Koffei

CVE-2026-43499

🟠 Medium

Linux kernel rtmutex and futex code. The record also carries three Siemens ProductCERT advisories, so the flaw is being tracked into Siemens products as well as distributions · Vulnerable from 2.6.39 onward (commit 8161239a). Fixed in 5.10.261, 5.15.212, 6.1.175, 6.6.140, 6.12.86, 6.18.27, 7.0.4 and 7.1 mainline

CVSS (Koffei note)
7.8
CVSS (v3.1, CVE List)
7.8
EPSS (2026-11-18)
5.372%
92.81th percentile

Exploitation

No confirmed exploitation. SSVC 22/08/2026: exploitation poc, automatable no, technical impact total ; not in CISA KEV. Neither CISA KEV nor its JSON feed could be reached in this run, so KEV status comes from the record's CISA-ADP container. Existing PoC (not opened), recorded from the proof-of-concept list supplied on 09/11/2026. Record and list agree here, both recording a proof of concept. The age of the bug is the notable part: it has been present since 2.6.39, which is 2011, so the fix had to be backported across eight stable branches

Timeline

  1. 21/05/2026CVE record published (CVE List)

Risk

7.8 High (v3.1, Linux as CNA) — AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H, EPSS 5.372% (92.71st percentile, scored 08/11/2026) — a notably high EPSS for a local kernel flaw, above the 92nd percentile (CWE-416 use after free, assigned by CISA-ADP: in the rtmutex code, remove_waiter() used the current task instead of the waiter's task during proxy-lock rollback from a futex requeue, corrupting the rbtree without the right lock and leaving a dangling pi_blocked_on pointer)

Fix

Yes — the per-branch stable releases above, or 7.1 mainline. An estate on a long-term branch must take the matching point release, since there is no single version that covers every branch

Official description

In the Linux kernel, the following vulnerability has been resolved: rtmutex: Use waiter::task instead of current in remove_waiter() remove_waiter() is used by the slowlock paths, but it is also used for proxy-lock rollback in rt_mutex_start_proxy_lock() when invoked from futex_requeue(). In the latter case waiter::task is not current, but remove_waiter() operates on current for the dequeue operation. That results in several problems: 1) the rbtree dequeue happens without waiter::task::pi_lock being held 2) the waiter task's pi_blocked_on state is not cleared, which leaves a dangling pointer primed for UAF around. 3) rt_mutex_adjust_prio_chain() operates on the wrong top priority waiter task Use waiter::task instead of current in all related operations in remove_waiter() to cure those problems. [ tglx: Fixup rt_mutex_adjust_prio_chain(), add a comment and amend the changelog ]

CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H