1 .\" Copyright (c) 1992 Drew Eckhardt, March 28, 1992
2 .\" and Copyright (c) 2002, 2004, 2005, 2008, 2010 Michael Kerrisk
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.
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.
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15 .\" manual page may be incorrect or out-of-date.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
22 .\" Formatted or processed versions of this manual, if unaccompanied by
23 .\" the source, must acknowledge the copyright and authors of this work.
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 .\" HÃ¥vard Lygre <hklygre@online.no>
36 .\" Modified 2001-04-17 by Michael Kerrisk <mtk.manpages@gmail.com>
37 .\" Modified 2002-06-13 by Michael Kerrisk <mtk.manpages@gmail.com>
38 .\" Added note on nonstandard behavior when SIGCHLD is ignored.
39 .\" Modified 2002-07-09 by Michael Kerrisk <mtk.manpages@gmail.com>
40 .\" Enhanced descriptions of 'resource' values
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@gmail.com>
44 .\" Added notes on CAP_SYS_RESOURCE
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.
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
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 .\" 2008-05-07, mtk / Peter Zijlstra, Added description of RLIMIT_RTTIME
62 .\" 2010-11-06, mtk: Added documentation of prlimit()
64 .TH GETRLIMIT 2 2014-01-22 "Linux" "Linux Programmer's Manual"
66 getrlimit, setrlimit, prlimit \- get/set resource limits
68 .B #include <sys/time.h>
70 .B #include <sys/resource.h>
72 .BI "int getrlimit(int " resource ", struct rlimit *" rlim );
74 .BI "int setrlimit(int " resource ", const struct rlimit *" rlim );
76 .BI "int prlimit(pid_t " pid ", int " resource \
77 ", const struct rlimit *" new_limit ,
79 .BI " struct rlimit *" old_limit );
82 Feature Test Macro Requirements for glibc (see
83 .BR feature_test_macros (7)):
87 _GNU_SOURCE && _FILE_OFFSET_BITS == 64
93 system calls get and set resource limits respectively.
94 Each resource has an associated soft and hard limit, as defined by the
101 rlim_t rlim_cur; /* Soft limit */
102 rlim_t rlim_max; /* Hard limit (ceiling for rlim_cur) */
107 The soft limit is the value that the kernel enforces for the
108 corresponding resource.
109 The hard limit acts as a ceiling for the soft limit:
110 an unprivileged process may set only its soft limit to a value in the
111 range from 0 up to the hard limit, and (irreversibly) lower its hard limit.
112 A privileged process (under Linux: one with the
114 capability) may make arbitrary changes to either limit value.
118 denotes no limit on a resource (both in the structure returned by
120 and in the structure passed to
125 argument must be one of:
128 The maximum size of the process's virtual memory (address space) in bytes.
129 .\" since 2.0.27 / 2.1.12
130 This limit affects calls to
135 which fail with the error
137 upon exceeding this limit.
138 Also automatic stack expansion will fail
141 that kills the process if no alternate stack
142 has been made available via
143 .BR sigaltstack (2)).
144 Since the value is a \fIlong\fP, on machines with a 32-bit \fIlong\fP
145 either this limit is at most 2 GiB, or this resource is unlimited.
151 When 0 no core dump files are created.
152 When nonzero, larger dumps are truncated to this size.
155 CPU time limit in seconds.
156 When the process reaches the soft limit, it is sent a
159 The default action for this signal is to terminate the process.
160 However, the signal can be caught, and the handler can return control to
162 If the process continues to consume CPU time, it will be sent
164 once per second until the hard limit is reached, at which time
167 (This latter point describes Linux behavior.
168 Implementations vary in how they treat processes which continue to
169 consume CPU time after reaching the soft limit.
170 Portable applications that need to catch this signal should
171 perform an orderly termination upon first receipt of
175 The maximum size of the process's data segment (initialized data,
176 uninitialized data, and heap).
177 This limit affects calls to
181 which fail with the error
183 upon encountering the soft limit of this resource.
186 The maximum size of files that the process may create.
187 Attempts to extend a file beyond this limit result in delivery of a
190 By default, this signal terminates a process, but a process can
191 catch this signal instead, in which case the relevant system call (e.g.,
197 .BR RLIMIT_LOCKS " (Early Linux 2.4 only)"
198 .\" to be precise: Linux 2.4.0-test9; no longer in 2.4.25 / 2.5.65
199 A limit on the combined number of
203 leases that this process may establish.
206 The maximum number of bytes of memory that may be locked
208 In effect this limit is rounded down to the nearest multiple
209 of the system page size.
218 Since Linux 2.6.9 it also affects the
221 operation, where it sets a maximum on the total bytes in
222 shared memory segments (see
224 that may be locked by the real user ID of the calling process.
228 locks are accounted for separately from the per-process memory
235 a process can lock bytes up to this limit in each of these
237 In Linux kernels before 2.6.9, this limit controlled the amount of
238 memory that could be locked by a privileged process.
239 Since Linux 2.6.9, no limits are placed on the amount of memory
240 that a privileged process may lock, and this limit instead governs
241 the amount of memory that an unprivileged process may lock.
243 .BR RLIMIT_MSGQUEUE " (since Linux 2.6.8)"
244 Specifies the limit on the number of bytes that can be allocated
245 for POSIX message queues for the real user ID of the calling process.
246 This limit is enforced for
248 Each message queue that the user creates counts (until it is removed)
249 against this limit according to the formula:
252 bytes = attr.mq_maxmsg * sizeof(struct msg_msg *) +
253 attr.mq_maxmsg * attr.mq_msgsize
260 structure specified as the fourth argument to
263 The first addend in the formula, which includes
264 .I "sizeof(struct msg_msg\ *)"
265 (4 bytes on Linux/i386), ensures that the user cannot
266 create an unlimited number of zero-length messages (such messages
267 nevertheless each consume some system memory for bookkeeping overhead).
269 .BR RLIMIT_NICE " (since Linux 2.6.12, but see BUGS below)"
270 Specifies a ceiling to which the process's nice value can be raised using
274 The actual ceiling for the nice value is calculated as
275 .IR "20\ \-\ rlim_cur" .
276 (This strangeness occurs because negative numbers cannot be specified
277 as resource limit values, since they typically have special meanings.
280 typically is the same as \-1.)
283 Specifies a value one greater than the maximum file descriptor number
284 that can be opened by this process.
290 to exceed this limit yield the error
292 (Historically, this limit was named
297 The maximum number of processes (or, more precisely on Linux, threads)
298 that can be created for the real user ID of the calling process.
299 Upon encountering this limit,
305 Specifies the limit (in pages) of the process's resident set
306 (the number of virtual pages resident in RAM).
307 This limit has effect only in Linux 2.4.x, x < 30, and there
308 affects only calls to
312 .\" As at kernel 2.6.12, this limit still does nothing in 2.6 though
313 .\" talk of making it do something has surfaced from time to time in LKML
316 .BR RLIMIT_RTPRIO " (since Linux 2.6.12, but see BUGS)"
317 Specifies a ceiling on the real-time priority that may be set for
319 .BR sched_setscheduler (2)
321 .BR sched_setparam (2).
323 .BR RLIMIT_RTTIME " (since Linux 2.6.25)"
324 Specifies a limit (in microseconds)
325 on the amount of CPU time that a process scheduled
326 under a real-time scheduling policy may consume without making a blocking
328 For the purpose of this limit,
329 each time a process makes a blocking system call,
330 the count of its consumed CPU time is reset to zero.
331 The CPU time count is not reset if the process continues trying to
332 use the CPU but is preempted, its time slice expires, or it calls
335 Upon reaching the soft limit, the process is sent a
338 If the process catches or ignores this signal and
339 continues consuming CPU time, then
341 will be generated once each second until the hard limit is reached,
342 at which point the process is sent a
346 The intended use of this limit is to stop a runaway
347 real-time process from locking up the system.
349 .BR RLIMIT_SIGPENDING " (since Linux 2.6.8)"
350 Specifies the limit on the number of signals
351 that may be queued for the real user ID of the calling process.
352 Both standard and real-time signals are counted for the purpose of
354 However, the limit is enforced only for
356 it is always possible to use
358 to queue one instance of any of the signals that are not already
359 queued to the process.
360 .\" This replaces the /proc/sys/kernel/rtsig-max system-wide limit
361 .\" that was present in kernels <= 2.6.7. MTK Dec 04
364 The maximum size of the process stack, in bytes.
365 Upon reaching this limit, a
368 To handle this signal, a process must employ an alternate signal stack
369 .RB ( sigaltstack (2)).
372 this limit also determines the amount of space used for the process's
373 command-line arguments and environment variables; for details, see
376 .\" commit c022a0acad534fd5f5d5f17280f6d4d135e74e81
377 .\" Author: Jiri Slaby <jslaby@suse.cz>
378 .\" Date: Tue May 4 18:03:50 2010 +0200
381 system call combines and extends the functionality of
385 It can be used to both set and get the resource limits of an arbitrary process.
389 argument has the same meaning as for
396 argument is a not NULL, then the
398 structure to which it points is used to set new values for
399 the soft and hard limits for
403 argument is a not NULL, then a successful call to
405 places the previous soft and hard limits for
409 structure pointed to by
414 argument specifies the ID of the process on which the call is to operate.
417 is 0, then the call applies to the calling process.
418 To set or get the resources of a process other than itself,
419 the caller must have the
422 real, effective, and saved set user IDs of the target process
423 must match the real user ID of the caller
425 the real, effective, and saved set group IDs of the target process
426 must match the real group ID of the caller.
427 .\" FIXME this permission check is strange
428 .\" Asked about this on LKML, 7 Nov 2010
429 .\" "Inconsistent credential checking in prlimit() syscall"
431 On success, these system calls return 0.
432 On error, \-1 is returned, and
434 is set appropriately.
438 A pointer argument points to a location
439 outside the accessible address space.
442 The value specified in
451 .IR rlim\->rlim_max .
454 An unprivileged process tried to raise the hard limit; the
456 capability is required to do this.
457 Or, the caller tried to increase the hard
459 limit above the current kernel maximum
461 Or, the calling process did not have permission to set limits
462 for the process specified by
466 Could not find a process with the ID specified in
471 system call is available since Linux 2.6.36.
472 Library support is available since glibc 2.13.
476 SVr4, 4.3BSD, POSIX.1-2001.
484 derive from BSD and are not specified in POSIX.1-2001;
485 they are present on the BSDs and Linux, but on few other implementations.
487 derives from BSD and is not specified in POSIX.1-2001;
488 it is nevertheless present on most implementations.
489 .BR RLIMIT_MSGQUEUE ,
497 A child process created via
499 inherits its parent's resource limits.
500 Resource limits are preserved across
503 Lowering the soft limit for a resource below the process's
504 current consumption of that resource will succeed
505 (but will prevent the process from further increasing
506 its consumption of the resource).
508 One can set the resource limits of the shell using the built-in
514 The shell's resource limits are inherited by the processes that
515 it creates to execute commands.
517 Since Linux 2.6.24, the resource limits of any process can be inspected via
518 .IR /proc/[pid]/limits ;
522 Ancient systems provided a
524 function with a similar purpose to
526 For backward compatibility, glibc also provides
528 All new applications should be written using
531 .\" FIXME prlimit() does not suffer
532 .\" https://bugzilla.kernel.org/show_bug.cgi?id=5042
533 .\" http://sources.redhat.com/bugzilla/show_bug.cgi?id=12201
534 .\" Since versions 2.13, glibc has library implementations of
535 .\" getrlimit() and setrlimit() that use prlimit() to work around
537 In older Linux kernels, the
541 signals delivered when a process encountered the soft and hard
543 limits were delivered one (CPU) second later than they should have been.
544 This was fixed in kernel 2.6.8.
546 In 2.6.x kernels before 2.6.17, a
548 limit of 0 is wrongly treated as "no limit" (like
550 Since Linux 2.6.17, setting a limit of 0 does have an effect,
551 but is actually treated as a limit of 1 second.
552 .\" see http://marc.theaimsgroup.com/?l=linux-kernel&m=114008066530167&w=2
554 A kernel bug means that
555 .\" See https://lwn.net/Articles/145008/
557 does not work in kernel 2.6.12; the problem is fixed in kernel 2.6.13.
559 In kernel 2.6.12, there was an off-by-one mismatch
560 between the priority ranges returned by
564 This had the effect that the actual ceiling for the nice value
566 .IR "19\ \-\ rlim_cur" .
567 This was fixed in kernel 2.6.13.
568 .\" see http://marc.theaimsgroup.com/?l=linux-kernel&m=112256338703880&w=2
571 .\" The relevant patch, sent to LKML, seems to be
572 .\" http://thread.gmane.org/gmane.linux.kernel/273462
573 .\" From: Roland McGrath <roland <at> redhat.com>
574 .\" Subject: [PATCH 7/7] make RLIMIT_CPU/SIGXCPU per-process
575 .\" Date: 2005-01-23 23:27:46 GMT
576 if a process reaches its soft
578 limit and has a handler installed for
580 then, in addition to invoking the signal handler,
581 the kernel increases the soft limit by one second.
582 This behavior repeats if the process continues to consume CPU time,
583 until the hard limit is reached,
584 at which point the process is killed.
585 Other implementations
586 .\" Tested Solaris 10, FreeBSD 9, OpenBSD 5.0
589 soft limit in this manner,
590 and the Linux behavior is probably not standards conformant;
591 portable applications should avoid relying on this Linux-specific behavior.
592 .\" FIXME https://bugzilla.kernel.org/show_bug.cgi?id=50951
595 limit exhibits the same behavior when the soft limit is encountered.
597 Kernels before 2.4.22 did not diagnose the error
604 .IR rlim\->rlim_max .
606 The program below demonstrates the use of
611 #define _FILE_OFFSET_BITS 64
616 #include <sys/resource.h>
618 #define errExit(msg) do { perror(msg); exit(EXIT_FAILURE); \\
622 main(int argc, char *argv[])
624 struct rlimit old, new;
628 if (!(argc == 2 || argc == 4)) {
629 fprintf(stderr, "Usage: %s <pid> [<new\-soft\-limit> "
630 "<new\-hard\-limit>]\\n", argv[0]);
634 pid = atoi(argv[1]); /* PID of target process */
638 new.rlim_cur = atoi(argv[2]);
639 new.rlim_max = atoi(argv[3]);
643 /* Set CPU time limit of target process; retrieve and display
646 if (prlimit(pid, RLIMIT_CPU, newp, &old) == \-1)
647 errExit("prlimit\-1");
648 printf("Previous limits: soft=%lld; hard=%lld\\n",
649 (long long) old.rlim_cur, (long long) old.rlim_max);
651 /* Retrieve and display new CPU time limit */
653 if (prlimit(pid, RLIMIT_CPU, NULL, &old) == \-1)
654 errExit("prlimit\-2");
655 printf("New limits: soft=%lld; hard=%lld\\n",
656 (long long) old.rlim_cur, (long long) old.rlim_max);
677 .BR capabilities (7),
680 This page is part of release 3.67 of the Linux
683 A description of the project,
684 information about reporting bugs,
685 and the latest version of this page,
687 \%http://www.kernel.org/doc/man\-pages/.