From b6fcc3867d746c181d253b110236985b9b2ee2aa Mon Sep 17 00:00:00 2001 From: Wayne Lin Date: Fri, 6 Dec 2024 13:51:46 +0800 Subject: [PATCH 01/16] drm/amd/display: Add support to configure CRC window on specific CRC instance [Why] Have the need to specify the CRC window on specific CRC engine. dc_stream_configure_crc() today calculates CRC on crc engine 0 only and always resets CRC engine at first. [How] Add index parameter to dc_stream_configure_crc() for selecting the desired crc engine. Additionally, add another parameter to specify whether to skip the default reset of crc engine. Reviewed-by: HaoPing Liu Signed-off-by: Wayne Lin Signed-off-by: Tom Chung Tested-by: Daniel Wheeler Signed-off-by: Alex Deucher --- .../drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c | 2 +- drivers/gpu/drm/amd/display/dc/core/dc.c | 11 +- drivers/gpu/drm/amd/display/dc/dc_stream.h | 4 +- .../dc/dce110/dce110_timing_generator.c | 169 ++++++++++++------ .../dc/dce120/dce120_timing_generator.c | 90 +++++++--- .../amd/display/dc/inc/hw/timing_generator.h | 3 + .../amd/display/dc/optc/dcn10/dcn10_optc.c | 86 ++++++--- .../amd/display/dc/optc/dcn35/dcn35_optc.c | 101 ++++++++--- 8 files changed, 329 insertions(+), 137 deletions(-) diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c index 7c15082c92ed..a2e36b882a4f 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c @@ -510,7 +510,7 @@ int amdgpu_dm_crtc_configure_crc_source(struct drm_crtc *crtc, /* Enable or disable CRTC CRC generation */ if (dm_is_crc_source_crtc(source) || source == AMDGPU_DM_PIPE_CRC_SOURCE_NONE) { if (!dc_stream_configure_crc(stream_state->ctx->dc, - stream_state, NULL, enable, enable)) { + stream_state, NULL, enable, enable, 0, true)) { ret = -EINVAL; goto unlock; } diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c index cb6293d8ba77..cecaadf741ad 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc.c @@ -687,15 +687,17 @@ dc_stream_forward_multiple_crc_window(struct dc_stream_state *stream, * @enable: Enable CRC if true, disable otherwise. * @continuous: Capture CRC on every frame if true. Otherwise, only capture * once. + * @idx: Capture CRC on which CRC engine instance + * @reset: Reset CRC engine before the configuration * - * By default, only CRC0 is configured, and the entire frame is used to - * calculate the CRC. + * By default, the entire frame is used to calculate the CRC. * * Return: %false if the stream is not found or CRC capture is not supported; * %true if the stream has been configured. */ bool dc_stream_configure_crc(struct dc *dc, struct dc_stream_state *stream, - struct crc_params *crc_window, bool enable, bool continuous) + struct crc_params *crc_window, bool enable, bool continuous, + uint8_t idx, bool reset) { struct pipe_ctx *pipe; struct crc_params param; @@ -739,6 +741,9 @@ bool dc_stream_configure_crc(struct dc *dc, struct dc_stream_state *stream, param.continuous_mode = continuous; param.enable = enable; + param.crc_eng_inst = idx; + param.reset = reset; + tg = pipe->stream_res.tg; /* Only call if supported */ diff --git a/drivers/gpu/drm/amd/display/dc/dc_stream.h b/drivers/gpu/drm/amd/display/dc/dc_stream.h index dbf0832eb706..3e303c7808fb 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_stream.h +++ b/drivers/gpu/drm/amd/display/dc/dc_stream.h @@ -550,7 +550,9 @@ bool dc_stream_configure_crc(struct dc *dc, struct dc_stream_state *stream, struct crc_params *crc_window, bool enable, - bool continuous); + bool continuous, + uint8_t idx, + bool reset); bool dc_stream_get_crc(struct dc *dc, struct dc_stream_state *stream, diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator.c index e32dd96a99cb..61b0807693fb 100644 --- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator.c +++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_timing_generator.c @@ -2127,65 +2127,126 @@ bool dce110_configure_crc(struct timing_generator *tg, cntl_addr = CRTC_REG(mmCRTC_CRC_CNTL); - /* First, disable CRC before we configure it. */ - dm_write_reg(tg->ctx, cntl_addr, 0); + if (!params->enable || params->reset) + /* First, disable CRC before we configure it. */ + dm_write_reg(tg->ctx, cntl_addr, 0); if (!params->enable) return true; /* Program frame boundaries */ - /* Window A x axis start and end. */ - value = 0; - addr = CRTC_REG(mmCRTC_CRC0_WINDOWA_X_CONTROL); - set_reg_field_value(value, params->windowa_x_start, - CRTC_CRC0_WINDOWA_X_CONTROL, - CRTC_CRC0_WINDOWA_X_START); - set_reg_field_value(value, params->windowa_x_end, - CRTC_CRC0_WINDOWA_X_CONTROL, - CRTC_CRC0_WINDOWA_X_END); - dm_write_reg(tg->ctx, addr, value); - - /* Window A y axis start and end. */ - value = 0; - addr = CRTC_REG(mmCRTC_CRC0_WINDOWA_Y_CONTROL); - set_reg_field_value(value, params->windowa_y_start, - CRTC_CRC0_WINDOWA_Y_CONTROL, - CRTC_CRC0_WINDOWA_Y_START); - set_reg_field_value(value, params->windowa_y_end, - CRTC_CRC0_WINDOWA_Y_CONTROL, - CRTC_CRC0_WINDOWA_Y_END); - dm_write_reg(tg->ctx, addr, value); - - /* Window B x axis start and end. */ - value = 0; - addr = CRTC_REG(mmCRTC_CRC0_WINDOWB_X_CONTROL); - set_reg_field_value(value, params->windowb_x_start, - CRTC_CRC0_WINDOWB_X_CONTROL, - CRTC_CRC0_WINDOWB_X_START); - set_reg_field_value(value, params->windowb_x_end, - CRTC_CRC0_WINDOWB_X_CONTROL, - CRTC_CRC0_WINDOWB_X_END); - dm_write_reg(tg->ctx, addr, value); - - /* Window B y axis start and end. */ - value = 0; - addr = CRTC_REG(mmCRTC_CRC0_WINDOWB_Y_CONTROL); - set_reg_field_value(value, params->windowb_y_start, - CRTC_CRC0_WINDOWB_Y_CONTROL, - CRTC_CRC0_WINDOWB_Y_START); - set_reg_field_value(value, params->windowb_y_end, - CRTC_CRC0_WINDOWB_Y_CONTROL, - CRTC_CRC0_WINDOWB_Y_END); - dm_write_reg(tg->ctx, addr, value); - - /* Set crc mode and selection, and enable. Only using CRC0*/ - value = 0; - set_reg_field_value(value, params->continuous_mode ? 1 : 0, - CRTC_CRC_CNTL, CRTC_CRC_CONT_EN); - set_reg_field_value(value, params->selection, - CRTC_CRC_CNTL, CRTC_CRC0_SELECT); - set_reg_field_value(value, 1, CRTC_CRC_CNTL, CRTC_CRC_EN); - dm_write_reg(tg->ctx, cntl_addr, value); + switch (params->crc_eng_inst) { + case 0: + /* Window A x axis start and end. */ + value = 0; + addr = CRTC_REG(mmCRTC_CRC0_WINDOWA_X_CONTROL); + set_reg_field_value(value, params->windowa_x_start, + CRTC_CRC0_WINDOWA_X_CONTROL, + CRTC_CRC0_WINDOWA_X_START); + set_reg_field_value(value, params->windowa_x_end, + CRTC_CRC0_WINDOWA_X_CONTROL, + CRTC_CRC0_WINDOWA_X_END); + dm_write_reg(tg->ctx, addr, value); + + /* Window A y axis start and end. */ + value = 0; + addr = CRTC_REG(mmCRTC_CRC0_WINDOWA_Y_CONTROL); + set_reg_field_value(value, params->windowa_y_start, + CRTC_CRC0_WINDOWA_Y_CONTROL, + CRTC_CRC0_WINDOWA_Y_START); + set_reg_field_value(value, params->windowa_y_end, + CRTC_CRC0_WINDOWA_Y_CONTROL, + CRTC_CRC0_WINDOWA_Y_END); + dm_write_reg(tg->ctx, addr, value); + + /* Window B x axis start and end. */ + value = 0; + addr = CRTC_REG(mmCRTC_CRC0_WINDOWB_X_CONTROL); + set_reg_field_value(value, params->windowb_x_start, + CRTC_CRC0_WINDOWB_X_CONTROL, + CRTC_CRC0_WINDOWB_X_START); + set_reg_field_value(value, params->windowb_x_end, + CRTC_CRC0_WINDOWB_X_CONTROL, + CRTC_CRC0_WINDOWB_X_END); + dm_write_reg(tg->ctx, addr, value); + + /* Window B y axis start and end. */ + value = 0; + addr = CRTC_REG(mmCRTC_CRC0_WINDOWB_Y_CONTROL); + set_reg_field_value(value, params->windowb_y_start, + CRTC_CRC0_WINDOWB_Y_CONTROL, + CRTC_CRC0_WINDOWB_Y_START); + set_reg_field_value(value, params->windowb_y_end, + CRTC_CRC0_WINDOWB_Y_CONTROL, + CRTC_CRC0_WINDOWB_Y_END); + dm_write_reg(tg->ctx, addr, value); + + /* Set crc mode and selection, and enable.*/ + value = 0; + set_reg_field_value(value, params->continuous_mode ? 1 : 0, + CRTC_CRC_CNTL, CRTC_CRC_CONT_EN); + set_reg_field_value(value, params->selection, + CRTC_CRC_CNTL, CRTC_CRC0_SELECT); + set_reg_field_value(value, 1, CRTC_CRC_CNTL, CRTC_CRC_EN); + dm_write_reg(tg->ctx, cntl_addr, value); + break; + case 1: + /* Window A x axis start and end. */ + value = 0; + addr = CRTC_REG(mmCRTC_CRC1_WINDOWA_X_CONTROL); + set_reg_field_value(value, params->windowa_x_start, + CRTC_CRC1_WINDOWA_X_CONTROL, + CRTC_CRC1_WINDOWA_X_START); + set_reg_field_value(value, params->windowa_x_end, + CRTC_CRC1_WINDOWA_X_CONTROL, + CRTC_CRC1_WINDOWA_X_END); + dm_write_reg(tg->ctx, addr, value); + + /* Window A y axis start and end. */ + value = 0; + addr = CRTC_REG(mmCRTC_CRC1_WINDOWA_Y_CONTROL); + set_reg_field_value(value, params->windowa_y_start, + CRTC_CRC1_WINDOWA_Y_CONTROL, + CRTC_CRC1_WINDOWA_Y_START); + set_reg_field_value(value, params->windowa_y_end, + CRTC_CRC1_WINDOWA_Y_CONTROL, + CRTC_CRC1_WINDOWA_Y_END); + dm_write_reg(tg->ctx, addr, value); + + /* Window B x axis start and end. */ + value = 0; + addr = CRTC_REG(mmCRTC_CRC1_WINDOWB_X_CONTROL); + set_reg_field_value(value, params->windowb_x_start, + CRTC_CRC1_WINDOWB_X_CONTROL, + CRTC_CRC1_WINDOWB_X_START); + set_reg_field_value(value, params->windowb_x_end, + CRTC_CRC1_WINDOWB_X_CONTROL, + CRTC_CRC1_WINDOWB_X_END); + dm_write_reg(tg->ctx, addr, value); + + /* Window B y axis start and end. */ + value = 0; + addr = CRTC_REG(mmCRTC_CRC1_WINDOWB_Y_CONTROL); + set_reg_field_value(value, params->windowb_y_start, + CRTC_CRC1_WINDOWB_Y_CONTROL, + CRTC_CRC1_WINDOWB_Y_START); + set_reg_field_value(value, params->windowb_y_end, + CRTC_CRC1_WINDOWB_Y_CONTROL, + CRTC_CRC1_WINDOWB_Y_END); + dm_write_reg(tg->ctx, addr, value); + + /* Set crc mode and selection, and enable.*/ + value = 0; + set_reg_field_value(value, params->continuous_mode ? 1 : 0, + CRTC_CRC_CNTL, CRTC_CRC_CONT_EN); + set_reg_field_value(value, params->selection, + CRTC_CRC_CNTL, CRTC_CRC1_SELECT); + set_reg_field_value(value, 1, CRTC_CRC_CNTL, CRTC_CRC_EN); + dm_write_reg(tg->ctx, cntl_addr, value); + break; + default: + return false; + } return true; } diff --git a/drivers/gpu/drm/amd/display/dc/dce120/dce120_timing_generator.c b/drivers/gpu/drm/amd/display/dc/dce120/dce120_timing_generator.c index 4984adce077e..31c4f44ceaac 100644 --- a/drivers/gpu/drm/amd/display/dc/dce120/dce120_timing_generator.c +++ b/drivers/gpu/drm/amd/display/dc/dce120/dce120_timing_generator.c @@ -1100,39 +1100,73 @@ static bool dce120_configure_crc(struct timing_generator *tg, if (!dce120_is_tg_enabled(tg)) return false; - /* First, disable CRC before we configure it. */ - dm_write_reg_soc15(tg->ctx, mmCRTC0_CRTC_CRC_CNTL, - tg110->offsets.crtc, 0); + if (!params->enable || params->reset) + /* First, disable CRC before we configure it. */ + dm_write_reg_soc15(tg->ctx, mmCRTC0_CRTC_CRC_CNTL, + tg110->offsets.crtc, 0); if (!params->enable) return true; /* Program frame boundaries */ - /* Window A x axis start and end. */ - CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC0_WINDOWA_X_CONTROL, - CRTC_CRC0_WINDOWA_X_START, params->windowa_x_start, - CRTC_CRC0_WINDOWA_X_END, params->windowa_x_end); - - /* Window A y axis start and end. */ - CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC0_WINDOWA_Y_CONTROL, - CRTC_CRC0_WINDOWA_Y_START, params->windowa_y_start, - CRTC_CRC0_WINDOWA_Y_END, params->windowa_y_end); - - /* Window B x axis start and end. */ - CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC0_WINDOWB_X_CONTROL, - CRTC_CRC0_WINDOWB_X_START, params->windowb_x_start, - CRTC_CRC0_WINDOWB_X_END, params->windowb_x_end); - - /* Window B y axis start and end. */ - CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC0_WINDOWB_Y_CONTROL, - CRTC_CRC0_WINDOWB_Y_START, params->windowb_y_start, - CRTC_CRC0_WINDOWB_Y_END, params->windowb_y_end); - - /* Set crc mode and selection, and enable. Only using CRC0*/ - CRTC_REG_UPDATE_3(CRTC0_CRTC_CRC_CNTL, - CRTC_CRC_EN, params->continuous_mode ? 1 : 0, - CRTC_CRC0_SELECT, params->selection, - CRTC_CRC_EN, 1); + switch (params->crc_eng_inst) { + case 0: + /* Window A x axis start and end. */ + CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC0_WINDOWA_X_CONTROL, + CRTC_CRC0_WINDOWA_X_START, params->windowa_x_start, + CRTC_CRC0_WINDOWA_X_END, params->windowa_x_end); + + /* Window A y axis start and end. */ + CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC0_WINDOWA_Y_CONTROL, + CRTC_CRC0_WINDOWA_Y_START, params->windowa_y_start, + CRTC_CRC0_WINDOWA_Y_END, params->windowa_y_end); + + /* Window B x axis start and end. */ + CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC0_WINDOWB_X_CONTROL, + CRTC_CRC0_WINDOWB_X_START, params->windowb_x_start, + CRTC_CRC0_WINDOWB_X_END, params->windowb_x_end); + + /* Window B y axis start and end. */ + CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC0_WINDOWB_Y_CONTROL, + CRTC_CRC0_WINDOWB_Y_START, params->windowb_y_start, + CRTC_CRC0_WINDOWB_Y_END, params->windowb_y_end); + + /* Set crc mode and selection, and enable.*/ + CRTC_REG_UPDATE_3(CRTC0_CRTC_CRC_CNTL, + CRTC_CRC_CONT_EN, params->continuous_mode ? 1 : 0, + CRTC_CRC0_SELECT, params->selection, + CRTC_CRC_EN, 1); + break; + case 1: + /* Window A x axis start and end. */ + CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC1_WINDOWA_X_CONTROL, + CRTC_CRC1_WINDOWA_X_START, params->windowa_x_start, + CRTC_CRC1_WINDOWA_X_END, params->windowa_x_end); + + /* Window A y axis start and end. */ + CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC1_WINDOWA_Y_CONTROL, + CRTC_CRC1_WINDOWA_Y_START, params->windowa_y_start, + CRTC_CRC1_WINDOWA_Y_END, params->windowa_y_end); + + /* Window B x axis start and end. */ + CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC1_WINDOWB_X_CONTROL, + CRTC_CRC1_WINDOWB_X_START, params->windowb_x_start, + CRTC_CRC1_WINDOWB_X_END, params->windowb_x_end); + + /* Window B y axis start and end. */ + CRTC_REG_UPDATE_2(CRTC0_CRTC_CRC1_WINDOWB_Y_CONTROL, + CRTC_CRC1_WINDOWB_Y_START, params->windowb_y_start, + CRTC_CRC1_WINDOWB_Y_END, params->windowb_y_end); + + /* Set crc mode and selection, and enable */ + CRTC_REG_UPDATE_3(CRTC0_CRTC_CRC_CNTL, + CRTC_CRC_CONT_EN, params->continuous_mode ? 1 : 0, + CRTC_CRC1_SELECT, params->selection, + CRTC_CRC_EN, 1); + break; + default: + return false; + } return true; } diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h b/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h index 4807b9f9f5a5..9885cb3c310f 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h +++ b/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h @@ -141,6 +141,9 @@ struct crc_params { bool continuous_mode; bool enable; + + uint8_t crc_eng_inst; + bool reset; }; /** diff --git a/drivers/gpu/drm/amd/display/dc/optc/dcn10/dcn10_optc.c b/drivers/gpu/drm/amd/display/dc/optc/dcn10/dcn10_optc.c index a296bb89b106..19d5ebc6763c 100644 --- a/drivers/gpu/drm/amd/display/dc/optc/dcn10/dcn10_optc.c +++ b/drivers/gpu/drm/amd/display/dc/optc/dcn10/dcn10_optc.c @@ -1470,37 +1470,71 @@ bool optc1_configure_crc(struct timing_generator *optc, if (!optc1_is_tg_enabled(optc)) return false; - REG_WRITE(OTG_CRC_CNTL, 0); + if (!params->enable || params->reset) + REG_WRITE(OTG_CRC_CNTL, 0); if (!params->enable) return true; /* Program frame boundaries */ - /* Window A x axis start and end. */ - REG_UPDATE_2(OTG_CRC0_WINDOWA_X_CONTROL, - OTG_CRC0_WINDOWA_X_START, params->windowa_x_start, - OTG_CRC0_WINDOWA_X_END, params->windowa_x_end); - - /* Window A y axis start and end. */ - REG_UPDATE_2(OTG_CRC0_WINDOWA_Y_CONTROL, - OTG_CRC0_WINDOWA_Y_START, params->windowa_y_start, - OTG_CRC0_WINDOWA_Y_END, params->windowa_y_end); - - /* Window B x axis start and end. */ - REG_UPDATE_2(OTG_CRC0_WINDOWB_X_CONTROL, - OTG_CRC0_WINDOWB_X_START, params->windowb_x_start, - OTG_CRC0_WINDOWB_X_END, params->windowb_x_end); - - /* Window B y axis start and end. */ - REG_UPDATE_2(OTG_CRC0_WINDOWB_Y_CONTROL, - OTG_CRC0_WINDOWB_Y_START, params->windowb_y_start, - OTG_CRC0_WINDOWB_Y_END, params->windowb_y_end); - - /* Set crc mode and selection, and enable. Only using CRC0*/ - REG_UPDATE_3(OTG_CRC_CNTL, - OTG_CRC_CONT_EN, params->continuous_mode ? 1 : 0, - OTG_CRC0_SELECT, params->selection, - OTG_CRC_EN, 1); + switch (params->crc_eng_inst) { + case 0: + /* Window A x axis start and end. */ + REG_UPDATE_2(OTG_CRC0_WINDOWA_X_CONTROL, + OTG_CRC0_WINDOWA_X_START, params->windowa_x_start, + OTG_CRC0_WINDOWA_X_END, params->windowa_x_end); + + /* Window A y axis start and end. */ + REG_UPDATE_2(OTG_CRC0_WINDOWA_Y_CONTROL, + OTG_CRC0_WINDOWA_Y_START, params->windowa_y_start, + OTG_CRC0_WINDOWA_Y_END, params->windowa_y_end); + + /* Window B x axis start and end. */ + REG_UPDATE_2(OTG_CRC0_WINDOWB_X_CONTROL, + OTG_CRC0_WINDOWB_X_START, params->windowb_x_start, + OTG_CRC0_WINDOWB_X_END, params->windowb_x_end); + + /* Window B y axis start and end. */ + REG_UPDATE_2(OTG_CRC0_WINDOWB_Y_CONTROL, + OTG_CRC0_WINDOWB_Y_START, params->windowb_y_start, + OTG_CRC0_WINDOWB_Y_END, params->windowb_y_end); + + /* Set crc mode and selection, and enable.*/ + REG_UPDATE_3(OTG_CRC_CNTL, + OTG_CRC_CONT_EN, params->continuous_mode ? 1 : 0, + OTG_CRC0_SELECT, params->selection, + OTG_CRC_EN, 1); + break; + case 1: + /* Window A x axis start and end. */ + REG_UPDATE_2(OTG_CRC1_WINDOWA_X_CONTROL, + OTG_CRC1_WINDOWA_X_START, params->windowa_x_start, + OTG_CRC1_WINDOWA_X_END, params->windowa_x_end); + + /* Window A y axis start and end. */ + REG_UPDATE_2(OTG_CRC1_WINDOWA_Y_CONTROL, + OTG_CRC1_WINDOWA_Y_START, params->windowa_y_start, + OTG_CRC1_WINDOWA_Y_END, params->windowa_y_end); + + /* Window B x axis start and end. */ + REG_UPDATE_2(OTG_CRC1_WINDOWB_X_CONTROL, + OTG_CRC1_WINDOWB_X_START, params->windowb_x_start, + OTG_CRC1_WINDOWB_X_END, params->windowb_x_end); + + /* Window B y axis start and end. */ + REG_UPDATE_2(OTG_CRC1_WINDOWB_Y_CONTROL, + OTG_CRC1_WINDOWB_Y_START, params->windowb_y_start, + OTG_CRC1_WINDOWB_Y_END, params->windowb_y_end); + + /* Set crc mode and selection, and enable.*/ + REG_UPDATE_3(OTG_CRC_CNTL, + OTG_CRC_CONT_EN, params->continuous_mode ? 1 : 0, + OTG_CRC1_SELECT, params->selection, + OTG_CRC_EN, 1); + break; + default: + return false; + } return true; } diff --git a/drivers/gpu/drm/amd/display/dc/optc/dcn35/dcn35_optc.c b/drivers/gpu/drm/amd/display/dc/optc/dcn35/dcn35_optc.c index dfa9364fe5a6..d21e82b927d0 100644 --- a/drivers/gpu/drm/amd/display/dc/optc/dcn35/dcn35_optc.c +++ b/drivers/gpu/drm/amd/display/dc/optc/dcn35/dcn35_optc.c @@ -183,34 +183,87 @@ static bool optc35_configure_crc(struct timing_generator *optc, { struct optc *optc1 = DCN10TG_FROM_TG(optc); + /* Cannot configure crc on a CRTC that is disabled */ if (!optc1_is_tg_enabled(optc)) return false; - REG_WRITE(OTG_CRC_CNTL, 0); + + if (!params->enable || params->reset) + REG_WRITE(OTG_CRC_CNTL, 0); + if (!params->enable) return true; - REG_UPDATE_2(OTG_CRC0_WINDOWA_X_CONTROL, - OTG_CRC0_WINDOWA_X_START, params->windowa_x_start, - OTG_CRC0_WINDOWA_X_END, params->windowa_x_end); - REG_UPDATE_2(OTG_CRC0_WINDOWA_Y_CONTROL, - OTG_CRC0_WINDOWA_Y_START, params->windowa_y_start, - OTG_CRC0_WINDOWA_Y_END, params->windowa_y_end); - REG_UPDATE_2(OTG_CRC0_WINDOWB_X_CONTROL, - OTG_CRC0_WINDOWB_X_START, params->windowb_x_start, - OTG_CRC0_WINDOWB_X_END, params->windowb_x_end); - REG_UPDATE_2(OTG_CRC0_WINDOWB_Y_CONTROL, - OTG_CRC0_WINDOWB_Y_START, params->windowb_y_start, - OTG_CRC0_WINDOWB_Y_END, params->windowb_y_end); - if (optc1->base.ctx->dc->debug.otg_crc_db && optc1->tg_mask->OTG_CRC_WINDOW_DB_EN != 0) { - REG_UPDATE_4(OTG_CRC_CNTL, - OTG_CRC_CONT_EN, params->continuous_mode ? 1 : 0, - OTG_CRC0_SELECT, params->selection, - OTG_CRC_EN, 1, - OTG_CRC_WINDOW_DB_EN, 1); - } else - REG_UPDATE_3(OTG_CRC_CNTL, - OTG_CRC_CONT_EN, params->continuous_mode ? 1 : 0, - OTG_CRC0_SELECT, params->selection, - OTG_CRC_EN, 1); + + /* Program frame boundaries */ + switch (params->crc_eng_inst) { + case 0: + /* Window A x axis start and end. */ + REG_UPDATE_2(OTG_CRC0_WINDOWA_X_CONTROL, + OTG_CRC0_WINDOWA_X_START, params->windowa_x_start, + OTG_CRC0_WINDOWA_X_END, params->windowa_x_end); + + /* Window A y axis start and end. */ + REG_UPDATE_2(OTG_CRC0_WINDOWA_Y_CONTROL, + OTG_CRC0_WINDOWA_Y_START, params->windowa_y_start, + OTG_CRC0_WINDOWA_Y_END, params->windowa_y_end); + + /* Window B x axis start and end. */ + REG_UPDATE_2(OTG_CRC0_WINDOWB_X_CONTROL, + OTG_CRC0_WINDOWB_X_START, params->windowb_x_start, + OTG_CRC0_WINDOWB_X_END, params->windowb_x_end); + + /* Window B y axis start and end. */ + REG_UPDATE_2(OTG_CRC0_WINDOWB_Y_CONTROL, + OTG_CRC0_WINDOWB_Y_START, params->windowb_y_start, + OTG_CRC0_WINDOWB_Y_END, params->windowb_y_end); + + if (optc1->base.ctx->dc->debug.otg_crc_db && optc1->tg_mask->OTG_CRC_WINDOW_DB_EN != 0) + REG_UPDATE_4(OTG_CRC_CNTL, + OTG_CRC_CONT_EN, params->continuous_mode ? 1 : 0, + OTG_CRC0_SELECT, params->selection, + OTG_CRC_EN, 1, + OTG_CRC_WINDOW_DB_EN, 1); + else + REG_UPDATE_3(OTG_CRC_CNTL, + OTG_CRC_CONT_EN, params->continuous_mode ? 1 : 0, + OTG_CRC0_SELECT, params->selection, + OTG_CRC_EN, 1); + break; + case 1: + /* Window A x axis start and end. */ + REG_UPDATE_2(OTG_CRC1_WINDOWA_X_CONTROL, + OTG_CRC1_WINDOWA_X_START, params->windowa_x_start, + OTG_CRC1_WINDOWA_X_END, params->windowa_x_end); + + /* Window A y axis start and end. */ + REG_UPDATE_2(OTG_CRC1_WINDOWA_Y_CONTROL, + OTG_CRC1_WINDOWA_Y_START, params->windowa_y_start, + OTG_CRC1_WINDOWA_Y_END, params->windowa_y_end); + + /* Window B x axis start and end. */ + REG_UPDATE_2(OTG_CRC1_WINDOWB_X_CONTROL, + OTG_CRC1_WINDOWB_X_START, params->windowb_x_start, + OTG_CRC1_WINDOWB_X_END, params->windowb_x_end); + + /* Window B y axis start and end. */ + REG_UPDATE_2(OTG_CRC1_WINDOWB_Y_CONTROL, + OTG_CRC1_WINDOWB_Y_START, params->windowb_y_start, + OTG_CRC1_WINDOWB_Y_END, params->windowb_y_end); + + if (optc1->base.ctx->dc->debug.otg_crc_db && optc1->tg_mask->OTG_CRC_WINDOW_DB_EN != 0) + REG_UPDATE_4(OTG_CRC_CNTL, + OTG_CRC_CONT_EN, params->continuous_mode ? 1 : 0, + OTG_CRC1_SELECT, params->selection, + OTG_CRC_EN, 1, + OTG_CRC_WINDOW_DB_EN, 1); + else + REG_UPDATE_3(OTG_CRC_CNTL, + OTG_CRC_CONT_EN, params->continuous_mode ? 1 : 0, + OTG_CRC1_SELECT, params->selection, + OTG_CRC_EN, 1); + break; + default: + return false; + } return true; } -- 2.51.0 From 44cea2bb9c872594e538412eb9c780b391eb112c Mon Sep 17 00:00:00 2001 From: Wayne Lin Date: Thu, 31 Oct 2024 15:57:40 +0800 Subject: [PATCH 02/16] drm/amd/display: Extend secure display to support DisplayCRC mode [Why] For the legacy secure display, it involves PSP + DMUB to confgiure and retrieve the CRC/ROI result. Have requirement to support mode which all handled by driver only. [How] Add another "DisplayCRC" mode, which doesn't involve PSP + DMUB. All things are handled by the driver only Reviewed-by: HaoPing Liu Signed-off-by: Wayne Lin Signed-off-by: Tom Chung Tested-by: Daniel Wheeler Signed-off-by: Alex Deucher --- .../drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c | 56 +++++++++++++------ .../drm/amd/display/amdgpu_dm/amdgpu_dm_crc.h | 9 +++ 2 files changed, 49 insertions(+), 16 deletions(-) diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c index a2e36b882a4f..2412b9d7c86f 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c @@ -316,7 +316,7 @@ static void amdgpu_dm_set_crc_window_default(struct drm_crtc *crtc, struct dc_st spin_unlock_irqrestore(&drm_dev->event_lock, flags); /* Disable secure_display if it was enabled */ - if (was_activated) { + if (was_activated && dm->secure_display_ctx.op_mode == LEGACY_MODE) { /* stop ROI update on this crtc */ flush_work(&dm->secure_display_ctx.crtc_ctx[crtc->index].notify_ta_work); flush_work(&dm->secure_display_ctx.crtc_ctx[crtc->index].forward_roi_work); @@ -691,7 +691,8 @@ int amdgpu_dm_crtc_set_crc_source(struct drm_crtc *crtc, const char *src_name) #if defined(CONFIG_DRM_AMD_SECURE_DISPLAY) /* Initialize phy id mapping table for secure display*/ - if (!dm->secure_display_ctx.phy_mapping_updated) + if (dm->secure_display_ctx.op_mode == LEGACY_MODE && + !dm->secure_display_ctx.phy_mapping_updated) update_phy_id_mapping(adev); #endif @@ -774,6 +775,7 @@ void amdgpu_dm_crtc_handle_crc_window_irq(struct drm_crtc *crtc) bool forward_roi_change = false; bool notify_ta = false; bool all_crc_ready = true; + struct dc_stream_state *stream_state; int i; if (crtc == NULL) @@ -782,6 +784,7 @@ void amdgpu_dm_crtc_handle_crc_window_irq(struct drm_crtc *crtc) acrtc = to_amdgpu_crtc(crtc); adev = drm_to_adev(crtc->dev); drm_dev = crtc->dev; + stream_state = to_dm_crtc_state(crtc->state)->stream; spin_lock_irqsave(&drm_dev->event_lock, flags1); cur_crc_src = acrtc->dm_irq_params.crc_src; @@ -807,6 +810,17 @@ void amdgpu_dm_crtc_handle_crc_window_irq(struct drm_crtc *crtc) } for (i = 0; i < MAX_CRC_WINDOW_NUM; i++) { + struct crc_params crc_window = { + .windowa_x_start = acrtc->dm_irq_params.window_param[i].x_start, + .windowa_y_start = acrtc->dm_irq_params.window_param[i].y_start, + .windowa_x_end = acrtc->dm_irq_params.window_param[i].x_end, + .windowa_y_end = acrtc->dm_irq_params.window_param[i].y_end, + .windowb_x_start = acrtc->dm_irq_params.window_param[i].x_start, + .windowb_y_start = acrtc->dm_irq_params.window_param[i].y_start, + .windowb_x_end = acrtc->dm_irq_params.window_param[i].x_end, + .windowb_y_end = acrtc->dm_irq_params.window_param[i].y_end, + }; + crtc_ctx->roi[i].enable = acrtc->dm_irq_params.window_param[i].enable; if (!acrtc->dm_irq_params.window_param[i].enable) { @@ -821,15 +835,20 @@ void amdgpu_dm_crtc_handle_crc_window_irq(struct drm_crtc *crtc) } if (acrtc->dm_irq_params.window_param[i].update_win) { - /* prepare work for dmub to update ROI */ - crtc_ctx->roi[i].rect.x = acrtc->dm_irq_params.window_param[i].x_start; - crtc_ctx->roi[i].rect.y = acrtc->dm_irq_params.window_param[i].y_start; - crtc_ctx->roi[i].rect.width = acrtc->dm_irq_params.window_param[i].x_end - - acrtc->dm_irq_params.window_param[i].x_start; - crtc_ctx->roi[i].rect.height = acrtc->dm_irq_params.window_param[i].y_end - - acrtc->dm_irq_params.window_param[i].y_start; - - forward_roi_change = true; + crtc_ctx->roi[i].rect.x = crc_window.windowa_x_start; + crtc_ctx->roi[i].rect.y = crc_window.windowa_y_start; + crtc_ctx->roi[i].rect.width = crc_window.windowa_x_end - + crc_window.windowa_x_start; + crtc_ctx->roi[i].rect.height = crc_window.windowa_y_end - + crc_window.windowa_y_start; + + if (adev->dm.secure_display_ctx.op_mode == LEGACY_MODE) + /* forward task to dmub to update ROI */ + forward_roi_change = true; + else if (adev->dm.secure_display_ctx.op_mode == DISPLAY_CRC_MODE) + /* update ROI via dm*/ + dc_stream_configure_crc(stream_state->ctx->dc, stream_state, + &crc_window, true, true, i, false); reset_crc_frame_count[i] = true; @@ -843,14 +862,18 @@ void amdgpu_dm_crtc_handle_crc_window_irq(struct drm_crtc *crtc) acrtc->dm_irq_params.window_param[i].skip_frame_cnt = 1; crtc_ctx->crc_info.crc[i].crc_ready = false; } else { - struct dc_stream_state *stream_state = to_dm_crtc_state(crtc->state)->stream; - if (!dc_stream_get_crc(stream_state->ctx->dc, stream_state, i, &crc_r[i], &crc_g[i], &crc_b[i])) DRM_ERROR("Secure Display: fail to get crc from engine %d\n", i); - /* prepare work for psp to read ROI/CRC and send to I2C */ - notify_ta = true; + if (adev->dm.secure_display_ctx.op_mode == LEGACY_MODE) + /* forward task to psp to read ROI/CRC and output via I2C */ + notify_ta = true; + else if (adev->dm.secure_display_ctx.op_mode == DISPLAY_CRC_MODE) + /* Avoid ROI window get changed, keep overwriting. */ + dc_stream_configure_crc(stream_state->ctx->dc, stream_state, + &crc_window, true, true, i, false); + /* crc ready for psp to read out */ crtc_ctx->crc_info.crc[i].crc_ready = true; } @@ -914,6 +937,7 @@ void amdgpu_dm_crtc_secure_display_create_contexts(struct amdgpu_device *adev) } adev->dm.secure_display_ctx.crtc_ctx = crtc_ctx; - return; + + adev->dm.secure_display_ctx.op_mode = DISPLAY_CRC_MODE; } #endif diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.h b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.h index 0c96f2db1352..3da056c8d20b 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.h +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.h @@ -42,6 +42,14 @@ enum amdgpu_dm_pipe_crc_source { #ifdef CONFIG_DRM_AMD_SECURE_DISPLAY #define MAX_CRTC 6 +enum secure_display_mode { + /* via dmub + psp */ + LEGACY_MODE = 0, + /* driver directly */ + DISPLAY_CRC_MODE, + SECURE_DISPLAY_MODE_MAX, +}; + struct phy_id_mapping { bool assigned; bool is_mst; @@ -98,6 +106,7 @@ struct secure_display_context { struct secure_display_crtc_context *crtc_ctx; /* Whether dmub support multiple ROI setting */ bool support_mul_roi; + enum secure_display_mode op_mode; bool phy_mapping_updated; int phy_id_mapping_cnt; struct phy_id_mapping phy_id_mapping[MAX_CRTC]; -- 2.51.0 From e2c4c6c10542ccfe4a0830bb6c9fd5b177b7bbb7 Mon Sep 17 00:00:00 2001 From: Alex Hung Date: Tue, 17 Dec 2024 11:51:58 -0700 Subject: [PATCH 03/16] drm/amd/display: Initialize denominator defaults to 1 [WHAT & HOW] Variables, used as denominators and maybe not assigned to other values, should be initialized to non-zero to avoid DIVIDE_BY_ZERO, as reported by Coverity. Reviewed-by: Austin Zheng Reviewed-by: Rodrigo Siqueira Signed-off-by: Alex Hung Signed-off-by: Tom Chung Tested-by: Daniel Wheeler Signed-off-by: Alex Deucher --- .../dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c index b9ec243cf9ba..a3ffb8c95d25 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c @@ -600,8 +600,8 @@ static void CalculateBytePerPixelAndBlockSizes( { *BytePerPixelDETY = 0; *BytePerPixelDETC = 0; - *BytePerPixelY = 0; - *BytePerPixelC = 0; + *BytePerPixelY = 1; + *BytePerPixelC = 1; if (SourcePixelFormat == dml2_444_64) { *BytePerPixelDETY = 8; -- 2.51.0 From 3ea943991d09a8d5cdddec5356d2694bd8060f69 Mon Sep 17 00:00:00 2001 From: Dillon Varone Date: Thu, 12 Dec 2024 17:01:34 -0500 Subject: [PATCH 04/16] drm/amd/display: Add SMU interface to get UMC count for dcn401 [WHY&HOW] BIOS table will not always contain accurate UMC channel info when harvesting is enabled, so get the correct info from SMU. Reviewed-by: Alvin Lee Signed-off-by: Dillon Varone Signed-off-by: Tom Chung Tested-by: Daniel Wheeler Signed-off-by: Alex Deucher --- .../amd/display/dc/clk_mgr/dcn401/dalsmc.h | 4 +- .../dc/clk_mgr/dcn401/dcn401_clk_mgr.c | 9 ++++ .../clk_mgr/dcn401/dcn401_clk_mgr_smu_msg.c | 14 +++++ .../clk_mgr/dcn401/dcn401_clk_mgr_smu_msg.h | 1 + .../dc/dml2/dml21/dml21_translation_helper.c | 9 +++- .../dc/resource/dcn401/dcn401_resource.c | 54 +++++++++++-------- 6 files changed, 65 insertions(+), 26 deletions(-) diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dalsmc.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dalsmc.h index dbfdd3487da5..2e0d34fd7512 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dalsmc.h +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dalsmc.h @@ -43,7 +43,9 @@ #define DALSMC_MSG_ActiveUclkFclk 0x18 #define DALSMC_MSG_IdleUclkFclk 0x19 #define DALSMC_MSG_SetUclkPstateAllow 0x1A -#define DALSMC_Message_Count 0x1B +#define DALSMC_MSG_SubvpUclkFclk 0x1B +#define DALSMC_MSG_GetNumUmcChannels 0x1C +#define DALSMC_Message_Count 0x1D typedef enum { FCLK_SWITCH_DISALLOW, diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr.c index 204ebdd99415..8082bb877611 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr.c @@ -1403,6 +1403,15 @@ static void dcn401_get_memclk_states_from_smu(struct clk_mgr *clk_mgr_base) if (clk_mgr->dpm_present && !num_levels) clk_mgr->dpm_present = false; + clk_mgr_base->bw_params->num_channels = dcn401_smu_get_num_of_umc_channels(clk_mgr); + if (clk_mgr_base->ctx->dc_bios) { + /* use BIOS values if none provided by PMFW */ + if (clk_mgr_base->bw_params->num_channels == 0) { + clk_mgr_base->bw_params->num_channels = clk_mgr_base->ctx->dc_bios->vram_info.num_chans; + } + clk_mgr_base->bw_params->dram_channel_width_bytes = clk_mgr_base->ctx->dc_bios->vram_info.dram_channel_width_bytes; + } + /* Refresh bounding box */ clk_mgr_base->ctx->dc->res_pool->funcs->update_bw_bounding_box( clk_mgr->base.ctx->dc, clk_mgr_base->bw_params); diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr_smu_msg.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr_smu_msg.c index b02a41179b41..21c35528f61f 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr_smu_msg.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr_smu_msg.c @@ -25,6 +25,9 @@ #ifndef DALSMC_MSG_SubvpUclkFclk #define DALSMC_MSG_SubvpUclkFclk 0x1B #endif +#ifndef DALSMC_MSG_GetNumUmcChannels +#define DALSMC_MSG_GetNumUmcChannels 0x1C +#endif /* * Function to be used instead of REG_WAIT macro because the wait ends when @@ -334,3 +337,14 @@ void dcn401_smu_set_num_of_displays(struct clk_mgr_internal *clk_mgr, uint32_t n dcn401_smu_send_msg_with_param(clk_mgr, DALSMC_MSG_NumOfDisplays, num_displays, NULL); } + +unsigned int dcn401_smu_get_num_of_umc_channels(struct clk_mgr_internal *clk_mgr) +{ + unsigned int response = 0; + + dcn401_smu_send_msg_with_param(clk_mgr, DALSMC_MSG_GetNumUmcChannels, 0, &response); + + smu_print("SMU Get Num UMC Channels: num_umc_channels = %d\n", response); + + return response; +} diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr_smu_msg.h b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr_smu_msg.h index 42cf7885a7cb..e02eb1294b37 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr_smu_msg.h +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr_smu_msg.h @@ -28,5 +28,6 @@ bool dcn401_smu_set_subvp_uclk_fclk_hardmin(struct clk_mgr_internal *clk_mgr, uint16_t fclk_freq_mhz); void dcn401_smu_set_min_deep_sleep_dcef_clk(struct clk_mgr_internal *clk_mgr, uint32_t freq_mhz); void dcn401_smu_set_num_of_displays(struct clk_mgr_internal *clk_mgr, uint32_t num_displays); +unsigned int dcn401_smu_get_num_of_umc_channels(struct clk_mgr_internal *clk_mgr); #endif /* __DCN401_CLK_MGR_SMU_MSG_H_ */ diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.c index 5d46f4e24f6b..b9c6b45f6872 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.c @@ -294,12 +294,17 @@ void dml21_apply_soc_bb_overrides(struct dml2_initialize_instance_in_out *dml_in dml_soc_bb->power_management_parameters.stutter_exit_latency_us = (in_dc->ctx->dc_bios->bb_info.dram_sr_exit_latency_100ns + 9) / 10; - if (in_dc->ctx->dc_bios->vram_info.num_chans) { + if (dc_bw_params->num_channels) { + dml_clk_table->dram_config.channel_count = dc_bw_params->num_channels; + dml_soc_bb->mall_allocated_for_dcn_mbytes = in_dc->caps.mall_size_total / 1048576; + } else if (in_dc->ctx->dc_bios->vram_info.num_chans) { dml_clk_table->dram_config.channel_count = in_dc->ctx->dc_bios->vram_info.num_chans; dml_soc_bb->mall_allocated_for_dcn_mbytes = in_dc->caps.mall_size_total / 1048576; } - if (in_dc->ctx->dc_bios->vram_info.dram_channel_width_bytes) { + if (dc_bw_params->dram_channel_width_bytes) { + dml_clk_table->dram_config.channel_width_bytes = dc_bw_params->dram_channel_width_bytes; + } else if (in_dc->ctx->dc_bios->vram_info.dram_channel_width_bytes) { dml_clk_table->dram_config.channel_width_bytes = in_dc->ctx->dc_bios->vram_info.dram_channel_width_bytes; } diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn401/dcn401_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn401/dcn401_resource.c index 0767adaa6389..c1ebc6b1c937 100644 --- a/drivers/gpu/drm/amd/display/dc/resource/dcn401/dcn401_resource.c +++ b/drivers/gpu/drm/amd/display/dc/resource/dcn401/dcn401_resource.c @@ -1297,6 +1297,29 @@ static struct hpo_dp_link_encoder *dcn401_hpo_dp_link_encoder_create( return &hpo_dp_enc31->base; } +static unsigned int dcn401_calc_num_avail_chans_for_mall(struct dc *dc, unsigned int num_chans) +{ + unsigned int num_available_chans = 1; + + /* channels for MALL must be a power of 2 */ + while (num_chans > 1) { + num_available_chans = (num_available_chans << 1); + num_chans = (num_chans >> 1); + } + + /* cannot be odd */ + num_available_chans &= ~1; + + /* clamp to max available channels for MALL per ASIC */ + if (ASICREV_IS_GC_12_0_0_A0(dc->ctx->asic_id.hw_internal_rev)) { + num_available_chans = num_available_chans > 16 ? 16 : num_available_chans; + } else if (ASICREV_IS_GC_12_0_1_A0(dc->ctx->asic_id.hw_internal_rev)) { + num_available_chans = num_available_chans > 8 ? 8 : num_available_chans; + } + + return num_available_chans; +} + static struct dce_hwseq *dcn401_hwseq_create( struct dc_context *ctx) { @@ -1592,6 +1615,14 @@ static void dcn401_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *b memcpy(dml2_opt, &dc->dml2_options, sizeof(dc->dml2_options)); + /* re-calculate the available MALL size if required */ + if (bw_params->num_channels > 0) { + dc->caps.max_cab_allocation_bytes = dcn401_calc_num_avail_chans_for_mall( + dc, bw_params->num_channels) * + dc->caps.mall_size_per_mem_channel * 1024 * 1024; + dc->caps.mall_size_total = dc->caps.max_cab_allocation_bytes; + } + DC_FP_START(); dcn401_update_bw_bounding_box_fpu(dc, bw_params); @@ -1714,29 +1745,6 @@ static unsigned int dcn401_get_vstartup_for_pipe(struct pipe_ctx *pipe_ctx) return pipe_ctx->global_sync.dcn4x.vstartup_lines; } -static unsigned int dcn401_calc_num_avail_chans_for_mall(struct dc *dc, unsigned int num_chans) -{ - unsigned int num_available_chans = 1; - - /* channels for MALL must be a power of 2 */ - while (num_chans > 1) { - num_available_chans = (num_available_chans << 1); - num_chans = (num_chans >> 1); - } - - /* cannot be odd */ - num_available_chans &= ~1; - - /* clamp to max available channels for MALL per ASIC */ - if (ASICREV_IS_GC_12_0_0_A0(dc->ctx->asic_id.hw_internal_rev)) { - num_available_chans = num_available_chans > 16 ? 16 : num_available_chans; - } else if (ASICREV_IS_GC_12_0_1_A0(dc->ctx->asic_id.hw_internal_rev)) { - num_available_chans = num_available_chans > 8 ? 8 : num_available_chans; - } - - return num_available_chans; -} - static struct resource_funcs dcn401_res_pool_funcs = { .destroy = dcn401_destroy_resource_pool, .link_enc_create = dcn401_link_encoder_create, -- 2.51.0 From c7ccfc0d4241a834c25a9a9e1e78b388b4445d23 Mon Sep 17 00:00:00 2001 From: Nicholas Susanto Date: Thu, 19 Dec 2024 14:15:37 -0500 Subject: [PATCH 05/16] Revert "drm/amd/display: Enable urgent latency adjustments for DCN35" Revert commit 284f141f5ce5 ("drm/amd/display: Enable urgent latency adjustments for DCN35") [Why & How] Urgent latency increase caused 2.8K OLED monitor caused it to block this panel support P0. Reverting this change does not reintroduce the netflix corruption issue which it fixed. Fixes: 284f141f5ce5 ("drm/amd/display: Enable urgent latency adjustments for DCN35") Reviewed-by: Charlene Liu Signed-off-by: Nicholas Susanto Signed-off-by: Tom Chung Tested-by: Daniel Wheeler Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c index beed7adbbd43..47d785204f29 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c @@ -195,9 +195,9 @@ struct _vcs_dpi_soc_bounding_box_st dcn3_5_soc = { .dcn_downspread_percent = 0.5, .gpuvm_min_page_size_bytes = 4096, .hostvm_min_page_size_bytes = 4096, - .do_urgent_latency_adjustment = 1, + .do_urgent_latency_adjustment = 0, .urgent_latency_adjustment_fabric_clock_component_us = 0, - .urgent_latency_adjustment_fabric_clock_reference_mhz = 3000, + .urgent_latency_adjustment_fabric_clock_reference_mhz = 0, }; void dcn35_build_wm_range_table_fpu(struct clk_mgr *clk_mgr) -- 2.51.0 From f5860c88cdfe7300d08c1aef881bba0cac369e34 Mon Sep 17 00:00:00 2001 From: Yiling Chen Date: Fri, 20 Dec 2024 17:12:54 +0800 Subject: [PATCH 06/16] drm/amd/display: Remove unnecessary eDP power down [why] When first time of link training is fail, eDP would be powered down and would not be powered up for next retry link training. It causes that all of retry link linking would be fail. [how] We has extracted both power up and down sequence from enable/disable link output function before DCN32. We remov eDP power down in dcn32_disable_link_output(). Reviewed-by: Charlene Liu Signed-off-by: Yiling Chen Signed-off-by: Tom Chung Tested-by: Daniel Wheeler Signed-off-by: Alex Deucher --- .../gpu/drm/amd/display/dc/hwss/dcn32/dcn32_hwseq.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn32/dcn32_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn32/dcn32_hwseq.c index 2e6ecf939fbd..ee4de9ddfef4 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/dcn32/dcn32_hwseq.c +++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn32/dcn32_hwseq.c @@ -1399,12 +1399,12 @@ void dcn32_disable_link_output(struct dc_link *link, link_hwss->disable_link_output(link, link_res, signal); link->phy_state.symclk_state = SYMCLK_OFF_TX_OFF; - - if (signal == SIGNAL_TYPE_EDP && - link->dc->hwss.edp_power_control && - !link->skip_implict_edp_power_control) - link->dc->hwss.edp_power_control(link, false); - else if (dmcu != NULL && dmcu->funcs->unlock_phy) + /* + * Add the logic to extract BOTH power up and power down sequences + * from enable/disable link output and only call edp panel control + * in enable_link_dp and disable_link_dp once. + */ + if (dmcu != NULL && dmcu->funcs->unlock_phy) dmcu->funcs->unlock_phy(dmcu); dc->link_srv->dp_trace_source_sequence(link, DPCD_SOURCE_SEQ_AFTER_DISABLE_LINK_PHY); -- 2.51.0 From aa6713fa2046f4c09bf3013dd1420ae15603ca6f Mon Sep 17 00:00:00 2001 From: Leo Li Date: Mon, 9 Dec 2024 12:58:33 -0500 Subject: [PATCH 07/16] drm/amd/display: Do not wait for PSR disable on vbl enable [Why] Outside of a modeset/link configuration change, we should not have to wait for the panel to exit PSR. Depending on the panel and it's state, it may take multiple frames for it to exit PSR. Therefore, waiting in all scenarios may cause perceived stuttering, especially in combination with faster vblank shutdown. [How] PSR1 disable is hooked up to the vblank enable event, and vice versa. In case of vblank enable, do not wait for panel to exit PSR, but still wait in all other cases. We also avoid a call to unnecessarily change power_opts on disable - this ends up sending another command to dmcub fw. When testing against IGT, some crc tests like kms_plane_alpha_blend and amd_hotplug were failing due to CRC timeouts. This was found to be caused by the early return before HW has fully exited PSR1. Fix this by first making sure we grab a vblank reference, then waiting for panel to exit PSR1, before programming hw for CRC generation. Fixes: 58a261bfc967 ("drm/amd/display: use a more lax vblank enable policy for older ASICs") Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/3743 Reviewed-by: Tom Chung Signed-off-by: Leo Li Signed-off-by: Tom Chung Tested-by: Daniel Wheeler Signed-off-by: Alex Deucher --- .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 4 +-- .../drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c | 25 ++++++++----- .../amd/display/amdgpu_dm/amdgpu_dm_crtc.c | 2 +- .../amd/display/amdgpu_dm/amdgpu_dm_debugfs.c | 2 +- .../drm/amd/display/amdgpu_dm/amdgpu_dm_psr.c | 35 +++++++++++++++++-- .../drm/amd/display/amdgpu_dm/amdgpu_dm_psr.h | 3 +- 6 files changed, 54 insertions(+), 17 deletions(-) diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c index 1db955c287ae..3ad548254d6c 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c @@ -9180,7 +9180,7 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state, acrtc_state->stream->link->psr_settings.psr_dirty_rects_change_timestamp_ns = timestamp_ns; if (acrtc_state->stream->link->psr_settings.psr_allow_active) - amdgpu_dm_psr_disable(acrtc_state->stream); + amdgpu_dm_psr_disable(acrtc_state->stream, true); mutex_unlock(&dm->dc_lock); } } @@ -9350,7 +9350,7 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state, if (acrtc_state->stream->link->replay_settings.replay_allow_active) amdgpu_dm_replay_disable(acrtc_state->stream); if (acrtc_state->stream->link->psr_settings.psr_allow_active) - amdgpu_dm_psr_disable(acrtc_state->stream); + amdgpu_dm_psr_disable(acrtc_state->stream, true); } mutex_unlock(&dm->dc_lock); diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c index 2412b9d7c86f..033bd817d871 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c @@ -30,6 +30,7 @@ #include "amdgpu_dm.h" #include "dc.h" #include "amdgpu_securedisplay.h" +#include "amdgpu_dm_psr.h" static const char *const pipe_crc_sources[] = { "none", @@ -507,6 +508,10 @@ int amdgpu_dm_crtc_configure_crc_source(struct drm_crtc *crtc, mutex_lock(&adev->dm.dc_lock); + /* For PSR1, check that the panel has exited PSR */ + if (stream_state->link->psr_settings.psr_version < DC_PSR_VERSION_SU_1) + amdgpu_dm_psr_wait_disable(stream_state); + /* Enable or disable CRTC CRC generation */ if (dm_is_crc_source_crtc(source) || source == AMDGPU_DM_PIPE_CRC_SOURCE_NONE) { if (!dc_stream_configure_crc(stream_state->ctx->dc, @@ -644,6 +649,17 @@ int amdgpu_dm_crtc_set_crc_source(struct drm_crtc *crtc, const char *src_name) } + /* + * Reading the CRC requires the vblank interrupt handler to be + * enabled. Keep a reference until CRC capture stops. + */ + enabled = amdgpu_dm_is_valid_crc_source(cur_crc_src); + if (!enabled && enable) { + ret = drm_crtc_vblank_get(crtc); + if (ret) + goto cleanup; + } + #if defined(CONFIG_DRM_AMD_SECURE_DISPLAY) /* Reset secure_display when we change crc source from debugfs */ amdgpu_dm_set_crc_window_default(crtc, crtc_state->stream); @@ -654,16 +670,7 @@ int amdgpu_dm_crtc_set_crc_source(struct drm_crtc *crtc, const char *src_name) goto cleanup; } - /* - * Reading the CRC requires the vblank interrupt handler to be - * enabled. Keep a reference until CRC capture stops. - */ - enabled = amdgpu_dm_is_valid_crc_source(cur_crc_src); if (!enabled && enable) { - ret = drm_crtc_vblank_get(crtc); - if (ret) - goto cleanup; - if (dm_is_crc_source_dprx(source)) { if (drm_dp_start_crc(aux, crtc)) { DRM_DEBUG_DRIVER("dp start crc failed\n"); diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crtc.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crtc.c index 8bc73922e3a6..36a830a7440f 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crtc.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crtc.c @@ -142,7 +142,7 @@ static void amdgpu_dm_crtc_set_panel_sr_feature( amdgpu_dm_replay_enable(vblank_work->stream, true); } else if (vblank_enabled) { if (link->psr_settings.psr_version < DC_PSR_VERSION_SU_1 && is_sr_active) - amdgpu_dm_psr_disable(vblank_work->stream); + amdgpu_dm_psr_disable(vblank_work->stream, false); } else if (link->psr_settings.psr_feature_enabled && allow_sr_entry && !is_sr_active && !is_crc_window_active) { diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c index a872e047d199..049046c60462 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c @@ -3693,7 +3693,7 @@ static int crc_win_update_set(void *data, u64 val) /* PSR may write to OTG CRC window control register, * so close it before starting secure_display. */ - amdgpu_dm_psr_disable(acrtc->dm_irq_params.stream); + amdgpu_dm_psr_disable(acrtc->dm_irq_params.stream, true); spin_lock_irq(&adev_to_drm(adev)->event_lock); diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_psr.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_psr.c index f40240aafe98..45858bf1523d 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_psr.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_psr.c @@ -201,14 +201,13 @@ void amdgpu_dm_psr_enable(struct dc_stream_state *stream) * * Return: true if success */ -bool amdgpu_dm_psr_disable(struct dc_stream_state *stream) +bool amdgpu_dm_psr_disable(struct dc_stream_state *stream, bool wait) { - unsigned int power_opt = 0; bool psr_enable = false; DRM_DEBUG_DRIVER("Disabling psr...\n"); - return dc_link_set_psr_allow_active(stream->link, &psr_enable, true, false, &power_opt); + return dc_link_set_psr_allow_active(stream->link, &psr_enable, wait, false, NULL); } /* @@ -251,3 +250,33 @@ bool amdgpu_dm_psr_is_active_allowed(struct amdgpu_display_manager *dm) return allow_active; } + +/** + * amdgpu_dm_psr_wait_disable() - Wait for eDP panel to exit PSR + * @stream: stream state attached to the eDP link + * + * Waits for a max of 500ms for the eDP panel to exit PSR. + * + * Return: true if panel exited PSR, false otherwise. + */ +bool amdgpu_dm_psr_wait_disable(struct dc_stream_state *stream) +{ + enum dc_psr_state psr_state = PSR_STATE0; + struct dc_link *link = stream->link; + int retry_count; + + if (link == NULL) + return false; + + for (retry_count = 0; retry_count <= 1000; retry_count++) { + dc_link_get_psr_state(link, &psr_state); + if (psr_state == PSR_STATE0) + break; + udelay(500); + } + + if (retry_count == 1000) + return false; + + return true; +} diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_psr.h b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_psr.h index cd2d45c2b5ef..e2366321a3c1 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_psr.h +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_psr.h @@ -34,8 +34,9 @@ void amdgpu_dm_set_psr_caps(struct dc_link *link); void amdgpu_dm_psr_enable(struct dc_stream_state *stream); bool amdgpu_dm_link_setup_psr(struct dc_stream_state *stream); -bool amdgpu_dm_psr_disable(struct dc_stream_state *stream); +bool amdgpu_dm_psr_disable(struct dc_stream_state *stream, bool wait); bool amdgpu_dm_psr_disable_all(struct amdgpu_display_manager *dm); bool amdgpu_dm_psr_is_active_allowed(struct amdgpu_display_manager *dm); +bool amdgpu_dm_psr_wait_disable(struct dc_stream_state *stream); #endif /* AMDGPU_DM_AMDGPU_DM_PSR_H_ */ -- 2.51.0 From 4caacd1671b7a013ad04cd8b6398f002540bdd4d Mon Sep 17 00:00:00 2001 From: Leo Li Date: Wed, 11 Dec 2024 12:06:24 -0500 Subject: [PATCH 08/16] drm/amd/display: Do not elevate mem_type change to full update [Why] There should not be any need to revalidate bandwidth on memory placement change, since the fb is expected to be pinned to DCN-accessable memory before scanout. For APU it's DRAM, and DGPU, it's VRAM. However, async flips + memory type change needs to be rejected. [How] Do not set lock_and_validation_needed on mem_type change. Instead, reject an async_flip request if the crtc's buffer(s) changed mem_type. This may fix stuttering/corruption experienced with PSR SU and PSR1 panels, if the compositor allocates fbs in both VRAM carveout and GTT and flips between them. Fixes: a7c0cad0dc06 ("drm/amd/display: ensure async flips are only accepted for fast updates") Reviewed-by: Tom Chung Signed-off-by: Leo Li Signed-off-by: Tom Chung Tested-by: Daniel Wheeler Signed-off-by: Alex Deucher --- .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 29 +++++++++++++++---- 1 file changed, 23 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c index 3ad548254d6c..9121abe0e5ef 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c @@ -11464,6 +11464,25 @@ static int dm_crtc_get_cursor_mode(struct amdgpu_device *adev, return 0; } +static bool amdgpu_dm_crtc_mem_type_changed(struct drm_device *dev, + struct drm_atomic_state *state, + struct drm_crtc_state *crtc_state) +{ + struct drm_plane *plane; + struct drm_plane_state *new_plane_state, *old_plane_state; + + drm_for_each_plane_mask(plane, dev, crtc_state->plane_mask) { + new_plane_state = drm_atomic_get_plane_state(state, plane); + old_plane_state = drm_atomic_get_plane_state(state, plane); + + if (old_plane_state->fb && new_plane_state->fb && + get_mem_type(old_plane_state->fb) != get_mem_type(new_plane_state->fb)) + return true; + } + + return false; +} + /** * amdgpu_dm_atomic_check() - Atomic check implementation for AMDgpu DM. * @@ -11661,10 +11680,6 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev, /* Remove exiting planes if they are modified */ for_each_oldnew_plane_in_descending_zpos(state, plane, old_plane_state, new_plane_state) { - if (old_plane_state->fb && new_plane_state->fb && - get_mem_type(old_plane_state->fb) != - get_mem_type(new_plane_state->fb)) - lock_and_validation_needed = true; ret = dm_update_plane_state(dc, state, plane, old_plane_state, @@ -11959,9 +11974,11 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev, /* * Only allow async flips for fast updates that don't change - * the FB pitch, the DCC state, rotation, etc. + * the FB pitch, the DCC state, rotation, mem_type, etc. */ - if (new_crtc_state->async_flip && lock_and_validation_needed) { + if (new_crtc_state->async_flip && + (lock_and_validation_needed || + amdgpu_dm_crtc_mem_type_changed(dev, state, new_crtc_state))) { drm_dbg_atomic(crtc->dev, "[CRTC:%d:%s] async flips are only supported for fast updates\n", crtc->base.id, crtc->name); -- 2.51.0 From 92d100378c4b2e7f566518a1c60ce990a845ed26 Mon Sep 17 00:00:00 2001 From: Karthi Kandasamy Date: Fri, 20 Dec 2024 08:46:45 +0100 Subject: [PATCH 09/16] drm/amd/display: Remove unused read_ono_state function from Hwss module [Why] The functions read_ono_state are no longer in use and have been identified as redundant. Removing them helps streamline the codebase and improve maintainability by eliminating unnecessary code. [How] These unused functions were removed from Hwss module, ensuring that no functionality is affected, and the code is simplified. Reviewed-by: Martin Leung Signed-off-by: Karthi Kandasamy Signed-off-by: Tom Chung Tested-by: Daniel Wheeler Signed-off-by: Alex Deucher --- .../amd/display/dc/hwss/dcn401/dcn401_hwseq.c | 85 ------------------- .../amd/display/dc/hwss/dcn401/dcn401_hwseq.h | 2 - 2 files changed, 87 deletions(-) diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c index d5bbd3531c61..555a9f590cd7 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c +++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c @@ -127,91 +127,6 @@ void dcn401_program_gamut_remap(struct pipe_ctx *pipe_ctx) mpc->funcs->set_gamut_remap(mpc, mpcc_id, &mpc_adjust); } -struct ips_ono_region_state dcn401_read_ono_state(struct dc *dc, uint8_t region) -{ - struct dce_hwseq *hws = dc->hwseq; - struct ips_ono_region_state state = {0, 0}; - - switch (region) { - case 0: - /* dccg, dio, dcio */ - REG_GET_2(DOMAIN22_PG_STATUS, - DOMAIN_DESIRED_PWR_STATE, &state.desire_pwr_state, - DOMAIN_PGFSM_PWR_STATUS, &state.current_pwr_state); - break; - case 1: - /* dchubbub, dchvm, dchubbubmem */ - REG_GET_2(DOMAIN23_PG_STATUS, - DOMAIN_DESIRED_PWR_STATE, &state.desire_pwr_state, - DOMAIN_PGFSM_PWR_STATUS, &state.current_pwr_state); - break; - case 2: - /* mpc, opp, optc, dwb */ - REG_GET_2(DOMAIN24_PG_STATUS, - DOMAIN_DESIRED_PWR_STATE, &state.desire_pwr_state, - DOMAIN_PGFSM_PWR_STATUS, &state.current_pwr_state); - break; - case 3: - /* hpo */ - REG_GET_2(DOMAIN25_PG_STATUS, - DOMAIN_DESIRED_PWR_STATE, &state.desire_pwr_state, - DOMAIN_PGFSM_PWR_STATUS, &state.current_pwr_state); - break; - case 4: - /* dchubp0, dpp0 */ - REG_GET_2(DOMAIN0_PG_STATUS, - DOMAIN_DESIRED_PWR_STATE, &state.desire_pwr_state, - DOMAIN_PGFSM_PWR_STATUS, &state.current_pwr_state); - break; - case 5: - /* dsc0 */ - REG_GET_2(DOMAIN16_PG_STATUS, - DOMAIN_DESIRED_PWR_STATE, &state.desire_pwr_state, - DOMAIN_PGFSM_PWR_STATUS, &state.current_pwr_state); - break; - case 6: - /* dchubp1, dpp1 */ - REG_GET_2(DOMAIN1_PG_STATUS, - DOMAIN_DESIRED_PWR_STATE, &state.desire_pwr_state, - DOMAIN_PGFSM_PWR_STATUS, &state.current_pwr_state); - break; - case 7: - /* dsc1 */ - REG_GET_2(DOMAIN17_PG_STATUS, - DOMAIN_DESIRED_PWR_STATE, &state.desire_pwr_state, - DOMAIN_PGFSM_PWR_STATUS, &state.current_pwr_state); - break; - case 8: - /* dchubp2, dpp2 */ - REG_GET_2(DOMAIN2_PG_STATUS, - DOMAIN_DESIRED_PWR_STATE, &state.desire_pwr_state, - DOMAIN_PGFSM_PWR_STATUS, &state.current_pwr_state); - break; - case 9: - /* dsc2 */ - REG_GET_2(DOMAIN18_PG_STATUS, - DOMAIN_DESIRED_PWR_STATE, &state.desire_pwr_state, - DOMAIN_PGFSM_PWR_STATUS, &state.current_pwr_state); - break; - case 10: - /* dchubp3, dpp3 */ - REG_GET_2(DOMAIN3_PG_STATUS, - DOMAIN_DESIRED_PWR_STATE, &state.desire_pwr_state, - DOMAIN_PGFSM_PWR_STATUS, &state.current_pwr_state); - break; - case 11: - /* dsc3 */ - REG_GET_2(DOMAIN19_PG_STATUS, - DOMAIN_DESIRED_PWR_STATE, &state.desire_pwr_state, - DOMAIN_PGFSM_PWR_STATUS, &state.current_pwr_state); - break; - default: - break; - } - - return state; -} - void dcn401_init_hw(struct dc *dc) { struct abm **abms = dc->res_pool->multiple_abms; diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.h b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.h index 17a06f299b68..17cea748789e 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.h +++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.h @@ -63,8 +63,6 @@ void dcn401_set_cursor_position(struct pipe_ctx *pipe_ctx); bool dcn401_apply_idle_power_optimizations(struct dc *dc, bool enable); -struct ips_ono_region_state dcn401_read_ono_state(struct dc *dc, - uint8_t region); void dcn401_wait_for_dcc_meta_propagation(const struct dc *dc, const struct pipe_ctx *top_pipe_to_program); -- 2.51.0 From 4e5a9bcc9b6aa3c11c7649c8e9727d662f069af8 Mon Sep 17 00:00:00 2001 From: Robin Chen Date: Wed, 25 Dec 2024 13:29:31 +0800 Subject: [PATCH 10/16] drm/amd/display: Add a new flag for replay low hz [Why & How] Add a new flag in replay_config to indicate the replay low hz status. Reviewed-by: Allen Li Signed-off-by: Robin Chen Signed-off-by: Tom Chung Tested-by: Daniel Wheeler Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/display/dc/dc_types.h | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/gpu/drm/amd/display/dc/dc_types.h b/drivers/gpu/drm/amd/display/dc/dc_types.h index 0aef2d84c3a4..d7242dd83def 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_types.h +++ b/drivers/gpu/drm/amd/display/dc/dc_types.h @@ -1119,6 +1119,8 @@ struct replay_config { union replay_error_status replay_error_status; /* Replay Low Hz enable Options */ union replay_low_refresh_rate_enable_options low_rr_enable_options; + /* Replay coasting vtotal is within low refresh rate range. */ + bool low_rr_activated; }; /* Replay feature flags*/ -- 2.51.0 From 0524dd3a4f0e462067a3fcd13b7c5086f9dc1ece Mon Sep 17 00:00:00 2001 From: Dennis Chan Date: Mon, 23 Sep 2024 10:12:05 +0800 Subject: [PATCH 11/16] drm/amd/display: Revised for Replay Pseudo vblank control [why & how] Revised Replay Full screen video Pseudo vblank control. Reviewed-by: Allen Li Signed-off-by: Dennis Chan Signed-off-by: Tom Chung Tested-by: Daniel Wheeler Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/display/dc/dc_types.h | 4 ++-- drivers/gpu/drm/amd/display/modules/power/power_helpers.c | 4 ++-- drivers/gpu/drm/amd/display/modules/power/power_helpers.h | 2 +- 3 files changed, 5 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/amd/display/dc/dc_types.h b/drivers/gpu/drm/amd/display/dc/dc_types.h index d7242dd83def..0c2aa91f0a11 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_types.h +++ b/drivers/gpu/drm/amd/display/dc/dc_types.h @@ -1145,8 +1145,8 @@ struct replay_settings { uint32_t defer_update_coasting_vtotal_table[PR_COASTING_TYPE_NUM]; /* Maximum link off frame count */ uint32_t link_off_frame_count; - /* Replay pseudo vtotal for abm + ips on full screen video which can improve ips residency */ - uint16_t abm_with_ips_on_full_screen_video_pseudo_vtotal; + /* Replay pseudo vtotal for low refresh rate*/ + uint16_t low_rr_full_screen_video_pseudo_vtotal; /* Replay last pseudo vtotal set to DMUB */ uint16_t last_pseudo_vtotal; /* Replay desync error */ diff --git a/drivers/gpu/drm/amd/display/modules/power/power_helpers.c b/drivers/gpu/drm/amd/display/modules/power/power_helpers.c index f4b86a63f25d..29ccd3532d13 100644 --- a/drivers/gpu/drm/amd/display/modules/power/power_helpers.c +++ b/drivers/gpu/drm/amd/display/modules/power/power_helpers.c @@ -996,9 +996,9 @@ void set_replay_coasting_vtotal(struct dc_link *link, link->replay_settings.coasting_vtotal_table[type] = vtotal; } -void set_replay_ips_full_screen_video_src_vtotal(struct dc_link *link, uint16_t vtotal) +void set_replay_low_rr_full_screen_video_src_vtotal(struct dc_link *link, uint16_t vtotal) { - link->replay_settings.abm_with_ips_on_full_screen_video_pseudo_vtotal = vtotal; + link->replay_settings.low_rr_full_screen_video_pseudo_vtotal = vtotal; } void calculate_replay_link_off_frame_count(struct dc_link *link, diff --git a/drivers/gpu/drm/amd/display/modules/power/power_helpers.h b/drivers/gpu/drm/amd/display/modules/power/power_helpers.h index 94ade0ef18a3..758a8aa31fbe 100644 --- a/drivers/gpu/drm/amd/display/modules/power/power_helpers.h +++ b/drivers/gpu/drm/amd/display/modules/power/power_helpers.h @@ -62,7 +62,7 @@ void set_replay_defer_update_coasting_vtotal(struct dc_link *link, uint32_t vtotal); void update_replay_coasting_vtotal_from_defer(struct dc_link *link, enum replay_coasting_vtotal_type type); -void set_replay_ips_full_screen_video_src_vtotal(struct dc_link *link, uint16_t vtotal); +void set_replay_low_rr_full_screen_video_src_vtotal(struct dc_link *link, uint16_t vtotal); void calculate_replay_link_off_frame_count(struct dc_link *link, uint16_t vtotal, uint16_t htotal); -- 2.51.0 From a2b5a9956269f4c1a09537177f18ab0229fe79f7 Mon Sep 17 00:00:00 2001 From: Tom Chung Date: Tue, 1 Oct 2024 17:13:07 +0800 Subject: [PATCH 12/16] drm/amd/display: Use HW lock mgr for PSR1 [Why] Without the dmub hw lock, it may cause the lock timeout issue while do modeset on PSR1 eDP panel. [How] Allow dmub hw lock for PSR1. Reviewed-by: Sun peng Li Signed-off-by: Tom Chung Tested-by: Daniel Wheeler Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/display/dc/dce/dmub_hw_lock_mgr.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/display/dc/dce/dmub_hw_lock_mgr.c b/drivers/gpu/drm/amd/display/dc/dce/dmub_hw_lock_mgr.c index bf636b28e3e1..5bb8b78bf250 100644 --- a/drivers/gpu/drm/amd/display/dc/dce/dmub_hw_lock_mgr.c +++ b/drivers/gpu/drm/amd/display/dc/dce/dmub_hw_lock_mgr.c @@ -63,7 +63,8 @@ void dmub_hw_lock_mgr_inbox0_cmd(struct dc_dmub_srv *dmub_srv, bool should_use_dmub_lock(struct dc_link *link) { - if (link->psr_settings.psr_version == DC_PSR_VERSION_SU_1) + if (link->psr_settings.psr_version == DC_PSR_VERSION_SU_1 || + link->psr_settings.psr_version == DC_PSR_VERSION_1) return true; if (link->replay_settings.replay_feature_enabled) -- 2.51.0 From ec6d8d49f44b4d0cbf8674f81374ea4df04228da Mon Sep 17 00:00:00 2001 From: Austin Zheng Date: Fri, 20 Dec 2024 12:35:20 -0500 Subject: [PATCH 13/16] drm/amd/display: Apply DML21 Patches [Why & How] Add several DML21 fixes Reviewed-by: Wenjing Liu Signed-off-by: Austin Zheng Signed-off-by: Tom Chung Tested-by: Daniel Wheeler Signed-off-by: Alex Deucher --- .../src/dml2_core/dml2_core_dcn4_calcs.c | 107 ++++++++++++++++-- .../src/dml2_core/dml2_core_shared_types.h | 6 +- .../dml21/src/dml2_core/dml2_core_utils.c | 2 +- .../dml2/dml21/src/dml2_pmo/dml2_pmo_dcn3.c | 6 +- .../dml21/src/dml2_pmo/dml2_pmo_dcn4_fams2.c | 10 +- .../dml21/src/dml2_top/dml2_top_interfaces.c | 1 - .../dml2/dml21/src/dml2_top/dml2_top_legacy.c | 6 + .../dml2/dml21/src/dml2_top/dml2_top_soc15.c | 1 + .../dml2/dml21/src/dml2_top/dml2_top_soc15.h | 1 + .../src/inc/dml2_internal_shared_types.h | 2 - 10 files changed, 124 insertions(+), 18 deletions(-) diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c index a3ffb8c95d25..c4dbf27abaf8 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c @@ -3528,10 +3528,9 @@ static void CalculateUrgentBurstFactor( dml2_printf("DML::%s: UrgentBurstFactorChroma = %f\n", __func__, *UrgentBurstFactorChroma); dml2_printf("DML::%s: NotEnoughUrgentLatencyHiding = %d\n", __func__, *NotEnoughUrgentLatencyHiding); #endif - } -static void CalculateDCFCLKDeepSleep( +static void CalculateDCFCLKDeepSleepTdlut( const struct dml2_display_cfg *display_cfg, unsigned int NumberOfActiveSurfaces, unsigned int BytePerPixelY[], @@ -3546,6 +3545,10 @@ static void CalculateDCFCLKDeepSleep( double ReadBandwidthChroma[], unsigned int ReturnBusWidth, + double dispclk, + unsigned int tdlut_bytes_to_deliver[], + double prefetch_swath_time_us[], + // Output double *DCFClkDeepSleep) { @@ -3580,6 +3583,22 @@ static void CalculateDCFCLKDeepSleep( } DCFClkDeepSleepPerSurface[k] = math_max2(DCFClkDeepSleepPerSurface[k], pixel_rate_mhz / 16); + // adjust for 3dlut delivery time + if (display_cfg->plane_descriptors[k].tdlut.setup_for_tdlut && tdlut_bytes_to_deliver[k] > 0) { + double tdlut_required_deepsleep_dcfclk = (double) tdlut_bytes_to_deliver[k] / 64.0 / prefetch_swath_time_us[k]; + + dml2_printf("DML::%s: k=%d, DCFClkDeepSleepPerSurface = %f\n", __func__, k, DCFClkDeepSleepPerSurface[k]); + dml2_printf("DML::%s: k=%d, tdlut_bytes_to_deliver = %d\n", __func__, k, tdlut_bytes_to_deliver[k]); + dml2_printf("DML::%s: k=%d, prefetch_swath_time_us = %f\n", __func__, k, prefetch_swath_time_us[k]); + dml2_printf("DML::%s: k=%d, tdlut_required_deepsleep_dcfclk = %f\n", __func__, k, tdlut_required_deepsleep_dcfclk); + + // increase the deepsleep dcfclk to match the original dispclk throughput rate + if (tdlut_required_deepsleep_dcfclk > DCFClkDeepSleepPerSurface[k]) { + DCFClkDeepSleepPerSurface[k] = math_max2(DCFClkDeepSleepPerSurface[k], tdlut_required_deepsleep_dcfclk); + DCFClkDeepSleepPerSurface[k] = math_max2(DCFClkDeepSleepPerSurface[k], dispclk / 4.0); + } + } + #ifdef __DML_VBA_DEBUG__ dml2_printf("DML::%s: k=%u, PixelClock = %f\n", __func__, k, pixel_rate_mhz); dml2_printf("DML::%s: k=%u, DCFClkDeepSleepPerSurface = %f\n", __func__, k, DCFClkDeepSleepPerSurface[k]); @@ -3602,9 +3621,56 @@ static void CalculateDCFCLKDeepSleep( for (unsigned int k = 0; k < NumberOfActiveSurfaces; ++k) { *DCFClkDeepSleep = math_max2(*DCFClkDeepSleep, DCFClkDeepSleepPerSurface[k]); } + dml2_printf("DML::%s: DCFClkDeepSleep = %f (final)\n", __func__, *DCFClkDeepSleep); } +static void CalculateDCFCLKDeepSleep( + const struct dml2_display_cfg *display_cfg, + unsigned int NumberOfActiveSurfaces, + unsigned int BytePerPixelY[], + unsigned int BytePerPixelC[], + unsigned int SwathWidthY[], + unsigned int SwathWidthC[], + unsigned int DPPPerSurface[], + double PSCL_THROUGHPUT[], + double PSCL_THROUGHPUT_CHROMA[], + double Dppclk[], + double ReadBandwidthLuma[], + double ReadBandwidthChroma[], + unsigned int ReturnBusWidth, + + // Output + double *DCFClkDeepSleep) +{ + double zero_double[DML2_MAX_PLANES]; + unsigned int zero_integer[DML2_MAX_PLANES]; + + memset(zero_double, 0, DML2_MAX_PLANES * sizeof(double)); + memset(zero_integer, 0, DML2_MAX_PLANES * sizeof(unsigned int)); + + CalculateDCFCLKDeepSleepTdlut( + display_cfg, + NumberOfActiveSurfaces, + BytePerPixelY, + BytePerPixelC, + SwathWidthY, + SwathWidthC, + DPPPerSurface, + PSCL_THROUGHPUT, + PSCL_THROUGHPUT_CHROMA, + Dppclk, + ReadBandwidthLuma, + ReadBandwidthChroma, + ReturnBusWidth, + 0, + zero_integer, //tdlut_bytes_to_deliver, + zero_double, //prefetch_swath_time_us, + + // Output + DCFClkDeepSleep); +} + static double CalculateWriteBackDelay( enum dml2_source_format_class WritebackPixelFormat, double WritebackHRatio, @@ -4604,6 +4670,7 @@ static void calculate_tdlut_setting( *p->tdlut_groups_per_2row_ub = 0; *p->tdlut_opt_time = 0; *p->tdlut_drain_time = 0; + *p->tdlut_bytes_to_deliver = 0; *p->tdlut_bytes_per_group = 0; *p->tdlut_pte_bytes_per_frame = 0; *p->tdlut_bytes_per_frame = 0; @@ -4672,6 +4739,7 @@ static void calculate_tdlut_setting( *p->tdlut_groups_per_2row_ub = (unsigned int)math_ceil2((double) *p->tdlut_bytes_per_frame / *p->tdlut_bytes_per_group, 1); *p->tdlut_opt_time = (*p->tdlut_bytes_per_frame - p->cursor_buffer_size * 1024) / tdlut_drain_rate; *p->tdlut_drain_time = p->cursor_buffer_size * 1024 / tdlut_drain_rate; + *p->tdlut_bytes_to_deliver = (unsigned int) (p->cursor_buffer_size * 1024.0); } #ifdef __DML_VBA_DEBUG__ @@ -4692,6 +4760,7 @@ static void calculate_tdlut_setting( dml2_printf("DML::%s: tdlut_delivery_cycles = %u\n", __func__, tdlut_delivery_cycles); dml2_printf("DML::%s: tdlut_opt_time = %f\n", __func__, *p->tdlut_opt_time); dml2_printf("DML::%s: tdlut_drain_time = %f\n", __func__, *p->tdlut_drain_time); + dml2_printf("DML::%s: tdlut_bytes_to_deliver = %d\n", __func__, *p->tdlut_bytes_to_deliver); dml2_printf("DML::%s: tdlut_groups_per_2row_ub = %d\n", __func__, *p->tdlut_groups_per_2row_ub); #endif } @@ -5700,6 +5769,7 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch s->cursor_prefetch_bytes = (unsigned int)math_max2(p->cursor_bytes_per_chunk, 4 * p->cursor_bytes_per_line); *p->prefetch_cursor_bw = p->num_cursors * s->cursor_prefetch_bytes / (s->LinesToRequestPrefetchPixelData * s->LineTime); + *p->prefetch_swath_time_us = (s->LinesToRequestPrefetchPixelData * s->LineTime); #ifdef __DML_VBA_DEBUG__ dml2_printf("DML::%s: TimeForFetchingVM = %f\n", __func__, s->TimeForFetchingVM); @@ -5710,6 +5780,7 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch dml2_printf("DML::%s: dst_y_per_row_vblank = %f\n", __func__, *p->dst_y_per_row_vblank); dml2_printf("DML::%s: LinesToRequestPrefetchPixelData = %f\n", __func__, s->LinesToRequestPrefetchPixelData); dml2_printf("DML::%s: PrefetchSourceLinesY = %f\n", __func__, p->PrefetchSourceLinesY); + dml2_printf("DML::%s: prefetch_swath_time_us = %f\n", __func__, *p->prefetch_swath_time_us); dml2_printf("DML::%s: cursor_bytes_per_chunk = %d\n", __func__, p->cursor_bytes_per_chunk); dml2_printf("DML::%s: cursor_bytes_per_line = %d\n", __func__, p->cursor_bytes_per_line); @@ -8817,6 +8888,7 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out calculate_tdlut_setting_params->tdlut_groups_per_2row_ub = &s->tdlut_groups_per_2row_ub[k]; calculate_tdlut_setting_params->tdlut_opt_time = &s->tdlut_opt_time[k]; calculate_tdlut_setting_params->tdlut_drain_time = &s->tdlut_drain_time[k]; + calculate_tdlut_setting_params->tdlut_bytes_to_deliver = &s->tdlut_bytes_to_deliver[k]; calculate_tdlut_setting_params->tdlut_bytes_per_group = &s->tdlut_bytes_per_group[k]; calculate_tdlut_setting(&mode_lib->scratch, calculate_tdlut_setting_params); @@ -9009,6 +9081,7 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out CalculatePrefetchSchedule_params->prefetch_sw_bytes = &s->prefetch_sw_bytes[k]; CalculatePrefetchSchedule_params->Tpre_rounded = &s->Tpre_rounded[k]; CalculatePrefetchSchedule_params->Tpre_oto = &s->Tpre_oto[k]; + CalculatePrefetchSchedule_params->prefetch_swath_time_us = &s->prefetch_swath_time_us[k]; mode_lib->ms.NoTimeForPrefetch[k] = CalculatePrefetchSchedule(&mode_lib->scratch, CalculatePrefetchSchedule_params); @@ -9017,6 +9090,27 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out dml2_printf("DML::%s: k=%d, dst_y_per_row_vblank = %f\n", __func__, k, *CalculatePrefetchSchedule_params->dst_y_per_row_vblank); } // for k num_planes + CalculateDCFCLKDeepSleepTdlut( + display_cfg, + mode_lib->ms.num_active_planes, + mode_lib->ms.BytePerPixelY, + mode_lib->ms.BytePerPixelC, + mode_lib->ms.SwathWidthY, + mode_lib->ms.SwathWidthC, + mode_lib->ms.NoOfDPP, + mode_lib->ms.PSCL_FACTOR, + mode_lib->ms.PSCL_FACTOR_CHROMA, + mode_lib->ms.RequiredDPPCLK, + mode_lib->ms.vactive_sw_bw_l, + mode_lib->ms.vactive_sw_bw_c, + mode_lib->soc.return_bus_width_bytes, + mode_lib->ms.RequiredDISPCLK, + s->tdlut_bytes_to_deliver, + s->prefetch_swath_time_us, + + /* Output */ + &mode_lib->ms.dcfclk_deepsleep); + for (k = 0; k < mode_lib->ms.num_active_planes; k++) { if (mode_lib->ms.dst_y_prefetch[k] < 2.0 || mode_lib->ms.LinesForVM[k] >= 32.0 @@ -10368,12 +10462,6 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex dml2_assert(s->SOCCLK > 0); #ifdef __DML_VBA_DEBUG__ - // dml2_printf_dml_display_cfg_timing(&display_cfg->timing, s->num_active_planes); - // dml2_printf_dml_display_cfg_plane(&display_cfg->plane, s->num_active_planes); - // dml2_printf_dml_display_cfg_surface(&display_cfg->surface, s->num_active_planes); - // dml2_printf_dml_display_cfg_output(&display_cfg->output, s->num_active_planes); - // dml2_printf_dml_display_cfg_hw_resource(&display_cfg->hw, s->num_active_planes); - dml2_printf("DML::%s: num_active_planes = %u\n", __func__, s->num_active_planes); dml2_printf("DML::%s: num_active_pipes = %u\n", __func__, mode_lib->mp.num_active_pipes); dml2_printf("DML::%s: Dcfclk = %f\n", __func__, mode_lib->mp.Dcfclk); @@ -10832,8 +10920,8 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex calculate_tdlut_setting_params->tdlut_groups_per_2row_ub = &s->tdlut_groups_per_2row_ub[k]; calculate_tdlut_setting_params->tdlut_opt_time = &s->tdlut_opt_time[k]; calculate_tdlut_setting_params->tdlut_drain_time = &s->tdlut_drain_time[k]; + calculate_tdlut_setting_params->tdlut_bytes_to_deliver = &s->tdlut_bytes_to_deliver[k]; calculate_tdlut_setting_params->tdlut_bytes_per_group = &s->tdlut_bytes_per_group[k]; - calculate_tdlut_setting(&mode_lib->scratch, calculate_tdlut_setting_params); } @@ -11219,6 +11307,7 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex CalculatePrefetchSchedule_params->prefetch_sw_bytes = &s->prefetch_sw_bytes[k]; CalculatePrefetchSchedule_params->Tpre_rounded = &s->Tpre_rounded[k]; CalculatePrefetchSchedule_params->Tpre_oto = &s->Tpre_oto[k]; + CalculatePrefetchSchedule_params->prefetch_swath_time_us = &s->dummy_single[0]; mode_lib->mp.NoTimeToPrefetch[k] = CalculatePrefetchSchedule(&mode_lib->scratch, CalculatePrefetchSchedule_params); diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_shared_types.h b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_shared_types.h index 4f54e54102ef..23c0fca5515f 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_shared_types.h +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_shared_types.h @@ -958,6 +958,7 @@ struct dml2_core_calcs_mode_support_locals { unsigned int tdlut_groups_per_2row_ub[DML2_MAX_PLANES]; double tdlut_opt_time[DML2_MAX_PLANES]; double tdlut_drain_time[DML2_MAX_PLANES]; + unsigned int tdlut_bytes_to_deliver[DML2_MAX_PLANES]; unsigned int tdlut_bytes_per_group[DML2_MAX_PLANES]; unsigned int cursor_bytes_per_chunk[DML2_MAX_PLANES]; @@ -979,6 +980,7 @@ struct dml2_core_calcs_mode_support_locals { enum dml2_source_format_class pixel_format[DML2_MAX_PLANES]; unsigned int lb_source_lines_l[DML2_MAX_PLANES]; unsigned int lb_source_lines_c[DML2_MAX_PLANES]; + double prefetch_swath_time_us[DML2_MAX_PLANES]; }; struct dml2_core_calcs_mode_programming_locals { @@ -1042,6 +1044,7 @@ struct dml2_core_calcs_mode_programming_locals { unsigned int tdlut_groups_per_2row_ub[DML2_MAX_PLANES]; double tdlut_opt_time[DML2_MAX_PLANES]; double tdlut_drain_time[DML2_MAX_PLANES]; + unsigned int tdlut_bytes_to_deliver[DML2_MAX_PLANES]; unsigned int tdlut_bytes_per_group[DML2_MAX_PLANES]; unsigned int cursor_bytes_per_chunk[DML2_MAX_PLANES]; @@ -1809,6 +1812,7 @@ struct dml2_core_calcs_CalculatePrefetchSchedule_params { unsigned int *VReadyOffsetPix; double *prefetch_cursor_bw; double *prefetch_sw_bytes; + double *prefetch_swath_time_us; }; struct dml2_core_calcs_CheckGlobalPrefetchAdmissibility_params { @@ -1993,6 +1997,7 @@ struct dml2_core_calcs_calculate_tdlut_setting_params { unsigned int *tdlut_groups_per_2row_ub; double *tdlut_opt_time; double *tdlut_drain_time; + unsigned int *tdlut_bytes_to_deliver; unsigned int *tdlut_bytes_per_group; }; @@ -2137,7 +2142,6 @@ struct dml2_core_calcs_mode_programming_ex { const struct core_display_cfg_support_info *cfg_support_info; int min_clk_index; struct dml2_display_cfg_programming *programming; - }; #endif diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_utils.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_utils.c index 1548dfc68b8e..456b3f8a6d38 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_utils.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_utils.c @@ -556,7 +556,7 @@ bool dml2_core_utils_is_dual_plane(enum dml2_source_format_class source_format) { bool ret_val = 0; - if ((source_format == dml2_420_12) || (source_format == dml2_420_8) || (source_format == dml2_420_10) || (source_format == dml2_rgbe_alpha)) + if (dml2_core_utils_is_420(source_format) || dml2_core_utils_is_422_planar(source_format) || (source_format == dml2_rgbe_alpha)) ret_val = 1; return ret_val; diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn3.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn3.c index 655d1ef56814..e763c8e45da8 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn3.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn3.c @@ -347,8 +347,12 @@ static int find_highest_odm_load_stream_index( int odm_load, highest_odm_load = -1, highest_odm_load_index = -1; for (i = 0; i < display_config->num_streams; i++) { - odm_load = display_config->stream_descriptors[i].timing.pixel_clock_khz + if (mode_support_result->cfg_support_info.stream_support_info[i].odms_used > 0) + odm_load = display_config->stream_descriptors[i].timing.pixel_clock_khz / mode_support_result->cfg_support_info.stream_support_info[i].odms_used; + else + odm_load = 0; + if (odm_load > highest_odm_load) { highest_odm_load_index = i; highest_odm_load = odm_load; diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn4_fams2.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn4_fams2.c index 94609f9c18eb..a3324f7b9ba6 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn4_fams2.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn4_fams2.c @@ -813,8 +813,12 @@ static int find_highest_odm_load_stream_index( int odm_load, highest_odm_load = -1, highest_odm_load_index = -1; for (i = 0; i < display_config->num_streams; i++) { - odm_load = display_config->stream_descriptors[i].timing.pixel_clock_khz + if (mode_support_result->cfg_support_info.stream_support_info[i].odms_used > 0) + odm_load = display_config->stream_descriptors[i].timing.pixel_clock_khz / mode_support_result->cfg_support_info.stream_support_info[i].odms_used; + else + odm_load = 0; + if (odm_load > highest_odm_load) { highest_odm_load_index = i; highest_odm_load = odm_load; @@ -1372,7 +1376,7 @@ static bool is_config_schedulable( if (j_disallow_us < jp1_disallow_us) { /* swap as A < B */ swap(s->pmo_dcn4.sorted_group_gtl_disallow_index[j], - s->pmo_dcn4.sorted_group_gtl_disallow_index[j+1]); + s->pmo_dcn4.sorted_group_gtl_disallow_index[j + 1]); swapped = true; } } @@ -1431,7 +1435,7 @@ static bool is_config_schedulable( if (j_period_us < jp1_period_us) { /* swap as A < B */ swap(s->pmo_dcn4.sorted_group_gtl_period_index[j], - s->pmo_dcn4.sorted_group_gtl_period_index[j+1]); + s->pmo_dcn4.sorted_group_gtl_period_index[j + 1]); swapped = true; } } diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_interfaces.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_interfaces.c index 5f6dfc24df69..f88931ccbc5e 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_interfaces.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_interfaces.c @@ -15,7 +15,6 @@ bool dml2_initialize_instance(struct dml2_initialize_instance_in_out *in_out) { switch (in_out->options.project_id) { case dml2_project_dcn4x_stage1: - return false; case dml2_project_dcn4x_stage2: case dml2_project_dcn4x_stage2_auto_drr_svp: return dml2_top_soc15_initialize_instance(in_out); diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_legacy.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_legacy.c index db0a30fdb58d..5e14d85821e2 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_legacy.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_legacy.c @@ -2,3 +2,9 @@ // // Copyright 2024 Advanced Micro Devices, Inc. +#include "dml2_top_legacy.h" +#include "dml2_top_soc15.h" +#include "dml2_core_factory.h" +#include "dml2_pmo_factory.h" +#include "display_mode_core_structs.h" + diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_soc15.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_soc15.c index b39029c0e56f..a8f58f8448e4 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_soc15.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_soc15.c @@ -545,6 +545,7 @@ bool dml2_top_mcache_validate_admissability(struct top_mcache_validate_admissabi if (odm_combine_factor > 1) { max_per_pipe_vp_p0 = plane->surface.plane0.width; temp = (unsigned int)math_ceil(plane->composition.scaler_info.plane0.h_ratio * stream->timing.h_active / odm_combine_factor); + if (temp < max_per_pipe_vp_p0) max_per_pipe_vp_p0 = temp; diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_soc15.h b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_soc15.h index 6fda201af898..53bd8602f9ef 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_soc15.h +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_soc15.h @@ -1,6 +1,7 @@ // SPDX-License-Identifier: MIT // // Copyright 2024 Advanced Micro Devices, Inc. + #ifndef __DML2_TOP_SOC15_H__ #define __DML2_TOP_SOC15_H__ #include "dml2_internal_shared_types.h" diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/inc/dml2_internal_shared_types.h b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/inc/dml2_internal_shared_types.h index d94b310d6eec..7fb6026bcb49 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/inc/dml2_internal_shared_types.h +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/inc/dml2_internal_shared_types.h @@ -357,8 +357,6 @@ struct dml2_pmo_pstate_strategy { enum dml2_pstate_method per_stream_pstate_method[DML2_MAX_PLANES]; bool allow_state_increase; }; - - struct dml2_core_mode_support_in_out { /* * Inputs -- 2.51.0 From 230dced3e2b712017c03ce0afb3f8c48d8af20ee Mon Sep 17 00:00:00 2001 From: Peichen Huang Date: Wed, 25 Dec 2024 15:51:34 +0800 Subject: [PATCH 14/16] drm/amd/display: improve dpia pre-train [WHY] We see unstable DP LL 4.2.1.3 test result with dpia pre-train. It is because the outbox interrupt mechanism can not handle HPD immediately and require some improvement. [HOW] 1. not enable link if hpd_pending is true. 2. abort pre-train if training failed and hpd_pending is true. 3. check if 2 lane supported when it is alt mode Reviewed-by: Wenjing Liu Reviewed-by: Meenakshikumar Somasundaram Signed-off-by: Peichen Huang Signed-off-by: Tom Chung Tested-by: Daniel Wheeler Signed-off-by: Alex Deucher --- .../amd/display/dc/dio/dcn31/dcn31_dio_link_encoder.c | 5 +++-- drivers/gpu/drm/amd/display/dc/link/link_dpms.c | 9 +++++++++ .../amd/display/dc/link/protocols/link_dp_capability.c | 8 ++++++++ 3 files changed, 20 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/amd/display/dc/dio/dcn31/dcn31_dio_link_encoder.c b/drivers/gpu/drm/amd/display/dc/dio/dcn31/dcn31_dio_link_encoder.c index b2cea59ba5d4..9a92f73d5b7f 100644 --- a/drivers/gpu/drm/amd/display/dc/dio/dcn31/dcn31_dio_link_encoder.c +++ b/drivers/gpu/drm/amd/display/dc/dio/dcn31/dcn31_dio_link_encoder.c @@ -653,8 +653,9 @@ void dcn31_link_encoder_get_max_link_cap(struct link_encoder *enc, struct dc_lin if (!query_dp_alt_from_dmub(enc, &cmd)) return; - if (cmd.query_dp_alt.data.is_usb && - cmd.query_dp_alt.data.is_dp4 == 0) + if (cmd.query_dp_alt.data.is_dp_alt_disable == 0 && + cmd.query_dp_alt.data.is_usb && + cmd.query_dp_alt.data.is_dp4 == 0) link_settings->lane_count = MIN(LANE_COUNT_TWO, link_settings->lane_count); return; diff --git a/drivers/gpu/drm/amd/display/dc/link/link_dpms.c b/drivers/gpu/drm/amd/display/dc/link/link_dpms.c index b8cfeb98e229..ec7de9c01fab 100644 --- a/drivers/gpu/drm/amd/display/dc/link/link_dpms.c +++ b/drivers/gpu/drm/amd/display/dc/link/link_dpms.c @@ -2548,6 +2548,15 @@ void link_set_dpms_on( if (pipe_ctx->stream->dpms_off) return; + /* For Dp tunneling link, a pending HPD means that we have a race condition between processing + * current link and processing the pending HPD. If we enable the link now, we may end up with a + * link that is not actually connected to a sink. So we skip enabling the link in this case. + */ + if (link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA && link->is_hpd_pending) { + DC_LOG_DEBUG("%s, Link%d HPD is pending, not enable it.\n", __func__, link->link_index); + return; + } + /* Have to setup DSC before DIG FE and BE are connected (which happens before the * link training). This is to make sure the bandwidth sent to DIG BE won't be * bigger than what the link and/or DIG BE can handle. VBID[6]/CompressedStream_flag diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c index d0fbf9c44a29..44c3023a7731 100644 --- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c +++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c @@ -2312,6 +2312,14 @@ bool dp_verify_link_cap_with_retries( } else { link->verified_link_cap = last_verified_link_cap; } + + /* For Dp tunneling link, a pending HPD means that we have a race condition between processing + * current link and processing the pending HPD. Since the training is failed, we should just brak + * the loop so that we have chance to process the pending HPD. + */ + if (link->ep_type == DISPLAY_ENDPOINT_USB4_DPIA && link->is_hpd_pending) + break; + fsleep(10 * 1000); } -- 2.51.0 From 0ae47e971b9add8f7b8f8d55ac5f407f6f346758 Mon Sep 17 00:00:00 2001 From: Charlene Liu Date: Tue, 26 Nov 2024 12:09:25 -0500 Subject: [PATCH 15/16] drm/amd/display: avoid reset DTBCLK at clock init [why & how] this is to init to HW real DTBCLK. and use real HW DTBCLK status to update internal logic state Reviewed-by: Nicholas Kazlauskas Reviewed-by: Martin Leung Signed-off-by: Charlene Liu Signed-off-by: Ausef Yousof Signed-off-by: Tom Chung Tested-by: Daniel Wheeler Signed-off-by: Alex Deucher --- .../display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c | 18 ++++++++++++------ 1 file changed, 12 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c index 2a74140d7ebf..1f974ea3b0c6 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c @@ -401,6 +401,7 @@ void dcn35_update_clocks(struct clk_mgr *clk_mgr_base, if (clk_mgr_base->clks.dtbclk_en && !new_clocks->dtbclk_en) { if (clk_mgr->base.ctx->dc->config.allow_0_dtb_clk) dcn35_smu_set_dtbclk(clk_mgr, false); + clk_mgr_base->clks.dtbclk_en = new_clocks->dtbclk_en; } /* check that we're not already in lower */ @@ -418,11 +419,17 @@ void dcn35_update_clocks(struct clk_mgr *clk_mgr_base, } if (!clk_mgr_base->clks.dtbclk_en && new_clocks->dtbclk_en) { - dcn35_smu_set_dtbclk(clk_mgr, true); - clk_mgr_base->clks.dtbclk_en = new_clocks->dtbclk_en; + int actual_dtbclk = 0; dcn35_update_clocks_update_dtb_dto(clk_mgr, context, new_clocks->ref_dtbclk_khz); - clk_mgr_base->clks.ref_dtbclk_khz = new_clocks->ref_dtbclk_khz; + dcn35_smu_set_dtbclk(clk_mgr, true); + + actual_dtbclk = REG_READ(CLK1_CLK4_CURRENT_CNT); + + if (actual_dtbclk) { + clk_mgr_base->clks.ref_dtbclk_khz = new_clocks->ref_dtbclk_khz; + clk_mgr_base->clks.dtbclk_en = new_clocks->dtbclk_en; + } } /* check that we're not already in D0 */ @@ -584,12 +591,10 @@ static bool dcn35_is_spll_ssc_enabled(struct clk_mgr *clk_mgr_base) static void init_clk_states(struct clk_mgr *clk_mgr) { - struct clk_mgr_internal *clk_mgr_int = TO_CLK_MGR_INTERNAL(clk_mgr); uint32_t ref_dtbclk = clk_mgr->clks.ref_dtbclk_khz; + memset(&(clk_mgr->clks), 0, sizeof(struct dc_clocks)); - if (clk_mgr_int->smu_ver >= SMU_VER_THRESHOLD) - clk_mgr->clks.dtbclk_en = true; // request DTBCLK disable on first commit clk_mgr->clks.ref_dtbclk_khz = ref_dtbclk; // restore ref_dtbclk clk_mgr->clks.p_state_change_support = true; clk_mgr->clks.prev_p_state_change_support = true; @@ -600,6 +605,7 @@ static void init_clk_states(struct clk_mgr *clk_mgr) void dcn35_init_clocks(struct clk_mgr *clk_mgr) { struct clk_mgr_internal *clk_mgr_int = TO_CLK_MGR_INTERNAL(clk_mgr); + init_clk_states(clk_mgr); // to adjust dp_dto reference clock if ssc is enable otherwise to apply dprefclk -- 2.51.0 From 812a33a65d00e3d813f5ed2c9923569acd0b445c Mon Sep 17 00:00:00 2001 From: Ryan Seto Date: Mon, 6 Jan 2025 11:44:48 -0500 Subject: [PATCH 16/16] drm/amd/display: 3.2.316 This version brings along following fixes: - Add some feature for secure display - Add replay desync error count tracking and reset - Update chip_cap defines and usage - Remove unnecessary eDP power down - Fix some stuttering/corruption issue on PSR panel - Cleanup and refactoring DML2.1 Acked-by: Wayne Lin Reviewed-by: Martin Leung Signed-off-by: Ryan Seto Signed-off-by: Tom Chung Tested-by: Daniel Wheeler Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/display/dc/dc.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/display/dc/dc.h b/drivers/gpu/drm/amd/display/dc/dc.h index 5fb15c6315a4..053481ab69ef 100644 --- a/drivers/gpu/drm/amd/display/dc/dc.h +++ b/drivers/gpu/drm/amd/display/dc/dc.h @@ -55,7 +55,7 @@ struct aux_payload; struct set_config_cmd_payload; struct dmub_notification; -#define DC_VER "3.2.315" +#define DC_VER "3.2.316" #define MAX_SURFACES 4 #define MAX_PLANES 6 -- 2.51.0