GHSA-9w4r-xc43-v236 json
Bare LF in chunked trailers can consume pipelined requests or stall body decoding
published yesterday modified 18h ago also RUSTSEC-2026-0552
- cvss
- —
- not scored
- cisa kev
- Not listed
- not in CISA's catalog
- epss
- —
- no score, EPSS covers CVEs only
- public exploits
- 0
- none known
Summary
Affected versions of `actix-http` skip a bare line feed (LF) inside an HTTP/1.1 chunked trailer instead of rejecting the invalid line ending. Trailer lines must end with a carriage return and line feed (CRLF).
For example, after `0\r\nX\n`, the parser continues reading trailer data. If a request follows on the same connection, its head can be consumed as part of the trailer, so that request is never dispatched. If no more bytes arrive, body decoding waits for more input instead of reporting an error. A remote client can trigger these effects without authentication.
Request smuggling additionally requires an HTTP intermediary that forwards the malformed body but uses a different request boundary. The tested HTTP intermediaries either rejected the input or consumed and normalized it before forwarding. Proxy access-control bypass, cache poisoning, and effects on other users were not demonstrated.
Upgrade to `actix-http` 3.21.1 or later. The fix rejects bare LF and invalid control bytes in trailers before consuming the following request, while preserving partial reads and valid CRLF-terminated trailers.
Patch / fix
Is GHSA-9w4r-xc43-v236 in your dependencies? Fencer shows which of your repositories are affected as soon as an advisory lands, and opens the fix as a pull request.
Check with FencerAffected packages
| ecosystem | package | affected | fixed in |
|---|---|---|---|
| crates.io | actix-http | >=0.0.0-0, <3.21.1 |
3.21.1 |
References
Sources
Fencer Intel keeps one entry per vulnerability: the CVE id wins, then GHSA, and records that alias each other are merged, except that two CVEs always keep separate entries. The same data is available as JSON.