On Wed, 25 Apr 2012 11:05:27 +0300 Felipe Balbi <balbi@xxxxxx> wrote: > Hi, > > On Wed, Apr 25, 2012 at 05:33:11PM +1000, NeilBrown wrote: > > Some USB chargers tie the ID pin low via various resistors. > > So they can cause VBUS to be high and ID to be low. > > > > The 'A' end of an OTG cable never receives VBUS, it only ever generates it. > > this isn't entirely true. Have you considered Accessory Charger > Adapters ? I confess that I probably did get lost amid the maze of twisty standards - all different. Looking at http://en.wikipedia.org/wiki/USB_On-The-Go it seems that a USB Accessory Charger Adapter can present 3 states including: A charger and a B-device are attached. The OTG device is allowed to charge and enter host mode. which would mean that the TWL4030 would see that A end of an OTG cable, and VBUS asserted. However that appears to be selected if the ID resistance is 36.5Kohms, while others are selected for 68Kohm and 124Kohm. But the twl4030 cannot detect that distinction. The cut-offs are Ground, 102K, 200K, 440K, Floating so it seems this is a standard that post-dates TWL4030. It also seems to be specific to OTG Micro plugs, and I have an OTG Mini plug (does TWL4030 support Micro plugs? Does it care?) > > > So if we see VBUS and are not generating it, this must be a charger, > > not the A end of an OTG cable, so in that case, ignore the fact that > > ID is low. > > wrong. > That may well be. However we need some way to tell twl4030_charger.c either USB_EVENT_VBUS or USB_EVENT_CHARGER when a charger is plugged in. I guess we could just punt to user-space: provide all the measurements through sysfs and allow user-space to enable the charger and select the desired current? Or should this just go in the too-hard basket for now? Thanks, NeilBrown
Attachment:
signature.asc
Description: PGP signature
