On 28 May 2014 23:08, Mark Brown [off-list ref] wrote:

Whatever - I don't think the particular code makes any practical
difference.  We would need to audit existing users who don't have a
REGULATOR dependency for breakage though.

Exactly what kind of drivers are we looking to fix here? These might
be the possible cases:
- We are checking 'regulator pointer' before calling and don't need to
handle anything there..
- drivers depend on CONFIG_REGULATOR and so again we don't need
to handle anything
- None of above are true and drivers aren't checking return value of
regulator_set_voltage()
OR
They are checking it and failing when it failed..

What do we want to do in these cases?