Skip to content

Integrate SPKI denylist into revocation checks for all callers - #271

Draft
ctz wants to merge 3 commits into
mainfrom
jbp-spki-denylist
Draft

ctz wants to merge 3 commits into
mainfrom
jbp-spki-denylist

Conversation

@ctz

@ctz ctz commented Oct 21, 2026

Copy link
Copy Markdown
Member

rustls-upki required no action here, as it already relies on the rustls-webpki crate which does this check.

$ cargo -q run -p rustls-upki --example simpleclient dv-g3b-emudhra.we-have-a-problem.com
Error: Custom { kind: InvalidData, error: InvalidCertificate(Other(OtherError(CertDistrusted))) }

For the others, after this change:

$ curl -s --cacert cacert-2026-09-25.pem -w '%{CERTS}' https://g3.emsign.filippo.io  | upki revocation check
Error: certificate has a compromised key (the private key is 156)

Location:
    upki-cli/src/bin/upki.rs:74:25
$ openssl s_client -connect dv-g3b-emudhra.we-have-a-problem.com:443 -verify 1 -verifyCAfile cacert-2026-09-25.pem
CONNECTED(00000003)
80D71977B37E0000:error:4080004D:upki::certificate's key is compromised:upki-openssl/src/lib.rs:174:a certificate in the chain has a known-compromised key
80D71977B37E0000:error:0A000086:SSL routines:tls_post_process_server_certificate:certificate verify failed:../ssl/statem/statem_clnt.c:1889:

@codspeed

codspeed Bot commented Oct 21, 2026 •

Copy link
Copy Markdown

Merging this PR will regress 1 benchmark

1 warning

⚠️ 1 benchmark is dominated by system calls

System calls cannot be simulated consistently and are left out of the measure. The reported value understates the real cost. Switch to the Walltime instrument to measure them.

Learn more about system calls

See affected benchmarks

⚡ 1 improved benchmark
❌ 1 regressed benchmark
✅ 1 untouched benchmark

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

Benchmark BASE HEAD Efficiency
❌ revocation-input-from-certs 41 µs 52.9 µs -22.46%
⚡ revocation-check 25.5 ms 16.9 ms +51.07%

Tip

Investigate this regression by commenting @codspeedbot fix this regression on this PR, or directly use the CodSpeed MCP with your agent.


Comparing jbp-spki-denylist (22a3c42) with main (c2ba5e3)

Open in CodSpeed

@ctz
ctz marked this pull request as draft October 21, 2026 16:05
@ctz
ctz force-pushed the jbp-spki-denylist branch from 8d82cb2 to 22a3c42 Compare October 21, 2026 16:06
@cpu

cpu commented Oct 28, 2026

Copy link
Copy Markdown
Member

For the rustls/webpki version of this mechanism I used the raw SPKI DER because we didn't have a handle on a digest algorithm without a bunch of refactoring. In this crate since there's the graviola SHA256 impl on-hand wouldn't it be nicer to block by SPKI SHA256 to cut down on the size of the data we have to embed for the blocklist? Something like: db944b3

@orangepizza

Copy link
Copy Markdown

could you add global crl (with issuer-serial number pair) to allow digest root program provided revocation lists like onecrl/crlset?

@djc

djc commented Nov 2, 2026

Copy link
Copy Markdown
Member

could you add global crl (with issuer-serial number pair) to allow digest root program provided revocation lists like onecrl/crlset?

OneCRL support is discussed here:

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants