[keycloak-user] Unable to get SAML ForceAuthn to work

[Neil Russell]

Hi John, 

No worries. Can you (or anyone else) confirm if Keycloak supports ForceAuthn when acting as the identity provider? I've applied a fix locally that appears to be handling the 1 correctly but after a bit more digging it doesn't look like AuthnRequestType.IsForceAuthn() is referenced during the processing of a login request. 

Thanks,
Neil

On 8/29/19 5:25 PM, Neil Russell wrote:
> Hi John,
> 
> The lexical space for a boolean in the document you referenced is defined as:
>     -An instance of a datatype that is defined as ·boolean· can have the following legal literals {true, false, 1, 0}.
> 
> That document seems to confirm that 1 or 0 is compliant.

Right you are, my bad. Thanks for the clarification.

> -----Original Message-----
> From: John Dennis <jdennis at redhat.com>
> Sent: Thursday, August 29, 2019 1:00 PM
> To: Neil Russell <nrussell at egbc.ca>; 'keycloak-user at lists.jboss.org' <keycloak-user at lists.jboss.org>
> Subject: Re: [keycloak-user] Unable to get SAML ForceAuthn to work
> 
> On 8/29/19 3:03 PM, Neil Russell wrote:
>> Hey,
>>
>> I'm trying to get ForceAuthn to work with a third party who is using Shibboleth but have been unable to get it to force re-authentication if I have an existing session. I've inspected the SAML request and ForceAuthn is being passed in the request, one issue is that Shibboleth passes ForceAuthn="1" instead of ForceAuthn="true" and the parser doesn't appear to handle that. I made a fix to the StaxParserUtil class to try and get it working but even though I can now see that parser is returning true when the ForceAuthn attribute is read I'm still not getting the expected behaviour and I'm not sure where to look next.
>>
>> Any suggestions would be appreciated, am I looking in completely the wrong place?
> 
> The ForceAuthn attribute is defined as an xsi:boolean. The XML schema
> (https://www.w3.org/TR/xmlschema-2/#boolean) defines a boolean as either "true" or "false", it's case sensitive, no other values are permitted.
> Sounds like the Shibboleth SP is non-compliant.
> 
> 
> --
> John Dennis
> 


-- 
John Dennis