Re: [PATCH 3/4] drm/i915/pps: add intel_pps_dp_power_down()

[Jani Nikula <jani.nikula@intel.com>]

On Wed, 04 Sep 2024, Ville Syrjälä <ville.syrjala@linux.intel.com> wrote:
> On Wed, Sep 04, 2024 at 05:02:33PM +0300, Jani Nikula wrote:
>> Add intel_pps_dp_power_down() and move the VLV/CHV active pipe clear
>> there from intel_dp_link_down(), hiding the PPS pipe details inside PPS
>> code.
>> 
>> Signed-off-by: Jani Nikula <jani.nikula@intel.com>
>> ---
>>  drivers/gpu/drm/i915/display/g4x_dp.c    |  9 +--------
>>  drivers/gpu/drm/i915/display/intel_pps.c | 16 ++++++++++++++++
>>  drivers/gpu/drm/i915/display/intel_pps.h |  1 +
>>  3 files changed, 18 insertions(+), 8 deletions(-)
>> 
>> diff --git a/drivers/gpu/drm/i915/display/g4x_dp.c b/drivers/gpu/drm/i915/display/g4x_dp.c
>> index 1f9812223263..98405fbd0e04 100644
>> --- a/drivers/gpu/drm/i915/display/g4x_dp.c
>> +++ b/drivers/gpu/drm/i915/display/g4x_dp.c
>> @@ -475,14 +475,7 @@ intel_dp_link_down(struct intel_encoder *encoder,
>>  		intel_set_pch_fifo_underrun_reporting(dev_priv, PIPE_A, true);
>>  	}
>>  
>> -	msleep(intel_dp->pps.panel_power_down_delay);
>> -
>> -	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
>> -		intel_wakeref_t wakeref;
>> -
>> -		with_intel_pps_lock(intel_dp, wakeref)
>> -			intel_dp->pps.active_pipe = INVALID_PIPE;
>> -	}
>> +	intel_pps_dp_power_down(intel_dp);
>>  }
>>  
>>  static void g4x_dp_audio_enable(struct intel_encoder *encoder,
>> diff --git a/drivers/gpu/drm/i915/display/intel_pps.c b/drivers/gpu/drm/i915/display/intel_pps.c
>> index 9e54acfea992..a7f7e5e1f3aa 100644
>> --- a/drivers/gpu/drm/i915/display/intel_pps.c
>> +++ b/drivers/gpu/drm/i915/display/intel_pps.c
>> @@ -1599,6 +1599,22 @@ static void pps_init_registers(struct intel_dp *intel_dp, bool force_disable_vdd
>>  		    (intel_de_read(display, regs.pp_ctrl) & BXT_POWER_CYCLE_DELAY_MASK));
>>  }
>>  
>> +/* Call on all DP, not just eDP */
>> +void intel_pps_dp_power_down(struct intel_dp *intel_dp)
>
> The name seems to imply this powers down something, which it doesn't.

Agreed.

>
>> +{
>> +	struct intel_display *display = to_intel_display(intel_dp);
>> +	struct drm_i915_private *i915 = to_i915(display->drm);
>> +
>> +	msleep(intel_dp->pps.panel_power_down_delay);
>
> I can't actually figure out why this msleep() even exists. We already
> waited for the power down delay (by waiting for the pps state machine)
> when we turned off the panel power, so this seems completely redundant.
>
> The whole pre-ddi modeset sequence is a bit weird because the port
> enable/disable essentially happens on the wrong side of panel power
> enable/disable. But I guess that's just how the hw works and the PPS
> somehow makes sure things happen in the right order. And I suppose
> the magic PPS register write protect thing even prevents doing it
> in the opposite order (although IIRC we always disable the write
> protect mechanism).
>
>> +
>> +	if (IS_VALLEYVIEW(i915) || IS_CHERRYVIEW(i915)) {
>> +		intel_wakeref_t wakeref;
>> +
>> +		with_intel_pps_lock(intel_dp, wakeref)
>> +			intel_dp->pps.active_pipe = INVALID_PIPE;
>> +	}
>
> This part is basically the counterpart to vlv_pps_init(),
> so the function naming should probably reflect that somehow.
> vlv_pps_port_{enable,disable}() or something like that?

I ended up doing things a bit differently across the entire series,
properly isolating pps_pipe/active_pipe to VLV/CHV code only, and adding
functions for that stuff.

I left the above msleep() where it is now. Didn't dare touch it yet, and
it should be a separate change to remove it.

New version of the whole series at [1].

BR,
Jani.


[1] https://lore.kernel.org/all/cover.1725883885.git.jani.nikula@intel.com/





>
>> +}
>> +
>>  /* Call on all DP, not just eDP */
>>  void intel_pps_dp_encoder_reset(struct intel_dp *intel_dp)
>>  {
>> diff --git a/drivers/gpu/drm/i915/display/intel_pps.h b/drivers/gpu/drm/i915/display/intel_pps.h
>> index 8dbea05f498d..42f0377a93a8 100644
>> --- a/drivers/gpu/drm/i915/display/intel_pps.h
>> +++ b/drivers/gpu/drm/i915/display/intel_pps.h
>> @@ -45,6 +45,7 @@ void intel_pps_init_late(struct intel_dp *intel_dp);
>>  
>>  void intel_pps_dp_init(struct intel_dp *intel_dp);
>>  void intel_pps_dp_encoder_reset(struct intel_dp *intel_dp);
>> +void intel_pps_dp_power_down(struct intel_dp *intel_dp);
>>  void intel_pps_reset_all(struct intel_display *display);
>>  
>>  void vlv_pps_init(struct intel_encoder *encoder,
>> -- 
>> 2.39.2

-- 
Jani Nikula, Intel