Re: [TLS] RFC-4366-bis and the unrecognized_name(112) alert

[Michael D'Errico <[e-mail]>]

I agree with Yngve that a server should send either an empty SNI
extension OR an unrecognized_name alert but not both.  However, I
disagree that the server SHOULD NOT send a warning alert since that
hides information from the client.

The possible cases are:

     1 - server understood the SNI and used it to select an appropriate
         certificate chain and other parameters
     2 - server understood the SNI but did not recognize it as one of
         its configured virtual hosts; however, the server is set up
         to use a default configuration in that case
     3 - server understood the SNI but did not recognize it as one of
         its configured virtual hosts; there is no default configuration
         available so the handshake can not continue
     4 - server does not understand the SNI extension

The way my server reacts to each of these cases is:

     1 - add an empty SNI extension to ServerHello
     2 - send a warning unrecognized_name alert
     3 - send a fatal unrecognized_name alert
     4 - send nothing

Yngve would prefer that nothing is sent in case 2, but then a client
can not distinguish it from case 4.

Mike



Yngve Nysaeter Pettersen wrote:
> Hi,
> 
> Actual text depends on what one want the extension and the alert to do.
> 
> Relevant text:
> 
>    A server that receives a client hello containing the "server_name"
>    extension MAY use the information contained in the extension to guide
>    its selection of an appropriate certificate to return to the client,
>    and/or other aspects of security policy.  In this event, the server
>    SHALL include an extension of type "server_name" in the (extended)
>    server hello.  The "extension_data" field of this extension SHALL be
>    empty.
> 
>    If the server understood the client hello extension but does not
>    recognize the server name, it SHOULD send an "unrecognized_name"
>    alert (which MAY be fatal).
> 
> What needs to be specified is what a warning with this alert should mean 
> to the process of establishing the connection, and how it should be used 
> in combination with the returned extension. IMO only one of them may be 
> returned.
> 
> My preferred option is that the alert is only sent when the server does 
> not want to continue the negotiation, that is, a fatal alert, and not 
> return a server_name extension in case it did not know the hostname, but 
> is willing to continue.
> 
> If that is the route we go then the last paragraph above could be 
> changed to
> 
>    If the server understood the client hello extension but does not
>    recognize the server name, and is not willing to establish a connection,
>    it MUST send a fatal "unrecognized_name" alert. If it is willing to
>    establish the connection it MUST NOT return a "server_name" extension,
>    and SHOULD NOT send a "unrecognized_name" warning.
> 
> One can also add a note that some current implementations will send a 
> warning _and_ include the "server_name" extension (that is the current 
> status on the server I observed with the configuration issue), and that 
> some clients escalate the warning to Fatal.
> 
> 
> On Thu, 03 Jun 2010 16:02:43 +0200, Joseph Salowey (jsalowey) 
> <[e-mail]> wrote:
> 
>> If we are going to make a change then we need some suggested text.
>>
>> Joe
>>
>>> -----Original Message-----
>>> From: [e-mail]
>>> [mailto:[e-mail]] On Behalf Of Nikos
>>> Mavrogiannopoulos
>>> Sent: Thursday, June 03, 2010 12:24 AM
>>> To: Joseph Salowey (jsalowey)
>>> Cc: Michael D'Errico; [e-mail]; [e-mail]
>>> Subject: Re: [TLS] RFC-4366-bis and the unrecognized_name(112) alert
>>>
>>> I believe that it is quite bad practice to delegate to application
>>> protocol. This is protocol stack signalling and shouldn't be
>>> application specific. Otherwise interoperability problems occur. A
>>> protocol spec should be clear on what it does.
>>>
>>> regards,
>>> Nikos
>>>
>>> On Thu, Jun 3, 2010 at 7:22 AM, Joseph Salowey (jsalowey)
>>> <[e-mail]> wrote:
>>> > It seems that the behavior is dependent upon the application.  I'm
>>> not
>>> > sure we would come up with useful text here.  In the interest of
>>> getting
>>> > this document published I suggest we leave it as is.
>>> >
>>> > Joe
>>> >
>>> >> -----Original Message-----
>>> >> From: [e-mail] [mailto:[e-mail]] On Behalf
>>> Of
>>> >> Michael D'Errico
>>> >> Sent: Tuesday, May 25, 2010 10:58 AM
>>> >> To: [e-mail]
>>> >> Cc: [e-mail]
>>> >> Subject: Re: [TLS] RFC-4366-bis and the unrecognized_name(112) alert
>>> >>
>>> >> Martin Rex wrote:
>>> >> > Michael D'Errico wrote:
>>> >> >> In my server code, the SNI is checked to see if there is a
>>> matching
>>> >> >> virtual host with that domain name.  If there is, then no alert
>>> is
>>> >> >> sent.  If there is no matching virtual host, then it checks
>>> whether
>>> >> >> there is a default virtual host set up.  If there is a default,
>>> > then
>>> >> >> an unrecognized_name alert is sent with the warning level.  When
>>> no
>>> >> >> default is configured, the alert sent is fatal since the
>>> handshake
>>> >> >> can not continue.
>>> >> >>
>>> >> >> The warning lets the client know that there was not a match, but
>>> >> that
>>> >> >> the server can still continue using its default.
>>> >> >
>>> >> >
>>> >> > While this behaviour appears quite sensible/plausible, it will
>>> lead
>>> >> > to the behaviour in the wild that Yngve is reporting.
>>> >> >
>>> >> > An application which
>>> >> > does not configure any SNI characteristics, is not using SNI, and
>>> >> > for these, the server TLS implementation should _NOT_ be sending
>>> >> > SNI mismatch TLS alerts unless the application explicitly requests
>>> >> so.
>>> >>
>>> >> My server code will not send an alert unless the client (a) sent an
>>> >> SNI and (b) that SNI did not map to a virtual host.
>>> >>
>>> >> So that should not cause a problem.
>>> >>
>>> >> It might be a good idea to clarify when sending an unrecognized_name
>>> >> alert would be appropriate and clarify that it can be just a warning
>>> >> and what that means.
>>> >>
>>> >> Mike