Re: rfc: high-resolution timer framework

[Alexander Shishkin <alexander.shishkin%teleca.com@localhost>]

On Thu, Nov 06, 2008 at 05:26:51PM +0000, Mindaugas Rasiukevicius wrote:
> Alexander Shishkin <alexander.shishkin%teleca.com@localhost> wrote:
> > On Thu, Nov 06, 2008 at 02:38:55PM +0100, Joerg Sonnenberger wrote:
> > > On Wed, Nov 05, 2008 at 12:10:28PM +0000, Mindaugas Rasiukevicius wrote:
> > > > Any concerns on this patch?
> > > 
> > > Hooking this into timecounters is a completely wrong abstraction.
> > > I'm fine with the userland side, but the kernel internals are simply
> > > wrong.
> >
> > Personally, I don't see why. Or, IOW, reasons for introducing a new
> > interface for the same sort of thing.
> > And this can always be changed, I take it, there's a long way ahead before
> > 6.0, so why shouldn't we have this in the tree and tested early?
> 
> Agree.
> 
> There are probably better ways than hooking this into the time-counters,
> however, I do not think it is a fundamental issue. If Joerg has the intention
> to make it better - then the whole release cycle time is available for that.
> 

I've rediffed the patchset against recent changes in current [1].

[1]: http://koowaldah.org/people/ash/netbsd/PENDING-CURRENT/

Regards,
--
Alex