Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 25 Feb 2004 12:56:48 +0100
From:      "Sonny T. Larsen" <sonny@unix.dk>
To:        freebsd-mobile@freebsd.org
Subject:   Timeout from an with Cisco Aironet MPI350
Message-ID:  <20040225115648.GA25557@unix.dk>

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

I am running 4.9-STABLE on my new IBM T40, and I just put in a Cisco
Aironet MPI350-card, since I have used regular 350 cards with the an
driver with excellent results over the last year or so.

The problem is, that although the card seems to be identified correctly,
and I am able to ifconfig the interface, I am getting a lot of timeouts
from the an driver.

Performance is very low, and the machine more or less freezes up when
the timeouts happen.

I have collected a bit of info below:

uname -a:
FreeBSD narya.staff.tdk.net 4.9-STABLE FreeBSD 4.9-STABLE #5: Mon Feb 23
13:38:28 CET 2004 sonny@narya.staff.tdk.net:/usr/src/sys/compile/NARYA  i386

>From dmesg:

an0: <Cisco Aironet MPI350> port 0x8000-0x80ff mem 0xc0400000-0xc07fffff,
0xc0200000-0xc0203fff irq 11 at device 2.0 on pci2
an0: got RSSI <-> dBM map
an0: Ethernet address: 00:02:8a:dc:90:83

Feb 25 12:19:59 narya /kernel: an0: device timeout
Feb 25 12:20:33 narya last message repeated 5 times
Feb 25 12:21:33 narya last message repeated 9 times


This is the output from a ping (10.253.0.1 is a box on the same
network):

64 bytes from 10.253.0.1: icmp_seq=0 ttl=255 time=2.156 ms
64 bytes from 10.253.0.1: icmp_seq=2 ttl=255 time=4758.609 ms
64 bytes from 10.253.0.1: icmp_seq=3 ttl=255 time=3749.048 ms
64 bytes from 10.253.0.1: icmp_seq=4 ttl=255 time=2739.507 ms
64 bytes from 10.253.0.1: icmp_seq=5 ttl=255 time=1729.866 ms
64 bytes from 10.253.0.1: icmp_seq=6 ttl=255 time=998.952 ms
64 bytes from 10.253.0.1: icmp_seq=8 ttl=255 time=4766.989 ms
64 bytes from 10.253.0.1: icmp_seq=9 ttl=255 time=3759.227 ms
64 bytes from 10.253.0.1: icmp_seq=10 ttl=255 time=2749.646 ms
64 bytes from 10.253.0.1: icmp_seq=11 ttl=255 time=1740.125 ms

For comparison, here is a ping done from the laptop, but using a
different card (with the wi-driver):

64 bytes from 10.253.0.1: icmp_seq=0 ttl=255 time=7.482 ms
64 bytes from 10.253.0.1: icmp_seq=1 ttl=255 time=3.290 ms
64 bytes from 10.253.0.1: icmp_seq=2 ttl=255 time=3.153 ms
64 bytes from 10.253.0.1: icmp_seq=3 ttl=255 time=2.997 ms
64 bytes from 10.253.0.1: icmp_seq=4 ttl=255 time=3.965 ms
64 bytes from 10.253.0.1: icmp_seq=5 ttl=255 time=4.122 ms
64 bytes from 10.253.0.1: icmp_seq=6 ttl=255 time=3.702 ms
64 bytes from 10.253.0.1: icmp_seq=7 ttl=255 time=3.048 ms

In short, packet-loss, timeouts and very bad performance with the
an-driver - strange, since it works nicely with a regular Cisco-card
(AIR-PCM352).

Any ideas?

I saw somebody mention, that maybe downgrading the firmware on the card
would help - although I am willing to find a colleague with a
windows-box to do this, I am somewhat perplexed.

It does look more like a driver-problem to me - and the suggested
firmware (5.0001) is very old indeed.

TIA,

-- 
Bye, Sonny!

"Respect is fine, but actually I've always wanted to be feared."



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20040225115648.GA25557>