On Thu, Oct 12, 2017 at 3:09 AM, Linus Walleij [off-list ref] wrote:

On Wed, Oct 11, 2017 at 2:00 PM, Maxime Ripard
[off-list ref] wrote:

quoted

What about not enforcing any muxing state when we want to mux to the
"ldo" function? We just leave it to whatever value it is, that way we
keep it under the regulator framework's control, and we don't disrupt
anything when the pin is requested.

In a way since setting the bits one way means "LDO on" and another
setting means "LDO off" those bits should be handled by the
regulator framework when used as a regulator, not pin control.

So I would say yes.

I agree. That would be the best solution.

ChenYu