Next-in-Thread Next Message

Question Ping returns BAD CHECKSUM on Cabletron Roamabout 

Forum: PCMCIA Wireless Network Adapter Issues
Date: 2000, Jul 29
From: Jesse Byler jbyler

Hello all,

I have a Cabletron Roamabout 802.11 DS wireless NIC, which I
understand is the same card as the Lucent Wavelan IEEE, in
an embedded linux system running RedHat 6.2.  The hardware
consists of an Ampro CoreModule/4GV PC/104 CPU, a
Diamond-MM-16 A/D card, and a PCMCIA card for which I don't
know the manufacturer off-hand.  All 3 modules are PC/104
boards.  The software started out as a standard RedHat 6.2
installation (kernel 2.2.14-5.0).  I compiled my own kernel
from the 2.2.14 sources and added the PCMCIA cardservices
package version 3.1.15 with Lucent's wavelan2_cs binary
driver version 6.02.  I'm trying to get this card to talk to
an Apple Airport, which has a Lucent WaveLAN IEEE/silver
card inside it.

Actually, the Cabletron card *does* talk to the airport.  I
can sometimes even connect to the embedded system using ssh,
though usually outgoing ssh (from the embedded system)
doesn't get very far.  Most network connections will hang
for long periods of time and then sometimes continue,
sometimes time out.  The key piece of evidence is that ping
shows a high corruption rate on the echo replies it
receives.  Here is a portion of a ping transcript:

PING 10.0.1.4 (10.0.1.4) from 10.0.1.21 : 56(84) bytes of data.
64 bytes from 10.0.1.4: icmp_seq=0 ttl=255 time=9.1 ms
64 bytes from 10.0.1.4: icmp_seq=1 ttl=255 time=10.2 ms (BAD
CHECKSUM!)
wrong data byte #28 should be 0x1c but was 0x1e
        bd 14 81 39 16 84 3 0 8 9 a b c d e f 10 11 12 13 14 15 16 17
        18 19 1a 1
b 1e 1f 20 21
        22 23 24 25 26 27 28 29 2a 2b 2c 2d 2e 2f 30 31
64 bytes from 10.0.1.4: icmp_seq=2 ttl=255 time=8.0 ms
64 bytes from 10.0.1.4: icmp_seq=3 ttl=255 time=8.9 ms
64 bytes from 10.0.1.4: icmp_seq=4 ttl=255 time=8.0 ms (BAD CHECKSUM!)
wrong data byte #48 should be 0x30 but was 0x32
        c0 14 81 39 cd 83 3 0 8 9 a b c d e f 10 11 12 13 14 15 16 17
        18 19 1a 1
b 1c 1d 1e 1f
        20 21 22 23 24 25 26 27 28 29 2a 2b 2c 2d 2e 2f
64 bytes from 10.0.1.4: icmp_seq=5 ttl=255 time=8.0 ms (BAD CHECKSUM!)
wrong data byte #32 should be 0x20 but was 0x22
        c1 14 81 39 54 92 3 0 8 9 a b c d e f 10 11 12 13 14 15 16 17
        18 19 1a 1
b 1c 1d 1e 1f
        22 23 24 25 26 27 28 29 2a 2b 2c 2d 2e 2f 30 31
64 bytes from 10.0.1.4: icmp_seq=6 ttl=255 time=7.9 ms
64 bytes from 10.0.1.4: icmp_seq=7 ttl=255 time=7.9 ms

<...>

--- 10.0.1.4 ping statistics ---
100 packets transmitted, 79 packets received, +21 corrupted, 21%
packet loss
round-trip min/avg/max = 4.8/6.0/15.4 ms

This corruption does *not* appear if I ping myself, but it
does appear when I ping any other host (local wireless net 
or internet with NAT).  I don't know for sure that the ping
client I have on the MacOS laptops which make up the 
remainder of the wireless network checks the checksum, but
I don't see any BAD CHECKSUM errors with it.

I know the airport works as there is an existing wireless
network using it, generally with no problems.  The Cabletron
card works fine when installed on a Windows 95 system.  I've
tried both the binary driver from Lucent (latest version)
and the GPL driver from Andreas Neuhaus.  I've tried with
and without 40-bit WEP encryption.  In all cases the
behaviour is similar, though the severity of the problem
changes.  Sometimes almost no network connections succeed;
other times I can do alot.

Anyone have any ideas what the problem might be?  =) Perhaps
the Cabletron isn't really exactly the same as the Lucent
Wavelan IEEE?  Maybe the card is flaky?  Perhaps the PCMCIA
board is causing the problem?  Or is it a software issue?
Thanks in advance!

-Jesse

Next-in-Thread Next Message

Add Message to: "Ping returns BAD CHECKSUM on Cabletron Roamabout"

Members Subscribe Admin Mode Show Frames Help for HyperNews at pcmcia-cs.sourceforge.net 1.10