1 .\" Copyright (C) 2003 Davide Libenzi
2 .\" Davide Libenzi <davidel@xmailserver.org>
4 .\" %%%LICENSE_START(GPLv2+_SW_3_PARA)
5 .\" This program is free software; you can redistribute it and/or modify
6 .\" it under the terms of the GNU General Public License as published by
7 .\" the Free Software Foundation; either version 2 of the License, or
8 .\" (at your option) any later version.
10 .\" This program is distributed in the hope that it will be useful,
11 .\" but WITHOUT ANY WARRANTY; without even the implied warranty of
12 .\" MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 .\" GNU General Public License for more details.
15 .\" You should have received a copy of the GNU General Public
16 .\" License along with this manual; if not, see
17 .\" <http://www.gnu.org/licenses/>.
20 .\" 2007-04-30: mtk, Added description of epoll_pwait()
22 .TH EPOLL_WAIT 2 2017-07-13 "Linux" "Linux Programmer's Manual"
24 epoll_wait, epoll_pwait \- wait for an I/O event on an epoll file descriptor
27 .B #include <sys/epoll.h>
29 .BI "int epoll_wait(int " epfd ", struct epoll_event *" events ,
30 .BI " int " maxevents ", int " timeout );
31 .BI "int epoll_pwait(int " epfd ", struct epoll_event *" events ,
32 .BI " int " maxevents ", int " timeout ,
33 .BI " const sigset_t *" sigmask );
38 system call waits for events on the
40 instance referred to by the file descriptor
42 The memory area pointed to by
44 will contain the events that will be available for the caller.
51 argument must be greater than zero.
55 argument specifies the number of milliseconds that
58 Time is measured against the
61 The call will block until either:
63 a file descriptor delivers an event;
65 the call is interrupted by a signal handler; or
71 interval will be rounded up to the system clock granularity,
72 and kernel scheduling delays mean that the blocking interval
73 may overrun by a small amount.
78 to block indefinitely, while specifying a
82 to return immediately, even if no events are available.
90 typedef union epoll_data {
98 uint32_t events; /* Epoll events */
99 epoll_data_t data; /* User data variable */
106 field of each returned structure contains the same data as was specified
107 in the most recent call to
109 .RB ( EPOLL_CTL_ADD ", " EPOLL_CTL_MOD )
110 for the corresponding open file description.
113 field contains the returned event bit field.
115 The relationship between
119 is analogous to the relationship between
126 allows an application to safely wait until either a file descriptor
127 becomes ready or until a signal is caught.
134 ready = epoll_pwait(epfd, &events, maxevents, timeout, &sigmask);
139 executing the following calls:
144 pthread_sigmask(SIG_SETMASK, &sigmask, &origmask);
145 ready = epoll_wait(epfd, &events, maxevents, timeout);
146 pthread_sigmask(SIG_SETMASK, &origmask, NULL);
151 argument may be specified as NULL, in which case
158 returns the number of file descriptors ready for the requested I/O, or zero
159 if no file descriptor became ready during the requested
162 When an error occurs,
166 is set appropriately.
171 is not a valid file descriptor.
174 The memory area pointed to by
176 is not accessible with write permissions.
179 The call was interrupted by a signal handler before either (1) any of the
180 requested events occurred or (2) the
191 is less than or equal to zero.
194 was added to the kernel in version 2.6.
195 .\" To be precise: kernel 2.5.44.
196 .\" The interface should be finalized by Linux kernel 2.5.66.
197 Library support is provided in glibc starting with version 2.3.2.
200 was added to Linux in kernel 2.6.19.
201 Library support is provided in glibc starting with version 2.6.
206 While one thread is blocked in a call to
208 it is possible for another thread to add a file descriptor to the waited-upon
211 If the new file descriptor becomes ready,
216 For a discussion of what may happen if a file descriptor in an
218 instance being monitored by
220 is closed in another thread, see
223 In kernels before 2.6.37, a
225 value larger than approximately
227 milliseconds is treated as \-1 (i.e., infinity).
228 Thus, for example, on a system where
233 this means that timeouts greater than 35.79 minutes are treated as infinity.
234 .SS C library/kernel differences
237 system call has a sixth argument,
238 .IR "size_t sigsetsize" ,
239 which specifies the size in bytes of the
244 wrapper function specifies this argument as a fixed value
246 .IR sizeof(sigset_t) ).
248 .BR epoll_create (2),