| 1 | /* x86-64 __mpn_lshift -- |
| 2 | Copyright (C) 2007-2021 Free Software Foundation, Inc. |
| 3 | This file is part of the GNU MP Library. |
| 4 | |
| 5 | The GNU MP Library is free software; you can redistribute it and/or modify |
| 6 | it under the terms of the GNU Lesser General Public License as published by |
| 7 | the Free Software Foundation; either version 2.1 of the License, or (at your |
| 8 | option) any later version. |
| 9 | |
| 10 | The GNU MP Library is distributed in the hope that it will be useful, but |
| 11 | WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY |
| 12 | or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public |
| 13 | License for more details. |
| 14 | |
| 15 | You should have received a copy of the GNU Lesser General Public License |
| 16 | along with the GNU MP Library; see the file COPYING.LIB. If not, |
| 17 | see <https://www.gnu.org/licenses/>. */ |
| 18 | |
| 19 | #include "sysdep.h" |
| 20 | #include "asm-syntax.h" |
| 21 | |
| 22 | #define rp %rdi |
| 23 | #define up %rsi |
| 24 | #define n %rdx |
| 25 | #define cnt %cl |
| 26 | |
| 27 | .text |
| 28 | ENTRY (__mpn_lshift) |
| 29 | lea -8(rp,n,8), rp |
| 30 | lea -8(up,n,8), up |
| 31 | |
| 32 | mov %edx, %eax |
| 33 | and $3, %eax |
| 34 | jne L(nb00) |
| 35 | L(b00): /* n = 4, 8, 12, ... */ |
| 36 | mov (up), %r10 |
| 37 | mov -8(up), %r11 |
| 38 | xor %eax, %eax |
| 39 | shld %cl, %r10, %rax |
| 40 | mov -16(up), %r8 |
| 41 | lea 24(rp), rp |
| 42 | sub $4, n |
| 43 | jmp L(00) |
| 44 | |
| 45 | L(nb00):/* n = 1, 5, 9, ... */ |
| 46 | cmp $2, %eax |
| 47 | jae L(nb01) |
| 48 | L(b01): mov (up), %r9 |
| 49 | xor %eax, %eax |
| 50 | shld %cl, %r9, %rax |
| 51 | sub $2, n |
| 52 | jb L(le1) |
| 53 | mov -8(up), %r10 |
| 54 | mov -16(up), %r11 |
| 55 | lea -8(up), up |
| 56 | lea 16(rp), rp |
| 57 | jmp L(01) |
| 58 | L(le1): shl %cl, %r9 |
| 59 | mov %r9, (rp) |
| 60 | ret |
| 61 | |
| 62 | L(nb01):/* n = 2, 6, 10, ... */ |
| 63 | jne L(b11) |
| 64 | L(b10): mov (up), %r8 |
| 65 | mov -8(up), %r9 |
| 66 | xor %eax, %eax |
| 67 | shld %cl, %r8, %rax |
| 68 | sub $3, n |
| 69 | jb L(le2) |
| 70 | mov -16(up), %r10 |
| 71 | lea -16(up), up |
| 72 | lea 8(rp), rp |
| 73 | jmp L(10) |
| 74 | L(le2): shld %cl, %r9, %r8 |
| 75 | mov %r8, (rp) |
| 76 | shl %cl, %r9 |
| 77 | mov %r9, -8(rp) |
| 78 | ret |
| 79 | |
| 80 | .p2align 4 /* performance critical! */ |
| 81 | L(b11): /* n = 3, 7, 11, ... */ |
| 82 | mov (up), %r11 |
| 83 | mov -8(up), %r8 |
| 84 | xor %eax, %eax |
| 85 | shld %cl, %r11, %rax |
| 86 | mov -16(up), %r9 |
| 87 | lea -24(up), up |
| 88 | sub $4, n |
| 89 | jb L(end) |
| 90 | |
| 91 | .p2align 4 |
| 92 | L(top): shld %cl, %r8, %r11 |
| 93 | mov (up), %r10 |
| 94 | mov %r11, (rp) |
| 95 | L(10): shld %cl, %r9, %r8 |
| 96 | mov -8(up), %r11 |
| 97 | mov %r8, -8(rp) |
| 98 | L(01): shld %cl, %r10, %r9 |
| 99 | mov -16(up), %r8 |
| 100 | mov %r9, -16(rp) |
| 101 | L(00): shld %cl, %r11, %r10 |
| 102 | mov -24(up), %r9 |
| 103 | mov %r10, -24(rp) |
| 104 | add $-32, up |
| 105 | lea -32(rp), rp |
| 106 | sub $4, n |
| 107 | jnc L(top) |
| 108 | |
| 109 | L(end): shld %cl, %r8, %r11 |
| 110 | mov %r11, (rp) |
| 111 | shld %cl, %r9, %r8 |
| 112 | mov %r8, -8(rp) |
| 113 | shl %cl, %r9 |
| 114 | mov %r9, -16(rp) |
| 115 | ret |
| 116 | END (__mpn_lshift) |
| 117 | |