Doesn't look like we want to certify SHA-1 verification at all, even though it's allowed for legacy usage. SHA-1 verification is blocked with config in FIPS, but for certification we should either hard-block or indicator-disapprove. I propose doing the latter downstream. Impact should be minimal if we block it by default anyway. Upstream test that checks for verification being approved: https://gitlab.com/gnutls/gnutls/-/blob/ebfe675f15bfb52d61451e96d0f73d792a2b9a9b/tests/fips-test.c#L457
(In reply to Alexander Sosedkin from comment #0) > Doesn't look like we want to certify SHA-1 verification at all, even though > it's allowed for legacy usage. > SHA-1 verification is blocked with config in FIPS, but for certification we > should either hard-block or indicator-disapprove. > I propose doing the latter downstream. Impact should be minimal if we block > it by default anyway. DSA verification is also allowed for legacy use but we mark them non-approved in upstream, so I guess we could do the same for SHA-1 signature verification.