SUMMARY: Giant packet from ff:ff:ff:ff:ff:ff

From: Morten K. Barfoed (morten@dsri.dk)
Date: Wed Feb 22 1995 - 19:23:49 CST


Dear sun-managers:

My original posting:

>
>Dear Sun-managers.
>
>I got the following message on the console:
>
>le0: Receive: giant packet from ff:ff:ff:ff:ff:ff
>le0: Receive: STP in rmd cleared
>
>..what does it mean ??
>

I got a lot of very quick answers, and the helpers are
too many to list.

There was a general consensus, which I can best sum up by
listing a couple of answers:

- The message is from device le0, the Lance Ethernet driver (Lance is
- just the name of the chip on which the Ethernet interface is based).
-
- A "giant packet" is an Ethernet packet longer than the maximum allowed
- by the Ethernet protocol. No properly functioning node should
- generate such packets, so the cause is either a malfunctioning node,
- or some other electrical problem in the network. For instance, at
- least one particular type of Ethernet interface card for the IBM PC
- generates a giant packet when its power is turned off.
-
- The Sun will just ignore the giant packet, so as long as they happen
- infrequently, they are harmless. If they occur often, it is
- probably a sign that something is wrong with your network (i.e. a
- loose connection). In that case you should find and fix it before it
- gets work and takes down the whole network. Unfortunately, the
- message offers no indication of the source of the giant packet since
- the all `f's from address is obvisouly bogus. The only way to locate
- the cause is usually to subdivide the network into smaller and smaller
- pieces until you determine the cause by a process of elimination.

(From mike@trdlnk.com (Michael Sullivan))

Different people reported to have had similar problems, and few were
worried about it, as long as the messages weren't too many.

- Something on your network is sending packets larger than the ethernet
- standard, probably all 1's and hence the ff:ff:ff:ff:ff:ff address. It
- happens to me every time there is a power cut (luckily less frequently
- than in the past). At the beginning I was worried about it and I think I
- tracked it down to one board in a MultiConnect Repeater. I even bought a
- spare board for when it broke. I still have it two years later..
-
- The answers vary - you should see the summaries on aurora.latech.edu
- searching for giant as a keyword (at least you can by gopher, I haven't
- tried by www), but in general, if it is frequent, something may be
- breaking on the net; if it is infrequent ignore it.

(From Richard Butler <richard@biocell.irmkant.rm.cnr.it>)

Another one:

- We used to get that regularly until we changed wiring. We had thick
- ethernet cable (the old vampire-tap stuff), then a transceiver and AUI
- cable to a 10BaseT hub, then 10BaseT wires to the workstations. All
- of these workstations constantly complained about ff:ff:ff:ff:ff:ff.
- When we changed our wiring to entirely level 5 UTP, this problem went
- away. The stories that I have told is the combination of old thick
- cable and 10BaseT causes these ghost giant packets, but we had this in
- several locations and it only occurred in this one lab. I personally
- voted for ghosts. ;^) In any case, if you have thick cable in use
- there, you might try replacing it and see if that clears up the problem.

(From "Christopher L. Barnard" <cbarnard@cs.uchicago.edu>)

And:

- >From the le(4s) man page:
- le%d: Receive: STP in rmd cleared
- The driver has received a packet that straddles
- multiple receive buffers and therefore consumes
- more than one of the LANCE chip's receive descrip-
- tors. Provided that all stations on the Ethernet
- are operating according to the Ethernet specifica-
- tion, this error "should never happen," since the
- driver allocates its receive buffers to be large
- enough to hold packets of the largest permitted
- size. Most likely, some other station on the net
- is transmitting packets whose lengths exceed the
- maximum permitted for Ethernet.
-
-
- We occasionally see this on one of our segments which has several
- old Sun 3/50s with potentially flakey ethernet hardware.
-
- I've also heard that this can be caused by large electrical disturbances
- near the network cable. Check that your cable connectors are not
- grounding nor that your cables are passing near some large transformer.

(From maclean@mnementh.cs.mcgill.ca (Matthew SAMS))

Suggestions with a different flavour did turn up, here they are:

- Your NVRAM is bad, you need to replace it.
(From Luis Vallejo <vallejo@osg.emc.com>)

- The giant packet message is the result of receiving a packet over 4,096
- bytes. Patch 101954-06 for 4.1.3_U1 corrects the problem but you can
- reset it with and ifconfig or a reboot.
(From ALAN.K.RUSSELL@x400gw.ameritech.com)

Now, I can inform you that the error is rather infrequent here, it has
happened before, but less than once a month I believe. So, with all
the really valuable information above, I won't have problems sleeping
over this one.

Thanks again you guys and girls:

Morten Krabbe Barfoed

Danish Space Research Institute phone: +45 42 88 22 77 (switch-board)
Gl. Lundtoftevej 7 phone: +45 45 87 40 77 - 161 (direct)
DK 2800 Lyngby FAX: +45 45 93 02 83
Denmark e-mail: morten@dsri.dk



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:10:17 CDT