git: ebd1ab3805d3 - main - security/wolfssl: Update to 5.9.2
- Go to: [ bottom of page ] [ top of archives ] [ this month ]
Date: Sat, 27 Jun 2026 09:08:02 UTC
The branch main has been updated by fox:
URL: https://cgit.FreeBSD.org/ports/commit/?id=ebd1ab3805d31a82613a1db818e6dc2bf5cdf75d
commit ebd1ab3805d31a82613a1db818e6dc2bf5cdf75d
Author: Santhosh Raju <fox@FreeBSD.org>
AuthorDate: 2026-06-27 08:43:03 +0000
Commit: Santhosh Raju <fox@FreeBSD.org>
CommitDate: 2026-06-27 08:46:05 +0000
security/wolfssl: Update to 5.9.2
Removed patches which has been merged by upstream
- https://github.com/wolfSSL/wolfssl/pull/10097
Removed autoreconf, since it is no longer required
Changes since 5.9.1:
To download the release bundle of wolfSSL visit the download page at
www.wolfssl.com/download/
PR stands for Pull Request, and PR references a GitHub pull request
number where the code change was added.
NOTE: The pre-standardization Dilithium API has been renamed to its
FIPS 204 ML-DSA name; the legacy dilithium.h header and
wc_dilithium_* names remain available through a temporary
compatibility shim.
NOTE: The SLH-DSA Hash sign/verify APIs now require a caller-supplied
pre-hashed digest rather than the raw message (see Enhancements
below).
NOTE: liboqs integrations for ML-KEM, ML-DSA, and SLH-DSA (SPHINCS+)
have been removed in favor of the native implementations; the
deprecated liblms and libxmss integrations have also been removed.
NOTE (BREAKING RFC 6960 4.2.2.2): OCSP responder authorization is
now strictly enforced. Removes the non-compliant
CheckOcspResponderChain() fallback, which authorized any OCSP
responder cert issued by an ancestor of the target's issuer; RFC 6960
4.2.2.2 requires direct issuance by the CA identified in the request.
Also removes the now-unused WOLFSSL_NO_OCSP_ISSUER_CHAIN_CHECK macro
and the vp parameter from CheckOcspResponder().
Vulnerabilities
* [High CVE-2026-11310] X.509 trust-chain bypass in the OpenSSL
compatibility certificate verifier (wolfSSL_X509_verify_cert()).
This affects only builds with --enable-opensslextra
(OPENSSL_EXTRA) and whose application validates certificates by
calling X509_verify_cert() (OpenSSL compatibility layer function)
with caller-supplied untrusted intermediate certificates; for
those users it is critical, otherwise the library is unaffected.
In particular, native wolfSSL TLS/DTLS usage is not impacted.
wolfSSL's X509_verify_cert() temporarily loads each
caller-supplied untrusted intermediate into the certificate
manager but failed to drop them before the trusted-store check,
so an untrusted intermediate could anchor the path itself. An
attacker can present a chain that never reaches a configured
trust anchor and have it accepted, resulting in acceptance of an
attacker-controlled certificate. This is certificate verification
independent of TLS (e.g. S/MIME/CMS, code/firmware signing,
JWT/JWS x5c), is not specific to any key type or algorithm, and
a single untrusted intermediate suffices. The default wolfSSL TLS
handshake (WOLFSSL_VERIFY_PEER) is not affected; only TLS
applications doing manual or deferred peer verification through
this API are, which also requires --enable-sessioncerts.
Affected: v5.8.4, v5.9.0 and v5.9.1 (introduced by commit
025dbc34); v5.8.2 and earlier are not. Thanks to Corban Villa,
Sohee Kim and Austin Chu (UC Berkeley, Sky Lab). Fixed in
PR 10674.
* [High CVE-2026-11999] X.509 trust-chain bypass (path-depth
exhaustion) in the OpenSSL compatibility certificate verifier
(wolfSSL_X509_verify_cert()). This affects only builds with
--enable-opensslextra whose application calls X509_verify_cert()
with caller-supplied untrusted intermediates; for those users it
is critical, otherwise the library is unaffected. Native wolfSSL
TLS/DTLS usage is not impacted. X509_verify_cert() returned
success based only on the last verified link rather than on
reaching a trust anchor: when the supplied chain is deeper than
the verifier's maximum path depth (default 100), path building
runs out of depth while still walking untrusted intermediates and
the chain is accepted even though it never reaches a configured
trust anchor, allowing acceptance of an attacker-controlled
certificate. The default TLS handshake (WOLFSSL_VERIFY_PEER) is
not affected; only applications doing manual or deferred
verification through this API are. Affected versions: v5.7.4
through v5.9.1, introduced in commit 17c9e92b7 (first released in
v5.7.4); v5.7.2 and earlier are not affected. Thanks to Corban
Villa, Sohee Kim and Austin Chu (UC Berkeley, Sky Lab). Fixed in
PR 10674.
* [High CVE-2026-6679] A heap buffer overflow could occur in the
DTLS 1.3 ACK serialization path before the connecting peer is
authenticated. The buffer overflow was due to an integer
truncation when computing the length of the ACK record-number
list, causing an undersized buffer to be allocated and then
overrun. This affects builds using DTLS 1.3 and wolfSSL version
5.9.0 and earlier. A fix was added to the 5.9.1 release. Thanks
to Nicholas Carlini from Anthropic for the report. Fixed in
PR 10116.
* [High CVE-2026-55958] Out-of-bounds write in the Renesas TSIP
TLS 1.3 transcript buffer. In tsip_StoreMessage() the capacity
check guarding the fixed message bag (MSGBAG_SIZE) sets an error
code but fails to return, so execution falls through to an
XMEMCPY that writes past the end of the buffer once the
accumulated TLS 1.3 handshake transcript exceeds MSGBAG_SIZE
(8 KB), corrupting adjacent heap state and potentially causing a
remote denial of service crash. The bag is sized to hold a normal
handshake, so this is reached only by an unusually large but
valid certificate chain, or by a malicious or man-in-the-middle
server sending an oversized handshake message to a client that
does not strictly verify the chain. This only affects builds
using the Renesas TSIP TLS port (WOLFSSL_RENESAS_TSIP_TLS) as a
TLS 1.3 client on Renesas MCUs with TSIP hardware enabled, and
is rated High within those builds. All other configurations are
unaffected. Thanks to NVIDIA Project Vanessa for the report.
Fixed in PR 10705.
* [High CVE-2026-55960] Un-negotiated Raw Public Key (RFC 7250)
accepted in place of an X.509 certificate, bypassing chain
validation. A raw public key has no chain, so
ParseCertRelative() accepts it without performing any trust
verification; it must therefore only be accepted when RPK was
actually negotiated for that peer. The check now defaults the
expected type to X.509 (per RFC 7250/8446) when no type was
negotiated, comparing against the received server certificate
type on the client and the selected client certificate type on
the server, and rejects any mismatch, including an un-negotiated
raw public key, with UNSUPPORTED_CERTIFICATE. Only affects builds
with Raw Public Key support (HAVE_RPK) enabled - disabled by
default in a standalone build, but included in --enable-all.
Thanks to NVIDIA Project Vanessa for the report. Fixed in
PR 10702.
* [High CVE-2026-55961] wolfSSL_PKCS7_verify() returning success
for a degenerate (certs-only) PKCS#7 object that contains no
signer. Such an object has empty signerInfos, so the underlying
signed-data verification succeeds without authenticating any
content. The compatibility-layer verify path now rejects the
object when no signer signature has actually been verified, so a
PKCS#7 carrying no valid signature is no longer reported as
verified. This is enforced regardless of the PKCS7_NOVERIFY
flag, which only suppresses signer certificate chain validation
and was never intended to waive the requirement that a signature
exist. Only affects OpenSSL compatibility builds that call the
PKCS7_verify() compatibility API on potentially degenerate PKCS#7
bundles. Thanks to NVIDIA Project Vanessa for the report. Fixed
in PR 10702.
* [Med CVE-2026-6731] X.509 name constraint bypass via the Subject
Common Name when treated as a DNS-type name. A certificate whose
Subject CN violates an issuing CA's DNS name constraints could be
accepted. Thanks to d0sf3t (Aradex) for the report. Fixed in
PR 10223.
* [Med CVE-2026-6091] Partial-chain certificate verification may
accept chains that terminate at a peer-supplied, untrusted
intermediate certificate rather than a trusted anchor. An
attacker could present a chain that ends at an intermediate they
control and have it accepted as valid. Thanks to Dikai Zou for
the report. Fixed in PR 10170.
* [Med CVE-2026-6094] Heap buffer overread in
wc_PKCS7_DecodeEnvelopedData when parsing crafted PKCS7
EnvelopedData. This could theoretically be triggered by
attacker-supplied data delivered via S/MIME or CMS. Thanks to
Dikai Zou for the report. Fixed in PR 10128.
* [Med CVE-2026-6329] PKCS#12 MAC verification uses an
attacker-controlled comparison length, weakening the integrity
check on the MAC and allowing a mismatched MAC to be accepted.
Thanks to Nicholas Carlini from Anthropic for the report. Fixed
in PR 10192.
* [Med CVE-2026-6330] The ML-KEM ARM64 NEON ciphertext comparison
only compares half of the input, breaking the Fujisaki-Okamoto
transform's implicit rejection and weakening IND-CCA2 security
on that code path. Thanks to Nicholas Carlini from Anthropic for
the report. Fixed in PR 10192.
* [Med CVE-2026-8720] wc_Blake2bHmacFinal and wc_Blake2sHmacFinal
discard the message when the key length exceeds the block size,
producing a MAC that is independent of the input. This bug is
specific to the HMAC-BLAKE2 API's that were added in wolfSSL
version 5.9.0. Fixed in PR 10447.
* [Med CVE-2026-10097] ML-KEM-1024 x64 AVX2 implicit rejection
failure in the Fujisaki-Okamoto transform breaks IND-CCA2
security, allowing decapsulation to deviate from the
implicit-rejection behavior required by the standard. Thanks to
007bsd for the report. Fixed in PR 10430.
* [Med CVE-2026-10098] OCSP CertID serial-number
length-confusion in wolfSSL_OCSP_resp_find_status allows a
same-issuer SingleResponse whose serial is a prefix of the
target serial to be reported as the revocation status of a
different certificate. Thanks to Kim Youngjoon (Team-Atlanta and
Georgia Institute of Technology) for the report. Fixed in
PR 10554.
* [Med CVE-2026-10592] Certificates with wildcard DNS SANs
(e.g. *.example.com) bypassed CA name-constraint checks. Thanks
to tonghuaroot for the report. Fixed in PR 10549.
* [Med CVE-2026-7532] iPAddress name constraints bypass when
WOLFSSL_IP_ALT_NAME is not defined. IP address name constraints
are not enforced in that configuration, allowing a certificate to
bypass an issuing CA's IP address constraints. Thanks to Ankur
Tyagi of Cisco Talos (TALOS-2026-2409) for the report. Fixed in
PR 10354.
* [Med CVE-2026-6291] Bleichenbacher padding oracle in PKCS#7
KTRI decryption. When decrypting PKCS#7 EnvelopedData using RSA
PKCS#1 v1.5 key transport, wolfSSL returned distinguishable error
codes depending on whether RSA padding validation failed versus
whether the decrypted content was malformed. An attacker able to
submit crafted EnvelopedData messages and observe error responses
could use this as a padding oracle to incrementally recover the
encrypted Content Encryption Key (CEK). The fix generates a
deterministic pseudo-random fake CEK on padding failure (via
HMAC-SHA256) and proceeds with decryption identically, using
constant-time operations throughout, so that all failure paths
produce the same error regardless of padding validity. Found with
internal wolfSSL review. Fixed in PR 10203.
* [Med CVE-2026-7511] PKCS7_verify signer confusion allows forged
signatures, where the signer associated with a signature is not
correctly bound, permitting a forged signature to be accepted.
Thanks to Nicholas Carlini from Anthropic for the report. Fixed
in PR 10203.
* [Med CVE-2026-11703] Fixed missing SNI/ALPN binding on stateful
(session-ID) resumption, which previously skipped the binding
check performed for ticket-based resumption. A cached session
could be resumed under a different SNI/ALPN than originally
negotiated and, where client-authentication policy differs across
virtual hosts, carry the cached peer-authentication state into a
context it was not established for. Resumption now verifies the
SNI/ALPN binding for all paths and declines (falling back to a
full handshake) on mismatch. Thanks to Dikai Zou for the report.
Fixed in PR 10489.
* [Med CVE-2026-55962] TLS 1.3 post-handshake authentication
(PHA) issue where a server could accept a client's Finished
message without the client having sent a Certificate and
CertificateVerify. The post-handshake-auth exemption that allows
an empty/absent peer certificate was only intended for the
initial handshake, but it was also being applied while a
post-handshake CertificateRequest was still outstanding. The
check is now scoped to the initial handshake only: on the server,
once a post-handshake CertificateRequest has been sent
(certReqCtx is set), a peer certificate and a valid
CertificateVerify are required again before the Finished is
accepted, with empty-certificate handling following the
configured verify mode (FAIL_IF_NO_PEER_CERT) just as during
first-handshake client authentication. Only affects TLS 1.3
servers built with post-handshake authentication support
(WOLFSSL_POST_HANDSHAKE_AUTH / --enable-postauth, included in
--enable-all) that enable WOLFSSL_VERIFY_POST_HANDSHAKE and
request a client certificate after the handshake via
wolfSSL_request_certificate(). Clients, and servers that do not
use post-handshake authentication, are unaffected. Thanks to
NVIDIA Project Vanessa for the report. Fixed in PR 10702.
* [Med CVE-2026-55964] Chain intermediate CA:TRUE without
keyCertSign accepted as a signing CA. Intermediate CA
certificates are required to have the keyCertSign key usage when
a Key Usage extension is present, but chain-supplied temporary
CAs (WOLFSSL_TEMP_CA) added while building a certificate path
were previously exempted from this check, so an intermediate
asserting CA:TRUE but lacking keyCertSign was accepted as a
signing CA. The check now applies to chain-supplied temporary CAs
as well; only operator-loaded root certificates (WOLFSSL_USER_CA)
and self-signed roots remain exempt. Per RFC 5280 an absent Key
Usage extension implies all usages, so the requirement is
enforced only when the extension is actually present
(extKeyUsageSet). Affects the OpenSSL-compatibility
certificate-path-building path (X509_verify_cert / X509_STORE,
OPENSSL_EXTRA/OPENSSL_ALL), where untrusted chain intermediates
are added as temporary CAs; native (non-OpenSSL-compat)
certificate verification does not create temporary CAs and is
unaffected. Within those builds, the check applies unless
ALLOW_INVALID_CERTSIGN is defined. Thanks to NVIDIA Project
Vanessa for the report. Fixed in PR 10702.
* [Low CVE-2026-6092] When HAVE_ENCRYPT_THEN_MAC is configured,
the implementation could fall back to MAC-then-Encrypt rather
than enforcing Encrypt-then-MAC. Thanks to Marcin Olejnik
(Rockwell Automation) for the report. Fixed in PR 10167.
* [Low CVE-2026-6331] HMAC zero-length tag forgery in
EVP_DigestVerifyFinal, where a zero-length tag could be accepted
as valid during HMAC verification. Thanks to Nicholas Carlini
from Anthropic for the report. Fixed in PR 10192.
* [Low CVE-2026-6681] The PKCS#7 decode path ignores the
caller-supplied output buffer size (outputSz), allowing decoded
content to be written past the bounds of the provided buffer.
This affects wolfSSL 5.9.0 and earlier and was fixed in the
5.9.1 release. Thanks to Nicholas Carlini from Anthropic for the
report. Fixed in PR 10116.
* [Low CVE-2026-10512] The X25519 x86_64 assembly implementation
fails to clear the most significant bit during the final modular
reduction, so the computed result may not be fully reduced modulo
the field prime 2^255 - 19. This can leave the field element in
a non-canonical form, producing an incorrect result from the
scalar multiplication and potentially a wrong shared secret.
Thanks to Haruki Oyama for the report. Fixed in PR 10536.
* [Low CVE-2026-6678] Integer underflow in wc_PKCS7_DecryptOri
when handling crafted Other Recipient Info, leading to incorrect
length handling during decryption. Thanks to Dikai Zou for the
report. Fixed in PR 10203.
* [Low CVE-2026-7531] Use-after-free in PQC hybrid key-share
handling. This is an incomplete-fix follow-up to CVE-2026-5460
(released in 5.9.1): a malicious TLS 1.3 server sending a
truncated PQC hybrid KeyShare can still trigger the error cleanup
path to operate on freed memory. Thanks to Thai Duong
(Calif.io / Anthropic) for the report. Fixed in PR 10327.
* [Low CVE-2026-6325] Out-of-bounds write in
SetSuitesHashSigAlgo when processing an oversized signature
algorithms list, allowing a write past the bounds of the
destination buffer. Thanks to Muhammad Arya Arjuna Habibullah
(Pelioro) for the report. Fixed in PR 10204.
* [Low CVE-2026-6412] Certificate policy and RFC 8446 compliance
concerns regarding the continued acceptance of SHA-1/MD5 in
certificate processing. Thanks to Xiangdong Li (Student, Beijing
University of Posts and Telecommunications [BUPT]) for the
report. Fixed in PR 10222.
* [Low CVE-2026-6450] A CRL critical extension bypass exists in
ParseCRL_Extensions where critical extensions are not properly
enforced, allowing a crafted CRL with an unhandled critical
extension to be accepted. This only affects builds with CRL
support enabled and where a crafted CRL had a trusted signature
when parsed. Thanks to Oleh Konko (@1seal) for the report.
Fixed in PR 10239.
* [Low CVE-2026-12340] Out-of-bounds heap read during SM2/SM3
certificate signature verification. When parsing a certificate
with an SM3wSM2 signature, the Subject Key Identifier computation
reads the trailing 65 bytes of the public key without checking
that the key is at least that long. A public key shorter than
65 bytes results in an out-of-bounds heap read, leading to a
potential crash (denial of service); there is no out-of-bounds
write. Note this only affects builds with SM2 support
(--enable-sm2 or --enable-all). Thanks to David Pokora, Trail of
Bits (in collaboration with Anthropic). Fixed in PR 10641.
* [Low CVE-2026-55967] AES-GCM encryption/decryption with
extremely large cumulative single message sizes (>64 GiB) were
not properly rejected by the streaming APIs, allowing counter
wrap, keystream reuse, and consequent plaintext recovery. Thanks
to NVIDIA Project Vanessa for the report. Fixed in PR 10709.
Enhancements
* BREAKING (FIPS 205 SLH-DSA): wc_SlhDsaKey_SignHash,
wc_SlhDsaKey_SignHashDeterministic,
wc_SlhDsaKey_SignHashWithRandom, and wc_SlhDsaKey_VerifyHash now
take the caller-pre-hashed message digest via hash/hashSz
parameters (renamed from msg/msgSz), aligned with ML-DSA's
wc_dilithium_sign_ctx_hash / wc_dilithium_verify_ctx_hash
semantics, and NIST ACVP signatureInterface=external /
preHash=preHash test vectors. hashSz must equal
wc_HashGetDigestSize(hashType) (32 bytes for SHAKE128, 64 bytes
for SHAKE256 per FIPS 205 Section 10.2.2); otherwise
BAD_LENGTH_E is returned. Migration: hash the message yourself
before the call (callers using positional arguments are
source-compatible; only the parameter names changed). Caveat:
callers who today pass a raw message whose length happens to
equal the digest size for the chosen hashType (e.g., signing a
32-byte handle/IV/seed with WC_HASH_TYPE_SHA256) will not trip
BAD_LENGTH_E; the resulting signature is syntactically valid but
is over the wrong bytes. The pre-existing
wc_SlhDsaKey_SignMsgDeterministic and
wc_SlhDsaKey_SignMsgWithRandom retain their M'-supplied-directly
contract (FIPS 205 internal interface, Algorithm 19); their input
validation is hardened with the same NULL/length/MISSING_KEY
checks as the *Hash* family. wc_SlhDsaKey_VerifyMsg is
unchanged. All three gain doxygen coverage. (PR 10450, PR 10465)
* Behavioral change (RSA-PSS trailerField enforcement):
DecodeRsaPssParams (and its public wrapper wc_DecodeRsaPssParams)
now enforces RFC 8017 A.2.3, which mandates trailerField ==
trailerFieldBC(1). In the default build (i.e., without
WOLFSSL_NO_ASN_STRICT), any certificate or CMS/PKCS#7 structure
whose RSA-PSS parameters contain a trailerField value other than
1 is now rejected with ASN_PARSE_E. Previously, any positive
integer value was silently accepted. This affects all call paths
that decode RSA-PSS algorithm parameters, including X.509
certificate parsing and PKCS#7 signature verification. Users who
need to interoperate with non-conformant peers can define
WOLFSSL_NO_ASN_STRICT to restore the previous permissive
behavior. (PR 10595)
* Renamed the post-quantum signature implementation from its
pre-standardization name Dilithium to its NIST-standardized name
ML-DSA (FIPS 204), mirroring the earlier Kyber -> ML-KEM rename
in wc_mlkem.{h,c}. The legacy <wolfssl/wolfcrypt/dilithium.h>
header, dilithium_key type, wc_dilithium_* / wc_Dilithium_*
functions, and HAVE_DILITHIUM / WOLFSSL_DILITHIUM_* /
WC_DILITHIUM_* build gates remain available through a temporary
compatibility shim, so application code keeps compiling
unchanged. See doc/dilithium-to-mldsa-migration.md for the full
list of renamed symbols, the new WOLFSSL_MLDSA cmake option /
--enable-mldsa configure switch, and the migration steps for
moving consumer code to the canonical API. (PR 10436, PR 10497,
PR 10516)
* TLS 1.3: zero traffic key staging buffers in SetKeysSide() once
a CryptoCB callback has imported the AES key into a Secure
Element (aes->devCtx != NULL). Clears
keys->{client,server}_write_key on the provisioned side(s) after
cipher init succeeds. The static IV buffers
(keys->{client,server}_write_IV, keys->aead_{enc,dec}_imp_IV)
are intentionally left intact because BuildTls13Nonce() reads
them on every AEAD record to construct the per-record nonce.
Scoped to TLS 1.3, non-DTLS, non-QUIC; requires WOLF_CRYPTO_CB
and WOLF_CRYPTO_CB_AES_SETKEY. (PR 10246)
* The wc_AesCmacVerify/_ex API were hardened to more closely
conform to NIST SP 800-38B MAC length guidance, and these verify
functions will now correctly enforce bounds on tag length checks.
As a result, Cmac verification that previously were erroneously
passing will now return MAC_CMP_FAILED_E or BAD_FUNC_ARG.
(PR 10462)
New Features
* Added wolfCrypt SRAM PUF (Physically Unclonable Function)
support, deriving device-unique keys from SRAM power-on state
using a BCH fuzzy extractor and HKDF
(wc_PufInit/wc_PufEnroll/wc_PufReconstruct). by @dgarske
(PR 10066)
* Added SHE (Secure Hardware Extension) support to wolfCrypt with
software CMD_LOAD_KEY message generation/verification (M1-M5)
and optional crypto callback hardware offload. by @night1rider
(PR 10009)
* Added WOLF_CRYPTO_CB_SETKEY and WOLF_CRYPTO_CB_EXPORT_KEY
generic crypto callbacks to bridge raw key bytes to and from
hardware key stores. by @night1rider (PR 9851)
* Added key id/label constructors
(wc_InitCmac_Id/_Label, wc_AesNew_Id/_Label,
wc_NewRsaKey_Id/_Label) so algorithms can forward a hardware
key-slot identifier. by @night1rider (PR 10072)
* Added RFC 8773(bis) cert_with_extern_psk support for TLS 1.3,
with API and handshake tests. by @Frauschi (PR 10085)
* Added native TLS-ALPN-01 ACME challenge certificate support via
the RFC 8737 id-pe-acmeIdentifier extension. by @lealem47
(PR 10334)
* Extended WOLFSSL_SP_NONBLOCK to RSA and Diffie-Hellman for the
C/Small 2048/3072/4096 backends so handshakes never block for
long on a single big-integer operation. by @dgarske (PR 10394)
* Added ML-KEM and ML-DSA support to the C# wrapper. by @dgarske
(PR 10191)
* Added an HPKE (RFC 9180) C# wrapper. by @dgarske (PR 10171)
* Added --enable-wolfzfs support for the wolfCrypt OpenZFS patch
in both kernel-space and user-space builds. by @philljj
(PR 10397)
* Added a minimal DTLS 1.3 client-only build via WOLFSSL_DTLS_ONLY
and an --enable-dtls13 autoconf cascade. by @julek-wolfssl
(PR 10353)
* Raised the --enable-context-extra-user-data ex_data index limit
to 9999 and increased external cookie/extra-user-data maximum
sizes for large-scale deployments. by @Roy-Carter (PR 10236)
* Published wolfSSL's external Security Policy and a structured
Vulnerability Report Template
(SECURITY-POLICY.md, SECURITY-REPORT-TEMPLATE.md).
by @ColtonWilley (PR 10284)
* Extended the OpenSSL compatibility layer for libevent
integration, adding struct-tag compatibility defines,
BIO_get_init, and short-form alert string helpers.
by @Roy-Carter (PR 10158, PR 10160)
* Add support for WOLF_CRYPTO_CB_ONLY_SHA512. by @rizlik (PR 10550)
* Support RFC 9802 LMS and XMSS in X.509 certificate and CSR
generation. by @Frauschi (PR 10572)
Post-Quantum Cryptography (PQC)
* Added a SHA-512 DRBG and FIPS module-boundary wrappers for
ML-KEM, ML-DSA, LMS, XMSS, and SLH-DSA as part of the upcoming
post-quantum FIPS submission. by @kaleb-himes (PR 9843)
* Replaced the liboqs-based pre-standardization SPHINCS+ with the
native FIPS 205 SLH-DSA implementation throughout the
certificate/ASN.1/X.509 layers. by @Frauschi (PR 10261)
* Removed the liboqs integrations for ML-KEM and ML-DSA in favor
of the native implementations. by @Frauschi (PR 10293)
* Removed the deprecated liblms and libxmss integrations now that
wolfCrypt has its own LMS/XMSS implementations. by @Frauschi
(PR 10292)
* Wired RFC 9802 HSS/LMS and XMSS/XMSS^MT stateful hash-based
signatures into X.509 certificate verification. by @Frauschi
(PR 10406)
* Added crypto callbacks (MakeKey/Sign/Verify/SigsLeft) and
id/label support for LMS and XMSS. by @padelsbach (PR 10380)
* Added CryptoCb support for SLH-DSA. by @Frauschi (PR 10466)
* Added dynamic key allocation for ML-KEM
(WOLFSSL_MLKEM_DYNAMIC_KEYS) to right-size key buffers and
reduce handshake memory use on constrained systems. by @Frauschi
(PR 10179, PR 10206)
* Added dynamic key allocation for ML-DSA
(WOLFSSL_DILITHIUM_DYNAMIC_KEYS). by @Frauschi (PR 10180)
* Added ML-KEM support for PKCS#11 via the PKCS#11 3.2
encapsulate/decapsulate interface. by @Frauschi (PR 10077)
* Added ML-DSA to X509_get_pubkey and EVP_PKEY_base_id.
by @kojo1 (PR 9965)
* Added ML-DSA SPKI/PKCS#8 DER support to
d2i_PUBKEY/d2i_PrivateKey, EVP_PKCS82PKEY, and
X509_check_private_key. by @cconlon (PR 10310, PR 10483)
* Added Ed25519/Ed448 support to the EVP_PKEY layer and fixed a
d2i key-probe bug affecting RSA/ECC/DH. by @lealem47 (PR 10135)
* Fixed PQC key exchange when a ClientHello offers multiple KEM
key shares. by @Frauschi (PR 10299)
* Fixed ML-DSA signing when WC_DILITHIUM_CACHE_MATRIX_A is
enabled. by @embhorn (PR 10400)
* Fixed ML-KEM AVX2 assembly (5-bit decompression and final-block
ciphertext comparison). by @SparkiDev (PR 10430)
* Various ML-KEM correctness and key-state validation
improvements. by @SparkiDev (PR 10405)
* Reduced the ML-DSA verify-only key object size and allowed
SHAKE-only builds without the SHA-3 APIs. by @SparkiDev
(PR 10420)
* LMS fixes and improvements including a 32-bit signature length,
hash-algorithm parameter accessors, and empty-message signing.
by @SparkiDev (PR 10448)
* Migrate internal ML-KEM consumers to canonical wc_MlKemKey API.
by @Frauschi (PR 10571)
* Add PQ documentation for LMS, ML-DSA, ML-KEM, XMSS.
by @kaleb-himes (PR 10514)
* Various leak / alloc and zeroization fixes for SLH-DSA.
by @Frauschi (PR 10698)
TLS/DTLS
* Multiple TLS 1.3 Encrypted Client Hello (ECH) compliance fixes
per RFC 9849, covering rejection handling, ECHConfig parsing,
and inner ClientHello validation. by @sebastian-carpenter
(PR 10141)
* Added client-side ECH ech_outer_extensions encoding to shrink
the HPKE-sealed inner ClientHello. by @sebastian-carpenter
(PR 10306)
* Added maximum_name_length padding for the ECH ClientHelloInner
and a new wolfSSL_CTX_GenerateEchConfigEx() API.
by @sebastian-carpenter (PR 10326)
* Added opt-in ECH trial decryption
(wolfSSL_CTX_SetEchEnableTrialDecrypt()) and ECH
connection-status reporting. by @sebastian-carpenter (PR 10469)
* Evict the session from cache after an accepted 0-RTT resumption
to prevent early-data replay (RFC 8446 section 8).
by @julek-wolfssl (PR 10221)
* Gate 0-RTT acceptance on a cache-backed resumption ticket and
stop auto-advertising 0-RTT unless the application requests it.
by @julek-wolfssl (PR 10289)
* Send a missing_extension alert when the SNI extension is absent
in TLS 1.3. by @jackctj117 (PR 10332)
* Reject extensions in a TLS 1.3 Certificate message that were
not offered in the prior ClientHello/CertificateRequest
(RFC 8446 section 4.4.2). by @gasbytes (PR 10338)
* Error out on unknown/unsolicited extensions in TLS 1.3 response
messages while still tolerating GREASE in NewSessionTicket.
by @Frauschi (PR 10186)
* Fixed encrypt-then-mac handling on the non-resumption path.
by @embhorn (PR 10167)
* Fixed a TLS 1.3 AEAD/KeyUpdate limit that used 16-bit counters
where 32-bit values were required. by @SparkiDev (PR 10513)
* Decoupled the speculative ClientHello key share from
preferredGroup[0] via WOLFSSL_KEY_SHARE_DEFAULT_GROUP to reduce
HelloRetryRequests. by @Frauschi (PR 10435)
* Fixed DTLS 1.3 unnecessary client retransmission after
HelloRetryRequest and improved server robustness. by @rizlik
(PR 10349)
* Free and NULL the DTLS 1.3 cipher slot on init failure.
by @gasbytes (PR 10360)
* Added a WOLFSSL_DTLS13_5_9_0_COMPAT compatibility mode for
interoperating with pre-5.9.0 DTLS 1.3 clients. by @rizlik
(PR 10492)
* Added a missing WOLFSSL_QUIC_MAX_RECORD_CAPACITY check on the
QUIC early-data path. by @gasbytes (PR 10201)
* Default WOLFSSL_MAX_SIGALGO to 128. by @julek-wolfssl (PR 10528)
* Added NID_X25519 and NID_X448 support to the EVP layer.
by @julek-wolfssl (PR 10552)
* Enabled the all-zero shared-secret check for Curve25519/Curve448
by default and ensured post_handshake_auth was offered before
accepting a post-handshake CertificateRequest.
by @kareem-wolfssl (PR 10374)
* Fixed a DupSSL (write-dup) issue with Poly1305 authentication.
by @embhorn (PR 10337)
* Various X509 enhancements and fixes. by @SparkiDev (PR 10548)
* Cache AEAD record overhead on WOLFSSL. by @julek-wolfssl
(PR 10476)
* Add keylog support for TLS 1.3 ECH. by @sebastian-carpenter
(PR 10259)
* Enhance OCSP responder authorization. by @rlm2002 (PR 10532)
* Allow RSA client certs on ECDHE-ECDSA mutual auth.
by @julek-wolfssl (PR 10553)
* Enforce only 1 protocolname in serverhello. by @anhu (PR 10443)
* Fix tls_bench DTLS mode failures. by @miyazakh (PR 10606)
* Fix cipher property NIDs for SSL_get_current_cipher and add PSK
kx mapping. by @julek-wolfssl (PR 10639)
ASN and Certificate Parsing
* Allow serial number 0 for self-signed root CA certificates while
keeping the RFC 5280 check for others. by @jackctj117 (PR 9567)
* Fixed peer certificate verification with IP address SAN entries.
by @embhorn (PR 10169)
* Fixed partial chain verification to only terminate at a
certificate present in the original trust set. by @embhorn
(PR 10170)
* Apply DNS name constraints to the Subject CN when no SAN is
present. by @rlm2002 (PR 10223)
* Fixed IDNA wildcard matching. Thanks to Andrew Chin, SSLab at
Georgia Institute of Technology. by @embhorn (PR 10331)
* Honor the otherName GeneralName form in name-constraint checks
and add WOLFSSL_X509_STORE_ALLOW_NON_CA_INTERMEDIATE for
backward compatibility. by @embhorn (PR 10339)
* Always parse and store iPAddress and registeredID GeneralNames
for name-constraint enforcement. by @embhorn (PR 10354) Thanks
to Ankur Tyagi of Cisco Talos for the report.
* Hardened X.509 chain validation, session-ticket peer-cert
binding, and peer-cert restore, and reject embedded NUL in
dNSName/rfc822Name/URI SAN entries. by @julek-wolfssl (PR 10279)
* Restrict domain-name pattern matching and case folding to valid
FQDNs via a new IsValidFQDN() helper. by @douzzer (PR 10183,
PR 10419)
* Report a certificate verify failure for MD5-signed certificates
during chain verification. by @embhorn (PR 10222)
* Make SHA-1-with-RSA the last-resort signature type, as SHA-1
signatures are deprecated. by @SparkiDev (PR 10173)
* Tightened the key length check in wc_SignatureGetSize() from >=
to exact equality. by @miyazakh (PR 10122)
* Fixed a malformed AKID extension produced by
wolfSSL_X509_set_authority_key_id(). by @cconlon (PR 10370)
* Set the PKCS#8 (RFC 5958) version correctly when a private key
bundles the optional publicKey field. by @cconlon (PR 10427)
* Bind OCSP responder authorization to the CertID issuerKeyHash.
by @julek-wolfssl (PR 10303)
* Fixed multi-CRL PEM bundles being truncated to the first CRL in
wolfSSL_PEM_read_bio_X509_CRL. by @julek-wolfssl (PR 10336)
* Reject CRLs with unrecognized critical extensions and critical
entry extensions per RFC 5280. by @gasbytes (PR 10239, PR 10274)
* Validate DSA parameters when verifying a DSA key.
by @kareem-wolfssl (PR 10381). Thanks to Kr0emer for the report.
* Validate parameters in wolfSSL_EC_POINT_hex2point. Thanks to
breakingbad6, Yanzhao Shen, and Yingpei Zeng (Hangzhou Dianzi
University). by @embhorn (PR 10445)
* Adjusted the CA pathlen check and added glitch hardening when a
trust anchor is included in the certificate chain.
by @JacobBarthelmeh (PR 10296)
* Added ascii-digit validation in the ASN.1 time decode functions.
by @padelsbach (PR 10276)
* Allow a SubjectInfoAccess extension without an
id-ad-caRepository entry. by @holtrop-wolfssl (PR 10368)
* Tighten RFC 8017 A.2.3 compliance for RSA-PSS. by @miyazakh
(PR 10595)
Hardware and Embedded Ports
* Reworked the Microchip TA-100 port for cryptoauthlib v3.6.0,
with AES-GCM now working. by @danielinux (PR 9702)
* Added wolfCrypt hardware crypto support for NXP LPC55S69.
by @twcook86 (PR 10278)
* Added STM32U3 hardware crypto support (AES, hash, TRNG).
by @dgarske (PR 10361)
* Fixed STM32 AES hardware crypto when WOLFSSL_ARMASM is set.
by @dgarske (PR 10311)
* Allow custom time functions on STM32 by not forcing NO_ASN_TIME
when XTIME is defined. by @lealem47 (PR 10295)
* Fixed NXP DCP AES multiblock CBC encrypt/decrypt and the
in-place case. by @padelsbach (PR 10307)
* Fixed Octeon/Cavium AES-GCM AAD GHASH and non-12-byte IV J0
derivation bugs. by @ejohnstown (PR 10439) and @JacobBarthelmeh
(PR 10471)
* SE050 fixes plus new SE050_RSA_NO_VERIFY/SE050_ECDSA_NO_VERIFY/
SE050_ECDSA_NO_ECDHE options and simulator CI. by @rizlik
(PR 10219)
* Improved MQX/Fusion RTOS compatibility for XINET_PTON (MQX RTCS
support, IPv6 buffer sizing). by @josepho0918 (PR 10075)
* Added Zephyr 4.3 default TLS-socket support. by @ColtonWilley
(PR 10268)
* Fixed wolfSSL support on the CHERI RISC-V architecture.
by @wbeasley-thegoodpenguin (PR 10272)
* Revived and hardened the IoT-Safe memory-TLS example.
by @danielinux (PR 10457)
* Fixed the NXP CAAM build when WOLFSSL_HASH_KEEP is not defined.
by @JacobBarthelmeh (PR 10459)
* Set hashType in port hash implementations so SHA-512 modes can
be distinguished. by @padelsbach (PR 10193)
* Fixed the LMS/XMSS crypto callback software fallback to
propagate CRYPTOCB_UNAVAILABLE. by @padelsbach (PR 10456)
* Fixed a PIC32 hardware-acceleration stack-pointer free bug and
added PIC32MZ emulator tests. by @LinuxJedi (PR 10495)
* Numerous hardware/embedded port hardening fixes (Xilinx/AMD,
PSoC6, TROPIC01, devcrypto, RX64, STM32, KCAPI, QAT) from
Fenrir review. by @JacobBarthelmeh (PR 10467, PR 10470,
PR 10508) and @dgarske (PR 10496)
* Implements the SHA accelerator for MAX32666 as bare-metal.
by @mattia-moffa (PR 10431)
Rust Wrapper
* Added rand_core, aead, and cipher crate trait implementations.
by @holtrop-wolfssl (PR 10070)
* Added digest and signature crate trait implementations.
by @holtrop-wolfssl (PR 10248)
* Added password-hash, kem, and mac crate trait implementations.
by @holtrop-wolfssl (PR 10305)
* Added zeroize-on-drop and fixed Fenrir findings across the crate.
by @holtrop-wolfssl (PR 10205)
* Ensured memory safety for the C RNG struct and added
functionality found during boringtun integration.
by @holtrop-wolfssl (PR 10402)
* Added buffer-size checks to the ChaCha20-Poly1305 one-shot
wrappers. by @holtrop-wolfssl (PR 10267)
* Add scrypt KDF and RSA-OAEP support. by @holtrop-wolfssl
(PR 10556)
Build System and Portability
* Removed unused m4 macros and updated AX_PTHREAD to the latest
autoconf-archive macro. by @BrianAker (PR 10106)
* Guard <stddef.h> behind NO_STDDEF_H and userspace
fcntl/filesystem code behind WOLFSSL_KERNEL_MODE. by @philljj
(PR 10280, PR 10521)
* Support building --enable-opensslextra with NO_BIO and
NO_FILESYSTEM. by @JacobBarthelmeh (PR 10393)
* Allow --enable-writedup when DTLS is disabled. by @mattia-moffa
(PR 10527)
* Set WOLFSSL_USE_ALIGN automatically for ARM user_settings.h
builds. by @embhorn (PR 10487)
* Added WC_LINUXKM_USE_HEAP_WRAPPERS heap wrappers and AES-CCM
LKCAPI shims for the Linux kernel module. by @douzzer (PR 10512,
PR 10194)
* linuxkm enhancements for wolfGuard support. by @douzzer
(PR 10590)
* Added a software CryptoCb test device and
WOLF_CRYPTO_CB_ONLY_* test infrastructure (SHA-256, SHA-224,
AES). by @rizlik (PR 10351, PR 10500)
* Added configure and CMake options for WOLF_CRYPTO_CB_RSA_PAD.
by @kareem-wolfssl (PR 10428)
* Fixed mem_track.h compilation on multi-threaded non-Linux builds.
by @LinuxJedi (PR 10453)
* Fixed CUDA builds with WOLFSSL_AES_SMALL_TABLES. by @embhorn
(PR 10366)
* Skip testsuite.test for --enable-leantls builds. by @miyazakh
(PR 10510)
* RPM packaging fixes for examples/cmake artifacts and RHEL10 LTO
builds. by @space88man (PR 10270, PR 10356) and @twcook86
(PR 10275)
* Fixed the SGX build to not require fcntl.h. by @JacobBarthelmeh
(PR 10524)
* Various linuxkm Fenrir fixes. by @douzzer (PR 10688)
* Various bsdkm fixes and cleanup. by @philljj (PR 10565)
Bug Fixes
* AES-CCM encryption/decryption with 13 or 12 byte nonces and
large messages (>1 MiB and >256 MiB respectively) were not
properly rejected at argument validation time, allowing counter
wrap, keystream reuse, and consequent plaintext recovery. Thanks
to NVIDIA Project Vanessa for the report. (PR 10709)
* Out-of-bounds write in wolfSSL_get_finished() and
wolfSSL_get_peer_finished(). The functions validated the caller
buffer against TLS_FINISHED_SZ (12 bytes) but copied the full
Finished message, which for TLS 1.3 is the handshake hash size
(32 or 48 bytes), overrunning a correctly sized 12-byte buffer.
The output length is now validated before the copy. Thanks to
Qiushi Wu from IBM Research for the report. (PR 10576)
* Fix for rejecting the import of a public key that is the
identity element with Ed448. Use of a public key that is the
identity element allows for crafting a forged signature that
verifies with that specific public key. wolfSSL does not generate
an Ed448 key pair that is the identity element. This
defense-in-depth hardening for Ed448 and has been available
since release 5.9.1. Thanks to Nicholas Carlini from Anthropic
for the report. (PR 10116)
* Fix for !aNULL after an explicit ADH suite leaving the anon
suite available. Thanks to NVIDIA Project Vanessa for the
report. (PR 10714)
* P-521 ECDH SP-math validates 65 bytes but writes 66 leading to
a 1-byte heap overflow. Thanks to NVIDIA Project Vanessa for the
report. (PR 10702)
* Fix for FreeBSD kernel module to have additional sanity checks
on IV size when copying. This avoids a stack overflow in
--enable-freebsdkm-crypto-register builds used for preliminary
testing. Thanks to NVIDIA Project Vanessa for the report.
(PR 10695)
* Added an upper limit to the PBKDF iteration count and changed
iterations <= 0 to fail closed instead of silently clamping to
1. by @anhu (PR 10050) and @miyazakh (PR 10504)
* Added MP integer size bounds checks in SizeASN_Items.
by @anhu (PR 10051)
* Ensure certificates are correctly added to the cert manager.
by @anhu (PR 10073)
* Use O_CLOEXEC at file creation to harden against multithread
races. by @embhorn (PR 10162)
* BIO improvements and fixes, including an out-of-bounds read and
an inverted ctrl_pending check. by @rizlik (PR 10164)
* Guard against negative lengths in BIO, I/O callbacks, and
PKCS12 PBKDF. by @ColtonWilley (PR 10208)
* Fixed a dangling secure_renegotiation pointer after
TLSX_FreeAll. by @ColtonWilley (PR 10210)
* Added missing NULL checks across public API entry points.
by @ColtonWilley (PR 10216)
* Fixed NULL dereferences, a DSA SignFinal overflow, and the i2d
contract across EVP/OCSP/X509. by @ColtonWilley (PR 10217)
* Enforce the minimum auth tag size in wc_AesGcmDecryptFinal.
by @yosuke-wolfssl (PR 10175)
* AES-EAX and AES-SIV lifecycle and NULL-check fixes.
by @SparkiDev (PR 10174)
* Added a C implementation for ARM AES-GCM small-table builds
without NEON. by @SparkiDev (PR 10176)
* SP integer fixes for negative numbers, zero-used edge cases,
bounds, and truncation, plus added testing. by @SparkiDev
(PR 10235, PR 10529) and @embhorn (PR 10478) Thanks to Kr0emer
for reporting these issues and testing the fixes.
* Fixed the Curve25519 private-key clamp check to enforce RFC 7748
rule 3. by @MarkAtwood (PR 10363)
* Fixed the Blake2 oversized-key path. by @mattia-moffa (PR 10447)
* Fixed a SAKKE heap buffer overflow and a correctness gap in
sakke_hash_to_range, plus a wc_export_int size check.
by @JeremiahM37 (PR 10442) and @philljj (PR 10444)
* Added checks to PKCS7_VerifySignedData to prevent out-of-bounds
access. Thanks to Feng Ning / Innora. by @padelsbach (PR 10441)
* Various PKCS#7 fixes. by @Frauschi (PR 10203) and
@kareem-wolfssl (PR 10128)
* Various PKCS#12 fixes including strict-aliasing, zeroization,
and overflow checks. by @rlm2002 (PR 10378)
* Fixed an ECC temporary-key leak and undefined behavior, and an
ECC validation regression. by @Frauschi (PR 10346, PR 10260)
* Fixed a double free in wolfSSL_X509_set_ext. by @padelsbach
(PR 10481)
* Reset the SHA-3 hashType between ML-KEM cryptocb calls.
by @night1rider (PR 10211)
* Plumb the caller heap into CMAC before the cryptocb fires.
by @night1rider (PR 10401)
* Fixed inverted AllocDer checks in the alt private key id/label
paths. by @night1rider (PR 10168)
* TLS extension bounds-checking fixes. Thanks to Suryansh
Mansharamani (Plainshift AI). by @embhorn (PR 10220)
* Hardened TLSX_KeyShare_ProcessPqcHybridClient against double
free and wrong-key NULLing. by @embhorn (PR 10327) and
@kareem-wolfssl (PR 10493)
* Various missing bounds and length checks. by @kareem-wolfssl
(PR 10142, PR 10277)
* Fixed the NO_VERIFY_OID build in GetOID. Thanks to @cpsource.
by @JeremiahM37 (PR 10440)
* Ensure large buffers are heap-allocated under
WOLFSSL_SMALL_STACK. by @Frauschi (PR 10245)
* Fixed private key lock issues. by @kaleb-himes (PR 10446)
* Improved HPKE return codes to use defined error values.
by @sebastian-carpenter (PR 10455)
* MD4 and MD2 public APIs now return int instead of void,
alongside static-analysis fixes for SECO, devcrypto, and ARIA.
by @JacobBarthelmeh (PR 10460)
* Extensive wolfCrypt input-validation, key-zeroization, and
side-channel hardening across RSA, ECC, EdDSA, Curve25519/448,
DH, SRP, PKCS#7, KDF, LMS, and PQ key handling.
by @JeremiahM37 (PR 10136, PR 10231, PR 10238, PR 10264,
PR 10304, PR 10340, PR 10386, PR 10392, PR 10413, PR 10426,
PR 10468), @julek-wolfssl (PR 10230, PR 10247, PR 10398), and
@aidangarske (PR 10155)
* Fixed a word32 size overflow in wc_DeCompressDynamic that could
under-allocate a buffer and cause a heap overflow on --with-libz
+ PKCS7-compressed builds. by @JeremiahM37 (PR 10413)
* Resolved numerous Coverity static-analysis findings. by @rlm2002
(PR 10129, PR 10165, PR 10376, PR 10418, PR 10482)
* Enforce MAX_ENTROPY_BITS upper bound in wc_Entropy_Get().
by @miyazakh (PR 10593)
* Various EVP API bug fixes. by @MarkAtwood (PR 10364, PR 9987)
* Various ppc and armv8 asm and linuxkm fixes. by @douzzer
(PR 10600)
* Fixes for Zephyr secure sockets integration. by @Frauschi
(PR 10583)
* Add signed-length validation to d2i, PEM, and buffer-load APIs
to harden against potential out of bounds case. by @ColtonWilley
(PR 10207)
* Reject duplicate certificatePolicies certificate extensions
(RFC 5280), previously accepted as last-wins. by @Frauschi
(PR 10714)
* Strengthen subgroup check in wc_DhAgree by @philljj (PR 10560).
Thanks to Muhammad Arya Arjuna Habibullah (Pelioro) for the
report.
* Fix missing private key zeroization in ML-KEM by @anhu
(PR 10665). Thanks to Uday Devaraj, SYNE Lab, Syracuse
University.
* Force-zero wc_AesSivDecrypt output buffer on authentication
failure. by @holtrop-wolfssl (PR 10668)
* Ed448: check for public key presence on export.
by @holtrop-wolfssl (PR 10656)
* Hardening fixes in wolfSSL_strnstr and mp_get_digit. Thanks to
Dominik Blain / COBALT Security for the bug report.
by @padelsbach (PR 10138)
Documentation and Maintenance
* Added documentation for the new OCSP responder and certificate
accessor APIs. by @julek-wolfssl (PR 10147)
* Improved API documentation argument descriptions for the
sigalgs/groups/cipher-list/options setters. by @kojo1 (PR 10382)
* Added doxygen coverage for the SHE API. by @night1rider (PR 10243)
* Added the BN_bn2binpad API. by @julek-wolfssl (PR 10148)
* Added TLS 1.3 decryption fuzz tests, Monte Carlo / unaligned /
in-place cipher tests, and expanded negative test coverage for
handshake, AEAD, PKCS7, PSS, DSA, DRBG, and PQ paths.
by @SparkiDev (PR 10461, PR 10213, PR 10226) and @JeremiahM37
(PR 10166, PR 10291)
* Various improvements to CI efficiency and parallelism.
by @julek-wolfssl (PR 10667, PR 10685, PR 10701, PR 10731)
---
security/wolfssl/Makefile | 5 ++---
security/wolfssl/distinfo | 6 +++---
security/wolfssl/files/patch-cmake_config.in | 15 ---------------
security/wolfssl/files/patch-cmake_options.h.in | 15 ---------------
security/wolfssl/files/patch-configure.ac | 16 ----------------
security/wolfssl/pkg-plist | 16 +++++++---------
6 files changed, 12 insertions(+), 61 deletions(-)
diff --git a/security/wolfssl/Makefile b/security/wolfssl/Makefile
index a458dae3a447..375c99fd2981 100644
--- a/security/wolfssl/Makefile
+++ b/security/wolfssl/Makefile
@@ -1,6 +1,5 @@
PORTNAME= wolfssl
-PORTVERSION= 5.9.1
-PORTREVISION= 1
+PORTVERSION= 5.9.2
CATEGORIES= security devel
MASTER_SITES= https://www.wolfssl.com/ \
LOCAL/fox
@@ -12,7 +11,7 @@ WWW= https://www.wolfssl.com/
LICENSE= GPLv3
LICENSE_FILE= ${WRKSRC}/COPYING
-USES= cpe libtool zip autoreconf
+USES= cpe libtool zip
USE_LDCONFIG= yes
GNU_CONFIGURE= yes
diff --git a/security/wolfssl/distinfo b/security/wolfssl/distinfo
index b150053a7498..31a6fa8fa9ba 100644
--- a/security/wolfssl/distinfo
+++ b/security/wolfssl/distinfo
@@ -1,3 +1,3 @@
-TIMESTAMP = 1775971830
-SHA256 (wolfssl-5.9.1.zip) = e9d84299739ea16ae18340aee6b8de9e584b5206a751e1df088cd0459cfe5e45
-SIZE (wolfssl-5.9.1.zip) = 28487279
+TIMESTAMP = 1782539423
+SHA256 (wolfssl-5.9.2.zip) = 591cc52625b9bda3cee8af4cba2cf0718b5e466dce5c0604af730d6d37a6d2fa
+SIZE (wolfssl-5.9.2.zip) = 29803913
diff --git a/security/wolfssl/files/patch-cmake_config.in b/security/wolfssl/files/patch-cmake_config.in
deleted file mode 100644
index a44796b6d171..000000000000
--- a/security/wolfssl/files/patch-cmake_config.in
+++ /dev/null
@@ -1,15 +0,0 @@
-Bug report: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=294287
-Upstream: https://github.com/wolfSSL/wolfssl/pull/10097
-
---- cmake/config.in.orig 2026-03-18 23:01:31 UTC
-+++ cmake/config.in
-@@ -19,9 +19,6 @@
- /* Define to 1 if you have the `gmtime_r' function. */
- #cmakedefine HAVE_GMTIME_R @HAVE_GMTIME_R@
-
--/* Define to 1 if you have the <limits.h> header file. */
--#cmakedefine HAVE_LIMITS_H @HAVE_LIMITS_H@
--
- /* Define to 1 if you have the <pcap/pcap.h> header file. */
- #cmakedefine HAVE_PCAP_PCAP_H @HAVE_PCAP_PCAP_H@
-
diff --git a/security/wolfssl/files/patch-cmake_options.h.in b/security/wolfssl/files/patch-cmake_options.h.in
deleted file mode 100644
index ee9196d69f0d..000000000000
--- a/security/wolfssl/files/patch-cmake_options.h.in
+++ /dev/null
@@ -1,15 +0,0 @@
-Bug report: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=294287
-Upstream: https://github.com/wolfSSL/wolfssl/pull/10097
-
---- cmake/options.h.in.orig 2026-03-18 23:01:31 UTC
-+++ cmake/options.h.in
-@@ -33,6 +33,9 @@ extern "C" {
- #endif
-
- #ifndef WOLFSSL_OPTIONS_IGNORE_SYS
-+/* Since types.h depends on HAVE_LIMITS_H, we must define it in options.h. */
-+#undef HAVE_LIMITS_H
-+#cmakedefine HAVE_LIMITS_H @HAVE_LIMITS_H@
- #undef _GNU_SOURCE
- #cmakedefine _GNU_SOURCE
- #undef _POSIX_THREADS
diff --git a/security/wolfssl/files/patch-configure.ac b/security/wolfssl/files/patch-configure.ac
deleted file mode 100644
index 4d94428448cf..000000000000
--- a/security/wolfssl/files/patch-configure.ac
+++ /dev/null
@@ -1,16 +0,0 @@
-Bug report: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=294287
-Upstream: https://github.com/wolfSSL/wolfssl/pull/10097
-
---- configure.ac.orig 2026-03-18 23:01:31 UTC
-+++ configure.ac
-@@ -185,7 +185,9 @@ AC_ARG_ENABLE([freebsdkm-crypto-register],
- [ENABLED_BSDKM_REGISTER=no]
- )
-
--AC_CHECK_HEADERS([arpa/inet.h fcntl.h limits.h netdb.h netinet/in.h stddef.h time.h sys/ioctl.h sys/socket.h sys/time.h errno.h sys/un.h ctype.h sys/random.h])
-+AC_CHECK_HEADERS([arpa/inet.h fcntl.h netdb.h netinet/in.h stddef.h time.h sys/ioctl.h sys/socket.h sys/time.h errno.h sys/un.h ctype.h sys/random.h])
-+# Special case: Since types.h depends on HAVE_LIMITS_H, we must define it in options.h.
-+AC_CHECK_HEADER([limits.h], [AM_CPPFLAGS="$AM_CPPFLAGS -DHAVE_LIMITS_H=1"], [])
- AC_CHECK_LIB([network],[socket])
- AC_C_BIGENDIAN
- AC_C___ATOMIC
diff --git a/security/wolfssl/pkg-plist b/security/wolfssl/pkg-plist
index 396a9770438b..05ffec14c4e7 100644
*** 75 LINES SKIPPED ***