ML-KEM fixes

[Frauschi]

Various ML-KEM related fixes (extracted from #9732):

• DTLS 1.3 cookie and CH frag handling
• static memory handling
• Fix memory leak in TLS server PQC handling in case of ECH
• Make sure hybrids are actually tested in testsuite
• Add tests for individual ML-KEM levels (follow up on Fix switch case handling in TLSX_IsGroupSupported function #9777)

All of these changes have already been reviewed and approved as part of #9732.

[Copilot]

Pull request overview

Updates ML-KEM/DTLS 1.3 behavior and static-memory sizing, fixes a TLS server-side PQC/ECH-related leak, and expands CI/test coverage to ensure ML-KEM and hybrid groups are exercised.

Changes:

• Enable/auto-configure DTLS 1.3 CH fragmentation for ML-KEM and adjust DTLS cookie/HRR initialization.
• Fix PQC key share memory handling (free before overwrite; cleanup when group unsupported) and tune static-memory bucket sizing.
• Expand tests/CI to cover hybrid groups and individual ML-KEM parameter sets.

Reviewed changes

Copilot reviewed 12 out of 12 changed files in this pull request and generated 6 comments.

Show a summary per file

File	Description
wolfssl/wolfcrypt/settings.h	Auto-enables DTLS CH fragmentation when ML-KEM is enabled with DTLS 1.3.
wolfssl/wolfcrypt/memory.h	Adjusts static-memory buckets/distribution for ML-KEM builds.
wolfssl/ssl.h	Updates draft references for ML-KEM and hybrid key agreement groups.
wolfssl/internal.h	Increases max curve-name buffer size to fit new hybrid group names.
tests/suites.c	Extends “available group” detection to ML-KEM + hybrids; increases static memory for suite tests.
tests/api/test_tls13.c	Ensures tests pick an actually-enabled ML-KEM/Kyber level.
tests/api.c	Updates static memory sizing gates; improves DTLS1.3 PQC frag test group selection and validation.
src/tls.c	Fixes PQC key share pubKey overwrite leak; clears key share state when group unsupported.
src/internal.c	Initializes DTLS1.3 HRR cookie and enables CH fragmentation for ML-KEM during init.
examples/server/server.c	Adjusts static-memory sizing condition for ML-KEM.
examples/client/client.c	Ensures DTLS+PQC runs with SupportedGroups constrained to preserve desired KEM after HRR.
.github/workflows/pq-all.yml	Adds CI runs for individual ML-KEM levels (512/768/1024).

Comments suppressed due to low confidence (1)

src/internal.c:1

• This introduces new default behavior (enabling DTLS 1.3 CH fragmentation during initialization when ML-KEM is enabled). The existing DTLS1.3 PQC frag test in tests/api.c still calls wolfSSL_dtls13_allow_ch_frag() explicitly, so it won’t catch regressions in this init-time path. Consider adding/adjusting a test to exercise a DTLS 1.3 ML-KEM handshake without manually enabling CH fragmentation, validating that init-time enabling works.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

You can also share your feedback on Copilot code review. Take the survey.

[Frauschi]

Addressed Copilot issues.