> -----Original Message-----
> From: Ville Syrjälä <[email protected]>
> Sent: Friday, December 12, 2025 10:05 PM
> To: Garg, Nemesa <[email protected]>
> Cc: [email protected]; [email protected]; Syrjala,
> Ville <[email protected]>; Shankar, Uma <[email protected]>
> Subject: Re: [PATCH 3/5] drm/i915/display: Pass dsb_commit to CASF helpers
>
> On Tue, Dec 09, 2025 at 11:55:25AM +0530, Nemesa Garg wrote:
> > Incase of non-modeset enable the casf, update the strength or disable
> > the casf through dsb.
> >
> > Signed-off-by: Nemesa Garg <[email protected]>
> > ---
> > drivers/gpu/drm/i915/display/intel_display.c | 12 ++++++++++++
> > 1 file changed, 12 insertions(+)
> >
> > diff --git a/drivers/gpu/drm/i915/display/intel_display.c
> > b/drivers/gpu/drm/i915/display/intel_display.c
> > index 256103d55409..7edfc8c2ae21 100644
> > --- a/drivers/gpu/drm/i915/display/intel_display.c
> > +++ b/drivers/gpu/drm/i915/display/intel_display.c
> > @@ -7300,6 +7300,8 @@ static void intel_atomic_dsb_finish(struct
> intel_atomic_state *state,
> > struct intel_display *display = to_intel_display(state);
> > struct intel_crtc_state *new_crtc_state =
> > intel_atomic_get_new_crtc_state(state, crtc);
> > + struct intel_crtc_state *old_crtc_state =
> > + intel_atomic_get_old_crtc_state(state, crtc);
> > unsigned int size = new_crtc_state->plane_color_changed ? 8192 :
> > 1024;
> >
> > if (!new_crtc_state->use_flipq &&
> > @@ -7332,6 +7334,16 @@ static void intel_atomic_dsb_finish(struct
> intel_atomic_state *state,
> > if (intel_crtc_needs_color_update(new_crtc_state))
> > intel_color_commit_noarm(new_crtc_state-
> >dsb_commit,
> > new_crtc_state);
> > + if (intel_casf_enabling(new_crtc_state, old_crtc_state))
> > + intel_casf_enable(new_crtc_state->dsb_commit,
> > + new_crtc_state);
> > + else if (new_crtc_state->hw.casf_params.strength !=
> > + old_crtc_state->hw.casf_params.strength)
> > + intel_casf_update_strength(new_crtc_state-
> >dsb_commit,
> > + new_crtc_state);
> > + if (intel_casf_disabling(old_crtc_state, new_crtc_state))
> > + intel_casf_disable(new_crtc_state->dsb_commit,
> > + new_crtc_state);
>
> I don't want to see this casf stuff becoming even more special. It needs to be
> made less special. So I think it should get completely buried inside the
> existing pipe scaler/pfit codepaths.
>
Hi Ville,
For sharpness we need to call casf_enable during modeset as it will enable all
registers related to scaler and sharpness and then we need to call
casf_update_strength during fastest
as only sharpness_strength needs to be updated.
Do I need to add something like this :
if (new_crtc_state->pch_pfit.enabled ||
crtc_state->hw.casf_params.casf_enable )
skl_pfit_enable(new_crtc_state);
and then call intel_casf_enable() and casf_update_strength() from
skl_pfit_enable();
> > intel_crtc_planes_update_noarm(new_crtc_state-
> >dsb_commit,
> > state, crtc);
> >
> > --
> > 2.25.1
>
> --
> Ville Syrjälä
> Intel
