1 .\" Copyright (c) 1991, 1993
2 .\" The Regents of the University of California. All rights reserved.
4 .\" Redistribution and use in source and binary forms, with or without
5 .\" modification, are permitted provided that the following conditions
7 .\" 1. Redistributions of source code must retain the above copyright
8 .\" notice, this list of conditions and the following disclaimer.
9 .\" 2. Redistributions in binary form must reproduce the above copyright
10 .\" notice, this list of conditions and the following disclaimer in the
11 .\" documentation and/or other materials provided with the distribution.
12 .\" 3. Neither the name of the University nor the names of its contributors
13 .\" may be used to endorse or promote products derived from this software
14 .\" without specific prior written permission.
16 .\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
17 .\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18 .\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
19 .\" ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
20 .\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
21 .\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
22 .\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
23 .\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
24 .\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
25 .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
28 .\" @(#)mmap.2 8.4 (Berkeley) 5/11/95
36 .Nd allocate memory, or map files or devices into memory
42 .Fn mmap "void *addr" "size_t len" "int prot" "int flags" "int fd" "off_t offset"
46 system call causes the pages starting at
48 and continuing for at most
50 bytes to be mapped from the object described by
52 starting at byte offset
56 is not a multiple of the page size, the mapped region may extend past the
58 Any such extension beyond the end of the mapped object will be zero-filled.
62 references a regular file or a shared memory object, the range of
67 bytes must be legitimate for the possible (not necessarily
68 current) offsets in the object.
71 value cannot be negative.
72 If the object is truncated and the process later accesses a page that
73 is wholly within the truncated region, the access is aborted and a
75 signal is delivered to the process.
79 references a device file, the interpretation of the
81 value is device specific and defined by the device driver.
82 The virtual memory subsystem does not impose any restrictitions on the
84 value in this case, passing it unchanged to the driver.
88 is non-zero, it is used as a hint to the system.
89 (As a convenience to the system, the actual address of the region may differ
90 from the address supplied.)
93 is zero, an address will be selected by the system.
94 The actual starting address of the region is returned.
97 deletes any previous mapping in the allocated address range.
99 The protections (region accessibility) are specified in the
103 the following values:
105 .Bl -tag -width PROT_WRITE -compact
107 Pages may not be accessed.
111 Pages may be written.
113 Pages may be executed.
118 argument specifies the type of the mapped object, mapping options and
119 whether modifications made to the mapped copy of the page are private
120 to the process or are to be shared with other references.
121 Sharing, mapping type and options are specified in the
125 the following values:
126 .Bl -tag -width MAP_PREFAULT_READ
128 Request a region in the first 2GB of the current process's address space.
129 If a suitable region cannot be found,
132 This flag is only available on 64-bit platforms.
133 .It Dv MAP_ALIGNED Ns Pq Fa n
134 Align the region on a requested boundary.
135 If a suitable region cannot be found,
140 argument specifies the binary logarithm of the desired alignment.
141 .It Dv MAP_ALIGNED_SUPER
142 Align the region to maximize the potential use of large
145 If a suitable region cannot be found,
148 The system will choose a suitable page size based on the size of
150 The page size used as well as the alignment of the region may both be
151 affected by properties of the file being mapped.
153 the physical address of existing pages of a file may require a specific
155 The region is not guaranteed to be aligned on any specific boundary.
157 Map anonymous memory not associated with any specific file.
158 The file descriptor used for creating
165 .\"Mapped from a regular file or character-special device memory.
167 This flag is identical to
169 and is provided for compatibility.
171 This flag can only be used in combination with
173 Please see the definition of
175 for the description of its effect.
177 Do not permit the system to select a different address than the one
179 If the specified address cannot be used,
186 must be a multiple of the page size.
189 is not specified, a successful
191 request replaces any previous mappings for the process'
192 pages in the range from
200 is specified, the request will fail if a mapping
201 already exists within the range.
203 Instead of a mapping, create a guard of the specified size.
204 Guards allow a process to create reservations in its address space,
205 which can later be replaced by actual mappings.
208 will not create mappings in the address range of a guard unless
209 the request specifies
211 Guards can be destroyed with
213 Any memory access by a thread to the guarded range results
216 signal to that thread.
218 Region is not included in a core file.
220 Causes data dirtied via this VM map to be flushed to physical media
221 only when necessary (usually by the pager) rather than gratuitously.
222 Typically this prevents the update daemons from flushing pages dirtied
223 through such maps and thus allows efficient sharing of memory across
224 unassociated processes using a file-backed shared memory map.
226 this option any VM pages you dirty may be flushed to disk every so often
227 (every 30-60 seconds usually) which can create performance problems if you
228 do not need that to occur (such as when you are using shared file-backed
229 mmap regions for IPC purposes).
230 Dirty data will be flushed automatically when all mappings of an object are
231 removed and all descriptors referencing the object are closed.
232 Note that VM/file system coherency is
233 maintained whether you use
236 This option is not portable
239 platforms (yet), though some may implement the same behavior
243 Extending a file with
245 thus creating a big hole, and then filling the hole by modifying a shared
247 can lead to severe file fragmentation.
248 In order to avoid such fragmentation you should always pre-allocate the
249 file's backing store by
251 zero's into the newly extended area prior to modifying the area via your
253 The fragmentation problem is especially sensitive to
255 pages, because pages may be flushed to disk in a totally random order.
257 The same applies when using
259 to implement a file-based shared memory store.
260 It is recommended that you create the backing store by
262 zero's to the backing file rather than
265 You can test file fragmentation by observing the KB/t (kilobytes per
266 transfer) results from an
268 while reading a large file sequentially, e.g.,\& using
269 .Dq Li dd if=filename of=/dev/null bs=32k .
273 system call will flush all dirty data and metadata associated with a file,
274 including dirty NOSYNC VM data, to physical media.
279 system call generally do not flush dirty NOSYNC VM data.
282 system call is usually not needed since
284 implements a coherent file system buffer cache.
286 used to associate dirty VM pages with file system buffers and thus cause
287 them to be flushed to physical media sooner rather than later.
288 .It Dv MAP_PREFAULT_READ
289 Immediately update the calling process's lowest-level virtual address
290 translation structures, such as its page table, so that every memory
291 resident page within the region is mapped for read access.
292 Ordinarily these structures are updated lazily.
293 The effect of this option is to eliminate any soft faults that would
294 otherwise occur on the initial read accesses to the region.
295 Although this option does not preclude
299 it does not eliminate soft faults on the initial write accesses to the
302 Modifications are private.
304 Modifications are shared.
317 must include at least
323 a memory region that grows to at most
325 bytes in size, starting from the stack top and growing down.
327 stack top is the starting address returned by the call, plus
330 The bottom of the stack at maximum growth is the starting
331 address returned by the call.
336 Guards prevent inadvertent use of the regions into which those
337 stacks can grow without requiring mapping the whole stack in advance.
342 system call does not unmap pages, see
344 for further information.
346 Although this implementation does not impose any alignment restrictions on
349 argument, a portable program must only use page-aligned values.
351 Large page mappings require that the pages backing an object be
352 aligned in matching blocks in both the virtual address space and RAM.
353 The system will automatically attempt to use large page mappings when
354 mapping an object that is already backed by large pages in RAM by
355 aligning the mapping request in the virtual address space to match the
356 alignment of the large physical pages.
357 The system may also use large page mappings when mapping portions of an
358 object that are not yet backed by pages in RAM.
360 .Dv MAP_ALIGNED_SUPER
361 flag is an optimization that will align the mapping request to the
362 size of a large page similar to
364 except that the system will override this alignment if an object already
365 uses large pages so that the mapping will be consistent with the existing
367 This flag is mostly useful for maximizing the use of large pages on the
368 first mapping of objects that do not yet have pages present in RAM.
370 Upon successful completion,
372 returns a pointer to the mapped region.
373 Otherwise, a value of
377 is set to indicate the error.
387 was specified as part of the
391 was not open for reading.
396 were specified as part of the
402 was not open for writing.
407 is not a valid open file descriptor.
409 An invalid (negative) value was passed in the
413 referenced a regular file or shared memory.
415 An invalid value was passed in the
419 An undefined option was set in the
437 At least one of these flags must be included.
440 was specified and the
442 argument was not page aligned, or part of the desired address space
443 resides out of the valid address space for a user process.
449 were specified and part of the desired address space resides outside
450 of the first 2GB of user address space.
458 was specified and the desired alignment was either larger than the
459 virtual address size of the machine or smaller than a page.
462 was specified and the
467 was specified and the
475 were specified, but the requested region is already used by a mapping.
483 was specified, but the
485 argument was not zero, the
487 argument was not -1, or the
493 was specified together with one of the flags
496 .Dv MAP_PREFAULT_READ ,
502 has not been specified and
504 did not reference a regular or character special file.
507 was specified and the
509 argument was not available.
511 was specified and insufficient memory was available.