The cryptocb dispatcher opened and closed a fresh PKCS#11 session around
each HMAC invocation. PKCS#11 sign operations are session-scoped, so a
multi-call HMAC (wc_HmacUpdate then wc_HmacFinal, which arrive as
separate cryptocb dispatches) had its C_SignFinal land on a session
that never saw a C_SignInit, returning CKR_OPERATION_NOT_INITIALIZED
and surfacing as WC_HW_E. This broke any code path that drives Update
and Final separately under PKCS#11 routing.
Cache the PKCS#11 session handle on Hmac.devCtx (cast through wc_ptr_t,
matching the existing pattern for cached PKCS#11 object handles) and
rebuild the Pkcs11Session on the stack. The session is opened on the
first dispatch when the operation enters
WC_HMAC_INNER_HASH_KEYED_DEV state and released when it leaves that
state (Final completed or hard error).
RFC 8446 section 8 requires any server instance to accept 0-RTT for a
given ClientHello at most once. Prior to this change wolfSSL's behaviour
diverged from that requirement in several ways:
* ctx->maxEarlyDataSz defaulted to MAX_EARLY_DATA_SZ whenever the
library was built with WOLFSSL_EARLY_DATA, so servers auto-
advertised 0-RTT in NewSessionTicket without the application
asking. RFC 8446 E.5 says 0-RTT MUST NOT be enabled unless
specifically requested.
* The post-accept eviction is compiled out under NO_SESSION_CACHE,
so builds without the cache accepted 0-RTT with no replay defence.
* Stateless self-encrypted tickets do not carry a session ID on the
stateless DoClientTicket decrypt path, so wolfSSL_SSL_CTX_remove_
session could not locate them to evict.
* wolfSSL_SSL_CTX_remove_session always returned 0 on success
regardless of whether the session was actually in the cache,
diverging from OpenSSL's SSL_CTX_remove_session (1 on success,
0 on not-found).
Changes:
* src/internal.c: ctx->maxEarlyDataSz defaults to 0; applications
must opt in with wolfSSL_CTX_set_max_early_data.
* src/tls13.c: #error when WOLFSSL_EARLY_DATA is built with
HAVE_SESSION_TICKET and NO_SESSION_CACHE. Escape hatch
WOLFSSL_EARLY_DATA_NO_ANTI_REPLAY for deployments that take
application-layer responsibility.
* wolfssl/internal.h: imply WOLFSSL_TICKET_HAVE_ID from
WOLFSSL_EARLY_DATA so stateless-ticket issuance populates the
cache under an ID that eviction can find.
* src/ssl_sess.c: wolfSSL_SSL_CTX_remove_session returns 1 when the
session was found (internal-cache hit, or ctx->rem_sess_cb fired
for an external cache), 0 otherwise. Matches OpenSSL semantics.
* src/tls13.c: the 0-RTT acceptance condition in CheckPreSharedKeys
now calls wolfSSL_SSL_CTX_remove_session and checks its return:
the eviction is the check. If the session was in the cache, 0-RTT
is accepted and the single-use requirement is satisfied. If not,
the early_data extension is rejected through the normal path so
the record layer correctly skips in-flight 0-RTT records.
WOLFSSL_MSG at each rejection site.
* doc/dox_comments/header_files/ssl.h: document runtime opt-in.
* tests: four new tests —
test_tls13_0rtt_default_off (fails without default-to-0 fix),
test_tls13_0rtt_stateless_replay (fails without TICKET_HAVE_ID
implication and remove_session gate),
test_tls13_remove_session_return (fails without return-value fix),
test_tls13_0rtt_ext_cache_eviction (fails without ext-cache
counts-as-found fix).
test_tls13_early_data explicitly opts in via
wolfSSL_CTX_set_max_early_data.
tests/api.c: two SSL_CTX_remove_session == 0 assertions updated
to == 1.
ReadPemFromBioToBuffer slurps the entire BIO in one shot, so iterative
callers like wolfSSL_PEM_read_bio_X509_CRL (and by extension
wolfSSL_X509_load_crl_file's BIO branch) saw EOF after the first block
and silently dropped every CRL after the first in a multi-CRL bundle.
Refactor wolfSSL_PEM_read_bio_X509_CRL to delegate to
wolfSSL_PEM_X509_X509_CRL_X509_PKEY_read_bio, which already reads one
PEM BEGIN/END pair per call and leaves the BIO positioned just past the
END line. Loop over it so we skip past intervening cert/key blocks and
return the next CRL in the stream — matching OpenSSL's
PEM_read_bio_X509_CRL, verified against OpenSSL 3.0.13 with cases
{cert,CRL}, {CRL,cert}, {CRL,cert,CRL}, {key,CRL}, {CRL,key,CRL}: in
each case OpenSSL skips non-CRL blocks until EOF.
When the caller passes a non-NULL `x` whose `*x` is already populated,
free the previous CRL before overwriting the slot — matching the
d2i_X509_CRL reuse contract the old body relied on.
To keep both helpers visible at the new call site, drop their `static`
qualifier (wolfSSL_PEM_X509_X509_CRL_X509_PKEY_read_bio for the per-block
read, wolfSSL_X509_PKEY_free to free defensively-allocated keys parsed
from intervening non-CRL blocks). Their definitions in src/x509.c and
declarations in wolfssl/internal.h are widened from OPENSSL_ALL to
OPENSSL_EXTRA || OPENSSL_ALL so the OPENSSL_EXTRA-only build (which
compiles wolfSSL_PEM_read_bio_X509_CRL) links cleanly. The unrelated
INFO_read_bio / INFO_read_bio_X509_INFO group below them keeps its
OPENSSL_ALL gate because it depends on wolfSSL_X509_INFO_new/free that
are still OPENSSL_ALL-only.
Also register the previously-orphaned test_wolfSSL_X509_load_crl_file
(its slot in TEST_OSSL_X509_LOOKUP_DECLS was a duplicated
test_wolfSSL_X509_LOOKUP_ctrl_hash_dir entry), update its assertion for
crl2.pem (which already contains two CRLs) to expect 2 instead of 1, and
add a multi-CRL bundle case that builds a memory BIO from
crl.pem + server-cert.pem + crl2.pem and asserts that the reader walks
past the cert and returns all 3 CRLs before NULL.
* fix aes_eax_test() for NO_MALLOC (use WC_*_VAR() to allocate eax context).
* in slhdsa_test(), gate the profusely verbose TestDumpData() clauses on WC_SLHDSA_VERBOSE_DEBUG.
* fix smallstackcache memory leaks in sha256 and sha512 contexts -- don't init or copy over a context that's been inited but not freed, and make sure to explicitly free any context that's been inited or copied over.
* fix uninited-var warnings in slhdsakey_wots_sign(), slhdsakey_xmss_sign(), and slhdsakey_fors_sign() (the uninited-var scenario depends on corrupt arg(s) resulting in zero iterations).
- tls.c: TLSX_CertWithExternPsk_GetSize takes word16*, but length was
widened to word32 in TLSX_GetSize. Use the hsz staging variable like
the other cases so WOLFSSL_CERT_WITH_EXTERN_PSK builds compile.
- tls.c: silence -Wunused-variable for hsz in builds where every case
that consumes it (TLS 1.3, PSK, ETM, early data, PHA, cookie, cert
with extern PSK) is compiled out, e.g. user_settings_tls12.h.
- test_tls_ext.c: assert session->ticketLen > 0 before mutating
ticketAdd in the ticket-age out-of-window test so it fails loudly if
no NewSessionTicket was received.
DoClientTicketCheck's ticket-age bounds (-1000 ms low bound and
MAX_TICKET_AGE_DIFF*1000+1000 ms high bound) were never exercised by
any integration test, so mutations of the constants went undetected.
Establish a TLS 1.3 session, read the NewSessionTicket, then shift the
client's cached ageAdd by well over 1 second so the server's
unobfuscated diff falls outside the valid window on resumption. The
server must reject the PSK — session_reused stays 0.
DoTls13ClientHello enforces RFC 8446 Section 4.1.4 by comparing the
cipher suite in the second ClientHello to the hrrCipherSuite cached on
the server from the HelloRetryRequest. No existing test covers the
mismatch branch, so a deletion of the check would silently allow a
client to switch cipher suite between CH1 and CH2. Drive a partial
handshake until the server has emitted the HRR, then flip the cached
hrrCipherSuite on the server; processing CH2 must surface
INVALID_PARAMETER.
wolfSSL_TicketKeyCb is the built-in ticket callback registered by the
OpenSSL-compat wolfSSL_CTX_set_tlsext_ticket_key_cb API. Its
ConstantCompare of the ticket HMAC was never reached in any test, so a
deletion of the check would silently accept forged tickets. New test
sets up the compat callback, establishes a TLS 1.2 session, saves it,
flips a byte of the encrypted ticket, and asserts the resumption
attempt does not complete.
Cover both branches of TLSX_SecureRenegotiation_Parse's ConstantCompare
against the cached Finished verify_data: a single memio test loops
over client-side and server-side corruption, renegotiates, and
asserts the offending peer surfaces SECURE_RENEGOTIATION_E.
Tls13IntegrityOnly_Decrypt was completely untouched by existing tests,
so any mutation of its ConstantCompare would pass CI. Add a memio
TLS 1.3 handshake over TLS13-SHA256-SHA256 (integrity-only NULL cipher),
then corrupt the final byte of the next record body via an IORecv
wrapper and assert the server surfaces DECRYPT_ERROR.
Cover the Poly1305 ConstantCompare tag check in ChachaAEADDecrypt that
no existing test was hitting (VERIFY_MAC_ERROR never expected in the
suite). A memio-based TLS 1.2 handshake over
ECDHE-RSA-CHACHA20-POLY1305 completes, the server's IORecv is then
replaced with a wrapper that flips the final byte of the next record
body so the forged Poly1305 tag no longer matches. The server's
wolfSSL_read must surface VERIFY_MAC_ERROR.
Covers the HandleResumeHistory check that RFC 7627 Section 5.3 requires:
if the original session used Extended Master Secret, the server MUST
abort when a resumption ClientHello is received without EMS. The new
memio test performs a TLS 1.2 handshake with EMS, saves the session,
disables EMS on a fresh client, resumes with the saved session, and
asserts the server returns EXT_MASTER_SECRET_NEEDED_E.
The CA Names extension size accumulator was a word16. With enough
CA entries (or large DER-encoded names) the running total can wrap
silently, leaving TLSX_CA_Names_Write to overflow an undersized
extension buffer. Match TLSX_SNI_GetSize: use a word32 accumulator
and return 0 when the total exceeds WOLFSSL_MAX_16BIT.
Both TLSX_PreSharedKey_GetSize and TLSX_PreSharedKey_GetSizeBinders
accumulate per-identity bytes into a word16. With enough PSK entries
(or large binderLen/identityLen values) the accumulator wraps silently
and the caller allocates an undersized extension buffer, which
TLSX_PreSharedKey_Write then overflows.
Switch both accumulators to word32 and return LENGTH_ERROR when the
total would exceed the 16-bit wire length field.
Mirror the TLSX_SNI_GetSize pattern: accumulate into a word32 and
return 0 when the aggregate size exceeds WOLFSSL_MAX_16BIT so large
idSz values or many TCA entries no longer silently wrap to a small
value that undersizes the TLSX_TCA_Write output buffer.
Match the TLSX_SNI_GetSize pattern: use a word32 accumulator and return
0 if the aggregate size exceeds WOLFSSL_MAX_16BIT, so a large number of
ALPN entries can no longer silently wrap the length computation.
wolfssl/wolfcrypt/hash.h: when WOLFSSL_NOSHA512_{224,256}, gate out prototypes for wc_Sha512_{224,256}Hash[_ex](), to shift build failures from link-time to compile-time.
.github/workflows/pq-all.yml: for the --enable-sp-math scenario, --disable-quic (QUIC unit tests fail on that combo);
wolfcrypt/test/test.c: add WC_MAYBE_UNUSED to ecdsa_test_deterministic_k_rs(), to fix armel sp-math build.
* allow for fips-dev in v7|ready|dev ENABLED_SHA256_DRBG and ENABLED_SHA512_DRBG setup and change from AC_MSG_WARN to AC_MSG_ERROR if user tries to disable outside fips-dev;
* set ENABLED_SHA512_DRBG=no in lean-aesgcm setup;
wolfcrypt/test/test.c: suppress concurrency-mt-unsafe in myFipsCb();
.wolfssl_known_macro_extras: fix lexical order.
* add "-Wnull-dereference" to all existing "-pedantic -Wdeclaration-after-statement" configs;
* add an --enable-sp-math config to .github/workflows/pq-all.yml and .github/workflows/multi-arch.yml.
Eric Blankenhorn
Lealem Amedie
Juliusz Sosinowicz
David Garske
Daniel Pouzzner
JacobBarthelmeh