Combine psa_pake_set_password_key() with psa_pake_setup()

[amtkarm1]

Description

Fixes #9320

Gatekeeper note: interdependent with Mbed-TLS/tf-psa-crypto#215 - handle with care!

PR checklist

• [X] changelog provided
• [X] development PR provided HERE
• [X] TF-PSA-Crypto PR provided Mbed-TLS/tf-psa-crypto#215
• [X] framework PR not required
• [x] 3.6 PR not required because: no backport required
• [x] 2.28 PR not required because: no backport required
• tests provided

[mpg]

Unfortunately this now has a conflict in tf-psa-crypto. Suggested resolution:

1. Update the associated crypto PR in order to fix the code style issues - quite probably rebase it too, because incompatible changes have happened in the meantime.
2. Update this PR, to point to the tip of the crypto PR (again, also rebase to avoid possible issues cause by incompatible changes that happened in the meantime).
3. Wait for CI to complete and make sure reviewers re-review promptly.
4. Make sure a gatekeeper is aware of the situation once both PRs are re-approved with green CI on the TLS PR, so the PRs can be merged in close succession (with announcements on slack, to be handled by the gatekeeper) before they grow another conflict.

[gilles-peskine-arm]

Sorry, there's a submodule conflict again.

[amtkarm1]

Similar CI issue to previous PR, should be fixed in the next run!

[bjwtaylor]

I've just done a checkout of this one and the tf-psa-crypto submodule pointer it still pointing to the fork. Have I missed something in the merge strategy, or does this still need to be adjusted?

[mpg]

You've not missed anything in the merge strategy, it is indeed supposed to point the the current tip of development.

And I think it does: on the github diff page if I hover or click on "47 files" in "Submodule tf-psa-crypto updated 47 files" I can see the new commit is fc1dca61954ee58701a47ba24cc27004e05440b2 which indeed looks like current development of tf-psa-crypto.

Perhaps you need to run git submodule update tf-psa-crypto in your local checkout?