diff options
author | jhb <jhb@FreeBSD.org> | 2011-12-14 22:22:19 +0000 |
---|---|---|
committer | jhb <jhb@FreeBSD.org> | 2011-12-14 22:22:19 +0000 |
commit | 6d1299a388490b22f5d0619237868efafa3f3428 (patch) | |
tree | 99542359d8ee93e187d91428fb4d8a6e2d87bc67 /share | |
parent | 83d1e21dd2a8af164c5792531d349adad212dee8 (diff) | |
download | FreeBSD-src-6d1299a388490b22f5d0619237868efafa3f3428.zip FreeBSD-src-6d1299a388490b22f5d0619237868efafa3f3428.tar.gz |
Add a helper API to allow in-kernel code to map portions of shared memory
objects created by shm_open(2) into the kernel's address space. This
provides a convenient way for creating shared memory buffers between
userland and the kernel without requiring custom character devices.
Diffstat (limited to 'share')
-rw-r--r-- | share/man/man9/Makefile | 2 | ||||
-rw-r--r-- | share/man/man9/shm_map.9 | 187 |
2 files changed, 189 insertions, 0 deletions
diff --git a/share/man/man9/Makefile b/share/man/man9/Makefile index a577854..b150b64 100644 --- a/share/man/man9/Makefile +++ b/share/man/man9/Makefile @@ -234,6 +234,7 @@ MAN= accept_filter.9 \ sema.9 \ sf_buf.9 \ sglist.9 \ + shm_map.9 \ signal.9 \ sleep.9 \ sleepqueue.9 \ @@ -1111,6 +1112,7 @@ MLINKS+=sglist.9 sglist_alloc.9 \ sglist.9 sglist_reset.9 \ sglist.9 sglist_slice.9 \ sglist.9 sglist_split.9 +MLINKS+=shm_map.9 shm_unmap.9 MLINKS+=signal.9 cursig.9 \ signal.9 execsigs.9 \ signal.9 issignal.9 \ diff --git a/share/man/man9/shm_map.9 b/share/man/man9/shm_map.9 new file mode 100644 index 0000000..6f61147 --- /dev/null +++ b/share/man/man9/shm_map.9 @@ -0,0 +1,187 @@ +.\" +.\" Copyright (c) 2011 Advanced Computing Technologies 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. +.\" +.\" $FreeBSD$ +.\" +.Dd December 14, 2011 +.Dt SHM_MAP 9 +.Os +.Sh NAME +.Nm shm_map , +.Nm 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 +.Nm shm_map +and +.Nm 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 +.Nm shm_map +when calling +.Nm shm_unmap . +Unmapping only a portion of the region is not permitted. +.Pp +The +.Nm 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 +.Nm shm_unmap +function unmaps a region previously mapped by +.Nm 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 +.Nm shm_map . +.Pp +Note that +.Nm 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 +.Nm shm_unmap +on the region in the future. +.Sh RETURN VALUES +The +.Nm shm_map +and +.Nm 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 +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 +.Nm 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 +.Nm 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 . |