Re: [TLS] Decryption_failed alert in TLS 1.1

[Eric Rescorla <ekr at networkresonance.com>]

At Thu, 30 Apr 2009 11:41:45 +0200,
<Pasi.Eronen at nokia.com> wrote:
> 
> Hi,
> 
> I'm going through unverified errata for various SEC area RFCs,
> and came to errata ID 117 for RFC 4346 (TLS 1.1), available
> from here:
> 
> http://www.rfc-editor.org/errata_search.php?rfc=4346
> 
> Currently, Section 7.2.2 of RFC 4346 says:
> 
>    bad_record_mac
>       This alert is returned if a record is received with an incorrect
>       MAC.  This alert also MUST be returned if an alert is sent because
>       a TLSCiphertext decrypted in an invalid way: either it wasn't an
>       even multiple of the block length, or its padding values, when
>       checked, weren't correct.  This message is always fatal.
> 
> and then continues:
> 
>    decryption_failed
>       This alert MAY be returned if a TLSCiphertext decrypted in an
>       invalid way: either it wasn't an even multiple of the block
>       length, or its padding values, when checked, weren't correct.
>       This message is always fatal.
> 
>    Note: Differentiating between bad_record_mac and decryption_failed
>          alerts may permit certain attacks against CBC mode as used in
>          TLS [CBCATT].  It is preferable to uniformly use the
>          bad_record_mac alert to hide the specific type of the error.
> 
> These two contradict each other; it first says "bad_record_mac MUST be
> sent", but then says decryption_failed "MAY be returned", and using
> bad_record_mac is "preferable".
> 
> Does anyone recall what the intent here was? Early drafts (until -09) 
> said "bad_record_mac SHOULD be returned", but that was later changed 
> to "MUST". Should the errata fix this to something like this?
>
>    decryption_failed
>       This alert was used in TLS 1.0 if a TLSCiphertext decrypted
>       in an invalid way. It MUST NOT be sent in TLS 1.1.
> 
> (TLS 1.2 is very clear about this; decryption_failed MUST NOT
> be sent.)

Yeah, I think we just did s/SHOULD/MUST/ and didn't remove the
corresponding MAY

-Ekr