[PATCH] drm/amd/display: Extends Tune min clk for MPO for RV

Pratik Vishwakarma pvishwak at amd.com
Fri Nov 27 04:23:55 UTC 2020


Friendly ping.
Please help review.

Regards
Pratik
On 25/11/20 10:02 am, Pratik Vishwakarma wrote:
> On 25/11/20 1:38 am, Harry Wentland wrote:
>> On 2020-11-24 10:55 a.m., Pratik Vishwakarma wrote:
>>> [Why]
>>> Changes in video resolution during playback cause
>>> dispclk to ramp higher but sets incompatile fclk
>>> and dcfclk values for MPO.
>>>
>>> [How]
>>> Check for MPO and set proper min clk values
>>> for this case also. This was missed during previous
>>> patch.
>>>
>>> Signed-off-by: Pratik Vishwakarma <Pratik.Vishwakarma at amd.com>
>>> ---
>>>   .../display/dc/clk_mgr/dcn10/rv1_clk_mgr.c    | 19 
>>> ++++++++++++++++---
>>>   1 file changed, 16 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 75b8240ed059..ed087a9e73bb 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
>>> @@ -275,9 +275,22 @@ 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
>>
>> Why do we want to only do this when MPO is enabled?
>>
>> Harry
>
> Hi Harry,
>
> When MPO is enabled and system moves to lower clock state, clock 
> values are not sufficient and we see flash lines across entire screen.
>
> This issue is not observed when MPO is disabled or not active.
>
> Regards,
>
> Pratik
>
>>
>>> +            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