Date: Thu, 29 Jun 2000 22:42:38 -0700 (PDT) From: Peter Van Epp <vanepp@sfu.ca> To: freebsd-hardware@freebsd.org Subject: Trap 12 kernel panics with Allied Telesyn Ethernet cards Message-ID: <200006300542.WAA09904@fraser.sfu.ca>
next in thread | raw e-mail | index | archive | help
I sent this to questions a few days ago. There has been no response, so
either there is no interest or its the wrong list. If this isn't the correct
list would some kind soul redirect me to an appropriate one please? While
Intel Etherexpress cards work just fine, I'd like to see this fixed if possible
(or at least know that hardware is broken and isn't fixable if thats the case).
I'm seeing trap 12 kernel panics from a FreeBSD 4-0 Release box when
I have a Allied Telesyn tx2500 / Realtek 10/100 Ethernet card in the machine.
It seems the card is the likely culprit because replacing it with a pair of
Intel Etherexpresses has cleared the problem. The box is an Intel 440-BX2
motherboard with a P3 600 and 256 Megs of ECC ram. There is an argus tcp audit
server running on the box sniffing on a 100meg segment out to our OC3
connection to the Internet. It appears to be input traffic related, since the
box doesn't panic at smaller traffic volumes (I have one on my downtown campus
also on a 100 link but with much lower traffic that has been up 11 days:
% ifconfig -a
rl0: flags=8943<UP,BROADCAST,RUNNING,PROMISC,SIMPLEX,MULTICAST> mtu 1500
inet 192.168.1.191 netmask 0xffffff00 broadcast 192.168.1.255
ether 00:a0:d2:05:a9:10
media: autoselect (none) status: active
supported media: autoselect 100baseTX <full-duplex> 100baseTX 10baseT/UT
% w
8:34PM up 11 days, 11:31, 1 user, load averages: 0.02, 0.01, 0.00
I expect that I can capture traffic from the link using tcpdump and
feed it back at a test box using tcpreplay and blow the driver up at will
(although I haven't verified that yet). I can certainly continue breaking it
by attaching it to the live link.
So that brings up the question(s) is there a driver maintainer
interested in trying to fix the driver? That would raise the sub question are
there later changes that might fix this and thus I should try and figure out
how to get a test machine tracking the latest 4.x code via SUP, or move to
the 4.0 STABLE release (which I didn't see anywhere in a look on
ftp.freebsd.org) and see if that helps? I'm interested in getting the driver
inproved (the cards are cheap and will output data using tcpreplay at the same
90 megs per second the Etherexpress will, with the 3c0m 905B being down at 85
megs). I have zero experience in FreeBSD kernel debugging and don't necessarily
have a lot of time but I'm willing to test fixes try if there is someone
interested in trying to make them.
Peter Van Epp / Operations and Technical Support
Simon Fraser University, Burnaby, B.C. Canada
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hardware" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200006300542.WAA09904>