Linux kernel BPF: bpf_get_stackid runs get_perf_callchain preemptible, exposing its trace buffer

Impact

bpf_get_stackid() called get_perf_callchain() and then read the returned per-CPU trace-entry buffer without disabling preemption, so the task can be scheduled off and the buffer reused underneath it. The upstream record is a one-paragraph fix with no exploitation analysis; NVD scores it as local high-impact, but the reachable consequence is a corrupted or attacker-influenced stack-trace buffer rather than a demonstrated escalation path. It matters on GPU nodes only insofar as fleet-wide observability agents (profilers, continuous-profiling daemons, eBPF security agents) run stack-map programs on every host, so the affected code path is loaded on essentially all of them.

Who can reach it

Local, and privileged: loading a BPF program that calls bpf_get_stackid requires CAP_BPF plus CAP_PERFMON (or root). Unprivileged tenants cannot reach it unless the node permits unprivileged BPF.

What to do

Pick up the stable-kernel update that adds the preempt_disable() around the callchain use. Kernel patch means drain and reboot the node; there is no module reload or daemon restart that fixes it. No mitigation is needed on nodes where only trusted observability agents hold CAP_BPF.

References

Related entries

• Linux kernel mm/migrate_device: compound folio at the end of a range writes past the PFN arrayCVE-2026-89967 · Linux kernel mm/migrate_device (device-private page migration)High
• Linux kernel megaraid_sas: NVMe PRP list overruns the chain frame and corrupts other in-flight commandsCVE-2026-90008 · Linux kernel megaraid_sas (NVMe PRP chain frame)High
• Linux kernel SCSI bsg: TOCTOU on the shared io_uring SQE overflows the SCSI command bufferCVE-2026-90009 · Linux kernel SCSI bsg (io_uring passthrough command setup)High
• Linux kernel bsg: sense data copy ignores max_response_len and overruns the user bufferCVE-2026-90010 · Linux kernel SCSI bsg (io_uring sense copy)High
• Linux kernel page allocator: unsafe spin_trylock in NMI context on uniprocessor buildsCVE-2026-90046 · Linux kernel mm/page_alloc (free_pages_nolock in NMI on UP builds)High
• Linux kernel virtio-gpu: unvalidated EDID block offset lets a malicious backend read past a kernel bufferCVE-2026-68255 · Linux kernel drm/virtio (virtio_get_edid_block response bounds)High