[PATCH 1/5] drm/encoder_slave: make mode_valid accept const struct drm_display_mode

Laurent Pinchart laurent.pinchart at ideasonboard.com
Mon Nov 18 09:43:58 UTC 2024


On Mon, Nov 18, 2024 at 11:26:03AM +0200, Jani Nikula wrote:
> On Mon, 18 Nov 2024, Dmitry Baryshkov wrote:
> > On Mon, 18 Nov 2024 at 01:33, Laurent Pinchart wrote:
> >> On Mon, Nov 18, 2024 at 01:22:12AM +0200, Dmitry Baryshkov wrote:
> >> > On Sun, 17 Nov 2024 at 22:54, Laurent Pinchart wrote:
> >> > > On Fri, Nov 15, 2024 at 11:09:26PM +0200, Dmitry Baryshkov wrote:
> >> > > > The mode_valid() callbacks of drm_encoder, drm_crtc and drm_bridge
> >> > > > accept const struct drm_display_mode argument. Change the mode_valid
> >> > > > callback of drm_encoder_slave to also accept const argument.
> >> > > >
> >> > > > Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov at linaro.org>
> >> > >
> >> > > Reviewed-by: Laurent Pinchart <laurent.pinchart+renesas at ideasonboard.com>
> >> > >
> >> > > On a side note, there's only two I2C slave encoder drivers left... I
> >> > > wonder if we could so something about them. The ch7006 and sil164
> >> > > drivers seem to be used by nouveau only, could they be moved to
> >> > > drivers/gpu/drm/nouveau/ ? We would move the whole drm_encoder_slave
> >> > > implementation there too, and leave it to die (or get taken out of limbo
> >> > > and fixed) with dispnv04.
> >> >
> >> > Or it might be better to switch to drm_bridge. Currently we also have
> >> > sil164 (sub)drivers in ast and i915 drivers. I don't know if there is
> >> > any common code to share or not. If there is some, it might be nice to
> >> > use common framework.
> >>
> >> That would require porting nouveau and i915 to drm_bridge. As much as
> >> I'd love to see that happening, I won't hold my breath.
> >
> > Me neither. Probably moving those two and drm_encoder_slave to nouveau
> > is really the best course for now.
> 
> Granted, the dvo part of i915 is ugly, but it's also only relevant for
> the oldest hardware i915 supports. Like 20 years old. Not sure there's
> much return on investment in big refactoring, more risk that it breaks
> without nobody noticing. Just let it be in i915?

That's my opinion too. The gain is not worth the risk.

-- 
Regards,

Laurent Pinchart


More information about the Freedreno mailing list