update to liboqs 0.9.1

[aparcar]

The release updated the Classic McEliece implemented and since we use it, might be a good idea to update the lib.

[koraa]

This is fascinating

thread 'fuzz_handle_msg' panicked at /home/runner/work/rosenpass/rosenpass/secret-memory/src/secret.rs:152:31:
source slice length (13568) does not match destination slice length (13608)
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

Does the fuzzer crash reproduce when you run locally?

[koraa]

Oh by the way, it is possible that the updated OQS actually has new versions of McEliece & Kyber, in which case this crash would make sense

[aparcar]

Oh by the way, it is possible that the updated OQS actually has new versions of McEliece & Kyber, in which case this crash would make sense

How so?

[aparcar]

In other words, does this mean with a bump all deployed setups would break?

[koraa]

Oh by the way, it is possible that the updated OQS actually has new versions of McEliece & Kyber, in which case this crash would make sense

How so?

Well, occasionally OQS just updates their implementations

In other words, does this mean with a bump all deployed setups would break?

I would say: All deployed setups should update.

Anway. Lets fix and merge this but please add info to the commit indicating whether the algorithm versions where changed.

[prabhpreet]

Looking into the cargo fuzz failure, will add any findings here

[aparcar]

Great thanks

[prabhpreet]

The secret key size and the cipher text length have changed in the McEliece KEM we are using.

https://github.com/open-quantum-safe/liboqs/blob/7c3a0e9aa7f9568e4dcafaf908ff8aa0008f0b71/src/kem/classic_mceliece/kem_classic_mceliece.h#L32-L33

So the slice size turns out to be smaller when copying the slice into Secret for the handle_msg.rs test

https://github.com/aparcar/rosenpass/blob/5c84df41454a36bf3a0241583273e6f694414aa0/fuzz/fuzz_targets/handle_msg.rs#L10

I've added some updates to use constants instead of manually specifying slice sizes in these tests for KEMs on this branch:

https://github.com/prabhpreet/rosenpass/tree/liboqs-910

@aparcar I don't have permissions to write to your branch. Can you add the changes from the branch and see if the pipeline passes?

[prabhpreet]

@aparcar Tests are passing with changes- please feel free to review #292 and incorporate the changes or review that PR and close this one

[aparcar]

Superseded by #292