CVE-2026-90317: bpf: Invalidate RCU pointers after final spin unlock
In the Linux kernel, the following vulnerability has been resolved:
bpf: Invalidate RCU pointers after final spin unlock
In a sleepable BPF program, a spin lock can provide the only RCU protection for a kptr. The final bpfspinunlock() ends that protection, but the verifier leaves the pointer valid. Another CPU can then free the object before the pointer is used. A capability-limited runtime PoC triggered a taskstruct use-after-free in bpfgettaskstack().
Record whether the program is in an RCU-protected context before releasing the lock. Invalidate RCU-protected pointers only when the unlock leaves the final such context. This preserves valid pointers in non-sleepable programs and inside an explicit RCU read-side section.
Affected Software
Event History
Frequently Asked Questions
Which BPF programs are exposed to this issue?
The issue affects sleepable BPF programs where a spin lock is the only RCU protection for a kptr. It does not invalidate otherwise valid pointers in non-sleepable programs.
What condition makes the use-after-free possible?
The final bpf_spin_unlock() must end the last RCU-protected context while the verifier still considers an RCU-protected pointer valid. Another CPU can then free the referenced object before that pointer is used.
Does an explicit RCU read-side section prevent the pointer from being invalidated at unlock?
Yes. The fix preserves valid pointers when execution remains inside an explicit RCU read-side section after the spin lock is released.