GPU VulnDB

Database/Kernel, userspace & hypervisor

Linux kernel BPF: resilient lock release in rbtree callback causes use-after-free

CVSS 7.8CVE-2026-98060Kernel, userspace & hypervisorcurated

Impact

The BPF verifier did not stop a BPF rbtree comparison callback from releasing the tree's root lock through the resilient-lock kfuncs. Another CPU could then remove and free a node while the tree walk was still using it, a use-after-free in kernel memory that could lead to a crash or memory corruption. Only BPF programs that use rbtrees guarded by resilient locks can reach it. Unprivileged BPF is disabled on most distributions, so the realistic exposure is processes or containers holding CAP_BPF/CAP_SYS_ADMIN, such as observability or networking agents.

Who can reach it

A local user or container able to load BPF programs (CAP_BPF or equivalent). Not reachable by an ordinary unprivileged tenant where unprivileged BPF is disabled.

What to do

Install a kernel that contains the stable fix commits listed below, then drain and reboot each node. Until then, make sure unprivileged BPF stays disabled (kernel.unprivileged_bpf_disabled) and do not give tenant containers CAP_BPF or CAP_SYS_ADMIN.

References

Related entries

All Kernel, userspace & hypervisor entries

This entry is curated: imported from vendor advisories with machine assistance, not yet individually verified. Confirm against your vendor's advisory before acting, and report anything wrong.