No subject

Tue Dec 2 02:44:59 GMT 2003

from both CPU's and despite several installs of new kernels I'm still having
the same problems. Having spent some time on Google tonight this appears to
be the only relevant piece of information I could find:

"Why does the clock drift so rapidly when I run linux SMP?

This is known problem with IRQ handling and long kernel locks in the 2.0
series kernels. Consider upgrading to a later 2.2 kernel.

>From Jakob Oestergaard: Or, consider running xntpd. That should keep your
clock right on time. (I think that I've heard that enabling RTC in the
kernel also fixes the clock drift. It works for me! but I'm not sure whether
that's general or I'm just being lucky)

There are some kernel fixes in the later 2.2.x series that may fix this."

That was at:

As you can see this piece of information is a little outdated. I am
currently attempting to run the 2.4.9-21smp kernel as provided by Red Hat.
So far I am still having the same problems. I'm in the middle of compiling
my own kernel to see if that will make a difference but I'm not to hopeful
on that.

The other thing I found which may help me out, yet I am completely unable to
understand, is some things that can be modified in the /proc system.

In the /proc/irq directories apparently you can set parameters in the
prof_cpu_mask file or in the individual IRQ directories (from what I can
tell my systems RTC IRQ is 8) there is a file called smp_affinity which can
apparently be used to control which CPU has access to that IRQ. I'm guessing
that it may be possible set it so only one of the CPU's has access to RTC
interrupt and maybe that will cause the time to be kept slightly more

My main problem is that I am very confused on how that bitmask works. Is
there anyone out there with good knowledge or the /proc filesystem that may
know the answer? Or at least know wether I'm barking up entirely the wrong
tree and all this hassle could be just caused by some weird hardware issue
with my particular motherboard.

Any help would be greatly appreciated



More information about the linux mailing list