CVE-2024-43869 — perf: Fix event leak upon exec and file release
In the Linux kernel, the following vulnerability has been resolved: perf: Fix event leak upon exec and file release The perf pending task work is never waited upon the matchin…
Published: 2024-08-21 · CVSS: Unscored
In the Linux kernel, the following vulnerability has been resolved: perf: Fix event leak upon exec and file release The perf pending task work is never waited upon the matching event release. In the case of a child event, released via free_event() directly, this can potentially result in a leaked event, such as in the following scenario that doesn't even require a weak IRQ work implementation to trigger: schedule() prepare_task_switch() =======> <NMI> perf_event_overflow() event->pending_sigtrap = ... irq_work_queue(&event->pending_irq) <======= </NMI> perf_event_task_sched_out() event_sched_out() event->pending_sigtrap = 0; atomic_long_inc_not_zero(&event->refcount) task_work_add(&event->pending_task) finish_lock_switch() =======> <IRQ> perf_pending_irq() //do nothing, rely on pending task work <======= </IRQ> begin_new_exec() perf_event_exit_task() perf_event_exit_event() // If is child event free_event() WARN(atomic_long_cmpxchg(&event->refcount, 1, 0) != 1) // event is leaked Similar scenarios can also happen with perf_event_remove_on_exec() or simply against concurrent perf_event_release(). Fix this with synchonizing against the possibly remaining pending task work while freeing the event, just like is done with remaining pending IRQ work. This means that the pending task callback neither need nor should hold a reference to the event, preventing it from ever beeing freed.
Source: CVE Program · Revision 8b16575c39f1e1810c5e863010da073d2d3285e7