Re: silo overflows

From: Roman Kurakin <rik_at_cronyx.ru>
Date: Fri, 01 Oct 2004 00:24:04 +0400
Giorgos Keramidas:

>I've been seeing the following for a while now.  Any hints about ways to
>track down why this happens or how to fix it?
>
>Sep 25 19:18:36 gothmog kernel: sio1: 2 more silo overflows (total 357)
>Sep 25 19:18:45 gothmog kernel: sio1: 2 more silo overflows (total 359)
>Sep 25 19:18:46 gothmog kernel: sio1: 1 more silo overflow (total 360)
>Sep 25 19:37:03 gothmog kernel: sio1: 3 more silo overflows (total 363)
>Sep 25 19:37:14 gothmog kernel: sio1: 1 more silo overflow (total 364)
>
>This is with a relatively recent 6.0-CURRENT on a Celeron _at_ 1800MHz.
>
>$ uname -v
>FreeBSD 6.0-CURRENT #0: Fri Sep 24 00:04:47 EEST 2004     \
>sysop_at_gothmog.gr:/usr/obj/usr/src/sys/GENERIC
>
This is not a bug. This is normal. This would be "fixed" after interrupt
optimisation in sio and other drivers/parts of system. BUG section of
sio(4) describes all reasons for that.
But you also should understand that you can always produce a load
on any system that would lead to problems like this one. I didn't try
this on FreeBSD, but Linux system was loaded by myself so that it
was unable to deliver interrupt to my driver within a ~0.5-1 secconds,
if my memory doesn't fails me.
I hadn't try to load it more, but I am sure it is possible and that I can
reproduce this on any other system (except realtime one I believe ;-).

Best regards,
                        Roman Kurakin

>_______________________________________________
>freebsd-current_at_freebsd.org mailing list
>http://lists.freebsd.org/mailman/listinfo/freebsd-current
>To unsubscribe, send any mail to "freebsd-current-unsubscribe_at_freebsd.org"
>
>
>  
>
Received on Thu Sep 30 2004 - 18:35:09 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:38:14 UTC