[Letux-kernel] a weird thing in phy-twl4030

H. Nikolaus Schaller hns at goldelico.com
Fri Aug 5 18:27:05 CEST 2016


> Am 05.08.2016 um 18:10 schrieb Andreas Kemnade <andreas at kemnade.info>:
> 
> Hi,
> 
> I discovered a quite weird thing in
> phy-twl4030:
> 
> There are two separate mechanism which power up/down things.
> 1. via pm_runtime:
> twl4030_usb_runtime_suspend()
> twl4030_usb_runtime_resume()
> 
> 2, via phy_power_off()/phy_power on() (which is used from
> drivers/usb/musb/omap2430.c
> twl4030_phy_power_on()
> twl4030_phy_power_off()
> 
> The infamous OTG_EN bit is enabled by the latter pair. phy_power_off()
> and phy_power_on() is not called in a balanced way.
> Moving things between these two might fix things.
> But I do not know the rationale behind that division yet.

Maybe purely historical.

I remember that Neil did submit a lot of fixes and it did appear to me that not all of them were accepted.
That might have left out something in an incomplete state.

> I am writing about that now to get some more brains involved.

Yes, is certainly the best way since pm is not (yet) our strength here...

> 
> Regards,
> Andreas

BR,
Nikolaus




More information about the Letux-kernel mailing list