Re: wait_on_page_bit_common(TASK_KILLABLE, EXCLUSIVE) can miss wakeup?

[Oleg Nesterov]

On 06/30, Oleg Nesterov wrote:
>
> may be someting like this

or this ...

        for (;;) {
                int intr = 0;

                spin_lock_irq(&q->lock);
                if (signal_pending_state(state, current)) {
                        /* see the comment in prepare_to_wait_event() */
                        list_del_init(&wait->entry);
                        intr = 1;
                } else {
                        if (likely(list_empty(&wait->entry))) {
                                __add_wait_queue_entry_tail(q, wait);
                                SetPageWaiters(page);
                        }
                        set_current_state(state);
                }
                spin_unlock_irq(&q->lock);

                bit_is_set = test_bit(bit_nr, &page->flags);
                if (behavior == DROP)
                        put_page(page);

                if (!bit_is_set) {
                        if (behavior != EXCLUSIVE)
                                break;
                        if (!test_and_set_bit_lock(bit_nr, &page->flags))
                                break;
                }

                if (intr) {
                        ret = -EINTR;
                        break;
                }

                io_schedule();

                if (behavior == DROP) {
                        /*
                         * We can no longer safely access page->flags:
                         * even if CONFIG_MEMORY_HOTREMOVE is not enabled,
                         * there is a risk of waiting forever on a page reused
                         * for something that keeps it locked indefinitely.
                         * But best check for -EINTR before breaking.
                         */
                        if (signal_pending_state(state, current))
                                ret = -EINTR;
                        break;
                }
        }

Oleg.