Update the pkalgs domain to use PSA macros in depends.py

[gabor-mezei-arm]

Description

In depends.py use PSA macros for the pkalgs domain.

Resolve #9145
Depends on #9292 (merged)

PR checklist

Please remove the segment/s on either side of the | symbol as appropriate, and add any relevant link/s to the end of the line.
If the provided content is part of the present PR remove the # symbol.

• changelog not required because: test related changes
• development PR provided this is
• framework PR not required
• 3.6 PR not required because: 4.0 related changes
• 2.28 PR not required because: 4.0 related changes
• tests provided

[minosgalanakis]

LGTM

[mpg]

@gabor-mezei-arm This has a conflict.

@ronald-cron-arm Kind reminder that this is waiting for your review :)

[mpg]

@ronald-cron-arm Are you still planning to review this? We're trying to close off our DI task and we have one that depends on this.

[ronald-cron-arm]

@mpg I can see that you already pinged me about the review of this PR but I somehow missed it, sorry about that. I looked at this PR sometimes ago, I can finalize my review of it.

[ronald-cron-arm]

@gabor-mezei-arm could you rebase to resolve the conflict? Thanks.

[mpg]

@mpg I can see that you already pinged me about the review of this PR but I somehow missed it, sorry about that.

No worries. Quite ironically, I missed the notification about your reply, and only saw it today while checking the status of the PR :)

I looked at this PR sometimes ago, I can finalize my review of it.

Great, thanks!

[ronald-cron-arm]

This PR changes the definitions of the pkalgs complementaty domains to be based on PSA_WANT_ symbols instead of MBEDTLS_ ones, in preparation of the removal of the MBEDTLS_ symbols as configuration options. The MbedTLS configurations that are tested as part of the pkalgs complementary domain remain the same. Thus this can be merged as it is to me.

After that step, we would probably want to adapt the definition of the pkalgs domain to the specifics of the PSA_WANT_ symbols compared to the MBEDTLS_ ones. For example, having a test where only PSA_WANT_ALG_RSA_OAEP is disabled and another one where PSA_WANT_ALG_RSA_PSS is disabled, whereas currently we only have one test that disables both. We would also probably want to keep the test where both are disabled and to have the three cases we need to add support for grouping as discussed in #9634.

[mpg]

I'm not super familiar with the concept of grouping as discussed in 9634. @gabor-mezei-arm can you create (an) issue(s) to track the follow-up(s) that Ronald is suggesting? Then add it/them to the 4.0 board, probably as SHOULDs in the "PSA Crypto always on" topic? Thanks in advance.

[minosgalanakis]

LGTM