On Mon, Apr 29, 2019 at 04:20:30PM +0200, Oleg Nesterov wrote:
On 04/29, Joel Fernandes wrote:
However, in your code above, it is avoided because we get:
Task A (poller) Task B (exiting task being polled)
poll() called add_wait_queue() exit_state is set to non-zero read exit_state remove_wait_queue() wake_up_all()
just to clarify... No, sys_poll() path doesn't do remove_wait_queue() until it returns to user mode, and that is why we can't race with set-exit_code + wake_up().
I didn't follow what you mean, the removal from the waitqueue happens in free_poll_entry() called from poll_freewait() which happens from do_sys_poll() which is before the syscall returns to user mode. Could you explain more?
pidfd_poll() can race with the exiting task, miss exit_code != 0, and return zero. However, do_poll() won't block after that and pidfd_poll() will be called again.
Here also I didn't follow what you mean. If exit_code is read as 0 in pidfd_poll(), then in do_poll() the count will be 0 and it will block in poll_schedule_timeout(). Right? But above you're saying it wont block. Also if you could show a timing diagram of this different race you're talking about, that will make things clear. It is a bit hard for me to picture otherwise.
Also, I will use task_pid() for getting the pid from the task, as you suggest in the other thread.
thanks,
- Joel