Re: MC 7304 ipv4v6

[Bjørn Mork <bjorn mork no>]

Dan Williams <dcbw redhat com> writes:

> Well, if we're going to add rawip support, why bother emulating a
> netdev at all?

(assuming you mean etherdev - we'll need the netdev in any case :)

Good question. Asked before, but I have fewer and fewer arguments. I
know Marcel H has wanted a raw device from the beginning.

> I guess because it's convenient and we don't have to
> destroy the existing wwan0 and create some other tun-type device
> instead.  But still, seems odd that we'd set raw-ip mode and then
> emulate ethernet when all that's really going back and forth is IP.

Yes, it is odd and it does create a few problems which the driver tries
to hide.  It would be more honest to just let userspace decide what to
do, e.g, wrt bridging.

So when the firmware finally gave up messing with the L2 headers, why
should the driver continue to do that? Make it raw IP all the way.  Yes,
I'm starting to lean to that side as well.

The only "real" objection I once had was that I don't think there are
any DHCP clients on Linux that will work with such an interface. But
that is a lousy excuse. Using DHCP here is another bad idea, trying to
make this link like something it isn't.  And if you still want to do it,
then you should be able to fake the headers using the tun driver.

You are right.  The only remaining problem is that it is more work :)

But I believe I implemented it the last time this was discussed.  I'll
see if I can dig up the old code for a demo.  Not going to rush this
anyway.  I'd like everyone to be happy with the result.  And it's going
to be immutable once it's in.


Bjørn