Re: [PATCH v2] pm_ops: add system quiesce/activate hooks

[Johannes Berg <johannes@sipsolutions.net>]

[-- Attachment #1.1: Type: text/plain, Size: 1173 bytes --]

On Thu, 2007-04-12 at 00:47 +0400, Dmitry Krivoschekov wrote:

> I doubt your reasoning is cogent enough. If you add such a global thing,
> you should prove it, or clearly explain what is the thing for.
> I don't believe that it's not possible to solve that decrementer case
> without this change. Can't you add appropriate platform_device which
> would register early enough, to meet that timing requirement when it
> gets suspended?

I think Ben explained that.

> >  static struct pm_ops pmac_pm_ops = {
> >  	.pm_disk_mode	= PM_DISK_SHUTDOWN,
> >  	.prepare	= pmac_pm_prepare,
> > +	.quiesce	= pm_default_quiesce,
> >  	.enter		= pmac_pm_enter,
> > +	.activate	= pm_default_activate,
> >  	.finish		= pmac_pm_finish,
> >  	.valid		= pmac_pm_valid,
> >  };
> 
> So, where is that callbacks that are so desirable for powermac?

If you'd looked at the current pmac_pm_valid, you'd see that it allows
nothing. Currently, pm_ops are totally broken on powermac. The callback
that is so desirable for powermac will be added when I actually make
powermac use the pm_ops infrastructure which currently isn't possible
due to this problem.

johannes

[-- Attachment #1.2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 190 bytes --]

[-- Attachment #2: Type: text/plain, Size: 0 bytes --]