Repository navigation
tls: bootstrap refuses a CCS inside a fragmented ServerHello - #13
Merged
Merged
Conversation
…essage RFC 8446 §5 (drop a CCS "at any time") and §5.1 (no record interleaved in a fragmented handshake message) conflict for a CCS between ServerHello fragments. Drop it only at a message boundary, as the connection dispatcher already does (#8); inside a message it is fatal with unexpected_message. BoringSSL and NSS also refuse it there. ERRATA entry added; not reported upstream. The implementation still drops it, so hello::bootstrap does not verify against this spec until the next commit. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
RFC 8446 §5.1; see docs/ERRATA.md. The Ccs arm of handle_record and record_step now drop only at a message boundary, matching bootstrap_spec. Tests updated; adds the certificate lemma used by the next commit. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Point four mutations at the module holding their killing lemma (hello::certificates, auth::certificates) so they no longer survive. Retarget alert-closure-any-level (restore warning-only user_canceled), missing-version-only-sh (now legacy-server-hello-extensions-without-supported-versions), has-ext-first-only (re-certified through the legacy ServerHello predicate) and server-hello-extensions-floor (floor raised to 6; empty extension vector). New lemmas in hello::certificates and lemma_server_hello. Remove certificate-context: 8b04e0b deliberately dropped the generic parser's empty-context check; the initial-authentication rule now lives in auth::spec::initial_certificate_context_valid, covered by auth-initial-server-certificate-empty-context. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
atiernan
marked this pull request as ready for review
October 27, 2026 07:12
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Stacked on #12. This applies #8's rule to the plaintext bootstrap as well: a compatibility CCS is dropped only at a handshake message boundary, and inside a fragmented ServerHello or HRR it is fatal with
unexpected_message.RFC 8446 §5 ("MUST simply drop it" at any time) and §5.1 (no record interleaved inside a fragmented handshake message) conflict for this case. We take the stricter reading. The new
docs/ERRATA.mdentry records this and is marked "Reported upstream: not yet". On this path, BoringSSL (Chrome) and NSS (Firefox) also refuse such a CCS.Commits
57896d7spec: the Ccs arm ofreceiveinbootstrap_spec.rs, plus a modelling-choice note and the ERRATA entry. It was committed before the implementation (rule 3).dd67840fix: theFrame::Ccsarm inbootstrap.rsand the matchingrecord_steparm inbootstrap_lemmas.rs. In the tests, the old test expected the drop, so it is replaced, and a new test covers a CCS at a boundary before ServerHello.923f001mutationbootstrap-ccs-dropped-mid-message, with certificate lemmalemma_ccs_fatal_mid_messageinhello/certificates.rs.Local results (per module; I re-ran these myself after cleaning the Verus cache)
hello::bootstrap15 verified, 0 errors.hello::bootstrap_lemmas4 verified, 0 errors.hello::certificates3 verified, 0 errors.cargo test -p clazel-tls: 86 passed, 0 failed. Strict clippy,cargo fmt --checkandcargo xtask checkare clean.cargo xtask mutatehas not been run.No public contract changed. No full verify has run, and
clazel-tlsstays at tier 0.🤖 Generated with Claude Code