Re: Ondemand governor by default in etch

[Mattia Dongili <malattia@linux.it>]

On Thu, Dec 07, 2006 at 03:22:35PM -0600, John Goerzen wrote:
> On Thu, Dec 07, 2006 at 07:28:50PM +0100, Mattia Dongili wrote:
> > arch/sparc64/kernel/us2e_cpufreq.c:     policy->cpuinfo.transition_latency = 0;
> > arch/sparc64/kernel/us3_cpufreq.c:      policy->cpuinfo.transition_latency = 0;
> > 
> > I'd say half of the supported cpus haven't ondemand/conservative
> > available (CPUFREQ_ETERNAL automatically disqualifies the two
> > governors). :)
> 
> Perhaps, but if 90% of people are running the half that do support it,
> then it doesn't matter too much.

eheh, I did expect this objection ;)

> > To automagically detect the thing you just have to load the proper cpu
> > driver (a script appeared here on the list), try to load all governors
> > and see what happens (AFAIR ondemand barfs if the transition latency is
> > too high, eventually just check /sys/..../scaling_available_governors).
> 
> Where would be the proper place to insert this logic?  A new package?
> An existing one?

Hummm, see #342014 and #396117 and #367307
and see also: http://lists.debian.org/debian-devel/2005/07/msg00026.html

> When you say that ondemand barfs, is that detectable at the point that
> you try to echo something to scaling_governor?  (That is, echo would
> return an error code?)

yes (should be EINVAL)
-- 
mattia
:wq!