Re: + clocksource-initialize-list-value.patch added to -mm tree

Re: + clocksource-initialize-list-value.patch added to -mm tree

[Thomas Gleixner]

On Fri, 2006-10-06 at 18:53 -0700, akpm@osdl.org wrote:
> A change to clocksource initialization.  It's optional for new clocksources,
> but prefered.  If the list field is initialized it allows clocksource_register
> to complete faster since it doesn't have the scan the list of clocks doing
> strcmp on each.

Either make it required or not.

> diff -puN arch/i386/kernel/hpet.c~clocksource-initialize-list-value arch/i386/kernel/hpet.c
> --- a/arch/i386/kernel/hpet.c~clocksource-initialize-list-value
> +++ a/arch/i386/kernel/hpet.c
> @@ -27,6 +27,7 @@ static struct clocksource clocksource_hp
>  	.mult		= 0, /* set below */
>  	.shift		= HPET_SHIFT,
>  	.is_continuous	= 1,
> +	.list		= CLOCKSOURCE_LIST_INIT(clocksource_hpet.list),
... 
> +/* Abstracted list initialization */
> +#define CLOCKSOURCE_LIST_INIT(x)	LIST_HEAD_INIT(x)
> +

Please use LIST_HEAD_INIT(). This is not an abstraction, this is an
obfuscation. 

	tglx

Re: + clocksource-initialize-list-value.patch added to -mm tree

[Daniel Walker]

On Sun, 2006-10-08 at 11:50 +0200, Thomas Gleixner wrote:
> On Fri, 2006-10-06 at 18:53 -0700, akpm@osdl.org wrote:
> > A change to clocksource initialization.  It's optional for new clocksources,
> > but prefered.  If the list field is initialized it allows clocksource_register
> > to complete faster since it doesn't have the scan the list of clocks doing
> > strcmp on each.
> 
> Either make it required or not.

This is for compatability with all those clocksources people are writing
that just haven't been released (assuming they will be eventually
released). I would love to make this required.

> > diff -puN arch/i386/kernel/hpet.c~clocksource-initialize-list-value arch/i386/kernel/hpet.c
> > --- a/arch/i386/kernel/hpet.c~clocksource-initialize-list-value
> > +++ a/arch/i386/kernel/hpet.c
> > @@ -27,6 +27,7 @@ static struct clocksource clocksource_hp
> >  	.mult		= 0, /* set below */
> >  	.shift		= HPET_SHIFT,
> >  	.is_continuous	= 1,
> > +	.list		= CLOCKSOURCE_LIST_INIT(clocksource_hpet.list),
> ... 
> > +/* Abstracted list initialization */
> > +#define CLOCKSOURCE_LIST_INIT(x)	LIST_HEAD_INIT(x)
> > +
> 
> Please use LIST_HEAD_INIT(). This is not an abstraction, this is an
> obfuscation. 

The reason it's there is to make the list->plist or plist->list
transition easier. I'd be happy to remove when I feel that's settled,
which may be right now.

Daniel

Re: + clocksource-initialize-list-value.patch added to -mm tree

[Arjan van de Ven]

On Sun, 2006-10-08 at 07:42 -0700, Daniel Walker wrote:
> On Sun, 2006-10-08 at 11:50 +0200, Thomas Gleixner wrote:
> > On Fri, 2006-10-06 at 18:53 -0700, akpm@osdl.org wrote:
> > > A change to clocksource initialization.  It's optional for new clocksources,
> > > but prefered.  If the list field is initialized it allows clocksource_register
> > > to complete faster since it doesn't have the scan the list of clocks doing
> > > strcmp on each.
> > 
> > Either make it required or not.
> 
> This is for compatability with all those clocksources people are writing
> that just haven't been released (assuming they will be eventually
> released). I would love to make this required.

so do it. The ones that are in development are easily fixed and
consistency is worth a LOT in terms of API. Bonus points if the kernel
printk's a helpful message on abuse.

Re: + clocksource-initialize-list-value.patch added to -mm tree

[Daniel Walker]

On Sun, 2006-10-08 at 17:07 +0200, Arjan van de Ven wrote:
> 
> so do it. The ones that are in development are easily fixed and
> consistency is worth a LOT in terms of API. Bonus points if the kernel
> printk's a helpful message on abuse.
> 

I was kicking myself for the "clocksource" kernel parameter removal, but
I guess that's more akin to being exposed to userspace. 

The printks might just end up being an OOPS, but I can look into it.

Daniel