For the audio part I think it is just changing the mixer. I have not attempted this for HDMI since I don’t use HDMI audio. This is for a TK1, not a TX1, but some may overlap and be the same…check audio through amixer. If you run “man amixer”, and type “/list”, then a number of “list” functions will show up. That’s rather vague, but I’m afraid I have no other experience at the moment with HDMI audio. You can use “amixer cget name=“some setting”” to see a setting, “amixer cset=“some setting” value” to set something.
If the HDMI cable itself does not use any adapters and is truly HDMI, then I’d suspect the EDID is still not correct. Even when a Jetson has had trouble using EDID data it has always correctly reported what the EDID is.
Does this file exist? I’m curious if i2c is looking.
If you list i2c busses via “sudo i2cdetect -l” you should see this (this is an attempt to verify if it is possible for i2c to respond):
i2c-0 i2c Tegra I2C adapter I2C adapter
i2c-1 i2c Tegra I2C adapter I2C adapter
i2c-2 i2c Tegra I2C adapter I2C adapter
i2c-3 i2c Tegra I2C adapter I2C adapter
i2c-4 i2c Tegra I2C adapter I2C adapter
i2c-5 i2c Tegra I2C adapter I2C adapter
i2c-6 i2c Tegra I2C adapter I2C adapter
Are any of those missing?
Can anyone answer which of those i2c busses is used with EDID? I see the dts has “sor1” with this line in it, but I’m unsure of associating that with a particular controller (I’m hoping to see if that i2c bus can be queried directly to know if there is an i2c response at all, or if the response is just not valid…this should be possible by a direct query of the particular i2c):
nvidia,ddc-i2c-bus = <0x82>;
As for other 5" displays I do not know of any.