On Thu, 16 Jul 2026 15:06:31 GMT, Artur Barashev <[email protected]> wrote:

>> Improve the definition of the jdk.tls.disabledAlgorithm security property to 
>> specify how components of the cipher suite can be disabled. This has never 
>> been clearly specified.
>> 
>> ---------
>> - [x] I confirm that I make this contribution in accordance with the 
>> [OpenJDK Interim AI Policy](https://openjdk.org/legal/ai).
>
> src/java.base/share/conf/security/java.security line 749:
> 
>> 747: #     the key exchange algorithm "ECDH_ECDSA" will disable all cipher 
>> suites
>> 748: #     that contain ECDH_ECDSA as the key exchange algorithm. Similarly, 
>> the
>> 749: #     bulk encryption algorithm "3DES_EDE_CBC" will disable all cipher 
>> suites
> 
> - `3DES_EDE_CBC` is already disabled by default in 
> `jdk.tls.disabledAlgorithms`, should we use a more realistic example like 
> `AES_128_CBC`?
> - Should we also include an example for `HmacSHA256`? It's far from being 
> obvious that such algorithm disables all cipher suites anding with `_SHA256`.

The reason I chose 3DES_EDE_CBC is because it is a case where one can see 
already that we have implemented something that is not specified. But I'm not 
tied to it, I can switch it to AES_128_CBC.

As for HmacSHA256 disabling cipher suites with SHA256 as the Mac in the name, I 
agree that is not obvious. Let me think about it as it is a bit awkward to 
document. It's a corner case but might be useful where you want to disable a 
Mac but not the algorithm if used as the hash algorithm in other places, like 
certificate signatures.

-------------

PR Review Comment: https://git.openjdk.org/jdk/pull/31933#discussion_r3596735664

Reply via email to