Some display fixes for vega20 for stable. Fixes stability issues with certain combinations of monitors. Cherry-picked from master.
Alex Deucher (1): drm/amdgpu/display: fix 64 bit divide
Charlene Liu (1): drm/amd/display: dce11.x /dce12 update formula input
Zhan Liu (1): drm/amd/display: Add missing HBM support and raise Vega20's uclk.
.../dc/clk_mgr/dce110/dce110_clk_mgr.c | 27 ++++++++++++++++--- .../drm/amd/display/dc/dce/dce_mem_input.c | 4 +-- .../amd/display/dc/dce112/dce112_resource.c | 16 ++++++----- .../amd/display/dc/dce120/dce120_resource.c | 11 +++++--- drivers/gpu/drm/amd/display/dc/inc/resource.h | 2 ++ 5 files changed, 45 insertions(+), 15 deletions(-)
From: Charlene Liu charlene.liu@amd.com
[Description] 1. OUTSTANDING_REQUEST_LIMIT update from 0xFF to 0x1F (HW doc update) 2. using memory type to convert UMC's MCLK to Yclk.
Signed-off-by: Charlene Liu charlene.liu@amd.com Reviewed-by: Dmytro Laktyushkin Dmytro.Laktyushkin@amd.com Acked-by: Bhawanpreet Lakha Bhawanpreet.Lakha@amd.com Signed-off-by: Alex Deucher alexander.deucher@amd.com (cherry picked from commit c46e5df4ac898108da66a880c4e18f69c74f6c1b) cc: stable@vger.kernel.org # 5.3.x --- .../display/dc/clk_mgr/dce110/dce110_clk_mgr.c | 7 +++++-- .../gpu/drm/amd/display/dc/dce/dce_mem_input.c | 4 ++-- .../drm/amd/display/dc/dce112/dce112_resource.c | 16 ++++++++++------ .../drm/amd/display/dc/dce120/dce120_resource.c | 11 ++++++++--- drivers/gpu/drm/amd/display/dc/inc/resource.h | 2 ++ 5 files changed, 27 insertions(+), 13 deletions(-)
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dce110/dce110_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dce110/dce110_clk_mgr.c index 5cc3acccda2a..ee32d2c19305 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dce110/dce110_clk_mgr.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dce110/dce110_clk_mgr.c @@ -98,11 +98,14 @@ uint32_t dce110_get_min_vblank_time_us(const struct dc_state *context) struct dc_stream_state *stream = context->streams[j]; uint32_t vertical_blank_in_pixels = 0; uint32_t vertical_blank_time = 0; + uint32_t vertical_total_min = stream->timing.v_total; + struct dc_crtc_timing_adjust adjust = stream->adjust; + if (adjust.v_total_max != adjust.v_total_min) + vertical_total_min = adjust.v_total_min;
vertical_blank_in_pixels = stream->timing.h_total * - (stream->timing.v_total + (vertical_total_min - stream->timing.v_addressable); - vertical_blank_time = vertical_blank_in_pixels * 10000 / stream->timing.pix_clk_100hz;
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_mem_input.c b/drivers/gpu/drm/amd/display/dc/dce/dce_mem_input.c index a24a2bda8656..1596ddcb26e6 100644 --- a/drivers/gpu/drm/amd/display/dc/dce/dce_mem_input.c +++ b/drivers/gpu/drm/amd/display/dc/dce/dce_mem_input.c @@ -148,7 +148,7 @@ static void dce_mi_program_pte_vm( pte->min_pte_before_flip_horiz_scan;
REG_UPDATE(GRPH_PIPE_OUTSTANDING_REQUEST_LIMIT, - GRPH_PIPE_OUTSTANDING_REQUEST_LIMIT, 0xff); + GRPH_PIPE_OUTSTANDING_REQUEST_LIMIT, 0x7f);
REG_UPDATE_3(DVMM_PTE_CONTROL, DVMM_PAGE_WIDTH, page_width, @@ -157,7 +157,7 @@ static void dce_mi_program_pte_vm(
REG_UPDATE_2(DVMM_PTE_ARB_CONTROL, DVMM_PTE_REQ_PER_CHUNK, pte->pte_req_per_chunk, - DVMM_MAX_PTE_REQ_OUTSTANDING, 0xff); + DVMM_MAX_PTE_REQ_OUTSTANDING, 0x7f); }
static void program_urgency_watermark( diff --git a/drivers/gpu/drm/amd/display/dc/dce112/dce112_resource.c b/drivers/gpu/drm/amd/display/dc/dce112/dce112_resource.c index c6136e0ed1a4..7a04be74c9cf 100644 --- a/drivers/gpu/drm/amd/display/dc/dce112/dce112_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dce112/dce112_resource.c @@ -987,6 +987,10 @@ static void bw_calcs_data_update_from_pplib(struct dc *dc) struct dm_pp_clock_levels_with_latency mem_clks = {0}; struct dm_pp_wm_sets_with_clock_ranges clk_ranges = {0}; struct dm_pp_clock_levels clks = {0}; + int memory_type_multiplier = MEMORY_TYPE_MULTIPLIER_CZ; + + if (dc->bw_vbios && dc->bw_vbios->memory_type == bw_def_hbm) + memory_type_multiplier = MEMORY_TYPE_HBM;
/*do system clock TODO PPLIB: after PPLIB implement, * then remove old way @@ -1026,12 +1030,12 @@ static void bw_calcs_data_update_from_pplib(struct dc *dc) &clks);
dc->bw_vbios->low_yclk = bw_frc_to_fixed( - clks.clocks_in_khz[0] * MEMORY_TYPE_MULTIPLIER_CZ, 1000); + clks.clocks_in_khz[0] * memory_type_multiplier, 1000); dc->bw_vbios->mid_yclk = bw_frc_to_fixed( - clks.clocks_in_khz[clks.num_levels>>1] * MEMORY_TYPE_MULTIPLIER_CZ, + clks.clocks_in_khz[clks.num_levels>>1] * memory_type_multiplier, 1000); dc->bw_vbios->high_yclk = bw_frc_to_fixed( - clks.clocks_in_khz[clks.num_levels-1] * MEMORY_TYPE_MULTIPLIER_CZ, + clks.clocks_in_khz[clks.num_levels-1] * memory_type_multiplier, 1000);
return; @@ -1067,12 +1071,12 @@ static void bw_calcs_data_update_from_pplib(struct dc *dc) * YCLK = UMACLK*m_memoryTypeMultiplier */ dc->bw_vbios->low_yclk = bw_frc_to_fixed( - mem_clks.data[0].clocks_in_khz * MEMORY_TYPE_MULTIPLIER_CZ, 1000); + mem_clks.data[0].clocks_in_khz * memory_type_multiplier, 1000); dc->bw_vbios->mid_yclk = bw_frc_to_fixed( - mem_clks.data[mem_clks.num_levels>>1].clocks_in_khz * MEMORY_TYPE_MULTIPLIER_CZ, + mem_clks.data[mem_clks.num_levels>>1].clocks_in_khz * memory_type_multiplier, 1000); dc->bw_vbios->high_yclk = bw_frc_to_fixed( - mem_clks.data[mem_clks.num_levels-1].clocks_in_khz * MEMORY_TYPE_MULTIPLIER_CZ, + mem_clks.data[mem_clks.num_levels-1].clocks_in_khz * memory_type_multiplier, 1000);
/* Now notify PPLib/SMU about which Watermarks sets they should select diff --git a/drivers/gpu/drm/amd/display/dc/dce120/dce120_resource.c b/drivers/gpu/drm/amd/display/dc/dce120/dce120_resource.c index 4a6ba3173a5a..ae38c9c7277c 100644 --- a/drivers/gpu/drm/amd/display/dc/dce120/dce120_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dce120/dce120_resource.c @@ -847,6 +847,8 @@ static void bw_calcs_data_update_from_pplib(struct dc *dc) int i; unsigned int clk; unsigned int latency; + /*original logic in dal3*/ + int memory_type_multiplier = MEMORY_TYPE_MULTIPLIER_CZ;
/*do system clock*/ if (!dm_pp_get_clock_levels_by_type_with_latency( @@ -905,13 +907,16 @@ static void bw_calcs_data_update_from_pplib(struct dc *dc) * ALSO always convert UMA clock (from PPLIB) to YCLK (HW formula): * YCLK = UMACLK*m_memoryTypeMultiplier */ + if (dc->bw_vbios->memory_type == bw_def_hbm) + memory_type_multiplier = MEMORY_TYPE_HBM; + dc->bw_vbios->low_yclk = bw_frc_to_fixed( - mem_clks.data[0].clocks_in_khz * MEMORY_TYPE_MULTIPLIER_CZ, 1000); + mem_clks.data[0].clocks_in_khz * memory_type_multiplier, 1000); dc->bw_vbios->mid_yclk = bw_frc_to_fixed( - mem_clks.data[mem_clks.num_levels>>1].clocks_in_khz * MEMORY_TYPE_MULTIPLIER_CZ, + mem_clks.data[mem_clks.num_levels>>1].clocks_in_khz * memory_type_multiplier, 1000); dc->bw_vbios->high_yclk = bw_frc_to_fixed( - mem_clks.data[mem_clks.num_levels-1].clocks_in_khz * MEMORY_TYPE_MULTIPLIER_CZ, + mem_clks.data[mem_clks.num_levels-1].clocks_in_khz * memory_type_multiplier, 1000);
/* Now notify PPLib/SMU about which Watermarks sets they should select diff --git a/drivers/gpu/drm/amd/display/dc/inc/resource.h b/drivers/gpu/drm/amd/display/dc/inc/resource.h index 47f81072d7e9..c0424b4035a5 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/resource.h +++ b/drivers/gpu/drm/amd/display/dc/inc/resource.h @@ -31,6 +31,8 @@ #include "dm_pp_smu.h"
#define MEMORY_TYPE_MULTIPLIER_CZ 4 +#define MEMORY_TYPE_HBM 2 +
enum dce_version resource_parse_asic_id( struct hw_asic_id asic_id);
From: Zhan Liu zhan.liu@amd.com
[Why] When more than 2 displays are connected to the graphics card, only the minimum memory clock is needed. However, when more displays are connected, the minimum memory clock is not sufficient enough to support the overwhelming bandwidth. System will hang under this circumstance.
Also, the old code didn't address HBM cards, which has 2 pseudo channels. We need to add the HBM part here.
[How] When graphics card connects to 2 or more displays, switch to high memory clock. Also, choose memory multiplier based on whether its regular DRAM or HBM.
Signed-off-by: Zhan Liu zhan.liu@amd.com Reviewed-by: Roman Li Roman.Li@amd.com Acked-by: Leo Li sunpeng.li@amd.com Signed-off-by: Alex Deucher alexander.deucher@amd.com (cherry picked from commit c02d6a161395dfc0c2fdabb9e976a229017288d8) cc: stable@vger.kernel.org # 5.3.x --- .../display/dc/clk_mgr/dce110/dce110_clk_mgr.c | 18 ++++++++++++++++-- 1 file changed, 16 insertions(+), 2 deletions(-)
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dce110/dce110_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dce110/dce110_clk_mgr.c index ee32d2c19305..36277bca0326 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dce110/dce110_clk_mgr.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dce110/dce110_clk_mgr.c @@ -174,6 +174,10 @@ void dce11_pplib_apply_display_requirements( struct dc_state *context) { struct dm_pp_display_configuration *pp_display_cfg = &context->pp_display_cfg; + int memory_type_multiplier = MEMORY_TYPE_MULTIPLIER_CZ; + + if (dc->bw_vbios && dc->bw_vbios->memory_type == bw_def_hbm) + memory_type_multiplier = MEMORY_TYPE_HBM;
pp_display_cfg->all_displays_in_sync = context->bw_ctx.bw.dce.all_displays_in_sync; @@ -186,8 +190,18 @@ void dce11_pplib_apply_display_requirements( pp_display_cfg->cpu_pstate_separation_time = context->bw_ctx.bw.dce.blackout_recovery_time_us;
- pp_display_cfg->min_memory_clock_khz = context->bw_ctx.bw.dce.yclk_khz - / MEMORY_TYPE_MULTIPLIER_CZ; + /* + * TODO: determine whether the bandwidth has reached memory's limitation + * , then change minimum memory clock based on real-time bandwidth + * limitation. + */ + if (ASICREV_IS_VEGA20_P(dc->ctx->asic_id.hw_internal_rev) && (context->stream_count >= 2)) { + pp_display_cfg->min_memory_clock_khz = max(pp_display_cfg->min_memory_clock_khz, + (uint32_t) (dc->bw_vbios->high_yclk.value / memory_type_multiplier / 10000)); + } else { + pp_display_cfg->min_memory_clock_khz = context->bw_ctx.bw.dce.yclk_khz + / memory_type_multiplier; + }
pp_display_cfg->min_engine_clock_khz = determine_sclk_from_bounding_box( dc,
Use proper helper for 32 bit.
Reviewed-by: Harry Wentland harry.wentland@amd.com Signed-off-by: Alex Deucher alexander.deucher@amd.com (cherry picked from commit dd9212a885ca4a95443905c7c3781122a4d664e8) cc: stable@vger.kernel.org # 5.3.x --- .../gpu/drm/amd/display/dc/clk_mgr/dce110/dce110_clk_mgr.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dce110/dce110_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dce110/dce110_clk_mgr.c index 36277bca0326..b1e657e137a9 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dce110/dce110_clk_mgr.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dce110/dce110_clk_mgr.c @@ -197,7 +197,9 @@ void dce11_pplib_apply_display_requirements( */ if (ASICREV_IS_VEGA20_P(dc->ctx->asic_id.hw_internal_rev) && (context->stream_count >= 2)) { pp_display_cfg->min_memory_clock_khz = max(pp_display_cfg->min_memory_clock_khz, - (uint32_t) (dc->bw_vbios->high_yclk.value / memory_type_multiplier / 10000)); + (uint32_t) div64_s64( + div64_s64(dc->bw_vbios->high_yclk.value, + memory_type_multiplier), 10000)); } else { pp_display_cfg->min_memory_clock_khz = context->bw_ctx.bw.dce.yclk_khz / memory_type_multiplier;
On Wed, Oct 02, 2019 at 01:42:16PM -0500, Alex Deucher wrote:
Some display fixes for vega20 for stable. Fixes stability issues with certain combinations of monitors. Cherry-picked from master.
Alex Deucher (1): drm/amdgpu/display: fix 64 bit divide
Charlene Liu (1): drm/amd/display: dce11.x /dce12 update formula input
Zhan Liu (1): drm/amd/display: Add missing HBM support and raise Vega20's uclk.
.../dc/clk_mgr/dce110/dce110_clk_mgr.c | 27 ++++++++++++++++--- .../drm/amd/display/dc/dce/dce_mem_input.c | 4 +-- .../amd/display/dc/dce112/dce112_resource.c | 16 ++++++----- .../amd/display/dc/dce120/dce120_resource.c | 11 +++++--- drivers/gpu/drm/amd/display/dc/inc/resource.h | 2 ++ 5 files changed, 45 insertions(+), 15 deletions(-)
All now queued up, thanks.
greg k-h
linux-stable-mirror@lists.linaro.org
-
Alex Deucher -
Greg KH