On Fri, 2017-03-24 at 12:54 +0000, Mark Brown wrote:

On Wed, Mar 22, 2017 at 06:53:07PM +0200, Leonard Crestez wrote:

quoted

+	if (anatop_reg->bypass)
+		anatop_reg->rdesc.min_dropout_uV = 0;
+	else
+		anatop_reg->rdesc.min_dropout_uV = LDO_MIN_DROPOUT_UV;

No, this is completely broken - you can't expect to randomly change hthe
regulator description at runtime behind the back of the framework and
expect things to work.??If there is a need to do this we need an
interface for getting the current value and a way to notify of changes.

That said I would not expect the dropout voltage to be considered at
all when the regulator is bypassed, since the regulator is not
regulating it doesn't need any headroom.

It's a more complex solution but this could be handled in the core instead.
Basically the core would treat min_dropout_uV as zero if the regulator is
currently in bypass mode.

In theory a function could be added in regulator_ops to ask a regulator driver
what requirements it has for its supply but this does not seem necessary.