man2/poll.2

   1 .\" Copyright (C) 1997 Andries Brouwer (aeb@cwi.nl)
   2 .\" and Copyright (C) 2006, Michael Kerrisk <mtk.manpages@gmail.com>
   3 .\"
   4 .\" %%%LICENSE_START(VERBATIM)
   5 .\" Permission is granted to make and distribute verbatim copies of this
   6 .\" manual provided the copyright notice and this permission notice are
   7 .\" preserved on all copies.
   8 .\"
   9 .\" Permission is granted to copy and distribute modified versions of this
  10 .\" manual under the conditions for verbatim copying, provided that the
  11 .\" entire resulting derived work is distributed under the terms of a
  12 .\" permission notice identical to this one.
  13 .\"
  14 .\" Since the Linux kernel and libraries are constantly changing, this
  15 .\" manual page may be incorrect or out-of-date.  The author(s) assume no
  16 .\" responsibility for errors or omissions, or for damages resulting from
  17 .\" the use of the information contained herein.  The author(s) may not
  18 .\" have taken the same level of care in the production of this manual,
  19 .\" which is licensed free of charge, as they might when working
  20 .\" professionally.
  21 .\"
  22 .\" Formatted or processed versions of this manual, if unaccompanied by
  23 .\" the source, must acknowledge the copyright and authors of this work.
  24 .\" %%%LICENSE_END
  25 .\"
  26 .\" Additions from Richard Gooch <rgooch@atnf.CSIRO.AU> and aeb, 971207
  27 .\" 2006-03-13, mtk, Added ppoll() + various other rewordings
  28 .\" 2006-07-01, mtk, Added POLLRDHUP + various other wording and
  29 .\"     formatting changes.
  30 .\"
  31 .TH POLL 2 2015-12-28 "Linux" "Linux Programmer's Manual"
  32 .SH NAME
  33 poll, ppoll \- wait for some event on a file descriptor
  34 .SH SYNOPSIS
  35 .nf
  36 .B #include <poll.h>
  37 .sp
  38 .BI "int poll(struct pollfd *" fds ", nfds_t " nfds ", int " timeout );
  39 .sp
  40 .BR "#define _GNU_SOURCE" "         /* See feature_test_macros(7) */"
  41 .B #include <signal.h>
  42 .B #include <poll.h>
  43 .sp
  44 .BI "int ppoll(struct pollfd *" fds ", nfds_t " nfds ", "
  45 .BI "        const struct timespec *" tmo_p ", const sigset_t *" sigmask );
  46 .fi
  47 .SH DESCRIPTION
  48 .BR poll ()
  49 performs a similar task to
  50 .BR select (2):
  51 it waits for one of a set of file descriptors to become ready
  52 to perform I/O.
  53
  54 The set of file descriptors to be monitored is specified in the
  55 .I fds
  56 argument, which is an array of structures of the following form:
  57 .in +4n
  58 .nf
  59
  60 struct pollfd {
  61     int   fd;         /* file descriptor */
  62     short events;     /* requested events */
  63     short revents;    /* returned events */
  64 };
  65 .in
  66 .fi
  67 .PP
  68 The caller should specify the number of items in the
  69 .I fds
  70 array in
  71 .IR nfds .
  72
  73 The field
  74 .I fd
  75 contains a file descriptor for an open file.
  76 If this field is negative, then the corresponding
  77 .I events
  78 field is ignored and the
  79 .I revents
  80 field returns zero.
  81 (This provides an easy way of ignoring a
  82 file descriptor for a single
  83 .BR poll ()
  84 call: simply negate the
  85 .I fd
  86 field.
  87 Note, however, that this technique can't be used to ignore file descriptor 0.)
  88
  89 The field
  90 .I events
  91 is an input parameter, a bit mask specifying the events the application
  92 is interested in for the file descriptor
  93 .IR fd .
  94 This field may be specified as zero,
  95 in which case the only events that can be returned in
  96 .I revents
  97 are
  98 .BR POLLHUP ,
  99 .BR POLLERR ,
 100 and
 101 .B POLLNVAL
 102 (see below).
 103
 104 The field
 105 .I revents
 106 is an output parameter, filled by the kernel with the events that
 107 actually occurred.
 108 The bits returned in
 109 .I revents
 110 can include any of those specified in
 111 .IR events ,
 112 or one of the values
 113 .BR POLLERR ,
 114 .BR POLLHUP ,
 115 or
 116 .BR POLLNVAL .
 117 (These three bits are meaningless in the
 118 .I events
 119 field, and will be set in the
 120 .I revents
 121 field whenever the corresponding condition is true.)
 122
 123 If none of the events requested (and no error) has occurred for any
 124 of the file descriptors, then
 125 .BR poll ()
 126 blocks until one of the events occurs.
 127
 128 The
 129 .I timeout
 130 argument specifies the number of milliseconds that
 131 .BR poll ()
 132 should block waiting for a file descriptor to become ready.
 133 The call will block until either:
 134 .IP * 3
 135 a file descriptor becomes ready;
 136 .IP *
 137 the call is interrupted by a signal handler; or
 138 .IP *
 139 the timeout expires.
 140 .PP
 141 Note that the
 142 .I timeout
 143 interval will be rounded up to the system clock granularity,
 144 and kernel scheduling delays mean that the blocking interval
 145 may overrun by a small amount.
 146 Specifying a negative value in
 147 .I timeout
 148 means an infinite timeout.
 149 Specifying a
 150 .I timeout
 151 of zero causes
 152 .BR poll ()
 153 to return immediately, even if no file descriptors are ready.
 154
 155 The bits that may be set/returned in
 156 .I events
 157 and
 158 .I revents
 159 are defined in \fI<poll.h>\fP:
 160 .RS
 161 .TP
 162 .B POLLIN
 163 There is data to read.
 164 .TP
 165 .B POLLPRI
 166 There is urgent data to read (e.g., out-of-band data on TCP socket;
 167 pseudoterminal master in packet mode has seen state change in slave).
 168 .TP
 169 .B POLLOUT
 170 Writing is now possible, though a write larger that the available space
 171 in a socket or pipe will still block (unless
 172 .B O_NONBLOCK
 173 is set).
 174 .TP
 175 .BR POLLRDHUP " (since Linux 2.6.17)"
 176 Stream socket peer closed connection,
 177 or shut down writing half of connection.
 178 The
 179 .B _GNU_SOURCE
 180 feature test macro must be defined
 181 (before including
 182 .I any
 183 header files)
 184 in order to obtain this definition.
 185 .TP
 186 .B POLLERR
 187 Error condition (only returned in
 188 .IR revents ;
 189 ignored in
 190 .IR events ).
 191 .TP
 192 .B POLLHUP
 193 Hang up (only returned in
 194 .IR revents ;
 195 ignored in
 196 .IR events ).
 197 Note that when reading from a channel such as a pipe or a stream socket,
 198 this event merely indicates that the peer closed its end of the channel.
 199 Subsequent reads from the channel will return 0 (end of file)
 200 only after all outstanding data in the channel has been consumed.
 201 .TP
 202 .B POLLNVAL
 203 Invalid request:
 204 .I fd
 205 not open (only returned in
 206 .IR revents ;
 207 ignored in
 208 .IR events ).
 209 .RE
 210 .PP
 211 When compiling with
 212 .B _XOPEN_SOURCE
 213 defined, one also has the following,
 214 which convey no further information beyond the bits listed above:
 215 .RS
 216 .TP
 217 .B POLLRDNORM
 218 Equivalent to
 219 .BR POLLIN .
 220 .TP
 221 .B POLLRDBAND
 222 Priority band data can be read (generally unused on Linux).
 223 .\" POLLRDBAND is used in the DECnet protocol.
 224 .TP
 225 .B POLLWRNORM
 226 Equivalent to
 227 .BR POLLOUT .
 228 .TP
 229 .B POLLWRBAND
 230 Priority data may be written.
 231 .RE
 232 .PP
 233 Linux also knows about, but does not use
 234 .BR POLLMSG .
 235 .SS ppoll()
 236 The relationship between
 237 .BR poll ()
 238 and
 239 .BR ppoll ()
 240 is analogous to the relationship between
 241 .BR select (2)
 242 and
 243 .BR pselect (2):
 244 like
 245 .BR pselect (2),
 246 .BR ppoll ()
 247 allows an application to safely wait until either a file descriptor
 248 becomes ready or until a signal is caught.
 249 .PP
 250 Other than the difference in the precision of the
 251 .I timeout
 252 argument, the following
 253 .BR ppoll ()
 254 call:
 255 .nf
 256
 257     ready = ppoll(&fds, nfds, tmo_p, &sigmask);
 258
 259 .fi
 260 is equivalent to
 261 .I atomically
 262 executing the following calls:
 263 .nf
 264
 265     sigset_t origmask;
 266     int timeout;
 267
 268     timeout = (tmo_p == NULL) ? \-1 :
 269               (tmo_p\->tv_sec * 1000 + tmo_p\->tv_nsec / 1000000);
 270     pthread_sigmask(SIG_SETMASK, &sigmask, &origmask);
 271     ready = poll(&fds, nfds, timeout);
 272     pthread_sigmask(SIG_SETMASK, &origmask, NULL);
 273 .fi
 274 .PP
 275 See the description of
 276 .BR pselect (2)
 277 for an explanation of why
 278 .BR ppoll ()
 279 is necessary.
 280
 281 If the
 282 .I sigmask
 283 argument is specified as NULL, then
 284 no signal mask manipulation is performed
 285 (and thus
 286 .BR ppoll ()
 287 differs from
 288 .BR poll ()
 289 only in the precision of the
 290 .I timeout
 291 argument).
 292
 293 The
 294 .I tmo_p
 295 argument specifies an upper limit on the amount of time that
 296 .BR ppoll ()
 297 will block.
 298 This argument is a pointer to a structure of the following form:
 299 .in +4n
 300 .nf
 301
 302 struct timespec {
 303     long    tv_sec;         /* seconds */
 304     long    tv_nsec;        /* nanoseconds */
 305 };
 306 .fi
 307 .in
 308
 309 If
 310 .I tmo_p
 311 is specified as NULL, then
 312 .BR ppoll ()
 313 can block indefinitely.
 314 .SH RETURN VALUE
 315 On success, a positive number is returned; this is
 316 the number of structures which have nonzero
 317 .I revents
 318 fields (in other words, those descriptors with events or errors reported).
 319 A value of 0 indicates that the call timed out and no file
 320 descriptors were ready.
 321 On error, \-1 is returned, and
 322 .I errno
 323 is set appropriately.
 324 .SH ERRORS
 325 .TP
 326 .B EFAULT
 327 The array given as argument was not contained in the calling program's
 328 address space.
 329 .TP
 330 .B EINTR
 331 A signal occurred before any requested event; see
 332 .BR signal (7).
 333 .TP
 334 .B EINVAL
 335 The
 336 .I nfds
 337 value exceeds the
 338 .B RLIMIT_NOFILE
 339 value.
 340 .TP
 341 .B ENOMEM
 342 There was no space to allocate file descriptor tables.
 343 .SH VERSIONS
 344 The
 345 .BR poll ()
 346 system call was introduced in Linux 2.1.23.
 347 On older kernels that lack this system call,
 348 .\" library call was introduced in libc 5.4.28
 349 the glibc (and the old Linux libc)
 350 .BR poll ()
 351 wrapper function provides emulation using
 352 .BR select (2).
 353
 354 The
 355 .BR ppoll ()
 356 system call was added to Linux in kernel 2.6.16.
 357 The
 358 .BR ppoll ()
 359 library call was added in glibc 2.4.
 360 .SH CONFORMING TO
 361 .BR poll ()
 362 conforms to POSIX.1-2001 and POSIX.1-2008.
 363 .BR ppoll ()
 364 is Linux-specific.
 365 .\" NetBSD 3.0 has a pollts() which is like Linux ppoll().
 366 .SH NOTES
 367 On some other UNIX systems,
 368 .\" Darwin, according to a report by Jeremy Sequoia, relayed by Josh Triplett
 369 .BR poll ()
 370 can fail with the error
 371 .B EAGAIN
 372 if the system fails to allocate kernel-internal resources, rather than
 373 .B ENOMEM
 374 as Linux does.
 375 POSIX permits this behavior.
 376 Portable programs may wish to check for
 377 .B EAGAIN
 378 and loop, just as with
 379 .BR EINTR .
 380
 381 Some implementations define the nonstandard constant
 382 .B INFTIM
 383 with the value \-1 for use as a
 384 .IR timeout
 385 for
 386 .BR poll ().
 387 This constant is not provided in glibc.
 388
 389 For a discussion of what may happen if a file descriptor being monitored by
 390 .BR poll ()
 391 is closed in another thread, see
 392 .BR select (2).
 393 .SS C library/kernel differences
 394 The Linux
 395 .BR ppoll ()
 396 system call modifies its
 397 .I tmo_p
 398 argument.
 399 However, the glibc wrapper function hides this behavior
 400 by using a local variable for the timeout argument that
 401 is passed to the system call.
 402 Thus, the glibc
 403 .BR ppoll ()
 404 function does not modify its
 405 .I tmo_p
 406 argument.
 407
 408 The raw
 409 .BR ppoll ()
 410 system call has a fifth argument,
 411 .IR "size_t sigsetsize" ,
 412 which specifies the size in bytes of the
 413 .IR sigmask
 414 argument.
 415 The glibc
 416 .BR ppoll ()
 417 wrapper function specifies this argument as a fixed value
 418 (equal to
 419 .IR sizeof(sigset_t) ).
 420 .SH BUGS
 421 See the discussion of spurious readiness notifications under the
 422 BUGS section of
 423 .BR select (2).
 424 .SH SEE ALSO
 425 .BR restart_syscall (2),
 426 .BR select (2),
 427 .BR select_tut (2),
 428 .BR epoll (7),
 429 .BR time (7)