On Wed, Mar 02, 2016 at 03:58:31PM +0530, Viresh Kumar wrote:
On 02-03-16, 11:50, Mark Brown wrote:
I'm really not sure what I can say here that I've not already said repeatedly, what's new here?
Yeah, you have raised some genuine concerns earlier, like the sequence in which the voltages of multiple supplies should be changed, and if we should rely on some DT sequence for that.
I started it again, because we have a real case now (not like the virtual case earlier) where a platform has multiple supplies for a device (CPU). I just wanted to hear from all of you on how should we go ahead and solve it.
TI guys had some crazy ideas, like using OPP-notifiers to update the second regulator, etc and I really hated it :)
My main thought with this stuff is that if it's really complicated to write generic code that might be a sign that it's sensible to write a custom driver. We can always writee something later if we figure out some general patterns.