mirror of
https://git.hardenedbsd.org/hardenedbsd/HardenedBSD.git
synced 2024-11-28 03:42:04 +01:00
5a38af1fb8
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
176 lines
4.1 KiB
Groff
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.
|