Linux定时器 timerfd使用

时间:2015-10-16 14:53:29   收藏:0   阅读:1600

英文使用手册原汁原味,一手资料。

NAME
       timerfd_create, timerfd_settime, timerfd_gettime - timers that notify via file descriptors

SYNOPSIS

1 #include <sys/timerfd.h>
2 
3        int timerfd_create(int clockid, int flags);
4 
5        int timerfd_settime(int fd, int flags,
6                            const struct itimerspec *new_value,
7                            struct itimerspec *old_value);
8 
9        int timerfd_gettime(int fd, struct itimerspec *curr_value);

DESCRIPTION
       These  system  calls  create  and  operate  on  a timer that delivers timer expiration notifications via a file
       descriptor.  They provide an alternative to the use of setitimer(2) or timer_create(2), with the advantage that
       the file descriptor may be monitored by select(2), poll(2), and epoll(7).

       The  use  of  these  three  system  calls  is  analogous  to  the use of timer_create(2), timer_settime(2), and
       timer_gettime(2).  (There is no analog of timer_getoverrun(2), since that functionality is provided by read(2),
       as described below.)

   timerfd_create()
       timerfd_create()  creates  a  new  timer  object, and returns a file descriptor that refers to that timer.  The
       clockid argument specifies the clock that is used to mark the  progress  of  the  timer,  and  must  be  either
       CLOCK_REALTIME  or CLOCK_MONOTONIC.  CLOCK_REALTIME is a settable system-wide clock.  CLOCK_MONOTONIC is a non‐
       settable clock that is not affected by discontinuous changes in the system clock (e.g., manual changes to  sys‐
       tem time).  The current value of each of these clocks can be retrieved using clock_gettime(2).

       Starting  with  Linux  2.6.27,  the  following  values  may  be bitwise ORed in flags to change the behavior of
       timerfd_create():

       TFD_NONBLOCK  Set the O_NONBLOCK file status flag on the new open file  description.   Using  this  flag  saves
                     extra calls to fcntl(2) to achieve the same result.

       TFD_CLOEXEC   Set  the  close-on-exec (FD_CLOEXEC) flag on the new file descriptor.  See the description of the
                     O_CLOEXEC flag in open(2) for reasons why this may be useful.

       In Linux versions up to and including 2.6.26, flags must be specified as zero.

   timerfd_settime()
       timerfd_settime() arms (starts) or disarms (stops) the timer referred to by the file descriptor fd.

       The new_value argument specifies the initial expiration and interval for the timer.  The itimer structure  used
       for this argument contains two fields, each of which is in turn a structure of type timespec:

1            struct timespec {
2                time_t tv_sec;                /* Seconds */
3                long   tv_nsec;               /* Nanoseconds */
4            };
5 
6            struct itimerspec {
7                struct timespec it_interval;  /* Interval for periodic timer */
8                struct timespec it_value;     /* Initial expiration */
9            };

 new_value.it_value  specifies  the initial expiration of the timer, in seconds and nanoseconds.  Setting either
       field of new_value.it_value to a nonzero value arms the timer.  Setting both fields  of  new_value.it_value  to
       zero disarms the timer.

       Setting  one  or  both  fields  of new_value.it_interval to nonzero values specifies the period, in seconds and
       nanoseconds,  for  repeated  timer  expirations  after   the   initial   expiration.    If   both   fields   of
       new_value.it_interval are zero, the timer expires just once, at the time specified by new_value.it_value.

       The  flags argument is either 0, to start a relative timer (new_value.it_value specifies a time relative to the
       current value  of  the  clock  specified  by  clockid),  or  TFD_TIMER_ABSTIME,  to  start  an  absolute  timer
       (new_value.it_value  specifies  an  absolute  time  for the clock specified by clockid; that is, the timer will
       expire when the value of that clock reaches the value specified in new_value.it_value).

       If the old_value argument is not NULL, then the itimerspec structure that it points to is used  to  return  the
       setting of the timer that was current at the time of the call; see the description of timerfd_gettime() follow‐
       ing.

   timerfd_gettime()
       timerfd_gettime() returns, in curr_value, an itimerspec structure that contains  the  current  setting  of  the
       timer referred to by the file descriptor fd.

       The  it_value field returns the amount of time until the timer will next expire.  If both fields of this struc‐
       ture are zero, then the timer is currently disarmed.  This field always contains a relative  value,  regardless
       of whether the TFD_TIMER_ABSTIME flag was specified when setting the timer.

       The  it_interval  field returns the interval of the timer.  If both fields of this structure are zero, then the
       timer is set to expire just once, at the time specified by curr_value.it_value.

   Operating on a timer file descriptor
       The file descriptor returned by timerfd_create() supports the following operations:

       read(2)
              If the timer has already expired one  or  more  times  since  its  settings  were  last  modified  using
              timerfd_settime(),  or  since  the  last successful read(2), then the buffer given to read(2) returns an
              unsigned 8-byte integer (uint64_t) containing the  number  of  expirations  that  have  occurred.   (The
              returned value is in host byte order—that is, the native byte order for integers on the host machine.)

              If  no timer expirations have occurred at the time of the read(2), then the call either blocks until the
              next timer expiration, or fails with the error EAGAIN if the file descriptor has been  made  nonblocking
              (via the use of the fcntl(2) F_SETFL operation to set the O_NONBLOCK flag).

              A read(2) will fail with the error EINVAL if the size of the supplied buffer is less than 8 bytes.

       poll(2), select(2) (and similar)
              The file descriptor is readable (the select(2) readfds argument; the poll(2) POLLIN flag) if one or more
              timer expirations have occurred.

              The file descriptor also supports the other file-descriptor multiplexing APIs: pselect(2), ppoll(2), and
              epoll(7).

       close(2)
              When  the  file descriptor is no longer required it should be closed.  When all file descriptors associ‐
              ated with the same timer object have been closed, the timer is disarmed and its resources are  freed  by
              the kernel.

   fork(2) semantics
       After  a  fork(2),  the  child  inherits  a  copy of the file descriptor created by timerfd_create().  The file
       descriptor refers to the same underlying timer object as the corresponding file descriptor in the  parent,  and
       read(2)s in the child will return information about expirations of the timer.

   execve(2) semantics
       A  file  descriptor  created by timerfd_create() is preserved across execve(2), and continues to generate timer
       expirations if the timer was armed.

RETURN VALUE
       On success, timerfd_create() returns a new file descriptor.  On error, -1 is returned and errno is set to indi‐
       cate the error.

       timerfd_settime() and timerfd_gettime() return 0 on success; on error they return -1, and set errno to indicate
       the error.

ERRORS
       timerfd_create() can fail with the following errors:

       EINVAL The clockid argument is neither CLOCK_MONOTONIC nor CLOCK_REALTIME;

       EINVAL flags is invalid; or, in Linux 2.6.26 or earlier, flags is nonzero.

       EMFILE The per-process limit of open file descriptors has been reached.

       ENFILE The system-wide limit on the total number of open files has been reached.

       ENODEV Could not mount (internal) anonymous inode device.

       ENOMEM There was insufficient kernel memory to create the timer.

       timerfd_settime() and timerfd_gettime() can fail with the following errors:

       EBADF  fd is not a valid file descriptor.

       EFAULT new_value, old_value, or curr_value is not valid a pointer.

       EINVAL fd is not a valid timerfd file descriptor.

       timerfd_settime() can also fail with the following errors:

       EINVAL new_value is not properly initialized (one of the tv_nsec falls outside the range zero to 999,999,999).

       EINVAL flags is invalid.

VERSIONS
       These system calls are available on Linux since kernel 2.6.25.  Library support is provided by glibc since ver‐
       sion 2.8.

CONFORMING TO
       These system calls are Linux-specific.

BUGS
       Currently, timerfd_create() supports fewer types of clock IDs than timer_create(2).

EXAMPLE
       The following program creates a timer and then monitors its progress.  The program accepts up to three command-
       line arguments.  The first argument specifies the number of seconds for the initial expiration  of  the  timer.
       The  second argument specifies the interval for the timer, in seconds.  The third argument specifies the number
       of times the program should allow the timer to expire before terminating.  The second  and  third  command-line
       arguments are optional.

       The following shell session demonstrates the use of the program:

$ a.out 3 1 100
           0.000: timer started
           3.000: read: 1; total=1
           4.000: read: 1; total=2
           ^Z                  # type control-Z to suspend the program
           [1]+  Stopped                 ./timerfd3_demo 3 1 100
           $ fg                # Resume execution after a few seconds
           a.out 3 1 100
           9.660: read: 5; total=7
           10.000: read: 1; total=8
           11.000: read: 1; total=9
           ^C                  # type control-C to suspend the program

Program source

 1 #include <sys/timerfd.h>
 2        #include <time.h>
 3        #include <unistd.h>
 4        #include <stdlib.h>
 5        #include <stdio.h>
 6        #include <stdint.h>        /* Definition of uint64_t */
 7 
 8        #define handle_error(msg)  9                do { perror(msg); exit(EXIT_FAILURE); } while (0)
10 
11        static void
12        print_elapsed_time(void)
13        {
14            static struct timespec start;
15            struct timespec curr;
16            static int first_call = 1;
17            int secs, nsecs;
18 
19            if (first_call) {
20                first_call = 0;
21                if (clock_gettime(CLOCK_MONOTONIC, &start) == -1)
22                    handle_error("clock_gettime");
23            }
24 
25            if (clock_gettime(CLOCK_MONOTONIC, &curr) == -1)
26                handle_error("clock_gettime");
27 
28            secs = curr.tv_sec - start.tv_sec;
29            nsecs = curr.tv_nsec - start.tv_nsec;
30            if (nsecs < 0) {
31                secs--;
32                nsecs += 1000000000;
33            }
34            printf("%d.%03d: ", secs, (nsecs + 500000) / 1000000);
35        }
36 
37        int
38        main(int argc, char *argv[])
39        {
40            struct itimerspec new_value;
41            int max_exp, fd;
42            struct timespec now;
43            uint64_t exp, tot_exp;
44            ssize_t s;
45 
46            if ((argc != 2) && (argc != 4)) {
47                fprintf(stderr, "%s init-secs [interval-secs max-exp]\n",
48                        argv[0]);
49                exit(EXIT_FAILURE);
50            }
51 
52            if (clock_gettime(CLOCK_REALTIME, &now) == -1)
53                handle_error("clock_gettime");
54 
55            /* Create a CLOCK_REALTIME absolute timer with initial
56               expiration and interval as specified in command line */
57 
58            new_value.it_value.tv_sec = now.tv_sec + atoi(argv[1]);
59            new_value.it_value.tv_nsec = now.tv_nsec;
60            if (argc == 2) {
61                new_value.it_interval.tv_sec = 0;
62                max_exp = 1;
63            } else {
64                new_value.it_interval.tv_sec = atoi(argv[2]);
65                max_exp = atoi(argv[3]);
66            }
67            new_value.it_interval.tv_nsec = 0;
68 
69            fd = timerfd_create(CLOCK_REALTIME, 0);
70            if (fd == -1)
71                handle_error("timerfd_create");
72 
73            if (timerfd_settime(fd, TFD_TIMER_ABSTIME, &new_value, NULL) == -1)
74                handle_error("timerfd_settime");
75 
76            print_elapsed_time();
77            printf("timer started\n");
78 
79            for (tot_exp = 0; tot_exp < max_exp;) {
80                s = read(fd, &exp, sizeof(uint64_t));
81                if (s != sizeof(uint64_t))
82                    handle_error("read");
83 
84                tot_exp += exp;
85                print_elapsed_time();
86                printf("read: %llu; total=%llu\n",
87                        (unsigned long long) exp,
88                        (unsigned long long) tot_exp);
89            }
90 
91            exit(EXIT_SUCCESS);
92        }

SEE ALSO
       eventfd(2),   poll(2),   read(2),  select(2),  setitimer(2),  signalfd(2),  timer_create(2),  timer_gettime(2),
       timer_settime(2), epoll(7), time(7)

COLOPHON
       This page is part of release 3.74 of the Linux man-pages project.  A description of  the  project,  information
       about    reporting    bugs,    and    the    latest    version    of    this    page,    can    be   found   at
       http://www.kernel.org/doc/man-pages/.

评论(0
© 2014 mamicode.com 版权所有 京ICP备13008772号-2  联系我们:gaon5@hotmail.com
迷上了代码!