11:48 <phh> so, I've checked this pwm problem thingy on rk3288

11:48 <phh> and it looks like the pin is not being driven

11:49 <phh> setting it to gpio function gets it driven as expected

12:55 <ayaka> you met the same problem as I reported ?

13:50 <phh> I think so yes

13:50 <phh> though it's hard to debug precisely

15:05 <ayaka> then I could confirm it, it is ok, let me find some time to do that

15:07 <phh> to be more specific, I'm measuring 110mV, which was slowly moving down (like 10mV/s)

15:08 <phh> and I also get 110mV AC?!?

15:57 <dianders> phh: totally not paying attention, but saw this and remembered that there's a magic "change which IP block does PWM" bit on rk3288. Maybe you're not getting that set?

16:01 <dianders> Some reference here <http://lists.infradead.org/pipermail/linux-arm-kernel/2014-August/280715.html>. That got NAKed (needs to happen a different way), but should point you at the right bit...

16:02 <wens> 2 IP blocks for the same function? O.o

16:04 <phh> Wtf -_-'

16:04 <phh> I'll try that, thanks

17:29 <ayaka> yes, but the second pwm which named as dwc

17:29 <ayaka> seems never be used

21:16 <mmind00> ayaka wens: that is Rockchip's way of introducing new IP blocks :-) ... i.e. when a block is new, they seem to keep the old IP around for one soc generation to have a fallback if something fails in some horrific way ... see for example the i2c controller on rk3066 and rk3188 (which has two controllers) ... and on rk3288 and newer that old IP is gone now