Add PrivateKey object for signing, only

[mandelsoft]

Proposed Changes

Fully implementing rsa.Signer requires a private and public key. But if only signing is required, the private key would be
sufficient. Currently, the public key must be present and both must have the same Id. If one of those constraints is not satisfied signing with the private key is not possible.

Types of Changes

New Feature: introduce interface PrivateKey with sole signing method and provide appropriate context methods to find private keys.

Verification

Create keys without Id and use new FindPrivateKey method flavors.

Testing

Tests are added.

Linked Issues

User-Facing Change

Support finding private keys without matching public key usable for sole signing operation.

Further Comments

[IceManGreen]

Thank you @mandelsoft for your proposition, this is indeed a better approach for private keys in KMS.
Also, this pull request might be the solution to #112.

[solcates]

Looks good and safe to me. Thanks for the contribution.

[IceManGreen]

However, the code is still in addendum.go. The implementation should be in keys.go though.

@mandelsoft can you fix this ?

[Nicolas-Peiffer]

Thanks for this, @mandelsoft — and apologies it sat open so long.

---

In 2026 the big project for crypto11 is to add the support of PKCS#11 v3.2 and the support of the ML-KEM PQC algorithm:

• add support for pkcs11 v3.2 ml-kem #131
• We work on this branch mainly: pkcs11-v3.2-ml-kem.
• It will lead to a crypto11@v2, as there are many changes, including we are using a new PKCS11 implementation eclipse-keypont/pkcs11-go.
• Summary of the changes here: pkcs11-v3.2-ml-kem/CHANGELOG.md

The other big project for 2026 is giving + moving the crypto11 project under the Eclipse Foundation.

So this is an opportunity to look back at PR and issues, like yours.

---

Closing #111 as superseded by #115: crypto11@v1.6.x now carries this feature. #115 ("RSA support for asymmetric decryption", commit c3b42eb) needed private-key retrieval for its decryption path and landed the same design you proposed here — the PrivateKey interface (Sign + KeyType() uint) and all four lookups:

• Context.FindPrivateKey(id, label)
• Context.FindPrivateKeys(id, label)
• Context.FindPrivateKeyWithAttributes(attrs)
• Context.FindPrivateKeysWithAttributes(attrs)

It also adds RSAPrivateKey for Sign + Decrypt on a private half alone.

Two differences from your version, in the interest of full disclosure:

1. Not-found: yours returns (nil, nil); master returns an error. If you were relying on the nil-nil signal, say so — arguably master has this wrong, since its doc comment still promises "or nil if it cannot be found".
2. Tests: your TestFindingPrivateKeysWithAttributes and the FindPrivateKey(nil, nil) argument-validation assertions have no equivalent on master. Those are still worth having — happy to cherry-pick them with attribution, or send them as a small PR if you'd prefer to keep the credit.

So we added your tests on 62bbca8 and we mention you as contributor. This targets crypto11@v2

The implementation is redundant now, but you framed the problem first and the result matches your proposal.