[PATCH] drm/amd/display: Tune min clk values for MPO for RV
Kazlauskas, Nicholas
nicholas.kazlauskas at amd.com
Thu Oct 29 13:33:36 UTC 2020
On 2020-10-29 12:31 a.m., Pratik Vishwakarma wrote:
> [Why]
> Incorrect values were resulting in flash lines
> when MPO was enabled and system was left idle.
>
> [How]
> Increase min clk values only when MPO is enabled
> and display is active to not affect S3 power.
>
> Signed-off-by: Pratik Vishwakarma <Pratik.Vishwakarma at amd.com>
Fine for now, but I'd like to understand what DCFCLK state we're in
after this change.
Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas at amd.com>
Regards,
Nicholas Kazlauskas
> ---
> .../display/dc/clk_mgr/dcn10/rv1_clk_mgr.c | 35 +++++++++++++++++--
> 1 file changed, 32 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr.c
> index e133edc587d3..c388a003956b 100644
> --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr.c
> +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn10/rv1_clk_mgr.c
> @@ -187,6 +187,22 @@ static void ramp_up_dispclk_with_dpp(
> clk_mgr->base.clks.max_supported_dppclk_khz = new_clocks->max_supported_dppclk_khz;
> }
>
> +static bool is_mpo_enabled(struct dc_state *context)
> +{
> + int i;
> +
> + for (i = 0; i < context->stream_count; i++) {
> + if (context->stream_status[i].plane_count > 1) {
> + /*
> + * No need to check for all streams as MPO
> + * can be enabled on single stream only for RV.
> + */
Nitpick: We can return early here regardless since DCFCLK isn't per pipe.
> + return true;
> + }
> + }
> + return false;
> +}
> +
> static void rv1_update_clocks(struct clk_mgr *clk_mgr_base,
> struct dc_state *context,
> bool safe_to_lower)
> @@ -284,9 +300,22 @@ static void rv1_update_clocks(struct clk_mgr *clk_mgr_base,
> if (pp_smu->set_hard_min_fclk_by_freq &&
> pp_smu->set_hard_min_dcfclk_by_freq &&
> pp_smu->set_min_deep_sleep_dcfclk) {
> - pp_smu->set_hard_min_fclk_by_freq(&pp_smu->pp_smu, new_clocks->fclk_khz / 1000);
> - pp_smu->set_hard_min_dcfclk_by_freq(&pp_smu->pp_smu, new_clocks->dcfclk_khz / 1000);
> - pp_smu->set_min_deep_sleep_dcfclk(&pp_smu->pp_smu, (new_clocks->dcfclk_deep_sleep_khz + 999) / 1000);
> + // Only increase clocks when display is active and MPO is enabled
> + if (display_count && is_mpo_enabled(context)) {
> + pp_smu->set_hard_min_fclk_by_freq(&pp_smu->pp_smu,
> + ((new_clocks->fclk_khz / 1000) * 101) / 100);
> + pp_smu->set_hard_min_dcfclk_by_freq(&pp_smu->pp_smu,
> + ((new_clocks->dcfclk_khz / 1000) * 101) / 100);
> + pp_smu->set_min_deep_sleep_dcfclk(&pp_smu->pp_smu,
> + (new_clocks->dcfclk_deep_sleep_khz + 999) / 1000);
> + } else {
> + pp_smu->set_hard_min_fclk_by_freq(&pp_smu->pp_smu,
> + new_clocks->fclk_khz / 1000);
> + pp_smu->set_hard_min_dcfclk_by_freq(&pp_smu->pp_smu,
> + new_clocks->dcfclk_khz / 1000);
> + pp_smu->set_min_deep_sleep_dcfclk(&pp_smu->pp_smu,
> + (new_clocks->dcfclk_deep_sleep_khz + 999) / 1000);
> + }
> }
> }
> }
>
More information about the amd-gfx
mailing list