05:19 <ayaka> at the firefly rk3288, it need to reset the usb hub after boot, but I don't find the request gpio at the dwc2 driver

05:19 <ayaka> I wonder whether it is really be done.

08:31 <mmind00> ayaka: http://comments.gmane.org/gmane.linux.drivers.devicetree/172103 is meaning to solve that

08:32 <mmind00> ayaka: this will do not only reset handling but also regulator handling, if that is controllable

08:34 <mmind00> ayaka: because the reset-pin is a property of the usb hub itself and not the usb controller

10:54 <ayaka> mmind00, thank you

10:57 <ayaka> mmind00, also about the usb issue at rk3288, I have confirmed it is bug in hardware, thank you for telling me your issue

11:02 <ayaka> mmind00, so adding a pinctrl-0 to usb_host1 doesn't actually work is it? then I have to send the other patch to remove it

11:02 <ayaka> http://paste.fedoraproject.org/394257/46927167/

11:05 <mmind00> ayaka: it depends ... for a lot of boards it is enough to bring the reset pin into a known state (which those pinctrl entries do)

11:05 <mmind00> ayaka: on all boards I have (firefly, popmetal, miqi) it is enough to at least make the the usb host enumerate devices connected on boot

17:23 <ayaka> mmind00, ok I need to read the code again, I just miss where it is implement

17:44 <ayaka> Oh, maybe I mistake what it does, it won't actually does a reset, but just configure that pin to the state we want

17:45 <ayaka> so I don't see something like request gpio in dwc2 driver, am I?

17:47 <mmind00> ayaka: correct ... right now we only have the pinctrl that puts the reset gpio into a known state

17:49 <ayaka> mmind00, so it is necessary to put that into &usb_host1, the work would be done in &pinctrl?

17:49 <ayaka> is it necessary

17:52 <ayaka> anyway, the pwrseq-simple is really a good idea, so just fix what I want, I want to move the idea from rk3288 to a exynos 4412 board of mine

17:54 <mmind00> the position is arguable :-) ... in the usbhost you at least have that connection that the pin is usb-related and the pin is only set on probe of the usb host

17:55 <mmind00> ayaka: if you need the pin to be set _before_ probe, just change the state from "default" to "init"

17:55 <mmind00> ayaka: see the rk3288 tsadc for reference

17:57 <ayaka> mmind00, wow, I don't know there is a such tip, I think it must be implement in the of driver. Thank you for telling me that

17:58 <mmind00> ayaka: I think it was actually done, because the tsadc needed it ... if you want to look it up, it's in drivers/base/dd.c in the function really_probe()

18:00 <ayaka> I see thank you