CVE-2026-43074: eventpoll: defer struct eventpoll free to RCU grace period
In the Linux kernel, the following vulnerability has been resolved:
eventpoll: defer struct eventpoll free to RCU grace period
In certain situations, ep_free() in eventpoll.c will kfree the epi->ep
eventpoll struct while it still being used by another concurrent thread.
Defer the kfree() to an RCU callback to prevent UAF.
Security readout for executives and security teams
Plain-English summary
A Linux kernel race can free event-polling memory while another thread still uses it. A local, low-privileged user could potentially trigger severe confidentiality, integrity, or availability impact. The supplied CVSS score is 7.8 (high).
Executive priority
Treat as a high-priority kernel update, especially on shared servers, developer platforms, and untrusted-workload hosts. It is not presented as remotely exploitable or actively exploited, so prioritize according to local-code exposure and normal emergency kernel-change controls.
Technical view
Concurrent activity can cause ep_free() to release an eventpoll structure still referenced by another thread, creating a use-after-free condition. The kernel fix defers memory release through an RCU callback, ensuring an RCU grace period before reclamation.
Likely exposure
Systems running affected Linux kernels and allowing untrusted or low-privileged local code are the primary concern. The bundle's version data is ambiguous, so determine exposure using distribution advisories and whether the running kernel contains an applicable referenced fix.
Exploitation context
The CVSS vector describes local access, low privileges, low complexity, and no user interaction. The bundle marks this CVE as absent from KEV and provides no evidence of active exploitation or a public exploit. Absence from KEV does not prove exploitation is impossible.
Researcher notes
The evidence establishes a concurrency-driven use-after-free involving epi->ep during ep_free(). The remediation changes reclamation timing rather than removing eventpoll functionality. The bundle does not identify trigger conditions beyond certain concurrent situations, exploitation outcomes, public proof-of-concept availability, or reliable distribution-specific affected ranges.
Mitigation direction
Install a vendor-supported kernel update containing the applicable eventpoll RCU fix.
Reboot systems into the updated kernel after installation.
Prioritize multi-user hosts and systems executing untrusted local workloads.
Consult Linux distribution guidance because the supplied affected-version boundaries are unclear.
Validation and detection
Record each system's running kernel release, build, and distribution package revision.
Map kernel packages to distribution advisories or the referenced stable commits.
Confirm the installed update contains the eventpoll RCU fix.
Verify each system rebooted into the corrected kernel, not merely installed it.
Document unresolved systems where vendor fix status remains unavailable.
Generated from the cited source records. This long-tail analysis has not been individually reviewed by a named human.
Potential ATT&CK relevance
Conservative CVE-to-ATT&CK context
These mappings and lookup hints may be relevant to the vulnerability behavior, CWE, affected product, or exposure path. Glexia-inferred context is not an official MITRE, ATT&CK, CWE, or CVE Program mapping.
ATT&CK lookup starting points
Use these exact CWE pages and searches to review the Glexia ATT&CK library from this CVE's weakness and description context.
cve · low confidence lookup
CVE-2026-43074 mapping review
Open the CVE-to-ATT&CK bridge for reviewed, inferred, or future official mappings tied to this CVE.
These fields come from the CVE record and ADP containers, not from Glexia's Take. They preserve time-varying source decisions such as CISA SSVC, KEV status, CVSS metrics, and provider references.
1CVSS vectors
3Timeline events
0ADP providers
8Source links
CVSS vector scores
1 official score
We collect every scored CVSS vector available in the official CNA and ADP containers. When more than one version is present, the table keeps the source vectors side by side instead of collapsing them into the highest score.