* Fix bug in the output of the ct_verify function
A bug in the CCA transformation was reported on the pqc-forum on 2020-12-10
https://groups.google.com/a/list.nist.gov/g/pqc-forum/c/kSUKzDNc5ME
It was fixed today in 669522db63
.
This commit ports that fix to PQClean
* add note to SECURITY.md
* update upstream commit in META.yml
2.4 KiB
A note on the security of the included implementations and schemes
This project contains (reference) implementations of cryptographic libraries. We do not make any security claims about the code included in PQClean. In the current state, we distribute reference implementations with minor modifications. We did not perform any extensive security analyses. This code is suitable for experimental or scientific use. We recommend careful expert code review before using any of the included implementations in production environments.
See also the NIST PQC Forum for discussion about the cryptographic schemes included in PQClean.
Current and past security issues
We will attempt to document security vulnerabilities made known to us on a best-effort basis. If an issue is marked with a date, the issue has been resolved since any commits made after that date.
Again, we emphasise that the code in this repository has not seen any formal analysis or audit! Use at your own risk.
Open issues
- LEDAcryptKEM
leaktime
implementations are known to not be constant-time and expected to have timing side channel vulnerabilities.
2020-12-11
- The fix of the timing leak in the CCA transform of FrodoKEM in PR #303 was ineffective. The FrodoKEM team released another fix which was ported to PQClean in PR #367.
2020-06-19
- A potential timing leak was present in the FrodoKEM decapsulation routine, as identified by Guo, Johansson, and Nilsson. This was fixed in PR #303.
2019-09-24
- All Falcon implementations before PR #235 got merged were insecure. See EPRINT report 2019/893.
2019-09-10
- The included incremental
sha512
implementation was callingcrypto_hashblocks_sha256
before 2019-9-10. This lead to an insufficient security level of the results of this hash function. The function was not used in any implementations, though. See PR #232.