Subject: Re: strange clock behavior
To: None <port-sparc64@NetBSD.org>
From: Manuel Bouyer <firstname.lastname@example.org>
Date: 03/03/2006 22:42:25
On Fri, Mar 03, 2006 at 12:30:26PM +0100, Manuel Bouyer wrote:
> I've got a ultra5 running a 3.99.15 32bit kernel in a strange state.
> some processes where hung, sleeping on various channels. The problem seems
> to be, again, the clock not going on.
> [clock looping around after 10s]
> Any idea ? I can keep the system in this state for some more time, if
> someone has an idea at what to look next.
So with help from martin@, we found that there was no clock interrupt going
in. The 10s loop when looking at the time with date(1) is a side effect of
cc_microtime(). It's probably another manifestation of PR 21750.
I could enter ddb, but unfortunably got a hard hang after a few commands
(after setting a breakpoint on hardclock() to see if it was being called,
but the breakpoint didn't fire up - and the hang came a few seconds
What started this behavior is an accidental break on serial console
after swiching off/on the console server. The box sat in ddb for
several minutes before I entered "continue" at the prompt.
Manuel Bouyer <email@example.com>
NetBSD: 26 ans d'experience feront toujours la difference