Re: [Intel-gfx] [PATCH 11/21] drm/i915: split the dpll clock compute out from display vtable.

[Dave Airlie <airlied@gmail.com>]

On Wed, 8 Sept 2021 at 20:09, Jani Nikula <jani.nikula@linux.intel.com> wrote:
>
> On Wed, 08 Sep 2021, Dave Airlie <airlied@gmail.com> wrote:
> > From: Dave Airlie <airlied@redhat.com>
> >
> > this could be merged later but for now it's simple to split it out.
> > ---
> >  drivers/gpu/drm/i915/display/intel_display.c |  6 +++---
> >  drivers/gpu/drm/i915/display/intel_dpll.c    | 16 ++++++++--------
> >  drivers/gpu/drm/i915/i915_drv.h              |  8 +++++++-
> >  3 files changed, 18 insertions(+), 12 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/i915/display/intel_display.c b/drivers/gpu/drm/i915/display/intel_display.c
> > index b981a923cc2f..87950202f4ce 100644
> > --- a/drivers/gpu/drm/i915/display/intel_display.c
> > +++ b/drivers/gpu/drm/i915/display/intel_display.c
> > @@ -6768,10 +6768,10 @@ static int intel_crtc_atomic_check(struct intel_atomic_state *state,
> >               crtc_state->update_wm_post = true;
> >
> >       if (mode_changed && crtc_state->hw.enable &&
> > -         dev_priv->display.crtc_compute_clock &&
> > +         dev_priv->dpll_funcs.crtc_compute_clock &&
> >           !crtc_state->bigjoiner_slave &&
> >           !drm_WARN_ON(&dev_priv->drm, crtc_state->shared_dpll)) {
> > -             ret = dev_priv->display.crtc_compute_clock(crtc_state);
> > +             ret = dev_priv->dpll_funcs.crtc_compute_clock(crtc_state);
> >               if (ret)
> >                       return ret;
>
> It was there before, but yuck. Conditions like this with checks on the
> existence of a vfunc are really ugly. Could benefit from a wrapper - but
> that requires figuring out what the condition actually is. *facepalm*.
>
> >       }
> > @@ -8807,7 +8807,7 @@ static void intel_modeset_clear_plls(struct intel_atomic_state *state)
> >       struct intel_crtc *crtc;
> >       int i;
> >
> > -     if (!dev_priv->display.crtc_compute_clock)
> > +     if (!dev_priv->dpll_funcs.crtc_compute_clock)
> >               return;
> >
> >       for_each_new_intel_crtc_in_state(state, crtc, new_crtc_state, i) {
> > diff --git a/drivers/gpu/drm/i915/display/intel_dpll.c b/drivers/gpu/drm/i915/display/intel_dpll.c
> > index 210f91f4a576..9326c7cbb05c 100644
> > --- a/drivers/gpu/drm/i915/display/intel_dpll.c
> > +++ b/drivers/gpu/drm/i915/display/intel_dpll.c
> > @@ -1367,21 +1367,21 @@ void
> >  intel_dpll_init_clock_hook(struct drm_i915_private *dev_priv)
> >  {
> >       if (DISPLAY_VER(dev_priv) >= 9 || HAS_DDI(dev_priv))
> > -             dev_priv->display.crtc_compute_clock = hsw_crtc_compute_clock;
> > +             dev_priv->dpll_funcs.crtc_compute_clock = hsw_crtc_compute_clock;
> >       else if (HAS_PCH_SPLIT(dev_priv))
> > -             dev_priv->display.crtc_compute_clock = ilk_crtc_compute_clock;
> > +             dev_priv->dpll_funcs.crtc_compute_clock = ilk_crtc_compute_clock;
> >       else if (IS_CHERRYVIEW(dev_priv))
> > -             dev_priv->display.crtc_compute_clock = chv_crtc_compute_clock;
> > +             dev_priv->dpll_funcs.crtc_compute_clock = chv_crtc_compute_clock;
> >       else if (IS_VALLEYVIEW(dev_priv))
> > -             dev_priv->display.crtc_compute_clock = vlv_crtc_compute_clock;
> > +             dev_priv->dpll_funcs.crtc_compute_clock = vlv_crtc_compute_clock;
> >       else if (IS_G4X(dev_priv))
> > -             dev_priv->display.crtc_compute_clock = g4x_crtc_compute_clock;
> > +             dev_priv->dpll_funcs.crtc_compute_clock = g4x_crtc_compute_clock;
> >       else if (IS_PINEVIEW(dev_priv))
> > -             dev_priv->display.crtc_compute_clock = pnv_crtc_compute_clock;
> > +             dev_priv->dpll_funcs.crtc_compute_clock = pnv_crtc_compute_clock;
> >       else if (DISPLAY_VER(dev_priv) != 2)
> > -             dev_priv->display.crtc_compute_clock = i9xx_crtc_compute_clock;
> > +             dev_priv->dpll_funcs.crtc_compute_clock = i9xx_crtc_compute_clock;
> >       else
> > -             dev_priv->display.crtc_compute_clock = i8xx_crtc_compute_clock;
> > +             dev_priv->dpll_funcs.crtc_compute_clock = i8xx_crtc_compute_clock;
> >  }
> >
> >  static bool i9xx_has_pps(struct drm_i915_private *dev_priv)
> > diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
> > index 49b23ea46475..461ab0a0f088 100644
> > --- a/drivers/gpu/drm/i915/i915_drv.h
> > +++ b/drivers/gpu/drm/i915/i915_drv.h
> > @@ -393,6 +393,10 @@ struct drm_i915_fdi_link_train_funcs {
> >                              const struct intel_crtc_state *crtc_state);
> >  };
> >
> > +struct drm_i915_dpll_funcs {
>
> Nitpick, intel_dpll_funcs. Starting to spot the pattern? ;D
>
> Part of the point is that I think these may eventually move to their own
> headers, and I like to drive naming structs and functions after the file
> name. So, you'd find intel_dpll_* stuff in intel_dpll.[ch]. Or if they
> stay in i915_drv.h, at least that's the chrystal clear context.

I've got a follow up series to move some of them as an RFC locally,
so I'm quite happy to rename them all!

Dave.