Re: bq24190: What's the correct API to turn boost mode (OTG) on for the battery charger ?

[Mark A. Greer]

On Fri, May 16, 2014 at 12:35:34AM +0200, Laurent Pinchart wrote:
> Hi Mark,
> 
> Thank you for the quick reply.

Hi Laurent.  Sorry for the delay.

> On Thursday 15 May 2014 15:08:44 Mark A. Greer wrote:
> > On Thu, May 15, 2014 at 10:29:26PM +0200, Laurent Pinchart wrote:
> > > Hello,
> > 
> > [Adding Anton Vorontsov to CC list.]
> > 
> > Hi Laurent.
> > 
> > > I'm trying to enable battery charging on an OMAP4 board based on a twl6030
> > > PMIC with external bq24190 battery charger and bq27510 fuel gauge.
> > > 
> > > The system has an OTG USB port that can be used to charge the battery, and
> > > that can also be used in host mode. In that case the bq24190 needs to be
> > > switched to boost mode to provide the +5V power supply from the battery.
> > > 
> > > The bq24190 has a charge configuration register field that supports charge
> > > disabled, charge enabled and OTG (boost mode). The field is set by the
> > > bq24190 driver in response to setting the charge type :
> > > POWER_SUPPLY_CHARGE_TYPE_NONE will disable charing, and
> > > POWER_SUPPLY_CHARGE_TYPE_TRICKLE and
> > > POWER_SUPPLY_CHARGE_TYPE_FAST will enable it. However, OTG boost mode is
> > > not supported.
> > > 
> > > 
> > > The driver exposes most register fields as sysfs attributes (which doesn't
> > > sound very safe to me, but that's another story). I can thus enable OTG
> > > boost mode directly from userspace through the driver-specific API, but
> > > that just bypasses the power supply API. I'm thus not very fond of that
> > > solution.
> >
> > No, its not a good solution.  As indicated in the commit log, the sysfs
> > entries are there because there are just so many fields that don't map
> > well to the existing interface.
> > 
> > Maybe we should add support for a DT entry to enable exporting those
> > fields(disabled by default)?
> > 
> > > Possibly due to my really basic (not to say nonexistent) knowledge of the
> > > power supply subsystem I haven't found an API to expose this feature. I
> > > was wondering if someone had given a though regarding how to implement
> > > this properly.
> > 
> > What if we just added something like POWER_SUPPLY_CHARGE_TYPE_BOOST?
> 
> That was actually my first thought, and I assumed it would be rejected as 
> boost mode is not a charge type but a discharge type :-)

Er, right.  :(

> This being said, I 
> don't have enough experience with battery chargers to decide whether this is a 
> good solution. If you believe it is, I'll trust you on that.

No, you're right, that isn't a good solution.

I don't see a really good fit anywhere.  The easiest thing may be to add
a new POWER_SUPPLY_PROP_xxx power_supply_property member which can be set
to flip it into boost mode but I'm not sure if that's really the right
thing to do.

I hope Anton will comment on this.

> On a side note, is there any reference userspace implementation of a battery 
> manager ? The vendor BSP I've received with the board hardcodes calls to the 
> bq24190 in the MUSB driver to switch between OTG and trickle charge modes 
> based on USB cable connection/disconnection events. That's pretty ugly, and I 
> assume this should be handled in userspace - or is there a standard kernel 
> infrastructure for that ? 

I don't have a good feel for what's out there but the charger-manager may
be a good place to start:

Documentation/power/charger-manager.txt
drivers/power/charger-manager.c

HTH,

Mark
--
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/