libev/ev_epoll.c

   1 /*
   2  * libev epoll fd activity backend
   3  *
   4  * Copyright (c) 2007 Marc Alexander Lehmann <libev@schmorp.de>
   5  * All rights reserved.
   6  *
   7  * Redistribution and use in source and binary forms, with or without
   8  * modification, are permitted provided that the following conditions are
   9  * met:
  10  *
  11  *     * Redistributions of source code must retain the above copyright
  12  *       notice, this list of conditions and the following disclaimer.
  13  *
  14  *     * Redistributions in binary form must reproduce the above
  15  *       copyright notice, this list of conditions and the following
  16  *       disclaimer in the documentation and/or other materials provided
  17  *       with the distribution.
  18  *
  19  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
  20  * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
  21  * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
  22  * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
  23  * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
  24  * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
  25  * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
  26  * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
  27  * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
  28  * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
  29  * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
  30  */
  31
  32 /*
  33  * general notes about epoll:
  34  *
  35  * a) epoll silently removes fds from the fd set. as nothing tells us
  36  *    that an fd has been removed otherwise, we have to continually
  37  *    "rearm" fds that we suspect *might* have changed (same
  38  *    problem with kqueue, but much less costly there).
  39  * b) the fact that ADD != MOD creates a lot of extra syscalls due to a)
  40  *    and seems not to have any advantage.
  41  * c) the inability to handle fork or file descriptors (think dup)
  42  *    limits the applicability over poll, so this is not a generic
  43  *    poll replacement.
  44  *
  45  * lots of "weird code" and complication handling in this file is due
  46  * to these design problems with epoll, as we try very hard to avoid
  47  * epoll_ctl syscalls for common usage patterns.
  48  */
  49
  50 #include <sys/epoll.h>
  51
  52 static void
  53 epoll_modify (EV_P_ int fd, int oev, int nev)
  54 {
  55   struct epoll_event ev;
  56
  57   /*
  58    * we handle EPOLL_CTL_DEL by ignoring it here
  59    * on the assumption that the fd is gone anyways
  60    * if that is wrong, we have to handle the spurious
  61    * event in epoll_poll.
  62    */
  63   if (!nev)
  64     return;
  65
  66   ev.data.u64 = fd; /* use u64 to fully initialise the struct, for nicer strace etc. */
  67   ev.events   = (nev & EV_READ  ? EPOLLIN  : 0)
  68               | (nev & EV_WRITE ? EPOLLOUT : 0);
  69
  70   if (expect_true (!epoll_ctl (backend_fd, oev ? EPOLL_CTL_MOD : EPOLL_CTL_ADD, fd, &ev)))
  71     return;
  72
  73   if (expect_true (errno == ENOENT))
  74     {
  75       /* on ENOENT the fd went away, so try to do the right thing */
  76       if (!nev)
  77         return;
  78
  79       if (!epoll_ctl (backend_fd, EPOLL_CTL_ADD, fd, &ev))
  80         return;
  81     }
  82   else if (expect_true (errno == EEXIST))
  83     {
  84       /* on EEXIST we ignored a previous DEL */
  85       if (!epoll_ctl (backend_fd, EPOLL_CTL_MOD, fd, &ev))
  86         return;
  87     }
  88
  89   fd_kill (EV_A_ fd);
  90 }
  91
  92 static void
  93 epoll_poll (EV_P_ ev_tstamp timeout)
  94 {
  95   int i;
  96   int eventcnt = epoll_wait (backend_fd, epoll_events, epoll_eventmax, (int)ceil (timeout * 1000.));
  97
  98   if (expect_false (eventcnt < 0))
  99     {
 100       if (errno != EINTR)
 101         syserr ("(libev) epoll_wait");
 102
 103       return;
 104     }
 105
 106   for (i = 0; i < eventcnt; ++i)
 107     {
 108       struct epoll_event *ev = epoll_events + i;
 109
 110       int fd   = ev->data.u64;
 111       int got  = (ev->events & (EPOLLOUT | EPOLLERR | EPOLLHUP) ? EV_WRITE : 0)
 112                | (ev->events & (EPOLLIN  | EPOLLERR | EPOLLHUP) ? EV_READ  : 0);
 113       int want = anfds [fd].events;
 114
 115       if (expect_false (got & ~want))
 116         {
 117           /* we received an event but are not interested in it, try mod or del */
 118           ev->events = (want & EV_READ  ? EPOLLIN  : 0)
 119                      | (want & EV_WRITE ? EPOLLOUT : 0);
 120
 121           epoll_ctl (backend_fd, want ? EPOLL_CTL_MOD : EPOLL_CTL_DEL, fd, ev);
 122         }
 123
 124       fd_event (EV_A_ fd, got);
 125     }
 126
 127   /* if the receive array was full, increase its size */
 128   if (expect_false (eventcnt == epoll_eventmax))
 129     {
 130       ev_free (epoll_events);
 131       epoll_eventmax = array_nextsize (sizeof (struct epoll_event), epoll_eventmax, epoll_eventmax + 1);
 132       epoll_events = (struct epoll_event *)ev_malloc (sizeof (struct epoll_event) * epoll_eventmax);
 133     }
 134 }
 135
 136 int inline_size
 137 epoll_init (EV_P_ int flags)
 138 {
 139   backend_fd = epoll_create (256);
 140
 141   if (backend_fd < 0)
 142     return 0;
 143
 144   fcntl (backend_fd, F_SETFD, FD_CLOEXEC);
 145
 146   backend_fudge  = 0.; /* kernel sources seem to indicate this to be zero */
 147   backend_modify = epoll_modify;
 148   backend_poll   = epoll_poll;
 149
 150   epoll_eventmax = 64; /* intiial number of events receivable per poll */
 151   epoll_events = (struct epoll_event *)ev_malloc (sizeof (struct epoll_event) * epoll_eventmax);
 152
 153   return EVBACKEND_EPOLL;
 154 }
 155
 156 void inline_size
 157 epoll_destroy (EV_P)
 158 {
 159   ev_free (epoll_events);
 160 }
 161
 162 void inline_size
 163 epoll_fork (EV_P)
 164 {
 165   close (backend_fd);
 166
 167   while ((backend_fd = epoll_create (256)) < 0)
 168     syserr ("(libev) epoll_create");
 169
 170   fcntl (backend_fd, F_SETFD, FD_CLOEXEC);
 171
 172   fd_rearm_all (EV_A);
 173 }
 174