Re: [PATCH 2/3] PERF(kernel): Cleanup power events

[Thomas Renninger <trenn@suse.de>]

On Friday 12 November 2010 03:50:21 pm Jean Pihet wrote:
> On Fri, Nov 12, 2010 at 7:17 PM, Thomas Renninger <trenn@suse.de> wrote:
...
> >> > +
> >> > +#ifdef CONFIG_EVENT_POWER_TRACING_DEPRECATED
> >> > +
> >> >  #ifndef _TRACE_POWER_ENUM_
> >> >  #define _TRACE_POWER_ENUM_
> >> >  enum {
> >> > @@ -153,8 +214,32 @@ DEFINE_EVENT(power_domain, power_domain_target,
> >> >
> >> >        TP_ARGS(name, state, cpu_id)
> >> >  );
> >> > -
> >> > +#endif /* CONFIG_EVENT_POWER_TRACING_DEPRECATED */
> >> The clock and power_domain events have been recently introduced and so
> >> must be part of the new API. Can this #endif be moved right after the
> >> definition of power_end?
> > Oops, I pulled again meanwhile and the patches still patched without fuzz,
> > but probably with some offset.
> > I'll look at that and resend this one.
> Ok
Thanks for pointing this out. Because pre-processor conditionals only have 
been moved around it looks like my test build after pulling still succeeded,
while the #ifdefs/#endifs were rather messed up.

I adjusted these parts and successfully test-built on quite a lot .config 
flavors on i386, x86_64, different ppc, ia64 and s390.

> >> A string is needed here. Without it it is impossible to have the option
> >> unset.
> >> This does the trick: +bool "Deprecated power event trace API, to be
> >> removed" 
Adjusted, thanks.

> > I am currently rebuilding on several archs/flavors and hope to be able
> > to re-send this one today or on Tue.
Done.

@Ingo: If this does not go into x86/tip, but perf or whatever tree, it would
be great if you can ping me as soon as this stuff is in.
I want to cleanup the "double cpu_idle events" issues on top and make this
more architecture independent (throw cpu_idle events from cpuidle framework
instead of throwing very x86 specific mwait states, etc.).

Thanks,

       Thomas