1MLOCK(2)                   Linux Programmer's Manual                  MLOCK(2)
2
3
4

NAME

6       mlock, mlock2, munlock, mlockall, munlockall - lock and unlock memory
7

SYNOPSIS

9       #include <sys/mman.h>
10
11       int mlock(const void *addr, size_t len);
12       int mlock2(const void *addr, size_t len, unsigned int flags);
13       int munlock(const void *addr, size_t len);
14
15       int mlockall(int flags);
16       int munlockall(void);
17

DESCRIPTION

19       mlock(),  mlock2(),  and  mlockall()  lock  part  or all of the calling
20       process's virtual address space into RAM, preventing that  memory  from
21       being paged to the swap area.
22
23       munlock()  and  munlockall()  perform the converse operation, unlocking
24       part or all of the calling process's virtual  address  space,  so  that
25       pages  in  the  specified  virtual  address  range  may once more to be
26       swapped out if required by the kernel memory manager.
27
28       Memory locking and unlocking are performed in units of whole pages.
29
30   mlock(), mlock2(), and munlock()
31       mlock() locks pages in the address range starting at addr and  continu‐
32       ing  for len bytes.  All pages that contain a part of the specified ad‐
33       dress range are guaranteed to be resident in RAM when the call  returns
34       successfully;  the  pages are guaranteed to stay in RAM until later un‐
35       locked.
36
37       mlock2() also locks pages in the specified range starting at  addr  and
38       continuing for len bytes.  However, the state of the pages contained in
39       that range after the call returns successfully will depend on the value
40       in the flags argument.
41
42       The flags argument can be either 0 or the following constant:
43
44       MLOCK_ONFAULT
45              Lock pages that are currently resident and mark the entire range
46              so that the remaining nonresident pages are locked when they are
47              populated by a page fault.
48
49       If flags is 0, mlock2() behaves exactly the same as mlock().
50
51       munlock()  unlocks pages in the address range starting at addr and con‐
52       tinuing for len bytes.  After this call, all pages that contain a  part
53       of the specified memory range can be moved to external swap space again
54       by the kernel.
55
56   mlockall() and munlockall()
57       mlockall() locks all pages mapped into the address space of the calling
58       process.  This includes the pages of the code, data, and stack segment,
59       as well as shared libraries, user space kernel data, shared memory, and
60       memory-mapped files.  All mapped pages are guaranteed to be resident in
61       RAM when the call returns successfully; the  pages  are  guaranteed  to
62       stay in RAM until later unlocked.
63
64       The  flags  argument is constructed as the bitwise OR of one or more of
65       the following constants:
66
67       MCL_CURRENT
68              Lock all pages which are currently mapped into the address space
69              of the process.
70
71       MCL_FUTURE
72              Lock  all  pages which will become mapped into the address space
73              of the process in the future.  These could be, for instance, new
74              pages  required  by a growing heap and stack as well as new mem‐
75              ory-mapped files or shared memory regions.
76
77       MCL_ONFAULT (since Linux 4.4)
78              Used together with MCL_CURRENT, MCL_FUTURE, or both.   Mark  all
79              current  (with MCL_CURRENT) or future (with MCL_FUTURE) mappings
80              to lock pages when they are faulted in.  When used with MCL_CUR‐
81              RENT,  all  present  pages  are  locked, but mlockall() will not
82              fault in non-present pages.  When used with MCL_FUTURE, all  fu‐
83              ture mappings will be marked to lock pages when they are faulted
84              in, but they will not be populated by the lock when the  mapping
85              is created.  MCL_ONFAULT must be used with either MCL_CURRENT or
86              MCL_FUTURE or both.
87
88       If MCL_FUTURE has been specified,  then  a  later  system  call  (e.g.,
89       mmap(2),  sbrk(2), malloc(3)), may fail if it would cause the number of
90       locked bytes to exceed the permitted maximum (see below).  In the  same
91       circumstances,  stack  growth  may  likewise fail: the kernel will deny
92       stack expansion and deliver a SIGSEGV signal to the process.
93
94       munlockall() unlocks all pages mapped into the  address  space  of  the
95       calling process.
96

RETURN VALUE

98       On success, these system calls return 0.  On error, -1 is returned, er‐
99       rno is set to indicate the error, and no changes are made to any  locks
100       in the address space of the process.
101

ERRORS

103       EAGAIN (mlock(),  mlock2(), and munlock()) Some or all of the specified
104              address range could not be locked.
105
106       EINVAL (mlock(), mlock2(), and munlock()) The result  of  the  addition
107              addr+len  was  less  than  addr (e.g., the addition may have re‐
108              sulted in an overflow).
109
110       EINVAL (mlock2()) Unknown flags were specified.
111
112       EINVAL (mlockall()) Unknown flags were  specified  or  MCL_ONFAULT  was
113              specified without either MCL_FUTURE or MCL_CURRENT.
114
115       EINVAL (Not on Linux) addr was not a multiple of the page size.
116
117       ENOMEM (mlock(), mlock2(), and munlock()) Some of the specified address
118              range does not correspond to mapped pages in the  address  space
119              of the process.
120
121       ENOMEM (mlock(), mlock2(), and munlock()) Locking or unlocking a region
122              would result in the total number of mappings with  distinct  at‐
123              tributes  (e.g.,  locked  versus unlocked) exceeding the allowed
124              maximum.  (For example, unlocking a range in  the  middle  of  a
125              currently  locked  mapping  would  result in three mappings: two
126              locked mappings at each end and an unlocked mapping in the  mid‐
127              dle.)
128
129       ENOMEM (Linux  2.6.9 and later) the caller had a nonzero RLIMIT_MEMLOCK
130              soft resource limit, but tried to  lock  more  memory  than  the
131              limit  permitted.   This limit is not enforced if the process is
132              privileged (CAP_IPC_LOCK).
133
134       ENOMEM (Linux 2.4 and earlier) the calling process tried to  lock  more
135              than half of RAM.
136
137       EPERM  The caller is not privileged, but needs privilege (CAP_IPC_LOCK)
138              to perform the requested operation.
139
140       EPERM  (munlockall()) (Linux 2.6.8 and  earlier)  The  caller  was  not
141              privileged (CAP_IPC_LOCK).
142

VERSIONS

144       mlock2()  is available since Linux 4.4; glibc support was added in ver‐
145       sion 2.27.
146

CONFORMING TO

148       mlock(),  munlock(),  mlockall(),   and   munlockall():   POSIX.1-2001,
149       POSIX.1-2008, SVr4.
150
151       mlock2() is Linux specific.
152
153       On  POSIX  systems  on  which  mlock()  and  munlock()  are  available,
154       _POSIX_MEMLOCK_RANGE is defined in <unistd.h> and the number  of  bytes
155       in  a page can be determined from the constant PAGESIZE (if defined) in
156       <limits.h> or by calling sysconf(_SC_PAGESIZE).
157
158       On POSIX systems on which mlockall() and  munlockall()  are  available,
159       _POSIX_MEMLOCK  is  defined  in  <unistd.h>  to a value greater than 0.
160       (See also sysconf(3).)
161

NOTES

163       Memory locking has two  main  applications:  real-time  algorithms  and
164       high-security  data  processing.  Real-time applications require deter‐
165       ministic timing, and, like scheduling, paging is one major cause of un‐
166       expected program execution delays.  Real-time applications will usually
167       also switch to a real-time scheduler with sched_setscheduler(2).  Cryp‐
168       tographic security software often handles critical bytes like passwords
169       or secret keys as data structures.  As a result of  paging,  these  se‐
170       crets  could  be transferred onto a persistent swap store medium, where
171       they might be accessible to the enemy long after the security  software
172       has  erased  the secrets in RAM and terminated.  (But be aware that the
173       suspend mode on laptops and some desktop computers will save a copy  of
174       the system's RAM to disk, regardless of memory locks.)
175
176       Real-time processes that are using mlockall() to prevent delays on page
177       faults should reserve enough locked stack  pages  before  entering  the
178       time-critical  section, so that no page fault can be caused by function
179       calls.  This can be achieved by calling a  function  that  allocates  a
180       sufficiently large automatic variable (an array) and writes to the mem‐
181       ory occupied by this array in order to touch these stack  pages.   This
182       way,  enough  pages will be mapped for the stack and can be locked into
183       RAM.  The dummy writes ensure that not even copy-on-write  page  faults
184       can occur in the critical section.
185
186       Memory  locks  are not inherited by a child created via fork(2) and are
187       automatically removed  (unlocked)  during  an  execve(2)  or  when  the
188       process terminates.  The mlockall() MCL_FUTURE and MCL_FUTURE | MCL_ON‐
189       FAULT settings are not inherited by a child created via fork(2) and are
190       cleared during an execve(2).
191
192       Note  that  fork(2)  will prepare the address space for a copy-on-write
193       operation.  The consequence is that any write access that follows  will
194       cause  a  page  fault that in turn may cause high latencies for a real-
195       time process.  Therefore, it is crucial not to invoke fork(2) after  an
196       mlockall()  or mlock() operation—not even from a thread which runs at a
197       low priority within a process which also has a thread running  at  ele‐
198       vated priority.
199
200       The memory lock on an address range is automatically removed if the ad‐
201       dress range is unmapped via munmap(2).
202
203       Memory locks do not stack, that is, pages which have been  locked  sev‐
204       eral  times  by  calls  to mlock(), mlock2(), or mlockall() will be un‐
205       locked by a single call to munlock() for the corresponding range or  by
206       munlockall().   Pages  which are mapped to several locations or by sev‐
207       eral processes stay locked into RAM as long as they are locked at least
208       at one location or by at least one process.
209
210       If  a  call to mlockall() which uses the MCL_FUTURE flag is followed by
211       another call that does not specify this flag, the changes made  by  the
212       MCL_FUTURE call will be lost.
213
214       The mlock2() MLOCK_ONFAULT flag and the mlockall() MCL_ONFAULT flag al‐
215       low efficient memory locking for applications that deal with large map‐
216       pings where only a (small) portion of pages in the mapping are touched.
217       In such cases, locking all of the pages in a mapping would incur a sig‐
218       nificant penalty for memory locking.
219
220   Linux notes
221       Under  Linux, mlock(), mlock2(), and munlock() automatically round addr
222       down to the nearest page boundary.  However, the POSIX.1  specification
223       of  mlock() and munlock() allows an implementation to require that addr
224       is page aligned, so portable applications should ensure this.
225
226       The VmLck field of the Linux-specific /proc/[pid]/status file shows how
227       many  kilobytes  of  memory  the  process  with ID PID has locked using
228       mlock(), mlock2(), mlockall(), and mmap(2) MAP_LOCKED.
229
230   Limits and permissions
231       In Linux 2.6.8 and earlier, a process must be privileged (CAP_IPC_LOCK)
232       in  order to lock memory and the RLIMIT_MEMLOCK soft resource limit de‐
233       fines a limit on how much memory the process may lock.
234
235       Since Linux 2.6.9, no limits are placed on the amount of memory that  a
236       privileged  process can lock and the RLIMIT_MEMLOCK soft resource limit
237       instead defines a limit on how much memory an unprivileged process  may
238       lock.
239

BUGS

241       In  Linux  4.8  and earlier, a bug in the kernel's accounting of locked
242       memory for unprivileged processes (i.e.,  without  CAP_IPC_LOCK)  meant
243       that  if  the  region  specified by addr and len overlapped an existing
244       lock, then the already locked bytes  in  the  overlapping  region  were
245       counted  twice when checking against the limit.  Such double accounting
246       could incorrectly calculate a  "total  locked  memory"  value  for  the
247       process  that  exceeded  the RLIMIT_MEMLOCK limit, with the result that
248       mlock() and mlock2() would fail on requests that should have succeeded.
249       This bug was fixed in Linux 4.9.
250
251       In  the  2.4  series  Linux  kernels  up to and including 2.4.17, a bug
252       caused the mlockall() MCL_FUTURE flag to be inherited across a fork(2).
253       This was rectified in kernel 2.4.18.
254
255       Since  kernel 2.6.9, if a privileged process calls mlockall(MCL_FUTURE)
256       and later drops privileges (loses the CAP_IPC_LOCK capability  by,  for
257       example, setting its effective UID to a nonzero value), then subsequent
258       memory allocations (e.g., mmap(2), brk(2)) will fail if the RLIMIT_MEM‐
259       LOCK resource limit is encountered.
260

SEE ALSO

262       mincore(2),  mmap(2), setrlimit(2), shmctl(2), sysconf(3), proc(5), ca‐
263       pabilities(7)
264

COLOPHON

266       This page is part of release 5.13 of the Linux  man-pages  project.   A
267       description  of  the project, information about reporting bugs, and the
268       latest    version    of    this    page,    can     be     found     at
269       https://www.kernel.org/doc/man-pages/.
270
271
272
273Linux                             2021-08-27                          MLOCK(2)
Impressum