[Renesas RA6M4] Fix Renesas RA6M4 SCE session-key generation issues

[miyazakh]

Summary

Three fixes to the Renesas FSPSM (RA6M4 Secure Cryptography Engine) port, driven by cases
found in wolfSSL's internal issue tracker:

• renesas_fspsm_util.c — wc_fspsm_generateSessionKey():
  • Fix a NULL-pointer dereference when the AES-GCM deferred-key-generation path leaves the
    session key buffers NULL.
  • Fix a hardware mutex leak on out-of-memory during session-key setup.
• renesas_fspsm_aes.c — wc_fspsm_AesGcmDecrypt():
  • Fix a copy-paste error in an error message.

Details

NULL-pointer dereference in wc_fspsm_generateSessionKey (issue f5408)

When AES-GCM session-key generation takes the deferred path, key_client_aes and
key_server_aes are left NULL (the real per-record key is generated later, inside
wc_fspsm_AesGcmEncrypt/Decrypt). The subsequent XMEMCPY into
enc/dec->aes->ctx.wrapped_key did not check for this and copied from the NULL pointers
unconditionally, crashing — remotely triggerable via cipher suite choice.

Fix: guard the copy with if (key_client_aes != NULL && key_server_aes != NULL).

Hardware mutex leak on allocation failure (issues f5417, f5409 — duplicate)

Five return MEMORY_E; exit paths inside the wc_fspsm_hw_lock()'d block of
wc_fspsm_generateSessionKey (client/server AES key alloc, encrypt-side Aes/wrapped-key
alloc, decrypt-side Aes/wrapped-key alloc) returned without calling
wc_fspsm_hw_unlock(), leaving the hardware mutex permanently locked after an allocation
failure.

Fix: add wc_fspsm_hw_unlock(); before each of the 5 early returns.

(f5409 reported the same root cause under a different description; no additional change was
needed beyond the f5417 fix.)

Copy-paste error in wc_fspsm_AesGcmDecrypt error message (issue 1537)

The authTagSz-too-small error message read "GcmEncrypt authTagSz too small error" inside
the decrypt function — copy-pasted from the encrypt path. Message-only fix, no functional
impact: now reads "GcmDecrypt authTagSz too small error".

Testing

Run crypt and benchmark test on the board

Checklist

• [n/a] added tests
• [n/a] updated/added doxygen
• [n/a] updated appropriate READMEs
• [n/a] Updated manual and documentation

[Copilot]

Pull request overview

This PR fixes correctness issues in the Renesas RA6M4 FSPSM (SCE) TLS port related to session-key generation, specifically addressing a NULL-dereference in the AES-GCM deferred-key path, ensuring the hardware mutex is always released on allocation failures, and correcting a misleading decrypt-side log message.

Changes:

• Prevent NULL-pointer dereference by skipping session-key buffer copies when AES-GCM session-key generation is deferred.
• Fix multiple early-return out-of-memory paths to free partial allocations and always release the FSPSM hardware mutex.
• Correct a copy-paste error in the AES-GCM decrypt error message.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
wolfcrypt/src/port/Renesas/renesas_fspsm_util.c	Adds NULL-guarded key copy and ensures cleanup + hw unlock on allocation failures during session-key setup.
wolfcrypt/src/port/Renesas/renesas_fspsm_aes.c	Fixes an incorrect decrypt-side log message string.

---

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

[github-actions]

MemBrowse Memory Report

No memory changes detected for:

• gcc-arm-cortex-m0plus
• gcc-arm-cortex-m3
• gcc-arm-cortex-m4
• gcc-arm-cortex-m4-baremetal
• gcc-arm-cortex-m4-crypto-only
• gcc-arm-cortex-m4-dtls13
• gcc-arm-cortex-m4-min-ecc
• gcc-arm-cortex-m4-openssl-compat
• gcc-arm-cortex-m4-pkcs7
• gcc-arm-cortex-m4-pq
• gcc-arm-cortex-m4-rsa-only
• gcc-arm-cortex-m4-sp-math
• gcc-arm-cortex-m4-tls12
• gcc-arm-cortex-m4-tls13
• gcc-arm-cortex-m7
• gcc-arm-cortex-m7-pq
• gcc-arm-cortex-m7-tls13
• linuxkm-pie
• linuxkm-standard
• stm32-sim-stm32h753

[miyazakh]

retest this please