Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 03 Mar 1997 09:27:52 -0700
From:      Steve Passe <smp@csn.net>
To:        Steve Price <sprice@hiwaay.net>
Cc:        freebsd-smp@freefall.freebsd.org
Subject:   Re: Problems with Dual PPRO 180 SuperMicro P6DNE 
Message-ID:  <199703031627.JAA26089@clem.systemsix.com>
In-Reply-To: Your message of "Mon, 03 Mar 1997 01:57:54 CST." <331A8482.41C67EA6@hiwaay.net> 

next in thread | previous in thread | raw e-mail | index | archive | help
Hi,

>Yes, this works.  When you get a minute or two can you explain to
>me the downside of making this change, if any?  I don't want to seem
>like I'm looking a gift horse (or should that be super hacker :) in
>the mouth, but I would like to know nonetheless.  Will this make things
>run slower?  If so, how much?

it is a slight hit on performance, how much I couldn't say.  what is
going on is that your 8254 clock output is NOT connected to the APIC
so hard clock  INTs  were being lost.  These changes program
the APIC to get INTs from the 8259 PIC, which gets the 8254 clock INT.
The APIC has to do an INT ACK cycle, etc.  with the 8259 each tick.

the disturbing part is that your motherboard's mptable is stating that this INT
is attached to the APIC.  If it correctly stated the fact that it wasn't the 
software could properly handle the situation (someday anyways, right now
its not turned on).  Further, other users report success with this board
without needing these patches.  How old is your board?  Any chance of swapping
it with another?  I would suspect a hardware problem with this particular 
board (open trace, etc.)
--
Steve Passe	| powered by
smp@csn.net	|            FreeBSD




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199703031627.JAA26089>