On Thu, Aug 11, 2016 at 11:20:23AM +0200, Philipp Zabel wrote:
This is Russell's HDMI notification prototype [1], currently waiting for the HDMI CEC situation to resolve.
The use case for the notifications on MediaTek MT8173 is to let the (dis)connection notifications control an ALSA jack object.
No Signed-off-by since this is not my code, and still up for discussion.
Well, I have two drivers (both CEC drivers) which use this, and I still don't see any alternative solution to the problem that this patch is solving.
I don't think it's really a CEC problem - there's three bits to HDMI that need to track each others state - the video, audio and CEC paths.
While the video and audio paths may be one block, the CEC path may actually be a separate block. For example, the TDA998x devices integrate the HDMI video/audio block along with a TDA9950 on the same device - the TDA9950 being a CEC engine. The TDA9950 is also available as a separate device, and even when integrated with HDMI, it appears on the I2C bus as a seperate device.
So, splitting the functionality is definitely the right model. We just need some way to keep each blocks state in sync. What's provided in this patch is the simple solution which seems to work for the use cases we have.
I think, in light of no comments against this approach, and no other approach being available, that this is good enough justification to merge this, especially as it is blocking other work.
So... if people want to give me reviewed-by/acked-bys, I'll add them to my patch, and I'll post that and the dw-hdmi and tda9950 CEC drivers.