tech-kern archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: Broadcast traffic on vlans leaks into the parent interface on NetBSD-5.1



Roy Marples <roy%marples.name@localhost> writes:

> On 05/12/2012 8:39, David Laight wrote:>
>> Although DHCP has to do strange things in order to acquire the
>> original lease, renewing it should really only requires packets
>> with the current IP address.
>
> It's not strange, it's very well documented in RFC2131.
> If renewal fails, it will be broadcast to all nodes from it's IP
> address
> If that fails then we re-enter the discover phase, ie back to your
> "strange things".
>
> If any DHCP foo is "leaking" then it is indeed a bug within BPF.

I am pretty sure the bug is not in bpf, but in the wm driver.  One
shouldn't receive a packet with hardware vlan processing and then just
call bpf_tap, because it doesn't represent the bits that came in.

Easy improvements are

  decline to call bpf_tap if the metadata indicates vlan processing was
  done on this packet

  disable hw vlan processing (until this is fixed correctly)

  expose vlan offload as a capability to be enabled, like the rest.

   

Attachment: pgp9rBM5gxVdL.pgp
Description: PGP signature



Home | Main Index | Thread Index | Old Index