RFR: 8341346: Add support for exporting TLS Keying Material [v10]

[Sean Mullan]

On Fri, 9 May 2025 21:38:07 GMT, Bradford Wetmore <wetmore at openjdk.org> wrote:

>> Adds the RFC 5705/8446 TLS Key Exporters API/implementation to JSSE/SunJSSE respectively.
>> 
>> CSR is underway.
>> 
>> Tests include new unit tests for TLSv1-1.3.  Will run tier1-2, plus the JCK API (jck:api/java_security jck:api/javax_crypto jck:api/javax_net jck:api/javax_security jck:api/org_ietf jck:api/javax_xml/crypto)
>
> Bradford Wetmore has updated the pull request with a new target base due to a merge or a rebase. The pull request now contains 13 commits:
> 
>  - Merge branch 'master' into JDK-8341346
>  - Adjustments made for JDK-8350830
>  - Merge branch 'master' into JDK-8341346
>  - Rework to avoid PKCS11 data extraction problems, and enhanced input verification and unit testing
>  - More Codereview comments
>  - Updated to use the upcoming KDF (still in preview) + bits of JDK-8353578 for compilation)
>  - Add in the SharedSecrets SecretKeySpec clearing mechanism
>  - More codereview/CSR comments
>  - Merge branch 'master' into JDK-8341346
>  - Codereview comments.
>  - ... and 3 more: https://git.openjdk.org/jdk/compare/68a11850...bd227aa8

src/java.base/share/classes/javax/net/ssl/ExtendedSSLSession.java line 172:

> 170:      * algorithms defined in RFCs 5705/8446.
> 171:      * <P>
> 172:      * Note RFC 5705 (for (D)TLSv1.2 and earlier) calculates different EKM

s/Note/Note that/

(or just remove "Note" - don't think it is really necessary).

src/java.base/share/classes/javax/net/ssl/ExtendedSSLSession.java line 187:

> 185:      *
> 186:      * @implSpec The default implementation throws
> 187:      *           {@code UnsupportedOperationException}

Missing period at end of sentence.

src/java.base/share/classes/javax/net/ssl/ExtendedSSLSession.java line 192:

> 190:      *                {@code label} will be converted to a {@code byte[]}
> 191:      *                before the operation begins
> 192:      * @param context the context bytes used in the EKM calculation

Add something that says `null` is acceptable. i.e. ", or `null` ..."

src/java.base/share/classes/javax/net/ssl/ExtendedSSLSession.java line 195:

> 193:      * @param length  the number of bytes of EKM material needed
> 194:      *
> 195:      * @throws SSLKeyException if the key could not be generated

I prefer "cannot" (present tense).

src/java.base/share/classes/javax/net/ssl/ExtendedSSLSession.java line 204:

> 202:      *
> 203:      * @return a {@code SecretKey} that contains {@code length} bytes of the
> 204:      *         EKM material.

no period necessary.

src/java.base/share/classes/javax/net/ssl/ExtendedSSLSession.java line 225:

> 223:      * character encoding.
> 224:      * <P>
> 225:      * Depending on the chosen underlying key derivation mechanism, the

This reads like it should maybe be an `@apiNote`.

src/java.base/share/classes/javax/net/ssl/ExtendedSSLSession.java line 254:

> 252:      *
> 253:      * @return a byte array of size {@code length} that contains the EKM
> 254:      *         material, or null if the derived key material does not support

For "or null if the derived key material does not support encoding" why wouldn't an implementation throw UOE instead?

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

PR Review Comment: https://git.openjdk.org/jdk/pull/24976#discussion_r2084853417
PR Review Comment: https://git.openjdk.org/jdk/pull/24976#discussion_r2084889527
PR Review Comment: https://git.openjdk.org/jdk/pull/24976#discussion_r2084879618
PR Review Comment: https://git.openjdk.org/jdk/pull/24976#discussion_r2084887858
PR Review Comment: https://git.openjdk.org/jdk/pull/24976#discussion_r2084891743
PR Review Comment: https://git.openjdk.org/jdk/pull/24976#discussion_r2084895438
PR Review Comment: https://git.openjdk.org/jdk/pull/24976#discussion_r2084899365