Re: [TLS] Spec issue with RFC 7627 (EMS) and resumption

[David Benjamin <[e-mail]>]

At least for an erratum, I don't think it makes sense to change that as
part of this.

I think your question is conflating a few things. Let me try to untangle
this, as this document is little confusing. It seems to be describing, via
SHOULDs and MUSTs, three different implementation profiles concurrently:
1. A client or server that negotiates EMS when available, but continues to
be compatible with legacy peers, resumption and all.
2. A client or server that negotiates EMS when available, continues to be
compatible with legacy peers, but only allows full handshakes with them. I
forget how 3SHAKE worked, but I vaguely recall that cutting out legacy
resumptions avoided a source of issues
3. A client or server that requires EMS on all connections, and is not
compatible with legacy peers at all. This way you can rely on the EMS
bugfix being applied.

Because these are all interleaved, it's a very hard to tell what bits
correspond to this profile and what bits correspond to EMS itself. I think
this was overambitious. Anything but (1) was implausible for most
deployments in 2016, because EMS only just started existing. I expect it is
still implausible for most, though I don't have metrics for the web. Hence,
if we were completely redoing this document (probably not worth it?), I
think it should be reorganized. :-)

Now, as to your question, I don't believe this is right:

> If the client follows this guide, it falls-back to use a full handshake.
> If the client doesn't follow this (maybe, the client is not aware of RFC
> 7627), the server SHOULD aborts.

The client SHOULD you are referring to is for a client that implements (2).
If a client does not follow this, it implements (1). The server SHOULD,
however, says "If neither the original session *nor the new ClientHello*
uses the extension, the server SHOULD abort the handshake." That clause
doesn't apply to the client SHOULD in the first place because, whether the
client is (1) or (2), the ClientHello will signal EMS.

It does apply to legacy clients that do not implement EMS at all. Whether
the server continues, does a full handshake, or aborts depends on whether
the server wants to do (1), (2), or (3). Yes, if it picks (3) and aborts,
the server will not interoperate with a legacy client. That is exactly what
those servers would want. It is odd that the text only describes (1) and
(3) for the server in this case, not (2). But adding (2) in without a
rework will only make this complicated situation worse, so I do not propose
to address it as part of this fix.

On Tue, Oct 26, 2021 at 1:05 AM Achim Kraus <[e-mail]> wrote:

> Hi David,
>
> if you're on it, maybe it's worth to consider my question from January
> 2021 as well.
>
>  > If the client follows this guide, it falls-back to use a full handshake.
>  > If the client doesn't follow this (maybe, the client is not aware of RFC
> 7627), the server SHOULD aborts.
>
>  > Why SHOULD the server not (also) just fall-back to use a full handshake?
>
> For more details see:
>
> https://mailarchive.ietf.org/arch/msg/tls/gjBFHWwp1k-w1KdBkotp496zaf8/
>
> best regards
> Achim Kraus
>
> Am 26.10.21 um 00:51 schrieb David Benjamin:
> > Here's some possible replacement text for that paragraph:
> >
> > """
> > In some deployments, a legacy client or server may be exposed to a
> > session using extended master secret. For example, a group of servers
> > sharing a ticket encryption key may be in the process of enabling this
> > extension. If such a session is used in an abbreviated handshake without
> > the extension, a newer peer will fail the connection, as described in
> > Section 5.3. To avoid this, legacy servers SHOULD ignore such sessions
> > and continue with a full handshake, and legacy clients SHOULD NOT offer
> > such sessions in the ClientHello.
> >
> > This can be implemented by ensuring legacy implementations do not
> > recognize sessions using extended master secret. For example, the
> > sessions may have a higher internal version number or, if the older
> > implementation rejects unrecognized fields, include a new field. If this
> > is not possible, deployments may deploy a new session cache or ticket
> > encryption key alongside the new version.
> > """
> >
> > Unfortunately, "session using extended master secret" is a bit of a
> > mouthful, but section 5.4 didn't define a more concise term.
> >
> > On Mon, Oct 25, 2021 at 4:01 PM David Benjamin <[e-mail]
> > <mailto:[e-mail]>> wrote:
> >
> >     Hi all,
> >
> >     In diagnosing an interop issue, I noticed RFC 7627 did not describe
> >     the correct server behavior for EMS very well. Seemingly as a
> >     result, some server implementation has gotten this wrong. I'd like
> >     to fix this in the spec so this doesn't happen again. I think, at
> >     minimum, we need to replace the last paragraph of section 5.4.
> >
> >     The issue is a server that /doesn't/ implement EMS, when presented a
> >     ClientHello containing a ticket or session ID by a server that
> >     /did/ implement EMS, must ignore the session and continue with a
> >     full handshake. Failing to do so will trip the client check in
> >     Section 5.3, "If a client receives a ServerHello that accepts an
> >     abbreviated handshake, [...]". This is important to meet these three
> >     properties:
> >
> >     - If the client and server both support EMS, the connection must
> >     negotiate it.
> >     - On resumption, the EMS status of the connection must match the EMS
> >     status of the session
> >     - In order for EMS to be safely deployable, it must be possible to
> >     roll EMS out gradually, or roll it back, without breaking
> >     connections. This means a mixed pre-EMS and post-EMS server
> >     deployment must work.
> >
> >     Note that, although this behavior is only visible at the pre-EMS
> >     server (not directly in scope for this document), it is actually a
> >     requirement on the post-EMS server. When the post-EMS server issues
> >     a session, it must arrange for the pre-EMS server to ignore it. For
> >     example, if the pre-EMS server rejects sessions with unparsable
> >     fields (the safest option), the post-EMS server can add a new field
> >     to the session state serialization. Failing that, it can bump some
> >     internal version number. Another strategy is to rotate session
> >     ticket keys alongside the version, but this can be tricky the way
> >     deployments and software updates are often split.
> >
> >     There's an analogous, though less likely, client scenario that a
> >     pre-EMS client must not offer a post-EMS session. Otherwise it will
> >     run afoul of a server requirement. This can be relevant for clients
> >     that serialize their session cache.
> >
> >     As far as I can tell, RFC 7627 does not specify any of this. The
> >     first paragraph of section 5.4 talks about adding a flag, but
> >     doesn't talk about how pre-EMS servers interact with that flag. The
> >     last paragraph discusses this scenario, but says something very
> >     strange, if not plain wrong:
> >
> >         If the original session uses an extended master secret but the
> >         ClientHello or ServerHello in the abbreviated handshake does not
> >         include the extension, it MAY be safe to continue the abbreviated
> >         handshake since it is protected by the extended master secret of
> the
> >         original session.  This scenario may occur, for example, when a
> >         server that implements this extension establishes a session but
> the
> >         session is subsequently resumed at a different server that does
> not
> >         support the extension.  Since such situations are unusual and
> likely
> >         to be the result of transient or inadvertent misconfigurations,
> this
> >         document recommends that the client and server MUST abort such
> >         handshakes.
> >
> >     https://datatracker.ietf.org/doc/html/rfc7627#section-5.4
> >     <https://datatracker.ietf.org/doc/html/rfc7627#section-5.4>
> >
> >     First, the "MAY" is immediately contradicted by the following
> >     "MUST", and by section 5.3. It seems it should have been an English
> >     lowercase "may", not a normative RFC 2119 "MAY". It is also wrong in
> >     calling this situation "unusual and likely to be the result of
> >     transient or inadvertent misconfigurations". Rather, it is the
> >     natural transition state of any large server rollout. I think we
> >     need to delete that entire paragraph and replace it with text that
> >     describes the rules above. If we were doing a whole new version of
> >     the document, I think the text could do with reorganization. But
> >     that may not be worth doing, given folks should be using TLS 1.3 now.
> >
> >     Thoughts? I can put together some replacement text if folks agree.
> >     What would be the best way to do this? Just an erratum?
> >
> >     David
> >
> >
> > _______________________________________________
> > TLS mailing list
> > [e-mail]
> > https://www.ietf.org/mailman/listinfo/tls
> >
>
>