Re: [tcpm] Benjamin Kaduk's Discuss on draft-ietf-tcpm-rfc793bis-25: (with DISCUSS and COMMENT)

Benjamin Kaduk <kaduk@mit.edu> Fri, 25 February 2022 21:00 UTC

Return-Path: <kaduk@mit.edu>
X-Original-To: tcpm@ietfa.amsl.com
Delivered-To: tcpm@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 3500F3A0900; Fri, 25 Feb 2022 13:00:59 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.498
X-Spam-Level:
X-Spam-Status: No, score=-1.498 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, KHOP_HELO_FCRDNS=0.399, SPF_HELO_NONE=0.001, SPF_NONE=0.001, URIBL_BLOCKED=0.001] autolearn=no autolearn_force=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id RVreCsWuun7Q; Fri, 25 Feb 2022 13:00:53 -0800 (PST)
Received: from outgoing.mit.edu (outgoing-auth-1.mit.edu [18.9.28.11]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 828F33A08C1; Fri, 25 Feb 2022 13:00:51 -0800 (PST)
Received: from kduck.mit.edu ([24.16.140.251]) (authenticated bits=56) (User authenticated as kaduk@ATHENA.MIT.EDU) by outgoing.mit.edu (8.14.7/8.12.4) with ESMTP id 21PL0Wgk000356 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-GCM-SHA384 bits=256 verify=NOT); Fri, 25 Feb 2022 16:00:39 -0500
Date: Fri, 25 Feb 2022 13:00:31 -0800
From: Benjamin Kaduk <kaduk@mit.edu>
To: Wesley Eddy <wes@mti-systems.com>
Cc: The IESG <iesg@ietf.org>, draft-ietf-tcpm-rfc793bis@ietf.org, tcpm-chairs@ietf.org, tcpm@ietf.org, Michael Scharf <michael.scharf@hs-esslingen.de>
Message-ID: <20220225210031.GV12881@kduck.mit.edu>
References: <163236803976.28405.5643771942452620510@ietfa.amsl.com> <6f6e7b90-081a-74b4-b329-8879addcb8c4@mti-systems.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <6f6e7b90-081a-74b4-b329-8879addcb8c4@mti-systems.com>
Archived-At: <https://mailarchive.ietf.org/arch/msg/tcpm/Bx3PjaryrmbVPzyhCkiLHpb7MNc>
Subject: Re: [tcpm] Benjamin Kaduk's Discuss on draft-ietf-tcpm-rfc793bis-25: (with DISCUSS and COMMENT)
X-BeenThere: tcpm@ietf.org
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: TCP Maintenance and Minor Extensions Working Group <tcpm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/tcpm>, <mailto:tcpm-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/tcpm/>
List-Post: <mailto:tcpm@ietf.org>
List-Help: <mailto:tcpm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/tcpm>, <mailto:tcpm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 25 Feb 2022 21:01:00 -0000

Hi Wes,

I'm also sorry for the slow response -- your note arrived during a
particularly busy time, and since I was going to have to swap in a bunch of
state on the document to reply, I needed a bigger chunk of time in order to
produce something useful.

On Sun, Jan 09, 2022 at 11:02:12PM -0500, Wesley Eddy wrote:
> Hello, I'm looping back on this thread, since we need to close the loop 
> on the DISCUSS points, and address the other COMMENT contents.  Sorry 
> for the rather slow response, but I've finally been able to process 
> everything from your review:
> 
> 
> On 9/22/2021 11:34 PM, Benjamin Kaduk via Datatracker wrote:
> > Benjamin Kaduk has entered the following ballot position for
> > draft-ietf-tcpm-rfc793bis-25: Discuss
> >
> > ...
> > ----------------------------------------------------------------------
> > DISCUSS:
> > ----------------------------------------------------------------------
> >
> > Many thanks for taking on the task of producing a roll-up update for the
> > core TCP specification!  I am sure it was a lot of work, but I am happy
> > to see it done.
> >
> > That said, I do have a few points that I would like to have a bit more
> > discussion on before the document is published; I'm happy to see that
> > Warren already linked to
> > https://www.ietf.org/blog/handling-iesg-ballot-positions/  on the topic
> > of what a DISCUSS position can (and cannot) mean.
> >
> > (1) We incorporate some long-standing enhancements that improve the
> > security and robustness of TCP (in particular, random ISN and protection
> > against off-path in-window attacks come to mind), but only at SHOULD or
> > MAY requirements level.
> >
> > For example, we currently say:
> >
> >     A TCP implementation MUST use the above type of "clock" for clock-
> >     driven selection of initial sequence numbers (MUST-8), and SHOULD
> >     generate its Initial Sequence Numbers with the expression:
> >
> >     ISN = M + F(localip, localport, remoteip, remoteport, secretkey)
> >
> > and:
> >
> >           +  RFC 5961 [37] section 5 describes a potential blind data
> >              injection attack, and mitigation that implementations MAY
> >              choose to include (MAY-12).  TCP stacks that implement
> >              RFC 5961 MUST add an input check that the ACK value is
> >              [...]
> >
> > What prevents us from making a MUST-level requirement for randomized
> > ISNs?  Is it just the fact that it was only a SHOULD in RFC 6528 and a
> > perception that promoting to a MUST would be incompatible with retaining
> > Internet Standard status?
> >
> > Likewise, what prevents using stronger normative language (e.g., MUST)
> > for the RFC 5961 protections?
> >
> > It seems to me that these mechanisms are of general applicability and
> > provide significant value for use of TCP on the internet, even though
> > they are not fully robust and do not use cryptographic mechanisms.  If
> > there are scenarios where their use is harmful or even just not
> > applicable, that seems like an exceptional case that should get
> > documented so as to strengthen the general recommendation for the
> > non-exception cases.
> 
> On this question, I don't know if you (Ben) agreed, but the final 
> message I saw was from Joe:
> 
> https://mailarchive.ietf.org/arch/msg/tcpm/vDCi0xAf1Iayzsb4BS6DwlW3nhY/

That is the final message I see as well, where Joe included the statement
"Remember that one axiom of TCP is “think twice before changing TCP, then
don’t”, as a rule", which didn't really seem like a great sign that
continuing the conversation was likely to be productive, so I wanted to
wait and hear your thoughts as well before sending further replies.

> I think your question would be a good one to bring up in TCPM for future 
> work, but the working group was trying to avoid such changes in this 
> document.

I would like to see some stronger justification for pushing this change to
future work rather than incorporating it now.

In essence, I think that we require a fairly strong justification to
publish an Internet Standard in 2022 that says it's okay to adopt a data
model where a host has a global piece of state that it freely sends to
anyone who asks, where that piece of state can be used to attack/disrupt
all new connections that host makes, as opposed to just connections on the
5-tuple that asked.

The actual scope of utility of an ISN is local to an individual 5-tuple,
not global to a host, and false sharing of ISNs across connections adds
risk.  My stance as SEC AD is that the IETF should produce protocols that
are as secure as possible *subject to any constraints on them*.  Here, we
have the constraint of retaining compatibility with the existing deployment
base (which is a very compelling constraint!), so you do not see me
advocating for mandatory tcpcrypt (for example).  But I want to see some
explanation of what harm or risk there is in saying that the ISN is always
produced with the PRF of the 5-tuple (okay, 4-tuple since the protocol is
always TCP), to motivate a divergence from the more-secure behavior and
thereby justify retaining the behavior currently in the draft.  What
constraint are we subject to that prevents doing the right (security-wise)
thing?



> 
> >
> > (2) I think this is just a process question to ensure that the IESG
> > knows what we are approving at Internet Standard maturity, though it
> > is certainly possible that I misunderstand the situation.
> >
> > In Section 3.7.3 we see the normative statement (SHLD-6) that "when the
> > when the effective MTU of an interface varies packet-to- packet, TCP
> > implementations SHOULD use the smallest effective MTU of the interface
> > to calculate the value to advertise in the MSS option".  This seems to
> > originate in RFC 6691 (being obsoleted by this document), but RFC 6691
> > is only an Informational document and has not had an opportunity to
> > "accumulate experience at Proposed Standard before progressing", to
> > paraphrase RFC 6410.
> >
> > Similarly, Section 3.9.2 has (SHLD-23) "Generally, an application SHOULD
> > NOT change the DiffServ field value during the course of a connection
> > (SHLD-23)."  This is a bit harder to track down, as the DiffServ field
> > was not always known by that name.  I actually failed to find a directly
> > analogous previous statement of this guidance (presumably my error), and
> > thus don't know if it had any experience at the PS level or not.
> >
> > RFC 6410 seems pretty clear that some revisions are okay in Internet
> > Standards without such "bake time" at PS, but it does seem like
> > something that should be done consciously rather than by accident.
> 
> If I understood the thread earlier and results of the IESG telechat, 
> this point could be cleared?

Yes, this point can be considered resolved.

> 
> >
> > (3) This is also a process point for explicit consideration by the IESG.
> >
> > Appendix A.2 appears to discuss a few (rare) scenarios in which the
> > technical mechanisms of this document fail catastrophically (e.g.,
> > getting stuck in a SYN|ACK loop and failing to complete the handshake).
> > Does this meet the "resolved known design choices" and "no known
> > technical omission" bar required by RFC 2026 even for *proposed*
> > standard?
> >
> > (Note that RFC 2026 explicitly says that the IESG may waive this
> > requirement, at least for PS.)
> >
> >
> > (AFAICT one such scenario is reported at
> > https://www.rfc-editor.org/errata_search.php?eid=3305  , which the change
> > log for this document calls out as "not applicable due to other
> > changes"; I am not sure which "other changes" are intended, for this
> > case.)
> 
> If I understood the thread earlier and results of the IESG telechat, 
> this point could be cleared?

Yes, this point can be considered resolved.

> > (4) Another point mostly just to get explicit IESG acknowledgment
> > (elevating one of Lars' comments to DISCUSS level, essentially).
> >
> > As the changelog (and gen-art reviewer!) notes:
> >
> >     Early in the process of updating RFC 793, Scott Brim mentioned that
> >     this should include a PERPASS/privacy review.  This may be something
> >     for the chairs or AD to request during WGLC or IETF LC.
> >
> > I don't see any evidence to suggest that such a review actually
> > occurred.  Do we want to seek out such a targeted review before
> > progressing?
> 
> If I understood the thread earlier and results of the IESG telechat, 
> this point could be cleared?

Yes, this point can be considered resolved.

> > ----------------------------------------------------------------------
> > COMMENT:
> > ----------------------------------------------------------------------
> >
> > Thank you for the editorial changes so that we now talk about "a TCP
> > implementation" or a "remote TCP peer" rather than just "a TCP" or
> > "a remote TCP"!
> >
> > Abstract
> >
> >                                                   It also updates RFC
> >     5961 by adding a small clarification in reset handling while in the
> >     SYN-RECEIVED state.  [...]
> >
> > I'm not sure I found what this clarification was; is SYN-RECEIVED the
> > correct state?  The ad-hoc diff I constructed between RFC 793 and this
> > document shows identical text for the "If the RST bit is set" case when
> > currently in SYN-RECEIVED STATE.
> 
> The only trick is that RFC 5961 has an inconsistency in that its Section 
> 3.2 is first is talking about the change being applied in the 
> synchronized states, but SYN-RECEIVED is one of the states that it 
> applies to and is not technically a synchronized state. The 
> clarification is to just not use that term (synchronized state) in the 
> place where this logic is added.

Ah, that makes sense.  Thank you for checking and the explanation here.

> 
> >
> > Section 3.1
> >
> >     Options: [TCP Option]; Options#Size == (DOffset-5)*32; present
> >     only when DOffset > 5.
> >
> > My (later) nit-level notation comment aside, the given expression does
> > not seem to convey the size occupied by the options, but rather the
> > combined size of the options and the padding.
> 
> True; in the working copy I've added "note that this size expression 
> also includes any padding trailing the actual options present."  Does 
> that make sense for you?

Yes, that sounds good.

> 
> >
> > Section 3.2
> >
> >     A TCP Option is one of: an End of Option List Option, a No-Operation
> >     Option, or a Maximum Segment Size Option.
> >
> > The IANA registry lists some thirty-odd option kinds, so this sentence
> > just seems false without some additional qualifier ("defined by this
> > specification", etc.)
> 
> True, in the working copy, I changed it to "A TCP Option in the 
> mandatory option set".
> 
> 
> >
> > Section 3.4
> >
> >     In response to sending data the TCP endpoint will receive
> >     acknowledgments.  The following comparisons are needed to process the
> >     acknowledgments.
> >        [...]
> >        SEG.SEQ = first sequence number of a segment
> >
> >        SEG.LEN = the number of octets occupied by the data in the segment
> >        (counting SYN and FIN)
> >
> >        SEG.SEQ+SEG.LEN-1 = last sequence number of a segment
> >
> > It seems to me that this information from the incoming segment is not
> > part of processing the *acknowledgment*, but rather part of processing
> > the data received in that segment (a procedure discussed a few
> > paragraphs later).
> 
> I see what you mean, though these are used in validating any segments, 
> including pure acknowledgements.
> 
> 
> >
> >               This clock is a 32-bit counter that typically increments at
> >     least once every roughly 4 microseconds, [...]
> >     Maximum Segment Lifetime (MSL), generated ISNs will be unique, since
> >     it cycles approximately every 4.55 hours, which is much longer than
> >     the MSL.
> >
> > Once we put in the "at least" we allow arbitrarily faster clock updates,
> > and that puts the "approximately every 4.55 hours" estimate in question.
> > Very fast clock updates would cycle correspondingly faster.  Do we need
> > to place a lower limit on the clock update interval?  (On first look, it
> > seems like we might not, since the keyed PRF F() is providing most of
> > the protection from off-path guessing, and an attacker can always use
> > direct connections to estimate the clock cycle interval.  OTOH, if it
> > cycles so fast that it repeats within O(MSL), that might be problematic.)
> >
> >     parameters and some secret data.  For discussion of the selection of
> >     a specific hash algorithm and management of the secret key data,
> >     please see Section 3 of [41].
> >
> > The guidance in the referenced document seems a bit dated (it indicates
> > that MD5 is probably still okay for this purpose).  While the known
> > attacks on MD5 do not directly translate into an attack on ISN
> > generation, collisions can be found on as little as 64 bytes of input,
> > and all of the straightforward ways to use pure MD5 as a keyed hash for
> > this purpose have some undesirable properties.  I'm happy to note that
> > FreeBSD is using siphash for this purpose, which should be more than
> > adequate.  I expect that Linux and other major TCP stacks are already
> > doing something similar, so the guidance to use MD5 may be dated in
> > practice as well as in utility.
> >
> > I don't have a great proposal for where to put some updated guidance
> > (unless there's already some work underway in tcpm?); it is probably not
> > appropriate to put it inline here, so either an appendix or a separate
> > document seem plausible.
> 
> What you say makes sense.  Properly updating RFC 6528 is something we 
> can propose in TCPM based on your comments, which seems better to do in 
> a separate document.

Yes, that seems better placed in an update to 6528 than in this document.

> 
> >
> > Section 3.7.1
> >
> >     The MSS value to be sent in an MSS option should be equal to the
> >     effective MTU minus the fixed IP and TCP headers.  By ignoring both
> >     IP and TCP options when calculating the value for the MSS option, if
> >     there are any IP or TCP options to be sent in a packet, then the
> >     sender must decrease the size of the TCP data accordingly.  RFC 6691
> >     [42] discusses this in greater detail.
> >
> > I note that RFC 6691 is obsoleted by this document; it seems to me that
> > if we think there is useful content still in that document, we should
> > include such content in this document instead of referring to a document
> > we are calling obsolete.  (This is not the only place we do so, to be
> > clear, but I will try to mention it just once.  I do see the note that
> > we only claim to incorporate the normative portions of most of the
> > obsoleted specs, leaving the informational content alone.)
> 
> Understood, though referencing is very little work and allows someone 
> who is interested in more detail "why" to find it.  In contrast,  
> incorporating all this type of information into this document would be a 
> lot more work, and that content would not be needed or really even of 
> value for many readers.
> 
> 
> >
> > Section 3.8.4
> >
> >     An implementation SHOULD send a keep-alive segment with no data
> >     (SHLD-12); however, it MAY be configurable to send a keep-alive
> >     segment containing one garbage octet (MAY-6), for compatibility with
> >     erroneous TCP implementations.
> >
> > Such misbehaved TCP impelementations were misbehaved even in 1989 when
> > RFC 1122 was published -- do we have a sense for whether they are still
> > around to any significant degree?
> 
> That's a great question; I don't know, but would highly doubt they are 
> around anymore (at least not in production on the Internet).

Might be an interesting experiment, but it sounds like we don't have enough
data to justify changing the text here (yet).

> 
> >
> > Section 3.8.5
> >
> >     As a result of implementation differences and middlebox interactions,
> >     new applications SHOULD NOT employ the TCP urgent mechanism (SHLD-
> >     13).  However, TCP implementations MUST still include support for the
> >     urgent mechanism (MUST-30).  Details can be found in RFC 6093 [38].
> >
> > This "SHOULD NOT employ" has been in force for over a decade (RFC 6093
> > is dated January 2011).  How long do we have to wait until there are
> > sufficiently few implementations employing the urgent mechanism that it
> > no longer needs to be implemented?
> 
> When it was last looked at, there was at least one fairly prominent 
> application that relied on it, though they could do so semi-safely 
> because of being tied to an OS with the expected interpretation.
> 
> 
> >
> > Section 3.9.2.3
> >
> >        An incoming SYN with an invalid source address MUST be ignored
> >        either by TCP or by the IP layer (MUST-63) (Section 3.2.1.3 of
> >        [18]).
> >
> > Requirements of the form "A or B must do X" that are ambiguous about
> > whether A or B takes the action leave the risk that both will expect the
> > other party to take the action, and the action will fail to occur.  If
> > we're in a position to specifically require one (or both!) to check,
> > that leads to a more robust and verifiable system.  (I assume we're not
> > in such a position, but it can't hurt to check.)
> 
> Great point.  I guess we hope IP would discard it, but in case not, TCP 
> should.  If the implementation is handling both protocols at once (as in 
> some small hosts), there might not be much distinction.
> 
> 
> >
> > Section 4
> >
> >     Destination Address
> >             The network layer address of the remote endpoint.
> >     [...]
> >     Source Address
> >             The network layer address of the sending endpoint.
> >
> > These definitions don't seem to work in the context of a receiver
> > validating the TCP checksum, where the destination address is the local
> > endpoint's address and the source address is the remote endpoint's
> > address.  (I note that these definitions are different from what RFC 793
> > itself used.)
> 
> Interesting point on the destination (though the source one seems fine, 
> it's always the sending endpoint).  Maybe we should just change "remote" 
> to "receiving" in the destination one?

That's an easy fix; good idea.

> 
> >
> >     receive window
> >             This represents the sequence numbers the local (receiving)
> >             TCP endpoint is willing to receive.  Thus, the local TCP
> >             endpoint considers that segments overlapping the range
> >             RCV.NXT to RCV.NXT + RCV.WND - 1 carry acceptable data or
> >             control.  Segments containing sequence numbers entirely
> >             outside of this range are considered duplicates and
> >             discarded.
> >
> > Duplicates or injection attacks (when the sequence numbers in the
> > segment are too large).
> 
> ACK, I added "or injection attacks" in the working copy.
> 
> 
> >
> > Section 5
> >
> >     The collection of applicable RFC Errata that have been reported and
> >     either accepted or held for an update to RFC 793 were incorporated
> >     (Errata IDs: 573, 574, 700, 701, 1283, 1561, 1562, 1564, 1565, 1571,
> >     1572, 2296, 2297, 2298, 2748, 2749, 2934, 3213, 3300, 3301, 6222).
> >     Some errata were not applicable due to other changes (Errata IDs:
> >     572, 575, 1569, 3305, 3602).
> >
> > I think that EID 1565 belongs in the "not applicable due to other
> > changes" list, since the text it attempts to modify involves the
> > now-removed discussion of the IP "precedence" field.
> >
> > Similarly, EID 2296 also affected text about precedence and security
> > that is no longer present in a recognizable form.
> 
> Good points, I've moved those in the working copy.  Originally they were 
> fixed as described, but then other edits made them OBE.

I figured as much, yeah.

> 
> >
> >     The more secure Initial Sequence Number generation algorithm from RFC
> >     6528 was incorporated.  See RFC 6528 for discussion of the attacks
> >     that this mitigates, as well as advice on selecting PRF algorithms
> >     and managing secret key data.
> >
> > (As I mentioned up in §3.4, that guidance is no longer current.)
> 
> Is updating 6528 agreed to be possible future work?

Yes.  (Is this something where I should be expecting to take the action
item to write a draft, or do you think someone else might take up the pen?)

> 
> >
> > Section 9.1
> >
> > It's not clear to me that RFC 2675 ([5]) needs to be classified as
> > normative.
> 
> Yeah, I don't have any concern with making it informative, if you think 
> that's right?

>From looking at where we reference it (it's [25] now in the -27), I think
it looks informative.  I don't remember any other places where we talked
about jumbograms without a reference, which would be the only real concern.

> 
> >
> > The guidance at
> > https://www.ietf.org/about/groups/iesg/statements/normative-informative-references/
> > would suggest that RFC 5961 ([37]) should be classified as normative,
> > since we replicate its MUST-level requirements with the condition that
> > "TCP stacks that implement RFC 5961 MUST [...]", which would appear to
> > make that behavior an "optional feature".
> 
> Ok; I've made it so in the working copy.
> 
> 
> > Appendix A.1.2
> >
> >     The IP security option (IPSO) and compartment defined in [1] was
> >     refined in RFC 1038 that was later obsoleted by RFC 1108.  The
> >     Commercial IP Security Option (CIPSO) is defined in FIPS-188, and is
> >     supported by some vendors and operating systems.  RFC 1108 is now
> >
> > Should we mention that FIPS-188 is archived and withdrawn by NIST?
> > (I also didn't find much to define the actual IP option in the PDF I
> > found,
> > https://csrc.nist.gov/csrc/media/publications/fips/188/archive/1994-09-06/documents/fips188.pdf,
> > but I didn't look very hard.)
> 
> In the working copy, I've added "(withdrawn by NIST in 2015)" after the 
> mention of FIPS-188.
> 
> 
> > Appendix A.3
> >
> > It's fascinating to me that the preferred reference for this modified
> > Nagle algorithm is an Internet-Draft from 1999, vs something more
> > recent.
> 
> Me too!
> 
> 
> > Appendix B
> >
> >       Every 2nd full-sized segment or 2*RMSS ACK'd | SHLD-19|x| | | | |
> >
> > This 'x' seems to be in the "MUST" column, not the "SHOULD" column.
> 
> Thank you!  Fixed in working copy.
> 
> 
> >
> >     Time Stamp support                             | MAY-10 | | |x| | |
> >
> > How do we square timestamp support being a "MAY" with SHLD-4,
> > SHOULD-level guidance to use timestamps to reduce TIME-WAIT?
> 
> MAY-10 is the ***IP*** Timestamp option, and SHLD-4 is about the 
> ***TCP*** Timestamp option.

Oops, sorry for missing that distinction.

> 
> >
> >     Time Exceeded => tell ALP, don't abort         | MUST-56| | | | |x|
> >     Param Problem => tell ALP, don't abort         | MUST-56| | | | |x|
> >
> > Is there a double negative between "don't abort" and the 'x' being in
> > the "MUST NOT" column?
> 
> Agreed, that is confusing.  In the working copy, I'll change both of 
> those titles to "Abort on ..."
> 
> 
> > NITS
> >
> > I made essentially no attempt to de-duplicate the nit-level remarks
> > against the ballot positions from other ADs (in contrast to the other
> > comments, where I made some modest effort to de-duplicate).  My
> > apologies for the extra work to ignore the already-fixed items.
> >
> > Section 1
> >
> >     For several decades, RFC 793 plus a number of other documents have
> >     combined to serve as the core specification for TCP [48].  Over time,
> >     a number of errata have been filed against RFC 793, as well as
> >     deficiencies in security, performance, and many other aspects.  The
> >
> > A naive parse would say that this means a number of errata have been
> > filed against deficiencies.  I suspect the transition between errata and
> > deficiencies should refer to deficiencies having been "discovered" or
> > similar.
> 
> Good point.  In the working copy, I've broke that into 2 sentences, in 
> case you think this is more clear:
> 
>     Over time, a number of errata have been filed against RFC 793. There
>     have also been deficiencies found and resolved in security,
>     performance, and many other aspects.
> 
> 
> >     The purpose of this document is to bring together all of the IETF
> >     Standards Track changes that have been made to the base TCP
> >     functional specification and unify them into an update of RFC 793.
> >
> > It's a little surprising to see this described as an "update of RFC 793"
> > (vs. a "replacement of" or "updated version of") since the relationship
> > is Obsoletes, not Updates.  I might even consider "into a single
> > consolidated specification".
> 
> Ok, I changed to "updated version of" in the working draft.
> 
> 
> > Section 3.1
> >
> >     Options: [TCP Option]; Options#Size == (DOffset-5)*32; present
> >     only when DOffset > 5.
> >
> > The "Options#Size" notation seems confusing and is not using any
> > convention I'm aware of.  It does not appear in RFC 793 or any other RFC
> > that I can find, either.
> 
> This is changed now, as discussed in other threads.
> 
> 
> > Section 3.3.1
> >
> >     maintenance of a TCP connection requires the remembering of several
> >     variables.  We conceive of these variables being stored in a
> >
> > "the remembering of" is a fairly awkward phrase, where something like
> > just "remembering" or "maintaining state for" would flow more naturally.
> 
> Ok, changed to "maintaining state for" in the working copy.
> 
> 
> > Section 3.4
> >
> >     It is essential to remember that the actual sequence number space is
> >     finite, though very large.  This space ranges from 0 to 2**32 - 1.
> >
> > The sense of scale in the broader ecosystem may have evolved out from
> > under us; QUIC's 62-bit sequence space might be more along the lines of
> > "very large" these days, with a 32-bit space being merely "large".
> 
> Ok, removed "very" in the working copy.
> 
> 
> >
> >     A connection is defined by a pair of sockets.  Connections can be
> >
> > This is the first instance of the word "socket" in this document.
> > RFC 793 used the term much more prevalently, but this update has
> > (beneficially, IMO) moved away from that approach in favor of discussing
> > IP addresses and port numbers.  Might such a change be appropriate here
> > as well?  Regardless, we should probably have some introduction to what
> > we mean by "socket" if we are to retain any uses of the term, IMO, more
> > than just the glossary entry.
> 
> Good point.  I think it will help to borrow the sentences below from 793 
> and put them into the section 2.2 on "Key TCP Concepts":
> 
>      To allow for many processes within a single host to use TCP
>      communication facilities simultaneously, TCP provides a set of
>      addresses or ports within each host.  Concatenated with the network
>      and host addresses from the internet communication layer, this forms
>      a socket.

Thanks!

> >
> >     verify this SYN.  The three way handshake and the advantages of a
> >     clock-driven scheme are discussed in [68].
> >
> > I don't have access to the reference, but it's not clear from just it's
> > abstract whether "advantages of" or "advantages over" a clock-driven
> > scheme is the intended meaning.
> 
> We could clarify as "advantages of a clock-driven scheme for ISN 
> selection ...".  (I checked the paper and that's what it's referring to.)

Thanks for checking the reference.

> 
> >     explanation for this specification is given.  TCP implementors may
> >     violate the "quiet time" restriction, but only at the risk of causing
> >     some old data to be accepted as new or new data rejected as old
> >     duplicated by some receivers in the internet system.
> >
> > Maybe "old duplicated data"?  The current phrasing feels like it's
> > missing a word.
> 
> ACK, fixed in my working copy.
> 
> 
> >                                  Hosts that prefer to avoid waiting are
> >     willing to risk possible confusion of old and new packets at a given
> >     destination may choose not to wait for the "quiet time".
> >
> > I think this needs an "and", for "prefer to avoid waiting and are
> > willing to risk".
> 
> ACK, fixed in my working copy.
> 
> 
> >
> >     To summarize: every segment emitted occupies one or more sequence
> >     numbers in the sequence space, the numbers occupied by a segment are
> >     "busy" or "in use" until MSL seconds have passed, upon rebooting a
> >     block of space-time is occupied by the octets and SYN or FIN flags of
> >     the last emitted segment, if a new connection is started too soon and
> >     uses any of the sequence numbers in the space-time footprint of the
> >     last segment of the previous connection incarnation, there is a
> >     potential sequence number overlap area that could cause confusion at
> >     the receiver.
> >
> > This list seems to be missing an "and".
> > (Also, is it really only the last emitted segment that could cause
> > problems?)
> 
> ACK, fixed in my working copy.  You're right that it's not just the last 
> segment, but any that could potentially still be in-flight.
> 
> 
> >
> > Section 3.5
> >
> >                                                               It is the
> >     implementation of a trade-off between memory and messages to provide
> >     information for this checking.
> >
> > I'm not sure this reads well; is "the implementation of" needed?
> 
> Ok, removed in working copy.
> 
> 
> >
> >        If an incoming segment has a security level, or compartment that
> >        does not exactly match the level and compartment requested for the
> >        connection, a reset is sent and the connection goes to the CLOSED
> >        state.  The reset takes its sequence number from the ACK field of
> >
> > The comma in the first line is no longer needed (it was part of the list
> > when precedence was still part of the list).
> 
> ACK, fixed in working copy.
> 
> 
> > Section 3.6
> >
> >        In this case, a FIN segment can be constructed and placed on the
> >        outgoing segment queue.  No further SENDs from the user will be
> >        accepted by the TCP implementation, and it enters the FIN-WAIT-1
> >        state.  RECEIVEs are allowed in this state.  All segments
> >        preceding and including FIN will be retransmitted until
> >        acknowledged.  When the other TCP peer has both acknowledged the
> >        FIN and sent a FIN of its own, the first TCP peer can ACK this
> >        FIN.  Note that a TCP endpoint receiving a FIN will ACK but not
> >        send its own FIN until its user has CLOSED the connection also.
> >
> > Naming the two peers (e.g., A and B) can help avoid awkward grammatical
> > constructions like "can ACK this FIN" and improve clarity.
> 
> Unless you really think it's necessary, I prefer not to change this 
> original 793 text, since it doesn't seem so bad to me.
> 
> 
> > Section 3.8
> >
> >     segments may arrive due to network or TCP retransmission.  As
> >     discussed in the section on sequence numbers the TCP implementation
> >     performs certain tests on the sequence and acknowledgment numbers in
> >     the segments to verify their acceptability.
> >
> > comma after "sequence number".
> 
> ACK, fixed in working copy.
> 
> 
> >
> > Section 3.8.6.2.2
> >
> >     Note that the general effect of this algorithm is to advance RCV.WND
> >     in increments of Eff.snd.MSS (for realistic receive buffers:
> >     Eff.snd.MSS < RCV.BUFF/2).  Note also that the receiver must use its
> >     own Eff.snd.MSS, assuming it is the same as the sender's.
> >
> > I think the last sentence would be more clear if it was something like
> > "making the assumption that is the same" or "on the assumption that it
> > is the same".
> 
> ACK, changed to your suggested "making the assumption" in my working copy.
> 
> 
> > Section 3.8.6.3
> >
> >     Note that there are several current practices that further lead to a
> >     reduced number of ACKs, including generic receive offload (GRO), ACK
> >     compression, and ACK decimation [26].
> >
> > Reference [26] seems reasonable for ACK decimation and ACK compression,
> > but doesn't seem to cover GRO at all.
> 
> True; do you think we should try to find a proper GRO reference? I'm not 
> aware of there being a real standard or academic type of reference.

I wouldn't put much effort into looking.
It's a bit outside my normal field, so I don't have any immediate insight.
Google finds the DPDK docs, which might be stable and well-known, and
wikipedia's https://en.wikipedia.org/wiki/TCP_offload_engine entry has a
brief mention, with links to an LWN article and
https://books.google.com/books?id=C3wQBwAAQBAJ .  If we know someone with
a copy of the latter, it might be worth checking if it would be a useful
reference.

> 
> >
> > Section 3.9.1
> >
> >           If the PUSH flag is set, the application intends the data to be
> >           transmitted promptly to the receiver, and the PUSH bit will be
> >           set in the last TCP segment created from the buffer.  When an
> >           application issues a series of SEND calls without setting the
> >           PUSH flag, the TCP implementation MAY aggregate the data
> >           internally without sending it (MAY-16).
> >
> > There's a dedicated paragraph a few paragraphs later for when the PUSH
> > flag is not set; the last sentence might flow better there.
> 
> ACK.  Done in my working copy.
> 
> 
> >           Some TCP implementations have included a FLUSH call, which will
> >           empty the TCP send queue of any data that the user has issued
> >           SEND calls but is still to the right of the current send
> >           window.  That is, it flushes as much queued send data as
> >
> > I think "has issued SEND calls for" (add "for").
> 
> ACK, fixed in my working copy.
> 
> 
> > Section 3.9.2
> >
> >     When received options are passed up to TCP from the IP layer, TCP
> >     implementations MUST ignore options that it does not understand
> >     (MUST-50).
> >
> > singular/plural mismatch (it/implementations)
> 
> ACK, fixed in working copy.
> 
> 
> >
> > Section 3.9.2.2
> >
> >     Soft Errors
> >       For ICMP these include: Destination Unreachable -- codes 0, 1, 5,
> >       Time Exceeded -- codes 0, 1, and Parameter Problem.
> >
> >       For ICMPv6 these include: Destination Unreachable -- codes 0 and 3,
> >       Time Exceeded -- codes 0, 1, and Parameter Problem -- codes 0, 1,
> >       2.
> >
> >       Since these Unreachable messages indicate soft error conditions,
> >
> > I'm not entirely sure that I'd classify "parameter problem" as an
> > "unreachable" message per se.
> 
> I'm not sure I understand this comment.  This is the list from RFC 5461.

It looks like I was misparsing on first read, since there were commas
serving different roles.  ("Parameter Problem" is an element in the outer
list, matched with "Destination Unreachable" and "Time Exceeded", not one
of the codes for Time Exceeded.)  Using semicolons as delimiters for the
"outer" list, with commas for the lists of type codes, would help
distinguish them.

> 
> > Section 3.10
> >
> >     Please note in the following that all arithmetic on sequence numbers,
> >     acknowledgment numbers, windows, et cetera, is modulo 2**32 the size
> >     of the sequence number space.  Also note that "=<" means less than or
> >
> > Some punctuation around "the size of the sequence number space" seems in
> > order.
> 
> ACK, added parenthesis in working copy.
> 
> 
> >     equal to (modulo 2**32).
> >
> > [In formal mathematics this "less than or equal to, modulo N" operator
> > is not defined.  But it's probably okay in this context.]
> 
> Ok.
> 
> 
> > Section 3.10.1
> >
> >           the parameters of the incoming SYN segment.  Verify the
> >           security and DiffServ value requested are allowed for this
> >           user, if not return "error: precedence not allowed" or "error:
> >           security/compartment not allowed."  If passive enter the LISTEN
> >
> > It's surprising for the error string to mention "precedence" when the
> > predicate is DiffServ value.
> 
> This is a good point.  When the definition of those bits was changed, it 
> doesn't seem like the error message indicated here was correspondingly 
> changed.

Is this something we need to leave alone for backwards compatibility, or do
we have leeway to change the error message here?  (It looks like this text
is unchanged in the -27.)

> 
> >
> >           with "error: insufficient resources".  If Foreign socket was
> >           not specified, then return "error: remote socket unspecified".
> >
> > I suspect s/Foreign/remote/ was intended.  (Also occurs later, but I
> > will just note it once here.)
> 
> ACK, changed in my working copy.  Based on other comments, "foreign" in 
> general was changed to "remote".
> 
> 
> > Section 3.10.3
> >
> >        -  Since the remote side has already sent FIN, RECEIVEs must be
> >           satisfied by data already on hand, but not yet delivered to the
> >           user.  If no text is awaiting delivery, the RECEIVE will get a
> >           "error: connection closing" response.  Otherwise, any remaining
> >           text can be used to satisfy the RECEIVE.
> >
> > I think s/text/data/ should be applied on the last line (since it was
> > already applied on the second line).
> 
> ACK, changed in my working copy.
> 
> 
> >
> > Section 3.10.7.4
> >
> >     o  Segments are processed in sequence.  Initial tests on
> >        arrival are used to discard old duplicates, but further
> >        processing is done in SEG.SEQ order.  If a segment's
> >        contents straddle the boundary between old and new, only the
> >        new parts should be processed.
> >
> > Maybe s/should be/are/?  There's not really optionality about it...
> 
> ACK, changed in my working copy.
> 
> 
> >              *  If this connection was initiated with a passive OPEN
> >                 (i.e., came from the LISTEN state), then return this
> >                 connection to LISTEN state and return.  The user need
> >                 not be informed.  If this connection was initiated
> >                 with an active OPEN (i.e., came from SYN-SENT state)
> >                 then the connection was refused, signal the user
> >                 "connection refused".  In either case, all segments on
> >                 the retransmission queue should be removed.  And in
> >
> > IIUC, what's described here as "removed" is described elsewhere as
> > "flushed"; it would be good to use consistent terminology when possible.
> 
> Do you have a strong preference?

No strong preference, just a general desire to have consistent terminology
unless there is a difference in meaning intended.

> 
> >           +  Once in the ESTABLISHED state, it is possible to deliver
> >              segment text to user RECEIVE buffers.  Text from segments
> >              can be moved into buffers until either the buffer is full
> >              or the segment is empty.  If the segment empties and
> >              [...]
> >
> > As above, it seems like (case-insensitive) s/test/data/ would improve
> > consistency.
> 
> Changed to "data" in working copy.
> 
> 
> >
> > Section 4
> >
> >     internet datagram
> >             The unit of data exchanged between an internet module and the
> >             higher level protocol together with the internet header.
> >
> > "exchanged between an internet module and the higher level protocol"
> > sounds like a local operation; I would have expected the definition of
> > an *internet* datagram to involve transfer over the (inter)network.
> 
> Ok, can you suggest an alternative definition?  I agree this one from 
> the original 793 isn't terrific.

I'm somewhat reluctant to concoct a new definition from scratch at this
point in the process.  I see that RFC 1983 has "IP datagram" refer to just
"datagram", on which it says:

   datagram
      A self-contained, independent entity of data carrying sufficient
      information to be routed from the source to the destination
      computer without reliance on earlier exchanges between this source
      and destination computer and the transporting network.  See also:
      frame, packet.
      [Source: J. Postel]

If I was to overcome my reluctance and try to revise the existing text here
into something that makes more sense, it would be:

internet datagram
  A unit of data [from some higher layer protocol] exchanged between
  internet hosts, together with the internet header that allows the
  datagram to be routed from source to destination.

(I'm not sure whether I'd include the bit in square brackets.)
Feel free to use mine if you like, or go with Postel's (per RFC 1983) or
something else.

Many thanks again for all your tiresome work on this document.  The end is
in sight!

-Ben

> 
> >
> >     segment length
> >             The amount of sequence number space occupied by a segment,
> >             including any controls that occupy sequence space.
> >
> > Should we say that this is a field in the segment header?
> 
> I don't think so (it's not in the TCP header).
> 
> 
> >
> >     URG
> >             A control bit (urgent), occupying no sequence space, used to
> >             indicate that the receiving user should be notified to do
> >             urgent processing as long as there is data to be consumed
> >             with sequence numbers less than the value indicated in the
> >             urgent pointer.
> >
> > To me, "value indicated in" is synonymous with "value contained in",
> > which is problematic here since the urgent field is only 16 bits and
> > sequence numbers 32 bits.  "indicated by" would be an improvement,
> > though of course if we're willing to spend more words we can increase
> > clarity further.
> ACK, changed to "indicated by" in my working copy.
> 
> >
> > Appendix A.1
> >
> >     RFC 793 requires checking the IP security compartment and precedence
> >     on incoming TCP segments for consistency within a connection, and
> >
> > I think the past tense "required" would be more appropriate upon
> > publication of this document as an RFC obsoleting RFC 793.
> >
> ACK, fixed in my working copy.