diff options
Diffstat (limited to 'drivers/gpu/drm/amd/display/dc/dml')
19 files changed, 12788 insertions, 0 deletions
diff --git a/drivers/gpu/drm/amd/display/dc/dml/Makefile b/drivers/gpu/drm/amd/display/dc/dml/Makefile new file mode 100644 index 000000000000..87bab8e8139f --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/dml/Makefile @@ -0,0 +1,22 @@ +# +# Makefile for the 'utils' sub-component of DAL. +# It provides the general basic services required by other DAL +# subcomponents. + +CFLAGS_display_mode_vba.o := -mhard-float -msse -mpreferred-stack-boundary=4 +CFLAGS_display_mode_lib.o := -mhard-float -msse -mpreferred-stack-boundary=4 +CFLAGS_display_pipe_clocks.o := -mhard-float -msse -mpreferred-stack-boundary=4 +CFLAGS_display_rq_dlg_calc.o := -mhard-float -msse -mpreferred-stack-boundary=4 +CFLAGS_dml1_display_rq_dlg_calc.o := -mhard-float -msse -mpreferred-stack-boundary=4 +CFLAGS_display_rq_dlg_helpers.o := -mhard-float -msse -mpreferred-stack-boundary=4 +CFLAGS_soc_bounding_box.o := -mhard-float -msse -mpreferred-stack-boundary=4 +CFLAGS_dml_common_defs.o := -mhard-float -msse -mpreferred-stack-boundary=4 + + +DML = display_mode_lib.o display_rq_dlg_calc.o \ + display_rq_dlg_helpers.o dml1_display_rq_dlg_calc.o \ + soc_bounding_box.o dml_common_defs.o display_mode_vba.o + +AMD_DAL_DML = $(addprefix $(AMDDALPATH)/dc/dml/,$(DML)) + +AMD_DISPLAY_FILES += $(AMD_DAL_DML) diff --git a/drivers/gpu/drm/amd/display/dc/dml/dc_features.h b/drivers/gpu/drm/amd/display/dc/dml/dc_features.h new file mode 100644 index 000000000000..ea4cde952f4f --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/dml/dc_features.h @@ -0,0 +1,559 @@ +/* + * Copyright 2017 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 __DC_FEATURES_H__ +#define __DC_FEATURES_H__ + +// local features +#define DC__PRESENT 1 +#define DC__PRESENT__1 1 +#define DC__NUM_DPP 4 +#define DC__VOLTAGE_STATES 7 +#define DC__NUM_DPP__4 1 +#define DC__NUM_DPP__0_PRESENT 1 +#define DC__NUM_DPP__1_PRESENT 1 +#define DC__NUM_DPP__2_PRESENT 1 +#define DC__NUM_DPP__3_PRESENT 1 +#define DC__NUM_DPP__MAX 8 +#define DC__NUM_DPP__MAX__8 1 +#define DC__PIPE_10BIT 0 +#define DC__PIPE_10BIT__0 1 +#define DC__PIPE_10BIT__MAX 1 +#define DC__PIPE_10BIT__MAX__1 1 +#define DC__NUM_OPP 4 +#define DC__NUM_OPP__4 1 +#define DC__NUM_OPP__0_PRESENT 1 +#define DC__NUM_OPP__1_PRESENT 1 +#define DC__NUM_OPP__2_PRESENT 1 +#define DC__NUM_OPP__3_PRESENT 1 +#define DC__NUM_OPP__MAX 6 +#define DC__NUM_OPP__MAX__6 1 +#define DC__NUM_DSC 0 +#define DC__NUM_DSC__0 1 +#define DC__NUM_DSC__MAX 6 +#define DC__NUM_DSC__MAX__6 1 +#define DC__NUM_ABM 1 +#define DC__NUM_ABM__1 1 +#define DC__NUM_ABM__0_PRESENT 1 +#define DC__NUM_ABM__MAX 2 +#define DC__NUM_ABM__MAX__2 1 +#define DC__ODM_PRESENT 0 +#define DC__ODM_PRESENT__0 1 +#define DC__NUM_OTG 4 +#define DC__NUM_OTG__4 1 +#define DC__NUM_OTG__0_PRESENT 1 +#define DC__NUM_OTG__1_PRESENT 1 +#define DC__NUM_OTG__2_PRESENT 1 +#define DC__NUM_OTG__3_PRESENT 1 +#define DC__NUM_OTG__MAX 6 +#define DC__NUM_OTG__MAX__6 1 +#define DC__NUM_DWB 2 +#define DC__NUM_DWB__2 1 +#define DC__NUM_DWB__0_PRESENT 1 +#define DC__NUM_DWB__1_PRESENT 1 +#define DC__NUM_DWB__MAX 2 +#define DC__NUM_DWB__MAX__2 1 +#define DC__NUM_DIG 4 +#define DC__NUM_DIG__4 1 +#define DC__NUM_DIG__0_PRESENT 1 +#define DC__NUM_DIG__1_PRESENT 1 +#define DC__NUM_DIG__2_PRESENT 1 +#define DC__NUM_DIG__3_PRESENT 1 +#define DC__NUM_DIG__MAX 6 +#define DC__NUM_DIG__MAX__6 1 +#define DC__NUM_AUX 4 +#define DC__NUM_AUX__4 1 +#define DC__NUM_AUX__0_PRESENT 1 +#define DC__NUM_AUX__1_PRESENT 1 +#define DC__NUM_AUX__2_PRESENT 1 +#define DC__NUM_AUX__3_PRESENT 1 +#define DC__NUM_AUX__MAX 6 +#define DC__NUM_AUX__MAX__6 1 +#define DC__NUM_AUDIO_STREAMS 4 +#define DC__NUM_AUDIO_STREAMS__4 1 +#define DC__NUM_AUDIO_STREAMS__0_PRESENT 1 +#define DC__NUM_AUDIO_STREAMS__1_PRESENT 1 +#define DC__NUM_AUDIO_STREAMS__2_PRESENT 1 +#define DC__NUM_AUDIO_STREAMS__3_PRESENT 1 +#define DC__NUM_AUDIO_STREAMS__MAX 8 +#define DC__NUM_AUDIO_STREAMS__MAX__8 1 +#define DC__NUM_AUDIO_ENDPOINTS 6 +#define DC__NUM_AUDIO_ENDPOINTS__6 1 +#define DC__NUM_AUDIO_ENDPOINTS__0_PRESENT 1 +#define DC__NUM_AUDIO_ENDPOINTS__1_PRESENT 1 +#define DC__NUM_AUDIO_ENDPOINTS__2_PRESENT 1 +#define DC__NUM_AUDIO_ENDPOINTS__3_PRESENT 1 +#define DC__NUM_AUDIO_ENDPOINTS__4_PRESENT 1 +#define DC__NUM_AUDIO_ENDPOINTS__5_PRESENT 1 +#define DC__NUM_AUDIO_ENDPOINTS__MAX 8 +#define DC__NUM_AUDIO_ENDPOINTS__MAX__8 1 +#define DC__NUM_AUDIO_INPUT_STREAMS 0 +#define DC__NUM_AUDIO_INPUT_STREAMS__0 1 +#define DC__NUM_AUDIO_INPUT_STREAMS__MAX 8 +#define DC__NUM_AUDIO_INPUT_STREAMS__MAX__8 1 +#define DC__NUM_AUDIO_INPUT_ENDPOINTS 0 +#define DC__NUM_AUDIO_INPUT_ENDPOINTS__0 1 +#define DC__NUM_AUDIO_INPUT_ENDPOINTS__MAX 8 +#define DC__NUM_AUDIO_INPUT_ENDPOINTS__MAX__8 1 +#define DC__NUM_CURSOR 1 +#define DC__NUM_CURSOR__1 1 +#define DC__NUM_CURSOR__0_PRESENT 1 +#define DC__NUM_CURSOR__MAX 2 +#define DC__NUM_CURSOR__MAX__2 1 +#define DC__DIGITAL_BYPASS_PRESENT 0 +#define DC__DIGITAL_BYPASS_PRESENT__0 1 +#define DC__HCID_HWMAJVER 1 +#define DC__HCID_HWMAJVER__1 1 +#define DC__HCID_HWMINVER 0 +#define DC__HCID_HWMINVER__0 1 +#define DC__HCID_HWREV 0 +#define DC__HCID_HWREV__0 1 +#define DC__ROMSTRAP_PRESENT 0 +#define DC__ROMSTRAP_PRESENT__0 1 +#define DC__NUM_RBBMIF_DECODES 30 +#define DC__NUM_RBBMIF_DECODES__30 1 +#define DC__NUM_DBG_REGS 36 +#define DC__NUM_DBG_REGS__36 1 +#define DC__NUM_PIPES_UNDERLAY 0 +#define DC__NUM_PIPES_UNDERLAY__0 1 +#define DC__NUM_PIPES_UNDERLAY__MAX 2 +#define DC__NUM_PIPES_UNDERLAY__MAX__2 1 +#define DC__NUM_VCE_ENGINE 1 +#define DC__NUM_VCE_ENGINE__1 1 +#define DC__NUM_VCE_ENGINE__0_PRESENT 1 +#define DC__NUM_VCE_ENGINE__MAX 2 +#define DC__NUM_VCE_ENGINE__MAX__2 1 +#define DC__OTG_EXTERNAL_SYNC_PRESENT 0 +#define DC__OTG_EXTERNAL_SYNC_PRESENT__0 1 +#define DC__OTG_CRC_PRESENT 1 +#define DC__OTG_CRC_PRESENT__1 1 +#define DC__VIP_PRESENT 0 +#define DC__VIP_PRESENT__0 1 +#define DC__DTMTEST_PRESENT 0 +#define DC__DTMTEST_PRESENT__0 1 +#define DC__POWER_GATE_PRESENT 1 +#define DC__POWER_GATE_PRESENT__1 1 +#define DC__MEM_PG 1 +#define DC__MEM_PG__1 1 +#define DC__FMT_SRC_SEL_PRESENT 0 +#define DC__FMT_SRC_SEL_PRESENT__0 1 +#define DC__DIG_FEATURES__HDMI_PRESENT 1 +#define DC__DIG_FEATURES__HDMI_PRESENT__1 1 +#define DC__DIG_FEATURES__DP_PRESENT 1 +#define DC__DIG_FEATURES__DP_PRESENT__1 1 +#define DC__DIG_FEATURES__DP_MST_PRESENT 1 +#define DC__DIG_FEATURES__DP_MST_PRESENT__1 1 +#define DC__DIG_LP_FEATURES__HDMI_PRESENT 0 +#define DC__DIG_LP_FEATURES__HDMI_PRESENT__0 1 +#define DC__DIG_LP_FEATURES__DP_PRESENT 1 +#define DC__DIG_LP_FEATURES__DP_PRESENT__1 1 +#define DC__DIG_LP_FEATURES__DP_MST_PRESENT 0 +#define DC__DIG_LP_FEATURES__DP_MST_PRESENT__0 1 +#define DC__DIG_RESYNC_FIFO_SIZE 14 +#define DC__DIG_RESYNC_FIFO_SIZE__14 1 +#define DC__DIG_RESYNC_FIFO_SIZE__0_PRESENT 1 +#define DC__DIG_RESYNC_FIFO_SIZE__1_PRESENT 1 +#define DC__DIG_RESYNC_FIFO_SIZE__2_PRESENT 1 +#define DC__DIG_RESYNC_FIFO_SIZE__3_PRESENT 1 +#define DC__DIG_RESYNC_FIFO_SIZE__4_PRESENT 1 +#define DC__DIG_RESYNC_FIFO_SIZE__5_PRESENT 1 +#define DC__DIG_RESYNC_FIFO_SIZE__6_PRESENT 1 +#define DC__DIG_RESYNC_FIFO_SIZE__7_PRESENT 1 +#define DC__DIG_RESYNC_FIFO_SIZE__8_PRESENT 1 +#define DC__DIG_RESYNC_FIFO_SIZE__9_PRESENT 1 +#define DC__DIG_RESYNC_FIFO_SIZE__10_PRESENT 1 +#define DC__DIG_RESYNC_FIFO_SIZE__11_PRESENT 1 +#define DC__DIG_RESYNC_FIFO_SIZE__12_PRESENT 1 +#define DC__DIG_RESYNC_FIFO_SIZE__13_PRESENT 1 +#define DC__DIG_RESYNC_FIFO_SIZE__MAX 16 +#define DC__DIG_RESYNC_FIFO_SIZE__MAX__16 1 +#define DC__DAC_RESYNC_FIFO_SIZE 12 +#define DC__DAC_RESYNC_FIFO_SIZE__12 1 +#define DC__DAC_RESYNC_FIFO_SIZE__0_PRESENT 1 +#define DC__DAC_RESYNC_FIFO_SIZE__1_PRESENT 1 +#define DC__DAC_RESYNC_FIFO_SIZE__2_PRESENT 1 +#define DC__DAC_RESYNC_FIFO_SIZE__3_PRESENT 1 +#define DC__DAC_RESYNC_FIFO_SIZE__4_PRESENT 1 +#define DC__DAC_RESYNC_FIFO_SIZE__5_PRESENT 1 +#define DC__DAC_RESYNC_FIFO_SIZE__6_PRESENT 1 +#define DC__DAC_RESYNC_FIFO_SIZE__7_PRESENT 1 +#define DC__DAC_RESYNC_FIFO_SIZE__8_PRESENT 1 +#define DC__DAC_RESYNC_FIFO_SIZE__9_PRESENT 1 +#define DC__DAC_RESYNC_FIFO_SIZE__10_PRESENT 1 +#define DC__DAC_RESYNC_FIFO_SIZE__11_PRESENT 1 +#define DC__DAC_RESYNC_FIFO_SIZE__MAX 16 +#define DC__DAC_RESYNC_FIFO_SIZE__MAX__16 1 +#define DC__DVO_RESYNC_FIFO_SIZE 12 +#define DC__DVO_RESYNC_FIFO_SIZE__12 1 +#define DC__DVO_RESYNC_FIFO_SIZE__0_PRESENT 1 +#define DC__DVO_RESYNC_FIFO_SIZE__1_PRESENT 1 +#define DC__DVO_RESYNC_FIFO_SIZE__2_PRESENT 1 +#define DC__DVO_RESYNC_FIFO_SIZE__3_PRESENT 1 +#define DC__DVO_RESYNC_FIFO_SIZE__4_PRESENT 1 +#define DC__DVO_RESYNC_FIFO_SIZE__5_PRESENT 1 +#define DC__DVO_RESYNC_FIFO_SIZE__6_PRESENT 1 +#define DC__DVO_RESYNC_FIFO_SIZE__7_PRESENT 1 +#define DC__DVO_RESYNC_FIFO_SIZE__8_PRESENT 1 +#define DC__DVO_RESYNC_FIFO_SIZE__9_PRESENT 1 +#define DC__DVO_RESYNC_FIFO_SIZE__10_PRESENT 1 +#define DC__DVO_RESYNC_FIFO_SIZE__11_PRESENT 1 +#define DC__DVO_RESYNC_FIFO_SIZE__MAX 16 +#define DC__DVO_RESYNC_FIFO_SIZE__MAX__16 1 +#define DC__MEM_CDC_PRESENT 1 +#define DC__MEM_CDC_PRESENT__1 1 +#define DC__NUM_HPD 4 +#define DC__NUM_HPD__4 1 +#define DC__NUM_HPD__0_PRESENT 1 +#define DC__NUM_HPD__1_PRESENT 1 +#define DC__NUM_HPD__2_PRESENT 1 +#define DC__NUM_HPD__3_PRESENT 1 +#define DC__NUM_HPD__MAX 6 +#define DC__NUM_HPD__MAX__6 1 +#define DC__NUM_DDC_PAIRS 4 +#define DC__NUM_DDC_PAIRS__4 1 +#define DC__NUM_DDC_PAIRS__0_PRESENT 1 +#define DC__NUM_DDC_PAIRS__1_PRESENT 1 +#define DC__NUM_DDC_PAIRS__2_PRESENT 1 +#define DC__NUM_DDC_PAIRS__3_PRESENT 1 +#define DC__NUM_DDC_PAIRS__MAX 6 +#define DC__NUM_DDC_PAIRS__MAX__6 1 +#define DC__NUM_AUDIO_PLL 0 +#define DC__NUM_AUDIO_PLL__0 1 +#define DC__NUM_AUDIO_PLL__MAX 2 +#define DC__NUM_AUDIO_PLL__MAX__2 1 +#define DC__NUM_PIXEL_PLL 1 +#define DC__NUM_PIXEL_PLL__1 1 +#define DC__NUM_PIXEL_PLL__0_PRESENT 1 +#define DC__NUM_PIXEL_PLL__MAX 4 +#define DC__NUM_PIXEL_PLL__MAX__4 1 +#define DC__NUM_CASCADED_PLL 0 +#define DC__NUM_CASCADED_PLL__0 1 +#define DC__NUM_CASCADED_PLL__MAX 3 +#define DC__NUM_CASCADED_PLL__MAX__3 1 +#define DC__PIXCLK_FROM_PHYPLL 1 +#define DC__PIXCLK_FROM_PHYPLL__1 1 +#define DC__NB_STUTTER_MODE_PRESENT 0 +#define DC__NB_STUTTER_MODE_PRESENT__0 1 +#define DC__I2S0_AND_SPDIF0_PRESENT 0 +#define DC__I2S0_AND_SPDIF0_PRESENT__0 1 +#define DC__I2S1_PRESENT 0 +#define DC__I2S1_PRESENT__0 1 +#define DC__SPDIF1_PRESENT 0 +#define DC__SPDIF1_PRESENT__0 1 +#define DC__DSI_PRESENT 0 +#define DC__DSI_PRESENT__0 1 +#define DC__DACA_PRESENT 0 +#define DC__DACA_PRESENT__0 1 +#define DC__DACB_PRESENT 0 +#define DC__DACB_PRESENT__0 1 +#define DC__NUM_PIPES 4 +#define DC__NUM_PIPES__4 1 +#define DC__NUM_PIPES__0_PRESENT 1 +#define DC__NUM_PIPES__1_PRESENT 1 +#define DC__NUM_PIPES__2_PRESENT 1 +#define DC__NUM_PIPES__3_PRESENT 1 +#define DC__NUM_PIPES__MAX 6 +#define DC__NUM_PIPES__MAX__6 1 +#define DC__NUM_DIG_LP 0 +#define DC__NUM_DIG_LP__0 1 +#define DC__NUM_DIG_LP__MAX 2 +#define DC__NUM_DIG_LP__MAX__2 1 +#define DC__DPDEBUG_PRESENT 0 +#define DC__DPDEBUG_PRESENT__0 1 +#define DC__DISPLAY_WB_PRESENT 1 +#define DC__DISPLAY_WB_PRESENT__1 1 +#define DC__NUM_CWB 0 +#define DC__NUM_CWB__0 1 +#define DC__NUM_CWB__MAX 2 +#define DC__NUM_CWB__MAX__2 1 +#define DC__MVP_PRESENT 0 +#define DC__MVP_PRESENT__0 1 +#define DC__DVO_PRESENT 0 +#define DC__DVO_PRESENT__0 1 +#define DC__ABM_PRESENT 0 +#define DC__ABM_PRESENT__0 1 +#define DC__BPHYC_PLL_PRESENT 0 +#define DC__BPHYC_PLL_PRESENT__0 1 +#define DC__BPHYC_UNIPHY_PRESENT 0 +#define DC__BPHYC_UNIPHY_PRESENT__0 1 +#define DC__PHY_BROADCAST_PRESENT 0 +#define DC__PHY_BROADCAST_PRESENT__0 1 +#define DC__NUM_OF_DCRX_SD 0 +#define DC__NUM_OF_DCRX_SD__0 1 +#define DC__DVO_17BIT_MAPPING 0 +#define DC__DVO_17BIT_MAPPING__0 1 +#define DC__AVSYNC_PRESENT 0 +#define DC__AVSYNC_PRESENT__0 1 +#define DC__NUM_OF_DCRX_PORTS 0 +#define DC__NUM_OF_DCRX_PORTS__0 1 +#define DC__NUM_OF_DCRX_PORTS__MAX 1 +#define DC__NUM_OF_DCRX_PORTS__MAX__1 1 +#define DC__NUM_PHY 4 +#define DC__NUM_PHY__4 1 +#define DC__NUM_PHY__0_PRESENT 1 +#define DC__NUM_PHY__1_PRESENT 1 +#define DC__NUM_PHY__2_PRESENT 1 +#define DC__NUM_PHY__3_PRESENT 1 +#define DC__NUM_PHY__MAX 7 +#define DC__NUM_PHY__MAX__7 1 +#define DC__NUM_PHY_LP 0 +#define DC__NUM_PHY_LP__0 1 +#define DC__NUM_PHY_LP__MAX 2 +#define DC__NUM_PHY_LP__MAX__2 1 +#define DC__SYNC_CELL vid_sync_gf14lpp +#define DC__SYNC_CELL__VID_SYNC_GF14LPP 1 +#define DC__USE_NEW_VSS 1 +#define DC__USE_NEW_VSS__1 1 +#define DC__SYNC_CELL_DISPCLK_NUM_LATCHES 6 +#define DC__SYNC_CELL_DISPCLK_NUM_LATCHES__6 1 +#define DC__SYNC_CELL_DVOCLK_NUM_LATCHES 6 +#define DC__SYNC_CELL_DVOCLK_NUM_LATCHES__6 1 +#define DC__SYNC_CELL_PIXCLK_NUM_LATCHES 6 +#define DC__SYNC_CELL_PIXCLK_NUM_LATCHES__6 1 +#define DC__SYNC_CELL_SYMCLK_NUM_LATCHES 6 +#define DC__SYNC_CELL_SYMCLK_NUM_LATCHES__6 1 +#define DC__SYNC_CELL_DPPCLK_NUM_LATCHES 6 +#define DC__SYNC_CELL_DPPCLK_NUM_LATCHES__6 1 +#define DC__SYNC_CELL_DPREFCLK_NUM_LATCHES 6 +#define DC__SYNC_CELL_DPREFCLK_NUM_LATCHES__6 1 +#define DC__SYNC_CELL_REFCLK_NUM_LATCHES 6 +#define DC__SYNC_CELL_REFCLK_NUM_LATCHES__6 1 +#define DC__SYNC_CELL_PCIE_REFCLK_NUM_LATCHES 6 +#define DC__SYNC_CELL_PCIE_REFCLK_NUM_LATCHES__6 1 +#define DC__SYNC_CELL_MVPCLK_NUM_LATCHES 6 +#define DC__SYNC_CELL_MVPCLK_NUM_LATCHES__6 1 +#define DC__SYNC_CELL_SCLK_NUM_LATCHES 6 +#define DC__SYNC_CELL_SCLK_NUM_LATCHES__6 1 +#define DC__SYNC_CELL_DCEFCLK_NUM_LATCHES 6 +#define DC__SYNC_CELL_DCEFCLK_NUM_LATCHES__6 1 +#define DC__SYNC_CELL_AMCLK_NUM_LATCHES 6 +#define DC__SYNC_CELL_AMCLK_NUM_LATCHES__6 1 +#define DC__SYNC_CELL_DSICLK_NUM_LATCHES 6 +#define DC__SYNC_CELL_DSICLK_NUM_LATCHES__6 1 +#define DC__SYNC_CELL_BYTECLK_NUM_LATCHES 6 +#define DC__SYNC_CELL_BYTECLK_NUM_LATCHES__6 1 +#define DC__SYNC_CELL_ESCCLK_NUM_LATCHES 6 +#define DC__SYNC_CELL_ESCCLK_NUM_LATCHES__6 1 +#define DC__SYNC_CELL_DB_CLK_NUM_LATCHES 6 +#define DC__SYNC_CELL_DB_CLK_NUM_LATCHES__6 1 +#define UNIPHYA_PRESENT 1 +#define UNIPHYA_PRESENT__1 1 +#define DC__UNIPHYA_PRESENT 1 +#define DC__UNIPHYA_PRESENT__1 1 +#define UNIPHYB_PRESENT 1 +#define UNIPHYB_PRESENT__1 1 +#define DC__UNIPHYB_PRESENT 1 +#define DC__UNIPHYB_PRESENT__1 1 +#define UNIPHYC_PRESENT 1 +#define UNIPHYC_PRESENT__1 1 +#define DC__UNIPHYC_PRESENT 1 +#define DC__UNIPHYC_PRESENT__1 1 +#define UNIPHYD_PRESENT 1 +#define UNIPHYD_PRESENT__1 1 +#define DC__UNIPHYD_PRESENT 1 +#define DC__UNIPHYD_PRESENT__1 1 +#define UNIPHYE_PRESENT 0 +#define UNIPHYE_PRESENT__0 1 +#define DC__UNIPHYE_PRESENT 0 +#define DC__UNIPHYE_PRESENT__0 1 +#define UNIPHYF_PRESENT 0 +#define UNIPHYF_PRESENT__0 1 +#define DC__UNIPHYF_PRESENT 0 +#define DC__UNIPHYF_PRESENT__0 1 +#define UNIPHYG_PRESENT 0 +#define UNIPHYG_PRESENT__0 1 +#define DC__UNIPHYG_PRESENT 0 +#define DC__UNIPHYG_PRESENT__0 1 +#define DC__TMDS_LINK tmds_link_dual +#define DC__TMDS_LINK__TMDS_LINK_DUAL 1 +#define DC__WBSCL_PIXBW 8 +#define DC__WBSCL_PIXBW__8 1 +#define DC__DWB_CSC_PRESENT 0 +#define DC__DWB_CSC_PRESENT__0 1 +#define DC__DWB_LUMA_SCL_PRESENT 0 +#define DC__DWB_LUMA_SCL_PRESENT__0 1 +#define DC__DENTIST_INTERFACE_PRESENT 1 +#define DC__DENTIST_INTERFACE_PRESENT__1 1 +#define DC__GENERICA_PRESENT 1 +#define DC__GENERICA_PRESENT__1 1 +#define DC__GENERICB_PRESENT 1 +#define DC__GENERICB_PRESENT__1 1 +#define DC__GENERICC_PRESENT 0 +#define DC__GENERICC_PRESENT__0 1 +#define DC__GENERICD_PRESENT 0 +#define DC__GENERICD_PRESENT__0 1 +#define DC__GENERICE_PRESENT 0 +#define DC__GENERICE_PRESENT__0 1 +#define DC__GENERICF_PRESENT 0 +#define DC__GENERICF_PRESENT__0 1 +#define DC__GENERICG_PRESENT 0 +#define DC__GENERICG_PRESENT__0 1 +#define DC__UNIPHY_VOLTAGE_MODE 1 +#define DC__UNIPHY_VOLTAGE_MODE__1 1 +#define DC__BLON_TYPE dedicated +#define DC__BLON_TYPE__DEDICATED 1 +#define DC__UNIPHY_STAGGER_CH_PRESENT 1 +#define DC__UNIPHY_STAGGER_CH_PRESENT__1 1 +#define DC__XDMA_PRESENT 0 +#define DC__XDMA_PRESENT__0 1 +#define XDMA__PRESENT 0 +#define XDMA__PRESENT__0 1 +#define DC__DP_MEM_PG 0 +#define DC__DP_MEM_PG__0 1 +#define DP__MEM_PG 0 +#define DP__MEM_PG__0 1 +#define DC__AFMT_MEM_PG 0 +#define DC__AFMT_MEM_PG__0 1 +#define AFMT__MEM_PG 0 +#define AFMT__MEM_PG__0 1 +#define DC__HDMI_MEM_PG 0 +#define DC__HDMI_MEM_PG__0 1 +#define HDMI__MEM_PG 0 +#define HDMI__MEM_PG__0 1 +#define DC__I2C_MEM_PG 0 +#define DC__I2C_MEM_PG__0 1 +#define I2C__MEM_PG 0 +#define I2C__MEM_PG__0 1 +#define DC__DSCL_MEM_PG 0 +#define DC__DSCL_MEM_PG__0 1 +#define DSCL__MEM_PG 0 +#define DSCL__MEM_PG__0 1 +#define DC__CM_MEM_PG 0 +#define DC__CM_MEM_PG__0 1 +#define CM__MEM_PG 0 +#define CM__MEM_PG__0 1 +#define DC__OBUF_MEM_PG 0 +#define DC__OBUF_MEM_PG__0 1 +#define OBUF__MEM_PG 0 +#define OBUF__MEM_PG__0 1 +#define DC__WBIF_MEM_PG 1 +#define DC__WBIF_MEM_PG__1 1 +#define WBIF__MEM_PG 1 +#define WBIF__MEM_PG__1 1 +#define DC__VGA_MEM_PG 0 +#define DC__VGA_MEM_PG__0 1 +#define VGA__MEM_PG 0 +#define VGA__MEM_PG__0 1 +#define DC__FMT_MEM_PG 0 +#define DC__FMT_MEM_PG__0 1 +#define FMT__MEM_PG 0 +#define FMT__MEM_PG__0 1 +#define DC__ODM_MEM_PG 0 +#define DC__ODM_MEM_PG__0 1 +#define ODM__MEM_PG 0 +#define ODM__MEM_PG__0 1 +#define DC__DSI_MEM_PG 0 +#define DC__DSI_MEM_PG__0 1 +#define DSI__MEM_PG 0 +#define DSI__MEM_PG__0 1 +#define DC__AZ_MEM_PG 1 +#define DC__AZ_MEM_PG__1 1 +#define AZ__MEM_PG 1 +#define AZ__MEM_PG__1 1 +#define DC__WBSCL_MEM1P1024X64QS_MEM_PG 1 +#define DC__WBSCL_MEM1P1024X64QS_MEM_PG__1 1 +#define WBSCL_MEM1P1024X64QS__MEM_PG 1 +#define WBSCL_MEM1P1024X64QS__MEM_PG__1 1 +#define DC__WBSCL_MEM1P528X64QS_MEM_PG 1 +#define DC__WBSCL_MEM1P528X64QS_MEM_PG__1 1 +#define WBSCL_MEM1P528X64QS__MEM_PG 1 +#define WBSCL_MEM1P528X64QS__MEM_PG__1 1 +#define DC__DMCU_MEM1P1024X32BQS_MEM_PG 1 +#define DC__DMCU_MEM1P1024X32BQS_MEM_PG__1 1 +#define DMCU_MEM1P1024X32BQS__MEM_PG 1 +#define DMCU_MEM1P1024X32BQS__MEM_PG__1 1 +#define DC__HUBBUB_SDP_TAG_INT_MEM_PG 0 +#define DC__HUBBUB_SDP_TAG_INT_MEM_PG__0 1 +#define HUBBUB_SDP_TAG_INT__MEM_PG 0 +#define HUBBUB_SDP_TAG_INT__MEM_PG__0 1 +#define DC__HUBBUB_SDP_TAG_EXT_MEM_PG 0 +#define DC__HUBBUB_SDP_TAG_EXT_MEM_PG__0 1 +#define HUBBUB_SDP_TAG_EXT__MEM_PG 0 +#define HUBBUB_SDP_TAG_EXT__MEM_PG__0 1 +#define DC__HUBBUB_RET_ZERO_MEM_PG 0 +#define DC__HUBBUB_RET_ZERO_MEM_PG__0 1 +#define HUBBUB_RET_ZERO__MEM_PG 0 +#define HUBBUB_RET_ZERO__MEM_PG__0 1 +#define DC__HUBBUB_RET_ROB_MEM_PG 0 +#define DC__HUBBUB_RET_ROB_MEM_PG__0 1 +#define HUBBUB_RET_ROB__MEM_PG 0 +#define HUBBUB_RET_ROB__MEM_PG__0 1 +#define DC__HUBPRET_CUR_ROB_MEM_PG 0 +#define DC__HUBPRET_CUR_ROB_MEM_PG__0 1 +#define HUBPRET_CUR_ROB__MEM_PG 0 +#define HUBPRET_CUR_ROB__MEM_PG__0 1 +#define DC__HUBPRET_CUR_CDC_MEM_PG 0 +#define DC__HUBPRET_CUR_CDC_MEM_PG__0 1 +#define HUBPRET_CUR_CDC__MEM_PG 0 +#define HUBPRET_CUR_CDC__MEM_PG__0 1 +#define DC__HUBPREQ_MPTE_MEM_PG 0 +#define DC__HUBPREQ_MPTE_MEM_PG__0 1 +#define HUBPREQ_MPTE__MEM_PG 0 +#define HUBPREQ_MPTE__MEM_PG__0 1 +#define DC__HUBPREQ_META_MEM_PG 0 +#define DC__HUBPREQ_META_MEM_PG__0 1 +#define HUBPREQ_META__MEM_PG 0 +#define HUBPREQ_META__MEM_PG__0 1 +#define DC__HUBPREQ_DPTE_MEM_PG 0 +#define DC__HUBPREQ_DPTE_MEM_PG__0 1 +#define HUBPREQ_DPTE__MEM_PG 0 +#define HUBPREQ_DPTE__MEM_PG__0 1 +#define DC__HUBPRET_DET_MEM_PG 0 +#define DC__HUBPRET_DET_MEM_PG__0 1 +#define HUBPRET_DET__MEM_PG 0 +#define HUBPRET_DET__MEM_PG__0 1 +#define DC__HUBPRET_PIX_CDC_MEM_PG 0 +#define DC__HUBPRET_PIX_CDC_MEM_PG__0 1 +#define HUBPRET_PIX_CDC__MEM_PG 0 +#define HUBPRET_PIX_CDC__MEM_PG__0 1 +#define DC__TOP_BLKS__DCCG 1 +#define DC__TOP_BLKS__DCHUBBUB 1 +#define DC__TOP_BLKS__DCHUBP 1 +#define DC__TOP_BLKS__HDA 1 +#define DC__TOP_BLKS__DIO 1 +#define DC__TOP_BLKS__DCIO 1 +#define DC__TOP_BLKS__DMU 1 +#define DC__TOP_BLKS__DPP 1 +#define DC__TOP_BLKS__MPC 1 +#define DC__TOP_BLKS__OPP 1 +#define DC__TOP_BLKS__OPTC 1 +#define DC__TOP_BLKS__MMHUBBUB 1 +#define DC__TOP_BLKS__WB 1 +#define DC__TOP_BLKS__MAX 13 +#define DC__TOP_BLKS__MAX__13 1 +#define DC__DCHUBP_DPP_SF_PIXEL_CREDITS 9 +#define DC__DCHUBP_DPP_SF_PIXEL_CREDITS__9 1 +#define DC__DPP_MPC_SF_PIXEL_CREDITS 9 +#define DC__DPP_MPC_SF_PIXEL_CREDITS__9 1 +#define DC__MPC_OPP_SF_PIXEL_CREDITS 8 +#define DC__MPC_OPP_SF_PIXEL_CREDITS__8 1 +#define DC__OPP_OPTC_SF_PIXEL_CREDITS 8 +#define DC__OPP_OPTC_SF_PIXEL_CREDITS__8 1 +#define DC__SFR_SFT_ROUND_TRIP_DELAY 5 +#define DC__SFR_SFT_ROUND_TRIP_DELAY__5 1 +#define DC__REPEATER_PROJECT_MAX 8 +#define DC__REPEATER_PROJECT_MAX__8 1 +#define DC__SURFACE_422_CAPABLE 0 +#define DC__SURFACE_422_CAPABLE__0 1 +#endif diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_mode_enums.h b/drivers/gpu/drm/amd/display/dc/dml/display_mode_enums.h new file mode 100644 index 000000000000..b1ad3553f900 --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/dml/display_mode_enums.h @@ -0,0 +1,111 @@ +/* + * Copyright 2017 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 __DISPLAY_MODE_ENUMS_H__ +#define __DISPLAY_MODE_ENUMS_H__ + +enum output_encoder_class { + dm_dp = 0, dm_hdmi = 1, dm_wb = 2, dm_edp +}; +enum output_format_class { + dm_444 = 0, dm_420 = 1, dm_n422, dm_s422 +}; +enum source_format_class { + dm_444_16 = 0, + dm_444_32 = 1, + dm_444_64 = 2, + dm_420_8 = 3, + dm_420_10 = 4, + dm_422_8 = 5, + dm_422_10 = 6, + dm_444_8 = 7, + dm_mono_8, + dm_mono_16 +}; +enum output_bpc_class { + dm_out_6 = 0, dm_out_8 = 1, dm_out_10 = 2, dm_out_12 = 3, dm_out_16 = 4 +}; +enum scan_direction_class { + dm_horz = 0, dm_vert = 1 +}; +enum dm_swizzle_mode { + dm_sw_linear = 0, + dm_sw_256b_s = 1, + dm_sw_256b_d = 2, + dm_sw_SPARE_0 = 3, + dm_sw_SPARE_1 = 4, + dm_sw_4kb_s = 5, + dm_sw_4kb_d = 6, + dm_sw_SPARE_2 = 7, + dm_sw_SPARE_3 = 8, + dm_sw_64kb_s = 9, + dm_sw_64kb_d = 10, + dm_sw_SPARE_4 = 11, + dm_sw_SPARE_5 = 12, + dm_sw_var_s = 13, + dm_sw_var_d = 14, + dm_sw_SPARE_6 = 15, + dm_sw_SPARE_7 = 16, + dm_sw_64kb_s_t = 17, + dm_sw_64kb_d_t = 18, + dm_sw_SPARE_10 = 19, + dm_sw_SPARE_11 = 20, + dm_sw_4kb_s_x = 21, + dm_sw_4kb_d_x = 22, + dm_sw_SPARE_12 = 23, + dm_sw_SPARE_13 = 24, + dm_sw_64kb_s_x = 25, + dm_sw_64kb_d_x = 26, + dm_sw_SPARE_14 = 27, + dm_sw_SPARE_15 = 28, + dm_sw_var_s_x = 29, + dm_sw_var_d_x = 30, + dm_sw_64kb_r_x, + dm_sw_gfx7_2d_thin_lvp, + dm_sw_gfx7_2d_thin_gl +}; +enum lb_depth { + dm_lb_10 = 0, dm_lb_8 = 1, dm_lb_6 = 2, dm_lb_12 = 3, dm_lb_16 +}; +enum voltage_state { + dm_vmin = 0, dm_vmid = 1, dm_vnom = 2, dm_vmax = 3 +}; +enum source_macro_tile_size { + dm_4k_tile = 0, dm_64k_tile = 1, dm_256k_tile = 2 +}; +enum cursor_bpp { + dm_cur_2bit = 0, dm_cur_32bit = 1, dm_cur_64bit = 2 +}; +enum clock_change_support { + dm_dram_clock_change_uninitialized = 0, + dm_dram_clock_change_vactive, + dm_dram_clock_change_vblank, + dm_dram_clock_change_unsupported +}; + +enum output_standard { + dm_std_uninitialized = 0, dm_std_cvtr2, dm_std_cvt +}; + +#endif diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.c b/drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.c new file mode 100644 index 000000000000..4c31fa54af39 --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.c @@ -0,0 +1,138 @@ +/* + * Copyright 2017 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 "display_mode_lib.h" +#include "dc_features.h" + +static void set_soc_bounding_box(struct _vcs_dpi_soc_bounding_box_st *soc, enum dml_project project) +{ + if (project == DML_PROJECT_RAVEN1) { + soc->sr_exit_time_us = 9.0; + soc->sr_enter_plus_exit_time_us = 11.0; + soc->urgent_latency_us = 4.0; + soc->writeback_latency_us = 12.0; + soc->ideal_dram_bw_after_urgent_percent = 80.0; + soc->max_request_size_bytes = 256; + + soc->vmin.dcfclk_mhz = 300.0; + soc->vmin.dispclk_mhz = 608.0; + soc->vmin.dppclk_mhz = 435.0; + soc->vmin.dram_bw_per_chan_gbps = 12.8; + soc->vmin.phyclk_mhz = 540.0; + soc->vmin.socclk_mhz = 208.0; + + soc->vmid.dcfclk_mhz = 600.0; + soc->vmid.dispclk_mhz = 661.0; + soc->vmid.dppclk_mhz = 661.0; + soc->vmid.dram_bw_per_chan_gbps = 12.8; + soc->vmid.phyclk_mhz = 540.0; + soc->vmid.socclk_mhz = 208.0; + + soc->vnom.dcfclk_mhz = 600.0; + soc->vnom.dispclk_mhz = 661.0; + soc->vnom.dppclk_mhz = 661.0; + soc->vnom.dram_bw_per_chan_gbps = 38.4; + soc->vnom.phyclk_mhz = 810; + soc->vnom.socclk_mhz = 208.0; + + soc->vmax.dcfclk_mhz = 600.0; + soc->vmax.dispclk_mhz = 1086.0; + soc->vmax.dppclk_mhz = 661.0; + soc->vmax.dram_bw_per_chan_gbps = 38.4; + soc->vmax.phyclk_mhz = 810.0; + soc->vmax.socclk_mhz = 208.0; + + soc->downspread_percent = 0.5; + soc->dram_page_open_time_ns = 50.0; + soc->dram_rw_turnaround_time_ns = 17.5; + soc->dram_return_buffer_per_channel_bytes = 8192; + soc->round_trip_ping_latency_dcfclk_cycles = 128; + soc->urgent_out_of_order_return_per_channel_bytes = 256; + soc->channel_interleave_bytes = 256; + soc->num_banks = 8; + soc->num_chans = 2; + soc->vmm_page_size_bytes = 4096; + soc->dram_clock_change_latency_us = 17.0; + soc->writeback_dram_clock_change_latency_us = 23.0; + soc->return_bus_width_bytes = 64; + } else { + BREAK_TO_DEBUGGER(); /* Invalid Project Specified */ + } +} + +static void set_ip_params(struct _vcs_dpi_ip_params_st *ip, enum dml_project project) +{ + if (project == DML_PROJECT_RAVEN1) { + ip->rob_buffer_size_kbytes = 64; + ip->det_buffer_size_kbytes = 164; + ip->dpte_buffer_size_in_pte_reqs = 42; + ip->dpp_output_buffer_pixels = 2560; + ip->opp_output_buffer_lines = 1; + ip->pixel_chunk_size_kbytes = 8; + ip->pte_enable = 1; + ip->pte_chunk_size_kbytes = 2; + ip->meta_chunk_size_kbytes = 2; + ip->writeback_chunk_size_kbytes = 2; + ip->line_buffer_size_bits = 589824; + ip->max_line_buffer_lines = 12; + ip->IsLineBufferBppFixed = 0; + ip->LineBufferFixedBpp = -1; + ip->writeback_luma_buffer_size_kbytes = 12; + ip->writeback_chroma_buffer_size_kbytes = 8; + ip->max_num_dpp = 4; + ip->max_num_wb = 2; + ip->max_dchub_pscl_bw_pix_per_clk = 4; + ip->max_pscl_lb_bw_pix_per_clk = 2; + ip->max_lb_vscl_bw_pix_per_clk = 4; + ip->max_vscl_hscl_bw_pix_per_clk = 4; + ip->max_hscl_ratio = 4; + ip->max_vscl_ratio = 4; + ip->hscl_mults = 4; + ip->vscl_mults = 4; + ip->max_hscl_taps = 8; + ip->max_vscl_taps = 8; + ip->dispclk_ramp_margin_percent = 1; + ip->underscan_factor = 1.10; + ip->min_vblank_lines = 14; + ip->dppclk_delay_subtotal = 90; + ip->dispclk_delay_subtotal = 42; + ip->dcfclk_cstate_latency = 10; + ip->max_inter_dcn_tile_repeaters = 8; + ip->can_vstartup_lines_exceed_vsync_plus_back_porch_lines_minus_one = 0; + ip->bug_forcing_LC_req_same_size_fixed = 0; + } else { + BREAK_TO_DEBUGGER(); /* Invalid Project Specified */ + } +} + +void dml_init_instance(struct display_mode_lib *lib, enum dml_project project) +{ + if (lib->project != project) { + set_soc_bounding_box(&lib->soc, project); + set_ip_params(&lib->ip, project); + lib->project = project; + } +} + diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.h b/drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.h new file mode 100644 index 000000000000..26f4f2a3d90d --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.h @@ -0,0 +1,50 @@ +/* + * Copyright 2017 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 __DISPLAY_MODE_LIB_H__ +#define __DISPLAY_MODE_LIB_H__ + + +#include "dml_common_defs.h" +#include "soc_bounding_box.h" +#include "display_mode_vba.h" +#include "display_rq_dlg_calc.h" +#include "dml1_display_rq_dlg_calc.h" + +enum dml_project { + DML_PROJECT_UNDEFINED, + DML_PROJECT_RAVEN1 +}; + +struct display_mode_lib { + struct _vcs_dpi_ip_params_st ip; + struct _vcs_dpi_soc_bounding_box_st soc; + enum dml_project project; + struct vba_vars_st vba; + struct dal_logger *logger; +}; + +void dml_init_instance(struct display_mode_lib *lib, enum dml_project project); + +#endif diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_mode_structs.h b/drivers/gpu/drm/amd/display/dc/dml/display_mode_structs.h new file mode 100644 index 000000000000..baf182177736 --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/dml/display_mode_structs.h @@ -0,0 +1,557 @@ +/* + * Copyright 2017 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 __DISPLAY_MODE_STRUCTS_H__ +#define __DISPLAY_MODE_STRUCTS_H__ + +typedef struct _vcs_dpi_voltage_scaling_st voltage_scaling_st; +typedef struct _vcs_dpi_soc_bounding_box_st soc_bounding_box_st; +typedef struct _vcs_dpi_ip_params_st ip_params_st; +typedef struct _vcs_dpi_display_pipe_source_params_st display_pipe_source_params_st; +typedef struct _vcs_dpi_display_output_params_st display_output_params_st; +typedef struct _vcs_dpi_display_bandwidth_st display_bandwidth_st; +typedef struct _vcs_dpi_scaler_ratio_depth_st scaler_ratio_depth_st; +typedef struct _vcs_dpi_scaler_taps_st scaler_taps_st; +typedef struct _vcs_dpi_display_pipe_dest_params_st display_pipe_dest_params_st; +typedef struct _vcs_dpi_display_pipe_params_st display_pipe_params_st; +typedef struct _vcs_dpi_display_clocks_and_cfg_st display_clocks_and_cfg_st; +typedef struct _vcs_dpi_display_e2e_pipe_params_st display_e2e_pipe_params_st; +typedef struct _vcs_dpi_dchub_buffer_sizing_st dchub_buffer_sizing_st; +typedef struct _vcs_dpi_watermarks_perf_st watermarks_perf_st; +typedef struct _vcs_dpi_cstate_pstate_watermarks_st cstate_pstate_watermarks_st; +typedef struct _vcs_dpi_wm_calc_pipe_params_st wm_calc_pipe_params_st; +typedef struct _vcs_dpi_vratio_pre_st vratio_pre_st; +typedef struct _vcs_dpi_display_data_rq_misc_params_st display_data_rq_misc_params_st; +typedef struct _vcs_dpi_display_data_rq_sizing_params_st display_data_rq_sizing_params_st; +typedef struct _vcs_dpi_display_data_rq_dlg_params_st display_data_rq_dlg_params_st; +typedef struct _vcs_dpi_display_cur_rq_dlg_params_st display_cur_rq_dlg_params_st; +typedef struct _vcs_dpi_display_rq_dlg_params_st display_rq_dlg_params_st; +typedef struct _vcs_dpi_display_rq_sizing_params_st display_rq_sizing_params_st; +typedef struct _vcs_dpi_display_rq_misc_params_st display_rq_misc_params_st; +typedef struct _vcs_dpi_display_rq_params_st display_rq_params_st; +typedef struct _vcs_dpi_display_dlg_regs_st display_dlg_regs_st; +typedef struct _vcs_dpi_display_ttu_regs_st display_ttu_regs_st; +typedef struct _vcs_dpi_display_data_rq_regs_st display_data_rq_regs_st; +typedef struct _vcs_dpi_display_rq_regs_st display_rq_regs_st; +typedef struct _vcs_dpi_display_dlg_sys_params_st display_dlg_sys_params_st; +typedef struct _vcs_dpi_display_dlg_prefetch_param_st display_dlg_prefetch_param_st; +typedef struct _vcs_dpi_display_pipe_clock_st display_pipe_clock_st; +typedef struct _vcs_dpi_display_arb_params_st display_arb_params_st; + +struct _vcs_dpi_voltage_scaling_st { + int state; + double dscclk_mhz; + double dcfclk_mhz; + double socclk_mhz; + double dram_speed_mhz; + double fabricclk_mhz; + double dispclk_mhz; + double dram_bw_per_chan_gbps; + double phyclk_mhz; + double dppclk_mhz; +}; + +struct _vcs_dpi_soc_bounding_box_st { + double sr_exit_time_us; + double sr_enter_plus_exit_time_us; + double urgent_latency_us; + double writeback_latency_us; + double ideal_dram_bw_after_urgent_percent; + unsigned int max_request_size_bytes; + struct _vcs_dpi_voltage_scaling_st vmin; + struct _vcs_dpi_voltage_scaling_st vmid; + struct _vcs_dpi_voltage_scaling_st vnom; + struct _vcs_dpi_voltage_scaling_st vmax; + double downspread_percent; + double dram_page_open_time_ns; + double dram_rw_turnaround_time_ns; + double dram_return_buffer_per_channel_bytes; + double dram_channel_width_bytes; + double fabric_datapath_to_dcn_data_return_bytes; + double dcn_downspread_percent; + double dispclk_dppclk_vco_speed_mhz; + double dfs_vco_period_ps; + unsigned int round_trip_ping_latency_dcfclk_cycles; + unsigned int urgent_out_of_order_return_per_channel_bytes; + unsigned int channel_interleave_bytes; + unsigned int num_banks; + unsigned int num_chans; + unsigned int vmm_page_size_bytes; + double dram_clock_change_latency_us; + double writeback_dram_clock_change_latency_us; + unsigned int return_bus_width_bytes; + unsigned int voltage_override; + double xfc_bus_transport_time_us; + double xfc_xbuf_latency_tolerance_us; + struct _vcs_dpi_voltage_scaling_st clock_limits[7]; +}; + +struct _vcs_dpi_ip_params_st { + unsigned int max_inter_dcn_tile_repeaters; + unsigned int num_dsc; + unsigned int odm_capable; + unsigned int rob_buffer_size_kbytes; + unsigned int det_buffer_size_kbytes; + unsigned int dpte_buffer_size_in_pte_reqs; + unsigned int pde_proc_buffer_size_64k_reqs; + unsigned int dpp_output_buffer_pixels; + unsigned int opp_output_buffer_lines; + unsigned int pixel_chunk_size_kbytes; + unsigned char pte_enable; + unsigned int pte_chunk_size_kbytes; + unsigned int meta_chunk_size_kbytes; + unsigned int writeback_chunk_size_kbytes; + unsigned int line_buffer_size_bits; + unsigned int max_line_buffer_lines; + unsigned int writeback_luma_buffer_size_kbytes; + unsigned int writeback_chroma_buffer_size_kbytes; + unsigned int writeback_chroma_line_buffer_width_pixels; + unsigned int max_page_table_levels; + unsigned int max_num_dpp; + unsigned int max_num_otg; + unsigned int cursor_chunk_size; + unsigned int cursor_buffer_size; + unsigned int max_num_wb; + unsigned int max_dchub_pscl_bw_pix_per_clk; + unsigned int max_pscl_lb_bw_pix_per_clk; + unsigned int max_lb_vscl_bw_pix_per_clk; + unsigned int max_vscl_hscl_bw_pix_per_clk; + double max_hscl_ratio; + double max_vscl_ratio; + unsigned int hscl_mults; + unsigned int vscl_mults; + unsigned int max_hscl_taps; + unsigned int max_vscl_taps; + unsigned int xfc_supported; + unsigned int ptoi_supported; + unsigned int xfc_fill_constant_bytes; + double dispclk_ramp_margin_percent; + double xfc_fill_bw_overhead_percent; + double underscan_factor; + unsigned int min_vblank_lines; + unsigned int dppclk_delay_subtotal; + unsigned int dispclk_delay_subtotal; + unsigned int dcfclk_cstate_latency; + unsigned int dppclk_delay_scl; + unsigned int dppclk_delay_scl_lb_only; + unsigned int dppclk_delay_cnvc_formatter; + unsigned int dppclk_delay_cnvc_cursor; + unsigned int is_line_buffer_bpp_fixed; + unsigned int line_buffer_fixed_bpp; + unsigned int dcc_supported; + + unsigned int IsLineBufferBppFixed; + unsigned int LineBufferFixedBpp; + unsigned int can_vstartup_lines_exceed_vsync_plus_back_porch_lines_minus_one; + unsigned int bug_forcing_LC_req_same_size_fixed; +}; + +struct _vcs_dpi_display_xfc_params_st { + double xfc_tslv_vready_offset_us; + double xfc_tslv_vupdate_width_us; + double xfc_tslv_vupdate_offset_us; + int xfc_slv_chunk_size_bytes; +}; + +struct _vcs_dpi_display_pipe_source_params_st { + int source_format; + unsigned char dcc; + unsigned int dcc_override; + unsigned int dcc_rate; + unsigned char dcc_use_global; + unsigned char vm; + unsigned char vm_levels_force_en; + unsigned int vm_levels_force; + int source_scan; + int sw_mode; + int macro_tile_size; + unsigned char is_display_sw; + unsigned int viewport_width; + unsigned int viewport_height; + unsigned int viewport_y_y; + unsigned int viewport_y_c; + unsigned int viewport_width_c; + unsigned int viewport_height_c; + unsigned int data_pitch; + unsigned int data_pitch_c; + unsigned int meta_pitch; + unsigned int meta_pitch_c; + unsigned int cur0_src_width; + int cur0_bpp; + unsigned int cur1_src_width; + int cur1_bpp; + int num_cursors; + unsigned char is_hsplit; + unsigned char dynamic_metadata_enable; + unsigned int dynamic_metadata_lines_before_active; + unsigned int dynamic_metadata_xmit_bytes; + unsigned int hsplit_grp; + unsigned char xfc_enable; + unsigned char xfc_slave; + struct _vcs_dpi_display_xfc_params_st xfc_params; +}; +struct writeback_st { + int wb_src_height; + int wb_dst_width; + int wb_dst_height; + int wb_pixel_format; + int wb_htaps_luma; + int wb_vtaps_luma; + int wb_htaps_chroma; + int wb_vtaps_chroma; + int wb_hratio; + int wb_vratio; +}; + +struct _vcs_dpi_display_output_params_st { + int dp_lanes; + int output_bpp; + int dsc_enable; + int wb_enable; + int output_bpc; + int output_type; + int output_format; + int output_standard; + int dsc_slices; + struct writeback_st wb; +}; + +struct _vcs_dpi_display_bandwidth_st { + double total_bw_consumed_gbps; + double guaranteed_urgent_return_bw_gbps; +}; + +struct _vcs_dpi_scaler_ratio_depth_st { + double hscl_ratio; + double vscl_ratio; + double hscl_ratio_c; + double vscl_ratio_c; + double vinit; + double vinit_c; + double vinit_bot; + double vinit_bot_c; + int lb_depth; + int scl_enable; +}; + +struct _vcs_dpi_scaler_taps_st { + unsigned int htaps; + unsigned int vtaps; + unsigned int htaps_c; + unsigned int vtaps_c; +}; + +struct _vcs_dpi_display_pipe_dest_params_st { + unsigned int recout_width; + unsigned int recout_height; + unsigned int full_recout_width; + unsigned int full_recout_height; + unsigned int hblank_start; + unsigned int hblank_end; + unsigned int vblank_start; + unsigned int vblank_end; + unsigned int htotal; + unsigned int vtotal; + unsigned int vactive; + unsigned int hactive; + unsigned int vstartup_start; + unsigned int vupdate_offset; + unsigned int vupdate_width; + unsigned int vready_offset; + unsigned char interlaced; + unsigned char underscan; + double pixel_rate_mhz; + unsigned char synchronized_vblank_all_planes; + unsigned char otg_inst; + unsigned char odm_split_cnt; + unsigned char odm_combine; +}; + +struct _vcs_dpi_display_pipe_params_st { + display_pipe_source_params_st src; + display_pipe_dest_params_st dest; + scaler_ratio_depth_st scale_ratio_depth; + scaler_taps_st scale_taps; +}; + +struct _vcs_dpi_display_clocks_and_cfg_st { + int voltage; + double dppclk_mhz; + double refclk_mhz; + double dispclk_mhz; + double dcfclk_mhz; + double socclk_mhz; +}; + +struct _vcs_dpi_display_e2e_pipe_params_st { + display_pipe_params_st pipe; + display_output_params_st dout; + display_clocks_and_cfg_st clks_cfg; +}; + +struct _vcs_dpi_dchub_buffer_sizing_st { + unsigned int swath_width_y; + unsigned int swath_height_y; + unsigned int swath_height_c; + unsigned int detail_buffer_size_y; +}; + +struct _vcs_dpi_watermarks_perf_st { + double stutter_eff_in_active_region_percent; + double urgent_latency_supported_us; + double non_urgent_latency_supported_us; + double dram_clock_change_margin_us; + double dram_access_eff_percent; +}; + +struct _vcs_dpi_cstate_pstate_watermarks_st { + double cstate_exit_us; + double cstate_enter_plus_exit_us; + double pstate_change_us; +}; + +struct _vcs_dpi_wm_calc_pipe_params_st { + unsigned int num_dpp; + int voltage; + int output_type; + double dcfclk_mhz; + double socclk_mhz; + double dppclk_mhz; + double pixclk_mhz; + unsigned char interlace_en; + unsigned char pte_enable; + unsigned char dcc_enable; + double dcc_rate; + double bytes_per_pixel_c; + double bytes_per_pixel_y; + unsigned int swath_width_y; + unsigned int swath_height_y; + unsigned int swath_height_c; + unsigned int det_buffer_size_y; + double h_ratio; + double v_ratio; + unsigned int h_taps; + unsigned int h_total; + unsigned int v_total; + unsigned int v_active; + unsigned int e2e_index; + double display_pipe_line_delivery_time; + double read_bw; + unsigned int lines_in_det_y; + unsigned int lines_in_det_y_rounded_down_to_swath; + double full_det_buffering_time; + double dcfclk_deepsleep_mhz_per_plane; +}; + +struct _vcs_dpi_vratio_pre_st { + double vratio_pre_l; + double vratio_pre_c; +}; + +struct _vcs_dpi_display_data_rq_misc_params_st { + unsigned int full_swath_bytes; + unsigned int stored_swath_bytes; + unsigned int blk256_height; + unsigned int blk256_width; + unsigned int req_height; + unsigned int req_width; +}; + +struct _vcs_dpi_display_data_rq_sizing_params_st { + unsigned int chunk_bytes; + unsigned int min_chunk_bytes; + unsigned int meta_chunk_bytes; + unsigned int min_meta_chunk_bytes; + unsigned int mpte_group_bytes; + unsigned int dpte_group_bytes; +}; + +struct _vcs_dpi_display_data_rq_dlg_params_st { + unsigned int swath_width_ub; + unsigned int swath_height; + unsigned int req_per_swath_ub; + unsigned int meta_pte_bytes_per_frame_ub; + unsigned int dpte_req_per_row_ub; + unsigned int dpte_groups_per_row_ub; + unsigned int dpte_row_height; + unsigned int dpte_bytes_per_row_ub; + unsigned int meta_chunks_per_row_ub; + unsigned int meta_req_per_row_ub; + unsigned int meta_row_height; + unsigned int meta_bytes_per_row_ub; +}; + +struct _vcs_dpi_display_cur_rq_dlg_params_st { + unsigned char enable; + unsigned int swath_height; + unsigned int req_per_line; +}; + +struct _vcs_dpi_display_rq_dlg_params_st { + display_data_rq_dlg_params_st rq_l; + display_data_rq_dlg_params_st rq_c; + display_cur_rq_dlg_params_st rq_cur0; +}; + +struct _vcs_dpi_display_rq_sizing_params_st { + display_data_rq_sizing_params_st rq_l; + display_data_rq_sizing_params_st rq_c; +}; + +struct _vcs_dpi_display_rq_misc_params_st { + display_data_rq_misc_params_st rq_l; + display_data_rq_misc_params_st rq_c; +}; + +struct _vcs_dpi_display_rq_params_st { + unsigned char yuv420; + unsigned char yuv420_10bpc; + display_rq_misc_params_st misc; + display_rq_sizing_params_st sizing; + display_rq_dlg_params_st dlg; +}; + +struct _vcs_dpi_display_dlg_regs_st { + unsigned int refcyc_h_blank_end; + unsigned int dlg_vblank_end; + unsigned int min_dst_y_next_start; + unsigned int refcyc_per_htotal; + unsigned int refcyc_x_after_scaler; + unsigned int dst_y_after_scaler; + unsigned int dst_y_prefetch; + unsigned int dst_y_per_vm_vblank; + unsigned int dst_y_per_row_vblank; + unsigned int dst_y_per_vm_flip; + unsigned int dst_y_per_row_flip; + unsigned int ref_freq_to_pix_freq; + unsigned int vratio_prefetch; + unsigned int vratio_prefetch_c; + unsigned int refcyc_per_pte_group_vblank_l; + unsigned int refcyc_per_pte_group_vblank_c; + unsigned int refcyc_per_meta_chunk_vblank_l; + unsigned int refcyc_per_meta_chunk_vblank_c; + unsigned int refcyc_per_pte_group_flip_l; + unsigned int refcyc_per_pte_group_flip_c; + unsigned int refcyc_per_meta_chunk_flip_l; + unsigned int refcyc_per_meta_chunk_flip_c; + unsigned int dst_y_per_pte_row_nom_l; + unsigned int dst_y_per_pte_row_nom_c; + unsigned int refcyc_per_pte_group_nom_l; + unsigned int refcyc_per_pte_group_nom_c; + unsigned int dst_y_per_meta_row_nom_l; + unsigned int dst_y_per_meta_row_nom_c; + unsigned int refcyc_per_meta_chunk_nom_l; + unsigned int refcyc_per_meta_chunk_nom_c; + unsigned int refcyc_per_line_delivery_pre_l; + unsigned int refcyc_per_line_delivery_pre_c; + unsigned int refcyc_per_line_delivery_l; + unsigned int refcyc_per_line_delivery_c; + unsigned int chunk_hdl_adjust_cur0; + unsigned int chunk_hdl_adjust_cur1; + unsigned int vready_after_vcount0; + unsigned int dst_y_offset_cur0; + unsigned int dst_y_offset_cur1; + unsigned int xfc_reg_transfer_delay; + unsigned int xfc_reg_precharge_delay; + unsigned int xfc_reg_remote_surface_flip_latency; + unsigned int xfc_reg_prefetch_margin; + unsigned int dst_y_delta_drq_limit; +}; + +struct _vcs_dpi_display_ttu_regs_st { + unsigned int qos_level_low_wm; + unsigned int qos_level_high_wm; + unsigned int min_ttu_vblank; + unsigned int qos_level_flip; + unsigned int refcyc_per_req_delivery_l; + unsigned int refcyc_per_req_delivery_c; + unsigned int refcyc_per_req_delivery_cur0; + unsigned int refcyc_per_req_delivery_cur1; + unsigned int refcyc_per_req_delivery_pre_l; + unsigned int refcyc_per_req_delivery_pre_c; + unsigned int refcyc_per_req_delivery_pre_cur0; + unsigned int refcyc_per_req_delivery_pre_cur1; + unsigned int qos_level_fixed_l; + unsigned int qos_level_fixed_c; + unsigned int qos_level_fixed_cur0; + unsigned int qos_level_fixed_cur1; + unsigned int qos_ramp_disable_l; + unsigned int qos_ramp_disable_c; + unsigned int qos_ramp_disable_cur0; + unsigned int qos_ramp_disable_cur1; +}; + +struct _vcs_dpi_display_data_rq_regs_st { + unsigned int chunk_size; + unsigned int min_chunk_size; + unsigned int meta_chunk_size; + unsigned int min_meta_chunk_size; + unsigned int dpte_group_size; + unsigned int mpte_group_size; + unsigned int swath_height; + unsigned int pte_row_height_linear; +}; + +struct _vcs_dpi_display_rq_regs_st { + display_data_rq_regs_st rq_regs_l; + display_data_rq_regs_st rq_regs_c; + unsigned int drq_expansion_mode; + unsigned int prq_expansion_mode; + unsigned int mrq_expansion_mode; + unsigned int crq_expansion_mode; + unsigned int plane1_base_address; +}; + +struct _vcs_dpi_display_dlg_sys_params_st { + double t_mclk_wm_us; + double t_urg_wm_us; + double t_sr_wm_us; + double t_extra_us; + double mem_trip_us; + double t_srx_delay_us; + double deepsleep_dcfclk_mhz; + double total_flip_bw; + unsigned int total_flip_bytes; +}; + +struct _vcs_dpi_display_dlg_prefetch_param_st { + double prefetch_bw; + unsigned int flip_bytes; +}; + +struct _vcs_dpi_display_pipe_clock_st { + double dcfclk_mhz; + double dispclk_mhz; + double socclk_mhz; + double dscclk_mhz[6]; + double dppclk_mhz[6]; +}; + +struct _vcs_dpi_display_arb_params_st { + int max_req_outstanding; + int min_req_outstanding; + int sat_level_us; +}; + +#endif /*__DISPLAY_MODE_STRUCTS_H__*/ diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.c b/drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.c new file mode 100644 index 000000000000..ea661ee44674 --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.c @@ -0,0 +1,6124 @@ +/* + * Copyright 2017 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 "display_mode_lib.h" +#include "display_mode_vba.h" + +#include "dml_inline_defs.h" + +static const unsigned int NumberOfStates = DC__VOLTAGE_STATES; + +static void fetch_socbb_params(struct display_mode_lib *mode_lib); +static void fetch_ip_params(struct display_mode_lib *mode_lib); +static void fetch_pipe_params(struct display_mode_lib *mode_lib); +static void recalculate_params( + struct display_mode_lib *mode_lib, + const display_e2e_pipe_params_st *pipes, + unsigned int num_pipes); +static void recalculate(struct display_mode_lib *mode_lib); +static double adjust_ReturnBW( + struct display_mode_lib *mode_lib, + double ReturnBW, + bool DCCEnabledAnyPlane, + double ReturnBandwidthToDCN); +static void ModeSupportAndSystemConfiguration(struct display_mode_lib *mode_lib); +static void DisplayPipeConfiguration(struct display_mode_lib *mode_lib); +static void DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation( + struct display_mode_lib *mode_lib); +static unsigned int dscceComputeDelay( + unsigned int bpc, + double bpp, + unsigned int sliceWidth, + unsigned int numSlices, + enum output_format_class pixelFormat); +static unsigned int dscComputeDelay(enum output_format_class pixelFormat); +// Super monster function with some 45 argument +static bool CalculatePrefetchSchedule( + struct display_mode_lib *mode_lib, + double DPPCLK, + double DISPCLK, + double PixelClock, + double DCFClkDeepSleep, + unsigned int DSCDelay, + unsigned int DPPPerPlane, + bool ScalerEnabled, + unsigned int NumberOfCursors, + double DPPCLKDelaySubtotal, + double DPPCLKDelaySCL, + double DPPCLKDelaySCLLBOnly, + double DPPCLKDelayCNVCFormater, + double DPPCLKDelayCNVCCursor, + double DISPCLKDelaySubtotal, + unsigned int ScalerRecoutWidth, + enum output_format_class OutputFormat, + unsigned int VBlank, + unsigned int HTotal, + unsigned int MaxInterDCNTileRepeaters, + unsigned int VStartup, + unsigned int PageTableLevels, + bool VirtualMemoryEnable, + bool DynamicMetadataEnable, + unsigned int DynamicMetadataLinesBeforeActiveRequired, + unsigned int DynamicMetadataTransmittedBytes, + bool DCCEnable, + double UrgentLatency, + double UrgentExtraLatency, + double TCalc, + unsigned int PDEAndMetaPTEBytesFrame, + unsigned int MetaRowByte, + unsigned int PixelPTEBytesPerRow, + double PrefetchSourceLinesY, + unsigned int SwathWidthY, + double BytePerPixelDETY, + double VInitPreFillY, + unsigned int MaxNumSwathY, + double PrefetchSourceLinesC, + double BytePerPixelDETC, + double VInitPreFillC, + unsigned int MaxNumSwathC, + unsigned int SwathHeightY, + unsigned int SwathHeightC, + double TWait, + bool XFCEnabled, + double XFCRemoteSurfaceFlipDelay, + bool InterlaceEnable, + bool ProgressiveToInterlaceUnitInOPP, + double *DSTXAfterScaler, + double *DSTYAfterScaler, + double *DestinationLinesForPrefetch, + double *PrefetchBandwidth, + double *DestinationLinesToRequestVMInVBlank, + double *DestinationLinesToRequestRowInVBlank, + double *VRatioPrefetchY, + double *VRatioPrefetchC, + double *RequiredPrefetchPixDataBW, + unsigned int *VStartupRequiredWhenNotEnoughTimeForDynamicMetadata, + double *Tno_bw, + unsigned int *VUpdateOffsetPix, + unsigned int *VUpdateWidthPix, + unsigned int *VReadyOffsetPix); +static double RoundToDFSGranularityUp(double Clock, double VCOSpeed); +static double RoundToDFSGranularityDown(double Clock, double VCOSpeed); +static double CalculatePrefetchSourceLines( + struct display_mode_lib *mode_lib, + double VRatio, + double vtaps, + bool Interlace, + bool ProgressiveToInterlaceUnitInOPP, + unsigned int SwathHeight, + unsigned int ViewportYStart, + double *VInitPreFill, + unsigned int *MaxNumSwath); +static unsigned int CalculateVMAndRowBytes( + struct display_mode_lib *mode_lib, + bool DCCEnable, + unsigned int BlockHeight256Bytes, + unsigned int BlockWidth256Bytes, + enum source_format_class SourcePixelFormat, + unsigned int SurfaceTiling, + unsigned int BytePerPixel, + enum scan_direction_class ScanDirection, + unsigned int ViewportWidth, + unsigned int ViewportHeight, + unsigned int SwathWidthY, + bool VirtualMemoryEnable, + unsigned int VMMPageSize, + unsigned int PTEBufferSizeInRequests, + unsigned int PDEProcessingBufIn64KBReqs, + unsigned int Pitch, + unsigned int DCCMetaPitch, + unsigned int *MacroTileWidth, + unsigned int *MetaRowByte, + unsigned int *PixelPTEBytesPerRow, + bool *PTEBufferSizeNotExceeded, + unsigned int *dpte_row_height, + unsigned int *meta_row_height); +static double CalculateTWait( + unsigned int PrefetchMode, + double DRAMClockChangeLatency, + double UrgentLatency, + double SREnterPlusExitTime); +static double CalculateRemoteSurfaceFlipDelay( + struct display_mode_lib *mode_lib, + double VRatio, + double SwathWidth, + double Bpp, + double LineTime, + double XFCTSlvVupdateOffset, + double XFCTSlvVupdateWidth, + double XFCTSlvVreadyOffset, + double XFCXBUFLatencyTolerance, + double XFCFillBWOverhead, + double XFCSlvChunkSize, + double XFCBusTransportTime, + double TCalc, + double TWait, + double *SrcActiveDrainRate, + double *TInitXFill, + double *TslvChk); +static double CalculateWriteBackDISPCLK( + enum source_format_class WritebackPixelFormat, + double PixelClock, + double WritebackHRatio, + double WritebackVRatio, + unsigned int WritebackLumaHTaps, + unsigned int WritebackLumaVTaps, + unsigned int WritebackChromaHTaps, + unsigned int WritebackChromaVTaps, + double WritebackDestinationWidth, + unsigned int HTotal, + unsigned int WritebackChromaLineBufferWidth); +static void CalculateActiveRowBandwidth( + bool VirtualMemoryEnable, + enum source_format_class SourcePixelFormat, + double VRatio, + bool DCCEnable, + double LineTime, + unsigned int MetaRowByteLuma, + unsigned int MetaRowByteChroma, + unsigned int meta_row_height_luma, + unsigned int meta_row_height_chroma, + unsigned int PixelPTEBytesPerRowLuma, + unsigned int PixelPTEBytesPerRowChroma, + unsigned int dpte_row_height_luma, + unsigned int dpte_row_height_chroma, + double *meta_row_bw, + double *dpte_row_bw, + double *qual_row_bw); +static void CalculateFlipSchedule( + struct display_mode_lib *mode_lib, + double UrgentExtraLatency, + double UrgentLatency, + unsigned int MaxPageTableLevels, + bool VirtualMemoryEnable, + double BandwidthAvailableForImmediateFlip, + unsigned int TotImmediateFlipBytes, + enum source_format_class SourcePixelFormat, + unsigned int ImmediateFlipBytes, + double LineTime, + double Tno_bw, + double VRatio, + double PDEAndMetaPTEBytesFrame, + unsigned int MetaRowByte, + unsigned int PixelPTEBytesPerRow, + bool DCCEnable, + unsigned int dpte_row_height, + unsigned int meta_row_height, + double qual_row_bw, + double *DestinationLinesToRequestVMInImmediateFlip, + double *DestinationLinesToRequestRowInImmediateFlip, + double *final_flip_bw, + bool *ImmediateFlipSupportedForPipe); +static double CalculateWriteBackDelay( + enum source_format_class WritebackPixelFormat, + double WritebackHRatio, + double WritebackVRatio, + unsigned int WritebackLumaHTaps, + unsigned int WritebackLumaVTaps, + unsigned int WritebackChromaHTaps, + unsigned int WritebackChromaVTaps, + unsigned int WritebackDestinationWidth); +static void PixelClockAdjustmentForProgressiveToInterlaceUnit(struct display_mode_lib *mode_lib); +static unsigned int CursorBppEnumToBits(enum cursor_bpp ebpp); +static void ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_lib); + +void set_prefetch_mode( + struct display_mode_lib *mode_lib, + bool cstate_en, + bool pstate_en, + bool ignore_viewport_pos, + bool immediate_flip_support) +{ + unsigned int prefetch_mode; + + if (cstate_en && pstate_en) + prefetch_mode = 0; + else if (cstate_en) + prefetch_mode = 1; + else + prefetch_mode = 2; + if (prefetch_mode != mode_lib->vba.PrefetchMode + || ignore_viewport_pos != mode_lib->vba.IgnoreViewportPositioning + || immediate_flip_support != mode_lib->vba.ImmediateFlipSupport) { + DTRACE( + " Prefetch mode has changed from %i to %i. Recalculating.", + prefetch_mode, + mode_lib->vba.PrefetchMode); + mode_lib->vba.PrefetchMode = prefetch_mode; + mode_lib->vba.IgnoreViewportPositioning = ignore_viewport_pos; + mode_lib->vba.ImmediateFlipSupport = immediate_flip_support; + recalculate(mode_lib); + } +} + +unsigned int dml_get_voltage_level( + struct display_mode_lib *mode_lib, + const display_e2e_pipe_params_st *pipes, + unsigned int num_pipes) +{ + bool need_recalculate = memcmp(&mode_lib->soc, &mode_lib->vba.soc, sizeof(mode_lib->vba.soc)) != 0 + || memcmp(&mode_lib->ip, &mode_lib->vba.ip, sizeof(mode_lib->vba.ip)) != 0 + || num_pipes != mode_lib->vba.cache_num_pipes + || memcmp(pipes, mode_lib->vba.cache_pipes, + sizeof(display_e2e_pipe_params_st) * num_pipes) != 0; + + mode_lib->vba.soc = mode_lib->soc; + mode_lib->vba.ip = mode_lib->ip; + memcpy(mode_lib->vba.cache_pipes, pipes, sizeof(*pipes) * num_pipes); + mode_lib->vba.cache_num_pipes = num_pipes; + + if (need_recalculate && pipes[0].clks_cfg.dppclk_mhz != 0) + recalculate(mode_lib); + else { + fetch_socbb_params(mode_lib); + fetch_ip_params(mode_lib); + fetch_pipe_params(mode_lib); + } + ModeSupportAndSystemConfigurationFull(mode_lib); + + return mode_lib->vba.VoltageLevel; +} + +#define dml_get_attr_func(attr, var) double get_##attr(struct display_mode_lib *mode_lib, const display_e2e_pipe_params_st *pipes, unsigned int num_pipes) \ +{ \ + recalculate_params(mode_lib, pipes, num_pipes); \ + return var; \ +} + +dml_get_attr_func(clk_dcf_deepsleep, mode_lib->vba.DCFClkDeepSleep); +dml_get_attr_func(wm_urgent, mode_lib->vba.UrgentWatermark); +dml_get_attr_func(wm_memory_trip, mode_lib->vba.MemoryTripWatermark); +dml_get_attr_func(wm_writeback_urgent, mode_lib->vba.WritebackUrgentWatermark); +dml_get_attr_func(wm_stutter_exit, mode_lib->vba.StutterExitWatermark); +dml_get_attr_func(wm_stutter_enter_exit, mode_lib->vba.StutterEnterPlusExitWatermark); +dml_get_attr_func(wm_dram_clock_change, mode_lib->vba.DRAMClockChangeWatermark); +dml_get_attr_func(wm_writeback_dram_clock_change, mode_lib->vba.WritebackDRAMClockChangeWatermark); +dml_get_attr_func(wm_xfc_underflow, mode_lib->vba.UrgentWatermark); // xfc_underflow maps to urgent +dml_get_attr_func(stutter_efficiency, mode_lib->vba.StutterEfficiency); +dml_get_attr_func(stutter_efficiency_no_vblank, mode_lib->vba.StutterEfficiencyNotIncludingVBlank); +dml_get_attr_func(urgent_latency, mode_lib->vba.MinUrgentLatencySupportUs); +dml_get_attr_func(urgent_extra_latency, mode_lib->vba.UrgentExtraLatency); +dml_get_attr_func(nonurgent_latency, mode_lib->vba.NonUrgentLatencyTolerance); +dml_get_attr_func( + dram_clock_change_latency, + mode_lib->vba.MinActiveDRAMClockChangeLatencySupported); +dml_get_attr_func(dispclk_calculated, mode_lib->vba.DISPCLK_calculated); +dml_get_attr_func(total_data_read_bw, mode_lib->vba.TotalDataReadBandwidth); +dml_get_attr_func(return_bw, mode_lib->vba.ReturnBW); +dml_get_attr_func(tcalc, mode_lib->vba.TCalc); + +#define dml_get_pipe_attr_func(attr, var) double get_##attr(struct display_mode_lib *mode_lib, const display_e2e_pipe_params_st *pipes, unsigned int num_pipes, unsigned int which_pipe) \ +{\ + unsigned int which_plane; \ + recalculate_params(mode_lib, pipes, num_pipes); \ + which_plane = mode_lib->vba.pipe_plane[which_pipe]; \ + return var[which_plane]; \ +} + +dml_get_pipe_attr_func(dsc_delay, mode_lib->vba.DSCDelay); +dml_get_pipe_attr_func(dppclk_calculated, mode_lib->vba.DPPCLK_calculated); +dml_get_pipe_attr_func(dscclk_calculated, mode_lib->vba.DSCCLK_calculated); +dml_get_pipe_attr_func(min_ttu_vblank, mode_lib->vba.MinTTUVBlank); +dml_get_pipe_attr_func(vratio_prefetch_l, mode_lib->vba.VRatioPrefetchY); +dml_get_pipe_attr_func(vratio_prefetch_c, mode_lib->vba.VRatioPrefetchC); +dml_get_pipe_attr_func(dst_x_after_scaler, mode_lib->vba.DSTXAfterScaler); +dml_get_pipe_attr_func(dst_y_after_scaler, mode_lib->vba.DSTYAfterScaler); +dml_get_pipe_attr_func(dst_y_per_vm_vblank, mode_lib->vba.DestinationLinesToRequestVMInVBlank); +dml_get_pipe_attr_func(dst_y_per_row_vblank, mode_lib->vba.DestinationLinesToRequestRowInVBlank); +dml_get_pipe_attr_func(dst_y_prefetch, mode_lib->vba.DestinationLinesForPrefetch); +dml_get_pipe_attr_func(dst_y_per_vm_flip, mode_lib->vba.DestinationLinesToRequestVMInImmediateFlip); +dml_get_pipe_attr_func( + dst_y_per_row_flip, + mode_lib->vba.DestinationLinesToRequestRowInImmediateFlip); + +dml_get_pipe_attr_func(xfc_transfer_delay, mode_lib->vba.XFCTransferDelay); +dml_get_pipe_attr_func(xfc_precharge_delay, mode_lib->vba.XFCPrechargeDelay); +dml_get_pipe_attr_func(xfc_remote_surface_flip_latency, mode_lib->vba.XFCRemoteSurfaceFlipLatency); +dml_get_pipe_attr_func(xfc_prefetch_margin, mode_lib->vba.XFCPrefetchMargin); + +unsigned int get_vstartup_calculated( + struct display_mode_lib *mode_lib, + const display_e2e_pipe_params_st *pipes, + unsigned int num_pipes, + unsigned int which_pipe) +{ + unsigned int which_plane; + + recalculate_params(mode_lib, pipes, num_pipes); + which_plane = mode_lib->vba.pipe_plane[which_pipe]; + return mode_lib->vba.VStartup[which_plane]; +} + +double get_total_immediate_flip_bytes( + struct display_mode_lib *mode_lib, + const display_e2e_pipe_params_st *pipes, + unsigned int num_pipes) +{ + recalculate_params(mode_lib, pipes, num_pipes); + return mode_lib->vba.TotImmediateFlipBytes; +} + +double get_total_immediate_flip_bw( + struct display_mode_lib *mode_lib, + const display_e2e_pipe_params_st *pipes, + unsigned int num_pipes) +{ + recalculate_params(mode_lib, pipes, num_pipes); + return mode_lib->vba.ImmediateFlipBW; +} + +double get_total_prefetch_bw( + struct display_mode_lib *mode_lib, + const display_e2e_pipe_params_st *pipes, + unsigned int num_pipes) +{ + unsigned int k; + double total_prefetch_bw = 0.0; + + recalculate_params(mode_lib, pipes, num_pipes); + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) + total_prefetch_bw += mode_lib->vba.PrefetchBandwidth[k]; + return total_prefetch_bw; +} + +static void fetch_socbb_params(struct display_mode_lib *mode_lib) +{ + soc_bounding_box_st *soc = &mode_lib->vba.soc; + unsigned int i; + + // SOC Bounding Box Parameters + mode_lib->vba.ReturnBusWidth = soc->return_bus_width_bytes; + mode_lib->vba.NumberOfChannels = soc->num_chans; + mode_lib->vba.PercentOfIdealDRAMAndFabricBWReceivedAfterUrgLatency = + soc->ideal_dram_bw_after_urgent_percent; // there's always that one bastard variable that's so long it throws everything out of alignment! + mode_lib->vba.UrgentLatency = soc->urgent_latency_us; + mode_lib->vba.RoundTripPingLatencyCycles = soc->round_trip_ping_latency_dcfclk_cycles; + mode_lib->vba.UrgentOutOfOrderReturnPerChannel = + soc->urgent_out_of_order_return_per_channel_bytes; + mode_lib->vba.WritebackLatency = soc->writeback_latency_us; + mode_lib->vba.SRExitTime = soc->sr_exit_time_us; + mode_lib->vba.SREnterPlusExitTime = soc->sr_enter_plus_exit_time_us; + mode_lib->vba.DRAMClockChangeLatency = soc->dram_clock_change_latency_us; + mode_lib->vba.Downspreading = soc->downspread_percent; + mode_lib->vba.DRAMChannelWidth = soc->dram_channel_width_bytes; // new! + mode_lib->vba.FabricDatapathToDCNDataReturn = soc->fabric_datapath_to_dcn_data_return_bytes; // new! + mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading = soc->dcn_downspread_percent; // new + mode_lib->vba.DISPCLKDPPCLKVCOSpeed = soc->dispclk_dppclk_vco_speed_mhz; // new + mode_lib->vba.VMMPageSize = soc->vmm_page_size_bytes; + // Set the voltage scaling clocks as the defaults. Most of these will + // be set to different values by the test + for (i = 0; i < DC__VOLTAGE_STATES; i++) + if (soc->clock_limits[i].state == mode_lib->vba.VoltageLevel) + break; + + mode_lib->vba.DCFCLK = soc->clock_limits[i].dcfclk_mhz; + mode_lib->vba.SOCCLK = soc->clock_limits[i].socclk_mhz; + mode_lib->vba.DRAMSpeed = soc->clock_limits[i].dram_speed_mhz; + mode_lib->vba.FabricClock = soc->clock_limits[i].fabricclk_mhz; + + mode_lib->vba.XFCBusTransportTime = soc->xfc_bus_transport_time_us; + mode_lib->vba.XFCXBUFLatencyTolerance = soc->xfc_xbuf_latency_tolerance_us; + + mode_lib->vba.SupportGFX7CompatibleTilingIn32bppAnd64bpp = false; + mode_lib->vba.MaxHSCLRatio = 4; + mode_lib->vba.MaxVSCLRatio = 4; + mode_lib->vba.MaxNumWriteback = 0; /*TODO*/ + mode_lib->vba.WritebackLumaAndChromaScalingSupported = true; + mode_lib->vba.Cursor64BppSupport = true; + for (i = 0; i <= DC__VOLTAGE_STATES; i++) { + mode_lib->vba.DCFCLKPerState[i] = soc->clock_limits[i].dcfclk_mhz; + mode_lib->vba.FabricClockPerState[i] = soc->clock_limits[i].fabricclk_mhz; + mode_lib->vba.SOCCLKPerState[i] = soc->clock_limits[i].socclk_mhz; + mode_lib->vba.PHYCLKPerState[i] = soc->clock_limits[i].phyclk_mhz; + mode_lib->vba.MaxDppclk[i] = soc->clock_limits[i].dppclk_mhz; + mode_lib->vba.MaxDSCCLK[i] = soc->clock_limits[i].dscclk_mhz; + mode_lib->vba.DRAMSpeedPerState[i] = soc->clock_limits[i].dram_speed_mhz; + mode_lib->vba.MaxDispclk[i] = soc->clock_limits[i].dispclk_mhz; + } +} + +static void fetch_ip_params(struct display_mode_lib *mode_lib) +{ + ip_params_st *ip = &mode_lib->vba.ip; + + // IP Parameters + mode_lib->vba.MaxNumDPP = ip->max_num_dpp; + mode_lib->vba.MaxNumOTG = ip->max_num_otg; + mode_lib->vba.CursorChunkSize = ip->cursor_chunk_size; + mode_lib->vba.CursorBufferSize = ip->cursor_buffer_size; + + mode_lib->vba.MaxDCHUBToPSCLThroughput = ip->max_dchub_pscl_bw_pix_per_clk; + mode_lib->vba.MaxPSCLToLBThroughput = ip->max_pscl_lb_bw_pix_per_clk; + mode_lib->vba.ROBBufferSizeInKByte = ip->rob_buffer_size_kbytes; + mode_lib->vba.DETBufferSizeInKByte = ip->det_buffer_size_kbytes; + mode_lib->vba.PixelChunkSizeInKByte = ip->pixel_chunk_size_kbytes; + mode_lib->vba.MetaChunkSize = ip->meta_chunk_size_kbytes; + mode_lib->vba.PTEChunkSize = ip->pte_chunk_size_kbytes; + mode_lib->vba.WritebackChunkSize = ip->writeback_chunk_size_kbytes; + mode_lib->vba.LineBufferSize = ip->line_buffer_size_bits; + mode_lib->vba.MaxLineBufferLines = ip->max_line_buffer_lines; + mode_lib->vba.PTEBufferSizeInRequests = ip->dpte_buffer_size_in_pte_reqs; + mode_lib->vba.DPPOutputBufferPixels = ip->dpp_output_buffer_pixels; + mode_lib->vba.OPPOutputBufferLines = ip->opp_output_buffer_lines; + mode_lib->vba.WritebackInterfaceLumaBufferSize = ip->writeback_luma_buffer_size_kbytes; + mode_lib->vba.WritebackInterfaceChromaBufferSize = ip->writeback_chroma_buffer_size_kbytes; + mode_lib->vba.WritebackChromaLineBufferWidth = + ip->writeback_chroma_line_buffer_width_pixels; + mode_lib->vba.MaxPageTableLevels = ip->max_page_table_levels; + mode_lib->vba.MaxInterDCNTileRepeaters = ip->max_inter_dcn_tile_repeaters; + mode_lib->vba.NumberOfDSC = ip->num_dsc; + mode_lib->vba.ODMCapability = ip->odm_capable; + mode_lib->vba.DISPCLKRampingMargin = ip->dispclk_ramp_margin_percent; + + mode_lib->vba.XFCSupported = ip->xfc_supported; + mode_lib->vba.XFCFillBWOverhead = ip->xfc_fill_bw_overhead_percent; + mode_lib->vba.XFCFillConstant = ip->xfc_fill_constant_bytes; + mode_lib->vba.DPPCLKDelaySubtotal = ip->dppclk_delay_subtotal; + mode_lib->vba.DPPCLKDelaySCL = ip->dppclk_delay_scl; + mode_lib->vba.DPPCLKDelaySCLLBOnly = ip->dppclk_delay_scl_lb_only; + mode_lib->vba.DPPCLKDelayCNVCFormater = ip->dppclk_delay_cnvc_formatter; + mode_lib->vba.DPPCLKDelayCNVCCursor = ip->dppclk_delay_cnvc_cursor; + mode_lib->vba.DISPCLKDelaySubtotal = ip->dispclk_delay_subtotal; + + mode_lib->vba.ProgressiveToInterlaceUnitInOPP = ip->ptoi_supported; + + mode_lib->vba.PDEProcessingBufIn64KBReqs = ip->pde_proc_buffer_size_64k_reqs; +} + +static void fetch_pipe_params(struct display_mode_lib *mode_lib) +{ + display_e2e_pipe_params_st *pipes = mode_lib->vba.cache_pipes; + ip_params_st *ip = &mode_lib->vba.ip; + + unsigned int OTGInstPlane[DC__NUM_DPP__MAX]; + unsigned int j, k; + bool PlaneVisited[DC__NUM_DPP__MAX]; + bool visited[DC__NUM_DPP__MAX]; + + // Convert Pipes to Planes + for (k = 0; k < mode_lib->vba.cache_num_pipes; ++k) + visited[k] = false; + + mode_lib->vba.NumberOfActivePlanes = 0; + for (j = 0; j < mode_lib->vba.cache_num_pipes; ++j) { + display_pipe_source_params_st *src = &pipes[j].pipe.src; + display_pipe_dest_params_st *dst = &pipes[j].pipe.dest; + scaler_ratio_depth_st *scl = &pipes[j].pipe.scale_ratio_depth; + scaler_taps_st *taps = &pipes[j].pipe.scale_taps; + display_output_params_st *dout = &pipes[j].dout; + display_clocks_and_cfg_st *clks = &pipes[j].clks_cfg; + + if (visited[j]) + continue; + visited[j] = true; + + mode_lib->vba.pipe_plane[j] = mode_lib->vba.NumberOfActivePlanes; + + mode_lib->vba.DPPPerPlane[mode_lib->vba.NumberOfActivePlanes] = 1; + mode_lib->vba.SourceScan[mode_lib->vba.NumberOfActivePlanes] = + (enum scan_direction_class) (src->source_scan); + mode_lib->vba.ViewportWidth[mode_lib->vba.NumberOfActivePlanes] = + src->viewport_width; + mode_lib->vba.ViewportHeight[mode_lib->vba.NumberOfActivePlanes] = + src->viewport_height; + mode_lib->vba.ViewportYStartY[mode_lib->vba.NumberOfActivePlanes] = + src->viewport_y_y; + mode_lib->vba.ViewportYStartC[mode_lib->vba.NumberOfActivePlanes] = + src->viewport_y_c; + mode_lib->vba.PitchY[mode_lib->vba.NumberOfActivePlanes] = src->data_pitch; + mode_lib->vba.PitchC[mode_lib->vba.NumberOfActivePlanes] = src->data_pitch_c; + mode_lib->vba.DCCMetaPitchY[mode_lib->vba.NumberOfActivePlanes] = src->meta_pitch; + mode_lib->vba.HRatio[mode_lib->vba.NumberOfActivePlanes] = scl->hscl_ratio; + mode_lib->vba.VRatio[mode_lib->vba.NumberOfActivePlanes] = scl->vscl_ratio; + mode_lib->vba.ScalerEnabled[mode_lib->vba.NumberOfActivePlanes] = scl->scl_enable; + mode_lib->vba.Interlace[mode_lib->vba.NumberOfActivePlanes] = dst->interlaced; + if (mode_lib->vba.Interlace[mode_lib->vba.NumberOfActivePlanes]) + mode_lib->vba.VRatio[mode_lib->vba.NumberOfActivePlanes] *= 2.0; + mode_lib->vba.htaps[mode_lib->vba.NumberOfActivePlanes] = taps->htaps; + mode_lib->vba.vtaps[mode_lib->vba.NumberOfActivePlanes] = taps->vtaps; + mode_lib->vba.HTAPsChroma[mode_lib->vba.NumberOfActivePlanes] = taps->htaps_c; + mode_lib->vba.VTAPsChroma[mode_lib->vba.NumberOfActivePlanes] = taps->vtaps_c; + mode_lib->vba.HTotal[mode_lib->vba.NumberOfActivePlanes] = dst->htotal; + mode_lib->vba.VTotal[mode_lib->vba.NumberOfActivePlanes] = dst->vtotal; + mode_lib->vba.DCCEnable[mode_lib->vba.NumberOfActivePlanes] = + src->dcc_use_global ? + ip->dcc_supported : src->dcc && ip->dcc_supported; + mode_lib->vba.DCCRate[mode_lib->vba.NumberOfActivePlanes] = src->dcc_rate; + mode_lib->vba.SourcePixelFormat[mode_lib->vba.NumberOfActivePlanes] = + (enum source_format_class) (src->source_format); + mode_lib->vba.HActive[mode_lib->vba.NumberOfActivePlanes] = dst->hactive; + mode_lib->vba.VActive[mode_lib->vba.NumberOfActivePlanes] = dst->vactive; + mode_lib->vba.SurfaceTiling[mode_lib->vba.NumberOfActivePlanes] = + (enum dm_swizzle_mode) (src->sw_mode); + mode_lib->vba.ScalerRecoutWidth[mode_lib->vba.NumberOfActivePlanes] = + dst->recout_width; // TODO: or should this be full_recout_width???...maybe only when in hsplit mode? + mode_lib->vba.ODMCombineEnabled[mode_lib->vba.NumberOfActivePlanes] = + dst->odm_combine; + mode_lib->vba.OutputFormat[mode_lib->vba.NumberOfActivePlanes] = + (enum output_format_class) (dout->output_format); + mode_lib->vba.Output[mode_lib->vba.NumberOfActivePlanes] = + (enum output_encoder_class) (dout->output_type); + mode_lib->vba.OutputBpp[mode_lib->vba.NumberOfActivePlanes] = dout->output_bpp; + mode_lib->vba.OutputLinkDPLanes[mode_lib->vba.NumberOfActivePlanes] = + dout->dp_lanes; + mode_lib->vba.DSCEnabled[mode_lib->vba.NumberOfActivePlanes] = dout->dsc_enable; + mode_lib->vba.NumberOfDSCSlices[mode_lib->vba.NumberOfActivePlanes] = + dout->dsc_slices; + mode_lib->vba.DSCInputBitPerComponent[mode_lib->vba.NumberOfActivePlanes] = + dout->output_bpc == 0 ? 12 : dout->output_bpc; + mode_lib->vba.WritebackEnable[mode_lib->vba.NumberOfActivePlanes] = dout->wb_enable; + mode_lib->vba.WritebackSourceHeight[mode_lib->vba.NumberOfActivePlanes] = + dout->wb.wb_src_height; + mode_lib->vba.WritebackDestinationWidth[mode_lib->vba.NumberOfActivePlanes] = + dout->wb.wb_dst_width; + mode_lib->vba.WritebackDestinationHeight[mode_lib->vba.NumberOfActivePlanes] = + dout->wb.wb_dst_height; + mode_lib->vba.WritebackPixelFormat[mode_lib->vba.NumberOfActivePlanes] = + (enum source_format_class) (dout->wb.wb_pixel_format); + mode_lib->vba.WritebackLumaHTaps[mode_lib->vba.NumberOfActivePlanes] = + dout->wb.wb_htaps_luma; + mode_lib->vba.WritebackLumaVTaps[mode_lib->vba.NumberOfActivePlanes] = + dout->wb.wb_vtaps_luma; + mode_lib->vba.WritebackChromaHTaps[mode_lib->vba.NumberOfActivePlanes] = + dout->wb.wb_htaps_chroma; + mode_lib->vba.WritebackChromaVTaps[mode_lib->vba.NumberOfActivePlanes] = + dout->wb.wb_vtaps_chroma; + mode_lib->vba.WritebackHRatio[mode_lib->vba.NumberOfActivePlanes] = + dout->wb.wb_hratio; + mode_lib->vba.WritebackVRatio[mode_lib->vba.NumberOfActivePlanes] = + dout->wb.wb_vratio; + + mode_lib->vba.DynamicMetadataEnable[mode_lib->vba.NumberOfActivePlanes] = + src->dynamic_metadata_enable; + mode_lib->vba.DynamicMetadataLinesBeforeActiveRequired[mode_lib->vba.NumberOfActivePlanes] = + src->dynamic_metadata_lines_before_active; + mode_lib->vba.DynamicMetadataTransmittedBytes[mode_lib->vba.NumberOfActivePlanes] = + src->dynamic_metadata_xmit_bytes; + + mode_lib->vba.XFCEnabled[mode_lib->vba.NumberOfActivePlanes] = src->xfc_enable + && ip->xfc_supported; + mode_lib->vba.XFCSlvChunkSize = src->xfc_params.xfc_slv_chunk_size_bytes; + mode_lib->vba.XFCTSlvVupdateOffset = src->xfc_params.xfc_tslv_vupdate_offset_us; + mode_lib->vba.XFCTSlvVupdateWidth = src->xfc_params.xfc_tslv_vupdate_width_us; + mode_lib->vba.XFCTSlvVreadyOffset = src->xfc_params.xfc_tslv_vready_offset_us; + mode_lib->vba.PixelClock[mode_lib->vba.NumberOfActivePlanes] = dst->pixel_rate_mhz; + mode_lib->vba.DPPCLK[mode_lib->vba.NumberOfActivePlanes] = clks->dppclk_mhz; + if (ip->is_line_buffer_bpp_fixed) + mode_lib->vba.LBBitPerPixel[mode_lib->vba.NumberOfActivePlanes] = + ip->line_buffer_fixed_bpp; + else { + unsigned int lb_depth; + + switch (scl->lb_depth) { + case dm_lb_6: + lb_depth = 18; + break; + case dm_lb_8: + lb_depth = 24; + break; + case dm_lb_10: + lb_depth = 30; + break; + case dm_lb_12: + lb_depth = 36; + break; + case dm_lb_16: + lb_depth = 48; + break; + default: + lb_depth = 36; + } + mode_lib->vba.LBBitPerPixel[mode_lib->vba.NumberOfActivePlanes] = lb_depth; + } + mode_lib->vba.NumberOfCursors[mode_lib->vba.NumberOfActivePlanes] = 0; + // The DML spreadsheet assumes that the two cursors utilize the same amount of bandwidth. We'll + // calculate things a little more accurately + for (k = 0; k < DC__NUM_CURSOR__MAX; ++k) { + switch (k) { + case 0: + mode_lib->vba.CursorBPP[mode_lib->vba.NumberOfActivePlanes][0] = + CursorBppEnumToBits( + (enum cursor_bpp) (src->cur0_bpp)); + mode_lib->vba.CursorWidth[mode_lib->vba.NumberOfActivePlanes][0] = + src->cur0_src_width; + if (src->cur0_src_width > 0) + mode_lib->vba.NumberOfCursors[mode_lib->vba.NumberOfActivePlanes]++; + break; + case 1: + mode_lib->vba.CursorBPP[mode_lib->vba.NumberOfActivePlanes][1] = + CursorBppEnumToBits( + (enum cursor_bpp) (src->cur1_bpp)); + mode_lib->vba.CursorWidth[mode_lib->vba.NumberOfActivePlanes][1] = + src->cur1_src_width; + if (src->cur1_src_width > 0) + mode_lib->vba.NumberOfCursors[mode_lib->vba.NumberOfActivePlanes]++; + break; + default: + dml_print( + "ERROR: Number of cursors specified exceeds supported maximum\n") + ; + } + } + + OTGInstPlane[mode_lib->vba.NumberOfActivePlanes] = dst->otg_inst; + + if (dst->odm_combine && !src->is_hsplit) + dml_print( + "ERROR: ODM Combine is specified but is_hsplit has not be specified for pipe %i\n", + j); + + if (src->is_hsplit) { + for (k = j + 1; k < mode_lib->vba.cache_num_pipes; ++k) { + display_pipe_source_params_st *src_k = &pipes[k].pipe.src; + display_output_params_st *dout_k = &pipes[k].dout; + + if (src_k->is_hsplit && !visited[k] + && src->hsplit_grp == src_k->hsplit_grp) { + mode_lib->vba.pipe_plane[k] = + mode_lib->vba.NumberOfActivePlanes; + mode_lib->vba.DPPPerPlane[mode_lib->vba.NumberOfActivePlanes]++; + if (mode_lib->vba.SourceScan[mode_lib->vba.NumberOfActivePlanes] + == dm_horz) + mode_lib->vba.ViewportWidth[mode_lib->vba.NumberOfActivePlanes] += + src_k->viewport_width; + else + mode_lib->vba.ViewportHeight[mode_lib->vba.NumberOfActivePlanes] += + src_k->viewport_height; + + mode_lib->vba.NumberOfDSCSlices[mode_lib->vba.NumberOfActivePlanes] += + dout_k->dsc_slices; + visited[k] = true; + } + } + } + + mode_lib->vba.NumberOfActivePlanes++; + } + + // handle overlays through dml_ml->vba.BlendingAndTiming + // dml_ml->vba.BlendingAndTiming tells you which instance to look at to get timing, the so called 'master' + + for (j = 0; j < mode_lib->vba.NumberOfActivePlanes; ++j) + PlaneVisited[j] = false; + + for (j = 0; j < mode_lib->vba.NumberOfActivePlanes; ++j) { + for (k = j + 1; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + if (!PlaneVisited[k] && OTGInstPlane[j] == OTGInstPlane[k]) { + // doesn't matter, so choose the smaller one + mode_lib->vba.BlendingAndTiming[j] = j; + PlaneVisited[j] = true; + mode_lib->vba.BlendingAndTiming[k] = j; + PlaneVisited[k] = true; + } + } + + if (!PlaneVisited[j]) { + mode_lib->vba.BlendingAndTiming[j] = j; + PlaneVisited[j] = true; + } + } + + // TODO: dml_ml->vba.ODMCombineEnabled => 2 * dml_ml->vba.DPPPerPlane...actually maybe not since all pipes are specified + // Do we want the dscclk to automatically be halved? Guess not since the value is specified + + mode_lib->vba.SynchronizedVBlank = pipes[0].pipe.dest.synchronized_vblank_all_planes; + for (k = 1; k < mode_lib->vba.cache_num_pipes; ++k) + ASSERT(mode_lib->vba.SynchronizedVBlank == pipes[k].pipe.dest.synchronized_vblank_all_planes); + + mode_lib->vba.VirtualMemoryEnable = false; + mode_lib->vba.OverridePageTableLevels = 0; + + for (k = 0; k < mode_lib->vba.cache_num_pipes; ++k) { + mode_lib->vba.VirtualMemoryEnable = mode_lib->vba.VirtualMemoryEnable + || !!pipes[k].pipe.src.vm; + mode_lib->vba.OverridePageTableLevels = + (pipes[k].pipe.src.vm_levels_force_en + && mode_lib->vba.OverridePageTableLevels + < pipes[k].pipe.src.vm_levels_force) ? + pipes[k].pipe.src.vm_levels_force : + mode_lib->vba.OverridePageTableLevels; + } + + if (mode_lib->vba.OverridePageTableLevels) + mode_lib->vba.MaxPageTableLevels = mode_lib->vba.OverridePageTableLevels; + + mode_lib->vba.VirtualMemoryEnable = mode_lib->vba.VirtualMemoryEnable && !!ip->pte_enable; + + mode_lib->vba.FabricAndDRAMBandwidth = dml_min( + mode_lib->vba.DRAMSpeed * mode_lib->vba.NumberOfChannels + * mode_lib->vba.DRAMChannelWidth, + mode_lib->vba.FabricClock * mode_lib->vba.FabricDatapathToDCNDataReturn) + / 1000.0; + + // TODO: Must be consistent across all pipes + // DCCProgrammingAssumesScanDirectionUnknown = src.dcc_scan_dir_unknown; +} + +static void recalculate(struct display_mode_lib *mode_lib) +{ + ModeSupportAndSystemConfiguration(mode_lib); + PixelClockAdjustmentForProgressiveToInterlaceUnit(mode_lib); + DisplayPipeConfiguration(mode_lib); + DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation(mode_lib); +} + +// in wm mode we pull the parameters needed from the display_e2e_pipe_params_st structs +// rather than working them out as in recalculate_ms +static void recalculate_params( + struct display_mode_lib *mode_lib, + const display_e2e_pipe_params_st *pipes, + unsigned int num_pipes) +{ + // This is only safe to use memcmp because there are non-POD types in struct display_mode_lib + if (memcmp(&mode_lib->soc, &mode_lib->vba.soc, sizeof(mode_lib->vba.soc)) != 0 + || memcmp(&mode_lib->ip, &mode_lib->vba.ip, sizeof(mode_lib->vba.ip)) != 0 + || num_pipes != mode_lib->vba.cache_num_pipes + || memcmp( + pipes, + mode_lib->vba.cache_pipes, + sizeof(display_e2e_pipe_params_st) * num_pipes) != 0) { + mode_lib->vba.soc = mode_lib->soc; + mode_lib->vba.ip = mode_lib->ip; + memcpy(mode_lib->vba.cache_pipes, pipes, sizeof(*pipes) * num_pipes); + mode_lib->vba.cache_num_pipes = num_pipes; + recalculate(mode_lib); + } +} + +static void ModeSupportAndSystemConfiguration(struct display_mode_lib *mode_lib) +{ + soc_bounding_box_st *soc = &mode_lib->vba.soc; + unsigned int i, k; + unsigned int total_pipes = 0; + + mode_lib->vba.VoltageLevel = mode_lib->vba.cache_pipes[0].clks_cfg.voltage; + for (i = 1; i < mode_lib->vba.cache_num_pipes; ++i) + ASSERT(mode_lib->vba.VoltageLevel == -1 || mode_lib->vba.VoltageLevel == mode_lib->vba.cache_pipes[i].clks_cfg.voltage); + + mode_lib->vba.DCFCLK = mode_lib->vba.cache_pipes[0].clks_cfg.dcfclk_mhz; + mode_lib->vba.SOCCLK = mode_lib->vba.cache_pipes[0].clks_cfg.socclk_mhz; + + if (mode_lib->vba.cache_pipes[0].clks_cfg.dispclk_mhz > 0.0) + mode_lib->vba.DISPCLK = mode_lib->vba.cache_pipes[0].clks_cfg.dispclk_mhz; + else + mode_lib->vba.DISPCLK = soc->clock_limits[mode_lib->vba.VoltageLevel].dispclk_mhz; + + fetch_socbb_params(mode_lib); + fetch_ip_params(mode_lib); + fetch_pipe_params(mode_lib); + + // Total Available Pipes Support Check + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) + total_pipes += mode_lib->vba.DPPPerPlane[k]; + ASSERT(total_pipes <= DC__NUM_DPP__MAX); +} + +static double adjust_ReturnBW( + struct display_mode_lib *mode_lib, + double ReturnBW, + bool DCCEnabledAnyPlane, + double ReturnBandwidthToDCN) +{ + double CriticalCompression; + + if (DCCEnabledAnyPlane + && ReturnBandwidthToDCN + > mode_lib->vba.DCFCLK * mode_lib->vba.ReturnBusWidth / 4.0) + ReturnBW = + dml_min( + ReturnBW, + ReturnBandwidthToDCN * 4 + * (1.0 + - mode_lib->vba.UrgentLatency + / ((mode_lib->vba.ROBBufferSizeInKByte + - mode_lib->vba.PixelChunkSizeInKByte) + * 1024 + / ReturnBandwidthToDCN + - mode_lib->vba.DCFCLK + * mode_lib->vba.ReturnBusWidth + / 4) + + mode_lib->vba.UrgentLatency)); + + CriticalCompression = 2.0 * mode_lib->vba.ReturnBusWidth * mode_lib->vba.DCFCLK + * mode_lib->vba.UrgentLatency + / (ReturnBandwidthToDCN * mode_lib->vba.UrgentLatency + + (mode_lib->vba.ROBBufferSizeInKByte + - mode_lib->vba.PixelChunkSizeInKByte) + * 1024); + + if (DCCEnabledAnyPlane && CriticalCompression > 1.0 && CriticalCompression < 4.0) + ReturnBW = + dml_min( + ReturnBW, + 4.0 * ReturnBandwidthToDCN + * (mode_lib->vba.ROBBufferSizeInKByte + - mode_lib->vba.PixelChunkSizeInKByte) + * 1024 + * mode_lib->vba.ReturnBusWidth + * mode_lib->vba.DCFCLK + * mode_lib->vba.UrgentLatency + / dml_pow( + (ReturnBandwidthToDCN + * mode_lib->vba.UrgentLatency + + (mode_lib->vba.ROBBufferSizeInKByte + - mode_lib->vba.PixelChunkSizeInKByte) + * 1024), + 2)); + + return ReturnBW; +} + +static unsigned int dscceComputeDelay( + unsigned int bpc, + double bpp, + unsigned int sliceWidth, + unsigned int numSlices, + enum output_format_class pixelFormat) +{ + // valid bpc = source bits per component in the set of {8, 10, 12} + // valid bpp = increments of 1/16 of a bit + // min = 6/7/8 in N420/N422/444, respectively + // max = such that compression is 1:1 + //valid sliceWidth = number of pixels per slice line, must be less than or equal to 5184/numSlices (or 4096/numSlices in 420 mode) + //valid numSlices = number of slices in the horiziontal direction per DSC engine in the set of {1, 2, 3, 4} + //valid pixelFormat = pixel/color format in the set of {:N444_RGB, :S422, :N422, :N420} + + // fixed value + unsigned int rcModelSize = 8192; + + // N422/N420 operate at 2 pixels per clock + unsigned int pixelsPerClock, lstall, D, initalXmitDelay, w, s, ix, wx, p, l0, a, ax, l, + Delay, pixels; + + if (pixelFormat == dm_n422 || pixelFormat == dm_420) + pixelsPerClock = 2; + // #all other modes operate at 1 pixel per clock + else + pixelsPerClock = 1; + + //initial transmit delay as per PPS + initalXmitDelay = dml_round(rcModelSize / 2.0 / bpp / pixelsPerClock); + + //compute ssm delay + if (bpc == 8) + D = 81; + else if (bpc == 10) + D = 89; + else + D = 113; + + //divide by pixel per cycle to compute slice width as seen by DSC + w = sliceWidth / pixelsPerClock; + + //422 mode has an additional cycle of delay + if (pixelFormat == dm_s422) + s = 1; + else + s = 0; + + //main calculation for the dscce + ix = initalXmitDelay + 45; + wx = (w + 2) / 3; + p = 3 * wx - w; + l0 = ix / w; + a = ix + p * l0; + ax = (a + 2) / 3 + D + 6 + 1; + l = (ax + wx - 1) / wx; + if ((ix % w) == 0 && p != 0) + lstall = 1; + else + lstall = 0; + Delay = l * wx * (numSlices - 1) + ax + s + lstall + 22; + + //dsc processes 3 pixel containers per cycle and a container can contain 1 or 2 pixels + pixels = Delay * 3 * pixelsPerClock; + return pixels; +} + +static unsigned int dscComputeDelay(enum output_format_class pixelFormat) +{ + unsigned int Delay = 0; + + if (pixelFormat == dm_420) { + // sfr + Delay = Delay + 2; + // dsccif + Delay = Delay + 0; + // dscc - input deserializer + Delay = Delay + 3; + // dscc gets pixels every other cycle + Delay = Delay + 2; + // dscc - input cdc fifo + Delay = Delay + 12; + // dscc gets pixels every other cycle + Delay = Delay + 13; + // dscc - cdc uncertainty + Delay = Delay + 2; + // dscc - output cdc fifo + Delay = Delay + 7; + // dscc gets pixels every other cycle + Delay = Delay + 3; + // dscc - cdc uncertainty + Delay = Delay + 2; + // dscc - output serializer + Delay = Delay + 1; + // sft + Delay = Delay + 1; + } else if (pixelFormat == dm_n422) { + // sfr + Delay = Delay + 2; + // dsccif + Delay = Delay + 1; + // dscc - input deserializer + Delay = Delay + 5; + // dscc - input cdc fifo + Delay = Delay + 25; + // dscc - cdc uncertainty + Delay = Delay + 2; + // dscc - output cdc fifo + Delay = Delay + 10; + // dscc - cdc uncertainty + Delay = Delay + 2; + // dscc - output serializer + Delay = Delay + 1; + // sft + Delay = Delay + 1; + } else { + // sfr + Delay = Delay + 2; + // dsccif + Delay = Delay + 0; + // dscc - input deserializer + Delay = Delay + 3; + // dscc - input cdc fifo + Delay = Delay + 12; + // dscc - cdc uncertainty + Delay = Delay + 2; + // dscc - output cdc fifo + Delay = Delay + 7; + // dscc - output serializer + Delay = Delay + 1; + // dscc - cdc uncertainty + Delay = Delay + 2; + // sft + Delay = Delay + 1; + } + + return Delay; +} + +static bool CalculatePrefetchSchedule( + struct display_mode_lib *mode_lib, + double DPPCLK, + double DISPCLK, + double PixelClock, + double DCFClkDeepSleep, + unsigned int DSCDelay, + unsigned int DPPPerPlane, + bool ScalerEnabled, + unsigned int NumberOfCursors, + double DPPCLKDelaySubtotal, + double DPPCLKDelaySCL, + double DPPCLKDelaySCLLBOnly, + double DPPCLKDelayCNVCFormater, + double DPPCLKDelayCNVCCursor, + double DISPCLKDelaySubtotal, + unsigned int ScalerRecoutWidth, + enum output_format_class OutputFormat, + unsigned int VBlank, + unsigned int HTotal, + unsigned int MaxInterDCNTileRepeaters, + unsigned int VStartup, + unsigned int PageTableLevels, + bool VirtualMemoryEnable, + bool DynamicMetadataEnable, + unsigned int DynamicMetadataLinesBeforeActiveRequired, + unsigned int DynamicMetadataTransmittedBytes, + bool DCCEnable, + double UrgentLatency, + double UrgentExtraLatency, + double TCalc, + unsigned int PDEAndMetaPTEBytesFrame, + unsigned int MetaRowByte, + unsigned int PixelPTEBytesPerRow, + double PrefetchSourceLinesY, + unsigned int SwathWidthY, + double BytePerPixelDETY, + double VInitPreFillY, + unsigned int MaxNumSwathY, + double PrefetchSourceLinesC, + double BytePerPixelDETC, + double VInitPreFillC, + unsigned int MaxNumSwathC, + unsigned int SwathHeightY, + unsigned int SwathHeightC, + double TWait, + bool XFCEnabled, + double XFCRemoteSurfaceFlipDelay, + bool InterlaceEnable, + bool ProgressiveToInterlaceUnitInOPP, + double *DSTXAfterScaler, + double *DSTYAfterScaler, + double *DestinationLinesForPrefetch, + double *PrefetchBandwidth, + double *DestinationLinesToRequestVMInVBlank, + double *DestinationLinesToRequestRowInVBlank, + double *VRatioPrefetchY, + double *VRatioPrefetchC, + double *RequiredPrefetchPixDataBW, + unsigned int *VStartupRequiredWhenNotEnoughTimeForDynamicMetadata, + double *Tno_bw, + unsigned int *VUpdateOffsetPix, + unsigned int *VUpdateWidthPix, + unsigned int *VReadyOffsetPix) +{ + bool MyError = false; + unsigned int DPPCycles, DISPCLKCycles; + double DSTTotalPixelsAfterScaler, TotalRepeaterDelayTime; + double Tdm, LineTime, Tsetup; + double dst_y_prefetch_equ; + double Tsw_oto; + double prefetch_bw_oto; + double Tvm_oto; + double Tr0_oto; + double Tpre_oto; + double dst_y_prefetch_oto; + double TimeForFetchingMetaPTE = 0; + double TimeForFetchingRowInVBlank = 0; + double LinesToRequestPrefetchPixelData = 0; + + if (ScalerEnabled) + DPPCycles = DPPCLKDelaySubtotal + DPPCLKDelaySCL; + else + DPPCycles = DPPCLKDelaySubtotal + DPPCLKDelaySCLLBOnly; + + DPPCycles = DPPCycles + DPPCLKDelayCNVCFormater + NumberOfCursors * DPPCLKDelayCNVCCursor; + + DISPCLKCycles = DISPCLKDelaySubtotal; + + if (DPPCLK == 0.0 || DISPCLK == 0.0) + return true; + + *DSTXAfterScaler = DPPCycles * PixelClock / DPPCLK + DISPCLKCycles * PixelClock / DISPCLK + + DSCDelay; + + if (DPPPerPlane > 1) + *DSTXAfterScaler = *DSTXAfterScaler + ScalerRecoutWidth; + + if (OutputFormat == dm_420 || (InterlaceEnable && ProgressiveToInterlaceUnitInOPP)) + *DSTYAfterScaler = 1; + else + *DSTYAfterScaler = 0; + + DSTTotalPixelsAfterScaler = ((double) (*DSTYAfterScaler * HTotal)) + *DSTXAfterScaler; + *DSTYAfterScaler = dml_floor(DSTTotalPixelsAfterScaler / HTotal, 1); + *DSTXAfterScaler = DSTTotalPixelsAfterScaler - ((double) (*DSTYAfterScaler * HTotal)); + + *VUpdateOffsetPix = dml_ceil(HTotal / 4.0, 1); + TotalRepeaterDelayTime = MaxInterDCNTileRepeaters * (2.0 / DPPCLK + 3.0 / DISPCLK); + *VUpdateWidthPix = (14.0 / DCFClkDeepSleep + 12.0 / DPPCLK + TotalRepeaterDelayTime) + * PixelClock; + + *VReadyOffsetPix = dml_max( + 150.0 / DPPCLK, + TotalRepeaterDelayTime + 20.0 / DCFClkDeepSleep + 10.0 / DPPCLK) + * PixelClock; + + Tsetup = (double) (*VUpdateOffsetPix + *VUpdateWidthPix + *VReadyOffsetPix) / PixelClock; + + LineTime = (double) HTotal / PixelClock; + + if (DynamicMetadataEnable) { + double Tdmbf, Tdmec, Tdmsks; + + Tdm = dml_max(0.0, UrgentExtraLatency - TCalc); + Tdmbf = DynamicMetadataTransmittedBytes / 4.0 / DISPCLK; + Tdmec = LineTime; + if (DynamicMetadataLinesBeforeActiveRequired == 0) + Tdmsks = VBlank * LineTime / 2.0; + else + Tdmsks = DynamicMetadataLinesBeforeActiveRequired * LineTime; + if (InterlaceEnable && !ProgressiveToInterlaceUnitInOPP) + Tdmsks = Tdmsks / 2; + if (VStartup * LineTime + < Tsetup + TWait + UrgentExtraLatency + Tdmbf + Tdmec + Tdmsks) { + MyError = true; + *VStartupRequiredWhenNotEnoughTimeForDynamicMetadata = (Tsetup + TWait + + UrgentExtraLatency + Tdmbf + Tdmec + Tdmsks) / LineTime; + } else + *VStartupRequiredWhenNotEnoughTimeForDynamicMetadata = 0.0; + } else + Tdm = 0; + + if (VirtualMemoryEnable) { + if (PageTableLevels == 4) + *Tno_bw = UrgentExtraLatency + UrgentLatency; + else if (PageTableLevels == 3) + *Tno_bw = UrgentExtraLatency; + else + *Tno_bw = 0; + } else if (DCCEnable) + *Tno_bw = LineTime; + else + *Tno_bw = LineTime / 4; + + dst_y_prefetch_equ = VStartup - dml_max(TCalc + TWait, XFCRemoteSurfaceFlipDelay) / LineTime + - (Tsetup + Tdm) / LineTime + - (*DSTYAfterScaler + *DSTXAfterScaler / HTotal); + + Tsw_oto = dml_max(PrefetchSourceLinesY, PrefetchSourceLinesC) * LineTime; + + prefetch_bw_oto = (MetaRowByte + PixelPTEBytesPerRow + + PrefetchSourceLinesY * SwathWidthY * dml_ceil(BytePerPixelDETY, 1) + + PrefetchSourceLinesC * SwathWidthY / 2 * dml_ceil(BytePerPixelDETC, 2)) + / Tsw_oto; + + if (VirtualMemoryEnable == true) { + Tvm_oto = + dml_max( + *Tno_bw + PDEAndMetaPTEBytesFrame / prefetch_bw_oto, + dml_max( + UrgentExtraLatency + + UrgentLatency + * (PageTableLevels + - 1), + LineTime / 4.0)); + } else + Tvm_oto = LineTime / 4.0; + + if ((VirtualMemoryEnable == true || DCCEnable == true)) { + Tr0_oto = dml_max( + (MetaRowByte + PixelPTEBytesPerRow) / prefetch_bw_oto, + dml_max(UrgentLatency, dml_max(LineTime - Tvm_oto, LineTime / 4))); + } else + Tr0_oto = LineTime - Tvm_oto; + + Tpre_oto = Tvm_oto + Tr0_oto + Tsw_oto; + + dst_y_prefetch_oto = Tpre_oto / LineTime; + + if (dst_y_prefetch_oto < dst_y_prefetch_equ) + *DestinationLinesForPrefetch = dst_y_prefetch_oto; + else + *DestinationLinesForPrefetch = dst_y_prefetch_equ; + + *DestinationLinesForPrefetch = dml_floor(4.0 * (*DestinationLinesForPrefetch + 0.125), 1) + / 4; + + dml_print("DML: VStartup: %d\n", VStartup); + dml_print("DML: TCalc: %f\n", TCalc); + dml_print("DML: TWait: %f\n", TWait); + dml_print("DML: XFCRemoteSurfaceFlipDelay: %f\n", XFCRemoteSurfaceFlipDelay); + dml_print("DML: LineTime: %f\n", LineTime); + dml_print("DML: Tsetup: %f\n", Tsetup); + dml_print("DML: Tdm: %f\n", Tdm); + dml_print("DML: DSTYAfterScaler: %f\n", *DSTYAfterScaler); + dml_print("DML: DSTXAfterScaler: %f\n", *DSTXAfterScaler); + dml_print("DML: HTotal: %d\n", HTotal); + + *PrefetchBandwidth = 0; + *DestinationLinesToRequestVMInVBlank = 0; + *DestinationLinesToRequestRowInVBlank = 0; + *VRatioPrefetchY = 0; + *VRatioPrefetchC = 0; + *RequiredPrefetchPixDataBW = 0; + if (*DestinationLinesForPrefetch > 1) { + *PrefetchBandwidth = (PDEAndMetaPTEBytesFrame + 2 * MetaRowByte + + 2 * PixelPTEBytesPerRow + + PrefetchSourceLinesY * SwathWidthY * dml_ceil(BytePerPixelDETY, 1) + + PrefetchSourceLinesC * SwathWidthY / 2 + * dml_ceil(BytePerPixelDETC, 2)) + / (*DestinationLinesForPrefetch * LineTime - *Tno_bw); + if (VirtualMemoryEnable) { + TimeForFetchingMetaPTE = + dml_max( + *Tno_bw + + (double) PDEAndMetaPTEBytesFrame + / *PrefetchBandwidth, + dml_max( + UrgentExtraLatency + + UrgentLatency + * (PageTableLevels + - 1), + LineTime / 4)); + } else { + if (NumberOfCursors > 0 || XFCEnabled) + TimeForFetchingMetaPTE = LineTime / 4; + else + TimeForFetchingMetaPTE = 0.0; + } + + if ((VirtualMemoryEnable == true || DCCEnable == true)) { + TimeForFetchingRowInVBlank = + dml_max( + (MetaRowByte + PixelPTEBytesPerRow) + / *PrefetchBandwidth, + dml_max( + UrgentLatency, + dml_max( + LineTime + - TimeForFetchingMetaPTE, + LineTime + / 4.0))); + } else { + if (NumberOfCursors > 0 || XFCEnabled) + TimeForFetchingRowInVBlank = LineTime - TimeForFetchingMetaPTE; + else + TimeForFetchingRowInVBlank = 0.0; + } + + *DestinationLinesToRequestVMInVBlank = dml_floor( + 4.0 * (TimeForFetchingMetaPTE / LineTime + 0.125), + 1) / 4.0; + + *DestinationLinesToRequestRowInVBlank = dml_floor( + 4.0 * (TimeForFetchingRowInVBlank / LineTime + 0.125), + 1) / 4.0; + + LinesToRequestPrefetchPixelData = + *DestinationLinesForPrefetch + - ((NumberOfCursors > 0 || VirtualMemoryEnable + || DCCEnable) ? + (*DestinationLinesToRequestVMInVBlank + + *DestinationLinesToRequestRowInVBlank) : + 0.0); + + if (LinesToRequestPrefetchPixelData > 0) { + + *VRatioPrefetchY = (double) PrefetchSourceLinesY + / LinesToRequestPrefetchPixelData; + *VRatioPrefetchY = dml_max(*VRatioPrefetchY, 1.0); + if ((SwathHeightY > 4) && (VInitPreFillY > 3)) { + if (LinesToRequestPrefetchPixelData > (VInitPreFillY - 3.0) / 2.0) { + *VRatioPrefetchY = + dml_max( + (double) PrefetchSourceLinesY + / LinesToRequestPrefetchPixelData, + (double) MaxNumSwathY + * SwathHeightY + / (LinesToRequestPrefetchPixelData + - (VInitPreFillY + - 3.0) + / 2.0)); + *VRatioPrefetchY = dml_max(*VRatioPrefetchY, 1.0); + } else { + MyError = true; + *VRatioPrefetchY = 0; + } + } + + *VRatioPrefetchC = (double) PrefetchSourceLinesC + / LinesToRequestPrefetchPixelData; + *VRatioPrefetchC = dml_max(*VRatioPrefetchC, 1.0); + + if ((SwathHeightC > 4)) { + if (LinesToRequestPrefetchPixelData > (VInitPreFillC - 3.0) / 2.0) { + *VRatioPrefetchC = + dml_max( + *VRatioPrefetchC, + (double) MaxNumSwathC + * SwathHeightC + / (LinesToRequestPrefetchPixelData + - (VInitPreFillC + - 3.0) + / 2.0)); + *VRatioPrefetchC = dml_max(*VRatioPrefetchC, 1.0); + } else { + MyError = true; + *VRatioPrefetchC = 0; + } + } + + *RequiredPrefetchPixDataBW = + DPPPerPlane + * ((double) PrefetchSourceLinesY + / LinesToRequestPrefetchPixelData + * dml_ceil( + BytePerPixelDETY, + 1) + + (double) PrefetchSourceLinesC + / LinesToRequestPrefetchPixelData + * dml_ceil( + BytePerPixelDETC, + 2) + / 2) + * SwathWidthY / LineTime; + } else { + MyError = true; + *VRatioPrefetchY = 0; + *VRatioPrefetchC = 0; + *RequiredPrefetchPixDataBW = 0; + } + + } else { + MyError = true; + } + + if (MyError) { + *PrefetchBandwidth = 0; + TimeForFetchingMetaPTE = 0; + TimeForFetchingRowInVBlank = 0; + *DestinationLinesToRequestVMInVBlank = 0; + *DestinationLinesToRequestRowInVBlank = 0; + *DestinationLinesForPrefetch = 0; + LinesToRequestPrefetchPixelData = 0; + *VRatioPrefetchY = 0; + *VRatioPrefetchC = 0; + *RequiredPrefetchPixDataBW = 0; + } + + return MyError; +} + +static double RoundToDFSGranularityUp(double Clock, double VCOSpeed) +{ + return VCOSpeed * 4 / dml_floor(VCOSpeed * 4 / Clock, 1); +} + +static double RoundToDFSGranularityDown(double Clock, double VCOSpeed) +{ + return VCOSpeed * 4 / dml_ceil(VCOSpeed * 4 / Clock, 1); +} + +static double CalculatePrefetchSourceLines( + struct display_mode_lib *mode_lib, + double VRatio, + double vtaps, + bool Interlace, + bool ProgressiveToInterlaceUnitInOPP, + unsigned int SwathHeight, + unsigned int ViewportYStart, + double *VInitPreFill, + unsigned int *MaxNumSwath) +{ + unsigned int MaxPartialSwath; + + if (ProgressiveToInterlaceUnitInOPP) + *VInitPreFill = dml_floor((VRatio + vtaps + 1) / 2.0, 1); + else + *VInitPreFill = dml_floor((VRatio + vtaps + 1 + Interlace * 0.5 * VRatio) / 2.0, 1); + + if (!mode_lib->vba.IgnoreViewportPositioning) { + + *MaxNumSwath = dml_ceil((*VInitPreFill - 1.0) / SwathHeight, 1) + 1.0; + + if (*VInitPreFill > 1.0) + MaxPartialSwath = (unsigned int) (*VInitPreFill - 2) % SwathHeight; + else + MaxPartialSwath = (unsigned int) (*VInitPreFill + SwathHeight - 2) + % SwathHeight; + MaxPartialSwath = dml_max(1U, MaxPartialSwath); + + } else { + + if (ViewportYStart != 0) + dml_print( + "WARNING DML: using viewport y position of 0 even though actual viewport y position is non-zero in prefetch source lines calculation\n"); + + *MaxNumSwath = dml_ceil(*VInitPreFill / SwathHeight, 1); + + if (*VInitPreFill > 1.0) + MaxPartialSwath = (unsigned int) (*VInitPreFill - 1) % SwathHeight; + else + MaxPartialSwath = (unsigned int) (*VInitPreFill + SwathHeight - 1) + % SwathHeight; + } + + return *MaxNumSwath * SwathHeight + MaxPartialSwath; +} + +static unsigned int CalculateVMAndRowBytes( + struct display_mode_lib *mode_lib, + bool DCCEnable, + unsigned int BlockHeight256Bytes, + unsigned int BlockWidth256Bytes, + enum source_format_class SourcePixelFormat, + unsigned int SurfaceTiling, + unsigned int BytePerPixel, + enum scan_direction_class ScanDirection, + unsigned int ViewportWidth, + unsigned int ViewportHeight, + unsigned int SwathWidth, + bool VirtualMemoryEnable, + unsigned int VMMPageSize, + unsigned int PTEBufferSizeInRequests, + unsigned int PDEProcessingBufIn64KBReqs, + unsigned int Pitch, + unsigned int DCCMetaPitch, + unsigned int *MacroTileWidth, + unsigned int *MetaRowByte, + unsigned int *PixelPTEBytesPerRow, + bool *PTEBufferSizeNotExceeded, + unsigned int *dpte_row_height, + unsigned int *meta_row_height) +{ + unsigned int MetaRequestHeight; + unsigned int MetaRequestWidth; + unsigned int MetaSurfWidth; + unsigned int MetaSurfHeight; + unsigned int MPDEBytesFrame; + unsigned int MetaPTEBytesFrame; + unsigned int DCCMetaSurfaceBytes; + + unsigned int MacroTileSizeBytes; + unsigned int MacroTileHeight; + unsigned int DPDE0BytesFrame; + unsigned int ExtraDPDEBytesFrame; + unsigned int PDEAndMetaPTEBytesFrame; + + if (DCCEnable == true) { + MetaRequestHeight = 8 * BlockHeight256Bytes; + MetaRequestWidth = 8 * BlockWidth256Bytes; + if (ScanDirection == dm_horz) { + *meta_row_height = MetaRequestHeight; + MetaSurfWidth = dml_ceil((double) SwathWidth - 1, MetaRequestWidth) + + MetaRequestWidth; + *MetaRowByte = MetaSurfWidth * MetaRequestHeight * BytePerPixel / 256.0; + } else { + *meta_row_height = MetaRequestWidth; + MetaSurfHeight = dml_ceil((double) SwathWidth - 1, MetaRequestHeight) + + MetaRequestHeight; + *MetaRowByte = MetaSurfHeight * MetaRequestWidth * BytePerPixel / 256.0; + } + if (ScanDirection == dm_horz) { + DCCMetaSurfaceBytes = DCCMetaPitch + * (dml_ceil(ViewportHeight - 1, 64 * BlockHeight256Bytes) + + 64 * BlockHeight256Bytes) * BytePerPixel + / 256; + } else { + DCCMetaSurfaceBytes = DCCMetaPitch + * (dml_ceil( + (double) ViewportHeight - 1, + 64 * BlockHeight256Bytes) + + 64 * BlockHeight256Bytes) * BytePerPixel + / 256; + } + if (VirtualMemoryEnable == true) { + MetaPTEBytesFrame = (dml_ceil( + (double) (DCCMetaSurfaceBytes - VMMPageSize) + / (8 * VMMPageSize), + 1) + 1) * 64; + MPDEBytesFrame = 128 * (mode_lib->vba.MaxPageTableLevels - 1); + } else { + MetaPTEBytesFrame = 0; + MPDEBytesFrame = 0; + } + } else { + MetaPTEBytesFrame = 0; + MPDEBytesFrame = 0; + *MetaRowByte = 0; + } + + if (SurfaceTiling == dm_sw_linear) { + MacroTileSizeBytes = 256; + MacroTileHeight = 1; + } else if (SurfaceTiling == dm_sw_4kb_s || SurfaceTiling == dm_sw_4kb_s_x + || SurfaceTiling == dm_sw_4kb_d || SurfaceTiling == dm_sw_4kb_d_x) { + MacroTileSizeBytes = 4096; + MacroTileHeight = 4 * BlockHeight256Bytes; + } else if (SurfaceTiling == dm_sw_64kb_s || SurfaceTiling == dm_sw_64kb_s_t + || SurfaceTiling == dm_sw_64kb_s_x || SurfaceTiling == dm_sw_64kb_d + || SurfaceTiling == dm_sw_64kb_d_t || SurfaceTiling == dm_sw_64kb_d_x + || SurfaceTiling == dm_sw_64kb_r_x) { + MacroTileSizeBytes = 65536; + MacroTileHeight = 16 * BlockHeight256Bytes; + } else { + MacroTileSizeBytes = 262144; + MacroTileHeight = 32 * BlockHeight256Bytes; + } + *MacroTileWidth = MacroTileSizeBytes / BytePerPixel / MacroTileHeight; + + if (VirtualMemoryEnable == true && mode_lib->vba.MaxPageTableLevels > 1) { + if (ScanDirection == dm_horz) { + DPDE0BytesFrame = + 64 + * (dml_ceil( + ((Pitch + * (dml_ceil( + ViewportHeight + - 1, + MacroTileHeight) + + MacroTileHeight) + * BytePerPixel) + - MacroTileSizeBytes) + / (8 + * 2097152), + 1) + 1); + } else { + DPDE0BytesFrame = + 64 + * (dml_ceil( + ((Pitch + * (dml_ceil( + (double) SwathWidth + - 1, + MacroTileHeight) + + MacroTileHeight) + * BytePerPixel) + - MacroTileSizeBytes) + / (8 + * 2097152), + 1) + 1); + } + ExtraDPDEBytesFrame = 128 * (mode_lib->vba.MaxPageTableLevels - 2); + } else { + DPDE0BytesFrame = 0; + ExtraDPDEBytesFrame = 0; + } + + PDEAndMetaPTEBytesFrame = MetaPTEBytesFrame + MPDEBytesFrame + DPDE0BytesFrame + + ExtraDPDEBytesFrame; + + if (VirtualMemoryEnable == true) { + unsigned int PTERequestSize; + unsigned int PixelPTEReqHeight; + unsigned int PixelPTEReqWidth; + double FractionOfPTEReturnDrop; + unsigned int EffectivePDEProcessingBufIn64KBReqs; + + if (SurfaceTiling == dm_sw_linear) { + PixelPTEReqHeight = 1; + PixelPTEReqWidth = 8.0 * VMMPageSize / BytePerPixel; + PTERequestSize = 64; + FractionOfPTEReturnDrop = 0; + } else if (MacroTileSizeBytes == 4096) { + PixelPTEReqHeight = MacroTileHeight; + PixelPTEReqWidth = 8 * *MacroTileWidth; + PTERequestSize = 64; + if (ScanDirection == dm_horz) + FractionOfPTEReturnDrop = 0; + else + FractionOfPTEReturnDrop = 7 / 8; + } else if (VMMPageSize == 4096 && MacroTileSizeBytes > 4096) { + PixelPTEReqHeight = 16 * BlockHeight256Bytes; + PixelPTEReqWidth = 16 * BlockWidth256Bytes; + PTERequestSize = 128; + FractionOfPTEReturnDrop = 0; + } else { + PixelPTEReqHeight = MacroTileHeight; + PixelPTEReqWidth = 8 * *MacroTileWidth; + PTERequestSize = 64; + FractionOfPTEReturnDrop = 0; + } + + if (SourcePixelFormat == dm_420_8 || SourcePixelFormat == dm_420_10) + EffectivePDEProcessingBufIn64KBReqs = PDEProcessingBufIn64KBReqs / 2; + else + EffectivePDEProcessingBufIn64KBReqs = PDEProcessingBufIn64KBReqs; + + if (SurfaceTiling == dm_sw_linear) { + *dpte_row_height = + dml_min( + 128, + 1 + << (unsigned int) dml_floor( + dml_log2( + dml_min( + (double) PTEBufferSizeInRequests + * PixelPTEReqWidth, + EffectivePDEProcessingBufIn64KBReqs + * 65536.0 + / BytePerPixel) + / Pitch), + 1)); + *PixelPTEBytesPerRow = PTERequestSize + * (dml_ceil( + (double) (Pitch * *dpte_row_height - 1) + / PixelPTEReqWidth, + 1) + 1); + } else if (ScanDirection == dm_horz) { + *dpte_row_height = PixelPTEReqHeight; + *PixelPTEBytesPerRow = PTERequestSize + * (dml_ceil(((double) SwathWidth - 1) / PixelPTEReqWidth, 1) + + 1); + } else { + *dpte_row_height = dml_min(PixelPTEReqWidth, *MacroTileWidth); + *PixelPTEBytesPerRow = PTERequestSize + * (dml_ceil( + ((double) SwathWidth - 1) + / PixelPTEReqHeight, + 1) + 1); + } + if (*PixelPTEBytesPerRow * (1 - FractionOfPTEReturnDrop) + <= 64 * PTEBufferSizeInRequests) { + *PTEBufferSizeNotExceeded = true; + } else { + *PTEBufferSizeNotExceeded = false; + } + } else { + *PixelPTEBytesPerRow = 0; + *PTEBufferSizeNotExceeded = true; + } + + return PDEAndMetaPTEBytesFrame; +} + +static void DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation( + struct display_mode_lib *mode_lib) +{ + unsigned int j, k; + + mode_lib->vba.WritebackDISPCLK = 0.0; + mode_lib->vba.DISPCLKWithRamping = 0; + mode_lib->vba.DISPCLKWithoutRamping = 0; + mode_lib->vba.GlobalDPPCLK = 0.0; + + // dml_ml->vba.DISPCLK and dml_ml->vba.DPPCLK Calculation + // + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + if (mode_lib->vba.WritebackEnable[k]) { + mode_lib->vba.WritebackDISPCLK = + dml_max( + mode_lib->vba.WritebackDISPCLK, + CalculateWriteBackDISPCLK( + mode_lib->vba.WritebackPixelFormat[k], + mode_lib->vba.PixelClock[k], + mode_lib->vba.WritebackHRatio[k], + mode_lib->vba.WritebackVRatio[k], + mode_lib->vba.WritebackLumaHTaps[k], + mode_lib->vba.WritebackLumaVTaps[k], + mode_lib->vba.WritebackChromaHTaps[k], + mode_lib->vba.WritebackChromaVTaps[k], + mode_lib->vba.WritebackDestinationWidth[k], + mode_lib->vba.HTotal[k], + mode_lib->vba.WritebackChromaLineBufferWidth)); + } + } + + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + if (mode_lib->vba.HRatio[k] > 1) { + mode_lib->vba.PSCL_THROUGHPUT_LUMA[k] = dml_min( + mode_lib->vba.MaxDCHUBToPSCLThroughput, + mode_lib->vba.MaxPSCLToLBThroughput + * mode_lib->vba.HRatio[k] + / dml_ceil( + mode_lib->vba.htaps[k] + / 6.0, + 1)); + } else { + mode_lib->vba.PSCL_THROUGHPUT_LUMA[k] = dml_min( + mode_lib->vba.MaxDCHUBToPSCLThroughput, + mode_lib->vba.MaxPSCLToLBThroughput); + } + + mode_lib->vba.DPPCLKUsingSingleDPPLuma = + mode_lib->vba.PixelClock[k] + * dml_max( + mode_lib->vba.vtaps[k] / 6.0 + * dml_min( + 1.0, + mode_lib->vba.HRatio[k]), + dml_max( + mode_lib->vba.HRatio[k] + * mode_lib->vba.VRatio[k] + / mode_lib->vba.PSCL_THROUGHPUT_LUMA[k], + 1.0)); + + if ((mode_lib->vba.htaps[k] > 6 || mode_lib->vba.vtaps[k] > 6) + && mode_lib->vba.DPPCLKUsingSingleDPPLuma + < 2 * mode_lib->vba.PixelClock[k]) { + mode_lib->vba.DPPCLKUsingSingleDPPLuma = 2 * mode_lib->vba.PixelClock[k]; + } + + if ((mode_lib->vba.SourcePixelFormat[k] != dm_420_8 + && mode_lib->vba.SourcePixelFormat[k] != dm_420_10)) { + mode_lib->vba.PSCL_THROUGHPUT_CHROMA[k] = 0.0; + mode_lib->vba.DPPCLKUsingSingleDPP[k] = + mode_lib->vba.DPPCLKUsingSingleDPPLuma; + } else { + if (mode_lib->vba.HRatio[k] > 1) { + mode_lib->vba.PSCL_THROUGHPUT_CHROMA[k] = + dml_min( + mode_lib->vba.MaxDCHUBToPSCLThroughput, + mode_lib->vba.MaxPSCLToLBThroughput + * mode_lib->vba.HRatio[k] + / 2 + / dml_ceil( + mode_lib->vba.HTAPsChroma[k] + / 6.0, + 1.0)); + } else { + mode_lib->vba.PSCL_THROUGHPUT_CHROMA[k] = dml_min( + mode_lib->vba.MaxDCHUBToPSCLThroughput, + mode_lib->vba.MaxPSCLToLBThroughput); + } + mode_lib->vba.DPPCLKUsingSingleDPPChroma = + mode_lib->vba.PixelClock[k] + * dml_max( + mode_lib->vba.VTAPsChroma[k] + / 6.0 + * dml_min( + 1.0, + mode_lib->vba.HRatio[k] + / 2), + dml_max( + mode_lib->vba.HRatio[k] + * mode_lib->vba.VRatio[k] + / 4 + / mode_lib->vba.PSCL_THROUGHPUT_CHROMA[k], + 1.0)); + + if ((mode_lib->vba.HTAPsChroma[k] > 6 || mode_lib->vba.VTAPsChroma[k] > 6) + && mode_lib->vba.DPPCLKUsingSingleDPPChroma + < 2 * mode_lib->vba.PixelClock[k]) { + mode_lib->vba.DPPCLKUsingSingleDPPChroma = 2 + * mode_lib->vba.PixelClock[k]; + } + + mode_lib->vba.DPPCLKUsingSingleDPP[k] = dml_max( + mode_lib->vba.DPPCLKUsingSingleDPPLuma, + mode_lib->vba.DPPCLKUsingSingleDPPChroma); + } + } + + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + if (mode_lib->vba.BlendingAndTiming[k] != k) + continue; + if (mode_lib->vba.ODMCombineEnabled[k]) { + mode_lib->vba.DISPCLKWithRamping = + dml_max( + mode_lib->vba.DISPCLKWithRamping, + mode_lib->vba.PixelClock[k] / 2 + * (1 + + mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading + / 100) + * (1 + + mode_lib->vba.DISPCLKRampingMargin + / 100)); + mode_lib->vba.DISPCLKWithoutRamping = + dml_max( + mode_lib->vba.DISPCLKWithoutRamping, + mode_lib->vba.PixelClock[k] / 2 + * (1 + + mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading + / 100)); + } else if (!mode_lib->vba.ODMCombineEnabled[k]) { + mode_lib->vba.DISPCLKWithRamping = + dml_max( + mode_lib->vba.DISPCLKWithRamping, + mode_lib->vba.PixelClock[k] + * (1 + + mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading + / 100) + * (1 + + mode_lib->vba.DISPCLKRampingMargin + / 100)); + mode_lib->vba.DISPCLKWithoutRamping = + dml_max( + mode_lib->vba.DISPCLKWithoutRamping, + mode_lib->vba.PixelClock[k] + * (1 + + mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading + / 100)); + } + } + + mode_lib->vba.DISPCLKWithRamping = dml_max( + mode_lib->vba.DISPCLKWithRamping, + mode_lib->vba.WritebackDISPCLK); + mode_lib->vba.DISPCLKWithoutRamping = dml_max( + mode_lib->vba.DISPCLKWithoutRamping, + mode_lib->vba.WritebackDISPCLK); + + ASSERT(mode_lib->vba.DISPCLKDPPCLKVCOSpeed != 0); + mode_lib->vba.DISPCLKWithRampingRoundedToDFSGranularity = RoundToDFSGranularityUp( + mode_lib->vba.DISPCLKWithRamping, + mode_lib->vba.DISPCLKDPPCLKVCOSpeed); + mode_lib->vba.DISPCLKWithoutRampingRoundedToDFSGranularity = RoundToDFSGranularityUp( + mode_lib->vba.DISPCLKWithoutRamping, + mode_lib->vba.DISPCLKDPPCLKVCOSpeed); + mode_lib->vba.MaxDispclkRoundedToDFSGranularity = RoundToDFSGranularityDown( + mode_lib->vba.soc.clock_limits[NumberOfStates - 1].dispclk_mhz, + mode_lib->vba.DISPCLKDPPCLKVCOSpeed); + if (mode_lib->vba.DISPCLKWithoutRampingRoundedToDFSGranularity + > mode_lib->vba.MaxDispclkRoundedToDFSGranularity) { + mode_lib->vba.DISPCLK_calculated = + mode_lib->vba.DISPCLKWithoutRampingRoundedToDFSGranularity; + } else if (mode_lib->vba.DISPCLKWithRampingRoundedToDFSGranularity + > mode_lib->vba.MaxDispclkRoundedToDFSGranularity) { + mode_lib->vba.DISPCLK_calculated = mode_lib->vba.MaxDispclkRoundedToDFSGranularity; + } else { + mode_lib->vba.DISPCLK_calculated = + mode_lib->vba.DISPCLKWithRampingRoundedToDFSGranularity; + } + DTRACE(" dispclk_mhz (calculated) = %f", mode_lib->vba.DISPCLK_calculated); + + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + mode_lib->vba.DPPCLK_calculated[k] = mode_lib->vba.DPPCLKUsingSingleDPP[k] + / mode_lib->vba.DPPPerPlane[k] + * (1 + mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading / 100); + mode_lib->vba.GlobalDPPCLK = dml_max( + mode_lib->vba.GlobalDPPCLK, + mode_lib->vba.DPPCLK_calculated[k]); + } + mode_lib->vba.GlobalDPPCLK = RoundToDFSGranularityUp( + mode_lib->vba.GlobalDPPCLK, + mode_lib->vba.DISPCLKDPPCLKVCOSpeed); + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + mode_lib->vba.DPPCLK_calculated[k] = mode_lib->vba.GlobalDPPCLK / 255 + * dml_ceil( + mode_lib->vba.DPPCLK_calculated[k] * 255 + / mode_lib->vba.GlobalDPPCLK, + 1); + DTRACE(" dppclk_mhz[%i] (calculated) = %f", k, mode_lib->vba.DPPCLK_calculated[k]); + } + + // Urgent Watermark + mode_lib->vba.DCCEnabledAnyPlane = false; + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) + if (mode_lib->vba.DCCEnable[k]) + mode_lib->vba.DCCEnabledAnyPlane = true; + + mode_lib->vba.ReturnBandwidthToDCN = dml_min( + mode_lib->vba.ReturnBusWidth * mode_lib->vba.DCFCLK, + mode_lib->vba.FabricAndDRAMBandwidth * 1000) + * mode_lib->vba.PercentOfIdealDRAMAndFabricBWReceivedAfterUrgLatency / 100; + + mode_lib->vba.ReturnBW = mode_lib->vba.ReturnBandwidthToDCN; + mode_lib->vba.ReturnBW = adjust_ReturnBW( + mode_lib, + mode_lib->vba.ReturnBW, + mode_lib->vba.DCCEnabledAnyPlane, + mode_lib->vba.ReturnBandwidthToDCN); + + // Let's do this calculation again?? + mode_lib->vba.ReturnBandwidthToDCN = dml_min( + mode_lib->vba.ReturnBusWidth * mode_lib->vba.DCFCLK, + mode_lib->vba.FabricAndDRAMBandwidth * 1000); + mode_lib->vba.ReturnBW = adjust_ReturnBW( + mode_lib, + mode_lib->vba.ReturnBW, + mode_lib->vba.DCCEnabledAnyPlane, + mode_lib->vba.ReturnBandwidthToDCN); + + DTRACE(" dcfclk_mhz = %f", mode_lib->vba.DCFCLK); + DTRACE(" return_bw_to_dcn = %f", mode_lib->vba.ReturnBandwidthToDCN); + DTRACE(" return_bus_bw = %f", mode_lib->vba.ReturnBW); + + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + bool MainPlaneDoesODMCombine = false; + + if (mode_lib->vba.SourceScan[k] == dm_horz) + mode_lib->vba.SwathWidthSingleDPPY[k] = mode_lib->vba.ViewportWidth[k]; + else + mode_lib->vba.SwathWidthSingleDPPY[k] = mode_lib->vba.ViewportHeight[k]; + + if (mode_lib->vba.ODMCombineEnabled[k] == true) + MainPlaneDoesODMCombine = true; + for (j = 0; j < mode_lib->vba.NumberOfActivePlanes; ++j) + if (mode_lib->vba.BlendingAndTiming[k] == j + && mode_lib->vba.ODMCombineEnabled[j] == true) + MainPlaneDoesODMCombine = true; + + if (MainPlaneDoesODMCombine == true) + mode_lib->vba.SwathWidthY[k] = dml_min( + (double) mode_lib->vba.SwathWidthSingleDPPY[k], + dml_round( + mode_lib->vba.HActive[k] / 2.0 + * mode_lib->vba.HRatio[k])); + else + mode_lib->vba.SwathWidthY[k] = mode_lib->vba.SwathWidthSingleDPPY[k] + / mode_lib->vba.DPPPerPlane[k]; + } + + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + if (mode_lib->vba.SourcePixelFormat[k] == dm_444_64) { + mode_lib->vba.BytePerPixelDETY[k] = 8; + mode_lib->vba.BytePerPixelDETC[k] = 0; + } else if (mode_lib->vba.SourcePixelFormat[k] == dm_444_32) { + mode_lib->vba.BytePerPixelDETY[k] = 4; + mode_lib->vba.BytePerPixelDETC[k] = 0; + } else if (mode_lib->vba.SourcePixelFormat[k] == dm_444_16) { + mode_lib->vba.BytePerPixelDETY[k] = 2; + mode_lib->vba.BytePerPixelDETC[k] = 0; + } else if (mode_lib->vba.SourcePixelFormat[k] == dm_444_8) { + mode_lib->vba.BytePerPixelDETY[k] = 1; + mode_lib->vba.BytePerPixelDETC[k] = 0; + } else if (mode_lib->vba.SourcePixelFormat[k] == dm_420_8) { + mode_lib->vba.BytePerPixelDETY[k] = 1; + mode_lib->vba.BytePerPixelDETC[k] = 2; + } else { // dm_420_10 + mode_lib->vba.BytePerPixelDETY[k] = 4.0 / 3.0; + mode_lib->vba.BytePerPixelDETC[k] = 8.0 / 3.0; + } + } + + mode_lib->vba.TotalDataReadBandwidth = 0.0; + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + mode_lib->vba.ReadBandwidthPlaneLuma[k] = mode_lib->vba.SwathWidthSingleDPPY[k] + * dml_ceil(mode_lib->vba.BytePerPixelDETY[k], 1) + / (mode_lib->vba.HTotal[k] / mode_lib->vba.PixelClock[k]) + * mode_lib->vba.VRatio[k]; + mode_lib->vba.ReadBandwidthPlaneChroma[k] = mode_lib->vba.SwathWidthSingleDPPY[k] + / 2 * dml_ceil(mode_lib->vba.BytePerPixelDETC[k], 2) + / (mode_lib->vba.HTotal[k] / mode_lib->vba.PixelClock[k]) + * mode_lib->vba.VRatio[k] / 2; + DTRACE( + " read_bw[%i] = %fBps", + k, + mode_lib->vba.ReadBandwidthPlaneLuma[k] + + mode_lib->vba.ReadBandwidthPlaneChroma[k]); + mode_lib->vba.TotalDataReadBandwidth += mode_lib->vba.ReadBandwidthPlaneLuma[k] + + mode_lib->vba.ReadBandwidthPlaneChroma[k]; + } + + mode_lib->vba.TotalDCCActiveDPP = 0; + mode_lib->vba.TotalActiveDPP = 0; + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + mode_lib->vba.TotalActiveDPP = mode_lib->vba.TotalActiveDPP + + mode_lib->vba.DPPPerPlane[k]; + if (mode_lib->vba.DCCEnable[k]) + mode_lib->vba.TotalDCCActiveDPP = mode_lib->vba.TotalDCCActiveDPP + + mode_lib->vba.DPPPerPlane[k]; + } + + mode_lib->vba.UrgentRoundTripAndOutOfOrderLatency = + (mode_lib->vba.RoundTripPingLatencyCycles + 32) / mode_lib->vba.DCFCLK + + mode_lib->vba.UrgentOutOfOrderReturnPerChannel + * mode_lib->vba.NumberOfChannels + / mode_lib->vba.ReturnBW; + + mode_lib->vba.LastPixelOfLineExtraWatermark = 0; + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + double DataFabricLineDeliveryTimeLuma, DataFabricLineDeliveryTimeChroma; + + if (mode_lib->vba.VRatio[k] <= 1.0) + mode_lib->vba.DisplayPipeLineDeliveryTimeLuma[k] = + (double) mode_lib->vba.SwathWidthY[k] + * mode_lib->vba.DPPPerPlane[k] + / mode_lib->vba.HRatio[k] + / mode_lib->vba.PixelClock[k]; + else + mode_lib->vba.DisplayPipeLineDeliveryTimeLuma[k] = + (double) mode_lib->vba.SwathWidthY[k] + / mode_lib->vba.PSCL_THROUGHPUT_LUMA[k] + / mode_lib->vba.DPPCLK[k]; + + DataFabricLineDeliveryTimeLuma = mode_lib->vba.SwathWidthSingleDPPY[k] + * mode_lib->vba.SwathHeightY[k] + * dml_ceil(mode_lib->vba.BytePerPixelDETY[k], 1) + / (mode_lib->vba.ReturnBW * mode_lib->vba.ReadBandwidthPlaneLuma[k] + / mode_lib->vba.TotalDataReadBandwidth); + mode_lib->vba.LastPixelOfLineExtraWatermark = dml_max( + mode_lib->vba.LastPixelOfLineExtraWatermark, + DataFabricLineDeliveryTimeLuma + - mode_lib->vba.DisplayPipeLineDeliveryTimeLuma[k]); + + if (mode_lib->vba.BytePerPixelDETC[k] == 0) + mode_lib->vba.DisplayPipeLineDeliveryTimeChroma[k] = 0.0; + else if (mode_lib->vba.VRatio[k] / 2.0 <= 1.0) + mode_lib->vba.DisplayPipeLineDeliveryTimeChroma[k] = + mode_lib->vba.SwathWidthY[k] / 2.0 + * mode_lib->vba.DPPPerPlane[k] + / (mode_lib->vba.HRatio[k] / 2.0) + / mode_lib->vba.PixelClock[k]; + else + mode_lib->vba.DisplayPipeLineDeliveryTimeChroma[k] = + mode_lib->vba.SwathWidthY[k] / 2.0 + / mode_lib->vba.PSCL_THROUGHPUT_CHROMA[k] + / mode_lib->vba.DPPCLK[k]; + + DataFabricLineDeliveryTimeChroma = mode_lib->vba.SwathWidthSingleDPPY[k] / 2.0 + * mode_lib->vba.SwathHeightC[k] + * dml_ceil(mode_lib->vba.BytePerPixelDETC[k], 2) + / (mode_lib->vba.ReturnBW + * mode_lib->vba.ReadBandwidthPlaneChroma[k] + / mode_lib->vba.TotalDataReadBandwidth); + mode_lib->vba.LastPixelOfLineExtraWatermark = + dml_max( + mode_lib->vba.LastPixelOfLineExtraWatermark, + DataFabricLineDeliveryTimeChroma + - mode_lib->vba.DisplayPipeLineDeliveryTimeChroma[k]); + } + + mode_lib->vba.UrgentExtraLatency = mode_lib->vba.UrgentRoundTripAndOutOfOrderLatency + + (mode_lib->vba.TotalActiveDPP * mode_lib->vba.PixelChunkSizeInKByte + + mode_lib->vba.TotalDCCActiveDPP + * mode_lib->vba.MetaChunkSize) * 1024.0 + / mode_lib->vba.ReturnBW; + + if (mode_lib->vba.VirtualMemoryEnable) + mode_lib->vba.UrgentExtraLatency += mode_lib->vba.TotalActiveDPP + * mode_lib->vba.PTEChunkSize * 1024.0 / mode_lib->vba.ReturnBW; + + mode_lib->vba.UrgentWatermark = mode_lib->vba.UrgentLatency + + mode_lib->vba.LastPixelOfLineExtraWatermark + + mode_lib->vba.UrgentExtraLatency; + + DTRACE(" urgent_extra_latency = %fus", mode_lib->vba.UrgentExtraLatency); + DTRACE(" wm_urgent = %fus", mode_lib->vba.UrgentWatermark); + + mode_lib->vba.MemoryTripWatermark = mode_lib->vba.UrgentLatency; + + mode_lib->vba.TotalActiveWriteback = 0; + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + if (mode_lib->vba.WritebackEnable[k]) + mode_lib->vba.TotalActiveWriteback = mode_lib->vba.TotalActiveWriteback + 1; + } + + if (mode_lib->vba.TotalActiveWriteback <= 1) + mode_lib->vba.WritebackUrgentWatermark = mode_lib->vba.WritebackLatency; + else + mode_lib->vba.WritebackUrgentWatermark = mode_lib->vba.WritebackLatency + + mode_lib->vba.WritebackChunkSize * 1024.0 / 32 + / mode_lib->vba.SOCCLK; + + DTRACE(" wm_wb_urgent = %fus", mode_lib->vba.WritebackUrgentWatermark); + + // NB P-State/DRAM Clock Change Watermark + mode_lib->vba.DRAMClockChangeWatermark = mode_lib->vba.DRAMClockChangeLatency + + mode_lib->vba.UrgentWatermark; + + DTRACE(" wm_pstate_change = %fus", mode_lib->vba.DRAMClockChangeWatermark); + + DTRACE(" calculating wb pstate watermark"); + DTRACE(" total wb outputs %d", mode_lib->vba.TotalActiveWriteback); + DTRACE(" socclk frequency %f Mhz", mode_lib->vba.SOCCLK); + + if (mode_lib->vba.TotalActiveWriteback <= 1) + mode_lib->vba.WritebackDRAMClockChangeWatermark = + mode_lib->vba.DRAMClockChangeLatency + + mode_lib->vba.WritebackLatency; + else + mode_lib->vba.WritebackDRAMClockChangeWatermark = + mode_lib->vba.DRAMClockChangeLatency + + mode_lib->vba.WritebackLatency + + mode_lib->vba.WritebackChunkSize * 1024.0 / 32 + / mode_lib->vba.SOCCLK; + + DTRACE(" wm_wb_pstate %fus", mode_lib->vba.WritebackDRAMClockChangeWatermark); + + // Stutter Efficiency + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + mode_lib->vba.LinesInDETY[k] = mode_lib->vba.DETBufferSizeY[k] + / mode_lib->vba.BytePerPixelDETY[k] / mode_lib->vba.SwathWidthY[k]; + mode_lib->vba.LinesInDETYRoundedDownToSwath[k] = dml_floor( + mode_lib->vba.LinesInDETY[k], + mode_lib->vba.SwathHeightY[k]); + mode_lib->vba.FullDETBufferingTimeY[k] = + mode_lib->vba.LinesInDETYRoundedDownToSwath[k] + * (mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k]) + / mode_lib->vba.VRatio[k]; + if (mode_lib->vba.BytePerPixelDETC[k] > 0) { + mode_lib->vba.LinesInDETC[k] = mode_lib->vba.DETBufferSizeC[k] + / mode_lib->vba.BytePerPixelDETC[k] + / (mode_lib->vba.SwathWidthY[k] / 2); + mode_lib->vba.LinesInDETCRoundedDownToSwath[k] = dml_floor( + mode_lib->vba.LinesInDETC[k], + mode_lib->vba.SwathHeightC[k]); + mode_lib->vba.FullDETBufferingTimeC[k] = + mode_lib->vba.LinesInDETCRoundedDownToSwath[k] + * (mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k]) + / (mode_lib->vba.VRatio[k] / 2); + } else { + mode_lib->vba.LinesInDETC[k] = 0; + mode_lib->vba.LinesInDETCRoundedDownToSwath[k] = 0; + mode_lib->vba.FullDETBufferingTimeC[k] = 999999; + } + } + + mode_lib->vba.MinFullDETBufferingTime = 999999.0; + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + if (mode_lib->vba.FullDETBufferingTimeY[k] + < mode_lib->vba.MinFullDETBufferingTime) { + mode_lib->vba.MinFullDETBufferingTime = + mode_lib->vba.FullDETBufferingTimeY[k]; + mode_lib->vba.FrameTimeForMinFullDETBufferingTime = + (double) mode_lib->vba.VTotal[k] * mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k]; + } + if (mode_lib->vba.FullDETBufferingTimeC[k] + < mode_lib->vba.MinFullDETBufferingTime) { + mode_lib->vba.MinFullDETBufferingTime = + mode_lib->vba.FullDETBufferingTimeC[k]; + mode_lib->vba.FrameTimeForMinFullDETBufferingTime = + (double) mode_lib->vba.VTotal[k] * mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k]; + } + } + + mode_lib->vba.AverageReadBandwidthGBytePerSecond = 0.0; + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + if (mode_lib->vba.DCCEnable[k]) { + mode_lib->vba.AverageReadBandwidthGBytePerSecond = + mode_lib->vba.AverageReadBandwidthGBytePerSecond + + mode_lib->vba.ReadBandwidthPlaneLuma[k] + / mode_lib->vba.DCCRate[k] + / 1000 + + mode_lib->vba.ReadBandwidthPlaneChroma[k] + / mode_lib->vba.DCCRate[k] + / 1000; + } else { + mode_lib->vba.AverageReadBandwidthGBytePerSecond = + mode_lib->vba.AverageReadBandwidthGBytePerSecond + + mode_lib->vba.ReadBandwidthPlaneLuma[k] + / 1000 + + mode_lib->vba.ReadBandwidthPlaneChroma[k] + / 1000; + } + if (mode_lib->vba.DCCEnable[k]) { + mode_lib->vba.AverageReadBandwidthGBytePerSecond = + mode_lib->vba.AverageReadBandwidthGBytePerSecond + + mode_lib->vba.ReadBandwidthPlaneLuma[k] + / 1000 / 256 + + mode_lib->vba.ReadBandwidthPlaneChroma[k] + / 1000 / 256; + } + if (mode_lib->vba.VirtualMemoryEnable) { + mode_lib->vba.AverageReadBandwidthGBytePerSecond = + mode_lib->vba.AverageReadBandwidthGBytePerSecond + + mode_lib->vba.ReadBandwidthPlaneLuma[k] + / 1000 / 512 + + mode_lib->vba.ReadBandwidthPlaneChroma[k] + / 1000 / 512; + } + } + + mode_lib->vba.PartOfBurstThatFitsInROB = + dml_min( + mode_lib->vba.MinFullDETBufferingTime + * mode_lib->vba.TotalDataReadBandwidth, + mode_lib->vba.ROBBufferSizeInKByte * 1024 + * mode_lib->vba.TotalDataReadBandwidth + / (mode_lib->vba.AverageReadBandwidthGBytePerSecond + * 1000)); + mode_lib->vba.StutterBurstTime = mode_lib->vba.PartOfBurstThatFitsInROB + * (mode_lib->vba.AverageReadBandwidthGBytePerSecond * 1000) + / mode_lib->vba.TotalDataReadBandwidth / mode_lib->vba.ReturnBW + + (mode_lib->vba.MinFullDETBufferingTime + * mode_lib->vba.TotalDataReadBandwidth + - mode_lib->vba.PartOfBurstThatFitsInROB) + / (mode_lib->vba.DCFCLK * 64); + if (mode_lib->vba.TotalActiveWriteback == 0) { + mode_lib->vba.StutterEfficiencyNotIncludingVBlank = (1 + - (mode_lib->vba.SRExitTime + mode_lib->vba.StutterBurstTime) + / mode_lib->vba.MinFullDETBufferingTime) * 100; + } else { + mode_lib->vba.StutterEfficiencyNotIncludingVBlank = 0; + } + + mode_lib->vba.SmallestVBlank = 999999; + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + if (mode_lib->vba.SynchronizedVBlank || mode_lib->vba.NumberOfActivePlanes == 1) { + mode_lib->vba.VBlankTime = (double) (mode_lib->vba.VTotal[k] + - mode_lib->vba.VActive[k]) * mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k]; + } else { + mode_lib->vba.VBlankTime = 0; + } + mode_lib->vba.SmallestVBlank = dml_min( + mode_lib->vba.SmallestVBlank, + mode_lib->vba.VBlankTime); + } + + mode_lib->vba.StutterEfficiency = (mode_lib->vba.StutterEfficiencyNotIncludingVBlank / 100 + * (mode_lib->vba.FrameTimeForMinFullDETBufferingTime + - mode_lib->vba.SmallestVBlank) + + mode_lib->vba.SmallestVBlank) + / mode_lib->vba.FrameTimeForMinFullDETBufferingTime * 100; + + // dml_ml->vba.DCFCLK Deep Sleep + mode_lib->vba.DCFClkDeepSleep = 8.0; + + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; k++) { + if (mode_lib->vba.BytePerPixelDETC[k] > 0) { + mode_lib->vba.DCFCLKDeepSleepPerPlane = + dml_max( + 1.1 * mode_lib->vba.SwathWidthY[k] + * dml_ceil( + mode_lib->vba.BytePerPixelDETY[k], + 1) / 32 + / mode_lib->vba.DisplayPipeLineDeliveryTimeLuma[k], + 1.1 * mode_lib->vba.SwathWidthY[k] / 2.0 + * dml_ceil( + mode_lib->vba.BytePerPixelDETC[k], + 2) / 32 + / mode_lib->vba.DisplayPipeLineDeliveryTimeChroma[k]); + } else + mode_lib->vba.DCFCLKDeepSleepPerPlane = 1.1 * mode_lib->vba.SwathWidthY[k] + * dml_ceil(mode_lib->vba.BytePerPixelDETY[k], 1) / 64.0 + / mode_lib->vba.DisplayPipeLineDeliveryTimeLuma[k]; + mode_lib->vba.DCFCLKDeepSleepPerPlane = dml_max( + mode_lib->vba.DCFCLKDeepSleepPerPlane, + mode_lib->vba.PixelClock[k] / 16.0); + mode_lib->vba.DCFClkDeepSleep = dml_max( + mode_lib->vba.DCFClkDeepSleep, + mode_lib->vba.DCFCLKDeepSleepPerPlane); + + DTRACE( + " dcfclk_deepsleep_per_plane[%i] = %fMHz", + k, + mode_lib->vba.DCFCLKDeepSleepPerPlane); + } + + DTRACE(" dcfclk_deepsleep_mhz = %fMHz", mode_lib->vba.DCFClkDeepSleep); + + // Stutter Watermark + mode_lib->vba.StutterExitWatermark = mode_lib->vba.SRExitTime + + mode_lib->vba.LastPixelOfLineExtraWatermark + + mode_lib->vba.UrgentExtraLatency + 10 / mode_lib->vba.DCFClkDeepSleep; + mode_lib->vba.StutterEnterPlusExitWatermark = mode_lib->vba.SREnterPlusExitTime + + mode_lib->vba.LastPixelOfLineExtraWatermark + + mode_lib->vba.UrgentExtraLatency; + + DTRACE(" wm_cstate_exit = %fus", mode_lib->vba.StutterExitWatermark); + DTRACE(" wm_cstate_enter_exit = %fus", mode_lib->vba.StutterEnterPlusExitWatermark); + + // Urgent Latency Supported + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + mode_lib->vba.EffectiveDETPlusLBLinesLuma = + dml_floor( + mode_lib->vba.LinesInDETY[k] + + dml_min( + mode_lib->vba.LinesInDETY[k] + * mode_lib->vba.DPPCLK[k] + * mode_lib->vba.BytePerPixelDETY[k] + * mode_lib->vba.PSCL_THROUGHPUT_LUMA[k] + / (mode_lib->vba.ReturnBW + / mode_lib->vba.DPPPerPlane[k]), + (double) mode_lib->vba.EffectiveLBLatencyHidingSourceLinesLuma), + mode_lib->vba.SwathHeightY[k]); + + mode_lib->vba.UrgentLatencySupportUsLuma = mode_lib->vba.EffectiveDETPlusLBLinesLuma + * (mode_lib->vba.HTotal[k] / mode_lib->vba.PixelClock[k]) + / mode_lib->vba.VRatio[k] + - mode_lib->vba.EffectiveDETPlusLBLinesLuma + * mode_lib->vba.SwathWidthY[k] + * mode_lib->vba.BytePerPixelDETY[k] + / (mode_lib->vba.ReturnBW + / mode_lib->vba.DPPPerPlane[k]); + + if (mode_lib->vba.BytePerPixelDETC[k] > 0) { + mode_lib->vba.EffectiveDETPlusLBLinesChroma = + dml_floor( + mode_lib->vba.LinesInDETC[k] + + dml_min( + mode_lib->vba.LinesInDETC[k] + * mode_lib->vba.DPPCLK[k] + * mode_lib->vba.BytePerPixelDETC[k] + * mode_lib->vba.PSCL_THROUGHPUT_CHROMA[k] + / (mode_lib->vba.ReturnBW + / mode_lib->vba.DPPPerPlane[k]), + (double) mode_lib->vba.EffectiveLBLatencyHidingSourceLinesChroma), + mode_lib->vba.SwathHeightC[k]); + mode_lib->vba.UrgentLatencySupportUsChroma = + mode_lib->vba.EffectiveDETPlusLBLinesChroma + * (mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k]) + / (mode_lib->vba.VRatio[k] / 2) + - mode_lib->vba.EffectiveDETPlusLBLinesChroma + * (mode_lib->vba.SwathWidthY[k] + / 2) + * mode_lib->vba.BytePerPixelDETC[k] + / (mode_lib->vba.ReturnBW + / mode_lib->vba.DPPPerPlane[k]); + mode_lib->vba.UrgentLatencySupportUs[k] = dml_min( + mode_lib->vba.UrgentLatencySupportUsLuma, + mode_lib->vba.UrgentLatencySupportUsChroma); + } else { + mode_lib->vba.UrgentLatencySupportUs[k] = + mode_lib->vba.UrgentLatencySupportUsLuma; + } + } + + mode_lib->vba.MinUrgentLatencySupportUs = 999999; + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + mode_lib->vba.MinUrgentLatencySupportUs = dml_min( + mode_lib->vba.MinUrgentLatencySupportUs, + mode_lib->vba.UrgentLatencySupportUs[k]); + } + + // Non-Urgent Latency Tolerance + mode_lib->vba.NonUrgentLatencyTolerance = mode_lib->vba.MinUrgentLatencySupportUs + - mode_lib->vba.UrgentWatermark; + + // DSCCLK + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + if ((mode_lib->vba.BlendingAndTiming[k] != k) || !mode_lib->vba.DSCEnabled[k]) { + mode_lib->vba.DSCCLK_calculated[k] = 0.0; + } else { + if (mode_lib->vba.OutputFormat[k] == dm_420 + || mode_lib->vba.OutputFormat[k] == dm_n422) + mode_lib->vba.DSCFormatFactor = 2; + else + mode_lib->vba.DSCFormatFactor = 1; + if (mode_lib->vba.ODMCombineEnabled[k]) + mode_lib->vba.DSCCLK_calculated[k] = + mode_lib->vba.PixelClockBackEnd[k] / 6 + / mode_lib->vba.DSCFormatFactor + / (1 + - mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading + / 100); + else + mode_lib->vba.DSCCLK_calculated[k] = + mode_lib->vba.PixelClockBackEnd[k] / 3 + / mode_lib->vba.DSCFormatFactor + / (1 + - mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading + / 100); + } + } + + // DSC Delay + // TODO + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + double bpp = mode_lib->vba.OutputBpp[k]; + unsigned int slices = mode_lib->vba.NumberOfDSCSlices[k]; + + if (mode_lib->vba.DSCEnabled[k] && bpp != 0) { + if (!mode_lib->vba.ODMCombineEnabled[k]) { + mode_lib->vba.DSCDelay[k] = + dscceComputeDelay( + mode_lib->vba.DSCInputBitPerComponent[k], + bpp, + dml_ceil( + (double) mode_lib->vba.HActive[k] + / mode_lib->vba.NumberOfDSCSlices[k], + 1), + slices, + mode_lib->vba.OutputFormat[k]) + + dscComputeDelay( + mode_lib->vba.OutputFormat[k]); + } else { + mode_lib->vba.DSCDelay[k] = + 2 + * (dscceComputeDelay( + mode_lib->vba.DSCInputBitPerComponent[k], + bpp, + dml_ceil( + (double) mode_lib->vba.HActive[k] + / mode_lib->vba.NumberOfDSCSlices[k], + 1), + slices / 2.0, + mode_lib->vba.OutputFormat[k]) + + dscComputeDelay( + mode_lib->vba.OutputFormat[k])); + } + mode_lib->vba.DSCDelay[k] = mode_lib->vba.DSCDelay[k] + * mode_lib->vba.PixelClock[k] + / mode_lib->vba.PixelClockBackEnd[k]; + } else { + mode_lib->vba.DSCDelay[k] = 0; + } + } + + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) + for (j = 0; j < mode_lib->vba.NumberOfActivePlanes; ++j) // NumberOfPlanes + if (j != k && mode_lib->vba.BlendingAndTiming[k] == j + && mode_lib->vba.DSCEnabled[j]) + mode_lib->vba.DSCDelay[k] = mode_lib->vba.DSCDelay[j]; + + // Prefetch + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + unsigned int PDEAndMetaPTEBytesFrameY; + unsigned int PixelPTEBytesPerRowY; + unsigned int MetaRowByteY; + unsigned int MetaRowByteC; + unsigned int PDEAndMetaPTEBytesFrameC; + unsigned int PixelPTEBytesPerRowC; + + Calculate256BBlockSizes( + mode_lib->vba.SourcePixelFormat[k], + mode_lib->vba.SurfaceTiling[k], + dml_ceil(mode_lib->vba.BytePerPixelDETY[k], 1), + dml_ceil(mode_lib->vba.BytePerPixelDETC[k], 2), + &mode_lib->vba.BlockHeight256BytesY[k], + &mode_lib->vba.BlockHeight256BytesC[k], + &mode_lib->vba.BlockWidth256BytesY[k], + &mode_lib->vba.BlockWidth256BytesC[k]); + PDEAndMetaPTEBytesFrameY = CalculateVMAndRowBytes( + mode_lib, + mode_lib->vba.DCCEnable[k], + mode_lib->vba.BlockHeight256BytesY[k], + mode_lib->vba.BlockWidth256BytesY[k], + mode_lib->vba.SourcePixelFormat[k], + mode_lib->vba.SurfaceTiling[k], + dml_ceil(mode_lib->vba.BytePerPixelDETY[k], 1), + mode_lib->vba.SourceScan[k], + mode_lib->vba.ViewportWidth[k], + mode_lib->vba.ViewportHeight[k], + mode_lib->vba.SwathWidthY[k], + mode_lib->vba.VirtualMemoryEnable, + mode_lib->vba.VMMPageSize, + mode_lib->vba.PTEBufferSizeInRequests, + mode_lib->vba.PDEProcessingBufIn64KBReqs, + mode_lib->vba.PitchY[k], + mode_lib->vba.DCCMetaPitchY[k], + &mode_lib->vba.MacroTileWidthY[k], + &MetaRowByteY, + &PixelPTEBytesPerRowY, + &mode_lib->vba.PTEBufferSizeNotExceeded[mode_lib->vba.VoltageLevel], + &mode_lib->vba.dpte_row_height[k], + &mode_lib->vba.meta_row_height[k]); + mode_lib->vba.PrefetchSourceLinesY[k] = CalculatePrefetchSourceLines( + mode_lib, + mode_lib->vba.VRatio[k], + mode_lib->vba.vtaps[k], + mode_lib->vba.Interlace[k], + mode_lib->vba.ProgressiveToInterlaceUnitInOPP, + mode_lib->vba.SwathHeightY[k], + mode_lib->vba.ViewportYStartY[k], + &mode_lib->vba.VInitPreFillY[k], + &mode_lib->vba.MaxNumSwathY[k]); + + if ((mode_lib->vba.SourcePixelFormat[k] != dm_444_64 + && mode_lib->vba.SourcePixelFormat[k] != dm_444_32 + && mode_lib->vba.SourcePixelFormat[k] != dm_444_16 + && mode_lib->vba.SourcePixelFormat[k] != dm_444_8)) { + PDEAndMetaPTEBytesFrameC = + CalculateVMAndRowBytes( + mode_lib, + mode_lib->vba.DCCEnable[k], + mode_lib->vba.BlockHeight256BytesC[k], + mode_lib->vba.BlockWidth256BytesC[k], + mode_lib->vba.SourcePixelFormat[k], + mode_lib->vba.SurfaceTiling[k], + dml_ceil( + mode_lib->vba.BytePerPixelDETC[k], + 2), + mode_lib->vba.SourceScan[k], + mode_lib->vba.ViewportWidth[k] / 2, + mode_lib->vba.ViewportHeight[k] / 2, + mode_lib->vba.SwathWidthY[k] / 2, + mode_lib->vba.VirtualMemoryEnable, + mode_lib->vba.VMMPageSize, + mode_lib->vba.PTEBufferSizeInRequests, + mode_lib->vba.PDEProcessingBufIn64KBReqs, + mode_lib->vba.PitchC[k], + 0, + &mode_lib->vba.MacroTileWidthC[k], + &MetaRowByteC, + &PixelPTEBytesPerRowC, + &mode_lib->vba.PTEBufferSizeNotExceeded[mode_lib->vba.VoltageLevel], + &mode_lib->vba.dpte_row_height_chroma[k], + &mode_lib->vba.meta_row_height_chroma[k]); + mode_lib->vba.PrefetchSourceLinesC[k] = CalculatePrefetchSourceLines( + mode_lib, + mode_lib->vba.VRatio[k] / 2, + mode_lib->vba.VTAPsChroma[k], + mode_lib->vba.Interlace[k], + mode_lib->vba.ProgressiveToInterlaceUnitInOPP, + mode_lib->vba.SwathHeightC[k], + mode_lib->vba.ViewportYStartC[k], + &mode_lib->vba.VInitPreFillC[k], + &mode_lib->vba.MaxNumSwathC[k]); + } else { + PixelPTEBytesPerRowC = 0; + PDEAndMetaPTEBytesFrameC = 0; + MetaRowByteC = 0; + mode_lib->vba.MaxNumSwathC[k] = 0; + mode_lib->vba.PrefetchSourceLinesC[k] = 0; + } + + mode_lib->vba.PixelPTEBytesPerRow[k] = PixelPTEBytesPerRowY + PixelPTEBytesPerRowC; + mode_lib->vba.PDEAndMetaPTEBytesFrame[k] = PDEAndMetaPTEBytesFrameY + + PDEAndMetaPTEBytesFrameC; + mode_lib->vba.MetaRowByte[k] = MetaRowByteY + MetaRowByteC; + + CalculateActiveRowBandwidth( + mode_lib->vba.VirtualMemoryEnable, + mode_lib->vba.SourcePixelFormat[k], + mode_lib->vba.VRatio[k], + mode_lib->vba.DCCEnable[k], + mode_lib->vba.HTotal[k] / mode_lib->vba.PixelClock[k], + MetaRowByteY, + MetaRowByteC, + mode_lib->vba.meta_row_height[k], + mode_lib->vba.meta_row_height_chroma[k], + PixelPTEBytesPerRowY, + PixelPTEBytesPerRowC, + mode_lib->vba.dpte_row_height[k], + mode_lib->vba.dpte_row_height_chroma[k], + &mode_lib->vba.meta_row_bw[k], + &mode_lib->vba.dpte_row_bw[k], + &mode_lib->vba.qual_row_bw[k]); + } + + mode_lib->vba.TCalc = 24.0 / mode_lib->vba.DCFClkDeepSleep; + + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + if (mode_lib->vba.BlendingAndTiming[k] == k) { + if (mode_lib->vba.WritebackEnable[k] == true) { + mode_lib->vba.WritebackDelay[mode_lib->vba.VoltageLevel][k] = + mode_lib->vba.WritebackLatency + + CalculateWriteBackDelay( + mode_lib->vba.WritebackPixelFormat[k], + mode_lib->vba.WritebackHRatio[k], + mode_lib->vba.WritebackVRatio[k], + mode_lib->vba.WritebackLumaHTaps[k], + mode_lib->vba.WritebackLumaVTaps[k], + mode_lib->vba.WritebackChromaHTaps[k], + mode_lib->vba.WritebackChromaVTaps[k], + mode_lib->vba.WritebackDestinationWidth[k]) + / mode_lib->vba.DISPCLK; + } else + mode_lib->vba.WritebackDelay[mode_lib->vba.VoltageLevel][k] = 0; + for (j = 0; j < mode_lib->vba.NumberOfActivePlanes; ++j) { + if (mode_lib->vba.BlendingAndTiming[j] == k + && mode_lib->vba.WritebackEnable[j] == true) { + mode_lib->vba.WritebackDelay[mode_lib->vba.VoltageLevel][k] = + dml_max( + mode_lib->vba.WritebackDelay[mode_lib->vba.VoltageLevel][k], + mode_lib->vba.WritebackLatency + + CalculateWriteBackDelay( + mode_lib->vba.WritebackPixelFormat[j], + mode_lib->vba.WritebackHRatio[j], + mode_lib->vba.WritebackVRatio[j], + mode_lib->vba.WritebackLumaHTaps[j], + mode_lib->vba.WritebackLumaVTaps[j], + mode_lib->vba.WritebackChromaHTaps[j], + mode_lib->vba.WritebackChromaVTaps[j], + mode_lib->vba.WritebackDestinationWidth[j]) + / mode_lib->vba.DISPCLK); + } + } + } + } + + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) + for (j = 0; j < mode_lib->vba.NumberOfActivePlanes; ++j) + if (mode_lib->vba.BlendingAndTiming[k] == j) + mode_lib->vba.WritebackDelay[mode_lib->vba.VoltageLevel][k] = + mode_lib->vba.WritebackDelay[mode_lib->vba.VoltageLevel][j]; + + mode_lib->vba.VStartupLines = 13; + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + mode_lib->vba.MaxVStartupLines[k] = + mode_lib->vba.VTotal[k] - mode_lib->vba.VActive[k] + - dml_max( + 1.0, + dml_ceil( + mode_lib->vba.WritebackDelay[mode_lib->vba.VoltageLevel][k] + / (mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k]), + 1)); + } + + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) + mode_lib->vba.MaximumMaxVStartupLines = dml_max( + mode_lib->vba.MaximumMaxVStartupLines, + mode_lib->vba.MaxVStartupLines[k]); + + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + mode_lib->vba.cursor_bw[k] = 0.0; + for (j = 0; j < mode_lib->vba.NumberOfCursors[k]; ++j) + mode_lib->vba.cursor_bw[k] += mode_lib->vba.CursorWidth[k][j] + * mode_lib->vba.CursorBPP[k][j] / 8.0 + / (mode_lib->vba.HTotal[k] / mode_lib->vba.PixelClock[k]) + * mode_lib->vba.VRatio[k]; + } + + do { + double MaxTotalRDBandwidth = 0; + bool DestinationLineTimesForPrefetchLessThan2 = false; + bool VRatioPrefetchMoreThan4 = false; + bool prefetch_vm_bw_valid = true; + bool prefetch_row_bw_valid = true; + double TWait = CalculateTWait( + mode_lib->vba.PrefetchMode, + mode_lib->vba.DRAMClockChangeLatency, + mode_lib->vba.UrgentLatency, + mode_lib->vba.SREnterPlusExitTime); + + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + if (mode_lib->vba.XFCEnabled[k] == true) { + mode_lib->vba.XFCRemoteSurfaceFlipDelay = + CalculateRemoteSurfaceFlipDelay( + mode_lib, + mode_lib->vba.VRatio[k], + mode_lib->vba.SwathWidthY[k], + dml_ceil( + mode_lib->vba.BytePerPixelDETY[k], + 1), + mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k], + mode_lib->vba.XFCTSlvVupdateOffset, + mode_lib->vba.XFCTSlvVupdateWidth, + mode_lib->vba.XFCTSlvVreadyOffset, + mode_lib->vba.XFCXBUFLatencyTolerance, + mode_lib->vba.XFCFillBWOverhead, + mode_lib->vba.XFCSlvChunkSize, + mode_lib->vba.XFCBusTransportTime, + mode_lib->vba.TCalc, + TWait, + &mode_lib->vba.SrcActiveDrainRate, + &mode_lib->vba.TInitXFill, + &mode_lib->vba.TslvChk); + } else { + mode_lib->vba.XFCRemoteSurfaceFlipDelay = 0; + } + mode_lib->vba.ErrorResult[k] = + CalculatePrefetchSchedule( + mode_lib, + mode_lib->vba.DPPCLK[k], + mode_lib->vba.DISPCLK, + mode_lib->vba.PixelClock[k], + mode_lib->vba.DCFClkDeepSleep, + mode_lib->vba.DSCDelay[k], + mode_lib->vba.DPPPerPlane[k], + mode_lib->vba.ScalerEnabled[k], + mode_lib->vba.NumberOfCursors[k], + mode_lib->vba.DPPCLKDelaySubtotal, + mode_lib->vba.DPPCLKDelaySCL, + mode_lib->vba.DPPCLKDelaySCLLBOnly, + mode_lib->vba.DPPCLKDelayCNVCFormater, + mode_lib->vba.DPPCLKDelayCNVCCursor, + mode_lib->vba.DISPCLKDelaySubtotal, + (unsigned int) (mode_lib->vba.SwathWidthY[k] + / mode_lib->vba.HRatio[k]), + mode_lib->vba.OutputFormat[k], + mode_lib->vba.VTotal[k] + - mode_lib->vba.VActive[k], + mode_lib->vba.HTotal[k], + mode_lib->vba.MaxInterDCNTileRepeaters, + dml_min( + mode_lib->vba.VStartupLines, + mode_lib->vba.MaxVStartupLines[k]), + mode_lib->vba.MaxPageTableLevels, + mode_lib->vba.VirtualMemoryEnable, + mode_lib->vba.DynamicMetadataEnable[k], + mode_lib->vba.DynamicMetadataLinesBeforeActiveRequired[k], + mode_lib->vba.DynamicMetadataTransmittedBytes[k], + mode_lib->vba.DCCEnable[k], + mode_lib->vba.UrgentLatency, + mode_lib->vba.UrgentExtraLatency, + mode_lib->vba.TCalc, + mode_lib->vba.PDEAndMetaPTEBytesFrame[k], + mode_lib->vba.MetaRowByte[k], + mode_lib->vba.PixelPTEBytesPerRow[k], + mode_lib->vba.PrefetchSourceLinesY[k], + mode_lib->vba.SwathWidthY[k], + mode_lib->vba.BytePerPixelDETY[k], + mode_lib->vba.VInitPreFillY[k], + mode_lib->vba.MaxNumSwathY[k], + mode_lib->vba.PrefetchSourceLinesC[k], + mode_lib->vba.BytePerPixelDETC[k], + mode_lib->vba.VInitPreFillC[k], + mode_lib->vba.MaxNumSwathC[k], + mode_lib->vba.SwathHeightY[k], + mode_lib->vba.SwathHeightC[k], + TWait, + mode_lib->vba.XFCEnabled[k], + mode_lib->vba.XFCRemoteSurfaceFlipDelay, + mode_lib->vba.Interlace[k], + mode_lib->vba.ProgressiveToInterlaceUnitInOPP, + &mode_lib->vba.DSTXAfterScaler[k], + &mode_lib->vba.DSTYAfterScaler[k], + &mode_lib->vba.DestinationLinesForPrefetch[k], + &mode_lib->vba.PrefetchBandwidth[k], + &mode_lib->vba.DestinationLinesToRequestVMInVBlank[k], + &mode_lib->vba.DestinationLinesToRequestRowInVBlank[k], + &mode_lib->vba.VRatioPrefetchY[k], + &mode_lib->vba.VRatioPrefetchC[k], + &mode_lib->vba.RequiredPrefetchPixDataBW[k], + &mode_lib->vba.VStartupRequiredWhenNotEnoughTimeForDynamicMetadata, + &mode_lib->vba.Tno_bw[k], + &mode_lib->vba.VUpdateOffsetPix[k], + &mode_lib->vba.VUpdateWidthPix[k], + &mode_lib->vba.VReadyOffsetPix[k]); + if (mode_lib->vba.BlendingAndTiming[k] == k) { + mode_lib->vba.VStartup[k] = dml_min( + mode_lib->vba.VStartupLines, + mode_lib->vba.MaxVStartupLines[k]); + if (mode_lib->vba.VStartupRequiredWhenNotEnoughTimeForDynamicMetadata + != 0) { + mode_lib->vba.VStartup[k] = + mode_lib->vba.VStartupRequiredWhenNotEnoughTimeForDynamicMetadata; + } + } else { + mode_lib->vba.VStartup[k] = + dml_min( + mode_lib->vba.VStartupLines, + mode_lib->vba.MaxVStartupLines[mode_lib->vba.BlendingAndTiming[k]]); + } + } + + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + + if (mode_lib->vba.PDEAndMetaPTEBytesFrame[k] == 0) + mode_lib->vba.prefetch_vm_bw[k] = 0; + else if (mode_lib->vba.DestinationLinesToRequestVMInVBlank[k] > 0) { + mode_lib->vba.prefetch_vm_bw[k] = + (double) mode_lib->vba.PDEAndMetaPTEBytesFrame[k] + / (mode_lib->vba.DestinationLinesToRequestVMInVBlank[k] + * mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k]); + } else { + mode_lib->vba.prefetch_vm_bw[k] = 0; + prefetch_vm_bw_valid = false; + } + if (mode_lib->vba.MetaRowByte[k] + mode_lib->vba.PixelPTEBytesPerRow[k] + == 0) + mode_lib->vba.prefetch_row_bw[k] = 0; + else if (mode_lib->vba.DestinationLinesToRequestRowInVBlank[k] > 0) { + mode_lib->vba.prefetch_row_bw[k] = + (double) (mode_lib->vba.MetaRowByte[k] + + mode_lib->vba.PixelPTEBytesPerRow[k]) + / (mode_lib->vba.DestinationLinesToRequestRowInVBlank[k] + * mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k]); + } else { + mode_lib->vba.prefetch_row_bw[k] = 0; + prefetch_row_bw_valid = false; + } + + MaxTotalRDBandwidth = + MaxTotalRDBandwidth + mode_lib->vba.cursor_bw[k] + + dml_max( + mode_lib->vba.prefetch_vm_bw[k], + dml_max( + mode_lib->vba.prefetch_row_bw[k], + dml_max( + mode_lib->vba.ReadBandwidthPlaneLuma[k] + + mode_lib->vba.ReadBandwidthPlaneChroma[k], + mode_lib->vba.RequiredPrefetchPixDataBW[k]) + + mode_lib->vba.meta_row_bw[k] + + mode_lib->vba.dpte_row_bw[k])); + + if (mode_lib->vba.DestinationLinesForPrefetch[k] < 2) + DestinationLineTimesForPrefetchLessThan2 = true; + if (mode_lib->vba.VRatioPrefetchY[k] > 4 + || mode_lib->vba.VRatioPrefetchC[k] > 4) + VRatioPrefetchMoreThan4 = true; + } + + if (MaxTotalRDBandwidth <= mode_lib->vba.ReturnBW && prefetch_vm_bw_valid + && prefetch_row_bw_valid && !VRatioPrefetchMoreThan4 + && !DestinationLineTimesForPrefetchLessThan2) + mode_lib->vba.PrefetchModeSupported = true; + else { + mode_lib->vba.PrefetchModeSupported = false; + dml_print( + "DML: CalculatePrefetchSchedule ***failed***. Bandwidth violation. Results are NOT valid\n"); + } + + if (mode_lib->vba.PrefetchModeSupported == true) { + double final_flip_bw[DC__NUM_DPP__MAX]; + unsigned int ImmediateFlipBytes[DC__NUM_DPP__MAX]; + double total_dcn_read_bw_with_flip = 0; + + mode_lib->vba.BandwidthAvailableForImmediateFlip = mode_lib->vba.ReturnBW; + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + mode_lib->vba.BandwidthAvailableForImmediateFlip = + mode_lib->vba.BandwidthAvailableForImmediateFlip + - mode_lib->vba.cursor_bw[k] + - dml_max( + mode_lib->vba.ReadBandwidthPlaneLuma[k] + + mode_lib->vba.ReadBandwidthPlaneChroma[k] + + mode_lib->vba.qual_row_bw[k], + mode_lib->vba.PrefetchBandwidth[k]); + } + + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + ImmediateFlipBytes[k] = 0; + if ((mode_lib->vba.SourcePixelFormat[k] != dm_420_8 + && mode_lib->vba.SourcePixelFormat[k] != dm_420_10)) { + ImmediateFlipBytes[k] = + mode_lib->vba.PDEAndMetaPTEBytesFrame[k] + + mode_lib->vba.MetaRowByte[k] + + mode_lib->vba.PixelPTEBytesPerRow[k]; + } + } + mode_lib->vba.TotImmediateFlipBytes = 0; + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + if ((mode_lib->vba.SourcePixelFormat[k] != dm_420_8 + && mode_lib->vba.SourcePixelFormat[k] != dm_420_10)) { + mode_lib->vba.TotImmediateFlipBytes = + mode_lib->vba.TotImmediateFlipBytes + + ImmediateFlipBytes[k]; + } + } + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + CalculateFlipSchedule( + mode_lib, + mode_lib->vba.UrgentExtraLatency, + mode_lib->vba.UrgentLatency, + mode_lib->vba.MaxPageTableLevels, + mode_lib->vba.VirtualMemoryEnable, + mode_lib->vba.BandwidthAvailableForImmediateFlip, + mode_lib->vba.TotImmediateFlipBytes, + mode_lib->vba.SourcePixelFormat[k], + ImmediateFlipBytes[k], + mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k], + mode_lib->vba.VRatio[k], + mode_lib->vba.Tno_bw[k], + mode_lib->vba.PDEAndMetaPTEBytesFrame[k], + mode_lib->vba.MetaRowByte[k], + mode_lib->vba.PixelPTEBytesPerRow[k], + mode_lib->vba.DCCEnable[k], + mode_lib->vba.dpte_row_height[k], + mode_lib->vba.meta_row_height[k], + mode_lib->vba.qual_row_bw[k], + &mode_lib->vba.DestinationLinesToRequestVMInImmediateFlip[k], + &mode_lib->vba.DestinationLinesToRequestRowInImmediateFlip[k], + &final_flip_bw[k], + &mode_lib->vba.ImmediateFlipSupportedForPipe[k]); + } + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + total_dcn_read_bw_with_flip = + total_dcn_read_bw_with_flip + + mode_lib->vba.cursor_bw[k] + + dml_max( + mode_lib->vba.prefetch_vm_bw[k], + dml_max( + mode_lib->vba.prefetch_row_bw[k], + final_flip_bw[k] + + dml_max( + mode_lib->vba.ReadBandwidthPlaneLuma[k] + + mode_lib->vba.ReadBandwidthPlaneChroma[k], + mode_lib->vba.RequiredPrefetchPixDataBW[k]))); + } + mode_lib->vba.ImmediateFlipSupported = true; + if (total_dcn_read_bw_with_flip > mode_lib->vba.ReturnBW) { + mode_lib->vba.ImmediateFlipSupported = false; + } + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + if (mode_lib->vba.ImmediateFlipSupportedForPipe[k] == false) { + mode_lib->vba.ImmediateFlipSupported = false; + } + } + } else { + mode_lib->vba.ImmediateFlipSupported = false; + } + + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + if (mode_lib->vba.ErrorResult[k]) { + mode_lib->vba.PrefetchModeSupported = false; + dml_print( + "DML: CalculatePrefetchSchedule ***failed***. Prefetch schedule violation. Results are NOT valid\n"); + } + } + + mode_lib->vba.VStartupLines = mode_lib->vba.VStartupLines + 1; + } while (!((mode_lib->vba.PrefetchModeSupported + && (!mode_lib->vba.ImmediateFlipSupport + || mode_lib->vba.ImmediateFlipSupported)) + || mode_lib->vba.MaximumMaxVStartupLines < mode_lib->vba.VStartupLines)); + + //Display Pipeline Delivery Time in Prefetch + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + if (mode_lib->vba.VRatioPrefetchY[k] <= 1) { + mode_lib->vba.DisplayPipeLineDeliveryTimeLumaPrefetch[k] = + mode_lib->vba.SwathWidthY[k] * mode_lib->vba.DPPPerPlane[k] + / mode_lib->vba.HRatio[k] + / mode_lib->vba.PixelClock[k]; + } else { + mode_lib->vba.DisplayPipeLineDeliveryTimeLumaPrefetch[k] = + mode_lib->vba.SwathWidthY[k] + / mode_lib->vba.PSCL_THROUGHPUT_LUMA[k] + / mode_lib->vba.DPPCLK[k]; + } + if (mode_lib->vba.BytePerPixelDETC[k] == 0) { + mode_lib->vba.DisplayPipeLineDeliveryTimeChromaPrefetch[k] = 0; + } else { + if (mode_lib->vba.VRatioPrefetchC[k] <= 1) { + mode_lib->vba.DisplayPipeLineDeliveryTimeChromaPrefetch[k] = + mode_lib->vba.SwathWidthY[k] + * mode_lib->vba.DPPPerPlane[k] + / mode_lib->vba.HRatio[k] + / mode_lib->vba.PixelClock[k]; + } else { + mode_lib->vba.DisplayPipeLineDeliveryTimeChromaPrefetch[k] = + mode_lib->vba.SwathWidthY[k] + / mode_lib->vba.PSCL_THROUGHPUT_LUMA[k] + / mode_lib->vba.DPPCLK[k]; + } + } + } + + // Min TTUVBlank + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + if (mode_lib->vba.PrefetchMode == 0) { + mode_lib->vba.AllowDRAMClockChangeDuringVBlank[k] = true; + mode_lib->vba.AllowDRAMSelfRefreshDuringVBlank[k] = true; + mode_lib->vba.MinTTUVBlank[k] = dml_max( + mode_lib->vba.DRAMClockChangeWatermark, + dml_max( + mode_lib->vba.StutterEnterPlusExitWatermark, + mode_lib->vba.UrgentWatermark)); + } else if (mode_lib->vba.PrefetchMode == 1) { + mode_lib->vba.AllowDRAMClockChangeDuringVBlank[k] = false; + mode_lib->vba.AllowDRAMSelfRefreshDuringVBlank[k] = true; + mode_lib->vba.MinTTUVBlank[k] = dml_max( + mode_lib->vba.StutterEnterPlusExitWatermark, + mode_lib->vba.UrgentWatermark); + } else { + mode_lib->vba.AllowDRAMClockChangeDuringVBlank[k] = false; + mode_lib->vba.AllowDRAMSelfRefreshDuringVBlank[k] = false; + mode_lib->vba.MinTTUVBlank[k] = mode_lib->vba.UrgentWatermark; + } + if (!mode_lib->vba.DynamicMetadataEnable[k]) + mode_lib->vba.MinTTUVBlank[k] = mode_lib->vba.TCalc + + mode_lib->vba.MinTTUVBlank[k]; + } + + // DCC Configuration + mode_lib->vba.ActiveDPPs = 0; + // NB P-State/DRAM Clock Change Support + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + mode_lib->vba.ActiveDPPs = mode_lib->vba.ActiveDPPs + mode_lib->vba.DPPPerPlane[k]; + } + + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + double EffectiveLBLatencyHidingY; + double EffectiveLBLatencyHidingC; + double DPPOutputBufferLinesY; + double DPPOutputBufferLinesC; + double DPPOPPBufferingY; + double MaxDETBufferingTimeY; + double ActiveDRAMClockChangeLatencyMarginY; + + mode_lib->vba.LBLatencyHidingSourceLinesY = + dml_min( + mode_lib->vba.MaxLineBufferLines, + (unsigned int) dml_floor( + (double) mode_lib->vba.LineBufferSize + / mode_lib->vba.LBBitPerPixel[k] + / (mode_lib->vba.SwathWidthY[k] + / dml_max( + mode_lib->vba.HRatio[k], + 1.0)), + 1)) - (mode_lib->vba.vtaps[k] - 1); + + mode_lib->vba.LBLatencyHidingSourceLinesC = + dml_min( + mode_lib->vba.MaxLineBufferLines, + (unsigned int) dml_floor( + (double) mode_lib->vba.LineBufferSize + / mode_lib->vba.LBBitPerPixel[k] + / (mode_lib->vba.SwathWidthY[k] + / 2.0 + / dml_max( + mode_lib->vba.HRatio[k] + / 2, + 1.0)), + 1)) + - (mode_lib->vba.VTAPsChroma[k] - 1); + + EffectiveLBLatencyHidingY = mode_lib->vba.LBLatencyHidingSourceLinesY + / mode_lib->vba.VRatio[k] + * (mode_lib->vba.HTotal[k] / mode_lib->vba.PixelClock[k]); + + EffectiveLBLatencyHidingC = mode_lib->vba.LBLatencyHidingSourceLinesC + / (mode_lib->vba.VRatio[k] / 2) + * (mode_lib->vba.HTotal[k] / mode_lib->vba.PixelClock[k]); + + if (mode_lib->vba.SwathWidthY[k] > 2 * mode_lib->vba.DPPOutputBufferPixels) { + DPPOutputBufferLinesY = mode_lib->vba.DPPOutputBufferPixels + / mode_lib->vba.SwathWidthY[k]; + } else if (mode_lib->vba.SwathWidthY[k] > mode_lib->vba.DPPOutputBufferPixels) { + DPPOutputBufferLinesY = 0.5; + } else { + DPPOutputBufferLinesY = 1; + } + + if (mode_lib->vba.SwathWidthY[k] / 2 > 2 * mode_lib->vba.DPPOutputBufferPixels) { + DPPOutputBufferLinesC = mode_lib->vba.DPPOutputBufferPixels + / (mode_lib->vba.SwathWidthY[k] / 2); + } else if (mode_lib->vba.SwathWidthY[k] / 2 > mode_lib->vba.DPPOutputBufferPixels) { + DPPOutputBufferLinesC = 0.5; + } else { + DPPOutputBufferLinesC = 1; + } + + DPPOPPBufferingY = (mode_lib->vba.HTotal[k] / mode_lib->vba.PixelClock[k]) + * (DPPOutputBufferLinesY + mode_lib->vba.OPPOutputBufferLines); + MaxDETBufferingTimeY = mode_lib->vba.FullDETBufferingTimeY[k] + + (mode_lib->vba.LinesInDETY[k] + - mode_lib->vba.LinesInDETYRoundedDownToSwath[k]) + / mode_lib->vba.SwathHeightY[k] + * (mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k]); + + ActiveDRAMClockChangeLatencyMarginY = DPPOPPBufferingY + EffectiveLBLatencyHidingY + + MaxDETBufferingTimeY - mode_lib->vba.DRAMClockChangeWatermark; + + if (mode_lib->vba.ActiveDPPs > 1) { + ActiveDRAMClockChangeLatencyMarginY = + ActiveDRAMClockChangeLatencyMarginY + - (1 - 1 / (mode_lib->vba.ActiveDPPs - 1)) + * mode_lib->vba.SwathHeightY[k] + * (mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k]); + } + + if (mode_lib->vba.BytePerPixelDETC[k] > 0) { + double DPPOPPBufferingC = (mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k]) + * (DPPOutputBufferLinesC + + mode_lib->vba.OPPOutputBufferLines); + double MaxDETBufferingTimeC = + mode_lib->vba.FullDETBufferingTimeC[k] + + (mode_lib->vba.LinesInDETC[k] + - mode_lib->vba.LinesInDETCRoundedDownToSwath[k]) + / mode_lib->vba.SwathHeightC[k] + * (mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k]); + double ActiveDRAMClockChangeLatencyMarginC = DPPOPPBufferingC + + EffectiveLBLatencyHidingC + MaxDETBufferingTimeC + - mode_lib->vba.DRAMClockChangeWatermark; + + if (mode_lib->vba.ActiveDPPs > 1) { + ActiveDRAMClockChangeLatencyMarginC = + ActiveDRAMClockChangeLatencyMarginC + - (1 + - 1 + / (mode_lib->vba.ActiveDPPs + - 1)) + * mode_lib->vba.SwathHeightC[k] + * (mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k]); + } + mode_lib->vba.ActiveDRAMClockChangeLatencyMargin[k] = dml_min( + ActiveDRAMClockChangeLatencyMarginY, + ActiveDRAMClockChangeLatencyMarginC); + } else { + mode_lib->vba.ActiveDRAMClockChangeLatencyMargin[k] = + ActiveDRAMClockChangeLatencyMarginY; + } + + if (mode_lib->vba.WritebackEnable[k]) { + double WritebackDRAMClockChangeLatencyMargin; + + if (mode_lib->vba.WritebackPixelFormat[k] == dm_444_32) { + WritebackDRAMClockChangeLatencyMargin = + (double) (mode_lib->vba.WritebackInterfaceLumaBufferSize + + mode_lib->vba.WritebackInterfaceChromaBufferSize) + / (mode_lib->vba.WritebackDestinationWidth[k] + * mode_lib->vba.WritebackDestinationHeight[k] + / (mode_lib->vba.WritebackSourceHeight[k] + * mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k]) + * 4) + - mode_lib->vba.WritebackDRAMClockChangeWatermark; + } else if (mode_lib->vba.WritebackPixelFormat[k] == dm_420_10) { + WritebackDRAMClockChangeLatencyMargin = + dml_min( + (double) mode_lib->vba.WritebackInterfaceLumaBufferSize + * 8.0 / 10, + 2.0 + * mode_lib->vba.WritebackInterfaceChromaBufferSize + * 8 / 10) + / (mode_lib->vba.WritebackDestinationWidth[k] + * mode_lib->vba.WritebackDestinationHeight[k] + / (mode_lib->vba.WritebackSourceHeight[k] + * mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k])) + - mode_lib->vba.WritebackDRAMClockChangeWatermark; + } else { + WritebackDRAMClockChangeLatencyMargin = + dml_min( + (double) mode_lib->vba.WritebackInterfaceLumaBufferSize, + 2.0 + * mode_lib->vba.WritebackInterfaceChromaBufferSize) + / (mode_lib->vba.WritebackDestinationWidth[k] + * mode_lib->vba.WritebackDestinationHeight[k] + / (mode_lib->vba.WritebackSourceHeight[k] + * mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k])) + - mode_lib->vba.WritebackDRAMClockChangeWatermark; + } + mode_lib->vba.ActiveDRAMClockChangeLatencyMargin[k] = dml_min( + mode_lib->vba.ActiveDRAMClockChangeLatencyMargin[k], + WritebackDRAMClockChangeLatencyMargin); + } + } + + mode_lib->vba.MinActiveDRAMClockChangeMargin = 999999; + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + if (mode_lib->vba.ActiveDRAMClockChangeLatencyMargin[k] + < mode_lib->vba.MinActiveDRAMClockChangeMargin) { + mode_lib->vba.MinActiveDRAMClockChangeMargin = + mode_lib->vba.ActiveDRAMClockChangeLatencyMargin[k]; + } + } + + mode_lib->vba.MinActiveDRAMClockChangeLatencySupported = + mode_lib->vba.MinActiveDRAMClockChangeMargin + + mode_lib->vba.DRAMClockChangeLatency; + + if (mode_lib->vba.MinActiveDRAMClockChangeMargin > 0) { + mode_lib->vba.DRAMClockChangeSupport = dm_dram_clock_change_vactive; + } else { + if (mode_lib->vba.SynchronizedVBlank || mode_lib->vba.NumberOfActivePlanes == 1) { + mode_lib->vba.DRAMClockChangeSupport = dm_dram_clock_change_vblank; + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + if (!mode_lib->vba.AllowDRAMClockChangeDuringVBlank[k]) { + mode_lib->vba.DRAMClockChangeSupport = + dm_dram_clock_change_unsupported; + } + } + } else { + mode_lib->vba.DRAMClockChangeSupport = dm_dram_clock_change_unsupported; + } + } + + //XFC Parameters: + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + if (mode_lib->vba.XFCEnabled[k] == true) { + double TWait; + + mode_lib->vba.XFCSlaveVUpdateOffset[k] = mode_lib->vba.XFCTSlvVupdateOffset; + mode_lib->vba.XFCSlaveVupdateWidth[k] = mode_lib->vba.XFCTSlvVupdateWidth; + mode_lib->vba.XFCSlaveVReadyOffset[k] = mode_lib->vba.XFCTSlvVreadyOffset; + TWait = CalculateTWait( + mode_lib->vba.PrefetchMode, + mode_lib->vba.DRAMClockChangeLatency, + mode_lib->vba.UrgentLatency, + mode_lib->vba.SREnterPlusExitTime); + mode_lib->vba.XFCRemoteSurfaceFlipDelay = CalculateRemoteSurfaceFlipDelay( + mode_lib, + mode_lib->vba.VRatio[k], + mode_lib->vba.SwathWidthY[k], + dml_ceil(mode_lib->vba.BytePerPixelDETY[k], 1), + mode_lib->vba.HTotal[k] / mode_lib->vba.PixelClock[k], + mode_lib->vba.XFCTSlvVupdateOffset, + mode_lib->vba.XFCTSlvVupdateWidth, + mode_lib->vba.XFCTSlvVreadyOffset, + mode_lib->vba.XFCXBUFLatencyTolerance, + mode_lib->vba.XFCFillBWOverhead, + mode_lib->vba.XFCSlvChunkSize, + mode_lib->vba.XFCBusTransportTime, + mode_lib->vba.TCalc, + TWait, + &mode_lib->vba.SrcActiveDrainRate, + &mode_lib->vba.TInitXFill, + &mode_lib->vba.TslvChk); + mode_lib->vba.XFCRemoteSurfaceFlipLatency[k] = + dml_floor( + mode_lib->vba.XFCRemoteSurfaceFlipDelay + / (mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k]), + 1); + mode_lib->vba.XFCTransferDelay[k] = + dml_ceil( + mode_lib->vba.XFCBusTransportTime + / (mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k]), + 1); + mode_lib->vba.XFCPrechargeDelay[k] = + dml_ceil( + (mode_lib->vba.XFCBusTransportTime + + mode_lib->vba.TInitXFill + + mode_lib->vba.TslvChk) + / (mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k]), + 1); + mode_lib->vba.InitFillLevel = mode_lib->vba.XFCXBUFLatencyTolerance + * mode_lib->vba.SrcActiveDrainRate; + mode_lib->vba.FinalFillMargin = + (mode_lib->vba.DestinationLinesToRequestVMInVBlank[k] + + mode_lib->vba.DestinationLinesToRequestRowInVBlank[k]) + * mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k] + * mode_lib->vba.SrcActiveDrainRate + + mode_lib->vba.XFCFillConstant; + mode_lib->vba.FinalFillLevel = mode_lib->vba.XFCRemoteSurfaceFlipDelay + * mode_lib->vba.SrcActiveDrainRate + + mode_lib->vba.FinalFillMargin; + mode_lib->vba.RemainingFillLevel = dml_max( + 0.0, + mode_lib->vba.FinalFillLevel - mode_lib->vba.InitFillLevel); + mode_lib->vba.TFinalxFill = mode_lib->vba.RemainingFillLevel + / (mode_lib->vba.SrcActiveDrainRate + * mode_lib->vba.XFCFillBWOverhead / 100); + mode_lib->vba.XFCPrefetchMargin[k] = + mode_lib->vba.XFCRemoteSurfaceFlipDelay + + mode_lib->vba.TFinalxFill + + (mode_lib->vba.DestinationLinesToRequestVMInVBlank[k] + + mode_lib->vba.DestinationLinesToRequestRowInVBlank[k]) + * mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k]; + } else { + mode_lib->vba.XFCSlaveVUpdateOffset[k] = 0; + mode_lib->vba.XFCSlaveVupdateWidth[k] = 0; + mode_lib->vba.XFCSlaveVReadyOffset[k] = 0; + mode_lib->vba.XFCRemoteSurfaceFlipLatency[k] = 0; + mode_lib->vba.XFCPrechargeDelay[k] = 0; + mode_lib->vba.XFCTransferDelay[k] = 0; + mode_lib->vba.XFCPrefetchMargin[k] = 0; + } + } +} + +static void DisplayPipeConfiguration(struct display_mode_lib *mode_lib) +{ + double BytePerPixDETY; + double BytePerPixDETC; + double Read256BytesBlockHeightY; + double Read256BytesBlockHeightC; + double Read256BytesBlockWidthY; + double Read256BytesBlockWidthC; + double MaximumSwathHeightY; + double MaximumSwathHeightC; + double MinimumSwathHeightY; + double MinimumSwathHeightC; + double SwathWidth; + double SwathWidthGranularityY; + double SwathWidthGranularityC; + double RoundedUpMaxSwathSizeBytesY; + double RoundedUpMaxSwathSizeBytesC; + unsigned int j, k; + + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + bool MainPlaneDoesODMCombine = false; + + if (mode_lib->vba.SourcePixelFormat[k] == dm_444_64) { + BytePerPixDETY = 8; + BytePerPixDETC = 0; + } else if (mode_lib->vba.SourcePixelFormat[k] == dm_444_32) { + BytePerPixDETY = 4; + BytePerPixDETC = 0; + } else if (mode_lib->vba.SourcePixelFormat[k] == dm_444_16) { + BytePerPixDETY = 2; + BytePerPixDETC = 0; + } else if (mode_lib->vba.SourcePixelFormat[k] == dm_444_8) { + BytePerPixDETY = 1; + BytePerPixDETC = 0; + } else if (mode_lib->vba.SourcePixelFormat[k] == dm_420_8) { + BytePerPixDETY = 1; + BytePerPixDETC = 2; + } else { + BytePerPixDETY = 4.0 / 3.0; + BytePerPixDETC = 8.0 / 3.0; + } + + if ((mode_lib->vba.SourcePixelFormat[k] == dm_444_64 + || mode_lib->vba.SourcePixelFormat[k] == dm_444_32 + || mode_lib->vba.SourcePixelFormat[k] == dm_444_16 + || mode_lib->vba.SourcePixelFormat[k] == dm_444_8)) { + if (mode_lib->vba.SurfaceTiling[k] == dm_sw_linear) { + Read256BytesBlockHeightY = 1; + } else if (mode_lib->vba.SourcePixelFormat[k] == dm_444_64) { + Read256BytesBlockHeightY = 4; + } else if (mode_lib->vba.SourcePixelFormat[k] == dm_444_32 + || mode_lib->vba.SourcePixelFormat[k] == dm_444_16) { + Read256BytesBlockHeightY = 8; + } else { + Read256BytesBlockHeightY = 16; + } + Read256BytesBlockWidthY = 256 / dml_ceil(BytePerPixDETY, 1) + / Read256BytesBlockHeightY; + Read256BytesBlockHeightC = 0; + Read256BytesBlockWidthC = 0; + } else { + if (mode_lib->vba.SurfaceTiling[k] == dm_sw_linear) { + Read256BytesBlockHeightY = 1; + Read256BytesBlockHeightC = 1; + } else if (mode_lib->vba.SourcePixelFormat[k] == dm_420_8) { + Read256BytesBlockHeightY = 16; + Read256BytesBlockHeightC = 8; + } else { + Read256BytesBlockHeightY = 8; + Read256BytesBlockHeightC = 8; + } + Read256BytesBlockWidthY = 256 / dml_ceil(BytePerPixDETY, 1) + / Read256BytesBlockHeightY; + Read256BytesBlockWidthC = 256 / dml_ceil(BytePerPixDETC, 2) + / Read256BytesBlockHeightC; + } + + if (mode_lib->vba.SourceScan[k] == dm_horz) { + MaximumSwathHeightY = Read256BytesBlockHeightY; + MaximumSwathHeightC = Read256BytesBlockHeightC; + } else { + MaximumSwathHeightY = Read256BytesBlockWidthY; + MaximumSwathHeightC = Read256BytesBlockWidthC; + } + + if ((mode_lib->vba.SourcePixelFormat[k] == dm_444_64 + || mode_lib->vba.SourcePixelFormat[k] == dm_444_32 + || mode_lib->vba.SourcePixelFormat[k] == dm_444_16 + || mode_lib->vba.SourcePixelFormat[k] == dm_444_8)) { + if (mode_lib->vba.SurfaceTiling[k] == dm_sw_linear + || (mode_lib->vba.SourcePixelFormat[k] == dm_444_64 + && (mode_lib->vba.SurfaceTiling[k] + == dm_sw_4kb_s + || mode_lib->vba.SurfaceTiling[k] + == dm_sw_4kb_s_x + || mode_lib->vba.SurfaceTiling[k] + == dm_sw_64kb_s + || mode_lib->vba.SurfaceTiling[k] + == dm_sw_64kb_s_t + || mode_lib->vba.SurfaceTiling[k] + == dm_sw_64kb_s_x + || mode_lib->vba.SurfaceTiling[k] + == dm_sw_var_s + || mode_lib->vba.SurfaceTiling[k] + == dm_sw_var_s_x) + && mode_lib->vba.SourceScan[k] == dm_horz)) { + MinimumSwathHeightY = MaximumSwathHeightY; + } else if (mode_lib->vba.SourcePixelFormat[k] == dm_444_8 + && mode_lib->vba.SourceScan[k] != dm_horz) { + MinimumSwathHeightY = MaximumSwathHeightY; + } else { + MinimumSwathHeightY = MaximumSwathHeightY / 2.0; + } + MinimumSwathHeightC = MaximumSwathHeightC; + } else { + if (mode_lib->vba.SurfaceTiling[k] == dm_sw_linear) { + MinimumSwathHeightY = MaximumSwathHeightY; + MinimumSwathHeightC = MaximumSwathHeightC; + } else if (mode_lib->vba.SourcePixelFormat[k] == dm_420_8 + && mode_lib->vba.SourceScan[k] == dm_horz) { + MinimumSwathHeightY = MaximumSwathHeightY / 2.0; + MinimumSwathHeightC = MaximumSwathHeightC; + } else if (mode_lib->vba.SourcePixelFormat[k] == dm_420_10 + && mode_lib->vba.SourceScan[k] == dm_horz) { + MinimumSwathHeightC = MaximumSwathHeightC / 2.0; + MinimumSwathHeightY = MaximumSwathHeightY; + } else { + MinimumSwathHeightY = MaximumSwathHeightY; + MinimumSwathHeightC = MaximumSwathHeightC; + } + } + + if (mode_lib->vba.SourceScan[k] == dm_horz) { + SwathWidth = mode_lib->vba.ViewportWidth[k]; + } else { + SwathWidth = mode_lib->vba.ViewportHeight[k]; + } + + if (mode_lib->vba.ODMCombineEnabled[k] == true) { + MainPlaneDoesODMCombine = true; + } + for (j = 0; j < mode_lib->vba.NumberOfActivePlanes; ++j) { + if (mode_lib->vba.BlendingAndTiming[k] == j + && mode_lib->vba.ODMCombineEnabled[j] == true) { + MainPlaneDoesODMCombine = true; + } + } + + if (MainPlaneDoesODMCombine == true) { + SwathWidth = dml_min( + SwathWidth, + mode_lib->vba.HActive[k] / 2.0 * mode_lib->vba.HRatio[k]); + } else { + SwathWidth = SwathWidth / mode_lib->vba.DPPPerPlane[k]; + } + + SwathWidthGranularityY = 256 / dml_ceil(BytePerPixDETY, 1) / MaximumSwathHeightY; + RoundedUpMaxSwathSizeBytesY = (dml_ceil( + (double) (SwathWidth - 1), + SwathWidthGranularityY) + SwathWidthGranularityY) * BytePerPixDETY + * MaximumSwathHeightY; + if (mode_lib->vba.SourcePixelFormat[k] == dm_420_10) { + RoundedUpMaxSwathSizeBytesY = dml_ceil(RoundedUpMaxSwathSizeBytesY, 256) + + 256; + } + if (MaximumSwathHeightC > 0) { + SwathWidthGranularityC = 256.0 / dml_ceil(BytePerPixDETC, 2) + / MaximumSwathHeightC; + RoundedUpMaxSwathSizeBytesC = (dml_ceil( + (double) (SwathWidth / 2.0 - 1), + SwathWidthGranularityC) + SwathWidthGranularityC) + * BytePerPixDETC * MaximumSwathHeightC; + if (mode_lib->vba.SourcePixelFormat[k] == dm_420_10) { + RoundedUpMaxSwathSizeBytesC = dml_ceil( + RoundedUpMaxSwathSizeBytesC, + 256) + 256; + } + } else + RoundedUpMaxSwathSizeBytesC = 0.0; + + if (RoundedUpMaxSwathSizeBytesY + RoundedUpMaxSwathSizeBytesC + <= mode_lib->vba.DETBufferSizeInKByte * 1024.0 / 2.0) { + mode_lib->vba.SwathHeightY[k] = MaximumSwathHeightY; + mode_lib->vba.SwathHeightC[k] = MaximumSwathHeightC; + } else { + mode_lib->vba.SwathHeightY[k] = MinimumSwathHeightY; + mode_lib->vba.SwathHeightC[k] = MinimumSwathHeightC; + } + + if (mode_lib->vba.SwathHeightC[k] == 0) { + mode_lib->vba.DETBufferSizeY[k] = mode_lib->vba.DETBufferSizeInKByte * 1024; + mode_lib->vba.DETBufferSizeC[k] = 0; + } else if (mode_lib->vba.SwathHeightY[k] <= mode_lib->vba.SwathHeightC[k]) { + mode_lib->vba.DETBufferSizeY[k] = mode_lib->vba.DETBufferSizeInKByte + * 1024.0 / 2; + mode_lib->vba.DETBufferSizeC[k] = mode_lib->vba.DETBufferSizeInKByte + * 1024.0 / 2; + } else { + mode_lib->vba.DETBufferSizeY[k] = mode_lib->vba.DETBufferSizeInKByte + * 1024.0 * 2 / 3; + mode_lib->vba.DETBufferSizeC[k] = mode_lib->vba.DETBufferSizeInKByte + * 1024.0 / 3; + } + } +} + +bool Calculate256BBlockSizes( + enum source_format_class SourcePixelFormat, + enum dm_swizzle_mode SurfaceTiling, + unsigned int BytePerPixelY, + unsigned int BytePerPixelC, + unsigned int *BlockHeight256BytesY, + unsigned int *BlockHeight256BytesC, + unsigned int *BlockWidth256BytesY, + unsigned int *BlockWidth256BytesC) +{ + if ((SourcePixelFormat == dm_444_64 || SourcePixelFormat == dm_444_32 + || SourcePixelFormat == dm_444_16 + || SourcePixelFormat == dm_444_8)) { + if (SurfaceTiling == dm_sw_linear) { + *BlockHeight256BytesY = 1; + } else if (SourcePixelFormat == dm_444_64) { + *BlockHeight256BytesY = 4; + } else if (SourcePixelFormat == dm_444_8) { + *BlockHeight256BytesY = 16; + } else { + *BlockHeight256BytesY = 8; + } + *BlockWidth256BytesY = 256 / BytePerPixelY / *BlockHeight256BytesY; + *BlockHeight256BytesC = 0; + *BlockWidth256BytesC = 0; + } else { + if (SurfaceTiling == dm_sw_linear) { + *BlockHeight256BytesY = 1; + *BlockHeight256BytesC = 1; + } else if (SourcePixelFormat == dm_420_8) { + *BlockHeight256BytesY = 16; + *BlockHeight256BytesC = 8; + } else { + *BlockHeight256BytesY = 8; + *BlockHeight256BytesC = 8; + } + *BlockWidth256BytesY = 256 / BytePerPixelY / *BlockHeight256BytesY; + *BlockWidth256BytesC = 256 / BytePerPixelC / *BlockHeight256BytesC; + } + return true; +} + +static double CalculateTWait( + unsigned int PrefetchMode, + double DRAMClockChangeLatency, + double UrgentLatency, + double SREnterPlusExitTime) +{ + if (PrefetchMode == 0) { + return dml_max( + DRAMClockChangeLatency + UrgentLatency, + dml_max(SREnterPlusExitTime, UrgentLatency)); + } else if (PrefetchMode == 1) { + return dml_max(SREnterPlusExitTime, UrgentLatency); + } else { + return UrgentLatency; + } +} + +static double CalculateRemoteSurfaceFlipDelay( + struct display_mode_lib *mode_lib, + double VRatio, + double SwathWidth, + double Bpp, + double LineTime, + double XFCTSlvVupdateOffset, + double XFCTSlvVupdateWidth, + double XFCTSlvVreadyOffset, + double XFCXBUFLatencyTolerance, + double XFCFillBWOverhead, + double XFCSlvChunkSize, + double XFCBusTransportTime, + double TCalc, + double TWait, + double *SrcActiveDrainRate, + double *TInitXFill, + double *TslvChk) +{ + double TSlvSetup, AvgfillRate, result; + + *SrcActiveDrainRate = VRatio * SwathWidth * Bpp / LineTime; + TSlvSetup = XFCTSlvVupdateOffset + XFCTSlvVupdateWidth + XFCTSlvVreadyOffset; + *TInitXFill = XFCXBUFLatencyTolerance / (1 + XFCFillBWOverhead / 100); + AvgfillRate = *SrcActiveDrainRate * (1 + XFCFillBWOverhead / 100); + *TslvChk = XFCSlvChunkSize / AvgfillRate; + dml_print( + "DML::CalculateRemoteSurfaceFlipDelay: SrcActiveDrainRate: %f\n", + *SrcActiveDrainRate); + dml_print("DML::CalculateRemoteSurfaceFlipDelay: TSlvSetup: %f\n", TSlvSetup); + dml_print("DML::CalculateRemoteSurfaceFlipDelay: TInitXFill: %f\n", *TInitXFill); + dml_print("DML::CalculateRemoteSurfaceFlipDelay: AvgfillRate: %f\n", AvgfillRate); + dml_print("DML::CalculateRemoteSurfaceFlipDelay: TslvChk: %f\n", *TslvChk); + result = 2 * XFCBusTransportTime + TSlvSetup + TCalc + TWait + *TslvChk + *TInitXFill; // TODO: This doesn't seem to match programming guide + dml_print("DML::CalculateRemoteSurfaceFlipDelay: RemoteSurfaceFlipDelay: %f\n", result); + return result; +} + +static double CalculateWriteBackDISPCLK( + enum source_format_class WritebackPixelFormat, + double PixelClock, + double WritebackHRatio, + double WritebackVRatio, + unsigned int WritebackLumaHTaps, + unsigned int WritebackLumaVTaps, + unsigned int WritebackChromaHTaps, + unsigned int WritebackChromaVTaps, + double WritebackDestinationWidth, + unsigned int HTotal, + unsigned int WritebackChromaLineBufferWidth) +{ + double CalculateWriteBackDISPCLK = + 1.01 * PixelClock + * dml_max( + dml_ceil(WritebackLumaHTaps / 4.0, 1) + / WritebackHRatio, + dml_max( + (WritebackLumaVTaps + * dml_ceil( + 1.0 + / WritebackVRatio, + 1) + * dml_ceil( + WritebackDestinationWidth + / 4.0, + 1) + + dml_ceil( + WritebackDestinationWidth + / 4.0, + 1)) + / (double) HTotal + + dml_ceil( + 1.0 + / WritebackVRatio, + 1) + * (dml_ceil( + WritebackLumaVTaps + / 4.0, + 1) + + 4.0) + / (double) HTotal, + dml_ceil( + 1.0 + / WritebackVRatio, + 1) + * WritebackDestinationWidth + / (double) HTotal)); + if (WritebackPixelFormat != dm_444_32) { + CalculateWriteBackDISPCLK = + dml_max( + CalculateWriteBackDISPCLK, + 1.01 * PixelClock + * dml_max( + dml_ceil( + WritebackChromaHTaps + / 2.0, + 1) + / (2 + * WritebackHRatio), + dml_max( + (WritebackChromaVTaps + * dml_ceil( + 1 + / (2 + * WritebackVRatio), + 1) + * dml_ceil( + WritebackDestinationWidth + / 2.0 + / 2.0, + 1) + + dml_ceil( + WritebackDestinationWidth + / 2.0 + / WritebackChromaLineBufferWidth, + 1)) + / HTotal + + dml_ceil( + 1 + / (2 + * WritebackVRatio), + 1) + * (dml_ceil( + WritebackChromaVTaps + / 4.0, + 1) + + 4) + / HTotal, + dml_ceil( + 1.0 + / (2 + * WritebackVRatio), + 1) + * WritebackDestinationWidth + / 2.0 + / HTotal))); + } + return CalculateWriteBackDISPCLK; +} + +static double CalculateWriteBackDelay( + enum source_format_class WritebackPixelFormat, + double WritebackHRatio, + double WritebackVRatio, + unsigned int WritebackLumaHTaps, + unsigned int WritebackLumaVTaps, + unsigned int WritebackChromaHTaps, + unsigned int WritebackChromaVTaps, + unsigned int WritebackDestinationWidth) +{ + double CalculateWriteBackDelay = + dml_max( + dml_ceil(WritebackLumaHTaps / 4.0, 1) / WritebackHRatio, + WritebackLumaVTaps * dml_ceil(1.0 / WritebackVRatio, 1) + * dml_ceil( + WritebackDestinationWidth + / 4.0, + 1) + + dml_ceil(1.0 / WritebackVRatio, 1) + * (dml_ceil( + WritebackLumaVTaps + / 4.0, + 1) + 4)); + + if (WritebackPixelFormat != dm_444_32) { + CalculateWriteBackDelay = + dml_max( + CalculateWriteBackDelay, + dml_max( + dml_ceil( + WritebackChromaHTaps + / 2.0, + 1) + / (2 + * WritebackHRatio), + WritebackChromaVTaps + * dml_ceil( + 1 + / (2 + * WritebackVRatio), + 1) + * dml_ceil( + WritebackDestinationWidth + / 2.0 + / 2.0, + 1) + + dml_ceil( + 1 + / (2 + * WritebackVRatio), + 1) + * (dml_ceil( + WritebackChromaVTaps + / 4.0, + 1) + + 4))); + } + return CalculateWriteBackDelay; +} + +static void CalculateActiveRowBandwidth( + bool VirtualMemoryEnable, + enum source_format_class SourcePixelFormat, + double VRatio, + bool DCCEnable, + double LineTime, + unsigned int MetaRowByteLuma, + unsigned int MetaRowByteChroma, + unsigned int meta_row_height_luma, + unsigned int meta_row_height_chroma, + unsigned int PixelPTEBytesPerRowLuma, + unsigned int PixelPTEBytesPerRowChroma, + unsigned int dpte_row_height_luma, + unsigned int dpte_row_height_chroma, + double *meta_row_bw, + double *dpte_row_bw, + double *qual_row_bw) +{ + if (DCCEnable != true) { + *meta_row_bw = 0; + } else if (SourcePixelFormat == dm_420_8 || SourcePixelFormat == dm_420_10) { + *meta_row_bw = VRatio * MetaRowByteLuma / (meta_row_height_luma * LineTime) + + VRatio / 2 * MetaRowByteChroma + / (meta_row_height_chroma * LineTime); + } else { + *meta_row_bw = VRatio * MetaRowByteLuma / (meta_row_height_luma * LineTime); + } + + if (VirtualMemoryEnable != true) { + *dpte_row_bw = 0; + } else if (SourcePixelFormat == dm_420_8 || SourcePixelFormat == dm_420_10) { + *dpte_row_bw = VRatio * PixelPTEBytesPerRowLuma / (dpte_row_height_luma * LineTime) + + VRatio / 2 * PixelPTEBytesPerRowChroma + / (dpte_row_height_chroma * LineTime); + } else { + *dpte_row_bw = VRatio * PixelPTEBytesPerRowLuma / (dpte_row_height_luma * LineTime); + } + + if ((SourcePixelFormat == dm_420_8 || SourcePixelFormat == dm_420_10)) { + *qual_row_bw = *meta_row_bw + *dpte_row_bw; + } else { + *qual_row_bw = 0; + } +} + +static void CalculateFlipSchedule( + struct display_mode_lib *mode_lib, + double UrgentExtraLatency, + double UrgentLatency, + unsigned int MaxPageTableLevels, + bool VirtualMemoryEnable, + double BandwidthAvailableForImmediateFlip, + unsigned int TotImmediateFlipBytes, + enum source_format_class SourcePixelFormat, + unsigned int ImmediateFlipBytes, + double LineTime, + double Tno_bw, + double VRatio, + double PDEAndMetaPTEBytesFrame, + unsigned int MetaRowByte, + unsigned int PixelPTEBytesPerRow, + bool DCCEnable, + unsigned int dpte_row_height, + unsigned int meta_row_height, + double qual_row_bw, + double *DestinationLinesToRequestVMInImmediateFlip, + double *DestinationLinesToRequestRowInImmediateFlip, + double *final_flip_bw, + bool *ImmediateFlipSupportedForPipe) +{ + double min_row_time = 0.0; + + if (SourcePixelFormat == dm_420_8 || SourcePixelFormat == dm_420_10) { + *DestinationLinesToRequestVMInImmediateFlip = 0.0; + *DestinationLinesToRequestRowInImmediateFlip = 0.0; + *final_flip_bw = qual_row_bw; + *ImmediateFlipSupportedForPipe = true; + } else { + double TimeForFetchingMetaPTEImmediateFlip; + double TimeForFetchingRowInVBlankImmediateFlip; + + if (VirtualMemoryEnable == true) { + mode_lib->vba.ImmediateFlipBW = BandwidthAvailableForImmediateFlip + * ImmediateFlipBytes / TotImmediateFlipBytes; + TimeForFetchingMetaPTEImmediateFlip = + dml_max( + Tno_bw + + PDEAndMetaPTEBytesFrame + / mode_lib->vba.ImmediateFlipBW, + dml_max( + UrgentExtraLatency + + UrgentLatency + * (MaxPageTableLevels + - 1), + LineTime / 4.0)); + } else { + TimeForFetchingMetaPTEImmediateFlip = 0; + } + + *DestinationLinesToRequestVMInImmediateFlip = dml_floor( + 4.0 * (TimeForFetchingMetaPTEImmediateFlip / LineTime + 0.125), + 1) / 4.0; + + if ((VirtualMemoryEnable == true || DCCEnable == true)) { + mode_lib->vba.ImmediateFlipBW = BandwidthAvailableForImmediateFlip + * ImmediateFlipBytes / TotImmediateFlipBytes; + TimeForFetchingRowInVBlankImmediateFlip = dml_max( + (MetaRowByte + PixelPTEBytesPerRow) + / mode_lib->vba.ImmediateFlipBW, + dml_max(UrgentLatency, LineTime / 4.0)); + } else { + TimeForFetchingRowInVBlankImmediateFlip = 0; + } + + *DestinationLinesToRequestRowInImmediateFlip = dml_floor( + 4.0 * (TimeForFetchingRowInVBlankImmediateFlip / LineTime + 0.125), + 1) / 4.0; + + if (VirtualMemoryEnable == true) { + *final_flip_bw = + dml_max( + PDEAndMetaPTEBytesFrame + / (*DestinationLinesToRequestVMInImmediateFlip + * LineTime), + (MetaRowByte + PixelPTEBytesPerRow) + / (TimeForFetchingRowInVBlankImmediateFlip + * LineTime)); + } else if (MetaRowByte + PixelPTEBytesPerRow > 0) { + *final_flip_bw = (MetaRowByte + PixelPTEBytesPerRow) + / (TimeForFetchingRowInVBlankImmediateFlip * LineTime); + } else { + *final_flip_bw = 0; + } + + if (VirtualMemoryEnable && !DCCEnable) + min_row_time = dpte_row_height * LineTime / VRatio; + else if (!VirtualMemoryEnable && DCCEnable) + min_row_time = meta_row_height * LineTime / VRatio; + else + min_row_time = dml_min(dpte_row_height, meta_row_height) * LineTime + / VRatio; + + if (*DestinationLinesToRequestVMInImmediateFlip >= 8 + || *DestinationLinesToRequestRowInImmediateFlip >= 16 + || TimeForFetchingMetaPTEImmediateFlip + + 2 * TimeForFetchingRowInVBlankImmediateFlip + > min_row_time) + *ImmediateFlipSupportedForPipe = false; + else + *ImmediateFlipSupportedForPipe = true; + } +} + +static void PixelClockAdjustmentForProgressiveToInterlaceUnit(struct display_mode_lib *mode_lib) +{ + unsigned int k; + + //Progressive To dml_ml->vba.Interlace Unit Effect + for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) { + mode_lib->vba.PixelClockBackEnd[k] = mode_lib->vba.PixelClock[k]; + if (mode_lib->vba.Interlace[k] == 1 + && mode_lib->vba.ProgressiveToInterlaceUnitInOPP == true) { + mode_lib->vba.PixelClock[k] = 2 * mode_lib->vba.PixelClock[k]; + } + } +} + +static unsigned int CursorBppEnumToBits(enum cursor_bpp ebpp) +{ + switch (ebpp) { + case dm_cur_2bit: + return 2; + case dm_cur_32bit: + return 32; + case dm_cur_64bit: + return 64; + default: + return 0; + } +} + +static unsigned int TruncToValidBPP( + double DecimalBPP, + bool DSCEnabled, + enum output_encoder_class Output, + enum output_format_class Format, + unsigned int DSCInputBitPerComponent) +{ + if (Output == dm_hdmi) { + if (Format == dm_420) { + if (DecimalBPP >= 18) + return 18; + else if (DecimalBPP >= 15) + return 15; + else if (DecimalBPP >= 12) + return 12; + else + return 0; + } else if (Format == dm_444) { + if (DecimalBPP >= 36) + return 36; + else if (DecimalBPP >= 30) + return 30; + else if (DecimalBPP >= 24) + return 24; + else + return 0; + } else { + if (DecimalBPP / 1.5 >= 24) + return 24; + else if (DecimalBPP / 1.5 >= 20) + return 20; + else if (DecimalBPP / 1.5 >= 16) + return 16; + else + return 0; + } + } else { + if (DSCEnabled) { + if (Format == dm_420) { + if (DecimalBPP < 6) + return 0; + else if (DecimalBPP >= 1.5 * DSCInputBitPerComponent - 1 / 16) + return 1.5 * DSCInputBitPerComponent - 1 / 16; + else + return dml_floor(16 * DecimalBPP, 1) / 16; + } else if (Format == dm_n422) { + if (DecimalBPP < 7) + return 0; + else if (DecimalBPP >= 2 * DSCInputBitPerComponent - 1 / 16) + return 2 * DSCInputBitPerComponent - 1 / 16; + else + return dml_floor(16 * DecimalBPP, 1) / 16; + } else { + if (DecimalBPP < 8) + return 0; + else if (DecimalBPP >= 3 * DSCInputBitPerComponent - 1 / 16) + return 3 * DSCInputBitPerComponent - 1 / 16; + else + return dml_floor(16 * DecimalBPP, 1) / 16; + } + } else if (Format == dm_420) { + if (DecimalBPP >= 18) + return 18; + else if (DecimalBPP >= 15) + return 15; + else if (DecimalBPP >= 12) + return 12; + else + return 0; + } else if (Format == dm_s422 || Format == dm_n422) { + if (DecimalBPP >= 24) + return 24; + else if (DecimalBPP >= 20) + return 20; + else if (DecimalBPP >= 16) + return 16; + else + return 0; + } else { + if (DecimalBPP >= 36) + return 36; + else if (DecimalBPP >= 30) + return 30; + else if (DecimalBPP >= 24) + return 24; + else + return 0; + } + } +} + +static void ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_lib) +{ + int i; + unsigned int j, k; + /*MODE SUPPORT, VOLTAGE STATE AND SOC CONFIGURATION*/ + + /*Scale Ratio, taps Support Check*/ + + mode_lib->vba.ScaleRatioAndTapsSupport = true; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + if (mode_lib->vba.ScalerEnabled[k] == false + && ((mode_lib->vba.SourcePixelFormat[k] != dm_444_64 + && mode_lib->vba.SourcePixelFormat[k] != dm_444_32 + && mode_lib->vba.SourcePixelFormat[k] != dm_444_16 + && mode_lib->vba.SourcePixelFormat[k] != dm_mono_16 + && mode_lib->vba.SourcePixelFormat[k] != dm_mono_8) + || mode_lib->vba.HRatio[k] != 1.0 + || mode_lib->vba.htaps[k] != 1.0 + || mode_lib->vba.VRatio[k] != 1.0 + || mode_lib->vba.vtaps[k] != 1.0)) { + mode_lib->vba.ScaleRatioAndTapsSupport = false; + } else if (mode_lib->vba.vtaps[k] < 1.0 || mode_lib->vba.vtaps[k] > 8.0 + || mode_lib->vba.htaps[k] < 1.0 || mode_lib->vba.htaps[k] > 8.0 + || (mode_lib->vba.htaps[k] > 1.0 + && (mode_lib->vba.htaps[k] % 2) == 1) + || mode_lib->vba.HRatio[k] > mode_lib->vba.MaxHSCLRatio + || mode_lib->vba.VRatio[k] > mode_lib->vba.MaxVSCLRatio + || mode_lib->vba.HRatio[k] > mode_lib->vba.htaps[k] + || mode_lib->vba.VRatio[k] > mode_lib->vba.vtaps[k] + || (mode_lib->vba.SourcePixelFormat[k] != dm_444_64 + && mode_lib->vba.SourcePixelFormat[k] != dm_444_32 + && mode_lib->vba.SourcePixelFormat[k] != dm_444_16 + && mode_lib->vba.SourcePixelFormat[k] != dm_mono_16 + && mode_lib->vba.SourcePixelFormat[k] != dm_mono_8 + && (mode_lib->vba.HRatio[k] / 2.0 + > mode_lib->vba.HTAPsChroma[k] + || mode_lib->vba.VRatio[k] / 2.0 + > mode_lib->vba.VTAPsChroma[k]))) { + mode_lib->vba.ScaleRatioAndTapsSupport = false; + } + } + /*Source Format, Pixel Format and Scan Support Check*/ + + mode_lib->vba.SourceFormatPixelAndScanSupport = true; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + if ((mode_lib->vba.SurfaceTiling[k] == dm_sw_linear + && mode_lib->vba.SourceScan[k] != dm_horz) + || ((mode_lib->vba.SurfaceTiling[k] == dm_sw_4kb_d + || mode_lib->vba.SurfaceTiling[k] == dm_sw_4kb_d_x + || mode_lib->vba.SurfaceTiling[k] == dm_sw_64kb_d + || mode_lib->vba.SurfaceTiling[k] == dm_sw_64kb_d_t + || mode_lib->vba.SurfaceTiling[k] == dm_sw_64kb_d_x + || mode_lib->vba.SurfaceTiling[k] == dm_sw_var_d + || mode_lib->vba.SurfaceTiling[k] == dm_sw_var_d_x) + && mode_lib->vba.SourcePixelFormat[k] != dm_444_64) + || (mode_lib->vba.SurfaceTiling[k] == dm_sw_64kb_r_x + && (mode_lib->vba.SourcePixelFormat[k] == dm_mono_8 + || mode_lib->vba.SourcePixelFormat[k] + == dm_420_8 + || mode_lib->vba.SourcePixelFormat[k] + == dm_420_10)) + || (((mode_lib->vba.SurfaceTiling[k] + == dm_sw_gfx7_2d_thin_gl + || mode_lib->vba.SurfaceTiling[k] + == dm_sw_gfx7_2d_thin_lvp) + && !((mode_lib->vba.SourcePixelFormat[k] + == dm_444_64 + || mode_lib->vba.SourcePixelFormat[k] + == dm_444_32) + && mode_lib->vba.SourceScan[k] + == dm_horz + && mode_lib->vba.SupportGFX7CompatibleTilingIn32bppAnd64bpp + == true + && mode_lib->vba.DCCEnable[k] + == false)) + || (mode_lib->vba.DCCEnable[k] == true + && (mode_lib->vba.SurfaceTiling[k] + == dm_sw_linear + || mode_lib->vba.SourcePixelFormat[k] + == dm_420_8 + || mode_lib->vba.SourcePixelFormat[k] + == dm_420_10)))) { + mode_lib->vba.SourceFormatPixelAndScanSupport = false; + } + } + /*Bandwidth Support Check*/ + + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + if (mode_lib->vba.SourceScan[k] == dm_horz) { + mode_lib->vba.SwathWidthYSingleDPP[k] = mode_lib->vba.ViewportWidth[k]; + } else { + mode_lib->vba.SwathWidthYSingleDPP[k] = mode_lib->vba.ViewportHeight[k]; + } + if (mode_lib->vba.SourcePixelFormat[k] == dm_444_64) { + mode_lib->vba.BytePerPixelInDETY[k] = 8.0; + mode_lib->vba.BytePerPixelInDETC[k] = 0.0; + } else if (mode_lib->vba.SourcePixelFormat[k] == dm_444_32) { + mode_lib->vba.BytePerPixelInDETY[k] = 4.0; + mode_lib->vba.BytePerPixelInDETC[k] = 0.0; + } else if (mode_lib->vba.SourcePixelFormat[k] == dm_444_16 + || mode_lib->vba.SourcePixelFormat[k] == dm_mono_16) { + mode_lib->vba.BytePerPixelInDETY[k] = 2.0; + mode_lib->vba.BytePerPixelInDETC[k] = 0.0; + } else if (mode_lib->vba.SourcePixelFormat[k] == dm_mono_8) { + mode_lib->vba.BytePerPixelInDETY[k] = 1.0; + mode_lib->vba.BytePerPixelInDETC[k] = 0.0; + } else if (mode_lib->vba.SourcePixelFormat[k] == dm_420_8) { + mode_lib->vba.BytePerPixelInDETY[k] = 1.0; + mode_lib->vba.BytePerPixelInDETC[k] = 2.0; + } else { + mode_lib->vba.BytePerPixelInDETY[k] = 4.0 / 3; + mode_lib->vba.BytePerPixelInDETC[k] = 8.0 / 3; + } + } + mode_lib->vba.TotalReadBandwidthConsumedGBytePerSecond = 0.0; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + mode_lib->vba.ReadBandwidth[k] = mode_lib->vba.SwathWidthYSingleDPP[k] + * (dml_ceil(mode_lib->vba.BytePerPixelInDETY[k], 1.0) + * mode_lib->vba.VRatio[k] + + dml_ceil(mode_lib->vba.BytePerPixelInDETC[k], 2.0) + / 2.0 * mode_lib->vba.VRatio[k] / 2) + / (mode_lib->vba.HTotal[k] / mode_lib->vba.PixelClock[k]); + if (mode_lib->vba.DCCEnable[k] == true) { + mode_lib->vba.ReadBandwidth[k] = mode_lib->vba.ReadBandwidth[k] + * (1 + 1 / 256); + } + if (mode_lib->vba.VirtualMemoryEnable == true + && mode_lib->vba.SourceScan[k] != dm_horz + && (mode_lib->vba.SurfaceTiling[k] == dm_sw_4kb_s + || mode_lib->vba.SurfaceTiling[k] == dm_sw_4kb_s_x + || mode_lib->vba.SurfaceTiling[k] == dm_sw_4kb_d + || mode_lib->vba.SurfaceTiling[k] == dm_sw_4kb_d_x)) { + mode_lib->vba.ReadBandwidth[k] = mode_lib->vba.ReadBandwidth[k] + * (1 + 1 / 64); + } else if (mode_lib->vba.VirtualMemoryEnable == true + && mode_lib->vba.SourceScan[k] == dm_horz + && (mode_lib->vba.SourcePixelFormat[k] == dm_444_64 + || mode_lib->vba.SourcePixelFormat[k] == dm_444_32) + && (mode_lib->vba.SurfaceTiling[k] == dm_sw_64kb_s + || mode_lib->vba.SurfaceTiling[k] == dm_sw_64kb_s_t + || mode_lib->vba.SurfaceTiling[k] == dm_sw_64kb_s_x + || mode_lib->vba.SurfaceTiling[k] == dm_sw_64kb_d + || mode_lib->vba.SurfaceTiling[k] == dm_sw_64kb_d_t + || mode_lib->vba.SurfaceTiling[k] == dm_sw_64kb_d_x + || mode_lib->vba.SurfaceTiling[k] == dm_sw_64kb_r_x)) { + mode_lib->vba.ReadBandwidth[k] = mode_lib->vba.ReadBandwidth[k] + * (1 + 1 / 256); + } else if (mode_lib->vba.VirtualMemoryEnable == true) { + mode_lib->vba.ReadBandwidth[k] = mode_lib->vba.ReadBandwidth[k] + * (1 + 1 / 512); + } + mode_lib->vba.TotalReadBandwidthConsumedGBytePerSecond = + mode_lib->vba.TotalReadBandwidthConsumedGBytePerSecond + + mode_lib->vba.ReadBandwidth[k] / 1000.0; + } + mode_lib->vba.TotalWriteBandwidthConsumedGBytePerSecond = 0.0; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + if (mode_lib->vba.WritebackEnable[k] == true + && mode_lib->vba.WritebackPixelFormat[k] == dm_444_32) { + mode_lib->vba.WriteBandwidth[k] = mode_lib->vba.WritebackDestinationWidth[k] + * mode_lib->vba.WritebackDestinationHeight[k] + / (mode_lib->vba.WritebackSourceHeight[k] + * mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k]) * 4.0; + } else if (mode_lib->vba.WritebackEnable[k] == true + && mode_lib->vba.WritebackPixelFormat[k] == dm_420_10) { + mode_lib->vba.WriteBandwidth[k] = mode_lib->vba.WritebackDestinationWidth[k] + * mode_lib->vba.WritebackDestinationHeight[k] + / (mode_lib->vba.WritebackSourceHeight[k] + * mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k]) * 3.0; + } else if (mode_lib->vba.WritebackEnable[k] == true) { + mode_lib->vba.WriteBandwidth[k] = mode_lib->vba.WritebackDestinationWidth[k] + * mode_lib->vba.WritebackDestinationHeight[k] + / (mode_lib->vba.WritebackSourceHeight[k] + * mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k]) * 1.5; + } else { + mode_lib->vba.WriteBandwidth[k] = 0.0; + } + mode_lib->vba.TotalWriteBandwidthConsumedGBytePerSecond = + mode_lib->vba.TotalWriteBandwidthConsumedGBytePerSecond + + mode_lib->vba.WriteBandwidth[k] / 1000.0; + } + mode_lib->vba.TotalBandwidthConsumedGBytePerSecond = + mode_lib->vba.TotalReadBandwidthConsumedGBytePerSecond + + mode_lib->vba.TotalWriteBandwidthConsumedGBytePerSecond; + mode_lib->vba.DCCEnabledInAnyPlane = false; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + if (mode_lib->vba.DCCEnable[k] == true) { + mode_lib->vba.DCCEnabledInAnyPlane = true; + } + } + for (i = 0; i <= DC__VOLTAGE_STATES; i++) { + mode_lib->vba.FabricAndDRAMBandwidthPerState[i] = dml_min( + mode_lib->vba.DRAMSpeedPerState[i] * mode_lib->vba.NumberOfChannels + * mode_lib->vba.DRAMChannelWidth, + mode_lib->vba.FabricClockPerState[i] + * mode_lib->vba.FabricDatapathToDCNDataReturn) + / 1000; + mode_lib->vba.ReturnBWToDCNPerState = dml_min( + mode_lib->vba.ReturnBusWidth * mode_lib->vba.DCFCLKPerState[i], + mode_lib->vba.FabricAndDRAMBandwidthPerState[i] * 1000.0) + * mode_lib->vba.PercentOfIdealDRAMAndFabricBWReceivedAfterUrgLatency + / 100; + mode_lib->vba.ReturnBWPerState[i] = mode_lib->vba.ReturnBWToDCNPerState; + if (mode_lib->vba.DCCEnabledInAnyPlane == true + && mode_lib->vba.ReturnBWToDCNPerState + > mode_lib->vba.DCFCLKPerState[i] + * mode_lib->vba.ReturnBusWidth + / 4.0) { + mode_lib->vba.ReturnBWPerState[i] = + dml_min( + mode_lib->vba.ReturnBWPerState[i], + mode_lib->vba.ReturnBWToDCNPerState * 4.0 + * (1.0 + - mode_lib->vba.UrgentLatency + / ((mode_lib->vba.ROBBufferSizeInKByte + - mode_lib->vba.PixelChunkSizeInKByte) + * 1024.0 + / (mode_lib->vba.ReturnBWToDCNPerState + - mode_lib->vba.DCFCLKPerState[i] + * mode_lib->vba.ReturnBusWidth + / 4.0) + + mode_lib->vba.UrgentLatency))); + } + mode_lib->vba.CriticalPoint = + 2.0 * mode_lib->vba.ReturnBusWidth * mode_lib->vba.DCFCLKPerState[i] + * mode_lib->vba.UrgentLatency + / (mode_lib->vba.ReturnBWToDCNPerState + * mode_lib->vba.UrgentLatency + + (mode_lib->vba.ROBBufferSizeInKByte + - mode_lib->vba.PixelChunkSizeInKByte) + * 1024.0); + if (mode_lib->vba.DCCEnabledInAnyPlane == true && mode_lib->vba.CriticalPoint > 1.0 + && mode_lib->vba.CriticalPoint < 4.0) { + mode_lib->vba.ReturnBWPerState[i] = + dml_min( + mode_lib->vba.ReturnBWPerState[i], + dml_pow( + 4.0 + * mode_lib->vba.ReturnBWToDCNPerState + * (mode_lib->vba.ROBBufferSizeInKByte + - mode_lib->vba.PixelChunkSizeInKByte) + * 1024.0 + * mode_lib->vba.ReturnBusWidth + * mode_lib->vba.DCFCLKPerState[i] + * mode_lib->vba.UrgentLatency + / (mode_lib->vba.ReturnBWToDCNPerState + * mode_lib->vba.UrgentLatency + + (mode_lib->vba.ROBBufferSizeInKByte + - mode_lib->vba.PixelChunkSizeInKByte) + * 1024.0), + 2)); + } + mode_lib->vba.ReturnBWToDCNPerState = dml_min( + mode_lib->vba.ReturnBusWidth * mode_lib->vba.DCFCLKPerState[i], + mode_lib->vba.FabricAndDRAMBandwidthPerState[i] * 1000.0); + if (mode_lib->vba.DCCEnabledInAnyPlane == true + && mode_lib->vba.ReturnBWToDCNPerState + > mode_lib->vba.DCFCLKPerState[i] + * mode_lib->vba.ReturnBusWidth + / 4.0) { + mode_lib->vba.ReturnBWPerState[i] = + dml_min( + mode_lib->vba.ReturnBWPerState[i], + mode_lib->vba.ReturnBWToDCNPerState * 4.0 + * (1.0 + - mode_lib->vba.UrgentLatency + / ((mode_lib->vba.ROBBufferSizeInKByte + - mode_lib->vba.PixelChunkSizeInKByte) + * 1024.0 + / (mode_lib->vba.ReturnBWToDCNPerState + - mode_lib->vba.DCFCLKPerState[i] + * mode_lib->vba.ReturnBusWidth + / 4.0) + + mode_lib->vba.UrgentLatency))); + } + mode_lib->vba.CriticalPoint = + 2.0 * mode_lib->vba.ReturnBusWidth * mode_lib->vba.DCFCLKPerState[i] + * mode_lib->vba.UrgentLatency + / (mode_lib->vba.ReturnBWToDCNPerState + * mode_lib->vba.UrgentLatency + + (mode_lib->vba.ROBBufferSizeInKByte + - mode_lib->vba.PixelChunkSizeInKByte) + * 1024.0); + if (mode_lib->vba.DCCEnabledInAnyPlane == true && mode_lib->vba.CriticalPoint > 1.0 + && mode_lib->vba.CriticalPoint < 4.0) { + mode_lib->vba.ReturnBWPerState[i] = + dml_min( + mode_lib->vba.ReturnBWPerState[i], + dml_pow( + 4.0 + * mode_lib->vba.ReturnBWToDCNPerState + * (mode_lib->vba.ROBBufferSizeInKByte + - mode_lib->vba.PixelChunkSizeInKByte) + * 1024.0 + * mode_lib->vba.ReturnBusWidth + * mode_lib->vba.DCFCLKPerState[i] + * mode_lib->vba.UrgentLatency + / (mode_lib->vba.ReturnBWToDCNPerState + * mode_lib->vba.UrgentLatency + + (mode_lib->vba.ROBBufferSizeInKByte + - mode_lib->vba.PixelChunkSizeInKByte) + * 1024.0), + 2)); + } + } + for (i = 0; i <= DC__VOLTAGE_STATES; i++) { + if ((mode_lib->vba.TotalReadBandwidthConsumedGBytePerSecond * 1000.0 + <= mode_lib->vba.ReturnBWPerState[i]) + && (mode_lib->vba.TotalBandwidthConsumedGBytePerSecond * 1000.0 + <= mode_lib->vba.FabricAndDRAMBandwidthPerState[i] + * 1000.0 + * mode_lib->vba.PercentOfIdealDRAMAndFabricBWReceivedAfterUrgLatency + / 100.0)) { + mode_lib->vba.BandwidthSupport[i] = true; + } else { + mode_lib->vba.BandwidthSupport[i] = false; + } + } + /*Writeback Latency support check*/ + + mode_lib->vba.WritebackLatencySupport = true; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + if (mode_lib->vba.WritebackEnable[k] == true) { + if (mode_lib->vba.WritebackPixelFormat[k] == dm_444_32) { + if (mode_lib->vba.WriteBandwidth[k] + > (mode_lib->vba.WritebackInterfaceLumaBufferSize + + mode_lib->vba.WritebackInterfaceChromaBufferSize) + / mode_lib->vba.WritebackLatency) { + mode_lib->vba.WritebackLatencySupport = false; + } + } else { + if (mode_lib->vba.WriteBandwidth[k] + > 1.5 + * dml_min( + mode_lib->vba.WritebackInterfaceLumaBufferSize, + 2.0 + * mode_lib->vba.WritebackInterfaceChromaBufferSize) + / mode_lib->vba.WritebackLatency) { + mode_lib->vba.WritebackLatencySupport = false; + } + } + } + } + /*Re-ordering Buffer Support Check*/ + + for (i = 0; i <= DC__VOLTAGE_STATES; i++) { + mode_lib->vba.UrgentRoundTripAndOutOfOrderLatencyPerState[i] = + (mode_lib->vba.RoundTripPingLatencyCycles + 32.0) + / mode_lib->vba.DCFCLKPerState[i] + + mode_lib->vba.UrgentOutOfOrderReturnPerChannel + * mode_lib->vba.NumberOfChannels + / mode_lib->vba.ReturnBWPerState[i]; + if ((mode_lib->vba.ROBBufferSizeInKByte - mode_lib->vba.PixelChunkSizeInKByte) + * 1024.0 / mode_lib->vba.ReturnBWPerState[i] + > mode_lib->vba.UrgentRoundTripAndOutOfOrderLatencyPerState[i]) { + mode_lib->vba.ROBSupport[i] = true; + } else { + mode_lib->vba.ROBSupport[i] = false; + } + } + /*Writeback Mode Support Check*/ + + mode_lib->vba.TotalNumberOfActiveWriteback = 0; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + if (mode_lib->vba.WritebackEnable[k] == true) { + mode_lib->vba.TotalNumberOfActiveWriteback = + mode_lib->vba.TotalNumberOfActiveWriteback + 1; + } + } + mode_lib->vba.WritebackModeSupport = true; + if (mode_lib->vba.TotalNumberOfActiveWriteback > mode_lib->vba.MaxNumWriteback) { + mode_lib->vba.WritebackModeSupport = false; + } + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + if (mode_lib->vba.WritebackEnable[k] == true + && mode_lib->vba.Writeback10bpc420Supported != true + && mode_lib->vba.WritebackPixelFormat[k] == dm_420_10) { + mode_lib->vba.WritebackModeSupport = false; + } + } + /*Writeback Scale Ratio and Taps Support Check*/ + + mode_lib->vba.WritebackScaleRatioAndTapsSupport = true; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + if (mode_lib->vba.WritebackEnable[k] == true) { + if (mode_lib->vba.WritebackLumaAndChromaScalingSupported == false + && (mode_lib->vba.WritebackHRatio[k] != 1.0 + || mode_lib->vba.WritebackVRatio[k] != 1.0)) { + mode_lib->vba.WritebackScaleRatioAndTapsSupport = false; + } + if (mode_lib->vba.WritebackHRatio[k] > mode_lib->vba.WritebackMaxHSCLRatio + || mode_lib->vba.WritebackVRatio[k] + > mode_lib->vba.WritebackMaxVSCLRatio + || mode_lib->vba.WritebackHRatio[k] + < mode_lib->vba.WritebackMinHSCLRatio + || mode_lib->vba.WritebackVRatio[k] + < mode_lib->vba.WritebackMinVSCLRatio + || mode_lib->vba.WritebackLumaHTaps[k] + > mode_lib->vba.WritebackMaxHSCLTaps + || mode_lib->vba.WritebackLumaVTaps[k] + > mode_lib->vba.WritebackMaxVSCLTaps + || mode_lib->vba.WritebackHRatio[k] + > mode_lib->vba.WritebackLumaHTaps[k] + || mode_lib->vba.WritebackVRatio[k] + > mode_lib->vba.WritebackLumaVTaps[k] + || (mode_lib->vba.WritebackLumaHTaps[k] > 2.0 + && ((mode_lib->vba.WritebackLumaHTaps[k] % 2) + == 1)) + || (mode_lib->vba.WritebackPixelFormat[k] != dm_444_32 + && (mode_lib->vba.WritebackChromaHTaps[k] + > mode_lib->vba.WritebackMaxHSCLTaps + || mode_lib->vba.WritebackChromaVTaps[k] + > mode_lib->vba.WritebackMaxVSCLTaps + || 2.0 + * mode_lib->vba.WritebackHRatio[k] + > mode_lib->vba.WritebackChromaHTaps[k] + || 2.0 + * mode_lib->vba.WritebackVRatio[k] + > mode_lib->vba.WritebackChromaVTaps[k] + || (mode_lib->vba.WritebackChromaHTaps[k] > 2.0 + && ((mode_lib->vba.WritebackChromaHTaps[k] % 2) == 1))))) { + mode_lib->vba.WritebackScaleRatioAndTapsSupport = false; + } + if (mode_lib->vba.WritebackVRatio[k] < 1.0) { + mode_lib->vba.WritebackLumaVExtra = + dml_max(1.0 - 2.0 / dml_ceil(1.0 / mode_lib->vba.WritebackVRatio[k], 1.0), 0.0); + } else { + mode_lib->vba.WritebackLumaVExtra = -1; + } + if ((mode_lib->vba.WritebackPixelFormat[k] == dm_444_32 + && mode_lib->vba.WritebackLumaVTaps[k] + > (mode_lib->vba.WritebackLineBufferLumaBufferSize + + mode_lib->vba.WritebackLineBufferChromaBufferSize) + / 3.0 + / mode_lib->vba.WritebackDestinationWidth[k] + - mode_lib->vba.WritebackLumaVExtra) + || (mode_lib->vba.WritebackPixelFormat[k] == dm_420_8 + && mode_lib->vba.WritebackLumaVTaps[k] + > mode_lib->vba.WritebackLineBufferLumaBufferSize + / mode_lib->vba.WritebackDestinationWidth[k] + - mode_lib->vba.WritebackLumaVExtra) + || (mode_lib->vba.WritebackPixelFormat[k] == dm_420_10 + && mode_lib->vba.WritebackLumaVTaps[k] + > mode_lib->vba.WritebackLineBufferLumaBufferSize + * 8.0 / 10.0 + / mode_lib->vba.WritebackDestinationWidth[k] + - mode_lib->vba.WritebackLumaVExtra)) { + mode_lib->vba.WritebackScaleRatioAndTapsSupport = false; + } + if (2.0 * mode_lib->vba.WritebackVRatio[k] < 1) { + mode_lib->vba.WritebackChromaVExtra = 0.0; + } else { + mode_lib->vba.WritebackChromaVExtra = -1; + } + if ((mode_lib->vba.WritebackPixelFormat[k] == dm_420_8 + && mode_lib->vba.WritebackChromaVTaps[k] + > mode_lib->vba.WritebackLineBufferChromaBufferSize + / mode_lib->vba.WritebackDestinationWidth[k] + - mode_lib->vba.WritebackChromaVExtra) + || (mode_lib->vba.WritebackPixelFormat[k] == dm_420_10 + && mode_lib->vba.WritebackChromaVTaps[k] + > mode_lib->vba.WritebackLineBufferChromaBufferSize + * 8.0 / 10.0 + / mode_lib->vba.WritebackDestinationWidth[k] + - mode_lib->vba.WritebackChromaVExtra)) { + mode_lib->vba.WritebackScaleRatioAndTapsSupport = false; + } + } + } + /*Maximum DISPCLK/DPPCLK Support check*/ + + mode_lib->vba.WritebackRequiredDISPCLK = 0.0; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + if (mode_lib->vba.WritebackEnable[k] == true) { + mode_lib->vba.WritebackRequiredDISPCLK = + dml_max( + mode_lib->vba.WritebackRequiredDISPCLK, + CalculateWriteBackDISPCLK( + mode_lib->vba.WritebackPixelFormat[k], + mode_lib->vba.PixelClock[k], + mode_lib->vba.WritebackHRatio[k], + mode_lib->vba.WritebackVRatio[k], + mode_lib->vba.WritebackLumaHTaps[k], + mode_lib->vba.WritebackLumaVTaps[k], + mode_lib->vba.WritebackChromaHTaps[k], + mode_lib->vba.WritebackChromaVTaps[k], + mode_lib->vba.WritebackDestinationWidth[k], + mode_lib->vba.HTotal[k], + mode_lib->vba.WritebackChromaLineBufferWidth)); + } + } + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + if (mode_lib->vba.HRatio[k] > 1.0) { + mode_lib->vba.PSCL_FACTOR[k] = dml_min( + mode_lib->vba.MaxDCHUBToPSCLThroughput, + mode_lib->vba.MaxPSCLToLBThroughput + * mode_lib->vba.HRatio[k] + / dml_ceil( + mode_lib->vba.htaps[k] + / 6.0, + 1.0)); + } else { + mode_lib->vba.PSCL_FACTOR[k] = dml_min( + mode_lib->vba.MaxDCHUBToPSCLThroughput, + mode_lib->vba.MaxPSCLToLBThroughput); + } + if (mode_lib->vba.BytePerPixelInDETC[k] == 0.0) { + mode_lib->vba.PSCL_FACTOR_CHROMA[k] = 0.0; + mode_lib->vba.MinDPPCLKUsingSingleDPP[k] = + mode_lib->vba.PixelClock[k] + * dml_max3( + mode_lib->vba.vtaps[k] / 6.0 + * dml_min( + 1.0, + mode_lib->vba.HRatio[k]), + mode_lib->vba.HRatio[k] + * mode_lib->vba.VRatio[k] + / mode_lib->vba.PSCL_FACTOR[k], + 1.0); + if ((mode_lib->vba.htaps[k] > 6.0 || mode_lib->vba.vtaps[k] > 6.0) + && mode_lib->vba.MinDPPCLKUsingSingleDPP[k] + < 2.0 * mode_lib->vba.PixelClock[k]) { + mode_lib->vba.MinDPPCLKUsingSingleDPP[k] = 2.0 + * mode_lib->vba.PixelClock[k]; + } + } else { + if (mode_lib->vba.HRatio[k] / 2.0 > 1.0) { + mode_lib->vba.PSCL_FACTOR_CHROMA[k] = + dml_min( + mode_lib->vba.MaxDCHUBToPSCLThroughput, + mode_lib->vba.MaxPSCLToLBThroughput + * mode_lib->vba.HRatio[k] + / 2.0 + / dml_ceil( + mode_lib->vba.HTAPsChroma[k] + / 6.0, + 1.0)); + } else { + mode_lib->vba.PSCL_FACTOR_CHROMA[k] = dml_min( + mode_lib->vba.MaxDCHUBToPSCLThroughput, + mode_lib->vba.MaxPSCLToLBThroughput); + } + mode_lib->vba.MinDPPCLKUsingSingleDPP[k] = + mode_lib->vba.PixelClock[k] + * dml_max5( + mode_lib->vba.vtaps[k] / 6.0 + * dml_min( + 1.0, + mode_lib->vba.HRatio[k]), + mode_lib->vba.HRatio[k] + * mode_lib->vba.VRatio[k] + / mode_lib->vba.PSCL_FACTOR[k], + mode_lib->vba.VTAPsChroma[k] + / 6.0 + * dml_min( + 1.0, + mode_lib->vba.HRatio[k] + / 2.0), + mode_lib->vba.HRatio[k] + * mode_lib->vba.VRatio[k] + / 4.0 + / mode_lib->vba.PSCL_FACTOR_CHROMA[k], + 1.0); + if ((mode_lib->vba.htaps[k] > 6.0 || mode_lib->vba.vtaps[k] > 6.0 + || mode_lib->vba.HTAPsChroma[k] > 6.0 + || mode_lib->vba.VTAPsChroma[k] > 6.0) + && mode_lib->vba.MinDPPCLKUsingSingleDPP[k] + < 2.0 * mode_lib->vba.PixelClock[k]) { + mode_lib->vba.MinDPPCLKUsingSingleDPP[k] = 2.0 + * mode_lib->vba.PixelClock[k]; + } + } + } + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + Calculate256BBlockSizes( + mode_lib->vba.SourcePixelFormat[k], + mode_lib->vba.SurfaceTiling[k], + dml_ceil(mode_lib->vba.BytePerPixelInDETY[k], 1.0), + dml_ceil(mode_lib->vba.BytePerPixelInDETC[k], 2.0), + &mode_lib->vba.Read256BlockHeightY[k], + &mode_lib->vba.Read256BlockHeightC[k], + &mode_lib->vba.Read256BlockWidthY[k], + &mode_lib->vba.Read256BlockWidthC[k]); + if (mode_lib->vba.SourceScan[k] == dm_horz) { + mode_lib->vba.MaxSwathHeightY[k] = mode_lib->vba.Read256BlockHeightY[k]; + mode_lib->vba.MaxSwathHeightC[k] = mode_lib->vba.Read256BlockHeightC[k]; + } else { + mode_lib->vba.MaxSwathHeightY[k] = mode_lib->vba.Read256BlockWidthY[k]; + mode_lib->vba.MaxSwathHeightC[k] = mode_lib->vba.Read256BlockWidthC[k]; + } + if ((mode_lib->vba.SourcePixelFormat[k] == dm_444_64 + || mode_lib->vba.SourcePixelFormat[k] == dm_444_32 + || mode_lib->vba.SourcePixelFormat[k] == dm_444_16 + || mode_lib->vba.SourcePixelFormat[k] == dm_mono_16 + || mode_lib->vba.SourcePixelFormat[k] == dm_mono_8)) { + if (mode_lib->vba.SurfaceTiling[k] == dm_sw_linear + || (mode_lib->vba.SourcePixelFormat[k] == dm_444_64 + && (mode_lib->vba.SurfaceTiling[k] + == dm_sw_4kb_s + || mode_lib->vba.SurfaceTiling[k] + == dm_sw_4kb_s_x + || mode_lib->vba.SurfaceTiling[k] + == dm_sw_64kb_s + || mode_lib->vba.SurfaceTiling[k] + == dm_sw_64kb_s_t + || mode_lib->vba.SurfaceTiling[k] + == dm_sw_64kb_s_x + || mode_lib->vba.SurfaceTiling[k] + == dm_sw_var_s + || mode_lib->vba.SurfaceTiling[k] + == dm_sw_var_s_x) + && mode_lib->vba.SourceScan[k] == dm_horz)) { + mode_lib->vba.MinSwathHeightY[k] = mode_lib->vba.MaxSwathHeightY[k]; + } else { + mode_lib->vba.MinSwathHeightY[k] = mode_lib->vba.MaxSwathHeightY[k] + / 2.0; + } + mode_lib->vba.MinSwathHeightC[k] = mode_lib->vba.MaxSwathHeightC[k]; + } else { + if (mode_lib->vba.SurfaceTiling[k] == dm_sw_linear) { + mode_lib->vba.MinSwathHeightY[k] = mode_lib->vba.MaxSwathHeightY[k]; + mode_lib->vba.MinSwathHeightC[k] = mode_lib->vba.MaxSwathHeightC[k]; + } else if (mode_lib->vba.SourcePixelFormat[k] == dm_420_8 + && mode_lib->vba.SourceScan[k] == dm_horz) { + mode_lib->vba.MinSwathHeightY[k] = mode_lib->vba.MaxSwathHeightY[k] + / 2.0; + mode_lib->vba.MinSwathHeightC[k] = mode_lib->vba.MaxSwathHeightC[k]; + } else if (mode_lib->vba.SourcePixelFormat[k] == dm_420_10 + && mode_lib->vba.SourceScan[k] == dm_horz) { + mode_lib->vba.MinSwathHeightC[k] = mode_lib->vba.MaxSwathHeightC[k] + / 2.0; + mode_lib->vba.MinSwathHeightY[k] = mode_lib->vba.MaxSwathHeightY[k]; + } else { + mode_lib->vba.MinSwathHeightY[k] = mode_lib->vba.MaxSwathHeightY[k]; + mode_lib->vba.MinSwathHeightC[k] = mode_lib->vba.MaxSwathHeightC[k]; + } + } + if (mode_lib->vba.SurfaceTiling[k] == dm_sw_linear) { + mode_lib->vba.MaximumSwathWidthSupport = 8192.0; + } else { + mode_lib->vba.MaximumSwathWidthSupport = 5120.0; + } + mode_lib->vba.MaximumSwathWidthInDETBuffer = + dml_min( + mode_lib->vba.MaximumSwathWidthSupport, + mode_lib->vba.DETBufferSizeInKByte * 1024.0 / 2.0 + / (mode_lib->vba.BytePerPixelInDETY[k] + * mode_lib->vba.MinSwathHeightY[k] + + mode_lib->vba.BytePerPixelInDETC[k] + / 2.0 + * mode_lib->vba.MinSwathHeightC[k])); + if (mode_lib->vba.BytePerPixelInDETC[k] == 0.0) { + mode_lib->vba.MaximumSwathWidthInLineBuffer = + mode_lib->vba.LineBufferSize + * dml_max(mode_lib->vba.HRatio[k], 1.0) + / mode_lib->vba.LBBitPerPixel[k] + / (mode_lib->vba.vtaps[k] + + dml_max( + dml_ceil( + mode_lib->vba.VRatio[k], + 1.0) + - 2, + 0.0)); + } else { + mode_lib->vba.MaximumSwathWidthInLineBuffer = + dml_min( + mode_lib->vba.LineBufferSize + * dml_max( + mode_lib->vba.HRatio[k], + 1.0) + / mode_lib->vba.LBBitPerPixel[k] + / (mode_lib->vba.vtaps[k] + + dml_max( + dml_ceil( + mode_lib->vba.VRatio[k], + 1.0) + - 2, + 0.0)), + 2.0 * mode_lib->vba.LineBufferSize + * dml_max( + mode_lib->vba.HRatio[k] + / 2.0, + 1.0) + / mode_lib->vba.LBBitPerPixel[k] + / (mode_lib->vba.VTAPsChroma[k] + + dml_max( + dml_ceil( + mode_lib->vba.VRatio[k] + / 2.0, + 1.0) + - 2, + 0.0))); + } + mode_lib->vba.MaximumSwathWidth[k] = dml_min( + mode_lib->vba.MaximumSwathWidthInDETBuffer, + mode_lib->vba.MaximumSwathWidthInLineBuffer); + } + for (i = 0; i <= DC__VOLTAGE_STATES; i++) { + mode_lib->vba.MaxDispclkRoundedDownToDFSGranularity = RoundToDFSGranularityDown( + mode_lib->vba.MaxDispclk[i], + mode_lib->vba.DISPCLKDPPCLKVCOSpeed); + mode_lib->vba.MaxDppclkRoundedDownToDFSGranularity = RoundToDFSGranularityDown( + mode_lib->vba.MaxDppclk[i], + mode_lib->vba.DISPCLKDPPCLKVCOSpeed); + mode_lib->vba.RequiredDISPCLK[i] = 0.0; + mode_lib->vba.DISPCLK_DPPCLK_Support[i] = true; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + mode_lib->vba.PlaneRequiredDISPCLKWithoutODMCombine = + mode_lib->vba.PixelClock[k] + * (1.0 + + mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading + / 100.0) + * (1.0 + + mode_lib->vba.DISPCLKRampingMargin + / 100.0); + if (mode_lib->vba.ODMCapability == true + && mode_lib->vba.PlaneRequiredDISPCLKWithoutODMCombine + > mode_lib->vba.MaxDispclkRoundedDownToDFSGranularity) { + mode_lib->vba.ODMCombineEnablePerState[i][k] = true; + mode_lib->vba.PlaneRequiredDISPCLK = + mode_lib->vba.PlaneRequiredDISPCLKWithoutODMCombine + / 2.0; + } else { + mode_lib->vba.ODMCombineEnablePerState[i][k] = false; + mode_lib->vba.PlaneRequiredDISPCLK = + mode_lib->vba.PlaneRequiredDISPCLKWithoutODMCombine; + } + if (mode_lib->vba.MinDPPCLKUsingSingleDPP[k] + * (1.0 + + mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading + / 100.0) + <= mode_lib->vba.MaxDppclkRoundedDownToDFSGranularity + && mode_lib->vba.SwathWidthYSingleDPP[k] + <= mode_lib->vba.MaximumSwathWidth[k] + && mode_lib->vba.ODMCombineEnablePerState[i][k] == false) { + mode_lib->vba.NoOfDPP[i][k] = 1; + mode_lib->vba.RequiredDPPCLK[i][k] = + mode_lib->vba.MinDPPCLKUsingSingleDPP[k] + * (1.0 + + mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading + / 100.0); + } else { + mode_lib->vba.NoOfDPP[i][k] = 2; + mode_lib->vba.RequiredDPPCLK[i][k] = + mode_lib->vba.MinDPPCLKUsingSingleDPP[k] + * (1.0 + + mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading + / 100.0) + / 2.0; + } + mode_lib->vba.RequiredDISPCLK[i] = dml_max( + mode_lib->vba.RequiredDISPCLK[i], + mode_lib->vba.PlaneRequiredDISPCLK); + if ((mode_lib->vba.MinDPPCLKUsingSingleDPP[k] / mode_lib->vba.NoOfDPP[i][k] + * (1.0 + + mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading + / 100.0) + > mode_lib->vba.MaxDppclkRoundedDownToDFSGranularity) + || (mode_lib->vba.PlaneRequiredDISPCLK + > mode_lib->vba.MaxDispclkRoundedDownToDFSGranularity)) { + mode_lib->vba.DISPCLK_DPPCLK_Support[i] = false; + } + } + mode_lib->vba.TotalNumberOfActiveDPP[i] = 0.0; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + mode_lib->vba.TotalNumberOfActiveDPP[i] = + mode_lib->vba.TotalNumberOfActiveDPP[i] + + mode_lib->vba.NoOfDPP[i][k]; + } + if ((mode_lib->vba.MaxDispclk[i] == mode_lib->vba.MaxDispclk[DC__VOLTAGE_STATES] + && mode_lib->vba.MaxDppclk[i] + == mode_lib->vba.MaxDppclk[DC__VOLTAGE_STATES]) + && (mode_lib->vba.TotalNumberOfActiveDPP[i] + > mode_lib->vba.MaxNumDPP + || mode_lib->vba.DISPCLK_DPPCLK_Support[i] == false)) { + mode_lib->vba.RequiredDISPCLK[i] = 0.0; + mode_lib->vba.DISPCLK_DPPCLK_Support[i] = true; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + mode_lib->vba.PlaneRequiredDISPCLKWithoutODMCombine = + mode_lib->vba.PixelClock[k] + * (1.0 + + mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading + / 100.0); + if (mode_lib->vba.ODMCapability == true + && mode_lib->vba.PlaneRequiredDISPCLKWithoutODMCombine + > mode_lib->vba.MaxDispclkRoundedDownToDFSGranularity) { + mode_lib->vba.ODMCombineEnablePerState[i][k] = true; + mode_lib->vba.PlaneRequiredDISPCLK = + mode_lib->vba.PlaneRequiredDISPCLKWithoutODMCombine + / 2.0; + } else { + mode_lib->vba.ODMCombineEnablePerState[i][k] = false; + mode_lib->vba.PlaneRequiredDISPCLK = + mode_lib->vba.PlaneRequiredDISPCLKWithoutODMCombine; + } + if (mode_lib->vba.MinDPPCLKUsingSingleDPP[k] + * (1.0 + + mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading + / 100.0) + <= mode_lib->vba.MaxDppclkRoundedDownToDFSGranularity + && mode_lib->vba.SwathWidthYSingleDPP[k] + <= mode_lib->vba.MaximumSwathWidth[k] + && mode_lib->vba.ODMCombineEnablePerState[i][k] + == false) { + mode_lib->vba.NoOfDPP[i][k] = 1; + mode_lib->vba.RequiredDPPCLK[i][k] = + mode_lib->vba.MinDPPCLKUsingSingleDPP[k] + * (1.0 + + mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading + / 100.0); + } else { + mode_lib->vba.NoOfDPP[i][k] = 2; + mode_lib->vba.RequiredDPPCLK[i][k] = + mode_lib->vba.MinDPPCLKUsingSingleDPP[k] + * (1.0 + + mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading + / 100.0) + / 2.0; + } + mode_lib->vba.RequiredDISPCLK[i] = dml_max( + mode_lib->vba.RequiredDISPCLK[i], + mode_lib->vba.PlaneRequiredDISPCLK); + if ((mode_lib->vba.MinDPPCLKUsingSingleDPP[k] + / mode_lib->vba.NoOfDPP[i][k] + * (1.0 + + mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading + / 100.0) + > mode_lib->vba.MaxDppclkRoundedDownToDFSGranularity) + || (mode_lib->vba.PlaneRequiredDISPCLK + > mode_lib->vba.MaxDispclkRoundedDownToDFSGranularity)) { + mode_lib->vba.DISPCLK_DPPCLK_Support[i] = false; + } + } + mode_lib->vba.TotalNumberOfActiveDPP[i] = 0.0; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + mode_lib->vba.TotalNumberOfActiveDPP[i] = + mode_lib->vba.TotalNumberOfActiveDPP[i] + + mode_lib->vba.NoOfDPP[i][k]; + } + } + if (mode_lib->vba.TotalNumberOfActiveDPP[i] > mode_lib->vba.MaxNumDPP) { + mode_lib->vba.RequiredDISPCLK[i] = 0.0; + mode_lib->vba.DISPCLK_DPPCLK_Support[i] = true; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + mode_lib->vba.ODMCombineEnablePerState[i][k] = false; + if (mode_lib->vba.SwathWidthYSingleDPP[k] + <= mode_lib->vba.MaximumSwathWidth[k]) { + mode_lib->vba.NoOfDPP[i][k] = 1; + mode_lib->vba.RequiredDPPCLK[i][k] = + mode_lib->vba.MinDPPCLKUsingSingleDPP[k] + * (1.0 + + mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading + / 100.0); + } else { + mode_lib->vba.NoOfDPP[i][k] = 2; + mode_lib->vba.RequiredDPPCLK[i][k] = + mode_lib->vba.MinDPPCLKUsingSingleDPP[k] + * (1.0 + + mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading + / 100.0) + / 2.0; + } + if (!(mode_lib->vba.MaxDispclk[i] + == mode_lib->vba.MaxDispclk[DC__VOLTAGE_STATES] + && mode_lib->vba.MaxDppclk[i] + == mode_lib->vba.MaxDppclk[DC__VOLTAGE_STATES])) { + mode_lib->vba.PlaneRequiredDISPCLK = + mode_lib->vba.PixelClock[k] + * (1.0 + + mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading + / 100.0) + * (1.0 + + mode_lib->vba.DISPCLKRampingMargin + / 100.0); + } else { + mode_lib->vba.PlaneRequiredDISPCLK = + mode_lib->vba.PixelClock[k] + * (1.0 + + mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading + / 100.0); + } + mode_lib->vba.RequiredDISPCLK[i] = dml_max( + mode_lib->vba.RequiredDISPCLK[i], + mode_lib->vba.PlaneRequiredDISPCLK); + if ((mode_lib->vba.MinDPPCLKUsingSingleDPP[k] + / mode_lib->vba.NoOfDPP[i][k] + * (1.0 + + mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading + / 100.0) + > mode_lib->vba.MaxDppclkRoundedDownToDFSGranularity) + || (mode_lib->vba.PlaneRequiredDISPCLK + > mode_lib->vba.MaxDispclkRoundedDownToDFSGranularity)) { + mode_lib->vba.DISPCLK_DPPCLK_Support[i] = false; + } + } + mode_lib->vba.TotalNumberOfActiveDPP[i] = 0.0; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + mode_lib->vba.TotalNumberOfActiveDPP[i] = + mode_lib->vba.TotalNumberOfActiveDPP[i] + + mode_lib->vba.NoOfDPP[i][k]; + } + } + mode_lib->vba.RequiredDISPCLK[i] = dml_max( + mode_lib->vba.RequiredDISPCLK[i], + mode_lib->vba.WritebackRequiredDISPCLK); + if (mode_lib->vba.MaxDispclkRoundedDownToDFSGranularity + < mode_lib->vba.WritebackRequiredDISPCLK) { + mode_lib->vba.DISPCLK_DPPCLK_Support[i] = false; + } + } + /*Viewport Size Check*/ + + for (i = 0; i <= DC__VOLTAGE_STATES; i++) { + mode_lib->vba.ViewportSizeSupport[i] = true; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + if (mode_lib->vba.ODMCombineEnablePerState[i][k] == true) { + if (dml_min( + mode_lib->vba.SwathWidthYSingleDPP[k], + dml_round( + mode_lib->vba.HActive[k] / 2.0 + * mode_lib->vba.HRatio[k])) + > mode_lib->vba.MaximumSwathWidth[k]) { + mode_lib->vba.ViewportSizeSupport[i] = false; + } + } else { + if (mode_lib->vba.SwathWidthYSingleDPP[k] / 2.0 + > mode_lib->vba.MaximumSwathWidth[k]) { + mode_lib->vba.ViewportSizeSupport[i] = false; + } + } + } + } + /*Total Available Pipes Support Check*/ + + for (i = 0; i <= DC__VOLTAGE_STATES; i++) { + if (mode_lib->vba.TotalNumberOfActiveDPP[i] <= mode_lib->vba.MaxNumDPP) { + mode_lib->vba.TotalAvailablePipesSupport[i] = true; + } else { + mode_lib->vba.TotalAvailablePipesSupport[i] = false; + } + } + /*Total Available OTG Support Check*/ + + mode_lib->vba.TotalNumberOfActiveOTG = 0.0; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + if (mode_lib->vba.BlendingAndTiming[k] == k) { + mode_lib->vba.TotalNumberOfActiveOTG = mode_lib->vba.TotalNumberOfActiveOTG + + 1.0; + } + } + if (mode_lib->vba.TotalNumberOfActiveOTG <= mode_lib->vba.MaxNumOTG) { + mode_lib->vba.NumberOfOTGSupport = true; + } else { + mode_lib->vba.NumberOfOTGSupport = false; + } + /*Display IO and DSC Support Check*/ + + mode_lib->vba.NonsupportedDSCInputBPC = false; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + if (!(mode_lib->vba.DSCInputBitPerComponent[k] == 12.0 + || mode_lib->vba.DSCInputBitPerComponent[k] == 10.0 + || mode_lib->vba.DSCInputBitPerComponent[k] == 8.0)) { + mode_lib->vba.NonsupportedDSCInputBPC = true; + } + } + for (i = 0; i <= DC__VOLTAGE_STATES; i++) { + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + mode_lib->vba.RequiresDSC[i][k] = 0; + mode_lib->vba.RequiresFEC[i][k] = 0; + if (mode_lib->vba.BlendingAndTiming[k] == k) { + if (mode_lib->vba.Output[k] == dm_hdmi) { + mode_lib->vba.RequiresDSC[i][k] = 0; + mode_lib->vba.RequiresFEC[i][k] = 0; + mode_lib->vba.OutputBppPerState[i][k] = + TruncToValidBPP( + dml_min( + 600.0, + mode_lib->vba.PHYCLKPerState[i]) + / mode_lib->vba.PixelClockBackEnd[k] + * 24, + false, + mode_lib->vba.Output[k], + mode_lib->vba.OutputFormat[k], + mode_lib->vba.DSCInputBitPerComponent[k]); + } else if (mode_lib->vba.Output[k] == dm_dp + || mode_lib->vba.Output[k] == dm_edp) { + if (mode_lib->vba.Output[k] == dm_edp) { + mode_lib->vba.EffectiveFECOverhead = 0.0; + } else { + mode_lib->vba.EffectiveFECOverhead = + mode_lib->vba.FECOverhead; + } + if (mode_lib->vba.PHYCLKPerState[i] >= 270.0) { + mode_lib->vba.Outbpp = + TruncToValidBPP( + (1.0 + - mode_lib->vba.Downspreading + / 100.0) + * 270.0 + * mode_lib->vba.OutputLinkDPLanes[k] + / mode_lib->vba.PixelClockBackEnd[k] + * 8.0, + false, + mode_lib->vba.Output[k], + mode_lib->vba.OutputFormat[k], + mode_lib->vba.DSCInputBitPerComponent[k]); + mode_lib->vba.OutbppDSC = + TruncToValidBPP( + (1.0 + - mode_lib->vba.Downspreading + / 100.0) + * (1.0 + - mode_lib->vba.EffectiveFECOverhead + / 100.0) + * 270.0 + * mode_lib->vba.OutputLinkDPLanes[k] + / mode_lib->vba.PixelClockBackEnd[k] + * 8.0, + true, + mode_lib->vba.Output[k], + mode_lib->vba.OutputFormat[k], + mode_lib->vba.DSCInputBitPerComponent[k]); + if (mode_lib->vba.DSCEnabled[k] == true) { + mode_lib->vba.RequiresDSC[i][k] = true; + if (mode_lib->vba.Output[k] == dm_dp) { + mode_lib->vba.RequiresFEC[i][k] = + true; + } else { + mode_lib->vba.RequiresFEC[i][k] = + false; + } + mode_lib->vba.Outbpp = + mode_lib->vba.OutbppDSC; + } else { + mode_lib->vba.RequiresDSC[i][k] = false; + mode_lib->vba.RequiresFEC[i][k] = false; + } + mode_lib->vba.OutputBppPerState[i][k] = + mode_lib->vba.Outbpp; + } + if (mode_lib->vba.Outbpp == 0) { + mode_lib->vba.Outbpp = + TruncToValidBPP( + (1.0 + - mode_lib->vba.Downspreading + / 100.0) + * 540.0 + * mode_lib->vba.OutputLinkDPLanes[k] + / mode_lib->vba.PixelClockBackEnd[k] + * 8.0, + false, + mode_lib->vba.Output[k], + mode_lib->vba.OutputFormat[k], + mode_lib->vba.DSCInputBitPerComponent[k]); + mode_lib->vba.OutbppDSC = + TruncToValidBPP( + (1.0 + - mode_lib->vba.Downspreading + / 100.0) + * (1.0 + - mode_lib->vba.EffectiveFECOverhead + / 100.0) + * 540.0 + * mode_lib->vba.OutputLinkDPLanes[k] + / mode_lib->vba.PixelClockBackEnd[k] + * 8.0, + true, + mode_lib->vba.Output[k], + mode_lib->vba.OutputFormat[k], + mode_lib->vba.DSCInputBitPerComponent[k]); + if (mode_lib->vba.DSCEnabled[k] == true) { + mode_lib->vba.RequiresDSC[i][k] = true; + if (mode_lib->vba.Output[k] == dm_dp) { + mode_lib->vba.RequiresFEC[i][k] = + true; + } else { + mode_lib->vba.RequiresFEC[i][k] = + false; + } + mode_lib->vba.Outbpp = + mode_lib->vba.OutbppDSC; + } else { + mode_lib->vba.RequiresDSC[i][k] = false; + mode_lib->vba.RequiresFEC[i][k] = false; + } + mode_lib->vba.OutputBppPerState[i][k] = + mode_lib->vba.Outbpp; + } + if (mode_lib->vba.Outbpp == 0 + && mode_lib->vba.PHYCLKPerState[i] + >= 810.0) { + mode_lib->vba.Outbpp = + TruncToValidBPP( + (1.0 + - mode_lib->vba.Downspreading + / 100.0) + * 810.0 + * mode_lib->vba.OutputLinkDPLanes[k] + / mode_lib->vba.PixelClockBackEnd[k] + * 8.0, + false, + mode_lib->vba.Output[k], + mode_lib->vba.OutputFormat[k], + mode_lib->vba.DSCInputBitPerComponent[k]); + mode_lib->vba.OutbppDSC = + TruncToValidBPP( + (1.0 + - mode_lib->vba.Downspreading + / 100.0) + * (1.0 + - mode_lib->vba.EffectiveFECOverhead + / 100.0) + * 810.0 + * mode_lib->vba.OutputLinkDPLanes[k] + / mode_lib->vba.PixelClockBackEnd[k] + * 8.0, + true, + mode_lib->vba.Output[k], + mode_lib->vba.OutputFormat[k], + mode_lib->vba.DSCInputBitPerComponent[k]); + if (mode_lib->vba.DSCEnabled[k] == true + || mode_lib->vba.Outbpp == 0) { + mode_lib->vba.RequiresDSC[i][k] = true; + if (mode_lib->vba.Output[k] == dm_dp) { + mode_lib->vba.RequiresFEC[i][k] = + true; + } else { + mode_lib->vba.RequiresFEC[i][k] = + false; + } + mode_lib->vba.Outbpp = + mode_lib->vba.OutbppDSC; + } else { + mode_lib->vba.RequiresDSC[i][k] = false; + mode_lib->vba.RequiresFEC[i][k] = false; + } + mode_lib->vba.OutputBppPerState[i][k] = + mode_lib->vba.Outbpp; + } + } + } else { + mode_lib->vba.OutputBppPerState[i][k] = 0; + } + } + } + for (i = 0; i <= DC__VOLTAGE_STATES; i++) { + mode_lib->vba.DIOSupport[i] = true; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + if (mode_lib->vba.OutputBppPerState[i][k] == 0 + || (mode_lib->vba.OutputFormat[k] == dm_420 + && mode_lib->vba.ProgressiveToInterlaceUnitInOPP + == true)) { + mode_lib->vba.DIOSupport[i] = false; + } + } + } + for (i = 0; i <= DC__VOLTAGE_STATES; i++) { + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + mode_lib->vba.DSCCLKRequiredMoreThanSupported[i] = false; + if (mode_lib->vba.BlendingAndTiming[k] == k) { + if ((mode_lib->vba.Output[k] == dm_dp + || mode_lib->vba.Output[k] == dm_edp)) { + if (mode_lib->vba.OutputFormat[k] == dm_420 + || mode_lib->vba.OutputFormat[k] + == dm_n422) { + mode_lib->vba.DSCFormatFactor = 2; + } else { + mode_lib->vba.DSCFormatFactor = 1; + } + if (mode_lib->vba.RequiresDSC[i][k] == true) { + if (mode_lib->vba.ODMCombineEnablePerState[i][k] + == true) { + if (mode_lib->vba.PixelClockBackEnd[k] / 6.0 + / mode_lib->vba.DSCFormatFactor + > (1.0 + - mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading + / 100.0) + * mode_lib->vba.MaxDSCCLK[i]) { + mode_lib->vba.DSCCLKRequiredMoreThanSupported[i] = + true; + } + } else { + if (mode_lib->vba.PixelClockBackEnd[k] / 3.0 + / mode_lib->vba.DSCFormatFactor + > (1.0 + - mode_lib->vba.DISPCLKDPPCLKDSCCLKDownSpreading + / 100.0) + * mode_lib->vba.MaxDSCCLK[i]) { + mode_lib->vba.DSCCLKRequiredMoreThanSupported[i] = + true; + } + } + } + } + } + } + } + for (i = 0; i <= DC__VOLTAGE_STATES; i++) { + mode_lib->vba.NotEnoughDSCUnits[i] = false; + mode_lib->vba.TotalDSCUnitsRequired = 0.0; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + if (mode_lib->vba.RequiresDSC[i][k] == true) { + if (mode_lib->vba.ODMCombineEnablePerState[i][k] == true) { + mode_lib->vba.TotalDSCUnitsRequired = + mode_lib->vba.TotalDSCUnitsRequired + 2.0; + } else { + mode_lib->vba.TotalDSCUnitsRequired = + mode_lib->vba.TotalDSCUnitsRequired + 1.0; + } + } + } + if (mode_lib->vba.TotalDSCUnitsRequired > mode_lib->vba.NumberOfDSC) { + mode_lib->vba.NotEnoughDSCUnits[i] = true; + } + } + /*DSC Delay per state*/ + + for (i = 0; i <= DC__VOLTAGE_STATES; i++) { + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + if (mode_lib->vba.BlendingAndTiming[k] != k) { + mode_lib->vba.slices = 0; + } else if (mode_lib->vba.RequiresDSC[i][k] == 0 + || mode_lib->vba.RequiresDSC[i][k] == false) { + mode_lib->vba.slices = 0; + } else if (mode_lib->vba.PixelClockBackEnd[k] > 3200.0) { + mode_lib->vba.slices = dml_ceil( + mode_lib->vba.PixelClockBackEnd[k] / 400.0, + 4.0); + } else if (mode_lib->vba.PixelClockBackEnd[k] > 1360.0) { + mode_lib->vba.slices = 8.0; + } else if (mode_lib->vba.PixelClockBackEnd[k] > 680.0) { + mode_lib->vba.slices = 4.0; + } else if (mode_lib->vba.PixelClockBackEnd[k] > 340.0) { + mode_lib->vba.slices = 2.0; + } else { + mode_lib->vba.slices = 1.0; + } + if (mode_lib->vba.OutputBppPerState[i][k] == 0 + || mode_lib->vba.OutputBppPerState[i][k] == 0) { + mode_lib->vba.bpp = 0.0; + } else { + mode_lib->vba.bpp = mode_lib->vba.OutputBppPerState[i][k]; + } + if (mode_lib->vba.RequiresDSC[i][k] == true && mode_lib->vba.bpp != 0.0) { + if (mode_lib->vba.ODMCombineEnablePerState[i][k] == false) { + mode_lib->vba.DSCDelayPerState[i][k] = + dscceComputeDelay( + mode_lib->vba.DSCInputBitPerComponent[k], + mode_lib->vba.bpp, + dml_ceil( + mode_lib->vba.HActive[k] + / mode_lib->vba.slices, + 1.0), + mode_lib->vba.slices, + mode_lib->vba.OutputFormat[k]) + + dscComputeDelay( + mode_lib->vba.OutputFormat[k]); + } else { + mode_lib->vba.DSCDelayPerState[i][k] = + 2.0 + * (dscceComputeDelay( + mode_lib->vba.DSCInputBitPerComponent[k], + mode_lib->vba.bpp, + dml_ceil( + mode_lib->vba.HActive[k] + / mode_lib->vba.slices, + 1.0), + mode_lib->vba.slices + / 2, + mode_lib->vba.OutputFormat[k]) + + dscComputeDelay( + mode_lib->vba.OutputFormat[k])); + } + mode_lib->vba.DSCDelayPerState[i][k] = + mode_lib->vba.DSCDelayPerState[i][k] + * mode_lib->vba.PixelClock[k] + / mode_lib->vba.PixelClockBackEnd[k]; + } else { + mode_lib->vba.DSCDelayPerState[i][k] = 0.0; + } + } + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + for (j = 0; j <= mode_lib->vba.NumberOfActivePlanes - 1; j++) { + if (mode_lib->vba.BlendingAndTiming[k] == j + && mode_lib->vba.RequiresDSC[i][j] == true) { + mode_lib->vba.DSCDelayPerState[i][k] = + mode_lib->vba.DSCDelayPerState[i][j]; + } + } + } + } + /*Urgent Latency Support Check*/ + + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + for (i = 0; i <= DC__VOLTAGE_STATES; i++) { + if (mode_lib->vba.ODMCombineEnablePerState[i][k] == true) { + mode_lib->vba.SwathWidthYPerState[i][k] = + dml_min( + mode_lib->vba.SwathWidthYSingleDPP[k], + dml_round( + mode_lib->vba.HActive[k] + / 2.0 + * mode_lib->vba.HRatio[k])); + } else { + mode_lib->vba.SwathWidthYPerState[i][k] = + mode_lib->vba.SwathWidthYSingleDPP[k] + / mode_lib->vba.NoOfDPP[i][k]; + } + mode_lib->vba.SwathWidthGranularityY = 256.0 + / dml_ceil(mode_lib->vba.BytePerPixelInDETY[k], 1.0) + / mode_lib->vba.MaxSwathHeightY[k]; + mode_lib->vba.RoundedUpMaxSwathSizeBytesY = (dml_ceil( + mode_lib->vba.SwathWidthYPerState[i][k] - 1.0, + mode_lib->vba.SwathWidthGranularityY) + + mode_lib->vba.SwathWidthGranularityY) + * mode_lib->vba.BytePerPixelInDETY[k] + * mode_lib->vba.MaxSwathHeightY[k]; + if (mode_lib->vba.SourcePixelFormat[k] == dm_420_10) { + mode_lib->vba.RoundedUpMaxSwathSizeBytesY = dml_ceil( + mode_lib->vba.RoundedUpMaxSwathSizeBytesY, + 256.0) + 256; + } + if (mode_lib->vba.MaxSwathHeightC[k] > 0.0) { + mode_lib->vba.SwathWidthGranularityC = 256.0 + / dml_ceil(mode_lib->vba.BytePerPixelInDETC[k], 2.0) + / mode_lib->vba.MaxSwathHeightC[k]; + mode_lib->vba.RoundedUpMaxSwathSizeBytesC = (dml_ceil( + mode_lib->vba.SwathWidthYPerState[i][k] / 2.0 - 1.0, + mode_lib->vba.SwathWidthGranularityC) + + mode_lib->vba.SwathWidthGranularityC) + * mode_lib->vba.BytePerPixelInDETC[k] + * mode_lib->vba.MaxSwathHeightC[k]; + if (mode_lib->vba.SourcePixelFormat[k] == dm_420_10) { + mode_lib->vba.RoundedUpMaxSwathSizeBytesC = dml_ceil( + mode_lib->vba.RoundedUpMaxSwathSizeBytesC, + 256.0) + 256; + } + } else { + mode_lib->vba.RoundedUpMaxSwathSizeBytesC = 0.0; + } + if (mode_lib->vba.RoundedUpMaxSwathSizeBytesY + + mode_lib->vba.RoundedUpMaxSwathSizeBytesC + <= mode_lib->vba.DETBufferSizeInKByte * 1024.0 / 2.0) { + mode_lib->vba.SwathHeightYPerState[i][k] = + mode_lib->vba.MaxSwathHeightY[k]; + mode_lib->vba.SwathHeightCPerState[i][k] = + mode_lib->vba.MaxSwathHeightC[k]; + } else { + mode_lib->vba.SwathHeightYPerState[i][k] = + mode_lib->vba.MinSwathHeightY[k]; + mode_lib->vba.SwathHeightCPerState[i][k] = + mode_lib->vba.MinSwathHeightC[k]; + } + if (mode_lib->vba.BytePerPixelInDETC[k] == 0.0) { + mode_lib->vba.LinesInDETLuma = mode_lib->vba.DETBufferSizeInKByte + * 1024.0 / mode_lib->vba.BytePerPixelInDETY[k] + / mode_lib->vba.SwathWidthYPerState[i][k]; + mode_lib->vba.LinesInDETChroma = 0.0; + } else if (mode_lib->vba.SwathHeightYPerState[i][k] + <= mode_lib->vba.SwathHeightCPerState[i][k]) { + mode_lib->vba.LinesInDETLuma = mode_lib->vba.DETBufferSizeInKByte + * 1024.0 / 2.0 / mode_lib->vba.BytePerPixelInDETY[k] + / mode_lib->vba.SwathWidthYPerState[i][k]; + mode_lib->vba.LinesInDETChroma = mode_lib->vba.DETBufferSizeInKByte + * 1024.0 / 2.0 / mode_lib->vba.BytePerPixelInDETC[k] + / (mode_lib->vba.SwathWidthYPerState[i][k] / 2.0); + } else { + mode_lib->vba.LinesInDETLuma = mode_lib->vba.DETBufferSizeInKByte + * 1024.0 * 2.0 / 3.0 + / mode_lib->vba.BytePerPixelInDETY[k] + / mode_lib->vba.SwathWidthYPerState[i][k]; + mode_lib->vba.LinesInDETChroma = mode_lib->vba.DETBufferSizeInKByte + * 1024.0 / 3.0 / mode_lib->vba.BytePerPixelInDETY[k] + / (mode_lib->vba.SwathWidthYPerState[i][k] / 2.0); + } + mode_lib->vba.EffectiveLBLatencyHidingSourceLinesLuma = + dml_min( + mode_lib->vba.MaxLineBufferLines, + dml_floor( + mode_lib->vba.LineBufferSize + / mode_lib->vba.LBBitPerPixel[k] + / (mode_lib->vba.SwathWidthYPerState[i][k] + / dml_max( + mode_lib->vba.HRatio[k], + 1.0)), + 1.0)) + - (mode_lib->vba.vtaps[k] - 1.0); + mode_lib->vba.EffectiveLBLatencyHidingSourceLinesChroma = + dml_min( + mode_lib->vba.MaxLineBufferLines, + dml_floor( + mode_lib->vba.LineBufferSize + / mode_lib->vba.LBBitPerPixel[k] + / (mode_lib->vba.SwathWidthYPerState[i][k] + / 2.0 + / dml_max( + mode_lib->vba.HRatio[k] + / 2.0, + 1.0)), + 1.0)) + - (mode_lib->vba.VTAPsChroma[k] - 1.0); + mode_lib->vba.EffectiveDETLBLinesLuma = + dml_floor( + mode_lib->vba.LinesInDETLuma + + dml_min( + mode_lib->vba.LinesInDETLuma + * mode_lib->vba.RequiredDISPCLK[i] + * mode_lib->vba.BytePerPixelInDETY[k] + * mode_lib->vba.PSCL_FACTOR[k] + / mode_lib->vba.ReturnBWPerState[i], + mode_lib->vba.EffectiveLBLatencyHidingSourceLinesLuma), + mode_lib->vba.SwathHeightYPerState[i][k]); + mode_lib->vba.EffectiveDETLBLinesChroma = + dml_floor( + mode_lib->vba.LinesInDETChroma + + dml_min( + mode_lib->vba.LinesInDETChroma + * mode_lib->vba.RequiredDISPCLK[i] + * mode_lib->vba.BytePerPixelInDETC[k] + * mode_lib->vba.PSCL_FACTOR_CHROMA[k] + / mode_lib->vba.ReturnBWPerState[i], + mode_lib->vba.EffectiveLBLatencyHidingSourceLinesChroma), + mode_lib->vba.SwathHeightCPerState[i][k]); + if (mode_lib->vba.BytePerPixelInDETC[k] == 0.0) { + mode_lib->vba.UrgentLatencySupportUsPerState[i][k] = + mode_lib->vba.EffectiveDETLBLinesLuma + * (mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k]) + / mode_lib->vba.VRatio[k] + - mode_lib->vba.EffectiveDETLBLinesLuma + * mode_lib->vba.SwathWidthYPerState[i][k] + * dml_ceil( + mode_lib->vba.BytePerPixelInDETY[k], + 1.0) + / (mode_lib->vba.ReturnBWPerState[i] + / mode_lib->vba.NoOfDPP[i][k]); + } else { + mode_lib->vba.UrgentLatencySupportUsPerState[i][k] = + dml_min( + mode_lib->vba.EffectiveDETLBLinesLuma + * (mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k]) + / mode_lib->vba.VRatio[k] + - mode_lib->vba.EffectiveDETLBLinesLuma + * mode_lib->vba.SwathWidthYPerState[i][k] + * dml_ceil( + mode_lib->vba.BytePerPixelInDETY[k], + 1.0) + / (mode_lib->vba.ReturnBWPerState[i] + / mode_lib->vba.NoOfDPP[i][k]), + mode_lib->vba.EffectiveDETLBLinesChroma + * (mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k]) + / (mode_lib->vba.VRatio[k] + / 2.0) + - mode_lib->vba.EffectiveDETLBLinesChroma + * mode_lib->vba.SwathWidthYPerState[i][k] + / 2.0 + * dml_ceil( + mode_lib->vba.BytePerPixelInDETC[k], + 2.0) + / (mode_lib->vba.ReturnBWPerState[i] + / mode_lib->vba.NoOfDPP[i][k])); + } + } + } + for (i = 0; i <= DC__VOLTAGE_STATES; i++) { + mode_lib->vba.UrgentLatencySupport[i] = true; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + if (mode_lib->vba.UrgentLatencySupportUsPerState[i][k] + < mode_lib->vba.UrgentLatency / 1.0) { + mode_lib->vba.UrgentLatencySupport[i] = false; + } + } + } + /*Prefetch Check*/ + + for (i = 0; i <= DC__VOLTAGE_STATES; i++) { + mode_lib->vba.TotalNumberOfDCCActiveDPP[i] = 0.0; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + if (mode_lib->vba.DCCEnable[k] == true) { + mode_lib->vba.TotalNumberOfDCCActiveDPP[i] = + mode_lib->vba.TotalNumberOfDCCActiveDPP[i] + + mode_lib->vba.NoOfDPP[i][k]; + } + } + } + for (i = 0; i <= DC__VOLTAGE_STATES; i++) { + mode_lib->vba.ProjectedDCFCLKDeepSleep = 8.0; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + mode_lib->vba.ProjectedDCFCLKDeepSleep = dml_max( + mode_lib->vba.ProjectedDCFCLKDeepSleep, + mode_lib->vba.PixelClock[k] / 16.0); + if (mode_lib->vba.BytePerPixelInDETC[k] == 0.0) { + if (mode_lib->vba.VRatio[k] <= 1.0) { + mode_lib->vba.ProjectedDCFCLKDeepSleep = + dml_max( + mode_lib->vba.ProjectedDCFCLKDeepSleep, + 1.1 + * dml_ceil( + mode_lib->vba.BytePerPixelInDETY[k], + 1.0) + / 64.0 + * mode_lib->vba.HRatio[k] + * mode_lib->vba.PixelClock[k] + / mode_lib->vba.NoOfDPP[i][k]); + } else { + mode_lib->vba.ProjectedDCFCLKDeepSleep = + dml_max( + mode_lib->vba.ProjectedDCFCLKDeepSleep, + 1.1 + * dml_ceil( + mode_lib->vba.BytePerPixelInDETY[k], + 1.0) + / 64.0 + * mode_lib->vba.PSCL_FACTOR[k] + * mode_lib->vba.RequiredDPPCLK[i][k]); + } + } else { + if (mode_lib->vba.VRatio[k] <= 1.0) { + mode_lib->vba.ProjectedDCFCLKDeepSleep = + dml_max( + mode_lib->vba.ProjectedDCFCLKDeepSleep, + 1.1 + * dml_ceil( + mode_lib->vba.BytePerPixelInDETY[k], + 1.0) + / 32.0 + * mode_lib->vba.HRatio[k] + * mode_lib->vba.PixelClock[k] + / mode_lib->vba.NoOfDPP[i][k]); + } else { + mode_lib->vba.ProjectedDCFCLKDeepSleep = + dml_max( + mode_lib->vba.ProjectedDCFCLKDeepSleep, + 1.1 + * dml_ceil( + mode_lib->vba.BytePerPixelInDETY[k], + 1.0) + / 32.0 + * mode_lib->vba.PSCL_FACTOR[k] + * mode_lib->vba.RequiredDPPCLK[i][k]); + } + if (mode_lib->vba.VRatio[k] / 2.0 <= 1.0) { + mode_lib->vba.ProjectedDCFCLKDeepSleep = + dml_max( + mode_lib->vba.ProjectedDCFCLKDeepSleep, + 1.1 + * dml_ceil( + mode_lib->vba.BytePerPixelInDETC[k], + 2.0) + / 32.0 + * mode_lib->vba.HRatio[k] + / 2.0 + * mode_lib->vba.PixelClock[k] + / mode_lib->vba.NoOfDPP[i][k]); + } else { + mode_lib->vba.ProjectedDCFCLKDeepSleep = + dml_max( + mode_lib->vba.ProjectedDCFCLKDeepSleep, + 1.1 + * dml_ceil( + mode_lib->vba.BytePerPixelInDETC[k], + 2.0) + / 32.0 + * mode_lib->vba.PSCL_FACTOR_CHROMA[k] + * mode_lib->vba.RequiredDPPCLK[i][k]); + } + } + } + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + mode_lib->vba.PDEAndMetaPTEBytesPerFrameY = CalculateVMAndRowBytes( + mode_lib, + mode_lib->vba.DCCEnable[k], + mode_lib->vba.Read256BlockHeightY[k], + mode_lib->vba.Read256BlockWidthY[k], + mode_lib->vba.SourcePixelFormat[k], + mode_lib->vba.SurfaceTiling[k], + dml_ceil(mode_lib->vba.BytePerPixelInDETY[k], 1.0), + mode_lib->vba.SourceScan[k], + mode_lib->vba.ViewportWidth[k], + mode_lib->vba.ViewportHeight[k], + mode_lib->vba.SwathWidthYPerState[i][k], + mode_lib->vba.VirtualMemoryEnable, + mode_lib->vba.VMMPageSize, + mode_lib->vba.PTEBufferSizeInRequests, + mode_lib->vba.PDEProcessingBufIn64KBReqs, + mode_lib->vba.PitchY[k], + mode_lib->vba.DCCMetaPitchY[k], + &mode_lib->vba.MacroTileWidthY[k], + &mode_lib->vba.MetaRowBytesY, + &mode_lib->vba.DPTEBytesPerRowY, + &mode_lib->vba.PTEBufferSizeNotExceededY[i][k], + &mode_lib->vba.dpte_row_height[k], + &mode_lib->vba.meta_row_height[k]); + mode_lib->vba.PrefetchLinesY[k] = CalculatePrefetchSourceLines( + mode_lib, + mode_lib->vba.VRatio[k], + mode_lib->vba.vtaps[k], + mode_lib->vba.Interlace[k], + mode_lib->vba.ProgressiveToInterlaceUnitInOPP, + mode_lib->vba.SwathHeightYPerState[i][k], + mode_lib->vba.ViewportYStartY[k], + &mode_lib->vba.PrefillY[k], + &mode_lib->vba.MaxNumSwY[k]); + if ((mode_lib->vba.SourcePixelFormat[k] != dm_444_64 + && mode_lib->vba.SourcePixelFormat[k] != dm_444_32 + && mode_lib->vba.SourcePixelFormat[k] != dm_444_16 + && mode_lib->vba.SourcePixelFormat[k] != dm_mono_16 + && mode_lib->vba.SourcePixelFormat[k] != dm_mono_8)) { + mode_lib->vba.PDEAndMetaPTEBytesPerFrameC = CalculateVMAndRowBytes( + mode_lib, + mode_lib->vba.DCCEnable[k], + mode_lib->vba.Read256BlockHeightY[k], + mode_lib->vba.Read256BlockWidthY[k], + mode_lib->vba.SourcePixelFormat[k], + mode_lib->vba.SurfaceTiling[k], + dml_ceil(mode_lib->vba.BytePerPixelInDETC[k], 2.0), + mode_lib->vba.SourceScan[k], + mode_lib->vba.ViewportWidth[k] / 2.0, + mode_lib->vba.ViewportHeight[k] / 2.0, + mode_lib->vba.SwathWidthYPerState[i][k] / 2.0, + mode_lib->vba.VirtualMemoryEnable, + mode_lib->vba.VMMPageSize, + mode_lib->vba.PTEBufferSizeInRequests, + mode_lib->vba.PDEProcessingBufIn64KBReqs, + mode_lib->vba.PitchC[k], + 0.0, + &mode_lib->vba.MacroTileWidthC[k], + &mode_lib->vba.MetaRowBytesC, + &mode_lib->vba.DPTEBytesPerRowC, + &mode_lib->vba.PTEBufferSizeNotExceededC[i][k], + &mode_lib->vba.dpte_row_height_chroma[k], + &mode_lib->vba.meta_row_height_chroma[k]); + mode_lib->vba.PrefetchLinesC[k] = CalculatePrefetchSourceLines( + mode_lib, + mode_lib->vba.VRatio[k] / 2.0, + mode_lib->vba.VTAPsChroma[k], + mode_lib->vba.Interlace[k], + mode_lib->vba.ProgressiveToInterlaceUnitInOPP, + mode_lib->vba.SwathHeightCPerState[i][k], + mode_lib->vba.ViewportYStartC[k], + &mode_lib->vba.PrefillC[k], + &mode_lib->vba.MaxNumSwC[k]); + } else { + mode_lib->vba.PDEAndMetaPTEBytesPerFrameC = 0.0; + mode_lib->vba.MetaRowBytesC = 0.0; + mode_lib->vba.DPTEBytesPerRowC = 0.0; + mode_lib->vba.PrefetchLinesC[k] = 0.0; + mode_lib->vba.PTEBufferSizeNotExceededC[i][k] = true; + } + mode_lib->vba.PDEAndMetaPTEBytesPerFrame[k] = + mode_lib->vba.PDEAndMetaPTEBytesPerFrameY + + mode_lib->vba.PDEAndMetaPTEBytesPerFrameC; + mode_lib->vba.MetaRowBytes[k] = mode_lib->vba.MetaRowBytesY + + mode_lib->vba.MetaRowBytesC; + mode_lib->vba.DPTEBytesPerRow[k] = mode_lib->vba.DPTEBytesPerRowY + + mode_lib->vba.DPTEBytesPerRowC; + } + mode_lib->vba.ExtraLatency = + mode_lib->vba.UrgentRoundTripAndOutOfOrderLatencyPerState[i] + + (mode_lib->vba.TotalNumberOfActiveDPP[i] + * mode_lib->vba.PixelChunkSizeInKByte + + mode_lib->vba.TotalNumberOfDCCActiveDPP[i] + * mode_lib->vba.MetaChunkSize) + * 1024.0 + / mode_lib->vba.ReturnBWPerState[i]; + if (mode_lib->vba.VirtualMemoryEnable == true) { + mode_lib->vba.ExtraLatency = mode_lib->vba.ExtraLatency + + mode_lib->vba.TotalNumberOfActiveDPP[i] + * mode_lib->vba.PTEChunkSize * 1024.0 + / mode_lib->vba.ReturnBWPerState[i]; + } + mode_lib->vba.TimeCalc = 24.0 / mode_lib->vba.ProjectedDCFCLKDeepSleep; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + if (mode_lib->vba.BlendingAndTiming[k] == k) { + if (mode_lib->vba.WritebackEnable[k] == true) { + mode_lib->vba.WritebackDelay[i][k] = + mode_lib->vba.WritebackLatency + + CalculateWriteBackDelay( + mode_lib->vba.WritebackPixelFormat[k], + mode_lib->vba.WritebackHRatio[k], + mode_lib->vba.WritebackVRatio[k], + mode_lib->vba.WritebackLumaHTaps[k], + mode_lib->vba.WritebackLumaVTaps[k], + mode_lib->vba.WritebackChromaHTaps[k], + mode_lib->vba.WritebackChromaVTaps[k], + mode_lib->vba.WritebackDestinationWidth[k]) + / mode_lib->vba.RequiredDISPCLK[i]; + } else { + mode_lib->vba.WritebackDelay[i][k] = 0.0; + } + for (j = 0; j <= mode_lib->vba.NumberOfActivePlanes - 1; j++) { + if (mode_lib->vba.BlendingAndTiming[j] == k + && mode_lib->vba.WritebackEnable[j] + == true) { + mode_lib->vba.WritebackDelay[i][k] = + dml_max( + mode_lib->vba.WritebackDelay[i][k], + mode_lib->vba.WritebackLatency + + CalculateWriteBackDelay( + mode_lib->vba.WritebackPixelFormat[j], + mode_lib->vba.WritebackHRatio[j], + mode_lib->vba.WritebackVRatio[j], + mode_lib->vba.WritebackLumaHTaps[j], + mode_lib->vba.WritebackLumaVTaps[j], + mode_lib->vba.WritebackChromaHTaps[j], + mode_lib->vba.WritebackChromaVTaps[j], + mode_lib->vba.WritebackDestinationWidth[j]) + / mode_lib->vba.RequiredDISPCLK[i]); + } + } + } + } + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + for (j = 0; j <= mode_lib->vba.NumberOfActivePlanes - 1; j++) { + if (mode_lib->vba.BlendingAndTiming[k] == j) { + mode_lib->vba.WritebackDelay[i][k] = + mode_lib->vba.WritebackDelay[i][j]; + } + } + } + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + mode_lib->vba.MaximumVStartup[k] = + mode_lib->vba.VTotal[k] - mode_lib->vba.VActive[k] + - dml_max( + 1.0, + dml_ceil( + mode_lib->vba.WritebackDelay[i][k] + / (mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k]), + 1.0)); + } + mode_lib->vba.TWait = CalculateTWait( + mode_lib->vba.PrefetchMode, + mode_lib->vba.DRAMClockChangeLatency, + mode_lib->vba.UrgentLatency, + mode_lib->vba.SREnterPlusExitTime); + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + if (mode_lib->vba.XFCEnabled[k] == true) { + mode_lib->vba.XFCRemoteSurfaceFlipDelay = + CalculateRemoteSurfaceFlipDelay( + mode_lib, + mode_lib->vba.VRatio[k], + mode_lib->vba.SwathWidthYPerState[i][k], + dml_ceil( + mode_lib->vba.BytePerPixelInDETY[k], + 1.0), + mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k], + mode_lib->vba.XFCTSlvVupdateOffset, + mode_lib->vba.XFCTSlvVupdateWidth, + mode_lib->vba.XFCTSlvVreadyOffset, + mode_lib->vba.XFCXBUFLatencyTolerance, + mode_lib->vba.XFCFillBWOverhead, + mode_lib->vba.XFCSlvChunkSize, + mode_lib->vba.XFCBusTransportTime, + mode_lib->vba.TimeCalc, + mode_lib->vba.TWait, + &mode_lib->vba.SrcActiveDrainRate, + &mode_lib->vba.TInitXFill, + &mode_lib->vba.TslvChk); + } else { + mode_lib->vba.XFCRemoteSurfaceFlipDelay = 0.0; + } + mode_lib->vba.IsErrorResult[i][k] = + CalculatePrefetchSchedule( + mode_lib, + mode_lib->vba.RequiredDPPCLK[i][k], + mode_lib->vba.RequiredDISPCLK[i], + mode_lib->vba.PixelClock[k], + mode_lib->vba.ProjectedDCFCLKDeepSleep, + mode_lib->vba.DSCDelayPerState[i][k], + mode_lib->vba.NoOfDPP[i][k], + mode_lib->vba.ScalerEnabled[k], + mode_lib->vba.NumberOfCursors[k], + mode_lib->vba.DPPCLKDelaySubtotal, + mode_lib->vba.DPPCLKDelaySCL, + mode_lib->vba.DPPCLKDelaySCLLBOnly, + mode_lib->vba.DPPCLKDelayCNVCFormater, + mode_lib->vba.DPPCLKDelayCNVCCursor, + mode_lib->vba.DISPCLKDelaySubtotal, + mode_lib->vba.SwathWidthYPerState[i][k] + / mode_lib->vba.HRatio[k], + mode_lib->vba.OutputFormat[k], + mode_lib->vba.VTotal[k] + - mode_lib->vba.VActive[k], + mode_lib->vba.HTotal[k], + mode_lib->vba.MaxInterDCNTileRepeaters, + mode_lib->vba.MaximumVStartup[k], + mode_lib->vba.MaxPageTableLevels, + mode_lib->vba.VirtualMemoryEnable, + mode_lib->vba.DynamicMetadataEnable[k], + mode_lib->vba.DynamicMetadataLinesBeforeActiveRequired[k], + mode_lib->vba.DynamicMetadataTransmittedBytes[k], + mode_lib->vba.DCCEnable[k], + mode_lib->vba.UrgentLatency, + mode_lib->vba.ExtraLatency, + mode_lib->vba.TimeCalc, + mode_lib->vba.PDEAndMetaPTEBytesPerFrame[k], + mode_lib->vba.MetaRowBytes[k], + mode_lib->vba.DPTEBytesPerRow[k], + mode_lib->vba.PrefetchLinesY[k], + mode_lib->vba.SwathWidthYPerState[i][k], + mode_lib->vba.BytePerPixelInDETY[k], + mode_lib->vba.PrefillY[k], + mode_lib->vba.MaxNumSwY[k], + mode_lib->vba.PrefetchLinesC[k], + mode_lib->vba.BytePerPixelInDETC[k], + mode_lib->vba.PrefillC[k], + mode_lib->vba.MaxNumSwC[k], + mode_lib->vba.SwathHeightYPerState[i][k], + mode_lib->vba.SwathHeightCPerState[i][k], + mode_lib->vba.TWait, + mode_lib->vba.XFCEnabled[k], + mode_lib->vba.XFCRemoteSurfaceFlipDelay, + mode_lib->vba.Interlace[k], + mode_lib->vba.ProgressiveToInterlaceUnitInOPP, + mode_lib->vba.DSTXAfterScaler, + mode_lib->vba.DSTYAfterScaler, + &mode_lib->vba.LineTimesForPrefetch[k], + &mode_lib->vba.PrefetchBW[k], + &mode_lib->vba.LinesForMetaPTE[k], + &mode_lib->vba.LinesForMetaAndDPTERow[k], + &mode_lib->vba.VRatioPreY[i][k], + &mode_lib->vba.VRatioPreC[i][k], + &mode_lib->vba.RequiredPrefetchPixelDataBW[i][k], + &mode_lib->vba.VStartupRequiredWhenNotEnoughTimeForDynamicMetadata, + &mode_lib->vba.Tno_bw[k], + &mode_lib->vba.VUpdateOffsetPix[k], + &mode_lib->vba.VUpdateWidthPix[k], + &mode_lib->vba.VReadyOffsetPix[k]); + } + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + mode_lib->vba.cursor_bw[k] = mode_lib->vba.NumberOfCursors[k] + * mode_lib->vba.CursorWidth[k][0] + * mode_lib->vba.CursorBPP[k][0] / 8.0 + / (mode_lib->vba.HTotal[k] / mode_lib->vba.PixelClock[k]) + * mode_lib->vba.VRatio[k]; + } + mode_lib->vba.MaximumReadBandwidthWithPrefetch = 0.0; + mode_lib->vba.prefetch_vm_bw_valid = true; + mode_lib->vba.prefetch_row_bw_valid = true; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + if (mode_lib->vba.PDEAndMetaPTEBytesPerFrame[k] == 0.0) { + mode_lib->vba.prefetch_vm_bw[k] = 0.0; + } else if (mode_lib->vba.LinesForMetaPTE[k] > 0.0) { + mode_lib->vba.prefetch_vm_bw[k] = + mode_lib->vba.PDEAndMetaPTEBytesPerFrame[k] + / (mode_lib->vba.LinesForMetaPTE[k] + * mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k]); + } else { + mode_lib->vba.prefetch_vm_bw[k] = 0.0; + mode_lib->vba.prefetch_vm_bw_valid = false; + } + if (mode_lib->vba.MetaRowBytes[k] + mode_lib->vba.DPTEBytesPerRow[k] + == 0.0) { + mode_lib->vba.prefetch_row_bw[k] = 0.0; + } else if (mode_lib->vba.LinesForMetaAndDPTERow[k] > 0.0) { + mode_lib->vba.prefetch_row_bw[k] = (mode_lib->vba.MetaRowBytes[k] + + mode_lib->vba.DPTEBytesPerRow[k]) + / (mode_lib->vba.LinesForMetaAndDPTERow[k] + * mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k]); + } else { + mode_lib->vba.prefetch_row_bw[k] = 0.0; + mode_lib->vba.prefetch_row_bw_valid = false; + } + mode_lib->vba.MaximumReadBandwidthWithPrefetch = + mode_lib->vba.MaximumReadBandwidthWithPrefetch + + mode_lib->vba.cursor_bw[k] + + dml_max4( + mode_lib->vba.prefetch_vm_bw[k], + mode_lib->vba.prefetch_row_bw[k], + mode_lib->vba.ReadBandwidth[k], + mode_lib->vba.RequiredPrefetchPixelDataBW[i][k]); + } + mode_lib->vba.PrefetchSupported[i] = true; + if (mode_lib->vba.MaximumReadBandwidthWithPrefetch + > mode_lib->vba.ReturnBWPerState[i] + || mode_lib->vba.prefetch_vm_bw_valid == false + || mode_lib->vba.prefetch_row_bw_valid == false) { + mode_lib->vba.PrefetchSupported[i] = false; + } + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + if (mode_lib->vba.LineTimesForPrefetch[k] < 2.0 + || mode_lib->vba.LinesForMetaPTE[k] >= 8.0 + || mode_lib->vba.LinesForMetaAndDPTERow[k] >= 16.0 + || mode_lib->vba.IsErrorResult[i][k] == true) { + mode_lib->vba.PrefetchSupported[i] = false; + } + } + mode_lib->vba.VRatioInPrefetchSupported[i] = true; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + if (mode_lib->vba.VRatioPreY[i][k] > 4.0 + || mode_lib->vba.VRatioPreC[i][k] > 4.0 + || mode_lib->vba.IsErrorResult[i][k] == true) { + mode_lib->vba.VRatioInPrefetchSupported[i] = false; + } + } + if (mode_lib->vba.PrefetchSupported[i] == true + && mode_lib->vba.VRatioInPrefetchSupported[i] == true) { + mode_lib->vba.BandwidthAvailableForImmediateFlip = + mode_lib->vba.ReturnBWPerState[i]; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + mode_lib->vba.BandwidthAvailableForImmediateFlip = + mode_lib->vba.BandwidthAvailableForImmediateFlip + - mode_lib->vba.cursor_bw[k] + - dml_max( + mode_lib->vba.ReadBandwidth[k], + mode_lib->vba.PrefetchBW[k]); + } + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + mode_lib->vba.ImmediateFlipBytes[k] = 0.0; + if ((mode_lib->vba.SourcePixelFormat[k] != dm_420_8 + && mode_lib->vba.SourcePixelFormat[k] != dm_420_10)) { + mode_lib->vba.ImmediateFlipBytes[k] = + mode_lib->vba.PDEAndMetaPTEBytesPerFrame[k] + + mode_lib->vba.MetaRowBytes[k] + + mode_lib->vba.DPTEBytesPerRow[k]; + } + } + mode_lib->vba.TotImmediateFlipBytes = 0.0; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + if ((mode_lib->vba.SourcePixelFormat[k] != dm_420_8 + && mode_lib->vba.SourcePixelFormat[k] != dm_420_10)) { + mode_lib->vba.TotImmediateFlipBytes = + mode_lib->vba.TotImmediateFlipBytes + + mode_lib->vba.ImmediateFlipBytes[k]; + } + } + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + CalculateFlipSchedule( + mode_lib, + mode_lib->vba.ExtraLatency, + mode_lib->vba.UrgentLatency, + mode_lib->vba.MaxPageTableLevels, + mode_lib->vba.VirtualMemoryEnable, + mode_lib->vba.BandwidthAvailableForImmediateFlip, + mode_lib->vba.TotImmediateFlipBytes, + mode_lib->vba.SourcePixelFormat[k], + mode_lib->vba.ImmediateFlipBytes[k], + mode_lib->vba.HTotal[k] + / mode_lib->vba.PixelClock[k], + mode_lib->vba.VRatio[k], + mode_lib->vba.Tno_bw[k], + mode_lib->vba.PDEAndMetaPTEBytesPerFrame[k], + mode_lib->vba.MetaRowBytes[k], + mode_lib->vba.DPTEBytesPerRow[k], + mode_lib->vba.DCCEnable[k], + mode_lib->vba.dpte_row_height[k], + mode_lib->vba.meta_row_height[k], + mode_lib->vba.qual_row_bw[k], + &mode_lib->vba.DestinationLinesToRequestVMInImmediateFlip[k], + &mode_lib->vba.DestinationLinesToRequestRowInImmediateFlip[k], + &mode_lib->vba.final_flip_bw[k], + &mode_lib->vba.ImmediateFlipSupportedForPipe[k]); + } + mode_lib->vba.total_dcn_read_bw_with_flip = 0.0; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + mode_lib->vba.total_dcn_read_bw_with_flip = + mode_lib->vba.total_dcn_read_bw_with_flip + + mode_lib->vba.cursor_bw[k] + + dml_max3( + mode_lib->vba.prefetch_vm_bw[k], + mode_lib->vba.prefetch_row_bw[k], + mode_lib->vba.final_flip_bw[k] + + dml_max( + mode_lib->vba.ReadBandwidth[k], + mode_lib->vba.RequiredPrefetchPixelDataBW[i][k])); + } + mode_lib->vba.ImmediateFlipSupportedForState[i] = true; + if (mode_lib->vba.total_dcn_read_bw_with_flip + > mode_lib->vba.ReturnBWPerState[i]) { + mode_lib->vba.ImmediateFlipSupportedForState[i] = false; + } + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + if (mode_lib->vba.ImmediateFlipSupportedForPipe[k] == false) { + mode_lib->vba.ImmediateFlipSupportedForState[i] = false; + } + } + } else { + mode_lib->vba.ImmediateFlipSupportedForState[i] = false; + } + } + /*PTE Buffer Size Check*/ + + for (i = 0; i <= DC__VOLTAGE_STATES; i++) { + mode_lib->vba.PTEBufferSizeNotExceeded[i] = true; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + if (mode_lib->vba.PTEBufferSizeNotExceededY[i][k] == false + || mode_lib->vba.PTEBufferSizeNotExceededC[i][k] == false) { + mode_lib->vba.PTEBufferSizeNotExceeded[i] = false; + } + } + } + /*Cursor Support Check*/ + + mode_lib->vba.CursorSupport = true; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + if (mode_lib->vba.CursorWidth[k][0] > 0.0) { + if (dml_floor( + dml_floor( + mode_lib->vba.CursorBufferSize + - mode_lib->vba.CursorChunkSize, + mode_lib->vba.CursorChunkSize) * 1024.0 + / (mode_lib->vba.CursorWidth[k][0] + * mode_lib->vba.CursorBPP[k][0] + / 8.0), + 1.0) + * (mode_lib->vba.HTotal[k] / mode_lib->vba.PixelClock[k]) + / mode_lib->vba.VRatio[k] < mode_lib->vba.UrgentLatency + || (mode_lib->vba.CursorBPP[k][0] == 64.0 + && mode_lib->vba.Cursor64BppSupport == false)) { + mode_lib->vba.CursorSupport = false; + } + } + } + /*Valid Pitch Check*/ + + mode_lib->vba.PitchSupport = true; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + mode_lib->vba.AlignedYPitch[k] = dml_ceil( + dml_max(mode_lib->vba.PitchY[k], mode_lib->vba.ViewportWidth[k]), + mode_lib->vba.MacroTileWidthY[k]); + if (mode_lib->vba.AlignedYPitch[k] > mode_lib->vba.PitchY[k]) { + mode_lib->vba.PitchSupport = false; + } + if (mode_lib->vba.DCCEnable[k] == true) { + mode_lib->vba.AlignedDCCMetaPitch[k] = dml_ceil( + dml_max( + mode_lib->vba.DCCMetaPitchY[k], + mode_lib->vba.ViewportWidth[k]), + 64.0 * mode_lib->vba.Read256BlockWidthY[k]); + } else { + mode_lib->vba.AlignedDCCMetaPitch[k] = mode_lib->vba.DCCMetaPitchY[k]; + } + if (mode_lib->vba.AlignedDCCMetaPitch[k] > mode_lib->vba.DCCMetaPitchY[k]) { + mode_lib->vba.PitchSupport = false; + } + if (mode_lib->vba.SourcePixelFormat[k] != dm_444_64 + && mode_lib->vba.SourcePixelFormat[k] != dm_444_32 + && mode_lib->vba.SourcePixelFormat[k] != dm_444_16 + && mode_lib->vba.SourcePixelFormat[k] != dm_mono_16 + && mode_lib->vba.SourcePixelFormat[k] != dm_mono_8) { + mode_lib->vba.AlignedCPitch[k] = dml_ceil( + dml_max( + mode_lib->vba.PitchC[k], + mode_lib->vba.ViewportWidth[k] / 2.0), + mode_lib->vba.MacroTileWidthC[k]); + } else { + mode_lib->vba.AlignedCPitch[k] = mode_lib->vba.PitchC[k]; + } + if (mode_lib->vba.AlignedCPitch[k] > mode_lib->vba.PitchC[k]) { + mode_lib->vba.PitchSupport = false; + } + } + /*Mode Support, Voltage State and SOC Configuration*/ + + for (i = DC__VOLTAGE_STATES; i >= 0; i--) { + if (mode_lib->vba.ScaleRatioAndTapsSupport == true + && mode_lib->vba.SourceFormatPixelAndScanSupport == true + && mode_lib->vba.ViewportSizeSupport[i] == true + && mode_lib->vba.BandwidthSupport[i] == true + && mode_lib->vba.DIOSupport[i] == true + && mode_lib->vba.NotEnoughDSCUnits[i] == false + && mode_lib->vba.DSCCLKRequiredMoreThanSupported[i] == false + && mode_lib->vba.UrgentLatencySupport[i] == true + && mode_lib->vba.ROBSupport[i] == true + && mode_lib->vba.DISPCLK_DPPCLK_Support[i] == true + && mode_lib->vba.TotalAvailablePipesSupport[i] == true + && mode_lib->vba.NumberOfOTGSupport == true + && mode_lib->vba.WritebackModeSupport == true + && mode_lib->vba.WritebackLatencySupport == true + && mode_lib->vba.WritebackScaleRatioAndTapsSupport == true + && mode_lib->vba.CursorSupport == true + && mode_lib->vba.PitchSupport == true + && mode_lib->vba.PrefetchSupported[i] == true + && mode_lib->vba.VRatioInPrefetchSupported[i] == true + && mode_lib->vba.PTEBufferSizeNotExceeded[i] == true + && mode_lib->vba.NonsupportedDSCInputBPC == false) { + mode_lib->vba.ModeSupport[i] = true; + } else { + mode_lib->vba.ModeSupport[i] = false; + } + } + for (i = DC__VOLTAGE_STATES; i >= 0; i--) { + if (i == DC__VOLTAGE_STATES || mode_lib->vba.ModeSupport[i] == true) { + mode_lib->vba.VoltageLevel = i; + } + } + mode_lib->vba.DCFCLK = mode_lib->vba.DCFCLKPerState[mode_lib->vba.VoltageLevel]; + mode_lib->vba.DRAMSpeed = mode_lib->vba.DRAMSpeedPerState[mode_lib->vba.VoltageLevel]; + mode_lib->vba.FabricClock = mode_lib->vba.FabricClockPerState[mode_lib->vba.VoltageLevel]; + mode_lib->vba.SOCCLK = mode_lib->vba.SOCCLKPerState[mode_lib->vba.VoltageLevel]; + mode_lib->vba.FabricAndDRAMBandwidth = + mode_lib->vba.FabricAndDRAMBandwidthPerState[mode_lib->vba.VoltageLevel]; + mode_lib->vba.ImmediateFlipSupport = + mode_lib->vba.ImmediateFlipSupportedForState[mode_lib->vba.VoltageLevel]; + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + mode_lib->vba.DPPPerPlane[k] = mode_lib->vba.NoOfDPP[mode_lib->vba.VoltageLevel][k]; + } + for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) { + if (mode_lib->vba.BlendingAndTiming[k] == k) { + mode_lib->vba.ODMCombineEnabled[k] = + mode_lib->vba.ODMCombineEnablePerState[mode_lib->vba.VoltageLevel][k]; + } else { + mode_lib->vba.ODMCombineEnabled[k] = 0; + } + mode_lib->vba.DSCEnabled[k] = + mode_lib->vba.RequiresDSC[mode_lib->vba.VoltageLevel][k]; + mode_lib->vba.OutputBpp[k] = + mode_lib->vba.OutputBppPerState[mode_lib->vba.VoltageLevel][k]; + } +} diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.h b/drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.h new file mode 100644 index 000000000000..4112409cd974 --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.h @@ -0,0 +1,598 @@ +/* + * Copyright 2017 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 __DML2_DISPLAY_MODE_VBA_H__ +#define __DML2_DISPLAY_MODE_VBA_H__ + +#include "dml_common_defs.h" + +struct display_mode_lib; + +void set_prefetch_mode(struct display_mode_lib *mode_lib, + bool cstate_en, + bool pstate_en, + bool ignore_viewport_pos, + bool immediate_flip_support); + +#define dml_get_attr_decl(attr) double get_##attr(struct display_mode_lib *mode_lib, const display_e2e_pipe_params_st *pipes, unsigned int num_pipes) + +dml_get_attr_decl(clk_dcf_deepsleep); +dml_get_attr_decl(wm_urgent); +dml_get_attr_decl(wm_memory_trip); +dml_get_attr_decl(wm_writeback_urgent); +dml_get_attr_decl(wm_stutter_exit); +dml_get_attr_decl(wm_stutter_enter_exit); +dml_get_attr_decl(wm_dram_clock_change); +dml_get_attr_decl(wm_writeback_dram_clock_change); +dml_get_attr_decl(wm_xfc_underflow); +dml_get_attr_decl(stutter_efficiency_no_vblank); +dml_get_attr_decl(stutter_efficiency); +dml_get_attr_decl(urgent_latency); +dml_get_attr_decl(urgent_extra_latency); +dml_get_attr_decl(nonurgent_latency); +dml_get_attr_decl(dram_clock_change_latency); +dml_get_attr_decl(dispclk_calculated); +dml_get_attr_decl(total_data_read_bw); +dml_get_attr_decl(return_bw); +dml_get_attr_decl(tcalc); + +#define dml_get_pipe_attr_decl(attr) double get_##attr(struct display_mode_lib *mode_lib, const display_e2e_pipe_params_st *pipes, unsigned int num_pipes, unsigned int which_pipe) + +dml_get_pipe_attr_decl(dsc_delay); +dml_get_pipe_attr_decl(dppclk_calculated); +dml_get_pipe_attr_decl(dscclk_calculated); +dml_get_pipe_attr_decl(min_ttu_vblank); +dml_get_pipe_attr_decl(vratio_prefetch_l); +dml_get_pipe_attr_decl(vratio_prefetch_c); +dml_get_pipe_attr_decl(dst_x_after_scaler); +dml_get_pipe_attr_decl(dst_y_after_scaler); +dml_get_pipe_attr_decl(dst_y_per_vm_vblank); +dml_get_pipe_attr_decl(dst_y_per_row_vblank); +dml_get_pipe_attr_decl(dst_y_prefetch); +dml_get_pipe_attr_decl(dst_y_per_vm_flip); +dml_get_pipe_attr_decl(dst_y_per_row_flip); +dml_get_pipe_attr_decl(xfc_transfer_delay); +dml_get_pipe_attr_decl(xfc_precharge_delay); +dml_get_pipe_attr_decl(xfc_remote_surface_flip_latency); +dml_get_pipe_attr_decl(xfc_prefetch_margin); + +unsigned int get_vstartup_calculated( + struct display_mode_lib *mode_lib, + const display_e2e_pipe_params_st *pipes, + unsigned int num_pipes, + unsigned int which_pipe); + +double get_total_immediate_flip_bytes( + struct display_mode_lib *mode_lib, + const display_e2e_pipe_params_st *pipes, + unsigned int num_pipes); +double get_total_immediate_flip_bw( + struct display_mode_lib *mode_lib, + const display_e2e_pipe_params_st *pipes, + unsigned int num_pipes); +double get_total_prefetch_bw( + struct display_mode_lib *mode_lib, + const display_e2e_pipe_params_st *pipes, + unsigned int num_pipes); + +unsigned int dml_get_voltage_level( + struct display_mode_lib *mode_lib, + const display_e2e_pipe_params_st *pipes, + unsigned int num_pipes); + +bool Calculate256BBlockSizes( + enum source_format_class SourcePixelFormat, + enum dm_swizzle_mode SurfaceTiling, + unsigned int BytePerPixelY, + unsigned int BytePerPixelC, + unsigned int *BlockHeight256BytesY, + unsigned int *BlockHeight256BytesC, + unsigned int *BlockWidth256BytesY, + unsigned int *BlockWidth256BytesC); + + +struct vba_vars_st { + ip_params_st ip; + soc_bounding_box_st soc; + + unsigned int MaximumMaxVStartupLines; + double cursor_bw[DC__NUM_DPP__MAX]; + double meta_row_bw[DC__NUM_DPP__MAX]; + double dpte_row_bw[DC__NUM_DPP__MAX]; + double qual_row_bw[DC__NUM_DPP__MAX]; + double WritebackDISPCLK; + double PSCL_THROUGHPUT_LUMA[DC__NUM_DPP__MAX]; + double PSCL_THROUGHPUT_CHROMA[DC__NUM_DPP__MAX]; + double DPPCLKUsingSingleDPPLuma; + double DPPCLKUsingSingleDPPChroma; + double DPPCLKUsingSingleDPP[DC__NUM_DPP__MAX]; + double DISPCLKWithRamping; + double DISPCLKWithoutRamping; + double GlobalDPPCLK; + double DISPCLKWithRampingRoundedToDFSGranularity; + double DISPCLKWithoutRampingRoundedToDFSGranularity; + double MaxDispclkRoundedToDFSGranularity; + bool DCCEnabledAnyPlane; + double ReturnBandwidthToDCN; + unsigned int SwathWidthY[DC__NUM_DPP__MAX]; + unsigned int SwathWidthSingleDPPY[DC__NUM_DPP__MAX]; + double BytePerPixelDETY[DC__NUM_DPP__MAX]; + double BytePerPixelDETC[DC__NUM_DPP__MAX]; + double ReadBandwidthPlaneLuma[DC__NUM_DPP__MAX]; + double ReadBandwidthPlaneChroma[DC__NUM_DPP__MAX]; + unsigned int TotalActiveDPP; + unsigned int TotalDCCActiveDPP; + double UrgentRoundTripAndOutOfOrderLatency; + double DisplayPipeLineDeliveryTimeLuma[DC__NUM_DPP__MAX]; // WM + double DisplayPipeLineDeliveryTimeChroma[DC__NUM_DPP__MAX]; // WM + double LinesInDETY[DC__NUM_DPP__MAX]; // WM + double LinesInDETC[DC__NUM_DPP__MAX]; // WM + unsigned int LinesInDETYRoundedDownToSwath[DC__NUM_DPP__MAX]; // WM + unsigned int LinesInDETCRoundedDownToSwath[DC__NUM_DPP__MAX]; // WM + double FullDETBufferingTimeY[DC__NUM_DPP__MAX]; // WM + double FullDETBufferingTimeC[DC__NUM_DPP__MAX]; // WM + double MinFullDETBufferingTime; + double FrameTimeForMinFullDETBufferingTime; + double AverageReadBandwidthGBytePerSecond; + double PartOfBurstThatFitsInROB; + double StutterBurstTime; + //unsigned int NextPrefetchMode; + double VBlankTime; + double SmallestVBlank; + double DCFCLKDeepSleepPerPlane; + double EffectiveDETPlusLBLinesLuma; + double EffectiveDETPlusLBLinesChroma; + double UrgentLatencySupportUsLuma; + double UrgentLatencySupportUsChroma; + double UrgentLatencySupportUs[DC__NUM_DPP__MAX]; + unsigned int DSCFormatFactor; + unsigned int BlockHeight256BytesY[DC__NUM_DPP__MAX]; + unsigned int BlockHeight256BytesC[DC__NUM_DPP__MAX]; + unsigned int BlockWidth256BytesY[DC__NUM_DPP__MAX]; + unsigned int BlockWidth256BytesC[DC__NUM_DPP__MAX]; + double VInitPreFillY[DC__NUM_DPP__MAX]; + double VInitPreFillC[DC__NUM_DPP__MAX]; + unsigned int MaxNumSwathY[DC__NUM_DPP__MAX]; + unsigned int MaxNumSwathC[DC__NUM_DPP__MAX]; + double PrefetchSourceLinesY[DC__NUM_DPP__MAX]; + double PrefetchSourceLinesC[DC__NUM_DPP__MAX]; + double PixelPTEBytesPerRow[DC__NUM_DPP__MAX]; + double MetaRowByte[DC__NUM_DPP__MAX]; + unsigned int dpte_row_height[DC__NUM_DPP__MAX]; + unsigned int dpte_row_height_chroma[DC__NUM_DPP__MAX]; + unsigned int meta_row_height[DC__NUM_DPP__MAX]; + unsigned int meta_row_height_chroma[DC__NUM_DPP__MAX]; + + unsigned int MacroTileWidthY[DC__NUM_DPP__MAX]; + unsigned int MacroTileWidthC[DC__NUM_DPP__MAX]; + unsigned int MaxVStartupLines[DC__NUM_DPP__MAX]; + double WritebackDelay[DC__VOLTAGE_STATES + 1][DC__NUM_DPP__MAX]; + bool PrefetchModeSupported; + bool AllowDRAMClockChangeDuringVBlank[DC__NUM_DPP__MAX]; + bool AllowDRAMSelfRefreshDuringVBlank[DC__NUM_DPP__MAX]; + double RequiredPrefetchPixDataBW[DC__NUM_DPP__MAX]; + double XFCRemoteSurfaceFlipDelay; + double TInitXFill; + double TslvChk; + double SrcActiveDrainRate; + double Tno_bw[DC__NUM_DPP__MAX]; + bool ImmediateFlipSupported; + + double prefetch_vm_bw[DC__NUM_DPP__MAX]; + double prefetch_row_bw[DC__NUM_DPP__MAX]; + bool ImmediateFlipSupportedForPipe[DC__NUM_DPP__MAX]; + unsigned int VStartupLines; + double DisplayPipeLineDeliveryTimeLumaPrefetch[DC__NUM_DPP__MAX]; + double DisplayPipeLineDeliveryTimeChromaPrefetch[DC__NUM_DPP__MAX]; + unsigned int ActiveDPPs; + unsigned int LBLatencyHidingSourceLinesY; + unsigned int LBLatencyHidingSourceLinesC; + double ActiveDRAMClockChangeLatencyMargin[DC__NUM_DPP__MAX]; + double MinActiveDRAMClockChangeMargin; + double XFCSlaveVUpdateOffset[DC__NUM_DPP__MAX]; + double XFCSlaveVupdateWidth[DC__NUM_DPP__MAX]; + double XFCSlaveVReadyOffset[DC__NUM_DPP__MAX]; + double InitFillLevel; + double FinalFillMargin; + double FinalFillLevel; + double RemainingFillLevel; + double TFinalxFill; + + + // + // SOC Bounding Box Parameters + // + double SRExitTime; + double SREnterPlusExitTime; + double UrgentLatency; + double WritebackLatency; + double PercentOfIdealDRAMAndFabricBWReceivedAfterUrgLatency; + double NumberOfChannels; + double DRAMChannelWidth; + double FabricDatapathToDCNDataReturn; + double ReturnBusWidth; + double Downspreading; + double DISPCLKDPPCLKDSCCLKDownSpreading; + double DISPCLKDPPCLKVCOSpeed; + double RoundTripPingLatencyCycles; + double UrgentOutOfOrderReturnPerChannel; + unsigned int VMMPageSize; + double DRAMClockChangeLatency; + double XFCBusTransportTime; + double XFCXBUFLatencyTolerance; + + // + // IP Parameters + // + unsigned int ROBBufferSizeInKByte; + double DETBufferSizeInKByte; + unsigned int DPPOutputBufferPixels; + unsigned int OPPOutputBufferLines; + unsigned int PixelChunkSizeInKByte; + double ReturnBW; + bool VirtualMemoryEnable; + unsigned int MaxPageTableLevels; + unsigned int OverridePageTableLevels; + unsigned int PTEChunkSize; + unsigned int MetaChunkSize; + unsigned int WritebackChunkSize; + bool ODMCapability; + unsigned int NumberOfDSC; + unsigned int LineBufferSize; + unsigned int MaxLineBufferLines; + unsigned int WritebackInterfaceLumaBufferSize; + unsigned int WritebackInterfaceChromaBufferSize; + unsigned int WritebackChromaLineBufferWidth; + double MaxDCHUBToPSCLThroughput; + double MaxPSCLToLBThroughput; + unsigned int PTEBufferSizeInRequests; + double DISPCLKRampingMargin; + unsigned int MaxInterDCNTileRepeaters; + bool XFCSupported; + double XFCSlvChunkSize; + double XFCFillBWOverhead; + double XFCFillConstant; + double XFCTSlvVupdateOffset; + double XFCTSlvVupdateWidth; + double XFCTSlvVreadyOffset; + double DPPCLKDelaySubtotal; + double DPPCLKDelaySCL; + double DPPCLKDelaySCLLBOnly; + double DPPCLKDelayCNVCFormater; + double DPPCLKDelayCNVCCursor; + double DISPCLKDelaySubtotal; + bool ProgressiveToInterlaceUnitInOPP; + unsigned int PDEProcessingBufIn64KBReqs; + + // Pipe/Plane Parameters + int VoltageLevel; + double FabricAndDRAMBandwidth; + double FabricClock; + double DRAMSpeed; + double DISPCLK; + double SOCCLK; + double DCFCLK; + + unsigned int NumberOfActivePlanes; + unsigned int ViewportWidth[DC__NUM_DPP__MAX]; + unsigned int ViewportHeight[DC__NUM_DPP__MAX]; + unsigned int ViewportYStartY[DC__NUM_DPP__MAX]; + unsigned int ViewportYStartC[DC__NUM_DPP__MAX]; + unsigned int PitchY[DC__NUM_DPP__MAX]; + unsigned int PitchC[DC__NUM_DPP__MAX]; + double HRatio[DC__NUM_DPP__MAX]; + double VRatio[DC__NUM_DPP__MAX]; + unsigned int htaps[DC__NUM_DPP__MAX]; + unsigned int vtaps[DC__NUM_DPP__MAX]; + unsigned int HTAPsChroma[DC__NUM_DPP__MAX]; + unsigned int VTAPsChroma[DC__NUM_DPP__MAX]; + unsigned int HTotal[DC__NUM_DPP__MAX]; + unsigned int VTotal[DC__NUM_DPP__MAX]; + unsigned int DPPPerPlane[DC__NUM_DPP__MAX]; + double PixelClock[DC__NUM_DPP__MAX]; + double PixelClockBackEnd[DC__NUM_DPP__MAX]; + double DPPCLK[DC__NUM_DPP__MAX]; + bool DCCEnable[DC__NUM_DPP__MAX]; + unsigned int DCCMetaPitchY[DC__NUM_DPP__MAX]; + enum scan_direction_class SourceScan[DC__NUM_DPP__MAX]; + enum source_format_class SourcePixelFormat[DC__NUM_DPP__MAX]; + bool WritebackEnable[DC__NUM_DPP__MAX]; + double WritebackDestinationWidth[DC__NUM_DPP__MAX]; + double WritebackDestinationHeight[DC__NUM_DPP__MAX]; + double WritebackSourceHeight[DC__NUM_DPP__MAX]; + enum source_format_class WritebackPixelFormat[DC__NUM_DPP__MAX]; + unsigned int WritebackLumaHTaps[DC__NUM_DPP__MAX]; + unsigned int WritebackLumaVTaps[DC__NUM_DPP__MAX]; + unsigned int WritebackChromaHTaps[DC__NUM_DPP__MAX]; + unsigned int WritebackChromaVTaps[DC__NUM_DPP__MAX]; + double WritebackHRatio[DC__NUM_DPP__MAX]; + double WritebackVRatio[DC__NUM_DPP__MAX]; + unsigned int HActive[DC__NUM_DPP__MAX]; + unsigned int VActive[DC__NUM_DPP__MAX]; + bool Interlace[DC__NUM_DPP__MAX]; + enum dm_swizzle_mode SurfaceTiling[DC__NUM_DPP__MAX]; + unsigned int ScalerRecoutWidth[DC__NUM_DPP__MAX]; + bool DynamicMetadataEnable[DC__NUM_DPP__MAX]; + unsigned int DynamicMetadataLinesBeforeActiveRequired[DC__NUM_DPP__MAX]; + unsigned int DynamicMetadataTransmittedBytes[DC__NUM_DPP__MAX]; + double DCCRate[DC__NUM_DPP__MAX]; + bool ODMCombineEnabled[DC__NUM_DPP__MAX]; + double OutputBpp[DC__NUM_DPP__MAX]; + unsigned int NumberOfDSCSlices[DC__NUM_DPP__MAX]; + bool DSCEnabled[DC__NUM_DPP__MAX]; + unsigned int DSCDelay[DC__NUM_DPP__MAX]; + unsigned int DSCInputBitPerComponent[DC__NUM_DPP__MAX]; + enum output_format_class OutputFormat[DC__NUM_DPP__MAX]; + enum output_encoder_class Output[DC__NUM_DPP__MAX]; + unsigned int BlendingAndTiming[DC__NUM_DPP__MAX]; + bool SynchronizedVBlank; + unsigned int NumberOfCursors[DC__NUM_DPP__MAX]; + unsigned int CursorWidth[DC__NUM_DPP__MAX][DC__NUM_CURSOR__MAX]; + unsigned int CursorBPP[DC__NUM_DPP__MAX][DC__NUM_CURSOR__MAX]; + bool XFCEnabled[DC__NUM_DPP__MAX]; + bool ScalerEnabled[DC__NUM_DPP__MAX]; + + // Intermediates/Informational + bool ImmediateFlipSupport; + unsigned int SwathHeightY[DC__NUM_DPP__MAX]; + unsigned int SwathHeightC[DC__NUM_DPP__MAX]; + unsigned int DETBufferSizeY[DC__NUM_DPP__MAX]; + unsigned int DETBufferSizeC[DC__NUM_DPP__MAX]; + unsigned int LBBitPerPixel[DC__NUM_DPP__MAX]; + double LastPixelOfLineExtraWatermark; + double TotalDataReadBandwidth; + unsigned int TotalActiveWriteback; + unsigned int EffectiveLBLatencyHidingSourceLinesLuma; + unsigned int EffectiveLBLatencyHidingSourceLinesChroma; + double BandwidthAvailableForImmediateFlip; + unsigned int PrefetchMode; + bool IgnoreViewportPositioning; + double PrefetchBandwidth[DC__NUM_DPP__MAX]; + bool ErrorResult[DC__NUM_DPP__MAX]; + double PDEAndMetaPTEBytesFrame[DC__NUM_DPP__MAX]; + + // + // Calculated dml_ml->vba.Outputs + // + double DCFClkDeepSleep; + double UrgentWatermark; + double UrgentExtraLatency; + double MemoryTripWatermark; + double WritebackUrgentWatermark; + double StutterExitWatermark; + double StutterEnterPlusExitWatermark; + double DRAMClockChangeWatermark; + double WritebackDRAMClockChangeWatermark; + double StutterEfficiency; + double StutterEfficiencyNotIncludingVBlank; + double MinUrgentLatencySupportUs; + double NonUrgentLatencyTolerance; + double MinActiveDRAMClockChangeLatencySupported; + enum clock_change_support DRAMClockChangeSupport; + + // These are the clocks calcuated by the library but they are not actually + // used explicitly. They are fetched by tests and then possibly used. The + // ultimate values to use are the ones specified by the parameters to DML + double DISPCLK_calculated; + double DSCCLK_calculated[DC__NUM_DPP__MAX]; + double DPPCLK_calculated[DC__NUM_DPP__MAX]; + + unsigned int VStartup[DC__NUM_DPP__MAX]; + unsigned int VUpdateOffsetPix[DC__NUM_DPP__MAX]; + unsigned int VUpdateWidthPix[DC__NUM_DPP__MAX]; + unsigned int VReadyOffsetPix[DC__NUM_DPP__MAX]; + unsigned int VStartupRequiredWhenNotEnoughTimeForDynamicMetadata; + + double ImmediateFlipBW; + unsigned int TotImmediateFlipBytes; + double TCalc; + double MinTTUVBlank[DC__NUM_DPP__MAX]; + double VRatioPrefetchY[DC__NUM_DPP__MAX]; + double VRatioPrefetchC[DC__NUM_DPP__MAX]; + double DSTXAfterScaler[DC__NUM_DPP__MAX]; + double DSTYAfterScaler[DC__NUM_DPP__MAX]; + + double DestinationLinesToRequestVMInVBlank[DC__NUM_DPP__MAX]; + double DestinationLinesToRequestRowInVBlank[DC__NUM_DPP__MAX]; + double DestinationLinesForPrefetch[DC__NUM_DPP__MAX]; + double DestinationLinesToRequestRowInImmediateFlip[DC__NUM_DPP__MAX]; + double DestinationLinesToRequestVMInImmediateFlip[DC__NUM_DPP__MAX]; + + double XFCTransferDelay[DC__NUM_DPP__MAX]; + double XFCPrechargeDelay[DC__NUM_DPP__MAX]; + double XFCRemoteSurfaceFlipLatency[DC__NUM_DPP__MAX]; + double XFCPrefetchMargin[DC__NUM_DPP__MAX]; + + display_e2e_pipe_params_st cache_pipes[DC__NUM_DPP__MAX]; + unsigned int cache_num_pipes; + unsigned int pipe_plane[DC__NUM_DPP__MAX]; + + /* vba mode support */ + /*inputs*/ + bool SupportGFX7CompatibleTilingIn32bppAnd64bpp; + double MaxHSCLRatio; + double MaxVSCLRatio; + unsigned int MaxNumWriteback; + bool WritebackLumaAndChromaScalingSupported; + bool Cursor64BppSupport; + double DCFCLKPerState[DC__VOLTAGE_STATES + 1]; + double FabricClockPerState[DC__VOLTAGE_STATES + 1]; + double SOCCLKPerState[DC__VOLTAGE_STATES + 1]; + double PHYCLKPerState[DC__VOLTAGE_STATES + 1]; + double MaxDppclk[DC__VOLTAGE_STATES + 1]; + double MaxDSCCLK[DC__VOLTAGE_STATES + 1]; + double DRAMSpeedPerState[DC__VOLTAGE_STATES + 1]; + double MaxDispclk[DC__VOLTAGE_STATES + 1]; + + /*outputs*/ + bool ScaleRatioAndTapsSupport; + bool SourceFormatPixelAndScanSupport; + unsigned int SwathWidthYSingleDPP[DC__NUM_DPP__MAX]; + double BytePerPixelInDETY[DC__NUM_DPP__MAX]; + double BytePerPixelInDETC[DC__NUM_DPP__MAX]; + double TotalReadBandwidthConsumedGBytePerSecond; + double ReadBandwidth[DC__NUM_DPP__MAX]; + double TotalWriteBandwidthConsumedGBytePerSecond; + double WriteBandwidth[DC__NUM_DPP__MAX]; + double TotalBandwidthConsumedGBytePerSecond; + bool DCCEnabledInAnyPlane; + bool WritebackLatencySupport; + bool WritebackModeSupport; + bool Writeback10bpc420Supported; + bool BandwidthSupport[DC__VOLTAGE_STATES + 1]; + unsigned int TotalNumberOfActiveWriteback; + double CriticalPoint; + double ReturnBWToDCNPerState; + double FabricAndDRAMBandwidthPerState[DC__VOLTAGE_STATES + 1]; + double ReturnBWPerState[DC__VOLTAGE_STATES + 1]; + double UrgentRoundTripAndOutOfOrderLatencyPerState[DC__VOLTAGE_STATES + 1]; + bool ODMCombineEnablePerState[DC__VOLTAGE_STATES + 1][DC__NUM_DPP__MAX]; + bool PTEBufferSizeNotExceededY[DC__VOLTAGE_STATES + 1][DC__NUM_DPP__MAX]; + bool PTEBufferSizeNotExceededC[DC__VOLTAGE_STATES + 1][DC__NUM_DPP__MAX]; + bool PrefetchSupported[DC__VOLTAGE_STATES + 1]; + bool VRatioInPrefetchSupported[DC__VOLTAGE_STATES + 1]; + bool DISPCLK_DPPCLK_Support[DC__VOLTAGE_STATES + 1]; + bool TotalAvailablePipesSupport[DC__VOLTAGE_STATES + 1]; + bool UrgentLatencySupport[DC__VOLTAGE_STATES + 1]; + bool ModeSupport[DC__VOLTAGE_STATES + 1]; + bool DIOSupport[DC__VOLTAGE_STATES + 1]; + bool NotEnoughDSCUnits[DC__VOLTAGE_STATES + 1]; + bool DSCCLKRequiredMoreThanSupported[DC__VOLTAGE_STATES + 1]; + bool ROBSupport[DC__VOLTAGE_STATES + 1]; + bool PTEBufferSizeNotExceeded[DC__VOLTAGE_STATES + 1]; + bool RequiresDSC[DC__VOLTAGE_STATES + 1][DC__NUM_DPP__MAX]; + bool IsErrorResult[DC__VOLTAGE_STATES + 1][DC__NUM_DPP__MAX]; + bool ViewportSizeSupport[DC__VOLTAGE_STATES + 1]; + bool prefetch_vm_bw_valid; + bool prefetch_row_bw_valid; + bool NumberOfOTGSupport; + bool NonsupportedDSCInputBPC; + bool WritebackScaleRatioAndTapsSupport; + bool CursorSupport; + bool PitchSupport; + + double WritebackLineBufferLumaBufferSize; + double WritebackLineBufferChromaBufferSize; + double WritebackMinHSCLRatio; + double WritebackMinVSCLRatio; + double WritebackMaxHSCLRatio; + double WritebackMaxVSCLRatio; + double WritebackMaxHSCLTaps; + double WritebackMaxVSCLTaps; + unsigned int MaxNumDPP; + unsigned int MaxNumOTG; + double CursorBufferSize; + double CursorChunkSize; + unsigned int Mode; + unsigned int NoOfDPP[DC__VOLTAGE_STATES + 1][DC__NUM_DPP__MAX]; + double OutputLinkDPLanes[DC__NUM_DPP__MAX]; + double SwathWidthYPerState[DC__VOLTAGE_STATES + 1][DC__NUM_DPP__MAX]; + double SwathHeightYPerState[DC__VOLTAGE_STATES + 1][DC__NUM_DPP__MAX]; + double SwathHeightCPerState[DC__VOLTAGE_STATES + 1][DC__NUM_DPP__MAX]; + double UrgentLatencySupportUsPerState[DC__VOLTAGE_STATES + 1][DC__NUM_DPP__MAX]; + double VRatioPreY[DC__VOLTAGE_STATES + 1][DC__NUM_DPP__MAX]; + double VRatioPreC[DC__VOLTAGE_STATES + 1][DC__NUM_DPP__MAX]; + double RequiredPrefetchPixelDataBW[DC__VOLTAGE_STATES + 1][DC__NUM_DPP__MAX]; + double RequiredDPPCLK[DC__VOLTAGE_STATES + 1][DC__NUM_DPP__MAX]; + double RequiredDISPCLK[DC__VOLTAGE_STATES + 1]; + double TotalNumberOfActiveDPP[DC__VOLTAGE_STATES + 1]; + double TotalNumberOfDCCActiveDPP[DC__VOLTAGE_STATES + 1]; + double PrefetchBW[DC__NUM_DPP__MAX]; + double PDEAndMetaPTEBytesPerFrame[DC__NUM_DPP__MAX]; + double MetaRowBytes[DC__NUM_DPP__MAX]; + double DPTEBytesPerRow[DC__NUM_DPP__MAX]; + double PrefetchLinesY[DC__NUM_DPP__MAX]; + double PrefetchLinesC[DC__NUM_DPP__MAX]; + unsigned int MaxNumSwY[DC__NUM_DPP__MAX]; + unsigned int MaxNumSwC[DC__NUM_DPP__MAX]; + double PrefillY[DC__NUM_DPP__MAX]; + double PrefillC[DC__NUM_DPP__MAX]; + double LineTimesForPrefetch[DC__NUM_DPP__MAX]; + double LinesForMetaPTE[DC__NUM_DPP__MAX]; + double LinesForMetaAndDPTERow[DC__NUM_DPP__MAX]; + double MinDPPCLKUsingSingleDPP[DC__NUM_DPP__MAX]; + double RequiresFEC[DC__VOLTAGE_STATES + 1][DC__NUM_DPP__MAX]; + unsigned int OutputBppPerState[DC__VOLTAGE_STATES + 1][DC__NUM_DPP__MAX]; + double DSCDelayPerState[DC__VOLTAGE_STATES + 1][DC__NUM_DPP__MAX]; + unsigned int Read256BlockHeightY[DC__NUM_DPP__MAX]; + unsigned int Read256BlockWidthY[DC__NUM_DPP__MAX]; + unsigned int Read256BlockHeightC[DC__NUM_DPP__MAX]; + unsigned int Read256BlockWidthC[DC__NUM_DPP__MAX]; + unsigned int ImmediateFlipBytes[DC__NUM_DPP__MAX]; + double MaxSwathHeightY[DC__NUM_DPP__MAX]; + double MaxSwathHeightC[DC__NUM_DPP__MAX]; + double MinSwathHeightY[DC__NUM_DPP__MAX]; + double MinSwathHeightC[DC__NUM_DPP__MAX]; + double PSCL_FACTOR[DC__NUM_DPP__MAX]; + double PSCL_FACTOR_CHROMA[DC__NUM_DPP__MAX]; + double MaximumVStartup[DC__NUM_DPP__MAX]; + double AlignedDCCMetaPitch[DC__NUM_DPP__MAX]; + double AlignedYPitch[DC__NUM_DPP__MAX]; + double AlignedCPitch[DC__NUM_DPP__MAX]; + double MaximumSwathWidth[DC__NUM_DPP__MAX]; + double final_flip_bw[DC__NUM_DPP__MAX]; + double ImmediateFlipSupportedForState[DC__VOLTAGE_STATES + 1]; + + double WritebackLumaVExtra; + double WritebackChromaVExtra; + double WritebackRequiredDISPCLK; + double MaximumSwathWidthSupport; + double MaximumSwathWidthInDETBuffer; + double MaximumSwathWidthInLineBuffer; + double MaxDispclkRoundedDownToDFSGranularity; + double MaxDppclkRoundedDownToDFSGranularity; + double PlaneRequiredDISPCLKWithoutODMCombine; + double PlaneRequiredDISPCLK; + double TotalNumberOfActiveOTG; + double FECOverhead; + double EffectiveFECOverhead; + unsigned int Outbpp; + unsigned int OutbppDSC; + double TotalDSCUnitsRequired; + double bpp; + unsigned int slices; + double SwathWidthGranularityY; + double RoundedUpMaxSwathSizeBytesY; + double SwathWidthGranularityC; + double RoundedUpMaxSwathSizeBytesC; + double LinesInDETLuma; + double LinesInDETChroma; + double EffectiveDETLBLinesLuma; + double EffectiveDETLBLinesChroma; + double ProjectedDCFCLKDeepSleep; + double PDEAndMetaPTEBytesPerFrameY; + double PDEAndMetaPTEBytesPerFrameC; + unsigned int MetaRowBytesY; + unsigned int MetaRowBytesC; + unsigned int DPTEBytesPerRowC; + unsigned int DPTEBytesPerRowY; + double ExtraLatency; + double TimeCalc; + double TWait; + double MaximumReadBandwidthWithPrefetch; + double total_dcn_read_bw_with_flip; +}; + +#endif /* _DML2_DISPLAY_MODE_VBA_H_ */ diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_rq_dlg_calc.c b/drivers/gpu/drm/amd/display/dc/dml/display_rq_dlg_calc.c new file mode 100644 index 000000000000..8ba962df42e6 --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/dml/display_rq_dlg_calc.c @@ -0,0 +1,1763 @@ +/* + * Copyright 2017 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 "display_mode_lib.h" +#include "display_mode_vba.h" +#include "display_rq_dlg_calc.h" + +static void calculate_ttu_cursor(struct display_mode_lib *mode_lib, + double *refcyc_per_req_delivery_pre_cur, + double *refcyc_per_req_delivery_cur, + double refclk_freq_in_mhz, + double ref_freq_to_pix_freq, + double hscale_pixel_rate_l, + double hscl_ratio, + double vratio_pre_l, + double vratio_l, + unsigned int cur_width, + enum cursor_bpp cur_bpp); + +#include "dml_inline_defs.h" + +static unsigned int get_bytes_per_element(enum source_format_class source_format, bool is_chroma) +{ + unsigned int ret_val = 0; + + if (source_format == dm_444_16) { + if (!is_chroma) + ret_val = 2; + } else if (source_format == dm_444_32) { + if (!is_chroma) + ret_val = 4; + } else if (source_format == dm_444_64) { + if (!is_chroma) + ret_val = 8; + } else if (source_format == dm_420_8) { + if (is_chroma) + ret_val = 2; + else + ret_val = 1; + } else if (source_format == dm_420_10) { + if (is_chroma) + ret_val = 4; + else + ret_val = 2; + } else if (source_format == dm_444_8) { + ret_val = 1; + } + return ret_val; +} + +static bool is_dual_plane(enum source_format_class source_format) +{ + bool ret_val = 0; + + if ((source_format == dm_420_8) || (source_format == dm_420_10)) + ret_val = 1; + + return ret_val; +} + +static double get_refcyc_per_delivery(struct display_mode_lib *mode_lib, + double refclk_freq_in_mhz, + double pclk_freq_in_mhz, + bool odm_combine, + unsigned int recout_width, + unsigned int hactive, + double vratio, + double hscale_pixel_rate, + unsigned int delivery_width, + unsigned int req_per_swath_ub) +{ + double refcyc_per_delivery = 0.0; + + if (vratio <= 1.0) { + if (odm_combine) + refcyc_per_delivery = (double) refclk_freq_in_mhz + * dml_min((double) recout_width, (double) hactive / 2.0) + / pclk_freq_in_mhz / (double) req_per_swath_ub; + else + refcyc_per_delivery = (double) refclk_freq_in_mhz * (double) recout_width + / pclk_freq_in_mhz / (double) req_per_swath_ub; + } else { + refcyc_per_delivery = (double) refclk_freq_in_mhz * (double) delivery_width + / (double) hscale_pixel_rate / (double) req_per_swath_ub; + } + + dml_print("DML_DLG: %s: refclk_freq_in_mhz = %3.2f\n", __func__, refclk_freq_in_mhz); + dml_print("DML_DLG: %s: pclk_freq_in_mhz = %3.2f\n", __func__, pclk_freq_in_mhz); + dml_print("DML_DLG: %s: recout_width = %d\n", __func__, recout_width); + dml_print("DML_DLG: %s: vratio = %3.2f\n", __func__, vratio); + dml_print("DML_DLG: %s: req_per_swath_ub = %d\n", __func__, req_per_swath_ub); + dml_print("DML_DLG: %s: refcyc_per_delivery= %3.2f\n", __func__, refcyc_per_delivery); + + return refcyc_per_delivery; + +} + +static unsigned int get_blk_size_bytes(const enum source_macro_tile_size tile_size) +{ + if (tile_size == dm_256k_tile) + return (256 * 1024); + else if (tile_size == dm_64k_tile) + return (64 * 1024); + else + return (4 * 1024); +} + +static void extract_rq_sizing_regs(struct display_mode_lib *mode_lib, + display_data_rq_regs_st *rq_regs, + const display_data_rq_sizing_params_st rq_sizing) +{ + dml_print("DML_DLG: %s: rq_sizing param\n", __func__); + print__data_rq_sizing_params_st(mode_lib, rq_sizing); + + rq_regs->chunk_size = dml_log2(rq_sizing.chunk_bytes) - 10; + + if (rq_sizing.min_chunk_bytes == 0) + rq_regs->min_chunk_size = 0; + else + rq_regs->min_chunk_size = dml_log2(rq_sizing.min_chunk_bytes) - 8 + 1; + + rq_regs->meta_chunk_size = dml_log2(rq_sizing.meta_chunk_bytes) - 10; + if (rq_sizing.min_meta_chunk_bytes == 0) + rq_regs->min_meta_chunk_size = 0; + else + rq_regs->min_meta_chunk_size = dml_log2(rq_sizing.min_meta_chunk_bytes) - 6 + 1; + + rq_regs->dpte_group_size = dml_log2(rq_sizing.dpte_group_bytes) - 6; + rq_regs->mpte_group_size = dml_log2(rq_sizing.mpte_group_bytes) - 6; +} + +static void extract_rq_regs(struct display_mode_lib *mode_lib, + display_rq_regs_st *rq_regs, + const display_rq_params_st rq_param) +{ + unsigned int detile_buf_size_in_bytes = mode_lib->ip.det_buffer_size_kbytes * 1024; + unsigned int detile_buf_plane1_addr = 0; + + extract_rq_sizing_regs(mode_lib, &(rq_regs->rq_regs_l), rq_param.sizing.rq_l); + + rq_regs->rq_regs_l.pte_row_height_linear = dml_floor(dml_log2(rq_param.dlg.rq_l.dpte_row_height), + 1) - 3; + + if (rq_param.yuv420) { + extract_rq_sizing_regs(mode_lib, &(rq_regs->rq_regs_c), rq_param.sizing.rq_c); + rq_regs->rq_regs_c.pte_row_height_linear = dml_floor(dml_log2(rq_param.dlg.rq_c.dpte_row_height), + 1) - 3; + } + + rq_regs->rq_regs_l.swath_height = dml_log2(rq_param.dlg.rq_l.swath_height); + rq_regs->rq_regs_c.swath_height = dml_log2(rq_param.dlg.rq_c.swath_height); + + // FIXME: take the max between luma, chroma chunk size? + // okay for now, as we are setting chunk_bytes to 8kb anyways + if (rq_param.sizing.rq_l.chunk_bytes >= 32 * 1024) { //32kb + rq_regs->drq_expansion_mode = 0; + } else { + rq_regs->drq_expansion_mode = 2; + } + rq_regs->prq_expansion_mode = 1; + rq_regs->mrq_expansion_mode = 1; + rq_regs->crq_expansion_mode = 1; + + if (rq_param.yuv420) { + if ((double) rq_param.misc.rq_l.stored_swath_bytes + / (double) rq_param.misc.rq_c.stored_swath_bytes <= 1.5) { + detile_buf_plane1_addr = (detile_buf_size_in_bytes / 2.0 / 64.0); // half to chroma + } else { + detile_buf_plane1_addr = dml_round_to_multiple((unsigned int) ((2.0 * detile_buf_size_in_bytes) / 3.0), + 256, + 0) / 64.0; // 2/3 to chroma + } + } + rq_regs->plane1_base_address = detile_buf_plane1_addr; +} + +static void handle_det_buf_split(struct display_mode_lib *mode_lib, + display_rq_params_st *rq_param, + const display_pipe_source_params_st pipe_src_param) +{ + unsigned int total_swath_bytes = 0; + unsigned int swath_bytes_l = 0; + unsigned int swath_bytes_c = 0; + unsigned int full_swath_bytes_packed_l = 0; + unsigned int full_swath_bytes_packed_c = 0; + bool req128_l = 0; + bool req128_c = 0; + bool surf_linear = (pipe_src_param.sw_mode == dm_sw_linear); + bool surf_vert = (pipe_src_param.source_scan == dm_vert); + unsigned int log2_swath_height_l = 0; + unsigned int log2_swath_height_c = 0; + unsigned int detile_buf_size_in_bytes = mode_lib->ip.det_buffer_size_kbytes * 1024; + + full_swath_bytes_packed_l = rq_param->misc.rq_l.full_swath_bytes; + full_swath_bytes_packed_c = rq_param->misc.rq_c.full_swath_bytes; + + if (rq_param->yuv420_10bpc) { + full_swath_bytes_packed_l = dml_round_to_multiple(rq_param->misc.rq_l.full_swath_bytes * 2 / 3, + 256, + 1) + 256; + full_swath_bytes_packed_c = dml_round_to_multiple(rq_param->misc.rq_c.full_swath_bytes * 2 / 3, + 256, + 1) + 256; + } + + if (rq_param->yuv420) { + total_swath_bytes = 2 * full_swath_bytes_packed_l + 2 * full_swath_bytes_packed_c; + + if (total_swath_bytes <= detile_buf_size_in_bytes) { //full 256b request + req128_l = 0; + req128_c = 0; + swath_bytes_l = full_swath_bytes_packed_l; + swath_bytes_c = full_swath_bytes_packed_c; + } else { //128b request (for luma only for yuv420 8bpc) + req128_l = 1; + req128_c = 0; + swath_bytes_l = full_swath_bytes_packed_l / 2; + swath_bytes_c = full_swath_bytes_packed_c; + } + // Note: assumption, the config that pass in will fit into + // the detiled buffer. + } else { + total_swath_bytes = 2 * full_swath_bytes_packed_l; + + if (total_swath_bytes <= detile_buf_size_in_bytes) + req128_l = 0; + else + req128_l = 1; + + swath_bytes_l = total_swath_bytes; + swath_bytes_c = 0; + } + rq_param->misc.rq_l.stored_swath_bytes = swath_bytes_l; + rq_param->misc.rq_c.stored_swath_bytes = swath_bytes_c; + + if (surf_linear) { + log2_swath_height_l = 0; + log2_swath_height_c = 0; + } else if (!surf_vert) { + log2_swath_height_l = dml_log2(rq_param->misc.rq_l.blk256_height) - req128_l; + log2_swath_height_c = dml_log2(rq_param->misc.rq_c.blk256_height) - req128_c; + } else { + log2_swath_height_l = dml_log2(rq_param->misc.rq_l.blk256_width) - req128_l; + log2_swath_height_c = dml_log2(rq_param->misc.rq_c.blk256_width) - req128_c; + } + rq_param->dlg.rq_l.swath_height = 1 << log2_swath_height_l; + rq_param->dlg.rq_c.swath_height = 1 << log2_swath_height_c; + + dml_print("DML_DLG: %s: req128_l = %0d\n", __func__, req128_l); + dml_print("DML_DLG: %s: req128_c = %0d\n", __func__, req128_c); + dml_print("DML_DLG: %s: full_swath_bytes_packed_l = %0d\n", + __func__, + full_swath_bytes_packed_l); + dml_print("DML_DLG: %s: full_swath_bytes_packed_c = %0d\n", + __func__, + full_swath_bytes_packed_c); +} + +static void get_meta_and_pte_attr(struct display_mode_lib *mode_lib, + display_data_rq_dlg_params_st *rq_dlg_param, + display_data_rq_misc_params_st *rq_misc_param, + display_data_rq_sizing_params_st *rq_sizing_param, + unsigned int vp_width, + unsigned int vp_height, + unsigned int data_pitch, + unsigned int meta_pitch, + unsigned int source_format, + unsigned int tiling, + unsigned int macro_tile_size, + unsigned int source_scan, + unsigned int is_chroma) +{ + bool surf_linear = (tiling == dm_sw_linear); + bool surf_vert = (source_scan == dm_vert); + + unsigned int bytes_per_element; + unsigned int bytes_per_element_y = get_bytes_per_element((enum source_format_class)(source_format), + false); + unsigned int bytes_per_element_c = get_bytes_per_element((enum source_format_class)(source_format), + true); + + unsigned int blk256_width = 0; + unsigned int blk256_height = 0; + + unsigned int blk256_width_y = 0; + unsigned int blk256_height_y = 0; + unsigned int blk256_width_c = 0; + unsigned int blk256_height_c = 0; + unsigned int log2_bytes_per_element; + unsigned int log2_blk256_width; + unsigned int log2_blk256_height; + unsigned int blk_bytes; + unsigned int log2_blk_bytes; + unsigned int log2_blk_height; + unsigned int log2_blk_width; + unsigned int log2_meta_req_bytes; + unsigned int log2_meta_req_height; + unsigned int log2_meta_req_width; + unsigned int meta_req_width; + unsigned int meta_req_height; + unsigned int log2_meta_row_height; + unsigned int meta_row_width_ub; + unsigned int log2_meta_chunk_bytes; + unsigned int log2_meta_chunk_height; + + //full sized meta chunk width in unit of data elements + unsigned int log2_meta_chunk_width; + unsigned int log2_min_meta_chunk_bytes; + unsigned int min_meta_chunk_width; + unsigned int meta_chunk_width; + unsigned int meta_chunk_per_row_int; + unsigned int meta_row_remainder; + unsigned int meta_chunk_threshold; + unsigned int meta_blk_bytes; + unsigned int meta_blk_height; + unsigned int meta_blk_width; + unsigned int meta_surface_bytes; + unsigned int vmpg_bytes; + unsigned int meta_pte_req_per_frame_ub; + unsigned int meta_pte_bytes_per_frame_ub; + const unsigned int log2_vmpg_bytes = dml_log2(mode_lib->soc.vmm_page_size_bytes); + const unsigned int dpte_buf_in_pte_reqs = mode_lib->ip.dpte_buffer_size_in_pte_reqs; + const unsigned int pde_proc_buffer_size_64k_reqs = + mode_lib->ip.pde_proc_buffer_size_64k_reqs; + + unsigned int log2_vmpg_height = 0; + unsigned int log2_vmpg_width = 0; + unsigned int log2_dpte_req_height_ptes = 0; + unsigned int log2_dpte_req_height = 0; + unsigned int log2_dpte_req_width = 0; + unsigned int log2_dpte_row_height_linear = 0; + unsigned int log2_dpte_row_height = 0; + unsigned int log2_dpte_group_width = 0; + unsigned int dpte_row_width_ub = 0; + unsigned int dpte_req_height = 0; + unsigned int dpte_req_width = 0; + unsigned int dpte_group_width = 0; + unsigned int log2_dpte_group_bytes = 0; + unsigned int log2_dpte_group_length = 0; + unsigned int pde_buf_entries; + bool yuv420 = (source_format == dm_420_8 || source_format == dm_420_10); + + Calculate256BBlockSizes((enum source_format_class)(source_format), + (enum dm_swizzle_mode)(tiling), + bytes_per_element_y, + bytes_per_element_c, + &blk256_height_y, + &blk256_height_c, + &blk256_width_y, + &blk256_width_c); + + if (!is_chroma) { + blk256_width = blk256_width_y; + blk256_height = blk256_height_y; + bytes_per_element = bytes_per_element_y; + } else { + blk256_width = blk256_width_c; + blk256_height = blk256_height_c; + bytes_per_element = bytes_per_element_c; + } + + log2_bytes_per_element = dml_log2(bytes_per_element); + + dml_print("DML_DLG: %s: surf_linear = %d\n", __func__, surf_linear); + dml_print("DML_DLG: %s: surf_vert = %d\n", __func__, surf_vert); + dml_print("DML_DLG: %s: blk256_width = %d\n", __func__, blk256_width); + dml_print("DML_DLG: %s: blk256_height = %d\n", __func__, blk256_height); + + log2_blk256_width = dml_log2((double) blk256_width); + log2_blk256_height = dml_log2((double) blk256_height); + blk_bytes = surf_linear ? + 256 : get_blk_size_bytes((enum source_macro_tile_size) macro_tile_size); + log2_blk_bytes = dml_log2((double) blk_bytes); + log2_blk_height = 0; + log2_blk_width = 0; + + // remember log rule + // "+" in log is multiply + // "-" in log is divide + // "/2" is like square root + // blk is vertical biased + if (tiling != dm_sw_linear) + log2_blk_height = log2_blk256_height + + dml_ceil((double) (log2_blk_bytes - 8) / 2.0, 1); + else + log2_blk_height = 0; // blk height of 1 + + log2_blk_width = log2_blk_bytes - log2_bytes_per_element - log2_blk_height; + + if (!surf_vert) { + rq_dlg_param->swath_width_ub = dml_round_to_multiple(vp_width - 1, blk256_width, 1) + + blk256_width; + rq_dlg_param->req_per_swath_ub = rq_dlg_param->swath_width_ub >> log2_blk256_width; + } else { + rq_dlg_param->swath_width_ub = dml_round_to_multiple(vp_height - 1, blk256_height, 1) + + blk256_height; + rq_dlg_param->req_per_swath_ub = rq_dlg_param->swath_width_ub >> log2_blk256_height; + } + + if (!surf_vert) + rq_misc_param->full_swath_bytes = rq_dlg_param->swath_width_ub * blk256_height + * bytes_per_element; + else + rq_misc_param->full_swath_bytes = rq_dlg_param->swath_width_ub * blk256_width + * bytes_per_element; + + rq_misc_param->blk256_height = blk256_height; + rq_misc_param->blk256_width = blk256_width; + + // ------- + // meta + // ------- + log2_meta_req_bytes = 6; // meta request is 64b and is 8x8byte meta element + + // each 64b meta request for dcn is 8x8 meta elements and + // a meta element covers one 256b block of the the data surface. + log2_meta_req_height = log2_blk256_height + 3; // meta req is 8x8 byte, each byte represent 1 blk256 + log2_meta_req_width = log2_meta_req_bytes + 8 - log2_bytes_per_element + - log2_meta_req_height; + meta_req_width = 1 << log2_meta_req_width; + meta_req_height = 1 << log2_meta_req_height; + log2_meta_row_height = 0; + meta_row_width_ub = 0; + + // the dimensions of a meta row are meta_row_width x meta_row_height in elements. + // calculate upper bound of the meta_row_width + if (!surf_vert) { + log2_meta_row_height = log2_meta_req_height; + meta_row_width_ub = dml_round_to_multiple(vp_width - 1, meta_req_width, 1) + + meta_req_width; + rq_dlg_param->meta_req_per_row_ub = meta_row_width_ub / meta_req_width; + } else { + log2_meta_row_height = log2_meta_req_width; + meta_row_width_ub = dml_round_to_multiple(vp_height - 1, meta_req_height, 1) + + meta_req_height; + rq_dlg_param->meta_req_per_row_ub = meta_row_width_ub / meta_req_height; + } + rq_dlg_param->meta_bytes_per_row_ub = rq_dlg_param->meta_req_per_row_ub * 64; + + rq_dlg_param->meta_row_height = 1 << log2_meta_row_height; + + log2_meta_chunk_bytes = dml_log2(rq_sizing_param->meta_chunk_bytes); + log2_meta_chunk_height = log2_meta_row_height; + + //full sized meta chunk width in unit of data elements + log2_meta_chunk_width = log2_meta_chunk_bytes + 8 - log2_bytes_per_element + - log2_meta_chunk_height; + log2_min_meta_chunk_bytes = dml_log2(rq_sizing_param->min_meta_chunk_bytes); + min_meta_chunk_width = 1 + << (log2_min_meta_chunk_bytes + 8 - log2_bytes_per_element + - log2_meta_chunk_height); + meta_chunk_width = 1 << log2_meta_chunk_width; + meta_chunk_per_row_int = (unsigned int) (meta_row_width_ub / meta_chunk_width); + meta_row_remainder = meta_row_width_ub % meta_chunk_width; + meta_chunk_threshold = 0; + meta_blk_bytes = 4096; + meta_blk_height = blk256_height * 64; + meta_blk_width = meta_blk_bytes * 256 / bytes_per_element / meta_blk_height; + meta_surface_bytes = meta_pitch + * (dml_round_to_multiple(vp_height - 1, meta_blk_height, 1) + meta_blk_height) + * bytes_per_element / 256; + vmpg_bytes = mode_lib->soc.vmm_page_size_bytes; + meta_pte_req_per_frame_ub = (dml_round_to_multiple(meta_surface_bytes - vmpg_bytes, + 8 * vmpg_bytes, + 1) + 8 * vmpg_bytes) / (8 * vmpg_bytes); + meta_pte_bytes_per_frame_ub = meta_pte_req_per_frame_ub * 64; //64B mpte request + rq_dlg_param->meta_pte_bytes_per_frame_ub = meta_pte_bytes_per_frame_ub; + + dml_print("DML_DLG: %s: meta_blk_height = %d\n", __func__, meta_blk_height); + dml_print("DML_DLG: %s: meta_blk_width = %d\n", __func__, meta_blk_width); + dml_print("DML_DLG: %s: meta_surface_bytes = %d\n", __func__, meta_surface_bytes); + dml_print("DML_DLG: %s: meta_pte_req_per_frame_ub = %d\n", + __func__, + meta_pte_req_per_frame_ub); + dml_print("DML_DLG: %s: meta_pte_bytes_per_frame_ub = %d\n", + __func__, + meta_pte_bytes_per_frame_ub); + + if (!surf_vert) + meta_chunk_threshold = 2 * min_meta_chunk_width - meta_req_width; + else + meta_chunk_threshold = 2 * min_meta_chunk_width - meta_req_height; + + if (meta_row_remainder <= meta_chunk_threshold) + rq_dlg_param->meta_chunks_per_row_ub = meta_chunk_per_row_int + 1; + else + rq_dlg_param->meta_chunks_per_row_ub = meta_chunk_per_row_int + 2; + + // ------ + // dpte + // ------ + if (surf_linear) { + log2_vmpg_height = 0; // one line high + } else { + log2_vmpg_height = (log2_vmpg_bytes - 8) / 2 + log2_blk256_height; + } + log2_vmpg_width = log2_vmpg_bytes - log2_bytes_per_element - log2_vmpg_height; + + // only 3 possible shapes for dpte request in dimensions of ptes: 8x1, 4x2, 2x4. + if (surf_linear) { //one 64B PTE request returns 8 PTEs + log2_dpte_req_height_ptes = 0; + log2_dpte_req_width = log2_vmpg_width + 3; + log2_dpte_req_height = 0; + } else if (log2_blk_bytes == 12) { //4KB tile means 4kB page size + //one 64B req gives 8x1 PTEs for 4KB tile + log2_dpte_req_height_ptes = 0; + log2_dpte_req_width = log2_blk_width + 3; + log2_dpte_req_height = log2_blk_height + 0; + } else if ((log2_blk_bytes >= 16) && (log2_vmpg_bytes == 12)) { // tile block >= 64KB + //two 64B reqs of 2x4 PTEs give 16 PTEs to cover 64KB + log2_dpte_req_height_ptes = 4; + log2_dpte_req_width = log2_blk256_width + 4; // log2_64KB_width + log2_dpte_req_height = log2_blk256_height + 4; // log2_64KB_height + } else { //64KB page size and must 64KB tile block + //one 64B req gives 8x1 PTEs for 64KB tile + log2_dpte_req_height_ptes = 0; + log2_dpte_req_width = log2_blk_width + 3; + log2_dpte_req_height = log2_blk_height + 0; + } + + // The dpte request dimensions in data elements is dpte_req_width x dpte_req_height + // log2_vmpg_width is how much 1 pte represent, now calculating how much a 64b pte req represent + // That depends on the pte shape (i.e. 8x1, 4x2, 2x4) + //log2_dpte_req_height = log2_vmpg_height + log2_dpte_req_height_ptes; + //log2_dpte_req_width = log2_vmpg_width + log2_dpte_req_width_ptes; + dpte_req_height = 1 << log2_dpte_req_height; + dpte_req_width = 1 << log2_dpte_req_width; + + // calculate pitch dpte row buffer can hold + // round the result down to a power of two. + pde_buf_entries = yuv420 ? (pde_proc_buffer_size_64k_reqs >> 1) : pde_proc_buffer_size_64k_reqs; + if (surf_linear) { + unsigned int dpte_row_height; + + log2_dpte_row_height_linear = dml_floor(dml_log2(dml_min(64 * 1024 * pde_buf_entries + / bytes_per_element, + dpte_buf_in_pte_reqs + * dpte_req_width) + / data_pitch), + 1); + + ASSERT(log2_dpte_row_height_linear >= 3); + + if (log2_dpte_row_height_linear > 7) + log2_dpte_row_height_linear = 7; + + log2_dpte_row_height = log2_dpte_row_height_linear; + // For linear, the dpte row is pitch dependent and the pte requests wrap at the pitch boundary. + // the dpte_row_width_ub is the upper bound of data_pitch*dpte_row_height in elements with this unique buffering. + dpte_row_height = 1 << log2_dpte_row_height; + dpte_row_width_ub = dml_round_to_multiple(data_pitch * dpte_row_height - 1, + dpte_req_width, + 1) + dpte_req_width; + rq_dlg_param->dpte_req_per_row_ub = dpte_row_width_ub / dpte_req_width; + } else { + // the upper bound of the dpte_row_width without dependency on viewport position follows. + // for tiled mode, row height is the same as req height and row store up to vp size upper bound + if (!surf_vert) { + log2_dpte_row_height = log2_dpte_req_height; + dpte_row_width_ub = dml_round_to_multiple(vp_width - 1, dpte_req_width, 1) + + dpte_req_width; + rq_dlg_param->dpte_req_per_row_ub = dpte_row_width_ub / dpte_req_width; + } else { + log2_dpte_row_height = + (log2_blk_width < log2_dpte_req_width) ? + log2_blk_width : log2_dpte_req_width; + dpte_row_width_ub = dml_round_to_multiple(vp_height - 1, dpte_req_height, 1) + + dpte_req_height; + rq_dlg_param->dpte_req_per_row_ub = dpte_row_width_ub / dpte_req_height; + } + } + if (log2_blk_bytes >= 16 && log2_vmpg_bytes == 12) // tile block >= 64KB + rq_dlg_param->dpte_bytes_per_row_ub = rq_dlg_param->dpte_req_per_row_ub * 128; //2*64B dpte request + else + rq_dlg_param->dpte_bytes_per_row_ub = rq_dlg_param->dpte_req_per_row_ub * 64; //64B dpte request + + rq_dlg_param->dpte_row_height = 1 << log2_dpte_row_height; + + // the dpte_group_bytes is reduced for the specific case of vertical + // access of a tile surface that has dpte request of 8x1 ptes. + if (!surf_linear & (log2_dpte_req_height_ptes == 0) & surf_vert) //reduced, in this case, will have page fault within a group + rq_sizing_param->dpte_group_bytes = 512; + else + //full size + rq_sizing_param->dpte_group_bytes = 2048; + + //since pte request size is 64byte, the number of data pte requests per full sized group is as follows. + log2_dpte_group_bytes = dml_log2(rq_sizing_param->dpte_group_bytes); + log2_dpte_group_length = log2_dpte_group_bytes - 6; //length in 64b requests + + // full sized data pte group width in elements + if (!surf_vert) + log2_dpte_group_width = log2_dpte_group_length + log2_dpte_req_width; + else + log2_dpte_group_width = log2_dpte_group_length + log2_dpte_req_height; + + //But if the tile block >=64KB and the page size is 4KB, then each dPTE request is 2*64B + if ((log2_blk_bytes >= 16) && (log2_vmpg_bytes == 12)) // tile block >= 64KB + log2_dpte_group_width = log2_dpte_group_width - 1; + + dpte_group_width = 1 << log2_dpte_group_width; + + // since dpte groups are only aligned to dpte_req_width and not dpte_group_width, + // the upper bound for the dpte groups per row is as follows. + rq_dlg_param->dpte_groups_per_row_ub = dml_ceil((double) dpte_row_width_ub / dpte_group_width, + 1); +} + +static void get_surf_rq_param(struct display_mode_lib *mode_lib, + display_data_rq_sizing_params_st *rq_sizing_param, + display_data_rq_dlg_params_st *rq_dlg_param, + display_data_rq_misc_params_st *rq_misc_param, + const display_pipe_source_params_st pipe_src_param, + bool is_chroma) +{ + bool mode_422 = 0; + unsigned int vp_width = 0; + unsigned int vp_height = 0; + unsigned int data_pitch = 0; + unsigned int meta_pitch = 0; + unsigned int ppe = mode_422 ? 2 : 1; + + // FIXME check if ppe apply for both luma and chroma in 422 case + if (is_chroma) { + vp_width = pipe_src_param.viewport_width_c / ppe; + vp_height = pipe_src_param.viewport_height_c; + data_pitch = pipe_src_param.data_pitch_c; + meta_pitch = pipe_src_param.meta_pitch_c; + } else { + vp_width = pipe_src_param.viewport_width / ppe; + vp_height = pipe_src_param.viewport_height; + data_pitch = pipe_src_param.data_pitch; + meta_pitch = pipe_src_param.meta_pitch; + } + + rq_sizing_param->chunk_bytes = 8192; + + if (rq_sizing_param->chunk_bytes == 64 * 1024) + rq_sizing_param->min_chunk_bytes = 0; + else + rq_sizing_param->min_chunk_bytes = 1024; + + rq_sizing_param->meta_chunk_bytes = 2048; + rq_sizing_param->min_meta_chunk_bytes = 256; + + rq_sizing_param->mpte_group_bytes = 2048; + + get_meta_and_pte_attr(mode_lib, + rq_dlg_param, + rq_misc_param, + rq_sizing_param, + vp_width, + vp_height, + data_pitch, + meta_pitch, + pipe_src_param.source_format, + pipe_src_param.sw_mode, + pipe_src_param.macro_tile_size, + pipe_src_param.source_scan, + is_chroma); +} + +void dml_rq_dlg_get_rq_params(struct display_mode_lib *mode_lib, + display_rq_params_st *rq_param, + const display_pipe_source_params_st pipe_src_param) +{ + // get param for luma surface + rq_param->yuv420 = pipe_src_param.source_format == dm_420_8 + || pipe_src_param.source_format == dm_420_10; + rq_param->yuv420_10bpc = pipe_src_param.source_format == dm_420_10; + + get_surf_rq_param(mode_lib, + &(rq_param->sizing.rq_l), + &(rq_param->dlg.rq_l), + &(rq_param->misc.rq_l), + pipe_src_param, + 0); + + if (is_dual_plane((enum source_format_class)(pipe_src_param.source_format))) { + // get param for chroma surface + get_surf_rq_param(mode_lib, + &(rq_param->sizing.rq_c), + &(rq_param->dlg.rq_c), + &(rq_param->misc.rq_c), + pipe_src_param, + 1); + } + + // calculate how to split the det buffer space between luma and chroma + handle_det_buf_split(mode_lib, rq_param, pipe_src_param); + print__rq_params_st(mode_lib, *rq_param); +} + +void dml_rq_dlg_get_rq_reg(struct display_mode_lib *mode_lib, + display_rq_regs_st *rq_regs, + const display_pipe_source_params_st pipe_src_param) +{ + display_rq_params_st rq_param = {0}; + + memset(rq_regs, 0, sizeof(*rq_regs)); + dml_rq_dlg_get_rq_params(mode_lib, &rq_param, pipe_src_param); + extract_rq_regs(mode_lib, rq_regs, rq_param); + + print__rq_regs_st(mode_lib, *rq_regs); +} + +// Note: currently taken in as is. +// Nice to decouple code from hw register implement and extract code that are repeated for luma and chroma. +void dml_rq_dlg_get_dlg_params(struct display_mode_lib *mode_lib, + const display_e2e_pipe_params_st *e2e_pipe_param, + const unsigned int num_pipes, + const unsigned int pipe_idx, + display_dlg_regs_st *disp_dlg_regs, + display_ttu_regs_st *disp_ttu_regs, + const display_rq_dlg_params_st rq_dlg_param, + const display_dlg_sys_params_st dlg_sys_param, + const bool cstate_en, + const bool pstate_en, + const bool vm_en, + const bool ignore_viewport_pos, + const bool immediate_flip_support) +{ + const display_pipe_source_params_st *src = &e2e_pipe_param[pipe_idx].pipe.src; + const display_pipe_dest_params_st *dst = &e2e_pipe_param[pipe_idx].pipe.dest; + const display_output_params_st *dout = &e2e_pipe_param[pipe_idx].dout; + const display_clocks_and_cfg_st *clks = &e2e_pipe_param[pipe_idx].clks_cfg; + const scaler_ratio_depth_st *scl = &e2e_pipe_param[pipe_idx].pipe.scale_ratio_depth; + const scaler_taps_st *taps = &e2e_pipe_param[pipe_idx].pipe.scale_taps; + + // ------------------------- + // Section 1.15.2.1: OTG dependent Params + // ------------------------- + // Timing + unsigned int htotal = dst->htotal; +// unsigned int hblank_start = dst.hblank_start; // TODO: Remove + unsigned int hblank_end = dst->hblank_end; + unsigned int vblank_start = dst->vblank_start; + unsigned int vblank_end = dst->vblank_end; + unsigned int min_vblank = mode_lib->ip.min_vblank_lines; + + double dppclk_freq_in_mhz = clks->dppclk_mhz; + double dispclk_freq_in_mhz = clks->dispclk_mhz; + double refclk_freq_in_mhz = clks->refclk_mhz; + double pclk_freq_in_mhz = dst->pixel_rate_mhz; + bool interlaced = dst->interlaced; + + double ref_freq_to_pix_freq = refclk_freq_in_mhz / pclk_freq_in_mhz; + + double min_dcfclk_mhz; + double t_calc_us; + double min_ttu_vblank; + + double min_dst_y_ttu_vblank; + unsigned int dlg_vblank_start; + bool dual_plane; + bool mode_422; + unsigned int access_dir; + unsigned int vp_height_l; + unsigned int vp_width_l; + unsigned int vp_height_c; + unsigned int vp_width_c; + + // Scaling + unsigned int htaps_l; + unsigned int htaps_c; + double hratio_l; + double hratio_c; + double vratio_l; + double vratio_c; + bool scl_enable; + + double line_time_in_us; + // double vinit_l; + // double vinit_c; + // double vinit_bot_l; + // double vinit_bot_c; + + // unsigned int swath_height_l; + unsigned int swath_width_ub_l; + // unsigned int dpte_bytes_per_row_ub_l; + unsigned int dpte_groups_per_row_ub_l; + // unsigned int meta_pte_bytes_per_frame_ub_l; + // unsigned int meta_bytes_per_row_ub_l; + + // unsigned int swath_height_c; + unsigned int swath_width_ub_c; + // unsigned int dpte_bytes_per_row_ub_c; + unsigned int dpte_groups_per_row_ub_c; + + unsigned int meta_chunks_per_row_ub_l; + unsigned int meta_chunks_per_row_ub_c; + unsigned int vupdate_offset; + unsigned int vupdate_width; + unsigned int vready_offset; + + unsigned int dppclk_delay_subtotal; + unsigned int dispclk_delay_subtotal; + unsigned int pixel_rate_delay_subtotal; + + unsigned int vstartup_start; + unsigned int dst_x_after_scaler; + unsigned int dst_y_after_scaler; + double line_wait; + double dst_y_prefetch; + double dst_y_per_vm_vblank; + double dst_y_per_row_vblank; + double dst_y_per_vm_flip; + double dst_y_per_row_flip; + double min_dst_y_per_vm_vblank; + double min_dst_y_per_row_vblank; + double lsw; + double vratio_pre_l; + double vratio_pre_c; + unsigned int req_per_swath_ub_l; + unsigned int req_per_swath_ub_c; + unsigned int meta_row_height_l; + unsigned int meta_row_height_c; + unsigned int swath_width_pixels_ub_l; + unsigned int swath_width_pixels_ub_c; + unsigned int scaler_rec_in_width_l; + unsigned int scaler_rec_in_width_c; + unsigned int dpte_row_height_l; + unsigned int dpte_row_height_c; + double hscale_pixel_rate_l; + double hscale_pixel_rate_c; + double min_hratio_fact_l; + double min_hratio_fact_c; + double refcyc_per_line_delivery_pre_l; + double refcyc_per_line_delivery_pre_c; + double refcyc_per_line_delivery_l; + double refcyc_per_line_delivery_c; + + double refcyc_per_req_delivery_pre_l; + double refcyc_per_req_delivery_pre_c; + double refcyc_per_req_delivery_l; + double refcyc_per_req_delivery_c; + + unsigned int full_recout_width; + double xfc_transfer_delay; + double xfc_precharge_delay; + double xfc_remote_surface_flip_latency; + double xfc_dst_y_delta_drq_limit; + double xfc_prefetch_margin; + double refcyc_per_req_delivery_pre_cur0; + double refcyc_per_req_delivery_cur0; + double refcyc_per_req_delivery_pre_cur1; + double refcyc_per_req_delivery_cur1; + + memset(disp_dlg_regs, 0, sizeof(*disp_dlg_regs)); + memset(disp_ttu_regs, 0, sizeof(*disp_ttu_regs)); + + dml_print("DML_DLG: %s: cstate_en = %d\n", __func__, cstate_en); + dml_print("DML_DLG: %s: pstate_en = %d\n", __func__, pstate_en); + dml_print("DML_DLG: %s: vm_en = %d\n", __func__, vm_en); + dml_print("DML_DLG: %s: ignore_viewport_pos = %d\n", __func__, ignore_viewport_pos); + dml_print("DML_DLG: %s: immediate_flip_support = %d\n", __func__, immediate_flip_support); + + dml_print("DML_DLG: %s: dppclk_freq_in_mhz = %3.2f\n", __func__, dppclk_freq_in_mhz); + dml_print("DML_DLG: %s: dispclk_freq_in_mhz = %3.2f\n", __func__, dispclk_freq_in_mhz); + dml_print("DML_DLG: %s: refclk_freq_in_mhz = %3.2f\n", __func__, refclk_freq_in_mhz); + dml_print("DML_DLG: %s: pclk_freq_in_mhz = %3.2f\n", __func__, pclk_freq_in_mhz); + dml_print("DML_DLG: %s: interlaced = %d\n", __func__, interlaced); + ASSERT(ref_freq_to_pix_freq < 4.0); + + disp_dlg_regs->ref_freq_to_pix_freq = + (unsigned int) (ref_freq_to_pix_freq * dml_pow(2, 19)); + disp_dlg_regs->refcyc_per_htotal = (unsigned int) (ref_freq_to_pix_freq * (double) htotal + * dml_pow(2, 8)); + disp_dlg_regs->dlg_vblank_end = interlaced ? (vblank_end / 2) : vblank_end; // 15 bits + disp_dlg_regs->refcyc_h_blank_end = (unsigned int) ((double) hblank_end + * (double) ref_freq_to_pix_freq); + ASSERT(disp_dlg_regs->refcyc_h_blank_end < (unsigned int) dml_pow(2, 13)); + + min_dcfclk_mhz = dlg_sys_param.deepsleep_dcfclk_mhz; + set_prefetch_mode(mode_lib, cstate_en, pstate_en, ignore_viewport_pos, immediate_flip_support); + t_calc_us = get_tcalc(mode_lib, e2e_pipe_param, num_pipes); + min_ttu_vblank = get_min_ttu_vblank(mode_lib, e2e_pipe_param, num_pipes, pipe_idx); + + min_dst_y_ttu_vblank = min_ttu_vblank * pclk_freq_in_mhz / (double) htotal; + dlg_vblank_start = interlaced ? (vblank_start / 2) : vblank_start; + + disp_dlg_regs->min_dst_y_next_start = (unsigned int) (((double) dlg_vblank_start + + min_dst_y_ttu_vblank) * dml_pow(2, 2)); + ASSERT(disp_dlg_regs->min_dst_y_next_start < (unsigned int) dml_pow(2, 18)); + + dml_print("DML_DLG: %s: min_dcfclk_mhz = %3.2f\n", + __func__, + min_dcfclk_mhz); + dml_print("DML_DLG: %s: min_ttu_vblank = %3.2f\n", + __func__, + min_ttu_vblank); + dml_print("DML_DLG: %s: min_dst_y_ttu_vblank = %3.2f\n", + __func__, + min_dst_y_ttu_vblank); + dml_print("DML_DLG: %s: t_calc_us = %3.2f\n", + __func__, + t_calc_us); + dml_print("DML_DLG: %s: disp_dlg_regs->min_dst_y_next_start = 0x%0x\n", + __func__, + disp_dlg_regs->min_dst_y_next_start); + dml_print("DML_DLG: %s: ref_freq_to_pix_freq = %3.2f\n", + __func__, + ref_freq_to_pix_freq); + + // ------------------------- + // Section 1.15.2.2: Prefetch, Active and TTU + // ------------------------- + // Prefetch Calc + // Source +// dcc_en = src.dcc; + dual_plane = is_dual_plane((enum source_format_class)(src->source_format)); + mode_422 = 0; // FIXME + access_dir = (src->source_scan == dm_vert); // vp access direction: horizontal or vertical accessed +// bytes_per_element_l = get_bytes_per_element(source_format_class(src.source_format), 0); +// bytes_per_element_c = get_bytes_per_element(source_format_class(src.source_format), 1); + vp_height_l = src->viewport_height; + vp_width_l = src->viewport_width; + vp_height_c = src->viewport_height_c; + vp_width_c = src->viewport_width_c; + + // Scaling + htaps_l = taps->htaps; + htaps_c = taps->htaps_c; + hratio_l = scl->hscl_ratio; + hratio_c = scl->hscl_ratio_c; + vratio_l = scl->vscl_ratio; + vratio_c = scl->vscl_ratio_c; + scl_enable = scl->scl_enable; + + line_time_in_us = (htotal / pclk_freq_in_mhz); +// vinit_l = scl.vinit; +// vinit_c = scl.vinit_c; +// vinit_bot_l = scl.vinit_bot; +// vinit_bot_c = scl.vinit_bot_c; + +// unsigned int swath_height_l = rq_dlg_param.rq_l.swath_height; + swath_width_ub_l = rq_dlg_param.rq_l.swath_width_ub; +// unsigned int dpte_bytes_per_row_ub_l = rq_dlg_param.rq_l.dpte_bytes_per_row_ub; + dpte_groups_per_row_ub_l = rq_dlg_param.rq_l.dpte_groups_per_row_ub; +// unsigned int meta_pte_bytes_per_frame_ub_l = rq_dlg_param.rq_l.meta_pte_bytes_per_frame_ub; +// unsigned int meta_bytes_per_row_ub_l = rq_dlg_param.rq_l.meta_bytes_per_row_ub; + +// unsigned int swath_height_c = rq_dlg_param.rq_c.swath_height; + swath_width_ub_c = rq_dlg_param.rq_c.swath_width_ub; + // dpte_bytes_per_row_ub_c = rq_dlg_param.rq_c.dpte_bytes_per_row_ub; + dpte_groups_per_row_ub_c = rq_dlg_param.rq_c.dpte_groups_per_row_ub; + + meta_chunks_per_row_ub_l = rq_dlg_param.rq_l.meta_chunks_per_row_ub; + meta_chunks_per_row_ub_c = rq_dlg_param.rq_c.meta_chunks_per_row_ub; + vupdate_offset = dst->vupdate_offset; + vupdate_width = dst->vupdate_width; + vready_offset = dst->vready_offset; + + dppclk_delay_subtotal = mode_lib->ip.dppclk_delay_subtotal; + dispclk_delay_subtotal = mode_lib->ip.dispclk_delay_subtotal; + + if (scl_enable) + dppclk_delay_subtotal += mode_lib->ip.dppclk_delay_scl; + else + dppclk_delay_subtotal += mode_lib->ip.dppclk_delay_scl_lb_only; + + dppclk_delay_subtotal += mode_lib->ip.dppclk_delay_cnvc_formatter + + src->num_cursors * mode_lib->ip.dppclk_delay_cnvc_cursor; + + if (dout->dsc_enable) { + double dsc_delay = get_dsc_delay(mode_lib, e2e_pipe_param, num_pipes, pipe_idx); + + dispclk_delay_subtotal += dsc_delay; + } + + pixel_rate_delay_subtotal = dppclk_delay_subtotal * pclk_freq_in_mhz / dppclk_freq_in_mhz + + dispclk_delay_subtotal * pclk_freq_in_mhz / dispclk_freq_in_mhz; + + vstartup_start = dst->vstartup_start; + if (interlaced) { + if (vstartup_start / 2.0 + - (double) (vready_offset + vupdate_width + vupdate_offset) / htotal + <= vblank_end / 2.0) + disp_dlg_regs->vready_after_vcount0 = 1; + else + disp_dlg_regs->vready_after_vcount0 = 0; + } else { + if (vstartup_start + - (double) (vready_offset + vupdate_width + vupdate_offset) / htotal + <= vblank_end) + disp_dlg_regs->vready_after_vcount0 = 1; + else + disp_dlg_regs->vready_after_vcount0 = 0; + } + + // TODO: Where is this coming from? + if (interlaced) + vstartup_start = vstartup_start / 2; + + // TODO: What if this min_vblank doesn't match the value in the dml_config_settings.cpp? + if (vstartup_start >= min_vblank) { + dml_print("WARNING: DML_DLG: %s: vblank_start=%d vblank_end=%d\n", + __func__, + vblank_start, + vblank_end); + dml_print("WARNING: DML_DLG: %s: vstartup_start=%d should be less than min_vblank=%d\n", + __func__, + vstartup_start, + min_vblank); + min_vblank = vstartup_start + 1; + dml_print("WARNING: DML_DLG: %s: vstartup_start=%d should be less than min_vblank=%d\n", + __func__, + vstartup_start, + min_vblank); + } + + dst_x_after_scaler = get_dst_x_after_scaler(mode_lib, e2e_pipe_param, num_pipes, pipe_idx); + dst_y_after_scaler = get_dst_y_after_scaler(mode_lib, e2e_pipe_param, num_pipes, pipe_idx); + + dml_print("DML_DLG: %s: htotal = %d\n", __func__, htotal); + dml_print("DML_DLG: %s: pixel_rate_delay_subtotal = %d\n", + __func__, + pixel_rate_delay_subtotal); + dml_print("DML_DLG: %s: dst_x_after_scaler = %d\n", + __func__, + dst_x_after_scaler); + dml_print("DML_DLG: %s: dst_y_after_scaler = %d\n", + __func__, + dst_y_after_scaler); + + // Lwait + line_wait = mode_lib->soc.urgent_latency_us; + if (cstate_en) + line_wait = dml_max(mode_lib->soc.sr_enter_plus_exit_time_us, line_wait); + if (pstate_en) + line_wait = dml_max(mode_lib->soc.dram_clock_change_latency_us + + mode_lib->soc.urgent_latency_us, + line_wait); + line_wait = line_wait / line_time_in_us; + + dst_y_prefetch = get_dst_y_prefetch(mode_lib, e2e_pipe_param, num_pipes, pipe_idx); + dml_print("DML_DLG: %s: dst_y_prefetch (after rnd) = %3.2f\n", __func__, dst_y_prefetch); + + dst_y_per_vm_vblank = get_dst_y_per_vm_vblank(mode_lib, + e2e_pipe_param, + num_pipes, + pipe_idx); + dst_y_per_row_vblank = get_dst_y_per_row_vblank(mode_lib, + e2e_pipe_param, + num_pipes, + pipe_idx); + dst_y_per_vm_flip = get_dst_y_per_vm_flip(mode_lib, e2e_pipe_param, num_pipes, pipe_idx); + dst_y_per_row_flip = get_dst_y_per_row_flip(mode_lib, e2e_pipe_param, num_pipes, pipe_idx); + + min_dst_y_per_vm_vblank = 8.0; + min_dst_y_per_row_vblank = 16.0; + + // magic! + if (htotal <= 75) { + min_vblank = 300; + min_dst_y_per_vm_vblank = 100.0; + min_dst_y_per_row_vblank = 100.0; + } + + dml_print("DML_DLG: %s: dst_y_per_vm_vblank = %3.2f\n", __func__, dst_y_per_vm_vblank); + dml_print("DML_DLG: %s: dst_y_per_row_vblank = %3.2f\n", __func__, dst_y_per_row_vblank); + + ASSERT(dst_y_per_vm_vblank < min_dst_y_per_vm_vblank); + ASSERT(dst_y_per_row_vblank < min_dst_y_per_row_vblank); + + ASSERT(dst_y_prefetch > (dst_y_per_vm_vblank + dst_y_per_row_vblank)); + lsw = dst_y_prefetch - (dst_y_per_vm_vblank + dst_y_per_row_vblank); + + dml_print("DML_DLG: %s: lsw = %3.2f\n", __func__, lsw); + + vratio_pre_l = get_vratio_prefetch_l(mode_lib, e2e_pipe_param, num_pipes, pipe_idx); + vratio_pre_c = get_vratio_prefetch_c(mode_lib, e2e_pipe_param, num_pipes, pipe_idx); + + dml_print("DML_DLG: %s: vratio_pre_l=%3.2f\n", __func__, vratio_pre_l); + dml_print("DML_DLG: %s: vratio_pre_c=%3.2f\n", __func__, vratio_pre_c); + + // Active + req_per_swath_ub_l = rq_dlg_param.rq_l.req_per_swath_ub; + req_per_swath_ub_c = rq_dlg_param.rq_c.req_per_swath_ub; + meta_row_height_l = rq_dlg_param.rq_l.meta_row_height; + meta_row_height_c = rq_dlg_param.rq_c.meta_row_height; + swath_width_pixels_ub_l = 0; + swath_width_pixels_ub_c = 0; + scaler_rec_in_width_l = 0; + scaler_rec_in_width_c = 0; + dpte_row_height_l = rq_dlg_param.rq_l.dpte_row_height; + dpte_row_height_c = rq_dlg_param.rq_c.dpte_row_height; + + if (mode_422) { + swath_width_pixels_ub_l = swath_width_ub_l * 2; // *2 for 2 pixel per element + swath_width_pixels_ub_c = swath_width_ub_c * 2; + } else { + swath_width_pixels_ub_l = swath_width_ub_l * 1; + swath_width_pixels_ub_c = swath_width_ub_c * 1; + } + + hscale_pixel_rate_l = 0.; + hscale_pixel_rate_c = 0.; + min_hratio_fact_l = 1.0; + min_hratio_fact_c = 1.0; + + if (htaps_l <= 1) + min_hratio_fact_l = 2.0; + else if (htaps_l <= 6) { + if ((hratio_l * 2.0) > 4.0) + min_hratio_fact_l = 4.0; + else + min_hratio_fact_l = hratio_l * 2.0; + } else { + if (hratio_l > 4.0) + min_hratio_fact_l = 4.0; + else + min_hratio_fact_l = hratio_l; + } + + hscale_pixel_rate_l = min_hratio_fact_l * dppclk_freq_in_mhz; + + if (htaps_c <= 1) + min_hratio_fact_c = 2.0; + else if (htaps_c <= 6) { + if ((hratio_c * 2.0) > 4.0) + min_hratio_fact_c = 4.0; + else + min_hratio_fact_c = hratio_c * 2.0; + } else { + if (hratio_c > 4.0) + min_hratio_fact_c = 4.0; + else + min_hratio_fact_c = hratio_c; + } + + hscale_pixel_rate_c = min_hratio_fact_c * dppclk_freq_in_mhz; + + refcyc_per_line_delivery_pre_l = 0.; + refcyc_per_line_delivery_pre_c = 0.; + refcyc_per_line_delivery_l = 0.; + refcyc_per_line_delivery_c = 0.; + + refcyc_per_req_delivery_pre_l = 0.; + refcyc_per_req_delivery_pre_c = 0.; + refcyc_per_req_delivery_l = 0.; + refcyc_per_req_delivery_c = 0.; + + full_recout_width = 0; + // In ODM + if (src->is_hsplit) { + // This "hack" is only allowed (and valid) for MPC combine. In ODM + // combine, you MUST specify the full_recout_width...according to Oswin + if (dst->full_recout_width == 0 && !dst->odm_combine) { + dml_print("DML_DLG: %s: Warning: full_recout_width not set in hsplit mode\n", + __func__); + full_recout_width = dst->recout_width * 2; // assume half split for dcn1 + } else + full_recout_width = dst->full_recout_width; + } else + full_recout_width = dst->recout_width; + + // mpc_combine and odm_combine are mutually exclusive + refcyc_per_line_delivery_pre_l = get_refcyc_per_delivery(mode_lib, + refclk_freq_in_mhz, + pclk_freq_in_mhz, + dst->odm_combine, + full_recout_width, + dst->hactive, + vratio_pre_l, + hscale_pixel_rate_l, + swath_width_pixels_ub_l, + 1); // per line + + refcyc_per_line_delivery_l = get_refcyc_per_delivery(mode_lib, + refclk_freq_in_mhz, + pclk_freq_in_mhz, + dst->odm_combine, + full_recout_width, + dst->hactive, + vratio_l, + hscale_pixel_rate_l, + swath_width_pixels_ub_l, + 1); // per line + + dml_print("DML_DLG: %s: full_recout_width = %d\n", + __func__, + full_recout_width); + dml_print("DML_DLG: %s: hscale_pixel_rate_l = %3.2f\n", + __func__, + hscale_pixel_rate_l); + dml_print("DML_DLG: %s: refcyc_per_line_delivery_pre_l = %3.2f\n", + __func__, + refcyc_per_line_delivery_pre_l); + dml_print("DML_DLG: %s: refcyc_per_line_delivery_l = %3.2f\n", + __func__, + refcyc_per_line_delivery_l); + + if (dual_plane) { + refcyc_per_line_delivery_pre_c = get_refcyc_per_delivery(mode_lib, + refclk_freq_in_mhz, + pclk_freq_in_mhz, + dst->odm_combine, + full_recout_width, + dst->hactive, + vratio_pre_c, + hscale_pixel_rate_c, + swath_width_pixels_ub_c, + 1); // per line + + refcyc_per_line_delivery_c = get_refcyc_per_delivery(mode_lib, + refclk_freq_in_mhz, + pclk_freq_in_mhz, + dst->odm_combine, + full_recout_width, + dst->hactive, + vratio_c, + hscale_pixel_rate_c, + swath_width_pixels_ub_c, + 1); // per line + + dml_print("DML_DLG: %s: refcyc_per_line_delivery_pre_c = %3.2f\n", + __func__, + refcyc_per_line_delivery_pre_c); + dml_print("DML_DLG: %s: refcyc_per_line_delivery_c = %3.2f\n", + __func__, + refcyc_per_line_delivery_c); + } + + // TTU - Luma / Chroma + if (access_dir) { // vertical access + scaler_rec_in_width_l = vp_height_l; + scaler_rec_in_width_c = vp_height_c; + } else { + scaler_rec_in_width_l = vp_width_l; + scaler_rec_in_width_c = vp_width_c; + } + + refcyc_per_req_delivery_pre_l = get_refcyc_per_delivery(mode_lib, + refclk_freq_in_mhz, + pclk_freq_in_mhz, + dst->odm_combine, + full_recout_width, + dst->hactive, + vratio_pre_l, + hscale_pixel_rate_l, + scaler_rec_in_width_l, + req_per_swath_ub_l); // per req + refcyc_per_req_delivery_l = get_refcyc_per_delivery(mode_lib, + refclk_freq_in_mhz, + pclk_freq_in_mhz, + dst->odm_combine, + full_recout_width, + dst->hactive, + vratio_l, + hscale_pixel_rate_l, + scaler_rec_in_width_l, + req_per_swath_ub_l); // per req + + dml_print("DML_DLG: %s: refcyc_per_req_delivery_pre_l = %3.2f\n", + __func__, + refcyc_per_req_delivery_pre_l); + dml_print("DML_DLG: %s: refcyc_per_req_delivery_l = %3.2f\n", + __func__, + refcyc_per_req_delivery_l); + + ASSERT(refcyc_per_req_delivery_pre_l < dml_pow(2, 13)); + ASSERT(refcyc_per_req_delivery_l < dml_pow(2, 13)); + + if (dual_plane) { + refcyc_per_req_delivery_pre_c = get_refcyc_per_delivery(mode_lib, + refclk_freq_in_mhz, + pclk_freq_in_mhz, + dst->odm_combine, + full_recout_width, + dst->hactive, + vratio_pre_c, + hscale_pixel_rate_c, + scaler_rec_in_width_c, + req_per_swath_ub_c); // per req + refcyc_per_req_delivery_c = get_refcyc_per_delivery(mode_lib, + refclk_freq_in_mhz, + pclk_freq_in_mhz, + dst->odm_combine, + full_recout_width, + dst->hactive, + vratio_c, + hscale_pixel_rate_c, + scaler_rec_in_width_c, + req_per_swath_ub_c); // per req + + dml_print("DML_DLG: %s: refcyc_per_req_delivery_pre_c = %3.2f\n", + __func__, + refcyc_per_req_delivery_pre_c); + dml_print("DML_DLG: %s: refcyc_per_req_delivery_c = %3.2f\n", + __func__, + refcyc_per_req_delivery_c); + + ASSERT(refcyc_per_req_delivery_pre_c < dml_pow(2, 13)); + ASSERT(refcyc_per_req_delivery_c < dml_pow(2, 13)); + } + + // XFC + xfc_transfer_delay = get_xfc_transfer_delay(mode_lib, e2e_pipe_param, num_pipes, pipe_idx); + xfc_precharge_delay = get_xfc_precharge_delay(mode_lib, + e2e_pipe_param, + num_pipes, + pipe_idx); + xfc_remote_surface_flip_latency = get_xfc_remote_surface_flip_latency(mode_lib, + e2e_pipe_param, + num_pipes, + pipe_idx); + xfc_dst_y_delta_drq_limit = xfc_remote_surface_flip_latency; + xfc_prefetch_margin = get_xfc_prefetch_margin(mode_lib, + e2e_pipe_param, + num_pipes, + pipe_idx); + + // TTU - Cursor + refcyc_per_req_delivery_pre_cur0 = 0.0; + refcyc_per_req_delivery_cur0 = 0.0; + if (src->num_cursors > 0) { + calculate_ttu_cursor(mode_lib, + &refcyc_per_req_delivery_pre_cur0, + &refcyc_per_req_delivery_cur0, + refclk_freq_in_mhz, + ref_freq_to_pix_freq, + hscale_pixel_rate_l, + scl->hscl_ratio, + vratio_pre_l, + vratio_l, + src->cur0_src_width, + (enum cursor_bpp)(src->cur0_bpp)); + } + + refcyc_per_req_delivery_pre_cur1 = 0.0; + refcyc_per_req_delivery_cur1 = 0.0; + if (src->num_cursors > 1) { + calculate_ttu_cursor(mode_lib, + &refcyc_per_req_delivery_pre_cur1, + &refcyc_per_req_delivery_cur1, + refclk_freq_in_mhz, + ref_freq_to_pix_freq, + hscale_pixel_rate_l, + scl->hscl_ratio, + vratio_pre_l, + vratio_l, + src->cur1_src_width, + (enum cursor_bpp)(src->cur1_bpp)); + } + + // TTU - Misc + // all hard-coded + + // Assignment to register structures + disp_dlg_regs->dst_y_after_scaler = dst_y_after_scaler; // in terms of line + disp_dlg_regs->refcyc_x_after_scaler = dst_x_after_scaler * ref_freq_to_pix_freq; // in terms of refclk + ASSERT(disp_dlg_regs->refcyc_x_after_scaler < (unsigned int) dml_pow(2, 13)); + disp_dlg_regs->dst_y_prefetch = (unsigned int) (dst_y_prefetch * dml_pow(2, 2)); + disp_dlg_regs->dst_y_per_vm_vblank = (unsigned int) (dst_y_per_vm_vblank * dml_pow(2, 2)); + disp_dlg_regs->dst_y_per_row_vblank = (unsigned int) (dst_y_per_row_vblank * dml_pow(2, 2)); + disp_dlg_regs->dst_y_per_vm_flip = (unsigned int) (dst_y_per_vm_flip * dml_pow(2, 2)); + disp_dlg_regs->dst_y_per_row_flip = (unsigned int) (dst_y_per_row_flip * dml_pow(2, 2)); + + disp_dlg_regs->vratio_prefetch = (unsigned int) (vratio_pre_l * dml_pow(2, 19)); + disp_dlg_regs->vratio_prefetch_c = (unsigned int) (vratio_pre_c * dml_pow(2, 19)); + + 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); + ASSERT(disp_dlg_regs->refcyc_per_pte_group_vblank_l < (unsigned int) dml_pow(2, 13)); + + if (dual_plane) { + disp_dlg_regs->refcyc_per_pte_group_vblank_c = (unsigned int) (dst_y_per_row_vblank + * (double) htotal * ref_freq_to_pix_freq + / (double) dpte_groups_per_row_ub_c); + ASSERT(disp_dlg_regs->refcyc_per_pte_group_vblank_c + < (unsigned int) dml_pow(2, 13)); + } + + disp_dlg_regs->refcyc_per_meta_chunk_vblank_l = + (unsigned int) (dst_y_per_row_vblank * (double) htotal + * ref_freq_to_pix_freq / (double) meta_chunks_per_row_ub_l); + ASSERT(disp_dlg_regs->refcyc_per_meta_chunk_vblank_l < (unsigned int) dml_pow(2, 13)); + + disp_dlg_regs->refcyc_per_meta_chunk_vblank_c = + disp_dlg_regs->refcyc_per_meta_chunk_vblank_l; // dcc for 4:2:0 is not supported in dcn1.0. assigned to be the same as _l for now + + disp_dlg_regs->refcyc_per_pte_group_flip_l = (unsigned int) (dst_y_per_row_flip * htotal + * ref_freq_to_pix_freq) / dpte_groups_per_row_ub_l; + disp_dlg_regs->refcyc_per_meta_chunk_flip_l = (unsigned int) (dst_y_per_row_flip * htotal + * ref_freq_to_pix_freq) / meta_chunks_per_row_ub_l; + + if (dual_plane) { + disp_dlg_regs->refcyc_per_pte_group_flip_c = (unsigned int) (dst_y_per_row_flip + * htotal * ref_freq_to_pix_freq) / dpte_groups_per_row_ub_c; + disp_dlg_regs->refcyc_per_meta_chunk_flip_c = (unsigned int) (dst_y_per_row_flip + * htotal * ref_freq_to_pix_freq) / meta_chunks_per_row_ub_c; + } + + disp_dlg_regs->dst_y_per_pte_row_nom_l = (unsigned int) ((double) dpte_row_height_l + / (double) vratio_l * dml_pow(2, 2)); + ASSERT(disp_dlg_regs->dst_y_per_pte_row_nom_l < (unsigned int) dml_pow(2, 17)); + + if (dual_plane) { + disp_dlg_regs->dst_y_per_pte_row_nom_c = (unsigned int) ((double) dpte_row_height_c + / (double) vratio_c * dml_pow(2, 2)); + if (disp_dlg_regs->dst_y_per_pte_row_nom_c >= (unsigned int) dml_pow(2, 17)) { + dml_print("DML_DLG: %s: Warning dst_y_per_pte_row_nom_c %u larger than supported by register format U15.2 %u\n", + __func__, + disp_dlg_regs->dst_y_per_pte_row_nom_c, + (unsigned int) dml_pow(2, 17) - 1); + } + } + + disp_dlg_regs->dst_y_per_meta_row_nom_l = (unsigned int) ((double) meta_row_height_l + / (double) vratio_l * dml_pow(2, 2)); + ASSERT(disp_dlg_regs->dst_y_per_meta_row_nom_l < (unsigned int) dml_pow(2, 17)); + + disp_dlg_regs->dst_y_per_meta_row_nom_c = disp_dlg_regs->dst_y_per_meta_row_nom_l; // TODO: dcc for 4:2:0 is not supported in dcn1.0. assigned to be the same as _l for now + + disp_dlg_regs->refcyc_per_pte_group_nom_l = (unsigned int) ((double) dpte_row_height_l + / (double) vratio_l * (double) htotal * ref_freq_to_pix_freq + / (double) dpte_groups_per_row_ub_l); + if (disp_dlg_regs->refcyc_per_pte_group_nom_l >= (unsigned int) dml_pow(2, 23)) + disp_dlg_regs->refcyc_per_pte_group_nom_l = dml_pow(2, 23) - 1; + disp_dlg_regs->refcyc_per_meta_chunk_nom_l = (unsigned int) ((double) meta_row_height_l + / (double) vratio_l * (double) htotal * ref_freq_to_pix_freq + / (double) meta_chunks_per_row_ub_l); + if (disp_dlg_regs->refcyc_per_meta_chunk_nom_l >= (unsigned int) dml_pow(2, 23)) + disp_dlg_regs->refcyc_per_meta_chunk_nom_l = dml_pow(2, 23) - 1; + + if (dual_plane) { + disp_dlg_regs->refcyc_per_pte_group_nom_c = + (unsigned int) ((double) dpte_row_height_c / (double) vratio_c + * (double) htotal * ref_freq_to_pix_freq + / (double) dpte_groups_per_row_ub_c); + if (disp_dlg_regs->refcyc_per_pte_group_nom_c >= (unsigned int) dml_pow(2, 23)) + disp_dlg_regs->refcyc_per_pte_group_nom_c = dml_pow(2, 23) - 1; + + // TODO: Is this the right calculation? Does htotal need to be halved? + disp_dlg_regs->refcyc_per_meta_chunk_nom_c = + (unsigned int) ((double) meta_row_height_c / (double) vratio_c + * (double) htotal * ref_freq_to_pix_freq + / (double) meta_chunks_per_row_ub_c); + if (disp_dlg_regs->refcyc_per_meta_chunk_nom_c >= (unsigned int) dml_pow(2, 23)) + disp_dlg_regs->refcyc_per_meta_chunk_nom_c = dml_pow(2, 23) - 1; + } + + disp_dlg_regs->refcyc_per_line_delivery_pre_l = (unsigned int) dml_floor(refcyc_per_line_delivery_pre_l, + 1); + disp_dlg_regs->refcyc_per_line_delivery_l = (unsigned int) dml_floor(refcyc_per_line_delivery_l, + 1); + ASSERT(disp_dlg_regs->refcyc_per_line_delivery_pre_l < (unsigned int) dml_pow(2, 13)); + ASSERT(disp_dlg_regs->refcyc_per_line_delivery_l < (unsigned int) dml_pow(2, 13)); + + disp_dlg_regs->refcyc_per_line_delivery_pre_c = (unsigned int) dml_floor(refcyc_per_line_delivery_pre_c, + 1); + disp_dlg_regs->refcyc_per_line_delivery_c = (unsigned int) dml_floor(refcyc_per_line_delivery_c, + 1); + ASSERT(disp_dlg_regs->refcyc_per_line_delivery_pre_c < (unsigned int) dml_pow(2, 13)); + ASSERT(disp_dlg_regs->refcyc_per_line_delivery_c < (unsigned int) dml_pow(2, 13)); + + disp_dlg_regs->chunk_hdl_adjust_cur0 = 3; + disp_dlg_regs->dst_y_offset_cur0 = 0; + disp_dlg_regs->chunk_hdl_adjust_cur1 = 3; + disp_dlg_regs->dst_y_offset_cur1 = 0; + + disp_dlg_regs->xfc_reg_transfer_delay = xfc_transfer_delay; + disp_dlg_regs->xfc_reg_precharge_delay = xfc_precharge_delay; + disp_dlg_regs->xfc_reg_remote_surface_flip_latency = xfc_remote_surface_flip_latency; + disp_dlg_regs->xfc_reg_prefetch_margin = dml_ceil(xfc_prefetch_margin * refclk_freq_in_mhz, + 1); + + // slave has to have this value also set to off + if (src->xfc_enable && !src->xfc_slave) + disp_dlg_regs->dst_y_delta_drq_limit = dml_ceil(xfc_dst_y_delta_drq_limit, 1); + else + disp_dlg_regs->dst_y_delta_drq_limit = 0x7fff; // off + + disp_ttu_regs->refcyc_per_req_delivery_pre_l = (unsigned int) (refcyc_per_req_delivery_pre_l + * dml_pow(2, 10)); + disp_ttu_regs->refcyc_per_req_delivery_l = (unsigned int) (refcyc_per_req_delivery_l + * dml_pow(2, 10)); + disp_ttu_regs->refcyc_per_req_delivery_pre_c = (unsigned int) (refcyc_per_req_delivery_pre_c + * dml_pow(2, 10)); + disp_ttu_regs->refcyc_per_req_delivery_c = (unsigned int) (refcyc_per_req_delivery_c + * dml_pow(2, 10)); + disp_ttu_regs->refcyc_per_req_delivery_pre_cur0 = + (unsigned int) (refcyc_per_req_delivery_pre_cur0 * dml_pow(2, 10)); + disp_ttu_regs->refcyc_per_req_delivery_cur0 = (unsigned int) (refcyc_per_req_delivery_cur0 + * dml_pow(2, 10)); + disp_ttu_regs->refcyc_per_req_delivery_pre_cur1 = + (unsigned int) (refcyc_per_req_delivery_pre_cur1 * dml_pow(2, 10)); + disp_ttu_regs->refcyc_per_req_delivery_cur1 = (unsigned int) (refcyc_per_req_delivery_cur1 + * dml_pow(2, 10)); + disp_ttu_regs->qos_level_low_wm = 0; + ASSERT(disp_ttu_regs->qos_level_low_wm < dml_pow(2, 14)); + disp_ttu_regs->qos_level_high_wm = (unsigned int) (4.0 * (double) htotal + * ref_freq_to_pix_freq); + ASSERT(disp_ttu_regs->qos_level_high_wm < dml_pow(2, 14)); + + disp_ttu_regs->qos_level_flip = 14; + disp_ttu_regs->qos_level_fixed_l = 8; + disp_ttu_regs->qos_level_fixed_c = 8; + disp_ttu_regs->qos_level_fixed_cur0 = 8; + disp_ttu_regs->qos_ramp_disable_l = 0; + disp_ttu_regs->qos_ramp_disable_c = 0; + disp_ttu_regs->qos_ramp_disable_cur0 = 0; + + disp_ttu_regs->min_ttu_vblank = min_ttu_vblank * refclk_freq_in_mhz; + ASSERT(disp_ttu_regs->min_ttu_vblank < dml_pow(2, 24)); + + print__ttu_regs_st(mode_lib, *disp_ttu_regs); + print__dlg_regs_st(mode_lib, *disp_dlg_regs); +} + +void dml_rq_dlg_get_dlg_reg(struct display_mode_lib *mode_lib, + display_dlg_regs_st *dlg_regs, + display_ttu_regs_st *ttu_regs, + display_e2e_pipe_params_st *e2e_pipe_param, + const unsigned int num_pipes, + const unsigned int pipe_idx, + const bool cstate_en, + const bool pstate_en, + const bool vm_en, + const bool ignore_viewport_pos, + const bool immediate_flip_support) +{ + display_rq_params_st rq_param = {0}; + display_dlg_sys_params_st dlg_sys_param = {0}; + + // Get watermark and Tex. + dlg_sys_param.t_urg_wm_us = get_wm_urgent(mode_lib, e2e_pipe_param, num_pipes); + dlg_sys_param.deepsleep_dcfclk_mhz = get_clk_dcf_deepsleep(mode_lib, + e2e_pipe_param, + num_pipes); + dlg_sys_param.t_extra_us = get_urgent_extra_latency(mode_lib, e2e_pipe_param, num_pipes); + dlg_sys_param.mem_trip_us = get_wm_memory_trip(mode_lib, e2e_pipe_param, num_pipes); + dlg_sys_param.t_mclk_wm_us = get_wm_dram_clock_change(mode_lib, e2e_pipe_param, num_pipes); + dlg_sys_param.t_sr_wm_us = get_wm_stutter_enter_exit(mode_lib, e2e_pipe_param, num_pipes); + dlg_sys_param.total_flip_bw = get_total_immediate_flip_bw(mode_lib, + e2e_pipe_param, + num_pipes); + dlg_sys_param.total_flip_bytes = get_total_immediate_flip_bytes(mode_lib, + e2e_pipe_param, + num_pipes); + dlg_sys_param.t_srx_delay_us = mode_lib->ip.dcfclk_cstate_latency + / dlg_sys_param.deepsleep_dcfclk_mhz; // TODO: Deprecated + + print__dlg_sys_params_st(mode_lib, dlg_sys_param); + + // system parameter calculation done + + dml_print("DML_DLG: Calculation for pipe[%d] start\n\n", pipe_idx); + dml_rq_dlg_get_rq_params(mode_lib, &rq_param, e2e_pipe_param[pipe_idx].pipe.src); + dml_rq_dlg_get_dlg_params(mode_lib, + e2e_pipe_param, + num_pipes, + pipe_idx, + dlg_regs, + ttu_regs, + rq_param.dlg, + dlg_sys_param, + cstate_en, + pstate_en, + vm_en, + ignore_viewport_pos, + immediate_flip_support); + dml_print("DML_DLG: Calculation for pipe[%d] end\n", pipe_idx); +} + +void dml_rq_dlg_get_arb_params(struct display_mode_lib *mode_lib, display_arb_params_st *arb_param) +{ + memset(arb_param, 0, sizeof(*arb_param)); + arb_param->max_req_outstanding = 256; + arb_param->min_req_outstanding = 68; + arb_param->sat_level_us = 60; +} + +void calculate_ttu_cursor(struct display_mode_lib *mode_lib, + double *refcyc_per_req_delivery_pre_cur, + double *refcyc_per_req_delivery_cur, + double refclk_freq_in_mhz, + double ref_freq_to_pix_freq, + double hscale_pixel_rate_l, + double hscl_ratio, + double vratio_pre_l, + double vratio_l, + unsigned int cur_width, + enum cursor_bpp cur_bpp) +{ + unsigned int cur_src_width = cur_width; + unsigned int cur_req_size = 0; + unsigned int cur_req_width = 0; + double cur_width_ub = 0.0; + double cur_req_per_width = 0.0; + double hactive_cur = 0.0; + + ASSERT(cur_src_width <= 256); + + *refcyc_per_req_delivery_pre_cur = 0.0; + *refcyc_per_req_delivery_cur = 0.0; + if (cur_src_width > 0) { + unsigned int cur_bit_per_pixel = 0; + + if (cur_bpp == dm_cur_2bit) { + cur_req_size = 64; // byte + cur_bit_per_pixel = 2; + } else { // 32bit + cur_bit_per_pixel = 32; + if (cur_src_width >= 1 && cur_src_width <= 16) + cur_req_size = 64; + else if (cur_src_width >= 17 && cur_src_width <= 31) + cur_req_size = 128; + else + cur_req_size = 256; + } + + cur_req_width = (double) cur_req_size / ((double) cur_bit_per_pixel / 8.0); + cur_width_ub = dml_ceil((double) cur_src_width / (double) cur_req_width, 1) + * (double) cur_req_width; + cur_req_per_width = cur_width_ub / (double) cur_req_width; + hactive_cur = (double) cur_src_width / hscl_ratio; // FIXME: oswin to think about what to do for cursor + + if (vratio_pre_l <= 1.0) { + *refcyc_per_req_delivery_pre_cur = hactive_cur * ref_freq_to_pix_freq + / (double) cur_req_per_width; + } else { + *refcyc_per_req_delivery_pre_cur = (double) refclk_freq_in_mhz + * (double) cur_src_width / hscale_pixel_rate_l + / (double) cur_req_per_width; + } + + ASSERT(*refcyc_per_req_delivery_pre_cur < dml_pow(2, 13)); + + if (vratio_l <= 1.0) { + *refcyc_per_req_delivery_cur = hactive_cur * ref_freq_to_pix_freq + / (double) cur_req_per_width; + } else { + *refcyc_per_req_delivery_cur = (double) refclk_freq_in_mhz + * (double) cur_src_width / hscale_pixel_rate_l + / (double) cur_req_per_width; + } + + dml_print("DML_DLG: %s: cur_req_width = %d\n", + __func__, + cur_req_width); + dml_print("DML_DLG: %s: cur_width_ub = %3.2f\n", + __func__, + cur_width_ub); + dml_print("DML_DLG: %s: cur_req_per_width = %3.2f\n", + __func__, + cur_req_per_width); + dml_print("DML_DLG: %s: hactive_cur = %3.2f\n", + __func__, + hactive_cur); + dml_print("DML_DLG: %s: refcyc_per_req_delivery_pre_cur = %3.2f\n", + __func__, + *refcyc_per_req_delivery_pre_cur); + dml_print("DML_DLG: %s: refcyc_per_req_delivery_cur = %3.2f\n", + __func__, + *refcyc_per_req_delivery_cur); + + ASSERT(*refcyc_per_req_delivery_cur < dml_pow(2, 13)); + } +} + +unsigned int dml_rq_dlg_get_calculated_vstartup(struct display_mode_lib *mode_lib, + display_e2e_pipe_params_st *e2e_pipe_param, + const unsigned int num_pipes, + const unsigned int pipe_idx) +{ + unsigned int vstartup_pipe[DC__NUM_PIPES__MAX]; + bool visited[DC__NUM_PIPES__MAX]; + unsigned int pipe_inst = 0; + unsigned int i, j, k; + + for (k = 0; k < num_pipes; ++k) + visited[k] = false; + + for (i = 0; i < num_pipes; i++) { + if (e2e_pipe_param[i].pipe.src.is_hsplit && !visited[i]) { + unsigned int grp = e2e_pipe_param[i].pipe.src.hsplit_grp; + + for (j = i; j < num_pipes; j++) { + if (e2e_pipe_param[j].pipe.src.hsplit_grp == grp + && e2e_pipe_param[j].pipe.src.is_hsplit + && !visited[j]) { + vstartup_pipe[j] = get_vstartup_calculated(mode_lib, + e2e_pipe_param, + num_pipes, + pipe_inst); + visited[j] = true; + } + } + + pipe_inst++; + } + + if (!visited[i]) { + vstartup_pipe[i] = get_vstartup_calculated(mode_lib, + e2e_pipe_param, + num_pipes, + pipe_inst); + visited[i] = true; + pipe_inst++; + } + } + + return vstartup_pipe[pipe_idx]; + +} + +void dml_rq_dlg_get_row_heights(struct display_mode_lib *mode_lib, + unsigned int *o_dpte_row_height, + unsigned int *o_meta_row_height, + unsigned int vp_width, + unsigned int data_pitch, + int source_format, + int tiling, + int macro_tile_size, + int source_scan, + int is_chroma) +{ + display_data_rq_dlg_params_st rq_dlg_param; + display_data_rq_misc_params_st rq_misc_param; + display_data_rq_sizing_params_st rq_sizing_param; + + get_meta_and_pte_attr(mode_lib, + &rq_dlg_param, + &rq_misc_param, + &rq_sizing_param, + vp_width, + 0, // dummy + data_pitch, + 0, // dummy + source_format, + tiling, + macro_tile_size, + source_scan, + is_chroma); + + *o_dpte_row_height = rq_dlg_param.dpte_row_height; + *o_meta_row_height = rq_dlg_param.meta_row_height; +} diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_rq_dlg_calc.h b/drivers/gpu/drm/amd/display/dc/dml/display_rq_dlg_calc.h new file mode 100644 index 000000000000..efdd4c73d8f3 --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/dml/display_rq_dlg_calc.h @@ -0,0 +1,148 @@ +/* + * Copyright 2017 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 __DML2_DISPLAY_RQ_DLG_CALC_H__ +#define __DML2_DISPLAY_RQ_DLG_CALC_H__ + +#include "dml_common_defs.h" +#include "display_rq_dlg_helpers.h" + +struct display_mode_lib; + +// Function: dml_rq_dlg_get_rq_params +// Calculate requestor related parameters that register definition agnostic +// (i.e. this layer does try to separate real values from register definition) +// Input: +// pipe_src_param - pipe source configuration (e.g. vp, pitch, etc.) +// Output: +// rq_param - values that can be used to setup RQ (e.g. swath_height, plane1_addr, etc.) +// +void dml_rq_dlg_get_rq_params( + struct display_mode_lib *mode_lib, + display_rq_params_st *rq_param, + const display_pipe_source_params_st pipe_src_param); + +// Function: dml_rq_dlg_get_rq_reg +// Main entry point for test to get the register values out of this DML class. +// This function calls <get_rq_param> and <extract_rq_regs> fucntions to calculate +// and then populate the rq_regs struct +// Input: +// pipe_src_param - pipe source configuration (e.g. vp, pitch, etc.) +// Output: +// rq_regs - struct that holds all the RQ registers field value. +// See also: <display_rq_regs_st> +void dml_rq_dlg_get_rq_reg( + struct display_mode_lib *mode_lib, + display_rq_regs_st *rq_regs, + const display_pipe_source_params_st pipe_src_param); + +// Function: dml_rq_dlg_get_dlg_params +// Calculate deadline related parameters +// +void dml_rq_dlg_get_dlg_params(struct display_mode_lib *mode_lib, + const display_e2e_pipe_params_st *e2e_pipe_param, + const unsigned int num_pipes, + const unsigned int pipe_idx, + display_dlg_regs_st *disp_dlg_regs, + display_ttu_regs_st *disp_ttu_regs, + const display_rq_dlg_params_st rq_dlg_param, + const display_dlg_sys_params_st dlg_sys_param, + const bool cstate_en, + const bool pstate_en, + const bool vm_en, + const bool ignore_viewport_pos, + const bool immediate_flip_support); + +// Function: dml_rq_dlg_get_dlg_param_prefetch +// For flip_bw programming guide change, now dml needs to calculate the flip_bytes and prefetch_bw +// for ALL pipes and use this info to calculate the prefetch programming. +// Output: prefetch_param.prefetch_bw and flip_bytes +void dml_rq_dlg_get_dlg_params_prefetch( + struct display_mode_lib *mode_lib, + display_dlg_prefetch_param_st *prefetch_param, + display_rq_dlg_params_st rq_dlg_param, + display_dlg_sys_params_st dlg_sys_param, + display_e2e_pipe_params_st e2e_pipe_param, + const bool cstate_en, + const bool pstate_en, + const bool vm_en); + +// Function: dml_rq_dlg_get_dlg_reg +// Calculate and return DLG and TTU register struct given the system setting +// Output: +// dlg_regs - output DLG register struct +// ttu_regs - output DLG TTU register struct +// Input: +// e2e_pipe_param - "compacted" array of e2e pipe param struct +// num_pipes - num of active "pipe" or "route" +// pipe_idx - index that identifies the e2e_pipe_param that corresponding to this dlg +// cstate - 0: when calculate min_ttu_vblank it is assumed cstate is not required. 1: Normal mode, cstate is considered. +// Added for legacy or unrealistic timing tests. +void dml_rq_dlg_get_dlg_reg( + struct display_mode_lib *mode_lib, + display_dlg_regs_st *dlg_regs, + display_ttu_regs_st *ttu_regs, + display_e2e_pipe_params_st *e2e_pipe_param, + const unsigned int num_pipes, + const unsigned int pipe_idx, + const bool cstate_en, + const bool pstate_en, + const bool vm_en, + const bool ignore_viewport_pos, + const bool immediate_flip_support); + +// Function: dml_rq_dlg_get_calculated_vstartup +// Calculate and return vstartup +// Output: +// unsigned int vstartup +// Input: +// e2e_pipe_param - "compacted" array of e2e pipe param struct +// num_pipes - num of active "pipe" or "route" +// pipe_idx - index that identifies the e2e_pipe_param that corresponding to this dlg +// NOTE: this MUST be called after setting the prefetch mode! +unsigned int dml_rq_dlg_get_calculated_vstartup( + struct display_mode_lib *mode_lib, + display_e2e_pipe_params_st *e2e_pipe_param, + const unsigned int num_pipes, + const unsigned int pipe_idx); + +// Function: dml_rq_dlg_get_row_heights +// Calculate dpte and meta row heights +void dml_rq_dlg_get_row_heights( + struct display_mode_lib *mode_lib, + unsigned int *o_dpte_row_height, + unsigned int *o_meta_row_height, + unsigned int vp_width, + unsigned int data_pitch, + int source_format, + int tiling, + int macro_tile_size, + int source_scan, + int is_chroma); + +// Function: dml_rq_dlg_get_arb_params +void dml_rq_dlg_get_arb_params(struct display_mode_lib *mode_lib, display_arb_params_st *arb_param); + +#endif diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_rq_dlg_helpers.c b/drivers/gpu/drm/amd/display/dc/dml/display_rq_dlg_helpers.c new file mode 100644 index 000000000000..189052e911fc --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/dml/display_rq_dlg_helpers.c @@ -0,0 +1,392 @@ +/* + * Copyright 2017 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 "display_rq_dlg_helpers.h" + +void print__rq_params_st(struct display_mode_lib *mode_lib, display_rq_params_st rq_param) +{ + dml_print("DML_RQ_DLG_CALC: ***************************\n"); + dml_print("DML_RQ_DLG_CALC: DISPLAY_RQ_PARAM_ST\n"); + dml_print("DML_RQ_DLG_CALC: <LUMA>\n"); + print__data_rq_sizing_params_st(mode_lib, rq_param.sizing.rq_l); + dml_print("DML_RQ_DLG_CALC: <CHROMA> ===\n"); + print__data_rq_sizing_params_st(mode_lib, rq_param.sizing.rq_c); + + dml_print("DML_RQ_DLG_CALC: <LUMA>\n"); + print__data_rq_dlg_params_st(mode_lib, rq_param.dlg.rq_l); + dml_print("DML_RQ_DLG_CALC: <CHROMA>\n"); + print__data_rq_dlg_params_st(mode_lib, rq_param.dlg.rq_c); + + dml_print("DML_RQ_DLG_CALC: <LUMA>\n"); + print__data_rq_misc_params_st(mode_lib, rq_param.misc.rq_l); + dml_print("DML_RQ_DLG_CALC: <CHROMA>\n"); + print__data_rq_misc_params_st(mode_lib, rq_param.misc.rq_c); + dml_print("DML_RQ_DLG_CALC: ***************************\n"); +} + +void print__data_rq_sizing_params_st(struct display_mode_lib *mode_lib, display_data_rq_sizing_params_st rq_sizing) +{ + dml_print("DML_RQ_DLG_CALC: =====================================\n"); + dml_print("DML_RQ_DLG_CALC: DISPLAY_DATA_RQ_SIZING_PARAM_ST\n"); + dml_print("DML_RQ_DLG_CALC: chunk_bytes = %0d\n", rq_sizing.chunk_bytes); + dml_print("DML_RQ_DLG_CALC: min_chunk_bytes = %0d\n", rq_sizing.min_chunk_bytes); + dml_print("DML_RQ_DLG_CALC: meta_chunk_bytes = %0d\n", rq_sizing.meta_chunk_bytes); + dml_print( + "DML_RQ_DLG_CALC: min_meta_chunk_bytes = %0d\n", + rq_sizing.min_meta_chunk_bytes); + dml_print("DML_RQ_DLG_CALC: mpte_group_bytes = %0d\n", rq_sizing.mpte_group_bytes); + dml_print("DML_RQ_DLG_CALC: dpte_group_bytes = %0d\n", rq_sizing.dpte_group_bytes); + dml_print("DML_RQ_DLG_CALC: =====================================\n"); +} + +void print__data_rq_dlg_params_st(struct display_mode_lib *mode_lib, display_data_rq_dlg_params_st rq_dlg_param) +{ + dml_print("DML_RQ_DLG_CALC: =====================================\n"); + dml_print("DML_RQ_DLG_CALC: DISPLAY_DATA_RQ_DLG_PARAM_ST\n"); + dml_print( + "DML_RQ_DLG_CALC: swath_width_ub = %0d\n", + rq_dlg_param.swath_width_ub); + dml_print( + "DML_RQ_DLG_CALC: swath_height = %0d\n", + rq_dlg_param.swath_height); + dml_print( + "DML_RQ_DLG_CALC: req_per_swath_ub = %0d\n", + rq_dlg_param.req_per_swath_ub); + dml_print( + "DML_RQ_DLG_CALC: meta_pte_bytes_per_frame_ub = %0d\n", + rq_dlg_param.meta_pte_bytes_per_frame_ub); + dml_print( + "DML_RQ_DLG_CALC: dpte_req_per_row_ub = %0d\n", + rq_dlg_param.dpte_req_per_row_ub); + dml_print( + "DML_RQ_DLG_CALC: dpte_groups_per_row_ub = %0d\n", + rq_dlg_param.dpte_groups_per_row_ub); + dml_print( + "DML_RQ_DLG_CALC: dpte_row_height = %0d\n", + rq_dlg_param.dpte_row_height); + dml_print( + "DML_RQ_DLG_CALC: dpte_bytes_per_row_ub = %0d\n", + rq_dlg_param.dpte_bytes_per_row_ub); + dml_print( + "DML_RQ_DLG_CALC: meta_chunks_per_row_ub = %0d\n", + rq_dlg_param.meta_chunks_per_row_ub); + dml_print( + "DML_RQ_DLG_CALC: meta_req_per_row_ub = %0d\n", + rq_dlg_param.meta_req_per_row_ub); + dml_print( + "DML_RQ_DLG_CALC: meta_row_height = %0d\n", + rq_dlg_param.meta_row_height); + dml_print( + "DML_RQ_DLG_CALC: meta_bytes_per_row_ub = %0d\n", + rq_dlg_param.meta_bytes_per_row_ub); + dml_print("DML_RQ_DLG_CALC: =====================================\n"); +} + +void print__data_rq_misc_params_st(struct display_mode_lib *mode_lib, display_data_rq_misc_params_st rq_misc_param) +{ + dml_print("DML_RQ_DLG_CALC: =====================================\n"); + dml_print("DML_RQ_DLG_CALC: DISPLAY_DATA_RQ_MISC_PARAM_ST\n"); + dml_print( + "DML_RQ_DLG_CALC: full_swath_bytes = %0d\n", + rq_misc_param.full_swath_bytes); + dml_print( + "DML_RQ_DLG_CALC: stored_swath_bytes = %0d\n", + rq_misc_param.stored_swath_bytes); + dml_print("DML_RQ_DLG_CALC: blk256_width = %0d\n", rq_misc_param.blk256_width); + dml_print("DML_RQ_DLG_CALC: blk256_height = %0d\n", rq_misc_param.blk256_height); + dml_print("DML_RQ_DLG_CALC: req_width = %0d\n", rq_misc_param.req_width); + dml_print("DML_RQ_DLG_CALC: req_height = %0d\n", rq_misc_param.req_height); + dml_print("DML_RQ_DLG_CALC: =====================================\n"); +} + +void print__rq_dlg_params_st(struct display_mode_lib *mode_lib, display_rq_dlg_params_st rq_dlg_param) +{ + dml_print("DML_RQ_DLG_CALC: =====================================\n"); + dml_print("DML_RQ_DLG_CALC: DISPLAY_RQ_DLG_PARAM_ST\n"); + dml_print("DML_RQ_DLG_CALC: <LUMA>\n"); + print__data_rq_dlg_params_st(mode_lib, rq_dlg_param.rq_l); + dml_print("DML_RQ_DLG_CALC: <CHROMA>\n"); + print__data_rq_dlg_params_st(mode_lib, rq_dlg_param.rq_c); + dml_print("DML_RQ_DLG_CALC: =====================================\n"); +} + +void print__dlg_sys_params_st(struct display_mode_lib *mode_lib, display_dlg_sys_params_st dlg_sys_param) +{ + dml_print("DML_RQ_DLG_CALC: =====================================\n"); + dml_print("DML_RQ_DLG_CALC: DISPLAY_RQ_DLG_PARAM_ST\n"); + dml_print("DML_RQ_DLG_CALC: t_mclk_wm_us = %3.2f\n", dlg_sys_param.t_mclk_wm_us); + dml_print("DML_RQ_DLG_CALC: t_urg_wm_us = %3.2f\n", dlg_sys_param.t_urg_wm_us); + dml_print("DML_RQ_DLG_CALC: t_sr_wm_us = %3.2f\n", dlg_sys_param.t_sr_wm_us); + dml_print("DML_RQ_DLG_CALC: t_extra_us = %3.2f\n", dlg_sys_param.t_extra_us); + dml_print( + "DML_RQ_DLG_CALC: t_srx_delay_us = %3.2f\n", + dlg_sys_param.t_srx_delay_us); + dml_print( + "DML_RQ_DLG_CALC: deepsleep_dcfclk_mhz = %3.2f\n", + dlg_sys_param.deepsleep_dcfclk_mhz); + dml_print( + "DML_RQ_DLG_CALC: total_flip_bw = %3.2f\n", + dlg_sys_param.total_flip_bw); + dml_print( + "DML_RQ_DLG_CALC: total_flip_bytes = %i\n", + dlg_sys_param.total_flip_bytes); + dml_print("DML_RQ_DLG_CALC: =====================================\n"); +} + +void print__data_rq_regs_st(struct display_mode_lib *mode_lib, display_data_rq_regs_st rq_regs) +{ + dml_print("DML_RQ_DLG_CALC: =====================================\n"); + dml_print("DML_RQ_DLG_CALC: DISPLAY_DATA_RQ_REGS_ST\n"); + dml_print("DML_RQ_DLG_CALC: chunk_size = 0x%0x\n", rq_regs.chunk_size); + dml_print("DML_RQ_DLG_CALC: min_chunk_size = 0x%0x\n", rq_regs.min_chunk_size); + dml_print("DML_RQ_DLG_CALC: meta_chunk_size = 0x%0x\n", rq_regs.meta_chunk_size); + dml_print( + "DML_RQ_DLG_CALC: min_meta_chunk_size = 0x%0x\n", + rq_regs.min_meta_chunk_size); + dml_print("DML_RQ_DLG_CALC: dpte_group_size = 0x%0x\n", rq_regs.dpte_group_size); + dml_print("DML_RQ_DLG_CALC: mpte_group_size = 0x%0x\n", rq_regs.mpte_group_size); + dml_print("DML_RQ_DLG_CALC: swath_height = 0x%0x\n", rq_regs.swath_height); + dml_print( + "DML_RQ_DLG_CALC: pte_row_height_linear = 0x%0x\n", + rq_regs.pte_row_height_linear); + dml_print("DML_RQ_DLG_CALC: =====================================\n"); +} + +void print__rq_regs_st(struct display_mode_lib *mode_lib, display_rq_regs_st rq_regs) +{ + dml_print("DML_RQ_DLG_CALC: =====================================\n"); + dml_print("DML_RQ_DLG_CALC: DISPLAY_RQ_REGS_ST\n"); + dml_print("DML_RQ_DLG_CALC: <LUMA>\n"); + print__data_rq_regs_st(mode_lib, rq_regs.rq_regs_l); + dml_print("DML_RQ_DLG_CALC: <CHROMA>\n"); + print__data_rq_regs_st(mode_lib, rq_regs.rq_regs_c); + dml_print("DML_RQ_DLG_CALC: drq_expansion_mode = 0x%0x\n", rq_regs.drq_expansion_mode); + dml_print("DML_RQ_DLG_CALC: prq_expansion_mode = 0x%0x\n", rq_regs.prq_expansion_mode); + dml_print("DML_RQ_DLG_CALC: mrq_expansion_mode = 0x%0x\n", rq_regs.mrq_expansion_mode); + dml_print("DML_RQ_DLG_CALC: crq_expansion_mode = 0x%0x\n", rq_regs.crq_expansion_mode); + dml_print("DML_RQ_DLG_CALC: plane1_base_address = 0x%0x\n", rq_regs.plane1_base_address); + dml_print("DML_RQ_DLG_CALC: =====================================\n"); +} + +void print__dlg_regs_st(struct display_mode_lib *mode_lib, display_dlg_regs_st dlg_regs) +{ + dml_print("DML_RQ_DLG_CALC: =====================================\n"); + dml_print("DML_RQ_DLG_CALC: DISPLAY_DLG_REGS_ST\n"); + dml_print( + "DML_RQ_DLG_CALC: refcyc_h_blank_end = 0x%0x\n", + dlg_regs.refcyc_h_blank_end); + dml_print( + "DML_RQ_DLG_CALC: dlg_vblank_end = 0x%0x\n", + dlg_regs.dlg_vblank_end); + dml_print( + "DML_RQ_DLG_CALC: min_dst_y_next_start = 0x%0x\n", + dlg_regs.min_dst_y_next_start); + dml_print( + "DML_RQ_DLG_CALC: refcyc_per_htotal = 0x%0x\n", + dlg_regs.refcyc_per_htotal); + dml_print( + "DML_RQ_DLG_CALC: refcyc_x_after_scaler = 0x%0x\n", + dlg_regs.refcyc_x_after_scaler); + dml_print( + "DML_RQ_DLG_CALC: dst_y_after_scaler = 0x%0x\n", + dlg_regs.dst_y_after_scaler); + dml_print( + "DML_RQ_DLG_CALC: dst_y_prefetch = 0x%0x\n", + dlg_regs.dst_y_prefetch); + dml_print( + "DML_RQ_DLG_CALC: dst_y_per_vm_vblank = 0x%0x\n", + dlg_regs.dst_y_per_vm_vblank); + dml_print( + "DML_RQ_DLG_CALC: dst_y_per_row_vblank = 0x%0x\n", + dlg_regs.dst_y_per_row_vblank); + dml_print( + "DML_RQ_DLG_CALC: dst_y_per_vm_flip = 0x%0x\n", + dlg_regs.dst_y_per_vm_flip); + dml_print( + "DML_RQ_DLG_CALC: dst_y_per_row_flip = 0x%0x\n", + dlg_regs.dst_y_per_row_flip); + dml_print( + "DML_RQ_DLG_CALC: ref_freq_to_pix_freq = 0x%0x\n", + dlg_regs.ref_freq_to_pix_freq); + dml_print( + "DML_RQ_DLG_CALC: vratio_prefetch = 0x%0x\n", + dlg_regs.vratio_prefetch); + dml_print( + "DML_RQ_DLG_CALC: vratio_prefetch_c = 0x%0x\n", + dlg_regs.vratio_prefetch_c); + dml_print( + "DML_RQ_DLG_CALC: refcyc_per_pte_group_vblank_l = 0x%0x\n", + dlg_regs.refcyc_per_pte_group_vblank_l); + dml_print( + "DML_RQ_DLG_CALC: refcyc_per_pte_group_vblank_c = 0x%0x\n", + dlg_regs.refcyc_per_pte_group_vblank_c); + dml_print( + "DML_RQ_DLG_CALC: refcyc_per_meta_chunk_vblank_l = 0x%0x\n", + dlg_regs.refcyc_per_meta_chunk_vblank_l); + dml_print( + "DML_RQ_DLG_CALC: refcyc_per_meta_chunk_vblank_c = 0x%0x\n", + dlg_regs.refcyc_per_meta_chunk_vblank_c); + dml_print( + "DML_RQ_DLG_CALC: refcyc_per_pte_group_flip_l = 0x%0x\n", + dlg_regs.refcyc_per_pte_group_flip_l); + dml_print( + "DML_RQ_DLG_CALC: refcyc_per_pte_group_flip_c = 0x%0x\n", + dlg_regs.refcyc_per_pte_group_flip_c); + dml_print( + "DML_RQ_DLG_CALC: refcyc_per_meta_chunk_flip_l = 0x%0x\n", + dlg_regs.refcyc_per_meta_chunk_flip_l); + dml_print( + "DML_RQ_DLG_CALC: refcyc_per_meta_chunk_flip_c = 0x%0x\n", + dlg_regs.refcyc_per_meta_chunk_flip_c); + dml_print( + "DML_RQ_DLG_CALC: dst_y_per_pte_row_nom_l = 0x%0x\n", + dlg_regs.dst_y_per_pte_row_nom_l); + dml_print( + "DML_RQ_DLG_CALC: dst_y_per_pte_row_nom_c = 0x%0x\n", + dlg_regs.dst_y_per_pte_row_nom_c); + dml_print( + "DML_RQ_DLG_CALC: refcyc_per_pte_group_nom_l = 0x%0x\n", + dlg_regs.refcyc_per_pte_group_nom_l); + dml_print( + "DML_RQ_DLG_CALC: refcyc_per_pte_group_nom_c = 0x%0x\n", + dlg_regs.refcyc_per_pte_group_nom_c); + dml_print( + "DML_RQ_DLG_CALC: dst_y_per_meta_row_nom_l = 0x%0x\n", + dlg_regs.dst_y_per_meta_row_nom_l); + dml_print( + "DML_RQ_DLG_CALC: dst_y_per_meta_row_nom_c = 0x%0x\n", + dlg_regs.dst_y_per_meta_row_nom_c); + dml_print( + "DML_RQ_DLG_CALC: refcyc_per_meta_chunk_nom_l = 0x%0x\n", + dlg_regs.refcyc_per_meta_chunk_nom_l); + dml_print( + "DML_RQ_DLG_CALC: refcyc_per_meta_chunk_nom_c = 0x%0x\n", + dlg_regs.refcyc_per_meta_chunk_nom_c); + dml_print( + "DML_RQ_DLG_CALC: refcyc_per_line_delivery_pre_l = 0x%0x\n", + dlg_regs.refcyc_per_line_delivery_pre_l); + dml_print( + "DML_RQ_DLG_CALC: refcyc_per_line_delivery_pre_c = 0x%0x\n", + dlg_regs.refcyc_per_line_delivery_pre_c); + dml_print( + "DML_RQ_DLG_CALC: refcyc_per_line_delivery_l = 0x%0x\n", + dlg_regs.refcyc_per_line_delivery_l); + dml_print( + "DML_RQ_DLG_CALC: refcyc_per_line_delivery_c = 0x%0x\n", + dlg_regs.refcyc_per_line_delivery_c); + dml_print( + "DML_RQ_DLG_CALC: chunk_hdl_adjust_cur0 = 0x%0x\n", + dlg_regs.chunk_hdl_adjust_cur0); + dml_print( + "DML_RQ_DLG_CALC: dst_y_offset_cur1 = 0x%0x\n", + dlg_regs.dst_y_offset_cur1); + dml_print( + "DML_RQ_DLG_CALC: chunk_hdl_adjust_cur1 = 0x%0x\n", + dlg_regs.chunk_hdl_adjust_cur1); + dml_print( + "DML_RQ_DLG_CALC: vready_after_vcount0 = 0x%0x\n", + dlg_regs.vready_after_vcount0); + dml_print( + "DML_RQ_DLG_CALC: dst_y_delta_drq_limit = 0x%0x\n", + dlg_regs.dst_y_delta_drq_limit); + dml_print( + "DML_RQ_DLG_CALC: xfc_reg_transfer_delay = 0x%0x\n", + dlg_regs.xfc_reg_transfer_delay); + dml_print( + "DML_RQ_DLG_CALC: xfc_reg_precharge_delay = 0x%0x\n", + dlg_regs.xfc_reg_precharge_delay); + dml_print( + "DML_RQ_DLG_CALC: xfc_reg_remote_surface_flip_latency = 0x%0x\n", + dlg_regs.xfc_reg_remote_surface_flip_latency); + + dml_print("DML_RQ_DLG_CALC: =====================================\n"); +} + +void print__ttu_regs_st(struct display_mode_lib *mode_lib, display_ttu_regs_st ttu_regs) +{ + dml_print("DML_RQ_DLG_CALC: =====================================\n"); + dml_print("DML_RQ_DLG_CALC: DISPLAY_TTU_REGS_ST\n"); + dml_print( + "DML_RQ_DLG_CALC: qos_level_low_wm = 0x%0x\n", + ttu_regs.qos_level_low_wm); + dml_print( + "DML_RQ_DLG_CALC: qos_level_high_wm = 0x%0x\n", + ttu_regs.qos_level_high_wm); + dml_print( + "DML_RQ_DLG_CALC: min_ttu_vblank = 0x%0x\n", + ttu_regs.min_ttu_vblank); + dml_print( + "DML_RQ_DLG_CALC: qos_level_flip = 0x%0x\n", + ttu_regs.qos_level_flip); + dml_print( + "DML_RQ_DLG_CALC: refcyc_per_req_delivery_pre_l = 0x%0x\n", + ttu_regs.refcyc_per_req_delivery_pre_l); + dml_print( + "DML_RQ_DLG_CALC: refcyc_per_req_delivery_l = 0x%0x\n", + ttu_regs.refcyc_per_req_delivery_l); + dml_print( + "DML_RQ_DLG_CALC: refcyc_per_req_delivery_pre_c = 0x%0x\n", + ttu_regs.refcyc_per_req_delivery_pre_c); + dml_print( + "DML_RQ_DLG_CALC: refcyc_per_req_delivery_c = 0x%0x\n", + ttu_regs.refcyc_per_req_delivery_c); + dml_print( + "DML_RQ_DLG_CALC: refcyc_per_req_delivery_cur0 = 0x%0x\n", + ttu_regs.refcyc_per_req_delivery_cur0); + dml_print( + "DML_RQ_DLG_CALC: refcyc_per_req_delivery_pre_cur0 = 0x%0x\n", + ttu_regs.refcyc_per_req_delivery_pre_cur0); + dml_print( + "DML_RQ_DLG_CALC: refcyc_per_req_delivery_cur1 = 0x%0x\n", + ttu_regs.refcyc_per_req_delivery_cur1); + dml_print( + "DML_RQ_DLG_CALC: refcyc_per_req_delivery_pre_cur1 = 0x%0x\n", + ttu_regs.refcyc_per_req_delivery_pre_cur1); + dml_print( + "DML_RQ_DLG_CALC: qos_level_fixed_l = 0x%0x\n", + ttu_regs.qos_level_fixed_l); + dml_print( + "DML_RQ_DLG_CALC: qos_ramp_disable_l = 0x%0x\n", + ttu_regs.qos_ramp_disable_l); + dml_print( + "DML_RQ_DLG_CALC: qos_level_fixed_c = 0x%0x\n", + ttu_regs.qos_level_fixed_c); + dml_print( + "DML_RQ_DLG_CALC: qos_ramp_disable_c = 0x%0x\n", + ttu_regs.qos_ramp_disable_c); + dml_print( + "DML_RQ_DLG_CALC: qos_level_fixed_cur0 = 0x%0x\n", + ttu_regs.qos_level_fixed_cur0); + dml_print( + "DML_RQ_DLG_CALC: qos_ramp_disable_cur0 = 0x%0x\n", + ttu_regs.qos_ramp_disable_cur0); + dml_print( + "DML_RQ_DLG_CALC: qos_level_fixed_cur1 = 0x%0x\n", + ttu_regs.qos_level_fixed_cur1); + dml_print( + "DML_RQ_DLG_CALC: qos_ramp_disable_cur1 = 0x%0x\n", + ttu_regs.qos_ramp_disable_cur1); + dml_print("DML_RQ_DLG_CALC: =====================================\n"); +} diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_rq_dlg_helpers.h b/drivers/gpu/drm/amd/display/dc/dml/display_rq_dlg_helpers.h new file mode 100644 index 000000000000..1f24db830737 --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/dml/display_rq_dlg_helpers.h @@ -0,0 +1,47 @@ +/* + * Copyright 2017 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 __DISPLAY_RQ_DLG_HELPERS_H__ +#define __DISPLAY_RQ_DLG_HELPERS_H__ + +#include "dml_common_defs.h" +#include "display_mode_lib.h" + +/* Function: Printer functions + * Print various struct + */ +void print__rq_params_st(struct display_mode_lib *mode_lib, display_rq_params_st rq_param); +void print__data_rq_sizing_params_st(struct display_mode_lib *mode_lib, display_data_rq_sizing_params_st rq_sizing); +void print__data_rq_dlg_params_st(struct display_mode_lib *mode_lib, display_data_rq_dlg_params_st rq_dlg_param); +void print__data_rq_misc_params_st(struct display_mode_lib *mode_lib, display_data_rq_misc_params_st rq_misc_param); +void print__rq_dlg_params_st(struct display_mode_lib *mode_lib, display_rq_dlg_params_st rq_dlg_param); +void print__dlg_sys_params_st(struct display_mode_lib *mode_lib, display_dlg_sys_params_st dlg_sys_param); + +void print__data_rq_regs_st(struct display_mode_lib *mode_lib, display_data_rq_regs_st data_rq_regs); +void print__rq_regs_st(struct display_mode_lib *mode_lib, display_rq_regs_st rq_regs); +void print__dlg_regs_st(struct display_mode_lib *mode_lib, display_dlg_regs_st dlg_regs); +void print__ttu_regs_st(struct display_mode_lib *mode_lib, display_ttu_regs_st ttu_regs); + +#endif diff --git a/drivers/gpu/drm/amd/display/dc/dml/dml1_display_rq_dlg_calc.c b/drivers/gpu/drm/amd/display/dc/dml/dml1_display_rq_dlg_calc.c new file mode 100644 index 000000000000..1e4b1e383401 --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/dml/dml1_display_rq_dlg_calc.c @@ -0,0 +1,1905 @@ +/* + * Copyright 2017 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 "dml1_display_rq_dlg_calc.h" +#include "display_mode_lib.h" + +#include "dml_inline_defs.h" + +static unsigned int get_bytes_per_element(enum source_format_class source_format, bool is_chroma) +{ + unsigned int ret_val = 0; + + if (source_format == dm_444_16) { + if (!is_chroma) + ret_val = 2; + } else if (source_format == dm_444_32) { + if (!is_chroma) + ret_val = 4; + } else if (source_format == dm_444_64) { + if (!is_chroma) + ret_val = 8; + } else if (source_format == dm_420_8) { + if (is_chroma) + ret_val = 2; + else + ret_val = 1; + } else if (source_format == dm_420_10) { + if (is_chroma) + ret_val = 4; + else + ret_val = 2; + } + return ret_val; +} + +static bool is_dual_plane(enum source_format_class source_format) +{ + bool ret_val = 0; + + if ((source_format == dm_420_8) || (source_format == dm_420_10)) + ret_val = 1; + + return ret_val; +} + +static void get_blk256_size( + unsigned int *blk256_width, + unsigned int *blk256_height, + unsigned int bytes_per_element) +{ + if (bytes_per_element == 1) { + *blk256_width = 16; + *blk256_height = 16; + } else if (bytes_per_element == 2) { + *blk256_width = 16; + *blk256_height = 8; + } else if (bytes_per_element == 4) { + *blk256_width = 8; + *blk256_height = 8; + } else if (bytes_per_element == 8) { + *blk256_width = 8; + *blk256_height = 4; + } +} + +static double get_refcyc_per_delivery( + struct display_mode_lib *mode_lib, + double refclk_freq_in_mhz, + double pclk_freq_in_mhz, + unsigned int recout_width, + double vratio, + double hscale_pixel_rate, + unsigned int delivery_width, + unsigned int req_per_swath_ub) +{ + double refcyc_per_delivery = 0.0; + + if (vratio <= 1.0) { + refcyc_per_delivery = (double) refclk_freq_in_mhz * (double) recout_width + / pclk_freq_in_mhz / (double) req_per_swath_ub; + } else { + refcyc_per_delivery = (double) refclk_freq_in_mhz * (double) delivery_width + / (double) hscale_pixel_rate / (double) req_per_swath_ub; + } + + DTRACE("DLG: %s: refclk_freq_in_mhz = %3.2f", __func__, refclk_freq_in_mhz); + DTRACE("DLG: %s: pclk_freq_in_mhz = %3.2f", __func__, pclk_freq_in_mhz); + DTRACE("DLG: %s: recout_width = %d", __func__, recout_width); + DTRACE("DLG: %s: vratio = %3.2f", __func__, vratio); + DTRACE("DLG: %s: req_per_swath_ub = %d", __func__, req_per_swath_ub); + DTRACE("DLG: %s: refcyc_per_delivery= %3.2f", __func__, refcyc_per_delivery); + + return refcyc_per_delivery; + +} + +static double get_vratio_pre( + struct display_mode_lib *mode_lib, + unsigned int max_num_sw, + unsigned int max_partial_sw, + unsigned int swath_height, + double vinit, + double l_sw) +{ + double prefill = dml_floor(vinit, 1); + double vratio_pre = 1.0; + + vratio_pre = (max_num_sw * swath_height + max_partial_sw) / l_sw; + + if (swath_height > 4) { + double tmp0 = (max_num_sw * swath_height) / (l_sw - (prefill - 3.0) / 2.0); + + if (tmp0 > vratio_pre) + vratio_pre = tmp0; + } + + DTRACE("DLG: %s: max_num_sw = %0d", __func__, max_num_sw); + DTRACE("DLG: %s: max_partial_sw = %0d", __func__, max_partial_sw); + DTRACE("DLG: %s: swath_height = %0d", __func__, swath_height); + DTRACE("DLG: %s: vinit = %3.2f", __func__, vinit); + DTRACE("DLG: %s: vratio_pre = %3.2f", __func__, vratio_pre); + + if (vratio_pre < 1.0) { + DTRACE("WARNING_DLG: %s: vratio_pre=%3.2f < 1.0, set to 1.0", __func__, vratio_pre); + vratio_pre = 1.0; + } + + if (vratio_pre > 4.0) { + DTRACE( + "WARNING_DLG: %s: vratio_pre=%3.2f > 4.0 (max scaling ratio). set to 4.0", + __func__, + vratio_pre); + vratio_pre = 4.0; + } + + return vratio_pre; +} + +static void get_swath_need( + struct display_mode_lib *mode_lib, + unsigned int *max_num_sw, + unsigned int *max_partial_sw, + unsigned int swath_height, + double vinit) +{ + double prefill = dml_floor(vinit, 1); + unsigned int max_partial_sw_int; + + DTRACE("DLG: %s: swath_height = %0d", __func__, swath_height); + DTRACE("DLG: %s: vinit = %3.2f", __func__, vinit); + + ASSERT(prefill > 0.0 && prefill <= 8.0); + + *max_num_sw = (unsigned int) (dml_ceil((prefill - 1.0) / (double) swath_height, 1) + 1.0); /* prefill has to be >= 1 */ + max_partial_sw_int = + (prefill == 1) ? + (swath_height - 1) : + ((unsigned int) (prefill - 2.0) % swath_height); + *max_partial_sw = (max_partial_sw_int < 1) ? 1 : max_partial_sw_int; /* ensure minimum of 1 is used */ + + DTRACE("DLG: %s: max_num_sw = %0d", __func__, *max_num_sw); + DTRACE("DLG: %s: max_partial_sw = %0d", __func__, *max_partial_sw); +} + +static unsigned int get_blk_size_bytes(const enum source_macro_tile_size tile_size) +{ + if (tile_size == dm_256k_tile) + return (256 * 1024); + else if (tile_size == dm_64k_tile) + return (64 * 1024); + else + return (4 * 1024); +} + +static void extract_rq_sizing_regs( + struct display_mode_lib *mode_lib, + struct _vcs_dpi_display_data_rq_regs_st *rq_regs, + const struct _vcs_dpi_display_data_rq_sizing_params_st rq_sizing) +{ + DTRACE("DLG: %s: rq_sizing param", __func__); + print__data_rq_sizing_params_st(mode_lib, rq_sizing); + + rq_regs->chunk_size = dml_log2(rq_sizing.chunk_bytes) - 10; + + if (rq_sizing.min_chunk_bytes == 0) + rq_regs->min_chunk_size = 0; + else + rq_regs->min_chunk_size = dml_log2(rq_sizing.min_chunk_bytes) - 8 + 1; + + rq_regs->meta_chunk_size = dml_log2(rq_sizing.meta_chunk_bytes) - 10; + if (rq_sizing.min_meta_chunk_bytes == 0) + rq_regs->min_meta_chunk_size = 0; + else + rq_regs->min_meta_chunk_size = dml_log2(rq_sizing.min_meta_chunk_bytes) - 6 + 1; + + rq_regs->dpte_group_size = dml_log2(rq_sizing.dpte_group_bytes) - 6; + rq_regs->mpte_group_size = dml_log2(rq_sizing.mpte_group_bytes) - 6; +} + +void dml1_extract_rq_regs( + struct display_mode_lib *mode_lib, + struct _vcs_dpi_display_rq_regs_st *rq_regs, + const struct _vcs_dpi_display_rq_params_st rq_param) +{ + unsigned int detile_buf_size_in_bytes = mode_lib->ip.det_buffer_size_kbytes * 1024; + unsigned int detile_buf_plane1_addr = 0; + + extract_rq_sizing_regs(mode_lib, &(rq_regs->rq_regs_l), rq_param.sizing.rq_l); + if (rq_param.yuv420) + extract_rq_sizing_regs(mode_lib, &(rq_regs->rq_regs_c), rq_param.sizing.rq_c); + + rq_regs->rq_regs_l.swath_height = dml_log2(rq_param.dlg.rq_l.swath_height); + rq_regs->rq_regs_c.swath_height = dml_log2(rq_param.dlg.rq_c.swath_height); + + /* FIXME: take the max between luma, chroma chunk size? + * okay for now, as we are setting chunk_bytes to 8kb anyways + */ + if (rq_param.sizing.rq_l.chunk_bytes >= 32 * 1024) { /*32kb */ + rq_regs->drq_expansion_mode = 0; + } else { + rq_regs->drq_expansion_mode = 2; + } + rq_regs->prq_expansion_mode = 1; + rq_regs->mrq_expansion_mode = 1; + rq_regs->crq_expansion_mode = 1; + + if (rq_param.yuv420) { + if ((double) rq_param.misc.rq_l.stored_swath_bytes + / (double) rq_param.misc.rq_c.stored_swath_bytes <= 1.5) { + detile_buf_plane1_addr = (detile_buf_size_in_bytes / 2.0 / 64.0); /* half to chroma */ + } else { + detile_buf_plane1_addr = dml_round_to_multiple( + (unsigned int) ((2.0 * detile_buf_size_in_bytes) / 3.0), + 256, + 0) / 64.0; /* 2/3 to chroma */ + } + } + rq_regs->plane1_base_address = detile_buf_plane1_addr; +} + +static void handle_det_buf_split( + struct display_mode_lib *mode_lib, + struct _vcs_dpi_display_rq_params_st *rq_param, + const struct _vcs_dpi_display_pipe_source_params_st pipe_src_param) +{ + unsigned int total_swath_bytes = 0; + unsigned int swath_bytes_l = 0; + unsigned int swath_bytes_c = 0; + unsigned int full_swath_bytes_packed_l = 0; + unsigned int full_swath_bytes_packed_c = 0; + bool req128_l = 0; + bool req128_c = 0; + bool surf_linear = (pipe_src_param.sw_mode == dm_sw_linear); + bool surf_vert = (pipe_src_param.source_scan == dm_vert); + unsigned int log2_swath_height_l = 0; + unsigned int log2_swath_height_c = 0; + unsigned int detile_buf_size_in_bytes = mode_lib->ip.det_buffer_size_kbytes * 1024; + + full_swath_bytes_packed_l = rq_param->misc.rq_l.full_swath_bytes; + full_swath_bytes_packed_c = rq_param->misc.rq_c.full_swath_bytes; + + if (rq_param->yuv420_10bpc) { + full_swath_bytes_packed_l = dml_round_to_multiple( + rq_param->misc.rq_l.full_swath_bytes * 2 / 3, + 256, + 1) + 256; + full_swath_bytes_packed_c = dml_round_to_multiple( + rq_param->misc.rq_c.full_swath_bytes * 2 / 3, + 256, + 1) + 256; + } + + if (rq_param->yuv420) { + total_swath_bytes = 2 * full_swath_bytes_packed_l + 2 * full_swath_bytes_packed_c; + + if (total_swath_bytes <= detile_buf_size_in_bytes) { /*full 256b request */ + req128_l = 0; + req128_c = 0; + swath_bytes_l = full_swath_bytes_packed_l; + swath_bytes_c = full_swath_bytes_packed_c; + } else { /*128b request (for luma only for yuv420 8bpc) */ + req128_l = 1; + req128_c = 0; + swath_bytes_l = full_swath_bytes_packed_l / 2; + swath_bytes_c = full_swath_bytes_packed_c; + } + + /* Bug workaround, luma and chroma req size needs to be the same. (see: DEGVIDCN10-137) + * TODO: Remove after rtl fix + */ + if (req128_l == 1) { + req128_c = 1; + DTRACE("DLG: %s: bug workaround DEGVIDCN10-137", __func__); + } + + /* Note: assumption, the config that pass in will fit into + * the detiled buffer. + */ + } else { + total_swath_bytes = 2 * full_swath_bytes_packed_l; + + if (total_swath_bytes <= detile_buf_size_in_bytes) + req128_l = 0; + else + req128_l = 1; + + swath_bytes_l = total_swath_bytes; + swath_bytes_c = 0; + } + rq_param->misc.rq_l.stored_swath_bytes = swath_bytes_l; + rq_param->misc.rq_c.stored_swath_bytes = swath_bytes_c; + + if (surf_linear) { + log2_swath_height_l = 0; + log2_swath_height_c = 0; + } else if (!surf_vert) { + log2_swath_height_l = dml_log2(rq_param->misc.rq_l.blk256_height) - req128_l; + log2_swath_height_c = dml_log2(rq_param->misc.rq_c.blk256_height) - req128_c; + } else { + log2_swath_height_l = dml_log2(rq_param->misc.rq_l.blk256_width) - req128_l; + log2_swath_height_c = dml_log2(rq_param->misc.rq_c.blk256_width) - req128_c; + } + rq_param->dlg.rq_l.swath_height = 1 << log2_swath_height_l; + rq_param->dlg.rq_c.swath_height = 1 << log2_swath_height_c; + + DTRACE("DLG: %s: req128_l = %0d", __func__, req128_l); + DTRACE("DLG: %s: req128_c = %0d", __func__, req128_c); + DTRACE("DLG: %s: full_swath_bytes_packed_l = %0d", __func__, full_swath_bytes_packed_l); + DTRACE("DLG: %s: full_swath_bytes_packed_c = %0d", __func__, full_swath_bytes_packed_c); +} + +/* Need refactor. */ +static void dml1_rq_dlg_get_row_heights( + struct display_mode_lib *mode_lib, + unsigned int *o_dpte_row_height, + unsigned int *o_meta_row_height, + unsigned int vp_width, + unsigned int data_pitch, + int source_format, + int tiling, + int macro_tile_size, + int source_scan, + int is_chroma) +{ + bool surf_linear = (tiling == dm_sw_linear); + bool surf_vert = (source_scan == dm_vert); + + unsigned int bytes_per_element = get_bytes_per_element( + (enum source_format_class) source_format, + is_chroma); + unsigned int log2_bytes_per_element = dml_log2(bytes_per_element); + unsigned int blk256_width = 0; + unsigned int blk256_height = 0; + + unsigned int log2_blk256_height; + unsigned int blk_bytes; + unsigned int log2_blk_bytes; + unsigned int log2_blk_height; + unsigned int log2_blk_width; + unsigned int log2_meta_req_bytes; + unsigned int log2_meta_req_height; + unsigned int log2_meta_req_width; + unsigned int log2_meta_row_height; + unsigned int log2_vmpg_bytes; + unsigned int dpte_buf_in_pte_reqs; + unsigned int log2_vmpg_height; + unsigned int log2_vmpg_width; + unsigned int log2_dpte_req_height_ptes; + unsigned int log2_dpte_req_width_ptes; + unsigned int log2_dpte_req_height; + unsigned int log2_dpte_req_width; + unsigned int log2_dpte_row_height_linear; + unsigned int log2_dpte_row_height; + unsigned int dpte_req_width; + + if (surf_linear) { + blk256_width = 256; + blk256_height = 1; + } else { + get_blk256_size(&blk256_width, &blk256_height, bytes_per_element); + } + + log2_blk256_height = dml_log2((double) blk256_height); + blk_bytes = surf_linear ? + 256 : get_blk_size_bytes((enum source_macro_tile_size) macro_tile_size); + log2_blk_bytes = dml_log2((double) blk_bytes); + log2_blk_height = 0; + log2_blk_width = 0; + + /* remember log rule + * "+" in log is multiply + * "-" in log is divide + * "/2" is like square root + * blk is vertical biased + */ + if (tiling != dm_sw_linear) + log2_blk_height = log2_blk256_height + + dml_ceil((double) (log2_blk_bytes - 8) / 2.0, 1); + else + log2_blk_height = 0; /* blk height of 1 */ + + log2_blk_width = log2_blk_bytes - log2_bytes_per_element - log2_blk_height; + + /* ------- */ + /* meta */ + /* ------- */ + log2_meta_req_bytes = 6; /* meta request is 64b and is 8x8byte meta element */ + + /* each 64b meta request for dcn is 8x8 meta elements and + * a meta element covers one 256b block of the the data surface. + */ + log2_meta_req_height = log2_blk256_height + 3; /* meta req is 8x8 */ + log2_meta_req_width = log2_meta_req_bytes + 8 - log2_bytes_per_element + - log2_meta_req_height; + log2_meta_row_height = 0; + + /* the dimensions of a meta row are meta_row_width x meta_row_height in elements. + * calculate upper bound of the meta_row_width + */ + if (!surf_vert) + log2_meta_row_height = log2_meta_req_height; + else + log2_meta_row_height = log2_meta_req_width; + + *o_meta_row_height = 1 << log2_meta_row_height; + + /* ------ */ + /* dpte */ + /* ------ */ + log2_vmpg_bytes = dml_log2(mode_lib->soc.vmm_page_size_bytes); + dpte_buf_in_pte_reqs = mode_lib->ip.dpte_buffer_size_in_pte_reqs; + + log2_vmpg_height = 0; + log2_vmpg_width = 0; + log2_dpte_req_height_ptes = 0; + log2_dpte_req_width_ptes = 0; + log2_dpte_req_height = 0; + log2_dpte_req_width = 0; + log2_dpte_row_height_linear = 0; + log2_dpte_row_height = 0; + dpte_req_width = 0; /* 64b dpte req width in data element */ + + if (surf_linear) + log2_vmpg_height = 0; /* one line high */ + else + log2_vmpg_height = (log2_vmpg_bytes - 8) / 2 + log2_blk256_height; + log2_vmpg_width = log2_vmpg_bytes - log2_bytes_per_element - log2_vmpg_height; + + /* only 3 possible shapes for dpte request in dimensions of ptes: 8x1, 4x2, 2x4. */ + if (log2_blk_bytes <= log2_vmpg_bytes) + log2_dpte_req_height_ptes = 0; + else if (log2_blk_height - log2_vmpg_height >= 2) + log2_dpte_req_height_ptes = 2; + else + log2_dpte_req_height_ptes = log2_blk_height - log2_vmpg_height; + log2_dpte_req_width_ptes = 3 - log2_dpte_req_height_ptes; + + ASSERT((log2_dpte_req_width_ptes == 3 && log2_dpte_req_height_ptes == 0) || /* 8x1 */ + (log2_dpte_req_width_ptes == 2 && log2_dpte_req_height_ptes == 1) || /* 4x2 */ + (log2_dpte_req_width_ptes == 1 && log2_dpte_req_height_ptes == 2)); /* 2x4 */ + + /* the dpte request dimensions in data elements is dpte_req_width x dpte_req_height + * log2_wmpg_width is how much 1 pte represent, now trying to calculate how much 64b pte req represent + */ + log2_dpte_req_height = log2_vmpg_height + log2_dpte_req_height_ptes; + log2_dpte_req_width = log2_vmpg_width + log2_dpte_req_width_ptes; + dpte_req_width = 1 << log2_dpte_req_width; + + /* calculate pitch dpte row buffer can hold + * round the result down to a power of two. + */ + if (surf_linear) { + log2_dpte_row_height_linear = dml_floor( + dml_log2(dpte_buf_in_pte_reqs * dpte_req_width / data_pitch), + 1); + + ASSERT(log2_dpte_row_height_linear >= 3); + + if (log2_dpte_row_height_linear > 7) + log2_dpte_row_height_linear = 7; + + log2_dpte_row_height = log2_dpte_row_height_linear; + } else { + /* the upper bound of the dpte_row_width without dependency on viewport position follows. */ + if (!surf_vert) + log2_dpte_row_height = log2_dpte_req_height; + else + log2_dpte_row_height = + (log2_blk_width < log2_dpte_req_width) ? + log2_blk_width : log2_dpte_req_width; + } + + /* From programming guide: + * There is a special case of saving only half of ptes returned due to buffer space limits. + * this case applies to 4 and 8bpe in horizontal access of a vp_width greater than 2560+16 + * when the pte request is 2x4 ptes (which happens when vmpg_bytes =4kb and tile blk_bytes >=64kb). + */ + if (!surf_vert && vp_width > (2560 + 16) && bytes_per_element >= 4 && log2_vmpg_bytes == 12 + && log2_blk_bytes >= 16) + log2_dpte_row_height = log2_dpte_row_height - 1; /*half of the full height */ + + *o_dpte_row_height = 1 << log2_dpte_row_height; +} + +static void get_surf_rq_param( + struct display_mode_lib *mode_lib, + struct _vcs_dpi_display_data_rq_sizing_params_st *rq_sizing_param, + struct _vcs_dpi_display_data_rq_dlg_params_st *rq_dlg_param, + struct _vcs_dpi_display_data_rq_misc_params_st *rq_misc_param, + const struct _vcs_dpi_display_pipe_source_params_st pipe_src_param, + bool is_chroma) +{ + bool mode_422 = 0; + unsigned int vp_width = 0; + unsigned int vp_height = 0; + unsigned int data_pitch = 0; + unsigned int meta_pitch = 0; + unsigned int ppe = mode_422 ? 2 : 1; + bool surf_linear; + bool surf_vert; + unsigned int bytes_per_element; + unsigned int log2_bytes_per_element; + unsigned int blk256_width; + unsigned int blk256_height; + unsigned int log2_blk256_width; + unsigned int log2_blk256_height; + unsigned int blk_bytes; + unsigned int log2_blk_bytes; + unsigned int log2_blk_height; + unsigned int log2_blk_width; + unsigned int log2_meta_req_bytes; + unsigned int log2_meta_req_height; + unsigned int log2_meta_req_width; + unsigned int meta_req_width; + unsigned int meta_req_height; + unsigned int log2_meta_row_height; + unsigned int meta_row_width_ub; + unsigned int log2_meta_chunk_bytes; + unsigned int log2_meta_chunk_height; + unsigned int log2_meta_chunk_width; + unsigned int log2_min_meta_chunk_bytes; + unsigned int min_meta_chunk_width; + unsigned int meta_chunk_width; + unsigned int meta_chunk_per_row_int; + unsigned int meta_row_remainder; + unsigned int meta_chunk_threshold; + unsigned int meta_blk_bytes; + unsigned int meta_blk_height; + unsigned int meta_blk_width; + unsigned int meta_surface_bytes; + unsigned int vmpg_bytes; + unsigned int meta_pte_req_per_frame_ub; + unsigned int meta_pte_bytes_per_frame_ub; + unsigned int log2_vmpg_bytes; + unsigned int dpte_buf_in_pte_reqs; + unsigned int log2_vmpg_height; + unsigned int log2_vmpg_width; + unsigned int log2_dpte_req_height_ptes; + unsigned int log2_dpte_req_width_ptes; + unsigned int log2_dpte_req_height; + unsigned int log2_dpte_req_width; + unsigned int log2_dpte_row_height_linear; + unsigned int log2_dpte_row_height; + unsigned int log2_dpte_group_width; + unsigned int dpte_row_width_ub; + unsigned int dpte_row_height; + unsigned int dpte_req_height; + unsigned int dpte_req_width; + unsigned int dpte_group_width; + unsigned int log2_dpte_group_bytes; + unsigned int log2_dpte_group_length; + unsigned int func_meta_row_height, func_dpte_row_height; + + /* FIXME check if ppe apply for both luma and chroma in 422 case */ + if (is_chroma) { + vp_width = pipe_src_param.viewport_width_c / ppe; + vp_height = pipe_src_param.viewport_height_c; + data_pitch = pipe_src_param.data_pitch_c; + meta_pitch = pipe_src_param.meta_pitch_c; + } else { + vp_width = pipe_src_param.viewport_width / ppe; + vp_height = pipe_src_param.viewport_height; + data_pitch = pipe_src_param.data_pitch; + meta_pitch = pipe_src_param.meta_pitch; + } + + rq_sizing_param->chunk_bytes = 8192; + + if (rq_sizing_param->chunk_bytes == 64 * 1024) + rq_sizing_param->min_chunk_bytes = 0; + else + rq_sizing_param->min_chunk_bytes = 1024; + + rq_sizing_param->meta_chunk_bytes = 2048; + rq_sizing_param->min_meta_chunk_bytes = 256; + + rq_sizing_param->mpte_group_bytes = 2048; + + surf_linear = (pipe_src_param.sw_mode == dm_sw_linear); + surf_vert = (pipe_src_param.source_scan == dm_vert); + + bytes_per_element = get_bytes_per_element( + (enum source_format_class) pipe_src_param.source_format, + is_chroma); + log2_bytes_per_element = dml_log2(bytes_per_element); + blk256_width = 0; + blk256_height = 0; + + if (surf_linear) { + blk256_width = 256 / bytes_per_element; + blk256_height = 1; + } else { + get_blk256_size(&blk256_width, &blk256_height, bytes_per_element); + } + + DTRACE("DLG: %s: surf_linear = %d", __func__, surf_linear); + DTRACE("DLG: %s: surf_vert = %d", __func__, surf_vert); + DTRACE("DLG: %s: blk256_width = %d", __func__, blk256_width); + DTRACE("DLG: %s: blk256_height = %d", __func__, blk256_height); + + log2_blk256_width = dml_log2((double) blk256_width); + log2_blk256_height = dml_log2((double) blk256_height); + blk_bytes = + surf_linear ? 256 : get_blk_size_bytes( + (enum source_macro_tile_size) pipe_src_param.macro_tile_size); + log2_blk_bytes = dml_log2((double) blk_bytes); + log2_blk_height = 0; + log2_blk_width = 0; + + /* remember log rule + * "+" in log is multiply + * "-" in log is divide + * "/2" is like square root + * blk is vertical biased + */ + if (pipe_src_param.sw_mode != dm_sw_linear) + log2_blk_height = log2_blk256_height + + dml_ceil((double) (log2_blk_bytes - 8) / 2.0, 1); + else + log2_blk_height = 0; /* blk height of 1 */ + + log2_blk_width = log2_blk_bytes - log2_bytes_per_element - log2_blk_height; + + if (!surf_vert) { + rq_dlg_param->swath_width_ub = dml_round_to_multiple(vp_width - 1, blk256_width, 1) + + blk256_width; + rq_dlg_param->req_per_swath_ub = rq_dlg_param->swath_width_ub >> log2_blk256_width; + } else { + rq_dlg_param->swath_width_ub = dml_round_to_multiple( + vp_height - 1, + blk256_height, + 1) + blk256_height; + rq_dlg_param->req_per_swath_ub = rq_dlg_param->swath_width_ub >> log2_blk256_height; + } + + if (!surf_vert) + rq_misc_param->full_swath_bytes = rq_dlg_param->swath_width_ub * blk256_height + * bytes_per_element; + else + rq_misc_param->full_swath_bytes = rq_dlg_param->swath_width_ub * blk256_width + * bytes_per_element; + + rq_misc_param->blk256_height = blk256_height; + rq_misc_param->blk256_width = blk256_width; + + /* ------- */ + /* meta */ + /* ------- */ + log2_meta_req_bytes = 6; /* meta request is 64b and is 8x8byte meta element */ + + /* each 64b meta request for dcn is 8x8 meta elements and + * a meta element covers one 256b block of the the data surface. + */ + log2_meta_req_height = log2_blk256_height + 3; /* meta req is 8x8 byte, each byte represent 1 blk256 */ + log2_meta_req_width = log2_meta_req_bytes + 8 - log2_bytes_per_element + - log2_meta_req_height; + meta_req_width = 1 << log2_meta_req_width; + meta_req_height = 1 << log2_meta_req_height; + log2_meta_row_height = 0; + meta_row_width_ub = 0; + + /* the dimensions of a meta row are meta_row_width x meta_row_height in elements. + * calculate upper bound of the meta_row_width + */ + if (!surf_vert) { + log2_meta_row_height = log2_meta_req_height; + meta_row_width_ub = dml_round_to_multiple(vp_width - 1, meta_req_width, 1) + + meta_req_width; + rq_dlg_param->meta_req_per_row_ub = meta_row_width_ub / meta_req_width; + } else { + log2_meta_row_height = log2_meta_req_width; + meta_row_width_ub = dml_round_to_multiple(vp_height - 1, meta_req_height, 1) + + meta_req_height; + rq_dlg_param->meta_req_per_row_ub = meta_row_width_ub / meta_req_height; + } + rq_dlg_param->meta_bytes_per_row_ub = rq_dlg_param->meta_req_per_row_ub * 64; + + log2_meta_chunk_bytes = dml_log2(rq_sizing_param->meta_chunk_bytes); + log2_meta_chunk_height = log2_meta_row_height; + + /*full sized meta chunk width in unit of data elements */ + log2_meta_chunk_width = log2_meta_chunk_bytes + 8 - log2_bytes_per_element + - log2_meta_chunk_height; + log2_min_meta_chunk_bytes = dml_log2(rq_sizing_param->min_meta_chunk_bytes); + min_meta_chunk_width = 1 + << (log2_min_meta_chunk_bytes + 8 - log2_bytes_per_element + - log2_meta_chunk_height); + meta_chunk_width = 1 << log2_meta_chunk_width; + meta_chunk_per_row_int = (unsigned int) (meta_row_width_ub / meta_chunk_width); + meta_row_remainder = meta_row_width_ub % meta_chunk_width; + meta_chunk_threshold = 0; + meta_blk_bytes = 4096; + meta_blk_height = blk256_height * 64; + meta_blk_width = meta_blk_bytes * 256 / bytes_per_element / meta_blk_height; + meta_surface_bytes = meta_pitch + * (dml_round_to_multiple(vp_height - 1, meta_blk_height, 1) + + meta_blk_height) * bytes_per_element / 256; + vmpg_bytes = mode_lib->soc.vmm_page_size_bytes; + meta_pte_req_per_frame_ub = (dml_round_to_multiple( + meta_surface_bytes - vmpg_bytes, + 8 * vmpg_bytes, + 1) + 8 * vmpg_bytes) / (8 * vmpg_bytes); + meta_pte_bytes_per_frame_ub = meta_pte_req_per_frame_ub * 64; /*64B mpte request */ + rq_dlg_param->meta_pte_bytes_per_frame_ub = meta_pte_bytes_per_frame_ub; + + DTRACE("DLG: %s: meta_blk_height = %d", __func__, meta_blk_height); + DTRACE("DLG: %s: meta_blk_width = %d", __func__, meta_blk_width); + DTRACE("DLG: %s: meta_surface_bytes = %d", __func__, meta_surface_bytes); + DTRACE("DLG: %s: meta_pte_req_per_frame_ub = %d", __func__, meta_pte_req_per_frame_ub); + DTRACE("DLG: %s: meta_pte_bytes_per_frame_ub = %d", __func__, meta_pte_bytes_per_frame_ub); + + if (!surf_vert) + meta_chunk_threshold = 2 * min_meta_chunk_width - meta_req_width; + else + meta_chunk_threshold = 2 * min_meta_chunk_width - meta_req_height; + + if (meta_row_remainder <= meta_chunk_threshold) + rq_dlg_param->meta_chunks_per_row_ub = meta_chunk_per_row_int + 1; + else + rq_dlg_param->meta_chunks_per_row_ub = meta_chunk_per_row_int + 2; + + rq_dlg_param->meta_row_height = 1 << log2_meta_row_height; + + /* ------ */ + /* dpte */ + /* ------ */ + log2_vmpg_bytes = dml_log2(mode_lib->soc.vmm_page_size_bytes); + dpte_buf_in_pte_reqs = mode_lib->ip.dpte_buffer_size_in_pte_reqs; + + log2_vmpg_height = 0; + log2_vmpg_width = 0; + log2_dpte_req_height_ptes = 0; + log2_dpte_req_width_ptes = 0; + log2_dpte_req_height = 0; + log2_dpte_req_width = 0; + log2_dpte_row_height_linear = 0; + log2_dpte_row_height = 0; + log2_dpte_group_width = 0; + dpte_row_width_ub = 0; + dpte_row_height = 0; + dpte_req_height = 0; /* 64b dpte req height in data element */ + dpte_req_width = 0; /* 64b dpte req width in data element */ + dpte_group_width = 0; + log2_dpte_group_bytes = 0; + log2_dpte_group_length = 0; + + if (surf_linear) + log2_vmpg_height = 0; /* one line high */ + else + log2_vmpg_height = (log2_vmpg_bytes - 8) / 2 + log2_blk256_height; + log2_vmpg_width = log2_vmpg_bytes - log2_bytes_per_element - log2_vmpg_height; + + /* only 3 possible shapes for dpte request in dimensions of ptes: 8x1, 4x2, 2x4. */ + if (log2_blk_bytes <= log2_vmpg_bytes) + log2_dpte_req_height_ptes = 0; + else if (log2_blk_height - log2_vmpg_height >= 2) + log2_dpte_req_height_ptes = 2; + else + log2_dpte_req_height_ptes = log2_blk_height - log2_vmpg_height; + log2_dpte_req_width_ptes = 3 - log2_dpte_req_height_ptes; + + /* Ensure we only have the 3 shapes */ + ASSERT((log2_dpte_req_width_ptes == 3 && log2_dpte_req_height_ptes == 0) || /* 8x1 */ + (log2_dpte_req_width_ptes == 2 && log2_dpte_req_height_ptes == 1) || /* 4x2 */ + (log2_dpte_req_width_ptes == 1 && log2_dpte_req_height_ptes == 2)); /* 2x4 */ + + /* The dpte request dimensions in data elements is dpte_req_width x dpte_req_height + * log2_vmpg_width is how much 1 pte represent, now calculating how much a 64b pte req represent + * That depends on the pte shape (i.e. 8x1, 4x2, 2x4) + */ + log2_dpte_req_height = log2_vmpg_height + log2_dpte_req_height_ptes; + log2_dpte_req_width = log2_vmpg_width + log2_dpte_req_width_ptes; + dpte_req_height = 1 << log2_dpte_req_height; + dpte_req_width = 1 << log2_dpte_req_width; + + /* calculate pitch dpte row buffer can hold + * round the result down to a power of two. + */ + if (surf_linear) { + log2_dpte_row_height_linear = dml_floor( + dml_log2(dpte_buf_in_pte_reqs * dpte_req_width / data_pitch), + 1); + + ASSERT(log2_dpte_row_height_linear >= 3); + + if (log2_dpte_row_height_linear > 7) + log2_dpte_row_height_linear = 7; + + log2_dpte_row_height = log2_dpte_row_height_linear; + rq_dlg_param->dpte_row_height = 1 << log2_dpte_row_height; + + /* For linear, the dpte row is pitch dependent and the pte requests wrap at the pitch boundary. + * the dpte_row_width_ub is the upper bound of data_pitch*dpte_row_height in elements with this unique buffering. + */ + dpte_row_width_ub = dml_round_to_multiple( + data_pitch * dpte_row_height - 1, + dpte_req_width, + 1) + dpte_req_width; + rq_dlg_param->dpte_req_per_row_ub = dpte_row_width_ub / dpte_req_width; + } else { + /* for tiled mode, row height is the same as req height and row store up to vp size upper bound */ + if (!surf_vert) { + log2_dpte_row_height = log2_dpte_req_height; + dpte_row_width_ub = dml_round_to_multiple(vp_width - 1, dpte_req_width, 1) + + dpte_req_width; + rq_dlg_param->dpte_req_per_row_ub = dpte_row_width_ub / dpte_req_width; + } else { + log2_dpte_row_height = + (log2_blk_width < log2_dpte_req_width) ? + log2_blk_width : log2_dpte_req_width; + dpte_row_width_ub = dml_round_to_multiple(vp_height - 1, dpte_req_height, 1) + + dpte_req_height; + rq_dlg_param->dpte_req_per_row_ub = dpte_row_width_ub / dpte_req_height; + } + rq_dlg_param->dpte_row_height = 1 << log2_dpte_row_height; + } + rq_dlg_param->dpte_bytes_per_row_ub = rq_dlg_param->dpte_req_per_row_ub * 64; + + /* From programming guide: + * There is a special case of saving only half of ptes returned due to buffer space limits. + * this case applies to 4 and 8bpe in horizontal access of a vp_width greater than 2560+16 + * when the pte request is 2x4 ptes (which happens when vmpg_bytes =4kb and tile blk_bytes >=64kb). + */ + if (!surf_vert && vp_width > (2560 + 16) && bytes_per_element >= 4 && log2_vmpg_bytes == 12 + && log2_blk_bytes >= 16) { + log2_dpte_row_height = log2_dpte_row_height - 1; /*half of the full height */ + rq_dlg_param->dpte_row_height = 1 << log2_dpte_row_height; + } + + /* the dpte_group_bytes is reduced for the specific case of vertical + * access of a tile surface that has dpte request of 8x1 ptes. + */ + if (!surf_linear & (log2_dpte_req_height_ptes == 0) & surf_vert) /*reduced, in this case, will have page fault within a group */ + rq_sizing_param->dpte_group_bytes = 512; + else + /*full size */ + rq_sizing_param->dpte_group_bytes = 2048; + + /*since pte request size is 64byte, the number of data pte requests per full sized group is as follows. */ + log2_dpte_group_bytes = dml_log2(rq_sizing_param->dpte_group_bytes); + log2_dpte_group_length = log2_dpte_group_bytes - 6; /*length in 64b requests */ + + /* full sized data pte group width in elements */ + if (!surf_vert) + log2_dpte_group_width = log2_dpte_group_length + log2_dpte_req_width; + else + log2_dpte_group_width = log2_dpte_group_length + log2_dpte_req_height; + + dpte_group_width = 1 << log2_dpte_group_width; + + /* since dpte groups are only aligned to dpte_req_width and not dpte_group_width, + * the upper bound for the dpte groups per row is as follows. + */ + rq_dlg_param->dpte_groups_per_row_ub = dml_ceil( + (double) dpte_row_width_ub / dpte_group_width, + 1); + + dml1_rq_dlg_get_row_heights( + mode_lib, + &func_dpte_row_height, + &func_meta_row_height, + vp_width, + data_pitch, + pipe_src_param.source_format, + pipe_src_param.sw_mode, + pipe_src_param.macro_tile_size, + pipe_src_param.source_scan, + is_chroma); + + /* Just a check to make sure this function and the new one give the same + * result. The standalone get_row_heights() function is based off of the + * code in this function so the same changes need to be made to both. + */ + if (rq_dlg_param->meta_row_height != func_meta_row_height) { + DTRACE( + "MISMATCH: rq_dlg_param->meta_row_height = %d", + rq_dlg_param->meta_row_height); + DTRACE("MISMATCH: func_meta_row_height = %d", func_meta_row_height); + ASSERT(0); + } + + if (rq_dlg_param->dpte_row_height != func_dpte_row_height) { + DTRACE( + "MISMATCH: rq_dlg_param->dpte_row_height = %d", + rq_dlg_param->dpte_row_height); + DTRACE("MISMATCH: func_dpte_row_height = %d", func_dpte_row_height); + ASSERT(0); + } +} + +void dml1_rq_dlg_get_rq_params( + struct display_mode_lib *mode_lib, + struct _vcs_dpi_display_rq_params_st *rq_param, + const struct _vcs_dpi_display_pipe_source_params_st pipe_src_param) +{ + /* get param for luma surface */ + rq_param->yuv420 = pipe_src_param.source_format == dm_420_8 + || pipe_src_param.source_format == dm_420_10; + rq_param->yuv420_10bpc = pipe_src_param.source_format == dm_420_10; + + get_surf_rq_param( + mode_lib, + &(rq_param->sizing.rq_l), + &(rq_param->dlg.rq_l), + &(rq_param->misc.rq_l), + pipe_src_param, + 0); + + if (is_dual_plane((enum source_format_class) pipe_src_param.source_format)) { + /* get param for chroma surface */ + get_surf_rq_param( + mode_lib, + &(rq_param->sizing.rq_c), + &(rq_param->dlg.rq_c), + &(rq_param->misc.rq_c), + pipe_src_param, + 1); + } + + /* calculate how to split the det buffer space between luma and chroma */ + handle_det_buf_split(mode_lib, rq_param, pipe_src_param); + print__rq_params_st(mode_lib, *rq_param); +} + +/* Note: currently taken in as is. + * Nice to decouple code from hw register implement and extract code that are repeated for luma and chroma. + */ +void dml1_rq_dlg_get_dlg_params( + struct display_mode_lib *mode_lib, + struct _vcs_dpi_display_dlg_regs_st *disp_dlg_regs, + struct _vcs_dpi_display_ttu_regs_st *disp_ttu_regs, + const struct _vcs_dpi_display_rq_dlg_params_st rq_dlg_param, + const struct _vcs_dpi_display_dlg_sys_params_st dlg_sys_param, + const struct _vcs_dpi_display_e2e_pipe_params_st e2e_pipe_param, + const bool cstate_en, + const bool pstate_en, + const bool vm_en, + const bool iflip_en) +{ + /* Timing */ + unsigned int htotal = e2e_pipe_param.pipe.dest.htotal; + unsigned int hblank_end = e2e_pipe_param.pipe.dest.hblank_end; + unsigned int vblank_start = e2e_pipe_param.pipe.dest.vblank_start; + unsigned int vblank_end = e2e_pipe_param.pipe.dest.vblank_end; + bool interlaced = e2e_pipe_param.pipe.dest.interlaced; + unsigned int min_vblank = mode_lib->ip.min_vblank_lines; + + double pclk_freq_in_mhz = e2e_pipe_param.pipe.dest.pixel_rate_mhz; + double refclk_freq_in_mhz = e2e_pipe_param.clks_cfg.refclk_mhz; + double dppclk_freq_in_mhz = e2e_pipe_param.clks_cfg.dppclk_mhz; + double dispclk_freq_in_mhz = e2e_pipe_param.clks_cfg.dispclk_mhz; + + double ref_freq_to_pix_freq; + double prefetch_xy_calc_in_dcfclk; + double min_dcfclk_mhz; + double t_calc_us; + double min_ttu_vblank; + double min_dst_y_ttu_vblank; + unsigned int dlg_vblank_start; + bool dcc_en; + bool dual_plane; + bool mode_422; + unsigned int access_dir; + unsigned int bytes_per_element_l; + unsigned int bytes_per_element_c; + unsigned int vp_height_l; + unsigned int vp_width_l; + unsigned int vp_height_c; + unsigned int vp_width_c; + unsigned int htaps_l; + unsigned int htaps_c; + double hratios_l; + double hratios_c; + double vratio_l; + double vratio_c; + double line_time_in_us; + double vinit_l; + double vinit_c; + double vinit_bot_l; + double vinit_bot_c; + unsigned int swath_height_l; + unsigned int swath_width_ub_l; + unsigned int dpte_bytes_per_row_ub_l; + unsigned int dpte_groups_per_row_ub_l; + unsigned int meta_pte_bytes_per_frame_ub_l; + unsigned int meta_bytes_per_row_ub_l; + unsigned int swath_height_c; + unsigned int swath_width_ub_c; + unsigned int dpte_bytes_per_row_ub_c; + unsigned int dpte_groups_per_row_ub_c; + unsigned int meta_chunks_per_row_ub_l; + unsigned int vupdate_offset; + unsigned int vupdate_width; + unsigned int vready_offset; + unsigned int dppclk_delay_subtotal; + unsigned int dispclk_delay_subtotal; + unsigned int pixel_rate_delay_subtotal; + unsigned int vstartup_start; + unsigned int dst_x_after_scaler; + unsigned int dst_y_after_scaler; + double line_wait; + double line_o; + double line_setup; + double line_calc; + double dst_y_prefetch; + double t_pre_us; + unsigned int vm_bytes; + unsigned int meta_row_bytes; + unsigned int max_num_sw_l; + unsigned int max_num_sw_c; + unsigned int max_partial_sw_l; + unsigned int max_partial_sw_c; + double max_vinit_l; + double max_vinit_c; + unsigned int lsw_l; + unsigned int lsw_c; + unsigned int sw_bytes_ub_l; + unsigned int sw_bytes_ub_c; + unsigned int sw_bytes; + unsigned int dpte_row_bytes; + double prefetch_bw; + double flip_bw; + double t_vm_us; + double t_r0_us; + double dst_y_per_vm_vblank; + double dst_y_per_row_vblank; + double min_dst_y_per_vm_vblank; + double min_dst_y_per_row_vblank; + double lsw; + double vratio_pre_l; + double vratio_pre_c; + unsigned int req_per_swath_ub_l; + unsigned int req_per_swath_ub_c; + unsigned int meta_row_height_l; + unsigned int swath_width_pixels_ub_l; + unsigned int swath_width_pixels_ub_c; + unsigned int scaler_rec_in_width_l; + unsigned int scaler_rec_in_width_c; + unsigned int dpte_row_height_l; + unsigned int dpte_row_height_c; + double hscale_pixel_rate_l; + double hscale_pixel_rate_c; + double min_hratio_fact_l; + double min_hratio_fact_c; + double refcyc_per_line_delivery_pre_l; + double refcyc_per_line_delivery_pre_c; + double refcyc_per_line_delivery_l; + double refcyc_per_line_delivery_c; + double refcyc_per_req_delivery_pre_l; + double refcyc_per_req_delivery_pre_c; + double refcyc_per_req_delivery_l; + double refcyc_per_req_delivery_c; + double refcyc_per_req_delivery_pre_cur0; + double refcyc_per_req_delivery_cur0; + unsigned int full_recout_width; + double hratios_cur0; + unsigned int cur0_src_width; + enum cursor_bpp cur0_bpp; + unsigned int cur0_req_size; + unsigned int cur0_req_width; + double cur0_width_ub; + double cur0_req_per_width; + double hactive_cur0; + + memset(disp_dlg_regs, 0, sizeof(*disp_dlg_regs)); + memset(disp_ttu_regs, 0, sizeof(*disp_ttu_regs)); + + DTRACE("DLG: %s: cstate_en = %d", __func__, cstate_en); + DTRACE("DLG: %s: pstate_en = %d", __func__, pstate_en); + DTRACE("DLG: %s: vm_en = %d", __func__, vm_en); + DTRACE("DLG: %s: iflip_en = %d", __func__, iflip_en); + + /* ------------------------- */ + /* Section 1.5.2.1: OTG dependent Params */ + /* ------------------------- */ + DTRACE("DLG: %s: dppclk_freq_in_mhz = %3.2f", __func__, dppclk_freq_in_mhz); + DTRACE("DLG: %s: dispclk_freq_in_mhz = %3.2f", __func__, dispclk_freq_in_mhz); + DTRACE("DLG: %s: refclk_freq_in_mhz = %3.2f", __func__, refclk_freq_in_mhz); + DTRACE("DLG: %s: pclk_freq_in_mhz = %3.2f", __func__, pclk_freq_in_mhz); + DTRACE("DLG: %s: interlaced = %d", __func__, interlaced); + + ref_freq_to_pix_freq = refclk_freq_in_mhz / pclk_freq_in_mhz; + ASSERT(ref_freq_to_pix_freq < 4.0); + disp_dlg_regs->ref_freq_to_pix_freq = + (unsigned int) (ref_freq_to_pix_freq * dml_pow(2, 19)); + disp_dlg_regs->refcyc_per_htotal = (unsigned int) (ref_freq_to_pix_freq * (double) htotal + * dml_pow(2, 8)); + disp_dlg_regs->refcyc_h_blank_end = (unsigned int) ((double) hblank_end + * (double) ref_freq_to_pix_freq); + ASSERT(disp_dlg_regs->refcyc_h_blank_end < (unsigned int) dml_pow(2, 13)); + disp_dlg_regs->dlg_vblank_end = interlaced ? (vblank_end / 2) : vblank_end; /* 15 bits */ + + prefetch_xy_calc_in_dcfclk = 24.0; /* FIXME: ip_param */ + min_dcfclk_mhz = dlg_sys_param.deepsleep_dcfclk_mhz; + t_calc_us = prefetch_xy_calc_in_dcfclk / min_dcfclk_mhz; + min_ttu_vblank = dlg_sys_param.t_urg_wm_us; + if (cstate_en) + min_ttu_vblank = dml_max(dlg_sys_param.t_sr_wm_us, min_ttu_vblank); + if (pstate_en) + min_ttu_vblank = dml_max(dlg_sys_param.t_mclk_wm_us, min_ttu_vblank); + min_ttu_vblank = min_ttu_vblank + t_calc_us; + + min_dst_y_ttu_vblank = min_ttu_vblank * pclk_freq_in_mhz / (double) htotal; + dlg_vblank_start = interlaced ? (vblank_start / 2) : vblank_start; + + disp_dlg_regs->min_dst_y_next_start = (unsigned int) (((double) dlg_vblank_start + + min_dst_y_ttu_vblank) * dml_pow(2, 2)); + ASSERT(disp_dlg_regs->min_dst_y_next_start < (unsigned int) dml_pow(2, 18)); + + DTRACE("DLG: %s: min_dcfclk_mhz = %3.2f", __func__, min_dcfclk_mhz); + DTRACE("DLG: %s: min_ttu_vblank = %3.2f", __func__, min_ttu_vblank); + DTRACE( + "DLG: %s: min_dst_y_ttu_vblank = %3.2f", + __func__, + min_dst_y_ttu_vblank); + DTRACE("DLG: %s: t_calc_us = %3.2f", __func__, t_calc_us); + DTRACE( + "DLG: %s: disp_dlg_regs->min_dst_y_next_start = 0x%0x", + __func__, + disp_dlg_regs->min_dst_y_next_start); + DTRACE( + "DLG: %s: ref_freq_to_pix_freq = %3.2f", + __func__, + ref_freq_to_pix_freq); + + /* ------------------------- */ + /* Section 1.5.2.2: Prefetch, Active and TTU */ + /* ------------------------- */ + /* Prefetch Calc */ + /* Source */ + dcc_en = e2e_pipe_param.pipe.src.dcc; + dual_plane = is_dual_plane( + (enum source_format_class) e2e_pipe_param.pipe.src.source_format); + mode_422 = 0; /* FIXME */ + access_dir = (e2e_pipe_param.pipe.src.source_scan == dm_vert); /* vp access direction: horizontal or vertical accessed */ + bytes_per_element_l = get_bytes_per_element( + (enum source_format_class) e2e_pipe_param.pipe.src.source_format, + 0); + bytes_per_element_c = get_bytes_per_element( + (enum source_format_class) e2e_pipe_param.pipe.src.source_format, + 1); + vp_height_l = e2e_pipe_param.pipe.src.viewport_height; + vp_width_l = e2e_pipe_param.pipe.src.viewport_width; + vp_height_c = e2e_pipe_param.pipe.src.viewport_height_c; + vp_width_c = e2e_pipe_param.pipe.src.viewport_width_c; + + /* Scaling */ + htaps_l = e2e_pipe_param.pipe.scale_taps.htaps; + htaps_c = e2e_pipe_param.pipe.scale_taps.htaps_c; + hratios_l = e2e_pipe_param.pipe.scale_ratio_depth.hscl_ratio; + hratios_c = e2e_pipe_param.pipe.scale_ratio_depth.hscl_ratio_c; + vratio_l = e2e_pipe_param.pipe.scale_ratio_depth.vscl_ratio; + vratio_c = e2e_pipe_param.pipe.scale_ratio_depth.vscl_ratio_c; + + line_time_in_us = (htotal / pclk_freq_in_mhz); + vinit_l = e2e_pipe_param.pipe.scale_ratio_depth.vinit; + vinit_c = e2e_pipe_param.pipe.scale_ratio_depth.vinit_c; + vinit_bot_l = e2e_pipe_param.pipe.scale_ratio_depth.vinit_bot; + vinit_bot_c = e2e_pipe_param.pipe.scale_ratio_depth.vinit_bot_c; + + swath_height_l = rq_dlg_param.rq_l.swath_height; + swath_width_ub_l = rq_dlg_param.rq_l.swath_width_ub; + dpte_bytes_per_row_ub_l = rq_dlg_param.rq_l.dpte_bytes_per_row_ub; + dpte_groups_per_row_ub_l = rq_dlg_param.rq_l.dpte_groups_per_row_ub; + meta_pte_bytes_per_frame_ub_l = rq_dlg_param.rq_l.meta_pte_bytes_per_frame_ub; + meta_bytes_per_row_ub_l = rq_dlg_param.rq_l.meta_bytes_per_row_ub; + + swath_height_c = rq_dlg_param.rq_c.swath_height; + swath_width_ub_c = rq_dlg_param.rq_c.swath_width_ub; + dpte_bytes_per_row_ub_c = rq_dlg_param.rq_c.dpte_bytes_per_row_ub; + dpte_groups_per_row_ub_c = rq_dlg_param.rq_c.dpte_groups_per_row_ub; + + meta_chunks_per_row_ub_l = rq_dlg_param.rq_l.meta_chunks_per_row_ub; + vupdate_offset = e2e_pipe_param.pipe.dest.vupdate_offset; + vupdate_width = e2e_pipe_param.pipe.dest.vupdate_width; + vready_offset = e2e_pipe_param.pipe.dest.vready_offset; + + dppclk_delay_subtotal = mode_lib->ip.dppclk_delay_subtotal; + dispclk_delay_subtotal = mode_lib->ip.dispclk_delay_subtotal; + pixel_rate_delay_subtotal = dppclk_delay_subtotal * pclk_freq_in_mhz / dppclk_freq_in_mhz + + dispclk_delay_subtotal * pclk_freq_in_mhz / dispclk_freq_in_mhz; + + vstartup_start = e2e_pipe_param.pipe.dest.vstartup_start; + + if (interlaced) + vstartup_start = vstartup_start / 2; + + if (vstartup_start >= min_vblank) { + DTRACE( + "WARNING_DLG: %s: vblank_start=%d vblank_end=%d", + __func__, + vblank_start, + vblank_end); + DTRACE( + "WARNING_DLG: %s: vstartup_start=%d should be less than min_vblank=%d", + __func__, + vstartup_start, + min_vblank); + min_vblank = vstartup_start + 1; + DTRACE( + "WARNING_DLG: %s: vstartup_start=%d should be less than min_vblank=%d", + __func__, + vstartup_start, + min_vblank); + } + + dst_x_after_scaler = 0; + dst_y_after_scaler = 0; + + if (e2e_pipe_param.pipe.src.is_hsplit) + dst_x_after_scaler = pixel_rate_delay_subtotal + + e2e_pipe_param.pipe.dest.recout_width; + else + dst_x_after_scaler = pixel_rate_delay_subtotal; + + if (e2e_pipe_param.dout.output_format == dm_420) + dst_y_after_scaler = 1; + else + dst_y_after_scaler = 0; + + if (dst_x_after_scaler >= htotal) { + dst_x_after_scaler = dst_x_after_scaler - htotal; + dst_y_after_scaler = dst_y_after_scaler + 1; + } + + DTRACE("DLG: %s: htotal = %d", __func__, htotal); + DTRACE( + "DLG: %s: pixel_rate_delay_subtotal = %d", + __func__, + pixel_rate_delay_subtotal); + DTRACE("DLG: %s: dst_x_after_scaler = %d", __func__, dst_x_after_scaler); + DTRACE("DLG: %s: dst_y_after_scaler = %d", __func__, dst_y_after_scaler); + + line_wait = mode_lib->soc.urgent_latency_us; + if (cstate_en) + line_wait = dml_max(mode_lib->soc.sr_enter_plus_exit_time_us, line_wait); + if (pstate_en) + line_wait = dml_max( + mode_lib->soc.dram_clock_change_latency_us + + mode_lib->soc.urgent_latency_us, + line_wait); + line_wait = line_wait / line_time_in_us; + + line_o = (double) dst_y_after_scaler + dst_x_after_scaler / (double) htotal; + line_setup = (double) (vupdate_offset + vupdate_width + vready_offset) / (double) htotal; + line_calc = t_calc_us / line_time_in_us; + + DTRACE( + "DLG: %s: soc.sr_enter_plus_exit_time_us = %3.2f", + __func__, + (double) mode_lib->soc.sr_enter_plus_exit_time_us); + DTRACE( + "DLG: %s: soc.dram_clock_change_latency_us = %3.2f", + __func__, + (double) mode_lib->soc.dram_clock_change_latency_us); + DTRACE( + "DLG: %s: soc.urgent_latency_us = %3.2f", + __func__, + mode_lib->soc.urgent_latency_us); + + DTRACE("DLG: %s: swath_height_l = %d", __func__, swath_height_l); + if (dual_plane) + DTRACE("DLG: %s: swath_height_c = %d", __func__, swath_height_c); + + DTRACE( + "DLG: %s: t_srx_delay_us = %3.2f", + __func__, + (double) dlg_sys_param.t_srx_delay_us); + DTRACE("DLG: %s: line_time_in_us = %3.2f", __func__, (double) line_time_in_us); + DTRACE("DLG: %s: vupdate_offset = %d", __func__, vupdate_offset); + DTRACE("DLG: %s: vupdate_width = %d", __func__, vupdate_width); + DTRACE("DLG: %s: vready_offset = %d", __func__, vready_offset); + DTRACE("DLG: %s: line_time_in_us = %3.2f", __func__, line_time_in_us); + DTRACE("DLG: %s: line_wait = %3.2f", __func__, line_wait); + DTRACE("DLG: %s: line_o = %3.2f", __func__, line_o); + DTRACE("DLG: %s: line_setup = %3.2f", __func__, line_setup); + DTRACE("DLG: %s: line_calc = %3.2f", __func__, line_calc); + + dst_y_prefetch = ((double) min_vblank - 1.0) + - (line_setup + line_calc + line_wait + line_o); + DTRACE("DLG: %s: dst_y_prefetch (before rnd) = %3.2f", __func__, dst_y_prefetch); + ASSERT(dst_y_prefetch >= 2.0); + + dst_y_prefetch = dml_floor(4.0 * (dst_y_prefetch + 0.125), 1) / 4; + DTRACE("DLG: %s: dst_y_prefetch (after rnd) = %3.2f", __func__, dst_y_prefetch); + + t_pre_us = dst_y_prefetch * line_time_in_us; + vm_bytes = 0; + meta_row_bytes = 0; + + if (dcc_en && vm_en) + vm_bytes = meta_pte_bytes_per_frame_ub_l; + if (dcc_en) + meta_row_bytes = meta_bytes_per_row_ub_l; + + max_num_sw_l = 0; + max_num_sw_c = 0; + max_partial_sw_l = 0; + max_partial_sw_c = 0; + + max_vinit_l = interlaced ? dml_max(vinit_l, vinit_bot_l) : vinit_l; + max_vinit_c = interlaced ? dml_max(vinit_c, vinit_bot_c) : vinit_c; + + get_swath_need(mode_lib, &max_num_sw_l, &max_partial_sw_l, swath_height_l, max_vinit_l); + if (dual_plane) + get_swath_need( + mode_lib, + &max_num_sw_c, + &max_partial_sw_c, + swath_height_c, + max_vinit_c); + + lsw_l = max_num_sw_l * swath_height_l + max_partial_sw_l; + lsw_c = max_num_sw_c * swath_height_c + max_partial_sw_c; + sw_bytes_ub_l = lsw_l * swath_width_ub_l * bytes_per_element_l; + sw_bytes_ub_c = lsw_c * swath_width_ub_c * bytes_per_element_c; + sw_bytes = 0; + dpte_row_bytes = 0; + + if (vm_en) { + if (dual_plane) + dpte_row_bytes = dpte_bytes_per_row_ub_l + dpte_bytes_per_row_ub_c; + else + dpte_row_bytes = dpte_bytes_per_row_ub_l; + } else { + dpte_row_bytes = 0; + } + + if (dual_plane) + sw_bytes = sw_bytes_ub_l + sw_bytes_ub_c; + else + sw_bytes = sw_bytes_ub_l; + + DTRACE("DLG: %s: sw_bytes_ub_l = %d", __func__, sw_bytes_ub_l); + DTRACE("DLG: %s: sw_bytes_ub_c = %d", __func__, sw_bytes_ub_c); + DTRACE("DLG: %s: sw_bytes = %d", __func__, sw_bytes); + DTRACE("DLG: %s: vm_bytes = %d", __func__, vm_bytes); + DTRACE("DLG: %s: meta_row_bytes = %d", __func__, meta_row_bytes); + DTRACE("DLG: %s: dpte_row_bytes = %d", __func__, dpte_row_bytes); + + prefetch_bw = (vm_bytes + 2 * dpte_row_bytes + 2 * meta_row_bytes + sw_bytes) / t_pre_us; + flip_bw = ((vm_bytes + dpte_row_bytes + meta_row_bytes) * dlg_sys_param.total_flip_bw) + / (double) dlg_sys_param.total_flip_bytes; + t_vm_us = line_time_in_us / 4.0; + if (vm_en && dcc_en) { + t_vm_us = dml_max( + dlg_sys_param.t_extra_us, + dml_max((double) vm_bytes / prefetch_bw, t_vm_us)); + + if (iflip_en && !dual_plane) { + t_vm_us = dml_max(mode_lib->soc.urgent_latency_us, t_vm_us); + if (flip_bw > 0.) + t_vm_us = dml_max(vm_bytes / flip_bw, t_vm_us); + } + } + + t_r0_us = dml_max(dlg_sys_param.t_extra_us - t_vm_us, line_time_in_us - t_vm_us); + + if (vm_en || dcc_en) { + t_r0_us = dml_max( + (double) (dpte_row_bytes + meta_row_bytes) / prefetch_bw, + dlg_sys_param.t_extra_us); + t_r0_us = dml_max((double) (line_time_in_us - t_vm_us), t_r0_us); + + if (iflip_en && !dual_plane) { + t_r0_us = dml_max(mode_lib->soc.urgent_latency_us * 2.0, t_r0_us); + if (flip_bw > 0.) + t_r0_us = dml_max( + (dpte_row_bytes + meta_row_bytes) / flip_bw, + t_r0_us); + } + } + + disp_dlg_regs->dst_y_after_scaler = dst_y_after_scaler; /* in terms of line */ + disp_dlg_regs->refcyc_x_after_scaler = dst_x_after_scaler * ref_freq_to_pix_freq; /* in terms of refclk */ + ASSERT(disp_dlg_regs->refcyc_x_after_scaler < (unsigned int) dml_pow(2, 13)); + DTRACE( + "DLG: %s: disp_dlg_regs->dst_y_after_scaler = 0x%0x", + __func__, + disp_dlg_regs->dst_y_after_scaler); + DTRACE( + "DLG: %s: disp_dlg_regs->refcyc_x_after_scaler = 0x%0x", + __func__, + disp_dlg_regs->refcyc_x_after_scaler); + + disp_dlg_regs->dst_y_prefetch = (unsigned int) (dst_y_prefetch * dml_pow(2, 2)); + DTRACE( + "DLG: %s: disp_dlg_regs->dst_y_prefetch = %d", + __func__, + disp_dlg_regs->dst_y_prefetch); + + dst_y_per_vm_vblank = 0.0; + dst_y_per_row_vblank = 0.0; + + dst_y_per_vm_vblank = t_vm_us / line_time_in_us; + dst_y_per_vm_vblank = dml_floor(4.0 * (dst_y_per_vm_vblank + 0.125), 1) / 4.0; + disp_dlg_regs->dst_y_per_vm_vblank = (unsigned int) (dst_y_per_vm_vblank * dml_pow(2, 2)); + + dst_y_per_row_vblank = t_r0_us / line_time_in_us; + dst_y_per_row_vblank = dml_floor(4.0 * (dst_y_per_row_vblank + 0.125), 1) / 4.0; + disp_dlg_regs->dst_y_per_row_vblank = (unsigned int) (dst_y_per_row_vblank * dml_pow(2, 2)); + + DTRACE("DLG: %s: lsw_l = %d", __func__, lsw_l); + DTRACE("DLG: %s: lsw_c = %d", __func__, lsw_c); + DTRACE("DLG: %s: dpte_bytes_per_row_ub_l = %d", __func__, dpte_bytes_per_row_ub_l); + DTRACE("DLG: %s: dpte_bytes_per_row_ub_c = %d", __func__, dpte_bytes_per_row_ub_c); + + DTRACE("DLG: %s: prefetch_bw = %3.2f", __func__, prefetch_bw); + DTRACE("DLG: %s: flip_bw = %3.2f", __func__, flip_bw); + DTRACE("DLG: %s: t_pre_us = %3.2f", __func__, t_pre_us); + DTRACE("DLG: %s: t_vm_us = %3.2f", __func__, t_vm_us); + DTRACE("DLG: %s: t_r0_us = %3.2f", __func__, t_r0_us); + DTRACE("DLG: %s: dst_y_per_vm_vblank = %3.2f", __func__, dst_y_per_vm_vblank); + DTRACE("DLG: %s: dst_y_per_row_vblank = %3.2f", __func__, dst_y_per_row_vblank); + DTRACE("DLG: %s: dst_y_prefetch = %3.2f", __func__, dst_y_prefetch); + + min_dst_y_per_vm_vblank = 8.0; + min_dst_y_per_row_vblank = 16.0; + if (htotal <= 75) { + min_vblank = 300; + min_dst_y_per_vm_vblank = 100.0; + min_dst_y_per_row_vblank = 100.0; + } + + ASSERT(dst_y_per_vm_vblank < min_dst_y_per_vm_vblank); + ASSERT(dst_y_per_row_vblank < min_dst_y_per_row_vblank); + + ASSERT(dst_y_prefetch > (dst_y_per_vm_vblank + dst_y_per_row_vblank)); + lsw = dst_y_prefetch - (dst_y_per_vm_vblank + dst_y_per_row_vblank); + + DTRACE("DLG: %s: lsw = %3.2f", __func__, lsw); + + vratio_pre_l = get_vratio_pre( + mode_lib, + max_num_sw_l, + max_partial_sw_l, + swath_height_l, + max_vinit_l, + lsw); + vratio_pre_c = 1.0; + if (dual_plane) + vratio_pre_c = get_vratio_pre( + mode_lib, + max_num_sw_c, + max_partial_sw_c, + swath_height_c, + max_vinit_c, + lsw); + + DTRACE("DLG: %s: vratio_pre_l=%3.2f", __func__, vratio_pre_l); + DTRACE("DLG: %s: vratio_pre_c=%3.2f", __func__, vratio_pre_c); + + ASSERT(vratio_pre_l <= 4.0); + if (vratio_pre_l >= 4.0) + disp_dlg_regs->vratio_prefetch = (unsigned int) dml_pow(2, 21) - 1; + else + disp_dlg_regs->vratio_prefetch = (unsigned int) (vratio_pre_l * dml_pow(2, 19)); + + ASSERT(vratio_pre_c <= 4.0); + if (vratio_pre_c >= 4.0) + disp_dlg_regs->vratio_prefetch_c = (unsigned int) dml_pow(2, 21) - 1; + else + disp_dlg_regs->vratio_prefetch_c = (unsigned int) (vratio_pre_c * dml_pow(2, 19)); + + 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); + ASSERT(disp_dlg_regs->refcyc_per_pte_group_vblank_l < (unsigned int) dml_pow(2, 13)); + + disp_dlg_regs->refcyc_per_pte_group_vblank_c = + (unsigned int) (dst_y_per_row_vblank * (double) htotal + * ref_freq_to_pix_freq / (double) dpte_groups_per_row_ub_c); + ASSERT(disp_dlg_regs->refcyc_per_pte_group_vblank_c < (unsigned int) dml_pow(2, 13)); + + disp_dlg_regs->refcyc_per_meta_chunk_vblank_l = + (unsigned int) (dst_y_per_row_vblank * (double) htotal + * ref_freq_to_pix_freq / (double) meta_chunks_per_row_ub_l); + ASSERT(disp_dlg_regs->refcyc_per_meta_chunk_vblank_l < (unsigned int) dml_pow(2, 13)); + + disp_dlg_regs->refcyc_per_meta_chunk_vblank_c = + disp_dlg_regs->refcyc_per_meta_chunk_vblank_l;/* dcc for 4:2:0 is not supported in dcn1.0. assigned to be the same as _l for now */ + + /* Active */ + req_per_swath_ub_l = rq_dlg_param.rq_l.req_per_swath_ub; + req_per_swath_ub_c = rq_dlg_param.rq_c.req_per_swath_ub; + meta_row_height_l = rq_dlg_param.rq_l.meta_row_height; + swath_width_pixels_ub_l = 0; + swath_width_pixels_ub_c = 0; + scaler_rec_in_width_l = 0; + scaler_rec_in_width_c = 0; + dpte_row_height_l = rq_dlg_param.rq_l.dpte_row_height; + dpte_row_height_c = rq_dlg_param.rq_c.dpte_row_height; + + disp_dlg_regs->dst_y_per_pte_row_nom_l = (unsigned int) ((double) dpte_row_height_l + / (double) vratio_l * dml_pow(2, 2)); + ASSERT(disp_dlg_regs->dst_y_per_pte_row_nom_l < (unsigned int) dml_pow(2, 17)); + + disp_dlg_regs->dst_y_per_pte_row_nom_c = (unsigned int) ((double) dpte_row_height_c + / (double) vratio_c * dml_pow(2, 2)); + ASSERT(disp_dlg_regs->dst_y_per_pte_row_nom_c < (unsigned int) dml_pow(2, 17)); + + disp_dlg_regs->dst_y_per_meta_row_nom_l = (unsigned int) ((double) meta_row_height_l + / (double) vratio_l * dml_pow(2, 2)); + ASSERT(disp_dlg_regs->dst_y_per_meta_row_nom_l < (unsigned int) dml_pow(2, 17)); + + disp_dlg_regs->dst_y_per_meta_row_nom_c = disp_dlg_regs->dst_y_per_meta_row_nom_l; /* dcc for 4:2:0 is not supported in dcn1.0. assigned to be the same as _l for now */ + + disp_dlg_regs->refcyc_per_pte_group_nom_l = (unsigned int) ((double) dpte_row_height_l + / (double) vratio_l * (double) htotal * ref_freq_to_pix_freq + / (double) dpte_groups_per_row_ub_l); + if (disp_dlg_regs->refcyc_per_pte_group_nom_l >= (unsigned int) dml_pow(2, 23)) + disp_dlg_regs->refcyc_per_pte_group_nom_l = dml_pow(2, 23) - 1; + + disp_dlg_regs->refcyc_per_pte_group_nom_c = (unsigned int) ((double) dpte_row_height_c + / (double) vratio_c * (double) htotal * ref_freq_to_pix_freq + / (double) dpte_groups_per_row_ub_c); + if (disp_dlg_regs->refcyc_per_pte_group_nom_c >= (unsigned int) dml_pow(2, 23)) + disp_dlg_regs->refcyc_per_pte_group_nom_c = dml_pow(2, 23) - 1; + + disp_dlg_regs->refcyc_per_meta_chunk_nom_l = (unsigned int) ((double) meta_row_height_l + / (double) vratio_l * (double) htotal * ref_freq_to_pix_freq + / (double) meta_chunks_per_row_ub_l); + if (disp_dlg_regs->refcyc_per_meta_chunk_nom_l >= (unsigned int) dml_pow(2, 23)) + disp_dlg_regs->refcyc_per_meta_chunk_nom_l = dml_pow(2, 23) - 1; + + if (mode_422) { + swath_width_pixels_ub_l = swath_width_ub_l * 2; /* *2 for 2 pixel per element */ + swath_width_pixels_ub_c = swath_width_ub_c * 2; + } else { + swath_width_pixels_ub_l = swath_width_ub_l * 1; + swath_width_pixels_ub_c = swath_width_ub_c * 1; + } + + hscale_pixel_rate_l = 0.; + hscale_pixel_rate_c = 0.; + min_hratio_fact_l = 1.0; + min_hratio_fact_c = 1.0; + + if (htaps_l <= 1) + min_hratio_fact_l = 2.0; + else if (htaps_l <= 6) { + if ((hratios_l * 2.0) > 4.0) + min_hratio_fact_l = 4.0; + else + min_hratio_fact_l = hratios_l * 2.0; + } else { + if (hratios_l > 4.0) + min_hratio_fact_l = 4.0; + else + min_hratio_fact_l = hratios_l; + } + + hscale_pixel_rate_l = min_hratio_fact_l * dppclk_freq_in_mhz; + + if (htaps_c <= 1) + min_hratio_fact_c = 2.0; + else if (htaps_c <= 6) { + if ((hratios_c * 2.0) > 4.0) + min_hratio_fact_c = 4.0; + else + min_hratio_fact_c = hratios_c * 2.0; + } else { + if (hratios_c > 4.0) + min_hratio_fact_c = 4.0; + else + min_hratio_fact_c = hratios_c; + } + + hscale_pixel_rate_c = min_hratio_fact_c * dppclk_freq_in_mhz; + + refcyc_per_line_delivery_pre_l = 0.; + refcyc_per_line_delivery_pre_c = 0.; + refcyc_per_line_delivery_l = 0.; + refcyc_per_line_delivery_c = 0.; + + refcyc_per_req_delivery_pre_l = 0.; + refcyc_per_req_delivery_pre_c = 0.; + refcyc_per_req_delivery_l = 0.; + refcyc_per_req_delivery_c = 0.; + refcyc_per_req_delivery_pre_cur0 = 0.; + refcyc_per_req_delivery_cur0 = 0.; + + full_recout_width = 0; + if (e2e_pipe_param.pipe.src.is_hsplit) { + if (e2e_pipe_param.pipe.dest.full_recout_width == 0) { + DTRACE("DLG: %s: Warningfull_recout_width not set in hsplit mode", __func__); + full_recout_width = e2e_pipe_param.pipe.dest.recout_width * 2; /* assume half split for dcn1 */ + } else + full_recout_width = e2e_pipe_param.pipe.dest.full_recout_width; + } else + full_recout_width = e2e_pipe_param.pipe.dest.recout_width; + + refcyc_per_line_delivery_pre_l = get_refcyc_per_delivery( + mode_lib, + refclk_freq_in_mhz, + pclk_freq_in_mhz, + full_recout_width, + vratio_pre_l, + hscale_pixel_rate_l, + swath_width_pixels_ub_l, + 1); /* per line */ + + refcyc_per_line_delivery_l = get_refcyc_per_delivery( + mode_lib, + refclk_freq_in_mhz, + pclk_freq_in_mhz, + full_recout_width, + vratio_l, + hscale_pixel_rate_l, + swath_width_pixels_ub_l, + 1); /* per line */ + + DTRACE("DLG: %s: full_recout_width = %d", __func__, full_recout_width); + DTRACE("DLG: %s: hscale_pixel_rate_l = %3.2f", __func__, hscale_pixel_rate_l); + DTRACE( + "DLG: %s: refcyc_per_line_delivery_pre_l = %3.2f", + __func__, + refcyc_per_line_delivery_pre_l); + DTRACE( + "DLG: %s: refcyc_per_line_delivery_l = %3.2f", + __func__, + refcyc_per_line_delivery_l); + + disp_dlg_regs->refcyc_per_line_delivery_pre_l = (unsigned int) dml_floor( + refcyc_per_line_delivery_pre_l, + 1); + disp_dlg_regs->refcyc_per_line_delivery_l = (unsigned int) dml_floor( + refcyc_per_line_delivery_l, + 1); + ASSERT(disp_dlg_regs->refcyc_per_line_delivery_pre_l < (unsigned int) dml_pow(2, 13)); + ASSERT(disp_dlg_regs->refcyc_per_line_delivery_l < (unsigned int) dml_pow(2, 13)); + + if (dual_plane) { + refcyc_per_line_delivery_pre_c = get_refcyc_per_delivery( + mode_lib, + refclk_freq_in_mhz, + pclk_freq_in_mhz, + full_recout_width, + vratio_pre_c, + hscale_pixel_rate_c, + swath_width_pixels_ub_c, + 1); /* per line */ + + refcyc_per_line_delivery_c = get_refcyc_per_delivery( + mode_lib, + refclk_freq_in_mhz, + pclk_freq_in_mhz, + full_recout_width, + vratio_c, + hscale_pixel_rate_c, + swath_width_pixels_ub_c, + 1); /* per line */ + + DTRACE( + "DLG: %s: refcyc_per_line_delivery_pre_c = %3.2f", + __func__, + refcyc_per_line_delivery_pre_c); + DTRACE( + "DLG: %s: refcyc_per_line_delivery_c = %3.2f", + __func__, + refcyc_per_line_delivery_c); + + disp_dlg_regs->refcyc_per_line_delivery_pre_c = (unsigned int) dml_floor( + refcyc_per_line_delivery_pre_c, + 1); + disp_dlg_regs->refcyc_per_line_delivery_c = (unsigned int) dml_floor( + refcyc_per_line_delivery_c, + 1); + ASSERT(disp_dlg_regs->refcyc_per_line_delivery_pre_c < (unsigned int) dml_pow(2, 13)); + ASSERT(disp_dlg_regs->refcyc_per_line_delivery_c < (unsigned int) dml_pow(2, 13)); + } + disp_dlg_regs->chunk_hdl_adjust_cur0 = 3; + + /* TTU - Luma / Chroma */ + if (access_dir) { /* vertical access */ + scaler_rec_in_width_l = vp_height_l; + scaler_rec_in_width_c = vp_height_c; + } else { + scaler_rec_in_width_l = vp_width_l; + scaler_rec_in_width_c = vp_width_c; + } + + refcyc_per_req_delivery_pre_l = get_refcyc_per_delivery( + mode_lib, + refclk_freq_in_mhz, + pclk_freq_in_mhz, + full_recout_width, + vratio_pre_l, + hscale_pixel_rate_l, + scaler_rec_in_width_l, + req_per_swath_ub_l); /* per req */ + refcyc_per_req_delivery_l = get_refcyc_per_delivery( + mode_lib, + refclk_freq_in_mhz, + pclk_freq_in_mhz, + full_recout_width, + vratio_l, + hscale_pixel_rate_l, + scaler_rec_in_width_l, + req_per_swath_ub_l); /* per req */ + + DTRACE( + "DLG: %s: refcyc_per_req_delivery_pre_l = %3.2f", + __func__, + refcyc_per_req_delivery_pre_l); + DTRACE( + "DLG: %s: refcyc_per_req_delivery_l = %3.2f", + __func__, + refcyc_per_req_delivery_l); + + disp_ttu_regs->refcyc_per_req_delivery_pre_l = (unsigned int) (refcyc_per_req_delivery_pre_l + * dml_pow(2, 10)); + disp_ttu_regs->refcyc_per_req_delivery_l = (unsigned int) (refcyc_per_req_delivery_l + * dml_pow(2, 10)); + + ASSERT(refcyc_per_req_delivery_pre_l < dml_pow(2, 13)); + ASSERT(refcyc_per_req_delivery_l < dml_pow(2, 13)); + + if (dual_plane) { + refcyc_per_req_delivery_pre_c = get_refcyc_per_delivery( + mode_lib, + refclk_freq_in_mhz, + pclk_freq_in_mhz, + full_recout_width, + vratio_pre_c, + hscale_pixel_rate_c, + scaler_rec_in_width_c, + req_per_swath_ub_c); /* per req */ + refcyc_per_req_delivery_c = get_refcyc_per_delivery( + mode_lib, + refclk_freq_in_mhz, + pclk_freq_in_mhz, + full_recout_width, + vratio_c, + hscale_pixel_rate_c, + scaler_rec_in_width_c, + req_per_swath_ub_c); /* per req */ + + DTRACE( + "DLG: %s: refcyc_per_req_delivery_pre_c = %3.2f", + __func__, + refcyc_per_req_delivery_pre_c); + DTRACE( + "DLG: %s: refcyc_per_req_delivery_c = %3.2f", + __func__, + refcyc_per_req_delivery_c); + + disp_ttu_regs->refcyc_per_req_delivery_pre_c = + (unsigned int) (refcyc_per_req_delivery_pre_c * dml_pow(2, 10)); + disp_ttu_regs->refcyc_per_req_delivery_c = (unsigned int) (refcyc_per_req_delivery_c + * dml_pow(2, 10)); + + ASSERT(refcyc_per_req_delivery_pre_c < dml_pow(2, 13)); + ASSERT(refcyc_per_req_delivery_c < dml_pow(2, 13)); + } + + /* TTU - Cursor */ + hratios_cur0 = e2e_pipe_param.pipe.scale_ratio_depth.hscl_ratio; + cur0_src_width = e2e_pipe_param.pipe.src.cur0_src_width; /* cursor source width */ + cur0_bpp = (enum cursor_bpp) e2e_pipe_param.pipe.src.cur0_bpp; + cur0_req_size = 0; + cur0_req_width = 0; + cur0_width_ub = 0.0; + cur0_req_per_width = 0.0; + hactive_cur0 = 0.0; + + ASSERT(cur0_src_width <= 256); + + if (cur0_src_width > 0) { + unsigned int cur0_bit_per_pixel = 0; + + if (cur0_bpp == dm_cur_2bit) { + cur0_req_size = 64; /* byte */ + cur0_bit_per_pixel = 2; + } else { /* 32bit */ + cur0_bit_per_pixel = 32; + if (cur0_src_width >= 1 && cur0_src_width <= 16) + cur0_req_size = 64; + else if (cur0_src_width >= 17 && cur0_src_width <= 31) + cur0_req_size = 128; + else + cur0_req_size = 256; + } + + cur0_req_width = (double) cur0_req_size / ((double) cur0_bit_per_pixel / 8.0); + cur0_width_ub = dml_ceil((double) cur0_src_width / (double) cur0_req_width, 1) + * (double) cur0_req_width; + cur0_req_per_width = cur0_width_ub / (double) cur0_req_width; + hactive_cur0 = (double) cur0_src_width / hratios_cur0; /* FIXME: oswin to think about what to do for cursor */ + + if (vratio_pre_l <= 1.0) { + refcyc_per_req_delivery_pre_cur0 = hactive_cur0 * ref_freq_to_pix_freq + / (double) cur0_req_per_width; + } else { + refcyc_per_req_delivery_pre_cur0 = (double) refclk_freq_in_mhz + * (double) cur0_src_width / hscale_pixel_rate_l + / (double) cur0_req_per_width; + } + + disp_ttu_regs->refcyc_per_req_delivery_pre_cur0 = + (unsigned int) (refcyc_per_req_delivery_pre_cur0 * dml_pow(2, 10)); + ASSERT(refcyc_per_req_delivery_pre_cur0 < dml_pow(2, 13)); + + if (vratio_l <= 1.0) { + refcyc_per_req_delivery_cur0 = hactive_cur0 * ref_freq_to_pix_freq + / (double) cur0_req_per_width; + } else { + refcyc_per_req_delivery_cur0 = (double) refclk_freq_in_mhz + * (double) cur0_src_width / hscale_pixel_rate_l + / (double) cur0_req_per_width; + } + + DTRACE("DLG: %s: cur0_req_width = %d", __func__, cur0_req_width); + DTRACE( + "DLG: %s: cur0_width_ub = %3.2f", + __func__, + cur0_width_ub); + DTRACE( + "DLG: %s: cur0_req_per_width = %3.2f", + __func__, + cur0_req_per_width); + DTRACE( + "DLG: %s: hactive_cur0 = %3.2f", + __func__, + hactive_cur0); + DTRACE( + "DLG: %s: refcyc_per_req_delivery_pre_cur0 = %3.2f", + __func__, + refcyc_per_req_delivery_pre_cur0); + DTRACE( + "DLG: %s: refcyc_per_req_delivery_cur0 = %3.2f", + __func__, + refcyc_per_req_delivery_cur0); + + disp_ttu_regs->refcyc_per_req_delivery_cur0 = + (unsigned int) (refcyc_per_req_delivery_cur0 * dml_pow(2, 10)); + ASSERT(refcyc_per_req_delivery_cur0 < dml_pow(2, 13)); + } else { + disp_ttu_regs->refcyc_per_req_delivery_pre_cur0 = 0; + disp_ttu_regs->refcyc_per_req_delivery_cur0 = 0; + } + + /* TTU - Misc */ + disp_ttu_regs->qos_level_low_wm = 0; + ASSERT(disp_ttu_regs->qos_level_low_wm < dml_pow(2, 14)); + disp_ttu_regs->qos_level_high_wm = (unsigned int) (4.0 * (double) htotal + * ref_freq_to_pix_freq); + ASSERT(disp_ttu_regs->qos_level_high_wm < dml_pow(2, 14)); + + disp_ttu_regs->qos_level_flip = 14; + disp_ttu_regs->qos_level_fixed_l = 8; + disp_ttu_regs->qos_level_fixed_c = 8; + disp_ttu_regs->qos_level_fixed_cur0 = 8; + disp_ttu_regs->qos_ramp_disable_l = 0; + disp_ttu_regs->qos_ramp_disable_c = 0; + disp_ttu_regs->qos_ramp_disable_cur0 = 0; + + disp_ttu_regs->min_ttu_vblank = min_ttu_vblank * refclk_freq_in_mhz; + ASSERT(disp_ttu_regs->min_ttu_vblank < dml_pow(2, 24)); + + print__ttu_regs_st(mode_lib, *disp_ttu_regs); + print__dlg_regs_st(mode_lib, *disp_dlg_regs); +} diff --git a/drivers/gpu/drm/amd/display/dc/dml/dml1_display_rq_dlg_calc.h b/drivers/gpu/drm/amd/display/dc/dml/dml1_display_rq_dlg_calc.h new file mode 100644 index 000000000000..987d7671cd0f --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/dml/dml1_display_rq_dlg_calc.h @@ -0,0 +1,67 @@ +/* + * Copyright 2017 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 __DISPLAY_RQ_DLG_CALC_H__ +#define __DISPLAY_RQ_DLG_CALC_H__ + +#include "dml_common_defs.h" +#include "display_rq_dlg_helpers.h" + +struct display_mode_lib; + +void dml1_extract_rq_regs( + struct display_mode_lib *mode_lib, + struct _vcs_dpi_display_rq_regs_st *rq_regs, + const struct _vcs_dpi_display_rq_params_st rq_param); +/* Function: dml_rq_dlg_get_rq_params + * Calculate requestor related parameters that register definition agnostic + * (i.e. this layer does try to separate real values from register definition) + * Input: + * pipe_src_param - pipe source configuration (e.g. vp, pitch, etc.) + * Output: + * rq_param - values that can be used to setup RQ (e.g. swath_height, plane1_addr, etc.) + */ +void dml1_rq_dlg_get_rq_params( + struct display_mode_lib *mode_lib, + struct _vcs_dpi_display_rq_params_st *rq_param, + const struct _vcs_dpi_display_pipe_source_params_st pipe_src_param); + + +/* Function: dml_rq_dlg_get_dlg_params + * Calculate deadline related parameters + */ +void dml1_rq_dlg_get_dlg_params( + struct display_mode_lib *mode_lib, + struct _vcs_dpi_display_dlg_regs_st *dlg_regs, + struct _vcs_dpi_display_ttu_regs_st *ttu_regs, + const struct _vcs_dpi_display_rq_dlg_params_st rq_dlg_param, + const struct _vcs_dpi_display_dlg_sys_params_st dlg_sys_param, + const struct _vcs_dpi_display_e2e_pipe_params_st e2e_pipe_param, + const bool cstate_en, + const bool pstate_en, + const bool vm_en, + const bool iflip_en); + +#endif diff --git a/drivers/gpu/drm/amd/display/dc/dml/dml_common_defs.c b/drivers/gpu/drm/amd/display/dc/dml/dml_common_defs.c new file mode 100644 index 000000000000..b953b02a1512 --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/dml/dml_common_defs.c @@ -0,0 +1,43 @@ +/* + * Copyright 2017 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 "dml_common_defs.h" +#include "../calcs/dcn_calc_math.h" + +#include "dml_inline_defs.h" + +double dml_round(double a) +{ + double round_pt = 0.5; + double ceil = dml_ceil(a, 1); + double floor = dml_floor(a, 1); + + if (a - floor >= round_pt) + return ceil; + else + return floor; +} + + diff --git a/drivers/gpu/drm/amd/display/dc/dml/dml_common_defs.h b/drivers/gpu/drm/amd/display/dc/dml/dml_common_defs.h new file mode 100644 index 000000000000..b2847bc469fe --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/dml/dml_common_defs.h @@ -0,0 +1,39 @@ +/* + * Copyright 2017 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 __DC_COMMON_DEFS_H__ +#define __DC_COMMON_DEFS_H__ + +#include "dm_services.h" +#include "dc_features.h" +#include "display_mode_structs.h" +#include "display_mode_enums.h" + +#define dml_print(str, ...) {dm_logger_write(mode_lib->logger, LOG_DML, str, ##__VA_ARGS__); } +#define DTRACE(str, ...) {dm_logger_write(mode_lib->logger, LOG_DML, str, ##__VA_ARGS__); } + +double dml_round(double a); + +#endif /* __DC_COMMON_DEFS_H__ */ 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 new file mode 100644 index 000000000000..e68086b8a22f --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/dml/dml_inline_defs.h @@ -0,0 +1,121 @@ +/* + * Copyright 2017 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_INLINE_DEFS_H__ +#define __DML_INLINE_DEFS_H__ + +#include "dml_common_defs.h" +#include "../calcs/dcn_calc_math.h" + +static inline double dml_min(double a, double b) +{ + return (double) dcn_bw_min2(a, b); +} + +static inline double dml_max(double a, double b) +{ + return (double) dcn_bw_max2(a, b); +} + +static inline double dml_max3(double a, double b, double c) +{ + return dml_max(dml_max(a, b), c); +} + +static inline double dml_max4(double a, double b, double c, double d) +{ + return dml_max(dml_max(a, b), dml_max(c, d)); +} + +static inline double dml_max5(double a, double b, double c, double d, double e) +{ + return dml_max(dml_max4(a, b, c, d), e); +} + +static inline double dml_ceil(double a, double granularity) +{ + return (double) dcn_bw_ceil2(a, granularity); +} + +static inline double dml_floor(double a, double granularity) +{ + return (double) dcn_bw_floor2(a, granularity); +} + +static inline int dml_log2(double x) +{ + return dml_round((double)dcn_bw_log(x, 2)); +} + +static inline double dml_pow(double a, int exp) +{ + return (double) dcn_bw_pow(a, exp); +} + +static inline double dml_fmod(double f, int val) +{ + return (double) dcn_bw_mod(f, val); +} + +static inline double dml_ceil_2(double f) +{ + return (double) dcn_bw_ceil2(f, 2); +} + +static inline double dml_ceil_ex(double x, double granularity) +{ + return (double) dcn_bw_ceil2(x, granularity); +} + +static inline double dml_floor_ex(double x, double granularity) +{ + return (double) dcn_bw_floor2(x, granularity); +} + +static inline double dml_log(double x, double base) +{ + return (double) dcn_bw_log(x, base); +} + +static inline unsigned int dml_round_to_multiple(unsigned int num, + unsigned int multiple, + bool up) +{ + unsigned int remainder; + + if (multiple == 0) + return num; + + remainder = num % multiple; + + if (remainder == 0) + return num; + + if (up) + return (num + multiple - remainder); + else + return (num - remainder); +} +#endif diff --git a/drivers/gpu/drm/amd/display/dc/dml/soc_bounding_box.c b/drivers/gpu/drm/amd/display/dc/dml/soc_bounding_box.c new file mode 100644 index 000000000000..bc7d8c707221 --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/dml/soc_bounding_box.c @@ -0,0 +1,69 @@ +/* + * Copyright 2017 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 "soc_bounding_box.h" +#include "display_mode_lib.h" +#include "dc_features.h" + +#include "dml_inline_defs.h" +void dml_socbb_set_latencies(soc_bounding_box_st *to_box, soc_bounding_box_st *from_box) +{ + to_box->dram_clock_change_latency_us = from_box->dram_clock_change_latency_us; + to_box->sr_exit_time_us = from_box->sr_exit_time_us; + to_box->sr_enter_plus_exit_time_us = from_box->sr_enter_plus_exit_time_us; + to_box->urgent_latency_us = from_box->urgent_latency_us; + to_box->writeback_latency_us = from_box->writeback_latency_us; +} + +voltage_scaling_st dml_socbb_voltage_scaling( + const soc_bounding_box_st *soc, + enum voltage_state voltage) +{ + const voltage_scaling_st *voltage_state; + const voltage_scaling_st * const voltage_end = soc->clock_limits + DC__VOLTAGE_STATES; + + for (voltage_state = soc->clock_limits; + voltage_state < voltage_end && voltage_state->state != voltage; + voltage_state++) { + } + + if (voltage_state < voltage_end) + return *voltage_state; + return soc->clock_limits[DC__VOLTAGE_STATES - 1]; +} + +double dml_socbb_return_bw_mhz(soc_bounding_box_st *box, enum voltage_state voltage) +{ + double return_bw; + + voltage_scaling_st state = dml_socbb_voltage_scaling(box, voltage); + + return_bw = dml_min((double) box->return_bus_width_bytes * state.dcfclk_mhz, + state.dram_bw_per_chan_gbps * 1000.0 * (double) box->num_chans + * box->ideal_dram_bw_after_urgent_percent / 100.0); + + return_bw = dml_min((double) box->return_bus_width_bytes * state.fabricclk_mhz, return_bw); + + return return_bw; +} diff --git a/drivers/gpu/drm/amd/display/dc/dml/soc_bounding_box.h b/drivers/gpu/drm/amd/display/dc/dml/soc_bounding_box.h new file mode 100644 index 000000000000..7a65206a6d21 --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/dml/soc_bounding_box.h @@ -0,0 +1,35 @@ +/* + * Copyright 2017 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 __SOC_BOUNDING_BOX_H__ +#define __SOC_BOUNDING_BOX_H__ + +#include "dml_common_defs.h" + +void dml_socbb_set_latencies(soc_bounding_box_st *to_box, soc_bounding_box_st *from_box); +voltage_scaling_st dml_socbb_voltage_scaling(const soc_bounding_box_st *box, enum voltage_state voltage); +double dml_socbb_return_bw_mhz(soc_bounding_box_st *box, enum voltage_state voltage); + +#endif |