Hi, today I was able to invest some time to look again into the MIPI panel issue.
The PWM backlight output is actively driven to low by the panel. This means that I have backlight voltage if I program the OMAP5 with a pull-up and have no panel connected. If I connect the panel, the signal is 0V actively turning off the backlight. This can only be driven by the panel.
So the question is as simple: which knobs turn on the backlight?
To simplify further experiments I have added a /sys interface to the R63311 driver [1] so that I can send arbitrary commands and read back values.
It works e.g. as this:
echo 51ff >/sys/devices/ocp.3/dss.10/0.dsi/display0/dcs should send 0x51 0xff and turn backlight to full intensity
or
echo 52r >/sys/devices/ocp.3/dss.10/0.dsi/display0/dcs should send a read command 0x52 and read 1 byte back
The interesting thing is that I can really find out that register 52 returns two bytes (0x0000 = backlight permanently off) because 52r and 52rrr fails. But I can't see the value written by the 51 command :(
And if I write 0xffff to 51 the panel controller appears to hang. So this might be an invalid value which locks up the R63311.
I think the panel communication really works - but we are still missing the interpretation of the bit patterns in the max. 255 registers/commands.
So the next step is to do some trial&error on that...
BR, Nikolaus
[1]: http://git.openpandora.org/cgi-bin/gitweb.cgi?p=pyra-kernel.git;a=commit;h=0...