Tool/software: Linux
It seems that when a Samsung 75" monitor (UN75J6300AF ) is controlled by CEC from an TMS320DM8168 based HDMI source , it goes from standby state to powering up when the HDMI source is booting up.During this process, somehow EDID was not read properly from this Samsung model.
Only 160 bytes of the 256 bytes EDID data were obtained by the TI81xx hdmi driver.
The driver checks EDID byte 164-166 to determine if the monitor is capable of HDMI.
In this case, the check failed and TMS320DM8168 based HDMI source only drove DVI signal to the Samsung monitor without audio packets.
According to the HDMI spec , The Samsung UN75J6300AF can de assert the HPD line to indicate the E-EDID need to be re-read and only when the HPD is high the HDMI source shell assume the EDID is readable , so I would like to know why the TMS320DM8168 is not aborting the EDID read transaction or trying to re-read the EDID after the HPD is reasserted.
See attached Scope snapshot .