| 1 |
/* Generic unsigned 32 bit division implementation. |
| 2 |
Copyright (C) 2009-2025 Free Software Foundation, Inc. |
| 3 |
Contributed by Embecosm on behalf of Adapteva, Inc. |
| 4 |
|
| 5 |
This file is part of GCC. |
| 6 |
|
| 7 |
This file is free software; you can redistribute it and/or modify it |
| 8 |
under the terms of the GNU General Public License as published by the |
| 9 |
Free Software Foundation; either version 3, or (at your option) any |
| 10 |
later version. |
| 11 |
|
| 12 |
This file is distributed in the hope that it will be useful, but |
| 13 |
WITHOUT ANY WARRANTY; without even the implied warranty of |
| 14 |
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
| 15 |
General Public License for more details. |
| 16 |
|
| 17 |
Under Section 7 of GPL version 3, you are granted additional |
| 18 |
permissions described in the GCC Runtime Library Exception, version |
| 19 |
3.1, as published by the Free Software Foundation. |
| 20 |
|
| 21 |
You should have received a copy of the GNU General Public License and |
| 22 |
a copy of the GCC Runtime Library Exception along with this program; |
| 23 |
see the files COPYING3 and COPYING.RUNTIME respectively. If not, see |
| 24 |
<http://www.gnu.org/licenses/>. */ |
| 25 |
|
| 26 |
typedef union { unsigned int i; float f; } fu; |
| 27 |
|
| 28 |
unsigned int __udivsi3 (unsigned int a, unsigned int b); |
| 29 |
|
| 30 |
unsigned int |
| 31 |
__udivsi3 (unsigned int a, unsigned int b) |
| 32 |
{ |
| 33 |
unsigned int d, t, s0, s1, s2, r0, r1; |
| 34 |
fu u0, u1, u2, u1b, u2b; |
| 35 |
|
| 36 |
if (b > a) |
| 37 |
return 0; |
| 38 |
if ((int) b < 0) |
| 39 |
return 1; |
| 40 |
|
| 41 |
/* Assuming B is nonzero, compute S0 such that 0 <= S0, |
| 42 |
(B << S0+1) does not overflow, |
| 43 |
A < 4.01 * (B << S0), with S0 chosen as small as possible |
| 44 |
without taking to much time calculating. */ |
| 45 |
#ifdef CONVERT_UNSIGNED |
| 46 |
u0.f = a; |
| 47 |
u1.f = b; |
| 48 |
#else /* !CONVERT_UNSIGNED */ |
| 49 |
u0.f = (int) a; |
| 50 |
u1.f = (int) b; |
| 51 |
#ifdef CONCISE |
| 52 |
if ((int) a < 0) |
| 53 |
u0.i = (a >> 8) - 0x00800000 + 0x3f800000 + (31 << 23); |
| 54 |
#else /* To use flag setting / cmove, this can be written as: */ |
| 55 |
{ |
| 56 |
unsigned c = 0xff800000 - 0x4f000000; |
| 57 |
t = (int)a >> 8; |
| 58 |
if (t >= c) |
| 59 |
u0.i = (t - c); |
| 60 |
} |
| 61 |
#endif |
| 62 |
#endif /* !CONVERT_UNSIGNED */ |
| 63 |
s0 = u0.i + 1 /* Compensate for rounding errors. */ |
| 64 |
- 0x00800000 /* adjust by one */ ; |
| 65 |
s0 = s0 - u1.i; |
| 66 |
s0 = (int)s0 >= 0 ? s0 : 0; |
| 67 |
s0 >>= 23; |
| 68 |
|
| 69 |
b <<= s0; |
| 70 |
r1 = 0; |
| 71 |
|
| 72 |
r0 = 1 << s0; |
| 73 |
a = ((t=a) - b); |
| 74 |
if (a <= t) |
| 75 |
{ |
| 76 |
r1 += r0; |
| 77 |
a = ((t=a) - b); |
| 78 |
if (a <= t) |
| 79 |
do { |
| 80 |
r1 += r0; |
| 81 |
a = ((t=a) - b); |
| 82 |
} while (a <= t); |
| 83 |
} |
| 84 |
a += b; |
| 85 |
d = b - 1; |
| 86 |
|
| 87 |
#define STEP(n) case n: a += a; t = a - d; if (t <= a) a = t; |
| 88 |
switch (s0) |
| 89 |
{ |
| 90 |
STEP (31) |
| 91 |
STEP (30) |
| 92 |
STEP (29) |
| 93 |
STEP (28) |
| 94 |
STEP (27) |
| 95 |
STEP (26) |
| 96 |
STEP (25) |
| 97 |
STEP (24) |
| 98 |
STEP (23) |
| 99 |
STEP (22) |
| 100 |
STEP (21) |
| 101 |
STEP (20) |
| 102 |
STEP (19) |
| 103 |
STEP (18) |
| 104 |
STEP (17) |
| 105 |
STEP (16) |
| 106 |
STEP (15) |
| 107 |
STEP (14) |
| 108 |
STEP (13) |
| 109 |
STEP (12) |
| 110 |
STEP (11) |
| 111 |
STEP (10) |
| 112 |
STEP (9) |
| 113 |
STEP (8) |
| 114 |
STEP (7) |
| 115 |
STEP (6) |
| 116 |
STEP (5) |
| 117 |
STEP (4) |
| 118 |
STEP (3) |
| 119 |
STEP (2) |
| 120 |
STEP (1) |
| 121 |
case 0: ; |
| 122 |
} |
| 123 |
r0 = r1 | (r0-1 & a); |
| 124 |
return r0; |
| 125 |
} |