HDMI EDID

jas-rk

I noticed in the firefly-rk3288.dts that i2c5 is disabled although the schematics shows it is wired for support. Does this mean that its not possible to get EDID for HDMI? The reason for my question is because with my test 3.14 kernel I'm not seeing any valid EDID information returned from the hdmi driver and wanted to know if EDID should be possible for HDMI.

jas-rk

Well, I found out the answer myself?by turning on i2c5 it can detect EDID on my kernel.




Another question is VGA mapped to lcd0 and HDMI to lcd1 under Android ?

jas-rk

Even with i2c5 disabled in the dtb, hdmi edid is still detected in Android, so I guessing that it is turned on somewhere in the kernel.

What puzzling me at the moment is what needs to be powered in order to get the hdmi interface working. I can see VDD10_LCD and VCC_18 need to be turned on (via?ACT8846). However with just these 2 ?turned on I'm not getting any interrupts generated from the HDMI controller when plugging in a hdmi source/cable (hot plugin) which leaves me to believe the interface is not fully powered. This is using my test kernel. I also noticed on the VGA side:




1. There is a?gpio (gpio0 GPIO_C1) line to power down the ?SDA7123 which I assume doesn't effect the hdmi interface.

2. The Android VGA drivers seems to set a number of outputs on the ACT8846 when probed which I assume aren't required for hdmi.




I would appreciate if the flyfire team could comment.

jas-rk

I'd be interested to know why the Android kernel is polling instead of using interrupts?for hpd detection, possibly because of?the interface VGA??

For those of you who may be interested, I've confirmed that the hdmi interface does work using an interrupt handler for hpd, furthermore?EDID detection also works via i2c5. What is important is that?VDD10_LCD and?VCC_18 need to be powered for hdmi to work.