The sequence number 2^64-1 is itself RFC 5246 6.1-legal; only the wrap to 0
is forbidden. GetSEQIncrement reads the current counter then post-increments
it, so the check refuses the final legal sequence number to avoid the
wrapping post-increment. Document that this last value is deliberately
sacrificed rather than implying 2^64-1 is itself unusable.
DoAlert evicted the cached session from the fatal-alert handling that runs
before the plaintext-under-encryption validation, so a forged TLS 1.3
plaintext alert injected on an established connection evicted the session
(forcing a full handshake on reconnect) even though the alert is then
rejected as PARSE_ERROR. The unexpected_message teardown sent in response
also evicted through the SendAlert hook.
Move the receive-side eviction past the validation, into the branch that
processes a genuine alert, and have InvalidateSessionOnFatalAlert refuse to
evict for a TLS 1.3 plaintext alert received while encryption is on (the
current record was not decrypted) - covering both the receive path and the
unexpected_message teardown sent in response. RFC 8446 6.2 does not require
TLS 1.3 invalidation, so this loses nothing; TLS 1.2 (RFC 5246 7.2.2) is
unaffected.
The client's resumed-session EMS (F-5807) and cipher-suite (F-5811) checks
were enforced in CompleteServerHello at ServerHello-parse time. For stateless
ticket resumption the client sends an empty session ID and cannot yet tell
whether the server accepted the ticket (RFC 5077 3.4): a server that declines
the ticket falls back to a full handshake under a freshly negotiated
suite/EMS state, which these checks wrongly aborted with MATCH_SUITE_ERROR,
breaking the RFC 5077 ticket-decline fallback to a full handshake.
Move both checks into CheckResumptionConsistency and run it only once
resumption is confirmed - from whichever the server sends first in the
abbreviated flight: a renewed NewSessionTicket (before SetupSession refreshes
the cached suite/EMS to the current values) or its ChangeCipherSpec. By then
the "Not resuming as thought" path has cleared 'resuming' for any ticket
decline, so the full-handshake fallback proceeds.
Add test_tls12_resume_ticket_decline_fallback (ticket declined by a fresh
server CTX, full handshake under a different suite must succeed) and gate
test_tls12_resume_ticket_wrong_suite on WOLFSSL_NO_DEF_TICKET_ENC_CB so it
skips rather than fails in builds without the default ticket encryption
callback.
Address review on PR #10582:
- The client-side extended_master_secret consistency check skipped all
session-ticket resumptions, leaving a generic ticket resumption open to
an undetected EMS downgrade by a malicious server or MITM. The client
retains the EMS state for ticket sessions too (SetupSession), so the
check now applies to ticket resumption as well, mirroring the adjacent
cipher-suite check. Only EAP-FAST style resumption - where the
session-secret callback supplies the master secret for an opaque PAC
ticket - is exempt, matched precisely via ssl->sessionSecretCb just as
the callback invocation in DoServerHello does.
- Add test_tls_ems_resumption_server_downgrade, exercising the
client-direction downgrade (server resumes but omits EMS from its
ServerHello) for both session-ID and session-ticket resumption. This
client-side branch previously had no test coverage.
The TLS 1.2 client only compared the ServerHello suite against the
cached session suite for session-ID resumption; ticket resumption was
skipped on the assumption the suite is bound in the ticket. But the
ticket is opaque to the client, so it must enforce the match itself -
otherwise a server could resume a ticket under a different (weaker)
suite the client offered and the downgrade would go undetected
(RFC 5246 7.4.1.3).
The check is skipped only when the client retained no suite for the
session (cipherSuite0/cipherSuite both zero), so there is nothing to
compare against - as for EAP-FAST, whose PAC is a TLS ticket whose keys
come from the session-secret callback and which never populates the
cached suite. (0,0) is TLS_NULL_WITH_NULL_NULL, never negotiated, so it
unambiguously means "no retained suite". The EMS check remains
ticket-gated.
Add memio regression tests: a ticket resumption under a different
(retained) suite is rejected with MATCH_SUITE_ERROR, and a resumption
whose cached suite was not retained still succeeds.
In the WOLFSSL_OP_NO_RENEGOTIATION refusal path, WOLFSSL_LEAVE logged a
hard-coded 0 while the function actually returned SendAlert()'s result.
Capture the return value first so the trace reflects reality (e.g. when
SendAlert fails due to write backpressure) and return it.
The server validated client_verify_data only inside
TLSX_SecureRenegotiation_Parse, which never runs when the renegotiation_info
extension is absent, so a renegotiation ClientHello that omitted it was never
checked. Track a per-handshake renegInfoSeen flag and, after parsing the
renegotiation ClientHello extensions, abort with handshake_failure if the
extension was absent (RFC 5746 3.7). Also reject an SCSV received during
renegotiation (RFC 5746 3.5).
The documented 'reject peer-initiated renegotiation' option was accepted and
stored but never consulted. Now DoHelloRequest replies with a no_renegotiation
warning instead of starting SCR when the bit is set (client side), and the
server refuses a renegotiation ClientHello with a no_renegotiation warning
instead of resetting handshake state.
DoAlert marked a connection closed on a received fatal alert but left the
established session in the resumption cache, and the send path did the same,
so a session whose connection ended in a fatal alert remained resumable. Per
RFC 5246 Section 7.2.2 the session identifier MUST be invalidated; evict the
established session from the cache on both receipt and transmission of a fatal
alert via the new InvalidateSessionOnFatalAlert helper.
GetSEQIncrement silently rolled the 64-bit write sequence counter from 2^64-1
back to 0, reusing sequence number 0 with the same keys. Per RFC 5246 Section
6.1 sequence numbers MUST NOT wrap. BuildMessage now refuses to emit a TLS 1.2
record once the write sequence number has reached its maximum, returning the
new SEQUENCE_NUMBER_E error so the caller renegotiates or closes instead.
On session-ID resumption the client only checked that the server's selected
suite was in its offered list, not that it equaled the resumed session's
suite, so a server could resume the session ID under a different cipher suite.
Per RFC 5246 Section 7.4.1.2 / F.1.4 a resumed session reuses its negotiated
suite; abort with a fatal illegal_parameter on mismatch.
CompleteServerHello's resumption branch derived keys from the cached master
secret without checking the resumed session's extended_master_secret state
against the abbreviated ServerHello, letting a MITM strip EMS on resumption.
Per RFC 7627 Section 5.3, abort with a fatal handshake_failure when the cached
session's EMS flag does not match the ServerHello EMS state.
FreeCiphers released the DTLS 1.3 record-number protection ChaCha contexts
with XFREE only, leaving key material in freed heap memory. ForceZero both
contexts before freeing, matching the regular TLS ChaCha path in
FreeCiphersSide, and also zeroize a partially-set key in
Dtls13InitChaChaCipher when wc_Chacha_SetKey fails.
DoTls13CertificateRequest advanced past the certificate_request_context and
extensions blocks but never verified the whole message body was consumed,
silently ignoring trailing bytes. RFC 8446 Section 4.3.2 fixes the wire
format; enforce that the consumed length equals the message size and return
BUFFER_ERROR (decode_error) otherwise.
DoTls13EncryptedExtensions only bounds-checked the extensions length against
the message size, silently ignoring any trailing bytes. RFC 8446 Section 4.3.1
defines the message as solely the extensions block, so enforce length equality
and return BUFFER_ERROR (decode_error) on a mismatch.
Extend test_wolfSSL_RSA_verify and test_wolfSSL_RSA_padding_add_PKCS1_PSS
with negative cases that flip a byte in the signature/encoding and in the
hash, asserting verification fails. This guards the XMEMCMP-based signature
acceptance decision in wolfSSL_RSA_verify_mgf against regressions that would
let any decryption result of matching length pass as valid.
Change to ignore plaintext alerts when
WOLFSSL_TLS13_IGNORE_PT_ALERT_ON_ENC is defined only until first
encrypted message from peer is seen.
Negative testing added.
Leak fixes: free existing ssl->buffers.key before overwriting in SetSSL_CTX() (internal.c) and wolfSSL_set_SSL_CTX() (ssl.c)
UAF fix: wc_CheckRsaKey() — mp_memzero_check(tmp) moved before the free (rsa.c)
Build guards: #ifndef NO_ED25519/ED448_VERIFY around forged-sig test data (test_ed25519/ed448.c); guard equal()/cmov() for verify-only builds (ge_operations.c); guard unused pointers under WOLFSSL_MLDSA_SIGN_SMALL_MEM_PRECALC (wc_mldsa.c)
Test cleanups (test.c): fix UB from out-of-range enum in hash_test(), always free AES dec object, fix der buffer declaration under small-stack builds
Removed WOLFSSL_PSK_ID_PROTECTION from use as it is now on by default.
Always check whether the server has a certificate (not a CA chain).
If there is a certificate then continue, otherwise, report a binder
error.
Added test to ensure binder error returned and alert sent when no
NO_CERT. test_tls13_bad_psk_binder already tested no certificate.
Allowed memio test harness to be built when NO_CERT is defined.
F-1379
Better handling of the lazy mutex initialization to use atomics where
available.
Improved atomic code when no system support:
- add types
- used types in functions
Add --no-ec to unit.test to not run wolfCrypt tests.
Juliusz Sosinowicz
Daniel Pouzzner
David Garske
Sean Parkinson
Mattia Moffa
Michael Rogov Papernov