Re: [TLS] Connection ID in TLS

[Benjamin Kaduk <[e-mail]>]

The issue is that the "TLSCiphertext.length" is the same field as
the "uint16 length", so you are saying that this field has the value
of "cid_length plus itself", which is impossible in integers modulo
[a value larger than cid_length].  In the "formal" grammar, you'd
need to define a new field.

-Ben

On Tue, Mar 20, 2018 at 11:45:59PM +0000, John Mattsson wrote:
> Correct, I just copied pasted the length of the arrays, should be length = cid_length + encrypted_record.length.
> 
> The example was taken from draft-ietf-tls-tls13-27. If I understand correctly, It seems like the same circular definition is done there as well....
> 
> -----------------------------------------------------
>       struct {
>           ContentType opaque_type = application_data; /* 23 */
>           ProtocolVersion legacy_record_version = 0x0303; /* TLS v1.2 */
>           uint16 length;
>           opaque encrypted_record[TLSCiphertext.length];
>       } TLSCiphertext;
> -----------------------------------------------------
> 
> Shouldn't this be...
> 
> -----------------------------------------------------
>       struct {
>           ContentType opaque_type = application_data; /* 23 */
>           ProtocolVersion legacy_record_version = 0x0303; /* TLS v1.2 */
>           uint16 length;
>           opaque encrypted_record[encrypted_record.length];
>       } TLSCiphertext;
> -----------------------------------------------------
> 
> Does this mean the 
> On 2018-03-20, 19:29, "Stephen Checkoway" <[e-mail]> wrote:
> 
>     
>     
>     > On Mar 20, 2018, at 11:38, John Mattsson <[e-mail]> wrote:
>     > 
>     > I think Connection ID is an important enabler for end-to-end security with (D)TLS. There seems to be important use cases for connection ID in TLS as well, see https://www.ietf.org/mailman/listinfo/atlas. At the Monday afternoon TLS session, it was stated that Connection ID in TLS was unemployable in the wild due to middleboxes. Couldn't that be solved by placing the cid field after the length field? E.g.
>     > 
>     >   struct {
>     >      ContentType opaque_type = application_data; /* 23 */
>     >      ProtocolVersion legacy_record_version = 0x0303; /* TLS v1.2 */
>     >      uint16 length;
>     >      opaque cid[cid_length];               // New field
>     >      opaque encrypted_record[TLSCiphertext.length];
>     >   } TLSCiphertext;
>     > 
>     >   length  The sum of cid_length and TLSCiphertext.length
>     
>     This defines length in terms of itself since length is TLSCiphertext.length.
>     
>     -- 
>     Stephen Checkoway
>     
>     
>     
>     
>     
>     
> 
> _______________________________________________
> TLS mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/tls