Add lint enforcing the restrictions on subject DN fields for mailbox validated SMIME certificates #713
Conversation
…validated SMIME certificates
…ger export the underlying struct as per other lints in zlint
v3/lints/cabf_smime_br/mailbox_validated_enforce_subject_field_restrictions.go
Outdated
Show resolved
Hide resolved
v3/lints/cabf_smime_br/mailbox_validated_enforce_subject_field_restrictions.go
Outdated
Show resolved
Hide resolved
v3/lints/cabf_smime_br/mailbox_validated_enforce_subject_field_restrictions.go
Show resolved
Hide resolved
…a, remove unused test certificates
v3/lints/cabf_smime_br/mailbox_validated_enforce_subject_field_restrictions.go
Outdated
Show resolved
Hide resolved
v3/lints/cabf_smime_br/mailbox_validated_enforce_subject_field_restrictions.go
Show resolved
Hide resolved
| @@ -221,6 +221,9 @@ func (l *CertificateLint) Execute(cert *x509.Certificate, config Configuration) | |||
| if l.Source == CABFBaselineRequirements && !util.IsServerAuthCert(cert) { | |||
| return &LintResult{Status: NA} | |||
| } | |||
| if l.Source == CABFSMIMEBaselineRequirements && !util.IsEmailProtectionCert(cert) { | |||
…_restrictions.go comment to list relevant policy OIDs Co-authored-by: Christopher Henderson <chris@chenderson.org>
…s of mailbox field presence lint
@robplee I'll take on this investigation as it's an interesting question. The test corpus from a query against the censys.io database, so I am as yet unsure of whether-or-not these certificates will be available to us. My guess would be |
|
I believe this PR sets some basic precedents and expectations required to start plowing through #712, so what else needs to be done here do you think? Does this need further review or just testing against a large corpus of SMIME certs? I could try to help run against an internal corpus here to at least get some initial testing results if that's needed before we feel comfortable merging this PR. |
I have actually been pecking away at this over the past several weeks. We got a new batch of certs 100k+ large from Censys, but it's been a lot of work to format the corpus, deduplicate, sanitize, and smoke that the new failures make sense.
In the interest of time I think that we would be fine merging this now. If things go truly south in the immediate absence of more test certs then it will be easy to soft "back out" changes by simply not running SMIME checks, |
I have actually been pecking away at this over the past several weeks. We got a new batch of certs 100k+ large from Censys, but it's been a lot of work to format the corpus, deduplicate, sanitize, and smoke that the new failures make sense.
In the interest of time I think that we would be fine merging this now. If things go truly south in the immediate absence of more test certs then it will be easy to soft "back out" changes by simply not running SMIME checks, |
|
Would it be meaningful to change
to:
because this is under Section 7.1.4.2 which regards subscriber certificates? |
This PR is the first to attempt to address something from the big todo list in #712. It's a simple lint enforcing the field MAY/SHALL NOT table for subject DN fields in mailbox validated certificates. I think the changes are fairly straightforward so we should be able to discover if there are any extra things needed for us to start supporting SMIME BR linting.
My current wonder is about the TestCorpus for integration testing as I assume it won't have any applicable certs in it but I guess that's something we can try to work out now?