On 17.01.2003 at 08:25:06, Jeremy Warren <[EMAIL PROTECTED]> wrote:

> All of our guests are running the kernel timer patch and this if the first
> time I have seen this, but it may be something to do with high utilization
> over time....

At the last SHARE, Ulrich Weigand mentioned that there was a \"second version\"
of the timer patch that was meant to address issues related to reduced
performance under high load.  I cannot reall if the new version was available,
coming soon, or what.

As I understood, in the second version the kernel would perform the normal
\"jiffies\" operation while the CPU was doing something, and revert to the \"no
timer\" operation when the system was idle.  Sounds like this might help these
reported clock drift issues (based on no reported clock drift on systems with
the timer patch backed off).

Is the updated timer patch out there?  Do these folk need to apply it (i.e. they
have the \"first version\")?  Worse, have they already applied it and it is
causing their problems?

Cheers,
Vic Cross

Reply via email to