On Fri, Jul 22, 2016 at 08:16:42AM -0700, Viresh Kumar wrote: > > Long term as I was mentioning in the other thread I think it'd be good > > if the current target() drivers were modified to supply resolve_freq(), > > and that cpufreq_register_driver() were again changed to require it for > > those drivers. > > There is no need for us to force this, its really optional for such > platforms. Worst case, schedutil wouldn't work at the best, so what? > Its a platform driver's choice, isn't it ?
This would be in the context of then being able to remove the additional if statement from the hot path. To reply to the suggestion of using likely() here, I'm partial to solving it without that because I'm guessing likely() has to be an architecture-dependent thing? It seems cleaner to me if the existing few target() drivers were augmented to provide the resolve_freq() calback and its presence required. But it's certainly not a big deal and won't affect any platforms I'm involved with, at least for now. Maybe I could work on those target() drivers if things change.