1 // SPDX-License-Identifier: GPL-2.0 1 // SPDX-License-Identifier: GPL-2.0
2 /* 2 /*
3 * rational fractions 3 * rational fractions
4 * 4 *
5 * Copyright (C) 2009 emlix GmbH, Oskar Schirm 5 * Copyright (C) 2009 emlix GmbH, Oskar Schirmer <oskar@scara.com>
6 * Copyright (C) 2019 Trent Piepho <tpiepho@gm 6 * Copyright (C) 2019 Trent Piepho <tpiepho@gmail.com>
7 * 7 *
8 * helper functions when coping with rational 8 * helper functions when coping with rational numbers
9 */ 9 */
10 10
11 #include <linux/rational.h> 11 #include <linux/rational.h>
12 #include <linux/compiler.h> 12 #include <linux/compiler.h>
13 #include <linux/export.h> 13 #include <linux/export.h>
14 #include <linux/minmax.h> 14 #include <linux/minmax.h>
15 #include <linux/limits.h> 15 #include <linux/limits.h>
16 #include <linux/module.h> <<
17 16
18 /* 17 /*
19 * calculate best rational approximation for a 18 * calculate best rational approximation for a given fraction
20 * taking into account restricted register siz 19 * taking into account restricted register size, e.g. to find
21 * appropriate values for a pll with 5 bit den 20 * appropriate values for a pll with 5 bit denominator and
22 * 8 bit numerator register fields, trying to 21 * 8 bit numerator register fields, trying to set up with a
23 * frequency ratio of 3.1415, one would say: 22 * frequency ratio of 3.1415, one would say:
24 * 23 *
25 * rational_best_approximation(31415, 10000, 24 * rational_best_approximation(31415, 10000,
26 * (1 << 8) - 1, (1 << 5) - 1, &n 25 * (1 << 8) - 1, (1 << 5) - 1, &n, &d);
27 * 26 *
28 * you may look at given_numerator as a fixed 27 * you may look at given_numerator as a fixed point number,
29 * with the fractional part size described in 28 * with the fractional part size described in given_denominator.
30 * 29 *
31 * for theoretical background, see: 30 * for theoretical background, see:
32 * https://en.wikipedia.org/wiki/Continued_fra 31 * https://en.wikipedia.org/wiki/Continued_fraction
33 */ 32 */
34 33
35 void rational_best_approximation( 34 void rational_best_approximation(
36 unsigned long given_numerator, unsigne 35 unsigned long given_numerator, unsigned long given_denominator,
37 unsigned long max_numerator, unsigned 36 unsigned long max_numerator, unsigned long max_denominator,
38 unsigned long *best_numerator, unsigne 37 unsigned long *best_numerator, unsigned long *best_denominator)
39 { 38 {
40 /* n/d is the starting rational, which 39 /* n/d is the starting rational, which is continually
41 * decreased each iteration using the 40 * decreased each iteration using the Euclidean algorithm.
42 * 41 *
43 * dp is the value of d from the prior 42 * dp is the value of d from the prior iteration.
44 * 43 *
45 * n2/d2, n1/d1, and n0/d0 are our suc 44 * n2/d2, n1/d1, and n0/d0 are our successively more accurate
46 * approximations of the rational. Th 45 * approximations of the rational. They are, respectively,
47 * the current, previous, and two prio 46 * the current, previous, and two prior iterations of it.
48 * 47 *
49 * a is current term of the continued 48 * a is current term of the continued fraction.
50 */ 49 */
51 unsigned long n, d, n0, d0, n1, d1, n2 50 unsigned long n, d, n0, d0, n1, d1, n2, d2;
52 n = given_numerator; 51 n = given_numerator;
53 d = given_denominator; 52 d = given_denominator;
54 n0 = d1 = 0; 53 n0 = d1 = 0;
55 n1 = d0 = 1; 54 n1 = d0 = 1;
56 55
57 for (;;) { 56 for (;;) {
58 unsigned long dp, a; 57 unsigned long dp, a;
59 58
60 if (d == 0) 59 if (d == 0)
61 break; 60 break;
62 /* Find next term in continued 61 /* Find next term in continued fraction, 'a', via
63 * Euclidean algorithm. 62 * Euclidean algorithm.
64 */ 63 */
65 dp = d; 64 dp = d;
66 a = n / d; 65 a = n / d;
67 d = n % d; 66 d = n % d;
68 n = dp; 67 n = dp;
69 68
70 /* Calculate the current ratio 69 /* Calculate the current rational approximation (aka
71 * convergent), n2/d2, using t 70 * convergent), n2/d2, using the term just found and
72 * the two prior approximation 71 * the two prior approximations.
73 */ 72 */
74 n2 = n0 + a * n1; 73 n2 = n0 + a * n1;
75 d2 = d0 + a * d1; 74 d2 = d0 + a * d1;
76 75
77 /* If the current convergent e 76 /* If the current convergent exceeds the maxes, then
78 * return either the previous 77 * return either the previous convergent or the
79 * largest semi-convergent, th 78 * largest semi-convergent, the final term of which is
80 * found below as 't'. 79 * found below as 't'.
81 */ 80 */
82 if ((n2 > max_numerator) || (d 81 if ((n2 > max_numerator) || (d2 > max_denominator)) {
83 unsigned long t = ULON 82 unsigned long t = ULONG_MAX;
84 83
85 if (d1) 84 if (d1)
86 t = (max_denom 85 t = (max_denominator - d0) / d1;
87 if (n1) 86 if (n1)
88 t = min(t, (ma 87 t = min(t, (max_numerator - n0) / n1);
89 88
90 /* This tests if the s 89 /* This tests if the semi-convergent is closer than the previous
91 * convergent. If d1 90 * convergent. If d1 is zero there is no previous convergent as this
92 * is the 1st iteratio 91 * is the 1st iteration, so always choose the semi-convergent.
93 */ 92 */
94 if (!d1 || 2u * t > a 93 if (!d1 || 2u * t > a || (2u * t == a && d0 * dp > d1 * d)) {
95 n1 = n0 + t * 94 n1 = n0 + t * n1;
96 d1 = d0 + t * 95 d1 = d0 + t * d1;
97 } 96 }
98 break; 97 break;
99 } 98 }
100 n0 = n1; 99 n0 = n1;
101 n1 = n2; 100 n1 = n2;
102 d0 = d1; 101 d0 = d1;
103 d1 = d2; 102 d1 = d2;
104 } 103 }
105 *best_numerator = n1; 104 *best_numerator = n1;
106 *best_denominator = d1; 105 *best_denominator = d1;
107 } 106 }
108 107
109 EXPORT_SYMBOL(rational_best_approximation); 108 EXPORT_SYMBOL(rational_best_approximation);
110 <<
111 MODULE_DESCRIPTION("Rational fraction support <<
112 MODULE_LICENSE("GPL v2"); <<
113 109
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.