TEAP Chaining and Partial Success Policies

[Jan Kříž]

Hi Alan,

Thank you very much for addressing this.

To give a bit more context on why I went down this rabbit hole: I was
researching options for handling "partial failures", specifically the
"chicken and egg" scenario where a user certificate is missing on a
freshly provisioned device. That led me to a "Wires and Wi-Fi" blog
post describing how TEAP is based on EAP-FASTv2 and how EAP chaining
allows the server to make decisions based on combined states,
including "User Failed and Machine Succeeded"
(https://www.wiresandwi.fi/blog/windows-network-authentication-sequence).

According to that explanation, Windows can perform machine-only
authentication during startup, even when the user certificate does not
exist yet, and the RADIUS server can return an Access-Accept with a
restricted authorization profile. This gives the device enough limited
network access to reach the PKI and enroll the missing user
certificate.

I also found a Cisco ISE guide showing this workflow in practice, with
explicit rules for "Machine authenticated, User not" that grant
reduced access, distinct from the full "User and Machine" success
state. (https://www.ise-support.com/2020/05/29/using-teap-for-eap-chaining/).

Based on these examples, it appears that several commercial platforms
intentionally treat the session as successful overall (sending
Access-Accept), even when one inner method fails, and rely on the
authorization layer to map that combined state to the appropriate
access level. This also implies that the Windows TEAP supplicant is
willing to accept a TEAP Success result even when an inner EAP
conversation completed with a failure.

> You will have to update the "eap_teap.c" code to allow TEAP to continue when one of the inner methods fails.  Right now, an inner reject results in the entire TEAP session failing.
> You can't edit the configuration to allow this.  When the TEAP code was written, I was assuming that failure means failure.  If the Windows TEAP implementation is OK with the above work flow, then that's a surprising bit of information.

Given that this "partial success with restricted access" behavior
seems to have become a de facto standard method for handling similar
scenarios, I wanted to ask whether FreeRADIUS would consider
supporting this mode?. Or is the project's stance that FreeRADIUS
should follow the RFC strictly until the specification explicitly
allows this behavior?

Best regards, Jan K.

On Tue, Dec 9, 2025 at 5:14 PM Alan DeKok via Freeradius-Users
<freeradius-users at lists.freeradius.org> wrote:
>
> On Dec 9, 2025, at 11:06 AM, Jan Kříž <jan.kriz1867 at gmail.com> wrote:
> > I'm following up on a thread from with the subject of "EAP-TEAP not
> > doing 2nd inner Method" from December 2024
> > (https://lists.freeradius.org/pipermail/freeradius-users/2024-December/105155.html)
> > about TEAP and setting network policies based on partial success
> > (e.g., Machine cert succeeds, but User cert fails).
>
>   I really don't see how that is supported by anything.  I've spent about 3 years going over the TEAP specifications, and the TEAP implementations.  I don't see anything in them which allows that work flow.
>
>   The inner authentications are bound to the outer TLS session via various crypto magic.  The result is that the outer authentication succeeds only when the inner authentication succeeds.
>
>   So I don't see any way in the specification which could allow machine EAP-TLS to succeed, fail at user EAP-TLS, but still allow the user online.
>
>   I could very well be missing something.  But if I am, then this behavior is not explicitly allowed by the specification.  So it's an accidental outcome.  The specification needs to be updated to mention this, and then either explicitly allow it, or forbid it.
>
> > I agree that the client side is a huge hurdle here, but I’m confused
> > because some commercial platforms like Cisco ISE and Aruba ClearPass
> > explicitly advertise and allow admins to configure different
> > VLANs/ACLs precisely for that partial success scenario.
> >
> > Since the commercial vendors seem to have found a way to achieve this,
> > are they relying on some vendor-specific extensions or a different,
> > non-TEAP chaining method entirely? Is it possible to configure this in
> > FreeRADIUS at all?
>
>   You'll have to try it and see.  I don't have time right now to test it, unfortunately.
>
> > Any insight into how FreeRADIUS could be configured to accurately
> > process the final result and differentiate between Machine-only,
> > User-only success and Full success would be hugely helpful.
>
>   You will have to update the "eap_teap.c" code to allow TEAP to continue when one of the inner methods fails.  Right now, an inner reject results in the entire TEAP session failing.
>
>   You can't edit the configuration to allow this.  When the TEAP code was written, I was assuming that failure means failure.  If the Windows TEAP implementation is OK with the above work flow, then that's a surprising bit of information.
>
>   Alan DeKok.
>
> -
> List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html