Re: [Intel-gfx] [PATCH 1/2] drm/i915: Check lane count when determining FEC support

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

On Wed, 20 Sep 2023, Ville Syrjälä <ville.syrjala@linux.intel.com> wrote:
> On Wed, Sep 20, 2023 at 12:23:34PM +0300, Jani Nikula wrote:
>> On Wed, 13 Sep 2023, Ville Syrjala <ville.syrjala@linux.intel.com> wrote:
>> > From: Ville Syrjälä <ville.syrjala@linux.intel.com>
>> >
>> > ICL doesn't support FEC with a x1 DP link. Make sure
>> > we don't try to enable FEC in such cases.
>> 
>> The question is, should we rather require x2 link for FEC?
>> 
>> I suppose x1 link with DSC+FEC is an unlikely scenario with our current
>> link bandwidth policy, so probably not a big deal.
>
> I think currently we just smash lane_count to max when using DSC.
> So doesn't really matter currently. But something to keep in mind 
> if/when we tune the policy.

The patch is actually a bit subtle. Or the existing code is.

The paths under intel_edp_dsc_compute_pipe_bpp() and
intel_dp_dsc_compute_pipe_bpp() *assume* FEC is enabled/disabled
depending on the case. They don't look at fec_enable. Any checks for
fec_enable in there would be late.

Let's say eDP gains the ability to do FEC. I don't even remember what
the DP spec says about that. But having to check for fec_enable would
trip over, because it's not initialized yet.

So the patch highlights one aspect to keep in mind, but a little bit
hides another point to keep in mind. Damned if you do, damned if you
don't...

BR,
Jani.


>
>> 
>> BR,
>> Jani.
>> 
>> >
>> > Requires a bit of reordering to make sure we've computed lane_count
>> > before checking it.
>> >
>> > Cc: Luca Coelho <luciano.coelho@intel.com>
>> > Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
>> > ---
>> >  drivers/gpu/drm/i915/display/intel_dp.c | 21 +++++++++++----------
>> >  1 file changed, 11 insertions(+), 10 deletions(-)
>> >
>> > diff --git a/drivers/gpu/drm/i915/display/intel_dp.c b/drivers/gpu/drm/i915/display/intel_dp.c
>> > index 55ba6eeaa810..2cde8ac513bb 100644
>> > --- a/drivers/gpu/drm/i915/display/intel_dp.c
>> > +++ b/drivers/gpu/drm/i915/display/intel_dp.c
>> > @@ -1363,7 +1363,8 @@ static bool intel_dp_source_supports_fec(struct intel_dp *intel_dp,
>> >  	if (DISPLAY_VER(dev_priv) >= 12)
>> >  		return true;
>> >  
>> > -	if (DISPLAY_VER(dev_priv) == 11 && encoder->port != PORT_A)
>> > +	if (DISPLAY_VER(dev_priv) == 11 &&
>> > +	    encoder->port != PORT_A && pipe_config->lane_count != 1)
>> >  		return true;
>> >  
>> >  	return false;
>> > @@ -2105,15 +2106,6 @@ int intel_dp_dsc_compute_config(struct intel_dp *intel_dp,
>> >  		&pipe_config->hw.adjusted_mode;
>> >  	int ret;
>> >  
>> > -	pipe_config->fec_enable = !intel_dp_is_edp(intel_dp) &&
>> > -		intel_dp_supports_fec(intel_dp, pipe_config);
>> > -
>> > -	if (!intel_dp_supports_dsc(intel_dp, pipe_config))
>> > -		return -EINVAL;
>> > -
>> > -	if (!intel_dp_dsc_supports_format(intel_dp, pipe_config->output_format))
>> > -		return -EINVAL;
>> > -
>> >  	/*
>> >  	 * compute pipe bpp is set to false for DP MST DSC case
>> >  	 * and compressed_bpp is calculated same time once
>> > @@ -2134,6 +2126,15 @@ int intel_dp_dsc_compute_config(struct intel_dp *intel_dp,
>> >  		}
>> >  	}
>> >  
>> > +	pipe_config->fec_enable = !intel_dp_is_edp(intel_dp) &&
>> > +		intel_dp_supports_fec(intel_dp, pipe_config);
>> > +
>> > +	if (!intel_dp_supports_dsc(intel_dp, pipe_config))
>> > +		return -EINVAL;
>> > +
>> > +	if (!intel_dp_dsc_supports_format(intel_dp, pipe_config->output_format))
>> > +		return -EINVAL;
>> > +
>> >  	/* Calculate Slice count */
>> >  	if (intel_dp_is_edp(intel_dp)) {
>> >  		pipe_config->dsc.slice_count =
>> 
>> -- 
>> Jani Nikula, Intel

-- 
Jani Nikula, Intel