Hi, I have upgraded the sources to 3.18-rc1 [1] and it appears to work as far as things can be expected to work [2].
Linux gta04 3.18.0-rc1-gta04+ #840 SMP Tue Oct 21 08:27:15 CEST 2014 armv7l GNU/Linux
I have noticed (but not fixed) one issue that drives the OMAP5 hot (by intensively running a background thread):
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 910 root -51 0 0 0 0 S 34.5 0.0 0:26.11 irq/92-4807c000 231 root 20 0 0 0 0 D 20.2 0.0 0:11.56 kworker/0:1 1171 root 20 0 0 0 0 R 13.9 0.0 0:03.84 kworker/1:3 1279 root 20 0 0 0 0 S 5.3 0.0 0:10.27 kworker/0:3
It appears that some irq is permantently being triggered to do something on i2c5. It disappears if I revert my DT patch that tries to enable the tca6424 interrupts. So it is not a problem in mainline kernel but ours. Most likely a wrong DT config for the gpio3_78 where the tca6424 INT line is connected to and that triggers i2c5 to read the interrupt source from the tca6424. Could be something as simple as making the gpio interrupt edge triggered.
I already have found that it should be possible to check the interrupt line status at R226 (if I can find it on the board :)
BR, Nikolaus
[1]: http://git.goldelico.com/?p=gta04-kernel.git;a=shortlog;h=refs/heads/3.18-rc... [2]: http://projects.goldelico.com/p/gta04-kernel/issues/