.\"
.\" Copyright (c) 2011 Hudson River Trading LLC
.\" Written by: John H. Baldwin <jhb@FreeBSD.org>
.\" All rights reserved.
.\"
.\" Redistribution and use in source and binary forms, with or without
.\" modification, are permitted provided that the following conditions
.\" are met:
.\" 1. Redistributions of source code must retain the above copyright
.\"    notice, this list of conditions and the following disclaimer.
.\" 2. Redistributions in binary form must reproduce the above copyright
.\"    notice, this list of conditions and the following disclaimer in the
.\"    documentation and/or other materials provided with the distribution.
.\"
.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
.\" ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
.\" SUCH DAMAGE.
.\"
.Dd December 14, 2011
.Dt SHM_MAP 9
.Os
.Sh NAME
.Nm shm_map , shm_unmap
.Nd "map shared memory objects into the kernel's address space"
.Sh SYNOPSIS
.In sys/types.h
.In sys/mman.h
.Ft int
.Fn shm_map "struct file *fp" "size_t size" "off_t offset" "void **memp"
.Ft int
.Fn shm_unmap "struct file *fp" "void *mem" "size_t size"
.Sh DESCRIPTION
The
.Fn shm_map
and
.Fn shm_unmap
functions provide an API for mapping shared memory objects into the kernel.
Shared memory objects are created by
.Xr shm_open 2 .
These objects can then be passed into the kernel via file descriptors.
.Pp
A shared memory object cannot be shrunk while it is mapped into the kernel.
This is to avoid invalidating any pages that may be wired into the kernel's
address space.
Shared memory objects can still be grown while mapped into the kernel.
.Pp
To simplify the accounting needed to enforce the above requirement,
callers of this API are required to unmap the entire region mapped by
.Fn shm_map
when calling
.Fn shm_unmap .
Unmapping only a portion of the region is not permitted.
.Pp
The
.Fn shm_map
function locates the shared memory object associated with the open file
.Fa fp .
It maps the region of that object described by
.Fa offset
and
.Fa size
into the kernel's address space.
If it succeeds,
.Fa *memp
will be set to the start of the mapping.
All pages for the range will be wired into memory upon successful return.
.Pp
The
.Fn shm_unmap
function unmaps a region previously mapped by
.Fn shm_map .
The
.Fa mem
argument should match the value previously returned in
.Fa *memp ,
and the
.Fa size
argument should match the value passed to
.Fn shm_map .
.Pp
Note that
.Fn shm_map
will not hold an extra reference on the open file
.Fa fp
for the lifetime of the mapping.
Instead,
the calling code is required to do this if it wishes to use
.Fn shm_unmap
on the region in the future.
.Sh RETURN VALUES
The
.Fn shm_map
and
.Fn shm_unmap
functions return zero on success or an error on failure.
.Sh EXAMPLES
The following function accepts a file descriptor for a shared memory
object.
It maps the first sixteen kilobytes of the object into the kernel,
performs some work on that address,
and then unmaps the address before returning.
.Bd -literal -offset indent
int
shm_example(int fd)
{
	struct file *fp;
	void *mem;
	int error;

	error = fget(curthread, fd, CAP_MMAP, &fp);
	if (error)
		return (error);
	error = shm_map(fp, 16384, 0, &mem);
	if (error) {
		fdrop(fp, curthread);
		return (error);
	}

	/* Do something with 'mem'. */

	error = shm_unmap(fp, mem, 16384);
	fdrop(fp, curthread);
	return (error);
}
.Ed
.Sh ERRORS
The
.Fn shm_map
function returns the following errors on failure:
.Bl -tag -width Er
.It Bq Er EINVAL
The open file
.Fa fp
is not a shared memory object.
.It Bq Er EINVAL
The requested region described by
.Fa offset
and
.Fa size
extends beyond the end of the shared memory object.
.It Bq Er ENOMEM
Insufficient address space was available.
.It Bq Er EACCES
The shared memory object could not be mapped due to a protection error.
.It Bq Er EINVAL
The shared memory object could not be mapped due to some other VM error.
.El
.Pp
The
.Fn shm_unmap
function returns the following errors on failure:
.Bl -tag -width Er
.It Bq Er EINVAL
The open file
.Fa fp
is not a shared memory object.
.It Bq Er EINVAL
The address range described by
.Fa mem
and
.Fa size
is not a valid address range.
.It Bq Er EINVAL
The address range described by
.Fa mem
and
.Fa size
is not backed by the shared memory object associated with the open file
.Fa fp ,
or the address range does not cover the entire mapping of the object.
.El
.Sh SEE ALSO
.Xr shm_open 2
.Sh HISTORY
This API was first introduced in
.Fx 10.0 .