Hello again,
> I even started concurrent builds of multiple packages, tarred /
> untarred stuff - the machine got slow, yes, but it was stable, up and
> running with no panics whatsoever.
Okay, as it seems now, I was a bit optimistic. I was not able to
reproduce the issue by creating heavy load on wednesday, but today,
sitting in front of the U60 and using it for normal business, I
witnessed another "cannot send_ipi"-panic.
I was killing some processes when it happened, and the machine's load
was as low as it can be.
The U60 is currently running on a kernel using your second
DELAY()-suggestion,
DELAY(i & 7);
It seems that this setting does not really solve the problem. Is it
persist with this solution, I will compile the other suggestion,
DELAY(i < 200 ? 1 : 25);
and check that one under "daily use"-conditions to see whether it
occurs then, too, or not.
I'll keep you posted.
--
Best regards / Mit freundlichen Gruessen
Volkmar Seifert
=====================================================================
eMail-Address : Volkmar Seifert <vs%nifelheim.info@localhost>
PGP-Fingerprint: 45AE 9258 C0B2 2F8D D529 3978 5283 3FD1 E97E 64B2
Website : http://www.dimensionv.de/
Blog : http://blog.nifelheim.info/tech/
=====================================================================
Attachment:
pgp892sGhk5Ba.pgp
Description: PGP signature