To: Mark Brown <broonie@kernel.org>,
        Javier Martinez Canillas <javier@osg.samsung.com>
From: Martin Fuzzey <mfuzzey@parkeon.com>
Subject: Regulator: drivers that need to know their supply
Cc: linux-kernel@vger.kernel.org
Message-ID: <57065FB5.6040707@parkeon.com>
Date: Thu, 7 Apr 2016 15:25:09 +0200
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:38.0) Gecko/20100101
 Thunderbird/38.3.0
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1576
Lines: 56

Hi,

I am working on a driver for the tps22993 load switch.

This chip has configurable slew rates but no voltage setting
(as it's just a switch).

To implement the .enable_time() method I need the voltage (which is
the supply's voltage).

The problem is that, when my regulator is configured as always-on the
supply is not known when .enable_time() is called from regulator_register()
(it does work if the regulator is not always-on)

This is similar to the problem fixed by
     5e3ca2b349b1 regulator: Try to resolve regulators supplies on 
registration
and the follow up deadlock fix.

I tried adding a new function:

struct regulator *rdev_get_supply(struct regulator_dev *rdev)
{
     int rc;

     rc = regulator_resolve_supply(rdev);
     if (rc)
         return ERR_PTR(rc);
     return rdev->supply;
}
EXPORT_SYMBOL_GPL(rdev_get_supply);

Unfortunately that doesn't work.
First of all supply_name is not set at that point
(since it is set in regulator_register() AFTER set_machine_constraints())

Even after swapping the call order in regulator_register() it still doesn't
work for my configuration since the parent supply is probed later
so rdev_get_supply() returns -EDEFER which just results in
a warning message from _regulator_do_enable() when it propagates
back from .enable_time()

     3V3_LCD: enable_time() failed: -517

And even without that problem there is a risk of another
deadlock since the driver callbacks are called from
regulator_register() with the regulator_list mutex held
in the always_on case.

Any ideas how to do this?

Regards,

Martin