1 | /* ix87 specific implementation of arctanh function. |
2 | Copyright (C) 1996-2022 Free Software Foundation, Inc. |
3 | This file is part of the GNU C Library. |
4 | |
5 | The GNU C Library is free software; you can redistribute it and/or |
6 | modify it under the terms of the GNU Lesser General Public |
7 | License as published by the Free Software Foundation; either |
8 | version 2.1 of the License, or (at your option) any later version. |
9 | |
10 | The GNU C Library is distributed in the hope that it will be useful, |
11 | but WITHOUT ANY WARRANTY; without even the implied warranty of |
12 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
13 | Lesser General Public License for more details. |
14 | |
15 | You should have received a copy of the GNU Lesser General Public |
16 | License along with the GNU C Library; if not, see |
17 | <https://www.gnu.org/licenses/>. */ |
18 | |
19 | #include <machine/asm.h> |
20 | #include <i386-math-asm.h> |
21 | #include <libm-alias-finite.h> |
22 | |
23 | .section .rodata |
24 | |
25 | .align ALIGNARG(4) |
26 | .type half,@object |
27 | half: .double 0.5 |
28 | ASM_SIZE_DIRECTIVE(half) |
29 | .type one,@object |
30 | one: .double 1.0 |
31 | ASM_SIZE_DIRECTIVE(one) |
32 | .type limit,@object |
33 | limit: .double 0.29 |
34 | ASM_SIZE_DIRECTIVE(limit) |
35 | .align ALIGNARG(4) |
36 | .type ln2_2,@object |
37 | ln2_2: .tfloat 0.3465735902799726547086160 |
38 | ASM_SIZE_DIRECTIVE(ln2_2) |
39 | |
40 | DEFINE_FLT_MIN |
41 | |
42 | #ifdef PIC |
43 | #define MO(op) op##@GOTOFF(%edx) |
44 | #else |
45 | #define MO(op) op |
46 | #endif |
47 | |
48 | .text |
49 | ENTRY(__ieee754_atanhf) |
50 | movl 4(%esp), %ecx |
51 | |
52 | movl %ecx, %eax |
53 | andl $0x7fffffff, %eax |
54 | cmpl $0x7f800000, %eax |
55 | ja 5f |
56 | |
57 | #ifdef PIC |
58 | LOAD_PIC_REG (dx) |
59 | #endif |
60 | |
61 | andl $0x80000000, %ecx // ECX == 0 iff X >= 0 |
62 | |
63 | fldt MO(ln2_2) // 0.5*ln2 |
64 | xorl %ecx, 4(%esp) |
65 | flds 4(%esp) // |x| : 0.5*ln2 |
66 | fcoml MO(half) // |x| : 0.5*ln2 |
67 | fld %st(0) // |x| : |x| : 0.5*ln2 |
68 | fnstsw // |x| : |x| : 0.5*ln2 |
69 | sahf |
70 | jae 2f |
71 | fadd %st, %st(1) // |x| : 2*|x| : 0.5*ln2 |
72 | fld %st // |x| : |x| : 2*|x| : 0.5*ln2 |
73 | fsubrl MO(one) // 1-|x| : |x| : 2*|x| : 0.5*ln2 |
74 | fxch // |x| : 1-|x| : 2*|x| : 0.5*ln2 |
75 | fmul %st(2) // 2*|x|^2 : 1-|x| : 2*|x| : 0.5*ln2 |
76 | fdivp // (2*|x|^2)/(1-|x|) : 2*|x| : 0.5*ln2 |
77 | faddp // 2*|x|+(2*|x|^2)/(1-|x|) : 0.5*ln2 |
78 | fcoml MO(limit) // 2*|x|+(2*|x|^2)/(1-|x|) : 0.5*ln2 |
79 | fnstsw // 2*|x|+(2*|x|^2)/(1-|x|) : 0.5*ln2 |
80 | sahf |
81 | jae 4f |
82 | fyl2xp1 // 0.5*ln2*ld(1+2*|x|+(2*|x|^2)/(1-|x|)) |
83 | FLT_CHECK_FORCE_UFLOW_NONNEG |
84 | jecxz 3f |
85 | fchs // 0.5*ln2*ld(1+2*x+(2*x^2)/(1-x)) |
86 | 3: ret |
87 | |
88 | .align ALIGNARG(4) |
89 | 4: faddl MO(one) // 1+2*|x|+(2*|x|^2)/(1-|x|) : 0.5*ln2 |
90 | fyl2x // 0.5*ln2*ld(1+2*|x|+(2*|x|^2)/(1-|x|)) |
91 | jecxz 3f |
92 | fchs // 0.5*ln2*ld(1+2*x+(2*x^2)/(1-x)) |
93 | 3: ret |
94 | |
95 | .align ALIGNARG(4) |
96 | 2: faddl MO(one) // 1+|x| : |x| : 0.5*ln2 |
97 | fxch // |x| : 1+|x| : 0.5*ln2 |
98 | fsubrl MO(one) // 1-|x| : 1+|x| : 0.5*ln2 |
99 | fdivrp // (1+|x|)/(1-|x|) : 0.5*ln2 |
100 | fyl2x // 0.5*ln2*ld((1+|x|)/(1-|x|)) |
101 | jecxz 3f |
102 | fchs // 0.5*ln2*ld((1+x)/(1-x)) |
103 | 3: ret |
104 | |
105 | // x == NaN |
106 | 5: flds 4(%esp) |
107 | ret |
108 | END(__ieee754_atanhf) |
109 | libm_alias_finite (__ieee754_atanhf, __atanhf) |
110 | |