On 1 Dec 2008, at 11:06, Magnus Westerlund wrote:
Hesham Soliman skrev:
Thanks Colin,
I agree with your rationale but I wonder if we need to support
every broken device out there. In any case, if we have to, I
prefer to require encryption than to add the XORED address option.
I'd like to hear what people in MEXT think about this, comments
from MEXT?
I don't think we should go to far to correct this bug in
implementations. My understanding is that this have been observed
but is
not a common behavior as these generic ALGs do break things
randomly and
horribly. I would only do something about if it has no minimal
impact on
the specification and any deployed implementations. Otherwise the
fix is
likely to cause more grief than the few devices that are broken.
Right - the issue should be documented, and if there is a trivial fix
it should be recommended, but this isn't a big issue.
--
Colin Perkins
http://csperkins.org/
_______________________________________________
Ietf@xxxxxxxx
https://www.ietf.org/mailman/listinfo/ietf