Re: IPSEC transport mode w/2.2.x kernels and large packets

Alan Cox (alan@lxorguk.ukuu.org.uk)
Sat, 7 Aug 1999 20:39:27 +0100 (BST)


> > mode the hop by hop lapb RR frames murder the link more than IP retransmits
>
> I see. Well, are such networks used somewhere for transit?

Yes. You can cross most of europe on net 44.

> I mean: if host attached to ax.25 advertises MTU=128 doing TCP,
> we need not pmtu discovery. The only case, when it is bad is AX.25
> using as transit between two high MTU networks.

That occurs a fair bit

ethernet<->radiogw------------------------radiogw<->ethernet

is common

> Seriously, Alan. I am about to disable PMTU discovery to decrease
> MTU under 556(512+20+20) and to fall to fragmentation instead.

That seems very sensible as a general policy.

> but it does not help, because big servers will not believe
> to PMTU discovery for clients sitting on ethernets behind AX.25 links...

Hopefully they will take the MSS 128 hint even though they are not obliged too.
Unlike the ICMP its relatively trustable - and since the packet has the
sequence space if its not you have other problems.

Alan

-
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/