xref: /illumos-gate/usr/src/man/man3c/mlock.3c (revision 8119dad84d6416f13557b0ba8e2aaf9064cbcfd3)
1.\"
2.\" Copyright 2022 Oxide Computer Company
3.\" Copyright (c) 2007, Sun Microsystems, Inc.  All Rights Reserved.
4.\" Copyright 1989 AT&T
5.\"
6.\" The contents of this file are subject to the terms of the
7.\" Common Development and Distribution License (the "License").
8.\" You may not use this file except in compliance with the License.
9.\"
10.\" You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
11.\" or http://www.opensolaris.org/os/licensing.
12.\" See the License for the specific language governing permissions
13.\" and limitations under the License.
14.\"
15.\" When distributing Covered Code, include this CDDL HEADER in each
16.\" file and include the License file at usr/src/OPENSOLARIS.LICENSE.
17.\" If applicable, add the following below this CDDL HEADER, with the
18.\" fields enclosed by brackets "[]" replaced with your own identifying
19.\" information: Portions Copyright [yyyy] [name of copyright owner]
20.\"
21.Dd March 13, 2022
22.Dt MLOCK 3C
23.Os
24.Sh NAME
25.Nm mlock ,
26.Nm munlock
27.Nd lock or unlock pages in memory
28.Sh SYNOPSIS
29.In sys/mman.h
30.Ft int
31.Fo mlock
32.Fa "const void *addr"
33.Fa "size_t len"
34.Fc
35.Ft int
36.Fo munlock
37.Fa "const void *addr"
38.Fa "size_t len"
39.Fc
40.Sh DESCRIPTION
41The
42.Fn mlock
43function uses the mappings established for the address range
44.Pf [ Fa addr ,
45.Fa addr
46+
47.Fa len Ns
48) to identify pages to be locked in memory.
49If the page identified by a mapping changes, such as occurs when a copy of a
50writable
51.Dv MAP_PRIVATE
52page is made upon the first store, the lock will be transferred to the newly
53copied private page.
54.Pp
55The
56.Fn munlock
57function removes locks established with
58.Fn mlock .
59.Pp
60A given page may be locked multiple times by executing an
61.Fn mlock
62through different mappings.
63That is, if two different processes lock the same page, then the page will
64remain locked until both processes remove their locks.
65However, within a given mapping, page locks do not nest \(em multiple
66.Fn mlock
67operations on the same address in the same process will all be removed with a
68single
69.Fn munlock .
70Of course, a page locked in one process and mapped in another
71.Pq or visible through a different mapping in the locking process
72is still locked in memory.
73This fact can be used to create applications that do nothing other than lock
74important data in memory, thereby avoiding page I/O faults on references from
75other processes in the system.
76.Pp
77The contents of the locked pages will not be transferred to or from disk except
78when explicitly requested by one of the locking processes.
79This guarantee applies only to the mapped data, and not to any associated data
80structures
81.Pq file descriptors and on-disk metadata, among others .
82.Pp
83If the mapping through which an
84.Fn mlock
85has been performed is removed, an
86.Fn munlock
87is implicitly performed.
88An
89.Fn munlock
90is also performed implicitly when a page is deleted through file removal or
91truncation.
92.Pp
93Locks established with
94.Fn mlock
95are not inherited by a child process after
96a
97.Xr fork 2
98and are not nested.
99.Pp
100Attempts to
101.Fn mlock
102more memory than a system-specific limit will fail.
103.Sh RETURN VALUES
104Upon successful completion, the
105.Fn mlock
106and
107.Fn munlock
108functions return
109.Sy 0 .
110Otherwise, no changes are made to any locks in the address space of the process,
111the functions return
112.Sy -1
113and set
114.Va errno
115to indicate the error.
116.Sh ERRORS
117The
118.Fn mlock
119and
120.Fn munlock
121functions will fail if:
122.Bl -tag -width Er
123.It Er EINVAL
124The
125.Fa addr
126argument is not a multiple of the page size as returned by
127.Xr sysconf 3C .
128.It Er ENOMEM
129Addresses in the range
130.Pf [ Fa addr ,
131.Fa addr
132+
133.Fa len Ns
134) are invalid for the address space of a process, or specify one or more pages
135which are not mapped.
136.It Er ENOSYS
137The system does not support this memory locking interface.
138.It Er EPERM
139The
140.Brq PRIV_PROC_LOCK_MEMORY
141privilege is not asserted in the effective set of the calling process.
142.El
143.Pp
144The
145.Fn mlock
146function will fail if:
147.Bl -tag -width Er
148.It Er EAGAIN
149.ad
150Some or all of the memory identified by the range
151.Pf [ Fa addr ,
152.Fa addr
153+
154.Fa len Ns
155) could not be locked because of insufficient system resources or because of a
156limit or resource control on locked memory.
157.El
158.Sh USAGE
159Because of the impact on system resources, the use of
160.Fn mlock
161and
162.Fn munlock
163is restricted to users with the
164.Brq PRIV_PROC_LOCK_MEMORY
165privilege.
166.Sh INTERFACE STABILITY
167.Sy Committed
168.Sh MT-LEVEL
169.Sy MT-Safe
170.Sh SEE ALSO
171.Xr fork 2 ,
172.Xr memcntl 2 ,
173.Xr mmap 2 ,
174.Xr mlockall 3C ,
175.Xr plock 3C ,
176.Xr sysconf 3C ,
177.Xr attributes 7 ,
178.Xr standards 7
179