How to deal with EPOLLERR and EPOLLHUP?

Question

I read the code in libevent epoll, here is the code:

       if (what & (EPOLLHUP|EPOLLERR)) {
            ev = EV_READ | EV_WRITE;
        } else {
            if (what & EPOLLIN)
                ev |= EV_READ;
            if (what & EPOLLOUT)
                ev |= EV_WRITE;
            if (what & EPOLLRDHUP)
                ev |= EV_CLOSED;
        }

To my understanding, when EPOLLERR or EPOLLHUP happens, the connection should be closed. But in the above code, when EPOLLHUP|EPOLLERR encounters, the event mask is set to EV_READ | EV_WRITE. So my question is :

What makes EPOLLERR and EPOLLHUP happen?
When EPOLLERR and EPOLLHUP happen, what should the program do in the event handle function? And please explain the reason behind it in detail.

Thanks in advance!

Here: https://stackoverflow.com/questions/52976152/tcp-when-is-epollhup-generated — anonymous, Aug 11 '23 at 19:25

score 9 · Answer 1 · answered Mar 23 '15 at 09:18

What makes EPOLLERR and EPOLLHUP happen?

man epoll_ctl

   EPOLLERR
          Error condition happened on the associated file descriptor.
          epoll_wait(2) will always wait for this event; it is not
          necessary to set it in events.

   EPOLLHUP
          Hang up happened on the associated file descriptor.
          epoll_wait(2) will always wait for this event; it is not
          necessary to set it in events.

When EPOLLERR and EPOLLHUP happen, what should the program do in the event handle function?

As libevent passes the events EV_READ | EV_WRITE in this case, the callback function calls e. g. recv(), which likely returns 0 when the peer has performed an orderly shutdown (EPOLLHUP), or -1 if an error occurred (EPOLLERR); the program might then clean up the connection, and if it's a client, possibly reestablish it.

How to deal with EPOLLERR and EPOLLHUP?

1 Answers1