.\"
.\" Copyright (c) 2004-2005
.\"	Hartmut Brandt.
.\" 	All rights reserved.
.\" Copyright (c) 2001-2003
.\"	Fraunhofer Institute for Open Communication Systems (FhG Fokus).
.\" 	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.
.\"
.\" Author: Hartmut Brandt <harti@freebsd.org>
.\"
.\" $Begemot: libunimsg/man/unimsg.3,v 1.4 2005/06/15 11:37:10 brandt_h Exp $
.\"
.Dd June 14, 2005
.Dt UNIMSG 3
.Os
.Sh NAME
.Nm uni_msg_len ,
.Nm uni_msg_space ,
.Nm uni_msg_leading ,
.Nm uni_msg_size ,
.Nm uni_msg_ensure ,
.Nm uni_msg_append ,
.Nm uni_msg_extend ,
.Nm uni_msg_alloc ,
.Nm uni_msg_build ,
.Nm uni_msg_destroy ,
.Nm uni_msg_strip32 ,
.Nm uni_msg_get32 ,
.Nm uni_msg_append32 ,
.Nm uni_msg_append8 ,
.Nm uni_msg_trail32 ,
.Nm uni_msg_dup
.Nd "ATM signalling library - message buffers"
.Sh LIBRARY
Begemot ATM signalling library
.Pq libunimsg, -lunimsg
.Sh SYNOPSIS
.In uni4/unimsg.h
.Ft size_t
.Fn uni_msg_len "const struct uni_msg *msg"
.Ft size_t
.Fn uni_msg_space "const struct uni_msg *msg"
.Ft size_t
.Fn uni_msg_leading "const struct uni_msg *msg"
.Ft size_t
.Fn uni_msg_size "const struct uni_msg *msg"
.Ft int
.Fn uni_msg_ensure "struct uni_msg *msg" "size_t bytes"
.Ft int
.Fn uni_msg_append "struct uni_msg *msg" "void *buf" "size_t buflen"
.Ft int
.Fn uni_msg_extend "struct uni_msg *msg" "size_t bytes"
.Ft struct uni_msg *
.Fn uni_msg_alloc "size_t space"
.Ft struct uni_msg *
.Fn uni_msg_build "void *buf" "..."
.Ft void
.Fn uni_msg_destroy "struct uni_msg *msg"
.Ft u_int
.Fn uni_msg_strip32 "struct uni_msg *msg"
.Ft u_int
.Fn uni_msg_get32 "struct uni_msg *msg"
.Ft int
.Fn uni_msg_append32 "struct uni_msg *msg" "u_int value"
.Ft int
.Fn uni_msg_append8 "struct uni_msg *msg" "u_int byte"
.Ft u_int
.Fn uni_msg_trail32 "const struct uni_msg *msg" "int n"
.Ft struct uni_msg *
.Fn uni_msg_dup "const struct uni_msg *msg"
.Sh DESCRIPTION
These functions are used to manipulate variable sized message.
They are
inspired by BSD mbufs and SysV stream buffers, but somewhat simplified because
signalling generally is a low bandwidth task.
All the functions operation on a
.Li uni_msg
data structure:
.Bd -literal -offset indent
struct uni_msg {
	u_char	*b_wptr;	/* tail pointer */
	u_char	*b_rptr;	/* head pointer */
	u_char	*b_buf;		/* data buffer */
	u_char	*b_lim;		/* end of data buffer */
};
.Ed
.Pp
The field
.Fa b_buf
points to the begin of a memory block that is used to store the actual message
and the field
.Fa b_lim
points just to the first byte behind that buffer.
This buffer is allocated
separate from the structure itself and the user calling any of the above
functions with a non const
.Vt struct uni_msg
argument should expect the buffer to be reallocated and hence not hold pointers
into the buffer accross call to these functions.
The pointer
.Fa b_rptr
points to the first used byte in the message and
.Fa b_wptr
to the first unused byte behind all used bytes.
If the message is empty, both pointers point to the same place somewhere in
the allocated buffer.
.Pp
There are several functions and macros that return various sizes and lengths.
The macro
.Fn uni_msg_len
returns the actual size of the message (the number of used bytes).
The macro
.Fn uni_msg_space
returns the number of bytes that are left unused behind the used space.
The macro
.Fn uni_msg_leading
returns the number of bytes that are unused before the used space and the
macro
.Fn uni_msg_size
returns the maximum size of the message (that is the size of the allocated
buffer).
.Pp
Two functions may be used to create new messages: The function
.Fn uni_msg_alloc
allocates the message structure and a buffer to hold at least
.Ar space
bytes (In fact it allocates a couple of bytes more).
If the allocation fails NULL is returned.
The pointers are setup so that there is no leading space in the buffer.
The function
.Fn uni_msg_build
constructs a new message from a variable number of buffers.
The arguments are pairs of
.Vt void *
pointers to buffers and
.Vt size_t
buffer sizes, terminated by a NULL pointer.
The function computes the total resulting message size, allocates a message
and copies all the buffers into the message.
The message is built to have no leading space.
If the allocation fails, NULL is returned.
.Pp
The function
.Fn uni_msg_destroy
deallocates the buffer pointed to by the message and the message itself.
It is save to pass a message with a NULL buffer, but not a NULL message.
.Pp
The function
.Fn uni_msg_dup
returns a copy of a message with exact the same leading space.
.Pp
A number of functions are used to add bytes to an existing message.
The function
.Fn uni_msg_extend
extends the message buffer to have space for at least
.Ar bytes
additional byte at the end.
The leading space does not change.
This function may reallocate the message buffer.
The function returns 0 on success and ENOMEM if the reallocation fails.
In this case the message buffer is not changed.
The macro
.Fn uni_msg_ensure
checks whether the message has space for additional
.Ar bytes
bytes.
If not it calls
.Fn uni_msg_extend
to make the message buffer larger.
The macro returns 0 on success or ENOMEM
if there is not enough space and the reallocation fails.
In this case the message buffer is not changed.
The function
.Fn uni_msg_append
appends
.Ar buflen
bytes from the buffer pointed to by
.Ar buf
to the message.
The function
.Fn uni_msg_append8
appends one byte to the message and the function
.Fn uni_msg_append32
appends a 32-bit value in network byte order to the message
.Fa ( b_wptr
needs not to be aligned).
All three functions call
.Fn uni_msg_ensure
to make sure, that the buffer contents fit into the message.
They return 0 on success and ENOMEM if the buffer is too small and
the reallocation fails.
In this case the message buffer is not changed.
.Pp
A number of functions can be used to retrieve parts of the message.
The function
.Fn uni_msg_strip32
returns the last four bytes of the message as a 32-bit integer assumed to
be in network byte order.
It adjusts
.Fa b_wptr
to remove these four bytes from the message.
.Fa b_wptr
does not need to be aligned.
The function
.Fn uni_msg_get32
returns the first four bytes of the message as a 32-bit integer assumed to
be in network byte order.
It adjusts
.Fa b_rptr
to remove these four bytes from the message.
.Fa b_rptr
does not need to be aligned.
The function
.Fn uni_msg_trail32
returns the
.Fa n 'th
32-bit integer from the buffer counted from the end of the buffer.
The integer is assumed to be in network byte order.
A value of -1 for
.Fa n
returns the last four bytes of the buffer, a value of -2 the four bytes
just before the last four bytes and so on.
All three functions do not check that the message is large enough.
.Sh SEE ALSO
.Xr libunimsg 3 ,
.Xr mbuf 9
.Sh AUTHORS
.An Hartmut Brandt Aq harti@freebsd.org