[cconlon]

This PR includes 5 Fenrir fixes:

• F-15725: Correct a comment in the native receive callback.
• F-15752: Exclude native build objects from make dist output.
• F-15699: Check RSA verify, sign check and decrypt callback return lengths against the buffer size, with tests.
• F-15700: Check atomic decrypt callback padding sizes against the record size, with tests.
• F-15742: Document that callback ByteBuffer arguments are only valid during the callback.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[cconlon]

@claude review

[wolfSSL-Fenrir-bot]

Fenrir Automated Review — PR #445

Scan targets checked: wolfssljni-src, wolfssljni-bugs
Not reviewed (no scan target covers them): Makefile
Coverage: 1 of 17 in-scope changed file(s) opened by the reviewer; not opened: native/com_wolfssl_WolfSSLSession.c, src/java/com/wolfssl/WolfSSLByteBufferIORecvCallback.java, src/java/com/wolfssl/WolfSSLByteBufferIOSendCallback.java, src/java/com/wolfssl/WolfSSLDecryptVerifyCallback.java, src/java/com/wolfssl/WolfSSLEccSharedSecretCallback.java, src/java/com/wolfssl/WolfSSLEccSignCallback.java, src/java/com/wolfssl/WolfSSLEccVerifyCallback.java, src/java/com/wolfssl/WolfSSLMacEncryptCallback.java, src/java/com/wolfssl/WolfSSLRsaDecCallback.java, src/java/com/wolfssl/WolfSSLRsaEncCallback.java and 6 more

Fenrir result: Approved ✅

No new issues found in the changed files.

Advisory only — this automated result does not count as a GitHub approval.

Review tier: Lite

[claude]

I reviewed this PR and didn't find any bugs. The new bounds checks are in JNI crypto callback paths whose output lengths and pointers go straight into wolfSSL, so a maintainer who knows those wolfSSL callback contracts should still take a look.

What was reviewed:

• The padSz checks in NativeDecryptVerifyCb and NativeVerifyDecryptCb. Values that are negative or larger than decSz are now rejected before the cast to unsigned int.
• The positive-return length checks against sigSz/inSz in the RSA verify, sign-check, PSS sign-check and decrypt callbacks. Each one runs before *out is set, so an oversized return never reaches wolfSSL.
• The Makefile dist cleanup of .o/.d files, the receive-callback comment fix, the ByteBuffer lifetime Javadoc, and the new oversized-return JUnit tests, including the pkCbHandshakeLoop change to also return the error code.

Extended reasoning...

The change adds input validation to the native JNI callbacks in com_wolfssl_WolfSSLContext.c, which are the paths where Java-supplied lengths and padding sizes go back into wolfSSL's TLS record and RSA handling. It also fixes a comment, tidies the dist target, adds Javadoc to 14 callback interfaces, and adds about 270 lines of handshake-based tests. Memory-safety checks in crypto callbacks are security-sensitive, so this is not a mechanical change to auto-approve, even though the checks look correct and are covered by tests. An earlier automated reviewer left a COMMENTED review whose content isn't visible here.