Re: 64-bit fields in struct net_device_stats

[Zwane Mwaikambo]

On Sun, 15 Jun 2003, Zwane Mwaikambo wrote:

> On Sun, 15 Jun 2003, Jeff wrote:
> 
> > I now realize, that locking is out of the question. Also, it has been
> 
> Well spinlocks in particular would be particularly ugly here and cause 
> horrid cache line ping pong. Other methods of synchronization would have 
> to be looked at.
> 
> > suggested to use per cpu stats and overflow into a global counter. (Thanks
> > Zwane) This might be a better idea - the problem is, the counter won't be
> > 100% accurate at all times. The degree of inaccuracy will vary with the
> > threshold value. On the other hand, if the threshold is relatively low, no
> > one will notice the difference these days.
> 
> This would be one method of doing updates and for stats it would be fine, 
> however feel free to look into other ways...

Looking closer i'd say go for it, there isn't a real locking issue, if 
you're looking for something to do this should be rather fun with not 
that much breakage, of course you'd also have to update some userland 
tools/utilities.

have fun,
	Zwane
-- 
function.linuxpower.ca