[lkml]   [2008]   [Oct]   [15]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: [PATCH 53/85] i7300_idle driver v1.55

> > --- a/arch/x86/kernel/process_64.c
> > +++ b/arch/x86/kernel/process_64.c
> > @@ -62,6 +62,13 @@ void idle_notifier_register(struct notifier_block *n)
> > {
> > atomic_notifier_chain_register(&idle_notifier, n);
> > }
> > +EXPORT_SYMBOL_GPL(idle_notifier_register);
> > +
> > +void idle_notifier_unregister(struct notifier_block *n)
> > +{
> > + atomic_notifier_chain_unregister(&idle_notifier, n);
> > +}
> > +EXPORT_SYMBOL_GPL(idle_notifier_unregister);
> hm, such x86 infrastructure changes should be submitted via the x86
> tree, and you should at least have Cc:-ed the maintainers.

I agree that this patch did not hit the list in the conventional way.
I apologize for that, and I thank you, Ingo, for noticing the patch.

> The thing is, we are _getting rid_ of the idle notifiers, not extending
> them. The last thing we need is random opaque stuff getting called in
> weird ordering when we enter/exit idle state. We want all that be
> visible and have explicit, in-source-code ordering.

The patch on the table is basically a platform (chipset) idle hook.
When the system is very idle, it saves energy in the memory sub-system.
Linux customers are asking for this capability b/c it allows them
to save energy and save money.

We considered putting a platform hook into the cpuidle code,
but it seemed simpler this way -- since the new platform hook
would end up looking almost exactly like the idle notifier
that we have already. So x86-wise, all we did here is
expose the register/unregister so the driver can be modular.

I'm certainly open to suggestions.


 \ /
  Last update: 2008-10-16 00:35    [W:0.190 / U:9.816 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site