LWN: Comments on "CPU frequency governors and remote callbacks"

CPU frequency governors and remote callbacks

marcH — Fri, 08 Sep 2017 08:29:03 +0000

> Mostly it's a question of power intent and complexity in system design,

I've often heard that being called... "architecture" (among many other design things)

CPU frequency governors and remote callbacks

mwsealey — Thu, 07 Sep 2017 16:47:50 +0000

You could find ARM systems that do not implement it, too. And even though you may have 8 potential clock inputs in the design, you might supply them all with the same source clock giving you a granularity of "I can set the frequency of 7 other CPUs, but have to consider myself".

Mostly it's a question of power intent and complexity in system design, more than it ever is capability of an architecture or microarchitecture. Whoever is designing the clock, reset and power controllers is the person to ask ;)

CPU frequency governors and remote callbacks

dark_knight — Thu, 07 Sep 2017 04:03:45 +0000

> Other architectures may want to consider updating their cpufreq drivers to set policy->dvfs_possible_from_any_cpu field to true
> if they can support cross-CPU frequency changes.

From the point of view of a newbie like me, it would be interesting to know which other architectures, besides ARM, allow this (if any!).