1 |
/* |
2 |
* libev epoll fd activity backend |
3 |
* |
4 |
* Copyright (c) 2007,2008 Marc Alexander Lehmann <libev@schmorp.de> |
5 |
* All rights reserved. |
6 |
* |
7 |
* Redistribution and use in source and binary forms, with or without modifica- |
8 |
* tion, are permitted provided that the following conditions are met: |
9 |
* |
10 |
* 1. Redistributions of source code must retain the above copyright notice, |
11 |
* this list of conditions and the following disclaimer. |
12 |
* |
13 |
* 2. Redistributions in binary form must reproduce the above copyright |
14 |
* notice, this list of conditions and the following disclaimer in the |
15 |
* documentation and/or other materials provided with the distribution. |
16 |
* |
17 |
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR IMPLIED |
18 |
* WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MER- |
19 |
* CHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO |
20 |
* EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPE- |
21 |
* CIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, |
22 |
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; |
23 |
* OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, |
24 |
* WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTH- |
25 |
* ERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED |
26 |
* OF THE POSSIBILITY OF SUCH DAMAGE. |
27 |
* |
28 |
* Alternatively, the contents of this file may be used under the terms of |
29 |
* the GNU General Public License ("GPL") version 2 or any later version, |
30 |
* in which case the provisions of the GPL are applicable instead of |
31 |
* the above. If you wish to allow the use of your version of this file |
32 |
* only under the terms of the GPL and not to allow others to use your |
33 |
* version of this file under the BSD license, indicate your decision |
34 |
* by deleting the provisions above and replace them with the notice |
35 |
* and other provisions required by the GPL. If you do not delete the |
36 |
* provisions above, a recipient may use your version of this file under |
37 |
* either the BSD or the GPL. |
38 |
*/ |
39 |
|
40 |
/* |
41 |
* general notes about epoll: |
42 |
* |
43 |
* a) epoll silently removes fds from the fd set. as nothing tells us |
44 |
* that an fd has been removed otherwise, we have to continually |
45 |
* "rearm" fds that we suspect *might* have changed (same |
46 |
* problem with kqueue, but much less costly there). |
47 |
* b) the fact that ADD != MOD creates a lot of extra syscalls due to a) |
48 |
* and seems not to have any advantage. |
49 |
* c) the inability to handle fork or file descriptors (think dup) |
50 |
* limits the applicability over poll, so this is not a generic |
51 |
* poll replacement. |
52 |
* |
53 |
* lots of "weird code" and complication handling in this file is due |
54 |
* to these design problems with epoll, as we try very hard to avoid |
55 |
* epoll_ctl syscalls for common usage patterns and handle the breakage |
56 |
* ensuing from receiving events for closed and otherwise long gone |
57 |
* file descriptors. |
58 |
*/ |
59 |
|
60 |
#include <sys/epoll.h> |
61 |
|
62 |
static void |
63 |
epoll_modify (EV_P_ int fd, int oev, int nev) |
64 |
{ |
65 |
struct epoll_event ev; |
66 |
unsigned char oldmask; |
67 |
|
68 |
/* |
69 |
* we handle EPOLL_CTL_DEL by ignoring it here |
70 |
* on the assumption that the fd is gone anyways |
71 |
* if that is wrong, we have to handle the spurious |
72 |
* event in epoll_poll. |
73 |
* the fd is later added, we try to ADD it, and, if that |
74 |
* fails, we assume it still has the same eventmask. |
75 |
*/ |
76 |
if (!nev) |
77 |
return; |
78 |
|
79 |
oldmask = anfds [fd].emask; |
80 |
anfds [fd].emask = nev; |
81 |
|
82 |
/* store the generation counter in the upper 32 bits */ |
83 |
ev.data.u64 = fd | ((uint64_t)++anfds [fd].egen << 32); |
84 |
ev.events = (nev & EV_READ ? EPOLLIN : 0) |
85 |
| (nev & EV_WRITE ? EPOLLOUT : 0); |
86 |
|
87 |
if (expect_true (!epoll_ctl (backend_fd, oev ? EPOLL_CTL_MOD : EPOLL_CTL_ADD, fd, &ev))) |
88 |
return; |
89 |
|
90 |
if (expect_true (errno == ENOENT)) |
91 |
{ |
92 |
/* if ENOENT then the fd went away, so try to do the right thing */ |
93 |
if (!nev) |
94 |
goto dec_egen; |
95 |
|
96 |
if (!epoll_ctl (backend_fd, EPOLL_CTL_ADD, fd, &ev)) |
97 |
return; |
98 |
} |
99 |
else if (expect_true (errno == EEXIST)) |
100 |
{ |
101 |
/* EEXIST means we ignored a previous DEL, but the fd is still active */ |
102 |
/* if the kernel mask is the same as the new mask, we assume it hasn't changed */ |
103 |
if (oldmask == nev) |
104 |
goto dec_egen; |
105 |
|
106 |
if (!epoll_ctl (backend_fd, EPOLL_CTL_MOD, fd, &ev)) |
107 |
return; |
108 |
} |
109 |
|
110 |
fd_kill (EV_A_ fd); |
111 |
|
112 |
dec_egen: |
113 |
/* we didn't successfully call epoll_ctl, so decrement the generation counter again */ |
114 |
--anfds [fd].egen; |
115 |
} |
116 |
|
117 |
static void |
118 |
epoll_poll (EV_P_ ev_tstamp timeout) |
119 |
{ |
120 |
int i; |
121 |
int eventcnt = epoll_wait (backend_fd, epoll_events, epoll_eventmax, (int)ceil (timeout * 1000.)); |
122 |
|
123 |
if (expect_false (eventcnt < 0)) |
124 |
{ |
125 |
if (errno != EINTR) |
126 |
ev_syserr ("(libev) epoll_wait"); |
127 |
|
128 |
return; |
129 |
} |
130 |
|
131 |
for (i = 0; i < eventcnt; ++i) |
132 |
{ |
133 |
struct epoll_event *ev = epoll_events + i; |
134 |
|
135 |
int fd = (uint32_t)ev->data.u64; /* mask out the lower 32 bits */ |
136 |
int want = anfds [fd].events; |
137 |
int got = (ev->events & (EPOLLOUT | EPOLLERR | EPOLLHUP) ? EV_WRITE : 0) |
138 |
| (ev->events & (EPOLLIN | EPOLLERR | EPOLLHUP) ? EV_READ : 0); |
139 |
|
140 |
/* check for spurious notification */ |
141 |
if (expect_false (anfds [fd].egen != (unsigned char)(ev->data.u64 >> 32))) |
142 |
{ |
143 |
/* recreate kernel state */ |
144 |
postfork = 1; |
145 |
continue; |
146 |
} |
147 |
|
148 |
if (expect_false (got & ~want)) |
149 |
{ |
150 |
anfds [fd].emask = want; |
151 |
|
152 |
/* we received an event but are not interested in it, try mod or del */ |
153 |
/* I don't think we ever need MOD, but let's handle it anyways */ |
154 |
ev->events = (want & EV_READ ? EPOLLIN : 0) |
155 |
| (want & EV_WRITE ? EPOLLOUT : 0); |
156 |
|
157 |
if (epoll_ctl (backend_fd, want ? EPOLL_CTL_MOD : EPOLL_CTL_DEL, fd, ev)) |
158 |
{ |
159 |
postfork = 1; /* an error occured, recreate kernel state */ |
160 |
continue; |
161 |
} |
162 |
} |
163 |
|
164 |
fd_event (EV_A_ fd, got); |
165 |
} |
166 |
|
167 |
/* if the receive array was full, increase its size */ |
168 |
if (expect_false (eventcnt == epoll_eventmax)) |
169 |
{ |
170 |
ev_free (epoll_events); |
171 |
epoll_eventmax = array_nextsize (sizeof (struct epoll_event), epoll_eventmax, epoll_eventmax + 1); |
172 |
epoll_events = (struct epoll_event *)ev_malloc (sizeof (struct epoll_event) * epoll_eventmax); |
173 |
} |
174 |
} |
175 |
|
176 |
int inline_size |
177 |
epoll_init (EV_P_ int flags) |
178 |
{ |
179 |
backend_fd = epoll_create (256); |
180 |
|
181 |
if (backend_fd < 0) |
182 |
return 0; |
183 |
|
184 |
fcntl (backend_fd, F_SETFD, FD_CLOEXEC); |
185 |
|
186 |
backend_fudge = 0.; /* kernel sources seem to indicate this to be zero */ |
187 |
backend_modify = epoll_modify; |
188 |
backend_poll = epoll_poll; |
189 |
|
190 |
epoll_eventmax = 64; /* initial number of events receivable per poll */ |
191 |
epoll_events = (struct epoll_event *)ev_malloc (sizeof (struct epoll_event) * epoll_eventmax); |
192 |
|
193 |
return EVBACKEND_EPOLL; |
194 |
} |
195 |
|
196 |
void inline_size |
197 |
epoll_destroy (EV_P) |
198 |
{ |
199 |
ev_free (epoll_events); |
200 |
} |
201 |
|
202 |
void inline_size |
203 |
epoll_fork (EV_P) |
204 |
{ |
205 |
close (backend_fd); |
206 |
|
207 |
while ((backend_fd = epoll_create (256)) < 0) |
208 |
ev_syserr ("(libev) epoll_create"); |
209 |
|
210 |
fcntl (backend_fd, F_SETFD, FD_CLOEXEC); |
211 |
|
212 |
fd_rearm_all (EV_A); |
213 |
} |
214 |
|