Re: [PATCH 12/17] drm/i915/dp: Simplify the helper get_max_compressed_bpp_with_joiner

["Ville Syrjälä" <ville.syrjala@linux.intel.com>]

On Fri, Sep 27, 2024 at 03:31:29PM +0300, Ville Syrjälä wrote:
> On Fri, Sep 27, 2024 at 02:08:26PM +0530, Ankit Nautiyal wrote:
> > Streamline the helper to get max compressed bpp for joiner case, to
> > effectively use num of pipes joined. This will make the addition of
> > ultrajoiner limitations easier and improve redability.
> > 
> > Signed-off-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>
> > ---
> >  drivers/gpu/drm/i915/display/intel_dp.c | 20 ++++++++------------
> >  1 file changed, 8 insertions(+), 12 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/i915/display/intel_dp.c b/drivers/gpu/drm/i915/display/intel_dp.c
> > index 7db037f631f7..e7fe100ef8db 100644
> > --- a/drivers/gpu/drm/i915/display/intel_dp.c
> > +++ b/drivers/gpu/drm/i915/display/intel_dp.c
> > @@ -882,7 +882,8 @@ static u32 bigjoiner_bw_max_bpp(struct intel_display *display, u32 mode_clock,
> >  	return max_bpp;
> >  }
> >  
> > -static u32 small_joiner_ram_max_bpp(struct intel_display *display, u32 mode_hdisplay)
> > +static u32 small_joiner_ram_max_bpp(struct intel_display *display, u32 mode_hdisplay,
> > +				    int num_joined_pipes)
> >  {
> >  	struct drm_i915_private *i915 = to_i915(display->drm);
> >  	u32 max_bpp;
> > @@ -890,6 +891,8 @@ static u32 small_joiner_ram_max_bpp(struct intel_display *display, u32 mode_hdis
> >  	/* Small Joiner Check: output bpp <= joiner RAM (bits) / Horiz. width */
> >  	max_bpp = small_joiner_ram_size_bits(i915) / mode_hdisplay;
> >  
> > +	max_bpp *= num_joined_pipes;
> > +
> 
> lgtm, with the caveat that this part needs a bit of adjustment
> after the *=2 gets moved here when respinning the previous patch.
> 
> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
> 
> >  	return max_bpp;
> >  }
> >  
> > @@ -899,19 +902,12 @@ u32 get_max_compressed_bpp_with_joiner(struct drm_i915_private *i915,
> >  				       int num_joined_pipes)
> >  {
> >  	struct intel_display *display = to_intel_display(&i915->drm);
> > -	u32 max_bpp_small_joiner_ram;
> > -
> > -	max_bpp_small_joiner_ram = small_joiner_ram_max_bpp(display, mode_hdisplay);
> > -
> > -	if (num_joined_pipes == 2) {
> > -		u32 max_bpp_bigjoiner = bigjoiner_bw_max_bpp(display, mode_clock, num_joined_pipes);
> > -
> > -		max_bpp_small_joiner_ram *= 2;
> > +	u32 max_bpp;
> >  
> > -		return min(max_bpp_small_joiner_ram, max_bpp_bigjoiner);
> > -	}
> > +	max_bpp = small_joiner_ram_max_bpp(display, mode_hdisplay, num_joined_pipes);
> > +	max_bpp = min(max_bpp, bigjoiner_bw_max_bpp(display, mode_clock, num_joined_pipes));

Actually, this isn't quite right. We now call bigjoiner_bw_max_bpp()
unconditioanlly (which I kinda do like), but it doesn't handle the
num_joined_pipes==1 case.

Hmm, the fact that we use min() on these makes it a bit more awkward to
handle it all in there. We'd probably have to return <TYPE>_MAX from
there to make sure the min() ignores it.

The alternative is to call these conditionally based on
num_joined_pipes. I suppose that's not all that terrible.

> >  
> > -	return max_bpp_small_joiner_ram;
> > +	return max_bpp;
> >  }
> >  
> >  u16 intel_dp_dsc_get_max_compressed_bpp(struct drm_i915_private *i915,
> > -- 
> > 2.45.2
> 
> -- 
> Ville Syrjälä
> Intel

-- 
Ville Syrjälä
Intel