answered
2016-10-17 15:05:04 +0200
I kind of managed to reproduce this. But since it required really careful fiddling with the usb cable, I wonder if what I saw is the same thing that has been reported?
The usb connection state & mode selection is done by usb-moded process. The inputs it sees and actions it takes can be debugged for example by logging into the device via ssh over wlan and then running usb-moded in verbose debug mode:
systemctl stop usb-moded ; usb_moded -DT
When charger is connected usb-moded is supposed to see dedicated charger profile udev event, for example:
CONNECTED - POWER_SUPPLY_TYPE = Unknown
CONNECTED - POWER_SUPPLY_TYPE = USB_DCP
If cable is inserted slowly, the initial detection can go wrong (leads to ui level problem - the mode selection dialog stays there even though mode can't really be seletected), but charging still works.
CONNECTED - POWER_SUPPLY_TYPE = Unknown
CONNECTED - POWER_SUPPLY_TYPE = USB
CONNECTED - POWER_SUPPLY_TYPE = USB_DCP
However, if I really carefully do the insertion, it seems that it is possible to end up not getting the final USB_DCP event -> usb moded is left thinking that it is connected to a pc instead of charger. In which case the charging just does not commence regardless of the actions taken by usb-moded.
I tried this with a Jolla1 charger, few Samsung/Nokia chargers and couple of no-name cheap ones: If I'm careful enough, I can make it happen with all chargers. But if I just plug the cable in, charging works with all of them.
Looks like you need to report this BUG to Finalwire, not Jolla :D
coderus ( 2016-08-26 18:22:15 +0200 )editSadly to know that....
Schturman ( 2016-08-26 19:56:45 +0200 )edit