You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
One of the x509 requirements is that every cert's Issuer field must be byte-for-byte identical to its issuer's Subject field. The go crypto/x509 CreateCertfiicate function is supposed to guarantee this.
However, we do not have a lint that checks this, because the zlint lints are contextless: we don't have a way to supply the issuer cert to compare the new cert's Issuer bytes against.
Since we're already thinking about lint-like checks which require more context to operate (see #6964 and #6945), we should consider this one too.
The text was updated successfully, but these errors were encountered:
In #7005 several safety
checks were added to the `ceremony` tool:
This change extracts the `RawSubject` to `RawIssuer` DER byte comparison
into the `//linter` package proper so that it can serve both `//ca` and
`//cmd/ceremony`.
Adds a helper function `verifyTBSCertificateDeterminism` to `//ca`
similar to an existing check in `//cmd/ceremony`. This code is not
shared because we want `//cmd/ceremony` to largely stand alone from
boulder proper. The helper performs a byte comparison on the
`RawTBSCertificate` DER bytes for a given linting certificate and leaf
certificate. The goal is to verify that `x509.CreateCertificate` was
deterministic and produced identical DER bytes after each signing
operation.
Fixes#6965
This idea came out of remediation for https://bugzilla.mozilla.org/show_bug.cgi?id=1838667
One of the x509 requirements is that every cert's Issuer field must be byte-for-byte identical to its issuer's Subject field. The go crypto/x509 CreateCertfiicate function is supposed to guarantee this.
However, we do not have a lint that checks this, because the zlint lints are contextless: we don't have a way to supply the issuer cert to compare the new cert's Issuer bytes against.
Since we're already thinking about lint-like checks which require more context to operate (see #6964 and #6945), we should consider this one too.
The text was updated successfully, but these errors were encountered: