Skip to content

Take rustls 0.23.45 for RUSTSEC-2026-0285 - #79

Merged
sotashimozono merged 1 commit into
mainfrom
rustls-advisory
Sep 15, 2026
Merged

sotashimozono merged 1 commit into
mainfrom
rustls-advisory

Conversation

@sotashimozono

Copy link
Copy Markdown
Member

The 0.5.1 release PR went red on cargo audit and cargo deny — the advisory scan doing the job it exists for.

RUSTSEC-2026-0285. rustls accepted TLS 1.3 handshake messages sent at the wrong encryption level when they followed a key-changing message in the same record; a plaintext EncryptedExtensions packed in with the ServerHello was taken. RFC 8446 §5.1 requires terminating with unexpected_message instead.

rustls v0.23.44 -> v0.23.45
  └── tokio-rustls v0.26.5

Not exploitable here in the way the headline suggests: the transcript is still authenticated, so nobody alters or completes a handshake with it, and this daemon's TLS is a server on loopback dialled by a browser on the same machine. Fixed anyway — "our deployment happens not to reach it" is a reason to be unhurried, not a reason to ship a known-defective TLS stack in a mode whose whole purpose is being a real secure context.

Verified past the lockfile: the https e2e completes a real handshake against a real SSH host and reports isSecureContext, which is what would break if the bump had.

ok  under https the daemon reports what handshakes have done
ok  and over https the alias origin is a secure context

230 tests, 65 e2e checks, clippy clean.

The release for 0.5.1 went red on `cargo audit` and `cargo deny`, which is the
weekly advisory scan doing the job it exists for.

RUSTSEC-2026-0285: rustls accepted TLS 1.3 handshake messages sent at the wrong
encryption level when they followed a key-changing message in the same record --
a plaintext `EncryptedExtensions` packed in with the `ServerHello` was taken.
RFC 8446 section 5.1 requires terminating with `unexpected_message` instead.

Not exploitable here in the way the headline suggests. The transcript is still
authenticated, so nobody alters or completes a handshake with it; and this
daemon's TLS is a server on loopback that a browser on the same machine dials.
Fixing it anyway, because "our deployment happens not to reach it" is a reason
to be unhurried, not a reason to ship a known-defective TLS stack in a mode
whose entire purpose is being a real secure context.

Verified past the lockfile: the https e2e completes a real handshake against a
real host and reports `isSecureContext`, which is the thing that would break if
the bump had.

Signed-off-by: Souta <shimozono-sota631@g.ecc.u-tokyo.ac.jp>
@codecov

codecov Bot commented Sep 15, 2026

Copy link
Copy Markdown

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

@sotashimozono
sotashimozono merged commit be94ad0 into main Sep 15, 2026
23 checks passed
@sotashimozono
sotashimozono deleted the rustls-advisory branch September 15, 2026 07:40
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.

1 participant