freebsd-dev/share/man/man9/zero_copy.9

170 lines
6.7 KiB
Groff

.\"
.\" Copyright (c) 2002 Kenneth D. Merry.
.\" 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,
.\" without modification, immediately at the beginning of the file.
.\" 2. The name of the author may not be used to endorse or promote products
.\" derived from this software without specific prior written permission.
.\"
.\" 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 5, 2004
.Dt ZERO_COPY 9
.Os
.Sh NAME
.Nm zero_copy ,
.Nm zero_copy_sockets
.Nd "zero copy sockets code"
.Sh SYNOPSIS
.Cd "options ZERO_COPY_SOCKETS"
.Sh DESCRIPTION
The
.Fx
kernel includes a facility for eliminating data copies on
socket reads and writes.
.Pp
This code is collectively known as the zero copy sockets code, because during
normal network I/O, data will not be copied by the CPU at all.
Rather it
will be DMAed from the user's buffer to the NIC (for sends), or DMAed from
the NIC to a buffer that will then be given to the user (receives).
.Pp
The zero copy sockets code uses the standard socket read and write
semantics, and therefore has some limitations and restrictions that
programmers should be aware of when trying to take advantage of this
functionality.
.Pp
For sending data, there are no special requirements or capabilities that
the sending NIC must have.
The data written to the socket, though, must be
at least a page in size and page aligned in order to be mapped into the
kernel.
If it does not meet the page size and alignment constraints, it
will be copied into the kernel, as is normally the case with socket I/O.
.Pp
The user should be careful not to overwrite buffers that have been written
to the socket before the data has been freed by the kernel, and the
copy-on-write mapping cleared.
If a buffer is overwritten before it has
been given up by the kernel, the data will be copied, and no savings in CPU
utilization and memory bandwidth utilization will be realized.
.Pp
The
.Xr socket 2
API does not really give the user any indication of when his data has
actually been sent over the wire, or when the data has been freed from
kernel buffers.
For protocols like TCP, the data will be kept around in
the kernel until it has been acknowledged by the other side; it must be
kept until the acknowledgement is received in case retransmission is required.
.Pp
From an application standpoint, the best way to guarantee that the data has
been sent out over the wire and freed by the kernel (for TCP-based sockets)
is to set a socket buffer size (see the
.Dv SO_SNDBUF
socket option in the
.Xr setsockopt 2
man page) appropriate for the application and network environment and then
make sure you have sent out twice as much data as the socket buffer size
before reusing a buffer.
For TCP, the send and receive socket buffer sizes
generally directly correspond to the TCP window size.
.Pp
For receiving data, in order to take advantage of the zero copy receive
code, the user must have a NIC that is configured for an MTU greater than
the architecture page size.
(E.g., for alpha this would be 8KB, for i386,
it would be 4KB.)
Additionally, in order for zero copy receive to work,
packet payloads must be at least a page in size and page aligned.
.Pp
Achieving page aligned payloads requires a NIC that can split an incoming
packet into multiple buffers.
It also generally requires some sort of
intelligence on the NIC to make sure that the payload starts in its own
buffer.
This is called
.Dq "header splitting" .
Currently the only NICs with
support for header splitting are Alteon Tigon 2 based boards running
slightly modified firmware.
The
.Fx
.Xr ti 4
driver includes modified firmware for Tigon 2 boards only.
Header
splitting code can be written, however, for any NIC that allows putting
received packets into multiple buffers and that has enough programmability
to determine that the header should go into one buffer and the payload into
another.
.Pp
You can also do a form of header splitting that does not require any NIC
modifications if your NIC is at least capable of splitting packets into
multiple buffers.
This requires that you optimize the NIC driver for your
most common packet header size.
If that size (ethernet + IP + TCP headers)
is generally 66 bytes, for instance, you would set the first buffer in a
set for a particular packet to be 66 bytes long, and then subsequent
buffers would be a page in size.
For packets that have headers that are
exactly 66 bytes long, your payload will be page aligned.
.Pp
The other requirement for zero copy receive to work is that the buffer that
is the destination for the data read from a socket must be at least a page
in size and page aligned.
.Pp
Obviously the requirements for receive side zero copy are impossible to
meet without NIC hardware that is programmable enough to do header
splitting of some sort.
Since most NICs are not that programmable, or their
manufacturers will not share the source code to their firmware, this approach
to zero copy receive is not widely useful.
.Pp
There are other approaches, such as RDMA and TCP Offload, that may
potentially help alleviate the CPU overhead associated with copying data
out of the kernel.
Most known techniques require some sort of support at
the NIC level to work, and describing such techniques is beyond the scope
of this manual page.
.Pp
The zero copy send and zero copy receive code can be individually turned
off via the
.Va kern.ipc.zero_copy.send
and
.Va kern.ipc.zero_copy.receive
.Nm sysctl
variables respectively.
.Sh SEE ALSO
.Xr sendfile 2 ,
.Xr socket 2 ,
.Xr ti 4
.Sh HISTORY
The zero copy sockets code first appeared in
.Fx 5.0 ,
although it has
been in existence in patch form since at least mid-1999.
.Sh AUTHORS
.An -nosplit
The zero copy sockets code was originally written by
.An Andrew Gallatin Aq gallatin@FreeBSD.org
and substantially modified and updated by
.An Kenneth Merry Aq ken@FreeBSD.org .