unsigned char remap_array[DML2_MAX_PLANES];
- unsigned char remap_array_size = 0;
+ unsigned int remap_array[DML2_MAX_PLANES];
+ unsigned int remap_array_size = 0;
// Create a remap array to enable simple iteration through only masked stream indicies
for (i = 0; i < display_config->display_config.num_streams; i++) {
@@ -347,8 +347,12 @@ static int find_highest_odm_load_stream_index(
int odm_load, highest_odm_load = -1, highest_odm_load_index = -1;
for (i = 0; i < display_config->num_streams; i++) {
- odm_load = display_config->stream_descriptors[i].timing.pixel_clock_khz
+ if (mode_support_result->cfg_support_info.stream_support_info[i].odms_used > 0)
+ odm_load = display_config->stream_descriptors[i].timing.pixel_clock_khz
/ mode_support_result->cfg_support_info.stream_support_info[i].odms_used;
+ else
+ odm_load = 0;
+
if (odm_load > highest_odm_load) {
highest_odm_load_index = i;
highest_odm_load = odm_load;
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn4_fams2.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn4_fams2.c
index 92269f0e50ed..a3324f7b9ba6 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn4_fams2.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn4_fams2.c
@@ -13,32 +13,32 @@ static const double MIN_BLANK_STUTTER_FACTOR = 3.0;
static const struct dml2_pmo_pstate_strategy base_strategy_list_1_display[] = {
// VActive Preferred
{
- .per_stream_pstate_method = { dml2_pmo_pstate_strategy_vactive, dml2_pmo_pstate_strategy_na, dml2_pmo_pstate_strategy_na, dml2_pmo_pstate_strategy_na },
+ .per_stream_pstate_method = { dml2_pstate_method_vactive, dml2_pstate_method_na, dml2_pstate_method_na, dml2_pstate_method_na },
.allow_state_increase = true,
},
// Then SVP
{
- .per_stream_pstate_method = { dml2_pmo_pstate_strategy_fw_svp, dml2_pmo_pstate_strategy_na, dml2_pmo_pstate_strategy_na, dml2_pmo_pstate_strategy_na },
+ .per_stream_pstate_method = { dml2_pstate_method_fw_svp, dml2_pstate_method_na, dml2_pstate_method_na, dml2_pstate_method_na },
.allow_state_increase = true,
},
// Then VBlank
{
- .per_stream_pstate_method = { dml2_pmo_pstate_strategy_vblank, dml2_pmo_pstate_strategy_na, dml2_pmo_pstate_strategy_na, dml2_pmo_pstate_strategy_na },
+ .per_stream_pstate_method = { dml2_pstate_method_vblank, dml2_pstate_method_na, dml2_pstate_method_na, dml2_pstate_method_na },
.allow_state_increase = false,
},
// Then DRR
{
- .per_stream_pstate_method = { dml2_pmo_pstate_strategy_fw_drr, dml2_pmo_pstate_strategy_na, dml2_pmo_pstate_strategy_na, dml2_pmo_pstate_strategy_na },
+ .per_stream_pstate_method = { dml2_pstate_method_fw_drr, dml2_pstate_method_na, dml2_pstate_method_na, dml2_pstate_method_na },
.allow_state_increase = true,
},
// Finally VBlank, but allow base clocks for latency to increase
/*
{
- .per_stream_pstate_method = { dml2_pmo_pstate_strategy_vblank, dml2_pmo_pstate_strategy_na, dml2_pmo_pstate_strategy_na, dml2_pmo_pstate_strategy_na },
+ .per_stream_pstate_method = { dml2_pstate_method_vblank, dml2_pstate_method_na, dml2_pstate_method_na, dml2_pstate_method_na },
.allow_state_increase = true,
},
*/
@@ -49,56 +49,56 @@ static const int base_strategy_list_1_display_size = sizeof(base_strategy_list_1
static const struct dml2_pmo_pstate_strategy base_strategy_list_2_display[] = {
// VActive only is preferred
{
- .per_stream_pstate_method = { dml2_pmo_pstate_strategy_vactive, dml2_pmo_pstate_strategy_vactive, dml2_pmo_pstate_strategy_na, dml2_pmo_pstate_strategy_na },
+ .per_stream_pstate_method = { dml2_pstate_method_vactive, dml2_pstate_method_vactive, dml2_pstate_method_na, dml2_pstate_method_na },
.allow_state_increase = true,
},
// Then VActive + VBlank
{
- .per_stream_pstate_method = { dml2_pmo_pstate_strategy_vactive, dml2_pmo_pstate_strategy_vblank, dml2_pmo_pstate_strategy_na, dml2_pmo_pstate_strategy_na },
+ .per_stream_pstate_method = { dml2_pstate_method_vactive, dml2_pstate_method_vblank, dml2_pstate_method_na, dml2_pstate_method_na },
.allow_state_increase = false,
},
// Then VBlank only
{
- .per_stream_pstate_method = { dml2_pmo_pstate_strategy_vblank, dml2_pmo_pstate_strategy_vblank, dml2_pmo_pstate_strategy_na, dml2_pmo_pstate_strategy_na },
+ .per_stream_pstate_method = { dml2_pstate_method_vblank, dml2_pstate_method_vblank, dml2_pstate_method_na, dml2_pstate_method_na },
.allow_state_increase = false,
},
// Then SVP + VBlank
{
- .per_stream_pstate_method = { dml2_pmo_pstate_strategy_fw_svp, dml2_pmo_pstate_strategy_vblank, dml2_pmo_pstate_strategy_na, dml2_pmo_pstate_strategy_na },
+ .per_stream_pstate_method = { dml2_pstate_method_fw_svp, dml2_pstate_method_vblank, dml2_pstate_method_na, dml2_pstate_method_na },
.allow_state_increase = false,
},
// Then SVP + DRR
{
- .per_stream_pstate_method = { dml2_pmo_pstate_strategy_fw_svp, dml2_pmo_pstate_strategy_fw_drr, dml2_pmo_pstate_strategy_na, dml2_pmo_pstate_strategy_na },
+ .per_stream_pstate_method = { dml2_pstate_method_fw_svp, dml2_pstate_method_fw_drr, dml2_pstate_method_na, dml2_pstate_method_na },
.allow_state_increase = true,
},
// Then SVP + SVP
{
- .per_stream_pstate_method = { dml2_pmo_pstate_strategy_fw_svp, dml2_pmo_pstate_strategy_fw_svp, dml2_pmo_pstate_strategy_na, dml2_pmo_pstate_strategy_na },
+ .per_stream_pstate_method = { dml2_pstate_method_fw_svp, dml2_pstate_method_fw_svp, dml2_pstate_method_na, dml2_pstate_method_na },
.allow_state_increase = true,
},
// Then DRR + VActive
{
- .per_stream_pstate_method = { dml2_pmo_pstate_strategy_vactive, dml2_pmo_pstate_strategy_fw_drr, dml2_pmo_pstate_strategy_na, dml2_pmo_pstate_strategy_na },
+ .per_stream_pstate_method = { dml2_pstate_method_vactive, dml2_pstate_method_fw_drr, dml2_pstate_method_na, dml2_pstate_method_na },
.allow_state_increase = true,
},
// Then DRR + DRR
{
- .per_stream_pstate_method = { dml2_pmo_pstate_strategy_fw_drr, dml2_pmo_pstate_strategy_fw_drr, dml2_pmo_pstate_strategy_na, dml2_pmo_pstate_strategy_na },
+ .per_stream_pstate_method = { dml2_pstate_method_fw_drr, dml2_pstate_method_fw_drr, dml2_pstate_method_na, dml2_pstate_method_na },
.allow_state_increase = true,
},
// Finally VBlank, but allow base clocks for latency to increase
/*
{
- .per_stream_pstate_method = { dml2_pmo_pstate_strategy_vblank, dml2_pmo_pstate_strategy_vblank, dml2_pmo_pstate_strategy_na, dml2_pmo_pstate_strategy_na },
+ .per_stream_pstate_method = { dml2_pstate_method_vblank, dml2_pstate_method_vblank, dml2_pstate_method_na, dml2_pstate_method_na },
.allow_state_increase = true,
},
*/
@@ -109,32 +109,32 @@ static const int base_strategy_list_2_display_size = sizeof(base_strategy_list_2
static const struct dml2_pmo_pstate_strategy base_strategy_list_3_display[] = {
// All VActive
{
- .per_stream_pstate_method = { dml2_pmo_pstate_strategy_vactive, dml2_pmo_pstate_strategy_vactive, dml2_pmo_pstate_strategy_vactive, dml2_pmo_pstate_strategy_na },
+ .per_stream_pstate_method = { dml2_pstate_method_vactive, dml2_pstate_method_vactive, dml2_pstate_method_vactive, dml2_pstate_method_na },
.allow_state_increase = true,
},
// VActive + 1 VBlank
{
- .per_stream_pstate_method = { dml2_pmo_pstate_strategy_vactive, dml2_pmo_pstate_strategy_vactive, dml2_pmo_pstate_strategy_vblank, dml2_pmo_pstate_strategy_na },
+ .per_stream_pstate_method = { dml2_pstate_method_vactive, dml2_pstate_method_vactive, dml2_pstate_method_vblank, dml2_pstate_method_na },
.allow_state_increase = false,
},
// All VBlank
{
- .per_stream_pstate_method = { dml2_pmo_pstate_strategy_vblank, dml2_pmo_pstate_strategy_vblank, dml2_pmo_pstate_strategy_vblank, dml2_pmo_pstate_strategy_na },
+ .per_stream_pstate_method = { dml2_pstate_method_vblank, dml2_pstate_method_vblank, dml2_pstate_method_vblank, dml2_pstate_method_na },
.allow_state_increase = false,
},
// All DRR
{
- .per_stream_pstate_method = { dml2_pmo_pstate_strategy_fw_drr, dml2_pmo_pstate_strategy_fw_drr, dml2_pmo_pstate_strategy_fw_drr, dml2_pmo_pstate_strategy_na },
+ .per_stream_pstate_method = { dml2_pstate_method_fw_drr, dml2_pstate_method_fw_drr, dml2_pstate_method_fw_drr, dml2_pstate_method_na },
.allow_state_increase = true,
},
// All VBlank, with state increase allowed
/*
{
- .per_stream_pstate_method = { dml2_pmo_pstate_strategy_vblank, dml2_pmo_pstate_strategy_vblank, dml2_pmo_pstate_strategy_vblank, dml2_pmo_pstate_strategy_na },
+ .per_stream_pstate_method = { dml2_pstate_method_vblank, dml2_pstate_method_vblank, dml2_pstate_method_vblank, dml2_pstate_method_na },
.allow_state_increase = true,
},
*/
@@ -145,32 +145,32 @@ static const int base_strategy_list_3_display_size = sizeof(base_strategy_list_3
static const struct dml2_pmo_pstate_strategy base_strategy_list_4_display[] = {
// All VActive
{
- .per_stream_pstate_method = { dml2_pmo_pstate_strategy_vactive, dml2_pmo_pstate_strategy_vactive, dml2_pmo_pstate_strategy_vactive, dml2_pmo_pstate_strategy_vactive },
+ .per_stream_pstate_method = { dml2_pstate_method_vactive, dml2_pstate_method_vactive, dml2_pstate_method_vactive, dml2_pstate_method_vactive },
.allow_state_increase = true,
},
// VActive + 1 VBlank
{
- .per_stream_pstate_method = { dml2_pmo_pstate_strategy_vactive, dml2_pmo_pstate_strategy_vactive, dml2_pmo_pstate_strategy_vactive, dml2_pmo_pstate_strategy_vblank },
+ .per_stream_pstate_method = { dml2_pstate_method_vactive, dml2_pstate_method_vactive, dml2_pstate_method_vactive, dml2_pstate_method_vblank },
.allow_state_increase = false,
},
// All Vblank
{
- .per_stream_pstate_method = { dml2_pmo_pstate_strategy_vblank, dml2_pmo_pstate_strategy_vblank, dml2_pmo_pstate_strategy_vblank, dml2_pmo_pstate_strategy_vblank },
+ .per_stream_pstate_method = { dml2_pstate_method_vblank, dml2_pstate_method_vblank, dml2_pstate_method_vblank, dml2_pstate_method_vblank },
.allow_state_increase = false,
},
// All DRR
{
- .per_stream_pstate_method = { dml2_pmo_pstate_strategy_fw_drr, dml2_pmo_pstate_strategy_fw_drr, dml2_pmo_pstate_strategy_fw_drr, dml2_pmo_pstate_strategy_fw_drr },
+ .per_stream_pstate_method = { dml2_pstate_method_fw_drr, dml2_pstate_method_fw_drr, dml2_pstate_method_fw_drr, dml2_pstate_method_fw_drr },
.allow_state_increase = true,
},
// All VBlank, with state increase allowed
/*
{
- .per_stream_pstate_method = { dml2_pmo_pstate_strategy_vblank, dml2_pmo_pstate_strategy_vblank, dml2_pmo_pstate_strategy_vblank, dml2_pmo_pstate_strategy_vblank },
+ .per_stream_pstate_method = { dml2_pstate_method_vblank, dml2_pstate_method_vblank, dml2_pstate_method_vblank, dml2_pstate_method_vblank },
.allow_state_increase = true,
},
*/
@@ -355,29 +355,30 @@ bool pmo_dcn4_fams2_optimize_dcc_mcache(struct dml2_pmo_optimize_dcc_mcache_in_o
return result;
}
-static enum dml2_pmo_pstate_method convert_strategy_to_drr_variant(const enum dml2_pmo_pstate_method base_strategy)
+static enum dml2_pstate_method convert_strategy_to_drr_variant(const enum dml2_pstate_method base_strategy)
{
- enum dml2_pmo_pstate_method variant_strategy = 0;
+ enum dml2_pstate_method variant_strategy = 0;
switch (base_strategy) {
- case dml2_pmo_pstate_strategy_vactive:
- variant_strategy = dml2_pmo_pstate_strategy_fw_vactive_drr;
+ case dml2_pstate_method_vactive:
+ variant_strategy = dml2_pstate_method_fw_vactive_drr;
break;
- case dml2_pmo_pstate_strategy_vblank:
- variant_strategy = dml2_pmo_pstate_strategy_fw_vblank_drr;
+ case dml2_pstate_method_vblank:
+ variant_strategy = dml2_pstate_method_fw_vblank_drr;
break;
- case dml2_pmo_pstate_strategy_fw_svp:
- variant_strategy = dml2_pmo_pstate_strategy_fw_svp_drr;
+ case dml2_pstate_method_fw_svp:
+ variant_strategy = dml2_pstate_method_fw_svp_drr;
break;
- case dml2_pmo_pstate_strategy_fw_vactive_drr:
- case dml2_pmo_pstate_strategy_fw_vblank_drr:
- case dml2_pmo_pstate_strategy_fw_svp_drr:
- case dml2_pmo_pstate_strategy_fw_drr:
- case dml2_pmo_pstate_strategy_reserved_hw:
- case dml2_pmo_pstate_strategy_reserved_fw:
- case dml2_pmo_pstate_strategy_reserved_fw_drr_clamped:
- case dml2_pmo_pstate_strategy_reserved_fw_drr_var:
- case dml2_pmo_pstate_strategy_na:
+ case dml2_pstate_method_fw_vactive_drr:
+ case dml2_pstate_method_fw_vblank_drr:
+ case dml2_pstate_method_fw_svp_drr:
+ case dml2_pstate_method_fw_drr:
+ case dml2_pstate_method_reserved_hw:
+ case dml2_pstate_method_reserved_fw:
+ case dml2_pstate_method_reserved_fw_drr_clamped:
+ case dml2_pstate_method_reserved_fw_drr_var:
+ case dml2_pstate_method_count:
+ case dml2_pstate_method_na:
default:
/* no variant for this mode */
variant_strategy = base_strategy;
@@ -419,23 +420,22 @@ static unsigned int get_num_expanded_strategies(
static void insert_strategy_into_expanded_list(
const struct dml2_pmo_pstate_strategy *per_stream_pstate_strategy,
- int stream_count,
- struct dml2_pmo_init_data *init_data)
+ const int stream_count,
+ struct dml2_pmo_pstate_strategy *expanded_strategy_list,
+ unsigned int *num_expanded_strategies)
{
- struct dml2_pmo_pstate_strategy *expanded_strategy_list = NULL;
-
- expanded_strategy_list = get_expanded_strategy_list(init_data, stream_count);
+ if (expanded_strategy_list && num_expanded_strategies) {
+ memcpy(&expanded_strategy_list[*num_expanded_strategies], per_stream_pstate_strategy, sizeof(struct dml2_pmo_pstate_strategy));
- if (expanded_strategy_list) {
- memcpy(&expanded_strategy_list[init_data->pmo_dcn4.num_expanded_strategies_per_list[stream_count - 1]], per_stream_pstate_strategy, sizeof(struct dml2_pmo_pstate_strategy));
-
- init_data->pmo_dcn4.num_expanded_strategies_per_list[stream_count - 1]++;
+ (*num_expanded_strategies)++;
}
}
-static void expand_base_strategy(struct dml2_pmo_instance *pmo,
+static void expand_base_strategy(
const struct dml2_pmo_pstate_strategy *base_strategy,
- unsigned int stream_count)
+ const unsigned int stream_count,
+ struct dml2_pmo_pstate_strategy *expanded_strategy_list,
+ unsigned int *num_expanded_strategies)
{
bool skip_to_next_stream;
bool expanded_strategy_added;
@@ -473,7 +473,7 @@ static void expand_base_strategy(struct dml2_pmo_instance *pmo,
if (i >= stream_count - 1) {
/* insert into strategy list */
- insert_strategy_into_expanded_list(&cur_strategy_list, stream_count, &pmo->init_data);
+ insert_strategy_into_expanded_list(&cur_strategy_list, stream_count, expanded_strategy_list, num_expanded_strategies);
expanded_strategy_added = true;
} else {
/* skip to next stream */
@@ -512,9 +512,9 @@ static void expand_base_strategy(struct dml2_pmo_instance *pmo,
static bool is_variant_method_valid(const struct dml2_pmo_pstate_strategy *base_strategy,
const struct dml2_pmo_pstate_strategy *variant_strategy,
- unsigned int num_streams_per_base_method[PMO_DCN4_MAX_DISPLAYS],
- unsigned int num_streams_per_variant_method[PMO_DCN4_MAX_DISPLAYS],
- unsigned int stream_count)
+ const unsigned int num_streams_per_base_method[PMO_DCN4_MAX_DISPLAYS],
+ const unsigned int num_streams_per_variant_method[PMO_DCN4_MAX_DISPLAYS],
+ const unsigned int stream_count)
{
bool valid = true;
unsigned int i;
@@ -522,7 +522,7 @@ static bool is_variant_method_valid(const struct dml2_pmo_pstate_strategy *base_
/* check all restrictions are met */
for (i = 0; i < stream_count; i++) {
/* vblank + vblank_drr variants are invalid */
- if (base_strategy->per_stream_pstate_method[i] == dml2_pmo_pstate_strategy_vblank &&
+ if (base_strategy->per_stream_pstate_method[i] == dml2_pstate_method_vblank &&
((num_streams_per_base_method[i] > 0 && num_streams_per_variant_method[i] > 0) ||
num_streams_per_variant_method[i] > 1)) {
valid = false;
@@ -533,9 +533,12 @@ static bool is_variant_method_valid(const struct dml2_pmo_pstate_strategy *base_
return valid;
}
-static void expand_variant_strategy(struct dml2_pmo_instance *pmo,
+static void expand_variant_strategy(
const struct dml2_pmo_pstate_strategy *base_strategy,
- unsigned int stream_count)
+ const unsigned int stream_count,
+ const bool should_permute,
+ struct dml2_pmo_pstate_strategy *expanded_strategy_list,
+ unsigned int *num_expanded_strategies)
{
bool variant_found;
unsigned int i, j;
@@ -544,7 +547,7 @@ static void expand_variant_strategy(struct dml2_pmo_instance *pmo,
unsigned int num_streams_per_method[PMO_DCN4_MAX_DISPLAYS] = { 0 };
unsigned int num_streams_per_base_method[PMO_DCN4_MAX_DISPLAYS] = { 0 };
unsigned int num_streams_per_variant_method[PMO_DCN4_MAX_DISPLAYS] = { 0 };
- enum dml2_pmo_pstate_method per_stream_variant_method[DML2_MAX_PLANES];
+ enum dml2_pstate_method per_stream_variant_method[DML2_MAX_PLANES];
struct dml2_pmo_pstate_strategy variant_strategy = { 0 };
/* determine number of displays per method */
@@ -585,7 +588,13 @@ static void expand_variant_strategy(struct dml2_pmo_instance *pmo,
}
if (variant_found && is_variant_method_valid(base_strategy, &variant_strategy, num_streams_per_base_method, num_streams_per_variant_method, stream_count)) {
- expand_base_strategy(pmo, &variant_strategy, stream_count);
+ if (should_permute) {
+ /* permutations are permitted, proceed to expand */
+ expand_base_strategy(&variant_strategy, stream_count, expanded_strategy_list, num_expanded_strategies);
+ } else {
+ /* no permutations allowed, so add to list now */
+ insert_strategy_into_expanded_list(&variant_strategy, stream_count, expanded_strategy_list, num_expanded_strategies);
+ }
}
/* rollback to earliest method with bases remaining */
@@ -612,18 +621,19 @@ static void expand_variant_strategy(struct dml2_pmo_instance *pmo,
}
}
-static void expand_base_strategies(
- struct dml2_pmo_instance *pmo,
- const struct dml2_pmo_pstate_strategy *base_strategies_list,
- const unsigned int num_base_strategies,
- unsigned int stream_count)
+void pmo_dcn4_fams2_expand_base_pstate_strategies(
+ const struct dml2_pmo_pstate_strategy *base_strategies_list,
+ const unsigned int num_base_strategies,
+ const unsigned int stream_count,
+ struct dml2_pmo_pstate_strategy *expanded_strategy_list,
+ unsigned int *num_expanded_strategies)
{
unsigned int i;
/* expand every explicit base strategy (except all DRR) */
for (i = 0; i < num_base_strategies; i++) {
- expand_base_strategy(pmo, &base_strategies_list[i], stream_count);
- expand_variant_strategy(pmo, &base_strategies_list[i], stream_count);
+ expand_base_strategy(&base_strategies_list[i], stream_count, expanded_strategy_list, num_expanded_strategies);
+ expand_variant_strategy(&base_strategies_list[i], stream_count, true, expanded_strategy_list, num_expanded_strategies);
}
}
@@ -652,25 +662,45 @@ bool pmo_dcn4_fams2_initialize(struct dml2_pmo_initialize_in_out *in_out)
DML2_ASSERT(base_strategy_list_1_display_size <= PMO_DCN4_MAX_BASE_STRATEGIES);
/* populate list */
- expand_base_strategies(pmo, base_strategy_list_1_display, base_strategy_list_1_display_size, 1);
+ pmo_dcn4_fams2_expand_base_pstate_strategies(
+ base_strategy_list_1_display,
+ base_strategy_list_1_display_size,
+ i,
+ pmo->init_data.pmo_dcn4.expanded_strategy_list_1_display,
+ &pmo->init_data.pmo_dcn4.num_expanded_strategies_per_list[i - 1]);
break;
case 2:
DML2_ASSERT(base_strategy_list_2_display_size <= PMO_DCN4_MAX_BASE_STRATEGIES);
/* populate list */
- expand_base_strategies(pmo, base_strategy_list_2_display, base_strategy_list_2_display_size, 2);
+ pmo_dcn4_fams2_expand_base_pstate_strategies(
+ base_strategy_list_2_display,
+ base_strategy_list_2_display_size,
+ i,
+ pmo->init_data.pmo_dcn4.expanded_strategy_list_2_display,
+ &pmo->init_data.pmo_dcn4.num_expanded_strategies_per_list[i - 1]);
break;
case 3:
DML2_ASSERT(base_strategy_list_3_display_size <= PMO_DCN4_MAX_BASE_STRATEGIES);
/* populate list */
- expand_base_strategies(pmo, base_strategy_list_3_display, base_strategy_list_3_display_size, 3);
+ pmo_dcn4_fams2_expand_base_pstate_strategies(
+ base_strategy_list_3_display,
+ base_strategy_list_3_display_size,
+ i,
+ pmo->init_data.pmo_dcn4.expanded_strategy_list_3_display,
+ &pmo->init_data.pmo_dcn4.num_expanded_strategies_per_list[i - 1]);
break;
case 4:
DML2_ASSERT(base_strategy_list_4_display_size <= PMO_DCN4_MAX_BASE_STRATEGIES);
/* populate list */
- expand_base_strategies(pmo, base_strategy_list_4_display, base_strategy_list_4_display_size, 4);
+ pmo_dcn4_fams2_expand_base_pstate_strategies(
+ base_strategy_list_4_display,
+ base_strategy_list_4_display_size,
+ i,
+ pmo->init_data.pmo_dcn4.expanded_strategy_list_4_display,
+ &pmo->init_data.pmo_dcn4.num_expanded_strategies_per_list[i - 1]);
break;
}
}
@@ -783,8 +813,12 @@ static int find_highest_odm_load_stream_index(
int odm_load, highest_odm_load = -1, highest_odm_load_index = -1;
for (i = 0; i < display_config->num_streams; i++) {
- odm_load = display_config->stream_descriptors[i].timing.pixel_clock_khz
+ if (mode_support_result->cfg_support_info.stream_support_info[i].odms_used > 0)
+ odm_load = display_config->stream_descriptors[i].timing.pixel_clock_khz
/ mode_support_result->cfg_support_info.stream_support_info[i].odms_used;
+ else
+ odm_load = 0;
+
if (odm_load > highest_odm_load) {
highest_odm_load_index = i;
highest_odm_load = odm_load;
@@ -941,11 +975,8 @@ static void build_synchronized_timing_groups(
/* find synchronizable timing groups */
for (j = i + 1; j < display_config->display_config.num_streams; j++) {
if (memcmp(master_timing,
- &display_config->display_config.stream_descriptors[j].timing,
- sizeof(struct dml2_timing_cfg)) == 0 &&
- display_config->display_config.stream_descriptors[i].output.output_encoder == display_config->display_config.stream_descriptors[j].output.output_encoder &&
- (display_config->display_config.stream_descriptors[i].output.output_encoder != dml2_hdmi || //hdmi requires formats match
- display_config->display_config.stream_descriptors[i].output.output_format == display_config->display_config.stream_descriptors[j].output.output_format)) {
+ &display_config->display_config.stream_descriptors[j].timing,
+ sizeof(struct dml2_timing_cfg)) == 0) {
set_bit_in_bitfield(&pmo->scratch.pmo_dcn4.synchronized_timing_group_masks[timing_group_idx], j);
set_bit_in_bitfield(&stream_mapped_mask, j);
}
@@ -959,7 +990,7 @@ static bool all_timings_support_vactive(const struct dml2_pmo_instance *pmo,
const struct display_configuation_with_meta *display_config,
unsigned int mask)
{
- unsigned char i;
+ unsigned int i;
bool valid = true;
// Create a remap array to enable simple iteration through only masked stream indicies
@@ -1008,7 +1039,7 @@ static bool all_timings_support_drr(const struct dml2_pmo_instance *pmo,
const struct display_configuation_with_meta *display_config,
unsigned int mask)
{
- unsigned char i;
+ unsigned int i;
for (i = 0; i < DML2_MAX_PLANES; i++) {
const struct dml2_stream_parameters *stream_descriptor;
const struct dml2_fams2_meta *stream_fams2_meta;
@@ -1050,7 +1081,7 @@ static bool all_timings_support_svp(const struct dml2_pmo_instance *pmo,
const struct dml2_plane_parameters *plane_descriptor;
const struct dml2_fams2_meta *stream_fams2_meta;
unsigned int microschedule_vlines;
- unsigned char i;
+ unsigned int i;
unsigned int num_planes_per_stream[DML2_MAX_PLANES] = { 0 };
@@ -1106,24 +1137,73 @@ static void insert_into_candidate_list(const struct dml2_pmo_pstate_strategy *ps
scratch->pmo_dcn4.num_pstate_candidates++;
}
-static bool all_planes_match_method(const struct display_configuation_with_meta *display_cfg, int plane_mask, enum dml2_pmo_pstate_method method)
+static enum dml2_pstate_method uclk_pstate_strategy_override_to_pstate_method(const enum dml2_uclk_pstate_change_strategy override_strategy)
{
- unsigned char i;
- enum dml2_uclk_pstate_change_strategy matching_strategy = (enum dml2_uclk_pstate_change_strategy) dml2_pmo_pstate_strategy_na;
+ enum dml2_pstate_method method = dml2_pstate_method_na;
- if (method == dml2_pmo_pstate_strategy_vactive || method == dml2_pmo_pstate_strategy_fw_vactive_drr)
- matching_strategy = dml2_uclk_pstate_change_strategy_force_vactive;
- else if (method == dml2_pmo_pstate_strategy_vblank || method == dml2_pmo_pstate_strategy_fw_vblank_drr)
- matching_strategy = dml2_uclk_pstate_change_strategy_force_vblank;
- else if (method == dml2_pmo_pstate_strategy_fw_svp)
- matching_strategy = dml2_uclk_pstate_change_strategy_force_mall_svp;
- else if (method == dml2_pmo_pstate_strategy_fw_drr)
- matching_strategy = dml2_uclk_pstate_change_strategy_force_drr;
+ switch (override_strategy) {
+ case dml2_uclk_pstate_change_strategy_force_vactive:
+ method = dml2_pstate_method_vactive;
+ break;
+ case dml2_uclk_pstate_change_strategy_force_vblank:
+ method = dml2_pstate_method_vblank;
+ break;
+ case dml2_uclk_pstate_change_strategy_force_drr:
+ method = dml2_pstate_method_fw_drr;
+ break;
+ case dml2_uclk_pstate_change_strategy_force_mall_svp:
+ method = dml2_pstate_method_fw_svp;
+ break;
+ case dml2_uclk_pstate_change_strategy_force_mall_full_frame:
+ case dml2_uclk_pstate_change_strategy_auto:
+ default:
+ method = dml2_pstate_method_na;
+ }
+
+ return method;
+}
+
+static enum dml2_uclk_pstate_change_strategy pstate_method_to_uclk_pstate_strategy_override(const enum dml2_pstate_method method)
+{
+ enum dml2_uclk_pstate_change_strategy override_strategy = dml2_uclk_pstate_change_strategy_auto;
+
+ switch (method) {
+ case dml2_pstate_method_vactive:
+ case dml2_pstate_method_fw_vactive_drr:
+ override_strategy = dml2_uclk_pstate_change_strategy_force_vactive;
+ break;
+ case dml2_pstate_method_vblank:
+ case dml2_pstate_method_fw_vblank_drr:
+ override_strategy = dml2_uclk_pstate_change_strategy_force_vblank;
+ break;
+ case dml2_pstate_method_fw_svp:
+ case dml2_pstate_method_fw_svp_drr:
+ override_strategy = dml2_uclk_pstate_change_strategy_force_mall_svp;
+ break;
+ case dml2_pstate_method_fw_drr:
+ override_strategy = dml2_uclk_pstate_change_strategy_force_drr;
+ break;
+ case dml2_pstate_method_reserved_hw:
+ case dml2_pstate_method_reserved_fw:
+ case dml2_pstate_method_reserved_fw_drr_clamped:
+ case dml2_pstate_method_reserved_fw_drr_var:
+ case dml2_pstate_method_count:
+ case dml2_pstate_method_na:
+ default:
+ override_strategy = dml2_uclk_pstate_change_strategy_auto;
+ }
+
+ return override_strategy;
+}
+
+static bool all_planes_match_method(const struct display_configuation_with_meta *display_cfg, int plane_mask, enum dml2_pstate_method method)
+{
+ unsigned int i;
for (i = 0; i < DML2_MAX_PLANES; i++) {
if (is_bit_set_in_bitfield(plane_mask, i)) {
if (display_cfg->display_config.plane_descriptors[i].overrides.uclk_pstate_change_strategy != dml2_uclk_pstate_change_strategy_auto &&
- display_cfg->display_config.plane_descriptors[i].overrides.uclk_pstate_change_strategy != matching_strategy)
+ display_cfg->display_config.plane_descriptors[i].overrides.uclk_pstate_change_strategy != pstate_method_to_uclk_pstate_strategy_override(method))
return false;
}
}
@@ -1149,32 +1229,33 @@ static void build_method_scheduling_params(
static struct dml2_fams2_per_method_common_meta *get_per_method_common_meta(
struct dml2_pmo_instance *pmo,
- enum dml2_pmo_pstate_method stream_pstate_method,
+ enum dml2_pstate_method stream_pstate_method,
int stream_idx)
{
struct dml2_fams2_per_method_common_meta *stream_method_fams2_meta = NULL;
switch (stream_pstate_method) {
- case dml2_pmo_pstate_strategy_vactive:
- case dml2_pmo_pstate_strategy_fw_vactive_drr:
+ case dml2_pstate_method_vactive:
+ case dml2_pstate_method_fw_vactive_drr:
stream_method_fams2_meta = &pmo->scratch.pmo_dcn4.stream_fams2_meta[stream_idx].method_vactive.common;
break;
- case dml2_pmo_pstate_strategy_vblank:
- case dml2_pmo_pstate_strategy_fw_vblank_drr:
+ case dml2_pstate_method_vblank:
+ case dml2_pstate_method_fw_vblank_drr:
stream_method_fams2_meta = &pmo->scratch.pmo_dcn4.stream_fams2_meta[stream_idx].method_vblank.common;
break;
- case dml2_pmo_pstate_strategy_fw_svp:
- case dml2_pmo_pstate_strategy_fw_svp_drr:
+ case dml2_pstate_method_fw_svp:
+ case dml2_pstate_method_fw_svp_drr:
stream_method_fams2_meta = &pmo->scratch.pmo_dcn4.stream_fams2_meta[stream_idx].method_subvp.common;
break;
- case dml2_pmo_pstate_strategy_fw_drr:
+ case dml2_pstate_method_fw_drr:
stream_method_fams2_meta = &pmo->scratch.pmo_dcn4.stream_fams2_meta[stream_idx].method_drr.common;
break;
- case dml2_pmo_pstate_strategy_reserved_hw:
- case dml2_pmo_pstate_strategy_reserved_fw:
- case dml2_pmo_pstate_strategy_reserved_fw_drr_clamped:
- case dml2_pmo_pstate_strategy_reserved_fw_drr_var:
- case dml2_pmo_pstate_strategy_na:
+ case dml2_pstate_method_reserved_hw:
+ case dml2_pstate_method_reserved_fw:
+ case dml2_pstate_method_reserved_fw_drr_clamped:
+ case dml2_pstate_method_reserved_fw_drr_var:
+ case dml2_pstate_method_count:
+ case dml2_pstate_method_na:
default:
stream_method_fams2_meta = NULL;
}
@@ -1215,7 +1296,7 @@ static bool is_timing_group_schedulable(
if (is_bit_set_in_bitfield(pmo->scratch.pmo_dcn4.synchronized_timing_group_masks[timing_group_idx], i)) {
stream_method_fams2_meta = get_per_method_common_meta(pmo, pstate_strategy->per_stream_pstate_method[i], i);
if (!stream_method_fams2_meta)
- return false;
+ continue;
if (group_fams2_meta->allow_start_otg_vline < stream_method_fams2_meta->allow_start_otg_vline) {
/* set group allow start to larger otg vline */
@@ -1295,7 +1376,7 @@ static bool is_config_schedulable(
if (j_disallow_us < jp1_disallow_us) {
/* swap as A < B */
swap(s->pmo_dcn4.sorted_group_gtl_disallow_index[j],
- s->pmo_dcn4.sorted_group_gtl_disallow_index[j+1]);
+ s->pmo_dcn4.sorted_group_gtl_disallow_index[j + 1]);
swapped = true;
}
}
@@ -1354,7 +1435,7 @@ static bool is_config_schedulable(
if (j_period_us < jp1_period_us) {
/* swap as A < B */
swap(s->pmo_dcn4.sorted_group_gtl_period_index[j],
- s->pmo_dcn4.sorted_group_gtl_period_index[j+1]);
+ s->pmo_dcn4.sorted_group_gtl_period_index[j + 1]);
swapped = true;
}
}
@@ -1413,7 +1494,7 @@ static bool is_config_schedulable(
static bool stream_matches_drr_policy(struct dml2_pmo_instance *pmo,
const struct display_configuation_with_meta *display_cfg,
- const enum dml2_pmo_pstate_method stream_pstate_method,
+ const enum dml2_pstate_method stream_pstate_method,
unsigned int stream_index)
{
const struct dml2_stream_parameters *stream_descriptor = &display_cfg->display_config.stream_descriptors[stream_index];
@@ -1468,7 +1549,7 @@ static bool validate_pstate_support_strategy_cofunctionality(struct dml2_pmo_ins
{
struct dml2_pmo_scratch *s = &pmo->scratch;
- unsigned char stream_index = 0;
+ unsigned int stream_index = 0;
unsigned int svp_count = 0;
unsigned int svp_stream_mask = 0;
@@ -1494,19 +1575,19 @@ static bool validate_pstate_support_strategy_cofunctionality(struct dml2_pmo_ins
strategy_matches_drr_requirements &=
stream_matches_drr_policy(pmo, display_cfg, pstate_strategy->per_stream_pstate_method[stream_index], stream_index);
- if (pstate_strategy->per_stream_pstate_method[stream_index] == dml2_pmo_pstate_strategy_fw_svp ||
- pstate_strategy->per_stream_pstate_method[stream_index] == dml2_pmo_pstate_strategy_fw_svp_drr) {
+ if (pstate_strategy->per_stream_pstate_method[stream_index] == dml2_pstate_method_fw_svp ||
+ pstate_strategy->per_stream_pstate_method[stream_index] == dml2_pstate_method_fw_svp_drr) {
svp_count++;
set_bit_in_bitfield(&svp_stream_mask, stream_index);
- } else if (pstate_strategy->per_stream_pstate_method[stream_index] == dml2_pmo_pstate_strategy_fw_drr) {
+ } else if (pstate_strategy->per_stream_pstate_method[stream_index] == dml2_pstate_method_fw_drr) {
drr_count++;
set_bit_in_bitfield(&drr_stream_mask, stream_index);
- } else if (pstate_strategy->per_stream_pstate_method[stream_index] == dml2_pmo_pstate_strategy_vactive ||
- pstate_strategy->per_stream_pstate_method[stream_index] == dml2_pmo_pstate_strategy_fw_vactive_drr) {
+ } else if (pstate_strategy->per_stream_pstate_method[stream_index] == dml2_pstate_method_vactive ||
+ pstate_strategy->per_stream_pstate_method[stream_index] == dml2_pstate_method_fw_vactive_drr) {
vactive_count++;
set_bit_in_bitfield(&vactive_stream_mask, stream_index);
- } else if (pstate_strategy->per_stream_pstate_method[stream_index] == dml2_pmo_pstate_strategy_vblank ||
- pstate_strategy->per_stream_pstate_method[stream_index] == dml2_pmo_pstate_strategy_fw_vblank_drr) {
+ } else if (pstate_strategy->per_stream_pstate_method[stream_index] == dml2_pstate_method_vblank ||
+ pstate_strategy->per_stream_pstate_method[stream_index] == dml2_pstate_method_fw_vblank_drr) {
vblank_count++;
set_bit_in_bitfield(&vblank_stream_mask, stream_index);
}
@@ -1532,7 +1613,7 @@ static bool validate_pstate_support_strategy_cofunctionality(struct dml2_pmo_ins
static int get_vactive_pstate_margin(const struct display_configuation_with_meta *display_cfg, int plane_mask)
{
- unsigned char i;
+ unsigned int i;
int min_vactive_margin_us = 0xFFFFFFF;
for (i = 0; i < DML2_MAX_PLANES; i++) {
@@ -1625,7 +1706,7 @@ static void build_fams2_meta_per_stream(struct dml2_pmo_instance *pmo,
/* for single stream, guarantee at least an instant of allow */
stream_fams2_meta->method_vactive.max_vactive_det_fill_delay_otg_vlines = (unsigned int)math_floor(
math_max2(0.0,
- timing->v_active - stream_fams2_meta->min_allow_width_otg_vlines - stream_fams2_meta->dram_clk_change_blackout_otg_vlines));
+ timing->v_active - math_max2(1.0, stream_fams2_meta->min_allow_width_otg_vlines) - stream_fams2_meta->dram_clk_change_blackout_otg_vlines));
} else {
/* for multi stream, bound to a max fill time defined by IP caps */
stream_fams2_meta->method_vactive.max_vactive_det_fill_delay_otg_vlines =
@@ -1738,8 +1819,10 @@ bool pmo_dcn4_fams2_init_for_pstate_support(struct dml2_pmo_init_for_pstate_supp
struct display_configuation_with_meta *display_config;
const struct dml2_plane_parameters *plane_descriptor;
const struct dml2_pmo_pstate_strategy *strategy_list = NULL;
+ struct dml2_pmo_pstate_strategy override_base_strategy = { 0 };
unsigned int strategy_list_size = 0;
- unsigned char plane_index, stream_index, i;
+ unsigned int plane_index, stream_index, i;
+ bool build_override_strategy = true;
state->performed = true;
in_out->base_display_config->stage3.min_clk_index_for_latency = in_out->base_display_config->stage1.min_clk_index_for_latency;
@@ -1763,7 +1846,11 @@ bool pmo_dcn4_fams2_init_for_pstate_support(struct dml2_pmo_init_for_pstate_supp
set_bit_in_bitfield(&s->pmo_dcn4.stream_plane_mask[plane_descriptor->stream_index], plane_index);
- state->pstate_switch_modes[plane_index] = dml2_uclk_pstate_support_method_vactive;
+ state->pstate_switch_modes[plane_index] = dml2_pstate_method_vactive;
+
+ build_override_strategy &= plane_descriptor->overrides.uclk_pstate_change_strategy != dml2_uclk_pstate_change_strategy_auto;
+ override_base_strategy.per_stream_pstate_method[plane_descriptor->stream_index] =
+ uclk_pstate_strategy_override_to_pstate_method(plane_descriptor->overrides.uclk_pstate_change_strategy);
}
// Figure out which streams can do vactive, and also build up implicit SVP and FAMS2 meta
@@ -1781,13 +1868,30 @@ bool pmo_dcn4_fams2_init_for_pstate_support(struct dml2_pmo_init_for_pstate_supp
/* get synchronized timing groups */
build_synchronized_timing_groups(pmo, display_config);
- strategy_list = get_expanded_strategy_list(&pmo->init_data, display_config->display_config.num_streams);
- if (!strategy_list)
- return false;
-
- strategy_list_size = get_num_expanded_strategies(&pmo->init_data, display_config->display_config.num_streams);
+ if (build_override_strategy) {
+ /* build expanded override strategy list (no permutations) */
+ override_base_strategy.allow_state_increase = true;
+ s->pmo_dcn4.num_expanded_override_strategies = 0;
+ insert_strategy_into_expanded_list(&override_base_strategy,
+ display_config->display_config.num_streams,
+ s->pmo_dcn4.expanded_override_strategy_list,
+ &s->pmo_dcn4.num_expanded_override_strategies);
+ expand_variant_strategy(&override_base_strategy,
+ display_config->display_config.num_streams,
+ false,
+ s->pmo_dcn4.expanded_override_strategy_list,
+ &s->pmo_dcn4.num_expanded_override_strategies);
+
+ /* use override strategy list */
+ strategy_list = s->pmo_dcn4.expanded_override_strategy_list;
+ strategy_list_size = s->pmo_dcn4.num_expanded_override_strategies;
+ } else {
+ /* use predefined strategy list */
+ strategy_list = get_expanded_strategy_list(&pmo->init_data, display_config->display_config.num_streams);
+ strategy_list_size = get_num_expanded_strategies(&pmo->init_data, display_config->display_config.num_streams);
+ }
- if (strategy_list_size == 0)
+ if (!strategy_list || strategy_list_size == 0)
return false;
s->pmo_dcn4.num_pstate_candidates = 0;
@@ -1799,7 +1903,7 @@ bool pmo_dcn4_fams2_init_for_pstate_support(struct dml2_pmo_init_for_pstate_supp
}
if (s->pmo_dcn4.num_pstate_candidates > 0) {
- s->pmo_dcn4.pstate_strategy_candidates[s->pmo_dcn4.num_pstate_candidates - 1].allow_state_increase = true;
+ s->pmo_dcn4.pstate_strategy_candidates[s->pmo_dcn4.num_pstate_candidates-1].allow_state_increase = true;
s->pmo_dcn4.cur_pstate_candidate = -1;
return true;
} else {
@@ -1832,7 +1936,7 @@ static void reset_display_configuration(struct display_configuation_with_meta *d
// Reset strategy to auto
plane->overrides.uclk_pstate_change_strategy = dml2_uclk_pstate_change_strategy_auto;
- display_config->stage3.pstate_switch_modes[plane_index] = dml2_uclk_pstate_support_method_not_supported;
+ display_config->stage3.pstate_switch_modes[plane_index] = dml2_pstate_method_na;
}
}
@@ -1840,7 +1944,7 @@ static void setup_planes_for_drr_by_mask(struct display_configuation_with_meta *
struct dml2_pmo_instance *pmo,
int plane_mask)
{
- unsigned char plane_index;
+ unsigned int plane_index;
struct dml2_plane_parameters *plane;
for (plane_index = 0; plane_index < display_config->display_config.num_planes; plane_index++) {
@@ -1849,7 +1953,7 @@ static void setup_planes_for_drr_by_mask(struct display_configuation_with_meta *
plane->overrides.uclk_pstate_change_strategy = dml2_uclk_pstate_change_strategy_force_drr;
- display_config->stage3.pstate_switch_modes[plane_index] = dml2_uclk_pstate_support_method_fw_drr;
+ display_config->stage3.pstate_switch_modes[plane_index] = dml2_pstate_method_fw_drr;
}
}
@@ -1861,13 +1965,13 @@ static void setup_planes_for_svp_by_mask(struct display_configuation_with_meta *
{
struct dml2_pmo_scratch *scratch = &pmo->scratch;
- unsigned char plane_index;
+ unsigned int plane_index;
int stream_index = -1;
for (plane_index = 0; plane_index < display_config->display_config.num_planes; plane_index++) {
if (is_bit_set_in_bitfield(plane_mask, plane_index)) {
stream_index = (char)display_config->display_config.plane_descriptors[plane_index].stream_index;
- display_config->stage3.pstate_switch_modes[plane_index] = dml2_uclk_pstate_support_method_fw_subvp_phantom;
+ display_config->stage3.pstate_switch_modes[plane_index] = dml2_pstate_method_fw_svp;
}
}
@@ -1884,13 +1988,13 @@ static void setup_planes_for_svp_drr_by_mask(struct display_configuation_with_me
{
struct dml2_pmo_scratch *scratch = &pmo->scratch;
- unsigned char plane_index;
+ unsigned int plane_index;
int stream_index = -1;
for (plane_index = 0; plane_index < display_config->display_config.num_planes; plane_index++) {
if (is_bit_set_in_bitfield(plane_mask, plane_index)) {
stream_index = (char)display_config->display_config.plane_descriptors[plane_index].stream_index;
- display_config->stage3.pstate_switch_modes[plane_index] = dml2_uclk_pstate_support_method_fw_subvp_phantom_drr;
+ display_config->stage3.pstate_switch_modes[plane_index] = dml2_pstate_method_fw_svp_drr;
}
}
@@ -1905,7 +2009,7 @@ static void setup_planes_for_vblank_by_mask(struct display_configuation_with_met
struct dml2_pmo_instance *pmo,
int plane_mask)
{
- unsigned char plane_index;
+ unsigned int plane_index;
struct dml2_plane_parameters *plane;
for (plane_index = 0; plane_index < display_config->display_config.num_planes; plane_index++) {
@@ -1915,7 +2019,7 @@ static void setup_planes_for_vblank_by_mask(struct display_configuation_with_met
plane->overrides.reserved_vblank_time_ns = (long)math_max2(pmo->soc_bb->power_management_parameters.dram_clk_change_blackout_us * 1000.0,
plane->overrides.reserved_vblank_time_ns);
- display_config->stage3.pstate_switch_modes[plane_index] = dml2_uclk_pstate_support_method_vblank;
+ display_config->stage3.pstate_switch_modes[plane_index] = dml2_pstate_method_vblank;
}
}
@@ -1925,7 +2029,7 @@ static void setup_planes_for_vblank_drr_by_mask(struct display_configuation_with
struct dml2_pmo_instance *pmo,
int plane_mask)
{
- unsigned char plane_index;
+ unsigned int plane_index;
struct dml2_plane_parameters *plane;
for (plane_index = 0; plane_index < display_config->display_config.num_planes; plane_index++) {
@@ -1933,7 +2037,7 @@ static void setup_planes_for_vblank_drr_by_mask(struct display_configuation_with
plane = &display_config->display_config.plane_descriptors[plane_index];
plane->overrides.reserved_vblank_time_ns = (long)(pmo->soc_bb->power_management_parameters.dram_clk_change_blackout_us * 1000);
- display_config->stage3.pstate_switch_modes[plane_index] = dml2_uclk_pstate_support_method_fw_vblank_drr;
+ display_config->stage3.pstate_switch_modes[plane_index] = dml2_pstate_method_fw_vblank_drr;
}
}
}
@@ -1942,14 +2046,14 @@ static void setup_planes_for_vactive_by_mask(struct display_configuation_with_me
struct dml2_pmo_instance *pmo,
int plane_mask)
{
- unsigned char plane_index;
+ unsigned int plane_index;
unsigned int stream_index;
for (plane_index = 0; plane_index < display_config->display_config.num_planes; plane_index++) {
if (is_bit_set_in_bitfield(plane_mask, plane_index)) {
stream_index = display_config->display_config.plane_descriptors[plane_index].stream_index;
- display_config->stage3.pstate_switch_modes[plane_index] = dml2_uclk_pstate_support_method_vactive;
+ display_config->stage3.pstate_switch_modes[plane_index] = dml2_pstate_method_vactive;
if (!pmo->options->disable_vactive_det_fill_bw_pad) {
display_config->display_config.plane_descriptors[plane_index].overrides.max_vactive_det_fill_delay_us =
@@ -1963,14 +2067,14 @@ static void setup_planes_for_vactive_drr_by_mask(struct display_configuation_wit
struct dml2_pmo_instance *pmo,
int plane_mask)
{
- unsigned char plane_index;
+ unsigned int plane_index;
unsigned int stream_index;
for (plane_index = 0; plane_index < display_config->display_config.num_planes; plane_index++) {
if (is_bit_set_in_bitfield(plane_mask, plane_index)) {
stream_index = display_config->display_config.plane_descriptors[plane_index].stream_index;
- display_config->stage3.pstate_switch_modes[plane_index] = dml2_uclk_pstate_support_method_fw_vactive_drr;
+ display_config->stage3.pstate_switch_modes[plane_index] = dml2_pstate_method_fw_vactive_drr;
if (!pmo->options->disable_vactive_det_fill_bw_pad) {
display_config->display_config.plane_descriptors[plane_index].overrides.max_vactive_det_fill_delay_us =
@@ -1992,26 +2096,26 @@ static bool setup_display_config(struct display_configuation_with_meta *display_
for (stream_index = 0; stream_index < display_config->display_config.num_streams; stream_index++) {
- if (pmo->scratch.pmo_dcn4.pstate_strategy_candidates[strategy_index].per_stream_pstate_method[stream_index] == dml2_pmo_pstate_strategy_na) {
+ if (pmo->scratch.pmo_dcn4.pstate_strategy_candidates[strategy_index].per_stream_pstate_method[stream_index] == dml2_pstate_method_na) {
success = false;
break;
- } else if (scratch->pmo_dcn4.pstate_strategy_candidates[strategy_index].per_stream_pstate_method[stream_index] == dml2_pmo_pstate_strategy_vactive) {
+ } else if (scratch->pmo_dcn4.pstate_strategy_candidates[strategy_index].per_stream_pstate_method[stream_index] == dml2_pstate_method_vactive) {
setup_planes_for_vactive_by_mask(display_config, pmo, scratch->pmo_dcn4.stream_plane_mask[stream_index]);
- } else if (scratch->pmo_dcn4.pstate_strategy_candidates[strategy_index].per_stream_pstate_method[stream_index] == dml2_pmo_pstate_strategy_vblank) {
+ } else if (scratch->pmo_dcn4.pstate_strategy_candidates[strategy_index].per_stream_pstate_method[stream_index] == dml2_pstate_method_vblank) {
setup_planes_for_vblank_by_mask(display_config, pmo, scratch->pmo_dcn4.stream_plane_mask[stream_index]);
- } else if (scratch->pmo_dcn4.pstate_strategy_candidates[strategy_index].per_stream_pstate_method[stream_index] == dml2_pmo_pstate_strategy_fw_svp) {
+ } else if (scratch->pmo_dcn4.pstate_strategy_candidates[strategy_index].per_stream_pstate_method[stream_index] == dml2_pstate_method_fw_svp) {
fams2_required = true;
setup_planes_for_svp_by_mask(display_config, pmo, scratch->pmo_dcn4.stream_plane_mask[stream_index]);
- } else if (scratch->pmo_dcn4.pstate_strategy_candidates[strategy_index].per_stream_pstate_method[stream_index] == dml2_pmo_pstate_strategy_fw_vactive_drr) {
+ } else if (scratch->pmo_dcn4.pstate_strategy_candidates[strategy_index].per_stream_pstate_method[stream_index] == dml2_pstate_method_fw_vactive_drr) {
fams2_required = true;
setup_planes_for_vactive_drr_by_mask(display_config, pmo, scratch->pmo_dcn4.stream_plane_mask[stream_index]);
- } else if (scratch->pmo_dcn4.pstate_strategy_candidates[strategy_index].per_stream_pstate_method[stream_index] == dml2_pmo_pstate_strategy_fw_vblank_drr) {
+ } else if (scratch->pmo_dcn4.pstate_strategy_candidates[strategy_index].per_stream_pstate_method[stream_index] == dml2_pstate_method_fw_vblank_drr) {
fams2_required = true;
setup_planes_for_vblank_drr_by_mask(display_config, pmo, scratch->pmo_dcn4.stream_plane_mask[stream_index]);
- } else if (scratch->pmo_dcn4.pstate_strategy_candidates[strategy_index].per_stream_pstate_method[stream_index] == dml2_pmo_pstate_strategy_fw_svp_drr) {
+ } else if (scratch->pmo_dcn4.pstate_strategy_candidates[strategy_index].per_stream_pstate_method[stream_index] == dml2_pstate_method_fw_svp_drr) {
fams2_required = true;
setup_planes_for_svp_drr_by_mask(display_config, pmo, scratch->pmo_dcn4.stream_plane_mask[stream_index]);
- } else if (scratch->pmo_dcn4.pstate_strategy_candidates[strategy_index].per_stream_pstate_method[stream_index] == dml2_pmo_pstate_strategy_fw_drr) {
+ } else if (scratch->pmo_dcn4.pstate_strategy_candidates[strategy_index].per_stream_pstate_method[stream_index] == dml2_pstate_method_fw_drr) {
fams2_required = true;
setup_planes_for_drr_by_mask(display_config, pmo, scratch->pmo_dcn4.stream_plane_mask[stream_index]);
}
@@ -2031,7 +2135,7 @@ static bool setup_display_config(struct display_configuation_with_meta *display_
static int get_minimum_reserved_time_us_for_planes(struct display_configuation_with_meta *display_config, int plane_mask)
{
int min_time_us = 0xFFFFFF;
- unsigned char plane_index = 0;
+ unsigned int plane_index = 0;
for (plane_index = 0; plane_index < display_config->display_config.num_planes; plane_index++) {
if (is_bit_set_in_bitfield(plane_mask, plane_index)) {
@@ -2066,34 +2170,34 @@ bool pmo_dcn4_fams2_test_for_pstate_support(struct dml2_pmo_test_for_pstate_supp
for (stream_index = 0; stream_index < in_out->base_display_config->display_config.num_streams; stream_index++) {
struct dml2_fams2_meta *stream_fams2_meta = &s->pmo_dcn4.stream_fams2_meta[stream_index];
- if (s->pmo_dcn4.pstate_strategy_candidates[s->pmo_dcn4.cur_pstate_candidate].per_stream_pstate_method[stream_index] == dml2_pmo_pstate_strategy_vactive ||
- s->pmo_dcn4.pstate_strategy_candidates[s->pmo_dcn4.cur_pstate_candidate].per_stream_pstate_method[stream_index] == dml2_pmo_pstate_strategy_fw_vactive_drr) {
+ if (s->pmo_dcn4.pstate_strategy_candidates[s->pmo_dcn4.cur_pstate_candidate].per_stream_pstate_method[stream_index] == dml2_pstate_method_vactive ||
+ s->pmo_dcn4.pstate_strategy_candidates[s->pmo_dcn4.cur_pstate_candidate].per_stream_pstate_method[stream_index] == dml2_pstate_method_fw_vactive_drr) {
if (get_vactive_pstate_margin(in_out->base_display_config, s->pmo_dcn4.stream_plane_mask[stream_index]) < (MIN_VACTIVE_MARGIN_PCT * in_out->instance->soc_bb->power_management_parameters.dram_clk_change_blackout_us) ||
get_vactive_det_fill_latency_delay_us(in_out->base_display_config, s->pmo_dcn4.stream_plane_mask[stream_index]) > stream_fams2_meta->method_vactive.max_vactive_det_fill_delay_us) {
p_state_supported = false;
break;
}
- } else if (s->pmo_dcn4.pstate_strategy_candidates[s->pmo_dcn4.cur_pstate_candidate].per_stream_pstate_method[stream_index] == dml2_pmo_pstate_strategy_vblank ||
- s->pmo_dcn4.pstate_strategy_candidates[s->pmo_dcn4.cur_pstate_candidate].per_stream_pstate_method[stream_index] == dml2_pmo_pstate_strategy_fw_vblank_drr) {
+ } else if (s->pmo_dcn4.pstate_strategy_candidates[s->pmo_dcn4.cur_pstate_candidate].per_stream_pstate_method[stream_index] == dml2_pstate_method_vblank ||
+ s->pmo_dcn4.pstate_strategy_candidates[s->pmo_dcn4.cur_pstate_candidate].per_stream_pstate_method[stream_index] == dml2_pstate_method_fw_vblank_drr) {
if (get_minimum_reserved_time_us_for_planes(in_out->base_display_config, s->pmo_dcn4.stream_plane_mask[stream_index]) <
REQUIRED_RESERVED_TIME ||
get_vactive_pstate_margin(in_out->base_display_config, s->pmo_dcn4.stream_plane_mask[stream_index]) < MIN_VACTIVE_MARGIN_VBLANK) {
p_state_supported = false;
break;
}
- } else if (s->pmo_dcn4.pstate_strategy_candidates[s->pmo_dcn4.cur_pstate_candidate].per_stream_pstate_method[stream_index] == dml2_pmo_pstate_strategy_fw_svp ||
- s->pmo_dcn4.pstate_strategy_candidates[s->pmo_dcn4.cur_pstate_candidate].per_stream_pstate_method[stream_index] == dml2_pmo_pstate_strategy_fw_svp_drr) {
+ } else if (s->pmo_dcn4.pstate_strategy_candidates[s->pmo_dcn4.cur_pstate_candidate].per_stream_pstate_method[stream_index] == dml2_pstate_method_fw_svp ||
+ s->pmo_dcn4.pstate_strategy_candidates[s->pmo_dcn4.cur_pstate_candidate].per_stream_pstate_method[stream_index] == dml2_pstate_method_fw_svp_drr) {
if (in_out->base_display_config->stage3.stream_svp_meta[stream_index].valid == false) {
p_state_supported = false;
break;
}
- } else if (s->pmo_dcn4.pstate_strategy_candidates[s->pmo_dcn4.cur_pstate_candidate].per_stream_pstate_method[stream_index] == dml2_pmo_pstate_strategy_fw_drr) {
- if (!all_planes_match_method(in_out->base_display_config, s->pmo_dcn4.stream_plane_mask[stream_index], dml2_pmo_pstate_strategy_fw_drr) ||
+ } else if (s->pmo_dcn4.pstate_strategy_candidates[s->pmo_dcn4.cur_pstate_candidate].per_stream_pstate_method[stream_index] == dml2_pstate_method_fw_drr) {
+ if (!all_planes_match_method(in_out->base_display_config, s->pmo_dcn4.stream_plane_mask[stream_index], dml2_pstate_method_fw_drr) ||
get_vactive_pstate_margin(in_out->base_display_config, s->pmo_dcn4.stream_plane_mask[stream_index]) < MIN_VACTIVE_MARGIN_DRR) {
p_state_supported = false;
break;
}
- } else if (s->pmo_dcn4.pstate_strategy_candidates[s->pmo_dcn4.cur_pstate_candidate].per_stream_pstate_method[stream_index] == dml2_pmo_pstate_strategy_na) {
+ } else if (s->pmo_dcn4.pstate_strategy_candidates[s->pmo_dcn4.cur_pstate_candidate].per_stream_pstate_method[stream_index] == dml2_pstate_method_na) {
p_state_supported = false;
break;
}
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn4_fams2.h b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn4_fams2.h
index 0c25bd3e9ac0..6baab7ad6ecc 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn4_fams2.h
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn4_fams2.h
@@ -23,4 +23,11 @@ bool pmo_dcn4_fams2_init_for_stutter(struct dml2_pmo_init_for_stutter_in_out *in
bool pmo_dcn4_fams2_test_for_stutter(struct dml2_pmo_test_for_stutter_in_out *in_out);
bool pmo_dcn4_fams2_optimize_for_stutter(struct dml2_pmo_optimize_for_stutter_in_out *in_out);
+void pmo_dcn4_fams2_expand_base_pstate_strategies(
+ const struct dml2_pmo_pstate_strategy *base_strategies_list,
+ const unsigned int num_base_strategies,
+ const unsigned int stream_count,
+ struct dml2_pmo_pstate_strategy *expanded_strategy_list,
+ unsigned int *num_expanded_strategies);
+
#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_factory.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_factory.c
index add51d41a515..7ed0242a4b33 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_factory.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_factory.c
@@ -72,7 +72,6 @@ bool dml2_pmo_create(enum dml2_project_id project_id, struct dml2_pmo_instance *
out->init_for_stutter = pmo_dcn4_fams2_init_for_stutter;
out->test_for_stutter = pmo_dcn4_fams2_test_for_stutter;
out->optimize_for_stutter = pmo_dcn4_fams2_optimize_for_stutter;
-
result = true;
break;
case dml2_project_invalid:
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_interfaces.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_interfaces.c
new file mode 100644
index 000000000000..f88931ccbc5e
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_interfaces.c
@@ -0,0 +1,50 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+#include "dml_top.h"
+#include "dml2_internal_shared_types.h"
+#include "dml2_top_soc15.h"
+
+unsigned int dml2_get_instance_size_bytes(void)
+{
+ return sizeof(struct dml2_instance);
+}
+
+bool dml2_initialize_instance(struct dml2_initialize_instance_in_out *in_out)
+{
+ switch (in_out->options.project_id) {
+ case dml2_project_dcn4x_stage1:
+ case dml2_project_dcn4x_stage2:
+ case dml2_project_dcn4x_stage2_auto_drr_svp:
+ return dml2_top_soc15_initialize_instance(in_out);
+ case dml2_project_invalid:
+ default:
+ return false;
+ }
+}
+
+bool dml2_check_mode_supported(struct dml2_check_mode_supported_in_out *in_out)
+{
+ if (!in_out->dml2_instance->funcs.check_mode_supported)
+ return false;
+
+ return in_out->dml2_instance->funcs.check_mode_supported(in_out);
+}
+
+bool dml2_build_mode_programming(struct dml2_build_mode_programming_in_out *in_out)
+{
+ if (!in_out->dml2_instance->funcs.build_mode_programming)
+ return false;
+
+ return in_out->dml2_instance->funcs.build_mode_programming(in_out);
+}
+
+bool dml2_build_mcache_programming(struct dml2_build_mcache_programming_in_out *in_out)
+{
+ if (!in_out->dml2_instance->funcs.build_mcache_programming)
+ return false;
+
+ return in_out->dml2_instance->funcs.build_mcache_programming(in_out);
+}
+
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_legacy.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_legacy.c
new file mode 100644
index 000000000000..5e14d85821e2
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_legacy.c
@@ -0,0 +1,10 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+#include "dml2_top_legacy.h"
+#include "dml2_top_soc15.h"
+#include "dml2_core_factory.h"
+#include "dml2_pmo_factory.h"
+#include "display_mode_core_structs.h"
+
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_legacy.h b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_legacy.h
new file mode 100644
index 000000000000..14d0ae03dce6
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_legacy.h
@@ -0,0 +1,9 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+#ifndef __DML2_TOP_LEGACY_H__
+#define __DML2_TOP_LEGACY_H__
+#include "dml2_internal_shared_types.h"
+bool dml2_top_legacy_initialize_instance(struct dml2_initialize_instance_in_out *in_out);
+#endif /* __DML2_TOP_LEGACY_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_optimization.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_optimization.c
deleted file mode 100644
index d0e026d981b5..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_optimization.c
+++ /dev/null
@@ -1,307 +0,0 @@
-// SPDX-License-Identifier: MIT
-//
-// Copyright 2024 Advanced Micro Devices, Inc.
-
-#include "dml2_top_optimization.h"
-#include "dml2_internal_shared_types.h"
-#include "dml_top_mcache.h"
-
-static void copy_display_configuration_with_meta(struct display_configuation_with_meta *dst, const struct display_configuation_with_meta *src)
-{
- memcpy(dst, src, sizeof(struct display_configuation_with_meta));
-}
-
-bool dml2_top_optimization_init_function_min_clk_for_latency(const struct optimization_init_function_params *params)
-{
- struct dml2_optimization_stage1_state *state = &params->display_config->stage1;
-
- state->performed = true;
-
- return true;
-}
-
-bool dml2_top_optimization_test_function_min_clk_for_latency(const struct optimization_test_function_params *params)
-{
- struct dml2_optimization_stage1_state *state = &params->display_config->stage1;
-
- return state->min_clk_index_for_latency == 0;
-}
-
-bool dml2_top_optimization_optimize_function_min_clk_for_latency(const struct optimization_optimize_function_params *params)
-{
- bool result = false;
-
- if (params->display_config->stage1.min_clk_index_for_latency > 0) {
- copy_display_configuration_with_meta(params->optimized_display_config, params->display_config);
- params->optimized_display_config->stage1.min_clk_index_for_latency--;
- result = true;
- }
-
- return result;
-}
-
-bool dml2_top_optimization_test_function_mcache(const struct optimization_test_function_params *params)
-{
- struct dml2_optimization_test_function_locals *l = params->locals;
- bool mcache_success = false;
- bool result = false;
-
- memset(l, 0, sizeof(struct dml2_optimization_test_function_locals));
-
- l->test_mcache.calc_mcache_count_params.dml2_instance = params->dml;
- l->test_mcache.calc_mcache_count_params.display_config = &params->display_config->display_config;
- l->test_mcache.calc_mcache_count_params.mcache_allocations = params->display_config->stage2.mcache_allocations;
-
- result = dml2_top_mcache_calc_mcache_count_and_offsets(&l->test_mcache.calc_mcache_count_params); // use core to get the basic mcache_allocations
-
- if (result) {
- l->test_mcache.assign_global_mcache_ids_params.allocations = params->display_config->stage2.mcache_allocations;
- l->test_mcache.assign_global_mcache_ids_params.num_allocations = params->display_config->display_config.num_planes;
-
- dml2_top_mcache_assign_global_mcache_ids(&l->test_mcache.assign_global_mcache_ids_params);
-
- l->test_mcache.validate_admissibility_params.dml2_instance = params->dml;
- l->test_mcache.validate_admissibility_params.display_cfg = &params->display_config->display_config;
- l->test_mcache.validate_admissibility_params.mcache_allocations = params->display_config->stage2.mcache_allocations;
- l->test_mcache.validate_admissibility_params.cfg_support_info = &params->display_config->mode_support_result.cfg_support_info;
-
- mcache_success = dml2_top_mcache_validate_admissability(&l->test_mcache.validate_admissibility_params); // also find the shift to make mcache allocation works
-
- memcpy(params->display_config->stage2.per_plane_mcache_support, l->test_mcache.validate_admissibility_params.per_plane_status, sizeof(bool) * DML2_MAX_PLANES);
- }
-
- return mcache_success;
-}
-
-bool dml2_top_optimization_optimize_function_mcache(const struct optimization_optimize_function_params *params)
-{
- struct dml2_optimization_optimize_function_locals *l = params->locals;
- bool optimize_success = false;
-
- if (params->last_candidate_supported == false)
- return false;
-
- copy_display_configuration_with_meta(params->optimized_display_config, params->display_config);
-
- l->optimize_mcache.optimize_mcache_params.instance = &params->dml->pmo_instance;
- l->optimize_mcache.optimize_mcache_params.dcc_mcache_supported = params->display_config->stage2.per_plane_mcache_support;
- l->optimize_mcache.optimize_mcache_params.display_config = &params->display_config->display_config;
- l->optimize_mcache.optimize_mcache_params.optimized_display_cfg = &params->optimized_display_config->display_config;
- l->optimize_mcache.optimize_mcache_params.cfg_support_info = &params->optimized_display_config->mode_support_result.cfg_support_info;
-
- optimize_success = params->dml->pmo_instance.optimize_dcc_mcache(&l->optimize_mcache.optimize_mcache_params);
-
- return optimize_success;
-}
-
-bool dml2_top_optimization_init_function_vmin(const struct optimization_init_function_params *params)
-{
- struct dml2_optimization_init_function_locals *l = params->locals;
-
- l->vmin.init_params.instance = &params->dml->pmo_instance;
- l->vmin.init_params.base_display_config = params->display_config;
- return params->dml->pmo_instance.init_for_vmin(&l->vmin.init_params);
-}
-
-bool dml2_top_optimization_test_function_vmin(const struct optimization_test_function_params *params)
-{
- struct dml2_optimization_test_function_locals *l = params->locals;
-
- l->test_vmin.pmo_test_vmin_params.instance = &params->dml->pmo_instance;
- l->test_vmin.pmo_test_vmin_params.display_config = params->display_config;
- l->test_vmin.pmo_test_vmin_params.vmin_limits = &params->dml->soc_bbox.vmin_limit;
- return params->dml->pmo_instance.test_for_vmin(&l->test_vmin.pmo_test_vmin_params);
-}
-
-bool dml2_top_optimization_optimize_function_vmin(const struct optimization_optimize_function_params *params)
-{
- struct dml2_optimization_optimize_function_locals *l = params->locals;
-
- if (params->last_candidate_supported == false)
- return false;
-
- l->optimize_vmin.pmo_optimize_vmin_params.instance = &params->dml->pmo_instance;
- l->optimize_vmin.pmo_optimize_vmin_params.base_display_config = params->display_config;
- l->optimize_vmin.pmo_optimize_vmin_params.optimized_display_config = params->optimized_display_config;
- return params->dml->pmo_instance.optimize_for_vmin(&l->optimize_vmin.pmo_optimize_vmin_params);
-}
-
-bool dml2_top_optimization_perform_optimization_phase(struct dml2_optimization_phase_locals *l, const struct optimization_phase_params *params)
-{
- bool test_passed = false;
- bool optimize_succeeded = true;
- bool candidate_validation_passed = true;
- struct optimization_init_function_params init_params = { 0 };
- struct optimization_test_function_params test_params = { 0 };
- struct optimization_optimize_function_params optimize_params = { 0 };
-
- if (!params->dml ||
- !params->optimize_function ||
- !params->test_function ||
- !params->display_config ||
- !params->optimized_display_config)
- return false;
-
- copy_display_configuration_with_meta(&l->cur_candidate_display_cfg, params->display_config);
-
- init_params.locals = &l->init_function_locals;
- init_params.dml = params->dml;
- init_params.display_config = &l->cur_candidate_display_cfg;
-
- if (params->init_function && !params->init_function(&init_params))
- return false;
-
- test_params.locals = &l->test_function_locals;
- test_params.dml = params->dml;
- test_params.display_config = &l->cur_candidate_display_cfg;
-
- test_passed = params->test_function(&test_params);
-
- while (!test_passed && optimize_succeeded) {
- memset(&optimize_params, 0, sizeof(struct optimization_optimize_function_params));
-
- optimize_params.locals = &l->optimize_function_locals;
- optimize_params.dml = params->dml;
- optimize_params.display_config = &l->cur_candidate_display_cfg;
- optimize_params.optimized_display_config = &l->next_candidate_display_cfg;
- optimize_params.last_candidate_supported = candidate_validation_passed;
-
- optimize_succeeded = params->optimize_function(&optimize_params);
-
- if (optimize_succeeded) {
- l->mode_support_params.instance = &params->dml->core_instance;
- l->mode_support_params.display_cfg = &l->next_candidate_display_cfg;
- l->mode_support_params.min_clk_table = &params->dml->min_clk_table;
-
- if (l->next_candidate_display_cfg.stage3.performed)
- l->mode_support_params.min_clk_index = l->next_candidate_display_cfg.stage3.min_clk_index_for_latency;
- else
- l->mode_support_params.min_clk_index = l->next_candidate_display_cfg.stage1.min_clk_index_for_latency;
-
- candidate_validation_passed = params->dml->core_instance.mode_support(&l->mode_support_params);
-
- l->next_candidate_display_cfg.mode_support_result = l->mode_support_params.mode_support_result;
- }
-
- if (optimize_succeeded && candidate_validation_passed) {
- memset(&test_params, 0, sizeof(struct optimization_test_function_params));
- test_params.locals = &l->test_function_locals;
- test_params.dml = params->dml;
- test_params.display_config = &l->next_candidate_display_cfg;
- test_passed = params->test_function(&test_params);
-
- copy_display_configuration_with_meta(&l->cur_candidate_display_cfg, &l->next_candidate_display_cfg);
-
- // If optimization is not all or nothing, then store partial progress in output
- if (!params->all_or_nothing)
- copy_display_configuration_with_meta(params->optimized_display_config, &l->next_candidate_display_cfg);
- }
- }
-
- if (test_passed)
- copy_display_configuration_with_meta(params->optimized_display_config, &l->cur_candidate_display_cfg);
-
- return test_passed;
-}
-
-bool dml2_top_optimization_perform_optimization_phase_1(struct dml2_optimization_phase_locals *l, const struct optimization_phase_params *params)
-{
- int highest_state, lowest_state, cur_state;
- bool supported = false;
-
- if (!params->dml ||
- !params->optimize_function ||
- !params->test_function ||
- !params->display_config ||
- !params->optimized_display_config)
- return false;
-
- copy_display_configuration_with_meta(&l->cur_candidate_display_cfg, params->display_config);
- highest_state = l->cur_candidate_display_cfg.stage1.min_clk_index_for_latency;
- lowest_state = 0;
-
- while (highest_state > lowest_state) {
- cur_state = (highest_state + lowest_state) / 2;
-
- l->mode_support_params.instance = &params->dml->core_instance;
- l->mode_support_params.display_cfg = &l->cur_candidate_display_cfg;
- l->mode_support_params.min_clk_table = &params->dml->min_clk_table;
- l->mode_support_params.min_clk_index = cur_state;
-
- supported = params->dml->core_instance.mode_support(&l->mode_support_params);
-
- if (supported) {
- l->cur_candidate_display_cfg.mode_support_result = l->mode_support_params.mode_support_result;
- highest_state = cur_state;
- } else {
- lowest_state = cur_state + 1;
- }
- }
- l->cur_candidate_display_cfg.stage1.min_clk_index_for_latency = lowest_state;
-
- copy_display_configuration_with_meta(params->optimized_display_config, &l->cur_candidate_display_cfg);
-
- return true;
-}
-
-bool dml2_top_optimization_init_function_uclk_pstate(const struct optimization_init_function_params *params)
-{
- struct dml2_optimization_init_function_locals *l = params->locals;
-
- l->uclk_pstate.init_params.instance = &params->dml->pmo_instance;
- l->uclk_pstate.init_params.base_display_config = params->display_config;
-
- return params->dml->pmo_instance.init_for_uclk_pstate(&l->uclk_pstate.init_params);
-}
-
-bool dml2_top_optimization_test_function_uclk_pstate(const struct optimization_test_function_params *params)
-{
- struct dml2_optimization_test_function_locals *l = params->locals;
-
- l->uclk_pstate.test_params.instance = &params->dml->pmo_instance;
- l->uclk_pstate.test_params.base_display_config = params->display_config;
-
- return params->dml->pmo_instance.test_for_uclk_pstate(&l->uclk_pstate.test_params);
-}
-
-bool dml2_top_optimization_optimize_function_uclk_pstate(const struct optimization_optimize_function_params *params)
-{
- struct dml2_optimization_optimize_function_locals *l = params->locals;
-
- l->uclk_pstate.optimize_params.instance = &params->dml->pmo_instance;
- l->uclk_pstate.optimize_params.base_display_config = params->display_config;
- l->uclk_pstate.optimize_params.optimized_display_config = params->optimized_display_config;
- l->uclk_pstate.optimize_params.last_candidate_failed = !params->last_candidate_supported;
-
- return params->dml->pmo_instance.optimize_for_uclk_pstate(&l->uclk_pstate.optimize_params);
-}
-
-bool dml2_top_optimization_init_function_stutter(const struct optimization_init_function_params *params)
-{
- struct dml2_optimization_init_function_locals *l = params->locals;
-
- l->uclk_pstate.init_params.instance = &params->dml->pmo_instance;
- l->uclk_pstate.init_params.base_display_config = params->display_config;
-
- return params->dml->pmo_instance.init_for_stutter(&l->stutter.stutter_params);
-}
-
-bool dml2_top_optimization_test_function_stutter(const struct optimization_test_function_params *params)
-{
- struct dml2_optimization_test_function_locals *l = params->locals;
-
- l->stutter.stutter_params.instance = &params->dml->pmo_instance;
- l->stutter.stutter_params.base_display_config = params->display_config;
- return params->dml->pmo_instance.test_for_stutter(&l->stutter.stutter_params);
-}
-
-bool dml2_top_optimization_optimize_function_stutter(const struct optimization_optimize_function_params *params)
-{
- struct dml2_optimization_optimize_function_locals *l = params->locals;
-
- l->stutter.stutter_params.instance = &params->dml->pmo_instance;
- l->stutter.stutter_params.base_display_config = params->display_config;
- l->stutter.stutter_params.optimized_display_config = params->optimized_display_config;
- l->stutter.stutter_params.last_candidate_failed = !params->last_candidate_supported;
- return params->dml->pmo_instance.optimize_for_stutter(&l->stutter.stutter_params);
-}
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_optimization.h b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_optimization.h
deleted file mode 100644
index 9f22ab33eab1..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_optimization.h
+++ /dev/null
@@ -1,33 +0,0 @@
-// SPDX-License-Identifier: MIT
-//
-// Copyright 2024 Advanced Micro Devices, Inc.
-
-#ifndef __DML2_TOP_OPTIMIZATION_H__
-#define __DML2_TOP_OPTIMIZATION_H__
-
-#include "dml2_external_lib_deps.h"
-#include "dml2_internal_shared_types.h"
-
-bool dml2_top_optimization_perform_optimization_phase(struct dml2_optimization_phase_locals *l, const struct optimization_phase_params *params);
-bool dml2_top_optimization_perform_optimization_phase_1(struct dml2_optimization_phase_locals *l, const struct optimization_phase_params *params);
-
-bool dml2_top_optimization_init_function_min_clk_for_latency(const struct optimization_init_function_params *params);
-bool dml2_top_optimization_test_function_min_clk_for_latency(const struct optimization_test_function_params *params);
-bool dml2_top_optimization_optimize_function_min_clk_for_latency(const struct optimization_optimize_function_params *params);
-
-bool dml2_top_optimization_test_function_mcache(const struct optimization_test_function_params *params);
-bool dml2_top_optimization_optimize_function_mcache(const struct optimization_optimize_function_params *params);
-
-bool dml2_top_optimization_init_function_uclk_pstate(const struct optimization_init_function_params *params);
-bool dml2_top_optimization_test_function_uclk_pstate(const struct optimization_test_function_params *params);
-bool dml2_top_optimization_optimize_function_uclk_pstate(const struct optimization_optimize_function_params *params);
-
-bool dml2_top_optimization_init_function_vmin(const struct optimization_init_function_params *params);
-bool dml2_top_optimization_test_function_vmin(const struct optimization_test_function_params *params);
-bool dml2_top_optimization_optimize_function_vmin(const struct optimization_optimize_function_params *params);
-
-bool dml2_top_optimization_init_function_stutter(const struct optimization_init_function_params *params);
-bool dml2_top_optimization_test_function_stutter(const struct optimization_test_function_params *params);
-bool dml2_top_optimization_optimize_function_stutter(const struct optimization_optimize_function_params *params);
-
-#endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_soc15.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_soc15.c
new file mode 100644
index 000000000000..a8f58f8448e4
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_soc15.c
@@ -0,0 +1,1178 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+#include "dml2_top_soc15.h"
+#include "dml2_mcg_factory.h"
+#include "dml2_dpmm_factory.h"
+#include "dml2_core_factory.h"
+#include "dml2_pmo_factory.h"
+#include "lib_float_math.h"
+#include "dml2_debug.h"
+static void setup_unoptimized_display_config_with_meta(const struct dml2_instance *dml, struct display_configuation_with_meta *out, const struct dml2_display_cfg *display_config)
+{
+ memcpy(&out->display_config, display_config, sizeof(struct dml2_display_cfg));
+ out->stage1.min_clk_index_for_latency = dml->min_clk_table.dram_bw_table.num_entries - 1; //dml->min_clk_table.clean_me_up.soc_bb.num_states - 1;
+}
+
+static void setup_speculative_display_config_with_meta(const struct dml2_instance *dml, struct display_configuation_with_meta *out, const struct dml2_display_cfg *display_config)
+{
+ memcpy(&out->display_config, display_config, sizeof(struct dml2_display_cfg));
+ out->stage1.min_clk_index_for_latency = 0;
+}
+
+static void copy_display_configuration_with_meta(struct display_configuation_with_meta *dst, const struct display_configuation_with_meta *src)
+{
+ memcpy(dst, src, sizeof(struct display_configuation_with_meta));
+}
+
+static bool dml2_top_optimization_init_function_min_clk_for_latency(const struct optimization_init_function_params *params)
+{
+ struct dml2_optimization_stage1_state *state = &params->display_config->stage1;
+
+ state->performed = true;
+
+ return true;
+}
+
+static bool dml2_top_optimization_test_function_min_clk_for_latency(const struct optimization_test_function_params *params)
+{
+ struct dml2_optimization_stage1_state *state = &params->display_config->stage1;
+
+ return state->min_clk_index_for_latency == 0;
+}
+
+static bool dml2_top_optimization_optimize_function_min_clk_for_latency(const struct optimization_optimize_function_params *params)
+{
+ bool result = false;
+
+ if (params->display_config->stage1.min_clk_index_for_latency > 0) {
+ copy_display_configuration_with_meta(params->optimized_display_config, params->display_config);
+ params->optimized_display_config->stage1.min_clk_index_for_latency--;
+ result = true;
+ }
+
+ return result;
+}
+
+static bool dml2_top_optimization_test_function_mcache(const struct optimization_test_function_params *params)
+{
+ struct dml2_optimization_test_function_locals *l = params->locals;
+ bool mcache_success = false;
+ bool result = false;
+
+ memset(l, 0, sizeof(struct dml2_optimization_test_function_locals));
+
+ l->test_mcache.calc_mcache_count_params.dml2_instance = params->dml;
+ l->test_mcache.calc_mcache_count_params.display_config = &params->display_config->display_config;
+ l->test_mcache.calc_mcache_count_params.mcache_allocations = params->display_config->stage2.mcache_allocations;
+
+ result = dml2_top_mcache_calc_mcache_count_and_offsets(&l->test_mcache.calc_mcache_count_params); // use core to get the basic mcache_allocations
+
+ if (result) {
+ l->test_mcache.assign_global_mcache_ids_params.allocations = params->display_config->stage2.mcache_allocations;
+ l->test_mcache.assign_global_mcache_ids_params.num_allocations = params->display_config->display_config.num_planes;
+
+ dml2_top_mcache_assign_global_mcache_ids(&l->test_mcache.assign_global_mcache_ids_params);
+
+ l->test_mcache.validate_admissibility_params.dml2_instance = params->dml;
+ l->test_mcache.validate_admissibility_params.display_cfg = &params->display_config->display_config;
+ l->test_mcache.validate_admissibility_params.mcache_allocations = params->display_config->stage2.mcache_allocations;
+ l->test_mcache.validate_admissibility_params.cfg_support_info = &params->display_config->mode_support_result.cfg_support_info;
+
+ mcache_success = dml2_top_mcache_validate_admissability(&l->test_mcache.validate_admissibility_params); // also find the shift to make mcache allocation works
+
+ memcpy(params->display_config->stage2.per_plane_mcache_support, l->test_mcache.validate_admissibility_params.per_plane_status, sizeof(bool) * DML2_MAX_PLANES);
+ }
+
+ return mcache_success;
+}
+
+static bool dml2_top_optimization_optimize_function_mcache(const struct optimization_optimize_function_params *params)
+{
+ struct dml2_optimization_optimize_function_locals *l = params->locals;
+ bool optimize_success = false;
+
+ if (params->last_candidate_supported == false)
+ return false;
+
+ copy_display_configuration_with_meta(params->optimized_display_config, params->display_config);
+
+ l->optimize_mcache.optimize_mcache_params.instance = &params->dml->pmo_instance;
+ l->optimize_mcache.optimize_mcache_params.dcc_mcache_supported = params->display_config->stage2.per_plane_mcache_support;
+ l->optimize_mcache.optimize_mcache_params.display_config = &params->display_config->display_config;
+ l->optimize_mcache.optimize_mcache_params.optimized_display_cfg = &params->optimized_display_config->display_config;
+ l->optimize_mcache.optimize_mcache_params.cfg_support_info = &params->optimized_display_config->mode_support_result.cfg_support_info;
+
+ optimize_success = params->dml->pmo_instance.optimize_dcc_mcache(&l->optimize_mcache.optimize_mcache_params);
+
+ return optimize_success;
+}
+
+static bool dml2_top_optimization_init_function_vmin(const struct optimization_init_function_params *params)
+{
+ struct dml2_optimization_init_function_locals *l = params->locals;
+
+ l->vmin.init_params.instance = &params->dml->pmo_instance;
+ l->vmin.init_params.base_display_config = params->display_config;
+ return params->dml->pmo_instance.init_for_vmin(&l->vmin.init_params);
+}
+
+static bool dml2_top_optimization_test_function_vmin(const struct optimization_test_function_params *params)
+{
+ struct dml2_optimization_test_function_locals *l = params->locals;
+
+ l->test_vmin.pmo_test_vmin_params.instance = &params->dml->pmo_instance;
+ l->test_vmin.pmo_test_vmin_params.display_config = params->display_config;
+ l->test_vmin.pmo_test_vmin_params.vmin_limits = &params->dml->soc_bbox.vmin_limit;
+ return params->dml->pmo_instance.test_for_vmin(&l->test_vmin.pmo_test_vmin_params);
+}
+
+static bool dml2_top_optimization_optimize_function_vmin(const struct optimization_optimize_function_params *params)
+{
+ struct dml2_optimization_optimize_function_locals *l = params->locals;
+
+ if (params->last_candidate_supported == false)
+ return false;
+
+ l->optimize_vmin.pmo_optimize_vmin_params.instance = &params->dml->pmo_instance;
+ l->optimize_vmin.pmo_optimize_vmin_params.base_display_config = params->display_config;
+ l->optimize_vmin.pmo_optimize_vmin_params.optimized_display_config = params->optimized_display_config;
+ return params->dml->pmo_instance.optimize_for_vmin(&l->optimize_vmin.pmo_optimize_vmin_params);
+}
+
+static bool dml2_top_optimization_init_function_uclk_pstate(const struct optimization_init_function_params *params)
+{
+ struct dml2_optimization_init_function_locals *l = params->locals;
+
+ l->uclk_pstate.init_params.instance = &params->dml->pmo_instance;
+ l->uclk_pstate.init_params.base_display_config = params->display_config;
+
+ return params->dml->pmo_instance.init_for_uclk_pstate(&l->uclk_pstate.init_params);
+}
+
+static bool dml2_top_optimization_test_function_uclk_pstate(const struct optimization_test_function_params *params)
+{
+ struct dml2_optimization_test_function_locals *l = params->locals;
+
+ l->uclk_pstate.test_params.instance = &params->dml->pmo_instance;
+ l->uclk_pstate.test_params.base_display_config = params->display_config;
+
+ return params->dml->pmo_instance.test_for_uclk_pstate(&l->uclk_pstate.test_params);
+}
+
+static bool dml2_top_optimization_optimize_function_uclk_pstate(const struct optimization_optimize_function_params *params)
+{
+ struct dml2_optimization_optimize_function_locals *l = params->locals;
+
+ l->uclk_pstate.optimize_params.instance = &params->dml->pmo_instance;
+ l->uclk_pstate.optimize_params.base_display_config = params->display_config;
+ l->uclk_pstate.optimize_params.optimized_display_config = params->optimized_display_config;
+ l->uclk_pstate.optimize_params.last_candidate_failed = !params->last_candidate_supported;
+
+ return params->dml->pmo_instance.optimize_for_uclk_pstate(&l->uclk_pstate.optimize_params);
+}
+
+static bool dml2_top_optimization_init_function_stutter(const struct optimization_init_function_params *params)
+{
+ struct dml2_optimization_init_function_locals *l = params->locals;
+
+ l->uclk_pstate.init_params.instance = &params->dml->pmo_instance;
+ l->uclk_pstate.init_params.base_display_config = params->display_config;
+
+ return params->dml->pmo_instance.init_for_stutter(&l->stutter.stutter_params);
+}
+
+static bool dml2_top_optimization_test_function_stutter(const struct optimization_test_function_params *params)
+{
+ struct dml2_optimization_test_function_locals *l = params->locals;
+
+ l->stutter.stutter_params.instance = &params->dml->pmo_instance;
+ l->stutter.stutter_params.base_display_config = params->display_config;
+ return params->dml->pmo_instance.test_for_stutter(&l->stutter.stutter_params);
+}
+
+static bool dml2_top_optimization_optimize_function_stutter(const struct optimization_optimize_function_params *params)
+{
+ struct dml2_optimization_optimize_function_locals *l = params->locals;
+
+ l->stutter.stutter_params.instance = &params->dml->pmo_instance;
+ l->stutter.stutter_params.base_display_config = params->display_config;
+ l->stutter.stutter_params.optimized_display_config = params->optimized_display_config;
+ l->stutter.stutter_params.last_candidate_failed = !params->last_candidate_supported;
+ return params->dml->pmo_instance.optimize_for_stutter(&l->stutter.stutter_params);
+}
+
+static bool dml2_top_optimization_perform_optimization_phase(struct dml2_optimization_phase_locals *l, const struct optimization_phase_params *params)
+{
+ bool test_passed = false;
+ bool optimize_succeeded = true;
+ bool candidate_validation_passed = true;
+ struct optimization_init_function_params init_params = { 0 };
+ struct optimization_test_function_params test_params = { 0 };
+ struct optimization_optimize_function_params optimize_params = { 0 };
+
+ if (!params->dml ||
+ !params->optimize_function ||
+ !params->test_function ||
+ !params->display_config ||
+ !params->optimized_display_config)
+ return false;
+
+ copy_display_configuration_with_meta(&l->cur_candidate_display_cfg, params->display_config);
+
+ init_params.locals = &l->init_function_locals;
+ init_params.dml = params->dml;
+ init_params.display_config = &l->cur_candidate_display_cfg;
+
+ if (params->init_function && !params->init_function(&init_params))
+ return false;
+
+ test_params.locals = &l->test_function_locals;
+ test_params.dml = params->dml;
+ test_params.display_config = &l->cur_candidate_display_cfg;
+
+ test_passed = params->test_function(&test_params);
+
+ while (!test_passed && optimize_succeeded) {
+ memset(&optimize_params, 0, sizeof(struct optimization_optimize_function_params));
+
+ optimize_params.locals = &l->optimize_function_locals;
+ optimize_params.dml = params->dml;
+ optimize_params.display_config = &l->cur_candidate_display_cfg;
+ optimize_params.optimized_display_config = &l->next_candidate_display_cfg;
+ optimize_params.last_candidate_supported = candidate_validation_passed;
+
+ optimize_succeeded = params->optimize_function(&optimize_params);
+
+ if (optimize_succeeded) {
+ l->mode_support_params.instance = &params->dml->core_instance;
+ l->mode_support_params.display_cfg = &l->next_candidate_display_cfg;
+ l->mode_support_params.min_clk_table = &params->dml->min_clk_table;
+
+ if (l->next_candidate_display_cfg.stage3.performed)
+ l->mode_support_params.min_clk_index = l->next_candidate_display_cfg.stage3.min_clk_index_for_latency;
+ else
+ l->mode_support_params.min_clk_index = l->next_candidate_display_cfg.stage1.min_clk_index_for_latency;
+ candidate_validation_passed = params->dml->core_instance.mode_support(&l->mode_support_params);
+ l->next_candidate_display_cfg.mode_support_result = l->mode_support_params.mode_support_result;
+ }
+
+ if (optimize_succeeded && candidate_validation_passed) {
+ memset(&test_params, 0, sizeof(struct optimization_test_function_params));
+ test_params.locals = &l->test_function_locals;
+ test_params.dml = params->dml;
+ test_params.display_config = &l->next_candidate_display_cfg;
+ test_passed = params->test_function(&test_params);
+
+ copy_display_configuration_with_meta(&l->cur_candidate_display_cfg, &l->next_candidate_display_cfg);
+
+ // If optimization is not all or nothing, then store partial progress in output
+ if (!params->all_or_nothing)
+ copy_display_configuration_with_meta(params->optimized_display_config, &l->next_candidate_display_cfg);
+ }
+ }
+
+ if (test_passed)
+ copy_display_configuration_with_meta(params->optimized_display_config, &l->cur_candidate_display_cfg);
+
+ return test_passed;
+}
+
+static bool dml2_top_optimization_perform_optimization_phase_1(struct dml2_optimization_phase_locals *l, const struct optimization_phase_params *params)
+{
+ int highest_state, lowest_state, cur_state;
+ bool supported = false;
+
+ if (!params->dml ||
+ !params->optimize_function ||
+ !params->test_function ||
+ !params->display_config ||
+ !params->optimized_display_config)
+ return false;
+
+ copy_display_configuration_with_meta(&l->cur_candidate_display_cfg, params->display_config);
+ highest_state = l->cur_candidate_display_cfg.stage1.min_clk_index_for_latency;
+ lowest_state = 0;
+
+ while (highest_state > lowest_state) {
+ cur_state = (highest_state + lowest_state) / 2;
+
+ l->mode_support_params.instance = &params->dml->core_instance;
+ l->mode_support_params.display_cfg = &l->cur_candidate_display_cfg;
+ l->mode_support_params.min_clk_table = &params->dml->min_clk_table;
+ l->mode_support_params.min_clk_index = cur_state;
+ supported = params->dml->core_instance.mode_support(&l->mode_support_params);
+
+ if (supported) {
+ l->cur_candidate_display_cfg.mode_support_result = l->mode_support_params.mode_support_result;
+ highest_state = cur_state;
+ } else {
+ lowest_state = cur_state + 1;
+ }
+ }
+ l->cur_candidate_display_cfg.stage1.min_clk_index_for_latency = lowest_state;
+
+ copy_display_configuration_with_meta(params->optimized_display_config, &l->cur_candidate_display_cfg);
+
+ return true;
+}
+
+/*
+* Takes an input set of mcache boundaries and finds the appropriate setting of cache programming.
+* Returns true if a valid set of programming can be made, and false otherwise. "Valid" means
+* that the horizontal viewport does not span more than 2 cache slices.
+*
+* It optionally also can apply a constant shift to all the cache boundaries.
+*/
+static const uint32_t MCACHE_ID_UNASSIGNED = 0xF;
+static const uint32_t SPLIT_LOCATION_UNDEFINED = 0xFFFF;
+
+static bool calculate_first_second_splitting(const int *mcache_boundaries, int num_boundaries, int shift,
+ int pipe_h_vp_start, int pipe_h_vp_end, int *first_offset, int *second_offset)
+{
+ const int MAX_VP = 0xFFFFFF;
+ int left_cache_id;
+ int right_cache_id;
+ int range_start;
+ int range_end;
+ bool success = false;
+
+ if (num_boundaries <= 1) {
+ if (first_offset && second_offset) {
+ *first_offset = 0;
+ *second_offset = -1;
+ }
+ success = true;
+ return success;
+ } else {
+ range_start = 0;
+ for (left_cache_id = 0; left_cache_id < num_boundaries; left_cache_id++) {
+ range_end = mcache_boundaries[left_cache_id] - shift - 1;
+
+ if (range_start <= pipe_h_vp_start && pipe_h_vp_start <= range_end)
+ break;
+
+ range_start = range_end + 1;
+ }
+
+ range_end = MAX_VP;
+ for (right_cache_id = num_boundaries - 1; right_cache_id >= -1; right_cache_id--) {
+ if (right_cache_id >= 0)
+ range_start = mcache_boundaries[right_cache_id] - shift;
+ else
+ range_start = 0;
+
+ if (range_start <= pipe_h_vp_end && pipe_h_vp_end <= range_end) {
+ break;
+ }
+ range_end = range_start - 1;
+ }
+ right_cache_id = (right_cache_id + 1) % num_boundaries;
+
+ if (right_cache_id == left_cache_id) {
+ if (first_offset && second_offset) {
+ *first_offset = left_cache_id;
+ *second_offset = -1;
+ }
+ success = true;
+ } else if (right_cache_id == (left_cache_id + 1) % num_boundaries) {
+ if (first_offset && second_offset) {
+ *first_offset = left_cache_id;
+ *second_offset = right_cache_id;
+ }
+ success = true;
+ }
+ }
+
+ return success;
+}
+
+/*
+* For a given set of pipe start/end x positions, checks to see it can support the input mcache splitting.
+* It also attempts to "optimize" by finding a shift if the default 0 shift does not work.
+*/
+static bool find_shift_for_valid_cache_id_assignment(int *mcache_boundaries, unsigned int num_boundaries,
+ int *pipe_vp_startx, int *pipe_vp_endx, unsigned int pipe_count, int shift_granularity, int *shift)
+{
+ int max_shift = 0xFFFF;
+ unsigned int pipe_index;
+ unsigned int i, slice_width;
+ bool success = false;
+
+ for (i = 0; i < num_boundaries; i++) {
+ if (i == 0)
+ slice_width = mcache_boundaries[i];
+ else
+ slice_width = mcache_boundaries[i] - mcache_boundaries[i - 1];
+
+ if (max_shift > (int)slice_width) {
+ max_shift = slice_width;
+ }
+ }
+
+ for (*shift = 0; *shift <= max_shift; *shift += shift_granularity) {
+ success = true;
+ for (pipe_index = 0; pipe_index < pipe_count; pipe_index++) {
+ if (!calculate_first_second_splitting(mcache_boundaries, num_boundaries, *shift,
+ pipe_vp_startx[pipe_index], pipe_vp_endx[pipe_index], 0, 0)) {
+ success = false;
+ break;
+ }
+ }
+ if (success)
+ break;
+ }
+
+ return success;
+}
+
+/*
+* Counts the number of elements inside input array within the given span length.
+* Formally, what is the size of the largest subset of the array where the largest and smallest element
+* differ no more than the span.
+*/
+static unsigned int count_elements_in_span(int *array, unsigned int array_size, unsigned int span)
+{
+ unsigned int i;
+ unsigned int span_start_value;
+ unsigned int span_start_index;
+ unsigned int greatest_element_count;
+
+ if (array_size == 0)
+ return 1;
+
+ if (span == 0)
+ return array_size > 0 ? 1 : 0;
+
+ span_start_value = 0;
+ span_start_index = 0;
+ greatest_element_count = 0;
+
+ while (span_start_index < array_size) {
+ for (i = span_start_index; i < array_size; i++) {
+ if (array[i] - span_start_value <= span) {
+ if (i - span_start_index + 1 > greatest_element_count) {
+ greatest_element_count = i - span_start_index + 1;
+ }
+ } else
+ break;
+ }
+
+ span_start_index++;
+
+ if (span_start_index < array_size) {
+ span_start_value = array[span_start_index - 1] + 1;
+ }
+ }
+
+ return greatest_element_count;
+}
+
+static bool calculate_h_split_for_scaling_transform(int full_vp_width, int h_active, int num_pipes,
+ enum dml2_scaling_transform scaling_transform, int *pipe_vp_x_start, int *pipe_vp_x_end)
+{
+ int i, slice_width;
+ const char MAX_SCL_VP_OVERLAP = 3;
+ bool success = false;
+
+ switch (scaling_transform) {
+ case dml2_scaling_transform_centered:
+ case dml2_scaling_transform_aspect_ratio:
+ case dml2_scaling_transform_fullscreen:
+ slice_width = full_vp_width / num_pipes;
+ for (i = 0; i < num_pipes; i++) {
+ pipe_vp_x_start[i] = i * slice_width;
+ pipe_vp_x_end[i] = (i + 1) * slice_width - 1;
+
+ if (pipe_vp_x_start[i] < MAX_SCL_VP_OVERLAP)
+ pipe_vp_x_start[i] = 0;
+ else
+ pipe_vp_x_start[i] -= MAX_SCL_VP_OVERLAP;
+
+ if (pipe_vp_x_end[i] > full_vp_width - MAX_SCL_VP_OVERLAP - 1)
+ pipe_vp_x_end[i] = full_vp_width - 1;
+ else
+ pipe_vp_x_end[i] += MAX_SCL_VP_OVERLAP;
+ }
+ break;
+ case dml2_scaling_transform_explicit:
+ default:
+ success = false;
+ break;
+ }
+
+ return success;
+}
+
+bool dml2_top_mcache_validate_admissability(struct top_mcache_validate_admissability_in_out *params)
+{
+ struct dml2_instance *dml = (struct dml2_instance *)params->dml2_instance;
+ struct dml2_top_mcache_validate_admissability_locals *l = &dml->scratch.mcache_validate_admissability_locals;
+
+ const int MAX_PIXEL_OVERLAP = 6;
+ int max_per_pipe_vp_p0 = 0;
+ int max_per_pipe_vp_p1 = 0;
+ int temp, p0shift, p1shift;
+ unsigned int plane_index = 0;
+ unsigned int i;
+ unsigned int odm_combine_factor;
+ unsigned int mpc_combine_factor;
+ unsigned int num_dpps;
+ unsigned int num_boundaries;
+ enum dml2_scaling_transform scaling_transform;
+ const struct dml2_plane_parameters *plane;
+ const struct dml2_stream_parameters *stream;
+
+ bool p0pass = false;
+ bool p1pass = false;
+ bool all_pass = true;
+
+ for (plane_index = 0; plane_index < params->display_cfg->num_planes; plane_index++) {
+ if (!params->display_cfg->plane_descriptors[plane_index].surface.dcc.enable)
+ continue;
+
+ plane = &params->display_cfg->plane_descriptors[plane_index];
+ stream = &params->display_cfg->stream_descriptors[plane->stream_index];
+
+ num_dpps = odm_combine_factor = params->cfg_support_info->stream_support_info[plane->stream_index].odms_used;
+
+ if (odm_combine_factor == 1)
+ num_dpps = mpc_combine_factor = (unsigned int)params->cfg_support_info->plane_support_info[plane_index].dpps_used;
+ else
+ mpc_combine_factor = 1;
+
+ if (odm_combine_factor > 1) {
+ max_per_pipe_vp_p0 = plane->surface.plane0.width;
+ temp = (unsigned int)math_ceil(plane->composition.scaler_info.plane0.h_ratio * stream->timing.h_active / odm_combine_factor);
+
+ if (temp < max_per_pipe_vp_p0)
+ max_per_pipe_vp_p0 = temp;
+
+ max_per_pipe_vp_p1 = plane->surface.plane1.width;
+ temp = (unsigned int)math_ceil(plane->composition.scaler_info.plane1.h_ratio * stream->timing.h_active / odm_combine_factor);
+
+ if (temp < max_per_pipe_vp_p1)
+ max_per_pipe_vp_p1 = temp;
+ } else {
+ max_per_pipe_vp_p0 = plane->surface.plane0.width / mpc_combine_factor;
+ max_per_pipe_vp_p1 = plane->surface.plane1.width / mpc_combine_factor;
+ }
+
+ max_per_pipe_vp_p0 += 2 * MAX_PIXEL_OVERLAP;
+ max_per_pipe_vp_p1 += MAX_PIXEL_OVERLAP;
+
+ p0shift = 0;
+ p1shift = 0;
+
+ // The last element in the unshifted boundary array will always be the first pixel outside the
+ // plane, which means theres no mcache associated with it, so -1
+ num_boundaries = params->mcache_allocations[plane_index].num_mcaches_plane0 == 0 ? 0 : params->mcache_allocations[plane_index].num_mcaches_plane0 - 1;
+ if ((count_elements_in_span(params->mcache_allocations[plane_index].mcache_x_offsets_plane0,
+ num_boundaries, max_per_pipe_vp_p0) <= 1) && (num_boundaries <= num_dpps)) {
+ p0pass = true;
+ }
+ num_boundaries = params->mcache_allocations[plane_index].num_mcaches_plane1 == 0 ? 0 : params->mcache_allocations[plane_index].num_mcaches_plane1 - 1;
+ if ((count_elements_in_span(params->mcache_allocations[plane_index].mcache_x_offsets_plane1,
+ num_boundaries, max_per_pipe_vp_p1) <= 1) && (num_boundaries <= num_dpps)) {
+ p1pass = true;
+ }
+
+ if (!p0pass || !p1pass) {
+ if (odm_combine_factor > 1) {
+ num_dpps = odm_combine_factor;
+ scaling_transform = plane->composition.scaling_transform;
+ } else {
+ num_dpps = mpc_combine_factor;
+ scaling_transform = dml2_scaling_transform_fullscreen;
+ }
+
+ if (!p0pass) {
+ if (plane->composition.viewport.stationary) {
+ calculate_h_split_for_scaling_transform(plane->surface.plane0.width,
+ stream->timing.h_active, num_dpps, scaling_transform,
+ &l->plane0.pipe_vp_startx[plane_index], &l->plane0.pipe_vp_endx[plane_index]);
+ p0pass = find_shift_for_valid_cache_id_assignment(params->mcache_allocations[plane_index].mcache_x_offsets_plane0,
+ params->mcache_allocations[plane_index].num_mcaches_plane0,
+ &l->plane0.pipe_vp_startx[plane_index], &l->plane0.pipe_vp_endx[plane_index], num_dpps,
+ params->mcache_allocations[plane_index].shift_granularity.p0, &p0shift);
+ }
+ }
+ if (!p1pass) {
+ if (plane->composition.viewport.stationary) {
+ calculate_h_split_for_scaling_transform(plane->surface.plane1.width,
+ stream->timing.h_active, num_dpps, scaling_transform,
+ &l->plane0.pipe_vp_startx[plane_index], &l->plane0.pipe_vp_endx[plane_index]);
+ p1pass = find_shift_for_valid_cache_id_assignment(params->mcache_allocations[plane_index].mcache_x_offsets_plane1,
+ params->mcache_allocations[plane_index].num_mcaches_plane1,
+ &l->plane1.pipe_vp_startx[plane_index], &l->plane1.pipe_vp_endx[plane_index], num_dpps,
+ params->mcache_allocations[plane_index].shift_granularity.p1, &p1shift);
+ }
+ }
+ }
+
+ if (p0pass && p1pass) {
+ for (i = 0; i < params->mcache_allocations[plane_index].num_mcaches_plane0; i++) {
+ params->mcache_allocations[plane_index].mcache_x_offsets_plane0[i] -= p0shift;
+ }
+ for (i = 0; i < params->mcache_allocations[plane_index].num_mcaches_plane1; i++) {
+ params->mcache_allocations[plane_index].mcache_x_offsets_plane1[i] -= p1shift;
+ }
+ }
+
+ params->per_plane_status[plane_index] = p0pass && p1pass;
+ all_pass &= p0pass && p1pass;
+ }
+
+ return all_pass;
+}
+
+static void reset_mcache_allocations(struct dml2_hubp_pipe_mcache_regs *per_plane_pipe_mcache_regs)
+{
+ // Initialize all entries to special valid MCache ID and special valid split coordinate
+ per_plane_pipe_mcache_regs->main.p0.mcache_id_first = MCACHE_ID_UNASSIGNED;
+ per_plane_pipe_mcache_regs->main.p0.mcache_id_second = MCACHE_ID_UNASSIGNED;
+ per_plane_pipe_mcache_regs->main.p0.split_location = SPLIT_LOCATION_UNDEFINED;
+
+ per_plane_pipe_mcache_regs->mall.p0.mcache_id_first = MCACHE_ID_UNASSIGNED;
+ per_plane_pipe_mcache_regs->mall.p0.mcache_id_second = MCACHE_ID_UNASSIGNED;
+ per_plane_pipe_mcache_regs->mall.p0.split_location = SPLIT_LOCATION_UNDEFINED;
+
+ per_plane_pipe_mcache_regs->main.p1.mcache_id_first = MCACHE_ID_UNASSIGNED;
+ per_plane_pipe_mcache_regs->main.p1.mcache_id_second = MCACHE_ID_UNASSIGNED;
+ per_plane_pipe_mcache_regs->main.p1.split_location = SPLIT_LOCATION_UNDEFINED;
+
+ per_plane_pipe_mcache_regs->mall.p1.mcache_id_first = MCACHE_ID_UNASSIGNED;
+ per_plane_pipe_mcache_regs->mall.p1.mcache_id_second = MCACHE_ID_UNASSIGNED;
+ per_plane_pipe_mcache_regs->mall.p1.split_location = SPLIT_LOCATION_UNDEFINED;
+}
+
+void dml2_top_mcache_assign_global_mcache_ids(struct top_mcache_assign_global_mcache_ids_in_out *params)
+{
+ int i;
+ unsigned int j;
+ int next_unused_cache_id = 0;
+
+ for (i = 0; i < params->num_allocations; i++) {
+ if (!params->allocations[i].valid)
+ continue;
+
+ for (j = 0; j < params->allocations[i].num_mcaches_plane0; j++) {
+ params->allocations[i].global_mcache_ids_plane0[j] = next_unused_cache_id++;
+ }
+ for (j = 0; j < params->allocations[i].num_mcaches_plane1; j++) {
+ params->allocations[i].global_mcache_ids_plane1[j] = next_unused_cache_id++;
+ }
+
+ // The "psuedo-last" slice is always wrapped around
+ params->allocations[i].global_mcache_ids_plane0[params->allocations[i].num_mcaches_plane0] =
+ params->allocations[i].global_mcache_ids_plane0[0];
+ params->allocations[i].global_mcache_ids_plane1[params->allocations[i].num_mcaches_plane1] =
+ params->allocations[i].global_mcache_ids_plane1[0];
+
+ // If we need dedicated caches for mall requesting, then we assign them here.
+ if (params->allocations[i].requires_dedicated_mall_mcache) {
+ for (j = 0; j < params->allocations[i].num_mcaches_plane0; j++) {
+ params->allocations[i].global_mcache_ids_mall_plane0[j] = next_unused_cache_id++;
+ }
+ for (j = 0; j < params->allocations[i].num_mcaches_plane1; j++) {
+ params->allocations[i].global_mcache_ids_mall_plane1[j] = next_unused_cache_id++;
+ }
+
+ // The "psuedo-last" slice is always wrapped around
+ params->allocations[i].global_mcache_ids_mall_plane0[params->allocations[i].num_mcaches_plane0] =
+ params->allocations[i].global_mcache_ids_mall_plane0[0];
+ params->allocations[i].global_mcache_ids_mall_plane1[params->allocations[i].num_mcaches_plane1] =
+ params->allocations[i].global_mcache_ids_mall_plane1[0];
+ }
+
+ // If P0 and P1 are sharing caches, then it means the largest mcache IDs for p0 and p1 can be the same
+ // since mcache IDs are always ascending, then it means the largest mcacheID of p1 should be the
+ // largest mcacheID of P0
+ if (params->allocations[i].num_mcaches_plane0 > 0 && params->allocations[i].num_mcaches_plane1 > 0 &&
+ params->allocations[i].last_slice_sharing.plane0_plane1) {
+ params->allocations[i].global_mcache_ids_plane1[params->allocations[i].num_mcaches_plane1 - 1] =
+ params->allocations[i].global_mcache_ids_plane0[params->allocations[i].num_mcaches_plane0 - 1];
+ }
+
+ // If we need dedicated caches handle last slice sharing
+ if (params->allocations[i].requires_dedicated_mall_mcache) {
+ if (params->allocations[i].num_mcaches_plane0 > 0 && params->allocations[i].num_mcaches_plane1 > 0 &&
+ params->allocations[i].last_slice_sharing.plane0_plane1) {
+ params->allocations[i].global_mcache_ids_mall_plane1[params->allocations[i].num_mcaches_plane1 - 1] =
+ params->allocations[i].global_mcache_ids_mall_plane0[params->allocations[i].num_mcaches_plane0 - 1];
+ }
+ // If mall_comb_mcache_l is set then it means that largest mcache ID for MALL p0 can be same as regular read p0
+ if (params->allocations[i].num_mcaches_plane0 > 0 && params->allocations[i].last_slice_sharing.mall_comb_mcache_p0) {
+ params->allocations[i].global_mcache_ids_mall_plane0[params->allocations[i].num_mcaches_plane0 - 1] =
+ params->allocations[i].global_mcache_ids_plane0[params->allocations[i].num_mcaches_plane0 - 1];
+ }
+ // If mall_comb_mcache_c is set then it means that largest mcache ID for MALL p1 can be same as regular
+ // read p1 (which can be same as regular read p0 if plane0_plane1 is also set)
+ if (params->allocations[i].num_mcaches_plane1 > 0 && params->allocations[i].last_slice_sharing.mall_comb_mcache_p1) {
+ params->allocations[i].global_mcache_ids_mall_plane1[params->allocations[i].num_mcaches_plane1 - 1] =
+ params->allocations[i].global_mcache_ids_plane1[params->allocations[i].num_mcaches_plane1 - 1];
+ }
+ }
+
+ // If you don't need dedicated mall mcaches, the mall mcache assignments are identical to the normal requesting
+ if (!params->allocations[i].requires_dedicated_mall_mcache) {
+ memcpy(params->allocations[i].global_mcache_ids_mall_plane0, params->allocations[i].global_mcache_ids_plane0,
+ sizeof(params->allocations[i].global_mcache_ids_mall_plane0));
+ memcpy(params->allocations[i].global_mcache_ids_mall_plane1, params->allocations[i].global_mcache_ids_plane1,
+ sizeof(params->allocations[i].global_mcache_ids_mall_plane1));
+ }
+ }
+}
+
+bool dml2_top_mcache_calc_mcache_count_and_offsets(struct top_mcache_calc_mcache_count_and_offsets_in_out *params)
+{
+ struct dml2_instance *dml = (struct dml2_instance *)params->dml2_instance;
+ struct dml2_top_mcache_verify_mcache_size_locals *l = &dml->scratch.mcache_verify_mcache_size_locals;
+
+ unsigned int total_mcaches_required;
+ unsigned int i;
+ bool result = false;
+
+ if (dml->soc_bbox.num_dcc_mcaches == 0) {
+ return true;
+ }
+
+ total_mcaches_required = 0;
+ l->calc_mcache_params.instance = &dml->core_instance;
+ for (i = 0; i < params->display_config->num_planes; i++) {
+ if (!params->display_config->plane_descriptors[i].surface.dcc.enable) {
+ memset(&params->mcache_allocations[i], 0, sizeof(struct dml2_mcache_surface_allocation));
+ continue;
+ }
+
+ l->calc_mcache_params.plane_descriptor = &params->display_config->plane_descriptors[i];
+ l->calc_mcache_params.mcache_allocation = &params->mcache_allocations[i];
+ l->calc_mcache_params.plane_index = i;
+
+ if (!dml->core_instance.calculate_mcache_allocation(&l->calc_mcache_params)) {
+ result = false;
+ break;
+ }
+
+ if (params->mcache_allocations[i].valid) {
+ total_mcaches_required += params->mcache_allocations[i].num_mcaches_plane0 + params->mcache_allocations[i].num_mcaches_plane1;
+ if (params->mcache_allocations[i].last_slice_sharing.plane0_plane1)
+ total_mcaches_required--;
+ }
+ }
+ dml2_printf("DML_CORE_DCN3::%s: plane_%d, total_mcaches_required=%d\n", __func__, i, total_mcaches_required);
+
+ if (total_mcaches_required > dml->soc_bbox.num_dcc_mcaches) {
+ result = false;
+ } else {
+ result = true;
+ }
+
+ return result;
+}
+
+static bool dml2_top_soc15_check_mode_supported(struct dml2_check_mode_supported_in_out *in_out)
+{
+ struct dml2_instance *dml = (struct dml2_instance *)in_out->dml2_instance;
+ struct dml2_check_mode_supported_locals *l = &dml->scratch.check_mode_supported_locals;
+ struct dml2_display_cfg_programming *dpmm_programming = &dml->dpmm_instance.dpmm_scratch.programming;
+
+ bool result = false;
+ bool mcache_success = false;
+ memset(dpmm_programming, 0, sizeof(struct dml2_display_cfg_programming));
+
+ setup_unoptimized_display_config_with_meta(dml, &l->base_display_config_with_meta, in_out->display_config);
+
+ l->mode_support_params.instance = &dml->core_instance;
+ l->mode_support_params.display_cfg = &l->base_display_config_with_meta;
+ l->mode_support_params.min_clk_table = &dml->min_clk_table;
+ l->mode_support_params.min_clk_index = l->base_display_config_with_meta.stage1.min_clk_index_for_latency;
+ result = dml->core_instance.mode_support(&l->mode_support_params);
+ l->base_display_config_with_meta.mode_support_result = l->mode_support_params.mode_support_result;
+
+ if (result) {
+ struct optimization_phase_params mcache_phase = {
+ .dml = dml,
+ .display_config = &l->base_display_config_with_meta,
+ .test_function = dml2_top_optimization_test_function_mcache,
+ .optimize_function = dml2_top_optimization_optimize_function_mcache,
+ .optimized_display_config = &l->optimized_display_config_with_meta,
+ .all_or_nothing = false,
+ };
+ mcache_success = dml2_top_optimization_perform_optimization_phase(&l->optimization_phase_locals, &mcache_phase);
+ }
+
+ /*
+ * Call DPMM to map all requirements to minimum clock state
+ */
+ if (result) {
+ l->dppm_map_mode_params.min_clk_table = &dml->min_clk_table;
+ l->dppm_map_mode_params.display_cfg = &l->base_display_config_with_meta;
+ l->dppm_map_mode_params.programming = dpmm_programming;
+ l->dppm_map_mode_params.soc_bb = &dml->soc_bbox;
+ l->dppm_map_mode_params.ip = &dml->core_instance.clean_me_up.mode_lib.ip;
+ result = dml->dpmm_instance.map_mode_to_soc_dpm(&l->dppm_map_mode_params);
+ }
+
+ in_out->is_supported = mcache_success;
+ result = result && in_out->is_supported;
+
+ return result;
+}
+
+static bool dml2_top_soc15_build_mode_programming(struct dml2_build_mode_programming_in_out *in_out)
+{
+ struct dml2_instance *dml = (struct dml2_instance *)in_out->dml2_instance;
+ struct dml2_build_mode_programming_locals *l = &dml->scratch.build_mode_programming_locals;
+
+ bool result = false;
+ bool mcache_success = false;
+ bool uclk_pstate_success = false;
+ bool vmin_success = false;
+ bool stutter_success = false;
+ unsigned int i;
+
+ memset(l, 0, sizeof(struct dml2_build_mode_programming_locals));
+ memset(in_out->programming, 0, sizeof(struct dml2_display_cfg_programming));
+
+ memcpy(&in_out->programming->display_config, in_out->display_config, sizeof(struct dml2_display_cfg));
+
+ setup_speculative_display_config_with_meta(dml, &l->base_display_config_with_meta, in_out->display_config);
+
+ l->mode_support_params.instance = &dml->core_instance;
+ l->mode_support_params.display_cfg = &l->base_display_config_with_meta;
+ l->mode_support_params.min_clk_table = &dml->min_clk_table;
+ l->mode_support_params.min_clk_index = l->base_display_config_with_meta.stage1.min_clk_index_for_latency;
+ result = dml->core_instance.mode_support(&l->mode_support_params);
+
+ l->base_display_config_with_meta.mode_support_result = l->mode_support_params.mode_support_result;
+
+ if (!result) {
+ setup_unoptimized_display_config_with_meta(dml, &l->base_display_config_with_meta, in_out->display_config);
+
+ l->mode_support_params.instance = &dml->core_instance;
+ l->mode_support_params.display_cfg = &l->base_display_config_with_meta;
+ l->mode_support_params.min_clk_table = &dml->min_clk_table;
+ l->mode_support_params.min_clk_index = l->base_display_config_with_meta.stage1.min_clk_index_for_latency;
+ result = dml->core_instance.mode_support(&l->mode_support_params);
+ l->base_display_config_with_meta.mode_support_result = l->mode_support_params.mode_support_result;
+
+ if (!result) {
+ l->informative_params.instance = &dml->core_instance;
+ l->informative_params.programming = in_out->programming;
+ l->informative_params.mode_is_supported = false;
+ dml->core_instance.populate_informative(&l->informative_params);
+
+ return false;
+ }
+
+ /*
+ * Phase 1: Determine minimum clocks to satisfy latency requirements for this mode
+ */
+ memset(&l->min_clock_for_latency_phase, 0, sizeof(struct optimization_phase_params));
+ l->min_clock_for_latency_phase.dml = dml;
+ l->min_clock_for_latency_phase.display_config = &l->base_display_config_with_meta;
+ l->min_clock_for_latency_phase.init_function = dml2_top_optimization_init_function_min_clk_for_latency;
+ l->min_clock_for_latency_phase.test_function = dml2_top_optimization_test_function_min_clk_for_latency;
+ l->min_clock_for_latency_phase.optimize_function = dml2_top_optimization_optimize_function_min_clk_for_latency;
+ l->min_clock_for_latency_phase.optimized_display_config = &l->optimized_display_config_with_meta;
+ l->min_clock_for_latency_phase.all_or_nothing = false;
+
+ dml2_top_optimization_perform_optimization_phase_1(&l->optimization_phase_locals, &l->min_clock_for_latency_phase);
+
+ memcpy(&l->base_display_config_with_meta, &l->optimized_display_config_with_meta, sizeof(struct display_configuation_with_meta));
+ }
+
+ /*
+ * Phase 2: Satisfy DCC mcache requirements
+ */
+ memset(&l->mcache_phase, 0, sizeof(struct optimization_phase_params));
+ l->mcache_phase.dml = dml;
+ l->mcache_phase.display_config = &l->base_display_config_with_meta;
+ l->mcache_phase.test_function = dml2_top_optimization_test_function_mcache;
+ l->mcache_phase.optimize_function = dml2_top_optimization_optimize_function_mcache;
+ l->mcache_phase.optimized_display_config = &l->optimized_display_config_with_meta;
+ l->mcache_phase.all_or_nothing = true;
+
+ mcache_success = dml2_top_optimization_perform_optimization_phase(&l->optimization_phase_locals, &l->mcache_phase);
+
+ if (!mcache_success) {
+ l->informative_params.instance = &dml->core_instance;
+ l->informative_params.programming = in_out->programming;
+ l->informative_params.mode_is_supported = false;
+
+ dml->core_instance.populate_informative(&l->informative_params);
+
+ in_out->programming->informative.failed_mcache_validation = true;
+ return false;
+ }
+
+ memcpy(&l->base_display_config_with_meta, &l->optimized_display_config_with_meta, sizeof(struct display_configuation_with_meta));
+
+ /*
+ * Phase 3: Optimize for Pstate
+ */
+ memset(&l->uclk_pstate_phase, 0, sizeof(struct optimization_phase_params));
+ l->uclk_pstate_phase.dml = dml;
+ l->uclk_pstate_phase.display_config = &l->base_display_config_with_meta;
+ l->uclk_pstate_phase.init_function = dml2_top_optimization_init_function_uclk_pstate;
+ l->uclk_pstate_phase.test_function = dml2_top_optimization_test_function_uclk_pstate;
+ l->uclk_pstate_phase.optimize_function = dml2_top_optimization_optimize_function_uclk_pstate;
+ l->uclk_pstate_phase.optimized_display_config = &l->optimized_display_config_with_meta;
+ l->uclk_pstate_phase.all_or_nothing = true;
+
+ uclk_pstate_success = dml2_top_optimization_perform_optimization_phase(&l->optimization_phase_locals, &l->uclk_pstate_phase);
+
+ if (uclk_pstate_success) {
+ memcpy(&l->base_display_config_with_meta, &l->optimized_display_config_with_meta, sizeof(struct display_configuation_with_meta));
+ l->base_display_config_with_meta.stage3.success = true;
+ }
+
+ /*
+ * Phase 4: Optimize for Vmin
+ */
+ memset(&l->vmin_phase, 0, sizeof(struct optimization_phase_params));
+ l->vmin_phase.dml = dml;
+ l->vmin_phase.display_config = &l->base_display_config_with_meta;
+ l->vmin_phase.init_function = dml2_top_optimization_init_function_vmin;
+ l->vmin_phase.test_function = dml2_top_optimization_test_function_vmin;
+ l->vmin_phase.optimize_function = dml2_top_optimization_optimize_function_vmin;
+ l->vmin_phase.optimized_display_config = &l->optimized_display_config_with_meta;
+ l->vmin_phase.all_or_nothing = false;
+
+ vmin_success = dml2_top_optimization_perform_optimization_phase(&l->optimization_phase_locals, &l->vmin_phase);
+
+ if (l->optimized_display_config_with_meta.stage4.performed) {
+ /*
+ * when performed is true, optimization has applied to
+ * optimized_display_config_with_meta and it has passed mode
+ * support. However it may or may not pass the test function to
+ * reach actual Vmin. As long as voltage is optimized even if it
+ * doesn't reach Vmin level, there is still power benefit so in
+ * this case we will still copy this optimization into base
+ * display config.
+ */
+ memcpy(&l->base_display_config_with_meta, &l->optimized_display_config_with_meta, sizeof(struct display_configuation_with_meta));
+ l->base_display_config_with_meta.stage4.success = vmin_success;
+ }
+
+ /*
+ * Phase 5: Optimize for Stutter
+ */
+ memset(&l->stutter_phase, 0, sizeof(struct optimization_phase_params));
+ l->stutter_phase.dml = dml;
+ l->stutter_phase.display_config = &l->base_display_config_with_meta;
+ l->stutter_phase.init_function = dml2_top_optimization_init_function_stutter;
+ l->stutter_phase.test_function = dml2_top_optimization_test_function_stutter;
+ l->stutter_phase.optimize_function = dml2_top_optimization_optimize_function_stutter;
+ l->stutter_phase.optimized_display_config = &l->optimized_display_config_with_meta;
+ l->stutter_phase.all_or_nothing = true;
+
+ stutter_success = dml2_top_optimization_perform_optimization_phase(&l->optimization_phase_locals, &l->stutter_phase);
+
+ if (stutter_success) {
+ memcpy(&l->base_display_config_with_meta, &l->optimized_display_config_with_meta, sizeof(struct display_configuation_with_meta));
+ l->base_display_config_with_meta.stage5.success = true;
+ }
+
+ /*
+ * Populate mcache programming
+ */
+ for (i = 0; i < in_out->display_config->num_planes; i++) {
+ in_out->programming->plane_programming[i].mcache_allocation = l->base_display_config_with_meta.stage2.mcache_allocations[i];
+ }
+
+ /*
+ * Call DPMM to map all requirements to minimum clock state
+ */
+ if (result) {
+ l->dppm_map_mode_params.min_clk_table = &dml->min_clk_table;
+ l->dppm_map_mode_params.display_cfg = &l->base_display_config_with_meta;
+ l->dppm_map_mode_params.programming = in_out->programming;
+ l->dppm_map_mode_params.soc_bb = &dml->soc_bbox;
+ l->dppm_map_mode_params.ip = &dml->core_instance.clean_me_up.mode_lib.ip;
+ result = dml->dpmm_instance.map_mode_to_soc_dpm(&l->dppm_map_mode_params);
+ if (!result)
+ in_out->programming->informative.failed_dpmm = true;
+ }
+
+ if (result) {
+ l->mode_programming_params.instance = &dml->core_instance;
+ l->mode_programming_params.display_cfg = &l->base_display_config_with_meta;
+ l->mode_programming_params.cfg_support_info = &l->base_display_config_with_meta.mode_support_result.cfg_support_info;
+ l->mode_programming_params.programming = in_out->programming;
+ result = dml->core_instance.mode_programming(&l->mode_programming_params);
+ if (!result)
+ in_out->programming->informative.failed_mode_programming = true;
+ }
+
+ if (result) {
+ l->dppm_map_watermarks_params.core = &dml->core_instance;
+ l->dppm_map_watermarks_params.display_cfg = &l->base_display_config_with_meta;
+ l->dppm_map_watermarks_params.programming = in_out->programming;
+ result = dml->dpmm_instance.map_watermarks(&l->dppm_map_watermarks_params);
+ }
+
+ l->informative_params.instance = &dml->core_instance;
+ l->informative_params.programming = in_out->programming;
+ l->informative_params.mode_is_supported = result;
+
+ dml->core_instance.populate_informative(&l->informative_params);
+
+ return result;
+}
+
+bool dml2_top_soc15_build_mcache_programming(struct dml2_build_mcache_programming_in_out *params)
+{
+ bool success = true;
+ int config_index, pipe_index;
+ int first_offset, second_offset;
+ int free_per_plane_reg_index = 0;
+
+ memset(params->per_plane_pipe_mcache_regs, 0, DML2_MAX_PLANES * DML2_MAX_DCN_PIPES * sizeof(struct dml2_hubp_pipe_mcache_regs *));
+
+ for (config_index = 0; config_index < params->num_configurations; config_index++) {
+ for (pipe_index = 0; pipe_index < params->mcache_configurations[config_index].num_pipes; pipe_index++) {
+ // Allocate storage for the mcache regs
+ params->per_plane_pipe_mcache_regs[config_index][pipe_index] = &params->mcache_regs_set[free_per_plane_reg_index++];
+
+ reset_mcache_allocations(params->per_plane_pipe_mcache_regs[config_index][pipe_index]);
+
+ if (params->mcache_configurations[config_index].plane_descriptor->surface.dcc.enable) {
+ // P0 always enabled
+ if (!calculate_first_second_splitting(params->mcache_configurations[config_index].mcache_allocation->mcache_x_offsets_plane0,
+ params->mcache_configurations[config_index].mcache_allocation->num_mcaches_plane0,
+ 0,
+ params->mcache_configurations[config_index].pipe_configurations[pipe_index].plane0.viewport_x_start,
+ params->mcache_configurations[config_index].pipe_configurations[pipe_index].plane0.viewport_x_start +
+ params->mcache_configurations[config_index].pipe_configurations[pipe_index].plane0.viewport_width - 1,
+ &first_offset, &second_offset)) {
+ success = false;
+ break;
+ }
+
+ params->per_plane_pipe_mcache_regs[config_index][pipe_index]->main.p0.mcache_id_first =
+ params->mcache_configurations[config_index].mcache_allocation->global_mcache_ids_plane0[first_offset];
+
+ params->per_plane_pipe_mcache_regs[config_index][pipe_index]->mall.p0.mcache_id_first =
+ params->mcache_configurations[config_index].mcache_allocation->global_mcache_ids_mall_plane0[first_offset];
+
+ if (second_offset >= 0) {
+ params->per_plane_pipe_mcache_regs[config_index][pipe_index]->main.p0.mcache_id_second =
+ params->mcache_configurations[config_index].mcache_allocation->global_mcache_ids_plane0[second_offset];
+ params->per_plane_pipe_mcache_regs[config_index][pipe_index]->main.p0.split_location =
+ params->mcache_configurations[config_index].mcache_allocation->mcache_x_offsets_plane0[first_offset] - 1;
+
+ params->per_plane_pipe_mcache_regs[config_index][pipe_index]->mall.p0.mcache_id_second =
+ params->mcache_configurations[config_index].mcache_allocation->global_mcache_ids_mall_plane0[second_offset];
+ params->per_plane_pipe_mcache_regs[config_index][pipe_index]->mall.p0.split_location =
+ params->mcache_configurations[config_index].mcache_allocation->mcache_x_offsets_plane0[first_offset] - 1;
+ }
+
+ // Populate P1 if enabled
+ if (params->mcache_configurations[config_index].pipe_configurations[pipe_index].plane1_enabled) {
+ if (!calculate_first_second_splitting(params->mcache_configurations[config_index].mcache_allocation->mcache_x_offsets_plane1,
+ params->mcache_configurations[config_index].mcache_allocation->num_mcaches_plane1,
+ 0,
+ params->mcache_configurations[config_index].pipe_configurations[pipe_index].plane1.viewport_x_start,
+ params->mcache_configurations[config_index].pipe_configurations[pipe_index].plane1.viewport_x_start +
+ params->mcache_configurations[config_index].pipe_configurations[pipe_index].plane1.viewport_width - 1,
+ &first_offset, &second_offset)) {
+ success = false;
+ break;
+ }
+
+ params->per_plane_pipe_mcache_regs[config_index][pipe_index]->main.p1.mcache_id_first =
+ params->mcache_configurations[config_index].mcache_allocation->global_mcache_ids_plane1[first_offset];
+
+ params->per_plane_pipe_mcache_regs[config_index][pipe_index]->mall.p1.mcache_id_first =
+ params->mcache_configurations[config_index].mcache_allocation->global_mcache_ids_mall_plane1[first_offset];
+
+ if (second_offset >= 0) {
+ params->per_plane_pipe_mcache_regs[config_index][pipe_index]->main.p1.mcache_id_second =
+ params->mcache_configurations[config_index].mcache_allocation->global_mcache_ids_plane1[second_offset];
+ params->per_plane_pipe_mcache_regs[config_index][pipe_index]->main.p1.split_location =
+ params->mcache_configurations[config_index].mcache_allocation->mcache_x_offsets_plane1[first_offset] - 1;
+
+ params->per_plane_pipe_mcache_regs[config_index][pipe_index]->mall.p1.mcache_id_second =
+ params->mcache_configurations[config_index].mcache_allocation->global_mcache_ids_mall_plane1[second_offset];
+ params->per_plane_pipe_mcache_regs[config_index][pipe_index]->mall.p1.split_location =
+ params->mcache_configurations[config_index].mcache_allocation->mcache_x_offsets_plane1[first_offset] - 1;
+ }
+ }
+ }
+ }
+ }
+
+ return success;
+}
+
+static const struct dml2_top_funcs soc15_funcs = {
+ .check_mode_supported = dml2_top_soc15_check_mode_supported,
+ .build_mode_programming = dml2_top_soc15_build_mode_programming,
+ .build_mcache_programming = dml2_top_soc15_build_mcache_programming,
+};
+
+bool dml2_top_soc15_initialize_instance(struct dml2_initialize_instance_in_out *in_out)
+{
+ struct dml2_instance *dml = (struct dml2_instance *)in_out->dml2_instance;
+ struct dml2_initialize_instance_locals *l = &dml->scratch.initialize_instance_locals;
+ struct dml2_core_initialize_in_out core_init_params = { 0 };
+ struct dml2_mcg_build_min_clock_table_params_in_out mcg_build_min_clk_params = { 0 };
+ struct dml2_pmo_initialize_in_out pmo_init_params = { 0 };
+ bool result = false;
+
+ memset(l, 0, sizeof(struct dml2_initialize_instance_locals));
+ memset(dml, 0, sizeof(struct dml2_instance));
+
+ memcpy(&dml->ip_caps, &in_out->ip_caps, sizeof(struct dml2_ip_capabilities));
+ memcpy(&dml->soc_bbox, &in_out->soc_bb, sizeof(struct dml2_soc_bb));
+
+ dml->project_id = in_out->options.project_id;
+ dml->pmo_options = in_out->options.pmo_options;
+
+ // Initialize All Components
+ result = dml2_mcg_create(in_out->options.project_id, &dml->mcg_instance);
+
+ if (result)
+ result = dml2_dpmm_create(in_out->options.project_id, &dml->dpmm_instance);
+
+ if (result)
+ result = dml2_core_create(in_out->options.project_id, &dml->core_instance);
+
+ if (result) {
+ mcg_build_min_clk_params.soc_bb = &in_out->soc_bb;
+ mcg_build_min_clk_params.min_clk_table = &dml->min_clk_table;
+ result = dml->mcg_instance.build_min_clock_table(&mcg_build_min_clk_params);
+ }
+
+ if (result) {
+ core_init_params.project_id = in_out->options.project_id;
+ core_init_params.instance = &dml->core_instance;
+ core_init_params.minimum_clock_table = &dml->min_clk_table;
+ core_init_params.explicit_ip_bb = in_out->overrides.explicit_ip_bb;
+ core_init_params.explicit_ip_bb_size = in_out->overrides.explicit_ip_bb_size;
+ core_init_params.ip_caps = &in_out->ip_caps;
+ core_init_params.soc_bb = &in_out->soc_bb;
+ result = dml->core_instance.initialize(&core_init_params);
+
+ if (core_init_params.explicit_ip_bb && core_init_params.explicit_ip_bb_size > 0) {
+ memcpy(&dml->ip_caps, &in_out->ip_caps, sizeof(struct dml2_ip_capabilities));
+ }
+ }
+
+ if (result)
+ result = dml2_pmo_create(in_out->options.project_id, &dml->pmo_instance);
+
+ if (result) {
+ pmo_init_params.instance = &dml->pmo_instance;
+ pmo_init_params.soc_bb = &dml->soc_bbox;
+ pmo_init_params.ip_caps = &dml->ip_caps;
+ pmo_init_params.mcg_clock_table_size = dml->min_clk_table.dram_bw_table.num_entries;
+ pmo_init_params.options = &dml->pmo_options;
+ dml->pmo_instance.initialize(&pmo_init_params);
+ }
+ dml->funcs = soc15_funcs;
+ return result;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml_top_mcache.h b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_soc15.h
index 7b1f6f7143d0..53bd8602f9ef 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml_top_mcache.h
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_soc15.h
@@ -2,22 +2,13 @@
//
// Copyright 2024 Advanced Micro Devices, Inc.
-#ifndef __DML_TOP_MCACHE_H__
-#define __DML_TOP_MCACHE_H__
-
-#include "dml2_external_lib_deps.h"
-#include "dml_top_display_cfg_types.h"
-#include "dml_top_types.h"
+#ifndef __DML2_TOP_SOC15_H__
+#define __DML2_TOP_SOC15_H__
#include "dml2_internal_shared_types.h"
+bool dml2_top_soc15_initialize_instance(struct dml2_initialize_instance_in_out *in_out);
bool dml2_top_mcache_calc_mcache_count_and_offsets(struct top_mcache_calc_mcache_count_and_offsets_in_out *params);
-
void dml2_top_mcache_assign_global_mcache_ids(struct top_mcache_assign_global_mcache_ids_in_out *params);
-
bool dml2_top_mcache_validate_admissability(struct top_mcache_validate_admissability_in_out *params);
-
-bool dml2_top_mcache_build_mcache_programming(struct dml2_build_mcache_programming_in_out *params);
-
-bool dml2_top_mcache_unit_test(void);
-
-#endif
+bool dml2_top_soc15_build_mcache_programming(struct dml2_build_mcache_programming_in_out *params);
+#endif /* __DML2_TOP_SOC15_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml_top_mcache.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml_top_mcache.c
deleted file mode 100644
index a342ebfbe4e7..000000000000
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml_top_mcache.c
+++ /dev/null
@@ -1,549 +0,0 @@
-// SPDX-License-Identifier: MIT
-//
-// Copyright 2024 Advanced Micro Devices, Inc.
-
-#include "dml2_debug.h"
-
-#include "dml_top_mcache.h"
-#include "lib_float_math.h"
-
-#include "dml2_internal_shared_types.h"
-
-/*
-* Takes an input set of mcache boundaries and finds the appropriate setting of cache programming.
-* Returns true if a valid set of programming can be made, and false otherwise. "Valid" means
-* that the horizontal viewport does not span more than 2 cache slices.
-*
-* It optionally also can apply a constant shift to all the cache boundaries.
-*/
-static const uint32_t MCACHE_ID_UNASSIGNED = 0xF;
-static const uint32_t SPLIT_LOCATION_UNDEFINED = 0xFFFF;
-
-static bool calculate_first_second_splitting(const int *mcache_boundaries, int num_boundaries, int shift,
- int pipe_h_vp_start, int pipe_h_vp_end, int *first_offset, int *second_offset)
-{
- const int MAX_VP = 0xFFFFFF;
- int left_cache_id;
- int right_cache_id;
- int range_start;
- int range_end;
- bool success = false;
-
- if (num_boundaries <= 1) {
- if (first_offset && second_offset) {
- *first_offset = 0;
- *second_offset = -1;
- }
- success = true;
- return success;
- } else {
- range_start = 0;
- for (left_cache_id = 0; left_cache_id < num_boundaries; left_cache_id++) {
- range_end = mcache_boundaries[left_cache_id] - shift - 1;
-
- if (range_start <= pipe_h_vp_start && pipe_h_vp_start <= range_end)
- break;
-
- range_start = range_end + 1;
- }
-
- range_end = MAX_VP;
- for (right_cache_id = num_boundaries - 1; right_cache_id >= -1; right_cache_id--) {
- if (right_cache_id >= 0)
- range_start = mcache_boundaries[right_cache_id] - shift;
- else
- range_start = 0;
-
- if (range_start <= pipe_h_vp_end && pipe_h_vp_end <= range_end) {
- break;
- }
- range_end = range_start - 1;
- }
- right_cache_id = (right_cache_id + 1) % num_boundaries;
-
- if (right_cache_id == left_cache_id) {
- if (first_offset && second_offset) {
- *first_offset = left_cache_id;
- *second_offset = -1;
- }
- success = true;
- } else if (right_cache_id == (left_cache_id + 1) % num_boundaries) {
- if (first_offset && second_offset) {
- *first_offset = left_cache_id;
- *second_offset = right_cache_id;
- }
- success = true;
- }
- }
-
- return success;
-}
-
-/*
-* For a given set of pipe start/end x positions, checks to see it can support the input mcache splitting.
-* It also attempts to "optimize" by finding a shift if the default 0 shift does not work.
-*/
-static bool find_shift_for_valid_cache_id_assignment(int *mcache_boundaries, unsigned int num_boundaries,
- int *pipe_vp_startx, int *pipe_vp_endx, unsigned int pipe_count, int shift_granularity, int *shift)
-{
- int max_shift = 0xFFFF;
- unsigned int pipe_index;
- unsigned int i, slice_width;
- bool success = false;
-
- for (i = 0; i < num_boundaries; i++) {
- if (i == 0)
- slice_width = mcache_boundaries[i];
- else
- slice_width = mcache_boundaries[i] - mcache_boundaries[i - 1];
-
- if (max_shift > (int)slice_width) {
- max_shift = slice_width;
- }
- }
-
- for (*shift = 0; *shift <= max_shift; *shift += shift_granularity) {
- success = true;
- for (pipe_index = 0; pipe_index < pipe_count; pipe_index++) {
- if (!calculate_first_second_splitting(mcache_boundaries, num_boundaries, *shift,
- pipe_vp_startx[pipe_index], pipe_vp_endx[pipe_index], 0, 0)) {
- success = false;
- break;
- }
- }
- if (success)
- break;
- }
-
- return success;
-}
-
-/*
-* Counts the number of elements inside input array within the given span length.
-* Formally, what is the size of the largest subset of the array where the largest and smallest element
-* differ no more than the span.
-*/
-static unsigned int count_elements_in_span(int *array, unsigned int array_size, unsigned int span)
-{
- unsigned int i;
- unsigned int span_start_value;
- unsigned int span_start_index;
- unsigned int greatest_element_count;
-
- if (array_size == 0)
- return 1;
-
- if (span == 0)
- return array_size > 0 ? 1 : 0;
-
- span_start_value = 0;
- span_start_index = 0;
- greatest_element_count = 0;
-
- while (span_start_index < array_size) {
- for (i = span_start_index; i < array_size; i++) {
- if (array[i] - span_start_value <= span) {
- if (i - span_start_index + 1 > greatest_element_count) {
- greatest_element_count = i - span_start_index + 1;
- }
- } else
- break;
- }
-
- span_start_index++;
-
- if (span_start_index < array_size) {
- span_start_value = array[span_start_index - 1] + 1;
- }
- }
-
- return greatest_element_count;
-}
-
-static bool calculate_h_split_for_scaling_transform(int full_vp_width, int h_active, int num_pipes,
- enum dml2_scaling_transform scaling_transform, int *pipe_vp_x_start, int *pipe_vp_x_end)
-{
- int i, slice_width;
- const char MAX_SCL_VP_OVERLAP = 3;
- bool success = false;
-
- switch (scaling_transform) {
- case dml2_scaling_transform_centered:
- case dml2_scaling_transform_aspect_ratio:
- case dml2_scaling_transform_fullscreen:
- slice_width = full_vp_width / num_pipes;
- for (i = 0; i < num_pipes; i++) {
- pipe_vp_x_start[i] = i * slice_width;
- pipe_vp_x_end[i] = (i + 1) * slice_width - 1;
-
- if (pipe_vp_x_start[i] < MAX_SCL_VP_OVERLAP)
- pipe_vp_x_start[i] = 0;
- else
- pipe_vp_x_start[i] -= MAX_SCL_VP_OVERLAP;
-
- if (pipe_vp_x_end[i] > full_vp_width - MAX_SCL_VP_OVERLAP - 1)
- pipe_vp_x_end[i] = full_vp_width - 1;
- else
- pipe_vp_x_end[i] += MAX_SCL_VP_OVERLAP;
- }
- break;
- case dml2_scaling_transform_explicit:
- default:
- success = false;
- break;
- }
-
- return success;
-}
-
-bool dml2_top_mcache_validate_admissability(struct top_mcache_validate_admissability_in_out *params)
-{
- struct dml2_instance *dml = (struct dml2_instance *)params->dml2_instance;
- struct dml2_top_mcache_validate_admissability_locals *l = &dml->scratch.mcache_validate_admissability_locals;
-
- const int MAX_PIXEL_OVERLAP = 6;
- int max_per_pipe_vp_p0 = 0;
- int max_per_pipe_vp_p1 = 0;
- int temp, p0shift, p1shift;
- unsigned int plane_index = 0;
- unsigned int i;
- unsigned int odm_combine_factor;
- unsigned int mpc_combine_factor;
- unsigned int num_dpps;
- unsigned int num_boundaries;
- enum dml2_scaling_transform scaling_transform;
- const struct dml2_plane_parameters *plane;
- const struct dml2_stream_parameters *stream;
-
- bool p0pass = false;
- bool p1pass = false;
- bool all_pass = true;
-
- for (plane_index = 0; plane_index < params->display_cfg->num_planes; plane_index++) {
- if (!params->display_cfg->plane_descriptors[plane_index].surface.dcc.enable)
- continue;
-
- plane = &params->display_cfg->plane_descriptors[plane_index];
- stream = &params->display_cfg->stream_descriptors[plane->stream_index];
-
- num_dpps = odm_combine_factor = params->cfg_support_info->stream_support_info[plane->stream_index].odms_used;
-
- if (odm_combine_factor == 1)
- num_dpps = mpc_combine_factor = (unsigned int)params->cfg_support_info->plane_support_info[plane_index].dpps_used;
- else
- mpc_combine_factor = 1;
-
- if (odm_combine_factor > 1) {
- max_per_pipe_vp_p0 = plane->surface.plane0.width;
- temp = (unsigned int)math_ceil(plane->composition.scaler_info.plane0.h_ratio * stream->timing.h_active / odm_combine_factor);
-
- if (temp < max_per_pipe_vp_p0)
- max_per_pipe_vp_p0 = temp;
-
- max_per_pipe_vp_p1 = plane->surface.plane1.width;
- temp = (unsigned int)math_ceil(plane->composition.scaler_info.plane1.h_ratio * stream->timing.h_active / odm_combine_factor);
-
- if (temp < max_per_pipe_vp_p1)
- max_per_pipe_vp_p1 = temp;
- } else {
- max_per_pipe_vp_p0 = plane->surface.plane0.width / mpc_combine_factor;
- max_per_pipe_vp_p1 = plane->surface.plane1.width / mpc_combine_factor;
- }
-
- max_per_pipe_vp_p0 += 2 * MAX_PIXEL_OVERLAP;
- max_per_pipe_vp_p1 += MAX_PIXEL_OVERLAP;
-
- p0shift = 0;
- p1shift = 0;
-
- // The last element in the unshifted boundary array will always be the first pixel outside the
- // plane, which means theres no mcache associated with it, so -1
- num_boundaries = params->mcache_allocations[plane_index].num_mcaches_plane0 == 0 ? 0 : params->mcache_allocations[plane_index].num_mcaches_plane0 - 1;
- if ((count_elements_in_span(params->mcache_allocations[plane_index].mcache_x_offsets_plane0,
- num_boundaries, max_per_pipe_vp_p0) <= 1) && (num_boundaries <= num_dpps)) {
- p0pass = true;
- }
- num_boundaries = params->mcache_allocations[plane_index].num_mcaches_plane1 == 0 ? 0 : params->mcache_allocations[plane_index].num_mcaches_plane1 - 1;
- if ((count_elements_in_span(params->mcache_allocations[plane_index].mcache_x_offsets_plane1,
- num_boundaries, max_per_pipe_vp_p1) <= 1) && (num_boundaries <= num_dpps)) {
- p1pass = true;
- }
-
- if (!p0pass || !p1pass) {
- if (odm_combine_factor > 1) {
- num_dpps = odm_combine_factor;
- scaling_transform = plane->composition.scaling_transform;
- } else {
- num_dpps = mpc_combine_factor;
- scaling_transform = dml2_scaling_transform_fullscreen;
- }
-
- if (!p0pass) {
- if (plane->composition.viewport.stationary) {
- calculate_h_split_for_scaling_transform(plane->surface.plane0.width,
- stream->timing.h_active, num_dpps, scaling_transform,
- &l->plane0.pipe_vp_startx[plane_index], &l->plane0.pipe_vp_endx[plane_index]);
- p0pass = find_shift_for_valid_cache_id_assignment(params->mcache_allocations[plane_index].mcache_x_offsets_plane0,
- params->mcache_allocations[plane_index].num_mcaches_plane0,
- &l->plane0.pipe_vp_startx[plane_index], &l->plane0.pipe_vp_endx[plane_index], num_dpps,
- params->mcache_allocations[plane_index].shift_granularity.p0, &p0shift);
- }
- }
- if (!p1pass) {
- if (plane->composition.viewport.stationary) {
- calculate_h_split_for_scaling_transform(plane->surface.plane1.width,
- stream->timing.h_active, num_dpps, scaling_transform,
- &l->plane0.pipe_vp_startx[plane_index], &l->plane0.pipe_vp_endx[plane_index]);
- p1pass = find_shift_for_valid_cache_id_assignment(params->mcache_allocations[plane_index].mcache_x_offsets_plane1,
- params->mcache_allocations[plane_index].num_mcaches_plane1,
- &l->plane1.pipe_vp_startx[plane_index], &l->plane1.pipe_vp_endx[plane_index], num_dpps,
- params->mcache_allocations[plane_index].shift_granularity.p1, &p1shift);
- }
- }
- }
-
- if (p0pass && p1pass) {
- for (i = 0; i < params->mcache_allocations[plane_index].num_mcaches_plane0; i++) {
- params->mcache_allocations[plane_index].mcache_x_offsets_plane0[i] -= p0shift;
- }
- for (i = 0; i < params->mcache_allocations[plane_index].num_mcaches_plane1; i++) {
- params->mcache_allocations[plane_index].mcache_x_offsets_plane1[i] -= p1shift;
- }
- }
-
- params->per_plane_status[plane_index] = p0pass && p1pass;
- all_pass &= p0pass && p1pass;
- }
-
- return all_pass;
-}
-
-static void reset_mcache_allocations(struct dml2_hubp_pipe_mcache_regs *per_plane_pipe_mcache_regs)
-{
- // Initialize all entries to special valid MCache ID and special valid split coordinate
- per_plane_pipe_mcache_regs->main.p0.mcache_id_first = MCACHE_ID_UNASSIGNED;
- per_plane_pipe_mcache_regs->main.p0.mcache_id_second = MCACHE_ID_UNASSIGNED;
- per_plane_pipe_mcache_regs->main.p0.split_location = SPLIT_LOCATION_UNDEFINED;
-
- per_plane_pipe_mcache_regs->mall.p0.mcache_id_first = MCACHE_ID_UNASSIGNED;
- per_plane_pipe_mcache_regs->mall.p0.mcache_id_second = MCACHE_ID_UNASSIGNED;
- per_plane_pipe_mcache_regs->mall.p0.split_location = SPLIT_LOCATION_UNDEFINED;
-
- per_plane_pipe_mcache_regs->main.p1.mcache_id_first = MCACHE_ID_UNASSIGNED;
- per_plane_pipe_mcache_regs->main.p1.mcache_id_second = MCACHE_ID_UNASSIGNED;
- per_plane_pipe_mcache_regs->main.p1.split_location = SPLIT_LOCATION_UNDEFINED;
-
- per_plane_pipe_mcache_regs->mall.p1.mcache_id_first = MCACHE_ID_UNASSIGNED;
- per_plane_pipe_mcache_regs->mall.p1.mcache_id_second = MCACHE_ID_UNASSIGNED;
- per_plane_pipe_mcache_regs->mall.p1.split_location = SPLIT_LOCATION_UNDEFINED;
-}
-
-bool dml2_top_mcache_build_mcache_programming(struct dml2_build_mcache_programming_in_out *params)
-{
- bool success = true;
- int config_index, pipe_index;
- int first_offset, second_offset;
- int free_per_plane_reg_index = 0;
-
- memset(params->per_plane_pipe_mcache_regs, 0, DML2_MAX_PLANES * DML2_MAX_DCN_PIPES * sizeof(struct dml2_hubp_pipe_mcache_regs *));
-
- for (config_index = 0; config_index < params->num_configurations; config_index++) {
- for (pipe_index = 0; pipe_index < params->mcache_configurations[config_index].num_pipes; pipe_index++) {
- // Allocate storage for the mcache regs
- params->per_plane_pipe_mcache_regs[config_index][pipe_index] = &params->mcache_regs_set[free_per_plane_reg_index++];
-
- reset_mcache_allocations(params->per_plane_pipe_mcache_regs[config_index][pipe_index]);
-
- if (params->mcache_configurations[config_index].plane_descriptor->surface.dcc.enable) {
- // P0 always enabled
- if (!calculate_first_second_splitting(params->mcache_configurations[config_index].mcache_allocation->mcache_x_offsets_plane0,
- params->mcache_configurations[config_index].mcache_allocation->num_mcaches_plane0,
- 0,
- params->mcache_configurations[config_index].pipe_configurations[pipe_index].plane0.viewport_x_start,
- params->mcache_configurations[config_index].pipe_configurations[pipe_index].plane0.viewport_x_start +
- params->mcache_configurations[config_index].pipe_configurations[pipe_index].plane0.viewport_width - 1,
- &first_offset, &second_offset)) {
- success = false;
- break;
- }
-
- params->per_plane_pipe_mcache_regs[config_index][pipe_index]->main.p0.mcache_id_first =
- params->mcache_configurations[config_index].mcache_allocation->global_mcache_ids_plane0[first_offset];
-
- params->per_plane_pipe_mcache_regs[config_index][pipe_index]->mall.p0.mcache_id_first =
- params->mcache_configurations[config_index].mcache_allocation->global_mcache_ids_mall_plane0[first_offset];
-
- if (second_offset >= 0) {
- params->per_plane_pipe_mcache_regs[config_index][pipe_index]->main.p0.mcache_id_second =
- params->mcache_configurations[config_index].mcache_allocation->global_mcache_ids_plane0[second_offset];
- params->per_plane_pipe_mcache_regs[config_index][pipe_index]->main.p0.split_location =
- params->mcache_configurations[config_index].mcache_allocation->mcache_x_offsets_plane0[first_offset] - 1;
-
- params->per_plane_pipe_mcache_regs[config_index][pipe_index]->mall.p0.mcache_id_second =
- params->mcache_configurations[config_index].mcache_allocation->global_mcache_ids_mall_plane0[second_offset];
- params->per_plane_pipe_mcache_regs[config_index][pipe_index]->mall.p0.split_location =
- params->mcache_configurations[config_index].mcache_allocation->mcache_x_offsets_plane0[first_offset] - 1;
- }
-
- // Populate P1 if enabled
- if (params->mcache_configurations[config_index].pipe_configurations[pipe_index].plane1_enabled) {
- if (!calculate_first_second_splitting(params->mcache_configurations[config_index].mcache_allocation->mcache_x_offsets_plane1,
- params->mcache_configurations[config_index].mcache_allocation->num_mcaches_plane1,
- 0,
- params->mcache_configurations[config_index].pipe_configurations[pipe_index].plane1.viewport_x_start,
- params->mcache_configurations[config_index].pipe_configurations[pipe_index].plane1.viewport_x_start +
- params->mcache_configurations[config_index].pipe_configurations[pipe_index].plane1.viewport_width - 1,
- &first_offset, &second_offset)) {
- success = false;
- break;
- }
-
- params->per_plane_pipe_mcache_regs[config_index][pipe_index]->main.p1.mcache_id_first =
- params->mcache_configurations[config_index].mcache_allocation->global_mcache_ids_plane1[first_offset];
-
- params->per_plane_pipe_mcache_regs[config_index][pipe_index]->mall.p1.mcache_id_first =
- params->mcache_configurations[config_index].mcache_allocation->global_mcache_ids_mall_plane1[first_offset];
-
- if (second_offset >= 0) {
- params->per_plane_pipe_mcache_regs[config_index][pipe_index]->main.p1.mcache_id_second =
- params->mcache_configurations[config_index].mcache_allocation->global_mcache_ids_plane1[second_offset];
- params->per_plane_pipe_mcache_regs[config_index][pipe_index]->main.p1.split_location =
- params->mcache_configurations[config_index].mcache_allocation->mcache_x_offsets_plane1[first_offset] - 1;
-
- params->per_plane_pipe_mcache_regs[config_index][pipe_index]->mall.p1.mcache_id_second =
- params->mcache_configurations[config_index].mcache_allocation->global_mcache_ids_mall_plane1[second_offset];
- params->per_plane_pipe_mcache_regs[config_index][pipe_index]->mall.p1.split_location =
- params->mcache_configurations[config_index].mcache_allocation->mcache_x_offsets_plane1[first_offset] - 1;
- }
- }
- }
- }
- }
-
- return success;
-}
-
-void dml2_top_mcache_assign_global_mcache_ids(struct top_mcache_assign_global_mcache_ids_in_out *params)
-{
- int i;
- unsigned int j;
- int next_unused_cache_id = 0;
-
- for (i = 0; i < params->num_allocations; i++) {
- if (!params->allocations[i].valid)
- continue;
-
- for (j = 0; j < params->allocations[i].num_mcaches_plane0; j++) {
- params->allocations[i].global_mcache_ids_plane0[j] = next_unused_cache_id++;
- }
- for (j = 0; j < params->allocations[i].num_mcaches_plane1; j++) {
- params->allocations[i].global_mcache_ids_plane1[j] = next_unused_cache_id++;
- }
-
- // The "psuedo-last" slice is always wrapped around
- params->allocations[i].global_mcache_ids_plane0[params->allocations[i].num_mcaches_plane0] =
- params->allocations[i].global_mcache_ids_plane0[0];
- params->allocations[i].global_mcache_ids_plane1[params->allocations[i].num_mcaches_plane1] =
- params->allocations[i].global_mcache_ids_plane1[0];
-
- // If we need dedicated caches for mall requesting, then we assign them here.
- if (params->allocations[i].requires_dedicated_mall_mcache) {
- for (j = 0; j < params->allocations[i].num_mcaches_plane0; j++) {
- params->allocations[i].global_mcache_ids_mall_plane0[j] = next_unused_cache_id++;
- }
- for (j = 0; j < params->allocations[i].num_mcaches_plane1; j++) {
- params->allocations[i].global_mcache_ids_mall_plane1[j] = next_unused_cache_id++;
- }
-
- // The "psuedo-last" slice is always wrapped around
- params->allocations[i].global_mcache_ids_mall_plane0[params->allocations[i].num_mcaches_plane0] =
- params->allocations[i].global_mcache_ids_mall_plane0[0];
- params->allocations[i].global_mcache_ids_mall_plane1[params->allocations[i].num_mcaches_plane1] =
- params->allocations[i].global_mcache_ids_mall_plane1[0];
- }
-
- // If P0 and P1 are sharing caches, then it means the largest mcache IDs for p0 and p1 can be the same
- // since mcache IDs are always ascending, then it means the largest mcacheID of p1 should be the
- // largest mcacheID of P0
- if (params->allocations[i].num_mcaches_plane0 > 0 && params->allocations[i].num_mcaches_plane1 > 0 &&
- params->allocations[i].last_slice_sharing.plane0_plane1) {
- params->allocations[i].global_mcache_ids_plane1[params->allocations[i].num_mcaches_plane1 - 1] =
- params->allocations[i].global_mcache_ids_plane0[params->allocations[i].num_mcaches_plane0 - 1];
- }
-
- // If we need dedicated caches handle last slice sharing
- if (params->allocations[i].requires_dedicated_mall_mcache) {
- if (params->allocations[i].num_mcaches_plane0 > 0 && params->allocations[i].num_mcaches_plane1 > 0 &&
- params->allocations[i].last_slice_sharing.plane0_plane1) {
- params->allocations[i].global_mcache_ids_mall_plane1[params->allocations[i].num_mcaches_plane1 - 1] =
- params->allocations[i].global_mcache_ids_mall_plane0[params->allocations[i].num_mcaches_plane0 - 1];
- }
- // If mall_comb_mcache_l is set then it means that largest mcache ID for MALL p0 can be same as regular read p0
- if (params->allocations[i].num_mcaches_plane0 > 0 && params->allocations[i].last_slice_sharing.mall_comb_mcache_p0) {
- params->allocations[i].global_mcache_ids_mall_plane0[params->allocations[i].num_mcaches_plane0 - 1] =
- params->allocations[i].global_mcache_ids_plane0[params->allocations[i].num_mcaches_plane0 - 1];
- }
- // If mall_comb_mcache_c is set then it means that largest mcache ID for MALL p1 can be same as regular
- // read p1 (which can be same as regular read p0 if plane0_plane1 is also set)
- if (params->allocations[i].num_mcaches_plane1 > 0 && params->allocations[i].last_slice_sharing.mall_comb_mcache_p1) {
- params->allocations[i].global_mcache_ids_mall_plane1[params->allocations[i].num_mcaches_plane1 - 1] =
- params->allocations[i].global_mcache_ids_plane1[params->allocations[i].num_mcaches_plane1 - 1];
- }
- }
-
- // If you don't need dedicated mall mcaches, the mall mcache assignments are identical to the normal requesting
- if (!params->allocations[i].requires_dedicated_mall_mcache) {
- memcpy(params->allocations[i].global_mcache_ids_mall_plane0, params->allocations[i].global_mcache_ids_plane0,
- sizeof(params->allocations[i].global_mcache_ids_mall_plane0));
- memcpy(params->allocations[i].global_mcache_ids_mall_plane1, params->allocations[i].global_mcache_ids_plane1,
- sizeof(params->allocations[i].global_mcache_ids_mall_plane1));
- }
- }
-}
-
-bool dml2_top_mcache_calc_mcache_count_and_offsets(struct top_mcache_calc_mcache_count_and_offsets_in_out *params)
-{
- struct dml2_instance *dml = (struct dml2_instance *)params->dml2_instance;
- struct dml2_top_mcache_verify_mcache_size_locals *l = &dml->scratch.mcache_verify_mcache_size_locals;
-
- unsigned int total_mcaches_required;
- unsigned int i;
- bool result = false;
-
- if (dml->soc_bbox.num_dcc_mcaches == 0) {
- return true;
- }
-
- total_mcaches_required = 0;
- l->calc_mcache_params.instance = &dml->core_instance;
- for (i = 0; i < params->display_config->num_planes; i++) {
- if (!params->display_config->plane_descriptors[i].surface.dcc.enable) {
- memset(&params->mcache_allocations[i], 0, sizeof(struct dml2_mcache_surface_allocation));
- continue;
- }
-
- l->calc_mcache_params.plane_descriptor = &params->display_config->plane_descriptors[i];
- l->calc_mcache_params.mcache_allocation = &params->mcache_allocations[i];
- l->calc_mcache_params.plane_index = i;
-
- if (!dml->core_instance.calculate_mcache_allocation(&l->calc_mcache_params)) {
- result = false;
- break;
- }
-
- if (params->mcache_allocations[i].valid) {
- total_mcaches_required += params->mcache_allocations[i].num_mcaches_plane0 + params->mcache_allocations[i].num_mcaches_plane1;
- if (params->mcache_allocations[i].last_slice_sharing.plane0_plane1)
- total_mcaches_required--;
- }
- }
- dml2_printf("DML_CORE_DCN3::%s: plane_%d, total_mcaches_required=%d\n", __func__, i, total_mcaches_required);
-
- if (total_mcaches_required > dml->soc_bbox.num_dcc_mcaches) {
- result = false;
- } else {
- result = true;
- }
-
- return result;
-}
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/inc/dml2_debug.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/inc/dml2_debug.c
index e9b8e10695ae..f95c7ff56f15 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/inc/dml2_debug.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/inc/dml2_debug.c
@@ -4,6 +4,11 @@
#include "dml2_debug.h"
+int dml2_log_internal(const char *format, ...)
+{
+ return 0;
+}
+
int dml2_printf(const char *format, ...)
{
#ifdef _DEBUG
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/inc/dml2_debug.h b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/inc/dml2_debug.h
index d51a1b6c62f2..a27792b56f7e 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/inc/dml2_debug.h
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/inc/dml2_debug.h
@@ -8,9 +8,53 @@
#ifdef _DEBUG
#define DML2_ASSERT(condition) dml2_assert(condition)
#else
-#define DML2_ASSERT(condition)
+#define DML2_ASSERT(condition) ((void)0)
+#endif
+/*
+ * DML_LOG_FATAL - fatal errors for unrecoverable DML states until a restart.
+ * DML_LOG_ERROR - unexpected but recoverable failures inside DML
+ * DML_LOG_WARN - unexpected inputs or events to DML
+ * DML_LOG_INFO - high level tracing of DML interfaces
+ * DML_LOG_DEBUG - detailed tracing of DML internal components
+ * DML_LOG_VERBOSE - detailed tracing of DML calculation procedure
+ */
+#if !defined(DML_LOG_LEVEL)
+#if defined(_DEBUG) && defined(_DEBUG_PRINTS)
+/* for backward compatibility with old macros */
+#define DML_LOG_LEVEL 5
+#else
+#define DML_LOG_LEVEL 0
+#endif
+#endif
+
+#define DML_LOG_FATAL(fmt, ...) dml2_log_internal(fmt, ## __VA_ARGS__)
+#if DML_LOG_LEVEL >= 1
+#define DML_LOG_ERROR(fmt, ...) dml2_log_internal(fmt, ## __VA_ARGS__)
+#else
+#define DML_LOG_ERROR(fmt, ...) ((void)0)
+#endif
+#if DML_LOG_LEVEL >= 2
+#define DML_LOG_WARN(fmt, ...) dml2_log_internal(fmt, ## __VA_ARGS__)
+#else
+#define DML_LOG_WARN(fmt, ...) ((void)0)
+#endif
+#if DML_LOG_LEVEL >= 3
+#define DML_LOG_INFO(fmt, ...) dml2_log_internal(fmt, ## __VA_ARGS__)
+#else
+#define DML_LOG_INFO(fmt, ...) ((void)0)
+#endif
+#if DML_LOG_LEVEL >= 4
+#define DML_LOG_DEBUG(fmt, ...) dml2_log_internal(fmt, ## __VA_ARGS__)
+#else
+#define DML_LOG_DEBUG(fmt, ...) ((void)0)
+#endif
+#if DML_LOG_LEVEL >= 5
+#define DML_LOG_VERBOSE(fmt, ...) dml2_log_internal(fmt, ## __VA_ARGS__)
+#else
+#define DML_LOG_VERBOSE(fmt, ...) ((void)0)
#endif
+int dml2_log_internal(const char *format, ...);
int dml2_printf(const char *format, ...);
void dml2_assert(int condition);
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/inc/dml2_internal_shared_types.h b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/inc/dml2_internal_shared_types.h
index aeac9f159fa5..7fb6026bcb49 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/inc/dml2_internal_shared_types.h
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/inc/dml2_internal_shared_types.h
@@ -8,7 +8,6 @@
#include "dml2_external_lib_deps.h"
#include "dml_top_types.h"
#include "dml2_core_shared_types.h"
-
/*
* DML2 MCG Types and Interfaces
*/
@@ -63,7 +62,6 @@ struct dml2_mcg_build_min_clock_table_params_in_out {
*/
struct dml2_mcg_min_clock_table *min_clk_table;
};
-
struct dml2_mcg_instance {
bool (*build_min_clock_table)(struct dml2_mcg_build_min_clock_table_params_in_out *in_out);
bool (*unit_test)(void);
@@ -81,7 +79,6 @@ struct dml2_dpmm_map_mode_to_soc_dpm_params_in_out {
struct dml2_soc_bb *soc_bb;
struct dml2_mcg_min_clock_table *min_clk_table;
const struct display_configuation_with_meta *display_cfg;
-
struct {
bool perform_pseudo_map;
struct dml2_core_internal_soc_bb *soc_bb;
@@ -309,7 +306,7 @@ struct dml2_optimization_stage3_state {
// The pstate support mode for each plane
// The number of valid elements == display_cfg.num_planes
// The indexing of pstate_switch_modes matches plane_descriptors[]
- enum dml2_uclk_pstate_support_method pstate_switch_modes[DML2_MAX_PLANES];
+ enum dml2_pstate_method pstate_switch_modes[DML2_MAX_PLANES];
// Meta-data for implicit SVP generation, indexed by stream index
struct dml2_implicit_svp_meta stream_svp_meta[DML2_MAX_PLANES];
@@ -356,6 +353,10 @@ struct display_configuation_with_meta {
struct dml2_optimization_stage5_state stage5;
};
+struct dml2_pmo_pstate_strategy {
+ enum dml2_pstate_method per_stream_pstate_method[DML2_MAX_PLANES];
+ bool allow_state_increase;
+};
struct dml2_core_mode_support_in_out {
/*
* Inputs
@@ -365,7 +366,6 @@ struct dml2_core_mode_support_in_out {
struct dml2_mcg_min_clock_table *min_clk_table;
int min_clk_index;
-
/*
* Outputs
*/
@@ -395,7 +395,6 @@ struct dml2_core_mode_programming_in_out {
struct dml2_core_instance *instance;
const struct display_configuation_with_meta *display_cfg;
const struct core_display_cfg_support_info *cfg_support_info;
-
/*
* Outputs (also Input the clk freq are also from programming struct)
*/
@@ -445,6 +444,7 @@ struct dml2_core_internal_state_intermediates {
struct dml2_core_mode_support_locals {
struct dml2_core_calcs_mode_support_ex mode_support_ex_params;
struct dml2_display_cfg svp_expanded_display_cfg;
+ struct dml2_calculate_mcache_allocation_in_out calc_mcache_allocation_params;
};
struct dml2_core_mode_programming_locals {
@@ -600,34 +600,11 @@ struct dml2_pmo_optimize_for_stutter_in_out {
struct display_configuation_with_meta *optimized_display_config;
};
-enum dml2_pmo_pstate_method {
- dml2_pmo_pstate_strategy_na = 0,
- /* hw exclusive modes */
- dml2_pmo_pstate_strategy_vactive = 1,
- dml2_pmo_pstate_strategy_vblank = 2,
- dml2_pmo_pstate_strategy_reserved_hw = 5,
- /* fw assisted exclusive modes */
- dml2_pmo_pstate_strategy_fw_svp = 6,
- dml2_pmo_pstate_strategy_reserved_fw = 10,
- /* fw assisted modes requiring drr modulation */
- dml2_pmo_pstate_strategy_fw_vactive_drr = 11,
- dml2_pmo_pstate_strategy_fw_vblank_drr = 12,
- dml2_pmo_pstate_strategy_fw_svp_drr = 13,
- dml2_pmo_pstate_strategy_reserved_fw_drr_clamped = 20,
- dml2_pmo_pstate_strategy_fw_drr = 21,
- dml2_pmo_pstate_strategy_reserved_fw_drr_var = 22,
-};
-
-struct dml2_pmo_pstate_strategy {
- enum dml2_pmo_pstate_method per_stream_pstate_method[DML2_MAX_PLANES];
- bool allow_state_increase;
-};
-
-#define PMO_NO_DRR_STRATEGY_MASK (((1 << (dml2_pmo_pstate_strategy_reserved_fw - dml2_pmo_pstate_strategy_na + 1)) - 1) << dml2_pmo_pstate_strategy_na)
-#define PMO_DRR_STRATEGY_MASK (((1 << (dml2_pmo_pstate_strategy_reserved_fw_drr_var - dml2_pmo_pstate_strategy_fw_vactive_drr + 1)) - 1) << dml2_pmo_pstate_strategy_fw_vactive_drr)
-#define PMO_DRR_CLAMPED_STRATEGY_MASK (((1 << (dml2_pmo_pstate_strategy_reserved_fw_drr_clamped - dml2_pmo_pstate_strategy_fw_vactive_drr + 1)) - 1) << dml2_pmo_pstate_strategy_fw_vactive_drr)
-#define PMO_DRR_VAR_STRATEGY_MASK (((1 << (dml2_pmo_pstate_strategy_reserved_fw_drr_var - dml2_pmo_pstate_strategy_fw_drr + 1)) - 1) << dml2_pmo_pstate_strategy_fw_drr)
-#define PMO_FW_STRATEGY_MASK (((1 << (dml2_pmo_pstate_strategy_reserved_fw_drr_var - dml2_pmo_pstate_strategy_fw_svp + 1)) - 1) << dml2_pmo_pstate_strategy_fw_svp)
+#define PMO_NO_DRR_STRATEGY_MASK (((1 << (dml2_pstate_method_reserved_fw - dml2_pstate_method_na + 1)) - 1) << dml2_pstate_method_na)
+#define PMO_DRR_STRATEGY_MASK (((1 << (dml2_pstate_method_reserved_fw_drr_var - dml2_pstate_method_fw_vactive_drr + 1)) - 1) << dml2_pstate_method_fw_vactive_drr)
+#define PMO_DRR_CLAMPED_STRATEGY_MASK (((1 << (dml2_pstate_method_reserved_fw_drr_clamped - dml2_pstate_method_fw_vactive_drr + 1)) - 1) << dml2_pstate_method_fw_vactive_drr)
+#define PMO_DRR_VAR_STRATEGY_MASK (((1 << (dml2_pstate_method_reserved_fw_drr_var - dml2_pstate_method_fw_drr + 1)) - 1) << dml2_pstate_method_fw_drr)
+#define PMO_FW_STRATEGY_MASK (((1 << (dml2_pstate_method_reserved_fw_drr_var - dml2_pstate_method_fw_svp + 1)) - 1) << dml2_pstate_method_fw_svp)
#define PMO_DCN4_MAX_DISPLAYS 4
#define PMO_DCN4_MAX_NUM_VARIANTS 2
@@ -645,6 +622,8 @@ struct dml2_pmo_scratch {
int stream_mask;
} pmo_dcn3;
struct {
+ struct dml2_pmo_pstate_strategy expanded_override_strategy_list[2 * 2 * 2 * 2];
+ unsigned int num_expanded_override_strategies;
struct dml2_pmo_pstate_strategy pstate_strategy_candidates[DML2_PMO_PSTATE_CANDIDATE_LIST_SIZE];
int num_pstate_candidates;
int cur_pstate_candidate;
@@ -706,7 +685,6 @@ struct dml2_pmo_instance {
int mpc_combine_limit;
int odm_combine_limit;
int mcg_clock_table_size;
-
union {
struct {
struct {
@@ -963,7 +941,13 @@ struct dml2_top_mcache_validate_admissability_locals {
struct dml2_top_display_cfg_support_info {
const struct dml2_display_cfg *display_config;
struct core_display_cfg_support_info core_info;
- enum dml2_pstate_support_method per_plane_pstate_method[DML2_MAX_PLANES];
+};
+
+struct dml2_top_funcs {
+ bool (*check_mode_supported)(struct dml2_check_mode_supported_in_out *in_out);
+ bool (*build_mode_programming)(struct dml2_build_mode_programming_in_out *in_out);
+ bool (*build_mcache_programming)(struct dml2_build_mcache_programming_in_out *in_out);
+ bool (*unit_test)(void);
};
struct dml2_instance {
@@ -978,8 +962,8 @@ struct dml2_instance {
struct dml2_ip_capabilities ip_caps;
struct dml2_mcg_min_clock_table min_clk_table;
-
struct dml2_pmo_options pmo_options;
+ struct dml2_top_funcs funcs;
struct {
struct dml2_initialize_instance_locals initialize_instance_locals;
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_mall_phantom.c b/drivers/gpu/drm/amd/display/dc/dml2/dml2_mall_phantom.c
index 3d29169dd6bb..6b3b8803e0ae 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml2_mall_phantom.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml2_mall_phantom.c
@@ -813,7 +813,7 @@ static bool remove_all_phantom_planes_for_stream(struct dml2_context *ctx, struc
{
int i, old_plane_count;
struct dc_stream_status *stream_status = NULL;
- struct dc_plane_state *del_planes[MAX_SURFACE_NUM] = { 0 };
+ struct dc_plane_state *del_planes[MAX_SURFACES] = { 0 };
for (i = 0; i < context->stream_count; i++)
if (context->streams[i] == stream) {
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c b/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c
index bde4250853b1..b416320873e1 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c
@@ -553,13 +553,53 @@ void dml2_init_soc_states(struct dml2_context *dml2, const struct dc *in_dc,
}
}
- dml2_policy_build_synthetic_soc_states(s, p);
- if (dml2->v20.dml_core_ctx.project == dml_project_dcn35) {
- // Override last out_state with data from last in_state
- // This will ensure that out_state contains max fclk
- memcpy(&p->out_states->state_array[p->out_states->num_states - 1],
- &p->in_states->state_array[p->in_states->num_states - 1],
- sizeof(struct soc_state_bounding_box_st));
+ if (dml2->v20.dml_core_ctx.project == dml_project_dcn35 ||
+ dml2->v20.dml_core_ctx.project == dml_project_dcn351) {
+ int max_dcfclk_mhz = 0, max_dispclk_mhz = 0, max_dppclk_mhz = 0, max_phyclk_mhz = 0,
+ max_dtbclk_mhz = 0, max_fclk_mhz = 0, max_uclk_mhz = 0, max_socclk_mhz = 0;
+
+ for (i = 0; i < p->in_states->num_states; i++) {
+ if (p->in_states->state_array[i].dcfclk_mhz > max_dcfclk_mhz)
+ max_dcfclk_mhz = (int)p->in_states->state_array[i].dcfclk_mhz;
+ if (p->in_states->state_array[i].fabricclk_mhz > max_fclk_mhz)
+ max_fclk_mhz = (int)p->in_states->state_array[i].fabricclk_mhz;
+ if (p->in_states->state_array[i].socclk_mhz > max_socclk_mhz)
+ max_socclk_mhz = (int)p->in_states->state_array[i].socclk_mhz;
+ if (p->in_states->state_array[i].dram_speed_mts > max_uclk_mhz)
+ max_uclk_mhz = (int)p->in_states->state_array[i].dram_speed_mts;
+ if (p->in_states->state_array[i].dispclk_mhz > max_dispclk_mhz)
+ max_dispclk_mhz = (int)p->in_states->state_array[i].dispclk_mhz;
+ if (p->in_states->state_array[i].dppclk_mhz > max_dppclk_mhz)
+ max_dppclk_mhz = (int)p->in_states->state_array[i].dppclk_mhz;
+ if (p->in_states->state_array[i].phyclk_mhz > max_phyclk_mhz)
+ max_phyclk_mhz = (int)p->in_states->state_array[i].phyclk_mhz;
+ if (p->in_states->state_array[i].dtbclk_mhz > max_dtbclk_mhz)
+ max_dtbclk_mhz = (int)p->in_states->state_array[i].dtbclk_mhz;
+ }
+
+ for (i = 0; i < p->in_states->num_states; i++) {
+ /* Independent states - including base (unlisted) parameters from state 0. */
+ p->out_states->state_array[i] = p->in_states->state_array[0];
+
+ p->out_states->state_array[i].dispclk_mhz = max_dispclk_mhz;
+ p->out_states->state_array[i].dppclk_mhz = max_dppclk_mhz;
+ p->out_states->state_array[i].dtbclk_mhz = max_dtbclk_mhz;
+ p->out_states->state_array[i].phyclk_mhz = max_phyclk_mhz;
+
+ p->out_states->state_array[i].dscclk_mhz = max_dispclk_mhz / 3.0;
+ p->out_states->state_array[i].phyclk_mhz = max_phyclk_mhz;
+ p->out_states->state_array[i].dtbclk_mhz = max_dtbclk_mhz;
+
+ /* Dependent states. */
+ p->out_states->state_array[i].dram_speed_mts = p->in_states->state_array[i].dram_speed_mts;
+ p->out_states->state_array[i].fabricclk_mhz = p->in_states->state_array[i].fabricclk_mhz;
+ p->out_states->state_array[i].socclk_mhz = p->in_states->state_array[i].socclk_mhz;
+ p->out_states->state_array[i].dcfclk_mhz = p->in_states->state_array[i].dcfclk_mhz;
+ }
+
+ p->out_states->num_states = p->in_states->num_states;
+ } else {
+ dml2_policy_build_synthetic_soc_states(s, p);
}
}
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.c b/drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.c
index 9190c1328d5b..68b882d28195 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.c
@@ -531,14 +531,21 @@ static bool optimize_pstate_with_svp_and_drr(struct dml2_context *dml2, struct d
static bool call_dml_mode_support_and_programming(struct dc_state *context)
{
unsigned int result = 0;
- unsigned int min_state;
+ unsigned int min_state = 0;
int min_state_for_g6_temp_read = 0;
+
+
+ if (!context)
+ return false;
+
struct dml2_context *dml2 = context->bw_ctx.dml2;
struct dml2_wrapper_scratch *s = &dml2->v20.scratch;
- min_state_for_g6_temp_read = calculate_lowest_supported_state_for_temp_read(dml2, context);
+ if (!context->streams[0]->sink->link->dc->caps.is_apu) {
+ min_state_for_g6_temp_read = calculate_lowest_supported_state_for_temp_read(dml2, context);
- ASSERT(min_state_for_g6_temp_read >= 0);
+ ASSERT(min_state_for_g6_temp_read >= 0);
+ }
if (!dml2->config.use_native_pstate_optimization) {
result = optimize_pstate_with_svp_and_drr(dml2, context);
@@ -549,14 +556,20 @@ static bool call_dml_mode_support_and_programming(struct dc_state *context)
/* Upon trying to sett certain frequencies in FRL, min_state_for_g6_temp_read is reported as -1. This leads to an invalid value of min_state causing crashes later on.
* Use the default logic for min_state only when min_state_for_g6_temp_read is a valid value. In other cases, use the value calculated by the DML directly.
*/
- if (min_state_for_g6_temp_read >= 0)
- min_state = min_state_for_g6_temp_read > s->mode_support_params.out_lowest_state_idx ? min_state_for_g6_temp_read : s->mode_support_params.out_lowest_state_idx;
- else
- min_state = s->mode_support_params.out_lowest_state_idx;
-
- if (result)
- result = dml_mode_programming(&dml2->v20.dml_core_ctx, min_state, &s->cur_display_config, true);
+ if (!context->streams[0]->sink->link->dc->caps.is_apu) {
+ if (min_state_for_g6_temp_read >= 0)
+ min_state = min_state_for_g6_temp_read > s->mode_support_params.out_lowest_state_idx ? min_state_for_g6_temp_read : s->mode_support_params.out_lowest_state_idx;
+ else
+ min_state = s->mode_support_params.out_lowest_state_idx;
+ }
+ if (result) {
+ if (!context->streams[0]->sink->link->dc->caps.is_apu) {
+ result = dml_mode_programming(&dml2->v20.dml_core_ctx, min_state, &s->cur_display_config, true);
+ } else {
+ result = dml_mode_programming(&dml2->v20.dml_core_ctx, s->mode_support_params.out_lowest_state_idx, &s->cur_display_config, true);
+ }
+ }
return result;
}
@@ -685,6 +698,8 @@ static bool dml2_validate_only(struct dc_state *context)
build_unoptimized_policy_settings(dml2->v20.dml_core_ctx.project, &dml2->v20.dml_core_ctx.policy);
map_dc_state_into_dml_display_cfg(dml2, context, &dml2->v20.scratch.cur_display_config);
+ if (!dml2->config.skip_hw_state_mapping)
+ dml2_apply_det_buffer_allocation_policy(dml2, &dml2->v20.scratch.cur_display_config);
result = pack_and_call_dml_mode_support_ex(dml2,
&dml2->v20.scratch.cur_display_config,
@@ -732,11 +747,10 @@ static inline struct dml2_context *dml2_allocate_memory(void)
static void dml2_init(const struct dc *in_dc, const struct dml2_configuration_options *config, struct dml2_context **dml2)
{
- // TODO : Temporarily add DCN_VERSION_3_2 for N-1 validation. Remove DCN_VERSION_3_2 after N-1 validation phase is complete.
- if ((in_dc->debug.using_dml21) && (in_dc->ctx->dce_version == DCN_VERSION_4_01 || in_dc->ctx->dce_version == DCN_VERSION_3_2)) {
- dml21_reinit(in_dc, dml2, config);
+ if ((in_dc->debug.using_dml21) && (in_dc->ctx->dce_version == DCN_VERSION_4_01)) {
+ dml21_reinit(in_dc, dml2, config);
return;
- }
+ }
// Store config options
(*dml2)->config = *config;
@@ -771,10 +785,8 @@ static void dml2_init(const struct dc *in_dc, const struct dml2_configuration_op
bool dml2_create(const struct dc *in_dc, const struct dml2_configuration_options *config, struct dml2_context **dml2)
{
- // TODO : Temporarily add DCN_VERSION_3_2 for N-1 validation. Remove DCN_VERSION_3_2 after N-1 validation phase is complete.
- if ((in_dc->debug.using_dml21) && (in_dc->ctx->dce_version == DCN_VERSION_4_01 || in_dc->ctx->dce_version == DCN_VERSION_3_2)) {
+ if ((in_dc->debug.using_dml21) && (in_dc->ctx->dce_version == DCN_VERSION_4_01))
return dml21_create(in_dc, dml2, config);
- }
// Allocate Mode Lib Ctx
*dml2 = dml2_allocate_memory();
@@ -842,8 +854,7 @@ void dml2_reinit(const struct dc *in_dc,
const struct dml2_configuration_options *config,
struct dml2_context **dml2)
{
- // TODO : Temporarily add DCN_VERSION_3_2 for N-1 validation. Remove DCN_VERSION_3_2 after N-1 validation phase is complete.
- if ((in_dc->debug.using_dml21) && (in_dc->ctx->dce_version == DCN_VERSION_4_01 || in_dc->ctx->dce_version == DCN_VERSION_3_2)) {
+ if ((in_dc->debug.using_dml21) && (in_dc->ctx->dce_version == DCN_VERSION_4_01)) {
dml21_reinit(in_dc, dml2, config);
return;
}
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml_display_rq_dlg_calc.c b/drivers/gpu/drm/amd/display/dc/dml2/dml_display_rq_dlg_calc.c
index 377ef6d01ae5..00d22e542469 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml_display_rq_dlg_calc.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml_display_rq_dlg_calc.c
@@ -427,18 +427,6 @@ void dml_rq_dlg_get_dlg_reg(dml_display_dlg_regs_st *disp_dlg_regs,
dml_print("DML_DLG: %s: disp_dlg_regs->dst_y_per_vm_flip = 0x%x\n", __func__, disp_dlg_regs->dst_y_per_vm_flip);
dml_print("DML_DLG: %s: disp_dlg_regs->dst_y_per_row_flip = 0x%x\n", __func__, disp_dlg_regs->dst_y_per_row_flip);
- // hack for FPGA
- /* NOTE: We dont have getenv defined in driver and it does not make any sense in the driver */
- /*char* fpga_env = getenv("FPGA_FPDIV");
- if(fpga_env !=NULL)
- {
- if(disp_dlg_regs->vratio_prefetch >= (dml_uint_t)dml_pow(2, 22))
- {
- disp_dlg_regs->vratio_prefetch = (dml_uint_t)dml_pow(2, 22)-1;
- dml_print("FPGA msg: vratio_prefetch exceed the max value, the register field is [21:0]\n");
- }
- }*/
-
disp_dlg_regs->refcyc_per_vm_group_vblank = (dml_uint_t)(dml_get_refcyc_per_vm_group_vblank_in_us(mode_lib, pipe_idx) * refclk_freq_in_mhz);
disp_dlg_regs->refcyc_per_vm_group_flip = (dml_uint_t)(dml_get_refcyc_per_vm_group_flip_in_us(mode_lib, pipe_idx) * refclk_freq_in_mhz);
disp_dlg_regs->refcyc_per_vm_req_vblank = (dml_uint_t)(dml_get_refcyc_per_vm_req_vblank_in_us(mode_lib, pipe_idx) * refclk_freq_in_mhz * dml_pow(2, 10));
diff --git a/drivers/gpu/drm/amd/display/dc/dsc/dc_dsc.c b/drivers/gpu/drm/amd/display/dc/dsc/dc_dsc.c
index d9aaebfa3a0a..11535922b5ff 100644
--- a/drivers/gpu/drm/amd/display/dc/dsc/dc_dsc.c
+++ b/drivers/gpu/drm/amd/display/dc/dsc/dc_dsc.c
@@ -30,6 +30,9 @@
#include "rc_calc.h"
#include "fixed31_32.h"
+#define DC_LOGGER \
+ dsc->ctx->logger
+
/* This module's internal functions */
/* default DSC policy target bitrate limit is 16bpp */
@@ -480,6 +483,48 @@ bool dc_dsc_compute_bandwidth_range(
return is_dsc_possible;
}
+void dc_dsc_dump_encoder_caps(const struct display_stream_compressor *dsc,
+ const struct dc_crtc_timing *timing)
+{
+ struct dsc_enc_caps dsc_enc_caps;
+
+ get_dsc_enc_caps(dsc, &dsc_enc_caps, timing->pix_clk_100hz);
+
+ DC_LOG_DSC("dsc encoder caps:");
+ DC_LOG_DSC("\tdsc_version 0x%x", dsc_enc_caps.dsc_version);
+ DC_LOG_DSC("\tslice_caps 0x%x", dsc_enc_caps.slice_caps.raw);
+ DC_LOG_DSC("\tlb_bit_depth %d", dsc_enc_caps.lb_bit_depth);
+ DC_LOG_DSC("\tis_block_pred_supported %d", dsc_enc_caps.is_block_pred_supported);
+ DC_LOG_DSC("\tcolor_formats 0x%x", dsc_enc_caps.color_formats.raw);
+ DC_LOG_DSC("\tcolor_depth 0x%x", dsc_enc_caps.color_depth.raw);
+ DC_LOG_DSC("\tmax_total_throughput_mps %d", dsc_enc_caps.max_total_throughput_mps);
+ DC_LOG_DSC("\tmax_slice_width %d", dsc_enc_caps.max_slice_width);
+ DC_LOG_DSC("\tbpp_increment_div %d", dsc_enc_caps.bpp_increment_div);
+}
+
+void dc_dsc_dump_decoder_caps(const struct display_stream_compressor *dsc,
+ const struct dsc_dec_dpcd_caps *dsc_sink_caps)
+{
+ DC_LOG_DSC("dsc decoder caps:");
+ DC_LOG_DSC("\tis_dsc_supported %d", dsc_sink_caps->is_dsc_supported);
+ DC_LOG_DSC("\tdsc_version 0x%x", dsc_sink_caps->dsc_version);
+ DC_LOG_DSC("\trc_buffer_size %d", dsc_sink_caps->rc_buffer_size);
+ DC_LOG_DSC("\tslice_caps1 0x%x", dsc_sink_caps->slice_caps1.raw);
+ DC_LOG_DSC("\tslice_caps2 0x%x", dsc_sink_caps->slice_caps2.raw);
+ DC_LOG_DSC("\tlb_bit_depth %d", dsc_sink_caps->lb_bit_depth);
+ DC_LOG_DSC("\tis_block_pred_supported %d", dsc_sink_caps->is_block_pred_supported);
+ DC_LOG_DSC("\tedp_max_bits_per_pixel %d", dsc_sink_caps->edp_max_bits_per_pixel);
+ DC_LOG_DSC("\tcolor_formats 0x%x", dsc_sink_caps->color_formats.raw);
+ DC_LOG_DSC("\tthroughput_mode_0_mps %d", dsc_sink_caps->throughput_mode_0_mps);
+ DC_LOG_DSC("\tthroughput_mode_1_mps %d", dsc_sink_caps->throughput_mode_1_mps);
+ DC_LOG_DSC("\tmax_slice_width %d", dsc_sink_caps->max_slice_width);
+ DC_LOG_DSC("\tbpp_increment_div %d", dsc_sink_caps->bpp_increment_div);
+ DC_LOG_DSC("\tbranch_overall_throughput_0_mps %d", dsc_sink_caps->branch_overall_throughput_0_mps);
+ DC_LOG_DSC("\tbranch_overall_throughput_1_mps %d", dsc_sink_caps->branch_overall_throughput_1_mps);
+ DC_LOG_DSC("\tbranch_max_line_width %d", dsc_sink_caps->branch_max_line_width);
+ DC_LOG_DSC("\tis_dp %d", dsc_sink_caps->is_dp);
+}
+
static void get_dsc_enc_caps(
const struct display_stream_compressor *dsc,
struct dsc_enc_caps *dsc_enc_caps,
diff --git a/drivers/gpu/drm/amd/display/dc/dwb/dcn30/dcn30_dwb.c b/drivers/gpu/drm/amd/display/dc/dwb/dcn30/dcn30_dwb.c
index fae98cf52020..bc058f682438 100644
--- a/drivers/gpu/drm/amd/display/dc/dwb/dcn30/dcn30_dwb.c
+++ b/drivers/gpu/drm/amd/display/dc/dwb/dcn30/dcn30_dwb.c
@@ -270,16 +270,3 @@ void dcn30_dwbc_construct(struct dcn30_dwbc *dwbc30,
dwbc30->dwbc_shift = dwbc_shift;
dwbc30->dwbc_mask = dwbc_mask;
}
-
-void dwb3_set_host_read_rate_control(struct dwbc *dwbc, bool host_read_delay)
-{
- struct dcn30_dwbc *dwbc30 = TO_DCN30_DWBC(dwbc);
-
- /*
- * Set maximum delay of host read access to DWBSCL LUT or OGAM LUT if there are no
- * idle cycles in HW pipeline (in number of clock cycles times 4)
- */
- REG_UPDATE(DWB_HOST_READ_CONTROL, DWB_HOST_READ_RATE_CONTROL, host_read_delay);
-
- DC_LOG_DWB("%s dwb3_rate_control at inst = %d", __func__, dwbc->inst);
-}
diff --git a/drivers/gpu/drm/amd/display/dc/dwb/dcn30/dcn30_dwb.h b/drivers/gpu/drm/amd/display/dc/dwb/dcn30/dcn30_dwb.h
index 0f3f7c5fbaec..7f053f49ec6a 100644
--- a/drivers/gpu/drm/amd/display/dc/dwb/dcn30/dcn30_dwb.h
+++ b/drivers/gpu/drm/amd/display/dc/dwb/dcn30/dcn30_dwb.h
@@ -914,7 +914,6 @@ bool dwb3_ogam_set_input_transfer_func(
struct dwbc *dwbc,
const struct dc_transfer_func *in_transfer_func_dwb_ogam);
-void dwb3_set_host_read_rate_control(struct dwbc *dwbc, bool host_read_delay);
#endif
diff --git a/drivers/gpu/drm/amd/display/dc/hubp/dcn10/dcn10_hubp.c b/drivers/gpu/drm/amd/display/dc/hubp/dcn10/dcn10_hubp.c
index 22ac2b7e49ae..8364c9f9231a 100644
--- a/drivers/gpu/drm/amd/display/dc/hubp/dcn10/dcn10_hubp.c
+++ b/drivers/gpu/drm/amd/display/dc/hubp/dcn10/dcn10_hubp.c
@@ -140,7 +140,7 @@ void hubp1_vready_workaround(struct hubp *hubp,
void hubp1_program_tiling(
struct hubp *hubp,
- const union dc_tiling_info *info,
+ const struct dc_tiling_info *info,
const enum surface_pixel_format pixel_format)
{
struct dcn10_hubp *hubp1 = TO_DCN10_HUBP(hubp);
@@ -518,6 +518,20 @@ bool hubp1_program_surface_flip_and_addr(
return true;
}
+void hubp1_clear_tiling(struct hubp *hubp)
+{
+ struct dcn10_hubp *hubp1 = TO_DCN10_HUBP(hubp);
+
+ REG_UPDATE(DCHUBP_REQ_SIZE_CONFIG, SWATH_HEIGHT, 0);
+ REG_UPDATE(DCSURF_TILING_CONFIG, SW_MODE, DC_SW_LINEAR);
+
+ REG_UPDATE_4(DCSURF_SURFACE_CONTROL,
+ PRIMARY_SURFACE_DCC_EN, 0,
+ PRIMARY_SURFACE_DCC_IND_64B_BLK, 0,
+ SECONDARY_SURFACE_DCC_EN, 0,
+ SECONDARY_SURFACE_DCC_IND_64B_BLK, 0);
+}
+
void hubp1_dcc_control(struct hubp *hubp, bool enable,
enum hubp_ind_block_size independent_64b_blks)
{
@@ -535,7 +549,7 @@ void hubp1_dcc_control(struct hubp *hubp, bool enable,
void hubp1_program_surface_config(
struct hubp *hubp,
enum surface_pixel_format format,
- union dc_tiling_info *tiling_info,
+ struct dc_tiling_info *tiling_info,
struct plane_size *plane_size,
enum dc_rotation_angle rotation,
struct dc_plane_dcc_param *dcc,
@@ -1363,6 +1377,7 @@ static const struct hubp_funcs dcn10_hubp_funcs = {
.hubp_disable_control = hubp1_disable_control,
.hubp_get_underflow_status = hubp1_get_underflow_status,
.hubp_init = hubp1_init,
+ .hubp_clear_tiling = hubp1_clear_tiling,
.dmdata_set_attributes = NULL,
.dmdata_load = NULL,
diff --git a/drivers/gpu/drm/amd/display/dc/hubp/dcn10/dcn10_hubp.h b/drivers/gpu/drm/amd/display/dc/hubp/dcn10/dcn10_hubp.h
index 69119b2fdce2..a85dc3be786f 100644
--- a/drivers/gpu/drm/amd/display/dc/hubp/dcn10/dcn10_hubp.h
+++ b/drivers/gpu/drm/amd/display/dc/hubp/dcn10/dcn10_hubp.h
@@ -706,7 +706,7 @@ struct dcn10_hubp {
void hubp1_program_surface_config(
struct hubp *hubp,
enum surface_pixel_format format,
- union dc_tiling_info *tiling_info,
+ struct dc_tiling_info *tiling_info,
struct plane_size *plane_size,
enum dc_rotation_angle rotation,
struct dc_plane_dcc_param *dcc,
@@ -739,7 +739,7 @@ void hubp1_program_rotation(
void hubp1_program_tiling(
struct hubp *hubp,
- const union dc_tiling_info *info,
+ const struct dc_tiling_info *info,
const enum surface_pixel_format pixel_format);
void hubp1_dcc_control(struct hubp *hubp,
@@ -794,4 +794,6 @@ void hubp1_soft_reset(struct hubp *hubp, bool reset);
void hubp1_set_flip_int(struct hubp *hubp);
+void hubp1_clear_tiling(struct hubp *hubp);
+
#endif
diff --git a/drivers/gpu/drm/amd/display/dc/hubp/dcn20/dcn20_hubp.c b/drivers/gpu/drm/amd/display/dc/hubp/dcn20/dcn20_hubp.c
index 0637e4c552d8..c74f6a3313a2 100644
--- a/drivers/gpu/drm/amd/display/dc/hubp/dcn20/dcn20_hubp.c
+++ b/drivers/gpu/drm/amd/display/dc/hubp/dcn20/dcn20_hubp.c
@@ -310,7 +310,7 @@ void hubp2_setup_interdependent(
*/
static void hubp2_program_tiling(
struct dcn20_hubp *hubp2,
- const union dc_tiling_info *info,
+ const struct dc_tiling_info *info,
const enum surface_pixel_format pixel_format)
{
REG_UPDATE_3(DCSURF_ADDR_CONFIG,
@@ -406,6 +406,20 @@ void hubp2_program_rotation(
H_MIRROR_EN, mirror);
}
+void hubp2_clear_tiling(struct hubp *hubp)
+{
+ struct dcn20_hubp *hubp2 = TO_DCN20_HUBP(hubp);
+
+ REG_UPDATE(DCHUBP_REQ_SIZE_CONFIG, SWATH_HEIGHT, 0);
+ REG_UPDATE(DCSURF_TILING_CONFIG, SW_MODE, DC_SW_LINEAR);
+
+ REG_UPDATE_4(DCSURF_SURFACE_CONTROL,
+ PRIMARY_SURFACE_DCC_EN, 0,
+ PRIMARY_SURFACE_DCC_IND_64B_BLK, 0,
+ SECONDARY_SURFACE_DCC_EN, 0,
+ SECONDARY_SURFACE_DCC_IND_64B_BLK, 0);
+}
+
void hubp2_dcc_control(struct hubp *hubp, bool enable,
enum hubp_ind_block_size independent_64b_blks)
{
@@ -536,7 +550,7 @@ void hubp2_program_pixel_format(
void hubp2_program_surface_config(
struct hubp *hubp,
enum surface_pixel_format format,
- union dc_tiling_info *tiling_info,
+ struct dc_tiling_info *tiling_info,
struct plane_size *plane_size,
enum dc_rotation_angle rotation,
struct dc_plane_dcc_param *dcc,
@@ -1676,6 +1690,7 @@ static struct hubp_funcs dcn20_hubp_funcs = {
.hubp_in_blank = hubp1_in_blank,
.hubp_soft_reset = hubp1_soft_reset,
.hubp_set_flip_int = hubp1_set_flip_int,
+ .hubp_clear_tiling = hubp2_clear_tiling,
};
diff --git a/drivers/gpu/drm/amd/display/dc/hubp/dcn20/dcn20_hubp.h b/drivers/gpu/drm/amd/display/dc/hubp/dcn20/dcn20_hubp.h
index 18e194507e36..6968087a3605 100644
--- a/drivers/gpu/drm/amd/display/dc/hubp/dcn20/dcn20_hubp.h
+++ b/drivers/gpu/drm/amd/display/dc/hubp/dcn20/dcn20_hubp.h
@@ -382,7 +382,7 @@ void hubp2_program_pixel_format(
void hubp2_program_surface_config(
struct hubp *hubp,
enum surface_pixel_format format,
- union dc_tiling_info *tiling_info,
+ struct dc_tiling_info *tiling_info,
struct plane_size *plane_size,
enum dc_rotation_angle rotation,
struct dc_plane_dcc_param *dcc,
@@ -409,6 +409,8 @@ void hubp2_read_state_common(struct hubp *hubp);
void hubp2_read_state(struct hubp *hubp);
+void hubp2_clear_tiling(struct hubp *hubp);
+
#endif /* __DC_MEM_INPUT_DCN20_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/hubp/dcn201/dcn201_hubp.c b/drivers/gpu/drm/amd/display/dc/hubp/dcn201/dcn201_hubp.c
index cd2bfcc51276..65c628078ca2 100644
--- a/drivers/gpu/drm/amd/display/dc/hubp/dcn201/dcn201_hubp.c
+++ b/drivers/gpu/drm/amd/display/dc/hubp/dcn201/dcn201_hubp.c
@@ -42,7 +42,7 @@
static void hubp201_program_surface_config(
struct hubp *hubp,
enum surface_pixel_format format,
- union dc_tiling_info *tiling_info,
+ struct dc_tiling_info *tiling_info,
struct plane_size *plane_size,
enum dc_rotation_angle rotation,
struct dc_plane_dcc_param *dcc,
@@ -131,6 +131,7 @@ static struct hubp_funcs dcn201_hubp_funcs = {
.hubp_clear_underflow = hubp1_clear_underflow,
.hubp_set_flip_control_surface_gsl = hubp2_set_flip_control_surface_gsl,
.hubp_init = hubp1_init,
+ .hubp_clear_tiling = hubp1_clear_tiling,
};
bool dcn201_hubp_construct(
diff --git a/drivers/gpu/drm/amd/display/dc/hubp/dcn21/dcn21_hubp.c b/drivers/gpu/drm/amd/display/dc/hubp/dcn21/dcn21_hubp.c
index e13d69a22c1c..edbdb8c88d5c 100644
--- a/drivers/gpu/drm/amd/display/dc/hubp/dcn21/dcn21_hubp.c
+++ b/drivers/gpu/drm/amd/display/dc/hubp/dcn21/dcn21_hubp.c
@@ -837,6 +837,7 @@ static struct hubp_funcs dcn21_hubp_funcs = {
.hubp_init = hubp21_init,
.validate_dml_output = hubp21_validate_dml_output,
.hubp_set_flip_int = hubp1_set_flip_int,
+ .hubp_clear_tiling = hubp1_clear_tiling,
};
bool hubp21_construct(
diff --git a/drivers/gpu/drm/amd/display/dc/hubp/dcn30/dcn30_hubp.c b/drivers/gpu/drm/amd/display/dc/hubp/dcn30/dcn30_hubp.c
index 60a64d290352..12b282ed7067 100644
--- a/drivers/gpu/drm/amd/display/dc/hubp/dcn30/dcn30_hubp.c
+++ b/drivers/gpu/drm/amd/display/dc/hubp/dcn30/dcn30_hubp.c
@@ -318,7 +318,7 @@ bool hubp3_program_surface_flip_and_addr(
void hubp3_program_tiling(
struct dcn20_hubp *hubp2,
- const union dc_tiling_info *info,
+ const struct dc_tiling_info *info,
const enum surface_pixel_format pixel_format)
{
REG_UPDATE_4(DCSURF_ADDR_CONFIG,
@@ -334,6 +334,22 @@ void hubp3_program_tiling(
}
+void hubp3_clear_tiling(struct hubp *hubp)
+{
+ struct dcn20_hubp *hubp2 = TO_DCN20_HUBP(hubp);
+
+ REG_UPDATE(DCHUBP_REQ_SIZE_CONFIG, SWATH_HEIGHT, 0);
+ REG_UPDATE(DCSURF_TILING_CONFIG, SW_MODE, DC_SW_LINEAR);
+
+ REG_UPDATE_6(DCSURF_SURFACE_CONTROL,
+ PRIMARY_SURFACE_DCC_EN, 0,
+ PRIMARY_SURFACE_DCC_IND_BLK, 0,
+ PRIMARY_SURFACE_DCC_IND_BLK_C, 0,
+ SECONDARY_SURFACE_DCC_EN, 0,
+ SECONDARY_SURFACE_DCC_IND_BLK, 0,
+ SECONDARY_SURFACE_DCC_IND_BLK_C, 0);
+}
+
void hubp3_dcc_control(struct hubp *hubp, bool enable,
enum hubp_ind_block_size blk_size)
{
@@ -395,7 +411,7 @@ void hubp3_dmdata_set_attributes(
void hubp3_program_surface_config(
struct hubp *hubp,
enum surface_pixel_format format,
- union dc_tiling_info *tiling_info,
+ struct dc_tiling_info *tiling_info,
struct plane_size *plane_size,
enum dc_rotation_angle rotation,
struct dc_plane_dcc_param *dcc,
@@ -512,6 +528,7 @@ static struct hubp_funcs dcn30_hubp_funcs = {
.hubp_in_blank = hubp1_in_blank,
.hubp_soft_reset = hubp1_soft_reset,
.hubp_set_flip_int = hubp1_set_flip_int,
+ .hubp_clear_tiling = hubp3_clear_tiling,
};
bool hubp3_construct(
diff --git a/drivers/gpu/drm/amd/display/dc/hubp/dcn30/dcn30_hubp.h b/drivers/gpu/drm/amd/display/dc/hubp/dcn30/dcn30_hubp.h
index b010531a7fe8..b7d7adf0b58c 100644
--- a/drivers/gpu/drm/amd/display/dc/hubp/dcn30/dcn30_hubp.h
+++ b/drivers/gpu/drm/amd/display/dc/hubp/dcn30/dcn30_hubp.h
@@ -264,7 +264,7 @@ bool hubp3_program_surface_flip_and_addr(
void hubp3_program_surface_config(
struct hubp *hubp,
enum surface_pixel_format format,
- union dc_tiling_info *tiling_info,
+ struct dc_tiling_info *tiling_info,
struct plane_size *plane_size,
enum dc_rotation_angle rotation,
struct dc_plane_dcc_param *dcc,
@@ -280,7 +280,7 @@ void hubp3_setup(
void hubp3_program_tiling(
struct dcn20_hubp *hubp2,
- const union dc_tiling_info *info,
+ const struct dc_tiling_info *info,
const enum surface_pixel_format pixel_format);
void hubp3_dcc_control(struct hubp *hubp, bool enable,
@@ -297,6 +297,8 @@ void hubp3_read_state(struct hubp *hubp);
void hubp3_init(struct hubp *hubp);
+void hubp3_clear_tiling(struct hubp *hubp);
+
#endif /* __DC_HUBP_DCN30_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/hubp/dcn31/dcn31_hubp.c b/drivers/gpu/drm/amd/display/dc/hubp/dcn31/dcn31_hubp.c
index 8394e8c06919..46b804ed05fb 100644
--- a/drivers/gpu/drm/amd/display/dc/hubp/dcn31/dcn31_hubp.c
+++ b/drivers/gpu/drm/amd/display/dc/hubp/dcn31/dcn31_hubp.c
@@ -96,6 +96,7 @@ static struct hubp_funcs dcn31_hubp_funcs = {
.hubp_set_flip_int = hubp1_set_flip_int,
.hubp_in_blank = hubp1_in_blank,
.program_extended_blank = hubp31_program_extended_blank,
+ .hubp_clear_tiling = hubp3_clear_tiling,
};
bool hubp31_construct(
diff --git a/drivers/gpu/drm/amd/display/dc/hubp/dcn32/dcn32_hubp.c b/drivers/gpu/drm/amd/display/dc/hubp/dcn32/dcn32_hubp.c
index ca5b4b28a664..8b5bd73b8094 100644
--- a/drivers/gpu/drm/amd/display/dc/hubp/dcn32/dcn32_hubp.c
+++ b/drivers/gpu/drm/amd/display/dc/hubp/dcn32/dcn32_hubp.c
@@ -201,7 +201,8 @@ static struct hubp_funcs dcn32_hubp_funcs = {
.hubp_update_force_cursor_pstate_disallow = hubp32_update_force_cursor_pstate_disallow,
.phantom_hubp_post_enable = hubp32_phantom_hubp_post_enable,
.hubp_update_mall_sel = hubp32_update_mall_sel,
- .hubp_prepare_subvp_buffering = hubp32_prepare_subvp_buffering
+ .hubp_prepare_subvp_buffering = hubp32_prepare_subvp_buffering,
+ .hubp_clear_tiling = hubp3_clear_tiling,
};
bool hubp32_construct(
diff --git a/drivers/gpu/drm/amd/display/dc/hubp/dcn35/dcn35_hubp.c b/drivers/gpu/drm/amd/display/dc/hubp/dcn35/dcn35_hubp.c
index d1f05b82b3dd..faf37febc6fb 100644
--- a/drivers/gpu/drm/amd/display/dc/hubp/dcn35/dcn35_hubp.c
+++ b/drivers/gpu/drm/amd/display/dc/hubp/dcn35/dcn35_hubp.c
@@ -172,7 +172,7 @@ void hubp35_program_pixel_format(
void hubp35_program_surface_config(
struct hubp *hubp,
enum surface_pixel_format format,
- union dc_tiling_info *tiling_info,
+ struct dc_tiling_info *tiling_info,
struct plane_size *plane_size,
enum dc_rotation_angle rotation,
struct dc_plane_dcc_param *dcc,
@@ -216,6 +216,7 @@ static struct hubp_funcs dcn35_hubp_funcs = {
.hubp_set_flip_int = hubp1_set_flip_int,
.hubp_in_blank = hubp1_in_blank,
.program_extended_blank = hubp31_program_extended_blank_value,
+ .hubp_clear_tiling = hubp3_clear_tiling,
};
bool hubp35_construct(
diff --git a/drivers/gpu/drm/amd/display/dc/hubp/dcn35/dcn35_hubp.h b/drivers/gpu/drm/amd/display/dc/hubp/dcn35/dcn35_hubp.h
index 586b43aa5834..d913f80b3130 100644
--- a/drivers/gpu/drm/amd/display/dc/hubp/dcn35/dcn35_hubp.h
+++ b/drivers/gpu/drm/amd/display/dc/hubp/dcn35/dcn35_hubp.h
@@ -65,7 +65,7 @@ void hubp35_program_pixel_format(
void hubp35_program_surface_config(
struct hubp *hubp,
enum surface_pixel_format format,
- union dc_tiling_info *tiling_info,
+ struct dc_tiling_info *tiling_info,
struct plane_size *plane_size,
enum dc_rotation_angle rotation,
struct dc_plane_dcc_param *dcc,
diff --git a/drivers/gpu/drm/amd/display/dc/hubp/dcn401/dcn401_hubp.c b/drivers/gpu/drm/amd/display/dc/hubp/dcn401/dcn401_hubp.c
index b1ebf5053b4f..28ceceaf9e31 100644
--- a/drivers/gpu/drm/amd/display/dc/hubp/dcn401/dcn401_hubp.c
+++ b/drivers/gpu/drm/amd/display/dc/hubp/dcn401/dcn401_hubp.c
@@ -40,7 +40,7 @@
#define FN(reg_name, field_name) \
hubp2->hubp_shift->field_name, hubp2->hubp_mask->field_name
-static void hubp401_program_3dlut_fl_addr(struct hubp *hubp,
+void hubp401_program_3dlut_fl_addr(struct hubp *hubp,
const struct dc_plane_address address)
{
struct dcn20_hubp *hubp2 = TO_DCN20_HUBP(hubp);
@@ -49,14 +49,14 @@ static void hubp401_program_3dlut_fl_addr(struct hubp *hubp,
REG_WRITE(HUBP_3DLUT_ADDRESS_LOW, address.lut3d.addr.low_part);
}
-static void hubp401_program_3dlut_fl_dlg_param(struct hubp *hubp, int refcyc_per_3dlut_group)
+void hubp401_program_3dlut_fl_dlg_param(struct hubp *hubp, int refcyc_per_3dlut_group)
{
struct dcn20_hubp *hubp2 = TO_DCN20_HUBP(hubp);
REG_UPDATE(HUBP_3DLUT_DLG_PARAM, REFCYC_PER_3DLUT_GROUP, refcyc_per_3dlut_group);
}
-static void hubp401_enable_3dlut_fl(struct hubp *hubp, bool enable)
+void hubp401_enable_3dlut_fl(struct hubp *hubp, bool enable)
{
struct dcn20_hubp *hubp2 = TO_DCN20_HUBP(hubp);
@@ -72,28 +72,28 @@ int hubp401_get_3dlut_fl_done(struct hubp *hubp)
return ret;
}
-static void hubp401_program_3dlut_fl_addressing_mode(struct hubp *hubp, enum hubp_3dlut_fl_addressing_mode addr_mode)
+void hubp401_program_3dlut_fl_addressing_mode(struct hubp *hubp, enum hubp_3dlut_fl_addressing_mode addr_mode)
{
struct dcn20_hubp *hubp2 = TO_DCN20_HUBP(hubp);
REG_UPDATE(HUBP_3DLUT_CONTROL, HUBP_3DLUT_ADDRESSING_MODE, addr_mode);
}
-static void hubp401_program_3dlut_fl_width(struct hubp *hubp, enum hubp_3dlut_fl_width width)
+void hubp401_program_3dlut_fl_width(struct hubp *hubp, enum hubp_3dlut_fl_width width)
{
struct dcn20_hubp *hubp2 = TO_DCN20_HUBP(hubp);
REG_UPDATE(HUBP_3DLUT_CONTROL, HUBP_3DLUT_WIDTH, width);
}
-static void hubp401_program_3dlut_fl_tmz_protected(struct hubp *hubp, bool protection_enabled)
+void hubp401_program_3dlut_fl_tmz_protected(struct hubp *hubp, bool protection_enabled)
{
struct dcn20_hubp *hubp2 = TO_DCN20_HUBP(hubp);
REG_UPDATE(HUBP_3DLUT_CONTROL, HUBP_3DLUT_TMZ, protection_enabled ? 1 : 0);
}
-static void hubp401_program_3dlut_fl_crossbar(struct hubp *hubp,
+void hubp401_program_3dlut_fl_crossbar(struct hubp *hubp,
enum hubp_3dlut_fl_crossbar_bit_slice bit_slice_y_g,
enum hubp_3dlut_fl_crossbar_bit_slice bit_slice_cb_b,
enum hubp_3dlut_fl_crossbar_bit_slice bit_slice_cr_r)
@@ -106,21 +106,21 @@ static void hubp401_program_3dlut_fl_crossbar(struct hubp *hubp,
HUBP_3DLUT_CROSSBAR_SELECT_CR_R, bit_slice_cr_r);
}
-static void hubp401_update_3dlut_fl_bias_scale(struct hubp *hubp, uint16_t bias, uint16_t scale)
+void hubp401_update_3dlut_fl_bias_scale(struct hubp *hubp, uint16_t bias, uint16_t scale)
{
struct dcn20_hubp *hubp2 = TO_DCN20_HUBP(hubp);
REG_UPDATE_2(_3DLUT_FL_BIAS_SCALE, HUBP0_3DLUT_FL_BIAS, bias, HUBP0_3DLUT_FL_SCALE, scale);
}
-static void hubp401_program_3dlut_fl_mode(struct hubp *hubp, enum hubp_3dlut_fl_mode mode)
+void hubp401_program_3dlut_fl_mode(struct hubp *hubp, enum hubp_3dlut_fl_mode mode)
{
struct dcn20_hubp *hubp2 = TO_DCN20_HUBP(hubp);
REG_UPDATE(_3DLUT_FL_CONFIG, HUBP0_3DLUT_FL_MODE, mode);
}
-static void hubp401_program_3dlut_fl_format(struct hubp *hubp, enum hubp_3dlut_fl_format format)
+void hubp401_program_3dlut_fl_format(struct hubp *hubp, enum hubp_3dlut_fl_format format)
{
struct dcn20_hubp *hubp2 = TO_DCN20_HUBP(hubp);
@@ -145,30 +145,44 @@ void hubp401_init(struct hubp *hubp)
}
void hubp401_vready_at_or_After_vsync(struct hubp *hubp,
- struct _vcs_dpi_display_pipe_dest_params_st *pipe_dest)
+ union dml2_global_sync_programming *pipe_global_sync,
+ struct dc_crtc_timing *timing)
{
- uint32_t value = 0;
+ unsigned int vstartup_lines = pipe_global_sync->dcn4x.vstartup_lines;
+ unsigned int vupdate_offset_pixels = pipe_global_sync->dcn4x.vupdate_offset_pixels;
+ unsigned int vupdate_width_pixels = pipe_global_sync->dcn4x.vupdate_vupdate_width_pixels;
+ unsigned int vready_offset_pixels = pipe_global_sync->dcn4x.vready_offset_pixels;
+ unsigned int htotal = timing->h_total;
+ unsigned int vblank_start = 0;
+ unsigned int vblank_end = 0;
+ unsigned int pixel_width = 0;
+ uint32_t reg_value = 0;
+ bool is_vready_at_or_after_vsync = false;
struct dcn20_hubp *hubp2 = TO_DCN20_HUBP(hubp);
+
/*
* if (VSTARTUP_START - (VREADY_OFFSET+VUPDATE_WIDTH+VUPDATE_OFFSET)/htotal) <= OTG_V_BLANK_END
* Set HUBP_VREADY_AT_OR_AFTER_VSYNC = 1
* else
* Set HUBP_VREADY_AT_OR_AFTER_VSYNC = 0
*/
- if (pipe_dest->htotal != 0) {
- if ((pipe_dest->vstartup_start - (pipe_dest->vready_offset+pipe_dest->vupdate_width
- + pipe_dest->vupdate_offset) / pipe_dest->htotal) <= pipe_dest->vblank_end) {
- value = 1;
- } else
- value = 0;
+ if (htotal != 0) {
+ vblank_start = timing->v_total - timing->v_front_porch;
+ vblank_end = vblank_start - timing->v_addressable - timing->v_border_top - timing->v_border_bottom;
+ pixel_width = vready_offset_pixels + vupdate_width_pixels + vupdate_offset_pixels;
+
+ is_vready_at_or_after_vsync = (vstartup_lines - pixel_width / htotal) <= vblank_end;
+
+ if (is_vready_at_or_after_vsync)
+ reg_value = 1;
}
- REG_UPDATE(DCHUBP_CNTL, HUBP_VREADY_AT_OR_AFTER_VSYNC, value);
+ REG_UPDATE(DCHUBP_CNTL, HUBP_VREADY_AT_OR_AFTER_VSYNC, reg_value);
}
void hubp401_program_requestor(
struct hubp *hubp,
- struct _vcs_dpi_display_rq_regs_st *rq_regs)
+ struct dml2_display_rq_regs *rq_regs)
{
struct dcn20_hubp *hubp2 = TO_DCN20_HUBP(hubp);
@@ -196,8 +210,8 @@ void hubp401_program_requestor(
void hubp401_program_deadline(
struct hubp *hubp,
- struct _vcs_dpi_display_dlg_regs_st *dlg_attr,
- struct _vcs_dpi_display_ttu_regs_st *ttu_attr)
+ struct dml2_display_dlg_regs *dlg_attr,
+ struct dml2_display_ttu_regs *ttu_attr)
{
struct dcn20_hubp *hubp2 = TO_DCN20_HUBP(hubp);
@@ -294,66 +308,64 @@ void hubp401_program_deadline(
void hubp401_setup(
struct hubp *hubp,
- struct _vcs_dpi_display_dlg_regs_st *dlg_attr,
- struct _vcs_dpi_display_ttu_regs_st *ttu_attr,
- struct _vcs_dpi_display_rq_regs_st *rq_regs,
- struct _vcs_dpi_display_pipe_dest_params_st *pipe_dest)
+ struct dml2_dchub_per_pipe_register_set *pipe_regs,
+ union dml2_global_sync_programming *pipe_global_sync,
+ struct dc_crtc_timing *timing)
{
/* otg is locked when this func is called. Register are double buffered.
* disable the requestors is not needed
*/
- hubp401_vready_at_or_After_vsync(hubp, pipe_dest);
- hubp401_program_requestor(hubp, rq_regs);
- hubp401_program_deadline(hubp, dlg_attr, ttu_attr);
+ hubp401_vready_at_or_After_vsync(hubp, pipe_global_sync, timing);
+ hubp401_program_requestor(hubp, &pipe_regs->rq_regs);
+ hubp401_program_deadline(hubp, &pipe_regs->dlg_regs, &pipe_regs->ttu_regs);
}
void hubp401_setup_interdependent(
struct hubp *hubp,
- struct _vcs_dpi_display_dlg_regs_st *dlg_attr,
- struct _vcs_dpi_display_ttu_regs_st *ttu_attr)
+ struct dml2_dchub_per_pipe_register_set *pipe_regs)
{
struct dcn20_hubp *hubp2 = TO_DCN20_HUBP(hubp);
REG_SET_2(PREFETCH_SETTINGS, 0,
- DST_Y_PREFETCH, dlg_attr->dst_y_prefetch,
- VRATIO_PREFETCH, dlg_attr->vratio_prefetch);
+ DST_Y_PREFETCH, pipe_regs->dlg_regs.dst_y_prefetch,
+ VRATIO_PREFETCH, pipe_regs->dlg_regs.vratio_prefetch);
REG_SET(PREFETCH_SETTINGS_C, 0,
- VRATIO_PREFETCH_C, dlg_attr->vratio_prefetch_c);
+ VRATIO_PREFETCH_C, pipe_regs->dlg_regs.vratio_prefetch_c);
REG_SET_2(VBLANK_PARAMETERS_0, 0,
- DST_Y_PER_VM_VBLANK, dlg_attr->dst_y_per_vm_vblank,
- DST_Y_PER_ROW_VBLANK, dlg_attr->dst_y_per_row_vblank);
+ DST_Y_PER_VM_VBLANK, pipe_regs->dlg_regs.dst_y_per_vm_vblank,
+ DST_Y_PER_ROW_VBLANK, pipe_regs->dlg_regs.dst_y_per_row_vblank);
REG_SET_2(FLIP_PARAMETERS_0, 0,
- DST_Y_PER_VM_FLIP, dlg_attr->dst_y_per_vm_flip,
- DST_Y_PER_ROW_FLIP, dlg_attr->dst_y_per_row_flip);
+ DST_Y_PER_VM_FLIP, pipe_regs->dlg_regs.dst_y_per_vm_flip,
+ DST_Y_PER_ROW_FLIP, pipe_regs->dlg_regs.dst_y_per_row_flip);
REG_SET(VBLANK_PARAMETERS_3, 0,
- REFCYC_PER_META_CHUNK_VBLANK_L, dlg_attr->refcyc_per_meta_chunk_vblank_l);
+ REFCYC_PER_META_CHUNK_VBLANK_L, pipe_regs->dlg_regs.refcyc_per_meta_chunk_vblank_l);
REG_SET(VBLANK_PARAMETERS_4, 0,
- REFCYC_PER_META_CHUNK_VBLANK_C, dlg_attr->refcyc_per_meta_chunk_vblank_c);
+ REFCYC_PER_META_CHUNK_VBLANK_C, pipe_regs->dlg_regs.refcyc_per_meta_chunk_vblank_c);
REG_SET(FLIP_PARAMETERS_2, 0,
- REFCYC_PER_META_CHUNK_FLIP_L, dlg_attr->refcyc_per_meta_chunk_flip_l);
+ REFCYC_PER_META_CHUNK_FLIP_L, pipe_regs->dlg_regs.refcyc_per_meta_chunk_flip_l);
REG_SET_2(PER_LINE_DELIVERY_PRE, 0,
- REFCYC_PER_LINE_DELIVERY_PRE_L, dlg_attr->refcyc_per_line_delivery_pre_l,
- REFCYC_PER_LINE_DELIVERY_PRE_C, dlg_attr->refcyc_per_line_delivery_pre_c);
+ REFCYC_PER_LINE_DELIVERY_PRE_L, pipe_regs->dlg_regs.refcyc_per_line_delivery_pre_l,
+ REFCYC_PER_LINE_DELIVERY_PRE_C, pipe_regs->dlg_regs.refcyc_per_line_delivery_pre_c);
REG_SET(DCN_SURF0_TTU_CNTL1, 0,
REFCYC_PER_REQ_DELIVERY_PRE,
- ttu_attr->refcyc_per_req_delivery_pre_l);
+ pipe_regs->ttu_regs.refcyc_per_req_delivery_pre_l);
REG_SET(DCN_SURF1_TTU_CNTL1, 0,
REFCYC_PER_REQ_DELIVERY_PRE,
- ttu_attr->refcyc_per_req_delivery_pre_c);
+ pipe_regs->ttu_regs.refcyc_per_req_delivery_pre_c);
REG_SET(DCN_CUR0_TTU_CNTL1, 0,
- REFCYC_PER_REQ_DELIVERY_PRE, ttu_attr->refcyc_per_req_delivery_pre_cur0);
+ REFCYC_PER_REQ_DELIVERY_PRE, pipe_regs->ttu_regs.refcyc_per_req_delivery_pre_cur0);
REG_SET_2(DCN_GLOBAL_TTU_CNTL, 0,
- MIN_TTU_VBLANK, ttu_attr->min_ttu_vblank,
- QoS_LEVEL_FLIP, ttu_attr->qos_level_flip);
+ MIN_TTU_VBLANK, pipe_regs->ttu_regs.min_ttu_vblank,
+ QoS_LEVEL_FLIP, pipe_regs->ttu_regs.qos_level_flip);
}
@@ -508,6 +520,18 @@ bool hubp401_program_surface_flip_and_addr(
return true;
}
+void hubp401_clear_tiling(struct hubp *hubp)
+{
+ struct dcn20_hubp *hubp2 = TO_DCN20_HUBP(hubp);
+
+ REG_UPDATE(DCHUBP_REQ_SIZE_CONFIG, SWATH_HEIGHT, 0);
+ REG_UPDATE(DCSURF_TILING_CONFIG, SW_MODE, DC_SW_LINEAR);
+
+ REG_UPDATE_2(DCSURF_SURFACE_CONTROL,
+ PRIMARY_SURFACE_DCC_EN, 0,
+ SECONDARY_SURFACE_DCC_EN, 0);
+}
+
void hubp401_dcc_control(struct hubp *hubp,
struct dc_plane_dcc_param *dcc)
{
@@ -520,7 +544,7 @@ void hubp401_dcc_control(struct hubp *hubp,
void hubp401_program_tiling(
struct dcn20_hubp *hubp2,
- const union dc_tiling_info *info,
+ const struct dc_tiling_info *info,
const enum surface_pixel_format pixel_format)
{
/* DCSURF_ADDR_CONFIG still shows up in reg spec, but does not need to be programmed for DCN4x
@@ -568,7 +592,7 @@ void hubp401_program_size(
void hubp401_program_surface_config(
struct hubp *hubp,
enum surface_pixel_format format,
- union dc_tiling_info *tiling_info,
+ struct dc_tiling_info *tiling_info,
struct plane_size *plane_size,
enum dc_rotation_angle rotation,
struct dc_plane_dcc_param *dcc,
@@ -969,8 +993,8 @@ static struct hubp_funcs dcn401_hubp_funcs = {
.hubp_program_surface_flip_and_addr = hubp401_program_surface_flip_and_addr,
.hubp_program_surface_config = hubp401_program_surface_config,
.hubp_is_flip_pending = hubp2_is_flip_pending,
- .hubp_setup = hubp401_setup,
- .hubp_setup_interdependent = hubp401_setup_interdependent,
+ .hubp_setup2 = hubp401_setup,
+ .hubp_setup_interdependent2 = hubp401_setup_interdependent,
.hubp_set_vm_system_aperture_settings = hubp3_set_vm_system_aperture_settings,
.set_blank = hubp2_set_blank,
.set_blank_regs = hubp2_set_blank_regs,
@@ -1004,7 +1028,8 @@ static struct hubp_funcs dcn401_hubp_funcs = {
.hubp_program_3dlut_fl_width = hubp401_program_3dlut_fl_width,
.hubp_program_3dlut_fl_tmz_protected = hubp401_program_3dlut_fl_tmz_protected,
.hubp_program_3dlut_fl_crossbar = hubp401_program_3dlut_fl_crossbar,
- .hubp_get_3dlut_fl_done = hubp401_get_3dlut_fl_done
+ .hubp_get_3dlut_fl_done = hubp401_get_3dlut_fl_done,
+ .hubp_clear_tiling = hubp2_clear_tiling,
};
bool hubp401_construct(
diff --git a/drivers/gpu/drm/amd/display/dc/hubp/dcn401/dcn401_hubp.h b/drivers/gpu/drm/amd/display/dc/hubp/dcn401/dcn401_hubp.h
index e52fdb5b0cd0..6e1d4c90ddd4 100644
--- a/drivers/gpu/drm/amd/display/dc/hubp/dcn401/dcn401_hubp.h
+++ b/drivers/gpu/drm/amd/display/dc/hubp/dcn401/dcn401_hubp.h
@@ -256,29 +256,15 @@
void hubp401_update_mall_sel(struct hubp *hubp, uint32_t mall_sel, bool c_cursor);
-void hubp401_vready_at_or_After_vsync(struct hubp *hubp,
- struct _vcs_dpi_display_pipe_dest_params_st *pipe_dest);
-
-void hubp401_program_requestor(
- struct hubp *hubp,
- struct _vcs_dpi_display_rq_regs_st *rq_regs);
-
-void hubp401_program_deadline(
- struct hubp *hubp,
- struct _vcs_dpi_display_dlg_regs_st *dlg_attr,
- struct _vcs_dpi_display_ttu_regs_st *ttu_attr);
-
void hubp401_setup(
struct hubp *hubp,
- struct _vcs_dpi_display_dlg_regs_st *dlg_attr,
- struct _vcs_dpi_display_ttu_regs_st *ttu_attr,
- struct _vcs_dpi_display_rq_regs_st *rq_regs,
- struct _vcs_dpi_display_pipe_dest_params_st *pipe_dest);
+ struct dml2_dchub_per_pipe_register_set *pipe_regs,
+ union dml2_global_sync_programming *pipe_global_sync,
+ struct dc_crtc_timing *timing);
void hubp401_setup_interdependent(
struct hubp *hubp,
- struct _vcs_dpi_display_dlg_regs_st *dlg_attr,
- struct _vcs_dpi_display_ttu_regs_st *ttu_attr);
+ struct dml2_dchub_per_pipe_register_set *pipe_regs);
bool hubp401_program_surface_flip_and_addr(
struct hubp *hubp,
@@ -290,7 +276,7 @@ void hubp401_dcc_control(struct hubp *hubp,
void hubp401_program_tiling(
struct dcn20_hubp *hubp2,
- const union dc_tiling_info *info,
+ const struct dc_tiling_info *info,
const enum surface_pixel_format pixel_format);
void hubp401_program_size(
@@ -302,7 +288,7 @@ void hubp401_program_size(
void hubp401_program_surface_config(
struct hubp *hubp,
enum surface_pixel_format format,
- union dc_tiling_info *tiling_info,
+ struct dc_tiling_info *tiling_info,
struct plane_size *plane_size,
enum dc_rotation_angle rotation,
struct dc_plane_dcc_param *dcc,
@@ -340,4 +326,42 @@ int hubp401_get_3dlut_fl_done(struct hubp *hubp);
void hubp401_set_unbounded_requesting(struct hubp *hubp, bool enable);
+void hubp401_update_3dlut_fl_bias_scale(struct hubp *hubp, uint16_t bias, uint16_t scale);
+
+void hubp401_program_3dlut_fl_crossbar(struct hubp *hubp,
+ enum hubp_3dlut_fl_crossbar_bit_slice bit_slice_y_g,
+ enum hubp_3dlut_fl_crossbar_bit_slice bit_slice_cb_b,
+ enum hubp_3dlut_fl_crossbar_bit_slice bit_slice_cr_r);
+
+void hubp401_program_3dlut_fl_tmz_protected(struct hubp *hubp, bool protection_enabled);
+
+void hubp401_program_3dlut_fl_width(struct hubp *hubp, enum hubp_3dlut_fl_width width);
+
+void hubp401_program_3dlut_fl_addressing_mode(struct hubp *hubp, enum hubp_3dlut_fl_addressing_mode addr_mode);
+
+void hubp401_enable_3dlut_fl(struct hubp *hubp, bool enable);
+
+void hubp401_program_3dlut_fl_dlg_param(struct hubp *hubp, int refcyc_per_3dlut_group);
+
+void hubp401_program_3dlut_fl_addr(struct hubp *hubp, const struct dc_plane_address address);
+
+void hubp401_program_3dlut_fl_format(struct hubp *hubp, enum hubp_3dlut_fl_format format);
+
+void hubp401_program_3dlut_fl_mode(struct hubp *hubp, enum hubp_3dlut_fl_mode mode);
+
+void hubp401_clear_tiling(struct hubp *hubp);
+
+void hubp401_vready_at_or_After_vsync(struct hubp *hubp,
+ union dml2_global_sync_programming *pipe_global_sync,
+ struct dc_crtc_timing *timing);
+
+void hubp401_program_requestor(
+ struct hubp *hubp,
+ struct dml2_display_rq_regs *rq_regs);
+
+void hubp401_program_deadline(
+ struct hubp *hubp,
+ struct dml2_display_dlg_regs *dlg_attr,
+ struct dml2_display_ttu_regs *ttu_attr);
+
#endif /* __DC_HUBP_DCN401_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c
index b029ec1b26d3..a5e18ab72394 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c
@@ -1288,7 +1288,7 @@ static void dcn20_power_on_plane_resources(
}
}
-static void dcn20_enable_plane(struct dc *dc, struct pipe_ctx *pipe_ctx,
+void dcn20_enable_plane(struct dc *dc, struct pipe_ctx *pipe_ctx,
struct dc_state *context)
{
//if (dc->debug.sanity_checks) {
@@ -1467,7 +1467,7 @@ void dcn20_pipe_control_lock(
}
}
-static void dcn20_detect_pipe_changes(struct dc_state *old_state,
+void dcn20_detect_pipe_changes(struct dc_state *old_state,
struct dc_state *new_state,
struct pipe_ctx *old_pipe,
struct pipe_ctx *new_pipe)
@@ -1655,7 +1655,7 @@ static void dcn20_detect_pipe_changes(struct dc_state *old_state,
}
}
-static void dcn20_update_dchubp_dpp(
+void dcn20_update_dchubp_dpp(
struct dc *dc,
struct pipe_ctx *pipe_ctx,
struct dc_state *context)
@@ -1678,25 +1678,41 @@ static void dcn20_update_dchubp_dpp(
* VTG is within DCHUBBUB which is commond block share by each pipe HUBP.
* VTG is 1:1 mapping with OTG. Each pipe HUBP will select which VTG
*/
+
if (pipe_ctx->update_flags.bits.hubp_rq_dlg_ttu) {
hubp->funcs->hubp_vtg_sel(hubp, pipe_ctx->stream_res.tg->inst);
- hubp->funcs->hubp_setup(
- hubp,
- &pipe_ctx->dlg_regs,
- &pipe_ctx->ttu_regs,
- &pipe_ctx->rq_regs,
- &pipe_ctx->pipe_dlg_param);
+ if (hubp->funcs->hubp_setup2) {
+ hubp->funcs->hubp_setup2(
+ hubp,
+ &pipe_ctx->hubp_regs,
+ &pipe_ctx->global_sync,
+ &pipe_ctx->stream->timing);
+ } else {
+ hubp->funcs->hubp_setup(
+ hubp,
+ &pipe_ctx->dlg_regs,
+ &pipe_ctx->ttu_regs,
+ &pipe_ctx->rq_regs,
+ &pipe_ctx->pipe_dlg_param);
+ }
}
if (pipe_ctx->update_flags.bits.unbounded_req && hubp->funcs->set_unbounded_requesting)
hubp->funcs->set_unbounded_requesting(hubp, pipe_ctx->unbounded_req);
- if (pipe_ctx->update_flags.bits.hubp_interdependent)
- hubp->funcs->hubp_setup_interdependent(
- hubp,
- &pipe_ctx->dlg_regs,
- &pipe_ctx->ttu_regs);
+ if (pipe_ctx->update_flags.bits.hubp_interdependent) {
+ if (hubp->funcs->hubp_setup_interdependent2) {
+ hubp->funcs->hubp_setup_interdependent2(
+ hubp,
+ &pipe_ctx->hubp_regs);
+ } else {
+ hubp->funcs->hubp_setup_interdependent(
+ hubp,
+ &pipe_ctx->dlg_regs,
+ &pipe_ctx->ttu_regs);
+ }
+ }
if (pipe_ctx->update_flags.bits.enable ||
pipe_ctx->update_flags.bits.plane_changed ||
@@ -1756,10 +1772,9 @@ static void dcn20_update_dchubp_dpp(
&pipe_ctx->plane_res.scl_data.viewport_c);
viewport_changed = true;
}
- if (hubp->funcs->hubp_program_mcache_id_and_split_coordinate)
- hubp->funcs->hubp_program_mcache_id_and_split_coordinate(
- hubp,
- &pipe_ctx->mcache_regs);
+
+ if (hubp->funcs->hubp_program_mcache_id_and_split_coordinate)
+ hubp->funcs->hubp_program_mcache_id_and_split_coordinate(hubp, &pipe_ctx->mcache_regs);
/* Any updates are handled in dc interface, just need to apply existing for plane enable */
if ((pipe_ctx->update_flags.bits.enable || pipe_ctx->update_flags.bits.opp_changed ||
@@ -1838,7 +1853,7 @@ static void dcn20_update_dchubp_dpp(
hubp->funcs->phantom_hubp_post_enable(hubp);
}
-static int calculate_vready_offset_for_group(struct pipe_ctx *pipe)
+static int dcn20_calculate_vready_offset_for_group(struct pipe_ctx *pipe)
{
struct pipe_ctx *other_pipe;
int vready_offset = pipe->pipe_dlg_param.vready_offset;
@@ -1864,6 +1879,30 @@ static int calculate_vready_offset_for_group(struct pipe_ctx *pipe)
return vready_offset;
}
+static void dcn20_program_tg(
+ struct dc *dc,
+ struct pipe_ctx *pipe_ctx,
+ struct dc_state *context,
+ struct dce_hwseq *hws)
+{
+ pipe_ctx->stream_res.tg->funcs->program_global_sync(
+ pipe_ctx->stream_res.tg,
+ dcn20_calculate_vready_offset_for_group(pipe_ctx),
+ pipe_ctx->pipe_dlg_param.vstartup_start,
+ pipe_ctx->pipe_dlg_param.vupdate_offset,
+ pipe_ctx->pipe_dlg_param.vupdate_width,
+ pipe_ctx->pipe_dlg_param.pstate_keepout);
+
+ if (dc_state_get_pipe_subvp_type(context, pipe_ctx) != SUBVP_PHANTOM)
+ pipe_ctx->stream_res.tg->funcs->wait_for_state(pipe_ctx->stream_res.tg, CRTC_STATE_VACTIVE);
+
+ pipe_ctx->stream_res.tg->funcs->set_vtg_params(
+ pipe_ctx->stream_res.tg, &pipe_ctx->stream->timing, true);
+
+ if (hws->funcs.setup_vupdate_interrupt)
+ hws->funcs.setup_vupdate_interrupt(dc, pipe_ctx);
+}
+
static void dcn20_program_pipe(
struct dc *dc,
struct pipe_ctx *pipe_ctx,
@@ -1874,33 +1913,17 @@ static void dcn20_program_pipe(
/* Only need to unblank on top pipe */
if (resource_is_pipe_type(pipe_ctx, OTG_MASTER)) {
if (pipe_ctx->update_flags.bits.enable ||
- pipe_ctx->update_flags.bits.odm ||
- pipe_ctx->stream->update_flags.bits.abm_level)
+ pipe_ctx->update_flags.bits.odm ||
+ pipe_ctx->stream->update_flags.bits.abm_level)
hws->funcs.blank_pixel_data(dc, pipe_ctx,
- !pipe_ctx->plane_state ||
- !pipe_ctx->plane_state->visible);
+ !pipe_ctx->plane_state ||
+ !pipe_ctx->plane_state->visible);
}
/* Only update TG on top pipe */
if (pipe_ctx->update_flags.bits.global_sync && !pipe_ctx->top_pipe
- && !pipe_ctx->prev_odm_pipe) {
- pipe_ctx->stream_res.tg->funcs->program_global_sync(
- pipe_ctx->stream_res.tg,
- calculate_vready_offset_for_group(pipe_ctx),
- pipe_ctx->pipe_dlg_param.vstartup_start,
- pipe_ctx->pipe_dlg_param.vupdate_offset,
- pipe_ctx->pipe_dlg_param.vupdate_width,
- pipe_ctx->pipe_dlg_param.pstate_keepout);
-
- if (dc_state_get_pipe_subvp_type(context, pipe_ctx) != SUBVP_PHANTOM)
- pipe_ctx->stream_res.tg->funcs->wait_for_state(pipe_ctx->stream_res.tg, CRTC_STATE_VACTIVE);
-
- pipe_ctx->stream_res.tg->funcs->set_vtg_params(
- pipe_ctx->stream_res.tg, &pipe_ctx->stream->timing, true);
-
- if (hws->funcs.setup_vupdate_interrupt)
- hws->funcs.setup_vupdate_interrupt(dc, pipe_ctx);
- }
+ && !pipe_ctx->prev_odm_pipe)
+ dcn20_program_tg(dc, pipe_ctx, context, hws);
if (pipe_ctx->update_flags.bits.odm)
hws->funcs.update_odm(dc, context, pipe_ctx);
@@ -1931,22 +1954,22 @@ static void dcn20_program_pipe(
dcn20_update_dchubp_dpp(dc, pipe_ctx, context);
if (pipe_ctx->plane_state && (pipe_ctx->update_flags.bits.enable ||
- pipe_ctx->plane_state->update_flags.bits.hdr_mult))
+ pipe_ctx->plane_state->update_flags.bits.hdr_mult))
hws->funcs.set_hdr_multiplier(pipe_ctx);
if (hws->funcs.populate_mcm_luts) {
if (pipe_ctx->plane_state) {
hws->funcs.populate_mcm_luts(dc, pipe_ctx, pipe_ctx->plane_state->mcm_luts,
- pipe_ctx->plane_state->lut_bank_a);
+ pipe_ctx->plane_state->lut_bank_a);
pipe_ctx->plane_state->lut_bank_a = !pipe_ctx->plane_state->lut_bank_a;
}
}
if (pipe_ctx->plane_state &&
- (pipe_ctx->plane_state->update_flags.bits.in_transfer_func_change ||
- pipe_ctx->plane_state->update_flags.bits.gamma_change ||
- pipe_ctx->plane_state->update_flags.bits.lut_3d ||
- pipe_ctx->update_flags.bits.enable))
+ (pipe_ctx->plane_state->update_flags.bits.in_transfer_func_change ||
+ pipe_ctx->plane_state->update_flags.bits.gamma_change ||
+ pipe_ctx->plane_state->update_flags.bits.lut_3d ||
+ pipe_ctx->update_flags.bits.enable))
hws->funcs.set_input_transfer_func(dc, pipe_ctx, pipe_ctx->plane_state);
/* dcn10_translate_regamma_to_hw_format takes 750us to finish
@@ -1954,10 +1977,10 @@ static void dcn20_program_pipe(
* updating on slave planes
*/
if (pipe_ctx->update_flags.bits.enable ||
- pipe_ctx->update_flags.bits.plane_changed ||
- pipe_ctx->stream->update_flags.bits.out_tf ||
- (pipe_ctx->plane_state &&
- pipe_ctx->plane_state->update_flags.bits.output_tf_change))
+ pipe_ctx->update_flags.bits.plane_changed ||
+ pipe_ctx->stream->update_flags.bits.out_tf ||
+ (pipe_ctx->plane_state &&
+ pipe_ctx->plane_state->update_flags.bits.output_tf_change))
hws->funcs.set_output_transfer_func(dc, pipe_ctx, pipe_ctx->stream);
/* If the pipe has been enabled or has a different opp, we
@@ -1966,7 +1989,7 @@ static void dcn20_program_pipe(
* causes a different pipe to be chosen to odm combine with.
*/
if (pipe_ctx->update_flags.bits.enable
- || pipe_ctx->update_flags.bits.opp_changed) {
+ || pipe_ctx->update_flags.bits.opp_changed) {
pipe_ctx->stream_res.opp->funcs->opp_set_dyn_expansion(
pipe_ctx->stream_res.opp,
@@ -1996,14 +2019,14 @@ static void dcn20_program_pipe(
memset(&params, 0, sizeof(params));
odm_opp->funcs->opp_program_bit_depth_reduction(odm_opp, &params);
dc->hwss.set_disp_pattern_generator(dc,
- pipe_ctx,
- pipe_ctx->stream_res.test_pattern_params.test_pattern,
- pipe_ctx->stream_res.test_pattern_params.color_space,
- pipe_ctx->stream_res.test_pattern_params.color_depth,
- NULL,
- pipe_ctx->stream_res.test_pattern_params.width,
- pipe_ctx->stream_res.test_pattern_params.height,
- pipe_ctx->stream_res.test_pattern_params.offset);
+ pipe_ctx,
+ pipe_ctx->stream_res.test_pattern_params.test_pattern,
+ pipe_ctx->stream_res.test_pattern_params.color_space,
+ pipe_ctx->stream_res.test_pattern_params.color_depth,
+ NULL,
+ pipe_ctx->stream_res.test_pattern_params.width,
+ pipe_ctx->stream_res.test_pattern_params.height,
+ pipe_ctx->stream_res.test_pattern_params.offset);
}
}
@@ -2012,11 +2035,12 @@ void dcn20_program_front_end_for_ctx(
struct dc_state *context)
{
int i;
- struct dce_hwseq *hws = dc->hwseq;
- DC_LOGGER_INIT(dc->ctx->logger);
unsigned int prev_hubp_count = 0;
unsigned int hubp_count = 0;
- struct pipe_ctx *pipe;
+ struct dce_hwseq *hws = dc->hwseq;
+ struct pipe_ctx *pipe = NULL;
+
+ DC_LOGGER_INIT(dc->ctx->logger);
if (resource_is_pipe_topology_changed(dc->current_state, context))
resource_log_pipe_topology_update(dc, context);
@@ -2029,7 +2053,7 @@ void dcn20_program_front_end_for_ctx(
ASSERT(!pipe->plane_state->triplebuffer_flips);
/*turn off triple buffer for full update*/
dc->hwss.program_triplebuffer(
- dc, pipe, pipe->plane_state->triplebuffer_flips);
+ dc, pipe, pipe->plane_state->triplebuffer_flips);
}
}
}
@@ -2044,30 +2068,31 @@ void dcn20_program_front_end_for_ctx(
if (prev_hubp_count == 0 && hubp_count > 0) {
if (dc->res_pool->hubbub->funcs->force_pstate_change_control)
dc->res_pool->hubbub->funcs->force_pstate_change_control(
- dc->res_pool->hubbub, true, false);
+ dc->res_pool->hubbub, true, false);
udelay(500);
}
/* Set pipe update flags and lock pipes */
for (i = 0; i < dc->res_pool->pipe_count; i++)
dcn20_detect_pipe_changes(dc->current_state, context, &dc->current_state->res_ctx.pipe_ctx[i],
- &context->res_ctx.pipe_ctx[i]);
+ &context->res_ctx.pipe_ctx[i]);
/* When disabling phantom pipes, turn on phantom OTG first (so we can get double
* buffer updates properly)
*/
for (i = 0; i < dc->res_pool->pipe_count; i++) {
struct dc_stream_state *stream = dc->current_state->res_ctx.pipe_ctx[i].stream;
+
pipe = &dc->current_state->res_ctx.pipe_ctx[i];
if (context->res_ctx.pipe_ctx[i].update_flags.bits.disable && stream &&
- dc_state_get_pipe_subvp_type(dc->current_state, pipe) == SUBVP_PHANTOM) {
+ dc_state_get_pipe_subvp_type(dc->current_state, pipe) == SUBVP_PHANTOM) {
struct timing_generator *tg = dc->current_state->res_ctx.pipe_ctx[i].stream_res.tg;
if (tg->funcs->enable_crtc) {
- if (dc->hwseq->funcs.blank_pixel_data) {
+ if (dc->hwseq->funcs.blank_pixel_data)
dc->hwseq->funcs.blank_pixel_data(dc, pipe, true);
- }
+
tg->funcs->enable_crtc(tg);
}
}
@@ -2075,15 +2100,15 @@ void dcn20_program_front_end_for_ctx(
/* OTG blank before disabling all front ends */
for (i = 0; i < dc->res_pool->pipe_count; i++)
if (context->res_ctx.pipe_ctx[i].update_flags.bits.disable
- && !context->res_ctx.pipe_ctx[i].top_pipe
- && !context->res_ctx.pipe_ctx[i].prev_odm_pipe
- && context->res_ctx.pipe_ctx[i].stream)
+ && !context->res_ctx.pipe_ctx[i].top_pipe
+ && !context->res_ctx.pipe_ctx[i].prev_odm_pipe
+ && context->res_ctx.pipe_ctx[i].stream)
hws->funcs.blank_pixel_data(dc, &context->res_ctx.pipe_ctx[i], true);
/* Disconnect mpcc */
for (i = 0; i < dc->res_pool->pipe_count; i++)
if (context->res_ctx.pipe_ctx[i].update_flags.bits.disable
- || context->res_ctx.pipe_ctx[i].update_flags.bits.opp_changed) {
+ || context->res_ctx.pipe_ctx[i].update_flags.bits.opp_changed) {
struct hubbub *hubbub = dc->res_pool->hubbub;
/* Phantom pipe DET should be 0, but if a pipe in use is being transitioned to phantom
@@ -2093,13 +2118,18 @@ void dcn20_program_front_end_for_ctx(
* DET allocation.
*/
if ((context->res_ctx.pipe_ctx[i].update_flags.bits.disable ||
- (context->res_ctx.pipe_ctx[i].plane_state && dc_state_get_pipe_subvp_type(context, &context->res_ctx.pipe_ctx[i]) == SUBVP_PHANTOM))) {
+ (context->res_ctx.pipe_ctx[i].plane_state &&
+ dc_state_get_pipe_subvp_type(context, &context->res_ctx.pipe_ctx[i])
+ == SUBVP_PHANTOM))) {
if (hubbub->funcs->program_det_size)
- hubbub->funcs->program_det_size(hubbub, dc->current_state->res_ctx.pipe_ctx[i].plane_res.hubp->inst, 0);
+ hubbub->funcs->program_det_size(hubbub,
+ dc->current_state->res_ctx.pipe_ctx[i].plane_res.hubp->inst, 0);
if (dc->res_pool->hubbub->funcs->program_det_segments)
- dc->res_pool->hubbub->funcs->program_det_segments(hubbub, dc->current_state->res_ctx.pipe_ctx[i].plane_res.hubp->inst, 0);
+ dc->res_pool->hubbub->funcs->program_det_segments(
+ hubbub, dc->current_state->res_ctx.pipe_ctx[i].plane_res.hubp->inst, 0);
}
- hws->funcs.plane_atomic_disconnect(dc, dc->current_state, &dc->current_state->res_ctx.pipe_ctx[i]);
+ hws->funcs.plane_atomic_disconnect(dc, dc->current_state,
+ &dc->current_state->res_ctx.pipe_ctx[i]);
DC_LOG_DC("Reset mpcc for pipe %d\n", dc->current_state->res_ctx.pipe_ctx[i].pipe_idx);
}
@@ -2107,9 +2137,9 @@ void dcn20_program_front_end_for_ctx(
for (i = 0; i < dc->res_pool->pipe_count; i++) {
pipe = &context->res_ctx.pipe_ctx[i];
if (resource_is_pipe_type(pipe, OTG_MASTER) &&
- !resource_is_pipe_type(pipe, DPP_PIPE) &&
- pipe->update_flags.bits.odm &&
- hws->funcs.update_odm)
+ !resource_is_pipe_type(pipe, DPP_PIPE) &&
+ pipe->update_flags.bits.odm &&
+ hws->funcs.update_odm)
hws->funcs.update_odm(dc, context, pipe);
}
@@ -2127,25 +2157,28 @@ void dcn20_program_front_end_for_ctx(
else {
/* Don't program phantom pipes in the regular front end programming sequence.
* There is an MPO transition case where a pipe being used by a video plane is
- * transitioned directly to be a phantom pipe when closing the MPO video. However
- * the phantom pipe will program a new HUBP_VTG_SEL (update takes place right away),
- * but the MPO still exists until the double buffered update of the main pipe so we
- * will get a frame of underflow if the phantom pipe is programmed here.
+ * transitioned directly to be a phantom pipe when closing the MPO video.
+ * However the phantom pipe will program a new HUBP_VTG_SEL (update takes place
+ * right away) but the MPO still exists until the double buffered update of the
+ * main pipe so we will get a frame of underflow if the phantom pipe is
+ * programmed here.
*/
- if (pipe->stream && dc_state_get_pipe_subvp_type(context, pipe) != SUBVP_PHANTOM)
+ if (pipe->stream &&
+ dc_state_get_pipe_subvp_type(context, pipe) != SUBVP_PHANTOM)
dcn20_program_pipe(dc, pipe, context);
}
pipe = pipe->bottom_pipe;
}
}
+
/* Program secondary blending tree and writeback pipes */
pipe = &context->res_ctx.pipe_ctx[i];
if (!pipe->top_pipe && !pipe->prev_odm_pipe
- && pipe->stream && pipe->stream->num_wb_info > 0
- && (pipe->update_flags.raw || (pipe->plane_state && pipe->plane_state->update_flags.raw)
- || pipe->stream->update_flags.raw)
- && hws->funcs.program_all_writeback_pipes_in_tree)
+ && pipe->stream && pipe->stream->num_wb_info > 0
+ && (pipe->update_flags.raw || (pipe->plane_state && pipe->plane_state->update_flags.raw)
+ || pipe->stream->update_flags.raw)
+ && hws->funcs.program_all_writeback_pipes_in_tree)
hws->funcs.program_all_writeback_pipes_in_tree(dc, pipe->stream, context);
/* Avoid underflow by check of pipe line read when adding 2nd plane. */
@@ -2164,7 +2197,7 @@ void dcn20_program_front_end_for_ctx(
* buffered pending status clear and reset opp head pipe's none double buffered
* registers to their initial state.
*/
-static void post_unlock_reset_opp(struct dc *dc,
+void dcn20_post_unlock_reset_opp(struct dc *dc,
struct pipe_ctx *opp_head)
{
struct display_stream_compressor *dsc = opp_head->stream_res.dsc;
@@ -2201,16 +2234,17 @@ void dcn20_post_unlock_program_front_end(
struct dc *dc,
struct dc_state *context)
{
- int i;
- const unsigned int TIMEOUT_FOR_PIPE_ENABLE_US = 100000;
+ // Timeout for pipe enable
+ unsigned int timeout_us = 100000;
unsigned int polling_interval_us = 1;
struct dce_hwseq *hwseq = dc->hwseq;
+ int i;
for (i = 0; i < dc->res_pool->pipe_count; i++)
if (resource_is_pipe_type(&dc->current_state->res_ctx.pipe_ctx[i], OPP_HEAD) &&
- !resource_is_pipe_type(&context->res_ctx.pipe_ctx[i], OPP_HEAD))
- post_unlock_reset_opp(dc,
- &dc->current_state->res_ctx.pipe_ctx[i]);
+ !resource_is_pipe_type(&context->res_ctx.pipe_ctx[i], OPP_HEAD))
+ dcn20_post_unlock_reset_opp(dc,
+ &dc->current_state->res_ctx.pipe_ctx[i]);
for (i = 0; i < dc->res_pool->pipe_count; i++)
if (context->res_ctx.pipe_ctx[i].update_flags.bits.disable)
@@ -2226,11 +2260,12 @@ void dcn20_post_unlock_program_front_end(
struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
// Don't check flip pending on phantom pipes
if (pipe->plane_state && !pipe->top_pipe && pipe->update_flags.bits.enable &&
- dc_state_get_pipe_subvp_type(context, pipe) != SUBVP_PHANTOM) {
+ dc_state_get_pipe_subvp_type(context, pipe) != SUBVP_PHANTOM) {
struct hubp *hubp = pipe->plane_res.hubp;
int j = 0;
- for (j = 0; j < TIMEOUT_FOR_PIPE_ENABLE_US / polling_interval_us
- && hubp->funcs->hubp_is_flip_pending(hubp); j++)
+
+ for (j = 0; j < timeout_us / polling_interval_us
+ && hubp->funcs->hubp_is_flip_pending(hubp); j++)
udelay(polling_interval_us);
}
}
@@ -2244,15 +2279,14 @@ void dcn20_post_unlock_program_front_end(
* before we've transitioned to 2:1 or 4:1
*/
if (resource_is_pipe_type(old_pipe, OTG_MASTER) && resource_is_pipe_type(pipe, OTG_MASTER) &&
- resource_get_odm_slice_count(old_pipe) < resource_get_odm_slice_count(pipe) &&
- dc_state_get_pipe_subvp_type(context, pipe) != SUBVP_PHANTOM) {
+ resource_get_odm_slice_count(old_pipe) < resource_get_odm_slice_count(pipe) &&
+ dc_state_get_pipe_subvp_type(context, pipe) != SUBVP_PHANTOM) {
int j = 0;
struct timing_generator *tg = pipe->stream_res.tg;
-
if (tg->funcs->get_optc_double_buffer_pending) {
- for (j = 0; j < TIMEOUT_FOR_PIPE_ENABLE_US / polling_interval_us
- && tg->funcs->get_optc_double_buffer_pending(tg); j++)
+ for (j = 0; j < timeout_us / polling_interval_us
+ && tg->funcs->get_optc_double_buffer_pending(tg); j++)
udelay(polling_interval_us);
}
}
@@ -2260,7 +2294,7 @@ void dcn20_post_unlock_program_front_end(
if (dc->res_pool->hubbub->funcs->force_pstate_change_control)
dc->res_pool->hubbub->funcs->force_pstate_change_control(
- dc->res_pool->hubbub, false, false);
+ dc->res_pool->hubbub, false, false);
for (i = 0; i < dc->res_pool->pipe_count; i++) {
struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
@@ -2291,11 +2325,11 @@ void dcn20_post_unlock_program_front_end(
return;
/* P-State support transitions:
- * Natural -> FPO: P-State disabled in prepare, force disallow anytime is safe
- * FPO -> Natural: Unforce anytime after FW disable is safe (P-State will assert naturally)
- * Unsupported -> FPO: P-State enabled in optimize, force disallow anytime is safe
- * FPO -> Unsupported: P-State disabled in prepare, unforce disallow anytime is safe
- * FPO <-> SubVP: Force disallow is maintained on the FPO / SubVP pipes
+ * Natural -> FPO: P-State disabled in prepare, force disallow anytime is safe
+ * FPO -> Natural: Unforce anytime after FW disable is safe (P-State will assert naturally)
+ * Unsupported -> FPO: P-State enabled in optimize, force disallow anytime is safe
+ * FPO -> Unsupported: P-State disabled in prepare, unforce disallow anytime is safe
+ * FPO <-> SubVP: Force disallow is maintained on the FPO / SubVP pipes
*/
if (hwseq->funcs.update_force_pstate)
dc->hwseq->funcs.update_force_pstate(dc, context);
@@ -2310,12 +2344,11 @@ void dcn20_post_unlock_program_front_end(
if (hwseq->wa.DEGVIDCN21)
dc->res_pool->hubbub->funcs->apply_DEDCN21_147_wa(dc->res_pool->hubbub);
-
/* WA for stutter underflow during MPO transitions when adding 2nd plane */
if (hwseq->wa.disallow_self_refresh_during_multi_plane_transition) {
if (dc->current_state->stream_status[0].plane_count == 1 &&
- context->stream_status[0].plane_count > 1) {
+ context->stream_status[0].plane_count > 1) {
struct timing_generator *tg = dc->res_pool->timing_generators[0];
@@ -2463,7 +2496,7 @@ bool dcn20_update_bandwidth(
pipe_ctx->stream_res.tg->funcs->program_global_sync(
pipe_ctx->stream_res.tg,
- calculate_vready_offset_for_group(pipe_ctx),
+ dcn20_calculate_vready_offset_for_group(pipe_ctx),
pipe_ctx->pipe_dlg_param.vstartup_start,
pipe_ctx->pipe_dlg_param.vupdate_offset,
pipe_ctx->pipe_dlg_param.vupdate_width,
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.h b/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.h
index 5c874f7b0683..9d1ad3b29ca5 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.h
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.h
@@ -154,6 +154,21 @@ void dcn20_setup_gsl_group_as_lock(
const struct dc *dc,
struct pipe_ctx *pipe_ctx,
bool enable);
-
+void dcn20_detect_pipe_changes(
+ struct dc_state *old_state,
+ struct dc_state *new_state,
+ struct pipe_ctx *old_pipe,
+ struct pipe_ctx *new_pipe);
+void dcn20_enable_plane(
+ struct dc *dc,
+ struct pipe_ctx *pipe_ctx,
+ struct dc_state *context);
+void dcn20_update_dchubp_dpp(
+ struct dc *dc,
+ struct pipe_ctx *pipe_ctx,
+ struct dc_state *context);
+void dcn20_post_unlock_reset_opp(
+ struct dc *dc,
+ struct pipe_ctx *opp_head);
#endif /* __DC_HWSS_DCN20_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn30/dcn30_init.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn30/dcn30_init.c
index 0e8d32e3dbae..c32764aef884 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dcn30/dcn30_init.c
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn30/dcn30_init.c
@@ -86,7 +86,6 @@ static const struct hw_sequencer_funcs dcn30_funcs = {
.enable_writeback = dcn30_enable_writeback,
.disable_writeback = dcn30_disable_writeback,
.update_writeback = dcn30_update_writeback,
- .mmhubbub_warmup = dcn30_mmhubbub_warmup,
.dmdata_status_done = dcn20_dmdata_status_done,
.program_dmdata_engine = dcn30_program_dmdata_engine,
.set_dmdata_attributes = dcn20_set_dmdata_attributes,
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn301/dcn301_init.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn301/dcn301_init.c
index 780ce4c064aa..dcb27cdbce73 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dcn301/dcn301_init.c
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn301/dcn301_init.c
@@ -86,7 +86,6 @@ static const struct hw_sequencer_funcs dcn301_funcs = {
.enable_writeback = dcn30_enable_writeback,
.disable_writeback = dcn30_disable_writeback,
.update_writeback = dcn30_update_writeback,
- .mmhubbub_warmup = dcn30_mmhubbub_warmup,
.dmdata_status_done = dcn20_dmdata_status_done,
.program_dmdata_engine = dcn30_program_dmdata_engine,
.set_dmdata_attributes = dcn20_set_dmdata_attributes,
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn31/dcn31_init.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn31/dcn31_init.c
index 5f8f45b48720..fb2ffb637931 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dcn31/dcn31_init.c
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn31/dcn31_init.c
@@ -89,7 +89,6 @@ static const struct hw_sequencer_funcs dcn31_funcs = {
.enable_writeback = dcn30_enable_writeback,
.disable_writeback = dcn30_disable_writeback,
.update_writeback = dcn30_update_writeback,
- .mmhubbub_warmup = dcn30_mmhubbub_warmup,
.dmdata_status_done = dcn20_dmdata_status_done,
.program_dmdata_engine = dcn30_program_dmdata_engine,
.set_dmdata_attributes = dcn20_set_dmdata_attributes,
@@ -98,7 +97,7 @@ static const struct hw_sequencer_funcs dcn31_funcs = {
.set_flip_control_gsl = dcn20_set_flip_control_gsl,
.get_vupdate_offset_from_vsync = dcn10_get_vupdate_offset_from_vsync,
.calc_vupdate_position = dcn10_calc_vupdate_position,
- .set_backlight_level = dcn31_set_backlight_level,
+ .set_backlight_level = dcn21_set_backlight_level,
.set_abm_immediate_disable = dcn21_set_abm_immediate_disable,
.set_pipe = dcn21_set_pipe,
.enable_lvds_link_output = dce110_enable_lvds_link_output,
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn314/dcn314_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn314/dcn314_hwseq.c
index 9b88eb72086d..be26c925fdfa 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dcn314/dcn314_hwseq.c
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn314/dcn314_hwseq.c
@@ -162,6 +162,8 @@ void dcn314_update_odm(struct dc *dc, struct dc_state *context, struct pipe_ctx
int opp_inst[MAX_PIPES] = {0};
int odm_slice_width = resource_get_odm_slice_dst_width(pipe_ctx, false);
int last_odm_slice_width = resource_get_odm_slice_dst_width(pipe_ctx, true);
+ struct mpc *mpc = dc->res_pool->mpc;
+ int i;
opp_cnt = get_odm_config(pipe_ctx, opp_inst);
@@ -174,6 +176,16 @@ void dcn314_update_odm(struct dc *dc, struct dc_state *context, struct pipe_ctx
pipe_ctx->stream_res.tg->funcs->set_odm_bypass(
pipe_ctx->stream_res.tg, &pipe_ctx->stream->timing);
+ if (mpc->funcs->set_out_rate_control) {
+ for (i = 0; i < opp_cnt; ++i) {
+ mpc->funcs->set_out_rate_control(
+ mpc, opp_inst[i],
+ false,
+ 0,
+ NULL);
+ }
+ }
+
for (odm_pipe = pipe_ctx->next_odm_pipe; odm_pipe; odm_pipe = odm_pipe->next_odm_pipe) {
odm_pipe->stream_res.opp->funcs->opp_pipe_clock_control(
odm_pipe->stream_res.opp,
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn314/dcn314_init.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn314/dcn314_init.c
index 6bdfbf22ce87..21ef03a76229 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dcn314/dcn314_init.c
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn314/dcn314_init.c
@@ -91,7 +91,6 @@ static const struct hw_sequencer_funcs dcn314_funcs = {
.enable_writeback = dcn30_enable_writeback,
.disable_writeback = dcn30_disable_writeback,
.update_writeback = dcn30_update_writeback,
- .mmhubbub_warmup = dcn30_mmhubbub_warmup,
.dmdata_status_done = dcn20_dmdata_status_done,
.program_dmdata_engine = dcn30_program_dmdata_engine,
.set_dmdata_attributes = dcn20_set_dmdata_attributes,
@@ -100,7 +99,7 @@ static const struct hw_sequencer_funcs dcn314_funcs = {
.set_flip_control_gsl = dcn20_set_flip_control_gsl,
.get_vupdate_offset_from_vsync = dcn10_get_vupdate_offset_from_vsync,
.calc_vupdate_position = dcn10_calc_vupdate_position,
- .set_backlight_level = dcn31_set_backlight_level,
+ .set_backlight_level = dcn21_set_backlight_level,
.set_abm_immediate_disable = dcn21_set_abm_immediate_disable,
.set_pipe = dcn21_set_pipe,
.enable_lvds_link_output = dce110_enable_lvds_link_output,
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn32/dcn32_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn32/dcn32_hwseq.c
index d7f8b2dcaa6b..ee4de9ddfef4 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dcn32/dcn32_hwseq.c
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn32/dcn32_hwseq.c
@@ -985,6 +985,7 @@ void dcn32_init_hw(struct dc *dc)
dc->caps.dmub_caps.subvp_psr = dc->ctx->dmub_srv->dmub->feature_caps.subvp_psr_support;
dc->caps.dmub_caps.gecc_enable = dc->ctx->dmub_srv->dmub->feature_caps.gecc_enable;
dc->caps.dmub_caps.mclk_sw = dc->ctx->dmub_srv->dmub->feature_caps.fw_assisted_mclk_switch_ver;
+ dc->caps.dmub_caps.aux_backlight_support = dc->ctx->dmub_srv->dmub->feature_caps.abm_aux_backlight_support;
/* for DCN401 testing only */
dc->caps.dmub_caps.fams_ver = dc->ctx->dmub_srv->dmub->feature_caps.fw_assisted_mclk_switch_ver;
@@ -1049,7 +1050,8 @@ void dcn32_update_dsc_on_stream(struct pipe_ctx *pipe_ctx, bool enable)
}
/* Enable DSC hw block */
- dsc_cfg.pic_width = (stream->timing.h_addressable + stream->timing.h_border_left + stream->timing.h_border_right) / opp_cnt;
+ dsc_cfg.pic_width = (stream->timing.h_addressable + pipe_ctx->hblank_borrow +
+ stream->timing.h_border_left + stream->timing.h_border_right) / opp_cnt;
dsc_cfg.pic_height = stream->timing.v_addressable + stream->timing.v_border_top + stream->timing.v_border_bottom;
dsc_cfg.pixel_encoding = stream->timing.pixel_encoding;
dsc_cfg.color_depth = stream->timing.display_color_depth;
@@ -1397,12 +1399,12 @@ void dcn32_disable_link_output(struct dc_link *link,
link_hwss->disable_link_output(link, link_res, signal);
link->phy_state.symclk_state = SYMCLK_OFF_TX_OFF;
-
- if (signal == SIGNAL_TYPE_EDP &&
- link->dc->hwss.edp_power_control &&
- !link->skip_implict_edp_power_control)
- link->dc->hwss.edp_power_control(link, false);
- else if (dmcu != NULL && dmcu->funcs->unlock_phy)
+ /*
+ * Add the logic to extract BOTH power up and power down sequences
+ * from enable/disable link output and only call edp panel control
+ * in enable_link_dp and disable_link_dp once.
+ */
+ if (dmcu != NULL && dmcu->funcs->unlock_phy)
dmcu->funcs->unlock_phy(dmcu);
dc->link_srv->dp_trace_source_sequence(link, DPCD_SOURCE_SEQ_AFTER_DISABLE_LINK_PHY);
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn32/dcn32_init.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn32/dcn32_init.c
index 5ecee7e320da..e4d149eff10f 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dcn32/dcn32_init.c
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn32/dcn32_init.c
@@ -87,7 +87,6 @@ static const struct hw_sequencer_funcs dcn32_funcs = {
.enable_writeback = dcn30_enable_writeback,
.disable_writeback = dcn30_disable_writeback,
.update_writeback = dcn30_update_writeback,
- .mmhubbub_warmup = dcn30_mmhubbub_warmup,
.dmdata_status_done = dcn20_dmdata_status_done,
.program_dmdata_engine = dcn30_program_dmdata_engine,
.set_dmdata_attributes = dcn20_set_dmdata_attributes,
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.c
index e599cdc465bf..59fc1c114fbe 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.c
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.c
@@ -426,6 +426,8 @@ void dcn35_update_odm(struct dc *dc, struct dc_state *context, struct pipe_ctx *
int opp_inst[MAX_PIPES] = {0};
int odm_slice_width = resource_get_odm_slice_dst_width(pipe_ctx, false);
int last_odm_slice_width = resource_get_odm_slice_dst_width(pipe_ctx, true);
+ struct mpc *mpc = dc->res_pool->mpc;
+ int i;
opp_cnt = get_odm_config(pipe_ctx, opp_inst);
@@ -438,6 +440,16 @@ void dcn35_update_odm(struct dc *dc, struct dc_state *context, struct pipe_ctx *
pipe_ctx->stream_res.tg->funcs->set_odm_bypass(
pipe_ctx->stream_res.tg, &pipe_ctx->stream->timing);
+ if (mpc->funcs->set_out_rate_control) {
+ for (i = 0; i < opp_cnt; ++i) {
+ mpc->funcs->set_out_rate_control(
+ mpc, opp_inst[i],
+ false,
+ 0,
+ NULL);
+ }
+ }
+
for (odm_pipe = pipe_ctx->next_odm_pipe; odm_pipe; odm_pipe = odm_pipe->next_odm_pipe) {
odm_pipe->stream_res.opp->funcs->opp_pipe_clock_control(
odm_pipe->stream_res.opp,
@@ -1020,8 +1032,13 @@ void dcn35_calc_blocks_to_gate(struct dc *dc, struct dc_state *context,
if (pipe_ctx->plane_res.dpp || pipe_ctx->stream_res.opp)
update_state->pg_pipe_res_update[PG_MPCC][pipe_ctx->plane_res.mpcc_inst] = false;
- if (pipe_ctx->stream_res.dsc)
+ if (pipe_ctx->stream_res.dsc) {
update_state->pg_pipe_res_update[PG_DSC][pipe_ctx->stream_res.dsc->inst] = false;
+ if (dc->caps.sequential_ono) {
+ update_state->pg_pipe_res_update[PG_HUBP][pipe_ctx->stream_res.dsc->inst] = false;
+ update_state->pg_pipe_res_update[PG_DPP][pipe_ctx->stream_res.dsc->inst] = false;
+ }
+ }
if (pipe_ctx->stream_res.opp)
update_state->pg_pipe_res_update[PG_OPP][pipe_ctx->stream_res.opp->inst] = false;
@@ -1579,3 +1596,37 @@ bool dcn35_is_dp_dig_pixel_rate_div_policy(struct pipe_ctx *pipe_ctx)
return false;
}
+
+/*
+ * Set powerup to true for every pipe to match pre-OS configuration.
+ */
+static void dcn35_calc_blocks_to_ungate_for_hw_release(struct dc *dc, struct pg_block_update *update_state)
+{
+ int i = 0, j = 0;
+
+ memset(update_state, 0, sizeof(struct pg_block_update));
+
+ for (i = 0; i < dc->res_pool->pipe_count; i++)
+ for (j = 0; j < PG_HW_PIPE_RESOURCES_NUM_ELEMENT; j++)
+ update_state->pg_pipe_res_update[j][i] = true;
+
+ update_state->pg_res_update[PG_HPO] = true;
+ update_state->pg_res_update[PG_DWB] = true;
+}
+
+/*
+ * The purpose is to power up all gatings to restore optimization to pre-OS env.
+ * Re-use hwss func and existing PG&RCG flags to decide powerup sequence.
+ */
+void dcn35_hardware_release(struct dc *dc)
+{
+ struct pg_block_update pg_update_state;
+
+ dcn35_calc_blocks_to_ungate_for_hw_release(dc, &pg_update_state);
+
+ if (dc->hwss.root_clock_control)
+ dc->hwss.root_clock_control(dc, &pg_update_state, true);
+ /*power up required HW block*/
+ if (dc->hwss.hw_block_power_up)
+ dc->hwss.hw_block_power_up(dc, &pg_update_state);
+}
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.h b/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.h
index e27b3609020f..0b1d6f608edd 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.h
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.h
@@ -99,4 +99,6 @@ void dcn35_set_long_vblank(struct pipe_ctx **pipe_ctx,
bool dcn35_is_dp_dig_pixel_rate_div_policy(struct pipe_ctx *pipe_ctx);
+void dcn35_hardware_release(struct dc *dc);
+
#endif /* __DC_HWSS_DCN35_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_init.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_init.c
index fd67779c27a9..c7acaf97974c 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_init.c
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_init.c
@@ -92,7 +92,6 @@ static const struct hw_sequencer_funcs dcn35_funcs = {
.enable_writeback = dcn30_enable_writeback,
.disable_writeback = dcn30_disable_writeback,
.update_writeback = dcn30_update_writeback,
- .mmhubbub_warmup = dcn30_mmhubbub_warmup,
.dmdata_status_done = dcn20_dmdata_status_done,
.program_dmdata_engine = dcn30_program_dmdata_engine,
.set_dmdata_attributes = dcn20_set_dmdata_attributes,
@@ -123,6 +122,11 @@ static const struct hw_sequencer_funcs dcn35_funcs = {
.root_clock_control = dcn35_root_clock_control,
.set_long_vtotal = dcn35_set_long_vblank,
.calculate_pix_rate_divider = dcn32_calculate_pix_rate_divider,
+ .hardware_release = dcn35_hardware_release,
+ .detect_pipe_changes = dcn20_detect_pipe_changes,
+ .enable_plane = dcn20_enable_plane,
+ .update_dchubp_dpp = dcn20_update_dchubp_dpp,
+ .post_unlock_reset_opp = dcn20_post_unlock_reset_opp,
};
static const struct hwseq_private_funcs dcn35_private_funcs = {
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn351/dcn351_init.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn351/dcn351_init.c
index 3c275a1eff58..4f73e7f551ac 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dcn351/dcn351_init.c
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn351/dcn351_init.c
@@ -91,7 +91,6 @@ static const struct hw_sequencer_funcs dcn351_funcs = {
.enable_writeback = dcn30_enable_writeback,
.disable_writeback = dcn30_disable_writeback,
.update_writeback = dcn30_update_writeback,
- .mmhubbub_warmup = dcn30_mmhubbub_warmup,
.dmdata_status_done = dcn20_dmdata_status_done,
.program_dmdata_engine = dcn30_program_dmdata_engine,
.set_dmdata_attributes = dcn20_set_dmdata_attributes,
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c
index 5de11e2837c0..555a9f590cd7 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c
@@ -3,6 +3,7 @@
// Copyright 2024 Advanced Micro Devices, Inc.
#include "dm_services.h"
+#include "basics/dc_common.h"
#include "dm_helpers.h"
#include "core_types.h"
#include "resource.h"
@@ -126,91 +127,6 @@ void dcn401_program_gamut_remap(struct pipe_ctx *pipe_ctx)
mpc->funcs->set_gamut_remap(mpc, mpcc_id, &mpc_adjust);
}
-struct ips_ono_region_state dcn401_read_ono_state(struct dc *dc, uint8_t region)
-{
- struct dce_hwseq *hws = dc->hwseq;
- struct ips_ono_region_state state = {0, 0};
-
- switch (region) {
- case 0:
- /* dccg, dio, dcio */
- REG_GET_2(DOMAIN22_PG_STATUS,
- DOMAIN_DESIRED_PWR_STATE, &state.desire_pwr_state,
- DOMAIN_PGFSM_PWR_STATUS, &state.current_pwr_state);
- break;
- case 1:
- /* dchubbub, dchvm, dchubbubmem */
- REG_GET_2(DOMAIN23_PG_STATUS,
- DOMAIN_DESIRED_PWR_STATE, &state.desire_pwr_state,
- DOMAIN_PGFSM_PWR_STATUS, &state.current_pwr_state);
- break;
- case 2:
- /* mpc, opp, optc, dwb */
- REG_GET_2(DOMAIN24_PG_STATUS,
- DOMAIN_DESIRED_PWR_STATE, &state.desire_pwr_state,
- DOMAIN_PGFSM_PWR_STATUS, &state.current_pwr_state);
- break;
- case 3:
- /* hpo */
- REG_GET_2(DOMAIN25_PG_STATUS,
- DOMAIN_DESIRED_PWR_STATE, &state.desire_pwr_state,
- DOMAIN_PGFSM_PWR_STATUS, &state.current_pwr_state);
- break;
- case 4:
- /* dchubp0, dpp0 */
- REG_GET_2(DOMAIN0_PG_STATUS,
- DOMAIN_DESIRED_PWR_STATE, &state.desire_pwr_state,
- DOMAIN_PGFSM_PWR_STATUS, &state.current_pwr_state);
- break;
- case 5:
- /* dsc0 */
- REG_GET_2(DOMAIN16_PG_STATUS,
- DOMAIN_DESIRED_PWR_STATE, &state.desire_pwr_state,
- DOMAIN_PGFSM_PWR_STATUS, &state.current_pwr_state);
- break;
- case 6:
- /* dchubp1, dpp1 */
- REG_GET_2(DOMAIN1_PG_STATUS,
- DOMAIN_DESIRED_PWR_STATE, &state.desire_pwr_state,
- DOMAIN_PGFSM_PWR_STATUS, &state.current_pwr_state);
- break;
- case 7:
- /* dsc1 */
- REG_GET_2(DOMAIN17_PG_STATUS,
- DOMAIN_DESIRED_PWR_STATE, &state.desire_pwr_state,
- DOMAIN_PGFSM_PWR_STATUS, &state.current_pwr_state);
- break;
- case 8:
- /* dchubp2, dpp2 */
- REG_GET_2(DOMAIN2_PG_STATUS,
- DOMAIN_DESIRED_PWR_STATE, &state.desire_pwr_state,
- DOMAIN_PGFSM_PWR_STATUS, &state.current_pwr_state);
- break;
- case 9:
- /* dsc2 */
- REG_GET_2(DOMAIN18_PG_STATUS,
- DOMAIN_DESIRED_PWR_STATE, &state.desire_pwr_state,
- DOMAIN_PGFSM_PWR_STATUS, &state.current_pwr_state);
- break;
- case 10:
- /* dchubp3, dpp3 */
- REG_GET_2(DOMAIN3_PG_STATUS,
- DOMAIN_DESIRED_PWR_STATE, &state.desire_pwr_state,
- DOMAIN_PGFSM_PWR_STATUS, &state.current_pwr_state);
- break;
- case 11:
- /* dsc3 */
- REG_GET_2(DOMAIN19_PG_STATUS,
- DOMAIN_DESIRED_PWR_STATE, &state.desire_pwr_state,
- DOMAIN_PGFSM_PWR_STATUS, &state.current_pwr_state);
- break;
- default:
- break;
- }
-
- return state;
-}
-
void dcn401_init_hw(struct dc *dc)
{
struct abm **abms = dc->res_pool->multiple_abms;
@@ -435,7 +351,8 @@ void dcn401_init_hw(struct dc *dc)
dc->caps.dmub_caps.psr = dc->ctx->dmub_srv->dmub->feature_caps.psr;
dc->caps.dmub_caps.mclk_sw = dc->ctx->dmub_srv->dmub->feature_caps.fw_assisted_mclk_switch_ver > 0;
dc->caps.dmub_caps.fams_ver = dc->ctx->dmub_srv->dmub->feature_caps.fw_assisted_mclk_switch_ver;
- dc->debug.fams2_config.bits.enable &= dc->ctx->dmub_srv->dmub->feature_caps.fw_assisted_mclk_switch_ver == 2;
+ dc->debug.fams2_config.bits.enable &=
+ dc->caps.dmub_caps.fams_ver == dc->debug.fams_version.ver; // sw & fw fams versions must match for support
if ((!dc->debug.fams2_config.bits.enable && dc->res_pool->funcs->update_bw_bounding_box)
|| res_pool->ref_clocks.dchub_ref_clock_inKhz / 1000 != current_dchub_ref_freq) {
/* update bounding box if FAMS2 disabled, or if dchub clk has changed */
@@ -820,7 +737,8 @@ enum dc_status dcn401_enable_stream_timing(
int opp_cnt = 1;
int opp_inst[MAX_PIPES] = {0};
struct pipe_ctx *opp_heads[MAX_PIPES] = {0};
- bool manual_mode;
+ struct dc_crtc_timing patched_crtc_timing = stream->timing;
+ bool manual_mode = false;
unsigned int tmds_div = PIXEL_RATE_DIV_NA;
unsigned int unused_div = PIXEL_RATE_DIV_NA;
int odm_slice_width;
@@ -874,16 +792,20 @@ enum dc_status dcn401_enable_stream_timing(
if (dc->hwseq->funcs.PLAT_58856_wa && (!dc_is_dp_signal(stream->signal)))
dc->hwseq->funcs.PLAT_58856_wa(context, pipe_ctx);
+ /* if we are borrowing from hblank, h_addressable needs to be adjusted */
+ if (dc->debug.enable_hblank_borrow)
+ patched_crtc_timing.h_addressable = patched_crtc_timing.h_addressable + pipe_ctx->hblank_borrow;
+
pipe_ctx->stream_res.tg->funcs->program_timing(
- pipe_ctx->stream_res.tg,
- &stream->timing,
- pipe_ctx->pipe_dlg_param.vready_offset,
- pipe_ctx->pipe_dlg_param.vstartup_start,
- pipe_ctx->pipe_dlg_param.vupdate_offset,
- pipe_ctx->pipe_dlg_param.vupdate_width,
- pipe_ctx->pipe_dlg_param.pstate_keepout,
- pipe_ctx->stream->signal,
- true);
+ pipe_ctx->stream_res.tg,
+ &patched_crtc_timing,
+ (unsigned int)pipe_ctx->global_sync.dcn4x.vready_offset_pixels,
+ (unsigned int)pipe_ctx->global_sync.dcn4x.vstartup_lines,
+ (unsigned int)pipe_ctx->global_sync.dcn4x.vupdate_offset_pixels,
+ (unsigned int)pipe_ctx->global_sync.dcn4x.vupdate_vupdate_width_pixels,
+ (unsigned int)pipe_ctx->global_sync.dcn4x.pstate_keepout_start_lines,
+ pipe_ctx->stream->signal,
+ true);
for (i = 0; i < opp_cnt; i++) {
opp_heads[i]->stream_res.opp->funcs->opp_pipe_clock_control(
@@ -2007,3 +1929,730 @@ void dcn401_reset_hw_ctx_wrap(
}
}
}
+
+static unsigned int dcn401_calculate_vready_offset_for_group(struct pipe_ctx *pipe)
+{
+ struct pipe_ctx *other_pipe;
+ unsigned int vready_offset = pipe->global_sync.dcn4x.vready_offset_pixels;
+
+ /* Always use the largest vready_offset of all connected pipes */
+ for (other_pipe = pipe->bottom_pipe; other_pipe != NULL; other_pipe = other_pipe->bottom_pipe) {
+ if (other_pipe->global_sync.dcn4x.vready_offset_pixels > vready_offset)
+ vready_offset = other_pipe->global_sync.dcn4x.vready_offset_pixels;
+ }
+ for (other_pipe = pipe->top_pipe; other_pipe != NULL; other_pipe = other_pipe->top_pipe) {
+ if (other_pipe->global_sync.dcn4x.vready_offset_pixels > vready_offset)
+ vready_offset = other_pipe->global_sync.dcn4x.vready_offset_pixels;
+ }
+ for (other_pipe = pipe->next_odm_pipe; other_pipe != NULL; other_pipe = other_pipe->next_odm_pipe) {
+ if (other_pipe->global_sync.dcn4x.vready_offset_pixels > vready_offset)
+ vready_offset = other_pipe->global_sync.dcn4x.vready_offset_pixels;
+ }
+ for (other_pipe = pipe->prev_odm_pipe; other_pipe != NULL; other_pipe = other_pipe->prev_odm_pipe) {
+ if (other_pipe->global_sync.dcn4x.vready_offset_pixels > vready_offset)
+ vready_offset = other_pipe->global_sync.dcn4x.vready_offset_pixels;
+ }
+
+ return vready_offset;
+}
+
+static void dcn401_program_tg(
+ struct dc *dc,
+ struct pipe_ctx *pipe_ctx,
+ struct dc_state *context,
+ struct dce_hwseq *hws)
+{
+ pipe_ctx->stream_res.tg->funcs->program_global_sync(
+ pipe_ctx->stream_res.tg,
+ dcn401_calculate_vready_offset_for_group(pipe_ctx),
+ (unsigned int)pipe_ctx->global_sync.dcn4x.vstartup_lines,
+ (unsigned int)pipe_ctx->global_sync.dcn4x.vupdate_offset_pixels,
+ (unsigned int)pipe_ctx->global_sync.dcn4x.vupdate_vupdate_width_pixels,
+ (unsigned int)pipe_ctx->global_sync.dcn4x.pstate_keepout_start_lines);
+
+ if (dc_state_get_pipe_subvp_type(context, pipe_ctx) != SUBVP_PHANTOM)
+ pipe_ctx->stream_res.tg->funcs->wait_for_state(pipe_ctx->stream_res.tg, CRTC_STATE_VACTIVE);
+
+ pipe_ctx->stream_res.tg->funcs->set_vtg_params(
+ pipe_ctx->stream_res.tg, &pipe_ctx->stream->timing, true);
+
+ if (hws->funcs.setup_vupdate_interrupt)
+ hws->funcs.setup_vupdate_interrupt(dc, pipe_ctx);
+}
+
+static void dcn401_program_pipe(
+ struct dc *dc,
+ struct pipe_ctx *pipe_ctx,
+ struct dc_state *context)
+{
+ struct dce_hwseq *hws = dc->hwseq;
+
+ /* Only need to unblank on top pipe */
+ if (resource_is_pipe_type(pipe_ctx, OTG_MASTER)) {
+ if (pipe_ctx->update_flags.bits.enable ||
+ pipe_ctx->update_flags.bits.odm ||
+ pipe_ctx->stream->update_flags.bits.abm_level)
+ hws->funcs.blank_pixel_data(dc, pipe_ctx,
+ !pipe_ctx->plane_state ||
+ !pipe_ctx->plane_state->visible);
+ }
+
+ /* Only update TG on top pipe */
+ if (pipe_ctx->update_flags.bits.global_sync && !pipe_ctx->top_pipe
+ && !pipe_ctx->prev_odm_pipe)
+ dcn401_program_tg(dc, pipe_ctx, context, hws);
+
+ if (pipe_ctx->update_flags.bits.odm)
+ hws->funcs.update_odm(dc, context, pipe_ctx);
+
+ if (pipe_ctx->update_flags.bits.enable) {
+ if (hws->funcs.enable_plane)
+ hws->funcs.enable_plane(dc, pipe_ctx, context);
+ else
+ dc->hwss.enable_plane(dc, pipe_ctx, context);
+
+ if (dc->res_pool->hubbub->funcs->force_wm_propagate_to_pipes)
+ dc->res_pool->hubbub->funcs->force_wm_propagate_to_pipes(dc->res_pool->hubbub);
+ }
+
+ if (pipe_ctx->update_flags.bits.det_size) {
+ if (dc->res_pool->hubbub->funcs->program_det_size)
+ dc->res_pool->hubbub->funcs->program_det_size(
+ dc->res_pool->hubbub, pipe_ctx->plane_res.hubp->inst, pipe_ctx->det_buffer_size_kb);
+ if (dc->res_pool->hubbub->funcs->program_det_segments)
+ dc->res_pool->hubbub->funcs->program_det_segments(
+ dc->res_pool->hubbub, pipe_ctx->plane_res.hubp->inst, pipe_ctx->hubp_regs.det_size);
+ }
+
+ if (pipe_ctx->update_flags.raw ||
+ (pipe_ctx->plane_state && pipe_ctx->plane_state->update_flags.raw) ||
+ pipe_ctx->stream->update_flags.raw)
+ dc->hwss.update_dchubp_dpp(dc, pipe_ctx, context);
+
+ if (pipe_ctx->plane_state && (pipe_ctx->update_flags.bits.enable ||
+ pipe_ctx->plane_state->update_flags.bits.hdr_mult))
+ hws->funcs.set_hdr_multiplier(pipe_ctx);
+
+ if (hws->funcs.populate_mcm_luts) {
+ if (pipe_ctx->plane_state) {
+ hws->funcs.populate_mcm_luts(dc, pipe_ctx, pipe_ctx->plane_state->mcm_luts,
+ pipe_ctx->plane_state->lut_bank_a);
+ pipe_ctx->plane_state->lut_bank_a = !pipe_ctx->plane_state->lut_bank_a;
+ }
+ }
+
+ if (pipe_ctx->plane_state &&
+ (pipe_ctx->plane_state->update_flags.bits.in_transfer_func_change ||
+ pipe_ctx->plane_state->update_flags.bits.gamma_change ||
+ pipe_ctx->plane_state->update_flags.bits.lut_3d ||
+ pipe_ctx->update_flags.bits.enable))
+ hws->funcs.set_input_transfer_func(dc, pipe_ctx, pipe_ctx->plane_state);
+
+ /* dcn10_translate_regamma_to_hw_format takes 750us to finish
+ * only do gamma programming for powering on, internal memcmp to avoid
+ * updating on slave planes
+ */
+ if (pipe_ctx->update_flags.bits.enable ||
+ pipe_ctx->update_flags.bits.plane_changed ||
+ pipe_ctx->stream->update_flags.bits.out_tf ||
+ (pipe_ctx->plane_state &&
+ pipe_ctx->plane_state->update_flags.bits.output_tf_change))
+ hws->funcs.set_output_transfer_func(dc, pipe_ctx, pipe_ctx->stream);
+
+ /* If the pipe has been enabled or has a different opp, we
+ * should reprogram the fmt. This deals with cases where
+ * interation between mpc and odm combine on different streams
+ * causes a different pipe to be chosen to odm combine with.
+ */
+ if (pipe_ctx->update_flags.bits.enable
+ || pipe_ctx->update_flags.bits.opp_changed) {
+
+ pipe_ctx->stream_res.opp->funcs->opp_set_dyn_expansion(
+ pipe_ctx->stream_res.opp,
+ COLOR_SPACE_YCBCR601,
+ pipe_ctx->stream->timing.display_color_depth,
+ pipe_ctx->stream->signal);
+
+ pipe_ctx->stream_res.opp->funcs->opp_program_fmt(
+ pipe_ctx->stream_res.opp,
+ &pipe_ctx->stream->bit_depth_params,
+ &pipe_ctx->stream->clamping);
+ }
+
+ /* Set ABM pipe after other pipe configurations done */
+ if ((pipe_ctx->plane_state && pipe_ctx->plane_state->visible)) {
+ if (pipe_ctx->stream_res.abm) {
+ dc->hwss.set_pipe(pipe_ctx);
+ pipe_ctx->stream_res.abm->funcs->set_abm_level(pipe_ctx->stream_res.abm,
+ pipe_ctx->stream->abm_level);
+ }
+ }
+
+ if (pipe_ctx->update_flags.bits.test_pattern_changed) {
+ struct output_pixel_processor *odm_opp = pipe_ctx->stream_res.opp;
+ struct bit_depth_reduction_params params;
+
+ memset(&params, 0, sizeof(params));
+ odm_opp->funcs->opp_program_bit_depth_reduction(odm_opp, &params);
+ dc->hwss.set_disp_pattern_generator(dc,
+ pipe_ctx,
+ pipe_ctx->stream_res.test_pattern_params.test_pattern,
+ pipe_ctx->stream_res.test_pattern_params.color_space,
+ pipe_ctx->stream_res.test_pattern_params.color_depth,
+ NULL,
+ pipe_ctx->stream_res.test_pattern_params.width,
+ pipe_ctx->stream_res.test_pattern_params.height,
+ pipe_ctx->stream_res.test_pattern_params.offset);
+ }
+}
+
+void dcn401_program_front_end_for_ctx(
+ struct dc *dc,
+ struct dc_state *context)
+{
+ int i;
+ unsigned int prev_hubp_count = 0;
+ unsigned int hubp_count = 0;
+ struct dce_hwseq *hws = dc->hwseq;
+ struct pipe_ctx *pipe = NULL;
+
+ DC_LOGGER_INIT(dc->ctx->logger);
+
+ if (resource_is_pipe_topology_changed(dc->current_state, context))
+ resource_log_pipe_topology_update(dc, context);
+
+ if (dc->hwss.program_triplebuffer != NULL && dc->debug.enable_tri_buf) {
+ for (i = 0; i < dc->res_pool->pipe_count; i++) {
+ pipe = &context->res_ctx.pipe_ctx[i];
+
+ if (!pipe->top_pipe && !pipe->prev_odm_pipe && pipe->plane_state) {
+ if (pipe->plane_state->triplebuffer_flips)
+ BREAK_TO_DEBUGGER();
+
+ /*turn off triple buffer for full update*/
+ dc->hwss.program_triplebuffer(
+ dc, pipe, pipe->plane_state->triplebuffer_flips);
+ }
+ }
+ }
+
+ for (i = 0; i < dc->res_pool->pipe_count; i++) {
+ if (dc->current_state->res_ctx.pipe_ctx[i].plane_state)
+ prev_hubp_count++;
+ if (context->res_ctx.pipe_ctx[i].plane_state)
+ hubp_count++;
+ }
+
+ if (prev_hubp_count == 0 && hubp_count > 0) {
+ if (dc->res_pool->hubbub->funcs->force_pstate_change_control)
+ dc->res_pool->hubbub->funcs->force_pstate_change_control(
+ dc->res_pool->hubbub, true, false);
+ udelay(500);
+ }
+
+ /* Set pipe update flags and lock pipes */
+ for (i = 0; i < dc->res_pool->pipe_count; i++)
+ dc->hwss.detect_pipe_changes(dc->current_state, context, &dc->current_state->res_ctx.pipe_ctx[i],
+ &context->res_ctx.pipe_ctx[i]);
+
+ /* When disabling phantom pipes, turn on phantom OTG first (so we can get double
+ * buffer updates properly)
+ */
+ for (i = 0; i < dc->res_pool->pipe_count; i++) {
+ struct dc_stream_state *stream = dc->current_state->res_ctx.pipe_ctx[i].stream;
+
+ pipe = &dc->current_state->res_ctx.pipe_ctx[i];
+
+ if (context->res_ctx.pipe_ctx[i].update_flags.bits.disable && stream &&
+ dc_state_get_pipe_subvp_type(dc->current_state, pipe) == SUBVP_PHANTOM) {
+ struct timing_generator *tg = dc->current_state->res_ctx.pipe_ctx[i].stream_res.tg;
+
+ if (tg->funcs->enable_crtc) {
+ if (dc->hwseq->funcs.blank_pixel_data)
+ dc->hwseq->funcs.blank_pixel_data(dc, pipe, true);
+
+ tg->funcs->enable_crtc(tg);
+ }
+ }
+ }
+ /* OTG blank before disabling all front ends */
+ for (i = 0; i < dc->res_pool->pipe_count; i++)
+ if (context->res_ctx.pipe_ctx[i].update_flags.bits.disable
+ && !context->res_ctx.pipe_ctx[i].top_pipe
+ && !context->res_ctx.pipe_ctx[i].prev_odm_pipe
+ && context->res_ctx.pipe_ctx[i].stream)
+ hws->funcs.blank_pixel_data(dc, &context->res_ctx.pipe_ctx[i], true);
+
+
+ /* Disconnect mpcc */
+ for (i = 0; i < dc->res_pool->pipe_count; i++)
+ if (context->res_ctx.pipe_ctx[i].update_flags.bits.disable
+ || context->res_ctx.pipe_ctx[i].update_flags.bits.opp_changed) {
+ struct hubbub *hubbub = dc->res_pool->hubbub;
+
+ /* Phantom pipe DET should be 0, but if a pipe in use is being transitioned to phantom
+ * then we want to do the programming here (effectively it's being disabled). If we do
+ * the programming later the DET won't be updated until the OTG for the phantom pipe is
+ * turned on (i.e. in an MCLK switch) which can come in too late and cause issues with
+ * DET allocation.
+ */
+ if ((context->res_ctx.pipe_ctx[i].update_flags.bits.disable ||
+ (context->res_ctx.pipe_ctx[i].plane_state &&
+ dc_state_get_pipe_subvp_type(context, &context->res_ctx.pipe_ctx[i]) ==
+ SUBVP_PHANTOM))) {
+ if (hubbub->funcs->program_det_size)
+ hubbub->funcs->program_det_size(hubbub,
+ dc->current_state->res_ctx.pipe_ctx[i].plane_res.hubp->inst, 0);
+ if (dc->res_pool->hubbub->funcs->program_det_segments)
+ dc->res_pool->hubbub->funcs->program_det_segments(
+ hubbub, dc->current_state->res_ctx.pipe_ctx[i].plane_res.hubp->inst, 0);
+ }
+ hws->funcs.plane_atomic_disconnect(dc, dc->current_state,
+ &dc->current_state->res_ctx.pipe_ctx[i]);
+ DC_LOG_DC("Reset mpcc for pipe %d\n", dc->current_state->res_ctx.pipe_ctx[i].pipe_idx);
+ }
+
+ /* update ODM for blanked OTG master pipes */
+ for (i = 0; i < dc->res_pool->pipe_count; i++) {
+ pipe = &context->res_ctx.pipe_ctx[i];
+ if (resource_is_pipe_type(pipe, OTG_MASTER) &&
+ !resource_is_pipe_type(pipe, DPP_PIPE) &&
+ pipe->update_flags.bits.odm &&
+ hws->funcs.update_odm)
+ hws->funcs.update_odm(dc, context, pipe);
+ }
+
+ /*
+ * Program all updated pipes, order matters for mpcc setup. Start with
+ * top pipe and program all pipes that follow in order
+ */
+ for (i = 0; i < dc->res_pool->pipe_count; i++) {
+ pipe = &context->res_ctx.pipe_ctx[i];
+
+ if (pipe->plane_state && !pipe->top_pipe) {
+ while (pipe) {
+ if (hws->funcs.program_pipe)
+ hws->funcs.program_pipe(dc, pipe, context);
+ else {
+ /* Don't program phantom pipes in the regular front end programming sequence.
+ * There is an MPO transition case where a pipe being used by a video plane is
+ * transitioned directly to be a phantom pipe when closing the MPO video.
+ * However the phantom pipe will program a new HUBP_VTG_SEL (update takes place
+ * right away) but the MPO still exists until the double buffered update of the
+ * main pipe so we will get a frame of underflow if the phantom pipe is
+ * programmed here.
+ */
+ if (pipe->stream &&
+ dc_state_get_pipe_subvp_type(context, pipe) != SUBVP_PHANTOM)
+ dcn401_program_pipe(dc, pipe, context);
+ }
+
+ pipe = pipe->bottom_pipe;
+ }
+ }
+
+ /* Program secondary blending tree and writeback pipes */
+ pipe = &context->res_ctx.pipe_ctx[i];
+ if (!pipe->top_pipe && !pipe->prev_odm_pipe
+ && pipe->stream && pipe->stream->num_wb_info > 0
+ && (pipe->update_flags.raw || (pipe->plane_state && pipe->plane_state->update_flags.raw)
+ || pipe->stream->update_flags.raw)
+ && hws->funcs.program_all_writeback_pipes_in_tree)
+ hws->funcs.program_all_writeback_pipes_in_tree(dc, pipe->stream, context);
+
+ /* Avoid underflow by check of pipe line read when adding 2nd plane. */
+ if (hws->wa.wait_hubpret_read_start_during_mpo_transition &&
+ !pipe->top_pipe &&
+ pipe->stream &&
+ pipe->plane_res.hubp->funcs->hubp_wait_pipe_read_start &&
+ dc->current_state->stream_status[0].plane_count == 1 &&
+ context->stream_status[0].plane_count > 1) {
+ pipe->plane_res.hubp->funcs->hubp_wait_pipe_read_start(pipe->plane_res.hubp);
+ }
+ }
+}
+
+void dcn401_post_unlock_program_front_end(
+ struct dc *dc,
+ struct dc_state *context)
+{
+ // Timeout for pipe enable
+ unsigned int timeout_us = 100000;
+ unsigned int polling_interval_us = 1;
+ struct dce_hwseq *hwseq = dc->hwseq;
+ int i;
+
+ DC_LOGGER_INIT(dc->ctx->logger);
+
+ for (i = 0; i < dc->res_pool->pipe_count; i++)
+ if (resource_is_pipe_type(&dc->current_state->res_ctx.pipe_ctx[i], OPP_HEAD) &&
+ !resource_is_pipe_type(&context->res_ctx.pipe_ctx[i], OPP_HEAD))
+ dc->hwss.post_unlock_reset_opp(dc,
+ &dc->current_state->res_ctx.pipe_ctx[i]);
+
+ for (i = 0; i < dc->res_pool->pipe_count; i++)
+ if (context->res_ctx.pipe_ctx[i].update_flags.bits.disable)
+ dc->hwss.disable_plane(dc, dc->current_state, &dc->current_state->res_ctx.pipe_ctx[i]);
+
+ /*
+ * If we are enabling a pipe, we need to wait for pending clear as this is a critical
+ * part of the enable operation otherwise, DM may request an immediate flip which
+ * will cause HW to perform an "immediate enable" (as opposed to "vsync enable") which
+ * is unsupported on DCN.
+ */
+ for (i = 0; i < dc->res_pool->pipe_count; i++) {
+ struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+ // Don't check flip pending on phantom pipes
+ if (pipe->plane_state && !pipe->top_pipe && pipe->update_flags.bits.enable &&
+ dc_state_get_pipe_subvp_type(context, pipe) != SUBVP_PHANTOM) {
+ struct hubp *hubp = pipe->plane_res.hubp;
+ int j = 0;
+
+ for (j = 0; j < timeout_us / polling_interval_us
+ && hubp->funcs->hubp_is_flip_pending(hubp); j++)
+ udelay(polling_interval_us);
+ }
+ }
+
+ for (i = 0; i < dc->res_pool->pipe_count; i++) {
+ struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+ struct pipe_ctx *old_pipe = &dc->current_state->res_ctx.pipe_ctx[i];
+
+ /* When going from a smaller ODM slice count to larger, we must ensure double
+ * buffer update completes before we return to ensure we don't reduce DISPCLK
+ * before we've transitioned to 2:1 or 4:1
+ */
+ if (resource_is_pipe_type(old_pipe, OTG_MASTER) && resource_is_pipe_type(pipe, OTG_MASTER) &&
+ resource_get_odm_slice_count(old_pipe) < resource_get_odm_slice_count(pipe) &&
+ dc_state_get_pipe_subvp_type(context, pipe) != SUBVP_PHANTOM) {
+ int j = 0;
+ struct timing_generator *tg = pipe->stream_res.tg;
+
+ if (tg->funcs->get_optc_double_buffer_pending) {
+ for (j = 0; j < timeout_us / polling_interval_us
+ && tg->funcs->get_optc_double_buffer_pending(tg); j++)
+ udelay(polling_interval_us);
+ }
+ }
+ }
+
+ if (dc->res_pool->hubbub->funcs->force_pstate_change_control)
+ dc->res_pool->hubbub->funcs->force_pstate_change_control(
+ dc->res_pool->hubbub, false, false);
+
+
+ for (i = 0; i < dc->res_pool->pipe_count; i++) {
+ struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+
+ if (pipe->plane_state && !pipe->top_pipe) {
+ /* Program phantom pipe here to prevent a frame of underflow in the MPO transition
+ * case (if a pipe being used for a video plane transitions to a phantom pipe, it
+ * can underflow due to HUBP_VTG_SEL programming if done in the regular front end
+ * programming sequence).
+ */
+ while (pipe) {
+ if (pipe->stream && dc_state_get_pipe_subvp_type(context, pipe) == SUBVP_PHANTOM) {
+ /* When turning on the phantom pipe we want to run through the
+ * entire enable sequence, so apply all the "enable" flags.
+ */
+ if (dc->hwss.apply_update_flags_for_phantom)
+ dc->hwss.apply_update_flags_for_phantom(pipe);
+ if (dc->hwss.update_phantom_vp_position)
+ dc->hwss.update_phantom_vp_position(dc, context, pipe);
+ dcn401_program_pipe(dc, pipe, context);
+ }
+ pipe = pipe->bottom_pipe;
+ }
+ }
+ }
+
+ if (!hwseq)
+ return;
+
+ /* P-State support transitions:
+ * Natural -> FPO: P-State disabled in prepare, force disallow anytime is safe
+ * FPO -> Natural: Unforce anytime after FW disable is safe (P-State will assert naturally)
+ * Unsupported -> FPO: P-State enabled in optimize, force disallow anytime is safe
+ * FPO -> Unsupported: P-State disabled in prepare, unforce disallow anytime is safe
+ * FPO <-> SubVP: Force disallow is maintained on the FPO / SubVP pipes
+ */
+ if (hwseq->funcs.update_force_pstate)
+ dc->hwseq->funcs.update_force_pstate(dc, context);
+
+ /* Only program the MALL registers after all the main and phantom pipes
+ * are done programming.
+ */
+ if (hwseq->funcs.program_mall_pipe_config)
+ hwseq->funcs.program_mall_pipe_config(dc, context);
+
+ /* WA to apply WM setting*/
+ if (hwseq->wa.DEGVIDCN21)
+ dc->res_pool->hubbub->funcs->apply_DEDCN21_147_wa(dc->res_pool->hubbub);
+
+
+ /* WA for stutter underflow during MPO transitions when adding 2nd plane */
+ if (hwseq->wa.disallow_self_refresh_during_multi_plane_transition) {
+
+ if (dc->current_state->stream_status[0].plane_count == 1 &&
+ context->stream_status[0].plane_count > 1) {
+
+ struct timing_generator *tg = dc->res_pool->timing_generators[0];
+
+ dc->res_pool->hubbub->funcs->allow_self_refresh_control(dc->res_pool->hubbub, false);
+
+ hwseq->wa_state.disallow_self_refresh_during_multi_plane_transition_applied = true;
+ hwseq->wa_state.disallow_self_refresh_during_multi_plane_transition_applied_on_frame =
+ tg->funcs->get_frame_count(tg);
+ }
+ }
+}
+
+bool dcn401_update_bandwidth(
+ struct dc *dc,
+ struct dc_state *context)
+{
+ int i;
+ struct dce_hwseq *hws = dc->hwseq;
+
+ /* recalculate DML parameters */
+ if (!dc->res_pool->funcs->validate_bandwidth(dc, context, false))
+ return false;
+
+ /* apply updated bandwidth parameters */
+ dc->hwss.prepare_bandwidth(dc, context);
+
+ /* update hubp configs for all pipes */
+ for (i = 0; i < dc->res_pool->pipe_count; i++) {
+ struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
+
+ if (pipe_ctx->plane_state == NULL)
+ continue;
+
+ if (pipe_ctx->top_pipe == NULL) {
+ bool blank = !is_pipe_tree_visible(pipe_ctx);
+
+ pipe_ctx->stream_res.tg->funcs->program_global_sync(
+ pipe_ctx->stream_res.tg,
+ dcn401_calculate_vready_offset_for_group(pipe_ctx),
+ (unsigned int)pipe_ctx->global_sync.dcn4x.vstartup_lines,
+ (unsigned int)pipe_ctx->global_sync.dcn4x.vupdate_offset_pixels,
+ (unsigned int)pipe_ctx->global_sync.dcn4x.vupdate_vupdate_width_pixels,
+ (unsigned int)pipe_ctx->global_sync.dcn4x.pstate_keepout_start_lines);
+
+ pipe_ctx->stream_res.tg->funcs->set_vtg_params(
+ pipe_ctx->stream_res.tg, &pipe_ctx->stream->timing, false);
+
+ if (pipe_ctx->prev_odm_pipe == NULL)
+ hws->funcs.blank_pixel_data(dc, pipe_ctx, blank);
+
+ if (hws->funcs.setup_vupdate_interrupt)
+ hws->funcs.setup_vupdate_interrupt(dc, pipe_ctx);
+ }
+
+ if (pipe_ctx->plane_res.hubp->funcs->hubp_setup2)
+ pipe_ctx->plane_res.hubp->funcs->hubp_setup2(
+ pipe_ctx->plane_res.hubp,
+ &pipe_ctx->hubp_regs,
+ &pipe_ctx->global_sync,
+ &pipe_ctx->stream->timing);
+ }
+
+ return true;
+}
+
+void dcn401_detect_pipe_changes(struct dc_state *old_state,
+ struct dc_state *new_state,
+ struct pipe_ctx *old_pipe,
+ struct pipe_ctx *new_pipe)
+{
+ bool old_is_phantom = dc_state_get_pipe_subvp_type(old_state, old_pipe) == SUBVP_PHANTOM;
+ bool new_is_phantom = dc_state_get_pipe_subvp_type(new_state, new_pipe) == SUBVP_PHANTOM;
+
+ unsigned int old_pipe_vready_offset_pixels = old_pipe->global_sync.dcn4x.vready_offset_pixels;
+ unsigned int new_pipe_vready_offset_pixels = new_pipe->global_sync.dcn4x.vready_offset_pixels;
+ unsigned int old_pipe_vstartup_lines = old_pipe->global_sync.dcn4x.vstartup_lines;
+ unsigned int new_pipe_vstartup_lines = new_pipe->global_sync.dcn4x.vstartup_lines;
+ unsigned int old_pipe_vupdate_offset_pixels = old_pipe->global_sync.dcn4x.vupdate_offset_pixels;
+ unsigned int new_pipe_vupdate_offset_pixels = new_pipe->global_sync.dcn4x.vupdate_offset_pixels;
+ unsigned int old_pipe_vupdate_width_pixels = old_pipe->global_sync.dcn4x.vupdate_vupdate_width_pixels;
+ unsigned int new_pipe_vupdate_width_pixels = new_pipe->global_sync.dcn4x.vupdate_vupdate_width_pixels;
+
+ new_pipe->update_flags.raw = 0;
+
+ /* If non-phantom pipe is being transitioned to a phantom pipe,
+ * set disable and return immediately. This is because the pipe
+ * that was previously in use must be fully disabled before we
+ * can "enable" it as a phantom pipe (since the OTG will certainly
+ * be different). The post_unlock sequence will set the correct
+ * update flags to enable the phantom pipe.
+ */
+ if (old_pipe->plane_state && !old_is_phantom &&
+ new_pipe->plane_state && new_is_phantom) {
+ new_pipe->update_flags.bits.disable = 1;
+ return;
+ }
+
+ if (resource_is_pipe_type(new_pipe, OTG_MASTER) &&
+ resource_is_odm_topology_changed(new_pipe, old_pipe))
+ /* Detect odm changes */
+ new_pipe->update_flags.bits.odm = 1;
+
+ /* Exit on unchanged, unused pipe */
+ if (!old_pipe->plane_state && !new_pipe->plane_state)
+ return;
+ /* Detect pipe enable/disable */
+ if (!old_pipe->plane_state && new_pipe->plane_state) {
+ new_pipe->update_flags.bits.enable = 1;
+ new_pipe->update_flags.bits.mpcc = 1;
+ new_pipe->update_flags.bits.dppclk = 1;
+ new_pipe->update_flags.bits.hubp_interdependent = 1;
+ new_pipe->update_flags.bits.hubp_rq_dlg_ttu = 1;
+ new_pipe->update_flags.bits.unbounded_req = 1;
+ new_pipe->update_flags.bits.gamut_remap = 1;
+ new_pipe->update_flags.bits.scaler = 1;
+ new_pipe->update_flags.bits.viewport = 1;
+ new_pipe->update_flags.bits.det_size = 1;
+ if (new_pipe->stream->test_pattern.type != DP_TEST_PATTERN_VIDEO_MODE &&
+ new_pipe->stream_res.test_pattern_params.width != 0 &&
+ new_pipe->stream_res.test_pattern_params.height != 0)
+ new_pipe->update_flags.bits.test_pattern_changed = 1;
+ if (!new_pipe->top_pipe && !new_pipe->prev_odm_pipe) {
+ new_pipe->update_flags.bits.odm = 1;
+ new_pipe->update_flags.bits.global_sync = 1;
+ }
+ return;
+ }
+
+ /* For SubVP we need to unconditionally enable because any phantom pipes are
+ * always removed then newly added for every full updates whenever SubVP is in use.
+ * The remove-add sequence of the phantom pipe always results in the pipe
+ * being blanked in enable_stream_timing (DPG).
+ */
+ if (new_pipe->stream && dc_state_get_pipe_subvp_type(new_state, new_pipe) == SUBVP_PHANTOM)
+ new_pipe->update_flags.bits.enable = 1;
+
+ /* Phantom pipes are effectively disabled, if the pipe was previously phantom
+ * we have to enable
+ */
+ if (old_pipe->plane_state && old_is_phantom &&
+ new_pipe->plane_state && !new_is_phantom)
+ new_pipe->update_flags.bits.enable = 1;
+
+ if (old_pipe->plane_state && !new_pipe->plane_state) {
+ new_pipe->update_flags.bits.disable = 1;
+ return;
+ }
+
+ /* Detect plane change */
+ if (old_pipe->plane_state != new_pipe->plane_state)
+ new_pipe->update_flags.bits.plane_changed = true;
+
+ /* Detect top pipe only changes */
+ if (resource_is_pipe_type(new_pipe, OTG_MASTER)) {
+ /* Detect global sync changes */
+ if ((old_pipe_vready_offset_pixels != new_pipe_vready_offset_pixels)
+ || (old_pipe_vstartup_lines != new_pipe_vstartup_lines)
+ || (old_pipe_vupdate_offset_pixels != new_pipe_vupdate_offset_pixels)
+ || (old_pipe_vupdate_width_pixels != new_pipe_vupdate_width_pixels))
+ new_pipe->update_flags.bits.global_sync = 1;
+ }
+
+ if (old_pipe->det_buffer_size_kb != new_pipe->det_buffer_size_kb)
+ new_pipe->update_flags.bits.det_size = 1;
+
+ /*
+ * Detect opp / tg change, only set on change, not on enable
+ * Assume mpcc inst = pipe index, if not this code needs to be updated
+ * since mpcc is what is affected by these. In fact all of our sequence
+ * makes this assumption at the moment with how hubp reset is matched to
+ * same index mpcc reset.
+ */
+ if (old_pipe->stream_res.opp != new_pipe->stream_res.opp)
+ new_pipe->update_flags.bits.opp_changed = 1;
+ if (old_pipe->stream_res.tg != new_pipe->stream_res.tg)
+ new_pipe->update_flags.bits.tg_changed = 1;
+
+ /*
+ * Detect mpcc blending changes, only dpp inst and opp matter here,
+ * mpccs getting removed/inserted update connected ones during their own
+ * programming
+ */
+ if (old_pipe->plane_res.dpp != new_pipe->plane_res.dpp
+ || old_pipe->stream_res.opp != new_pipe->stream_res.opp)
+ new_pipe->update_flags.bits.mpcc = 1;
+
+ /* Detect dppclk change */
+ if (old_pipe->plane_res.bw.dppclk_khz != new_pipe->plane_res.bw.dppclk_khz)
+ new_pipe->update_flags.bits.dppclk = 1;
+
+ /* Check for scl update */
+ if (memcmp(&old_pipe->plane_res.scl_data, &new_pipe->plane_res.scl_data, sizeof(struct scaler_data)))
+ new_pipe->update_flags.bits.scaler = 1;
+ /* Check for vp update */
+ if (memcmp(&old_pipe->plane_res.scl_data.viewport, &new_pipe->plane_res.scl_data.viewport, sizeof(struct rect))
+ || memcmp(&old_pipe->plane_res.scl_data.viewport_c,
+ &new_pipe->plane_res.scl_data.viewport_c, sizeof(struct rect)))
+ new_pipe->update_flags.bits.viewport = 1;
+
+ /* Detect dlg/ttu/rq updates */
+ {
+ struct dml2_display_dlg_regs old_dlg_regs = old_pipe->hubp_regs.dlg_regs;
+ struct dml2_display_ttu_regs old_ttu_regs = old_pipe->hubp_regs.ttu_regs;
+ struct dml2_display_rq_regs old_rq_regs = old_pipe->hubp_regs.rq_regs;
+ struct dml2_display_dlg_regs *new_dlg_regs = &new_pipe->hubp_regs.dlg_regs;
+ struct dml2_display_ttu_regs *new_ttu_regs = &new_pipe->hubp_regs.ttu_regs;
+ struct dml2_display_rq_regs *new_rq_regs = &new_pipe->hubp_regs.rq_regs;
+
+ /* Detect pipe interdependent updates */
+ if ((old_dlg_regs.dst_y_prefetch != new_dlg_regs->dst_y_prefetch)
+ || (old_dlg_regs.vratio_prefetch != new_dlg_regs->vratio_prefetch)
+ || (old_dlg_regs.vratio_prefetch_c != new_dlg_regs->vratio_prefetch_c)
+ || (old_dlg_regs.dst_y_per_vm_vblank != new_dlg_regs->dst_y_per_vm_vblank)
+ || (old_dlg_regs.dst_y_per_row_vblank != new_dlg_regs->dst_y_per_row_vblank)
+ || (old_dlg_regs.dst_y_per_vm_flip != new_dlg_regs->dst_y_per_vm_flip)
+ || (old_dlg_regs.dst_y_per_row_flip != new_dlg_regs->dst_y_per_row_flip)
+ || (old_dlg_regs.refcyc_per_meta_chunk_vblank_l != new_dlg_regs->refcyc_per_meta_chunk_vblank_l)
+ || (old_dlg_regs.refcyc_per_meta_chunk_vblank_c != new_dlg_regs->refcyc_per_meta_chunk_vblank_c)
+ || (old_dlg_regs.refcyc_per_meta_chunk_flip_l != new_dlg_regs->refcyc_per_meta_chunk_flip_l)
+ || (old_dlg_regs.refcyc_per_line_delivery_pre_l != new_dlg_regs->refcyc_per_line_delivery_pre_l)
+ || (old_dlg_regs.refcyc_per_line_delivery_pre_c != new_dlg_regs->refcyc_per_line_delivery_pre_c)
+ || (old_ttu_regs.refcyc_per_req_delivery_pre_l != new_ttu_regs->refcyc_per_req_delivery_pre_l)
+ || (old_ttu_regs.refcyc_per_req_delivery_pre_c != new_ttu_regs->refcyc_per_req_delivery_pre_c)
+ || (old_ttu_regs.refcyc_per_req_delivery_pre_cur0 !=
+ new_ttu_regs->refcyc_per_req_delivery_pre_cur0)
+ || (old_ttu_regs.min_ttu_vblank != new_ttu_regs->min_ttu_vblank)
+ || (old_ttu_regs.qos_level_flip != new_ttu_regs->qos_level_flip)) {
+ old_dlg_regs.dst_y_prefetch = new_dlg_regs->dst_y_prefetch;
+ old_dlg_regs.vratio_prefetch = new_dlg_regs->vratio_prefetch;
+ old_dlg_regs.vratio_prefetch_c = new_dlg_regs->vratio_prefetch_c;
+ old_dlg_regs.dst_y_per_vm_vblank = new_dlg_regs->dst_y_per_vm_vblank;
+ old_dlg_regs.dst_y_per_row_vblank = new_dlg_regs->dst_y_per_row_vblank;
+ old_dlg_regs.dst_y_per_vm_flip = new_dlg_regs->dst_y_per_vm_flip;
+ old_dlg_regs.dst_y_per_row_flip = new_dlg_regs->dst_y_per_row_flip;
+ old_dlg_regs.refcyc_per_meta_chunk_vblank_l = new_dlg_regs->refcyc_per_meta_chunk_vblank_l;
+ old_dlg_regs.refcyc_per_meta_chunk_vblank_c = new_dlg_regs->refcyc_per_meta_chunk_vblank_c;
+ old_dlg_regs.refcyc_per_meta_chunk_flip_l = new_dlg_regs->refcyc_per_meta_chunk_flip_l;
+ old_dlg_regs.refcyc_per_line_delivery_pre_l = new_dlg_regs->refcyc_per_line_delivery_pre_l;
+ old_dlg_regs.refcyc_per_line_delivery_pre_c = new_dlg_regs->refcyc_per_line_delivery_pre_c;
+ old_ttu_regs.refcyc_per_req_delivery_pre_l = new_ttu_regs->refcyc_per_req_delivery_pre_l;
+ old_ttu_regs.refcyc_per_req_delivery_pre_c = new_ttu_regs->refcyc_per_req_delivery_pre_c;
+ old_ttu_regs.refcyc_per_req_delivery_pre_cur0 = new_ttu_regs->refcyc_per_req_delivery_pre_cur0;
+ old_ttu_regs.min_ttu_vblank = new_ttu_regs->min_ttu_vblank;
+ old_ttu_regs.qos_level_flip = new_ttu_regs->qos_level_flip;
+ new_pipe->update_flags.bits.hubp_interdependent = 1;
+ }
+ /* Detect any other updates to ttu/rq/dlg */
+ if (memcmp(&old_dlg_regs, new_dlg_regs, sizeof(old_dlg_regs)) ||
+ memcmp(&old_ttu_regs, new_ttu_regs, sizeof(old_ttu_regs)) ||
+ memcmp(&old_rq_regs, new_rq_regs, sizeof(old_rq_regs)))
+ new_pipe->update_flags.bits.hubp_rq_dlg_ttu = 1;
+ }
+
+ if (old_pipe->unbounded_req != new_pipe->unbounded_req)
+ new_pipe->update_flags.bits.unbounded_req = 1;
+
+ if (memcmp(&old_pipe->stream_res.test_pattern_params,
+ &new_pipe->stream_res.test_pattern_params, sizeof(struct test_pattern_params))) {
+ new_pipe->update_flags.bits.test_pattern_changed = 1;
+ }
+}
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.h b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.h
index 28a513dfc005..17cea748789e 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.h
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.h
@@ -63,8 +63,6 @@ void dcn401_set_cursor_position(struct pipe_ctx *pipe_ctx);
bool dcn401_apply_idle_power_optimizations(struct dc *dc, bool enable);
-struct ips_ono_region_state dcn401_read_ono_state(struct dc *dc,
- uint8_t region);
void dcn401_wait_for_dcc_meta_propagation(const struct dc *dc,
const struct pipe_ctx *top_pipe_to_program);
@@ -96,5 +94,12 @@ void dcn401_reset_hw_ctx_wrap(
struct dc *dc,
struct dc_state *context);
void dcn401_perform_3dlut_wa_unlock(struct pipe_ctx *pipe_ctx);
-
+void dcn401_program_front_end_for_ctx(struct dc *dc, struct dc_state *context);
+void dcn401_post_unlock_program_front_end(struct dc *dc, struct dc_state *context);
+bool dcn401_update_bandwidth(struct dc *dc, struct dc_state *context);
+void dcn401_detect_pipe_changes(
+ struct dc_state *old_state,
+ struct dc_state *new_state,
+ struct pipe_ctx *old_pipe,
+ struct pipe_ctx *new_pipe);
#endif /* __DC_HWSS_DCN401_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_init.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_init.c
index 23e4f208152e..44cb376f97c1 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_init.c
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_init.c
@@ -17,9 +17,9 @@ static const struct hw_sequencer_funcs dcn401_funcs = {
.init_hw = dcn401_init_hw,
.apply_ctx_to_hw = dce110_apply_ctx_to_hw,
.apply_ctx_for_surface = NULL,
- .program_front_end_for_ctx = dcn20_program_front_end_for_ctx,
+ .program_front_end_for_ctx = dcn401_program_front_end_for_ctx,
.wait_for_pending_cleared = dcn10_wait_for_pending_cleared,
- .post_unlock_program_front_end = dcn20_post_unlock_program_front_end,
+ .post_unlock_program_front_end = dcn401_post_unlock_program_front_end,
.update_plane_addr = dcn20_update_plane_addr,
.update_dchub = dcn10_update_dchub,
.update_pending_status = dcn10_update_pending_status,
@@ -42,7 +42,7 @@ static const struct hw_sequencer_funcs dcn401_funcs = {
.cursor_lock = dcn10_cursor_lock,
.prepare_bandwidth = dcn401_prepare_bandwidth,
.optimize_bandwidth = dcn401_optimize_bandwidth,
- .update_bandwidth = dcn20_update_bandwidth,
+ .update_bandwidth = dcn401_update_bandwidth,
.set_drr = dcn10_set_drr,
.get_position = dcn10_get_position,
.set_static_screen_control = dcn31_set_static_screen_control,
@@ -66,7 +66,6 @@ static const struct hw_sequencer_funcs dcn401_funcs = {
.enable_writeback = dcn30_enable_writeback,
.disable_writeback = dcn30_disable_writeback,
.update_writeback = dcn30_update_writeback,
- .mmhubbub_warmup = dcn30_mmhubbub_warmup,
.dmdata_status_done = dcn20_dmdata_status_done,
.program_dmdata_engine = dcn30_program_dmdata_engine,
.set_dmdata_attributes = dcn20_set_dmdata_attributes,
@@ -100,6 +99,10 @@ static const struct hw_sequencer_funcs dcn401_funcs = {
.fams2_global_control_lock_fast = dcn401_fams2_global_control_lock_fast,
.program_outstanding_updates = dcn401_program_outstanding_updates,
.wait_for_all_pending_updates = dcn30_wait_for_all_pending_updates,
+ .detect_pipe_changes = dcn401_detect_pipe_changes,
+ .enable_plane = dcn20_enable_plane,
+ .update_dchubp_dpp = dcn20_update_dchubp_dpp,
+ .post_unlock_reset_opp = dcn20_post_unlock_reset_opp,
};
static const struct hwseq_private_funcs dcn401_private_funcs = {
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/hw_sequencer.h b/drivers/gpu/drm/amd/display/dc/hwss/hw_sequencer.h
index 66fdc5805d0a..a7d66cfd93c9 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/hw_sequencer.h
+++ b/drivers/gpu/drm/amd/display/dc/hwss/hw_sequencer.h
@@ -194,7 +194,6 @@ enum block_sequence_func {
DMUB_SUBVP_SAVE_SURF_ADDR,
HUBP_WAIT_FOR_DCC_META_PROP,
DMUB_FAMS2_GLOBAL_CONTROL_LOCK_FAST,
-
};
struct block_sequence {
@@ -331,10 +330,6 @@ struct hw_sequencer_funcs {
void (*disable_writeback)(struct dc *dc,
unsigned int dwb_pipe_inst);
- bool (*mmhubbub_warmup)(struct dc *dc,
- unsigned int num_dwb,
- struct dc_writeback_info *wb_info);
-
/* Clock Related */
enum dc_status (*set_clock)(struct dc *dc,
enum dc_clock_type clock_type,
@@ -462,6 +457,18 @@ struct hw_sequencer_funcs {
struct dc_state *context);
void (*setup_hpo_hw_control)(const struct dce_hwseq *hws, bool enable);
void (*wait_for_all_pending_updates)(const struct pipe_ctx *pipe_ctx);
+ void (*detect_pipe_changes)(struct dc_state *old_state,
+ struct dc_state *new_state,
+ struct pipe_ctx *old_pipe,
+ struct pipe_ctx *new_pipe);
+ void (*enable_plane)(struct dc *dc,
+ struct pipe_ctx *pipe_ctx,
+ struct dc_state *context);
+ void (*update_dchubp_dpp)(struct dc *dc,
+ struct pipe_ctx *pipe_ctx,
+ struct dc_state *context);
+ void (*post_unlock_reset_opp)(struct dc *dc,
+ struct pipe_ctx *opp_head);
};
void color_space_to_black_color(
@@ -489,11 +496,12 @@ void get_hdr_visual_confirm_color(
void get_mpctree_visual_confirm_color(
struct pipe_ctx *pipe_ctx,
struct tg_color *color);
-
+void get_vabc_visual_confirm_color(
+ struct pipe_ctx *pipe_ctx,
+ struct tg_color *color);
void get_subvp_visual_confirm_color(
struct pipe_ctx *pipe_ctx,
struct tg_color *color);
-
void get_fams2_visual_confirm_color(
struct dc *dc,
struct dc_state *context,
diff --git a/drivers/gpu/drm/amd/display/dc/inc/core_types.h b/drivers/gpu/drm/amd/display/dc/inc/core_types.h
index 8597e866bfe6..d558efc6e12f 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/core_types.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/core_types.h
@@ -45,9 +45,6 @@
#define MAX_SVP_PHANTOM_STREAMS 2
#define MAX_SVP_PHANTOM_PLANES 2
-void enable_surface_flip_reporting(struct dc_plane_state *plane_state,
- uint32_t controller_id);
-
#include "grph_object_id.h"
#include "link_encoder.h"
#include "stream_encoder.h"
@@ -219,6 +216,8 @@ struct resource_funcs {
* Get indicator of power from a context that went through full validation
*/
int (*get_power_profile)(const struct dc_state *context);
+ unsigned int (*get_det_buffer_size)(const struct dc_state *context);
+ unsigned int (*get_vstartup_for_pipe)(struct pipe_ctx *pipe_ctx);
};
struct audio_support{
@@ -467,6 +466,7 @@ struct pipe_ctx {
unsigned int surface_size_in_mall_bytes;
struct dml2_dchub_per_pipe_register_set hubp_regs;
struct dml2_hubp_pipe_mcache_regs mcache_regs;
+ union dml2_global_sync_programming global_sync;
struct dwbc *dwbc;
struct mcif_wb *mcif_wb;
@@ -477,6 +477,8 @@ struct pipe_ctx {
/* subvp_index: only valid if the pipe is a SUBVP_MAIN*/
uint8_t subvp_index;
struct pixel_rate_divider pixel_rate_divider;
+ /* pixels borrowed from hblank to hactive */
+ uint8_t hblank_borrow;
};
/* Data used for dynamic link encoder assignment.
@@ -539,7 +541,8 @@ struct dcn_bw_output {
bool legacy_svp_drr_stream_index_valid;
struct dml2_mcache_surface_allocation mcache_allocations[DML2_MAX_PLANES];
struct dmub_cmd_fams2_global_config fams2_global_config;
- struct dmub_fams2_stream_static_state fams2_stream_params[DML2_MAX_PLANES];
+ union dmub_cmd_fams2_config fams2_stream_base_params[DML2_MAX_PLANES];
+ union dmub_cmd_fams2_config fams2_stream_sub_params[DML2_MAX_PLANES];
struct dml2_display_arb_regs arb_regs;
};
diff --git a/drivers/gpu/drm/amd/display/dc/inc/dcn_calcs.h b/drivers/gpu/drm/amd/display/dc/inc/dcn_calcs.h
index 55529c5f471c..d19a595c2be4 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/dcn_calcs.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/dcn_calcs.h
@@ -624,10 +624,6 @@ bool dcn_validate_bandwidth(
struct dc_state *context,
bool fast_validate);
-unsigned int dcn_find_dcfclk_suits_all(
- const struct dc *dc,
- struct dc_clocks *clocks);
-
void dcn_get_soc_clks(
struct dc *dc,
int *min_fclk_khz,
diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/clk_mgr.h b/drivers/gpu/drm/amd/display/dc/inc/hw/clk_mgr.h
index 2d06067ff36d..c14d64687a3d 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/hw/clk_mgr.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/hw/clk_mgr.h
@@ -306,6 +306,9 @@ struct clk_mgr_funcs {
*/
void (*set_hard_min_memclk)(struct clk_mgr *clk_mgr, bool current_mode);
+ int (*get_hard_min_memclk)(struct clk_mgr *clk_mgr);
+ int (*get_hard_min_fclk)(struct clk_mgr *clk_mgr);
+
/* Send message to PMFW to set hard max memclk frequency to highest DPM */
void (*set_hard_max_memclk)(struct clk_mgr *clk_mgr);
diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/clk_mgr_internal.h b/drivers/gpu/drm/amd/display/dc/inc/hw/clk_mgr_internal.h
index c2dd061892f4..7a1ca1e98059 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/hw/clk_mgr_internal.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/hw/clk_mgr_internal.h
@@ -166,6 +166,41 @@ enum dentist_divider_range {
CLK_SR_DCN32(CLK1_CLK4_CURRENT_CNT), \
CLK_SR_DCN32(CLK4_CLK0_CURRENT_CNT)
+#define CLK_REG_LIST_DCN35() \
+ CLK_SR_DCN35(CLK1_CLK_PLL_REQ), \
+ CLK_SR_DCN35(CLK1_CLK0_DFS_CNTL), \
+ CLK_SR_DCN35(CLK1_CLK1_DFS_CNTL), \
+ CLK_SR_DCN35(CLK1_CLK2_DFS_CNTL), \
+ CLK_SR_DCN35(CLK1_CLK3_DFS_CNTL), \
+ CLK_SR_DCN35(CLK1_CLK4_DFS_CNTL), \
+ CLK_SR_DCN35(CLK1_CLK5_DFS_CNTL), \
+ CLK_SR_DCN35(CLK1_CLK0_CURRENT_CNT), \
+ CLK_SR_DCN35(CLK1_CLK1_CURRENT_CNT), \
+ CLK_SR_DCN35(CLK1_CLK2_CURRENT_CNT), \
+ CLK_SR_DCN35(CLK1_CLK3_CURRENT_CNT), \
+ CLK_SR_DCN35(CLK1_CLK4_CURRENT_CNT), \
+ CLK_SR_DCN35(CLK1_CLK5_CURRENT_CNT), \
+ CLK_SR_DCN35(CLK1_CLK0_BYPASS_CNTL), \
+ CLK_SR_DCN35(CLK1_CLK1_BYPASS_CNTL), \
+ CLK_SR_DCN35(CLK1_CLK2_BYPASS_CNTL), \
+ CLK_SR_DCN35(CLK1_CLK3_BYPASS_CNTL), \
+ CLK_SR_DCN35(CLK1_CLK4_BYPASS_CNTL),\
+ CLK_SR_DCN35(CLK1_CLK5_BYPASS_CNTL), \
+ CLK_SR_DCN35(CLK1_CLK0_DS_CNTL), \
+ CLK_SR_DCN35(CLK1_CLK1_DS_CNTL), \
+ CLK_SR_DCN35(CLK1_CLK2_DS_CNTL), \
+ CLK_SR_DCN35(CLK1_CLK3_DS_CNTL), \
+ CLK_SR_DCN35(CLK1_CLK4_DS_CNTL), \
+ CLK_SR_DCN35(CLK1_CLK5_DS_CNTL), \
+ CLK_SR_DCN35(CLK1_CLK0_ALLOW_DS), \
+ CLK_SR_DCN35(CLK1_CLK1_ALLOW_DS), \
+ CLK_SR_DCN35(CLK1_CLK2_ALLOW_DS), \
+ CLK_SR_DCN35(CLK1_CLK3_ALLOW_DS), \
+ CLK_SR_DCN35(CLK1_CLK4_ALLOW_DS), \
+ CLK_SR_DCN35(CLK1_CLK5_ALLOW_DS), \
+ CLK_SR_DCN35(CLK5_spll_field_8), \
+ SR(DENTIST_DISPCLK_CNTL), \
+
#define CLK_COMMON_MASK_SH_LIST_DCN32(mask_sh) \
CLK_COMMON_MASK_SH_LIST_DCN20_BASE(mask_sh),\
CLK_SF(CLK1_CLK_PLL_REQ, FbMult_int, mask_sh),\
@@ -236,6 +271,7 @@ struct clk_mgr_registers {
uint32_t CLK1_CLK2_DFS_CNTL;
uint32_t CLK1_CLK3_DFS_CNTL;
uint32_t CLK1_CLK4_DFS_CNTL;
+ uint32_t CLK1_CLK5_DFS_CNTL;
uint32_t CLK2_CLK2_DFS_CNTL;
uint32_t CLK1_CLK0_CURRENT_CNT;
@@ -243,11 +279,34 @@ struct clk_mgr_registers {
uint32_t CLK1_CLK2_CURRENT_CNT;
uint32_t CLK1_CLK3_CURRENT_CNT;
uint32_t CLK1_CLK4_CURRENT_CNT;
+ uint32_t CLK1_CLK5_CURRENT_CNT;
uint32_t CLK0_CLK0_DFS_CNTL;
uint32_t CLK0_CLK1_DFS_CNTL;
uint32_t CLK0_CLK3_DFS_CNTL;
uint32_t CLK0_CLK4_DFS_CNTL;
+ uint32_t CLK1_CLK0_BYPASS_CNTL;
+ uint32_t CLK1_CLK1_BYPASS_CNTL;
+ uint32_t CLK1_CLK2_BYPASS_CNTL;
+ uint32_t CLK1_CLK3_BYPASS_CNTL;
+ uint32_t CLK1_CLK4_BYPASS_CNTL;
+ uint32_t CLK1_CLK5_BYPASS_CNTL;
+
+ uint32_t CLK1_CLK0_DS_CNTL;
+ uint32_t CLK1_CLK1_DS_CNTL;
+ uint32_t CLK1_CLK2_DS_CNTL;
+ uint32_t CLK1_CLK3_DS_CNTL;
+ uint32_t CLK1_CLK4_DS_CNTL;
+ uint32_t CLK1_CLK5_DS_CNTL;
+
+ uint32_t CLK1_CLK0_ALLOW_DS;
+ uint32_t CLK1_CLK1_ALLOW_DS;
+ uint32_t CLK1_CLK2_ALLOW_DS;
+ uint32_t CLK1_CLK3_ALLOW_DS;
+ uint32_t CLK1_CLK4_ALLOW_DS;
+ uint32_t CLK1_CLK5_ALLOW_DS;
+ uint32_t CLK5_spll_field_8;
+
};
struct clk_mgr_shift {
diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/hubp.h b/drivers/gpu/drm/amd/display/dc/inc/hw/hubp.h
index 16580d624278..2a530a4a39f7 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/hw/hubp.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/hw/hubp.h
@@ -42,6 +42,7 @@
#include "cursor_reg_cache.h"
#include "dml2/dml21/inc/dml_top_dchub_registers.h"
+#include "dml2/dml21/inc/dml_top_types.h"
#define OPP_ID_INVALID 0xf
#define MAX_TTU 0xffffff
@@ -144,11 +145,21 @@ struct hubp_funcs {
struct _vcs_dpi_display_rq_regs_st *rq_regs,
struct _vcs_dpi_display_pipe_dest_params_st *pipe_dest);
+ void (*hubp_setup2)(
+ struct hubp *hubp,
+ struct dml2_dchub_per_pipe_register_set *pipe_regs,
+ union dml2_global_sync_programming *pipe_global_sync,
+ struct dc_crtc_timing *timing);
+
void (*hubp_setup_interdependent)(
struct hubp *hubp,
struct _vcs_dpi_display_dlg_regs_st *dlg_regs,
struct _vcs_dpi_display_ttu_regs_st *ttu_regs);
+ void (*hubp_setup_interdependent2)(
+ struct hubp *hubp,
+ struct dml2_dchub_per_pipe_register_set *pipe_regs);
+
void (*dcc_control)(struct hubp *hubp, bool enable,
enum hubp_ind_block_size blk_size);
@@ -165,7 +176,7 @@ struct hubp_funcs {
void (*hubp_program_pte_vm)(
struct hubp *hubp,
enum surface_pixel_format format,
- union dc_tiling_info *tiling_info,
+ struct dc_tiling_info *tiling_info,
enum dc_rotation_angle rotation);
void (*hubp_set_vm_system_aperture_settings)(
@@ -179,7 +190,7 @@ struct hubp_funcs {
void (*hubp_program_surface_config)(
struct hubp *hubp,
enum surface_pixel_format format,
- union dc_tiling_info *tiling_info,
+ struct dc_tiling_info *tiling_info,
struct plane_size *plane_size,
enum dc_rotation_angle rotation,
struct dc_plane_dcc_param *dcc,
@@ -275,6 +286,7 @@ struct hubp_funcs {
enum hubp_3dlut_fl_crossbar_bit_slice bit_slice_cb_b,
enum hubp_3dlut_fl_crossbar_bit_slice bit_slice_cr_r);
int (*hubp_get_3dlut_fl_done)(struct hubp *hubp);
+ void (*hubp_clear_tiling)(struct hubp *hubp);
};
#endif
diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/link_encoder.h b/drivers/gpu/drm/amd/display/dc/inc/hw/link_encoder.h
index af9183f5d69b..08c16ba52a51 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/hw/link_encoder.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/hw/link_encoder.h
@@ -168,6 +168,14 @@ struct link_encoder_funcs {
struct link_encoder *enc,
enum encoder_type_select sel,
uint32_t hpo_inst);
+ void (*enable_dpia_output)(struct link_encoder *enc,
+ const struct dc_link_settings *link_settings,
+ uint8_t dpia_id,
+ uint8_t digmode,
+ uint8_t fec_rdy);
+ void (*disable_dpia_output)(struct link_encoder *link_enc,
+ uint8_t dpia_id,
+ uint8_t digmode);
};
/*
diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/mem_input.h b/drivers/gpu/drm/amd/display/dc/inc/hw/mem_input.h
index a8b44f398ce6..42fbc70f7056 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/hw/mem_input.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/hw/mem_input.h
@@ -150,7 +150,7 @@ struct mem_input_funcs {
void (*mem_input_program_pte_vm)(
struct mem_input *mem_input,
enum surface_pixel_format format,
- union dc_tiling_info *tiling_info,
+ struct dc_tiling_info *tiling_info,
enum dc_rotation_angle rotation);
void (*mem_input_set_vm_system_aperture_settings)(
@@ -164,7 +164,7 @@ struct mem_input_funcs {
void (*mem_input_program_surface_config)(
struct mem_input *mem_input,
enum surface_pixel_format format,
- union dc_tiling_info *tiling_info,
+ struct dc_tiling_info *tiling_info,
struct plane_size *plane_size,
enum dc_rotation_angle rotation,
struct dc_plane_dcc_param *dcc,
@@ -187,6 +187,8 @@ struct mem_input_funcs {
const struct dc_cursor_position *pos,
const struct dc_cursor_mi_param *param);
+ void (*mem_input_clear_tiling)(
+ struct mem_input *mem_input);
};
#endif
diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/optc.h b/drivers/gpu/drm/amd/display/dc/inc/hw/optc.h
index 03cbcbb36f1c..6fdc9809280c 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/hw/optc.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/hw/optc.h
@@ -210,7 +210,7 @@ void optc1_enable_crtc_reset(struct timing_generator *optc,
bool optc1_configure_crc(struct timing_generator *optc, const struct crc_params *params);
-bool optc1_get_crc(struct timing_generator *optc,
+bool optc1_get_crc(struct timing_generator *optc, uint8_t idx,
uint32_t *r_cr,
uint32_t *g_y,
uint32_t *b_cb);
diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h b/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h
index b74e18cc1e66..9885cb3c310f 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h
@@ -141,6 +141,9 @@ struct crc_params {
bool continuous_mode;
bool enable;
+
+ uint8_t crc_eng_inst;
+ bool reset;
};
/**
@@ -291,7 +294,7 @@ struct timing_generator_funcs {
* @get_crc: Get CRCs for the given timing generator. Return false if
* CRCs are not enabled (via configure_crc).
*/
- bool (*get_crc)(struct timing_generator *tg,
+ bool (*get_crc)(struct timing_generator *tg, uint8_t idx,
uint32_t *r_cr, uint32_t *g_y, uint32_t *b_cb);
void (*program_manual_trigger)(struct timing_generator *optc);
diff --git a/drivers/gpu/drm/amd/display/dc/inc/link.h b/drivers/gpu/drm/amd/display/dc/inc/link.h
index f04292086c08..fd1f9d3db039 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/link.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/link.h
@@ -148,6 +148,10 @@ struct link_service {
const struct dc_stream_state *stream,
const unsigned int num_streams);
+ uint32_t (*dp_required_hblank_size_bytes)(
+ const struct dc_link *link,
+ struct dp_audio_bandwidth_params *audio_params);
+
/*************************** DPMS *************************************/
void (*set_dpms_on)(struct dc_state *state, struct pipe_ctx *pipe_ctx);
diff --git a/drivers/gpu/drm/amd/display/dc/irq/dcn201/irq_service_dcn201.c b/drivers/gpu/drm/amd/display/dc/irq/dcn201/irq_service_dcn201.c
index 4fb9cd6708d5..1d61d475d36f 100644
--- a/drivers/gpu/drm/amd/display/dc/irq/dcn201/irq_service_dcn201.c
+++ b/drivers/gpu/drm/amd/display/dc/irq/dcn201/irq_service_dcn201.c
@@ -30,8 +30,8 @@
#include "../dce110/irq_service_dce110.h"
#include "irq_service_dcn201.h"
-#include "dcn/dcn_2_0_3_offset.h"
-#include "dcn/dcn_2_0_3_sh_mask.h"
+#include "dcn/dcn_2_0_1_offset.h"
+#include "dcn/dcn_2_0_1_sh_mask.h"
#include "cyan_skillfish_ip_offset.h"
#include "soc15_hw_ip.h"
diff --git a/drivers/gpu/drm/amd/display/dc/link/accessories/link_dp_cts.c b/drivers/gpu/drm/amd/display/dc/link/accessories/link_dp_cts.c
index ff8fe1a94965..96febabf464a 100644
--- a/drivers/gpu/drm/amd/display/dc/link/accessories/link_dp_cts.c
+++ b/drivers/gpu/drm/amd/display/dc/link/accessories/link_dp_cts.c
@@ -251,7 +251,7 @@ static void dp_test_send_phy_test_pattern(struct dc_link *link)
link_training_settings.lttpr_mode = dp_decide_lttpr_mode(link, &link->cur_link_settings);
- if ((link->chip_caps & EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN) &&
+ if (((link->chip_caps & AMD_EXT_DISPLAY_PATH_CAPS__EXT_CHIP_MASK) == AMD_EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN) &&
link_training_settings.lttpr_mode == LTTPR_MODE_TRANSPARENT)
dp_fixed_vs_pe_read_lane_adjust(
link,
@@ -646,7 +646,7 @@ bool dp_set_test_pattern(
if (IS_DP_PHY_PATTERN(test_pattern)) {
/* Set DPCD Lane Settings before running test pattern */
if (p_link_settings != NULL) {
- if ((link->chip_caps & EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN) &&
+ if (((link->chip_caps & AMD_EXT_DISPLAY_PATH_CAPS__EXT_CHIP_MASK) == AMD_EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN) &&
p_link_settings->lttpr_mode == LTTPR_MODE_TRANSPARENT) {
dp_fixed_vs_pe_set_retimer_lane_settings(
link,
diff --git a/drivers/gpu/drm/amd/display/dc/link/hwss/link_hwss_dio.c b/drivers/gpu/drm/amd/display/dc/link/hwss/link_hwss_dio.c
index 3e47a6735912..06faa461067b 100644
--- a/drivers/gpu/drm/amd/display/dc/link/hwss/link_hwss_dio.c
+++ b/drivers/gpu/drm/amd/display/dc/link/hwss/link_hwss_dio.c
@@ -164,7 +164,9 @@ void disable_dio_link_output(struct dc_link *link,
{
struct link_encoder *link_enc = link_enc_cfg_get_link_enc(link);
- link_enc->funcs->disable_output(link_enc, signal);
+ if (link_enc != NULL)
+ link_enc->funcs->disable_output(link_enc, signal);
+
link->dc->link_srv->dp_trace_source_sequence(link,
DPCD_SOURCE_SEQ_AFTER_DISABLE_LINK_PHY);
}
diff --git a/drivers/gpu/drm/amd/display/dc/link/hwss/link_hwss_dio_fixed_vs_pe_retimer.c b/drivers/gpu/drm/amd/display/dc/link/hwss/link_hwss_dio_fixed_vs_pe_retimer.c
index 348ea4cb832d..a6d1d7641ab4 100644
--- a/drivers/gpu/drm/amd/display/dc/link/hwss/link_hwss_dio_fixed_vs_pe_retimer.c
+++ b/drivers/gpu/drm/amd/display/dc/link/hwss/link_hwss_dio_fixed_vs_pe_retimer.c
@@ -187,7 +187,7 @@ static const struct link_hwss dio_fixed_vs_pe_retimer_link_hwss = {
bool requires_fixed_vs_pe_retimer_dio_link_hwss(const struct dc_link *link)
{
- return (link->chip_caps & EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN);
+ return ((link->chip_caps & AMD_EXT_DISPLAY_PATH_CAPS__EXT_CHIP_MASK) == AMD_EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN);
}
const struct link_hwss *get_dio_fixed_vs_pe_retimer_link_hwss(void)
diff --git a/drivers/gpu/drm/amd/display/dc/link/hwss/link_hwss_dpia.c b/drivers/gpu/drm/amd/display/dc/link/hwss/link_hwss_dpia.c
index 6499807af72a..36adf95744fe 100644
--- a/drivers/gpu/drm/amd/display/dc/link/hwss/link_hwss_dpia.c
+++ b/drivers/gpu/drm/amd/display/dc/link/hwss/link_hwss_dpia.c
@@ -77,17 +77,74 @@ static void set_dio_dpia_lane_settings(struct dc_link *link,
{
}
+static void enable_dpia_link_output(struct dc_link *link,
+ const struct link_resource *link_res,
+ enum signal_type signal,
+ enum clock_source_id clock_source,
+ const struct dc_link_settings *link_settings)
+{
+ struct link_encoder *link_enc = link_enc_cfg_get_link_enc(link);
+
+ if (link_enc != NULL) {
+ if (link->dc->config.enable_dpia_pre_training && link_enc->funcs->enable_dpia_output) {
+ uint8_t fec_rdy = link->dc->link_srv->dp_should_enable_fec(link);
+ uint8_t digmode = dc_is_dp_sst_signal(signal) ? DIG_SST_MODE : DIG_MST_MODE;
+
+ link_enc->funcs->enable_dpia_output(
+ link_enc,
+ link_settings,
+ link->ddc_hw_inst,
+ digmode,
+ fec_rdy);
+ } else {
+ if (dc_is_dp_sst_signal(signal))
+ link_enc->funcs->enable_dp_output(
+ link_enc,
+ link_settings,
+ clock_source);
+ else
+ link_enc->funcs->enable_dp_mst_output(
+ link_enc,
+ link_settings,
+ clock_source);
+ }
+
+ }
+
+ link->dc->link_srv->dp_trace_source_sequence(link,
+ DPCD_SOURCE_SEQ_AFTER_ENABLE_LINK_PHY);
+}
+
+static void disable_dpia_link_output(struct dc_link *link,
+ const struct link_resource *link_res,
+ enum signal_type signal)
+{
+ struct link_encoder *link_enc = link_enc_cfg_get_link_enc(link);
+
+ if (link_enc != NULL) {
+ if (link->dc->config.enable_dpia_pre_training && link_enc->funcs->disable_dpia_output) {
+ uint8_t digmode = dc_is_dp_sst_signal(signal) ? DIG_SST_MODE : DIG_MST_MODE;
+
+ link_enc->funcs->disable_dpia_output(link_enc, link->ddc_hw_inst, digmode);
+ } else
+ link_enc->funcs->disable_output(link_enc, signal);
+ }
+
+ link->dc->link_srv->dp_trace_source_sequence(link,
+ DPCD_SOURCE_SEQ_AFTER_DISABLE_LINK_PHY);
+}
+
static const struct link_hwss dpia_link_hwss = {
.setup_stream_encoder = setup_dio_stream_encoder,
.reset_stream_encoder = reset_dio_stream_encoder,
.setup_stream_attribute = setup_dio_stream_attribute,
- .disable_link_output = disable_dio_link_output,
+ .disable_link_output = disable_dpia_link_output,
.setup_audio_output = setup_dio_audio_output,
.enable_audio_packet = enable_dio_audio_packet,
.disable_audio_packet = disable_dio_audio_packet,
.ext = {
.set_throttled_vcp_size = set_dio_throttled_vcp_size,
- .enable_dp_link_output = enable_dio_dp_link_output,
+ .enable_dp_link_output = enable_dpia_link_output,
.set_dp_link_test_pattern = set_dio_dpia_link_test_pattern,
.set_dp_lane_settings = set_dio_dpia_lane_settings,
.update_stream_allocation_table = update_dpia_stream_allocation_table,
diff --git a/drivers/gpu/drm/amd/display/dc/link/hwss/link_hwss_dpia.h b/drivers/gpu/drm/amd/display/dc/link/hwss/link_hwss_dpia.h
index ad16ec5d9bb7..259e0f4775e1 100644
--- a/drivers/gpu/drm/amd/display/dc/link/hwss/link_hwss_dpia.h
+++ b/drivers/gpu/drm/amd/display/dc/link/hwss/link_hwss_dpia.h
@@ -27,6 +27,9 @@
#include "link_hwss.h"
+#define DIG_SST_MODE 0
+#define DIG_MST_MODE 5
+
const struct link_hwss *get_dpia_link_hwss(void);
bool can_use_dpia_link_hwss(const struct dc_link *link,
const struct link_resource *link_res);
diff --git a/drivers/gpu/drm/amd/display/dc/link/link_detection.c b/drivers/gpu/drm/amd/display/dc/link/link_detection.c
index e026c728042a..550e1a098fa2 100644
--- a/drivers/gpu/drm/amd/display/dc/link/link_detection.c
+++ b/drivers/gpu/drm/amd/display/dc/link/link_detection.c
@@ -829,7 +829,8 @@ static bool should_verify_link_capability_destructively(struct dc_link *link,
if (link->dc->debug.skip_detection_link_training ||
dc_is_embedded_signal(link->local_sink->sink_signal) ||
- link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA) {
+ (link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA &&
+ !link->dc->config.enable_dpia_pre_training)) {
destrictive = false;
} else if (link_dp_get_encoding_format(&max_link_cap) ==
DP_8b_10b_ENCODING) {
diff --git a/drivers/gpu/drm/amd/display/dc/link/link_dpms.c b/drivers/gpu/drm/amd/display/dc/link/link_dpms.c
index 41cab9ad6885..ec7de9c01fab 100644
--- a/drivers/gpu/drm/amd/display/dc/link/link_dpms.c
+++ b/drivers/gpu/drm/amd/display/dc/link/link_dpms.c
@@ -772,6 +772,20 @@ static bool dp_set_dsc_on_rx(struct pipe_ctx *pipe_ctx, bool enable)
return result;
}
+static bool dp_set_hblank_reduction_on_rx(struct pipe_ctx *pipe_ctx)
+{
+ struct dc *dc = pipe_ctx->stream->ctx->dc;
+ struct dc_stream_state *stream = pipe_ctx->stream;
+ bool result = false;
+
+ if (dc_is_virtual_signal(stream->signal))
+ result = true;
+ else
+ result = dm_helpers_dp_write_hblank_reduction(dc->ctx, stream);
+ return result;
+}
+
+
/* The stream with these settings can be sent (unblanked) only after DSC was enabled on RX first,
* i.e. after dp_enable_dsc_on_rx() had been called
*/
@@ -808,7 +822,8 @@ void link_set_dsc_on_stream(struct pipe_ctx *pipe_ctx, bool enable)
enum optc_dsc_mode optc_dsc_mode;
/* Enable DSC hw block */
- dsc_cfg.pic_width = (stream->timing.h_addressable + stream->timing.h_border_left + stream->timing.h_border_right) / opp_cnt;
+ dsc_cfg.pic_width = (stream->timing.h_addressable + pipe_ctx->hblank_borrow +
+ stream->timing.h_border_left + stream->timing.h_border_right) / opp_cnt;
dsc_cfg.pic_height = stream->timing.v_addressable + stream->timing.v_border_top + stream->timing.v_border_bottom;
dsc_cfg.pixel_encoding = stream->timing.pixel_encoding;
dsc_cfg.color_depth = stream->timing.display_color_depth;
@@ -1952,11 +1967,15 @@ static void enable_link_hdmi(struct pipe_ctx *pipe_ctx)
stream->phy_pix_clk = stream->timing.pix_clk_100hz / 10;
if (stream->phy_pix_clk > 340000)
is_over_340mhz = true;
+ if (dc_is_tmds_signal(stream->signal) && stream->phy_pix_clk > 6000000UL) {
+ ASSERT(false);
+ return;
+ }
if (dc_is_hdmi_signal(pipe_ctx->stream->signal)) {
unsigned short masked_chip_caps = pipe_ctx->stream->link->chip_caps &
- EXT_DISPLAY_PATH_CAPS__EXT_CHIP_MASK;
- if (masked_chip_caps == EXT_DISPLAY_PATH_CAPS__HDMI20_TISN65DP159RSBT) {
+ AMD_EXT_DISPLAY_PATH_CAPS__EXT_CHIP_MASK;
+ if (masked_chip_caps == AMD_EXT_DISPLAY_PATH_CAPS__HDMI20_TISN65DP159RSBT) {
/* DP159, Retimer settings */
eng_id = pipe_ctx->stream_res.stream_enc->id;
@@ -1967,7 +1986,7 @@ static void enable_link_hdmi(struct pipe_ctx *pipe_ctx)
write_i2c_default_retimer_setting(pipe_ctx,
is_vga_mode, is_over_340mhz);
}
- } else if (masked_chip_caps == EXT_DISPLAY_PATH_CAPS__HDMI20_PI3EQX1204) {
+ } else if (masked_chip_caps == AMD_EXT_DISPLAY_PATH_CAPS__HDMI20_PI3EQX1204) {
/* PI3EQX1204, Redriver settings */
write_i2c_redriver_setting(pipe_ctx, is_over_340mhz);
}
@@ -2023,7 +2042,7 @@ static enum dc_status enable_link_dp(struct dc_state *state,
int lt_attempts = LINK_TRAINING_ATTEMPTS;
// Increase retry count if attempting DP1.x on FIXED_VS link
- if ((link->chip_caps & EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN) &&
+ if (((link->chip_caps & AMD_EXT_DISPLAY_PATH_CAPS__EXT_CHIP_MASK) == AMD_EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN) &&
link_dp_get_encoding_format(link_settings) == DP_8b_10b_ENCODING)
lt_attempts = 10;
@@ -2038,7 +2057,8 @@ static enum dc_status enable_link_dp(struct dc_state *state,
/* Train with fallback when enabling DPIA link. Conventional links are
* trained with fallback during sink detection.
*/
- if (link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA)
+ if (link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA &&
+ !link->dc->config.enable_dpia_pre_training)
do_fallback = true;
/*
@@ -2374,13 +2394,13 @@ void link_set_dpms_off(struct pipe_ctx *pipe_ctx)
enum engine_id eng_id = pipe_ctx->stream_res.stream_enc->id;
unsigned short masked_chip_caps = link->chip_caps &
- EXT_DISPLAY_PATH_CAPS__EXT_CHIP_MASK;
+ AMD_EXT_DISPLAY_PATH_CAPS__EXT_CHIP_MASK;
//Need to inform that sink is going to use legacy HDMI mode.
write_scdc_data(
link->ddc,
165000,//vbios only handles 165Mhz.
false);
- if (masked_chip_caps == EXT_DISPLAY_PATH_CAPS__HDMI20_TISN65DP159RSBT) {
+ if (masked_chip_caps == AMD_EXT_DISPLAY_PATH_CAPS__HDMI20_TISN65DP159RSBT) {
/* DP159, Retimer settings */
if (get_ext_hdmi_settings(pipe_ctx, eng_id, &settings))
write_i2c_retimer_setting(pipe_ctx,
@@ -2388,7 +2408,7 @@ void link_set_dpms_off(struct pipe_ctx *pipe_ctx)
else
write_i2c_default_retimer_setting(pipe_ctx,
false, false);
- } else if (masked_chip_caps == EXT_DISPLAY_PATH_CAPS__HDMI20_PI3EQX1204) {
+ } else if (masked_chip_caps == AMD_EXT_DISPLAY_PATH_CAPS__HDMI20_PI3EQX1204) {
/* PI3EQX1204, Redriver settings */
write_i2c_redriver_setting(pipe_ctx, false);
}
@@ -2528,6 +2548,15 @@ void link_set_dpms_on(
if (pipe_ctx->stream->dpms_off)
return;
+ /* For Dp tunneling link, a pending HPD means that we have a race condition between processing
+ * current link and processing the pending HPD. If we enable the link now, we may end up with a
+ * link that is not actually connected to a sink. So we skip enabling the link in this case.
+ */
+ if (link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA && link->is_hpd_pending) {
+ DC_LOG_DEBUG("%s, Link%d HPD is pending, not enable it.\n", __func__, link->link_index);
+ return;
+ }
+
/* Have to setup DSC before DIG FE and BE are connected (which happens before the
* link training). This is to make sure the bandwidth sent to DIG BE won't be
* bigger than what the link and/or DIG BE can handle. VBID[6]/CompressedStream_flag
@@ -2593,6 +2622,9 @@ void link_set_dpms_on(
}
}
+ if (dc_is_dp_signal(pipe_ctx->stream->signal))
+ dp_set_hblank_reduction_on_rx(pipe_ctx);
+
if (pipe_ctx->stream->link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA)
allocate_usb4_bandwidth(pipe_ctx->stream);
diff --git a/drivers/gpu/drm/amd/display/dc/link/link_factory.c b/drivers/gpu/drm/amd/display/dc/link/link_factory.c
index 5e1b5ab9fbc6..a7877d57a00f 100644
--- a/drivers/gpu/drm/amd/display/dc/link/link_factory.c
+++ b/drivers/gpu/drm/amd/display/dc/link/link_factory.c
@@ -101,6 +101,7 @@ static void construct_link_service_validation(struct link_service *link_srv)
link_srv->validate_mode_timing = link_validate_mode_timing;
link_srv->dp_link_bandwidth_kbps = dp_link_bandwidth_kbps;
link_srv->validate_dpia_bandwidth = link_validate_dpia_bandwidth;
+ link_srv->dp_required_hblank_size_bytes = dp_required_hblank_size_bytes;
}
/* link dpms owns the programming sequence of stream's dpms state associated
@@ -698,7 +699,7 @@ static bool construct_phy(struct dc_link *link,
link->chip_caps);
}
- if (link->chip_caps & EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN) {
+ if ((link->chip_caps & AMD_EXT_DISPLAY_PATH_CAPS__EXT_CHIP_MASK) == AMD_EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN) {
link->bios_forced_drive_settings.VOLTAGE_SWING =
(bios->integrated_info->ext_disp_conn_info.fixdpvoltageswing & 0x3);
link->bios_forced_drive_settings.PRE_EMPHASIS =
diff --git a/drivers/gpu/drm/amd/display/dc/link/link_validation.c b/drivers/gpu/drm/amd/display/dc/link/link_validation.c
index 60f15a9ba7a5..29606fda029d 100644
--- a/drivers/gpu/drm/amd/display/dc/link/link_validation.c
+++ b/drivers/gpu/drm/amd/display/dc/link/link_validation.c
@@ -409,3 +409,182 @@ bool link_validate_dpia_bandwidth(const struct dc_stream_state *stream, const un
return dpia_validate_usb4_bw(dpia_link, bw_needed, num_dpias);
}
+
+struct dp_audio_layout_config {
+ uint8_t layouts_per_sample_denom;
+ uint8_t symbols_per_layout;
+ uint8_t max_layouts_per_audio_sdp;
+};
+
+static void get_audio_layout_config(
+ uint32_t channel_count,
+ enum dp_link_encoding encoding,
+ struct dp_audio_layout_config *output)
+{
+ memset(output, 0, sizeof(struct dp_audio_layout_config));
+
+ /* Assuming L-PCM audio. Current implementation uses max 1 layout per SDP,
+ * with each layout being the same size (8ch layout).
+ */
+ if (encoding == DP_8b_10b_ENCODING) {
+ if (channel_count == 2) {
+ output->layouts_per_sample_denom = 4;
+ output->symbols_per_layout = 40;
+ output->max_layouts_per_audio_sdp = 1;
+ } else if (channel_count == 8 || channel_count == 6) {
+ output->layouts_per_sample_denom = 1;
+ output->symbols_per_layout = 40;
+ output->max_layouts_per_audio_sdp = 1;
+ }
+ } else if (encoding == DP_128b_132b_ENCODING) {
+ if (channel_count == 2) {
+ output->layouts_per_sample_denom = 4;
+ output->symbols_per_layout = 10;
+ output->max_layouts_per_audio_sdp = 1;
+ } else if (channel_count == 8 || channel_count == 6) {
+ output->layouts_per_sample_denom = 1;
+ output->symbols_per_layout = 10;
+ output->max_layouts_per_audio_sdp = 1;
+ }
+ }
+}
+
+static uint32_t get_av_stream_map_lane_count(
+ enum dp_link_encoding encoding,
+ enum dc_lane_count lane_count,
+ bool is_mst)
+{
+ uint32_t av_stream_map_lane_count = 0;
+
+ if (encoding == DP_8b_10b_ENCODING) {
+ if (!is_mst)
+ av_stream_map_lane_count = lane_count;
+ else
+ av_stream_map_lane_count = 4;
+ } else if (encoding == DP_128b_132b_ENCODING) {
+ av_stream_map_lane_count = 4;
+ }
+
+ ASSERT(av_stream_map_lane_count != 0);
+
+ return av_stream_map_lane_count;
+}
+
+static uint32_t get_audio_sdp_overhead(
+ enum dp_link_encoding encoding,
+ enum dc_lane_count lane_count,
+ bool is_mst)
+{
+ uint32_t audio_sdp_overhead = 0;
+
+ if (encoding == DP_8b_10b_ENCODING) {
+ if (is_mst)
+ audio_sdp_overhead = 16; /* 4 * 2 + 8 */
+ else
+ audio_sdp_overhead = lane_count * 2 + 8;
+ } else if (encoding == DP_128b_132b_ENCODING) {
+ audio_sdp_overhead = 10; /* 4 x 2.5 */
+ }
+
+ ASSERT(audio_sdp_overhead != 0);
+
+ return audio_sdp_overhead;
+}
+
+/* Current calculation only applicable for 8b/10b MST and 128b/132b SST/MST.
+ */
+static uint32_t calculate_overhead_hblank_bw_in_symbols(
+ uint32_t max_slice_h)
+{
+ uint32_t overhead_hblank_bw = 0; /* in stream symbols */
+
+ overhead_hblank_bw += max_slice_h * 4; /* EOC overhead */
+ overhead_hblank_bw += 12; /* Main link overhead (VBID, BS/BE) */
+
+ return overhead_hblank_bw;
+}
+
+uint32_t dp_required_hblank_size_bytes(
+ const struct dc_link *link,
+ struct dp_audio_bandwidth_params *audio_params)
+{
+ /* Main logic from dce_audio is duplicated here, with the main
+ * difference being:
+ * - Pre-determined lane count of 4
+ * - Assumed 16 dsc slices for worst case
+ * - Assumed SDP split disabled for worst case
+ * TODO: Unify logic from dce_audio to prevent duplicated logic.
+ */
+
+ const struct dc_crtc_timing *timing = audio_params->crtc_timing;
+ const uint32_t channel_count = audio_params->channel_count;
+ const uint32_t sample_rate_hz = audio_params->sample_rate_hz;
+ const enum dp_link_encoding link_encoding = audio_params->link_encoding;
+
+ // 8b/10b MST and 128b/132b are always 4 logical lanes.
+ const uint32_t lane_count = 4;
+ const bool is_mst = (link->connector_signal == SIGNAL_TYPE_DISPLAY_PORT);
+ // Maximum slice count is with ODM 4:1, 4 slices per DSC
+ const uint32_t max_slices_h = 16;
+
+ const uint32_t av_stream_map_lane_count = get_av_stream_map_lane_count(
+ link_encoding, lane_count, is_mst);
+ const uint32_t audio_sdp_overhead = get_audio_sdp_overhead(
+ link_encoding, lane_count, is_mst);
+ struct dp_audio_layout_config layout_config;
+
+ if (link_encoding == DP_8b_10b_ENCODING && link->connector_signal == SIGNAL_TYPE_DISPLAY_PORT)
+ return 0;
+
+ get_audio_layout_config(
+ channel_count, link_encoding, &layout_config);
+
+ /* DP spec recommends between 1.05 to 1.1 safety margin to prevent sample under-run */
+ struct fixed31_32 audio_sdp_margin = dc_fixpt_from_fraction(110, 100);
+ struct fixed31_32 horizontal_line_freq_khz = dc_fixpt_from_fraction(
+ timing->pix_clk_100hz, (long long)timing->h_total * 10);
+ struct fixed31_32 samples_per_line;
+ struct fixed31_32 layouts_per_line;
+ struct fixed31_32 symbols_per_sdp_max_layout;
+ struct fixed31_32 remainder;
+ uint32_t num_sdp_with_max_layouts;
+ uint32_t required_symbols_per_hblank;
+ uint32_t required_bytes_per_hblank = 0;
+
+ samples_per_line = dc_fixpt_from_fraction(sample_rate_hz, 1000);
+ samples_per_line = dc_fixpt_div(samples_per_line, horizontal_line_freq_khz);
+ layouts_per_line = dc_fixpt_div_int(samples_per_line, layout_config.layouts_per_sample_denom);
+ // HBlank expansion usage assumes SDP split disabled to allow for worst case.
+ layouts_per_line = dc_fixpt_from_int(dc_fixpt_ceil(layouts_per_line));
+
+ num_sdp_with_max_layouts = dc_fixpt_floor(
+ dc_fixpt_div_int(layouts_per_line, layout_config.max_layouts_per_audio_sdp));
+ symbols_per_sdp_max_layout = dc_fixpt_from_int(
+ layout_config.max_layouts_per_audio_sdp * layout_config.symbols_per_layout);
+ symbols_per_sdp_max_layout = dc_fixpt_add_int(symbols_per_sdp_max_layout, audio_sdp_overhead);
+ symbols_per_sdp_max_layout = dc_fixpt_mul(symbols_per_sdp_max_layout, audio_sdp_margin);
+ required_symbols_per_hblank = num_sdp_with_max_layouts;
+ required_symbols_per_hblank *= ((dc_fixpt_ceil(symbols_per_sdp_max_layout) + av_stream_map_lane_count) /
+ av_stream_map_lane_count) * av_stream_map_lane_count;
+
+ if (num_sdp_with_max_layouts != dc_fixpt_ceil(
+ dc_fixpt_div_int(layouts_per_line, layout_config.max_layouts_per_audio_sdp))) {
+ remainder = dc_fixpt_sub_int(layouts_per_line,
+ num_sdp_with_max_layouts * layout_config.max_layouts_per_audio_sdp);
+ remainder = dc_fixpt_mul_int(remainder, layout_config.symbols_per_layout);
+ remainder = dc_fixpt_add_int(remainder, audio_sdp_overhead);
+ remainder = dc_fixpt_mul(remainder, audio_sdp_margin);
+ required_symbols_per_hblank += ((dc_fixpt_ceil(remainder) + av_stream_map_lane_count) /
+ av_stream_map_lane_count) * av_stream_map_lane_count;
+ }
+
+ required_symbols_per_hblank += calculate_overhead_hblank_bw_in_symbols(max_slices_h);
+
+ if (link_encoding == DP_8b_10b_ENCODING)
+ required_bytes_per_hblank = required_symbols_per_hblank; // 8 bits per 8b/10b symbol
+ else if (link_encoding == DP_128b_132b_ENCODING)
+ required_bytes_per_hblank = required_symbols_per_hblank * 4; // 32 bits per 128b/132b symbol
+
+ return required_bytes_per_hblank;
+}
+
diff --git a/drivers/gpu/drm/amd/display/dc/link/link_validation.h b/drivers/gpu/drm/amd/display/dc/link/link_validation.h
index 595fb05946e9..bf398c49c3e8 100644
--- a/drivers/gpu/drm/amd/display/dc/link/link_validation.h
+++ b/drivers/gpu/drm/amd/display/dc/link/link_validation.h
@@ -37,4 +37,9 @@ uint32_t dp_link_bandwidth_kbps(
const struct dc_link *link,
const struct dc_link_settings *link_settings);
+
+uint32_t dp_required_hblank_size_bytes(
+ const struct dc_link *link,
+ struct dp_audio_bandwidth_params *audio_params);
+
#endif /* __LINK_VALIDATION_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_ddc.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_ddc.c
index d6d5bbf2108c..267180e7bc48 100644
--- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_ddc.c
+++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_ddc.c
@@ -505,7 +505,7 @@ bool try_to_configure_aux_timeout(struct ddc_service *ddc,
bool result = false;
struct ddc *ddc_pin = ddc->ddc_pin;
- if ((ddc->link->chip_caps & EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN) &&
+ if (((ddc->link->chip_caps & AMD_EXT_DISPLAY_PATH_CAPS__EXT_CHIP_MASK) == AMD_EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN) &&
!ddc->link->dc->debug.disable_fixed_vs_aux_timeout_wa &&
ddc->ctx->dce_version == DCN_VERSION_3_1) {
/* Fixed VS workaround for AUX timeout */
diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c
index 9dabaf682171..44c3023a7731 100644
--- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c
+++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c
@@ -1554,7 +1554,7 @@ enum dc_status dp_retrieve_lttpr_cap(struct dc_link *link)
/* If this chip cap is set, at least one retimer must exist in the chain
* Override count to 1 if we receive a known bad count (0 or an invalid value) */
- if ((link->chip_caps & EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN) &&
+ if (((link->chip_caps & AMD_EXT_DISPLAY_PATH_CAPS__EXT_CHIP_MASK) == AMD_EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN) &&
(dp_parse_lttpr_repeater_count(link->dpcd_caps.lttpr_caps.phy_repeater_cnt) == 0)) {
/* If you see this message consistently, either the host platform has FIXED_VS flag
* incorrectly configured or the sink device is returning an invalid count.
@@ -1632,13 +1632,6 @@ static bool retrieve_link_cap(struct dc_link *link)
sizeof(link->dpcd_caps.lttpr_caps.phy_repeater_cnt));
}
- /* Read DP tunneling information. */
- if (link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA) {
- status = dpcd_get_tunneling_device_data(link);
- if (status != DC_OK)
- dm_error("%s: Read tunneling device data failed.\n", __func__);
- }
-
dpcd_set_source_specific_data(link);
/* Sink may need to configure internals based on vendor, so allow some
* time before proceeding with possibly vendor specific transactions
@@ -1711,7 +1704,7 @@ static bool retrieve_link_cap(struct dc_link *link)
link->dpcd_caps.dprx_feature.raw = dpcd_dprx_data;
if (status != DC_OK)
- dm_error("%s: Read DPRX caps data failed.\n", __func__);
+ dm_error("%s: Read DPRX feature list failed.\n", __func__);
/* AdaptiveSyncCapability */
dpcd_dprx_data = 0;
@@ -1726,15 +1719,13 @@ static bool retrieve_link_cap(struct dc_link *link)
link->dpcd_caps.adaptive_sync_caps.dp_adap_sync_caps.raw = dpcd_dprx_data;
if (status != DC_OK)
- dm_error("%s: Read DPRX caps data failed. Addr:%#x\n",
+ dm_error("%s: Read DPRX feature list_1 failed. Addr:%#x\n",
__func__, DP_DPRX_FEATURE_ENUMERATION_LIST_CONT_1);
}
-
else {
link->dpcd_caps.dprx_feature.raw = 0;
}
-
/* Error condition checking...
* It is impossible for Sink to report Max Lane Count = 0.
* It is possible for Sink to report Max Link Rate = 0, if it is
@@ -1788,6 +1779,11 @@ static bool retrieve_link_cap(struct dc_link *link)
link->test_pattern_enabled = false;
link->compliance_test_state.raw = 0;
+ link->dpcd_caps.receive_port0_cap.raw[0] =
+ dpcd_data[DP_RECEIVE_PORT_0_CAP_0 - DP_DPCD_REV];
+ link->dpcd_caps.receive_port0_cap.raw[1] =
+ dpcd_data[DP_RECEIVE_PORT_0_BUFFER_SIZE - DP_DPCD_REV];
+
/* read sink count */
core_link_read_dpcd(link,
DP_SINK_COUNT,
@@ -1918,6 +1914,7 @@ static bool retrieve_link_cap(struct dc_link *link)
if (link->dpcd_caps.channel_coding_cap.bits.DP_128b_132b_SUPPORTED) {
DC_LOG_DP2("128b/132b encoding is supported at link %d", link->link_index);
+ /* Read 128b/132b suppoerted link rates */
core_link_read_dpcd(link,
DP_128B132B_SUPPORTED_LINK_RATES,
&link->dpcd_caps.dp_128b_132b_supported_link_rates.raw,
@@ -1965,6 +1962,13 @@ static bool retrieve_link_cap(struct dc_link *link)
link->dpcd_caps.max_uncompressed_pixel_rate_cap.raw,
sizeof(link->dpcd_caps.max_uncompressed_pixel_rate_cap.raw));
+ /* Read DP tunneling information. */
+ if (link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA) {
+ status = dpcd_get_tunneling_device_data(link);
+ if (status != DC_OK)
+ dm_error("%s: Read DP tunneling device data failed.\n", __func__);
+ }
+
retrieve_cable_id(link);
dpcd_write_cable_id_to_dprx(link);
@@ -2308,6 +2312,14 @@ bool dp_verify_link_cap_with_retries(
} else {
link->verified_link_cap = last_verified_link_cap;
}
+
+ /* For Dp tunneling link, a pending HPD means that we have a race condition between processing
+ * current link and processing the pending HPD. Since the training is failed, we should just brak
+ * the loop so that we have chance to process the pending HPD.
+ */
+ if (link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA && link->is_hpd_pending)
+ break;
+
fsleep(10 * 1000);
}
diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_irq_handler.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_irq_handler.c
index 48abeaa88678..a08403c022ea 100644
--- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_irq_handler.c
+++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_irq_handler.c
@@ -226,6 +226,8 @@ static void handle_hpd_irq_replay_sink(struct dc_link *link)
replay_configuration.bits.STATE_TRANSITION_ERROR_STATUS) {
bool allow_active;
+ link->replay_settings.config.replay_error_status.raw |= replay_error_status.raw;
+
if (link->replay_settings.config.force_disable_desync_error_check)
return;
@@ -237,6 +239,9 @@ static void handle_hpd_irq_replay_sink(struct dc_link *link)
&replay_configuration.raw,
sizeof(replay_configuration.raw));
+ /* Update desync error counter */
+ link->replay_settings.replay_desync_error_fail_count++;
+
/* Acknowledge and clear error bits */
dm_helpers_dp_write_dpcd(
link->ctx,
@@ -408,7 +413,8 @@ bool dp_handle_hpd_rx_irq(struct dc_link *link,
if (hpd_irq_dpcd_data.bytes.device_service_irq.bits.AUTOMATED_TEST) {
// Workaround for DP 1.4a LL Compliance CTS as USB4 has to share encoders unlike DP and USBC
- if (link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA)
+ if (link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA &&
+ !link->dc->config.enable_dpia_pre_training)
link->skip_fallback_on_link_loss = true;
device_service_clear.bits.AUTOMATED_TEST = 1;
diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_phy.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_phy.c
index bafa52a0165a..2c73ac87cd66 100644
--- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_phy.c
+++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_phy.c
@@ -104,7 +104,7 @@ void dp_set_hw_lane_settings(
// Don't return here if using FIXED_VS link HWSS and encoding is 128b/132b
if ((link_settings->lttpr_mode == LTTPR_MODE_NON_TRANSPARENT) &&
!is_immediate_downstream(link, offset) &&
- (!(link->chip_caps & EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN) ||
+ (!((link->chip_caps & AMD_EXT_DISPLAY_PATH_CAPS__EXT_CHIP_MASK) == AMD_EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN) ||
link_dp_get_encoding_format(&link_settings->link_settings) == DP_8b_10b_ENCODING))
return;
diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training.c
index 754c895e1bfb..88d4288cde0f 100644
--- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training.c
+++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training.c
@@ -739,7 +739,7 @@ void override_training_settings(
if (overrides->ffe_preset != NULL)
lt_settings->ffe_preset = overrides->ffe_preset;
/* Override HW lane settings with BIOS forced values if present */
- if ((link->chip_caps & EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN) &&
+ if ((link->chip_caps & AMD_EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN) &&
lt_settings->lttpr_mode == LTTPR_MODE_TRANSPARENT) {
lt_settings->voltage_swing = &link->bios_forced_drive_settings.VOLTAGE_SWING;
lt_settings->pre_emphasis = &link->bios_forced_drive_settings.PRE_EMPHASIS;
@@ -1574,7 +1574,7 @@ enum link_training_result dp_perform_link_training(
* Per DP specs starting from here, DPTX device shall not issue
* Non-LT AUX transactions inside training mode.
*/
- if ((link->chip_caps & EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN) && encoding == DP_8b_10b_ENCODING)
+ if (((link->chip_caps & AMD_EXT_DISPLAY_PATH_CAPS__EXT_CHIP_MASK) == AMD_EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN) && encoding == DP_8b_10b_ENCODING)
status = dp_perform_fixed_vs_pe_training_sequence(link, link_res, &lt_settings);
else if (encoding == DP_8b_10b_ENCODING)
status = dp_perform_8b_10b_link_training(link, link_res, &lt_settings);
diff --git a/drivers/gpu/drm/amd/display/dc/mpc/dcn30/dcn30_mpc.c b/drivers/gpu/drm/amd/display/dc/mpc/dcn30/dcn30_mpc.c
index fe26fde12eeb..85298b8a1b5e 100644
--- a/drivers/gpu/drm/amd/display/dc/mpc/dcn30/dcn30_mpc.c
+++ b/drivers/gpu/drm/amd/display/dc/mpc/dcn30/dcn30_mpc.c
@@ -110,6 +110,23 @@ void mpc3_disable_dwb_mux(
MPC_DWB0_MUX, 0xf);
}
+void mpc3_set_out_rate_control(
+ struct mpc *mpc,
+ int opp_id,
+ bool enable,
+ bool rate_2x_mode,
+ struct mpc_dwb_flow_control *flow_control)
+{
+ struct dcn30_mpc *mpc30 = TO_DCN30_MPC(mpc);
+
+ /* Always disable mpc out rate and flow control.
+ * MPC flow rate control is not needed for DCN30 and above.
+ */
+ REG_UPDATE_2(MUX[opp_id],
+ MPC_OUT_RATE_CONTROL_DISABLE, 1,
+ MPC_OUT_RATE_CONTROL, 0);
+}
+
enum dc_lut_mode mpc3_get_ogam_current(struct mpc *mpc, int mpcc_id)
{
/*Contrary to DCN2 and DCN1 wherein a single status register field holds this info;
@@ -1519,6 +1536,7 @@ static const struct mpc_funcs dcn30_mpc_funcs = {
.set_dwb_mux = mpc3_set_dwb_mux,
.disable_dwb_mux = mpc3_disable_dwb_mux,
.is_dwb_idle = mpc3_is_dwb_idle,
+ .set_out_rate_control = mpc3_set_out_rate_control,
.set_gamut_remap = mpc3_set_gamut_remap,
.program_shaper = mpc3_program_shaper,
.acquire_rmu = mpcc3_acquire_rmu,
diff --git a/drivers/gpu/drm/amd/display/dc/mpc/dcn30/dcn30_mpc.h b/drivers/gpu/drm/amd/display/dc/mpc/dcn30/dcn30_mpc.h
index ce93003dae01..103f29900a2c 100644
--- a/drivers/gpu/drm/amd/display/dc/mpc/dcn30/dcn30_mpc.h
+++ b/drivers/gpu/drm/amd/display/dc/mpc/dcn30/dcn30_mpc.h
@@ -1085,6 +1085,13 @@ bool mpc3_is_dwb_idle(
struct mpc *mpc,
int dwb_id);
+void mpc3_set_out_rate_control(
+ struct mpc *mpc,
+ int opp_id,
+ bool enable,
+ bool rate_2x_mode,
+ struct mpc_dwb_flow_control *flow_control);
+
void mpc3_power_on_ogam_lut(
struct mpc *mpc, int mpcc_id,
bool power_on);
diff --git a/drivers/gpu/drm/amd/display/dc/optc/dcn10/dcn10_optc.c b/drivers/gpu/drm/amd/display/dc/optc/dcn10/dcn10_optc.c
index 097d06023e64..19d5ebc6763c 100644
--- a/drivers/gpu/drm/amd/display/dc/optc/dcn10/dcn10_optc.c
+++ b/drivers/gpu/drm/amd/display/dc/optc/dcn10/dcn10_optc.c
@@ -302,7 +302,6 @@ void optc1_program_timing(
/* Enable stereo - only when we need to pack 3D frame. Other types
* of stereo handled in explicit call
*/
-
if (optc->funcs->is_two_pixels_per_container(&patched_crtc_timing) || optc1->opp_count == 2)
h_div = H_TIMING_DIV_BY2;
@@ -1471,37 +1470,71 @@ bool optc1_configure_crc(struct timing_generator *optc,
if (!optc1_is_tg_enabled(optc))
return false;
- REG_WRITE(OTG_CRC_CNTL, 0);
+ if (!params->enable || params->reset)
+ REG_WRITE(OTG_CRC_CNTL, 0);
if (!params->enable)
return true;
/* Program frame boundaries */
- /* Window A x axis start and end. */
- REG_UPDATE_2(OTG_CRC0_WINDOWA_X_CONTROL,
- OTG_CRC0_WINDOWA_X_START, params->windowa_x_start,
- OTG_CRC0_WINDOWA_X_END, params->windowa_x_end);
-
- /* Window A y axis start and end. */
- REG_UPDATE_2(OTG_CRC0_WINDOWA_Y_CONTROL,
- OTG_CRC0_WINDOWA_Y_START, params->windowa_y_start,
- OTG_CRC0_WINDOWA_Y_END, params->windowa_y_end);
-
- /* Window B x axis start and end. */
- REG_UPDATE_2(OTG_CRC0_WINDOWB_X_CONTROL,
- OTG_CRC0_WINDOWB_X_START, params->windowb_x_start,
- OTG_CRC0_WINDOWB_X_END, params->windowb_x_end);
-
- /* Window B y axis start and end. */
- REG_UPDATE_2(OTG_CRC0_WINDOWB_Y_CONTROL,
- OTG_CRC0_WINDOWB_Y_START, params->windowb_y_start,
- OTG_CRC0_WINDOWB_Y_END, params->windowb_y_end);
-
- /* Set crc mode and selection, and enable. Only using CRC0*/
- REG_UPDATE_3(OTG_CRC_CNTL,
- OTG_CRC_CONT_EN, params->continuous_mode ? 1 : 0,
- OTG_CRC0_SELECT, params->selection,
- OTG_CRC_EN, 1);
+ switch (params->crc_eng_inst) {
+ case 0:
+ /* Window A x axis start and end. */
+ REG_UPDATE_2(OTG_CRC0_WINDOWA_X_CONTROL,
+ OTG_CRC0_WINDOWA_X_START, params->windowa_x_start,
+ OTG_CRC0_WINDOWA_X_END, params->windowa_x_end);
+
+ /* Window A y axis start and end. */
+ REG_UPDATE_2(OTG_CRC0_WINDOWA_Y_CONTROL,
+ OTG_CRC0_WINDOWA_Y_START, params->windowa_y_start,
+ OTG_CRC0_WINDOWA_Y_END, params->windowa_y_end);
+
+ /* Window B x axis start and end. */
+ REG_UPDATE_2(OTG_CRC0_WINDOWB_X_CONTROL,
+ OTG_CRC0_WINDOWB_X_START, params->windowb_x_start,
+ OTG_CRC0_WINDOWB_X_END, params->windowb_x_end);
+
+ /* Window B y axis start and end. */
+ REG_UPDATE_2(OTG_CRC0_WINDOWB_Y_CONTROL,
+ OTG_CRC0_WINDOWB_Y_START, params->windowb_y_start,
+ OTG_CRC0_WINDOWB_Y_END, params->windowb_y_end);
+
+ /* Set crc mode and selection, and enable.*/
+ REG_UPDATE_3(OTG_CRC_CNTL,
+ OTG_CRC_CONT_EN, params->continuous_mode ? 1 : 0,
+ OTG_CRC0_SELECT, params->selection,
+ OTG_CRC_EN, 1);
+ break;
+ case 1:
+ /* Window A x axis start and end. */
+ REG_UPDATE_2(OTG_CRC1_WINDOWA_X_CONTROL,
+ OTG_CRC1_WINDOWA_X_START, params->windowa_x_start,
+ OTG_CRC1_WINDOWA_X_END, params->windowa_x_end);
+
+ /* Window A y axis start and end. */
+ REG_UPDATE_2(OTG_CRC1_WINDOWA_Y_CONTROL,
+ OTG_CRC1_WINDOWA_Y_START, params->windowa_y_start,
+ OTG_CRC1_WINDOWA_Y_END, params->windowa_y_end);
+
+ /* Window B x axis start and end. */
+ REG_UPDATE_2(OTG_CRC1_WINDOWB_X_CONTROL,
+ OTG_CRC1_WINDOWB_X_START, params->windowb_x_start,
+ OTG_CRC1_WINDOWB_X_END, params->windowb_x_end);
+
+ /* Window B y axis start and end. */
+ REG_UPDATE_2(OTG_CRC1_WINDOWB_Y_CONTROL,
+ OTG_CRC1_WINDOWB_Y_START, params->windowb_y_start,
+ OTG_CRC1_WINDOWB_Y_END, params->windowb_y_end);
+
+ /* Set crc mode and selection, and enable.*/
+ REG_UPDATE_3(OTG_CRC_CNTL,
+ OTG_CRC_CONT_EN, params->continuous_mode ? 1 : 0,
+ OTG_CRC1_SELECT, params->selection,
+ OTG_CRC_EN, 1);
+ break;
+ default:
+ return false;
+ }
return true;
}
@@ -1510,6 +1543,7 @@ bool optc1_configure_crc(struct timing_generator *optc,
* optc1_get_crc - Capture CRC result per component
*
* @optc: timing_generator instance.
+ * @idx: index of crc engine to get CRC from
* @r_cr: 16-bit primary CRC signature for red data.
* @g_y: 16-bit primary CRC signature for green data.
* @b_cb: 16-bit primary CRC signature for blue data.
@@ -1521,7 +1555,7 @@ bool optc1_configure_crc(struct timing_generator *optc,
* If CRC is disabled, return false; otherwise, return true, and the CRC
* results in the parameters.
*/
-bool optc1_get_crc(struct timing_generator *optc,
+bool optc1_get_crc(struct timing_generator *optc, uint8_t idx,
uint32_t *r_cr, uint32_t *g_y, uint32_t *b_cb)
{
uint32_t field = 0;
@@ -1533,14 +1567,30 @@ bool optc1_get_crc(struct timing_generator *optc,
if (!field)
return false;
- /* OTG_CRC0_DATA_RG has the CRC16 results for the red and green component */
- REG_GET_2(OTG_CRC0_DATA_RG,
- CRC0_R_CR, r_cr,
- CRC0_G_Y, g_y);
+ switch (idx) {
+ case 0:
+ /* OTG_CRC0_DATA_RG has the CRC16 results for the red and green component */
+ REG_GET_2(OTG_CRC0_DATA_RG,
+ CRC0_R_CR, r_cr,
+ CRC0_G_Y, g_y);
- /* OTG_CRC0_DATA_B has the CRC16 results for the blue component */
- REG_GET(OTG_CRC0_DATA_B,
- CRC0_B_CB, b_cb);
+ /* OTG_CRC0_DATA_B has the CRC16 results for the blue component */
+ REG_GET(OTG_CRC0_DATA_B,
+ CRC0_B_CB, b_cb);
+ break;
+ case 1:
+ /* OTG_CRC1_DATA_RG has the CRC16 results for the red and green component */
+ REG_GET_2(OTG_CRC1_DATA_RG,
+ CRC1_R_CR, r_cr,
+ CRC1_G_Y, g_y);
+
+ /* OTG_CRC1_DATA_B has the CRC16 results for the blue component */
+ REG_GET(OTG_CRC1_DATA_B,
+ CRC1_B_CB, b_cb);
+ break;
+ default:
+ return false;
+ }
return true;
}
diff --git a/drivers/gpu/drm/amd/display/dc/optc/dcn10/dcn10_optc.h b/drivers/gpu/drm/amd/display/dc/optc/dcn10/dcn10_optc.h
index 40757f20d73f..159172178d51 100644
--- a/drivers/gpu/drm/amd/display/dc/optc/dcn10/dcn10_optc.h
+++ b/drivers/gpu/drm/amd/display/dc/optc/dcn10/dcn10_optc.h
@@ -86,6 +86,12 @@
SRI(OTG_CRC0_WINDOWA_Y_CONTROL, OTG, inst),\
SRI(OTG_CRC0_WINDOWB_X_CONTROL, OTG, inst),\
SRI(OTG_CRC0_WINDOWB_Y_CONTROL, OTG, inst),\
+ SRI(OTG_CRC1_DATA_RG, OTG, inst),\
+ SRI(OTG_CRC1_DATA_B, OTG, inst),\
+ SRI(OTG_CRC1_WINDOWA_X_CONTROL, OTG, inst),\
+ SRI(OTG_CRC1_WINDOWA_Y_CONTROL, OTG, inst),\
+ SRI(OTG_CRC1_WINDOWB_X_CONTROL, OTG, inst),\
+ SRI(OTG_CRC1_WINDOWB_Y_CONTROL, OTG, inst),\
SR(GSL_SOURCE_SELECT),\
SRI(OTG_GLOBAL_CONTROL2, OTG, inst),\
SRI(OTG_TRIGA_MANUAL_TRIG, OTG, inst)
@@ -315,6 +321,7 @@ struct dcn_optc_registers {
SF(OTG0_OTG_GSL_CONTROL, OTG_GSL_CHECK_ALL_FIELDS, mask_sh),\
SF(OTG0_OTG_CRC_CNTL, OTG_CRC_CONT_EN, mask_sh),\
SF(OTG0_OTG_CRC_CNTL, OTG_CRC0_SELECT, mask_sh),\
+ SF(OTG0_OTG_CRC_CNTL, OTG_CRC1_SELECT, mask_sh),\
SF(OTG0_OTG_CRC_CNTL, OTG_CRC_EN, mask_sh),\
SF(OTG0_OTG_CRC0_DATA_RG, CRC0_R_CR, mask_sh),\
SF(OTG0_OTG_CRC0_DATA_RG, CRC0_G_Y, mask_sh),\
@@ -327,6 +334,17 @@ struct dcn_optc_registers {
SF(OTG0_OTG_CRC0_WINDOWB_X_CONTROL, OTG_CRC0_WINDOWB_X_END, mask_sh),\
SF(OTG0_OTG_CRC0_WINDOWB_Y_CONTROL, OTG_CRC0_WINDOWB_Y_START, mask_sh),\
SF(OTG0_OTG_CRC0_WINDOWB_Y_CONTROL, OTG_CRC0_WINDOWB_Y_END, mask_sh),\
+ SF(OTG0_OTG_CRC1_DATA_RG, CRC1_R_CR, mask_sh),\
+ SF(OTG0_OTG_CRC1_DATA_RG, CRC1_G_Y, mask_sh),\
+ SF(OTG0_OTG_CRC1_DATA_B, CRC1_B_CB, mask_sh),\
+ SF(OTG0_OTG_CRC1_WINDOWA_X_CONTROL, OTG_CRC1_WINDOWA_X_START, mask_sh),\
+ SF(OTG0_OTG_CRC1_WINDOWA_X_CONTROL, OTG_CRC1_WINDOWA_X_END, mask_sh),\
+ SF(OTG0_OTG_CRC1_WINDOWA_Y_CONTROL, OTG_CRC1_WINDOWA_Y_START, mask_sh),\
+ SF(OTG0_OTG_CRC1_WINDOWA_Y_CONTROL, OTG_CRC1_WINDOWA_Y_END, mask_sh),\
+ SF(OTG0_OTG_CRC1_WINDOWB_X_CONTROL, OTG_CRC1_WINDOWB_X_START, mask_sh),\
+ SF(OTG0_OTG_CRC1_WINDOWB_X_CONTROL, OTG_CRC1_WINDOWB_X_END, mask_sh),\
+ SF(OTG0_OTG_CRC1_WINDOWB_Y_CONTROL, OTG_CRC1_WINDOWB_Y_START, mask_sh),\
+ SF(OTG0_OTG_CRC1_WINDOWB_Y_CONTROL, OTG_CRC1_WINDOWB_Y_END, mask_sh),\
SF(GSL_SOURCE_SELECT, GSL0_READY_SOURCE_SEL, mask_sh),\
SF(GSL_SOURCE_SELECT, GSL1_READY_SOURCE_SEL, mask_sh),\
SF(GSL_SOURCE_SELECT, GSL2_READY_SOURCE_SEL, mask_sh),\
@@ -482,6 +500,7 @@ struct dcn_optc_registers {
type OTG_MASTER_UPDATE_LOCK_VUPDATE_KEEPOUT_EN;\
type OTG_CRC_CONT_EN;\
type OTG_CRC0_SELECT;\
+ type OTG_CRC1_SELECT;\
type OTG_CRC_EN;\
type CRC0_R_CR;\
type CRC0_G_Y;\
diff --git a/drivers/gpu/drm/amd/display/dc/optc/dcn35/dcn35_optc.c b/drivers/gpu/drm/amd/display/dc/optc/dcn35/dcn35_optc.c
index dfa9364fe5a6..d21e82b927d0 100644
--- a/drivers/gpu/drm/amd/display/dc/optc/dcn35/dcn35_optc.c
+++ b/drivers/gpu/drm/amd/display/dc/optc/dcn35/dcn35_optc.c
@@ -183,34 +183,87 @@ static bool optc35_configure_crc(struct timing_generator *optc,
{
struct optc *optc1 = DCN10TG_FROM_TG(optc);
+ /* Cannot configure crc on a CRTC that is disabled */
if (!optc1_is_tg_enabled(optc))
return false;
- REG_WRITE(OTG_CRC_CNTL, 0);
+
+ if (!params->enable || params->reset)
+ REG_WRITE(OTG_CRC_CNTL, 0);
+
if (!params->enable)
return true;
- REG_UPDATE_2(OTG_CRC0_WINDOWA_X_CONTROL,
- OTG_CRC0_WINDOWA_X_START, params->windowa_x_start,
- OTG_CRC0_WINDOWA_X_END, params->windowa_x_end);
- REG_UPDATE_2(OTG_CRC0_WINDOWA_Y_CONTROL,
- OTG_CRC0_WINDOWA_Y_START, params->windowa_y_start,
- OTG_CRC0_WINDOWA_Y_END, params->windowa_y_end);
- REG_UPDATE_2(OTG_CRC0_WINDOWB_X_CONTROL,
- OTG_CRC0_WINDOWB_X_START, params->windowb_x_start,
- OTG_CRC0_WINDOWB_X_END, params->windowb_x_end);
- REG_UPDATE_2(OTG_CRC0_WINDOWB_Y_CONTROL,
- OTG_CRC0_WINDOWB_Y_START, params->windowb_y_start,
- OTG_CRC0_WINDOWB_Y_END, params->windowb_y_end);
- if (optc1->base.ctx->dc->debug.otg_crc_db && optc1->tg_mask->OTG_CRC_WINDOW_DB_EN != 0) {
- REG_UPDATE_4(OTG_CRC_CNTL,
- OTG_CRC_CONT_EN, params->continuous_mode ? 1 : 0,
- OTG_CRC0_SELECT, params->selection,
- OTG_CRC_EN, 1,
- OTG_CRC_WINDOW_DB_EN, 1);
- } else
- REG_UPDATE_3(OTG_CRC_CNTL,
- OTG_CRC_CONT_EN, params->continuous_mode ? 1 : 0,
- OTG_CRC0_SELECT, params->selection,
- OTG_CRC_EN, 1);
+
+ /* Program frame boundaries */
+ switch (params->crc_eng_inst) {
+ case 0:
+ /* Window A x axis start and end. */
+ REG_UPDATE_2(OTG_CRC0_WINDOWA_X_CONTROL,
+ OTG_CRC0_WINDOWA_X_START, params->windowa_x_start,
+ OTG_CRC0_WINDOWA_X_END, params->windowa_x_end);
+
+ /* Window A y axis start and end. */
+ REG_UPDATE_2(OTG_CRC0_WINDOWA_Y_CONTROL,
+ OTG_CRC0_WINDOWA_Y_START, params->windowa_y_start,
+ OTG_CRC0_WINDOWA_Y_END, params->windowa_y_end);
+
+ /* Window B x axis start and end. */
+ REG_UPDATE_2(OTG_CRC0_WINDOWB_X_CONTROL,
+ OTG_CRC0_WINDOWB_X_START, params->windowb_x_start,
+ OTG_CRC0_WINDOWB_X_END, params->windowb_x_end);
+
+ /* Window B y axis start and end. */
+ REG_UPDATE_2(OTG_CRC0_WINDOWB_Y_CONTROL,
+ OTG_CRC0_WINDOWB_Y_START, params->windowb_y_start,
+ OTG_CRC0_WINDOWB_Y_END, params->windowb_y_end);
+
+ if (optc1->base.ctx->dc->debug.otg_crc_db && optc1->tg_mask->OTG_CRC_WINDOW_DB_EN != 0)
+ REG_UPDATE_4(OTG_CRC_CNTL,
+ OTG_CRC_CONT_EN, params->continuous_mode ? 1 : 0,
+ OTG_CRC0_SELECT, params->selection,
+ OTG_CRC_EN, 1,
+ OTG_CRC_WINDOW_DB_EN, 1);
+ else
+ REG_UPDATE_3(OTG_CRC_CNTL,
+ OTG_CRC_CONT_EN, params->continuous_mode ? 1 : 0,
+ OTG_CRC0_SELECT, params->selection,
+ OTG_CRC_EN, 1);
+ break;
+ case 1:
+ /* Window A x axis start and end. */
+ REG_UPDATE_2(OTG_CRC1_WINDOWA_X_CONTROL,
+ OTG_CRC1_WINDOWA_X_START, params->windowa_x_start,
+ OTG_CRC1_WINDOWA_X_END, params->windowa_x_end);
+
+ /* Window A y axis start and end. */
+ REG_UPDATE_2(OTG_CRC1_WINDOWA_Y_CONTROL,
+ OTG_CRC1_WINDOWA_Y_START, params->windowa_y_start,
+ OTG_CRC1_WINDOWA_Y_END, params->windowa_y_end);
+
+ /* Window B x axis start and end. */
+ REG_UPDATE_2(OTG_CRC1_WINDOWB_X_CONTROL,
+ OTG_CRC1_WINDOWB_X_START, params->windowb_x_start,
+ OTG_CRC1_WINDOWB_X_END, params->windowb_x_end);
+
+ /* Window B y axis start and end. */
+ REG_UPDATE_2(OTG_CRC1_WINDOWB_Y_CONTROL,
+ OTG_CRC1_WINDOWB_Y_START, params->windowb_y_start,
+ OTG_CRC1_WINDOWB_Y_END, params->windowb_y_end);
+
+ if (optc1->base.ctx->dc->debug.otg_crc_db && optc1->tg_mask->OTG_CRC_WINDOW_DB_EN != 0)
+ REG_UPDATE_4(OTG_CRC_CNTL,
+ OTG_CRC_CONT_EN, params->continuous_mode ? 1 : 0,
+ OTG_CRC1_SELECT, params->selection,
+ OTG_CRC_EN, 1,
+ OTG_CRC_WINDOW_DB_EN, 1);
+ else
+ REG_UPDATE_3(OTG_CRC_CNTL,
+ OTG_CRC_CONT_EN, params->continuous_mode ? 1 : 0,
+ OTG_CRC1_SELECT, params->selection,
+ OTG_CRC_EN, 1);
+ break;
+ default:
+ return false;
+ }
return true;
}
diff --git a/drivers/gpu/drm/amd/display/dc/optc/dcn401/dcn401_optc.c b/drivers/gpu/drm/amd/display/dc/optc/dcn401/dcn401_optc.c
index 783ca9acc762..338a0cad23a5 100644
--- a/drivers/gpu/drm/amd/display/dc/optc/dcn401/dcn401_optc.c
+++ b/drivers/gpu/drm/amd/display/dc/optc/dcn401/dcn401_optc.c
@@ -315,7 +315,7 @@ void optc401_set_drr(
struct drr_params amended_params = { 0 };
bool program_manual_trigger = false;
- if (dc->caps.dmub_caps.fams_ver >= 2 && dc->debug.fams2_config.bits.enable) {
+ if (dc->caps.dmub_caps.fams_ver == dc->debug.fams_version.ver && dc->debug.fams2_config.bits.enable) {
if (params != NULL &&
params->vertical_total_max > 0 &&
params->vertical_total_min > 0) {
@@ -380,7 +380,7 @@ void optc401_set_vtotal_min_max(struct timing_generator *optc, int vtotal_min, i
{
struct dc *dc = optc->ctx->dc;
- if (dc->caps.dmub_caps.fams_ver >= 2 && dc->debug.fams2_config.bits.enable) {
+ if (dc->caps.dmub_caps.fams_ver == dc->debug.fams_version.ver && dc->debug.fams2_config.bits.enable) {
/* FAMS2 */
dc_dmub_srv_fams2_drr_update(dc, optc->inst,
vtotal_min,
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn10/dcn10_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn10/dcn10_resource.c
index 770a380cc03d..e92f14d50adb 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn10/dcn10_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn10/dcn10_resource.c
@@ -1258,6 +1258,11 @@ struct stream_encoder *dcn10_find_first_free_match_stream_enc_for_link(
return NULL;
}
+unsigned int dcn10_get_vstartup_for_pipe(struct pipe_ctx *pipe_ctx)
+{
+ return pipe_ctx->pipe_dlg_param.vstartup_start;
+}
+
static const struct dc_cap_funcs cap_funcs = {
.get_dcc_compression_cap = dcn10_get_dcc_compression_cap
};
@@ -1272,7 +1277,8 @@ static const struct resource_funcs dcn10_res_pool_funcs = {
.validate_global = dcn10_validate_global,
.add_stream_to_ctx = dcn10_add_stream_to_ctx,
.patch_unknown_plane_state = dcn10_patch_unknown_plane_state,
- .find_first_free_match_stream_enc_for_link = dcn10_find_first_free_match_stream_enc_for_link
+ .find_first_free_match_stream_enc_for_link = dcn10_find_first_free_match_stream_enc_for_link,
+ .get_vstartup_for_pipe = dcn10_get_vstartup_for_pipe
};
static uint32_t read_pipe_fuses(struct dc_context *ctx)
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn10/dcn10_resource.h b/drivers/gpu/drm/amd/display/dc/resource/dcn10/dcn10_resource.h
index bf8e33cd8147..7bc1be53e800 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn10/dcn10_resource.h
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn10/dcn10_resource.h
@@ -51,6 +51,7 @@ struct stream_encoder *dcn10_find_first_free_match_stream_enc_for_link(
const struct resource_pool *pool,
struct dc_stream_state *stream);
+unsigned int dcn10_get_vstartup_for_pipe(struct pipe_ctx *pipe_ctx);
#endif /* __DC_RESOURCE_DCN10_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn20/dcn20_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn20/dcn20_resource.c
index 189d0c85872e..5c6dc710e96c 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn20/dcn20_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn20/dcn20_resource.c
@@ -1509,41 +1509,12 @@ bool dcn20_split_stream_for_odm(
next_odm_pipe->prev_odm_pipe = prev_odm_pipe;
if (prev_odm_pipe->plane_state) {
- struct scaler_data *sd = &prev_odm_pipe->plane_res.scl_data;
- int new_width;
-
- /* HACTIVE halved for odm combine */
- sd->h_active /= 2;
- /* Calculate new vp and recout for left pipe */
- /* Need at least 16 pixels width per side */
- if (sd->recout.x + 16 >= sd->h_active)
- return false;
- new_width = sd->h_active - sd->recout.x;
- sd->viewport.width -= dc_fixpt_floor(dc_fixpt_mul_int(
- sd->ratios.horz, sd->recout.width - new_width));
- sd->viewport_c.width -= dc_fixpt_floor(dc_fixpt_mul_int(
- sd->ratios.horz_c, sd->recout.width - new_width));
- sd->recout.width = new_width;
-
- /* Calculate new vp and recout for right pipe */
- sd = &next_odm_pipe->plane_res.scl_data;
- /* HACTIVE halved for odm combine */
- sd->h_active /= 2;
- /* Need at least 16 pixels width per side */
- if (new_width <= 16)
- return false;
- new_width = sd->recout.width + sd->recout.x - sd->h_active;
- sd->viewport.width -= dc_fixpt_floor(dc_fixpt_mul_int(
- sd->ratios.horz, sd->recout.width - new_width));
- sd->viewport_c.width -= dc_fixpt_floor(dc_fixpt_mul_int(
- sd->ratios.horz_c, sd->recout.width - new_width));
- sd->recout.width = new_width;
- sd->viewport.x += dc_fixpt_floor(dc_fixpt_mul_int(
- sd->ratios.horz, sd->h_active - sd->recout.x));
- sd->viewport_c.x += dc_fixpt_floor(dc_fixpt_mul_int(
- sd->ratios.horz_c, sd->h_active - sd->recout.x));
- sd->recout.x = 0;
+ if (!resource_build_scaling_params(prev_odm_pipe) ||
+ !resource_build_scaling_params(next_odm_pipe)) {
+ return false;
+ }
}
+
if (!next_odm_pipe->top_pipe)
next_odm_pipe->stream_res.opp = pool->opps[next_odm_pipe->pipe_idx];
else
@@ -2132,6 +2103,7 @@ bool dcn20_fast_validate_bw(
ASSERT(0);
}
}
+
/* Actual dsc count per stream dsc validation*/
if (!dcn20_validate_dsc(dc, context)) {
context->bw_ctx.dml.vba.ValidationStatus[context->bw_ctx.dml.vba.soc.num_states] =
@@ -2257,7 +2229,8 @@ static const struct resource_funcs dcn20_res_pool_funcs = {
.patch_unknown_plane_state = dcn20_patch_unknown_plane_state,
.set_mcif_arb_params = dcn20_set_mcif_arb_params,
.populate_dml_pipes = dcn20_populate_dml_pipes_from_context,
- .find_first_free_match_stream_enc_for_link = dcn10_find_first_free_match_stream_enc_for_link
+ .find_first_free_match_stream_enc_for_link = dcn10_find_first_free_match_stream_enc_for_link,
+ .get_vstartup_for_pipe = dcn10_get_vstartup_for_pipe
};
bool dcn20_dwbc_create(struct dc_context *ctx, struct resource_pool *pool)
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn201/dcn201_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn201/dcn201_resource.c
index d3d67d366523..43fa2cb117f3 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn201/dcn201_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn201/dcn201_resource.c
@@ -59,8 +59,8 @@
#include "cyan_skillfish_ip_offset.h"
-#include "dcn/dcn_2_0_3_offset.h"
-#include "dcn/dcn_2_0_3_sh_mask.h"
+#include "dcn/dcn_2_0_1_offset.h"
+#include "dcn/dcn_2_0_1_sh_mask.h"
#include "dpcs/dpcs_2_0_3_offset.h"
#include "dpcs/dpcs_2_0_3_sh_mask.h"
@@ -1079,7 +1079,8 @@ static struct resource_funcs dcn201_res_pool_funcs = {
.populate_dml_writeback_from_context = dcn201_populate_dml_writeback_from_context,
.patch_unknown_plane_state = dcn20_patch_unknown_plane_state,
.set_mcif_arb_params = dcn20_set_mcif_arb_params,
- .find_first_free_match_stream_enc_for_link = dcn10_find_first_free_match_stream_enc_for_link
+ .find_first_free_match_stream_enc_for_link = dcn10_find_first_free_match_stream_enc_for_link,
+ .get_vstartup_for_pipe = dcn10_get_vstartup_for_pipe
};
static bool dcn201_resource_construct(
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn21/dcn21_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn21/dcn21_resource.c
index 021ba8ac5c8c..2615c36d5ffe 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn21/dcn21_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn21/dcn21_resource.c
@@ -1378,6 +1378,7 @@ static const struct resource_funcs dcn21_res_pool_funcs = {
.find_first_free_match_stream_enc_for_link = dcn10_find_first_free_match_stream_enc_for_link,
.update_bw_bounding_box = dcn21_update_bw_bounding_box,
.get_panel_config_defaults = dcn21_get_panel_config_defaults,
+ .get_vstartup_for_pipe = dcn10_get_vstartup_for_pipe
};
static bool dcn21_resource_construct(
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn30/dcn30_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn30/dcn30_resource.c
index cd31e4f16c14..13202ce30d66 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn30/dcn30_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn30/dcn30_resource.c
@@ -2250,6 +2250,7 @@ static const struct resource_funcs dcn30_res_pool_funcs = {
.update_bw_bounding_box = dcn30_update_bw_bounding_box,
.patch_unknown_plane_state = dcn20_patch_unknown_plane_state,
.get_panel_config_defaults = dcn30_get_panel_config_defaults,
+ .get_vstartup_for_pipe = dcn10_get_vstartup_for_pipe
};
#define CTX ctx
@@ -2353,6 +2354,7 @@ static bool dcn30_resource_construct(
dc->caps.dp_hdmi21_pcon_support = true;
dc->caps.max_v_total = (1 << 15) - 1;
+ dc->caps.vtotal_limited_by_fp2 = true;
/* read VBIOS LTTPR caps */
{
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn301/dcn301_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn301/dcn301_resource.c
index a9816affd312..121a86a59833 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn301/dcn301_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn301/dcn301_resource.c
@@ -671,9 +671,9 @@ static const struct dc_plane_cap plane_cap = {
/* 6:1 downscaling ratio: 1000/6 = 166.666 */
.max_downscale_factor = {
- .argb8888 = 167,
- .nv12 = 167,
- .fp16 = 167
+ .argb8888 = 358,
+ .nv12 = 358,
+ .fp16 = 358
},
64,
64
@@ -693,7 +693,7 @@ static const struct dc_debug_options debug_defaults_drv = {
.disable_dcc = DCC_ENABLE,
.vsr_support = true,
.performance_trace = false,
- .max_downscale_src_width = 7680,/*upto 8K*/
+ .max_downscale_src_width = 4096,/*upto true 4k*/
.scl_reset_length10 = true,
.sanity_checks = false,
.underflow_assert_delay_us = 0xFFFFFFFF,
@@ -1400,7 +1400,8 @@ static struct resource_funcs dcn301_res_pool_funcs = {
.acquire_post_bldn_3dlut = dcn30_acquire_post_bldn_3dlut,
.release_post_bldn_3dlut = dcn30_release_post_bldn_3dlut,
.update_bw_bounding_box = dcn301_update_bw_bounding_box,
- .patch_unknown_plane_state = dcn20_patch_unknown_plane_state
+ .patch_unknown_plane_state = dcn20_patch_unknown_plane_state,
+ .get_vstartup_for_pipe = dcn10_get_vstartup_for_pipe
};
static bool dcn301_resource_construct(
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn302/dcn302_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn302/dcn302_resource.c
index 02af8b8f4d27..012c5fd52cb1 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn302/dcn302_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn302/dcn302_resource.c
@@ -1151,6 +1151,7 @@ static struct resource_funcs dcn302_res_pool_funcs = {
.update_bw_bounding_box = dcn302_update_bw_bounding_box,
.patch_unknown_plane_state = dcn20_patch_unknown_plane_state,
.get_panel_config_defaults = dcn302_get_panel_config_defaults,
+ .get_vstartup_for_pipe = dcn10_get_vstartup_for_pipe
};
static struct dc_cap_funcs cap_funcs = {
@@ -1233,6 +1234,7 @@ static bool dcn302_resource_construct(
dc->caps.extended_aux_timeout_support = true;
dc->caps.dmcub_support = true;
dc->caps.max_v_total = (1 << 15) - 1;
+ dc->caps.vtotal_limited_by_fp2 = true;
/* Color pipeline capabilities */
dc->caps.color.dpp.dcn_arch = 1;
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn303/dcn303_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn303/dcn303_resource.c
index 7002a8dd358a..a8d0b4686f9a 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn303/dcn303_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn303/dcn303_resource.c
@@ -1096,6 +1096,7 @@ static struct resource_funcs dcn303_res_pool_funcs = {
.update_bw_bounding_box = dcn303_update_bw_bounding_box,
.patch_unknown_plane_state = dcn20_patch_unknown_plane_state,
.get_panel_config_defaults = dcn303_get_panel_config_defaults,
+ .get_vstartup_for_pipe = dcn10_get_vstartup_for_pipe
};
static struct dc_cap_funcs cap_funcs = {
@@ -1178,6 +1179,7 @@ static bool dcn303_resource_construct(
dc->caps.extended_aux_timeout_support = true;
dc->caps.dmcub_support = true;
dc->caps.max_v_total = (1 << 15) - 1;
+ dc->caps.vtotal_limited_by_fp2 = true;
/* Color pipeline capabilities */
dc->caps.color.dpp.dcn_arch = 1;
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn31/dcn31_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn31/dcn31_resource.c
index c16cf1c8f7f9..911bd60d4fbc 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn31/dcn31_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn31/dcn31_resource.c
@@ -1720,6 +1720,12 @@ int dcn31_populate_dml_pipes_from_context(
return pipe_cnt;
}
+unsigned int dcn31_get_det_buffer_size(
+ const struct dc_state *context)
+{
+ return context->bw_ctx.dml.ip.det_buffer_size_kbytes;
+}
+
void dcn31_calculate_wm_and_dlg(
struct dc *dc, struct dc_state *context,
display_e2e_pipe_params_st *pipes,
@@ -1842,6 +1848,8 @@ static struct resource_funcs dcn31_res_pool_funcs = {
.update_bw_bounding_box = dcn31_update_bw_bounding_box,
.patch_unknown_plane_state = dcn20_patch_unknown_plane_state,
.get_panel_config_defaults = dcn31_get_panel_config_defaults,
+ .get_det_buffer_size = dcn31_get_det_buffer_size,
+ .get_vstartup_for_pipe = dcn10_get_vstartup_for_pipe
};
static struct clock_source *dcn30_clock_source_create(
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn31/dcn31_resource.h b/drivers/gpu/drm/amd/display/dc/resource/dcn31/dcn31_resource.h
index 901436591ed4..551ad912f7be 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn31/dcn31_resource.h
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn31/dcn31_resource.h
@@ -63,6 +63,9 @@ struct resource_pool *dcn31_create_resource_pool(
const struct dc_init_data *init_data,
struct dc *dc);
+unsigned int dcn31_get_det_buffer_size(
+ const struct dc_state *context);
+
/*temp: B0 specific before switch to dcn313 headers*/
#ifndef regPHYPLLF_PIXCLK_RESYNC_CNTL
#define regPHYPLLF_PIXCLK_RESYNC_CNTL 0x007e
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn314/dcn314_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn314/dcn314_resource.c
index c0f48c78e968..e3ba105034f8 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn314/dcn314_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn314/dcn314_resource.c
@@ -1777,6 +1777,8 @@ static struct resource_funcs dcn314_res_pool_funcs = {
.patch_unknown_plane_state = dcn20_patch_unknown_plane_state,
.get_panel_config_defaults = dcn314_get_panel_config_defaults,
.get_preferred_eng_id_dpia = dcn314_get_preferred_eng_id_dpia,
+ .get_det_buffer_size = dcn31_get_det_buffer_size,
+ .get_vstartup_for_pipe = dcn10_get_vstartup_for_pipe
};
static struct clock_source *dcn30_clock_source_create(
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn315/dcn315_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn315/dcn315_resource.c
index 6c3295259a81..14acef036b5a 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn315/dcn315_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn315/dcn315_resource.c
@@ -1845,6 +1845,8 @@ static struct resource_funcs dcn315_res_pool_funcs = {
.patch_unknown_plane_state = dcn20_patch_unknown_plane_state,
.get_panel_config_defaults = dcn315_get_panel_config_defaults,
.get_power_profile = dcn315_get_power_profile,
+ .get_det_buffer_size = dcn31_get_det_buffer_size,
+ .get_vstartup_for_pipe = dcn10_get_vstartup_for_pipe
};
static bool dcn315_resource_construct(
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn316/dcn316_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn316/dcn316_resource.c
index 6edaaadcb173..568094827212 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn316/dcn316_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn316/dcn316_resource.c
@@ -1719,6 +1719,8 @@ static struct resource_funcs dcn316_res_pool_funcs = {
.update_bw_bounding_box = dcn316_update_bw_bounding_box,
.patch_unknown_plane_state = dcn20_patch_unknown_plane_state,
.get_panel_config_defaults = dcn316_get_panel_config_defaults,
+ .get_det_buffer_size = dcn31_get_det_buffer_size,
+ .get_vstartup_for_pipe = dcn10_get_vstartup_for_pipe
};
static bool dcn316_resource_construct(
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn32/dcn32_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn32/dcn32_resource.c
index 01d1a11d5545..664302876019 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn32/dcn32_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn32/dcn32_resource.c
@@ -2066,6 +2066,7 @@ static struct resource_funcs dcn32_res_pool_funcs = {
.add_phantom_pipes = dcn32_add_phantom_pipes,
.build_pipe_pix_clk_params = dcn20_build_pipe_pix_clk_params,
.calculate_mall_ways_from_bytes = dcn32_calculate_mall_ways_from_bytes,
+ .get_vstartup_for_pipe = dcn10_get_vstartup_for_pipe
};
static uint32_t read_pipe_fuses(struct dc_context *ctx)
@@ -2189,6 +2190,7 @@ static bool dcn32_resource_construct(
dc->caps.dmcub_support = true;
dc->caps.seamless_odm = true;
dc->caps.max_v_total = (1 << 15) - 1;
+ dc->caps.vtotal_limited_by_fp2 = true;
/* Color pipeline capabilities */
dc->caps.color.dpp.dcn_arch = 1;
@@ -2803,6 +2805,7 @@ struct pipe_ctx *dcn32_acquire_free_pipe_as_secondary_opp_head(
free_pipe->plane_res.xfm = pool->transforms[free_pipe_idx];
free_pipe->plane_res.dpp = pool->dpps[free_pipe_idx];
free_pipe->plane_res.mpcc_inst = pool->dpps[free_pipe_idx]->inst;
+ free_pipe->hblank_borrow = otg_master->hblank_borrow;
if (free_pipe->stream->timing.flags.DSC == 1) {
dcn20_acquire_dsc(free_pipe->stream->ctx->dc,
&new_ctx->res_ctx,
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn321/dcn321_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn321/dcn321_resource.c
index 5cb74fd9cb7d..38d76434683e 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn321/dcn321_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn321/dcn321_resource.c
@@ -1624,6 +1624,7 @@ static struct resource_funcs dcn321_res_pool_funcs = {
.add_phantom_pipes = dcn32_add_phantom_pipes,
.build_pipe_pix_clk_params = dcn20_build_pipe_pix_clk_params,
.calculate_mall_ways_from_bytes = dcn32_calculate_mall_ways_from_bytes,
+ .get_vstartup_for_pipe = dcn10_get_vstartup_for_pipe
};
static uint32_t read_pipe_fuses(struct dc_context *ctx)
@@ -1742,6 +1743,7 @@ static bool dcn321_resource_construct(
dc->caps.extended_aux_timeout_support = true;
dc->caps.dmcub_support = true;
dc->caps.max_v_total = (1 << 15) - 1;
+ dc->caps.vtotal_limited_by_fp2 = true;
/* Color pipeline capabilities */
dc->caps.color.dpp.dcn_arch = 1;
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn35/dcn35_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn35/dcn35_resource.c
index 6cc2960b6104..8ee3d99ea2aa 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn35/dcn35_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn35/dcn35_resource.c
@@ -1752,6 +1752,13 @@ static bool dcn35_validate_bandwidth(struct dc *dc,
return out;
}
+enum dc_status dcn35_patch_unknown_plane_state(struct dc_plane_state *plane_state)
+{
+ plane_state->tiling_info.gfxversion = DcGfxVersion9;
+ dcn20_patch_unknown_plane_state(plane_state);
+ return DC_OK;
+}
+
static struct resource_funcs dcn35_res_pool_funcs = {
.destroy = dcn35_destroy_resource_pool,
@@ -1775,9 +1782,11 @@ static struct resource_funcs dcn35_res_pool_funcs = {
.acquire_post_bldn_3dlut = dcn30_acquire_post_bldn_3dlut,
.release_post_bldn_3dlut = dcn30_release_post_bldn_3dlut,
.update_bw_bounding_box = dcn35_update_bw_bounding_box_fpu,
- .patch_unknown_plane_state = dcn20_patch_unknown_plane_state,
+ .patch_unknown_plane_state = dcn35_patch_unknown_plane_state,
.get_panel_config_defaults = dcn35_get_panel_config_defaults,
.get_preferred_eng_id_dpia = dcn35_get_preferred_eng_id_dpia,
+ .get_det_buffer_size = dcn31_get_det_buffer_size,
+ .get_vstartup_for_pipe = dcn10_get_vstartup_for_pipe
};
static bool dcn35_resource_construct(
@@ -1849,6 +1858,7 @@ static bool dcn35_resource_construct(
dc->caps.zstate_support = true;
dc->caps.ips_support = true;
dc->caps.max_v_total = (1 << 15) - 1;
+ dc->caps.vtotal_limited_by_fp2 = true;
/* Color pipeline capabilities */
dc->caps.color.dpp.dcn_arch = 1;
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn35/dcn35_resource.h b/drivers/gpu/drm/amd/display/dc/resource/dcn35/dcn35_resource.h
index f97bb4cb3761..9d03a55d90cf 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn35/dcn35_resource.h
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn35/dcn35_resource.h
@@ -35,6 +35,7 @@
extern struct _vcs_dpi_ip_params_st dcn3_5_ip;
extern struct _vcs_dpi_soc_bounding_box_st dcn3_5_soc;
+enum dc_status dcn35_patch_unknown_plane_state(struct dc_plane_state *plane_state);
struct dcn35_resource_pool {
struct resource_pool base;
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn351/dcn351_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn351/dcn351_resource.c
index d87e2641cda1..14f7c3acdc96 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn351/dcn351_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn351/dcn351_resource.c
@@ -1754,9 +1754,11 @@ static struct resource_funcs dcn351_res_pool_funcs = {
.acquire_post_bldn_3dlut = dcn30_acquire_post_bldn_3dlut,
.release_post_bldn_3dlut = dcn30_release_post_bldn_3dlut,
.update_bw_bounding_box = dcn351_update_bw_bounding_box_fpu,
- .patch_unknown_plane_state = dcn20_patch_unknown_plane_state,
+ .patch_unknown_plane_state = dcn35_patch_unknown_plane_state,
.get_panel_config_defaults = dcn35_get_panel_config_defaults,
.get_preferred_eng_id_dpia = dcn351_get_preferred_eng_id_dpia,
+ .get_det_buffer_size = dcn31_get_det_buffer_size,
+ .get_vstartup_for_pipe = dcn10_get_vstartup_for_pipe
};
static bool dcn351_resource_construct(
@@ -1828,6 +1830,7 @@ static bool dcn351_resource_construct(
dc->caps.zstate_support = true;
dc->caps.ips_support = true;
dc->caps.max_v_total = (1 << 15) - 1;
+ dc->caps.vtotal_limited_by_fp2 = true;
/* Color pipeline capabilities */
dc->caps.color.dpp.dcn_arch = 1;
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn401/dcn401_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn401/dcn401_resource.c
index db93bac247c0..c1ebc6b1c937 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn401/dcn401_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn401/dcn401_resource.c
@@ -726,6 +726,10 @@ static const struct dc_debug_options debug_defaults_drv = {
.disable_unbounded_requesting = false,
.enable_legacy_fast_update = false,
.dcc_meta_propagation_delay_us = 10,
+ .fams_version = {
+ .minor = 1,
+ .major = 2,
+ }, //v2.1
.fams2_config = {
.bits = {
.enable = true,
@@ -733,7 +737,7 @@ static const struct dc_debug_options debug_defaults_drv = {
.enable_stall_recovery = true,
}
},
- .force_cositing = CHROMA_COSITING_TOPLEFT + 1,
+ .force_cositing = CHROMA_COSITING_NONE + 1,
};
static struct dce_aux *dcn401_aux_engine_create(
@@ -1293,6 +1297,29 @@ static struct hpo_dp_link_encoder *dcn401_hpo_dp_link_encoder_create(
return &hpo_dp_enc31->base;
}
+static unsigned int dcn401_calc_num_avail_chans_for_mall(struct dc *dc, unsigned int num_chans)
+{
+ unsigned int num_available_chans = 1;
+
+ /* channels for MALL must be a power of 2 */
+ while (num_chans > 1) {
+ num_available_chans = (num_available_chans << 1);
+ num_chans = (num_chans >> 1);
+ }
+
+ /* cannot be odd */
+ num_available_chans &= ~1;
+
+ /* clamp to max available channels for MALL per ASIC */
+ if (ASICREV_IS_GC_12_0_0_A0(dc->ctx->asic_id.hw_internal_rev)) {
+ num_available_chans = num_available_chans > 16 ? 16 : num_available_chans;
+ } else if (ASICREV_IS_GC_12_0_1_A0(dc->ctx->asic_id.hw_internal_rev)) {
+ num_available_chans = num_available_chans > 8 ? 8 : num_available_chans;
+ }
+
+ return num_available_chans;
+}
+
static struct dce_hwseq *dcn401_hwseq_create(
struct dc_context *ctx)
{
@@ -1588,6 +1615,14 @@ static void dcn401_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *b
memcpy(dml2_opt, &dc->dml2_options, sizeof(dc->dml2_options));
+ /* re-calculate the available MALL size if required */
+ if (bw_params->num_channels > 0) {
+ dc->caps.max_cab_allocation_bytes = dcn401_calc_num_avail_chans_for_mall(
+ dc, bw_params->num_channels) *
+ dc->caps.mall_size_per_mem_channel * 1024 * 1024;
+ dc->caps.mall_size_total = dc->caps.max_cab_allocation_bytes;
+ }
+
DC_FP_START();
dcn401_update_bw_bounding_box_fpu(dc, bw_params);
@@ -1605,6 +1640,7 @@ static void dcn401_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *b
enum dc_status dcn401_patch_unknown_plane_state(struct dc_plane_state *plane_state)
{
+ plane_state->tiling_info.gfxversion = DcGfxAddr3;
plane_state->tiling_info.gfx_addr3.swizzle = DC_ADDR3_SW_64KB_2D;
return DC_OK;
}
@@ -1704,27 +1740,9 @@ static int dcn401_get_power_profile(const struct dc_state *context)
return dpm_level;
}
-static unsigned int dcn401_calc_num_avail_chans_for_mall(struct dc *dc, unsigned int num_chans)
+static unsigned int dcn401_get_vstartup_for_pipe(struct pipe_ctx *pipe_ctx)
{
- unsigned int num_available_chans = 1;
-
- /* channels for MALL must be a power of 2 */
- while (num_chans > 1) {
- num_available_chans = (num_available_chans << 1);
- num_chans = (num_chans >> 1);
- }
-
- /* cannot be odd */
- num_available_chans &= ~1;
-
- /* clamp to max available channels for MALL per ASIC */
- if (ASICREV_IS_GC_12_0_0_A0(dc->ctx->asic_id.hw_internal_rev)) {
- num_available_chans = num_available_chans > 16 ? 16 : num_available_chans;
- } else if (ASICREV_IS_GC_12_0_1_A0(dc->ctx->asic_id.hw_internal_rev)) {
- num_available_chans = num_available_chans > 8 ? 8 : num_available_chans;
- }
-
- return num_available_chans;
+ return pipe_ctx->global_sync.dcn4x.vstartup_lines;
}
static struct resource_funcs dcn401_res_pool_funcs = {
@@ -1754,6 +1772,7 @@ static struct resource_funcs dcn401_res_pool_funcs = {
.build_pipe_pix_clk_params = dcn401_build_pipe_pix_clk_params,
.calculate_mall_ways_from_bytes = dcn32_calculate_mall_ways_from_bytes,
.get_power_profile = dcn401_get_power_profile,
+ .get_vstartup_for_pipe = dcn401_get_vstartup_for_pipe
};
static uint32_t read_pipe_fuses(struct dc_context *ctx)
@@ -1864,6 +1883,7 @@ static bool dcn401_resource_construct(
dc->caps.extended_aux_timeout_support = true;
dc->caps.dmcub_support = true;
dc->caps.max_v_total = (1 << 15) - 1;
+ dc->caps.vtotal_limited_by_fp2 = true;
if (ASICREV_IS_GC_12_0_1_A0(dc->ctx->asic_id.hw_internal_rev))
dc->caps.dcc_plane_width_limit = 7680;
diff --git a/drivers/gpu/drm/amd/display/dc/spl/dc_spl.c b/drivers/gpu/drm/amd/display/dc/spl/dc_spl.c
index 73a65913cb12..38a9a0d68058 100644
--- a/drivers/gpu/drm/amd/display/dc/spl/dc_spl.c
+++ b/drivers/gpu/drm/amd/display/dc/spl/dc_spl.c
@@ -11,6 +11,41 @@
#define IDENTITY_RATIO(ratio) (spl_fixpt_u2d19(ratio) == (1 << 19))
#define MIN_VIEWPORT_SIZE 12
+static bool spl_is_yuv420(enum spl_pixel_format format)
+{
+ if ((format >= SPL_PIXEL_FORMAT_420BPP8) &&
+ (format <= SPL_PIXEL_FORMAT_420BPP10))
+ return true;
+
+ return false;
+}
+
+static bool spl_is_rgb8(enum spl_pixel_format format)
+{
+ if (format == SPL_PIXEL_FORMAT_ARGB8888)
+ return true;
+
+ return false;
+}
+
+static bool spl_is_video_format(enum spl_pixel_format format)
+{
+ if (format >= SPL_PIXEL_FORMAT_VIDEO_BEGIN
+ && format <= SPL_PIXEL_FORMAT_VIDEO_END)
+ return true;
+ else
+ return false;
+}
+
+static bool spl_is_subsampled_format(enum spl_pixel_format format)
+{
+ if (format >= SPL_PIXEL_FORMAT_SUBSAMPLED_BEGIN
+ && format <= SPL_PIXEL_FORMAT_SUBSAMPLED_END)
+ return true;
+ else
+ return false;
+}
+
static struct spl_rect intersect_rec(const struct spl_rect *r0, const struct spl_rect *r1)
{
struct spl_rect rec;
@@ -137,15 +172,32 @@ static struct spl_rect calculate_mpc_slice_in_timing_active(
struct spl_in *spl_in,
struct spl_rect *plane_clip_rec)
{
- int mpc_slice_count = spl_in->basic_in.mpc_combine_h;
- int mpc_slice_idx = spl_in->basic_in.mpc_combine_v;
+ bool use_recout_width_aligned =
+ spl_in->basic_in.num_h_slices_recout_width_align.use_recout_width_aligned;
+ int mpc_slice_count =
+ spl_in->basic_in.num_h_slices_recout_width_align.num_slices_recout_width.mpc_num_h_slices;
+ int recout_width_align =
+ spl_in->basic_in.num_h_slices_recout_width_align.num_slices_recout_width.mpc_recout_width_align;
+ int mpc_slice_idx = spl_in->basic_in.mpc_h_slice_index;
int epimo = mpc_slice_count - plane_clip_rec->width % mpc_slice_count - 1;
struct spl_rect mpc_rec;
- mpc_rec.width = plane_clip_rec->width / mpc_slice_count;
- mpc_rec.x = plane_clip_rec->x + mpc_rec.width * mpc_slice_idx;
- mpc_rec.height = plane_clip_rec->height;
- mpc_rec.y = plane_clip_rec->y;
+ if (use_recout_width_aligned) {
+ mpc_rec.width = recout_width_align;
+ if ((mpc_rec.width * (mpc_slice_idx + 1)) > plane_clip_rec->width) {
+ mpc_rec.width = plane_clip_rec->width % recout_width_align;
+ mpc_rec.x = plane_clip_rec->x + recout_width_align * mpc_slice_idx;
+ } else
+ mpc_rec.x = plane_clip_rec->x + mpc_rec.width * mpc_slice_idx;
+ mpc_rec.height = plane_clip_rec->height;
+ mpc_rec.y = plane_clip_rec->y;
+
+ } else {
+ mpc_rec.width = plane_clip_rec->width / mpc_slice_count;
+ mpc_rec.x = plane_clip_rec->x + mpc_rec.width * mpc_slice_idx;
+ mpc_rec.height = plane_clip_rec->height;
+ mpc_rec.y = plane_clip_rec->y;
+ }
SPL_ASSERT(mpc_slice_count == 1 ||
spl_in->basic_out.view_format != SPL_VIEW_3D_SIDE_BY_SIDE ||
mpc_rec.width % 2 == 0);
@@ -391,8 +443,7 @@ static void spl_calculate_scaling_ratios(struct spl_in *spl_in,
spl_scratch->scl_data.ratios.horz_c = spl_scratch->scl_data.ratios.horz;
spl_scratch->scl_data.ratios.vert_c = spl_scratch->scl_data.ratios.vert;
- if (spl_in->basic_in.format == SPL_PIXEL_FORMAT_420BPP8
- || spl_in->basic_in.format == SPL_PIXEL_FORMAT_420BPP10) {
+ if (spl_is_yuv420(spl_in->basic_in.format)) {
spl_scratch->scl_data.ratios.horz_c.value /= 2;
spl_scratch->scl_data.ratios.vert_c.value /= 2;
}
@@ -529,23 +580,6 @@ static void spl_calculate_init_and_vp(bool flip_scan_dir,
*vp_offset = src_size - *vp_offset - *vp_size;
}
-static bool spl_is_yuv420(enum spl_pixel_format format)
-{
- if ((format >= SPL_PIXEL_FORMAT_420BPP8) &&
- (format <= SPL_PIXEL_FORMAT_420BPP10))
- return true;
-
- return false;
-}
-
-static bool spl_is_rgb8(enum spl_pixel_format format)
-{
- if (format == SPL_PIXEL_FORMAT_ARGB8888)
- return true;
-
- return false;
-}
-
/*Calculate inits and viewport */
static void spl_calculate_inits_and_viewports(struct spl_in *spl_in,
struct spl_scratch *spl_scratch)
@@ -556,8 +590,7 @@ static void spl_calculate_inits_and_viewports(struct spl_in *spl_in,
struct spl_rect recout_clip_in_recout_dst;
struct spl_rect overlap_in_active_timing;
struct spl_rect odm_slice = calculate_odm_slice_in_timing_active(spl_in);
- int vpc_div = (spl_in->basic_in.format == SPL_PIXEL_FORMAT_420BPP8
- || spl_in->basic_in.format == SPL_PIXEL_FORMAT_420BPP10) ? 2 : 1;
+ int vpc_div = spl_is_subsampled_format(spl_in->basic_in.format) ? 2 : 1;
bool orthogonal_rotation, flip_vert_scan_dir, flip_horz_scan_dir;
struct spl_fixed31_32 init_adj_h = spl_fixpt_zero;
struct spl_fixed31_32 init_adj_v = spl_fixpt_zero;
@@ -585,12 +618,7 @@ static void spl_calculate_inits_and_viewports(struct spl_in *spl_in,
&flip_vert_scan_dir,
&flip_horz_scan_dir);
- if (orthogonal_rotation) {
- spl_swap(src.width, src.height);
- spl_swap(flip_vert_scan_dir, flip_horz_scan_dir);
- }
-
- if (spl_is_yuv420(spl_in->basic_in.format)) {
+ if (spl_is_subsampled_format(spl_in->basic_in.format)) {
/* this gives the direction of the cositing (negative will move
* left, right otherwise)
*/
@@ -598,15 +626,15 @@ static void spl_calculate_inits_and_viewports(struct spl_in *spl_in,
switch (spl_in->basic_in.cositing) {
- case CHROMA_COSITING_LEFT:
- init_adj_h = spl_fixpt_zero;
+ case CHROMA_COSITING_TOPLEFT:
+ init_adj_h = spl_fixpt_from_fraction(sign, 4);
init_adj_v = spl_fixpt_from_fraction(sign, 4);
break;
- case CHROMA_COSITING_NONE:
+ case CHROMA_COSITING_LEFT:
init_adj_h = spl_fixpt_from_fraction(sign, 4);
- init_adj_v = spl_fixpt_from_fraction(sign, 4);
+ init_adj_v = spl_fixpt_zero;
break;
- case CHROMA_COSITING_TOPLEFT:
+ case CHROMA_COSITING_NONE:
default:
init_adj_h = spl_fixpt_zero;
init_adj_v = spl_fixpt_zero;
@@ -614,6 +642,12 @@ static void spl_calculate_inits_and_viewports(struct spl_in *spl_in,
}
}
+ if (orthogonal_rotation) {
+ spl_swap(src.width, src.height);
+ spl_swap(flip_vert_scan_dir, flip_horz_scan_dir);
+ spl_swap(init_adj_h, init_adj_v);
+ }
+
spl_calculate_init_and_vp(
flip_horz_scan_dir,
recout_clip_in_recout_dst.x,
@@ -678,7 +712,7 @@ static void spl_handle_3d_recout(struct spl_in *spl_in, struct spl_rect *recout)
* since 3d is special and needs to calculate vp as if there is no recout offset
* This may break with rotation, good thing we aren't mixing hw rotation and 3d
*/
- if (spl_in->basic_in.mpc_combine_v) {
+ if (spl_in->basic_in.mpc_h_slice_index) {
SPL_ASSERT(spl_in->basic_in.rotation == SPL_ROTATION_ANGLE_0 ||
(spl_in->basic_out.view_format != SPL_VIEW_3D_TOP_AND_BOTTOM &&
spl_in->basic_out.view_format != SPL_VIEW_3D_SIDE_BY_SIDE));
@@ -698,24 +732,6 @@ static void spl_clamp_viewport(struct spl_rect *viewport)
viewport->width = MIN_VIEWPORT_SIZE;
}
-static bool spl_dscl_is_420_format(enum spl_pixel_format format)
-{
- if (format == SPL_PIXEL_FORMAT_420BPP8 ||
- format == SPL_PIXEL_FORMAT_420BPP10)
- return true;
- else
- return false;
-}
-
-static bool spl_dscl_is_video_format(enum spl_pixel_format format)
-{
- if (format >= SPL_PIXEL_FORMAT_VIDEO_BEGIN
- && format <= SPL_PIXEL_FORMAT_VIDEO_END)
- return true;
- else
- return false;
-}
-
static enum scl_mode spl_get_dscl_mode(const struct spl_in *spl_in,
const struct spl_scaler_data *data,
bool enable_isharp, bool enable_easf)
@@ -732,8 +748,8 @@ static enum scl_mode spl_get_dscl_mode(const struct spl_in *spl_in,
&& !enable_isharp)
return SCL_MODE_SCALING_444_BYPASS;
- if (!spl_dscl_is_420_format(pixel_format)) {
- if (spl_dscl_is_video_format(pixel_format))
+ if (!spl_is_subsampled_format(pixel_format)) {
+ if (spl_is_video_format(pixel_format))
return SCL_MODE_SCALING_444_YCBCR_ENABLE;
else
return SCL_MODE_SCALING_444_RGB_ENABLE;
@@ -756,7 +772,7 @@ static bool spl_choose_lls_policy(enum spl_pixel_format format,
enum spl_transfer_func_predefined tf_predefined_type,
enum linear_light_scaling *lls_pref)
{
- if (spl_is_yuv420(format)) {
+ if (spl_is_video_format(format)) {
*lls_pref = LLS_PREF_NO;
if ((tf_type == SPL_TF_TYPE_PREDEFINED) ||
(tf_type == SPL_TF_TYPE_DISTRIBUTED_POINTS))
@@ -815,7 +831,7 @@ static bool enable_easf(struct spl_in *spl_in, struct spl_scratch *spl_scratch)
/* Check if video is in fullscreen mode */
static bool spl_is_video_fullscreen(struct spl_in *spl_in)
{
- if (spl_is_yuv420(spl_in->basic_in.format) && spl_in->is_fullscreen)
+ if (spl_is_video_format(spl_in->basic_in.format) && spl_in->is_fullscreen)
return true;
return false;
}
@@ -846,10 +862,10 @@ static bool spl_get_isharp_en(struct spl_in *spl_in,
* Apply sharpness to RGB and YUV (NV12/P010)
* surfaces based on policy setting
*/
- if (!spl_is_yuv420(spl_in->basic_in.format) &&
+ if (!spl_is_video_format(spl_in->basic_in.format) &&
(spl_in->sharpen_policy == SHARPEN_YUV))
return enable_isharp;
- else if ((spl_is_yuv420(spl_in->basic_in.format) && !fullscreen) &&
+ else if ((spl_is_video_format(spl_in->basic_in.format) && !fullscreen) &&
(spl_in->sharpen_policy == SHARPEN_RGB_FULLSCREEN_YUV))
return enable_isharp;
else if (!spl_in->is_fullscreen &&
@@ -882,8 +898,8 @@ static void spl_get_taps_non_adaptive_scaler(
if (in_taps->v_taps == 0) {
if (spl_fixpt_ceil(spl_scratch->scl_data.ratios.vert) > 1)
- spl_scratch->scl_data.taps.v_taps = spl_min(spl_fixpt_ceil(spl_fixpt_mul_int(
- spl_scratch->scl_data.ratios.vert, 2)), 8);
+ spl_scratch->scl_data.taps.v_taps = spl_min(2 * spl_fixpt_ceil(
+ spl_scratch->scl_data.ratios.vert), 8);
else
spl_scratch->scl_data.taps.v_taps = 4;
} else
@@ -891,8 +907,8 @@ static void spl_get_taps_non_adaptive_scaler(
if (in_taps->v_taps_c == 0) {
if (spl_fixpt_ceil(spl_scratch->scl_data.ratios.vert_c) > 1)
- spl_scratch->scl_data.taps.v_taps_c = spl_min(spl_fixpt_ceil(spl_fixpt_mul_int(
- spl_scratch->scl_data.ratios.vert_c, 2)), 8);
+ spl_scratch->scl_data.taps.v_taps_c = spl_min(2 * spl_fixpt_ceil(
+ spl_scratch->scl_data.ratios.vert_c), 8);
else
spl_scratch->scl_data.taps.v_taps_c = 4;
} else
@@ -932,7 +948,7 @@ static bool spl_get_optimal_number_of_taps(
int min_taps_y, min_taps_c;
enum lb_memory_config lb_config;
bool skip_easf = false;
- bool is_ycbcr = spl_dscl_is_video_format(spl_in->basic_in.format);
+ bool is_subsampled = spl_is_subsampled_format(spl_in->basic_in.format);
if (spl_scratch->scl_data.viewport.width > spl_scratch->scl_data.h_active &&
max_downscale_src_width != 0 &&
@@ -964,7 +980,7 @@ static bool spl_get_optimal_number_of_taps(
if (skip_easf)
spl_get_taps_non_adaptive_scaler(spl_scratch, in_taps);
else {
- if (spl_is_yuv420(spl_in->basic_in.format)) {
+ if (spl_is_video_format(spl_in->basic_in.format)) {
spl_scratch->scl_data.taps.h_taps = 6;
spl_scratch->scl_data.taps.v_taps = 6;
spl_scratch->scl_data.taps.h_taps_c = 4;
@@ -982,8 +998,7 @@ static bool spl_get_optimal_number_of_taps(
min_taps_c = spl_fixpt_ceil(spl_scratch->scl_data.ratios.vert_c);
/* Use LB_MEMORY_CONFIG_3 for 4:2:0 */
- if ((spl_in->basic_in.format == SPL_PIXEL_FORMAT_420BPP8)
- || (spl_in->basic_in.format == SPL_PIXEL_FORMAT_420BPP10))
+ if (spl_is_yuv420(spl_in->basic_in.format))
lb_config = LB_MEMORY_CONFIG_3;
else
lb_config = LB_MEMORY_CONFIG_0;
@@ -1039,13 +1054,11 @@ static bool spl_get_optimal_number_of_taps(
if (spl_scratch->scl_data.taps.h_taps_c == 5)
spl_scratch->scl_data.taps.h_taps_c = 4;
- if (spl_is_yuv420(spl_in->basic_in.format)) {
- if ((spl_scratch->scl_data.taps.h_taps <= 4) ||
- (spl_scratch->scl_data.taps.h_taps_c <= 3)) {
+ if (spl_is_video_format(spl_in->basic_in.format)) {
+ if (spl_scratch->scl_data.taps.h_taps <= 4) {
*enable_easf_v = false;
*enable_easf_h = false;
- } else if ((spl_scratch->scl_data.taps.v_taps <= 3) ||
- (spl_scratch->scl_data.taps.v_taps_c <= 3)) {
+ } else if (spl_scratch->scl_data.taps.v_taps <= 3) {
*enable_easf_v = false;
*enable_easf_h = true;
} else {
@@ -1086,10 +1099,10 @@ static bool spl_get_optimal_number_of_taps(
spl_scratch->scl_data.taps.h_taps = 1;
spl_scratch->scl_data.taps.v_taps = 1;
- if (IDENTITY_RATIO(spl_scratch->scl_data.ratios.horz_c) && !is_ycbcr)
+ if (IDENTITY_RATIO(spl_scratch->scl_data.ratios.horz_c) && !is_subsampled)
spl_scratch->scl_data.taps.h_taps_c = 1;
- if (IDENTITY_RATIO(spl_scratch->scl_data.ratios.vert_c) && !is_ycbcr)
+ if (IDENTITY_RATIO(spl_scratch->scl_data.ratios.vert_c) && !is_subsampled)
spl_scratch->scl_data.taps.v_taps_c = 1;
*enable_easf_v = false;
@@ -1103,11 +1116,11 @@ static bool spl_get_optimal_number_of_taps(
(IDENTITY_RATIO(spl_scratch->scl_data.ratios.vert)))
spl_scratch->scl_data.taps.v_taps = 1;
- if ((!*enable_easf_h) && !is_ycbcr &&
+ if ((!*enable_easf_h) && !is_subsampled &&
(IDENTITY_RATIO(spl_scratch->scl_data.ratios.horz_c)))
spl_scratch->scl_data.taps.h_taps_c = 1;
- if ((!*enable_easf_v) && !is_ycbcr &&
+ if ((!*enable_easf_v) && !is_subsampled &&
(IDENTITY_RATIO(spl_scratch->scl_data.ratios.vert_c)))
spl_scratch->scl_data.taps.v_taps_c = 1;
}
@@ -1118,7 +1131,7 @@ static bool spl_get_optimal_number_of_taps(
static void spl_set_black_color_data(enum spl_pixel_format format,
struct scl_black_color *scl_black_color)
{
- bool ycbcr = spl_dscl_is_video_format(format);
+ bool ycbcr = spl_is_video_format(format);
if (ycbcr) {
scl_black_color->offset_rgb_y = BLACK_OFFSET_RGB_Y;
scl_black_color->offset_rgb_cbcr = BLACK_OFFSET_CBCR;
@@ -1585,7 +1598,7 @@ static void spl_set_easf_data(struct spl_scratch *spl_scratch, struct spl_out *s
0x0; // fp1.5.10, C3 coefficient
}
- if (spl_is_yuv420(format)) { /* TODO: 0 = RGB, 1 = YUV */
+ if (spl_is_subsampled_format(format)) { /* TODO: 0 = RGB, 1 = YUV */
dscl_prog_data->easf_matrix_mode = 1;
/*
* 2-bit, BF3 chroma mode correction calculation mode
diff --git a/drivers/gpu/drm/amd/display/dc/spl/dc_spl_types.h b/drivers/gpu/drm/amd/display/dc/spl/dc_spl_types.h
index 55d557df4aa5..467af9dd90de 100644
--- a/drivers/gpu/drm/amd/display/dc/spl/dc_spl_types.h
+++ b/drivers/gpu/drm/amd/display/dc/spl/dc_spl_types.h
@@ -63,13 +63,13 @@ enum spl_pixel_format {
SPL_PIXEL_FORMAT_420BPP8,
SPL_PIXEL_FORMAT_420BPP10,
/*end of pixel format definition*/
- SPL_PIXEL_FORMAT_INVALID,
- SPL_PIXEL_FORMAT_422BPP8,
- SPL_PIXEL_FORMAT_422BPP10,
SPL_PIXEL_FORMAT_GRPH_BEGIN = SPL_PIXEL_FORMAT_INDEX8,
SPL_PIXEL_FORMAT_GRPH_END = SPL_PIXEL_FORMAT_FP16,
+ SPL_PIXEL_FORMAT_SUBSAMPLED_BEGIN = SPL_PIXEL_FORMAT_420BPP8,
+ SPL_PIXEL_FORMAT_SUBSAMPLED_END = SPL_PIXEL_FORMAT_420BPP10,
SPL_PIXEL_FORMAT_VIDEO_BEGIN = SPL_PIXEL_FORMAT_420BPP8,
SPL_PIXEL_FORMAT_VIDEO_END = SPL_PIXEL_FORMAT_420BPP10,
+ SPL_PIXEL_FORMAT_INVALID,
SPL_PIXEL_FORMAT_UNKNOWN
};
@@ -436,8 +436,14 @@ struct basic_in {
struct spl_rect clip_rect; // Clip rect
enum spl_rotation_angle rotation; // Rotation
bool horizontal_mirror; // Horizontal mirror
- int mpc_combine_h; // MPC Horizontal Combine Factor (split_count)
- int mpc_combine_v; // MPC Vertical Combine Factor (split_idx)
+ struct { // previous mpc_combine_h - split count
+ bool use_recout_width_aligned;
+ union {
+ int mpc_num_h_slices;
+ int mpc_recout_width_align;
+ } num_slices_recout_width;
+ } num_h_slices_recout_width_align;
+ int mpc_h_slice_index; // previous mpc_combine_v - split_idx
// Inputs for adaptive scaler - TODO
enum spl_transfer_func_type tf_type; /* Transfer function type */
enum spl_transfer_func_predefined tf_predefined_type; /* Transfer function predefined type */