On 29/06/18 09:17, Neil Armstrong wrote: > Hi Hans, > > On 03/04/2018 10:27, Hans Verkuil wrote: >> On 27/03/18 00:25, Rob Herring wrote: >>> On Fri, Mar 23, 2018 at 01:59:13PM +0100, Hans Verkuil wrote: >>>> From: Hans Verkuil <hans.verkuil@xxxxxxxxx> >>>> >>>> Some boards have both a DesignWare and their own CEC controller. >>>> The CEC pin is only hooked up to their own CEC controller and not >>>> to the DW controller. >>>> >>>> Add the cec-disable property to disable the DW CEC controller. >>>> >>>> This particular situation happens on Amlogic boards that have their >>>> own meson CEC controller. >>> >>> Seems like we could avoid this by describing how the CEC line is hooked >>> up which could be needed for other reasons. >> >> So there are three situations: >> >> 1) The cec pin is connected to the DW HDMI TX. That's already supported. >> 2) The cec pin is not connected at all, but the CEC IP is instantiated. >> We need the cec-disable property for that. This simply states that the >> CEC pin is not connected. >> 3) The cec pin is connected to an HDMI RX. We do not support this at the >> moment. If we want to support this, then we need a 'hdmi-rx' phandle >> that points to the HDMI receiver that the CEC pin is associated with. >> This will be similar to the already existing 'hdmi-phandle' property >> used to associate a CEC driver with an HDMI transmitter. In hindsight >> it would have been better if 'hdmi-phandle' was named 'hdmi-tx' :-( >> >> I can make a binding proposal for 3, but I have no hardware to test it, >> so I think it is better to add this only when someone has hardware. It >> will require quite a few changes to the driver and likely also the CEC core. > > Can't we simply add a property to override the HW config fields in this case ? > It will be then usable with any feature the is enabled by reading the config > bits like AHB Audio, I2c, CEC, ... and maybe many more in the future. Yes, we can do that as well. But is that a proper description of the hardware? The HW *does* support CEC, but in this case the CEC line is just not hooked up to anything. So overriding the HW config field doesn't really sound right to me. I still believe that the cec-disable property in order to describe situation 2 in the list above is the right approach. Perhaps giving it a different name (cec-not-connected?) helps overcome Rob's objections? Rob? Sorry for not following up on this earlier. Regards, Hans > > Neil > >> >> Regards, >> >> Hans >> >>> >>>> >>>> Signed-off-by: Hans Verkuil <hans.verkuil@xxxxxxxxx> >>>> Acked-by: Neil Armstrong <narmstrong@xxxxxxxxxxxx> >>>> --- >>>> Documentation/devicetree/bindings/display/bridge/dw_hdmi.txt | 3 +++ >>>> 1 file changed, 3 insertions(+) >> > > > _______________________________________________ > dri-devel mailing list > dri-devel@xxxxxxxxxxxxxxxxxxxxx > https://lists.freedesktop.org/mailman/listinfo/dri-devel > -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
