Re: [PATCH v5] examples: add Wycheproof validation app

[Stephen Hemminger <stephen@networkplumber.org>]

On Tue, 22 Sep 2026 16:05:57 +0000
Kai Ji <kai.ji@intel.com> wrote:

> Add a Wycheproof JSON vector validation example for cryptodev PMDs.
> 
> Support these algorithms when advertised by the selected PMD:
> - AEAD: AES-GCM, AES-CCM, SM4-GCM, ChaCha20-Poly1305
> - MAC: AES-CMAC, AES-GMAC, HMAC SHA-1/SHA-2/SHA-3/SM3
> - Asymmetric: DSA (P1363 verify), ECDSA (P1363 verify),
>   ECDH (ecpoint shared-secret compute)
> 
> Validate valid vectors against generated ciphertexts, tags, plaintexts,
> digests, shared secrets, or signature verification, and require the
> expected rejection for invalid vectors. Digest inputs for DSA and ECDSA
> use the symmetric auth path, selecting a separate symmetric-capable
> device when the target device is asymmetric-only.
> 
> Skip parameter combinations outside PMD capability ranges and identify
> recognized vector families without a compatible DPDK transform. A
> --debug option lists every failed or skipped vector.
> 
> Add Meson and standalone build integration, with usage documentation.
> 
> Signed-off-by: Kai Ji <kai.ji@intel.com>
> ---

AI review had some useful suggestions on this.

Wycheproof validation example (v5) - review

v4 -> v5 delta: drop unused APP_NAME, close the device on the
app_init() error path, and treat asym session-create failure as
-ENOTSUP (skip) instead of -EIO/-ENOMEM (fail) in run_dsa_verify(),
run_ecdh_ecpoint() and run_ecdsa_verify().

Applied on 6bbb7b3, built with -Dwerror=true, docutils clean. Same
crypto_openssl runs as v4: directory run passed=1902 failed=0,
--mbuf-dataroom 160 and --cryptodev-id 256 behave as in v4.

No Errors or Warnings.

Info
----

main.c:1222, 1480, 1715
    ret = -ENOTSUP;
  Session-create failure is now indistinguishable from an unsupported
  curve or key size: a PMD that advertises the xform but fails
  session setup for a curve it claims to support, or a -ENOMEM from
  an exhausted asym session pool, lands in skipped_capability and the
  exit status stays zero. Understood that the asym capability struct
  does not enumerate curves, so this is a documented trade-off; just
  noting that valid asym vectors can no longer fail at session setup.

main.c:823-824, 837, 957, 1047 (open since v3)
    memcmp(output, vector->ct, vector->ct_len) != 0
  memcmp(NULL, NULL, 0) for empty-message vectors; glibc declares
  memcmp nonnull. Guard with len != 0 &&.

main.c:182-183, 266 (open since v3)
  Positional struct initializers; use designated members.

doc/guides/sample_app_ug/wycheproof_validation.rst (open since v3)
  crypto_openssl advertises no ECDSA/ECDH xform capability, so the
  documented command line skips every ECDSA/ECDH vector. One sentence
  saying asymmetric coverage needs a PMD advertising those xforms.

Pre-existing, not introduced by this patch
  lib/cryptodev/rte_cryptodev.c:2411 rte_cryptodev_asym_session_create()
    ret = dev->dev_ops->asym_session_configure(dev, xforms, sess);
    if (ret < 0) {
        ...
        return ret;
    }
  On configure failure the object stays allocated from the pool and
  *session still points at it, unlike the sym variant which does
  rte_mempool_put() on its error_exit path. This app copes by freeing
  a non-NULL session at out: even when create returned an error; if
  the library is ever aligned with the sym behaviour that becomes a
  double put. One for the cryptodev maintainers, not this patch.

Review-Result: CLEAN