Strange timing problems with BETA7

[Marius Strobl]

On Mon, Oct 18, 2004 at 11:24:24AM +0200, Herve Boulouis wrote:
> Le 15/10/2004 à 17:46, Marius Strobl a écrit:
> > On Fri, Oct 15, 2004 at 05:33:29PM +0200, Herve Boulouis wrote:
> > > Le 15/10/2004 à 15:57, Marius Strobl a écrit:
> > > > 
> > > > Are there other problems like time elapsing too fast on that machine?
> > > 
> > > I've not really paid attention to that so I will keep an eye on it
> > > and I'll tell you in a few hours.
> > > 
> > 
> > Thanks. But unfortunately I got it backwards, time would have to elapse
> > too slow if it is the problem that I'm suspecting.
> 
> Well, after 2 days of uptime and some activity all the symptoms are gone,
> and no date problems.
> 
> The thing I don't understand is that I cvsuped on Oct 14, rebooted with
> new world & kernel, left the machine running for 12 hours, saw the problems
> (stucked ping & netstat, power cycle necessary to reboot, ...) reappear, rebooted the
> box Oct 15 and nothing since. The only thing I changed on the last reboot was
> to put back debug.mpsafenet to 1.
> 
> In all cases thanks to all for your help.
> 

There's potential for problems with the counter-timer which should be
fixed but this doesn't seem to have been the cause of your problems. I'm
not sure if debug.mpsafenet=0 can cause such effects. OTOH I'm running
-current with debug.mpsafenet=1 on an U60 (should be the same mainboard
as your Netra t 1125) and haven't seen such problems but in -current
John Baldwin also reworked how the kernel handles process times which
also could take of these problems.

Marius