[Letux-kernel] automatic charging working with 4.7

H. Nikolaus Schaller hns at goldelico.com
Tue Jul 26 22:10:34 CEST 2016


> Am 26.07.2016 um 21:07 schrieb Andreas Kemnade <andreas at kemnade.info>:
> Hi,
> I finally took the advice given to us about in the UART slave
> discussion:
> Fix powermanagement issues in userspace ;-)


> So I did:
> rmmod twl4030_charger
> sleep 2
> #seems to be already done
> twl-i2c.sh USB w 0xFE 0x01
> sleep 1
> twl-i2c.sh USB r 0xFF
> #should be 01
> twl-i2c.sh USB w 0xAD 0x20


that is what the phy-twl4030 should have done, but does not.

> sleep 1
> # now we have sometimes a working usb gadget
> #BCICTL1 (not needed)
> #twl-i2c.sh MAIN_CHARGE w 0x23 0x1f
> twl-i2c.sh MAIN_CHARGE w 0x10 0x74
> twl-i2c.sh PM_MASTER w 0x07 0x32
> sleep 2

would normally be done by twl4030_charger

> #charging starts
> #to have the vbus and iusb sysnodes in /sys/class/power_supply/twl4030_usb...
> modprobe twl4030_charger
> now we have even vbus measurement.
> twl-i2c.sh is from
> http://misc.andi.de1.cc/twl-i2c.sh
> Well, of course that is not the end solution, but maybe a step backward to move two steps forward.

Yes, it more or less mimics what the charging enable command was doing in 2.6.32 kernels...

So that is a good step forward!

a) we know for (almost) sure now that the problem is not in twl4030_charger
b) it must be in twl4030-phy (which does not handle OTG_EN) and power management (which we already know to be broken)

Great step towards a solution (but we don't know how many steps the ladder has in total :)

I think the next thing I will try is to check/change the status of OTG_EN in twl4030_charger...

Would be a simple workaround :)


More information about the Letux-kernel mailing list