man2/getrlimit.2

   1 .\" Hey Emacs! This file is -*- nroff -*- source.
   2 .\"
   3 .\" Copyright (c) 1992 Drew Eckhardt, March 28, 1992
   4 .\" and Copyright (c) 2002 Michael Kerrisk
   5 .\"
   6 .\" Permission is granted to make and distribute verbatim copies of this
   7 .\" manual provided the copyright notice and this permission notice are
   8 .\" preserved on all copies.
   9 .\"
  10 .\" Permission is granted to copy and distribute modified versions of this
  11 .\" manual under the conditions for verbatim copying, provided that the
  12 .\" entire resulting derived work is distributed under the terms of a
  13 .\" permission notice identical to this one.
  14 .\"
  15 .\" Since the Linux kernel and libraries are constantly changing, this
  16 .\" manual page may be incorrect or out-of-date.  The author(s) assume no
  17 .\" responsibility for errors or omissions, or for damages resulting from
  18 .\" the use of the information contained herein.  The author(s) may not
  19 .\" have taken the same level of care in the production of this manual,
  20 .\" which is licensed free of charge, as they might when working
  21 .\" professionally.
  22 .\"
  23 .\" Formatted or processed versions of this manual, if unaccompanied by
  24 .\" the source, must acknowledge the copyright and authors of this work.
  25 .\"
  26 .\" Modified by Michael Haardt <michael@moria.de>
  27 .\" Modified 1993-07-23 by Rik Faith <faith@cs.unc.edu>
  28 .\" Modified 1996-01-13 by Arnt Gulbrandsen <agulbra@troll.no>
  29 .\" Modified 1996-01-22 by aeb, following a remark by
  30 .\"          Tigran Aivazian <tigran@sco.com>
  31 .\" Modified 1996-04-14 by aeb, following a remark by
  32 .\"          Robert Bihlmeyer <robbe@orcus.ping.at>
  33 .\" Modified 1996-10-22 by Eric S. Raymond <esr@thyrsus.com>
  34 .\" Modified 2001-05-04 by aeb, following a remark by
  35 .\"          Håvard Lygre <hklygre@online.no>
  36 .\" Modified 2001-04-17 by Michael Kerrisk <mtk-manpages@gmx.net>
  37 .\" Modified 2002-06-13 by Michael Kerrisk <mtk-manpages@gmx.net>
  38 .\"     Added note on non-standard behaviour when SIGCHLD is ignored.
  39 .\" Modified 2002-07-09 by Michael Kerrisk <mtk-manpages@gmx.net>
  40 .\"     Enhanced descriptions of 'resource' values for [gs]etrlimit()
  41 .\" Modified 2003-11-28 by aeb, added RLIMIT_CORE
  42 .\" Modified 2004-03-26 by aeb, added RLIMIT_AS
  43 .\" Modified 2004-06-16 by Michael Kerrisk <mtk-manpages@gmx.net>
  44 .\"     Added notes on CAP_SYS_RESOURCE
  45 .\"
  46 .\" 2004-11-16 -- mtk: the getrlimit.2 page, which formally included
  47 .\" coverage of getrusage(2), has been split, so that the latter
  48 .\" is now covered in its own getrusage.2.
  49 .\"
  50 .\" Modified 2004-11-16, mtk: A few other minor changes
  51 .\" Modified 2004-11-23, mtk
  52 .\"     Added notes on RLIMIT_MEMLOCK, RLIMIT_NPROC, and RLIMIT_RSS
  53 .\"             to "CONFORMING TO"
  54 .\" Modified 2004-11-25, mtk
  55 .\"     Rewrote discussion on RLIMIT_MEMLOCK to incorporate kernel
  56 .\"             2.6.9 changes.
  57 .\"     Added note on RLIMIT_CPU error in older kernels
  58 .\" 2004-11-03, mtk, Added RLIMIT_SIGPENDING
  59 .\" 2005-07-13, mtk, documented RLIMIT_MSGQUEUE limit.
  60 .\" 2005-07-28, mtk, Added descriptions of RLIMIT_NICE and RLIMIT_RTPRIO
  61 .\"
  62 .TH GETRLIMIT 2 2005-09-20 "Linux" "Linux Programmer's Manual"
  63 .SH NAME
  64 getrlimit, setrlimit \- get/set resource limits
  65 .SH SYNOPSIS
  66 .B #include <sys/time.h>
  67 .br
  68 .B #include <sys/resource.h>
  69 .sp
  70 .BI "int getrlimit(int " resource ", struct rlimit *" rlim );
  71 .br
  72 .BI "int setrlimit(int " resource ", const struct rlimit *" rlim );
  73 .SH DESCRIPTION
  74 .BR getrlimit ()
  75 and
  76 .BR setrlimit ()
  77 get and set resource limits respectively.
  78 Each resource has an associated soft and hard limit, as defined by the
  79 .I rlimit
  80 structure (the
  81 .I rlim
  82 argument to both
  83 .BR getrlimit ()
  84 and
  85 .BR setrlimit ()):
  86 .PP
  87 .in +0.5i
  88 .nf
  89 struct rlimit {
  90     rlim_t rlim_cur;  /* Soft limit */
  91     rlim_t rlim_max;  /* Hard limit (ceiling for rlim_cur) */
  92 };
  93
  94 .fi
  95 .in -0.5i
  96 The soft limit is the value that the kernel enforces for the
  97 corresponding resource.
  98 The hard limit acts as a ceiling for the soft limit:
  99 an unprivileged process may only set its soft limit to a value in the
 100 range from 0 up to the hard limit, and (irreversibly) lower its hard limit.
 101 A privileged process (under Linux: one with the
 102 .B CAP_SYS_RESOURCE
 103 capability) may make arbitrary changes to either limit value.
 104 .PP
 105 The value
 106 .B RLIM_INFINITY
 107 denotes no limit on a resource (both in the structure returned by
 108 .BR getrlimit ()
 109 and in the structure passed to
 110 .BR setrlimit ()).
 111 .PP
 112 .I resource
 113 must be one of:
 114 .TP
 115 .B RLIMIT_AS
 116 The maximum size of the process's virtual memory (address space) in bytes.
 117 .\" since 2.0.27 / 2.1.12
 118 This limit affects calls to
 119 .BR brk (2),
 120 .BR mmap (2)
 121 and
 122 .BR mremap (2),
 123 which fail with the error
 124 .B ENOMEM
 125 upon exceeding this limit.
 126 Also automatic stack expansion will fail
 127 (and generate a
 128 .B SIGSEGV
 129 that kills the process if no alternate stack
 130 has been made available via
 131 .BR sigaltstack (2)).
 132 Since the value is a \fIlong\fP, on machines with a 32-bit \fIlong\fP
 133 either this limit is at most 2 GiB, or this resource is unlimited.
 134 .TP
 135 .B RLIMIT_CORE
 136 Maximum size of
 137 .I core
 138 file.
 139 When 0 no core dump files are created.
 140 When non-zero, larger dumps are truncated to this size.
 141 .TP
 142 .B RLIMIT_CPU
 143 CPU time limit in seconds.
 144 When the process reaches the soft limit, it is sent a
 145 .B SIGXCPU
 146 signal.
 147 The default action for this signal is to terminate the process.
 148 However, the signal can be caught, and the handler can return control to
 149 the main program.
 150 If the process continues to consume CPU time, it will be sent
 151 .B SIGXCPU
 152 once per second until the hard limit is reached, at which time
 153 it is sent
 154 .BR SIGKILL .
 155 (This latter point describes Linux 2.2 through 2.6 behaviour.
 156 Implementations vary in how they treat processes which continue to
 157 consume CPU time after reaching the soft limit.
 158 Portable applications that need to catch this signal should
 159 perform an orderly termination upon first receipt of
 160 .BR SIGXCPU .)
 161 .TP
 162 .B RLIMIT_DATA
 163 The maximum size of the process's data segment (initialized data,
 164 uninitialized data, and heap).
 165 This limit affects calls to
 166 .BR brk (2)
 167 and
 168 .BR sbrk (2),
 169 which fail with the error
 170 .B ENOMEM
 171 upon encountering the soft limit of this resource.
 172 .TP
 173 .B RLIMIT_FSIZE
 174 The maximum size of files that the process may create.
 175 Attempts to extend a file beyond this limit result in delivery of a
 176 .B SIGXFSZ
 177 signal.
 178 By default, this signal terminates a process, but a process can
 179 catch this signal instead, in which case the relevant system call (e.g.,
 180 .BR write (2)
 181 .BR truncate (2))
 182 fails with the error
 183 .BR EFBIG .
 184 .TP
 185 .BR RLIMIT_LOCKS " (Early Linux 2.4 only)"
 186 .\" to be precise: Linux 2.4.0-test9; no longer in 2.4.25 / 2.5.65
 187 A limit on the combined number of
 188 .BR flock (2)
 189 locks and
 190 .BR fcntl (2)
 191 leases that this process may establish.
 192 .TP
 193 .B RLIMIT_MEMLOCK
 194 The maximum number of bytes of memory that may be locked
 195 into RAM.
 196 In effect this limit is rounded down to the nearest multiple
 197 of the system page size.
 198 This limit affects
 199 .BR mlock (2)
 200 and
 201 .BR mlockall (2)
 202 and the
 203 .BR mmap (2)
 204 .B MAP_LOCKED
 205 operation.
 206 Since Linux 2.6.9 it also affects the
 207 .BR shmctl (2)
 208 .B SHM_LOCK
 209 operation, where it sets a maximum on the total bytes in
 210 shared memory segments (see
 211 .BR shmget (2))
 212 that may be locked by the real user ID of the calling process.
 213 The
 214 .BR shmctl (2)
 215 .B SHM_LOCK
 216 locks are accounted for separately from the per-process memory
 217 locks established by
 218 .BR mlock (2),
 219 .BR mlockall (2),
 220 and
 221 .BR mmap (2)
 222 .BR MAP_LOCKED ;
 223 a process can lock bytes up to this limit in each of these
 224 two categories.
 225 In Linux kernels before 2.6.9, this limit controlled the amount of
 226 memory that could be locked by a privileged process.
 227 Since Linux 2.6.9, no limits are placed on the amount of memory
 228 that a privileged process may lock, and this limit instead governs
 229 the amount of memory that an unprivileged process may lock.
 230 .TP
 231 .BR RLIMIT_MSGQUEUE " (Since Linux 2.6.8)"
 232 Specifies the limit on the number of bytes that can be allocated
 233 for POSIX message queues for the real user ID of the calling process.
 234 This limit is enforced for
 235 .BR mq_open (3).
 236 Each message queue that the user creates counts (until it is removed)
 237 against this limit according to the formula:
 238 .nf
 239
 240     bytes = attr.mq_maxmsg * sizeof(struct msg_msg *) +
 241             attr.mq_maxmsg * attr.mq_msgsize
 242
 243 .fi
 244 where
 245 .I attr
 246 is the
 247 .I mq_attr
 248 structure specified as the fourth argument to
 249 .BR mq_open (3).
 250
 251 The first addend in the formula, which includes
 252 .I "sizeof(struct msg_msg *)"
 253 (4 bytes on Linux/x86), ensures that the user cannot
 254 create an unlimited number of zero-length messages (such messages
 255 nevertheless each consume some system memory for bookkeeping overhead).
 256 .TP
 257 .BR RLIMIT_NICE " (since kernel 2.6.12, but see BUGS below)"
 258 Specifies a ceiling to which the process's nice value can be raised using
 259 .BR setpriority (2)
 260 or
 261 .BR nice (2).
 262 The actual ceiling for the nice value is calculated as
 263 .IR "20\ \-\ rlim_cur" .
 264 (This strangeness occurs because negative numbers cannot be specified
 265 as resource limit values, since they typically have special meanings.
 266 For example, RLIM_INFINITY typically is the same as \-1.)
 267 .TP
 268 .B RLIMIT_NOFILE
 269 Specifies a value one greater than the maximum file descriptor number
 270 that can be opened by this process.
 271 Attempts
 272 .RB ( open (2),
 273 .BR pipe (2),
 274 .BR dup (2),
 275 etc.)
 276 to exceed this limit yield the error
 277 .BR EMFILE .
 278 .TP
 279 .B RLIMIT_NPROC
 280 The maximum number of processes (or, more precisely on Linux, threads)
 281 that can be created for the real user ID of the calling process.
 282 Upon encountering this limit,
 283 .BR fork (2)
 284 fails with the error
 285 .BR EAGAIN .
 286 .TP
 287 .B RLIMIT_RSS
 288 Specifies the limit (in pages) of the process's resident set
 289 (the number of virtual pages resident in RAM).
 290 This limit only has effect in Linux 2.4.x, x < 30, and there only
 291 affects calls to
 292 .BR madvise (2)
 293 specifying
 294 .BR MADV_WILLNEED .
 295 .\" As at kernel 2.6.12, this limit still does nothing in 2.6 though
 296 .\" talk of making it do something has surfaced from time to time in LKML
 297 .\"       -- MTK, Jul 05
 298 .TP
 299 .BR RLIMIT_RTPRIO " (Since Linux 2.6.12, but see BUGS)"
 300 Specifies a ceiling on the real-time priority that may be set for
 301 this process using
 302 .BR sched_setscheduler (2)
 303 and
 304 .BR sched_setparam (2).
 305 .TP
 306 .BR RLIMIT_SIGPENDING " (Since Linux 2.6.8)"
 307 Specifies the limit on the number of signals
 308 that may be queued for the real user ID of the calling process.
 309 Both standard and real-time signals are counted for the purpose of
 310 checking this limit.
 311 However, the limit is only enforced for
 312 .BR sigqueue (2);
 313 it is always possible to use
 314 .BR kill (2)
 315 to queue one instance of any of the signals that are not already
 316 queued to the process.
 317 .\" This replaces the /proc/sys/kernel/rtsig-max system-wide limit
 318 .\" that was present in kernels <= 2.6.7.  MTK Dec 04
 319 .TP
 320 .B RLIMIT_STACK
 321 The maximum size of the process stack, in bytes.
 322 Upon reaching this limit, a
 323 .B SIGSEGV
 324 signal is generated.
 325 To handle this signal, a process must employ an alternate signal stack
 326 .RB ( sigaltstack (2)).
 327 .PP
 328 .B RLIMIT_OFILE
 329 is the BSD name for
 330 .BR RLIMIT_NOFILE .
 331 .SH "RETURN VALUE"
 332 On success, zero is returned.
 333 On error, \-1 is returned, and
 334 .I errno
 335 is set appropriately.
 336 .SH ERRORS
 337 .TP
 338 .B EFAULT
 339 .I rlim
 340 points outside the accessible address space.
 341 .TP
 342 .B EINVAL
 343 .I resource
 344 is not valid;
 345 or, for
 346 .BR setrlimit ():
 347 .IR rlim->rlim_cur
 348 was greater than
 349 .IR rlim->rlim_max .
 350 .TP
 351 .B EPERM
 352 An unprivileged process tried to use
 353 .BR setrlimit ()
 354 to
 355 increase a soft or hard limit above the current hard limit; the
 356 .B CAP_SYS_RESOURCE
 357 capability is required to do this.
 358 Or, the process tried to use
 359 .BR setrlimit ()
 360 to increase
 361 the soft or hard RLIMIT_NOFILE limit above the current kernel
 362 maximum (NR_OPEN).
 363 .SH CONFORMING TO
 364 SVr4, 4.3BSD, POSIX.1-2001.
 365 .BR RLIMIT_MEMLOCK
 366 and
 367 .BR RLIMIT_NPROC
 368 derive from BSD and are not specified in POSIX.1-2001;
 369 they are present on the BSDs and Linux, but on few other implementations.
 370 .BR RLIMIT_RSS
 371 derives from BSD and is not specified in POSIX.1-2001;
 372 it is nevertheless present on most implementations.
 373 .BR RLIMIT_MSGQUEUE ,
 374 .BR RLIMIT_NICE ,
 375 .BR RLIMIT_RTPRIO ,
 376 and
 377 .B RLIMIT_SIGPENDING
 378 are Linux specific.
 379 .SH NOTES
 380 A child process created via
 381 .BR fork (2)
 382 inherits its parents resource limits.
 383 Resource limits are preserved across
 384 .BR execve (2).
 385 .SH BUGS
 386 In older Linux kernels, the
 387 .B SIGXCPU
 388 and
 389 .B SIGKILL
 390 signals delivered when a process encountered the soft and hard
 391 .B RLIMIT_CPU
 392 limits were delivered one (CPU) second later than they should have been.
 393 This was fixed in kernel 2.6.8.
 394
 395 In 2.6.x kernels before 2.6.17, a
 396 .B RLIMIT_CPU
 397 limit of 0 is wrongly treated as "no limit" (like
 398 .BR RLIM_INFINITY ).
 399 Since kernel 2.6.17, setting a limit of 0 does have an effect,
 400 but is actually treated as a limit of 1 second.
 401 .\" see http://marc.theaimsgroup.com/?l=linux-kernel&m=114008066530167&w=2
 402
 403 A kernel bug means that
 404 .B RLIMIT_RTPRIO
 405 does not work in kernel 2.6.12; the problem is fixed in kernel 2.6.13.
 406
 407 In kernel 2.6.12, there was an off-by-one mismatch
 408 between the priority ranges returned by
 409 .BR getpriority (2)
 410 and
 411 .BR RLIMIT_NICE .
 412 This had the effect that actual ceiling for the nice value
 413 was calculated as
 414 .IR "19\ \-\ rlim_cur" .
 415 This was fixed in kernel 2.6.13.
 416 .\" see http://marc.theaimsgroup.com/?l=linux-kernel&m=112256338703880&w=2
 417
 418 Kernels before 2.4.22 did not diagnose the error
 419 .B EINVAL
 420 for
 421 .BR setrlimit ()
 422 when
 423 .IR rlim->rlim_cur
 424 was greater than
 425 .IR rlim->rlim_max .
 426 .SH "SEE ALSO"
 427 .BR dup (2),
 428 .BR fcntl (2),
 429 .BR fork (2),
 430 .BR getrusage (2),
 431 .BR mlock (2),
 432 .BR mmap (2),
 433 .BR open (2),
 434 .BR quotactl (2),
 435 .BR sbrk (2),
 436 .BR shmctl (2),
 437 .BR sigqueue (2),
 438 .BR malloc (3),
 439 .BR ulimit (3),
 440 .BR core (5),
 441 .BR capabilities (7),
 442 .BR signal (7)