CVE-2026-35190 json
Out-of-Bounds Read in TLS 1.3 Key-Share Processing with Suite B Certificate Flag
published 2d ago modified yesterday also EUVD-2026-126929
- cvss
- 5.3
- version 3.1, CNA
- cisa kev
- Not listed
- not in CISA's catalog
- epss
- 0.2%
- EPSS 0.2% chance of exploitation in 30 days, percentile 10%
- public exploits
- 0
- none known
Summary
Issue summary: A server configured with the Suite B certificate flag through SSL_CTX_set_cert_flags() and with TLS 1.3 enabled may read past the end of a fixed group list when it processes a ClientHello whose supported_groups and key_share contain no Suite B group.
Impact summary: A remote client can send a ClientHello that makes the server read past the end of a static array. The values read are never returned to the client, so no memory is disclosed; on builds with sanitizers or hardened allocators the read may crash the process, causing Denial of Service.
CWE: CWE-125: Out-of-bounds Read
Description: When the Suite B certificate flag is set, OpenSSL restricts the accepted groups to P-256 and P-384 by substituting a fixed two-element list for the configured group list. The tuple machinery used for key-share selection was not updated for this substitution: tls1_get_group_tuples() still returns the configured tuples, which by default describe eight groups. The server-side walk in tls_parse_ctos_key_share() then advances through the two-element list using those tuple counts without checking the cumulative offset against the number of groups present.
Only applications that call SSL_CTX_set_cert_flags() with SSL_CERT_FLAG_SUITEB_128_LOS and leave TLS 1.3 enabled are affected. The openssl s_server -suiteB_128 option sets an X.509 verification parameter instead, so the stock command line server is not vulnerable. Default configurations do not reach this code.
A client that offers only groups outside the Suite B set, such as X25519, forces every tuple to be walked and the read to run past the array. The bytes read are compared against client group IDs and never sent back, so no memory is disclosed. On builds with address sanitizers or hardened allocators the read can abort the process; in ordinary release builds it reads a few bytes of adjacent constant data.
FIPS impact: no The affected code is outside the FIPS module boundary.
Patch / fix
- OpenSSL OpenSSL 3.5.10, 4.0.4, 4.1.1
Is CVE-2026-35190 in your stack? Fencer shows which of your repositories are affected as soon as an advisory lands, and opens the fix as a pull request.
Check with FencerCISA SSVC
- Exploitation
- none
- Automatable
- yes
- Technical impact
- partial
- Assessed
- yesterday
Exploit prediction
0.21% chance of exploitation in the next 30 days, higher than 10% of all scored CVEs (FIRST EPSS, 2026-11-18).
CVSS scores
| assigned by | version | score | vector |
|---|---|---|---|
| CISA ADP | 3.1 | 5.3 | CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:N/A:N |
| CNA | 3.1 | 5.3 | CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:N/A:N |
Affected products
- OpenSSL OpenSSL
Classification
- Weakness
-
- CWE-125 Out-of-bounds Read
- Technology feeds
References
-
github.com
/openssl/openssl/commit/ade64619df04056b3e5fd1c1befe9425a7c80eb6
patch
-
github.com
/openssl/openssl/commit/9a812981b927c5714b7c2e5d138618105cea2152
patch
-
github.com
/openssl/openssl/commit/9491b0d9cc7950af1f3dc89499bd5433b9cd4f43
patch
-
openssl-library.org
/news/secadv/20261117.txt
vendor-advisory
Sources
- CVE List CVE-2026-35190
- NVD CVE-2026-35190
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.