xref: /freebsd/share/man/man9/sf_buf.9 (revision f4c129f5fb1669b284d76a838b1e509f97c2d299)
1.\"
2.\" Copyright (c) 2007 Seccuris Inc.
3.\" All rights reserved.
4.\"
5.\" This documentation was written by Robert N. M. Watson under contract to
6.\" Seccuris Inc.
7.\"
8.\" Redistribution and use in source and binary forms, with or without
9.\" modification, are permitted provided that the following conditions
10.\" are met:
11.\" 1. Redistributions of source code must retain the above copyright
12.\"    notice(s), this list of conditions and the following disclaimer as
13.\"    the first lines of this file unmodified other than the possible
14.\"    addition of one or more copyright notices.
15.\" 2. Redistributions in binary form must reproduce the above copyright
16.\"    notice(s), this list of conditions and the following disclaimer in the
17.\"    documentation and/or other materials provided with the distribution.
18.\"
19.\" THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDER(S) ``AS IS'' AND ANY
20.\" EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
21.\" WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
22.\" DISCLAIMED.  IN NO EVENT SHALL THE COPYRIGHT HOLDER(S) BE LIABLE FOR ANY
23.\" DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
24.\" (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
25.\" SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
26.\" CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
27.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
28.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH
29.\" DAMAGE.
30.\"
31.\" $FreeBSD$
32.\"
33.Dd January 28, 2007
34.Dt SF_BUF 9
35.Os
36.Sh NAME
37.Nm sf_buf
38.Nd manage temporary kernel address space mapping for memory pages
39.Sh SYNOPSIS
40.In sys/sf_buf.h
41.Ft struct sf_buf *
42.Fn sf_buf_alloc "struct vm_page *m" "int flags"
43.Ft void
44.Fn sf_buf_free "struct sf_buf *sf"
45.Ft vm_offset_t
46.Fn sf_buf_kva "struct sf_buf *sf"
47.Ft struct vm_page *
48.Fn sf_buf_page "struct sf_buf *sf"
49.Sh DESCRIPTION
50The
51.Nm
52interface, historically the
53.Xr sendfile 2
54buffer interface, allows kernel subsystems to manage temporary kernel address
55space mappings for physical memory pages.
56On systems with a direct memory map region (allowing all physical pages to be
57visible in the kernel address space at all times), the
58.Vt "struct sf_buf"
59will point to an address in the direct map region; on systems without a
60direct memory map region, the
61.Vt "struct sf_buf"
62will manage a temporary kernel address space mapping valid for the lifetime
63of the
64.Vt "struct sf_buf".
65.Pp
66Call
67.Fn sf_buf_alloc
68to allocate a
69.Vt "struct sf_buf"
70for a physical memory page.
71.Fn sf_buf_alloc
72is not responsible for arranging for the page to be present in physical
73memory; the caller should already have arranged for the page to be wired,
74i.e., by calling
75.Xr vm_page_wire 9 .
76Several flags may be passed to
77.Fn sf_buf_alloc :
78.Bl -tag -width SFB_CPUPRIVATE
79.It Dv SFB_CATCH
80Cause
81.Fn sf_buf_alloc
82to abort and return
83.Dv NULL
84if a signal is received waiting for a
85.Vt "struct sf_buf"
86to become available.
87.It Dv SFB_NOWAIT
88Cause
89.Fn sf_buf_alloc
90to return
91.Dv NULL
92rather than sleeping if a
93.Vt "struct sf_buf"
94is not immediately available.
95.It Dv SFB_CPUPRIVATE
96Cause
97.Fn sf_buf_alloc
98to only arrange that the temporary mapping be valid on the current CPU,
99avoiding unnecessary TLB shootdowns for mappings that will only be accessed
100on a single CPU at a time.
101The caller must ensure that accesses to the virtual address occur only on the
102CPU from which
103.Fn sf_buf_alloc
104was invoked, perhaps by using
105.Fn sched_pin .
106.El
107.Pp
108Call
109.Fn sf_buf_kva
110to return a kernel mapped address for the page.
111.Pp
112Call
113.Fn sf_buf_page
114to return a pointer to the page originally passed into
115.Fn sf_buf_alloc .
116.Pp
117Call
118.Fn sf_buf_free
119to release the
120.Vt "struct sf_buf"
121reference.
122The caller is responsible for releasing any wiring they have previously
123acquired on the physical page;
124.Fn sf_buf_free
125releases only the temporary kernel address space mapping, not the page
126itself.
127.Pp
128Uses of this interface include managing mappings of borrowed pages from user
129memory, such as in zero-copy socket I/O, or pages of memory from the buffer
130cache referenced by mbuf external storage for
131.Xr sendfile 2 .
132.Sh SEE ALSO
133.Xr sendfile 2 ,
134.Xr vm_page_wire 9
135.Sh AUTHORS
136.An -nosplit
137The
138.Vt "struct sf_buf"
139API was designed and implemented by
140.An Alan L. Cox .
141This manual page was written by
142.An Robert N. M. Watson .
143