Re: [PATCH 22/23] clocksource: new clock lookup method

[Daniel Walker <dwalker@mvista.com>]

On Wed, 2007-01-31 at 18:55 +0100, Thomas Gleixner wrote:
> On Wed, 2007-01-31 at 09:39 -0800, Daniel Walker wrote:
> > > please read my reply above! To repeat: such flags tend to get forgotten, 
> > > resulting in a less safe default behavior. Clock hardware and thus 
> > > clocksources are fundamentally fragile so we want to default to the 
> > > safest behavior. I.e. if the IS_CONTINOUS flag is 'forgotten', the clock 
> > > wont be usable as a clock verification base for example. The flag has to 
> > > be affirmatively set to mark the clocksource continous.
> > 
> > I don't see this as an issue .. Your assuming that not continuous clocks
> > will be prevalent which they aren't ..
> 
> And how does this change Ingo's statement ? Such beasts exist and having
> them default to the safe side is good. 
> 
> Also I really do not see the "huge burden" for a clocksource coder to
> add this flag. If he forgets it, then he can not use highres / dynticks
> and nothing breaks.

We're not talking about a rare system runtime configuration . If that
were the case I would agree safe is better..

I'm assuming that programmers will test their code, and others will
review the code .. Catering to any other situation doesn't make sense to
me. On top of that those clocks are rare, and not desirable ..

Daniel

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/