SMC-EtherEZ(Ultra) giving "bogus packet size" HELP, PLEASE!

SMC-EtherEZ(Ultra) giving "bogus packet size" HELP, PLEASE!

Post by Del Simmon » Mon, 04 Dec 1995 04:00:00



Thanks for taking the time to read this.

 I am having strange problems with my ethernet connection and any help any of you could
give me would be most appriceated! When I test the two boxes with SMC's EZStart disk in
the two-node test both cards test out fine so I don't think (notice i said "think". you
guys know more than me.) that there is a hardware conflict or problem. The problem
manifests itself in a dead network connection between two SMC EtherEZ cards when I boot
linux on both machines.

 If I boot Win NT or 95 or even Dos with packet drivers, Then the networking comes up like
a champ. So let me start by saying that I don't think (that "think" again) that the
gateway machine has any problems config wise. It routes packets to these other OS's
without any probs.

 It's only when I boot my second machine (pentium90) into Linux (Caldera Pre II with
kernel 1.3.45) that I have this problem.

My ip numbers so you can see my network config:

Network   : 168.121.255.72
Broadcast : 168.121.255.79
NetMask   : 255.255.255.248

The missing ip addresses are being used for ip-aliasing that has worked fine for months.
Mostly so that I can use them for Virtual Hosts under apache.

I have two Linux machines:
-----------------------------------------------------------------------
1)---------------------www.mojoski.com-(168.121.255.73)----------------
-----------------------------------------------------------------------
This is my primary gateway and web server. It stays on the net 24/7 and I rarely touch
it's net config. (haven't changed it in months.) Don't think the prob is here. It routes
to other OS's fine.

  486 DX2/66 - 32MB RAM

  Linux 1.2.3 with 2 network interfaces:

    A) PPP ISDN to Internet (168.121.17.102 - part of my ISP's network)
          aka mojoski.mindspring.com by the machines at mindspring.com

     and

    B) SMC EtherEZ card (168.121.255.73) aka mojoski.com

If I ping the second machine (see below) from here i get nothing. After the ping my "arp
-a" command gives the following output:
  ---------------------------------------------------------------------
  bash# /sbin/arp -a
  Address                 HW type         HW address              Flags
  168.121.255.75          10Mbps Ethernet 00:00:C0:68:3F:B4       C    
  ---------------------------------------------------------------------
which appears to be correct.

If i do a "netstat -i" then the output is:
  ---------------------------------------------------------------------
  bash# netstat -i
  Kernel Interface table
  Iface   MTU Met  RX-OK RX-ERR RX-DRP RX-OVR  TX-OK TX-ERR TX-DRP TX-OVR Flags
  lo     2000   0      0      0      0      0  13625      9      0      0 BLRU
  ppp0   1500   0 372417      6      0      0 412903      0      0      0 PRU
  sl0    1500   0      0      0      0      0      2      0      0      0 PRU
  eth0   1500   0  44143      0      0      0  40033      0      0      0 BRU
  ---------------------------------------------------------------------
and notice it shows no errors on eth0. (NOTE: SL0 is needed for diald)

If I do a "route" I get:
  ---------------------------------------------------------------------
  bash# /sbin/route
  Kernel routing table
  Destination     Gateway         Genmask         Flags MSS    Window Use Iface
  168.121.1.1     *               255.255.255.255 UH    1436   0    12755 ppp0
  168.121.255.72  *               255.255.255.248 U     1436   0    40170 eth0
  loopback        *               255.0.0.0       U     1936   0    14107 lo
  default         168.121.1.1     *               UG    1436   0   394810 ppp0
  ---------------------------------------------------------------------
which appears correct also, i guess. (remember it routes to 168.121.255.75 with no
prob if that machine is running windoze.)

Here is the "ifconfig" for this machine:
   --------------------------------------------------------------------
lo        Link encap:Local Loopback  
          inet addr:127.0.0.1  Bcast:127.255.255.255  Mask:255.0.0.0
          UP BROADCAST LOOPBACK RUNNING  MTU:2000  Metric:1
          RX packets:0 errors:0 dropped:0 overruns:0
          TX packets:13625 errors:9 dropped:0 overruns:0

ppp0      Link encap:Point-Point Protocol  
          inet addr:168.121.17.102  P-t-P:168.121.1.1  Mask:255.255.0.0
          UP POINTOPOINT RUNNING  MTU:1500  Metric:1
          RX packets:376116 errors:6 dropped:0 overruns:0
          TX packets:416858 errors:0 dropped:0 overruns:0

sl0       Link encap:Serial Line IP  
          inet addr:168.121.17.102  P-t-P:168.121.1.1  Mask:255.255.0.0
          UP POINTOPOINT RUNNING  MTU:1500  Metric:1
          RX packets:0 errors:0 dropped:0 overruns:0
          TX packets:2 errors:0 dropped:0 overruns:0

eth0      Link encap:10Mbps Ethernet  HWaddr 00:00:C0:C1:67:B6
          inet addr:168.121.255.73  Bcast:168.121.255.79  Mask:255.255.255.248
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:44152 errors:0 dropped:0 overruns:0
          TX packets:40077 errors:0 dropped:0 overruns:0
          Interrupt:11 Base address:0x310 Memory:c8000-ca000
  ---------------------------------------------------------------------

This machine's ppp works great.. it's online almost all of the time.
And remember the SLIP is just there for diald..

------------------------------------------------------------------------
2)--------------------pentium90.mojoski.com-(168.121.255.75)------------
------------------------------------------------------------------------
This is the machine that I can't keep on the net.

  Pentium 90 - 32MB RAM
  Linux 1.3.45 with another SMC EtherEZ  

This machine is the one I am trying to convert from Windoze NT/95 to Caldera Network
Desktop Preview II. The networking under linux worked great when I first loaded it. Then,
all of a sudden, it quit. (I have been trying to figure this one out myself for over two
weeks now.)

All of a sudden I started getting this error on my console and my messages file  and
everything quit working:
  -------------------------------------------------------------------------
  etho: bogus packet size: 53643, status=0x0 nxpg=0xee.
  eth0: next frame inconsistency, 0xee
  -------------------------------------------------------------------------
This error is the same everytime and appears every few minutes.

Here is where the funkiness begins. When I ping the gateway machine (168.121.255.73) from
this machine I get nothing again, of course, but check out the output of "arp -a":
  ----------------------------------------------------------------------
  Address                 HW type         HW address              Flags   Mask
  168.121.255.73          10Mbps Ethernet 00:00:00:00:00:00               *
  ----------------------------------------------------------------------
Note that the hardware address is wrong, as if the gateway machine hadn't answered the arp
request.

Here is output for "route" on pentium90:
  -----------------------------------------------------------------------
  Kernel routing table
  Destination     Gateway         Genmask         Flags MSS    Window Use Iface
  168.121.255.72  *               255.255.255.248 U     1500   0        2 eth0
  127.0.0.0       *               255.0.0.0       U     2000   0        2 lo
  default         168.121.255.73  *               UG    1500   0        1 eth0
  -----------------------------------------------------------------------

Here is the output of "netstat -i":
  -----------------------------------------------------------------------
  Kernel Interface table
  Iface   MTU Met  RX-OK RX-ERR RX-DRP RX-OVR  TX-OK TX-ERR TX-DRP TX-OVR Flags
  lo     2000   0     13      0      0      0     13      0      0      0 BLRU
  eth0   1500   0      0    115      0      0     13      0      0      0 BRU
  -----------------------------------------------------------------------
Note: recv errors on eth0

Here is the output of "ifconfig"
  -----------------------------------------------------------------------
lo        Link encap:Local Loopback
          inet addr:127.0.0.1  Bcast:127.255.255.255  Mask:255.0.0.0
          UP BROADCAST LOOPBACK RUNNING  MTU:2000  Metric:1
          RX packets:13 errors:0 dropped:0 overruns:0
          TX packets:13 errors:0 dropped:0 overruns:0

eth0      Link encap:10Mbps Ethernet  HWaddr 00:00:C0:68:3F:B4
          inet addr:168.121.255.75  Bcast:168.121.255.79  Mask:255.255.255.248
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:0 errors:115 dropped:0 overruns:0
          TX packets:13 errors:0 dropped:0 overruns:0
          Interrupt:11 Base address:0x290 Memory:c8000-ca000
  -----------------------------------------------------------------------

I am using Donald Becker's SMC-Ultra.c v1.12 (1/18/95) on both boxes and both drivers
apear to init correctly.

I have wiped the drives and reinstalled this Caldera at least 3 times. (I figured I must
have screwed up a config file somewhere so I was going to reset them.) Nothing seems to
help. I got sick of working on it and left pentium90 powered of for about a week and then
I came back and upgraded the kernel on pentium90 from 1.3.39 to 1.3.45 and it worked!!!!
That it it worked for about 8-10 hours, then it went back to doing the same old shit. :-(

I have done a "find" to get a list of files changed since the last time my networking
worked and I have checked all of those files. Nothing seems to be out of place. Please let
me know if you can think of a setting I should check.

Do I need to upgrade my kernel on the gateway?? Doesn't seem like I should have to since
it routes fine to Windoze.

Please let me know of any ideas you have. Please email responses to d...@mojoski.com

Thanks for taking to time to read this long boring post,
 Del Simmons

aka Mojoski

d...@mojoski.com
http://www.mojoski.com

 
 
 

SMC-EtherEZ(Ultra) giving "bogus packet size" HELP, PLEASE!

Post by roo » Sat, 09 Dec 1995 04:00:00


1.3.x are BETA, some are probably the quality of ALPHA unless you are a kernel
guru.  What happens with 1.2.13?  The ppp has to be updated with 1.3.4x  My SMC
runs 24x7 with ppp to Inet.



 
 
 

SMC-EtherEZ(Ultra) giving "bogus packet size" HELP, PLEASE!

Post by Mojos » Fri, 15 Dec 1995 04:00:00



>1.3.x are BETA, some are probably the quality of ALPHA unless you are a kernel
>guru.  What happens with 1.2.13?  The ppp has to be updated with 1.3.4x  My SMC
>runs 24x7 with ppp to Inet.


Unfortunately, The same thing happens with 1.2.13...
(I had to patch 1.2.13 with the Adaptec 2940 patch, but compiled with
no errors.)

  Any ideas?? HELP!!


 
 
 

1. got "bogus packet size" and lost connection

Hello,

I got a very strange problem : I am in a local ethernet network (no known
problem), and while not doing anything (I only had a remote X session
running), I lost my connection. I suspected an attack from someone else in
the network (it's in a student dorm), but I couln't find any evidence.
Anyway, I switched to one of the text consoles, and I was flooded with one
repeated message :

eth0 : bogus packet size: 65408, status=0, nxpg=0.
(I couldn't copy it, but that's what I remember)

Every console I switched to was flooded with this message.
I wait a few minutes, it didn't stop.
I unplugged my ethernet cable, and it stopped immediately.
I plugged back, and it came back immediately.
My X session was frozen (it was a X -query remote_host :0), and it seemed
impossible to use any of the text consoles (flood, flood...), so I
rebooted :(

After a cold reset (see further), it booted normally, the problem was
gone. I looked at /var/log/messages and found this :

Oct 11 21:26:32 saffroy kernel: h pointers  0 vs ff.
(about the same time my connection hanged)

So, am I victim of an evil attack ? Looks like a bug, anyway...
This problem is not really critical for me, but all this seems to me very
strange, so a friend of mine advised me to report it here.
Please cc: me your answers, I don't read this newsgroup usually, and there
is way too much traffic for me...

TIA for any help.

JMS

Background :
My system is a i486 DX4-100, motherboard Asus PCI486SP3 w/ on-board PCI
IDE controller, 8Mb RAM. It runs a Slackware 3.1, kernel 2.0.35. The
network device is a PCI NE2000 clone (chip RTL 8029). It has a strange
behaviour : after a session under Linux (or win95), I have to reset hard
the system, otherwise it hangs right after completing the memory check.
All parameters seem correct (IO=0xe800,IRQ 9), the pci-ne2k module seems
to detect it without trouble. I belong to this local network since over a
month now, with my PC on every day (uptime usually > 12h), and it's the
first problem I encounter.

2. Linux in the Embedded World

3. Cabletron e2100 - "bogus packet size..."

4. problems mklinux installation on 5300cs power book.

5. "weof" operation in "mt" command gives "permission denied"

6. hard drive always on

7. GETSERVBYNAME()????????????????????"""""""""""""

8. IPX System?

9. netstat -s output: "packets pruned" and "packets collapsed"

10. "Ping" packet size vs. MTU size?

11. """"""""My SoundBlast 16 pnp isn't up yet""""""""""""

12. "eth0: bogus packet" in linux 1.3.48

13. 3Com 501 and "bogus packets"