Extend the swdev callback to handle ECC operations: keygen, ECDH, sign,
verify, get-size, get-sig-size.
Relax WOLF_CRYPTO_CB_ONLY_ECC guards in the test suite so that tests run
under swdev, and wire wc_SwDev_Init/Cleanup into testsuite, client, and
server. Two tests are intentionally kept excluded even with swdev because
they call raw ECC math primitives (wc_ecc_mulmod, on-curve validation in
wc_ecc_import_x963) that are stripped below the cryptocb dispatch layer:
- ecc_mulmod_test (wolfcrypt/test/test.c)
- test_wc_ecc_import_x963_off_curve (tests/api/test_ecc.c)
Windows test code pre-picked a random port via GetRandomPort() (returning
a value in [49152, 65535]) before calling bind(), with no check that the
port was free and no retry on collision. Under load this occasionally
collided with an already-bound port and aborted the test with
"tcp bind failed", producing intermittent Jenkins failures (e.g. PRB
windows-test-v2 #17140 in the OCSP responder test).
The Unix path already does the right thing: bind to port 0 (OS-assigned
ephemeral) and read the port back via getsockname(). The same primitives
exist in Winsock 1.1, so drop the USE_WINDOWS_API guard around the
getsockname block in tcp_listen()/udp_accept() and remove the per-caller
GetRandomPort() workarounds in the OCSP responder, server example, and
the api.c / test_ossl_bio.c test sites. socklen_t is already typedef'd
as int on Windows in test.h.
GetRandomPort() itself is left in place since it is a static inline in a
shipped public test header.
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.
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.
wolfssl/ssl.h, examples/client/client.c, examples/server/server.c, src/bio.c, tests/api.c: add error tracing for WOLFSSL_SHUTDOWN_NOT_DONE.
tests/api.c: in test_wolfSSL_read_write_ex(), use WOLFSSL_SUCCESS rather than 1 for expected-success wolfSSL_shutdown()s, and add note that the wrong value is being returned (the test currently always fails, which is masked by an always-success retval).
* 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
* DTLS 1.3 cookie and CH frag handling
* static memory handling
* Fix memory leak in TLS server PQC handling in case of ECH
* Make sure hybrids are actually tested in testsuite
OCSP Responder Core API:
- Add new public API for creating and managing an OCSP responder
- Add public wrappers for internal OCSP request/response functions
- OcspRespCheck: fix check when authorized responder is loaded into CM
Header Cleanup:
- Remove circular dependency when including `#include <wolfssl/wolfcrypt/asn.h>` from wolfssl/wolfcrypt/ecc.h and wolfssl/wolfcrypt/rsa.h
OCSP Responder Example (examples/ocsp_responder/):
- Add a command-line OCSP responder for interoperability testing with OpenSSL's `openssl ocsp` client
Test Scripts (scripts/):
- ocsp-responder-openssl-interop.test: Tests wolfSSL OCSP responder with `openssl ocsp` client
- ocsp-stapling-with-wolfssl-responder.test: Tests wolfSSL OCSP responder when doing OCSP stapling
Certificate Infrastructure (certs/ocsp/):
- Add DER-format certificates and keys for OCSP testing
- Update renewcerts.sh to generate DER versions
Known Limitations (documented in src/ocsp.c header comment):
- Single request/response per OCSP exchange only
- Key-hash responder ID only (no name-based responder ID)
- No singleExtensions support
When both TLS 1.3 and Brainpool curves are enabled, three new groups can
be used for the ECDHE key exchange according to RFC 8734:
* WOLFSSL_ECC_BRAINPOOLP256R1TLS13 (31)
* WOLFSSL_ECC_BRAINPOOLP384R1TLS13 (32)
* WOLFSSL_ECC_BRAINPOOLP512R1TLS13 (33)
Also ensure that the existing TLS 1.2 curves are sent properly.
The TLS client application is updated to support handshakes via
Brainpool curves using the new argument "--bpKs".
add WC_GENERATE_SEED_DEFAULT, which defaults to wc_GenerateSeed if not overridden, and replace wc_GenerateSeed with WC_GENERATE_SEED_DEFAULT in various calls to wc_SetSeed_Cb();
linuxkm/linuxkm_wc_port.h: if FIPS <v6 and RDSEED, define WC_LINUXKM_RDSEED_IN_GLUE_LAYER and define WC_GENERATE_SEED_DEFAULT wc_linuxkm_GenerateSeed_IntelRD;
wolfcrypt/test/test.c: update rng_seed_test() with gating and vectors for FIPS v5 with HAVE_AMD_RDSEED or HAVE_INTEL_RDSEED;
wolfssl/wolfcrypt/types.h: add WC_HAVE_VECTOR_SPEEDUPS helper macro, and enlarge fallthrough definition coverage for DISABLE_VECTOR_REGISTERS.
Add support for X25519 and X448 based hybrid PQC + ECC key exchange
groups. Furthermore, two new combinations with SECP curves are added to
match OQS combinations.
This also incorporates the changed order of X25519 and X448 based
combinations to place the PQC material before the ECDH material. This is
motivated by the necessity to always have material of a FIPS approved
algorithm first.
Also, codepoints are updated to reflect the latest draft standards for
pure ML-KEM and some of the hybrids. With these changes and based on the
recent additions to both enable ML-KEM final and draft versions
simultaneously, a WolfSSL TLS server is now compatible with all recent
browsers that support either the draft version of ML-KEM (Chromium based
browsers and Firefox < version 132; only when the draft version is
enabled in the build) or the final version already (Firefox > version 132).
In the process of extending support, some code and logic cleanup
happened. Furthermore, some memory leaks within the hybrid code path have
been fixed.
Signed-off-by: Tobias Frauenschläger <tobias.frauenschlaeger@oth-regensburg.de>
* Fix for `TlsSessionCacheGetAndLock` that was not checking the sessionIDSz, so could return a pointer to an invalid session (if 0's). Resolves issue with `test_wolfSSL_CTX_sess_set_remove_cb` test.
* Fix cast warning with `HAVE_EX_DATA` in Windows VS.
* Fix openssl_extra without PKCS12.
* Refactor the EX data crypto and session API's to gate on `HAVE_EX_DATA_CRYPTO`.
* Grouped the EX data API's in ssl.h
* Moved API's in ssl.h to separate the compatibility ones from ours.
.github/workflows/codespell.yml: remove */README_jp.txt from "skip" list.
IDE/Renesas/cs+/Projects/t4_demo/README_jp.txt: convert from SHIFT_JIS to UTF-8.
cmake/options.h.in: use "#cmakedefine HAVE_PTHREAD 1" to avoid conflict with config.h.
configure.ac: add --enable-c89, and remove !ENABLED_OPENSSLEXTRA dependency from AM_CONDITIONAL([BUILD_CRYPTONLY],...).
wolfcrypt/src/asn.c: refactor SetOthername() for efficiency, and add PRAGMA_GCC to suppress false positive -Wstringop-overflow associated with -fstack-protector.
wolfssl/wolfcrypt/rsa.h: add WC_ prefixes to RSA_PKCS1_PADDING_SIZE and RSA_PKCS1_OAEP_PADDING_SIZE, and define unprefixed compat aliases only if !OPENSSL_COEXIST.
wolfssl/wolfcrypt/types.h:
#ifdef WOLF_C89, #define WC_BITFIELD unsigned;
enhance WOLF_ENUM_DUMMY_LAST_ELEMENT() to include the line number, to construct unique labels given a per-filename argument, to accommodate anonymous enums.
examples/asn1/asn1.c:
examples/client/client.c:
examples/pem/pem.c:
examples/server/server.c:
wolfcrypt/src/sp_dsp32.c:
wolfcrypt/src/wc_port.c:
wolfssl/test.h:
use XMALLOC/XREALLOC/XFREE consistently, not malloc/realloc/free.
wolfcrypt/benchmark/benchmark.c:
wolfcrypt/src/memory.c:
wolfcrypt/test/test.c:
wolfssl/wolfcrypt/mem_track.h:
wolfssl/wolfcrypt/settings.h:
wolfssl/wolfcrypt/types.h:
annotate intentional native heap access with "/* native heap */".
wolfcrypt/src/asn.c:
wolfssl/callbacks.h:
wolfssl/openssl/ec.h:
wolfssl/openssl/ssl.h:
wolfssl/wolfcrypt/hpke.h:
wolfssl/wolfcrypt/types.h:
fix enum trailing commas.
wolfssl/openssl/ec.h:
wolfssl/openssl/evp.h:
wolfssl/openssl/rsa.h:
wolfssl/openssl/ssl.h:
use WC_BITFIELD in bitfield elements, not byte or word16, to allow for pedantic C89 conformant builds.
wolfssl/openssl/ec.h:
wolfssl/openssl/evp.h:
wolfssl/openssl/pem.h:
wolfssl/openssl/ssl.h:
wolfssl/wolfcrypt/logging.h:
avoid variadic macros wherever possible, and where unavoidable, #ifdef WOLF_NO_VARIADIC_MACROS, define them with empty arg lists, rather than ..., to support Watcom compiler.
wolfssl/wolfcrypt/settings.h: if defined(__WATCOMC__), define WOLF_NO_VARIADIC_MACROS.
Make Kyber and ML-KEM individually available as well as at the same
time.
Modified TLS layer to support both Kyber and ML-KEM.
Added new identifiers in TLS layer for ML-KEM.
configure.ac:
* add --enable-all-osp to separate OSP meta-feature sets from --enable-all, allowing --enable-all --disable-all-osp --disable-opensslall (e.g. for testing OPENSSL_COEXIST).
* fix enable_all_crypto=yes in enable-all to be conditional on "$enable_all_crypto" = "".
* move enable_rsapss=yes from enable-all to enable-all-crypto.
examples/ and testsuite/: #undef OPENSSL_COEXIST unconditionally rather than only if defined(OPENSSL_EXTRA), to capture -DOPENSSL_EXTRA_X509_SMALL or any other such variants.
No functional changes.
Several compat symbols that were formerly enums are now macros.
All library source is refactored to use only native symbols in all code gated in with --enable-all-crypto --enable-opensslextra.
wolfcrypt/test/test.c is similarly refactored to use only native symbols.
examples/ and tests/ are unmodified except for header setup to disable OPENSSL_COEXIST and TEST_OPENSSL_COEXIST.
David Garske
rizlik
Tobias Frauenschläger
Juliusz Sosinowicz
Daniel Pouzzner
JacobBarthelmeh
sebastian-carpenter
Kareem
effbiae
Anthony Hu
Eric Blankenhorn
Sean Parkinson
Tobias Frauenschläger
jordan