path: root/share/man/man9/sosplice.9

.\"	$OpenBSD: sosplice.9,v 1.2 2011/03/12 18:31:41 bluhm Exp $
.\"
.\" Copyright (c) 2011 Alexander Bluhm <bluhm@openbsd.org>
.\"
.\" Permission to use, copy, modify, and distribute this software for any
.\" purpose with or without fee is hereby granted, provided that the above
.\" copyright notice and this permission notice appear in all copies.
.\"
.\" THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
.\" WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
.\" MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
.\" ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
.\" WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
.\" ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
.\" OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
.\"
.Dd $Mdocdate: March 12 2011 $
.Dt SOSPLICE 9
.Os
.Sh NAME
.Nm sosplice ,
.Nm somove
.Nd splice two sockets for zero-copy data transfer
.Sh SYNOPSIS
.Ft int
.Fn sosplice "struct socket *so" "int fd" "off_t max"
.Ft int
.Fn somove "struct socket *so" "int wait"
.Sh DESCRIPTION
The function
.Fn sosplice
is used to splice together a source and a drain socket.
The source socket is passed as the
.Fa so
argument;
the file descriptor of the drain is passed in
.Fa fd .
If
.Fa fd
is negative, an existing splicing gets dissolved.
If
.Fa max
is positive, at most that many bytes will get transferred.
Socket splicing can be invoked from user-land via the
.Xr setsockopt 2
system-call at the
.Dv SOL_SOCKET
level with the socket option
.Dv SO_SPLICE .
.Pp
Before connecting both sockets, several checks are executed.
See the
.Sx ERRORS
section for possible failures.
The connection between both sockets is implemented by setting these
additional fields in
.Vt struct socket :
.Pp
.Bl -dash -compact -offset indent
.It
.Vt struct socket Fa *so_splice
links from the source to the drain socket.
.It
.Vt struct socket Fa *so_spliceback
links back from the drain to the source socket.
.It
.Vt off_t Fa so_splicelen
counts the number of bytes spliced so far from this socket.
.It
.Vt off_t Fa so_splicemax
specifies the maximum number of bytes to splice from this socket if
non-zero.
.El
.Pp
After connecting both sockets,
.Fn sosplice
calls
.Fn somove
to transfer the mbufs already in the source receive buffer to the
drain send buffer.
Finally the socket buffer flag
.Dv SB_SPLICE
is set on both socket buffers, to indicate that the protocol layer
has to call
.Fn somove
whenever data or space is available.
.Pp
The function
.Fn somove
transfers data from the source's receive buffer to the drain's send
buffer.
It must be called at
.Xr splsoftnet 9
and
.Fa so
must be a spliced drain socket.
It may be necessary to split an mbuf to handle out-of-band data
inline or when the maximum splice length has been reached.
If
.Fa wait
is
.Dv M_WAIT ,
splitting mbufs will always succeed.
For
.Dv M_DONTWAIT
the out-of-band property might get lost or a short splice might
happen.
In the latter case, less than the given maximum number of bytes are
transferred and user-land has to cope with this.
Note that a short splice cannot happen if
.Fn somove
was called by
.Fn sosplice .
So a second
.Xr setsockopt 2
after a short splice pointing to the same maximum will always
succeed.
.Pp
Before transferring data,
.Fn somove
checks both sockets for errors and that the drain socket is connected.
If the drain cannot send anymore, an
.Er EPIPE
error is set on the source socket.
The data length to move is limited by the optional maximum splice
length and the space in the drain's send socket buffer.
Up to this amount of data is taken out of the source's receive
socket buffer.
.Pp
If the maximum splice length has been reached, an mbuf may get
split.
Otherwise an mbuf is either moved completely to the send buffer or
left in the receive buffer for later processing.
If SO_OOBINLINE is set, out-of-band data will get moved as such
although this might not be reliable.
The data is sent out to the drain socket via the protocol function.
If that fails and the drain socket cannot send anymore, an
.Er EPIPE
error is set on the source socket.
.Pp
Finally the socket splicing gets dissolved if the source socket
cannot receive anymore and its receive buffer is empty; or if the
drain socket cannot send anymore; or if the maximum has been reached;
or if an error occurred.
.Pp
If the socket buffer flag
.Dv SB_SPLICE
is set, the functions
.Fn sorwakeup
and
.Fn sowwakeup
will call
.Fn somove
to trigger the transfer when new data or buffer space is available.
While socket splicing is active, any
.Xr read 2
from the source socket will block and the wakeup will not be delivered
to the file descriptor.
A read event is signaled to user-land after dissolving.
.Sh RETURN VALUES
.Fn sosplice
returns 0 on success and otherwise the error number.
.Fn somove
returns 0 if socket splicing has been finished and 1 if it continues.
.Sh ERRORS
.Fn sosplice
will succeed unless:
.Bl -tag -width Er
.It Bq Er EBADF
The given file descriptor
.Fa fd
is not an active descriptor.
.It Bq Er EBUSY
The source or the drain socket is already spliced.
.It Bq Er EINVAL
The given maximum value
.Fa max
is negative.
.It Bq Er ENOTCONN
The source or the drain socket is neither connected nor in the
process of connecting to a peer.
.It Bq Er ENOTSOCK
The given file descriptor
.Fa fd
is not a socket.
.It Bq Er EOPNOTSUPP
The source or the drain socket is a listen socket.
.It Bq Er EPROTONOSUPPORT
The source socket's protocol layer does not have the
.Dv PR_SPLICE
flag set.
At the moment only TCP supports socket splicing.
.It Bq Er EPROTONOSUPPORT
The drain socket's protocol does not have the same
.Fa pr_usrreq
function as the source.
.It Bq Er EWOULDBLOCK
The source socket is non-blocking and the receive buffer is already
locked.
.El
.Sh SEE ALSO
.Xr setsockopt 2 ,
.Xr options 4
.Sh HISTORY
Socket splicing first appeared in
.Ox 4.9 .
.Sh AUTHORS
.An -nosplit
The idea for socket splicing originally came from
.An Markus Friedl Aq markus@openbsd.org ,
and
.An Alexander Bluhm Aq bluhm@openbsd.org
implemented it.