Suresh Krishnan a écrit :
Hi Alex,
On 18/02/09 11:56 AM, Alexandru Petrescu wrote:
Dear 6MANers,
May I comment on two things about rfc2464 IPv6 over EThernet.
4. Stateless Autoconfiguration
I think a better title for this would be "Forming an IID for Ethernet".
Because that's what the majority of the text of the section describes.
An IPv6 over foo document needs to define two parameters for SLAAC.
* IID formation
* DAD params (mainly DupAddrDetectTransmits)
RFC2464 defines the former while leaving the latter undefined and hence
using the default(1). So, I think it is defining SLAAC parameters. I do
not have a strong preference, but the way I see it, the current title
looks fine.
In this sense, I think that rfc2464 section should be titled "Parameters
for SLAAC" instead of "Stateless Autoconfiguration" (sounds as if the
IID is statelessly autoconfigured).
This rfc2464 section was used as inspiration at least for the 6LoWPAN
IPv6 over 802.15.4 RFC. Reading that RFC and finding a section titled
"Stateless Address Autoconfiguration" was very surprising to me. And it
doesn't define "DAD params" either.
It is certainly not worth a bis :-)
Maybe an errata?
And because it sounds too much as "Stateless Address Autoconfiguration"
described in rfc4862, which is much more than just forming an IID, and
has message exchanges.
An IPv6 address prefix used for stateless autoconfiguration [ACONF]
of an Ethernet interface must have a length of 64 bits.
I disagree with this. There's an implementation of SLAAC over
Ethernet whose prefix can be shorter than 64 and works ok. I suppose
there's at least another similar implementation.
Given that the EUI-64 identifier as described in RFC2464 is 64 bits long
and the following statement in RFC4862
"If the sum of the prefix length and interface identifier length
does not equal 128 bits, the Prefix Information option MUST be
ignored."
I think the statement in RFC2464 makes sense. I do not believe your
implementation follows Section 4 of RFC2464. i.e. You form your IID
differently.
No no, the IID is still 64bit. Just the prefix advertised in the RA is
shorter than 64bit. And the sum is indeed less than 128.
How about modifying the paragraph you cite above to say: "If the sum is
larger than 128..."
Alex
--------------------------------------------------------------------
IETF IPv6 working group mailing list
ipv6@ietf.org
Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------
