diff options
| author | Dave Airlie <airlied@redhat.com> | 2022-11-22 13:41:11 +1000 |
|---|---|---|
| committer | Dave Airlie <airlied@redhat.com> | 2022-11-22 13:41:11 +1000 |
| commit | fc58764bbf602b65a6f63c53e5fd6feae76c510c (patch) | |
| tree | 03f5448cf5d742b8fd2980e86a33636026557ac6 /drivers/gpu/drm/amd/display/dc/dml | |
| parent | 819683a1fc2f7e64017d50caf539e7bafcb37b81 (diff) | |
| parent | aec3bb3a01de09058fbebed4821ed7d07e1ed994 (diff) | |
| download | linux-fc58764bbf602b65a6f63c53e5fd6feae76c510c.tar.gz linux-fc58764bbf602b65a6f63c53e5fd6feae76c510c.tar.bz2 linux-fc58764bbf602b65a6f63c53e5fd6feae76c510c.zip | |
Merge tag 'amd-drm-next-6.2-2022-11-18' of https://gitlab.freedesktop.org/agd5f/linux into drm-next
amd-drm-next-6.2-2022-11-18:
amdgpu:
- SR-IOV fixes
- Clean up DC checks
- DCN 3.2.x fixes
- DCN 3.1.x fixes
- Don't enable degamma on asics which don't support it
- IP discovery fixes
- BACO fixes
- Fix vbios allocation handling when vkms is enabled
- Drop buggy tdr advanced mode GPU reset handling
- Fix the build when DCN is not set in kconfig
- MST DSC fixes
- Userptr fixes
- FRU and RAS EEPROM fixes
- VCN 4.x RAS support
- Aldrebaran CU occupancy reporting fix
- PSP ring cleanup
amdkfd:
- Memory limit fix
- Enable cooperative launch on gfx 10.3
amd-drm-next-6.2-2022-11-11:
amdgpu:
- SMU 13.x updates
- GPUVM TLB race fix
- DCN 3.1.4 updates
- DCN 3.2.x updates
- PSR fixes
- Kerneldoc fix
- Vega10 fan fix
- GPUVM locking fixes in error pathes
- BACO fix for Beige Goby
- EEPROM I2C address cleanup
- GFXOFF fix
- Fix DC memory leak in error pathes
- Flexible array updates
- Mtype fix for GPUVM PTEs
- Move Kconfig into amdgpu directory
- SR-IOV updates
- Fix possible memory leak in CS IOCTL error path
amdkfd:
- Fix possible memory overrun
- CRIU fixes
radeon:
- ACPI ref count fix
- HDA audio notifier support
- Move Kconfig into radeon directory
UAPI:
- Add new GEM_CREATE flags to help to transition more KFD functionality to the DRM UAPI.
These are used internally in the driver to align location based memory coherency
requirements from memory allocated in the KFD with how we manage GPUVM PTEs. They
are currently blocked in the GEM_CREATE IOCTL as we don't have a user right now.
They are just used internally in the kernel driver for now for existing KFD memory
allocations. So a change to the UAPI header, but no functional change in the UAPI.
From: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221118170807.6505-1-alexander.deucher@amd.com
Signed-off-by: Dave Airlie <airlied@redhat.com>
Diffstat (limited to 'drivers/gpu/drm/amd/display/dc/dml')
13 files changed, 175 insertions, 26 deletions
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn20/dcn20_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn20/dcn20_fpu.c index 602e885ed52c..75dbb7ee193b 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn20/dcn20_fpu.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn20/dcn20_fpu.c @@ -1296,6 +1296,8 @@ int dcn20_populate_dml_pipes_from_context( case SIGNAL_TYPE_DISPLAY_PORT_MST: case SIGNAL_TYPE_DISPLAY_PORT: pipes[pipe_cnt].dout.output_type = dm_dp; + if (is_dp_128b_132b_signal(&res_ctx->pipe_ctx[i])) + pipes[pipe_cnt].dout.output_type = dm_dp2p0; break; case SIGNAL_TYPE_EDP: pipes[pipe_cnt].dout.output_type = dm_edp; diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn31/dcn31_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn31/dcn31_fpu.c index 7dd0845d1bd9..12b23bd50e19 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn31/dcn31_fpu.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn31/dcn31_fpu.c @@ -483,7 +483,7 @@ void dcn31_calculate_wm_and_dlg_fp( int pipe_cnt, int vlevel) { - int i, pipe_idx, active_dpp_count = 0; + int i, pipe_idx, active_hubp_count = 0; double dcfclk = context->bw_ctx.dml.vba.DCFCLKState[vlevel][context->bw_ctx.dml.vba.maxMpcComb]; dc_assert_fp_enabled(); @@ -529,7 +529,7 @@ void dcn31_calculate_wm_and_dlg_fp( continue; if (context->res_ctx.pipe_ctx[i].plane_state) - active_dpp_count++; + active_hubp_count++; pipes[pipe_idx].clks_cfg.dispclk_mhz = get_dispclk_calculated(&context->bw_ctx.dml, pipes, pipe_cnt); pipes[pipe_idx].clks_cfg.dppclk_mhz = get_dppclk_calculated(&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx); @@ -547,9 +547,19 @@ void dcn31_calculate_wm_and_dlg_fp( } dcn20_calculate_dlg_params(dc, context, pipes, pipe_cnt, vlevel); - /* For 31x apu pstate change is only supported if possible in vactive or if there are no active dpps */ + /* For 31x apu pstate change is only supported if possible in vactive*/ context->bw_ctx.bw.dcn.clk.p_state_change_support = - context->bw_ctx.dml.vba.DRAMClockChangeSupport[vlevel][context->bw_ctx.dml.vba.maxMpcComb] == dm_dram_clock_change_vactive || !active_dpp_count; + context->bw_ctx.dml.vba.DRAMClockChangeSupport[vlevel][context->bw_ctx.dml.vba.maxMpcComb] == dm_dram_clock_change_vactive; + /* If DCN isn't making memory requests we can allow pstate change and lower clocks */ + if (!active_hubp_count) { + context->bw_ctx.bw.dcn.clk.socclk_khz = 0; + context->bw_ctx.bw.dcn.clk.dppclk_khz = 0; + context->bw_ctx.bw.dcn.clk.dcfclk_khz = 0; + context->bw_ctx.bw.dcn.clk.dcfclk_deep_sleep_khz = 0; + context->bw_ctx.bw.dcn.clk.dramclk_khz = 0; + context->bw_ctx.bw.dcn.clk.fclk_khz = 0; + context->bw_ctx.bw.dcn.clk.p_state_change_support = true; + } } void dcn31_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params) @@ -797,3 +807,8 @@ void dcn316_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_param else dml_init_instance(&dc->dml, &dcn3_16_soc, &dcn3_16_ip, DML_PROJECT_DCN31_FPGA); } + +int dcn_get_max_non_odm_pix_rate_100hz(struct _vcs_dpi_soc_bounding_box_st *soc) +{ + return soc->clock_limits[0].dispclk_mhz * 10000.0 / (1.0 + soc->dcn_downspread_percent / 100.0); +} diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn31/dcn31_fpu.h b/drivers/gpu/drm/amd/display/dc/dml/dcn31/dcn31_fpu.h index fd58b2561ec9..687d3522cc33 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn31/dcn31_fpu.h +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn31/dcn31_fpu.h @@ -46,5 +46,10 @@ void dcn31_calculate_wm_and_dlg_fp( void dcn31_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params); void dcn315_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params); void dcn316_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params); +int dcn_get_max_non_odm_pix_rate_100hz(struct _vcs_dpi_soc_bounding_box_st *soc); +int dcn31x_populate_dml_pipes_from_context(struct dc *dc, + struct dc_state *context, + display_e2e_pipe_params_st *pipes, + bool fast_validate); #endif /* __DCN31_FPU_H__*/ diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c b/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c index 45ab0ce50860..4e45c6d9ecdc 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn31/display_mode_vba_31.c @@ -1056,14 +1056,12 @@ static bool CalculatePrefetchSchedule( prefetch_bw_pr = dml_min(1, myPipe->VRatio) * prefetch_bw_pr; max_Tsw = dml_max(PrefetchSourceLinesY, PrefetchSourceLinesC) * LineTime; prefetch_sw_bytes = PrefetchSourceLinesY * swath_width_luma_ub * myPipe->BytePerPixelY + PrefetchSourceLinesC * swath_width_chroma_ub * myPipe->BytePerPixelC; - prefetch_bw_oto = dml_max(bytes_pp * myPipe->PixelClock / myPipe->DPPPerPlane, prefetch_sw_bytes / (dml_max(PrefetchSourceLinesY, PrefetchSourceLinesC) * LineTime)); prefetch_bw_oto = dml_max(prefetch_bw_pr, prefetch_sw_bytes / max_Tsw); min_Lsw = dml_max(1, dml_max(PrefetchSourceLinesY, PrefetchSourceLinesC) / max_vratio_pre); Lsw_oto = dml_ceil(4 * dml_max(prefetch_sw_bytes / prefetch_bw_oto / LineTime, min_Lsw), 1) / 4; Tsw_oto = Lsw_oto * LineTime; - prefetch_bw_oto = (PrefetchSourceLinesY * swath_width_luma_ub * myPipe->BytePerPixelY + PrefetchSourceLinesC * swath_width_chroma_ub * myPipe->BytePerPixelC) / Tsw_oto; #ifdef __DML_VBA_DEBUG__ dml_print("DML: HTotal: %d\n", myPipe->HTotal); @@ -5362,6 +5360,58 @@ void dml31_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l v->ModeSupport[i][j] = true; } else { v->ModeSupport[i][j] = false; +#ifdef __DML_VBA_DEBUG__ + if (v->ScaleRatioAndTapsSupport == false) + dml_print("DML SUPPORT: ScaleRatioAndTapsSupport failed"); + if (v->SourceFormatPixelAndScanSupport == false) + dml_print("DML SUPPORT: SourceFormatPixelAndScanSupport failed"); + if (v->ViewportSizeSupport[i][j] == false) + dml_print("DML SUPPORT: ViewportSizeSupport failed"); + if (v->LinkCapacitySupport[i] == false) + dml_print("DML SUPPORT: LinkCapacitySupport failed"); + if (v->ODMCombine4To1SupportCheckOK[i] == false) + dml_print("DML SUPPORT: DSC422NativeNotSupported failed"); + if (v->NotEnoughDSCUnits[i] == true) + dml_print("DML SUPPORT: NotEnoughDSCUnits"); + if (v->DTBCLKRequiredMoreThanSupported[i] == true) + dml_print("DML SUPPORT: DTBCLKRequiredMoreThanSupported"); + if (v->ROBSupport[i][j] == false) + dml_print("DML SUPPORT: ROBSupport failed"); + if (v->DISPCLK_DPPCLK_Support[i][j] == false) + dml_print("DML SUPPORT: DISPCLK_DPPCLK_Support failed"); + if (v->TotalAvailablePipesSupport[i][j] == false) + dml_print("DML SUPPORT: DSC422NativeNotSupported failed"); + if (EnoughWritebackUnits == false) + dml_print("DML SUPPORT: DSC422NativeNotSupported failed"); + if (v->WritebackLatencySupport == false) + dml_print("DML SUPPORT: WritebackLatencySupport failed"); + if (v->WritebackScaleRatioAndTapsSupport == false) + dml_print("DML SUPPORT: DSC422NativeNotSupported "); + if (v->CursorSupport == false) + dml_print("DML SUPPORT: DSC422NativeNotSupported failed"); + if (v->PitchSupport == false) + dml_print("DML SUPPORT: PitchSupport failed"); + if (ViewportExceedsSurface == true) + dml_print("DML SUPPORT: ViewportExceedsSurface failed"); + if (v->PrefetchSupported[i][j] == false) + dml_print("DML SUPPORT: PrefetchSupported failed"); + if (v->DynamicMetadataSupported[i][j] == false) + dml_print("DML SUPPORT: DSC422NativeNotSupported failed"); + if (v->TotalVerticalActiveBandwidthSupport[i][j] == false) + dml_print("DML SUPPORT: TotalVerticalActiveBandwidthSupport failed"); + if (v->VRatioInPrefetchSupported[i][j] == false) + dml_print("DML SUPPORT: VRatioInPrefetchSupported failed"); + if (v->PTEBufferSizeNotExceeded[i][j] == false) + dml_print("DML SUPPORT: PTEBufferSizeNotExceeded failed"); + if (v->NonsupportedDSCInputBPC == true) + dml_print("DML SUPPORT: NonsupportedDSCInputBPC failed"); + if (!((v->HostVMEnable == false + && v->ImmediateFlipRequirement[0] != dm_immediate_flip_required) + || v->ImmediateFlipSupportedForState[i][j] == true)) + dml_print("DML SUPPORT: ImmediateFlipRequirement failed"); + if (FMTBufferExceeded == true) + dml_print("DML SUPPORT: FMTBufferExceeded failed"); +#endif } } } diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn314/dcn314_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn314/dcn314_fpu.c index cf420ad2b8dc..1dd51c4b6804 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn314/dcn314_fpu.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn314/dcn314_fpu.c @@ -29,6 +29,7 @@ #include "dcn31/dcn31_hubbub.h" #include "dcn314_fpu.h" #include "dml/dcn20/dcn20_fpu.h" +#include "dml/dcn31/dcn31_fpu.h" #include "dml/display_mode_vba.h" struct _vcs_dpi_ip_params_st dcn3_14_ip = { @@ -146,8 +147,8 @@ struct _vcs_dpi_soc_bounding_box_st dcn3_14_soc = { }, }, .num_states = 5, - .sr_exit_time_us = 9.0, - .sr_enter_plus_exit_time_us = 11.0, + .sr_exit_time_us = 16.5, + .sr_enter_plus_exit_time_us = 18.5, .sr_exit_z8_time_us = 442.0, .sr_enter_plus_exit_z8_time_us = 560.0, .writeback_latency_us = 12.0, @@ -264,11 +265,8 @@ void dcn314_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_p dc->dml.soc.dispclk_dppclk_vco_speed_mhz = max_dispclk_mhz * 2; } - if ((int)(dcn3_14_soc.dram_clock_change_latency_us * 1000) - != dc->debug.dram_clock_change_latency_ns - && dc->debug.dram_clock_change_latency_ns) { - dcn3_14_soc.dram_clock_change_latency_us = dc->debug.dram_clock_change_latency_ns / 1000; - } + dcn20_patch_bounding_box(dc, &dcn3_14_soc); + if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) dml_init_instance(&dc->dml, &dcn3_14_soc, &dcn3_14_ip, DML_PROJECT_DCN314); else @@ -291,7 +289,7 @@ int dcn314_populate_dml_pipes_from_context_fpu(struct dc *dc, struct dc_state *c dc_assert_fp_enabled(); - dcn20_populate_dml_pipes_from_context(dc, context, pipes, fast_validate); + dcn31x_populate_dml_pipes_from_context(dc, context, pipes, fast_validate); for (i = 0, pipe_cnt = 0; i < dc->res_pool->pipe_count; i++) { struct dc_crtc_timing *timing; @@ -318,8 +316,6 @@ int dcn314_populate_dml_pipes_from_context_fpu(struct dc *dc, struct dc_state *c pipes[pipe_cnt].pipe.src.immediate_flip = true; pipes[pipe_cnt].pipe.src.unbounded_req_mode = false; - pipes[pipe_cnt].pipe.src.hostvm = dc->res_pool->hubbub->riommu_active; - pipes[pipe_cnt].pipe.src.gpuvm = true; pipes[pipe_cnt].pipe.src.dcc_fraction_of_zs_req_luma = 0; pipes[pipe_cnt].pipe.src.dcc_fraction_of_zs_req_chroma = 0; pipes[pipe_cnt].pipe.dest.vfront_porch = timing->v_front_porch; diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_mode_vba_314.c b/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_mode_vba_314.c index 53e3e7364ec6..41f0b4c1c72f 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_mode_vba_314.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_mode_vba_314.c @@ -1078,14 +1078,12 @@ static bool CalculatePrefetchSchedule( prefetch_bw_pr = dml_min(1, myPipe->VRatio) * prefetch_bw_pr; max_Tsw = dml_max(PrefetchSourceLinesY, PrefetchSourceLinesC) * LineTime; prefetch_sw_bytes = PrefetchSourceLinesY * swath_width_luma_ub * myPipe->BytePerPixelY + PrefetchSourceLinesC * swath_width_chroma_ub * myPipe->BytePerPixelC; - prefetch_bw_oto = dml_max(bytes_pp * myPipe->PixelClock / myPipe->DPPPerPlane, prefetch_sw_bytes / (dml_max(PrefetchSourceLinesY, PrefetchSourceLinesC) * LineTime)); prefetch_bw_oto = dml_max(prefetch_bw_pr, prefetch_sw_bytes / max_Tsw); min_Lsw = dml_max(1, dml_max(PrefetchSourceLinesY, PrefetchSourceLinesC) / max_vratio_pre); Lsw_oto = dml_ceil(4 * dml_max(prefetch_sw_bytes / prefetch_bw_oto / LineTime, min_Lsw), 1) / 4; Tsw_oto = Lsw_oto * LineTime; - prefetch_bw_oto = (PrefetchSourceLinesY * swath_width_luma_ub * myPipe->BytePerPixelY + PrefetchSourceLinesC * swath_width_chroma_ub * myPipe->BytePerPixelC) / Tsw_oto; #ifdef __DML_VBA_DEBUG__ dml_print("DML: HTotal: %d\n", myPipe->HTotal); diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c index 0d704e302d03..97b333b230d1 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c @@ -531,9 +531,12 @@ void dcn32_set_phantom_stream_timing(struct dc *dc, unsigned int i, pipe_idx; struct pipe_ctx *pipe; uint32_t phantom_vactive, phantom_bp, pstate_width_fw_delay_lines; + unsigned int num_dpp; unsigned int vlevel = context->bw_ctx.dml.vba.VoltageLevel; unsigned int dcfclk = context->bw_ctx.dml.vba.DCFCLKState[vlevel][context->bw_ctx.dml.vba.maxMpcComb]; unsigned int socclk = context->bw_ctx.dml.vba.SOCCLKPerState[vlevel]; + struct vba_vars_st *vba = &context->bw_ctx.dml.vba; + struct dc_stream_state *main_stream = ref_pipe->stream; dc_assert_fp_enabled(); @@ -569,13 +572,23 @@ void dcn32_set_phantom_stream_timing(struct dc *dc, phantom_vactive = get_subviewport_lines_needed_in_mall(&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx) + pstate_width_fw_delay_lines + dc->caps.subvp_swath_height_margin_lines; + // W/A for DCC corruption with certain high resolution timings. + // Determing if pipesplit is used. If so, add meta_row_height to the phantom vactive. + num_dpp = vba->NoOfDPP[vba->VoltageLevel][vba->maxMpcComb][vba->pipe_plane[pipe_idx]]; + phantom_vactive += num_dpp > 1 ? vba->meta_row_height[vba->pipe_plane[pipe_idx]] : 0; + // For backporch of phantom pipe, use vstartup of the main pipe phantom_bp = get_vstartup(&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx); phantom_stream->dst.y = 0; phantom_stream->dst.height = phantom_vactive; + /* When scaling, DML provides the end to end required number of lines for MALL. + * dst.height is always correct for this case, but src.height is not which causes a + * delta between main and phantom pipe scaling outputs. Need to adjust src.height on + * phantom for this case. + */ phantom_stream->src.y = 0; - phantom_stream->src.height = phantom_vactive; + phantom_stream->src.height = (double)phantom_vactive * (double)main_stream->src.height / (double)main_stream->dst.height; phantom_stream->timing.v_addressable = phantom_vactive; phantom_stream->timing.v_front_porch = 1; @@ -1228,7 +1241,7 @@ static void dcn32_calculate_dlg_params(struct dc *dc, struct dc_state *context, display_e2e_pipe_params_st *pipes, int pipe_cnt, int vlevel) { - int i, pipe_idx; + int i, pipe_idx, active_hubp_count = 0; bool usr_retraining_support = false; bool unbounded_req_enabled = false; @@ -1273,6 +1286,8 @@ static void dcn32_calculate_dlg_params(struct dc *dc, struct dc_state *context, for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) { if (!context->res_ctx.pipe_ctx[i].stream) continue; + if (context->res_ctx.pipe_ctx[i].plane_state) + active_hubp_count++; pipes[pipe_idx].pipe.dest.vstartup_start = get_vstartup(&context->bw_ctx.dml, pipes, pipe_cnt, pipe_idx); pipes[pipe_idx].pipe.dest.vupdate_offset = get_vupdate_offset(&context->bw_ctx.dml, pipes, pipe_cnt, @@ -1298,6 +1313,16 @@ static void dcn32_calculate_dlg_params(struct dc *dc, struct dc_state *context, context->res_ctx.pipe_ctx[i].pipe_dlg_param = pipes[pipe_idx].pipe.dest; pipe_idx++; } + /* If DCN isn't making memory requests we can allow pstate change and lower clocks */ + if (!active_hubp_count) { + context->bw_ctx.bw.dcn.clk.socclk_khz = 0; + context->bw_ctx.bw.dcn.clk.dppclk_khz = 0; + context->bw_ctx.bw.dcn.clk.dcfclk_khz = 0; + context->bw_ctx.bw.dcn.clk.dcfclk_deep_sleep_khz = 0; + context->bw_ctx.bw.dcn.clk.dramclk_khz = 0; + context->bw_ctx.bw.dcn.clk.fclk_khz = 0; + context->bw_ctx.bw.dcn.clk.p_state_change_support = true; + } /*save a original dppclock copy*/ context->bw_ctx.bw.dcn.clk.bw_dppclk_khz = context->bw_ctx.bw.dcn.clk.dppclk_khz; context->bw_ctx.bw.dcn.clk.bw_dispclk_khz = context->bw_ctx.bw.dcn.clk.dispclk_khz; @@ -1729,6 +1754,9 @@ bool dcn32_internal_validate_bw(struct dc *dc, } if (repopulate_pipes) { + int flag_max_mpc_comb = vba->maxMpcComb; + int flag_vlevel = vlevel; + pipe_cnt = dc->res_pool->funcs->populate_dml_pipes(dc, context, pipes, fast_validate); /* repopulate_pipes = 1 means the pipes were either split or merged. In this case @@ -1742,6 +1770,22 @@ bool dcn32_internal_validate_bw(struct dc *dc, if (vlevel == context->bw_ctx.dml.soc.num_states) { /* failed after DET size changes */ goto validate_fail; + } else if (flag_max_mpc_comb == 0 && + flag_max_mpc_comb != context->bw_ctx.dml.vba.maxMpcComb) { + /* check the context constructed with pipe split flags is still valid*/ + bool flags_valid = false; + for (int i = flag_vlevel; i < context->bw_ctx.dml.soc.num_states; i++) { + if (vba->ModeSupport[i][flag_max_mpc_comb]) { + vba->maxMpcComb = flag_max_mpc_comb; + vba->VoltageLevel = i; + vlevel = i; + flags_valid = true; + } + } + + /* this should never happen */ + if (!flags_valid) + goto validate_fail; } } *vlevel_out = vlevel; @@ -1800,6 +1844,12 @@ void dcn32_calculate_wm_and_dlg_fpu(struct dc *dc, struct dc_state *context, */ context->bw_ctx.dml.soc.dram_clock_change_latency_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_A].dml_input.pstate_latency_us; + /* For DCN32/321 need to validate with fclk pstate change latency equal to dummy so + * prefetch is scheduled correctly to account for dummy pstate. + */ + if (dummy_latency_index == 0) + context->bw_ctx.dml.soc.fclk_change_latency_us = + dc->clk_mgr->bw_params->dummy_pstate_table[dummy_latency_index].dummy_pstate_latency_us; dcn32_internal_validate_bw(dc, context, pipes, &pipe_cnt, &vlevel, false); maxMpcComb = context->bw_ctx.dml.vba.maxMpcComb; dcfclk_from_fw_based_mclk_switching = context->bw_ctx.dml.vba.DCFCLKState[vlevel][context->bw_ctx.dml.vba.maxMpcComb]; @@ -1987,6 +2037,10 @@ void dcn32_calculate_wm_and_dlg_fpu(struct dc *dc, struct dc_state *context, context->perf_params.stutter_period_us = context->bw_ctx.dml.vba.StutterPeriod; + if (context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching && dummy_latency_index == 0) + context->bw_ctx.dml.soc.fclk_change_latency_us = + dc->clk_mgr->bw_params->dummy_pstate_table[dummy_latency_index].dummy_pstate_latency_us; + dcn32_calculate_dlg_params(dc, context, pipes, pipe_cnt, vlevel); if (!pstate_en) @@ -1994,8 +2048,12 @@ void dcn32_calculate_wm_and_dlg_fpu(struct dc *dc, struct dc_state *context, context->bw_ctx.dml.soc.dram_clock_change_latency_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_A].dml_input.pstate_latency_us; - if (context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching) + if (context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching) { dcn30_setup_mclk_switch_using_fw_based_vblank_stretch(dc, context); + if (dummy_latency_index == 0) + context->bw_ctx.dml.soc.fclk_change_latency_us = + dc->clk_mgr->bw_params->wm_table.nv_entries[WM_A].dml_input.fclk_change_latency_us; + } } static void dcn32_get_optimal_dcfclk_fclk_for_uclk(unsigned int uclk_mts, @@ -2351,6 +2409,8 @@ void dcn32_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_pa /* DML DSC delay factor workaround */ dcn3_2_ip.dsc_delay_factor_wa = dc->debug.dsc_delay_factor_wa_x1000 / 1000.0; + dcn3_2_ip.min_prefetch_in_strobe_us = dc->debug.min_prefetch_in_strobe_ns / 1000.0; + /* Override dispclk_dppclk_vco_speed_mhz from Clk Mgr */ dcn3_2_soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0; dc->dml.soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0; diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_32.c b/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_32.c index 3d184679f129..e5c8f6a71b5b 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_32.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_32.c @@ -718,6 +718,8 @@ static void DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerforman do { MaxTotalRDBandwidth = 0; + DestinationLineTimesForPrefetchLessThan2 = false; + VRatioPrefetchMoreThanMax = false; #ifdef __DML_VBA_DEBUG__ dml_print("DML::%s: Start loop: VStartup = %d\n", __func__, mode_lib->vba.VStartupLines); #endif @@ -786,6 +788,8 @@ static void DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerforman v->SwathHeightY[k], v->SwathHeightC[k], TWait, + v->DRAMSpeedPerState[mode_lib->vba.VoltageLevel] <= MEM_STROBE_FREQ_MHZ ? + mode_lib->vba.ip.min_prefetch_in_strobe_us : 0, /* Output */ &v->DSTXAfterScaler[k], &v->DSTYAfterScaler[k], @@ -2282,7 +2286,7 @@ void dml32_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l && (mode_lib->vba.Output[k] == dm_dp || mode_lib->vba.Output[k] == dm_dp2p0 || mode_lib->vba.Output[k] == dm_edp || mode_lib->vba.Output[k] == dm_hdmi) - && mode_lib->vba.OutputBppPerState[i][k] == 0) { + && mode_lib->vba.OutputBppPerState[i][k] == 0 && (mode_lib->vba.UsesMALLForPStateChange[k] != dm_use_mall_pstate_change_phantom_pipe)) { mode_lib->vba.LinkCapacitySupport[i] = false; } } @@ -3192,6 +3196,7 @@ void dml32_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l mode_lib->vba.FCLKChangeLatency, mode_lib->vba.UrgLatency[i], mode_lib->vba.SREnterPlusExitTime); + memset(&v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull, 0, sizeof(DmlPipe)); v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.myPipe.Dppclk = mode_lib->vba.RequiredDPPCLK[i][j][k]; v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.myPipe.Dispclk = mode_lib->vba.RequiredDISPCLK[i][j]; v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.myPipe.PixelClock = mode_lib->vba.PixelClock[k]; @@ -3244,6 +3249,8 @@ void dml32_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_l v->swath_width_chroma_ub_this_state[k], v->SwathHeightYThisState[k], v->SwathHeightCThisState[k], v->TWait, + v->DRAMSpeedPerState[i] <= MEM_STROBE_FREQ_MHZ ? + mode_lib->vba.ip.min_prefetch_in_strobe_us : 0, /* Output */ &v->dummy_vars.dml32_ModeSupportAndSystemConfigurationFull.DSTXAfterScaler[k], diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_32.h b/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_32.h index c62e0991358b..c8b28c83ddf4 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_32.h +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_32.h @@ -46,9 +46,14 @@ // Prefetch schedule max vratio #define __DML_MAX_VRATIO_PRE__ 4.0 +#define __DML_VBA_MAX_DST_Y_PRE__ 63.75 + #define BPP_INVALID 0 #define BPP_BLENDED_PIPE 0xffffffff +#define MEM_STROBE_FREQ_MHZ 1600 +#define MEM_STROBE_MAX_DELIVERY_TIME_US 60.0 + struct display_mode_lib; void dml32_ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_lib); diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_util_32.c b/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_util_32.c index 968924c491c1..debe46b24a3e 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_util_32.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_util_32.c @@ -3417,6 +3417,7 @@ bool dml32_CalculatePrefetchSchedule( unsigned int SwathHeightY, unsigned int SwathHeightC, double TWait, + double TPreReq, /* Output */ double *DSTXAfterScaler, double *DSTYAfterScaler, @@ -3667,6 +3668,7 @@ bool dml32_CalculatePrefetchSchedule( dst_y_prefetch_equ = VStartup - (*TSetup + dml_max(TWait + TCalc, *Tdmdl)) / LineTime - (*DSTYAfterScaler + (double) *DSTXAfterScaler / (double) myPipe->HTotal); + dst_y_prefetch_equ = dml_min(dst_y_prefetch_equ, __DML_VBA_MAX_DST_Y_PRE__); #ifdef __DML_VBA_DEBUG__ dml_print("DML::%s: HTotal = %d\n", __func__, myPipe->HTotal); dml_print("DML::%s: min_Lsw = %f\n", __func__, min_Lsw); @@ -3726,7 +3728,8 @@ bool dml32_CalculatePrefetchSchedule( *VRatioPrefetchY = 0; *VRatioPrefetchC = 0; *RequiredPrefetchPixDataBWLuma = 0; - if (dst_y_prefetch_equ > 1) { + if (dst_y_prefetch_equ > 1 && + (Tpre_rounded >= TPreReq || dst_y_prefetch_equ == __DML_VBA_MAX_DST_Y_PRE__)) { double PrefetchBandwidth1; double PrefetchBandwidth2; double PrefetchBandwidth3; @@ -3872,7 +3875,11 @@ bool dml32_CalculatePrefetchSchedule( } if (dst_y_prefetch_oto < dst_y_prefetch_equ) { - *DestinationLinesForPrefetch = dst_y_prefetch_oto; + if (dst_y_prefetch_oto * LineTime < TPreReq) { + *DestinationLinesForPrefetch = dst_y_prefetch_equ; + } else { + *DestinationLinesForPrefetch = dst_y_prefetch_oto; + } TimeForFetchingMetaPTE = Tvm_oto; TimeForFetchingRowInVBlank = Tr0_oto; *PrefetchBandwidth = prefetch_bw_oto; @@ -4397,7 +4404,7 @@ void dml32_CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport( if (v->NumberOfActiveSurfaces > 1) { ActiveClockChangeLatencyHidingY = ActiveClockChangeLatencyHidingY - - (1 - 1 / v->NumberOfActiveSurfaces) * SwathHeightY[k] * v->HTotal[k] + - (1.0 - 1.0 / v->NumberOfActiveSurfaces) * SwathHeightY[k] * v->HTotal[k] / v->PixelClock[k] / v->VRatio[k]; } diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_util_32.h b/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_util_32.h index 2c3827546ac7..3989c2a28fae 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_util_32.h +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn32/display_mode_vba_util_32.h @@ -30,7 +30,7 @@ #include "os_types.h" #include "../dc_features.h" #include "../display_mode_structs.h" -#include "dml/display_mode_vba.h" +#include "../display_mode_vba.h" unsigned int dml32_dscceComputeDelay( unsigned int bpc, @@ -743,6 +743,7 @@ bool dml32_CalculatePrefetchSchedule( unsigned int SwathHeightY, unsigned int SwathHeightC, double TWait, + double TPreReq, /* Output */ double *DSTXAfterScaler, double *DSTYAfterScaler, diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn321/dcn321_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn321/dcn321_fpu.c index ec0486efab14..432b4ecd01a7 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn321/dcn321_fpu.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn321/dcn321_fpu.c @@ -544,6 +544,8 @@ void dcn321_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_p /* DML DSC delay factor workaround */ dcn3_21_ip.dsc_delay_factor_wa = dc->debug.dsc_delay_factor_wa_x1000 / 1000.0; + dcn3_21_ip.min_prefetch_in_strobe_us = dc->debug.min_prefetch_in_strobe_ns / 1000.0; + /* Override dispclk_dppclk_vco_speed_mhz from Clk Mgr */ dcn3_21_soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0; dc->dml.soc.dispclk_dppclk_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0; diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_mode_structs.h b/drivers/gpu/drm/amd/display/dc/dml/display_mode_structs.h index d7be01ac0751..64d602e6412f 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/display_mode_structs.h +++ b/drivers/gpu/drm/amd/display/dc/dml/display_mode_structs.h @@ -367,6 +367,7 @@ struct _vcs_dpi_ip_params_st { /* DM workarounds */ double dsc_delay_factor_wa; // TODO: Remove after implementing root cause fix + double min_prefetch_in_strobe_us; }; struct _vcs_dpi_display_xfc_params_st { |