>>>>> "Ben" == Ben Greear <greearb@candelatech.com> writes:
Ben> Jes Sorensen wrote:
>> Ehhm, are you trying to transmit packets which are outside the
>> Ethernet spec? That doesn't sound like a particularly good idea to
>> me. If you want to send an extra 4 bytes of date, you have to nick
>> that from the data area in order to be within spec.
Ben> The spec, or at least some spec, allows for 802.1Q VLAN pkts that
Ben> are 4 bytes bigger. I hear most hardware will support it: you
Ben> just have to make the drivers able to handle it. I know some
Ben> cards work, at least in the 2.2 series, but others don't. At
Ben> around the 2.3.48 timeframe, the tulip driver no longer worked
Ben> (It may have been fixed by now, I haven't tested lately.)
Do you know which standard this is?
Ben> In general, while ppl are working on drivers, if they can just be
Ben> aware that 1500 (+ 4 + overhead) is an option, it shouldn't be
Ben> hard to support.
That sounds really scary, my worry here is not really whether the NICs
support it, the interesting issue is whether all (or at least the
majority) of the switches, hubs and routers on the market will do
it. The only way to be sure is if the Ethernet standard allows for
this, if there is some proposal to allow this by the 802.1Q people,
but it's not in the original Ethernet standard, then it should
certainly only be an option. The think I don't like is that it will
require that you implement driver specific versions of
dev_change_mtu() for each driver.
Jes
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Thu Mar 23 2000 - 21:00:28 EST