1TIMERFD_CREATE(2) Linux Programmer's Manual TIMERFD_CREATE(2)
2
3
4
6 timerfd_create, timerfd_settime, timerfd_gettime - timers that notify
7 via file descriptors
8
10 #include <sys/timerfd.h>
11
12 int timerfd_create(int clockid, int flags);
13
14 int timerfd_settime(int fd, int flags,
15 const struct itimerspec *new_value,
16 struct itimerspec *old_value);
17
18 int timerfd_gettime(int fd, struct itimerspec *curr_value);
19
21 These system calls create and operate on a timer that delivers timer
22 expiration notifications via a file descriptor. They provide an alter‐
23 native to the use of setitimer(2) or timer_create(2), with the advan‐
24 tage that the file descriptor may be monitored by select(2), poll(2),
25 and epoll(7).
26
27 The use of these three system calls is analogous to the use of
28 timer_create(2), timer_settime(2), and timer_gettime(2). (There is no
29 analog of timer_getoverrun(2), since that functionality is provided by
30 read(2), as described below.)
31
32 timerfd_create()
33 timerfd_create() creates a new timer object, and returns a file
34 descriptor that refers to that timer. The clockid argument specifies
35 the clock that is used to mark the progress of the timer, and must be
36 either CLOCK_REALTIME or CLOCK_MONOTONIC. CLOCK_REALTIME is a settable
37 system-wide clock. CLOCK_MONOTONIC is a nonsettable clock that is not
38 affected by discontinuous changes in the system clock (e.g., manual
39 changes to system time). The current value of each of these clocks can
40 be retrieved using clock_gettime(2).
41
42 Starting with Linux 2.6.27, the following values may be bitwise ORed in
43 flags to change the behavior of timerfd_create():
44
45 TFD_NONBLOCK Set the O_NONBLOCK file status flag on the new open file
46 description. Using this flag saves extra calls to
47 fcntl(2) to achieve the same result.
48
49 TFD_CLOEXEC Set the close-on-exec (FD_CLOEXEC) flag on the new file
50 descriptor. See the description of the O_CLOEXEC flag in
51 open(2) for reasons why this may be useful.
52
53 In Linux versions up to and including 2.6.26, flags must be specified
54 as zero.
55
56 timerfd_settime()
57 timerfd_settime() arms (starts) or disarms (stops) the timer referred
58 to by the file descriptor fd.
59
60 The new_value argument specifies the initial expiration and interval
61 for the timer. The itimer structure used for this argument contains
62 two fields, each of which is in turn a structure of type timespec:
63
64 struct timespec {
65 time_t tv_sec; /* Seconds */
66 long tv_nsec; /* Nanoseconds */
67 };
68
69 struct itimerspec {
70 struct timespec it_interval; /* Interval for periodic timer */
71 struct timespec it_value; /* Initial expiration */
72 };
73
74 new_value.it_value specifies the initial expiration of the timer, in
75 seconds and nanoseconds. Setting either field of new_value.it_value to
76 a nonzero value arms the timer. Setting both fields of
77 new_value.it_value to zero disarms the timer.
78
79 Setting one or both fields of new_value.it_interval to nonzero values
80 specifies the period, in seconds and nanoseconds, for repeated timer
81 expirations after the initial expiration. If both fields of
82 new_value.it_interval are zero, the timer expires just once, at the
83 time specified by new_value.it_value.
84
85 The flags argument is either 0, to start a relative timer
86 (new_value.it_value specifies a time relative to the current value of
87 the clock specified by clockid), or TFD_TIMER_ABSTIME, to start an
88 absolute timer (new_value.it_value specifies an absolute time for the
89 clock specified by clockid; that is, the timer will expire when the
90 value of that clock reaches the value specified in new_value.it_value).
91
92 If the old_value argument is not NULL, then the itimerspec structure
93 that it points to is used to return the setting of the timer that was
94 current at the time of the call; see the description of timerfd_get‐
95 time() following.
96
97 timerfd_gettime()
98 timerfd_gettime() returns, in curr_value, an itimerspec structure that
99 contains the current setting of the timer referred to by the file
100 descriptor fd.
101
102 The it_value field returns the amount of time until the timer will next
103 expire. If both fields of this structure are zero, then the timer is
104 currently disarmed. This field always contains a relative value,
105 regardless of whether the TFD_TIMER_ABSTIME flag was specified when
106 setting the timer.
107
108 The it_interval field returns the interval of the timer. If both
109 fields of this structure are zero, then the timer is set to expire just
110 once, at the time specified by curr_value.it_value.
111
112 Operating on a timer file descriptor
113 The file descriptor returned by timerfd_create() supports the following
114 operations:
115
116 read(2)
117 If the timer has already expired one or more times since its
118 settings were last modified using timerfd_settime(), or since
119 the last successful read(2), then the buffer given to read(2)
120 returns an unsigned 8-byte integer (uint64_t) containing the
121 number of expirations that have occurred. (The returned value
122 is in host byte order, i.e., the native byte order for integers
123 on the host machine.)
124
125 If no timer expirations have occurred at the time of the
126 read(2), then the call either blocks until the next timer expi‐
127 ration, or fails with the error EAGAIN if the file descriptor
128 has been made nonblocking (via the use of the fcntl(2) F_SETFL
129 operation to set the O_NONBLOCK flag).
130
131 A read(2) will fail with the error EINVAL if the size of the
132 supplied buffer is less than 8 bytes.
133
134 poll(2), select(2) (and similar)
135 The file descriptor is readable (the select(2) readfds argument;
136 the poll(2) POLLIN flag) if one or more timer expirations have
137 occurred.
138
139 The file descriptor also supports the other file-descriptor mul‐
140 tiplexing APIs: pselect(2), ppoll(2), and epoll(7).
141
142 close(2)
143 When the file descriptor is no longer required it should be
144 closed. When all file descriptors associated with the same
145 timer object have been closed, the timer is disarmed and its
146 resources are freed by the kernel.
147
148 fork(2) semantics
149 After a fork(2), the child inherits a copy of the file descriptor cre‐
150 ated by timerfd_create(). The file descriptor refers to the same
151 underlying timer object as the corresponding file descriptor in the
152 parent, and read(2)s in the child will return information about expira‐
153 tions of the timer.
154
155 execve(2) semantics
156 A file descriptor created by timerfd_create() is preserved across
157 execve(2), and continues to generate timer expirations if the timer was
158 armed.
159
161 On success, timerfd_create() returns a new file descriptor. On error,
162 -1 is returned and errno is set to indicate the error.
163
164 timerfd_settime() and timerfd_gettime() return 0 on success; on error
165 they return -1, and set errno to indicate the error.
166
168 timerfd_create() can fail with the following errors:
169
170 EINVAL The clockid argument is neither CLOCK_MONOTONIC nor CLOCK_REAL‐
171 TIME;
172
173 EINVAL flags is invalid; or, in Linux 2.6.26 or earlier, flags is
174 nonzero.
175
176 EMFILE The per-process limit of open file descriptors has been reached.
177
178 ENFILE The system-wide limit on the total number of open files has been
179 reached.
180
181 ENODEV Could not mount (internal) anonymous inode device.
182
183 ENOMEM There was insufficient kernel memory to create the timer.
184
185 timerfd_settime() and timerfd_gettime() can fail with the following
186 errors:
187
188 EBADF fd is not a valid file descriptor.
189
190 EFAULT new_value, old_value, or curr_value is not valid a pointer.
191
192 EINVAL fd is not a valid timerfd file descriptor.
193
194 timerfd_settime() can also fail with the following errors:
195
196 EINVAL new_value is not properly initialized (one of the tv_nsec falls
197 outside the range zero to 999,999,999).
198
199 EINVAL flags is invalid.
200
202 These system calls are available on Linux since kernel 2.6.25. Library
203 support is provided by glibc since version 2.8.
204
206 These system calls are Linux-specific.
207
209 Currently, timerfd_create() supports fewer types of clock IDs than
210 timer_create(2).
211
213 The following program creates a timer and then monitors its progress.
214 The program accepts up to three command-line arguments. The first
215 argument specifies the number of seconds for the initial expiration of
216 the timer. The second argument specifies the interval for the timer,
217 in seconds. The third argument specifies the number of times the pro‐
218 gram should allow the timer to expire before terminating. The second
219 and third command-line arguments are optional.
220
221 The following shell session demonstrates the use of the program:
222
223 $ a.out 3 1 100
224 0.000: timer started
225 3.000: read: 1; total=1
226 4.000: read: 1; total=2
227 ^Z # type control-Z to suspend the program
228 [1]+ Stopped ./timerfd3_demo 3 1 100
229 $ fg # Resume execution after a few seconds
230 a.out 3 1 100
231 9.660: read: 5; total=7
232 10.000: read: 1; total=8
233 11.000: read: 1; total=9
234 ^C # type control-C to suspend the program
235
236 Program source
237
238 #include <sys/timerfd.h>
239 #include <time.h>
240 #include <unistd.h>
241 #include <stdlib.h>
242 #include <stdio.h>
243 #include <stdint.h> /* Definition of uint64_t */
244
245 #define handle_error(msg) \
246 do { perror(msg); exit(EXIT_FAILURE); } while (0)
247
248 static void
249 print_elapsed_time(void)
250 {
251 static struct timespec start;
252 struct timespec curr;
253 static int first_call = 1;
254 int secs, nsecs;
255
256 if (first_call) {
257 first_call = 0;
258 if (clock_gettime(CLOCK_MONOTONIC, &start) == -1)
259 handle_error("clock_gettime");
260 }
261
262 if (clock_gettime(CLOCK_MONOTONIC, &curr) == -1)
263 handle_error("clock_gettime");
264
265 secs = curr.tv_sec - start.tv_sec;
266 nsecs = curr.tv_nsec - start.tv_nsec;
267 if (nsecs < 0) {
268 secs--;
269 nsecs += 1000000000;
270 }
271 printf("%d.%03d: ", secs, (nsecs + 500000) / 1000000);
272 }
273
274 int
275 main(int argc, char *argv[])
276 {
277 struct itimerspec new_value;
278 int max_exp, fd;
279 struct timespec now;
280 uint64_t exp, tot_exp;
281 ssize_t s;
282
283 if ((argc != 2) && (argc != 4)) {
284 fprintf(stderr, "%s init-secs [interval-secs max-exp]\n",
285 argv[0]);
286 exit(EXIT_FAILURE);
287 }
288
289 if (clock_gettime(CLOCK_REALTIME, &now) == -1)
290 handle_error("clock_gettime");
291
292 /* Create a CLOCK_REALTIME absolute timer with initial
293 expiration and interval as specified in command line */
294
295 new_value.it_value.tv_sec = now.tv_sec + atoi(argv[1]);
296 new_value.it_value.tv_nsec = now.tv_nsec;
297 if (argc == 2) {
298 new_value.it_interval.tv_sec = 0;
299 max_exp = 1;
300 } else {
301 new_value.it_interval.tv_sec = atoi(argv[2]);
302 max_exp = atoi(argv[3]);
303 }
304 new_value.it_interval.tv_nsec = 0;
305
306 fd = timerfd_create(CLOCK_REALTIME, 0);
307 if (fd == -1)
308 handle_error("timerfd_create");
309
310 if (timerfd_settime(fd, TFD_TIMER_ABSTIME, &new_value, NULL) == -1)
311 handle_error("timerfd_settime");
312
313 print_elapsed_time();
314 printf("timer started\n");
315
316 for (tot_exp = 0; tot_exp < max_exp;) {
317 s = read(fd, &exp, sizeof(uint64_t));
318 if (s != sizeof(uint64_t))
319 handle_error("read");
320
321 tot_exp += exp;
322 print_elapsed_time();
323 printf("read: %llu; total=%llu\n",
324 (unsigned long long) exp,
325 (unsigned long long) tot_exp);
326 }
327
328 exit(EXIT_SUCCESS);
329 }
330
332 eventfd(2), poll(2), read(2), select(2), setitimer(2), signalfd(2),
333 timer_create(2), timer_gettime(2), timer_settime(2), epoll(7), time(7)
334
336 This page is part of release 3.53 of the Linux man-pages project. A
337 description of the project, and information about reporting bugs, can
338 be found at http://www.kernel.org/doc/man-pages/.
339
340
341
342Linux 2011-09-14 TIMERFD_CREATE(2)