diff options
Diffstat (limited to 'common/recipes-kernel/linux/linux-yocto-4.14.71/0387-drm-amdgpu-display-Add-calcs-code-for-DCN.patch')
-rw-r--r-- | common/recipes-kernel/linux/linux-yocto-4.14.71/0387-drm-amdgpu-display-Add-calcs-code-for-DCN.patch | 5868 |
1 files changed, 5868 insertions, 0 deletions
diff --git a/common/recipes-kernel/linux/linux-yocto-4.14.71/0387-drm-amdgpu-display-Add-calcs-code-for-DCN.patch b/common/recipes-kernel/linux/linux-yocto-4.14.71/0387-drm-amdgpu-display-Add-calcs-code-for-DCN.patch new file mode 100644 index 00000000..9c801151 --- /dev/null +++ b/common/recipes-kernel/linux/linux-yocto-4.14.71/0387-drm-amdgpu-display-Add-calcs-code-for-DCN.patch @@ -0,0 +1,5868 @@ +From 57c3468e7d35f6ca133558dd343fed4bfcce8eb1 Mon Sep 17 00:00:00 2001 +From: Harry Wentland <harry.wentland@amd.com> +Date: Mon, 8 May 2017 15:17:39 -0400 +Subject: [PATCH 0387/4131] drm/amdgpu/display: Add calcs code for DCN + +Bandwidth and scaling calculations for DCN. + +Signed-off-by: Harry Wentland <harry.wentland@amd.com> +Signed-off-by: Alex Deucher <alexander.deucher@amd.com> +--- + .../gpu/drm/amd/display/dc/calcs/dcn_calc_auto.c | 3629 ++++++++++++++++++++ + .../gpu/drm/amd/display/dc/calcs/dcn_calc_auto.h | 37 + + .../gpu/drm/amd/display/dc/calcs/dcn_calc_math.c | 104 + + .../gpu/drm/amd/display/dc/calcs/dcn_calc_math.h | 40 + + drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c | 1366 ++++++++ + drivers/gpu/drm/amd/display/dc/inc/dcn_calcs.h | 629 ++++ + 6 files changed, 5805 insertions(+) + create mode 100644 drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_auto.c + create mode 100644 drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_auto.h + create mode 100644 drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_math.c + create mode 100644 drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_math.h + create mode 100644 drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c + create mode 100644 drivers/gpu/drm/amd/display/dc/inc/dcn_calcs.h + +diff --git a/drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_auto.c b/drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_auto.c +new file mode 100644 +index 0000000..21b83e4 +--- /dev/null ++++ b/drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_auto.c +@@ -0,0 +1,3629 @@ ++/* ++ * 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 "dm_services.h" ++#include "dcn_calc_auto.h" ++#include "dcn_calc_math.h" ++ ++/*REVISION#247*/ ++void scaler_settings_calculation(struct dcn_bw_internal_vars *v) ++{ ++ int k; ++ /*scaler settings calculation*/ ++ ++ /*scale ratio calculation*/ ++ ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->allow_different_hratio_vratio == dcn_bw_yes) { ++ if (v->source_scan[k] == dcn_bw_hor) { ++ v->h_ratio[k] = v->viewport_width[k] / v->scaler_rec_out_width[k]; ++ v->v_ratio[k] = v->viewport_height[k] / v->scaler_recout_height[k]; ++ } else { ++ v->h_ratio[k] = v->viewport_height[k] / v->scaler_rec_out_width[k]; ++ v->v_ratio[k] = v->viewport_width[k] / v->scaler_recout_height[k]; ++ } ++ } else { ++ if (v->source_scan[k] == dcn_bw_hor) { ++ v->h_ratio[k] = dcn_bw_max2( ++ v->viewport_width[k] / v->scaler_rec_out_width[k], ++ v->viewport_height[k] / v->scaler_recout_height[k]); ++ } else { ++ v->h_ratio[k] = dcn_bw_max2( ++ v->viewport_height[k] / v->scaler_rec_out_width[k], ++ v->viewport_width[k] / v->scaler_recout_height[k]); ++ } ++ v->v_ratio[k] = v->h_ratio[k]; ++ } ++ if (v->interlace_output[k] == 1.0) { ++ v->v_ratio[k] = 2.0 * v->v_ratio[k]; ++ } ++ if ((v->underscan_output[k] == 1.0)) { ++ v->h_ratio[k] = v->h_ratio[k] * v->under_scan_factor; ++ v->v_ratio[k] = v->v_ratio[k] * v->under_scan_factor; ++ } ++ } ++ /*scaler taps calculation*/ ++ ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->h_ratio[k] > 1.0) { ++ v->acceptable_quality_hta_ps = dcn_bw_min2( ++ v->max_hscl_taps, ++ 2.0 * dcn_bw_ceil2(v->h_ratio[k], 1.0)); ++ } else if (v->h_ratio[k] < 1.0) { ++ v->acceptable_quality_hta_ps = 4.0; ++ } else { ++ v->acceptable_quality_hta_ps = 1.0; ++ } ++ if (v->ta_pscalculation == dcn_bw_override) { ++ v->htaps[k] = v->override_hta_ps[k]; ++ } else { ++ v->htaps[k] = v->acceptable_quality_hta_ps; ++ } ++ if (v->v_ratio[k] > 1.0) { ++ v->acceptable_quality_vta_ps = dcn_bw_min2( ++ v->max_vscl_taps, ++ 2.0 * dcn_bw_ceil2(v->v_ratio[k], 1.0)); ++ } else if (v->v_ratio[k] < 1.0) { ++ v->acceptable_quality_vta_ps = 4.0; ++ } else { ++ v->acceptable_quality_vta_ps = 1.0; ++ } ++ if (v->ta_pscalculation == dcn_bw_override) { ++ v->vtaps[k] = v->override_vta_ps[k]; ++ } else { ++ v->vtaps[k] = v->acceptable_quality_vta_ps; ++ } ++ if (v->source_pixel_format[k] == dcn_bw_rgb_sub_64 ++ || v->source_pixel_format[k] == dcn_bw_rgb_sub_32 ++ || v->source_pixel_format[k] == dcn_bw_rgb_sub_16) { ++ v->vta_pschroma[k] = 0.0; ++ v->hta_pschroma[k] = 0.0; ++ } else { ++ if (v->ta_pscalculation == dcn_bw_override) { ++ v->vta_pschroma[k] = v->override_vta_pschroma[k]; ++ v->hta_pschroma[k] = v->override_hta_pschroma[k]; ++ } else { ++ v->vta_pschroma[k] = v->acceptable_quality_vta_ps; ++ v->hta_pschroma[k] = v->acceptable_quality_hta_ps; ++ } ++ } ++ } ++} ++void mode_support_and_system_configuration(struct dcn_bw_internal_vars *v) ++{ ++ int i, j, k; ++ /*mode support, voltage state and soc configuration*/ ++ ++ /*scale ratio support check*/ ++ ++ v->scale_ratio_support = dcn_bw_yes; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->h_ratio[k] > v->max_hscl_ratio || v->v_ratio[k] > v->max_vscl_ratio ++ || v->h_ratio[k] > v->htaps[k] || v->v_ratio[k] > v->vtaps[k] ++ || (v->source_pixel_format[k] != dcn_bw_rgb_sub_64 ++ && v->source_pixel_format[k] != dcn_bw_rgb_sub_32 ++ && v->source_pixel_format[k] != dcn_bw_rgb_sub_16 ++ && (v->h_ratio[k] / 2.0 > v->hta_pschroma[k] ++ || v->v_ratio[k] / 2.0 ++ > v->vta_pschroma[k]))) { ++ v->scale_ratio_support = dcn_bw_no; ++ } ++ } ++ /*source format, pixel format and scan support check*/ ++ ++ v->source_format_pixel_and_scan_support = dcn_bw_yes; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if ((v->source_surface_mode[k] == dcn_bw_sw_linear ++ && v->source_scan[k] != dcn_bw_hor) ++ || ((v->source_surface_mode[k] == dcn_bw_sw_4_kb_d ++ || v->source_surface_mode[k] == dcn_bw_sw_4_kb_d_x ++ || v->source_surface_mode[k] == dcn_bw_sw_64_kb_d ++ || v->source_surface_mode[k] == dcn_bw_sw_64_kb_d_t ++ || v->source_surface_mode[k] == dcn_bw_sw_64_kb_d_x ++ || v->source_surface_mode[k] == dcn_bw_sw_var_d ++ || v->source_surface_mode[k] == dcn_bw_sw_var_d_x) ++ && v->source_pixel_format[k] != dcn_bw_rgb_sub_64)) { ++ v->source_format_pixel_and_scan_support = dcn_bw_no; ++ } ++ } ++ /*bandwidth support check*/ ++ ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->source_scan[k] == dcn_bw_hor) { ++ v->swath_width_ysingle_dpp[k] = v->viewport_width[k]; ++ } else { ++ v->swath_width_ysingle_dpp[k] = v->viewport_height[k]; ++ } ++ if (v->source_pixel_format[k] == dcn_bw_rgb_sub_64) { ++ v->byte_per_pixel_in_dety[k] = 8.0; ++ v->byte_per_pixel_in_detc[k] = 0.0; ++ } else if (v->source_pixel_format[k] == dcn_bw_rgb_sub_32) { ++ v->byte_per_pixel_in_dety[k] = 4.0; ++ v->byte_per_pixel_in_detc[k] = 0.0; ++ } else if (v->source_pixel_format[k] == dcn_bw_rgb_sub_16) { ++ v->byte_per_pixel_in_dety[k] = 2.0; ++ v->byte_per_pixel_in_detc[k] = 0.0; ++ } else if (v->source_pixel_format[k] == dcn_bw_yuv420_sub_8) { ++ v->byte_per_pixel_in_dety[k] = 1.0; ++ v->byte_per_pixel_in_detc[k] = 2.0; ++ } else { ++ v->byte_per_pixel_in_dety[k] = 4.0f / 3; ++ v->byte_per_pixel_in_detc[k] = 8.0f / 3; ++ } ++ } ++ v->total_read_bandwidth_consumed_gbyte_per_second = 0.0; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ v->read_bandwidth[k] = v->swath_width_ysingle_dpp[k] ++ * (dcn_bw_ceil2(v->byte_per_pixel_in_dety[k], 1.0) * v->v_ratio[k] ++ + dcn_bw_ceil2(v->byte_per_pixel_in_detc[k], 2.0) / 2.0 ++ * v->v_ratio[k] / 2) ++ / (v->htotal[k] / v->pixel_clock[k]); ++ if (v->dcc_enable[k] == dcn_bw_yes) { ++ v->read_bandwidth[k] = v->read_bandwidth[k] * (1 + 1 / 256); ++ } ++ if (v->pte_enable == dcn_bw_yes && v->source_scan[k] != dcn_bw_hor ++ && (v->source_surface_mode[k] == dcn_bw_sw_4_kb_s ++ || v->source_surface_mode[k] == dcn_bw_sw_4_kb_s_x ++ || v->source_surface_mode[k] == dcn_bw_sw_4_kb_d ++ || v->source_surface_mode[k] == dcn_bw_sw_4_kb_d_x)) { ++ v->read_bandwidth[k] = v->read_bandwidth[k] * (1 + 1 / 64); ++ } else if (v->pte_enable == dcn_bw_yes && v->source_scan[k] == dcn_bw_hor ++ && (v->source_pixel_format[k] == dcn_bw_rgb_sub_64 ++ || v->source_pixel_format[k] == dcn_bw_rgb_sub_32) ++ && (v->source_surface_mode[k] == dcn_bw_sw_64_kb_s ++ || v->source_surface_mode[k] == dcn_bw_sw_64_kb_s_t ++ || v->source_surface_mode[k] == dcn_bw_sw_64_kb_s_x ++ || v->source_surface_mode[k] == dcn_bw_sw_64_kb_d ++ || v->source_surface_mode[k] == dcn_bw_sw_64_kb_d_t ++ || v->source_surface_mode[k] == dcn_bw_sw_64_kb_d_x)) { ++ v->read_bandwidth[k] = v->read_bandwidth[k] * (1 + 1 / 256); ++ } else if (v->pte_enable == dcn_bw_yes) { ++ v->read_bandwidth[k] = v->read_bandwidth[k] * (1 + 1 / 512); ++ } ++ v->total_read_bandwidth_consumed_gbyte_per_second = ++ v->total_read_bandwidth_consumed_gbyte_per_second ++ + v->read_bandwidth[k] / 1000.0; ++ } ++ v->total_write_bandwidth_consumed_gbyte_per_second = 0.0; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->output[k] == dcn_bw_writeback && v->output_format[k] == dcn_bw_444) { ++ v->write_bandwidth[k] = v->scaler_rec_out_width[k] ++ / (v->htotal[k] / v->pixel_clock[k]) * 4.0; ++ } else if (v->output[k] == dcn_bw_writeback) { ++ v->write_bandwidth[k] = v->scaler_rec_out_width[k] ++ / (v->htotal[k] / v->pixel_clock[k]) * 1.5; ++ } else { ++ v->write_bandwidth[k] = 0.0; ++ } ++ v->total_write_bandwidth_consumed_gbyte_per_second = ++ v->total_write_bandwidth_consumed_gbyte_per_second ++ + v->write_bandwidth[k] / 1000.0; ++ } ++ v->total_bandwidth_consumed_gbyte_per_second = ++ v->total_read_bandwidth_consumed_gbyte_per_second ++ + v->total_write_bandwidth_consumed_gbyte_per_second; ++ v->dcc_enabled_in_any_plane = dcn_bw_no; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->dcc_enable[k] == dcn_bw_yes) { ++ v->dcc_enabled_in_any_plane = dcn_bw_yes; ++ } ++ } ++ for (i = 0; i <= number_of_states_plus_one; i++) { ++ v->return_bw_todcn_per_state = ++ dcn_bw_min2( ++ v->return_bus_width * v->dcfclk_per_state[i], ++ v->fabric_and_dram_bandwidth_per_state[i] * 1000.0 ++ * v->percent_of_ideal_drambw_received_after_urg_latency ++ / 100.0); ++ v->return_bw_per_state[i] = v->return_bw_todcn_per_state; ++ if (v->dcc_enabled_in_any_plane == dcn_bw_yes ++ && v->return_bw_todcn_per_state ++ > v->dcfclk_per_state[i] * v->return_bus_width ++ / 4.0) { ++ v->return_bw_per_state[i] = ++ dcn_bw_min2( ++ v->return_bw_per_state[i], ++ v->return_bw_todcn_per_state * 4.0 ++ * (1.0 ++ - v->urgent_latency ++ / ((v->rob_buffer_size_in_kbyte ++ - v->pixel_chunk_size_in_kbyte) ++ * 1024.0 ++ / (v->return_bw_todcn_per_state ++ - v->dcfclk_per_state[i] ++ * v->return_bus_width ++ / 4.0) ++ + v->urgent_latency))); ++ } ++ v->critical_point = 2.0 * v->return_bus_width * v->dcfclk_per_state[i] ++ * v->urgent_latency ++ / (v->return_bw_todcn_per_state * v->urgent_latency ++ + (v->rob_buffer_size_in_kbyte ++ - v->pixel_chunk_size_in_kbyte) ++ * 1024.0); ++ if (v->dcc_enabled_in_any_plane == dcn_bw_yes && v->critical_point > 1.0 ++ && v->critical_point < 4.0) { ++ v->return_bw_per_state[i] = ++ dcn_bw_min2( ++ v->return_bw_per_state[i], ++ dcn_bw_pow( ++ 4.0 ++ * v->return_bw_todcn_per_state ++ * (v->rob_buffer_size_in_kbyte ++ - v->pixel_chunk_size_in_kbyte) ++ * 1024.0 ++ * v->return_bus_width ++ * v->dcfclk_per_state[i] ++ * v->urgent_latency ++ / (v->return_bw_todcn_per_state ++ * v->urgent_latency ++ + (v->rob_buffer_size_in_kbyte ++ - v->pixel_chunk_size_in_kbyte) ++ * 1024.0), ++ 2)); ++ } ++ v->return_bw_todcn_per_state = dcn_bw_min2( ++ v->return_bus_width * v->dcfclk_per_state[i], ++ v->fabric_and_dram_bandwidth_per_state[i] * 1000.0); ++ if (v->dcc_enabled_in_any_plane == dcn_bw_yes ++ && v->return_bw_todcn_per_state ++ > v->dcfclk_per_state[i] * v->return_bus_width ++ / 4.0) { ++ v->return_bw_per_state[i] = ++ dcn_bw_min2( ++ v->return_bw_per_state[i], ++ v->return_bw_todcn_per_state * 4.0 ++ * (1.0 ++ - v->urgent_latency ++ / ((v->rob_buffer_size_in_kbyte ++ - v->pixel_chunk_size_in_kbyte) ++ * 1024.0 ++ / (v->return_bw_todcn_per_state ++ - v->dcfclk_per_state[i] ++ * v->return_bus_width ++ / 4.0) ++ + v->urgent_latency))); ++ } ++ v->critical_point = 2.0 * v->return_bus_width * v->dcfclk_per_state[i] ++ * v->urgent_latency ++ / (v->return_bw_todcn_per_state * v->urgent_latency ++ + (v->rob_buffer_size_in_kbyte ++ - v->pixel_chunk_size_in_kbyte) ++ * 1024.0); ++ if (v->dcc_enabled_in_any_plane == dcn_bw_yes && v->critical_point > 1.0 ++ && v->critical_point < 4.0) { ++ v->return_bw_per_state[i] = ++ dcn_bw_min2( ++ v->return_bw_per_state[i], ++ dcn_bw_pow( ++ 4.0 ++ * v->return_bw_todcn_per_state ++ * (v->rob_buffer_size_in_kbyte ++ - v->pixel_chunk_size_in_kbyte) ++ * 1024.0 ++ * v->return_bus_width ++ * v->dcfclk_per_state[i] ++ * v->urgent_latency ++ / (v->return_bw_todcn_per_state ++ * v->urgent_latency ++ + (v->rob_buffer_size_in_kbyte ++ - v->pixel_chunk_size_in_kbyte) ++ * 1024.0), ++ 2)); ++ } ++ } ++ for (i = 0; i <= number_of_states_plus_one; i++) { ++ if ((v->total_read_bandwidth_consumed_gbyte_per_second * 1000.0 ++ <= v->return_bw_per_state[i]) ++ && (v->total_bandwidth_consumed_gbyte_per_second * 1000.0 ++ <= v->fabric_and_dram_bandwidth_per_state[i] ++ * 1000.0 ++ * v->percent_of_ideal_drambw_received_after_urg_latency ++ / 100.0)) { ++ v->bandwidth_support[i] = dcn_bw_yes; ++ } else { ++ v->bandwidth_support[i] = dcn_bw_no; ++ } ++ } ++ /*writeback latency support check*/ ++ ++ v->writeback_latency_support = dcn_bw_yes; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->output[k] == dcn_bw_writeback && v->output_format[k] == dcn_bw_444 ++ && v->scaler_rec_out_width[k] / (v->htotal[k] / v->pixel_clock[k]) ++ * 4.0 ++ > (v->writeback_luma_buffer_size ++ + v->writeback_chroma_buffer_size) ++ * 1024.0 / v->write_back_latency) { ++ v->writeback_latency_support = dcn_bw_no; ++ } else if (v->output[k] == dcn_bw_writeback ++ && v->scaler_rec_out_width[k] / (v->htotal[k] / v->pixel_clock[k]) ++ > dcn_bw_min2( ++ v->writeback_luma_buffer_size, ++ 2.0 ++ * v->writeback_chroma_buffer_size) ++ * 1024.0 / v->write_back_latency) { ++ v->writeback_latency_support = dcn_bw_no; ++ } ++ } ++ /*re-ordering buffer support check*/ ++ ++ for (i = 0; i <= number_of_states_plus_one; i++) { ++ v->urgent_round_trip_and_out_of_order_latency_per_state[i] = ++ (v->round_trip_ping_latency_cycles + 32.0) / v->dcfclk_per_state[i] ++ + v->urgent_out_of_order_return_per_channel ++ * v->number_of_channels ++ / v->return_bw_per_state[i]; ++ if ((v->rob_buffer_size_in_kbyte - v->pixel_chunk_size_in_kbyte) * 1024.0 ++ / v->return_bw_per_state[i] ++ > v->urgent_round_trip_and_out_of_order_latency_per_state[i]) { ++ v->rob_support[i] = dcn_bw_yes; ++ } else { ++ v->rob_support[i] = dcn_bw_no; ++ } ++ } ++ /*display io support check*/ ++ ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->output[k] == dcn_bw_dp && v->dsc_capability == dcn_bw_yes) { ++ if (v->output_format[k] == dcn_bw_420) { ++ v->required_output_bw = v->pixel_clock[k] / 2.0; ++ } else { ++ v->required_output_bw = v->pixel_clock[k]; ++ } ++ } else if (v->output_format[k] == dcn_bw_420) { ++ v->required_output_bw = v->pixel_clock[k] * 3.0 / 2.0; ++ } else { ++ v->required_output_bw = v->pixel_clock[k] * 3.0; ++ } ++ if (v->output[k] == dcn_bw_hdmi) { ++ v->required_phyclk[k] = v->required_output_bw / 3.0; ++ } else if (v->output[k] == dcn_bw_dp) { ++ v->required_phyclk[k] = v->required_output_bw / 4.0; ++ } else { ++ v->required_phyclk[k] = 0.0; ++ } ++ } ++ for (i = 0; i <= number_of_states_plus_one; i++) { ++ v->dio_support[i] = dcn_bw_yes; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->required_phyclk[k] > v->phyclk_per_state[i] ++ || (v->output[k] == dcn_bw_hdmi ++ && v->required_phyclk[k] > 600.0)) { ++ v->dio_support[i] = dcn_bw_no; ++ } ++ } ++ } ++ /*total available writeback support check*/ ++ ++ v->total_number_of_active_writeback = 0.0; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->output[k] == dcn_bw_writeback) { ++ v->total_number_of_active_writeback = v->total_number_of_active_writeback ++ + 1.0; ++ } ++ } ++ if (v->total_number_of_active_writeback <= v->max_num_writeback) { ++ v->total_available_writeback_support = dcn_bw_yes; ++ } else { ++ v->total_available_writeback_support = dcn_bw_no; ++ } ++ /*maximum dispclk/dppclk support check*/ ++ ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->h_ratio[k] > 1.0) { ++ v->pscl_factor[k] = dcn_bw_min2( ++ v->max_dchub_topscl_throughput, ++ v->max_pscl_tolb_throughput * v->h_ratio[k] ++ / dcn_bw_ceil2(v->htaps[k] / 6.0, 1.0)); ++ } else { ++ v->pscl_factor[k] = dcn_bw_min2( ++ v->max_dchub_topscl_throughput, ++ v->max_pscl_tolb_throughput); ++ } ++ if (v->byte_per_pixel_in_detc[k] == 0.0) { ++ v->pscl_factor_chroma[k] = 0.0; ++ v->min_dppclk_using_single_dpp[k] = ++ v->pixel_clock[k] ++ * dcn_bw_max3( ++ v->vtaps[k] / 6.0 ++ * dcn_bw_min2( ++ 1.0, ++ v->h_ratio[k]), ++ v->h_ratio[k] ++ * v->v_ratio[k] ++ / v->pscl_factor[k], ++ 1.0); ++ } else { ++ if (v->h_ratio[k] / 2.0 > 1.0) { ++ v->pscl_factor_chroma[k] = ++ dcn_bw_min2( ++ v->max_dchub_topscl_throughput, ++ v->max_pscl_tolb_throughput ++ * v->h_ratio[k] ++ / 2.0 ++ / dcn_bw_ceil2( ++ v->hta_pschroma[k] ++ / 6.0, ++ 1.0)); ++ } else { ++ v->pscl_factor_chroma[k] = dcn_bw_min2( ++ v->max_dchub_topscl_throughput, ++ v->max_pscl_tolb_throughput); ++ } ++ v->min_dppclk_using_single_dpp[k] = ++ v->pixel_clock[k] ++ * dcn_bw_max5( ++ v->vtaps[k] / 6.0 ++ * dcn_bw_min2( ++ 1.0, ++ v->h_ratio[k]), ++ v->h_ratio[k] ++ * v->v_ratio[k] ++ / v->pscl_factor[k], ++ v->vta_pschroma[k] / 6.0 ++ * dcn_bw_min2( ++ 1.0, ++ v->h_ratio[k] ++ / 2.0), ++ v->h_ratio[k] ++ * v->v_ratio[k] ++ / 4.0 ++ / v->pscl_factor_chroma[k], ++ 1.0); ++ } ++ } ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if ((v->source_pixel_format[k] == dcn_bw_rgb_sub_64 ++ || v->source_pixel_format[k] == dcn_bw_rgb_sub_32 ++ || v->source_pixel_format[k] == dcn_bw_rgb_sub_16)) { ++ if (v->source_surface_mode[k] == dcn_bw_sw_linear) { ++ v->read256_block_height_y[k] = 1.0; ++ } else if (v->source_pixel_format[k] == dcn_bw_rgb_sub_64) { ++ v->read256_block_height_y[k] = 4.0; ++ } else { ++ v->read256_block_height_y[k] = 8.0; ++ } ++ v->read256_block_width_y[k] = 256.0 ++ / dcn_bw_ceil2(v->byte_per_pixel_in_dety[k], 1.0) ++ / v->read256_block_height_y[k]; ++ v->read256_block_height_c[k] = 0.0; ++ v->read256_block_width_c[k] = 0.0; ++ } else { ++ if (v->source_surface_mode[k] == dcn_bw_sw_linear) { ++ v->read256_block_height_y[k] = 1.0; ++ v->read256_block_height_c[k] = 1.0; ++ } else if (v->source_pixel_format[k] == dcn_bw_yuv420_sub_8) { ++ v->read256_block_height_y[k] = 16.0; ++ v->read256_block_height_c[k] = 8.0; ++ } else { ++ v->read256_block_height_y[k] = 8.0; ++ v->read256_block_height_c[k] = 8.0; ++ } ++ v->read256_block_width_y[k] = 256.0 ++ / dcn_bw_ceil2(v->byte_per_pixel_in_dety[k], 1.0) ++ / v->read256_block_height_y[k]; ++ v->read256_block_width_c[k] = 256.0 ++ / dcn_bw_ceil2(v->byte_per_pixel_in_detc[k], 2.0) ++ / v->read256_block_height_c[k]; ++ } ++ if (v->source_scan[k] == dcn_bw_hor) { ++ v->max_swath_height_y[k] = v->read256_block_height_y[k]; ++ v->max_swath_height_c[k] = v->read256_block_height_c[k]; ++ } else { ++ v->max_swath_height_y[k] = v->read256_block_width_y[k]; ++ v->max_swath_height_c[k] = v->read256_block_width_c[k]; ++ } ++ if ((v->source_pixel_format[k] == dcn_bw_rgb_sub_64 ++ || v->source_pixel_format[k] == dcn_bw_rgb_sub_32 ++ || v->source_pixel_format[k] == dcn_bw_rgb_sub_16)) { ++ if (v->source_surface_mode[k] == dcn_bw_sw_linear ++ || (v->source_pixel_format[k] == dcn_bw_rgb_sub_64 ++ && (v->source_surface_mode[k] ++ == dcn_bw_sw_4_kb_s ++ || v->source_surface_mode[k] ++ == dcn_bw_sw_4_kb_s_x ++ || v->source_surface_mode[k] ++ == dcn_bw_sw_64_kb_s ++ || v->source_surface_mode[k] ++ == dcn_bw_sw_64_kb_s_t ++ || v->source_surface_mode[k] ++ == dcn_bw_sw_64_kb_s_x ++ || v->source_surface_mode[k] ++ == dcn_bw_sw_var_s ++ || v->source_surface_mode[k] ++ == dcn_bw_sw_var_s_x) ++ && v->source_scan[k] == dcn_bw_hor)) { ++ v->min_swath_height_y[k] = v->max_swath_height_y[k]; ++ } else { ++ v->min_swath_height_y[k] = v->max_swath_height_y[k] / 2.0; ++ } ++ v->min_swath_height_c[k] = v->max_swath_height_c[k]; ++ } else { ++ if (v->source_surface_mode[k] == dcn_bw_sw_linear) { ++ v->min_swath_height_y[k] = v->max_swath_height_y[k]; ++ v->min_swath_height_c[k] = v->max_swath_height_c[k]; ++ } else if (v->source_pixel_format[k] == dcn_bw_yuv420_sub_8 ++ && v->source_scan[k] == dcn_bw_hor) { ++ v->min_swath_height_y[k] = v->max_swath_height_y[k] / 2.0; ++ if (v->bug_forcing_luma_and_chroma_request_to_same_size_fixed ++ == dcn_bw_yes) { ++ v->min_swath_height_c[k] = v->max_swath_height_c[k]; ++ } else { ++ v->min_swath_height_c[k] = v->max_swath_height_c[k] / 2.0; ++ } ++ } else if (v->source_pixel_format[k] == dcn_bw_yuv420_sub_10 ++ && v->source_scan[k] == dcn_bw_hor) { ++ v->min_swath_height_c[k] = v->max_swath_height_c[k] / 2.0; ++ if (v->bug_forcing_luma_and_chroma_request_to_same_size_fixed ++ == dcn_bw_yes) { ++ v->min_swath_height_y[k] = v->max_swath_height_y[k]; ++ } else { ++ v->min_swath_height_y[k] = v->max_swath_height_y[k] / 2.0; ++ } ++ } else { ++ v->min_swath_height_y[k] = v->max_swath_height_y[k]; ++ v->min_swath_height_c[k] = v->max_swath_height_c[k]; ++ } ++ } ++ if (v->source_surface_mode[k] == dcn_bw_sw_linear) { ++ v->maximum_swath_width = 8192.0; ++ } else { ++ v->maximum_swath_width = 5120.0; ++ } ++ v->number_of_dpp_required_for_det_size = ++ dcn_bw_ceil2( ++ v->swath_width_ysingle_dpp[k] ++ / dcn_bw_min2( ++ v->maximum_swath_width, ++ v->det_buffer_size_in_kbyte ++ * 1024.0 ++ / 2.0 ++ / (v->byte_per_pixel_in_dety[k] ++ * v->min_swath_height_y[k] ++ + v->byte_per_pixel_in_detc[k] ++ / 2.0 ++ * v->min_swath_height_c[k])), ++ 1.0); ++ if (v->byte_per_pixel_in_detc[k] == 0.0) { ++ v->number_of_dpp_required_for_lb_size = dcn_bw_ceil2( ++ (v->vtaps[k] ++ + dcn_bw_max2( ++ dcn_bw_ceil2(v->v_ratio[k], 1.0) ++ - 2, ++ 0.0)) ++ * v->swath_width_ysingle_dpp[k] ++ / dcn_bw_max2(v->h_ratio[k], 1.0) ++ * v->lb_bit_per_pixel[k] ++ / v->line_buffer_size, ++ 1.0); ++ } else { ++ v->number_of_dpp_required_for_lb_size = ++ dcn_bw_max2( ++ dcn_bw_ceil2( ++ (v->vtaps[k] ++ + dcn_bw_max2( ++ dcn_bw_ceil2( ++ v->v_ratio[k], ++ 1.0) ++ - 2, ++ 0.0)) ++ * v->swath_width_ysingle_dpp[k] ++ / dcn_bw_max2( ++ v->h_ratio[k], ++ 1.0) ++ * v->lb_bit_per_pixel[k] ++ / v->line_buffer_size, ++ 1.0), ++ dcn_bw_ceil2( ++ (v->vta_pschroma[k] ++ + dcn_bw_max2( ++ dcn_bw_ceil2( ++ v->v_ratio[k] ++ / 2.0, ++ 1.0) ++ - 2, ++ 0.0)) ++ * v->swath_width_ysingle_dpp[k] ++ / 2.0 ++ / dcn_bw_max2( ++ v->h_ratio[k] ++ / 2.0, ++ 1.0) ++ * v->lb_bit_per_pixel[k] ++ / v->line_buffer_size, ++ 1.0)); ++ } ++ v->number_of_dpp_required_for_det_and_lb_size[k] = dcn_bw_max2( ++ v->number_of_dpp_required_for_det_size, ++ v->number_of_dpp_required_for_lb_size); ++ } ++ for (i = 0; i <= number_of_states_plus_one; i++) { ++ for (j = 0; j <= 1; j++) { ++ v->total_number_of_active_dpp[i][j] = 0.0; ++ v->required_dispclk[i][j] = 0.0; ++ v->dispclk_dppclk_support[i][j] = dcn_bw_yes; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ v->min_dispclk_using_single_dpp = dcn_bw_max2( ++ v->pixel_clock[k], ++ v->min_dppclk_using_single_dpp[k] * (j + 1)) ++ * (1.0 + v->downspreading / 100.0); ++ if (v->odm_capability == dcn_bw_yes) { ++ v->min_dispclk_using_dual_dpp = dcn_bw_max2( ++ v->pixel_clock[k] / 2.0, ++ v->min_dppclk_using_single_dpp[k] / 2.0 ++ * (j + 1)) ++ * (1.0 + v->downspreading / 100.0); ++ } else { ++ v->min_dispclk_using_dual_dpp = dcn_bw_max2( ++ v->pixel_clock[k], ++ v->min_dppclk_using_single_dpp[k] / 2.0 ++ * (j + 1)) ++ * (1.0 + v->downspreading / 100.0); ++ } ++ if (i < number_of_states) { ++ v->min_dispclk_using_single_dpp = ++ v->min_dispclk_using_single_dpp ++ * (1.0 ++ + v->dispclk_ramping_margin ++ / 100.0); ++ v->min_dispclk_using_dual_dpp = ++ v->min_dispclk_using_dual_dpp ++ * (1.0 ++ + v->dispclk_ramping_margin ++ / 100.0); ++ } ++ if (v->min_dispclk_using_single_dpp ++ <= dcn_bw_min2( ++ v->max_dispclk[i], ++ (j + 1) * v->max_dppclk[i]) ++ && v->number_of_dpp_required_for_det_and_lb_size[k] ++ <= 1.0) { ++ v->no_of_dpp[i][j][k] = 1.0; ++ v->required_dispclk[i][j] = dcn_bw_max2( ++ v->required_dispclk[i][j], ++ v->min_dispclk_using_single_dpp); ++ } else if (v->min_dispclk_using_dual_dpp ++ <= dcn_bw_min2( ++ v->max_dispclk[i], ++ (j + 1) * v->max_dppclk[i])) { ++ v->no_of_dpp[i][j][k] = 2.0; ++ v->required_dispclk[i][j] = dcn_bw_max2( ++ v->required_dispclk[i][j], ++ v->min_dispclk_using_dual_dpp); ++ } else { ++ v->no_of_dpp[i][j][k] = 2.0; ++ v->required_dispclk[i][j] = dcn_bw_max2( ++ v->required_dispclk[i][j], ++ v->min_dispclk_using_dual_dpp); ++ v->dispclk_dppclk_support[i][j] = dcn_bw_no; ++ } ++ v->total_number_of_active_dpp[i][j] = ++ v->total_number_of_active_dpp[i][j] ++ + v->no_of_dpp[i][j][k]; ++ } ++ if (v->total_number_of_active_dpp[i][j] > v->max_num_dpp) { ++ v->total_number_of_active_dpp[i][j] = 0.0; ++ v->required_dispclk[i][j] = 0.0; ++ v->dispclk_dppclk_support[i][j] = dcn_bw_yes; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ v->min_dispclk_using_single_dpp = dcn_bw_max2( ++ v->pixel_clock[k], ++ v->min_dppclk_using_single_dpp[k] * (j + 1)) ++ * (1.0 + v->downspreading / 100.0); ++ v->min_dispclk_using_dual_dpp = dcn_bw_max2( ++ v->pixel_clock[k], ++ v->min_dppclk_using_single_dpp[k] / 2.0 ++ * (j + 1)) ++ * (1.0 + v->downspreading / 100.0); ++ if (i < number_of_states) { ++ v->min_dispclk_using_single_dpp = ++ v->min_dispclk_using_single_dpp ++ * (1.0 ++ + v->dispclk_ramping_margin ++ / 100.0); ++ v->min_dispclk_using_dual_dpp = ++ v->min_dispclk_using_dual_dpp ++ * (1.0 ++ + v->dispclk_ramping_margin ++ / 100.0); ++ } ++ if (v->number_of_dpp_required_for_det_and_lb_size[k] ++ <= 1.0) { ++ v->no_of_dpp[i][j][k] = 1.0; ++ v->required_dispclk[i][j] = dcn_bw_max2( ++ v->required_dispclk[i][j], ++ v->min_dispclk_using_single_dpp); ++ if (v->min_dispclk_using_single_dpp ++ > dcn_bw_min2( ++ v->max_dispclk[i], ++ (j + 1) ++ * v->max_dppclk[i])) { ++ v->dispclk_dppclk_support[i][j] = dcn_bw_no; ++ } ++ } else { ++ v->no_of_dpp[i][j][k] = 2.0; ++ v->required_dispclk[i][j] = dcn_bw_max2( ++ v->required_dispclk[i][j], ++ v->min_dispclk_using_dual_dpp); ++ if (v->min_dispclk_using_dual_dpp ++ > dcn_bw_min2( ++ v->max_dispclk[i], ++ (j + 1) ++ * v->max_dppclk[i])) { ++ v->dispclk_dppclk_support[i][j] = dcn_bw_no; ++ } ++ } ++ v->total_number_of_active_dpp[i][j] = ++ v->total_number_of_active_dpp[i][j] ++ + v->no_of_dpp[i][j][k]; ++ } ++ } ++ } ++ } ++ /*viewport size check*/ ++ ++ v->viewport_size_support = dcn_bw_yes; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->number_of_dpp_required_for_det_and_lb_size[k] > 2.0) { ++ v->viewport_size_support = dcn_bw_no; ++ } ++ } ++ /*total available pipes support check*/ ++ ++ for (i = 0; i <= number_of_states_plus_one; i++) { ++ for (j = 0; j <= 1; j++) { ++ if (v->total_number_of_active_dpp[i][j] <= v->max_num_dpp) { ++ v->total_available_pipes_support[i][j] = dcn_bw_yes; ++ } else { ++ v->total_available_pipes_support[i][j] = dcn_bw_no; ++ } ++ } ++ } ++ /*urgent latency support check*/ ++ ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ for (i = 0; i <= number_of_states_plus_one; i++) { ++ for (j = 0; j <= 1; j++) { ++ v->swath_width_yper_state[i][j][k] = v->swath_width_ysingle_dpp[k] ++ / v->no_of_dpp[i][j][k]; ++ v->swath_width_granularity_y = 256.0 ++ / dcn_bw_ceil2(v->byte_per_pixel_in_dety[k], 1.0) ++ / v->max_swath_height_y[k]; ++ v->rounded_up_max_swath_size_bytes_y = (dcn_bw_ceil2( ++ v->swath_width_yper_state[i][j][k] - 1.0, ++ v->swath_width_granularity_y) ++ + v->swath_width_granularity_y) ++ * v->byte_per_pixel_in_dety[k] ++ * v->max_swath_height_y[k]; ++ if (v->source_pixel_format[k] == dcn_bw_yuv420_sub_10) { ++ v->rounded_up_max_swath_size_bytes_y = dcn_bw_ceil2( ++ v->rounded_up_max_swath_size_bytes_y, ++ 256.0) + 256; ++ } ++ if (v->max_swath_height_c[k] > 0.0) { ++ v->swath_width_granularity_c = ++ 256.0 ++ / dcn_bw_ceil2( ++ v->byte_per_pixel_in_detc[k], ++ 2.0) ++ / v->max_swath_height_c[k]; ++ } ++ v->rounded_up_max_swath_size_bytes_c = (dcn_bw_ceil2( ++ v->swath_width_yper_state[i][j][k] / 2.0 - 1.0, ++ v->swath_width_granularity_c) ++ + v->swath_width_granularity_c) ++ * v->byte_per_pixel_in_detc[k] ++ * v->max_swath_height_c[k]; ++ if (v->source_pixel_format[k] == dcn_bw_yuv420_sub_10) { ++ v->rounded_up_max_swath_size_bytes_c = dcn_bw_ceil2( ++ v->rounded_up_max_swath_size_bytes_c, ++ 256.0) + 256; ++ } ++ if (v->rounded_up_max_swath_size_bytes_y ++ + v->rounded_up_max_swath_size_bytes_c ++ <= v->det_buffer_size_in_kbyte * 1024.0 / 2.0) { ++ v->swath_height_yper_state[i][j][k] = ++ v->max_swath_height_y[k]; ++ v->swath_height_cper_state[i][j][k] = ++ v->max_swath_height_c[k]; ++ } else { ++ v->swath_height_yper_state[i][j][k] = ++ v->min_swath_height_y[k]; ++ v->swath_height_cper_state[i][j][k] = ++ v->min_swath_height_c[k]; ++ } ++ if (v->byte_per_pixel_in_detc[k] == 0.0) { ++ v->lines_in_det_luma = v->det_buffer_size_in_kbyte * 1024.0 ++ / v->byte_per_pixel_in_dety[k] ++ / v->swath_width_yper_state[i][j][k]; ++ v->lines_in_det_chroma = 0.0; ++ } else if (v->swath_height_yper_state[i][j][k] ++ <= v->swath_height_cper_state[i][j][k]) { ++ v->lines_in_det_luma = v->det_buffer_size_in_kbyte * 1024.0 ++ / 2.0 / v->byte_per_pixel_in_dety[k] ++ / v->swath_width_yper_state[i][j][k]; ++ v->lines_in_det_chroma = ++ v->det_buffer_size_in_kbyte * 1024.0 / 2.0 ++ / v->byte_per_pixel_in_detc[k] ++ / (v->swath_width_yper_state[i][j][k] ++ / 2.0); ++ } else { ++ v->lines_in_det_luma = v->det_buffer_size_in_kbyte * 1024.0 ++ * 2.0 / 3.0 / v->byte_per_pixel_in_dety[k] ++ / v->swath_width_yper_state[i][j][k]; ++ v->lines_in_det_chroma = ++ v->det_buffer_size_in_kbyte * 1024.0 / 3.0 ++ / v->byte_per_pixel_in_dety[k] ++ / (v->swath_width_yper_state[i][j][k] ++ / 2.0); ++ } ++ v->effective_lb_latency_hiding_source_lines_luma = ++ dcn_bw_min2( ++ v->max_line_buffer_lines, ++ dcn_bw_floor2( ++ v->line_buffer_size ++ / v->lb_bit_per_pixel[k] ++ / (v->swath_width_yper_state[i][j][k] ++ / dcn_bw_max2( ++ v->h_ratio[k], ++ 1.0)), ++ 1.0)) ++ - (v->vtaps[k] - 1.0); ++ v->effective_lb_latency_hiding_source_lines_chroma = ++ dcn_bw_min2( ++ v->max_line_buffer_lines, ++ dcn_bw_floor2( ++ v->line_buffer_size ++ / v->lb_bit_per_pixel[k] ++ / (v->swath_width_yper_state[i][j][k] ++ / 2.0 ++ / dcn_bw_max2( ++ v->h_ratio[k] ++ / 2.0, ++ 1.0)), ++ 1.0)) ++ - (v->vta_pschroma[k] - 1.0); ++ v->effective_detlb_lines_luma = ++ dcn_bw_floor2( ++ v->lines_in_det_luma ++ + dcn_bw_min2( ++ v->lines_in_det_luma ++ * v->required_dispclk[i][j] ++ * v->byte_per_pixel_in_dety[k] ++ * v->pscl_factor[k] ++ / v->return_bw_per_state[i], ++ v->effective_lb_latency_hiding_source_lines_luma), ++ v->swath_height_yper_state[i][j][k]); ++ v->effective_detlb_lines_chroma = ++ dcn_bw_floor2( ++ v->lines_in_det_chroma ++ + dcn_bw_min2( ++ v->lines_in_det_chroma ++ * v->required_dispclk[i][j] ++ * v->byte_per_pixel_in_detc[k] ++ * v->pscl_factor_chroma[k] ++ / v->return_bw_per_state[i], ++ v->effective_lb_latency_hiding_source_lines_chroma), ++ v->swath_height_cper_state[i][j][k]); ++ if (v->byte_per_pixel_in_detc[k] == 0.0) { ++ v->urgent_latency_support_us_per_state[i][j][k] = ++ v->effective_detlb_lines_luma ++ * (v->htotal[k] ++ / v->pixel_clock[k]) ++ / v->v_ratio[k] ++ - v->effective_detlb_lines_luma ++ * v->swath_width_yper_state[i][j][k] ++ * dcn_bw_ceil2( ++ v->byte_per_pixel_in_dety[k], ++ 1.0) ++ / (v->return_bw_per_state[i] ++ / v->no_of_dpp[i][j][k]); ++ } else { ++ v->urgent_latency_support_us_per_state[i][j][k] = ++ dcn_bw_min2( ++ v->effective_detlb_lines_luma ++ * (v->htotal[k] ++ / v->pixel_clock[k]) ++ / v->v_ratio[k] ++ - v->effective_detlb_lines_luma ++ * v->swath_width_yper_state[i][j][k] ++ * dcn_bw_ceil2( ++ v->byte_per_pixel_in_dety[k], ++ 1.0) ++ / (v->return_bw_per_state[i] ++ / v->no_of_dpp[i][j][k]), ++ v->effective_detlb_lines_chroma ++ * (v->htotal[k] ++ / v->pixel_clock[k]) ++ / (v->v_ratio[k] ++ / 2.0) ++ - v->effective_detlb_lines_chroma ++ * v->swath_width_yper_state[i][j][k] ++ / 2.0 ++ * dcn_bw_ceil2( ++ v->byte_per_pixel_in_detc[k], ++ 2.0) ++ / (v->return_bw_per_state[i] ++ / v->no_of_dpp[i][j][k])); ++ } ++ } ++ } ++ } ++ for (i = 0; i <= number_of_states_plus_one; i++) { ++ for (j = 0; j <= 1; j++) { ++ v->urgent_latency_support[i][j] = dcn_bw_yes; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->urgent_latency_support_us_per_state[i][j][k] ++ < v->urgent_latency / 1.0) { ++ v->urgent_latency_support[i][j] = dcn_bw_no; ++ } ++ } ++ } ++ } ++ /*prefetch check*/ ++ ++ for (i = 0; i <= number_of_states_plus_one; i++) { ++ for (j = 0; j <= 1; j++) { ++ v->total_number_of_dcc_active_dpp[i][j] = 0.0; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->dcc_enable[k] == dcn_bw_yes) { ++ v->total_number_of_dcc_active_dpp[i][j] = ++ v->total_number_of_dcc_active_dpp[i][j] ++ + v->no_of_dpp[i][j][k]; ++ } ++ } ++ } ++ } ++ for (i = 0; i <= number_of_states_plus_one; i++) { ++ for (j = 0; j <= 1; j++) { ++ v->projected_dcfclk_deep_sleep = 8.0; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ v->projected_dcfclk_deep_sleep = dcn_bw_max2( ++ v->projected_dcfclk_deep_sleep, ++ v->pixel_clock[k] / 16.0); ++ if (v->byte_per_pixel_in_detc[k] == 0.0) { ++ if (v->v_ratio[k] <= 1.0) { ++ v->projected_dcfclk_deep_sleep = ++ dcn_bw_max2( ++ v->projected_dcfclk_deep_sleep, ++ 1.1 ++ * dcn_bw_ceil2( ++ v->byte_per_pixel_in_dety[k], ++ 1.0) ++ / 64.0 ++ * v->h_ratio[k] ++ * v->pixel_clock[k] ++ / v->no_of_dpp[i][j][k]); ++ } else { ++ v->projected_dcfclk_deep_sleep = ++ dcn_bw_max2( ++ v->projected_dcfclk_deep_sleep, ++ 1.1 ++ * dcn_bw_ceil2( ++ v->byte_per_pixel_in_dety[k], ++ 1.0) ++ / 64.0 ++ * v->pscl_factor[k] ++ * v->required_dispclk[i][j] ++ / (1 ++ + j)); ++ } ++ } else { ++ if (v->v_ratio[k] <= 1.0) { ++ v->projected_dcfclk_deep_sleep = ++ dcn_bw_max2( ++ v->projected_dcfclk_deep_sleep, ++ 1.1 ++ * dcn_bw_ceil2( ++ v->byte_per_pixel_in_dety[k], ++ 1.0) ++ / 32.0 ++ * v->h_ratio[k] ++ * v->pixel_clock[k] ++ / v->no_of_dpp[i][j][k]); ++ } else { ++ v->projected_dcfclk_deep_sleep = ++ dcn_bw_max2( ++ v->projected_dcfclk_deep_sleep, ++ 1.1 ++ * dcn_bw_ceil2( ++ v->byte_per_pixel_in_dety[k], ++ 1.0) ++ / 32.0 ++ * v->pscl_factor[k] ++ * v->required_dispclk[i][j] ++ / (1 ++ + j)); ++ } ++ if (v->v_ratio[k] / 2.0 <= 1.0) { ++ v->projected_dcfclk_deep_sleep = ++ dcn_bw_max2( ++ v->projected_dcfclk_deep_sleep, ++ 1.1 ++ * dcn_bw_ceil2( ++ v->byte_per_pixel_in_detc[k], ++ 2.0) ++ / 32.0 ++ * v->h_ratio[k] ++ / 2.0 ++ * v->pixel_clock[k] ++ / v->no_of_dpp[i][j][k]); ++ } else { ++ v->projected_dcfclk_deep_sleep = ++ dcn_bw_max2( ++ v->projected_dcfclk_deep_sleep, ++ 1.1 ++ * dcn_bw_ceil2( ++ v->byte_per_pixel_in_detc[k], ++ 2.0) ++ / 32.0 ++ * v->pscl_factor_chroma[k] ++ * v->required_dispclk[i][j] ++ / (1 ++ + j)); ++ } ++ } ++ } ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->dcc_enable[k] == dcn_bw_yes) { ++ v->meta_req_height_y = 8.0 * v->read256_block_height_y[k]; ++ v->meta_req_width_y = ++ 64.0 * 256.0 ++ / dcn_bw_ceil2( ++ v->byte_per_pixel_in_dety[k], ++ 1.0) ++ / v->meta_req_height_y; ++ v->meta_surface_width_y = dcn_bw_ceil2( ++ v->viewport_width[k] / v->no_of_dpp[i][j][k] ++ - 1.0, ++ v->meta_req_width_y) + v->meta_req_width_y; ++ v->meta_surface_height_y = dcn_bw_ceil2( ++ v->viewport_height[k] - 1.0, ++ v->meta_req_height_y) ++ + v->meta_req_height_y; ++ if (v->pte_enable == dcn_bw_yes) { ++ v->meta_pte_bytes_per_frame_y = ++ (dcn_bw_ceil2( ++ (v->meta_surface_width_y ++ * v->meta_surface_height_y ++ * dcn_bw_ceil2( ++ v->byte_per_pixel_in_dety[k], ++ 1.0) ++ / 256.0 ++ - 4096.0) ++ / 8.0 ++ / 4096.0, ++ 1.0) + 1) * 64.0; ++ } else { ++ v->meta_pte_bytes_per_frame_y = 0.0; ++ } ++ if (v->source_scan[k] == dcn_bw_hor) { ++ v->meta_row_bytes_y = ++ v->meta_surface_width_y ++ * v->meta_req_height_y ++ * dcn_bw_ceil2( ++ v->byte_per_pixel_in_dety[k], ++ 1.0) ++ / 256.0; ++ } else { ++ v->meta_row_bytes_y = ++ v->meta_surface_height_y ++ * v->meta_req_width_y ++ * dcn_bw_ceil2( ++ v->byte_per_pixel_in_dety[k], ++ 1.0) ++ / 256.0; ++ } ++ } else { ++ v->meta_pte_bytes_per_frame_y = 0.0; ++ v->meta_row_bytes_y = 0.0; ++ } ++ if (v->pte_enable == dcn_bw_yes) { ++ if (v->source_surface_mode[k] == dcn_bw_sw_linear) { ++ v->macro_tile_block_size_bytes_y = 256.0; ++ v->macro_tile_block_height_y = 1.0; ++ } else if (v->source_surface_mode[k] == dcn_bw_sw_4_kb_s ++ || v->source_surface_mode[k] ++ == dcn_bw_sw_4_kb_s_x ++ || v->source_surface_mode[k] ++ == dcn_bw_sw_4_kb_d ++ || v->source_surface_mode[k] ++ == dcn_bw_sw_4_kb_d_x) { ++ v->macro_tile_block_size_bytes_y = 4096.0; ++ v->macro_tile_block_height_y = 4.0 ++ * v->read256_block_height_y[k]; ++ } else if (v->source_surface_mode[k] == dcn_bw_sw_64_kb_s ++ || v->source_surface_mode[k] ++ == dcn_bw_sw_64_kb_s_t ++ || v->source_surface_mode[k] ++ == dcn_bw_sw_64_kb_s_x ++ || v->source_surface_mode[k] ++ == dcn_bw_sw_64_kb_d ++ || v->source_surface_mode[k] ++ == dcn_bw_sw_64_kb_d_t ++ || v->source_surface_mode[k] ++ == dcn_bw_sw_64_kb_d_x) { ++ v->macro_tile_block_size_bytes_y = 64.0 * 1024; ++ v->macro_tile_block_height_y = 16.0 ++ * v->read256_block_height_y[k]; ++ } else { ++ v->macro_tile_block_size_bytes_y = 256.0 * 1024; ++ v->macro_tile_block_height_y = 32.0 ++ * v->read256_block_height_y[k]; ++ } ++ if (v->macro_tile_block_size_bytes_y <= 65536.0) { ++ v->data_pte_req_height_y = ++ v->macro_tile_block_height_y; ++ } else { ++ v->data_pte_req_height_y = 16.0 ++ * v->read256_block_height_y[k]; ++ } ++ v->data_pte_req_width_y = ++ 4096.0 ++ / dcn_bw_ceil2( ++ v->byte_per_pixel_in_dety[k], ++ 1.0) ++ / v->data_pte_req_height_y ++ * 8; ++ if (v->source_surface_mode[k] == dcn_bw_sw_linear) { ++ v->dpte_bytes_per_row_y = ++ 64.0 ++ * (dcn_bw_ceil2( ++ (v->viewport_width[k] ++ / v->no_of_dpp[i][j][k] ++ * dcn_bw_min2( ++ 128.0, ++ dcn_bw_pow( ++ 2.0, ++ dcn_bw_floor2( ++ dcn_bw_log( ++ v->pte_buffer_size_in_requests ++ * v->data_pte_req_width_y ++ / (v->viewport_width[k] ++ / v->no_of_dpp[i][j][k]), ++ 2.0), ++ 1.0))) ++ - 1.0) ++ / v->data_pte_req_width_y, ++ 1.0) ++ + 1); ++ } else if (v->source_scan[k] == dcn_bw_hor) { ++ v->dpte_bytes_per_row_y = ++ 64.0 ++ * (dcn_bw_ceil2( ++ (v->viewport_width[k] ++ / v->no_of_dpp[i][j][k] ++ - 1.0) ++ / v->data_pte_req_width_y, ++ 1.0) ++ + 1); ++ } else { ++ v->dpte_bytes_per_row_y = ++ 64.0 ++ * (dcn_bw_ceil2( ++ (v->viewport_height[k] ++ - 1.0) ++ / v->data_pte_req_height_y, ++ 1.0) ++ + 1); ++ } ++ } else { ++ v->dpte_bytes_per_row_y = 0.0; ++ } ++ if ((v->source_pixel_format[k] != dcn_bw_rgb_sub_64 ++ && v->source_pixel_format[k] != dcn_bw_rgb_sub_32 ++ && v->source_pixel_format[k] != dcn_bw_rgb_sub_16)) { ++ if (v->dcc_enable[k] == dcn_bw_yes) { ++ v->meta_req_height_c = 8.0 ++ * v->read256_block_height_c[k]; ++ v->meta_req_width_c = ++ 64.0 * 256.0 ++ / dcn_bw_ceil2( ++ v->byte_per_pixel_in_detc[k], ++ 2.0) ++ / v->meta_req_height_c; ++ v->meta_surface_width_c = ++ dcn_bw_ceil2( ++ v->viewport_width[k] ++ / v->no_of_dpp[i][j][k] ++ / 2.0 ++ - 1.0, ++ v->meta_req_width_c) ++ + v->meta_req_width_c; ++ v->meta_surface_height_c = dcn_bw_ceil2( ++ v->viewport_height[k] / 2.0 - 1.0, ++ v->meta_req_height_c) ++ + v->meta_req_height_c; ++ if (v->pte_enable == dcn_bw_yes) { ++ v->meta_pte_bytes_per_frame_c = ++ (dcn_bw_ceil2( ++ (v->meta_surface_width_c ++ * v->meta_surface_height_c ++ * dcn_bw_ceil2( ++ v->byte_per_pixel_in_detc[k], ++ 2.0) ++ / 256.0 ++ - 4096.0) ++ / 8.0 ++ / 4096.0, ++ 1.0) + 1) ++ * 64.0; ++ } else { ++ v->meta_pte_bytes_per_frame_c = 0.0; ++ } ++ if (v->source_scan[k] == dcn_bw_hor) { ++ v->meta_row_bytes_c = ++ v->meta_surface_width_c ++ * v->meta_req_height_c ++ * dcn_bw_ceil2( ++ v->byte_per_pixel_in_detc[k], ++ 2.0) ++ / 256.0; ++ } else { ++ v->meta_row_bytes_c = ++ v->meta_surface_height_c ++ * v->meta_req_width_c ++ * dcn_bw_ceil2( ++ v->byte_per_pixel_in_detc[k], ++ 2.0) ++ / 256.0; ++ } ++ } else { ++ v->meta_pte_bytes_per_frame_c = 0.0; ++ v->meta_row_bytes_c = 0.0; ++ } ++ if (v->pte_enable == dcn_bw_yes) { ++ if (v->source_surface_mode[k] == dcn_bw_sw_linear) { ++ v->macro_tile_block_size_bytes_c = 256.0; ++ v->macro_tile_block_height_c = 1.0; ++ } else if (v->source_surface_mode[k] ++ == dcn_bw_sw_4_kb_s ++ || v->source_surface_mode[k] ++ == dcn_bw_sw_4_kb_s_x ++ || v->source_surface_mode[k] ++ == dcn_bw_sw_4_kb_d ++ || v->source_surface_mode[k] ++ == dcn_bw_sw_4_kb_d_x) { ++ v->macro_tile_block_size_bytes_c = 4096.0; ++ v->macro_tile_block_height_c = ++ 4.0 ++ * v->read256_block_height_c[k]; ++ } else if (v->source_surface_mode[k] ++ == dcn_bw_sw_64_kb_s ++ || v->source_surface_mode[k] ++ == dcn_bw_sw_64_kb_s_t ++ || v->source_surface_mode[k] ++ == dcn_bw_sw_64_kb_s_x ++ || v->source_surface_mode[k] ++ == dcn_bw_sw_64_kb_d ++ || v->source_surface_mode[k] ++ == dcn_bw_sw_64_kb_d_t ++ || v->source_surface_mode[k] ++ == dcn_bw_sw_64_kb_d_x) { ++ v->macro_tile_block_size_bytes_c = 64.0 ++ * 1024; ++ v->macro_tile_block_height_c = ++ 16.0 ++ * v->read256_block_height_c[k]; ++ } else { ++ v->macro_tile_block_size_bytes_c = 256.0 ++ * 1024; ++ v->macro_tile_block_height_c = ++ 32.0 ++ * v->read256_block_height_c[k]; ++ } ++ v->macro_tile_block_width_c = ++ v->macro_tile_block_size_bytes_c ++ / dcn_bw_ceil2( ++ v->byte_per_pixel_in_detc[k], ++ 2.0) ++ / v->macro_tile_block_height_c; ++ if (v->macro_tile_block_size_bytes_c <= 65536.0) { ++ v->data_pte_req_height_c = ++ v->macro_tile_block_height_c; ++ } else { ++ v->data_pte_req_height_c = ++ 16.0 ++ * v->read256_block_height_c[k]; ++ } ++ v->data_pte_req_width_c = ++ 4096.0 ++ / dcn_bw_ceil2( ++ v->byte_per_pixel_in_detc[k], ++ 2.0) ++ / v->data_pte_req_height_c ++ * 8; ++ if (v->source_surface_mode[k] == dcn_bw_sw_linear) { ++ v->dpte_bytes_per_row_c = ++ 64.0 ++ * (dcn_bw_ceil2( ++ (v->viewport_width[k] ++ / v->no_of_dpp[i][j][k] ++ / 2.0 ++ * dcn_bw_min2( ++ 128.0, ++ dcn_bw_pow( ++ 2.0, ++ dcn_bw_floor2( ++ dcn_bw_log( ++ v->pte_buffer_size_in_requests ++ * v->data_pte_req_width_c ++ / (v->viewport_width[k] ++ / v->no_of_dpp[i][j][k] ++ / 2.0), ++ 2.0), ++ 1.0))) ++ - 1.0) ++ / v->data_pte_req_width_c, ++ 1.0) ++ + 1); ++ } else if (v->source_scan[k] == dcn_bw_hor) { ++ v->dpte_bytes_per_row_c = ++ 64.0 ++ * (dcn_bw_ceil2( ++ (v->viewport_width[k] ++ / v->no_of_dpp[i][j][k] ++ / 2.0 ++ - 1.0) ++ / v->data_pte_req_width_c, ++ 1.0) ++ + 1); ++ } else { ++ v->dpte_bytes_per_row_c = ++ 64.0 ++ * (dcn_bw_ceil2( ++ (v->viewport_height[k] ++ / 2.0 ++ - 1.0) ++ / v->data_pte_req_height_c, ++ 1.0) ++ + 1); ++ } ++ } else { ++ v->dpte_bytes_per_row_c = 0.0; ++ } ++ } else { ++ v->dpte_bytes_per_row_c = 0.0; ++ v->meta_pte_bytes_per_frame_c = 0.0; ++ v->meta_row_bytes_c = 0.0; ++ } ++ v->dpte_bytes_per_row[k] = v->dpte_bytes_per_row_y ++ + v->dpte_bytes_per_row_c; ++ v->meta_pte_bytes_per_frame[k] = v->meta_pte_bytes_per_frame_y ++ + v->meta_pte_bytes_per_frame_c; ++ v->meta_row_bytes[k] = v->meta_row_bytes_y + v->meta_row_bytes_c; ++ v->v_init_y = (v->v_ratio[k] + v->vtaps[k] + 1.0 ++ + v->interlace_output[k] * 0.5 * v->v_ratio[k]) ++ / 2.0; ++ v->prefill_y[k] = dcn_bw_floor2(v->v_init_y, 1.0); ++ v->max_num_sw_y[k] = ++ dcn_bw_ceil2( ++ (v->prefill_y[k] - 1.0) ++ / v->swath_height_yper_state[i][j][k], ++ 1.0) + 1; ++ if (v->prefill_y[k] > 1.0) { ++ v->max_partial_sw_y = dcn_bw_mod( ++ (v->prefill_y[k] - 2.0), ++ v->swath_height_yper_state[i][j][k]); ++ } else { ++ v->max_partial_sw_y = ++ dcn_bw_mod( ++ (v->prefill_y[k] ++ + v->swath_height_yper_state[i][j][k] ++ - 2.0), ++ v->swath_height_yper_state[i][j][k]); ++ } ++ v->max_partial_sw_y = dcn_bw_max2(1.0, v->max_partial_sw_y); ++ v->prefetch_lines_y[k] = v->max_num_sw_y[k] ++ * v->swath_height_yper_state[i][j][k] ++ + v->max_partial_sw_y; ++ if ((v->source_pixel_format[k] != dcn_bw_rgb_sub_64 ++ && v->source_pixel_format[k] != dcn_bw_rgb_sub_32 ++ && v->source_pixel_format[k] != dcn_bw_rgb_sub_16)) { ++ v->v_init_c = (v->v_ratio[k] / 2.0 + v->vtaps[k] + 1.0 ++ + v->interlace_output[k] * 0.5 ++ * v->v_ratio[k] / 2.0) ++ / 2.0; ++ v->prefill_c[k] = dcn_bw_floor2(v->v_init_c, 1.0); ++ v->max_num_sw_c[k] = ++ dcn_bw_ceil2( ++ (v->prefill_c[k] - 1.0) ++ / v->swath_height_cper_state[i][j][k], ++ 1.0) + 1; ++ if (v->prefill_c[k] > 1.0) { ++ v->max_partial_sw_c = ++ dcn_bw_mod( ++ (v->prefill_c[k] ++ - 2.0), ++ v->swath_height_cper_state[i][j][k]); ++ } else { ++ v->max_partial_sw_c = ++ dcn_bw_mod( ++ (v->prefill_c[k] ++ + v->swath_height_cper_state[i][j][k] ++ - 2.0), ++ v->swath_height_cper_state[i][j][k]); ++ } ++ v->max_partial_sw_c = dcn_bw_max2(1.0, v->max_partial_sw_c); ++ v->prefetch_lines_c[k] = v->max_num_sw_c[k] ++ * v->swath_height_cper_state[i][j][k] ++ + v->max_partial_sw_c; ++ } else { ++ v->prefetch_lines_c[k] = 0.0; ++ } ++ v->dst_x_after_scaler = 90.0 * v->pixel_clock[k] ++ / (v->required_dispclk[i][j] / (j + 1)) ++ + 42.0 * v->pixel_clock[k] ++ / v->required_dispclk[i][j]; ++ if (v->no_of_dpp[i][j][k] > 1.0) { ++ v->dst_x_after_scaler = v->dst_x_after_scaler ++ + v->scaler_rec_out_width[k] / 2.0; ++ } ++ if (v->output_format[k] == dcn_bw_420) { ++ v->dst_y_after_scaler = 1.0; ++ } else { ++ v->dst_y_after_scaler = 0.0; ++ } ++ v->time_calc = 24.0 / v->projected_dcfclk_deep_sleep; ++ v->v_update_offset[k] = dcn_bw_ceil2(v->htotal[k] / 4.0, 1.0); ++ v->total_repeater_delay = v->max_inter_dcn_tile_repeaters ++ * (2.0 / (v->required_dispclk[i][j] / (j + 1)) ++ + 3.0 / v->required_dispclk[i][j]); ++ v->v_update_width[k] = (14.0 / v->projected_dcfclk_deep_sleep ++ + 12.0 / (v->required_dispclk[i][j] / (j + 1)) ++ + v->total_repeater_delay) * v->pixel_clock[k]; ++ v->v_ready_offset[k] = ++ dcn_bw_max2( ++ 150.0 ++ / (v->required_dispclk[i][j] ++ / (j ++ + 1)), ++ v->total_repeater_delay ++ + 20.0 ++ / v->projected_dcfclk_deep_sleep ++ + 10.0 ++ / (v->required_dispclk[i][j] ++ / (j ++ + 1))) ++ * v->pixel_clock[k]; ++ v->time_setup = (v->v_update_offset[k] + v->v_update_width[k] ++ + v->v_ready_offset[k]) / v->pixel_clock[k]; ++ v->extra_latency = ++ v->urgent_round_trip_and_out_of_order_latency_per_state[i] ++ + (v->total_number_of_active_dpp[i][j] ++ * v->pixel_chunk_size_in_kbyte ++ + v->total_number_of_dcc_active_dpp[i][j] ++ * v->meta_chunk_size) ++ * 1024.0 ++ / v->return_bw_per_state[i]; ++ if (v->pte_enable == dcn_bw_yes) { ++ v->extra_latency = v->extra_latency ++ + v->total_number_of_active_dpp[i][j] ++ * v->pte_chunk_size * 1024.0 ++ / v->return_bw_per_state[i]; ++ } ++ if (v->can_vstartup_lines_exceed_vsync_plus_back_porch_lines_minus_one ++ == dcn_bw_yes) { ++ v->maximum_vstartup = v->vtotal[k] - v->vactive[k] - 1.0; ++ } else { ++ v->maximum_vstartup = v->v_sync_plus_back_porch[k] - 1.0; ++ } ++ v->line_times_for_prefetch[k] = v->maximum_vstartup ++ - v->urgent_latency ++ / (v->htotal[k] / v->pixel_clock[k]) ++ - (v->time_calc + v->time_setup) ++ / (v->htotal[k] / v->pixel_clock[k]) ++ - (v->dst_y_after_scaler ++ + v->dst_x_after_scaler ++ / v->htotal[k]); ++ v->line_times_for_prefetch[k] = dcn_bw_floor2( ++ 4.0 * (v->line_times_for_prefetch[k] + 0.125), ++ 1.0) / 4; ++ v->prefetch_bw[k] = ++ (v->meta_pte_bytes_per_frame[k] ++ + 2.0 * v->meta_row_bytes[k] ++ + 2.0 * v->dpte_bytes_per_row[k] ++ + v->prefetch_lines_y[k] ++ * v->swath_width_yper_state[i][j][k] ++ * dcn_bw_ceil2( ++ v->byte_per_pixel_in_dety[k], ++ 1.0) ++ + v->prefetch_lines_c[k] ++ * v->swath_width_yper_state[i][j][k] ++ / 2.0 ++ * dcn_bw_ceil2( ++ v->byte_per_pixel_in_detc[k], ++ 2.0)) ++ / (v->line_times_for_prefetch[k] ++ * v->htotal[k] ++ / v->pixel_clock[k]); ++ } ++ v->bw_available_for_immediate_flip = v->return_bw_per_state[i]; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ v->bw_available_for_immediate_flip = ++ v->bw_available_for_immediate_flip ++ - dcn_bw_max2( ++ v->read_bandwidth[k], ++ v->prefetch_bw[k]); ++ } ++ v->total_immediate_flip_bytes[k] = 0.0; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if ((v->source_pixel_format[k] != dcn_bw_yuv420_sub_8 ++ && v->source_pixel_format[k] != dcn_bw_yuv420_sub_10)) { ++ v->total_immediate_flip_bytes[k] = ++ v->total_immediate_flip_bytes[k] ++ + v->meta_pte_bytes_per_frame[k] ++ + v->meta_row_bytes[k] ++ + v->dpte_bytes_per_row[k]; ++ } ++ } ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->pte_enable == dcn_bw_yes && v->dcc_enable[k] == dcn_bw_yes) { ++ v->time_for_meta_pte_with_immediate_flip = ++ dcn_bw_max5( ++ v->meta_pte_bytes_per_frame[k] ++ / v->prefetch_bw[k], ++ v->meta_pte_bytes_per_frame[k] ++ * v->total_immediate_flip_bytes[k] ++ / (v->bw_available_for_immediate_flip ++ * (v->meta_pte_bytes_per_frame[k] ++ + v->meta_row_bytes[k] ++ + v->dpte_bytes_per_row[k])), ++ v->extra_latency, ++ v->urgent_latency, ++ v->htotal[k] ++ / v->pixel_clock[k] ++ / 4.0); ++ v->time_for_meta_pte_without_immediate_flip = dcn_bw_max3( ++ v->meta_pte_bytes_per_frame[k] ++ / v->prefetch_bw[k], ++ v->extra_latency, ++ v->htotal[k] / v->pixel_clock[k] / 4.0); ++ } else { ++ v->time_for_meta_pte_with_immediate_flip = v->htotal[k] ++ / v->pixel_clock[k] / 4.0; ++ v->time_for_meta_pte_without_immediate_flip = v->htotal[k] ++ / v->pixel_clock[k] / 4.0; ++ } ++ if (v->pte_enable == dcn_bw_yes || v->dcc_enable[k] == dcn_bw_yes) { ++ v->time_for_meta_and_dpte_row_with_immediate_flip = ++ dcn_bw_max5( ++ (v->meta_row_bytes[k] ++ + v->dpte_bytes_per_row[k]) ++ / v->prefetch_bw[k], ++ (v->meta_row_bytes[k] ++ + v->dpte_bytes_per_row[k]) ++ * v->total_immediate_flip_bytes[k] ++ / (v->bw_available_for_immediate_flip ++ * (v->meta_pte_bytes_per_frame[k] ++ + v->meta_row_bytes[k] ++ + v->dpte_bytes_per_row[k])), ++ v->htotal[k] ++ / v->pixel_clock[k] ++ - v->time_for_meta_pte_with_immediate_flip, ++ v->extra_latency, ++ 2.0 * v->urgent_latency); ++ v->time_for_meta_and_dpte_row_without_immediate_flip = ++ dcn_bw_max3( ++ (v->meta_row_bytes[k] ++ + v->dpte_bytes_per_row[k]) ++ / v->prefetch_bw[k], ++ v->htotal[k] ++ / v->pixel_clock[k] ++ - v->time_for_meta_pte_without_immediate_flip, ++ v->extra_latency); ++ } else { ++ v->time_for_meta_and_dpte_row_with_immediate_flip = ++ dcn_bw_max2( ++ v->htotal[k] ++ / v->pixel_clock[k] ++ - v->time_for_meta_pte_with_immediate_flip, ++ v->extra_latency ++ - v->time_for_meta_pte_with_immediate_flip); ++ v->time_for_meta_and_dpte_row_without_immediate_flip = ++ dcn_bw_max2( ++ v->htotal[k] ++ / v->pixel_clock[k] ++ - v->time_for_meta_pte_without_immediate_flip, ++ v->extra_latency ++ - v->time_for_meta_pte_without_immediate_flip); ++ } ++ v->lines_for_meta_pte_with_immediate_flip[k] = ++ dcn_bw_floor2( ++ 4.0 ++ * (v->time_for_meta_pte_with_immediate_flip ++ / (v->htotal[k] ++ / v->pixel_clock[k]) ++ + 0.125), ++ 1.0) / 4; ++ v->lines_for_meta_pte_without_immediate_flip[k] = ++ dcn_bw_floor2( ++ 4.0 ++ * (v->time_for_meta_pte_without_immediate_flip ++ / (v->htotal[k] ++ / v->pixel_clock[k]) ++ + 0.125), ++ 1.0) / 4; ++ v->lines_for_meta_and_dpte_row_with_immediate_flip[k] = ++ dcn_bw_floor2( ++ 4.0 ++ * (v->time_for_meta_and_dpte_row_with_immediate_flip ++ / (v->htotal[k] ++ / v->pixel_clock[k]) ++ + 0.125), ++ 1.0) / 4; ++ v->lines_for_meta_and_dpte_row_without_immediate_flip[k] = ++ dcn_bw_floor2( ++ 4.0 ++ * (v->time_for_meta_and_dpte_row_without_immediate_flip ++ / (v->htotal[k] ++ / v->pixel_clock[k]) ++ + 0.125), ++ 1.0) / 4; ++ v->line_times_to_request_prefetch_pixel_data_with_immediate_flip = ++ v->line_times_for_prefetch[k] ++ - v->lines_for_meta_pte_with_immediate_flip[k] ++ - v->lines_for_meta_and_dpte_row_with_immediate_flip[k]; ++ v->line_times_to_request_prefetch_pixel_data_without_immediate_flip = ++ v->line_times_for_prefetch[k] ++ - v->lines_for_meta_pte_without_immediate_flip[k] ++ - v->lines_for_meta_and_dpte_row_without_immediate_flip[k]; ++ if (v->line_times_to_request_prefetch_pixel_data_with_immediate_flip ++ > 0.0) { ++ v->v_ratio_pre_ywith_immediate_flip[i][j][k] = ++ v->prefetch_lines_y[k] ++ / v->line_times_to_request_prefetch_pixel_data_with_immediate_flip; ++ if ((v->swath_height_yper_state[i][j][k] > 4.0)) { ++ if (v->line_times_to_request_prefetch_pixel_data_with_immediate_flip ++ - (v->prefill_y[k] - 3.0) / 2.0 ++ > 0.0) { ++ v->v_ratio_pre_ywith_immediate_flip[i][j][k] = ++ dcn_bw_max2( ++ v->v_ratio_pre_ywith_immediate_flip[i][j][k], ++ (v->max_num_sw_y[k] ++ * v->swath_height_yper_state[i][j][k]) ++ / (v->line_times_to_request_prefetch_pixel_data_with_immediate_flip ++ - (v->prefill_y[k] ++ - 3.0) ++ / 2.0)); ++ } else { ++ v->v_ratio_pre_ywith_immediate_flip[i][j][k] = ++ 999999.0; ++ } ++ } ++ v->v_ratio_pre_cwith_immediate_flip[i][j][k] = ++ v->prefetch_lines_c[k] ++ / v->line_times_to_request_prefetch_pixel_data_with_immediate_flip; ++ if ((v->swath_height_cper_state[i][j][k] > 4.0)) { ++ if (v->line_times_to_request_prefetch_pixel_data_with_immediate_flip ++ - (v->prefill_c[k] - 3.0) / 2.0 ++ > 0.0) { ++ v->v_ratio_pre_cwith_immediate_flip[i][j][k] = ++ dcn_bw_max2( ++ v->v_ratio_pre_cwith_immediate_flip[i][j][k], ++ (v->max_num_sw_c[k] ++ * v->swath_height_cper_state[i][j][k]) ++ / (v->line_times_to_request_prefetch_pixel_data_with_immediate_flip ++ - (v->prefill_c[k] ++ - 3.0) ++ / 2.0)); ++ } else { ++ v->v_ratio_pre_cwith_immediate_flip[i][j][k] = ++ 999999.0; ++ } ++ } ++ v->required_prefetch_pixel_data_bw_with_immediate_flip[i][j][k] = ++ v->no_of_dpp[i][j][k] ++ * (v->prefetch_lines_y[k] ++ / v->line_times_to_request_prefetch_pixel_data_with_immediate_flip ++ * dcn_bw_ceil2( ++ v->byte_per_pixel_in_dety[k], ++ 1.0) ++ + v->prefetch_lines_c[k] ++ / v->line_times_to_request_prefetch_pixel_data_with_immediate_flip ++ * dcn_bw_ceil2( ++ v->byte_per_pixel_in_detc[k], ++ 2.0) ++ / 2.0) ++ * v->swath_width_yper_state[i][j][k] ++ / (v->htotal[k] ++ / v->pixel_clock[k]); ++ } else { ++ v->v_ratio_pre_ywith_immediate_flip[i][j][k] = 999999.0; ++ v->v_ratio_pre_cwith_immediate_flip[i][j][k] = 999999.0; ++ v->required_prefetch_pixel_data_bw_with_immediate_flip[i][j][k] = ++ 999999.0; ++ } ++ if (v->line_times_to_request_prefetch_pixel_data_without_immediate_flip ++ > 0.0) { ++ v->v_ratio_pre_ywithout_immediate_flip[i][j][k] = ++ v->prefetch_lines_y[k] ++ / v->line_times_to_request_prefetch_pixel_data_without_immediate_flip; ++ if ((v->swath_height_yper_state[i][j][k] > 4.0)) { ++ if (v->line_times_to_request_prefetch_pixel_data_without_immediate_flip ++ - (v->prefill_y[k] - 3.0) / 2.0 ++ > 0.0) { ++ v->v_ratio_pre_ywithout_immediate_flip[i][j][k] = ++ dcn_bw_max2( ++ v->v_ratio_pre_ywithout_immediate_flip[i][j][k], ++ (v->max_num_sw_y[k] ++ * v->swath_height_yper_state[i][j][k]) ++ / (v->line_times_to_request_prefetch_pixel_data_without_immediate_flip ++ - (v->prefill_y[k] ++ - 3.0) ++ / 2.0)); ++ } else { ++ v->v_ratio_pre_ywithout_immediate_flip[i][j][k] = ++ 999999.0; ++ } ++ } ++ v->v_ratio_pre_cwithout_immediate_flip[i][j][k] = ++ v->prefetch_lines_c[k] ++ / v->line_times_to_request_prefetch_pixel_data_without_immediate_flip; ++ if ((v->swath_height_cper_state[i][j][k] > 4.0)) { ++ if (v->line_times_to_request_prefetch_pixel_data_without_immediate_flip ++ - (v->prefill_c[k] - 3.0) / 2.0 ++ > 0.0) { ++ v->v_ratio_pre_cwithout_immediate_flip[i][j][k] = ++ dcn_bw_max2( ++ v->v_ratio_pre_cwithout_immediate_flip[i][j][k], ++ (v->max_num_sw_c[k] ++ * v->swath_height_cper_state[i][j][k]) ++ / (v->line_times_to_request_prefetch_pixel_data_without_immediate_flip ++ - (v->prefill_c[k] ++ - 3.0) ++ / 2.0)); ++ } else { ++ v->v_ratio_pre_cwithout_immediate_flip[i][j][k] = ++ 999999.0; ++ } ++ } ++ v->required_prefetch_pixel_data_bw_without_immediate_flip[i][j][k] = ++ v->no_of_dpp[i][j][k] ++ * (v->prefetch_lines_y[k] ++ / v->line_times_to_request_prefetch_pixel_data_without_immediate_flip ++ * dcn_bw_ceil2( ++ v->byte_per_pixel_in_dety[k], ++ 1.0) ++ + v->prefetch_lines_c[k] ++ / v->line_times_to_request_prefetch_pixel_data_without_immediate_flip ++ * dcn_bw_ceil2( ++ v->byte_per_pixel_in_detc[k], ++ 2.0) ++ / 2.0) ++ * v->swath_width_yper_state[i][j][k] ++ / (v->htotal[k] ++ / v->pixel_clock[k]); ++ } else { ++ v->v_ratio_pre_ywithout_immediate_flip[i][j][k] = 999999.0; ++ v->v_ratio_pre_cwithout_immediate_flip[i][j][k] = 999999.0; ++ v->required_prefetch_pixel_data_bw_without_immediate_flip[i][j][k] = ++ 999999.0; ++ } ++ } ++ v->maximum_read_bandwidth_with_prefetch_with_immediate_flip = 0.0; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if ((v->source_pixel_format[k] != dcn_bw_yuv420_sub_8 ++ && v->source_pixel_format[k] != dcn_bw_yuv420_sub_10)) { ++ v->maximum_read_bandwidth_with_prefetch_with_immediate_flip = ++ v->maximum_read_bandwidth_with_prefetch_with_immediate_flip ++ + dcn_bw_max2( ++ v->read_bandwidth[k], ++ v->required_prefetch_pixel_data_bw_with_immediate_flip[i][j][k]) ++ + dcn_bw_max2( ++ v->meta_pte_bytes_per_frame[k] ++ / (v->lines_for_meta_pte_with_immediate_flip[k] ++ * v->htotal[k] ++ / v->pixel_clock[k]), ++ (v->meta_row_bytes[k] ++ + v->dpte_bytes_per_row[k]) ++ / (v->lines_for_meta_and_dpte_row_with_immediate_flip[k] ++ * v->htotal[k] ++ / v->pixel_clock[k])); ++ } else { ++ v->maximum_read_bandwidth_with_prefetch_with_immediate_flip = ++ v->maximum_read_bandwidth_with_prefetch_with_immediate_flip ++ + dcn_bw_max2( ++ v->read_bandwidth[k], ++ v->required_prefetch_pixel_data_bw_without_immediate_flip[i][j][k]); ++ } ++ } ++ v->maximum_read_bandwidth_with_prefetch_without_immediate_flip = 0.0; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ v->maximum_read_bandwidth_with_prefetch_without_immediate_flip = ++ v->maximum_read_bandwidth_with_prefetch_without_immediate_flip ++ + dcn_bw_max2( ++ v->read_bandwidth[k], ++ v->required_prefetch_pixel_data_bw_without_immediate_flip[i][j][k]); ++ } ++ v->prefetch_supported_with_immediate_flip[i][j] = dcn_bw_yes; ++ if (v->maximum_read_bandwidth_with_prefetch_with_immediate_flip ++ > v->return_bw_per_state[i]) { ++ v->prefetch_supported_with_immediate_flip[i][j] = dcn_bw_no; ++ } ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->line_times_for_prefetch[k] < 2.0 ++ || v->lines_for_meta_pte_with_immediate_flip[k] ++ >= 8.0 ++ || v->lines_for_meta_and_dpte_row_with_immediate_flip[k] ++ >= 16.0) { ++ v->prefetch_supported_with_immediate_flip[i][j] = dcn_bw_no; ++ } ++ } ++ v->prefetch_supported_without_immediate_flip[i][j] = dcn_bw_yes; ++ if (v->maximum_read_bandwidth_with_prefetch_without_immediate_flip ++ > v->return_bw_per_state[i]) { ++ v->prefetch_supported_without_immediate_flip[i][j] = dcn_bw_no; ++ } ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->line_times_for_prefetch[k] < 2.0 ++ || v->lines_for_meta_pte_without_immediate_flip[k] ++ >= 8.0 ++ || v->lines_for_meta_and_dpte_row_without_immediate_flip[k] ++ >= 16.0) { ++ v->prefetch_supported_without_immediate_flip[i][j] = ++ dcn_bw_no; ++ } ++ } ++ } ++ } ++ for (i = 0; i <= number_of_states_plus_one; i++) { ++ for (j = 0; j <= 1; j++) { ++ v->v_ratio_in_prefetch_supported_with_immediate_flip[i][j] = dcn_bw_yes; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if ((((v->source_pixel_format[k] != dcn_bw_yuv420_sub_8 ++ && v->source_pixel_format[k] != dcn_bw_yuv420_sub_10) ++ && (v->v_ratio_pre_ywith_immediate_flip[i][j][k] ++ > 4.0 ++ || v->v_ratio_pre_cwith_immediate_flip[i][j][k] ++ > 4.0)) ++ || ((v->source_pixel_format[k] ++ == dcn_bw_yuv420_sub_8 ++ || v->source_pixel_format[k] ++ == dcn_bw_yuv420_sub_10) ++ && (v->v_ratio_pre_ywithout_immediate_flip[i][j][k] ++ > 4.0 ++ || v->v_ratio_pre_cwithout_immediate_flip[i][j][k] ++ > 4.0)))) { ++ v->v_ratio_in_prefetch_supported_with_immediate_flip[i][j] = ++ dcn_bw_no; ++ } ++ } ++ v->v_ratio_in_prefetch_supported_without_immediate_flip[i][j] = dcn_bw_yes; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if ((v->v_ratio_pre_ywithout_immediate_flip[i][j][k] > 4.0 ++ || v->v_ratio_pre_cwithout_immediate_flip[i][j][k] ++ > 4.0)) { ++ v->v_ratio_in_prefetch_supported_without_immediate_flip[i][j] = ++ dcn_bw_no; ++ } ++ } ++ } ++ } ++ /*mode support, voltage state and soc configuration*/ ++ ++ for (i = number_of_states_plus_one; i >= 0; i--) { ++ for (j = 0; j <= 1; j++) { ++ if (v->scale_ratio_support == dcn_bw_yes ++ && v->source_format_pixel_and_scan_support == dcn_bw_yes ++ && v->viewport_size_support == dcn_bw_yes ++ && v->bandwidth_support[i] == dcn_bw_yes ++ && v->dio_support[i] == dcn_bw_yes ++ && v->urgent_latency_support[i][j] == dcn_bw_yes ++ && v->rob_support[i] == dcn_bw_yes ++ && v->dispclk_dppclk_support[i][j] == dcn_bw_yes ++ && v->total_available_pipes_support[i][j] == dcn_bw_yes ++ && v->total_available_writeback_support == dcn_bw_yes ++ && v->writeback_latency_support == dcn_bw_yes) { ++ if (v->prefetch_supported_with_immediate_flip[i][j] == dcn_bw_yes ++ && v->v_ratio_in_prefetch_supported_with_immediate_flip[i][j] ++ == dcn_bw_yes) { ++ v->mode_support_with_immediate_flip[i][j] = dcn_bw_yes; ++ } else { ++ v->mode_support_with_immediate_flip[i][j] = dcn_bw_no; ++ } ++ if (v->prefetch_supported_without_immediate_flip[i][j] == dcn_bw_yes ++ && v->v_ratio_in_prefetch_supported_without_immediate_flip[i][j] ++ == dcn_bw_yes) { ++ v->mode_support_without_immediate_flip[i][j] = dcn_bw_yes; ++ } else { ++ v->mode_support_without_immediate_flip[i][j] = dcn_bw_no; ++ } ++ } else { ++ v->mode_support_with_immediate_flip[i][j] = dcn_bw_no; ++ v->mode_support_without_immediate_flip[i][j] = dcn_bw_no; ++ } ++ } ++ } ++ for (i = number_of_states_plus_one; i >= 0; i--) { ++ if ((i == number_of_states_plus_one ++ || v->mode_support_with_immediate_flip[i][1] == dcn_bw_yes ++ || v->mode_support_with_immediate_flip[i][0] == dcn_bw_yes) ++ && i >= v->voltage_override_level) { ++ v->voltage_level_with_immediate_flip = i; ++ } ++ } ++ for (i = number_of_states_plus_one; i >= 0; i--) { ++ if ((i == number_of_states_plus_one ++ || v->mode_support_without_immediate_flip[i][1] == dcn_bw_yes ++ || v->mode_support_without_immediate_flip[i][0] == dcn_bw_yes) ++ && i >= v->voltage_override_level) { ++ v->voltage_level_without_immediate_flip = i; ++ } ++ } ++ if (v->voltage_level_with_immediate_flip == number_of_states_plus_one) { ++ v->immediate_flip_supported = dcn_bw_no; ++ v->voltage_level = v->voltage_level_without_immediate_flip; ++ } else { ++ v->immediate_flip_supported = dcn_bw_yes; ++ v->voltage_level = v->voltage_level_with_immediate_flip; ++ } ++ v->dcfclk = v->dcfclk_per_state[v->voltage_level]; ++ v->fabric_and_dram_bandwidth = v->fabric_and_dram_bandwidth_per_state[v->voltage_level]; ++ for (j = 0; j <= 1; j++) { ++ v->required_dispclk_per_ratio[j] = v->required_dispclk[v->voltage_level][j]; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ v->dpp_per_plane_per_ratio[j][k] = v->no_of_dpp[v->voltage_level][j][k]; ++ } ++ v->dispclk_dppclk_support_per_ratio[j] = ++ v->dispclk_dppclk_support[v->voltage_level][j]; ++ } ++ v->max_phyclk = v->phyclk_per_state[v->voltage_level]; ++} ++void display_pipe_configuration(struct dcn_bw_internal_vars *v) ++{ ++ int j, k; ++ /*display pipe configuration*/ ++ ++ for (j = 0; j <= 1; j++) { ++ v->total_number_of_active_dpp_per_ratio[j] = 0.0; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ v->total_number_of_active_dpp_per_ratio[j] = ++ v->total_number_of_active_dpp_per_ratio[j] ++ + v->dpp_per_plane_per_ratio[j][k]; ++ } ++ } ++ if ((v->dispclk_dppclk_support_per_ratio[0] == dcn_bw_yes ++ && v->dispclk_dppclk_support_per_ratio[1] == dcn_bw_no) ++ || (v->dispclk_dppclk_support_per_ratio[0] ++ == v->dispclk_dppclk_support_per_ratio[1] ++ && (v->total_number_of_active_dpp_per_ratio[0] ++ < v->total_number_of_active_dpp_per_ratio[1] ++ || (((v->total_number_of_active_dpp_per_ratio[0] ++ == v->total_number_of_active_dpp_per_ratio[1]) ++ && v->required_dispclk_per_ratio[0] ++ <= 0.5 ++ * v->required_dispclk_per_ratio[1]))))) { ++ v->dispclk_dppclk_ratio = 1; ++ v->final_error_message = v->error_message[0]; ++ } else { ++ v->dispclk_dppclk_ratio = 2; ++ v->final_error_message = v->error_message[1]; ++ } ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ v->dpp_per_plane[k] = v->dpp_per_plane_per_ratio[v->dispclk_dppclk_ratio - 1][k]; ++ } ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->source_pixel_format[k] == dcn_bw_rgb_sub_64) { ++ v->byte_per_pix_dety = 8.0; ++ v->byte_per_pix_detc = 0.0; ++ } else if (v->source_pixel_format[k] == dcn_bw_rgb_sub_32) { ++ v->byte_per_pix_dety = 4.0; ++ v->byte_per_pix_detc = 0.0; ++ } else if (v->source_pixel_format[k] == dcn_bw_rgb_sub_16) { ++ v->byte_per_pix_dety = 2.0; ++ v->byte_per_pix_detc = 0.0; ++ } else if (v->source_pixel_format[k] == dcn_bw_yuv420_sub_8) { ++ v->byte_per_pix_dety = 1.0; ++ v->byte_per_pix_detc = 2.0; ++ } else { ++ v->byte_per_pix_dety = 4.0f / 3; ++ v->byte_per_pix_detc = 8.0f / 3; ++ } ++ if ((v->source_pixel_format[k] == dcn_bw_rgb_sub_64 ++ || v->source_pixel_format[k] == dcn_bw_rgb_sub_32 ++ || v->source_pixel_format[k] == dcn_bw_rgb_sub_16)) { ++ if (v->source_surface_mode[k] == dcn_bw_sw_linear) { ++ v->read256_bytes_block_height_y = 1.0; ++ } else if (v->source_pixel_format[k] == dcn_bw_rgb_sub_64) { ++ v->read256_bytes_block_height_y = 4.0; ++ } else { ++ v->read256_bytes_block_height_y = 8.0; ++ } ++ v->read256_bytes_block_width_y = 256.0 / dcn_bw_ceil2(v->byte_per_pix_dety, 1.0) ++ / v->read256_bytes_block_height_y; ++ v->read256_bytes_block_height_c = 0.0; ++ v->read256_bytes_block_width_c = 0.0; ++ } else { ++ if (v->source_surface_mode[k] == dcn_bw_sw_linear) { ++ v->read256_bytes_block_height_y = 1.0; ++ v->read256_bytes_block_height_c = 1.0; ++ } else if (v->source_pixel_format[k] == dcn_bw_yuv420_sub_8) { ++ v->read256_bytes_block_height_y = 16.0; ++ v->read256_bytes_block_height_c = 8.0; ++ } else { ++ v->read256_bytes_block_height_y = 8.0; ++ v->read256_bytes_block_height_c = 8.0; ++ } ++ v->read256_bytes_block_width_y = 256.0 / dcn_bw_ceil2(v->byte_per_pix_dety, 1.0) ++ / v->read256_bytes_block_height_y; ++ v->read256_bytes_block_width_c = 256.0 / dcn_bw_ceil2(v->byte_per_pix_detc, 2.0) ++ / v->read256_bytes_block_height_c; ++ } ++ if (v->source_scan[k] == dcn_bw_hor) { ++ v->maximum_swath_height_y = v->read256_bytes_block_height_y; ++ v->maximum_swath_height_c = v->read256_bytes_block_height_c; ++ } else { ++ v->maximum_swath_height_y = v->read256_bytes_block_width_y; ++ v->maximum_swath_height_c = v->read256_bytes_block_width_c; ++ } ++ if ((v->source_pixel_format[k] == dcn_bw_rgb_sub_64 ++ || v->source_pixel_format[k] == dcn_bw_rgb_sub_32 ++ || v->source_pixel_format[k] == dcn_bw_rgb_sub_16)) { ++ if (v->source_surface_mode[k] == dcn_bw_sw_linear ++ || (v->source_pixel_format[k] == dcn_bw_rgb_sub_64 ++ && (v->source_surface_mode[k] ++ == dcn_bw_sw_4_kb_s ++ || v->source_surface_mode[k] ++ == dcn_bw_sw_4_kb_s_x ++ || v->source_surface_mode[k] ++ == dcn_bw_sw_64_kb_s ++ || v->source_surface_mode[k] ++ == dcn_bw_sw_64_kb_s_t ++ || v->source_surface_mode[k] ++ == dcn_bw_sw_64_kb_s_x ++ || v->source_surface_mode[k] ++ == dcn_bw_sw_var_s ++ || v->source_surface_mode[k] ++ == dcn_bw_sw_var_s_x) ++ && v->source_scan[k] == dcn_bw_hor)) { ++ v->minimum_swath_height_y = v->maximum_swath_height_y; ++ } else { ++ v->minimum_swath_height_y = v->maximum_swath_height_y / 2.0; ++ } ++ v->minimum_swath_height_c = v->maximum_swath_height_c; ++ } else { ++ if (v->source_surface_mode[k] == dcn_bw_sw_linear) { ++ v->minimum_swath_height_y = v->maximum_swath_height_y; ++ v->minimum_swath_height_c = v->maximum_swath_height_c; ++ } else if (v->source_pixel_format[k] == dcn_bw_yuv420_sub_8 ++ && v->source_scan[k] == dcn_bw_hor) { ++ v->minimum_swath_height_y = v->maximum_swath_height_y / 2.0; ++ if (v->bug_forcing_luma_and_chroma_request_to_same_size_fixed ++ == dcn_bw_yes) { ++ v->minimum_swath_height_c = v->maximum_swath_height_c; ++ } else { ++ v->minimum_swath_height_c = v->maximum_swath_height_c / 2.0; ++ } ++ } else if (v->source_pixel_format[k] == dcn_bw_yuv420_sub_10 ++ && v->source_scan[k] == dcn_bw_hor) { ++ v->minimum_swath_height_c = v->maximum_swath_height_c / 2.0; ++ if (v->bug_forcing_luma_and_chroma_request_to_same_size_fixed ++ == dcn_bw_yes) { ++ v->minimum_swath_height_y = v->maximum_swath_height_y; ++ } else { ++ v->minimum_swath_height_y = v->maximum_swath_height_y / 2.0; ++ } ++ } else { ++ v->minimum_swath_height_y = v->maximum_swath_height_y; ++ v->minimum_swath_height_c = v->maximum_swath_height_c; ++ } ++ } ++ if (v->source_scan[k] == dcn_bw_hor) { ++ v->swath_width = v->viewport_width[k] / v->dpp_per_plane[k]; ++ } else { ++ v->swath_width = v->viewport_height[k] / v->dpp_per_plane[k]; ++ } ++ v->swath_width_granularity_y = 256.0 / dcn_bw_ceil2(v->byte_per_pix_dety, 1.0) ++ / v->maximum_swath_height_y; ++ v->rounded_up_max_swath_size_bytes_y = (dcn_bw_ceil2( ++ v->swath_width - 1.0, ++ v->swath_width_granularity_y) + v->swath_width_granularity_y) ++ * v->byte_per_pix_dety * v->maximum_swath_height_y; ++ if (v->source_pixel_format[k] == dcn_bw_yuv420_sub_10) { ++ v->rounded_up_max_swath_size_bytes_y = dcn_bw_ceil2( ++ v->rounded_up_max_swath_size_bytes_y, ++ 256.0) + 256; ++ } ++ if (v->maximum_swath_height_c > 0.0) { ++ v->swath_width_granularity_c = 256.0 / dcn_bw_ceil2(v->byte_per_pix_detc, 2.0) ++ / v->maximum_swath_height_c; ++ } ++ v->rounded_up_max_swath_size_bytes_c = (dcn_bw_ceil2( ++ v->swath_width / 2.0 - 1.0, ++ v->swath_width_granularity_c) + v->swath_width_granularity_c) ++ * v->byte_per_pix_detc * v->maximum_swath_height_c; ++ if (v->source_pixel_format[k] == dcn_bw_yuv420_sub_10) { ++ v->rounded_up_max_swath_size_bytes_c = dcn_bw_ceil2( ++ v->rounded_up_max_swath_size_bytes_c, ++ 256.0) + 256; ++ } ++ if (v->rounded_up_max_swath_size_bytes_y + v->rounded_up_max_swath_size_bytes_c ++ <= v->det_buffer_size_in_kbyte * 1024.0 / 2.0) { ++ v->swath_height_y[k] = v->maximum_swath_height_y; ++ v->swath_height_c[k] = v->maximum_swath_height_c; ++ } else { ++ v->swath_height_y[k] = v->minimum_swath_height_y; ++ v->swath_height_c[k] = v->minimum_swath_height_c; ++ } ++ if (v->swath_height_c[k] == 0.0) { ++ v->det_buffer_size_y[k] = v->det_buffer_size_in_kbyte * 1024.0; ++ v->det_buffer_size_c[k] = 0.0; ++ } else if (v->swath_height_y[k] <= v->swath_height_c[k]) { ++ v->det_buffer_size_y[k] = v->det_buffer_size_in_kbyte * 1024.0 / 2.0; ++ v->det_buffer_size_c[k] = v->det_buffer_size_in_kbyte * 1024.0 / 2.0; ++ } else { ++ v->det_buffer_size_y[k] = v->det_buffer_size_in_kbyte * 1024.0 * 2.0 / 3.0; ++ v->det_buffer_size_c[k] = v->det_buffer_size_in_kbyte * 1024.0 / 3.0; ++ } ++ } ++} ++void dispclkdppclkdcfclk_deep_sleep_prefetch_parameters_watermarks_and_performance_calculation( ++ struct dcn_bw_internal_vars *v) ++{ ++ int k; ++ /*dispclk and dppclk calculation*/ ++ ++ v->dispclk_with_ramping = 0.0; ++ v->dispclk_without_ramping = 0.0; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->h_ratio[k] > 1.0) { ++ v->pscl_throughput[k] = dcn_bw_min2( ++ v->max_dchub_topscl_throughput, ++ v->max_pscl_tolb_throughput * v->h_ratio[k] ++ / dcn_bw_ceil2(v->htaps[k] / 6.0, 1.0)); ++ } else { ++ v->pscl_throughput[k] = dcn_bw_min2( ++ v->max_dchub_topscl_throughput, ++ v->max_pscl_tolb_throughput); ++ } ++ v->dppclk_using_single_dpp_luma = v->pixel_clock[k] ++ * dcn_bw_max3( ++ v->vtaps[k] / 6.0 * dcn_bw_min2(1.0, v->h_ratio[k]), ++ v->h_ratio[k] * v->v_ratio[k] ++ / v->pscl_throughput[k], ++ 1.0); ++ if ((v->source_pixel_format[k] != dcn_bw_yuv420_sub_8 ++ && v->source_pixel_format[k] != dcn_bw_yuv420_sub_10)) { ++ v->pscl_throughput_chroma[k] = 0.0; ++ v->dppclk_using_single_dpp = v->dppclk_using_single_dpp_luma; ++ } else { ++ if (v->h_ratio[k] > 1.0) { ++ v->pscl_throughput_chroma[k] = ++ dcn_bw_min2( ++ v->max_dchub_topscl_throughput, ++ v->max_pscl_tolb_throughput ++ * v->h_ratio[k] ++ / 2.0 ++ / dcn_bw_ceil2( ++ v->hta_pschroma[k] ++ / 6.0, ++ 1.0)); ++ } else { ++ v->pscl_throughput_chroma[k] = dcn_bw_min2( ++ v->max_dchub_topscl_throughput, ++ v->max_pscl_tolb_throughput); ++ } ++ v->dppclk_using_single_dpp_chroma = ++ v->pixel_clock[k] ++ * dcn_bw_max3( ++ v->vta_pschroma[k] / 6.0 ++ * dcn_bw_min2( ++ 1.0, ++ v->h_ratio[k] ++ / 2.0), ++ v->h_ratio[k] ++ * v->v_ratio[k] ++ / 4.0 ++ / v->pscl_throughput_chroma[k], ++ 1.0); ++ v->dppclk_using_single_dpp = dcn_bw_max2( ++ v->dppclk_using_single_dpp_luma, ++ v->dppclk_using_single_dpp_chroma); ++ } ++ if (v->odm_capable == dcn_bw_yes) { ++ v->dispclk_with_ramping = ++ dcn_bw_max2( ++ v->dispclk_with_ramping, ++ dcn_bw_max2( ++ v->dppclk_using_single_dpp ++ / v->dpp_per_plane[k] ++ * v->dispclk_dppclk_ratio, ++ v->pixel_clock[k] ++ / v->dpp_per_plane[k]) ++ * (1.0 ++ + v->downspreading ++ / 100.0) ++ * (1.0 ++ + v->dispclk_ramping_margin ++ / 100.0)); ++ v->dispclk_without_ramping = dcn_bw_max2( ++ v->dispclk_without_ramping, ++ dcn_bw_max2( ++ v->dppclk_using_single_dpp ++ / v->dpp_per_plane[k] ++ * v->dispclk_dppclk_ratio, ++ v->pixel_clock[k] / v->dpp_per_plane[k]) ++ * (1.0 + v->downspreading / 100.0)); ++ } else { ++ v->dispclk_with_ramping = ++ dcn_bw_max2( ++ v->dispclk_with_ramping, ++ dcn_bw_max2( ++ v->dppclk_using_single_dpp ++ / v->dpp_per_plane[k] ++ * v->dispclk_dppclk_ratio, ++ v->pixel_clock[k]) ++ * (1.0 ++ + v->downspreading ++ / 100.0) ++ * (1.0 ++ + v->dispclk_ramping_margin ++ / 100.0)); ++ v->dispclk_without_ramping = dcn_bw_max2( ++ v->dispclk_without_ramping, ++ dcn_bw_max2( ++ v->dppclk_using_single_dpp ++ / v->dpp_per_plane[k] ++ * v->dispclk_dppclk_ratio, ++ v->pixel_clock[k]) ++ * (1.0 + v->downspreading / 100.0)); ++ } ++ } ++ if (v->dispclk_without_ramping > v->max_dispclk[number_of_states]) { ++ v->dispclk = v->dispclk_without_ramping; ++ } else if (v->dispclk_with_ramping > v->max_dispclk[number_of_states]) { ++ v->dispclk = v->max_dispclk[number_of_states]; ++ } else { ++ v->dispclk = v->dispclk_with_ramping; ++ } ++ v->dppclk = v->dispclk / v->dispclk_dppclk_ratio; ++ /*urgent watermark*/ ++ ++ v->return_bandwidth_to_dcn = dcn_bw_min2( ++ v->return_bus_width * v->dcfclk, ++ v->fabric_and_dram_bandwidth * 1000.0 ++ * v->percent_of_ideal_drambw_received_after_urg_latency ++ / 100.0); ++ v->dcc_enabled_any_plane = dcn_bw_no; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->dcc_enable[k] == dcn_bw_yes) { ++ v->dcc_enabled_any_plane = dcn_bw_yes; ++ } ++ } ++ v->return_bw = v->return_bandwidth_to_dcn; ++ if (v->dcc_enabled_any_plane == dcn_bw_yes ++ && v->return_bandwidth_to_dcn > v->dcfclk * v->return_bus_width / 4.0) { ++ v->return_bw = ++ dcn_bw_min2( ++ v->return_bw, ++ v->return_bandwidth_to_dcn * 4.0 ++ * (1.0 ++ - v->urgent_latency ++ / ((v->rob_buffer_size_in_kbyte ++ - v->pixel_chunk_size_in_kbyte) ++ * 1024.0 ++ / (v->return_bandwidth_to_dcn ++ - v->dcfclk ++ * v->return_bus_width ++ / 4.0) ++ + v->urgent_latency))); ++ } ++ v->critical_compression = 2.0 * v->return_bus_width * v->dcfclk * v->urgent_latency ++ / (v->return_bandwidth_to_dcn * v->urgent_latency ++ + (v->rob_buffer_size_in_kbyte ++ - v->pixel_chunk_size_in_kbyte) * 1024.0); ++ if (v->dcc_enabled_any_plane == dcn_bw_yes && v->critical_compression > 1.0 ++ && v->critical_compression < 4.0) { ++ v->return_bw = ++ dcn_bw_min2( ++ v->return_bw, ++ dcn_bw_pow( ++ 4.0 * v->return_bandwidth_to_dcn ++ * (v->rob_buffer_size_in_kbyte ++ - v->pixel_chunk_size_in_kbyte) ++ * 1024.0 ++ * v->return_bus_width ++ * v->dcfclk ++ * v->urgent_latency ++ / (v->return_bandwidth_to_dcn ++ * v->urgent_latency ++ + (v->rob_buffer_size_in_kbyte ++ - v->pixel_chunk_size_in_kbyte) ++ * 1024.0), ++ 2)); ++ } ++ v->return_bandwidth_to_dcn = dcn_bw_min2( ++ v->return_bus_width * v->dcfclk, ++ v->fabric_and_dram_bandwidth * 1000.0); ++ if (v->dcc_enabled_any_plane == dcn_bw_yes ++ && v->return_bandwidth_to_dcn > v->dcfclk * v->return_bus_width / 4.0) { ++ v->return_bw = ++ dcn_bw_min2( ++ v->return_bw, ++ v->return_bandwidth_to_dcn * 4.0 ++ * (1.0 ++ - v->urgent_latency ++ / ((v->rob_buffer_size_in_kbyte ++ - v->pixel_chunk_size_in_kbyte) ++ * 1024.0 ++ / (v->return_bandwidth_to_dcn ++ - v->dcfclk ++ * v->return_bus_width ++ / 4.0) ++ + v->urgent_latency))); ++ } ++ v->critical_compression = 2.0 * v->return_bus_width * v->dcfclk * v->urgent_latency ++ / (v->return_bandwidth_to_dcn * v->urgent_latency ++ + (v->rob_buffer_size_in_kbyte ++ - v->pixel_chunk_size_in_kbyte) * 1024.0); ++ if (v->dcc_enabled_any_plane == dcn_bw_yes && v->critical_compression > 1.0 ++ && v->critical_compression < 4.0) { ++ v->return_bw = ++ dcn_bw_min2( ++ v->return_bw, ++ dcn_bw_pow( ++ 4.0 * v->return_bandwidth_to_dcn ++ * (v->rob_buffer_size_in_kbyte ++ - v->pixel_chunk_size_in_kbyte) ++ * 1024.0 ++ * v->return_bus_width ++ * v->dcfclk ++ * v->urgent_latency ++ / (v->return_bandwidth_to_dcn ++ * v->urgent_latency ++ + (v->rob_buffer_size_in_kbyte ++ - v->pixel_chunk_size_in_kbyte) ++ * 1024.0), ++ 2)); ++ } ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->source_scan[k] == dcn_bw_hor) { ++ v->swath_width_y[k] = v->viewport_width[k] / v->dpp_per_plane[k]; ++ } else { ++ v->swath_width_y[k] = v->viewport_height[k] / v->dpp_per_plane[k]; ++ } ++ } ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->source_pixel_format[k] == dcn_bw_rgb_sub_64) { ++ v->byte_per_pixel_dety[k] = 8.0; ++ v->byte_per_pixel_detc[k] = 0.0; ++ } else if (v->source_pixel_format[k] == dcn_bw_rgb_sub_32) { ++ v->byte_per_pixel_dety[k] = 4.0; ++ v->byte_per_pixel_detc[k] = 0.0; ++ } else if (v->source_pixel_format[k] == dcn_bw_rgb_sub_16) { ++ v->byte_per_pixel_dety[k] = 2.0; ++ v->byte_per_pixel_detc[k] = 0.0; ++ } else if (v->source_pixel_format[k] == dcn_bw_yuv420_sub_8) { ++ v->byte_per_pixel_dety[k] = 1.0; ++ v->byte_per_pixel_detc[k] = 2.0; ++ } else { ++ v->byte_per_pixel_dety[k] = 4.0f / 3; ++ v->byte_per_pixel_detc[k] = 8.0f / 3; ++ } ++ } ++ v->total_data_read_bandwidth = 0.0; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ v->read_bandwidth_plane_luma[k] = v->swath_width_y[k] * v->dpp_per_plane[k] ++ * dcn_bw_ceil2(v->byte_per_pixel_dety[k], 1.0) ++ / (v->htotal[k] / v->pixel_clock[k]) * v->v_ratio[k]; ++ v->read_bandwidth_plane_chroma[k] = v->swath_width_y[k] / 2.0 * v->dpp_per_plane[k] ++ * dcn_bw_ceil2(v->byte_per_pixel_detc[k], 2.0) ++ / (v->htotal[k] / v->pixel_clock[k]) * v->v_ratio[k] / 2.0; ++ v->total_data_read_bandwidth = v->total_data_read_bandwidth ++ + v->read_bandwidth_plane_luma[k] ++ + v->read_bandwidth_plane_chroma[k]; ++ } ++ v->total_active_dpp = 0.0; ++ v->total_dcc_active_dpp = 0.0; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ v->total_active_dpp = v->total_active_dpp + v->dpp_per_plane[k]; ++ if (v->dcc_enable[k] == dcn_bw_yes) { ++ v->total_dcc_active_dpp = v->total_dcc_active_dpp + v->dpp_per_plane[k]; ++ } ++ } ++ v->urgent_round_trip_and_out_of_order_latency = (v->round_trip_ping_latency_cycles + 32.0) ++ / v->dcfclk ++ + v->urgent_out_of_order_return_per_channel * v->number_of_channels ++ / v->return_bw; ++ v->last_pixel_of_line_extra_watermark = 0.0; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->v_ratio[k] <= 1.0) { ++ v->display_pipe_line_delivery_time_luma[k] = v->swath_width_y[k] ++ * v->dpp_per_plane[k] / v->h_ratio[k] / v->pixel_clock[k]; ++ } else { ++ v->display_pipe_line_delivery_time_luma[k] = v->swath_width_y[k] ++ / v->pscl_throughput[k] / v->dppclk; ++ } ++ v->data_fabric_line_delivery_time_luma = ++ v->swath_width_y[k] * v->swath_height_y[k] ++ * dcn_bw_ceil2(v->byte_per_pixel_dety[k], 1.0) ++ / (v->return_bw * v->read_bandwidth_plane_luma[k] ++ / v->dpp_per_plane[k] ++ / v->total_data_read_bandwidth); ++ v->last_pixel_of_line_extra_watermark = dcn_bw_max2( ++ v->last_pixel_of_line_extra_watermark, ++ v->data_fabric_line_delivery_time_luma ++ - v->display_pipe_line_delivery_time_luma[k]); ++ if (v->byte_per_pixel_detc[k] == 0.0) { ++ v->display_pipe_line_delivery_time_chroma[k] = 0.0; ++ } else { ++ if (v->v_ratio[k] / 2.0 <= 1.0) { ++ v->display_pipe_line_delivery_time_chroma[k] = v->swath_width_y[k] ++ / 2.0 * v->dpp_per_plane[k] / (v->h_ratio[k] / 2.0) ++ / v->pixel_clock[k]; ++ } else { ++ v->display_pipe_line_delivery_time_chroma[k] = v->swath_width_y[k] ++ / 2.0 / v->pscl_throughput_chroma[k] / v->dppclk; ++ } ++ v->data_fabric_line_delivery_time_chroma = v->swath_width_y[k] / 2.0 ++ * v->swath_height_c[k] ++ * dcn_bw_ceil2(v->byte_per_pixel_detc[k], 2.0) ++ / (v->return_bw * v->read_bandwidth_plane_chroma[k] ++ / v->dpp_per_plane[k] ++ / v->total_data_read_bandwidth); ++ v->last_pixel_of_line_extra_watermark = ++ dcn_bw_max2( ++ v->last_pixel_of_line_extra_watermark, ++ v->data_fabric_line_delivery_time_chroma ++ - v->display_pipe_line_delivery_time_chroma[k]); ++ } ++ } ++ v->urgent_extra_latency = v->urgent_round_trip_and_out_of_order_latency ++ + (v->total_active_dpp * v->pixel_chunk_size_in_kbyte ++ + v->total_dcc_active_dpp * v->meta_chunk_size) * 1024.0 ++ / v->return_bw; ++ if (v->pte_enable == dcn_bw_yes) { ++ v->urgent_extra_latency = v->urgent_extra_latency ++ + v->total_active_dpp * v->pte_chunk_size * 1024.0 / v->return_bw; ++ } ++ v->urgent_watermark = v->urgent_latency + v->last_pixel_of_line_extra_watermark ++ + v->urgent_extra_latency; ++ v->ptemeta_urgent_watermark = v->urgent_watermark + 2.0 * v->urgent_latency; ++ /*nb p-state/dram clock change watermark*/ ++ ++ v->dram_clock_change_watermark = v->dram_clock_change_latency + v->urgent_watermark; ++ v->total_active_writeback = 0.0; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->output[k] == dcn_bw_writeback) { ++ v->total_active_writeback = v->total_active_writeback + 1.0; ++ } ++ } ++ if (v->total_active_writeback <= 1.0) { ++ v->writeback_dram_clock_change_watermark = v->dram_clock_change_latency ++ + v->write_back_latency; ++ } else { ++ v->writeback_dram_clock_change_watermark = v->dram_clock_change_latency ++ + v->write_back_latency ++ + v->writeback_chunk_size * 1024.0 / 32.0 / v->socclk; ++ } ++ /*stutter efficiency*/ ++ ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ v->lines_in_dety[k] = v->det_buffer_size_y[k] / v->byte_per_pixel_dety[k] ++ / v->swath_width_y[k]; ++ v->lines_in_dety_rounded_down_to_swath[k] = dcn_bw_floor2( ++ v->lines_in_dety[k], ++ v->swath_height_y[k]); ++ v->full_det_buffering_time_y[k] = v->lines_in_dety_rounded_down_to_swath[k] ++ * (v->htotal[k] / v->pixel_clock[k]) / v->v_ratio[k]; ++ if (v->byte_per_pixel_detc[k] > 0.0) { ++ v->lines_in_detc[k] = v->det_buffer_size_c[k] / v->byte_per_pixel_detc[k] ++ / (v->swath_width_y[k] / 2.0); ++ v->lines_in_detc_rounded_down_to_swath[k] = dcn_bw_floor2( ++ v->lines_in_detc[k], ++ v->swath_height_c[k]); ++ v->full_det_buffering_time_c[k] = v->lines_in_detc_rounded_down_to_swath[k] ++ * (v->htotal[k] / v->pixel_clock[k]) ++ / (v->v_ratio[k] / 2.0); ++ } else { ++ v->lines_in_detc[k] = 0.0; ++ v->lines_in_detc_rounded_down_to_swath[k] = 0.0; ++ v->full_det_buffering_time_c[k] = 999999.0; ++ } ++ } ++ v->min_full_det_buffering_time = 999999.0; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->full_det_buffering_time_y[k] < v->min_full_det_buffering_time) { ++ v->min_full_det_buffering_time = v->full_det_buffering_time_y[k]; ++ v->frame_time_for_min_full_det_buffering_time = v->vtotal[k] * v->htotal[k] ++ / v->pixel_clock[k]; ++ } ++ if (v->full_det_buffering_time_c[k] < v->min_full_det_buffering_time) { ++ v->min_full_det_buffering_time = v->full_det_buffering_time_c[k]; ++ v->frame_time_for_min_full_det_buffering_time = v->vtotal[k] * v->htotal[k] ++ / v->pixel_clock[k]; ++ } ++ } ++ v->average_read_bandwidth_gbyte_per_second = 0.0; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->dcc_enable[k] == dcn_bw_yes) { ++ v->average_read_bandwidth_gbyte_per_second = ++ v->average_read_bandwidth_gbyte_per_second ++ + v->read_bandwidth_plane_luma[k] ++ / v->dcc_rate[k] / 1000.0 ++ + v->read_bandwidth_plane_chroma[k] ++ / v->dcc_rate[k] / 1000.0; ++ } else { ++ v->average_read_bandwidth_gbyte_per_second = ++ v->average_read_bandwidth_gbyte_per_second ++ + v->read_bandwidth_plane_luma[k] / 1000.0 ++ + v->read_bandwidth_plane_chroma[k] ++ / 1000.0; ++ } ++ if (v->dcc_enable[k] == dcn_bw_yes) { ++ v->average_read_bandwidth_gbyte_per_second = ++ v->average_read_bandwidth_gbyte_per_second ++ + v->read_bandwidth_plane_luma[k] / 1000.0 ++ / 256.0 ++ + v->read_bandwidth_plane_chroma[k] / 1000.0 ++ / 256.0; ++ } ++ if (v->pte_enable == dcn_bw_yes) { ++ v->average_read_bandwidth_gbyte_per_second = ++ v->average_read_bandwidth_gbyte_per_second ++ + v->read_bandwidth_plane_luma[k] / 1000.0 ++ / 512.0 ++ + v->read_bandwidth_plane_chroma[k] / 1000.0 ++ / 512.0; ++ } ++ } ++ v->part_of_burst_that_fits_in_rob = dcn_bw_min2( ++ v->min_full_det_buffering_time * v->total_data_read_bandwidth, ++ v->rob_buffer_size_in_kbyte * 1024.0 * v->total_data_read_bandwidth ++ / (v->average_read_bandwidth_gbyte_per_second * 1000.0)); ++ v->stutter_burst_time = v->part_of_burst_that_fits_in_rob ++ * (v->average_read_bandwidth_gbyte_per_second * 1000.0) ++ / v->total_data_read_bandwidth / v->return_bw ++ + (v->min_full_det_buffering_time * v->total_data_read_bandwidth ++ - v->part_of_burst_that_fits_in_rob) / (v->dcfclk * 64.0); ++ if (v->total_active_writeback == 0.0) { ++ v->stutter_efficiency_not_including_vblank = (1.0 ++ - (v->sr_exit_time + v->stutter_burst_time) ++ / v->min_full_det_buffering_time) * 100.0; ++ } else { ++ v->stutter_efficiency_not_including_vblank = 0.0; ++ } ++ v->smallest_vblank = 999999.0; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->synchronized_vblank == dcn_bw_yes || v->number_of_active_planes == 1) { ++ v->v_blank_time = (v->vtotal[k] - v->vactive[k]) * v->htotal[k] ++ / v->pixel_clock[k]; ++ } else { ++ v->v_blank_time = 0.0; ++ } ++ v->smallest_vblank = dcn_bw_min2(v->smallest_vblank, v->v_blank_time); ++ } ++ v->stutter_efficiency = (v->stutter_efficiency_not_including_vblank / 100.0 ++ * (v->frame_time_for_min_full_det_buffering_time - v->smallest_vblank) ++ + v->smallest_vblank) / v->frame_time_for_min_full_det_buffering_time ++ * 100.0; ++ /*dcfclk deep sleep*/ ++ ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->byte_per_pixel_detc[k] > 0.0) { ++ v->dcfclk_deep_sleep_per_plane[k] = ++ dcn_bw_max2( ++ 1.1 * v->swath_width_y[k] ++ * dcn_bw_ceil2( ++ v->byte_per_pixel_dety[k], ++ 1.0) / 32.0 ++ / v->display_pipe_line_delivery_time_luma[k], ++ 1.1 * v->swath_width_y[k] / 2.0 ++ * dcn_bw_ceil2( ++ v->byte_per_pixel_detc[k], ++ 2.0) / 32.0 ++ / v->display_pipe_line_delivery_time_chroma[k]); ++ } else { ++ v->dcfclk_deep_sleep_per_plane[k] = 1.1 * v->swath_width_y[k] ++ * dcn_bw_ceil2(v->byte_per_pixel_dety[k], 1.0) / 64.0 ++ / v->display_pipe_line_delivery_time_luma[k]; ++ } ++ v->dcfclk_deep_sleep_per_plane[k] = dcn_bw_max2( ++ v->dcfclk_deep_sleep_per_plane[k], ++ v->pixel_clock[k] / 16.0); ++ } ++ v->dcf_clk_deep_sleep = 8.0; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ v->dcf_clk_deep_sleep = dcn_bw_max2( ++ v->dcf_clk_deep_sleep, ++ v->dcfclk_deep_sleep_per_plane[k]); ++ } ++ /*stutter watermark*/ ++ ++ v->stutter_exit_watermark = v->sr_exit_time + v->last_pixel_of_line_extra_watermark ++ + v->urgent_extra_latency + 10.0 / v->dcf_clk_deep_sleep; ++ v->stutter_enter_plus_exit_watermark = v->sr_enter_plus_exit_time ++ + v->last_pixel_of_line_extra_watermark + v->urgent_extra_latency; ++ /*urgent latency supported*/ ++ ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ v->effective_det_plus_lb_lines_luma = ++ dcn_bw_floor2( ++ v->lines_in_dety[k] ++ + dcn_bw_min2( ++ v->lines_in_dety[k] ++ * v->dppclk ++ * v->byte_per_pixel_dety[k] ++ * v->pscl_throughput[k] ++ / (v->return_bw ++ / v->dpp_per_plane[k]), ++ v->effective_lb_latency_hiding_source_lines_luma), ++ v->swath_height_y[k]); ++ v->urgent_latency_support_us_luma = v->effective_det_plus_lb_lines_luma ++ * (v->htotal[k] / v->pixel_clock[k]) / v->v_ratio[k] ++ - v->effective_det_plus_lb_lines_luma * v->swath_width_y[k] ++ * v->byte_per_pixel_dety[k] ++ / (v->return_bw / v->dpp_per_plane[k]); ++ if (v->byte_per_pixel_detc[k] > 0.0) { ++ v->effective_det_plus_lb_lines_chroma = ++ dcn_bw_floor2( ++ v->lines_in_detc[k] ++ + dcn_bw_min2( ++ v->lines_in_detc[k] ++ * v->dppclk ++ * v->byte_per_pixel_detc[k] ++ * v->pscl_throughput_chroma[k] ++ / (v->return_bw ++ / v->dpp_per_plane[k]), ++ v->effective_lb_latency_hiding_source_lines_chroma), ++ v->swath_height_c[k]); ++ v->urgent_latency_support_us_chroma = v->effective_det_plus_lb_lines_chroma ++ * (v->htotal[k] / v->pixel_clock[k]) / (v->v_ratio[k] / 2.0) ++ - v->effective_det_plus_lb_lines_chroma ++ * (v->swath_width_y[k] / 2.0) ++ * v->byte_per_pixel_detc[k] ++ / (v->return_bw / v->dpp_per_plane[k]); ++ v->urgent_latency_support_us[k] = dcn_bw_min2( ++ v->urgent_latency_support_us_luma, ++ v->urgent_latency_support_us_chroma); ++ } else { ++ v->urgent_latency_support_us[k] = v->urgent_latency_support_us_luma; ++ } ++ } ++ v->min_urgent_latency_support_us = 999999.0; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ v->min_urgent_latency_support_us = dcn_bw_min2( ++ v->min_urgent_latency_support_us, ++ v->urgent_latency_support_us[k]); ++ } ++ /*non-urgent latency tolerance*/ ++ ++ v->non_urgent_latency_tolerance = v->min_urgent_latency_support_us - v->urgent_watermark; ++ /*prefetch*/ ++ ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if ((v->source_pixel_format[k] == dcn_bw_rgb_sub_64 ++ || v->source_pixel_format[k] == dcn_bw_rgb_sub_32 ++ || v->source_pixel_format[k] == dcn_bw_rgb_sub_16)) { ++ if (v->source_surface_mode[k] == dcn_bw_sw_linear) { ++ v->block_height256_bytes_y = 1.0; ++ } else if (v->source_pixel_format[k] == dcn_bw_rgb_sub_64) { ++ v->block_height256_bytes_y = 4.0; ++ } else { ++ v->block_height256_bytes_y = 8.0; ++ } ++ v->block_height256_bytes_c = 0.0; ++ } else { ++ if (v->source_surface_mode[k] == dcn_bw_sw_linear) { ++ v->block_height256_bytes_y = 1.0; ++ v->block_height256_bytes_c = 1.0; ++ } else if (v->source_pixel_format[k] == dcn_bw_yuv420_sub_8) { ++ v->block_height256_bytes_y = 16.0; ++ v->block_height256_bytes_c = 8.0; ++ } else { ++ v->block_height256_bytes_y = 8.0; ++ v->block_height256_bytes_c = 8.0; ++ } ++ } ++ if (v->dcc_enable[k] == dcn_bw_yes) { ++ v->meta_request_width_y = 64.0 * 256.0 ++ / dcn_bw_ceil2(v->byte_per_pixel_dety[k], 1.0) ++ / (8.0 * v->block_height256_bytes_y); ++ v->meta_surf_width_y = dcn_bw_ceil2( ++ v->swath_width_y[k] - 1.0, ++ v->meta_request_width_y) + v->meta_request_width_y; ++ v->meta_surf_height_y = dcn_bw_ceil2( ++ v->viewport_height[k] - 1.0, ++ 8.0 * v->block_height256_bytes_y) ++ + 8.0 * v->block_height256_bytes_y; ++ if (v->pte_enable == dcn_bw_yes) { ++ v->meta_pte_bytes_frame_y = ++ (dcn_bw_ceil2( ++ (v->meta_surf_width_y ++ * v->meta_surf_height_y ++ * dcn_bw_ceil2( ++ v->byte_per_pixel_dety[k], ++ 1.0) ++ / 256.0 - 4096.0) ++ / 8.0 / 4096.0, ++ 1.0) + 1) * 64.0; ++ } else { ++ v->meta_pte_bytes_frame_y = 0.0; ++ } ++ if (v->source_scan[k] == dcn_bw_hor) { ++ v->meta_row_byte_y = v->meta_surf_width_y * 8.0 ++ * v->block_height256_bytes_y ++ * dcn_bw_ceil2(v->byte_per_pixel_dety[k], 1.0) / 256.0; ++ } else { ++ v->meta_row_byte_y = v->meta_surf_height_y * v->meta_request_width_y ++ * dcn_bw_ceil2(v->byte_per_pixel_dety[k], 1.0) / 256.0; ++ } ++ } else { ++ v->meta_pte_bytes_frame_y = 0.0; ++ v->meta_row_byte_y = 0.0; ++ } ++ if (v->pte_enable == dcn_bw_yes) { ++ if (v->source_surface_mode[k] == dcn_bw_sw_linear) { ++ v->macro_tile_size_byte_y = 256.0; ++ v->macro_tile_height_y = 1.0; ++ } else if (v->source_surface_mode[k] == dcn_bw_sw_4_kb_s ++ || v->source_surface_mode[k] == dcn_bw_sw_4_kb_s_x ++ || v->source_surface_mode[k] == dcn_bw_sw_4_kb_d ++ || v->source_surface_mode[k] == dcn_bw_sw_4_kb_d_x) { ++ v->macro_tile_size_byte_y = 4096.0; ++ v->macro_tile_height_y = 4.0 * v->block_height256_bytes_y; ++ } else if (v->source_surface_mode[k] == dcn_bw_sw_64_kb_s ++ || v->source_surface_mode[k] == dcn_bw_sw_64_kb_s_t ++ || v->source_surface_mode[k] == dcn_bw_sw_64_kb_s_x ++ || v->source_surface_mode[k] == dcn_bw_sw_64_kb_d ++ || v->source_surface_mode[k] == dcn_bw_sw_64_kb_d_t ++ || v->source_surface_mode[k] == dcn_bw_sw_64_kb_d_x) { ++ v->macro_tile_size_byte_y = 64.0 * 1024; ++ v->macro_tile_height_y = 16.0 * v->block_height256_bytes_y; ++ } else { ++ v->macro_tile_size_byte_y = 256.0 * 1024; ++ v->macro_tile_height_y = 32.0 * v->block_height256_bytes_y; ++ } ++ if (v->macro_tile_size_byte_y <= 65536.0) { ++ v->pixel_pte_req_height_y = v->macro_tile_height_y; ++ } else { ++ v->pixel_pte_req_height_y = 16.0 * v->block_height256_bytes_y; ++ } ++ v->pixel_pte_req_width_y = 4096.0 / dcn_bw_ceil2(v->byte_per_pixel_dety[k], 1.0) ++ / v->pixel_pte_req_height_y * 8; ++ if (v->source_surface_mode[k] == dcn_bw_sw_linear) { ++ v->pixel_pte_bytes_per_row_y = ++ 64.0 ++ * (dcn_bw_ceil2( ++ (v->swath_width_y[k] ++ * dcn_bw_min2( ++ 128.0, ++ dcn_bw_pow( ++ 2.0, ++ dcn_bw_floor2( ++ dcn_bw_log( ++ v->pte_buffer_size_in_requests ++ * v->pixel_pte_req_width_y ++ / v->swath_width_y[k], ++ 2.0), ++ 1.0))) ++ - 1.0) ++ / v->pixel_pte_req_width_y, ++ 1.0) + 1); ++ } else if (v->source_scan[k] == dcn_bw_hor) { ++ v->pixel_pte_bytes_per_row_y = ++ 64.0 ++ * (dcn_bw_ceil2( ++ (v->swath_width_y[k] ++ - 1.0) ++ / v->pixel_pte_req_width_y, ++ 1.0) + 1); ++ } else { ++ v->pixel_pte_bytes_per_row_y = ++ 64.0 ++ * (dcn_bw_ceil2( ++ (v->viewport_height[k] ++ - 1.0) ++ / v->pixel_pte_req_height_y, ++ 1.0) + 1); ++ } ++ } else { ++ v->pixel_pte_bytes_per_row_y = 0.0; ++ } ++ if ((v->source_pixel_format[k] != dcn_bw_rgb_sub_64 ++ && v->source_pixel_format[k] != dcn_bw_rgb_sub_32 ++ && v->source_pixel_format[k] != dcn_bw_rgb_sub_16)) { ++ if (v->dcc_enable[k] == dcn_bw_yes) { ++ v->meta_request_width_c = 64.0 * 256.0 ++ / dcn_bw_ceil2(v->byte_per_pixel_detc[k], 2.0) ++ / (8.0 * v->block_height256_bytes_c); ++ v->meta_surf_width_c = dcn_bw_ceil2( ++ v->swath_width_y[k] / 2.0 - 1.0, ++ v->meta_request_width_c) + v->meta_request_width_c; ++ v->meta_surf_height_c = dcn_bw_ceil2( ++ v->viewport_height[k] / 2.0 - 1.0, ++ 8.0 * v->block_height256_bytes_c) ++ + 8.0 * v->block_height256_bytes_c; ++ if (v->pte_enable == dcn_bw_yes) { ++ v->meta_pte_bytes_frame_c = ++ (dcn_bw_ceil2( ++ (v->meta_surf_width_c ++ * v->meta_surf_height_c ++ * dcn_bw_ceil2( ++ v->byte_per_pixel_detc[k], ++ 2.0) ++ / 256.0 ++ - 4096.0) ++ / 8.0 ++ / 4096.0, ++ 1.0) + 1) * 64.0; ++ } else { ++ v->meta_pte_bytes_frame_c = 0.0; ++ } ++ if (v->source_scan[k] == dcn_bw_hor) { ++ v->meta_row_byte_c = v->meta_surf_width_c * 8.0 ++ * v->block_height256_bytes_c ++ * dcn_bw_ceil2(v->byte_per_pixel_detc[k], 2.0) ++ / 256.0; ++ } else { ++ v->meta_row_byte_c = v->meta_surf_height_c ++ * v->meta_request_width_c ++ * dcn_bw_ceil2(v->byte_per_pixel_detc[k], 2.0) ++ / 256.0; ++ } ++ } else { ++ v->meta_pte_bytes_frame_c = 0.0; ++ v->meta_row_byte_c = 0.0; ++ } ++ if (v->pte_enable == dcn_bw_yes) { ++ if (v->source_surface_mode[k] == dcn_bw_sw_linear) { ++ v->macro_tile_size_bytes_c = 256.0; ++ v->macro_tile_height_c = 1.0; ++ } else if (v->source_surface_mode[k] == dcn_bw_sw_4_kb_s ++ || v->source_surface_mode[k] == dcn_bw_sw_4_kb_s_x ++ || v->source_surface_mode[k] == dcn_bw_sw_4_kb_d ++ || v->source_surface_mode[k] ++ == dcn_bw_sw_4_kb_d_x) { ++ v->macro_tile_size_bytes_c = 4096.0; ++ v->macro_tile_height_c = 4.0 * v->block_height256_bytes_c; ++ } else if (v->source_surface_mode[k] == dcn_bw_sw_64_kb_s ++ || v->source_surface_mode[k] == dcn_bw_sw_64_kb_s_t ++ || v->source_surface_mode[k] == dcn_bw_sw_64_kb_s_x ++ || v->source_surface_mode[k] == dcn_bw_sw_64_kb_d ++ || v->source_surface_mode[k] == dcn_bw_sw_64_kb_d_t ++ || v->source_surface_mode[k] ++ == dcn_bw_sw_64_kb_d_x) { ++ v->macro_tile_size_bytes_c = 64.0 * 1024; ++ v->macro_tile_height_c = 16.0 * v->block_height256_bytes_c; ++ } else { ++ v->macro_tile_size_bytes_c = 256.0 * 1024; ++ v->macro_tile_height_c = 32.0 * v->block_height256_bytes_c; ++ } ++ if (v->macro_tile_size_bytes_c <= 65536.0) { ++ v->pixel_pte_req_height_c = v->macro_tile_height_c; ++ } else { ++ v->pixel_pte_req_height_c = 16.0 ++ * v->block_height256_bytes_c; ++ } ++ v->pixel_pte_req_width_c = 4096.0 ++ / dcn_bw_ceil2(v->byte_per_pixel_detc[k], 2.0) ++ / v->pixel_pte_req_height_c * 8; ++ if (v->source_surface_mode[k] == dcn_bw_sw_linear) { ++ v->pixel_pte_bytes_per_row_c = ++ 64.0 ++ * (dcn_bw_ceil2( ++ (v->swath_width_y[k] ++ / 2.0 ++ * dcn_bw_min2( ++ 128.0, ++ dcn_bw_pow( ++ 2.0, ++ dcn_bw_floor2( ++ dcn_bw_log( ++ v->pte_buffer_size_in_requests ++ * v->pixel_pte_req_width_c ++ / (v->swath_width_y[k] ++ / 2.0), ++ 2.0), ++ 1.0))) ++ - 1.0) ++ / v->pixel_pte_req_width_c, ++ 1.0) + 1); ++ } else if (v->source_scan[k] == dcn_bw_hor) { ++ v->pixel_pte_bytes_per_row_c = ++ 64.0 ++ * (dcn_bw_ceil2( ++ (v->swath_width_y[k] ++ / 2.0 ++ - 1.0) ++ / v->pixel_pte_req_width_c, ++ 1.0) + 1); ++ } else { ++ v->pixel_pte_bytes_per_row_c = ++ 64.0 ++ * (dcn_bw_ceil2( ++ (v->viewport_height[k] ++ / 2.0 ++ - 1.0) ++ / v->pixel_pte_req_height_c, ++ 1.0) + 1); ++ } ++ } else { ++ v->pixel_pte_bytes_per_row_c = 0.0; ++ } ++ } else { ++ v->pixel_pte_bytes_per_row_c = 0.0; ++ v->meta_pte_bytes_frame_c = 0.0; ++ v->meta_row_byte_c = 0.0; ++ } ++ v->pixel_pte_bytes_per_row[k] = v->pixel_pte_bytes_per_row_y ++ + v->pixel_pte_bytes_per_row_c; ++ v->meta_pte_bytes_frame[k] = v->meta_pte_bytes_frame_y + v->meta_pte_bytes_frame_c; ++ v->meta_row_byte[k] = v->meta_row_byte_y + v->meta_row_byte_c; ++ v->v_init_pre_fill_y[k] = dcn_bw_floor2( ++ (v->v_ratio[k] + v->vtaps[k] + 1.0 ++ + v->interlace_output[k] * 0.5 * v->v_ratio[k]) ++ / 2.0, ++ 1.0); ++ v->max_num_swath_y[k] = dcn_bw_ceil2( ++ (v->v_init_pre_fill_y[k] - 1.0) / v->swath_height_y[k], ++ 1.0) + 1; ++ if (v->v_init_pre_fill_y[k] > 1.0) { ++ v->max_partial_swath_y = dcn_bw_mod( ++ (v->v_init_pre_fill_y[k] - 2.0), ++ v->swath_height_y[k]); ++ } else { ++ v->max_partial_swath_y = dcn_bw_mod( ++ (v->v_init_pre_fill_y[k] + v->swath_height_y[k] - 2.0), ++ v->swath_height_y[k]); ++ } ++ v->max_partial_swath_y = dcn_bw_max2(1.0, v->max_partial_swath_y); ++ v->prefetch_source_lines_y[k] = v->max_num_swath_y[k] * v->swath_height_y[k] ++ + v->max_partial_swath_y; ++ if ((v->source_pixel_format[k] != dcn_bw_rgb_sub_64 ++ && v->source_pixel_format[k] != dcn_bw_rgb_sub_32 ++ && v->source_pixel_format[k] != dcn_bw_rgb_sub_16)) { ++ v->v_init_pre_fill_c[k] = dcn_bw_floor2( ++ (v->v_ratio[k] / 2.0 + v->vtaps[k] + 1.0 ++ + v->interlace_output[k] * 0.5 ++ * v->v_ratio[k] / 2.0) ++ / 2.0, ++ 1.0); ++ v->max_num_swath_c[k] = dcn_bw_ceil2( ++ (v->v_init_pre_fill_c[k] - 1.0) / v->swath_height_c[k], ++ 1.0) + 1; ++ if (v->v_init_pre_fill_c[k] > 1.0) { ++ v->max_partial_swath_c = dcn_bw_mod( ++ (v->v_init_pre_fill_c[k] - 2.0), ++ v->swath_height_c[k]); ++ } else { ++ v->max_partial_swath_c = dcn_bw_mod( ++ (v->v_init_pre_fill_c[k] + v->swath_height_c[k] ++ - 2.0), ++ v->swath_height_c[k]); ++ } ++ v->max_partial_swath_c = dcn_bw_max2(1.0, v->max_partial_swath_c); ++ } else { ++ v->max_num_swath_c[k] = 0.0; ++ v->max_partial_swath_c = 0.0; ++ } ++ v->prefetch_source_lines_c[k] = v->max_num_swath_c[k] * v->swath_height_c[k] ++ + v->max_partial_swath_c; ++ } ++ v->t_calc = 24.0 / v->dcf_clk_deep_sleep; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->can_vstartup_lines_exceed_vsync_plus_back_porch_lines_minus_one ++ == dcn_bw_yes) { ++ v->max_vstartup_lines[k] = v->vtotal[k] - v->vactive[k] - 1.0; ++ } else { ++ v->max_vstartup_lines[k] = v->v_sync_plus_back_porch[k] - 1.0; ++ } ++ } ++ v->next_prefetch_mode = 0.0; ++ do { ++ v->v_startup_lines = 13.0; ++ do { ++ v->planes_with_room_to_increase_vstartup_prefetch_bw_less_than_active_bw = ++ dcn_bw_yes; ++ v->planes_with_room_to_increase_vstartup_vratio_prefetch_more_than4 = ++ dcn_bw_no; ++ v->planes_with_room_to_increase_vstartup_destination_line_times_for_prefetch_less_than2 = ++ dcn_bw_no; ++ v->v_ratio_prefetch_more_than4 = dcn_bw_no; ++ v->destination_line_times_for_prefetch_less_than2 = dcn_bw_no; ++ v->prefetch_mode = v->next_prefetch_mode; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ v->dstx_after_scaler = 90.0 * v->pixel_clock[k] / v->dppclk ++ + 42.0 * v->pixel_clock[k] / v->dispclk; ++ if (v->dpp_per_plane[k] > 1.0) { ++ v->dstx_after_scaler = v->dstx_after_scaler ++ + v->scaler_rec_out_width[k] / 2.0; ++ } ++ if (v->output_format[k] == dcn_bw_420) { ++ v->dsty_after_scaler = 1.0; ++ } else { ++ v->dsty_after_scaler = 0.0; ++ } ++ v->v_update_offset_pix = dcn_bw_ceil2(v->htotal[k] / 4.0, 1.0); ++ v->total_repeater_delay_time = v->max_inter_dcn_tile_repeaters ++ * (2.0 / v->dppclk + 3.0 / v->dispclk); ++ v->v_update_width_pix = (14.0 / v->dcf_clk_deep_sleep ++ + 12.0 / v->dppclk + v->total_repeater_delay_time) ++ * v->pixel_clock[k]; ++ v->v_ready_offset_pix = dcn_bw_max2( ++ 150.0 / v->dppclk, ++ v->total_repeater_delay_time ++ + 20.0 / v->dcf_clk_deep_sleep ++ + 10.0 / v->dppclk) ++ * v->pixel_clock[k]; ++ v->t_setup = (v->v_update_offset_pix + v->v_update_width_pix ++ + v->v_ready_offset_pix) / v->pixel_clock[k]; ++ v->v_startup[k] = dcn_bw_min2( ++ v->v_startup_lines, ++ v->max_vstartup_lines[k]); ++ if (v->prefetch_mode == 0.0) { ++ v->t_wait = dcn_bw_max3( ++ v->dram_clock_change_latency ++ + v->urgent_latency, ++ v->sr_enter_plus_exit_time, ++ v->urgent_latency); ++ } else if (v->prefetch_mode == 1.0) { ++ v->t_wait = dcn_bw_max2( ++ v->sr_enter_plus_exit_time, ++ v->urgent_latency); ++ } else { ++ v->t_wait = v->urgent_latency; ++ } ++ v->destination_lines_for_prefetch[k] = ++ dcn_bw_floor2( ++ 4.0 ++ * (v->v_startup[k] ++ - v->t_wait ++ / (v->htotal[k] ++ / v->pixel_clock[k]) ++ - (v->t_calc ++ + v->t_setup) ++ / (v->htotal[k] ++ / v->pixel_clock[k]) ++ - (v->dsty_after_scaler ++ + v->dstx_after_scaler ++ / v->htotal[k]) ++ + 0.125), ++ 1.0) / 4; ++ if (v->destination_lines_for_prefetch[k] > 0.0) { ++ v->prefetch_bandwidth[k] = ++ (v->meta_pte_bytes_frame[k] ++ + 2.0 * v->meta_row_byte[k] ++ + 2.0 ++ * v->pixel_pte_bytes_per_row[k] ++ + v->prefetch_source_lines_y[k] ++ * v->swath_width_y[k] ++ * dcn_bw_ceil2( ++ v->byte_per_pixel_dety[k], ++ 1.0) ++ + v->prefetch_source_lines_c[k] ++ * v->swath_width_y[k] ++ / 2.0 ++ * dcn_bw_ceil2( ++ v->byte_per_pixel_detc[k], ++ 2.0)) ++ / (v->destination_lines_for_prefetch[k] ++ * v->htotal[k] ++ / v->pixel_clock[k]); ++ } else { ++ v->prefetch_bandwidth[k] = 999999.0; ++ } ++ } ++ v->bandwidth_available_for_immediate_flip = v->return_bw; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ v->bandwidth_available_for_immediate_flip = ++ v->bandwidth_available_for_immediate_flip ++ - dcn_bw_max2( ++ v->read_bandwidth_plane_luma[k] ++ + v->read_bandwidth_plane_chroma[k], ++ v->prefetch_bandwidth[k]); ++ } ++ v->tot_immediate_flip_bytes = 0.0; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->immediate_flip_supported == dcn_bw_yes ++ && (v->source_pixel_format[k] != dcn_bw_yuv420_sub_8 ++ && v->source_pixel_format[k] ++ != dcn_bw_yuv420_sub_10)) { ++ v->tot_immediate_flip_bytes = v->tot_immediate_flip_bytes ++ + v->meta_pte_bytes_frame[k] ++ + v->meta_row_byte[k] ++ + v->pixel_pte_bytes_per_row[k]; ++ } ++ } ++ v->max_rd_bandwidth = 0.0; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->pte_enable == dcn_bw_yes && v->dcc_enable[k] == dcn_bw_yes) { ++ if (v->immediate_flip_supported == dcn_bw_yes ++ && (v->source_pixel_format[k] ++ != dcn_bw_yuv420_sub_8 ++ && v->source_pixel_format[k] ++ != dcn_bw_yuv420_sub_10)) { ++ v->time_for_fetching_meta_pte = ++ dcn_bw_max5( ++ v->meta_pte_bytes_frame[k] ++ / v->prefetch_bandwidth[k], ++ v->meta_pte_bytes_frame[k] ++ * v->tot_immediate_flip_bytes ++ / (v->bandwidth_available_for_immediate_flip ++ * (v->meta_pte_bytes_frame[k] ++ + v->meta_row_byte[k] ++ + v->pixel_pte_bytes_per_row[k])), ++ v->urgent_extra_latency, ++ v->urgent_latency, ++ v->htotal[k] ++ / v->pixel_clock[k] ++ / 4.0); ++ } else { ++ v->time_for_fetching_meta_pte = ++ dcn_bw_max3( ++ v->meta_pte_bytes_frame[k] ++ / v->prefetch_bandwidth[k], ++ v->urgent_extra_latency, ++ v->htotal[k] ++ / v->pixel_clock[k] ++ / 4.0); ++ } ++ } else { ++ v->time_for_fetching_meta_pte = v->htotal[k] ++ / v->pixel_clock[k] / 4.0; ++ } ++ v->destination_lines_to_request_vm_inv_blank[k] = ++ dcn_bw_floor2( ++ 4.0 ++ * (v->time_for_fetching_meta_pte ++ / (v->htotal[k] ++ / v->pixel_clock[k]) ++ + 0.125), ++ 1.0) / 4; ++ if ((v->pte_enable == dcn_bw_yes || v->dcc_enable[k] == dcn_bw_yes)) { ++ if (v->immediate_flip_supported == dcn_bw_yes ++ && (v->source_pixel_format[k] ++ != dcn_bw_yuv420_sub_8 ++ && v->source_pixel_format[k] ++ != dcn_bw_yuv420_sub_10)) { ++ v->time_for_fetching_row_in_vblank = ++ dcn_bw_max5( ++ (v->meta_row_byte[k] ++ + v->pixel_pte_bytes_per_row[k]) ++ / v->prefetch_bandwidth[k], ++ (v->meta_row_byte[k] ++ + v->pixel_pte_bytes_per_row[k]) ++ * v->tot_immediate_flip_bytes ++ / (v->bandwidth_available_for_immediate_flip ++ * (v->meta_pte_bytes_frame[k] ++ + v->meta_row_byte[k] ++ + v->pixel_pte_bytes_per_row[k])), ++ v->urgent_extra_latency, ++ 2.0 ++ * v->urgent_latency, ++ v->htotal[k] ++ / v->pixel_clock[k] ++ - v->time_for_fetching_meta_pte); ++ } else { ++ v->time_for_fetching_row_in_vblank = ++ dcn_bw_max3( ++ (v->meta_row_byte[k] ++ + v->pixel_pte_bytes_per_row[k]) ++ / v->prefetch_bandwidth[k], ++ v->urgent_extra_latency, ++ v->htotal[k] ++ / v->pixel_clock[k] ++ - v->time_for_fetching_meta_pte); ++ } ++ } else { ++ v->time_for_fetching_row_in_vblank = ++ dcn_bw_max2( ++ v->urgent_extra_latency ++ - v->time_for_fetching_meta_pte, ++ v->htotal[k] ++ / v->pixel_clock[k] ++ - v->time_for_fetching_meta_pte); ++ } ++ v->destination_lines_to_request_row_in_vblank[k] = ++ dcn_bw_floor2( ++ 4.0 ++ * (v->time_for_fetching_row_in_vblank ++ / (v->htotal[k] ++ / v->pixel_clock[k]) ++ + 0.125), ++ 1.0) / 4; ++ v->lines_to_request_prefetch_pixel_data = ++ v->destination_lines_for_prefetch[k] ++ - v->destination_lines_to_request_vm_inv_blank[k] ++ - v->destination_lines_to_request_row_in_vblank[k]; ++ if (v->lines_to_request_prefetch_pixel_data > 0.0) { ++ v->v_ratio_prefetch_y[k] = v->prefetch_source_lines_y[k] ++ / v->lines_to_request_prefetch_pixel_data; ++ if ((v->swath_height_y[k] > 4.0)) { ++ if (v->lines_to_request_prefetch_pixel_data ++ > (v->v_init_pre_fill_y[k] - 3.0) ++ / 2.0) { ++ v->v_ratio_prefetch_y[k] = ++ dcn_bw_max2( ++ v->v_ratio_prefetch_y[k], ++ v->max_num_swath_y[k] ++ * v->swath_height_y[k] ++ / (v->lines_to_request_prefetch_pixel_data ++ - (v->v_init_pre_fill_y[k] ++ - 3.0) ++ / 2.0)); ++ } else { ++ v->v_ratio_prefetch_y[k] = 999999.0; ++ } ++ } ++ } else { ++ v->v_ratio_prefetch_y[k] = 999999.0; ++ } ++ v->v_ratio_prefetch_y[k] = dcn_bw_max2(v->v_ratio_prefetch_y[k], 1.0); ++ if (v->lines_to_request_prefetch_pixel_data > 0.0) { ++ v->v_ratio_prefetch_c[k] = v->prefetch_source_lines_c[k] ++ / v->lines_to_request_prefetch_pixel_data; ++ if ((v->swath_height_c[k] > 4.0)) { ++ if (v->lines_to_request_prefetch_pixel_data ++ > (v->v_init_pre_fill_c[k] - 3.0) ++ / 2.0) { ++ v->v_ratio_prefetch_c[k] = ++ dcn_bw_max2( ++ v->v_ratio_prefetch_c[k], ++ v->max_num_swath_c[k] ++ * v->swath_height_c[k] ++ / (v->lines_to_request_prefetch_pixel_data ++ - (v->v_init_pre_fill_c[k] ++ - 3.0) ++ / 2.0)); ++ } else { ++ v->v_ratio_prefetch_c[k] = 999999.0; ++ } ++ } ++ } else { ++ v->v_ratio_prefetch_c[k] = 999999.0; ++ } ++ v->v_ratio_prefetch_c[k] = dcn_bw_max2(v->v_ratio_prefetch_c[k], 1.0); ++ if (v->lines_to_request_prefetch_pixel_data > 0.0) { ++ v->required_prefetch_pix_data_bw = ++ v->dpp_per_plane[k] ++ * (v->prefetch_source_lines_y[k] ++ / v->lines_to_request_prefetch_pixel_data ++ * dcn_bw_ceil2( ++ v->byte_per_pixel_dety[k], ++ 1.0) ++ + v->prefetch_source_lines_c[k] ++ / v->lines_to_request_prefetch_pixel_data ++ * dcn_bw_ceil2( ++ v->byte_per_pixel_detc[k], ++ 2.0) ++ / 2.0) ++ * v->swath_width_y[k] ++ / (v->htotal[k] ++ / v->pixel_clock[k]); ++ } else { ++ v->required_prefetch_pix_data_bw = 999999.0; ++ } ++ v->max_rd_bandwidth = ++ v->max_rd_bandwidth ++ + dcn_bw_max2( ++ v->read_bandwidth_plane_luma[k] ++ + v->read_bandwidth_plane_chroma[k], ++ v->required_prefetch_pix_data_bw); ++ if (v->immediate_flip_supported == dcn_bw_yes ++ && (v->source_pixel_format[k] != dcn_bw_yuv420_sub_8 ++ && v->source_pixel_format[k] ++ != dcn_bw_yuv420_sub_10)) { ++ v->max_rd_bandwidth = ++ v->max_rd_bandwidth ++ + dcn_bw_max2( ++ v->meta_pte_bytes_frame[k] ++ / (v->destination_lines_to_request_vm_inv_blank[k] ++ * v->htotal[k] ++ / v->pixel_clock[k]), ++ (v->meta_row_byte[k] ++ + v->pixel_pte_bytes_per_row[k]) ++ / (v->destination_lines_to_request_row_in_vblank[k] ++ * v->htotal[k] ++ / v->pixel_clock[k])); ++ } ++ if (v->v_ratio_prefetch_y[k] > 4.0 ++ || v->v_ratio_prefetch_c[k] > 4.0) { ++ v->v_ratio_prefetch_more_than4 = dcn_bw_yes; ++ } ++ if (v->destination_lines_for_prefetch[k] < 2.0) { ++ v->destination_line_times_for_prefetch_less_than2 = ++ dcn_bw_yes; ++ } ++ if (v->max_vstartup_lines[k] > v->v_startup_lines) { ++ if (v->required_prefetch_pix_data_bw ++ > (v->read_bandwidth_plane_luma[k] ++ + v->read_bandwidth_plane_chroma[k])) { ++ v->planes_with_room_to_increase_vstartup_prefetch_bw_less_than_active_bw = ++ dcn_bw_no; ++ } ++ if (v->v_ratio_prefetch_y[k] > 4.0 ++ || v->v_ratio_prefetch_c[k] > 4.0) { ++ v->planes_with_room_to_increase_vstartup_vratio_prefetch_more_than4 = ++ dcn_bw_yes; ++ } ++ if (v->destination_lines_for_prefetch[k] < 2.0) { ++ v->planes_with_room_to_increase_vstartup_destination_line_times_for_prefetch_less_than2 = ++ dcn_bw_yes; ++ } ++ } ++ } ++ if (v->max_rd_bandwidth <= v->return_bw ++ && v->v_ratio_prefetch_more_than4 == dcn_bw_no ++ && v->destination_line_times_for_prefetch_less_than2 ++ == dcn_bw_no) { ++ v->prefetch_mode_supported = dcn_bw_yes; ++ } else { ++ v->prefetch_mode_supported = dcn_bw_no; ++ } ++ v->v_startup_lines = v->v_startup_lines + 1.0; ++ } while (!(v->prefetch_mode_supported == dcn_bw_yes ++ || (v->planes_with_room_to_increase_vstartup_prefetch_bw_less_than_active_bw ++ == dcn_bw_yes ++ && v->planes_with_room_to_increase_vstartup_vratio_prefetch_more_than4 ++ == dcn_bw_no ++ && v->planes_with_room_to_increase_vstartup_destination_line_times_for_prefetch_less_than2 ++ == dcn_bw_no))); ++ v->next_prefetch_mode = v->next_prefetch_mode + 1.0; ++ } while (!(v->prefetch_mode_supported == dcn_bw_yes || v->prefetch_mode == 2.0)); ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->v_ratio_prefetch_y[k] <= 1.0) { ++ v->display_pipe_line_delivery_time_luma_prefetch[k] = v->swath_width_y[k] ++ * v->dpp_per_plane[k] / v->h_ratio[k] / v->pixel_clock[k]; ++ } else { ++ v->display_pipe_line_delivery_time_luma_prefetch[k] = v->swath_width_y[k] ++ / v->pscl_throughput[k] / v->dppclk; ++ } ++ if (v->byte_per_pixel_detc[k] == 0.0) { ++ v->display_pipe_line_delivery_time_chroma_prefetch[k] = 0.0; ++ } else { ++ if (v->v_ratio_prefetch_c[k] <= 1.0) { ++ v->display_pipe_line_delivery_time_chroma_prefetch[k] = ++ v->swath_width_y[k] * v->dpp_per_plane[k] ++ / v->h_ratio[k] / v->pixel_clock[k]; ++ } else { ++ v->display_pipe_line_delivery_time_chroma_prefetch[k] = ++ v->swath_width_y[k] / v->pscl_throughput[k] ++ / v->dppclk; ++ } ++ } ++ } ++ /*min ttuv_blank*/ ++ ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->prefetch_mode == 0.0) { ++ v->allow_dram_clock_change_during_vblank[k] = dcn_bw_yes; ++ v->allow_dram_self_refresh_during_vblank[k] = dcn_bw_yes; ++ v->min_ttuv_blank[k] = v->t_calc ++ + dcn_bw_max3( ++ v->dram_clock_change_watermark, ++ v->stutter_enter_plus_exit_watermark, ++ v->urgent_watermark); ++ } else if (v->prefetch_mode == 1.0) { ++ v->allow_dram_clock_change_during_vblank[k] = dcn_bw_no; ++ v->allow_dram_self_refresh_during_vblank[k] = dcn_bw_yes; ++ v->min_ttuv_blank[k] = v->t_calc ++ + dcn_bw_max2( ++ v->stutter_enter_plus_exit_watermark, ++ v->urgent_watermark); ++ } else { ++ v->allow_dram_clock_change_during_vblank[k] = dcn_bw_no; ++ v->allow_dram_self_refresh_during_vblank[k] = dcn_bw_no; ++ v->min_ttuv_blank[k] = v->t_calc + v->urgent_watermark; ++ } ++ } ++ /*nb p-state/dram clock change support*/ ++ ++ v->active_dp_ps = 0.0; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ v->active_dp_ps = v->active_dp_ps + v->dpp_per_plane[k]; ++ } ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ v->lb_latency_hiding_source_lines_y = ++ dcn_bw_min2( ++ v->max_line_buffer_lines, ++ dcn_bw_floor2( ++ v->line_buffer_size ++ / v->lb_bit_per_pixel[k] ++ / (v->swath_width_y[k] ++ / dcn_bw_max2( ++ v->h_ratio[k], ++ 1.0)), ++ 1.0)) - (v->vtaps[k] - 1.0); ++ v->lb_latency_hiding_source_lines_c = ++ dcn_bw_min2( ++ v->max_line_buffer_lines, ++ dcn_bw_floor2( ++ v->line_buffer_size ++ / v->lb_bit_per_pixel[k] ++ / (v->swath_width_y[k] ++ / 2.0 ++ / dcn_bw_max2( ++ v->h_ratio[k] ++ / 2.0, ++ 1.0)), ++ 1.0)) - (v->vta_pschroma[k] - 1.0); ++ v->effective_lb_latency_hiding_y = v->lb_latency_hiding_source_lines_y ++ / v->v_ratio[k] * (v->htotal[k] / v->pixel_clock[k]); ++ v->effective_lb_latency_hiding_c = v->lb_latency_hiding_source_lines_c ++ / (v->v_ratio[k] / 2.0) * (v->htotal[k] / v->pixel_clock[k]); ++ if (v->swath_width_y[k] > 2.0 * v->dpp_output_buffer_pixels) { ++ v->dpp_output_buffer_lines_y = v->dpp_output_buffer_pixels ++ / v->swath_width_y[k]; ++ } else if (v->swath_width_y[k] > v->dpp_output_buffer_pixels) { ++ v->dpp_output_buffer_lines_y = 0.5; ++ } else { ++ v->dpp_output_buffer_lines_y = 1.0; ++ } ++ if (v->swath_width_y[k] / 2.0 > 2.0 * v->dpp_output_buffer_pixels) { ++ v->dpp_output_buffer_lines_c = v->dpp_output_buffer_pixels ++ / (v->swath_width_y[k] / 2.0); ++ } else if (v->swath_width_y[k] / 2.0 > v->dpp_output_buffer_pixels) { ++ v->dpp_output_buffer_lines_c = 0.5; ++ } else { ++ v->dpp_output_buffer_lines_c = 1.0; ++ } ++ v->dppopp_buffering_y = (v->htotal[k] / v->pixel_clock[k]) ++ * (v->dpp_output_buffer_lines_y + v->opp_output_buffer_lines); ++ v->max_det_buffering_time_y = v->full_det_buffering_time_y[k] ++ + (v->lines_in_dety[k] - v->lines_in_dety_rounded_down_to_swath[k]) ++ / v->swath_height_y[k] ++ * (v->htotal[k] / v->pixel_clock[k]); ++ v->active_dram_clock_change_latency_margin_y = v->dppopp_buffering_y ++ + v->effective_lb_latency_hiding_y + v->max_det_buffering_time_y ++ - v->dram_clock_change_watermark; ++ if (v->active_dp_ps > 1.0) { ++ v->active_dram_clock_change_latency_margin_y = ++ v->active_dram_clock_change_latency_margin_y ++ - (1.0 - 1.0 / (v->active_dp_ps - 1.0)) ++ * v->swath_height_y[k] ++ * (v->htotal[k] ++ / v->pixel_clock[k]); ++ } ++ if (v->byte_per_pixel_detc[k] > 0.0) { ++ v->dppopp_buffering_c = ++ (v->htotal[k] / v->pixel_clock[k]) ++ * (v->dpp_output_buffer_lines_c ++ + v->opp_output_buffer_lines); ++ v->max_det_buffering_time_c = v->full_det_buffering_time_c[k] ++ + (v->lines_in_detc[k] ++ - v->lines_in_detc_rounded_down_to_swath[k]) ++ / v->swath_height_c[k] ++ * (v->htotal[k] / v->pixel_clock[k]); ++ v->active_dram_clock_change_latency_margin_c = v->dppopp_buffering_c ++ + v->effective_lb_latency_hiding_c ++ + v->max_det_buffering_time_c ++ - v->dram_clock_change_watermark; ++ if (v->active_dp_ps > 1.0) { ++ v->active_dram_clock_change_latency_margin_c = ++ v->active_dram_clock_change_latency_margin_c ++ - (1.0 ++ - 1.0 ++ / (v->active_dp_ps ++ - 1.0)) ++ * v->swath_height_c[k] ++ * (v->htotal[k] ++ / v->pixel_clock[k]); ++ } ++ v->active_dram_clock_change_latency_margin[k] = dcn_bw_min2( ++ v->active_dram_clock_change_latency_margin_y, ++ v->active_dram_clock_change_latency_margin_c); ++ } else { ++ v->active_dram_clock_change_latency_margin[k] = ++ v->active_dram_clock_change_latency_margin_y; ++ } ++ if (v->output_format[k] == dcn_bw_444) { ++ v->writeback_dram_clock_change_latency_margin = ++ (v->writeback_luma_buffer_size ++ + v->writeback_chroma_buffer_size) * 1024.0 ++ / (v->scaler_rec_out_width[k] ++ / (v->htotal[k] ++ / v->pixel_clock[k]) ++ * 4.0) ++ - v->writeback_dram_clock_change_watermark; ++ } else { ++ v->writeback_dram_clock_change_latency_margin = dcn_bw_min2( ++ v->writeback_luma_buffer_size, ++ 2.0 * v->writeback_chroma_buffer_size) * 1024.0 ++ / (v->scaler_rec_out_width[k] ++ / (v->htotal[k] / v->pixel_clock[k])) ++ - v->writeback_dram_clock_change_watermark; ++ } ++ if (v->output[k] == dcn_bw_writeback) { ++ v->active_dram_clock_change_latency_margin[k] = dcn_bw_min2( ++ v->active_dram_clock_change_latency_margin[k], ++ v->writeback_dram_clock_change_latency_margin); ++ } ++ } ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->allow_dram_clock_change_during_vblank[k] == dcn_bw_yes) { ++ v->v_blank_dram_clock_change_latency_margin[k] = (v->vtotal[k] ++ - v->scaler_recout_height[k]) ++ * (v->htotal[k] / v->pixel_clock[k]) ++ - dcn_bw_max2( ++ v->dram_clock_change_watermark, ++ v->writeback_dram_clock_change_watermark); ++ } else { ++ v->v_blank_dram_clock_change_latency_margin[k] = 0.0; ++ } ++ } ++ v->min_active_dram_clock_change_margin = 999999.0; ++ v->v_blank_of_min_active_dram_clock_change_margin = 999999.0; ++ v->second_min_active_dram_clock_change_margin = 999999.0; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->active_dram_clock_change_latency_margin[k] ++ < v->min_active_dram_clock_change_margin) { ++ v->second_min_active_dram_clock_change_margin = ++ v->min_active_dram_clock_change_margin; ++ v->min_active_dram_clock_change_margin = ++ v->active_dram_clock_change_latency_margin[k]; ++ v->v_blank_of_min_active_dram_clock_change_margin = ++ v->v_blank_dram_clock_change_latency_margin[k]; ++ } else if (v->active_dram_clock_change_latency_margin[k] ++ < v->second_min_active_dram_clock_change_margin) { ++ v->second_min_active_dram_clock_change_margin = ++ v->active_dram_clock_change_latency_margin[k]; ++ } ++ } ++ v->min_vblank_dram_clock_change_margin = 999999.0; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->min_vblank_dram_clock_change_margin ++ > v->v_blank_dram_clock_change_latency_margin[k]) { ++ v->min_vblank_dram_clock_change_margin = ++ v->v_blank_dram_clock_change_latency_margin[k]; ++ } ++ } ++ if (v->synchronized_vblank == dcn_bw_yes || v->number_of_active_planes == 1) { ++ v->dram_clock_change_margin = dcn_bw_max2( ++ v->min_active_dram_clock_change_margin, ++ v->min_vblank_dram_clock_change_margin); ++ } else if (v->v_blank_of_min_active_dram_clock_change_margin ++ > v->min_active_dram_clock_change_margin) { ++ v->dram_clock_change_margin = dcn_bw_min2( ++ v->second_min_active_dram_clock_change_margin, ++ v->v_blank_of_min_active_dram_clock_change_margin); ++ } else { ++ v->dram_clock_change_margin = v->min_active_dram_clock_change_margin; ++ } ++ if (v->min_active_dram_clock_change_margin > 0.0) { ++ v->dram_clock_change_support = dcn_bw_supported_in_v_active; ++ } else if (v->dram_clock_change_margin > 0.0) { ++ v->dram_clock_change_support = dcn_bw_supported_in_v_blank; ++ } else { ++ v->dram_clock_change_support = dcn_bw_not_supported; ++ } ++ /*maximum bandwidth used*/ ++ ++ v->wr_bandwidth = 0.0; ++ for (k = 0; k <= v->number_of_active_planes - 1; k++) { ++ if (v->output[k] == dcn_bw_writeback && v->output_format[k] == dcn_bw_444) { ++ v->wr_bandwidth = v->wr_bandwidth ++ + v->scaler_rec_out_width[k] ++ / (v->htotal[k] / v->pixel_clock[k]) * 4.0; ++ } else if (v->output[k] == dcn_bw_writeback) { ++ v->wr_bandwidth = v->wr_bandwidth ++ + v->scaler_rec_out_width[k] ++ / (v->htotal[k] / v->pixel_clock[k]) * 1.5; ++ } ++ } ++ v->max_used_bw = v->max_rd_bandwidth + v->wr_bandwidth; ++} +diff --git a/drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_auto.h b/drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_auto.h +new file mode 100644 +index 0000000..03f06f6 +--- /dev/null ++++ b/drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_auto.h +@@ -0,0 +1,37 @@ ++/* ++ * 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 _DCN_CALC_AUTO_H_ ++#define _DCN_CALC_AUTO_H_ ++ ++#include "dcn_calcs.h" ++ ++void scaler_settings_calculation(struct dcn_bw_internal_vars *v); ++void mode_support_and_system_configuration(struct dcn_bw_internal_vars *v); ++void display_pipe_configuration(struct dcn_bw_internal_vars *v); ++void dispclkdppclkdcfclk_deep_sleep_prefetch_parameters_watermarks_and_performance_calculation( ++ struct dcn_bw_internal_vars *v); ++ ++#endif /* _DCN_CALC_AUTO_H_ */ +diff --git a/drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_math.c b/drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_math.c +new file mode 100644 +index 0000000..a184744 +--- /dev/null ++++ b/drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_math.c +@@ -0,0 +1,104 @@ ++/* ++ * 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 "dcn_calc_math.h" ++ ++float dcn_bw_mod(const float arg1, const float arg2) ++{ ++ return arg1 - arg1 * ((int) (arg1 / arg2)); ++} ++ ++float dcn_bw_min2(const float arg1, const float arg2) ++{ ++ return arg1 < arg2 ? arg1 : arg2; ++} ++ ++unsigned int dcn_bw_max(const unsigned int arg1, const unsigned int arg2) ++{ ++ return arg1 > arg2 ? arg1 : arg2; ++} ++float dcn_bw_max2(const float arg1, const float arg2) ++{ ++ return arg1 > arg2 ? arg1 : arg2; ++} ++ ++float dcn_bw_floor2(const float arg, const float significance) ++{ ++ if (significance == 0) ++ return 0; ++ return ((int) (arg / significance)) * significance; ++} ++ ++float dcn_bw_ceil2(const float arg, const float significance) ++{ ++ float flr = dcn_bw_floor2(arg, significance); ++ if (significance == 0) ++ return 0; ++ return flr + 0.00001 >= arg ? arg : flr + significance; ++} ++ ++float dcn_bw_max3(float v1, float v2, float v3) ++{ ++ return v3 > dcn_bw_max2(v1, v2) ? v3 : dcn_bw_max2(v1, v2); ++} ++ ++float dcn_bw_max5(float v1, float v2, float v3, float v4, float v5) ++{ ++ return dcn_bw_max3(v1, v2, v3) > dcn_bw_max2(v4, v5) ? dcn_bw_max3(v1, v2, v3) : dcn_bw_max2(v4, v5); ++} ++ ++float dcn_bw_pow(float a, float exp) ++{ ++ float temp; ++ /*ASSERT(exp == (int)exp);*/ ++ if ((int)exp == 0) ++ return 1; ++ temp = dcn_bw_pow(a, (int)(exp / 2)); ++ if (((int)exp % 2) == 0) { ++ return temp * temp; ++ } else { ++ if ((int)exp > 0) ++ return a * temp * temp; ++ else ++ return (temp * temp) / a; ++ } ++} ++ ++float dcn_bw_log(float a, float b) ++{ ++ int * const exp_ptr = (int *)(&a); ++ int x = *exp_ptr; ++ const int log_2 = ((x >> 23) & 255) - 128; ++ x &= ~(255 << 23); ++ x += 127 << 23; ++ *exp_ptr = x; ++ ++ a = ((-1.0f / 3) * a + 2) * a - 2.0f / 3; ++ ++ if (b > 2.00001 || b < 1.99999) ++ return (a + log_2) / dcn_bw_log(b, 2); ++ else ++ return (a + log_2); ++} +diff --git a/drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_math.h b/drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_math.h +new file mode 100644 +index 0000000..f46ab0e +--- /dev/null ++++ b/drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_math.h +@@ -0,0 +1,40 @@ ++/* ++ * 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 _DCN_CALC_MATH_H_ ++#define _DCN_CALC_MATH_H_ ++ ++float dcn_bw_mod(const float arg1, const float arg2); ++float dcn_bw_min2(const float arg1, const float arg2); ++unsigned int dcn_bw_max(const unsigned int arg1, const unsigned int arg2); ++float dcn_bw_max2(const float arg1, const float arg2); ++float dcn_bw_floor2(const float arg, const float significance); ++float dcn_bw_ceil2(const float arg, const float significance); ++float dcn_bw_max3(float v1, float v2, float v3); ++float dcn_bw_max5(float v1, float v2, float v3, float v4, float v5); ++float dcn_bw_pow(float a, float exp); ++float dcn_bw_log(float a, float b); ++ ++#endif /* _DCN_CALC_MATH_H_ */ +diff --git a/drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c b/drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c +new file mode 100644 +index 0000000..eb81edf +--- /dev/null ++++ b/drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c +@@ -0,0 +1,1366 @@ ++/* ++ * 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 "dm_services.h" ++#include "dcn_calcs.h" ++#include "dcn_calc_auto.h" ++#include "dc.h" ++#include "core_dc.h" ++#include "dal_asic_id.h" ++ ++#include "resource.h" ++#include "dcn10/dcn10_resource.h" ++#include "dcn_calc_math.h" ++ ++/* Defaults from spreadsheet rev#247 */ ++const struct dcn_soc_bounding_box dcn10_soc_defaults = { ++ .sr_exit_time = 17, /*us*/ /*update based on HW Request for 118773*/ ++ .sr_enter_plus_exit_time = 19, /*us*/ ++ .urgent_latency = 4, /*us*/ ++ .write_back_latency = 12, /*us*/ ++ .percent_of_ideal_drambw_received_after_urg_latency = 80, /*%*/ ++ .max_request_size = 256, /*bytes*/ ++ .dcfclkv_max0p9 = 600, /*MHz*/ ++ .dcfclkv_nom0p8 = 600, /*MHz*/ ++ .dcfclkv_mid0p72 = 300, /*MHz*/ ++ .dcfclkv_min0p65 = 300, /*MHz*/ ++ .max_dispclk_vmax0p9 = 1086, /*MHz*/ ++ .max_dispclk_vnom0p8 = 661, /*MHz*/ ++ .max_dispclk_vmid0p72 = 608, /*MHz*/ ++ .max_dispclk_vmin0p65 = 608, /*MHz*/ ++ .max_dppclk_vmax0p9 = 661, /*MHz*/ ++ .max_dppclk_vnom0p8 = 661, /*MHz*/ ++ .max_dppclk_vmid0p72 = 435, /*MHz*/ ++ .max_dppclk_vmin0p65 = 435, /*MHz*/ ++ .socclk = 208, /*MHz*/ ++ .fabric_and_dram_bandwidth_vmax0p9 = 38.4f, /*GB/s*/ ++ .fabric_and_dram_bandwidth_vnom0p8 = 34.1f, /*GB/s*/ ++ .fabric_and_dram_bandwidth_vmid0p72 = 29.8f, /*GB/s*/ ++ .fabric_and_dram_bandwidth_vmin0p65 = 12.8f, /*GB/s*/ ++ .phyclkv_max0p9 = 810, /*MHz*/ ++ .phyclkv_nom0p8 = 810, /*MHz*/ ++ .phyclkv_mid0p72 = 540, /*MHz*/ ++ .phyclkv_min0p65 = 540, /*MHz*/ ++ .downspreading = 0.5f, /*%*/ ++ .round_trip_ping_latency_cycles = 128, /*DCFCLK Cycles*/ ++ .urgent_out_of_order_return_per_channel = 256, /*bytes*/ ++ .number_of_channels = 2, ++ .vmm_page_size = 4096, /*bytes*/ ++ .dram_clock_change_latency = 17, /*us*/ ++ .return_bus_width = 64, /*bytes*/ ++}; ++ ++const struct dcn_ip_params dcn10_ip_defaults = { ++ .rob_buffer_size_in_kbyte = 64, ++ .det_buffer_size_in_kbyte = 164, ++ .dpp_output_buffer_pixels = 2560, ++ .opp_output_buffer_lines = 1, ++ .pixel_chunk_size_in_kbyte = 8, ++ .pte_enable = dcn_bw_yes, ++ .pte_chunk_size = 2, /*kbytes*/ ++ .meta_chunk_size = 2, /*kbytes*/ ++ .writeback_chunk_size = 2, /*kbytes*/ ++ .odm_capability = dcn_bw_no, ++ .dsc_capability = dcn_bw_no, ++ .line_buffer_size = 589824, /*bit*/ ++ .max_line_buffer_lines = 12, ++ .is_line_buffer_bpp_fixed = dcn_bw_no, ++ .line_buffer_fixed_bpp = dcn_bw_na, ++ .writeback_luma_buffer_size = 12, /*kbytes*/ ++ .writeback_chroma_buffer_size = 8, /*kbytes*/ ++ .max_num_dpp = 4, ++ .max_num_writeback = 2, ++ .max_dchub_topscl_throughput = 4, /*pixels/dppclk*/ ++ .max_pscl_tolb_throughput = 2, /*pixels/dppclk*/ ++ .max_lb_tovscl_throughput = 4, /*pixels/dppclk*/ ++ .max_vscl_tohscl_throughput = 4, /*pixels/dppclk*/ ++ .max_hscl_ratio = 4, ++ .max_vscl_ratio = 4, ++ .max_hscl_taps = 8, ++ .max_vscl_taps = 8, ++ .pte_buffer_size_in_requests = 42, ++ .dispclk_ramping_margin = 1, /*%*/ ++ .under_scan_factor = 1.11f, ++ .max_inter_dcn_tile_repeaters = 8, ++ .can_vstartup_lines_exceed_vsync_plus_back_porch_lines_minus_one = dcn_bw_no, ++ .bug_forcing_luma_and_chroma_request_to_same_size_fixed = dcn_bw_no, ++ .dcfclk_cstate_latency = 10 /*TODO clone of something else? sr_enter_plus_exit_time?*/ ++}; ++ ++static enum dcn_bw_defs tl_sw_mode_to_bw_defs(enum swizzle_mode_values sw_mode) ++{ ++ switch (sw_mode) { ++ case DC_SW_LINEAR: ++ return dcn_bw_sw_linear; ++ case DC_SW_4KB_S: ++ return dcn_bw_sw_4_kb_s; ++ case DC_SW_4KB_D: ++ return dcn_bw_sw_4_kb_d; ++ case DC_SW_64KB_S: ++ return dcn_bw_sw_64_kb_s; ++ case DC_SW_64KB_D: ++ return dcn_bw_sw_64_kb_d; ++ case DC_SW_VAR_S: ++ return dcn_bw_sw_var_s; ++ case DC_SW_VAR_D: ++ return dcn_bw_sw_var_d; ++ case DC_SW_64KB_S_T: ++ return dcn_bw_sw_64_kb_s_t; ++ case DC_SW_64KB_D_T: ++ return dcn_bw_sw_64_kb_d_t; ++ case DC_SW_4KB_S_X: ++ return dcn_bw_sw_4_kb_s_x; ++ case DC_SW_4KB_D_X: ++ return dcn_bw_sw_4_kb_d_x; ++ case DC_SW_64KB_S_X: ++ return dcn_bw_sw_64_kb_s_x; ++ case DC_SW_64KB_D_X: ++ return dcn_bw_sw_64_kb_d_x; ++ case DC_SW_VAR_S_X: ++ return dcn_bw_sw_var_s_x; ++ case DC_SW_VAR_D_X: ++ return dcn_bw_sw_var_d_x; ++ case DC_SW_256B_S: ++ case DC_SW_256_D: ++ case DC_SW_256_R: ++ case DC_SW_4KB_R: ++ case DC_SW_64KB_R: ++ case DC_SW_VAR_R: ++ case DC_SW_4KB_R_X: ++ case DC_SW_64KB_R_X: ++ case DC_SW_VAR_R_X: ++ default: ++ BREAK_TO_DEBUGGER(); /*not in formula*/ ++ return dcn_bw_sw_4_kb_s; ++ } ++} ++ ++static int tl_lb_bpp_to_int(enum lb_pixel_depth depth) ++{ ++ switch (depth) { ++ case LB_PIXEL_DEPTH_18BPP: ++ return 18; ++ case LB_PIXEL_DEPTH_24BPP: ++ return 24; ++ case LB_PIXEL_DEPTH_30BPP: ++ return 30; ++ case LB_PIXEL_DEPTH_36BPP: ++ return 36; ++ default: ++ return 30; ++ } ++} ++ ++static enum dcn_bw_defs tl_pixel_format_to_bw_defs(enum surface_pixel_format format) ++{ ++ switch (format) { ++ case SURFACE_PIXEL_FORMAT_GRPH_ARGB1555: ++ case SURFACE_PIXEL_FORMAT_GRPH_RGB565: ++ return dcn_bw_rgb_sub_16; ++ case SURFACE_PIXEL_FORMAT_GRPH_ARGB8888: ++ case SURFACE_PIXEL_FORMAT_GRPH_ABGR8888: ++ case SURFACE_PIXEL_FORMAT_GRPH_ARGB2101010: ++ case SURFACE_PIXEL_FORMAT_GRPH_ABGR2101010: ++ case SURFACE_PIXEL_FORMAT_GRPH_ABGR2101010_XR_BIAS: ++ return dcn_bw_rgb_sub_32; ++ case SURFACE_PIXEL_FORMAT_GRPH_ARGB16161616: ++ case SURFACE_PIXEL_FORMAT_GRPH_ARGB16161616F: ++ case SURFACE_PIXEL_FORMAT_GRPH_ABGR16161616F: ++ return dcn_bw_rgb_sub_64; ++ case SURFACE_PIXEL_FORMAT_VIDEO_420_YCbCr: ++ case SURFACE_PIXEL_FORMAT_VIDEO_420_YCrCb: ++ return dcn_bw_yuv420_sub_8; ++ case SURFACE_PIXEL_FORMAT_VIDEO_420_10bpc_YCbCr: ++ case SURFACE_PIXEL_FORMAT_VIDEO_420_10bpc_YCrCb: ++ return dcn_bw_yuv420_sub_10; ++ default: ++ return dcn_bw_rgb_sub_32; ++ } ++} ++ ++static void pipe_ctx_to_e2e_pipe_params ( ++ const struct pipe_ctx *pipe, ++ struct _vcs_dpi_display_pipe_params_st *input) ++{ ++ input->src.is_hsplit = false; ++ if (pipe->top_pipe != NULL && pipe->top_pipe->surface == pipe->surface) ++ input->src.is_hsplit = true; ++ else if (pipe->bottom_pipe != NULL && pipe->bottom_pipe->surface == pipe->surface) ++ input->src.is_hsplit = true; ++ ++ input->src.dcc = pipe->surface->public.dcc.enable; ++ input->src.dcc_rate = 1; ++ input->src.meta_pitch = pipe->surface->public.dcc.grph.meta_pitch; ++ input->src.source_scan = dm_horz; ++ input->src.sw_mode = pipe->surface->public.tiling_info.gfx9.swizzle; ++ ++ input->src.viewport_width = pipe->scl_data.viewport.width; ++ input->src.viewport_height = pipe->scl_data.viewport.height; ++ input->src.data_pitch = pipe->scl_data.viewport.width; ++ input->src.data_pitch_c = pipe->scl_data.viewport.width; ++ input->src.cur0_src_width = 128; /* TODO: Cursor calcs, not curently stored */ ++ input->src.cur0_bpp = 32; ++ ++ switch (pipe->surface->public.tiling_info.gfx9.swizzle) { ++ /* for 4/8/16 high tiles */ ++ case DC_SW_LINEAR: ++ input->src.is_display_sw = 1; ++ input->src.macro_tile_size = dm_4k_tile; ++ break; ++ case DC_SW_4KB_S: ++ case DC_SW_4KB_S_X: ++ input->src.is_display_sw = 0; ++ input->src.macro_tile_size = dm_4k_tile; ++ break; ++ case DC_SW_64KB_S: ++ case DC_SW_64KB_S_X: ++ input->src.is_display_sw = 0; ++ input->src.macro_tile_size = dm_64k_tile; ++ break; ++ case DC_SW_VAR_S: ++ case DC_SW_VAR_S_X: ++ input->src.is_display_sw = 0; ++ input->src.macro_tile_size = dm_256k_tile; ++ break; ++ ++ /* For 64bpp 2 high tiles */ ++ case DC_SW_4KB_D: ++ case DC_SW_4KB_D_X: ++ input->src.is_display_sw = 1; ++ input->src.macro_tile_size = dm_4k_tile; ++ break; ++ case DC_SW_64KB_D: ++ case DC_SW_64KB_D_X: ++ input->src.is_display_sw = 1; ++ input->src.macro_tile_size = dm_64k_tile; ++ break; ++ case DC_SW_VAR_D: ++ case DC_SW_VAR_D_X: ++ input->src.is_display_sw = 1; ++ input->src.macro_tile_size = dm_256k_tile; ++ break; ++ ++ /* Unsupported swizzle modes for dcn */ ++ case DC_SW_256B_S: ++ default: ++ ASSERT(0); /* Not supported */ ++ break; ++ } ++ ++ switch (pipe->surface->public.rotation) { ++ case ROTATION_ANGLE_0: ++ case ROTATION_ANGLE_180: ++ input->src.source_scan = dm_horz; ++ break; ++ case ROTATION_ANGLE_90: ++ case ROTATION_ANGLE_270: ++ input->src.source_scan = dm_vert; ++ break; ++ default: ++ ASSERT(0); /* Not supported */ ++ break; ++ } ++ ++ /* TODO: Fix pixel format mappings */ ++ switch (pipe->surface->public.format) { ++ case SURFACE_PIXEL_FORMAT_VIDEO_420_YCbCr: ++ case SURFACE_PIXEL_FORMAT_VIDEO_420_YCrCb: ++ input->src.source_format = dm_420_8; ++ input->src.viewport_width_c = input->src.viewport_width / 2; ++ input->src.viewport_height_c = input->src.viewport_height / 2; ++ break; ++ case SURFACE_PIXEL_FORMAT_VIDEO_420_10bpc_YCbCr: ++ case SURFACE_PIXEL_FORMAT_VIDEO_420_10bpc_YCrCb: ++ input->src.source_format = dm_420_10; ++ input->src.viewport_width_c = input->src.viewport_width / 2; ++ input->src.viewport_height_c = input->src.viewport_height / 2; ++ break; ++ case SURFACE_PIXEL_FORMAT_GRPH_ARGB16161616: ++ case SURFACE_PIXEL_FORMAT_GRPH_ARGB16161616F: ++ case SURFACE_PIXEL_FORMAT_GRPH_ABGR16161616F: ++ input->src.source_format = dm_444_64; ++ input->src.viewport_width_c = input->src.viewport_width; ++ input->src.viewport_height_c = input->src.viewport_height; ++ break; ++ default: ++ input->src.source_format = dm_444_32; ++ input->src.viewport_width_c = input->src.viewport_width; ++ input->src.viewport_height_c = input->src.viewport_height; ++ break; ++ } ++ ++ input->scale_taps.htaps = pipe->scl_data.taps.h_taps; ++ input->scale_ratio_depth.hscl_ratio = pipe->scl_data.ratios.horz.value/4294967296.0; ++ input->scale_ratio_depth.vscl_ratio = pipe->scl_data.ratios.vert.value/4294967296.0; ++ input->scale_ratio_depth.vinit = pipe->scl_data.inits.v.value/4294967296.0; ++ if (input->scale_ratio_depth.vinit < 1.0) ++ input->scale_ratio_depth.vinit = 1; ++ input->scale_taps.vtaps = pipe->scl_data.taps.v_taps; ++ input->scale_taps.vtaps_c = pipe->scl_data.taps.v_taps_c; ++ input->scale_taps.htaps_c = pipe->scl_data.taps.h_taps_c; ++ input->scale_ratio_depth.hscl_ratio_c = pipe->scl_data.ratios.horz_c.value/4294967296.0; ++ input->scale_ratio_depth.vscl_ratio_c = pipe->scl_data.ratios.vert_c.value/4294967296.0; ++ input->scale_ratio_depth.vinit_c = pipe->scl_data.inits.v_c.value/4294967296.0; ++ if (input->scale_ratio_depth.vinit_c < 1.0) ++ input->scale_ratio_depth.vinit_c = 1; ++ switch (pipe->scl_data.lb_params.depth) { ++ case LB_PIXEL_DEPTH_30BPP: ++ input->scale_ratio_depth.lb_depth = 30; break; ++ case LB_PIXEL_DEPTH_36BPP: ++ input->scale_ratio_depth.lb_depth = 36; break; ++ default: ++ input->scale_ratio_depth.lb_depth = 24; break; ++ } ++ ++ ++ input->dest.vactive = pipe->stream->public.timing.v_addressable; ++ ++ input->dest.recout_width = pipe->scl_data.recout.width; ++ input->dest.recout_height = pipe->scl_data.recout.height; ++ ++ input->dest.full_recout_width = pipe->scl_data.recout.width; ++ input->dest.full_recout_height = pipe->scl_data.recout.height; ++ ++ input->dest.htotal = pipe->stream->public.timing.h_total; ++ input->dest.hblank_start = input->dest.htotal - pipe->stream->public.timing.h_front_porch; ++ input->dest.hblank_end = input->dest.hblank_start ++ - pipe->stream->public.timing.h_addressable ++ - pipe->stream->public.timing.h_border_left ++ - pipe->stream->public.timing.h_border_right; ++ ++ input->dest.vtotal = pipe->stream->public.timing.v_total; ++ input->dest.vblank_start = input->dest.vtotal - pipe->stream->public.timing.v_front_porch; ++ input->dest.vblank_end = input->dest.vblank_start ++ - pipe->stream->public.timing.v_addressable ++ - pipe->stream->public.timing.v_border_bottom ++ - pipe->stream->public.timing.v_border_top; ++ ++ input->dest.vsync_plus_back_porch = pipe->stream->public.timing.v_total ++ - pipe->stream->public.timing.v_addressable ++ - pipe->stream->public.timing.v_front_porch; ++ input->dest.pixel_rate_mhz = pipe->stream->public.timing.pix_clk_khz/1000.0; ++ input->dest.vstartup_start = pipe->pipe_dlg_param.vstartup_start; ++ input->dest.vupdate_offset = pipe->pipe_dlg_param.vupdate_offset; ++ input->dest.vupdate_offset = pipe->pipe_dlg_param.vupdate_offset; ++ input->dest.vupdate_width = pipe->pipe_dlg_param.vupdate_width; ++ ++} ++ ++static void dcn_bw_calc_rq_dlg_ttu( ++ const struct core_dc *dc, ++ const struct dcn_bw_internal_vars *v, ++ struct pipe_ctx *pipe) ++{ ++ struct display_mode_lib *dml = (struct display_mode_lib *)(&dc->dml); ++ struct _vcs_dpi_display_dlg_regs_st *dlg_regs = &pipe->dlg_regs; ++ struct _vcs_dpi_display_ttu_regs_st *ttu_regs = &pipe->ttu_regs; ++ struct _vcs_dpi_display_rq_regs_st *rq_regs = &pipe->rq_regs; ++ struct _vcs_dpi_display_rq_params_st rq_param = {0}; ++ struct _vcs_dpi_display_dlg_sys_params_st dlg_sys_param = {0}; ++ struct _vcs_dpi_display_e2e_pipe_params_st input = { { { 0 } } }; ++ float total_active_bw = 0; ++ float total_prefetch_bw = 0; ++ int total_flip_bytes = 0; ++ int i; ++ ++ for (i = 0; i < number_of_planes; i++) { ++ total_active_bw += v->read_bandwidth[i]; ++ total_prefetch_bw += v->prefetch_bandwidth[i]; ++ total_flip_bytes += v->total_immediate_flip_bytes[i]; ++ } ++ dlg_sys_param.total_flip_bw = v->return_bw - dcn_bw_max2(total_active_bw, total_prefetch_bw); ++ if (dlg_sys_param.total_flip_bw < 0.0) ++ dlg_sys_param.total_flip_bw = 0; ++ ++ dlg_sys_param.t_mclk_wm_us = v->dram_clock_change_watermark; ++ dlg_sys_param.t_sr_wm_us = v->stutter_enter_plus_exit_watermark; ++ dlg_sys_param.t_urg_wm_us = v->urgent_watermark; ++ dlg_sys_param.t_extra_us = v->urgent_extra_latency; ++ dlg_sys_param.deepsleep_dcfclk_mhz = v->dcf_clk_deep_sleep; ++ dlg_sys_param.total_flip_bytes = total_flip_bytes; ++ ++ pipe_ctx_to_e2e_pipe_params(pipe, &input.pipe); ++ input.clks_cfg.dcfclk_mhz = v->dcfclk; ++ input.clks_cfg.dispclk_mhz = v->dispclk; ++ input.clks_cfg.dppclk_mhz = v->dppclk; ++ input.clks_cfg.refclk_mhz = dc->res_pool->ref_clock_inKhz/1000; ++ input.clks_cfg.socclk_mhz = v->socclk; ++ input.clks_cfg.voltage = v->voltage_level; ++// dc->dml.logger = pool->base.logger; ++ ++ /*todo: soc->sr_enter_plus_exit_time??*/ ++ dlg_sys_param.t_srx_delay_us = dc->dcn_ip.dcfclk_cstate_latency / v->dcf_clk_deep_sleep; ++ ++ dml_rq_dlg_get_rq_params(dml, &rq_param, input.pipe.src); ++ extract_rq_regs(dml, rq_regs, rq_param); ++ dml_rq_dlg_get_dlg_params( ++ dml, ++ dlg_regs, ++ ttu_regs, ++ rq_param.dlg, ++ dlg_sys_param, ++ input, ++ true, ++ true, ++ v->pte_enable == dcn_bw_yes, ++ pipe->surface->public.flip_immediate); ++} ++ ++static void dcn_dml_wm_override( ++ const struct dcn_bw_internal_vars *v, ++ struct display_mode_lib *dml, ++ struct validate_context *context, ++ const struct resource_pool *pool) ++{ ++ int i, in_idx, active_count; ++ ++ struct _vcs_dpi_display_e2e_pipe_params_st *input = dm_alloc(pool->pipe_count * ++ sizeof(struct _vcs_dpi_display_e2e_pipe_params_st)); ++ struct wm { ++ double urgent; ++ struct _vcs_dpi_cstate_pstate_watermarks_st cpstate; ++ double pte_meta_urgent; ++ } a; ++ ++ ++ for (i = 0, in_idx = 0; i < pool->pipe_count; i++) { ++ struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i]; ++ ++ if (!pipe->stream || !pipe->surface) ++ continue; ++ ++ input[in_idx].clks_cfg.dcfclk_mhz = v->dcfclk; ++ input[in_idx].clks_cfg.dispclk_mhz = v->dispclk; ++ input[in_idx].clks_cfg.dppclk_mhz = v->dppclk; ++ input[in_idx].clks_cfg.refclk_mhz = pool->ref_clock_inKhz / 1000; ++ input[in_idx].clks_cfg.socclk_mhz = v->socclk; ++ input[in_idx].clks_cfg.voltage = v->voltage_level; ++ pipe_ctx_to_e2e_pipe_params(pipe, &input[in_idx].pipe); ++ dml_rq_dlg_get_rq_reg( ++ dml, ++ &pipe->rq_regs, ++ input[in_idx].pipe.src); ++ in_idx++; ++ } ++ active_count = in_idx; ++ ++ a.urgent = dml_wm_urgent_e2e(dml, input, active_count); ++ a.cpstate = dml_wm_cstate_pstate_e2e(dml, input, active_count); ++ a.pte_meta_urgent = dml_wm_pte_meta_urgent(dml, a.urgent); ++ ++ context->watermarks.a.cstate_pstate.cstate_exit_ns = ++ a.cpstate.cstate_exit_us * 1000; ++ context->watermarks.a.cstate_pstate.cstate_enter_plus_exit_ns = ++ a.cpstate.cstate_enter_plus_exit_us * 1000; ++ context->watermarks.a.cstate_pstate.pstate_change_ns = ++ a.cpstate.pstate_change_us * 1000; ++ context->watermarks.a.pte_meta_urgent_ns = a.pte_meta_urgent * 1000; ++ context->watermarks.a.urgent_ns = a.urgent * 1000; ++ context->watermarks.b = context->watermarks.a; ++ context->watermarks.c = context->watermarks.a; ++ context->watermarks.d = context->watermarks.a; ++ ++ ++ for (i = 0, in_idx = 0; i < pool->pipe_count; i++) { ++ struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i]; ++ ++ if (!pipe->stream || !pipe->surface) ++ continue; ++ ++ dml_rq_dlg_get_dlg_reg(dml, ++ &pipe->dlg_regs, ++ &pipe->ttu_regs, ++ input, active_count, ++ in_idx, ++ true, ++ true, ++ v->pte_enable == dcn_bw_yes, ++ pipe->surface->public.flip_immediate); ++ in_idx++; ++ } ++ dm_free(input); ++} ++ ++static void split_stream_across_pipes( ++ struct resource_context *res_ctx, ++ const struct resource_pool *pool, ++ struct pipe_ctx *primary_pipe, ++ struct pipe_ctx *secondary_pipe) ++{ ++ if (!primary_pipe->surface) ++ return; ++ ++ secondary_pipe->stream = primary_pipe->stream; ++ secondary_pipe->tg = primary_pipe->tg; ++ ++ secondary_pipe->mi = pool->mis[secondary_pipe->pipe_idx]; ++ secondary_pipe->ipp = pool->ipps[secondary_pipe->pipe_idx]; ++ secondary_pipe->xfm = pool->transforms[secondary_pipe->pipe_idx]; ++ secondary_pipe->opp = pool->opps[secondary_pipe->pipe_idx]; ++ if (primary_pipe->bottom_pipe) { ++ secondary_pipe->bottom_pipe = primary_pipe->bottom_pipe; ++ secondary_pipe->bottom_pipe->top_pipe = secondary_pipe; ++ } ++ primary_pipe->bottom_pipe = secondary_pipe; ++ secondary_pipe->top_pipe = primary_pipe; ++ secondary_pipe->surface = primary_pipe->surface; ++ secondary_pipe->pipe_dlg_param = primary_pipe->pipe_dlg_param; ++ ++ resource_build_scaling_params(primary_pipe); ++ resource_build_scaling_params(secondary_pipe); ++} ++ ++static void calc_wm_sets_and_perf_params( ++ struct validate_context *context, ++ struct dcn_bw_internal_vars *v) ++{ ++ /* Calculate set A last to keep internal var state consistent for required config */ ++ if (v->voltage_level < 2) { ++ v->fabric_and_dram_bandwidth_per_state[1] = v->fabric_and_dram_bandwidth_vnom0p8; ++ v->fabric_and_dram_bandwidth_per_state[0] = v->fabric_and_dram_bandwidth_vnom0p8; ++ v->fabric_and_dram_bandwidth = v->fabric_and_dram_bandwidth_vnom0p8; ++ dispclkdppclkdcfclk_deep_sleep_prefetch_parameters_watermarks_and_performance_calculation(v); ++ ++ context->watermarks.b.cstate_pstate.cstate_exit_ns = ++ v->stutter_exit_watermark * 1000; ++ context->watermarks.b.cstate_pstate.cstate_enter_plus_exit_ns = ++ v->stutter_enter_plus_exit_watermark * 1000; ++ context->watermarks.b.cstate_pstate.pstate_change_ns = ++ v->dram_clock_change_watermark * 1000; ++ context->watermarks.b.pte_meta_urgent_ns = v->ptemeta_urgent_watermark * 1000; ++ context->watermarks.b.urgent_ns = v->urgent_watermark * 1000; ++ ++ v->dcfclk_per_state[1] = v->dcfclkv_nom0p8; ++ v->dcfclk_per_state[0] = v->dcfclkv_nom0p8; ++ v->dcfclk = v->dcfclkv_nom0p8; ++ dispclkdppclkdcfclk_deep_sleep_prefetch_parameters_watermarks_and_performance_calculation(v); ++ ++ context->watermarks.c.cstate_pstate.cstate_exit_ns = ++ v->stutter_exit_watermark * 1000; ++ context->watermarks.c.cstate_pstate.cstate_enter_plus_exit_ns = ++ v->stutter_enter_plus_exit_watermark * 1000; ++ context->watermarks.c.cstate_pstate.pstate_change_ns = ++ v->dram_clock_change_watermark * 1000; ++ context->watermarks.c.pte_meta_urgent_ns = v->ptemeta_urgent_watermark * 1000; ++ context->watermarks.c.urgent_ns = v->urgent_watermark * 1000; ++ } ++ ++ if (v->voltage_level < 3) { ++ v->fabric_and_dram_bandwidth_per_state[2] = v->fabric_and_dram_bandwidth_vmax0p9; ++ v->fabric_and_dram_bandwidth_per_state[1] = v->fabric_and_dram_bandwidth_vmax0p9; ++ v->fabric_and_dram_bandwidth_per_state[0] = v->fabric_and_dram_bandwidth_vmax0p9; ++ v->fabric_and_dram_bandwidth = v->fabric_and_dram_bandwidth_vmax0p9; ++ v->dcfclk_per_state[2] = v->dcfclkv_max0p9; ++ v->dcfclk_per_state[1] = v->dcfclkv_max0p9; ++ v->dcfclk_per_state[0] = v->dcfclkv_max0p9; ++ v->dcfclk = v->dcfclkv_max0p9; ++ dispclkdppclkdcfclk_deep_sleep_prefetch_parameters_watermarks_and_performance_calculation(v); ++ ++ context->watermarks.d.cstate_pstate.cstate_exit_ns = ++ v->stutter_exit_watermark * 1000; ++ context->watermarks.d.cstate_pstate.cstate_enter_plus_exit_ns = ++ v->stutter_enter_plus_exit_watermark * 1000; ++ context->watermarks.d.cstate_pstate.pstate_change_ns = ++ v->dram_clock_change_watermark * 1000; ++ context->watermarks.d.pte_meta_urgent_ns = v->ptemeta_urgent_watermark * 1000; ++ context->watermarks.d.urgent_ns = v->urgent_watermark * 1000; ++ } ++ ++ v->fabric_and_dram_bandwidth_per_state[2] = v->fabric_and_dram_bandwidth_vnom0p8; ++ v->fabric_and_dram_bandwidth_per_state[1] = v->fabric_and_dram_bandwidth_vmid0p72; ++ v->fabric_and_dram_bandwidth_per_state[0] = v->fabric_and_dram_bandwidth_vmin0p65; ++ v->fabric_and_dram_bandwidth = v->fabric_and_dram_bandwidth_per_state[v->voltage_level]; ++ v->dcfclk_per_state[2] = v->dcfclkv_nom0p8; ++ v->dcfclk_per_state[1] = v->dcfclkv_mid0p72; ++ v->dcfclk_per_state[0] = v->dcfclkv_min0p65; ++ v->dcfclk = v->dcfclk_per_state[v->voltage_level]; ++ dispclkdppclkdcfclk_deep_sleep_prefetch_parameters_watermarks_and_performance_calculation(v); ++ ++ context->watermarks.a.cstate_pstate.cstate_exit_ns = ++ v->stutter_exit_watermark * 1000; ++ context->watermarks.a.cstate_pstate.cstate_enter_plus_exit_ns = ++ v->stutter_enter_plus_exit_watermark * 1000; ++ context->watermarks.a.cstate_pstate.pstate_change_ns = ++ v->dram_clock_change_watermark * 1000; ++ context->watermarks.a.pte_meta_urgent_ns = v->ptemeta_urgent_watermark * 1000; ++ context->watermarks.a.urgent_ns = v->urgent_watermark * 1000; ++ if (v->voltage_level >= 2) { ++ context->watermarks.b = context->watermarks.a; ++ context->watermarks.c = context->watermarks.a; ++ } ++ if (v->voltage_level >= 3) ++ context->watermarks.d = context->watermarks.a; ++} ++ ++static void dcn_bw_apply_registry_override(struct core_dc *dc) ++{ ++ kernel_fpu_begin(); ++ if (dc->public.debug.sr_exit_time_ns) ++ dc->dcn_soc.sr_exit_time = dc->public.debug.sr_exit_time_ns / 1000.0; ++ if (dc->public.debug.sr_enter_plus_exit_time_ns) ++ dc->dcn_soc.sr_enter_plus_exit_time = ++ dc->public.debug.sr_enter_plus_exit_time_ns / 1000.0; ++ if (dc->public.debug.urgent_latency_ns) ++ dc->dcn_soc.urgent_latency = dc->public.debug.urgent_latency_ns / 1000.0; ++ if (dc->public.debug.percent_of_ideal_drambw) ++ dc->dcn_soc.percent_of_ideal_drambw_received_after_urg_latency = ++ dc->public.debug.percent_of_ideal_drambw; ++ if (dc->public.debug.dram_clock_change_latency_ns) ++ dc->dcn_soc.dram_clock_change_latency = ++ dc->public.debug.dram_clock_change_latency_ns / 1000.0; ++ kernel_fpu_end(); ++} ++ ++bool dcn_validate_bandwidth( ++ const struct core_dc *dc, ++ struct validate_context *context) ++{ ++ const struct resource_pool *pool = dc->res_pool; ++ struct dcn_bw_internal_vars *v = &context->dcn_bw_vars; ++ int i, input_idx; ++ int vesa_sync_start, asic_blank_end, asic_blank_start; ++ ++ dcn_bw_apply_registry_override(DC_TO_CORE(&dc->public)); ++ ++ memset(v, 0, sizeof(*v)); ++ kernel_fpu_begin(); ++ v->sr_exit_time = dc->dcn_soc.sr_exit_time; ++ v->sr_enter_plus_exit_time = dc->dcn_soc.sr_enter_plus_exit_time; ++ v->urgent_latency = dc->dcn_soc.urgent_latency; ++ v->write_back_latency = dc->dcn_soc.write_back_latency; ++ v->percent_of_ideal_drambw_received_after_urg_latency = ++ dc->dcn_soc.percent_of_ideal_drambw_received_after_urg_latency; ++ ++ v->dcfclkv_min0p65 = dc->dcn_soc.dcfclkv_min0p65; ++ v->dcfclkv_mid0p72 = dc->dcn_soc.dcfclkv_mid0p72; ++ v->dcfclkv_nom0p8 = dc->dcn_soc.dcfclkv_nom0p8; ++ v->dcfclkv_max0p9 = dc->dcn_soc.dcfclkv_max0p9; ++ ++ v->max_dispclk_vmin0p65 = dc->dcn_soc.max_dispclk_vmin0p65; ++ v->max_dispclk_vmid0p72 = dc->dcn_soc.max_dispclk_vmid0p72; ++ v->max_dispclk_vnom0p8 = dc->dcn_soc.max_dispclk_vnom0p8; ++ v->max_dispclk_vmax0p9 = dc->dcn_soc.max_dispclk_vmax0p9; ++ ++ v->max_dppclk_vmin0p65 = dc->dcn_soc.max_dppclk_vmin0p65; ++ v->max_dppclk_vmid0p72 = dc->dcn_soc.max_dppclk_vmid0p72; ++ v->max_dppclk_vnom0p8 = dc->dcn_soc.max_dppclk_vnom0p8; ++ v->max_dppclk_vmax0p9 = dc->dcn_soc.max_dppclk_vmax0p9; ++ ++ v->socclk = dc->dcn_soc.socclk; ++ ++ v->fabric_and_dram_bandwidth_vmin0p65 = dc->dcn_soc.fabric_and_dram_bandwidth_vmin0p65; ++ v->fabric_and_dram_bandwidth_vmid0p72 = dc->dcn_soc.fabric_and_dram_bandwidth_vmid0p72; ++ v->fabric_and_dram_bandwidth_vnom0p8 = dc->dcn_soc.fabric_and_dram_bandwidth_vnom0p8; ++ v->fabric_and_dram_bandwidth_vmax0p9 = dc->dcn_soc.fabric_and_dram_bandwidth_vmax0p9; ++ ++ v->phyclkv_min0p65 = dc->dcn_soc.phyclkv_min0p65; ++ v->phyclkv_mid0p72 = dc->dcn_soc.phyclkv_mid0p72; ++ v->phyclkv_nom0p8 = dc->dcn_soc.phyclkv_nom0p8; ++ v->phyclkv_max0p9 = dc->dcn_soc.phyclkv_max0p9; ++ ++ v->downspreading = dc->dcn_soc.downspreading; ++ v->round_trip_ping_latency_cycles = dc->dcn_soc.round_trip_ping_latency_cycles; ++ v->urgent_out_of_order_return_per_channel = dc->dcn_soc.urgent_out_of_order_return_per_channel; ++ v->number_of_channels = dc->dcn_soc.number_of_channels; ++ v->vmm_page_size = dc->dcn_soc.vmm_page_size; ++ v->dram_clock_change_latency = dc->dcn_soc.dram_clock_change_latency; ++ v->return_bus_width = dc->dcn_soc.return_bus_width; ++ ++ v->rob_buffer_size_in_kbyte = dc->dcn_ip.rob_buffer_size_in_kbyte; ++ v->det_buffer_size_in_kbyte = dc->dcn_ip.det_buffer_size_in_kbyte; ++ v->dpp_output_buffer_pixels = dc->dcn_ip.dpp_output_buffer_pixels; ++ v->opp_output_buffer_lines = dc->dcn_ip.opp_output_buffer_lines; ++ v->pixel_chunk_size_in_kbyte = dc->dcn_ip.pixel_chunk_size_in_kbyte; ++ v->pte_enable = dc->dcn_ip.pte_enable; ++ v->pte_chunk_size = dc->dcn_ip.pte_chunk_size; ++ v->meta_chunk_size = dc->dcn_ip.meta_chunk_size; ++ v->writeback_chunk_size = dc->dcn_ip.writeback_chunk_size; ++ v->odm_capability = dc->dcn_ip.odm_capability; ++ v->dsc_capability = dc->dcn_ip.dsc_capability; ++ v->line_buffer_size = dc->dcn_ip.line_buffer_size; ++ v->is_line_buffer_bpp_fixed = dc->dcn_ip.is_line_buffer_bpp_fixed; ++ v->line_buffer_fixed_bpp = dc->dcn_ip.line_buffer_fixed_bpp; ++ v->max_line_buffer_lines = dc->dcn_ip.max_line_buffer_lines; ++ v->writeback_luma_buffer_size = dc->dcn_ip.writeback_luma_buffer_size; ++ v->writeback_chroma_buffer_size = dc->dcn_ip.writeback_chroma_buffer_size; ++ v->max_num_dpp = dc->dcn_ip.max_num_dpp; ++ v->max_num_writeback = dc->dcn_ip.max_num_writeback; ++ v->max_dchub_topscl_throughput = dc->dcn_ip.max_dchub_topscl_throughput; ++ v->max_pscl_tolb_throughput = dc->dcn_ip.max_pscl_tolb_throughput; ++ v->max_lb_tovscl_throughput = dc->dcn_ip.max_lb_tovscl_throughput; ++ v->max_vscl_tohscl_throughput = dc->dcn_ip.max_vscl_tohscl_throughput; ++ v->max_hscl_ratio = dc->dcn_ip.max_hscl_ratio; ++ v->max_vscl_ratio = dc->dcn_ip.max_vscl_ratio; ++ v->max_hscl_taps = dc->dcn_ip.max_hscl_taps; ++ v->max_vscl_taps = dc->dcn_ip.max_vscl_taps; ++ v->under_scan_factor = dc->dcn_ip.under_scan_factor; ++ v->pte_buffer_size_in_requests = dc->dcn_ip.pte_buffer_size_in_requests; ++ v->dispclk_ramping_margin = dc->dcn_ip.dispclk_ramping_margin; ++ v->max_inter_dcn_tile_repeaters = dc->dcn_ip.max_inter_dcn_tile_repeaters; ++ v->can_vstartup_lines_exceed_vsync_plus_back_porch_lines_minus_one = ++ dc->dcn_ip.can_vstartup_lines_exceed_vsync_plus_back_porch_lines_minus_one; ++ v->bug_forcing_luma_and_chroma_request_to_same_size_fixed = ++ dc->dcn_ip.bug_forcing_luma_and_chroma_request_to_same_size_fixed; ++ ++ v->voltage[5] = dcn_bw_no_support; ++ v->voltage[4] = dcn_bw_v_max0p9; ++ v->voltage[3] = dcn_bw_v_max0p9; ++ v->voltage[2] = dcn_bw_v_nom0p8; ++ v->voltage[1] = dcn_bw_v_mid0p72; ++ v->voltage[0] = dcn_bw_v_min0p65; ++ v->fabric_and_dram_bandwidth_per_state[5] = v->fabric_and_dram_bandwidth_vmax0p9; ++ v->fabric_and_dram_bandwidth_per_state[4] = v->fabric_and_dram_bandwidth_vmax0p9; ++ v->fabric_and_dram_bandwidth_per_state[3] = v->fabric_and_dram_bandwidth_vmax0p9; ++ v->fabric_and_dram_bandwidth_per_state[2] = v->fabric_and_dram_bandwidth_vnom0p8; ++ v->fabric_and_dram_bandwidth_per_state[1] = v->fabric_and_dram_bandwidth_vmid0p72; ++ v->fabric_and_dram_bandwidth_per_state[0] = v->fabric_and_dram_bandwidth_vmin0p65; ++ v->dcfclk_per_state[5] = v->dcfclkv_max0p9; ++ v->dcfclk_per_state[4] = v->dcfclkv_max0p9; ++ v->dcfclk_per_state[3] = v->dcfclkv_max0p9; ++ v->dcfclk_per_state[2] = v->dcfclkv_nom0p8; ++ v->dcfclk_per_state[1] = v->dcfclkv_mid0p72; ++ v->dcfclk_per_state[0] = v->dcfclkv_min0p65; ++ v->max_dispclk[5] = v->max_dispclk_vmax0p9; ++ v->max_dispclk[4] = v->max_dispclk_vmax0p9; ++ v->max_dispclk[3] = v->max_dispclk_vmax0p9; ++ v->max_dispclk[2] = v->max_dispclk_vnom0p8; ++ v->max_dispclk[1] = v->max_dispclk_vmid0p72; ++ v->max_dispclk[0] = v->max_dispclk_vmin0p65; ++ v->max_dppclk[5] = v->max_dppclk_vmax0p9; ++ v->max_dppclk[4] = v->max_dppclk_vmax0p9; ++ v->max_dppclk[3] = v->max_dppclk_vmax0p9; ++ v->max_dppclk[2] = v->max_dppclk_vnom0p8; ++ v->max_dppclk[1] = v->max_dppclk_vmid0p72; ++ v->max_dppclk[0] = v->max_dppclk_vmin0p65; ++ v->phyclk_per_state[5] = v->phyclkv_max0p9; ++ v->phyclk_per_state[4] = v->phyclkv_max0p9; ++ v->phyclk_per_state[3] = v->phyclkv_max0p9; ++ v->phyclk_per_state[2] = v->phyclkv_nom0p8; ++ v->phyclk_per_state[1] = v->phyclkv_mid0p72; ++ v->phyclk_per_state[0] = v->phyclkv_min0p65; ++ ++ if (dc->public.debug.use_max_voltage) { ++ v->max_dppclk[1] = v->max_dppclk_vnom0p8; ++ v->max_dppclk[0] = v->max_dppclk_vnom0p8; ++ } ++ ++ if (v->voltage_override == dcn_bw_v_max0p9) { ++ v->voltage_override_level = number_of_states - 1; ++ } else if (v->voltage_override == dcn_bw_v_nom0p8) { ++ v->voltage_override_level = number_of_states - 2; ++ } else if (v->voltage_override == dcn_bw_v_mid0p72) { ++ v->voltage_override_level = number_of_states - 3; ++ } else { ++ v->voltage_override_level = 0; ++ } ++ v->synchronized_vblank = dcn_bw_no; ++ v->ta_pscalculation = dcn_bw_override; ++ v->allow_different_hratio_vratio = dcn_bw_yes; ++ ++ ++ for (i = 0, input_idx = 0; i < pool->pipe_count; i++) { ++ struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i]; ++ ++ if (!pipe->stream) ++ continue; ++ /* skip all but first of split pipes */ ++ if (pipe->top_pipe && pipe->top_pipe->surface == pipe->surface) ++ continue; ++ ++ v->underscan_output[input_idx] = false; /* taken care of in recout already*/ ++ v->interlace_output[input_idx] = false; ++ ++ v->htotal[input_idx] = pipe->stream->public.timing.h_total; ++ v->vtotal[input_idx] = pipe->stream->public.timing.v_total; ++ v->v_sync_plus_back_porch[input_idx] = pipe->stream->public.timing.v_total ++ - pipe->stream->public.timing.v_addressable ++ - pipe->stream->public.timing.v_front_porch; ++ v->vactive[input_idx] = pipe->stream->public.timing.v_addressable; ++ v->pixel_clock[input_idx] = pipe->stream->public.timing.pix_clk_khz / 1000.0f; ++ ++ ++ if (!pipe->surface){ ++ v->dcc_enable[input_idx] = dcn_bw_yes; ++ v->source_pixel_format[input_idx] = dcn_bw_rgb_sub_32; ++ v->source_surface_mode[input_idx] = dcn_bw_sw_4_kb_s; ++ v->lb_bit_per_pixel[input_idx] = 30; ++ v->viewport_width[input_idx] = pipe->stream->public.timing.h_addressable; ++ v->viewport_height[input_idx] = pipe->stream->public.timing.v_addressable; ++ v->scaler_rec_out_width[input_idx] = pipe->stream->public.timing.h_addressable; ++ v->scaler_recout_height[input_idx] = pipe->stream->public.timing.v_addressable; ++ v->override_hta_ps[input_idx] = 1; ++ v->override_vta_ps[input_idx] = 1; ++ v->override_hta_pschroma[input_idx] = 1; ++ v->override_vta_pschroma[input_idx] = 1; ++ v->source_scan[input_idx] = dcn_bw_hor; ++ ++ } else { ++ v->viewport_height[input_idx] = pipe->scl_data.viewport.height; ++ v->viewport_width[input_idx] = pipe->scl_data.viewport.width; ++ v->scaler_rec_out_width[input_idx] = pipe->scl_data.recout.width; ++ v->scaler_recout_height[input_idx] = pipe->scl_data.recout.height; ++ if (pipe->bottom_pipe && pipe->bottom_pipe->surface == pipe->surface) { ++ if (pipe->surface->public.rotation % 2 == 0) { ++ int viewport_end = pipe->scl_data.viewport.width ++ + pipe->scl_data.viewport.x; ++ int viewport_b_end = pipe->bottom_pipe->scl_data.viewport.width ++ + pipe->bottom_pipe->scl_data.viewport.x; ++ ++ if (viewport_end > viewport_b_end) ++ v->viewport_width[input_idx] = viewport_end ++ - pipe->bottom_pipe->scl_data.viewport.x; ++ else ++ v->viewport_width[input_idx] = viewport_b_end ++ - pipe->scl_data.viewport.x; ++ } else { ++ int viewport_end = pipe->scl_data.viewport.height ++ + pipe->scl_data.viewport.y; ++ int viewport_b_end = pipe->bottom_pipe->scl_data.viewport.height ++ + pipe->bottom_pipe->scl_data.viewport.y; ++ ++ if (viewport_end > viewport_b_end) ++ v->viewport_height[input_idx] = viewport_end ++ - pipe->bottom_pipe->scl_data.viewport.y; ++ else ++ v->viewport_height[input_idx] = viewport_b_end ++ - pipe->scl_data.viewport.y; ++ } ++ v->scaler_rec_out_width[input_idx] = pipe->scl_data.recout.width ++ + pipe->bottom_pipe->scl_data.recout.width; ++ } ++ ++ v->dcc_enable[input_idx] = pipe->surface->public.dcc.enable ? dcn_bw_yes : dcn_bw_no; ++ v->source_pixel_format[input_idx] = tl_pixel_format_to_bw_defs( ++ pipe->surface->public.format); ++ v->source_surface_mode[input_idx] = tl_sw_mode_to_bw_defs( ++ pipe->surface->public.tiling_info.gfx9.swizzle); ++ v->lb_bit_per_pixel[input_idx] = tl_lb_bpp_to_int(pipe->scl_data.lb_params.depth); ++ v->override_hta_ps[input_idx] = pipe->scl_data.taps.h_taps; ++ v->override_vta_ps[input_idx] = pipe->scl_data.taps.v_taps; ++ v->override_hta_pschroma[input_idx] = pipe->scl_data.taps.h_taps_c; ++ v->override_vta_pschroma[input_idx] = pipe->scl_data.taps.v_taps_c; ++ v->source_scan[input_idx] = (pipe->surface->public.rotation % 2) ? dcn_bw_vert : dcn_bw_hor; ++ } ++ if (v->is_line_buffer_bpp_fixed == dcn_bw_yes) ++ v->lb_bit_per_pixel[input_idx] = v->line_buffer_fixed_bpp; ++ v->dcc_rate[input_idx] = 1; /*TODO: Worst case? does this change?*/ ++ v->output_format[input_idx] = dcn_bw_444; ++ v->output[input_idx] = dcn_bw_dp; ++ ++ input_idx++; ++ } ++ v->number_of_active_planes = input_idx; ++ ++ scaler_settings_calculation(v); ++ mode_support_and_system_configuration(v); ++ ++ if (v->voltage_level != 5) { ++ float bw_consumed = v->total_bandwidth_consumed_gbyte_per_second; ++ if (bw_consumed < v->fabric_and_dram_bandwidth_vmin0p65) ++ bw_consumed = v->fabric_and_dram_bandwidth_vmin0p65; ++ else if (bw_consumed < v->fabric_and_dram_bandwidth_vmid0p72) ++ bw_consumed = v->fabric_and_dram_bandwidth_vmid0p72; ++ else if (bw_consumed < v->fabric_and_dram_bandwidth_vnom0p8) ++ bw_consumed = v->fabric_and_dram_bandwidth_vnom0p8; ++ else ++ bw_consumed = v->fabric_and_dram_bandwidth_vmax0p9; ++ ++ display_pipe_configuration(v); ++ calc_wm_sets_and_perf_params(context, v); ++ context->fclk_khz = (int)(bw_consumed * 1000000 / ++ (ddr4_dram_factor_single_Channel * v->number_of_channels)); ++ context->dram_ccm_us = (int)(v->dram_clock_change_margin); ++ context->min_active_dram_ccm_us = (int)(v->min_active_dram_clock_change_margin); ++ context->dcfclk_deep_sleep_khz = (int)(v->dcf_clk_deep_sleep * 1000); ++ context->dcfclk_khz = (int)(v->dcfclk * 1000); ++ context->dispclk_khz = (int)(v->dispclk * 1000); ++ if (dc->public.debug.max_disp_clk == true) ++ context->dispclk_khz = (int)(dc->dcn_soc.max_dispclk_vmax0p9 * 1000); ++ context->dppclk_khz = (int)(v->dppclk * 1000); ++ context->dppclk_div = (int)(v->dispclk_dppclk_ratio) == 2; ++ ++ for (i = 0, input_idx = 0; i < pool->pipe_count; i++) { ++ struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i]; ++ ++ /* skip inactive pipe */ ++ if (!pipe->stream) ++ continue; ++ /* skip all but first of split pipes */ ++ if (pipe->top_pipe && pipe->top_pipe->surface == pipe->surface) ++ continue; ++ ++ pipe->pipe_dlg_param.vupdate_width = v->v_update_width[input_idx]; ++ pipe->pipe_dlg_param.vupdate_offset = v->v_update_offset[input_idx]; ++ pipe->pipe_dlg_param.vready_offset = v->v_ready_offset[input_idx]; ++ pipe->pipe_dlg_param.vstartup_start = v->v_startup[input_idx]; ++ ++ pipe->pipe_dlg_param.htotal = pipe->stream->public.timing.h_total; ++ pipe->pipe_dlg_param.vtotal = pipe->stream->public.timing.v_total; ++ vesa_sync_start = pipe->stream->public.timing.v_addressable + ++ pipe->stream->public.timing.v_border_bottom + ++ pipe->stream->public.timing.v_front_porch; ++ ++ asic_blank_end = (pipe->stream->public.timing.v_total - ++ vesa_sync_start - ++ pipe->stream->public.timing.v_border_top) ++ * (pipe->stream->public.timing.flags.INTERLACE ? 1 : 0); ++ ++ asic_blank_start = asic_blank_end + ++ (pipe->stream->public.timing.v_border_top + ++ pipe->stream->public.timing.v_addressable + ++ pipe->stream->public.timing.v_border_bottom) ++ * (pipe->stream->public.timing.flags.INTERLACE ? 1 : 0); ++ ++ pipe->pipe_dlg_param.vblank_start = asic_blank_start; ++ pipe->pipe_dlg_param.vblank_end = asic_blank_end; ++ ++ if (pipe->surface) { ++ struct pipe_ctx *hsplit_pipe = pipe->bottom_pipe; ++ ++ if (v->dpp_per_plane[input_idx] == 2 || ++ (pipe->stream->public.timing.timing_3d_format == TIMING_3D_FORMAT_TOP_AND_BOTTOM || ++ pipe->stream->public.timing.timing_3d_format == TIMING_3D_FORMAT_SIDE_BY_SIDE)) { ++ if (hsplit_pipe && hsplit_pipe->surface == pipe->surface) { ++ /* update previously split pipe */ ++ hsplit_pipe->pipe_dlg_param.vupdate_width = v->v_update_width[input_idx]; ++ hsplit_pipe->pipe_dlg_param.vupdate_offset = v->v_update_offset[input_idx]; ++ hsplit_pipe->pipe_dlg_param.vready_offset = v->v_ready_offset[input_idx]; ++ hsplit_pipe->pipe_dlg_param.vstartup_start = v->v_startup[input_idx]; ++ ++ hsplit_pipe->pipe_dlg_param.htotal = pipe->stream->public.timing.h_total; ++ hsplit_pipe->pipe_dlg_param.vtotal = pipe->stream->public.timing.v_total; ++ hsplit_pipe->pipe_dlg_param.vblank_start = pipe->pipe_dlg_param.vblank_start; ++ hsplit_pipe->pipe_dlg_param.vblank_end = pipe->pipe_dlg_param.vblank_end; ++ } else { ++ /* pipe not split previously needs split */ ++ hsplit_pipe = find_idle_secondary_pipe(&context->res_ctx, pool); ++ ASSERT(hsplit_pipe); ++ split_stream_across_pipes( ++ &context->res_ctx, pool, ++ pipe, hsplit_pipe); ++ } ++ ++ dcn_bw_calc_rq_dlg_ttu(dc, v, hsplit_pipe); ++ } else if (hsplit_pipe && hsplit_pipe->surface == pipe->surface) { ++ /* merge previously split pipe */ ++ if (pipe->bottom_pipe->bottom_pipe) ++ pipe->bottom_pipe->bottom_pipe->top_pipe = pipe; ++ memset(pipe->bottom_pipe, 0, sizeof(*pipe->bottom_pipe)); ++ pipe->bottom_pipe = pipe->bottom_pipe->bottom_pipe; ++ resource_build_scaling_params(pipe); ++ } ++ /* for now important to do this after pipe split for building e2e params */ ++ dcn_bw_calc_rq_dlg_ttu(dc, v, pipe); ++ } ++ ++ input_idx++; ++ } ++ if (dc->public.debug.use_dml_wm) ++ dcn_dml_wm_override(v, (struct display_mode_lib *) ++ &dc->dml, context, pool); ++ } ++ ++ kernel_fpu_end(); ++ return v->voltage_level != 5; ++} ++ ++unsigned int dcn_find_normalized_clock_vdd_Level( ++ const struct core_dc *dc, ++ enum dm_pp_clock_type clocks_type, ++ int clocks_in_khz) ++{ ++ int vdd_level = dcn_bw_v_min0p65; ++ ++ if (clocks_in_khz == 0)/*todo some clock not in the considerations*/ ++ return vdd_level; ++ ++ switch (clocks_type) { ++ case DM_PP_CLOCK_TYPE_DISPLAY_CLK: ++ if (clocks_in_khz > dc->dcn_soc.max_dispclk_vmax0p9*1000) { ++ vdd_level = dcn_bw_v_max0p91; ++ BREAK_TO_DEBUGGER(); ++ } else if (clocks_in_khz > dc->dcn_soc.max_dispclk_vnom0p8*1000) { ++ vdd_level = dcn_bw_v_max0p9; ++ } else if (clocks_in_khz > dc->dcn_soc.max_dispclk_vmid0p72*1000) { ++ vdd_level = dcn_bw_v_nom0p8; ++ } else if (clocks_in_khz > dc->dcn_soc.max_dispclk_vmin0p65*1000) { ++ vdd_level = dcn_bw_v_mid0p72; ++ } else ++ vdd_level = dcn_bw_v_min0p65; ++ break; ++ case DM_PP_CLOCK_TYPE_DISPLAYPHYCLK: ++ if (clocks_in_khz > dc->dcn_soc.phyclkv_max0p9*1000) { ++ vdd_level = dcn_bw_v_max0p91; ++ BREAK_TO_DEBUGGER(); ++ } else if (clocks_in_khz > dc->dcn_soc.phyclkv_nom0p8*1000) { ++ vdd_level = dcn_bw_v_max0p9; ++ } else if (clocks_in_khz > dc->dcn_soc.phyclkv_mid0p72*1000) { ++ vdd_level = dcn_bw_v_nom0p8; ++ } else if (clocks_in_khz > dc->dcn_soc.phyclkv_min0p65*1000) { ++ vdd_level = dcn_bw_v_mid0p72; ++ } else ++ vdd_level = dcn_bw_v_min0p65; ++ break; ++ ++ case DM_PP_CLOCK_TYPE_DPPCLK: ++ if (clocks_in_khz > dc->dcn_soc.max_dppclk_vmax0p9*1000) { ++ vdd_level = dcn_bw_v_max0p91; ++ BREAK_TO_DEBUGGER(); ++ } else if (clocks_in_khz > dc->dcn_soc.max_dppclk_vnom0p8*1000) { ++ vdd_level = dcn_bw_v_max0p9; ++ } else if (clocks_in_khz > dc->dcn_soc.max_dppclk_vmid0p72*1000) { ++ vdd_level = dcn_bw_v_nom0p8; ++ } else if (clocks_in_khz > dc->dcn_soc.max_dppclk_vmin0p65*1000) { ++ vdd_level = dcn_bw_v_mid0p72; ++ } else ++ vdd_level = dcn_bw_v_min0p65; ++ break; ++ ++ case DM_PP_CLOCK_TYPE_MEMORY_CLK: ++ { ++ unsigned factor = (ddr4_dram_factor_single_Channel * dc->dcn_soc.number_of_channels); ++ if (clocks_in_khz > dc->dcn_soc.fabric_and_dram_bandwidth_vmax0p9*1000000/factor) { ++ vdd_level = dcn_bw_v_max0p91; ++ BREAK_TO_DEBUGGER(); ++ } else if (clocks_in_khz > dc->dcn_soc.fabric_and_dram_bandwidth_vnom0p8*1000000/factor) { ++ vdd_level = dcn_bw_v_max0p9; ++ } else if (clocks_in_khz > dc->dcn_soc.fabric_and_dram_bandwidth_vmid0p72*1000000/factor) { ++ vdd_level = dcn_bw_v_nom0p8; ++ } else if (clocks_in_khz > dc->dcn_soc.fabric_and_dram_bandwidth_vmin0p65*1000000/factor) { ++ vdd_level = dcn_bw_v_mid0p72; ++ } else ++ vdd_level = dcn_bw_v_min0p65; ++ } ++ break; ++ ++ case DM_PP_CLOCK_TYPE_DCFCLK: ++ if (clocks_in_khz > dc->dcn_soc.dcfclkv_max0p9*1000) { ++ vdd_level = dcn_bw_v_max0p91; ++ BREAK_TO_DEBUGGER(); ++ } else if (clocks_in_khz > dc->dcn_soc.dcfclkv_nom0p8*1000) { ++ vdd_level = dcn_bw_v_max0p9; ++ } else if (clocks_in_khz > dc->dcn_soc.dcfclkv_mid0p72*1000) { ++ vdd_level = dcn_bw_v_nom0p8; ++ } else if (clocks_in_khz > dc->dcn_soc.dcfclkv_min0p65*1000) { ++ vdd_level = dcn_bw_v_mid0p72; ++ } else ++ vdd_level = dcn_bw_v_min0p65; ++ break; ++ ++ default: ++ break; ++ } ++ return vdd_level; ++} ++ ++unsigned int dcn_find_dcfclk_suits_all( ++ const struct core_dc *dc, ++ struct clocks_value *clocks) ++{ ++ unsigned vdd_level, vdd_level_temp; ++ unsigned dcf_clk; ++ ++ /*find a common supported voltage level*/ ++ vdd_level = dcn_find_normalized_clock_vdd_Level( ++ dc, DM_PP_CLOCK_TYPE_DISPLAY_CLK, clocks->dispclk_in_khz); ++ vdd_level_temp = dcn_find_normalized_clock_vdd_Level( ++ dc, DM_PP_CLOCK_TYPE_DISPLAYPHYCLK, clocks->phyclk_in_khz); ++ ++ vdd_level = dcn_bw_max(vdd_level, vdd_level_temp); ++ vdd_level_temp = dcn_find_normalized_clock_vdd_Level( ++ dc, DM_PP_CLOCK_TYPE_DPPCLK, clocks->dppclk_in_khz); ++ vdd_level = dcn_bw_max(vdd_level, vdd_level_temp); ++ ++ vdd_level_temp = dcn_find_normalized_clock_vdd_Level( ++ dc, DM_PP_CLOCK_TYPE_MEMORY_CLK, clocks->dcfclock_in_khz); ++ vdd_level = dcn_bw_max(vdd_level, vdd_level_temp); ++ vdd_level_temp = dcn_find_normalized_clock_vdd_Level( ++ dc, DM_PP_CLOCK_TYPE_DCFCLK, clocks->dcfclock_in_khz); ++ ++ /*find that level conresponding dcfclk*/ ++ vdd_level = dcn_bw_max(vdd_level, vdd_level_temp); ++ if (vdd_level == dcn_bw_v_max0p91) { ++ BREAK_TO_DEBUGGER(); ++ dcf_clk = dc->dcn_soc.dcfclkv_max0p9*1000; ++ } else if (vdd_level == dcn_bw_v_max0p9) ++ dcf_clk = dc->dcn_soc.dcfclkv_max0p9*1000; ++ else if (vdd_level == dcn_bw_v_nom0p8) ++ dcf_clk = dc->dcn_soc.dcfclkv_nom0p8*1000; ++ else if (vdd_level == dcn_bw_v_mid0p72) ++ dcf_clk = dc->dcn_soc.dcfclkv_mid0p72*1000; ++ else ++ dcf_clk = dc->dcn_soc.dcfclkv_min0p65*1000; ++ ++ dm_logger_write(dc->ctx->logger, LOG_HW_MARKS, ++ "\tdcf_clk for voltage = %d\n", dcf_clk); ++ return dcf_clk; ++} ++ ++void dcn_bw_update_from_pplib(struct core_dc *dc) ++{ ++ struct dc_context *ctx = dc->ctx; ++ struct dm_pp_clock_levels_with_latency clks = {0}; ++ struct dm_pp_clock_levels_with_voltage clks2 = {0}; ++ ++ kernel_fpu_begin(); ++ dc->dcn_soc.number_of_channels = dc->ctx->asic_id.vram_width / ddr4_dram_width; ++ ASSERT(dc->dcn_soc.number_of_channels && dc->dcn_soc.number_of_channels < 3); ++ if (dc->dcn_soc.number_of_channels == 0)/*old sbios bug*/ ++ dc->dcn_soc.number_of_channels = 2; ++ ++ if (dm_pp_get_clock_levels_by_type_with_voltage( ++ ctx, DM_PP_CLOCK_TYPE_DISPLAY_CLK, &clks2) && ++ clks2.num_levels >= 3) { ++ dc->dcn_soc.max_dispclk_vmin0p65 = clks2.data[0].clocks_in_khz / 1000.0; ++ dc->dcn_soc.max_dispclk_vmid0p72 = clks2.data[clks2.num_levels - 3].clocks_in_khz / 1000.0; ++ dc->dcn_soc.max_dispclk_vnom0p8 = clks2.data[clks2.num_levels - 2].clocks_in_khz / 1000.0; ++ dc->dcn_soc.max_dispclk_vmax0p9 = clks2.data[clks2.num_levels - 1].clocks_in_khz / 1000.0; ++ } else ++ BREAK_TO_DEBUGGER(); ++/* ++ if (dm_pp_get_clock_levels_by_type_with_latency( ++ ctx, DM_PP_CLOCK_TYPE_MEMORY_CLK, &clks) && ++ clks.num_levels != 0) { ++ //this is to get DRAM data_rate ++ //FabricAndDRAMBandwidth = min(64*FCLK , Data rate * single_Channel_Width * number of channels); ++ }*/ ++ if (dm_pp_get_clock_levels_by_type_with_latency( ++ ctx, DM_PP_CLOCK_TYPE_FCLK, &clks) && ++ clks.num_levels != 0) { ++ ASSERT(clks.num_levels >= 3); ++ dc->dcn_soc.fabric_and_dram_bandwidth_vmin0p65 = dc->dcn_soc.number_of_channels * ++ (clks.data[0].clocks_in_khz / 1000.0) * ddr4_dram_factor_single_Channel / 1000.0; ++ if (clks.num_levels > 2) { ++ dc->dcn_soc.fabric_and_dram_bandwidth_vmid0p72 = dc->dcn_soc.number_of_channels * ++ (clks.data[clks.num_levels - 3].clocks_in_khz / 1000.0) * ddr4_dram_factor_single_Channel / 1000.0; ++ } else { ++ dc->dcn_soc.fabric_and_dram_bandwidth_vmid0p72 = dc->dcn_soc.number_of_channels * ++ (clks.data[clks.num_levels - 2].clocks_in_khz / 1000.0) * ddr4_dram_factor_single_Channel / 1000.0; ++ } ++ dc->dcn_soc.fabric_and_dram_bandwidth_vnom0p8 = dc->dcn_soc.number_of_channels * ++ (clks.data[clks.num_levels - 2].clocks_in_khz / 1000.0) * ddr4_dram_factor_single_Channel / 1000.0; ++ dc->dcn_soc.fabric_and_dram_bandwidth_vmax0p9 = dc->dcn_soc.number_of_channels * ++ (clks.data[clks.num_levels - 1].clocks_in_khz / 1000.0) * ddr4_dram_factor_single_Channel / 1000.0; ++ } else ++ BREAK_TO_DEBUGGER(); ++ if (dm_pp_get_clock_levels_by_type_with_latency( ++ ctx, DM_PP_CLOCK_TYPE_DCFCLK, &clks) && ++ clks.num_levels >= 3) { ++ dc->dcn_soc.dcfclkv_min0p65 = clks.data[0].clocks_in_khz / 1000.0; ++ dc->dcn_soc.dcfclkv_mid0p72 = clks.data[clks.num_levels - 3].clocks_in_khz / 1000.0; ++ dc->dcn_soc.dcfclkv_nom0p8 = clks.data[clks.num_levels - 2].clocks_in_khz / 1000.0; ++ dc->dcn_soc.dcfclkv_max0p9 = clks.data[clks.num_levels - 1].clocks_in_khz / 1000.0; ++ } else ++ BREAK_TO_DEBUGGER(); ++ if (dm_pp_get_clock_levels_by_type_with_voltage( ++ ctx, DM_PP_CLOCK_TYPE_DISPLAYPHYCLK, &clks2) && ++ clks2.num_levels >= 3) { ++ dc->dcn_soc.phyclkv_min0p65 = clks2.data[0].clocks_in_khz / 1000.0; ++ dc->dcn_soc.phyclkv_mid0p72 = clks2.data[clks2.num_levels - 3].clocks_in_khz / 1000.0; ++ dc->dcn_soc.phyclkv_nom0p8 = clks2.data[clks2.num_levels - 2].clocks_in_khz / 1000.0; ++ dc->dcn_soc.phyclkv_max0p9 = clks2.data[clks2.num_levels - 1].clocks_in_khz / 1000.0; ++ } else ++ BREAK_TO_DEBUGGER(); ++ if (dm_pp_get_clock_levels_by_type_with_latency( ++ ctx, DM_PP_CLOCK_TYPE_DPPCLK, &clks) && ++ clks.num_levels >= 3) { ++ dc->dcn_soc.max_dppclk_vmin0p65 = clks.data[0].clocks_in_khz / 1000.0; ++ dc->dcn_soc.max_dppclk_vmid0p72 = clks.data[clks.num_levels - 3].clocks_in_khz / 1000.0; ++ dc->dcn_soc.max_dppclk_vnom0p8 = clks.data[clks.num_levels - 2].clocks_in_khz / 1000.0; ++ dc->dcn_soc.max_dppclk_vmax0p9 = clks.data[clks.num_levels - 1].clocks_in_khz / 1000.0; ++ } ++ ++ if (dm_pp_get_clock_levels_by_type_with_latency( ++ ctx, DM_PP_CLOCK_TYPE_SOCCLK, &clks) && ++ clks.num_levels >= 3) { ++ dc->dcn_soc.socclk = clks.data[0].clocks_in_khz / 1000.0; ++ } else ++ BREAK_TO_DEBUGGER(); ++ kernel_fpu_end(); ++} ++ ++void dcn_bw_notify_pplib_of_wm_ranges(struct core_dc *dc) ++{ ++ struct dm_pp_wm_sets_with_clock_ranges_soc15 clk_ranges = {0}; ++ int max_fclk_khz, nom_fclk_khz, min_fclk_khz, max_dcfclk_khz, ++ nom_dcfclk_khz, min_dcfclk_khz, socclk_khz; ++ const int overdrive = 5000000; /* 5 GHz to cover Overdrive */ ++ unsigned factor = (ddr4_dram_factor_single_Channel * dc->dcn_soc.number_of_channels); ++ ++ kernel_fpu_begin(); ++ max_fclk_khz = dc->dcn_soc.fabric_and_dram_bandwidth_vmax0p9 * 1000000 / factor; ++ nom_fclk_khz = dc->dcn_soc.fabric_and_dram_bandwidth_vnom0p8 * 1000000 / factor; ++ min_fclk_khz = dc->dcn_soc.fabric_and_dram_bandwidth_vmin0p65 * 1000000 / factor; ++ max_dcfclk_khz = dc->dcn_soc.dcfclkv_max0p9 * 1000; ++ nom_dcfclk_khz = dc->dcn_soc.dcfclkv_nom0p8 * 1000; ++ min_dcfclk_khz = dc->dcn_soc.dcfclkv_min0p65 * 1000; ++ socclk_khz = dc->dcn_soc.socclk * 1000; ++ kernel_fpu_end(); ++ ++ /* Now notify PPLib/SMU about which Watermarks sets they should select ++ * depending on DPM state they are in. And update BW MGR GFX Engine and ++ * Memory clock member variables for Watermarks calculations for each ++ * Watermark Set ++ */ ++ /* SOCCLK does not affect anytihng but writeback for DCN so for now we dont ++ * care what the value is, hence min to overdrive level ++ */ ++ clk_ranges.num_wm_dmif_sets = 4; ++ clk_ranges.num_wm_mcif_sets = 4; ++ clk_ranges.wm_dmif_clocks_ranges[0].wm_set_id = WM_SET_A; ++ clk_ranges.wm_dmif_clocks_ranges[0].wm_min_dcfclk_clk_in_khz = min_dcfclk_khz; ++ clk_ranges.wm_dmif_clocks_ranges[0].wm_max_dcfclk_clk_in_khz = nom_dcfclk_khz - 1; ++ clk_ranges.wm_dmif_clocks_ranges[0].wm_min_memg_clk_in_khz = min_fclk_khz; ++ clk_ranges.wm_dmif_clocks_ranges[0].wm_max_mem_clk_in_khz = nom_fclk_khz - 1; ++ clk_ranges.wm_mcif_clocks_ranges[0].wm_set_id = WM_SET_A; ++ clk_ranges.wm_mcif_clocks_ranges[0].wm_min_socclk_clk_in_khz = socclk_khz; ++ clk_ranges.wm_mcif_clocks_ranges[0].wm_max_socclk_clk_in_khz = overdrive; ++ clk_ranges.wm_mcif_clocks_ranges[0].wm_min_memg_clk_in_khz = min_fclk_khz; ++ clk_ranges.wm_mcif_clocks_ranges[0].wm_max_mem_clk_in_khz = nom_fclk_khz - 1; ++ ++ clk_ranges.wm_dmif_clocks_ranges[1].wm_set_id = WM_SET_B; ++ clk_ranges.wm_dmif_clocks_ranges[1].wm_min_dcfclk_clk_in_khz = min_dcfclk_khz; ++ clk_ranges.wm_dmif_clocks_ranges[1].wm_max_dcfclk_clk_in_khz = nom_dcfclk_khz - 1; ++ clk_ranges.wm_dmif_clocks_ranges[1].wm_min_memg_clk_in_khz = nom_fclk_khz; ++ clk_ranges.wm_dmif_clocks_ranges[1].wm_max_mem_clk_in_khz = max_fclk_khz; ++ clk_ranges.wm_mcif_clocks_ranges[1].wm_set_id = WM_SET_B; ++ clk_ranges.wm_mcif_clocks_ranges[1].wm_min_socclk_clk_in_khz = socclk_khz; ++ clk_ranges.wm_mcif_clocks_ranges[1].wm_max_socclk_clk_in_khz = overdrive; ++ clk_ranges.wm_mcif_clocks_ranges[1].wm_min_memg_clk_in_khz = nom_fclk_khz; ++ clk_ranges.wm_mcif_clocks_ranges[1].wm_max_mem_clk_in_khz = max_fclk_khz; ++ ++ ++ clk_ranges.wm_dmif_clocks_ranges[2].wm_set_id = WM_SET_C; ++ clk_ranges.wm_dmif_clocks_ranges[2].wm_min_dcfclk_clk_in_khz = nom_dcfclk_khz; ++ clk_ranges.wm_dmif_clocks_ranges[2].wm_max_dcfclk_clk_in_khz = max_dcfclk_khz; ++ clk_ranges.wm_dmif_clocks_ranges[2].wm_min_memg_clk_in_khz = nom_fclk_khz; ++ clk_ranges.wm_dmif_clocks_ranges[2].wm_max_mem_clk_in_khz = max_fclk_khz; ++ clk_ranges.wm_mcif_clocks_ranges[2].wm_set_id = WM_SET_C; ++ clk_ranges.wm_mcif_clocks_ranges[2].wm_min_socclk_clk_in_khz = socclk_khz; ++ clk_ranges.wm_mcif_clocks_ranges[2].wm_max_socclk_clk_in_khz = overdrive; ++ clk_ranges.wm_mcif_clocks_ranges[2].wm_min_memg_clk_in_khz = nom_fclk_khz; ++ clk_ranges.wm_mcif_clocks_ranges[2].wm_max_mem_clk_in_khz = max_fclk_khz; ++ ++ clk_ranges.wm_dmif_clocks_ranges[3].wm_set_id = WM_SET_D; ++ clk_ranges.wm_dmif_clocks_ranges[3].wm_min_dcfclk_clk_in_khz = max_dcfclk_khz + 1; ++ clk_ranges.wm_dmif_clocks_ranges[3].wm_max_dcfclk_clk_in_khz = overdrive; ++ clk_ranges.wm_dmif_clocks_ranges[3].wm_min_memg_clk_in_khz = max_fclk_khz + 1; ++ clk_ranges.wm_dmif_clocks_ranges[3].wm_max_mem_clk_in_khz = overdrive; ++ clk_ranges.wm_mcif_clocks_ranges[3].wm_set_id = WM_SET_D; ++ clk_ranges.wm_mcif_clocks_ranges[3].wm_min_socclk_clk_in_khz = socclk_khz; ++ clk_ranges.wm_mcif_clocks_ranges[3].wm_max_socclk_clk_in_khz = overdrive; ++ clk_ranges.wm_mcif_clocks_ranges[3].wm_min_memg_clk_in_khz = max_fclk_khz + 1; ++ clk_ranges.wm_mcif_clocks_ranges[3].wm_max_mem_clk_in_khz = overdrive; ++ ++ /* Notify PP Lib/SMU which Watermarks to use for which clock ranges */ ++ dm_pp_notify_wm_clock_changes_soc15(dc->ctx, &clk_ranges); ++} ++ ++void dcn_bw_sync_calcs_and_dml(struct core_dc *dc) ++{ ++ kernel_fpu_begin(); ++ dc->dml.soc.vmin.socclk_mhz = dc->dcn_soc.socclk; ++ dc->dml.soc.vmid.socclk_mhz = dc->dcn_soc.socclk; ++ dc->dml.soc.vnom.socclk_mhz = dc->dcn_soc.socclk; ++ dc->dml.soc.vmax.socclk_mhz = dc->dcn_soc.socclk; ++ ++ dc->dml.soc.vmin.dcfclk_mhz = dc->dcn_soc.dcfclkv_min0p65; ++ dc->dml.soc.vmid.dcfclk_mhz = dc->dcn_soc.dcfclkv_mid0p72; ++ dc->dml.soc.vnom.dcfclk_mhz = dc->dcn_soc.dcfclkv_nom0p8; ++ dc->dml.soc.vmax.dcfclk_mhz = dc->dcn_soc.dcfclkv_max0p9; ++ ++ dc->dml.soc.vmin.dispclk_mhz = dc->dcn_soc.max_dispclk_vmin0p65; ++ dc->dml.soc.vmid.dispclk_mhz = dc->dcn_soc.max_dispclk_vmid0p72; ++ dc->dml.soc.vnom.dispclk_mhz = dc->dcn_soc.max_dispclk_vnom0p8; ++ dc->dml.soc.vmax.dispclk_mhz = dc->dcn_soc.max_dispclk_vmax0p9; ++ ++ dc->dml.soc.vmin.dppclk_mhz = dc->dcn_soc.max_dppclk_vmin0p65; ++ dc->dml.soc.vmid.dppclk_mhz = dc->dcn_soc.max_dppclk_vmid0p72; ++ dc->dml.soc.vnom.dppclk_mhz = dc->dcn_soc.max_dppclk_vnom0p8; ++ dc->dml.soc.vmax.dppclk_mhz = dc->dcn_soc.max_dppclk_vmax0p9; ++ ++ dc->dml.soc.vmin.phyclk_mhz = dc->dcn_soc.phyclkv_min0p65; ++ dc->dml.soc.vmid.phyclk_mhz = dc->dcn_soc.phyclkv_mid0p72; ++ dc->dml.soc.vnom.phyclk_mhz = dc->dcn_soc.phyclkv_nom0p8; ++ dc->dml.soc.vmax.phyclk_mhz = dc->dcn_soc.phyclkv_max0p9; ++ ++ dc->dml.soc.vmin.dram_bw_per_chan_gbps = dc->dcn_soc.fabric_and_dram_bandwidth_vmin0p65; ++ dc->dml.soc.vmid.dram_bw_per_chan_gbps = dc->dcn_soc.fabric_and_dram_bandwidth_vmid0p72; ++ dc->dml.soc.vnom.dram_bw_per_chan_gbps = dc->dcn_soc.fabric_and_dram_bandwidth_vnom0p8; ++ dc->dml.soc.vmax.dram_bw_per_chan_gbps = dc->dcn_soc.fabric_and_dram_bandwidth_vmax0p9; ++ ++ dc->dml.soc.sr_exit_time_us = dc->dcn_soc.sr_exit_time; ++ dc->dml.soc.sr_enter_plus_exit_time_us = dc->dcn_soc.sr_enter_plus_exit_time; ++ dc->dml.soc.urgent_latency_us = dc->dcn_soc.urgent_latency; ++ dc->dml.soc.writeback_latency_us = dc->dcn_soc.write_back_latency; ++ dc->dml.soc.ideal_dram_bw_after_urgent_percent = ++ dc->dcn_soc.percent_of_ideal_drambw_received_after_urg_latency; ++ dc->dml.soc.max_request_size_bytes = dc->dcn_soc.max_request_size; ++ dc->dml.soc.downspread_percent = dc->dcn_soc.downspreading; ++ dc->dml.soc.round_trip_ping_latency_dcfclk_cycles = ++ dc->dcn_soc.round_trip_ping_latency_cycles; ++ dc->dml.soc.urgent_out_of_order_return_per_channel_bytes = ++ dc->dcn_soc.urgent_out_of_order_return_per_channel; ++ dc->dml.soc.num_chans = dc->dcn_soc.number_of_channels; ++ dc->dml.soc.vmm_page_size_bytes = dc->dcn_soc.vmm_page_size; ++ dc->dml.soc.dram_clock_change_latency_us = dc->dcn_soc.dram_clock_change_latency; ++ dc->dml.soc.return_bus_width_bytes = dc->dcn_soc.return_bus_width; ++ ++ dc->dml.ip.rob_buffer_size_kbytes = dc->dcn_ip.rob_buffer_size_in_kbyte; ++ dc->dml.ip.det_buffer_size_kbytes = dc->dcn_ip.det_buffer_size_in_kbyte; ++ dc->dml.ip.dpp_output_buffer_pixels = dc->dcn_ip.dpp_output_buffer_pixels; ++ dc->dml.ip.opp_output_buffer_lines = dc->dcn_ip.opp_output_buffer_lines; ++ dc->dml.ip.pixel_chunk_size_kbytes = dc->dcn_ip.pixel_chunk_size_in_kbyte; ++ dc->dml.ip.pte_enable = dc->dcn_ip.pte_enable == dcn_bw_yes; ++ dc->dml.ip.pte_chunk_size_kbytes = dc->dcn_ip.pte_chunk_size; ++ dc->dml.ip.meta_chunk_size_kbytes = dc->dcn_ip.meta_chunk_size; ++ dc->dml.ip.writeback_chunk_size_kbytes = dc->dcn_ip.writeback_chunk_size; ++ dc->dml.ip.line_buffer_size_bits = dc->dcn_ip.line_buffer_size; ++ dc->dml.ip.max_line_buffer_lines = dc->dcn_ip.max_line_buffer_lines; ++ dc->dml.ip.IsLineBufferBppFixed = dc->dcn_ip.is_line_buffer_bpp_fixed == dcn_bw_yes; ++ dc->dml.ip.LineBufferFixedBpp = dc->dcn_ip.line_buffer_fixed_bpp; ++ dc->dml.ip.writeback_luma_buffer_size_kbytes = dc->dcn_ip.writeback_luma_buffer_size; ++ dc->dml.ip.writeback_chroma_buffer_size_kbytes = dc->dcn_ip.writeback_chroma_buffer_size; ++ dc->dml.ip.max_num_dpp = dc->dcn_ip.max_num_dpp; ++ dc->dml.ip.max_num_wb = dc->dcn_ip.max_num_writeback; ++ dc->dml.ip.max_dchub_pscl_bw_pix_per_clk = dc->dcn_ip.max_dchub_topscl_throughput; ++ dc->dml.ip.max_pscl_lb_bw_pix_per_clk = dc->dcn_ip.max_pscl_tolb_throughput; ++ dc->dml.ip.max_lb_vscl_bw_pix_per_clk = dc->dcn_ip.max_lb_tovscl_throughput; ++ dc->dml.ip.max_vscl_hscl_bw_pix_per_clk = dc->dcn_ip.max_vscl_tohscl_throughput; ++ dc->dml.ip.max_hscl_ratio = dc->dcn_ip.max_hscl_ratio; ++ dc->dml.ip.max_vscl_ratio = dc->dcn_ip.max_vscl_ratio; ++ dc->dml.ip.max_hscl_taps = dc->dcn_ip.max_hscl_taps; ++ dc->dml.ip.max_vscl_taps = dc->dcn_ip.max_vscl_taps; ++ /*pte_buffer_size_in_requests missing in dml*/ ++ dc->dml.ip.dispclk_ramp_margin_percent = dc->dcn_ip.dispclk_ramping_margin; ++ dc->dml.ip.underscan_factor = dc->dcn_ip.under_scan_factor; ++ dc->dml.ip.max_inter_dcn_tile_repeaters = dc->dcn_ip.max_inter_dcn_tile_repeaters; ++ dc->dml.ip.can_vstartup_lines_exceed_vsync_plus_back_porch_lines_minus_one = ++ dc->dcn_ip.can_vstartup_lines_exceed_vsync_plus_back_porch_lines_minus_one == dcn_bw_yes; ++ dc->dml.ip.bug_forcing_LC_req_same_size_fixed = ++ dc->dcn_ip.bug_forcing_luma_and_chroma_request_to_same_size_fixed == dcn_bw_yes; ++ dc->dml.ip.dcfclk_cstate_latency = dc->dcn_ip.dcfclk_cstate_latency; ++ kernel_fpu_end(); ++} +diff --git a/drivers/gpu/drm/amd/display/dc/inc/dcn_calcs.h b/drivers/gpu/drm/amd/display/dc/inc/dcn_calcs.h +new file mode 100644 +index 0000000..499bc11 +--- /dev/null ++++ b/drivers/gpu/drm/amd/display/dc/inc/dcn_calcs.h +@@ -0,0 +1,629 @@ ++/* ++ * 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 ++ * ++ */ ++ ++/** ++ * Bandwidth and Watermark calculations interface. ++ * (Refer to "DCEx_mode_support.xlsm" from Perforce.) ++ */ ++#ifndef __DCN_CALCS_H__ ++#define __DCN_CALCS_H__ ++ ++#include "bw_fixed.h" ++#include "display_clock.h" ++#include "../dml/display_mode_lib.h" ++ ++struct core_dc; ++struct validate_context; ++ ++/******************************************************************************* ++ * DCN data structures. ++ ******************************************************************************/ ++ ++#define number_of_planes 6 ++#define number_of_planes_minus_one 5 ++#define number_of_states 4 ++#define number_of_states_plus_one 5 ++ ++#define ddr4_dram_width 64 ++#define ddr4_dram_factor_single_Channel 16 ++enum dcn_bw_defs { ++ dcn_bw_v_min0p65, ++ dcn_bw_v_mid0p72, ++ dcn_bw_v_nom0p8, ++ dcn_bw_v_max0p9, ++ dcn_bw_v_max0p91, ++ dcn_bw_no_support = 5, ++ dcn_bw_yes, ++ dcn_bw_hor, ++ dcn_bw_vert, ++ dcn_bw_override, ++ dcn_bw_rgb_sub_64, ++ dcn_bw_rgb_sub_32, ++ dcn_bw_rgb_sub_16, ++ dcn_bw_no, ++ dcn_bw_sw_linear, ++ dcn_bw_sw_4_kb_d, ++ dcn_bw_sw_4_kb_d_x, ++ dcn_bw_sw_64_kb_d, ++ dcn_bw_sw_64_kb_d_t, ++ dcn_bw_sw_64_kb_d_x, ++ dcn_bw_sw_var_d, ++ dcn_bw_sw_var_d_x, ++ dcn_bw_yuv420_sub_8, ++ dcn_bw_sw_4_kb_s, ++ dcn_bw_sw_4_kb_s_x, ++ dcn_bw_sw_64_kb_s, ++ dcn_bw_sw_64_kb_s_t, ++ dcn_bw_sw_64_kb_s_x, ++ dcn_bw_writeback, ++ dcn_bw_444, ++ dcn_bw_dp, ++ dcn_bw_420, ++ dcn_bw_hdmi, ++ dcn_bw_sw_var_s, ++ dcn_bw_sw_var_s_x, ++ dcn_bw_yuv420_sub_10, ++ dcn_bw_supported_in_v_active, ++ dcn_bw_supported_in_v_blank, ++ dcn_bw_not_supported, ++ dcn_bw_na, ++}; ++ ++/*bounding box parameters*/ ++/*mode parameters*/ ++/*system configuration*/ ++/* display configuration*/ ++struct dcn_bw_internal_vars { ++ float voltage[number_of_states_plus_one + 1]; ++ float max_dispclk[number_of_states_plus_one + 1]; ++ float max_dppclk[number_of_states_plus_one + 1]; ++ float dcfclk_per_state[number_of_states_plus_one + 1]; ++ float phyclk_per_state[number_of_states_plus_one + 1]; ++ float fabric_and_dram_bandwidth_per_state[number_of_states_plus_one + 1]; ++ float sr_exit_time; ++ float sr_enter_plus_exit_time; ++ float dram_clock_change_latency; ++ float urgent_latency; ++ float write_back_latency; ++ float percent_of_ideal_drambw_received_after_urg_latency; ++ float dcfclkv_max0p9; ++ float dcfclkv_nom0p8; ++ float dcfclkv_mid0p72; ++ float dcfclkv_min0p65; ++ float max_dispclk_vmax0p9; ++ float max_dppclk_vmax0p9; ++ float max_dispclk_vnom0p8; ++ float max_dppclk_vnom0p8; ++ float max_dispclk_vmid0p72; ++ float max_dppclk_vmid0p72; ++ float max_dispclk_vmin0p65; ++ float max_dppclk_vmin0p65; ++ float socclk; ++ float fabric_and_dram_bandwidth_vmax0p9; ++ float fabric_and_dram_bandwidth_vnom0p8; ++ float fabric_and_dram_bandwidth_vmid0p72; ++ float fabric_and_dram_bandwidth_vmin0p65; ++ float round_trip_ping_latency_cycles; ++ float urgent_out_of_order_return_per_channel; ++ float number_of_channels; ++ float vmm_page_size; ++ float return_bus_width; ++ float rob_buffer_size_in_kbyte; ++ float det_buffer_size_in_kbyte; ++ float dpp_output_buffer_pixels; ++ float opp_output_buffer_lines; ++ float pixel_chunk_size_in_kbyte; ++ float pte_chunk_size; ++ float meta_chunk_size; ++ float writeback_chunk_size; ++ enum dcn_bw_defs odm_capability; ++ enum dcn_bw_defs dsc_capability; ++ float line_buffer_size; ++ enum dcn_bw_defs is_line_buffer_bpp_fixed; ++ float line_buffer_fixed_bpp; ++ float max_line_buffer_lines; ++ float writeback_luma_buffer_size; ++ float writeback_chroma_buffer_size; ++ float max_num_dpp; ++ float max_num_writeback; ++ float max_dchub_topscl_throughput; ++ float max_pscl_tolb_throughput; ++ float max_lb_tovscl_throughput; ++ float max_vscl_tohscl_throughput; ++ float max_hscl_ratio; ++ float max_vscl_ratio; ++ float max_hscl_taps; ++ float max_vscl_taps; ++ float under_scan_factor; ++ float phyclkv_max0p9; ++ float phyclkv_nom0p8; ++ float phyclkv_mid0p72; ++ float phyclkv_min0p65; ++ float pte_buffer_size_in_requests; ++ float dispclk_ramping_margin; ++ float downspreading; ++ float max_inter_dcn_tile_repeaters; ++ enum dcn_bw_defs can_vstartup_lines_exceed_vsync_plus_back_porch_lines_minus_one; ++ enum dcn_bw_defs bug_forcing_luma_and_chroma_request_to_same_size_fixed; ++ int mode; ++ float viewport_width[number_of_planes_minus_one + 1]; ++ float htotal[number_of_planes_minus_one + 1]; ++ float vtotal[number_of_planes_minus_one + 1]; ++ float v_sync_plus_back_porch[number_of_planes_minus_one + 1]; ++ float vactive[number_of_planes_minus_one + 1]; ++ float pixel_clock[number_of_planes_minus_one + 1]; /*MHz*/ ++ float viewport_height[number_of_planes_minus_one + 1]; ++ enum dcn_bw_defs dcc_enable[number_of_planes_minus_one + 1]; ++ float dcc_rate[number_of_planes_minus_one + 1]; ++ enum dcn_bw_defs source_scan[number_of_planes_minus_one + 1]; ++ float lb_bit_per_pixel[number_of_planes_minus_one + 1]; ++ enum dcn_bw_defs source_pixel_format[number_of_planes_minus_one + 1]; ++ enum dcn_bw_defs source_surface_mode[number_of_planes_minus_one + 1]; ++ enum dcn_bw_defs output_format[number_of_planes_minus_one + 1]; ++ enum dcn_bw_defs output[number_of_planes_minus_one + 1]; ++ float scaler_rec_out_width[number_of_planes_minus_one + 1]; ++ float scaler_recout_height[number_of_planes_minus_one + 1]; ++ float underscan_output[number_of_planes_minus_one + 1]; ++ float interlace_output[number_of_planes_minus_one + 1]; ++ float override_hta_ps[number_of_planes_minus_one + 1]; ++ float override_vta_ps[number_of_planes_minus_one + 1]; ++ float override_hta_pschroma[number_of_planes_minus_one + 1]; ++ float override_vta_pschroma[number_of_planes_minus_one + 1]; ++ float urgent_latency_support_us[number_of_planes_minus_one + 1]; ++ float h_ratio[number_of_planes_minus_one + 1]; ++ float v_ratio[number_of_planes_minus_one + 1]; ++ float htaps[number_of_planes_minus_one + 1]; ++ float vtaps[number_of_planes_minus_one + 1]; ++ float hta_pschroma[number_of_planes_minus_one + 1]; ++ float vta_pschroma[number_of_planes_minus_one + 1]; ++ enum dcn_bw_defs pte_enable; ++ enum dcn_bw_defs synchronized_vblank; ++ enum dcn_bw_defs ta_pscalculation; ++ int voltage_override_level; ++ int number_of_active_planes; ++ int voltage_level; ++ enum dcn_bw_defs immediate_flip_supported; ++ float dcfclk; ++ float max_phyclk; ++ float fabric_and_dram_bandwidth; ++ float dpp_per_plane_per_ratio[1 + 1][number_of_planes_minus_one + 1]; ++ enum dcn_bw_defs dispclk_dppclk_support_per_ratio[1 + 1]; ++ float required_dispclk_per_ratio[1 + 1]; ++ enum dcn_bw_defs error_message[1 + 1]; ++ int dispclk_dppclk_ratio; ++ float dpp_per_plane[number_of_planes_minus_one + 1]; ++ float det_buffer_size_y[number_of_planes_minus_one + 1]; ++ float det_buffer_size_c[number_of_planes_minus_one + 1]; ++ float swath_height_y[number_of_planes_minus_one + 1]; ++ float swath_height_c[number_of_planes_minus_one + 1]; ++ enum dcn_bw_defs final_error_message; ++ float frequency; ++ float header_line; ++ float header; ++ enum dcn_bw_defs voltage_override; ++ enum dcn_bw_defs allow_different_hratio_vratio; ++ float acceptable_quality_hta_ps; ++ float acceptable_quality_vta_ps; ++ float no_of_dpp[number_of_states_plus_one + 1][1 + 1][number_of_planes_minus_one + 1]; ++ float swath_width_yper_state[number_of_states_plus_one + 1][1 + 1][number_of_planes_minus_one + 1]; ++ float swath_height_yper_state[number_of_states_plus_one + 1][1 + 1][number_of_planes_minus_one + 1]; ++ float swath_height_cper_state[number_of_states_plus_one + 1][1 + 1][number_of_planes_minus_one + 1]; ++ float urgent_latency_support_us_per_state[number_of_states_plus_one + 1][1 + 1][number_of_planes_minus_one + 1]; ++ float v_ratio_pre_ywith_immediate_flip[number_of_states_plus_one + 1][1 + 1][number_of_planes_minus_one + 1]; ++ float v_ratio_pre_cwith_immediate_flip[number_of_states_plus_one + 1][1 + 1][number_of_planes_minus_one + 1]; ++ float required_prefetch_pixel_data_bw_with_immediate_flip[number_of_states_plus_one + 1][1 + 1][number_of_planes_minus_one + 1]; ++ float v_ratio_pre_ywithout_immediate_flip[number_of_states_plus_one + 1][1 + 1][number_of_planes_minus_one + 1]; ++ float v_ratio_pre_cwithout_immediate_flip[number_of_states_plus_one + 1][1 + 1][number_of_planes_minus_one + 1]; ++ float required_prefetch_pixel_data_bw_without_immediate_flip[number_of_states_plus_one + 1][1 + 1][number_of_planes_minus_one + 1]; ++ enum dcn_bw_defs prefetch_supported_with_immediate_flip[number_of_states_plus_one + 1][1 + 1]; ++ enum dcn_bw_defs prefetch_supported_without_immediate_flip[number_of_states_plus_one + 1][1 + 1]; ++ enum dcn_bw_defs v_ratio_in_prefetch_supported_with_immediate_flip[number_of_states_plus_one + 1][1 + 1]; ++ enum dcn_bw_defs v_ratio_in_prefetch_supported_without_immediate_flip[number_of_states_plus_one + 1][1 + 1]; ++ float required_dispclk[number_of_states_plus_one + 1][1 + 1]; ++ enum dcn_bw_defs dispclk_dppclk_support[number_of_states_plus_one + 1][1 + 1]; ++ enum dcn_bw_defs total_available_pipes_support[number_of_states_plus_one + 1][1 + 1]; ++ float total_number_of_active_dpp[number_of_states_plus_one + 1][1 + 1]; ++ float total_number_of_dcc_active_dpp[number_of_states_plus_one + 1][1 + 1]; ++ enum dcn_bw_defs urgent_latency_support[number_of_states_plus_one + 1][1 + 1]; ++ enum dcn_bw_defs mode_support_with_immediate_flip[number_of_states_plus_one + 1][1 + 1]; ++ enum dcn_bw_defs mode_support_without_immediate_flip[number_of_states_plus_one + 1][1 + 1]; ++ float return_bw_per_state[number_of_states_plus_one + 1]; ++ enum dcn_bw_defs dio_support[number_of_states_plus_one + 1]; ++ float urgent_round_trip_and_out_of_order_latency_per_state[number_of_states_plus_one + 1]; ++ enum dcn_bw_defs rob_support[number_of_states_plus_one + 1]; ++ enum dcn_bw_defs bandwidth_support[number_of_states_plus_one + 1]; ++ float prefetch_bw[number_of_planes_minus_one + 1]; ++ float meta_pte_bytes_per_frame[number_of_planes_minus_one + 1]; ++ float meta_row_bytes[number_of_planes_minus_one + 1]; ++ float dpte_bytes_per_row[number_of_planes_minus_one + 1]; ++ float prefetch_lines_y[number_of_planes_minus_one + 1]; ++ float prefetch_lines_c[number_of_planes_minus_one + 1]; ++ float max_num_sw_y[number_of_planes_minus_one + 1]; ++ float max_num_sw_c[number_of_planes_minus_one + 1]; ++ float line_times_for_prefetch[number_of_planes_minus_one + 1]; ++ float lines_for_meta_pte_with_immediate_flip[number_of_planes_minus_one + 1]; ++ float lines_for_meta_pte_without_immediate_flip[number_of_planes_minus_one + 1]; ++ float lines_for_meta_and_dpte_row_with_immediate_flip[number_of_planes_minus_one + 1]; ++ float lines_for_meta_and_dpte_row_without_immediate_flip[number_of_planes_minus_one + 1]; ++ float min_dppclk_using_single_dpp[number_of_planes_minus_one + 1]; ++ float swath_width_ysingle_dpp[number_of_planes_minus_one + 1]; ++ float byte_per_pixel_in_dety[number_of_planes_minus_one + 1]; ++ float byte_per_pixel_in_detc[number_of_planes_minus_one + 1]; ++ float number_of_dpp_required_for_det_and_lb_size[number_of_planes_minus_one + 1]; ++ float required_phyclk[number_of_planes_minus_one + 1]; ++ float read256_block_height_y[number_of_planes_minus_one + 1]; ++ float read256_block_width_y[number_of_planes_minus_one + 1]; ++ float read256_block_height_c[number_of_planes_minus_one + 1]; ++ float read256_block_width_c[number_of_planes_minus_one + 1]; ++ float max_swath_height_y[number_of_planes_minus_one + 1]; ++ float max_swath_height_c[number_of_planes_minus_one + 1]; ++ float min_swath_height_y[number_of_planes_minus_one + 1]; ++ float min_swath_height_c[number_of_planes_minus_one + 1]; ++ float read_bandwidth[number_of_planes_minus_one + 1]; ++ float write_bandwidth[number_of_planes_minus_one + 1]; ++ float pscl_factor[number_of_planes_minus_one + 1]; ++ float pscl_factor_chroma[number_of_planes_minus_one + 1]; ++ enum dcn_bw_defs scale_ratio_support; ++ enum dcn_bw_defs source_format_pixel_and_scan_support; ++ float total_read_bandwidth_consumed_gbyte_per_second; ++ float total_write_bandwidth_consumed_gbyte_per_second; ++ float total_bandwidth_consumed_gbyte_per_second; ++ enum dcn_bw_defs dcc_enabled_in_any_plane; ++ float return_bw_todcn_per_state; ++ float critical_point; ++ enum dcn_bw_defs writeback_latency_support; ++ float required_output_bw; ++ float total_number_of_active_writeback; ++ enum dcn_bw_defs total_available_writeback_support; ++ float maximum_swath_width; ++ float number_of_dpp_required_for_det_size; ++ float number_of_dpp_required_for_lb_size; ++ float min_dispclk_using_single_dpp; ++ float min_dispclk_using_dual_dpp; ++ enum dcn_bw_defs viewport_size_support; ++ float swath_width_granularity_y; ++ float rounded_up_max_swath_size_bytes_y; ++ float swath_width_granularity_c; ++ float rounded_up_max_swath_size_bytes_c; ++ float lines_in_det_luma; ++ float lines_in_det_chroma; ++ float effective_lb_latency_hiding_source_lines_luma; ++ float effective_lb_latency_hiding_source_lines_chroma; ++ float effective_detlb_lines_luma; ++ float effective_detlb_lines_chroma; ++ float projected_dcfclk_deep_sleep; ++ float meta_req_height_y; ++ float meta_req_width_y; ++ float meta_surface_width_y; ++ float meta_surface_height_y; ++ float meta_pte_bytes_per_frame_y; ++ float meta_row_bytes_y; ++ float macro_tile_block_size_bytes_y; ++ float macro_tile_block_height_y; ++ float data_pte_req_height_y; ++ float data_pte_req_width_y; ++ float dpte_bytes_per_row_y; ++ float meta_req_height_c; ++ float meta_req_width_c; ++ float meta_surface_width_c; ++ float meta_surface_height_c; ++ float meta_pte_bytes_per_frame_c; ++ float meta_row_bytes_c; ++ float macro_tile_block_size_bytes_c; ++ float macro_tile_block_height_c; ++ float macro_tile_block_width_c; ++ float data_pte_req_height_c; ++ float data_pte_req_width_c; ++ float dpte_bytes_per_row_c; ++ float v_init_y; ++ float max_partial_sw_y; ++ float v_init_c; ++ float max_partial_sw_c; ++ float dst_x_after_scaler; ++ float dst_y_after_scaler; ++ float time_calc; ++ float v_update_offset[number_of_planes_minus_one + 1]; ++ float total_repeater_delay; ++ float v_update_width[number_of_planes_minus_one + 1]; ++ float v_ready_offset[number_of_planes_minus_one + 1]; ++ float time_setup; ++ float extra_latency; ++ float maximum_vstartup; ++ float bw_available_for_immediate_flip; ++ float total_immediate_flip_bytes[number_of_planes_minus_one + 1]; ++ float time_for_meta_pte_with_immediate_flip; ++ float time_for_meta_pte_without_immediate_flip; ++ float time_for_meta_and_dpte_row_with_immediate_flip; ++ float time_for_meta_and_dpte_row_without_immediate_flip; ++ float line_times_to_request_prefetch_pixel_data_with_immediate_flip; ++ float line_times_to_request_prefetch_pixel_data_without_immediate_flip; ++ float maximum_read_bandwidth_with_prefetch_with_immediate_flip; ++ float maximum_read_bandwidth_with_prefetch_without_immediate_flip; ++ float voltage_level_with_immediate_flip; ++ float voltage_level_without_immediate_flip; ++ float total_number_of_active_dpp_per_ratio[1 + 1]; ++ float byte_per_pix_dety; ++ float byte_per_pix_detc; ++ float read256_bytes_block_height_y; ++ float read256_bytes_block_width_y; ++ float read256_bytes_block_height_c; ++ float read256_bytes_block_width_c; ++ float maximum_swath_height_y; ++ float maximum_swath_height_c; ++ float minimum_swath_height_y; ++ float minimum_swath_height_c; ++ float swath_width; ++ float prefetch_bandwidth[number_of_planes_minus_one + 1]; ++ float v_init_pre_fill_y[number_of_planes_minus_one + 1]; ++ float v_init_pre_fill_c[number_of_planes_minus_one + 1]; ++ float max_num_swath_y[number_of_planes_minus_one + 1]; ++ float max_num_swath_c[number_of_planes_minus_one + 1]; ++ float prefill_y[number_of_planes_minus_one + 1]; ++ float prefill_c[number_of_planes_minus_one + 1]; ++ float v_startup[number_of_planes_minus_one + 1]; ++ enum dcn_bw_defs allow_dram_clock_change_during_vblank[number_of_planes_minus_one + 1]; ++ float allow_dram_self_refresh_during_vblank[number_of_planes_minus_one + 1]; ++ float v_ratio_prefetch_y[number_of_planes_minus_one + 1]; ++ float v_ratio_prefetch_c[number_of_planes_minus_one + 1]; ++ float destination_lines_for_prefetch[number_of_planes_minus_one + 1]; ++ float destination_lines_to_request_vm_inv_blank[number_of_planes_minus_one + 1]; ++ float destination_lines_to_request_row_in_vblank[number_of_planes_minus_one + 1]; ++ float min_ttuv_blank[number_of_planes_minus_one + 1]; ++ float byte_per_pixel_dety[number_of_planes_minus_one + 1]; ++ float byte_per_pixel_detc[number_of_planes_minus_one + 1]; ++ float swath_width_y[number_of_planes_minus_one + 1]; ++ float lines_in_dety[number_of_planes_minus_one + 1]; ++ float lines_in_dety_rounded_down_to_swath[number_of_planes_minus_one + 1]; ++ float lines_in_detc[number_of_planes_minus_one + 1]; ++ float lines_in_detc_rounded_down_to_swath[number_of_planes_minus_one + 1]; ++ float full_det_buffering_time_y[number_of_planes_minus_one + 1]; ++ float full_det_buffering_time_c[number_of_planes_minus_one + 1]; ++ float active_dram_clock_change_latency_margin[number_of_planes_minus_one + 1]; ++ float v_blank_dram_clock_change_latency_margin[number_of_planes_minus_one + 1]; ++ float dcfclk_deep_sleep_per_plane[number_of_planes_minus_one + 1]; ++ float read_bandwidth_plane_luma[number_of_planes_minus_one + 1]; ++ float read_bandwidth_plane_chroma[number_of_planes_minus_one + 1]; ++ float display_pipe_line_delivery_time_luma[number_of_planes_minus_one + 1]; ++ float display_pipe_line_delivery_time_chroma[number_of_planes_minus_one + 1]; ++ float display_pipe_line_delivery_time_luma_prefetch[number_of_planes_minus_one + 1]; ++ float display_pipe_line_delivery_time_chroma_prefetch[number_of_planes_minus_one + 1]; ++ float pixel_pte_bytes_per_row[number_of_planes_minus_one + 1]; ++ float meta_pte_bytes_frame[number_of_planes_minus_one + 1]; ++ float meta_row_byte[number_of_planes_minus_one + 1]; ++ float prefetch_source_lines_y[number_of_planes_minus_one + 1]; ++ float prefetch_source_lines_c[number_of_planes_minus_one + 1]; ++ float pscl_throughput[number_of_planes_minus_one + 1]; ++ float pscl_throughput_chroma[number_of_planes_minus_one + 1]; ++ float output_bpphdmi[number_of_planes_minus_one + 1]; ++ float output_bppdp4_lane_hbr[number_of_planes_minus_one + 1]; ++ float output_bppdp4_lane_hbr2[number_of_planes_minus_one + 1]; ++ float output_bppdp4_lane_hbr3[number_of_planes_minus_one + 1]; ++ float max_vstartup_lines[number_of_planes_minus_one + 1]; ++ float dispclk_with_ramping; ++ float dispclk_without_ramping; ++ float dppclk_using_single_dpp_luma; ++ float dppclk_using_single_dpp; ++ float dppclk_using_single_dpp_chroma; ++ enum dcn_bw_defs odm_capable; ++ float dispclk; ++ float dppclk; ++ float return_bandwidth_to_dcn; ++ enum dcn_bw_defs dcc_enabled_any_plane; ++ float return_bw; ++ float critical_compression; ++ float total_data_read_bandwidth; ++ float total_active_dpp; ++ float total_dcc_active_dpp; ++ float urgent_round_trip_and_out_of_order_latency; ++ float last_pixel_of_line_extra_watermark; ++ float data_fabric_line_delivery_time_luma; ++ float data_fabric_line_delivery_time_chroma; ++ float urgent_extra_latency; ++ float urgent_watermark; ++ float ptemeta_urgent_watermark; ++ float dram_clock_change_watermark; ++ float total_active_writeback; ++ float writeback_dram_clock_change_watermark; ++ float min_full_det_buffering_time; ++ float frame_time_for_min_full_det_buffering_time; ++ float average_read_bandwidth_gbyte_per_second; ++ float part_of_burst_that_fits_in_rob; ++ float stutter_burst_time; ++ float stutter_efficiency_not_including_vblank; ++ float smallest_vblank; ++ float v_blank_time; ++ float stutter_efficiency; ++ float dcf_clk_deep_sleep; ++ float stutter_exit_watermark; ++ float stutter_enter_plus_exit_watermark; ++ float effective_det_plus_lb_lines_luma; ++ float urgent_latency_support_us_luma; ++ float effective_det_plus_lb_lines_chroma; ++ float urgent_latency_support_us_chroma; ++ float min_urgent_latency_support_us; ++ float non_urgent_latency_tolerance; ++ float block_height256_bytes_y; ++ float block_height256_bytes_c; ++ float meta_request_width_y; ++ float meta_surf_width_y; ++ float meta_surf_height_y; ++ float meta_pte_bytes_frame_y; ++ float meta_row_byte_y; ++ float macro_tile_size_byte_y; ++ float macro_tile_height_y; ++ float pixel_pte_req_height_y; ++ float pixel_pte_req_width_y; ++ float pixel_pte_bytes_per_row_y; ++ float meta_request_width_c; ++ float meta_surf_width_c; ++ float meta_surf_height_c; ++ float meta_pte_bytes_frame_c; ++ float meta_row_byte_c; ++ float macro_tile_size_bytes_c; ++ float macro_tile_height_c; ++ float pixel_pte_req_height_c; ++ float pixel_pte_req_width_c; ++ float pixel_pte_bytes_per_row_c; ++ float max_partial_swath_y; ++ float max_partial_swath_c; ++ float t_calc; ++ float next_prefetch_mode; ++ float v_startup_lines; ++ enum dcn_bw_defs planes_with_room_to_increase_vstartup_prefetch_bw_less_than_active_bw; ++ enum dcn_bw_defs planes_with_room_to_increase_vstartup_vratio_prefetch_more_than4; ++ enum dcn_bw_defs planes_with_room_to_increase_vstartup_destination_line_times_for_prefetch_less_than2; ++ enum dcn_bw_defs v_ratio_prefetch_more_than4; ++ enum dcn_bw_defs destination_line_times_for_prefetch_less_than2; ++ float prefetch_mode; ++ float dstx_after_scaler; ++ float dsty_after_scaler; ++ float v_update_offset_pix; ++ float total_repeater_delay_time; ++ float v_update_width_pix; ++ float v_ready_offset_pix; ++ float t_setup; ++ float t_wait; ++ float bandwidth_available_for_immediate_flip; ++ float tot_immediate_flip_bytes; ++ float max_rd_bandwidth; ++ float time_for_fetching_meta_pte; ++ float time_for_fetching_row_in_vblank; ++ float lines_to_request_prefetch_pixel_data; ++ float required_prefetch_pix_data_bw; ++ enum dcn_bw_defs prefetch_mode_supported; ++ float active_dp_ps; ++ float lb_latency_hiding_source_lines_y; ++ float lb_latency_hiding_source_lines_c; ++ float effective_lb_latency_hiding_y; ++ float effective_lb_latency_hiding_c; ++ float dpp_output_buffer_lines_y; ++ float dpp_output_buffer_lines_c; ++ float dppopp_buffering_y; ++ float max_det_buffering_time_y; ++ float active_dram_clock_change_latency_margin_y; ++ float dppopp_buffering_c; ++ float max_det_buffering_time_c; ++ float active_dram_clock_change_latency_margin_c; ++ float writeback_dram_clock_change_latency_margin; ++ float min_active_dram_clock_change_margin; ++ float v_blank_of_min_active_dram_clock_change_margin; ++ float second_min_active_dram_clock_change_margin; ++ float min_vblank_dram_clock_change_margin; ++ float dram_clock_change_margin; ++ float dram_clock_change_support; ++ float wr_bandwidth; ++ float max_used_bw; ++}; ++ ++struct dcn_soc_bounding_box { ++ float sr_exit_time; /*us*/ ++ float sr_enter_plus_exit_time; /*us*/ ++ float urgent_latency; /*us*/ ++ float write_back_latency; /*us*/ ++ float percent_of_ideal_drambw_received_after_urg_latency; /*%*/ ++ int max_request_size; /*bytes*/ ++ float dcfclkv_max0p9; /*MHz*/ ++ float dcfclkv_nom0p8; /*MHz*/ ++ float dcfclkv_mid0p72; /*MHz*/ ++ float dcfclkv_min0p65; /*MHz*/ ++ float max_dispclk_vmax0p9; /*MHz*/ ++ float max_dispclk_vmid0p72; /*MHz*/ ++ float max_dispclk_vnom0p8; /*MHz*/ ++ float max_dispclk_vmin0p65; /*MHz*/ ++ float max_dppclk_vmax0p9; /*MHz*/ ++ float max_dppclk_vnom0p8; /*MHz*/ ++ float max_dppclk_vmid0p72; /*MHz*/ ++ float max_dppclk_vmin0p65; /*MHz*/ ++ float socclk; /*MHz*/ ++ float fabric_and_dram_bandwidth_vmax0p9; /*GB/s*/ ++ float fabric_and_dram_bandwidth_vnom0p8; /*GB/s*/ ++ float fabric_and_dram_bandwidth_vmid0p72; /*GB/s*/ ++ float fabric_and_dram_bandwidth_vmin0p65; /*GB/s*/ ++ float phyclkv_max0p9; /*MHz*/ ++ float phyclkv_nom0p8; /*MHz*/ ++ float phyclkv_mid0p72; /*MHz*/ ++ float phyclkv_min0p65; /*MHz*/ ++ float downspreading; /*%*/ ++ int round_trip_ping_latency_cycles; /*DCFCLK Cycles*/ ++ int urgent_out_of_order_return_per_channel; /*bytes*/ ++ int number_of_channels; ++ int vmm_page_size; /*bytes*/ ++ float dram_clock_change_latency; /*us*/ ++ int return_bus_width; /*bytes*/ ++}; ++extern const struct dcn_soc_bounding_box dcn10_soc_defaults; ++ ++struct dcn_ip_params { ++ float rob_buffer_size_in_kbyte; ++ float det_buffer_size_in_kbyte; ++ float dpp_output_buffer_pixels; ++ float opp_output_buffer_lines; ++ float pixel_chunk_size_in_kbyte; ++ enum dcn_bw_defs pte_enable; ++ int pte_chunk_size; /*kbytes*/ ++ int meta_chunk_size; /*kbytes*/ ++ int writeback_chunk_size; /*kbytes*/ ++ enum dcn_bw_defs odm_capability; ++ enum dcn_bw_defs dsc_capability; ++ int line_buffer_size; /*bit*/ ++ int max_line_buffer_lines; ++ enum dcn_bw_defs is_line_buffer_bpp_fixed; ++ int line_buffer_fixed_bpp; ++ int writeback_luma_buffer_size; /*kbytes*/ ++ int writeback_chroma_buffer_size; /*kbytes*/ ++ int max_num_dpp; ++ int max_num_writeback; ++ int max_dchub_topscl_throughput; /*pixels/dppclk*/ ++ int max_pscl_tolb_throughput; /*pixels/dppclk*/ ++ int max_lb_tovscl_throughput; /*pixels/dppclk*/ ++ int max_vscl_tohscl_throughput; /*pixels/dppclk*/ ++ float max_hscl_ratio; ++ float max_vscl_ratio; ++ int max_hscl_taps; ++ int max_vscl_taps; ++ int pte_buffer_size_in_requests; ++ float dispclk_ramping_margin; /*%*/ ++ float under_scan_factor; ++ int max_inter_dcn_tile_repeaters; ++ enum dcn_bw_defs can_vstartup_lines_exceed_vsync_plus_back_porch_lines_minus_one; ++ enum dcn_bw_defs bug_forcing_luma_and_chroma_request_to_same_size_fixed; ++ int dcfclk_cstate_latency; ++}; ++extern const struct dcn_ip_params dcn10_ip_defaults; ++ ++bool dcn_validate_bandwidth( ++ const struct core_dc *dc, ++ struct validate_context *context); ++ ++unsigned int dcn_find_dcfclk_suits_all( ++ const struct core_dc *dc, ++ struct clocks_value *clocks); ++ ++void dcn_bw_update_from_pplib(struct core_dc *dc); ++void dcn_bw_notify_pplib_of_wm_ranges(struct core_dc *dc); ++void dcn_bw_sync_calcs_and_dml(struct core_dc *dc); ++ ++#endif /* __DCN_CALCS_H__ */ ++ +-- +2.7.4 + |