Implement RFC8773bis (draft-ietf-tls-8773bis-13)
cert_with_extern_psk for TLS 1.3, including protocol checks
and API support.
Includes unit tests for API and handshake behavior as well
as tests in the testsuite using extended examples.
Per RFC 8446 section 8, a server MUST ensure that any instance of it
would accept 0-RTT for the same 0-RTT handshake at most once. Without
this, the same ClientHello could be replayed to re-accept early data on
a subsequent connection.
After the PSK is authenticated (binder verified) in DoPreSharedKeys,
call wolfSSL_SSL_CTX_remove_session on ssl->session when the client
offered 0-RTT and the session permits it. That evicts the entry from
the internal cache (under the row's write lock) and invokes the
application's ctx->rem_sess_cb so any external cache can drop its copy
too. The session's timeout is also cleared so the live reference held
by the current handshake cannot be resumed again.
The mutation is paid only when the client actually included the
early_data extension on a 0-RTT-capable session, so normal resumptions
are unaffected and the existing remove-callback counts in
test_wolfSSL_CTX_add_session_ext_{tls13,dtls13} stay correct.
wolfSSL_SSL_CTX_remove_session was previously declared and defined only
under the OpenSSL compatibility layer. Because it is now called from
the core TLS 1.3 PSK path, the declaration in wolfssl/ssl.h and the
definition in src/ssl_sess.c are moved out of that block to match the
existing !NO_SESSION_CACHE gate under which the function is meaningful.
wolfSSL_SSL_get0_session stays in the compat block.
test_tls13_early_data_0rtt_replay verifies the behaviour. It does a
full TLS 1.3 handshake with stateful tickets (SSL_OP_NO_TICKET) and
max_early_data > 0, then tries to resume the saved session twice while
offering 0-RTT each time. A minimal single-slot external session cache
is wired up via wolfSSL_CTX_sess_set_{new,get,remove}_cb to confirm
both caches are cleared. Round 0 must resume and deliver the early
data, and rem_calls must hit 1 (the fix's single eviction). Round 1
must fall back to a full handshake (session_reused == 0), deliver no
early data, and leave rem_calls at 1.
Verified against multiple configurations (incl. --enable-all
--enable-earlydata, the no-compat -DHAVE_EXT_CACHE build, and the
os-check.yml combo). Valgrind under -g2 -O0 with OPENSSL_EXTRA +
HAVE_EXT_CACHE + HAVE_EX_DATA reports no errors and no
definitely-lost bytes.
Refs wolfSSL/wolfssl#10197
dtls13.c:
- Fix wrong return value in Dtls13SendFragmentedInternal error path (return outputSz instead of recordLength)
- Fix incomplete bounds check in Dtls13SendFragmented to account for DTLS_HANDSHAKE_HEADER_SZ
- Fix wrong WOLFSSL_ENTER trace string in Dtls13EpochCopyKeys
tls13.c:
- Remove wrong (byte) cast on cookie->len passed to TlsCheckCookie
- Add missing bounds check on PSK identityLen in SetupPskKey before copying to client_identity
- Fix data race on static header array in ExpectedResumptionSecret
- Add defensive underflow check in EncryptTls13 for consistency with DecryptTls13
- Fix wrong return variable in DTLS 1.3 Finished send error path (return dtlsRet instead of ret)
- Add missing SM3 case and default in Tls13_Exporter hash switch to prevent NULL dereference
- Initialize *outSz to 0 in wolfSSL_write_early_data to match wolfSSL_read_early_data
- Add bounds check for bindersLen against helloSz in CheckPreSharedKeys
- Fix resource leak and hash state corruption in ExpectedResumptionSecret error paths
- Fix memory leak of rsaSigBuf in dual-alg RSA+RSA CertificateVerify
- Guard against word32 underflow in inputLength - HANDSHAKE_HEADER_SZ in DoTls13HandShakeMsg
- Fix swapped side parameter in DeriveFinishedSecret for server-side Finished processing
- Fix no_mac fall-through in ssl_handshake_md to return NULL instead of wrong digest
- Fix strict aliasing violation in FindPsk PSK key size check
- Remove duplicate !ssl->options.dtls check in TLS 1.3 middlebox compat condition
tests:
- Add regression tests for wolfSSL_write_early_data outSz initialization and DTLS 1.3 Finished send error propagation
* Enable ML-KEM by default in build systems (autoconf and CMake)
* Only allow three to-be-standardized hybrid PQ/T combinations by
default
* Use X25519MLKEM768 as the default KeyShare in the ClientHello (if user
does not override that). When Curve25519 is disabled, then either
WOLFSSL_SECP384R1MLKEM1024 or WOLFSSL_SECP256R1MLKEM768 is used as
default depending on the ECC configuration
* Disable standalone ML-KEM in supported groups by default (enable with
--enable-tls-mlkem-standalone)
* Disable extra OQS-based hybrid PQ/T curves by default and gate
behind --enable-experimental (enable with --enable-extra-pqc-hybrids)
* Reorder the SupportedGroups extension to reflect the preferences
* Reorder the preferredGroup array to also reflect the same preferences
* Add async support for ML-KEM hybrids
CRL APIs not usable when NO_ASN_TIME defined.
WOLFSSL_TLS13 needs to be defined with HAVE_ECH.
When session ticket encrypted with CBC, must be a multiple of block
size.
Fix test define protection.
Fix ML-DSA protection of reduction functions.
Need !NO_RSA with WC_RSA_PSS.
Connection ID is not a DTLS 1.3 only extension.
Mostly combinations of NO_WOLFSSL_CLIENT, NO_WOLFSSL_SERVER and
WOLFSSL_NO_CLIENT_AUTH were failing.
Added configurations to CI loop.
wc_AesGcmDecryptFinal: use WC_AES_BLOCK_SIZE to satisfy compiler.
The specification states to return illegal_parameter when a message is
syntactically correct but semantically invalid. (RFC 8446 section 6,
Paragraph 5)
Change TLS 1.3 handling to return missing_extension alert code when
- KeyShare is present but SupportedGroups is missing and
- SupportedGroups is present but KeyShare is missing
Added tests for this.
- Add validation to ensure the cipher suite in the ServerHello matches the one specified in the HelloRetryRequest.
- test_TLSX_CA_NAMES_bad_extension: use the same ciphersuite in HRR and SH
x509.c: realloc may fail and therefore need to store result in a
temporary so the old pointer is not lost.
tls.c: free the name if it is not pushed on to the stack of peer CA
names. Failure to push can be from memory allocation failure.
aes.c: Don't compile XTS decrypt functions without HAVE_AES_DECRYPT.
Fix tests to have better pre-processor protection.
Eric Blankenhorn
JacobBarthelmeh
Tobias Frauenschläger
Jeremiah Mackey
Juliusz Sosinowicz
Reda Chouk
Anthony Hu
sebastian-carpenter
Sean Parkinson
Daniel Pouzzner
Marco Oliverio
Josh Holtrop
David Garske
Chris Conlon
Kareem