* [Medium] CVE-2023-6935: After review of the previous RSA timing fix in wolfSSL 5.6.4, additional changes were found to be required. A complete resistant change is delivered in this release. This fix is for the Marvin attack, leading to being able to decrypt a saved TLS connection and potentially forge a signature after probing with a very large number of trial connections. This issue is around RSA decryption and affects the optional static RSA cipher suites on the server side, which are considered weak, not recommended to be used and are off by default in wolfSSL (even with --enable-all). Static RSA cipher suites were also removed from the TLS 1.3 protocol and are only present in TLS 1.2 and lower. All padding versions of RSA decrypt are affected since the code under review is outside of the padding processing. Information about the private keys is NOT compromised in affected code. It is recommended to disable static RSA cipher suites and update the version of wolfSSL used if using RSA private decryption alone outside of TLS. Thanks to Hubert Kario for the report. The fix for this issue is located in the following GitHub Pull Request: https://github.com/wolfSSL/wolfssl/pull/6955.
* [Low] CVE-2023-6936: A potential heap overflow read is possible in servers connecting over TLS 1.3 when the optional WOLFSSL_CALLBACKS has been defined. The out of bounds read can occur when a server receives a malicious malformed ClientHello. Users should either discontinue use of WOLFSSL_CALLBACKS on the server side or update versions of wolfSSL to 5.6.6. Thanks to the tlspuffin fuzzer team for the report which was designed and developed by; Lucca Hirschi (Inria, LORIA), Steve Kremer (Inria, LORIA), and Max Ammann (Trail of Bits). The fix for this issue is located in the following GitHub Pull Request: https://github.com/wolfSSL/wolfssl/pull/6949.
* [Low] A side channel vulnerability with AES T-Tables is possible in a very controlled environment where precision sub-cache-line inspection can happen, such as inside an Intel SGX enclave. This can lead to recovery of the AES key. To prevent this type of attack, wolfSSL added an AES bitsliced implementation which can be enabled with the “--enable-aes-bitsliced” configure option. Thanks to Florian Sieck, Zhiyuan Zhang, Sebastian Berndt, Chitchanok Chuengsatiansup, Thomas Eisenbarth, and Yuval Yarom for the report (Universities of Lübeck, Melbourne, Adelaide and Bochum). The fix for this issue is located in the following GitHub Pull Request: https://github.com/wolfSSL/wolfssl/pull/6854.
* [Low] CVE-2023-6937: wolfSSL prior to 5.6.6 did not check that messages in a single (D)TLS record do not span key boundaries. As a result, it was possible to combine (D)TLS messages using different keys into one (D)TLS record. The most extreme edge case is that, in (D)TLS 1.3, it was possible that an unencrypted (D)TLS 1.3 record from the server containing first a ServerHello message and then the rest of the first server flight would be accepted by a wolfSSL client. In (D)TLS 1.3 the handshake is encrypted after the ServerHello but a wolfSSL client would accept an unencrypted flight from the server. This does not compromise key negotiation and authentication so it is assigned a low severity rating. Thanks to Johannes Wilson for the report (Sectra Communications and Linköping University). The fix for this issue is located in the following GitHub Pull Request: https://github.com/wolfSSL/wolfssl/pull/7029.
