On Mon, Jun 09, 2025 at 02:01:24PM -0400, Joel Fernandes wrote:
> During rcu_read_unlock_special(), if this happens during irq_exit(), we
> can lockup if an IPI is issued. This is because the IPI itself triggers
> the irq_exit() path causing a recursive lock up.
>
> This is precisely what Xiongfeng found when invoking a BPF program on
> the trace_tick_stop() tracepoint As shown in the trace below. Fix by
> using context-tracking to tell us if we're still in an IRQ.
> context-tracking keeps track of the IRQ until after the tracepoint, so
> it cures the issues.
>
> irq_exit()
> __irq_exit_rcu()
> /* in_hardirq() returns false after this */
> preempt_count_sub(HARDIRQ_OFFSET)
> tick_irq_exit()
@Frederic, while we are at it, what's the purpose of in_hardirq() in
tick_irq_exit()? For nested interrupt detection?
Regards,
Boqun
> tick_nohz_irq_exit()
> tick_nohz_stop_sched_tick()
> trace_tick_stop() /* a bpf prog is hooked on this trace point */
> __bpf_trace_tick_stop()
> bpf_trace_run2()
> rcu_read_unlock_special()
> /* will send a IPI to itself */
> irq_work_queue_on(&rdp->defer_qs_iw, rdp->cpu);
>
> A simple reproducer can also be obtained by doing the following in
> tick_irq_exit(). It will hang on boot without the patch:
>
> static inline void tick_irq_exit(void)
> {
> + rcu_read_lock();
> + WRITE_ONCE(current->rcu_read_unlock_special.b.need_qs, true);
> + rcu_read_unlock();
> +
>
> While at it, add some comments to this code.
>
> Reported-by: Xiongfeng Wang <wangxiongfe...@huawei.com>
> Closes:
> https://lore.kernel.org/all/9acd5f9f-6732-7701-6880-4b51190aa...@huawei.com/
> Tested-by: Xiongfeng Wang <wangxiongfe...@huawei.com>
> Signed-off-by: Joel Fernandes <joelagn...@nvidia.com>
[...]
