1/* Add two limb vectors of the same length > 0 and store sum in a third
2 limb vector.
3 Copyright (C) 1992-2022 Free Software Foundation, Inc.
4 This file is part of the GNU MP Library.
5
6 The GNU MP Library is free software; you can redistribute it and/or modify
7 it under the terms of the GNU Lesser General Public License as published by
8 the Free Software Foundation; either version 2.1 of the License, or (at your
9 option) any later version.
10
11 The GNU MP Library is distributed in the hope that it will be useful, but
12 WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
13 or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public
14 License for more details.
15
16 You should have received a copy of the GNU Lesser General Public License
17 along with the GNU MP Library; see the file COPYING.LIB. If not,
18 see <https://www.gnu.org/licenses/>. */
19
20#include <sysdep.h>
21#include "asm-syntax.h"
22
23#define PARMS 4+8 /* space for 2 saved regs */
24#define RES PARMS
25#define S1 RES+4
26#define S2 S1+4
27#define SIZE S2+4
28
29 .text
30ENTRY (__mpn_add_n)
31
32 pushl %edi
33 cfi_adjust_cfa_offset (4)
34 pushl %esi
35 cfi_adjust_cfa_offset (4)
36
37 movl RES(%esp),%edi
38 cfi_rel_offset (edi, 4)
39 movl S1(%esp),%esi
40 cfi_rel_offset (esi, 0)
41 movl S2(%esp),%edx
42 movl SIZE(%esp),%ecx
43
44#if IBT_ENABLED
45 pushl %ebx
46 cfi_adjust_cfa_offset (4)
47 cfi_rel_offset (ebx, 0)
48#endif
49
50 movl %ecx,%eax
51 shrl $3,%ecx /* compute count for unrolled loop */
52 negl %eax
53 andl $7,%eax /* get index where to start loop */
54 jz L(oop) /* necessary special case for 0 */
55 incl %ecx /* adjust loop count */
56 shll $2,%eax /* adjustment for pointers... */
57 subl %eax,%edi /* ... since they are offset ... */
58 subl %eax,%esi /* ... by a constant when we ... */
59 subl %eax,%edx /* ... enter the loop */
60 shrl $2,%eax /* restore previous value */
61#if IBT_ENABLED
62 leal -4(,%eax,4),%ebx /* Count for 4-byte endbr32 */
63#endif
64#ifdef PIC
65/* Calculate start address in loop for PIC. Due to limitations in some
66 assemblers, Loop-L0-3 cannot be put into the leal */
67 call L(0)
68 cfi_adjust_cfa_offset (4)
69L(0): leal (%eax,%eax,8),%eax
70 addl (%esp),%eax
71 addl $(L(oop)-L(0)-3),%eax
72 addl $4,%esp
73 cfi_adjust_cfa_offset (-4)
74#else
75/* Calculate start address in loop for non-PIC. */
76 leal (L(oop) - 3)(%eax,%eax,8),%eax
77#endif
78#if IBT_ENABLED
79 addl %ebx,%eax /* Adjust for endbr32 */
80#endif
81 jmp *%eax /* jump into loop */
82 ALIGN (3)
83L(oop): movl (%esi),%eax
84 adcl (%edx),%eax
85 movl %eax,(%edi)
86 _CET_ENDBR
87 movl 4(%esi),%eax
88 adcl 4(%edx),%eax
89 movl %eax,4(%edi)
90 _CET_ENDBR
91 movl 8(%esi),%eax
92 adcl 8(%edx),%eax
93 movl %eax,8(%edi)
94 _CET_ENDBR
95 movl 12(%esi),%eax
96 adcl 12(%edx),%eax
97 movl %eax,12(%edi)
98 _CET_ENDBR
99 movl 16(%esi),%eax
100 adcl 16(%edx),%eax
101 movl %eax,16(%edi)
102 _CET_ENDBR
103 movl 20(%esi),%eax
104 adcl 20(%edx),%eax
105 movl %eax,20(%edi)
106 _CET_ENDBR
107 movl 24(%esi),%eax
108 adcl 24(%edx),%eax
109 movl %eax,24(%edi)
110 _CET_ENDBR
111 movl 28(%esi),%eax
112 adcl 28(%edx),%eax
113 movl %eax,28(%edi)
114 leal 32(%edi),%edi
115 leal 32(%esi),%esi
116 leal 32(%edx),%edx
117 decl %ecx
118 jnz L(oop)
119
120 sbbl %eax,%eax
121 negl %eax
122
123#if IBT_ENABLED
124 popl %ebx
125 cfi_adjust_cfa_offset (-4)
126 cfi_restore (ebx)
127#endif
128 popl %esi
129 cfi_adjust_cfa_offset (-4)
130 cfi_restore (esi)
131 popl %edi
132 cfi_adjust_cfa_offset (-4)
133 cfi_restore (edi)
134
135 ret
136END (__mpn_add_n)
137

source code of glibc/sysdeps/i386/add_n.S