On Sat, Nov 14, 2015 at 10:16:33PM +0100, Pavel Machek wrote:
> HiOn Sat 2015-11-14 18:49:40, Mark Brown wrote:
> > On Sat, Nov 14, 2015 at 06:59:16PM +0100, Pavel Machek wrote:

> > > Well, mfd_core.c seems to call regulator_bulk_register_supply_alias()
> > > with device that does not have dev_name initialized.

> > OK, that'll be the problem then - we're not mapping the supply into the
> > individual child device but rather system wide, probably because that
> > mapping is being done too early, before we've actually created the
> > device.

> Take a look at mfd_add_device(). Yes, we do
> regulator_bulk_register_supply_alias() before doing
> platform_device_add().

Looking at this I suspect we need to re-add the code for matching
regulators on the actual struct device and do that since this is going
to be very error prone.

> I guess Charles Keepax should be listed there?

Possibly, up to them.

> > So if you look at this just templates out some boilerplate regulator API
> > client code which calls regulator_get() like any other client and then
> > hooks that regulator into the audio power management.

> Ok, so SND_SOC_DAPM_REGULATOR_SUPPLY() does not work, because I have
> two regulators, right? Is there similar macro I can use?

No?  What would make you say that?

> Do you have an example (filename, linenumber) of sound driver that
> gets this right? 

I'm not aware of any drivers that get this wrong.

> > I'm not sure how I can be any clearer that supply names are namespaced
> > by client device and that as a result fiddling around with the supply
> > name is not going to help anything.

> Well, you are saying that pretty clearly, but sound driver I seen
> assumes names are global, and /sys interface assumed the names are
> global. Give me an example I can look at and I should be able to
> figure it out... You are clear, but the kernel code seems to disagree
> with you.

Every single sound driver gets this right, none of them assume the name
is global.  What makes you say that they assume names are global?