DC Patches 15 Jul, 2019
cover DC Patches 15 Jul, 2019
From: Leo Li <[email protected]>
Lots of patches this time, sorry for the volume! Regular DC promotions were put on hold while Navi was being reviewed and upstreamed. These patches were what accumulated during that period.
Summary of change:
* Implement DPRX CRC source capture
* Implement CRC source readback
* Navi Bandwidth calcs refactor
* Misc fixes for Navi:
* Dual GPU boot black screen
* Flip timeouts
* Workaround for p-state hangs on high-bandwidth timings
Alvin Lee (4):
drm/amd/display: Disable Audio on reinitialize hardware drm/amd/display: Remove second initialization of pp_smu drm/amd/display: Wait for flip to complete drm/amd/display: Only enable audio if speaker allocation exists
Anthony Koo (1):
drm/amd/display: add monitor patch to add T7 delay
Aric Cyr (7):
drm/amd/display: 3.2.36
drm/amd/display: 3.2.37
drm/amd/display: 3.2.38
drm/amd/display: 3.2.39
drm/amd/display: 3.2.40
drm/amd/display: 3.2.41
drm/amd/display: 3.2.42
Charlene Liu (4):
drm/amd/display: Split out common HUBP registers and code drm/amd/display: Do not fill Null packet in the blank period drm/amd/display: add set and get clock for testing purposes drm/amd/display: add a option to force the clock at every mode change.
Chiawen Huang (1):
drm/amd/display: Add aux tracing log in dce
Dale Zhao (1):
drm/amd/display: handle active dongle port type is DP++ or DP case
David Francis (1):
drm/amd/display: Update drm_dsc to reflect native 4.2.0 DSC spec
David Galiffi (2):
drm/amd/display: Add ability to set preferred link training
parameters.
drm/amd/display: Incorrect Read Interval Time For CR Sequence
Derek Lai (2):
drm/amd/display: Read max down spread
drm/amd/display: allocate 4 ddc engines for RV2
Dingchen Zhang (3):
drm/amd/display: add functionality to grab DPRX CRC entries. drm/amd/display: add functionality to get pipe CRC source. drm/amd/display: add pipe CRC sources without disabling dithering.
Dmytro Laktyushkin (6):
drm/amd/display: fix dsc disable
drm/amd/display: Set default block_size, even in unexpected cases drm/amd/display: add hdmi2.1 dsc pps packet programming drm/amd/display: Remove dsc disable_ich flag programming. drm/amd/display: use min disp and dpp clk debug option for dcn2 drm/amd/display: add dcc programming for dual plane
Eric Bernstein (1):
drm/amd/display: Use helper for determining HDMI signal
Eric Yang (5):
drm/amd/display: move bw calc code into helpers drm/amd/display: early return when pipe_cnt is 0 in bw validation drm/amd/display: put back front end initialization sequence drm/amd/display: do not read link setting if edp not connected drm/amd/display: fix mpcc assert condition
Fatemeh Darbehani (2):
drm/amd/display: Change min_h_sync_width from 8 to 4 drm/amd/display: Add SMU version field to clk_mgr_internal
Harmanprit Tatla (1):
drm/amd/display: No audio endpoint for Dell MST display
Harry Wentland (1):
drm/amd/display: Remove unnecessary NULL check in
set_preferred_link_settings
Ilya Bakoulin (5):
drm/amd/display: Expose enc2_set_dynamic_metadata drm/amd/display: Check for valid stream_encode drm/amd/display: Fix some HUBP programming issues drm/amd/display: Cache the use_pitch_c conditional drm/amd/display: Fixes for some MPO cases
Joshua Aberback (1):
drm/amd/display: Add debug option to disable timing sync
Julian Parkin (3):
drm/amd/display: Poll for GPUVM context ready drm/amd/display: Fix dc_create failure handling and 666 color depths drm/amd/display: Clean up dynamic metadata logic
Jun Lei (6):
drm/amd/display: initialize p_state to proper value drm/amd/display: fix up HUBBUB hw programming for VM drm/amd/display: cap DCFCLK hardmin to 507 for NV10 drm/amd/display: swap system aperture high/low drm/amd/display: populate last calculated bb state with max clocks drm/amd/display: support "dummy pstate"
Krunoslav Kovac (1):
drm/amd/display: Optimize gamma calculations
Lewis Huang (1):
drm/amd/display: Add debug entry to destroy disconnected edp link
Murton Liu (4):
drm/amd/display: Clock does not lower in Updateplanes drm/amd/display: Implement generic MUX registers drm/amd/display: Hook up calls to do stereo mux and dig programming to
stereo control interface
drm/amd/display: Change offset_to_id to reflect what id_to_offset
returns
Nevenko Stupar (2):
drm/amd/display:Use Pixel clock in 100Hz units for HDMI Audio wall
clock DTO
drm/amd/display: Add DIG_CLOCK_PATTERN register
Nicholas Kazlauskas (4):
drm/amd/display: Copy max_clks_by_state after dce_clk_mgr_construct drm/amd/display: Set enabled to false at start of audio disable drm/amd/display: Copy GSL groups when committing a new context drm/amd/display: Force uclk to max for every state
Nikola Cornij (4):
drm/amd/display: Set one 4:2:0-related PPS field as recommended by DSC
spec
drm/amd/display: Power-gate all DSCs at driver init time drm/amd/display: Set FEC_READY always before link training drm/amd/display: Clear FEC_READY shadow register if DPCD write fails
Qingqing Zhuo (1):
drm/amd/display: Add CM_BYPASS via debug option
Reza Amini (1):
drm/amd/display: Implement DAL3 GPU Integer Scaling
Samson Tam (1):
drm/amd/display: skip retrain in dc_link_set_preferred_link_settings()
if using passive dongle
SivapiriyanKumarasamy (1):
drm/amd/display: Wait for backlight programming completion in set
backlight level
Su Sung Chung (2):
drm/amd/display: refactor dump_clk_registers drm/amd/display: fix not calling ppsmu to trigger PME
Tai Man (2):
drm/amd/display: use encoder's engine id to find matched free audio
device
drm/amd/display: Increase size of audios array
Tony Cheng (1):
drm/amd/display: avoid power gate domains that doesn't exist
Vitaly Prosyak (1):
drm/amd/display: Add MPC 3DLUT resource management
Wenjing Liu (1):
drm/amd/display: wait for the whole frame after global unlock
Zhan Liu (1):
drm/amd/display: drop ASSERT() if eDP panel is not connected
Zi Yu Liao (1):
drm/amd/display: fix DMCU hang when going into Modern Standby
hersen wu (1):
drm/amd/display: init res_pool dccg_ref, dchub_ref with xtalin_freq
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 1 + .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 60 +- .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h | 16 +- .../drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c | 139 +- .../drm/amd/display/amdgpu_dm/amdgpu_dm_crc.h | 61 + .../amd/display/amdgpu_dm/amdgpu_dm_pp_smu.c | 6 +- .../gpu/drm/amd/display/dc/calcs/dce_calcs.c | 2 +- .../gpu/drm/amd/display/dc/calcs/dcn_calcs.c | 2 +- .../dc/clk_mgr/dce110/dce110_clk_mgr.c | 4 +- .../dc/clk_mgr/dce112/dce112_clk_mgr.c | 4 +- .../dc/clk_mgr/dce120/dce120_clk_mgr.c | 4 +- .../display/dc/clk_mgr/dcn20/dcn20_clk_mgr.c | 39 +- .../display/dc/clk_mgr/dcn20/dcn20_clk_mgr.h | 5 + drivers/gpu/drm/amd/display/dc/core/dc.c | 96 +- .../gpu/drm/amd/display/dc/core/dc_debug.c | 40 +- drivers/gpu/drm/amd/display/dc/core/dc_link.c | 101 +- .../gpu/drm/amd/display/dc/core/dc_link_dp.c | 394 +- .../drm/amd/display/dc/core/dc_link_hwss.c | 45 +- .../gpu/drm/amd/display/dc/core/dc_resource.c | 69 +- .../gpu/drm/amd/display/dc/core/dc_stream.c | 26 +- .../gpu/drm/amd/display/dc/core/dc_surface.c | 3 +- drivers/gpu/drm/amd/display/dc/dc.h | 44 +- drivers/gpu/drm/amd/display/dc/dc_dp_types.h | 21 + drivers/gpu/drm/amd/display/dc/dc_hw_types.h | 56 +- drivers/gpu/drm/amd/display/dc/dc_link.h | 11 + drivers/gpu/drm/amd/display/dc/dc_types.h | 14 + drivers/gpu/drm/amd/display/dc/dce/dce_abm.c | 4 + .../gpu/drm/amd/display/dc/dce/dce_audio.c | 28 +- drivers/gpu/drm/amd/display/dc/dce/dce_aux.c | 9 +- .../gpu/drm/amd/display/dc/dce/dce_hwseq.h | 7 +- .../drm/amd/display/dc/dce/dce_mem_input.c | 10 +- .../amd/display/dc/dce/dce_stream_encoder.c | 30 +- .../display/dc/dce110/dce110_hw_sequencer.c | 70 +- .../display/dc/dce110/dce110_mem_input_v.c | 42 +- .../gpu/drm/amd/display/dc/dcn10/dcn10_hubp.c | 70 +- .../gpu/drm/amd/display/dc/dcn10/dcn10_hubp.h | 27 +- .../amd/display/dc/dcn10/dcn10_hw_sequencer.c | 85 +- .../dc/dcn10/dcn10_hw_sequencer_debug.c | 2 +- .../gpu/drm/amd/display/dc/dcn10/dcn10_mpc.c | 2 +- .../drm/amd/display/dc/dcn10/dcn10_resource.c | 2 +- .../display/dc/dcn10/dcn10_stream_encoder.c | 33 +- .../display/dc/dcn10/dcn10_stream_encoder.h | 16 +- .../drm/amd/display/dc/dcn20/dcn20_dpp_cm.c | 7 +- .../gpu/drm/amd/display/dc/dcn20/dcn20_dsc.c | 8 +- .../gpu/drm/amd/display/dc/dcn20/dcn20_dsc.h | 4 +- .../drm/amd/display/dc/dcn20/dcn20_hubbub.c | 29 +- .../gpu/drm/amd/display/dc/dcn20/dcn20_hubp.c | 688 ++- .../gpu/drm/amd/display/dc/dcn20/dcn20_hubp.h | 62 +- .../drm/amd/display/dc/dcn20/dcn20_hwseq.c | 221 +- .../drm/amd/display/dc/dcn20/dcn20_hwseq.h | 14 +- .../gpu/drm/amd/display/dc/dcn20/dcn20_mpc.c | 34 +- .../gpu/drm/amd/display/dc/dcn20/dcn20_optc.c | 2 +- .../drm/amd/display/dc/dcn20/dcn20_resource.c | 356 +- .../drm/amd/display/dc/dcn20/dcn20_resource.h | 12 + .../display/dc/dcn20/dcn20_stream_encoder.c | 2 +- .../display/dc/dcn20/dcn20_stream_encoder.h | 5 + .../gpu/drm/amd/display/dc/dcn20/dcn20_vmid.c | 35 + drivers/gpu/drm/amd/display/dc/dm_services.h | 1 + drivers/gpu/drm/amd/display/dc/dml/Makefile | 3 + .../dc/dml/dcn20/display_mode_vba_20v2.c | 5109 +++++++++++++++++ .../dc/dml/dcn20/display_mode_vba_20v2.h | 32 + .../dc/dml/dcn20/display_rq_dlg_calc_20v2.c | 1701 ++++++ .../dc/dml/dcn20/display_rq_dlg_calc_20v2.h | 74 + .../drm/amd/display/dc/dml/display_mode_lib.c | 12 + .../drm/amd/display/dc/dml/display_mode_lib.h | 1 + .../amd/display/dc/dml/display_mode_structs.h | 1 + .../drm/amd/display/dc/dml/display_mode_vba.c | 8 +- .../gpu/drm/amd/display/dc/dsc/drm_dsc_dc.c | 79 +- drivers/gpu/drm/amd/display/dc/gpio/Makefile | 2 +- .../display/dc/gpio/dcn10/hw_factory_dcn10.c | 42 +- .../display/dc/gpio/dcn20/hw_factory_dcn20.c | 41 +- .../dc/gpio/dcn20/hw_translate_dcn20.c | 2 +- .../dc/gpio/diagnostics/hw_factory_diag.c | 1 + .../drm/amd/display/dc/gpio/generic_regs.h | 66 + .../drm/amd/display/dc/gpio/gpio_service.c | 68 + .../gpu/drm/amd/display/dc/gpio/hw_factory.h | 3 + .../gpu/drm/amd/display/dc/gpio/hw_generic.c | 132 + .../gpu/drm/amd/display/dc/gpio/hw_generic.h | 46 + .../gpu/drm/amd/display/dc/inc/core_status.h | 3 + .../gpu/drm/amd/display/dc/inc/core_types.h | 2 +- .../gpu/drm/amd/display/dc/inc/dc_link_dp.h | 1 + .../gpu/drm/amd/display/dc/inc/hw/clk_mgr.h | 7 + .../amd/display/dc/inc/hw/clk_mgr_internal.h | 3 + .../gpu/drm/amd/display/dc/inc/hw/dchubbub.h | 4 +- drivers/gpu/drm/amd/display/dc/inc/hw/dpp.h | 1 + drivers/gpu/drm/amd/display/dc/inc/hw/hubp.h | 2 +- .../gpu/drm/amd/display/dc/inc/hw/hw_shared.h | 1 + .../gpu/drm/amd/display/dc/inc/hw/mem_input.h | 2 +- drivers/gpu/drm/amd/display/dc/inc/hw/mpc.h | 1 + .../gpu/drm/amd/display/dc/inc/hw_sequencer.h | 10 + .../gpu/drm/amd/display/dc/inc/link_hwss.h | 2 +- .../gpu/drm/amd/display/include/audio_types.h | 4 +- .../gpu/drm/amd/display/include/dpcd_defs.h | 2 +- .../display/include/gpio_service_interface.h | 18 +- .../amd/display/include/link_service_types.h | 17 +- .../amd/display/modules/color/color_gamma.c | 163 +- .../amd/display/modules/color/color_gamma.h | 9 + 97 files changed, 10040 insertions(+), 815 deletions(-) create mode 100644 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.h create mode 100644 drivers/gpu/drm/amd/display/dc/dml/dcn20/display_mode_vba_20v2.c create mode 100644 drivers/gpu/drm/amd/display/dc/dml/dcn20/display_mode_vba_20v2.h create mode 100644 drivers/gpu/drm/amd/display/dc/dml/dcn20/display_rq_dlg_calc_20v2.c create mode 100644 drivers/gpu/drm/amd/display/dc/dml/dcn20/display_rq_dlg_calc_20v2.h create mode 100644 drivers/gpu/drm/amd/display/dc/gpio/generic_regs.h create mode 100644 drivers/gpu/drm/amd/display/dc/gpio/hw_generic.c create mode 100644 drivers/gpu/drm/amd/display/dc/gpio/hw_generic.h
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
A message that is not in the archive (sent off-list or not yet received): [email protected]
PATCH 1/87 drm/amd/display: initialize p_state to proper value
From: Jun Lei <[email protected]>
[why]
On some modes SMU will be in infinite loop state at boot, this is because driver assumes p_state_support is false, but this is the opposite of the assumed boot state by SMU. we optimize away notifying SMU about no pstate, and so they will get stuck
[how]
when we init clk manager, init pstate to true, so it matches driver load assumption
--- drivers/gpu/drm/amd/display/dc/clk_mgr/dcn20/dcn20_clk_mgr.c | 2 ++ 1 file changed, 2 insertions(+)
drivers/gpu/drm/amd/display/dc/clk_mgr/dcn20/dcn20_clk_mgr.c +2 -0
@@ -301,6 +301,8 @@ void dcn2_update_clocks_fpga(struct clk_mgr *clk_mgr,void dcn2_init_clocks (struct clk_mgr *clk_mgr) {memset (&(clk_mgr->clks),0 ,sizeof (struct dc_clocks )); +// Assumption is that boot state always supports pstate + clk_mgr->clks.p_state_change_support = true; }void dcn2_enable_pme_wa (struct clk_mgr *clk_mgr_base)
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 2/87 drm/amd/display: Add ability to set preferred link training parameters.
In reply to a message not in the archiveFrom: David Galiffi <[email protected]>
[WHY]
To add support for OS requirement to set preferred link training parameters.
[HOW]
Create new structure of dp link training overrides. During link training processes, these values should be used instead of the default training parameters.
--- drivers/gpu/drm/amd/display/dc/core/dc_link.c | 46 ++- .../gpu/drm/amd/display/dc/core/dc_link_dp.c | 337 +++++++++++++----- .../drm/amd/display/dc/core/dc_link_hwss.c | 28 +- drivers/gpu/drm/amd/display/dc/dc_dp_types.h | 21 ++ drivers/gpu/drm/amd/display/dc/dc_link.h | 11 + .../gpu/drm/amd/display/dc/inc/link_hwss.h | 2 +- .../amd/display/include/link_service_types.h | 17 +- 7 files changed, 338 insertions(+), 124 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc_link.c +42 -4
@@ -1187,6 +1187,9 @@ static bool construct( link->ctx = dc_ctx; link->link_index = init_params->link_index; +memset (&link->preferred_training_settings,0 ,sizeof (struct dc_link_training_overrides )); +memset (&link->preferred_link_setting,0 ,sizeof (struct dc_link_settings )); + link->link_id = bios->funcs->get_connector_id (bios, init_params->connector_index);if (link->link_id.type !=OBJECT_TYPE_CONNECTOR ) {
@@ -1465,6 +1468,9 @@ static enum dc_status enable_link_dp(struct dc_link *link = stream->link;struct dc_link_settings link_settings = {0 };enum dp_panel_mode panel_mode; +#ifdef CONFIG_DRM_AMD_DC_DSC_SUPPORT +bool fec_enable; +#endif /* get link settings for video mode timing */ decide_link_settings (stream, &link_settings);
@@ -1509,10 +1515,20 @@ static enum dc_status enable_link_dp( skip_video_pattern = false;#ifdef CONFIG_DRM_AMD_DC_DSC_SUPPORT -dp_set_fec_ready (link, true); +if (link->preferred_training_settings.fec_enable !=NULL ) + fec_enable = *link->preferred_training_settings.fec_enable; +else + fec_enable = true; + +dp_set_fec_ready (link, fec_enable);#endif - if (perform_link_training_with_retries( +if (link->aux_access_disabled) { +dc_link_dp_perform_link_training_skip_aux (link, &link_settings); + + link->cur_link_settings = link_settings; + status =DC_OK ; + } else if (perform_link_training_with_retries( link, &link_settings, skip_video_pattern,
@@ -1524,7 +1540,7 @@ static enum dc_status enable_link_dp( status =DC_FAIL_DP_LINK_TRAINING ;#ifdef CONFIG_DRM_AMD_DC_DSC_SUPPORT -dp_set_fec_enable (link, true); +dp_set_fec_enable (link, fec_enable);#endif return status; }
@@ -3012,6 +3028,29 @@ void dc_link_set_preferred_link_settings(struct dc *dc,dp_retrain_link_dp_test (link, &store_settings, false); } +void dc_link_set_preferred_training_settings (struct dc *dc, +struct dc_link_settings *link_setting, +struct dc_link_training_overrides *lt_overrides, +struct dc_link *link, +bool skip_immediate_retrain) +{ +if (lt_overrides !=NULL ) + link->preferred_training_settings = *lt_overrides; +else +memset (&link->preferred_training_settings,0 ,sizeof (link->preferred_training_settings)); + +if (link_setting !=NULL ) { + link->preferred_link_setting = *link_setting; + }else { + link->preferred_link_setting.lane_count =LANE_COUNT_UNKNOWN ; + link->preferred_link_setting.link_rate =LINK_RATE_UNKNOWN ; + } + +/* Retrain now, or wait until next stream update to apply */ +if (skip_immediate_retrain == false) +dc_link_set_preferred_link_settings (dc, &link->preferred_link_setting, link); +} +void dc_link_enable_hpd (const struct dc_link *link) {dc_link_dp_enable_hpd (link);
@@ -3022,7 +3061,6 @@ void dc_link_disable_hpd(const struct dc_link *link)dc_link_dp_disable_hpd (link); } - void dc_link_set_test_pattern(struct dc_link *link, enum dp_test_pattern test_pattern, const struct link_training_settings *p_link_settings,
drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c +238 -99
557 lines not on this page · raw message
drivers/gpu/drm/amd/display/dc/core/dc_link_hwss.c +15 -13
@@ -25,10 +25,11 @@ enum dc_status core_link_read_dpcd( uint8_t *data, uint32_t size) { -if (!dm_helpers_dp_read_dpcd (link->ctx, - link, - address, data, size)) -return DC_ERROR_UNEXPECTED ; +if (!link->aux_access_disabled && + !dm_helpers_dp_read_dpcd (link->ctx, + link, address, data, size)) { +return DC_ERROR_UNEXPECTED ; + }return DC_OK ; }
@@ -39,10 +40,11 @@ enum dc_status core_link_write_dpcd(const uint8_t *data, uint32_t size) { - if (!dm_helpers_dp_write_dpcd(link->ctx, - link, - address, data, size)) - return DC_ERROR_UNEXPECTED; + if (!link->aux_access_disabled && + !dm_helpers_dp_write_dpcd(link->ctx, + link, address, data, size)) { + return DC_ERROR_UNEXPECTED; + }return DC_OK ; }
@@ -203,21 +205,21 @@ void dp_disable_link_phy_mst(struct dc_link *link, enum signal_type signal)bool dp_set_hw_training_pattern (struct dc_link *link, -enum hw_dp_training_pattern pattern) +enum dc_dp_training_pattern pattern) {enum dp_test_pattern test_pattern =DP_TEST_PATTERN_UNSUPPORTED ;switch (pattern) { -case HW_DP_TRAINING_PATTERN_1 : +case DP_TRAINING_PATTERN_SEQUENCE_1 : test_pattern =DP_TEST_PATTERN_TRAINING_PATTERN1 ;break ; -case HW_DP_TRAINING_PATTERN_2 : +case DP_TRAINING_PATTERN_SEQUENCE_2 : test_pattern =DP_TEST_PATTERN_TRAINING_PATTERN2 ;break ; -case HW_DP_TRAINING_PATTERN_3 : +case DP_TRAINING_PATTERN_SEQUENCE_3 : test_pattern =DP_TEST_PATTERN_TRAINING_PATTERN3 ;break ; -case HW_DP_TRAINING_PATTERN_4 : +case DP_TRAINING_PATTERN_SEQUENCE_4 : test_pattern =DP_TEST_PATTERN_TRAINING_PATTERN4 ;break ;default :
drivers/gpu/drm/amd/display/dc/dc_dp_types.h +21 -0
@@ -90,6 +90,13 @@ enum dc_post_cursor2 { POST_CURSOR2_MAX_LEVEL = POST_CURSOR2_LEVEL3, }; +enum dc_dp_training_pattern { +DP_TRAINING_PATTERN_SEQUENCE_1 =0 , +DP_TRAINING_PATTERN_SEQUENCE_2 , +DP_TRAINING_PATTERN_SEQUENCE_3 , +DP_TRAINING_PATTERN_SEQUENCE_4 , +}; +struct dc_link_settings {enum dc_lane_count lane_count;enum dc_link_rate link_rate;
@@ -109,6 +116,20 @@ struct dc_link_training_settings {struct dc_lane_settings lane_settings[LANE_COUNT_DP_MAX ]; }; +struct dc_link_training_overrides { +enum dc_voltage_swing *voltage_swing; +enum dc_pre_emphasis *pre_emphasis; +enum dc_post_cursor2 *post_cursor2; + +uint16_t *cr_pattern_time; +uint16_t *eq_pattern_time; +enum dc_dp_training_pattern *pattern_for_eq; + +enum dc_link_spread *downspread; +bool *alternate_scrambler_reset; +bool *enhanced_framing; +bool *fec_enable; +}; union dpcd_rev { struct {
drivers/gpu/drm/amd/display/dc/dc_link.h +11 -0
@@ -83,6 +83,7 @@ struct dc_link {bool is_hpd_filter_disabled;bool dp_ss_off;bool link_state_valid; +bool aux_access_disabled;/* caps is the same as reported_link_cap. link_traing use * reported_link_cap. Will clean up. TODO
@@ -92,6 +93,7 @@ struct dc_link {struct dc_link_settings cur_link_settings;struct dc_lane_settings cur_lane_setting;struct dc_link_settings preferred_link_setting; +struct dc_link_training_overrides preferred_training_settings;uint8_t ddc_hw_inst;
@@ -217,6 +219,10 @@ void dc_link_dp_set_drive_settings(struct dc_link *link, struct link_training_settings *lt_settings); +bool dc_link_dp_perform_link_training_skip_aux ( +struct dc_link *link, +const struct dc_link_settings *link_setting); + enum link_training_result dc_link_dp_perform_link_training( struct dc_link *link, const struct dc_link_settings *link_setting,
@@ -251,6 +257,11 @@ void dc_link_perform_link_training(struct dc *dc,void dc_link_set_preferred_link_settings (struct dc *dc,struct dc_link_settings *link_setting,struct dc_link *link); +void dc_link_set_preferred_training_settings (struct dc *dc, +struct dc_link_settings *link_setting, +struct dc_link_training_overrides *lt_overrides, +struct dc_link *link, +bool skip_immediate_retrain);void dc_link_enable_hpd (const struct dc_link *link);void dc_link_disable_hpd (const struct dc_link *link); void dc_link_set_test_pattern(struct dc_link *link,
drivers/gpu/drm/amd/display/dc/inc/link_hwss.h +1 -1
@@ -60,7 +60,7 @@ void dp_disable_link_phy_mst(struct dc_link *link, enum signal_type signal);bool dp_set_hw_training_pattern (struct dc_link *link, -enum hw_dp_training_pattern pattern); +enum dc_dp_training_pattern pattern); void dp_set_hw_lane_settings( struct dc_link *link,
drivers/gpu/drm/amd/display/include/link_service_types.h +10 -7
@@ -71,14 +71,17 @@ enum link_training_result {struct link_training_settings {struct dc_link_settings link_settings;struct dc_lane_settings lane_settings[LANE_COUNT_DP_MAX ]; -bool allow_invalid_msa_timing_param; -}; -enum hw_dp_training_pattern { -HW_DP_TRAINING_PATTERN_1 =0 , -HW_DP_TRAINING_PATTERN_2 , -HW_DP_TRAINING_PATTERN_3 , -HW_DP_TRAINING_PATTERN_4 +enum dc_voltage_swing *voltage_swing; +enum dc_pre_emphasis *pre_emphasis; +enum dc_post_cursor2 *post_cursor2; + +uint16_t cr_pattern_time; +uint16_t eq_pattern_time; +enum dc_dp_training_pattern pattern_for_eq; + +bool enhanced_framing; +bool allow_invalid_msa_timing_param; };/*TODO: Move this enum test harness*/
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 3/87 drm/amd/display: 3.2.36
In reply to a message not in the archiveFrom: Aric Cyr <[email protected]>
--- drivers/gpu/drm/amd/display/dc/dc.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/dc.h +1 -1
@@ -39,7 +39,7 @@#include "inc/hw/dmcu.h" #include "dml/display_mode_lib.h" -#define DC_VER "3.2.35" +#define DC_VER "3.2.36"#define MAX_SURFACES 3#define MAX_PLANES 6
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 4/87 drm/amd/display: fix up HUBBUB hw programming for VM
In reply to a message not in the archiveFrom: Jun Lei <[email protected]>
[why]
Some values were not being converted or bit-shifted properly for HW registers, causing black screen
[how]
Fix up the values before programming HW
--- .../gpu/drm/amd/display/dc/dcn20/dcn20_hubbub.c | 17 ++++++++--------- .../gpu/drm/amd/display/dc/inc/hw/dchubbub.h | 4 ++-- 2 files changed, 10 insertions(+), 11 deletions(-)
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hubbub.c +8 -9
@@ -366,25 +366,24 @@ int hubbub2_init_dchub_sys_ctx(struct hubbub *hubbub,struct dcn_vmid_page_table_config phys_config;REG_SET (DCN_VM_FB_LOCATION_BASE ,0 , -FB_BASE , pa_config->system_aperture.fb_base); +FB_BASE , pa_config->system_aperture.fb_base >>24 );REG_SET (DCN_VM_FB_LOCATION_TOP ,0 , -FB_TOP , pa_config->system_aperture.fb_top); +FB_TOP , pa_config->system_aperture.fb_top >>24 );REG_SET (DCN_VM_FB_OFFSET ,0 , -FB_OFFSET , pa_config->system_aperture.fb_offset); +FB_OFFSET , pa_config->system_aperture.fb_offset >>24 );REG_SET (DCN_VM_AGP_BOT ,0 , -AGP_BOT , pa_config->system_aperture.agp_bot); +AGP_BOT , pa_config->system_aperture.agp_bot >>24 );REG_SET (DCN_VM_AGP_TOP ,0 , -AGP_TOP , pa_config->system_aperture.agp_top); +AGP_TOP , pa_config->system_aperture.agp_top >>24 );REG_SET (DCN_VM_AGP_BASE ,0 , -AGP_BASE , pa_config->system_aperture.agp_base); +AGP_BASE , pa_config->system_aperture.agp_base >>24 );if (pa_config->gart_config.page_table_start_addr != pa_config->gart_config.page_table_end_addr) { - phys_config.depth =1 ; - phys_config.block_size =4096 ; phys_config.page_table_start_addr = pa_config->gart_config.page_table_start_addr >>12 ; phys_config.page_table_end_addr = pa_config->gart_config.page_table_end_addr >>12 ; phys_config.page_table_base_addr = pa_config->gart_config.page_table_base_addr; - + phys_config.depth =0 ; + phys_config.block_size =0 ;// Init VMID 0 based on PA config dcn20_vmid_setup (&hubbub1->vmid[0 ], &phys_config); }
drivers/gpu/drm/amd/display/dc/inc/hw/dchubbub.h +2 -2
@@ -61,8 +61,8 @@ enum dcn_hubbub_page_table_depth { };enum dcn_hubbub_page_table_block_size { -DCN_PAGE_TABLE_BLOCK_SIZE_4KB , -DCN_PAGE_TABLE_BLOCK_SIZE_64KB +DCN_PAGE_TABLE_BLOCK_SIZE_4KB =0 , +DCN_PAGE_TABLE_BLOCK_SIZE_64KB =4 };struct dcn_hubbub_phys_addr_config {
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 5/87 drm/amd/display: fix dsc disable
In reply to a message not in the archiveFrom: Dmytro Laktyushkin <[email protected]>
A regression caused dsc to never get disabled in certain situations.
--- drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c +4 -1
@@ -1715,8 +1715,11 @@ static void dcn20_reset_back_end_for_pipe(else if (pipe_ctx->stream_res.audio) { dc->hwss.disable_audio_stream (pipe_ctx,FREE_ACQUIRED_RESOURCE ); } - } +#ifdef CONFIG_DRM_AMD_DC_DSC_SUPPORT + else if (pipe_ctx->stream_res.dsc) + dp_set_dsc_enable(pipe_ctx, false); +#endif /* by upper caller loop, parent pipe: pipe0, will be reset last. * back end share by all pipes and will be disable only when disable
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 6/87 drm/amd/display: 3.2.37
In reply to a message not in the archiveFrom: Aric Cyr <[email protected]>
--- drivers/gpu/drm/amd/display/dc/dc.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/dc.h +1 -1
@@ -39,7 +39,7 @@#include "inc/hw/dmcu.h" #include "dml/display_mode_lib.h" -#define DC_VER "3.2.36" +#define DC_VER "3.2.37"#define MAX_SURFACES 3#define MAX_PLANES 6
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 7/87 drm/amd/display: move bw calc code into helpers
In reply to a message not in the archiveFrom: Eric Yang <[email protected]>
[Why]
For better readability and reusability
[How]
Move snippets of BW calculation code into helpers.
--- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 1 + .../display/dc/clk_mgr/dcn20/dcn20_clk_mgr.c | 2 - .../drm/amd/display/dc/dcn20/dcn20_resource.c | 250 +++++++++++------- .../drm/amd/display/dc/dcn20/dcn20_resource.h | 11 + .../amd/display/dc/inc/hw/clk_mgr_internal.h | 2 + 5 files changed, 167 insertions(+), 99 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +1 -0
@@ -126,6 +126,7 @@ static DEVICE_ATTR(pcie_replay_count, S_IRUGO, amdgpu_device_get_pcie_replay_count, NULL);static void amdgpu_device_get_pcie_info (struct amdgpu_device *adev); +static void amdgpu_device_parse_faked_did (struct amdgpu_device *adev);/** * amdgpu_device_is_px - Is the device is a dGPU with HG/PX power control
drivers/gpu/drm/amd/display/dc/clk_mgr/dcn20/dcn20_clk_mgr.c +0 -2
@@ -26,8 +26,6 @@#include "dccg.h" #include "clk_mgr_internal.h" - -#include "dcn20/dcn20_clk_mgr.h" #include "dce100/dce_clk_mgr.h" #include "reg_helper.h" #include "core_types.h"
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c +153 -97
301 lines not on this page · raw message
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.h +11 -0
@@ -116,6 +116,17 @@ void dcn20_set_mcif_arb_params(display_e2e_pipe_params_st *pipes, int pipe_cnt);bool dcn20_validate_bandwidth (struct dc *dc,struct dc_state *context,bool fast_validate); +bool dcn20_fast_validate_bw ( +struct dc *dc, +struct dc_state *context, +display_e2e_pipe_params_st *pipes, +int *pipe_split_from, +int *vlevel_out); +void dcn20_calculate_dlg_params ( +struct dc *dc,struct dc_state *context, +display_e2e_pipe_params_st *pipes, +int pipe_cnt, +int vlevel);enum dc_status dcn20_build_mapped_resource (const struct dc *dc,struct dc_state *context,struct dc_stream_state *stream);enum dc_status dcn20_add_stream_to_ctx (struct dc *dc,struct dc_state *new_ctx,struct dc_stream_state *dc_stream);
drivers/gpu/drm/amd/display/dc/inc/hw/clk_mgr_internal.h +2 -0
@@ -64,6 +64,8 @@ enum dentist_divider_range {*************************************************************************************** */ +/* Macros */ +#define TO_CLK_MGR_INTERNAL (clk_mgr)\ container_of(clk_mgr, struct clk_mgr_internal, base)
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
A message that is not in the archive (sent off-list or not yet received): [email protected]
Re: [PATCH 07/87] drm/amd/display: move bw calc code into helpers
Quoted 19 lines · On 2019-07-15 11:19 p.m., [email protected] wrote:
From: Eric Yang <[email protected]>
[Why]
For better readability and reusability[How]
Move snippets of BW calculation code into helpers.Signed-off-by Eric Yang <[email protected]>Reviewed-by Fatemeh Darbehani <[email protected]>Acked-by Leo Li <[email protected]>---
[...]drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +0
-0
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c static DEVICE_ATTR(pcie_replay_count, S_IRUGO, · line 129
Quoted 4 lines
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +1
-0@@ -126,6 +126,7 @@ static DEVICE_ATTR(pcie_replay_count, S_IRUGO, amdgpu_device_get_pcie_replay_count, NULL);static void amdgpu_device_get_pcie_info (struct amdgpu_device *adev); +static void amdgpu_device_parse_faked_did (struct amdgpu_device *adev);
This function isn't defined anywhere, producing a warning:
drivers/gpu/drm//amd/amdgpu/amdgpu_device.c:131:13: warning: ‘amdgpu_device_parse_faked_did’ declared ‘static’ but never defined [-Wunused-function]
static void amdgpu_device_parse_faked_did(struct amdgpu_device *adev);
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~Signature · 6 lines
Earthling Michel Dänzer | https://www.amd.com Libre software enthusiast | Mesa and X developer _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
A message that is not in the archive (sent off-list or not yet received): 66c6c4d7-34b3-b73b-4e21-f745fff470cd-otUistvHUpPR7s880joybQ@public.gmane.org
Re: [PATCH 07/87] drm/amd/display: move bw calc code into helpers
Quoted 22 lines · Michel Dänzer
Quoted 21 lines · On 2019-07-15 11:19 p.m., [email protected] wrote:
From: Eric Yang <[email protected]>
[Why]
For better readability and reusability[How]
Move snippets of BW calculation code into helpers.Signed-off-by Eric Yang <[email protected]>Reviewed-by Fatemeh Darbehani <[email protected]>Acked-by Leo Li <[email protected]>---
[...]drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +0
-0@@ -126,6 +126,7 @@ static DEVICE_ATTR(pcie_replay_count, S_IRUGO, amdgpu_device_get_pcie_replay_count, NULL);
Quoted 8 lines
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c static DEVICE_ATTR(pcie_replay_count, S_IRUGO, · line 129
Quoted 2 lines
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +1
-0static void amdgpu_device_get_pcie_info (struct amdgpu_device *adev); +static void amdgpu_device_parse_faked_did (struct amdgpu_device *adev);This function isn't defined anywhere, producing a warning:
drivers/gpu/drm//amd/amdgpu/amdgpu_device.c:131:13: warning: ‘amdgpu_device_parse_faked_did’ declared ‘static’ but never defined [-Wunused-function]
static void amdgpu_device_parse_faked_did(struct amdgpu_device *adev); ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Oops, I don't know how that managed to get in there. It's not even in the display dir...
Sent a fix.
Leo
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 8/87 drm/amd/display: cap DCFCLK hardmin to 507 for NV10
In reply to a message not in the archiveFrom: Jun Lei <[email protected]>
[why]
Due to limitation in SMU/PPLIB, it is not possible to know Fmax @ Vmin for DCFCLK. This causes issues at high display configurations where extra headroom of DCFCLK can enable P-state switching
[how]
Use existing override logic. If override not defined, then force min = 507
--- drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c | 4 ++++ 1 file changed, 4 insertions(+)
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c +4 -0
@@ -2704,6 +2704,10 @@ static void update_bounding_box(struct dc *dc, struct _vcs_dpi_soc_bounding_box_if (dc->bb_overrides.min_dcfclk_mhz >0 ) min_dcfclk = dc->bb_overrides.min_dcfclk_mhz; +else +// Accounting for SOC/DCF relationship, we can go as high as +// 506Mhz in Vmin. We need to code 507 since SMU will round down to 506. + min_dcfclk =507 ;for (i =0 ; i < num_states; i++) {int min_fclk_required_by_uclk;
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 9/87 drm/amd/display: No audio endpoint for Dell MST display
In reply to a message not in the archiveFrom: Harmanprit Tatla <[email protected]>
[Why]
There are certain MST displays (i.e. Dell P2715Q) that although have the MST feature set to off may still report it is a branch device and a non-zero value for downstream port present.
This can lead to us incorrectly classifying a dp dongle connection as being active and disabling the audio endpoint for the display.
[How]
Modified the placement and
condition used to assign
the is_branch_dev bit.
--- drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c +8 -1
@@ -2369,11 +2369,18 @@ static void get_active_converter_info( link->dpcd_caps.dongle_type =DISPLAY_DONGLE_NONE ;ddc_service_set_dongle_type (link->ddc, link->dpcd_caps.dongle_type); + link->dpcd_caps.is_branch_dev = false;return ; }/* DPCD 0x5 bit 0 = 1, it indicate it's branch device */ - link->dpcd_caps.is_branch_dev = ds_port.fields.PORT_PRESENT; +if (ds_port.fields.PORT_TYPE ==DOWNSTREAM_DP ) { + link->dpcd_caps.is_branch_dev = false; + } + +else { + link->dpcd_caps.is_branch_dev = ds_port.fields.PORT_PRESENT; + }switch (ds_port.fields.PORT_TYPE) {case DOWNSTREAM_VGA :
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 10/87 drm/amd/display: Add aux tracing log in dce
In reply to a message not in the archiveFrom: Chiawen Huang <[email protected]>
[Why]
dce was re-arch'd, therefore adding aux tracing log into new dce
[How]
The porting from submit_channel_request/process_channel_reply of aux_engine_dce110.c
--- drivers/gpu/drm/amd/display/dc/dce/dce_aux.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/dce/dce_aux.c +8 -1
@@ -27,6 +27,7 @@#include "core_types.h" #include "dce_aux.h" #include "dce/dce_11_0_sh_mask.h" +#include "dm_event_log.h" #define CTX \ aux110->base.ctx
@@ -249,6 +250,8 @@ static void submit_channel_request( }REG_UPDATE (AUX_SW_CONTROL ,AUX_SW_GO ,1 ); +EVENT_LOG_AUX_REQ (engine->ddc->pin_data->en,EVENT_LOG_AUX_ORIGIN_NATIVE , + request->action, request->address, request->length, request->data); } static int read_channel_reply(struct dce_aux *engine, uint32_t size,
@@ -477,9 +480,13 @@ int dce_aux_transfer_raw(struct ddc_service *ddc, *operation_result =get_channel_status (aux_engine, &returned_bytes);if (*operation_result ==AUX_CHANNEL_OPERATION_SUCCEEDED ) { -read_channel_reply (aux_engine, payload->length, +int bytes_replied =0 ; + bytes_replied =read_channel_reply (aux_engine, payload->length, payload->data, payload->reply, &status); +EVENT_LOG_AUX_REP (aux_engine->ddc->pin_data->en, +EVENT_LOG_AUX_ORIGIN_NATIVE , *payload->reply, + bytes_replied, payload->data); res = returned_bytes; }else { res =-1 ;
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 11/87 drm/amd/display: Update drm_dsc to reflect native 4.2.0 DSC spec
In reply to a message not in the archiveFrom: David Francis <[email protected]>
[Why]
Some parts of the DSC spec relating to 4.2.0 were not reflected in drm_dsc_compute_rc_parameters, causing unexpected config failures
[How]
Add nsl_bpg_offset and rbs_min computation
--- .../gpu/drm/amd/display/dc/dsc/drm_dsc_dc.c | 73 ++++++++++++++++++- 1 file changed, 69 insertions(+), 4 deletions(-)
drivers/gpu/drm/amd/display/dc/dsc/drm_dsc_dc.c +69 -4
@@ -15,6 +15,7 @@#define ERANGE -1#define DRM_DEBUG_KMS (msg)/* nothing */ #define cpu_to_be16 (__x) little_to_big(__x) +#define MAX (x, y) ((x) > (y) ? (x) : (y))static unsigned short little_to_big (int data) {
@@ -232,6 +233,38 @@ void drm_dsc_pps_payload_pack(struct drm_dsc_picture_parameter_set *pps_payload, }EXPORT_SYMBOL (drm_dsc_pps_payload_pack); +static int compute_offset (struct drm_dsc_config *vdsc_cfg,int pixels_per_group, +int groups_per_line,int grpcnt) +{ +int offset =0 ; +int grpcnt_id =DIV_ROUND_UP (vdsc_cfg->initial_xmit_delay, pixels_per_group); + +if (grpcnt <= grpcnt_id) + offset =DIV_ROUND_UP (grpcnt * pixels_per_group * vdsc_cfg->bits_per_pixel,16 ); +else + offset =DIV_ROUND_UP (grpcnt_id * pixels_per_group * vdsc_cfg->bits_per_pixel,16 ) + - (((grpcnt - grpcnt_id) * vdsc_cfg->slice_bpg_offset) >>11 ); + +if (grpcnt <= groups_per_line) + offset += grpcnt * vdsc_cfg->first_line_bpg_offset; +else + offset += groups_per_line * vdsc_cfg->first_line_bpg_offset + - (((grpcnt - groups_per_line) * vdsc_cfg->nfl_bpg_offset) >>11 ); + +if (vdsc_cfg->native_420) { +if (grpcnt <= groups_per_line) + offset -= (grpcnt * vdsc_cfg->nsl_bpg_offset) >>11 ; +else if (grpcnt <=2 * groups_per_line) + offset += (grpcnt - groups_per_line) * vdsc_cfg->second_line_bpg_offset + - ((groups_per_line * vdsc_cfg->nsl_bpg_offset) >>11 ); +else + offset += (grpcnt - groups_per_line) * vdsc_cfg->second_line_bpg_offset + - (((grpcnt - groups_per_line) * vdsc_cfg->nsl_bpg_offset) >>11 ); + } + +return offset; +} +/** * drm_dsc_compute_rc_parameters() - Write rate control * parameters to the dsc configuration defined in
@@ -251,6 +284,7 @@ int drm_dsc_compute_rc_parameters(struct drm_dsc_config *vdsc_cfg)unsigned long hrd_delay =0 ;unsigned long final_scale =0 ;unsigned long rbs_min =0 ; +unsigned long max_offset =0 ;if (vdsc_cfg->native_420 || vdsc_cfg->native_422) {/* Number of groups used to code each line of a slice */
@@ -329,6 +363,17 @@ int drm_dsc_compute_rc_parameters(struct drm_dsc_config *vdsc_cfg)return -ERANGE ; } +if (vdsc_cfg->slice_height >2 ) + vdsc_cfg->nsl_bpg_offset =DIV_ROUND_UP ((vdsc_cfg->second_line_bpg_offset <<11 ), + (vdsc_cfg->slice_height -1 )); +else + vdsc_cfg->nsl_bpg_offset =0 ; + +if (vdsc_cfg->nsl_bpg_offset >65535 ) { +DRM_DEBUG_KMS ("NslBpgOffset is too large for this slice height\n" ); +return -ERANGE ; + } +/* Number of groups used to code the entire slice */ groups_total = groups_per_line * vdsc_cfg->slice_height;
@@ -348,6 +393,7 @@ int drm_dsc_compute_rc_parameters(struct drm_dsc_config *vdsc_cfg) vdsc_cfg->scale_increment_interval = (vdsc_cfg->final_offset * (1 <<11 )) / ((vdsc_cfg->nfl_bpg_offset + + vdsc_cfg->nsl_bpg_offset + vdsc_cfg->slice_bpg_offset) * (final_scale -9 )); }else {
@@ -368,10 +414,29 @@ int drm_dsc_compute_rc_parameters(struct drm_dsc_config *vdsc_cfg)* bits/pixel (bpp) rate that is used by the encoder, * in steps of 1/16 of a bit per pixel */ - rbs_min = vdsc_cfg->rc_model_size - vdsc_cfg->initial_offset + -DIV_ROUND_UP (vdsc_cfg->initial_xmit_delay * - vdsc_cfg->bits_per_pixel,16 ) + - groups_per_line * vdsc_cfg->first_line_bpg_offset; +if (vdsc_cfg->dsc_version_minor ==2 && (vdsc_cfg->native_420 || vdsc_cfg->native_422)) { + + max_offset =compute_offset (vdsc_cfg,DSC_RC_PIXELS_PER_GROUP , groups_per_line, +DIV_ROUND_UP (vdsc_cfg->initial_xmit_delay, +DSC_RC_PIXELS_PER_GROUP )); + + max_offset =MAX (max_offset, +compute_offset (vdsc_cfg,DSC_RC_PIXELS_PER_GROUP , groups_per_line, +DIV_ROUND_UP (vdsc_cfg->initial_xmit_delay, + groups_per_line))); + + max_offset =MAX (max_offset, +compute_offset (vdsc_cfg,DSC_RC_PIXELS_PER_GROUP , groups_per_line, +DIV_ROUND_UP (vdsc_cfg->initial_xmit_delay, + groups_per_line *2 ))); + + rbs_min = vdsc_cfg->rc_model_size - vdsc_cfg->initial_offset + max_offset; + }else { + rbs_min = vdsc_cfg->rc_model_size - vdsc_cfg->initial_offset + +DIV_ROUND_UP (vdsc_cfg->initial_xmit_delay * + vdsc_cfg->bits_per_pixel,16 ) + + groups_per_line * vdsc_cfg->first_line_bpg_offset; + } hrd_delay =DIV_ROUND_UP ((rbs_min *16 ), vdsc_cfg->bits_per_pixel); vdsc_cfg->rc_bits = (hrd_delay * vdsc_cfg->bits_per_pixel) /16 ;
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 12/87 drm/amd/display: early return when pipe_cnt is 0 in bw validation
In reply to a message not in the archiveFrom: Eric Yang <[email protected]>
[Why]
Unintentionally introduced behaviour change from previous refactor, which causes clks to be 0 in no stream cases, which will cause divide by 0.
[How]
Skip calculation of clocks when no stream. Which is the same as old behaviour.
--- drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c | 8 +++++++- drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.h | 1 + 2 files changed, 8 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c +7 -1
@@ -2020,6 +2020,7 @@ bool dcn20_fast_validate_bw(struct dc *dc, struct dc_state *context, display_e2e_pipe_params_st *pipes, + int *pipe_cnt_out, int *pipe_split_from, int *vlevel_out) {
@@ -2072,6 +2073,8 @@ bool dcn20_fast_validate_bw( pipe_cnt =dcn20_populate_dml_pipes_from_context (dc, &context->res_ctx, pipes); + *pipe_cnt_out = pipe_cnt; +if (!pipe_cnt) { out = true; goto validate_out;
@@ -2432,7 +2435,10 @@ bool dcn20_validate_bandwidth(struct dc *dc, struct dc_state *context,BW_VAL_TRACE_COUNT (); - out =dcn20_fast_validate_bw (dc, context, pipes, pipe_split_from, &vlevel); + out =dcn20_fast_validate_bw (dc, context, pipes, &pipe_cnt, pipe_split_from, &vlevel); + +if (pipe_cnt ==0 ) + goto validate_out;if (!out) goto validate_fail;
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.h +1 -0
@@ -120,6 +120,7 @@ bool dcn20_fast_validate_bw(struct dc *dc, struct dc_state *context, display_e2e_pipe_params_st *pipes, + int *pipe_cnt_out, int *pipe_split_from, int *vlevel_out); void dcn20_calculate_dlg_params(
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 13/87 drm/amd/display: Set default block_size, even in unexpected cases
In reply to a message not in the archiveFrom: Dmytro Laktyushkin <[email protected]>
We're not expected to enter the default case, but not returning a default value here is incorrect.
--- drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hubbub.c | 1 + drivers/gpu/drm/amd/display/dc/inc/hw/dchubbub.h | 2 +- 2 files changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hubbub.c +1 -0
@@ -337,6 +337,7 @@ static enum dcn_hubbub_page_table_block_size page_table_block_size_to_hw(unsignebreak ;default :ASSERT (false); + block_size = page_table_block_size;break ; }
drivers/gpu/drm/amd/display/dc/inc/hw/dchubbub.h +1 -1
@@ -62,7 +62,7 @@ enum dcn_hubbub_page_table_depth {enum dcn_hubbub_page_table_block_size {DCN_PAGE_TABLE_BLOCK_SIZE_4KB =0 , -DCN_PAGE_TABLE_BLOCK_SIZE_64KB =4 +DCN_PAGE_TABLE_BLOCK_SIZE_64KB =4 , };struct dcn_hubbub_phys_addr_config {
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 14/87 drm/amd/display: Set one 4:2:0-related PPS field as recommended by DSC spec
In reply to a message not in the archiveFrom: Nikola Cornij <[email protected]>
[why]
'second_line_offset_adj' was mistakenly left at zero, even though DSC spec v1.2a recommends setting this field to 512 for 4:2:0.
[how]
Set 'second_line_offset_adj' to 512 for 4:2:0 and leave at zero otherwise
--- drivers/gpu/drm/amd/display/dc/dsc/drm_dsc_dc.c | 6 ++++++ 1 file changed, 6 insertions(+)
drivers/gpu/drm/amd/display/dc/dsc/drm_dsc_dc.c +6 -0
@@ -442,6 +442,12 @@ int drm_dsc_compute_rc_parameters(struct drm_dsc_config *vdsc_cfg) vdsc_cfg->rc_bits = (hrd_delay * vdsc_cfg->bits_per_pixel) /16 ; vdsc_cfg->initial_dec_delay = hrd_delay - vdsc_cfg->initial_xmit_delay; +/* As per DSC spec v1.2a recommendation: */ +if (vdsc_cfg->native_420) + vdsc_cfg->second_line_offset_adj =512 ; +else + vdsc_cfg->second_line_offset_adj =0 ; +return 0 ; }EXPORT_SYMBOL (drm_dsc_compute_rc_parameters);
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 15/87 drm/amd/display: swap system aperture high/low
In reply to a message not in the archiveFrom: Jun Lei <[email protected]>
[why]
Currently logical values are swapped in HW, causing system aperture to be undefined, so VA and PA cannot co-exist
[how]
program values correctly
--- drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c +2 -2
@@ -1128,8 +1128,8 @@ void dcn20_enable_plane( apt.sys_default.quad_part =0 ; - apt.sys_high.quad_part = dc->vm_pa_config.system_aperture.start_addr; - apt.sys_low.quad_part = dc->vm_pa_config.system_aperture.end_addr; + apt.sys_low.quad_part = dc->vm_pa_config.system_aperture.start_addr; + apt.sys_high.quad_part = dc->vm_pa_config.system_aperture.end_addr;// Program system aperture settings pipe_ctx->plane_res.hubp->funcs->hubp_set_vm_system_aperture_settings (pipe_ctx->plane_res.hubp, &apt);
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 16/87 drm/amd/display: skip retrain in dc_link_set_preferred_link_settings() if using passive dongle
In reply to a message not in the archiveFrom: Samson Tam <[email protected]>
[Why]
Fixes issue when we have a display connected using a passive dongle and then emulate over it using a DP connection at 1 x 1.62 Ghz. System hangs because register bus returns back 0xFFFFFFFF for all register reads after setting register DIG_BE_CNTL in dcn10_link_encoder_connect_dig_be_to_fe(). Hang occurs later when trying to do a register read.
[How]
At the start of the emulation, dc_link_set_preferred_link_settings() and dp_retrain_link_dp_test() is called, even though it is connected using a passive dongle.
Add an extra condition in dp_retrain_link_dp_test() to check for link->dongle_max_pix_clk > 0. This is the only way we know if the connection is using passive dongle so we don't retrain DP.
--- drivers/gpu/drm/amd/display/dc/core/dc_link.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/core/dc_link.c +3 -1
@@ -2998,8 +2998,10 @@ void dc_link_set_preferred_link_settings(struct dc *dc,/* Retrain with preferred link settings only relevant for * DP signal type +* Check for non-DP signal or if passive dongle present */ -if (!dc_is_dp_signal (link->connector_signal)) +if (!dc_is_dp_signal (link->connector_signal) || + link->dongle_max_pix_clk >0 )return ;for (i =0 ; i <MAX_PIPES ; i++) {
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 17/87 drm/amd/display: Split out common HUBP registers and code
In reply to a message not in the archiveFrom: Charlene Liu <[email protected]>
There are shared regs and code across DCN generations. Pull them out into a shared common location.
Also, expose some dcn20 init functions.
--- .../gpu/drm/amd/display/dc/dce/dce_hwseq.h | 3 +- .../gpu/drm/amd/display/dc/dcn10/dcn10_hubp.c | 50 ++++++++++++------- .../gpu/drm/amd/display/dc/dcn10/dcn10_hubp.h | 23 ++++++--- .../gpu/drm/amd/display/dc/dcn20/dcn20_hubp.h | 23 +++++---- .../drm/amd/display/dc/dcn20/dcn20_hwseq.c | 43 +++++++++++++++- .../drm/amd/display/dc/dcn20/dcn20_hwseq.h | 14 ++---- 6 files changed, 109 insertions(+), 47 deletions(-)
drivers/gpu/drm/amd/display/dc/dce/dce_hwseq.h +2 -1
@@ -696,7 +696,8 @@ struct dce_hwseq_registers {type D2VGA_MODE_ENABLE ; \type D3VGA_MODE_ENABLE ; \type D4VGA_MODE_ENABLE ; \ -type AZALIA_AUDIO_DTO_MODULE ; +type AZALIA_AUDIO_DTO_MODULE ;\ +type HPO_HDMISTREAMCLK_GATE_DIS ;struct dce_hwseq_shift {HWSEQ_REG_FIELD_LIST (uint8_t )
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.c +31 -19
70 lines not on this page · raw message
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.h +16 -7
@@ -249,7 +249,8 @@ .field_name = reg_name ## __ ## field_name ## post_fix/* Mask/shift struct generation macro for all ASICs (including those with reduced functionality) */ -#define HUBP_MASK_SH_LIST_DCN_COMMON (mask_sh)\ +/*1.x, 2.x, and 3.x*/ +#define HUBP_MASK_SH_LIST_DCN_SHARE_COMMON (mask_sh)\ HUBP_SF(HUBP0_DCHUBP_CNTL, HUBP_BLANK_EN, mask_sh),\ HUBP_SF(HUBP0_DCHUBP_CNTL, HUBP_TTU_DISABLE, mask_sh),\ HUBP_SF(HUBP0_DCHUBP_CNTL, HUBP_UNDERFLOW_STATUS, mask_sh),\
@@ -265,7 +266,6 @@HUBP_SF (HUBP0_DCSURF_ADDR_CONFIG ,MAX_COMPRESSED_FRAGS , mask_sh),\HUBP_SF (HUBP0_DCSURF_TILING_CONFIG ,SW_MODE , mask_sh),\HUBP_SF (HUBP0_DCSURF_TILING_CONFIG ,META_LINEAR , mask_sh),\ -HUBP_SF (HUBP0_DCSURF_TILING_CONFIG ,RB_ALIGNED , mask_sh),\HUBP_SF (HUBP0_DCSURF_TILING_CONFIG ,PIPE_ALIGNED , mask_sh),\HUBP_SF (HUBPREQ0_DCSURF_SURFACE_PITCH ,PITCH , mask_sh),\ HUBP_SF(HUBPREQ0_DCSURF_SURFACE_PITCH, META_PITCH, mask_sh),\
@@ -372,12 +372,17 @@HUBP_SF (HUBPREQ0_DCN_SURF0_TTU_CNTL0 , QoS_RAMP_DISABLE, mask_sh),\HUBP_SF (HUBPREQ0_DCN_SURF0_TTU_CNTL1 ,REFCYC_PER_REQ_DELIVERY_PRE , mask_sh),\HUBP_SF (HUBP0_HUBP_CLK_CNTL ,HUBP_CLOCK_ENABLE , mask_sh) - -#define HUBP_MASK_SH_LIST_DCN (mask_sh)\ - HUBP_MASK_SH_LIST_DCN_COMMON(mask_sh),\ +/*2.x and 1.x only*/ +#define HUBP_MASK_SH_LIST_DCN_COMMON (mask_sh)\ + HUBP_MASK_SH_LIST_DCN_SHARE_COMMON(mask_sh),\ + HUBP_SF(HUBP0_DCSURF_TILING_CONFIG, RB_ALIGNED, mask_sh),\ HUBP_SF(HUBP0_DCHUBP_REQ_SIZE_CONFIG, MPTE_GROUP_SIZE, mask_sh),\ HUBP_SF(HUBP0_DCHUBP_REQ_SIZE_CONFIG_C, MPTE_GROUP_SIZE_C, mask_sh) +/*2.x and 1.x only*/ +#define HUBP_MASK_SH_LIST_DCN (mask_sh)\ + HUBP_MASK_SH_LIST_DCN_COMMON(mask_sh) +/* Mask/shift struct generation macro for ASICs with VM */ #define HUBP_MASK_SH_LIST_DCN_VM (mask_sh)\ HUBP_SF(HUBPREQ0_NOM_PARAMETERS_0, DST_Y_PER_PTE_ROW_NOM_L, mask_sh),\
@@ -434,7 +439,7 @@HUBP_SF (CURSOR0_CURSOR_HOT_SPOT ,CURSOR_HOT_SPOT_Y , mask_sh), \ HUBP_SF(CURSOR0_CURSOR_DST_OFFSET, CURSOR_DST_X_OFFSET, mask_sh) -#define DCN_HUBP_REG_FIELD_LIST (type) \ +#define DCN_HUBP_REG_FIELD_BASE_LIST (type) \ type HUBP_BLANK_EN;\ type HUBP_DISABLE;\ type HUBP_TTU_DISABLE;\
@@ -459,7 +464,6 @@type ROTATION_ANGLE ;\type H_MIRROR_EN ;\type SURFACE_PIXEL_FORMAT ;\ -type ALPHA_PLANE_EN ;\type SURFACE_FLIP_TYPE ;\type SURFACE_FLIP_MODE_FOR_STEREOSYNC ;\type SURFACE_FLIP_IN_STEREOSYNC ;\
@@ -632,6 +636,10 @@type CURSOR_DST_X_OFFSET ; \ type OUTPUT_FP +#define DCN_HUBP_REG_FIELD_LIST (type) \ + DCN_HUBP_REG_FIELD_BASE_LIST(type);\ + type ALPHA_PLANE_EN +struct dcn_mi_registers {HUBP_COMMON_REG_VARIABLE_LIST ; };
@@ -760,5 +768,6 @@ void hubp1_vready_workaround(struct hubp *hubp, struct _vcs_dpi_display_pipe_dest_params_st *pipe_dest);void hubp1_init (struct hubp *hubp); +void hubp1_read_state_common (struct hubp *hubp);#endif
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hubp.h +14 -9
@@ -72,8 +72,8 @@SR (DCN_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB ),\ SR(DCN_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB) -#define HUBP_MASK_SH_LIST_DCN2_COMMON (mask_sh)\ - HUBP_MASK_SH_LIST_DCN(mask_sh),\ +#define HUBP_MASK_SH_LIST_DCN2_SHARE_COMMON (mask_sh)\ + HUBP_MASK_SH_LIST_DCN_SHARE_COMMON(mask_sh),\ HUBP_MASK_SH_LIST_DCN_VM(mask_sh),\ HUBP_SF(HUBP0_DCSURF_SURFACE_CONFIG, ROTATION_ANGLE, mask_sh),\ HUBP_SF(HUBP0_DCSURF_SURFACE_CONFIG, H_MIRROR_EN, mask_sh),\
@@ -127,13 +127,21 @@HUBP_SF (HUBPREQ0_DCSURF_FLIP_CONTROL2 ,SURFACE_TRIPLE_BUFFER_ENABLE , mask_sh),\HUBP_SF (HUBPREQ0_VMID_SETTINGS_0 ,VMID , mask_sh) +/*DCN2.x and DCN1.x*/ +#define HUBP_MASK_SH_LIST_DCN2_COMMON (mask_sh)\ + HUBP_MASK_SH_LIST_DCN2_SHARE_COMMON(mask_sh),\ + HUBP_SF(HUBP0_DCSURF_TILING_CONFIG, RB_ALIGNED, mask_sh),\ + HUBP_SF(HUBP0_DCHUBP_REQ_SIZE_CONFIG, MPTE_GROUP_SIZE, mask_sh),\ + HUBP_SF(HUBP0_DCHUBP_REQ_SIZE_CONFIG_C, MPTE_GROUP_SIZE_C, mask_sh) + +/*DCN2.0 specific*/ #define HUBP_MASK_SH_LIST_DCN20 (mask_sh)\ HUBP_MASK_SH_LIST_DCN2_COMMON(mask_sh),\ HUBP_SF(DCN_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB, DCN_VM_SYSTEM_APERTURE_DEFAULT_SYSTEM, mask_sh),\ HUBP_SF(DCN_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB, DCN_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB, mask_sh),\ HUBP_SF(DCN_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB, DCN_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB, mask_sh) - +/*DCN2.x */ #define DCN2_HUBP_REG_COMMON_VARIABLE_LIST \ HUBP_COMMON_REG_VARIABLE_LIST; \ uint32_t DMDATA_ADDRESS_HIGH; \
@@ -149,14 +157,11 @@uint32_t FLIP_PARAMETERS_2 ;\uint32_t DCN_CUR1_TTU_CNTL0 ;\uint32_t DCN_CUR1_TTU_CNTL1 ;\ -uint32_t VMID_SETTINGS_0 ;\ -uint32_t FLIP_PARAMETERS_3 ;\ -uint32_t FLIP_PARAMETERS_4 ;\ -uint32_t VBLANK_PARAMETERS_5 ;\ - uint32_t VBLANK_PARAMETERS_6 + uint32_t VMID_SETTINGS_0 +#define DCN2_HUBP_REG_FIELD_VARIABLE_LIST (type) \ - DCN_HUBP_REG_FIELD_LIST(type); \ + DCN_HUBP_REG_FIELD_BASE_LIST(type); \ type DMDATA_ADDRESS_HIGH;\ type DMDATA_MODE;\ type DMDATA_UPDATED;\
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c +41 -2
61 lines not on this page · raw message
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.h +5 -9
@@ -91,13 +91,9 @@ void dcn20_pipe_control_lock_global(void dcn20_setup_gsl_group_as_lock (const struct dc *dc,struct pipe_ctx *pipe_ctx,bool enable); -void dcn20_pipe_control_lock ( -struct dc *dc, -struct pipe_ctx *pipe, -bool lock); -void dcn20_disable_plane (struct dc *dc,struct pipe_ctx *pipe_ctx); -void dcn20_enable_plane ( -struct dc *dc, -struct pipe_ctx *pipe_ctx, -struct dc_state *context); +void dcn20_dccg_init (struct dce_hwseq *hws); +void dcn20_init_blank ( +struct dc *dc, +struct timing_generator *tg); +void dcn20_display_init (struct dc *dc);#endif /* __DC_HWSS_DCN20_H__ */
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 18/87 drm/amd/display: 3.2.38
In reply to a message not in the archiveFrom: Aric Cyr <[email protected]>
--- drivers/gpu/drm/amd/display/dc/dc.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/dc.h +1 -1
@@ -39,7 +39,7 @@#include "inc/hw/dmcu.h" #include "dml/display_mode_lib.h" -#define DC_VER "3.2.37" +#define DC_VER "3.2.38"#define MAX_SURFACES 3#define MAX_PLANES 6
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 19/87 drm/amd/display: Incorrect Read Interval Time For CR Sequence
In reply to a message not in the archiveFrom: David Galiffi <[email protected]>
[WHY]TRAINING_AUX_RD_INTERVAL (DPCD 000Eh) modifies the read interval for the EQ training sequence. CR read interval should remain 100 us. Currently, the CR interval is also being modified.
[HOW]lt_settings->cr_pattern_time should always be 100 us.
--- drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c +1 -1
@@ -1035,7 +1035,7 @@ static void initialize_training_settings(if (link->preferred_training_settings.cr_pattern_time !=NULL ) lt_settings->cr_pattern_time = *link->preferred_training_settings.cr_pattern_time;else - lt_settings->cr_pattern_time =get_training_aux_rd_interval (link,100 ); + lt_settings->cr_pattern_time =100 ;if (link->preferred_training_settings.eq_pattern_time !=NULL ) lt_settings->eq_pattern_time = *link->preferred_training_settings.eq_pattern_time;
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 20/87 drm/amd/display: Clock does not lower in Updateplanes
In reply to a message not in the archiveFrom: Murton Liu <[email protected]>
[why]
We reset the optimized_required in atomic_plane_disable
flag immediately after it is set in atomic_plane_disconnect, causing us to never have flag set during next flip in UpdatePlanes.
[how]
Optimize directly after each time plane is removed.
--- drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c | 6 ++++++ 1 file changed, 6 insertions(+)
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c +6 -0
@@ -2515,6 +2515,12 @@ static void dcn10_apply_ctx_for_surface(if (removed_pipe[i])dcn10_disable_plane (dc, &dc->current_state->res_ctx.pipe_ctx[i]); +for (i =0 ; i < dc->res_pool->pipe_count; i++) +if (removed_pipe[i]) { + dc->hwss.optimize_bandwidth (dc, context); +break ; + } +if (dc->hwseq->wa.DEGVIDCN10_254)hubbub1_wm_change_req_wa (dc->res_pool->hubbub); }
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 21/87 drm/amd/display: Copy max_clks_by_state after dce_clk_mgr_construct
In reply to a message not in the archiveFrom: Nicholas Kazlauskas <[email protected]>
[Why]
For DCE110, DCE112 and DCE120 the max_clks_by_state for the clk_mgr are copied from their respective table before the call to dce_clk_mgr_construct, but then dce_clk_mgr_construct overwrites these with the dce80_max_clks_by_state.
[How]
Copy these after we call dce_clk_mgr_construct so we're using the right tables.
--- .../gpu/drm/amd/display/dc/clk_mgr/dce110/dce110_clk_mgr.c | 4 ++-- .../gpu/drm/amd/display/dc/clk_mgr/dce112/dce112_clk_mgr.c | 4 ++-- .../gpu/drm/amd/display/dc/clk_mgr/dce120/dce120_clk_mgr.c | 4 ++-- 3 files changed, 6 insertions(+), 6 deletions(-)
drivers/gpu/drm/amd/display/dc/clk_mgr/dce110/dce110_clk_mgr.c +2 -2
@@ -262,12 +262,12 @@ void dce110_clk_mgr_construct(struct dc_context *ctx, struct clk_mgr_internal *clk_mgr) { +dce_clk_mgr_construct (ctx ,clk_mgr ); +memcpy (clk_mgr->max_clks_by_state, dce110_max_clks_by_state,sizeof (dce110_max_clks_by_state)); -dce_clk_mgr_construct (ctx, clk_mgr); - clk_mgr->regs = &disp_clk_regs; clk_mgr->clk_mgr_shift = &disp_clk_shift; clk_mgr->clk_mgr_mask = &disp_clk_mask;
drivers/gpu/drm/amd/display/dc/clk_mgr/dce112/dce112_clk_mgr.c +2 -2
@@ -226,12 +226,12 @@ void dce112_clk_mgr_construct(struct dc_context *ctx, struct clk_mgr_internal *clk_mgr) { +dce_clk_mgr_construct (ctx ,clk_mgr ); +memcpy (clk_mgr->max_clks_by_state, dce112_max_clks_by_state,sizeof (dce112_max_clks_by_state)); -dce_clk_mgr_construct (ctx, clk_mgr); - clk_mgr->regs = &disp_clk_regs; clk_mgr->clk_mgr_shift = &disp_clk_shift; clk_mgr->clk_mgr_mask = &disp_clk_mask;
drivers/gpu/drm/amd/display/dc/clk_mgr/dce120/dce120_clk_mgr.c +2 -2
@@ -127,12 +127,12 @@ static struct clk_mgr_funcs dce120_funcs = {void dce120_clk_mgr_construct (struct dc_context *ctx,struct clk_mgr_internal *clk_mgr) { +dce_clk_mgr_construct (ctx, clk_mgr); +memcpy (clk_mgr->max_clks_by_state, dce120_max_clks_by_state,sizeof (dce120_max_clks_by_state)); -dce_clk_mgr_construct (ctx, clk_mgr); - clk_mgr->base.dprefclk_khz =600000 ; clk_mgr->base.funcs = &dce120_funcs; }
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 22/87 drm/amd/display:Use Pixel clock in 100Hz units for HDMI Audio wall clock DTO
In reply to a message not in the archiveFrom: Nevenko Stupar <[email protected]>
[Why]
-Pass and use pixel clock in 100 Hz to Audio for HDMI audio DTO for Audio wall clock programming so audio DTO gets increased precision for timings with /1001 factor. -For HDMI TMDS for N and CTS ACR tables are based on 10 KHz units, these does not need to be modified as N and CTS values are still valid using current tables.
--- .../gpu/drm/amd/display/dc/dce/dce_audio.c | 28 ++++++++--------- .../amd/display/dc/dce/dce_stream_encoder.c | 30 +++++++++---------- .../display/dc/dce110/dce110_hw_sequencer.c | 22 +++++++------- .../display/dc/dcn10/dcn10_stream_encoder.c | 30 +++++++++---------- .../display/dc/dcn10/dcn10_stream_encoder.h | 4 +-- .../gpu/drm/amd/display/include/audio_types.h | 4 +-- 6 files changed, 59 insertions(+), 59 deletions(-)
drivers/gpu/drm/amd/display/dc/dce/dce_audio.c +14 -14
@@ -143,20 +143,20 @@ static void check_audio_bandwidth_hdmi(if (channel_count >2 ) {/* Based on HDMI spec 1.3 Table 7.5 */ -if ((crtc_info->requested_pixel_clock <=27000 ) && +if ((crtc_info->requested_pixel_clock_100Hz <=270000 ) && (crtc_info->v_active <=576 ) && !(crtc_info->interlaced) && !(crtc_info->pixel_repetition ==2 || crtc_info->pixel_repetition ==4 )) { limit_freq_to_48_khz = true; - }else if ((crtc_info->requested_pixel_clock <=27000 ) && + }else if ((crtc_info->requested_pixel_clock_100Hz <=270000 ) && (crtc_info->v_active <=576 ) && (crtc_info->interlaced) && (crtc_info->pixel_repetition ==2 )) { limit_freq_to_88_2_khz = true; - }else if ((crtc_info->requested_pixel_clock <=54000 ) && + }else if ((crtc_info->requested_pixel_clock_100Hz <=540000 ) && (crtc_info->v_active <=576 ) && !(crtc_info->interlaced)) { limit_freq_to_174_4_khz = true;
@@ -735,8 +735,8 @@ void dce_aud_az_configure(/* search pixel clock value for Azalia HDMI Audio */ static void get_azalia_clock_info_hdmi ( -uint32_t crtc_pixel_clock_in_khz, -uint32_t actual_pixel_clock_in_khz, +uint32_t crtc_pixel_clock_100hz, +uint32_t actual_pixel_clock_100Hz,struct azalia_clock_info *azalia_clock_info) {/* audio_dto_phase= 24 * 10,000;
@@ -747,11 +747,11 @@ static void get_azalia_clock_info_hdmi(/* audio_dto_module = PCLKFrequency * 10,000; * [khz] -> [100Hz] */ azalia_clock_info->audio_dto_module = - actual_pixel_clock_in_khz *10 ; + actual_pixel_clock_100Hz; }static void get_azalia_clock_info_dp ( -uint32_t requested_pixel_clock_in_khz, +uint32_t requested_pixel_clock_100Hz,const struct audio_pll_info *pll_info,struct azalia_clock_info *azalia_clock_info) {
@@ -790,15 +790,15 @@ void dce_aud_wall_dto_setup(/* calculate DTO settings */ get_azalia_clock_info_hdmi ( - crtc_info->requested_pixel_clock, - crtc_info->calculated_pixel_clock, + crtc_info->requested_pixel_clock_100Hz, + crtc_info->calculated_pixel_clock_100Hz, &clock_info); -DC_LOG_HW_AUDIO ("\n%s:Input::requested_pixel_clock = %d" \ -"calculated_pixel_clock =%d\n" \ +DC_LOG_HW_AUDIO ("\n%s:Input::requested_pixel_clock_100Hz = %d" \ +"calculated_pixel_clock_100Hz =%d\n" \"audio_dto_module = %d audio_dto_phase =%d \n\n" , __func__,\ - crtc_info->requested_pixel_clock,\ - crtc_info->calculated_pixel_clock,\ + crtc_info->requested_pixel_clock_100Hz,\ + crtc_info->calculated_pixel_clock_100Hz,\ clock_info.audio_dto_module,\ clock_info.audio_dto_phase);
@@ -831,7 +831,7 @@ void dce_aud_wall_dto_setup(calculate DTO settings */ get_azalia_clock_info_dp ( - crtc_info->requested_pixel_clock, + crtc_info->requested_pixel_clock_100Hz, pll_info, &clock_info);
drivers/gpu/drm/amd/display/dc/dce/dce_stream_encoder.c +15 -15
@@ -1249,13 +1249,13 @@ static uint32_t calc_max_audio_packets_per_line(static void get_audio_clock_info (enum dc_color_depth color_depth, -uint32_t crtc_pixel_clock_in_khz, -uint32_t actual_pixel_clock_in_khz, +uint32_t crtc_pixel_clock_100Hz, +uint32_t actual_pixel_clock_100Hz,struct audio_clock_info *audio_clock_info) {const struct audio_clock_info *clock_info;uint32_t index; -uint32_t crtc_pixel_clock_in_10khz = crtc_pixel_clock_in_khz /10 ; +uint32_t crtc_pixel_clock_in_10khz = crtc_pixel_clock_100Hz /100 ;uint32_t audio_array_size;switch (color_depth) {
@@ -1292,16 +1292,16 @@ static void get_audio_clock_info( }/* not found */ -if (actual_pixel_clock_in_khz ==0 ) - actual_pixel_clock_in_khz = crtc_pixel_clock_in_khz; +if (actual_pixel_clock_100Hz ==0 ) + actual_pixel_clock_100Hz = crtc_pixel_clock_100Hz;/* See HDMI spec the table entry under * pixel clock of "Other". */ audio_clock_info->pixel_clock_in_10khz = - actual_pixel_clock_in_khz /10 ; - audio_clock_info->cts_32khz = actual_pixel_clock_in_khz; - audio_clock_info->cts_44khz = actual_pixel_clock_in_khz; - audio_clock_info->cts_48khz = actual_pixel_clock_in_khz; + actual_pixel_clock_100Hz /100 ; + audio_clock_info->cts_32khz = actual_pixel_clock_100Hz /10 ; + audio_clock_info->cts_44khz = actual_pixel_clock_100Hz /10 ; + audio_clock_info->cts_48khz = actual_pixel_clock_100Hz /10 ; audio_clock_info->n_32khz =4096 ; audio_clock_info->n_44khz =6272 ;
@@ -1367,14 +1367,14 @@ static void dce110_se_setup_hdmi_audio(/* Program audio clock sample/regeneration parameters */ get_audio_clock_info (crtc_info->color_depth, - crtc_info->requested_pixel_clock, - crtc_info->calculated_pixel_clock, + crtc_info->requested_pixel_clock_100Hz, + crtc_info->calculated_pixel_clock_100Hz, &audio_clock_info);DC_LOG_HW_AUDIO ( -"\n%s:Input::requested_pixel_clock = %d" \ -"calculated_pixel_clock = %d \n" , __func__, \ - crtc_info->requested_pixel_clock, \ - crtc_info->calculated_pixel_clock); +"\n%s:Input::requested_pixel_clock_100Hz = %d" \ +"calculated_pixel_clock_100Hz = %d \n" , __func__, \ + crtc_info->requested_pixel_clock_100Hz, \ + crtc_info->calculated_pixel_clock_100Hz);/* HDMI_ACR_32_0__HDMI_ACR_CTS_32_MASK */ REG_UPDATE (HDMI_ACR_32_0 ,HDMI_ACR_CTS_32 , audio_clock_info.cts_32khz);
drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c +11 -11
@@ -1159,27 +1159,27 @@ static void build_audio_output( stream->timing.flags.INTERLACE; audio_output->crtc_info.refresh_rate = - (stream->timing.pix_clk_100hz*10000 )/ + (stream->timing.pix_clk_100hz*100 )/ (stream->timing.h_total*stream->timing.v_total); audio_output->crtc_info.color_depth = stream->timing.display_color_depth; - audio_output->crtc_info.requested_pixel_clock = - pipe_ctx->stream_res.pix_clk_params.requested_pix_clk_100hz /10 ; + audio_output->crtc_info.requested_pixel_clock_100Hz = + pipe_ctx->stream_res.pix_clk_params.requested_pix_clk_100hz; - audio_output->crtc_info.calculated_pixel_clock = - pipe_ctx->stream_res.pix_clk_params.requested_pix_clk_100hz /10 ; + audio_output->crtc_info.calculated_pixel_clock_100Hz = + pipe_ctx->stream_res.pix_clk_params.requested_pix_clk_100hz;/*for HDMI, audio ACR is with deep color ratio factor*/ if (dc_is_hdmi_signal (pipe_ctx->stream->signal) && - audio_output->crtc_info.requested_pixel_clock == - (stream->timing.pix_clk_100hz /10 )) { + audio_output->crtc_info.requested_pixel_clock_100Hz == + (stream->timing.pix_clk_100hz)) {if (pipe_ctx->stream_res.pix_clk_params.pixel_encoding ==PIXEL_ENCODING_YCBCR420 ) { - audio_output->crtc_info.requested_pixel_clock = - audio_output->crtc_info.requested_pixel_clock/2 ; - audio_output->crtc_info.calculated_pixel_clock = - pipe_ctx->stream_res.pix_clk_params.requested_pix_clk_100hz/20 ; + audio_output->crtc_info.requested_pixel_clock_100Hz = + audio_output->crtc_info.requested_pixel_clock_100Hz/2 ; + audio_output->crtc_info.calculated_pixel_clock_100Hz = + pipe_ctx->stream_res.pix_clk_params.requested_pix_clk_100hz/2 ; } }
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_stream_encoder.c +15 -15
@@ -1195,13 +1195,13 @@ static union audio_cea_channels speakers_to_channels(void get_audio_clock_info (enum dc_color_depth color_depth, -uint32_t crtc_pixel_clock_in_khz, -uint32_t actual_pixel_clock_in_khz, +uint32_t crtc_pixel_clock_100Hz, +uint32_t actual_pixel_clock_100Hz,struct audio_clock_info *audio_clock_info) {const struct audio_clock_info *clock_info;uint32_t index; -uint32_t crtc_pixel_clock_in_10khz = crtc_pixel_clock_in_khz /10 ; +uint32_t crtc_pixel_clock_in_10khz = crtc_pixel_clock_100Hz /100 ;uint32_t audio_array_size;switch (color_depth) {
@@ -1238,16 +1238,16 @@ void get_audio_clock_info( }/* not found */ -if (actual_pixel_clock_in_khz ==0 ) - actual_pixel_clock_in_khz = crtc_pixel_clock_in_khz; +if (actual_pixel_clock_100Hz ==0 ) + actual_pixel_clock_100Hz = crtc_pixel_clock_100Hz;/* See HDMI spec the table entry under * pixel clock of "Other". */ audio_clock_info->pixel_clock_in_10khz = - actual_pixel_clock_in_khz /10 ; - audio_clock_info->cts_32khz = actual_pixel_clock_in_khz; - audio_clock_info->cts_44khz = actual_pixel_clock_in_khz; - audio_clock_info->cts_48khz = actual_pixel_clock_in_khz; + actual_pixel_clock_100Hz /100 ; + audio_clock_info->cts_32khz = actual_pixel_clock_100Hz /10 ; + audio_clock_info->cts_44khz = actual_pixel_clock_100Hz /10 ; + audio_clock_info->cts_48khz = actual_pixel_clock_100Hz /10 ; audio_clock_info->n_32khz =4096 ; audio_clock_info->n_44khz =6272 ;
@@ -1307,14 +1307,14 @@ static void enc1_se_setup_hdmi_audio(/* Program audio clock sample/regeneration parameters */ get_audio_clock_info (crtc_info->color_depth, - crtc_info->requested_pixel_clock, - crtc_info->calculated_pixel_clock, + crtc_info->requested_pixel_clock_100Hz, + crtc_info->calculated_pixel_clock_100Hz, &audio_clock_info);DC_LOG_HW_AUDIO ( -"\n%s:Input::requested_pixel_clock = %d" \ -"calculated_pixel_clock = %d \n" , __func__, \ - crtc_info->requested_pixel_clock, \ - crtc_info->calculated_pixel_clock); +"\n%s:Input::requested_pixel_clock_100Hz = %d" \ +"calculated_pixel_clock_100Hz = %d \n" , __func__, \ + crtc_info->requested_pixel_clock_100Hz, \ + crtc_info->calculated_pixel_clock_100Hz);/* HDMI_ACR_32_0__HDMI_ACR_CTS_32_MASK */ REG_UPDATE (HDMI_ACR_32_0 ,HDMI_ACR_CTS_32 , audio_clock_info.cts_32khz);
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_stream_encoder.h +2 -2
@@ -605,8 +605,8 @@ void enc1_se_enable_dp_audio(void get_audio_clock_info (enum dc_color_depth color_depth, -uint32_t crtc_pixel_clock_in_khz, -uint32_t actual_pixel_clock_in_khz, +uint32_t crtc_pixel_clock_100Hz, +uint32_t actual_pixel_clock_100Hz,struct audio_clock_info *audio_clock_info);#endif /* __DC_STREAM_ENCODER_DCN10_H__ */
drivers/gpu/drm/amd/display/include/audio_types.h +2 -2
@@ -38,8 +38,8 @@ struct audio_crtc_info {uint32_t h_active;uint32_t v_active;uint32_t pixel_repetition; -uint32_t requested_pixel_clock;/* in KHz */ -uint32_t calculated_pixel_clock;/* in KHz */ +uint32_t requested_pixel_clock_100Hz;/* in 100Hz */ +uint32_t calculated_pixel_clock_100Hz;/* in 100Hz */ uint32_t refresh_rate;enum dc_color_depth color_depth;bool interlaced;
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 23/87 drm/amd/display: wait for the whole frame after global unlock
In reply to a message not in the archiveFrom: Wenjing Liu <[email protected]>
[why]
The current code will not wait for the entire frame
after global unlock.
This causes dsc dynamic target bpp update corruption when there is a surface update immediately happens after this.
[how]
Wait for the entire whole frame after unlock before continuing the rest of stream and surface update.
--- drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c | 2 ++ 1 file changed, 2 insertions(+)
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c +2 -0
@@ -1256,6 +1256,8 @@ void dcn20_pipe_control_lock_global( CRTC_STATE_VACTIVE); pipe->stream_res.tg->funcs->wait_for_state (pipe->stream_res.tg,CRTC_STATE_VBLANK ); + pipe->stream_res.tg->funcs->wait_for_state (pipe->stream_res.tg, +CRTC_STATE_VACTIVE ); pipe->stream_res.tg->funcs->lock_doublebuffer_disable ( pipe->stream_res.tg); }
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 24/87 drm/amd/display: refactor dump_clk_registers
In reply to a message not in the archiveFrom: Su Sung Chung <[email protected]>
[why]
for 2 purposes:
1. get raw register value dumped on the log, which will make it easier to talk to other team who only knows about the register
2. enable other HW to be able to use the same interface as raven to log clock register data
--- drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer_debug.c | 2 +- drivers/gpu/drm/amd/display/dc/dm_services.h | 1 + 2 files changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer_debug.c +1 -1
@@ -45,7 +45,7 @@#include "dcn10_cm_common.h" #include "clk_mgr.h" -static unsigned int snprintf_count (char *pBuf,unsigned int bufSize,char *fmt, ...) +unsigned int snprintf_count (char *pBuf,unsigned int bufSize,char *fmt, ...) {unsigned int ret_vsnprintf;unsigned int chars_printed;
drivers/gpu/drm/amd/display/dc/dm_services.h +1 -0
@@ -151,6 +151,7 @@ void generic_reg_wait(const struct dc_context *ctx, unsigned int delay_between_poll_us, unsigned int time_out_num_tries, const char *func_name, int line); +unsigned int snprintf_count (char *pBuf,unsigned int bufSize,char *fmt, ...);/* These macros need to be used with soc15 registers in order to retrieve * the actual offset.
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 25/87 drm/amd/display: add hdmi2.1 dsc pps packet programming
In reply to a message not in the archiveFrom: Dmytro Laktyushkin <[email protected]>
This change adds EMP packet programming for enabling dsc with hdmi. The packets are structured according to VESA HDMI 2.1x r2 spec, section 10.10.2.2.
--- drivers/gpu/drm/amd/display/dc/core/dc_link.c | 14 +++++++------- drivers/gpu/drm/amd/display/dc/core/dc_link_hwss.c | 13 ++++++------- drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c | 3 ++- drivers/gpu/drm/amd/display/dc/inc/dc_link_dp.h | 1 + 4 files changed, 16 insertions(+), 15 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc_link.c +7 -7
@@ -2769,10 +2769,10 @@ void core_link_enable_stream(allocate_mst_payload (pipe_ctx);#ifdef CONFIG_DRM_AMD_DC_DSC_SUPPORT -if (pipe_ctx->stream->timing.flags.DSC && - (dc_is_dp_signal (pipe_ctx->stream->signal) || -dc_is_virtual_signal (pipe_ctx->stream->signal))) { -dp_set_dsc_enable (pipe_ctx, true); +if (pipe_ctx->stream->timing.flags.DSC) { +if (dc_is_dp_signal (pipe_ctx->stream->signal) || +dc_is_virtual_signal (pipe_ctx->stream->signal)) +dp_set_dsc_enable (pipe_ctx, true); pipe_ctx->stream_res.tg->funcs->wait_for_state ( pipe_ctx->stream_res.tg,CRTC_STATE_VBLANK );
@@ -2833,9 +2833,9 @@ void core_link_disable_stream(struct pipe_ctx *pipe_ctx, int option)disable_link (pipe_ctx->stream->link, pipe_ctx->stream->signal);#ifdef CONFIG_DRM_AMD_DC_DSC_SUPPORT -if (pipe_ctx->stream->timing.flags.DSC && -dc_is_dp_signal (pipe_ctx->stream->signal)) { -dp_set_dsc_enable (pipe_ctx, false); +if (pipe_ctx->stream->timing.flags.DSC) { +if (dc_is_dp_signal (pipe_ctx->stream->signal)) +dp_set_dsc_enable (pipe_ctx, false); }#endif }
drivers/gpu/drm/amd/display/dc/core/dc_link_hwss.c +6 -7
@@ -396,7 +396,7 @@ static bool dp_set_dsc_on_rx(struct pipe_ctx *pipe_ctx, bool enable)/* This has to be done after DSC was enabled on RX first, i.e. after dp_enable_dsc_on_rx() had been called */ -static void dp_set_dsc_on_stream (struct pipe_ctx *pipe_ctx,bool enable) +void set_dsc_on_stream (struct pipe_ctx *pipe_ctx,bool enable) {struct display_stream_compressor *dsc = pipe_ctx->stream_res.dsc;struct dc *core_dc = pipe_ctx->stream->ctx->dc;
@@ -435,7 +435,7 @@ static void dp_set_dsc_on_stream(struct pipe_ctx *pipe_ctx, bool enable)dsc_optc_config_log (dsc, &dsc_optc_cfg);/* Enable DSC in encoder */ -if (!IS_FPGA_MAXIMUS_DC (core_dc->ctx->dce_environment) && pipe_ctx->stream_res.stream_enc->funcs->dp_set_dsc_config) +if (dc_is_dp_signal (stream->signal) && !IS_FPGA_MAXIMUS_DC (core_dc->ctx->dce_environment)) pipe_ctx->stream_res.stream_enc->funcs->dp_set_dsc_config(pipe_ctx->stream_res.stream_enc, optc_dsc_mode, dsc_optc_cfg.bytes_per_pixel,
@@ -454,11 +454,10 @@ static void dp_set_dsc_on_stream(struct pipe_ctx *pipe_ctx, bool enable) OPTC_DSC_DISABLED, 0, 0);/* disable DSC in stream encoder */ -if (!IS_FPGA_MAXIMUS_DC (core_dc->ctx->dce_environment)) { +if (dc_is_dp_signal (stream->signal) && !IS_FPGA_MAXIMUS_DC (core_dc->ctx->dce_environment)) pipe_ctx->stream_res.stream_enc->funcs->dp_set_dsc_config ( pipe_ctx->stream_res.stream_enc,OPTC_DSC_DISABLED ,0 ,0 ,NULL ); - }/* disable DSC block */ pipe_ctx->stream_res.dsc->funcs->dsc_disable (pipe_ctx->stream_res.dsc);
@@ -479,12 +478,12 @@ bool dp_set_dsc_enable(struct pipe_ctx *pipe_ctx, bool enable)if (enable) {if (dp_set_dsc_on_rx (pipe_ctx, true)) { -dp_set_dsc_on_stream (pipe_ctx, true); +set_dsc_on_stream (pipe_ctx, true); result = true; } }else {dp_set_dsc_on_rx (pipe_ctx, false); -dp_set_dsc_on_stream (pipe_ctx, false); +set_dsc_on_stream (pipe_ctx, false); result = true; } out:
@@ -500,7 +499,7 @@ bool dp_update_dsc_config(struct pipe_ctx *pipe_ctx)if (!dsc)return false; -dp_set_dsc_on_stream (pipe_ctx, true); +set_dsc_on_stream (pipe_ctx, true);return true; }
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c +2 -1
@@ -1758,8 +1758,9 @@ static void dcn20_reset_back_end_for_pipe( } }#ifdef CONFIG_DRM_AMD_DC_DSC_SUPPORT - else if (pipe_ctx->stream_res.dsc) + else if (pipe_ctx->stream_res.dsc) {dp_set_dsc_enable (pipe_ctx, false); + }#endif /* by upper caller loop, parent pipe: pipe0, will be reset last.
drivers/gpu/drm/amd/display/dc/inc/dc_link_dp.h +1 -0
@@ -66,6 +66,7 @@ void dp_enable_mst_on_sink(struct dc_link *link, bool enable);void dp_set_fec_ready (struct dc_link *link,bool ready);void dp_set_fec_enable (struct dc_link *link,bool enable);bool dp_set_dsc_enable (struct pipe_ctx *pipe_ctx,bool enable); +void set_dsc_on_stream (struct pipe_ctx *pipe_ctx,bool enable);bool dp_update_dsc_config (struct pipe_ctx *pipe_ctx);#endif
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 26/87 drm/amd/display: add monitor patch to add T7 delay
In reply to a message not in the archiveFrom: Anthony Koo <[email protected]>
[Why]
Specifically to one panel,
TCON is able to accept active video signal quickly, but the Source Driver requires 2-3 frames of extra time.
It is a Panel issue since TCON needs to take care of all Sink requirements including Source Driver. But in this case it does not.
Customer is asking to add fixed T7 delay as panel workaround.
[How]
Add monitor specific patch to add T7 delay
--- drivers/gpu/drm/amd/display/dc/core/dc_link_hwss.c | 4 ++++ drivers/gpu/drm/amd/display/dc/dc_types.h | 1 + 2 files changed, 5 insertions(+)
drivers/gpu/drm/amd/display/dc/core/dc_link_hwss.c +4 -0
@@ -162,6 +162,10 @@ bool edp_receiver_ready_T7(struct dc_link *link)break ;udelay (25 );//MAx T7 is 50ms }while (++tries <300 ); + +if (link->local_sink->edid_caps.panel_patch.extra_t7_ms >0 ) +udelay (link->local_sink->edid_caps.panel_patch.extra_t7_ms *1000 ); +return result; }
drivers/gpu/drm/amd/display/dc/dc_types.h +1 -0
@@ -202,6 +202,7 @@ struct dc_panel_patch {unsigned int dppowerup_delay;unsigned int extra_t12_ms;unsigned int extra_delay_backlight_off; +unsigned int extra_t7_ms; };struct dc_edid_caps {
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 27/87 drm/amd/display: Poll for GPUVM context ready
In reply to a message not in the archiveFrom: Julian Parkin <[email protected]>
[Why]
Hardware docs state that we must wait until the GPUVM context is ready after programming it.
[How]
Poll until the valid bit of PAGE_TABLE_BASE_ADDR_LO32 is set to 1 after programming it.
--- .../gpu/drm/amd/display/dc/dcn20/dcn20_vmid.c | 35 +++++++++++++++++++ 1 file changed, 35 insertions(+)
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_vmid.c +35 -0
@@ -36,6 +36,38 @@#define FN (reg_name, field_name) \ vmid->shifts->field_name, vmid->masks->field_name +static void dcn20_wait_for_vmid_ready (struct dcn20_vmid *vmid) +{ +/* According the hardware spec, we need to poll for the lowest +* bit of PAGE_TABLE_BASE_ADDR_LO32 = 1 any time a GPUVM +* context is updated. We can't use REG_WAIT here since we +* don't have a seperate field to wait on. +* +* TODO: Confirm timeout / poll interval with hardware team +*/ + +int max_times =10000 ; +int delay_us =5 ; +int i; + +for (i =0 ; i < max_times; ++i) { +uint32_t entry_lo32; + +REG_GET (PAGE_TABLE_BASE_ADDR_LO32 , +VM_CONTEXT0_PAGE_DIRECTORY_ENTRY_LO32 , + &entry_lo32); + +if (entry_lo32 &0x1 ) +return ; + +udelay (delay_us); + } + +/* VM setup timed out */ +DC_LOG_WARNING ("Timeout while waiting for GPUVM context update\n" ); +ASSERT (0 ); +} +void dcn20_vmid_setup (struct dcn20_vmid *vmid,const struct dcn_vmid_page_table_config *config) { REG_SET(PAGE_TABLE_START_ADDR_HI32, 0,
@@ -54,6 +86,9 @@ void dcn20_vmid_setup(struct dcn20_vmid *vmid, const struct dcn_vmid_page_table_REG_SET (PAGE_TABLE_BASE_ADDR_HI32 ,0 ,VM_CONTEXT0_PAGE_DIRECTORY_ENTRY_HI32 , (config->page_table_base_addr >>32 ) &0xFFFFFFFF ); +/* Note: per hardware spec PAGE_TABLE_BASE_ADDR_LO32 must be programmed last in sequence */ REG_SET (PAGE_TABLE_BASE_ADDR_LO32 ,0 ,VM_CONTEXT0_PAGE_DIRECTORY_ENTRY_LO32 , config->page_table_base_addr &0xFFFFFFFF ); + +dcn20_wait_for_vmid_ready (vmid); }
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 28/87 drm/amd/display: add functionality to grab DPRX CRC entries.
In reply to a message not in the archiveFrom: Dingchen Zhang <[email protected]>
[Why]
We need to compare DPRX CRCs with framebuffer CRCs for digital bypass mode.
[How]
Hook into DRM to grab DP receiver CRCs through drm_dp_start_crc.
--- .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 17 ++-- .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h | 16 +--- .../drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c | 91 ++++++++++++++----- .../drm/amd/display/amdgpu_dm/amdgpu_dm_crc.h | 56 ++++++++++++ 4 files changed, 136 insertions(+), 44 deletions(-) create mode 100644 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.h
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +11 -6
@@ -3736,7 +3736,7 @@ dm_crtc_duplicate_state(struct drm_crtc *crtc) state->abm_level = cur->abm_level; state->vrr_supported = cur->vrr_supported; state->freesync_config = cur->freesync_config; - state->crc_enabled = cur->crc_enabled; + state->crc_src = cur->crc_src; state->cm_has_degamma = cur->cm_has_degamma; state->cm_is_degamma_srgb = cur->cm_is_degamma_srgb;
@@ -5936,6 +5936,7 @@ static void amdgpu_dm_enable_crtc_interrupts(struct drm_device *dev,struct drm_crtc *crtc;struct drm_crtc_state *old_crtc_state, *new_crtc_state;int i; +enum amdgpu_dm_pipe_crc_source source;for_each_oldnew_crtc_in_state (state, crtc, old_crtc_state, new_crtc_state, i) {
@@ -5961,9 +5962,13 @@ static void amdgpu_dm_enable_crtc_interrupts(struct drm_device *dev,#ifdef CONFIG_DEBUG_FS /* The stream has changed so CRC capture needs to re-enabled. */ -if (dm_new_crtc_state->crc_enabled) { - dm_new_crtc_state->crc_enabled = false; -amdgpu_dm_crtc_set_crc_source (crtc,"auto" ); + source = dm_new_crtc_state->crc_src; +if (amdgpu_dm_is_valid_crc_source (source)) { + dm_new_crtc_state->crc_src =AMDGPU_DM_PIPE_CRC_SOURCE_NONE ; +if (source ==AMDGPU_DM_PIPE_CRC_SOURCE_CRTC ) +amdgpu_dm_crtc_set_crc_source (crtc,"crtc" ); +else if (source ==AMDGPU_DM_PIPE_CRC_SOURCE_DPRX ) +amdgpu_dm_crtc_set_crc_source (crtc,"dprx" ); }#endif }
@@ -6019,7 +6024,7 @@ static int amdgpu_dm_atomic_commit(struct drm_device *dev,* Drop the extra vblank reference added by CRC * capture if applicable. */ -if (dm_new_crtc_state->crc_enabled) +if (amdgpu_dm_is_valid_crc_source (dm_new_crtc_state->crc_src))drm_crtc_vblank_put (crtc);/*
@@ -6027,7 +6032,7 @@ static int amdgpu_dm_atomic_commit(struct drm_device *dev,* still a stream for the CRTC. */ if (!dm_new_crtc_state->stream) - dm_new_crtc_state->crc_enabled = false; + dm_new_crtc_state->crc_src =AMDGPU_DM_PIPE_CRC_SOURCE_NONE ;manage_dm_interrupts (adev, acrtc, false); }
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h +2 -14
@@ -47,6 +47,7 @@#include "irq_types.h" #include "signal_types.h" +#include "amdgpu_dm_crc.h" /* Forward declarations */ struct amdgpu_device ;
@@ -310,7 +311,7 @@ struct dm_crtc_state {bool interrupts_enabled;int crc_skip_count; -bool crc_enabled; +enum amdgpu_dm_pipe_crc_source crc_src;bool freesync_timing_changed;bool freesync_vrr_info_changed;
@@ -378,19 +379,6 @@ void dm_restore_drm_connector_state(struct drm_device *dev,void amdgpu_dm_update_freesync_caps (struct drm_connector *connector,struct edid *edid); -/* amdgpu_dm_crc.c */ -#ifdef CONFIG_DEBUG_FS -int amdgpu_dm_crtc_set_crc_source (struct drm_crtc *crtc,const char *src_name); -int amdgpu_dm_crtc_verify_crc_source (struct drm_crtc *crtc, -const char *src_name, -size_t *values_cnt); -void amdgpu_dm_crtc_handle_crc_irq (struct drm_crtc *crtc); -#else -#define amdgpu_dm_crtc_set_crc_source NULL -#define amdgpu_dm_crtc_verify_crc_source NULL -#define amdgpu_dm_crtc_handle_crc_irq (x) -#endif -#define MAX_COLOR_LUT_ENTRIES 4096/* Legacy gamm LUT users such as X doesn't like large LUT sizes */ #define MAX_COLOR_LEGACY_LUT_ENTRIES 256
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c +67 -24
135 lines not on this page · raw message
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.h new file +56 -0
56 lines not on this page · raw message
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 29/87 drm/amd/display: add functionality to get pipe CRC source.
In reply to a message not in the archiveFrom: Dingchen Zhang <[email protected]>
[Why]
We need to check the pipe crc source through debugfs for bypass mode test.
[How]
add implementation of amdgpu_dm_crtc_get_crc_sources and hook into drm_crtc callback get_crc_sources.
--- drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 1 + .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c | 14 ++++++++++++++ .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.h | 3 +++ 3 files changed, 18 insertions(+)
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +1 -0
@@ -3806,6 +3806,7 @@ static const struct drm_crtc_funcs amdgpu_dm_crtc_funcs = { .atomic_destroy_state = dm_crtc_destroy_state, .set_crc_source = amdgpu_dm_crtc_set_crc_source, .verify_crc_source = amdgpu_dm_crtc_verify_crc_source, + .get_crc_sources = amdgpu_dm_crtc_get_crc_sources, .enable_vblank = dm_enable_vblank, .disable_vblank = dm_disable_vblank, };
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c +14 -0
@@ -29,6 +29,13 @@#include "amdgpu_dm.h" #include "dc.h" +static const char *const pipe_crc_sources[] = { +"none" , +"crtc" , +"dprx" , +"auto" , +}; +static enum amdgpu_dm_pipe_crc_source dm_parse_crc_source (const char *source) {if (!source || !strcmp (source,"none" ))
@@ -41,6 +48,13 @@ static enum amdgpu_dm_pipe_crc_source dm_parse_crc_source(const char *source)return AMDGPU_DM_PIPE_CRC_SOURCE_INVALID ; } +const char *const *amdgpu_dm_crtc_get_crc_sources (struct drm_crtc *crtc, +size_t *count) +{ + *count =ARRAY_SIZE (pipe_crc_sources); +return pipe_crc_sources; +} +int amdgpu_dm_crtc_verify_crc_source (struct drm_crtc *crtc,const char *src_name,size_t *values_cnt)
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.h +3 -0
@@ -46,10 +46,13 @@ int amdgpu_dm_crtc_set_crc_source(struct drm_crtc *crtc, const char *src_name);int amdgpu_dm_crtc_verify_crc_source (struct drm_crtc *crtc,const char *src_name,size_t *values_cnt); +const char *const *amdgpu_dm_crtc_get_crc_sources (struct drm_crtc *crtc, +size_t *count);void amdgpu_dm_crtc_handle_crc_irq (struct drm_crtc *crtc);#else #define amdgpu_dm_crtc_set_crc_source NULL#define amdgpu_dm_crtc_verify_crc_source NULL +#define amdgpu_dm_crtc_get_crc_sources NULL#define amdgpu_dm_crtc_handle_crc_irq (x)#endif
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 30/87 drm/amd/display: Wait for backlight programming completion in set backlight level
In reply to a message not in the archiveFrom: SivapiriyanKumarasamy <[email protected]>
[WHY]
Currently we don't wait for blacklight programming completion in DMCU when setting backlight level. Some sequences such as PSR static screen event trigger reprogramming requires it to be complete.
[How]
Add generic wait for dmcu command completion in set backlight level.
--- drivers/gpu/drm/amd/display/dc/dce/dce_abm.c | 4 ++++ 1 file changed, 4 insertions(+)
drivers/gpu/drm/amd/display/dc/dce/dce_abm.c +4 -0
@@ -237,6 +237,10 @@ static void dmcu_set_backlight_level( s2 |= (backlight_8_bit <<ATOM_S2_CURRENT_BL_LEVEL_SHIFT );REG_WRITE (BIOS_SCRATCH_2 , s2); + +/* waitDMCUReadyForCmd */ +REG_WAIT (MASTER_COMM_CNTL_REG ,MASTER_COMM_INTERRUPT , +0 ,1 ,80000 ); }static void dce_abm_init (struct abm *abm)
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 31/87 drm/amd/display: 3.2.39
In reply to a message not in the archiveFrom: Aric Cyr <[email protected]>
--- drivers/gpu/drm/amd/display/dc/dc.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/dc.h +1 -1
@@ -39,7 +39,7 @@#include "inc/hw/dmcu.h" #include "dml/display_mode_lib.h" -#define DC_VER "3.2.38" +#define DC_VER "3.2.39"#define MAX_SURFACES 3#define MAX_PLANES 6
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 32/87 drm/amd/display: Expose enc2_set_dynamic_metadata
In reply to a message not in the archiveFrom: Ilya Bakoulin <[email protected]>
[Why]
Need to implement register programming for HDR dynamic metadata transmission and tests.
--- drivers/gpu/drm/amd/display/dc/dcn20/dcn20_stream_encoder.c | 2 +- drivers/gpu/drm/amd/display/dc/dcn20/dcn20_stream_encoder.h | 5 +++++ 2 files changed, 6 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_stream_encoder.c +1 -1
@@ -371,7 +371,7 @@ static void enc2_read_state(struct stream_encoder *enc, struct enc_state *s)* * Ensure the OTG master update lock is set when changing DME configuration. */ -static void enc2_set_dynamic_metadata (struct stream_encoder *enc, +void enc2_set_dynamic_metadata (struct stream_encoder *enc,bool enable_dme,uint32_t hubp_requestor_id,enum dynamic_metadata_mode dmdata_mode)
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_stream_encoder.h +5 -0
@@ -104,4 +104,9 @@ void enc2_stream_encoder_dp_unblank(struct stream_encoder *enc, const struct encoder_unblank_param *param); +void enc2_set_dynamic_metadata (struct stream_encoder *enc, +bool enable_dme, +uint32_t hubp_requestor_id, +enum dynamic_metadata_mode dmdata_mode); +#endif /* __DC_STREAM_ENCODER_DCN20_H__ */
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 33/87 drm/amd/display: Check for valid stream_encode
In reply to a message not in the archiveFrom: Ilya Bakoulin <[email protected]>
Before accessing it's vtable, check that stream_encoder is non-null.
--- drivers/gpu/drm/amd/display/dc/core/dc_stream.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/core/dc_stream.c +2 -1
@@ -609,7 +609,8 @@ bool dc_stream_set_dynamic_metadata(struct dc *dc, pipe_ctx->stream->dmdata_address = attr->address; -if (pipe_ctx->stream_res.stream_enc->funcs->set_dynamic_metadata !=NULL ) { +if (pipe_ctx->stream_res.stream_enc && + pipe_ctx->stream_res.stream_enc->funcs->set_dynamic_metadata !=NULL ) {if (pipe_ctx->stream->dmdata_address.quad_part !=0 ) {/* if using dynamic meta, don't set up generic infopackets */ pipe_ctx->stream_res.encoder_info_frame.hdrsmd.valid = false;
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 34/87 drm/amd/display: Fix some HUBP programming issues
In reply to a message not in the archiveFrom: Ilya Bakoulin <[email protected]>
[Why]
A hubp pointer was being passed to DCN1 functions, which expect the enclosing structure (for the purpose of container_of macros) to be dcn10_hubp, but the actual type was dcn20_hubp.
[How]
Copy existing DCN1 functions and alter them slightly for use with dcn20_hubp.
--- .../gpu/drm/amd/display/dc/dcn20/dcn20_hubp.c | 682 +++++++++++++++++- .../gpu/drm/amd/display/dc/dcn20/dcn20_hubp.h | 37 + 2 files changed, 705 insertions(+), 14 deletions(-)
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hubp.c +668 -14
@@ -156,7 +156,85 @@ void hubp2_program_deadline( {struct dcn20_hubp *hubp2 =TO_DCN20_HUBP (hubp); -hubp1_program_deadline (hubp, dlg_attr, ttu_attr); +/* DLG - Per hubp */ +REG_SET_2 (BLANK_OFFSET_0 ,0 , +REFCYC_H_BLANK_END , dlg_attr->refcyc_h_blank_end, +DLG_V_BLANK_END , dlg_attr->dlg_vblank_end); + +REG_SET (BLANK_OFFSET_1 ,0 , +MIN_DST_Y_NEXT_START , dlg_attr->min_dst_y_next_start); + +REG_SET (DST_DIMENSIONS ,0 , +REFCYC_PER_HTOTAL , dlg_attr->refcyc_per_htotal); + +REG_SET_2 (DST_AFTER_SCALER ,0 , +REFCYC_X_AFTER_SCALER , dlg_attr->refcyc_x_after_scaler, +DST_Y_AFTER_SCALER , dlg_attr->dst_y_after_scaler); + +REG_SET (REF_FREQ_TO_PIX_FREQ ,0 , +REF_FREQ_TO_PIX_FREQ , dlg_attr->ref_freq_to_pix_freq); + +/* DLG - Per luma/chroma */ +REG_SET (VBLANK_PARAMETERS_1 ,0 , +REFCYC_PER_PTE_GROUP_VBLANK_L , dlg_attr->refcyc_per_pte_group_vblank_l); + +if (REG (NOM_PARAMETERS_0 )) +REG_SET (NOM_PARAMETERS_0 ,0 , +DST_Y_PER_PTE_ROW_NOM_L , dlg_attr->dst_y_per_pte_row_nom_l); + +if (REG (NOM_PARAMETERS_1 )) +REG_SET (NOM_PARAMETERS_1 ,0 , +REFCYC_PER_PTE_GROUP_NOM_L , dlg_attr->refcyc_per_pte_group_nom_l); + +REG_SET (NOM_PARAMETERS_4 ,0 , +DST_Y_PER_META_ROW_NOM_L , dlg_attr->dst_y_per_meta_row_nom_l); + +REG_SET (NOM_PARAMETERS_5 ,0 , +REFCYC_PER_META_CHUNK_NOM_L , dlg_attr->refcyc_per_meta_chunk_nom_l); + +REG_SET_2 (PER_LINE_DELIVERY ,0 , +REFCYC_PER_LINE_DELIVERY_L , dlg_attr->refcyc_per_line_delivery_l, +REFCYC_PER_LINE_DELIVERY_C , dlg_attr->refcyc_per_line_delivery_c); + +REG_SET (VBLANK_PARAMETERS_2 ,0 , +REFCYC_PER_PTE_GROUP_VBLANK_C , dlg_attr->refcyc_per_pte_group_vblank_c); + +if (REG (NOM_PARAMETERS_2 )) +REG_SET (NOM_PARAMETERS_2 ,0 , +DST_Y_PER_PTE_ROW_NOM_C , dlg_attr->dst_y_per_pte_row_nom_c); + +if (REG (NOM_PARAMETERS_3 )) +REG_SET (NOM_PARAMETERS_3 ,0 , +REFCYC_PER_PTE_GROUP_NOM_C , dlg_attr->refcyc_per_pte_group_nom_c); + +REG_SET (NOM_PARAMETERS_6 ,0 , +DST_Y_PER_META_ROW_NOM_C , dlg_attr->dst_y_per_meta_row_nom_c); + +REG_SET (NOM_PARAMETERS_7 ,0 , +REFCYC_PER_META_CHUNK_NOM_C , dlg_attr->refcyc_per_meta_chunk_nom_c); + +/* TTU - per hubp */ +REG_SET_2 (DCN_TTU_QOS_WM ,0 , + QoS_LEVEL_LOW_WM, ttu_attr->qos_level_low_wm, + QoS_LEVEL_HIGH_WM, ttu_attr->qos_level_high_wm); + +/* TTU - per luma/chroma */ +/* Assumed surf0 is luma and 1 is chroma */ + +REG_SET_3 (DCN_SURF0_TTU_CNTL0 ,0 , +REFCYC_PER_REQ_DELIVERY , ttu_attr->refcyc_per_req_delivery_l, + QoS_LEVEL_FIXED, ttu_attr->qos_level_fixed_l, + QoS_RAMP_DISABLE, ttu_attr->qos_ramp_disable_l); + +REG_SET_3 (DCN_SURF1_TTU_CNTL0 ,0 , +REFCYC_PER_REQ_DELIVERY , ttu_attr->refcyc_per_req_delivery_c, + QoS_LEVEL_FIXED, ttu_attr->qos_level_fixed_c, + QoS_RAMP_DISABLE, ttu_attr->qos_ramp_disable_c); + +REG_SET_3 (DCN_CUR0_TTU_CNTL0 ,0 , +REFCYC_PER_REQ_DELIVERY , ttu_attr->refcyc_per_req_delivery_cur0, + QoS_LEVEL_FIXED, ttu_attr->qos_level_fixed_cur0, + QoS_RAMP_DISABLE, ttu_attr->qos_ramp_disable_cur0);REG_SET (FLIP_PARAMETERS_1 ,0 ,REFCYC_PER_PTE_GROUP_FLIP_L , dlg_attr->refcyc_per_pte_group_flip_l);
@@ -184,6 +262,39 @@ void hubp2_vready_at_or_After_vsync(struct hubp *hubp,REG_UPDATE (DCHUBP_CNTL ,HUBP_VREADY_AT_OR_AFTER_VSYNC , value); } +void hubp2_program_requestor ( +struct hubp *hubp, +struct _vcs_dpi_display_rq_regs_st *rq_regs) +{ +struct dcn20_hubp *hubp2 =TO_DCN20_HUBP (hubp); + +REG_UPDATE (HUBPRET_CONTROL , +DET_BUF_PLANE1_BASE_ADDRESS , rq_regs->plane1_base_address); +REG_SET_4 (DCN_EXPANSION_MODE ,0 , +DRQ_EXPANSION_MODE , rq_regs->drq_expansion_mode, +PRQ_EXPANSION_MODE , rq_regs->prq_expansion_mode, +MRQ_EXPANSION_MODE , rq_regs->mrq_expansion_mode, +CRQ_EXPANSION_MODE , rq_regs->crq_expansion_mode); +REG_SET_8 (DCHUBP_REQ_SIZE_CONFIG ,0 , +CHUNK_SIZE , rq_regs->rq_regs_l.chunk_size, +MIN_CHUNK_SIZE , rq_regs->rq_regs_l.min_chunk_size, +META_CHUNK_SIZE , rq_regs->rq_regs_l.meta_chunk_size, +MIN_META_CHUNK_SIZE , rq_regs->rq_regs_l.min_meta_chunk_size, +DPTE_GROUP_SIZE , rq_regs->rq_regs_l.dpte_group_size, +MPTE_GROUP_SIZE , rq_regs->rq_regs_l.mpte_group_size, +SWATH_HEIGHT , rq_regs->rq_regs_l.swath_height, +PTE_ROW_HEIGHT_LINEAR , rq_regs->rq_regs_l.pte_row_height_linear); +REG_SET_8 (DCHUBP_REQ_SIZE_CONFIG_C ,0 , +CHUNK_SIZE_C , rq_regs->rq_regs_c.chunk_size, +MIN_CHUNK_SIZE_C , rq_regs->rq_regs_c.min_chunk_size, +META_CHUNK_SIZE_C , rq_regs->rq_regs_c.meta_chunk_size, +MIN_META_CHUNK_SIZE_C , rq_regs->rq_regs_c.min_meta_chunk_size, +DPTE_GROUP_SIZE_C , rq_regs->rq_regs_c.dpte_group_size, +MPTE_GROUP_SIZE_C , rq_regs->rq_regs_c.mpte_group_size, +SWATH_HEIGHT_C , rq_regs->rq_regs_c.swath_height, +PTE_ROW_HEIGHT_LINEAR_C , rq_regs->rq_regs_c.pte_row_height_linear); +} + static void hubp2_setup( struct hubp *hubp, struct _vcs_dpi_display_dlg_regs_st *dlg_attr,
@@ -196,7 +307,7 @@ static void hubp2_setup(*/ hubp2_vready_at_or_After_vsync (hubp, pipe_dest); -hubp1_program_requestor (hubp, rq_regs); +hubp2_program_requestor (hubp, rq_regs);hubp2_program_deadline (hubp, dlg_attr, ttu_attr); }
@@ -283,6 +394,196 @@ static void hubp2_program_tiling( PIPE_ALIGNED, 0); } +void hubp2_program_size ( +struct hubp *hubp, +enum surface_pixel_format format, +const union plane_size *plane_size, +struct dc_plane_dcc_param *dcc) +{ +struct dcn20_hubp *hubp2 =TO_DCN20_HUBP (hubp); +uint32_t pitch, meta_pitch, pitch_c, meta_pitch_c; + +/* Program data and meta surface pitch (calculation from addrlib) +* 444 or 420 luma +*/ +if (format >=SURFACE_PIXEL_FORMAT_VIDEO_BEGIN && format <SURFACE_PIXEL_FORMAT_SUBSAMPLE_END ) { +ASSERT (plane_size->video.chroma_pitch !=0 ); +/* Chroma pitch zero can cause system hang! */ + + pitch = plane_size->video.luma_pitch -1 ; + meta_pitch = dcc->video.meta_pitch_l -1 ; + pitch_c = plane_size->video.chroma_pitch -1 ; + meta_pitch_c = dcc->video.meta_pitch_c -1 ; + }else { + pitch = plane_size->grph.surface_pitch -1 ; + meta_pitch = dcc->grph.meta_pitch -1 ; + pitch_c =0 ; + meta_pitch_c =0 ; + } + +if (!dcc->enable) { + meta_pitch =0 ; + meta_pitch_c =0 ; + } + +REG_UPDATE_2 (DCSURF_SURFACE_PITCH , +PITCH , pitch,META_PITCH , meta_pitch); + +if (format >=SURFACE_PIXEL_FORMAT_VIDEO_BEGIN ) +REG_UPDATE_2 (DCSURF_SURFACE_PITCH_C , +PITCH_C , pitch_c,META_PITCH_C , meta_pitch_c); +} + +void hubp2_program_rotation ( +struct hubp *hubp, +enum dc_rotation_angle rotation, +bool horizontal_mirror) +{ +struct dcn20_hubp *hubp2 =TO_DCN20_HUBP (hubp); +uint32_t mirror; + + +if (horizontal_mirror) + mirror =1 ; +else + mirror =0 ; + +/* Program rotation angle and horz mirror - no mirror */ +if (rotation ==ROTATION_ANGLE_0 ) +REG_UPDATE_2 (DCSURF_SURFACE_CONFIG , +ROTATION_ANGLE ,0 , +H_MIRROR_EN , mirror); +else if (rotation ==ROTATION_ANGLE_90 ) +REG_UPDATE_2 (DCSURF_SURFACE_CONFIG , +ROTATION_ANGLE ,1 , +H_MIRROR_EN , mirror); +else if (rotation ==ROTATION_ANGLE_180 ) +REG_UPDATE_2 (DCSURF_SURFACE_CONFIG , +ROTATION_ANGLE ,2 , +H_MIRROR_EN , mirror); +else if (rotation ==ROTATION_ANGLE_270 ) +REG_UPDATE_2 (DCSURF_SURFACE_CONFIG , +ROTATION_ANGLE ,3 , +H_MIRROR_EN , mirror); +} + +void hubp2_dcc_control (struct hubp *hubp,bool enable, +bool independent_64b_blks) +{ +uint32_t dcc_en = enable ?1 :0 ; +uint32_t dcc_ind_64b_blk = independent_64b_blks ?1 :0 ; +struct dcn20_hubp *hubp2 =TO_DCN20_HUBP (hubp); + +REG_UPDATE_4 (DCSURF_SURFACE_CONTROL , +PRIMARY_SURFACE_DCC_EN , dcc_en, +PRIMARY_SURFACE_DCC_IND_64B_BLK , dcc_ind_64b_blk, +SECONDARY_SURFACE_DCC_EN , dcc_en, +SECONDARY_SURFACE_DCC_IND_64B_BLK , dcc_ind_64b_blk); +} + +void hubp2_program_pixel_format ( +struct hubp *hubp, +enum surface_pixel_format format) +{ +struct dcn20_hubp *hubp2 =TO_DCN20_HUBP (hubp); +uint32_t red_bar =3 ; +uint32_t blue_bar =2 ; + +/* swap for ABGR format */ +if (format ==SURFACE_PIXEL_FORMAT_GRPH_ABGR8888 + || format ==SURFACE_PIXEL_FORMAT_GRPH_ABGR2101010 + || format ==SURFACE_PIXEL_FORMAT_GRPH_ABGR2101010_XR_BIAS + || format ==SURFACE_PIXEL_FORMAT_GRPH_ABGR16161616F ) { + red_bar =2 ; + blue_bar =3 ; + } + +REG_UPDATE_2 (HUBPRET_CONTROL , +CROSSBAR_SRC_CB_B , blue_bar, +CROSSBAR_SRC_CR_R , red_bar); + +/* Mapping is same as ipp programming (cnvc) */ + +switch (format) { +case SURFACE_PIXEL_FORMAT_GRPH_ARGB1555 : +REG_UPDATE (DCSURF_SURFACE_CONFIG , +SURFACE_PIXEL_FORMAT ,1 ); +break ; +case SURFACE_PIXEL_FORMAT_GRPH_RGB565 : +REG_UPDATE (DCSURF_SURFACE_CONFIG , +SURFACE_PIXEL_FORMAT ,3 ); +break ; +case SURFACE_PIXEL_FORMAT_GRPH_ARGB8888 : +case SURFACE_PIXEL_FORMAT_GRPH_ABGR8888 : +REG_UPDATE (DCSURF_SURFACE_CONFIG , +SURFACE_PIXEL_FORMAT ,8 ); +break ; +case SURFACE_PIXEL_FORMAT_GRPH_ARGB2101010 : +case SURFACE_PIXEL_FORMAT_GRPH_ABGR2101010 : +case SURFACE_PIXEL_FORMAT_GRPH_ABGR2101010_XR_BIAS : +REG_UPDATE (DCSURF_SURFACE_CONFIG , +SURFACE_PIXEL_FORMAT ,10 ); +break ; +case SURFACE_PIXEL_FORMAT_GRPH_ARGB16161616 : +REG_UPDATE (DCSURF_SURFACE_CONFIG , +SURFACE_PIXEL_FORMAT ,22 ); +break ; +case SURFACE_PIXEL_FORMAT_GRPH_ARGB16161616F : +case SURFACE_PIXEL_FORMAT_GRPH_ABGR16161616F :/*we use crossbar already*/ +REG_UPDATE (DCSURF_SURFACE_CONFIG , +SURFACE_PIXEL_FORMAT ,24 ); +break ; + +case SURFACE_PIXEL_FORMAT_VIDEO_420_YCbCr: +REG_UPDATE (DCSURF_SURFACE_CONFIG , +SURFACE_PIXEL_FORMAT ,65 ); +break ; +case SURFACE_PIXEL_FORMAT_VIDEO_420_YCrCb: +REG_UPDATE (DCSURF_SURFACE_CONFIG , +SURFACE_PIXEL_FORMAT ,64 ); +break ; +case SURFACE_PIXEL_FORMAT_VIDEO_420_10bpc_YCbCr: +REG_UPDATE (DCSURF_SURFACE_CONFIG , +SURFACE_PIXEL_FORMAT ,67 ); +break ; +case SURFACE_PIXEL_FORMAT_VIDEO_420_10bpc_YCrCb: +REG_UPDATE (DCSURF_SURFACE_CONFIG , +SURFACE_PIXEL_FORMAT ,66 ); +break ; +case SURFACE_PIXEL_FORMAT_VIDEO_AYCrCb8888: +REG_UPDATE (DCSURF_SURFACE_CONFIG , +SURFACE_PIXEL_FORMAT ,12 ); +break ; +#if defined(CONFIG_DRM_AMD_DC_DCN2_0 ) +case SURFACE_PIXEL_FORMAT_GRPH_RGB111110_FIX : +REG_UPDATE (DCSURF_SURFACE_CONFIG , +SURFACE_PIXEL_FORMAT ,112 ); +break ; +case SURFACE_PIXEL_FORMAT_GRPH_BGR101111_FIX : +REG_UPDATE (DCSURF_SURFACE_CONFIG , +SURFACE_PIXEL_FORMAT ,113 ); +break ; +case SURFACE_PIXEL_FORMAT_VIDEO_ACrYCb2101010: +REG_UPDATE (DCSURF_SURFACE_CONFIG , +SURFACE_PIXEL_FORMAT ,114 ); +break ; +case SURFACE_PIXEL_FORMAT_GRPH_RGB111110_FLOAT : +REG_UPDATE (DCSURF_SURFACE_CONFIG , +SURFACE_PIXEL_FORMAT ,118 ); +break ; +case SURFACE_PIXEL_FORMAT_GRPH_BGR101111_FLOAT : +REG_UPDATE (DCSURF_SURFACE_CONFIG , +SURFACE_PIXEL_FORMAT ,119 ); +break ; +#endif +default : +BREAK_TO_DEBUGGER (); +break ; + } + +/* don't see the need of program the xbar in DCN 1.0 */ +} + void hubp2_program_surface_config( struct hubp *hubp, enum surface_pixel_format format,
@@ -295,11 +596,11 @@ void hubp2_program_surface_config( {struct dcn20_hubp *hubp2 =TO_DCN20_HUBP (hubp); -hubp1_dcc_control (hubp, dcc->enable, dcc->grph.independent_64b_blks); +hubp2_dcc_control (hubp, dcc->enable, dcc->grph.independent_64b_blks);hubp2_program_tiling (hubp2, tiling_info, format); -hubp1_program_size (hubp, format, plane_size, dcc); -hubp1_program_rotation (hubp, rotation, horizontal_mirror); -hubp1_program_pixel_format (hubp, format); +hubp2_program_size (hubp, format, plane_size, dcc); +hubp2_program_rotation (hubp, rotation, horizontal_mirror); +hubp2_program_pixel_format (hubp, format); } enum cursor_lines_per_chunk hubp2_get_lines_per_chunk(
@@ -652,28 +953,381 @@ void hubp2_set_flip_control_surface_gsl(struct hubp *hubp, bool enable)REG_UPDATE (DCSURF_FLIP_CONTROL2 ,SURFACE_GSL_ENABLE , enable ?1 :0 ); } +bool hubp2_is_flip_pending (struct hubp *hubp) +{ +uint32_t flip_pending =0 ; +struct dcn20_hubp *hubp2 =TO_DCN20_HUBP (hubp); +struct dc_plane_address earliest_inuse_address; + +REG_GET (DCSURF_FLIP_CONTROL , +SURFACE_FLIP_PENDING , &flip_pending); + +REG_GET (DCSURF_SURFACE_EARLIEST_INUSE , +SURFACE_EARLIEST_INUSE_ADDRESS , &earliest_inuse_address.grph.addr.low_part); + +REG_GET (DCSURF_SURFACE_EARLIEST_INUSE_HIGH , +SURFACE_EARLIEST_INUSE_ADDRESS_HIGH , &earliest_inuse_address.grph.addr.high_part); + +if (flip_pending) +return true; + +if (earliest_inuse_address.grph.addr.quad_part != hubp->request_address.grph.addr.quad_part) +return true; + +return false; +} + +void hubp2_set_blank (struct hubp *hubp,bool blank) +{ +struct dcn20_hubp *hubp2 =TO_DCN20_HUBP (hubp); +uint32_t blank_en = blank ?1 :0 ; + +REG_UPDATE_2 (DCHUBP_CNTL , +HUBP_BLANK_EN , blank_en, +HUBP_TTU_DISABLE , blank_en); + +if (blank) { +uint32_t reg_val =REG_READ (DCHUBP_CNTL ); + +if (reg_val) { +/* init sequence workaround: in case HUBP is +* power gated, this wait would timeout. +* +* we just wrote reg_val to non-0, if it stay 0 +* it means HUBP is gated +*/ +REG_WAIT (DCHUBP_CNTL , +HUBP_NO_OUTSTANDING_REQ ,1 , +1 ,200 ); + } + + hubp->mpcc_id =0xf ; + hubp->opp_id =OPP_ID_INVALID ; + } +} + +void hubp2_cursor_set_position ( +struct hubp *hubp, +const struct dc_cursor_position *pos, +const struct dc_cursor_mi_param *param) +{ +struct dcn20_hubp *hubp2 =TO_DCN20_HUBP (hubp); +int src_x_offset = pos->x - pos->x_hotspot - param->viewport.x; +int src_y_offset = pos->y - pos->y_hotspot - param->viewport.y; +int x_hotspot = pos->x_hotspot; +int y_hotspot = pos->y_hotspot; +uint32_t dst_x_offset; +uint32_t cur_en = pos->enable ?1 :0 ; + +/* +* Guard aganst cursor_set_position() from being called with invalid +* attributes +* +* TODO: Look at combining cursor_set_position() and +* cursor_set_attributes() into cursor_update() +*/ +if (hubp->curs_attr.address.quad_part ==0 ) +return ; + +if (param->rotation ==ROTATION_ANGLE_90 || param->rotation ==ROTATION_ANGLE_270 ) { + src_x_offset = pos->y - pos->y_hotspot - param->viewport.x; + y_hotspot = pos->x_hotspot; + x_hotspot = pos->y_hotspot; + } + +if (param->mirror) { + x_hotspot = param->viewport.width - x_hotspot; + src_x_offset = param->viewport.x + param->viewport.width - src_x_offset; + } + + dst_x_offset = (src_x_offset >=0 ) ? src_x_offset :0 ; + dst_x_offset *= param->ref_clk_khz; + dst_x_offset /= param->pixel_clk_khz; + +ASSERT (param->h_scale_ratio.value); + +if (param->h_scale_ratio.value) + dst_x_offset =dc_fixpt_floor (dc_fixpt_div ( +dc_fixpt_from_int (dst_x_offset), + param->h_scale_ratio)); + +if (src_x_offset >= (int )param->viewport.width) + cur_en =0 ;/* not visible beyond right edge*/ + +if (src_x_offset + (int )hubp->curs_attr.width <=0 ) + cur_en =0 ;/* not visible beyond left edge*/ + +if (src_y_offset >= (int )param->viewport.height) + cur_en =0 ;/* not visible beyond bottom edge*/ + +if (src_y_offset + (int )hubp->curs_attr.height <=0 ) + cur_en =0 ;/* not visible beyond top edge*/ + +if (cur_en &®_READ (CURSOR_SURFACE_ADDRESS ) ==0 ) + hubp->funcs->set_cursor_attributes (hubp, &hubp->curs_attr); + +REG_UPDATE (CURSOR_CONTROL , +CURSOR_ENABLE , cur_en); + +REG_SET_2 (CURSOR_POSITION ,0 , +CURSOR_X_POSITION , pos->x, +CURSOR_Y_POSITION , pos->y); + +REG_SET_2 (CURSOR_HOT_SPOT ,0 , +CURSOR_HOT_SPOT_X , x_hotspot, +CURSOR_HOT_SPOT_Y , y_hotspot); + +REG_SET (CURSOR_DST_OFFSET ,0 , +CURSOR_DST_X_OFFSET , dst_x_offset); +/* TODO Handle surface pixel formats other than 4:4:4 */ +} + +void hubp2_clk_cntl (struct hubp *hubp,bool enable) +{ +struct dcn20_hubp *hubp2 =TO_DCN20_HUBP (hubp); +uint32_t clk_enable = enable ?1 :0 ; + +REG_UPDATE (HUBP_CLK_CNTL ,HUBP_CLOCK_ENABLE , clk_enable); +} + +void hubp2_vtg_sel (struct hubp *hubp,uint32_t otg_inst) +{ +struct dcn20_hubp *hubp2 =TO_DCN20_HUBP (hubp); + +REG_UPDATE (DCHUBP_CNTL ,HUBP_VTG_SEL , otg_inst); +} + +void hubp2_clear_underflow (struct hubp *hubp) +{ +struct dcn20_hubp *hubp2 =TO_DCN20_HUBP (hubp); + +REG_UPDATE (DCHUBP_CNTL ,HUBP_UNDERFLOW_CLEAR ,1 ); +} + +void hubp2_read_state_common (struct hubp *hubp) +{ +struct dcn20_hubp *hubp2 =TO_DCN20_HUBP (hubp); +struct dcn_hubp_state *s = &hubp2->state; +struct _vcs_dpi_display_dlg_regs_st *dlg_attr = &s->dlg_attr; +struct _vcs_dpi_display_ttu_regs_st *ttu_attr = &s->ttu_attr; +struct _vcs_dpi_display_rq_regs_st *rq_regs = &s->rq_regs; + +/* Requester */ +REG_GET (HUBPRET_CONTROL , +DET_BUF_PLANE1_BASE_ADDRESS , &rq_regs->plane1_base_address); +REG_GET_4 (DCN_EXPANSION_MODE , +DRQ_EXPANSION_MODE , &rq_regs->drq_expansion_mode, +PRQ_EXPANSION_MODE , &rq_regs->prq_expansion_mode, +MRQ_EXPANSION_MODE , &rq_regs->mrq_expansion_mode, +CRQ_EXPANSION_MODE , &rq_regs->crq_expansion_mode); + +/* DLG - Per hubp */ +REG_GET_2 (BLANK_OFFSET_0 , +REFCYC_H_BLANK_END , &dlg_attr->refcyc_h_blank_end, +DLG_V_BLANK_END , &dlg_attr->dlg_vblank_end); + +REG_GET (BLANK_OFFSET_1 , +MIN_DST_Y_NEXT_START , &dlg_attr->min_dst_y_next_start); + +REG_GET (DST_DIMENSIONS , +REFCYC_PER_HTOTAL , &dlg_attr->refcyc_per_htotal); + +REG_GET_2 (DST_AFTER_SCALER , +REFCYC_X_AFTER_SCALER , &dlg_attr->refcyc_x_after_scaler, +DST_Y_AFTER_SCALER , &dlg_attr->dst_y_after_scaler); + +if (REG (PREFETCH_SETTINS )) +REG_GET_2 (PREFETCH_SETTINS , +DST_Y_PREFETCH , &dlg_attr->dst_y_prefetch, +VRATIO_PREFETCH , &dlg_attr->vratio_prefetch); +else +REG_GET_2 (PREFETCH_SETTINGS , +DST_Y_PREFETCH , &dlg_attr->dst_y_prefetch, +VRATIO_PREFETCH , &dlg_attr->vratio_prefetch); + +REG_GET_2 (VBLANK_PARAMETERS_0 , +DST_Y_PER_VM_VBLANK , &dlg_attr->dst_y_per_vm_vblank, +DST_Y_PER_ROW_VBLANK , &dlg_attr->dst_y_per_row_vblank); + +REG_GET (REF_FREQ_TO_PIX_FREQ , +REF_FREQ_TO_PIX_FREQ , &dlg_attr->ref_freq_to_pix_freq); + +/* DLG - Per luma/chroma */ +REG_GET (VBLANK_PARAMETERS_1 , +REFCYC_PER_PTE_GROUP_VBLANK_L , &dlg_attr->refcyc_per_pte_group_vblank_l); + +REG_GET (VBLANK_PARAMETERS_3 , +REFCYC_PER_META_CHUNK_VBLANK_L , &dlg_attr->refcyc_per_meta_chunk_vblank_l); + +if (REG (NOM_PARAMETERS_0 )) +REG_GET (NOM_PARAMETERS_0 , +DST_Y_PER_PTE_ROW_NOM_L , &dlg_attr->dst_y_per_pte_row_nom_l); + +if (REG (NOM_PARAMETERS_1 )) +REG_GET (NOM_PARAMETERS_1 , +REFCYC_PER_PTE_GROUP_NOM_L , &dlg_attr->refcyc_per_pte_group_nom_l); + +REG_GET (NOM_PARAMETERS_4 , +DST_Y_PER_META_ROW_NOM_L , &dlg_attr->dst_y_per_meta_row_nom_l); + +REG_GET (NOM_PARAMETERS_5 , +REFCYC_PER_META_CHUNK_NOM_L , &dlg_attr->refcyc_per_meta_chunk_nom_l); + +REG_GET_2 (PER_LINE_DELIVERY_PRE , +REFCYC_PER_LINE_DELIVERY_PRE_L , &dlg_attr->refcyc_per_line_delivery_pre_l, +REFCYC_PER_LINE_DELIVERY_PRE_C , &dlg_attr->refcyc_per_line_delivery_pre_c); + +REG_GET_2 (PER_LINE_DELIVERY , +REFCYC_PER_LINE_DELIVERY_L , &dlg_attr->refcyc_per_line_delivery_l, +REFCYC_PER_LINE_DELIVERY_C , &dlg_attr->refcyc_per_line_delivery_c); + +if (REG (PREFETCH_SETTINS_C )) +REG_GET (PREFETCH_SETTINS_C , +VRATIO_PREFETCH_C , &dlg_attr->vratio_prefetch_c); +else +REG_GET (PREFETCH_SETTINGS_C , +VRATIO_PREFETCH_C , &dlg_attr->vratio_prefetch_c); + +REG_GET (VBLANK_PARAMETERS_2 , +REFCYC_PER_PTE_GROUP_VBLANK_C , &dlg_attr->refcyc_per_pte_group_vblank_c); + +REG_GET (VBLANK_PARAMETERS_4 , +REFCYC_PER_META_CHUNK_VBLANK_C , &dlg_attr->refcyc_per_meta_chunk_vblank_c); + +if (REG (NOM_PARAMETERS_2 )) +REG_GET (NOM_PARAMETERS_2 , +DST_Y_PER_PTE_ROW_NOM_C , &dlg_attr->dst_y_per_pte_row_nom_c); + +if (REG (NOM_PARAMETERS_3 )) +REG_GET (NOM_PARAMETERS_3 , +REFCYC_PER_PTE_GROUP_NOM_C , &dlg_attr->refcyc_per_pte_group_nom_c); + +REG_GET (NOM_PARAMETERS_6 , +DST_Y_PER_META_ROW_NOM_C , &dlg_attr->dst_y_per_meta_row_nom_c); + +REG_GET (NOM_PARAMETERS_7 , +REFCYC_PER_META_CHUNK_NOM_C , &dlg_attr->refcyc_per_meta_chunk_nom_c); + +/* TTU - per hubp */ +REG_GET_2 (DCN_TTU_QOS_WM , + QoS_LEVEL_LOW_WM, &ttu_attr->qos_level_low_wm, + QoS_LEVEL_HIGH_WM, &ttu_attr->qos_level_high_wm); + +REG_GET_2 (DCN_GLOBAL_TTU_CNTL , +MIN_TTU_VBLANK , &ttu_attr->min_ttu_vblank, + QoS_LEVEL_FLIP, &ttu_attr->qos_level_flip); + +/* TTU - per luma/chroma */ +/* Assumed surf0 is luma and 1 is chroma */ + +REG_GET_3 (DCN_SURF0_TTU_CNTL0 , +REFCYC_PER_REQ_DELIVERY , &ttu_attr->refcyc_per_req_delivery_l, + QoS_LEVEL_FIXED, &ttu_attr->qos_level_fixed_l, + QoS_RAMP_DISABLE, &ttu_attr->qos_ramp_disable_l); + +REG_GET (DCN_SURF0_TTU_CNTL1 , +REFCYC_PER_REQ_DELIVERY_PRE , + &ttu_attr->refcyc_per_req_delivery_pre_l); + +REG_GET_3 (DCN_SURF1_TTU_CNTL0 , +REFCYC_PER_REQ_DELIVERY , &ttu_attr->refcyc_per_req_delivery_c, + QoS_LEVEL_FIXED, &ttu_attr->qos_level_fixed_c, + QoS_RAMP_DISABLE, &ttu_attr->qos_ramp_disable_c); + +REG_GET (DCN_SURF1_TTU_CNTL1 , +REFCYC_PER_REQ_DELIVERY_PRE , + &ttu_attr->refcyc_per_req_delivery_pre_c); + +/* Rest of hubp */ +REG_GET (DCSURF_SURFACE_CONFIG , +SURFACE_PIXEL_FORMAT , &s->pixel_format); + +REG_GET (DCSURF_SURFACE_EARLIEST_INUSE_HIGH , +SURFACE_EARLIEST_INUSE_ADDRESS_HIGH , &s->inuse_addr_hi); + +REG_GET (DCSURF_SURFACE_EARLIEST_INUSE , +SURFACE_EARLIEST_INUSE_ADDRESS , &s->inuse_addr_lo); + +REG_GET_2 (DCSURF_PRI_VIEWPORT_DIMENSION , +PRI_VIEWPORT_WIDTH , &s->viewport_width, +PRI_VIEWPORT_HEIGHT , &s->viewport_height); + +REG_GET_2 (DCSURF_SURFACE_CONFIG , +ROTATION_ANGLE , &s->rotation_angle, +H_MIRROR_EN , &s->h_mirror_en); + +REG_GET (DCSURF_TILING_CONFIG , +SW_MODE , &s->sw_mode); + +REG_GET (DCSURF_SURFACE_CONTROL , +PRIMARY_SURFACE_DCC_EN , &s->dcc_en); + +REG_GET_3 (DCHUBP_CNTL , +HUBP_BLANK_EN , &s->blank_en, +HUBP_TTU_DISABLE , &s->ttu_disable, +HUBP_UNDERFLOW_STATUS , &s->underflow_status); + +REG_GET (DCN_GLOBAL_TTU_CNTL , +MIN_TTU_VBLANK , &s->min_ttu_vblank); + +REG_GET_2 (DCN_TTU_QOS_WM , + QoS_LEVEL_LOW_WM, &s->qos_level_low_wm, + QoS_LEVEL_HIGH_WM, &s->qos_level_high_wm); + +} + +void hubp2_read_state (struct hubp *hubp) +{ +struct dcn20_hubp *hubp2 =TO_DCN20_HUBP (hubp); +struct dcn_hubp_state *s = &hubp2->state; +struct _vcs_dpi_display_rq_regs_st *rq_regs = &s->rq_regs; + +hubp2_read_state_common (hubp); + +REG_GET_8 (DCHUBP_REQ_SIZE_CONFIG , +CHUNK_SIZE , &rq_regs->rq_regs_l.chunk_size, +MIN_CHUNK_SIZE , &rq_regs->rq_regs_l.min_chunk_size, +META_CHUNK_SIZE , &rq_regs->rq_regs_l.meta_chunk_size, +MIN_META_CHUNK_SIZE , &rq_regs->rq_regs_l.min_meta_chunk_size, +DPTE_GROUP_SIZE , &rq_regs->rq_regs_l.dpte_group_size, +MPTE_GROUP_SIZE , &rq_regs->rq_regs_l.mpte_group_size, +SWATH_HEIGHT , &rq_regs->rq_regs_l.swath_height, +PTE_ROW_HEIGHT_LINEAR , &rq_regs->rq_regs_l.pte_row_height_linear); + +REG_GET_8 (DCHUBP_REQ_SIZE_CONFIG_C , +CHUNK_SIZE_C , &rq_regs->rq_regs_c.chunk_size, +MIN_CHUNK_SIZE_C , &rq_regs->rq_regs_c.min_chunk_size, +META_CHUNK_SIZE_C , &rq_regs->rq_regs_c.meta_chunk_size, +MIN_META_CHUNK_SIZE_C , &rq_regs->rq_regs_c.min_meta_chunk_size, +DPTE_GROUP_SIZE_C , &rq_regs->rq_regs_c.dpte_group_size, +MPTE_GROUP_SIZE_C , &rq_regs->rq_regs_c.mpte_group_size, +SWATH_HEIGHT_C , &rq_regs->rq_regs_c.swath_height, +PTE_ROW_HEIGHT_LINEAR_C , &rq_regs->rq_regs_c.pte_row_height_linear); + +} +static struct hubp_funcs dcn20_hubp_funcs = { .hubp_enable_tripleBuffer = hubp2_enable_triplebuffer, .hubp_is_triplebuffer_enabled = hubp2_is_triplebuffer_enabled, .hubp_program_surface_flip_and_addr = hubp2_program_surface_flip_and_addr, .hubp_program_surface_config = hubp2_program_surface_config, - .hubp_is_flip_pending = hubp1_is_flip_pending, + .hubp_is_flip_pending = hubp2_is_flip_pending, .hubp_setup = hubp2_setup, .hubp_setup_interdependent = hubp2_setup_interdependent, .hubp_set_vm_system_aperture_settings = hubp2_set_vm_system_aperture_settings, - .set_blank = hubp1_set_blank, - .dcc_control = hubp1_dcc_control, + .set_blank = hubp2_set_blank, + .dcc_control = hubp2_dcc_control, .hubp_update_dchub = hubp2_update_dchub, .mem_program_viewport = min_set_viewport, .set_cursor_attributes = hubp2_cursor_set_attributes, - .set_cursor_position = hubp1_cursor_set_position, - .hubp_clk_cntl = hubp1_clk_cntl, - .hubp_vtg_sel = hubp1_vtg_sel, + .set_cursor_position = hubp2_cursor_set_position, + .hubp_clk_cntl = hubp2_clk_cntl, + .hubp_vtg_sel = hubp2_vtg_sel, .dmdata_set_attributes = hubp2_dmdata_set_attributes, .dmdata_load = hubp2_dmdata_load, .dmdata_status_done = hubp2_dmdata_status_done, - .hubp_read_state = hubp1_read_state, - .hubp_clear_underflow = hubp1_clear_underflow, + .hubp_read_state = hubp2_read_state, + .hubp_clear_underflow = hubp2_clear_underflow, .hubp_set_flip_control_surface_gsl = hubp2_set_flip_control_surface_gsl, .hubp_init = hubp1_init, };
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hubp.h +37 -0
@@ -267,6 +267,24 @@ bool hubp2_program_surface_flip_and_addr(const struct dc_plane_address *address, bool flip_immediate); +void hubp2_dcc_control (struct hubp *hubp,bool enable, +bool independent_64b_blks); + +void hubp2_program_size ( +struct hubp *hubp, +enum surface_pixel_format format, +const union plane_size *plane_size, +struct dc_plane_dcc_param *dcc); + +void hubp2_program_rotation ( +struct hubp *hubp, +enum dc_rotation_angle rotation, +bool horizontal_mirror); + +void hubp2_program_pixel_format ( +struct hubp *hubp, +enum surface_pixel_format format); + void hubp2_program_surface_config( struct hubp *hubp, enum surface_pixel_format format,
@@ -277,6 +295,25 @@ void hubp2_program_surface_config(bool horizontal_mirror, unsigned int compat_level); +bool hubp2_is_flip_pending (struct hubp *hubp); + +void hubp2_set_blank (struct hubp *hubp,bool blank); + +void hubp2_cursor_set_position ( +struct hubp *hubp, +const struct dc_cursor_position *pos, +const struct dc_cursor_mi_param *param); + +void hubp2_clk_cntl (struct hubp *hubp,bool enable); + +void hubp2_vtg_sel (struct hubp *hubp,uint32_t otg_inst); + +void hubp2_clear_underflow (struct hubp *hubp); + +void hubp2_read_state_common (struct hubp *hubp); + +void hubp2_read_state (struct hubp *hubp); +#endif /* __DC_MEM_INPUT_DCN20_H__ */
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 35/87 drm/amd/display: Read max down spread
In reply to a message not in the archiveFrom: Derek Lai <[email protected]>
[Why]
When launch D10.2, driver will write DPCD 0x107 with 0x00
[How]
Read MAX_DOWNSPREAD (0x0003h) then keep in current link settings
--- drivers/gpu/drm/amd/display/dc/core/dc_link.c | 7 +++++++ 1 file changed, 7 insertions(+)
drivers/gpu/drm/amd/display/dc/core/dc_link.c +7 -0
@@ -530,6 +530,7 @@ static void read_edp_current_link_settings_on_detect(struct dc_link *link)uint32_t read_dpcd_retry_cnt =10 ;enum dc_status status =DC_ERROR_UNEXPECTED ;int i; +union max_down_spread max_down_spread = { {0 } };// Read DPCD 00101h to find out the number of lanes currently set for (i =0 ; i < read_dpcd_retry_cnt; i++) {
@@ -574,6 +575,12 @@ static void read_edp_current_link_settings_on_detect(struct dc_link *link) link->cur_link_settings.link_rate = link_bw_set; link->cur_link_settings.use_link_rate_set = false; } +// Read DPCD 00003h to find the max down spread. +core_link_read_dpcd (link,DP_MAX_DOWNSPREAD , + &max_down_spread.raw,sizeof (max_down_spread)); + link->cur_link_settings.link_spread = + max_down_spread.bits.MAX_DOWN_SPREAD ? +LINK_SPREAD_05_DOWNSPREAD_30KHZ :LINK_SPREAD_DISABLED ; } static bool detect_dp(
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 36/87 drm/amd/display: Remove dsc disable_ich flag programming.
In reply to a message not in the archiveFrom: Dmytro Laktyushkin <[email protected]>
Current default is sufficient for a flag that does not change.
--- drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dsc.c | 8 +++++--- drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dsc.h | 4 ++-- 2 files changed, 7 insertions(+), 5 deletions(-)
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dsc.c +5 -3
@@ -436,7 +436,7 @@ static void dsc_init_reg_values(struct dsc_reg_values *reg_vals) reg_vals->ich_reset_at_eol =0 ; reg_vals->alternate_ich_encoding_en =0 ; reg_vals->rc_buffer_model_size =0 ; - reg_vals->disable_ich =0 ; +/*reg_vals->disable_ich = 0;*/ reg_vals->dsc_dbg_en =0 ;for (i =0 ; i <4 ; i++)
@@ -518,9 +518,11 @@ static void dsc_write_to_registers(struct display_stream_compressor *dsc, constALTERNATE_ICH_ENCODING_EN , reg_vals->alternate_ich_encoding_en, NUMBER_OF_SLICES_IN_VERTICAL_DIRECTION, reg_vals->num_slices_v - 1); -REG_SET_2 (DSCC_CONFIG1 ,0 , +REG_SET (DSCC_CONFIG1 ,0 , +DSCC_RATE_CONTROL_BUFFER_MODEL_SIZE , reg_vals->rc_buffer_model_size); +/*REG_SET_2(DSCC_CONFIG1, 0, DSCC_RATE_CONTROL_BUFFER_MODEL_SIZE, reg_vals->rc_buffer_model_size, -DSCC_DISABLE_ICH , reg_vals->disable_ich); +DSCC_DISABLE_ICH, reg_vals->disable_ich);*/ REG_SET_4(DSCC_INTERRUPT_CONTROL_STATUS, 0, DSCC_RATE_CONTROL_BUFFER_MODEL0_OVERFLOW_OCCURRED_INT_EN, reg_vals->rc_buffer_model_overflow_int_en[0],
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dsc.h +2 -2
@@ -103,7 +103,7 @@DSC_SF (DSCC0_DSCC_CONFIG0 ,ALTERNATE_ICH_ENCODING_EN , mask_sh), \DSC_SF (DSCC0_DSCC_CONFIG0 ,NUMBER_OF_SLICES_IN_VERTICAL_DIRECTION , mask_sh), \DSC_SF (DSCC0_DSCC_CONFIG1 ,DSCC_RATE_CONTROL_BUFFER_MODEL_SIZE , mask_sh), \ -DSC_SF (DSCC0_DSCC_CONFIG1 ,DSCC_DISABLE_ICH , mask_sh), \ +/*DSC_SF(DSCC0_DSCC_CONFIG1, DSCC_DISABLE_ICH, mask_sh),*/ \DSC_SF (DSCC0_DSCC_STATUS ,DSCC_DOUBLE_BUFFER_REG_UPDATE_PENDING , mask_sh), \DSC_SF (DSCC0_DSCC_INTERRUPT_CONTROL_STATUS ,DSCC_RATE_BUFFER0_OVERFLOW_OCCURRED , mask_sh), \ DSC_SF(DSCC0_DSCC_INTERRUPT_CONTROL_STATUS, DSCC_RATE_BUFFER1_OVERFLOW_OCCURRED, mask_sh), \
@@ -278,7 +278,7 @@type ALTERNATE_ICH_ENCODING_EN ; \type NUMBER_OF_SLICES_IN_VERTICAL_DIRECTION ; \type DSCC_RATE_CONTROL_BUFFER_MODEL_SIZE ; \ -type DSCC_DISABLE_ICH ; \ +/*type DSCC_DISABLE_ICH;*/ \type DSCC_DOUBLE_BUFFER_REG_UPDATE_PENDING ; \type DSCC_RATE_BUFFER0_OVERFLOW_OCCURRED ; \type DSCC_RATE_BUFFER1_OVERFLOW_OCCURRED ; \
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 37/87 drm/amd/display: Power-gate all DSCs at driver init time
In reply to a message not in the archiveFrom: Nikola Cornij <[email protected]>
[why]
DSC should be powered-on only on as-needed basis, i.e. if the mode requires it
[how]
Loop over all the DSCs at driver init time and power-gate each
--- drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c | 4 ++++ 1 file changed, 4 insertions(+)
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c +4 -0
@@ -599,6 +599,10 @@ static void dcn20_init_hw(struct dc *dc) } } +/* Power gate DSCs */ +for (i =0 ; i < res_pool->res_cap->num_dsc; i++) +dcn20_dsc_pg_control (hws, res_pool->dscs[i]->inst, false); +/* Blank pixel data with OPP DPG */ for (i =0 ; i < dc->res_pool->timing_generator_count; i++) {struct timing_generator *tg = dc->res_pool->timing_generators[i];
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 38/87 drm/amd/display: Disable Audio on reinitialize hardware
In reply to a message not in the archiveFrom: Alvin Lee <[email protected]>
[Why]
When we recover from hang, we do not want to skip the audio enable call.
[How]
Disable audio in dc_reinitialize_hardware
--- .../display/dc/clk_mgr/dcn20/dcn20_clk_mgr.c | 1 + .../display/dc/dce110/dce110_hw_sequencer.c | 20 +++++++++++++++---- 2 files changed, 17 insertions(+), 4 deletions(-)
drivers/gpu/drm/amd/display/dc/clk_mgr/dcn20/dcn20_clk_mgr.c +1 -0
@@ -331,6 +331,7 @@ void dcn20_clk_mgr_construct( struct dccg *dccg) { clk_mgr->base.ctx = ctx; + clk_mgr->pp_smu = pp_smu; clk_mgr->base.funcs = &dcn2_funcs; clk_mgr->regs = &clk_mgr_regs; clk_mgr->clk_mgr_shift = &clk_mgr_shift;
drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c +16 -4
@@ -962,11 +962,17 @@ void hwss_edp_backlight_control(void dce110_enable_audio_stream (struct pipe_ctx *pipe_ctx) {/* notify audio driver for audio modes of monitor */ -struct dc *core_dc = pipe_ctx->stream->ctx->dc; +struct dc *core_dc;struct pp_smu_funcs *pp_smu =NULL ; -struct clk_mgr *clk_mgr = core_dc->clk_mgr; +struct clk_mgr *clk_mgr;unsigned int i, num_audio =1 ; +if (!pipe_ctx->stream) +return ; + + core_dc = pipe_ctx->stream->ctx->dc; + clk_mgr = core_dc->clk_mgr; +if (pipe_ctx->stream_res.audio && pipe_ctx->stream_res.audio->enabled == true)return ;
@@ -996,9 +1002,15 @@ void dce110_enable_audio_stream(struct pipe_ctx *pipe_ctx)void dce110_disable_audio_stream (struct pipe_ctx *pipe_ctx,int option) { -struct dc *dc = pipe_ctx->stream->ctx->dc; +struct dc *dc;struct pp_smu_funcs *pp_smu =NULL ; -struct clk_mgr *clk_mgr = dc->clk_mgr; +struct clk_mgr *clk_mgr; + +if (!pipe_ctx || !pipe_ctx->stream) +return ; + + dc = pipe_ctx->stream->ctx->dc; + clk_mgr = dc->clk_mgr;if (pipe_ctx->stream_res.audio && pipe_ctx->stream_res.audio->enabled == false)return ;
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 39/87 drm/amd/display: fix DMCU hang when going into Modern Standby
In reply to a message not in the archiveFrom: Zi Yu Liao <[email protected]>
[why]
When the system is going into suspend, set_backlight gets called after the eDP got blanked. Since smooth brightness is enabled, the driver will make a call into the DMCU to ramp the brightness. The DMCU would try to enable ABM to do so. But since the display is blanked, this ends up causing ABM1_ACE_DBUF_REG_UPDATE_PENDING to get stuck at 1, which results in a dead lock in the DMCU firmware.
[how]
Disable brightness ramping when the eDP display is blanked.
--- drivers/gpu/drm/amd/display/dc/core/dc_link.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/core/dc_link.c +8 -1
@@ -2350,7 +2350,7 @@ bool dc_link_set_backlight_level(const struct dc_link *link,if (core_dc->current_state->res_ctx.pipe_ctx[i].stream) {if (core_dc->current_state->res_ctx. pipe_ctx[i].stream->link - == link) + == link) {/* DMCU -1 for all controller id values, * therefore +1 here */
@@ -2358,6 +2358,13 @@ bool dc_link_set_backlight_level(const struct dc_link *link, core_dc->current_state-> res_ctx.pipe_ctx[i].stream_res.tg->inst +1 ; + +/* Disable brightness ramping when the display is blanked +* as it can hang the DMCU +*/ +if (core_dc->current_state->res_ctx.pipe_ctx[i].plane_state ==NULL ) + frame_ramp =0 ; + } } } abm->funcs->set_backlight_level_pwm(
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 40/87 drm/amd/display: Do not fill Null packet in the blank period
In reply to a message not in the archiveFrom: Charlene Liu <[email protected]>
[Description]
Do not fill Null packet in the blank period for new packet gen This is based on HW IP team recommended default setting change.
--- drivers/gpu/drm/amd/display/dc/dcn10/dcn10_stream_encoder.c | 3 ++- drivers/gpu/drm/amd/display/dc/dcn10/dcn10_stream_encoder.h | 2 ++ 2 files changed, 4 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_stream_encoder.c +2 -1
@@ -511,11 +511,12 @@ void enc1_stream_encoder_hdmi_set_stream_attribute(enc1_stream_encoder_set_stream_attribute_helper (enc1, crtc_timing);/* setup HDMI engine */ -REG_UPDATE_5 (HDMI_CONTROL , +REG_UPDATE_6 (HDMI_CONTROL ,HDMI_PACKET_GEN_VERSION ,1 ,HDMI_KEEPOUT_MODE ,1 ,HDMI_DEEP_COLOR_ENABLE ,0 ,HDMI_DATA_SCRAMBLE_EN ,0 , +HDMI_NO_EXTRA_NULL_PACKET_FILLED ,1 ,HDMI_CLOCK_CHANNEL_RATE ,0 );
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_stream_encoder.h +2 -0
@@ -189,6 +189,7 @@ struct dcn10_stream_enc_registers {SE_SF (DIG0_HDMI_CONTROL ,HDMI_DEEP_COLOR_ENABLE , mask_sh),\SE_SF (DIG0_HDMI_CONTROL ,HDMI_DEEP_COLOR_DEPTH , mask_sh),\SE_SF (DIG0_HDMI_CONTROL ,HDMI_DATA_SCRAMBLE_EN , mask_sh),\ +SE_SF (DIG0_HDMI_CONTROL ,HDMI_NO_EXTRA_NULL_PACKET_FILLED , mask_sh),\SE_SF (DIG0_HDMI_VBI_PACKET_CONTROL ,HDMI_GC_CONT , mask_sh),\SE_SF (DIG0_HDMI_VBI_PACKET_CONTROL ,HDMI_GC_SEND , mask_sh),\ SE_SF(DIG0_HDMI_VBI_PACKET_CONTROL, HDMI_NULL_SEND, mask_sh),\
@@ -374,6 +375,7 @@ struct dcn10_stream_enc_registers {type HDMI_GC_SEND ;\type HDMI_NULL_SEND ;\type HDMI_DATA_SCRAMBLE_EN ;\ +type HDMI_NO_EXTRA_NULL_PACKET_FILLED ;\type HDMI_AUDIO_INFO_SEND ;\type AFMT_AUDIO_INFO_UPDATE ;\type HDMI_AUDIO_INFO_LINE ;\
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 41/87 drm/amd/display: Remove unnecessary NULL check in set_preferred_link_settings
In reply to a message not in the archiveFrom: Harry Wentland <[email protected]>
[Why]link_stream is never NULL here as we've dereferenced it a couple lines before and have done so for a couple months now.
[How]
- Drop the NULL check.
- Initialize where we know link_stream is non-NULL
--- drivers/gpu/drm/amd/display/dc/core/dc_link.c | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc_link.c +4 -5
@@ -3021,8 +3021,10 @@ void dc_link_set_preferred_link_settings(struct dc *dc,for (i =0 ; i <MAX_PIPES ; i++) { pipe = &dc->current_state->res_ctx.pipe_ctx[i];if (pipe->stream && pipe->stream->link) { -if (pipe->stream->link == link) +if (pipe->stream->link == link) { + link_stream = pipe->stream;break ; + } } }
@@ -3030,14 +3032,11 @@ void dc_link_set_preferred_link_settings(struct dc *dc,if (i ==MAX_PIPES )return ; - link_stream = link->dc->current_state->res_ctx.pipe_ctx[i].stream; -/* Cannot retrain link if backend is off */ if (link_stream->dpms_off)return ; -if (link_stream) -decide_link_settings (link_stream, &store_settings); +decide_link_settings (link_stream, &store_settings);if ((store_settings.lane_count !=LANE_COUNT_UNKNOWN ) && (store_settings.link_rate !=LINK_RATE_UNKNOWN ))
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 42/87 drm/amd/display: use encoder's engine id to find matched free audio device
In reply to a message not in the archiveFrom: Tai Man <[email protected]>
[Why]
On some platforms, the encoder id 3 is not populated. So the encoders are not stored in right order as index (id: 0, 1, 2, 4, 5) at pool. This would cause encoders id 4 & id 5 to fail when finding corresponding audio device, defaulting to the first available audio device. As result, we cannot stream audio into two DP ports with encoders id 4 & id 5.
[How]
It need to create enough audio device objects (0 - 5) to perform matching. Then use encoder engine id to find matched audio device.
--- drivers/gpu/drm/amd/display/dc/core/dc_resource.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/core/dc_resource.c +7 -1
@@ -265,7 +265,7 @@ bool resource_construct(* PORT_CONNECTIVITY == 1 (as instructed by HW team). */ update_num_audio (&straps, &num_audio, &pool->audio_support); -for (i =0 ; i < pool->pipe_count && i < num_audio; i++) { +for (i =0 ; i < caps->num_audio; i++) {struct audio *aud = create_funcs->create_audio (ctx, i);if (aud ==NULL ) {
@@ -1676,6 +1676,12 @@ static struct audio *find_first_free_audio(return pool->audios[i]; } } + +/* use engine id to find free audio */ +if ((id < pool->audio_count) && (res_ctx->is_audio_acquired[id] == false)) { +return pool->audios[id]; + } +/*not found the matching one, first come first serve*/ for (i =0 ; i < pool->audio_count; i++) {if (res_ctx->is_audio_acquired[i] == false) {
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 43/87 drm/amd/display: fix not calling ppsmu to trigger PME
In reply to a message not in the archiveFrom: Su Sung Chung <[email protected]>
[why]dcn20_clk_mgr_construct was not initializing pp_smu, and PME call gets filtered out by the null check
[how]
initialize pp_smu dcn20_clk_mgr_construct
--- drivers/gpu/drm/amd/display/dc/clk_mgr/dcn20/dcn20_clk_mgr.c | 2 ++ 1 file changed, 2 insertions(+)
drivers/gpu/drm/amd/display/dc/clk_mgr/dcn20/dcn20_clk_mgr.c +2 -0
@@ -346,6 +346,8 @@ void dcn20_clk_mgr_construct( clk_mgr->base.dprefclk_khz =700000 ;// 700 MHz planned if VCO is 3.85 GHz, will be retrieved + clk_mgr->pp_smu = pp_smu; +if (IS_FPGA_MAXIMUS_DC (ctx->dce_environment)) { dcn2_funcs.update_clocks = dcn2_update_clocks_fpga; clk_mgr->dentist_vco_freq_khz =3850000 ;
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 44/87 drm/amd/display: Change min_h_sync_width from 8 to 4
In reply to a message not in the archiveFrom: Fatemeh Darbehani <[email protected]>
[Why]
Some display's hsync width is lower than the minimum dcn20 is set to support right now. This will cause optc1_validate_timing to fail which eventually will result in wrong set mode. This was set to 8 as per HW team's request for no valid reason.
[How]
Changing min_h_sync_width to 4 will let us validate timing for preffered mode and light up the headset. This change was made to Vega 10 before for a similar issue.
--- drivers/gpu/drm/amd/display/dc/dcn20/dcn20_optc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_optc.c +1 -1
@@ -535,7 +535,7 @@ void dcn20_timing_generator_init(struct optc *optc1) optc1->min_h_blank =32 ; optc1->min_v_blank =3 ; optc1->min_v_blank_interlace =5 ; - optc1->min_h_sync_width =8 ; + optc1->min_h_sync_width =4 ;// Minimum HSYNC = 8 pixels asked By HW in the first place for no actual reason. Oculus Rift S will not light up with 8 as it's hsyncWidth is 6. Changing it to 4 to fix that issue. optc1->min_v_sync_width =1 ; optc1->comb_opp_id =0xf ; }
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 45/87 drm/amd/display: Remove second initialization of pp_smu
In reply to a message not in the archiveFrom: Alvin Lee <[email protected]>
[why]
We initialize pp_smu twice
[how]
Remove second initialization of pp_smu
--- drivers/gpu/drm/amd/display/dc/clk_mgr/dcn20/dcn20_clk_mgr.c | 2 -- 1 file changed, 2 deletions(-)
drivers/gpu/drm/amd/display/dc/clk_mgr/dcn20/dcn20_clk_mgr.c +0 -2
@@ -346,8 +346,6 @@ void dcn20_clk_mgr_construct( clk_mgr->base.dprefclk_khz =700000 ;// 700 MHz planned if VCO is 3.85 GHz, will be retrieved - clk_mgr->pp_smu = pp_smu; -if (IS_FPGA_MAXIMUS_DC (ctx->dce_environment)) { dcn2_funcs.update_clocks = dcn2_update_clocks_fpga; clk_mgr->dentist_vco_freq_khz =3850000 ;
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 46/87 drm/amd/display: 3.2.40
In reply to a message not in the archiveFrom: Aric Cyr <[email protected]>
--- drivers/gpu/drm/amd/display/dc/dc.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/dc.h +1 -1
@@ -39,7 +39,7 @@#include "inc/hw/dmcu.h" #include "dml/display_mode_lib.h" -#define DC_VER "3.2.39" +#define DC_VER "3.2.40"#define MAX_SURFACES 3#define MAX_PLANES 6
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 47/87 drm/amd/display: Wait for flip to complete
In reply to a message not in the archiveFrom: Alvin Lee <[email protected]>
[why]
In pipe split issue occurs when we program immediate flip while vsync flip is pending
[how]
Don't program immediate flip until flip is no longer pending
--- drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c | 11 +++++++++++ 1 file changed, 11 insertions(+)
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c +11 -0
@@ -1283,6 +1283,17 @@ void dcn20_pipe_control_lock(if (pipe->plane_state !=NULL ) flip_immediate = pipe->plane_state->flip_immediate; +if (flip_immediate && lock) { +while (pipe->plane_res.hubp->funcs->hubp_is_flip_pending (pipe->plane_res.hubp)) { +udelay (1 ); + } + +if (pipe->bottom_pipe !=NULL ) +while (pipe->bottom_pipe->plane_res.hubp->funcs->hubp_is_flip_pending (pipe->bottom_pipe->plane_res.hubp)) { +udelay (1 ); + } + } +/* In flip immediate and pipe splitting case, we need to use GSL * for synchronization. Only do setup on locking and on flip type change. */
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 48/87 drm/amd/display: Implement generic MUX registers
In reply to a message not in the archiveFrom: Murton Liu <[email protected]>
[Why]
Logic & structures for generic regs does not exist in DC currently.
[How]
Implement register masks/shifts and relevant functions for generic mux, similar to existing HPD and DDC objects.
--- drivers/gpu/drm/amd/display/dc/gpio/Makefile | 2 +- .../display/dc/gpio/dcn20/hw_factory_dcn20.c | 41 +++++- .../dc/gpio/diagnostics/hw_factory_diag.c | 1 + .../drm/amd/display/dc/gpio/generic_regs.h | 66 +++++++++ .../gpu/drm/amd/display/dc/gpio/hw_factory.h | 3 + .../gpu/drm/amd/display/dc/gpio/hw_generic.c | 132 ++++++++++++++++++ .../gpu/drm/amd/display/dc/gpio/hw_generic.h | 46 ++++++ 7 files changed, 288 insertions(+), 3 deletions(-) create mode 100644 drivers/gpu/drm/amd/display/dc/gpio/generic_regs.h create mode 100644 drivers/gpu/drm/amd/display/dc/gpio/hw_generic.c create mode 100644 drivers/gpu/drm/amd/display/dc/gpio/hw_generic.h
drivers/gpu/drm/amd/display/dc/gpio/Makefile +1 -1
@@ -24,7 +24,7 @@ # It provides the control and status of HW GPIO pins. GPIO = gpio_base.o gpio_service.o hw_factory.o \ - hw_gpio.o hw_hpd.o hw_ddc.o hw_translate.o + hw_gpio.o hw_hpd.o hw_ddc.o hw_generic.o hw_translate.o AMD_DAL_GPIO = $(addprefix $(AMDDALPATH)/dc/gpio/,$(GPIO))
drivers/gpu/drm/amd/display/dc/gpio/dcn20/hw_factory_dcn20.c +39 -2
68 lines not on this page · raw message
drivers/gpu/drm/amd/display/dc/gpio/diagnostics/hw_factory_diag.c +1 -0
@@ -38,6 +38,7 @@#include "../hw_gpio.h" #include "../hw_ddc.h" #include "../hw_hpd.h" +#include "../hw_generic.h" /* function table */ static const struct hw_factory_funcs funcs = {
drivers/gpu/drm/amd/display/dc/gpio/generic_regs.h new file +66 -0
66 lines not on this page · raw message
drivers/gpu/drm/amd/display/dc/gpio/hw_factory.h +3 -0
@@ -63,6 +63,9 @@ struct hw_factory {void (*define_ddc_registers)(struct hw_gpio_pin *pin,uint32_t en); +void (*define_generic_registers)( +struct hw_gpio_pin *pin, +uint32_t en); } *funcs; };
drivers/gpu/drm/amd/display/dc/gpio/hw_generic.c new file +132 -0
132 lines not on this page · raw message
drivers/gpu/drm/amd/display/dc/gpio/hw_generic.h new file +46 -0
46 lines not on this page · raw message
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 49/87 drm/amd/display: Use helper for determining HDMI signal
In reply to a message not in the archiveFrom: Eric Bernstein <[email protected]>
Use helper to determine if HDMI signal when processing avmute.
--- drivers/gpu/drm/amd/display/dc/core/dc_link.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/core/dc_link.c +1 -1
@@ -2858,7 +2858,7 @@ void core_link_set_avmute(struct pipe_ctx *pipe_ctx, bool enable) {struct dc *core_dc = pipe_ctx->stream->ctx->dc; -if (pipe_ctx->stream->signal !=SIGNAL_TYPE_HDMI_TYPE_A ) +if (!dc_is_hdmi_signal (pipe_ctx->stream->signal))return ; core_dc->hwss.set_avmute (pipe_ctx, enable);
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 50/87 drm/amd/display: Set FEC_READY always before link training
In reply to a message not in the archiveFrom: Nikola Cornij <[email protected]>
[why]
Right now we FEC_READY is set only before the final link training, i.e. at mode set time. This means FEC_READY won't be set when doing link training as a response to HPD. It also fails UCD400 FEC test in DP compliance.
[how]
Move FEC_READY setup to link training.
--- drivers/gpu/drm/amd/display/dc/core/dc_link.c | 14 +++++--------- .../gpu/drm/amd/display/dc/core/dc_link_dp.c | 18 +++++++++++++++--- 2 files changed, 20 insertions(+), 12 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc_link.c +5 -9
@@ -1521,15 +1521,6 @@ static enum dc_status enable_link_dp(if (link_settings.link_rate ==LINK_RATE_LOW ) skip_video_pattern = false; -#ifdef CONFIG_DRM_AMD_DC_DSC_SUPPORT -if (link->preferred_training_settings.fec_enable !=NULL ) - fec_enable = *link->preferred_training_settings.fec_enable; -else - fec_enable = true; - -dp_set_fec_ready (link, fec_enable); -#endif -if (link->aux_access_disabled) {dc_link_dp_perform_link_training_skip_aux (link, &link_settings);
@@ -1547,6 +1538,11 @@ static enum dc_status enable_link_dp( status =DC_FAIL_DP_LINK_TRAINING ;#ifdef CONFIG_DRM_AMD_DC_DSC_SUPPORT +if (link->preferred_training_settings.fec_enable !=NULL ) + fec_enable = *link->preferred_training_settings.fec_enable; +else + fec_enable = true; +dp_set_fec_enable (link, fec_enable);#endif return status;
drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c +15 -3
@@ -1179,14 +1179,26 @@ enum link_training_result dc_link_dp_perform_link_training( bool skip_video_pattern) {enum link_training_result status =LINK_TRAINING_SUCCESS ; -struct link_training_settings lt_settings; +#ifdef CONFIG_DRM_AMD_DC_DSC_SUPPORT +bool fec_enable; +#endif initialize_training_settings (link, link_setting, <_settings);/* 1. set link rate, lane count and spread. */ dpcd_set_link_settings (link, <_settings); +#ifdef CONFIG_DRM_AMD_DC_DSC_SUPPORT +if (link->preferred_training_settings.fec_enable !=NULL ) + fec_enable = *link->preferred_training_settings.fec_enable; +else + fec_enable = true; + +dp_set_fec_ready (link, fec_enable); +#endif + +/* 2. perform link training (set link training done * to false is done as well) */
@@ -3153,7 +3165,7 @@ void dp_set_fec_ready(struct dc_link *link, bool ready)if (link_enc->funcs->fec_set_ready && link->dpcd_caps.fec_cap.bits.FEC_CAPABLE) { -if (link->fec_state == dc_link_fec_not_ready && ready) { +if (ready) { fec_config =1 ; if (core_link_write_dpcd(link, DP_FEC_CONFIGURATION,
@@ -3164,7 +3176,7 @@ void dp_set_fec_ready(struct dc_link *link, bool ready) }else {dm_error ("dpcd write failed to set fec_ready" ); } - }else if (link->fec_state == dc_link_fec_ready && !ready) { + }else if (link->fec_state == dc_link_fec_ready) { fec_config =0 ; core_link_write_dpcd(link, DP_FEC_CONFIGURATION,
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 51/87 drm/amd/display: put back front end initialization sequence
In reply to a message not in the archiveFrom: Eric Yang <[email protected]>
[Why]
Seamless boot optimization removed proper front end power off sequence. In driver disable enable case, this causes driver to power gate hubp and dpp while there is still memory fetching going on, this can cause invalid memory requests to be generated which will hang data fabric.
[How]
Put back proper front end power off sequence
--- .../drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c | 15 +-------------- 1 file changed, 1 insertion(+), 14 deletions(-)
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c +1 -14
@@ -1194,16 +1194,7 @@ static void dcn10_init_hw(struct dc *dc)* everything down. */ if (dcb->funcs->is_accelerated_mode (dcb) || dc->config.power_down_display_on_boot) { -for (i =0 ; i < dc->res_pool->pipe_count; i++) { -struct hubp *hubp = dc->res_pool->hubps[i]; -struct dpp *dpp = dc->res_pool->dpps[i]; - - hubp->funcs->hubp_init (hubp); - dc->res_pool->opps[i]->mpc_tree_params.opp_id = dc->res_pool->opps[i]->inst; -plane_atomic_power_down (dc, dpp, hubp); - } - -apply_DEGVIDCN10_253_wa (dc); + dc->hwss.init_pipes (dc, dc->current_state); }for (i =0 ; i < dc->res_pool->audio_count; i++) {
@@ -1374,10 +1365,6 @@ static bool dcn10_set_input_transfer_func(struct pipe_ctx *pipe_ctx,return result; } - - - -static bool dcn10_set_output_transfer_func (struct pipe_ctx *pipe_ctx,const struct dc_stream_state *stream)
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 52/87 drm/amd/display: Optimize gamma calculations
In reply to a message not in the archiveFrom: Krunoslav Kovac <[email protected]>
[Why&How]
1. Stack usage is pretty high as fixed31_32 struct is 8 bytes and we have functions with >30 vars on the stack.
2. Optimize gamma calculation by reducing number of calls to dc_fixpt_pow Our X points are divided into 32 regions wth 16 pts each. Each region is 2x the previous, meaning x[i] = 2*x[i-16] for i>=16. Using (2x)^gamma = 2^gamma * x^gamma, we can recursively compute powers of gamma, we just need first 16 pts to start it up. dc_fixpt_pow() is expensive, it computes x^y by doing exp(y*logx) Exp is done by Taylor series approximation, and log by Newton-like approximation that also uses exp internally. In short, it's significantly heavier than run-of-the-mill addition/subtraction/multiply.
--- drivers/gpu/drm/amd/display/dc/dc_hw_types.h | 1 - .../amd/display/modules/color/color_gamma.c | 163 +++++++++++------- .../amd/display/modules/color/color_gamma.h | 9 + 3 files changed, 111 insertions(+), 62 deletions(-)
drivers/gpu/drm/amd/display/dc/dc_hw_types.h +0 -1
@@ -482,7 +482,6 @@ struct dc_gamma {* is_logical_identity indicates the given gamma ramp regardless of type is identity. */ bool is_identity; -bool is_logical_identity; };/* Used by both ipp amd opp functions*/
drivers/gpu/drm/amd/display/modules/color/color_gamma.c +102 -61
256 lines not on this page · raw message
drivers/gpu/drm/amd/display/modules/color/color_gamma.h +9 -0
@@ -82,6 +82,15 @@ struct freesync_hdr_tf_params {unsigned int skip_tm;// skip tm }; +struct translate_from_linear_space_args { +struct fixed31_32 arg; +struct fixed31_32 a0; +struct fixed31_32 a1; +struct fixed31_32 a2; +struct fixed31_32 a3; +struct fixed31_32 gamma; +}; +void setup_x_points_distribution (void );void precompute_pq (void );void precompute_de_pq (void );
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 53/87 drm/amd/display: Clear FEC_READY shadow register if DPCD write fails
In reply to a message not in the archiveFrom: Nikola Cornij <[email protected]>
[why]
As a fail-safe, in case 'set FEC_READY' DPCD write fails, a HW shadow register should be cleared and the internal FEC stat should be set to 'not ready'. This is to make sure HW settings will be consistent with FEC_READY state on the RX.
--- drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c | 2 ++ 1 file changed, 2 insertions(+)
drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c +2 -0
@@ -3174,6 +3174,8 @@ void dp_set_fec_ready(struct dc_link *link, bool ready) link_enc->funcs->fec_set_ready (link_enc, true); link->fec_state = dc_link_fec_ready; }else { + link->link_enc->funcs->fec_set_ready (link->link_enc, false); + link->fec_state = dc_link_fec_not_ready;dm_error ("dpcd write failed to set fec_ready" ); } }else if (link->fec_state == dc_link_fec_ready) {
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 54/87 drm/amd/display: Add debug option to disable timing sync
In reply to a message not in the archiveFrom: Joshua Aberback <[email protected]>
[Why]
We want a debug option to disable timing sync for testing.
[How]
New dc debug option that must be false to call program_timing_sync
--- drivers/gpu/drm/amd/display/dc/core/dc.c | 2 +- drivers/gpu/drm/amd/display/dc/dc.h | 1 + 2 files changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/core/dc.c +1 -1
@@ -1060,7 +1060,7 @@ static enum dc_status dc_commit_state_no_check(struct dc *dc, struct dc_state *cif (result !=DC_OK )return result; -if (context->stream_count >1 ) { +if (context->stream_count >1 && !dc->debug.disable_timing_sync) {enable_timing_multisync (dc, context);program_timing_sync (dc, context); }
drivers/gpu/drm/amd/display/dc/dc.h +1 -0
@@ -378,6 +378,7 @@ struct dc_debug_options {* watermarks are not affected. */ unsigned int force_min_dcfclk_mhz; +bool disable_timing_sync; };struct dc_debug_data {
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 55/87 drm/amd/display: Add MPC 3DLUT resource management
In reply to a message not in the archiveFrom: Vitaly Prosyak <[email protected]>
[Why & How]
Number of 3DLUT's in MPC are not equal to number of pipes. Resource management is required.
Activate on FPGA entire tm solution which includes the following :hdr multiplier, shaper, 3dlut.
--- drivers/gpu/drm/amd/display/dc/core/dc_surface.c | 3 +-- drivers/gpu/drm/amd/display/dc/dc.h | 16 +++++++++++++++- .../gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c | 4 ++-- 3 files changed, 18 insertions(+), 5 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc_surface.c +1 -2
@@ -58,7 +58,6 @@ static void construct(struct dc_context *ctx, struct dc_plane_state *plane_state plane_state->lut3d_func =dc_create_3dlut_func ();if (plane_state->lut3d_func !=NULL ) { plane_state->lut3d_func->ctx = ctx; - plane_state->lut3d_func->initialized = false; } plane_state->blend_tf =dc_create_transfer_func ();if (plane_state->blend_tf !=NULL ) {
@@ -277,7 +276,7 @@ struct dc_3dlut *dc_create_3dlut_func(void) goto alloc_fail;kref_init (&lut->refcount); - lut->initialized = false; + lut->state.raw =0 ;return lut;
drivers/gpu/drm/amd/display/dc/dc.h +15 -1
@@ -616,12 +616,26 @@ struct dc_transfer_func {#if defined(CONFIG_DRM_AMD_DC_DCN2_0 ) +union dc_3dlut_state { +struct { +uint32_t initialized:1 ;/*if 3dlut is went through color module for initialization */ +uint32_t rmu_idx_valid:1 ;/*if mux settings are valid*/ +uint32_t rmu_mux_num:3 ;/*index of mux to use*/ +uint32_t mpc_rmu0_mux:4 ;/*select mpcc on mux, one of the following : mpcc0, mpcc1, mpcc2, mpcc3*/ +uint32_t mpc_rmu1_mux:4 ; +uint32_t mpc_rmu2_mux:4 ; +uint32_t reserved:15 ; + } bits; +uint32_t raw; +}; +struct dc_3dlut {struct kref refcount;struct tetrahedral_params lut_3d;uint32_t hdr_multiplier; -bool initialized; +bool initialized;/*remove after diag fix*/ +union dc_3dlut_state state;struct dc_context *ctx; };#endif
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c +2 -2
@@ -927,14 +927,14 @@ static bool dcn20_set_shaper_3dlut( result = dpp_base->funcs->dpp_program_shaper_lut (dpp_base, shaper_lut);if (plane_state->lut3d_func && - plane_state->lut3d_func->initialized == true) + plane_state->lut3d_func->state.bits.initialized ==1 ) result = dpp_base->funcs->dpp_program_3dlut (dpp_base, &plane_state->lut3d_func->lut_3d);else result = dpp_base->funcs->dpp_program_3dlut (dpp_base,NULL );if (plane_state->lut3d_func && - plane_state->lut3d_func->initialized == true && + plane_state->lut3d_func->state.bits.initialized ==1 && plane_state->lut3d_func->hdr_multiplier !=0 ) dpp_base->funcs->dpp_set_hdr_multiplier (dpp_base, plane_state->lut3d_func->hdr_multiplier);
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 56/87 drm/amd/display: Add CM_BYPASS via debug option
In reply to a message not in the archiveFrom: Qingqing Zhuo <[email protected]>
[Why]
bypass CM block and MPC ogam for debug or triage use.
[How]
create a new flag cm_bypass_mode, which will set both CM_CONTROL and MPCC_OGAM_MODE to bypass when set to 1.
--- drivers/gpu/drm/amd/display/dc/dc.h | 3 +++ drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dpp_cm.c | 7 ++++++- drivers/gpu/drm/amd/display/dc/dcn20/dcn20_mpc.c | 11 +++++++++++ drivers/gpu/drm/amd/display/dc/inc/hw/dpp.h | 1 + drivers/gpu/drm/amd/display/dc/inc/hw/mpc.h | 1 + 5 files changed, 22 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/dc.h +3 -0
@@ -379,6 +379,9 @@ struct dc_debug_options {*/ unsigned int force_min_dcfclk_mhz;bool disable_timing_sync; +#if defined(CONFIG_DRM_AMD_DC_DCN2_0 ) +bool cm_in_bypass; +#endif };struct dc_debug_data {
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dpp_cm.c +6 -1
@@ -52,7 +52,12 @@ static void dpp2_enable_cm_block( {struct dcn20_dpp *dpp =TO_DCN20_DPP (dpp_base); -REG_UPDATE (CM_CONTROL ,CM_BYPASS ,0 ); +unsigned int cm_bypass_mode =0 ; +//Temp, put CM in bypass mode +if (dpp_base->ctx->dc->debug.cm_in_bypass) + cm_bypass_mode =1 ; + +REG_UPDATE (CM_CONTROL ,CM_BYPASS , cm_bypass_mode); }
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_mpc.c +11 -0
@@ -368,6 +368,11 @@ void apply_DEDCN20_305_wa( {struct dcn20_mpc *mpc20 =TO_DCN20_MPC (mpc); +if (mpc->ctx->dc->debug.cm_in_bypass) { +REG_SET (MPCC_OGAM_MODE [mpcc_id],0 ,MPCC_OGAM_MODE ,0 ); +return ; + } +if (mpc->ctx->dc->work_arounds.dedcn20_305_wa == false) {/*hw fixed in new review*/ return ;
@@ -390,10 +395,16 @@ void mpc2_set_output_gamma(enum dc_lut_mode next_mode;struct dcn20_mpc *mpc20 =TO_DCN20_MPC (mpc); +if (mpc->ctx->dc->debug.cm_in_bypass) { +REG_SET (MPCC_OGAM_MODE [mpcc_id],0 ,MPCC_OGAM_MODE ,0 ); +return ; + } +if (params ==NULL ) {REG_SET (MPCC_OGAM_MODE [mpcc_id],0 ,MPCC_OGAM_MODE ,0 );return ; } + current_mode =mpc20_get_ogam_current (mpc, mpcc_id);if (current_mode ==LUT_BYPASS || current_mode ==LUT_RAM_A ) next_mode =LUT_RAM_B ;
drivers/gpu/drm/amd/display/dc/inc/hw/dpp.h +1 -0
@@ -42,6 +42,7 @@ struct dpp {#if defined(CONFIG_DRM_AMD_DC_DCN2_0 )struct pwl_params shaper_params; +bool cm_bypass_mode;#endif };
drivers/gpu/drm/amd/display/dc/inc/hw/mpc.h +1 -0
@@ -128,6 +128,7 @@ struct mpc {struct mpcc mpcc_array[MAX_MPCC ];#if defined(CONFIG_DRM_AMD_DC_DCN2_0 )struct pwl_params blender_params; +bool cm_bypass_mode;#endif };
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 57/87 drm/amd/display: Add DIG_CLOCK_PATTERN register
In reply to a message not in the archiveFrom: Nevenko Stupar <[email protected]>
Add this register for future use
--- .../drm/amd/display/dc/dcn10/dcn10_stream_encoder.h | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-)
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_stream_encoder.h +7 -3
@@ -89,7 +89,8 @@SRI (DP_VID_STREAM_CNTL ,DP , id), \SRI (DP_VID_TIMING ,DP , id), \SRI (DP_SEC_AUD_N ,DP , id), \ - SRI(DP_SEC_TIMESTAMP, DP, id) +SRI (DP_SEC_TIMESTAMP ,DP , id), \ + SRI(DIG_CLOCK_PATTERN, DIG, id)#define SE_DCN_REG_LIST (id)\ SE_COMMON_DCN_REG_LIST(id)
@@ -170,6 +171,7 @@ struct dcn10_stream_enc_registers {uint32_t HDMI_METADATA_PACKET_CONTROL ;uint32_t DP_SEC_FRAMING4 ;#endif +uint32_t DIG_CLOCK_PATTERN ; };
@@ -298,7 +300,8 @@ struct dcn10_stream_enc_registers {SE_SF (DP0_DP_MSA_TIMING_PARAM4 ,DP_MSA_VHEIGHT , mask_sh),\SE_SF (DIG0_HDMI_DB_CONTROL ,HDMI_DB_DISABLE , mask_sh),\SE_SF (DP0_DP_VID_TIMING ,DP_VID_N_MUL , mask_sh),\ - SE_SF(DIG0_DIG_FE_CNTL, DIG_SOURCE_SELECT, mask_sh) +SE_SF (DIG0_DIG_FE_CNTL ,DIG_SOURCE_SELECT , mask_sh),\ + SE_SF(DIG0_DIG_CLOCK_PATTERN, DIG_CLOCK_PATTERN, mask_sh)#define SE_COMMON_MASK_SH_LIST_SOC (mask_sh)\ SE_COMMON_MASK_SH_LIST_SOC_BASE(mask_sh)
@@ -460,7 +463,8 @@ struct dcn10_stream_enc_registers {type HDMI_DB_DISABLE ;\type DP_VID_N_MUL ;\type DP_VID_M_DOUBLE_VALUE_EN ;\ - type DIG_SOURCE_SELECT +type DIG_SOURCE_SELECT ;\ + type DIG_CLOCK_PATTERN#if defined(CONFIG_DRM_AMD_DC_DCN2_0 )#define SE_REG_FIELD_LIST_DCN2_0 (type) \
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 58/87 drm/amd/display: Cache the use_pitch_c conditional
In reply to a message not in the archiveFrom: Ilya Bakoulin <[email protected]>
For clarity, save the use_pitch_c logic to a bool
--- drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hubp.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hubp.c +6 -2
@@ -402,11 +402,14 @@ void hubp2_program_size( {struct dcn20_hubp *hubp2 =TO_DCN20_HUBP (hubp);uint32_t pitch, meta_pitch, pitch_c, meta_pitch_c; +bool use_pitch_c = false;/* Program data and meta surface pitch (calculation from addrlib) * 444 or 420 luma */ -if (format >=SURFACE_PIXEL_FORMAT_VIDEO_BEGIN && format <SURFACE_PIXEL_FORMAT_SUBSAMPLE_END ) { + use_pitch_c = format >=SURFACE_PIXEL_FORMAT_VIDEO_BEGIN + && format <SURFACE_PIXEL_FORMAT_SUBSAMPLE_END ; +if (use_pitch_c) {ASSERT (plane_size->video.chroma_pitch !=0 );/* Chroma pitch zero can cause system hang! */
@@ -429,7 +432,8 @@ void hubp2_program_size(REG_UPDATE_2 (DCSURF_SURFACE_PITCH ,PITCH , pitch,META_PITCH , meta_pitch); -if (format >=SURFACE_PIXEL_FORMAT_VIDEO_BEGIN ) + use_pitch_c = format >=SURFACE_PIXEL_FORMAT_VIDEO_BEGIN ; +if (use_pitch_c)REG_UPDATE_2 (DCSURF_SURFACE_PITCH_C ,PITCH_C , pitch_c,META_PITCH_C , meta_pitch_c); }
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 59/87 drm/amd/display: Fixes for some MPO cases
In reply to a message not in the archiveFrom: Ilya Bakoulin <[email protected]>
[Why]
Alpha could be improperly applied (only affecting half the frame) for some source pixel formats.
[How]
Change how alpha is enabled in MPC/DPP LB and change the bottom plane blend mode in MPC.
--- drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mpc.c | 2 +- drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mpc.c +1 -1
@@ -211,7 +211,7 @@ struct mpcc *mpc1_insert_plane( }else { new_mpcc->mpcc_bot =NULL ;REG_SET (MPCC_BOT_SEL [mpcc_id],0 ,MPCC_BOT_SEL ,0xf ); -REG_UPDATE (MPCC_CONTROL [mpcc_id],MPCC_MODE ,MPCC_BLEND_MODE_TOP_LAYER_PASSTHROUGH ); +REG_UPDATE (MPCC_CONTROL [mpcc_id],MPCC_MODE ,MPCC_BLEND_MODE_TOP_LAYER_ONLY ); }REG_SET (MPCC_TOP_SEL [mpcc_id],0 ,MPCC_TOP_SEL , dpp_id);REG_SET (MPCC_OPP_ID [mpcc_id],0 ,MPCC_OPP_ID , tree->opp_id);
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c +1 -1
@@ -1838,7 +1838,7 @@ static void dcn20_update_mpcc(struct dc *dc, struct pipe_ctx *pipe_ctx) {struct hubp *hubp = pipe_ctx->plane_res.hubp;struct mpcc_blnd_cfg blnd_cfg = { {0 } }; -bool per_pixel_alpha = pipe_ctx->plane_state->per_pixel_alpha && pipe_ctx->bottom_pipe; +bool per_pixel_alpha = pipe_ctx->plane_state->per_pixel_alpha;int mpcc_id;struct mpcc *new_mpcc;struct mpc *mpc = dc->res_pool->mpc;
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 60/87 drm/amd/display: 3.2.41
In reply to a message not in the archiveFrom: Aric Cyr <[email protected]>
--- drivers/gpu/drm/amd/display/dc/dc.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/dc.h +1 -1
@@ -39,7 +39,7 @@#include "inc/hw/dmcu.h" #include "dml/display_mode_lib.h" -#define DC_VER "3.2.40" +#define DC_VER "3.2.41"#define MAX_SURFACES 3#define MAX_PLANES 6
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 61/87 drm/amd/display: Hook up calls to do stereo mux and dig programming to stereo control interface
In reply to a message not in the archiveFrom: Murton Liu <[email protected]>
[Why]
Implementation of stereo mux register is complete, but unused. Need to call functions to write relevant configs.
[How]
Add function to write stereo config for enable/disable case and call in stereo control interface.
--- drivers/gpu/drm/amd/display/dc/core/dc.c | 47 ++++++++++++ drivers/gpu/drm/amd/display/dc/dc.h | 3 + .../amd/display/dc/dcn10/dcn10_hw_sequencer.c | 7 ++ .../display/dc/gpio/dcn10/hw_factory_dcn10.c | 42 ++++++++++- .../drm/amd/display/dc/gpio/gpio_service.c | 72 +++++++++++++++++++ .../display/include/gpio_service_interface.h | 18 ++++- 6 files changed, 186 insertions(+), 3 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc.c +47 -0
@@ -1234,6 +1234,53 @@ void dc_release_state(struct dc_state *context)kref_put (&context->refcount, dc_state_free); } +bool dc_set_generic_gpio_for_stereo (bool enable, +struct gpio_service *gpio_service) +{ +enum gpio_result gpio_result =GPIO_RESULT_NON_SPECIFIC_ERROR ; +struct gpio_pin_info pin_info; +struct gpio *generic; +struct gpio_generic_mux_config *config =kzalloc (sizeof (struct gpio_generic_mux_config ), +GFP_KERNEL ); + + pin_info =dal_gpio_get_generic_pin_info (gpio_service,GPIO_ID_GENERIC ,0 ); + +if (pin_info.mask ==0xFFFFFFFF || pin_info.offset ==0xFFFFFFFF ) { +kfree (config); +return false; + }else { + generic =dal_gpio_service_create_generic_mux ( + gpio_service, + pin_info.offset, + pin_info.mask); + } + +if (!generic) { +kfree (config); +return false; + } + + gpio_result =dal_gpio_open (generic,GPIO_MODE_OUTPUT ); + + config->enable_output_from_mux = enable; + config->mux_select =GPIO_SIGNAL_SOURCE_PASS_THROUGH_STEREO_SYNC ; + +if (gpio_result ==GPIO_RESULT_OK ) + gpio_result =dal_mux_setup_config (generic, config); + +if (gpio_result ==GPIO_RESULT_OK ) { +dal_gpio_close (generic); +dal_gpio_destroy_generic_mux (&generic); +kfree (config); +return true; + }else { +dal_gpio_close (generic); +dal_gpio_destroy_generic_mux (&generic); +kfree (config); +return false; + } +} +static bool is_surface_in_context (const struct dc_state *context,const struct dc_plane_state *plane_state)
drivers/gpu/drm/amd/display/dc/dc.h +3 -0
@@ -852,6 +852,9 @@ enum dc_status dc_validate_plane(struct dc *dc, const struct dc_plane_state *plavoid get_clock_requirements_for_state (struct dc_state *state,struct AsicStateEx *info); +bool dc_set_generic_gpio_for_stereo (bool enable, +struct gpio_service *gpio_service); +/* * fast_validate: we return after determining if we can support the new state, * but before we populate the programming info
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c +7 -0
@@ -2691,6 +2691,13 @@ static void dcn10_setup_stereo(struct pipe_ctx *pipe_ctx, struct dc *dc)dcn10_config_stereo_parameters (stream, &flags); +if (stream->timing.timing_3d_format ==TIMING_3D_FORMAT_SIDEBAND_FA ) { +if (!dc_set_generic_gpio_for_stereo (true, dc->ctx->gpio_service)) +dc_set_generic_gpio_for_stereo (false, dc->ctx->gpio_service); + }else { +dc_set_generic_gpio_for_stereo (false, dc->ctx->gpio_service); + } + pipe_ctx->stream_res.opp->funcs->opp_program_stereo( pipe_ctx->stream_res.opp, flags.PROGRAM_STEREO == 1 ? true:false,
drivers/gpu/drm/amd/display/dc/gpio/dcn10/hw_factory_dcn10.c +40 -2
64 lines not on this page · raw message
drivers/gpu/drm/amd/display/dc/gpio/gpio_service.c +72 -0
90 lines not on this page · raw message
drivers/gpu/drm/amd/display/include/gpio_service_interface.h +17 -1
@@ -51,13 +51,29 @@ struct gpio *dal_gpio_service_create_irq(uint32_t offset, uint32_t mask); +struct gpio *dal_gpio_service_create_generic_mux ( +struct gpio_service *service, +uint32_t offset, +uint32_t mask); + +void dal_gpio_destroy_generic_mux ( +struct gpio **mux); + +enum gpio_result dal_mux_setup_config ( +struct gpio *mux, +struct gpio_generic_mux_config *config); + +struct gpio_pin_info dal_gpio_get_generic_pin_info ( +struct gpio_service *service, +enum gpio_id id, +uint32_t en); +struct ddc *dal_gpio_create_ddc (struct gpio_service *service,uint32_t offset,uint32_t mask,struct gpio_ddc_hw_info *info); -void dal_gpio_destroy_ddc (struct ddc **ddc);
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 62/87 drm/amd/display: allocate 4 ddc engines for RV2
In reply to a message not in the archiveFrom: Derek Lai <[email protected]>
[Why]
Driver will create 0, 1, and 2 ddc engines for RV2, but some platforms used 0, 1, and 3.
[How]
Still allocate 4 ddc engines for RV2.
--- drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c +1 -1
@@ -506,7 +506,7 @@ static const struct resource_caps rv2_res_cap = { .num_audio = 3, .num_stream_encoder = 3, .num_pll = 3, - .num_ddc = 3, + .num_ddc = 4, };static const struct dc_plane_cap plane_cap = {
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 63/87 drm/amd/display: add set and get clock for testing purposes
In reply to a message not in the archiveFrom: Charlene Liu <[email protected]>
add dc_set_clock
add dc_get_clock
this is for testing and diagnostics to get/set DPPCLK and DISPCLK.
--- .../display/dc/clk_mgr/dcn20/dcn20_clk_mgr.c | 23 +++++++- .../display/dc/clk_mgr/dcn20/dcn20_clk_mgr.h | 5 ++ drivers/gpu/drm/amd/display/dc/core/dc.c | 11 ++++ drivers/gpu/drm/amd/display/dc/dc.h | 5 ++ drivers/gpu/drm/amd/display/dc/dc_types.h | 13 +++++ .../amd/display/dc/dcn10/dcn10_hw_sequencer.c | 53 ++++++++++++++++++- .../drm/amd/display/dc/dcn20/dcn20_resource.c | 5 ++ .../gpu/drm/amd/display/dc/inc/core_status.h | 3 ++ .../gpu/drm/amd/display/dc/inc/hw/clk_mgr.h | 7 +++ .../gpu/drm/amd/display/dc/inc/hw_sequencer.h | 9 ++++ 10 files changed, 132 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/display/dc/clk_mgr/dcn20/dcn20_clk_mgr.c +22 -1
@@ -316,11 +316,32 @@ void dcn2_enable_pme_wa(struct clk_mgr *clk_mgr_base) } } +void dcn2_get_clock (struct clk_mgr *clk_mgr, +struct dc_state *context, +enum dc_clock_type clock_type, +struct dc_clock_config *clock_cfg) +{ + +if (clock_type ==DC_CLOCK_TYPE_DISPCLK ) { + clock_cfg->max_clock_khz = context->bw_ctx.bw.dcn.clk.max_supported_dispclk_khz; + clock_cfg->min_clock_khz = DCN_MINIMUM_DISPCLK_Khz; + clock_cfg->current_clock_khz = clk_mgr->clks.dispclk_khz; + clock_cfg->bw_requirequired_clock_khz = context->bw_ctx.bw.dcn.clk.bw_dispclk_khz; + } +if (clock_type ==DC_CLOCK_TYPE_DPPCLK ) { + clock_cfg->max_clock_khz = context->bw_ctx.bw.dcn.clk.max_supported_dppclk_khz; + clock_cfg->min_clock_khz = DCN_MINIMUM_DPPCLK_Khz; + clock_cfg->current_clock_khz = clk_mgr->clks.dppclk_khz; + clock_cfg->bw_requirequired_clock_khz = context->bw_ctx.bw.dcn.clk.bw_dppclk_khz; + } +} +static struct clk_mgr_funcs dcn2_funcs = { .get_dp_ref_clk_frequency = dce12_get_dp_ref_freq_khz, .update_clocks = dcn2_update_clocks, .init_clocks = dcn2_init_clocks, - .enable_pme_wa = dcn2_enable_pme_wa + .enable_pme_wa = dcn2_enable_pme_wa, + .get_clock = dcn2_get_clock, };
drivers/gpu/drm/amd/display/dc/clk_mgr/dcn20/dcn20_clk_mgr.h +5 -0
@@ -45,4 +45,9 @@ void dcn20_clk_mgr_construct(struct dc_context *ctx,uint32_t dentist_get_did_from_divider (int divider); +void dcn2_get_clock (struct clk_mgr *clk_mgr, +struct dc_state *context, +enum dc_clock_type clock_type, +struct dc_clock_config *clock_cfg); +#endif //__DCN20_CLK_MGR_H__
drivers/gpu/drm/amd/display/dc/core/dc.c +11 -0
@@ -2429,3 +2429,14 @@ void get_clock_requirements_for_state(struct dc_state *state, struct AsicStateEx info->fClock = (unsigned int )state->bw_ctx.bw.dcn.clk.fclk_khz; info->phyClock = (unsigned int )state->bw_ctx.bw.dcn.clk.phyclk_khz; } +enum dc_status dc_set_clock (struct dc *dc,enum dc_clock_type clock_type,uint32_t clk_khz,uint32_t stepping) +{ +if (dc->hwss.set_clock) +return dc->hwss.set_clock (dc, clock_type, clk_khz, stepping); +return DC_ERROR_UNEXPECTED ; +} +void dc_get_clock (struct dc *dc,enum dc_clock_type clock_type,struct dc_clock_config *clock_cfg) +{ +if (dc->hwss.get_clock) + dc->hwss.get_clock (dc, clock_type, clock_cfg); +}
drivers/gpu/drm/amd/display/dc/dc.h +5 -0
@@ -252,7 +252,10 @@ enum wm_report_mode {struct dc_clocks {int dispclk_khz;int max_supported_dppclk_khz; +int max_supported_dispclk_khz;int dppclk_khz; +int bw_dppclk_khz;/*a copy of dppclk_khz*/ +int bw_dispclk_khz;int dcfclk_khz;int socclk_khz;int dcfclk_deep_sleep_khz;
@@ -1041,6 +1044,8 @@ unsigned int dc_get_target_backlight_pwm(struct dc *dc);bool dc_is_dmcu_initialized (struct dc *dc); +enum dc_status dc_set_clock (struct dc *dc,enum dc_clock_type clock_type,uint32_t clk_khz,uint32_t stepping); +void dc_get_clock (struct dc *dc,enum dc_clock_type clock_type,struct dc_clock_config *clock_cfg);#if defined(CONFIG_DRM_AMD_DC_DSC_SUPPORT )/******************************************************************************* * DSC Interfaces
drivers/gpu/drm/amd/display/dc/dc_types.h +13 -0
@@ -726,6 +726,19 @@ struct AsicStateEx {unsigned int phyClock; }; + +enum dc_clock_type { +DC_CLOCK_TYPE_DISPCLK =0 , +DC_CLOCK_TYPE_DPPCLK =1 , +}; + +struct dc_clock_config { +uint32_t max_clock_khz; +uint32_t min_clock_khz; +uint32_t bw_requirequired_clock_khz; +uint32_t current_clock_khz;/*current clock in use*/ +}; +#ifdef CONFIG_DRM_AMD_DC_DSC_SUPPORT /* DSC DPCD capabilities */ union dsc_slice_caps1 {
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c +52 -1
65 lines not on this page · raw message
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c +5 -0
@@ -2396,6 +2396,11 @@ void dcn20_calculate_dlg_params( context->res_ctx.pipe_ctx[i].pipe_dlg_param = pipes[pipe_idx].pipe.dest; pipe_idx++; } +/*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; + context->bw_ctx.bw.dcn.clk.max_supported_dppclk_khz = context->bw_ctx.dml.soc.clock_limits[vlevel].dppclk_mhz*1000 ; + context->bw_ctx.bw.dcn.clk.max_supported_dispclk_khz = context->bw_ctx.dml.soc.clock_limits[vlevel].dispclk_mhz*1000 ;for (i =0 , pipe_idx =0 ; i < dc->res_pool->pipe_count; i++) {bool cstate_en = context->bw_ctx.dml.vba.PrefetchMode[vlevel][context->bw_ctx.dml.vba.maxMpcComb] !=2 ;
drivers/gpu/drm/amd/display/dc/inc/core_status.h +3 -0
@@ -48,6 +48,9 @@ enum dc_status {DC_NO_DSC_RESOURCE =17 , #endifDC_FAIL_UNSUPPORTED_1 =18 , +DC_FAIL_CLK_EXCEED_MAX =21 , +DC_FAIL_CLK_BELOW_MIN =22 ,/*THIS IS MIN PER IP*/ +DC_FAIL_CLK_BELOW_CFG_REQUIRED =23 ,/*THIS IS hard_min in PPLIB*/ DC_ERROR_UNEXPECTED =-1 };
drivers/gpu/drm/amd/display/dc/inc/hw/clk_mgr.h +7 -0
@@ -28,6 +28,9 @@#include "dc.h" +#define DCN_MINIMUM_DISPCLK_Khz 100000 +#define DCN_MINIMUM_DPPCLK_Khz 100000 +/* Public interfaces */ struct clk_states {
@@ -51,6 +54,10 @@ struct clk_mgr_funcs {void (*init_clocks)(struct clk_mgr *clk_mgr);void (*enable_pme_wa) (struct clk_mgr *clk_mgr); +void (*get_clock)(struct clk_mgr *clk_mgr, +struct dc_state *context, +enum dc_clock_type clock_type, +struct dc_clock_config *clock_cfg); };struct clk_mgr {
drivers/gpu/drm/amd/display/dc/inc/hw_sequencer.h +9 -0
@@ -294,6 +294,15 @@ struct hw_sequencer_funcs {void (*disable_writeback)(struct dc *dc,unsigned int dwb_pipe_inst);#endif +enum dc_status (*set_clock)(struct dc *dc, +enum dc_clock_type clock_type, +uint32_t clk_khz, +uint32_t stepping); + +void (*get_clock)(struct dc *dc, +enum dc_clock_type clock_type, +struct dc_clock_config *clock_cfg); + }; void color_space_to_black_color(
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 64/87 drm/amd/display: Change offset_to_id to reflect what id_to_offset returns
In reply to a message not in the archiveFrom: Murton Liu <[email protected]>
id_to_offset does not point to the same reg offset that offset_to_id checks for, causing unintended asserts
--- .../drm/amd/display/dc/gpio/dcn20/hw_translate_dcn20.c | 2 +- drivers/gpu/drm/amd/display/dc/gpio/gpio_service.c | 10 +++------- 2 files changed, 4 insertions(+), 8 deletions(-)
drivers/gpu/drm/amd/display/dc/gpio/dcn20/hw_translate_dcn20.c +1 -1
@@ -71,7 +71,7 @@ static bool offset_to_id( {switch (offset) {/* GENERIC */ -case REG (DC_GENERICA ): +case REG (DC_GPIO_GENERIC_A ): *id =GPIO_ID_GENERIC ;switch (mask) {case DC_GPIO_GENERIC_A__DC_GPIO_GENERICA_A_MASK :
drivers/gpu/drm/amd/display/dc/gpio/gpio_service.c +3 -7
@@ -148,14 +148,10 @@ struct gpio *dal_gpio_service_create_generic_mux(uint32_t en;struct gpio *generic; -if (mask ==1 ) - en =GPIO_GENERIC_A ; -else if (mask ==0x00000100L ) - en =GPIO_GENERIC_B ; -else +if (!service->translate.funcs->offset_to_id (offset, mask, &id, &en)) { +ASSERT_CRITICAL (false);return NULL ; - - id =GPIO_ID_GENERIC ; + } generic =dal_gpio_create ( service, id, en,GPIO_PIN_OUTPUT_STATE_DEFAULT );
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 65/87 drm/amd/display: init res_pool dccg_ref, dchub_ref with xtalin_freq
In reply to a message not in the archiveFrom: hersen wu <[email protected]>
[WHY] dc sw clock implementation of navi10 and raven are not exact the same. dcccg, dchub reference clock initialization is done after dc calls vbios dispcontroller_init table. for raven family, before dispcontroller_init is called by dc, the ref clk values are referred by sw clock implementation and program asic register using wrong values. this causes dchub pstate error. This need provide valid ref clk values. for navi10, since dispcontroller_init is not called, dchubbub_global_timer_enable = 0, hubbub2_get_dchub_ref_freq will hit aeert. this need remove hubbub2_get_dchub_ref_freq from this location and move to dcn20_init_hw.
[HOW] for all asic, initialize dccg, dchub ref clk with data from vbios firmware table by default. for raven asic family, use these data from vbios, for asic which support sw dccg component, like navi10, read ref clk by sw dccg functions and update the ref clk.
--- .../gpu/drm/amd/display/dc/core/dc_resource.c | 42 +++++++------------ .../drm/amd/display/dc/dcn20/dcn20_hwseq.c | 25 +++++++++++ 2 files changed, 41 insertions(+), 26 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc_resource.c +16 -26
@@ -172,32 +172,22 @@ struct resource_pool *dc_create_resource_pool(struct dc *dc,if (res_pool !=NULL ) {struct dc_firmware_info fw_info = { {0 } }; -if (dc->ctx->dc_bios->funcs->get_firmware_info ( - dc->ctx->dc_bios, &fw_info) ==BP_RESULT_OK ) { - res_pool->ref_clocks.xtalin_clock_inKhz = fw_info.pll_info.crystal_frequency; - -if (IS_FPGA_MAXIMUS_DC (dc->ctx->dce_environment)) { -// On FPGA these dividers are currently not configured by GDB - res_pool->ref_clocks.dccg_ref_clock_inKhz = res_pool->ref_clocks.xtalin_clock_inKhz; - res_pool->ref_clocks.dchub_ref_clock_inKhz = res_pool->ref_clocks.xtalin_clock_inKhz; - }else if (res_pool->dccg && res_pool->hubbub) { -// If DCCG reference frequency cannot be determined (usually means not set to xtalin) then this is a critical error -// as this value must be known for DCHUB programming - (res_pool->dccg->funcs->get_dccg_ref_freq)(res_pool->dccg, - fw_info.pll_info.crystal_frequency, - &res_pool->ref_clocks.dccg_ref_clock_inKhz); - -// Similarly, if DCHUB reference frequency cannot be determined, then it is also a critical error - (res_pool->hubbub->funcs->get_dchub_ref_freq)(res_pool->hubbub, - res_pool->ref_clocks.dccg_ref_clock_inKhz, - &res_pool->ref_clocks.dchub_ref_clock_inKhz); - }else { -// Not all ASICs have DCCG sw component - res_pool->ref_clocks.dccg_ref_clock_inKhz = res_pool->ref_clocks.xtalin_clock_inKhz; - res_pool->ref_clocks.dchub_ref_clock_inKhz = res_pool->ref_clocks.xtalin_clock_inKhz; - } - }else -ASSERT_CRITICAL (false); +if (dc->ctx->dc_bios->funcs->get_firmware_info (dc->ctx->dc_bios, + &fw_info) ==BP_RESULT_OK ) { + res_pool->ref_clocks.xtalin_clock_inKhz = + fw_info.pll_info.crystal_frequency; +/* initialize with firmware data first, no all +* ASIC have DCCG SW component. FPGA or +* simulation need initialization of +* dccg_ref_clock_inKhz, dchub_ref_clock_inKhz +* with xtalin_clock_inKhz +*/ + res_pool->ref_clocks.dccg_ref_clock_inKhz = + res_pool->ref_clocks.xtalin_clock_inKhz; + res_pool->ref_clocks.dchub_ref_clock_inKhz = + res_pool->ref_clocks.xtalin_clock_inKhz; + }else +ASSERT_CRITICAL (false); }return res_pool;
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c +25 -0
@@ -562,6 +562,7 @@ static void dcn20_init_hw(struct dc *dc)struct dc_bios *dcb = dc->ctx->dc_bios;struct resource_pool *res_pool = dc->res_pool;struct dc_state *context = dc->current_state; +struct dc_firmware_info fw_info = { {0 } };if (dc->clk_mgr && dc->clk_mgr->funcs->init_clocks) dc->clk_mgr->funcs->init_clocks (dc->clk_mgr);
@@ -585,6 +586,30 @@ static void dcn20_init_hw(struct dc *dc) }else {if (!dcb->funcs->is_accelerated_mode (dcb)) {bios_golden_init (dc); +if (dc->ctx->dc_bios->funcs->get_firmware_info ( + dc->ctx->dc_bios, &fw_info) ==BP_RESULT_OK ) { + res_pool->ref_clocks.xtalin_clock_inKhz = fw_info.pll_info.crystal_frequency; + +if (!IS_FPGA_MAXIMUS_DC (dc->ctx->dce_environment)) { +if (res_pool->dccg && res_pool->hubbub) { + + (res_pool->dccg->funcs->get_dccg_ref_freq)(res_pool->dccg, + fw_info.pll_info.crystal_frequency, + &res_pool->ref_clocks.dccg_ref_clock_inKhz); + + (res_pool->hubbub->funcs->get_dchub_ref_freq)(res_pool->hubbub, + res_pool->ref_clocks.dccg_ref_clock_inKhz, + &res_pool->ref_clocks.dchub_ref_clock_inKhz); + }else { +// Not all ASICs have DCCG sw component + res_pool->ref_clocks.dccg_ref_clock_inKhz = + res_pool->ref_clocks.xtalin_clock_inKhz; + res_pool->ref_clocks.dchub_ref_clock_inKhz = + res_pool->ref_clocks.xtalin_clock_inKhz; + } + } + }else +ASSERT_CRITICAL (false);disable_vga (dc->hwseq); }
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 66/87 drm/amd/display: add a option to force the clock at every mode change.
In reply to a message not in the archiveFrom: Charlene Liu <[email protected]>
[Description]
This is for HW negative stress testing use.
force reset the dispclk and dppclk even the same clock already set in HW.
--- .../gpu/drm/amd/display/dc/clk_mgr/dcn20/dcn20_clk_mgr.c | 9 ++++++++- drivers/gpu/drm/amd/display/dc/dc.h | 1 + 2 files changed, 9 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/clk_mgr/dcn20/dcn20_clk_mgr.c +8 -1
@@ -151,7 +151,14 @@ void dcn2_update_clocks(struct clk_mgr *clk_mgr_base,bool enter_display_off = false;bool dpp_clock_lowered = false;struct dmcu *dmcu = clk_mgr_base->ctx->dc->res_pool->dmcu; +bool force_reset = false; +if (clk_mgr_base->clks.dispclk_khz ==0 || + dc->debug.force_clock_mode &0x1 ) { +//this is from resume or boot up, if forced_clock cfg option used, we bypass program dispclk and DPPCLK, but need set them for S3. + force_reset = true; +//force_clock_mode 0x1: force reset the clock even it is the same clock as long as it is in Passive level. + } display_count =clk_mgr_helper_get_active_display_cnt (dc, context);if (dc->res_pool->pp_smu) pp_smu = &dc->res_pool->pp_smu->nv_funcs;
@@ -223,7 +230,7 @@ void dcn2_update_clocks(struct clk_mgr *clk_mgr_base, update_dispclk = true; } -if (dc->config.forced_clocks == false) { +if (dc->config.forced_clocks == false || (force_reset && safe_to_lower)) {if (dpp_clock_lowered) {// if clock is being lowered, increase DTO before lowering refclk dcn20_update_clocks_update_dpp_dto (clk_mgr, context);
drivers/gpu/drm/amd/display/dc/dc.h +1 -0
@@ -385,6 +385,7 @@ struct dc_debug_options {#if defined(CONFIG_DRM_AMD_DC_DCN2_0 )bool cm_in_bypass;#endif +int force_clock_mode;/*every mode change.*/ };struct dc_debug_data {
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 67/87 drm/amd/display: use min disp and dpp clk debug option for dcn2
In reply to a message not in the archiveFrom: Dmytro Laktyushkin <[email protected]>
This allows to set a minimum display and dpp clock on dcn2+ HW by adjusting clocks used for dml calculations.
--- drivers/gpu/drm/amd/display/dc/dc.h | 1 + drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c | 8 ++++---- 2 files changed, 5 insertions(+), 4 deletions(-)
drivers/gpu/drm/amd/display/dc/dc.h +1 -0
@@ -344,6 +344,7 @@ struct dc_debug_options {bool disable_pplib_wm_range;enum wm_report_mode pplib_wm_report_mode;unsigned int min_disp_clk_khz; +unsigned int min_dpp_clk_khz;int sr_exit_time_dpm0_ns;int sr_enter_plus_exit_time_dpm0_ns;int sr_exit_time_ns;
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c +4 -4
@@ -2168,10 +2168,6 @@ bool dcn20_fast_validate_bw( }if (force_split && context->bw_ctx.dml.vba.NoOfDPP[vlevel][context->bw_ctx.dml.vba.maxMpcComb][pipe_idx] ==1 ) context->bw_ctx.dml.vba.RequiredDPPCLK[vlevel][context->bw_ctx.dml.vba.maxMpcComb][pipe_idx] /=2 ; -if (dc->config.forced_clocks == true) { - context->bw_ctx.dml.vba.RequiredDPPCLK[vlevel][context->bw_ctx.dml.vba.maxMpcComb][pipe_idx] = - context->bw_ctx.dml.soc.clock_limits[0 ].dppclk_mhz; - }if (!pipe->top_pipe && !pipe->plane_state && context->bw_ctx.dml.vba.ODMCombineEnabled[pipe_idx]) { hsplit_pipe =find_idle_secondary_pipe (&context->res_ctx, dc->res_pool, pipe);ASSERT (hsplit_pipe);
@@ -2291,6 +2287,10 @@ void dcn20_calculate_wm( pipes[pipe_cnt].clks_cfg.dispclk_mhz = context->bw_ctx.dml.soc.clock_limits[0 ].dispclk_mhz; pipes[pipe_cnt].clks_cfg.dppclk_mhz = context->bw_ctx.dml.soc.clock_limits[0 ].dppclk_mhz; } +if (dc->debug.min_disp_clk_khz > pipes[pipe_cnt].clks_cfg.dispclk_mhz *1000 ) + pipes[pipe_cnt].clks_cfg.dispclk_mhz = dc->debug.min_disp_clk_khz /1000.0 ; +if (dc->debug.min_dpp_clk_khz > pipes[pipe_cnt].clks_cfg.dppclk_mhz *1000 ) + pipes[pipe_cnt].clks_cfg.dppclk_mhz = dc->debug.min_dpp_clk_khz /1000.0 ; pipe_cnt++; }
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 68/87 drm/amd/display: add pipe CRC sources without disabling dithering.
In reply to a message not in the archiveFrom: Dingchen Zhang <[email protected]>
[Why]
need to verify the impact of spatial dithering on 8bpc bypass mode.
[How]
added CRC sources and configure dihter option from dc stream.
--- .../drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c | 58 ++++++++++++++----- .../drm/amd/display/amdgpu_dm/amdgpu_dm_crc.h | 6 +- 2 files changed, 48 insertions(+), 16 deletions(-)
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c +44 -14
@@ -32,7 +32,9 @@static const char *const pipe_crc_sources[] = {"none" ,"crtc" , +"crtc dither" ,"dprx" , +"dprx dither" ,"auto" , };
@@ -44,10 +46,33 @@ static enum amdgpu_dm_pipe_crc_source dm_parse_crc_source(const char *source)return AMDGPU_DM_PIPE_CRC_SOURCE_CRTC ;if (!strcmp (source,"dprx" ))return AMDGPU_DM_PIPE_CRC_SOURCE_DPRX ; +if (!strcmp (source,"crtc dither" )) +return AMDGPU_DM_PIPE_CRC_SOURCE_CRTC_DITHER ; +if (!strcmp (source,"dprx dither" )) +return AMDGPU_DM_PIPE_CRC_SOURCE_DPRX_DITHER ;return AMDGPU_DM_PIPE_CRC_SOURCE_INVALID ; } +static bool dm_is_crc_source_crtc (enum amdgpu_dm_pipe_crc_source src) +{ +return (src ==AMDGPU_DM_PIPE_CRC_SOURCE_CRTC ) || + (src ==AMDGPU_DM_PIPE_CRC_SOURCE_CRTC_DITHER ); +} + +static bool dm_is_crc_source_dprx (enum amdgpu_dm_pipe_crc_source src) +{ +return (src ==AMDGPU_DM_PIPE_CRC_SOURCE_DPRX ) || + (src ==AMDGPU_DM_PIPE_CRC_SOURCE_DPRX_DITHER ); +} + +static bool dm_need_crc_dither (enum amdgpu_dm_pipe_crc_source src) +{ +return (src ==AMDGPU_DM_PIPE_CRC_SOURCE_CRTC_DITHER ) || + (src ==AMDGPU_DM_PIPE_CRC_SOURCE_DPRX_DITHER ) || + (src ==AMDGPU_DM_PIPE_CRC_SOURCE_NONE ); +} +const char *const *amdgpu_dm_crtc_get_crc_sources (struct drm_crtc *crtc,size_t *count) {
@@ -101,14 +126,18 @@ int amdgpu_dm_crtc_set_crc_source(struct drm_crtc *crtc, const char *src_name)* USER REQ SRC | CURRENT SRC | BEHAVIOR * ----------------------------- * None | None | Do nothing -* None | CRTC | Disable CRTC CRC -* None | DPRX | Disable DPRX CRC, need 'aux' -* CRTC | XXXX | Enable CRTC CRC, configure DC strm -* DPRX | XXXX | Enable DPRX CRC, need 'aux' +* None | CRTC | Disable CRTC CRC, set default to dither +* None | DPRX | Disable DPRX CRC, need 'aux', set default to dither +* None | CRTC DITHER | Disable CRTC CRC +* None | DPRX DITHER | Disable DPRX CRC, need 'aux' +* CRTC | XXXX | Enable CRTC CRC, no dither +* DPRX | XXXX | Enable DPRX CRC, need 'aux', no dither +* CRTC DITHER | XXXX | Enable CRTC CRC, set dither +* DPRX DITHER | XXXX | Enable DPRX CRC, need 'aux', set dither */ -if (source ==AMDGPU_DM_PIPE_CRC_SOURCE_DPRX || +if (dm_is_crc_source_dprx (source) || (source ==AMDGPU_DM_PIPE_CRC_SOURCE_NONE && - crtc_state->crc_src ==AMDGPU_DM_PIPE_CRC_SOURCE_DPRX )) { +dm_is_crc_source_dprx (crtc_state->crc_src))) { aconn = stream_state->link->priv;if (!aconn) {
@@ -124,7 +153,7 @@ int amdgpu_dm_crtc_set_crc_source(struct drm_crtc *crtc, const char *src_name)mutex_unlock (&adev->dm.dc_lock);return -EINVAL ; } - }else if (source ==AMDGPU_DM_PIPE_CRC_SOURCE_CRTC ) { + }else if (dm_is_crc_source_crtc (source)) {if (!dc_stream_configure_crc (stream_state->ctx->dc, stream_state, enable, enable)) {mutex_unlock (&adev->dm.dc_lock);
@@ -132,10 +161,11 @@ int amdgpu_dm_crtc_set_crc_source(struct drm_crtc *crtc, const char *src_name) } } -/* When enabling CRC, we should also disable dithering. */ -dc_stream_set_dither_option (stream_state, - enable ?DITHER_OPTION_TRUN8 - :DITHER_OPTION_DEFAULT ); +/* configure dithering */ +if (!dm_need_crc_dither (source)) +dc_stream_set_dither_option (stream_state,DITHER_OPTION_TRUN8 ); +else if (!dm_need_crc_dither (crtc_state->crc_src)) +dc_stream_set_dither_option (stream_state,DITHER_OPTION_DEFAULT );mutex_unlock (&adev->dm.dc_lock);
@@ -146,7 +176,7 @@ int amdgpu_dm_crtc_set_crc_source(struct drm_crtc *crtc, const char *src_name) enabled =amdgpu_dm_is_valid_crc_source (crtc_state->crc_src);if (!enabled && enable) {drm_crtc_vblank_get (crtc); -if (source ==AMDGPU_DM_PIPE_CRC_SOURCE_DPRX ) { +if (dm_is_crc_source_dprx (source)) {if (drm_dp_start_crc (aux, crtc)) {DRM_DEBUG_DRIVER ("dp start crc failed\n" );return -EINVAL ;
@@ -154,7 +184,7 @@ int amdgpu_dm_crtc_set_crc_source(struct drm_crtc *crtc, const char *src_name) } }else if (enabled && !enable) {drm_crtc_vblank_put (crtc); -if (crtc_state->crc_src ==AMDGPU_DM_PIPE_CRC_SOURCE_DPRX ) { +if (dm_is_crc_source_dprx (source)) {if (drm_dp_stop_crc (aux)) {DRM_DEBUG_DRIVER ("dp stop crc failed\n" );return -EINVAL ;
@@ -203,7 +233,7 @@ void amdgpu_dm_crtc_handle_crc_irq(struct drm_crtc *crtc)return ; } -if (crtc_state->crc_src ==AMDGPU_DM_PIPE_CRC_SOURCE_CRTC ) { +if (dm_is_crc_source_crtc (crtc_state->crc_src)) {if (!dc_stream_get_crc (stream_state->ctx->dc, stream_state, &crcs[0 ], &crcs[1 ], &crcs[2 ]))return ;
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.h +4 -2
@@ -29,15 +29,17 @@enum amdgpu_dm_pipe_crc_source {AMDGPU_DM_PIPE_CRC_SOURCE_NONE =0 ,AMDGPU_DM_PIPE_CRC_SOURCE_CRTC , +AMDGPU_DM_PIPE_CRC_SOURCE_CRTC_DITHER ,AMDGPU_DM_PIPE_CRC_SOURCE_DPRX , +AMDGPU_DM_PIPE_CRC_SOURCE_DPRX_DITHER ,AMDGPU_DM_PIPE_CRC_SOURCE_MAX ,AMDGPU_DM_PIPE_CRC_SOURCE_INVALID =-1 , };static inline bool amdgpu_dm_is_valid_crc_source (enum amdgpu_dm_pipe_crc_source source) { -return (source ==AMDGPU_DM_PIPE_CRC_SOURCE_CRTC ) || - (source ==AMDGPU_DM_PIPE_CRC_SOURCE_DPRX ); +return (source >AMDGPU_DM_PIPE_CRC_SOURCE_NONE ) && + (source <AMDGPU_DM_PIPE_CRC_SOURCE_MAX ); }/* amdgpu_dm_crc.c */
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 69/87 drm/amd/display: Implement DAL3 GPU Integer Scaling
In reply to a message not in the archiveFrom: Reza Amini <[email protected]>
[WHY]
Users want to not have filtering when scaling by integer multiples to native timing.
[HOW]
If timing is a multiple integer of view, we set number of taps to 1 (effectivly closest neighbour).
--- .../gpu/drm/amd/display/dc/core/dc_resource.c | 16 ++++++++++++++++ drivers/gpu/drm/amd/display/dc/dc_hw_types.h | 1 + 2 files changed, 17 insertions(+)
drivers/gpu/drm/amd/display/dc/core/dc_resource.c +16 -0
@@ -937,7 +937,14 @@ static void calculate_inits_and_adj_vp(struct pipe_ctx *pipe_ctx) data->inits.v_c_bot =dc_fixpt_add (data->inits.v_c, data->ratios.vert_c); } +static bool are_rect_integer_multiples (struct rect src,struct rect dest) +{ +if (dest.width >= src.width && dest.width % src.width ==0 && + dest.height >= src.height && dest.height % src.height ==0 ) +return true; +return false; +}bool resource_build_scaling_params (struct pipe_ctx *pipe_ctx) {const struct dc_plane_state *plane_state = pipe_ctx->plane_state;
@@ -980,6 +987,15 @@ bool resource_build_scaling_params(struct pipe_ctx *pipe_ctx)if (pipe_ctx->plane_res.dpp !=NULL ) res = pipe_ctx->plane_res.dpp->funcs->dpp_get_optimal_number_of_taps ( pipe_ctx->plane_res.dpp, &pipe_ctx->plane_res.scl_data, &plane_state->scaling_quality); + +if (res && + plane_state->scaling_quality.integer_scaling && +are_rect_integer_multiples (pipe_ctx->plane_res.scl_data.viewport, + pipe_ctx->plane_res.scl_data.recout)) { + pipe_ctx->plane_res.scl_data.taps.v_taps =1 ; + pipe_ctx->plane_res.scl_data.taps.h_taps =1 ; + } +if (!res) {/* Try 24 bpp linebuffer */ pipe_ctx->plane_res.scl_data.lb_params.depth =LB_PIXEL_DEPTH_24BPP ;
drivers/gpu/drm/amd/display/dc/dc_hw_types.h +1 -0
@@ -614,6 +614,7 @@ struct scaling_taps {uint32_t h_taps;uint32_t v_taps_c;uint32_t h_taps_c; +bool integer_scaling; };enum dc_timing_standard {
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 70/87 drm/amd/display: add dcc programming for dual plane
In reply to a message not in the archiveFrom: Dmytro Laktyushkin <[email protected]>
Add dual plane dcc programming support for
surfaces.
Removes unions from plane size and dcc params as they serve no practical purpose only making our code more convoluted. This results in easy dual plane dcc and surface size programming.
--- .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 42 +++++++-------- .../gpu/drm/amd/display/dc/calcs/dce_calcs.c | 2 +- .../gpu/drm/amd/display/dc/calcs/dcn_calcs.c | 2 +- drivers/gpu/drm/amd/display/dc/core/dc.c | 10 ++-- .../gpu/drm/amd/display/dc/core/dc_debug.c | 40 +++++++------- drivers/gpu/drm/amd/display/dc/dc.h | 4 +- drivers/gpu/drm/amd/display/dc/dc_hw_types.h | 54 +++++++++---------- .../drm/amd/display/dc/dce/dce_mem_input.c | 10 ++-- .../display/dc/dce110/dce110_mem_input_v.c | 42 +++++++-------- .../gpu/drm/amd/display/dc/dcn10/dcn10_hubp.c | 20 +++---- .../gpu/drm/amd/display/dc/dcn10/dcn10_hubp.h | 4 +- .../amd/display/dc/dcn10/dcn10_hw_sequencer.c | 4 +- .../gpu/drm/amd/display/dc/dcn20/dcn20_hubp.c | 20 +++---- .../gpu/drm/amd/display/dc/dcn20/dcn20_hubp.h | 4 +- .../drm/amd/display/dc/dcn20/dcn20_resource.c | 12 ++--- drivers/gpu/drm/amd/display/dc/inc/hw/hubp.h | 2 +- .../gpu/drm/amd/display/dc/inc/hw/mem_input.h | 2 +- 17 files changed, 136 insertions(+), 138 deletions(-)
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +21 -21
@@ -2743,7 +2743,7 @@ fill_plane_dcc_attributes(struct amdgpu_device *adev,const struct amdgpu_framebuffer *afb, const enum surface_pixel_format format, const enum dc_rotation_angle rotation, - const union plane_size *plane_size, + const struct plane_size *plane_size, const union dc_tiling_info *tiling_info, const uint64_t info, struct dc_plane_dcc_param *dcc,
@@ -2769,8 +2769,8 @@ fill_plane_dcc_attributes(struct amdgpu_device *adev,return -EINVAL ; input.format = format; - input.surface_size.width = plane_size->grph.surface_size.width; - input.surface_size.height = plane_size->grph.surface_size.height; + input.surface_size.width = plane_size->surface_size.width; + input.surface_size.height = plane_size->surface_size.height; input.swizzle_mode = tiling_info->gfx9.swizzle;if (rotation ==ROTATION_ANGLE_0 || rotation ==ROTATION_ANGLE_180 )
@@ -2788,9 +2788,9 @@ fill_plane_dcc_attributes(struct amdgpu_device *adev,return -EINVAL ; dcc->enable =1 ; - dcc->grph.meta_pitch = + dcc->meta_pitch =AMDGPU_TILING_GET (info,DCC_PITCH_MAX ) +1 ; - dcc->grph.independent_64b_blks = i64b; + dcc->independent_64b_blks = i64b; dcc_address =get_dcc_address (afb->address, info); address->grph.meta_addr.low_part =lower_32_bits (dcc_address);
@@ -2806,7 +2806,7 @@ fill_plane_buffer_attributes(struct amdgpu_device *adev,const enum dc_rotation_angle rotation, const uint64_t tiling_flags, union dc_tiling_info *tiling_info, - union plane_size *plane_size, + struct plane_size *plane_size, struct dc_plane_dcc_param *dcc, struct dc_plane_address *address) {
@@ -2819,11 +2819,11 @@ fill_plane_buffer_attributes(struct amdgpu_device *adev,memset (address,0 ,sizeof (*address));if (format <SURFACE_PIXEL_FORMAT_VIDEO_BEGIN ) { - plane_size->grph.surface_size.x =0 ; - plane_size->grph.surface_size.y =0 ; - plane_size->grph.surface_size.width = fb->width; - plane_size->grph.surface_size.height = fb->height; - plane_size->grph.surface_pitch = + plane_size->surface_size.x =0 ; + plane_size->surface_size.y =0 ; + plane_size->surface_size.width = fb->width; + plane_size->surface_size.height = fb->height; + plane_size->surface_pitch = fb->pitches[0 ] / fb->format->cpp[0 ]; address->type =PLN_ADDR_TYPE_GRAPHICS ;
@@ -2832,20 +2832,20 @@ fill_plane_buffer_attributes(struct amdgpu_device *adev, }else if (format <SURFACE_PIXEL_FORMAT_INVALID ) {uint64_t chroma_addr = afb->address + fb->offsets[1 ]; - plane_size->video.luma_size.x =0 ; - plane_size->video.luma_size.y =0 ; - plane_size->video.luma_size.width = fb->width; - plane_size->video.luma_size.height = fb->height; - plane_size->video.luma_pitch = + plane_size->surface_size.x =0 ; + plane_size->surface_size.y =0 ; + plane_size->surface_size.width = fb->width; + plane_size->surface_size.height = fb->height; + plane_size->surface_pitch = fb->pitches[0 ] / fb->format->cpp[0 ]; - plane_size->video.chroma_size.x =0 ; - plane_size->video.chroma_size.y =0 ; + plane_size->chroma_size.x =0 ; + plane_size->chroma_size.y =0 ;/* TODO: set these based on surface format */ - plane_size->video.chroma_size.width = fb->width /2 ; - plane_size->video.chroma_size.height = fb->height /2 ; + plane_size->chroma_size.width = fb->width /2 ; + plane_size->chroma_size.height = fb->height /2 ; - plane_size->video.chroma_pitch = + plane_size->chroma_pitch = fb->pitches[1 ] / fb->format->cpp[1 ]; address->type =PLN_ADDR_TYPE_VIDEO_PROGRESSIVE ;
drivers/gpu/drm/amd/display/dc/calcs/dce_calcs.c +1 -1
@@ -2850,7 +2850,7 @@ static void populate_initial_data( data->src_height[num_displays *2 + j] =bw_int_to_fixed (pipe[i].bottom_pipe->plane_res.scl_data.viewport.height); data->src_width[num_displays *2 + j] =bw_int_to_fixed (pipe[i].bottom_pipe->plane_res.scl_data.viewport.width); data->pitch_in_pixels[num_displays *2 + j] =bw_int_to_fixed ( - pipe[i].bottom_pipe->plane_state->plane_size.grph.surface_pitch); + pipe[i].bottom_pipe->plane_state->plane_size.surface_pitch); data->h_taps[num_displays *2 + j] =bw_int_to_fixed (pipe[i].bottom_pipe->plane_res.scl_data.taps.h_taps); data->v_taps[num_displays *2 + j] =bw_int_to_fixed (pipe[i].bottom_pipe->plane_res.scl_data.taps.v_taps); data->h_scale_ratio[num_displays * 2 + j] = fixed31_32_to_bw_fixed(
drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c +1 -1
@@ -329,7 +329,7 @@ static void pipe_ctx_to_e2e_pipe_params (dcc_support_pixel_format (pipe->plane_state->format, &bpe) ?1 :0 ; } input->src.dcc_rate =1 ; - input->src.meta_pitch = pipe->plane_state->dcc.grph.meta_pitch; + input->src.meta_pitch = pipe->plane_state->dcc.meta_pitch; input->src.source_scan = dm_horz; input->src.sw_mode = pipe->plane_state->tiling_info.gfx9.swizzle;
drivers/gpu/drm/amd/display/dc/core/dc.c +5 -5
@@ -1347,8 +1347,8 @@ static enum surface_update_type get_plane_info_update_type(const struct dc_surfa }if (u->plane_info->dcc.enable != u->surface->dcc.enable - || u->plane_info->dcc.grph.independent_64b_blks != u->surface->dcc.grph.independent_64b_blks - || u->plane_info->dcc.grph.meta_pitch != u->surface->dcc.grph.meta_pitch) { + || u->plane_info->dcc.independent_64b_blks != u->surface->dcc.independent_64b_blks + || u->plane_info->dcc.meta_pitch != u->surface->dcc.meta_pitch) { update_flags->bits.dcc_change =1 ;elevate_update_type (&update_type,UPDATE_TYPE_MED ); }
@@ -1362,9 +1362,9 @@ static enum surface_update_type get_plane_info_update_type(const struct dc_surfaelevate_update_type (&update_type,UPDATE_TYPE_FULL ); } -if (u->plane_info->plane_size.grph.surface_pitch != u->surface->plane_size.grph.surface_pitch - || u->plane_info->plane_size.video.luma_pitch != u->surface->plane_size.video.luma_pitch - || u->plane_info->plane_size.video.chroma_pitch != u->surface->plane_size.video.chroma_pitch) { +if (u->plane_info->plane_size.surface_pitch != u->surface->plane_size.surface_pitch + || u->plane_info->plane_size.surface_pitch != u->surface->plane_size.surface_pitch + || u->plane_info->plane_size.chroma_pitch != u->surface->plane_size.chroma_pitch) { update_flags->bits.plane_size_change =1 ;elevate_update_type (&update_type,UPDATE_TYPE_MED ); }
drivers/gpu/drm/amd/display/dc/core/dc_debug.c +20 -20
@@ -115,16 +115,16 @@ void pre_surface_trace( plane_state->clip_rect.height);SURFACE_TRACE ( -"plane_state->plane_size.grph.surface_size.x = %d;\n" -"plane_state->plane_size.grph.surface_size.y = %d;\n" -"plane_state->plane_size.grph.surface_size.width = %d;\n" -"plane_state->plane_size.grph.surface_size.height = %d;\n" -"plane_state->plane_size.grph.surface_pitch = %d;\n" , - plane_state->plane_size.grph.surface_size.x, - plane_state->plane_size.grph.surface_size.y, - plane_state->plane_size.grph.surface_size.width, - plane_state->plane_size.grph.surface_size.height, - plane_state->plane_size.grph.surface_pitch); +"plane_state->plane_size.surface_size.x = %d;\n" +"plane_state->plane_size.surface_size.y = %d;\n" +"plane_state->plane_size.surface_size.width = %d;\n" +"plane_state->plane_size.surface_size.height = %d;\n" +"plane_state->plane_size.surface_pitch = %d;\n" , + plane_state->plane_size.surface_size.x, + plane_state->plane_size.surface_size.y, + plane_state->plane_size.surface_size.width, + plane_state->plane_size.surface_size.height, + plane_state->plane_size.surface_pitch); SURFACE_TRACE(
@@ -202,20 +202,20 @@ void update_surface_trace(SURFACE_TRACE ("plane_info->color_space = %d;\n" "plane_info->format = %d;\n" -"plane_info->plane_size.grph.surface_pitch = %d;\n" -"plane_info->plane_size.grph.surface_size.height = %d;\n" -"plane_info->plane_size.grph.surface_size.width = %d;\n" -"plane_info->plane_size.grph.surface_size.x = %d;\n" -"plane_info->plane_size.grph.surface_size.y = %d;\n" +"plane_info->plane_size.surface_pitch = %d;\n" +"plane_info->plane_size.surface_size.height = %d;\n" +"plane_info->plane_size.surface_size.width = %d;\n" +"plane_info->plane_size.surface_size.x = %d;\n" +"plane_info->plane_size.surface_size.y = %d;\n" "plane_info->rotation = %d;\n" "plane_info->stereo_format = %d;\n" , update->plane_info->color_space, update->plane_info->format, - update->plane_info->plane_size.grph.surface_pitch, - update->plane_info->plane_size.grph.surface_size.height, - update->plane_info->plane_size.grph.surface_size.width, - update->plane_info->plane_size.grph.surface_size.x, - update->plane_info->plane_size.grph.surface_size.y, + update->plane_info->plane_size.surface_pitch, + update->plane_info->plane_size.surface_size.height, + update->plane_info->plane_size.surface_size.width, + update->plane_info->plane_size.surface_size.x, + update->plane_info->plane_size.surface_size.y, update->plane_info->rotation, update->plane_info->stereo_format);
drivers/gpu/drm/amd/display/dc/dc.h +2 -2
@@ -705,7 +705,7 @@ struct dc_plane_state {struct rect dst_rect;struct rect clip_rect; -union plane_size plane_size; +struct plane_size plane_size;union dc_tiling_info tiling_info;struct dc_plane_dcc_param dcc;
@@ -754,7 +754,7 @@ struct dc_plane_state { };struct dc_plane_info { -union plane_size plane_size; +struct plane_size plane_size;union dc_tiling_info tiling_info;struct dc_plane_dcc_param dcc;enum surface_pixel_format format;
drivers/gpu/drm/amd/display/dc/dc_hw_types.h +26 -28
68 lines not on this page · raw message
drivers/gpu/drm/amd/display/dc/dce/dce_mem_input.c +5 -5
@@ -391,10 +391,10 @@ static void program_tiling(static void program_size_and_rotation (struct dce_mem_input *dce_mi,enum dc_rotation_angle rotation, -const union plane_size *plane_size) +const struct plane_size *plane_size) { -const struct rect *in_rect = &plane_size->grph.surface_size; -struct rect hw_rect = plane_size->grph.surface_size; +const struct rect *in_rect = &plane_size->surface_size; +struct rect hw_rect = plane_size->surface_size;const uint32_t rotation_angles[ROTATION_ANGLE_COUNT ] = { [ROTATION_ANGLE_0 ] =0 , [ROTATION_ANGLE_90 ] =1 ,
@@ -423,7 +423,7 @@ static void program_size_and_rotation( GRPH_Y_END, hw_rect.height);REG_SET (GRPH_PITCH ,0 , -GRPH_PITCH , plane_size->grph.surface_pitch); +GRPH_PITCH , plane_size->surface_pitch);REG_SET (HW_ROTATION ,0 ,GRPH_ROTATION_ANGLE , rotation_angles[rotation]);
@@ -505,7 +505,7 @@ static void dce_mi_program_surface_config(struct mem_input *mi, enum surface_pixel_format format, union dc_tiling_info *tiling_info, - union plane_size *plane_size, + struct plane_size *plane_size, enum dc_rotation_angle rotation, struct dc_plane_dcc_param *dcc, bool horizontal_mirror)
drivers/gpu/drm/amd/display/dc/dce110/dce110_mem_input_v.c +21 -21
93 lines not on this page · raw message
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.c +10 -10
@@ -163,7 +163,7 @@ void hubp1_program_tiling(void hubp1_program_size (struct hubp *hubp,enum surface_pixel_format format, -const union plane_size *plane_size, +const struct plane_size *plane_size,struct dc_plane_dcc_param *dcc) {struct dcn10_hubp *hubp1 =TO_DCN10_HUBP (hubp);
@@ -173,16 +173,16 @@ void hubp1_program_size(* 444 or 420 luma */ if (format >=SURFACE_PIXEL_FORMAT_VIDEO_BEGIN && format <SURFACE_PIXEL_FORMAT_SUBSAMPLE_END ) { -ASSERT (plane_size->video.chroma_pitch !=0 ); +ASSERT (plane_size->chroma_pitch !=0 );/* Chroma pitch zero can cause system hang! */ - pitch = plane_size->video.luma_pitch -1 ; - meta_pitch = dcc->video.meta_pitch_l -1 ; - pitch_c = plane_size->video.chroma_pitch -1 ; - meta_pitch_c = dcc->video.meta_pitch_c -1 ; + pitch = plane_size->surface_pitch -1 ; + meta_pitch = dcc->meta_pitch -1 ; + pitch_c = plane_size->chroma_pitch -1 ; + meta_pitch_c = dcc->meta_pitch_c -1 ; }else { - pitch = plane_size->grph.surface_pitch -1 ; - meta_pitch = dcc->grph.meta_pitch -1 ; + pitch = plane_size->surface_pitch -1 ; + meta_pitch = dcc->meta_pitch -1 ; pitch_c =0 ; meta_pitch_c =0 ; }
@@ -526,13 +526,13 @@ void hubp1_program_surface_config(struct hubp *hubp, enum surface_pixel_format format, union dc_tiling_info *tiling_info, - union plane_size *plane_size, + struct plane_size *plane_size, enum dc_rotation_angle rotation, struct dc_plane_dcc_param *dcc, bool horizontal_mirror, unsigned int compat_level) { - hubp1_dcc_control(hubp, dcc->enable, dcc->grph.independent_64b_blks); + hubp1_dcc_control(hubp, dcc->enable, dcc->independent_64b_blks);hubp1_program_tiling (hubp, tiling_info, format);hubp1_program_size (hubp, format, plane_size, dcc);hubp1_program_rotation (hubp, rotation, horizontal_mirror);
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.h +2 -2
@@ -685,7 +685,7 @@ void hubp1_program_surface_config(struct hubp *hubp, enum surface_pixel_format format, union dc_tiling_info *tiling_info, - union plane_size *plane_size, + struct plane_size *plane_size, enum dc_rotation_angle rotation, struct dc_plane_dcc_param *dcc, bool horizontal_mirror,
@@ -707,7 +707,7 @@ void hubp1_program_pixel_format(void hubp1_program_size (struct hubp *hubp,enum surface_pixel_format format, -const union plane_size *plane_size, +const struct plane_size *plane_size,struct dc_plane_dcc_param *dcc); void hubp1_program_rotation(
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c +2 -2
@@ -2131,7 +2131,7 @@ void update_dchubp_dpp(struct hubp *hubp = pipe_ctx->plane_res.hubp;struct dpp *dpp = pipe_ctx->plane_res.dpp;struct dc_plane_state *plane_state = pipe_ctx->plane_state; -union plane_size size = plane_state->plane_size; +struct plane_size size = plane_state->plane_size;unsigned int compat_level =0 ;/* depends on DML calculation, DPP clock value may change dynamically */
@@ -2177,7 +2177,7 @@ void update_dchubp_dpp( &pipe_ctx->ttu_regs); } - size.grph.surface_size = pipe_ctx->plane_res.scl_data.viewport; + size.surface_size = pipe_ctx->plane_res.scl_data.viewport;if (plane_state->update_flags.bits.full_update || plane_state->update_flags.bits.bpp_change)
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hubp.c +10 -10
@@ -397,7 +397,7 @@ static void hubp2_program_tiling(void hubp2_program_size (struct hubp *hubp,enum surface_pixel_format format, -const union plane_size *plane_size, +const struct plane_size *plane_size,struct dc_plane_dcc_param *dcc) {struct dcn20_hubp *hubp2 =TO_DCN20_HUBP (hubp);
@@ -410,16 +410,16 @@ void hubp2_program_size( use_pitch_c = format >=SURFACE_PIXEL_FORMAT_VIDEO_BEGIN && format <SURFACE_PIXEL_FORMAT_SUBSAMPLE_END ;if (use_pitch_c) { -ASSERT (plane_size->video.chroma_pitch !=0 ); +ASSERT (plane_size->chroma_pitch !=0 );/* Chroma pitch zero can cause system hang! */ - pitch = plane_size->video.luma_pitch -1 ; - meta_pitch = dcc->video.meta_pitch_l -1 ; - pitch_c = plane_size->video.chroma_pitch -1 ; - meta_pitch_c = dcc->video.meta_pitch_c -1 ; + pitch = plane_size->surface_pitch -1 ; + meta_pitch = dcc->meta_pitch -1 ; + pitch_c = plane_size->chroma_pitch -1 ; + meta_pitch_c = dcc->meta_pitch_c -1 ; }else { - pitch = plane_size->grph.surface_pitch -1 ; - meta_pitch = dcc->grph.meta_pitch -1 ; + pitch = plane_size->surface_pitch -1 ; + meta_pitch = dcc->meta_pitch -1 ; pitch_c =0 ; meta_pitch_c =0 ; }
@@ -592,7 +592,7 @@ void hubp2_program_surface_config(struct hubp *hubp, enum surface_pixel_format format, union dc_tiling_info *tiling_info, - union plane_size *plane_size, + struct plane_size *plane_size, enum dc_rotation_angle rotation, struct dc_plane_dcc_param *dcc, bool horizontal_mirror,
@@ -600,7 +600,7 @@ void hubp2_program_surface_config( {struct dcn20_hubp *hubp2 =TO_DCN20_HUBP (hubp); -hubp2_dcc_control (hubp, dcc->enable, dcc->grph.independent_64b_blks); +hubp2_dcc_control (hubp, dcc->enable, dcc->independent_64b_blks);hubp2_program_tiling (hubp2, tiling_info, format);hubp2_program_size (hubp, format, plane_size, dcc);hubp2_program_rotation (hubp, rotation, horizontal_mirror);
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hubp.h +2 -2
@@ -273,7 +273,7 @@ void hubp2_dcc_control(struct hubp *hubp, bool enable,void hubp2_program_size (struct hubp *hubp,enum surface_pixel_format format, -const union plane_size *plane_size, +const struct plane_size *plane_size,struct dc_plane_dcc_param *dcc); void hubp2_program_rotation(
@@ -289,7 +289,7 @@ void hubp2_program_surface_config(struct hubp *hubp, enum surface_pixel_format format, union dc_tiling_info *tiling_info, - union plane_size *plane_size, + struct plane_size *plane_size, enum dc_rotation_angle rotation, struct dc_plane_dcc_param *dcc, bool horizontal_mirror,
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c +6 -6
@@ -1820,13 +1820,13 @@ int dcn20_populate_dml_pipes_from_context( pipes[pipe_cnt].pipe.src.viewport_height = scl->viewport.height; pipes[pipe_cnt].pipe.src.viewport_height_c = scl->viewport_c.height;if (pln->format >=SURFACE_PIXEL_FORMAT_VIDEO_BEGIN ) { - pipes[pipe_cnt].pipe.src.data_pitch = pln->plane_size.video.luma_pitch; - pipes[pipe_cnt].pipe.src.data_pitch_c = pln->plane_size.video.chroma_pitch; - pipes[pipe_cnt].pipe.src.meta_pitch = pln->dcc.video.meta_pitch_l; - pipes[pipe_cnt].pipe.src.meta_pitch_c = pln->dcc.video.meta_pitch_c; + pipes[pipe_cnt].pipe.src.data_pitch = pln->plane_size.surface_pitch; + pipes[pipe_cnt].pipe.src.data_pitch_c = pln->plane_size.chroma_pitch; + pipes[pipe_cnt].pipe.src.meta_pitch = pln->dcc.meta_pitch; + pipes[pipe_cnt].pipe.src.meta_pitch_c = pln->dcc.meta_pitch_c; }else { - pipes[pipe_cnt].pipe.src.data_pitch = pln->plane_size.grph.surface_pitch; - pipes[pipe_cnt].pipe.src.meta_pitch = pln->dcc.grph.meta_pitch; + pipes[pipe_cnt].pipe.src.data_pitch = pln->plane_size.surface_pitch; + pipes[pipe_cnt].pipe.src.meta_pitch = pln->dcc.meta_pitch; } pipes[pipe_cnt].pipe.src.dcc = pln->dcc.enable; pipes[pipe_cnt].pipe.dest.recout_width = scl->recout.width;
drivers/gpu/drm/amd/display/dc/inc/hw/hubp.h +1 -1
@@ -103,7 +103,7 @@ struct hubp_funcs {struct hubp *hubp, enum surface_pixel_format format, union dc_tiling_info *tiling_info, - union plane_size *plane_size, + struct plane_size *plane_size, enum dc_rotation_angle rotation, struct dc_plane_dcc_param *dcc, bool horizontal_mirror,
drivers/gpu/drm/amd/display/dc/inc/hw/mem_input.h +1 -1
@@ -149,7 +149,7 @@ struct mem_input_funcs {struct mem_input *mem_input, enum surface_pixel_format format, union dc_tiling_info *tiling_info, - union plane_size *plane_size, + struct plane_size *plane_size, enum dc_rotation_angle rotation, struct dc_plane_dcc_param *dcc, bool horizontal_mirror);
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 71/87 drm/amd/display: populate last calculated bb state with max clocks
In reply to a message not in the archiveFrom: Jun Lei <[email protected]>
[why]update_bounding_box calculates intermediate bb states based on clock relationship however, the last state doesn't need to maintain a minimum relationship, but should actually contain maximum of every clock. otherwise maximum clocks are not usable
[how]
once the calculated bb is built, override the last state with max values
--- drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c | 4 ++++ 1 file changed, 4 insertions(+)
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c +4 -0
@@ -2746,6 +2746,10 @@ static void update_bounding_box(struct dc *dc, struct _vcs_dpi_soc_bounding_box_ num_calculated_states++; } + calculated_states[num_calculated_states -1 ].socclk_mhz = max_clocks->socClockInKhz /1000 ; + calculated_states[num_calculated_states -1 ].fabricclk_mhz = max_clocks->socClockInKhz /1000 ; + calculated_states[num_calculated_states -1 ].dcfclk_mhz = max_clocks->dcfClockInKhz /1000 ; +memcpy (bb->clock_limits, calculated_states,sizeof (bb->clock_limits)); bb->num_states = num_calculated_states;
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 72/87 drm/amd/display: Fix dc_create failure handling and 666 color depths
In reply to a message not in the archiveFrom: Julian Parkin <[email protected]>
[Why]
It is possible (but very unlikely) that constructing dc fails before current_state is created.
We support 666 color depth in some scenarios, but this isn't handled in get_norm_pix_clk. It uses exactly the same pixel clock as the 888 case.
[How]
Check for non null current_state before destructing.
Add case for 666 color depth to get_norm_pix_clk to avoid assertion.
--- drivers/gpu/drm/amd/display/dc/core/dc.c | 6 ++++-- drivers/gpu/drm/amd/display/dc/core/dc_resource.c | 1 + 2 files changed, 5 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc.c +4 -2
@@ -500,8 +500,10 @@ void dc_stream_set_static_screen_events(struct dc *dc,static void destruct (struct dc *dc) { -dc_release_state (dc->current_state); - dc->current_state =NULL ; +if (dc->current_state) { +dc_release_state (dc->current_state); + dc->current_state =NULL ; + }destroy_links (dc);
drivers/gpu/drm/amd/display/dc/core/dc_resource.c +1 -0
@@ -1852,6 +1852,7 @@ static int get_norm_pix_clk(const struct dc_crtc_timing *timing) pix_clk /=2 ;if (timing->pixel_encoding !=PIXEL_ENCODING_YCBCR422 ) {switch (timing->display_color_depth) { +case COLOR_DEPTH_666 :case COLOR_DEPTH_888 : normalized_pix_clk = pix_clk;break ;
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 73/87 drm/amd/display: Only enable audio if speaker allocation exists
In reply to a message not in the archiveFrom: Alvin Lee <[email protected]>
[Why]
In dm_helpers_parse_edid_caps, there is a corner case where no speakers can be allocated even though the audio mode count is greater than 0. Enabling audio when no speaker allocations exists can cause issues in the video stream.
[How]
Add a check to not enable audio unless one or more speaker allocations exist (since doing this can cause issues in the video stream).
--- drivers/gpu/drm/amd/display/dc/core/dc_resource.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/core/dc_resource.c +1 -1
@@ -1999,7 +1999,7 @@ enum dc_status resource_map_pool_resources(/* TODO: Add check if ASIC support and EDID audio */ if (!stream->converter_disable_audio &&dc_is_audio_capable_signal (pipe_ctx->stream->signal) && - stream->audio_info.mode_count) { + stream->audio_info.mode_count && stream->audio_info.flags.all) { pipe_ctx->stream_res.audio =find_first_free_audio ( &context->res_ctx, pool, pipe_ctx->stream_res.stream_enc->id);
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 74/87 drm/amd/display: Clean up dynamic metadata logic
In reply to a message not in the archiveFrom: Julian Parkin <[email protected]>
[Why]
Code to enable DCN20 dynamic metadata feature is duplicated in two places and was added to DCE110 enable stream.
[How]
Create DCN20 specific enable stream function for clarity, and add a hardware sequencer function to program dynamic metadata to avoid the duplicate code.
--- .../gpu/drm/amd/display/dc/core/dc_stream.c | 27 +++---- .../display/dc/dce110/dce110_hw_sequencer.c | 24 +----- .../drm/amd/display/dc/dcn20/dcn20_hwseq.c | 74 +++++++++++++++++++ .../gpu/drm/amd/display/dc/inc/hw_sequencer.h | 1 + 4 files changed, 86 insertions(+), 40 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc_stream.c +10 -17
@@ -563,6 +563,7 @@ bool dc_stream_get_scanoutpos(const struct dc_stream_state *stream,return ret; } +#if defined(CONFIG_DRM_AMD_DC_DCN2_0 )bool dc_stream_dmdata_status_done (struct dc *dc,struct dc_stream_state *stream) {
@@ -594,6 +595,14 @@ bool dc_stream_set_dynamic_metadata(struct dc *dc,struct hubp *hubp;int i; +/* Dynamic metadata is only supported on HDMI or DP */ +if (!dc_is_hdmi_signal (stream->signal) && !dc_is_dp_signal (stream->signal)) +return false; + +/* Check hardware support */ +if (!dc->hwss.program_dmdata_engine) +return false; +for (i =0 ; i <MAX_PIPES ; i++) { pipe_ctx = &dc->current_state->res_ctx.pipe_ctx[i];if (pipe_ctx->stream == stream)
@@ -609,23 +618,7 @@ bool dc_stream_set_dynamic_metadata(struct dc *dc, pipe_ctx->stream->dmdata_address = attr->address; -if (pipe_ctx->stream_res.stream_enc && - pipe_ctx->stream_res.stream_enc->funcs->set_dynamic_metadata !=NULL ) { -if (pipe_ctx->stream->dmdata_address.quad_part !=0 ) { -/* if using dynamic meta, don't set up generic infopackets */ - pipe_ctx->stream_res.encoder_info_frame.hdrsmd.valid = false; - pipe_ctx->stream_res.stream_enc->funcs->set_dynamic_metadata ( - pipe_ctx->stream_res.stream_enc, - true, pipe_ctx->plane_res.hubp->inst, -dc_is_dp_signal (pipe_ctx->stream->signal) ? - dmdata_dp : dmdata_hdmi); - }else - pipe_ctx->stream_res.stream_enc->funcs->set_dynamic_metadata ( - pipe_ctx->stream_res.stream_enc, - false, pipe_ctx->plane_res.hubp->inst, -dc_is_dp_signal (pipe_ctx->stream->signal) ? - dmdata_dp : dmdata_hdmi); - } + dc->hwss.program_dmdata_engine (pipe_ctx);if (hubp->funcs->dmdata_set_attributes !=NULL && pipe_ctx->stream->dmdata_address.quad_part !=0 ) {
drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c +1 -23
@@ -664,29 +664,7 @@ void dce110_enable_stream(struct pipe_ctx *pipe_ctx) link->link_enc->funcs->connect_dig_be_to_fe (link->link_enc, pipe_ctx->stream_res.stream_enc->id, true); -/* update AVI info frame (HDMI, DP)*/ -/* TODO: FPGA may change to hwss.update_info_frame */ - -#if defined(CONFIG_DRM_AMD_DC_DCN2_0 ) -if (pipe_ctx->stream_res.stream_enc->funcs->set_dynamic_metadata !=NULL && - pipe_ctx->plane_res.hubp !=NULL ) { -if (pipe_ctx->stream->dmdata_address.quad_part !=0 ) { -/* if using dynamic meta, don't set up generic infopackets */ - pipe_ctx->stream_res.encoder_info_frame.hdrsmd.valid = false; - pipe_ctx->stream_res.stream_enc->funcs->set_dynamic_metadata ( - pipe_ctx->stream_res.stream_enc, - true, pipe_ctx->plane_res.hubp->inst, -dc_is_dp_signal (pipe_ctx->stream->signal) ? - dmdata_dp : dmdata_hdmi); - }else - pipe_ctx->stream_res.stream_enc->funcs->set_dynamic_metadata ( - pipe_ctx->stream_res.stream_enc, - false, pipe_ctx->plane_res.hubp->inst, -dc_is_dp_signal (pipe_ctx->stream->signal) ? - dmdata_dp : dmdata_hdmi); - } -#endif -dce110_update_info_frame (pipe_ctx); + link->dc->hwss.update_info_frame (pipe_ctx);/* enable early control to avoid corruption on DP monitor*/ active_total_with_borders =
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c +74 -0
86 lines not on this page · raw message
drivers/gpu/drm/amd/display/dc/inc/hw_sequencer.h +1 -0
@@ -231,6 +231,7 @@ struct hw_sequencer_funcs {bool (*update_bandwidth)(struct dc *dc,struct dc_state *context); +void (*program_dmdata_engine)(struct pipe_ctx *pipe_ctx);bool (*dmdata_status_done)(struct pipe_ctx *pipe_ctx);#endif
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 75/87 drm/amd/display: Set enabled to false at start of audio disable
In reply to a message not in the archiveFrom: Nicholas Kazlauskas <[email protected]>
[Why]
In an effort to stop redundant calls to dce110_disable_audio_stream the audio->enabled flag was added to the audio resource struct. While this state probably shouldn't have been tracked on the audio struct itself it still works fine for some sequences.
However, it does not work for cases where we're freeing the audio resource (such as hotplugs) or when dynamic audio is enabled.
In these cases the pipe_ctx->stream_res.audio = NULL before we can set audio->enabled = false. The next time we acquire the audio resource such as on hotplug the audio will not be enabled for the stream since DC thinks it's still enabled.
Audio state tracking should cover this sequence.
[How]
Set audio->enabled = false at the start as long as we have pipe_ctx->stream_res.audio.
--- drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c +2 -2
@@ -996,6 +996,8 @@ void dce110_disable_audio_stream(struct pipe_ctx *pipe_ctx, int option) pipe_ctx->stream_res.stream_enc->funcs->audio_mute_control ( pipe_ctx->stream_res.stream_enc, true);if (pipe_ctx->stream_res.audio) { + pipe_ctx->stream_res.audio->enabled = false; +if (dc->res_pool->pp_smu) pp_smu = dc->res_pool->pp_smu;
@@ -1026,8 +1028,6 @@ void dce110_disable_audio_stream(struct pipe_ctx *pipe_ctx, int option)/* dal_audio_disable_azalia_audio_jack_presence(stream->audio, * stream->stream_engine_id); */ -if (pipe_ctx->stream_res.audio) - pipe_ctx->stream_res.audio->enabled = false; } }
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 76/87 drm/amd/display: drop ASSERT() if eDP panel is not connected
In reply to a message not in the archiveFrom: Zhan Liu <[email protected]>
[Why]
For boards that support eDP but do not have a physical eDP display connected an ASSERT will be thrown. This is not a critical failure and shouldn't be treated as such.
[How]
Drop the assertion.
--- drivers/gpu/drm/amd/display/dc/core/dc_link.c | 2 -- 1 file changed, 2 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc_link.c +0 -2
@@ -552,8 +552,6 @@ static void read_edp_current_link_settings_on_detect(struct dc_link *link)msleep (8 ); } -ASSERT (status ==DC_OK ); -// Read DPCD 00100h to find if standard link rates are set core_link_read_dpcd (link,DP_LINK_BW_SET , &link_bw_set,sizeof (link_bw_set));
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 77/87 drm/amd/display: 3.2.42
In reply to a message not in the archiveFrom: Aric Cyr <[email protected]>
--- drivers/gpu/drm/amd/display/dc/dc.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/dc.h +1 -1
@@ -39,7 +39,7 @@#include "inc/hw/dmcu.h" #include "dml/display_mode_lib.h" -#define DC_VER "3.2.41" +#define DC_VER "3.2.42"#define MAX_SURFACES 3#define MAX_PLANES 6
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 78/87 drm/amd/display: Increase size of audios array
In reply to a message not in the archiveFrom: Tai Man <[email protected]>
[Why]
The audios array defined in "struct resource_pool" is only 6 (MAX_PIPES) but the max number of audio devices (num_audio) is 7. In some projects, it will run out of audios array.
[How]
Incraese the audios array size to 7.
--- drivers/gpu/drm/amd/display/dc/inc/core_types.h | 2 +- drivers/gpu/drm/amd/display/dc/inc/hw/hw_shared.h | 1 + 2 files changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/inc/core_types.h +1 -1
@@ -212,7 +212,7 @@ struct resource_pool {struct clock_source *clock_sources[MAX_CLOCK_SOURCES ];unsigned int clk_src_count; -struct audio *audios[MAX_PIPES ]; +struct audio *audios[MAX_AUDIOS ];unsigned int audio_count;struct audio_support audio_support;
drivers/gpu/drm/amd/display/dc/inc/hw/hw_shared.h +1 -0
@@ -34,6 +34,7 @@* Data types shared between different Virtual HW blocks ******************************************************************************/ +#define MAX_AUDIOS 7#define MAX_PIPES 6#if defined(CONFIG_DRM_AMD_DC_DCN2_0 )#define MAX_DWB_PIPES 1
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 79/87 drm/amd/display: do not read link setting if edp not connected
In reply to a message not in the archiveFrom: Eric Yang <[email protected]>
[Why]
Previously assume eDP sink present if connector present. Do not need to enforce this restriction. Fix issue where driver attempt to read link setting even though no edp connected.
{How]
Only read link setting after reading connection status.
--- drivers/gpu/drm/amd/display/dc/core/dc_link.c | 8 +------- 1 file changed, 1 insertion(+), 7 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc_link.c +1 -7
@@ -720,13 +720,6 @@ bool dc_link_detect(struct dc_link *link, enum dc_detect_reason reason)return false; } -if (link->connector_signal ==SIGNAL_TYPE_EDP ) { -/* On detect, we want to make sure current link settings are -* up to date, especially if link was powered on by GOP. -*/ -read_edp_current_link_settings_on_detect (link); - } - prev_sink = link->local_sink;if (prev_sink !=NULL ) {dc_sink_retain (prev_sink);
@@ -768,6 +761,7 @@ bool dc_link_detect(struct dc_link *link, enum dc_detect_reason reason) }case SIGNAL_TYPE_EDP : { +read_edp_current_link_settings_on_detect (link);detect_edp_sink_caps (link); sink_caps.transaction_type =DDC_TRANSACTION_TYPE_I2C_OVER_AUX ;
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 80/87 drm/amd/display: fix mpcc assert condition
In reply to a message not in the archiveFrom: Eric Yang <[email protected]>
[Why]
In DCN2x asic, the MPCC status register definition changed, and our logic for assert is incorrect. disabled is valid state, where we should see idle and not busy, where as in not disabled state, we should see not idle.
[How]
Change assert condition to be more sensible.
--- .../gpu/drm/amd/display/dc/dcn20/dcn20_mpc.c | 23 +++++++++---------- 1 file changed, 11 insertions(+), 12 deletions(-)
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_mpc.c +11 -12
@@ -446,23 +446,22 @@ void mpc2_assert_mpcc_idle_before_connect(struct mpc *mpc, int mpcc_id) {struct dcn20_mpc *mpc20 =TO_DCN20_MPC (mpc);unsigned int top_sel, mpc_busy, mpc_idle, mpc_disabled; -REG_GET (MPCC_STATUS [mpcc_id],MPCC_DISABLED , &mpc_disabled); - -if (mpc_disabled) { -ASSERT (0 ); -return ; - }REG_GET (MPCC_TOP_SEL [mpcc_id],MPCC_TOP_SEL , &top_sel); -if (top_sel ==0xf ) { -REG_GET_2 (MPCC_STATUS [mpcc_id], -MPCC_BUSY , &mpc_busy, -MPCC_IDLE , &mpc_idle); +REG_GET_3 (MPCC_STATUS [mpcc_id], +MPCC_BUSY , &mpc_busy, +MPCC_IDLE , &mpc_idle, +MPCC_DISABLED , &mpc_disabled); -ASSERT (mpc_busy ==0 ); -ASSERT (mpc_idle ==1 ); +if (top_sel ==0xf ) { +ASSERT (!mpc_busy); +ASSERT (mpc_idle); +ASSERT (mpc_disabled); + }else { +ASSERT (!mpc_disabled); +ASSERT (!mpc_idle); } }
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 81/87 drm/amd/display: support "dummy pstate"
In reply to a message not in the archiveFrom: Jun Lei <[email protected]>
[why]
Existing support in DC for pstate only accounts for a single latency. This is sufficient when the variance of latency is small, or that pstate support isn't necessary for correct ASIC functionality.
Newer ASICs violate both existing assumptions. PState support is mandatory of correct ASIC functionality, but not all latencies have to be supported. Existing code supports a "full p state" which allows memory clock to change, but is hard for DCN to support (as it requires very large buffers). New code will now fall back to a "dummy p state" support when "full p state" cannot be support. This easy p state support should always be allowed.
[how]
Define a new latency in socBB. Add fallback logic to support it. Note DML is also updated to ensure that fallback will always work.
--- .../display/dc/clk_mgr/dcn20/dcn20_clk_mgr.c | 2 + drivers/gpu/drm/amd/display/dc/dc.h | 7 + .../drm/amd/display/dc/dcn20/dcn20_hubbub.c | 11 + .../drm/amd/display/dc/dcn20/dcn20_hwseq.c | 10 +- .../drm/amd/display/dc/dcn20/dcn20_resource.c | 59 +- drivers/gpu/drm/amd/display/dc/dml/Makefile | 3 + .../dc/dml/dcn20/display_mode_vba_20v2.c | 5109 +++++++++++++++++ .../dc/dml/dcn20/display_mode_vba_20v2.h | 32 + .../dc/dml/dcn20/display_rq_dlg_calc_20v2.c | 1701 ++++++ .../dc/dml/dcn20/display_rq_dlg_calc_20v2.h | 74 + .../drm/amd/display/dc/dml/display_mode_lib.c | 12 + .../drm/amd/display/dc/dml/display_mode_lib.h | 1 + .../amd/display/dc/dml/display_mode_structs.h | 1 + .../drm/amd/display/dc/dml/display_mode_vba.c | 8 +- 14 files changed, 7022 insertions(+), 8 deletions(-) create mode 100644 drivers/gpu/drm/amd/display/dc/dml/dcn20/display_mode_vba_20v2.c create mode 100644 drivers/gpu/drm/amd/display/dc/dml/dcn20/display_mode_vba_20v2.h create mode 100644 drivers/gpu/drm/amd/display/dc/dml/dcn20/display_rq_dlg_calc_20v2.c create mode 100644 drivers/gpu/drm/amd/display/dc/dml/dcn20/display_rq_dlg_calc_20v2.h
drivers/gpu/drm/amd/display/dc/clk_mgr/dcn20/dcn20_clk_mgr.c +2 -0
@@ -201,6 +201,7 @@ void dcn2_update_clocks(struct clk_mgr *clk_mgr_base, }if (should_update_pstate_support (safe_to_lower, new_clocks->p_state_change_support, clk_mgr_base->clks.p_state_change_support)) { + clk_mgr_base->clks.prev_p_state_change_support = clk_mgr_base->clks.p_state_change_support; clk_mgr_base->clks.p_state_change_support = new_clocks->p_state_change_support;if (pp_smu && pp_smu->set_pstate_handshake_support) pp_smu->set_pstate_handshake_support (&pp_smu->pp_smu, clk_mgr_base->clks.p_state_change_support);
@@ -308,6 +309,7 @@ void dcn2_init_clocks(struct clk_mgr *clk_mgr)memset (&(clk_mgr->clks),0 ,sizeof (struct dc_clocks ));// Assumption is that boot state always supports pstate clk_mgr->clks.p_state_change_support = true; + clk_mgr->clks.prev_p_state_change_support = true; }void dcn2_enable_pme_wa (struct clk_mgr *clk_mgr_base)
drivers/gpu/drm/amd/display/dc/dc.h +7 -0
@@ -121,6 +121,7 @@ struct dc_caps {struct dc_bug_wa {bool no_connect_phy_config;bool dedcn20_305_wa; +struct display_mode_lib alternate_dml; };#endif
@@ -263,6 +264,12 @@ struct dc_clocks {int phyclk_khz;int dramclk_khz;bool p_state_change_support; + +/* +* Elements below are not compared for the purposes of +* optimization required +*/ +bool prev_p_state_change_support; };struct dc_bw_validation_profile {
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hubbub.c +11 -0
@@ -26,6 +26,7 @@#include "dcn20_hubbub.h" #include "reg_helper.h" +#include "clk_mgr.h" #define REG (reg)\ hubbub1->regs->reg
@@ -553,6 +554,16 @@ static void hubbub2_program_watermarks(*/ hubbub1_program_urgent_watermarks (hubbub, watermarks, refclk_mhz, safe_to_lower);hubbub1_program_stutter_watermarks (hubbub, watermarks, refclk_mhz, safe_to_lower); + +/* +* There's a special case when going from p-state support to p-state unsupported +* here we are going to LOWER watermarks to go to dummy p-state only, but this has +* to be done prepare_bandwidth, not optimize +*/ +if (hubbub1->base.ctx->dc->clk_mgr->clks.prev_p_state_change_support == true && + hubbub1->base.ctx->dc->clk_mgr->clks.p_state_change_support == false) + safe_to_lower = true; +hubbub1_program_pstate_watermarks (hubbub, watermarks, refclk_mhz, safe_to_lower); REG_SET(DCHUBBUB_ARB_SAT_LEVEL, 0,
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c +5 -5
@@ -1443,16 +1443,16 @@ void dcn20_prepare_bandwidth( {struct hubbub *hubbub = dc->res_pool->hubbub; + dc->clk_mgr->funcs->update_clocks ( + dc->clk_mgr, + context, + false); +/* program dchubbub watermarks */ hubbub->funcs->program_watermarks (hubbub, &context->bw_ctx.bw.dcn.watermarks, dc->res_pool->ref_clocks.dchub_ref_clock_inKhz /1000 , false); - - dc->clk_mgr->funcs->update_clocks ( - dc->clk_mgr, - context, - false); } void dcn20_optimize_bandwidth(
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c +58 -1
77 lines not on this page · raw message
drivers/gpu/drm/amd/display/dc/dml/Makefile +3 -0
@@ -38,6 +38,8 @@ ifdef CONFIG_DRM_AMD_DC_DCN2_0 CFLAGS_display_mode_vba.o := $(dml_ccflags) CFLAGS_display_mode_vba_20.o := $(dml_ccflags) CFLAGS_display_rq_dlg_calc_20.o := $(dml_ccflags) +CFLAGS_display_mode_vba_20v2.o := $(dml_ccflags) +CFLAGS_display_rq_dlg_calc_20v2.o := $(dml_ccflags) endif ifdef CONFIG_DRM_AMD_DCN3AG CFLAGS_display_mode_vba_3ag.o := $(dml_ccflags)
@@ -51,6 +53,7 @@ DML = display_mode_lib.o display_rq_dlg_helpers.o dml1_display_rq_dlg_calc.o \ ifdef CONFIG_DRM_AMD_DC_DCN2_0 DML += display_mode_vba.o dcn20/display_rq_dlg_calc_20.o dcn20/display_mode_vba_20.o +DML += dcn20/display_rq_dlg_calc_20v2.o dcn20/display_mode_vba_20v2.o endif AMD_DAL_DML = $(addprefix $(AMDDALPATH)/dc/dml/,$(DML))
drivers/gpu/drm/amd/display/dc/dml/dcn20/display_mode_vba_20v2.c new file +5109 -0
2,831 lines not on this page · raw message
drivers/gpu/drm/amd/display/dc/dml/dcn20/display_mode_vba_20v2.h new file +32 -0
Lines of this file are not shown. View the raw message.
drivers/gpu/drm/amd/display/dc/dml/dcn20/display_rq_dlg_calc_20v2.c new file +1701 -0
1,701 lines not on this page · raw message
drivers/gpu/drm/amd/display/dc/dml/dcn20/display_rq_dlg_calc_20v2.h new file +74 -0
74 lines not on this page · raw message
drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.c +12 -0
Lines of this file are not shown. View the raw message.
drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.h +1 -0
Lines of this file are not shown. View the raw message.
drivers/gpu/drm/amd/display/dc/dml/display_mode_structs.h +1 -0
Lines of this file are not shown. View the raw message.
drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.c +6 -2
Lines of this file are not shown. View the raw message.
The rest of this diff is not shown. View the raw message.
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
Part of the diff is not shown here. View the raw message.
PATCH 82/87 drm/amd/display: Add SMU version field to clk_mgr_internal
In reply to a message not in the archiveFrom: Fatemeh Darbehani <[email protected]>
For some platforms, we need to know SMU version for driver/SMU compatibility. This change adds that field.
--- drivers/gpu/drm/amd/display/dc/inc/hw/clk_mgr_internal.h | 1 + 1 file changed, 1 insertion(+)
drivers/gpu/drm/amd/display/dc/inc/hw/clk_mgr_internal.h +1 -0
@@ -191,6 +191,7 @@ struct state_dependent_clocks {struct clk_mgr_internal {struct clk_mgr base; +int smu_ver;struct pp_smu_funcs *pp_smu;struct clk_mgr_internal_funcs *funcs;
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 83/87 drm/amd/display: avoid power gate domains that doesn't exist
In reply to a message not in the archiveFrom: Tony Cheng <[email protected]>
--- .../gpu/drm/amd/display/dc/dce/dce_hwseq.h | 4 ++-- .../drm/amd/display/dc/dcn20/dcn20_hwseq.c | 24 ++++++++++++------- 2 files changed, 17 insertions(+), 11 deletions(-)
drivers/gpu/drm/amd/display/dc/dce/dce_hwseq.h +2 -2
@@ -227,8 +227,8 @@SR (DOMAIN7_PG_CONFIG ), \SR (DOMAIN8_PG_CONFIG ), \SR (DOMAIN9_PG_CONFIG ), \ -SR (DOMAIN10_PG_CONFIG ), \ -SR (DOMAIN11_PG_CONFIG ), \ +/* SR(DOMAIN10_PG_CONFIG), Navi1x HUBP5 not powergate-able*/ \ +/* SR(DOMAIN11_PG_CONFIG), Navi1x DPP5 is not powergate-able */ \SR (DOMAIN16_PG_CONFIG ), \SR (DOMAIN17_PG_CONFIG ), \ SR(DOMAIN18_PG_CONFIG), \
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c +15 -9
@@ -94,25 +94,31 @@ static void enable_power_gating_plane(REG_UPDATE (DOMAIN2_PG_CONFIG ,DOMAIN2_POWER_FORCEON , force_on);REG_UPDATE (DOMAIN4_PG_CONFIG ,DOMAIN4_POWER_FORCEON , force_on);REG_UPDATE (DOMAIN6_PG_CONFIG ,DOMAIN6_POWER_FORCEON , force_on); -REG_UPDATE (DOMAIN8_PG_CONFIG ,DOMAIN8_POWER_FORCEON , force_on); -/*Do not power gate DCHUB5, should be left at HW default, power on permanently*/ -/*REG_UPDATE(DOMAIN10_PG_CONFIG, DOMAIN10_POWER_FORCEON, force_on);*/ +if (REG (DOMAIN8_PG_CONFIG )) +REG_UPDATE (DOMAIN8_PG_CONFIG ,DOMAIN8_POWER_FORCEON , force_on); +if (REG (DOMAIN10_PG_CONFIG )) +REG_UPDATE (DOMAIN10_PG_CONFIG ,DOMAIN8_POWER_FORCEON , force_on);/* DPP0/1/2/3/4/5 */ REG_UPDATE (DOMAIN1_PG_CONFIG ,DOMAIN1_POWER_FORCEON , force_on);REG_UPDATE (DOMAIN3_PG_CONFIG ,DOMAIN3_POWER_FORCEON , force_on);REG_UPDATE (DOMAIN5_PG_CONFIG ,DOMAIN5_POWER_FORCEON , force_on);REG_UPDATE (DOMAIN7_PG_CONFIG ,DOMAIN7_POWER_FORCEON , force_on); -REG_UPDATE (DOMAIN9_PG_CONFIG ,DOMAIN9_POWER_FORCEON , force_on); -/*Do not power gate DPP5, should be left at HW default, power on permanently*/ -/*REG_UPDATE(DOMAIN11_PG_CONFIG, DOMAIN11_POWER_FORCEON, force_on);*/ +if (REG (DOMAIN9_PG_CONFIG )) +REG_UPDATE (DOMAIN9_PG_CONFIG ,DOMAIN9_POWER_FORCEON , force_on); +if (REG (DOMAIN11_PG_CONFIG )) +REG_UPDATE (DOMAIN11_PG_CONFIG ,DOMAIN9_POWER_FORCEON , force_on); +/* DCS0/1/2/3/4/5 */ REG_UPDATE (DOMAIN16_PG_CONFIG ,DOMAIN16_POWER_FORCEON , force_on);REG_UPDATE (DOMAIN17_PG_CONFIG ,DOMAIN17_POWER_FORCEON , force_on);REG_UPDATE (DOMAIN18_PG_CONFIG ,DOMAIN18_POWER_FORCEON , force_on); -REG_UPDATE (DOMAIN19_PG_CONFIG ,DOMAIN19_POWER_FORCEON , force_on); -REG_UPDATE (DOMAIN20_PG_CONFIG ,DOMAIN20_POWER_FORCEON , force_on); -REG_UPDATE (DOMAIN21_PG_CONFIG ,DOMAIN21_POWER_FORCEON , force_on); +if (REG (DOMAIN19_PG_CONFIG )) +REG_UPDATE (DOMAIN19_PG_CONFIG ,DOMAIN19_POWER_FORCEON , force_on); +if (REG (DOMAIN20_PG_CONFIG )) +REG_UPDATE (DOMAIN20_PG_CONFIG ,DOMAIN20_POWER_FORCEON , force_on); +if (REG (DOMAIN21_PG_CONFIG )) +REG_UPDATE (DOMAIN21_PG_CONFIG ,DOMAIN21_POWER_FORCEON , force_on); }void dcn20_dccg_init (struct dce_hwseq *hws)
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 84/87 drm/amd/display: Add debug entry to destroy disconnected edp link
In reply to a message not in the archiveFrom: Lewis Huang <[email protected]>
Add a flag to dc_debug_options to determine if a disconnected edp link should be destroyed.
--- drivers/gpu/drm/amd/display/dc/core/dc.c | 20 ++++++++++++++++---- drivers/gpu/drm/amd/display/dc/dc.h | 1 + 2 files changed, 17 insertions(+), 4 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc.c +16 -4
@@ -178,13 +178,25 @@ static bool create_links( link =link_create (&link_init_params);if (link) { -if (dc->config.edp_not_connected && - link->connector_signal ==SIGNAL_TYPE_EDP ) { -link_destroy (&link); - }else { +bool should_destory_link = false; + +if (link->connector_signal ==SIGNAL_TYPE_EDP ) { +if (dc->config.edp_not_connected) + should_destory_link = true; +else if (dc->debug.remove_disconnect_edp) { +enum dc_connection_type type; +dc_link_detect_sink (link, &type); +if (type == dc_connection_none) + should_destory_link = true; + } + } + +if (!should_destory_link) { dc->links[dc->link_count] = link; link->dc = dc; ++dc->link_count; + }else { +link_destroy (&link); } } }
drivers/gpu/drm/amd/display/dc/dc.h +1 -0
@@ -378,6 +378,7 @@ struct dc_debug_options {bool scl_reset_length10;bool hdmi20_disable;bool skip_detection_link_training; +bool remove_disconnect_edp;unsigned int force_odm_combine;//bit vector based on otg inst unsigned int force_fclk_khz;bool disable_tri_buf;
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 85/87 drm/amd/display: Copy GSL groups when committing a new context
In reply to a message not in the archiveFrom: Nicholas Kazlauskas <[email protected]>
[Why]
DC configures the GSL group for the pipe when pipe_split is enabled and we're switching flip types (buffered <-> immediate flip) on DCN2.
In order to record what GSL group the pipe is using DC stores it in the pipe's stream_res. DM is not aware of this internal grouping, nor is DC resource.
So when DM creates a dc_state context and passes it to DC the current GSL group is lost - DM never knew about it in the first place.
After 3 immediate flips we run out of GSL groups and we're no longer able to correctly perform *any* flip for multi-pipe scenarios.
[How]
The gsl_group needs to be copied to the new context.
DM has no insight into GSL grouping and could even potentially create a brand new context without referencing current hardware state. So this makes the most sense to have happen in DC.
There are two places where DC can apply a new context:
- dc_commit_state
- dc_commit_updates_for_stream
But what's shared between both of these is apply_ctx_for_surface.
This logic only matters for DCN2, so it can be placed in dcn20_apply_ctx_for_surface. Before doing any locking (where the GSL group is setup) we can copy over the GSL groups before committing the new context.
--- drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c | 12 ++++++++++++ 1 file changed, 12 insertions(+)
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c +12 -0
@@ -1364,6 +1364,18 @@ static void dcn20_apply_ctx_for_surface(if (!top_pipe_to_program)return ; +/* Carry over GSL groups in case the context is changing. */ +for (i =0 ; i < dc->res_pool->pipe_count; i++) { +struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i]; +struct pipe_ctx *old_pipe_ctx = + &dc->current_state->res_ctx.pipe_ctx[i]; + +if (pipe_ctx->stream == stream && + pipe_ctx->stream == old_pipe_ctx->stream) + pipe_ctx->stream_res.gsl_group = + old_pipe_ctx->stream_res.gsl_group; + } + tg = top_pipe_to_program->stream_res.tg; interdependent_update = top_pipe_to_program->plane_state &&
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 86/87 drm/amd/display: handle active dongle port type is DP++ or DP case
In reply to a message not in the archiveFrom: Dale Zhao <[email protected]>
[Why]:
Some active dongles have DP++ port and DP port at the same time. Current code doesn't cover DP++ case and processes as default DVI case, in which audio is disabled. Because of dual mode, DP case is also treat as DVI case for the other port.
[How]:
According DP 1.4 spec, add DP++ procedure similar with HDMI case. Also add None dongle type for DP case.
--- .../gpu/drm/amd/display/dc/core/dc_link_dp.c | 30 ++++++++++++------- .../gpu/drm/amd/display/include/dpcd_defs.h | 2 +- 2 files changed, 21 insertions(+), 11 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c +20 -10
@@ -2398,8 +2398,8 @@ static void get_active_converter_info(case DOWNSTREAM_VGA : link->dpcd_caps.dongle_type =DISPLAY_DONGLE_DP_VGA_CONVERTER ;break ; -case DOWNSTREAM_DVI_HDMI : -/* At this point we don't know is it DVI or HDMI, +case DOWNSTREAM_DVI_HDMI_DP_PLUS_PLUS : +/* At this point we don't know is it DVI or HDMI or DP++, * assume DVI.*/ link->dpcd_caps.dongle_type =DISPLAY_DONGLE_DP_DVI_CONVERTER ;break ;
@@ -2416,6 +2416,10 @@ static void get_active_converter_info( det_caps, sizeof(det_caps));switch (port_caps->bits.DWN_STRM_PORTX_TYPE) { +/*Handle DP case as DONGLE_NONE*/ +case DOWN_STREAM_DETAILED_DP : + link->dpcd_caps.dongle_type =DISPLAY_DONGLE_NONE ; +break ;case DOWN_STREAM_DETAILED_VGA : link->dpcd_caps.dongle_type =DISPLAY_DONGLE_DP_VGA_CONVERTER ;
@@ -2425,6 +2429,8 @@ static void get_active_converter_info(DISPLAY_DONGLE_DP_DVI_CONVERTER ;break ;case DOWN_STREAM_DETAILED_HDMI : +case DOWN_STREAM_DETAILED_DP_PLUS_PLUS : +/*Handle DP++ active converter case, process DP++ case as HDMI case according DP1.4 spec*/ link->dpcd_caps.dongle_type =DISPLAY_DONGLE_DP_HDMI_CONVERTER ;
@@ -2440,14 +2446,18 @@ static void get_active_converter_info( link->dpcd_caps.dongle_caps.is_dp_hdmi_s3d_converter = hdmi_caps.bits.FRAME_SEQ_TO_FRAME_PACK; - link->dpcd_caps.dongle_caps.is_dp_hdmi_ycbcr422_pass_through = - hdmi_caps.bits.YCrCr422_PASS_THROUGH; - link->dpcd_caps.dongle_caps.is_dp_hdmi_ycbcr420_pass_through = - hdmi_caps.bits.YCrCr420_PASS_THROUGH; - link->dpcd_caps.dongle_caps.is_dp_hdmi_ycbcr422_converter = - hdmi_caps.bits.YCrCr422_CONVERSION; - link->dpcd_caps.dongle_caps.is_dp_hdmi_ycbcr420_converter = - hdmi_caps.bits.YCrCr420_CONVERSION; +/*YCBCR capability only for HDMI case*/ +if (port_caps->bits.DWN_STRM_PORTX_TYPE + ==DOWN_STREAM_DETAILED_HDMI ) { + link->dpcd_caps.dongle_caps.is_dp_hdmi_ycbcr422_pass_through = + hdmi_caps.bits.YCrCr422_PASS_THROUGH; + link->dpcd_caps.dongle_caps.is_dp_hdmi_ycbcr420_pass_through = + hdmi_caps.bits.YCrCr420_PASS_THROUGH; + link->dpcd_caps.dongle_caps.is_dp_hdmi_ycbcr422_converter = + hdmi_caps.bits.YCrCr422_CONVERSION; + link->dpcd_caps.dongle_caps.is_dp_hdmi_ycbcr420_converter = + hdmi_caps.bits.YCrCr420_CONVERSION; + } link->dpcd_caps.dongle_caps.dp_hdmi_max_bpc = translate_dpcd_max_bpc(
drivers/gpu/drm/amd/display/include/dpcd_defs.h +1 -1
@@ -43,7 +43,7 @@ enum dpcd_revision {enum dpcd_downstream_port_type {DOWNSTREAM_DP =0 ,DOWNSTREAM_VGA , -DOWNSTREAM_DVI_HDMI , +DOWNSTREAM_DVI_HDMI_DP_PLUS_PLUS ,/* DVI, HDMI, DP++ */ DOWNSTREAM_NONDDC /* has no EDID (TV,CV) */ };
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
PATCH 87/87 drm/amd/display: Force uclk to max for every state
In reply to a message not in the archiveFrom: Nicholas Kazlauskas <[email protected]>
Workaround for now to avoid underflow.
The uclk switch time should really be bumped up to 404, but doing so would expose p-state hang issues for higher bandwidth display configurations.
--- .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_pp_smu.c | 6 +++--- drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c | 10 ++++++++++ 2 files changed, 13 insertions(+), 3 deletions(-)
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_pp_smu.c +3 -3
@@ -912,11 +912,11 @@ void dm_pp_get_funcs(/* todo set_pme_wa_enable cause 4k@6ohz display not light up */ funcs->nv_funcs.set_pme_wa_enable =NULL ;/* todo debug waring message */ - funcs->nv_funcs.set_hard_min_uclk_by_freq =NULL ; + funcs->nv_funcs.set_hard_min_uclk_by_freq = pp_nv_set_hard_min_uclk_by_freq;/* todo compare data with window driver*/ - funcs->nv_funcs.get_maximum_sustainable_clocks =NULL ; + funcs->nv_funcs.get_maximum_sustainable_clocks = pp_nv_get_maximum_sustainable_clocks;/*todo compare data with window driver */ - funcs->nv_funcs.get_uclk_dpm_states =NULL ; + funcs->nv_funcs.get_uclk_dpm_states = pp_nv_get_uclk_dpm_states;break ;#endif default :
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c +10 -0
@@ -2707,6 +2707,9 @@ static void cap_soc_clocks( && max_clocks.uClockInKhz != 0) bb->clock_limits[i].dram_speed_mts = (max_clocks.uClockInKhz /1000 ) *16 ; +// HACK: Force every uclk to max for now to "disable" uclk switching. + bb->clock_limits[i].dram_speed_mts = (max_clocks.uClockInKhz /1000 ) *16 ; +if ((bb->clock_limits[i].fabricclk_mhz > (max_clocks.fabricClockInKhz /1000 )) && max_clocks.fabricClockInKhz !=0 ) bb->clock_limits[i].fabricclk_mhz = (max_clocks.fabricClockInKhz /1000 );
@@ -2922,6 +2925,8 @@ static bool init_soc_bounding_box(struct dc *dc,le32_to_cpu (bb->vmm_page_size_bytes); dcn2_0_soc.dram_clock_change_latency_us =fixed16_to_double_to_cpu (bb->dram_clock_change_latency_us); +// HACK!! Lower uclock latency switch time so we don't switch + dcn2_0_soc.dram_clock_change_latency_us =10 ; dcn2_0_soc.writeback_dram_clock_change_latency_us =fixed16_to_double_to_cpu (bb->writeback_dram_clock_change_latency_us); dcn2_0_soc.return_bus_width_bytes =
@@ -2963,6 +2968,7 @@ static bool init_soc_bounding_box(struct dc *dc,struct pp_smu_nv_clock_table max_clocks = {0 };unsigned int uclk_states[8 ] = {0 };unsigned int num_states =0 ; +int i;enum pp_smu_status status;bool clock_limits_available = false;bool uclk_states_available = false;
@@ -2984,6 +2990,10 @@ static bool init_soc_bounding_box(struct dc *dc, clock_limits_available = (status ==PP_SMU_RESULT_OK ); } +// HACK: Use the max uclk_states value for all elements. +for (i =0 ; i < num_states; i++) + uclk_states[i] = uclk_states[num_states -1 ]; +if (clock_limits_available && uclk_states_available && num_states)update_bounding_box (dc, &dcn2_0_soc, &max_clocks, uclk_states, num_states);else if (clock_limits_available)
Signature · 6 lines
2.22.0 _______________________________________________ amd-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/amd-gfx
A message that is not in the archive (sent off-list or not yet received): [email protected]
Liu, Zhan
Reviewed-by Zhan LiuRE: [PATCH 87/87] drm/amd/display: Force uclk to max for every state
-----Original Message-----
From: amd-gfx <[email protected]> On Behalf Of [email protected]
Sent: Monday, July 15, 2019 5:21 PM
To: [email protected]
Cc: Li, Sun peng (Leo) <[email protected]>; Kazlauskas, Nicholas <[email protected]>
Subject: [PATCH 87/87] drm/amd/display: Force uclk to max for every state
From: Nicholas Kazlauskas <[email protected]>
Workaround for now to avoid underflow.
The uclk switch time should really be bumped up to 404, but doing so would expose p-state hang issues for higher bandwidth display configurations.
--- .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_pp_smu.c | 6 +++--- drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c | 10 ++++++++++ 2 files changed, 13 insertions(+), 3 deletions(-)
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_pp_smu.c +5 -3
@@ -912,11 +912,11 @@ void dm_pp_get_funcs(/* todo set_pme_wa_enable cause 4k@6ohz display not light up */ funcs->nv_funcs.set_pme_wa_enable =NULL ;/* todo debug waring message */ - funcs->nv_funcs.set_hard_min_uclk_by_freq =NULL ; + funcs->nv_funcs.set_hard_min_uclk_by_freq = +pp_nv_set_hard_min_uclk_by_freq;/* todo compare data with window driver*/ - funcs->nv_funcs.get_maximum_sustainable_clocks =NULL ; + funcs->nv_funcs.get_maximum_sustainable_clocks = +pp_nv_get_maximum_sustainable_clocks;/*todo compare data with window driver */ - funcs->nv_funcs.get_uclk_dpm_states =NULL ; + funcs->nv_funcs.get_uclk_dpm_states = pp_nv_get_uclk_dpm_states;break ;
#endif
default:
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c +4 -0
@@ -2707,6 +2707,9 @@ static void cap_soc_clocks( && max_clocks.uClockInKhz != 0) bb->clock_limits[i].dram_speed_mts = (max_clocks.uClockInKhz /1000 ) *16 ; +// HACK: Force every uclk to max for now to "disable" uclk switching. + bb->clock_limits[i].dram_speed_mts = (max_clocks.uClockInKhz /1000 ) +*16 ; +if ((bb->clock_limits[i].fabricclk_mhz > (max_clocks.fabricClockInKhz /1000 )) && max_clocks.fabricClockInKhz !=0 )
bb->clock_limits[i].fabricclk_mhz = (max_clocks.fabricClockInKhz / 1000); @@ -2922,6 +2925,8 @@ static bool init_soc_bounding_box(struct dc *dc,
le32_to_cpu(bb->vmm_page_size_bytes);
dcn2_0_soc.dram_clock_change_latency_us =
fixed16_to_double_to_cpu(bb->dram_clock_change_latency_us);
+ // HACK!! Lower uclock latency switch time so we don't switch
+ dcn2_0_soc.dram_clock_change_latency_us = 10;
dcn2_0_soc.writeback_dram_clock_change_latency_us =
fixed16_to_double_to_cpu(bb->writeback_dram_clock_change_latency_us);
dcn2_0_soc.return_bus_width_bytes =
@@ -2963,6 +2968,7 @@ static bool init_soc_bounding_box(struct dc *dc,
struct pp_smu_nv_clock_table max_clocks = {0};
unsigned int uclk_states[8] = {0};
unsigned int num_states = 0;
+ int i;
enum pp_smu_status status;
bool clock_limits_available = false;
bool uclk_states_available = false;
@@ -2984,6 +2990,10 @@ static bool init_soc_bounding_box(struct dc *dc,
clock_limits_available = (status == PP_SMU_RESULT_OK);
}
+ // HACK: Use the max uclk_states value for all elements.
+ for (i = 0; i < num_states; i++)
+ uclk_states[i] = uclk_states[num_states - 1];
+
if (clock_limits_available && uclk_states_available && num_states)
update_bounding_box(dc, &dcn2_0_soc, &max_clocks, uclk_states, num_states);
else if (clock_limits_available)
--
2.22.0
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx