Messages in this thread | | | Date | Tue, 23 Feb 1999 15:58:08 +0100 | From | Volker Dormeyer <> | Subject | Re: ARP - hardtype 6 on ether device / [kernelpatch] |
| |
Hi!
> > I contacted cabletron again; they provided a firmware upgrade for > > the switches to me. The release note of the new firmware says that > > it will send to arp replies to the ethernet (one with hard type 6 > > (IEEE802) > > and another with hard type 1 (ethernet)) to support TCP/IP > > stacks which are not rfc1042 comlpiant, too. I installed the > > firmware - result it doesen't work much better than the old version. > > I can't see the two arp replies with tcpdump or such. > > I.e. they cheated?
Perhaps, yes
> > > > Again, I did some work on sniffering. I asked myself why are the > > other UNIX machines (AIX, DEC UNIX) are having no problems. > > And at least why are also WinNT and Win95 working with this > > cabletron-switch behaviour. > > Apparently they do not verify type.
Before I patched the kernel, I did some tests on the IBM AIX-Machines. I find out that AIX also does ARP type checking. In my tests the machine only accepted ARP type 1 an ARP type 6 packets on Ethernet. ARP types of i. e. 4 or 5 are discarded. IBM refers here to RFC1042.
> > the arp reply is discarded by the linux kernel but accepted by > > the other UNIX machines. > > Any suggestions on this? > > Suggestion is to remember this sad fact, not to hurry and to think a bit. > Linux checks address type for years and it did not result in any problems. > To change kernel only because some vendor issued one more buggy > switch is suspicious idea. >
I looked through some RFCs to find out something about this behaviour. There is the host requirements rfc [rfc1122] which says a little about ethernet and IEEE802 encapsulation in Section 2.3.3. Here is a short cut of this section:
----> snip <---- 2.3.3 Ethernet and IEEE 802 Encapsulation
The IP encapsulation for Ethernets is described in RFC-894 [LINK:3], while RFC-1042 [LINK:4] describes the IP encapsulation for IEEE 802 networks. RFC-1042 elaborates and replaces the discussion in Section 3.4 of [INTRO:2].
Every Internet host connected to a 10Mbps Ethernet cable:
o MUST be able to send and receive packets using RFC-894 encapsulation;
o SHOULD be able to receive RFC-1042 packets, intermixed with RFC-894 packets; and
o MAY be able to send packets using RFC-1042 encapsulation.
An Internet host that implements sending both the RFC-894 and the RFC-1042 encapsulations MUST provide a configuration switch to select which is sent, and this switch MUST default to RFC- 894.
Note that the standard IP encapsulation in RFC-1042 does not use the protocol id value (K1=6) that IEEE reserved for IP; instead, it uses a value (K1=170) that implies an extension (the "SNAP") which can be used to hold the Ether-Type field. An Internet system MUST NOT send 802 packets using K1=6.
Address translation from Internet addresses to link-layer addresses on Ethernet and IEEE 802 networks MUST be managed by the Address Resolution Protocol (ARP). ----> snip <----
So, why not to make linux accepting ARP type 6 on ethernet?
> Probably, we should remove check for address type at all, > but it requires thinking and digging in docs.
hm, I don't know exactly!?
-Volker-
- 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/
| |