LANCE/ethernet problem. (original v9n179)

[Craig Dowell]

In article <8509 at brazos.Rice.edu> nickw at sol1.harlqn.co.uk (Nick Walton) writes:
>X-Sun-Spots-Digest: Volume 9, Issue 194, message 9
>In v9n179, buengc!apollo at bu.edu (Douglas Chan) writes:
>> We're getting a lot of the following pairs of error messages:
>>  le0: Received packet with ENP bit in rmd cleared
>>  le0: Received packet with STP bit in rmd cleared

Nobody, that I have read, has gone into the details of what the Lance is
really telling us, sooooo ...

The rmd mentioned is a Receive Message Descriptor -- an element of the
Lance receive ring.  The rmd specifies things like: where is the receive
buffer, how long is it, does the Lance "own" it and includes a byte for
the Lance to write status.  ENP and STP are bits in that status byte.

The Lance will scatter a packet into multiple buffers if the first buffer
is not big enough to receive the whole packet.  STP (Start Of Packet)
means that the buffer related to this rmd is the first buffer of a
scattered packet.  ENP (ENd of Packet) means that this rmd points to the
last buffer to receive data from the packet.

Many systems don't allow scattered packet data and therefore any rmd
without STP and ENP set is an unusual (error?) condition.  Lance drivers
that I have seen will just pitch rmds without both STP and ENP set.  Why
the Sun needs to print the message?  Dunno.