xref: /freebsd/share/man/man9/shm_map.9 (revision 42249ef2)
1.\"
2.\" Copyright (c) 2011 Hudson River Trading LLC
3.\" Written by: John H. Baldwin <jhb@FreeBSD.org>
4.\" All rights reserved.
5.\"
6.\" Redistribution and use in source and binary forms, with or without
7.\" modification, are permitted provided that the following conditions
8.\" are met:
9.\" 1. Redistributions of source code must retain the above copyright
10.\"    notice, this list of conditions and the following disclaimer.
11.\" 2. Redistributions in binary form must reproduce the above copyright
12.\"    notice, this list of conditions and the following disclaimer in the
13.\"    documentation and/or other materials provided with the distribution.
14.\"
15.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
16.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
17.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
18.\" ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
19.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
20.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
21.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
22.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
23.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
24.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
25.\" SUCH DAMAGE.
26.\"
27.\" $FreeBSD$
28.\"
29.Dd December 14, 2011
30.Dt SHM_MAP 9
31.Os
32.Sh NAME
33.Nm shm_map , shm_unmap
34.Nd "map shared memory objects into the kernel's address space"
35.Sh SYNOPSIS
36.In sys/types.h
37.In sys/mman.h
38.Ft int
39.Fn shm_map "struct file *fp" "size_t size" "off_t offset" "void **memp"
40.Ft int
41.Fn shm_unmap "struct file *fp" "void *mem" "size_t size"
42.Sh DESCRIPTION
43The
44.Fn shm_map
45and
46.Fn shm_unmap
47functions provide an API for mapping shared memory objects into the kernel.
48Shared memory objects are created by
49.Xr shm_open 2 .
50These objects can then be passed into the kernel via file descriptors.
51.Pp
52A shared memory object cannot be shrunk while it is mapped into the kernel.
53This is to avoid invalidating any pages that may be wired into the kernel's
54address space.
55Shared memory objects can still be grown while mapped into the kernel.
56.Pp
57To simplify the accounting needed to enforce the above requirement,
58callers of this API are required to unmap the entire region mapped by
59.Fn shm_map
60when calling
61.Fn shm_unmap .
62Unmapping only a portion of the region is not permitted.
63.Pp
64The
65.Fn shm_map
66function locates the shared memory object associated with the open file
67.Fa fp .
68It maps the region of that object described by
69.Fa offset
70and
71.Fa size
72into the kernel's address space.
73If it succeeds,
74.Fa *memp
75will be set to the start of the mapping.
76All pages for the range will be wired into memory upon successful return.
77.Pp
78The
79.Fn shm_unmap
80function unmaps a region previously mapped by
81.Fn shm_map .
82The
83.Fa mem
84argument should match the value previously returned in
85.Fa *memp ,
86and the
87.Fa size
88argument should match the value passed to
89.Fn shm_map .
90.Pp
91Note that
92.Fn shm_map
93will not hold an extra reference on the open file
94.Fa fp
95for the lifetime of the mapping.
96Instead,
97the calling code is required to do this if it wishes to use
98.Fn shm_unmap
99on the region in the future.
100.Sh RETURN VALUES
101The
102.Fn shm_map
103and
104.Fn shm_unmap
105functions return zero on success or an error on failure.
106.Sh EXAMPLES
107The following function accepts a file descriptor for a shared memory
108object.
109It maps the first sixteen kilobytes of the object into the kernel,
110performs some work on that address,
111and then unmaps the address before returning.
112.Bd -literal -offset indent
113int
114shm_example(int fd)
115{
116	struct file *fp;
117	void *mem;
118	int error;
119
120	error = fget(curthread, fd, CAP_MMAP, &fp);
121	if (error)
122		return (error);
123	error = shm_map(fp, 16384, 0, &mem);
124	if (error) {
125		fdrop(fp, curthread);
126		return (error);
127	}
128
129	/* Do something with 'mem'. */
130
131	error = shm_unmap(fp, mem, 16384);
132	fdrop(fp, curthread);
133	return (error);
134}
135.Ed
136.Sh ERRORS
137The
138.Fn shm_map
139function returns the following errors on failure:
140.Bl -tag -width Er
141.It Bq Er EINVAL
142The open file
143.Fa fp
144is not a shared memory object.
145.It Bq Er EINVAL
146The requested region described by
147.Fa offset
148and
149.Fa size
150extends beyond the end of the shared memory object.
151.It Bq Er ENOMEM
152Insufficient address space was available.
153.It Bq Er EACCES
154The shared memory object could not be mapped due to a protection error.
155.It Bq Er EINVAL
156The shared memory object could not be mapped due to some other VM error.
157.El
158.Pp
159The
160.Fn shm_unmap
161function returns the following errors on failure:
162.Bl -tag -width Er
163.It Bq Er EINVAL
164The open file
165.Fa fp
166is not a shared memory object.
167.It Bq Er EINVAL
168The address range described by
169.Fa mem
170and
171.Fa size
172is not a valid address range.
173.It Bq Er EINVAL
174The address range described by
175.Fa mem
176and
177.Fa size
178is not backed by the shared memory object associated with the open file
179.Fa fp ,
180or the address range does not cover the entire mapping of the object.
181.El
182.Sh SEE ALSO
183.Xr shm_open 2
184.Sh HISTORY
185This API was first introduced in
186.Fx 10.0 .
187