> problem we are having is that the system clock under Linux is nowhere
> near accurate--it can be off by as much as several minutes per hour.
> The CMOS clock is not affected.
That suggests the crystals are out, or there are a lot of dropped interrupts.
The latter I've seen with SMP 2.0 kernels, but not with non SMP kernels.
> over a six minute period. A heavy load exacerbates the problem
> considerably (to the extent that xntpd seems to give up trying if the
> system is subjected to a heavy load for an extended period).
Thats suspicious. I'd expect load to make no odds. What I/O devices.
Also if you throw a DOS floppy in, type date, come back an hour later and
type date is the time ok ?
-
Linux SMP list: FIRST see FAQ at http://www.irisa.fr/prive/mentre/smp-faq/
To Unsubscribe: send "unsubscribe linux-smp" to [EMAIL PROTECTED]
