Re: [Idr] Bug in draft-ietf-idr-rfc5575bis, worth fixing?

["Dongjie (Jimmy)" <[e-mail]>]

Hi John,

Thanks for bringing this to list discussion.

I agree it would be no harm to clarify it as “Is a fragment other than the first”. (Sorry for the typo with “segment”, I mean fragment actually).

And since the description of the rule  “fragment offset is not 0” is added in 5575bis, perhaps it would be helpful to note whether the semantics aligns with RFC 5575, or is a small update?

Best regards,
Jie

From: Idr [mailto:[e-mail]] On Behalf Of John Scudder
Sent: Saturday, September 26, 2020 3:15 AM
To: idr@ietf. org <[e-mail]>
Cc: [e-mail]; Hares Susan <[e-mail]>
Subject: Re: [Idr] Bug in draft-ietf-idr-rfc5575bis, worth fixing?

Hi All,

Jie brought up another point in a unicast email. He points out that

The description about component type 12 (Fragment) has been updated, the meaning of the Bitmask values is added. For Bit-6 (IsF), the new description is:

   IsF -  Is a fragment - match if [RFC0791] IP Header Fragment Offset is not 0.

This description is IMO a little bit confusing, as it simply says “is a fragment”, but the matching rule does not include the case of the first fragment (MF=1, offset=0),. Perhaps it could either update the matching rules to also include the first fragment case, or it may update the description as “fragmented and not the first segment”.

Let’s also discuss this before we take the document off of hold and continue its processing. Speaking for myself, I don’t find the quoted text confusing since the rule specifically says “fragment offset is not 0”, but on the other hand it would do no harm to change the description to the even-more-correct “Is a fragment other than the first” (I didn’t use “segment” as Jie does, since that’s potentially a new source of confusion; “segment” is a TCP term, not an IPv4 term, and it’s a *different* IPv6 term).

Thanks,

—John