xref: /freebsd/share/man/man9/shm_map.9 (revision 2e3f49888ec8851bafb22011533217487764fdb0)
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.Dd December 14, 2011
28.Dt SHM_MAP 9
29.Os
30.Sh NAME
31.Nm shm_map , shm_unmap
32.Nd "map shared memory objects into the kernel's address space"
33.Sh SYNOPSIS
34.In sys/types.h
35.In sys/mman.h
36.Ft int
37.Fn shm_map "struct file *fp" "size_t size" "off_t offset" "void **memp"
38.Ft int
39.Fn shm_unmap "struct file *fp" "void *mem" "size_t size"
40.Sh DESCRIPTION
41The
42.Fn shm_map
43and
44.Fn shm_unmap
45functions provide an API for mapping shared memory objects into the kernel.
46Shared memory objects are created by
47.Xr shm_open 2 .
48These objects can then be passed into the kernel via file descriptors.
49.Pp
50A shared memory object cannot be shrunk while it is mapped into the kernel.
51This is to avoid invalidating any pages that may be wired into the kernel's
52address space.
53Shared memory objects can still be grown while mapped into the kernel.
54.Pp
55To simplify the accounting needed to enforce the above requirement,
56callers of this API are required to unmap the entire region mapped by
57.Fn shm_map
58when calling
59.Fn shm_unmap .
60Unmapping only a portion of the region is not permitted.
61.Pp
62The
63.Fn shm_map
64function locates the shared memory object associated with the open file
65.Fa fp .
66It maps the region of that object described by
67.Fa offset
68and
69.Fa size
70into the kernel's address space.
71If it succeeds,
72.Fa *memp
73will be set to the start of the mapping.
74All pages for the range will be wired into memory upon successful return.
75.Pp
76The
77.Fn shm_unmap
78function unmaps a region previously mapped by
79.Fn shm_map .
80The
81.Fa mem
82argument should match the value previously returned in
83.Fa *memp ,
84and the
85.Fa size
86argument should match the value passed to
87.Fn shm_map .
88.Pp
89Note that
90.Fn shm_map
91will not hold an extra reference on the open file
92.Fa fp
93for the lifetime of the mapping.
94Instead,
95the calling code is required to do this if it wishes to use
96.Fn shm_unmap
97on the region in the future.
98.Sh RETURN VALUES
99The
100.Fn shm_map
101and
102.Fn shm_unmap
103functions return zero on success or an error on failure.
104.Sh EXAMPLES
105The following function accepts a file descriptor for a shared memory
106object.
107It maps the first sixteen kilobytes of the object into the kernel,
108performs some work on that address,
109and then unmaps the address before returning.
110.Bd -literal -offset indent
111int
112shm_example(int fd)
113{
114	struct file *fp;
115	void *mem;
116	int error;
117
118	error = fget(curthread, fd, CAP_MMAP, &fp);
119	if (error)
120		return (error);
121	error = shm_map(fp, 16384, 0, &mem);
122	if (error) {
123		fdrop(fp, curthread);
124		return (error);
125	}
126
127	/* Do something with 'mem'. */
128
129	error = shm_unmap(fp, mem, 16384);
130	fdrop(fp, curthread);
131	return (error);
132}
133.Ed
134.Sh ERRORS
135The
136.Fn shm_map
137function returns the following errors on failure:
138.Bl -tag -width Er
139.It Bq Er EINVAL
140The open file
141.Fa fp
142is not a shared memory object.
143.It Bq Er EINVAL
144The requested region described by
145.Fa offset
146and
147.Fa size
148extends beyond the end of the shared memory object.
149.It Bq Er ENOMEM
150Insufficient address space was available.
151.It Bq Er EACCES
152The shared memory object could not be mapped due to a protection error.
153.It Bq Er EINVAL
154The shared memory object could not be mapped due to some other VM error.
155.El
156.Pp
157The
158.Fn shm_unmap
159function returns the following errors on failure:
160.Bl -tag -width Er
161.It Bq Er EINVAL
162The open file
163.Fa fp
164is not a shared memory object.
165.It Bq Er EINVAL
166The address range described by
167.Fa mem
168and
169.Fa size
170is not a valid address range.
171.It Bq Er EINVAL
172The address range described by
173.Fa mem
174and
175.Fa size
176is not backed by the shared memory object associated with the open file
177.Fa fp ,
178or the address range does not cover the entire mapping of the object.
179.El
180.Sh SEE ALSO
181.Xr shm_open 2
182.Sh HISTORY
183This API was first introduced in
184.Fx 10.0 .
185