Adapt existing sig fuzz harness including more algorithms

[nathaniel-brough]

Adapts existing fuzz-harness to support more algorithms.

This is a continuation of the effort described in #1215

• Does this PR change the input/output behaviour of a cryptographic algorithm (i.e., does it change known answer test values)? (If so, a version bump will be required from x.y.z to x.(y+1).0.)
• Does this PR change the list of algorithms available -- either adding, removing, or renaming? Does this PR otherwise change an API? (If so, PRs in fully supported downstream projects dependent on these, i.e., oqs-provider will also need to be ready for review and merge by the time this is merged.)

[nathaniel-brough]

The failing check doesn't appear to be related to this PR. I'm unsure but it kind of looks like a flaky performance check.

[cothan]

It's seem like you found a rare case that the verify operation failed to verify the signature.
I hope the bug lies in the implementation, not the algorithm itself. I don't know about CROSS signature but this one is concerning.

liboqs/tests/speed_sig.c

Line 29 in 1d92135

if (OQS_SIG_verify(sig, message, message_len, signature, signature_len, public_key) != OQS_SUCCESS) {

[nathaniel-brough]

Hmm ok. It doesn't look like the public/private keypair or the message is saved on failure here, so I guess it'll be a little difficult to reproduce.

While this seems unrelated to the PR, I'll see if I can modify my fuzzer in such a way to trigger the bug in a way that's reproducible.

[nathaniel-brough]

Ok I've updated this fuzz harness (locally) such that it might be able to find this bug in a deterministic way. Is there a preference on stacking changes or just drip feeding in changes so that each PR is atomic i.e. doing one thing?

[SWilson4]

Looping in @alexrow and @rtjk as the maintainers of CROSS.

[rtjk]

Thank you @SWilson4, we're taking a look at this verification failure and will respond soon.

[rtjk]

I managed to reproduce this very rare verification error in CROSS-RSDPG-128-small. I'm still looking for the root cause but I can confirm that it seems totally unrelated to the new fuzzing harness.

[SWilson4]

I managed to reproduce this very rare verification error in CROSS-RDSDP-128-small. I'm still looking for the root cause but I can confirm that it seems totaly unrelated to the new fuzzing harness.

Thanks for taking a look! I'll open a separate issue to track the error, then.

[baentsch]

LGTM.

[SWilson4]

Thanks for the contribution @nathaniel-brough!