Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!purdue!bu-cs!oliveb!amdahl!rtech!wrs!hwajin
From: hwajin@wrs.wrs.com (Hwajin Bae)
Newsgroups: comp.dcom.lans
Subject: Re: Sun interfaces - strange error report
Message-ID: <766@wrs.wrs.com>
Date: 28 Sep 89 20:51:55 GMT
References: <1366@xyzzy.UUCP>
Reply-To: hwajin@wrs.wrs.com (Hwajin Bae)
Distribution: usa
Organization: Wind River Systems, Emeryville, CA
Lines: 21

In article <1366@xyzzy.UUCP> waugh@dg-rtp.dg.com (Matthew Waugh) writes:
>	le0: Received packet with ENP bit in rmd cleared
>	le0: Received packet with STP bit in rmd cleared

When an ethernet frame received cannot fit into a single
buffer in the receive ring, LANCE will try to use data-chaining 
to use multiple buffers for a packet.  The first one will have 
the STP (start-of-packet) bit set to 1 and the last one will have 
the ENP (end-of-packet) bit set.  Only the last one will have the
valid ethernet frame length in the RMD.  One common problem with most LANCE
drivers is that they assume that there won't be any data-chaining
initiated by LANCE.  They usually drop packets that doesn't have both STP
and ENP bits set and not handle the input data-chaing initiated by
LANCE at all.  This works well only if each of your RMD buffer is big enough
to hold maximum ethernet frame size (1516).  If your RMD buffer unit
size is smaller than that -- say 1024, LANCE will occasionally try to
use data-chaining.  If your driver doesn't handle data-chaining on the
input side you will get these errors.
-- 
Hwajin Bae
Wind River Systems, Emeryville, CA