utilities for generic SetKey and ExportKey operations on HMAC, RSA, ECC,
and AES. Add wc_ecc_size/wc_ecc_sig_size callback hooks for hardware-only
keys. Integrate into configure.ac as --enable-cryptocbutils=setkey,export
options with CI test configurations in os-check.yml.
Add test handlers in test.c and api.c with export/import delegation
pattern, small-stack-safe allocations, custom curve support, and
DEBUG_CRYPTOCB helpers.
Refactor CMAC init to common function, add wc_AesNew_Id/Label API, do same for RSA new
functions, and add test init helpers for id[] support along with some test disable options
Fixed wc_AesEaxAuthDataUpdate to check eax for NULL before
dereferencing.
Fix AesSivCipher to delete/free AES if new/initialization succeeded.
Memsetting to 0 doesn't work when WC_DEBUG_CIPHER_LIFECYCLE is defined.
Added tests for:
- AES-EAX streaming
- AES-SIV
- Poly1305
- DES-CBC
wc_PKCS7_DecodeAuthEnvelopedData() accepted an attacker-controlled GCM tag
length from the mac OCTET STRING and did not validate it against the
parsed aes-ICVlen parameter. In parallel, wc_AesGcmDecrypt() accepted
very short tags on decrypt while encrypt enforced WOLFSSL_MIN_AUTH_TAG_SZ.
This made short-tag verification reachable through CMS AuthEnvelopedData
and weakened integrity checks by allowing tag truncation.
Fixes:
- validate parsed macSz range in AuthEnvelopedData decode
- require authTagSz to match parsed macSz
- reject undersized GCM tags in PKCS7 decode
- enforce WOLFSSL_MIN_AUTH_TAG_SZ in wc_AesGcmDecrypt() and
wc_AesGcmDecryptFinal()
Also add a regression test in pkcs7authenveloped vectors that truncates
the final MAC OCTET STRING length from 16 to 1 and verifies decode fails.
Reported by: Nicholas Carlini (Anthropic) & Thai Duong (Calif.io)
The guard `if (cmac->totalSz != 0)` was used to skip XOR-chaining on
the first block (where digest is all-zeros and the XOR is a no-op).
However, totalSz is word32 and wraps to zero after 2^28 block flushes
(4 GiB), causing the guard to erroneously fire again and discard the
live CBC-MAC chain state. Any two messages sharing a common suffix
beyond the 4 GiB mark then produce identical CMAC tags, enabling a
zero-work prefix-substitution forgery. The fix removes the guard,
making the XOR unconditional; the no-op property on the first block is
preserved because digest is zero-initialized by wc_InitCmac_ex.
Identified by: Nicholas Carlini (Anthropic) & Thai Duong (Calif.io)
Offload ML-DSA operations onto a PKCS#11 token via the cryptoCb
interface:
* Key generation
* Signature generation
* Signature verification
* Key import
Both the pure and pre-hash versions are supported. Not yet supported are
the pre-hash versions that also offload the hashing onto the token.
This also fixes casting errors introduced in #9780 due to usage of
uintptr_t, which is unavailable without including stdint.h on some
platforms. Use the wolfssl own wc_ptr_t instead.
Add full RSA-PSS (RSASSA-PSS) support to PKCS#7 SignedData
encoding and verification.
This change enables SignerInfo.signatureAlgorithm to use
id-RSASSA-PSS with explicit RSASSA-PSS-params (hash, MGF1,
salt length), as required by RFC 4055 and CMS profiles.
Key changes:
- Add RSA-PSS encode and verify paths for PKCS7 SignedData
- Encode full RSASSA-PSS AlgorithmIdentifier parameters
- Decode RSA-PSS parameters from SignerInfo for verification
- Treat RSA-PSS like ECDSA (sign raw digest, not DigestInfo)
- Fix certificate signatureAlgorithm parameter length handling
- Add API test coverage for RSA-PSS SignedData
This resolves failures when using RSA-PSS signer certificates
(e.g. -173 invalid signature algorithm) and maintains backward
compatibility with RSA PKCS#1 v1.5 and ECDSA.
Signed-off-by: Sameeh Jubran <sameeh@wolfssl.com>
* add prefetch_ptr flag argument to AesEncrypt_C() and AesDecrypt_C(), and call PreFetchTe() and PreFetchSBox() only if *prefetch_ptr is zero, whereupon it is set to 1;
* when C implementations are available, add prefetch_ptr arg to wc_AesEncrypt() and wc_AesDecrypt(), and pass it through;
* in functions that directly call the AES block encryption methods, opportunistically inhibit prefetch on all but the first call;
* move AES-specific code in wc_CmacUpdate() in cmac.c to wc_local_CmacUpdateAes() in aes.c to let it use conditional prefetching;
* add WC_ARG_NOT_NULL(), WC_ARGS_NOT_NULL(), and WC_ALL_ARGS_NOT_NULL attribute abstractions.
Latest version of clang with optimization turned right up will make the
pre-fetch functions return 0.
The pre-fetch functions are there to ensure tables are all in cache not
to calculate a value.
Add CryptoCB-based AES key import support to enable Secure Element
offload without exposing raw AES key material to wolfCrypt.
When WOLF_CRYPTO_CB_AES_SETKEY is defined, wolfCrypt invokes a CryptoCB
callback during AES key setup. Behavior is determined by the callback
return value:
- If callback returns 0: Key is imported to the device (aes->devCtx).
Key is NOT copied to wolfCrypt RAM; GCM H/M tables are NOT generated.
Full hardware offload is assumed.
- If callback returns CRYPTOCB_UNAVAILABLE: Device does not support
SetKey. Normal software path is used; key is copied to devKey for
optional encrypt/decrypt acceleration.
- Any other error: Propagated to the caller.
Key points:
- Add wc_CryptoCb_AesSetKey() callback for AES key import
- Update AES SetKey paths to call CryptoCB and branch on return value
- Skip GCM H/M table generation when callback succeeded (devCtx set)
- Preserve existing behavior when devId is INVALID_DEVID or
WOLF_CRYPTO_CB_AES_SETKEY is not defined
Testing:
- Add unit test for CryptoCB AES SetKey (verifies key isolation when
callback succeeds)
- Add end-to-end AES-GCM offload test (SetKey, Encrypt, Decrypt, Free
via CryptoCB)
- Tests use a mock SE with software AES to validate routing
Enable with: CPPFLAGS="-DWOLF_CRYPTO_CB_AES_SETKEY -DWOLF_CRYPTO_CB_FREE"
Signed-off-by: Sameeh Jubran <sameeh@wolfssl.com>
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.
Undefined behaviour when there are multiple volatile variables accessed
in the one C statement.
Changes to introduce non-volatile temporaries, split statement or make
variable non-volatile.
OpenSSL compatability layer expects aes->reg to be unmodified by AES-GCM
decrypt call. ARM32/Thumb2 assembly implementation modifies buffer.
Keep a copy and restore aes->reg after call.
Implementations of AES-ECB, AES-CBC, AES-CTR, AES-GCM, AES-XTS with base
instructions and NEON but not using crypto instructions.
Benchmark of AES-ECB added.
Updated AES tests.
Eric Blankenhorn
Jeremiah Mackey
night1rider
Daniel Pouzzner
David Garske
Sean Parkinson
Yosuke Shimizu
Tobias Frauenschläger
Ruby Martin
JacobBarthelmeh
Sameeh Jubran
aidan garske
Tesfa Mael
Joseph Chen