🚀 go-pugleaf

In article <1992Oct7.145853.28738@alw.nih.gov>, merike@alw.nih.gov (Merike Kaeo) writes:
> Am testing some equipment and on occasion see some void
> frames with SRC=DST=0.  To my understanding this is a 
> bogus frame.  Also, from what I understand, void frames 
> have data length of zero whereas these particular voids
> have varying sizes - up to max fddi length.  Any clue as
> to how I can track down who is doing this (other than
> disconnecting stations one-by-one) ?


I do not think there is any rule or even suggestion that void frames
have zero INFO bytes.  They can be any length that you need them to
be.  I seem to recall one router that liked occassional 40 byte voids.
I have used infrequent 107 byte voids in one FDDI implementation to
measure the ring latency.  107 was just big enough to let the node
processor get ready to receive it after sending it, making the
measurement more accurate; the exercise either produced a good value
or no value at all.

SRC=0 does seem bogus, unless you have a station with a MAC address of 0,
which itself would be bogus, provide the address is a long, 48-bit
address.  Nothing should strip the frame unless it recognizes its
source address.  (Such an orphan frame will eventually meet the same
fate as a fragment when it runs into a MAC that is transmitting.)


Vernon Schryver,  vjs@sgi.com