On Fri, Jul 14, 2017 at 11:56:18AM +0800, Chen-Yu Tsai wrote:
> On Thu, Jul 13, 2017 at 10:13 PM, Maxime Ripard
> <maxime.ripard@free-electrons.com> wrote:
> > Just like we did for the TCON enable and disable, for historical reasons we
> > used to rely on the encoders calling the TCON mode_set function, while the
> > CRTC has a callback for that.
> >
> > Let's implement it in order to reduce the boilerplate code.
> >
> > Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
> > ---
> >  drivers/gpu/drm/sun4i/sun4i_crtc.c          | 11 ++++-
> >  drivers/gpu/drm/sun4i/sun4i_hdmi_ddc_clk.c  |  1 +-
> >  drivers/gpu/drm/sun4i/sun4i_hdmi_enc.c      |  7 +---
> >  drivers/gpu/drm/sun4i/sun4i_hdmi_tmds_clk.c |  1 +-
> >  drivers/gpu/drm/sun4i/sun4i_rgb.c           | 15 +------
> >  drivers/gpu/drm/sun4i/sun4i_tcon.c          | 56 ++++++++++------------
> >  drivers/gpu/drm/sun4i/sun4i_tcon.h          | 10 +----
> >  drivers/gpu/drm/sun4i/sun4i_tv.c            |  6 +--
> >  8 files changed, 40 insertions(+), 67 deletions(-)
> >
> 
> [...]
> 
> > diff --git a/drivers/gpu/drm/sun4i/sun4i_tcon.c b/drivers/gpu/drm/sun4i/sun4i_tcon.c
> > index dc70bc2a42a5..c4407910dfaf 100644
> > --- a/drivers/gpu/drm/sun4i/sun4i_tcon.c
> > +++ b/drivers/gpu/drm/sun4i/sun4i_tcon.c
> > @@ -106,29 +106,6 @@ void sun4i_tcon_enable_vblank(struct sun4i_tcon *tcon, bool enable)
> >  }
> >  EXPORT_SYMBOL(sun4i_tcon_enable_vblank);
> >
> > -void sun4i_tcon_set_mux(struct sun4i_tcon *tcon, int channel,
> > -                       struct drm_encoder *encoder)
> > -{
> > -       u32 val;
> > -
> > -       if (!tcon->quirks->has_unknown_mux)
> > -               return;
> > -
> > -       if (channel != 1)
> > -               return;
> > -
> > -       if (encoder->encoder_type == DRM_MODE_ENCODER_TVDAC)
> > -               val = 1;
> > -       else
> > -               val = 0;
> > -
> > -       /*
> > -        * FIXME: Undocumented bits
> > -        */
> > -       regmap_write(tcon->regs, SUN4I_TCON_MUX_CTRL_REG, val);
> > -}
> > -EXPORT_SYMBOL(sun4i_tcon_set_mux);
> > -
> >  static int sun4i_tcon_get_clk_delay(struct drm_display_mode *mode,
> >                                     int channel)
> >  {
> > @@ -147,8 +124,8 @@ static int sun4i_tcon_get_clk_delay(struct drm_display_mode *mode,
> >         return delay;
> >  }
> >
> > -void sun4i_tcon0_mode_set(struct sun4i_tcon *tcon,
> > -                         struct drm_display_mode *mode)
> > +static void sun4i_tcon0_mode_set(struct sun4i_tcon *tcon,
> > +                                struct drm_display_mode *mode)
> 
> Nit on the side: maybe we could mark mode as constant?
> Since the function doesn't change it. Same applies to the
> other mode_set functions. But this could be left to another
> patch.

We totally should. I'll do it.

> >  {
> >         unsigned int bp, hsync, vsync;
> >         u8 clk_delay;
> > @@ -221,10 +198,9 @@ void sun4i_tcon0_mode_set(struct sun4i_tcon *tcon,
> >         /* Enable the output on the pins */
> >         regmap_write(tcon->regs, SUN4I_TCON0_IO_TRI_REG, 0);
> >  }
> > -EXPORT_SYMBOL(sun4i_tcon0_mode_set);
> >
> > -void sun4i_tcon1_mode_set(struct sun4i_tcon *tcon,
> > -                         struct drm_display_mode *mode)
> > +static void sun4i_tcon1_mode_set(struct sun4i_tcon *tcon,
> > +                                struct drm_display_mode *mode)
> >  {
> >         unsigned int bp, hsync, vsync, vtotal;
> >         u8 clk_delay;
> > @@ -312,7 +288,29 @@ void sun4i_tcon1_mode_set(struct sun4i_tcon *tcon,
> >                            SUN4I_TCON_GCTL_IOMAP_MASK,
> >                            SUN4I_TCON_GCTL_IOMAP_TCON1);
> >  }
> > -EXPORT_SYMBOL(sun4i_tcon1_mode_set);
> > +
> > +void sun4i_tcon_mode_set(struct sun4i_tcon *tcon, struct drm_encoder *encoder,
> > +                        struct drm_display_mode *mode)
> 
> (also mark encoder as const?)

Yep.

> > +{
> > +       switch (encoder->encoder_type) {
> > +       case DRM_MODE_ENCODER_NONE:
> > +               sun4i_tcon0_mode_set(tcon, mode);
> > +               break;
> > +       case DRM_MODE_ENCODER_TVDAC:
> > +               /*
> > +                * FIXME: Undocumented bits
> > +                */
> > +               if (tcon->quirks->has_unknown_mux)
> > +                       regmap_write(tcon->regs, SUN4I_TCON_MUX_CTRL_REG, 1);
> > +               /* Fallthrough */
> > +       case DRM_MODE_ENCODER_TMDS:
> > +               sun4i_tcon1_mode_set(tcon, mode);
> 
> IIRC you need to clear the mux bit here. So ...
> 
> > +               break;
> > +       default:
> > +               DRM_DEBUG_DRIVER("Unknown encoder type, doing nothing...\n");
> > +       }
> 
> I think keeping the muxing in a separate function would be cleaner.
> The above is already slightly messy if you add the bit clearing part.
> With all the other muxing possibilities in the other SoC this is
> going to get really messy.

Ok.

> > +}
> > +EXPORT_SYMBOL(sun4i_tcon_mode_set);
> >
> >  static void sun4i_tcon_finish_page_flip(struct drm_device *dev,
> >                                         struct sun4i_crtc *scrtc)
> 
> [...]
> 
> Thanks for working on this. Now we've decoupled the TCON/CRTC code
> from all the encoders.

Yeah, I still have mixed feelings about this, but it was the sensible
thing I guess.

Maxime

-- 
Maxime Ripard, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com