freebsd-dev/share/man/man3/Q_QADDQ.3
Edward Tomasz Napierala 5a38af1fb8 Introduce <sys/qmath.h>, a fixed-point math library from Netflix.
This makes it possible to perform mathematical operations
on
fractional values without using floating point. It operates on Q
numbers, which are integer-sized, opaque structures initialized
to hold a chosen number of integer and fractional
bits.


For a general description of the Q number system, see the "Fixed Point
Representation & Fractional Math" whitepaper[1]; for the actual
API see the qmath(3) man page.

This is one of dependencies for the upcoming stats(3) framework[2]
that will be applied to the TCP stack in a later commit.

1. https://www.superkits.net/whitepapers/Fixed%20Point%20Representation%20&%20Fractional%20Math.pdf
2. https://reviews.freebsd.org/D20477

Reviewed by:	bcr (man pages, earlier version), sef (earlier version)
Discussed with:	cem, dteske, imp, lstewart
Sponsored By:	Klara Inc, Netflix
Obtained from:	Netflix
Differential Revision:	https://reviews.freebsd.org/D20116
2019-08-27 11:46:22 +00:00

176 lines
4.1 KiB
Groff

.\"
.\" Copyright (c) 2018 Netflix, Inc.
.\" 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 July 8, 2018
.Dt Q_QADDQ 3
.Os
.Sh NAME
.Nm Q_QADDQ ,
.Nm Q_QDIVQ ,
.Nm Q_QMULQ ,
.Nm Q_QSUBQ ,
.Nm Q_NORMPREC ,
.Nm Q_QMAXQ ,
.Nm Q_QMINQ ,
.Nm Q_QCLONEQ ,
.Nm Q_CPYVALQ
.Nd fixed-point math functions which operate on two Q numbers
.Sh SYNOPSIS
.In sys/qmath.h
.Ft int
.Fn Q_QADDQ "QTYPE *a" "QTYPE b"
.Ft int
.Fn Q_QDIVQ "QTYPE *a" "QTYPE b"
.Ft int
.Fn Q_QMULQ "QTYPE *a" "QTYPE b"
.Ft int
.Fn Q_QSUBQ "QTYPE *a" "QTYPE b"
.Ft int
.Fn Q_NORMPREC "QTYPE *a" "QTYPE *b"
.Ft QTYPE
.Fn Q_QMAXQ "QTYPE a" "QTYPE b"
.Ft QTYPE
.Fn Q_QMINQ "QTYPE a" "QTYPE b"
.Ft int
.Fn Q_QCLONEQ "QTYPE *l" "QTYPE r"
.Ft int
.Fn Q_QCPYVALQ "QTYPE *l" "QTYPE r"
.Sh DESCRIPTION
The
.Fn Q_QADDQ ,
.Fn Q_QDIVQ ,
.Fn Q_QMULQ ,
and
.Fn Q_QSUBQ
functions add, divide, multiply or subtract
.Fa b
to/by/from
.Fa a
respectively, storing the result in
.Fa a .
Both arguments must be initialized with the same fractional radix point.
.Pp
The
.Fn Q_NORMPREC
function attempts to normalise the precision of
.Fa a
and
.Fa b
if they differ.
The greater of the two precisions is preferred if possible, unless that would
truncate integer component data for the other operand, in which case the highest
precision that preserves the integer component of both
.Fa a
and
.Fa b
is selected.
.Pp
The
.Fn Q_QMAXQ
and
.Fn Q_QMINQ
functions return the larger or smaller of
.Fa a
and
.Fa b
respectively.
.Pp
The
.Fn Q_QCLONEQ
and
.Fn Q_QCPYVALQ
functions attempt to store identical or representational copies of
.Fa r ,
in
.Fa l
respectively.
An identical Q number produced by cloning copies the control bits as well as the
verbatim integer/fractional bits.
A representational copy only copies the values of
.Fa r Ap s
integer and fractional bits, representing them in the bits available per
.Fa l Ap s
Q format.
.Pp
All of those functions operate on the following data types:
.Vt s8q_t ,
.Vt u8q_t ,
.Vt s16q_t ,
.Vt u16q_t ,
.Vt s32q_t ,
.Vt u32q_t ,
.Vt s64q_t ,
and
.Vt u64q_t ,
which are referred to generically as
.Fa QTYPE .
.Pp
For more details, see
.Xr qmath 3 .
.Sh RETURN VALUES
The
.Fn Q_QADDQ ,
.Fn Q_QDIVQ ,
.Fn Q_QMULQ ,
.Fn Q_QSUBQ
.Fn Q_NORMPREC ,
.Fn Q_QCLONEQ
and
.Fn Q_QCPYVALQ
functions return 0 on success, or an errno on failure.
.Er EINVAL
is returned for divide-by-zero.
.Er EOVERFLOW
and
.Er ERANGE
are returned for overflow and underflow respectively.
.Er ERANGE is also returned when the precision of arguments
does not match.
.Pp
The
.Fn Q_QMAXQ
and
.Fn Q_QMINQ
functions return the numerically larger or smaller of their two inputs
respectively.
.Sh SEE ALSO
.Xr errno 2 ,
.Xr qmath 3 ,
.Xr stdint 7
.Sh HISTORY
The
.Xr qmath 3
functions first appeared in
.Fx 13.0 .
.Sh AUTHORS
.An -nosplit
The
.Xr qmath 3
functions and this manual page were written by
.An Lawrence Stewart Aq Mt lstewart@FreeBSD.org
and sponsored by Netflix, Inc.