Zqiang caught a snafu in the rude grace period, if a too small grace period was reached, a UAF could occur. However, his patch used a WARN_ONCE() on schedule_on_each_cpu() OOM failures, which would not have handled the failure properly.
Instead of throwing a warning or relying on dynamic alloc_percpu(), replace the dynamic allocation with static per cpu work_struct instances. Using schedule_work_on() and flush_work() keeps everything in task context, guarantees the required context switch on every online CPU, and stays zero allocation so it cannot fail under OOM. Suggested-by: Zqiang <[email protected]> Signed-off-by: Bradley Morgan <[email protected]> --- kernel/rcu/tasks.h | 16 +++++++++++++++- 1 file changed, 15 insertions(+), 1 deletion(-) diff --git a/kernel/rcu/tasks.h b/kernel/rcu/tasks.h index 627295396cd9..f81c0ea0b92a 100644 --- a/kernel/rcu/tasks.h +++ b/kernel/rcu/tasks.h @@ -1348,11 +1348,25 @@ static void rcu_tasks_be_rude(struct work_struct *work) { } +static DEFINE_PER_CPU(struct work_struct, tasks_rude_work); + // Wait for one rude RCU-tasks grace period. static void rcu_tasks_rude_wait_gp(struct rcu_tasks *rtp) { + int cpu; + + guard(cpus_read_lock)(); rtp->n_ipis += cpumask_weight(cpu_online_mask); - schedule_on_each_cpu(rcu_tasks_be_rude); + + for_each_online_cpu(cpu) { + struct work_struct *work = &per_cpu(tasks_rude_work, cpu); + + INIT_WORK(work, rcu_tasks_be_rude); + schedule_work_on(cpu, work); + } + + for_each_online_cpu(cpu) + flush_work(&per_cpu(tasks_rude_work, cpu)); } static void call_rcu_tasks_rude(struct rcu_head *rhp, rcu_callback_t func); -- 2.53.0
